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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
84b433e4-1706-4f2d-97e4-a2aeeeaf9012 | diffucd-unsupervised-hyperspectral-image | 2305.1241 | null | https://arxiv.org/abs/2305.12410v1 | https://arxiv.org/pdf/2305.12410v1.pdf | DiffUCD:Unsupervised Hyperspectral Image Change Detection with Semantic Correlation Diffusion Model | Hyperspectral image change detection (HSI-CD) has emerged as a crucial research area in remote sensing due to its ability to detect subtle changes on the earth's surface. Recently, diffusional denoising probabilistic models (DDPM) have demonstrated remarkable performance in the generative domain. Apart from their image generation capability, the denoising process in diffusion models can comprehensively account for the semantic correlation of spectral-spatial features in HSI, resulting in the retrieval of semantically relevant features in the original image. In this work, we extend the diffusion model's application to the HSI-CD field and propose a novel unsupervised HSI-CD with semantic correlation diffusion model (DiffUCD). Specifically, the semantic correlation diffusion model (SCDM) leverages abundant unlabeled samples and fully accounts for the semantic correlation of spectral-spatial features, which mitigates pseudo change between multi-temporal images arising from inconsistent imaging conditions. Besides, objects with the same semantic concept at the same spatial location may exhibit inconsistent spectral signatures at different times, resulting in pseudo change. To address this problem, we propose a cross-temporal contrastive learning (CTCL) mechanism that aligns the spectral feature representations of unchanged samples. By doing so, the spectral difference invariant features caused by environmental changes can be obtained. Experiments conducted on three publicly available datasets demonstrate that the proposed method outperforms the other state-of-the-art unsupervised methods in terms of Overall Accuracy (OA), Kappa Coefficient (KC), and F1 scores, achieving improvements of approximately 3.95%, 8.13%, and 4.45%, respectively. Notably, our method can achieve comparable results to those fully supervised methods requiring numerous annotated samples. | ['Licheng Jiao', 'Huiyu Zhou', 'Guanchun Wang', 'Shunli Tian', 'Xiangrong Zhang'] | 2023-05-21 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 5.52568376e-01 -5.40290475e-01 2.48896793e-01 -2.70225942e-01
-8.39555025e-01 -6.62665784e-01 6.85998440e-01 8.22549313e-02
-6.79735392e-02 6.13171697e-01 1.86124906e-01 1.60386980e-01
-5.23253918e-01 -9.69743371e-01 -2.49203116e-01 -1.33523345e+00
1.11323759e-01 -1.25466853e-01 2.07363516e-01 -8.71135890e-02
2.23000839e-04 3.95712614e-01 -1.43782115e+00 -1.15167871e-01
1.35383129e+00 9.87055421e-01 4.60052907e-01 4.58764657e-02
7.66685605e-02 3.55698764e-01 -2.51738071e-01 1.92070976e-01
2.70562857e-01 -4.73540127e-01 -2.66924351e-01 4.43455696e-01
2.58805633e-01 -1.39855146e-01 -3.86372179e-01 1.62281430e+00
4.60821360e-01 2.30074987e-01 7.73630500e-01 -9.04131234e-01
-8.86841655e-01 2.42819190e-01 -8.92781556e-01 1.65002137e-01
-1.24103703e-01 1.27918035e-01 8.29200864e-01 -1.02125823e+00
4.00913239e-01 1.07555628e+00 6.12993002e-01 -6.36766776e-02
-1.31808817e+00 -5.77608526e-01 1.92017823e-01 2.54774570e-01
-1.55599391e+00 -1.54105291e-01 9.09122765e-01 -7.33371556e-01
3.35734129e-01 3.18219304e-01 5.57929277e-01 6.22466981e-01
1.31697237e-01 5.46625614e-01 1.49478805e+00 -1.96351066e-01
3.48918825e-01 -3.24566633e-01 -6.42537028e-02 3.26469392e-01
4.18192595e-01 1.55284822e-01 -5.08013368e-01 5.78395613e-02
4.34686363e-01 2.63702244e-01 -6.44071758e-01 -1.01959027e-01
-9.33177650e-01 6.73985839e-01 6.74061120e-01 3.24846357e-01
-6.27867818e-01 -2.84685940e-01 -3.35214175e-02 -1.22409686e-02
9.26261246e-01 2.03693863e-02 -2.08972678e-01 3.50113511e-01
-9.81265783e-01 -4.04842338e-03 1.01073951e-01 5.64251125e-01
9.31219697e-01 1.77099362e-01 -3.72015387e-01 8.17383468e-01
4.70687091e-01 1.02207100e+00 2.63060838e-01 -8.87293994e-01
1.73558056e-01 5.36011934e-01 1.44170448e-01 -1.20213521e+00
-1.40754119e-01 -7.59249151e-01 -1.34856308e+00 -5.29812500e-02
-1.98427290e-01 8.35720524e-02 -9.31290984e-01 1.73043692e+00
4.22648937e-01 3.02960277e-01 2.57426172e-01 8.75557244e-01
6.31245673e-01 8.56195688e-01 1.29233256e-01 -4.42591250e-01
1.05855680e+00 -5.64036846e-01 -8.10447693e-01 -3.27586770e-01
5.77818267e-02 -6.23467386e-01 7.36349165e-01 1.95126876e-01
-4.90936458e-01 -4.84545678e-01 -1.01123428e+00 4.08056349e-01
-2.19124675e-01 1.77606925e-01 4.24010247e-01 4.93797898e-01
-8.39745700e-01 5.38305163e-01 -7.81269193e-01 -3.53562564e-01
4.95224327e-01 -1.96309254e-01 -9.12872478e-02 -5.76948702e-01
-1.04431188e+00 4.18045104e-01 2.03642860e-01 4.45390642e-01
-9.52208757e-01 -8.38398874e-01 -6.05656445e-01 5.46339080e-02
3.52819741e-01 -3.84490162e-01 6.14712775e-01 -8.80256772e-01
-1.12090981e+00 4.46339935e-01 -3.40921074e-01 -5.97415753e-02
3.92301857e-01 -1.26205971e-02 -7.14937091e-01 3.82523447e-01
4.28729653e-01 3.72567832e-01 9.22102571e-01 -1.54584122e+00
-4.99490023e-01 -6.30275428e-01 -4.39090520e-01 2.63737828e-01
-3.68855506e-01 -3.44871521e-01 -2.04244450e-01 -9.81897116e-01
7.89392769e-01 -9.33477104e-01 -5.05982339e-02 2.60878742e-01
-3.25462341e-01 9.34005529e-02 1.00095558e+00 -8.03715885e-01
1.01452291e+00 -2.43302655e+00 1.56731755e-01 3.48719805e-01
9.83752161e-02 2.82960594e-01 -1.52180225e-01 2.84811288e-01
1.16267830e-01 7.39011765e-02 -9.68423605e-01 7.32303318e-03
-1.34451553e-01 2.15418503e-01 -1.82274878e-01 6.15529954e-01
3.28566283e-01 6.37564898e-01 -1.07315195e+00 -5.10141701e-02
1.55061275e-01 5.10178685e-01 -1.69045985e-01 -3.10888775e-02
-1.08523041e-01 7.66394138e-01 -5.47499776e-01 8.46377134e-01
1.25633466e+00 -1.66744575e-01 2.51278132e-01 -4.69085515e-01
-3.06230992e-01 -3.56285661e-01 -1.25547647e+00 1.49196780e+00
-2.50957981e-02 4.81417060e-01 9.23014209e-02 -8.50648940e-01
9.47565675e-01 1.51420504e-01 5.59962153e-01 -6.67229056e-01
-3.71751249e-01 3.62450093e-01 -2.77128160e-01 -5.16668737e-01
3.01879644e-01 -3.64021420e-01 3.15815806e-01 1.01205811e-01
-2.02391133e-01 -2.54084527e-01 -1.64515540e-01 7.07910880e-02
7.60246277e-01 1.29926652e-02 1.19809173e-01 -3.41808349e-01
5.85231602e-01 -1.22868158e-01 7.23727047e-01 7.32661247e-01
-3.37685615e-01 5.10879576e-01 -1.97596192e-01 5.30194752e-02
-5.72333932e-01 -1.19580650e+00 -2.48745725e-01 4.53905761e-01
5.01577139e-01 1.15793303e-01 -2.55333692e-01 -3.52376044e-01
-2.25548949e-02 6.25562370e-01 -4.69157159e-01 -2.43032634e-01
1.03825651e-01 -1.33400106e+00 2.72886574e-01 1.66851938e-01
1.12985480e+00 -4.93066490e-01 -1.76257193e-02 2.97605664e-01
-4.87715989e-01 -9.55643475e-01 -7.51779377e-02 -5.06580658e-02
-8.26412976e-01 -9.38894451e-01 -6.29550576e-01 -4.22786385e-01
6.65451705e-01 9.41715360e-01 4.72960800e-01 -4.51588959e-01
-1.82292223e-01 3.16562027e-01 -4.98684138e-01 -1.75666854e-01
-2.06909582e-01 -3.64062458e-01 -9.92621332e-02 4.63128477e-01
1.82759061e-01 -6.26774430e-01 -8.43850553e-01 2.88296759e-01
-1.50417614e+00 -1.55272990e-01 5.67729652e-01 8.33983660e-01
7.86286771e-01 7.14714766e-01 5.06897330e-01 -5.85985482e-01
3.45090657e-01 -7.48605669e-01 -5.51301420e-01 3.24170530e-01
-1.01243675e+00 -1.72710508e-01 2.11220101e-01 -2.48888627e-01
-1.64215398e+00 -1.43633960e-02 3.03937107e-01 -1.40700549e-01
-1.08447365e-01 1.04288232e+00 -2.40718290e-01 -1.82663620e-01
6.35298371e-01 6.68300569e-01 -6.56397194e-02 -5.20075083e-01
1.85439125e-01 6.23355269e-01 5.08485377e-01 -3.27536583e-01
1.22561181e+00 9.47325051e-01 5.46330474e-02 -9.40632939e-01
-1.00819170e+00 -7.41917253e-01 -4.08128709e-01 -3.56633931e-01
6.07458472e-01 -1.39323676e+00 1.53043523e-01 9.48205709e-01
-7.40740299e-01 -5.25757670e-02 -1.12782456e-01 6.03744388e-01
8.24097171e-02 7.16080725e-01 -3.44897866e-01 -7.93333411e-01
-2.78057367e-01 -8.88768315e-01 9.94015872e-01 3.97647828e-01
3.48232567e-01 -9.38254774e-01 -4.54924861e-03 3.16970050e-01
6.49200082e-01 3.79433244e-01 9.67820585e-01 -3.11854109e-02
-6.78525686e-01 3.22430693e-02 -4.75759715e-01 6.42330945e-01
6.22921944e-01 -1.01352632e-01 -1.12550056e+00 -3.58483762e-01
2.52327204e-01 1.91000968e-01 1.09354508e+00 6.43831074e-01
9.66729879e-01 -1.29369736e-01 -2.28185520e-01 5.44038177e-01
1.83086479e+00 1.01093054e-01 6.39900386e-01 3.33274722e-01
6.87658072e-01 6.48724616e-01 5.97594261e-01 6.35238051e-01
2.44955391e-01 2.59526968e-01 5.48127949e-01 -1.94420785e-01
-3.14505398e-01 -4.61602211e-02 3.13098997e-01 7.04368114e-01
2.15713140e-02 -2.66773880e-01 -7.60747433e-01 6.78894877e-01
-1.68766820e+00 -1.09999454e+00 -5.03957748e-01 2.08234906e+00
7.29776561e-01 -4.43431526e-01 -6.51411951e-01 7.26168454e-02
9.44514334e-01 5.61993122e-01 -8.23710561e-01 6.28843963e-01
-7.35566199e-01 8.97445604e-02 5.26426315e-01 3.54017019e-01
-1.12853801e+00 7.22651720e-01 5.02856207e+00 7.93294609e-01
-1.11976171e+00 2.87893504e-01 4.94456679e-01 3.15982521e-01
-5.68928480e-01 5.57270348e-02 -3.80077988e-01 4.04193312e-01
3.34045917e-01 -1.22660041e-01 4.64904696e-01 3.00132543e-01
7.67386556e-01 -2.89530754e-01 -3.30273688e-01 8.98186803e-01
3.49095166e-02 -1.00932789e+00 1.69915468e-01 1.15526048e-02
1.35625434e+00 2.20041499e-01 2.52896637e-01 -9.76455361e-02
1.45392939e-01 -5.69240987e-01 6.01604581e-01 8.27642858e-01
7.13854432e-01 -4.40871030e-01 7.41340756e-01 2.44761154e-01
-1.19747162e+00 -8.10466334e-02 -2.76023537e-01 2.23812088e-01
6.78162128e-02 1.29783261e+00 -2.54123241e-01 1.07105517e+00
8.56197834e-01 1.12065709e+00 -5.68513334e-01 1.05901086e+00
-4.70299542e-01 8.24535191e-01 -1.93172663e-01 5.29133499e-01
2.28902400e-01 -6.56975627e-01 7.57731318e-01 8.57103467e-01
8.71516109e-01 3.48260850e-01 6.20001331e-02 9.73541558e-01
-8.52102190e-02 -6.54917136e-02 -3.09372157e-01 1.33461263e-02
6.45311356e-01 1.24489129e+00 -4.81501937e-01 -8.47528577e-02
-2.65268475e-01 1.07441688e+00 -3.55882287e-01 6.88020647e-01
-6.10904038e-01 -1.54855728e-01 6.09035790e-01 -1.75165743e-01
4.53960329e-01 -2.65236139e-01 -1.19764343e-01 -1.14516294e+00
2.32953995e-01 -6.74765587e-01 1.94741145e-01 -1.03833187e+00
-1.64575624e+00 2.67607152e-01 -1.36111602e-01 -1.42259061e+00
4.32870686e-01 -1.59483448e-01 -5.05671144e-01 1.02050447e+00
-2.16350579e+00 -1.23487711e+00 -9.00593698e-01 5.59211195e-01
3.44085842e-01 6.30952343e-02 6.58208966e-01 1.85046285e-01
-6.90690339e-01 -2.16730952e-01 8.24655175e-01 -2.31618419e-01
7.43248045e-01 -9.91406977e-01 -1.80939585e-01 1.39465797e+00
-1.36370987e-01 2.42853805e-01 4.64539826e-01 -7.83552587e-01
-1.07458806e+00 -1.70722342e+00 4.67802435e-01 2.98890054e-01
7.28104413e-01 3.56790006e-01 -1.11008346e+00 1.39857516e-01
-1.79723457e-01 6.88177943e-02 7.31067657e-01 -4.27756667e-01
-4.37527597e-01 -4.96963829e-01 -1.22810030e+00 4.34053868e-01
9.21385288e-01 -6.77147746e-01 -1.94381312e-01 4.60527897e-01
4.45697308e-01 1.28995506e-02 -9.14030373e-01 6.01036072e-01
1.57337815e-01 -8.60293508e-01 8.29630196e-01 1.31659016e-01
3.52913857e-01 -8.80252242e-01 -5.88725567e-01 -1.46790516e+00
-7.06044853e-01 -2.37313196e-01 1.16414577e-01 1.41887188e+00
1.06423236e-01 -7.27549791e-01 1.49782717e-01 2.43605539e-01
-1.69259667e-01 -6.99053183e-02 -8.93574238e-01 -8.87023270e-01
-1.16563343e-01 -3.05857420e-01 5.60719967e-01 1.20580864e+00
-7.19042242e-01 -6.26156479e-02 -1.93597659e-01 9.67090964e-01
9.29821789e-01 2.92696953e-01 1.92412838e-01 -1.53407204e+00
-9.37222615e-02 -3.20523977e-01 -9.99757349e-02 -7.55384505e-01
1.08770221e-01 -9.18584347e-01 1.67906001e-01 -1.64584482e+00
4.33897138e-01 -4.61161643e-01 -3.45396191e-01 4.18040395e-01
-4.50204223e-01 2.93999195e-01 -1.51542082e-01 8.02803874e-01
-1.17393807e-01 9.47515607e-01 1.17711723e+00 -6.51569307e-01
-1.03154927e-01 -1.97458655e-01 -7.17317581e-01 4.97613370e-01
8.18506360e-01 -6.45291507e-01 -3.58925104e-01 -6.31987631e-01
7.51746818e-02 -3.36287290e-01 6.47760093e-01 -1.11514771e+00
2.14210302e-01 -4.07201767e-01 3.77734512e-01 -4.58335042e-01
4.47221920e-02 -7.82868862e-01 4.26473886e-01 4.14983660e-01
5.27517498e-02 -5.43581784e-01 6.82501420e-02 1.08462286e+00
-4.88773912e-01 -1.22409180e-01 8.39188039e-01 2.93417964e-02
-9.25611496e-01 5.22733271e-01 -4.83581781e-01 -2.09234685e-01
8.88721287e-01 -1.62893400e-01 -4.54655737e-01 -2.33114198e-01
-4.87808734e-01 2.62713805e-02 3.91321898e-01 1.92675337e-01
5.65419495e-01 -1.14166284e+00 -9.31000412e-01 1.11818127e-01
4.05664086e-01 1.21359810e-01 7.57834375e-01 9.19535339e-01
-2.66074240e-01 -1.04675591e-02 2.43590072e-01 -7.58526802e-01
-9.37958419e-01 7.98068643e-02 4.38289315e-01 -3.66123058e-02
-3.80582571e-01 6.06896043e-01 1.17077872e-01 -3.37548763e-01
-4.31570500e-01 -2.03341663e-01 4.17251401e-02 2.12329835e-01
3.40511173e-01 4.62123275e-01 1.99388891e-01 -6.60898089e-01
-2.40906104e-01 6.18099809e-01 3.41866583e-01 -2.10389420e-02
1.65358603e+00 -3.77193093e-01 -4.07995582e-01 2.75722712e-01
9.89826024e-01 1.19189397e-01 -1.57173777e+00 -7.62279868e-01
-1.15025476e-01 -5.60609281e-01 6.49453104e-01 -9.59423482e-01
-1.15073836e+00 5.33998907e-01 1.09963858e+00 1.28813088e-01
1.47002077e+00 -1.05654486e-01 4.08350348e-01 1.06220976e-01
1.76849619e-01 -9.67546940e-01 9.70650539e-02 3.67653757e-01
7.31585562e-01 -1.40226460e+00 2.05639303e-01 -5.77083111e-01
-5.36434472e-01 9.64667141e-01 1.15123421e-01 2.92099863e-01
7.23678291e-01 -2.53623366e-01 -2.83145066e-02 -3.69209141e-01
-1.02967925e-01 -3.17423463e-01 3.11249316e-01 6.87525928e-01
7.03315958e-02 3.10797721e-01 -8.81244615e-02 3.00998122e-01
3.68107915e-01 -2.18786061e-01 2.92210519e-01 8.30354035e-01
-5.78659356e-01 -8.00898910e-01 -5.28186262e-01 3.57846975e-01
-7.96380490e-02 -2.57971019e-01 -1.32046506e-01 2.96534836e-01
2.56882936e-01 1.38431931e+00 -2.04417333e-01 -1.91095784e-01
1.12779632e-01 -1.27693012e-01 1.16063952e-01 -4.81039226e-01
1.62971634e-02 3.85758251e-01 -3.97086889e-01 -2.43994728e-01
-1.08469164e+00 -9.56879318e-01 -9.56265628e-01 -2.57154717e-03
-4.08995897e-01 7.63680320e-03 5.59560835e-01 9.74720657e-01
4.92269337e-01 4.78961498e-01 1.02867734e+00 -6.73617303e-01
-5.15678465e-01 -1.05592835e+00 -1.13807607e+00 2.64252722e-01
3.47763389e-01 -7.36157715e-01 -7.41138995e-01 1.63044363e-01] | [10.10805606842041, -1.9497205018997192] |
8eee917f-24ca-451e-bcf9-332242602097 | with-a-little-push-nli-models-can-robustly | 2305.16819 | null | https://arxiv.org/abs/2305.16819v1 | https://arxiv.org/pdf/2305.16819v1.pdf | With a Little Push, NLI Models can Robustly and Efficiently Predict Faithfulness | Conditional language models still generate unfaithful output that is not supported by their input. These unfaithful generations jeopardize trust in real-world applications such as summarization or human-machine interaction, motivating a need for automatic faithfulness metrics. To implement such metrics, NLI models seem attractive, since they solve a strongly related task that comes with a wealth of prior research and data. But recent research suggests that NLI models require costly additional machinery to perform reliably across datasets, e.g., by running inference on a cartesian product of input and generated sentences, or supporting them with a question-generation/answering step. In this work we show that pure NLI models _can_ outperform more complex metrics when combining task-adaptive data augmentation with robust inference procedures. We propose: (1) Augmenting NLI training data to adapt NL inferences to the specificities of faithfulness prediction in dialogue; (2) Making use of both entailment and contradiction probabilities in NLI, and (3) Using Monte-Carlo dropout during inference. Applied to the TRUE benchmark, which combines faithfulness datasets across diverse domains and tasks, our approach strongly improves a vanilla NLI model and significantly outperforms previous work, while showing favourable computational cost. | ['Katja Markert', 'Anette Frank', 'Juri Opitz', 'Julius Steen'] | 2023-05-26 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 4.49966222e-01 6.71559632e-01 -2.22655714e-01 -4.83954608e-01
-9.10410881e-01 -5.81562340e-01 1.19425678e+00 2.77870297e-01
-4.51388806e-01 1.09307623e+00 7.74510741e-01 -4.69319910e-01
7.60466084e-02 -6.69583321e-01 -6.83831275e-01 -2.11682975e-01
3.89121175e-01 8.52417767e-01 1.93474237e-02 -3.94705236e-01
4.24800992e-01 -9.34558213e-02 -1.17380679e+00 2.89218098e-01
1.04688466e+00 6.16334081e-01 -2.15867147e-01 9.29089189e-01
-3.27024125e-02 1.33841860e+00 -7.90229023e-01 -9.00230587e-01
-2.25352064e-01 -8.77798975e-01 -1.33073437e+00 -1.04738384e-01
4.16162014e-01 -3.52653742e-01 2.84049809e-02 8.24472964e-01
4.44442630e-01 2.94152021e-01 6.69689775e-01 -1.00202763e+00
-7.18006849e-01 1.10547256e+00 -6.20710179e-02 9.41313058e-02
7.37767637e-01 4.99231726e-01 1.34443486e+00 -6.68639421e-01
6.21463835e-01 1.40624106e+00 8.06190789e-01 7.90809691e-01
-1.62430847e+00 -1.95750803e-01 -2.08822474e-01 2.95443058e-01
-8.13883960e-01 -9.37454581e-01 6.45789921e-01 -2.41588995e-01
1.33024323e+00 4.83447224e-01 4.11988109e-01 1.53762233e+00
1.25355154e-01 8.04105520e-01 1.15241146e+00 -3.98597151e-01
3.51213664e-01 1.74639881e-01 2.72828579e-01 6.15673125e-01
2.71150082e-01 -1.76646367e-01 -7.69352198e-01 -2.77199984e-01
2.35805228e-01 -4.90595639e-01 -2.87920803e-01 2.26260528e-01
-1.34171307e+00 1.09619379e+00 9.14179161e-02 2.82893896e-01
-3.14411849e-01 4.64009941e-02 4.33111966e-01 6.23439193e-01
5.63820899e-01 1.05946481e+00 -3.02471489e-01 -4.34490085e-01
-1.26124632e+00 4.51730639e-01 1.36116362e+00 7.54651785e-01
5.56800902e-01 -8.11843798e-02 -4.20699298e-01 7.85066426e-01
1.92228273e-01 2.45140046e-01 6.89243674e-01 -1.39392066e+00
4.85789359e-01 4.47145969e-01 2.12677687e-01 -8.42578173e-01
-3.25691193e-01 -1.07723221e-01 -9.63488102e-01 -1.51728332e-01
7.15661168e-01 -1.23767294e-01 -5.42470574e-01 2.12647152e+00
5.24882004e-02 2.18889373e-03 3.27057064e-01 6.66728020e-01
8.18825483e-01 7.64140010e-01 -6.18207715e-02 -3.63572598e-01
1.07411087e+00 -8.05734277e-01 -7.61647761e-01 -5.30650318e-01
7.75956154e-01 -7.25567341e-01 1.46244645e+00 5.56841135e-01
-1.73915505e+00 -2.21758813e-01 -1.03639221e+00 -4.56694484e-01
6.50468003e-03 -3.65607798e-01 7.38359988e-01 5.45819044e-01
-1.16313815e+00 8.23779941e-01 -5.92185020e-01 -1.46784022e-01
3.77127439e-01 1.77302539e-01 -3.67673308e-01 -2.26886302e-01
-1.41156316e+00 1.55860841e+00 3.03654164e-01 1.77941456e-01
-8.38402152e-01 -5.18452048e-01 -9.95852113e-01 3.24544460e-02
5.81694663e-01 -1.09763753e+00 1.47649956e+00 -8.34233701e-01
-1.85054183e+00 7.97943056e-01 2.46969704e-02 -6.96153283e-01
6.66005492e-01 -3.56653810e-01 1.87317222e-01 -3.68446745e-02
3.32637765e-02 5.82806587e-01 7.63134599e-01 -9.15900409e-01
-1.79822847e-01 -1.68457821e-01 3.63810211e-01 1.46247000e-01
-2.36844659e-01 -1.79169118e-01 -4.83492576e-02 -2.14839429e-01
-9.79493782e-02 -8.40988576e-01 -1.51695311e-01 -4.19979960e-01
-7.66569734e-01 -5.36090255e-01 1.23328060e-01 -7.48979092e-01
1.24240363e+00 -1.52836311e+00 4.10922050e-01 -1.13036923e-01
2.14129165e-01 4.32311803e-01 -1.15253448e-01 4.33670878e-01
3.20293546e-01 3.10089767e-01 -4.35978532e-01 -5.81109643e-01
3.67068440e-01 4.34437305e-01 -4.31683689e-01 2.59451091e-01
4.54816341e-01 1.03717172e+00 -8.64988923e-01 -7.29926407e-01
2.16084048e-01 1.01727434e-01 -9.86715376e-01 3.92113090e-01
-5.78432322e-01 2.63594687e-01 -2.91576926e-02 1.93080217e-01
2.44468391e-01 -4.32371706e-01 2.06436217e-01 1.16659135e-01
1.98941499e-01 9.72463369e-01 -9.53009367e-01 1.77482677e+00
-7.04744339e-01 4.66141224e-01 -6.52203858e-02 -1.06263196e+00
6.65754795e-01 3.62266660e-01 -3.77028942e-01 -4.36254323e-01
-5.65686915e-03 1.45023584e-01 2.47896209e-01 -5.41309655e-01
6.93239331e-01 -6.37263119e-01 -3.93538415e-01 7.49278367e-01
4.18092489e-01 -8.03721070e-01 2.33921349e-01 6.29310131e-01
1.03270400e+00 1.27430139e-02 4.38146532e-01 -2.21442774e-01
5.67831159e-01 -6.00696020e-02 4.97168630e-01 1.01009798e+00
-2.95997262e-02 4.46498185e-01 8.87763739e-01 -2.13052168e-01
-1.00562227e+00 -9.34624434e-01 6.25056252e-02 1.12851691e+00
-3.27150971e-01 -2.79625893e-01 -6.99025333e-01 -5.81749976e-01
-2.13595718e-01 1.57156217e+00 -5.74925900e-01 -5.20347774e-01
-5.40673435e-01 -6.91320658e-01 7.21626163e-01 3.52427304e-01
3.50363851e-01 -1.14946663e+00 -5.77710032e-01 3.71221393e-01
-6.02736712e-01 -8.91466141e-01 -1.65252835e-01 2.60216862e-01
-8.58613014e-01 -6.84851468e-01 -1.64096713e-01 -3.46926451e-01
1.56563774e-01 -1.60670981e-01 1.78963077e+00 2.46329173e-01
1.51639968e-01 2.64077634e-01 -1.82126746e-01 -1.51942939e-01
-9.15021539e-01 2.35852882e-01 -3.87092717e-02 -4.41370875e-01
5.21838553e-02 -7.08519816e-01 -3.86753619e-01 1.12727517e-02
-8.73466909e-01 3.43892127e-01 5.05913258e-01 1.17995703e+00
-1.19419672e-01 -5.63901365e-01 9.66460288e-01 -1.14886320e+00
9.47261035e-01 -5.41705012e-01 -1.29732504e-01 3.49484712e-01
-6.51484370e-01 3.70531946e-01 7.18928993e-01 -2.99463362e-01
-1.30334282e+00 -5.98464310e-01 -3.14973235e-01 1.77138090e-01
1.07328594e-01 5.82066000e-01 -1.62153766e-01 4.83270794e-01
8.89752567e-01 3.94765958e-02 2.91511357e-01 -4.78238761e-02
7.96234012e-01 5.22696555e-01 7.24649489e-01 -7.76988208e-01
5.17609894e-01 1.86514363e-01 -1.27205282e-01 -7.18607366e-01
-1.47956061e+00 1.79922041e-02 -3.44321966e-01 6.28544539e-02
5.76767147e-01 -5.72198212e-01 -7.87173152e-01 1.48752630e-01
-1.28829801e+00 -5.45403242e-01 -4.50219005e-01 3.41231108e-01
-7.18262732e-01 5.14593780e-01 -8.98267150e-01 -9.30757523e-01
-5.15466213e-01 -7.86859453e-01 5.72324753e-01 1.03978068e-01
-9.91184652e-01 -1.20114493e+00 2.48945653e-01 7.68109083e-01
6.38849914e-01 1.34822056e-01 1.06376958e+00 -1.04909050e+00
-2.61719316e-01 -2.92209953e-01 4.77270968e-02 7.63542354e-01
-1.94209129e-01 3.82907428e-02 -1.07437479e+00 9.16176736e-02
3.92513514e-01 -1.19488955e+00 9.24107730e-01 1.12848014e-01
7.54958272e-01 -8.23449910e-01 2.83415794e-01 2.49965549e-01
9.69706535e-01 -5.91954172e-01 6.69342935e-01 8.40760171e-02
3.82230222e-01 8.04382145e-01 3.50213647e-01 5.48936367e-01
6.20771885e-01 4.30636913e-01 2.87727058e-01 2.77168870e-01
-1.25755724e-02 -3.94227177e-01 3.45401555e-01 8.90498221e-01
-4.57945429e-02 -2.50861615e-01 -7.02332556e-01 5.53588569e-01
-2.02271008e+00 -1.25617802e+00 -1.86214060e-01 2.18253136e+00
1.61216319e+00 4.47026938e-01 -1.02742396e-01 1.77833989e-01
2.52152830e-01 3.27303320e-01 -4.07716185e-01 -8.29693794e-01
-2.55087018e-01 5.99306703e-01 -1.84745222e-01 1.09752822e+00
-5.83546460e-01 9.80086088e-01 5.98214197e+00 6.83329642e-01
-6.58783078e-01 1.14329301e-01 7.95897841e-01 -3.47020358e-01
-7.99876153e-01 1.91275522e-01 -4.79290098e-01 3.35004091e-01
1.27665710e+00 -3.35312843e-01 5.26475787e-01 4.65450227e-01
1.27103642e-01 -3.79045427e-01 -1.61011481e+00 3.82203400e-01
2.47280627e-01 -1.28704143e+00 1.01639867e-01 -3.32263559e-01
5.97689748e-01 -3.54819536e-01 -1.61665548e-02 5.22805512e-01
6.51753545e-01 -1.30535984e+00 6.83988035e-01 6.21646762e-01
4.27126437e-01 -6.28760278e-01 9.66736555e-01 7.51661241e-01
-1.59541607e-01 1.35997295e-01 -3.57248336e-01 -5.09467661e-01
1.86072007e-01 7.08469689e-01 -9.62731183e-01 2.95992047e-01
7.40424171e-02 4.39925253e-01 -5.16876817e-01 3.49196553e-01
-6.51699245e-01 1.06476104e+00 -3.92192304e-01 -3.19215298e-01
3.10284287e-01 -1.67445406e-01 3.06136906e-01 1.34523714e+00
-3.30366530e-02 5.51952459e-02 -1.67567149e-01 1.14209044e+00
-2.11362660e-01 -1.50749072e-01 -6.66105032e-01 -7.82989785e-02
4.71827120e-01 1.20913696e+00 -1.53072000e-01 -6.51409030e-01
-2.62856364e-01 1.07484674e+00 5.92292428e-01 -4.38455567e-02
-6.38547421e-01 -1.80590436e-01 3.24955404e-01 -2.42789119e-01
-1.74936131e-01 -1.52938992e-01 -7.17559338e-01 -1.14352930e+00
-9.81223360e-02 -8.60688806e-01 3.98098409e-01 -7.71912873e-01
-1.31080401e+00 2.28687912e-01 -1.89140383e-02 -4.80666697e-01
-9.89135146e-01 -1.65004462e-01 -7.55032957e-01 8.87964964e-01
-1.53222442e+00 -1.15007484e+00 1.38259493e-02 4.93881404e-01
4.44146216e-01 2.90079743e-01 9.80240583e-01 -3.05412203e-01
-5.38485885e-01 5.18835008e-01 -3.86912316e-01 -2.51011431e-01
7.67418563e-01 -1.44042003e+00 5.15083492e-01 8.18003953e-01
2.07067907e-01 8.51060033e-01 1.03927302e+00 -3.87143701e-01
-1.40620446e+00 -7.24457204e-01 1.68909466e+00 -9.43632483e-01
8.72960269e-01 -3.26112032e-01 -1.13176739e+00 8.37625742e-01
5.79129696e-01 -4.00540471e-01 8.08009088e-01 4.08338904e-01
-4.27128255e-01 2.33677909e-01 -1.09731019e+00 7.73965776e-01
9.56953406e-01 -5.65299094e-01 -1.29643166e+00 5.48990071e-01
6.69438541e-01 -2.19540447e-01 -6.86367154e-01 7.46768862e-02
3.04536998e-01 -1.12078166e+00 7.02101350e-01 -7.79538631e-01
9.93777335e-01 1.34662926e-01 -1.03755973e-01 -1.35253882e+00
-1.72238708e-01 -1.02785337e+00 -1.72589183e-01 1.31866074e+00
6.61024213e-01 -5.55617392e-01 4.48999345e-01 1.09733379e+00
-2.39384294e-01 -6.28229201e-01 -1.03305089e+00 -4.38354999e-01
4.00830358e-01 -5.26408374e-01 1.44812852e-01 9.00550306e-01
4.95038182e-01 9.98133600e-01 -5.67197323e-01 -3.26320738e-01
5.28656244e-01 1.08477607e-01 8.38077605e-01 -1.17549932e+00
-4.91855681e-01 -5.20575881e-01 1.88199177e-01 -9.18044686e-01
5.30102611e-01 -1.06149948e+00 2.32498094e-01 -1.62328792e+00
4.03125316e-01 -2.98240725e-02 2.54963666e-01 6.22022927e-01
-4.66194838e-01 2.93511957e-01 5.04714660e-02 1.78778842e-01
-7.95092463e-01 6.38480484e-01 9.59030926e-01 4.60779853e-03
1.20945536e-02 -7.41334027e-03 -1.10826695e+00 9.45593715e-01
7.52929688e-01 -2.38735467e-01 -4.08327371e-01 -3.50521296e-01
7.52402544e-01 3.00278813e-01 5.49219906e-01 -6.64854527e-01
2.65294403e-01 -8.59486908e-02 1.41080767e-01 -1.33187383e-01
4.27462429e-01 -1.85590610e-01 -2.84626216e-01 3.87122244e-01
-8.40655923e-01 -1.24333210e-01 -4.93070409e-02 4.27521765e-01
-3.53618078e-02 -5.53117275e-01 8.51786077e-01 -4.18396950e-01
-8.24257284e-02 -3.90426934e-01 -4.12632465e-01 7.02426314e-01
4.61248875e-01 -5.31985052e-02 -4.53675389e-01 -6.67711437e-01
-6.10819757e-01 1.42491385e-01 5.10492563e-01 -5.90193160e-02
6.20140433e-01 -9.73757625e-01 -1.12342477e+00 -9.27489903e-03
-1.40351161e-01 6.60745427e-02 7.59729147e-02 9.22937334e-01
-2.25318536e-01 2.90264666e-01 -2.55041700e-02 -4.23135072e-01
-9.26518798e-01 2.44472936e-01 2.59176176e-02 -5.91650963e-01
-4.63849097e-01 1.09407687e+00 -1.80808052e-01 -5.74307978e-01
1.57758896e-03 -4.38875347e-01 -4.99862768e-02 1.11535102e-01
3.37204158e-01 3.16100001e-01 -8.85594077e-03 -1.55496851e-01
-3.22725624e-01 -1.52200475e-01 -1.96252689e-01 -5.30989528e-01
1.30726087e+00 3.45688127e-02 -2.90651113e-01 6.34448349e-01
8.10344756e-01 -1.50555909e-01 -1.12114298e+00 -3.11405331e-01
2.06994355e-01 -2.40069911e-01 -1.41487509e-01 -1.10146642e+00
-3.37471992e-01 1.00984144e+00 -4.70939964e-01 3.38940889e-01
6.55539691e-01 8.01005140e-02 7.23879278e-01 6.54307723e-01
1.37941197e-01 -1.27165532e+00 3.36336344e-01 7.09349692e-01
1.03425336e+00 -1.26815701e+00 1.15989335e-01 1.98150218e-01
-9.49391544e-01 8.70132625e-01 3.83590817e-01 -1.19854592e-01
5.01445383e-02 9.74107906e-02 -2.15089396e-01 -5.90079688e-02
-1.44539857e+00 1.37021363e-01 1.14657208e-01 3.30867827e-01
7.59029746e-01 9.78196189e-02 -5.01075685e-01 8.34749997e-01
-6.36150837e-01 -5.08240946e-02 8.71219337e-01 5.96514463e-01
-5.12954712e-01 -7.88289845e-01 -4.98362407e-02 6.61705792e-01
-4.41900194e-01 -5.05321681e-01 -6.78398609e-01 6.19493186e-01
-2.79492885e-01 1.42617655e+00 -1.20332949e-01 -1.01870574e-01
1.86823174e-01 3.96202445e-01 5.94940603e-01 -7.22100317e-01
-8.53808224e-01 -4.72195834e-01 7.73221612e-01 -5.11396289e-01
-3.19810212e-01 -7.42047966e-01 -1.00594175e+00 -7.51223326e-01
-3.26781601e-01 2.35633910e-01 4.20671225e-01 1.19304967e+00
2.25047365e-01 1.49497151e-01 4.03860807e-01 -9.29108620e-01
-1.32170391e+00 -1.41904676e+00 -1.63406208e-01 6.30667984e-01
2.17260331e-01 -2.28819698e-01 -7.13862181e-01 -3.95246260e-02] | [12.214714050292969, 8.916519165039062] |
f0af3240-8e11-4dce-976e-e1c6162ad04c | labrad-or-lightweight-memory-scene-graphs-for | 2303.13293 | null | https://arxiv.org/abs/2303.13293v1 | https://arxiv.org/pdf/2303.13293v1.pdf | LABRAD-OR: Lightweight Memory Scene Graphs for Accurate Bimodal Reasoning in Dynamic Operating Rooms | Modern surgeries are performed in complex and dynamic settings, including ever-changing interactions between medical staff, patients, and equipment. The holistic modeling of the operating room (OR) is, therefore, a challenging but essential task, with the potential to optimize the performance of surgical teams and aid in developing new surgical technologies to improve patient outcomes. The holistic representation of surgical scenes as semantic scene graphs (SGG), where entities are represented as nodes and relations between them as edges, is a promising direction for fine-grained semantic OR understanding. We propose, for the first time, the use of temporal information for more accurate and consistent holistic OR modeling. Specifically, we introduce memory scene graphs, where the scene graphs of previous time steps act as the temporal representation guiding the current prediction. We design an end-to-end architecture that intelligently fuses the temporal information of our lightweight memory scene graphs with the visual information from point clouds and images. We evaluate our method on the 4D-OR dataset and demonstrate that integrating temporality leads to more accurate and consistent results achieving an +5% increase and a new SOTA of 0.88 in macro F1. This work opens the path for representing the entire surgery history with memory scene graphs and improves the holistic understanding in the OR. Introducing scene graphs as memory representations can offer a valuable tool for many temporal understanding tasks. | ['Nassir Navab', 'Chantal Pellegrini', 'Felix Holm', 'Tobias Czempiel', 'Ege Özsoy'] | 2023-03-23 | null | null | null | null | ['scene-graph-generation'] | ['computer-vision'] | [ 1.15402237e-01 1.22730307e-01 -2.26769179e-01 -3.03208113e-01
-2.83448517e-01 -3.88056636e-01 4.51518416e-01 7.25955248e-01
-2.50259391e-03 7.30633661e-02 4.76165891e-01 -4.54637945e-01
-5.25646985e-01 -7.93289602e-01 -5.07191658e-01 -5.80877244e-01
-1.80815876e-01 4.50027823e-01 2.92945921e-01 -1.45729095e-01
2.31311902e-01 7.69950271e-01 -1.45601761e+00 5.88453114e-01
5.41341007e-01 1.03859437e+00 5.08201361e-01 5.93391120e-01
-2.78083444e-01 9.32220340e-01 -2.35801503e-01 -6.10827021e-02
8.07206053e-03 -5.94457053e-02 -8.49188685e-01 5.20188501e-03
2.72655070e-01 3.85991111e-02 -5.37377298e-01 6.46365881e-01
3.63171786e-01 1.36427924e-01 -7.69809727e-03 -7.96780229e-01
-9.75909755e-02 2.87136883e-01 -3.24464291e-01 1.92374080e-01
2.51039654e-01 9.84228253e-02 1.86177567e-01 -6.85473084e-01
9.31144595e-01 1.16088152e+00 4.39568043e-01 3.77564400e-01
-9.87709939e-01 -3.92876655e-01 4.52099234e-01 3.31466109e-01
-1.08001184e+00 -1.29681945e-01 7.22422123e-01 -5.84379673e-01
1.02775049e+00 3.80155265e-01 1.09460366e+00 8.73989642e-01
8.79143178e-01 5.39324641e-01 7.16314614e-01 -2.44487107e-01
7.06686005e-02 -7.88516924e-02 2.81686515e-01 1.01574385e+00
9.69769210e-02 2.94815421e-01 -9.29455400e-01 1.15152903e-01
9.49391425e-01 7.62852311e-01 -2.68000126e-01 -6.21027589e-01
-1.48649335e+00 4.23536807e-01 9.56781566e-01 3.60017747e-01
-5.10650218e-01 4.47731376e-01 4.10876989e-01 -6.82147369e-02
2.65634716e-01 4.67465311e-01 -2.02207819e-01 -2.35211655e-01
-6.78341925e-01 -2.17769012e-01 5.27084768e-01 1.02545607e+00
3.35817009e-01 -2.88778454e-01 -1.33049428e-01 5.31968892e-01
-1.09447680e-01 5.95999025e-02 2.95360714e-01 -6.86335444e-01
5.08426316e-02 1.09231758e+00 -6.61519021e-02 -1.09894240e+00
-1.03456879e+00 -8.35051298e-01 -7.86142826e-01 5.15193678e-02
-2.33325720e-01 6.03305697e-01 -1.30808973e+00 1.35885835e+00
3.85966957e-01 3.81925434e-01 -1.47777230e-01 5.67137480e-01
9.86936927e-01 3.15148860e-01 3.86271596e-01 -1.68088049e-01
1.49444246e+00 -1.07300472e+00 -8.11347544e-01 -3.35741013e-01
9.48075950e-01 -6.44946873e-01 7.47644007e-01 2.19332397e-01
-1.05459070e+00 -4.62630689e-01 -8.25869739e-01 -8.87228549e-02
-4.66070175e-01 3.71429585e-02 9.54923511e-01 7.26738796e-02
-1.01081586e+00 5.81362665e-01 -1.52780521e+00 -4.17123705e-01
5.09667933e-01 3.71904284e-01 -6.17218375e-01 -4.82226074e-01
-4.88871336e-01 9.95095909e-01 3.17552686e-01 3.11245233e-01
-8.44232082e-01 -1.07052207e+00 -1.24440157e+00 -3.18754371e-03
3.79976690e-01 -1.05255771e+00 9.31438208e-01 -2.51832724e-01
-8.20774615e-01 9.60565090e-01 -2.75583297e-01 -3.49493086e-01
2.18842715e-01 -2.97003627e-01 -2.85977334e-01 7.90061206e-02
-2.87495553e-01 3.54876071e-01 3.64846528e-01 -1.33817589e+00
-2.95494825e-01 -6.83311045e-01 1.14256732e-01 2.61850595e-01
-9.10313874e-02 -4.52770293e-01 -7.32832968e-01 -7.19105780e-01
5.13917267e-01 -1.21709049e+00 -5.80394745e-01 4.32314783e-01
-1.80760741e-01 3.13195258e-01 7.87210941e-01 -7.85956383e-01
1.52657485e+00 -2.50324774e+00 3.88854057e-01 1.57330006e-01
4.98917669e-01 -9.00451615e-02 2.12862581e-01 5.35968184e-01
-2.62434900e-01 -1.38375327e-01 -3.28614586e-03 -5.23508847e-01
-5.12393892e-01 4.91100967e-01 -2.03551859e-01 2.77618706e-01
-2.33730912e-01 9.23608840e-01 -1.17486811e+00 -4.73841429e-01
5.89527249e-01 6.52505755e-01 -5.21011055e-01 2.80893117e-01
8.98749903e-02 7.11841702e-01 -3.69913220e-01 5.75232506e-01
2.96304971e-01 -6.14161491e-01 3.36936027e-01 -4.16061401e-01
-4.87415604e-02 2.22010121e-01 -6.46642685e-01 2.44295239e+00
-8.74529362e-01 4.88628477e-01 -1.54174998e-01 -5.45152903e-01
8.07520807e-01 1.56112671e-01 7.61651039e-01 -9.39535022e-01
1.05831355e-01 4.49555963e-02 -1.78483561e-01 -3.18880945e-01
4.70811099e-01 -3.72931242e-01 -1.67104319e-01 -5.15350774e-02
-1.11241087e-01 -2.67470360e-01 -8.21461231e-02 4.14579123e-01
1.06980908e+00 2.84034386e-02 4.48555797e-01 -1.76704764e-01
5.90421781e-02 4.56496805e-01 3.74717802e-01 3.93534452e-01
5.77199571e-02 4.50209022e-01 2.57845730e-01 -9.37497318e-01
-4.88739729e-01 -1.35655379e+00 -4.71125878e-02 4.99138683e-01
7.43926644e-01 -6.63551629e-01 -1.93878889e-01 -5.11649072e-01
-4.30389587e-03 7.58528709e-01 -8.66742074e-01 -7.58246064e-01
-6.96240664e-01 -2.56308228e-01 -4.26401943e-01 1.00621331e+00
-5.27576655e-02 -6.80372953e-01 -8.83512974e-01 3.45071912e-01
-1.53074622e-01 -1.12316000e+00 -3.32629114e-01 1.24900512e-01
-1.47829473e+00 -1.16870677e+00 -2.18118072e-01 -6.53983474e-01
8.96123409e-01 5.46465456e-01 1.16271901e+00 2.83620894e-01
-9.26492631e-01 7.24007368e-01 -2.93324322e-01 -4.86547083e-01
-2.62302428e-01 -3.97238553e-01 -1.84058934e-01 -3.51626545e-01
-2.00285554e-01 -4.20771867e-01 -1.06191957e+00 2.06278875e-01
-8.02117050e-01 8.31457436e-01 4.87597167e-01 7.09425688e-01
9.07772899e-01 -1.60229266e-01 -1.02458291e-01 -1.10315263e+00
-8.25193375e-02 -5.21112084e-01 -4.03360516e-01 3.52578104e-01
-5.29526055e-01 1.36943460e-01 4.23687309e-01 -1.23716235e-01
-9.19002593e-01 -5.96359791e-03 3.39747488e-01 -8.45372319e-01
1.91161782e-01 6.31922185e-01 3.48652810e-01 -8.27961713e-02
4.87206697e-01 8.30591365e-04 7.78639466e-02 -4.83514190e-01
3.58959019e-01 -2.15187326e-01 6.63359404e-01 -3.87204319e-01
4.05438602e-01 8.31536114e-01 4.34676021e-01 -3.72038573e-01
-8.11271846e-01 -8.09423149e-01 -6.43198907e-01 -4.90460306e-01
8.29226613e-01 -9.14308548e-01 -6.65493667e-01 3.22439298e-02
-1.14242661e+00 -3.35664839e-01 -2.69847900e-01 5.43742657e-01
-6.03808701e-01 3.83960493e-02 -4.72426981e-01 -4.37175304e-01
-2.87335455e-01 -1.27610624e+00 1.52701473e+00 1.19369619e-01
-3.36138457e-01 -1.26498580e+00 -7.54731447e-02 3.84974629e-01
4.55189437e-01 7.89877474e-01 1.14181769e+00 -1.27441928e-01
-1.03450286e+00 -3.71643305e-01 -1.25865340e-01 -2.27299109e-01
2.93055654e-01 -2.06494257e-01 -5.73762834e-01 -3.64018023e-01
-8.09749514e-02 5.08853376e-01 7.18926430e-01 3.45221162e-01
1.54533863e+00 -3.77513617e-02 -9.44512844e-01 7.56087065e-01
1.60280573e+00 4.27496850e-01 4.34403360e-01 8.23690221e-02
8.23574662e-01 6.45326138e-01 9.50273454e-01 6.95142031e-01
4.51552331e-01 6.99535787e-01 7.65904069e-01 -2.55093962e-01
-4.42106754e-01 -1.34911224e-01 -1.79654285e-01 9.38823879e-01
-7.83207715e-02 9.60715022e-03 -1.35303783e+00 6.79675043e-01
-2.07045984e+00 -7.05697060e-01 -3.89666297e-02 2.19428086e+00
3.58531266e-01 1.28307743e-02 -5.59207320e-01 -2.75761634e-01
3.11006904e-01 1.75163329e-01 -3.47855330e-01 -2.63154030e-01
4.51164752e-01 4.08496708e-01 3.54118884e-01 3.30390960e-01
-8.84213328e-01 7.81740069e-01 5.40263414e+00 3.79345328e-01
-1.31500137e+00 6.42822981e-02 4.52551395e-01 -5.77794135e-01
-1.42822236e-01 -5.42084947e-02 -3.35022360e-01 5.49174435e-02
7.81234980e-01 -4.40753251e-01 2.57587790e-01 9.20218170e-01
2.14415565e-01 -1.73121095e-01 -1.40633559e+00 1.27280331e+00
3.79561782e-02 -1.90270042e+00 2.57045358e-01 -3.69177535e-02
5.78989029e-01 -2.83621222e-01 1.77851453e-01 -2.57711131e-02
1.31309271e-01 -1.09549701e+00 6.28974438e-01 1.02773488e+00
6.60579979e-01 -4.72818166e-01 6.86009288e-01 1.61173478e-01
-1.53716183e+00 -1.47249952e-01 1.52124330e-01 3.86006147e-01
2.72744060e-01 4.21080709e-01 -1.07656431e+00 9.67654467e-01
7.54607737e-01 1.09523475e+00 -3.93157393e-01 1.17426074e+00
3.09955984e-01 -8.96020532e-02 -2.05937959e-02 4.66238916e-01
1.23275563e-01 2.73510218e-01 4.47413385e-01 1.04613853e+00
4.16081607e-01 2.55937159e-01 2.49023989e-01 4.10661250e-01
1.67154506e-01 -8.92907530e-02 -8.40968668e-01 -2.35309321e-02
1.56907633e-01 1.13626528e+00 -8.91439557e-01 -2.30753899e-01
-2.05440015e-01 8.13563704e-01 2.87778914e-01 2.01326579e-01
-7.89942205e-01 1.63379997e-01 9.19858932e-01 3.18826884e-01
-2.39891559e-01 -7.15292454e-01 -6.45718098e-01 -8.04682672e-01
4.36532609e-02 -3.80052447e-01 7.51694977e-01 -8.48242819e-01
-6.64413571e-01 7.06327319e-01 -8.16305280e-02 -1.65253711e+00
-3.22684385e-02 -7.95598626e-01 -4.64898318e-01 6.51699126e-01
-1.45752037e+00 -1.25289464e+00 -1.17428887e+00 6.43718243e-01
6.50410593e-01 3.81852359e-01 1.05625999e+00 1.07543036e-01
-2.47995332e-01 -5.45770116e-02 -2.98949599e-01 -1.87366575e-01
4.89231080e-01 -1.00323474e+00 3.06161493e-02 5.63250542e-01
-3.03312410e-02 9.18022752e-01 6.28445268e-01 -6.16043091e-01
-1.69263697e+00 -1.20012724e+00 4.78253961e-01 -6.93113685e-01
4.00063574e-01 -2.19041750e-01 -7.27970541e-01 7.73203254e-01
-3.56484860e-01 1.54420674e-01 8.66693437e-01 1.01009227e-01
-2.09337994e-01 -1.60626113e-01 -8.12283099e-01 7.76975691e-01
1.47618330e+00 -4.98733073e-01 -3.03585738e-01 5.31409562e-01
1.11689055e+00 -1.02106631e+00 -1.21584535e+00 8.25511575e-01
4.89091694e-01 -8.91126275e-01 1.09021568e+00 -5.62707603e-01
3.78872514e-01 -2.30324253e-01 -2.89118383e-02 -1.13815522e+00
-1.60467789e-01 -2.33938321e-01 -1.40046060e-01 3.20551783e-01
8.97509083e-02 -4.03087616e-01 6.74958408e-01 8.14809740e-01
-7.59909928e-01 -1.12092042e+00 -7.59514809e-01 -5.54210365e-01
-5.94424725e-01 -5.89391172e-01 3.54317605e-01 7.61283517e-01
8.71779397e-02 -3.70335311e-01 6.04012236e-02 3.85278761e-01
3.99154067e-01 5.58928907e-01 4.72921044e-01 -1.17758405e+00
-6.02347776e-02 -5.08358240e-01 -7.89216816e-01 -6.22580707e-01
-3.33668053e-01 -1.00774193e+00 -1.32264271e-01 -2.17864752e+00
1.39772862e-01 -6.15053773e-01 -5.28394461e-01 5.29626548e-01
4.10218947e-02 -1.73415020e-01 3.05170178e-01 5.13466448e-02
-5.25577128e-01 4.19427544e-01 1.51836908e+00 -4.17583659e-02
-1.98554575e-01 -1.61804467e-01 -4.39078301e-01 5.87228537e-01
3.21942985e-01 -3.85885119e-01 -5.75056612e-01 -5.97454011e-01
5.07360362e-02 4.83128607e-01 4.91818070e-01 -1.01549125e+00
7.27133572e-01 -1.38225898e-01 1.27947107e-01 -6.26999140e-01
6.53658032e-01 -1.19744146e+00 7.19715059e-01 7.74913549e-01
1.23146281e-01 2.42589802e-01 6.82384670e-01 8.50382984e-01
-5.00516951e-01 5.89394689e-01 6.31390274e-01 -2.85495758e-01
-1.07417774e+00 5.81716359e-01 2.78777629e-01 -3.99237454e-01
1.32319295e+00 -3.33828777e-01 -4.30186987e-01 4.77430299e-02
-1.29801595e+00 3.59113783e-01 5.61279595e-01 8.63946378e-01
1.07618368e+00 -1.08948636e+00 -1.93844154e-01 2.67971426e-01
4.78629768e-01 4.73277867e-01 1.03447616e+00 1.11287665e+00
-8.65173161e-01 4.52017426e-01 -1.65437490e-01 -9.12481606e-01
-1.58493972e+00 8.41662943e-01 3.06302786e-01 -5.29484332e-01
-9.63815451e-01 7.54721344e-01 8.48812997e-01 4.30549355e-03
2.71911412e-01 -4.89625245e-01 -6.93676472e-02 -1.62766889e-01
4.82427090e-01 1.39699847e-01 2.68379837e-01 -2.91516840e-01
-5.59479773e-01 7.02754259e-01 -1.49335802e-01 4.28587347e-01
1.46145189e+00 7.37115890e-02 -1.97563648e-01 6.27071083e-01
9.89766717e-01 -2.94590086e-01 -9.35560405e-01 -1.99615121e-01
-4.69033513e-03 -5.69281518e-01 1.48442760e-01 -8.98032784e-01
-1.10612857e+00 9.59794462e-01 6.13664389e-01 -2.07786813e-01
1.32101989e+00 3.36161971e-01 7.38944292e-01 -6.09084666e-02
9.45141852e-01 -4.36212212e-01 3.48696202e-01 1.86568543e-01
1.23593080e+00 -1.04204214e+00 -2.41109333e-03 -8.49386513e-01
-6.17840111e-01 1.25956964e+00 5.30621231e-01 1.46751434e-01
8.42684269e-01 3.01927328e-01 9.78397578e-03 -7.44604826e-01
-9.29577649e-01 6.22685812e-02 6.18658364e-01 2.34849960e-01
2.95753628e-01 3.07862163e-01 1.42690465e-01 4.03296381e-01
-1.03877760e-01 -2.52694637e-02 1.11773133e-01 1.17350626e+00
-6.55103996e-02 -8.97994697e-01 6.23548180e-02 3.96409631e-01
-3.23460996e-02 -1.39257953e-01 1.03964120e-01 8.18915009e-01
-1.08958557e-01 5.76911569e-01 1.94524229e-01 -4.94943291e-01
7.39101231e-01 -7.35710785e-02 4.63056147e-01 -9.29202199e-01
-6.09910250e-01 -2.41866156e-01 4.16187271e-02 -1.34440506e+00
-2.36218259e-01 -2.90042847e-01 -1.66519928e+00 -2.52124548e-01
1.76309437e-01 -2.96200067e-01 1.01950181e+00 6.05665386e-01
7.42026150e-01 1.24716020e+00 3.30998927e-01 -8.52091074e-01
2.42869213e-01 -2.85813332e-01 -2.82969475e-01 4.45368886e-01
3.85598242e-01 -9.49045181e-01 8.48953128e-02 2.02858876e-02] | [14.064148902893066, -3.353748321533203] |
c1bbeb75-773a-4e31-aac7-2d7cd53bf63d | towards-controllable-and-personalized-review | 1910.03506 | null | https://arxiv.org/abs/1910.03506v2 | https://arxiv.org/pdf/1910.03506v2.pdf | Towards Controllable and Personalized Review Generation | In this paper, we propose a novel model RevGAN that automatically generates controllable and personalized user reviews based on the arbitrarily given sentimental and stylistic information. RevGAN utilizes the combination of three novel components, including self-attentive recursive autoencoders, conditional discriminators, and personalized decoders. We test its performance on the several real-world datasets, where our model significantly outperforms state-of-the-art generation models in terms of sentence quality, coherence, personalization and human evaluations. We also empirically show that the generated reviews could not be easily distinguished from the organically produced reviews and that they follow the same statistical linguistics laws. | ['Alexander Tuzhilin', 'Pan Li'] | 2019-09-30 | towards-controllable-and-personalized-review-1 | https://aclanthology.org/D19-1319 | https://aclanthology.org/D19-1319.pdf | ijcnlp-2019-11 | ['review-generation'] | ['natural-language-processing'] | [-2.65586853e-01 5.46556890e-01 -3.44939642e-02 -5.62177181e-01
-6.15245998e-01 -5.79548776e-01 8.28770697e-01 -3.48734915e-01
-1.29216775e-01 1.20337188e+00 5.63008308e-01 4.84219342e-02
4.97793287e-01 -9.13834810e-01 -5.05688429e-01 -3.49301726e-01
3.40484262e-01 5.76127827e-01 -3.72910678e-01 -8.14370871e-01
3.12778801e-01 -4.06436265e-01 -1.37949359e+00 1.29810214e-01
1.32683396e+00 5.55407941e-01 7.49589503e-02 9.90225077e-01
7.00358972e-02 7.27971554e-01 -9.75634694e-01 -1.05590582e+00
-1.11609392e-01 -9.32508707e-01 -5.27071893e-01 -4.99772392e-02
-8.91928747e-02 -2.76001602e-01 -1.73081800e-01 9.99790132e-01
7.34828115e-01 5.02549054e-04 8.39249253e-01 -6.21462345e-01
-1.99754691e+00 1.34248686e+00 -1.81408450e-01 -2.93703437e-01
2.89362490e-01 3.63471746e-01 1.29736984e+00 -6.69030130e-01
7.32849181e-01 1.26268041e+00 4.12771910e-01 1.08976638e+00
-8.55571628e-01 -6.70227647e-01 2.42096052e-01 -1.40011236e-01
-9.02675867e-01 -4.24379349e-01 7.99674213e-01 -1.65062934e-01
7.82969296e-01 4.77376357e-02 6.73788965e-01 1.58031964e+00
6.72358453e-01 7.05496013e-01 1.09640384e+00 -3.19363773e-01
3.49117309e-01 6.24987960e-01 3.06384657e-02 5.15026927e-01
3.40927303e-01 -1.38292769e-02 -5.69967031e-01 2.25977033e-01
2.94517696e-01 -4.48373854e-01 -1.94227159e-01 5.68408430e-01
-9.15793300e-01 1.08230054e+00 -2.37832367e-02 2.38439977e-01
-5.08753896e-01 2.47610405e-01 2.65471131e-01 3.59208018e-01
1.00449002e+00 8.40355814e-01 -7.52777696e-01 -2.16712669e-01
-6.10230088e-01 4.68553342e-02 8.82784307e-01 1.19788897e+00
6.06590331e-01 5.04284978e-01 -4.79402393e-01 5.21088839e-01
3.47290397e-01 1.06731451e+00 1.20284355e+00 -6.48453236e-01
1.47644311e-01 4.43275869e-01 1.46271974e-01 -9.23990488e-01
-5.76260611e-02 -5.27428508e-01 -9.89245594e-01 -5.10085046e-01
-4.74243939e-01 -8.74480128e-01 -8.24509263e-01 1.78373337e+00
-1.54920220e-01 -2.27173299e-01 5.76649368e-01 4.37550873e-01
1.28323793e+00 8.16049695e-01 -9.62265488e-03 -2.36026734e-01
1.06013191e+00 -1.11202896e+00 -1.15957355e+00 -3.03430349e-01
3.76275867e-01 -6.63404405e-01 1.33525908e+00 2.72730887e-01
-1.30314505e+00 -7.35843062e-01 -1.16668844e+00 6.73961639e-02
-4.01758999e-01 3.42568845e-01 6.65374875e-01 8.19755495e-01
-1.27808058e+00 7.79560864e-01 -3.10535759e-01 -4.58134152e-02
6.38457611e-02 2.01026589e-01 3.78955342e-02 5.41245282e-01
-1.60557938e+00 9.02228355e-01 6.45990893e-02 -2.10544020e-02
-9.09858406e-01 -1.88075677e-01 -1.05251825e+00 -2.23098788e-02
-1.46624371e-01 -1.02154577e+00 1.47861910e+00 -1.20020497e+00
-2.45427585e+00 6.55835867e-01 -1.01854943e-01 -3.04457664e-01
1.82610050e-01 -3.63955438e-01 -6.34069502e-01 -2.44011849e-01
2.08021197e-02 4.82439041e-01 1.02124786e+00 -1.25569570e+00
-3.27843070e-01 1.21908292e-01 2.82310676e-02 1.99273407e-01
-7.44448125e-01 -1.19178317e-01 -8.31290931e-02 -5.76352179e-01
-8.40763509e-01 -1.10980415e+00 -3.65159929e-01 -1.10136569e+00
-8.71526480e-01 -3.41369808e-01 3.27499598e-01 -5.59996188e-01
1.21513581e+00 -1.71908963e+00 8.76241773e-02 -1.58994451e-01
1.98495120e-01 2.81159639e-01 -4.51868176e-01 4.02069181e-01
2.92613715e-01 7.35621870e-01 6.95275664e-02 -8.28508914e-01
3.09043854e-01 -2.32726093e-02 -3.37516546e-01 3.22779678e-02
1.68505520e-01 1.37393296e+00 -1.18923366e+00 -4.42142695e-01
-2.64404237e-01 4.75294471e-01 -4.89635199e-01 5.33264399e-01
-4.29757088e-01 3.56422842e-01 -4.66585129e-01 1.75530761e-01
2.94956207e-01 -3.99430633e-01 1.46865144e-01 2.14856848e-01
1.44689590e-01 5.59405327e-01 -4.85808432e-01 1.35630536e+00
-6.42347515e-01 7.61899412e-01 -4.97042835e-01 -2.63839126e-01
1.31638253e+00 5.21905899e-01 -1.72385201e-01 -6.73710346e-01
5.13819158e-01 1.77587923e-02 -3.22453499e-01 -3.74047667e-01
1.16619158e+00 -9.68638584e-02 -3.18550318e-01 9.48400676e-01
1.82216376e-01 -5.88876247e-01 3.16795051e-01 3.72355938e-01
6.51979983e-01 -8.84158835e-02 3.29634726e-01 -2.75837094e-01
3.86299044e-01 -4.63668227e-01 3.95789862e-01 7.40816832e-01
-2.53008958e-02 7.21067965e-01 5.50589204e-01 2.22480875e-02
-1.10995042e+00 -8.64398181e-01 2.97450304e-01 8.93379927e-01
7.23441616e-02 -6.83308780e-01 -1.02852798e+00 -8.12847614e-01
-5.87893665e-01 1.14849126e+00 -9.76690888e-01 -4.83383209e-01
-1.16166234e-01 -8.33257079e-01 3.43511224e-01 3.03684324e-01
4.06616330e-01 -1.50352001e+00 -9.63950381e-02 6.56204745e-02
-2.24609599e-01 -1.04763031e+00 -7.96697497e-01 -2.10358858e-01
-5.73175192e-01 -5.93653560e-01 -4.51636910e-01 -6.58699214e-01
7.15201497e-01 -8.77717510e-02 1.57507908e+00 5.37904575e-02
3.29397887e-01 3.80495042e-01 -8.05538476e-01 -8.11662138e-01
-8.98824573e-01 5.79953671e-01 1.61965415e-01 -5.90024106e-02
3.11334640e-01 -5.05259514e-01 -7.22698033e-01 1.01100262e-02
-7.64239192e-01 1.63357779e-01 5.72808206e-01 8.43263209e-01
3.64723772e-01 1.64157059e-02 1.20053804e+00 -1.23940527e+00
1.61543369e+00 -4.66516495e-01 -1.22813597e-01 2.99669981e-01
-1.04983890e+00 1.66246891e-01 1.09799314e+00 -5.49049497e-01
-1.23901093e+00 -3.50483894e-01 -2.87497900e-02 3.70652318e-01
1.51033357e-01 5.94775975e-01 -5.45367263e-02 5.64663827e-01
8.58363628e-01 4.60334629e-01 -1.71439260e-01 -3.66475247e-02
9.08680141e-01 9.78194594e-01 3.19888860e-01 -3.51797253e-01
8.16090763e-01 9.01454017e-02 -6.62488103e-01 -5.18619657e-01
-1.03927422e+00 3.97620231e-01 -4.13403034e-01 -1.29444972e-01
9.82460439e-01 -9.66332257e-01 -4.75240380e-01 5.39079428e-01
-1.51732075e+00 -4.06410336e-01 -6.54576302e-01 1.67552859e-01
-4.06498343e-01 1.92305624e-01 -9.59949255e-01 -8.31875384e-01
-1.31743562e+00 -9.28280056e-01 8.39908540e-01 7.66807258e-01
-4.19208020e-01 -1.18912017e+00 3.51922840e-01 5.85666075e-02
6.48043573e-01 -6.71748817e-03 6.15393817e-01 -6.91161752e-01
-1.12248912e-01 -3.57427806e-01 2.33209476e-01 9.06652629e-01
1.88307941e-01 5.82310319e-01 -8.89796495e-01 5.18725291e-02
-3.79681885e-02 -4.68649715e-01 6.18126214e-01 2.30592847e-01
6.61903918e-01 -7.03214407e-01 2.09091887e-01 3.19286764e-01
1.07251930e+00 1.42195776e-01 8.05516779e-01 1.17979851e-02
4.90159422e-01 5.91952503e-01 1.55780882e-01 6.28150702e-01
8.24756742e-01 -5.93495853e-02 2.74065107e-01 1.52728885e-01
1.26362974e-02 -5.85738838e-01 8.84000599e-01 1.71754599e+00
2.45918091e-02 -8.44280660e-01 -2.27956936e-01 6.63025498e-01
-1.81799781e+00 -1.04708600e+00 -1.08392373e-01 1.52520227e+00
1.23647904e+00 2.08077937e-01 -3.28552127e-01 -3.92054319e-01
8.36624622e-01 2.74797976e-01 -6.24711633e-01 -1.02151239e+00
-3.94713879e-01 4.69608277e-01 2.39039809e-01 4.49525833e-01
-6.85719490e-01 1.36347449e+00 7.83746576e+00 5.24972022e-01
-8.78053069e-01 1.21027708e-01 9.77363884e-01 -7.57724270e-02
-9.64918017e-01 -2.70053625e-01 -9.44873095e-01 7.15717494e-01
1.27657568e+00 -6.72626734e-01 3.40857655e-01 9.94612396e-01
2.62361109e-01 2.83921152e-01 -9.02492285e-01 5.46621978e-01
4.89561468e-01 -1.22199476e+00 2.12708458e-01 -1.89692780e-01
1.46811974e+00 -1.55861855e-01 5.63029051e-01 5.62587261e-01
1.13999319e+00 -1.19810379e+00 7.27186799e-01 7.12868750e-01
8.34721029e-01 -8.44433963e-01 8.81531417e-01 2.01710567e-01
-5.36312342e-01 1.77393079e-01 -4.30616647e-01 -2.07292974e-01
3.54683042e-01 8.50365937e-01 -5.43215215e-01 3.98060083e-01
4.08466071e-01 1.00774837e+00 -6.35934532e-01 1.63240105e-01
-1.07145452e+00 8.82672608e-01 3.00358772e-01 -7.40814626e-01
2.33284347e-02 -2.19181925e-01 1.63434967e-01 1.24415207e+00
3.34185660e-01 1.14534190e-02 -4.09918994e-01 9.90993381e-01
-4.30066824e-01 2.28071243e-01 -4.99846607e-01 -2.06619591e-01
3.61913115e-01 1.43610811e+00 -1.92082182e-01 -5.79963028e-01
2.27495693e-02 1.06749451e+00 2.60161161e-01 3.13925594e-01
-7.45939851e-01 -6.45423889e-01 5.63645601e-01 -4.22543615e-01
3.77848417e-01 -6.44515976e-02 -6.16275549e-01 -1.39274561e+00
-2.06791282e-01 -1.05950367e+00 -2.38078833e-01 -9.32970941e-01
-1.65600443e+00 1.11526084e+00 -5.99125743e-01 -8.82513165e-01
-5.90909660e-01 -3.29258114e-01 -8.78315210e-01 7.35729098e-01
-1.48215497e+00 -1.05677414e+00 -1.79084495e-01 3.47288221e-01
5.09794772e-01 -5.36192119e-01 1.10441279e+00 -2.26940140e-02
-7.35717654e-01 8.24246168e-01 2.23890081e-01 5.63794300e-02
8.67013276e-01 -1.43905020e+00 8.57886374e-01 8.10413837e-01
-1.82401881e-01 7.05220520e-01 9.70164657e-01 -7.35940278e-01
-1.01368511e+00 -1.18864691e+00 1.29289424e+00 -5.11128604e-01
7.12482154e-01 -4.33528721e-01 -2.86060482e-01 6.93780422e-01
1.01632202e+00 -6.76747501e-01 9.44427133e-01 1.57029286e-01
-1.26581818e-01 7.26726502e-02 -8.22060943e-01 1.02489603e+00
8.66811275e-01 -4.28236693e-01 -4.91868407e-01 4.70142037e-01
1.33595967e+00 -1.52732939e-01 -6.64560854e-01 1.41380489e-01
3.56190264e-01 -9.17766035e-01 2.96954840e-01 -6.45885408e-01
1.41620982e+00 1.39024422e-01 1.77289620e-01 -2.04845190e+00
-4.78245735e-01 -9.41579401e-01 -2.50743955e-01 1.56133795e+00
8.37186396e-01 -5.76400697e-01 5.77807248e-01 6.43456042e-01
-2.21718565e-01 -7.24975944e-01 -3.23886782e-01 -5.37346900e-01
2.45312750e-01 -6.79006204e-02 9.86670792e-01 6.80317402e-01
2.61003703e-01 1.10566592e+00 -8.97823334e-01 -4.13880467e-01
2.05922946e-01 6.86419532e-02 8.89033914e-01 -9.09658492e-01
-5.00552237e-01 -5.09425163e-01 1.11885838e-01 -1.23258936e+00
5.27613163e-01 -6.51088595e-01 2.96216011e-01 -1.67297900e+00
2.07418427e-01 -4.55021746e-02 -9.18348432e-02 4.60439362e-02
-6.19580448e-01 1.62923262e-01 4.30217534e-02 -1.86505064e-01
-9.33536947e-01 1.17946994e+00 1.76964247e+00 -1.53785959e-01
-3.17596674e-01 -2.65591070e-02 -1.66637933e+00 3.94875377e-01
1.14519751e+00 -4.03924733e-01 -6.64213538e-01 -4.86120224e-01
8.23970139e-01 -2.43547440e-01 -3.43068004e-01 -5.87071896e-01
-1.57107471e-03 -1.22717589e-01 6.49274066e-02 -3.65151405e-01
1.09888732e-01 -7.55898654e-02 -2.90114313e-01 1.38452306e-01
-7.75829196e-01 3.76085460e-01 -1.68798313e-01 4.65448201e-01
-4.54381183e-02 -3.32774282e-01 4.22369093e-01 -2.79408216e-01
7.82827511e-02 2.04414576e-01 -6.47751331e-01 2.71262527e-01
7.87837744e-01 3.32884818e-01 -6.55006230e-01 -1.12450504e+00
-1.85414732e-01 2.01265469e-01 3.50744814e-01 6.55175388e-01
7.36429691e-01 -1.11414766e+00 -1.21011865e+00 -1.92494452e-01
-5.54641746e-02 -1.27996251e-01 2.24130794e-01 4.76098694e-02
-2.54893243e-01 3.84029955e-01 1.11016326e-01 8.18647165e-03
-8.13541591e-01 4.59841371e-01 1.97651401e-01 -7.18103886e-01
-2.27863625e-01 8.80042493e-01 1.67778060e-02 -6.42259002e-01
-2.65165269e-01 -8.32839012e-02 -5.20108998e-01 -6.41619340e-02
5.30787289e-01 7.09355250e-02 -3.17440778e-01 -4.26499635e-01
2.37644196e-01 3.35571885e-01 -2.95267463e-01 -2.26123527e-01
1.25285625e+00 -3.78144026e-01 -1.16966061e-01 4.65211481e-01
9.12615240e-01 2.05930382e-01 -9.99922633e-01 -1.59384996e-01
-4.29962486e-01 6.08540401e-02 -3.31893504e-01 -9.04958844e-01
-1.19553852e+00 5.40221989e-01 -2.05112379e-02 4.99215037e-01
8.01600575e-01 -7.48051330e-02 1.24928045e+00 3.29055667e-01
-1.91626232e-02 -1.43541574e+00 4.67698604e-01 8.88103843e-01
7.92893708e-01 -1.24251664e+00 3.49997617e-02 2.16461532e-02
-1.26307988e+00 7.83566296e-01 7.79129744e-01 -4.38127875e-01
6.71782494e-01 5.86087704e-02 3.38852942e-01 -3.76969166e-02
-1.31434429e+00 2.48262677e-02 2.03419104e-01 8.05900633e-01
7.90462554e-01 4.09035325e-01 -7.56073236e-01 1.21920264e+00
-1.18519902e+00 -3.42389405e-01 1.11116242e+00 2.62848526e-01
-3.34038645e-01 -1.06716537e+00 2.43661314e-01 3.82830530e-01
-5.52536368e-01 -5.34939289e-01 -6.62978172e-01 3.48441035e-01
-1.07137106e-01 1.38312030e+00 -1.26781315e-01 -8.44966292e-01
1.29098684e-01 -1.51259825e-01 4.71089482e-02 -8.61052096e-01
-9.91636992e-01 -3.45153660e-01 2.77805239e-01 1.27992677e-02
-2.58796155e-01 -4.17361975e-01 -1.08676553e+00 -5.68723619e-01
-4.74529356e-01 4.17441428e-01 8.85022938e-01 9.51855958e-01
6.84069455e-01 8.17753375e-01 1.04899800e+00 -5.56849718e-01
-6.97137535e-01 -1.26414084e+00 -5.62759519e-01 4.52329665e-01
1.32422492e-01 1.08152680e-01 -5.27391911e-01 2.71770686e-01] | [11.99781322479248, 9.06580638885498] |
e8ac01fd-1171-4b24-b77c-da7c5424ba69 | hands-on-wireless-sensing-with-wi-fi-a | 2206.09532 | null | https://arxiv.org/abs/2206.09532v1 | https://arxiv.org/pdf/2206.09532v1.pdf | Hands-on Wireless Sensing with Wi-Fi: A Tutorial | With the rapid development of wireless communication technology, wireless access points (AP) and internet of things (IoT) devices have been widely deployed in our surroundings. Various types of wireless signals (e.g., Wi-Fi, LoRa, LTE) are filling out our living and working spaces. Previous researches reveal the fact that radio waves are modulated by the spatial structure during the propagation process (e.g., reflection, diffraction, and scattering) and superimposed on the receiver. This observation allows us to reconstruct the surrounding environment based on received wireless signals, called "wireless sensing". Wireless sensing is an emerging technology that enables a wide range of applications, such as gesture recognition for human-computer interaction, vital signs monitoring for health care, and intrusion detection for security management. Compared with other sensing paradigms, such as vision-based and IMU-based sensing, wireless sensing solutions have unique advantages such as high coverage, pervasiveness, low cost, and robustness under adverse light and texture scenarios. Besides, wireless sensing solutions are generally lightweight in terms of both computation overhead and device size. This tutorial takes Wi-Fi sensing as an example. It introduces both the theoretical principles and the code implementation of data collection, signal processing, features extraction, and model design. In addition, this tutorial highlights state-of-the-art deep learning models (e.g., CNN, RNN, and adversarial learning models) and their applications in wireless sensing systems. We hope this tutorial will help people in other research fields to break into wireless sensing research and learn more about its theories, designs, and implementation skills, promoting prosperity in the wireless sensing research field. | ['Guidong Zhang', 'Guoxuan Chi', 'Yi Zhang', 'Zheng Yang'] | 2022-06-20 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 5.52242756e-01 -3.13113570e-01 -2.52849132e-01 -1.33871600e-01
-9.68658477e-02 -2.12927565e-01 2.42904294e-02 -3.52522999e-01
-5.16895235e-01 7.56870270e-01 2.74800062e-02 -3.49252999e-01
-2.34000549e-01 -1.19943523e+00 -2.10367307e-01 -1.17244184e+00
-1.57708719e-01 -5.25130332e-01 2.36465782e-02 -1.01915374e-01
-1.63280129e-01 4.42739338e-01 -1.30864334e+00 -3.15806359e-01
4.33561772e-01 1.54649270e+00 3.44473541e-01 5.72376192e-01
-6.29323870e-02 5.66741109e-01 -7.26019263e-01 1.53360844e-01
-7.47603625e-02 -3.36947665e-02 -1.76345512e-01 -6.67181194e-01
-4.49370295e-01 -4.27695990e-01 -7.81828940e-01 9.80948985e-01
8.87732089e-01 -1.10969238e-01 2.58035392e-01 -1.03501642e+00
-4.62536275e-01 6.50724351e-01 -6.39006436e-01 3.57124768e-02
5.22366583e-01 -1.97916076e-01 1.68439895e-01 -4.78236079e-01
-4.74630557e-02 6.92860782e-01 1.00508058e+00 8.85050952e-01
-4.35043424e-01 -1.22184718e+00 -1.45189092e-01 1.14622548e-01
-1.43505716e+00 -4.16357487e-01 8.20513606e-01 7.26499557e-02
6.03819609e-01 3.95736188e-01 5.71782768e-01 1.50901949e+00
3.97614598e-01 7.48754680e-01 7.50333786e-01 -5.38814843e-01
2.69313067e-01 -1.41706377e-01 -9.46967676e-02 5.97386301e-01
3.77349079e-01 4.13670927e-01 -2.82914281e-01 -1.32000055e-02
6.93778276e-01 6.52321517e-01 -3.78940195e-01 1.49983987e-01
-1.36567867e+00 3.30747068e-01 6.12414598e-01 7.22558260e-01
-5.30052602e-01 5.21999478e-01 -5.83047159e-02 2.47472152e-01
-3.00841406e-02 -7.51420809e-03 -3.44199121e-01 -2.27507159e-01
-6.98981762e-01 -3.97662699e-01 7.79734433e-01 8.69643688e-01
3.72811168e-01 3.14267725e-01 5.81314825e-02 7.45364010e-01
7.16346383e-01 1.47483325e+00 6.88835144e-01 -7.28820026e-01
4.13920045e-01 -8.38888958e-02 1.80020303e-01 -1.08391154e+00
-9.51097250e-01 -3.62409949e-01 -1.49406242e+00 -1.65839836e-01
-7.50343800e-02 -8.86410117e-01 -6.24050617e-01 1.74258721e+00
1.54009938e-01 5.72597265e-01 1.33017093e-01 6.32970631e-01
1.07775950e+00 5.46847403e-01 7.94508755e-02 -2.53385156e-01
1.24516189e+00 -3.06296587e-01 -7.49365807e-01 -1.89318120e-01
7.89309740e-02 -5.49822092e-01 4.88852561e-01 3.27725917e-01
-6.51998639e-01 -6.81242347e-01 -1.30062950e+00 6.29951715e-01
-3.43641251e-01 -1.22774795e-01 8.93802285e-01 1.30264568e+00
-5.89578211e-01 1.41923219e-01 -1.28959727e+00 -5.42262673e-01
4.84104067e-01 3.93812507e-01 3.46783012e-01 -1.33637577e-01
-1.29005587e+00 3.69554460e-01 -9.09773260e-02 2.27988094e-01
-5.93838692e-01 -3.71106714e-01 -5.27176142e-01 -5.95692024e-02
-1.61258578e-02 -5.26924729e-01 1.05966163e+00 -3.89327824e-01
-1.57850873e+00 1.67285889e-01 -1.91406727e-01 -1.96943268e-01
-3.08114123e-02 -4.46252190e-02 -1.17020273e+00 -4.25121710e-02
-1.92942679e-01 1.67449310e-01 7.25215316e-01 -9.82373059e-01
-5.33464074e-01 -3.05284798e-01 -2.97750197e-02 -2.63225675e-01
-5.71393847e-01 1.73246395e-02 2.07573906e-01 -4.44733292e-01
5.34681797e-01 -7.45371938e-01 -2.66574293e-01 1.81726776e-02
-3.56432855e-01 2.51743525e-01 1.22783995e+00 -1.60160527e-01
1.15101779e+00 -2.20885158e+00 -5.45614481e-01 5.50259292e-01
2.68189520e-01 3.43403608e-01 -7.80302286e-02 3.49039853e-01
3.63050193e-01 5.64595051e-02 -4.54912968e-02 1.58038232e-02
-1.74176231e-01 1.71571180e-01 -1.83127254e-01 4.88183439e-01
-6.87726736e-01 9.48561013e-01 -1.01167476e+00 -1.69989049e-01
4.40885991e-01 8.12491119e-01 9.01726037e-02 -4.13058810e-02
4.52720672e-01 7.99727976e-01 -1.07827115e+00 8.58423948e-01
7.38546610e-01 -1.77435279e-01 -2.86070891e-02 -1.33030951e-01
-1.22810662e-01 5.24562411e-02 -1.14359367e+00 1.41930425e+00
-9.80032265e-01 5.59070945e-01 4.53517288e-01 -1.19889092e+00
1.01629579e+00 7.04847991e-01 9.30416107e-01 -1.11805069e+00
3.88500273e-01 2.14046061e-01 -3.45928162e-01 -9.68026102e-01
-5.61851971e-02 -3.21998186e-02 -1.23010904e-01 6.95509553e-01
-4.22178358e-01 3.70534211e-01 -3.43938708e-01 -3.39720637e-01
1.40029120e+00 -1.59004688e-01 2.68621922e-01 1.85548425e-01
4.37083632e-01 -6.74359560e-01 4.04529393e-01 9.47928846e-01
-2.79016614e-01 1.05908446e-01 -7.56272435e-01 -3.95440221e-01
-1.36957169e-01 -1.16036928e+00 -1.86341763e-01 9.72556889e-01
6.81617141e-01 2.24045485e-01 -4.56160545e-01 -8.98093823e-03
-1.54617280e-01 1.29573837e-01 -1.86304897e-01 -2.20076367e-01
-5.33478737e-01 -9.94959533e-01 9.34355795e-01 5.62610030e-01
1.32042050e+00 -1.28551280e+00 -1.07462561e+00 5.02874434e-01
-4.93615746e-01 -1.18608546e+00 1.27083570e-01 1.40616104e-01
-7.09699094e-01 -8.38286340e-01 -7.22349763e-01 -7.89833128e-01
3.79595608e-01 7.80283749e-01 6.99154437e-01 3.35867375e-01
-2.38205433e-01 7.03422308e-01 -4.82853830e-01 -1.03383470e+00
8.63979384e-02 -1.13975825e-02 4.03396368e-01 1.16464552e-02
7.59305894e-01 -8.18464577e-01 -8.34542751e-01 5.12578726e-01
-8.02147865e-01 -5.29384971e-01 7.00944960e-01 5.26668131e-01
2.26212233e-01 5.73183835e-01 5.67259848e-01 -2.97656447e-01
5.86493075e-01 -5.84642291e-01 -3.31185609e-01 2.59272158e-01
-3.41467023e-01 -4.88795251e-01 4.42886502e-01 -5.27002096e-01
-1.03597772e+00 -2.59503692e-01 -4.26434964e-01 2.67411828e-01
-4.21453416e-01 6.45666182e-01 -3.51445645e-01 -5.45967162e-01
7.46764779e-01 2.55248249e-01 -1.47584423e-01 -2.99874723e-01
-2.89584786e-01 1.33642256e+00 3.93928140e-01 -2.37438455e-01
9.74129677e-01 7.78601050e-01 -1.31462282e-02 -1.32999623e+00
-6.55017674e-01 -4.52792078e-01 -3.38452384e-02 -3.50646228e-01
7.62490809e-01 -7.88577676e-01 -9.49670792e-01 7.34086037e-01
-1.03145444e+00 -3.02695811e-01 8.77026096e-02 8.44548702e-01
-7.34069198e-02 1.98453724e-01 -5.77345669e-01 -1.10058534e+00
-4.03744578e-01 -7.44849682e-01 7.24312007e-01 7.56343603e-01
-2.07457274e-01 -1.07354867e+00 -1.29625546e-02 2.73386121e-01
1.18288755e+00 3.28154176e-01 3.61986488e-01 1.93773717e-01
-5.18214941e-01 -4.33399767e-01 -4.71129492e-02 -6.71639591e-02
4.13684160e-01 -5.67899227e-01 -1.22961140e+00 -1.82078242e-01
1.99497849e-01 5.80629855e-02 7.08062887e-01 7.84590423e-01
1.46391928e+00 -2.80183613e-01 -1.01513183e+00 9.23568428e-01
1.34912181e+00 6.46565259e-01 8.66494536e-01 7.78972730e-02
4.93545949e-01 1.45102799e-01 2.76361853e-01 4.72753406e-01
4.00878824e-02 3.47898006e-01 6.21652782e-01 -3.19414914e-01
1.22332998e-01 -4.23434153e-02 1.75560161e-01 5.81183374e-01
-4.56070185e-01 -4.49695855e-01 -6.80615008e-01 -5.15824603e-03
-1.49327779e+00 -1.22228086e+00 3.26496474e-02 2.08231902e+00
2.14946464e-01 -7.99897984e-02 -1.77735567e-01 4.51551855e-01
6.18264258e-01 1.49737448e-01 -6.98693216e-01 3.38269174e-02
7.60316849e-02 4.91118759e-01 9.06978071e-01 3.20482939e-01
-1.16048527e+00 3.54029864e-01 6.44569635e+00 3.42545062e-01
-1.60205436e+00 2.69456029e-01 5.11611737e-02 2.38965765e-01
-3.29668730e-01 -8.33503067e-01 -3.89915675e-01 4.95370388e-01
6.88607752e-01 5.20061612e-01 6.97773874e-01 6.52470052e-01
3.75497639e-01 -3.86286192e-02 -4.91770804e-01 1.35994709e+00
-2.43286729e-01 -1.18413055e+00 -7.06044674e-01 9.17009562e-02
4.22520518e-01 5.30999601e-01 1.11808721e-02 2.22643211e-01
5.93923293e-02 -9.88996267e-01 3.54824066e-01 5.36706150e-01
1.03020692e+00 -4.46107715e-01 1.02654743e+00 4.41019207e-01
-1.49621105e+00 -2.95690030e-01 -3.22059482e-01 -4.24195707e-01
1.43075541e-01 8.56342912e-01 -4.59396467e-02 4.37519521e-01
1.02195907e+00 5.92069626e-01 2.12071851e-01 9.42072749e-01
-2.88830936e-01 8.40413809e-01 -5.78476906e-01 -6.80931747e-01
-2.38086972e-02 -7.26354076e-03 3.39850932e-01 1.00775135e+00
6.65420175e-01 3.94959301e-01 7.55274808e-03 6.21267140e-01
5.93056381e-02 -5.40437520e-01 -6.99758887e-01 2.84103066e-01
9.09317493e-01 1.13387632e+00 -7.01297462e-01 1.51724771e-01
-7.84258544e-01 5.22023439e-01 -8.04378867e-01 7.82051802e-01
-8.07381034e-01 -8.62610936e-01 6.97822392e-01 -6.22484125e-02
1.06453598e-01 -4.72146630e-01 -4.19199765e-01 -9.13730860e-01
-2.48163454e-02 -3.96116138e-01 -1.00890279e-01 -4.93571848e-01
-1.06166708e+00 2.04416692e-01 -2.79742241e-01 -1.42007256e+00
1.74805596e-01 -5.58757126e-01 -8.13818157e-01 6.23813093e-01
-1.67114544e+00 -7.26701975e-01 -9.48232234e-01 8.61421943e-01
1.53563153e-02 -1.75618231e-01 1.17225456e+00 5.58702528e-01
-3.46007794e-01 5.67295134e-01 4.24261272e-01 6.01968884e-01
1.17004409e-01 -4.72918272e-01 1.10018313e-01 7.22552001e-01
6.40492961e-02 7.07354963e-01 3.50338161e-01 -2.29472265e-01
-1.53470707e+00 -9.90201056e-01 5.12786329e-01 1.78939868e-02
4.78145987e-01 -1.19999960e-01 -2.76545733e-01 4.10331339e-01
-1.87704653e-01 1.85048789e-01 8.80666852e-01 -3.12906392e-02
-8.34364370e-02 -6.40140593e-01 -1.53612065e+00 4.79266554e-01
1.07409155e+00 -3.22691530e-01 -4.24033813e-02 4.64280620e-02
2.76451617e-01 -2.09330305e-01 -6.02926016e-01 3.29966754e-01
1.24890411e+00 -6.98401451e-01 1.20890868e+00 1.63959995e-01
-2.27683499e-01 -1.52761146e-01 -4.17547286e-01 -1.01344812e+00
-3.34767193e-01 -7.65045643e-01 -2.81259000e-01 9.24790025e-01
2.41282925e-01 -1.07582629e+00 8.00941288e-01 2.68903196e-01
2.15816185e-01 -6.31430447e-01 -1.03426194e+00 -6.19620800e-01
-4.80308205e-01 -9.39796150e-01 1.00198424e+00 6.24791086e-01
8.25368986e-02 1.63368642e-01 -3.77720833e-01 4.41024631e-01
8.14798892e-01 -2.60426253e-01 4.76925462e-01 -1.46921206e+00
-1.95077166e-01 -2.53299534e-01 -4.00393248e-01 -1.52946210e+00
-3.85422081e-01 -4.23234701e-01 2.89141177e-03 -1.76641035e+00
-3.22636604e-01 -9.91856217e-01 -7.79519379e-01 4.01282877e-01
3.21455628e-01 6.07746422e-01 -2.40091801e-01 1.16414443e-01
-4.63226944e-01 1.40106887e-01 1.17991340e+00 -5.92396319e-01
-2.73510903e-01 6.82602108e-01 -8.05120349e-01 8.23938012e-01
1.18663156e+00 -2.30487153e-01 -2.95644253e-01 -6.50529385e-01
2.78596908e-01 1.53565139e-01 3.85203004e-01 -1.47112334e+00
3.49950016e-01 -1.81623653e-01 6.41332567e-01 -6.89927265e-02
4.07814413e-01 -1.48100770e+00 1.23183848e-02 8.58120143e-01
6.96370006e-02 -4.69061613e-01 -1.91178367e-01 5.55371463e-01
6.12654816e-03 1.59499850e-02 5.99342942e-01 -8.30276534e-02
-7.41831601e-01 4.44628149e-01 -7.51791596e-01 -4.41442847e-01
8.79124045e-01 -5.37489891e-01 -3.03452671e-01 -9.26740050e-01
-3.79256308e-01 1.04461405e-02 -2.14261070e-01 2.75184125e-01
8.45396340e-01 -1.24467623e+00 -2.41201118e-01 5.23873150e-01
-8.01439136e-02 -2.30519712e-01 1.26009032e-01 9.32316422e-01
-3.06068987e-01 4.11122084e-01 -5.49998023e-02 -5.54399967e-01
-1.08743536e+00 6.96729124e-02 4.65326607e-01 5.88605851e-02
-5.19778788e-01 7.55936861e-01 -4.04867291e-01 -2.64507562e-01
6.52222395e-01 -3.76135737e-01 -2.08877876e-01 -5.33186972e-01
9.47616637e-01 5.62270343e-01 -6.82809427e-02 -2.84814984e-01
-6.57665372e-01 1.08050489e+00 6.76164150e-01 2.07660615e-01
1.07779551e+00 -4.06078339e-01 2.42705852e-01 1.76520362e-01
9.12646115e-01 2.08717614e-01 -7.64928997e-01 -2.33547539e-01
-3.25178683e-01 -1.01996377e-01 2.20280334e-01 -7.67507493e-01
-1.24929786e+00 8.35626781e-01 1.09624362e+00 6.85057759e-01
1.35315156e+00 -8.46945420e-02 1.30977023e+00 7.10530698e-01
1.28093755e+00 -9.88389313e-01 -9.72027406e-02 3.30128938e-01
2.54694283e-01 -1.13897145e+00 -9.96226892e-02 -1.84632570e-01
2.12445900e-01 1.07429028e+00 1.13842383e-01 1.66643739e-01
1.40726840e+00 8.28507483e-01 5.36046445e-01 7.99509287e-02
1.17838785e-01 -1.69407740e-01 -3.11712772e-01 1.20987034e+00
4.71538186e-01 3.62114847e-01 7.26231113e-02 5.47662914e-01
-2.28194743e-01 1.68547824e-01 3.51804905e-02 1.05672228e+00
-7.01954782e-01 -9.07158434e-01 -5.71235538e-01 6.06568933e-01
-7.06921339e-01 1.87093377e-01 2.52229989e-01 4.66550916e-01
4.20814484e-01 1.71482170e+00 -2.38995373e-01 -8.61751676e-01
1.62649721e-01 -3.25843662e-01 4.39928353e-01 -1.56053096e-01
-5.03304824e-02 -1.75186336e-01 -2.24950522e-01 -6.31064892e-01
-9.54032838e-01 -1.80940270e-01 -1.31057465e+00 -4.24244910e-01
-3.33955795e-01 3.36822644e-02 9.87430274e-01 1.21021974e+00
1.00233421e-01 7.56727040e-01 7.12488651e-01 -8.59959245e-01
-1.39598072e-01 -9.85997856e-01 -8.21141362e-01 -1.56682506e-01
6.01956487e-01 -7.12932289e-01 -2.27183446e-01 -3.60032499e-01] | [6.569962501525879, 0.8358908295631409] |
f06f831a-b896-4dfc-9eb8-7b861a4eb9a1 | neural-code-summarization | 2103.01025 | null | https://arxiv.org/abs/2103.01025v2 | https://arxiv.org/pdf/2103.01025v2.pdf | Neural Code Summarization | Code summarization is the task of generating readable summaries that are semantically meaningful and can accurately describe the presumed task of a software. Program comprehension has become one of the most tedious tasks for knowledge transfer. As the codebase evolves over time, the description needs to be manually updated each time with the changes made. An automatic approach is proposed to infer such captions based on benchmarked and custom datasets with comparison between the original and generated results. | ['Piyush Shrivastava'] | 2021-02-26 | null | null | null | null | ['code-summarization'] | ['computer-code'] | [ 3.96229029e-01 4.54035401e-01 -2.21827373e-01 -3.73723209e-01
-1.03597248e+00 -6.03372991e-01 3.18131834e-01 9.39436197e-01
1.85685292e-01 7.38476276e-01 7.07766831e-01 -1.87638581e-01
7.00025633e-02 -4.69564408e-01 -7.63116002e-01 3.84596050e-01
-9.36472267e-02 3.51088822e-01 3.29078346e-01 -3.19793016e-01
8.20585668e-01 -2.25890487e-01 -1.61844850e+00 5.21688163e-01
1.42450893e+00 1.78924352e-01 5.70512235e-01 9.56276000e-01
-6.55149996e-01 1.01575017e+00 -1.11100733e+00 -4.66521621e-01
-4.17502671e-01 -6.61832869e-01 -1.30552626e+00 -2.26267383e-01
2.31074870e-01 7.33883530e-02 1.88812554e-01 1.47637975e+00
1.73040912e-01 -6.72286004e-02 4.07907516e-01 -1.16966522e+00
-7.59651244e-01 1.10652792e+00 -2.86307096e-01 1.06053948e-01
8.89576197e-01 -1.03542954e-01 6.71995044e-01 -4.59962100e-01
8.22038829e-01 8.81128073e-01 6.20838702e-01 7.02676654e-01
-9.95149612e-01 -1.55428961e-01 -3.30107957e-02 3.57203215e-01
-9.32547152e-01 -3.61896753e-01 7.28538930e-01 -7.03505218e-01
1.33520353e+00 5.02500594e-01 5.17457128e-01 6.38151884e-01
3.01622093e-01 3.92832458e-01 2.46424913e-01 -5.70693314e-01
1.60993367e-01 3.88567507e-01 7.10509598e-01 8.31403852e-01
7.01371312e-01 -8.41258943e-01 -4.02957827e-01 -5.44184983e-01
-7.79932588e-02 -2.61512607e-01 -6.93796098e-01 -1.62618473e-01
-1.00686657e+00 4.53277349e-01 1.32137677e-02 4.88721251e-01
-3.31259489e-01 4.48968828e-01 8.76265466e-01 1.91074669e-01
4.15136814e-01 1.20097017e+00 -5.12163460e-01 -8.16906452e-01
-1.00277615e+00 2.14258358e-01 1.31487834e+00 1.41216576e+00
8.10881317e-01 -2.31136948e-01 -3.92886430e-01 5.67267120e-01
1.50906593e-01 2.68582791e-01 8.37093115e-01 -6.35165751e-01
7.26314664e-01 1.26705217e+00 4.18620259e-01 -1.15832114e+00
1.28125027e-01 -3.88721257e-01 -2.38810241e-01 -8.12670365e-02
-3.26602697e-01 1.46682650e-01 -5.25483608e-01 1.27062094e+00
-2.37800807e-01 -2.39314824e-01 2.77031839e-01 2.24951312e-01
1.05950546e+00 8.21789920e-01 1.16982117e-01 -4.17294472e-01
1.10977781e+00 -9.81220305e-01 -8.68816495e-01 -4.47251290e-01
6.65830791e-01 -7.44537830e-01 1.07092988e+00 1.21154040e-01
-9.37327147e-01 -5.77324927e-01 -1.31621253e+00 8.79743770e-02
-2.54455626e-01 1.75789922e-01 3.78323108e-01 3.83031577e-01
-9.35820520e-01 5.62419474e-01 -7.43097603e-01 -5.32175124e-01
2.71857530e-01 -1.19564600e-01 -9.07516778e-02 6.22415021e-02
-6.54317319e-01 9.44648504e-01 8.01773310e-01 -2.15990752e-01
-7.71876097e-01 -8.73791099e-01 -8.86979222e-01 3.10109049e-01
4.00486946e-01 -8.36237788e-01 1.78008878e+00 -1.18110704e+00
-1.17302549e+00 5.66589355e-01 -4.67476785e-01 -3.20708275e-01
1.62899703e-01 -4.29056704e-01 -2.83328265e-01 -2.12109536e-01
2.37070128e-01 3.85677591e-02 5.18955290e-01 -1.38027442e+00
-3.28364819e-01 -1.67308509e-01 1.17814712e-01 4.88311425e-02
-1.97525784e-01 1.47894174e-01 -4.40217733e-01 -5.58941007e-01
-4.45133418e-01 -6.46378458e-01 -7.31754899e-02 -5.54866910e-01
-5.45537949e-01 -3.93823743e-01 7.69191861e-01 -1.29332125e+00
1.99421740e+00 -2.01168180e+00 4.10722554e-01 -2.12576687e-01
1.74733773e-01 2.09273115e-01 -9.61647481e-02 7.20655859e-01
-4.38506156e-02 4.47501898e-01 -6.63961768e-01 1.18653856e-01
1.42966677e-02 -2.64424473e-01 -4.91387516e-01 -3.16343516e-01
1.81160122e-01 8.61291230e-01 -1.18698776e+00 -5.31518936e-01
-3.14017862e-01 -7.49332458e-02 -4.56392109e-01 5.67151845e-01
-7.85066903e-01 2.03338668e-01 -7.23246336e-01 3.81807894e-01
2.47253880e-01 -3.17635417e-01 -1.42665952e-01 -1.73416167e-01
-7.09056407e-02 2.54115164e-01 -5.74004054e-01 1.97768784e+00
-7.16733515e-01 8.76419902e-01 -5.98278403e-01 -7.34271049e-01
9.95675743e-01 4.50310677e-01 -1.37716308e-01 -4.08526987e-01
-3.09596986e-01 1.05595581e-01 -3.78827065e-01 -1.44621944e+00
7.64188647e-01 3.21149051e-01 -4.36215609e-01 6.95703387e-01
-2.94610023e-01 -5.06955683e-01 4.66049880e-01 4.49986458e-01
1.35700607e+00 2.63313264e-01 9.45134759e-01 -1.43084645e-01
6.40407801e-01 8.36170733e-01 5.02236605e-01 5.60565472e-01
2.52701193e-01 3.54227245e-01 6.89003348e-01 -3.80933404e-01
-9.88249242e-01 -5.37667990e-01 4.74202424e-01 9.39449072e-01
2.37047136e-01 -7.77004957e-01 -1.14258730e+00 -8.71001005e-01
-1.70359671e-01 1.45967960e+00 -7.02539682e-01 -4.33676541e-01
-6.21127903e-01 -3.43392789e-01 4.95187402e-01 5.23578346e-01
3.11646849e-01 -1.24623299e+00 -1.06841409e+00 2.77737528e-01
-4.75191981e-01 -4.62343186e-01 -5.58437824e-01 -1.75312042e-01
-7.13568211e-01 -1.36827326e+00 -1.81420311e-01 -7.44864643e-01
1.00032437e+00 -1.19545385e-02 1.53889930e+00 6.72959089e-01
-3.48670661e-01 4.75617379e-01 -7.52778471e-01 -6.02937698e-01
-1.37442088e+00 7.90344104e-02 -8.45137060e-01 -7.88669050e-01
2.44724438e-01 -3.38837296e-01 -5.57501130e-02 -2.16384277e-01
-1.01812279e+00 5.41221380e-01 5.05484164e-01 4.60428476e-01
3.28746885e-01 -1.97873302e-02 6.61716819e-01 -1.39208663e+00
1.17115295e+00 -5.58975279e-01 -4.39810663e-01 1.03365338e+00
-5.32574415e-01 5.40704668e-01 8.88350427e-01 -7.84322247e-02
-1.28836071e+00 6.08176664e-02 3.32153857e-01 3.03520441e-01
-2.43478063e-02 1.06398106e+00 -7.71516934e-02 2.84407139e-01
9.10315573e-01 5.28219283e-01 -3.56715262e-01 -4.20853496e-01
2.64228910e-01 8.58782113e-01 4.87883598e-01 -5.95566511e-01
6.15580738e-01 -2.95453280e-01 -7.23409653e-01 -6.10540509e-01
-5.19023359e-01 -1.54297769e-01 -7.44636953e-01 -3.27614158e-01
5.96670210e-01 -3.29481661e-01 2.26318184e-03 -1.44466590e-02
-1.64278471e+00 -8.19323957e-02 -4.00075257e-01 -1.35525659e-01
-5.55118501e-01 5.50702631e-01 -3.73370945e-02 -3.92423689e-01
-7.75366843e-01 -1.31153345e+00 8.10186267e-01 3.59793365e-01
-8.70367706e-01 -8.22516322e-01 4.78251278e-01 1.61619335e-01
7.01012850e-01 5.28533399e-01 1.56280196e+00 -7.18049109e-01
-5.20214438e-01 -4.81172830e-01 -4.19331640e-02 4.15383205e-02
5.43553233e-01 4.58997697e-01 -5.48880458e-01 -1.22604012e-01
-2.30839536e-01 5.45336911e-03 2.71331489e-01 -1.47457346e-01
1.34791386e+00 -7.52019525e-01 -5.08976340e-01 -3.51433344e-02
1.47074223e+00 4.41920340e-01 4.41703290e-01 2.19596565e-01
5.78366935e-01 4.21288967e-01 5.33609509e-01 3.03922802e-01
4.16599482e-01 4.98756170e-01 3.91580611e-01 6.78213596e-01
-3.56030643e-01 -4.20967996e-01 3.57987672e-01 1.34058225e+00
3.04873466e-01 -2.86458164e-01 -1.18746507e+00 7.96102822e-01
-1.91447973e+00 -9.02826846e-01 -1.75715178e-01 2.16761851e+00
9.80872035e-01 -5.84746785e-02 -3.53914291e-01 -2.85620183e-01
7.51404405e-01 -7.78073147e-02 -8.65584433e-01 -6.26231372e-01
4.24045354e-01 -1.12252519e-01 -5.33357039e-02 3.16205412e-01
-4.93738115e-01 6.16765618e-01 6.43695402e+00 2.76675165e-01
-8.98918569e-01 2.88356319e-02 -2.24612262e-02 5.36906064e-01
-7.08095491e-01 3.43031913e-01 -2.46958390e-01 4.96544391e-01
1.18241107e+00 -1.50524688e+00 2.04843298e-01 1.05977023e+00
1.73428297e-01 -4.30175483e-01 -1.50891948e+00 6.74722672e-01
5.33718765e-01 -1.39050198e+00 4.66279835e-01 -6.94119573e-01
7.66720295e-01 -3.00920516e-01 -4.13692921e-01 3.77465427e-01
4.11219031e-01 -9.42520976e-01 6.56354547e-01 8.64544034e-01
6.09700441e-01 -5.81957459e-01 9.45441186e-01 5.11626959e-01
-1.11082697e+00 1.25701576e-01 -2.70199299e-01 4.12511826e-02
-4.13428014e-03 4.25338626e-01 -1.22808063e+00 7.96175182e-01
4.46266860e-01 1.02610791e+00 -1.21927452e+00 1.37995422e+00
-3.55036318e-01 5.83218098e-01 4.09090310e-01 -5.49431026e-01
-4.33482230e-01 3.01643670e-01 7.35011935e-01 1.49029577e+00
5.78900158e-01 -1.41041592e-01 -9.73641779e-03 1.12128186e+00
-2.30598286e-01 3.29339951e-01 -7.55099297e-01 -4.55557048e-01
6.09524548e-01 8.71482670e-01 -4.57262725e-01 -7.07725585e-01
-2.42243841e-01 1.05372500e+00 2.75335073e-01 2.18855768e-01
-6.24520838e-01 -9.05080795e-01 1.67697638e-01 -1.26056835e-01
1.46908704e-02 1.26374573e-01 -2.21224710e-01 -1.15312576e+00
4.19578135e-01 -7.68435657e-01 3.22310090e-01 -1.11338937e+00
-8.62726331e-01 8.93572986e-01 1.84012815e-01 -1.33090532e+00
-2.70712018e-01 2.61616975e-01 -9.31045771e-01 7.59708941e-01
-1.14445817e+00 -6.22233033e-01 -9.31481779e-01 -2.46837780e-01
1.18738747e+00 -8.32066089e-02 1.05715585e+00 -1.77009031e-02
-6.14846647e-01 3.21379453e-01 -4.01486382e-02 -1.02722012e-01
5.21544337e-01 -1.33920443e+00 5.72512448e-01 1.19267678e+00
-1.87744811e-01 1.07233989e+00 1.31370640e+00 -1.03637481e+00
-1.35455847e+00 -1.52009881e+00 9.36620414e-01 -5.47245502e-01
6.03998482e-01 6.43941686e-02 -1.49358499e+00 6.99334860e-01
5.77244461e-01 -6.33340418e-01 7.67615616e-01 -3.74047279e-01
-3.45049918e-01 8.92461017e-02 -7.91974664e-01 1.59733579e-01
7.09050536e-01 -5.92726409e-01 -1.22425091e+00 4.03349578e-01
9.73866045e-01 -4.08788413e-01 -7.64197528e-01 4.17142585e-02
6.91371262e-02 -6.88430309e-01 1.66534483e-01 -5.46843588e-01
7.71690726e-01 -7.50741661e-01 1.15038693e-01 -1.63596296e+00
1.33838847e-01 -6.65131390e-01 -2.75429152e-02 1.36239672e+00
6.57069504e-01 -2.52901763e-01 4.12490755e-01 6.93326771e-01
-4.30681497e-01 -2.56933749e-01 -4.94831592e-01 -5.35050094e-01
-4.44189876e-01 -1.17554151e-01 7.32984245e-01 8.73126090e-01
7.51411855e-01 3.57083201e-01 8.49826857e-02 1.33568838e-01
4.33025569e-01 3.40905756e-01 9.14622128e-01 -1.35307825e+00
-1.63650289e-01 -4.87541586e-01 -3.87036920e-01 -6.94936156e-01
3.08272690e-01 -1.09754717e+00 3.29361498e-01 -2.01577520e+00
9.89964068e-01 2.40062267e-01 3.70446414e-01 4.57741946e-01
-4.48384881e-01 -6.40924871e-01 -1.22820571e-01 3.77425402e-01
-9.83906269e-01 4.02236104e-01 7.65739501e-01 -4.73889738e-01
-3.63444239e-01 1.84334591e-01 -9.34065282e-01 4.66836780e-01
7.27214456e-01 -7.34877646e-01 -6.48042262e-01 -5.88834882e-01
6.00018919e-01 4.93842363e-01 -8.94758403e-02 -1.13433492e+00
5.28366148e-01 -1.24056280e-01 -3.66040021e-01 -4.57912654e-01
-4.08563137e-01 -6.84180677e-01 5.74735940e-01 8.35164905e-01
-8.38102758e-01 6.08675539e-01 5.06634116e-01 6.59768641e-01
-4.83904839e-01 -8.30077767e-01 4.08305496e-01 -1.84420049e-01
-1.02650940e+00 -1.16666019e-01 -4.15683985e-01 2.28974566e-01
1.32156742e+00 -2.72222012e-01 -7.40442991e-01 -2.66033918e-01
-2.65052766e-01 3.12793702e-01 7.83811033e-01 7.52168953e-01
7.50262082e-01 -8.71295273e-01 -7.92446434e-01 -2.33540833e-01
7.37211704e-01 2.08147783e-02 5.29685691e-02 1.63484424e-01
-8.46035779e-01 4.71909165e-01 -2.75518149e-01 -2.32280225e-01
-1.39522648e+00 4.68563110e-01 2.54285574e-01 -2.11639658e-01
-6.61284924e-01 3.32235456e-01 -2.85504967e-01 -4.09471840e-02
1.07942536e-01 -4.13864970e-01 -7.78131902e-01 -2.57178158e-01
9.36145425e-01 3.30242187e-01 1.87643766e-01 -2.39229158e-01
-2.35960886e-01 4.51423466e-01 -2.38460138e-01 4.48539972e-01
1.51577353e+00 -2.59734727e-02 -6.37275934e-01 8.10609400e-01
1.06771123e+00 -1.29805267e-01 -7.25415409e-01 -1.90452840e-02
6.24452293e-01 -3.00636202e-01 -2.51269341e-01 -1.04321957e+00
-4.25905764e-01 4.87211108e-01 1.14119276e-01 3.96891713e-01
9.14048254e-01 1.21304072e-01 5.27749300e-01 7.65081048e-01
4.43700016e-01 -7.40762889e-01 1.21088654e-01 4.29945797e-01
1.22197700e+00 -1.09116185e+00 3.46945785e-02 -4.67143416e-01
-5.62122226e-01 1.50116420e+00 8.02476346e-01 3.54833990e-01
1.81556255e-01 3.00597884e-02 -1.64281845e-01 -4.38189149e-01
-8.24609697e-01 4.83584553e-01 3.00735772e-01 6.85501695e-01
5.84629595e-01 2.81589478e-02 -4.08358902e-01 6.84479833e-01
-1.55348077e-01 2.17478737e-01 1.20802259e+00 1.23499072e+00
-7.26802945e-01 -9.65477228e-01 -4.21656929e-02 7.54369438e-01
-1.31482631e-01 -3.93957831e-02 -7.62605131e-01 5.08031547e-01
-4.88395184e-01 7.99160361e-01 -2.96725094e-01 -1.79948434e-01
7.45721996e-01 2.53621280e-01 2.52641529e-01 -1.35502243e+00
-7.64852643e-01 -9.32541132e-01 2.44355187e-01 -4.79097039e-01
-3.86590987e-01 -5.48800230e-01 -1.50979948e+00 9.14333388e-02
-2.50234336e-01 7.46017635e-01 7.51096427e-01 8.25587690e-01
5.81095159e-01 8.28593314e-01 3.26893777e-01 -3.51118267e-01
-2.91814864e-01 -1.02897704e+00 1.44447580e-01 7.65221894e-01
3.97398025e-01 -2.26552099e-01 -1.89114407e-01 7.63952553e-01] | [7.623711109161377, 7.939359188079834] |
38a89164-2a1b-403b-846b-e8832175f275 | object-preserving-siamese-network-for-single | 2301.12057 | null | https://arxiv.org/abs/2301.12057v1 | https://arxiv.org/pdf/2301.12057v1.pdf | Object Preserving Siamese Network for Single Object Tracking on Point Clouds | Obviously, the object is the key factor of the 3D single object tracking (SOT) task. However, previous Siamese-based trackers overlook the negative effects brought by randomly dropped object points during backbone sampling, which hinder trackers to predict accurate bounding boxes (BBoxes). Exploring an approach that seeks to maximize the preservation of object points and their object-aware features is of particular significance. Motivated by this, we propose an Object Preserving Siamese Network (OPSNet), which can significantly maintain object integrity and boost tracking performance. Firstly, the object highlighting module enhances the object-aware features and extracts discriminative features from template and search area. Then, the object-preserved sampling selects object candidates to obtain object-preserved search area seeds and drop the background points that contribute less to tracking. Finally, the object localization network precisely locates 3D BBoxes based on the object-preserved search area seeds. Extensive experiments demonstrate our method outperforms the state-of-the-art performance (9.4% and 2.5% success gain on KITTI and Waymo Open Dataset respectively). | ['Zhengwei Hu', 'Jingchao Peng', 'Zhongze Wang', 'Haitao Zhao', 'Kaijie Zhao'] | 2023-01-28 | null | null | null | null | ['3d-single-object-tracking'] | ['computer-vision'] | [-3.23770136e-01 -1.94923401e-01 -3.87681931e-01 5.70612736e-02
-5.43014705e-01 -6.06815577e-01 3.52178603e-01 -1.43332496e-01
-4.84884113e-01 5.80555320e-01 -1.51235744e-01 3.35988045e-01
-2.13973388e-01 -2.69428879e-01 -7.75616884e-01 -8.04655612e-01
-3.67809087e-01 4.91010576e-01 1.13783658e+00 1.14829473e-01
3.31455320e-01 9.57544148e-01 -1.41468859e+00 -3.57602328e-01
7.43213117e-01 1.25098431e+00 2.58534938e-01 2.10519284e-01
-2.38269139e-02 4.04273480e-01 -8.33013296e-01 -9.87519175e-02
6.77866817e-01 1.31107837e-01 5.54435253e-02 -1.50704846e-01
1.10573161e+00 -2.84599334e-01 -5.27579129e-01 1.30971766e+00
5.12953043e-01 9.23642814e-02 3.33753079e-01 -1.65865242e+00
-4.15687114e-01 2.08822206e-01 -9.53909874e-01 5.32936513e-01
-1.65817469e-01 4.91699785e-01 7.32330739e-01 -9.31001186e-01
6.53812289e-01 1.31076074e+00 6.77302182e-01 6.93934321e-01
-1.02132583e+00 -9.94351387e-01 3.25602382e-01 1.52764991e-02
-1.62456727e+00 -4.17298496e-01 8.28745902e-01 -2.99729824e-01
5.25138438e-01 3.88156503e-01 7.62099206e-01 6.13698006e-01
2.00067222e-01 1.03637683e+00 6.38725519e-01 1.63091481e-01
-1.15755081e-01 1.33625790e-01 1.53683081e-01 6.97608471e-01
8.79013717e-01 4.49816585e-01 -8.09717417e-01 -1.33059382e-01
6.71101570e-01 2.15303361e-01 -2.34502852e-01 -1.07669580e+00
-1.45659244e+00 3.63249779e-01 9.41997766e-01 7.78426751e-02
-2.15462402e-01 2.74214834e-01 1.02556907e-01 -1.81101933e-01
2.05890760e-01 1.37793094e-01 -2.67428398e-01 4.94640470e-02
-1.08426321e+00 4.94876683e-01 2.12415069e-01 1.40001118e+00
5.62933624e-01 1.54472277e-01 -6.72995269e-01 3.54908824e-01
6.28156364e-01 1.00322115e+00 -2.18972743e-01 -7.49601185e-01
4.55642343e-01 9.79679286e-01 4.06709969e-01 -1.04959846e+00
-2.26712748e-01 -8.89706016e-01 -4.19148922e-01 3.61981630e-01
5.68404198e-01 8.42707679e-02 -9.59569573e-01 1.64221942e+00
1.01983678e+00 3.04832190e-01 -3.99842680e-01 1.38187063e+00
7.82766640e-01 3.39176744e-01 1.13367677e-01 -5.16168848e-02
1.40137661e+00 -9.55869436e-01 -4.83751476e-01 -2.40890637e-01
3.67597729e-01 -7.24824607e-01 4.78338778e-01 -1.49499640e-01
-1.00017869e+00 -6.24886513e-01 -9.61851656e-01 1.89095899e-01
-1.86475456e-01 3.23273867e-01 2.08280206e-01 5.56031764e-01
-6.62284553e-01 3.50214899e-01 -8.94434988e-01 -2.07750216e-01
9.96140122e-01 5.71683228e-01 -9.48965624e-02 2.31647447e-01
-6.56295180e-01 8.81525159e-01 3.72561753e-01 1.98506638e-01
-1.26792538e+00 -1.11566198e+00 -5.21942139e-01 -7.91059136e-02
8.44131649e-01 -3.99725348e-01 8.36367071e-01 -3.47188741e-01
-9.36085820e-01 7.06483901e-01 -6.60751686e-02 -6.23149395e-01
7.12288678e-01 -3.44264895e-01 -1.35999218e-01 1.21310435e-01
2.62156039e-01 9.52226281e-01 9.96327877e-01 -1.32261145e+00
-1.12158000e+00 -7.02259779e-01 -3.63106489e-01 1.40259042e-01
-1.91797614e-01 9.93572697e-02 -6.80025518e-01 -6.23540521e-01
3.29076499e-01 -1.02827883e+00 -4.96294051e-02 8.90642941e-01
-4.38041985e-01 -5.64046323e-01 1.45644641e+00 -3.89790088e-01
1.16600728e+00 -2.28216338e+00 -1.51650414e-01 1.85332328e-01
6.09290957e-01 3.97143066e-01 -9.95151605e-03 -1.55907691e-01
3.65215272e-01 -2.40502670e-01 3.44775230e-01 -1.83494925e-01
9.66621786e-02 -1.50858939e-01 -1.35254338e-01 8.19886088e-01
1.93076983e-01 1.13859129e+00 -8.20493579e-01 -9.22597468e-01
2.95611471e-01 3.81414920e-01 -3.31924975e-01 -4.89680655e-02
-3.37104768e-01 2.92613626e-01 -1.00624645e+00 1.16801691e+00
1.02191544e+00 -9.85013992e-02 -3.53516757e-01 -4.71169919e-01
-5.42217195e-01 -1.07567556e-01 -1.24993813e+00 1.34550714e+00
4.60036159e-01 4.78667200e-01 1.62928611e-01 -2.61168212e-01
1.03956842e+00 -2.34827384e-01 7.53793597e-01 -5.11833847e-01
1.44911185e-01 2.26092100e-01 2.28749681e-02 -1.17190845e-01
6.05819106e-01 4.22435135e-01 2.09588572e-01 -4.74272016e-03
-2.38757059e-01 3.97900701e-01 5.21567604e-03 3.34400475e-01
8.70321810e-01 4.00695741e-01 -3.15117277e-02 -3.98105174e-01
4.69709903e-01 2.88525522e-01 7.91320860e-01 7.13634849e-01
-9.27301228e-01 2.15627626e-01 1.13701656e-01 -4.41007704e-01
-8.50702286e-01 -1.15351462e+00 -2.19031990e-01 9.00561631e-01
9.02025402e-01 -9.84014720e-02 -4.81645107e-01 -9.23077822e-01
5.66144645e-01 3.08473945e-01 -5.51891088e-01 -2.94494480e-01
-8.60876977e-01 -1.03771135e-01 4.41404104e-01 5.42874992e-01
4.00557846e-01 -8.75661194e-01 -1.01438963e+00 3.96985561e-02
1.13191083e-01 -8.77286255e-01 -9.14852500e-01 -2.65115034e-02
-9.82199252e-01 -1.07711077e+00 -8.51834178e-01 -6.17554128e-01
7.33888626e-01 7.91708529e-01 6.65324926e-01 1.58323839e-01
-2.11578012e-01 4.30050194e-02 -1.30695879e-01 -4.57833111e-01
1.26175761e-01 9.20429677e-02 2.31571376e-01 -2.15849318e-02
4.88227427e-01 -4.69640233e-02 -7.84208536e-01 6.91990674e-01
-3.10043842e-01 -3.15872669e-01 7.16239095e-01 3.24318230e-01
8.63809168e-01 -1.18094593e-01 8.06895122e-02 2.90980116e-02
-1.08894862e-01 -1.76464066e-01 -1.19121480e+00 2.38317922e-01
-3.19815546e-01 -8.77144411e-02 1.00911379e-01 -8.48867655e-01
-6.70511246e-01 2.78119683e-01 5.90984166e-01 -1.09006310e+00
1.26529902e-01 -4.30314779e-01 -2.26541102e-01 -7.28073657e-01
3.65377426e-01 3.65375459e-01 1.48380762e-02 -5.87727308e-01
1.70239657e-02 1.92453668e-01 4.69227999e-01 -3.52103412e-01
1.39443123e+00 8.38906348e-01 2.22832873e-01 -5.60555816e-01
-9.76359904e-01 -4.83126253e-01 -4.77465779e-01 -7.29830444e-01
8.43753219e-01 -9.23017502e-01 -1.05295515e+00 1.31283224e-01
-1.00947022e+00 2.67039090e-01 -4.18488592e-01 5.40767789e-01
-1.81723878e-01 1.29645109e-01 -7.94603974e-02 -1.17189157e+00
-3.85419041e-01 -9.60743666e-01 1.42824042e+00 7.53611386e-01
2.37161145e-01 -2.99596071e-01 -1.21114150e-01 4.95754965e-02
3.67098123e-01 2.66763926e-01 1.86306119e-01 -5.55794179e-01
-1.54581547e+00 -3.29336673e-01 -5.18955529e-01 -2.91377157e-01
-2.02701688e-01 -6.06353097e-02 -6.51203871e-01 -6.85841560e-01
-2.90459394e-01 1.21658981e-01 7.05270767e-01 6.77554607e-01
9.37057674e-01 1.13612644e-01 -1.05247974e+00 6.37644351e-01
1.20454228e+00 1.30658701e-01 1.98506057e-01 3.56062442e-01
6.74129248e-01 2.60269135e-01 1.17771268e+00 3.03982645e-01
1.49943098e-01 9.72897053e-01 8.90968442e-01 1.29401997e-01
-3.93019229e-01 -5.61095834e-01 3.51864368e-01 1.61760285e-01
1.31378070e-01 1.02021053e-01 -5.16633153e-01 7.82324135e-01
-1.78568041e+00 -7.97003031e-01 -4.00202066e-01 2.21679354e+00
4.91643429e-01 5.29353499e-01 4.37885910e-01 -4.36026812e-01
1.12750578e+00 2.80641764e-01 -8.44515026e-01 9.69390035e-01
-1.59902811e-01 -3.84421527e-01 8.96057308e-01 9.69154239e-02
-1.16853988e+00 1.03833413e+00 4.86389637e+00 1.14047468e+00
-6.61440551e-01 1.47235066e-01 -1.24023952e-01 -6.32723033e-01
3.31562847e-01 2.27840780e-03 -1.77710533e+00 6.86803222e-01
3.39484602e-01 -1.49497241e-01 -6.73587993e-03 1.27884948e+00
2.43601665e-01 -6.46428987e-02 -8.14359665e-01 8.53382051e-01
-3.16661261e-02 -1.16989219e+00 -1.47359639e-01 2.61781961e-01
5.74390888e-01 1.04812413e-01 1.21688254e-01 1.59372136e-01
-3.45988907e-02 -2.72848368e-01 1.26770127e+00 5.62701225e-01
4.49061632e-01 -6.13593936e-01 4.14401859e-01 2.54083365e-01
-1.75965941e+00 -8.03118646e-02 -6.25299156e-01 5.57630181e-01
2.44734511e-01 3.62348378e-01 -6.33575559e-01 3.28026325e-01
1.04401457e+00 6.89229846e-01 -7.79860020e-01 1.87482393e+00
5.62015027e-02 3.37844014e-01 -7.70036340e-01 -3.05508524e-01
3.60220611e-01 9.10639316e-02 1.33760774e+00 7.24125504e-01
2.35969618e-01 -1.52793750e-01 3.76511633e-01 1.04664683e+00
2.71241628e-02 -2.52112359e-01 -4.05511290e-01 2.04944894e-01
7.56935954e-01 1.25137150e+00 -8.86665642e-01 -2.26798177e-01
-1.06604405e-01 3.76534283e-01 2.07608029e-01 1.38860568e-01
-1.27855384e+00 -2.45771691e-01 8.61303866e-01 3.70434135e-01
8.52556944e-01 -3.27976823e-01 8.64061806e-03 -8.91748130e-01
3.21922153e-01 -4.97067362e-01 3.07531893e-01 -6.12369418e-01
-1.11751866e+00 3.34496528e-01 -6.55737743e-02 -1.70949578e+00
7.22414792e-01 -4.50952530e-01 -3.48220974e-01 6.15427315e-01
-1.55880761e+00 -1.27873135e+00 -6.17552757e-01 5.62677920e-01
4.98371512e-01 -7.74637738e-04 -2.85203040e-01 6.27362430e-01
-5.00319421e-01 5.92337430e-01 -5.30208424e-02 9.61735845e-02
6.57179117e-01 -8.32886279e-01 2.21509680e-01 8.61254930e-01
1.06889315e-01 7.47525394e-01 6.06486559e-01 -1.27015555e+00
-1.59652960e+00 -1.25906134e+00 6.77351773e-01 -8.50233972e-01
6.24412298e-01 -3.78000528e-01 -8.07344258e-01 5.05666733e-01
-3.43449205e-01 3.38479489e-01 -1.09784462e-01 -3.79622132e-01
-7.04208240e-02 -3.53414118e-01 -1.04322779e+00 6.99666858e-01
1.54137373e+00 1.47867411e-01 -4.87695068e-01 2.75130928e-01
8.51904929e-01 -6.24462724e-01 -6.28643751e-01 4.14208502e-01
6.43294930e-01 -6.29495084e-01 1.24754381e+00 -6.01156235e-01
-4.33661371e-01 -1.18993926e+00 1.41674474e-01 -6.72713339e-01
-3.99990767e-01 -6.53940678e-01 -7.38712907e-01 1.11506951e+00
1.37623772e-02 -3.34166080e-01 1.20444429e+00 2.85212457e-01
-2.67753571e-01 -5.33462584e-01 -1.32304156e+00 -1.18927109e+00
-4.85600263e-01 4.62333672e-02 6.19396031e-01 4.85763818e-01
-6.95044219e-01 -1.11392200e-01 -4.33230966e-01 4.50583369e-01
1.40961850e+00 3.08179080e-01 1.05404449e+00 -1.47126281e+00
4.59429353e-01 -5.19728363e-01 -4.63260949e-01 -1.51260364e+00
-2.27491438e-01 -7.42771924e-01 1.85250387e-01 -9.40787315e-01
2.38126963e-01 -8.39356184e-01 -3.60157073e-01 3.55935246e-01
-2.85157025e-01 3.13199282e-01 5.90820193e-01 4.84694898e-01
-1.27937198e+00 6.99586034e-01 1.51661873e+00 -9.53610986e-02
-1.79072067e-01 2.69624323e-01 -3.48961443e-01 4.27900791e-01
3.96766305e-01 -1.00531685e+00 1.50451317e-01 -2.40324691e-01
-3.65254134e-01 -3.42061520e-01 8.49323273e-01 -1.23802304e+00
6.96510196e-01 -1.66065991e-01 6.42928720e-01 -1.52161205e+00
3.61525953e-01 -1.29316878e+00 2.22243965e-01 8.32767010e-01
-4.38601114e-02 -2.28859574e-01 3.02301675e-01 7.73806512e-01
2.14531213e-01 -1.37048051e-01 1.04545081e+00 1.47440732e-01
-6.82413876e-01 7.31532812e-01 2.08766177e-01 6.53694421e-02
1.49531019e+00 -7.07201660e-01 -3.44787329e-01 3.42375845e-01
-2.90379018e-01 6.82024539e-01 4.85014319e-01 6.62657738e-01
5.11827767e-01 -1.56897032e+00 -5.15442014e-01 2.71848291e-01
5.92013486e-02 9.48284641e-02 1.12059116e-01 1.22881997e+00
-1.64192438e-01 5.55843234e-01 -1.71810970e-01 -1.13568223e+00
-1.55280507e+00 5.88611484e-01 4.30026501e-01 1.13410456e-02
-8.40945363e-01 9.77313876e-01 1.36295035e-01 1.49634406e-02
6.16871834e-01 -3.60122740e-01 1.10231087e-01 -1.23486288e-01
4.81529444e-01 5.52937210e-01 -4.17955130e-01 -6.94464743e-01
-7.57133305e-01 8.70680869e-01 -1.74768820e-01 3.93963516e-01
1.19236135e+00 -1.01152882e-01 2.26449862e-01 -1.73489407e-01
7.11104453e-01 6.20404258e-02 -1.90160918e+00 -4.18866605e-01
1.08793601e-01 -1.01957715e+00 2.34257039e-02 -5.58029830e-01
-1.24451149e+00 5.59106350e-01 8.53518426e-01 1.37720525e-01
5.70102036e-01 2.73146212e-01 8.28078926e-01 1.92345440e-01
5.21871090e-01 -9.46156085e-01 1.48101136e-01 2.57370710e-01
6.10238314e-01 -1.13270819e+00 2.22936958e-01 -4.10193771e-01
-3.22965443e-01 6.33774698e-01 1.15148020e+00 -2.27344930e-01
3.16621810e-01 3.12448919e-01 -2.55134940e-01 -4.55819160e-01
-4.05606449e-01 -4.19410527e-01 5.53117573e-01 7.29168594e-01
-3.56433213e-01 -3.34144354e-01 -8.98393989e-03 3.78884345e-01
1.54924139e-01 -2.07541749e-01 -3.12514305e-01 1.04863548e+00
-9.70710397e-01 -5.19205570e-01 -8.31176102e-01 4.31902856e-01
-1.54333472e-01 2.49939665e-01 -4.35689360e-01 1.08202577e+00
3.12487125e-01 3.36778343e-01 -2.03942992e-02 -1.26922384e-01
4.91846383e-01 -3.19504112e-01 3.86782527e-01 -1.59504414e-01
-6.11678481e-01 2.94353634e-01 -2.77352720e-01 -7.32648134e-01
-3.78627628e-01 -7.76582837e-01 -1.17713451e+00 -1.64384395e-01
-8.70718122e-01 1.15565740e-01 6.48430526e-01 5.38353562e-01
5.45232534e-01 4.16233480e-01 3.04345340e-01 -1.00789011e+00
-7.41313219e-01 -6.86180294e-01 -4.87862647e-01 1.76887438e-01
4.50815588e-01 -1.30934048e+00 -3.27190787e-01 -2.61422426e-01] | [6.498702049255371, -2.259793758392334] |
21accc8a-9abe-4bd0-b8c1-57fd5e33434d | pdq-tmk-pdqf-a-test-drive-of-facebooks | 1912.07745 | null | https://arxiv.org/abs/1912.07745v1 | https://arxiv.org/pdf/1912.07745v1.pdf | PDQ & TMK + PDQF -- A Test Drive of Facebook's Perceptual Hashing Algorithms | Efficient and reliable automated detection of modified image and multimedia files has long been a challenge for law enforcement, compounded by the harm caused by repeated exposure to psychologically harmful materials. In August 2019 Facebook open-sourced their PDQ and TMK + PDQF algorithms for image and video similarity measurement, respectively. In this report, we review the algorithms' performance on detecting commonly encountered transformations on real-world case data, sourced from contemporary investigations. We also provide a reference implementation to demonstrate the potential application and integration of such algorithms within existing law enforcement systems. | ['Janis Dalins', 'Douglas Boudry', 'Campbell Wilson'] | 2019-12-16 | null | null | null | null | ['video-similarity'] | ['computer-vision'] | [ 4.88544703e-01 -2.74160177e-01 -2.29189172e-01 -1.79131135e-01
-7.64962316e-01 -7.79912472e-01 4.94392961e-01 5.67096949e-01
-5.01835346e-01 3.94763589e-01 2.04757527e-01 -1.81036845e-01
-4.96867687e-01 -4.41608936e-01 -1.39198110e-01 -2.40473315e-01
-6.27597347e-02 2.83272509e-02 1.80825651e-01 -3.92211042e-02
1.05069768e+00 7.18919575e-01 -1.54006612e+00 3.48686159e-01
6.31629884e-01 5.30129492e-01 -1.10013625e-02 6.99470282e-01
2.94119835e-01 6.74471140e-01 -8.43497872e-01 -1.25339818e+00
3.79485816e-01 -9.19788182e-02 -8.22842896e-01 2.65431672e-01
8.24266553e-01 -9.02859330e-01 -5.82379639e-01 1.32740390e+00
7.18851924e-01 2.77765095e-01 5.47739685e-01 -1.37007618e+00
-6.59385741e-01 -4.05765399e-02 -6.81383073e-01 9.11029696e-01
1.07297909e+00 4.40852076e-01 5.60835779e-01 -5.52439928e-01
1.14354241e+00 1.35482371e+00 6.28055990e-01 1.90294012e-01
-1.05616343e+00 -5.75622022e-01 -5.43080688e-01 8.42030108e-01
-1.48143423e+00 -6.49387419e-01 5.51129043e-01 -6.57681525e-01
1.27564526e+00 5.83758652e-01 4.77674872e-01 1.00448024e+00
8.72569606e-02 3.30981016e-01 8.36156785e-01 -2.01576397e-01
-7.98094347e-02 4.13935632e-01 2.21048385e-01 6.65916681e-01
4.69216317e-01 8.77066180e-02 -4.55338866e-01 -9.02555406e-01
3.69122684e-01 -2.06988066e-01 -8.18752125e-02 3.81308973e-01
-5.71720779e-01 7.37469256e-01 -2.69049585e-01 5.02579749e-01
-4.22342420e-01 -2.78936058e-01 6.50056660e-01 1.46457732e-01
4.02581573e-01 4.94703948e-01 5.87441206e-01 -7.56853163e-01
-1.01334929e+00 3.11128020e-01 2.42275566e-01 5.65542817e-01
5.61107397e-01 -4.65064943e-01 -4.18296367e-01 8.51852953e-01
-7.75408000e-02 5.30786276e-01 3.47507149e-01 -1.35727382e+00
3.81713271e-01 4.78134155e-01 1.30305707e-01 -1.90450025e+00
-9.02787745e-02 2.41828989e-02 -2.00388283e-01 2.06207559e-01
1.65834039e-01 4.19558644e-01 -3.06203902e-01 1.07562387e+00
3.07536364e-01 3.13473433e-01 -4.38099086e-01 3.12666357e-01
4.63611662e-01 8.06302428e-02 2.70558268e-01 -6.14013433e-01
9.70667064e-01 -3.18458110e-01 -1.07704520e+00 1.22992449e-01
6.00408673e-01 -1.04772258e+00 1.04668117e+00 5.52550375e-01
-1.34608459e+00 -2.63010591e-01 -7.67400861e-01 -2.49363426e-02
-3.92864257e-01 -4.26400006e-01 3.69327873e-01 1.34588575e+00
-7.87721574e-01 9.41083252e-01 -2.43041858e-01 -7.59686172e-01
9.46640670e-01 1.30054265e-01 -5.54360509e-01 -3.23737809e-03
-1.06040597e+00 1.15275359e+00 3.53290021e-01 -4.08129722e-01
-3.86273175e-01 -3.90365601e-01 -4.64969635e-01 -2.14973837e-01
4.42602754e-01 -1.09969929e-01 1.12537503e+00 -3.94412160e-01
-9.79545057e-01 1.68523669e+00 -1.01045862e-01 -3.39486778e-01
4.12489384e-01 6.44380152e-02 -8.11259031e-01 7.15662479e-01
3.27619165e-01 1.47751316e-01 1.09614480e+00 -6.50826395e-01
-4.59560186e-01 -4.28209394e-01 6.38884604e-02 8.37786272e-02
-7.39337504e-01 9.42194164e-01 -1.61848098e-01 -6.06328964e-01
-6.14870131e-01 -6.57857180e-01 2.21292023e-02 1.35510499e-02
-3.10905457e-01 -7.66558573e-02 7.85041928e-01 -1.06786454e+00
1.76079726e+00 -2.61190748e+00 -5.73859394e-01 4.73961145e-01
1.96256593e-01 9.36913192e-01 -1.01358347e-01 5.30888975e-01
-5.14482558e-02 3.36694419e-01 -4.03258167e-02 6.26130626e-02
7.07651973e-02 7.50309676e-02 -1.87427223e-01 8.46745431e-01
-2.86677301e-01 6.54752016e-01 -1.02510798e+00 -7.36390531e-01
2.57680684e-01 1.59785002e-02 -4.14605111e-01 -5.43180071e-02
4.88829255e-01 -5.13671152e-02 -3.53105217e-02 7.37469256e-01
9.12803292e-01 2.95831889e-01 -7.39291161e-02 -1.62429716e-02
-1.02776058e-01 1.17220685e-01 -9.33577180e-01 1.19568872e+00
2.30078787e-01 7.92184889e-01 1.09922603e-01 -5.76902449e-01
5.57862163e-01 3.57824445e-01 3.93909574e-01 -7.94446349e-01
3.75549883e-01 -1.87537391e-02 -3.50416452e-01 -1.26718879e+00
7.69979835e-01 -1.09166652e-01 1.98236883e-01 8.11456859e-01
-2.59148866e-01 -7.53361285e-02 5.16519725e-01 3.30208391e-01
1.34928048e+00 -6.06053889e-01 7.38244355e-01 4.53022212e-01
4.27984834e-01 8.53713453e-02 3.38926971e-01 8.23957920e-01
-1.04746521e+00 1.47884756e-01 4.28540319e-01 -7.87462369e-02
-8.80255342e-01 -9.32407498e-01 -1.54331401e-01 9.60409641e-01
9.15841237e-02 -6.86086416e-01 -1.04278672e+00 -6.56044245e-01
-1.85221471e-02 9.80195284e-01 -2.98254102e-01 -4.16892737e-01
-1.91374585e-01 -5.15187085e-01 9.18424547e-01 -1.43881306e-01
5.12411535e-01 -1.14235711e+00 -6.33678555e-01 1.33533940e-01
-4.51383233e-01 -1.24908948e+00 -4.89860028e-01 -1.03700995e+00
-3.82361710e-01 -1.56811380e+00 -3.00788432e-02 -1.60269514e-01
3.27793479e-01 8.48814368e-01 4.61272061e-01 4.44403023e-01
-9.73809838e-01 8.32463026e-01 -3.00681025e-01 -3.54724169e-01
-7.19660163e-01 -5.41570902e-01 3.70272964e-01 1.12419359e-01
7.30898082e-01 -5.57998538e-01 -1.89367756e-01 2.49853626e-01
-1.21372747e+00 -4.98583347e-01 -8.83722082e-02 -1.16762377e-01
4.57491465e-02 3.43407452e-01 3.13741177e-01 -7.43651152e-01
1.13831592e+00 -4.63397920e-01 -2.96978652e-01 2.01137289e-01
-6.02751136e-01 -8.49519968e-01 -6.81528896e-02 -2.30742320e-01
-1.07867932e+00 -2.56420314e-01 -1.07107754e-03 -3.22249770e-01
-4.27101731e-01 2.86507428e-01 1.20904692e-01 -3.10341626e-01
9.86473501e-01 -1.47786245e-01 2.96524838e-02 -2.78065711e-01
1.79855675e-01 1.14394653e+00 8.17852259e-01 8.75432417e-02
1.07966912e+00 6.12923801e-01 -1.97666660e-01 -9.63917792e-01
-5.01663566e-01 -7.55066812e-01 -6.01248801e-01 -6.49271369e-01
7.82445788e-01 -2.41611049e-01 -9.74503756e-01 6.41222596e-01
-1.27573931e+00 1.78547591e-01 5.06891385e-02 3.46744001e-01
-5.20736635e-01 1.12567115e+00 -3.91762227e-01 -8.45496655e-01
-2.84617245e-01 -7.50030637e-01 3.51698637e-01 -6.24415167e-02
-6.23752177e-01 -6.87279463e-01 3.27731013e-01 9.55382407e-01
3.18716496e-01 4.20041531e-01 7.21904755e-01 -6.42235041e-01
7.08449306e-03 -5.52473843e-01 -2.68745869e-01 6.04785800e-01
2.29606733e-01 2.04359248e-01 -1.01540112e+00 -1.56749815e-01
1.88707381e-01 -8.26772898e-02 2.70312756e-01 -7.35331476e-02
1.03591573e+00 -4.95494723e-01 -2.78819948e-01 2.11442426e-01
1.08074498e+00 5.16069055e-01 9.80966747e-01 5.67191899e-01
2.02454597e-01 6.50947571e-01 8.79760087e-01 7.36189008e-01
-2.40127668e-01 7.26686358e-01 3.58363569e-01 3.46489817e-01
-1.14496119e-01 -1.09731063e-01 3.54557544e-01 1.61073789e-01
-4.31425422e-01 -1.61956340e-01 -8.38561535e-01 1.75496787e-01
-1.66407752e+00 -1.90859377e+00 -2.93957919e-01 2.20977044e+00
3.06050867e-01 8.36774800e-03 3.68793607e-01 3.89292628e-01
1.41948247e+00 4.34217677e-02 -3.75294060e-01 -6.51785791e-01
-1.35076478e-01 4.48686257e-02 2.95357198e-01 3.92439455e-01
-1.07779300e+00 6.47633016e-01 7.78521872e+00 1.35913122e+00
-6.81726396e-01 4.94486451e-01 4.61708724e-01 -3.88033569e-01
2.36593083e-01 -3.53033185e-01 -3.23521286e-01 5.16570747e-01
1.16187739e+00 -7.27884114e-01 5.69270492e-01 7.98998475e-01
6.05409265e-01 -3.64786088e-01 -6.84740782e-01 1.37290478e+00
5.46474636e-01 -1.30330789e+00 -1.84783772e-01 2.82818973e-01
2.30769783e-01 -4.25959080e-01 3.58531088e-01 -8.20921883e-02
-5.27843893e-01 -7.72331953e-01 3.86597008e-01 1.97597727e-01
9.09814656e-01 -6.05930984e-01 3.17078322e-01 1.85534835e-01
-5.99517047e-01 -2.17585698e-01 -2.84732699e-01 -1.56319946e-01
3.06002289e-01 4.04899240e-01 -1.08686948e+00 2.07372203e-01
4.28050041e-01 7.66534567e-01 -7.96326101e-01 1.23547113e+00
-1.16611660e-01 4.62678134e-01 1.27349589e-02 2.98320979e-01
4.53495048e-02 -6.17798604e-02 1.00046146e+00 1.45277619e+00
2.65500814e-01 3.52098793e-01 -2.52421945e-01 5.91686308e-01
2.73850709e-01 2.23089412e-01 -9.29280102e-01 -1.96655571e-01
5.77338338e-01 1.41984284e+00 -6.68436885e-01 -4.48589057e-01
-2.05300197e-01 1.17696273e+00 2.40633234e-01 -2.48137768e-02
-9.90478754e-01 -4.51861888e-01 6.85564220e-01 4.96967673e-01
-2.46291429e-01 -7.91364163e-02 5.70077971e-02 -8.42760921e-01
-1.16907038e-01 -9.01903629e-01 7.73311853e-01 -8.15329492e-01
-1.26183605e+00 3.21201146e-01 4.68157887e-01 -1.31707585e+00
-2.48468935e-01 -4.00136560e-01 -7.56034791e-01 3.97704542e-01
-8.12462330e-01 -6.42712653e-01 -1.41492322e-01 8.77028108e-01
9.62492526e-02 -3.64917308e-01 7.65380204e-01 5.82217634e-01
-6.35238528e-01 6.58004761e-01 -1.20872132e-01 -2.08116531e-01
9.48760092e-01 -6.08821154e-01 9.97232124e-02 9.86281991e-01
-1.35859326e-01 6.94457650e-01 6.66101098e-01 -8.79421651e-01
-9.81157660e-01 -9.24752295e-01 9.19494510e-01 -4.66375977e-01
7.66617954e-01 -1.47764340e-01 -8.99819195e-01 5.80899775e-01
2.93862164e-01 -5.60651064e-01 1.20954871e+00 -3.61777693e-01
-5.07018864e-01 1.98261470e-01 -1.93862700e+00 5.91116130e-01
1.30552304e+00 -7.63773561e-01 -6.62998199e-01 9.90615070e-01
1.39899090e-01 -1.41108677e-01 -7.17273593e-01 -2.67786890e-01
3.47898096e-01 -1.55510426e+00 9.95395422e-01 -4.46384698e-01
1.00473717e-01 1.36883721e-01 1.90539420e-01 -6.80009782e-01
-7.56255746e-01 -1.14052689e+00 2.15412930e-01 1.22086048e+00
-1.07098266e-01 -5.24964333e-01 5.17371655e-01 9.92901444e-01
1.36666432e-01 -8.42332542e-02 -1.23591149e+00 -9.33496535e-01
-5.98470271e-01 -8.40264499e-01 2.93487936e-01 1.23571789e+00
6.13985598e-01 -1.83376998e-01 -3.78437072e-01 1.20672129e-01
5.71480691e-01 -4.51674491e-01 6.51379824e-01 -7.74942338e-01
-3.04503560e-01 -5.75527966e-01 -1.13653588e+00 7.77303204e-02
1.66543141e-01 -8.87792766e-01 -6.99161530e-01 -1.09658468e+00
7.89989054e-01 2.07379773e-01 6.90849265e-03 4.98797655e-01
-1.20827325e-01 7.78787553e-01 3.76815498e-01 2.17018783e-01
-5.26790023e-01 -2.73797810e-01 9.16914105e-01 -8.51037800e-02
8.48526135e-03 -1.75987512e-01 -7.29384482e-01 8.08815122e-01
8.59478533e-01 -6.43271625e-01 -4.33310300e-01 1.70062843e-03
3.93949181e-01 -2.02993706e-01 5.45269012e-01 -1.05917716e+00
1.32345095e-01 -5.54509401e-01 -1.23434789e-01 -4.38171178e-01
4.30950791e-01 -5.86199284e-01 3.35352778e-01 6.44170225e-01
-2.52400339e-01 1.23474896e-01 2.06900060e-01 5.20593822e-01
-5.70262549e-03 -8.31823945e-01 8.21884573e-01 -2.13935245e-02
-4.59477842e-01 1.64022177e-01 -9.50254261e-01 -3.49903047e-01
1.56756091e+00 -6.51858509e-01 -6.31506443e-01 -7.58269846e-01
-6.64651155e-01 -4.09693152e-01 4.13133949e-01 3.78000587e-01
7.94531405e-01 -1.19936121e+00 -5.07542789e-01 -1.50027499e-02
1.20023742e-01 -1.34600437e+00 4.97898310e-01 9.02444005e-01
-4.87134337e-01 1.64621487e-01 -5.75403392e-01 7.43042724e-03
-2.05859184e+00 1.05192804e+00 -8.72588828e-02 8.00121129e-02
-2.08707020e-01 3.87575358e-01 -5.53612530e-01 2.39778206e-01
5.99727072e-02 4.55781758e-01 -2.68736959e-01 6.24766424e-02
1.08238816e+00 1.02626002e+00 2.35575251e-02 -9.30351436e-01
-4.05896366e-01 2.48477280e-01 -2.40387946e-01 -1.60757318e-01
1.10452306e+00 -2.51298070e-01 -5.25055647e-01 -1.93971798e-01
1.36131740e+00 8.85411724e-02 -2.45951459e-01 5.52591383e-02
-9.60844457e-02 -1.32172000e+00 -1.16155781e-01 -5.45890331e-01
-6.56586111e-01 4.79795247e-01 6.87213480e-01 5.06134808e-01
1.15564334e+00 -2.21967176e-01 9.91695821e-01 4.44461882e-01
2.38829747e-01 -1.45845413e+00 3.40835690e-01 1.32617444e-01
9.23063397e-01 -9.68626380e-01 2.77447045e-01 -7.35258341e-01
-5.26642442e-01 1.15348256e+00 2.04927221e-01 1.42031670e-01
4.66010600e-01 -2.92736769e-01 -2.10556425e-02 -4.98837650e-01
-1.81984723e-01 1.26360074e-01 4.77689877e-02 9.16030467e-01
1.53820038e-01 -1.41548902e-01 -6.92252457e-01 1.16716385e-01
-2.31242076e-01 -9.22903121e-02 8.51393282e-01 1.16739202e+00
-6.32963896e-01 -1.16951776e+00 -8.58502030e-01 6.46626174e-01
-5.86882770e-01 1.27923816e-01 -6.08400285e-01 5.27395785e-01
4.09377337e-01 1.33876920e+00 -1.73746482e-01 -5.06948113e-01
3.83274615e-01 -2.43017569e-01 4.75087732e-01 -5.50946057e-01
-1.01161194e+00 -2.19941780e-01 3.53075445e-01 -8.75731111e-01
-6.30905688e-01 -1.12121236e+00 -7.61030734e-01 -7.20583797e-01
2.41171774e-02 -1.89902037e-01 5.81358790e-01 7.88082778e-01
5.49701214e-01 -3.53325933e-01 5.02116799e-01 -7.60128140e-01
-4.05680984e-01 -7.61947155e-01 -8.91354084e-01 9.89543915e-01
2.06712191e-03 -5.37539423e-01 -3.83785576e-01 1.44979060e-01] | [12.49780559539795, 1.0658634901046753] |
080acfa4-e501-4b06-8dfd-68e894b982ae | ammunition-component-classification-using | 2208.12863 | null | https://arxiv.org/abs/2208.12863v1 | https://arxiv.org/pdf/2208.12863v1.pdf | Ammunition Component Classification Using Deep Learning | Ammunition scrap inspection is an essential step in the process of recycling ammunition metal scrap. Most ammunition is composed of a number of components, including case, primer, powder, and projectile. Ammo scrap containing energetics is considered to be potentially dangerous and should be separated before the recycling process. Manually inspecting each piece of scrap is tedious and time-consuming. We have gathered a dataset of ammunition components with the goal of applying artificial intelligence for classifying safe and unsafe scrap pieces automatically. First, two training datasets are manually created from visual and x-ray images of ammo. Second, the x-ray dataset is augmented using the spatial transforms of histogram equalization, averaging, sharpening, power law, and Gaussian blurring in order to compensate for the lack of sufficient training data. Lastly, the representative YOLOv4 object detection method is applied to detect the ammo components and classify the scrap pieces into safe and unsafe classes, respectively. The trained models are tested against unseen data in order to evaluate the performance of the applied method. The experiments demonstrate the feasibility of ammo component detection and classification using deep learning. The datasets and the pre-trained models are available at https://github.com/hadi-ghnd/Scrap-Classification. | ['Hang Shi', 'Chengjun Liu', 'Hadi Ghahremannezhad'] | 2022-08-26 | null | null | null | null | ['component-classification'] | ['natural-language-processing'] | [ 9.52920243e-02 -4.48497325e-01 3.82673085e-01 -5.17255217e-02
-4.12958443e-01 -5.89019001e-01 4.13814723e-01 1.52868524e-01
-1.34512037e-01 4.98550981e-01 -1.12899333e-01 -1.29288450e-01
-3.20884824e-01 -1.02667987e+00 -4.56040382e-01 -9.26113427e-01
7.88559094e-02 6.59488916e-01 -1.61449052e-02 -1.28796190e-01
5.73765934e-01 9.31966603e-01 -1.48808670e+00 6.50082111e-01
9.07993555e-01 1.15976286e+00 3.80733877e-01 4.07056868e-01
4.47763577e-02 6.49999857e-01 -6.52150273e-01 -2.27443933e-01
4.23367441e-01 -3.99058729e-01 -5.84606886e-01 2.96852559e-01
-8.13366100e-02 -3.97844404e-01 -2.11492628e-01 1.07863009e+00
3.57594490e-01 2.98203558e-01 1.12558353e+00 -8.78954470e-01
-1.02351749e+00 2.72934407e-01 -4.70040321e-01 1.98009774e-01
1.23928376e-01 2.06770003e-01 5.83366692e-01 -9.64206219e-01
3.88728976e-01 1.12288570e+00 3.18984658e-01 2.90836990e-01
-7.00388491e-01 -6.91914916e-01 -2.12098137e-01 5.44288695e-01
-1.05033851e+00 1.57495499e-01 8.00288498e-01 -7.35446632e-01
6.07678413e-01 4.46369469e-01 6.55024886e-01 7.44262457e-01
4.60147262e-01 6.94194317e-01 1.15126467e+00 -5.35005450e-01
2.88395047e-01 3.16042125e-01 5.04492402e-01 5.15996993e-01
3.89501214e-01 -6.87762722e-02 1.43442780e-01 2.01291338e-01
4.53580022e-01 2.30853483e-01 -3.55715334e-01 2.08277270e-01
-6.34325087e-01 7.58255363e-01 5.71774364e-01 3.43783826e-01
-4.38078165e-01 -3.59806210e-01 5.17159104e-01 1.25821054e-01
3.41169685e-01 3.98520470e-01 7.16908649e-02 2.96724647e-01
-5.44423282e-01 -2.10678927e-03 5.77057421e-01 7.16832340e-01
4.17374611e-01 4.97422069e-02 2.08631873e-01 1.18859470e+00
2.36933053e-01 3.12177122e-01 4.49654073e-01 -5.54161251e-01
3.34615678e-01 8.36546957e-01 2.55123563e-02 -6.33788824e-01
-2.96730965e-01 -5.77070490e-02 -7.68784761e-01 6.96224928e-01
9.27753970e-02 1.21967383e-01 -1.24578071e+00 4.18484718e-01
2.25910574e-01 -5.54172873e-01 -1.74168393e-01 1.22720599e+00
7.40929008e-01 8.82278502e-01 -5.65227494e-02 1.68257281e-01
1.63689613e+00 -7.91016459e-01 -7.31828451e-01 -1.34977594e-01
-1.00417286e-01 -7.93422341e-01 9.41859365e-01 8.36883307e-01
-8.94376695e-01 -4.83925760e-01 -1.38368428e+00 3.31560761e-04
-5.00544906e-01 4.99176800e-01 5.64273477e-01 6.35491192e-01
-2.43957285e-02 7.75678396e-01 -7.50111759e-01 -4.05730568e-02
4.25555915e-01 4.46310826e-02 -4.01369572e-01 -3.59777719e-01
-9.73480880e-01 9.52502251e-01 4.48850214e-01 6.98184490e-01
-9.05916333e-01 -2.35484198e-01 -7.78321207e-01 3.01384069e-02
9.40540954e-02 -1.79680772e-02 9.39897418e-01 -7.33812451e-01
-1.04164040e+00 5.90756297e-01 4.65855688e-01 -1.19872272e-01
7.17267156e-01 -4.49463487e-01 -4.07013655e-01 4.36108589e-01
5.12450896e-02 -1.63154900e-01 7.88646042e-01 -1.35991895e+00
-6.07511342e-01 -5.32694221e-01 5.51676750e-02 2.24622730e-02
4.58372794e-02 2.08734393e-01 -7.88937360e-02 -7.19098985e-01
3.70747894e-01 -8.09193015e-01 2.14854613e-01 -2.71446019e-01
-4.98374939e-01 -8.82353913e-03 8.24401498e-01 -1.14374340e+00
6.25930309e-01 -2.09766364e+00 -2.60005385e-01 2.29486212e-01
-2.48039857e-01 1.48692042e-01 2.39362746e-01 4.94848728e-01
-4.17922139e-01 -1.12204693e-01 -3.71326208e-01 2.87420839e-01
1.84576362e-02 -1.98569581e-01 -1.85876235e-01 6.47287011e-01
3.97369772e-01 3.06330979e-01 -7.12852836e-01 -1.20325543e-01
4.18513596e-01 6.72572017e-01 1.30050570e-01 -7.16512352e-02
2.72110058e-03 4.56209332e-01 -1.41107202e-01 1.27577174e+00
9.32014108e-01 1.92889705e-01 -2.22996548e-02 -5.08128405e-01
-1.07665278e-01 1.52044594e-01 -1.23642862e+00 6.58874512e-01
-4.56360340e-01 4.83839601e-01 2.10542092e-03 -8.42455387e-01
1.01436889e+00 1.63280591e-01 3.70878637e-01 -1.05416536e+00
4.15734977e-01 4.45830047e-01 -1.05220549e-01 -8.49360883e-01
4.42351669e-01 -6.16687596e-01 1.07140496e-01 3.66809189e-01
-1.74167275e-01 -6.22486919e-02 4.40214336e-01 -1.68263510e-01
7.96194553e-01 -1.83962677e-02 -1.32365659e-01 -1.96874470e-01
4.38798815e-01 3.47791463e-02 4.64749664e-01 1.75977156e-01
1.51387691e-01 9.26545382e-01 2.13812321e-01 -4.82421488e-01
-1.07936966e+00 -1.23772836e+00 -5.32308578e-01 7.36295521e-01
2.79925734e-01 4.34662968e-01 -6.15586817e-01 -4.23517853e-01
6.50715157e-02 9.19707060e-01 -6.14754140e-01 -8.66247118e-02
-6.40471518e-01 -1.11060405e+00 4.98447288e-03 6.90568984e-01
5.95649779e-01 -1.34444594e+00 -5.47722638e-01 1.74360275e-01
-3.67455813e-03 -6.97975695e-01 -2.35027343e-01 2.74608940e-01
-5.30325055e-01 -1.66718745e+00 -7.57911265e-01 -7.16364324e-01
9.46030080e-01 3.12169969e-01 6.10293627e-01 2.79630959e-01
-8.40311706e-01 1.76001457e-03 -5.69800198e-01 -4.67808217e-01
-5.01798868e-01 -5.92885911e-01 -7.62672871e-02 1.72371000e-01
3.20535839e-01 -1.82983249e-01 -7.19091833e-01 4.68319535e-01
-9.02792394e-01 -9.23215076e-02 6.27108991e-01 6.59503758e-01
4.26492125e-01 4.30883855e-01 4.26777333e-01 -7.63170660e-01
5.92899680e-01 -4.16206747e-01 -5.40964663e-01 1.32338554e-02
-2.34005526e-01 -3.66678834e-01 7.30105102e-01 -2.30366245e-01
-1.26379704e+00 -2.33600810e-01 1.61154959e-02 -1.63382843e-01
-2.06870511e-01 1.68003246e-01 -2.10896611e-01 2.10423559e-01
5.77134311e-01 3.44918221e-02 -2.62108803e-01 -7.19829082e-01
-1.44823879e-01 1.05405414e+00 7.21930385e-01 -4.28062618e-01
5.00412345e-01 6.09752834e-01 -1.65019244e-01 -8.66448224e-01
-5.67108929e-01 -4.57779825e-01 -4.37199235e-01 -5.98416805e-01
9.84674335e-01 -4.69187528e-01 -7.35863805e-01 6.20684326e-01
-1.06548870e+00 4.13322672e-02 9.29414704e-02 6.58373654e-01
-1.42968029e-01 5.04349113e-01 -1.03650212e+00 -1.06827581e+00
-8.13452363e-01 -1.02190340e+00 6.83701992e-01 2.87492961e-01
2.09570646e-01 -3.02938074e-01 -1.09796248e-01 6.90865099e-01
6.17687684e-03 3.73458326e-01 1.21007311e+00 -5.79867959e-01
-4.57007408e-01 -6.15663946e-01 -3.01083446e-01 8.61396968e-01
3.46118033e-01 5.91006782e-03 -8.17049026e-01 -2.98611581e-01
4.51018512e-01 -5.85227571e-02 9.48529363e-01 5.30390032e-02
8.75454307e-01 1.54778305e-02 7.30550066e-02 2.17335388e-01
1.45677221e+00 9.47461307e-01 7.49581397e-01 6.81582749e-01
5.41135252e-01 7.32993960e-01 9.75802541e-01 2.94130951e-01
-1.61056027e-01 1.84087753e-01 6.32305503e-01 3.15301299e-01
-6.15015179e-02 3.24683428e-01 2.49378785e-01 6.23884797e-01
-5.61899722e-01 -3.18679363e-01 -9.52408731e-01 6.06189907e-01
-1.32275283e+00 -1.04998493e+00 -4.32425469e-01 2.06390667e+00
3.00858766e-01 1.55256361e-01 -1.22484550e-01 5.08612335e-01
9.61027086e-01 -4.20338273e-01 -4.15024191e-01 -3.27902973e-01
1.21609896e-01 1.54184431e-01 4.20723826e-01 2.99181044e-01
-1.10584223e+00 6.60993010e-02 5.33676863e+00 4.08167869e-01
-1.07786250e+00 -1.00449912e-01 3.84974211e-01 -1.51996389e-01
-5.96965365e-02 -2.20119923e-01 -4.47825432e-01 9.47763622e-01
3.65356177e-01 2.46803194e-01 4.40716386e-01 7.71953285e-01
1.30104542e-01 -3.84628892e-01 -9.67363656e-01 6.97391987e-01
5.61444163e-02 -9.42050755e-01 -2.01640174e-01 -2.21674696e-01
4.29794937e-01 -1.73558250e-01 -2.82106578e-01 1.12559177e-01
7.45733678e-02 -9.07241285e-01 1.03219688e+00 5.96298754e-01
2.35132962e-01 -9.66358125e-01 1.06035030e+00 8.95994529e-02
-9.81792748e-01 -5.02536416e-01 -3.22379261e-01 2.08978698e-01
2.18475670e-01 6.62337244e-01 -5.74360490e-01 7.63450801e-01
8.82633448e-01 7.72580579e-02 -3.99208456e-01 1.13002515e+00
-4.10306901e-01 1.79748341e-01 -1.95318252e-01 6.28237473e-03
5.14959730e-02 -7.73932993e-01 3.19020122e-01 1.18730080e+00
3.30646008e-01 5.87617904e-02 -1.56072333e-01 8.82132590e-01
1.98422834e-01 -1.88442811e-01 -3.12881291e-01 5.58286533e-02
3.60517591e-01 1.13602793e+00 -1.02394724e+00 -2.43587717e-01
-2.04136148e-01 7.95635700e-01 -2.21162096e-01 2.24051297e-01
-8.83692503e-01 -7.69935369e-01 4.85824645e-02 3.03897202e-01
2.70478725e-01 -1.22980729e-01 -5.45226872e-01 -7.78847873e-01
4.06468719e-01 -6.78659022e-01 4.38110143e-01 -9.38491940e-01
-1.46164954e+00 5.58753550e-01 7.10007399e-02 -1.50193918e+00
3.13267082e-01 -9.46829438e-01 -8.79325330e-01 8.63995016e-01
-1.19999313e+00 -1.09023654e+00 -7.22797692e-01 -5.83532155e-02
8.52815032e-01 -2.67678916e-01 2.80148119e-01 6.16557598e-01
-7.39189744e-01 -1.48591399e-01 3.32633674e-01 2.05178961e-01
2.92923689e-01 -1.20161641e+00 1.97648361e-01 6.57076180e-01
-4.28286523e-01 4.74395961e-01 7.55401909e-01 -8.24839711e-01
-1.27498841e+00 -1.13834512e+00 1.00333579e-01 -1.39154792e-01
4.50856686e-01 -4.23709095e-01 -1.20438445e+00 6.60133123e-01
6.08311817e-02 -3.95888507e-01 5.64791143e-01 -4.95089710e-01
7.26060197e-02 1.31318241e-01 -1.28047550e+00 2.51053184e-01
4.82243538e-01 -3.36205900e-01 -8.45012009e-01 6.10274673e-01
-1.19493149e-01 -1.75577164e-01 -7.54375100e-01 2.80142516e-01
4.59664285e-01 -9.37401056e-01 9.86865282e-01 -2.42407039e-01
7.48814940e-01 -4.53612506e-01 -5.99123491e-03 -9.44633842e-01
-4.45340276e-01 -1.18669914e-03 1.49289012e-01 1.27437782e+00
8.97409674e-03 -5.20361066e-01 5.91971934e-01 4.81178939e-01
-3.93067747e-01 -5.80770314e-01 -7.40934968e-01 -9.36832368e-01
-1.11074284e-01 -2.22973198e-01 4.03377891e-01 9.16265965e-01
-2.01548502e-01 2.00523883e-01 -1.14603750e-01 3.74058217e-01
7.35462129e-01 2.62431920e-01 2.93118298e-01 -1.12126386e+00
-9.81283635e-02 -2.54813612e-01 -4.19603616e-01 -1.42894804e-01
-2.52965033e-01 -1.06228685e+00 8.98998901e-02 -1.95177495e+00
8.63147378e-02 -2.62874275e-01 -2.47373670e-01 3.88798356e-01
-9.78181791e-03 3.93605381e-01 3.43434572e-01 1.13375269e-01
6.36956990e-02 7.01306105e-01 9.97841001e-01 -6.04557216e-01
-8.58101100e-02 1.37953922e-01 -4.31138009e-01 8.52103412e-01
9.59784329e-01 -4.48146045e-01 6.78975927e-03 -3.68534535e-01
-1.38759911e-02 -1.65993288e-01 6.06022120e-01 -9.11376059e-01
-1.20392263e-01 -3.82606150e-03 5.53752959e-01 -8.63185465e-01
2.95494616e-01 -1.03372788e+00 3.15180749e-01 5.80979645e-01
2.78307706e-01 -1.48820803e-01 6.56416565e-02 3.08771342e-01
-1.05777599e-01 -8.94318998e-01 1.22484958e+00 -3.06625754e-01
-6.00670695e-01 -1.25503361e-01 -3.79864693e-01 -6.82469785e-01
1.14779377e+00 -4.17848498e-01 -5.46285212e-01 2.27276132e-01
-7.43834078e-01 1.97079927e-01 3.76234293e-01 2.92936921e-01
8.25976372e-01 -1.25826669e+00 -6.25562251e-01 1.43622220e-01
-2.17946455e-01 5.31456098e-02 6.53491795e-01 6.13029718e-01
-1.28802633e+00 1.49347540e-02 -7.21082449e-01 -2.33483717e-01
-1.09440219e+00 7.35614419e-01 1.85327336e-01 4.73139100e-02
-9.83758450e-01 4.11017656e-01 1.30573794e-01 -2.30918661e-01
-3.57375853e-02 -1.91828296e-01 -2.90881217e-01 2.87569940e-01
6.42621338e-01 7.98600137e-01 4.80271757e-01 -2.88003772e-01
-2.61019826e-01 5.55745125e-01 -3.26099902e-01 2.66285092e-01
1.60845828e+00 2.87080288e-01 -3.81799132e-01 2.22642988e-01
1.06668639e+00 -4.87335324e-02 -9.90350604e-01 4.19443101e-01
-2.70271786e-02 -5.93020737e-01 -1.52360633e-01 -8.31205368e-01
-1.06650245e+00 9.75943089e-01 6.94901049e-01 4.30919260e-01
1.13386106e+00 -5.38433455e-02 5.36119640e-01 3.57743591e-01
3.17270398e-01 -1.35922647e+00 3.50296274e-02 3.08925986e-01
1.39104807e+00 -9.57515776e-01 4.64007288e-01 -3.89474332e-01
-7.13463187e-01 1.24365127e+00 5.88958323e-01 -4.15777355e-01
2.64056236e-01 1.56583056e-01 1.01036705e-01 -5.12193143e-01
-1.45929053e-01 1.02542177e-01 2.64712244e-01 4.95638758e-01
2.21345454e-01 -6.89606443e-02 -5.28645694e-01 8.18742394e-01
-8.70989487e-02 -2.09102720e-01 6.42182112e-01 1.37981164e+00
-5.73190928e-01 -5.18305898e-01 -8.69300485e-01 5.91906905e-01
-4.71086144e-01 3.05081785e-01 -2.63177544e-01 8.77723694e-01
1.69613466e-01 7.79129267e-01 7.09924847e-02 -1.46212563e-01
5.89786828e-01 -7.24704713e-02 3.68851274e-01 -4.78686899e-01
-6.71438694e-01 -3.79881412e-02 1.38863444e-01 -7.82388598e-02
4.94198874e-02 -2.71560490e-01 -1.56711543e+00 2.34105736e-02
-4.45494980e-01 5.28565347e-02 1.07653689e+00 7.26572692e-01
-2.63057321e-01 7.79248893e-01 8.01799715e-01 -1.12637711e+00
-3.41830760e-01 -1.05469692e+00 -9.62948143e-01 6.84587300e-01
1.58918649e-01 -7.99529672e-01 -4.95757878e-01 1.62505180e-01] | [7.435753345489502, 1.7795143127441406] |
6729b9da-0ac1-4fc1-b771-c342eb820db9 | efficient-debiasing-with-contrastive-weight | 2210.05247 | null | https://arxiv.org/abs/2210.05247v3 | https://arxiv.org/pdf/2210.05247v3.pdf | Training Debiased Subnetworks with Contrastive Weight Pruning | Neural networks are often biased to spuriously correlated features that provide misleading statistical evidence that does not generalize. This raises an interesting question: ``Does an optimal unbiased functional subnetwork exist in a severely biased network? If so, how to extract such subnetwork?" While empirical evidence has been accumulated about the existence of such unbiased subnetworks, these observations are mainly based on the guidance of ground-truth unbiased samples. Thus, it is unexplored how to discover the optimal subnetworks with biased training datasets in practice. To address this, here we first present our theoretical insight that alerts potential limitations of existing algorithms in exploring unbiased subnetworks in the presence of strong spurious correlations. We then further elucidate the importance of bias-conflicting samples on structure learning. Motivated by these observations, we propose a Debiased Contrastive Weight Pruning (DCWP) algorithm, which probes unbiased subnetworks without expensive group annotations. Experimental results demonstrate that our approach significantly outperforms state-of-the-art debiasing methods despite its considerable reduction in the number of parameters. | ['Jong Chul Ye', 'Sang Wan Lee', 'Sangmin Lee', 'Geon Yeong Park'] | 2022-10-11 | training-debiased-subnetworks-with | http://openaccess.thecvf.com//content/CVPR2023/html/Park_Training_Debiased_Subnetworks_With_Contrastive_Weight_Pruning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Park_Training_Debiased_Subnetworks_With_Contrastive_Weight_Pruning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['facial-attribute-classification'] | ['computer-vision'] | [ 7.65834928e-01 3.99847656e-01 -5.16867876e-01 -5.06329715e-01
-5.87924004e-01 -4.17372137e-01 2.23716334e-01 -2.56222934e-01
-3.23712766e-01 1.53021300e+00 1.99934721e-01 -3.71539056e-01
-6.09684646e-01 -5.82446873e-01 -9.65809643e-01 -9.86407280e-01
-1.68079343e-02 4.42149609e-01 3.73493284e-01 9.42012668e-02
3.44314128e-01 3.27296704e-01 -1.53959954e+00 1.44599348e-01
9.71333206e-01 7.03555882e-01 -1.52048141e-01 1.71168193e-01
6.56894520e-02 4.07237142e-01 -6.37712717e-01 -4.45346206e-01
9.57555845e-02 -7.33978271e-01 -8.52378786e-01 -9.17055756e-02
5.87805033e-01 -1.50282115e-01 1.75716374e-02 1.34363759e+00
5.42788982e-01 6.03599288e-02 6.45657241e-01 -1.13524747e+00
-3.32081228e-01 1.12231827e+00 -6.59134984e-01 5.40023863e-01
-4.71413970e-01 2.28851795e-01 1.50502646e+00 -3.79533768e-01
8.77679825e-01 1.06217813e+00 5.69047987e-01 5.52961051e-01
-1.73005295e+00 -9.98455346e-01 5.80261111e-01 -1.24368049e-01
-1.11454868e+00 -5.60846865e-01 9.77453768e-01 -3.54387999e-01
6.30610943e-01 3.98989677e-01 7.14571714e-01 1.72396517e+00
-1.58607304e-01 6.88922048e-01 1.01387084e+00 -4.35009263e-02
4.52151418e-01 -1.08881012e-01 5.53399146e-01 7.10212350e-01
1.23125231e+00 2.88914114e-01 -9.01193857e-01 -4.92371619e-01
6.49962723e-01 -3.95957649e-01 -5.82241297e-01 -5.95429420e-01
-1.22364438e+00 7.78744936e-01 2.86644429e-01 3.17644030e-01
-2.06057310e-01 3.03833485e-01 4.08936560e-01 2.98177630e-01
6.46483362e-01 1.03930223e+00 -6.35381103e-01 6.22518808e-02
-1.13339353e+00 2.86096513e-01 6.09982252e-01 6.69739664e-01
7.23740578e-01 7.52588212e-02 6.75860643e-02 6.45536602e-01
6.25855997e-02 -2.83435695e-02 2.58843392e-01 -9.58317161e-01
1.99058533e-01 5.76302290e-01 1.26089985e-02 -9.51341331e-01
-4.07257557e-01 -1.11742985e+00 -1.03535903e+00 -6.82915226e-02
7.77727544e-01 -3.69441926e-01 -4.13083166e-01 2.07168245e+00
3.20823461e-01 3.00357848e-01 -3.43725920e-01 9.60050464e-01
5.11433840e-01 -1.89265400e-01 -1.91961631e-01 -7.13684037e-02
9.81061816e-01 -4.35738742e-01 -4.19006795e-01 -2.52007455e-01
3.73317003e-01 -1.22273967e-01 9.36646581e-01 4.13678229e-01
-8.95233572e-01 -8.57267529e-02 -1.12098873e+00 2.18861043e-01
-2.76138097e-01 6.35977685e-02 1.01756144e+00 9.18185592e-01
-6.73658907e-01 9.85229611e-01 -6.38790607e-01 -2.63615459e-01
8.85125041e-01 5.40083170e-01 -2.05007404e-01 2.04444557e-01
-1.17744684e+00 5.08504987e-01 4.21616763e-01 -4.54706233e-03
-7.77300179e-01 -1.02820098e+00 -3.19070518e-01 1.32789969e-01
7.33818412e-01 -1.00247657e+00 7.13240921e-01 -1.08052862e+00
-1.05881667e+00 6.81319833e-01 -3.35480601e-01 -5.73481381e-01
5.41832387e-01 1.25521095e-02 -1.27710849e-01 1.08803518e-03
7.29410350e-02 5.76518476e-01 7.48995245e-01 -1.29007161e+00
-4.88421559e-01 -5.48622966e-01 -3.73302735e-02 -2.45357692e-01
-3.62649739e-01 -2.80624121e-01 2.61629429e-02 -7.27949202e-01
4.93625522e-01 -7.12701082e-01 -3.04480433e-01 1.72642246e-01
-1.04537106e+00 1.27882525e-01 2.34551564e-01 -6.66548908e-02
9.37468588e-01 -1.94932127e+00 1.38712034e-01 4.84894276e-01
7.03666151e-01 -1.45344302e-01 -2.87904423e-02 -4.14730199e-02
-3.85432690e-01 4.51379210e-01 -3.69270325e-01 1.30023211e-01
-1.38705745e-01 6.33228868e-02 -5.01610816e-01 7.96558022e-01
2.81530052e-01 6.88285410e-01 -1.03342462e+00 -3.59312356e-01
-4.03797895e-01 1.64291948e-01 -8.36053073e-01 -2.27300465e-01
-1.65508419e-01 6.20665610e-01 -4.83335346e-01 8.52787852e-01
5.12534559e-01 -5.35209835e-01 5.15970707e-01 -3.97451401e-01
-8.60819668e-02 4.49974597e-01 -1.04812121e+00 1.08298790e+00
4.62494850e-01 7.66565740e-01 7.82241970e-02 -1.37335670e+00
7.85221636e-01 -1.82931609e-02 3.41972530e-01 7.71347880e-02
4.10371065e-01 3.49001259e-01 3.35312903e-01 -1.57404006e-01
2.28489950e-01 -3.31158072e-01 1.99069962e-01 5.00605464e-01
1.40063509e-01 3.25968266e-01 1.63225204e-01 -1.31278217e-01
1.45187831e+00 -4.91756089e-02 3.64978909e-01 -7.07019210e-01
4.79557998e-02 -2.17609685e-02 6.93959773e-01 1.01868105e+00
-1.49158731e-01 6.07553244e-01 1.12314379e+00 -1.13064870e-01
-8.27196121e-01 -9.20070887e-01 -2.33536661e-01 1.07015026e+00
1.41221493e-01 3.53442580e-02 -6.49631917e-01 -7.90925264e-01
1.33884579e-01 5.81804574e-01 -9.51063991e-01 -3.33796203e-01
-2.44816080e-01 -1.27099085e+00 7.84394026e-01 3.70805830e-01
3.38077575e-01 -5.47171295e-01 -3.51981193e-01 -2.60548089e-02
-6.64616600e-02 -6.61632538e-01 -7.91720077e-02 6.71947181e-01
-1.26881194e+00 -1.49885511e+00 -6.40641153e-01 -2.75276810e-01
1.13990414e+00 3.56204242e-01 1.05946302e+00 3.44305933e-02
-1.14678770e-01 -3.87167543e-01 1.41754746e-01 -3.70106936e-01
-1.69751540e-01 5.81642210e-01 -1.68446600e-02 -1.39210239e-01
4.03630972e-01 -1.05371022e+00 -4.99154419e-01 4.69401300e-01
-7.11539984e-01 -1.41494170e-01 8.60545933e-01 1.14075029e+00
4.63111758e-01 1.53568089e-01 9.04442906e-01 -1.44474590e+00
7.61687875e-01 -5.83638072e-01 -7.27794170e-01 6.83140606e-02
-5.68498254e-01 5.53954542e-01 4.20705169e-01 -5.51414490e-01
-8.89052689e-01 -2.62299895e-01 5.14729545e-02 -5.33930399e-02
-1.13233432e-01 3.20981711e-01 -3.15377057e-01 -8.90708622e-03
9.29166555e-01 -2.32304428e-02 -3.35751660e-02 -5.48636377e-01
1.38972282e-01 2.09957868e-01 3.95371646e-01 -9.18246448e-01
5.17345369e-01 7.89531350e-01 2.75069535e-01 -8.38185668e-01
-1.13799906e+00 -4.91237283e-01 -4.91511792e-01 -6.03127517e-02
2.10560709e-01 -5.38170576e-01 -7.80144215e-01 -9.15581733e-02
-9.49032187e-01 -1.10393852e-01 -3.68947893e-01 3.55501533e-01
-3.92723352e-01 3.32094729e-01 -1.16272464e-01 -7.83163249e-01
-1.49176002e-01 -1.07317603e+00 7.94737458e-01 -2.56807595e-01
-9.06840503e-01 -6.86791241e-01 5.40797897e-02 3.58803570e-01
2.08058596e-01 2.84266025e-01 1.27231526e+00 -9.60302591e-01
-7.92261302e-01 3.40351537e-02 -3.10405642e-01 1.77143559e-01
1.15427421e-02 5.77096157e-02 -9.92731988e-01 -3.12708654e-02
-2.58272737e-01 -2.75752753e-01 1.33179092e+00 5.75168908e-01
1.32403827e+00 -2.53113568e-01 -4.99349326e-01 6.18384719e-01
1.00075543e+00 -3.63198042e-01 2.24710241e-01 9.78810266e-02
3.09727997e-01 7.56668985e-01 2.16433212e-01 3.07224095e-01
-3.78248781e-01 2.24362284e-01 6.41811550e-01 1.11176066e-01
5.52810505e-02 -2.07127586e-01 -7.84789920e-02 2.70725369e-01
-1.70337364e-01 -2.61783779e-01 -4.49195474e-01 7.08335698e-01
-1.79523206e+00 -9.79770422e-01 -2.02489600e-01 2.11955476e+00
1.08561373e+00 5.03870308e-01 1.91089027e-02 3.13610673e-01
8.20006371e-01 2.73951411e-01 -9.42467511e-01 1.27439976e-01
-3.47813547e-01 7.65864924e-02 7.07724392e-01 1.73334301e-01
-7.72905946e-01 6.02585077e-01 6.55098152e+00 7.39296377e-01
-9.92418587e-01 -7.20850006e-02 7.96818316e-01 -4.16874200e-01
-5.89616895e-01 1.15080334e-01 -1.00270939e+00 4.46886569e-01
5.11697531e-01 5.01757078e-02 3.42885762e-01 8.82075131e-01
2.28589654e-01 -1.07059002e-01 -1.10403526e+00 5.48534334e-01
-2.02478379e-01 -1.42935574e+00 3.50667862e-03 2.84276724e-01
9.21701193e-01 3.41567621e-02 1.36080012e-01 8.29557423e-03
6.45979762e-01 -8.26595008e-01 5.01527905e-01 5.75141132e-01
3.42272460e-01 -8.20277870e-01 6.78257525e-01 3.51416767e-01
-3.26564372e-01 1.32816777e-01 -6.10863268e-01 1.19556263e-01
-2.91977048e-01 1.34298694e+00 -8.91074896e-01 1.58628151e-01
4.84538525e-01 4.66198862e-01 -5.34001529e-01 1.07419145e+00
-3.83556455e-01 9.21573997e-01 -3.66911232e-01 -2.76709914e-01
5.40968105e-02 5.64595126e-02 7.56103218e-01 7.56565094e-01
1.73482940e-01 -3.29716563e-01 -5.77655911e-01 1.45134854e+00
-3.33529174e-01 -2.45467395e-01 -6.47834063e-01 -1.86419904e-01
4.71864462e-01 9.52825785e-01 -9.08962727e-01 -9.50456262e-02
1.25684068e-02 5.27395785e-01 5.73166966e-01 4.52014476e-01
-6.54571235e-01 -1.34524837e-01 9.54772055e-01 1.56763345e-01
3.66727561e-01 1.69268474e-01 -6.02841914e-01 -1.06763172e+00
-1.76951572e-01 -8.06426883e-01 3.41836482e-01 -3.20070833e-01
-1.53164434e+00 8.95894617e-02 8.02352661e-05 -6.95152640e-01
1.75962806e-01 -5.10523319e-01 -3.76825005e-01 4.77404356e-01
-1.28327405e+00 -4.14332479e-01 -6.60344362e-02 -1.36150178e-02
2.38697931e-01 5.45353331e-02 3.65755290e-01 4.12843339e-02
-8.45241487e-01 4.94585872e-01 3.14683527e-01 -1.61543816e-01
7.49386072e-01 -1.09405720e+00 1.85461313e-01 7.02113152e-01
-9.74780396e-02 1.16133285e+00 1.07107592e+00 -8.55483413e-01
-9.75141764e-01 -1.15896761e+00 6.75267696e-01 -3.02692294e-01
6.79325223e-01 -2.84781575e-01 -9.16506231e-01 6.40804887e-01
-2.98741221e-01 2.76146475e-02 9.57385480e-01 6.09181523e-01
-3.20768625e-01 -2.82883309e-02 -1.19894612e+00 1.02625680e+00
1.61525345e+00 -4.04614629e-03 -5.20897150e-01 8.32788944e-02
5.64100444e-01 1.13784991e-01 -4.98424172e-01 5.06309867e-01
6.21709168e-01 -1.17179036e+00 1.00745630e+00 -9.49403226e-01
3.41881752e-01 -5.28989621e-02 -2.39557063e-04 -1.20521820e+00
-1.11319169e-01 -5.83093762e-01 2.70785373e-02 9.66631293e-01
7.38793194e-01 -7.34741509e-01 1.20312774e+00 4.23987329e-01
1.07895255e-01 -9.38814878e-01 -1.17571819e+00 -9.10089493e-01
-1.05203558e-02 -3.60825092e-01 7.13188708e-01 1.09648836e+00
7.62437508e-02 3.25761229e-01 -2.20479324e-01 5.84633388e-02
8.48015010e-01 2.78166085e-01 6.94232702e-01 -1.59392440e+00
-4.84750688e-01 -7.81848550e-01 -8.77506286e-02 -1.00433469e+00
4.30970520e-01 -7.66888857e-01 1.63597371e-02 -8.88636529e-01
3.77148718e-01 -1.90292746e-01 -2.72333622e-01 2.69886404e-01
-2.96218097e-01 2.00523108e-01 -6.00584626e-01 4.15067039e-02
-6.61769867e-01 5.13674021e-01 1.07964349e+00 -1.05777591e-01
-4.88198698e-02 6.25463575e-02 -1.33978248e+00 9.43698525e-01
9.65256453e-01 -6.88905239e-01 -4.15787280e-01 1.03421263e-01
5.93825519e-01 -5.70980847e-01 4.20065433e-01 -6.94795012e-01
1.61485579e-02 -3.44456226e-01 3.29601228e-01 -6.18792653e-01
2.52045840e-01 -5.89227021e-01 6.81689233e-02 3.58037472e-01
-6.81489766e-01 -4.98560458e-01 -6.28622472e-02 8.96288395e-01
2.12680250e-01 -4.95462000e-01 8.12654316e-01 -2.38780484e-01
-9.00385007e-02 -4.82568964e-02 -4.30868953e-01 2.92837709e-01
5.76001048e-01 -1.71570450e-01 -5.90061545e-01 -7.62920231e-02
-4.35037434e-01 -1.61518425e-01 3.13670486e-01 -9.75023136e-02
3.46248418e-01 -8.73820305e-01 -2.81651258e-01 1.50217444e-01
6.98265806e-02 -4.28962298e-02 -8.87601376e-02 9.98727620e-01
2.38212897e-03 4.24680650e-01 8.27636719e-02 -3.95511836e-01
-1.27029765e+00 3.23671043e-01 2.43578792e-01 -2.49857366e-01
-2.67139465e-01 1.02548563e+00 2.73182273e-01 -3.36540014e-01
2.62431175e-01 -5.75483382e-01 1.75005391e-01 2.49065652e-01
1.87604070e-01 6.19630218e-01 -1.13227079e-02 -2.38157570e-01
-1.46967188e-01 4.79082912e-02 -3.04880049e-02 1.30989209e-01
1.39659905e+00 1.06013395e-01 -2.33119220e-01 4.42391515e-01
9.62793291e-01 5.78427538e-02 -1.15448272e+00 -1.58344373e-01
1.71495467e-01 -4.63712126e-01 2.66089141e-02 -6.04581177e-01
-1.14485443e+00 6.84211493e-01 -4.92964126e-02 1.50068104e-01
6.93734109e-01 9.29103568e-02 3.27829033e-01 6.54274583e-01
3.08819413e-01 -9.23473597e-01 4.58405428e-02 3.00773144e-01
6.93296134e-01 -9.86861169e-01 3.63244265e-01 -5.63200414e-01
-2.31978297e-01 8.29214931e-01 6.48845613e-01 -8.93582255e-02
4.54040021e-01 -5.88083733e-03 -4.82771993e-01 -5.56056619e-01
-8.51672173e-01 -2.30535775e-01 8.46088454e-02 5.41668415e-01
2.54184127e-01 -2.46014342e-01 -4.17656273e-01 8.52111816e-01
-3.61163497e-01 -9.93429124e-02 2.42624342e-01 7.38470495e-01
-6.92537904e-01 -8.55198324e-01 -3.69070679e-01 9.50036168e-01
-5.32013595e-01 -1.23010300e-01 -9.47854221e-01 8.38221788e-01
8.81406814e-02 5.90189815e-01 -2.09099516e-01 -1.19201116e-01
4.94227670e-02 1.81469038e-01 4.45790261e-01 -5.58013856e-01
-4.74711448e-01 -6.13150150e-02 4.03560370e-01 -3.70739162e-01
-4.01267737e-01 -6.96160257e-01 -7.52640307e-01 -2.17958882e-01
-8.40979636e-01 2.98303226e-03 2.90811867e-01 1.00129330e+00
4.29142863e-01 8.44772696e-01 7.20278174e-02 -5.46329498e-01
-7.02847123e-01 -7.89728165e-01 -6.75402999e-01 -4.69923392e-02
3.91376883e-01 -9.14979935e-01 -7.94076085e-01 -2.61340380e-01] | [8.559683799743652, 4.114323616027832] |
c63bfa2a-6d6f-4d8b-b6fd-8ab226c34db6 | a-survey-on-deep-learning-of-small-sample-in | 1908.00473 | null | https://arxiv.org/abs/1908.00473v1 | https://arxiv.org/pdf/1908.00473v1.pdf | A Survey on Deep Learning of Small Sample in Biomedical Image Analysis | The success of deep learning has been witnessed as a promising technique for computer-aided biomedical image analysis, due to end-to-end learning framework and availability of large-scale labelled samples. However, in many cases of biomedical image analysis, deep learning techniques suffer from the small sample learning (SSL) dilemma caused mainly by lack of annotations. To be more practical for biomedical image analysis, in this paper we survey the key SSL techniques that help relieve the suffering of deep learning by combining with the development of related techniques in computer vision applications. In order to accelerate the clinical usage of biomedical image analysis based on deep learning techniques, we intentionally expand this survey to include the explanation methods for deep models that are important to clinical decision making. We survey the key SSL techniques by dividing them into five categories: (1) explanation techniques, (2) weakly supervised learning techniques, (3) transfer learning techniques, (4) active learning techniques, and (5) miscellaneous techniques involving data augmentation, domain knowledge, traditional shallow methods and attention mechanism. These key techniques are expected to effectively support the application of deep learning in clinical biomedical image analysis, and furtherly improve the analysis performance, especially when large-scale annotated samples are not available. We bulid demos at https://github.com/PengyiZhang/MIADeepSSL. | ['Xiaoqiong Li', 'Yulin Deng', 'Xiaoying Tang', 'Pengyi Zhang', 'Yunxin Zhong'] | 2019-08-01 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [ 3.26786369e-01 6.03117466e-01 -4.04215485e-01 -3.69188160e-01
-7.96855688e-01 1.04306728e-01 1.99812636e-01 3.69319946e-01
-4.00281608e-01 8.33879530e-01 2.53824055e-01 -3.91363800e-01
-1.70824826e-01 -4.79251564e-01 -4.18922931e-01 -1.00645244e+00
-1.52804092e-01 5.44961154e-01 -7.33708441e-02 8.43869597e-02
1.57157287e-01 5.32008290e-01 -1.14236891e+00 4.47228998e-01
9.70549643e-01 9.53212500e-01 7.76578709e-02 3.98548663e-01
-3.81246209e-01 1.08412683e+00 -3.98482114e-01 -2.54729509e-01
-1.25788793e-01 -3.99203956e-01 -1.07702017e+00 9.85303894e-02
-1.98133215e-01 -3.71915311e-01 -7.03141913e-02 9.93378997e-01
8.25186193e-01 -2.03777701e-02 8.11655521e-01 -1.26672149e+00
-5.13017952e-01 4.26227599e-01 -6.62840307e-01 3.67360592e-01
-2.00372078e-02 1.39824450e-01 4.98432726e-01 -1.14606678e+00
3.65927219e-01 9.63146865e-01 7.25677073e-01 9.14063692e-01
-6.32867813e-01 -8.35594714e-01 7.15044066e-02 6.00386560e-01
-1.04769969e+00 -6.25803471e-01 8.56072962e-01 -4.12204921e-01
7.50433266e-01 1.50545612e-01 6.34862483e-01 1.18496752e+00
2.21677884e-01 1.38737905e+00 7.72270262e-01 -6.54768944e-01
3.11839759e-01 5.30087687e-02 3.98523599e-01 8.81885350e-01
1.29100531e-01 7.32257813e-02 -3.74709666e-01 -3.05077940e-01
7.45341241e-01 3.75575542e-01 -3.28451306e-01 -2.70792872e-01
-1.02402222e+00 9.47903395e-01 6.55667543e-01 3.38282615e-01
-4.04259235e-01 -2.70081669e-01 6.75609469e-01 9.29923877e-02
6.43087924e-01 4.93419886e-01 -7.68277705e-01 2.13092580e-01
-7.53138661e-01 -1.10577628e-01 3.41435164e-01 7.40239978e-01
4.98262376e-01 1.11648358e-01 1.77348759e-02 9.76251721e-01
4.73525494e-01 -4.94450517e-02 9.46916342e-01 -5.16435683e-01
1.72062501e-01 9.13174152e-01 -3.19791555e-01 -6.75144136e-01
-6.05953932e-01 -4.03832585e-01 -1.29274106e+00 5.63665777e-02
1.38790071e-01 -1.21613346e-01 -1.08104312e+00 1.22699773e+00
2.81351000e-01 2.49736130e-01 8.29823166e-02 6.79393649e-01
1.45117009e+00 3.94487560e-01 3.31296474e-01 -4.97371107e-01
1.21344435e+00 -1.16381931e+00 -1.05009389e+00 -2.67999619e-01
1.05406320e+00 -5.43826103e-01 7.88571417e-01 5.15387952e-01
-8.75007153e-01 -5.12689412e-01 -9.02188897e-01 -8.36753398e-02
-5.27371168e-01 2.53208000e-02 9.52048779e-01 2.86903262e-01
-8.11452210e-01 2.42794096e-01 -1.03783369e+00 -2.88190216e-01
1.22767484e+00 7.03087151e-01 -4.61422384e-01 -1.66256875e-01
-1.20863438e+00 6.95970118e-01 3.63810986e-01 4.09597665e-01
-8.06513190e-01 -7.20355928e-01 -9.06290770e-01 -5.36315292e-02
2.86599606e-01 -5.93933821e-01 1.20684683e+00 -1.16790175e+00
-1.12376225e+00 1.11835897e+00 1.97085459e-03 -4.61985976e-01
3.73118490e-01 -3.16433311e-01 -1.96733162e-01 2.27175891e-01
6.19390197e-02 7.76440263e-01 3.63351792e-01 -1.01225924e+00
-3.59782100e-01 -6.36678994e-01 -3.00382107e-01 8.40075389e-02
-6.61257565e-01 -8.14348683e-02 -3.42245132e-01 -7.29077101e-01
-3.69030796e-02 -7.15820253e-01 -6.42941713e-01 3.07319611e-01
-4.36574817e-01 -4.68001723e-01 9.92622852e-01 -5.92577159e-01
8.13951552e-01 -2.19279957e+00 -1.47510797e-01 -1.91617727e-01
6.10649824e-01 7.98686504e-01 1.03917077e-01 2.50011533e-01
-3.57612163e-01 6.00431412e-02 -3.63323718e-01 -2.44853556e-01
-5.21511197e-01 1.90362781e-01 5.34964120e-03 4.31948781e-01
2.85739660e-01 1.19279838e+00 -8.92880261e-01 -8.45333457e-01
3.27761650e-01 3.08697373e-01 -3.52387965e-01 4.35964465e-01
1.13424323e-01 8.03126812e-01 -6.06374383e-01 9.91201937e-01
4.44559962e-01 -5.59360862e-01 -8.71037692e-02 -4.24455255e-01
3.03021759e-01 -1.00680389e-01 -5.55107832e-01 1.67608297e+00
-1.70889899e-01 5.11980355e-01 -1.95675325e-02 -1.59135664e+00
8.21228862e-01 7.39240468e-01 8.17344368e-01 -5.77286959e-01
3.40606481e-01 2.81666160e-01 1.29759371e-01 -9.57103729e-01
-2.81250447e-01 -3.07541281e-01 2.89480895e-01 1.40978117e-02
1.47364989e-01 4.58016574e-01 -1.50893927e-01 2.20369428e-01
9.77259457e-01 -1.27435401e-01 9.32810068e-01 -1.47711754e-01
5.24750173e-01 1.96060110e-02 6.81446195e-01 4.13163811e-01
-7.55776048e-01 4.89694178e-01 2.76328862e-01 -6.27566338e-01
-8.18780661e-01 -7.15534151e-01 -3.30810130e-01 6.46959543e-01
-3.47683355e-02 -5.83668984e-02 -5.26319981e-01 -9.31289554e-01
-2.56949276e-01 1.93950638e-01 -8.30150485e-01 -4.51771647e-01
-4.15060520e-01 -1.14613640e+00 4.89048898e-01 9.09706175e-01
6.92978263e-01 -1.29578555e+00 -3.21996897e-01 1.64986536e-01
-7.75910541e-02 -8.60018790e-01 4.75647412e-02 3.68547171e-01
-1.30440211e+00 -1.30889058e+00 -1.05550265e+00 -1.04050756e+00
9.73563313e-01 3.03505212e-02 9.31747079e-01 6.24464154e-01
-7.01541126e-01 1.08009167e-01 -6.25195861e-01 -9.60741103e-01
-2.27264538e-01 -1.87162254e-02 -1.25482038e-01 -2.35658407e-01
6.08260810e-01 -1.95725441e-01 -8.77712011e-01 1.13145754e-01
-7.71933317e-01 2.20547959e-01 9.89030242e-01 1.23690462e+00
5.63732207e-01 -2.06734270e-01 8.90869260e-01 -1.43145227e+00
2.67057776e-01 -5.25814235e-01 1.90110859e-02 6.14626966e-02
-6.66992664e-01 -4.33112442e-01 4.68616843e-01 -3.23997974e-01
-1.00261354e+00 1.68353617e-01 -4.67328250e-01 -3.73630971e-01
-2.80484438e-01 6.60871983e-01 -1.40267998e-01 2.83223353e-02
8.94664049e-01 9.68521982e-02 3.27653885e-01 -3.79617602e-01
-3.43018435e-02 1.00621307e+00 1.67947695e-01 -9.13463682e-02
1.51980400e-01 4.39721525e-01 -8.93162787e-02 -7.51216769e-01
-1.09916210e+00 -6.91744089e-01 -6.81756437e-01 -1.97247788e-01
8.40354919e-01 -8.32095146e-01 -4.16166455e-01 4.72287953e-01
-1.01381588e+00 -4.01041478e-01 -2.03163922e-01 6.79899395e-01
-4.65472609e-01 5.11735380e-01 -8.16662788e-01 -7.52394617e-01
-8.05407047e-01 -1.38581419e+00 9.66296315e-01 3.05065215e-01
-4.02973652e-01 -1.25337875e+00 6.77626440e-03 6.83492064e-01
2.50225425e-01 3.29936475e-01 1.13846552e+00 -9.55650628e-01
-1.90066025e-01 -4.18479919e-01 -2.51769543e-01 3.60993594e-01
2.92866647e-01 -2.68181562e-01 -1.16333914e+00 -2.72998631e-01
3.46147157e-02 -5.97966433e-01 8.17681432e-01 6.03906751e-01
1.70940268e+00 -2.36552581e-01 -6.61544561e-01 4.87018287e-01
1.18035078e+00 4.90370214e-01 8.27900529e-01 2.51190454e-01
6.44100547e-01 6.30374193e-01 6.54287994e-01 1.19875081e-01
6.34386903e-03 2.82068133e-01 5.70965469e-01 -9.58633304e-01
-1.86897531e-01 2.20541507e-01 -1.19139798e-01 1.01325440e+00
-6.88223913e-02 -1.01308890e-01 -1.17006266e+00 6.23000979e-01
-2.00432301e+00 -6.49184942e-01 -2.96205252e-01 1.92586076e+00
9.22435045e-01 1.18425325e-01 -7.10298568e-02 5.69922328e-01
6.18963420e-01 -3.33011776e-01 -6.86712325e-01 -1.50542304e-01
2.16590613e-01 2.81387657e-01 1.66414276e-01 7.08639398e-02
-1.17326510e+00 6.82559490e-01 6.09367895e+00 9.55186784e-01
-1.10262883e+00 3.07899415e-01 1.02572393e+00 2.42120698e-01
3.59934092e-01 -4.60873932e-01 -5.47248125e-01 4.10729676e-01
6.83321118e-01 1.63448080e-01 -3.76651466e-01 1.09990335e+00
2.96874017e-01 6.16436675e-02 -1.13027883e+00 1.10554338e+00
-4.58578430e-02 -1.54293072e+00 1.00014940e-01 9.40440372e-02
5.60946047e-01 -1.26485780e-01 -1.04264349e-01 4.40655857e-01
-2.41297498e-01 -1.11805952e+00 -6.28260002e-02 2.83703595e-01
7.24939346e-01 -5.22197604e-01 1.36894596e+00 4.39537495e-01
-7.48341620e-01 -2.56553054e-01 -3.67956996e-01 1.17343981e-02
-7.04305768e-02 7.10019112e-01 -1.02697778e+00 5.55783689e-01
6.45178139e-01 9.53783631e-01 -3.77549201e-01 1.27549160e+00
-8.92824754e-02 8.44396412e-01 3.05969298e-01 4.77331951e-02
1.26693487e-01 2.89119303e-01 1.08996280e-01 1.08043361e+00
-1.21912055e-01 3.58619213e-01 1.15012884e-01 5.87070048e-01
-1.22510292e-01 2.41094187e-01 -5.46692312e-01 -2.21517026e-01
2.41908073e-01 1.26095212e+00 -9.92599130e-01 -4.77489293e-01
-3.67031068e-01 7.77839482e-01 1.37374014e-01 2.00497597e-01
-5.14815748e-01 -2.98250616e-01 3.98315154e-02 1.27990529e-01
-2.04161823e-01 2.29956970e-01 -4.78573143e-01 -9.56101477e-01
-4.02541071e-01 -9.79590774e-01 5.10318339e-01 -7.21987367e-01
-1.49051666e+00 5.40938795e-01 -1.40804529e-01 -1.26401675e+00
-1.80020127e-02 -9.10614192e-01 -5.86359918e-01 6.08936965e-01
-1.45865810e+00 -1.13984752e+00 -4.64292347e-01 5.79773366e-01
9.27869499e-01 -5.10346472e-01 1.07853472e+00 6.65206969e-01
-7.24511087e-01 6.38834000e-01 -2.15599705e-02 5.37895560e-01
7.02300131e-01 -1.06641853e+00 -1.07831685e-02 2.98498422e-01
-1.08523071e-01 7.50227034e-01 3.76845539e-01 -4.13785249e-01
-1.06472445e+00 -1.10515642e+00 6.73080325e-01 -2.24797025e-01
2.71363169e-01 -1.73579440e-01 -1.08423233e+00 6.59665644e-01
4.22473997e-02 2.26827875e-01 1.26993501e+00 2.27731466e-01
2.98613459e-01 -7.56056830e-02 -1.22006595e+00 4.28265303e-01
6.89532459e-01 -1.47054628e-01 -6.04946434e-01 5.84929764e-01
3.78115118e-01 -3.56117725e-01 -7.42988169e-01 8.00929487e-01
4.17611390e-01 -7.42930472e-01 8.40174079e-01 -9.00347471e-01
6.16944969e-01 1.24326460e-01 2.85558850e-01 -8.55265737e-01
-4.09962565e-01 -5.09580113e-02 -2.66112119e-01 9.54484224e-01
4.91641700e-01 -4.70773935e-01 1.19268680e+00 4.84616190e-01
-5.34453094e-01 -1.61518049e+00 -8.11503470e-01 -3.39438349e-01
9.47168320e-02 -1.72318816e-01 1.46594465e-01 1.29424894e+00
1.21542372e-01 3.97887886e-01 -2.49082491e-01 -2.48556450e-01
5.92914701e-01 -3.53634387e-01 5.24334669e-01 -1.24385381e+00
-8.26718733e-02 -3.65065455e-01 -6.68068111e-01 -8.33055913e-01
-6.56117126e-02 -7.69579589e-01 5.10262772e-02 -1.81674111e+00
5.84963441e-01 -4.03471589e-01 -4.33885992e-01 7.97626197e-01
-3.53279114e-01 6.12799644e-01 -3.05113375e-01 5.01919091e-01
-5.23198366e-01 4.82408971e-01 1.27615905e+00 -2.34648719e-01
6.70147762e-02 1.24673106e-01 -8.85892034e-01 9.53845799e-01
7.81123519e-01 -4.59350854e-01 -5.49484491e-01 -2.74774134e-01
-4.65014428e-02 8.55707005e-02 3.75751942e-01 -7.12709248e-01
1.89631268e-01 9.92657468e-02 7.31754363e-01 -5.61503351e-01
2.11594924e-01 -7.58350730e-01 -2.60562748e-01 8.36322725e-01
-5.80888510e-01 -4.28260982e-01 1.40502557e-01 5.14184892e-01
-4.41420645e-01 -2.90208280e-01 9.95195985e-01 -4.96116966e-01
-7.86828637e-01 6.55109823e-01 -4.46522772e-01 -2.01302871e-01
1.02339530e+00 -5.42728662e-01 -5.09119742e-02 -6.53344840e-02
-1.00915194e+00 2.71181434e-01 -1.20216206e-01 1.87576920e-01
8.01391602e-01 -1.27396667e+00 -5.37071407e-01 5.19597381e-02
2.33059347e-01 4.08198297e-01 4.42647129e-01 1.23588109e+00
-6.01185858e-01 4.97619510e-01 -2.19657317e-01 -6.54618502e-01
-1.42725706e+00 5.59914649e-01 2.98134357e-01 -3.51810336e-01
-7.94698179e-01 1.08660114e+00 6.53242588e-01 -2.35465288e-01
5.41123986e-01 -1.22658640e-01 -5.33998787e-01 -4.04406972e-02
6.01852775e-01 1.22788846e-01 4.77780610e-01 -2.52166241e-01
-5.01228690e-01 1.70706242e-01 -5.10851622e-01 5.05914092e-01
1.43874538e+00 6.02100827e-02 2.33299192e-02 4.86336350e-01
1.27297950e+00 -5.73974252e-01 -7.50811577e-01 -2.93834448e-01
1.11778006e-02 -1.33010089e-01 2.84534216e-01 -1.00138342e+00
-1.27121699e+00 1.22054708e+00 9.89566267e-01 -4.61115269e-03
1.19189692e+00 -2.65168063e-02 8.18164051e-01 3.03741246e-01
2.96508372e-01 -8.30437422e-01 3.99551660e-01 -2.83678286e-02
7.09024310e-01 -1.70553803e+00 3.13045204e-01 -4.51950759e-01
-6.29209101e-01 1.17488861e+00 6.40774608e-01 1.31889775e-01
7.98696637e-01 2.60834783e-01 3.08705658e-01 -4.41389441e-01
-5.85459232e-01 4.17033955e-03 2.27171466e-01 7.10852861e-01
9.37392831e-01 -2.60597497e-01 -3.23002666e-01 7.61079550e-01
3.38839829e-01 4.36923414e-01 1.23455517e-01 8.93748701e-01
-4.14107263e-01 -1.00321007e+00 -1.52527615e-01 7.65056252e-01
-6.01220012e-01 -1.46031931e-01 -3.43133926e-01 9.16724622e-01
3.65303367e-01 7.85520017e-01 -9.13392678e-02 1.83366258e-02
9.85715166e-02 -1.10707238e-01 3.97465229e-01 -8.26237440e-01
-4.16480988e-01 1.34690404e-01 -1.18437514e-01 -9.37156379e-02
-6.06092274e-01 -3.31750125e-01 -1.41225708e+00 5.52895926e-02
-5.91687799e-01 2.49966055e-01 2.68733919e-01 1.22543550e+00
2.36016914e-01 6.58971965e-01 3.20130229e-01 -5.96826375e-01
-2.11763620e-01 -1.20103800e+00 -4.87607598e-01 1.61125898e-01
2.39641637e-01 -7.35328853e-01 -2.47050971e-01 2.81511754e-01] | [14.846482276916504, -2.2302772998809814] |
1b2f2f92-eedd-4863-9e9c-a8a2633d2a68 | nbd-gap-non-blind-image-deblurring-without | 2209.09498 | null | https://arxiv.org/abs/2209.09498v1 | https://arxiv.org/pdf/2209.09498v1.pdf | NBD-GAP: Non-Blind Image Deblurring Without Clean Target Images | In recent years, deep neural network-based restoration methods have achieved state-of-the-art results in various image deblurring tasks. However, one major drawback of deep learning-based deblurring networks is that large amounts of blurry-clean image pairs are required for training to achieve good performance. Moreover, deep networks often fail to perform well when the blurry images and the blur kernels during testing are very different from the ones used during training. This happens mainly because of the overfitting of the network parameters on the training data. In this work, we present a method that addresses these issues. We view the non-blind image deblurring problem as a denoising problem. To do so, we perform Wiener filtering on a pair of blurry images with the corresponding blur kernels. This results in a pair of images with colored noise. Hence, the deblurring problem is translated into a denoising problem. We then solve the denoising problem without using explicit clean target images. Extensive experiments are conducted to show that our method achieves results that are on par to the state-of-the-art non-blind deblurring works. | ['Vishal M. Patel', 'Rajeev Yasarla', 'Nithin Gopalakrishnan Nair'] | 2022-09-20 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 9.03371200e-02 -7.39291191e-01 4.64376509e-01 -6.03034832e-02
-5.06009400e-01 -3.10306311e-01 4.89668697e-01 -6.89393222e-01
-2.74235517e-01 8.40744674e-01 4.74536240e-01 -1.28554642e-01
-6.43616393e-02 -3.42817187e-01 -6.55060649e-01 -1.00404692e+00
2.06135660e-01 -1.33886606e-01 -1.77867904e-01 -7.56447464e-02
2.68217802e-01 1.50847226e-01 -1.31971705e+00 2.75204591e-02
1.24335575e+00 7.56672800e-01 4.64979291e-01 8.65972996e-01
1.93247855e-01 8.70906949e-01 -8.23526919e-01 -1.60195142e-01
3.34913731e-01 -7.20977128e-01 -6.98505759e-01 2.18587369e-01
6.80930495e-01 -9.04356122e-01 -8.81511271e-01 1.64502263e+00
6.09072030e-01 2.40227461e-01 4.61025417e-01 -7.38170624e-01
-1.28269255e+00 3.23928654e-01 -7.58173943e-01 3.41584712e-01
-1.33052558e-01 1.28717631e-01 4.07043636e-01 -9.95680034e-01
1.85814589e-01 1.13202417e+00 8.27235460e-01 6.02743864e-01
-1.17567742e+00 -4.90162641e-01 -3.44673127e-01 5.92340231e-01
-1.21207321e+00 -7.57313728e-01 8.09327841e-01 -3.66097242e-01
3.98525268e-01 2.84112096e-01 3.23170602e-01 9.84145999e-01
2.40738049e-01 4.49684769e-01 1.26238739e+00 -2.57885754e-01
3.53991091e-01 -3.11437666e-01 2.05002725e-01 2.67297268e-01
4.65657294e-01 1.93237886e-01 -2.21161112e-01 1.33956090e-01
1.02316833e+00 2.56011665e-01 -9.78948832e-01 -8.20715353e-02
-1.17257369e+00 4.81580079e-01 7.57557273e-01 5.60844600e-01
-5.82810700e-01 3.43604296e-01 1.30438104e-01 4.71048623e-01
6.63749337e-01 4.49408621e-01 -1.96774662e-01 -2.64732521e-02
-1.26487386e+00 1.14167921e-01 5.57241321e-01 3.66280049e-01
6.36503220e-01 2.74144650e-01 -1.08420841e-01 1.14254832e+00
1.65827617e-01 4.58554089e-01 7.58430183e-01 -9.75205541e-01
2.09945768e-01 -1.27889886e-01 5.43247163e-01 -1.09325302e+00
9.03246403e-02 -6.12764895e-01 -1.61204767e+00 2.75591105e-01
5.70239604e-01 -2.55819678e-01 -1.14398086e+00 1.56464767e+00
-2.29764208e-01 6.96461737e-01 3.02160949e-01 1.51066458e+00
6.42267048e-01 7.39716232e-01 -4.36346114e-01 -2.87076712e-01
1.25341129e+00 -1.32807386e+00 -1.16784668e+00 -5.70969105e-01
1.09265121e-02 -1.00768268e+00 8.26235354e-01 4.06160176e-01
-1.11442161e+00 -7.58377373e-01 -1.17377806e+00 1.73036437e-02
1.33809624e-02 1.60179272e-01 2.02016383e-01 4.99427080e-01
-1.24699056e+00 7.66948819e-01 -5.88114560e-01 -1.63247123e-01
3.86782765e-01 -9.48032439e-02 -3.52438807e-01 -6.68557584e-01
-1.31772220e+00 1.20376706e+00 8.19199607e-02 6.89824641e-01
-1.10620892e+00 -4.08226639e-01 -6.39108777e-01 2.51932591e-01
8.44835863e-02 -7.42621005e-01 1.28039896e+00 -1.19345200e+00
-1.25988722e+00 4.92183596e-01 -1.65255621e-01 -3.91679436e-01
6.68416560e-01 -7.40381598e-01 -4.56263244e-01 4.48157154e-02
-8.67514163e-02 6.38794005e-02 1.43288350e+00 -1.59364271e+00
-1.53041661e-01 -2.22387061e-01 -2.48798370e-01 1.42487854e-01
-1.05711848e-01 -1.10964313e-01 -4.37978148e-01 -9.86289561e-01
1.27119526e-01 -6.41083479e-01 -1.19425647e-01 -2.99627453e-01
-3.87357235e-01 4.72470701e-01 8.57520700e-01 -1.17072105e+00
1.11883080e+00 -2.34085131e+00 2.36826330e-01 -3.52002323e-01
5.10074735e-01 6.13325059e-01 -4.12906617e-01 2.02144757e-01
-3.50623101e-01 -3.65062915e-02 -3.31159502e-01 -4.87992436e-01
-4.04243380e-01 6.88811392e-02 -3.67742598e-01 8.23311448e-01
-1.17212757e-01 7.98406243e-01 -8.97388756e-01 1.10636473e-01
3.12453389e-01 6.32949650e-01 -1.05446965e-01 5.50916910e-01
1.79054186e-01 6.59371614e-01 1.10873103e-01 2.41341308e-01
1.25158727e+00 -2.65257865e-01 -1.33895546e-01 -4.44057047e-01
1.71000510e-02 -8.64186212e-02 -9.25053179e-01 1.67833543e+00
-4.79398191e-01 1.04668379e+00 3.77619654e-01 -1.01362717e+00
6.21493995e-01 4.33694869e-01 1.29543737e-01 -3.89616042e-01
1.15675539e-01 3.83454651e-01 2.02739872e-02 -8.90177846e-01
5.86831152e-01 -2.91625291e-01 5.28188825e-01 3.79813045e-01
-1.29304945e-01 -1.16072416e-01 -8.74862373e-02 -1.61827832e-01
1.12830412e+00 -2.22454891e-01 -1.19267203e-01 -2.56791055e-01
4.93896931e-01 -3.07642341e-01 4.41991299e-01 1.09239638e+00
-2.66764522e-01 1.17862713e+00 6.47178069e-02 -4.45228845e-01
-1.19657266e+00 -8.30317497e-01 9.00703445e-02 4.01537001e-01
4.26108778e-01 1.24316156e-01 -1.17541838e+00 -2.41875857e-01
-4.43362713e-01 6.18953943e-01 -5.92445195e-01 -2.42558956e-01
-5.87126136e-01 -1.03611684e+00 4.32484925e-01 1.75619617e-01
1.05162001e+00 -8.01327527e-01 -2.11648658e-01 1.79174155e-01
-4.72065032e-01 -1.05351162e+00 -7.44780183e-01 -1.57856271e-01
-9.15931940e-01 -1.08706379e+00 -1.57451487e+00 -9.44850862e-01
8.71483505e-01 9.90152597e-01 8.82272184e-01 1.98098198e-01
-9.71828122e-03 -2.11483270e-01 -2.32910186e-01 3.00409555e-01
-6.44809008e-01 -4.46841419e-01 3.93410698e-02 2.07401589e-01
1.63452663e-02 -6.72621906e-01 -9.31380510e-01 3.25151980e-01
-1.36367452e+00 1.37642026e-01 6.24199033e-01 1.20529389e+00
-9.02531967e-02 6.39216423e-01 4.14679766e-01 -2.78707057e-01
1.03921342e+00 -3.76825839e-01 -4.05747116e-01 2.86876708e-01
-5.33484757e-01 1.08421162e-01 6.57900333e-01 -7.07538366e-01
-1.09343612e+00 -4.53340232e-01 1.66072518e-01 -7.44050622e-01
-2.52767831e-01 6.73873186e-01 1.34024322e-01 2.96319202e-02
7.40745604e-01 4.23840880e-01 1.96276262e-01 -9.08526838e-01
3.72381538e-01 9.18826580e-01 8.85466099e-01 7.17964442e-03
9.74562109e-01 2.05269441e-01 -3.29504013e-01 -7.80657947e-01
-7.75380969e-01 -2.74465412e-01 -2.95715451e-01 -1.43504158e-01
6.67648137e-01 -1.04547334e+00 -2.62232482e-01 1.46000016e+00
-1.49314451e+00 -3.66996706e-01 4.74675149e-02 5.48635840e-01
-2.01278940e-01 8.09742451e-01 -9.92555380e-01 -6.35334671e-01
-3.59225065e-01 -1.31774938e+00 5.37388206e-01 4.35057133e-01
2.97663778e-01 -9.55063820e-01 1.43757761e-01 3.85550946e-01
1.01636016e+00 -1.63019329e-01 5.93549013e-01 -8.62146728e-04
-4.18631524e-01 -2.45064005e-01 -6.94537580e-01 8.80708575e-01
5.81288517e-01 -5.10199904e-01 -9.79383111e-01 -5.24145186e-01
6.41594052e-01 3.10364533e-02 1.08687389e+00 7.01851308e-01
1.00257564e+00 -4.74643230e-01 1.23076126e-01 6.05594575e-01
1.45577371e+00 -7.58820027e-02 1.01604569e+00 3.18252504e-01
7.63117433e-01 2.47576118e-01 6.35795221e-02 2.64962930e-02
-4.02665585e-02 6.16765082e-01 4.23334450e-01 -3.50409389e-01
-5.71466506e-01 1.18179724e-01 3.21685731e-01 7.85669923e-01
-9.96202813e-04 -3.47528309e-01 -7.62276113e-01 7.06288457e-01
-1.96165347e+00 -9.58022594e-01 -3.50631893e-01 2.22259641e+00
1.13182330e+00 -3.21680069e-01 -4.21587080e-01 6.06583171e-02
1.09328234e+00 3.92755955e-01 -5.39723933e-01 1.10766947e-01
-2.36063734e-01 -5.43814152e-02 4.01984304e-01 7.07675397e-01
-1.07991314e+00 8.21369946e-01 6.03263283e+00 7.00763285e-01
-1.19707870e+00 3.15014869e-01 7.27704167e-01 2.30967432e-01
4.75394502e-02 -1.13274463e-01 2.57120300e-02 6.93482578e-01
7.32401013e-01 9.97364223e-02 1.15368724e+00 3.54247153e-01
5.02137005e-01 -2.87780821e-01 -7.32719719e-01 1.42815530e+00
1.84509888e-01 -1.12594986e+00 -3.12736750e-01 -1.90622583e-01
1.02016282e+00 1.55238643e-01 9.76264253e-02 -1.64296627e-01
2.19732225e-01 -1.24546576e+00 5.28470814e-01 7.16782570e-01
7.86023855e-01 -1.80573285e-01 1.02558208e+00 2.81869859e-01
-4.53813076e-01 8.30727294e-02 -4.68783677e-01 -2.62295425e-01
2.23912582e-01 1.29777980e+00 -1.76475674e-01 5.23178279e-01
7.51053154e-01 8.37492228e-01 -3.98157299e-01 1.45932126e+00
-3.88527989e-01 5.76782465e-01 1.87229529e-01 5.09030402e-01
4.70665321e-02 -3.54159147e-01 5.91719329e-01 9.85232472e-01
6.00049615e-01 -6.88931197e-02 -5.60454011e-01 1.03457630e+00
-3.25283200e-01 -5.59453905e-01 -3.39366794e-01 1.17204480e-01
2.06277266e-01 9.96327817e-01 -2.89512277e-01 -4.98060137e-01
-2.76509672e-01 1.57177913e+00 -1.08775161e-01 8.70549560e-01
-6.71163797e-01 -3.95649284e-01 9.52646136e-01 -3.12003404e-01
2.92828262e-01 -3.26269448e-01 -2.49772176e-01 -1.73046982e+00
-1.28624275e-01 -1.09429240e+00 -3.71080786e-01 -1.25324261e+00
-1.58237994e+00 7.33179808e-01 -3.57684195e-01 -1.17963803e+00
7.64853880e-02 -2.40064338e-01 -6.80488050e-01 1.28336048e+00
-1.66317439e+00 -6.19679689e-01 -7.93278456e-01 3.09337318e-01
7.02082396e-01 6.78639933e-02 4.89999056e-01 5.05414844e-01
-6.33350551e-01 3.86271998e-02 6.10492766e-01 1.39653414e-01
9.63890493e-01 -1.08642399e+00 6.98205054e-01 1.49167621e+00
-1.88716203e-01 5.77565014e-01 1.09184289e+00 -5.40853620e-01
-1.24204528e+00 -9.16197300e-01 6.15230918e-01 -5.04337102e-02
6.13750458e-01 9.22888443e-02 -1.27922404e+00 2.25402847e-01
7.51443446e-01 1.85908780e-01 3.49480882e-02 -3.71247083e-01
-2.29639098e-01 -8.23852122e-02 -9.58221197e-01 5.82818031e-01
6.59627736e-01 -5.10765910e-01 -6.72130466e-01 1.15071826e-01
4.87938404e-01 -4.47412670e-01 -3.96856368e-01 2.18516842e-01
2.68069208e-01 -1.19240165e+00 1.03149867e+00 -2.52822399e-01
8.42556238e-01 -5.08744955e-01 1.27448604e-01 -1.91297054e+00
-5.36232471e-01 -6.60238743e-01 -3.20208669e-01 9.52534497e-01
-1.19799577e-01 -6.61985517e-01 3.14432651e-01 3.02711397e-01
4.53174636e-02 -2.13361248e-01 -6.95825040e-01 -6.81258440e-01
3.21694873e-02 6.42823875e-02 4.65836346e-01 1.09848595e+00
-4.90857124e-01 2.34918028e-01 -1.02097404e+00 2.73498535e-01
9.27218139e-01 1.96765512e-02 5.69762051e-01 -8.29748631e-01
-2.64401317e-01 -4.94823486e-01 3.82108688e-02 -1.37465096e+00
-6.71998113e-02 -2.34746456e-01 5.81017375e-01 -1.64151454e+00
3.23848009e-01 -2.12715715e-01 -2.21295521e-01 2.43985295e-01
-7.12804854e-01 4.47347969e-01 1.04953639e-01 6.95798576e-01
-8.95059407e-02 5.68157613e-01 1.38487899e+00 -2.68842340e-01
-1.02084242e-01 -1.01471908e-01 -7.70227969e-01 4.16326374e-01
9.24306154e-01 -3.24355125e-01 -1.87130779e-01 -1.04718196e+00
-1.10649191e-01 1.52741835e-01 6.55549228e-01 -1.06434262e+00
2.03914434e-01 -6.98277121e-03 3.46460968e-01 -1.88440084e-01
2.16506511e-01 -7.13076890e-01 3.89886230e-01 4.69853044e-01
-2.43925437e-01 -1.29216462e-01 1.46320596e-01 5.78023136e-01
-4.25453663e-01 -5.00696003e-01 1.01075602e+00 -1.33645967e-01
-4.06350821e-01 -4.39437442e-02 -4.53373939e-01 -1.80452406e-01
4.00237411e-01 4.24753549e-03 -7.20907450e-01 -7.30681598e-01
-5.21975160e-01 -2.64664471e-01 7.21070051e-01 3.52826267e-01
6.41290069e-01 -1.11404848e+00 -9.08705592e-01 3.04639339e-01
-4.87254500e-01 -3.75257693e-02 5.57292879e-01 1.03934753e+00
-6.69475794e-01 6.79384172e-02 -2.65709907e-01 -2.10535690e-01
-1.11748672e+00 6.22769058e-01 6.36025310e-01 -9.45504457e-02
-5.22423208e-01 7.05158949e-01 3.05610567e-01 7.16727525e-02
1.71304226e-01 -1.80288747e-01 3.62707227e-02 -3.45959127e-01
8.56578171e-01 5.04878163e-01 1.17860563e-01 -4.84678745e-01
-1.38403382e-02 4.59887594e-01 -1.91940181e-02 -3.02374717e-02
1.42313588e+00 -5.00355244e-01 -4.85258907e-01 -6.47900924e-02
1.18295121e+00 8.61510728e-03 -1.49418390e+00 -3.45688105e-01
-3.90819401e-01 -7.12347746e-01 6.84778571e-01 -9.22923625e-01
-1.41222835e+00 8.66851568e-01 9.00889695e-01 4.10940617e-01
1.37813020e+00 -3.75866354e-01 9.59164798e-01 5.27710505e-02
-1.49839044e-01 -6.75102651e-01 1.37155995e-01 4.38560069e-01
1.13616586e+00 -1.32940876e+00 -7.83808008e-02 -5.15125983e-04
-2.38473609e-01 1.16068733e+00 2.59520203e-01 -2.89582670e-01
4.76530939e-01 3.08161741e-03 3.74757946e-01 4.43877280e-02
-2.23992810e-01 6.00141427e-03 2.85973459e-01 3.69154096e-01
1.28424063e-01 -1.28678858e-01 -1.06136322e-01 1.85718969e-01
1.95511758e-01 2.42577538e-01 7.39728749e-01 5.38325489e-01
-4.42765117e-01 -8.64185572e-01 -9.99429822e-01 1.37379318e-01
-5.26982605e-01 -4.32181269e-01 -5.54127991e-02 1.01875149e-01
-1.71677634e-01 1.26223540e+00 -1.99337989e-01 -3.36652637e-01
-4.06056419e-02 -3.31846297e-01 4.26815003e-01 -1.84714377e-01
-2.35933676e-01 1.53039351e-01 -3.16985816e-01 -2.94998467e-01
-5.20781517e-01 -2.50063807e-01 -5.70944071e-01 -5.94795048e-01
-5.88057637e-01 5.88719994e-02 5.95017076e-01 9.28843796e-01
3.69676679e-01 6.16602242e-01 6.51032507e-01 -1.24687052e+00
-7.55872786e-01 -1.32842946e+00 -6.32908762e-01 5.43512940e-01
1.04890943e+00 -3.44600141e-01 -7.51259327e-01 3.38367581e-01] | [11.601874351501465, -2.6881301403045654] |
ffa9d278-68d5-4a54-a8c6-dd9008abc453 | lidar-flow-dense-scene-flow-estimation-from | 1910.14453 | null | https://arxiv.org/abs/1910.14453v2 | https://arxiv.org/pdf/1910.14453v2.pdf | LiDAR-Flow: Dense Scene Flow Estimation from Sparse LiDAR and Stereo Images | We propose a new approach called LiDAR-Flow to robustly estimate a dense scene flow by fusing a sparse LiDAR with stereo images. We take the advantage of the high accuracy of LiDAR to resolve the lack of information in some regions of stereo images due to textureless objects, shadows, ill-conditioned light environment and many more. Additionally, this fusion can overcome the difficulty of matching unstructured 3D points between LiDAR-only scans. Our LiDAR-Flow approach consists of three main steps; each of them exploits LiDAR measurements. First, we build strong seeds from LiDAR to enhance the robustness of matches between stereo images. The imagery part seeks the motion matches and increases the density of scene flow estimation. Then, a consistency check employs LiDAR seeds to remove the possible mismatches. Finally, LiDAR measurements constraint the edge-preserving interpolation method to fill the remaining gaps. In our evaluation we investigate the individual processing steps of our LiDAR-Flow approach and demonstrate the superior performance compared to image-only approach. | ['Oliver Wasenmüller', 'René Schuster', 'Ramy Battrawy', 'Didier Stricker', 'Qing Rao'] | 2019-10-31 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [ 3.62054527e-01 -5.79796314e-01 3.94608639e-02 -3.15994352e-01
-5.89494169e-01 -4.30889279e-01 3.85240763e-01 -1.75846238e-02
-3.85696381e-01 9.07437563e-01 -2.65286397e-02 -1.42894655e-01
-1.11322261e-01 -9.89544451e-01 -4.02616948e-01 -3.30855697e-01
2.81259626e-01 5.56025803e-01 9.28854465e-01 -1.85588319e-02
5.19390941e-01 9.21829939e-01 -1.91999424e+00 -5.62640466e-02
1.24017406e+00 9.37114835e-01 5.23262739e-01 4.05675679e-01
-6.99060023e-01 3.79665315e-01 4.43635285e-02 -4.33043204e-02
6.39171720e-01 -2.83590127e-02 -4.13449705e-01 2.97831267e-01
8.63273203e-01 -7.07572639e-01 -1.64431065e-01 1.13040519e+00
2.94446021e-01 8.45512301e-02 4.38586712e-01 -1.29307973e+00
1.33092836e-01 -2.82876104e-01 -1.13098693e+00 2.05793530e-01
4.74451572e-01 2.49519572e-01 5.84509790e-01 -1.31024563e+00
6.84026718e-01 1.24504268e+00 6.65773988e-01 -1.66059345e-01
-1.06735837e+00 -8.48273337e-01 -1.78343177e-01 1.56317919e-01
-1.63834786e+00 -6.53403103e-01 9.81409907e-01 -4.93875444e-01
3.61675769e-01 1.07353911e-01 6.82181835e-01 2.72561342e-01
1.91769358e-02 1.77361965e-01 1.12772751e+00 -3.42530280e-01
2.91666985e-02 1.14489280e-01 6.25446439e-02 8.77137780e-01
4.94736522e-01 4.08383578e-01 -6.20872855e-01 -1.35189041e-01
9.65747476e-01 1.98818758e-01 -3.94807607e-01 -5.37566960e-01
-1.04199994e+00 5.55473804e-01 4.23967838e-01 -1.47707006e-02
-3.84916961e-01 -1.08494170e-01 1.40344605e-01 -1.37524262e-01
2.51619726e-01 -1.95857257e-01 -6.19381703e-02 7.62669444e-02
-1.22261322e+00 7.40949735e-02 2.61089593e-01 1.03741813e+00
1.54197896e+00 3.09322178e-02 3.48118305e-01 5.43339491e-01
3.16550016e-01 7.06284404e-01 5.70500046e-02 -1.16788077e+00
8.07798743e-01 6.41025722e-01 5.16766869e-02 -1.14985454e+00
-1.62494443e-02 -1.28331497e-01 -7.59916067e-01 3.43769252e-01
3.11906517e-01 2.09925443e-01 -6.29151106e-01 9.85768557e-01
6.97066188e-01 6.35588169e-01 -2.18531340e-01 7.94622362e-01
6.68071210e-01 4.81881291e-01 -1.93989128e-01 -3.29869539e-01
1.11455810e+00 -5.52015722e-01 -5.19487083e-01 -1.81703776e-01
1.54041916e-01 -1.16978765e+00 8.21699500e-01 1.28527600e-02
-1.03683734e+00 -8.28850985e-01 -9.37818527e-01 -7.04591423e-02
-2.01196626e-01 7.63007766e-03 4.14537281e-01 4.20716375e-01
-7.20123112e-01 4.42231059e-01 -5.76835692e-01 -2.04167888e-01
3.17435354e-01 2.36785293e-01 -3.99916083e-01 -4.02015686e-01
-8.15708041e-01 6.10719144e-01 3.12790364e-01 2.34545141e-01
-2.73695737e-01 -8.50679934e-01 -1.00694215e+00 -7.09619597e-02
4.19450283e-01 -7.79516757e-01 6.57335162e-01 -4.12154645e-01
-1.01135933e+00 7.35925555e-01 -7.22013414e-01 -1.46430299e-01
7.60358930e-01 -1.55445546e-01 7.77905360e-02 2.31501907e-01
4.03703779e-01 7.00770080e-01 6.56720340e-01 -1.40789580e+00
-9.04892683e-01 -4.30175900e-01 -2.43391544e-01 2.38469973e-01
8.86083245e-02 -4.06885922e-01 -5.40033937e-01 -1.38922960e-01
5.47981441e-01 -6.44169629e-01 -3.03961009e-01 3.76324743e-01
-1.34470433e-01 2.90720969e-01 1.26466990e+00 -2.82076448e-01
9.97066498e-01 -2.13224816e+00 -4.69474375e-01 2.40893304e-01
-3.01594716e-02 1.59698471e-01 -8.86647031e-03 3.77479315e-01
3.33641499e-01 -1.55140057e-01 -4.37010944e-01 -4.78901327e-01
-6.39809847e-01 3.34972948e-01 -4.72383648e-01 4.75584716e-01
3.17007393e-01 4.85688955e-01 -7.01622427e-01 -1.10790813e+00
8.35968196e-01 4.91161495e-01 -4.56417412e-01 3.11844237e-02
5.29149175e-02 6.28217757e-01 -6.39077723e-01 8.73459637e-01
1.41399705e+00 2.33280718e-01 -3.02785516e-01 -5.16833603e-01
-7.99995482e-01 2.48573348e-01 -1.88776028e+00 1.74317384e+00
-4.03457403e-01 4.26882595e-01 4.34386939e-01 -5.31302392e-01
1.04599667e+00 -9.12789926e-02 7.10551977e-01 -4.54615593e-01
-2.61994172e-02 3.94863039e-01 -3.46185744e-01 -4.36970472e-01
7.25743711e-01 -6.30387217e-02 4.56027001e-01 2.51491874e-01
-4.12477016e-01 -6.84727490e-01 9.83860642e-02 1.28024668e-01
7.81128824e-01 2.89309293e-01 2.48091653e-01 -2.18326777e-01
9.91151750e-01 2.14250743e-01 7.74660051e-01 4.86995906e-01
-2.56034344e-01 7.57204056e-01 -1.65461540e-01 -3.80054653e-01
-9.54462230e-01 -1.11316347e+00 -4.08727229e-01 1.39231458e-01
5.46340764e-01 -2.91035026e-01 -1.53684437e-01 -4.21694309e-01
2.79529452e-01 1.57637790e-01 -1.42086864e-01 3.87725979e-01
-7.48684466e-01 -3.16170037e-01 -3.75462184e-03 4.24517542e-01
1.02061093e+00 -6.68046832e-01 -9.07439411e-01 8.10928941e-02
-2.47758955e-01 -1.34018004e+00 -4.75608468e-01 -2.40912572e-01
-1.22905779e+00 -1.14075518e+00 -2.49271110e-01 -5.78755260e-01
5.72612762e-01 9.62225199e-01 8.90258729e-01 3.86546522e-01
-4.52196807e-01 1.03248872e-01 -7.71514606e-03 -6.83663264e-02
-3.05670537e-02 -2.09983945e-01 -1.11535728e-01 2.33794432e-02
1.57944813e-01 -7.80377507e-01 -6.09584928e-01 4.37628120e-01
-6.32353604e-01 5.56268468e-02 3.91743034e-01 5.72276294e-01
7.11637974e-01 1.57790348e-01 -1.10909259e-02 -5.11281252e-01
1.53870583e-01 -1.36594936e-01 -9.85132277e-01 -6.82211146e-02
-2.57469267e-01 -7.76698515e-02 1.85102969e-01 -9.47953388e-02
-1.05339801e+00 6.58362389e-01 1.65295508e-02 -7.08956480e-01
-1.55622408e-01 1.25104591e-01 -7.12737963e-02 -3.18558872e-01
2.67811328e-01 2.08217680e-01 4.36753854e-02 -4.12909955e-01
1.90998748e-01 6.30685985e-01 7.21831441e-01 -5.97802103e-01
1.31586730e+00 1.05649745e+00 4.91973877e-01 -1.20041490e+00
-5.18829942e-01 -8.92223120e-01 -1.06265366e+00 -3.73346120e-01
7.62773693e-01 -8.65803778e-01 -6.05007827e-01 2.42487505e-01
-1.40680790e+00 2.89548248e-01 -2.82162547e-01 7.27996826e-01
-3.25304657e-01 7.94680178e-01 -4.16331202e-01 -1.12231338e+00
-1.91118136e-01 -1.23996949e+00 1.25871825e+00 3.03664923e-01
1.61771536e-01 -6.29222810e-01 1.87534858e-02 3.76773924e-01
1.45355627e-01 1.74325794e-01 5.35041451e-01 3.32392782e-01
-1.26410937e+00 9.10371915e-02 -5.71281135e-01 6.13114201e-02
3.17512691e-01 4.13508117e-01 -1.10088074e+00 -1.43110538e-02
-5.75706735e-02 5.68084344e-02 7.69055724e-01 2.76688010e-01
7.03172207e-01 4.00972754e-01 -4.31502581e-01 7.32968450e-01
1.79538965e+00 1.93739384e-01 7.09247828e-01 2.74998754e-01
7.25195467e-01 8.08025241e-01 1.05216038e+00 3.09923619e-01
4.35962081e-01 6.32329941e-01 5.09994268e-01 -2.11656466e-01
-2.00305223e-01 -3.64462018e-01 2.63551883e-02 6.81438446e-01
-6.04900122e-02 2.49709353e-01 -8.15211415e-01 5.21639466e-01
-1.70059383e+00 -9.83751178e-01 -7.34172881e-01 2.53586149e+00
4.03396457e-01 2.04480737e-01 -1.78045154e-01 2.96120763e-01
7.78523862e-01 1.96584195e-01 -3.38120878e-01 -6.86264262e-02
-2.19005179e-02 2.53961563e-01 6.91348910e-01 9.91745532e-01
-8.61059010e-01 1.00600183e+00 5.75356293e+00 5.37324548e-01
-9.75106537e-01 -7.39505738e-02 -9.25021898e-03 2.94496655e-01
-3.56355608e-01 3.91263276e-01 -9.96445537e-01 5.15756667e-01
2.01234102e-01 8.21335167e-02 -5.08900732e-03 4.10131037e-01
4.97086108e-01 -7.79800594e-01 -5.87927639e-01 1.06765485e+00
-6.48205578e-02 -1.26814842e+00 3.86806461e-03 2.99472064e-01
6.03102505e-01 -1.02196634e-02 -3.17260861e-01 -3.30217361e-01
4.53285910e-02 -5.44935167e-01 5.12481749e-01 5.37977993e-01
6.95904016e-01 -6.41588032e-01 4.76427943e-01 5.54398656e-01
-1.78173780e+00 1.02393359e-01 -5.47402740e-01 -1.61129758e-01
6.24023497e-01 1.07395864e+00 -6.88983679e-01 8.08847249e-01
5.66105664e-01 6.83913052e-01 -4.40659255e-01 1.27816975e+00
-9.97792091e-03 -5.51265553e-02 -6.07561827e-01 5.75427353e-01
1.00465305e-01 -8.30715477e-01 6.34209454e-01 9.13119435e-01
5.99775434e-01 1.59908578e-01 5.11916399e-01 9.80570138e-01
3.41869920e-01 1.04154564e-01 -9.86215413e-01 5.96575022e-01
8.19321811e-01 1.20029223e+00 -7.17842937e-01 -3.51102293e-01
-5.41037261e-01 8.47357750e-01 -6.59518735e-03 1.03719540e-01
-5.60613215e-01 -4.40504365e-02 6.16231382e-01 6.42136931e-01
1.92741007e-01 -6.92269564e-01 -4.25563991e-01 -1.08675408e+00
1.73538357e-01 -2.63431102e-01 1.66395873e-01 -1.03982258e+00
-9.13584054e-01 3.24098676e-01 5.96944764e-02 -1.58577859e+00
-9.38322768e-02 -1.57118186e-01 -7.23753035e-01 1.10383749e+00
-2.03178382e+00 -1.06893373e+00 -7.94552743e-01 7.38330245e-01
7.41423309e-01 3.48895192e-01 2.79721141e-01 4.81206030e-01
-3.09360147e-01 -2.87873149e-01 -2.25321904e-01 -1.15656272e-01
7.70287752e-01 -7.62333333e-01 1.64030656e-01 1.16013038e+00
-2.31659482e-03 5.27382731e-01 4.56849843e-01 -1.02034247e+00
-1.14709914e+00 -8.36208344e-01 9.77919459e-01 -1.46061599e-01
3.97807091e-01 -2.47347131e-01 -1.02385092e+00 3.81083667e-01
-9.69111249e-02 2.21887767e-01 4.35447581e-02 -3.49477649e-01
-3.56818573e-03 -3.43712807e-01 -1.18838918e+00 2.92887151e-01
1.05704367e+00 -4.62345958e-01 -5.86281657e-01 -1.52143225e-01
3.55031163e-01 -4.18370664e-01 -5.09265602e-01 6.38481557e-01
4.92417455e-01 -1.52200043e+00 1.23652565e+00 9.33447629e-02
2.43530124e-01 -8.57698739e-01 -1.51830226e-01 -8.19835663e-01
1.08163645e-02 -3.47923160e-01 3.41714293e-01 1.43258274e+00
-1.70753524e-01 -5.58850825e-01 9.48208809e-01 2.93347985e-01
-4.41535972e-02 -1.55991822e-01 -9.91630435e-01 -5.71454406e-01
-4.76400793e-01 -3.65342259e-01 5.00812769e-01 8.69779944e-01
-5.29013991e-01 3.05027574e-01 -1.82033479e-01 3.51251125e-01
1.09023428e+00 5.10884643e-01 1.00711715e+00 -1.53019702e+00
8.86463150e-02 3.86766461e-03 -3.99857402e-01 -1.14550507e+00
2.99057215e-02 -5.29213250e-01 9.51708108e-02 -1.40096927e+00
4.97636087e-02 -8.82967234e-01 3.28275025e-01 -3.50669995e-02
-2.06984133e-01 4.81797874e-01 2.93634027e-01 4.37426090e-01
-5.04508279e-02 4.29133505e-01 1.31461108e+00 1.68349668e-01
-3.75943899e-01 -4.77968007e-02 -1.44681811e-01 8.53727877e-01
5.06314814e-01 -5.12460709e-01 -4.43458378e-01 -4.22816515e-01
-2.21451223e-01 2.29794800e-01 5.02054572e-01 -1.25955486e+00
4.07653034e-01 -3.63728940e-01 3.90752971e-01 -1.35159183e+00
5.66708446e-01 -1.17949569e+00 3.26764017e-01 5.04438698e-01
4.09284234e-01 2.76272446e-01 1.17063589e-01 4.47639644e-01
-3.23083162e-01 -4.36083257e-01 9.98887002e-01 -3.04042161e-01
-7.42405772e-01 4.89880085e-01 -3.75343226e-02 -1.79755360e-01
9.11678612e-01 -7.93355644e-01 -3.68083902e-02 -3.08786303e-01
-1.47926748e-01 2.66985148e-01 7.05458283e-01 5.89573793e-02
8.26814115e-01 -1.25560701e+00 -5.28495371e-01 6.86367810e-01
-1.22498488e-02 4.43682104e-01 1.90047830e-01 9.35568213e-01
-7.69921064e-01 2.62577981e-01 -2.32811317e-01 -1.01131892e+00
-1.50733447e+00 3.92717510e-01 1.71059266e-01 -8.44220519e-02
-6.71701252e-01 3.15677613e-01 1.68213204e-01 -1.97844177e-01
-7.66854882e-02 -3.17450374e-01 -4.14045416e-02 4.19901982e-02
4.35361624e-01 6.03327394e-01 5.92955500e-02 -9.51384723e-01
-4.68706250e-01 1.43465710e+00 2.84699887e-01 -2.95890272e-01
9.15943384e-01 -4.67731923e-01 -3.02837510e-02 2.09086657e-01
1.03458142e+00 5.25916636e-01 -1.25222719e+00 -2.29208827e-01
-2.69425869e-01 -1.13045871e+00 1.70663550e-01 -7.18371347e-02
-1.08233202e+00 1.19562745e+00 4.49407071e-01 -9.79604572e-02
1.04484642e+00 -3.85327667e-01 8.20669949e-01 -7.12785171e-03
4.99656796e-01 -8.34736705e-01 -2.65517443e-01 4.95956391e-01
4.65356886e-01 -1.19574714e+00 3.46107751e-01 -1.07487476e+00
-2.39454910e-01 1.24356234e+00 5.57925999e-01 -7.69705474e-02
5.06178737e-01 4.58855867e-01 -1.90421902e-02 -8.70382637e-02
-2.61542648e-01 -6.83582008e-01 -1.56052606e-02 7.05526233e-01
4.73545538e-03 -5.46851218e-01 -3.03082079e-01 -4.95166630e-01
-4.89930920e-02 2.40492955e-01 4.36700642e-01 9.16083992e-01
-8.16826165e-01 -1.15594387e+00 -7.31787264e-01 1.18558079e-01
1.33832142e-01 -7.72985369e-02 -7.17496723e-02 8.76950264e-01
3.33353788e-01 9.11492586e-01 4.65044886e-01 -1.58884317e-01
4.68708098e-01 -1.65741250e-01 4.45217818e-01 -4.18605059e-01
-2.53031671e-01 3.44259918e-01 -1.44901752e-01 -7.69686282e-01
-6.50696695e-01 -8.44509482e-01 -1.18537807e+00 -2.93753952e-01
-4.12233889e-01 1.16842061e-01 6.82969511e-01 6.54893517e-01
2.56034523e-01 -8.75594765e-02 8.82076800e-01 -9.51832354e-01
-2.18640387e-01 -5.43561339e-01 -5.20885706e-01 3.01564693e-01
4.44392145e-01 -9.01149273e-01 -6.29751325e-01 1.08210117e-01] | [8.626466751098633, -2.481154203414917] |
47b3232b-e81f-45f3-a869-936981c318c1 | collective-intelligence-for-object | 2211.15136 | null | https://arxiv.org/abs/2211.15136v3 | https://arxiv.org/pdf/2211.15136v3.pdf | Collective Intelligence for 2D Push Manipulations with Mobile Robots | While natural systems often present collective intelligence that allows them to self-organize and adapt to changes, the equivalent is missing in most artificial systems. We explore the possibility of such a system in the context of cooperative 2D push manipulations using mobile robots. Although conventional works demonstrate potential solutions for the problem in restricted settings, they have computational and learning difficulties. More importantly, these systems do not possess the ability to adapt when facing environmental changes. In this work, we show that by distilling a planner derived from a differentiable soft-body physics simulator into an attention-based neural network, our multi-robot push manipulation system achieves better performance than baselines. In addition, our system also generalizes to configurations not seen during training and is able to adapt toward task completions when external turbulence and environmental changes are applied. Supplementary videos can be found on our project website: https://sites.google.com/view/ciom/home | ['Yujin Tang', 'Shixiang Shane Gu', 'Yutaka Matsuo', 'Hiroki Furuta', 'Jumpei Arima', 'Tatsuya Matsushima', 'So Kuroki'] | 2022-11-28 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [-1.14889920e-01 3.05994749e-01 3.20362672e-02 4.36640196e-02
-1.52534798e-01 -6.63535058e-01 4.87939179e-01 -3.88315544e-02
-3.74355018e-01 7.68596828e-01 -8.63907801e-04 -1.36820376e-01
-4.50944960e-01 -5.22446871e-01 -9.30698216e-01 -5.76101780e-01
-3.35956275e-01 8.28697383e-01 3.20841998e-01 -7.48377144e-01
2.58107275e-01 4.29462701e-01 -1.44765317e+00 -4.69224751e-01
9.26611364e-01 2.21539348e-01 8.83777022e-01 7.65274286e-01
4.88985926e-01 5.61308682e-01 -2.07559392e-01 4.26824987e-01
6.63793147e-01 -1.47978246e-01 -7.58649707e-01 -1.95647799e-03
2.01207563e-01 -1.12400621e-01 -5.82482815e-01 7.75304615e-01
6.14675105e-01 8.40851665e-01 5.56555629e-01 -1.30472195e+00
-6.64033353e-01 3.92557621e-01 -1.06568627e-01 2.00670779e-01
3.81190628e-01 6.69524968e-01 6.33417189e-01 -5.35344541e-01
6.22009218e-01 1.22926521e+00 6.05861962e-01 6.60399556e-01
-1.16467011e+00 -2.60628730e-01 2.51960397e-01 6.33190274e-02
-8.96982074e-01 -4.71334308e-01 5.16786456e-01 -3.63898069e-01
1.22955155e+00 -1.35674596e-01 7.16855943e-01 1.18445516e+00
5.81582308e-01 6.58856392e-01 8.33991289e-01 -1.03110924e-01
3.89630824e-01 -2.63101816e-01 -2.19618902e-01 7.51654267e-01
3.65177929e-01 1.11904658e-01 -4.59856212e-01 -1.37257865e-02
8.55142236e-01 -3.59765068e-02 -2.80884117e-01 -7.95917571e-01
-1.53678131e+00 4.04424757e-01 6.70062602e-01 2.27520600e-01
-7.57605195e-01 5.24588704e-01 7.73645118e-02 4.08239603e-01
-1.44390268e-02 1.18263412e+00 -5.31908393e-01 -2.41421133e-01
-2.79950619e-01 5.55955887e-01 1.02713406e+00 1.05713391e+00
5.64466298e-01 4.36980743e-03 1.99672639e-01 4.77300167e-01
7.25332871e-02 5.30203342e-01 5.32402992e-01 -1.43203747e+00
4.01607424e-01 4.14971113e-01 4.01706994e-01 -1.05549788e+00
-9.56122816e-01 -4.16804403e-01 -4.08542931e-01 3.67547691e-01
2.54621983e-01 -4.92810607e-01 -5.43572247e-01 1.78381348e+00
6.33000612e-01 -5.62703013e-02 2.77343422e-01 1.03732944e+00
2.34673396e-01 4.43164229e-01 -1.64172858e-01 1.74732536e-01
8.52017999e-01 -1.31751084e+00 -6.19805694e-01 -6.13637924e-01
4.12954330e-01 -2.02408761e-01 1.04202902e+00 2.74908811e-01
-1.38252926e+00 -5.00489593e-01 -1.00077379e+00 3.62056233e-02
-2.57088065e-01 -3.66979927e-01 7.25083470e-01 -1.77686363e-02
-1.34844828e+00 8.66668224e-01 -1.44143856e+00 -9.02059317e-01
1.15166984e-01 5.47450244e-01 -4.30071801e-01 2.38208830e-01
-8.92039239e-01 1.34286785e+00 2.81582773e-01 8.71090740e-02
-1.05545139e+00 -3.38773698e-01 -6.24425411e-01 -8.27463120e-02
7.95984328e-01 -1.13338053e+00 1.64283323e+00 -6.54435098e-01
-1.95111144e+00 1.58849373e-01 1.13381565e-01 -3.06523591e-01
5.69297373e-01 -3.38427067e-01 4.72433418e-01 4.55858223e-02
6.92207515e-02 6.66185439e-01 5.29118299e-01 -1.16866529e+00
-4.28010553e-01 -1.71686932e-01 4.36099648e-01 8.44006598e-01
-1.92682579e-01 -3.67587686e-01 -1.89183652e-01 -2.91455150e-01
9.32487175e-02 -1.49197650e+00 -6.85544610e-01 -6.48054481e-02
-9.24113616e-02 -1.41708508e-01 5.55275202e-01 -2.22289622e-01
3.68963182e-01 -1.84282279e+00 9.17698443e-01 -2.02474490e-01
5.58494553e-02 9.86818373e-02 -2.37343922e-01 8.82552683e-01
3.61341089e-01 1.13976069e-01 -1.51620865e-01 -3.93486142e-01
2.79080868e-01 3.77588511e-01 -8.23024139e-02 4.30394679e-01
2.51616277e-02 9.33558702e-01 -1.19915557e+00 1.65100936e-02
-4.09880280e-02 1.78671330e-01 -8.48727942e-01 -2.74953749e-02
-4.55083370e-01 8.97743046e-01 -7.81879425e-01 3.37067366e-01
2.58723259e-01 -2.46686131e-01 2.19201431e-01 6.13207519e-01
-2.24443823e-01 1.02408081e-01 -1.10128236e+00 1.85413015e+00
-4.97945160e-01 5.46356797e-01 4.90393251e-01 -1.06089520e+00
6.57037437e-01 2.11271182e-01 4.87058014e-01 -4.03723061e-01
1.43480793e-01 2.38387764e-01 2.38814041e-01 -8.64929497e-01
8.17024529e-01 5.95935248e-02 -7.10900947e-02 5.07652402e-01
1.36381965e-02 -6.34017050e-01 3.91377620e-02 7.04497918e-02
1.59457064e+00 3.30632985e-01 1.62513465e-01 -2.14330524e-01
1.77703023e-01 3.50304246e-01 4.32184845e-01 1.12074268e+00
-3.35904747e-01 3.50467861e-01 1.44441158e-01 -2.64206678e-01
-1.07621646e+00 -9.48173940e-01 1.41428649e-01 1.18377948e+00
6.01024508e-01 -1.39077261e-01 -5.29204905e-01 -9.24788043e-02
2.01104730e-01 6.10085607e-01 -4.49380100e-01 -1.20060362e-01
-8.33289802e-01 -5.47737300e-01 2.41097540e-01 4.12167996e-01
4.73302484e-01 -1.45286763e+00 -1.23688817e+00 2.77893424e-01
-2.05791056e-01 -1.03234601e+00 -1.36915758e-01 2.26725802e-01
-7.28576839e-01 -9.72182572e-01 -5.47566593e-01 -8.07311833e-01
5.62123954e-01 5.12137294e-01 5.91522455e-01 2.53202438e-01
-9.66402292e-02 1.03360105e+00 -4.10479903e-01 -4.85808671e-01
-3.66681248e-01 4.13195342e-01 6.71410024e-01 -6.99001670e-01
-4.00637031e-01 -8.03950369e-01 -6.46637142e-01 4.97048557e-01
-7.08617926e-01 1.95372358e-01 4.66866076e-01 8.21213484e-01
3.18144381e-01 3.54752876e-02 5.66610992e-01 -1.21512733e-01
9.03612077e-01 -7.56161332e-01 -3.33168447e-01 -8.77286047e-02
-2.91642517e-01 -4.32827249e-02 6.57898903e-01 -5.67454338e-01
-9.25452352e-01 2.28866175e-01 3.45363319e-01 -2.38898605e-01
-3.61940980e-01 6.23782516e-01 2.44663343e-01 -3.36342990e-01
7.59719133e-01 1.26791984e-01 2.50848740e-01 -3.04387093e-01
3.82994711e-01 4.25163954e-01 4.33672249e-01 -8.99767518e-01
7.93906569e-01 3.21738809e-01 1.47814661e-01 -7.46122181e-01
-2.01258808e-01 -1.39130265e-01 -7.49428570e-01 -3.09552491e-01
6.82640254e-01 -6.50320530e-01 -8.11435819e-01 5.49847007e-01
-9.18066025e-01 -1.17500854e+00 -2.56684810e-01 5.17589331e-01
-1.13908446e+00 3.47752392e-01 -5.27757704e-01 -6.67506039e-01
2.22715344e-02 -1.10557842e+00 8.46770108e-01 5.30542791e-01
-1.41787380e-01 -9.03803408e-01 2.47997165e-01 1.83301061e-01
8.73298943e-01 2.75835961e-01 3.63819778e-01 -6.08099580e-01
-5.95764816e-01 -2.00392410e-01 5.58569849e-01 -2.61869043e-01
1.28249034e-01 -2.12238178e-01 -3.23598117e-01 -5.72201669e-01
-1.09805152e-01 -4.62338150e-01 6.33159280e-01 3.52108091e-01
7.28073359e-01 -3.56060088e-01 -7.17691481e-01 4.19232965e-01
8.63377333e-01 2.05645472e-01 1.10393189e-01 6.70651615e-01
3.29240918e-01 7.68213153e-01 5.53759634e-01 3.06498677e-01
7.25184500e-01 8.40073287e-01 7.17064977e-01 3.54024202e-01
1.31707564e-01 -6.51715621e-02 5.31576991e-01 6.17848039e-01
-3.88564736e-01 -4.76339579e-01 -1.28702390e+00 5.67105711e-01
-2.27333808e+00 -9.42132056e-01 3.24763134e-02 1.80320907e+00
4.08171684e-01 -1.13636963e-01 2.14284565e-02 -3.97519946e-01
5.44783831e-01 -9.72347781e-02 -9.30522919e-01 -2.04020709e-01
8.71040896e-02 -1.44010335e-01 5.22063315e-01 5.29288471e-01
-1.00960338e+00 1.09541118e+00 6.10440969e+00 1.18790366e-01
-9.75647628e-01 6.97705224e-02 -7.34481886e-02 -4.47205424e-01
2.03959957e-01 9.08219293e-02 -2.59519786e-01 1.54377103e-01
7.74016500e-01 -2.33677521e-01 8.93601120e-01 7.10613191e-01
3.09469372e-01 -3.26176971e-01 -9.95886445e-01 3.39493901e-01
-1.28062680e-01 -9.25579906e-01 -6.33573055e-01 1.51900828e-01
5.54079354e-01 5.22875190e-01 1.44335285e-01 4.64431494e-01
7.13076770e-01 -8.37837994e-01 9.17752087e-01 6.78955734e-01
8.53475556e-02 -1.90899223e-01 3.50958794e-01 9.90777075e-01
-7.20233798e-01 -5.15701950e-01 -4.09449011e-01 -7.14683473e-01
5.35031974e-01 -1.32603720e-01 -9.86506879e-01 4.24039602e-01
6.65419161e-01 6.42935753e-01 -3.89633507e-01 1.27459240e+00
-1.50238350e-01 2.28840321e-01 -6.37350440e-01 -6.05757535e-01
1.56480730e-01 -1.82308987e-01 1.22813404e+00 6.46955132e-01
4.75448459e-01 3.00000846e-01 5.21496654e-01 6.50109172e-01
3.39465916e-01 -1.56257480e-01 -9.22937572e-01 -8.62171501e-03
3.26672286e-01 1.14386499e+00 -8.45628381e-01 -7.79973194e-02
5.92441745e-02 9.49236035e-01 6.36008739e-01 4.74884152e-01
-7.55364060e-01 -3.32543850e-01 6.77521169e-01 -1.03782199e-01
3.66593748e-01 -9.69170570e-01 2.43712179e-02 -1.18534350e+00
4.63610850e-02 -7.05219626e-01 -1.76372856e-01 -1.01306808e+00
-1.05119312e+00 3.96865189e-01 9.42434147e-02 -9.72389340e-01
-2.00045183e-01 -4.86402810e-01 -4.93088573e-01 3.87484103e-01
-1.19410193e+00 -9.82614756e-01 -5.07580519e-01 4.28184271e-01
6.21405303e-01 -1.78125978e-01 7.09547520e-01 -2.03035563e-01
-5.85405290e-01 3.81359481e-03 2.37789214e-01 -3.69070232e-01
6.13413870e-01 -1.25425684e+00 4.99816895e-01 6.48362279e-01
-2.86349595e-01 7.77971029e-01 1.06859994e+00 -8.25962365e-01
-1.85726118e+00 -7.07540452e-01 2.39274397e-01 -6.67230606e-01
8.36685717e-01 -4.01256353e-01 -9.51641083e-01 9.41496730e-01
2.53691614e-01 -2.04227399e-02 1.99653551e-01 2.00306833e-01
2.46462107e-01 5.33554912e-01 -9.70327139e-01 8.98092031e-01
1.54956639e+00 2.23238915e-02 -6.65855289e-01 6.05063379e-01
7.66767323e-01 -9.30357754e-01 -6.47418439e-01 3.02348107e-01
5.04024625e-01 -5.77735066e-01 9.04008031e-01 -8.21970165e-01
2.70183504e-01 -4.22381997e-01 -1.14718221e-01 -1.75441217e+00
-6.53858066e-01 -9.68820870e-01 -9.92154405e-02 6.33749723e-01
3.58208954e-01 -1.07310700e+00 3.09151798e-01 5.77453911e-01
-5.58025241e-01 -5.73722601e-01 -1.05604160e+00 -9.65716422e-01
4.13626015e-01 -1.65774345e-01 3.46161216e-01 7.73540854e-01
5.76044798e-01 1.34094670e-01 -2.42879480e-01 5.99448800e-01
2.23550379e-01 -1.31003678e-01 1.06618249e+00 -1.03605938e+00
-5.32383323e-01 -5.86903870e-01 -2.40684301e-01 -1.02958524e+00
4.34335291e-01 -8.76714528e-01 4.89501387e-01 -1.75290978e+00
-1.45003125e-02 -7.73973227e-01 9.84387174e-02 5.68482876e-01
5.31569533e-02 -2.39351541e-01 6.13066077e-01 5.09684741e-01
-8.59466493e-01 7.70908594e-01 1.51575518e+00 9.40799043e-02
-6.39082789e-01 1.01950057e-02 -5.71334839e-01 6.15456700e-01
1.36932600e+00 -3.62458646e-01 -2.84533679e-01 -6.72071755e-01
3.98347229e-01 2.87898958e-01 4.38309133e-01 -1.32288444e+00
7.43200958e-01 -4.59562272e-01 -1.02678291e-01 2.15181321e-01
5.59246659e-01 -6.26623333e-01 2.16126919e-01 7.22706795e-01
-3.90632629e-01 3.22400451e-01 3.75464112e-01 8.86932373e-01
3.88429999e-01 -2.87184268e-01 2.80646086e-01 -4.10776019e-01
-4.79244381e-01 4.46614251e-02 -8.99809480e-01 -1.52335599e-01
1.17097223e+00 -1.00981928e-01 -6.40440583e-01 -4.38379526e-01
-7.72817135e-01 8.27844024e-01 9.43661988e-01 5.46327531e-01
2.50900000e-01 -8.52662563e-01 -3.86499703e-01 -1.41286507e-01
-1.66399702e-01 3.66818547e-01 1.93315417e-01 1.11843705e+00
-6.83316350e-01 2.55989283e-01 -5.38097322e-01 -4.20714349e-01
-6.88357532e-01 4.44727540e-01 6.05088830e-01 9.50907543e-02
-8.63260090e-01 5.54224789e-01 8.23430568e-02 -8.28989267e-01
-6.19385252e-03 -3.65062803e-01 1.70599390e-02 -5.08054197e-01
1.20781772e-01 2.61069745e-01 -2.45102331e-01 -4.59032297e-01
-3.20904762e-01 4.47782129e-01 3.28814417e-01 -2.82028049e-01
1.63624632e+00 -3.47385019e-01 1.76727429e-01 3.11779290e-01
3.80004525e-01 -1.35167152e-01 -1.54802239e+00 -9.14101377e-02
-1.56924844e-01 -2.13038936e-01 -2.09948570e-01 -6.08224034e-01
-5.93109071e-01 5.96326709e-01 8.82429779e-02 1.90929756e-01
7.14862823e-01 1.11654542e-01 5.13804674e-01 1.25896633e+00
6.41634226e-01 -1.09154952e+00 4.77124214e-01 1.11579561e+00
1.17719018e+00 -1.20152175e+00 -1.63311154e-01 2.14613657e-02
-7.46770442e-01 1.07050705e+00 8.28961551e-01 -4.26906496e-01
3.57456148e-01 2.64347672e-01 -2.27069065e-01 -2.37243876e-01
-1.03781450e+00 -1.17231421e-01 -2.05952644e-01 9.71875548e-01
-6.75896183e-02 -6.42205477e-02 -7.61258751e-02 3.46997738e-01
-4.96414602e-01 -2.16027439e-01 1.07915807e+00 1.30082095e+00
-7.49359310e-01 -6.20179772e-01 -1.61866620e-01 -6.99132234e-02
1.35834515e-01 4.96273696e-01 -3.72794420e-01 7.56396413e-01
-1.48193240e-01 7.55122900e-01 6.10267036e-02 -2.30566904e-01
4.13818240e-01 -3.44260409e-02 5.93098581e-01 -7.26144433e-01
-5.34326613e-01 -3.47469538e-01 -4.13107872e-02 -6.05453610e-01
-4.54785615e-01 -1.13231373e+00 -1.43902326e+00 -1.57261521e-01
-1.73431724e-01 -7.60872439e-02 7.09970772e-01 6.67568982e-01
6.87185228e-01 6.41262531e-01 2.50533998e-01 -1.66635537e+00
-6.45646095e-01 -9.07339454e-01 -1.44551709e-01 6.35133535e-02
5.38235724e-01 -1.13802183e+00 -3.74186188e-01 -1.50330395e-01] | [4.474392414093018, 1.1183671951293945] |
e84c2c75-0ca6-48c7-a236-a7f144569a04 | transforming-graphs-for-enhanced-attribute | 2306.11307 | null | https://arxiv.org/abs/2306.11307v2 | https://arxiv.org/pdf/2306.11307v2.pdf | Transforming Graphs for Enhanced Attribute-Based Clustering: An Innovative Graph Transformer Method | Graph Representation Learning (GRL) is an influential methodology, enabling a more profound understanding of graph-structured data and aiding graph clustering, a critical task across various domains. The recent incursion of attention mechanisms, originally an artifact of Natural Language Processing (NLP), into the realm of graph learning has spearheaded a notable shift in research trends. Consequently, Graph Attention Networks (GATs) and Graph Attention Auto-Encoders have emerged as preferred tools for graph clustering tasks. Yet, these methods primarily employ a local attention mechanism, thereby curbing their capacity to apprehend the intricate global dependencies between nodes within graphs. Addressing these impediments, this study introduces an innovative method known as the Graph Transformer Auto-Encoder for Graph Clustering (GTAGC). By melding the Graph Auto-Encoder with the Graph Transformer, GTAGC is adept at capturing global dependencies between nodes. This integration amplifies the graph representation and surmounts the constraints posed by the local attention mechanism. The architecture of GTAGC encompasses graph embedding, integration of the Graph Transformer within the autoencoder structure, and a clustering component. It strategically alternates between graph embedding and clustering, thereby tailoring the Graph Transformer for clustering tasks, whilst preserving the graph's global structural information. Through extensive experimentation on diverse benchmark datasets, GTAGC has exhibited superior performance against existing state-of-the-art graph clustering methodologies. This pioneering approach represents a novel contribution to the field of graph clustering, paving the way for promising avenues in future research. | ['Jiacheng Liu', 'Li Tao', 'Yinan Chen', 'Jiayun Wu', 'Shuo Han'] | 2023-06-20 | null | null | null | null | ['graph-clustering', 'graph-embedding', 'graph-attention', 'graph-learning', 'clustering', 'graph-representation-learning'] | ['graphs', 'graphs', 'graphs', 'graphs', 'methodology', 'methodology'] | [-2.13507041e-02 4.57809687e-01 -2.28757590e-01 -1.45410985e-01
-4.80464138e-02 -4.66966271e-01 6.99431002e-01 5.36499560e-01
-1.80908620e-01 9.58096460e-02 3.48220378e-01 -4.89713430e-01
-1.84988931e-01 -9.43646550e-01 -4.55987692e-01 -8.40878904e-01
-3.30300063e-01 3.72527808e-01 -1.48685679e-01 -6.41435161e-02
2.46410817e-02 4.68369246e-01 -1.24605310e+00 2.07993258e-02
6.77274168e-01 5.20156145e-01 1.20066643e-01 4.50185716e-01
-3.14346313e-01 9.83642578e-01 -4.64277297e-01 -5.49055874e-01
-4.95853880e-03 -4.41912889e-01 -7.46055722e-01 1.82526678e-01
1.46296829e-01 2.35083327e-01 -7.30093122e-01 1.05386400e+00
2.99058944e-01 2.50892520e-01 4.49561179e-01 -1.36311603e+00
-1.24797440e+00 9.06096339e-01 -5.92994809e-01 2.27347091e-01
2.04534739e-01 -8.99892393e-03 1.60106826e+00 -4.43237007e-01
4.83993530e-01 1.17042780e+00 6.81841731e-01 2.91610092e-01
-1.35792232e+00 -3.18215013e-01 3.95029187e-01 2.88886815e-01
-1.54522288e+00 -1.23676728e-03 1.09940732e+00 -5.31863809e-01
1.16922319e+00 7.72821307e-02 7.89423287e-01 9.37073350e-01
3.69610995e-01 7.19210088e-01 6.76992476e-01 -4.62724090e-01
2.79472768e-01 -1.60398379e-01 3.00640404e-01 7.13033855e-01
4.07610744e-01 -5.80355935e-02 -1.28957495e-01 5.12130372e-02
6.54267788e-01 1.88827619e-01 -3.53866011e-01 -5.57719469e-01
-7.46124923e-01 1.05288529e+00 9.05407131e-01 6.50896370e-01
-4.26323980e-01 2.07205415e-01 5.24254203e-01 2.70457774e-01
3.53880197e-01 5.12418866e-01 -6.25045523e-02 1.71057418e-01
-4.52975124e-01 -1.33746848e-01 6.50717497e-01 7.86226332e-01
7.98021615e-01 2.45866656e-01 -6.47725090e-02 5.95458806e-01
4.01784480e-01 9.42269899e-03 4.56271708e-01 -3.01184118e-01
2.16672674e-01 1.35322213e+00 -7.56585479e-01 -1.42261517e+00
-3.63809288e-01 -6.86077893e-01 -1.06646550e+00 -9.45057943e-02
4.80329432e-02 2.12344408e-01 -8.81966531e-01 1.81637716e+00
2.52017915e-01 2.02051684e-01 -2.27628216e-01 7.00457394e-01
6.10577106e-01 4.94332790e-01 1.60545960e-01 1.72013655e-01
1.30065513e+00 -8.70387316e-01 -5.54271996e-01 -2.74731696e-01
5.81499219e-01 -1.84596151e-01 1.08536148e+00 9.69261751e-02
-6.08007431e-01 -5.45641303e-01 -1.08055615e+00 -3.87858450e-02
-6.91682458e-01 -4.00916219e-01 9.56662297e-01 4.52855855e-01
-1.46880901e+00 5.05485475e-01 -9.64453876e-01 -4.88887817e-01
5.18082023e-01 5.74590564e-01 -5.75655401e-01 -1.56888798e-01
-1.04216230e+00 6.52371168e-01 4.51596648e-01 2.45950758e-01
-5.48825979e-01 -2.54332185e-01 -1.17732668e+00 4.56937760e-01
6.12505019e-01 -6.38926446e-01 6.67987525e-01 -9.71524894e-01
-1.15092731e+00 7.15569198e-01 1.68350324e-01 -7.27654755e-01
-1.73477203e-01 1.09411344e-01 -4.81714696e-01 1.83309883e-01
1.78564079e-02 4.42523777e-01 1.02366960e+00 -1.10091937e+00
-1.50731713e-01 -4.70254272e-01 -1.60068534e-02 2.10731238e-01
-5.19189894e-01 -2.16752678e-01 -5.56398034e-01 -8.12558651e-01
-7.53146932e-02 -7.78317392e-01 -3.42123568e-01 -6.35629058e-01
-4.93507326e-01 -4.56115961e-01 9.63727295e-01 -3.86401415e-01
1.56488276e+00 -2.24684191e+00 5.53316295e-01 1.89136595e-01
8.85696292e-01 3.03759545e-01 -1.93022937e-01 8.67935300e-01
-4.10779387e-01 2.53734022e-01 -4.02992040e-01 -2.97674805e-01
1.91022545e-01 3.92586797e-01 -7.61312917e-02 5.76076090e-01
3.13429534e-01 1.37366223e+00 -8.62143099e-01 -1.98865861e-01
3.86286527e-01 7.19606161e-01 -7.09977806e-01 2.67570347e-01
-2.08667353e-01 1.33717924e-01 -2.73928195e-01 4.99382287e-01
3.43370944e-01 -6.25169039e-01 4.53675866e-01 -1.88206166e-01
9.43490192e-02 7.27676600e-02 -8.50288630e-01 1.44685435e+00
-8.90382528e-02 6.12053871e-01 1.45444512e-01 -1.39870620e+00
8.43128204e-01 2.03257963e-01 6.55642509e-01 -5.59676707e-01
4.49370861e-01 -3.53674084e-01 2.72440404e-01 -2.74795383e-01
2.63900548e-01 -1.80050090e-01 -1.14645518e-01 5.69783151e-01
4.49798673e-01 2.49929816e-01 2.46468917e-01 5.63409984e-01
1.32391465e+00 -1.76142767e-01 4.72053468e-01 -2.93051869e-01
5.27578175e-01 -1.95188850e-01 2.40237519e-01 4.76921409e-01
-2.44821623e-01 2.78052241e-01 6.66511834e-01 -5.89992642e-01
-7.91731417e-01 -9.38239098e-01 3.46992373e-01 9.73442316e-01
-8.41865018e-02 -8.27052116e-01 -7.32264519e-01 -8.27699423e-01
1.38280153e-01 2.85756588e-01 -9.37796175e-01 -5.81628621e-01
-5.65975130e-01 -6.14220858e-01 4.03846651e-01 5.90699255e-01
2.29114980e-01 -1.14933419e+00 -2.93205827e-01 2.77219206e-01
1.16900176e-01 -1.16588807e+00 -3.75891030e-01 4.25687551e-01
-6.84362233e-01 -1.30298913e+00 -3.38828027e-01 -9.77112412e-01
7.22383022e-01 3.59780312e-01 1.26380563e+00 4.38945949e-01
-3.18321139e-01 5.59953272e-01 -6.16348922e-01 -1.51035875e-01
-2.32056394e-01 3.26622963e-01 -1.65443406e-01 2.51805335e-01
5.87866008e-01 -8.97907197e-01 -4.08752739e-01 -1.33993670e-01
-1.13805306e+00 -1.00822702e-01 5.77513039e-01 8.36934984e-01
4.40082669e-01 5.43150544e-01 4.30431098e-01 -9.63130891e-01
8.16153169e-01 -6.99265957e-01 -5.95815480e-01 6.31429404e-02
-7.94946074e-01 1.99626386e-01 9.49466586e-01 -1.58219472e-01
-5.58043718e-01 -2.68770337e-01 6.43661153e-03 -6.14665031e-01
-1.12412564e-01 9.15686429e-01 -3.98743480e-01 -4.39404510e-02
2.44915605e-01 1.18565880e-01 1.04227193e-01 -2.99188852e-01
6.25757098e-01 3.67443860e-01 5.52602887e-01 -4.06619132e-01
8.44674706e-01 5.31351030e-01 -1.83756024e-01 -9.31514263e-01
-4.21049714e-01 -5.39369822e-01 -6.66657150e-01 -7.89574012e-02
1.08647466e+00 -6.89073086e-01 -6.24752641e-01 1.70448840e-01
-8.21445942e-01 -4.04923528e-01 -2.58974969e-01 1.70628726e-01
-2.59347141e-01 5.00516653e-01 -6.87811315e-01 -5.87699950e-01
-1.44726723e-01 -1.07083392e+00 9.56365287e-01 6.33295476e-02
-3.06690484e-01 -1.42882323e+00 1.66489258e-01 2.33097479e-01
2.48722672e-01 4.63972837e-01 1.27676356e+00 -8.59320581e-01
-5.35252750e-01 -3.61048341e-01 -2.33308733e-01 2.19599545e-01
2.21794739e-01 -1.44370392e-01 -8.22686076e-01 -3.81593287e-01
-1.89940900e-01 1.37301534e-01 7.98027694e-01 2.70492256e-01
1.06423938e+00 -3.61019850e-01 -3.08566868e-01 5.87842584e-01
1.56576717e+00 2.68269479e-01 4.60343093e-01 2.16821626e-01
1.26540840e+00 5.05627334e-01 -2.87470698e-01 1.78833082e-01
6.94576323e-01 5.37358403e-01 6.73107624e-01 -1.92116871e-01
-1.55352756e-01 -4.03758764e-01 1.89788580e-01 1.13452673e+00
1.10306852e-01 -5.45937300e-01 -1.16946709e+00 5.25848389e-01
-1.84376836e+00 -8.19157958e-01 -6.98640347e-02 1.83999801e+00
3.29882562e-01 7.85406232e-02 6.43348843e-02 3.56687456e-01
8.07852387e-01 4.94290799e-01 -3.30292076e-01 -6.45082593e-01
2.57160310e-02 2.78825074e-01 1.81700110e-01 5.16778111e-01
-1.06652677e+00 1.13779545e+00 5.77047539e+00 3.57029587e-01
-1.00466299e+00 -2.17239648e-01 4.12216336e-01 5.54060876e-01
-3.74443769e-01 6.22738563e-02 -3.54414284e-01 4.54060912e-01
9.87863302e-01 -2.81797707e-01 7.25105703e-01 7.57785141e-01
-8.45904276e-02 2.64476568e-01 -1.13622284e+00 7.59042680e-01
1.99439272e-01 -1.12939918e+00 2.55130887e-01 3.01870048e-01
3.49141836e-01 1.75961077e-01 2.15898268e-02 5.10836065e-01
5.67647934e-01 -1.35436511e+00 2.87741423e-01 9.40930024e-02
4.51144129e-01 -7.37156212e-01 8.68722081e-01 3.62815335e-02
-1.52088928e+00 -2.04430208e-01 -9.52194184e-02 -3.04796726e-01
1.38223525e-02 5.45720041e-01 -7.79867291e-01 8.03700805e-01
5.62576175e-01 1.01745844e+00 -6.31041169e-01 5.82578003e-01
-3.55358005e-01 6.06747389e-01 1.97961763e-01 1.70743108e-01
5.31537533e-01 -4.96441394e-01 6.00350976e-01 1.07762253e+00
-2.53731817e-01 -4.14949954e-02 1.57898501e-01 9.77619171e-01
-2.87112564e-01 -2.06883270e-02 -9.22768176e-01 -7.79396474e-01
3.19537789e-01 1.27811921e+00 -1.03872144e+00 -2.47740537e-01
-5.14547169e-01 8.93509448e-01 6.42396092e-01 4.65692997e-01
-8.37617815e-01 -4.88719285e-01 5.47292948e-01 1.15691289e-01
4.24681216e-01 -4.34407055e-01 -2.42037386e-01 -1.00021434e+00
-9.84205157e-02 -9.70458984e-01 6.66227579e-01 -4.70362782e-01
-1.25727654e+00 9.08389628e-01 -1.83281064e-01 -6.83161378e-01
2.13796310e-02 -5.58592081e-01 -6.39186978e-01 5.18200755e-01
-1.28212774e+00 -1.20048344e+00 -3.28140020e-01 6.97150707e-01
3.00988615e-01 -7.45331272e-02 8.08637202e-01 3.93667787e-01
-7.94117570e-01 4.26710159e-01 -1.55704051e-01 3.24084342e-01
2.05547616e-01 -1.49817610e+00 5.21839261e-01 9.79307353e-01
5.72797596e-01 8.83679509e-01 3.96119654e-01 -5.38397789e-01
-1.62991500e+00 -1.23090374e+00 7.91295052e-01 -4.58925098e-01
9.70891178e-01 -7.42400467e-01 -1.11699975e+00 9.18536007e-01
3.81387949e-01 -9.65711777e-04 7.52087533e-01 2.54301608e-01
-5.10749817e-01 -4.07152437e-02 -6.55272782e-01 7.08456397e-01
9.12958801e-01 -8.83767724e-01 -4.86734927e-01 2.01943889e-02
9.65369821e-01 -3.56815234e-02 -8.86625171e-01 1.87885508e-01
2.14215472e-01 -9.65853631e-01 8.76836181e-01 -5.25222361e-01
3.60269517e-01 -2.58752048e-01 -8.48629922e-02 -1.33945560e+00
-6.78622663e-01 -7.57105172e-01 -3.77773017e-01 1.18655062e+00
7.86098093e-02 -6.87777817e-01 7.62235343e-01 3.02050829e-01
-4.03040409e-01 -8.61961424e-01 -6.01760209e-01 -5.57555079e-01
-9.80074778e-02 -2.87321985e-01 6.75351262e-01 1.26377654e+00
1.41128242e-01 6.91028476e-01 1.13325305e-02 1.44598171e-01
5.46921790e-01 3.31950225e-02 6.56998336e-01 -1.31484902e+00
-3.96364450e-01 -8.46310735e-01 -7.80926824e-01 -6.33909106e-01
4.16772157e-01 -1.36795533e+00 -2.44442359e-01 -1.68161190e+00
5.89095466e-02 -2.72536688e-02 -4.16287363e-01 6.22305632e-01
-2.34779865e-01 1.03941612e-01 2.08326072e-01 -1.26509726e-01
-5.57385325e-01 6.44138217e-01 9.47001159e-01 -3.05261046e-01
-2.15144515e-01 -4.15418535e-01 -1.19424045e+00 5.02326429e-01
9.20736730e-01 -3.69634449e-01 -6.93490684e-01 -4.84265804e-01
1.99264064e-01 -3.42744797e-01 3.12401772e-01 -9.32865381e-01
4.06570911e-01 1.89915910e-01 -1.23857260e-02 -3.28290820e-01
-5.57078943e-02 -1.00455976e+00 1.40017822e-01 3.44414592e-01
-3.84990871e-02 3.70920777e-01 1.62498370e-01 9.13791597e-01
-4.80352193e-01 4.59338188e-01 6.84271812e-01 -1.46096155e-01
-8.21850598e-01 4.93971527e-01 -3.52635205e-01 8.80321115e-02
1.01800776e+00 -4.39244479e-01 -6.64631501e-02 -2.49853730e-01
-7.99708962e-01 2.74687439e-01 4.80995983e-01 5.44927657e-01
4.88102853e-01 -1.17774570e+00 -3.32068086e-01 5.72783351e-01
7.86803290e-02 4.46670912e-02 1.39225140e-01 8.41314971e-01
-2.68931448e-01 5.00943840e-01 -1.63492896e-02 -4.83907253e-01
-9.79778945e-01 9.10783291e-01 2.07455739e-01 -4.43657875e-01
-9.39861774e-01 8.29710841e-01 4.54230219e-01 -3.84817034e-01
2.74466872e-01 -3.62229407e-01 -2.80724645e-01 2.71622166e-02
1.62231684e-01 2.65801817e-01 1.33070901e-01 -7.05250084e-01
-3.79407644e-01 3.72053653e-01 -1.27966896e-01 3.97395045e-01
1.53042150e+00 -1.62021413e-01 -3.99688512e-01 2.92721510e-01
1.15470910e+00 -8.68677720e-02 -1.06866443e+00 -4.01863903e-02
3.52735668e-01 -3.60214487e-02 1.18762545e-01 -4.95761931e-01
-1.17093992e+00 9.43363249e-01 -5.84493317e-02 7.01786637e-01
1.09644759e+00 1.78410783e-01 6.42971635e-01 3.53716724e-02
2.25020587e-01 -7.40271211e-01 2.51231372e-01 5.19250572e-01
7.21306086e-01 -1.11747444e+00 -1.21582881e-01 -2.66646117e-01
-5.68108916e-01 8.95516753e-01 5.65288305e-01 -4.31271762e-01
6.36470735e-01 3.04176390e-01 -2.20113769e-01 -7.40345657e-01
-6.53684258e-01 -2.64546275e-01 3.41116488e-01 6.70710802e-01
4.74345326e-01 1.68843701e-01 7.63317049e-02 6.67805016e-01
-7.92338774e-02 -5.57870030e-01 2.43056595e-01 6.50827587e-01
-2.67996073e-01 -1.15775752e+00 -3.21535878e-02 4.09198284e-01
-3.20901722e-01 -2.88423508e-01 -8.54712605e-01 1.11488748e+00
-1.21785887e-02 8.13383579e-01 1.18556313e-01 -5.25829256e-01
9.61041376e-02 -2.68199760e-02 1.59528598e-01 -6.85972869e-01
-6.79244101e-01 -4.61652987e-02 -2.73026198e-01 -6.54851437e-01
-2.57047921e-01 -2.02755705e-01 -1.20077264e+00 -3.32045078e-01
-3.68860751e-01 5.20091832e-01 2.98169762e-01 6.54883504e-01
7.20817029e-01 8.86312723e-01 4.55110252e-01 -6.78452373e-01
-7.02276975e-02 -6.57262921e-01 -6.86790943e-01 3.37312132e-01
2.68317968e-01 -5.98455548e-01 -3.43509972e-01 -1.23814806e-01] | [7.069454669952393, 6.260171890258789] |
d2fbed87-ee64-4f6b-b059-0660e06c37cb | towards-an-embodied-semantic-fovea-semantic | 1807.10561 | null | http://arxiv.org/abs/1807.10561v1 | http://arxiv.org/pdf/1807.10561v1.pdf | Towards an Embodied Semantic Fovea: Semantic 3D scene reconstruction from ego-centric eye-tracker videos | Incorporating the physical environment is essential for a complete
understanding of human behavior in unconstrained every-day tasks. This is
especially important in ego-centric tasks where obtaining 3 dimensional
information is both limiting and challenging with the current 2D video analysis
methods proving insufficient. Here we demonstrate a proof-of-concept system
which provides real-time 3D mapping and semantic labeling of the local
environment from an ego-centric RGB-D video-stream with 3D gaze point
estimation from head mounted eye tracking glasses. We augment existing work in
Semantic Simultaneous Localization And Mapping (Semantic SLAM) with collected
gaze vectors. Our system can then find and track objects both inside and
outside the user field-of-view in 3D from multiple perspectives with reasonable
accuracy. We validate our concept by producing a semantic map from images of
the NYUv2 dataset while simultaneously estimating gaze position and gaze
classes from recorded gaze data of the dataset images. | ['A. Aldo Faisal', 'Pavel Orlov', 'Stefan Leutenegger', 'Noyan Songur', 'Mickey Li'] | 2018-07-27 | null | null | null | null | ['3d-scene-reconstruction', 'semantic-slam'] | ['computer-vision', 'computer-vision'] | [-1.84113551e-02 -1.51535317e-01 8.02311450e-02 -6.57790005e-01
-3.32548440e-01 -7.32326031e-01 1.97651237e-01 -2.13842452e-01
-4.25034076e-01 2.97944605e-01 1.31296158e-01 -2.69973092e-02
-7.85732344e-02 4.09674123e-02 -7.78700233e-01 -2.86848098e-01
1.37581781e-01 4.25838232e-01 2.78572172e-01 -1.87914401e-01
5.71197510e-01 6.27941430e-01 -2.26262093e+00 -1.45602629e-01
2.41608739e-01 1.00714099e+00 4.85058129e-01 1.01329303e+00
2.78036028e-01 5.63751936e-01 -1.07918613e-01 2.80993253e-01
2.87902713e-01 -2.12907404e-01 -5.90136588e-01 4.48390007e-01
1.18688810e+00 -6.51142716e-01 -9.02509466e-02 9.96870637e-01
4.61379558e-01 1.81878075e-01 3.88558954e-02 -1.72940457e+00
-8.63736570e-02 -7.18437791e-01 -6.71088099e-01 2.83561349e-01
1.34907544e+00 -2.15564109e-02 5.48066974e-01 -7.11060822e-01
8.64003003e-01 1.06498194e+00 4.36404169e-01 6.70611799e-01
-1.03652203e+00 -5.84095776e-01 2.10281685e-01 2.89910376e-01
-1.58623147e+00 -8.65944505e-01 5.89295447e-01 -4.79147434e-01
1.10045922e+00 1.41349509e-01 6.73187435e-01 9.29793715e-01
1.74618922e-02 5.27861297e-01 1.21365762e+00 -4.60112512e-01
1.95359603e-01 3.36974412e-01 5.07244691e-02 9.05741811e-01
2.01052800e-01 -8.53191242e-02 -1.39626896e+00 2.87622541e-01
6.86763167e-01 2.76326418e-01 -4.54265505e-01 -1.01226425e+00
-1.19668937e+00 1.10264830e-01 5.12714267e-01 -6.85672760e-02
-3.26625496e-01 1.53024256e-01 -3.32762152e-01 2.30120301e-01
7.63681471e-01 1.20326377e-01 -4.78233337e-01 -5.81896603e-01
-9.59566653e-01 3.54365975e-01 4.18600649e-01 1.56001592e+00
9.13240016e-01 -5.73348582e-01 5.24033248e-01 -4.40834351e-02
7.25385904e-01 9.29891169e-01 2.44883284e-01 -1.19557309e+00
1.81388482e-01 8.33443761e-01 5.31373799e-01 -7.49979675e-01
-7.53914952e-01 1.87099665e-01 4.77141589e-01 5.65343738e-01
5.53255737e-01 4.57823724e-02 -7.28441060e-01 1.62835598e+00
9.48348105e-01 1.78748339e-01 -4.19786870e-01 1.28997874e+00
5.21862507e-01 -1.93824828e-01 -2.41006702e-01 -6.75204322e-02
1.56455982e+00 -5.06184578e-01 -9.36167181e-01 -3.97762924e-01
8.03643823e-01 -5.88689685e-01 1.06870806e+00 3.84184241e-01
-8.79534185e-01 -3.19520056e-01 -8.95187378e-01 -4.65066880e-01
-4.01916087e-01 -5.63726295e-03 4.16493088e-01 6.92962289e-01
-1.50676501e+00 -6.74966052e-02 -1.29870403e+00 -1.05514276e+00
2.66222686e-01 6.53420329e-01 -9.34681416e-01 -1.93096101e-02
-4.75406408e-01 1.03588092e+00 -3.87408435e-02 2.67701410e-02
-6.40413284e-01 -5.27216673e-01 -1.11093211e+00 -2.96359718e-01
5.08340061e-01 -4.56879735e-01 1.38050532e+00 -5.52542627e-01
-1.34817815e+00 1.26517296e+00 -9.08294320e-01 2.16013156e-02
1.74055368e-01 -6.44611955e-01 -2.36811176e-01 2.39324510e-01
1.55576751e-01 6.60580277e-01 7.61635244e-01 -1.06929886e+00
-8.32630277e-01 -1.27611184e+00 1.14521705e-01 7.86899269e-01
-1.86444432e-01 1.86811060e-01 -4.09694105e-01 4.28520709e-01
6.08013630e-01 -1.25901449e+00 3.89562219e-01 5.00048995e-01
-1.88505352e-01 8.27203542e-02 1.23980141e+00 -4.08038318e-01
7.60972977e-01 -1.83140516e+00 1.42877236e-01 -9.56418179e-03
5.23284793e-01 -2.50955462e-01 4.79426503e-01 -1.13360202e-02
1.30314633e-01 -4.07352805e-01 5.15589952e-01 -1.00174201e+00
-7.90873244e-02 -1.03869982e-01 -3.33930925e-02 9.43282247e-01
-3.46204698e-01 7.90785909e-01 -1.12563229e+00 -4.60197777e-01
5.90579629e-01 6.64782941e-01 -5.57581484e-01 3.80309969e-01
-1.02531034e-02 6.60416305e-01 -3.53388071e-01 1.00372446e+00
5.05577385e-01 -4.08218384e-01 -1.03012949e-01 7.27798883e-03
-3.69393945e-01 3.07230383e-01 -1.05121732e+00 2.53182530e+00
-2.13517368e-01 1.17342794e+00 1.16447657e-01 -1.97545663e-01
3.92307162e-01 2.03290507e-01 3.96709800e-01 -6.65677130e-01
3.83452624e-01 -1.07955281e-02 -5.33922851e-01 -8.46634984e-01
9.00324166e-01 1.70829117e-01 1.49416402e-01 7.05837786e-01
1.38801321e-01 2.28453893e-03 -3.74331981e-01 1.17333956e-01
9.91551936e-01 8.22545230e-01 2.17443481e-01 -2.29619607e-01
2.42771640e-01 2.33045146e-01 -1.04331724e-01 2.61844426e-01
-5.65879881e-01 7.02529490e-01 2.55243462e-02 -2.90074140e-01
-8.95984113e-01 -9.77488518e-01 -5.18433452e-02 1.12390053e+00
5.89066625e-01 -4.96150076e-01 -9.80143964e-01 -5.84791780e-01
-1.61943540e-01 5.31024814e-01 -8.23657453e-01 2.35208943e-01
-1.35060027e-01 8.12619179e-03 -4.98519093e-02 2.21731454e-01
2.58020103e-01 -5.97416759e-01 -1.44334531e+00 -5.57314813e-01
-1.67797536e-01 -1.34962034e+00 -4.58610415e-01 -2.34857976e-01
-5.58954656e-01 -1.44792485e+00 -4.34620559e-01 -3.93305063e-01
1.07108843e+00 9.99977350e-01 7.22856522e-01 -2.78862536e-01
-2.35326007e-01 9.97313142e-01 -1.76375240e-01 -5.32170177e-01
2.47606099e-01 -2.69453645e-01 7.19935179e-01 3.64801660e-02
1.19715679e+00 -2.53356904e-01 -6.78172946e-01 5.57548702e-01
-1.30215466e-01 1.07868403e-01 -2.45755613e-01 -3.00986171e-02
4.63184237e-01 -4.27776009e-01 -3.69336277e-01 -4.74130571e-01
2.69781888e-01 -3.69809628e-01 -9.47527647e-01 4.18350212e-02
-5.44479132e-01 -3.52271855e-01 -3.84478033e-01 -1.03380293e-01
-9.03968215e-01 3.81766707e-01 4.16774213e-01 -9.11666930e-01
-6.13259196e-01 -1.65207684e-01 -8.54780003e-02 -4.19289470e-01
8.88045013e-01 -1.02833740e-01 3.68055940e-01 -3.21143210e-01
2.60670543e-01 1.02658558e+00 4.80494022e-01 -7.75579065e-02
3.99460614e-01 1.17079604e+00 1.90473333e-01 -8.57640266e-01
-9.95977104e-01 -8.84378076e-01 -1.39452612e+00 -5.82301736e-01
1.00088596e+00 -1.09892082e+00 -1.40500498e+00 3.39678973e-01
-9.64578450e-01 -1.39020830e-01 1.04737125e-01 7.18718588e-01
-8.98865223e-01 -1.01038843e-01 2.36442074e-01 -1.01694012e+00
1.83989063e-01 -1.32110524e+00 1.85652435e+00 4.71872419e-01
-5.38170993e-01 -8.68905008e-01 2.97354367e-02 4.55624163e-01
4.71994132e-02 2.56143361e-01 -1.85095638e-01 -4.42259349e-02
-7.05751181e-01 -3.44007224e-01 -6.95340335e-02 -2.85865873e-01
1.79747775e-01 -4.78873312e-01 -1.44726992e+00 -6.94562718e-02
1.27485052e-01 -1.79670423e-01 1.57986279e-03 5.06189227e-01
4.32236016e-01 3.12993616e-01 -5.95605969e-01 8.21045756e-01
1.26629341e+00 -5.24152853e-02 1.54318333e-01 5.36921740e-01
1.05758369e+00 9.21799839e-01 8.65433753e-01 2.74101913e-01
8.66881430e-01 1.08273089e+00 5.75364292e-01 1.70980468e-01
-1.32853195e-01 -2.86229074e-01 2.89615959e-01 1.15955986e-01
-6.25988320e-02 -1.94073990e-01 -1.08898520e+00 4.66228038e-01
-1.83637130e+00 -7.71152675e-01 -2.06507549e-01 2.42684960e+00
3.07695776e-01 -2.35946119e-01 3.06548774e-01 -4.15305980e-02
5.32578647e-01 -2.97322601e-01 -7.53558576e-01 -3.23644206e-02
3.41651499e-01 -3.85303617e-01 8.22904050e-01 6.43491149e-01
-9.15286481e-01 1.00799811e+00 6.15423012e+00 -2.74646640e-01
-1.01180089e+00 2.20297411e-01 -2.74678051e-01 -9.15119052e-01
1.07508585e-01 -1.02147818e-01 -1.11298335e+00 4.29779738e-01
1.02897692e+00 2.23136023e-01 5.24298251e-01 7.57877648e-01
3.69820178e-01 -7.51035154e-01 -1.40324700e+00 1.46229076e+00
6.12911701e-01 -8.06776404e-01 -8.97133946e-01 5.70792258e-01
3.16063613e-01 4.96211141e-01 4.17995453e-02 -4.26007837e-01
-1.17918678e-01 -7.12283254e-01 8.67946684e-01 6.53859377e-01
1.11652780e+00 -3.44870031e-01 9.16606709e-02 5.66065550e-01
-7.77008653e-01 4.43286262e-02 8.61129239e-02 -3.44983101e-01
4.15123343e-01 -2.13070691e-01 -1.20448899e+00 -1.48474738e-01
1.12069869e+00 1.02642691e+00 -7.26644635e-01 8.46705914e-01
-1.93447530e-01 -1.06057145e-01 -6.71096563e-01 7.50961453e-02
-8.03663880e-02 -1.62368059e-01 7.50015736e-01 3.20613503e-01
5.26782990e-01 2.31265277e-01 -2.45939970e-01 7.07894742e-01
1.37995988e-01 -5.57214022e-01 -9.88686025e-01 4.62214559e-01
5.01096070e-01 9.45311487e-01 -8.11423838e-01 1.45328715e-02
-2.40608826e-01 1.08907020e+00 1.27914667e-01 4.45807338e-01
-4.96438503e-01 -3.03044766e-01 1.30948448e+00 4.69154894e-01
1.87276509e-02 -7.58797109e-01 -1.14161380e-01 -1.32239068e+00
1.10733896e-01 -2.20971480e-01 -9.21868384e-02 -1.60023355e+00
-3.28056484e-01 4.70630765e-01 2.44299710e-01 -1.56836009e+00
-5.26120663e-01 -7.33443856e-01 2.46302322e-01 9.52456832e-01
-1.56919265e+00 -1.22268534e+00 -1.03391695e+00 9.81928825e-01
3.76362860e-01 -9.48384684e-03 8.51598084e-01 7.12992921e-02
-2.60189831e-01 3.10081273e-01 -1.00591451e-01 -4.39971060e-01
9.69439447e-01 -1.18922174e+00 4.52431500e-01 7.73515522e-01
2.99795598e-01 8.99869859e-01 8.28247547e-01 -6.18083358e-01
-1.93773031e+00 -4.05993253e-01 9.87511337e-01 -1.68400300e+00
3.93905133e-01 -9.44956064e-01 -4.69960868e-01 1.07357538e+00
-7.06239119e-02 1.91437513e-01 5.63069761e-01 2.98536897e-01
-2.62116641e-01 1.29379228e-01 -1.29558682e+00 4.56259072e-01
1.53918457e+00 -9.92667139e-01 -4.77340132e-01 4.29563016e-01
6.48714721e-01 -1.16380048e+00 -4.48807180e-01 -1.39642820e-01
9.68665063e-01 -1.04601669e+00 9.50693786e-01 -3.74670953e-01
-2.92684853e-01 -7.25259900e-01 -3.54065031e-01 -1.00463033e+00
2.55939454e-01 -6.92248762e-01 -3.36816609e-01 6.02338910e-01
-2.27981448e-01 -5.12067974e-01 1.03884804e+00 1.07809520e+00
2.15824574e-01 -3.53667438e-02 -7.88011968e-01 -3.73316556e-01
-1.09111166e+00 -7.94778824e-01 6.56577408e-01 5.95997572e-01
1.97005808e-01 7.86448792e-02 -1.75927371e-01 5.01351058e-01
9.31577921e-01 -8.55186954e-02 1.18819404e+00 -1.38149357e+00
4.87288773e-01 6.30824491e-02 -1.06533468e+00 -1.08396840e+00
3.09190035e-01 -3.97900254e-01 5.90480380e-02 -1.20645022e+00
-1.90832689e-01 -2.15439335e-01 1.61856022e-02 3.28166336e-01
2.97711819e-01 6.09477282e-01 2.63209944e-03 5.62194407e-01
-9.52713668e-01 1.80460200e-01 9.21651840e-01 6.20262980e-01
-1.85092047e-01 -1.33650437e-01 -4.38653201e-01 6.87936008e-01
3.10136229e-01 -4.85164285e-01 -7.50494838e-01 -7.48827279e-01
5.42780101e-01 -1.06203265e-01 6.29731894e-01 -9.75466967e-01
6.78938091e-01 -5.00023849e-02 4.56080556e-01 -8.93221617e-01
8.64665508e-01 -1.19361532e+00 -7.18756467e-02 -8.57093185e-02
-1.86439678e-01 1.99733317e-01 2.12805837e-01 6.69236064e-01
1.94171965e-01 2.39162028e-01 1.67257696e-01 -8.79680440e-02
-1.17821860e+00 3.23030084e-01 1.26880929e-01 -1.99109390e-01
1.26093984e+00 -7.98467517e-01 -3.35867703e-01 -4.82761919e-01
-7.86552191e-01 2.72148669e-01 1.39528537e+00 6.94578946e-01
6.77917004e-01 -1.15300977e+00 -5.00053279e-02 9.42547023e-01
5.03440559e-01 2.15599999e-01 1.55803591e-01 1.00369430e+00
-6.73703730e-01 8.06666970e-01 -3.63368034e-01 -1.32605147e+00
-1.73537874e+00 4.31234539e-01 3.53168845e-01 1.02113545e+00
-5.96679866e-01 8.64457965e-01 2.31390059e-01 -3.48949790e-01
2.68935174e-01 -1.78698465e-01 -1.90452665e-01 -7.24981129e-02
8.69547606e-01 3.86105269e-01 2.85695225e-01 -1.32571554e+00
-7.41809964e-01 9.47061658e-01 3.16687077e-01 -4.43815172e-01
1.09080505e+00 -1.20349050e+00 -1.21055273e-02 5.75000405e-01
1.30266690e+00 -1.84317470e-01 -1.53693676e+00 -3.56663615e-01
-1.47954777e-01 -1.05899119e+00 5.11935890e-01 -3.99358511e-01
-3.95785987e-01 1.06307232e+00 9.31787133e-01 -4.62075137e-02
1.08736968e+00 2.15512231e-01 3.64734977e-01 4.27586943e-01
9.36949670e-01 -9.06970918e-01 -3.84496093e-01 2.13881209e-01
3.55576456e-01 -1.65214670e+00 2.35007867e-01 -1.18226923e-01
-6.08845055e-01 7.24486828e-01 6.35337830e-01 1.75120234e-01
8.02583039e-01 4.52595837e-02 2.80686408e-01 -7.74968803e-01
-4.51983243e-01 -4.78278488e-01 4.67885375e-01 8.19961905e-01
1.58698231e-01 -1.42735928e-01 6.99229658e-01 -2.54433125e-01
-4.07476723e-01 3.75323184e-02 5.98508954e-01 1.22885203e+00
-3.41601640e-01 -6.45856977e-01 -5.28929412e-01 -1.63277648e-02
-1.58187911e-01 2.11547107e-01 -2.14756817e-01 8.34825099e-01
-8.16500038e-02 1.10388362e+00 5.06751657e-01 -3.13994646e-01
3.25808585e-01 2.49132276e-01 1.02114999e+00 -8.16415131e-01
8.25118925e-03 5.52961864e-02 -2.37839118e-01 -1.41677296e+00
-7.67563879e-01 -1.08104479e+00 -1.03212845e+00 -1.95639625e-01
-2.85619050e-01 -3.98895830e-01 1.58269012e+00 8.73933792e-01
8.30012321e-01 8.93811360e-02 3.89778972e-01 -1.48803937e+00
4.16060351e-02 -6.48168027e-01 -7.94688702e-01 2.31487438e-01
1.07779145e+00 -1.11034334e+00 -2.67550856e-01 3.95232260e-01] | [14.105823516845703, 0.08291333168745041] |
054cf603-e8fb-431c-949f-1130e235d71a | sfi-swin-symmetric-face-inpainting-with-swin | 2301.0313 | null | https://arxiv.org/abs/2301.03130v1 | https://arxiv.org/pdf/2301.03130v1.pdf | SFI-Swin: Symmetric Face Inpainting with Swin Transformer by Distinctly Learning Face Components Distributions | Image inpainting consists of filling holes or missing parts of an image. Inpainting face images with symmetric characteristics is more challenging than inpainting a natural scene. None of the powerful existing models can fill out the missing parts of an image while considering the symmetry and homogeneity of the picture. Moreover, the metrics that assess a repaired face image quality cannot measure the preservation of symmetry between the rebuilt and existing parts of a face. In this paper, we intend to solve the symmetry problem in the face inpainting task by using multiple discriminators that check each face organ's reality separately and a transformer-based network. We also propose "symmetry concentration score" as a new metric for measuring the symmetry of a repaired face image. The quantitative and qualitative results show the superiority of our proposed method compared to some of the recently proposed algorithms in terms of the reality, symmetry, and homogeneity of the inpainted parts. | ['Shadrokh Samavi', 'Shahram Shirani', 'Nader Karimi', 'MohammadHossein Givkashi', 'Mohammadreza Naderi'] | 2023-01-09 | null | null | null | null | ['face-image-quality', 'facial-inpainting', 'image-inpainting'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 8.27988535e-02 3.97053838e-01 1.13685600e-01 -3.30505013e-01
-3.19785595e-01 -2.01272935e-01 2.20457613e-01 -2.50074297e-01
-3.21418494e-02 7.70031631e-01 3.05652648e-01 4.48462784e-01
-2.33414307e-01 -7.47761428e-01 -5.17654181e-01 -5.96917748e-01
4.34907883e-01 2.89265215e-01 -1.41354531e-01 -2.01592654e-01
1.87616035e-01 8.66256535e-01 -1.61630809e+00 3.13584238e-01
7.96267569e-01 1.15647590e+00 9.93848816e-02 1.78181320e-01
-9.73190069e-02 5.81261873e-01 -4.75335598e-01 -6.63658917e-01
5.14561236e-01 -7.72117078e-01 -7.65048265e-01 3.17303509e-01
6.28254950e-01 -3.61213028e-01 -1.36273578e-01 1.20997393e+00
3.13659638e-01 -2.37487331e-01 4.20262635e-01 -1.40937030e+00
-5.36971450e-01 2.33754039e-01 -7.45990694e-01 -4.66305809e-03
5.60200155e-01 -1.09302990e-01 4.89010155e-01 -7.98312008e-01
8.36948633e-01 1.21856117e+00 6.82100773e-01 6.13608360e-01
-1.06443238e+00 -5.72463155e-01 -2.62539595e-01 2.14050785e-01
-1.36432374e+00 -7.17471838e-01 1.24716520e+00 -2.49313250e-01
1.82481617e-01 3.95462900e-01 5.03743112e-01 5.14334679e-01
3.02337080e-01 1.93836719e-01 1.21494567e+00 -3.03303152e-01
5.74084669e-02 2.79046655e-01 -5.54530442e-01 8.09499025e-01
8.19311738e-02 1.82211306e-02 -5.56916595e-01 4.46651317e-02
9.44007039e-01 2.31879562e-01 -4.31530684e-01 -4.14889425e-01
-7.59811640e-01 5.60754597e-01 3.57182682e-01 4.98551935e-01
-4.90636766e-01 -3.10404569e-01 2.02277631e-01 3.86553884e-01
6.76925838e-01 3.58405292e-01 -9.42542553e-02 3.08299750e-01
-1.50978124e+00 -3.55115719e-02 4.06630069e-01 5.63930392e-01
7.80075490e-01 4.34657559e-02 -1.58711866e-01 7.27944672e-01
1.60738844e-02 -2.75612790e-02 2.97452420e-01 -1.26694202e+00
1.38855278e-01 8.37709665e-01 -1.28967240e-01 -1.30780458e+00
-1.00882851e-01 -3.50379914e-01 -1.10033131e+00 5.96350670e-01
2.23044693e-01 1.64878964e-01 -5.00846386e-01 1.90963376e+00
4.45567995e-01 1.18435949e-01 -1.33022845e-01 7.56365955e-01
8.96173656e-01 4.75138783e-01 -2.94199198e-01 -7.10476637e-01
1.10905755e+00 -1.01491308e+00 -1.00917363e+00 -4.99433875e-02
-4.85042445e-02 -1.04443204e+00 7.94354379e-01 3.69567126e-01
-1.70669436e+00 -7.18870282e-01 -8.98399234e-01 -1.85887992e-01
1.55599505e-01 1.95055604e-01 3.66799533e-01 6.08400941e-01
-1.34807742e+00 6.08277261e-01 -2.79794037e-01 -2.16396987e-01
4.96783733e-01 2.82722354e-01 -9.97020721e-01 -2.37635612e-01
-6.17192745e-01 8.47899973e-01 6.92561194e-02 2.22938046e-01
-7.19665706e-01 -7.72912681e-01 -6.99041247e-01 3.70842993e-01
8.70662853e-02 -6.12271667e-01 6.96023107e-01 -1.29601622e+00
-1.21145749e+00 1.00582981e+00 -1.38134569e-01 7.09764957e-02
9.57274795e-01 3.28941882e-01 -3.05846214e-01 4.19445455e-01
1.43345729e-01 7.40262628e-01 8.84764314e-01 -1.40646112e+00
-3.11194837e-01 -7.60173142e-01 -1.04572691e-01 1.66147113e-01
-2.60133386e-01 3.58908549e-02 -3.97458166e-01 -6.88648224e-01
4.36198175e-01 -3.03308547e-01 1.83220237e-01 7.31775701e-01
-1.26526147e-01 2.15605959e-01 1.14499354e+00 -1.20656335e+00
8.68978381e-01 -2.29804039e+00 1.68402344e-01 1.85692266e-01
2.48483747e-01 1.20723687e-01 -3.61160636e-01 3.56973410e-01
-2.06035063e-01 -1.69617869e-02 -4.25971448e-01 -7.23015904e-01
-4.29401577e-01 2.74471641e-01 -4.42925021e-02 6.82051957e-01
2.14850485e-01 5.49649596e-01 -5.12395442e-01 -7.71406233e-01
7.77199715e-02 8.63380790e-01 -5.64348578e-01 3.03785622e-01
1.18895486e-01 6.91031396e-01 8.10349882e-02 8.04984868e-01
1.27771223e+00 2.26032466e-01 1.34774253e-01 -2.82300949e-01
2.38928087e-02 -2.78383672e-01 -1.17069983e+00 1.73210621e+00
-3.77212107e-01 3.86091918e-01 4.94839072e-01 -8.13667655e-01
9.52490747e-01 4.72021490e-01 6.15558147e-01 -9.24098611e-01
1.54359415e-01 1.60467163e-01 -2.18678504e-01 -5.30302465e-01
3.35057169e-01 -3.67925704e-01 4.84002501e-01 5.12817085e-01
1.57975525e-01 -7.30879232e-02 1.37632489e-01 -1.50571344e-03
7.01733112e-01 -1.02009602e-01 1.45266145e-01 -9.88958105e-02
6.02886081e-01 -7.81005561e-01 6.09654725e-01 2.24589452e-01
-2.22783461e-01 1.21092522e+00 8.11679304e-01 -4.14113104e-01
-1.13782573e+00 -9.26157892e-01 -8.12003314e-02 4.13106531e-01
8.09554160e-02 -1.33988127e-01 -9.45420206e-01 -5.29443085e-01
-1.78459063e-01 1.93389341e-01 -8.48552048e-01 -2.45788619e-01
-4.64895457e-01 -3.44895869e-01 3.71904641e-01 2.35701412e-01
6.90182626e-01 -1.14538491e+00 -4.05577511e-01 -2.57241219e-01
-7.16894448e-01 -1.01129377e+00 -5.22786796e-01 -5.66757143e-01
-1.03637064e+00 -1.20978189e+00 -8.41358840e-01 -1.03460836e+00
1.15522134e+00 3.40119332e-01 8.68661821e-01 5.57899415e-01
-3.25840473e-01 -1.35002494e-01 -1.73886284e-01 1.54669493e-01
-6.03780925e-01 -4.86851752e-01 -8.78022760e-02 3.03644300e-01
-6.31361127e-01 -8.09912384e-01 -6.72624588e-01 3.46061587e-01
-1.17073631e+00 9.16788429e-02 5.45795977e-01 7.12705553e-01
4.62002337e-01 4.86766636e-01 2.17186645e-01 -6.11815572e-01
6.11193717e-01 -2.88816690e-01 -1.75613329e-01 3.91714483e-01
-6.68555856e-01 -8.38216469e-02 5.17401159e-01 -1.83859885e-01
-1.19902599e+00 -2.53815576e-02 -1.53688833e-01 -5.71021855e-01
1.75821111e-01 1.19755648e-01 -3.78617227e-01 -3.98035258e-01
2.81821460e-01 3.62296522e-01 4.90740567e-01 -4.59915400e-01
8.01746547e-02 1.83412313e-01 6.41858399e-01 -3.25016320e-01
6.23775482e-01 8.03584278e-01 2.35952973e-01 -4.55338866e-01
-3.57879996e-01 -2.46964753e-01 -5.15912771e-01 -4.29551870e-01
5.53348124e-01 -5.72064459e-01 -8.05645406e-01 4.16496933e-01
-1.33166766e+00 4.12229270e-01 -5.21389782e-01 7.11786523e-02
-4.86883879e-01 9.24312234e-01 -5.63161492e-01 -7.71257639e-01
-3.85699093e-01 -1.03059912e+00 1.03911209e+00 3.66875321e-01
1.58270314e-01 -5.78717649e-01 -2.82930695e-02 5.17112970e-01
4.04462844e-01 3.21377873e-01 8.37072551e-01 1.07061252e-01
-2.58279353e-01 -3.19556773e-01 -3.38393837e-01 6.45204663e-01
3.39371145e-01 4.47076596e-02 -9.44132447e-01 -2.77705848e-01
6.42407119e-01 -9.45191681e-02 6.59845173e-01 4.09210056e-01
9.67350900e-01 -6.27092719e-01 1.03663847e-01 6.72681928e-01
1.38867414e+00 2.11615011e-01 1.26145482e+00 -1.40379936e-01
2.05589131e-01 1.06466031e+00 3.76031041e-01 4.65036273e-01
-8.38671327e-02 4.88013774e-01 6.38789356e-01 -4.22489434e-01
-3.96926582e-01 -4.38738853e-01 2.51834899e-01 5.42223454e-01
-1.40099525e-01 -1.79630339e-01 -3.26477855e-01 4.61460084e-01
-1.47236025e+00 -1.26549315e+00 1.27958313e-01 2.20316648e+00
7.49559999e-01 -1.70060992e-01 -1.16311677e-01 4.22566742e-01
9.44590032e-01 -9.98225510e-02 -9.12413597e-02 -2.95456558e-01
-3.14585805e-01 2.71390617e-01 -9.85092297e-02 6.27433062e-01
-6.87413037e-01 4.88054991e-01 6.41251040e+00 8.19333017e-01
-1.00154328e+00 1.97099656e-01 9.41990018e-01 3.12671550e-02
-2.21238568e-01 1.90030679e-01 -2.81083822e-01 5.65099359e-01
2.60390550e-01 1.67586863e-01 3.88213903e-01 6.19283080e-01
2.34994248e-01 -3.33012998e-01 -9.19771075e-01 1.11447155e+00
6.86543524e-01 -1.34014177e+00 3.07899624e-01 4.04770039e-02
7.46011972e-01 -8.51384759e-01 3.48883390e-01 -3.29302728e-01
-6.35121226e-01 -1.13960993e+00 6.92384183e-01 9.07982647e-01
8.76964688e-01 -8.25496137e-01 7.95952141e-01 2.72606343e-01
-8.50066304e-01 -3.88419665e-02 -4.10689741e-01 -4.66868188e-03
9.74570811e-02 6.09749913e-01 -6.22444451e-01 6.46530628e-01
6.61347747e-01 1.75077677e-01 -6.74553454e-01 1.11683297e+00
-6.01343885e-02 -8.24797750e-02 -1.28723592e-01 8.07552755e-01
-2.65327036e-01 -4.42480475e-01 5.01061440e-01 5.29349506e-01
5.40585637e-01 -1.63039041e-03 -5.11922836e-01 1.20065057e+00
-3.00358117e-01 3.27741504e-01 -5.73446453e-01 1.45928338e-01
1.18521631e-01 1.33005989e+00 -7.85861433e-01 -1.41612634e-01
-1.00398220e-01 1.22398698e+00 -1.51232937e-02 -4.62385975e-02
-6.88356757e-01 -6.59095049e-02 3.52394253e-01 3.43918264e-01
6.92584738e-02 3.05172563e-01 -3.47966641e-01 -9.92330968e-01
6.14738643e-01 -8.01822245e-01 1.58901855e-01 -1.09001195e+00
-1.00975513e+00 8.65203917e-01 -3.27244192e-01 -1.35334194e+00
6.09906241e-02 -9.81612727e-02 -8.06812704e-01 7.19274819e-01
-1.33102202e+00 -1.37847507e+00 -5.89021385e-01 6.43851101e-01
4.28989589e-01 -1.79704964e-01 5.73552012e-01 4.81545150e-01
-4.04917181e-01 6.51681900e-01 -2.39076242e-01 -6.23353831e-02
7.18918443e-01 -5.34287453e-01 -2.26077318e-01 9.64037716e-01
-2.09381223e-01 3.94948334e-01 7.40829706e-01 -6.45498931e-01
-9.48759019e-01 -8.14081848e-01 1.22205484e+00 6.92733675e-02
-2.34125793e-01 -4.78873178e-02 -8.33758652e-01 3.89205188e-01
2.84049362e-01 9.61897746e-02 2.58004040e-01 -4.61640805e-01
-5.37355781e-01 -4.39899534e-01 -1.92353666e+00 3.40208888e-01
9.14764166e-01 -3.73922169e-01 -3.59307736e-01 1.29488111e-01
2.84734756e-01 -1.10429693e-02 -6.71960950e-01 4.64978307e-01
7.09773183e-01 -1.57237411e+00 1.00416243e+00 -3.00140411e-01
7.26918757e-01 -2.65882999e-01 -1.00396713e-02 -9.27887917e-01
-9.36861187e-02 -4.58503097e-01 3.33324790e-01 1.35826790e+00
-6.79935440e-02 -3.92114908e-01 8.96141350e-01 2.14258954e-01
-6.74277693e-02 -4.00219798e-01 -1.17684305e+00 -5.17795026e-01
-1.88515976e-01 1.83878541e-01 6.99911654e-01 9.95631099e-01
-4.08426635e-02 -1.78395927e-01 -5.90907335e-01 -3.01044565e-02
5.96573293e-01 4.95336130e-02 3.83002371e-01 -1.33297431e+00
-6.57762587e-02 -3.50353539e-01 -6.30911171e-01 -2.25059509e-01
1.72897503e-01 -8.07370782e-01 -2.98048615e-01 -1.43074369e+00
6.06921732e-01 -1.03597008e-01 -2.85804980e-02 3.97743016e-01
2.98367351e-01 6.60164177e-01 1.41777381e-01 1.89402536e-01
-1.38086349e-01 7.46182859e-01 1.64044130e+00 -1.21411003e-01
1.14616804e-01 -2.33935878e-01 -7.94969916e-01 6.03342652e-01
7.09931016e-01 -4.84732270e-01 -4.62923259e-01 -3.17629308e-01
1.52187631e-01 5.80661058e-01 6.09603047e-01 -1.19318855e+00
2.77513176e-01 -1.19379625e-01 6.87515676e-01 -3.34989548e-01
6.76147580e-01 -9.47478831e-01 5.81191957e-01 5.09705663e-01
-1.28296345e-01 3.39041561e-01 -4.74834852e-02 1.92909807e-01
-5.60229063e-01 -2.04481155e-01 1.35075748e+00 -2.99121767e-01
-2.01719970e-01 2.38912508e-01 2.08493501e-01 -2.09762648e-01
1.09330094e+00 -5.88482440e-01 -1.57198310e-01 -5.25927484e-01
-6.57069683e-01 -3.32509011e-01 7.21519947e-01 2.66214758e-01
9.96994615e-01 -1.14204443e+00 -7.45352745e-01 6.07710004e-01
-1.92514390e-01 -2.99497008e-01 6.74403727e-01 9.67520833e-01
-8.69821012e-01 4.34878357e-02 -1.03078306e+00 -1.91029191e-01
-1.59131730e+00 5.49583316e-01 4.26432252e-01 -1.38997853e-01
-3.16899180e-01 6.07587874e-01 3.14208895e-01 -1.39238223e-01
2.81450987e-01 -5.98211400e-02 -1.11773014e-01 1.33741703e-02
4.74817157e-01 4.34170187e-01 1.64006159e-01 -8.17954183e-01
-2.08988652e-01 6.04898393e-01 7.44156837e-02 1.56830736e-02
1.30916405e+00 -3.12307745e-01 -8.61526310e-01 -1.56162560e-01
1.14134240e+00 1.93607926e-01 -1.03389859e+00 1.02042563e-01
-6.02064073e-01 -7.54616499e-01 -6.36080578e-02 -6.94490314e-01
-1.64064062e+00 8.06107163e-01 7.83888459e-01 -1.17216520e-01
1.52811205e+00 -9.20770839e-02 7.48118103e-01 -3.31667840e-01
1.83329254e-01 -9.66370940e-01 3.17036390e-01 -1.38171300e-01
1.29047370e+00 -9.54208434e-01 2.82368176e-02 -6.39507592e-01
-4.40810174e-01 1.08431792e+00 5.60155630e-01 -5.46520203e-02
6.59935474e-01 1.65807903e-01 -8.61009434e-02 -1.32202148e-01
-4.24797595e-01 1.57864437e-01 1.49860680e-01 5.67811191e-01
2.50673264e-01 -3.30060035e-01 -7.72407353e-01 2.69107342e-01
-2.39105552e-01 -6.91146450e-03 6.15902722e-01 5.76078176e-01
-1.82318971e-01 -1.02571714e+00 -4.51719224e-01 7.46312141e-02
-3.48974228e-01 2.25467786e-01 -5.07874250e-01 5.62152624e-01
5.91972530e-01 9.93918657e-01 -3.13908909e-03 -1.99896038e-01
1.58509016e-01 -1.59062091e-02 7.49130487e-01 -3.10415298e-01
-4.66126293e-01 -1.48908878e-02 -3.85489792e-01 -5.42527497e-01
-5.24675071e-01 -3.99090052e-01 -9.40627992e-01 -5.87765932e-01
-1.21477015e-01 9.12550688e-02 7.14046717e-01 6.70939445e-01
3.94095927e-01 2.45810568e-01 8.36377084e-01 -6.36599362e-01
-2.79106051e-01 -7.62273312e-01 -7.98076093e-01 5.24991751e-01
3.33320916e-01 -6.45999372e-01 -4.02949870e-01 8.81930441e-03] | [12.706623077392578, -0.0973181501030922] |
0fca832b-e6c5-405d-bec0-ec2f08075a36 | pointinverter-point-cloud-reconstruction-and | 2211.08702 | null | https://arxiv.org/abs/2211.08702v1 | https://arxiv.org/pdf/2211.08702v1.pdf | PointInverter: Point Cloud Reconstruction and Editing via a Generative Model with Shape Priors | In this paper, we propose a new method for mapping a 3D point cloud to the latent space of a 3D generative adversarial network. Our generative model for 3D point clouds is based on SP-GAN, a state-of-the-art sphere-guided 3D point cloud generator. We derive an efficient way to encode an input 3D point cloud to the latent space of the SP-GAN. Our point cloud encoder can resolve the point ordering issue during inversion, and thus can determine the correspondences between points in the generated 3D point cloud and those in the canonical sphere used by the generator. We show that our method outperforms previous GAN inversion methods for 3D point clouds, achieving state-of-the-art results both quantitatively and qualitatively. Our code is available at https://github.com/hkust-vgd/point_inverter. | ['Sai-Kit Yeung', 'Duc Thanh Nguyen', 'Binh-Son Hua', 'Jaeyeon Kim'] | 2022-11-16 | null | null | null | null | ['point-cloud-reconstruction'] | ['computer-vision'] | [-1.37800887e-01 3.27798247e-01 1.51980713e-01 6.10621870e-02
-9.79246855e-01 -9.15906191e-01 8.52061152e-01 -5.98459899e-01
3.05396140e-01 4.66189325e-01 -1.41012251e-01 -3.87053937e-01
3.92738611e-01 -1.29718125e+00 -1.22727001e+00 -6.78511322e-01
3.83063376e-01 1.14753675e+00 -1.39681160e-01 -1.66139409e-01
1.73166811e-01 1.05018055e+00 -1.29090071e+00 -1.58162206e-01
7.59678662e-01 6.64919734e-01 -4.51851897e-02 7.44844556e-01
-1.14127569e-01 -1.75579323e-03 -5.55970669e-01 -3.70666385e-01
7.53402650e-01 -5.34877837e-01 -5.79358459e-01 3.70439328e-02
4.43987876e-01 -2.56395489e-01 -4.25647050e-01 9.38616633e-01
2.61543781e-01 -1.71550855e-01 1.00012243e+00 -1.56598544e+00
-8.22606564e-01 -8.11169967e-02 -3.87198389e-01 -4.68556643e-01
3.55879813e-01 9.90188792e-02 6.58023000e-01 -1.06910896e+00
8.72187495e-01 1.16075051e+00 6.31748736e-01 5.52236795e-01
-1.35071528e+00 -8.49643528e-01 -3.79177690e-01 -3.52666318e-01
-1.66943049e+00 -1.14486545e-01 1.13748276e+00 -7.23699033e-01
9.13192511e-01 3.05417746e-01 1.09319580e+00 9.73052979e-01
1.99436322e-01 2.81309545e-01 8.09607208e-01 -3.93202305e-01
2.10512221e-01 -2.16944128e-01 -6.06258631e-01 7.33367920e-01
5.80350347e-02 3.12274635e-01 -3.59005511e-01 -3.06283027e-01
1.60662603e+00 1.73611507e-01 3.66582051e-02 -8.31751406e-01
-1.23597527e+00 9.97128069e-01 7.74701834e-01 -1.03414468e-02
-4.31470901e-01 5.51483691e-01 -3.84273112e-01 8.33517462e-02
9.14340377e-01 2.93752581e-01 -2.67530810e-02 -1.12881579e-01
-8.19014847e-01 5.54132819e-01 5.70495546e-01 1.45998013e+00
1.01428008e+00 1.04771920e-01 1.42626747e-01 2.26195946e-01
6.28512919e-01 1.11039972e+00 -2.86494106e-01 -1.27246428e+00
3.85821611e-01 4.83505040e-01 -4.47838195e-02 -8.35194230e-01
2.69609541e-01 -2.54831642e-01 -8.95040572e-01 6.14781797e-01
-1.19548663e-01 6.88520670e-02 -1.14765048e+00 1.43919003e+00
4.04425263e-01 5.86399019e-01 -1.71962082e-02 7.11710930e-01
6.60211086e-01 6.99620485e-01 -5.99903643e-01 3.72890383e-01
8.81088257e-01 -4.85030919e-01 -7.29224309e-02 3.04052811e-02
2.22670615e-01 -7.35066533e-01 8.91862631e-01 -2.42856592e-01
-1.28690481e+00 -5.02886653e-01 -9.52540755e-01 -2.77730465e-01
-7.40970448e-02 -1.56407058e-01 3.75014573e-01 3.41472089e-01
-1.15384138e+00 4.04586673e-01 -1.15438092e+00 -9.85012576e-02
6.69492602e-01 2.54707158e-01 -4.03958470e-01 5.63461930e-02
-8.39375973e-01 6.47977710e-01 -2.63886284e-02 -2.13578880e-01
-1.06577206e+00 -9.17612970e-01 -8.73746991e-01 -1.93177164e-01
-1.58643112e-01 -1.44459617e+00 1.20367324e+00 -2.34568864e-01
-1.47949493e+00 1.38884282e+00 -4.34558183e-01 -2.62276858e-01
5.67838371e-01 -1.95807517e-02 1.59252286e-01 6.47444800e-02
4.03690666e-01 8.53047490e-01 8.94145608e-01 -1.68601334e+00
-1.09807946e-01 -4.18531835e-01 -1.14160687e-01 2.10603371e-01
7.48942375e-01 -5.58531821e-01 -5.49777389e-01 -5.84632456e-01
5.31719983e-01 -1.38278186e+00 -1.34104267e-01 1.21551417e-01
-8.46141875e-01 5.82412668e-02 9.68205273e-01 -2.89687723e-01
1.41934961e-01 -2.08112192e+00 5.03356695e-01 4.54212457e-01
3.84072483e-01 -1.60683319e-01 1.27038389e-01 4.45861220e-01
-3.05790186e-01 4.03748780e-01 -4.83424604e-01 -8.33669007e-01
1.85405478e-01 2.65425026e-01 -7.87570536e-01 6.07092023e-01
2.50097901e-01 1.28284085e+00 -7.53373742e-01 -4.00299430e-02
4.62817132e-01 8.85916650e-01 -6.19455636e-01 3.66790444e-01
-3.66621822e-01 9.45544124e-01 -5.05780697e-01 5.67708671e-01
9.35531974e-01 -2.58116156e-01 -5.51611245e-01 4.97884117e-02
6.60602748e-02 3.03465188e-01 -6.62297904e-01 2.20164490e+00
-3.20112109e-01 4.31678385e-01 -2.63994157e-01 -3.80019248e-01
1.22984540e+00 3.72253835e-01 5.16828656e-01 -9.27471071e-02
-1.00634163e-02 2.68635631e-01 -4.58148271e-01 3.46080363e-01
3.41919273e-01 -3.44769955e-01 -2.34599724e-01 3.99435580e-01
9.56127942e-02 -1.35940158e+00 -4.85038161e-01 2.07337111e-01
8.37654948e-01 4.79775488e-01 -4.25197519e-02 6.68297112e-02
1.25751644e-01 7.83859789e-02 1.68232381e-01 5.47590017e-01
6.05998993e-01 1.07836282e+00 4.95043606e-01 -8.29217657e-02
-1.55814433e+00 -1.69701695e+00 -7.91943744e-02 -1.69740587e-01
2.79181208e-02 -3.47805142e-01 -7.12318063e-01 -5.05565584e-01
3.89640838e-01 1.01781380e+00 -6.41183555e-01 -1.00087911e-01
-4.52743500e-01 8.03498402e-02 4.56327975e-01 3.71850073e-01
3.27074081e-01 -7.28320003e-01 -1.40274897e-01 -1.90980777e-01
4.98464890e-02 -9.96257007e-01 -3.80827218e-01 -8.02323148e-02
-1.12045050e+00 -8.76121104e-01 -6.36981726e-01 -5.77170610e-01
1.19928193e+00 1.83722839e-01 1.30911160e+00 -1.02375463e-01
1.69442520e-01 4.68590468e-01 -1.05302610e-01 -6.85752749e-01
-8.12280595e-01 1.95014104e-01 -7.67575651e-02 -4.45935428e-01
3.06255400e-01 -1.04680419e+00 -4.76471275e-01 3.66004944e-01
-7.98500717e-01 4.01326656e-01 1.17536418e-01 4.60697949e-01
1.22755396e+00 -2.59023964e-01 -1.59137592e-01 -7.97433555e-01
1.63524032e-01 -3.22078705e-01 -9.97101068e-01 -3.71690571e-01
-3.45716327e-02 1.86441690e-02 2.13017359e-01 1.88811541e-01
-3.29668373e-01 2.56026238e-01 -4.87456948e-01 -1.20183253e+00
-2.45556384e-01 -2.52270866e-02 -3.15690219e-01 -3.36375713e-01
5.55862665e-01 2.65561849e-01 2.85235167e-01 -4.18805003e-01
7.14278162e-01 2.96966046e-01 6.59435034e-01 -5.54907978e-01
1.71123660e+00 7.71961272e-01 4.46409613e-01 -4.91316617e-01
-4.91249353e-01 -2.11950943e-01 -1.02065468e+00 -9.53632407e-03
1.00473416e+00 -1.05724001e+00 -3.99406105e-01 5.83672583e-01
-1.49934590e+00 -4.18262482e-01 -7.45389640e-01 -2.51508188e-02
-1.17951334e+00 -5.89660332e-02 -2.61521876e-01 -3.28553736e-01
-3.24225664e-01 -1.35762227e+00 1.69073796e+00 -2.11992308e-01
-1.90915570e-01 -9.50775146e-01 3.48619550e-01 1.40204519e-01
2.36682091e-02 8.19068551e-01 7.79835522e-01 -1.30648185e-02
-1.26952362e+00 -3.77888203e-01 9.64155644e-02 2.55272627e-01
3.63195091e-01 5.57753034e-02 -9.47246611e-01 -2.24727124e-01
1.59400001e-01 2.09737912e-01 4.19883400e-01 4.56435442e-01
1.21096385e+00 -1.31249964e-01 -4.37241435e-01 1.35686624e+00
1.43466592e+00 -1.83160417e-02 9.26746309e-01 1.12571016e-01
1.10162437e+00 -1.13202073e-01 2.39304990e-01 1.17115475e-01
3.56095880e-01 8.53546679e-01 8.00398946e-01 -1.90693289e-01
-1.38413250e-01 -9.00839925e-01 -6.50931075e-02 8.82707238e-01
-2.99955130e-01 -2.28630215e-01 -1.07841575e+00 4.11049008e-01
-1.62247956e+00 -6.80095255e-01 -5.46789058e-02 2.04273772e+00
5.10554671e-01 -1.97544843e-02 -1.27968654e-01 6.52409196e-02
5.67336440e-01 1.90047398e-01 -5.08325160e-01 -1.16024770e-01
-5.52595444e-02 6.34647012e-01 5.93320787e-01 7.86394179e-01
-7.78712809e-01 1.02749145e+00 6.03800201e+00 5.29469907e-01
-1.03259921e+00 1.85142517e-01 3.32992047e-01 -5.59218079e-02
-8.63203764e-01 1.71742544e-01 -6.80117369e-01 4.72077250e-01
4.66137230e-01 -4.39838529e-01 4.31862146e-01 7.47820437e-01
-9.16756094e-02 4.81761009e-01 -1.22028136e+00 1.03476584e+00
-1.14408629e-02 -1.77875507e+00 2.85232127e-01 5.46141148e-01
1.00075209e+00 2.66130686e-01 2.64290690e-01 -2.14084387e-01
4.82918561e-01 -1.16660464e+00 8.04056823e-01 5.91210723e-01
1.48335910e+00 -8.24671626e-01 2.52950549e-01 4.12308425e-01
-7.89999664e-01 7.98141837e-01 -3.09629261e-01 2.39506871e-01
4.24010634e-01 6.87509775e-01 -1.11310339e+00 7.13211060e-01
4.93185580e-01 8.79791021e-01 -1.57675475e-01 7.16917992e-01
-7.85963535e-01 4.48973387e-01 -5.82357764e-01 4.84778672e-01
1.60805061e-01 -5.13032675e-01 1.13929439e+00 4.78391618e-01
9.50277030e-01 9.30586681e-02 -1.69907823e-01 1.56380856e+00
-4.01219606e-01 -4.99707371e-01 -1.24918485e+00 9.44555402e-02
6.19135559e-01 7.40299463e-01 -5.02265394e-01 -6.71552792e-02
7.09110498e-02 1.01532173e+00 1.61200948e-02 3.00910920e-01
-8.43622148e-01 -2.51984626e-01 9.97644186e-01 4.35143352e-01
5.56703031e-01 -7.63270557e-01 -5.86118877e-01 -1.13391805e+00
7.63298273e-02 -4.57910717e-01 -3.34864795e-01 -1.33166409e+00
-1.11689711e+00 5.52402556e-01 1.08033553e-01 -1.62107694e+00
-5.60009181e-01 -4.00763690e-01 -6.57083213e-01 1.33390760e+00
-1.13207865e+00 -1.41590023e+00 -5.21322370e-01 5.51799953e-01
-1.53375017e-02 -2.69645393e-01 1.06864893e+00 -1.62556708e-01
2.54532695e-01 2.70896703e-01 7.68307643e-03 8.75882134e-02
1.31728411e-01 -1.30764878e+00 1.48175728e+00 8.84463310e-01
3.95499378e-01 4.31650072e-01 5.53237557e-01 -8.72453630e-01
-1.47802973e+00 -1.18486905e+00 4.56627190e-01 -1.06311882e+00
3.44079763e-01 -6.63303196e-01 -7.06979215e-01 1.15413356e+00
-4.62039523e-02 -1.05627865e-01 4.54875946e-01 -3.65961343e-01
-2.21390262e-01 2.00162679e-01 -1.36309123e+00 5.96110463e-01
1.27216446e+00 -7.98099756e-01 -4.97638434e-01 5.46264112e-01
1.05498254e+00 -1.27908337e+00 -8.95182908e-01 2.33408570e-01
-7.27830175e-03 -7.14728057e-01 1.34273040e+00 -1.48779541e-01
7.75149524e-01 -6.58986688e-01 -2.19538793e-01 -1.72417533e+00
-3.49904388e-01 -7.47666180e-01 -2.23475948e-01 7.80566931e-01
1.91387683e-01 -8.18768322e-01 1.21308506e+00 1.11836031e-01
-3.87373805e-01 -5.98939717e-01 -1.23720825e+00 -6.48448706e-01
4.35443759e-01 -4.61220711e-01 1.31778204e+00 7.15706110e-01
-8.17504883e-01 9.76811424e-02 -9.03404783e-03 5.13268709e-01
9.14219677e-01 3.28772098e-01 1.32546830e+00 -1.05220807e+00
-1.85065463e-01 -3.85208547e-01 -8.85220408e-01 -1.21861744e+00
3.29585046e-01 -1.21180236e+00 -1.55481339e-01 -1.72667408e+00
-3.90027702e-01 -6.72453225e-01 3.99743497e-01 3.85224968e-01
2.10403696e-01 5.77480972e-01 3.87569964e-01 4.04420137e-01
2.36479864e-01 7.82005727e-01 1.51619709e+00 -9.25114080e-02
-1.13112323e-01 1.85252935e-01 -6.14371002e-01 4.51766372e-01
7.83094287e-01 -6.17092192e-01 -3.30102384e-01 -8.02817106e-01
1.23849764e-01 1.59046352e-01 9.45231974e-01 -1.00107610e+00
-2.62044221e-02 -3.00862044e-02 4.54735637e-01 -1.13047099e+00
8.18593681e-01 -9.28512990e-01 9.02234375e-01 2.62000620e-01
2.08882034e-01 8.69204998e-02 3.28072280e-01 2.69084156e-01
-8.83945152e-02 3.45674455e-02 7.23351955e-01 -1.86353892e-01
-2.64546345e-03 8.59402359e-01 3.12547952e-01 -8.04078020e-03
1.02341950e+00 -2.26290345e-01 -3.09733421e-01 -5.45616090e-01
-5.62512040e-01 -4.52565998e-02 1.33939755e+00 3.22115123e-01
8.28383982e-01 -1.80973887e+00 -9.05890048e-01 7.97512472e-01
1.05866328e-01 1.05685997e+00 -6.49404079e-02 1.42114028e-01
-1.04256642e+00 3.78140002e-01 -1.41622633e-01 -1.11573863e+00
-8.42506945e-01 1.57596573e-01 5.72339356e-01 1.76003516e-01
-8.13646436e-01 9.08382833e-01 2.36972153e-01 -8.09201360e-01
-3.63614827e-01 -3.65999401e-01 8.09970796e-01 -6.98630452e-01
-1.03284329e-01 3.25061195e-02 5.73915541e-02 -8.63073051e-01
-3.05188149e-01 1.02292275e+00 4.80040580e-01 -2.68021643e-01
1.44512784e+00 4.67547536e-01 -2.64716178e-01 4.13993537e-01
1.35482967e+00 1.65295556e-01 -1.27960980e+00 4.28518206e-02
-9.07068193e-01 -7.91403174e-01 -5.73362261e-02 -2.13350549e-01
-1.16798973e+00 7.93807626e-01 2.21352890e-01 1.54519260e-01
7.67623663e-01 4.26585346e-01 7.54439294e-01 -7.03081861e-02
6.70521736e-01 -1.09521441e-01 -2.56956935e-01 5.88237166e-01
1.26669145e+00 -7.60297239e-01 6.24366850e-02 -7.96351135e-01
-3.54840159e-01 7.38479614e-01 1.36536047e-01 -7.53659487e-01
8.03722799e-01 2.42099151e-01 -3.38794813e-02 -4.30174470e-01
-3.90169173e-01 1.98316321e-01 3.33817750e-01 9.93108809e-01
-2.42371857e-02 2.58551687e-01 6.66515768e-01 -9.82349217e-02
-1.01965177e+00 6.92559630e-02 2.96429694e-01 7.06753016e-01
1.05890125e-01 -1.39789307e+00 -4.90889996e-01 1.87643185e-01
1.44066319e-01 1.62836947e-02 -3.98123503e-01 7.14485347e-01
-1.19184088e-02 2.19814211e-01 5.76368392e-01 -4.03581083e-01
4.91560966e-01 -1.53984372e-02 7.98619866e-01 -7.55268276e-01
-3.32134664e-02 -8.88084620e-02 -4.74312603e-01 -5.03606856e-01
-2.48498052e-01 -7.65829384e-01 -1.20338142e+00 -5.56395650e-01
6.37683645e-02 2.16284037e-01 8.91258597e-01 4.25756276e-01
8.25483799e-01 2.49221146e-01 9.26672101e-01 -1.40711820e+00
-1.34525314e-01 -5.67696154e-01 -5.13982415e-01 4.15576249e-01
4.05891240e-01 -7.08798707e-01 -5.39407849e-01 -8.43702536e-03] | [8.858694076538086, -3.663846492767334] |
a7116806-c141-4eb4-bb16-dabd6e0e6ee6 | program-analysis-of-probabilistic-programs | 2204.06868 | null | https://arxiv.org/abs/2204.06868v1 | https://arxiv.org/pdf/2204.06868v1.pdf | Program Analysis of Probabilistic Programs | Probabilistic programming is a growing area that strives to make statistical analysis more accessible, by separating probabilistic modelling from probabilistic inference. In practice this decoupling is difficult. No single inference algorithm can be used as a probabilistic programming back-end that is simultaneously reliable, efficient, black-box, and general. Probabilistic programming languages often choose a single algorithm to apply to a given problem, thus inheriting its limitations. While substantial work has been done both to formalise probabilistic programming and to improve efficiency of inference, there has been little work that makes use of the available program structure, by formally analysing it, to better utilise the underlying inference algorithm. This dissertation presents three novel techniques (both static and dynamic), which aim to improve probabilistic programming using program analysis. The techniques analyse a probabilistic program and adapt it to make inference more efficient, sometimes in a way that would have been tedious or impossible to do by hand. | ['Maria I. Gorinova'] | 2022-04-14 | null | null | null | null | ['probabilistic-programming'] | ['methodology'] | [ 1.87185854e-01 1.92864791e-01 -1.10203266e-01 -5.83306968e-01
-7.12864101e-01 -7.62509942e-01 6.19107723e-01 2.92030305e-01
-2.98354268e-01 6.24939561e-01 -1.37494221e-01 -9.93139505e-01
-4.09533173e-01 -1.07929003e+00 -2.85814613e-01 -8.01262140e-01
7.51940683e-02 8.16034257e-01 5.85996866e-01 1.19689219e-01
4.89900053e-01 5.87407291e-01 -1.98749959e+00 2.64053028e-02
5.54948449e-01 1.76527724e-01 1.27847478e-01 8.71996105e-01
-4.97484237e-01 5.26447475e-01 -4.61276025e-01 -6.14456952e-01
-1.74036488e-01 -4.41480398e-01 -8.43726456e-01 -4.91155297e-01
-3.04879397e-01 -7.12038428e-02 2.56521404e-01 1.00717509e+00
3.68427455e-01 -3.44450702e-03 6.96250737e-01 -1.46638751e+00
-4.46477756e-02 1.00691187e+00 -6.32718682e-01 -1.30066305e-01
6.87926233e-01 -1.94292650e-01 7.23876595e-01 -2.97607571e-01
3.81008595e-01 1.37130451e+00 6.95545912e-01 2.15334222e-01
-1.61177158e+00 -6.10187471e-01 -2.91551083e-01 -2.93330163e-01
-1.57958126e+00 -1.73282266e-01 6.78737462e-01 -5.31774998e-01
9.80267406e-01 5.93301654e-01 5.23701191e-01 6.86855137e-01
1.91393390e-01 6.69667661e-01 1.35681438e+00 -9.11531866e-01
3.32063645e-01 5.57368875e-01 4.05381829e-01 4.01355207e-01
3.86269569e-01 3.89868528e-01 -2.24753961e-01 -4.41892952e-01
2.62179673e-01 -3.32698226e-01 6.35917932e-02 -4.89829838e-01
-8.39655578e-01 8.06690812e-01 -4.90582883e-01 5.55181324e-01
-3.38997580e-02 3.04743141e-01 5.99885345e-01 1.59280553e-01
1.94748551e-01 3.48507352e-02 -4.68034118e-01 -8.51879299e-01
-1.20311475e+00 7.81070769e-01 1.35112989e+00 6.47244871e-01
6.79472387e-01 -1.79826528e-01 2.55254060e-01 6.06196284e-01
9.88390803e-01 6.16937578e-01 -1.79564413e-02 -9.51837540e-01
1.19172232e-02 4.73946124e-01 5.17906621e-02 -1.21959591e+00
-6.47719949e-02 1.57900050e-01 -2.54722148e-01 8.85835230e-01
7.21702933e-01 2.57249996e-02 -5.40110290e-01 1.39991295e+00
4.32867885e-01 -4.62277114e-01 -1.19626764e-02 3.36349308e-01
3.87996495e-01 7.88639903e-01 3.50684822e-01 -4.65618074e-02
1.22309840e+00 -1.11211345e-01 -5.40201366e-01 1.50014594e-01
9.24777806e-01 -9.57442105e-01 9.19117868e-01 7.89870858e-01
-1.03863430e+00 -1.09749563e-01 -1.06863785e+00 1.03378728e-01
-6.03491187e-01 -4.56822306e-01 8.32122982e-01 1.59434247e+00
-8.48018289e-01 5.46508193e-01 -1.21455586e+00 -3.73921245e-02
-8.62003341e-02 4.01010334e-01 -1.75167277e-01 6.96485266e-02
-9.99994457e-01 1.19394636e+00 8.61545205e-01 -1.27146661e-01
-2.26289913e-01 -6.26721978e-01 -8.80336940e-01 -1.32583855e-02
4.44757879e-01 -4.90658879e-01 1.30517673e+00 -7.95391560e-01
-1.65476358e+00 7.59035528e-01 -4.50607687e-01 -2.40086645e-01
4.64153588e-01 4.35408913e-02 -1.88856959e-01 -3.39719564e-01
-9.14527848e-02 1.80876046e-01 3.19602966e-01 -1.34496915e+00
-5.85104227e-01 -5.29552281e-01 1.51818484e-01 -1.90369919e-01
2.22897053e-01 6.90382779e-01 -2.13816643e-01 -4.33423370e-01
5.13242185e-02 -8.88983667e-01 -3.22428823e-01 -4.10893232e-01
-3.19938242e-01 -4.13650513e-01 6.69454634e-01 -5.28270781e-01
1.42985702e+00 -1.98674047e+00 6.54132152e-03 7.40679920e-01
-9.23515782e-02 9.61683318e-02 6.46800578e-01 5.82901478e-01
-9.93838012e-02 4.68399256e-01 -7.41520703e-01 -1.31801084e-01
5.14639676e-01 5.58768094e-01 -3.15689176e-01 3.73197973e-01
-1.15642488e-01 4.43136662e-01 -9.30470824e-01 -7.44661331e-01
5.19287109e-01 5.27425706e-01 -3.39554578e-01 -1.47131160e-01
-2.81731635e-01 -1.78184763e-01 -5.03546119e-01 6.54749274e-01
8.80189657e-01 5.25899112e-01 4.92577434e-01 4.79693711e-01
-3.79294693e-01 2.37528607e-01 -1.94672132e+00 1.30738604e+00
-5.20760238e-01 4.49418843e-01 -2.67296098e-02 -8.86224985e-01
1.01629663e+00 2.48695463e-01 2.50342280e-01 -1.46268010e-01
5.12233414e-02 3.45833927e-01 -2.16468811e-01 -4.23216790e-01
4.97236997e-01 -4.61602986e-01 -3.20584536e-01 8.87360632e-01
-1.80631757e-01 -5.66221178e-01 3.79547656e-01 4.04640473e-02
9.02374625e-01 7.04374731e-01 4.84826803e-01 -4.64902133e-01
6.35170460e-01 1.56860486e-01 4.04885083e-01 8.35507274e-01
1.64253846e-01 2.70062715e-01 8.17833960e-01 -2.80809134e-01
-9.30960119e-01 -1.20966983e+00 -3.72817814e-01 1.20742965e+00
-4.18103337e-01 -5.07230878e-01 -6.94975734e-01 -5.64864278e-01
-1.79084778e-01 1.17645562e+00 -3.01342458e-01 2.30563164e-01
-3.03936809e-01 -8.83526742e-01 6.15451217e-01 3.81235957e-01
5.18353619e-02 -9.42944169e-01 -9.31069434e-01 4.11516070e-01
2.41781488e-01 -3.79998982e-01 2.42604733e-01 4.00018871e-01
-8.99495006e-01 -7.60630131e-01 -3.13388616e-01 -2.16271356e-01
3.51211846e-01 -8.60558674e-02 1.32319725e+00 4.95809615e-02
-1.31191323e-02 2.57247061e-01 -3.27600151e-01 -8.23435068e-01
-8.20040166e-01 3.73705029e-02 -3.99633586e-01 -6.62563980e-01
7.95116186e-01 -7.84617484e-01 4.16124344e-01 1.48857117e-01
-1.21179497e+00 -3.13124985e-01 5.38065553e-01 6.66435242e-01
2.34888062e-01 7.06906021e-01 4.72809114e-02 -1.18330491e+00
6.24197483e-01 -4.39717293e-01 -9.27748144e-01 2.37735897e-01
-7.03234971e-01 4.45373714e-01 1.48932576e-01 2.79667694e-02
-1.18223405e+00 3.16134632e-01 -4.76813346e-01 3.32675397e-01
-5.03483415e-01 9.92756367e-01 -4.45239186e-01 1.21752322e-01
6.85647190e-01 1.76934153e-01 5.76844625e-02 -2.54103422e-01
3.62385213e-01 7.03408420e-01 4.02444005e-01 -1.13957560e+00
8.59369695e-01 2.25167066e-01 1.13870665e-01 -7.32182801e-01
-3.37577194e-01 -5.44897974e-01 -7.50544667e-01 -1.01735488e-01
5.02082586e-01 -1.43867821e-01 -6.54446244e-01 2.61384428e-01
-1.05263889e+00 -1.41818598e-01 -2.65965257e-02 2.11987168e-01
-5.66756070e-01 5.24801672e-01 1.83885038e-01 -1.52051675e+00
1.91914141e-01 -1.24008381e+00 9.53163505e-01 2.60748658e-02
-7.65679479e-01 -1.07897866e+00 4.62582171e-01 -2.44818311e-02
2.81466395e-01 3.73152345e-01 8.70654702e-01 -6.10382140e-01
-2.99595267e-01 -5.66267669e-01 -6.47463053e-02 9.19257030e-02
-2.40715697e-01 9.39339697e-01 -9.50616896e-01 3.02801371e-01
-2.59287879e-02 1.18469767e-01 2.04383105e-01 1.48700178e-01
1.02649426e+00 -1.89489201e-01 -6.18642032e-01 1.20824799e-02
1.56478953e+00 3.82302821e-01 1.03602338e+00 4.61643100e-01
2.94043154e-01 8.85675490e-01 4.82367396e-01 1.44352373e-02
5.27932763e-01 6.93628132e-01 4.67617922e-02 3.12556356e-01
4.14930969e-01 -2.03392860e-02 3.30018371e-01 4.98455137e-01
-2.39395514e-01 1.26236036e-01 -1.52955365e+00 7.61184633e-01
-1.84693837e+00 -1.31234026e+00 -5.29392600e-01 2.36146951e+00
1.16905487e+00 5.81547022e-01 2.99307734e-01 7.98111737e-01
4.54611748e-01 -9.20081064e-02 3.45748544e-01 -1.03764105e+00
3.36148769e-01 5.35828888e-01 3.38459581e-01 6.67283356e-01
-9.08912301e-01 4.44522351e-01 6.59232044e+00 8.21945369e-01
-7.14882910e-01 -1.68101396e-02 5.25072329e-02 4.38412279e-01
-8.10606778e-01 5.19184709e-01 -8.00370812e-01 5.24098456e-01
1.43386710e+00 -2.00365290e-01 2.69749463e-01 1.21596920e+00
1.18462130e-01 -7.67530024e-01 -1.14901006e+00 5.63175797e-01
-2.85991102e-01 -1.08658683e+00 -2.87228942e-01 8.38601664e-02
3.52527589e-01 -4.70996529e-01 -2.91671365e-01 4.20752138e-01
8.30354929e-01 -1.15048981e+00 7.82148600e-01 7.26153195e-01
-8.88055339e-02 -1.21628630e+00 1.24786317e+00 5.24030685e-01
-8.52498233e-01 2.35379890e-01 1.05722614e-01 -1.89613983e-01
2.08131433e-01 1.02505100e+00 -6.92748249e-01 6.88602269e-01
8.56553376e-01 -1.73296183e-02 -2.16626003e-01 1.17132032e+00
-3.99601310e-01 6.19725227e-01 -7.21320927e-01 -3.90618861e-01
3.44362743e-02 -1.72554687e-01 6.88913465e-01 1.68719649e+00
6.36777878e-02 -2.41506323e-01 5.51144257e-02 1.08759236e+00
7.33233273e-01 4.95782704e-04 -8.41674566e-01 -2.66207810e-02
4.96524930e-01 9.88057613e-01 -8.99897397e-01 -1.94472894e-01
-3.88487637e-01 2.55788594e-01 -2.88146794e-01 -2.03606542e-02
-7.90304184e-01 -7.58511901e-01 2.49503449e-01 9.15809795e-02
1.90540906e-02 -6.12313926e-01 -7.94651091e-01 -5.07162094e-01
8.85869488e-02 -8.59638035e-01 3.73562038e-01 -7.42132246e-01
-1.03421474e+00 1.74131036e-01 7.35580087e-01 -3.89966011e-01
-8.50025177e-01 -5.23627341e-01 -5.36259472e-01 1.26556993e+00
-9.32938099e-01 -1.03329694e+00 1.75493136e-01 5.21368496e-02
-7.06841722e-02 3.93191963e-01 1.14054155e+00 6.65362999e-02
-2.68553317e-01 4.81755972e-01 -7.58814663e-02 -1.61912873e-01
3.62674594e-01 -1.59375262e+00 -1.28464952e-01 1.02127278e+00
2.36822311e-02 1.05905962e+00 1.08246279e+00 -5.04399061e-01
-1.49215138e+00 -4.53289360e-01 1.28898716e+00 -8.21777344e-01
9.29278016e-01 -3.61590415e-01 -1.04559326e+00 4.67172951e-01
-1.26866221e-01 -4.84252691e-01 6.74915791e-01 6.41911626e-01
-4.68736351e-01 2.74191834e-02 -1.22068787e+00 6.78900659e-01
3.11547041e-01 -7.35824943e-01 -1.02237868e+00 1.34441711e-03
1.80651546e-01 -3.15094143e-01 -1.02844715e+00 1.65553436e-01
9.75147665e-01 -1.01492608e+00 1.01241386e+00 -1.28926054e-01
2.04192430e-01 -7.93967068e-01 -1.89701036e-01 -7.68637300e-01
2.37904966e-01 -6.23092353e-01 4.27659648e-03 1.66068864e+00
6.53354883e-01 -8.64084661e-01 7.25065291e-01 1.23789513e+00
-2.28298847e-02 -4.09065664e-01 -8.51779938e-01 -8.15335989e-01
3.74517828e-01 -1.23379695e+00 7.16613770e-01 7.03688085e-01
3.29548538e-01 -1.48320004e-01 1.54756829e-01 2.06027821e-01
5.65244675e-01 4.97233868e-02 1.02181339e+00 -1.42981327e+00
-4.27284211e-01 -9.37618852e-01 -6.31550014e-01 -3.55547875e-01
1.35541409e-01 -6.58152401e-01 3.73726636e-01 -1.42752206e+00
2.25790605e-01 -7.63327539e-01 1.19589947e-01 6.61362648e-01
-1.30017474e-01 -1.92444096e-03 -1.56593740e-01 -7.58052096e-02
-3.68072242e-02 -1.37533873e-01 4.12725627e-01 -3.04511772e-03
-2.58888245e-01 2.68819511e-01 -8.14660847e-01 8.32791865e-01
8.51492822e-01 -1.03754985e+00 -3.67581040e-01 -2.45976690e-02
7.37740397e-01 -4.81986627e-02 4.33747321e-01 -8.61881077e-01
3.63206863e-01 -3.25424850e-01 2.84646545e-02 -8.28771770e-01
-6.64996058e-02 -9.30344403e-01 7.26342380e-01 2.63981909e-01
-1.45205304e-01 3.15273739e-02 4.27372634e-01 3.52371633e-01
-1.67461663e-01 -1.09739745e+00 5.76080561e-01 -9.57481936e-02
-6.11803174e-01 -3.12753171e-01 -7.34354615e-01 -3.40080500e-01
1.19127345e+00 -4.34674054e-01 2.99905688e-01 -5.22201955e-02
-7.51919925e-01 -3.82651649e-02 8.00483644e-01 -1.01756312e-01
1.08432047e-01 -7.94793606e-01 -4.94956315e-01 -1.81581289e-01
-5.24308942e-02 2.10668594e-02 1.41985878e-01 6.63371384e-01
-7.05987334e-01 3.88854563e-01 -4.30659996e-03 -6.09240174e-01
-1.31107688e+00 7.12978601e-01 8.22167322e-02 -2.94492006e-01
-4.27560091e-01 6.29676819e-01 -5.72080970e-01 -7.89415777e-01
1.88178018e-01 -2.39265531e-01 -2.28429511e-02 -8.79951864e-02
7.00774133e-01 3.18766952e-01 -5.26973829e-02 -2.54785031e-01
-4.50355440e-01 4.12670672e-01 3.12172264e-01 -6.51975751e-01
1.23437881e+00 -3.02979536e-02 -5.89356482e-01 7.43538380e-01
6.43409967e-01 2.98706174e-01 -6.21418953e-01 1.46364048e-01
6.14287853e-01 -5.60733438e-01 2.37960279e-01 -8.66862893e-01
-4.42621112e-01 8.65380406e-01 2.36769587e-01 7.62861848e-01
9.52268779e-01 -2.06064761e-01 2.02668514e-02 2.12347031e-01
5.46858609e-01 -9.29408193e-01 -9.61346686e-01 3.52882057e-01
6.71683908e-01 -9.32121038e-01 3.38000625e-01 -5.18138826e-01
-3.34747136e-01 1.43830943e+00 -7.47836754e-02 1.28263563e-01
1.00192523e+00 8.09977233e-01 -4.27093625e-01 -1.82402417e-01
-2.60541290e-01 -1.63472220e-02 -3.58006218e-03 1.01606655e+00
5.61394036e-01 2.51733005e-01 -3.57030749e-01 5.27409971e-01
-3.79538506e-01 3.97808313e-01 4.39602673e-01 1.32443416e+00
-2.43100286e-01 -1.59271491e+00 -7.75641620e-01 2.22341061e-01
-6.91927969e-01 -4.68634404e-02 -1.79305911e-01 1.21226156e+00
-4.80045704e-03 9.55971897e-01 -2.94914544e-01 -2.21215665e-01
1.07445024e-01 2.77957380e-01 5.15962541e-01 -6.74632013e-01
-4.54116583e-01 -2.81840980e-01 4.59880799e-01 -2.06837535e-01
-7.41128504e-01 -1.02575290e+00 -1.06220233e+00 -7.81189799e-01
-2.62747794e-01 4.04619098e-01 1.20770669e+00 1.13256192e+00
-1.59832314e-01 4.28469777e-01 1.10644959e-01 -6.47830367e-01
-4.57790822e-01 -5.69590032e-01 -5.50574362e-01 -4.87273514e-01
-2.15145424e-01 -8.03136826e-01 -2.69570976e-01 8.90934169e-02] | [8.494254112243652, 6.6041693687438965] |
dc633c7a-6c1b-4adc-8d1e-4ee87c79afd6 | direct-acoustics-to-word-models-for-english | 1703.07754 | null | http://arxiv.org/abs/1703.07754v1 | http://arxiv.org/pdf/1703.07754v1.pdf | Direct Acoustics-to-Word Models for English Conversational Speech Recognition | Recent work on end-to-end automatic speech recognition (ASR) has shown that
the connectionist temporal classification (CTC) loss can be used to convert
acoustics to phone or character sequences. Such systems are used with a
dictionary and separately-trained Language Model (LM) to produce word
sequences. However, they are not truly end-to-end in the sense of mapping
acoustics directly to words without an intermediate phone representation. In
this paper, we present the first results employing direct acoustics-to-word CTC
models on two well-known public benchmark tasks: Switchboard and CallHome.
These models do not require an LM or even a decoder at run-time and hence
recognize speech with minimal complexity. However, due to the large number of
word output units, CTC word models require orders of magnitude more data to
train reliably compared to traditional systems. We present some techniques to
mitigate this issue. Our CTC word model achieves a word error rate of
13.0%/18.8% on the Hub5-2000 Switchboard/CallHome test sets without any LM or
decoder compared with 9.6%/16.0% for phone-based CTC with a 4-gram LM. We also
present rescoring results on CTC word model lattices to quantify the
performance benefits of a LM, and contrast the performance of word and phone
CTC models. | ['George Saon', 'Bhuvana Ramabhadran', 'Kartik Audhkhasi', 'Michael Picheny', 'David Nahamoo'] | 2017-03-22 | null | null | null | null | ['english-conversational-speech-recognition'] | ['speech'] | [ 4.16230977e-01 -7.29848370e-02 7.26830885e-02 -2.78732359e-01
-1.59889853e+00 -6.87619567e-01 5.19624472e-01 -1.45913705e-01
-6.50325835e-01 5.89501083e-01 1.72719285e-01 -9.67636347e-01
6.28548861e-01 -1.79980218e-01 -6.18790090e-01 -7.06018269e-01
5.71275204e-02 4.78785872e-01 2.17198908e-01 -1.11575857e-01
-2.42378622e-01 6.11629337e-02 -1.03640437e+00 3.13071638e-01
4.99036968e-01 6.42768741e-01 6.69197381e-01 1.14792776e+00
-4.14665900e-02 4.41587329e-01 -8.57348680e-01 -1.85500041e-01
1.78269953e-01 -7.09974766e-01 -4.99843091e-01 -1.55183777e-01
2.89318264e-01 -3.03386569e-01 -3.98610502e-01 7.90888011e-01
8.17219257e-01 3.56953323e-01 5.61232507e-01 -5.55649221e-01
-4.21492159e-01 6.82022154e-01 -6.61602542e-02 1.52416229e-01
2.46811390e-01 1.52968436e-01 1.21422517e+00 -1.09306133e+00
1.67064726e-01 1.24977326e+00 4.07464057e-01 7.66072273e-01
-1.21136582e+00 -7.37619519e-01 1.19026430e-01 5.76689607e-03
-1.57409739e+00 -1.17361200e+00 2.72836864e-01 -2.39073321e-01
1.80808520e+00 4.07141030e-01 2.74156988e-01 1.33309710e+00
3.05481441e-02 4.31189537e-01 8.71894956e-01 -7.94194758e-01
2.34805122e-01 1.40671507e-01 1.39281422e-01 5.38253248e-01
-4.25186336e-01 4.40224648e-01 -7.44216561e-01 -7.08453655e-02
3.04990679e-01 -6.66108370e-01 -2.46075228e-01 6.07411146e-01
-1.12009263e+00 7.18477547e-01 -1.78069502e-01 2.35767812e-01
-1.22833252e-01 5.00022829e-01 4.73428249e-01 3.40422869e-01
4.35886770e-01 1.62667960e-01 -4.69971031e-01 -5.30713201e-01
-1.11300623e+00 -5.33207536e-01 7.75753140e-01 9.80559647e-01
3.03748757e-01 7.58628190e-01 -1.57131270e-01 1.20511150e+00
3.84072900e-01 8.13110590e-01 8.19150627e-01 -5.66413879e-01
5.55915952e-01 -6.30208492e-01 -1.57750368e-01 -2.28810549e-01
3.16708237e-02 -6.38275027e-01 -5.70329428e-01 -1.90653354e-02
1.34445056e-01 -2.21003622e-01 -1.36658108e+00 1.76159132e+00
-1.37685746e-01 5.61436355e-01 2.66913772e-01 5.17373681e-01
4.66731191e-01 1.26396978e+00 -1.08599946e-01 -3.07128966e-01
1.27534759e+00 -1.08665919e+00 -8.94760370e-01 -5.45254469e-01
9.07206655e-01 -9.89212394e-01 1.32981718e+00 4.36883509e-01
-1.26130438e+00 -6.95303440e-01 -1.24112344e+00 -5.26959598e-02
-2.54000127e-01 2.35496581e-01 -1.44866318e-01 9.29401100e-01
-1.36942637e+00 2.24448964e-01 -1.06146789e+00 -2.51482844e-01
-3.86770129e-01 3.27150464e-01 -3.92715223e-02 1.27209589e-01
-1.24827743e+00 9.30090845e-01 3.92942071e-01 1.53495371e-02
-1.21670222e+00 -3.69756013e-01 -7.91931033e-01 4.94337082e-02
-7.45485201e-02 -1.86930776e-01 1.63003421e+00 -6.13454282e-01
-2.00405025e+00 5.95856249e-01 -7.48022139e-01 -7.78188169e-01
1.78410515e-01 -1.91317350e-01 -8.36604893e-01 -1.44052073e-01
-3.40783775e-01 6.16521299e-01 8.34622681e-01 -8.96863103e-01
-7.28339672e-01 3.00471187e-01 -5.16942799e-01 1.12001769e-01
-3.89074415e-01 2.21066713e-01 -6.49967790e-01 -8.12860847e-01
-1.68826967e-01 -1.10069191e+00 -8.57973620e-02 -7.50055969e-01
-4.91323709e-01 -2.15334430e-01 5.67192852e-01 -1.06069422e+00
1.43213904e+00 -2.38760734e+00 -2.12223276e-01 1.74444482e-01
-5.15595675e-01 5.03613234e-01 -4.18839753e-01 5.46647429e-01
-4.33527455e-02 2.32002422e-01 -3.91553968e-01 -9.23626184e-01
-2.66894195e-02 2.87841618e-01 -7.04872489e-01 3.28910530e-01
2.52881408e-01 6.94688320e-01 -7.02636361e-01 -1.41355932e-01
4.18924034e-01 6.48094177e-01 -4.59883779e-01 1.07034191e-01
-1.52932078e-01 3.16465139e-01 2.68596768e-01 2.10895032e-01
1.44185737e-01 2.90015608e-01 2.79119849e-01 2.79227316e-01
-1.04342047e-02 1.29523826e+00 -7.08880126e-01 1.62729263e+00
-1.07080233e+00 9.04547334e-01 6.18550070e-02 -7.37012148e-01
9.43255246e-01 8.42001855e-01 -1.54122129e-01 -7.47830033e-01
-9.07926485e-02 7.27126300e-01 2.37677038e-01 1.44102931e-01
4.41073418e-01 -3.11542004e-01 -6.13350794e-02 2.92681575e-01
3.62012275e-02 -3.05004150e-01 -1.29822060e-01 -4.82359864e-02
1.22164381e+00 -3.53173971e-01 -2.05476722e-03 -4.50581200e-02
4.25685734e-01 -1.29844427e-01 3.96591187e-01 7.12150991e-01
1.07805014e-01 9.01017547e-01 -1.70330137e-01 3.12411875e-01
-9.69275892e-01 -1.26239276e+00 -3.77855008e-03 1.18468952e+00
-4.29885834e-01 -4.65741158e-01 -6.09907091e-01 -2.14133054e-01
-4.30215925e-01 1.38407588e+00 1.93554416e-01 -2.38301650e-01
-9.48771477e-01 -4.77470666e-01 1.09820640e+00 4.42622215e-01
-6.60406351e-02 -1.05041325e+00 5.99773936e-02 6.18161738e-01
-2.22734004e-01 -1.49719095e+00 -1.01457953e+00 6.47824645e-01
-6.69908166e-01 -1.34661496e-01 -6.65429950e-01 -1.16484976e+00
3.39043617e-01 1.61742717e-01 9.72400606e-01 -4.02845442e-02
2.08996892e-01 5.92515506e-02 -5.12592733e-01 -1.64681390e-01
-9.83088613e-01 3.16506475e-01 3.90349001e-01 -5.03751598e-02
2.15841413e-01 -5.21057844e-01 -1.81152210e-01 3.62661690e-01
-7.59291291e-01 -5.12163267e-02 4.53081161e-01 9.93299186e-01
5.61677694e-01 -1.76905349e-01 8.45260084e-01 -3.96283001e-01
5.80239236e-01 -2.58907706e-01 -7.03203976e-01 1.10571846e-01
-5.38927853e-01 7.96119031e-03 8.33952427e-01 -5.15972853e-01
-9.24180746e-01 2.47106135e-01 -8.75895977e-01 -3.30550104e-01
6.14327192e-02 5.10290623e-01 -6.55942559e-02 4.56868172e-01
5.83734870e-01 5.05410314e-01 -3.14648002e-01 -6.40742123e-01
3.96948278e-01 1.23525405e+00 8.27498615e-01 -3.22584122e-01
8.74239087e-01 1.24637522e-01 -7.11006224e-01 -1.04873931e+00
-4.47720498e-01 -6.10469878e-01 -2.56767362e-01 2.37463355e-01
9.31026757e-01 -1.05321455e+00 -2.68634617e-01 2.66181231e-01
-1.40337563e+00 -7.27491975e-01 -1.68190286e-01 8.59557211e-01
-5.08679807e-01 2.92628139e-01 -9.14715886e-01 -1.00546646e+00
-5.00884950e-01 -1.06382823e+00 9.99114335e-01 -3.98486346e-01
-3.76561433e-01 -1.00383723e+00 -6.03417261e-03 1.98936939e-01
6.12500727e-01 -5.58282197e-01 8.54494810e-01 -6.22441411e-01
-2.85285592e-01 -2.16386676e-01 3.13791931e-01 9.00044560e-01
1.25705421e-01 -2.03932300e-01 -1.36759663e+00 -5.98653436e-01
-1.86084174e-02 -1.38454631e-01 8.12084675e-01 3.32329452e-01
8.75709832e-01 -2.52597630e-01 -1.39642194e-01 5.18671691e-01
1.09357619e+00 7.06432283e-01 6.60073519e-01 -2.26535469e-01
5.25179863e-01 1.98997498e-01 3.98851752e-01 5.81654124e-02
2.63900310e-02 9.48704481e-01 -4.26949225e-02 -5.87541871e-02
-6.23605132e-01 -5.21300077e-01 1.21362913e+00 1.88455665e+00
6.66698098e-01 -7.92516291e-01 -1.03016436e+00 7.45592892e-01
-1.26427460e+00 -4.99315351e-01 -1.52379692e-01 2.51914907e+00
1.05526495e+00 4.25929815e-01 -1.22724988e-01 1.79312468e-01
7.38261163e-01 6.80573434e-02 -2.41360098e-01 -7.10069895e-01
-1.05567694e-01 6.94021881e-01 6.91935241e-01 1.15419304e+00
-5.67947745e-01 1.37361634e+00 6.20629644e+00 1.35836184e+00
-1.18197715e+00 4.97386426e-01 4.87333328e-01 -1.95037499e-01
-2.47860447e-01 5.60989790e-02 -1.01366103e+00 3.69571239e-01
2.03285718e+00 1.00381508e-01 6.83648705e-01 3.88739347e-01
4.12263036e-01 1.83983073e-01 -1.37741923e+00 1.03592110e+00
-8.76999572e-02 -9.91026521e-01 -4.46604565e-02 2.22164858e-03
5.31209946e-01 4.70289260e-01 1.52456895e-01 4.58798409e-01
3.70839894e-01 -1.41164029e+00 9.55059886e-01 -9.56996009e-02
1.45062816e+00 -6.43600643e-01 5.29000461e-01 3.60512257e-01
-1.43673372e+00 4.34929043e-01 -2.84276158e-01 -3.14596947e-03
5.57878256e-01 4.94783491e-01 -1.37987638e+00 2.00604275e-01
1.87315300e-01 2.12711543e-01 -1.22259386e-01 8.01584303e-01
-1.86686426e-01 1.55007243e+00 -5.25709093e-01 -1.24840587e-01
4.57615852e-01 4.64123674e-02 6.15614414e-01 1.73962247e+00
5.81761062e-01 -1.78266063e-01 7.84760807e-03 3.63587141e-01
-9.67563093e-02 -1.78815657e-03 -3.42375040e-01 -2.99392581e-01
8.25799167e-01 6.60153866e-01 -4.95067656e-01 -3.71202826e-01
-3.11635405e-01 1.31247723e+00 1.02603704e-01 5.90747118e-01
-8.46398532e-01 -5.18491745e-01 7.12384880e-01 -1.15975618e-01
4.17648524e-01 -8.89158249e-01 -1.08514369e-01 -7.71072149e-01
-3.39515358e-02 -1.01328599e+00 -9.94673222e-02 -5.21609902e-01
-1.12823951e+00 9.48551536e-01 -3.59561056e-01 -1.05956292e+00
-7.24636376e-01 -6.39536142e-01 -5.72369158e-01 1.25180542e+00
-1.55370867e+00 -6.25269771e-01 3.23295474e-01 3.68305534e-01
9.88829255e-01 -2.06694871e-01 1.14403701e+00 5.08484364e-01
-5.18051982e-01 1.02080643e+00 4.96810466e-01 2.11958095e-01
6.31180584e-01 -1.12887526e+00 1.21010447e+00 9.63440716e-01
6.10602200e-01 5.40540695e-01 5.43090940e-01 -4.79532927e-01
-1.34024000e+00 -1.20384073e+00 1.21187079e+00 -6.24935925e-01
5.24531722e-01 -9.37158644e-01 -8.83937895e-01 6.54793203e-01
4.26266909e-01 -1.23299100e-01 5.22554219e-01 3.42589170e-02
-2.31742725e-01 2.35554930e-02 -5.35062492e-01 7.58383512e-01
1.04936755e+00 -1.08176398e+00 -4.59373981e-01 3.68171632e-01
1.27428007e+00 -3.21145386e-01 -4.28725630e-01 1.23630472e-01
3.55326205e-01 -4.20783550e-01 7.82961965e-01 -3.00004721e-01
-2.48890638e-01 -3.53856474e-01 -5.69662333e-01 -1.51208556e+00
-2.66769715e-02 -1.09591591e+00 9.72362310e-02 1.32484984e+00
1.02548945e+00 -8.31081510e-01 3.88514578e-01 5.19165285e-02
-6.92763925e-01 -2.81497478e-01 -1.36996484e+00 -1.40057886e+00
3.07128549e-01 -1.12713444e+00 1.79796666e-01 5.64715028e-01
-1.23982109e-01 6.57270670e-01 -3.51273566e-01 2.28163809e-01
7.58519769e-02 -4.96327430e-01 3.34560007e-01 -5.01133680e-01
-6.67491853e-01 -3.58617038e-01 -1.55609744e-02 -1.53401697e+00
2.92393327e-01 -1.14656925e+00 7.17786491e-01 -1.25640583e+00
-3.63953084e-01 -5.08722723e-01 -4.34456885e-01 3.90523046e-01
4.01170440e-02 1.73201144e-01 1.60274029e-01 3.31740938e-02
-8.06107298e-02 7.18484104e-01 4.51857269e-01 -1.46974148e-02
-2.58770287e-01 -5.51221073e-02 -1.10852674e-01 3.84797037e-01
1.02555382e+00 -6.68221951e-01 -4.41901594e-01 -7.44980991e-01
-2.53141135e-01 1.85931861e-01 -2.00341959e-02 -1.04130936e+00
2.89915651e-01 2.92547375e-01 -2.61929065e-01 -5.78819215e-01
8.06282759e-01 -4.93110329e-01 8.34764242e-02 5.70458829e-01
-3.94186974e-01 1.28584132e-01 3.82618368e-01 4.77005392e-01
-3.22145939e-01 -2.36862123e-01 8.80138993e-01 2.49374613e-01
-4.62808430e-01 -1.84087425e-01 -8.52311373e-01 1.16064668e-01
5.27699351e-01 -5.24243433e-03 1.10683618e-02 -6.66733623e-01
-5.37176132e-01 -9.76721793e-02 -3.41257313e-03 5.45212328e-01
6.96772158e-01 -1.22054172e+00 -9.92540419e-01 2.44793713e-01
-1.50830179e-01 -3.95688325e-01 -2.17305481e-01 5.28608501e-01
-3.28520596e-01 8.83186221e-01 5.87341011e-01 -5.48510373e-01
-1.24936771e+00 9.37120691e-02 5.32819211e-01 -8.57885182e-02
-4.36014503e-01 1.07792127e+00 2.33143851e-01 -4.01529133e-01
6.50299430e-01 -4.87214029e-01 5.06751299e-01 -3.39377642e-01
3.65511864e-01 1.59261182e-01 4.76015121e-01 -5.51791966e-01
-5.58977127e-01 2.92846352e-01 -3.09366100e-02 -1.04056668e+00
7.20012784e-01 -2.30449304e-01 4.81282234e-01 6.29857600e-01
1.42055035e+00 3.39190334e-01 -8.92670095e-01 -2.70948112e-01
1.73843667e-01 7.94537831e-03 2.20912993e-01 -7.52820671e-01
-7.70409346e-01 1.19123507e+00 6.32405162e-01 3.83458287e-02
9.25679982e-01 -7.83040822e-02 1.42809212e+00 4.38502073e-01
2.40767911e-01 -1.25755394e+00 -4.75957170e-02 1.06524038e+00
8.46248567e-01 -9.83442426e-01 -8.69692147e-01 -2.46402413e-01
-4.72465724e-01 8.68338346e-01 2.63347864e-01 1.55899107e-01
5.71415663e-01 5.89034855e-01 3.01752895e-01 3.70333284e-01
-1.08398652e+00 -4.11631972e-01 2.25067690e-01 4.40177262e-01
5.40572762e-01 2.73573935e-01 -8.54861364e-02 3.99037391e-01
-6.64732695e-01 -5.70053637e-01 4.68253970e-01 6.67352736e-01
-5.43278277e-01 -1.24046099e+00 -3.39591205e-01 1.89796358e-01
-5.16554654e-01 -8.78333271e-01 -2.75699496e-01 2.28274420e-01
-4.87733930e-01 1.65793955e+00 1.41259328e-01 -7.29645014e-01
1.56397119e-01 6.19532049e-01 -7.98284933e-02 -9.85347927e-01
-5.87614000e-01 5.19283235e-01 5.05707204e-01 -1.49002880e-01
2.57549763e-01 -4.99160439e-01 -1.53339887e+00 -2.87550539e-01
-5.46894670e-01 2.68327028e-01 9.60750878e-01 8.26519072e-01
2.72064716e-01 7.54198909e-01 6.24733925e-01 -4.99997348e-01
-7.92433143e-01 -1.17310214e+00 -3.66787344e-01 -1.39816999e-01
5.96515954e-01 8.74599218e-02 -5.41236341e-01 2.02332795e-01] | [14.443574905395508, 6.734960556030273] |
0acb3126-4610-4e5b-9f9b-46e6fd583758 | puzzlenet-scene-text-detection-by-segment | 2002.11371 | null | https://arxiv.org/abs/2002.11371v1 | https://arxiv.org/pdf/2002.11371v1.pdf | PuzzleNet: Scene Text Detection by Segment Context Graph Learning | Recently, a series of decomposition-based scene text detection methods has achieved impressive progress by decomposing challenging text regions into pieces and linking them in a bottom-up manner. However, most of them merely focus on linking independent text pieces while the context information is underestimated. In the puzzle game, the solver often put pieces together in a logical way according to the contextual information of each piece, in order to arrive at the correct solution. Inspired by it, we propose a novel decomposition-based method, termed Puzzle Networks (PuzzleNet), to address the challenging scene text detection task in this work. PuzzleNet consists of the Segment Proposal Network (SPN) that predicts the candidate text segments fitting arbitrary shape of text region, and the two-branch Multiple-Similarity Graph Convolutional Network (MSGCN) that models both appearance and geometry correlations between each segment to its contextual ones. By building segments as context graphs, MSGCN effectively employs segment context to predict combinations of segments. Final detections of polygon shape are produced by merging segments according to the predicted combinations. Evaluations on three benchmark datasets, ICDAR15, MSRA-TD500 and SCUT-CTW1500, have demonstrated that our method can achieve better or comparable performance than current state-of-the-arts, which is beneficial from the exploitation of segment context graph. | ['Yiqing Hu', 'Hao Liu', 'Deqiang Jiang', 'Bo Ren', 'Antai Guo'] | 2020-02-26 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 1.84108585e-01 -1.16722196e-01 1.03403352e-01 9.00952965e-02
-7.00911582e-01 -5.99489331e-01 5.33589482e-01 1.60323590e-01
2.49950334e-01 9.42048356e-02 1.60109967e-01 -1.63869247e-01
1.02139920e-01 -8.94715548e-01 -7.20156312e-01 -3.85259241e-01
2.29635671e-01 7.12143898e-01 7.46720612e-01 -3.61008018e-01
5.24893939e-01 4.06016588e-01 -1.26244032e+00 6.02029741e-01
1.12484610e+00 1.03274429e+00 4.35752720e-01 5.30992270e-01
-5.34094393e-01 6.46871209e-01 -3.66137862e-01 -5.05355716e-01
5.14774144e-01 -2.68949032e-01 -6.17668986e-01 4.01204944e-01
4.45250124e-01 -1.44042850e-01 -5.07303119e-01 7.64464736e-01
3.64295542e-01 -4.47530076e-02 6.59077585e-01 -9.73458230e-01
-2.35875785e-01 7.44505048e-01 -1.07014203e+00 -9.80332792e-02
3.55805635e-01 2.01815531e-01 1.20951998e+00 -9.33160067e-01
6.67126775e-01 1.04364085e+00 9.75219727e-01 -5.37290983e-02
-1.03723419e+00 -4.38447118e-01 5.32276452e-01 3.73715609e-01
-1.45894599e+00 1.20845951e-01 1.10305178e+00 -3.75642866e-01
8.09470415e-01 4.39663619e-01 6.86079264e-01 8.74337196e-01
-8.79265144e-02 1.28015089e+00 4.95424628e-01 -3.36845666e-01
-1.54217714e-02 -3.09357584e-01 9.23344567e-02 7.61247039e-01
2.06665799e-01 -3.72585118e-01 -4.94341314e-01 -9.02058631e-02
8.25227857e-01 -2.53920317e-01 -3.47323388e-01 -6.71174884e-01
-1.20074809e+00 7.26878405e-01 6.12903297e-01 2.93872297e-01
-3.28523487e-01 -1.06077515e-01 5.19555748e-01 -3.15035611e-01
4.80070323e-01 3.36500436e-01 -2.48126462e-01 3.16101432e-01
-1.15528774e+00 3.44089329e-01 6.54826164e-01 1.13021183e+00
5.99108160e-01 -2.43620649e-02 -3.31302583e-01 9.52690721e-01
2.08694428e-01 2.93252230e-01 1.33404449e-01 -3.74248087e-01
1.08432722e+00 1.24804628e+00 -1.55971169e-01 -1.38028061e+00
-6.33657873e-01 -6.38550401e-01 -9.54895496e-01 -1.82735741e-01
3.72857362e-01 7.63731673e-02 -1.04368508e+00 1.07038009e+00
5.83971739e-01 2.04885393e-01 -2.82291234e-01 1.07130909e+00
1.04118848e+00 6.93441510e-01 -2.30625600e-01 3.15866858e-01
1.47162485e+00 -1.44206703e+00 -1.98455900e-01 -3.99601489e-01
6.87927902e-01 -1.03281784e+00 8.65463018e-01 6.71110332e-01
-1.03339481e+00 -5.70186555e-01 -1.05806899e+00 -1.77127406e-01
-1.35372728e-01 5.29393792e-01 4.00100946e-01 2.74433017e-01
-8.33892524e-01 5.15747666e-01 -5.18517613e-01 -4.45020080e-01
4.83106762e-01 2.26494625e-01 -9.71248299e-02 -1.81030989e-01
-6.20457828e-01 3.30458134e-01 5.17927945e-01 4.19696987e-01
-5.86278975e-01 -5.61171949e-01 -7.52284229e-01 3.23401958e-01
7.59902358e-01 -9.34559822e-01 8.41562808e-01 -8.57430995e-01
-1.12563896e+00 6.28872693e-01 7.75426179e-02 -4.06138092e-01
8.41023743e-01 -2.28212979e-02 -1.15841947e-01 3.75160009e-01
1.63275078e-01 6.34996653e-01 8.92554760e-01 -1.37611389e+00
-7.94046044e-01 -3.70915800e-01 -8.30920786e-02 4.79462236e-01
3.01693799e-03 -5.29501215e-02 -1.04773307e+00 -7.66224921e-01
6.53113246e-01 -6.61777437e-01 -3.50502849e-01 -1.35434657e-01
-1.00593078e+00 -2.84343719e-01 9.67436731e-01 -7.93322206e-01
1.28461695e+00 -1.83450758e+00 2.78824478e-01 2.03556046e-01
4.05522704e-01 2.94129848e-01 -2.22983167e-01 6.08026326e-01
-3.37481163e-02 4.10180278e-02 -3.33469659e-01 -6.45372570e-01
2.42810678e-02 -7.61378407e-02 -5.27310073e-01 4.43203181e-01
1.01571009e-01 1.00478005e+00 -5.22989213e-01 -7.02975094e-01
4.57855552e-01 1.59630269e-01 -5.08700430e-01 -3.31233740e-02
-7.59468496e-01 1.17399991e-01 -5.64455450e-01 7.66078591e-01
1.14446700e+00 -2.54417717e-01 2.19194412e-01 -3.83741379e-01
-2.35192955e-01 1.07331298e-01 -1.31626821e+00 1.66384661e+00
2.60938760e-02 6.25196934e-01 5.69546372e-02 -1.08779287e+00
1.09182858e+00 -9.19148773e-02 4.70302820e-01 -5.89394152e-01
1.41340092e-01 -1.53156677e-02 -2.33576521e-01 -5.99922001e-01
8.38571310e-01 4.47592974e-01 -7.98420385e-02 1.53334334e-01
-4.26837057e-01 -2.07167313e-01 3.36048126e-01 3.23185831e-01
1.12990510e+00 4.37527537e-01 1.11469164e-01 4.69108491e-04
5.45288503e-01 3.99341017e-01 4.31617200e-01 6.91278100e-01
1.49752468e-01 1.15947628e+00 6.55398011e-01 -7.50106514e-01
-1.10854840e+00 -8.48132074e-01 9.18468907e-02 8.24974835e-01
5.18597901e-01 -6.74050927e-01 -7.76443422e-01 -7.38197565e-01
-1.39501035e-01 4.53266382e-01 -6.61076605e-01 3.08161914e-01
-9.33429539e-01 -6.76002443e-01 3.77985388e-01 6.93926573e-01
7.35553920e-01 -1.02735281e+00 -3.73087138e-01 2.42337763e-01
-5.60041010e-01 -1.27561939e+00 -6.41853631e-01 7.61129633e-02
-7.16850221e-01 -1.22278953e+00 -5.72586894e-01 -8.68162692e-01
7.67833710e-01 7.78123379e-01 1.00046217e+00 4.16667551e-01
-4.23946083e-01 2.74632778e-02 -6.50189221e-01 -2.26009246e-02
-9.75968242e-02 2.02113941e-01 -5.50131619e-01 1.09206639e-01
-1.07772343e-01 -3.83752137e-01 -6.52914345e-01 6.34031832e-01
-7.77531683e-01 6.76133990e-01 5.91634512e-01 5.97529769e-01
6.93146348e-01 6.00494817e-02 -1.97278857e-02 -6.60020351e-01
2.79825032e-01 -3.97157460e-01 -5.83029687e-01 4.40526903e-01
-2.38411706e-02 -3.00466776e-01 8.95462632e-01 -2.68906325e-01
-9.42232847e-01 2.38944694e-01 -3.04215271e-02 -5.53967237e-01
-2.30060369e-01 4.55623031e-01 -3.23014110e-01 -7.78540149e-02
4.59854126e-01 6.05738580e-01 -6.45630300e-01 -4.20951247e-01
2.78575212e-01 3.94453853e-01 7.24971831e-01 -6.98979080e-01
7.82585382e-01 5.26827157e-01 7.06816241e-02 -8.76134038e-01
-6.78561926e-01 -8.53076696e-01 -8.77517998e-01 -2.23504141e-01
8.42254460e-01 -7.94289529e-01 -3.97534370e-01 6.06680989e-01
-1.47095788e+00 -3.44955146e-01 7.06296712e-02 -2.46692039e-02
-5.13983011e-01 9.12466466e-01 -4.66667622e-01 -7.33871996e-01
-4.12739694e-01 -1.03096950e+00 1.46358395e+00 1.61984906e-01
1.57951787e-01 -6.77543759e-01 -2.01805800e-01 5.76342046e-01
-1.25663608e-01 3.74233544e-01 9.59655523e-01 -5.61099470e-01
-1.00058758e+00 -2.48692006e-01 -5.91970265e-01 -2.99887896e-01
-3.09597075e-01 2.47021362e-01 -6.04414165e-01 2.89215217e-03
-1.89752087e-01 1.57801405e-01 9.84576344e-01 5.03238618e-01
1.21164703e+00 -2.66448874e-02 -5.47140539e-01 8.60087335e-01
1.47389972e+00 -4.72855289e-03 8.45938146e-01 3.67234379e-01
1.14187002e+00 4.50524300e-01 5.93309760e-01 5.60626388e-01
6.10138655e-01 9.30335462e-01 8.05869520e-01 -2.73931116e-01
-2.38060653e-01 -3.78919959e-01 8.16595852e-02 5.10722697e-01
-7.94580858e-03 -6.72217965e-01 -1.06244123e+00 4.37697917e-01
-2.20630288e+00 -7.21759319e-01 -1.03902364e+00 1.87745798e+00
8.69378820e-02 3.04435998e-01 3.28147262e-01 5.85790835e-02
9.59250093e-01 7.18131363e-02 -5.05520582e-01 -1.48036063e-01
-3.82629514e-01 -1.37472674e-01 3.46555382e-01 4.37418669e-02
-1.27986979e+00 1.23588777e+00 5.05836296e+00 1.29420114e+00
-8.80848825e-01 -2.81639159e-01 7.36780941e-01 2.64857113e-01
-1.11640595e-01 4.56172861e-02 -7.50239134e-01 3.34633380e-01
-2.39980370e-02 3.16641480e-01 4.00448710e-01 8.98188353e-01
2.87104636e-01 -4.33130473e-01 -8.07234168e-01 9.13608968e-01
3.24067414e-01 -1.59505939e+00 2.33931318e-01 -1.55312747e-01
7.90314317e-01 4.77414168e-02 -1.88670039e-01 1.75321981e-01
1.40944496e-01 -8.02985907e-01 1.04410660e+00 3.54400665e-01
4.74078208e-01 -6.50298178e-01 6.20758712e-01 5.50713539e-01
-1.84278905e+00 -1.11755751e-01 -5.81132948e-01 1.92164019e-01
1.27298638e-01 5.19136667e-01 -1.09545207e+00 1.12741005e+00
5.41319370e-01 6.65282726e-01 -7.45967925e-01 1.53362644e+00
-3.08508992e-01 6.32535398e-01 -4.27337378e-01 1.75216179e-02
3.97816032e-01 -4.04481620e-01 7.05444634e-01 1.20527959e+00
3.06132406e-01 1.18976263e-02 4.88855749e-01 1.08845651e+00
-2.20710896e-02 3.85876805e-01 -3.33164841e-01 4.07130867e-01
2.25402310e-01 1.58158839e+00 -1.42916656e+00 -3.20319533e-01
-2.75728852e-01 1.02015364e+00 4.72180605e-01 3.21150161e-02
-1.05545151e+00 -1.67322680e-01 4.94727679e-02 2.78611362e-01
8.76188695e-01 -3.34971160e-01 -8.44575703e-01 -1.04566550e+00
2.96756268e-01 -7.17916846e-01 4.23817396e-01 -1.06413460e+00
-1.17411983e+00 6.66021645e-01 -3.38592917e-01 -1.55268764e+00
4.10206467e-01 -4.41696644e-01 -1.10711920e+00 6.47300541e-01
-1.22925937e+00 -1.60368061e+00 -4.93955702e-01 4.60237741e-01
1.06267834e+00 1.20584041e-01 2.51464039e-01 -2.71381764e-03
-9.20808434e-01 5.46617389e-01 2.01522075e-02 3.40833694e-01
4.16442931e-01 -1.10306418e+00 8.67045581e-01 8.81218791e-01
3.56166840e-01 9.10932124e-02 4.95888740e-01 -9.98124242e-01
-1.26037645e+00 -1.30006921e+00 5.07966995e-01 -2.96279043e-01
5.95807731e-01 -6.48387790e-01 -9.52158213e-01 3.92206699e-01
9.58450213e-02 -1.28584817e-01 5.73077947e-02 -1.34289980e-01
-3.82278055e-01 2.27749452e-01 -8.50759327e-01 8.47570002e-01
1.22062099e+00 6.67912362e-04 -4.82502818e-01 5.26990235e-01
7.11091757e-01 -6.82926893e-01 -2.97459126e-01 3.48082870e-01
2.30572477e-01 -1.20336843e+00 1.01477408e+00 -2.57447511e-01
7.80432701e-01 -5.53808153e-01 -3.17650996e-02 -1.04505074e+00
-3.73486340e-01 -4.52500850e-01 4.67125922e-02 1.05059516e+00
2.18735680e-01 -4.53055762e-02 1.08668602e+00 1.60911903e-01
-7.44531572e-01 -8.66714895e-01 -8.61904919e-01 -6.29693687e-01
-1.33587033e-01 -6.09782398e-01 7.81335056e-01 8.46740484e-01
1.02729090e-02 3.29841316e-01 -3.21288764e-01 4.21540797e-01
4.71087158e-01 5.54814816e-01 1.14813292e+00 -1.21822858e+00
-3.25736105e-01 -7.81660199e-01 -3.49283338e-01 -1.49884212e+00
-2.68829674e-01 -8.76165390e-01 9.59352106e-02 -1.89333546e+00
2.45624244e-01 -5.80226302e-01 3.90426606e-01 3.56794715e-01
-4.06557590e-01 3.06207836e-01 4.05248493e-01 1.54463857e-01
-1.00983131e+00 4.18464392e-01 1.33379865e+00 -2.53672600e-01
-3.69654953e-01 -3.05108130e-02 -5.60496509e-01 9.14548635e-01
7.50516772e-01 -1.48273826e-01 -9.01624113e-02 -5.12321055e-01
3.24964166e-01 2.42614746e-01 4.63799715e-01 -1.17671514e+00
5.03112316e-01 -3.64628956e-02 3.73248816e-01 -1.52054620e+00
1.56184375e-01 -7.39637733e-01 6.94973990e-02 1.30175397e-01
7.95788690e-02 -1.42798275e-01 3.27492863e-01 5.29779017e-01
1.15138404e-01 -3.92694861e-01 4.64597851e-01 -4.70282100e-02
-7.17731893e-01 3.20401043e-01 -1.37526363e-01 -6.79494627e-03
1.01336098e+00 -5.59531808e-01 -5.42678237e-01 -6.43580779e-02
-5.01995921e-01 5.33438742e-01 4.72339779e-01 3.11841667e-01
6.67254031e-01 -1.09998417e+00 -8.77155721e-01 1.75274193e-01
1.43155098e-01 6.66933179e-01 6.58325255e-01 1.01815307e+00
-8.85149777e-01 3.95021528e-01 2.06759870e-01 -9.45009708e-01
-1.29631114e+00 7.24635065e-01 2.22144961e-01 -7.47057736e-01
-1.03936434e+00 7.87328839e-01 6.92528725e-01 -2.65202343e-01
1.49366304e-01 -5.38123608e-01 -1.64000154e-01 -7.20068812e-02
1.89314634e-01 3.28054875e-01 1.92542747e-01 -5.89937627e-01
-1.10675827e-01 9.64992583e-01 -1.03530146e-01 4.59408164e-01
1.33103442e+00 -6.07504025e-02 2.11247546e-03 -2.22534448e-01
6.27274871e-01 1.49466902e-01 -1.29632366e+00 -2.20327407e-01
-7.38558844e-02 -3.42424423e-01 -1.64148733e-01 -8.72302532e-01
-1.16090322e+00 8.33292782e-01 -6.75426945e-02 4.18893814e-01
1.03906715e+00 8.45989510e-02 8.71884048e-01 1.87671512e-01
2.96980739e-01 -1.05882788e+00 2.96967000e-01 5.88368356e-01
1.11568654e+00 -9.73543882e-01 1.50757238e-01 -9.83917356e-01
-6.17166936e-01 1.46873546e+00 8.67931247e-01 -2.53113180e-01
1.72199175e-01 2.13381261e-01 -3.61102641e-01 -3.91428798e-01
-4.70978558e-01 -2.07260728e-01 4.70716298e-01 5.07781804e-01
-1.19916812e-01 1.50096714e-01 -9.61775780e-02 8.85256827e-01
-2.63335168e-01 -4.52127844e-01 4.45285529e-01 5.04999399e-01
-6.18219495e-01 -8.80254090e-01 -8.72298598e-01 4.90931123e-01
4.19269837e-02 -2.23139986e-01 -6.29850805e-01 9.35819089e-01
3.68760526e-01 7.76883841e-01 2.19096914e-01 -4.73830312e-01
4.51159418e-01 -4.26992208e-01 2.91399866e-01 -5.81884205e-01
-7.62462735e-01 4.59972411e-01 8.54805410e-02 -3.93193096e-01
5.87792881e-02 -7.48493016e-01 -1.28692651e+00 -2.16430202e-01
-6.01531386e-01 -2.52330244e-01 5.13470531e-01 9.04286087e-01
3.34491402e-01 5.78278363e-01 7.29177713e-01 -1.11596680e+00
-2.46400665e-02 -7.97249138e-01 -4.57083225e-01 1.44584268e-01
-1.67076588e-01 -4.80030954e-01 -1.78342909e-01 -1.14570290e-01] | [12.07944393157959, 2.2892282009124756] |
0657e740-e1a7-434b-a5aa-ddc4f955e1ab | high-resolution-and-fast-face-completion-via | null | null | https://openreview.net/forum?id=Hkxx3o0qFX | https://openreview.net/pdf?id=Hkxx3o0qFX | High Resolution and Fast Face Completion via Progressively Attentive GANs | Face completion is a challenging task with the difficulty level increasing significantly with respect to high resolution, the complexity of "holes" and the controllable attributes of filled-in fragments. Our system addresses the challenges by learning a fully end-to-end framework that trains generative adversarial networks (GANs) progressively from low resolution to high resolution with conditional vectors encoding controllable attributes. We design a novel coarse-to-fine attentive module network architecture. Our model is encouraged to attend on finer details while the network is growing to a higher resolution, thus being capable of showing progressive attention to different frequency components in a coarse-to-fine way. We term the module Frequency-oriented Attentive Module (FAM). Our system can complete faces with large structural and appearance variations using a single feed-forward pass of computation with mean inference time of 0.54 seconds for images at 1024x1024 resolution. A pilot human study shows our approach outperforms state-of-the-art face completion methods. The code will be released upon publication. | ['Christopher G. Healey', 'Shaoliang Nie', 'Zeyuan Chen', 'Tianfu Wu'] | null | null | null | null | iclr-2019-5 | ['facial-inpainting'] | ['computer-vision'] | [ 3.31894666e-01 4.20149803e-01 4.27166820e-01 -4.60849166e-01
-8.32872629e-01 -2.56738633e-01 5.91004312e-01 -6.03945613e-01
-1.79959700e-01 7.98503995e-01 3.07176828e-01 2.98274159e-01
3.20475840e-04 -8.71831179e-01 -9.27894294e-01 -3.33505809e-01
-1.70847818e-01 6.39801085e-01 -2.71425724e-01 -2.37122133e-01
-6.01123162e-02 6.11868918e-01 -1.49842560e+00 6.24432325e-01
6.83964729e-01 6.98209822e-01 2.40304530e-01 9.30355489e-01
1.85506791e-01 8.19345176e-01 -4.89003181e-01 -6.01254880e-01
3.69496822e-01 -1.79496944e-01 -7.04486430e-01 1.79919273e-01
1.08816159e+00 -5.15358388e-01 -3.31617475e-01 9.43098724e-01
6.80782974e-01 4.59146537e-02 4.37861145e-01 -1.01289392e+00
-9.17741120e-01 2.61971831e-01 -6.98156774e-01 3.84714827e-02
5.74384153e-01 3.95528734e-01 5.62502563e-01 -1.25475764e+00
7.20911086e-01 1.63230753e+00 8.90921056e-01 1.06760764e+00
-1.62068224e+00 -6.95207894e-01 1.66493729e-01 -1.14058852e-01
-1.36888182e+00 -7.79731989e-01 6.75020218e-01 -4.02491719e-01
1.07864082e+00 2.34616727e-01 4.08584952e-01 1.26393127e+00
1.60792619e-01 1.04575329e-01 1.14943492e+00 -8.89108256e-02
2.82552779e-01 -8.99866894e-02 -6.14414752e-01 8.17699492e-01
-5.31955399e-02 1.37078419e-01 -5.55142999e-01 5.38175032e-02
1.38833737e+00 9.63283703e-02 -1.89077124e-01 1.76629219e-02
-6.54356718e-01 8.14247012e-01 6.91424131e-01 7.17486888e-02
-6.66771829e-01 5.12558103e-01 3.49632874e-02 4.23899531e-01
5.09404898e-01 4.22213614e-01 -4.38082099e-01 2.04718128e-01
-1.23860025e+00 3.06728095e-01 6.11716628e-01 8.91206503e-01
8.50299537e-01 3.66194218e-01 -4.09329951e-01 8.02687883e-01
1.67285606e-01 3.37893367e-01 3.23417187e-02 -1.32894957e+00
2.45670721e-01 1.34320021e-01 3.24409045e-02 -6.86204374e-01
-2.15161189e-01 -4.29839194e-01 -1.12443781e+00 8.24190021e-01
2.36160576e-01 -3.19927871e-01 -1.31094742e+00 2.08139396e+00
3.56366098e-01 3.65598738e-01 -2.32978269e-01 8.41147780e-01
7.93387949e-01 5.40054560e-01 3.73858601e-01 -1.56880785e-02
1.52121484e+00 -8.48789096e-01 -6.76087618e-01 -2.60146856e-01
-3.91176432e-01 -8.42221975e-01 1.17432928e+00 3.44715863e-01
-1.62347710e+00 -1.02896988e+00 -7.39941955e-01 -2.41103470e-01
-2.32745960e-01 -3.43762571e-03 8.06780875e-01 5.26071012e-01
-1.62644982e+00 7.45313942e-01 -6.99937522e-01 -2.55873233e-01
9.50848937e-01 6.13261282e-01 -6.46432340e-01 -3.39315742e-01
-9.94952261e-01 6.95591509e-01 -1.21413760e-01 -3.56750228e-02
-1.31829894e+00 -1.21350467e+00 -8.57815087e-01 1.87809065e-01
5.23716807e-02 -1.15811431e+00 8.95525277e-01 -1.15982950e+00
-1.51147020e+00 6.99083865e-01 -8.01307783e-02 -1.65206373e-01
6.62228167e-01 -3.40717971e-01 -3.51650387e-01 3.93577874e-01
4.15701903e-02 1.20553696e+00 1.61852205e+00 -1.39806545e+00
-2.97793269e-01 -3.50599259e-01 3.23933631e-01 5.40929027e-02
-2.04615667e-01 7.64495432e-02 -4.38589275e-01 -9.69834268e-01
-2.31632307e-01 -7.44342506e-01 -1.74606308e-01 2.70485759e-01
-1.10936515e-01 1.16781116e-01 7.74314761e-01 -8.37541878e-01
8.32107782e-01 -1.92200840e+00 2.78713793e-01 -8.50697458e-02
4.84608054e-01 8.74082372e-02 -4.20940101e-01 1.94080696e-01
-3.08314264e-01 2.77999323e-02 -3.69465023e-01 -7.21937537e-01
-1.69204339e-01 9.49495584e-02 -1.60651743e-01 2.38849148e-01
6.86491907e-01 9.19950783e-01 -7.87035286e-01 -2.81034142e-01
1.17610127e-01 1.17746723e+00 -1.08184421e+00 4.95380253e-01
-2.14956388e-01 6.29713595e-01 -9.29197483e-03 7.57099748e-01
1.06495118e+00 -2.18239367e-01 -1.80429280e-01 -3.86364937e-01
1.48804292e-01 -1.24099813e-01 -9.25349772e-01 1.95662129e+00
-7.48788536e-01 5.12539804e-01 4.77719009e-01 -2.86072940e-01
5.31050980e-01 4.91091460e-01 1.49582282e-01 -4.10403401e-01
-2.19517305e-01 -2.39834055e-01 -1.93256959e-01 -2.81094879e-01
5.34595370e-01 -2.35664159e-01 2.25958347e-01 1.27445295e-01
3.72222304e-01 -2.03976527e-01 -9.37191322e-02 2.74218708e-01
1.19707596e+00 2.46233761e-01 -1.78865179e-01 -2.48253718e-01
3.13817471e-01 -5.53340912e-01 2.69529998e-01 5.01671135e-01
9.35190469e-02 9.05247629e-01 3.65620553e-01 -5.67429900e-01
-1.25737262e+00 -1.31169653e+00 -8.00491963e-03 1.26972628e+00
-5.10984361e-01 -2.19297558e-01 -1.12184608e+00 -4.00556564e-01
-6.35183156e-02 3.11157644e-01 -1.22321141e+00 7.65322372e-02
-6.67029738e-01 -3.56736481e-01 4.05861795e-01 6.03780091e-01
6.22486234e-01 -1.58334494e+00 -3.26479256e-01 1.54887304e-01
1.35495856e-01 -1.11228800e+00 -6.19046926e-01 -1.59585774e-01
-8.01971138e-01 -6.97365642e-01 -8.99906218e-01 -8.35492134e-01
1.07600093e+00 -2.11462274e-01 1.52167749e+00 9.03052911e-02
-7.00104117e-01 4.79047269e-01 1.01638570e-01 -7.53141865e-02
-2.38856152e-01 -7.58804604e-02 4.61693555e-02 7.55620152e-02
-1.01834998e-01 -9.76113677e-01 -9.62034225e-01 -6.63745254e-02
-1.04402864e+00 5.60372621e-02 8.22301269e-01 8.37144852e-01
5.17848432e-01 -8.85942802e-02 5.37557900e-01 -1.08676004e+00
6.06409013e-01 -3.11306536e-01 -3.67159635e-01 -9.40304548e-02
-2.47692212e-01 4.50239740e-02 7.25998223e-01 -6.00208998e-01
-1.19693804e+00 2.09815845e-01 -3.48220944e-01 -8.13175499e-01
-2.40376219e-01 -8.46198797e-02 -3.03181678e-01 -2.81854212e-01
7.61043489e-01 -5.87588735e-02 -1.56497769e-02 -3.89347494e-01
5.85585177e-01 -4.64070700e-02 9.34091568e-01 -7.02921987e-01
1.09345329e+00 5.43690324e-01 -3.72008421e-02 -7.24588692e-01
-6.13861322e-01 2.94618785e-01 -5.44102013e-01 -2.96303779e-01
8.96679640e-01 -1.18282700e+00 -7.58766890e-01 4.10402238e-01
-1.04173863e+00 -5.90821147e-01 -5.57870746e-01 -4.89026941e-02
-5.40765226e-01 -8.04142430e-02 -9.62805212e-01 -4.69821781e-01
-5.63357294e-01 -8.11542392e-01 1.22210264e+00 3.54463428e-01
-3.02779585e-01 -8.70262682e-01 -6.59279451e-02 2.27933705e-01
8.80618036e-01 4.43009228e-01 5.80595613e-01 1.21695608e-01
-7.18345940e-01 9.12089050e-02 -2.34891310e-01 2.45316982e-01
-4.54963967e-02 -1.39742643e-01 -1.33666611e+00 -7.02175021e-01
3.34443897e-02 -5.04775584e-01 9.08623338e-01 4.38090116e-01
1.31180513e+00 -6.48419619e-01 4.27504592e-02 8.75370085e-01
1.53013742e+00 -4.38588887e-01 1.14614081e+00 -1.59480751e-01
8.39835823e-01 5.12673259e-01 1.42329633e-01 4.44394737e-01
1.73486605e-01 4.63921726e-01 5.88478982e-01 -4.34013873e-01
-5.83101153e-01 -3.71388406e-01 1.88798979e-01 1.61793873e-01
-2.69358575e-01 1.38985664e-01 -4.60696071e-01 4.29030299e-01
-1.30105472e+00 -1.21283019e+00 3.57896179e-01 1.94732022e+00
9.33288157e-01 5.42051010e-02 -3.22320685e-02 -1.06918298e-01
7.43878663e-01 -3.43432557e-03 -4.79807705e-01 -4.81227130e-01
1.91676795e-01 8.35391939e-01 2.27686968e-02 8.66508365e-01
-8.37262392e-01 1.15303040e+00 6.40183163e+00 8.63281965e-01
-1.06331122e+00 1.59069031e-01 9.28229451e-01 -2.61583805e-01
-3.09317172e-01 -2.67463326e-01 -6.10414505e-01 5.23647189e-01
7.83420265e-01 5.48660278e-01 8.64335895e-01 7.62206435e-01
-1.62206084e-01 1.45694971e-01 -1.25812447e+00 9.60660517e-01
6.62285183e-03 -1.32990813e+00 3.79993707e-01 5.38829938e-02
9.26970184e-01 -2.26290330e-01 4.65477765e-01 3.95244390e-01
3.86582226e-01 -1.67245495e+00 5.38189948e-01 7.83349335e-01
1.61044264e+00 -8.51045549e-01 1.63131446e-01 -1.03419669e-01
-1.28177714e+00 -6.15478493e-03 -5.14125288e-01 -3.11140832e-03
-9.03657544e-03 3.65961820e-01 -5.63908041e-01 9.08211768e-02
8.18455338e-01 1.14658035e-01 -5.40587723e-01 3.97656292e-01
-1.96841270e-01 2.21369818e-01 -1.33959875e-01 6.85784101e-01
-1.06315017e-01 7.02763945e-02 2.80538917e-01 1.25323081e+00
3.08719814e-01 3.31016392e-01 1.17633743e-02 1.06196606e+00
-4.64588374e-01 -4.95998800e-01 -3.77311260e-01 3.05521309e-01
5.29268622e-01 1.51670206e+00 -4.37427491e-01 -3.26448768e-01
-2.15417564e-01 1.40982831e+00 6.67092085e-01 3.54885161e-01
-7.67916024e-01 -2.82656431e-01 7.13462234e-01 4.91259992e-01
5.81511736e-01 9.76273194e-02 -2.56555736e-01 -7.30224669e-01
-2.69250553e-02 -9.82230127e-01 5.53520098e-02 -9.21774507e-01
-1.42141914e+00 1.06524241e+00 -3.55250657e-01 -7.81715512e-01
-3.26175630e-01 -2.94329166e-01 -6.65928543e-01 1.21903539e+00
-1.37514746e+00 -1.55839455e+00 -7.02567399e-01 9.61868882e-01
6.61855280e-01 -1.61088005e-01 1.16785288e+00 4.02544051e-01
-2.32400090e-01 9.56151247e-01 -5.56543112e-01 1.95275486e-01
6.84550881e-01 -1.16602576e+00 6.79281235e-01 7.42754817e-01
-1.32688448e-01 6.57517314e-01 7.62662590e-01 -5.60449481e-01
-1.28063715e+00 -1.34192359e+00 7.09701836e-01 -5.44729412e-01
9.28500071e-02 -4.91142780e-01 -9.71415281e-01 5.62344193e-01
4.56507236e-01 4.31483209e-01 3.38576257e-01 3.33755575e-02
-6.85993552e-01 -2.30921164e-01 -1.73925710e+00 5.23307681e-01
1.06901741e+00 -6.79896116e-01 -2.61569321e-01 1.23632059e-01
7.26077318e-01 -3.76865417e-01 -9.37530935e-01 3.94095480e-01
5.24567604e-01 -1.07384062e+00 1.15282798e+00 -4.10962850e-01
7.65232742e-01 -2.06411839e-01 5.95976450e-02 -1.27811730e+00
-7.66480207e-01 -9.38941956e-01 -4.22889024e-01 1.16820097e+00
1.59993753e-01 -3.26241910e-01 8.39734077e-01 5.43306649e-01
-6.08511604e-02 -7.62397885e-01 -8.95372450e-01 -2.30890363e-01
-5.02896309e-02 2.21290607e-02 6.21269345e-01 9.06291842e-01
-5.61166227e-01 3.83827001e-01 -5.89138567e-01 3.36815476e-01
7.73107827e-01 -1.70801952e-01 7.43702054e-01 -1.03025460e+00
-4.89511847e-01 -2.04244405e-01 -2.92141974e-01 -7.63835490e-01
-2.21069939e-02 -5.21652639e-01 -1.11586541e-01 -1.24949574e+00
3.56608689e-01 -2.36404523e-01 9.60979462e-02 4.44665104e-01
-2.34625340e-01 9.05547798e-01 1.38862789e-01 -3.08173019e-02
-2.78415531e-01 4.27471966e-01 1.40059900e+00 -2.88345013e-02
-4.28245924e-02 -2.84512460e-01 -7.66523361e-01 6.30686700e-01
7.17026949e-01 -2.12060750e-01 -4.42441314e-01 -6.37505651e-01
2.27551046e-03 1.86914071e-01 5.25738895e-01 -1.34774578e+00
2.96093412e-02 1.60041004e-01 1.17466319e+00 -2.61017829e-01
8.34871650e-01 -7.33619809e-01 4.56476152e-01 4.37319309e-01
-3.32161307e-01 1.54568568e-01 4.32090163e-01 4.21356380e-01
1.72380760e-01 2.90342540e-01 1.18016696e+00 -3.75347793e-01
-3.74147862e-01 7.39407897e-01 -7.77800083e-02 7.70717040e-02
7.91404009e-01 -1.84030950e-01 -8.95760059e-02 -5.23666322e-01
-1.14106238e+00 -1.20085195e-01 6.48565829e-01 3.77385646e-01
7.78500676e-01 -1.37951899e+00 -1.11870718e+00 5.81032217e-01
-3.72835308e-01 -1.55209808e-03 5.86498678e-01 1.45532221e-01
-5.44871211e-01 1.43603487e-02 -5.98169327e-01 -4.47846472e-01
-1.19430912e+00 6.69922590e-01 3.25510502e-01 -3.65342885e-01
-6.34389758e-01 1.22543252e+00 4.22656834e-01 -1.57215536e-01
2.13137537e-01 1.90036803e-01 -1.13391489e-01 -1.09240159e-01
7.09213793e-01 2.37351418e-01 -6.17824048e-02 -5.11194944e-01
-1.09640680e-01 6.54332995e-01 -2.21601382e-01 -1.70043543e-01
1.38628507e+00 7.17914402e-02 -1.49929494e-01 -3.27167332e-01
1.04991674e+00 9.97596681e-02 -1.98341763e+00 9.94318910e-03
-6.96645737e-01 -5.78987896e-01 -2.20617741e-01 -1.00685418e+00
-1.37808287e+00 7.89050758e-01 8.01728904e-01 -8.79257396e-02
1.32583892e+00 -6.67815208e-02 5.08336008e-01 -8.25443938e-02
3.03600103e-01 -8.70381296e-01 4.07094359e-01 3.79638135e-01
1.34507549e+00 -1.07595181e+00 4.71283123e-02 -4.08683330e-01
-3.59755725e-01 7.31500268e-01 9.03372645e-01 -5.31735659e-01
6.29566908e-01 5.90867937e-01 -2.41214141e-01 -2.77565479e-01
-8.61734271e-01 4.45572324e-02 2.33716264e-01 9.47488487e-01
5.10960519e-01 2.79353969e-02 3.20208669e-01 2.12670952e-01
-3.32484394e-01 -4.71259691e-02 2.07404763e-01 5.65726340e-01
-1.96250737e-01 -8.83748412e-01 -5.10597348e-01 3.45818996e-01
-6.87973380e-01 -4.02799070e-01 -5.28261391e-03 6.56957328e-01
3.04489374e-01 7.72137761e-01 3.15154254e-01 -1.36901915e-01
1.70636386e-01 -7.34642819e-02 7.92064965e-01 -5.97607017e-01
-6.93224370e-01 4.36466858e-02 -1.65016711e-01 -7.88437605e-01
-1.21968970e-01 -3.86217088e-01 -9.93081987e-01 -6.27828479e-01
2.86699623e-01 -2.15735659e-01 4.24352974e-01 4.31155652e-01
6.50577843e-01 9.93980050e-01 6.94305301e-01 -1.44992805e+00
-3.26345414e-01 -1.34096777e+00 -1.94943398e-01 5.19572139e-01
5.18097579e-01 -3.63431513e-01 -1.37636259e-01 2.86352009e-01] | [12.637664794921875, -0.2049209177494049] |
e027c5f7-0a30-4fdd-b40c-acccb71188d7 | cross-domain-iterative-network-for | 2305.10326 | null | https://arxiv.org/abs/2305.10326v1 | https://arxiv.org/pdf/2305.10326v1.pdf | Cross-domain Iterative Network for Simultaneous Denoising, Limited-angle Reconstruction, and Attenuation Correction of Low-dose Cardiac SPECT | Single-Photon Emission Computed Tomography (SPECT) is widely applied for the diagnosis of ischemic heart diseases. Low-dose (LD) SPECT aims to minimize radiation exposure but leads to increased image noise. Limited-angle (LA) SPECT enables faster scanning and reduced hardware costs but results in lower reconstruction accuracy. Additionally, computed tomography (CT)-derived attenuation maps ($\mu$-maps) are commonly used for SPECT attenuation correction (AC), but it will cause extra radiation exposure and SPECT-CT misalignments. In addition, the majority of SPECT scanners in the market are not hybrid SPECT/CT scanners. Although various deep learning methods have been introduced to separately address these limitations, the solution for simultaneously addressing these challenges still remains highly under-explored and challenging. To this end, we propose a Cross-domain Iterative Network (CDI-Net) for simultaneous denoising, LA reconstruction, and CT-free AC in cardiac SPECT. In CDI-Net, paired projection- and image-domain networks are end-to-end connected to fuse the emission and anatomical information across domains and iterations. Adaptive Weight Recalibrators (AWR) adjust the multi-channel input features to enhance prediction accuracy. Our experiments using clinical data showed that CDI-Net produced more accurate $\mu$-maps, projections, and reconstructions compared to existing approaches that addressed each task separately. Ablation studies demonstrated the significance of cross-domain and cross-iteration connections, as well as AWR, in improving the reconstruction performance. | ['Chi Liu', 'Albert J. Sinusas', 'Qiong Liu', 'Xueqi Guo', 'Huidong Xie', 'Bo Zhou', 'Xiongchao Chen'] | 2023-05-17 | null | null | null | null | ['computed-tomography-ct'] | ['methodology'] | [ 1.91677049e-01 -3.03230226e-01 -9.19456258e-02 -4.88952577e-01
-1.05574369e+00 -2.62233198e-01 -4.32832837e-02 -1.29074305e-01
-5.35415828e-01 7.27676988e-01 1.57303348e-01 -3.51277351e-01
-7.51203373e-02 -7.11839139e-01 -3.84735912e-01 -6.50477946e-01
1.59915596e-01 5.29316604e-01 3.99900794e-01 2.78336257e-01
-3.35704535e-01 5.60645640e-01 -4.76843894e-01 4.56282973e-01
7.39744902e-01 8.23446512e-01 7.07828581e-01 5.58073401e-01
1.14969946e-02 7.06284046e-01 -5.64365499e-02 -1.82057142e-01
2.42905214e-01 -8.23141634e-01 -5.95605850e-01 -2.25786179e-01
5.60957566e-02 -6.46146715e-01 -5.51595986e-01 7.55719841e-01
1.05835462e+00 -2.29915589e-01 4.74714786e-01 -6.62614644e-01
-2.39661872e-01 6.70205653e-01 -6.71905816e-01 7.64491498e-01
-1.95471436e-01 3.60209614e-01 2.61433691e-01 -9.60314453e-01
4.62087274e-01 4.80271637e-01 1.02839887e+00 5.08174241e-01
-1.32587528e+00 -8.22071493e-01 -2.19974861e-01 1.68036625e-01
-1.20006883e+00 2.60371193e-02 7.11578071e-01 -3.63792062e-01
8.90439689e-01 1.83992431e-01 1.04193187e+00 9.18457508e-01
5.31818092e-01 3.60842019e-01 1.01048636e+00 -1.34712547e-01
-1.87024400e-01 -6.13673627e-02 -1.22449972e-01 5.73736250e-01
1.46158472e-01 9.59354565e-02 -1.59924701e-01 -7.35340789e-02
1.24821913e+00 2.77844042e-01 -7.36084461e-01 -3.45620394e-01
-1.34082699e+00 7.31956482e-01 7.92266250e-01 3.79349887e-01
-5.07815719e-01 1.38327733e-01 6.51268423e-01 -2.70221457e-02
3.58643681e-01 2.53159910e-01 -3.78731132e-01 -1.84423123e-02
-9.37133670e-01 -3.11109964e-02 2.02598497e-01 4.84739512e-01
1.74731642e-01 2.46872649e-01 -2.51432151e-01 1.02628469e+00
3.02637488e-01 3.76710832e-01 6.79021657e-01 -8.81183267e-01
5.23779154e-01 5.06620526e-01 -3.44936132e-01 -5.43526173e-01
-9.04904962e-01 -8.11769187e-01 -1.33461058e+00 1.08837105e-01
4.65907305e-01 -2.85241216e-01 -1.03351367e+00 1.73221612e+00
1.65675268e-01 3.05540681e-01 -3.96464497e-01 1.43021381e+00
8.48800063e-01 4.48249668e-01 4.01841313e-01 -4.99312848e-01
1.32952034e+00 -7.74323404e-01 -6.13237798e-01 -2.01402411e-01
7.93222189e-01 -7.38634408e-01 1.13390434e+00 2.14256778e-01
-1.29521823e+00 -2.82901973e-01 -1.01841736e+00 9.56844762e-02
4.29018795e-01 1.13509648e-01 3.80634516e-01 8.12492013e-01
-7.55819321e-01 5.20124078e-01 -1.22234714e+00 1.66000485e-01
7.80378520e-01 4.33590621e-01 -7.39537477e-02 -2.13083744e-01
-1.12455559e+00 1.10234880e+00 1.40769975e-02 1.25170007e-01
-8.30556989e-01 -1.33067477e+00 -3.82409155e-01 -8.92995819e-02
2.72375047e-01 -1.12216592e+00 1.14434183e+00 -8.64055097e-01
-1.54770470e+00 5.67480385e-01 1.18453905e-01 -4.15682405e-01
7.18560636e-01 -2.76328065e-02 -3.50304902e-01 4.45615262e-01
1.41828179e-01 5.31162381e-01 4.95778978e-01 -9.07849491e-01
-1.32484928e-01 -5.83581328e-01 -6.29377902e-01 4.10024673e-01
-7.18132630e-02 1.04867302e-01 -3.95155460e-01 -6.93360269e-01
6.64136648e-01 -7.68327177e-01 -4.31398660e-01 3.04773152e-01
1.96282547e-02 4.72000927e-01 7.51624346e-01 -8.81321311e-01
1.14164317e+00 -1.92512476e+00 -6.23059943e-02 1.70194745e-01
4.80642438e-01 2.72099286e-01 1.20354488e-01 -3.55749846e-01
-3.65540981e-01 2.75676064e-02 -4.14885372e-01 2.89009735e-02
-7.29393363e-01 1.00613311e-01 1.22329317e-01 6.54685199e-01
-2.32845828e-01 1.05424047e+00 -5.65533698e-01 -5.32586098e-01
4.75259572e-01 6.45533800e-01 -5.95748127e-01 -6.03824966e-02
2.58136302e-01 1.13617396e+00 -3.97663027e-01 4.65275556e-01
7.39993513e-01 -4.06125724e-01 4.50244159e-01 -5.45538783e-01
-3.28347944e-02 1.40970856e-01 -8.78547549e-01 1.79722273e+00
-5.11977792e-01 1.84577689e-01 1.02390964e-02 -9.41012800e-01
5.61282694e-01 5.98774433e-01 9.78263676e-01 -1.23402166e+00
3.65052223e-01 4.55289185e-01 3.16385537e-01 -4.07120258e-01
-1.65139213e-01 -7.18455017e-01 3.57877374e-01 3.95433873e-01
-2.60222882e-01 -3.02724719e-01 -3.67401659e-01 6.07346855e-02
9.80337858e-01 -2.39798310e-03 1.03521332e-01 -1.23131119e-01
2.89988577e-01 -5.42737544e-02 7.90082276e-01 7.55514741e-01
-3.02277327e-01 1.27017236e+00 3.26281279e-01 -6.33623958e-01
-1.26719213e+00 -1.26898634e+00 -5.66628814e-01 4.37827975e-01
1.74320295e-01 -1.50625378e-01 -5.98261774e-01 -6.19460940e-01
-3.67373168e-01 5.05649567e-01 -1.06388070e-01 9.78652108e-03
-1.00759649e+00 -1.33906078e+00 6.99728131e-01 9.03605700e-01
4.63902116e-01 -7.67854750e-01 -8.59438896e-01 6.58511102e-01
-5.57711542e-01 -9.74409819e-01 -5.01014888e-01 3.25102180e-01
-1.46538556e+00 -1.03188765e+00 -9.83468711e-01 -4.67712760e-01
5.44570029e-01 1.24716103e-01 1.20215440e+00 -5.97051196e-02
-3.95923644e-01 4.01264168e-02 -1.38933033e-01 -3.17930490e-01
-2.16370553e-01 -8.98244679e-02 -1.96921960e-01 -4.62616056e-01
-1.71211094e-01 -8.69044542e-01 -1.09969461e+00 5.14447510e-01
-8.06558013e-01 2.70877987e-01 6.20780706e-01 1.13099861e+00
1.04716849e+00 -3.96139503e-01 6.21775985e-01 -9.87747610e-01
4.81372327e-01 -2.60170490e-01 -5.49593151e-01 2.32412349e-02
-7.06199288e-01 -3.89376074e-01 6.05975866e-01 -2.69781530e-01
-1.08422709e+00 1.40134513e-01 -4.70830262e-01 -3.74760807e-01
1.60877958e-01 5.94681919e-01 1.81712076e-01 -2.08476499e-01
8.39856446e-01 1.91189185e-01 1.23618446e-01 -1.55175745e-01
-1.44620851e-01 1.82641804e-01 3.61283064e-01 -2.05351427e-01
2.24569738e-01 4.90815997e-01 1.44189462e-01 -4.51122433e-01
-6.45798206e-01 -1.99900880e-01 -6.17650628e-01 -3.22148502e-01
1.00595665e+00 -1.06300092e+00 -4.52700287e-01 4.10675585e-01
-9.87621427e-01 -2.73943245e-01 -4.01264429e-01 1.11618757e+00
-2.99895108e-01 5.20948946e-01 -7.47369230e-01 -1.68491304e-01
-6.93365932e-01 -1.67977870e+00 4.58472967e-01 4.10527550e-02
-2.53583193e-01 -6.73577964e-01 9.95957032e-02 5.04735053e-01
5.88302493e-01 1.75760284e-01 1.13402748e+00 -1.95148498e-01
-6.05021536e-01 -4.08945605e-02 -5.15208185e-01 5.54284751e-01
-1.60282925e-01 -6.36492491e-01 -6.68833733e-01 -1.20346271e-01
3.95307869e-01 -2.45196521e-01 6.83152735e-01 1.21932304e+00
1.49295521e+00 3.31000000e-01 -7.72520900e-02 1.10371280e+00
1.37922812e+00 3.74414235e-01 7.74867058e-01 2.79129505e-01
7.90051043e-01 -1.12995185e-01 1.31582856e-01 4.48790520e-01
1.92327678e-01 7.00497925e-01 3.64661545e-01 -4.84011561e-01
-4.29187834e-01 9.15560201e-02 -1.66929841e-01 8.10734451e-01
-2.88385332e-01 2.50997953e-02 -1.03552306e+00 3.41128170e-01
-1.43860936e+00 -6.93263888e-01 -7.48657644e-01 1.99024665e+00
6.82477176e-01 4.21388773e-03 3.06974296e-02 -1.59139037e-01
4.66625959e-01 -1.43538803e-01 -7.21165657e-01 -3.91076095e-02
-5.45073301e-02 5.63502192e-01 6.60450399e-01 1.61619902e-01
-8.02849054e-01 1.30857199e-01 6.38399458e+00 5.80194652e-01
-1.42992187e+00 8.38514924e-01 9.16154146e-01 -4.04539108e-01
-1.88991055e-01 -9.07297283e-02 -3.36864084e-01 4.44220990e-01
7.02637732e-01 3.37354898e-01 6.01583943e-02 6.94263041e-01
4.73440349e-01 -4.17238414e-01 -7.53366530e-01 1.07818639e+00
-1.20018721e-01 -1.40312076e+00 -2.68768609e-01 -1.41256481e-01
6.75308526e-01 4.64865655e-01 -5.23715513e-03 1.08735144e-01
-2.61767566e-01 -9.18252230e-01 5.11280179e-01 3.15159202e-01
1.13437450e+00 -7.64300942e-01 9.77066815e-01 2.65085995e-01
-9.38195765e-01 -3.81426625e-02 -2.00023964e-01 1.89993978e-01
7.35905707e-01 8.18070173e-01 -8.00636411e-01 5.87971210e-01
8.34921896e-01 4.47498471e-01 -1.76907867e-01 1.02895558e+00
-5.45777641e-02 6.19805217e-01 -4.04021859e-01 4.88089234e-01
8.14202651e-02 -3.52313280e-01 4.60260749e-01 1.01541698e+00
4.47616667e-01 4.61866885e-01 1.01190455e-01 1.01506293e+00
-1.43712297e-01 6.42870888e-02 -1.03804566e-01 6.06044948e-01
-2.17673406e-02 1.06313932e+00 -9.55602348e-01 -3.73371065e-01
-5.66893339e-01 7.37415433e-01 -9.50488448e-03 2.74939090e-01
-1.13187873e+00 2.97864377e-01 1.04022831e-01 7.74691999e-01
-6.85036555e-02 -1.17920503e-01 -9.92867589e-01 -1.08483291e+00
-5.91729134e-02 -5.86759329e-01 4.71494615e-01 -7.58081794e-01
-1.20539355e+00 6.46170557e-01 -5.55321015e-03 -1.23795164e+00
1.50189862e-01 -2.43786424e-01 -5.19095182e-01 1.14709914e+00
-1.63055182e+00 -8.71774077e-01 -4.41745609e-01 6.26959741e-01
5.01746356e-01 1.31203786e-01 7.17709363e-01 8.26030493e-01
-4.98974591e-01 4.04997975e-01 7.14414492e-02 -3.44080999e-02
6.66583359e-01 -9.73009765e-01 -1.63986877e-01 7.11692929e-01
-2.81711102e-01 3.42339039e-01 2.50541836e-01 -7.96387970e-01
-9.77319479e-01 -1.10946143e+00 3.35102975e-01 -2.54530758e-01
2.83140957e-01 2.26506412e-01 -8.64764512e-01 5.75541317e-01
-6.68378845e-02 5.87890506e-01 5.38405538e-01 -4.03085083e-01
1.62477478e-01 -2.11607605e-01 -1.22023249e+00 3.17400455e-01
7.44316697e-01 -2.51607090e-01 -2.12645277e-01 3.54052007e-01
2.16432542e-01 -8.94482315e-01 -9.17110682e-01 6.16856039e-01
6.11311078e-01 -1.11879504e+00 1.27552140e+00 -4.77548921e-03
4.97898638e-01 -1.37469858e-01 1.60572797e-01 -1.10710800e+00
-4.20912653e-01 1.21288225e-01 2.73505211e-01 4.45627689e-01
4.81528878e-01 -6.21195972e-01 8.67435217e-01 5.44899702e-01
-6.64808512e-01 -9.58482206e-01 -1.20434427e+00 -3.31805915e-01
1.66928500e-01 -7.78178453e-01 8.44158605e-02 8.40903759e-01
-2.79882371e-01 2.13521391e-01 -4.16219234e-01 1.30031064e-01
5.32526374e-01 -2.17171058e-01 1.16829164e-01 -9.30926621e-01
-4.03375834e-01 -3.31998497e-01 -8.28260332e-02 -8.60764563e-01
-3.73443514e-01 -1.22985780e+00 -3.03732485e-01 -1.57576740e+00
6.03202641e-01 -5.61669648e-01 -4.12269503e-01 1.96663052e-01
-1.09417260e-01 7.11107969e-01 1.05144575e-01 3.07255447e-01
-1.22129016e-01 3.47373664e-01 1.82031369e+00 1.64880455e-01
-3.34087282e-01 6.54913560e-02 -4.15141165e-01 9.50490355e-01
7.60894477e-01 -8.64291728e-01 -4.94162530e-01 -6.59096181e-01
1.11622423e-01 6.11850083e-01 6.94562852e-01 -9.28248048e-01
1.34826243e-01 2.54696399e-01 9.55744088e-01 -7.63856590e-01
2.65202075e-01 -9.63889599e-01 4.16617364e-01 6.42439306e-01
-5.35499528e-02 1.49399772e-01 1.06439427e-01 2.74278432e-01
-2.52786428e-01 -1.38698280e-01 1.34219921e+00 -4.64519560e-01
-9.80467126e-02 4.63466704e-01 -5.86075723e-01 1.67169441e-02
8.77566636e-01 -3.96044999e-01 1.34571999e-01 -2.32634649e-01
-9.52609658e-01 -7.08756745e-02 -5.28911725e-02 -2.70772904e-01
9.32120562e-01 -1.12805736e+00 -6.62050068e-01 1.53313279e-01
-3.01386029e-01 2.26637259e-01 9.71114516e-01 1.78880882e+00
-9.41430211e-01 2.02117175e-01 -2.57708728e-01 -9.32907104e-01
-1.06553137e+00 3.18083586e-03 7.77844489e-01 -6.91532791e-01
-1.20621228e+00 8.10727656e-01 2.97614098e-01 -3.70202422e-01
-1.59124628e-01 -3.95758837e-01 2.22240195e-01 -2.31410876e-01
2.50465393e-01 3.29939038e-01 3.03110689e-01 -3.89656633e-01
-2.95114130e-01 5.12538493e-01 -1.01823315e-01 -6.76424205e-02
1.53066611e+00 -2.20794573e-01 1.11596987e-01 -2.54726168e-02
9.40347970e-01 -5.84832907e-01 -1.25164115e+00 -1.75365344e-01
-5.14970779e-01 -3.74631464e-01 5.96706986e-01 -1.12615371e+00
-1.44402742e+00 1.00152826e+00 1.15357673e+00 -5.10383964e-01
1.33110189e+00 -2.95100719e-01 1.04572713e+00 -1.60906285e-01
9.32036713e-02 -8.83677185e-01 1.96393624e-01 1.53965458e-01
6.43799603e-01 -1.15543675e+00 2.49992996e-01 -4.94485527e-01
-9.48683977e-01 1.20820308e+00 6.59583151e-01 2.58818325e-02
6.20500505e-01 6.02707148e-01 2.80632526e-01 -2.48613834e-01
-1.56726807e-01 4.67145711e-01 2.50158459e-03 3.86395633e-01
6.65225565e-01 -1.61492433e-02 -5.65110505e-01 7.55260050e-01
2.38875717e-01 2.33242065e-01 4.01650518e-01 8.89444351e-01
-1.60427451e-01 -1.01241469e+00 -5.13338745e-01 7.90956795e-01
-6.85088754e-01 -2.16412231e-01 3.67481083e-01 6.30203724e-01
2.48627797e-01 3.52512270e-01 -2.59927869e-01 7.22058266e-02
4.83679920e-01 6.99021593e-02 5.34791172e-01 -3.81567240e-01
-9.57929909e-01 5.63674092e-01 -2.33867079e-01 -5.02303958e-01
-3.63165915e-01 -5.55495739e-01 -1.39984918e+00 -8.11410993e-02
-3.13910782e-01 -3.50945324e-01 7.17141271e-01 8.75576854e-01
1.03556991e-01 1.09715235e+00 3.83931190e-01 -5.27493060e-01
-3.50595891e-01 -8.68172705e-01 -6.45182967e-01 1.67486712e-01
-1.20869689e-01 -4.55265492e-01 -1.95237361e-02 1.64591067e-03] | [13.504866600036621, -2.524232864379883] |
66396cb8-3256-4570-8b4d-dd061cb6f4a5 | variable-length-hashing | 1603.05414 | null | http://arxiv.org/abs/1603.05414v1 | http://arxiv.org/pdf/1603.05414v1.pdf | Variable-Length Hashing | Hashing has emerged as a popular technique for large-scale similarity search.
Most learning-based hashing methods generate compact yet correlated hash codes.
However, this redundancy is storage-inefficient. Hence we propose a lossless
variable-length hashing (VLH) method that is both storage- and
search-efficient. Storage efficiency is achieved by converting the fixed-length
hash code into a variable-length code. Search efficiency is obtained by using a
multiple hash table structure. With VLH, we are able to deliberately add
redundancy into hash codes to improve retrieval performance with little
sacrifice in storage efficiency or search complexity. In particular, we propose
a block K-means hashing (B-KMH) method to obtain significantly improved
retrieval performance with no increase in storage and marginal increase in
computational cost. | ['XiaoLi Li', 'Hong Wei Ng', 'Pierre Moulin', 'Honghai Yu'] | 2016-03-17 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-1.32472277e-01 -3.12438667e-01 -4.06454116e-01 -6.40936475e-03
-1.32741153e+00 -4.90932286e-01 3.91485900e-01 7.10171223e-01
-6.36484087e-01 7.21850336e-01 1.64123047e-02 -1.21629670e-01
-2.02718481e-01 -1.11853182e+00 -3.77876788e-01 -7.82937229e-01
-2.01963797e-01 2.26996675e-01 5.94234824e-01 -1.98962819e-02
5.39983094e-01 5.43117046e-01 -1.77410376e+00 1.40313888e-02
4.52201128e-01 9.60399985e-01 3.52381289e-01 4.41387534e-01
4.78342324e-02 3.57667893e-01 -3.87005597e-01 -3.89287859e-01
4.59483445e-01 -3.64120126e-01 -7.43031263e-01 -6.74553037e-01
1.34702146e-01 -2.94177741e-01 -7.48475254e-01 6.60860062e-01
7.98558533e-01 1.75689980e-01 4.94234920e-01 -1.12212348e+00
-4.39745396e-01 5.80778658e-01 -4.62101400e-01 3.30875367e-02
4.56359297e-01 -3.19210649e-01 1.34484315e+00 -9.68297780e-01
4.34475213e-01 1.03680861e+00 7.90944517e-01 1.61450237e-01
-1.19581449e+00 -5.38480341e-01 -8.33487630e-01 5.45926154e-01
-2.28268290e+00 -3.11787158e-01 5.85896552e-01 1.66613907e-01
9.79425132e-01 4.50482219e-01 7.61665523e-01 2.25762665e-01
2.41482139e-01 6.81197822e-01 7.09963083e-01 -6.28778100e-01
3.88169676e-01 -3.48592103e-02 -9.54556614e-02 8.15583110e-01
5.15492916e-01 1.04262874e-01 -6.43268883e-01 -4.04269427e-01
4.24304128e-01 2.47922257e-01 -5.22754975e-02 -5.34717798e-01
-1.10122907e+00 1.23301959e+00 6.35936022e-01 2.30321914e-01
-3.03811550e-01 4.55392212e-01 5.12046874e-01 4.12796259e-01
-1.95973486e-01 3.27970266e-01 8.04528520e-02 -1.24743611e-01
-1.31024206e+00 4.50087100e-01 5.90942144e-01 9.28390980e-01
1.11208105e+00 -3.13152581e-01 -3.55834395e-01 8.08468878e-01
1.58021644e-01 7.96647727e-01 7.43136764e-01 -7.93023527e-01
7.08357170e-02 3.20189267e-01 -6.78808019e-02 -1.20713091e+00
-2.53175735e-01 -3.14478837e-02 -7.31680989e-01 -2.72719771e-01
-1.03569224e-01 5.00675023e-01 -5.22456706e-01 1.66798377e+00
3.68224293e-01 -2.37160429e-01 4.33517387e-03 5.98608315e-01
4.93302196e-01 8.38855267e-01 -2.42243886e-01 -4.62599456e-01
1.38634253e+00 -9.66373920e-01 -7.14500248e-01 2.05048501e-01
1.12841821e+00 -9.10174072e-01 8.03681672e-01 -4.82695214e-02
-1.39362395e+00 -3.27842057e-01 -1.23735082e+00 -3.17270398e-01
-4.07742232e-01 -5.07421076e-01 5.81641614e-01 7.24641025e-01
-1.17495012e+00 3.70969713e-01 -6.44502640e-01 -6.99470341e-02
1.54531915e-02 4.91401911e-01 -4.84675705e-01 -2.83034831e-01
-1.33695173e+00 8.90872121e-01 5.64128160e-01 -5.44425726e-01
-3.30413848e-01 -5.56243062e-01 -9.64550197e-01 3.60351086e-01
-3.31824534e-02 -3.28239977e-01 1.06283128e+00 2.79769599e-01
-9.48584378e-01 5.64354539e-01 -3.99291992e-01 -6.44420147e-01
-2.52983849e-02 1.10570341e-01 -1.37818128e-01 6.18543267e-01
-7.97887146e-02 8.30489516e-01 7.47195721e-01 -7.93316782e-01
-4.48532432e-01 -1.17472976e-01 -4.28374708e-01 6.06740862e-02
-8.00185859e-01 -4.71882597e-02 -4.10267681e-01 -7.27896452e-01
2.85028398e-01 -1.21816361e+00 8.66888557e-03 3.11294466e-01
1.60149381e-01 -1.37369186e-01 9.09925163e-01 -3.62296283e-01
1.82118845e+00 -2.21931386e+00 -1.36622116e-01 3.70304525e-01
3.24049890e-02 4.82395530e-01 -2.50007331e-01 9.93760228e-01
4.45982456e-01 -1.39895752e-01 -9.80458502e-03 -2.82296568e-01
1.88076235e-02 2.79895842e-01 -3.37052792e-01 5.26381314e-01
-2.72261709e-01 1.00138998e+00 -7.10199714e-01 -9.90483522e-01
6.38132170e-02 6.86801255e-01 -7.97248960e-01 1.99675962e-01
4.29229826e-01 -5.42026937e-01 -1.73807874e-01 7.59709537e-01
6.19567394e-01 -3.48101258e-01 1.24901205e-01 -2.29032576e-01
-2.12111231e-02 4.54956621e-01 -1.09919631e+00 1.53169560e+00
-5.26430368e-01 3.60828131e-01 -2.13926107e-01 -7.68695354e-01
1.16671133e+00 2.61923134e-01 3.85348320e-01 -1.04512107e+00
-1.87556148e-01 4.56250310e-01 -4.93757278e-01 3.33220675e-03
8.74811709e-01 -1.85738489e-01 -1.77231655e-01 6.16572440e-01
-3.64298970e-01 -1.26292214e-01 1.12518080e-01 1.98596373e-01
1.29025888e+00 -8.09395373e-01 4.77336049e-01 -1.69321999e-01
6.23946428e-01 -1.20625213e-01 2.74283558e-01 7.72664487e-01
-1.53613701e-01 6.51526272e-01 -2.68949896e-01 -3.16343814e-01
-1.34941149e+00 -1.07546461e+00 -3.33063573e-01 9.29144681e-01
3.80954146e-01 -1.02120268e+00 -4.20229465e-01 -7.23990500e-02
2.96597749e-01 7.96008781e-02 -2.10060403e-01 -5.25506079e-01
-7.65554249e-01 -3.67286623e-01 7.07011104e-01 4.25718009e-01
3.84748042e-01 -1.01869130e+00 -7.86891222e-01 2.79568166e-01
-2.35491440e-01 -7.46838093e-01 -7.59118736e-01 4.24480796e-01
-8.19801927e-01 -5.14770865e-01 -7.03576565e-01 -1.04356456e+00
3.29129189e-01 7.61475503e-01 6.97307587e-01 4.49162513e-01
-6.68157339e-01 1.49634913e-01 -6.70397401e-01 2.34669194e-01
-2.19362602e-01 3.53162766e-01 8.47521871e-02 -6.51204407e-01
3.90767604e-01 -4.82492596e-01 -9.78933454e-01 3.98507804e-01
-1.13599050e+00 -4.93728936e-01 7.08716333e-01 1.06762314e+00
6.34124577e-01 1.76307082e-01 4.19116527e-01 -3.02630737e-02
4.96218354e-01 -2.28664756e-01 -4.78893936e-01 2.47817039e-01
-1.13021839e+00 4.99049932e-01 6.78976655e-01 -4.22139108e-01
-1.68059945e-01 9.36735719e-02 -1.04661547e-01 -3.58368933e-01
5.03621638e-01 5.79287827e-01 2.89204955e-01 -5.37437797e-01
4.77145076e-01 9.03539300e-01 2.48770654e-01 -3.35826457e-01
3.44158053e-01 8.25096190e-01 3.80047411e-01 -3.17760438e-01
9.38749015e-01 3.96297216e-01 4.49832946e-01 -6.20950282e-01
-3.19780797e-01 -7.26298869e-01 -5.20626962e-01 2.84586459e-01
3.65110487e-01 -9.97133076e-01 -8.06607246e-01 6.62033781e-02
-6.16389990e-01 5.01298718e-02 -1.43228039e-01 3.82829309e-01
-5.70406139e-01 7.39047348e-01 -7.61642456e-01 -6.08938456e-01
-6.81418657e-01 -9.37161207e-01 1.10778391e+00 -9.15427729e-02
-2.88949251e-01 -6.28779888e-01 2.77428716e-01 1.54407755e-01
5.86662650e-01 -3.14586908e-01 9.29634631e-01 -4.82898921e-01
-7.49457896e-01 -5.76519310e-01 -3.64733696e-01 -2.17257082e-01
4.71061952e-02 -4.55642521e-01 -4.69084889e-01 -8.62003922e-01
-4.69316393e-01 -6.18874550e-01 7.40242839e-01 -5.75997978e-02
1.05302000e+00 -5.11899650e-01 -2.95814067e-01 4.99853760e-01
1.58676243e+00 2.55721867e-01 7.70471275e-01 5.57660639e-01
2.67841041e-01 1.90140978e-01 8.28296959e-01 8.84370267e-01
3.26438844e-01 1.02803993e+00 -4.61658500e-02 1.46616608e-01
-1.86449751e-01 -5.13685107e-01 1.66418210e-01 1.14059758e+00
5.81367731e-01 -1.46545202e-01 -7.93045163e-01 8.95527363e-01
-1.72791946e+00 -1.14682281e+00 1.50397941e-01 2.45755816e+00
1.23511410e+00 -1.47251531e-01 2.07313702e-01 7.30367243e-01
6.80442095e-01 1.79362476e-01 -2.19419703e-01 -5.19939125e-01
-1.12214200e-01 4.02022034e-01 5.46689987e-01 3.81763041e-01
-8.33009839e-01 7.52676129e-01 7.37507820e+00 1.34666443e+00
-9.79954779e-01 -1.30089551e-01 -7.08007738e-02 -2.24651411e-01
-3.65697592e-01 2.34037079e-02 -1.11558461e+00 5.80902696e-01
1.17952240e+00 -4.71617639e-01 3.62978131e-01 1.04558420e+00
-3.48303884e-01 7.52700819e-03 -8.39846075e-01 1.35727394e+00
9.88047943e-02 -1.49239290e+00 4.60880011e-01 2.10215494e-01
3.28706890e-01 -4.42618698e-01 1.41856402e-01 1.63741425e-01
-2.68654913e-01 -8.06586683e-01 5.43670952e-01 -1.67171910e-01
9.76706266e-01 -1.07555485e+00 7.00153708e-01 -1.17070843e-02
-1.68927884e+00 -2.25653931e-01 -6.46875620e-01 3.79881449e-02
3.06798369e-01 4.99501079e-01 -7.27461040e-01 1.26101062e-01
6.41565204e-01 1.57404289e-01 -4.60062951e-01 1.31023204e+00
3.09206367e-01 1.98925152e-01 -4.61873889e-01 -2.74599314e-01
3.47421020e-01 1.19032636e-01 1.63855553e-01 1.32352471e+00
6.68468952e-01 1.01469710e-01 1.89628169e-01 1.34792686e-01
-9.43937153e-02 2.12876901e-01 -7.57355690e-01 1.08902371e-02
1.30763566e+00 1.06005955e+00 -6.84713781e-01 -2.81150132e-01
-4.30506259e-01 9.28799987e-01 2.57398963e-01 -3.80684823e-01
-6.47683859e-01 -9.31669056e-01 5.78098059e-01 2.75417596e-01
9.73468304e-01 -3.99716824e-01 2.82410644e-02 -7.16188550e-01
-2.30077170e-02 -5.96557558e-01 5.74811995e-01 -4.22708094e-01
-8.64637256e-01 3.90840739e-01 1.40177589e-02 -1.23001564e+00
-2.85267800e-01 5.60483076e-02 1.40687034e-01 5.60081661e-01
-1.56414139e+00 -9.11322415e-01 -1.34353116e-01 8.14257860e-01
1.93815783e-01 2.22711284e-02 8.92572939e-01 6.64470553e-01
5.52276634e-02 1.31986237e+00 4.72872317e-01 -1.78888217e-01
9.46131110e-01 -8.57907593e-01 2.35420749e-01 4.20376897e-01
4.85682413e-02 9.98829365e-01 5.66565573e-01 -3.75510931e-01
-1.84862244e+00 -8.78579557e-01 1.46470380e+00 8.07174817e-02
5.72936416e-01 -2.77273357e-01 -1.09745395e+00 9.64181200e-02
-2.94746071e-01 1.81301251e-01 1.06208909e+00 -2.10078791e-01
-8.32791686e-01 -3.68117571e-01 -1.41126609e+00 5.50983548e-01
7.50731289e-01 -1.04698730e+00 -6.00988150e-01 1.07780375e-01
8.69762301e-01 -6.39042165e-03 -1.17705095e+00 4.54898655e-01
9.28277433e-01 -7.23607004e-01 1.49462259e+00 1.65090710e-01
7.27996752e-02 -3.77215117e-01 -5.69774330e-01 -4.80068386e-01
-6.76502228e-01 -5.80623448e-01 -4.38320369e-01 8.37888896e-01
1.18477419e-01 -4.85910773e-01 8.07227075e-01 5.26360929e-01
2.85907239e-01 -9.22862291e-01 -1.08857965e+00 -1.33807743e+00
-1.20103583e-02 1.72162041e-01 6.65790737e-01 6.75359845e-01
4.94370550e-01 7.42366686e-02 -6.24735296e-01 -1.53509244e-01
7.42455542e-01 5.11368394e-01 5.39410889e-01 -1.06149018e+00
-1.27846777e-01 -1.91426829e-01 -6.86532974e-01 -1.15575206e+00
-1.08788028e-01 -8.27309847e-01 4.89461534e-02 -9.46155190e-01
5.60532212e-01 -7.25015998e-01 -4.63207126e-01 6.15272522e-01
-3.04287132e-02 7.28988945e-01 2.72726148e-01 6.42411411e-01
-7.79055238e-01 7.48790443e-01 8.58946204e-01 -1.49708120e-02
-5.44676743e-02 -4.43135381e-01 -2.86642760e-01 -1.76171735e-01
6.28403723e-01 -7.25656450e-01 -4.73603159e-01 -1.32839577e-02
2.29493141e-01 3.64107221e-01 -2.97146030e-02 -1.06376314e+00
5.32350004e-01 2.05854714e-01 -1.25864921e-02 -8.64453614e-01
5.51007807e-01 -6.69247150e-01 2.49231488e-01 1.00917125e+00
-4.79660600e-01 5.54903030e-01 -6.54245168e-02 4.25600737e-01
-4.83202368e-01 -5.70283830e-01 9.38199043e-01 1.42780319e-01
-4.44552809e-01 2.12622151e-01 -3.82993698e-01 -4.31238472e-01
9.20375109e-01 -3.90745580e-01 -2.10683912e-01 -2.87716269e-01
-8.35129470e-02 -9.13514048e-02 7.26539016e-01 1.28265530e-01
8.36290479e-01 -1.69560230e+00 -3.62995267e-01 3.48953456e-01
2.59989768e-01 -3.36762398e-01 -1.26529755e-02 3.46982092e-01
-6.97135448e-01 9.45282757e-01 -2.06127048e-01 -3.79150689e-01
-1.62268722e+00 9.78651524e-01 -5.78581035e-01 -3.65049988e-01
-7.05065370e-01 7.82481551e-01 -4.57595110e-01 -3.27520035e-02
2.78723776e-01 1.97006389e-01 2.67649740e-01 1.08359113e-01
9.59955931e-01 4.94966805e-01 1.43099204e-01 -4.04510498e-01
-3.94991934e-01 7.57014930e-01 -2.61957675e-01 -1.86999694e-01
1.07116032e+00 -3.72471660e-01 -1.83253974e-01 3.32288146e-02
1.82240939e+00 -1.15452580e-01 -4.61459666e-01 -4.36192244e-01
5.31340800e-02 -7.94901431e-01 2.15451494e-01 -9.34175849e-02
-7.88627446e-01 5.64208627e-01 6.09265864e-01 1.42138317e-01
1.17173779e+00 -1.14870608e-01 1.60639620e+00 8.48371506e-01
7.98884630e-01 -8.75613332e-01 2.38348290e-01 3.71826231e-01
4.10037637e-01 -1.09474921e+00 2.97738075e-01 -1.47118315e-01
-1.21280886e-01 9.34724331e-01 2.07847834e-01 9.85256061e-02
6.25581682e-01 2.63385862e-01 -2.61263758e-01 -9.37678590e-02
-7.73741424e-01 -1.95269868e-01 5.21338359e-02 4.17151600e-01
1.86026245e-01 -1.93089470e-01 -8.45210612e-01 -5.55164181e-02
-3.36236387e-01 -2.04650730e-01 1.76528111e-01 1.47276020e+00
-8.74469519e-01 -1.54138803e+00 -3.91913325e-01 1.37805969e-01
-2.97684193e-01 -3.42499614e-01 3.93780693e-02 5.20228088e-01
-5.26990771e-01 7.12885439e-01 1.63648859e-01 -5.43458104e-01
-2.81241804e-01 1.25041813e-01 4.05432165e-01 -2.52557248e-01
-4.77305859e-01 5.11430465e-02 -4.74936426e-01 -6.47035480e-01
-1.60950944e-01 -3.81709695e-01 -1.35355806e+00 -7.30942607e-01
-5.41083455e-01 6.70951903e-01 7.73373544e-01 4.40448731e-01
5.67174494e-01 -3.51957798e-01 9.74419832e-01 -3.85920793e-01
-7.85149753e-01 -3.55923057e-01 -7.65445471e-01 2.96539247e-01
4.71654832e-01 -5.87039590e-01 -4.87626940e-01 -2.28796139e-01] | [11.178378105163574, 1.133490800857544] |
0290c4d0-295f-4a72-b21c-ca6a4b942bb8 | developing-a-fidelity-evaluation-approach-for | 2106.08492 | null | https://arxiv.org/abs/2106.08492v1 | https://arxiv.org/pdf/2106.08492v1.pdf | Developing a Fidelity Evaluation Approach for Interpretable Machine Learning | Although modern machine learning and deep learning methods allow for complex and in-depth data analytics, the predictive models generated by these methods are often highly complex, and lack transparency. Explainable AI (XAI) methods are used to improve the interpretability of these complex models, and in doing so improve transparency. However, the inherent fitness of these explainable methods can be hard to evaluate. In particular, methods to evaluate the fidelity of the explanation to the underlying black box require further development, especially for tabular data. In this paper, we (a) propose a three phase approach to developing an evaluation method; (b) adapt an existing evaluation method primarily for image and text data to evaluate models trained on tabular data; and (c) evaluate two popular explainable methods using this evaluation method. Our evaluations suggest that the internal mechanism of the underlying predictive model, the internal mechanism of the explainable method used and model and data complexity all affect explanation fidelity. Given that explanation fidelity is so sensitive to context and tools and data used, we could not clearly identify any specific explainable method as being superior to another. | ['Renuka Sindhgatta', 'Catarina Moreira', 'Chun Ouyang', 'Mythreyi Velmurugan'] | 2021-06-16 | null | null | null | null | ['explanation-fidelity-evaluation'] | ['methodology'] | [ 7.64078498e-02 5.62064648e-01 -3.39279324e-01 -5.18137932e-01
-4.32376742e-01 -7.30333686e-01 8.56371403e-01 2.88827032e-01
3.20984304e-01 5.73647201e-01 4.73854274e-01 -8.69663477e-01
-5.18273830e-01 -4.19115394e-01 -6.66644037e-01 1.45283702e-03
1.72949404e-01 6.36595726e-01 -4.12503123e-01 1.10906765e-01
6.61677063e-01 4.05194581e-01 -1.58886850e+00 6.88621819e-01
1.30376089e+00 8.04465950e-01 -1.66877732e-01 7.00466633e-01
-3.58481586e-01 1.17641878e+00 -7.54566371e-01 -5.28866470e-01
7.00503588e-02 -6.12731397e-01 -1.07151437e+00 -9.20523256e-02
5.15355289e-01 -4.15817946e-01 2.69257605e-01 4.94954765e-01
-1.10821620e-01 -3.52263123e-01 9.95532572e-01 -1.61275136e+00
-1.04231846e+00 7.74366736e-01 -2.39026062e-02 -2.78098043e-02
3.73986453e-01 3.01513106e-01 1.09565759e+00 -9.26452637e-01
6.20115161e-01 1.27882755e+00 8.08429241e-01 5.49737990e-01
-1.37946022e+00 -7.24305868e-01 2.81097949e-01 1.06825903e-01
-9.83746409e-01 -3.75727683e-01 5.11792719e-01 -7.84451604e-01
1.03675330e+00 5.26145637e-01 9.32931185e-01 9.12344456e-01
4.12160516e-01 5.21778882e-01 1.21392989e+00 -4.32663739e-01
2.10900083e-01 5.23499668e-01 1.78583503e-01 8.08650911e-01
7.04076767e-01 2.72832394e-01 -6.55756414e-01 -1.69403613e-01
5.14240682e-01 3.78271006e-02 -2.11230293e-01 -3.50767851e-01
-9.63271260e-01 1.08575749e+00 5.33192873e-01 1.62842482e-01
-2.10490197e-01 3.12019169e-01 1.02809332e-01 2.92684287e-01
4.99628276e-01 1.26980829e+00 -6.35594070e-01 -3.30644995e-01
-1.08560443e+00 5.06883860e-01 1.11097705e+00 8.40877533e-01
7.37050891e-01 2.73485314e-02 -9.16653797e-02 2.37183094e-01
6.17888331e-01 1.19296081e-01 1.50617614e-01 -1.06019819e+00
4.11216319e-01 1.23507595e+00 1.52641699e-01 -1.18892658e+00
-5.23418725e-01 -3.40650558e-01 -4.14115667e-01 5.73762596e-01
3.80826145e-01 1.54713437e-01 -9.72350240e-01 1.21402752e+00
-1.65820956e-01 -8.60081375e-01 4.81329709e-02 8.92384171e-01
1.16174626e+00 5.86727977e-01 2.81764269e-01 2.33444184e-01
9.82759595e-01 -1.18372130e+00 -7.70303845e-01 -4.16405171e-01
1.03263438e+00 -6.93352520e-01 1.59695888e+00 3.90414834e-01
-1.09248221e+00 -4.87452000e-01 -1.11253190e+00 -3.74269277e-01
-5.52483439e-01 8.98823440e-02 9.62251067e-01 4.76288587e-01
-9.50130880e-01 7.18794763e-01 -7.53473043e-01 -1.41689435e-01
4.53942597e-01 3.12638074e-01 -3.60390693e-01 6.81451485e-02
-9.12510455e-01 1.32019269e+00 2.83749312e-01 -8.82477500e-04
-7.12441206e-01 -8.23340774e-01 -7.61101365e-01 4.60950464e-01
2.07887948e-01 -9.32071149e-01 1.23999250e+00 -1.43202555e+00
-9.76405323e-01 4.85231400e-01 -1.58086121e-01 -2.09002480e-01
5.64704895e-01 -4.18835372e-01 -5.88906072e-02 -2.46793851e-01
6.75811023e-02 7.61193454e-01 5.42421043e-01 -1.49231768e+00
-4.43060070e-01 -2.73336828e-01 2.73892134e-01 2.01033846e-01
-1.48173437e-01 -1.15032308e-01 -1.85230970e-02 -4.26539600e-01
7.61984736e-02 -8.87026489e-01 -4.25637811e-02 2.58808941e-01
-3.70504558e-01 -1.37410000e-01 6.36722803e-01 -6.49893820e-01
1.40029514e+00 -1.88403320e+00 -5.62430033e-03 1.97904333e-02
7.47425377e-01 2.07313448e-02 1.19507819e-01 5.37027001e-01
-1.28077760e-01 9.72578824e-01 -4.19491082e-02 -2.33890787e-01
2.45208412e-01 9.63389650e-02 -2.90594548e-01 1.39816920e-03
3.44372749e-01 9.14857268e-01 -5.48592985e-01 -5.04249692e-01
2.17192367e-01 5.92714369e-01 -8.21111917e-01 2.22489849e-01
-2.95884937e-01 2.31877476e-01 -3.74076158e-01 6.28504038e-01
2.81482965e-01 -5.75527728e-01 -1.27945647e-01 -1.71173736e-01
-2.94677228e-01 6.61492229e-01 -9.40741599e-01 1.16803432e+00
-3.47487181e-01 1.00511718e+00 -4.11739230e-01 -2.85057962e-01
1.01453984e+00 2.46301726e-01 1.12136036e-01 -5.13996124e-01
-2.64755905e-01 4.40345109e-01 2.29023218e-01 -5.01180649e-01
5.81642985e-01 -1.38501346e-01 2.65698075e-01 6.56235099e-01
-3.75421554e-01 -5.02499998e-01 -1.12367809e-01 2.14219049e-01
9.55273211e-01 3.23722601e-01 6.91919684e-01 -3.73386204e-01
-7.59980932e-04 6.15342259e-01 3.29262346e-01 9.52808678e-01
1.63761084e-03 7.80268550e-01 7.64049888e-01 -1.12714207e+00
-1.19361603e+00 -5.24582267e-01 -4.30095717e-02 7.16650605e-01
-6.48292005e-02 -8.89942586e-01 -7.15721965e-01 -7.77553916e-01
5.60030229e-02 1.30328012e+00 -7.78281391e-01 -1.35093361e-01
-2.79359464e-02 -2.29005411e-01 2.94188231e-01 6.04340792e-01
2.27275521e-01 -1.01277995e+00 -1.06531680e+00 5.33672944e-02
-5.52135780e-02 -5.41873634e-01 -1.52296379e-01 3.33423674e-01
-1.11814499e+00 -1.08481848e+00 8.18582699e-02 4.19996530e-02
7.50219047e-01 2.17662394e-01 1.52633929e+00 7.21134424e-01
1.08274840e-01 2.43676901e-01 -4.03058469e-01 -9.19390321e-01
-6.94671512e-01 1.20076887e-01 -3.81716847e-01 -7.52208531e-01
3.32809687e-01 7.73666473e-03 -5.93908131e-01 3.14439416e-01
-9.49302733e-01 6.74779773e-01 5.98631263e-01 7.52257824e-01
3.74674827e-01 -1.57215089e-01 -1.68507230e-02 -1.17610729e+00
9.24094737e-01 -3.51838946e-01 -3.22906971e-01 4.25737619e-01
-1.46834910e+00 3.47492665e-01 4.87325221e-01 -1.79790854e-01
-8.54079187e-01 -2.19969656e-02 2.31040984e-01 -9.89024248e-03
-4.52991538e-02 1.04860938e+00 5.81858717e-02 9.05254781e-02
1.04866779e+00 -4.83541101e-01 2.23385077e-03 -3.65155727e-01
3.63356054e-01 3.81882131e-01 4.76990677e-02 -4.04666722e-01
7.19010472e-01 2.35871837e-01 -2.67155558e-01 -8.25366527e-02
-7.31550574e-01 3.77466500e-01 -6.34193361e-01 -2.09640875e-01
6.52720630e-01 -5.51187515e-01 -5.01960397e-01 -3.56726348e-01
-1.30459666e+00 -4.16092992e-01 -3.06232333e-01 5.28964520e-01
-4.39324677e-01 -1.32601067e-01 -1.36229530e-01 -8.20036888e-01
-4.99473721e-01 -1.33854818e+00 9.49359894e-01 1.86587095e-01
-1.15679145e+00 -1.21880376e+00 -1.00248769e-01 6.34439111e-01
6.73595846e-01 4.85965163e-01 1.37067854e+00 -8.67812574e-01
-7.10507154e-01 -3.53524148e-01 -4.10483569e-01 -1.41552418e-01
7.14305788e-02 7.05177665e-01 -8.66436481e-01 2.11820543e-01
-5.51810712e-02 -3.77237350e-01 4.45782095e-01 4.33801204e-01
1.23185897e+00 -8.71505141e-01 -3.32368970e-01 5.66013157e-01
1.09730148e+00 6.12979010e-02 4.90454227e-01 8.16053629e-01
8.82428527e-01 9.71247494e-01 6.58564746e-01 2.03457162e-01
5.72457194e-01 5.37936032e-01 6.12969995e-01 -4.69890505e-01
-1.72978360e-02 -4.84806478e-01 2.33644173e-01 4.07802999e-01
-1.53157547e-01 -2.06512570e-01 -1.34314573e+00 1.93054602e-01
-1.94452369e+00 -8.78094614e-01 -4.14403886e-01 1.87338245e+00
6.37765288e-01 2.86706954e-01 -7.92476535e-02 5.67598864e-02
3.37056629e-02 -2.03201205e-01 -5.55320382e-01 -1.07702231e+00
-8.55116732e-03 -3.05067539e-01 1.92049354e-01 5.62509060e-01
-4.25848544e-01 7.89595425e-01 7.65520239e+00 1.60925448e-01
-1.01192689e+00 -3.17476809e-01 9.00066614e-01 -8.42259750e-02
-1.02930498e+00 4.81014192e-01 -4.28908139e-01 2.15216488e-01
1.02775979e+00 -2.42107853e-01 4.62093443e-01 1.13999200e+00
4.75370586e-01 6.26699850e-02 -1.67392063e+00 5.77503145e-01
-1.71340808e-01 -1.64423609e+00 4.34525818e-01 2.30841041e-01
5.96968293e-01 -4.86409009e-01 1.45957574e-01 1.97323233e-01
2.15718493e-01 -1.82090580e+00 1.01382804e+00 5.70053458e-01
5.12605309e-01 -5.78132451e-01 7.39413142e-01 7.57624358e-02
-6.75665855e-01 -5.27949929e-02 -1.74101830e-01 -6.13047957e-01
-3.66858780e-01 3.34827125e-01 -1.01532924e+00 1.84816211e-01
8.37584376e-01 5.68301141e-01 -9.19854820e-01 6.40512288e-01
-4.15750176e-01 8.03508282e-01 1.07170209e-01 -1.29692480e-01
1.65591836e-01 1.95458427e-01 3.12270373e-01 1.10665011e+00
2.64997363e-01 3.62957492e-02 -3.10859650e-01 1.58613336e+00
2.70966738e-01 2.02433951e-02 -9.72052395e-01 -4.75104719e-01
5.16960323e-01 1.11704826e+00 -5.50931096e-01 -2.29521215e-01
-4.54717755e-01 5.20000637e-01 3.50776851e-01 2.83838749e-01
-7.05967844e-01 9.95987281e-02 5.31286001e-01 3.68710726e-01
-8.75836387e-02 -1.23992309e-01 -1.07595897e+00 -8.61432970e-01
-2.62798625e-03 -1.48237205e+00 3.07924747e-01 -1.40364218e+00
-9.55422580e-01 7.29578733e-01 1.22354552e-01 -1.00713933e+00
-4.65185881e-01 -5.31597853e-01 -5.43194234e-01 9.65972066e-01
-1.22305703e+00 -1.20639038e+00 -6.43842995e-01 1.53374493e-01
3.79664779e-01 -1.49071589e-01 8.92093778e-01 -3.75883579e-01
-2.85019577e-01 3.34268063e-01 -1.25312507e-01 -4.31716144e-01
5.57251632e-01 -1.40456831e+00 5.81872880e-01 4.80371952e-01
2.16393083e-01 1.17505097e+00 1.15508962e+00 -7.83919156e-01
-1.31576800e+00 -6.66087508e-01 1.12575758e+00 -1.06800389e+00
4.44146723e-01 -1.26315385e-01 -1.04733849e+00 8.42868090e-01
2.24572405e-01 -5.10930419e-01 8.26479793e-01 5.74526668e-01
-4.23984975e-01 1.25258356e-01 -1.06154490e+00 7.58809388e-01
7.59243190e-01 -3.63755286e-01 -5.55168927e-01 1.23967320e-01
7.13510990e-01 -3.91050369e-01 -8.22693825e-01 4.13722873e-01
8.04278672e-01 -1.28344202e+00 4.83314008e-01 -7.85429716e-01
1.20129037e+00 -2.31349617e-01 6.02106713e-02 -1.31884468e+00
-4.69913393e-01 -5.31344533e-01 -1.43521309e-01 1.07731974e+00
1.01425898e+00 -5.04883170e-01 7.05120862e-01 1.58083153e+00
-2.59500831e-01 -9.19962943e-01 -3.35674107e-01 -3.85658771e-01
1.21718183e-01 -5.27347982e-01 1.02863264e+00 9.89240408e-01
1.23051874e-01 3.27566266e-01 -2.49229878e-01 -1.20369688e-01
1.91848725e-01 2.22471699e-01 1.14587188e+00 -1.34786892e+00
-1.08560048e-01 -6.24149024e-01 -2.79509515e-01 -4.62704480e-01
-2.23202333e-01 -6.56808376e-01 -2.00569361e-01 -1.97409916e+00
5.13137877e-01 -4.12744641e-01 1.05046682e-01 7.53724217e-01
-4.30541098e-01 -2.93564051e-01 3.59460682e-01 7.76798248e-01
-2.62987345e-01 2.47616902e-01 1.03857267e+00 -1.16779856e-01
-3.11442137e-01 -3.14910680e-01 -1.40842891e+00 8.95918250e-01
6.50137246e-01 -6.89778805e-01 -5.99065959e-01 -7.52775192e-01
6.41328752e-01 -2.22008288e-01 5.64025104e-01 -7.44427085e-01
4.29767445e-02 -4.64447767e-01 5.86073160e-01 -3.24800551e-01
2.52473075e-02 -8.55127037e-01 5.53426504e-01 5.73273897e-01
-7.68246651e-01 4.17593867e-01 5.03615558e-01 1.20026991e-01
-2.05461353e-01 -2.80034184e-01 2.61153966e-01 -8.92053545e-02
-3.10743272e-01 -6.59272671e-02 -2.32116595e-01 -1.56881645e-01
6.33266151e-01 -6.77256346e-01 -6.22744024e-01 -6.91855490e-01
-2.03826368e-01 1.65036142e-01 7.66517341e-01 4.61865634e-01
5.93035460e-01 -1.23110461e+00 -5.49811482e-01 -4.00798656e-02
3.42849523e-01 -9.11915898e-02 -2.56944746e-01 4.90702480e-01
-7.21402466e-01 6.71880901e-01 -3.24200600e-01 -3.87687445e-01
-1.19611490e+00 4.35862839e-01 3.36398780e-01 -1.51532531e-01
-4.99542981e-01 6.21443033e-01 3.20108175e-01 -4.65814263e-01
1.62610039e-01 -6.08048499e-01 -1.68261603e-01 -2.04933196e-01
3.23633343e-01 3.30131710e-01 -1.90631524e-01 -1.99272186e-01
-2.59915292e-01 1.90518394e-01 -2.16221705e-01 -1.33086488e-01
1.42922843e+00 4.04968578e-03 3.43807638e-02 5.58239818e-01
6.88478708e-01 -1.84701741e-01 -1.27998888e+00 4.17500138e-01
1.00244038e-01 -5.69299877e-01 3.34485471e-02 -1.26945364e+00
-6.76933050e-01 1.03237653e+00 2.26006672e-01 6.05562866e-01
8.48685026e-01 -4.19243947e-02 5.05987555e-02 2.36373752e-01
-2.17514396e-01 -9.20452356e-01 -2.04673386e-03 6.20268434e-02
1.38465834e+00 -1.53040946e+00 5.19117951e-01 -2.29539722e-01
-9.36081052e-01 1.44596386e+00 8.27834189e-01 5.54823160e-01
2.47631595e-01 7.34733194e-02 2.69970328e-01 -7.28174388e-01
-1.04098678e+00 4.30995673e-01 7.77037621e-01 4.18957949e-01
1.01798761e+00 5.57148419e-02 -3.02289844e-01 6.35701299e-01
-6.64627731e-01 1.33161679e-01 4.55164582e-01 6.73766434e-01
-2.57443190e-01 -6.70549989e-01 -5.35552084e-01 6.64830983e-01
-1.75945669e-01 -1.63089365e-01 -1.32798386e+00 1.25501108e+00
-2.48092134e-02 1.17335474e+00 -2.05024555e-01 -6.30135536e-01
2.35104576e-01 -1.09264068e-01 8.36274773e-02 -5.82002699e-01
-8.31095636e-01 -3.75840396e-01 3.16644788e-01 -7.14817226e-01
-6.63865805e-02 -5.95575094e-01 -1.25648284e+00 -7.22625852e-01
-2.82786429e-01 1.61360383e-01 1.02322209e+00 1.16140866e+00
6.18219972e-01 4.59957540e-01 4.70915958e-02 -5.56994617e-01
-4.60168362e-01 -8.35789382e-01 5.40165380e-02 5.09745061e-01
2.33776510e-01 -6.67782128e-01 -5.58834374e-01 7.90362060e-02] | [8.831499099731445, 5.9264726638793945] |
19d3f41d-a056-4746-a02f-76755af81d65 | an-image-source-method-framework-for | 1906.12227 | null | http://arxiv.org/abs/1906.12227v1 | http://arxiv.org/pdf/1906.12227v1.pdf | An Image Source Method Framework for Arbitrary Reflecting Boundaries | We propose a theoretical framework for the image source method that
generalizes to arbitrary reflecting boundaries, e.g. boundaries that are curved
or even with certain openings. Furthermore, it can seamlessly incorporate
boundary absorption, source directivity, and nonspecular reflections. This
framework is based on the notion of reflection paths that allows the
introduction of the concepts of validity and visibility of virtual sources.
These definitions facilitate the determination, for a given source and receiver
location, of the distribution of virtual sources that explain the boundary
effects of a wide range of reflecting surfaces. The structure of the set of
virtual sources is then more general than just punctual virtual sources. Due to
this more diverse configuration of image sources, we represent the room impulse
response as an integral involving the temporal excitation signal against a
measure determined by the source and receiver locations, and the original
boundary. The latter smoothly enables, in an analytically tractable manner, the
incorporation of more general boundary shapes as well as directivity of sources
and boundary absorption while, at the same time, maintaining the conceptual
benefits of the image source method. | [] | 2019-06-28 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 6.12997115e-01 1.61367342e-01 4.47734714e-01 -7.45599642e-02
-5.33585370e-01 -8.00046027e-01 7.06805587e-01 -1.43286616e-01
2.56967187e-01 4.75571126e-01 3.59119833e-01 -3.61912876e-01
-3.30829889e-01 -1.01501977e+00 -4.55362201e-01 -8.08229625e-01
-1.46567538e-01 -1.02833569e-01 2.11681426e-01 -3.94231290e-01
1.16879471e-01 5.97617745e-01 -1.58882654e+00 8.31927434e-02
6.70682192e-01 1.05632365e+00 1.12820700e-01 7.31707573e-01
-2.08776981e-01 5.02655029e-01 -6.66204333e-01 -1.12110758e-02
1.45291835e-01 -4.57193047e-01 -4.68023777e-01 2.02519462e-01
1.38349637e-01 -1.15258761e-01 -3.95663753e-02 6.89795732e-01
3.04569006e-01 1.82038695e-01 7.29304850e-01 -9.03178990e-01
-1.08651125e+00 2.39663512e-01 -4.25764740e-01 -8.61899853e-02
8.92445385e-01 -2.81192541e-01 7.83699632e-01 -9.02606308e-01
4.16493386e-01 1.01214850e+00 7.88245857e-01 2.94686794e-01
-1.05173659e+00 -5.56276273e-03 -7.47199878e-02 -2.64541477e-01
-1.35903060e+00 -5.64182997e-01 8.64171207e-01 -1.95259303e-01
4.40650403e-01 8.17844987e-01 4.19210970e-01 8.31474066e-01
1.46548927e-01 7.16456547e-02 1.08371854e+00 -1.00599313e+00
3.02604645e-01 5.05843818e-01 3.77772063e-01 4.80687290e-01
5.80431856e-02 1.75175220e-01 -3.13155330e-03 -3.60104918e-01
7.56038427e-01 -6.26982152e-02 -9.70960796e-01 -3.50244641e-01
-1.01349902e+00 4.36118066e-01 1.42148316e-01 5.52382529e-01
-1.53955057e-01 -9.15930271e-02 -2.31654257e-01 1.17380008e-01
3.28872830e-01 2.08556086e-01 4.18346841e-03 4.43433583e-01
-6.76008642e-01 9.85176340e-02 9.84055459e-01 9.72963512e-01
7.44170725e-01 5.42857572e-02 1.11815976e-02 7.79179752e-01
4.72941518e-01 1.02104676e+00 -4.20539886e-01 -8.71931434e-01
1.27848968e-01 1.83724403e-01 7.24065125e-01 -8.10615778e-01
-4.86199617e-01 -5.59785426e-01 -5.20446837e-01 5.64004719e-01
6.76049173e-01 -1.04123667e-01 -7.92277575e-01 1.82126725e+00
5.41589439e-01 5.31480946e-02 2.46809945e-01 7.76715100e-01
7.14760482e-01 7.90624738e-01 -5.14082193e-01 -4.59038228e-01
1.19492376e+00 -4.04728115e-01 -8.27139914e-01 1.94805130e-01
2.08258703e-01 -9.27340627e-01 1.03339469e+00 4.13055539e-01
-1.34115648e+00 -3.03173721e-01 -9.58429813e-01 1.20010480e-01
-1.68733686e-01 -2.88489550e-01 2.62640208e-01 1.08881414e+00
-1.43633235e+00 2.18559563e-01 -2.89602607e-01 -1.93269197e-02
-3.67433250e-01 -4.08024117e-02 2.32438609e-01 2.56746896e-02
-1.06643975e+00 7.04640627e-01 -5.43116868e-01 4.01342899e-01
-5.23441911e-01 -1.10860157e+00 -6.59022748e-01 4.48328815e-03
2.68086612e-01 -4.77208436e-01 1.07230937e+00 -1.29038382e+00
-1.45932353e+00 6.56537473e-01 -2.34686255e-01 1.68965697e-01
5.77141881e-01 5.76403178e-02 -8.97611260e-01 2.51444727e-01
-9.16896239e-02 -3.17549497e-01 7.61366546e-01 -2.03043914e+00
2.08818316e-02 -2.10997760e-01 1.78758398e-01 1.82050750e-01
1.30425274e-01 -5.85177504e-02 -3.41016613e-02 -1.14729218e-01
2.54110843e-01 -7.22249985e-01 -1.44193903e-01 3.54992738e-03
-6.71296716e-01 1.77358270e-01 5.87547064e-01 -5.23762226e-01
8.38972926e-01 -2.44681191e+00 -8.85628760e-02 7.92481303e-01
1.52759373e-01 -2.11783364e-01 -1.00575134e-01 8.34543705e-01
-2.03055069e-01 5.42815179e-02 -4.79923904e-01 -8.35088491e-02
-3.03642103e-03 -3.14018995e-01 -5.85575044e-01 6.58509851e-01
-2.06788555e-01 4.59552020e-01 -5.02326906e-01 -1.38141569e-02
2.21020848e-01 1.06890142e+00 -2.64432222e-01 -4.89360169e-02
9.73525345e-02 4.97542620e-01 -6.52011037e-01 2.42208242e-01
1.02477074e+00 2.64529902e-02 -1.45355344e-01 1.72416374e-01
-7.55882978e-01 1.04712576e-01 -1.55302489e+00 1.08737826e+00
-6.84249043e-01 3.69215399e-01 6.63151801e-01 -5.16501904e-01
1.14094770e+00 5.47004938e-01 4.81621534e-01 -7.51584470e-01
-6.71634302e-02 4.19076443e-01 -5.69790065e-01 -4.29226577e-01
5.81528246e-01 -5.13878584e-01 8.16066787e-02 5.55572391e-01
-4.15196091e-01 4.28454019e-03 -5.04979968e-01 2.56993845e-02
8.30922365e-01 2.70696096e-02 1.64963573e-01 -5.15921414e-01
6.28506303e-01 -2.24536702e-01 7.79660419e-03 7.03313231e-01
1.97751373e-01 7.89513111e-01 1.51237413e-01 1.28231272e-01
-9.10894394e-01 -1.68898666e+00 -6.83881760e-01 7.13219285e-01
5.85180461e-01 7.74143785e-02 -7.00409472e-01 3.57117474e-01
-1.65453896e-01 8.59369218e-01 -5.42784154e-01 -1.14912000e-02
-5.15102446e-01 -4.47324902e-01 3.83090943e-01 -5.23007251e-02
3.00097108e-01 -7.96582222e-01 -6.87025726e-01 8.97670537e-03
-4.41929281e-01 -8.44225466e-01 -1.31953597e-01 -1.60935402e-01
-6.57783568e-01 -9.66598034e-01 -1.12160623e+00 -4.53267276e-01
4.34342086e-01 6.63354754e-01 1.10429275e+00 1.96711496e-01
-4.69407290e-01 1.30549872e+00 -4.12524343e-01 -3.33899587e-01
-5.44195592e-01 -9.53049183e-01 -2.51526803e-01 3.34142447e-01
-3.74335051e-01 -6.87024355e-01 -6.76648259e-01 4.18087661e-01
-8.49899352e-01 -2.94175386e-01 -1.42654836e-01 1.65579870e-01
4.34537590e-01 3.60642634e-02 5.62696636e-01 -2.98797458e-01
3.67511660e-01 -6.02090597e-01 -5.57553113e-01 3.73558223e-01
-1.16391830e-01 -2.78825074e-01 6.56291366e-01 -1.71289086e-01
-1.65829563e+00 -6.58534050e-01 -1.48266435e-01 3.39658529e-01
-3.19979399e-01 8.66886787e-03 -3.16655397e-01 -3.38458747e-01
8.81089866e-01 1.38470605e-01 -3.17904174e-01 -3.86523843e-01
5.55929720e-01 4.91757780e-01 3.83736253e-01 -4.86417502e-01
7.32121468e-01 9.06109869e-01 1.86274022e-01 -1.47423685e+00
-5.08137524e-01 -4.38821495e-01 -3.50991070e-01 -3.90205652e-01
7.51705766e-01 -4.89681363e-01 -6.49130404e-01 4.49708343e-01
-1.23995221e+00 -3.42659026e-01 -6.25087321e-01 4.87652391e-01
-6.61466122e-01 3.74089032e-01 -3.32861692e-01 -1.47929561e+00
1.90863654e-01 -7.20787764e-01 7.93242157e-01 2.16521904e-01
8.27614367e-02 -1.46793568e+00 1.70062199e-01 1.09661609e-01
5.21737039e-01 5.68946302e-01 1.04269946e+00 -1.71032660e-02
-8.92372489e-01 2.86322497e-02 -1.00518577e-01 9.84096900e-02
1.84957087e-01 1.72003098e-02 -1.30082536e+00 8.37532282e-02
6.85023725e-01 2.18348026e-01 7.12395191e-01 8.33450437e-01
3.93131763e-01 -1.09547287e-01 -3.92257243e-01 5.61688542e-01
1.95066643e+00 3.01247060e-01 9.60547447e-01 8.69321898e-02
3.90423864e-01 8.99948180e-01 9.46880504e-03 5.52827537e-01
1.49225116e-01 6.69163525e-01 5.16682208e-01 -5.34672618e-01
-1.09430820e-01 2.00671569e-01 2.39197567e-01 2.96115547e-01
-4.08419490e-01 -6.10142589e-01 -6.50037646e-01 4.99602526e-01
-1.06867802e+00 -1.07757485e+00 -8.32804501e-01 2.69455028e+00
3.43897820e-01 -3.50596339e-01 -4.40583900e-02 3.14387560e-01
7.80818522e-01 2.75148124e-01 -1.18529208e-01 -6.35983109e-01
-1.61477268e-01 1.50063798e-01 4.04660136e-01 1.03795660e+00
-3.89370054e-01 2.81332694e-02 7.49480343e+00 2.53085166e-01
-8.81381989e-01 2.55131811e-01 1.13757960e-01 1.96655005e-01
-1.21407866e+00 -1.55561656e-01 -6.22252762e-01 -2.55675036e-02
7.08710730e-01 -1.97649980e-03 5.93534529e-01 1.84252903e-01
4.48008180e-01 -3.63121748e-01 -8.45665097e-01 4.26227063e-01
1.61008351e-03 -9.53675449e-01 -2.00386226e-01 5.51459827e-02
6.20567560e-01 -3.76734585e-01 2.52181500e-01 -4.46891963e-01
-6.86795935e-02 -9.47304428e-01 1.02391768e+00 7.13294387e-01
7.56071925e-01 -4.70506638e-01 1.50914332e-02 2.41780862e-01
-1.16668177e+00 -1.14326216e-01 -9.87879038e-02 -5.14497608e-02
5.20319879e-01 9.32440042e-01 -8.10230076e-02 7.04936206e-01
3.79244059e-01 -8.67986381e-02 2.64318943e-01 1.00700581e+00
-7.58716092e-02 4.36411917e-01 -5.17786384e-01 1.90035835e-01
1.44824713e-01 -4.48632836e-01 8.70363951e-01 1.24113560e+00
6.74313903e-01 2.60640889e-01 -7.42914677e-02 1.28499675e+00
3.97034109e-01 1.59218013e-01 -8.14876199e-01 7.68437982e-01
3.54769617e-01 1.09484482e+00 -7.03222096e-01 1.79171339e-01
-5.51932275e-01 6.89318955e-01 -3.99278700e-01 1.04029524e+00
-9.83239472e-01 -3.85635734e-01 5.62237024e-01 5.44653475e-01
2.65431739e-02 -4.65483576e-01 -2.73541749e-01 -6.69445753e-01
3.50306571e-01 -3.01974714e-01 -1.28870532e-01 -8.65454793e-01
-8.88066113e-01 3.56074601e-01 2.69842178e-01 -1.06276035e+00
4.05523658e-01 -5.06942451e-01 -1.02688110e+00 1.20820868e+00
-1.88711536e+00 -8.47458243e-01 -3.83588731e-01 9.92972076e-01
-1.67494386e-01 6.12330258e-01 8.69158506e-01 6.16698302e-02
9.86615717e-02 1.29914880e-01 5.53036630e-01 -2.95012414e-01
2.89437979e-01 -1.10891771e+00 -2.50349659e-02 7.04267800e-01
-1.42304018e-01 6.60236955e-01 8.88574719e-01 -2.27930605e-01
-1.26054728e+00 -7.93725789e-01 7.01280296e-01 -3.83087784e-01
4.49527621e-01 -4.27797794e-01 -7.88721740e-01 6.32164538e-01
1.69030905e-01 -3.00199121e-01 7.69206524e-01 1.94886431e-01
-4.05793637e-01 1.51207456e-02 -1.22457337e+00 6.58665657e-01
1.11176348e+00 -3.90294999e-01 -4.39535856e-01 4.01924461e-01
5.86789608e-01 -2.16501385e-01 -6.19553149e-01 7.39750490e-02
4.95402664e-01 -1.46881521e+00 1.39772367e+00 2.57022113e-01
1.94263965e-01 -1.17222331e-01 -2.77601272e-01 -1.16348481e+00
-4.23522234e-01 -7.57501721e-01 5.78232825e-01 1.30312157e+00
5.97105384e-01 -1.12074554e+00 -8.32711086e-02 5.83831012e-01
-4.81449634e-01 -2.96457112e-01 -1.03051186e+00 -8.67599964e-01
8.31281543e-02 -5.83642364e-01 4.40621197e-01 7.05896199e-01
1.13364398e-01 -2.39433557e-01 -2.70478845e-01 5.49696565e-01
9.67269242e-01 2.71327257e-01 1.29928753e-01 -1.22890270e+00
-4.55945343e-01 -1.78947434e-01 9.98125523e-02 -1.10400665e+00
-2.71905273e-01 -7.51583159e-01 2.49152675e-01 -1.74772596e+00
-3.12212408e-01 -1.04802895e+00 -1.54458195e-01 -2.56494999e-01
2.96810955e-01 1.53927013e-01 -1.21260375e-01 2.16950923e-01
6.43876493e-02 3.63225967e-01 1.51564705e+00 3.67826551e-01
-6.43002391e-01 2.38736629e-01 -6.63097620e-01 8.44032168e-01
7.53728747e-01 -1.34701684e-01 -5.83378613e-01 -3.45546663e-01
4.32986408e-01 3.10651153e-01 9.27935362e-01 -8.19086552e-01
-2.44469970e-01 -1.77456394e-01 2.09370375e-01 -9.50327814e-02
5.57698488e-01 -7.10987091e-01 3.40745062e-01 -4.16768342e-02
-1.99903652e-01 -5.57425797e-01 1.04776271e-01 6.28988922e-01
2.33289286e-01 -3.26054633e-01 7.63882816e-01 -2.81215340e-01
-3.26288462e-01 -1.98329553e-01 -4.69053239e-01 1.22539990e-01
9.90032434e-01 -8.44392002e-01 -4.82043356e-01 -4.76525098e-01
-7.38665104e-01 -3.58028084e-01 4.92341459e-01 2.44488642e-02
5.74183106e-01 -1.16298640e+00 -4.97968227e-01 2.25368977e-01
-1.50036916e-01 -2.96872646e-01 4.74211872e-01 9.25359786e-01
-2.35216111e-01 1.72967970e-01 9.87346936e-03 -4.51549888e-01
-1.15031838e+00 5.41496575e-01 6.90727532e-01 1.61659628e-01
-9.37875330e-01 7.12908089e-01 8.02825272e-01 -2.34678522e-01
-1.43306166e-01 -2.83002049e-01 -2.54346162e-01 -2.64805377e-01
6.92589819e-01 6.01492405e-01 -2.03576267e-01 -6.91641390e-01
-3.41358840e-01 1.10627234e+00 7.97697783e-01 -4.02403802e-01
8.59255612e-01 -5.80350637e-01 -1.42977282e-01 7.45538950e-01
8.17729175e-01 8.16453934e-01 -1.08983517e+00 1.51555112e-03
-4.63002712e-01 -2.58289605e-01 -1.76237643e-01 -9.49178398e-01
-6.27094328e-01 8.21571290e-01 3.85510564e-01 8.17613423e-01
1.25786805e+00 1.27340570e-01 4.23937231e-01 -1.64400503e-01
5.23982227e-01 -7.21100569e-01 -1.29283234e-01 3.15118015e-01
1.02643025e+00 -4.33680564e-01 -4.00076300e-01 -1.03244543e+00
-5.94105609e-02 1.27607501e+00 -2.40308776e-01 -8.57684240e-02
8.83448243e-01 6.94943607e-01 2.90022105e-01 -1.31023630e-01
-2.71772929e-02 -9.10778642e-02 1.43369064e-01 9.12418485e-01
5.44437170e-01 2.12960243e-01 -2.38067165e-01 1.38409764e-01
1.28216878e-01 -5.46861477e-02 8.59285295e-01 7.23434150e-01
-8.04301023e-01 -8.25566053e-01 -1.04880238e+00 -2.88732171e-01
-3.72391105e-01 -1.42595604e-01 -1.02053434e-01 6.97242677e-01
1.51875064e-01 1.44794488e+00 1.97845865e-02 9.30587202e-02
5.92297912e-01 4.54428010e-02 5.84813476e-01 -3.70590866e-01
-2.86769778e-01 9.52041745e-02 9.60427821e-02 -2.50883013e-01
-7.06259191e-01 -4.64862794e-01 -1.50042832e+00 -1.84649959e-01
-5.62599838e-01 1.65462330e-01 7.48430073e-01 8.80889416e-01
5.63629810e-03 6.13307416e-01 9.06814218e-01 -7.66271472e-01
-4.85018611e-01 -4.99906272e-01 -1.09841454e+00 1.63925305e-01
8.14396143e-01 -5.24093509e-01 -6.62815690e-01 -3.89165580e-02] | [15.143163681030273, 5.674461364746094] |
de361482-886b-4b3a-bfb5-30745c020217 | fusiondepth-complement-self-supervised | 2305.06036 | null | https://arxiv.org/abs/2305.06036v1 | https://arxiv.org/pdf/2305.06036v1.pdf | FusionDepth: Complement Self-Supervised Monocular Depth Estimation with Cost Volume | Multi-view stereo depth estimation based on cost volume usually works better than self-supervised monocular depth estimation except for moving objects and low-textured surfaces. So in this paper, we propose a multi-frame depth estimation framework which monocular depth can be refined continuously by multi-frame sequential constraints, leveraging a Bayesian fusion layer within several iterations. Both monocular and multi-view networks can be trained with no depth supervision. Our method also enhances the interpretability when combining monocular estimation with multi-view cost volume. Detailed experiments show that our method surpasses state-of-the-art unsupervised methods utilizing single or multiple frames at test time on KITTI benchmark. | ['Yong liu', 'Ying Chen', 'Shang Xu', 'Jianlin Liu', 'Zhuofei Huang'] | 2023-05-10 | null | null | null | null | ['stereo-depth-estimation', 'monocular-depth-estimation'] | ['computer-vision', 'computer-vision'] | [ 7.17280898e-03 5.71994595e-02 -3.96242619e-01 -6.15056396e-01
-5.08755744e-01 -3.16684008e-01 5.53754508e-01 -5.98483741e-01
-5.09412706e-01 9.38470960e-01 1.59374624e-01 1.60056695e-01
2.20395043e-01 -7.75866032e-01 -7.45669603e-01 -6.17124259e-01
4.38558549e-01 3.88473958e-01 6.46075666e-01 3.60613644e-01
3.83462638e-01 2.64618188e-01 -1.68574166e+00 4.75005448e-01
4.97089744e-01 1.05877686e+00 5.09567022e-01 8.35551739e-01
6.51643276e-02 9.43918049e-01 -1.42767549e-01 -5.62828071e-02
3.99603695e-01 4.38595079e-02 -6.24404430e-01 3.84678066e-01
9.96458948e-01 -1.11227548e+00 -4.84067947e-01 9.88810360e-01
4.25963253e-01 3.27672884e-02 4.19191808e-01 -8.47707927e-01
-7.73305520e-02 8.45783278e-02 -7.06494689e-01 3.41873139e-01
7.38106370e-01 -3.10551375e-02 8.14016521e-01 -1.08704448e+00
8.76346409e-01 1.34229851e+00 3.65866363e-01 5.36560655e-01
-8.10273528e-01 -2.79571086e-01 6.71939790e-01 3.23818475e-01
-1.04322779e+00 -4.48677391e-01 8.74109507e-01 -4.04404253e-01
1.26696432e+00 -1.39422819e-01 7.06945419e-01 1.04185641e+00
5.54729462e-01 1.05818141e+00 1.14695358e+00 -1.86976075e-01
9.29988176e-02 -1.20364226e-01 -1.21105500e-01 9.24450099e-01
2.25407481e-01 4.88208145e-01 -9.42585886e-01 1.19195171e-01
1.10626781e+00 1.30889624e-01 -4.09437537e-01 -7.17872322e-01
-1.19499767e+00 7.22191334e-01 1.22146644e-02 -4.29558307e-02
-1.22292034e-01 2.43121088e-01 2.45347336e-01 1.22296996e-01
1.00790071e+00 -1.00852758e-01 -7.85569549e-01 -8.70890692e-02
-9.40493345e-01 -8.42301250e-02 4.86704767e-01 9.92582262e-01
1.16019332e+00 1.44205824e-01 4.07492280e-01 5.97445965e-01
5.73997200e-01 3.29426557e-01 4.56646144e-01 -1.41550756e+00
5.64602435e-01 5.00302076e-01 -6.78823516e-03 -5.77772856e-01
-5.01104534e-01 -2.40555853e-01 -8.23224783e-01 5.47109902e-01
4.18854028e-01 -2.25512996e-01 -9.07068253e-01 1.27198172e+00
6.27565563e-01 1.41235113e-01 5.81940077e-02 1.00464642e+00
9.24456716e-01 3.46568823e-01 -6.85932517e-01 -2.80092925e-01
1.02911830e+00 -1.27759993e+00 -6.76137805e-01 -5.63656569e-01
4.12990540e-01 -5.57731569e-01 5.78331590e-01 1.04640317e+00
-1.29819226e+00 -7.07912505e-01 -1.17751873e+00 -2.92494476e-01
-3.89102772e-02 3.22053768e-02 7.85416603e-01 5.93991041e-01
-1.12282658e+00 5.83983600e-01 -1.18539870e+00 1.14917960e-02
3.48207772e-01 3.01668584e-01 -5.58821559e-01 -5.02524316e-01
-8.79469097e-01 6.60044014e-01 4.70656484e-01 3.20058353e-02
-1.10156405e+00 -4.51882064e-01 -1.41536248e+00 -5.48675776e-01
5.08811712e-01 -1.24297059e+00 1.08114922e+00 -7.35700011e-01
-1.91849875e+00 8.84550691e-01 -5.09038687e-01 -1.88584000e-01
7.22780764e-01 -6.96901441e-01 1.16925627e-01 6.43884242e-01
1.27440974e-01 9.03610110e-01 8.14452052e-01 -1.29664540e+00
-9.40265357e-01 -8.00907493e-01 4.57039207e-01 7.31271982e-01
5.05895354e-02 -4.56861198e-01 -7.49763429e-01 -1.24290563e-01
7.69533098e-01 -6.44753575e-01 -1.01135701e-01 2.32596412e-01
-4.06409681e-01 -2.56262068e-02 9.07097161e-01 -3.16906631e-01
7.52991080e-01 -1.69547033e+00 4.48051482e-01 -4.37022984e-01
3.60056758e-01 -2.97521800e-01 4.36313212e-01 -2.07452685e-01
2.26812735e-01 -4.21789587e-01 3.52877602e-02 -7.64808178e-01
-3.23638856e-01 3.70697796e-01 2.34616593e-01 8.55264843e-01
-1.54542506e-01 5.42092681e-01 -1.01118076e+00 -6.41330898e-01
8.81696880e-01 4.43448156e-01 -7.01404154e-01 4.21190262e-01
-1.71494797e-01 7.28279769e-01 -2.29027450e-01 9.04351592e-01
8.62291396e-01 -5.11259854e-01 6.23571426e-02 -2.19086185e-01
-9.82205123e-02 9.16366652e-03 -1.37528777e+00 2.31887054e+00
-7.09483802e-01 7.09468246e-01 1.07171617e-01 -5.26992440e-01
5.43052375e-01 2.19590306e-01 4.64737326e-01 -3.31257552e-01
-2.31507793e-02 2.27426514e-02 -5.95612407e-01 -4.39404726e-01
5.28790355e-01 -2.19544455e-01 3.72330546e-01 2.56891072e-01
1.11058570e-01 -5.39095461e-01 3.87284756e-02 5.32537028e-02
7.90826261e-01 7.78959453e-01 4.76048529e-01 1.53829142e-01
7.26410210e-01 -5.67402482e-01 7.72385299e-01 4.81577158e-01
-2.46464461e-01 1.00485408e+00 1.89066470e-01 -6.72960699e-01
-7.87369132e-01 -1.07319260e+00 -3.30334902e-01 6.23558879e-01
5.37277341e-01 -1.81789145e-01 -3.49150687e-01 -8.20792556e-01
-1.39511719e-01 2.14227200e-01 -6.78468645e-01 4.69972670e-01
-3.20049375e-01 -6.25694335e-01 7.73981363e-02 7.18739808e-01
7.94118464e-01 -5.95026374e-01 -9.55970168e-01 1.24944180e-01
-2.11315855e-01 -1.70674098e+00 -2.46950224e-01 2.77413994e-01
-1.31469512e+00 -1.06356168e+00 -8.53354573e-01 -3.23106349e-01
5.06260395e-01 5.30139506e-01 1.37493348e+00 -2.66536832e-01
1.67565867e-01 5.57268441e-01 -6.91247880e-02 -9.39728022e-02
7.63247535e-02 -4.48955297e-02 -2.03696266e-03 -1.84789807e-01
2.14535937e-01 -7.31797457e-01 -8.83666933e-01 5.75439274e-01
-8.54754746e-01 5.15838325e-01 2.39383802e-01 9.31485415e-01
8.66341293e-01 -1.30495861e-01 -1.29357399e-02 -8.58088255e-01
-5.55469155e-01 -2.75286674e-01 -8.97702992e-01 -1.10977612e-01
-4.86100793e-01 -7.01301768e-02 2.99843431e-01 -6.54581636e-02
-1.44156802e+00 3.72250229e-01 7.16014579e-02 -7.17172205e-01
-3.68468434e-01 1.64716437e-01 -2.92743742e-01 -1.59922019e-01
4.05692399e-01 1.24697432e-01 -2.26074964e-01 -2.25022435e-01
2.17399761e-01 3.95522952e-01 3.22165191e-01 -3.50027233e-01
5.12970984e-01 1.27047777e+00 1.64945453e-01 -7.22979307e-01
-1.14861369e+00 -7.04077423e-01 -1.15327573e+00 -4.21963692e-01
1.11954904e+00 -1.59460342e+00 -5.67699552e-01 9.08717215e-01
-1.30015111e+00 -3.85213166e-01 2.76726931e-01 7.89645076e-01
-7.87540793e-01 9.00656104e-01 -8.42902303e-01 -8.91232431e-01
2.84819640e-02 -1.08639455e+00 1.57358968e+00 1.05002068e-01
1.83486253e-01 -1.38513196e+00 -4.89918552e-02 7.22979605e-01
-1.64036199e-01 2.61417627e-01 1.12639301e-01 3.31009746e-01
-1.12272561e+00 2.59686619e-01 -2.26521865e-01 4.13503796e-01
1.99639425e-01 -1.24074072e-01 -1.48511970e+00 -3.19752455e-01
4.19278264e-01 -5.53209126e-01 1.04508960e+00 8.25825095e-01
1.12747192e+00 3.01543951e-01 -1.79567263e-01 1.20180786e+00
1.72270942e+00 1.62738919e-01 7.01244116e-01 5.31435609e-01
1.24562788e+00 5.94706059e-01 8.56868505e-01 6.87923491e-01
7.70916402e-01 5.81007838e-01 7.16821790e-01 3.12547050e-02
8.84067789e-02 -1.49119347e-02 3.62582207e-01 5.95869780e-01
-6.97523877e-02 -3.44593018e-01 -6.81288898e-01 5.15329421e-01
-1.77085257e+00 -8.64909887e-01 -1.91519380e-01 1.99108624e+00
4.68788356e-01 3.21952850e-01 -2.07164213e-01 1.24207176e-01
3.09073597e-01 4.69319403e-01 -7.93056250e-01 5.93324564e-02
-4.08588916e-01 -3.02435875e-01 7.26786315e-01 9.58748460e-01
-1.15642142e+00 8.85728359e-01 6.40881062e+00 4.24936354e-01
-9.39311743e-01 1.18468761e-01 9.53779161e-01 -3.61467630e-01
-3.57335687e-01 -9.71812233e-02 -1.08365262e+00 8.10737088e-02
4.22498167e-01 4.45171773e-01 9.26438496e-02 9.74033058e-01
6.25790730e-02 -1.01470709e+00 -1.40681469e+00 1.44754493e+00
4.57588404e-01 -1.28333771e+00 -2.40464196e-01 1.11673221e-01
1.16407323e+00 5.15958726e-01 -2.66008288e-01 -1.11333765e-01
2.08296731e-01 -6.96140587e-01 5.14314473e-01 3.88095468e-01
7.62134075e-01 -4.69117761e-01 7.47777045e-01 4.77790177e-01
-1.22366166e+00 1.47285834e-01 -4.02711570e-01 -4.54393834e-01
4.41748917e-01 8.84843469e-01 -3.13256174e-01 9.11850631e-01
8.55444729e-01 1.42838359e+00 -2.76459575e-01 6.14251256e-01
-3.51019979e-01 -1.38509274e-01 -3.70906323e-01 4.21272904e-01
2.09219500e-01 -7.02995360e-02 4.55578148e-01 6.18151903e-01
2.29280517e-01 1.65447369e-01 1.13392688e-01 5.20274103e-01
3.26305568e-01 -3.28028291e-01 -7.68887103e-01 6.34149790e-01
-1.05028756e-01 1.04540050e+00 -5.85787356e-01 -6.34712517e-01
-8.49191546e-01 1.21289456e+00 2.53379941e-01 3.00640255e-01
-5.70450246e-01 3.90758276e-01 4.03613687e-01 -3.79764475e-02
4.19277340e-01 -3.67572814e-01 -5.62409639e-01 -1.78504038e+00
1.05960794e-01 -3.65163028e-01 3.39155883e-01 -1.15334463e+00
-1.06695342e+00 6.07602417e-01 2.09028035e-01 -1.44857585e+00
-5.18334925e-01 -8.93778622e-01 -1.37395740e-01 5.59414744e-01
-2.19099951e+00 -7.73214161e-01 -8.28957438e-01 6.49326682e-01
9.49401855e-01 -6.97038919e-02 4.20109779e-01 1.73928365e-01
-3.42162639e-01 1.21550523e-01 -2.25886554e-01 -3.00103664e-01
6.98347151e-01 -1.34912837e+00 3.09316635e-01 7.81297445e-01
3.86432000e-02 1.07617527e-02 3.68487448e-01 -6.10691011e-01
-1.32132399e+00 -8.18220377e-01 4.17057216e-01 -7.28811562e-01
2.79869586e-01 -1.50327027e-01 -6.81854486e-01 6.90411031e-01
2.47387350e-01 3.53822082e-01 3.15259546e-01 -2.42784005e-02
-4.50287908e-02 4.26619872e-03 -1.05607069e+00 1.90428376e-01
1.24204636e+00 -5.98726571e-01 -2.91772395e-01 3.47952485e-01
5.45876443e-01 -1.12343395e+00 -8.60628963e-01 5.81426024e-01
5.99147320e-01 -1.74707305e+00 1.03204811e+00 2.03079611e-01
7.58987606e-01 -2.88233578e-01 -5.08991718e-01 -1.00821805e+00
1.24471538e-01 -4.07937109e-01 -4.44779336e-01 4.88685012e-01
-3.87915187e-02 -5.26444435e-01 1.36455107e+00 2.58651316e-01
-2.11582243e-01 -8.99553597e-01 -1.11719275e+00 -5.92124939e-01
-1.94528371e-01 -6.47532523e-01 6.54214248e-02 7.01488256e-01
-3.69998753e-01 3.62099558e-01 -5.79055250e-01 3.46450597e-01
1.01840425e+00 1.02732666e-01 9.79984224e-01 -1.17047787e+00
-7.28034914e-01 9.28828418e-02 -5.29575109e-01 -1.68426085e+00
2.81092614e-01 -2.90450186e-01 -5.15685938e-02 -1.48496699e+00
2.73537517e-01 1.93830431e-01 3.29810157e-02 5.11705354e-02
-1.88401416e-01 3.90271008e-01 -2.54335403e-01 -3.58226225e-02
-6.71725571e-01 7.29989707e-01 1.60303688e+00 1.37299612e-01
-4.09497246e-02 -1.39324024e-01 -9.16165859e-02 1.28118849e+00
4.37336653e-01 -2.30314627e-01 -7.49765158e-01 -8.04954231e-01
2.97715783e-01 6.47456110e-01 2.53624409e-01 -1.11418343e+00
1.87484965e-01 -2.57361501e-01 7.11313248e-01 -1.12179494e+00
8.90762746e-01 -6.91450953e-01 -1.66827247e-01 2.15371892e-01
2.51570344e-01 1.58452988e-02 6.92203194e-02 8.36258709e-01
-4.13346529e-01 -9.17768404e-02 8.59030664e-01 -5.29352963e-01
-9.99888480e-01 4.53440756e-01 -2.17633650e-01 -4.71482687e-02
7.82371521e-01 -7.47138739e-01 -1.21681161e-01 -6.08701110e-01
-7.18165338e-01 1.52482420e-01 7.24386096e-01 2.53085673e-01
1.07410026e+00 -1.09796274e+00 -4.74884212e-01 3.13791603e-01
8.07959586e-02 7.06150651e-01 4.65623915e-01 6.62456930e-01
-6.65681005e-01 4.68404204e-01 -1.43700108e-01 -1.39219689e+00
-1.23652875e+00 1.05913080e-01 6.38803303e-01 -3.72218788e-01
-6.36272073e-01 9.55097675e-01 8.81930172e-01 -3.97683680e-01
2.33995184e-01 -5.45942664e-01 -2.39989050e-02 -1.70078009e-01
4.55629557e-01 2.90149122e-01 -4.21393141e-02 -4.83146399e-01
-2.55786657e-01 1.00019920e+00 -5.76257296e-02 -3.40560794e-01
1.34689879e+00 -7.11316228e-01 9.17365104e-02 8.86481702e-01
1.22447562e+00 -3.30424786e-01 -1.99268103e+00 -2.42270932e-01
-5.56625366e-01 -8.25035512e-01 5.25023520e-01 -4.63525623e-01
-1.28615606e+00 8.93929541e-01 5.76045275e-01 -5.69967270e-01
1.10722542e+00 -2.37870961e-01 6.56135798e-01 4.47820753e-01
7.63377309e-01 -1.17251503e+00 5.14754891e-01 6.41138256e-01
5.13199866e-01 -1.90725446e+00 4.52983409e-01 -7.02434123e-01
-4.54319417e-01 1.25570941e+00 1.00726891e+00 -1.18282922e-01
8.40128481e-01 2.12958470e-01 1.72905684e-01 -1.73147634e-01
-9.98808086e-01 8.16437304e-02 2.79774845e-01 4.49577510e-01
2.29920372e-01 -4.29224730e-01 1.57983169e-01 -1.20391421e-01
1.76809222e-01 5.73462201e-03 7.79914558e-01 8.94792199e-01
-3.41364592e-01 -8.92452121e-01 -3.43426436e-01 1.89731851e-01
-4.43997741e-01 -1.02460980e-01 2.96494248e-03 8.18339407e-01
1.63422555e-01 1.02678084e+00 1.09986216e-01 -2.96956062e-01
-9.57598165e-02 -2.12833703e-01 8.78582776e-01 -8.40920210e-01
-1.29011735e-01 4.53789264e-01 1.59664586e-01 -1.01512241e+00
-1.03417742e+00 -8.23378623e-01 -9.51477945e-01 -6.05534129e-02
-4.11485523e-01 -5.65095425e-01 6.22296333e-01 1.25995994e+00
-4.35307845e-02 3.81482661e-01 6.86572731e-01 -1.37297642e+00
8.76599103e-02 -8.27122033e-01 -5.45359015e-01 -3.94372456e-02
5.45713127e-01 -8.32695961e-01 -7.58693457e-01 4.65013795e-02] | [8.792044639587402, -2.4210281372070312] |
e82a5951-8ab8-4801-818e-addb1081fe00 | buda-boundless-unsupervised-domain-adaptation | 2004.0113 | null | https://arxiv.org/abs/2004.01130v2 | https://arxiv.org/pdf/2004.01130v2.pdf | Handling new target classes in semantic segmentation with domain adaptation | In this work, we define and address a novel domain adaptation (DA) problem in semantic scene segmentation, where the target domain not only exhibits a data distribution shift w.r.t. the source domain, but also includes novel classes that do not exist in the latter. Different to "open-set" and "universal domain adaptation", which both regard all objects from new classes as "unknown", we aim at explicit test-time prediction for these new classes. To reach this goal, we propose a framework that leverages domain adaptation and zero-shot learning techniques to enable "boundless" adaptation in the target domain. It relies on a novel architecture, along with a dedicated learning scheme, to bridge the source-target domain gap while learning how to map new classes' labels to relevant visual representations. The performance is further improved using self-training on target-domain pseudo-labels. For validation, we consider different domain adaptation set-ups, namely synthetic-2-real, country-2-country and dataset-2-dataset. Our framework outperforms the baselines by significant margins, setting competitive standards on all benchmarks for the new task. Code and models are available at https://github.com/valeoai/buda. | ['Patrick Pérez', 'Tuan-Hung Vu', 'Matthieu Cord', 'Maxime Bucher'] | 2020-04-02 | null | null | null | null | ['universal-domain-adaptation'] | ['computer-vision'] | [ 3.55256349e-01 1.32087335e-01 -3.31432253e-01 -5.02361119e-01
-9.73232210e-01 -7.23860025e-01 5.59487462e-01 6.15983875e-03
-4.65457290e-01 8.55545402e-01 -1.15049861e-01 -8.15344080e-02
2.65893966e-01 -6.75578177e-01 -8.01439047e-01 -6.39995933e-01
3.84415388e-01 7.69771516e-01 7.20686495e-01 -1.95661396e-01
-8.27650353e-02 7.64425695e-02 -1.37448156e+00 3.64284605e-01
9.34911072e-01 8.68802845e-01 3.28846872e-01 4.28757489e-01
-3.22889775e-01 4.96041089e-01 -6.69985950e-01 -4.62272078e-01
3.71159673e-01 -3.57201606e-01 -1.01680374e+00 3.21656406e-01
3.95499527e-01 -2.94339135e-02 -1.47940949e-01 1.10722768e+00
5.44457257e-01 3.76104027e-01 8.13716769e-01 -1.19722939e+00
-7.29057372e-01 2.90452391e-01 -8.02812397e-01 1.72953695e-01
1.25514343e-01 3.19653988e-01 7.10097730e-01 -8.69576216e-01
9.23560560e-01 1.14919913e+00 5.73116183e-01 1.10921109e+00
-1.32383478e+00 -8.13192666e-01 4.80281085e-01 4.06472296e-01
-1.26845646e+00 -3.62670541e-01 9.03740346e-01 -6.34781480e-01
5.50010562e-01 -5.11795133e-02 1.86853737e-01 1.68072367e+00
-4.12525713e-01 9.50145066e-01 1.11369932e+00 -4.87608254e-01
4.73009646e-01 3.83450747e-01 3.13593537e-01 9.03992876e-02
2.35174466e-02 3.63013484e-02 -1.29969865e-01 2.65603423e-01
3.25418472e-01 -1.78992584e-01 -2.50595689e-01 -9.93839920e-01
-9.88221467e-01 7.72619903e-01 4.08819258e-01 1.85888559e-01
5.86617514e-02 -3.55125636e-01 6.24434590e-01 1.96681932e-01
7.69650161e-01 2.91717619e-01 -9.60425556e-01 4.42691259e-02
-6.64397061e-01 1.40829533e-01 6.56265974e-01 1.19251478e+00
9.92921591e-01 -3.57378051e-02 -2.54199654e-01 1.26313889e+00
-9.78475320e-04 3.93736541e-01 7.63292193e-01 -7.53445446e-01
6.19612217e-01 5.54347932e-01 1.70678124e-01 -3.75624031e-01
-3.81104082e-01 -5.83739400e-01 -5.73495567e-01 1.91488728e-01
7.29442358e-01 -3.44232500e-01 -1.47411501e+00 1.89101994e+00
7.23986030e-01 5.00275910e-01 3.52643669e-01 8.44532430e-01
8.28785956e-01 5.71934700e-01 2.72854716e-01 1.51597023e-01
1.21446157e+00 -1.21035469e+00 -3.80761772e-01 -4.83858794e-01
6.85097218e-01 -4.11826700e-01 1.49214828e+00 1.99385568e-01
-5.70519328e-01 -7.16596663e-01 -9.79232252e-01 1.17477313e-01
-8.69559884e-01 -9.77898538e-02 2.27375865e-01 5.58344722e-01
-6.56054616e-01 3.70166898e-01 -4.76549387e-01 -6.11459970e-01
8.24359059e-01 8.86998177e-02 -2.52954781e-01 -2.99845457e-01
-1.16448677e+00 7.22118020e-01 8.14496219e-01 -6.29767954e-01
-1.11086261e+00 -9.58268344e-01 -9.35901165e-01 -1.76467210e-01
6.61277652e-01 -5.04946232e-01 1.44238126e+00 -1.39626849e+00
-1.32076418e+00 1.23704469e+00 1.63509715e-02 -5.72897732e-01
6.14138305e-01 -5.20162471e-02 -6.76426470e-01 2.74134111e-02
5.07442176e-01 8.96079481e-01 8.75309646e-01 -1.64920890e+00
-9.62675035e-01 -3.77056420e-01 1.80842467e-02 2.05239564e-01
-2.83651203e-01 -2.39790067e-01 -5.89060724e-01 -7.08989084e-01
-1.67467654e-01 -8.84859562e-01 -1.93252966e-01 -2.63349235e-01
-3.74161094e-01 -1.25046104e-01 9.97646987e-01 -5.26557744e-01
9.12306845e-01 -2.25231004e+00 -3.99740599e-02 -5.87465093e-02
-4.61033033e-03 5.57980001e-01 -2.59794682e-01 -7.55327716e-02
-2.72617072e-01 -9.51642022e-02 -6.67684317e-01 -2.69246936e-01
-1.12177715e-01 4.44998801e-01 -1.86405867e-01 1.69846877e-01
3.86990577e-01 8.02346349e-01 -1.08093393e+00 -4.11517054e-01
1.95240676e-01 4.63784263e-02 -4.48195517e-01 8.16592202e-02
-4.80564237e-01 8.10075700e-01 -2.10145921e-01 7.20505476e-01
9.17539358e-01 -3.44579756e-01 1.06071897e-01 9.51271728e-02
2.00138405e-01 -1.62223175e-01 -1.18621695e+00 2.03292227e+00
-3.49509150e-01 3.08695436e-01 -2.27583095e-01 -1.18281519e+00
9.95717645e-01 9.43141654e-02 2.92699963e-01 -9.27985966e-01
6.95289224e-02 2.73037821e-01 -7.83955902e-02 -4.46727693e-01
2.94803411e-01 -2.56156981e-01 -3.41323674e-01 -1.57820340e-02
4.60116148e-01 6.48952127e-02 3.30610663e-01 1.73811466e-02
9.18649018e-01 5.05408585e-01 3.25489640e-01 -1.80875599e-01
4.81842071e-01 2.93681204e-01 9.27749097e-01 5.03505588e-01
-5.92859983e-01 8.01310301e-01 5.02546191e-01 -1.93610907e-01
-1.10443628e+00 -1.22910929e+00 -1.81451976e-01 1.38271177e+00
4.67663825e-01 1.80218816e-01 -8.26913416e-01 -1.27335429e+00
7.98226446e-02 1.00005746e+00 -8.49718034e-01 -2.58332193e-01
-3.26216221e-01 -5.83810687e-01 3.44883978e-01 4.70295876e-01
5.22497356e-01 -9.76631105e-01 -5.05263329e-01 1.91371232e-01
-1.94420248e-01 -1.29080772e+00 -3.18854660e-01 3.96506220e-01
-7.02378690e-01 -1.21162820e+00 -1.12913740e+00 -9.30809259e-01
5.82564414e-01 1.54745147e-01 1.14744258e+00 -6.77752614e-01
-1.81842074e-01 4.67743158e-01 -7.05112636e-01 -5.93418121e-01
-4.08157259e-01 2.02736050e-01 -2.35947058e-01 3.43117952e-01
6.23255074e-01 -4.35098350e-01 -5.03499091e-01 4.12406683e-01
-8.72024119e-01 6.88018501e-02 4.11589235e-01 9.00114894e-01
7.44228303e-01 -1.02228858e-01 8.20744693e-01 -1.27762580e+00
1.74778849e-01 -9.57072735e-01 -4.92770880e-01 2.71919280e-01
-4.75918919e-01 -8.98804329e-03 6.29399002e-01 -7.24847019e-01
-1.32009685e+00 2.17871919e-01 -1.02299713e-02 -6.83011770e-01
-6.89052224e-01 4.49811220e-02 -6.34988964e-01 2.60122836e-01
9.75416303e-01 1.02854259e-01 -2.67807245e-01 -7.99054682e-01
5.80783904e-01 5.82855105e-01 7.28212714e-01 -7.38682210e-01
7.63210952e-01 5.28074622e-01 -5.41835845e-01 -5.24899125e-01
-9.64468837e-01 -7.73308396e-01 -9.71510589e-01 1.89277325e-02
8.74928236e-01 -1.08624375e+00 2.88136959e-01 5.18063664e-01
-1.03818727e+00 -6.83710635e-01 -8.11011016e-01 2.07990751e-01
-5.29159904e-01 2.68928796e-01 -1.58934534e-01 -3.55153710e-01
2.39185020e-02 -1.01286268e+00 9.32550251e-01 3.57887745e-01
-7.07041174e-02 -1.06871963e+00 2.19370306e-01 2.58020043e-01
1.29590621e-02 4.39292222e-01 6.97964668e-01 -1.14597487e+00
-1.25788122e-01 2.17099227e-02 -3.97284657e-01 6.92654550e-01
1.71224684e-01 -4.05429274e-01 -1.22879958e+00 -3.00166488e-01
-2.31211871e-01 -4.41855133e-01 9.27481890e-01 2.51377076e-01
1.05222678e+00 -1.02039166e-01 -5.48051059e-01 7.47808039e-01
1.48720551e+00 3.80990922e-01 5.47004521e-01 5.95655143e-01
7.46413648e-01 7.30631948e-01 8.72735739e-01 4.66083139e-01
6.03926122e-01 7.24113882e-01 3.53151411e-01 -1.91211641e-01
-5.94101489e-01 -2.82678068e-01 4.97207269e-02 2.97894567e-01
4.03799295e-01 -2.27226958e-01 -1.26773822e+00 1.09443676e+00
-1.89240539e+00 -6.12785459e-01 3.37437727e-02 2.25093985e+00
9.06831205e-01 2.12630719e-01 3.88384730e-01 -1.58388704e-01
9.64690804e-01 -5.18933404e-03 -1.05679691e+00 -2.55730212e-01
-1.80101007e-01 4.03222442e-01 5.53500116e-01 3.13884735e-01
-1.52503908e+00 1.32879162e+00 4.95658350e+00 1.02833605e+00
-1.00747597e+00 5.86060762e-01 6.74712896e-01 6.76515028e-02
-4.16173451e-02 1.86228510e-02 -8.39308441e-01 6.14482641e-01
8.84537578e-01 -1.53672472e-01 1.89213045e-02 1.07153392e+00
-2.19146788e-01 5.44818826e-02 -1.02130759e+00 8.94064665e-01
2.81206388e-02 -9.75720644e-01 4.94290888e-02 -1.94404051e-01
9.64231491e-01 1.78275079e-01 5.07896096e-02 8.47903788e-01
4.99705732e-01 -5.12402773e-01 6.02888346e-01 1.65698871e-01
1.07901847e+00 -6.75399780e-01 4.87908840e-01 3.80294740e-01
-1.05297482e+00 -2.80117363e-01 -4.51424956e-01 1.77082956e-01
-1.17669299e-01 4.06107843e-01 -8.44784439e-01 6.85355425e-01
8.77314746e-01 9.23746467e-01 -7.68654108e-01 1.06507039e+00
-2.45051712e-01 5.55391550e-01 -7.97839276e-03 4.90677536e-01
2.17147529e-01 1.98326558e-02 5.02698481e-01 1.37938035e+00
1.75181597e-01 -1.04361944e-01 2.04230994e-01 7.99258471e-01
-1.29634455e-01 1.27088279e-01 -6.95478261e-01 3.83277833e-01
3.65640432e-01 1.13999116e+00 -8.02139342e-01 -6.27165437e-01
-5.26412189e-01 1.18730474e+00 2.85562307e-01 6.06935263e-01
-1.00979877e+00 -4.09399748e-01 7.21387923e-01 2.21840575e-01
4.94295686e-01 1.36333928e-01 -2.55192578e-01 -1.35377908e+00
-9.63905677e-02 -7.72018015e-01 6.96279824e-01 -5.97170830e-01
-1.49525011e+00 4.35069263e-01 2.18311414e-01 -1.51004481e+00
-1.59317613e-01 -6.48443639e-01 -3.43483001e-01 6.65599823e-01
-1.72779763e+00 -1.27724683e+00 -2.79706687e-01 9.05174434e-01
9.05980408e-01 -3.65564644e-01 7.08345830e-01 5.05272925e-01
-5.76691210e-01 8.55193257e-01 3.82913381e-01 2.72764891e-01
9.92243230e-01 -1.35738671e+00 5.19946635e-01 8.83133292e-01
8.84212330e-02 -1.03882499e-01 5.31545341e-01 -5.89231610e-01
-5.61300278e-01 -1.48021078e+00 3.90004277e-01 -7.09776521e-01
7.00960457e-01 -5.47459483e-01 -1.12022829e+00 7.58919597e-01
3.84116210e-02 2.94256270e-01 5.68022192e-01 -1.28685171e-02
-6.86769128e-01 -1.56169742e-01 -1.24609065e+00 4.77708906e-01
1.05634367e+00 -1.98604777e-01 -7.06114888e-01 3.42401385e-01
1.02175820e+00 -4.71399039e-01 -6.18836641e-01 6.11893773e-01
1.56518325e-01 -7.00662613e-01 1.06084526e+00 -8.47943664e-01
3.67974430e-01 -2.40257174e-01 -2.55810231e-01 -1.49239016e+00
-2.71440655e-01 -1.16249092e-01 -1.03553124e-01 1.46876669e+00
4.61378127e-01 -6.72551334e-01 8.64985585e-01 2.48271778e-01
-3.44211102e-01 -3.92499059e-01 -1.03405428e+00 -1.22270238e+00
5.09755373e-01 -4.42825556e-01 6.05542719e-01 1.25817525e+00
-4.47470754e-01 5.48605800e-01 -2.12630749e-01 1.65513262e-01
6.40978754e-01 -1.21178501e-03 8.50535512e-01 -1.28615618e+00
-3.51498932e-01 -3.24870944e-01 -2.71965891e-01 -8.98277104e-01
3.46852899e-01 -9.86407995e-01 7.45675936e-02 -1.36602378e+00
1.64931193e-01 -5.47766387e-01 -5.28603911e-01 7.13573456e-01
-2.17380524e-01 1.76049247e-01 2.15859324e-01 1.61294475e-01
-8.82538617e-01 5.71273386e-01 1.15080070e+00 -2.98362732e-01
-3.18532944e-01 6.60944581e-02 -7.27077305e-01 7.47608662e-01
1.04674888e+00 -5.58829665e-01 -5.67350328e-01 -4.28179681e-01
-3.57893914e-01 -3.26720476e-01 3.33170116e-01 -1.35906208e+00
-1.61103174e-01 -1.71573535e-01 2.78102994e-01 -3.06457430e-01
2.40227863e-01 -9.60473955e-01 -2.18651146e-01 8.05794522e-02
-3.23706537e-01 -5.34871638e-01 4.76011485e-01 8.12039375e-01
-1.82805359e-01 -2.60661930e-01 1.25458837e+00 -1.53893158e-01
-1.55613816e+00 3.59361827e-01 1.58717349e-01 5.78855276e-01
1.46441436e+00 -3.78671765e-01 -4.18249428e-01 4.54605296e-02
-1.10880303e+00 3.31557989e-01 5.33409476e-01 6.91121221e-01
3.67989272e-01 -1.24369860e+00 -7.49626696e-01 9.09449905e-02
7.21438169e-01 4.00097251e-01 5.03029466e-01 5.22513449e-01
-2.13638321e-01 4.92453165e-02 -2.95689583e-01 -7.25610793e-01
-1.01540065e+00 9.29814517e-01 4.05508518e-01 -3.35256100e-01
-4.99775052e-01 9.84637439e-01 6.83422327e-01 -9.00284767e-01
1.63940325e-01 -7.59870112e-02 -2.34322071e-01 1.62145123e-01
3.29850286e-01 1.84674621e-01 -1.44033253e-01 -6.06995583e-01
-3.47320527e-01 4.59862441e-01 -1.91876188e-01 5.85896112e-02
1.04930365e+00 -3.25617343e-01 4.64867145e-01 6.61632657e-01
1.15777183e+00 -3.11271966e-01 -1.66442966e+00 -6.24379277e-01
4.33778048e-01 -3.96614462e-01 -5.14102638e-01 -1.28707206e+00
-9.17255700e-01 1.05649996e+00 1.05063212e+00 -2.23152071e-01
1.20150924e+00 2.25513235e-01 7.54283309e-01 -8.35159719e-02
3.21148276e-01 -1.46878469e+00 1.98239416e-01 6.80148780e-01
5.23943245e-01 -1.65769291e+00 -4.21791345e-01 -2.34729484e-01
-9.85462487e-01 7.14467883e-01 1.05105674e+00 -5.69926901e-03
4.90669221e-01 -7.25879818e-02 1.67869911e-01 1.16661489e-01
-5.10457754e-01 -6.94172144e-01 2.81526893e-01 1.18572581e+00
1.36728249e-02 2.55401671e-01 7.07698837e-02 8.56888711e-01
1.90877184e-01 -6.13600016e-03 3.03168744e-01 8.00025642e-01
-4.06468570e-01 -1.23686194e+00 -3.48200887e-01 2.50624448e-01
-1.21126398e-01 5.34579009e-02 -3.03714961e-01 1.06741941e+00
6.05716765e-01 8.05675328e-01 1.17891401e-01 -3.93439323e-01
6.45873308e-01 3.66496980e-01 2.40418389e-01 -9.72550452e-01
-2.02436119e-01 4.22383100e-03 8.02487414e-03 -2.87536502e-01
-2.95958787e-01 -6.08906150e-01 -1.22833562e+00 1.09799005e-01
4.74447533e-02 -3.03538330e-02 4.21497762e-01 7.51960456e-01
4.39505637e-01 7.26508796e-01 4.38486338e-01 -5.30048192e-01
-4.72773284e-01 -8.93791854e-01 -6.31357670e-01 7.69177198e-01
1.82580054e-01 -9.01412606e-01 -1.49986222e-01 2.15850562e-01] | [9.816248893737793, 1.8266125917434692] |
ba651242-728b-40e8-9b07-260df449213f | pacs-a-dataset-for-physical-audiovisual | 2203.1113 | null | https://arxiv.org/abs/2203.11130v3 | https://arxiv.org/pdf/2203.11130v3.pdf | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning | In order for AI to be safely deployed in real-world scenarios such as hospitals, schools, and the workplace, it must be able to robustly reason about the physical world. Fundamental to this reasoning is physical common sense: understanding the physical properties and affordances of available objects, how they can be manipulated, and how they interact with other objects. Physical commonsense reasoning is fundamentally a multi-sensory task, since physical properties are manifested through multiple modalities - two of them being vision and acoustics. Our paper takes a step towards real-world physical commonsense reasoning by contributing PACS: the first audiovisual benchmark annotated for physical commonsense attributes. PACS contains 13,400 question-answer pairs, involving 1,377 unique physical commonsense questions and 1,526 videos. Our dataset provides new opportunities to advance the research field of physical reasoning by bringing audio as a core component of this multimodal problem. Using PACS, we evaluate multiple state-of-the-art models on our new challenging task. While some models show promising results (70% accuracy), they all fall short of human performance (95% accuracy). We conclude the paper by demonstrating the importance of multimodal reasoning and providing possible avenues for future research. | ['Louis-Philippe Morency', 'Ruslan Salakhutdinov', 'Paul Pu Liang', 'Peter Wu', 'Samuel Yu'] | 2022-03-21 | null | null | null | null | ['physical-commonsense-reasoning'] | ['reasoning'] | [ 5.33304870e-01 1.76085904e-01 3.88054959e-02 -1.51674315e-01
-7.27938652e-01 -6.60300553e-01 6.19000614e-01 3.32205683e-01
-1.28149286e-01 5.72208941e-01 4.81750309e-01 -2.10426360e-01
-4.81121719e-01 -6.24235570e-01 -7.98945308e-01 -3.79224002e-01
1.41520485e-01 2.07238451e-01 3.52160692e-01 -4.60456878e-01
2.57043362e-01 5.10815442e-01 -1.85381842e+00 6.74369395e-01
4.20941383e-01 1.14984071e+00 8.57309997e-02 1.15538967e+00
3.86332609e-02 1.57456815e+00 -5.22982836e-01 -4.82322305e-01
-3.30858111e-01 -3.77188295e-01 -1.19417500e+00 -2.89082766e-01
8.75005186e-01 -3.08852226e-01 -7.31946528e-01 9.34191585e-01
3.56180370e-01 6.10384047e-01 7.04672992e-01 -1.53876817e+00
-1.10646629e+00 4.47490364e-01 -4.85505871e-02 1.76664904e-01
1.17267334e+00 4.03480738e-01 1.37153232e+00 -2.71763057e-01
3.16477984e-01 1.57615864e+00 4.31734413e-01 6.22647822e-01
-9.09400582e-01 -5.36038458e-01 -1.85098812e-01 8.26442122e-01
-1.06978154e+00 -6.28799021e-01 7.35584140e-01 -4.63614464e-01
1.07664454e+00 7.08492219e-01 6.73443735e-01 1.35526204e+00
4.10023890e-02 8.51278186e-01 8.62334073e-01 -2.87969649e-01
3.92428577e-01 -2.58418351e-01 -2.20787823e-02 7.40302026e-01
-7.60053098e-02 -2.31819123e-01 -1.03993297e+00 -2.83984631e-01
5.19985139e-01 -2.10369915e-01 -2.57053077e-01 -1.98181555e-01
-1.79883361e+00 2.05265343e-01 5.23682654e-01 2.08206043e-01
-3.91217858e-01 8.67694199e-01 2.74839580e-01 3.82020622e-02
-4.39664543e-01 7.25098133e-01 -1.73485205e-01 -7.10727692e-01
-1.93338528e-01 4.84777033e-01 6.60909653e-01 7.33593166e-01
5.61181344e-02 -2.91989893e-01 -3.36002856e-02 7.83683300e-01
4.01557148e-01 1.00858355e+00 3.58955227e-02 -1.59597623e+00
5.44692576e-01 3.91313821e-01 1.55597508e-01 -1.23320544e+00
-4.50422287e-01 2.65112579e-01 -3.29361826e-01 -3.41449752e-02
5.25811255e-01 3.43597651e-01 -4.63574350e-01 1.78540063e+00
4.03244346e-01 4.26946044e-01 1.12485066e-01 9.52957571e-01
1.48056936e+00 5.32752872e-01 3.67547184e-01 2.87003070e-01
1.72774339e+00 -5.24879634e-01 -6.57269657e-01 -1.36153266e-01
2.97773659e-01 -7.89269030e-01 1.48708177e+00 7.34059453e-01
-1.09723544e+00 -4.09510791e-01 -9.18734610e-01 -4.10761535e-01
-4.73972559e-01 -4.41223502e-01 8.51241589e-01 4.69261080e-01
-4.62494284e-01 4.69944924e-01 -6.68780446e-01 -3.96439135e-01
4.97677922e-01 -4.49562408e-02 -4.98547852e-01 -1.56081453e-01
-1.27717042e+00 9.38070834e-01 1.26100019e-01 -1.16452828e-01
-6.91344261e-01 -7.75996029e-01 -1.01293576e+00 -5.86661063e-02
4.99303907e-01 -8.25335562e-01 1.44643033e+00 -1.43420890e-01
-1.23784089e+00 7.35333025e-01 2.30411842e-01 2.87919119e-02
2.28594109e-01 -4.15225327e-01 -7.65996456e-01 6.56636417e-01
1.19557129e-02 7.58246183e-01 5.16843081e-01 -1.27561748e+00
-3.86369646e-01 -1.51205331e-01 5.54347396e-01 5.34732305e-02
-1.83137760e-01 -7.81617537e-02 -2.65911804e-03 -3.56045932e-01
1.08546473e-01 -1.03516924e+00 3.73634130e-01 4.26824749e-01
-6.32179081e-01 -1.62306085e-01 6.64232016e-01 -4.58039016e-01
7.62084603e-01 -2.35326004e+00 3.25041980e-01 -1.24065265e-01
3.38755280e-01 -2.17771336e-01 -5.86612448e-02 5.82506657e-01
1.99299268e-02 -6.94174320e-02 -9.81869083e-03 -1.44190207e-01
6.37197196e-01 2.85748631e-01 -6.46197438e-01 3.89901102e-01
1.24721713e-01 8.90427709e-01 -1.34276581e+00 -7.10164011e-01
5.83808959e-01 6.28731728e-01 -3.38308692e-01 1.15724728e-02
-4.84064400e-01 3.99265975e-01 -5.36304057e-01 7.80037940e-01
1.60519183e-01 -3.27284902e-01 -2.52938569e-01 -3.27042103e-01
4.06539857e-01 4.42113340e-01 -1.24695790e+00 1.87725782e+00
-4.68792945e-01 8.08441043e-01 -2.29961991e-01 -5.75024962e-01
2.63066441e-01 3.73367190e-01 5.87576210e-01 -7.71517098e-01
7.10747987e-02 -1.02683522e-01 3.81524079e-02 -1.29905832e+00
4.87019330e-01 -4.30727750e-01 -4.59934622e-01 4.95462269e-01
-1.26008466e-01 -1.04172206e+00 5.98101467e-02 4.98325229e-01
1.42132890e+00 8.16255435e-02 2.57253200e-01 2.79933542e-01
2.97088534e-01 -6.47331476e-02 -8.24480429e-02 6.05604470e-01
-4.20938283e-01 5.08499384e-01 3.42724025e-01 -2.67158359e-01
-5.07157624e-01 -1.71748781e+00 -1.34575933e-01 1.29915226e+00
4.27733809e-01 -3.44152063e-01 -3.53398383e-01 -2.41426021e-01
7.41083845e-02 1.13364446e+00 -7.36188948e-01 -3.78193289e-01
-1.74305901e-01 -1.32202625e-01 8.52374375e-01 6.75433159e-01
4.57458586e-01 -1.47544158e+00 -1.12527478e+00 -7.58700371e-02
-7.97161877e-01 -1.68595719e+00 -7.53232241e-02 -1.99772298e-01
-3.50250930e-01 -1.37273920e+00 -4.88247313e-02 -1.43372014e-01
-1.18809052e-01 2.38720417e-01 1.25318205e+00 2.05526814e-01
-6.74257576e-01 1.12520301e+00 -3.42386693e-01 -6.37760401e-01
-5.74643791e-01 -6.05446577e-01 4.15064320e-02 -1.89665452e-01
1.64667234e-01 -6.06755316e-01 -4.12135273e-01 1.48081407e-02
-8.64597201e-01 6.17515221e-02 3.09074342e-01 2.57119745e-01
2.97294945e-01 2.33782716e-02 3.00441384e-01 -1.28640220e-01
4.85386789e-01 -5.12765288e-01 1.64790690e-01 3.12936693e-01
3.97865593e-01 -1.02671355e-01 4.25407320e-01 -6.12459958e-01
-9.67322826e-01 -1.88768014e-01 -2.07408741e-02 -1.71020910e-01
-6.29930258e-01 1.10984631e-01 -1.41965702e-01 8.98142755e-02
8.20483446e-01 -1.17613636e-02 -4.22314972e-01 -2.20407158e-01
7.02296138e-01 4.99829561e-01 1.10022461e+00 -1.09296370e+00
4.76836950e-01 6.95392311e-01 4.88502592e-01 -1.05552125e+00
-1.09904671e+00 -4.99918163e-01 -4.76024240e-01 -5.15900671e-01
1.15528262e+00 -5.89707673e-01 -1.60344076e+00 1.34091362e-01
-1.16163146e+00 -3.85566562e-01 -3.91813099e-01 5.94718874e-01
-9.55813766e-01 5.80170810e-01 -5.93169987e-01 -9.33967471e-01
2.57871121e-01 -9.68083262e-01 1.46419239e+00 -5.24001978e-02
-8.91319871e-01 -8.20930123e-01 3.45812440e-02 1.10399032e+00
8.64361599e-02 4.61840123e-01 1.16054320e+00 -2.29116231e-01
-4.85845238e-01 9.88213867e-02 -4.19534415e-01 9.80881751e-02
1.55362174e-01 1.55598700e-01 -1.31985176e+00 4.86458749e-01
-3.18134487e-01 -8.68217826e-01 6.82627916e-01 2.85085235e-02
1.35847592e+00 4.49109562e-02 -5.30207567e-02 1.05060302e-02
9.64134812e-01 6.00008480e-02 7.02895463e-01 2.83303633e-02
8.11959803e-01 6.02752090e-01 4.43876326e-01 5.12915373e-01
7.60550976e-01 5.38774252e-01 9.60549653e-01 3.90178233e-01
-4.38280284e-01 -1.30688056e-01 1.58402205e-01 5.72277963e-01
-2.89576173e-01 -2.17507675e-01 -1.09962201e+00 4.55391288e-01
-1.70091140e+00 -1.55838478e+00 -4.27167088e-01 1.71630216e+00
8.10761094e-01 4.66728956e-02 2.72456352e-02 6.12487912e-01
1.29401967e-01 -4.45644837e-03 -4.89133865e-01 -4.27741855e-01
1.35701254e-01 5.20615429e-02 -2.93490320e-01 3.91368330e-01
-1.08855760e+00 4.34900999e-01 6.57007885e+00 5.84313393e-01
-7.50490487e-01 -2.15408318e-02 -6.91197813e-02 -3.50487381e-01
-3.80554795e-01 -5.19383669e-01 -2.23154575e-02 3.23066056e-01
7.59807110e-01 5.21781385e-01 8.24739695e-01 4.42969441e-01
-3.30480337e-02 -4.73356098e-01 -1.51903498e+00 1.39601862e+00
1.73646107e-01 -1.13300717e+00 -1.82207674e-01 -2.14004084e-01
3.86495769e-01 -1.37951061e-01 1.20141558e-01 1.53438985e-01
3.67524058e-01 -1.24989831e+00 1.01597977e+00 9.28551674e-01
5.00957668e-01 -3.32117558e-01 1.48150265e-01 3.12200963e-01
-1.08017528e+00 -1.44835278e-01 6.03167266e-02 -3.20127487e-01
1.37414008e-01 3.56947243e-01 -4.64737743e-01 2.66817719e-01
8.90076816e-01 4.40408736e-01 -4.19301629e-01 7.57716596e-01
-2.56896287e-01 3.48901093e-01 -4.01582211e-01 -2.55154252e-01
-1.83917340e-02 5.13417304e-01 5.34298301e-01 1.14951789e+00
1.37443826e-01 5.66258371e-01 -6.06203750e-02 9.99435663e-01
3.11578680e-02 -2.20616385e-01 -6.46712840e-01 -3.00039381e-01
4.40141499e-01 9.16171670e-01 -4.99025613e-01 -3.35570663e-01
-2.04352528e-01 8.01121831e-01 -2.76599335e-03 1.91328943e-01
-1.27509034e+00 -3.62036824e-01 8.49137902e-01 -3.20201844e-01
-2.43353933e-01 -3.43463331e-01 -2.49677643e-01 -9.26029086e-01
1.68695018e-01 -8.20787787e-01 4.53201294e-01 -1.63232458e+00
-1.43630922e+00 2.15767045e-02 2.26315200e-01 -9.14264381e-01
9.21527594e-02 -8.44805062e-01 -2.05818921e-01 2.10909143e-01
-1.18596625e+00 -1.34722710e+00 -7.54457593e-01 8.02471340e-01
3.04493368e-01 2.77637541e-01 1.08299494e+00 9.50496867e-02
6.45864829e-02 -9.57747921e-03 -5.35653770e-01 -1.48680061e-01
6.01386189e-01 -1.24940681e+00 3.24887820e-02 1.75597534e-01
5.31841815e-01 4.96984631e-01 9.56443489e-01 -4.13587451e-01
-1.81869328e+00 -4.53078419e-01 6.22153342e-01 -1.12284720e+00
1.06162739e+00 -2.46978387e-01 -7.47308612e-01 4.92857069e-01
7.66490176e-02 1.22152410e-01 9.95637357e-01 2.34025016e-01
-8.85589480e-01 1.82904601e-01 -1.25604987e+00 7.49324322e-01
1.33817673e+00 -9.69559848e-01 -1.09866941e+00 5.46893060e-01
9.19510126e-01 -5.24756134e-01 -9.67345595e-01 1.46067604e-01
8.84421051e-01 -1.02582359e+00 1.43879008e+00 -8.95377994e-01
9.04516041e-01 -1.61976561e-01 -7.79825568e-01 -9.07524705e-01
-1.04376569e-01 -4.35050428e-01 -3.70077968e-01 8.44071746e-01
-4.93548065e-02 -3.01403910e-01 2.71649629e-01 8.94163668e-01
-8.48143324e-02 -5.37592232e-01 -9.84769523e-01 -7.83517897e-01
-1.85422152e-01 -1.36195874e+00 7.76049614e-01 9.87731516e-01
5.68269193e-01 3.06910098e-01 -1.04951732e-01 3.01048309e-01
5.81347406e-01 1.23617267e-02 6.76414847e-01 -1.20169199e+00
-5.09641707e-01 -6.36454225e-01 -7.43788362e-01 -6.86592638e-01
3.42489719e-01 -5.75774133e-01 3.03419009e-02 -1.76833963e+00
2.98088163e-01 -1.94672540e-01 -3.97562265e-01 6.35891438e-01
1.08673856e-01 4.12933230e-01 5.19658029e-01 -2.51432687e-01
-9.11697447e-01 4.21622932e-01 1.67020261e+00 -3.56313288e-01
2.24497408e-01 -3.91750902e-01 -6.24423802e-01 9.20185804e-01
5.00582576e-01 -1.19551457e-01 -5.39106965e-01 -3.19403201e-01
6.96268559e-01 3.03055912e-01 1.28660226e+00 -1.13753867e+00
1.80139646e-01 -7.53949702e-01 1.98914498e-01 -4.22911137e-01
1.16336250e+00 -1.06119609e+00 -4.81817573e-02 2.21933499e-01
-6.21906519e-01 -2.67149329e-01 2.90899724e-01 5.59488118e-01
1.69037938e-01 -8.12598616e-02 4.84116495e-01 8.68911147e-02
-9.95521724e-01 -2.96672106e-01 -2.33727500e-01 4.38759446e-01
9.89928961e-01 -1.08198531e-01 -6.25795901e-01 -6.07460082e-01
-9.08686101e-01 -2.78297141e-02 1.72260046e-01 6.36036217e-01
7.22644329e-01 -1.34237218e+00 -3.49909663e-01 -3.00713837e-01
5.00613987e-01 -3.56437653e-01 5.68609059e-01 6.39691472e-01
-3.58941823e-01 3.69204521e-01 -1.10665105e-01 -5.47580600e-01
-1.25186193e+00 6.16624773e-01 1.63650885e-01 5.71658969e-01
-6.91823900e-01 9.52276587e-01 4.22836319e-02 -3.16457957e-01
3.04597437e-01 -7.31188893e-01 9.34233069e-02 -3.47065747e-01
6.57787502e-01 5.72212756e-01 -2.69216210e-01 -6.51442409e-01
-7.31944561e-01 7.20843315e-01 7.99430728e-01 -2.64306545e-01
1.16660571e+00 4.64217700e-02 -1.59826919e-01 9.60027874e-01
9.73292947e-01 1.61062598e-01 -7.92985737e-01 -2.07022764e-02
-4.30226803e-01 -5.57211936e-01 -1.86658278e-01 -1.13925481e+00
-5.51409543e-01 1.20100808e+00 1.52428016e-01 3.98701847e-01
9.93076563e-01 6.36794388e-01 9.32219446e-01 8.89085174e-01
4.30807680e-01 -1.11546838e+00 7.60173857e-01 2.92556792e-01
1.15768707e+00 -1.31691778e+00 8.39761272e-02 -5.57754576e-01
-7.38765061e-01 1.17405295e+00 3.13318193e-01 2.29117244e-01
5.25068164e-01 1.69480577e-01 -1.73801497e-01 -6.77987933e-01
-7.84076929e-01 -1.09587684e-01 5.84466875e-01 6.75007582e-01
3.73101681e-01 3.91132712e-01 6.87881052e-01 4.96743947e-01
-6.52426422e-01 -8.10860097e-02 3.17489654e-01 8.04617584e-01
-4.23750818e-01 -3.79324555e-01 -6.43277228e-01 2.06475973e-01
-1.66962355e-01 1.12123035e-01 -5.19387782e-01 6.43978119e-01
2.91609615e-01 1.42170298e+00 -1.18383616e-01 -2.83626169e-01
5.76729834e-01 8.41210485e-02 1.01438165e+00 -4.52383429e-01
-1.67590618e-01 -6.65822804e-01 3.38232011e-01 -7.97325611e-01
-5.04310071e-01 -8.41432273e-01 -1.77521372e+00 -4.90928322e-01
4.94349636e-02 -5.08524179e-01 7.07385838e-01 1.23812211e+00
8.64407700e-03 9.64465559e-01 -1.42280325e-01 -6.49297297e-01
-5.37096262e-01 -5.17618775e-01 -4.32631999e-01 7.74422884e-01
6.30445540e-01 -1.02711618e+00 -3.07671547e-01 3.28578144e-01] | [10.47431755065918, 1.2280741930007935] |
7bede14f-f89f-416b-ab5e-d7a0979be0ef | enhancing-diversity-in-teacher-student | 2011.13776 | null | https://arxiv.org/abs/2011.13776v1 | https://arxiv.org/pdf/2011.13776v1.pdf | Enhancing Diversity in Teacher-Student Networks via Asymmetric branches for Unsupervised Person Re-identification | The objective of unsupervised person re-identification (Re-ID) is to learn discriminative features without labor-intensive identity annotations. State-of-the-art unsupervised Re-ID methods assign pseudo labels to unlabeled images in the target domain and learn from these noisy pseudo labels. Recently introduced Mean Teacher Model is a promising way to mitigate the label noise. However, during the training, self-ensembled teacher-student networks quickly converge to a consensus which leads to a local minimum. We explore the possibility of using an asymmetric structure inside neural network to address this problem. First, asymmetric branches are proposed to extract features in different manners, which enhances the feature diversity in appearance signatures. Then, our proposed cross-branch supervision allows one branch to get supervision from the other branch, which transfers distinct knowledge and enhances the weight diversity between teacher and student networks. Extensive experiments show that our proposed method can significantly surpass the performance of previous work on both unsupervised domain adaptation and fully unsupervised Re-ID tasks. | ['Francois Bremond', 'Benoit Lagadec', 'Hao Chen'] | 2020-11-27 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 3.34969908e-01 5.16511537e-02 -1.46256328e-01 -7.03318119e-01
-3.47304195e-01 -5.08594215e-01 5.94725072e-01 -1.18770264e-01
-7.03544974e-01 7.76380002e-01 5.34656122e-02 3.70780438e-01
-4.57791612e-02 -5.00226855e-01 -4.27382916e-01 -9.15115118e-01
3.72865140e-01 6.53218567e-01 1.14804417e-01 2.27906108e-02
6.91226348e-02 2.85882801e-01 -1.56938517e+00 -4.17198539e-02
1.05370462e+00 6.78775132e-01 -1.80632584e-02 1.39086694e-01
-8.79670158e-02 5.44372916e-01 -4.80713725e-01 -5.82206964e-01
3.27069253e-01 -5.07481754e-01 -8.41181993e-01 3.30977529e-01
6.25816464e-01 -2.46915206e-01 -2.05338240e-01 1.42504787e+00
5.69121540e-01 3.95077467e-01 9.53073978e-01 -1.13521516e+00
-9.23975110e-01 5.60833752e-01 -8.29354227e-01 -4.01041061e-02
-2.95347311e-02 -2.02774927e-01 6.69469416e-01 -7.61122823e-01
4.45304602e-01 1.09679687e+00 7.19997048e-01 1.01043880e+00
-1.51104069e+00 -9.58898008e-01 3.46235842e-01 2.86668807e-01
-1.59584916e+00 -4.45192784e-01 9.29677188e-01 -4.48486567e-01
4.08312529e-01 -1.47311628e-01 3.04845840e-01 1.16338813e+00
-7.39616513e-01 7.13715136e-01 1.31354392e+00 -5.75628340e-01
-8.94460380e-02 5.78051805e-01 4.47482586e-01 6.23382926e-01
1.95695907e-01 1.00474574e-01 -5.74187458e-01 -6.60005659e-02
5.61561644e-01 1.68495238e-01 -5.02046049e-02 -4.84214455e-01
-9.62108552e-01 4.36709464e-01 3.83718699e-01 2.77978212e-01
-7.68724531e-02 -2.28397578e-01 3.94753456e-01 2.66756237e-01
2.89510369e-01 1.56904489e-01 -2.48384804e-01 1.86571926e-01
-7.82484829e-01 -5.59455417e-02 5.04884839e-01 7.87886679e-01
1.01591277e+00 -9.53380615e-02 -1.94809645e-01 1.23028457e+00
2.29271844e-01 3.93045008e-01 7.26132810e-01 -6.63863897e-01
2.66241699e-01 7.47119188e-01 1.45866975e-01 -8.20103645e-01
-2.75591552e-01 -5.95316589e-01 -1.14629579e+00 2.44029596e-01
6.32540464e-01 -1.54421136e-01 -1.06471622e+00 1.87844253e+00
3.92365247e-01 6.04536057e-01 7.65089691e-02 1.01540053e+00
7.23842800e-01 1.40063673e-01 2.67357439e-01 -1.01947084e-01
1.21109164e+00 -1.25067067e+00 -3.72170717e-01 -3.71195599e-02
5.18214881e-01 -4.12974507e-01 6.30219936e-01 3.65715474e-01
-5.17616987e-01 -1.01295936e+00 -1.01622593e+00 2.40130275e-01
-2.47465983e-01 5.07644475e-01 2.35173732e-01 7.73568392e-01
-8.37507129e-01 5.50369561e-01 -4.40270483e-01 -3.41070801e-01
4.34549600e-01 6.71980500e-01 -6.37858391e-01 -5.34150451e-02
-9.91414607e-01 5.73107541e-01 5.95324397e-01 1.13086604e-01
-9.37420547e-01 -3.65349531e-01 -5.64458489e-01 4.67910208e-02
2.89206654e-01 -4.26396132e-01 1.04882431e+00 -1.56208313e+00
-1.62511575e+00 1.08870018e+00 -2.10940525e-01 -2.42864355e-01
3.43973041e-01 -1.14738710e-01 -5.07067382e-01 -2.84921527e-02
2.00815603e-01 7.20627844e-01 1.00422788e+00 -1.53061974e+00
-8.02918553e-01 -5.91246903e-01 -3.21137398e-01 3.63417774e-01
-7.77519107e-01 -9.67767462e-02 -3.27408671e-01 -6.58531964e-01
6.07647076e-02 -1.08041894e+00 -1.33150905e-01 -3.28198344e-01
-3.33727866e-01 -5.98255694e-01 6.62530899e-01 -6.07079089e-01
9.73791599e-01 -2.25245953e+00 1.13487326e-01 4.83625054e-01
3.57817680e-01 4.35115337e-01 -1.80145964e-01 1.76184624e-02
-2.44245499e-01 -1.17662385e-01 -1.08004212e-01 -5.76248825e-01
-1.93012074e-01 2.46968925e-01 -3.90936136e-02 4.64818895e-01
-3.86186386e-03 4.72854793e-01 -8.48621666e-01 -5.78143239e-01
3.41970325e-02 2.64543742e-01 -3.08060497e-01 3.38363051e-01
2.37880543e-01 9.20124650e-01 -3.92649949e-01 4.42296833e-01
8.75670254e-01 -2.21932292e-01 4.08543766e-01 -2.37337828e-01
3.88896950e-02 -2.56753862e-02 -1.36089349e+00 1.71462846e+00
-1.70203254e-01 2.41282672e-01 -5.53323589e-02 -1.57457840e+00
1.16866505e+00 2.77360857e-01 3.79484653e-01 -5.37315130e-01
1.34893954e-01 5.79601787e-02 -1.25750273e-01 -2.70715773e-01
2.19143540e-01 -9.95765626e-02 1.22903861e-01 5.98639727e-01
4.43683118e-01 8.12181592e-01 -1.93105289e-03 -4.82721515e-02
4.88238662e-01 2.40477264e-01 1.00842319e-01 -1.93633080e-01
9.60375249e-01 -2.26571530e-01 9.32753980e-01 8.49222958e-01
-3.53326410e-01 5.34195364e-01 -1.10583745e-01 -5.28535783e-01
-9.33490634e-01 -1.07623279e+00 -1.24310534e-02 1.43056929e+00
5.19912064e-01 -4.25279280e-03 -8.01257432e-01 -1.14980876e+00
1.10833556e-03 2.34638378e-01 -5.50077319e-01 -1.13017231e-01
-5.07172406e-01 -7.12858915e-01 6.15181029e-01 5.90853095e-01
7.87474275e-01 -8.68158460e-01 5.01001664e-02 1.33926168e-01
-2.56308585e-01 -9.51092482e-01 -4.89765465e-01 1.93376034e-01
-7.43226409e-01 -9.53804135e-01 -1.06698477e+00 -1.18045592e+00
1.23890722e+00 3.77086937e-01 7.48599052e-01 1.62737295e-02
-8.62422958e-02 3.70217800e-01 -3.51535082e-01 -1.31914631e-01
-1.86593235e-01 1.92289442e-01 5.94158351e-01 6.01725817e-01
8.41805995e-01 -8.54884267e-01 -5.09456754e-01 5.22044718e-01
-6.00981772e-01 -9.90007520e-02 4.98641163e-01 1.21228158e+00
5.50411224e-01 7.65617415e-02 7.51295090e-01 -9.27362263e-01
4.13064599e-01 -1.68455780e-01 -4.83631104e-01 6.36428595e-01
-8.91107082e-01 4.57561731e-01 8.24737608e-01 -8.13745618e-01
-1.52530909e+00 5.02930701e-01 2.33161569e-01 -2.91973859e-01
-5.43382704e-01 5.26269060e-03 -1.60119355e-01 -2.62443841e-01
7.00588465e-01 3.83486688e-01 -3.97634991e-02 -8.46751750e-01
2.93815792e-01 9.85711932e-01 8.42610836e-01 -8.36221635e-01
1.08082569e+00 6.14734411e-01 -2.54137933e-01 -3.82212400e-01
-9.29534256e-01 -7.13521481e-01 -1.06571937e+00 -8.16237330e-02
7.22699344e-01 -1.13188136e+00 -8.11055779e-01 7.17423797e-01
-9.14461792e-01 6.45662174e-02 -1.81141168e-01 4.65246171e-01
-2.18346063e-02 6.21925473e-01 -3.54242444e-01 -7.87888229e-01
-3.16218287e-01 -8.91934812e-01 7.16645777e-01 8.61635566e-01
1.49553875e-02 -8.50299120e-01 1.35627553e-01 4.03681219e-01
5.12009524e-02 -2.47613892e-01 5.16503751e-01 -1.11251104e+00
-1.02757320e-01 -3.26379249e-03 -4.54486310e-01 5.71254373e-01
3.91528040e-01 -5.49371004e-01 -1.21708834e+00 -4.66389030e-01
-2.19600871e-01 -5.35337567e-01 9.45831597e-01 -9.80023816e-02
1.02496159e+00 -1.81448519e-01 -5.54913878e-01 6.96680784e-01
1.25795019e+00 -2.01039985e-02 2.03556076e-01 5.10405481e-01
9.48092639e-01 7.88587868e-01 4.02991682e-01 4.62042958e-01
4.94349718e-01 6.91561997e-01 -2.92304680e-02 -3.24695647e-01
-2.32525729e-03 -3.98924112e-01 3.11085343e-01 6.77389860e-01
-3.79833668e-01 1.44555151e-01 -6.70099497e-01 6.46589339e-01
-1.90843332e+00 -8.21715415e-01 1.72300115e-01 2.25677514e+00
9.65616941e-01 -1.77980945e-01 2.96451032e-01 -6.03367537e-02
1.10443461e+00 -1.81619629e-01 -5.91078579e-01 5.19774901e-03
-7.42684072e-03 9.22074988e-02 6.35915339e-01 2.17632636e-01
-1.32625532e+00 9.42062616e-01 5.68133068e+00 9.84419763e-01
-1.00418127e+00 2.66597927e-01 4.34247762e-01 2.04190493e-01
9.74390209e-02 1.25697218e-02 -1.14900303e+00 5.20002782e-01
5.64087152e-01 -8.58515874e-03 4.14892048e-01 1.03134549e+00
-2.13287681e-01 -1.82515848e-02 -1.06426227e+00 1.36795545e+00
3.06477398e-01 -7.53589869e-01 9.72648337e-02 -5.11557497e-02
1.06330907e+00 -2.86174297e-01 9.15987417e-03 2.03107432e-01
7.12294161e-01 -9.13245678e-01 2.98083633e-01 5.13972878e-01
9.34031785e-01 -9.09828901e-01 7.86020279e-01 4.48976666e-01
-1.10103416e+00 -2.25792915e-01 -5.47013938e-01 8.37325975e-02
1.48725035e-02 3.58138055e-01 -8.23908806e-01 5.93141317e-01
8.36542785e-01 7.88965344e-01 -6.97824121e-01 8.45452368e-01
-2.89627314e-01 6.00179732e-01 -2.87007153e-01 3.15828532e-01
-4.31351438e-02 -3.64102185e-01 3.46525133e-01 1.08812213e+00
1.12213403e-01 -1.78481974e-02 4.74150658e-01 5.94391644e-01
-1.34460256e-01 -5.21461144e-02 -3.35707337e-01 1.41369656e-01
3.91855121e-01 1.24420595e+00 -4.68888581e-01 -5.44845283e-01
-4.47991222e-01 1.45758355e+00 6.37295723e-01 5.17622173e-01
-5.73321879e-01 -3.11755657e-01 4.74795371e-01 -1.13716751e-01
2.27908328e-01 -1.16759688e-01 -8.41379073e-03 -1.25822854e+00
-8.42671692e-02 -7.47229517e-01 6.45311654e-01 -3.26617271e-01
-1.73809910e+00 4.01694536e-01 -1.57128677e-01 -1.32199216e+00
-1.51713207e-01 -4.90800858e-01 -4.14533019e-01 7.61412084e-01
-1.61985266e+00 -1.33312547e+00 -5.64582705e-01 8.36556792e-01
1.98730692e-01 -6.12869263e-01 9.15689170e-01 4.73220974e-01
-7.06597924e-01 1.15549970e+00 5.14948189e-01 5.82997501e-01
1.31140900e+00 -1.17653263e+00 -1.03758968e-01 7.22823620e-01
3.66287678e-01 6.65952682e-01 2.59969711e-01 -6.07987821e-01
-7.71865785e-01 -1.01453400e+00 6.46036685e-01 -2.49036327e-01
2.42977768e-01 -2.12603509e-01 -8.93886268e-01 6.27327740e-01
-9.20019969e-02 -7.98971280e-02 8.69927943e-01 3.60070825e-01
-6.29321337e-01 -4.86217827e-01 -1.13887167e+00 4.52237040e-01
1.16564941e+00 -4.79958802e-01 -6.33964419e-01 6.95111305e-02
2.00867072e-01 1.20373322e-02 -7.54089057e-01 3.28229606e-01
6.39927208e-01 -8.40661108e-01 9.73811984e-01 -5.35884619e-01
-8.02371353e-02 -5.44951320e-01 3.41144174e-01 -1.32474041e+00
-4.48859990e-01 -4.01797205e-01 2.40380660e-01 1.83552849e+00
6.13880642e-02 -6.80794716e-01 9.84569967e-01 5.92865229e-01
2.83898205e-01 -1.03704445e-01 -7.70903349e-01 -9.33714628e-01
-1.86223269e-01 2.16624156e-01 5.18286407e-01 1.19470429e+00
-1.22966692e-01 3.84562582e-01 -7.00121284e-01 2.83162326e-01
1.00661194e+00 -2.44731177e-02 8.21013391e-01 -1.59784544e+00
-1.92049921e-01 -1.16895847e-01 -3.53713304e-01 -1.28547621e+00
4.34950411e-01 -1.08047581e+00 5.11805564e-02 -9.85234678e-01
5.87860465e-01 -7.63147473e-01 -7.54544377e-01 6.42016530e-01
-3.40966940e-01 4.23730880e-01 2.02928614e-02 4.54973817e-01
-7.52965093e-01 5.38298905e-01 9.64189947e-01 -3.36906761e-01
-3.00763369e-01 -8.81595258e-03 -7.52793372e-01 7.22730935e-01
6.98566258e-01 -7.56874263e-01 -3.83215427e-01 -4.10155863e-01
-3.36897045e-01 -4.73829508e-01 4.41065311e-01 -1.06387126e+00
5.37274539e-01 -2.05612760e-02 7.24979997e-01 -2.66910821e-01
1.24895386e-01 -9.06820416e-01 -3.71348187e-02 1.06829688e-01
-3.65506232e-01 -3.79067719e-01 -2.35945702e-01 6.91094518e-01
-3.02205801e-01 -5.25889277e-01 8.85604620e-01 -2.15839714e-01
-9.20079470e-01 2.53083825e-01 -1.18891478e-01 -6.19826466e-02
7.83097446e-01 -3.89317214e-01 -7.14737251e-02 -2.30314270e-01
-7.89304376e-01 3.16577256e-01 5.17596126e-01 3.61848503e-01
3.63051444e-01 -1.38326192e+00 -7.02479899e-01 3.17431360e-01
1.56984270e-01 -1.33387093e-02 3.97795051e-01 6.40337586e-01
1.20326012e-01 5.82304150e-02 -4.69481736e-01 -6.82196796e-01
-1.44357407e+00 5.50314009e-01 3.22301537e-01 -3.68481904e-01
-4.13319677e-01 9.36694920e-01 4.23536688e-01 -7.22559988e-01
3.61337304e-01 5.02175450e-01 -4.89143461e-01 8.15083236e-02
5.63318551e-01 3.26213330e-01 -3.85405213e-01 -8.73680353e-01
-2.73110032e-01 8.83164346e-01 -4.98089164e-01 -6.27687722e-02
1.18956399e+00 -2.82626450e-01 -4.88360450e-02 1.15946624e-02
1.03580916e+00 -1.17327958e-01 -1.19045067e+00 -6.94001913e-01
1.18950695e-01 -2.54268318e-01 -4.16862935e-01 -7.25668252e-01
-8.62930954e-01 7.84208059e-01 1.09385693e+00 -3.14927220e-01
1.22081769e+00 -1.94131270e-01 7.79495001e-01 5.43411672e-01
2.54880369e-01 -1.58782589e+00 2.04826891e-01 3.38759780e-01
2.24382922e-01 -1.51506591e+00 -8.81083123e-03 -2.72127092e-01
-7.30245054e-01 9.65721607e-01 1.04077590e+00 -7.77067170e-02
4.72879201e-01 -8.90638977e-02 1.77276120e-01 2.21574694e-01
-5.83681688e-02 -4.22912180e-01 3.85847181e-01 9.46107924e-01
1.66519731e-01 -6.98442105e-03 -2.87229002e-01 8.99215519e-01
1.37721851e-01 -5.47340922e-02 5.67441061e-02 5.20449579e-01
-3.33358467e-01 -1.55028057e+00 -5.16128063e-01 1.69043884e-01
-3.41334790e-01 3.32976580e-02 -4.10674185e-01 4.31340009e-01
5.52541375e-01 8.68678629e-01 -3.63397598e-03 -5.58752894e-01
2.07773671e-02 3.05713952e-01 4.53947991e-01 -6.08224690e-01
-5.31746566e-01 -7.39227235e-02 -1.26183778e-01 -9.89491865e-02
-7.76080668e-01 -4.15601939e-01 -1.12706363e+00 1.09151900e-01
-5.16402125e-01 3.35875779e-01 3.47061992e-01 1.06782770e+00
3.11388433e-01 2.10253745e-01 7.48273194e-01 -6.95985734e-01
-7.42044151e-01 -1.04736614e+00 -6.85115099e-01 8.52717161e-01
2.73883075e-01 -8.95487189e-01 -3.86976928e-01 4.89416361e-01] | [14.80106258392334, 1.0843011140823364] |
d55522bb-6430-4e7d-a5c0-f8c8eed5aef0 | knowledge-distillation-for-end-to-endperson | 1909.01058 | null | https://arxiv.org/abs/1909.01058v2 | https://arxiv.org/pdf/1909.01058v2.pdf | Knowledge Distillation for End-to-End Person Search | We introduce knowledge distillation for end-to-end person search. End-to-End methods are the current state-of-the-art for person search that solve both detection and re-identification jointly. These approaches for joint optimization show their largest drop in performance due to a sub-optimal detector. We propose two distinct approaches for extra supervision of end-to-end person search methods in a teacher-student setting. The first is adopted from state-of-the-art knowledge distillation in object detection. We employ this to supervise the detector of our person search model at various levels using a specialized detector. The second approach is new, simple and yet considerably more effective. This distills knowledge from a teacher re-identification technique via a pre-computed look-up table of ID features. It relaxes the learning of identification features and allows the student to focus on the detection task. This procedure not only helps fixing the sub-optimal detector training in the joint optimization and simultaneously improving the person search, but also closes the performance gap between the teacher and the student for model compression in this case. Overall, we demonstrate significant improvements for two recent state-of-the-art methods using our proposed knowledge distillation approach on two benchmark datasets. Moreover, on the model compression task our approach brings the performance of smaller models on par with the larger models. | ['Fabio Galasso', 'Sikandar Amin', 'Bharti Munjal'] | 2019-09-03 | null | null | null | null | ['person-search'] | ['computer-vision'] | [ 7.35172778e-02 4.02756482e-02 -1.15501307e-01 -1.81827620e-01
-1.16273129e+00 -4.95501637e-01 4.67664242e-01 2.56122887e-01
-1.09267700e+00 4.56849664e-01 -7.00786635e-02 4.48478870e-02
-3.92809331e-01 -4.30677265e-01 -7.40239501e-01 -7.03551710e-01
3.27395290e-01 1.30055475e+00 5.07023633e-01 7.38563836e-02
-9.68131870e-02 2.89163232e-01 -1.78263795e+00 -2.70907376e-02
8.31242561e-01 7.85620987e-01 1.69347823e-01 9.65646327e-01
4.53002632e-01 3.63766938e-01 -5.60483158e-01 -9.11133230e-01
5.23849845e-01 -1.78460583e-01 -8.77496481e-01 1.41215697e-01
1.18559003e+00 -4.25702989e-01 -6.52658463e-01 8.37267339e-01
1.12418520e+00 4.13796902e-01 4.34973031e-01 -9.43852961e-01
-4.07537997e-01 5.80283105e-01 -6.46569312e-01 4.72016901e-01
4.26005810e-01 3.24863732e-01 9.26499248e-01 -8.63669753e-01
2.57902682e-01 1.50389695e+00 6.89657390e-01 7.07702041e-01
-1.40551782e+00 -7.08301723e-01 1.42305881e-01 3.95585746e-01
-1.63017738e+00 -4.57642049e-01 1.16099186e-01 -3.29964072e-01
1.15525770e+00 3.44878197e-01 6.14435256e-01 7.56179333e-01
-8.86101902e-01 1.12783587e+00 8.14242542e-01 -7.35878348e-01
-1.25617251e-01 4.49704617e-01 4.07818168e-01 1.10538709e+00
3.58066320e-01 4.00407553e-01 -7.68819988e-01 -3.95474613e-01
5.16348839e-01 -1.53723210e-01 -2.17745960e-01 -5.32015562e-01
-8.03746462e-01 7.20111310e-01 1.36819124e-01 -5.31405322e-02
-2.70797670e-01 1.86925977e-01 3.52288902e-01 1.05811357e-01
1.84502140e-01 1.57272577e-01 -5.91735601e-01 -1.98777393e-01
-1.31874466e+00 7.29045868e-01 9.28277373e-01 7.20546305e-01
5.60372591e-01 -4.91625220e-01 -7.68726051e-01 8.99433494e-01
9.89165306e-02 5.44478297e-01 3.73081565e-01 -6.96502745e-01
5.39187372e-01 5.30089378e-01 6.80203140e-02 -4.90487218e-01
-2.44864792e-01 -8.73186588e-01 -3.53606701e-01 -8.51843730e-02
7.04299390e-01 -1.45380452e-01 -1.11663866e+00 1.85267508e+00
7.98525989e-01 6.68905020e-01 -9.23388898e-02 7.58020103e-01
7.70549238e-01 6.49033859e-02 1.25818178e-01 1.63697638e-02
1.91711080e+00 -1.45725799e+00 -2.40963131e-01 -2.43289039e-01
6.13161683e-01 -4.32984382e-01 6.41441286e-01 4.14045185e-01
-1.40831304e+00 -7.91355491e-01 -7.71515131e-01 -1.80215552e-01
-3.10872048e-01 5.24403095e-01 3.34821820e-01 1.04726160e+00
-1.00994503e+00 4.93322730e-01 -7.82761633e-01 -4.56575274e-01
4.72389966e-01 7.50047326e-01 -3.22955132e-01 1.80263817e-02
-9.26515818e-01 9.12248373e-01 6.20551467e-01 -3.97979915e-01
-8.10773730e-01 -1.04652524e+00 -6.73459113e-01 4.00432974e-01
9.92629290e-01 -1.18559730e+00 1.54737866e+00 -2.41248846e-01
-1.19532537e+00 1.17271888e+00 -2.87167728e-01 -7.35834062e-01
7.77668178e-01 -4.99307394e-01 1.04154699e-01 2.25037873e-01
7.95871392e-02 6.48896575e-01 8.43880892e-01 -1.02755249e+00
-1.14781904e+00 -5.54037273e-01 -1.25103220e-01 4.02500093e-01
-4.70862895e-01 2.71093458e-01 -1.30747616e+00 -4.48773861e-01
-3.41982365e-01 -9.56888318e-01 -1.93089247e-01 -4.34193760e-02
-4.34134483e-01 -7.64474154e-01 4.95313227e-01 -7.32300341e-01
1.46120489e+00 -1.93112910e+00 2.25584775e-01 2.93692172e-01
2.85082042e-01 7.28260815e-01 -5.18256575e-02 1.94388166e-01
1.08786318e-02 -1.71594962e-01 1.30087268e-02 -1.17254698e+00
1.86247319e-01 8.66088271e-02 7.85235986e-02 5.13570666e-01
-2.18054116e-01 8.97841692e-01 -7.82551050e-01 -6.79145038e-01
4.42191847e-02 3.48665297e-01 -7.74494350e-01 3.12811852e-01
3.98007259e-02 2.12912545e-01 -2.50523418e-01 4.46540892e-01
5.19932985e-01 -2.79361129e-01 -7.78261349e-02 6.59102798e-02
1.96038008e-01 2.12222487e-01 -1.61852288e+00 1.70289993e+00
-1.17359541e-01 1.57155156e-01 9.14690495e-02 -1.02096128e+00
3.21171194e-01 1.45519793e-01 3.06543112e-01 -4.52097952e-01
-1.75451055e-01 1.06807210e-01 -6.49498478e-02 -2.95193106e-01
4.12596345e-01 3.52972925e-01 2.88583618e-02 2.84624159e-01
4.00507331e-01 3.04905057e-01 5.38438797e-01 3.20006371e-01
9.95743811e-01 -1.54208625e-02 2.11289153e-01 -4.25078981e-02
9.26253021e-01 -3.15168560e-01 5.06671011e-01 1.38203835e+00
-1.71609651e-02 2.72134751e-01 -4.56424011e-03 -2.95999497e-01
-8.06609154e-01 -9.05916274e-01 -2.61105858e-02 1.61066890e+00
-6.27832115e-02 -6.11759722e-01 -1.07075715e+00 -9.39672291e-01
5.07044375e-01 4.35764313e-01 -5.32839298e-01 -1.26855999e-01
-7.86453605e-01 -5.79221666e-01 9.06963110e-01 5.23402333e-01
5.10116935e-01 -6.63648605e-01 -7.70761192e-01 5.23502305e-02
-9.60298181e-02 -1.30982888e+00 -8.13872337e-01 1.95426404e-01
-4.73546684e-01 -1.09882379e+00 -8.87505054e-01 -8.10728014e-01
7.07321107e-01 4.79213893e-02 1.24447477e+00 2.84004599e-01
-6.09428346e-01 7.59841144e-01 -3.98793593e-02 -5.87965906e-01
-1.94890976e-01 3.12905371e-01 3.48325312e-01 -1.87262654e-01
6.74055219e-01 -1.56581923e-01 -6.94066346e-01 3.30423981e-01
-4.43817884e-01 -3.46331865e-01 5.26511431e-01 8.76984119e-01
6.87601566e-01 2.30284452e-01 -6.61349967e-02 -6.40008330e-01
4.27898109e-01 6.99966028e-02 -7.92942703e-01 5.42145550e-01
-9.11271632e-01 4.01823699e-01 2.32092664e-01 -5.80377340e-01
-8.06932509e-01 3.45078886e-01 8.68743286e-02 -5.88952124e-01
-1.96827263e-01 3.14922109e-02 1.07185103e-01 -2.44489923e-01
6.17040396e-01 4.04788584e-01 -1.41048044e-01 -9.34428930e-01
2.41585672e-01 4.46569979e-01 1.10142338e+00 -7.04762161e-01
1.06051147e+00 3.12553644e-01 -7.52629191e-02 -5.34522712e-01
-9.41140473e-01 -1.13903439e+00 -6.47365689e-01 3.10145527e-01
6.19987428e-01 -1.13749695e+00 -1.28427505e+00 4.97766018e-01
-1.03202462e+00 -1.96975321e-01 -3.91655385e-01 2.61532545e-01
-3.65770459e-01 6.63504064e-01 -2.88815379e-01 -1.00008225e+00
-6.51685596e-01 -1.07859182e+00 1.40108573e+00 4.29472148e-01
-6.03018180e-02 -6.84403956e-01 2.01071665e-01 7.07986116e-01
5.34619903e-03 -3.55913132e-01 5.82022429e-01 -1.09162128e+00
-5.06280303e-01 -4.00595844e-01 -1.30721971e-01 3.87671255e-02
-4.23765361e-01 -7.38663435e-01 -1.04467630e+00 -6.88643575e-01
-3.40768278e-01 -4.11656439e-01 1.48171031e+00 1.27695248e-01
1.11478519e+00 -3.41839492e-01 -7.30622888e-01 6.38978422e-01
1.25101209e+00 -2.21175253e-01 2.76273996e-01 3.69840354e-01
7.43108988e-01 5.11223435e-01 5.13487220e-01 4.49606091e-01
6.85823083e-01 1.31711113e+00 -1.17644621e-02 -1.87939227e-01
-3.67068946e-01 -4.48728293e-01 2.52682775e-01 -2.61479318e-02
-8.98785442e-02 -1.11125186e-01 -8.41595531e-01 7.73834527e-01
-2.24617982e+00 -1.04585063e+00 2.10135192e-01 2.55728889e+00
1.15560567e+00 -9.36127827e-02 8.13688040e-01 1.01444624e-01
4.87763226e-01 -3.44411373e-01 -5.45320213e-01 1.10696428e-01
2.49366984e-01 4.22519416e-01 6.91633463e-01 6.32934034e-01
-1.37798655e+00 1.13155591e+00 5.55823326e+00 1.22662961e+00
-4.82678294e-01 3.33358496e-01 2.56893247e-01 -3.94104332e-01
5.75370669e-01 -1.42519951e-01 -1.81602490e+00 3.47250074e-01
9.03270185e-01 -5.39121917e-03 5.10654449e-01 9.64380980e-01
-1.50283575e-01 -3.35551023e-01 -1.62612641e+00 1.43854141e+00
2.63365299e-01 -9.24533367e-01 -1.39013439e-01 8.33554268e-02
5.65026104e-01 -2.04925135e-01 1.81501001e-01 5.52341580e-01
2.92471766e-01 -8.25691223e-01 6.64299846e-01 4.16898936e-01
5.77432334e-01 -8.88092935e-01 5.61491191e-01 7.72413015e-01
-1.25211954e+00 -3.99776965e-01 -2.40715891e-01 3.07949722e-01
1.42288595e-01 2.54794896e-01 -9.55089092e-01 3.55045348e-01
8.44881952e-01 -3.35895717e-02 -9.44098294e-01 1.26311326e+00
-2.13527754e-01 4.50316727e-01 -7.35731363e-01 2.19322294e-01
3.94039005e-02 2.47739956e-01 6.82450354e-01 1.52526379e+00
-9.90005136e-02 -5.75005338e-02 6.57577574e-01 6.31172478e-01
-1.23699769e-01 -1.97051808e-01 3.06591894e-02 3.03343534e-01
5.76703191e-01 9.77475107e-01 -3.36150259e-01 -7.65030980e-01
-1.78998172e-01 1.17014360e+00 6.13151312e-01 1.00641012e-01
-8.16482484e-01 -1.52656943e-01 4.84246850e-01 2.42379561e-01
7.50867069e-01 1.65269464e-01 7.26610199e-02 -9.54140961e-01
1.60694674e-01 -1.15851009e+00 9.06932771e-01 -1.60436821e-03
-1.05802143e+00 6.41888157e-02 4.83770072e-01 -4.87821490e-01
-5.50889790e-01 -3.64247531e-01 -3.18739533e-01 9.41242814e-01
-1.55581748e+00 -1.18966603e+00 -2.23990858e-01 7.58398294e-01
5.59240162e-01 -3.38200837e-01 7.98967659e-01 5.26367307e-01
-6.97797537e-01 1.53272545e+00 -2.15769410e-01 2.39006475e-01
8.45740795e-01 -1.51472139e+00 4.62350875e-01 9.94451880e-01
3.65623504e-01 7.37516463e-01 6.86434627e-01 -6.29544199e-01
-1.31643128e+00 -6.79060876e-01 1.06558120e+00 -8.18223536e-01
2.43890613e-01 -4.06650633e-01 -6.48309648e-01 4.48657393e-01
-7.86992982e-02 -1.96482867e-01 5.52327216e-01 4.13623214e-01
-2.80893594e-01 -1.10093713e-01 -1.21186769e+00 3.74987304e-01
1.04814613e+00 -5.01789570e-01 -5.87016761e-01 5.81575096e-01
5.09887815e-01 -6.56518698e-01 -5.46045005e-01 1.30014196e-01
5.55844367e-01 -7.56887436e-01 1.51256287e+00 -6.06847584e-01
-1.74228519e-01 -2.20396668e-01 3.66318882e-01 -9.07278240e-01
-4.63710576e-01 -8.80640209e-01 -7.15001047e-01 1.16182721e+00
3.13902467e-01 -3.98407906e-01 1.17108941e+00 7.70573258e-01
3.59517008e-01 -8.76075625e-01 -1.02998281e+00 -1.04227138e+00
-2.64313400e-01 -1.15515620e-01 4.19964135e-01 3.49088937e-01
-2.95062542e-01 3.88876796e-01 -5.23768604e-01 4.94533509e-01
1.13900936e+00 -1.80480689e-01 1.19572234e+00 -1.29639041e+00
-1.16112971e+00 -4.08120751e-01 -2.84725308e-01 -1.37923765e+00
5.29574864e-02 -8.75932336e-01 -8.59475359e-02 -1.21159542e+00
7.80729592e-01 -2.18602508e-01 -2.82595783e-01 7.36960948e-01
-6.85808837e-01 -8.16443190e-03 4.34865087e-01 1.23246923e-01
-8.68856728e-01 1.73999399e-01 7.11592913e-01 -6.78787380e-02
-3.54825556e-01 2.83466786e-01 -6.41802788e-01 5.63557148e-01
3.24165165e-01 -6.92035258e-01 -2.22162917e-01 -3.67550373e-01
-1.66934371e-01 -2.06871808e-01 6.86401606e-01 -9.84950483e-01
9.53556597e-01 4.61495548e-01 3.97053570e-01 -6.55827761e-01
5.11265099e-01 -7.11030960e-01 -2.74823517e-01 7.56419599e-01
-2.13156655e-01 -6.68855309e-02 2.28836328e-01 7.38243878e-01
1.34924978e-01 -4.75348473e-01 9.24586594e-01 -1.71722934e-01
-5.85031867e-01 4.14377779e-01 3.28009486e-01 4.49762158e-02
8.86293828e-01 -4.12382394e-01 -7.10688308e-02 -2.44710654e-01
-7.33435988e-01 6.17201865e-01 2.83993594e-02 2.35280856e-01
2.46544272e-01 -8.65294337e-01 -8.89568567e-01 2.24516630e-01
-5.00785261e-02 1.00691102e-01 1.54411271e-01 9.99682903e-01
-3.11542153e-02 6.32180452e-01 4.46525216e-01 -6.74461961e-01
-2.06417775e+00 7.05763459e-01 3.32202047e-01 -9.57085788e-01
-3.15110326e-01 1.49500179e+00 8.38054270e-02 -3.78107250e-01
9.38580394e-01 1.16876081e-01 -6.52221292e-02 9.23178792e-02
7.36611187e-01 7.46831477e-01 -1.10316472e-02 -4.67921674e-01
-4.39920992e-01 7.26192951e-01 -5.86283863e-01 -1.47195533e-01
9.83992279e-01 4.52159941e-02 2.87288457e-01 -3.07494193e-01
8.70151043e-01 4.30069156e-02 -1.03174698e+00 -5.52576482e-01
-2.10580863e-02 -5.22841573e-01 1.22579895e-01 -1.12108278e+00
-7.53731847e-01 4.60707158e-01 9.21229482e-01 -1.49508491e-01
1.02966261e+00 2.35921025e-01 8.43976259e-01 6.74627125e-01
3.16218138e-01 -1.33020997e+00 4.48722169e-02 3.86925817e-01
4.36435401e-01 -1.27785742e+00 4.18927819e-01 -4.79028821e-01
-3.88196111e-01 6.84040487e-01 6.88495696e-01 1.37154430e-01
3.40876073e-01 1.18339583e-02 -4.51930910e-01 -2.30141684e-01
-6.23252928e-01 -6.76029265e-01 6.28556073e-01 4.32662487e-01
-1.24927863e-01 -3.51825394e-02 -1.25349537e-01 6.34373724e-01
-2.39599854e-01 -6.62124082e-02 -3.42625260e-01 7.02361822e-01
-4.34596956e-01 -1.37862229e+00 -5.85387588e-01 2.64413625e-01
-7.00159431e-01 -2.22253829e-01 -4.37532246e-01 7.45703459e-01
5.59997499e-01 9.72160935e-01 -1.71855375e-01 -1.32342294e-01
5.61117172e-01 2.38293201e-01 7.48833835e-01 -7.03664780e-01
-9.96530771e-01 1.21149402e-02 1.29571542e-01 -5.32906592e-01
-9.45672467e-02 -7.76023865e-01 -1.01773870e+00 -1.37020230e-01
-6.37701571e-01 2.72128403e-01 4.12902057e-01 9.63896930e-01
4.56670791e-01 2.39695162e-01 1.81006983e-01 -6.56457841e-01
-1.31398654e+00 -8.67650688e-01 -2.93971002e-01 2.81238317e-01
3.60296726e-01 -7.98980832e-01 -1.75277218e-01 4.06248420e-02] | [14.831236839294434, 0.8226412534713745] |
3cf81688-0415-4d4b-9b82-9b8b7bed54ea | a-new-split-for-evaluating-true-zero-shot | 2107.13029 | null | https://arxiv.org/abs/2107.13029v2 | https://arxiv.org/pdf/2107.13029v2.pdf | A New Split for Evaluating True Zero-Shot Action Recognition | Zero-shot action recognition is the task of classifying action categories that are not available in the training set. In this setting, the standard evaluation protocol is to use existing action recognition datasets(e.g. UCF101) and randomly split the classes into seen and unseen. However, most recent work builds on representations pre-trained on the Kinetics dataset, where classes largely overlap with classes in the zero-shot evaluation datasets. As a result, classes which are supposed to be unseen, are present during supervised pre-training, invalidating the condition of the zero-shot setting. A similar concern was previously noted several years ago for image based zero-shot recognition but has not been considered by the zero-shot action recognition community. In this paper, we propose a new split for true zero-shot action recognition with no overlap between unseen test classes and training or pre-training classes. We benchmark several recent approaches on the proposed True Zero-Shot(TruZe) Split for UCF101 and HMDB51, with zero-shot and generalized zero-shot evaluation. In our extensive analysis, we find that our TruZesplits are significantly harder than comparable random splits as nothing is leaking from pre-training, i.e. unseen performance is consistently lower,up to 8.9% for zero-shot action recognition. In an additional evaluation we also find that similar issues exist in the splits used in few-shot action recognition, here we see differences of up to 17.1%. We publish oursplits1and hope that our benchmark analysis will change how the field is evaluating zero- and few-shot action recognition moving forward. | ['Marcus Rohrbach', 'Frank Keller', 'Kiyoon Kim', 'Laura Sevilla-Lara', 'Shreyank N Gowda'] | 2021-07-27 | null | null | null | null | ['zero-shot-action-recognition', 'few-shot-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 7.12120950e-01 2.01822415e-01 -3.48384589e-01 -2.49706745e-01
-9.45001721e-01 -2.97818989e-01 7.75494516e-01 -2.42687508e-01
-4.85675871e-01 8.33278775e-01 4.58052814e-01 8.99538621e-02
-7.84217268e-02 -7.10807681e-01 -5.92481494e-01 -8.90288651e-01
1.71411008e-01 4.50800031e-01 6.87986553e-01 -1.90770403e-01
2.63909996e-01 8.39794576e-02 -2.20371485e+00 8.13365817e-01
1.26590475e-01 8.53946030e-01 -3.91137600e-01 8.97603691e-01
1.24078050e-01 1.15094995e+00 -9.26969349e-01 -2.24228874e-01
4.02554750e-01 -8.57554138e-01 -1.26713789e+00 3.67949724e-01
6.21764958e-01 -5.54244578e-01 -3.00117999e-01 6.80199385e-01
5.22614837e-01 4.36835766e-01 7.00389862e-01 -1.53002644e+00
-4.66510326e-01 3.61782163e-01 -2.16524541e-01 3.31281960e-01
5.48781216e-01 4.21857357e-01 9.24432814e-01 -6.90874279e-01
8.99906039e-01 1.13384807e+00 5.27436912e-01 9.29863930e-01
-1.04492080e+00 -4.20594156e-01 -9.06797349e-02 5.79782546e-01
-1.15137804e+00 -7.61867464e-01 2.89145261e-01 -4.73395348e-01
1.25909495e+00 4.13651586e-01 2.43452907e-01 1.69965196e+00
1.79121140e-02 1.04298282e+00 9.68973994e-01 -5.90216100e-01
5.03412306e-01 -9.46012735e-02 5.79469800e-01 1.14075273e-01
8.30750763e-02 2.37980708e-01 -4.42572773e-01 -4.10138629e-02
4.41076756e-01 2.90794879e-01 -2.84358382e-01 -5.56543887e-01
-1.19651604e+00 6.61910951e-01 2.56890245e-02 6.93855107e-01
-2.25321814e-01 3.13018113e-02 7.74141312e-01 4.49044198e-01
2.41596729e-01 2.41471872e-01 -2.80297786e-01 -6.22952163e-01
-8.73514652e-01 3.17356884e-02 7.05465138e-01 6.87914193e-01
6.71986341e-01 4.33425829e-02 -6.99711978e-01 8.46004009e-01
-3.06154042e-01 9.92000476e-03 1.11367059e+00 -1.03521538e+00
2.22858310e-01 4.89177704e-01 1.68201461e-01 -4.26049709e-01
-6.80576563e-02 1.60360083e-01 -4.90092129e-01 4.69021142e-01
5.05617321e-01 1.19512983e-01 -1.47409010e+00 1.51264989e+00
5.12758382e-02 4.67172235e-01 4.38246816e-01 8.25423360e-01
7.74135172e-01 1.83007225e-01 1.35475978e-01 1.36916395e-02
1.19363534e+00 -9.53294337e-01 -5.74620008e-01 -5.24206087e-02
1.10114408e+00 -4.11974162e-01 1.02447462e+00 4.28847313e-01
-5.51684201e-01 -6.20711803e-01 -1.18958473e+00 2.75590956e-01
-7.76464045e-01 -8.63239393e-02 5.75596571e-01 8.16769421e-01
-8.10022056e-01 9.95071471e-01 -8.23533833e-01 -8.92245412e-01
6.36564910e-01 1.69098347e-01 -6.66009307e-01 -2.92533845e-01
-1.13914895e+00 9.40772116e-01 5.08472979e-01 -5.63394189e-01
-1.29049277e+00 -3.88582855e-01 -8.23554635e-01 -1.53690293e-01
7.55008578e-01 -6.75365403e-02 1.35818100e+00 -1.23674357e+00
-1.28796077e+00 1.08312714e+00 1.26837462e-01 -6.67721033e-01
5.10724068e-01 -1.01642003e-02 -6.21146858e-01 1.37337998e-01
1.94051057e-01 6.04528666e-01 6.93233967e-01 -9.08026695e-01
-6.23366952e-01 -3.91994864e-01 4.54474062e-01 -8.29134881e-02
2.05822084e-02 2.60865930e-02 -1.51877329e-02 -5.68343580e-01
-1.67402849e-01 -8.56134117e-01 9.72633734e-02 -2.33262300e-01
-2.30564311e-01 -2.10853890e-01 9.80391860e-01 -5.98748475e-02
1.00041842e+00 -2.29487157e+00 -2.53133684e-01 -4.19670790e-01
-1.02810055e-01 5.29742897e-01 -3.81854177e-01 4.77494240e-01
-7.02151775e-01 1.06677592e-01 -2.62505263e-01 -1.18519843e-01
1.07590191e-01 6.82011187e-01 -2.61856616e-01 4.66526359e-01
2.53096342e-01 8.43907535e-01 -9.59546924e-01 -2.67569125e-01
4.91499186e-01 1.77905783e-01 -3.53394151e-01 -3.95516586e-03
2.06621200e-01 1.33484811e-01 -6.08934946e-02 9.78301108e-01
2.00942338e-01 -6.57875910e-02 -8.84554908e-02 1.09813452e-01
9.00631025e-02 2.20767148e-02 -1.17907894e+00 1.55277467e+00
1.68402363e-02 5.94835758e-01 -5.90008080e-01 -1.26243591e+00
4.79634285e-01 6.57343030e-01 6.75125778e-01 -6.53384089e-01
2.42821127e-01 -1.16166603e-02 1.35837063e-01 -3.69839400e-01
3.65717173e-01 -5.37457824e-01 -1.12165622e-01 5.38472354e-01
6.05991662e-01 3.46768975e-01 5.10960340e-01 5.55735119e-02
1.65532994e+00 3.58671844e-01 5.52864015e-01 5.60400188e-02
3.70110691e-01 1.58632189e-01 5.04088163e-01 9.17277813e-01
-6.90951884e-01 1.10108542e+00 5.31297386e-01 -3.24625701e-01
-7.59536445e-01 -9.18024361e-01 -2.95641094e-01 1.23192799e+00
-9.11739394e-02 -5.19954085e-01 -8.32324922e-01 -1.01979792e+00
-2.12792620e-01 9.19924736e-01 -1.02492881e+00 -5.21238267e-01
-1.83530524e-01 -6.94682062e-01 6.48027837e-01 7.49856234e-01
2.76378900e-01 -1.32661736e+00 -9.54229116e-01 6.37346357e-02
9.15719122e-02 -9.16345119e-01 4.13350984e-02 4.37097162e-01
-6.45229161e-01 -1.48652506e+00 -8.56363416e-01 -2.25755811e-01
3.39261234e-01 3.58545005e-01 1.06306016e+00 -4.32663038e-03
-5.88349283e-01 7.99204409e-01 -1.01439440e+00 -3.87976170e-01
-4.23237354e-01 -3.55378151e-01 3.32509279e-02 1.91859081e-01
9.66943085e-01 -4.20634270e-01 -3.07034075e-01 7.24830866e-01
-1.03942621e+00 -3.78731698e-01 5.51590502e-01 8.61581028e-01
3.69883418e-01 -1.00716569e-01 3.83019656e-01 -8.46748114e-01
1.18177824e-01 -5.47280014e-01 1.15472391e-01 3.88422370e-01
-3.43024164e-01 -9.38110724e-02 3.11334610e-01 -6.47922099e-01
-7.28614271e-01 -2.09800608e-04 6.70316219e-02 -8.91402066e-01
-7.12692857e-01 -8.93369466e-02 -1.70206353e-01 4.28390317e-02
8.97925019e-01 1.30386382e-01 -1.29488826e-01 -5.08759916e-01
1.22113332e-01 8.35328937e-01 3.32211733e-01 -2.15388924e-01
3.53070557e-01 5.35093248e-01 -3.59702200e-01 -8.74631643e-01
-8.60201061e-01 -8.10422122e-01 -8.64058673e-01 -1.43507406e-01
1.10108554e+00 -7.18901813e-01 -2.26375833e-01 5.46153307e-01
-7.25956738e-01 -3.86043936e-01 -1.07400131e+00 4.98062909e-01
-9.57097173e-01 5.06058931e-01 -2.78193116e-01 -8.98371577e-01
1.35750994e-01 -1.00419283e+00 1.28689325e+00 -1.31017387e-01
-5.42221308e-01 -6.25055730e-01 2.44790256e-01 4.60688442e-01
1.55462593e-01 4.08071607e-01 4.32784051e-01 -1.07680261e+00
-1.83568910e-01 -4.66373861e-01 1.28866211e-01 6.40018702e-01
3.05692762e-01 -1.48539230e-01 -1.27355421e+00 -2.04844117e-01
-7.20615759e-02 -9.01702046e-01 1.16601527e+00 9.01735276e-02
9.12474573e-01 1.21024892e-01 -1.63075045e-01 7.66763911e-02
1.28526068e+00 3.82541984e-01 1.18226004e+00 2.16294825e-01
2.69270360e-01 4.75620180e-01 9.64389563e-01 5.45109153e-01
-2.85144836e-01 7.23534524e-01 3.04959327e-01 6.32331893e-02
-3.73904288e-01 -5.45598939e-02 6.90499067e-01 -9.90905566e-04
-3.83128375e-01 -3.61781031e-01 -8.14435542e-01 5.92716694e-01
-1.93733656e+00 -1.56419492e+00 3.29912826e-02 2.54480815e+00
4.57375765e-01 1.36714041e-01 2.66015738e-01 6.42666876e-01
7.14176595e-01 3.15341651e-01 -4.35509592e-01 -2.61003613e-01
-4.25552167e-02 5.00878572e-01 4.03630614e-01 2.34052271e-01
-1.38296747e+00 7.95512795e-01 6.50640965e+00 8.40651512e-01
-9.10774410e-01 4.00055557e-01 4.49127883e-01 -3.71553004e-01
4.52075034e-01 7.31769651e-02 -7.90582597e-01 3.33559632e-01
1.19053888e+00 -7.24550858e-02 3.53191048e-02 9.40463006e-01
-1.88857764e-01 -3.20351601e-01 -1.31700826e+00 1.06795657e+00
5.33881426e-01 -8.58767509e-01 -5.34103205e-03 5.00816293e-02
5.05131543e-01 1.45912349e-01 -4.43452269e-01 1.04670823e+00
3.46450806e-01 -9.15528238e-01 4.76203173e-01 4.39702123e-01
8.56702030e-01 -3.62587243e-01 7.95169950e-01 3.00497919e-01
-8.23138297e-01 -2.96907544e-01 -5.56804597e-01 -2.53663361e-01
-3.08136977e-02 2.23732665e-02 -4.93742108e-01 5.43091297e-01
5.51554620e-01 9.11938488e-01 -6.60871089e-01 9.54382777e-01
3.86351254e-03 4.89953786e-01 3.25619243e-02 3.25275779e-01
3.87899041e-01 2.03953385e-01 4.45721835e-01 9.51101303e-01
2.51322329e-01 1.79989681e-01 2.49974206e-01 2.31270671e-01
2.14671761e-01 -9.06043723e-02 -7.29340971e-01 -2.96519458e-01
-3.89322698e-01 8.55995357e-01 -7.42203653e-01 -8.01253140e-01
-5.69166243e-01 1.25761819e+00 -1.03858836e-01 2.86067247e-01
-9.74608362e-01 -3.95507127e-01 7.16445088e-01 2.26708293e-01
4.70609874e-01 3.79991472e-01 3.42712015e-01 -1.50418031e+00
-1.00357845e-01 -8.68666053e-01 7.55018294e-01 -7.45202243e-01
-1.26486385e+00 3.95698756e-01 2.44398117e-01 -1.69778562e+00
-5.18085003e-01 -8.60731721e-01 -6.44157588e-01 3.10216457e-01
-1.01630795e+00 -7.89882183e-01 -2.63564825e-01 5.91166377e-01
9.04244661e-01 -1.00232318e-01 1.15022027e+00 3.11949253e-01
-6.75867319e-01 4.07544881e-01 -7.73871914e-02 3.14801186e-01
8.29160988e-01 -1.04412651e+00 3.88010964e-02 7.01272666e-01
4.15664613e-01 2.49042258e-01 6.26810610e-01 -4.54529941e-01
-9.99209166e-01 -1.00872922e+00 6.77093387e-01 -8.55232358e-01
6.89038634e-01 -2.26744235e-01 -9.42301989e-01 7.78850615e-01
5.64528108e-02 3.08425367e-01 9.06409860e-01 6.02643099e-03
-3.52526546e-01 2.42782056e-01 -1.10275197e+00 2.23655969e-01
1.30659783e+00 -4.50435996e-01 -1.06641328e+00 4.35608089e-01
3.05089384e-01 6.95564747e-02 -6.61912858e-01 5.60754538e-01
6.55576110e-01 -1.31589353e+00 8.35377455e-01 -1.17543054e+00
4.53607023e-01 -7.30446428e-02 -5.43620050e-01 -1.01498032e+00
-1.95541695e-01 -5.34099936e-02 -8.31221342e-02 1.03987288e+00
1.68822661e-01 -5.12221336e-01 9.57550049e-01 6.26058400e-01
-9.04330462e-02 -5.22484064e-01 -1.09154689e+00 -1.29269254e+00
-1.40503213e-01 -4.59466726e-01 2.21515715e-01 1.07389760e+00
3.98039930e-02 2.71498621e-01 -4.77507770e-01 -5.07688582e-01
3.20990503e-01 -1.36744112e-01 9.82320309e-01 -1.11181712e+00
-6.14422679e-01 -2.98275918e-01 -1.20410323e+00 -5.07405043e-01
6.15436807e-02 -6.93463802e-01 5.86268678e-02 -1.40113211e+00
4.23330337e-01 2.89642185e-01 -8.15730572e-01 1.07450235e+00
1.21288277e-01 5.20710170e-01 2.00066760e-01 1.86739057e-01
-9.46561456e-01 5.33053458e-01 9.14233029e-01 -2.52423465e-01
7.76430592e-02 -5.90830594e-02 -3.28494310e-01 6.05827391e-01
6.43819511e-01 -5.88050246e-01 -6.01233184e-01 1.80702731e-01
-6.51754797e-01 -1.83275566e-01 6.41102910e-01 -1.56959307e+00
-1.93010986e-01 -1.82766274e-01 2.87151992e-01 -3.31537247e-01
5.45668602e-01 -7.33773112e-01 5.99713624e-02 4.52984482e-01
-2.49619156e-01 -6.24506950e-01 -3.57808806e-02 6.43569291e-01
-3.55803370e-01 -5.67718029e-01 8.30572426e-01 -4.57397997e-01
-1.22408271e+00 1.11425661e-01 -4.22436029e-01 1.32512867e-01
1.42900240e+00 -8.31921935e-01 -4.30239707e-01 -2.56949514e-01
-1.21205556e+00 -2.35881165e-01 6.81356668e-01 6.13463700e-01
3.88039589e-01 -1.26351392e+00 -5.19834518e-01 1.31120905e-01
6.16773665e-01 -7.28988945e-01 4.40640152e-01 1.13850188e+00
-1.35568995e-02 3.72819483e-01 -5.23415625e-01 -4.91164565e-01
-1.51785791e+00 6.95673227e-01 2.75153100e-01 -1.24823749e-01
-6.86057150e-01 5.27872562e-01 1.56655014e-01 -1.59879208e-01
2.75897533e-01 -1.39756441e-01 -1.53682053e-01 3.69344175e-01
8.90556693e-01 4.76505369e-01 1.47564054e-01 -8.18893194e-01
-4.30214405e-01 2.71322191e-01 2.12496482e-02 2.47834381e-02
1.33061838e+00 4.65395361e-01 4.12084758e-01 9.66872394e-01
1.29715204e+00 -6.10659540e-01 -1.03527617e+00 1.12104543e-01
1.10594645e-01 -7.20130503e-01 -2.02106014e-01 -6.85923457e-01
-7.87133873e-01 9.04683888e-01 9.10207629e-01 1.83991477e-01
1.00560665e+00 1.34431079e-01 4.77313548e-01 5.32945812e-01
6.19091272e-01 -1.28000796e+00 1.82099521e-01 4.49455202e-01
7.92831659e-01 -1.44305634e+00 -7.46781901e-02 -2.05632553e-01
-7.70177245e-01 8.92774761e-01 6.62731767e-01 -1.66809142e-01
2.65352994e-01 2.35929657e-02 -7.01767802e-02 -2.56156594e-01
-7.85841763e-01 -7.16903329e-01 -1.38258291e-02 6.82924986e-01
3.41384500e-01 -1.82698723e-02 -3.24507684e-01 5.68480551e-01
2.84738332e-01 3.66321176e-01 6.73607230e-01 1.48474777e+00
-5.42865574e-01 -1.20860720e+00 -3.00592989e-01 5.80219269e-01
-3.43253404e-01 1.60396278e-01 -7.26377487e-01 9.52598393e-01
3.39231610e-01 9.75245118e-01 1.00869015e-01 -5.93207121e-01
4.86986816e-01 7.06566751e-01 5.36514759e-01 -9.55576062e-01
-4.13033277e-01 -3.82544637e-01 3.07869017e-01 -9.96289790e-01
-6.71457589e-01 -6.65848613e-01 -1.09850132e+00 1.55799210e-01
-2.32058346e-01 -4.99702990e-02 1.46483660e-01 9.36789572e-01
2.68833399e-01 4.82211500e-01 2.14528754e-01 -8.79584312e-01
-6.90940201e-01 -1.11113894e+00 -7.19228983e-01 8.73494387e-01
5.06719016e-02 -1.15697896e+00 -7.47503877e-01 6.85433671e-02] | [8.517664909362793, 0.9436069130897522] |
7db06643-a352-4bec-92ea-4fc8029da6d9 | from-noisy-prediction-to-true-label-noisy | 2205.0069 | null | https://arxiv.org/abs/2205.00690v3 | https://arxiv.org/pdf/2205.00690v3.pdf | From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model | Noisy labels are inevitable yet problematic in machine learning society. It ruins the generalization of a classifier by making the classifier over-fitted to noisy labels. Existing methods on noisy label have focused on modifying the classifier during the training procedure. It has two potential problems. First, these methods are not applicable to a pre-trained classifier without further access to training. Second, it is not easy to train a classifier and regularize all negative effects from noisy labels, simultaneously. We suggest a new branch of method, Noisy Prediction Calibration (NPC) in learning with noisy labels. Through the introduction and estimation of a new type of transition matrix via generative model, NPC corrects the noisy prediction from the pre-trained classifier to the true label as a post-processing scheme. We prove that NPC theoretically aligns with the transition matrix based methods. Yet, NPC empirically provides more accurate pathway to estimate true label, even without involvement in classifier learning. Also, NPC is applicable to any classifier trained with noisy label methods, if training instances and its predictions are available. Our method, NPC, boosts the classification performances of all baseline models on both synthetic and real-world datasets. The implemented code is available at https://github.com/BaeHeeSun/NPC. | ['Il-Chul Moon', 'Kyungwoo Song', 'Byeonghu Na', 'JoonHo Jang', 'Seungjae Shin', 'HeeSun Bae'] | 2022-05-02 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 5.98868787e-01 1.86941847e-01 -1.38819769e-01 -5.72753072e-01
-9.96123433e-01 -7.02608466e-01 4.82088089e-01 1.89667046e-02
-2.85676479e-01 8.58172953e-01 -2.36212283e-01 -2.57199645e-01
-2.42499672e-02 -7.28339672e-01 -8.46410990e-01 -9.15458441e-01
4.25100297e-01 3.46818417e-01 8.71635750e-02 1.48546547e-01
3.72955687e-02 7.66403824e-02 -1.59448075e+00 3.09032559e-01
6.95089400e-01 1.09103012e+00 1.18795253e-01 5.11328757e-01
-2.31635030e-02 7.55182862e-01 -5.72274387e-01 -5.89678049e-01
2.97020465e-01 -3.80282998e-01 -8.74870300e-01 1.98260739e-01
1.64424732e-01 1.89008638e-01 2.38398418e-01 1.28936911e+00
2.99084783e-01 6.19591698e-02 7.85664022e-01 -1.21674705e+00
-6.20172441e-01 8.75341415e-01 -3.25165033e-01 -2.14692488e-01
2.59177327e-01 -1.28307715e-01 7.94816554e-01 -8.72793376e-01
3.39794070e-01 1.02613926e+00 1.10873449e+00 7.01055110e-01
-1.33621669e+00 -7.56629109e-01 1.40230104e-01 -1.46922506e-02
-1.33427083e+00 -2.28607118e-01 7.19078183e-01 -4.72332537e-01
3.95994425e-01 4.90024626e-01 4.24723834e-01 1.41499627e+00
8.01418559e-04 4.18011367e-01 1.56309164e+00 -5.99540532e-01
3.55130553e-01 5.16702056e-01 4.97906148e-01 3.88710737e-01
2.34265506e-01 4.35012251e-01 -3.16222310e-01 -2.46692300e-01
1.25981465e-01 -1.52782887e-01 -2.80301690e-01 -8.09878409e-02
-9.45210874e-01 7.04543054e-01 2.15692446e-02 1.16138920e-01
-1.06239375e-02 1.45122677e-01 2.92118639e-01 3.14387560e-01
6.73028588e-01 4.38710660e-01 -7.62354612e-01 -2.19590545e-01
-7.61958063e-01 -8.14058334e-02 8.09854805e-01 9.83978689e-01
7.97087073e-01 -1.85880318e-01 -2.74210852e-02 9.19893026e-01
2.82341331e-01 5.04443169e-01 6.50157988e-01 -8.91475737e-01
5.63722141e-02 5.56271195e-01 -6.35260865e-02 -7.81139016e-01
-5.00143826e-01 -8.60721827e-01 -8.86024415e-01 -7.09444955e-02
5.46982825e-01 -1.64998963e-01 -7.73401260e-01 1.71590483e+00
5.18526733e-01 5.11057198e-01 1.20031163e-01 4.65888828e-01
6.43513739e-01 5.08241355e-01 8.13820213e-02 -4.20706034e-01
1.13349390e+00 -8.87138426e-01 -8.60090077e-01 -3.53792399e-01
1.01976562e+00 -9.19415236e-01 1.40865672e+00 7.63333082e-01
-4.76662368e-01 -6.71580255e-01 -8.72006357e-01 1.71252802e-01
-4.37931776e-01 3.08505446e-01 4.73676324e-01 1.17095554e+00
-7.55001664e-01 6.34836555e-01 -6.76638246e-01 -2.17940450e-01
3.28217566e-01 5.91701150e-01 -2.66588390e-01 -2.19839901e-01
-1.07734489e+00 9.18612838e-01 4.21474904e-01 2.06576273e-01
-5.98703265e-01 -3.50670427e-01 -6.75015390e-01 -2.17903540e-01
5.28506279e-01 -3.69454175e-01 1.68004704e+00 -1.29852951e+00
-1.63992727e+00 6.71333075e-01 -1.51322395e-01 -6.58586845e-02
6.04641914e-01 -1.33088231e-01 -2.90311843e-01 -5.75711250e-01
3.84434760e-02 4.62003618e-01 8.31055760e-01 -1.65818703e+00
-5.39110661e-01 -1.92892641e-01 -2.31905445e-01 -1.17901608e-01
-2.55590230e-01 -2.26479381e-01 -1.14872023e-01 -5.76561928e-01
3.84857953e-01 -1.14410186e+00 -1.91027343e-01 -4.48911905e-01
-5.59222102e-01 -3.49689126e-01 9.53121006e-01 -2.01041296e-01
1.13237190e+00 -2.09603119e+00 -4.38164562e-01 3.84989053e-01
8.23740214e-02 2.69725621e-01 2.92171035e-02 1.69621632e-01
-2.92781830e-01 3.87944728e-01 -4.19486910e-01 -5.01997471e-01
-5.54193817e-02 5.31355321e-01 -1.49837047e-01 5.16841590e-01
2.48100042e-01 6.56609595e-01 -1.01529968e+00 -4.15811092e-01
-4.67587896e-02 3.40804636e-01 -3.67247880e-01 1.30287051e-01
-2.14398205e-02 6.42907739e-01 -2.12925598e-01 5.62385201e-01
8.57962430e-01 -3.17450851e-01 3.56222302e-01 -8.84385183e-02
2.04153478e-01 3.03097755e-01 -1.51023853e+00 1.21560252e+00
-4.15690720e-01 2.10364014e-01 -3.92944723e-01 -1.22511053e+00
1.05588305e+00 3.52785349e-01 1.06367059e-01 -1.63452908e-01
3.65871489e-01 4.48098928e-01 -1.91054732e-01 -6.49011493e-01
2.19247162e-01 -3.12628001e-01 -1.43161505e-01 3.34216654e-01
1.95016935e-01 -1.61793008e-01 1.39335245e-01 -1.13991044e-01
9.73394036e-01 2.67215580e-01 3.84944975e-01 -7.02988356e-02
4.33605790e-01 -1.99968293e-01 7.88962781e-01 9.07939613e-01
-2.11370975e-01 5.89192569e-01 4.16590184e-01 -9.18781385e-02
-8.24824393e-01 -6.73767805e-01 -5.85449576e-01 1.06291568e+00
-5.65134571e-04 -3.64491135e-01 -8.14001560e-01 -1.00356293e+00
-2.80686349e-01 5.76734245e-01 -6.46366477e-01 -3.15238804e-01
-1.43984452e-01 -1.09463131e+00 5.52224219e-01 3.63696814e-01
4.13947612e-01 -6.49690151e-01 4.71319444e-02 1.85888007e-01
-2.81942666e-01 -1.13344705e+00 -3.11424524e-01 7.83151329e-01
-7.38706946e-01 -1.18676364e+00 -1.36939034e-01 -8.44170094e-01
8.19407701e-01 6.17106557e-02 9.48237777e-01 3.07867169e-01
4.63406712e-01 1.95529386e-01 -6.66742980e-01 -4.90531832e-01
-8.55403960e-01 2.81384408e-01 1.50527596e-01 8.23870227e-02
5.27148545e-01 -4.64395970e-01 -1.78401358e-02 5.60022235e-01
-7.74488091e-01 6.23981021e-02 4.46643710e-01 1.25440216e+00
7.26804256e-01 1.65318534e-01 8.08697164e-01 -1.48327160e+00
4.39507425e-01 -5.82848966e-01 -6.53906345e-01 2.99690843e-01
-1.03246403e+00 1.15846120e-01 6.09510481e-01 -7.12008715e-01
-9.49734211e-01 6.58304334e-01 -2.75106698e-01 3.25341187e-02
-3.69022936e-01 4.67103034e-01 -3.46317053e-01 7.85574038e-03
9.80849922e-01 -7.38918856e-02 -2.22831324e-01 -5.34839332e-01
1.59863338e-01 1.10539305e+00 3.42720002e-01 -6.67762935e-01
8.55212092e-01 1.35054499e-01 1.25204369e-01 -3.23061377e-01
-1.34174299e+00 -5.82224131e-01 -8.42128873e-01 -1.97312027e-01
3.47441733e-01 -7.24857450e-01 -5.86355269e-01 5.39896011e-01
-9.84460354e-01 -3.26867014e-01 -3.26205552e-01 5.16313493e-01
-3.20473403e-01 3.54515612e-01 -4.29360688e-01 -9.72530901e-01
-4.82769916e-03 -1.31579566e+00 9.18622732e-01 6.97689578e-02
-1.90551773e-01 -9.39314008e-01 -1.61809456e-02 4.92335826e-01
7.83400461e-02 9.28695649e-02 5.87389052e-01 -9.08748865e-01
-2.18475774e-01 -4.93540168e-01 -1.06231763e-03 9.38448131e-01
1.13817491e-01 1.00287177e-01 -1.49931800e+00 -1.33191511e-01
3.46439987e-01 -3.91914427e-01 5.97613394e-01 1.88290492e-01
1.18979561e+00 -2.90027887e-01 -2.82328457e-01 3.68582040e-01
1.29357564e+00 1.27930924e-01 3.41338426e-01 2.67373472e-01
6.92107081e-01 4.65366572e-01 6.01219535e-01 1.57240361e-01
2.51003176e-01 5.79647362e-01 2.70699173e-01 1.51870295e-01
-5.20665050e-02 -3.17360312e-01 3.92772108e-01 1.11722040e+00
1.81448624e-01 -2.72906840e-01 -1.12635136e+00 9.96826142e-02
-1.85227096e+00 -5.47621012e-01 -6.55126452e-01 2.18470144e+00
1.23523593e+00 1.41973823e-01 -3.43787581e-01 7.46612251e-01
8.01910698e-01 -6.07684255e-01 -3.70810121e-01 -1.94889188e-01
3.68912667e-02 1.97062567e-01 5.05790293e-01 5.85787296e-01
-1.32501459e+00 8.74446988e-01 6.20238829e+00 1.05048966e+00
-1.00934696e+00 5.06332755e-01 8.10980082e-01 2.42145970e-01
-7.20048025e-02 2.35176265e-01 -1.08477390e+00 5.66336334e-01
1.18239701e+00 2.60670185e-01 2.77574807e-01 9.61275339e-01
2.80610353e-01 -3.49806964e-01 -1.24710107e+00 8.56955886e-01
2.00636405e-02 -8.12761426e-01 -3.28062057e-01 -6.82418123e-02
9.82378542e-01 -2.31786907e-01 8.19456801e-02 5.41520178e-01
3.89032692e-01 -1.06733882e+00 7.20383942e-01 4.42048728e-01
5.93857586e-01 -7.07119286e-01 1.29173875e+00 8.06084871e-01
-6.82883322e-01 -2.14039296e-01 -3.07163686e-01 -2.31621057e-01
-3.21707338e-01 1.01413298e+00 -1.01615536e+00 3.25702429e-01
5.17939091e-01 5.43644726e-01 -9.46672738e-01 9.38231051e-01
-4.57435697e-01 1.14475060e+00 -3.68069619e-01 1.26297966e-01
-2.54257787e-02 -8.21539611e-02 9.16066468e-02 1.08945215e+00
3.72587025e-01 -4.64615226e-02 4.00992960e-01 5.45414329e-01
-1.01433299e-01 2.10755050e-01 -4.64560002e-01 2.06565261e-01
5.38990200e-01 1.19177294e+00 -8.00241888e-01 -2.96482563e-01
-2.94981241e-01 5.30545592e-01 2.48874918e-01 2.41292074e-01
-8.28582466e-01 5.23406528e-02 -1.20453060e-01 -6.01378875e-03
-2.25289941e-01 9.67101678e-02 -7.84367800e-01 -1.01256657e+00
4.31537889e-02 -1.13864660e+00 1.60514295e-01 -6.38645411e-01
-1.52618289e+00 5.60552239e-01 -1.39797136e-01 -1.32906330e+00
-3.82439122e-02 -5.40022373e-01 -1.63817331e-01 6.22712076e-01
-1.40698361e+00 -1.20105577e+00 -4.38591987e-01 3.26665431e-01
3.77247065e-01 1.97260037e-01 1.04138792e+00 4.46934491e-01
-7.81195879e-01 8.24459255e-01 3.55483353e-01 2.57145595e-02
9.58189964e-01 -1.26462483e+00 -8.75911042e-02 8.69560361e-01
1.18983403e-01 3.21263731e-01 7.79023290e-01 -5.91155529e-01
-7.82072842e-01 -1.28892088e+00 9.02384937e-01 -7.52836764e-01
5.66422760e-01 -3.92265558e-01 -1.02213168e+00 7.06934810e-01
-2.57425576e-01 2.49991864e-01 8.96208823e-01 3.50782543e-01
-3.34841192e-01 -1.24292403e-01 -1.13404202e+00 1.11589216e-01
8.10806811e-01 -4.03772950e-01 -3.10862869e-01 7.54207194e-01
5.81585228e-01 -5.70360839e-01 -7.94527411e-01 4.32785600e-01
3.20653021e-01 -7.28924334e-01 4.04415488e-01 -2.92849183e-01
1.98001176e-01 -2.96151966e-01 -3.17817569e-01 -1.37915719e+00
-6.73623011e-02 -4.21705365e-01 8.47452283e-02 1.45584226e+00
8.17587674e-01 -7.56403744e-01 7.53391087e-01 6.89494193e-01
-2.20397323e-01 -8.92766476e-01 -8.64103079e-01 -8.90644014e-01
4.85959053e-02 -9.83879685e-01 5.71332991e-01 1.14382041e+00
-7.90937096e-02 4.03340280e-01 -4.47799027e-01 3.25100601e-01
4.61137265e-01 -3.61717552e-01 7.28570819e-01 -1.40855014e+00
-4.30802286e-01 -3.84944603e-02 -2.76769191e-01 -8.52323234e-01
3.10124874e-01 -1.04651177e+00 3.36348057e-01 -1.02867877e+00
2.71250814e-01 -8.90135109e-01 -1.61991179e-01 8.41811478e-01
-3.21299076e-01 4.81498212e-01 -5.89242391e-02 2.87880927e-01
-4.24750209e-01 3.12894136e-01 1.31941378e+00 2.43610088e-02
-6.20361939e-02 5.58561683e-01 -7.02595055e-01 8.95102203e-01
1.16670358e+00 -1.24464357e+00 -4.74009097e-01 5.52309901e-02
5.08324981e-01 -4.69780058e-01 2.86705583e-01 -1.10383236e+00
3.58177423e-02 -8.33475664e-02 2.43690133e-01 -2.48937264e-01
1.66092843e-01 -9.69240189e-01 3.86509180e-01 2.64664501e-01
-4.92867291e-01 -1.92610443e-01 -1.43083572e-01 6.96112275e-01
-1.41024634e-01 -9.52390075e-01 9.29436088e-01 1.02352891e-02
-2.33770519e-01 -7.61575326e-02 -3.50423366e-01 -1.07949250e-03
9.66262519e-01 -1.76758587e-01 -3.33216697e-01 -1.80416957e-01
-8.90465438e-01 -2.92703439e-03 3.15298587e-01 1.11486845e-01
1.91931978e-01 -1.20426941e+00 -3.14485729e-01 1.98686451e-01
4.62953970e-02 1.94710642e-01 -1.04856625e-01 9.06613469e-01
-1.57753140e-01 6.45674765e-02 3.70792836e-01 -6.91677153e-01
-1.35318100e+00 5.30716002e-01 4.30553436e-01 -3.76088709e-01
-2.82395720e-01 8.73901486e-01 -1.87437043e-01 -8.63168955e-01
4.19883758e-01 -5.75953186e-01 -1.75632522e-01 2.85451091e-03
1.77796483e-01 1.86149716e-01 4.60284948e-01 -6.59872115e-01
-3.24366242e-03 5.15968978e-01 4.75516878e-02 1.39167264e-01
1.07983398e+00 -2.31245607e-01 -4.91195954e-02 9.25240755e-01
1.14008915e+00 -9.81095433e-02 -1.04600072e+00 -3.00635129e-01
3.02608907e-01 -2.95280367e-01 1.91272438e-01 -9.13224876e-01
-9.61077869e-01 6.98935568e-01 8.52874815e-01 3.39762628e-01
1.07769966e+00 -1.98010042e-01 4.50977266e-01 5.32008767e-01
2.81668425e-01 -1.19817626e+00 1.59086601e-03 6.10101640e-01
4.87330556e-01 -1.64437926e+00 -1.27160043e-01 -8.57089758e-01
-6.56401694e-01 9.55845416e-01 5.17101467e-01 1.55596212e-01
1.03578711e+00 4.32884276e-01 4.20628697e-01 3.43655944e-02
-7.50267148e-01 -1.01865023e-01 1.62068322e-01 6.77110612e-01
4.74725306e-01 1.76186353e-01 -4.59017456e-01 1.02578008e+00
-3.46775830e-01 9.34617594e-02 8.02477479e-01 8.59709918e-01
-3.44980925e-01 -1.53849113e+00 -6.19589806e-01 5.66902041e-01
-5.04445374e-01 -6.35383204e-02 -2.93773651e-01 5.14689267e-01
8.63609254e-01 1.29387677e+00 -3.38026494e-01 -6.72080457e-01
3.30571234e-01 5.80247581e-01 9.51737463e-02 -9.77604330e-01
-6.62142098e-01 1.16790704e-01 7.78406039e-02 -1.11161202e-01
-6.57973707e-01 -6.43338501e-01 -1.02444541e+00 -1.54060364e-01
-1.03813183e+00 3.64081711e-01 7.23559678e-01 1.06572342e+00
9.16251987e-02 4.12094653e-01 7.82444715e-01 -6.48621619e-01
-8.84840608e-01 -1.24400055e+00 -6.57793105e-01 3.11153531e-01
1.99794725e-01 -8.28753650e-01 -7.33071089e-01 3.29006732e-01] | [9.407529830932617, 4.004250526428223] |
7733ff97-6220-4fe4-b73a-16cc1c0bec8e | codec-complex-document-and-entity-collection | 2205.04546 | null | https://arxiv.org/abs/2205.04546v2 | https://arxiv.org/pdf/2205.04546v2.pdf | CODEC: Complex Document and Entity Collection | CODEC is a document and entity ranking benchmark that focuses on complex research topics. We target essay-style information needs of social science researchers, i.e. "How has the UK's Open Banking Regulation benefited Challenger Banks?". CODEC includes 42 topics developed by researchers and a new focused web corpus with semantic annotations including entity links. This resource includes expert judgments on 17,509 documents and entities (416.9 per topic) from diverse automatic and interactive manual runs. The manual runs include 387 query reformulations, providing data for query performance prediction and automatic rewriting evaluation. CODEC includes analysis of state-of-the-art systems, including dense retrieval and neural re-ranking. The results show the topics are challenging with headroom for document and entity ranking improvement. Query expansion with entity information shows significant gains in document ranking, demonstrating the resource's value for evaluating and improving entity-oriented search. We also show that the manual query reformulations significantly improve document ranking and entity ranking performance. Overall, CODEC provides challenging research topics to support the development and evaluation of entity-centric search methods. | ['Jeffrey Dalton', 'Sean MacAvaney', 'Sophie Fischer', 'Carlos Gemmell', 'Paul Owoicho', 'Iain Mackie'] | 2022-05-09 | null | null | null | null | ['document-ranking'] | ['natural-language-processing'] | [-5.05058408e-01 1.21391609e-01 -6.78324044e-01 -1.59563705e-01
-1.34749520e+00 -9.03706968e-01 8.90373051e-01 7.93333113e-01
-8.94031584e-01 6.78277671e-01 8.90888929e-01 -4.53913063e-01
-4.55551773e-01 -6.47816777e-01 -3.78218085e-01 2.07162142e-01
1.69562493e-02 1.10505509e+00 2.24316314e-01 -7.46580720e-01
9.88358974e-01 6.75280392e-02 -9.41375315e-01 5.70532441e-01
1.10100532e+00 4.61963415e-01 -1.33299172e-01 6.16601646e-01
-2.68056452e-01 6.28770530e-01 -8.65525067e-01 -9.79435802e-01
-1.17980495e-01 2.65352756e-01 -1.38980675e+00 -9.73251998e-01
6.12365961e-01 3.17912102e-02 -2.39816427e-01 8.26052189e-01
8.80698323e-01 1.39680117e-01 8.73523891e-01 -6.93065464e-01
-1.36846781e+00 1.06410670e+00 -2.24020392e-01 7.62783229e-01
6.38247132e-01 -4.34935600e-01 2.01475549e+00 -1.19012487e+00
1.26093471e+00 1.05278850e+00 2.82565951e-01 1.76273763e-01
-9.66646552e-01 -6.50515497e-01 -2.59621799e-01 4.30466771e-01
-1.24076402e+00 -4.79237169e-01 2.68575460e-01 -4.45256561e-01
1.81507599e+00 3.83721083e-01 3.19415778e-01 1.03077102e+00
-1.08162805e-01 5.57715237e-01 7.13658333e-01 -6.67331040e-01
-1.71921536e-01 3.52636307e-01 8.31272900e-01 4.59154546e-01
5.93597531e-01 -2.49560744e-01 -5.05349576e-01 -4.37837809e-01
5.29915802e-02 -2.26598427e-01 -3.16974103e-01 7.39924237e-02
-1.39931428e+00 7.64405370e-01 3.86749029e-01 4.32977647e-01
-3.64870697e-01 -4.86950912e-02 6.82318389e-01 4.66948420e-01
4.75107223e-01 1.73124897e+00 -5.98690927e-01 -2.06470191e-01
-7.76418507e-01 6.12231135e-01 1.38630235e+00 1.13168359e+00
5.17121792e-01 -6.56996369e-01 -7.06361055e-01 1.33268571e+00
1.79338306e-01 7.96881616e-01 7.34266222e-01 -1.32325172e+00
8.95023763e-01 9.13482785e-01 3.06695700e-01 -1.28950918e+00
-5.20637095e-01 -8.84358466e-01 -3.34619582e-01 -8.20097685e-01
9.33523625e-02 -5.25749773e-02 -4.02961075e-01 1.22713280e+00
-2.90979922e-01 -7.40426660e-01 2.57352322e-01 4.63609815e-01
1.34303081e+00 5.89541972e-01 2.07502186e-01 -1.76693350e-02
1.52747917e+00 -1.10193062e+00 -7.68769920e-01 1.61710251e-02
1.19294703e+00 -1.06842101e+00 9.77618933e-01 -1.46065637e-01
-1.17763734e+00 -1.59872994e-01 -6.75176561e-01 -4.87582535e-01
-1.07256114e+00 1.47141978e-01 4.99911726e-01 3.63327265e-01
-1.13198543e+00 3.17779213e-01 -1.40634805e-01 -5.26358604e-01
6.52242899e-02 2.28419006e-02 -1.12240434e-01 -9.86355022e-02
-1.85371244e+00 1.38196647e+00 2.25638583e-01 -4.18741435e-01
-3.09000224e-01 -1.09088266e+00 -5.90167761e-01 5.61393261e-01
2.64256924e-01 -6.49609983e-01 1.31786263e+00 3.08058292e-01
-7.77875662e-01 9.47711051e-01 6.95249822e-04 -5.35701394e-01
6.66402504e-02 -4.61409122e-01 -6.86401963e-01 2.49585062e-01
7.69616008e-01 6.11810148e-01 -1.44974515e-01 -6.91676617e-01
-5.92471898e-01 -9.09030735e-02 1.76567137e-01 5.52710533e-01
-6.92323625e-01 5.31989813e-01 -8.76649022e-01 -6.85775518e-01
-2.97934383e-01 -8.36449027e-01 1.32289350e-01 -8.82189810e-01
-3.37349921e-01 -9.46355939e-01 3.79922032e-01 -7.82176554e-01
1.81237924e+00 -1.43420732e+00 -7.25928172e-02 2.85141587e-01
3.58614385e-01 4.37942520e-02 -2.36968249e-01 8.60245287e-01
5.83865531e-02 8.33365262e-01 6.14607096e-01 -1.09818801e-01
4.24270600e-01 -3.76465321e-01 -6.02799475e-01 -1.82734832e-01
-2.48649210e-01 1.33675802e+00 -9.28644657e-01 -8.84844780e-01
-5.49024105e-01 6.71410281e-03 -6.08894169e-01 -2.63694823e-01
3.51566784e-02 -4.59502578e-01 -5.54207683e-01 1.04165506e+00
-9.02202278e-02 -7.56225467e-01 -8.21599290e-02 -2.13238820e-01
7.05852509e-02 7.38843441e-01 -4.89952207e-01 1.38588119e+00
-3.72087449e-01 9.00544822e-01 -4.80868608e-01 -3.76883119e-01
6.45740807e-01 2.69374490e-01 5.67176998e-01 -1.26656497e+00
-3.46671343e-01 5.29515445e-01 -3.64419103e-01 -5.23169219e-01
1.21367800e+00 6.75619125e-01 -4.27561641e-01 7.74503291e-01
-4.13056975e-03 -6.99450299e-02 7.26955652e-01 9.68163073e-01
1.47017348e+00 -4.25381511e-01 -8.84590447e-02 -4.94003266e-01
3.93082470e-01 3.79986376e-01 2.91765988e-01 1.06564772e+00
1.60329957e-02 1.87450841e-01 2.21631229e-01 -1.14186376e-01
-1.14877522e+00 -6.30975246e-01 -2.75519997e-01 1.78839803e+00
1.09700866e-01 -9.05298829e-01 -5.40659547e-01 -7.02342570e-01
3.57588291e-01 7.62265265e-01 -6.76196396e-01 -2.44097933e-01
-6.74238861e-01 -5.51578522e-01 6.94991469e-01 5.59610009e-01
4.31259692e-01 -1.14166510e+00 -2.05062345e-01 1.88063115e-01
-8.10933471e-01 -1.18150198e+00 -7.37035096e-01 -6.72598556e-02
-6.58240914e-01 -1.06276202e+00 -1.03048706e+00 -7.80333817e-01
1.27153277e-01 2.10490942e-01 2.00269747e+00 1.50470704e-01
-2.26358980e-01 8.39889407e-01 -6.49991155e-01 -3.22816521e-01
-1.23746775e-01 9.18514788e-01 -1.19666763e-01 -1.15111291e+00
7.26075470e-01 2.31588110e-02 -8.90558839e-01 5.99744201e-01
-4.87825841e-01 -3.98675680e-01 7.62261271e-01 9.63435173e-01
1.57079726e-01 -3.73765618e-01 1.13545787e+00 -1.18957198e+00
1.41507053e+00 -6.55239761e-01 -4.05183822e-01 1.14922750e+00
-1.66764116e+00 1.16260514e-01 6.42066300e-02 8.69091041e-03
-1.20693231e+00 -8.51004481e-01 3.03125352e-01 2.42555261e-01
4.79007602e-01 1.02866995e+00 5.76677382e-01 1.94510385e-01
1.31698763e+00 -1.80322468e-01 -6.12079799e-01 -5.95841408e-01
6.39684319e-01 7.33717144e-01 4.68424886e-01 -8.89752507e-01
5.03350019e-01 -9.69849825e-02 -4.90412563e-01 -3.97988170e-01
-9.87444282e-01 -1.16035450e+00 -1.57366648e-01 -9.98656377e-02
7.35135794e-01 -1.01717222e+00 -5.45668840e-01 -1.65755432e-02
-1.19380462e+00 1.50591265e-02 -2.47204989e-01 2.98833132e-01
2.29122154e-02 2.21973717e-01 -9.48284090e-01 -7.46029168e-02
-9.20765162e-01 -1.01583183e+00 1.19369519e+00 3.39018255e-01
-4.23065215e-01 -1.06774294e+00 5.46091676e-01 8.53457153e-01
5.79681635e-01 -4.34658945e-01 1.19180810e+00 -1.28502095e+00
-3.38824689e-01 -4.70319539e-01 -3.27568144e-01 -6.40030056e-02
-3.08313400e-01 9.70064942e-03 -4.63842809e-01 -2.42691800e-01
-9.59059238e-01 -6.09815717e-01 8.93843293e-01 -3.80428992e-02
8.76560390e-01 -3.36590767e-01 -5.33084095e-01 2.91541386e-02
1.09228027e+00 4.67971899e-02 4.51090664e-01 8.52720857e-01
3.60876471e-01 7.49304295e-01 5.69719553e-01 3.03701311e-01
7.01664090e-01 5.95761955e-01 -1.74917996e-01 2.72719741e-01
-9.41432938e-02 -1.56137675e-01 1.26839608e-01 1.00744951e+00
-1.04438730e-01 -3.57431203e-01 -1.34552383e+00 6.84556723e-01
-1.61942220e+00 -1.05154848e+00 -1.97701510e-02 1.75445163e+00
1.27809262e+00 8.27635229e-02 -2.51635730e-01 -5.60734093e-01
6.78345323e-01 6.67775720e-02 -4.72834498e-01 -2.47992650e-01
-4.76240009e-01 4.15928721e-01 5.04796386e-01 5.47725141e-01
-9.09488976e-01 1.13072062e+00 6.84509802e+00 9.68512416e-01
-3.01578045e-01 -2.00460598e-01 5.20931125e-01 7.42585734e-02
-7.17376590e-01 1.46821409e-01 -1.41757548e+00 2.93034077e-01
1.00539911e+00 -1.00219905e+00 1.06982365e-01 8.32963884e-01
-3.98050189e-01 8.06478038e-02 -1.07996035e+00 5.77201128e-01
2.51358896e-01 -1.77710319e+00 1.94866851e-01 1.51859343e-01
1.01903391e+00 2.96088070e-01 7.28046745e-02 1.28065121e+00
8.98335457e-01 -8.89488041e-01 2.30320841e-01 7.38644004e-01
6.29646838e-01 -2.56455153e-01 9.82741117e-01 3.05045769e-02
-7.97524214e-01 -1.76759630e-01 -4.59675670e-01 5.33897936e-01
-1.27101570e-01 4.39891994e-01 -1.02167988e+00 3.79741013e-01
1.03400171e+00 9.77142274e-01 -1.29469717e+00 9.19466615e-01
5.69157563e-02 4.93855745e-01 -1.09688967e-01 -5.97825646e-01
1.06348008e-01 -5.48685342e-03 5.89051247e-01 1.80526614e+00
3.31523381e-02 1.27596036e-01 -1.40990049e-01 6.79024577e-01
-8.25226784e-01 6.58538818e-01 -5.61867297e-01 -3.65085095e-01
1.10597062e+00 1.31255758e+00 -3.09553653e-01 -6.92007601e-01
-2.21635774e-01 5.17670453e-01 6.25704467e-01 5.72405100e-01
-5.02836525e-01 -9.87647772e-01 6.40865937e-02 -3.11753988e-01
-1.60339609e-01 1.32387996e-01 -2.46341437e-01 -1.30737388e+00
7.73820952e-02 -8.61911416e-01 9.55036700e-01 -8.11421871e-01
-1.47321963e+00 3.78775567e-01 7.96140060e-02 -8.31405818e-01
-5.05068898e-01 -3.88692677e-01 -3.53860617e-01 8.89785886e-01
-1.61297798e+00 -8.16013515e-01 -9.86152068e-02 2.08721563e-01
5.27170002e-01 -6.02728248e-01 8.75336289e-01 6.63421154e-01
-5.69645822e-01 8.00759375e-01 4.66946930e-01 3.46233636e-01
1.57888281e+00 -1.67368853e+00 2.60972321e-01 4.01698261e-01
1.87582880e-01 1.23047197e+00 3.22931409e-01 -8.36227953e-01
-8.95458102e-01 -7.45260775e-01 1.68115985e+00 -1.21979499e+00
9.04125988e-01 1.61312744e-01 -8.59098434e-01 4.92736101e-01
6.66805208e-01 -4.92040277e-01 9.83485758e-01 8.59044731e-01
-5.74867725e-01 -2.91090738e-02 -5.60156643e-01 4.55653846e-01
7.81586528e-01 -9.25979733e-01 -1.08431113e+00 8.94132257e-01
8.78349423e-01 -1.90926120e-01 -1.31794059e+00 4.41487342e-01
5.82527757e-01 -3.30745280e-01 1.43438375e+00 -1.24448264e+00
7.05963314e-01 1.94594756e-01 -1.32223710e-01 -1.20983016e+00
-6.52829230e-01 -2.42196709e-01 -4.99106765e-01 1.27696335e+00
1.20838094e+00 -5.37879169e-01 6.48803055e-01 9.86380219e-01
-1.54670984e-01 -8.64473939e-01 -3.19747448e-01 -7.33298123e-01
7.00202882e-01 2.08436385e-01 4.12420511e-01 1.25015688e+00
4.92798477e-01 7.84384310e-01 4.53331262e-01 -3.27305466e-01
2.60824412e-01 1.84871286e-01 4.23447579e-01 -1.28947937e+00
-6.08756691e-02 -1.06983435e+00 1.83752403e-01 -9.09294307e-01
3.71462911e-01 -1.23458755e+00 -4.97774869e-01 -1.68518484e+00
5.71864367e-01 -6.00217044e-01 -3.79707396e-01 3.86717498e-01
-5.04085422e-01 3.52308638e-02 -2.66012978e-02 8.06005776e-01
-1.27193034e+00 4.23461735e-01 9.66932058e-01 -4.24959064e-01
1.27510786e-01 -3.35912883e-01 -1.10073662e+00 1.68338642e-01
3.83895606e-01 -2.94584036e-01 -1.64002895e-01 -6.68968976e-01
1.06143606e+00 3.28821279e-02 3.76055650e-02 -4.60766733e-01
8.75026524e-01 6.28825501e-02 2.64265269e-01 -5.10919333e-01
-7.85824805e-02 -3.91584307e-01 -4.44758594e-01 2.23730609e-01
-1.25208127e+00 6.05378211e-01 -2.29674295e-01 4.93273646e-01
-3.77790332e-01 -5.48882365e-01 2.08479613e-01 -7.00829849e-02
-6.09663188e-01 -4.82942536e-02 -7.32559338e-02 9.97877061e-01
3.58478725e-01 4.38745528e-01 -1.28150403e+00 -4.80728209e-01
-6.43114746e-01 8.06404471e-01 -4.15657042e-03 6.88129842e-01
3.53739142e-01 -1.32448101e+00 -1.07455003e+00 -7.49130785e-01
6.61189020e-01 -4.06063795e-01 -1.49881423e-01 4.91124541e-01
-4.14424866e-01 1.29652369e+00 1.35413334e-01 -9.94169638e-02
-1.20089412e+00 1.60485968e-01 -6.64449763e-04 -1.08740568e+00
-2.25841135e-01 7.89639652e-01 -4.74648505e-01 -6.73755527e-01
3.06111157e-01 2.02728547e-02 -8.19294274e-01 4.34496343e-01
4.67680067e-01 8.32155406e-01 2.81403720e-01 -2.11161375e-01
-1.21635608e-01 3.13136697e-01 -7.18520403e-01 -3.50557476e-01
1.22607136e+00 -1.42309055e-01 -3.99375975e-01 4.03685048e-02
1.61476815e+00 2.11732477e-01 1.69010177e-01 -6.26502395e-01
6.28266394e-01 -5.20600602e-02 8.77411738e-02 -1.49058795e+00
-5.95907211e-01 5.20510495e-01 3.65317076e-01 3.26006562e-01
5.29386759e-01 1.23724431e-01 6.30989432e-01 1.45317316e+00
1.59512565e-01 -1.71438503e+00 1.99268892e-01 1.03974962e+00
1.10497701e+00 -1.43239343e+00 2.46936306e-01 3.10827792e-02
-5.67279160e-01 9.94572401e-01 7.07663059e-01 3.55829179e-01
7.70717680e-01 -4.25032377e-01 -1.43961802e-01 -8.57852101e-01
-8.63750041e-01 1.08046308e-01 8.67246568e-01 -7.91974440e-02
7.79750645e-01 -1.13436833e-01 -6.89122736e-01 5.76837301e-01
-4.70735401e-01 -5.24736404e-01 1.89913809e-01 7.93589234e-01
-3.71336550e-01 -1.07737267e+00 1.20356604e-01 9.06687915e-01
-7.34922886e-01 -6.29989207e-01 -6.69513345e-01 6.28012955e-01
-6.23395026e-01 9.20659482e-01 -1.42562985e-01 -2.02503100e-01
5.12093544e-01 3.27719063e-01 -1.79436564e-01 -7.33502686e-01
-9.11027610e-01 -2.90732563e-01 7.14732945e-01 -5.29173434e-01
-1.85379565e-01 -8.55983436e-01 -9.14747417e-01 -1.20926000e-01
-7.22502887e-01 9.34782207e-01 6.78625107e-01 3.50188702e-01
9.53228652e-01 2.85260826e-01 5.11522651e-01 4.50038053e-02
-7.29100227e-01 -1.34077525e+00 -1.33102447e-01 6.03463829e-01
-1.78357810e-01 -6.64629877e-01 -2.72858441e-01 -1.89919770e-02] | [11.271039009094238, 7.8105292320251465] |
cafde926-c2ec-4023-9191-48a935ba3e2f | self-supervised-representation-learning-from-4 | 2103.12676 | null | https://arxiv.org/abs/2103.12676v2 | https://arxiv.org/pdf/2103.12676v2.pdf | Self-supervised representation learning from 12-lead ECG data | Clinical 12-lead electrocardiography (ECG) is one of the most widely encountered kinds of biosignals. Despite the increased availability of public ECG datasets, label scarcity remains a central challenge in the field. Self-supervised learning represents a promising way to alleviate this issue. In this work, we put forward the first comprehensive assessment of self-supervised representation learning from clinical 12-lead ECG data. To this end, we adapt state-of-the-art self-supervised methods based on instance discrimination and latent forecasting to the ECG domain. In a first step, we learn contrastive representations and evaluate their quality based on linear evaluation performance on a recently established, comprehensive, clinical ECG classification task. In a second step, we analyze the impact of self-supervised pretraining on finetuned ECG classifiers as compared to purely supervised performance. For the best-performing method, an adaptation of contrastive predictive coding, we find a linear evaluation performance only 0.5% below supervised performance. For the finetuned models, we find improvements in downstream performance of roughly 1% compared to supervised performance, label efficiency, as well as robustness against physiological noise. This work clearly establishes the feasibility of extracting discriminative representations from ECG data via self-supervised learning and the numerous advantages when finetuning such representations on downstream tasks as compared to purely supervised training. As first comprehensive assessment of its kind in the ECG domain carried out exclusively on publicly available datasets, we hope to establish a first step towards reproducible progress in the rapidly evolving field of representation learning for biosignals. | ['Nils Strodthoff', 'Temesgen Mehari'] | 2021-03-23 | null | null | null | null | ['ecg-classification', 'electrocardiography-ecg'] | ['medical', 'methodology'] | [ 7.71794736e-01 1.93107337e-01 -1.26002848e-01 -5.46404719e-01
-1.11411798e+00 -5.32130718e-01 1.79765165e-01 6.06413782e-01
-4.60630655e-01 8.73233914e-01 1.23051912e-01 -2.37847522e-01
-5.10746300e-01 -4.35601890e-01 -4.56380635e-01 -7.25198090e-01
-4.34074730e-01 6.29396856e-01 -1.65414006e-01 -6.18073940e-02
-9.27388147e-02 4.77818429e-01 -1.34715486e+00 4.53533649e-01
7.74195254e-01 1.05367124e+00 -3.29438031e-01 5.95115781e-01
3.89454216e-01 6.49608791e-01 -4.95956719e-01 -7.95392618e-02
6.64620250e-02 -7.31781721e-01 -7.91114926e-01 -6.92077950e-02
1.81460738e-01 9.32835490e-02 8.01301226e-02 5.45013309e-01
1.03723550e+00 -2.26360470e-01 8.72087836e-01 -6.93757296e-01
-2.75975317e-01 8.12514484e-01 -2.78454926e-02 5.12297630e-01
1.25749275e-01 2.74853379e-01 9.07943130e-01 -5.11181831e-01
5.24273515e-01 4.72177178e-01 1.14466548e+00 4.98333901e-01
-1.64888275e+00 -5.72852910e-01 -2.25917920e-01 4.04263288e-02
-1.42126369e+00 -6.22507393e-01 8.27633023e-01 -6.87395155e-01
8.71573627e-01 2.03111559e-01 4.83383179e-01 1.04799402e+00
1.38288394e-01 3.52069139e-01 1.43099010e+00 -5.07341683e-01
2.69410461e-01 1.42123714e-01 2.08666772e-01 5.40579677e-01
1.82092011e-01 3.68446440e-01 -3.69161457e-01 -3.69144350e-01
4.73441243e-01 -1.13223173e-01 -2.47161672e-01 -4.35790032e-01
-1.34788609e+00 5.74425995e-01 2.59705067e-01 5.53710043e-01
-4.12424177e-01 2.48718802e-02 7.54148245e-01 4.27955419e-01
6.45809472e-01 8.42682004e-01 -7.15125918e-01 -3.56603295e-01
-1.35194731e+00 -7.60224909e-02 6.63108885e-01 3.29490811e-01
3.82823795e-01 1.38148442e-01 -3.07008654e-01 9.44234610e-01
-2.63450760e-02 2.37865150e-01 8.24288368e-01 -6.71515644e-01
1.08769685e-01 5.18175483e-01 -3.02676469e-01 -7.21378267e-01
-7.49980211e-01 -1.14992154e+00 -1.12533629e+00 9.23244879e-02
4.09543365e-01 -2.02335581e-01 -7.27089584e-01 1.57524467e+00
-1.36072740e-01 1.55070990e-01 4.92895991e-02 7.46325970e-01
8.05085480e-01 1.46902055e-01 2.91159928e-01 -7.08943605e-01
1.18999887e+00 -3.84805053e-01 -4.82146055e-01 1.41714975e-01
8.58495355e-01 -3.87242377e-01 7.26809263e-01 5.87416947e-01
-7.99145520e-01 -7.45905817e-01 -1.13940299e+00 3.78455907e-01
-1.01185977e-01 2.86209315e-01 6.35348618e-01 8.74308884e-01
-8.59945536e-01 1.15974998e+00 -8.54030907e-01 -3.12782049e-01
6.80895090e-01 4.50834900e-01 -5.31766415e-01 1.54701769e-01
-1.31782281e+00 8.41318071e-01 2.91992515e-01 3.04574538e-02
-7.75113821e-01 -8.71627986e-01 -6.07274652e-01 4.41794991e-02
1.30774394e-01 -5.27482748e-01 7.75043309e-01 -1.01050174e+00
-1.34083676e+00 1.24201965e+00 1.67887360e-01 -8.16453338e-01
5.88750482e-01 -3.81601974e-02 -4.98406619e-01 2.23328322e-01
5.93168437e-02 2.57518083e-01 8.56081665e-01 -1.01387680e+00
-1.58351287e-01 -3.65129918e-01 -4.64708030e-01 -9.01450440e-02
-2.49071777e-01 -2.23936036e-01 1.57583043e-01 -7.28168070e-01
1.94646075e-01 -8.89935195e-01 -3.36224705e-01 -3.13065946e-01
-1.27056196e-01 -1.05478510e-01 8.38114619e-02 -3.82422388e-01
1.31926513e+00 -2.34582472e+00 7.94621482e-02 2.38602757e-01
4.14744169e-01 3.62575382e-01 1.15460470e-01 3.64387423e-01
-6.11382902e-01 7.31651932e-02 -5.71770608e-01 -3.22052062e-01
-2.45505556e-01 7.89417326e-02 -2.51690924e-01 6.40287638e-01
3.53039354e-01 9.86616254e-01 -9.38664973e-01 -4.52909559e-01
1.07392468e-01 4.08578217e-01 -3.67880762e-01 1.16218172e-01
2.28539154e-01 8.66433144e-01 -1.66083187e-01 3.85941923e-01
1.84557617e-01 -4.00106221e-01 4.27524269e-01 -2.41285816e-01
1.60959482e-01 2.64808089e-01 -8.54953110e-01 1.88196588e+00
-2.57705986e-01 4.09256399e-01 -6.59583449e-01 -1.48890996e+00
1.06012249e+00 6.38689220e-01 9.44889367e-01 -4.88982707e-01
4.03583646e-02 3.69251549e-01 3.09714258e-01 -3.84789526e-01
-2.09553063e-01 -7.22665429e-01 -4.42346334e-02 4.23133761e-01
3.41455936e-01 1.36933938e-01 -4.65454124e-02 -1.38198406e-01
1.27917743e+00 2.64736503e-01 6.15361154e-01 -3.60727459e-01
4.61850494e-01 -8.84212106e-02 4.36231762e-01 8.85175169e-01
-3.36463302e-01 9.69318569e-01 5.29243410e-01 -6.77028000e-01
-6.03387415e-01 -9.68355536e-01 -6.77959859e-01 8.57403398e-01
-5.76439381e-01 -5.91096461e-01 -5.28310716e-01 -8.17451656e-01
1.14988275e-01 2.18535393e-01 -6.78382397e-01 -4.42257017e-01
-4.99500185e-01 -1.15217066e+00 1.00314355e+00 6.18906438e-01
1.25256792e-01 -1.04339194e+00 -8.44838142e-01 3.90200824e-01
1.08702416e-02 -8.77844572e-01 1.87749967e-01 7.96606839e-01
-1.37203825e+00 -9.94653046e-01 -9.55563068e-01 -5.09553611e-01
3.65853935e-01 -4.89857882e-01 1.30182695e+00 2.26976406e-02
-5.49109161e-01 2.64954269e-01 -3.51578265e-01 -4.15302277e-01
-4.65929359e-01 2.92788297e-01 1.46567076e-01 1.98840305e-01
1.18308656e-01 -8.04468036e-01 -7.20769763e-01 4.99563962e-02
-5.70602238e-01 -3.73860270e-01 7.29499698e-01 1.04110253e+00
7.47259974e-01 -1.74252406e-01 1.02485836e+00 -1.38607907e+00
5.81010044e-01 -4.86575395e-01 -1.38301060e-01 2.02783220e-03
-1.18466365e+00 1.50597945e-01 6.83572888e-01 -2.35704929e-01
-5.76550245e-01 3.57964039e-01 -3.23945075e-01 -2.11067498e-01
-4.93532628e-01 7.10698307e-01 3.42905939e-01 9.44582745e-02
1.30088782e+00 2.03632250e-01 9.84802023e-02 -4.51129019e-01
1.51041940e-01 6.44299746e-01 4.10733402e-01 -4.67782438e-01
5.29153287e-01 3.65451068e-01 2.83767313e-01 -6.47532821e-01
-7.75447667e-01 -6.29127324e-01 -9.19249535e-01 -1.99166760e-02
6.99688375e-01 -7.95418441e-01 -3.97920221e-01 1.21600054e-01
-5.18334389e-01 -3.72303545e-01 -8.06203842e-01 3.97424102e-01
-7.77685404e-01 5.77054024e-01 -4.18614417e-01 -8.54198933e-01
-4.81578976e-01 -7.72288859e-01 1.02827430e+00 -5.04723549e-01
-5.44821560e-01 -1.05918610e+00 3.14080983e-01 2.89371192e-01
5.13872504e-01 5.48910797e-01 9.20552790e-01 -8.86065185e-01
2.96230447e-02 -3.65264535e-01 9.05018002e-02 6.77261710e-01
3.19160849e-01 -5.34722745e-01 -1.28338325e+00 -3.81659627e-01
-1.90761611e-02 -4.31284010e-01 1.18448150e+00 3.45133871e-01
1.12097526e+00 2.00523481e-01 -2.10088432e-01 6.62023127e-01
1.18492949e+00 2.94882916e-02 5.78731954e-01 1.17596477e-01
3.78205776e-01 3.72205406e-01 3.62003803e-01 4.38572168e-01
-1.02456911e-02 4.76807624e-01 4.68307026e-02 -4.65972275e-01
-2.11664051e-01 -2.48129703e-02 -5.44927418e-02 6.91538036e-01
-4.75852877e-01 3.42097878e-01 -1.06009459e+00 3.22093308e-01
-1.66027200e+00 -8.77693057e-01 -1.13874666e-01 2.49339032e+00
9.27983701e-01 3.44681531e-01 3.56755674e-01 8.59864056e-01
3.26193243e-01 -5.25279157e-02 -3.81866336e-01 -2.54622370e-01
-2.04906940e-01 6.25227034e-01 3.11864942e-01 4.18456569e-02
-1.32828116e+00 5.17030239e-01 6.53413773e+00 4.46762681e-01
-1.32982957e+00 2.66140044e-01 8.89004648e-01 1.07580811e-01
1.99438214e-01 3.56682134e-03 -3.55116069e-01 3.87501091e-01
1.34944344e+00 1.92697361e-01 1.21184327e-01 6.70773208e-01
2.54819930e-01 2.74994493e-01 -1.32337296e+00 1.20855975e+00
2.12677196e-01 -1.31268287e+00 -3.28273505e-01 -3.14550586e-02
5.53124666e-01 5.82302883e-02 -1.34082481e-01 4.40602511e-01
-5.25897980e-01 -1.29004157e+00 2.88437307e-01 6.16717041e-01
1.35037458e+00 -4.71746385e-01 1.05899549e+00 2.58630008e-01
-7.15413034e-01 -1.63763568e-01 -1.87049747e-01 -2.57934093e-01
4.59493622e-02 8.95537138e-01 -8.02259266e-01 6.01493359e-01
4.90244657e-01 1.11657202e+00 -5.99014878e-01 1.01864493e+00
-4.94762994e-02 1.11262286e+00 -1.59735344e-02 4.09464777e-01
-2.58464307e-01 2.32587814e-01 3.14424813e-01 1.48817241e+00
-5.27610220e-02 1.17910422e-01 1.87351450e-01 5.88319719e-01
5.47931753e-02 4.11872804e-01 -5.59530377e-01 -1.31380886e-01
-1.25885710e-01 1.11671078e+00 -7.33856499e-01 -4.66215074e-01
3.00690513e-02 9.13167298e-01 3.27123016e-01 1.85771510e-01
-6.66864038e-01 -2.42529452e-01 1.20740205e-01 3.48958284e-01
-1.41158244e-02 1.59493968e-01 -6.51639640e-01 -1.16490078e+00
-1.18572898e-01 -9.42808390e-01 6.66206181e-01 -3.07444483e-01
-1.52303672e+00 8.03324759e-01 -9.11513194e-02 -1.51024270e+00
-3.60011667e-01 -4.51568127e-01 -1.17210843e-01 8.87904227e-01
-1.57506406e+00 -8.85623634e-01 -2.25959584e-01 3.31123412e-01
2.31506348e-01 -2.62063533e-01 1.53776729e+00 3.93519521e-01
-2.40500376e-01 6.60243034e-01 6.58144802e-02 1.00415096e-01
9.91195500e-01 -1.23742330e+00 -5.50907925e-02 3.96744967e-01
3.65772545e-01 6.85179591e-01 4.28225338e-01 -3.68610919e-01
-9.39118505e-01 -9.15910661e-01 9.27340508e-01 -5.50946474e-01
3.60605955e-01 -1.59211114e-01 -8.73332083e-01 4.29691166e-01
-2.79819965e-01 4.92713451e-01 1.20368040e+00 4.57230836e-01
-3.35590422e-01 -4.23116356e-01 -8.80280912e-01 1.69725362e-02
1.02597630e+00 -6.74663842e-01 -8.55433762e-01 3.72083366e-01
6.26615137e-02 -2.82079458e-01 -1.32626522e+00 8.31659377e-01
7.76444137e-01 -9.79862630e-01 9.61932063e-01 -7.42547929e-01
3.60355318e-01 -9.57721621e-02 8.42540115e-02 -1.27984035e+00
-5.74084520e-01 -6.79576874e-01 8.28873739e-02 9.80418503e-01
5.37933290e-01 -7.12723017e-01 8.51397395e-01 1.38931528e-01
-1.79798946e-01 -9.87404287e-01 -9.56197798e-01 -6.75268888e-01
2.07753971e-01 -5.49817562e-01 3.94276120e-02 1.01209509e+00
2.52529472e-01 5.44033229e-01 -4.91995186e-01 -1.63060054e-01
6.38361633e-01 1.99661180e-01 4.38919812e-01 -1.62236428e+00
-7.58768916e-01 -1.77480176e-01 -8.98966074e-01 -5.33682287e-01
-5.96021395e-03 -1.39937961e+00 -8.50178078e-02 -1.35740566e+00
2.67046273e-01 -6.45120323e-01 -1.16912103e+00 5.45196056e-01
-1.86953962e-01 6.93220258e-01 2.78563779e-02 3.59242469e-01
-5.16988039e-01 -1.28767593e-02 8.01913619e-01 -1.01892829e-01
-3.05826157e-01 1.56293079e-01 -7.77523398e-01 4.98677373e-01
9.86839414e-01 -7.05980480e-01 -5.17572045e-01 3.62863652e-02
7.54957572e-02 2.55880028e-01 2.62732893e-01 -1.35449195e+00
-1.93720788e-01 3.63836646e-01 6.33196294e-01 -6.25717565e-02
1.77893475e-01 -5.95972419e-01 4.54757139e-02 6.88933134e-01
-7.14189470e-01 -1.90917432e-01 1.49462923e-01 5.69018662e-01
-2.42320433e-01 2.37973649e-02 7.65107691e-01 -2.34645009e-01
-3.64933908e-01 4.72339801e-02 -3.39505255e-01 2.76452273e-01
7.36020207e-01 -2.58793265e-01 8.67822319e-02 -2.03085259e-01
-1.36036098e+00 -3.26647401e-01 -3.00163459e-02 1.63718373e-01
5.90192735e-01 -8.42881620e-01 -9.41569984e-01 4.60347801e-01
4.52528596e-01 -4.73502100e-01 3.11679244e-01 1.26617622e+00
-3.09354842e-01 5.33182025e-01 -3.94820690e-01 -8.92446578e-01
-1.19462371e+00 4.87009108e-01 4.25867617e-01 -4.22520965e-01
-7.47781038e-01 5.91141522e-01 -6.72636703e-02 -1.67917967e-01
2.03321964e-01 -4.22418743e-01 -4.24671859e-01 1.46648914e-01
1.93994358e-01 2.75285661e-01 5.33837140e-01 -4.22277004e-01
-5.65943420e-01 4.62941676e-01 2.55823612e-01 1.68562129e-01
1.41414046e+00 2.60455817e-01 7.67484531e-02 8.98360133e-01
1.12732792e+00 -1.30179241e-01 -8.91485631e-01 1.62226155e-01
1.65236130e-01 -2.84613576e-02 -2.86170449e-02 -1.11653006e+00
-8.56758118e-01 1.04668248e+00 1.11408138e+00 7.46444836e-02
1.22156894e+00 -1.01204641e-01 2.81041026e-01 3.00012380e-01
4.55211252e-01 -7.67383993e-01 -6.36997074e-02 1.31468952e-01
6.71711385e-01 -1.29431665e+00 1.90192744e-01 -3.20039272e-01
-6.88095927e-01 1.09592247e+00 -8.47391337e-02 -2.92522520e-01
8.23987544e-01 1.13092631e-01 3.15988541e-01 -1.86135456e-01
-6.66619956e-01 -1.71069086e-01 4.37759936e-01 7.07913816e-01
9.59190547e-01 1.75722301e-01 -6.02607012e-01 7.62877703e-01
1.03843315e-02 4.20634985e-01 2.17507854e-01 7.48589516e-01
-1.73110161e-02 -1.26973462e+00 4.80763055e-02 7.89279044e-01
-7.94822335e-01 -2.23922074e-01 -2.21409380e-01 4.03345585e-01
2.74380833e-01 8.78883064e-01 -3.07836682e-01 -3.40283483e-01
4.17952508e-01 5.11604726e-01 4.27448362e-01 -8.95603776e-01
-9.88368571e-01 1.07162423e-01 1.38791323e-01 -3.59612495e-01
-5.57816803e-01 -6.28198981e-01 -1.13884246e+00 6.33982420e-01
-1.44466355e-01 2.27002829e-01 4.89862591e-01 8.34155738e-01
5.35801589e-01 8.60170424e-01 4.16821599e-01 -6.47377729e-01
-9.02331710e-01 -9.61074412e-01 -6.51041746e-01 7.88746178e-01
3.41350377e-01 -5.77277720e-01 -3.99945229e-01 2.61772394e-01] | [14.293757438659668, 3.3287816047668457] |
10d9c1b6-a5dc-4c8f-aa73-e30b1ee7983e | mitosis-detection-for-breast-cancer-pathology | 2109.01526 | null | https://arxiv.org/abs/2109.01526v3 | https://arxiv.org/pdf/2109.01526v3.pdf | Mitosis Detection for Breast Cancer Pathology Images Using UV-Net | The difficulty of detecting mitosis and its similarity to non-mitosis objects has remained a challenge in computational pathology. The lack of publicly available data has added more complexity. Deep learning algorithms have shown potentials in mitosis detection tasks. However, they face challenges when applied to pathology images with dense medium and diverse dataset. This paper introduces an optimized UV-Net architecture, developed to focus on mitosis details with high-resolution through feature preservation. Stain normalization methods are used to generalize the trained network. An F1 score of 0.6721 is achieved using this network. | ['April Khademi', 'Susan Done', 'Salar Razavi', 'Samir Mitha', 'Seyed H. Mirjahanmardi'] | 2021-09-02 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 3.67929250e-01 1.24187358e-01 -3.90259549e-02 -1.88913822e-01
-8.36036623e-01 -3.12497079e-01 5.69077313e-01 4.44533646e-01
-7.77807832e-01 1.07247221e+00 -5.35439923e-02 1.01777144e-01
1.89149171e-01 -6.47953093e-01 -2.56975502e-01 -1.39578080e+00
1.01292357e-01 3.92401755e-01 4.26273078e-01 8.73622298e-02
9.40720662e-02 8.19181383e-01 -1.12201238e+00 2.59626776e-01
4.38314915e-01 6.57517791e-01 2.58597702e-01 1.14536548e+00
-2.17977408e-02 5.74703813e-01 -5.99572599e-01 -3.13688278e-01
-8.26626644e-02 -2.19376966e-01 -9.26582515e-01 -4.24581096e-02
4.51612294e-01 -2.89651722e-01 -4.89209503e-01 1.02869403e+00
8.98456872e-01 -2.84645587e-01 9.63840187e-01 -8.84428024e-01
-4.83055115e-01 2.31638223e-01 -5.71813166e-01 9.21143115e-01
-1.39244422e-01 1.76714152e-01 6.73847198e-01 -5.03102362e-01
1.31393540e+00 7.25369930e-01 7.70066142e-01 8.14067423e-01
-1.30724621e+00 -1.50686309e-01 -7.67307818e-01 1.14482120e-01
-1.18656623e+00 -9.33330730e-02 4.89366472e-01 -3.72714460e-01
1.26910424e+00 2.49812365e-01 6.19978011e-01 1.12394702e+00
7.64719129e-01 5.59167802e-01 1.16657996e+00 -1.63381800e-01
6.20494364e-03 -6.03575222e-02 2.03061447e-01 3.90593529e-01
5.73674619e-01 -3.05017501e-01 -3.85553420e-01 3.06385458e-01
9.16622519e-01 2.25205287e-01 -4.17655736e-01 -5.14133312e-02
-1.34822965e+00 6.14526212e-01 5.44829726e-01 9.65420723e-01
1.91348661e-02 6.00999929e-02 6.70414805e-01 4.69811894e-02
3.67211878e-01 3.99962246e-01 -3.94222677e-01 -6.26901630e-03
-9.58555520e-01 9.09883454e-02 3.82011890e-01 3.63523811e-01
3.35783929e-01 -8.94343257e-02 -2.96592265e-01 4.77346092e-01
7.93512911e-02 2.51655370e-01 6.66757762e-01 -5.74687958e-01
-2.44633287e-01 9.23639178e-01 -5.45369506e-01 -8.74659836e-01
-1.03469288e+00 -6.46928906e-01 -1.51442981e+00 4.85118657e-01
9.23784852e-01 2.86001354e-01 -1.31849384e+00 1.23970687e+00
4.55805987e-01 7.91345071e-03 1.26292601e-01 8.93293798e-01
1.16931701e+00 2.79121816e-01 6.81404620e-02 -7.33689219e-02
1.37220752e+00 -6.77684426e-01 -9.67434525e-01 -1.56500582e-02
7.36616015e-01 -8.06660295e-01 8.10582817e-01 1.44258291e-01
-9.53910530e-01 -1.66775361e-01 -1.19613540e+00 -7.66761482e-01
-9.07470524e-01 1.11403741e-01 7.62911260e-01 3.31117719e-01
-1.38365257e+00 6.32171631e-01 -1.04154682e+00 -9.27074313e-01
9.55710411e-01 7.49834418e-01 -7.32491195e-01 1.37839928e-01
-5.17778873e-01 9.04389739e-01 5.81923068e-01 4.65208553e-02
-6.70895576e-01 -8.28933656e-01 -6.94079936e-01 1.89905301e-01
-2.66994685e-01 -8.57660592e-01 9.91082311e-01 -2.81417817e-01
-1.22636545e+00 1.49789727e+00 5.66603020e-02 -3.11904967e-01
5.76636314e-01 6.77118123e-01 5.17747249e-04 4.22804177e-01
-5.62845804e-02 1.02602279e+00 4.21824269e-02 -8.43126237e-01
-6.96279228e-01 -3.87126297e-01 -2.80954063e-01 -7.59820789e-02
-4.12632972e-01 -9.19180438e-02 -3.67661208e-01 -6.67938769e-01
8.55037794e-02 -5.27773559e-01 -9.63960886e-02 5.43730855e-01
-2.67886907e-01 -5.62529685e-03 1.02746785e+00 -7.50498950e-01
5.80369055e-01 -1.97555256e+00 7.67778233e-02 -4.32456806e-02
4.71029073e-01 3.66509885e-01 6.92064166e-02 -5.44453934e-02
3.31421942e-02 1.27112284e-01 -1.60730764e-01 -3.85288566e-01
-1.66089967e-01 2.42549732e-01 5.15882134e-01 8.74879181e-01
4.37606573e-01 1.17000413e+00 -8.31187248e-01 -8.27757001e-01
2.31662065e-01 9.83390689e-01 -1.26288861e-01 6.88280091e-02
2.60709703e-01 3.37089717e-01 2.41455317e-01 1.28884399e+00
8.19872499e-01 -6.40827179e-01 1.97944894e-01 -2.62483597e-01
2.28723809e-01 -2.89297014e-01 -6.30073130e-01 1.49291170e+00
2.27451205e-01 1.07025039e+00 3.88224334e-01 -8.07540596e-01
6.43772185e-01 2.07348689e-01 3.40223312e-01 -5.23792863e-01
6.46097720e-01 2.82230854e-01 2.45371744e-01 -4.52219427e-01
2.48684883e-01 -2.90324450e-01 4.02190953e-01 5.37139587e-02
5.61392665e-01 -6.35040924e-02 6.25524819e-01 5.17030843e-02
1.29431593e+00 -3.64589423e-01 3.96800578e-01 -3.69773626e-01
5.35048664e-01 -8.92646983e-02 6.21417046e-01 3.13838214e-01
-6.97892666e-01 8.91848505e-01 6.28727376e-01 -7.50876009e-01
-1.26270640e+00 -1.02719378e+00 -3.82921904e-01 5.23918808e-01
-8.66319612e-02 8.59484449e-02 -6.79903030e-01 -6.70391738e-01
-1.18441574e-01 -4.13025230e-01 -1.13590896e+00 -2.32385565e-02
-5.97775996e-01 -1.25452912e+00 6.70310855e-01 6.07914150e-01
5.56739151e-01 -9.74722981e-01 -4.83978570e-01 5.84732853e-02
8.16392675e-02 -1.06487620e+00 -1.74167976e-01 5.95622838e-01
-8.00198555e-01 -1.37066078e+00 -1.13039863e+00 -1.31061637e+00
1.03784108e+00 -4.16581817e-02 1.10574889e+00 3.30457270e-01
-1.05180645e+00 -2.61466026e-01 1.04202203e-01 -4.00730073e-01
-2.94229180e-01 3.25197637e-01 -3.21776420e-01 -5.29076040e-01
5.59618771e-01 -3.51751089e-01 -7.76284695e-01 -1.85883835e-01
-1.15872443e+00 -8.58433172e-02 8.42847705e-01 1.23371911e+00
1.02147186e+00 3.28774862e-02 3.84697646e-01 -8.67554128e-01
2.66204774e-01 -6.49819523e-02 -1.85596034e-01 4.82565578e-04
-3.99592727e-01 -4.59027141e-01 2.94637352e-01 -5.03624380e-01
-7.95717299e-01 1.98500022e-01 -1.89125642e-01 7.83878192e-02
-1.97711542e-01 2.64561355e-01 7.35999737e-03 -6.31022096e-01
5.47606409e-01 -1.25832781e-02 2.85141230e-01 1.01371042e-01
-4.01387215e-01 3.35676104e-01 9.55764592e-01 1.53535396e-01
5.05629599e-01 9.11514282e-01 5.67402065e-01 -7.59713054e-01
-6.05782390e-01 -6.52635753e-01 -7.20403194e-01 -2.20326141e-01
1.06177878e+00 -5.50535202e-01 -7.62727916e-01 9.01750624e-01
-8.52865994e-01 -3.72371197e-01 -2.73836315e-01 2.88300455e-01
-3.84328604e-01 3.06078583e-01 -1.27880299e+00 -1.29342347e-01
-7.29415536e-01 -9.23938751e-01 9.73378003e-01 8.50069344e-01
-2.55477339e-01 -1.25949287e+00 3.85533422e-01 4.40535754e-01
4.32756573e-01 8.33271444e-01 8.26574981e-01 -5.77491641e-01
-5.62839627e-01 -3.61986160e-01 -4.66047764e-01 -3.09178848e-02
4.33213741e-01 5.34747303e-01 -1.38045061e+00 -3.87549818e-01
-4.00075704e-01 -1.95598722e-01 1.09116244e+00 8.15176487e-01
1.08497214e+00 1.57358155e-01 -7.29463279e-01 9.16075885e-01
1.51637232e+00 -4.64653783e-02 6.54081762e-01 7.31160581e-01
3.15740883e-01 3.93249780e-01 2.04806343e-01 -2.25595444e-01
-9.33643617e-03 -3.27622071e-02 4.69209433e-01 -8.06110203e-01
-6.30221009e-01 3.85417670e-01 -4.53588217e-01 5.17741561e-01
1.92800146e-02 -2.93392181e-01 -1.16097832e+00 8.53830636e-01
-1.53360903e+00 -8.50375593e-01 -1.57306865e-01 1.41709137e+00
8.34567726e-01 3.38909715e-01 -8.47449154e-02 4.86619532e-01
6.76647842e-01 -2.50803441e-01 -5.44854403e-01 -2.22430080e-01
-7.34561920e-01 2.13160440e-01 2.33233795e-01 2.41902962e-01
-1.37094569e+00 5.67581058e-01 7.26482153e+00 9.34375703e-01
-1.19495595e+00 1.15906067e-01 1.11723244e+00 -5.74014969e-02
1.23583399e-01 -6.24136627e-01 -6.81891561e-01 3.40386540e-01
4.28897858e-01 -1.56632155e-01 -3.44637126e-01 3.20722729e-01
-5.42247519e-02 -2.90854394e-01 -1.02308452e+00 1.00533485e+00
9.99400988e-02 -1.73864269e+00 -1.07346646e-01 4.00860637e-01
7.43346870e-01 -1.94230396e-02 2.99544781e-01 1.09724045e-01
-1.82105303e-01 -1.24818826e+00 -3.12751919e-01 5.12826264e-01
8.71040285e-01 -8.27105105e-01 1.69070792e+00 -5.16471826e-02
-9.77044404e-01 2.10858986e-01 -5.17335296e-01 -4.34836233e-03
-1.00535326e-01 7.39454925e-01 -1.46479261e+00 1.19781554e-01
7.91642368e-01 5.95978856e-01 -8.08694422e-01 1.52751863e+00
3.17671090e-01 2.12476254e-01 -2.96967804e-01 -2.64396608e-01
5.02374806e-02 1.34892061e-01 1.49720088e-01 1.81389439e+00
3.33496958e-01 -3.56215417e-01 -1.27682135e-01 6.86514318e-01
-2.44978115e-01 5.49960323e-02 -4.93212551e-01 -2.47740492e-01
-1.26488596e-01 1.93851197e+00 -1.70217574e+00 -1.72994897e-01
-9.40126181e-02 6.72790170e-01 4.49543327e-01 1.28129527e-01
-4.08464432e-01 -5.03696501e-01 6.20339274e-01 1.35020629e-01
8.62705857e-02 1.66602135e-01 -5.08218646e-01 -8.06620777e-01
-3.85722220e-01 -4.97277111e-01 5.69312871e-01 -5.22643566e-01
-1.41585851e+00 3.81920099e-01 -6.78279817e-01 -7.37630188e-01
2.55477250e-01 -1.03650701e+00 -7.37543106e-01 5.54943144e-01
-1.61600018e+00 -1.44221282e+00 -4.35800225e-01 2.32199460e-01
3.72818053e-01 -7.50467256e-02 9.50416267e-01 3.84578645e-01
-6.49053574e-01 5.84068179e-01 2.04206705e-01 2.39102393e-01
8.78569603e-01 -1.77676666e+00 -6.08957820e-02 7.28958249e-01
-5.46218216e-01 4.14045870e-01 7.05583274e-01 -5.55796564e-01
-8.31865251e-01 -1.01626074e+00 1.17582238e+00 -2.90319383e-01
5.42667091e-01 -3.70360523e-01 -9.93097603e-01 3.49674642e-01
4.01510954e-01 4.91405904e-01 1.09196293e+00 -3.60926151e-01
-2.70723253e-02 1.47093624e-01 -1.75548851e+00 4.99191493e-01
5.61470568e-01 -4.12735552e-01 -7.80831203e-02 4.24995750e-01
3.14326346e-01 -7.64424622e-01 -1.12868536e+00 4.63025093e-01
2.95172602e-01 -8.59912038e-01 8.67822528e-01 -8.38318020e-02
3.15741777e-01 -4.05842930e-01 3.41525495e-01 -9.56516683e-01
-6.36227727e-01 -2.16287255e-01 -2.45325699e-01 1.21814167e+00
1.69305533e-01 -3.42067093e-01 1.42662215e+00 -2.93027086e-04
-1.89028233e-01 -1.12863123e+00 -1.21034598e+00 -3.40606004e-01
3.13846260e-01 5.32264888e-01 1.53574631e-01 8.57623398e-01
1.40633747e-01 1.16673693e-01 1.33359462e-01 -7.53685907e-02
6.54466689e-01 -2.77749628e-01 3.81200284e-01 -1.38187432e+00
1.45217821e-01 -7.47024417e-01 -1.16042137e+00 -1.22055039e-01
7.71148577e-02 -9.43814218e-01 -3.05520982e-01 -1.62871480e+00
8.20722103e-01 2.63999522e-01 -3.82063180e-01 5.21679819e-01
-1.83695957e-01 9.91286159e-01 -1.73481777e-01 -8.92378911e-02
-4.97031897e-01 4.38619927e-02 1.57575107e+00 -6.92671895e-01
3.27535361e-01 -5.33648431e-01 -4.84131783e-01 6.89586818e-01
1.21837509e+00 -2.84854621e-01 -4.40653972e-02 1.56155461e-02
1.49110287e-01 -5.82035601e-01 2.18495265e-01 -1.39378381e+00
5.38152874e-01 1.87270194e-01 1.14891601e+00 -9.02981639e-01
4.17821735e-01 -6.08057499e-01 1.57688066e-01 7.37948895e-01
-1.63858622e-01 -2.28674725e-01 2.60258526e-01 3.68926585e-01
-2.04420730e-01 -6.22256622e-02 9.53772426e-01 -7.86126852e-02
-2.73205101e-01 2.67487794e-01 -6.63305283e-01 -3.12399805e-01
1.25099015e+00 -7.21259654e-01 -8.11788201e-01 2.28717670e-01
-8.01844954e-01 2.27254033e-01 7.63892710e-01 -1.46735698e-01
5.49150169e-01 -1.29274964e+00 -4.75705445e-01 7.29605928e-02
5.29651903e-02 3.62203330e-01 4.36815411e-01 9.76736903e-01
-1.14195037e+00 5.54525077e-01 -7.24799454e-01 -9.10411775e-01
-1.79277611e+00 3.76954615e-01 5.98457813e-01 -7.43348241e-01
-6.51383817e-01 1.11232853e+00 3.37687321e-02 -4.68573362e-01
3.47837389e-01 -5.21265388e-01 -5.04658520e-01 -3.04161496e-02
7.44162202e-01 3.80578309e-01 3.04101229e-01 -4.79984909e-01
-3.93721521e-01 5.09181976e-01 -4.32836950e-01 6.47024810e-01
1.52835429e+00 6.86249584e-02 -3.33489001e-01 2.78492570e-01
1.17154348e+00 -5.34672081e-01 -1.15685010e+00 8.63549784e-02
-3.95180918e-02 -2.46643379e-01 1.28427714e-01 -9.44703698e-01
-1.31530893e+00 7.85970688e-01 9.57103193e-01 4.77828756e-02
9.74931717e-01 1.71722472e-03 7.75545120e-01 4.13809091e-01
-6.42113313e-02 -1.07823622e+00 8.39745477e-02 4.78199095e-01
3.13092142e-01 -1.48641980e+00 -7.30579868e-02 -5.39954305e-01
1.78949252e-01 1.41472471e+00 8.37143064e-01 -5.95072284e-02
3.42912614e-01 8.89950097e-01 3.54277670e-01 -3.70995671e-01
-6.39044404e-01 -2.12105438e-01 -1.17764063e-01 1.10501659e+00
7.56505966e-01 -2.63971448e-01 -4.48912889e-01 3.47291946e-01
3.41235958e-02 2.13575676e-01 8.21962297e-01 9.97674406e-01
-3.33565563e-01 -8.00160885e-01 -2.27301329e-01 5.98617613e-01
-9.66432869e-01 3.30155343e-01 -5.83811820e-01 9.92317319e-01
1.80901289e-01 3.61805648e-01 3.24142188e-01 1.44412875e-01
-2.28917927e-01 -2.77736247e-01 5.72239578e-01 -3.51726860e-01
-7.02000678e-01 3.03852022e-01 -2.65429854e-01 -1.90479636e-01
-5.72236419e-01 -4.53088909e-01 -1.52442074e+00 -5.02574801e-01
-9.38712582e-02 -3.20241183e-01 4.21820790e-01 5.27546585e-01
1.90517306e-01 9.46612239e-01 -1.10448219e-01 -8.42868507e-01
2.55927108e-02 -9.67789769e-01 -9.43223000e-01 3.09350520e-01
5.84338248e-01 -4.42792356e-01 -4.89228249e-01 3.96602780e-01] | [15.076482772827148, -3.1472885608673096] |
d69c6239-565b-41b0-a341-05eff5a41385 | reformulating-ctr-prediction-learning | 2304.13643 | null | https://arxiv.org/abs/2304.13643v1 | https://arxiv.org/pdf/2304.13643v1.pdf | Reformulating CTR Prediction: Learning Invariant Feature Interactions for Recommendation | Click-Through Rate (CTR) prediction plays a core role in recommender systems, serving as the final-stage filter to rank items for a user. The key to addressing the CTR task is learning feature interactions that are useful for prediction, which is typically achieved by fitting historical click data with the Empirical Risk Minimization (ERM) paradigm. Representative methods include Factorization Machines and Deep Interest Network, which have achieved wide success in industrial applications. However, such a manner inevitably learns unstable feature interactions, i.e., the ones that exhibit strong correlations in historical data but generalize poorly for future serving. In this work, we reformulate the CTR task -- instead of pursuing ERM on historical data, we split the historical data chronologically into several periods (a.k.a, environments), aiming to learn feature interactions that are stable across periods. Such feature interactions are supposed to generalize better to predict future behavior data. Nevertheless, a technical challenge is that existing invariant learning solutions like Invariant Risk Minimization are not applicable, since the click data entangles both environment-invariant and environment-specific correlations. To address this dilemma, we propose Disentangled Invariant Learning (DIL) which disentangles feature embeddings to capture the two types of correlations separately. To improve the modeling efficiency, we further design LightDIL which performs the disentanglement at the higher level of the feature field. Extensive experiments demonstrate the effectiveness of DIL in learning stable feature interactions for CTR. We release the code at https://github.com/zyang1580/DIL. | ['Yongdong Zhang', 'Xiangnan He', 'Dingxian Wang', 'Wenjie Wang', 'Fuli Feng', 'Tianhao Shi', 'Yang Zhang'] | 2023-04-26 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-2.86133915e-01 -5.85250139e-01 -1.56148911e-01 -4.68968958e-01
-4.30043906e-01 -5.46647251e-01 4.46302682e-01 -2.37854123e-02
-2.46239334e-01 5.91063440e-01 3.20837796e-01 -4.05887038e-01
-6.12753093e-01 -7.58735478e-01 -6.30765140e-01 -7.34858334e-01
-3.31944585e-01 1.61283668e-02 -1.20071627e-01 -4.82808143e-01
8.01276267e-02 1.72672823e-01 -1.64185035e+00 2.02329695e-01
9.43738401e-01 9.20130432e-01 1.39079198e-01 4.37284917e-01
1.87781140e-01 3.04226786e-01 -1.26003072e-01 -5.15636325e-01
4.97346878e-01 -1.28742486e-01 -3.22209448e-01 -4.04441237e-01
2.08842278e-01 -1.98174328e-01 -6.25934839e-01 1.06449556e+00
2.34754264e-01 4.11789626e-01 5.29185891e-01 -1.27522099e+00
-8.56097579e-01 7.73367345e-01 -3.33798051e-01 2.26407334e-01
1.37372568e-01 2.10284982e-02 1.63325489e+00 -9.35315073e-01
1.30373567e-01 9.95526135e-01 3.89463097e-01 4.40285653e-01
-1.32231593e+00 -8.16698015e-01 6.02450788e-01 3.32022786e-01
-1.25103092e+00 -1.87002551e-02 8.68402004e-01 -5.59833884e-01
2.92514741e-01 7.57053435e-01 6.26894236e-01 1.47595835e+00
3.48524988e-01 1.03734314e+00 9.51076806e-01 9.89840403e-02
1.65364206e-01 1.56146809e-01 6.24356866e-01 3.24169070e-01
3.08998406e-01 1.58071429e-01 -7.73744702e-01 -1.28442958e-01
4.64452356e-01 6.48304343e-01 -4.18132424e-01 -3.92422289e-01
-1.04456437e+00 9.07781899e-01 5.84087372e-01 3.04217428e-01
-1.00766286e-01 -2.37249345e-01 1.84975430e-01 6.16924405e-01
4.72360790e-01 6.86828852e-01 -8.65578711e-01 -1.16013765e-01
-5.49801111e-01 3.57817352e-01 6.28199100e-01 6.07999444e-01
7.39249170e-01 -2.13838860e-01 -3.43631089e-01 5.88766456e-01
2.93260843e-01 2.59663999e-01 5.38768888e-01 -3.82588416e-01
3.75637531e-01 6.97722912e-01 3.25152636e-01 -1.17390406e+00
-3.00967962e-01 -9.83519554e-01 -8.41731846e-01 -2.27716476e-01
3.04292083e-01 -2.94192638e-02 -5.23209810e-01 1.93621242e+00
2.26950794e-01 2.46627033e-01 -2.39843816e-01 9.34352219e-01
2.25486338e-01 4.28858668e-01 -1.69052660e-01 -1.79811731e-01
1.16275311e+00 -9.87851679e-01 -5.98303258e-01 9.48940590e-02
5.08119404e-01 -5.44469118e-01 1.36777067e+00 7.24407613e-01
-6.06854677e-01 -6.13455474e-01 -1.12473607e+00 1.38536230e-01
-6.46562696e-01 2.08897457e-01 1.02807796e+00 4.82274294e-01
-5.45802295e-01 9.77110147e-01 -8.39284301e-01 1.36353031e-01
5.82888350e-02 4.37116057e-01 -1.10846929e-01 1.04722129e-02
-1.62761533e+00 3.95780563e-01 9.01437029e-02 2.72105187e-01
-5.16255677e-01 -1.01185524e+00 -5.77615440e-01 2.04105079e-01
4.09969836e-01 -6.27091348e-01 1.02268255e+00 -6.05138004e-01
-1.31635356e+00 -1.27120381e-02 5.96355321e-03 -3.38967353e-01
5.61487675e-01 -7.01987088e-01 -6.83819890e-01 -6.12181664e-01
-1.03803404e-01 -1.82220191e-01 9.93362308e-01 -9.87287164e-01
-6.68534040e-01 -5.73873520e-01 2.30055213e-01 -2.19352380e-03
-8.48882854e-01 -3.98495823e-01 -3.08053374e-01 -6.68805301e-01
1.76940602e-03 -1.02635741e+00 -3.04015338e-01 -3.25393379e-01
-3.94819617e-01 -3.70439321e-01 5.93637407e-01 -5.42488813e-01
1.71260643e+00 -2.17898226e+00 2.06539765e-01 2.06605166e-01
4.24789816e-01 2.01438174e-01 -1.61888257e-01 3.82776856e-01
-2.00851128e-01 1.24580018e-01 3.60551625e-01 -3.61121267e-01
2.48891205e-01 4.23013344e-02 -5.95206499e-01 2.89179593e-01
-1.81544144e-02 9.70912874e-01 -9.84936833e-01 2.17016667e-01
5.12510911e-02 3.62070739e-01 -8.54951322e-01 2.18589976e-01
-3.66280764e-01 5.72397053e-01 -7.97922909e-01 3.38278949e-01
6.99067414e-01 -3.57905984e-01 9.88886282e-02 -3.65176350e-01
-3.85720283e-01 2.95044750e-01 -1.23079264e+00 1.65524936e+00
-6.61728501e-01 2.38005579e-01 -2.39871711e-01 -8.55284214e-01
7.42669702e-01 -9.56060644e-03 7.52607703e-01 -6.62603557e-01
-3.98230702e-02 1.50953338e-01 1.62192330e-01 -3.98583829e-01
6.81922793e-01 2.03517288e-01 -2.02263653e-01 4.41365957e-01
-1.07267015e-01 6.37775421e-01 2.42752079e-02 2.43819416e-01
1.11774707e+00 -8.13340861e-03 -7.35284016e-02 -1.28277242e-01
4.59983379e-01 -4.69876140e-01 6.39750957e-01 7.45475411e-01
-7.77643733e-03 3.56962353e-01 4.27676737e-01 -5.27521849e-01
-7.55009174e-01 -1.04353166e+00 -2.53123313e-01 1.22012424e+00
3.13776940e-01 -7.32858181e-01 -2.92336103e-02 -8.44029307e-01
1.72636628e-01 6.00057900e-01 -7.84276903e-01 -5.45352399e-01
-3.24980170e-01 -7.75777042e-01 -5.75976409e-02 3.41979921e-01
1.74863786e-01 -5.23378968e-01 -4.45238315e-02 2.11928934e-01
2.33537406e-02 -5.55797458e-01 -6.11199021e-01 3.91528279e-01
-5.17887592e-01 -9.33548868e-01 -4.59510654e-01 -2.18314439e-01
4.13477808e-01 5.45787573e-01 8.28429759e-01 -5.58927990e-02
-2.56779253e-01 2.52635926e-01 -5.17238617e-01 -5.82080223e-02
1.74574077e-01 2.12455049e-01 5.83447695e-01 4.13362533e-01
4.70558316e-01 -7.97647774e-01 -1.08790469e+00 5.33891797e-01
-9.51167822e-01 -1.65595859e-01 6.80845320e-01 1.11014342e+00
3.63223940e-01 3.26095194e-01 5.39108634e-01 -8.92273724e-01
8.02020669e-01 -8.47627103e-01 -4.63220090e-01 4.02386785e-01
-7.60920823e-01 2.01614708e-01 9.65551019e-01 -6.61676764e-01
-8.29872251e-01 -1.66153923e-01 -2.15280950e-02 -4.58461404e-01
1.67851105e-01 7.57129788e-01 -2.19601765e-01 1.74159229e-01
5.84208727e-01 2.09414393e-01 -3.56802523e-01 -7.40393877e-01
4.75965798e-01 6.10090792e-01 5.55066802e-02 -6.01718187e-01
1.15993035e+00 2.30057538e-01 -1.87248915e-01 -4.88950253e-01
-1.16770256e+00 -7.16564476e-01 -4.63711500e-01 -5.98636121e-02
5.55692255e-01 -8.87152731e-01 -7.59674251e-01 1.37149105e-02
-6.85085833e-01 4.30126190e-02 -1.85532600e-01 7.56793201e-01
-2.90089548e-01 1.67348683e-01 -3.26191932e-01 -7.30448365e-01
-1.72810361e-01 -9.54118490e-01 8.88522565e-01 2.95359403e-01
-2.35744361e-02 -8.38852346e-01 3.94388109e-01 2.04758465e-01
4.57106858e-01 -4.86805327e-02 8.89971793e-01 -9.92851675e-01
-7.79053748e-01 -3.82337868e-01 -1.90952756e-02 3.12007785e-01
2.66484290e-01 -1.12845182e-01 -8.58522356e-01 -4.84126061e-01
-1.82584133e-02 -1.68646634e-01 9.99346673e-01 8.81437808e-02
1.35932231e+00 -3.52339953e-01 -2.54426301e-01 5.94306827e-01
1.15916991e+00 -7.80605450e-02 2.68144339e-01 1.24536991e-01
6.34486616e-01 4.23889369e-01 9.08558846e-01 7.16935396e-01
2.97817737e-01 8.78066242e-01 4.50861156e-01 2.75669038e-01
2.91845918e-01 -6.26262128e-01 5.56927145e-01 1.10163426e+00
-6.08249102e-04 -1.22967266e-01 -2.66386926e-01 1.53401941e-01
-2.13437438e+00 -9.31056321e-01 -7.25646317e-02 2.39651394e+00
7.59661078e-01 1.41539440e-01 1.12064227e-01 -1.79617293e-02
4.49028105e-01 2.22560033e-01 -8.52858305e-01 -1.21391043e-02
1.63278759e-01 1.66146718e-02 1.72071084e-01 2.59453803e-01
-1.14677012e+00 5.42393804e-01 4.40495682e+00 8.91305745e-01
-1.20139790e+00 7.90717974e-02 4.28027391e-01 -1.77390203e-01
-6.81370080e-01 1.11218400e-01 -8.76877129e-01 5.92884839e-01
7.23615527e-01 -2.62033910e-01 6.46066248e-01 1.00123250e+00
3.03135902e-01 4.10108030e-01 -1.24110067e+00 9.77159619e-01
-2.43100986e-01 -1.00938833e+00 2.07431037e-02 1.70994937e-01
7.14996755e-01 -3.80825214e-02 6.47637665e-01 8.47086310e-01
2.41803139e-01 -8.02323163e-01 4.60947990e-01 7.79198050e-01
3.30881804e-01 -5.53768575e-01 4.42067832e-01 4.44039404e-01
-1.17353737e+00 -3.56467366e-01 -5.37779689e-01 -8.14992487e-02
-3.99647690e-02 8.81432772e-01 -2.71818161e-01 8.14753234e-01
6.62651300e-01 1.13483346e+00 -6.80278122e-01 1.07160878e+00
-2.99600940e-02 4.05077934e-01 -1.22631416e-01 -7.17858970e-02
-8.52924362e-02 -5.81326306e-01 6.40030205e-01 6.51905537e-01
4.23887551e-01 -1.17406689e-01 9.73172635e-02 8.80648851e-01
-3.86906527e-02 1.05336234e-01 -5.12610257e-01 -1.61664039e-01
2.41500631e-01 1.41502905e+00 -2.34826401e-01 2.85849690e-01
-4.93718684e-01 9.43539083e-01 3.17715049e-01 4.13241982e-01
-9.90194321e-01 -2.35457420e-01 1.07131743e+00 -2.67274547e-02
3.89912754e-01 -3.75001907e-01 -2.74544191e-02 -1.84861994e+00
2.29532957e-01 -8.60058606e-01 3.03402126e-01 -2.22566113e-01
-1.64761543e+00 4.59054112e-01 -3.28747869e-01 -1.68614483e+00
2.07417235e-01 -6.80188775e-01 -6.52825058e-01 7.36283600e-01
-1.54760063e+00 -1.19176602e+00 -1.44203126e-01 7.53748953e-01
6.11443579e-01 2.74069197e-02 6.22887492e-01 6.06634200e-01
-9.48153436e-01 9.68701005e-01 4.44396675e-01 -3.22198093e-01
8.27624977e-01 -1.27894354e+00 8.46921876e-02 6.49367571e-01
5.31701446e-01 1.43819761e+00 7.31541336e-01 -5.54875374e-01
-1.73516726e+00 -1.17469561e+00 6.73078001e-01 -4.95594770e-01
1.18466747e+00 -6.58636093e-01 -7.64489412e-01 5.22446156e-01
-3.36596251e-01 4.30414341e-02 9.08725739e-01 7.79359102e-01
-6.38167441e-01 -3.17477971e-01 -5.85408807e-01 9.70435679e-01
1.16395259e+00 -5.30194640e-01 -2.49730274e-01 4.35541928e-01
9.85370636e-01 -2.99526619e-05 -1.02178741e+00 3.83871436e-01
8.74426007e-01 -9.66305852e-01 9.27561581e-01 -1.02647996e+00
2.19320476e-01 -2.64142513e-01 -2.69316643e-01 -1.45723021e+00
-5.19321620e-01 -8.48238230e-01 -4.03983921e-01 1.12945330e+00
5.11288226e-01 -7.84545243e-01 6.71524167e-01 6.85603023e-01
-1.25745565e-01 -1.00252521e+00 -6.87073231e-01 -8.66616488e-01
-7.43114725e-02 -5.00042200e-01 6.50653124e-01 8.35781991e-01
1.56397149e-02 3.83707672e-01 -7.31712878e-01 2.63882339e-01
4.03418481e-01 5.71113706e-01 7.85251081e-01 -1.42113030e+00
-6.99656367e-01 -3.78931224e-01 -2.40896478e-01 -1.37651300e+00
6.74531534e-02 -8.94288480e-01 -3.48739684e-01 -8.13677728e-01
2.01749131e-01 -6.18235230e-01 -8.25730145e-01 1.77694291e-01
-2.97250599e-01 -2.01355487e-01 1.33714974e-01 2.82506019e-01
-8.60547602e-01 9.21282530e-01 1.17659223e+00 -1.52448416e-01
-3.39961410e-01 6.88663960e-01 -9.15511906e-01 4.67083573e-01
9.05675888e-01 -3.09056967e-01 -5.54336548e-01 -1.83904022e-01
6.40794277e-01 -1.81345835e-01 1.49147570e-01 -6.86374843e-01
1.99855193e-01 -3.02003980e-01 3.15466732e-01 -5.74772596e-01
3.17520082e-01 -9.24877286e-01 1.18261606e-01 3.61777246e-01
-5.67060530e-01 7.63869435e-02 -3.28880101e-01 1.13614941e+00
-1.71517670e-01 8.34756568e-02 3.09556693e-01 1.58653155e-01
-4.32604849e-01 7.85208106e-01 7.35525414e-02 -1.57337144e-01
7.55675852e-01 2.80586243e-01 -3.86547655e-01 -3.86037886e-01
-8.95804048e-01 3.64945531e-01 2.12636515e-01 8.78042519e-01
5.02735496e-01 -1.33835411e+00 -3.41757596e-01 3.82895172e-01
3.14325958e-01 -3.57090026e-01 4.98297244e-01 9.24794018e-01
2.31873989e-01 5.71638525e-01 5.49987145e-02 -2.52582967e-01
-8.43995035e-01 6.83481514e-01 7.74626210e-02 -5.82111835e-01
-2.77061492e-01 7.21338451e-01 2.84667373e-01 -3.63437772e-01
2.06079423e-01 -3.17476302e-01 -2.99917728e-01 2.15143293e-01
5.13916135e-01 2.51945972e-01 2.94405073e-01 -2.30415538e-01
-1.14347421e-01 3.07957411e-01 -5.13272047e-01 1.90948457e-01
1.50062811e+00 -2.42205873e-01 1.06074572e-01 6.34682357e-01
1.43796337e+00 2.15088055e-01 -1.12295520e+00 -3.69112641e-01
8.97523109e-03 -6.55958176e-01 -1.43068535e-02 -5.88278055e-01
-9.82044220e-01 7.51626074e-01 6.47646964e-01 5.59418142e-01
8.79487038e-01 -1.71123266e-01 9.07879174e-01 3.94615471e-01
5.19777060e-01 -8.38421404e-01 2.69254446e-01 5.31200111e-01
7.68329024e-01 -1.25878751e+00 7.23099243e-03 -1.03681214e-01
-4.51187223e-01 1.04242122e+00 6.82875156e-01 -9.36779454e-02
1.15174854e+00 -3.43388468e-01 -3.51123095e-01 -6.36649430e-02
-8.78486693e-01 -2.13643745e-01 6.39525414e-01 1.18772030e-01
3.22874963e-01 2.89417833e-01 -5.78508914e-01 1.26546001e+00
-2.28614345e-01 -3.45263034e-01 1.36283502e-01 6.03786051e-01
-1.47756919e-01 -1.41950464e+00 2.43453294e-01 6.92871869e-01
-3.47591072e-01 -1.14938118e-01 -1.59420937e-01 4.50498730e-01
1.09633893e-01 8.59316289e-01 -3.76878828e-01 -1.02396393e+00
4.20492321e-01 -4.18620072e-02 1.22680981e-02 -6.11577749e-01
-5.85343778e-01 -1.31668985e-01 -2.52894640e-01 -7.52290308e-01
4.25979756e-02 -6.94090307e-01 -6.37131333e-01 -8.67396891e-02
-5.92099309e-01 3.82372946e-01 6.14611387e-01 7.00277805e-01
5.34743190e-01 5.51945090e-01 1.31159258e+00 -7.31212437e-01
-1.21314180e+00 -6.76530898e-01 -8.14804077e-01 5.71879268e-01
4.57350820e-01 -8.97458613e-01 -7.71400094e-01 -4.50568140e-01] | [10.103321075439453, 5.448392391204834] |
e0ffa32f-2ecd-4cac-8d20-f545672e963c | high-speed-and-high-quality-text-to-lip | 2107.06831 | null | https://arxiv.org/abs/2107.06831v2 | https://arxiv.org/pdf/2107.06831v2.pdf | Parallel and High-Fidelity Text-to-Lip Generation | As a key component of talking face generation, lip movements generation determines the naturalness and coherence of the generated talking face video. Prior literature mainly focuses on speech-to-lip generation while there is a paucity in text-to-lip (T2L) generation. T2L is a challenging task and existing end-to-end works depend on the attention mechanism and autoregressive (AR) decoding manner. However, the AR decoding manner generates current lip frame conditioned on frames generated previously, which inherently hinders the inference speed, and also has a detrimental effect on the quality of generated lip frames due to error propagation. This encourages the research of parallel T2L generation. In this work, we propose a parallel decoding model for fast and high-fidelity text-to-lip generation (ParaLip). Specifically, we predict the duration of the encoded linguistic features and model the target lip frames conditioned on the encoded linguistic features with their duration in a non-autoregressive manner. Furthermore, we incorporate the structural similarity index loss and adversarial learning to improve perceptual quality of generated lip frames and alleviate the blurry prediction problem. Extensive experiments conducted on GRID and TCD-TIMIT datasets demonstrate the superiority of proposed methods. Video samples are available via \url{https://paralip.github.io/}. | ['Nicholas Yuan', 'Baoxing Huai', 'Wencan Huang', 'Zhou Zhao', 'Yi Ren', 'Zhiying Zhu', 'Jinglin Liu'] | 2021-07-14 | null | null | null | null | ['text-to-face-generation', 'talking-face-generation'] | ['computer-vision', 'computer-vision'] | [ 1.18016742e-01 3.85368466e-02 -2.84782082e-01 -2.73819000e-01
-8.95270944e-01 -2.09341228e-01 5.20921409e-01 -6.75545812e-01
2.05831751e-01 7.19889224e-01 6.66425467e-01 1.40065653e-02
3.70239735e-01 -3.90312523e-01 -6.44578457e-01 -9.16199625e-01
3.37722957e-01 -2.41466150e-01 -1.35461137e-01 1.57504976e-01
1.75948679e-01 2.19284192e-01 -1.52821147e+00 4.30256754e-01
8.12235296e-01 1.22515261e+00 3.01572800e-01 8.51526082e-01
-1.54657170e-01 8.93715620e-01 -4.87805516e-01 -6.12722278e-01
1.58802956e-01 -6.69565618e-01 -2.06166014e-01 9.33214277e-02
2.58371472e-01 -4.86405313e-01 -6.14176035e-01 1.00260186e+00
1.00237679e+00 5.88555969e-02 6.61957562e-01 -1.65238667e+00
-7.22999394e-01 3.69515419e-01 -5.80742598e-01 8.58383253e-03
4.76473093e-01 6.11013353e-01 6.00078583e-01 -1.03874648e+00
3.79299104e-01 1.65410745e+00 5.98476231e-01 8.31799626e-01
-8.57859313e-01 -1.09073210e+00 -1.53660223e-01 4.07402545e-01
-1.43566310e+00 -1.55002105e+00 1.08431387e+00 -2.78421134e-01
3.77917647e-01 5.66784013e-03 2.65698910e-01 1.33015704e+00
2.37092644e-01 8.88802886e-01 9.23305571e-01 -3.06293070e-01
-1.40676990e-01 1.70605723e-02 -6.51123583e-01 4.64516848e-01
-5.07880270e-01 2.51281142e-01 -8.34653258e-01 2.21207529e-01
5.46161771e-01 -3.63301814e-01 -5.01735568e-01 2.33327836e-01
-9.49122012e-01 6.35464430e-01 -3.13058794e-02 -5.03339916e-02
-3.37262392e-01 1.96052521e-01 5.16950309e-01 -5.39571680e-02
5.86752415e-01 -3.04083914e-01 -6.95826933e-02 -3.69138628e-01
-9.99840200e-01 7.84668177e-02 4.65774238e-01 1.04941082e+00
4.23798233e-01 3.90281886e-01 -5.69993913e-01 1.09456098e+00
6.71663642e-01 8.04304242e-01 4.96370792e-01 -1.09333038e+00
6.45059347e-01 7.27167027e-03 -3.82063538e-02 -1.03013170e+00
3.42531838e-02 6.50083274e-02 -9.14841294e-01 1.11816064e-01
1.62764087e-01 -5.19421518e-01 -6.69599175e-01 1.96185255e+00
3.27888161e-01 5.79088092e-01 1.59760460e-01 8.96906078e-01
8.63391936e-01 9.92357254e-01 1.06876098e-01 -7.97421396e-01
9.77152288e-01 -9.02172744e-01 -1.33135712e+00 7.97029436e-02
1.78647876e-01 -1.27041674e+00 9.95804310e-01 1.16276734e-01
-1.41677809e+00 -9.05635476e-01 -5.46030343e-01 -1.87915668e-01
1.66451767e-01 4.06415761e-01 2.24406168e-01 6.02567494e-01
-1.09532690e+00 1.02076009e-01 -3.85896295e-01 7.96641260e-02
6.22156501e-01 5.11021130e-02 -1.46596983e-01 -1.03767663e-01
-1.23971677e+00 3.32866639e-01 3.05015240e-02 2.40697429e-01
-8.74036312e-01 -7.09237754e-01 -8.91267538e-01 -1.58384368e-01
1.86371133e-01 -4.97355133e-01 1.26155794e+00 -1.00108230e+00
-1.95327687e+00 4.63191628e-01 -6.56138182e-01 -3.23307723e-01
7.36696422e-01 -1.14839273e-02 -4.99662608e-01 2.99655765e-01
8.26942176e-02 1.01049757e+00 1.33808982e+00 -1.26646912e+00
-5.54031730e-01 -1.16161741e-01 -2.70167172e-01 4.79404151e-01
-1.42861962e-01 -3.16625014e-02 -6.14383280e-01 -8.86553645e-01
-3.29275668e-01 -8.06931138e-01 4.30860221e-01 1.82959169e-01
-3.62207085e-01 -4.10991639e-01 1.02209091e+00 -9.54152763e-01
1.25024235e+00 -2.35271621e+00 -1.20020941e-01 -5.44450760e-01
-9.05173719e-02 3.27121437e-01 -2.38960713e-01 2.89907336e-01
7.50136599e-02 2.27884516e-01 8.77838489e-03 -6.05319440e-01
4.90780175e-02 -1.34518340e-01 -5.23905337e-01 3.65099192e-01
2.46062949e-01 8.59849274e-01 -7.34000444e-01 -9.01631653e-01
3.85114372e-01 9.86334205e-01 -5.86587846e-01 5.63003600e-01
-1.65981695e-01 5.05361974e-01 -2.30320007e-01 6.87435210e-01
8.73057842e-01 3.13761622e-01 -2.86979854e-01 -6.01111650e-01
-1.96723398e-02 1.57542348e-01 -6.89802408e-01 1.58472157e+00
-7.91639566e-01 8.21007729e-01 1.95588395e-01 -6.03929460e-01
9.33539748e-01 7.16785550e-01 4.65588689e-01 -5.98569155e-01
1.64243519e-01 -4.18327488e-02 -7.56609067e-02 -9.14936423e-01
2.81866580e-01 -2.13163257e-01 3.33698809e-01 4.79704365e-02
-8.88216123e-02 -2.44206175e-01 5.93132451e-02 -1.35346904e-01
4.08148646e-01 2.90050715e-01 -9.33836550e-02 2.02318579e-01
7.21554577e-01 -7.84420311e-01 6.00180149e-01 9.40204039e-02
-3.81233156e-01 7.08239555e-01 4.94606405e-01 2.22843856e-01
-9.93657649e-01 -1.07497787e+00 -1.55209392e-01 1.00084591e+00
9.90076438e-02 -2.68974274e-01 -9.18909848e-01 -3.11550617e-01
-3.59501302e-01 8.71889710e-01 -3.79969746e-01 -2.03714624e-01
-3.93685848e-01 -2.71330267e-01 8.08180869e-01 1.94304451e-01
6.34692788e-01 -1.42400336e+00 -1.25238180e-01 2.21776605e-01
-7.84104049e-01 -1.29902732e+00 -1.10665393e+00 -8.05544198e-01
-3.26873094e-01 -7.17113495e-01 -9.79328096e-01 -7.78981984e-01
4.44237828e-01 1.26407161e-01 7.17737257e-01 -2.35163242e-01
-1.22560710e-01 1.38290480e-01 -1.48962677e-01 -3.76984000e-01
-7.13759005e-01 -3.54052782e-01 1.30568266e-01 5.40222049e-01
3.29430439e-02 -5.88055551e-01 -8.00592124e-01 2.97761112e-01
-7.43331730e-01 4.46395606e-01 5.00504017e-01 8.42091680e-01
4.55096543e-01 1.72573566e-01 8.94468009e-01 -1.76600948e-01
8.22256982e-01 -5.56312323e-01 -3.13869655e-01 2.78664324e-02
-3.10210705e-01 -2.95838922e-01 7.89513469e-01 -7.22009897e-01
-1.39673615e+00 -2.03361064e-01 -3.95285219e-01 -1.03178728e+00
-1.14938349e-01 6.65050326e-03 -5.93777895e-01 3.02935958e-01
3.68004385e-03 6.26773834e-01 2.36247107e-01 -2.30331540e-01
3.19381267e-01 1.08344364e+00 5.41115582e-01 -3.67132902e-01
4.18124288e-01 2.53815204e-01 -1.35010868e-01 -1.07221568e+00
-3.96176249e-01 -8.97330418e-02 -7.07616284e-02 -4.71967310e-01
8.63827646e-01 -9.95946229e-01 -1.09257305e+00 7.61425674e-01
-1.32200944e+00 -4.03523475e-01 -4.20207158e-02 3.88386428e-01
-9.16707933e-01 4.31422174e-01 -5.15152335e-01 -1.23052204e+00
-4.66163367e-01 -1.41314662e+00 1.25289118e+00 3.13758999e-01
9.13359746e-02 -6.53927565e-01 -2.24854127e-01 6.96014166e-01
3.44768822e-01 -2.48936731e-02 4.52276617e-01 1.27507836e-01
-4.80180651e-01 6.45767013e-03 -2.78691262e-01 5.03094912e-01
3.82241815e-01 1.20201334e-01 -1.14923656e+00 -7.69906044e-02
-5.92263043e-02 -3.86227846e-01 5.48439682e-01 7.10768223e-01
1.24679458e+00 -8.93511236e-01 -1.63896218e-01 6.18493438e-01
9.48190629e-01 4.21024084e-01 8.14896882e-01 -5.99970877e-01
5.78967273e-01 7.08080173e-01 5.99248052e-01 5.75395584e-01
4.74897712e-01 8.21348250e-01 1.45169377e-01 2.27349326e-02
-7.71538198e-01 -6.73906446e-01 7.39981174e-01 8.91214550e-01
1.59622014e-01 -6.77453637e-01 -6.08622789e-01 5.93373537e-01
-1.51441443e+00 -1.21357214e+00 1.93994537e-01 1.98118854e+00
1.13342774e+00 -1.08692817e-01 -3.15424092e-02 2.33493075e-01
9.87509966e-01 3.10379684e-01 -6.35471940e-01 -2.40237564e-01
-1.74328893e-01 -7.33775422e-02 6.59596696e-02 8.02295029e-01
-7.59273529e-01 1.03242302e+00 5.21641302e+00 1.37951541e+00
-1.30165613e+00 2.21057892e-01 1.03504407e+00 -1.54038042e-01
-1.73861980e-01 -3.15727353e-01 -9.35273349e-01 1.05671883e+00
9.63723838e-01 -4.30618614e-01 4.83243018e-01 4.96291548e-01
9.87680495e-01 -1.89786460e-02 -7.32144475e-01 1.34008086e+00
2.46821702e-01 -1.10560012e+00 1.98263854e-01 -4.33872081e-02
5.21630526e-01 -3.96206766e-01 3.65991086e-01 1.62871629e-01
-2.44785339e-01 -1.10321379e+00 7.72839665e-01 7.33793557e-01
1.36206305e+00 -7.57732034e-01 3.87009740e-01 1.54580891e-01
-1.40412891e+00 -3.12118605e-02 -2.41097435e-01 3.47110957e-01
4.28547084e-01 3.03865552e-01 -1.01244450e+00 2.19579279e-01
4.23773080e-01 7.02329516e-01 -1.17537431e-01 7.23717034e-01
-2.29721829e-01 7.69135237e-01 -1.18865386e-01 1.71885088e-01
-1.22285828e-01 7.90708698e-03 7.85591245e-01 1.13517594e+00
4.36400533e-01 8.06034356e-02 -7.21839070e-02 7.81915188e-01
-3.57422411e-01 2.17031464e-01 -7.34837711e-01 -7.49267116e-02
6.57775760e-01 9.35331106e-01 -1.22769110e-01 -7.83534572e-02
-1.61149547e-01 8.58193934e-01 -3.05591285e-01 4.93489355e-01
-1.00277281e+00 -3.02496344e-01 8.12041104e-01 2.95770615e-01
8.62464234e-02 8.36119503e-02 3.83295678e-02 -8.87165964e-01
8.07597861e-02 -9.49043334e-01 -1.94583580e-01 -1.11335433e+00
-1.08439398e+00 5.87429225e-01 -2.55740255e-01 -1.26410139e+00
-5.27387917e-01 3.44104134e-02 -5.39797306e-01 1.00608337e+00
-1.74399662e+00 -1.22831917e+00 -2.95031786e-01 9.03751671e-01
1.20300639e+00 -2.28108555e-01 4.28335309e-01 3.26585859e-01
-5.34129262e-01 9.73206818e-01 -6.49476945e-02 6.76609725e-02
8.69928598e-01 -4.81551439e-01 8.42559114e-02 7.24531531e-01
-4.24926579e-01 1.18348591e-01 6.70607388e-01 -5.31584859e-01
-1.29771972e+00 -1.27440953e+00 8.60745668e-01 3.57249603e-02
3.00239295e-01 -1.83416024e-01 -6.56813085e-01 2.33449757e-01
3.96476060e-01 -6.82100803e-02 4.67343241e-01 -7.08656907e-01
3.24258721e-03 -3.08777750e-01 -1.15265775e+00 8.23332369e-01
9.11243498e-01 -6.22990727e-01 -1.41085744e-01 1.69406295e-01
7.69544005e-01 -2.30135098e-01 -7.27600098e-01 5.00415862e-01
6.50066674e-01 -1.06497169e+00 8.59045088e-01 1.20300362e-02
5.61487913e-01 -2.20925644e-01 -1.41987815e-01 -9.93584752e-01
2.21113294e-01 -1.19333816e+00 -2.31875673e-01 1.80424416e+00
1.01906575e-01 -4.88213390e-01 5.10669529e-01 3.88552874e-01
-7.09300712e-02 -8.40355515e-01 -9.97082531e-01 -4.65237826e-01
-1.06037445e-02 -4.55676049e-01 5.14942527e-01 4.86685634e-01
-1.85074866e-01 3.71628314e-01 -8.41635585e-01 8.19200277e-02
6.81776285e-01 -1.36143222e-01 9.39943552e-01 -5.75697482e-01
3.97534110e-02 -4.56077456e-01 -1.35731772e-01 -1.30051947e+00
5.07935345e-01 -6.29836440e-01 3.51038784e-01 -1.12986994e+00
-1.04480900e-01 -4.14998174e-01 1.41757622e-01 2.40622029e-01
-2.36975223e-01 2.51176298e-01 3.97439271e-01 2.81580269e-01
-2.95171052e-01 1.03930068e+00 1.51598966e+00 -1.00198144e-03
-2.17698261e-01 1.24524206e-01 -3.80280077e-01 5.50147772e-01
8.73673975e-01 -1.71598569e-01 -6.74424469e-01 -2.80316442e-01
-4.01252359e-01 7.63507724e-01 1.96504205e-01 -9.82219219e-01
1.88769460e-01 -1.68595225e-01 2.68196881e-01 -6.02352500e-01
9.21243846e-01 -5.21640360e-01 4.60354500e-02 8.92659575e-02
-4.42867845e-01 -1.81192577e-01 2.49316379e-01 4.98481929e-01
-2.57959902e-01 8.47950727e-02 1.05644977e+00 2.47085348e-01
-2.58718997e-01 7.00537920e-01 -1.90588772e-01 2.68654257e-01
8.98363173e-01 -2.93249905e-01 1.40668303e-02 -8.90265584e-01
-4.13910300e-01 1.01232119e-02 2.33470336e-01 6.66757107e-01
8.56781244e-01 -1.44216561e+00 -1.00049031e+00 3.75370830e-01
-2.91180938e-01 -1.63650572e-01 5.17638266e-01 7.40828991e-01
-7.71198943e-02 3.77662450e-01 -8.12984109e-02 -5.66896141e-01
-1.29127717e+00 4.93738800e-01 1.97162494e-01 1.21389732e-01
-4.07443523e-01 7.10530460e-01 5.48533797e-01 1.66079625e-01
4.04278457e-01 1.88592449e-01 -1.94367081e-01 7.25459903e-02
6.01178765e-01 4.06518340e-01 -4.23960805e-01 -1.10200560e+00
-5.34200408e-02 6.05053544e-01 2.00479284e-01 -1.77013472e-01
8.33463669e-01 -6.33561552e-01 2.49867052e-01 4.24752235e-01
1.32377720e+00 3.03577691e-01 -1.58551323e+00 7.16744065e-02
-7.38999844e-01 -5.81073642e-01 1.59466133e-01 -5.23164451e-01
-1.22079265e+00 1.18384933e+00 6.38263702e-01 -1.53777644e-01
1.40532732e+00 -2.34903142e-01 1.04272318e+00 -3.48811835e-01
-6.92646876e-02 -7.72648156e-01 2.75593609e-01 1.59687012e-01
1.12220597e+00 -1.16091812e+00 -5.21710515e-01 -4.66149598e-01
-6.97409272e-01 9.89158809e-01 4.38103527e-01 3.46681356e-01
6.53596044e-01 3.84697825e-01 2.21450686e-01 5.55888176e-01
-9.62808788e-01 -7.50812516e-02 5.37229665e-02 7.44356692e-01
5.00184119e-01 -1.14408880e-01 -1.37398958e-01 4.08708364e-01
-3.48556668e-01 4.01254326e-01 2.45194107e-01 2.87965178e-01
2.93527008e-03 -8.32072377e-01 -4.52182740e-01 1.15046479e-01
-6.76865876e-01 -4.25236881e-01 1.43880889e-01 3.12405854e-01
2.51436740e-01 1.42230201e+00 -9.43087507e-03 -2.01593772e-01
-3.76833752e-02 1.68886065e-01 2.47967497e-01 -6.14841236e-03
3.20760161e-02 4.16321367e-01 -9.35306624e-02 -3.67847711e-01
-2.63556778e-01 -6.04653358e-01 -1.01501060e+00 -5.14388382e-01
-2.67897397e-01 7.42576737e-03 6.18402302e-01 5.90012908e-01
6.06230795e-01 3.80071133e-01 1.05925536e+00 -1.05530345e+00
-3.79353374e-01 -1.24244201e+00 -2.70348161e-01 4.92624819e-01
5.22984564e-01 -6.19722843e-01 -4.59331214e-01 5.85367858e-01] | [13.277749061584473, -0.37426289916038513] |
d91afcf2-6de9-4d54-b067-1fe312706018 | exploring-non-verbal-predicates-in-semantic | 2307.0187 | null | https://arxiv.org/abs/2307.01870v1 | https://arxiv.org/pdf/2307.01870v1.pdf | Exploring Non-Verbal Predicates in Semantic Role Labeling: Challenges and Opportunities | Although we have witnessed impressive progress in Semantic Role Labeling (SRL), most of the research in the area is carried out assuming that the majority of predicates are verbs. Conversely, predicates can also be expressed using other parts of speech, e.g., nouns and adjectives. However, non-verbal predicates appear in the benchmarks we commonly use to measure progress in SRL less frequently than in some real-world settings -- newspaper headlines, dialogues, and tweets, among others. In this paper, we put forward a new PropBank dataset which boasts wide coverage of multiple predicate types. Thanks to it, we demonstrate empirically that standard benchmarks do not provide an accurate picture of the current situation in SRL and that state-of-the-art systems are still incapable of transferring knowledge across different predicate types. Having observed these issues, we also present a novel, manually-annotated challenge set designed to give equal importance to verbal, nominal, and adjectival predicate-argument structures. We use such dataset to investigate whether we can leverage different linguistic resources to promote knowledge transfer. In conclusion, we claim that SRL is far from "solved", and its integration with other semantic tasks might enable significant improvements in the future, especially for the long tail of non-verbal predicates, thereby facilitating further research on SRL for non-verbal predicates. | ['Roberto Navigli', 'Simone Conia', 'Riccardo Orlando'] | 2023-07-04 | null | null | null | null | ['transfer-learning', 'semantic-role-labeling'] | ['miscellaneous', 'natural-language-processing'] | [ 3.02640289e-01 5.95702589e-01 -6.89590335e-01 -4.37793553e-01
-5.89133382e-01 -1.06957400e+00 9.18313444e-01 7.60001421e-01
-6.49750471e-01 1.24448466e+00 8.22510302e-01 -3.02655637e-01
-1.09435245e-01 -8.83726120e-01 -7.24259257e-01 -2.86408186e-01
1.20099805e-01 8.85918319e-01 6.01089895e-01 -9.41257417e-01
-1.77924410e-02 1.96997374e-01 -1.54927409e+00 4.76847470e-01
4.61093277e-01 6.63912356e-01 -1.05726756e-01 -1.61755458e-01
-3.48172039e-01 1.21309078e+00 -6.80099726e-01 -5.16723216e-01
-6.11380562e-02 -1.18248917e-01 -1.42025685e+00 -4.28094089e-01
1.34465203e-01 3.40591937e-01 2.61320919e-02 8.62327754e-01
4.19029683e-01 1.78458959e-01 4.67801869e-01 -1.20167387e+00
-5.01696587e-01 1.23539388e+00 -6.99503273e-02 2.82419503e-01
8.74790668e-01 -1.28729209e-01 1.78451478e+00 -3.49864662e-01
1.04935372e+00 1.51821470e+00 7.10671425e-01 6.06298447e-01
-1.18514681e+00 -5.55874407e-01 1.62620112e-01 2.48186886e-01
-9.07650709e-01 -3.40469807e-01 7.16326296e-01 -2.96632946e-01
1.04041946e+00 2.00555667e-01 4.20696616e-01 1.19464004e+00
-5.09671926e-01 7.66765356e-01 1.21497047e+00 -5.64734697e-01
-1.82192460e-01 1.96164593e-01 3.25787902e-01 3.80569547e-01
3.78485978e-01 -1.52975291e-01 -6.82713985e-01 -5.79486266e-02
3.38380009e-01 -7.24091351e-01 -3.69871795e-01 -1.31950751e-01
-1.33410192e+00 1.00558126e+00 1.77005351e-01 8.02749515e-01
-2.07775548e-01 7.13956170e-03 8.70414674e-01 3.79891962e-01
6.10790610e-01 7.53938496e-01 -9.18421149e-01 -4.22725230e-01
-5.10213315e-01 5.31607985e-01 1.14968383e+00 8.62354100e-01
5.14267087e-01 -4.15706813e-01 2.35762633e-02 9.93373394e-01
7.58469850e-02 4.74824533e-02 4.72869664e-01 -9.43493724e-01
6.62651658e-01 7.31747508e-01 2.86767036e-01 -9.38880742e-01
-6.85823083e-01 5.95955690e-03 -9.36230570e-02 -3.98158669e-01
7.74286151e-01 -2.65816925e-03 -3.20790410e-01 1.93373072e+00
3.58141690e-01 -9.44112763e-02 3.53199840e-01 6.90332234e-01
1.15346372e+00 5.57662129e-01 5.68688571e-01 -8.62251371e-02
1.69590807e+00 -6.46508396e-01 -5.52252233e-01 -4.09428120e-01
8.90563965e-01 -5.20163000e-01 1.46167731e+00 2.92928033e-02
-9.34308589e-01 -3.00976425e-01 -9.39071774e-01 -2.44251058e-01
-6.55920982e-01 -4.93401289e-01 1.00312042e+00 4.56976026e-01
-7.56534994e-01 5.78871846e-01 -4.93973821e-01 -5.05801439e-01
3.55699390e-01 8.43628123e-02 -5.96707344e-01 -6.90663084e-02
-1.75658464e+00 1.16504586e+00 1.02955556e+00 -5.39136350e-01
-4.28170949e-01 -8.81913543e-01 -1.04031432e+00 -7.08255395e-02
6.89137161e-01 -4.20937002e-01 1.39515960e+00 -9.12796497e-01
-1.14548361e+00 1.43886340e+00 -1.40547669e-02 -6.82831883e-01
2.70279229e-01 -1.87335998e-01 -4.43039626e-01 2.42279768e-02
4.92051721e-01 6.79168165e-01 2.89031476e-01 -1.20417416e+00
-9.19063807e-01 -2.99732715e-01 6.89553678e-01 4.32156086e-01
-1.68468162e-01 6.88592136e-01 1.61195561e-01 -5.73528290e-01
-1.54508486e-01 -7.42953539e-01 1.79896399e-01 -4.55704123e-01
-1.40739158e-01 -8.94092262e-01 5.63777566e-01 -4.10427570e-01
1.15270293e+00 -2.14783335e+00 -9.28456336e-03 -1.87073886e-01
-3.93087082e-02 2.89398551e-01 1.20283335e-01 7.24003613e-01
-3.12794149e-01 2.96860188e-01 -2.67682135e-01 -2.03562919e-02
2.08595142e-01 8.07424426e-01 -5.17860830e-01 3.40496421e-01
1.84457317e-01 1.02839506e+00 -1.25437438e+00 -5.16943634e-01
-2.17952067e-03 2.10607633e-01 -4.23962325e-01 -2.19460994e-01
-7.57187903e-01 5.91345370e-01 -3.81662995e-01 4.56240207e-01
5.70315160e-02 -2.55015612e-01 3.97957832e-01 -3.34814414e-02
-2.49422848e-01 1.22365713e+00 -8.99625957e-01 1.50289285e+00
-3.93718898e-01 5.32234132e-01 -1.55769691e-01 -1.32330358e+00
5.47551394e-01 5.15933454e-01 5.21554112e-01 -6.11295521e-01
1.06450126e-01 4.05677348e-01 2.47417450e-01 -2.78278142e-01
5.15061855e-01 -5.07300198e-01 -3.89833242e-01 1.92967370e-01
-2.93384045e-02 -3.83702725e-01 5.99921048e-01 -1.45686522e-01
1.10212767e+00 1.72009483e-01 7.49206066e-01 -3.89633626e-01
7.61199713e-01 5.51532984e-01 4.36497808e-01 4.47068423e-01
-1.56901464e-01 1.71009853e-01 6.86087549e-01 -1.85213178e-01
-7.94642687e-01 -8.42023551e-01 -3.99336278e-01 1.54686618e+00
1.75148129e-01 -5.57820141e-01 -4.26710546e-01 -9.93787587e-01
6.97227940e-02 1.01280963e+00 -4.24090743e-01 2.98427343e-01
-9.21408772e-01 -6.61194444e-01 8.88077140e-01 4.75201845e-01
3.76854718e-01 -1.50977683e+00 -6.07298672e-01 4.90800500e-01
-2.60840714e-01 -1.46369159e+00 2.06516013e-01 1.85772672e-01
-4.86329705e-01 -1.27619445e+00 -1.43309698e-01 -9.05372381e-01
8.93822238e-02 -1.45497143e-01 1.65219426e+00 -8.21769238e-02
3.28896314e-01 3.08981478e-01 -8.01579237e-01 -7.55241215e-01
-5.54821014e-01 6.63299933e-02 -2.79190123e-01 -4.87003148e-01
5.41492403e-01 -5.21743417e-01 -1.47248194e-01 2.68445145e-02
-7.39560068e-01 -2.12394342e-01 4.85553555e-02 6.58661246e-01
3.59072775e-01 2.93538362e-01 6.42054379e-01 -1.59459460e+00
6.12631083e-01 -6.71562076e-01 -2.76256353e-01 3.14327283e-03
-2.40290448e-01 4.47774082e-02 6.86505318e-01 -2.52161682e-01
-1.13254774e+00 -4.46610749e-01 -4.85181093e-01 5.56193590e-01
-3.84878188e-01 8.56583059e-01 -2.59983271e-01 3.14325988e-01
7.89994955e-01 -1.32412225e-01 -1.43755004e-01 -4.30042475e-01
5.52751839e-01 4.38008875e-01 4.13846076e-01 -1.07092261e+00
7.31013060e-01 5.33000469e-01 -4.80561554e-02 -9.14926946e-01
-1.57276058e+00 -8.00038815e-01 -2.64259934e-01 2.99239814e-01
8.28444541e-01 -1.02245617e+00 -6.29704595e-01 9.18869153e-02
-1.19280469e+00 -5.09814501e-01 -7.38729835e-01 8.97840559e-02
-6.69689715e-01 2.79627889e-01 -6.96099937e-01 -3.12577993e-01
-2.58703567e-02 -7.12387264e-01 1.02588642e+00 -1.22607008e-01
-6.70414448e-01 -1.35338151e+00 -8.73971879e-02 5.09288073e-01
3.34351957e-01 2.46783778e-01 1.40456653e+00 -1.27636743e+00
1.53779332e-02 1.71174541e-01 -3.08868468e-01 3.40206116e-01
1.72295883e-01 -4.75402504e-01 -7.56762385e-01 -1.42195392e-02
-7.65699372e-02 -5.53525746e-01 6.09457910e-01 -5.75273857e-02
7.97913909e-01 -3.93295974e-01 -2.63641149e-01 -1.30685687e-01
1.23385024e+00 1.85985014e-01 6.06861472e-01 7.28068173e-01
5.23988008e-01 9.99211371e-01 1.02978754e+00 1.94896758e-01
8.65497589e-01 5.96484721e-01 2.54903376e-01 6.78265691e-02
-2.47289285e-01 -3.72585922e-01 2.23903224e-01 4.64691877e-01
-8.58980641e-02 -2.67407626e-01 -1.07081246e+00 7.39054382e-01
-1.75005996e+00 -9.88664031e-01 -2.14160010e-01 1.84147942e+00
1.45007038e+00 4.67709213e-01 3.18397522e-01 3.94629151e-01
4.80649889e-01 4.72465247e-01 5.15986569e-02 -1.35569513e-01
-4.13531095e-01 6.09038949e-01 4.70557630e-01 4.27525610e-01
-1.32277071e+00 1.43947434e+00 5.74289370e+00 8.10285509e-01
-9.60139215e-01 2.91878223e-01 1.16142742e-01 6.70598328e-01
-4.64049965e-01 2.36855298e-01 -1.09745932e+00 3.95404845e-01
9.38189387e-01 -1.31503403e-01 2.35918581e-01 7.02335238e-01
-7.15074167e-02 -6.08979650e-02 -1.24594438e+00 6.94003105e-01
-4.73329015e-02 -1.26676631e+00 -6.24413528e-02 -3.20374757e-01
3.43644381e-01 1.68742597e-01 -3.58494818e-01 6.56940520e-01
5.66210806e-01 -1.07352078e+00 8.83291245e-01 -1.92422360e-01
5.87643325e-01 -3.33110094e-01 8.01934183e-01 3.67730916e-01
-9.78698730e-01 -6.34663254e-02 -2.50510782e-01 -3.69638324e-01
2.32276201e-01 3.38934362e-01 -9.01640534e-01 6.00383520e-01
4.09634233e-01 9.39747751e-01 -3.64074022e-01 5.98252058e-01
-7.68451393e-01 8.74108851e-01 -2.64002681e-01 -1.03081532e-01
3.14437479e-01 1.44229054e-01 5.69196701e-01 1.33959556e+00
-1.73326626e-01 1.46976605e-01 3.94017816e-01 3.82667601e-01
-1.75675586e-01 5.15574038e-01 -6.69656873e-01 -2.87588328e-01
5.69975853e-01 7.75904834e-01 -7.15538740e-01 -2.76304096e-01
-7.74959445e-01 4.05834973e-01 3.94002944e-01 3.94382421e-03
-7.76760876e-01 1.06836278e-02 6.20036066e-01 4.50233817e-01
2.71194540e-02 -1.69093892e-01 -1.49616495e-01 -9.37009096e-01
-3.73738743e-02 -9.50520933e-01 8.25806677e-01 -6.64608061e-01
-1.61061680e+00 2.34248266e-01 4.62919056e-01 -7.52313197e-01
-3.21706802e-01 -8.32456827e-01 5.87477088e-02 5.40108860e-01
-2.05730653e+00 -1.17575479e+00 1.19330280e-01 4.72776562e-01
5.92816651e-01 1.71250522e-01 9.08644319e-01 3.88162583e-01
5.59382141e-03 1.76832020e-01 -5.54887176e-01 2.76525319e-01
8.12877178e-01 -1.34430325e+00 3.09588641e-01 3.92385572e-01
3.24071527e-01 4.98739570e-01 9.84734833e-01 -5.17082393e-01
-8.57313514e-01 -7.99347818e-01 1.31829822e+00 -8.03809524e-01
1.21610069e+00 -1.30011797e-01 -9.99700069e-01 1.00917983e+00
1.33814186e-01 5.88406473e-02 5.97079396e-01 5.03386676e-01
-4.95908350e-01 3.46578747e-01 -1.01584673e+00 3.98833454e-01
1.39638472e+00 -5.32983243e-01 -1.35700190e+00 4.73884523e-01
1.23381245e+00 -5.41248441e-01 -9.13181305e-01 7.43929744e-01
8.76170248e-02 -5.75691640e-01 1.04973066e+00 -8.45315397e-01
4.58988845e-01 -2.80336380e-01 -3.77038151e-01 -1.26677561e+00
6.11357391e-02 -2.98117161e-01 2.61668980e-01 1.50621152e+00
7.45197117e-01 -7.56127119e-01 7.32286215e-01 4.24801946e-01
-3.88920188e-01 -4.03150111e-01 -9.68903482e-01 -7.85009682e-01
2.53910542e-01 -7.31068373e-01 4.80525076e-01 1.29284668e+00
1.03767492e-01 8.84317815e-01 2.18346193e-01 -1.55355081e-01
1.16693728e-01 1.15210004e-01 4.86923218e-01 -1.64498305e+00
-3.22754860e-01 -4.21025842e-01 -4.69974011e-01 -7.77790666e-01
8.36790919e-01 -1.26495099e+00 -7.09723905e-02 -1.66335297e+00
-9.65310484e-02 -9.17439938e-01 -2.43679807e-02 9.45219517e-01
-3.73045765e-02 1.39221981e-01 7.19362572e-02 -4.76930290e-02
-6.70528233e-01 2.64403045e-01 1.21726203e+00 4.43966500e-02
-7.85764083e-02 -1.42245740e-01 -1.05940759e+00 9.42327023e-01
7.46433556e-01 -5.04246771e-01 -3.96482944e-01 -2.55955309e-01
5.50934076e-01 -1.92210957e-01 1.57609165e-01 -6.86339319e-01
-2.19913796e-01 -3.35101664e-01 -3.21989298e-01 4.03310079e-03
2.77346283e-01 -8.32889438e-01 -1.60726279e-01 7.71208992e-03
-3.75137717e-01 -1.19771846e-01 3.05385768e-01 2.61299461e-01
-5.35838187e-01 -3.83392483e-01 4.76818562e-01 -3.95900369e-01
-1.10608673e+00 -1.42057508e-01 -9.34270993e-02 9.59203243e-01
8.24525654e-01 -4.65783179e-02 -6.94764793e-01 -1.71834692e-01
-6.65798485e-01 1.31123140e-01 3.99557322e-01 4.22672987e-01
9.67830941e-02 -8.64538372e-01 -8.18843365e-01 -2.37447500e-01
3.65452170e-01 9.80515778e-03 -2.07682714e-01 6.93906724e-01
-5.67324698e-01 6.16399527e-01 2.15219590e-03 -2.42045060e-01
-1.12793732e+00 4.68698800e-01 -2.46332809e-02 -5.93187153e-01
-5.77757716e-01 9.32593703e-01 -1.82560142e-02 -6.03312612e-01
1.32231921e-01 -1.89519539e-01 -6.98628843e-01 5.40344238e-01
1.87202632e-01 -7.57190585e-02 5.24626337e-02 -9.66788948e-01
-4.62554932e-01 1.75026566e-01 2.12525949e-01 -9.02548879e-02
1.54819202e+00 9.76998061e-02 -4.79640275e-01 6.81964338e-01
8.74114096e-01 4.02369440e-01 -5.62743902e-01 -3.47065866e-01
6.55348003e-01 -2.00374693e-01 -5.39895892e-01 -8.23025405e-01
-3.81156325e-01 4.53929812e-01 -1.51433796e-01 6.04491055e-01
8.86880994e-01 5.74039578e-01 7.84868538e-01 2.77663469e-01
7.90006101e-01 -1.11248398e+00 -2.01487288e-01 8.97790015e-01
8.31899881e-01 -1.15489113e+00 -1.00315273e-01 -9.68766689e-01
-7.34328985e-01 7.06760049e-01 2.97143668e-01 1.35757858e-02
4.60330337e-01 1.13985971e-01 7.88280182e-03 -5.35542071e-01
-6.59728229e-01 -6.46276534e-01 4.87435050e-02 6.01764977e-01
9.55142021e-01 2.99745202e-01 -7.79202282e-01 4.52434719e-01
-8.16086173e-01 -2.91376889e-01 4.90759164e-01 9.49831545e-01
-3.19435567e-01 -1.53340924e+00 -4.64349724e-02 2.83803165e-01
-9.69519556e-01 -3.45519572e-01 -3.56031746e-01 1.31713867e+00
1.54514566e-01 9.76350605e-01 -6.11170344e-02 2.13797912e-01
6.89913869e-01 2.83119798e-01 6.57168269e-01 -1.34390104e+00
-7.31103897e-01 -5.18466532e-01 1.03754020e+00 -4.61174726e-01
-9.71418202e-01 -7.55445540e-01 -1.53497291e+00 -7.99528062e-02
1.55631965e-02 3.67361754e-01 4.68659610e-01 1.14238930e+00
-1.52221620e-01 3.42202157e-01 -1.83906749e-01 -4.37158316e-01
-3.78546387e-01 -7.64529884e-01 -4.19986784e-01 7.94985116e-01
4.01273556e-02 -1.00551009e+00 -3.44390988e-01 -1.98725000e-01] | [10.30292797088623, 9.364359855651855] |
af9c7816-3a9a-4262-88c9-5b6706390221 | a-deep-learning-approach-to-language | null | null | https://aclanthology.org/W19-3630 | https://aclanthology.org/W19-3630.pdf | A Deep Learning Approach to Language-independent Gender Prediction on Twitter | This work presents a set of experiments conducted to predict the gender of Twitter users based on language-independent features extracted from the text of the users{'} tweets. The experiments were performed on a version of TwiSty dataset including tweets written by the users of six different languages: Portuguese, French, Dutch, English, German, and Italian. Logistic regression (LR), and feed-forward neural networks (FFNN) with back-propagation were used to build models in two different settings: Inter-Lingual (IL) and Cross-Lingual (CL). In the IL setting, the training and testing were performed on the same language whereas in the CL, Italian and German datasets were set aside and only used as test sets and the rest were combined to compose training and development sets. In the IL, the highest accuracy score belongs to LR whereas, in the CL, FFNN with three hidden layers yields the highest score. The results show that neural network based models underperform traditional models when the size of the training set is small; however, they beat traditional models by a non-trivial margin, when they are fed with large enough data. Finally, the feature analysis confirms that men and women have different writing styles independent of their language. | ['REYHANEH HASHEMPOUR'] | 2019-08-01 | null | null | null | ws-2019-8 | ['gender-prediction'] | ['computer-vision'] | [-4.02512342e-01 -7.14087412e-02 -3.88096869e-01 -5.13178289e-01
-3.61639522e-02 -3.53140891e-01 9.50384319e-01 3.14593822e-01
-8.88124824e-01 8.46148789e-01 1.33045211e-01 -3.27069789e-01
1.54273480e-01 -8.01238477e-01 -3.66206735e-01 -4.64623928e-01
8.81975889e-02 6.36594236e-01 -1.80059355e-02 -3.70201528e-01
1.34319246e-01 3.95478904e-01 -1.45105052e+00 2.00164244e-01
6.56631947e-01 8.07452142e-01 -2.76251912e-01 5.41666746e-01
-1.19828686e-01 8.16940486e-01 -4.54138994e-01 -7.28363693e-01
2.05863714e-01 -1.60677955e-01 -6.22520804e-01 -2.72437960e-01
1.15366891e-01 -3.51295248e-02 -9.04697832e-03 6.61750317e-01
4.91934031e-01 4.38879710e-03 5.37422895e-01 -1.16770530e+00
-4.91557866e-01 8.95364285e-01 -6.21117115e-01 -1.15137264e-01
2.74956435e-01 -2.07878634e-01 7.27603257e-01 -1.08961630e+00
6.33925021e-01 1.33525085e+00 8.31808805e-01 4.97227252e-01
-1.14343154e+00 -8.33298266e-01 1.08620971e-01 -5.44029735e-02
-1.41148460e+00 -4.58910435e-01 5.51201224e-01 -6.00543320e-01
6.92065179e-01 2.02982396e-01 5.07511616e-01 1.30160654e+00
8.07492509e-02 4.27105010e-01 1.40698457e+00 -6.38369977e-01
4.38820384e-02 1.02609289e+00 3.42108250e-01 4.38222796e-01
2.73940563e-01 -1.67420611e-01 -5.24551570e-01 -1.49718821e-01
2.06118867e-01 -2.92250901e-01 1.18199140e-01 1.67202920e-01
-7.44340062e-01 1.17077363e+00 1.97610408e-01 8.05604339e-01
-3.01453620e-01 -5.00932336e-01 7.02259481e-01 4.13469642e-01
7.98964441e-01 3.48741025e-01 -6.51312947e-01 -5.01763262e-02
-1.02700388e+00 2.56998003e-01 1.23023629e+00 3.96382362e-01
7.41972029e-01 -7.81193599e-02 5.20495363e-02 9.62742150e-01
1.41718119e-01 4.61335868e-01 1.00477731e+00 -2.49352500e-01
6.66177273e-01 7.54194975e-01 6.12403005e-02 -1.26090693e+00
-6.47038996e-01 -5.37216008e-01 -8.72857988e-01 -2.29907885e-01
7.37952352e-01 -6.34477913e-01 -5.93147218e-01 1.68966019e+00
8.87309760e-02 -2.86480427e-01 8.95731151e-02 4.95362043e-01
1.10200572e+00 6.18156433e-01 7.18169138e-02 -2.11960152e-01
1.31470811e+00 -8.72079730e-01 -6.12480879e-01 -3.78515631e-01
5.44887006e-01 -6.14478409e-01 8.20395112e-01 4.65544045e-01
-9.16503191e-01 -7.82365263e-01 -8.60393465e-01 7.43431151e-02
-8.64695489e-01 5.19052684e-01 2.80732214e-01 8.76702905e-01
-9.48786199e-01 5.51850259e-01 -4.31757331e-01 -4.82752740e-01
1.17636219e-01 4.92993623e-01 -4.94688362e-01 2.89899021e-01
-1.37225556e+00 9.23184395e-01 3.05416256e-01 2.09847525e-01
-7.07884505e-02 -1.57440215e-01 -6.60315156e-01 -1.98985845e-01
-1.44100800e-01 -2.72047371e-02 9.19015527e-01 -1.52206469e+00
-1.53148162e+00 1.12816584e+00 -1.29008070e-01 -4.56534803e-01
9.51393604e-01 -1.30344138e-01 -5.21944106e-01 -3.80290151e-01
2.82179266e-01 4.81274784e-01 5.80427647e-01 -1.07570946e+00
-6.41626656e-01 -3.60652983e-01 -1.28160998e-01 -1.88263014e-01
-6.66585028e-01 2.84980297e-01 -3.81536275e-01 -3.68085176e-01
-1.51133642e-01 -1.04757011e+00 1.22880600e-01 -7.52272964e-01
-5.54743230e-01 -3.54971647e-01 5.04492640e-01 -9.09720182e-01
1.55449855e+00 -2.25518131e+00 -1.84642512e-03 4.77141231e-01
8.14956874e-02 6.77650571e-02 2.18477920e-01 6.10141516e-01
-3.92860845e-02 2.04435229e-01 1.65834039e-01 -5.78781009e-01
-1.25270292e-01 7.04052225e-02 8.17824155e-02 3.66742641e-01
3.18937488e-02 5.04734755e-01 -6.17742777e-01 -6.33952856e-01
-9.57248732e-02 2.99686432e-01 -4.08555090e-01 6.95629790e-03
1.66949838e-01 4.83053714e-01 -2.54374862e-01 2.29788721e-01
4.89161849e-01 2.69386042e-02 3.37974191e-01 1.19380169e-01
-4.61987436e-01 1.38522938e-01 -1.02440464e+00 9.01168227e-01
-7.12810516e-01 8.71970654e-01 1.51425945e-02 -7.62145042e-01
1.29681170e+00 2.03859910e-01 2.64508605e-01 -6.65096819e-01
4.35971141e-01 4.69867468e-01 1.80893779e-01 -5.72833598e-01
6.10295594e-01 6.48979694e-02 -4.08162266e-01 4.78712916e-01
-9.00767073e-02 2.99171269e-01 7.39200056e-01 -3.23868901e-01
4.06979501e-01 -8.74181911e-02 7.49626756e-02 -1.73165530e-01
9.30263698e-01 -4.68153864e-01 4.78721589e-01 5.49589634e-01
-8.02645162e-02 3.56947809e-01 8.79530132e-01 -6.61326885e-01
-8.98979723e-01 -5.87281287e-01 -3.33975285e-01 1.56876647e+00
-3.31889391e-01 -3.91901314e-01 -8.57684433e-01 -5.65159857e-01
3.98322335e-03 6.51947379e-01 -8.69572163e-01 1.15979753e-01
-6.09978259e-01 -1.04917157e+00 7.55442679e-01 2.36732036e-01
6.17381632e-01 -1.10072052e+00 -3.14618051e-01 -2.03252994e-02
-1.97961196e-01 -1.28775465e+00 4.32994068e-02 2.16904551e-01
-4.54049557e-01 -9.07574594e-01 -6.13917291e-01 -8.86896789e-01
4.73370135e-01 -6.01399362e-01 9.40759420e-01 -1.56882137e-01
1.95843399e-01 -2.84879923e-01 -2.90592521e-01 -6.73103631e-01
-4.49597746e-01 6.64751947e-01 9.92390290e-02 4.27080870e-01
6.18023932e-01 -4.29282576e-01 -3.72628793e-02 2.23598197e-01
-4.08180833e-01 -5.00151701e-02 4.72610474e-01 6.60479009e-01
-5.53548299e-02 5.99045530e-02 4.74865168e-01 -1.16600931e+00
7.14427233e-01 -7.20038891e-01 -2.66308069e-01 -1.54822230e-01
-5.17650485e-01 -1.18362255e-01 9.45913076e-01 -6.26981556e-01
-7.97921419e-01 -2.17198268e-01 -1.77042425e-01 8.74086544e-02
-1.54120311e-01 6.37638390e-01 8.97294804e-02 3.35780621e-01
7.18590915e-01 1.69869754e-02 -5.70371673e-02 -6.05888307e-01
-2.32680663e-01 1.01485038e+00 2.01981485e-01 -2.30129376e-01
5.68676770e-01 3.99936475e-02 -2.84133345e-01 -9.98370349e-01
-6.64888442e-01 -3.95162888e-02 -9.61656988e-01 -1.51363432e-01
8.18783164e-01 -8.14792573e-01 -5.53874075e-01 7.74173021e-01
-8.01766038e-01 -3.08181107e-01 4.71448526e-02 6.79568112e-01
-1.39881233e-02 -2.08700821e-01 -6.97883725e-01 -8.06859374e-01
-2.88073957e-01 -9.09748018e-01 7.23467708e-01 3.49733680e-01
-5.99975586e-01 -1.35344434e+00 -8.93798918e-02 2.37432063e-01
3.96882921e-01 3.92697901e-01 9.56010044e-01 -1.32663929e+00
4.94580716e-01 -5.73748171e-01 -2.18622968e-01 3.19499493e-01
-1.01809902e-02 4.16149616e-01 -1.12455440e+00 -1.62723124e-01
-2.91434199e-01 -3.83677751e-01 6.07823193e-01 2.20361114e-01
9.75453734e-01 -3.39520544e-01 -1.26809463e-01 2.82691240e-01
1.24100649e+00 -1.91976994e-01 1.90617397e-01 5.71410179e-01
4.83589262e-01 8.16842139e-01 1.06456093e-01 5.11561573e-01
7.23798692e-01 4.97085363e-01 -4.92420420e-02 -3.29537958e-01
3.46117765e-01 -2.51895577e-01 5.13092518e-01 8.25676560e-01
-3.99269044e-01 -2.02533171e-01 -1.01193678e+00 2.77845412e-01
-1.66358292e+00 -8.58018935e-01 -4.26106513e-01 2.28884292e+00
9.06309605e-01 4.91443634e-01 5.96343040e-01 5.76005578e-01
6.43754005e-01 5.92908524e-02 1.72680151e-02 -9.65075314e-01
-1.75104693e-01 2.17688218e-01 5.32638550e-01 4.66044813e-01
-1.40765870e+00 8.04456949e-01 6.00580502e+00 6.09247446e-01
-1.81712091e+00 7.62195140e-02 8.82751346e-01 -6.08359762e-02
1.76543385e-01 -3.85573000e-01 -1.02955687e+00 7.01564372e-01
1.35255957e+00 5.15104346e-02 2.96069980e-01 6.97167754e-01
3.93173307e-01 -2.27029845e-01 -8.50059986e-01 8.16254735e-01
1.01919159e-01 -6.75105155e-01 -3.16198021e-01 7.91092124e-03
5.83358645e-01 1.94535583e-01 -6.36560842e-02 6.31677806e-01
6.59184083e-02 -1.13562500e+00 9.30062532e-01 6.01576507e-01
6.54178083e-01 -7.47673273e-01 1.25575674e+00 7.81848133e-01
-7.38119304e-01 -3.80637050e-01 1.55977218e-03 -3.51985633e-01
-3.03668082e-01 5.26846528e-01 -8.21138680e-01 4.32878643e-01
6.98365808e-01 6.80682361e-01 -1.07721555e+00 3.80275726e-01
-1.23895235e-01 6.94053769e-01 -2.43813440e-01 -3.71581733e-01
3.07687014e-01 -1.04378037e-01 4.58397567e-02 1.57463920e+00
1.03400812e-01 -6.33974254e-01 3.37448537e-01 6.12478495e-01
-7.38422722e-02 5.69510937e-01 -4.52156544e-01 -1.41155692e-02
3.36430490e-01 1.42614508e+00 -7.88007259e-01 -1.68410450e-01
-4.69415933e-01 5.24589062e-01 2.99212277e-01 2.16973260e-01
-7.16068029e-01 -4.15459901e-01 1.65452026e-02 3.83952826e-01
-1.01590557e-02 -1.05435155e-01 -5.04713297e-01 -8.65312755e-01
-1.22792378e-01 -6.89368129e-01 2.53691941e-01 -3.17280173e-01
-1.32333422e+00 6.97907329e-01 8.26934651e-02 -9.16158259e-01
-5.90128481e-01 -8.27445924e-01 -5.16817749e-01 1.10750079e+00
-1.23802042e+00 -1.25123799e+00 -3.03058237e-01 3.10155898e-01
2.75252670e-01 -5.32400846e-01 8.46140385e-01 6.90960467e-01
-8.26672256e-01 7.78869152e-01 1.79482326e-01 7.08648324e-01
7.17574835e-01 -1.11014855e+00 -1.83272548e-03 2.70059675e-01
-1.62964746e-01 5.75673461e-01 5.43362558e-01 -4.24221754e-01
-8.12429309e-01 -1.07429862e+00 1.80875278e+00 -2.65146852e-01
5.60765982e-01 -6.43822551e-01 -5.22618353e-01 5.68723798e-01
1.15472749e-01 -1.54691964e-01 8.53507161e-01 3.60324800e-01
-1.28875524e-01 -1.25704512e-01 -1.03722620e+00 2.66296595e-01
3.03680420e-01 -4.54014093e-01 -2.35522449e-01 2.21543401e-01
5.44315130e-02 -2.93007821e-01 -9.39486206e-01 -3.76544669e-02
1.08001101e+00 -1.17072332e+00 5.43881416e-01 -4.52540010e-01
8.64938855e-01 1.54065505e-01 -6.38057524e-03 -1.09421551e+00
-2.32065976e-01 -1.03767991e-01 4.54194784e-01 1.53452229e+00
7.85968482e-01 -8.46358895e-01 4.69781995e-01 3.27657849e-01
4.63509440e-01 -8.08677316e-01 -7.37449527e-01 -5.70705533e-01
4.86082464e-01 -3.86590034e-01 6.08028471e-01 9.96237278e-01
-2.65892893e-01 6.21607840e-01 -3.44487816e-01 -5.35376191e-01
-1.15338359e-02 -8.81186053e-02 8.82427096e-01 -1.64348936e+00
-7.57856071e-02 -6.24526381e-01 -1.48722962e-01 -2.61084527e-01
4.70083743e-01 -8.32841635e-01 -2.84762800e-01 -1.04199672e+00
9.51372907e-02 -6.25738740e-01 3.08122914e-02 6.28516197e-01
6.78023845e-02 7.06571817e-01 2.51425594e-01 2.14166850e-01
-1.19344287e-01 7.09456876e-02 7.36615956e-01 5.57283089e-02
-5.44127941e-01 4.38871622e-01 -5.27238369e-01 1.00156403e+00
7.43123114e-01 -2.40753263e-01 5.86249419e-02 -1.30324319e-01
4.77972120e-01 -2.82540470e-01 1.06912039e-01 -7.66024292e-01
-1.13417692e-01 5.64741418e-02 7.54787624e-01 -2.48841643e-01
2.00856566e-01 -6.39353812e-01 -1.37642264e-01 4.71197575e-01
-3.83259594e-01 1.89838871e-01 8.73366594e-02 -5.59102036e-02
-2.45364428e-01 -2.25343853e-01 7.33528793e-01 -4.33849171e-02
-2.49490172e-01 2.55245380e-02 -4.44085807e-01 1.10445991e-01
7.91657984e-01 -3.40696692e-01 7.76009485e-02 -4.51635748e-01
-9.47205126e-01 4.63091247e-02 3.23292762e-01 6.71883523e-01
-9.51951593e-02 -1.16095984e+00 -1.00386965e+00 5.92104971e-01
7.58543005e-03 -5.28106093e-01 -2.55699515e-01 1.05373645e+00
-3.97475958e-01 3.91135156e-01 -2.22970113e-01 -4.27581668e-01
-1.24901259e+00 4.95270267e-02 3.27332675e-01 -6.59640312e-01
4.55735885e-02 7.34769702e-01 -2.26189896e-01 -9.26469982e-01
1.91523299e-01 -7.03667104e-02 -8.41008961e-01 6.94137037e-01
2.30839536e-01 4.87805903e-01 2.42385060e-01 -1.19248772e+00
-2.24629521e-01 4.01207417e-01 1.70190662e-01 -2.13372305e-01
1.40105951e+00 -1.10419118e-03 -3.55478197e-01 1.01071930e+00
1.19217229e+00 5.47849238e-01 -4.29901719e-01 -8.34200606e-02
1.06823131e-01 -9.87167656e-02 -1.30380601e-01 -7.14928865e-01
-9.21884060e-01 8.00813437e-01 4.15144712e-01 5.30531168e-01
8.60396147e-01 -3.84477884e-01 3.25543553e-01 8.16892833e-02
1.22828588e-01 -1.45716715e+00 -4.07157898e-01 9.21429694e-01
7.47284710e-01 -1.15042627e+00 1.35131106e-02 -1.16930820e-01
-6.29837990e-01 1.44841838e+00 4.97575551e-01 -1.45807475e-01
8.93115580e-01 1.11299880e-01 3.64656970e-02 2.03726023e-01
-5.96120179e-01 4.45795245e-02 3.72663707e-01 2.25063011e-01
9.77083623e-01 8.06593448e-02 -6.91843748e-01 1.02066576e+00
-9.45887864e-01 5.28263599e-02 5.19932806e-01 6.55018747e-01
-3.17679271e-02 -1.08506191e+00 -4.48754817e-01 5.74228168e-01
-7.29400456e-01 4.38850559e-02 -6.82133734e-01 1.20798051e+00
7.20150292e-01 1.00893009e+00 2.67789692e-01 -5.21101952e-01
1.44832194e-01 2.92153120e-01 1.99450999e-01 -4.32662904e-01
-1.12680411e+00 -1.84782252e-01 4.51085299e-01 -8.95733293e-03
-2.58696586e-01 -7.56768465e-01 -1.03252375e+00 -6.21722281e-01
-3.37934762e-01 3.03596556e-01 8.02281976e-01 1.06030154e+00
-7.53062442e-02 1.52455583e-01 1.05678308e+00 -8.00865769e-01
-3.50007296e-01 -1.40542352e+00 -5.96944153e-01 4.83731031e-01
2.74029225e-01 -3.55561078e-01 -5.24918139e-01 1.64154358e-02] | [9.47107219696045, 10.361076354980469] |
62e8f7fe-f749-4c08-b8f6-4f7ad688768a | a-generalised-directional-laplacian | 1708.04816 | null | http://arxiv.org/abs/1708.04816v1 | http://arxiv.org/pdf/1708.04816v1.pdf | A Generalised Directional Laplacian Distribution: Estimation, Mixture Models and Audio Source Separation | Directional or Circular statistics are pertaining to the analysis and
interpretation of directions or rotations. In this work, a novel probability
distribution is proposed to model multidimensional sparse directional data. The
Generalised Directional Laplacian Distribution (DLD) is a hybrid between the
Laplacian distribution and the von Mises-Fisher distribution. The
distribution's parameters are estimated using Maximum-Likelihood Estimation
over a set of training data points. Mixtures of Directional Laplacian
Distributions (MDLD) are also introduced in order to model multiple
concentrations of sparse directional data. The author explores the application
of the derived DLD mixture model to cluster sound sources that exist in an
underdetermined instantaneous sound mixture. The proposed model can solve the
general K x L (K<L) underdetermined instantaneous source separation problem,
offering a fast and stable solution. | ['Nikolaos Mitianoudis'] | 2017-08-16 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [-1.79526061e-01 -4.18061703e-01 1.04379875e-03 -1.19642556e-01
-9.65912461e-01 -6.98897839e-01 3.09830576e-01 -5.98105550e-01
5.40785715e-02 6.66953504e-01 5.18235028e-01 -1.80307075e-01
-6.44266427e-01 -4.88680303e-02 -9.19136554e-02 -1.13286853e+00
-1.86054394e-01 1.64141342e-01 -7.09332526e-02 5.17779827e-01
2.74938971e-01 8.31597090e-01 -1.41797709e+00 -1.37591407e-01
7.81941056e-01 5.50872207e-01 3.45939875e-01 1.09582269e+00
-1.34039834e-01 5.22246838e-01 -5.16001165e-01 6.98314086e-02
6.00090623e-02 -5.08480787e-01 -6.51373267e-02 -7.10147154e-03
3.31606954e-01 -5.16431332e-02 -2.08139122e-01 1.17848122e+00
8.83459866e-01 2.18895540e-01 1.45794928e+00 -1.40588510e+00
-5.85314453e-01 4.17997062e-01 -7.58017778e-01 4.26338285e-01
4.78270710e-01 -7.10266471e-01 5.30104458e-01 -1.27445078e+00
2.02715173e-01 1.38864112e+00 7.42930353e-01 7.46699981e-04
-1.15616572e+00 -4.94611233e-01 -3.25226516e-01 -7.06207827e-02
-1.89905632e+00 -6.51300669e-01 1.27514279e+00 -5.85402668e-01
5.89422047e-01 1.48336500e-01 2.73249686e-01 8.18727195e-01
1.98880792e-01 8.66167486e-01 9.91231263e-01 -5.30969441e-01
2.87784785e-01 6.94249496e-02 -1.33955916e-02 4.14190918e-01
5.28689682e-01 -2.02408969e-01 -6.75217271e-01 -6.74777150e-01
9.93618548e-01 -5.40811121e-01 -2.64226466e-01 -4.59252805e-01
-1.17755592e+00 6.53213859e-01 -3.03620309e-01 6.99863255e-01
-5.84481895e-01 -1.19681783e-01 -2.43373036e-01 -1.12157717e-01
4.80417311e-01 1.25805894e-02 -1.49629846e-01 -2.53976464e-01
-1.18052220e+00 1.30401686e-01 1.13583815e+00 1.03110945e+00
5.90270340e-01 6.04863584e-01 2.16847599e-01 1.02025461e+00
1.09741402e+00 1.35344374e+00 3.05254519e-01 -1.08187294e+00
2.13960320e-01 -2.20071934e-02 1.72913864e-01 -1.46728742e+00
-6.31245434e-01 -6.12021863e-01 -1.15223897e+00 -6.38490990e-02
2.61429459e-01 -4.32535380e-01 -4.51358914e-01 1.68591464e+00
3.56941164e-01 5.36445141e-01 2.94466406e-01 8.32579970e-01
6.80630624e-01 8.29253197e-01 -3.39082152e-01 -6.21022761e-01
7.03971624e-01 -2.44101569e-01 -1.33078063e+00 6.40309304e-02
7.36713633e-02 -1.30960333e+00 4.34041977e-01 7.13540971e-01
-9.81956542e-01 -4.50561345e-01 -6.28275812e-01 2.05704361e-01
-1.78049952e-01 4.92861390e-01 3.53188992e-01 1.01490581e+00
-1.18354166e+00 -1.58819661e-01 -7.41915047e-01 1.70142315e-02
-2.08894953e-01 7.55044743e-02 -1.48098975e-01 -1.23261005e-01
-8.33936393e-01 5.25541008e-01 -3.33607882e-01 3.63308072e-01
-7.04390109e-01 -5.64873099e-01 -7.40818620e-01 -2.03687370e-01
-4.31000918e-01 -3.64610314e-01 7.89101303e-01 -3.66447538e-01
-1.57479477e+00 3.03214967e-01 -7.53273964e-01 3.39731872e-01
1.15493990e-01 -2.21654981e-01 -8.63841534e-01 3.12055349e-01
3.17439765e-01 4.83113863e-02 1.18508506e+00 -1.63972270e+00
-1.80405200e-01 -1.84946761e-01 -9.06483948e-01 3.25759649e-01
6.41087443e-02 -1.91299126e-01 -4.23676014e-01 -9.39102769e-01
8.31598639e-01 -8.76703680e-01 -2.35301182e-02 -5.39483249e-01
-6.62737310e-01 -6.36405572e-02 8.16832364e-01 -8.62476289e-01
1.46994901e+00 -2.40922570e+00 3.04848135e-01 7.37393737e-01
-6.30010366e-02 -1.99731097e-01 -9.73411351e-02 5.12617350e-01
-1.53011203e-01 -3.01284850e-01 -4.48131114e-01 -4.20063615e-01
1.40137792e-01 -6.58670068e-02 -1.99081331e-01 8.98715615e-01
-4.10281777e-01 1.37137190e-01 -6.82838321e-01 -4.71178830e-01
1.40704840e-01 9.69699860e-01 -5.22734463e-01 1.91982061e-01
6.18244588e-01 5.34804165e-01 -2.28347316e-01 6.76503241e-01
1.23486352e+00 2.16974750e-01 7.06712529e-02 -5.14453411e-01
-3.02484870e-01 -2.72706181e-01 -2.20424962e+00 1.70146704e+00
-2.34709993e-01 8.45887840e-01 6.84506893e-01 -7.75430381e-01
1.09600520e+00 6.51316106e-01 7.79150903e-01 7.09089413e-02
-4.54493351e-02 4.78992522e-01 -2.43997518e-02 -9.27892983e-01
3.84372532e-01 -3.71341825e-01 -4.59265150e-02 4.15985405e-01
2.36322731e-01 -5.52934587e-01 4.49887142e-02 2.86969513e-01
5.54685652e-01 -8.25986415e-02 2.45204031e-01 -6.93802714e-01
6.27909601e-01 -5.43108642e-01 4.57192391e-01 6.29696071e-01
-3.52554135e-02 5.49244761e-01 1.87988833e-01 4.16040719e-01
-5.88598788e-01 -1.34564435e+00 -3.53537470e-01 6.86584115e-01
1.10575430e-01 -5.18832430e-02 -6.12777829e-01 1.64224342e-01
-1.81549024e-02 9.79610026e-01 2.62775086e-02 3.93963829e-02
-2.72407860e-01 -1.04476237e+00 7.83583939e-01 1.03614911e-01
2.55012274e-01 -2.86360860e-01 -8.87473449e-02 1.25365555e-01
-4.26000446e-01 -8.43066454e-01 -4.47846413e-01 2.06412986e-01
-6.71460211e-01 -8.09312940e-01 -1.06218004e+00 -7.71644652e-01
5.90271771e-01 4.62992251e-01 7.24231362e-01 -1.06229603e+00
-7.38844648e-02 1.22468591e+00 -1.50401548e-01 -3.58131468e-01
-3.41202348e-01 -3.96274120e-01 6.69137180e-01 4.44011003e-01
3.90600204e-01 -8.52367640e-01 -2.93879420e-01 3.93447936e-01
-8.12030911e-01 -5.89511871e-01 5.77593267e-01 2.81431973e-01
5.30816138e-01 6.50711179e-01 6.10311747e-01 -1.91299841e-01
1.07635808e+00 -8.09784532e-01 -4.36692923e-01 2.32094340e-02
-2.12549105e-01 1.53582260e-01 1.80025443e-01 -6.68454528e-01
-1.56738257e+00 -3.62362713e-02 -3.96778584e-02 -6.46779478e-01
-6.79640949e-01 4.76231992e-01 -3.58677775e-01 -1.43913954e-01
4.18674827e-01 4.81830150e-01 -2.30695814e-01 -9.41158056e-01
5.13608992e-01 8.76981437e-01 6.30909741e-01 -5.60152531e-01
5.75078309e-01 3.50749284e-01 2.47750908e-01 -1.61367965e+00
-2.04480976e-01 -8.82880807e-01 -9.13966537e-01 -3.28598827e-01
8.74009311e-01 -8.20492506e-01 -3.97614479e-01 7.97268450e-01
-1.13991117e+00 -1.65270735e-02 -4.81605381e-02 1.26806700e+00
-5.30849576e-01 4.88989919e-01 -2.92785734e-01 -1.44198358e+00
4.39807534e-01 -1.04036272e+00 1.07364726e+00 2.33516350e-01
-1.55669183e-01 -1.46279490e+00 5.36985219e-01 -2.32192129e-01
5.10733783e-01 -1.35663062e-01 6.03084624e-01 -4.00970519e-01
-1.44288987e-01 -2.89349586e-01 1.63173854e-01 4.17805314e-01
4.98994261e-01 7.20676184e-02 -9.09350514e-01 -1.62112907e-01
4.42480028e-01 3.14914137e-01 4.22218472e-01 1.05752027e+00
5.49894810e-01 -1.79953039e-01 -3.84266943e-01 6.23929501e-01
1.23844147e+00 3.65190148e-01 2.12639004e-01 -3.75855535e-01
6.93956375e-01 5.17865241e-01 2.10925072e-01 5.52253246e-01
3.42102766e-01 3.27474982e-01 2.23002378e-02 6.49686232e-02
-2.29382589e-01 -1.32182136e-01 2.73620814e-01 1.41851830e+00
-1.67175978e-01 -3.65420163e-01 -7.80717134e-01 5.78695536e-01
-1.27536285e+00 -1.12477005e+00 -6.41326725e-01 2.06685686e+00
5.11415958e-01 -4.55679923e-01 -1.12116396e-01 2.67605871e-01
8.11923802e-01 1.83395937e-01 -2.88458139e-01 -7.03557953e-02
-5.22037029e-01 -2.46693924e-01 4.05105174e-01 1.02583051e+00
-1.14362538e+00 5.78859448e-01 7.73587656e+00 1.07130980e+00
-9.68147218e-01 4.66282368e-02 -1.42000332e-01 2.30166718e-01
-4.61805612e-01 -2.55576700e-01 -8.00149918e-01 3.56441915e-01
7.49441028e-01 -7.70783350e-02 3.63799989e-01 5.66328108e-01
4.48228747e-01 -3.53433937e-01 -4.94376957e-01 1.52046573e+00
4.97406691e-01 -5.38796127e-01 -1.61234155e-01 2.03288957e-01
9.88146305e-01 -2.04626411e-01 3.18041474e-01 -4.27343607e-01
2.04584137e-01 -5.50468802e-01 5.54783762e-01 1.07884061e+00
6.57730579e-01 -6.39172256e-01 3.57655406e-01 6.50727928e-01
-1.26970744e+00 1.26757817e-02 -4.41763818e-01 2.44361833e-01
6.93809748e-01 1.18856049e+00 -6.48729980e-01 4.74776030e-01
5.69122672e-01 7.33929634e-01 -3.50087844e-02 1.24034584e+00
-8.63722116e-02 8.54232907e-01 -7.60704637e-01 2.52319485e-01
5.85695840e-02 -7.40167558e-01 1.23130620e+00 1.57152510e+00
1.07816994e+00 8.99591893e-02 -1.95832208e-01 6.89336061e-01
5.62462091e-01 1.07005343e-01 -8.14175367e-01 3.81081849e-02
5.93038857e-01 1.21940696e+00 -6.08123183e-01 -8.83045197e-02
-4.29057181e-01 7.12647438e-01 -4.88701969e-01 1.02425790e+00
-4.98224705e-01 -1.99489683e-01 6.32755518e-01 -2.38430768e-01
2.02001169e-01 -7.52209961e-01 -2.13423312e-01 -1.10241473e+00
-5.81843853e-01 -6.43082917e-01 1.69557512e-01 -8.62342298e-01
-1.50851798e+00 3.86800468e-01 3.88112903e-01 -1.48708093e+00
-2.82404631e-01 -5.20826399e-01 -5.78783154e-01 1.01695335e+00
-1.20576930e+00 -7.29625106e-01 3.22188810e-02 9.49173450e-01
2.82395214e-01 -5.84774315e-01 7.48547256e-01 5.47166467e-01
-3.92317802e-01 1.57526389e-01 5.77666879e-01 -3.39604110e-01
8.17145586e-01 -1.45089686e+00 -5.24109721e-01 9.20941651e-01
1.38869017e-01 7.82657743e-01 1.13665402e+00 -6.54416442e-01
-1.49306619e+00 -5.98782063e-01 7.76156962e-01 -4.79954444e-02
5.22184551e-01 4.75857891e-02 -7.99821019e-01 4.14271802e-01
1.63116992e-01 -3.44146699e-01 1.34293270e+00 -2.75304168e-01
-1.21757329e-01 1.05251998e-01 -1.11533260e+00 9.94492322e-02
6.21330976e-01 -6.25219166e-01 -6.44193530e-01 2.73811638e-01
-2.04958878e-02 -3.27546126e-03 -9.09594059e-01 1.48606643e-01
4.27120239e-01 -9.68000889e-01 1.24145710e+00 3.86591814e-02
-3.50067466e-01 -5.88832796e-01 -7.96716034e-01 -1.51267445e+00
-7.60372400e-01 -8.19108665e-01 -1.25615060e-01 1.34779727e+00
3.18493098e-01 -5.01475036e-01 1.59983248e-01 2.24410966e-01
-1.78814128e-01 4.37033921e-02 -1.21161008e+00 -5.86168230e-01
1.14926035e-02 -8.51368666e-01 2.07945451e-01 1.03293419e+00
-1.65747218e-02 1.65153936e-01 -6.21195138e-01 8.51699948e-01
1.21611285e+00 -1.05510511e-01 5.28012037e-01 -1.23096037e+00
-1.91628531e-01 -4.86363679e-01 -9.11502987e-02 -1.51135135e+00
3.24646682e-01 -7.74667382e-01 2.50741839e-01 -1.57814705e+00
-2.69075006e-01 -4.88966852e-01 -7.46099725e-02 -3.10268223e-01
8.41394737e-02 1.35508284e-01 -2.09412992e-01 3.55091900e-01
-1.91639036e-01 5.78018904e-01 9.83220518e-01 1.37924299e-01
-4.65955049e-01 4.39707786e-01 -4.26733643e-01 1.04348731e+00
6.06615961e-01 -5.02970159e-01 -6.91595316e-01 -3.39161813e-01
1.79749429e-01 3.29193920e-01 6.66083694e-02 -9.02235389e-01
5.92947781e-01 -9.98970959e-03 2.56861269e-01 -9.17743325e-01
4.94872361e-01 -9.79132473e-01 3.36362571e-01 -1.50924250e-01
-1.13951303e-01 -1.65105358e-01 4.57192734e-02 7.12306976e-01
-5.52499473e-01 -1.60246983e-01 6.89788043e-01 2.79076129e-01
-4.34566855e-01 -1.00954592e-01 -8.25547695e-01 -2.45029613e-01
7.63384104e-01 5.79144321e-02 6.62076771e-02 -8.42835128e-01
-1.02631044e+00 -4.24197108e-01 -1.75365895e-01 3.91736105e-02
6.45552218e-01 -1.61739016e+00 -7.14636564e-01 6.37886882e-01
-5.63430309e-01 -1.48120880e-01 5.96189618e-01 1.30453122e+00
-4.24030721e-01 4.67486441e-01 2.41319001e-01 -7.10261166e-01
-8.43804538e-01 3.29932064e-01 3.01555753e-01 3.09957057e-01
-1.61379158e-01 1.04684949e+00 1.74907312e-01 -6.35540962e-01
1.21910840e-01 -1.31449595e-01 -4.21811283e-01 3.45679939e-01
3.59808952e-01 1.10364330e+00 -4.66955125e-01 -1.23901415e+00
-5.98244965e-01 1.06525958e+00 9.49918151e-01 -6.90522671e-01
1.10824704e+00 -6.67857289e-01 -5.06359637e-01 8.28937531e-01
1.49930322e+00 8.55102241e-01 -1.05101335e+00 -1.20043054e-01
-1.59156799e-01 -4.17710066e-01 2.62874544e-01 -5.19627929e-01
-8.00301552e-01 9.31270778e-01 5.87403715e-01 4.88686234e-01
1.13667285e+00 1.61325902e-01 3.86164673e-02 2.57926527e-02
-2.69317292e-02 -8.62737298e-01 -8.45285729e-02 4.49251860e-01
1.05935061e+00 -8.34948659e-01 -1.83412164e-01 -4.75001216e-01
-6.67857826e-01 1.20957351e+00 -2.01631170e-02 -6.70583919e-02
1.43545306e+00 8.24904323e-01 2.96653509e-01 6.36151507e-02
-1.10215686e-01 6.59114346e-02 7.77258039e-01 9.64231312e-01
5.87693989e-01 1.65922821e-01 -2.43767992e-01 7.01868236e-01
-1.90872714e-01 -3.53074461e-01 3.64723682e-01 5.46476662e-01
-6.15034103e-01 -6.79846525e-01 -1.13599372e+00 7.56052211e-02
-2.96499908e-01 3.91944721e-02 -3.32929753e-02 2.19370097e-01
1.17405176e-01 1.40413737e+00 -4.67734374e-02 -3.53718586e-02
1.20190807e-01 2.68910021e-01 1.62337929e-01 -2.36092750e-02
7.06020117e-01 1.35239542e+00 -2.02649981e-01 -2.22371563e-01
-9.44073737e-01 -1.01017654e+00 -9.09254849e-01 6.04212359e-02
-2.34958559e-01 5.26958287e-01 9.89344299e-01 6.69396758e-01
1.99586898e-01 4.56178576e-01 5.57735622e-01 -1.03925121e+00
-2.42322892e-01 -1.06300187e+00 -1.45550394e+00 -5.30408584e-02
5.58219790e-01 -7.88129389e-01 -9.44319606e-01 3.73738080e-01] | [15.23755931854248, 5.664351463317871] |
1e583fa8-70db-4a97-bdef-f7f52c80ae4b | underwater-single-channel-acoustic-signal | null | null | https://asa.scitation.org/doi/abs/10.1121/10.0009852 | https://asa.scitation.org/doi/10.1121/10.0009852 | Underwater single-channel acoustic signal multitarget recognition using convolutional neural networks | The radiated noise from ships is of great significance to target recognition, and several deep learning methods have been developed for the recognition of underwater acoustic signals. Previous studies have focused on single-target recognition, with relatively few reports on multitarget recognition. This paper proposes a deep learning-based single-channel multitarget underwater acoustic signal recognition method for an unknown number of targets in the specified category. The proposed method allows the two subproblems of recognizing the unique class and duplicate categories of multiple targets to be solved. These two tasks are essentially multilabel binary classification and multilabel multiple value classification, respectively. In this paper, we describe the use of real-valued and complex-valued ResNet and DenseNet convolutional networks to recognize synthetic mixed multitarget signals, which was superimposed from individual target signals. We compare the performance of various features, including the original audio signal, complex-valued short-time Fourier transform (STFT) spectrum, magnitude STFT spectrum, logarithmic mel spectrum, and mel frequency cepstral coefficients. The experimental results show that our method can effectively recognize synthetic multitarget ship signals when the magnitude STFT spectrum, complex-valued STFT spectrum, and log-mel spectrum are used as network inputs. | ['Kejun Wang', 'Qinggang Sun'] | 2022-03-31 | null | null | null | the-journal-of-the-acoustical-society-of | ['audio-multiple-target-classification'] | ['audio'] | [ 5.09712875e-01 -4.97853667e-01 7.10426450e-01 -4.89527613e-01
-1.27568400e+00 -6.47009671e-01 2.53712118e-01 -6.01828285e-02
-7.74631977e-01 5.81027925e-01 1.38766006e-01 2.02287966e-03
-2.49504715e-01 -7.07072496e-01 -5.70568383e-01 -1.08989596e+00
-5.07853866e-01 -1.18528284e-01 1.00262374e-01 -3.40267420e-01
8.55955184e-02 5.60580075e-01 -1.70919716e+00 4.88641053e-01
3.12067896e-01 1.37581336e+00 9.37101170e-02 1.16378069e+00
1.92222372e-01 9.33354676e-01 -1.29869938e+00 -1.30654931e-01
5.82545042e-01 -4.80959266e-01 -4.36350591e-02 -5.74509859e-01
5.93713403e-01 -2.52481848e-01 -2.05677956e-01 1.38026917e+00
1.09372854e+00 2.89684325e-01 7.72792459e-01 -1.10368943e+00
-3.13226342e-01 8.34616482e-01 -2.62501299e-01 4.65167165e-01
1.45877327e-03 -1.59663826e-01 9.45948184e-01 -8.09735954e-01
-4.91861343e-01 1.43722582e+00 1.18035603e+00 1.60203621e-01
-5.15059590e-01 -1.07874656e+00 -3.11875641e-01 3.08751225e-01
-1.31438494e+00 -4.06753123e-01 5.92536807e-01 -5.01710236e-01
8.92239809e-01 3.61951858e-01 6.12873316e-01 7.76193798e-01
5.16372263e-01 5.76714754e-01 1.04944003e+00 -3.34237337e-01
4.98047918e-02 -4.49569643e-01 -2.37330943e-02 3.54275763e-01
-4.66590524e-02 5.28707922e-01 -2.70008594e-01 -2.58055329e-01
4.72834766e-01 -1.19660117e-01 -4.15439129e-01 5.55158019e-01
-1.07734525e+00 1.01476550e+00 1.52769849e-01 3.61841738e-01
-2.65984595e-01 3.03811282e-01 4.68583494e-01 8.03785205e-01
3.19184303e-01 2.78286219e-01 -6.13301277e-01 -6.09754864e-03
-5.48737824e-01 -4.36827214e-03 9.94612694e-01 4.82470125e-01
7.67104685e-01 8.67167950e-01 1.97328717e-01 9.88921046e-01
4.40222532e-01 1.25829864e+00 9.41867769e-01 -3.85781825e-01
6.15686923e-02 -1.08596444e-01 -4.09496650e-02 -1.19522250e+00
-7.23699808e-01 -5.17605782e-01 -8.71613324e-01 3.85792285e-01
1.89371258e-01 -7.34623432e-01 -8.76568854e-01 1.37097919e+00
-5.45151383e-02 6.77769780e-01 9.56144094e-01 9.73823190e-01
1.41284359e+00 7.27843702e-01 -1.95028573e-01 -2.04860225e-01
9.98018086e-01 -6.17472351e-01 -7.37762690e-01 -3.09422761e-01
2.89590567e-01 -9.24124122e-01 3.38316739e-01 3.21822375e-01
-3.95351619e-01 -6.43753588e-01 -1.25571775e+00 6.94301128e-01
-3.59396815e-01 1.42576188e-01 2.16060281e-01 9.14452732e-01
-6.18835151e-01 4.09210265e-01 -3.95796448e-01 3.88037115e-01
4.48234333e-03 4.55934733e-01 -5.28793514e-01 3.33409496e-02
-1.62560141e+00 9.98850942e-01 3.78696114e-01 4.55251962e-01
-1.32772315e+00 -3.39822292e-01 -1.20280552e+00 1.04023091e-01
-3.44512910e-01 3.55775535e-01 1.45005941e+00 -1.20837855e+00
-1.44117188e+00 2.15762392e-01 4.38743055e-01 -3.36309820e-01
-2.90343408e-02 1.11262619e-01 -9.96207058e-01 7.46180937e-02
5.55827189e-03 -8.68391097e-02 9.08356488e-01 -9.19697464e-01
-9.71357226e-01 3.26231169e-03 -1.07519310e-02 3.29505146e-01
-3.50801736e-01 2.36557841e-01 6.38201416e-01 -8.03645015e-01
1.20560549e-01 -3.76835883e-01 -1.03047058e-01 -2.19063967e-01
-1.86796151e-02 3.62701565e-02 8.17915201e-01 -5.58235884e-01
4.54481959e-01 -2.56548643e+00 -1.41796172e-01 8.47224072e-02
-3.82686049e-01 2.95223624e-01 -7.52643883e-01 2.15748921e-01
-1.25747830e-01 -2.85220861e-01 -1.66248262e-01 6.25025257e-02
-1.23233654e-01 2.14511633e-01 -3.19077730e-01 8.33798766e-01
-2.52214428e-02 4.53587770e-01 -9.33303654e-01 2.92626135e-02
3.79192596e-03 4.61815447e-01 2.60217428e-01 2.61828125e-01
1.82611704e-01 2.34361421e-02 -1.05195127e-01 8.66168976e-01
8.73115897e-01 5.75751126e-01 -2.45289072e-01 -7.38722980e-01
-4.06000376e-01 -3.69210005e-01 -1.34245038e+00 8.81160557e-01
-3.32352340e-01 9.02754605e-01 3.17779660e-01 -1.06023717e+00
1.38291752e+00 4.75156337e-01 3.85953546e-01 -6.18135512e-01
5.03208458e-01 4.02624577e-01 2.15866685e-01 -6.91067994e-01
4.65119779e-01 -7.63602495e-01 -3.60834986e-01 1.02247104e-01
1.62138268e-02 -1.95494980e-01 -1.75577939e-01 -5.14723003e-01
8.64738584e-01 -6.13581121e-01 2.15707630e-01 -1.27961561e-01
2.81430066e-01 -1.03421785e-01 6.80691719e-01 8.67092252e-01
-1.17872804e-01 5.16603947e-01 -1.62414014e-01 -5.42427599e-01
-4.60199356e-01 -7.58524776e-01 -7.21917525e-02 1.34198034e+00
3.93308312e-01 4.67429698e-01 -1.20782165e-03 -2.31478408e-01
4.88782451e-02 2.13202596e-01 -5.67281961e-01 -2.16888070e-01
-7.36776471e-01 -1.03067863e+00 1.41790605e+00 5.27435482e-01
3.35007012e-01 -1.16600990e+00 -6.71433330e-01 4.03589666e-01
-9.93138403e-02 -8.66034329e-01 -3.17714989e-01 6.67079389e-01
-2.43425116e-01 -1.18504882e+00 -7.80880034e-01 -1.29392385e+00
-1.77921299e-02 1.90108851e-01 4.55400467e-01 -3.64177346e-01
-2.67396092e-01 4.21963125e-01 -6.69255078e-01 -7.89066315e-01
-4.43532139e-01 -7.55556464e-01 2.35143259e-01 5.23687363e-01
5.02584651e-02 -3.93239170e-01 -1.42438501e-01 2.68669307e-01
-1.05871010e+00 -6.84997320e-01 1.04401600e+00 1.05028200e+00
2.87360817e-01 1.60006344e-01 7.91548669e-01 1.11845098e-02
7.73697019e-01 -4.86157686e-01 -4.09079880e-01 2.60569245e-01
5.42681634e-01 -3.27530354e-01 8.20586264e-01 -1.07861066e+00
-5.80141008e-01 2.37270489e-01 -4.23835546e-01 -4.63830233e-01
-1.57854259e-01 1.14869308e+00 2.04950254e-02 -7.64454603e-01
8.58943045e-01 7.80377567e-01 5.57629317e-02 -3.96898657e-01
-1.16747580e-01 1.11210537e+00 6.83826804e-01 -6.17267713e-02
5.91208160e-01 2.60265261e-01 -2.67014951e-01 -1.17998600e+00
-7.63440132e-01 -5.89751959e-01 -2.47217268e-01 -6.05332077e-01
7.05238998e-01 -1.29985857e+00 -6.83801413e-01 1.14609993e+00
-1.21952724e+00 -2.27509484e-01 2.84514353e-02 1.04577696e+00
-7.80554786e-02 5.47091067e-01 -6.66082561e-01 -9.42087948e-01
-4.90767360e-01 -1.06302869e+00 9.77245808e-01 4.11607891e-01
5.64753413e-01 -7.73297608e-01 2.06264123e-01 -2.10572019e-01
6.72533333e-01 3.36371869e-01 3.82415503e-01 -1.15915191e+00
2.25848798e-02 -5.80510736e-01 2.27278732e-02 7.42067158e-01
3.20248932e-01 -4.05269384e-01 -9.83010292e-01 -5.68505883e-01
2.32330263e-01 -6.74278975e-01 8.65547597e-01 3.89889568e-01
4.67860758e-01 -3.07195485e-01 2.17684880e-01 6.99598074e-01
1.39310157e+00 6.91092849e-01 3.42858583e-01 9.82704610e-02
4.49117601e-01 2.85570204e-01 6.45802498e-01 6.58784449e-01
3.06344330e-01 1.57730475e-01 8.07922959e-01 -1.70643516e-02
1.13188490e-01 5.39401829e-01 6.30750060e-01 9.60076272e-01
1.01507358e-01 -3.59930843e-01 -7.38251984e-01 6.96793795e-01
-1.28372788e+00 -1.13613629e+00 -7.74824470e-02 1.75767493e+00
5.77048063e-01 -2.94859290e-01 -3.99466753e-01 1.46394461e-01
9.15725768e-01 1.05947942e-01 -6.61237597e-01 1.24630220e-01
-6.20061696e-01 3.36922377e-01 7.10642457e-01 4.58333850e-01
-1.64799309e+00 4.61802900e-01 6.29647779e+00 9.83949125e-01
-1.53930509e+00 -1.50317565e-01 9.48953331e-02 2.42014214e-01
3.37691940e-02 -6.55222118e-01 -7.45045424e-01 9.94786844e-02
1.07691526e+00 -1.49332136e-01 1.04527622e-01 6.86398447e-01
-5.35179712e-02 2.81832188e-01 -6.84946239e-01 1.03795648e+00
4.48328972e-01 -9.21753109e-01 3.79559137e-02 -3.58734846e-01
2.41829604e-01 4.73705053e-01 3.40459310e-02 5.23601651e-01
5.95226049e-01 -1.09508872e+00 1.10628450e+00 4.65453804e-01
9.02795434e-01 -5.80975652e-01 1.18537354e+00 3.06502849e-01
-1.63231993e+00 -4.83663261e-01 -7.51791060e-01 -2.26746261e-01
-5.21589071e-03 2.74162501e-01 -6.61681354e-01 6.49335563e-01
8.17265153e-01 7.76988566e-01 -5.00740223e-02 1.46886063e+00
1.56688422e-01 5.74914455e-01 -5.48384666e-01 -4.83824193e-01
2.74103940e-01 1.89604536e-01 4.47620839e-01 1.60006452e+00
7.84768701e-01 4.10120606e-01 4.58267391e-01 -1.69272572e-01
2.83897281e-01 1.82792038e-01 -2.72983342e-01 -9.89563465e-02
3.87077659e-01 1.13872898e+00 -3.65091741e-01 -1.97444469e-01
-4.55784500e-01 2.71858394e-01 -2.46456996e-01 2.34497011e-01
-7.00108647e-01 -1.05722558e+00 8.09038043e-01 -8.02536666e-01
5.20898938e-01 -1.02863312e-01 1.29581690e-01 -8.41721117e-01
-3.06857467e-01 -9.38746870e-01 5.48380554e-01 -5.77549636e-01
-1.59293091e+00 1.04106700e+00 -2.68189967e-01 -1.86877537e+00
-1.24220788e-01 -7.04146564e-01 -7.16759384e-01 9.77153540e-01
-1.82574093e+00 -1.03489316e+00 -5.13938606e-01 6.78140163e-01
5.30543745e-01 -6.68779075e-01 1.26730752e+00 5.28889060e-01
-3.82719077e-02 7.77891576e-01 5.85249603e-01 8.23923290e-01
4.96257126e-01 -7.38924265e-01 1.57533418e-02 6.49242163e-01
-9.89559442e-02 -1.55949906e-01 6.03133619e-01 -3.60525548e-01
-1.39606524e+00 -1.27211094e+00 3.05543661e-01 4.15824115e-01
8.85348797e-01 -1.08463811e-02 -6.14824176e-01 4.10202712e-01
2.70506404e-02 2.70319790e-01 1.17024767e+00 -7.31210768e-01
-3.96899581e-01 -2.38217160e-01 -1.15544558e+00 -3.81045379e-02
8.80362839e-02 -2.67908126e-01 -7.85793185e-01 2.99069762e-01
6.49975240e-01 -4.80210662e-01 -1.04641616e+00 8.21432948e-01
7.18233168e-01 -5.37734032e-01 9.16040719e-01 -2.22208261e-01
-1.06271423e-01 -6.43782377e-01 -8.83887351e-01 -1.75858891e+00
-3.81730556e-01 -6.12866618e-02 4.64352638e-01 4.57404554e-01
4.59721148e-01 -7.25296021e-01 2.87554860e-01 -3.96944314e-01
-8.51028502e-01 -2.56494641e-01 -1.40579355e+00 -8.44650924e-01
-1.81407966e-02 -5.05353332e-01 4.70192373e-01 8.48315060e-01
-1.81653261e-01 5.62282391e-02 -8.85832250e-01 9.39778924e-01
7.61283398e-01 3.63487273e-01 4.13555264e-01 -1.53372002e+00
-4.62262630e-01 -1.87192082e-01 -8.32777143e-01 -1.08063495e+00
5.10284975e-02 -6.14761293e-01 7.54292309e-01 -1.33062708e+00
-3.61593634e-01 -3.61620672e-02 -5.42063832e-01 7.39944339e-01
3.14765126e-01 6.10118806e-01 -5.91461249e-02 8.05108342e-03
-2.87258625e-01 7.80675232e-01 8.55372965e-01 -7.68164337e-01
1.63061529e-01 3.16154420e-01 -3.92878622e-01 6.54703796e-01
7.56703138e-01 -6.38041496e-01 1.16956078e-01 -5.90882957e-01
-2.30474651e-01 3.63373190e-01 2.01336920e-01 -1.29474449e+00
6.12877190e-01 -2.38106549e-01 3.04876328e-01 -5.92322946e-01
7.55907476e-01 -6.96222544e-01 1.39107570e-01 7.51533091e-01
-2.77855337e-01 -1.34168491e-01 4.69440997e-01 7.19566822e-01
-8.23056877e-01 -5.13776779e-01 9.81076360e-01 -3.14176649e-01
-1.19686687e+00 5.10945357e-02 -9.86140549e-01 -2.12229893e-01
7.72647440e-01 -3.45947444e-01 -5.06366134e-01 -4.31011349e-01
-8.08941424e-01 -8.92696530e-02 -5.41091561e-01 4.91985291e-01
1.27976322e+00 -1.10533690e+00 -1.27416360e+00 4.14768815e-01
3.03591918e-02 -2.97330767e-01 4.78672624e-01 4.93034691e-01
-6.30532801e-01 -1.74409285e-01 -4.21274245e-01 -4.65824664e-01
-1.49485147e+00 -3.84480566e-01 9.43633378e-01 3.60084951e-01
-2.13480741e-01 1.20709908e+00 9.49402004e-02 -6.72399461e-01
5.04477561e-01 -2.75110990e-01 -6.85650289e-01 2.67775059e-01
8.77606034e-01 1.26766920e-01 -1.03007182e-01 -9.21967030e-01
-2.73190051e-01 8.87186348e-01 2.14803621e-01 6.10510856e-02
1.61285949e+00 3.49985749e-01 -3.31821889e-01 4.30059731e-01
1.37690687e+00 -2.25971028e-01 -1.06840444e+00 -2.21088827e-01
-3.82491678e-01 -2.17757538e-01 -6.38831109e-02 -9.24300969e-01
-1.09336662e+00 6.45948172e-01 8.80380929e-01 4.38874424e-01
1.25322568e+00 -2.87309080e-01 5.70046663e-01 7.52680600e-01
4.12319690e-01 -5.95719218e-01 2.42660284e-01 1.11709309e+00
8.87101710e-01 -1.05060267e+00 -3.87829602e-01 -1.73881333e-02
-5.18021047e-01 1.57289743e+00 5.82043350e-01 -1.87069327e-02
9.78471100e-01 8.90230477e-01 9.31470931e-01 4.85202335e-02
-2.58461028e-01 -1.67176232e-01 5.60197942e-02 6.97798729e-01
8.66249651e-02 7.80629665e-02 3.45754623e-01 8.46589625e-01
-4.32361096e-01 -7.31498122e-01 8.69082212e-01 9.78349328e-01
-1.15079272e+00 -4.86406177e-01 -7.81650484e-01 6.13144934e-01
-7.10535169e-01 -3.36618304e-01 -2.83249523e-02 2.72767097e-01
4.98978235e-02 1.29177535e+00 4.03723903e-02 -9.79748309e-01
6.82359993e-01 -3.33707899e-01 8.44459981e-03 -3.74504179e-01
-5.95834494e-01 1.55027762e-01 4.42159874e-03 1.80013075e-01
-6.68789387e-01 -3.57842118e-01 -1.24192488e+00 4.18996334e-01
-5.55662692e-01 6.35987937e-01 7.56293356e-01 8.49231184e-01
-2.52523571e-01 4.83253360e-01 8.16506386e-01 -1.31617296e+00
-9.75484669e-01 -1.49069023e+00 -8.64824951e-01 5.05720265e-02
9.65001523e-01 -4.64411646e-01 -9.47879612e-01 -7.45223090e-02] | [8.538020133972168, -1.1572321653366089] |
2d2478a8-71d6-46cc-bd5a-08d642a40ae0 | building-powerful-and-equivariant-graph | 2006.15107 | null | https://arxiv.org/abs/2006.15107v3 | https://arxiv.org/pdf/2006.15107v3.pdf | Building powerful and equivariant graph neural networks with structural message-passing | Message-passing has proved to be an effective way to design graph neural networks, as it is able to leverage both permutation equivariance and an inductive bias towards learning local structures in order to achieve good generalization. However, current message-passing architectures have a limited representation power and fail to learn basic topological properties of graphs. We address this problem and propose a powerful and equivariant message-passing framework based on two ideas: first, we propagate a one-hot encoding of the nodes, in addition to the features, in order to learn a local context matrix around each node. This matrix contains rich local information about both features and topology and can eventually be pooled to build node representations. Second, we propose methods for the parametrization of the message and update functions that ensure permutation equivariance. Having a representation that is independent of the specific choice of the one-hot encoding permits inductive reasoning and leads to better generalization properties. Experimentally, our model can predict various graph topological properties on synthetic data more accurately than previous methods and achieves state-of-the-art results on molecular graph regression on the ZINC dataset. | ['Andreas Loukas', 'Pascal Frossard', 'Clement Vignac'] | 2020-06-26 | null | http://proceedings.neurips.cc/paper/2020/hash/a32d7eeaae19821fd9ce317f3ce952a7-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/a32d7eeaae19821fd9ce317f3ce952a7-Paper.pdf | neurips-2020-12 | ['graph-regression'] | ['graphs'] | [ 3.31617445e-01 7.16138929e-02 -3.80826384e-01 -5.63945293e-01
-3.27097118e-01 -7.01003194e-01 7.77106464e-01 8.28944027e-01
-2.51249135e-01 6.84748054e-01 1.29022971e-01 -4.29752469e-01
-4.29445416e-01 -1.05706322e+00 -1.08595300e+00 -8.29066813e-01
-5.65218687e-01 5.28138638e-01 3.10834765e-01 -5.11751711e-01
3.33398670e-01 6.39866173e-01 -1.22508597e+00 2.60377437e-01
7.87027121e-01 5.96096992e-01 1.00369819e-01 7.93049753e-01
4.18711305e-02 6.63390815e-01 1.82543264e-03 -2.24573091e-01
1.84294637e-02 -3.94604772e-01 -1.07697940e+00 -5.01785219e-01
5.20615995e-01 9.45406705e-02 -4.35556263e-01 7.97066271e-01
3.23972434e-01 7.70949572e-02 8.20448101e-01 -8.66136789e-01
-3.37111473e-01 7.38663733e-01 -1.93620458e-01 2.02100739e-01
6.17975183e-02 7.25784898e-02 1.45982897e+00 -4.31591898e-01
1.05988050e+00 1.00895250e+00 7.59931684e-01 3.08709383e-01
-1.80848825e+00 -3.85504812e-01 2.07932696e-01 2.01650441e-01
-1.19390512e+00 -2.68443942e-01 8.87565970e-01 -3.49133193e-01
7.61668265e-01 2.71004349e-01 7.82829940e-01 9.52253819e-01
5.64571559e-01 3.01311761e-01 7.91024804e-01 -3.01998258e-01
2.75724560e-01 -2.03774631e-01 2.03697637e-01 1.11561775e+00
2.56057978e-01 -1.50373042e-01 -4.95348394e-01 -2.93915331e-01
7.43793428e-01 3.18548712e-03 -1.83934659e-01 -9.05905604e-01
-1.22887659e+00 9.85351264e-01 1.02512121e+00 4.47883546e-01
-2.56951988e-01 5.09065926e-01 2.70513892e-01 4.12197858e-01
1.56640336e-01 9.22192156e-01 -4.02996868e-01 1.69783503e-01
-4.77615118e-01 2.56204098e-01 8.05604041e-01 3.89628738e-01
1.12548018e+00 -3.09888124e-01 -1.12454467e-01 6.37130678e-01
2.12524161e-01 5.78283556e-02 2.15282932e-01 -4.96136665e-01
3.36796820e-01 7.52570271e-01 -4.39308912e-01 -1.29098308e+00
-7.79126525e-01 -8.01728368e-01 -8.57164979e-01 3.15128937e-02
4.48802382e-01 1.73626959e-01 -8.73033345e-01 2.32539344e+00
3.21337402e-01 -3.72188799e-02 -2.26428479e-01 5.35829306e-01
5.29130340e-01 5.60510695e-01 1.30259246e-01 2.31029391e-01
1.10211384e+00 -3.80510002e-01 3.30465436e-02 -2.48248726e-01
9.67152536e-01 -3.10749352e-01 7.54014254e-01 1.41172737e-01
-9.43377316e-01 -1.06206477e-01 -1.28609860e+00 -9.43315923e-02
-6.08579636e-01 -2.31792226e-01 1.03020036e+00 4.45017755e-01
-9.89988863e-01 1.13710070e+00 -8.38040352e-01 -3.80565107e-01
4.34454411e-01 7.53086925e-01 -5.05601346e-01 -5.15948758e-02
-1.01284635e+00 6.50627911e-01 4.73616004e-01 -9.56293792e-02
-7.71153033e-01 -8.00362051e-01 -6.81596160e-01 5.72494119e-02
1.68993413e-01 -9.19227958e-01 6.72027051e-01 -6.91817462e-01
-1.32059240e+00 5.64525008e-01 7.87331536e-02 -5.01927137e-01
2.29343385e-01 5.59088826e-01 9.03890431e-02 1.49084032e-01
-2.39291519e-01 7.87893772e-01 6.19182110e-01 -9.28953588e-01
-2.43809327e-01 -4.25735891e-01 2.63971329e-01 3.71607742e-03
-3.94336492e-01 -6.84582889e-01 -2.44618863e-01 -5.22519469e-01
4.41685915e-01 -1.01824927e+00 -3.26299220e-01 3.74118313e-02
-4.18068171e-01 5.09186648e-02 4.91715163e-01 -2.64173388e-01
1.02376103e+00 -1.93727148e+00 6.20053828e-01 6.84187770e-01
6.17892742e-01 6.28285808e-03 -2.56187409e-01 6.54047549e-01
-1.90978423e-01 1.35123491e-01 -3.27815711e-01 1.02175348e-01
-1.24976605e-01 2.65651256e-01 -1.29066840e-01 5.75344622e-01
4.70055163e-01 1.05001342e+00 -9.54315543e-01 -1.45926550e-01
1.26609609e-01 6.60537899e-01 -9.19700086e-01 -9.06888172e-02
-5.35416186e-01 3.79537612e-01 -4.98482734e-01 6.85187802e-02
4.74141806e-01 -6.18264377e-01 7.67298996e-01 -2.80826300e-01
8.13975483e-02 5.88630497e-01 -1.05318880e+00 1.59194601e+00
-3.68040383e-01 4.68476772e-01 -3.78592126e-02 -1.22219908e+00
9.45688903e-01 -3.10185291e-02 5.11006832e-01 -5.71065903e-01
2.14859337e-01 7.76687413e-02 1.46591216e-01 -7.72585347e-02
2.42673039e-01 -7.19421878e-02 5.77670038e-02 4.58017319e-01
2.93164879e-01 1.11873792e-02 2.30051056e-01 4.29247320e-01
1.32982266e+00 -1.14256643e-01 3.26246023e-02 -5.85542560e-01
5.60024798e-01 -3.21364880e-01 4.93932486e-01 7.58188605e-01
2.68527746e-01 4.01477456e-01 9.87487912e-01 -6.74407065e-01
-1.15770221e+00 -9.43334579e-01 -1.49031088e-01 1.27164614e+00
1.26946092e-01 -7.53591359e-01 -6.37907267e-01 -5.35818815e-01
1.26487575e-02 2.28203624e-01 -9.60012197e-01 -6.07906818e-01
-6.75209820e-01 -9.19747412e-01 4.75878865e-01 4.76475745e-01
9.50682759e-02 -5.52315354e-01 -1.38100669e-01 3.96544725e-01
1.16044559e-01 -8.70445073e-01 -3.35199952e-01 6.47450864e-01
-9.72371221e-01 -1.25620675e+00 -2.46900454e-01 -7.88704216e-01
9.08268273e-01 4.40137349e-02 1.05651319e+00 3.97722244e-01
-3.29403281e-01 8.91895741e-02 -6.54585138e-02 1.45997165e-03
-4.95445549e-01 7.24825740e-01 -1.81466907e-01 1.45937249e-01
-9.33739170e-02 -9.34429228e-01 -5.99134386e-01 1.80921823e-01
-1.02207172e+00 2.68778116e-01 6.39261127e-01 9.97073233e-01
4.44210947e-01 -8.13122019e-02 4.71665800e-01 -1.08164763e+00
5.40706098e-01 -4.14176613e-01 -5.24597824e-01 2.54749507e-01
-4.38263685e-01 7.85894156e-01 9.71267164e-01 -3.41278285e-01
-4.79046077e-01 1.76666468e-01 -9.13642198e-02 2.23416418e-01
1.53310835e-01 6.01920664e-01 4.13118191e-02 -5.66391826e-01
7.19106376e-01 1.05864316e-01 1.14744574e-01 -3.61805826e-01
4.78013098e-01 -8.48091692e-02 3.28174651e-01 -8.86777520e-01
6.99513793e-01 4.34153199e-01 6.86585844e-01 -1.08042777e+00
-2.06555203e-01 -8.39411840e-02 -9.31189477e-01 6.16752170e-02
6.44001544e-01 -5.20300984e-01 -1.01034224e+00 4.08543766e-01
-8.02110255e-01 -2.85985887e-01 -4.03727219e-03 2.96853095e-01
-4.96593863e-01 1.42987728e-01 -6.11575544e-01 -3.52789640e-01
-9.01034474e-02 -1.23323965e+00 7.14819252e-01 -7.43780062e-02
-2.81148846e-03 -1.31373250e+00 1.40328169e-01 -2.26198792e-01
5.81860721e-01 4.73285347e-01 1.57179666e+00 -6.68903530e-01
-8.74195457e-01 -1.24532826e-01 -1.20608702e-01 -1.65875778e-01
-1.95213586e-01 1.04138881e-01 -6.14334285e-01 -3.83519948e-01
-5.50073683e-01 -2.33162090e-01 1.21159720e+00 2.17569888e-01
1.10892999e+00 -3.63150477e-01 -4.94562536e-01 7.09537983e-01
1.40744996e+00 -4.46984202e-01 4.82208312e-01 8.29840973e-02
8.16300452e-01 2.99667954e-01 -2.99090713e-01 1.86083049e-01
5.34835815e-01 7.29686201e-01 5.18510580e-01 -5.90752997e-02
3.71235162e-02 -5.73548853e-01 9.90613922e-02 8.52734208e-01
-4.62800004e-02 3.60063128e-02 -9.79519427e-01 1.62528753e-01
-1.74135804e+00 -7.37104595e-01 4.66729281e-03 2.27884698e+00
9.69524324e-01 3.97594780e-01 2.27102712e-01 1.41555414e-01
3.90578926e-01 4.41208124e-01 -3.74329358e-01 -2.91501641e-01
-1.56239256e-01 2.87622094e-01 7.70293236e-01 6.11707747e-01
-8.67959619e-01 8.66826653e-01 6.80751085e+00 5.89665830e-01
-1.42166805e+00 -2.18080223e-01 5.86655259e-01 2.04159454e-01
-7.59571373e-01 1.88854948e-01 -6.04170740e-01 1.55020714e-01
9.46140468e-01 1.74668908e-01 6.17764711e-01 4.01147038e-01
-2.26256266e-01 2.75965631e-01 -1.22791064e+00 4.22229409e-01
-1.66632578e-01 -1.88941205e+00 2.93520600e-01 1.72796175e-01
5.18111289e-01 2.55208194e-01 -8.93549919e-02 1.12471776e-03
3.55650455e-01 -1.04274023e+00 4.58296716e-01 6.33308053e-01
5.00947416e-01 -8.14416289e-01 3.91934782e-01 2.39569142e-01
-1.25848222e+00 -1.38434470e-01 -4.88398314e-01 -1.79314800e-02
-3.12253743e-01 4.98999655e-01 -1.03525853e+00 4.19067204e-01
1.94107369e-01 6.97923303e-01 -8.37652683e-01 9.70159173e-01
-1.80075914e-01 4.53671366e-01 -5.08165777e-01 -3.44711840e-01
4.04943854e-01 -1.71886012e-01 4.82669890e-01 1.09333968e+00
-1.93090767e-01 -1.21362910e-01 2.64028430e-01 9.33009207e-01
-4.20075566e-01 2.56944180e-01 -6.34867847e-01 -2.16375634e-01
3.43522340e-01 1.08397400e+00 -8.59307528e-01 -4.29041833e-02
-3.13957632e-02 4.28064495e-01 9.11567032e-01 3.21559608e-01
-5.08886158e-01 -4.40093398e-01 7.15055943e-01 2.37122536e-01
4.69652295e-01 -6.75976217e-01 -2.19943881e-01 -1.11291683e+00
-3.67630832e-02 -7.50376523e-01 3.70782644e-01 -3.19168240e-01
-9.37742472e-01 3.58715892e-01 -2.19303250e-01 -5.49183309e-01
-4.47733216e-02 -8.57958615e-01 -7.21700370e-01 5.08841515e-01
-1.40216506e+00 -1.24538469e+00 -6.21913373e-02 2.57835180e-01
-1.26332477e-01 1.87389310e-02 8.63423288e-01 9.91713330e-02
-5.74620068e-01 5.39607525e-01 3.23851079e-01 2.60292310e-02
4.48627800e-01 -1.32223332e+00 5.09723186e-01 5.92218399e-01
4.36754078e-01 9.24197733e-01 6.61624253e-01 -4.62547272e-01
-2.03064156e+00 -9.34041142e-01 6.79108441e-01 -5.27310312e-01
7.35314727e-01 -7.52844214e-01 -9.83809173e-01 6.01352572e-01
-3.28828752e-01 1.07316762e-01 4.64812994e-01 4.13880348e-01
-7.76756048e-01 -3.48342150e-01 -8.57021987e-01 5.95071375e-01
1.05429709e+00 -6.76805615e-01 -1.64805129e-01 3.40495288e-01
6.64518833e-01 -2.01110423e-01 -1.05903590e+00 3.41557056e-01
5.76136708e-01 -7.19068289e-01 1.04073703e+00 -1.01662397e+00
2.96918333e-01 -2.68660486e-01 -5.22330590e-02 -1.22855914e+00
-7.70582795e-01 -6.72942698e-01 -1.84772220e-02 9.08499241e-01
8.60551357e-01 -9.20365930e-01 7.44069576e-01 2.93390721e-01
1.04275711e-01 -9.90421355e-01 -7.34639466e-01 -4.83493507e-01
3.49373251e-01 -1.86616972e-01 7.28544354e-01 9.08062994e-01
6.71018437e-02 5.95976949e-01 -2.60008126e-01 4.36223857e-02
6.01450145e-01 5.06805107e-02 7.42604613e-01 -1.44231904e+00
-2.73094207e-01 -7.58627355e-01 -7.93462515e-01 -1.00206399e+00
2.28022829e-01 -1.48058403e+00 -1.93051606e-01 -1.25946569e+00
2.10895106e-01 -7.91724801e-01 -4.35237795e-01 5.70828438e-01
3.83760817e-02 2.54506528e-01 -1.07986636e-01 -8.75861868e-02
-5.18921673e-01 4.98518795e-01 1.05020988e+00 -1.52855724e-01
-2.20219120e-01 -3.26108098e-01 -8.39681566e-01 3.45873684e-01
7.34766781e-01 -4.40799087e-01 -3.35614115e-01 -4.07948971e-01
5.98743796e-01 -2.32613027e-01 4.68245625e-01 -9.52341795e-01
3.94432068e-01 -1.83838099e-01 5.62792301e-01 -1.70573071e-01
2.80501574e-01 -5.05706131e-01 8.40936154e-02 5.58501661e-01
-7.01812148e-01 -2.02830255e-01 -5.31939119e-02 7.07342744e-01
1.95521578e-01 -9.59811956e-02 9.14808631e-01 4.58276086e-02
-2.85070300e-01 5.98379612e-01 -9.66150463e-02 -1.61128619e-03
5.98828316e-01 1.39600411e-01 -4.20296222e-01 -2.54671663e-01
-4.88365084e-01 1.28302634e-01 6.79538488e-01 3.65480036e-01
2.58519530e-01 -1.19631231e+00 -5.79371452e-01 6.27557755e-01
1.57721996e-01 -3.41494977e-01 -1.01469494e-01 7.79814780e-01
-6.06222928e-01 4.98486787e-01 -3.22830051e-01 -7.31089354e-01
-1.06201518e+00 4.64899987e-01 4.43607360e-01 -2.50453681e-01
-7.04765260e-01 5.65575302e-01 3.21618356e-02 -5.55973411e-01
3.92945223e-02 -3.85907859e-01 -6.22884482e-02 -1.15568638e-01
3.10110331e-01 8.39104727e-02 3.26730996e-01 -5.82428932e-01
-3.74793977e-01 6.90448821e-01 -2.09747642e-01 2.29624271e-01
1.55320811e+00 2.57648349e-01 -2.43960604e-01 2.87446856e-01
1.36228144e+00 -1.78018093e-01 -9.83670712e-01 -2.81215966e-01
7.32829198e-02 -3.40644866e-01 6.97031468e-02 -4.77232128e-01
-9.20690060e-01 7.97887981e-01 1.30290672e-01 3.00801367e-01
5.72006822e-01 6.02722354e-02 4.68744516e-01 7.37207651e-01
4.00615871e-01 -8.90033424e-01 8.00276250e-02 5.98528147e-01
5.94668269e-01 -9.60479915e-01 2.03659981e-02 -4.12195176e-01
-1.02892928e-02 1.34498906e+00 1.87722683e-01 -2.42952451e-01
6.27485633e-01 -5.59462309e-02 -4.69079256e-01 -3.61443490e-01
-1.01380301e+00 3.81579809e-02 4.61066276e-01 4.47774291e-01
6.01600707e-01 1.51807606e-01 -2.61426300e-01 1.33329660e-01
-1.66172177e-01 -3.76338869e-01 3.82288098e-01 8.64189446e-01
-6.10258222e-01 -1.42556036e+00 7.47885257e-02 5.28296530e-01
-2.85339683e-01 7.43054673e-02 -5.89104831e-01 6.52284861e-01
-2.09263086e-01 3.64280283e-01 -9.70164686e-02 -4.36657757e-01
7.31312558e-02 1.67723134e-01 7.53963530e-01 -4.40309405e-01
-5.15252352e-01 -3.69512111e-01 1.55544907e-01 -5.08752465e-01
4.81960624e-02 -3.91062260e-01 -1.17403233e+00 -6.20840728e-01
-2.07574159e-01 2.27462441e-01 8.86304617e-01 9.47594762e-01
5.84121644e-01 5.73005021e-01 6.19198084e-01 -8.83091211e-01
-4.37162459e-01 -4.85250771e-01 -4.16648895e-01 5.45250237e-01
4.67080176e-01 -6.43846571e-01 -1.99259430e-01 -3.49813014e-01] | [6.742607116699219, 6.043454170227051] |
548cee94-8b2d-4970-b847-a61d24f27590 | syntax-controlled-knowledge-graph-to-text | 2207.00719 | null | https://arxiv.org/abs/2207.00719v1 | https://arxiv.org/pdf/2207.00719v1.pdf | Syntax Controlled Knowledge Graph-to-Text Generation with Order and Semantic Consistency | The knowledge graph (KG) stores a large amount of structural knowledge, while it is not easy for direct human understanding. Knowledge graph-to-text (KG-to-text) generation aims to generate easy-to-understand sentences from the KG, and at the same time, maintains semantic consistency between generated sentences and the KG. Existing KG-to-text generation methods phrase this task as a sequence-to-sequence generation task with linearized KG as input and consider the consistency issue of the generated texts and KG through a simple selection between decoded sentence word and KG node word at each time step. However, the linearized KG order is commonly obtained through a heuristic search without data-driven optimization. In this paper, we optimize the knowledge description order prediction under the order supervision extracted from the caption and further enhance the consistency of the generated sentences and KG through syntactic and semantic regularization. We incorporate the Part-of-Speech (POS) syntactic tags to constrain the positions to copy words from the KG and employ a semantic context scoring function to evaluate the semantic fitness for each word in its local context when decoding each word in the generated sentence. Extensive experiments are conducted on two datasets, WebNLG and DART, and achieve state-of-the-art performances. | ['Huijuan Xu', 'Fengyu Zhou', 'Chongfeng Fan', 'Jin Liu'] | 2022-07-02 | null | https://aclanthology.org/2022.findings-naacl.95 | https://aclanthology.org/2022.findings-naacl.95.pdf | findings-naacl-2022-7 | ['kg-to-text'] | ['natural-language-processing'] | [ 4.09248859e-01 7.53962398e-01 -1.67189062e-01 -4.91000384e-01
-8.40514183e-01 -3.93467933e-01 3.02767247e-01 1.43578678e-01
-1.68462902e-01 9.97783124e-01 4.77559239e-01 -1.68310314e-01
5.44255413e-02 -1.02457559e+00 -1.01569521e+00 -5.68790793e-01
4.49309528e-01 5.70959449e-01 7.43834376e-02 -2.45489299e-01
2.33088687e-01 -3.07714731e-01 -1.23580825e+00 6.28991425e-01
1.50607383e+00 7.90151119e-01 8.92777085e-01 5.59986770e-01
-6.08642399e-01 7.64535129e-01 -7.06018686e-01 -8.74054909e-01
-1.74692616e-01 -8.68170023e-01 -9.78423238e-01 2.12967560e-01
5.43517470e-02 2.02476978e-01 -3.65611464e-02 1.09868360e+00
4.95880485e-01 3.19336094e-02 5.73253334e-01 -1.02305651e+00
-1.04415894e+00 1.28420222e+00 -2.05675855e-01 -9.03203487e-02
5.05130887e-01 8.83126184e-02 1.19274032e+00 -8.57988000e-01
9.19067979e-01 1.29577541e+00 1.86714709e-01 5.28521895e-01
-9.14712012e-01 -3.36861819e-01 2.76923656e-01 5.27641118e-01
-1.49221969e+00 -3.14258814e-01 7.26031303e-01 -1.27984047e-01
1.36942041e+00 6.14952073e-02 6.29830658e-01 1.01709580e+00
1.37798861e-01 1.02048910e+00 5.09050429e-01 -7.17591763e-01
2.23001525e-01 6.12185374e-02 -9.54847634e-02 9.75653291e-01
2.33860806e-01 -3.67916435e-01 -9.06505704e-01 5.17726243e-02
2.87366688e-01 -7.17414021e-01 -4.09340590e-01 -6.64573610e-02
-1.13140523e+00 8.09639990e-01 2.11498812e-01 1.68382168e-01
-3.67417216e-01 1.13143764e-01 3.25627744e-01 2.49755895e-03
3.35928172e-01 2.97787964e-01 -5.95644474e-01 -1.58602789e-01
-7.63376892e-01 3.34172279e-01 9.07388806e-01 1.46402347e+00
8.06864083e-01 2.96071749e-02 -6.10059440e-01 6.84781194e-01
4.17055070e-01 6.62842751e-01 7.08976090e-01 -3.55522007e-01
1.06468272e+00 6.40699804e-01 -2.58995444e-01 -1.07609820e+00
-5.49012460e-02 -4.68973905e-01 -7.02317774e-01 -7.02117562e-01
-5.55210523e-02 -2.37990975e-01 -9.87765133e-01 1.74377429e+00
3.39952171e-01 6.74320534e-02 5.84594488e-01 6.57095015e-01
9.70568597e-01 9.48430896e-01 1.21963009e-01 -4.29494530e-01
1.22039938e+00 -1.19497120e+00 -8.01920116e-01 -6.60545051e-01
1.02014232e+00 -5.63275218e-01 1.09527028e+00 4.56691384e-02
-1.00361836e+00 -3.88224602e-01 -9.52616751e-01 -8.32177550e-02
-2.80608773e-01 3.58423918e-01 1.87797099e-01 2.91742921e-01
-8.65927756e-01 4.74797100e-01 -5.01386166e-01 -1.52241424e-01
9.75511000e-02 -6.19600900e-03 -5.86812384e-02 -1.70863718e-01
-1.70412004e+00 7.79704750e-01 1.16844630e+00 2.95430690e-01
-4.56612796e-01 -4.56583768e-01 -1.22448647e+00 4.36454080e-02
6.30192280e-01 -1.22381687e+00 1.09546804e+00 -9.78416681e-01
-1.57395828e+00 5.79851568e-01 -4.92967069e-01 -3.83020937e-01
2.49686390e-01 -6.68490902e-02 -4.13715810e-01 7.14236125e-02
3.49688172e-01 6.96426809e-01 8.97301137e-01 -1.10632348e+00
-6.25145555e-01 -1.91087693e-01 -2.86361098e-01 6.30048752e-01
-1.11964300e-01 -1.88901767e-01 -8.26256633e-01 -7.69078135e-01
8.88360813e-02 -7.18804181e-01 -1.17561080e-01 -7.67634392e-01
-8.83534610e-01 -3.60590786e-01 2.85286397e-01 -1.08640873e+00
1.47175753e+00 -1.84481180e+00 3.80719125e-01 2.41929576e-01
-3.31090063e-01 1.94800645e-01 -2.59647608e-01 5.59705794e-01
3.74843538e-01 1.94163412e-01 -2.09778637e-01 -2.74242133e-01
1.15598887e-01 3.41962099e-01 -3.06640327e-01 -3.02265704e-01
2.55934894e-01 1.29625022e+00 -1.16977942e+00 -8.29065681e-01
-1.82843834e-01 1.07799053e-01 -5.23845255e-01 3.12881768e-01
-7.05463409e-01 1.56714529e-01 -7.44695663e-01 2.70301282e-01
2.54895836e-01 -4.34675634e-01 5.08017421e-01 -2.26204485e-01
4.29474115e-01 6.02847993e-01 -8.52254450e-01 1.80662882e+00
-5.36116183e-01 4.25346904e-02 -4.82475698e-01 -9.03171062e-01
9.74475563e-01 1.76826328e-01 -2.62650013e-01 -6.90968335e-01
-5.66925332e-02 2.42929578e-01 -7.76780397e-02 -6.14700973e-01
8.20444167e-01 -1.71329245e-01 -3.27912092e-01 2.07435057e-01
4.30456102e-02 -2.92524576e-01 6.04799390e-01 4.99325395e-01
9.15370822e-01 2.30888575e-01 2.86362410e-01 1.06392197e-01
5.36785662e-01 4.56609987e-02 6.43068194e-01 7.64386237e-01
5.54178238e-01 4.84612674e-01 5.60487747e-01 1.88260198e-01
-8.19269061e-01 -7.81988978e-01 4.15824205e-01 7.63558149e-01
1.25924766e-01 -6.84575319e-01 -9.86833036e-01 -1.01071751e+00
-8.20629373e-02 1.48168397e+00 -3.77003580e-01 -5.19725323e-01
-5.32943249e-01 -6.29190385e-01 6.15530431e-01 5.12290716e-01
3.93293500e-01 -1.45029593e+00 1.24284029e-01 5.27828753e-01
-6.30051196e-01 -1.40708601e+00 -7.53330350e-01 -1.52674094e-01
-4.97125655e-01 -8.34745526e-01 -4.44002867e-01 -1.04694200e+00
1.07314932e+00 -1.97513208e-01 1.16139436e+00 1.07714711e-02
7.01078447e-03 -6.58814684e-02 -8.47651243e-01 -1.60492837e-01
-8.50600541e-01 2.92056143e-01 -2.22929403e-01 -2.58486941e-02
3.31151634e-02 -1.54588580e-01 -1.86499491e-01 5.47935553e-02
-9.47915852e-01 7.24620104e-01 7.31542349e-01 7.76350319e-01
8.06709707e-01 2.52422720e-01 8.40835035e-01 -8.41507733e-01
8.26375425e-01 -2.82811493e-01 -2.57474482e-01 7.11802661e-01
-5.81191421e-01 4.11970139e-01 9.29942846e-01 -7.91880563e-02
-1.22493589e+00 -1.04370683e-01 -2.24721447e-01 -1.62033141e-01
2.81500310e-01 1.06159163e+00 -6.49133623e-01 4.78125751e-01
3.12530428e-01 9.06900227e-01 -3.38647395e-01 -1.96722627e-01
9.04376030e-01 7.26796329e-01 4.65535522e-01 -6.61483228e-01
6.71981156e-01 -1.39949158e-01 -2.95745194e-01 -6.45245731e-01
-1.23650134e+00 -7.95980841e-02 -4.80834275e-01 -5.15973046e-02
8.79845738e-01 -7.86429644e-01 -2.19272956e-01 4.28079635e-01
-1.43159735e+00 -3.23049873e-01 -2.34548017e-01 1.25720188e-01
-5.27031362e-01 6.35858893e-01 -4.37077433e-01 -5.97932458e-01
-7.07843900e-01 -7.79130161e-01 1.17848670e+00 2.10107937e-01
-1.25695139e-01 -9.70901012e-01 -3.19345236e-01 6.00333035e-01
-7.27123246e-02 -1.65617749e-01 1.24195075e+00 -5.53526938e-01
-5.65360725e-01 -3.55343446e-02 -1.48528725e-01 5.43002427e-01
2.24297166e-01 -2.12721109e-01 -3.61554623e-01 -1.49908915e-01
-4.84643668e-01 -3.94084007e-01 8.63726854e-01 1.10902295e-01
1.03510010e+00 -8.64154100e-01 -3.09818625e-01 3.22192907e-01
1.28285873e+00 4.99475002e-02 7.47213423e-01 4.66720872e-02
9.40385342e-01 5.70648789e-01 7.57146001e-01 3.12678427e-01
6.96414709e-01 6.76805377e-01 1.30525734e-02 4.95957732e-01
-2.60143876e-01 -9.24383998e-01 6.27323866e-01 1.36127412e+00
5.28614342e-01 -8.40903938e-01 -7.68065214e-01 6.10388160e-01
-2.01100636e+00 -7.69711256e-01 -2.45781522e-02 2.01947927e+00
1.26748347e+00 2.08844855e-01 -4.94733453e-01 -1.62907556e-01
9.31473851e-01 7.13697001e-02 -2.66655236e-01 -2.79100567e-01
-3.83499652e-01 -1.66448981e-01 2.66642243e-01 6.82054222e-01
-5.50107479e-01 1.53695858e+00 5.29205656e+00 1.37646294e+00
-8.28430772e-01 -1.75464019e-01 5.35566151e-01 -6.47123680e-02
-6.15255654e-01 1.65067226e-01 -1.33064902e+00 7.65532494e-01
7.50481486e-01 -6.73858643e-01 3.64225984e-01 6.82294607e-01
4.03671831e-01 -7.42219016e-03 -8.08469296e-01 8.00792158e-01
3.89723331e-01 -1.39846087e+00 6.76252782e-01 -4.16448951e-01
8.06666613e-01 -3.97162974e-01 -4.17176723e-01 3.70384604e-01
2.42861509e-01 -8.29400659e-01 1.01062679e+00 5.99284708e-01
5.67678154e-01 -8.15405965e-01 7.01321959e-01 6.26985013e-01
-1.25250196e+00 2.14175165e-01 -4.83364552e-01 2.83853054e-01
6.89199328e-01 9.08060133e-01 -1.25917947e+00 1.09246743e+00
2.75435627e-01 6.93530440e-01 -5.20015776e-01 4.75482702e-01
-9.52787399e-01 5.19755363e-01 -7.09209889e-02 -5.11910558e-01
2.90158361e-01 -1.31687388e-01 6.34162664e-01 1.27764869e+00
5.26604712e-01 3.19952630e-02 3.59084547e-01 9.70106363e-01
-1.87128916e-01 4.89457875e-01 -3.42424184e-01 -3.82623345e-01
4.73362237e-01 1.05023503e+00 -5.76842904e-01 -5.48182964e-01
-1.47580519e-01 1.34108198e+00 6.41756415e-01 3.05110484e-01
-6.21576071e-01 -6.13097489e-01 2.42781654e-01 -4.62401323e-02
6.75908506e-01 -5.38570136e-02 -1.23375334e-01 -1.13772047e+00
4.92467076e-01 -6.69627726e-01 3.19449246e-01 -1.21167564e+00
-1.26624298e+00 5.64487159e-01 -8.00254345e-02 -8.02143395e-01
-5.76074958e-01 -1.58435881e-01 -4.07583952e-01 9.97951150e-01
-1.47363806e+00 -1.13393438e+00 -8.63780640e-03 2.19801694e-01
7.09822953e-01 -4.26310301e-02 5.80130577e-01 -1.33020133e-01
-6.71743929e-01 6.57505691e-01 9.64850411e-02 1.47065341e-01
2.38839895e-01 -1.16326284e+00 6.84593081e-01 8.81056488e-01
3.14083844e-02 4.28365648e-01 6.50646985e-01 -1.32666421e+00
-1.31842852e+00 -1.51482093e+00 1.45913661e+00 -1.12052649e-01
5.18625319e-01 -3.88598204e-01 -8.39285791e-01 5.35591841e-01
1.26938090e-01 -5.36516249e-01 4.74467248e-01 -1.90979674e-01
-4.89544589e-03 2.46548802e-01 -7.28351593e-01 7.74010956e-01
1.26764464e+00 -3.13979805e-01 -7.80844867e-01 5.13030946e-01
1.14751232e+00 -5.99766254e-01 -5.65654337e-01 2.55278170e-01
6.40536612e-03 -4.08853859e-01 5.24998069e-01 -6.80686116e-01
6.85131252e-01 -5.02530813e-01 8.82774964e-02 -1.68731570e+00
-2.58131325e-01 -7.16530621e-01 -2.25985765e-01 1.49705911e+00
9.56202149e-01 -2.58561879e-01 8.95337045e-01 3.95393997e-01
-4.69188124e-01 -8.34926009e-01 -8.26325417e-01 -1.08237946e+00
-2.50255436e-01 -4.02442634e-01 6.25998318e-01 6.03972316e-01
4.72194165e-01 8.33577812e-01 -2.71649688e-01 2.06660971e-01
5.15429914e-01 1.41623497e-01 5.20192146e-01 -5.71547568e-01
-4.10507858e-01 -2.04427004e-01 -2.03326166e-01 -1.25595284e+00
6.28503084e-01 -1.43141806e+00 5.14453590e-01 -2.01738906e+00
2.70517856e-01 -1.73707306e-01 2.03468442e-01 7.28895664e-01
-6.98940992e-01 -4.08289254e-01 1.36713907e-01 -3.13305631e-02
-7.82449067e-01 1.15404069e+00 1.45559120e+00 -1.82660833e-01
-2.53740728e-01 -4.38353300e-01 -9.30264950e-01 4.33496535e-01
6.74910188e-01 -4.61580992e-01 -7.75436521e-01 -5.85900724e-01
6.69678748e-01 2.46543184e-01 6.38581663e-02 -6.36546612e-01
1.92589447e-01 -2.88668096e-01 1.16988286e-01 -5.19484460e-01
-1.62624456e-02 -2.50314474e-01 7.29112029e-02 2.38945603e-01
-4.84195977e-01 -2.07136899e-01 9.45908874e-02 7.43658662e-01
-1.82935983e-01 -6.24289691e-01 4.24910247e-01 -1.99135274e-01
-8.60262871e-01 8.18094984e-02 -1.77735060e-01 4.98775810e-01
8.86065781e-01 -2.04341158e-01 -3.36010724e-01 -4.73015964e-01
-5.76374829e-01 5.30554414e-01 2.39549428e-01 6.00688994e-01
8.79392266e-01 -1.29011142e+00 -8.86956155e-01 1.34204060e-01
3.33123446e-01 2.81939149e-01 2.93155223e-01 3.00640613e-01
-3.15453380e-01 5.16771734e-01 4.13650393e-01 -1.77176148e-01
-1.19614708e+00 3.69492561e-01 1.45415202e-01 -5.77344477e-01
-5.72582245e-01 8.50921869e-01 6.30061552e-02 -4.69111353e-01
-5.60023747e-02 -2.44995803e-01 -1.23504758e-01 -2.15076968e-01
5.01944065e-01 1.18389595e-02 1.38574824e-01 -5.79818904e-01
-1.54887244e-01 2.66633034e-01 -3.54595095e-01 -1.65662631e-01
9.96065259e-01 -3.22593659e-01 -1.75352558e-01 1.99407451e-02
1.00428820e+00 -1.15311749e-01 -7.96442449e-01 -2.97644854e-01
1.35106504e-01 -2.41500542e-01 -1.14253439e-01 -9.87973809e-01
-9.15881097e-01 4.03577387e-01 -3.50539386e-01 -6.98585361e-02
8.16466808e-01 2.42087692e-01 1.17018688e+00 5.79242527e-01
3.01145911e-01 -1.40944481e+00 2.14612201e-01 7.98329890e-01
1.00070286e+00 -7.83540845e-01 -2.04594299e-01 -9.64422584e-01
-1.04347301e+00 7.67337143e-01 7.21101284e-01 4.63423789e-01
2.69352108e-01 -9.67827439e-02 -1.07135810e-01 5.35930917e-02
-9.17822719e-01 -2.04258546e-01 3.60122144e-01 6.32810891e-01
1.82192579e-01 1.26960659e-02 -5.10099530e-01 9.06973422e-01
-5.69373846e-01 4.49541137e-02 4.05348241e-01 8.89743209e-01
-6.08114004e-01 -1.20526958e+00 1.59618571e-01 5.52087069e-01
-1.81957886e-01 -5.93792617e-01 -5.10958910e-01 1.80616066e-01
-2.64949538e-02 1.05183864e+00 -3.05469900e-01 -2.24940166e-01
3.39985371e-01 3.73757601e-01 3.93742740e-01 -9.05222952e-01
-3.22944283e-01 -1.40652850e-01 5.09472072e-01 -1.78278640e-01
6.99873716e-02 -5.17868936e-01 -1.70066643e+00 -1.05668880e-01
-4.55370337e-01 3.42278987e-01 5.66231430e-01 1.24300385e+00
6.26674771e-01 5.93983948e-01 4.95470554e-01 -3.15034300e-01
-4.64618772e-01 -9.74765718e-01 -7.17178881e-01 5.90314507e-01
-4.24957335e-01 -1.51418731e-01 -2.31174663e-01 3.32849473e-01] | [11.796923637390137, 8.896486282348633] |
15a7bb33-8c88-4021-93f7-e99f38244c50 | firl-fast-imitation-and-policy-reuse-learning | 2203.00251 | null | https://arxiv.org/abs/2203.00251v2 | https://arxiv.org/pdf/2203.00251v2.pdf | A Versatile Agent for Fast Learning from Human Instructors | In recent years, a myriad of superlative works on intelligent robotics policies have been done, thanks to advances in machine learning. However, inefficiency and lack of transfer ability hindered algorithms from pragmatic applications, especially in human-robot collaboration, when few-shot fast learning and high flexibility become a wherewithal. To surmount this obstacle, we refer to a "Policy Pool", containing pre-trained skills that can be easily accessed and reused. An agent is employed to govern the "Policy Pool" by unfolding requisite skills in a flexible sequence, contingent on task specific predilection. This predilection can be automatically interpreted from one or few human expert demonstrations. Under this hierarchical setting, our algorithm is able to pick up a sparse-reward, multi-stage knack with only one demonstration in a Mini-Grid environment, showing the potential for instantly mastering complex robotics skills from human instructors. Additionally, the innate quality of our algorithm also allows for lifelong learning, making it a versatile agent. | ['Chee-Meng Chew', 'Marcelo Ang', 'Jiayi Tan', 'Haofeng Liu', 'Zedong Zhang', 'YiWen Chen'] | 2022-03-01 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 4.18697484e-02 2.99360156e-01 -3.03171545e-01 -1.13758177e-01
-3.53934020e-01 -8.55731905e-01 7.43636250e-01 -1.27289698e-01
-7.02806592e-01 1.11722660e+00 -6.18869588e-02 -4.52684760e-01
-4.28016663e-01 -2.97869444e-01 -7.08440542e-01 -6.17293298e-01
-2.13570103e-01 6.86083794e-01 2.68404365e-01 -4.78042632e-01
3.16610307e-01 6.78049624e-01 -1.88048124e+00 -2.10517183e-01
1.12281060e+00 4.45805460e-01 1.08660579e+00 7.08845377e-01
1.45105273e-01 1.04413760e+00 -3.90506744e-01 1.06782876e-01
2.43248820e-01 -2.89827406e-01 -9.77232158e-01 6.59363866e-02
-4.70993519e-02 -3.50710392e-01 -1.20960452e-01 7.56567061e-01
4.21903998e-01 4.27284300e-01 4.28437799e-01 -1.18170726e+00
-1.44855008e-01 6.48036718e-01 -9.42308679e-02 4.68814373e-02
6.71393514e-01 7.65004873e-01 8.68480682e-01 -4.00059074e-01
7.29564548e-01 1.13504767e+00 3.46541941e-01 6.37979925e-01
-1.08072889e+00 -3.46840024e-01 1.93840832e-01 2.61317223e-01
-6.95161641e-01 -1.99890301e-01 4.96345520e-01 -5.52045405e-01
9.50184941e-01 -5.90526015e-02 1.17970920e+00 1.05778515e+00
2.64043093e-01 9.85780478e-01 1.30723214e+00 -4.36118662e-01
5.06762981e-01 1.75037384e-01 -3.83587301e-01 9.10126507e-01
-3.86653980e-03 4.49383169e-01 -5.24231076e-01 7.83616379e-02
1.03361106e+00 4.34108041e-02 -2.17445031e-01 -9.72885966e-01
-1.41901207e+00 5.74571550e-01 2.66339242e-01 2.34982401e-01
-4.56183106e-01 8.42740387e-02 3.41558129e-01 8.38694453e-01
-4.00558263e-01 1.19547784e+00 -6.51212454e-01 -6.64396942e-01
-5.03747106e-01 2.91087449e-01 1.17741406e+00 9.43085015e-01
8.14527273e-01 -5.10082170e-02 2.01416165e-01 5.26633441e-01
-8.55412036e-02 2.28229672e-01 5.99049687e-01 -1.65626752e+00
1.66877955e-01 4.61137742e-01 5.54613411e-01 -4.06571329e-01
-4.82395291e-01 -2.60039657e-01 -3.45427990e-01 8.77892673e-01
5.35683036e-01 -3.36992115e-01 -8.47580254e-01 1.56261694e+00
5.64309359e-01 -1.07755408e-01 -2.16987450e-02 9.79242086e-01
-6.04885165e-03 4.15977508e-01 1.49569273e-01 -8.83060470e-02
1.16691446e+00 -1.05992043e+00 -1.73248291e-01 -3.94033372e-01
6.55079544e-01 -5.17705321e-01 1.49434614e+00 7.31143534e-01
-9.97067153e-01 -4.57318366e-01 -8.61965656e-01 1.35795251e-01
-2.54701048e-01 -3.28932434e-01 9.35842812e-01 1.72562093e-01
-8.70809019e-01 9.70382392e-01 -1.04148638e+00 -4.43362176e-01
2.34566242e-01 5.37360489e-01 -4.71240073e-01 -1.41105324e-01
-7.64909327e-01 1.37422454e+00 4.95505273e-01 -2.64591575e-01
-1.31812906e+00 -5.98909497e-01 -6.45169258e-01 -1.75467192e-03
8.82810175e-01 -7.60657966e-01 1.67553568e+00 -1.04711270e+00
-2.10376978e+00 5.89375675e-01 4.45057213e-01 -2.50923812e-01
6.46012962e-01 -2.19096065e-01 2.35924125e-01 2.15775400e-01
1.13789789e-01 8.66864681e-01 8.25055957e-01 -1.08848524e+00
-8.79570484e-01 -9.27358307e-03 5.94033897e-01 7.41961479e-01
-1.06813960e-01 -8.35049972e-02 3.25719379e-02 -9.91479605e-02
-2.19913155e-01 -1.16128790e+00 -5.72898448e-01 -8.69262293e-02
6.01313710e-02 -4.74701256e-01 5.84186018e-01 -7.10926577e-02
5.97670794e-01 -2.03541183e+00 6.56558752e-01 -1.65448189e-01
-4.14169952e-02 2.36006483e-01 -1.36121005e-01 4.96240050e-01
2.42549971e-01 -2.26165771e-01 4.53393627e-03 7.06228018e-02
2.67546684e-01 4.83646691e-01 -5.13769127e-02 1.41212925e-01
-5.51553592e-02 8.06019962e-01 -1.52139974e+00 -5.25861442e-01
2.47408077e-01 4.20395024e-02 -5.84620595e-01 5.29479921e-01
-4.88425970e-01 9.70033884e-01 -7.28079319e-01 4.63025719e-01
-2.14583024e-01 -3.48814070e-01 3.96213531e-01 6.96408987e-01
-3.00500572e-01 4.12069350e-01 -1.12340593e+00 2.08286023e+00
-7.77196646e-01 2.96102762e-01 5.52328765e-01 -9.42384243e-01
6.20564103e-01 2.44410828e-01 5.07621825e-01 -3.55682939e-01
1.20672109e-02 3.05391461e-01 3.18408251e-01 -7.71324575e-01
5.45435905e-01 -1.61399260e-01 -1.23446859e-01 6.72942877e-01
1.25679135e-01 -7.40643799e-01 4.49010998e-01 1.62218109e-01
1.27342737e+00 8.50778699e-01 3.89641464e-01 -2.34314278e-01
1.26651779e-01 2.94488430e-01 3.15772861e-01 1.07608163e+00
-4.48593885e-01 1.83055084e-02 3.67578745e-01 -4.99148309e-01
-1.17025912e+00 -7.69992948e-01 3.20765376e-01 1.67898631e+00
-1.73946656e-02 -9.34387818e-02 -4.56806183e-01 -6.02602422e-01
-6.15691161e-03 4.25297469e-01 -1.87713593e-01 -1.89460050e-02
-8.70834351e-01 -6.86226264e-02 1.04638245e-02 2.18136892e-01
3.04344118e-01 -1.75964546e+00 -1.59346032e+00 4.56448317e-01
2.05232129e-01 -8.42808247e-01 -2.83439577e-01 6.88825905e-01
-8.45455885e-01 -9.38556254e-01 -6.54539585e-01 -9.72730041e-01
5.34873128e-01 3.09285492e-01 1.15554619e+00 9.82217342e-02
-3.99215132e-01 7.40909100e-01 -2.33234808e-01 -2.73814410e-01
-5.50940156e-01 2.91377604e-01 3.65351737e-01 -7.32914269e-01
-1.69423014e-01 -1.00945330e+00 -5.27753592e-01 1.82769135e-01
-4.72580165e-01 2.23948136e-01 1.01839960e+00 9.39270437e-01
2.77745813e-01 -6.11755811e-02 5.66700280e-01 -6.16737008e-01
7.21468210e-01 -4.78780061e-01 -7.51893520e-01 1.03143312e-01
-7.04328418e-01 3.07581872e-01 8.19853425e-01 -7.67805159e-01
-9.67691123e-01 4.11568791e-01 2.66920477e-01 -2.75206149e-01
-3.66694480e-01 4.42595601e-01 1.38520434e-01 3.01662534e-02
6.99595749e-01 2.33864948e-01 3.67051870e-01 -2.30755866e-01
7.24337459e-01 3.92327189e-01 5.89841723e-01 -1.23334980e+00
5.42154789e-01 -1.04268886e-01 -1.14197172e-01 -6.18407905e-01
-7.01345742e-01 -2.71345079e-01 -6.99641347e-01 -2.65948772e-01
5.72132587e-01 -8.26818049e-01 -1.19944680e+00 2.18911201e-01
-7.73826301e-01 -1.08128548e+00 -5.69041252e-01 4.89478141e-01
-9.52282012e-01 7.34711066e-02 -5.22820294e-01 -8.34137559e-01
2.09703520e-01 -1.32162678e+00 6.47074282e-01 5.21680415e-01
-5.07277966e-01 -7.80740321e-01 6.51068538e-02 1.51400670e-01
4.86934304e-01 -3.17354053e-02 6.62222803e-01 -4.75381166e-01
-8.62757146e-01 -3.83556522e-02 2.03796476e-01 6.93988800e-02
5.50149083e-02 -3.00375462e-01 -4.70025122e-01 -4.54689950e-01
-1.78965539e-01 -1.09461474e+00 2.47061580e-01 -1.13999754e-01
9.35077071e-01 -3.61234456e-01 -3.41829300e-01 2.21279904e-01
9.83522892e-01 2.26237386e-01 1.39813378e-01 7.23341882e-01
4.64950979e-01 6.52024150e-01 7.73040354e-01 2.77186066e-01
4.38473403e-01 5.58541477e-01 4.97571439e-01 3.02501172e-01
5.94735853e-02 -3.24717194e-01 3.70859623e-01 6.08487427e-01
-3.52109522e-01 2.54608095e-01 -9.23043251e-01 4.96073335e-01
-2.01217127e+00 -1.06372368e+00 5.58916450e-01 1.94705141e+00
1.02744234e+00 2.79441744e-01 2.98372090e-01 -2.43364334e-01
3.20036829e-01 7.70072686e-03 -8.60916853e-01 -3.97005826e-01
5.06767333e-01 7.94530511e-02 1.64939806e-01 4.13077295e-01
-8.01123798e-01 1.18975782e+00 7.06111336e+00 6.65456116e-01
-1.01005757e+00 -9.67979580e-02 1.00957930e-01 -6.99647889e-02
-9.54073481e-03 1.45408228e-01 -5.32809734e-01 5.00111580e-01
8.37148070e-01 -1.76165640e-01 9.79376614e-01 1.21248484e+00
2.24700570e-01 -5.10502636e-01 -1.17358959e+00 5.66185236e-01
-4.23568517e-01 -1.02109313e+00 -4.37455148e-01 1.20680638e-01
6.61801457e-01 6.27539903e-02 1.68095976e-02 6.90493584e-01
1.24447751e+00 -1.13067091e+00 5.83490133e-01 3.75851005e-01
6.88990295e-01 -5.89029729e-01 1.77664831e-01 1.08569121e+00
-7.15214133e-01 -6.08771563e-01 -2.44361266e-01 -5.34038484e-01
-4.77672517e-02 4.47915196e-02 -1.11272848e+00 3.22353065e-01
5.19826293e-01 3.77368003e-01 3.46870795e-02 8.67549956e-01
-5.50908208e-01 1.68421686e-01 -3.16060036e-01 -5.56797862e-01
4.60901797e-01 -2.92724043e-01 4.72718894e-01 8.78975153e-01
2.36621708e-01 3.61110002e-01 7.43247688e-01 5.02507389e-01
2.70356566e-01 -1.77991256e-01 -8.19829285e-01 -9.32693630e-02
6.06207371e-01 1.45754516e+00 -7.20444500e-01 -3.02989304e-01
-2.56402139e-02 9.35613155e-01 7.02020347e-01 1.75225019e-01
-5.42182684e-01 -1.71085641e-01 5.90809166e-01 -2.14275792e-01
2.72607118e-01 -5.97051382e-01 -8.87336880e-02 -1.06592953e+00
-2.74939328e-01 -1.24986100e+00 2.12628096e-02 -8.64416242e-01
-8.11804891e-01 3.25730741e-01 -3.50661650e-02 -1.19124854e+00
-5.79247832e-01 -5.11762202e-01 -4.87739056e-01 4.47305650e-01
-1.41257954e+00 -8.45093906e-01 -2.12106124e-01 5.15122056e-01
7.01667786e-01 -4.46798563e-01 9.28415835e-01 -1.90440118e-01
-3.44405562e-01 1.32158726e-01 -7.89995641e-02 -3.69829804e-01
5.66102445e-01 -1.47487438e+00 2.42393911e-02 4.71644640e-01
-4.06884849e-02 6.87997341e-01 9.87224460e-01 -3.95436406e-01
-1.67801738e+00 -5.60942352e-01 4.03702646e-01 -6.34282649e-01
9.05896723e-01 -1.38736546e-01 -7.05655754e-01 7.16938376e-01
1.76277667e-01 -2.82615960e-01 1.07522659e-01 3.25131506e-01
-1.67211711e-01 -7.54668145e-03 -8.66520882e-01 8.68659139e-01
1.16574454e+00 -3.85439098e-01 -9.20972228e-01 4.83219206e-01
6.42777205e-01 -7.08619893e-01 -6.45452976e-01 8.79121795e-02
5.70912957e-01 -8.94962370e-01 7.75781512e-01 -7.32496858e-01
5.71715295e-01 -2.28761062e-01 2.89943516e-01 -1.54617310e+00
-4.43592638e-01 -1.23387289e+00 -4.86350432e-02 6.22547925e-01
3.00793856e-01 -5.08657694e-01 7.40425229e-01 7.03336775e-01
-4.47409302e-01 -8.28140140e-01 -6.29562676e-01 -9.80741262e-01
4.26033847e-02 -8.31332803e-02 4.47395444e-01 9.34334576e-01
5.75301528e-01 5.42338014e-01 -4.65264261e-01 -2.53705978e-01
3.75887573e-01 3.40045571e-01 1.06754398e+00 -1.34535098e+00
-7.46493340e-01 -5.70005715e-01 1.76533177e-01 -1.40878999e+00
2.38194585e-01 -8.27573657e-01 5.21348894e-01 -1.31412625e+00
6.65859729e-02 -7.72167206e-01 -2.21938327e-01 7.43829429e-01
-5.03697470e-02 -3.59247833e-01 3.85892719e-01 3.25897932e-01
-9.08938885e-01 2.47422397e-01 1.69971764e+00 1.67472854e-01
-4.89649355e-01 2.31328961e-02 -6.60211980e-01 8.16102326e-01
8.42500389e-01 -2.83138067e-01 -6.38752222e-01 -2.42637664e-01
1.77778780e-01 3.76346380e-01 1.66434169e-01 -1.09279168e+00
5.12767136e-01 -7.05518901e-01 1.65985718e-01 2.04085469e-01
4.07331824e-01 -7.63566196e-01 3.65751795e-02 6.66053534e-01
-4.57063049e-01 4.17757742e-02 -1.08867809e-01 5.21748900e-01
1.82782710e-01 -3.57815772e-01 6.97769046e-01 -7.95414448e-01
-9.42065120e-01 1.13503143e-01 -7.14227796e-01 -8.53893906e-02
1.28382194e+00 -3.43342632e-01 -6.47162795e-02 -3.54785055e-01
-7.92852640e-01 6.14857793e-01 6.48633063e-01 2.25034550e-01
2.22724378e-01 -7.96571851e-01 -2.61173606e-01 -5.13863331e-03
-1.51480779e-01 2.68404841e-01 -7.50832856e-02 6.42347991e-01
-2.63727337e-01 2.31160432e-01 -6.97435617e-01 -5.36688328e-01
-9.47340071e-01 5.94644368e-01 1.78693399e-01 -4.37733710e-01
-9.63636994e-01 7.97174573e-01 -1.83034793e-01 -6.31230175e-01
3.92303646e-01 -1.69680759e-01 -9.45549756e-02 -1.45151153e-01
3.72260064e-01 1.65330708e-01 -4.39754695e-01 1.86087430e-01
-8.32153782e-02 2.78284043e-01 -4.55355011e-02 -3.40825289e-01
1.49347734e+00 -1.01578999e-02 9.58013237e-02 5.50787270e-01
3.37833166e-01 -2.26949587e-01 -2.00547886e+00 5.78551777e-02
1.84892192e-01 -2.22203702e-01 -4.13183182e-01 -1.08465719e+00
-2.80374825e-01 6.01445079e-01 1.05749659e-01 3.85773987e-01
7.25657165e-01 -1.12709612e-01 4.49925154e-01 1.02221537e+00
1.11860287e+00 -1.31469822e+00 5.53727686e-01 8.08275282e-01
7.16549635e-01 -1.29496050e+00 -7.73389358e-04 8.13091546e-02
-8.66300166e-01 1.05731618e+00 7.57986009e-01 -2.37017974e-01
1.95921287e-01 2.23209396e-01 -3.11807077e-03 -8.13184753e-02
-1.00394082e+00 -1.83114037e-01 -2.84198523e-01 7.63709724e-01
9.93516371e-02 6.82355538e-02 -3.18416432e-02 7.41180554e-02
-4.81601924e-01 1.33890927e-01 8.13235760e-01 1.24880147e+00
-1.09364676e+00 -1.09098184e+00 -9.00026560e-02 1.85954556e-01
-3.43284570e-02 3.05036783e-01 -1.24852978e-01 9.21145916e-01
1.51959183e-02 7.42565274e-01 -2.02887893e-01 -3.32234576e-02
2.43464217e-01 2.98185438e-01 8.22225332e-01 -9.31676686e-01
-5.69960237e-01 -2.85956472e-01 2.35868692e-02 -7.59877741e-01
-1.91369176e-01 -5.70365310e-01 -1.37861931e+00 -2.76988506e-01
8.04405753e-03 2.64188468e-01 6.80214345e-01 1.13423598e+00
3.53524357e-01 5.28046131e-01 5.26092231e-01 -1.22729230e+00
-1.03519952e+00 -8.40846419e-01 -4.41133738e-01 1.10723913e-01
4.00594205e-01 -7.67917335e-01 -2.49707535e-01 4.79101278e-02] | [4.218785762786865, 1.3027145862579346] |
024d522a-922d-408f-925d-f53defd989a9 | continual-one-shot-learning-of-hidden-spike | 1708.09072 | null | http://arxiv.org/abs/1708.09072v1 | http://arxiv.org/pdf/1708.09072v1.pdf | Continual One-Shot Learning of Hidden Spike-Patterns with Neural Network Simulation Expansion and STDP Convergence Predictions | This paper presents a constructive algorithm that achieves successful
one-shot learning of hidden spike-patterns in a competitive detection task. It
has previously been shown (Masquelier et al., 2008) that spike-timing-dependent
plasticity (STDP) and lateral inhibition can result in neurons competitively
tuned to repeating spike-patterns concealed in high rates of overall
presynaptic activity. One-shot construction of neurons with synapse weights
calculated as estimates of converged STDP outcomes results in immediate
selective detection of hidden spike-patterns. The capability of continual
learning is demonstrated through the successful one-shot detection of new sets
of spike-patterns introduced after long intervals in the simulation time.
Simulation expansion (Lightheart et al., 2013) has been proposed as an approach
to the development of constructive algorithms that are compatible with
simulations of biological neural networks. A simulation of a biological neural
network may have orders of magnitude fewer neurons and connections than the
related biological neural systems; therefore, simulated neural networks can be
assumed to be a subset of a larger neural system. The constructive algorithm is
developed using simulation expansion concepts to perform an operation
equivalent to the exchange of neurons between the simulation and the larger
hypothetical neural system. The dynamic selection of neurons to simulate within
a larger neural system (hypothetical or stored in memory) may be a starting
point for a wide range of developments and applications in machine learning and
the simulation of biology. | ['Tien-Fu Lu', 'Steven Grainger', 'Toby Lightheart'] | 2017-08-30 | null | null | null | null | ['neural-network-simulation'] | ['computer-code'] | [ 5.08159757e-01 6.07861057e-02 6.21893346e-01 1.46698728e-02
9.72358808e-02 -3.58398259e-01 7.50833273e-01 -1.07299851e-03
-7.10245490e-01 1.08469176e+00 -5.87243319e-01 -2.34029204e-01
-3.80123295e-02 -8.21007967e-01 -9.51250553e-01 -1.18776882e+00
-3.56234998e-01 5.12594938e-01 7.15405941e-01 -3.05306256e-01
5.39371133e-01 8.34969699e-01 -2.20842409e+00 3.88346404e-01
4.32141006e-01 4.69538838e-01 7.30808794e-01 6.53638303e-01
-7.96933472e-02 4.83535439e-01 -5.85116386e-01 2.16043115e-01
2.29965478e-01 -8.60871136e-01 -9.65010673e-02 -1.60379961e-01
-3.33278596e-01 4.78184730e-01 1.89434081e-01 8.50586236e-01
5.11705995e-01 -1.23838492e-01 5.58695376e-01 -1.12479472e+00
-1.87315211e-01 8.54041100e-01 -2.77682841e-02 3.08595270e-01
1.36057278e-02 3.78770888e-01 7.02802539e-02 -1.20390451e+00
7.87104607e-01 8.11947167e-01 7.03585386e-01 8.01940143e-01
-1.65468574e+00 -8.50656867e-01 -2.59636700e-01 -1.72885135e-01
-1.56936896e+00 -2.93399841e-01 3.98771495e-01 -2.46296644e-01
1.16894031e+00 3.45842898e-01 1.59348559e+00 9.11484897e-01
6.03817225e-01 4.44864124e-01 1.33981907e+00 -5.59813797e-01
9.60597754e-01 2.63758935e-02 -4.65361625e-02 1.78047076e-01
5.98040104e-01 6.99201047e-01 -6.11955345e-01 7.20381066e-02
1.17120421e+00 -8.63791853e-02 -3.28518860e-02 -1.53599620e-01
-9.99581695e-01 5.35488784e-01 2.85331666e-01 6.39042497e-01
-5.19475162e-01 3.03584844e-01 1.05374187e-01 5.72228432e-01
-1.61636278e-01 7.13986218e-01 -2.53765106e-01 4.91106659e-02
-1.19547975e+00 4.59454387e-01 8.88590991e-01 4.33447450e-01
7.48346686e-01 4.90802646e-01 2.71959662e-01 4.68880951e-01
3.05980202e-02 3.19890618e-01 6.31897271e-01 -8.96714211e-01
-4.18084979e-01 6.16290152e-01 -1.96161300e-01 -4.64103103e-01
-4.01890337e-01 -6.82725608e-01 -8.80902231e-01 1.01687682e+00
3.79412651e-01 -3.01350076e-02 -9.94268358e-01 1.70426893e+00
-1.07070804e-01 5.16665816e-01 2.10018575e-01 3.98835450e-01
2.00314060e-01 8.31029236e-01 -2.31620595e-01 -7.44781315e-01
6.34650707e-01 -2.77177155e-01 -3.96404117e-01 -8.06232840e-02
3.93342286e-01 -2.30888203e-01 4.36699212e-01 4.30801392e-01
-1.45933664e+00 -4.08774465e-01 -1.19113326e+00 7.67199576e-01
-6.99254215e-01 -6.02905810e-01 4.21548516e-01 3.41507345e-01
-1.31589758e+00 7.51708388e-01 -9.30121958e-01 -4.46795553e-01
4.30907637e-01 6.64406598e-01 -4.58529685e-03 4.86615211e-01
-1.00819457e+00 1.01993537e+00 5.60862303e-01 5.94948232e-02
-1.02420175e+00 -5.85665941e-01 -2.21529171e-01 8.89336318e-02
-1.72123164e-01 -7.07744956e-01 7.47407079e-01 -1.20688868e+00
-1.30107248e+00 1.14122558e+00 -6.70307875e-02 -9.90688741e-01
2.72737354e-01 6.32796705e-01 -1.07941730e-02 -1.03093766e-01
-2.25153297e-01 8.82936776e-01 4.78319108e-01 -1.60315478e+00
-4.65650290e-01 -2.69838959e-01 -5.65280139e-01 7.54731297e-02
3.05910319e-01 -8.86648893e-02 2.52157152e-01 -6.39930964e-01
5.41096270e-01 -9.39197004e-01 -4.00925457e-01 2.34691892e-02
9.89146531e-02 2.82310069e-01 3.98059219e-01 2.24583074e-01
7.86199689e-01 -1.98019850e+00 2.63620496e-01 4.41002250e-01
4.32880111e-02 3.01477611e-01 -8.87055919e-02 6.82369947e-01
-2.14641571e-01 8.89247004e-03 -6.03168547e-01 -4.55102250e-02
-5.76086760e-01 3.96140248e-01 -2.53791034e-01 2.67088711e-01
1.88752577e-01 8.95586967e-01 -6.65437102e-01 -1.45923898e-01
-1.01991728e-01 4.37136292e-01 -8.70574564e-02 -9.14850608e-02
4.99351993e-02 4.49295849e-01 1.46494716e-01 4.24858451e-01
4.52469319e-01 9.94864572e-03 2.22337283e-02 4.96793896e-01
-8.66445899e-01 -2.86932644e-02 -1.19418836e+00 1.07614946e+00
4.65987548e-02 5.48383653e-01 -1.56757474e-01 -1.15975189e+00
1.32821333e+00 3.84288400e-01 2.05066741e-01 -8.77184927e-01
3.26409072e-01 6.89941227e-01 6.75090373e-01 1.84593037e-01
-3.97880435e-01 -2.95215189e-01 3.13748151e-01 6.94392383e-01
1.36859179e-01 -3.80275637e-01 4.49401319e-01 -2.25542337e-02
1.09746230e+00 -1.64277732e-01 2.62998581e-01 -5.55348754e-01
4.22399789e-01 -9.86786559e-03 4.66713876e-01 9.94334757e-01
8.21968615e-02 4.72477108e-01 2.38695309e-01 -2.78615296e-01
-1.34896457e+00 -1.32356572e+00 -3.88124973e-01 7.04659462e-01
-5.48266061e-02 3.16801667e-01 -7.76907682e-01 7.05238760e-01
-1.35626599e-01 7.68225193e-01 -9.26437616e-01 -4.68624711e-01
-6.73543572e-01 -7.82071471e-01 7.51448631e-01 2.69359469e-01
2.14569494e-01 -1.76474023e+00 -1.36064732e+00 5.62397778e-01
6.88129306e-01 -3.07996064e-01 5.68678498e-01 1.08112419e+00
-1.25541747e+00 -6.87768340e-01 -8.15478981e-01 -1.23952246e+00
9.90295291e-01 1.63347572e-02 6.15752280e-01 4.27460521e-01
-6.16109788e-01 -5.23252077e-02 -1.03021882e-01 -6.88198864e-01
-6.90491855e-01 -4.01798129e-01 3.23938310e-01 -3.39283973e-01
1.95826203e-01 -1.11759758e+00 -3.22122842e-01 1.33468300e-01
-9.86320615e-01 3.59002560e-01 5.04956245e-01 9.80860233e-01
9.59971607e-01 -1.61801606e-01 7.29528904e-01 -6.92535818e-01
6.15859091e-01 -2.44456261e-01 -7.44233847e-01 7.27618635e-02
-7.25798011e-01 1.40225872e-01 4.64178145e-01 -8.12027931e-01
-6.89107835e-01 1.75281420e-01 3.75190563e-02 -1.61811247e-01
-5.46985641e-02 5.57206631e-01 4.27902997e-01 -4.75327313e-01
8.23362172e-01 1.10253596e+00 4.00705993e-01 -3.32762487e-02
-1.99626163e-01 -1.37598470e-01 4.74827468e-01 -1.92300171e-01
7.40307808e-01 3.37407082e-01 2.14679852e-01 -5.65389991e-01
2.47073397e-01 9.67890844e-02 -7.67625213e-01 -6.08951569e-01
2.66063273e-01 -5.23137808e-01 -5.83811879e-01 1.03149045e+00
-1.07134640e+00 -5.44841528e-01 -6.40776515e-01 3.90077025e-01
-8.45171809e-01 -3.11081529e-01 -5.25739610e-01 -1.14900517e+00
-1.05909385e-01 -8.19213092e-01 1.89290285e-01 4.02664185e-01
-1.91136450e-01 -7.53618658e-01 5.08400261e-01 -7.89675653e-01
4.73528177e-01 1.23355567e-01 7.81902611e-01 -7.84776747e-01
-6.11682773e-01 -2.29804173e-01 4.76338625e-01 -3.79648455e-03
-3.46265078e-01 5.18655181e-01 -8.67806077e-01 -2.45422825e-01
3.11937660e-01 -1.47831455e-01 1.06053972e+00 6.38268113e-01
4.83152866e-01 -4.33284342e-02 -6.56231940e-01 3.38615268e-01
1.57636583e+00 7.81616330e-01 1.04290271e+00 5.52540183e-01
-2.01650977e-01 6.78291440e-01 1.90813497e-01 2.36065179e-01
-5.81160605e-01 1.25048414e-01 4.11073089e-01 -1.57293603e-02
8.77172276e-02 -8.45994987e-03 7.23565072e-02 1.00844049e+00
-3.38573039e-01 6.83297403e-04 -9.47637379e-01 4.77218181e-01
-1.77498364e+00 -1.43236411e+00 -2.04203218e-01 2.56482506e+00
9.83079553e-01 7.07136631e-01 1.87322423e-02 5.32924354e-01
1.01455879e+00 -6.90271914e-01 -6.71828210e-01 -5.26747227e-01
-7.51341343e-01 4.24907744e-01 3.87170732e-01 4.19015825e-01
-2.04114780e-01 6.68629766e-01 7.42821550e+00 5.11723876e-01
-1.23403907e+00 -2.53807157e-01 4.95413542e-01 -3.17930281e-01
-1.36641145e-01 1.33165672e-01 -8.50797176e-01 6.68608069e-01
1.22292876e+00 -6.14828408e-01 5.48259974e-01 2.85975367e-01
2.35349789e-01 -4.32589352e-01 -1.03918695e+00 7.53855884e-01
-2.32854843e-01 -1.93008971e+00 3.07997223e-02 1.64481774e-01
9.11407828e-01 8.41807872e-02 6.41577691e-02 2.12350056e-01
1.68096915e-01 -1.01435876e+00 8.06628406e-01 8.77513230e-01
3.36562127e-01 -6.93079054e-01 5.94646752e-01 7.64118731e-01
-9.81407464e-01 -1.80572912e-01 -4.42712814e-01 -4.79606181e-01
1.72863647e-01 3.50645602e-01 -6.09867632e-01 -4.14897799e-01
4.97968405e-01 2.78411001e-01 -4.46864426e-01 1.54032075e+00
2.23106310e-01 6.13839746e-01 -5.79719782e-01 -4.54766065e-01
7.95955807e-02 -7.49917626e-02 6.79963589e-01 1.02319181e+00
5.47732353e-01 1.70210063e-01 -6.03830516e-01 1.35558534e+00
3.28131884e-01 -1.19195558e-01 -8.34443808e-01 -2.93474775e-02
8.99304271e-01 8.99727106e-01 -1.38554358e+00 -3.89690518e-01
1.18183978e-01 6.12171471e-01 1.00244030e-01 3.07704628e-01
-6.51149452e-01 -1.38987020e-01 1.91737935e-01 3.15231800e-01
3.65901172e-01 -9.38072503e-02 -8.30832481e-01 -5.26281953e-01
-4.55302864e-01 -4.72884029e-01 -3.29058975e-01 -8.56727302e-01
-9.35350239e-01 5.76161861e-01 -1.66657850e-01 -1.07322323e+00
-3.42642277e-01 -4.06253338e-01 -1.12614691e+00 8.50908339e-01
-6.02373958e-01 -5.11659920e-01 1.40189916e-01 5.40401876e-01
4.18629408e-01 -4.39748377e-01 1.01754785e+00 -2.13141561e-01
-3.20627868e-01 1.87433213e-01 4.76097882e-01 -3.45041603e-01
-1.83184519e-02 -9.32578385e-01 4.39859539e-01 7.01769888e-01
7.20582157e-02 8.65782261e-01 1.07744408e+00 -8.66247475e-01
-9.52730238e-01 -7.04641283e-01 8.72531474e-01 1.60880581e-01
5.70551753e-01 -5.96281409e-01 -1.31344235e+00 3.05873990e-01
1.07875548e-01 -3.13755304e-01 6.09263480e-01 -5.30843019e-01
1.24165095e-01 2.74225767e-03 -1.04507589e+00 9.39490855e-01
9.31466997e-01 -1.10738494e-01 -7.00164676e-01 -1.88897818e-01
1.87964290e-01 9.71574187e-02 -2.80942202e-01 3.14545631e-01
7.45696425e-01 -1.04582918e+00 7.25685000e-01 -1.91788092e-01
-6.93421438e-02 -3.05435240e-01 2.48375431e-01 -1.12517810e+00
-3.74396652e-01 -4.87252980e-01 1.65067673e-01 8.57331395e-01
6.45885706e-01 -1.09409642e+00 5.46428680e-01 2.48940036e-01
-6.04013987e-02 -7.79872894e-01 -1.24977899e+00 -1.00224948e+00
3.47663760e-01 -5.47602810e-02 5.42759383e-03 5.00874460e-01
2.03389481e-01 -7.17272460e-02 2.05849022e-01 -3.22099715e-01
6.22160852e-01 -2.22545981e-01 2.34249085e-01 -1.56466150e+00
-2.49221206e-01 -8.69379222e-01 -6.62201703e-01 -3.28210950e-01
-2.31988281e-01 -8.25902045e-01 3.68784368e-01 -1.27511418e+00
1.23679921e-01 -2.67036498e-01 -6.59214318e-01 2.63715506e-01
4.30453986e-01 3.11921090e-01 1.95255697e-01 5.22786796e-01
7.82075673e-02 1.03609078e-01 1.01051700e+00 2.74782330e-01
-2.92498589e-01 6.55416846e-02 -1.21378571e-01 5.83074272e-01
8.79946053e-01 -9.13457513e-01 -3.01999599e-01 2.59744495e-01
5.66988468e-01 9.61696543e-03 5.12220085e-01 -1.48693120e+00
7.14956045e-01 3.81340422e-02 4.67301399e-01 -7.65768230e-01
3.24435622e-01 -5.71854115e-01 7.20643044e-01 1.37108517e+00
-5.01592934e-01 1.81921244e-01 4.36878860e-01 4.65936065e-01
4.75265831e-02 -5.62536001e-01 1.01582038e+00 -5.54065347e-01
-5.22449493e-01 -2.72412688e-01 -1.49913681e+00 -3.64924192e-01
1.33159244e+00 -1.08219290e+00 -1.16293706e-01 -1.13701243e-02
-1.11319149e+00 -1.73648566e-01 5.73080361e-01 -1.14770643e-01
9.17549670e-01 -1.02153039e+00 -6.78162575e-01 4.90743250e-01
-3.24274868e-01 -4.51593935e-01 -2.18949184e-01 8.40179026e-01
-7.12503552e-01 2.94626713e-01 -1.13889813e+00 -4.70864713e-01
-1.17140484e+00 4.16117728e-01 5.44059634e-01 2.32772470e-01
-3.75413209e-01 1.11615515e+00 1.06517792e-01 1.55293718e-02
7.75792152e-02 -5.17596677e-03 -1.37012288e-01 4.54290919e-02
6.12339616e-01 1.28287032e-01 -2.38427401e-01 -2.44372338e-01
-2.50214696e-01 3.66944522e-01 6.27269521e-02 -5.46190262e-01
1.60999966e+00 1.95198525e-02 -4.17816848e-01 1.31543100e+00
5.93618691e-01 -4.10324574e-01 -1.21362674e+00 1.94796517e-01
-1.92219168e-02 1.72748297e-01 -3.95957053e-01 -9.75600958e-01
-7.44103909e-01 9.04893279e-01 7.69093812e-01 3.50157052e-01
9.62598205e-01 -1.89976513e-01 2.54809976e-01 4.83969182e-01
8.14959943e-01 -1.06345367e+00 1.06145434e-01 5.84128261e-01
9.18926120e-01 -5.66018879e-01 -3.62319410e-01 1.00955509e-01
-2.93397933e-01 1.37348902e+00 5.22369146e-01 -7.50117302e-01
6.60290897e-01 1.01491177e+00 -2.24248767e-01 -9.60689411e-02
-1.26232195e+00 -2.46966660e-01 -3.44445050e-01 8.78286362e-01
2.49549702e-01 -1.04016103e-01 -7.17583776e-01 3.13511372e-01
1.10967956e-01 4.03005838e-01 7.24802971e-01 1.08533418e+00
-1.13203609e+00 -8.11282754e-01 -3.59101385e-01 2.88460910e-01
9.05130357e-02 -3.88566017e-01 -4.82941866e-01 7.98482537e-01
2.29757637e-01 3.91017705e-01 4.13229525e-01 -1.81032509e-01
-2.02647001e-01 3.23063374e-01 7.69977868e-01 -6.86126411e-01
-8.38158071e-01 1.75146699e-01 -4.54460084e-01 -5.11238985e-02
-4.47912425e-01 -9.11600173e-01 -1.95754743e+00 -2.54748762e-01
-2.26822957e-01 -3.01697757e-03 6.81394637e-01 9.41966236e-01
2.62376647e-02 5.21716654e-01 3.29388380e-01 -1.05483162e+00
-6.99269325e-02 -5.88769197e-01 -9.00538743e-01 2.98153870e-02
-1.30236953e-01 -6.86270952e-01 -6.96010649e-01 3.02502871e-01] | [8.070778846740723, 2.8253636360168457] |
8e4bdc6b-6902-4aee-bdf9-744e8555dd4a | knowledge-authoring-for-rules-and-actions | 2305.07763 | null | https://arxiv.org/abs/2305.07763v1 | https://arxiv.org/pdf/2305.07763v1.pdf | Knowledge Authoring for Rules and Actions | Knowledge representation and reasoning (KRR) systems describe and reason with complex concepts and relations in the form of facts and rules. Unfortunately, wide deployment of KRR systems runs into the problem that domain experts have great difficulty constructing correct logical representations of their domain knowledge. Knowledge engineers can help with this construction process, but there is a deficit of such specialists. The earlier Knowledge Authoring Logic Machine (KALM) based on Controlled Natural Language (CNL) was shown to have very high accuracy for authoring facts and questions. More recently, KALMFL, a successor of KALM, replaced CNL with factual English, which is much less restrictive and requires very little training from users. However, KALMFL has limitations in representing certain types of knowledge, such as authoring rules for multi-step reasoning or understanding actions with timestamps. To address these limitations, we propose KALMRA to enable authoring of rules and actions. Our evaluation using the UTI guidelines benchmark shows that KALMRA achieves a high level of correctness (100%) on rule authoring. When used for authoring and reasoning with actions, KALMRA achieves more than 99.3% correctness on the bAbI benchmark, demonstrating its effectiveness in more sophisticated KRR jobs. Finally, we illustrate the logical reasoning capabilities of KALMRA by drawing attention to the problems faced by the recently made famous AI, ChatGPT. | ['Michael Kifer', 'Paul Fodor', 'Yuheng Wang'] | 2023-05-12 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-3.01568836e-01 6.60110593e-01 -2.62334734e-01 -3.31589073e-01
-2.81508058e-01 -7.93936193e-01 8.01199734e-01 1.77489758e-01
-4.04907987e-02 9.31441247e-01 6.12055250e-02 -1.09610116e+00
-6.75704956e-01 -1.27217102e+00 -7.73693979e-01 2.80867457e-01
-2.60149734e-03 8.85793746e-01 5.29510498e-01 -7.24055648e-01
1.17623080e-02 4.86798942e-01 -1.59042478e+00 4.88399506e-01
7.50693142e-01 8.72740626e-01 -2.89875060e-01 6.27955735e-01
-5.93393028e-01 1.96594393e+00 -6.89412355e-01 -6.00699663e-01
1.07025869e-01 5.58029748e-02 -1.63548374e+00 -5.20974755e-01
3.04450870e-01 -2.01244503e-01 -2.87009835e-01 6.82207465e-01
-2.54561514e-01 2.41974354e-01 3.61611575e-01 -1.65899932e+00
-7.73626685e-01 1.31411469e+00 1.37691185e-01 -1.97388530e-02
9.91472661e-01 -1.17226236e-01 1.29262090e+00 -4.09360975e-01
7.98607767e-01 1.46290410e+00 6.48114026e-01 5.65590978e-01
-8.31925094e-01 -3.97945076e-01 1.65462300e-01 5.70044994e-01
-1.65112782e+00 -2.76041597e-01 2.07487564e-03 -2.78266937e-01
1.72067189e+00 7.06516504e-01 4.40666288e-01 3.38140279e-01
9.90905911e-02 6.36721373e-01 8.60270083e-01 -7.17584372e-01
1.55133173e-01 3.52555633e-01 4.61235821e-01 8.96576226e-01
6.75380051e-01 -1.97753638e-01 -5.53745568e-01 -4.07818913e-01
8.79882216e-01 -3.11825067e-01 -9.19326469e-02 -3.69140804e-02
-1.14382887e+00 7.71451592e-01 -3.87289748e-02 6.13079607e-01
-1.36353254e-01 3.78549427e-01 3.23684692e-01 4.53924030e-01
-3.78383607e-01 8.77689242e-01 -7.57803202e-01 -3.55601728e-01
-6.51603818e-01 6.11572027e-01 1.54828548e+00 1.30108941e+00
2.97638535e-01 -4.55817655e-02 1.48766577e-01 3.11313897e-01
4.32147384e-01 4.53733593e-01 2.72545099e-01 -1.45389938e+00
3.04271281e-01 9.42285359e-01 4.18155491e-01 -1.07006097e+00
-3.62224579e-01 3.71008851e-02 -4.56285179e-01 1.98039815e-01
5.24883211e-01 1.52231827e-01 -7.07059503e-01 1.40200508e+00
1.02888264e-01 -1.52788768e-02 7.78739929e-01 5.70535541e-01
8.47870290e-01 7.99447596e-01 1.21491469e-01 -3.75559330e-02
1.44150257e+00 -5.23421884e-01 -6.51557863e-01 -9.43503603e-02
8.19088876e-01 -1.83109999e-01 8.12048256e-01 7.76503623e-01
-1.29131842e+00 -1.15100302e-01 -8.28974128e-01 -2.55556226e-01
-8.66525412e-01 -3.58898550e-01 1.45073497e+00 4.19512689e-01
-9.08278704e-01 2.68973589e-01 -4.01077062e-01 -4.48816657e-01
9.90002006e-02 2.27018654e-01 -3.42873603e-01 -4.55421537e-01
-1.82362783e+00 1.25930858e+00 9.92903233e-01 -4.50942256e-02
-5.08596539e-01 -1.02015793e+00 -9.74754155e-01 1.58625320e-01
1.09034109e+00 -7.40418792e-01 1.53705621e+00 -1.98676229e-01
-1.33421612e+00 5.97669423e-01 1.36815444e-01 -9.79371607e-01
4.38960373e-01 -3.68047029e-01 -1.03589272e+00 6.64546341e-02
4.59733680e-02 3.35191160e-01 1.93702102e-01 -9.64675725e-01
-9.65679705e-01 1.32476375e-01 9.77155924e-01 -1.22039497e-01
3.32816094e-01 1.01638608e-01 -3.42164040e-01 -8.15372393e-02
6.45004511e-02 -4.41106856e-01 -7.80131966e-02 -4.36079919e-01
-9.81501341e-02 -4.71040636e-01 4.84958380e-01 -3.55249882e-01
1.48998463e+00 -1.64639330e+00 -2.16412038e-01 7.11876810e-01
2.86009848e-01 4.05971915e-01 3.28707188e-01 5.97764373e-01
-1.18040897e-01 5.36389887e-01 1.32494241e-01 9.32385027e-01
4.24575686e-01 8.24715197e-01 -9.42231178e-01 -3.48407894e-01
4.70663607e-02 1.11235225e+00 -1.05000937e+00 -7.43226409e-01
2.78734118e-01 2.50636451e-02 -5.09950578e-01 -1.66231334e-01
-8.97059739e-01 -4.24792886e-01 -4.38082159e-01 7.79568911e-01
6.73961267e-02 -5.20880520e-01 7.39577174e-01 1.02857113e-01
7.11535383e-03 3.40433538e-01 -1.60148311e+00 1.40936124e+00
-5.43175757e-01 3.18686575e-01 -2.64954448e-01 -6.41312003e-01
7.46605277e-01 6.52640641e-01 9.86896455e-02 -3.77660304e-01
-2.61447191e-01 2.38104627e-01 -8.00418034e-02 -6.33412302e-01
5.84102809e-01 -2.09362119e-01 -3.17149907e-01 4.81783241e-01
-2.35643029e-01 -5.49343526e-01 7.06574440e-01 8.45122516e-01
1.36362290e+00 1.13986999e-01 7.72664249e-01 -9.35862735e-02
7.37163961e-01 7.09849596e-01 5.29740155e-01 1.18070054e+00
4.44033593e-01 -7.01969489e-02 5.09131372e-01 -7.59817481e-01
-5.04198790e-01 -7.66141355e-01 -5.65785244e-02 9.46590185e-01
-5.21960929e-02 -1.11407661e+00 -1.18933253e-01 -7.24615753e-01
4.02554274e-01 1.52326167e+00 -1.16664901e-01 -8.77560899e-02
-4.60583270e-01 -7.99280256e-02 1.06369567e+00 7.75971413e-01
6.12645864e-01 -1.10083055e+00 -9.25397694e-01 4.25528497e-01
-4.06410962e-01 -1.35796309e+00 4.62295800e-01 -1.14954775e-02
-5.29121578e-01 -1.55932677e+00 3.78415734e-01 -1.50944874e-01
4.56067652e-01 -3.88959907e-02 1.45849013e+00 5.02496004e-01
-2.57803082e-01 8.85511220e-01 -5.24641156e-01 -6.59152925e-01
-6.93987370e-01 -2.44205654e-01 -1.05120428e-03 -8.10568094e-01
5.61267614e-01 -4.12293553e-01 2.20222265e-01 3.76303494e-01
-1.11110425e+00 1.62099615e-01 4.05240148e-01 3.49389344e-01
1.31247640e-01 1.02406657e+00 5.22545934e-01 -1.41096461e+00
6.84316039e-01 -4.00307685e-01 -6.90162420e-01 8.76618385e-01
-8.27743590e-01 2.54364789e-01 8.31490219e-01 -9.64355469e-02
-1.23996007e+00 -5.08324027e-01 8.52055475e-02 2.07136348e-02
-2.29038045e-01 1.04088676e+00 -7.62033835e-02 2.35245764e-01
9.42116857e-01 -3.16486992e-02 -4.17976737e-01 2.38931496e-02
6.89722121e-01 4.37351555e-01 8.32575977e-01 -1.45893276e+00
8.53855431e-01 1.41140386e-01 1.99619099e-01 -3.88273805e-01
-1.03332114e+00 -2.16007531e-01 -1.55081972e-01 7.55264983e-02
3.28004181e-01 -7.66090274e-01 -1.26982558e+00 -1.20173581e-01
-1.01171553e+00 -5.39561152e-01 -6.87699914e-01 2.19770625e-01
-7.83397555e-01 2.42878422e-01 -4.70082372e-01 -9.58013177e-01
-2.04379201e-01 -4.46254522e-01 3.49232823e-01 -1.24771319e-01
-6.95759952e-01 -9.25871313e-01 -3.28848928e-01 6.07503712e-01
4.27717596e-01 1.48935199e-01 1.51487863e+00 -8.60803068e-01
-5.14624119e-01 -4.36520487e-01 -1.97734714e-01 2.76296884e-01
1.39586534e-02 1.13337278e-01 -3.16382706e-01 3.67505759e-01
-5.30118704e-01 -3.44687253e-01 4.73843776e-02 -2.37104282e-01
9.65516627e-01 -4.70205307e-01 -1.06060512e-01 -1.64397389e-01
1.34381938e+00 3.68893564e-01 1.01428235e+00 4.64270502e-01
2.80537426e-01 3.65955085e-01 9.53523755e-01 3.83317262e-01
7.22847700e-01 4.33527261e-01 4.28936305e-03 5.81831574e-01
1.71617895e-01 -3.19534987e-01 5.22508174e-02 1.46742284e-01
-6.98864877e-01 5.99775352e-02 -1.49374914e+00 3.76241148e-01
-2.34068704e+00 -1.29924512e+00 -1.59091756e-01 1.74953866e+00
1.19460189e+00 4.76277560e-01 -2.60673404e-01 5.43485940e-01
2.39502758e-01 -3.16616595e-01 -3.00838381e-01 -6.27201319e-01
8.50342028e-03 2.43299678e-01 2.76504427e-01 7.51940250e-01
-5.66165864e-01 1.22269464e+00 6.88648939e+00 6.42484725e-01
-4.09456313e-01 -2.96960324e-01 -4.19451416e-01 2.23522916e-01
-3.81768107e-01 4.58198309e-01 -9.65891302e-01 -1.21457018e-01
1.06764150e+00 -6.06263041e-01 8.69246185e-01 1.01108313e+00
-1.81987822e-01 -3.47828716e-01 -1.27555251e+00 8.95918727e-01
-1.69651747e-01 -1.82832754e+00 4.82922196e-01 -4.35736090e-01
5.03963351e-01 -7.12993681e-01 -6.52343214e-01 8.03725421e-01
1.15434158e+00 -1.31905878e+00 8.77669811e-01 9.36152399e-01
5.19272566e-01 -7.98097074e-01 8.11291635e-01 4.06602651e-01
-1.14603293e+00 -3.79209340e-01 -3.15998912e-01 -4.61271793e-01
2.84169186e-02 5.69755137e-01 -1.04318082e+00 1.14016318e+00
5.53729236e-01 5.02250612e-01 -4.11741763e-01 4.47752297e-01
-5.79016685e-01 4.32681918e-01 -4.69305217e-01 2.45385900e-01
8.09992999e-02 1.51580006e-01 7.70018920e-02 1.18420362e+00
1.15563171e-02 7.51367867e-01 5.42162359e-02 8.28318596e-01
1.89936683e-01 -3.15452188e-01 -5.86963654e-01 -3.95389467e-01
6.00376725e-01 9.06434417e-01 -3.77041042e-01 -8.87451649e-01
-3.72444242e-01 2.35322386e-01 1.53660014e-01 4.21893150e-01
-8.01754296e-01 -5.12353837e-01 4.73943502e-01 3.96767944e-01
-6.25557378e-02 -3.71437460e-01 7.75501430e-02 -1.08453739e+00
-7.72946551e-02 -1.19065857e+00 8.41444075e-01 -1.34083486e+00
-1.00363147e+00 3.22687745e-01 6.80654883e-01 -6.39608562e-01
-6.21059597e-01 -7.55641878e-01 -3.00743252e-01 6.10449195e-01
-1.43079734e+00 -1.16855431e+00 -1.97749630e-01 8.64113092e-01
-2.85517313e-02 -6.02782406e-02 1.01486361e+00 2.24903867e-01
-1.13655254e-01 3.55090685e-02 -8.02031159e-01 5.88656254e-02
4.34640825e-01 -1.53594816e+00 1.09270878e-01 6.89234853e-01
5.22808805e-02 1.32796776e+00 7.88198531e-01 -7.85934448e-01
-1.74823284e+00 -1.07716596e+00 1.22358942e+00 -8.18823576e-01
8.19729209e-01 1.91432297e-01 -8.97765815e-01 1.42626119e+00
-2.00523049e-01 -1.15574509e-01 5.28207481e-01 3.36979121e-01
-7.05653489e-01 -1.50140464e-01 -1.25184011e+00 6.07020020e-01
1.11644650e+00 -5.66150784e-01 -1.32541180e+00 5.43513238e-01
7.10923254e-01 -4.72672105e-01 -1.23237562e+00 2.82106251e-01
5.10317624e-01 -7.15814531e-01 9.34515476e-01 -1.17375600e+00
2.83472687e-01 -9.81057882e-01 -3.65405649e-01 -6.97623372e-01
-2.05781311e-01 -6.19583845e-01 -5.25508225e-01 1.16613853e+00
4.94834512e-01 -8.74082625e-01 2.76551574e-01 1.42158377e+00
5.59986383e-02 -4.56180990e-01 -4.14280415e-01 -9.66207623e-01
-2.50703603e-01 -1.08428538e+00 1.01411259e+00 1.07716477e+00
6.47580504e-01 2.89908230e-01 -5.87810241e-02 3.08578998e-01
2.54650384e-01 3.49694192e-01 7.24969149e-01 -1.58599126e+00
-2.68104166e-01 -5.97331673e-02 -3.36142033e-01 -5.39847732e-01
3.52073312e-01 -1.09106147e+00 -2.43197814e-01 -2.17333889e+00
-2.80866712e-01 -5.53674519e-01 1.07420504e-01 1.28862476e+00
4.68406051e-01 -4.09542263e-01 1.45844668e-01 3.39818001e-02
-8.20993304e-01 -3.74496937e-01 9.16366339e-01 -1.41399220e-01
-7.53708705e-02 -2.52212882e-01 -9.82994556e-01 9.62801337e-01
5.75780511e-01 -1.06061578e-01 -6.14396214e-01 -2.56612480e-01
1.23284233e+00 1.97778776e-01 5.59020996e-01 -1.11596465e+00
5.81729412e-01 -7.16373324e-01 -5.68410978e-02 -3.25484663e-01
9.84568801e-03 -1.20609450e+00 7.42710233e-01 2.04379588e-01
-3.11811745e-01 -7.97368363e-02 2.34305233e-01 1.13543704e-01
-3.52639705e-01 -4.37310308e-01 3.04165989e-01 -5.27643025e-01
-1.20147479e+00 -3.37292433e-01 -5.07780254e-01 2.31861964e-01
1.08622169e+00 -1.06797040e-01 -6.65003419e-01 -3.14666867e-01
-7.14501083e-01 5.85268795e-01 1.35559887e-01 9.34397131e-02
6.69302821e-01 -1.03343940e+00 -2.95424521e-01 -1.20320320e-01
1.75499365e-01 2.95927733e-01 -2.27429252e-02 5.17256379e-01
-8.48993123e-01 9.01774526e-01 -3.97000425e-02 1.06667720e-01
-9.57679093e-01 6.42567873e-01 3.49671811e-01 -4.74763989e-01
-7.57557213e-01 4.14485782e-01 -4.44310635e-01 -4.73833263e-01
8.95405114e-02 -7.94880331e-01 -1.63140520e-01 -2.41199434e-01
7.88560033e-01 4.76909965e-01 1.44602478e-01 9.43275467e-02
-6.85486495e-01 2.60599196e-01 -1.57134637e-01 -5.10168150e-02
1.22034967e+00 2.48537675e-01 -4.71927792e-01 3.06316346e-01
4.44212109e-02 1.19047493e-01 -1.48853660e-01 -3.89399081e-01
2.88915902e-01 -2.88871258e-01 -2.35661909e-01 -1.36001062e+00
-4.43679065e-01 2.92836607e-01 -4.35184300e-01 5.42187512e-01
6.68165028e-01 8.71150792e-02 6.75653517e-01 1.29049158e+00
1.01318407e+00 -1.11009324e+00 -5.25665522e-01 9.07733679e-01
1.00137901e+00 -7.98224509e-01 3.27030927e-01 -4.50171769e-01
-7.49311686e-01 1.19379032e+00 6.40913248e-01 3.12785327e-01
3.79651725e-01 4.74608421e-01 8.84567425e-02 -4.95904982e-01
-1.08087373e+00 -1.64293855e-01 -1.83997676e-01 4.23557401e-01
1.53616607e-01 2.18588799e-01 -1.97944179e-01 8.32749963e-01
-5.02544105e-01 7.06588268e-01 7.77583420e-01 1.49593961e+00
-4.11786050e-01 -1.11952913e+00 -5.48405588e-01 4.91221279e-01
-3.65205824e-01 -9.07244086e-02 -3.22034657e-01 1.19228208e+00
5.28840646e-02 1.13462639e+00 -3.57812971e-01 -1.96383089e-01
6.13822341e-01 4.84861314e-01 6.19782686e-01 -8.78837466e-01
-6.86414063e-01 -9.42018330e-01 7.82297552e-01 -6.66219234e-01
-3.29632670e-01 -1.82410404e-01 -1.94953871e+00 -7.17164934e-01
-5.53892739e-02 7.64155149e-01 1.60272956e-01 1.00326443e+00
2.42185354e-01 5.07160425e-01 -2.61530399e-01 3.02919835e-01
-5.16471863e-01 -4.60622698e-01 -7.66500294e-01 3.28560442e-01
-3.38755727e-01 -5.18583834e-01 -1.10274918e-01 2.51088738e-01] | [9.098252296447754, 7.218609809875488] |
854e2ef9-ae9f-430d-a242-f4312cc2fe94 | designing-a-3d-aware-stylenerf-encoder-for | 2302.09467 | null | https://arxiv.org/abs/2302.09467v1 | https://arxiv.org/pdf/2302.09467v1.pdf | Designing a 3D-Aware StyleNeRF Encoder for Face Editing | GAN inversion has been exploited in many face manipulation tasks, but 2D GANs often fail to generate multi-view 3D consistent images. The encoders designed for 2D GANs are not able to provide sufficient 3D information for the inversion and editing. Therefore, 3D-aware GAN inversion is proposed to increase the 3D editing capability of GANs. However, the 3D-aware GAN inversion remains under-explored. To tackle this problem, we propose a 3D-aware (3Da) encoder for GAN inversion and face editing based on the powerful StyleNeRF model. Our proposed 3Da encoder combines a parametric 3D face model with a learnable detail representation model to generate geometry, texture and view direction codes. For more flexible face manipulation, we then design a dual-branch StyleFlow module to transfer the StyleNeRF codes with disentangled geometry and texture flows. Extensive experiments demonstrate that we realize 3D consistent face manipulation in both facial attribute editing and texture transfer. Furthermore, for video editing, we make the sequence of frame codes share a common canonical manifold, which improves the temporal consistency of the edited attributes. | ['Jing Dong', 'Bo Peng', 'Wei Wang', 'Songlin Yang'] | 2023-02-19 | null | null | null | null | ['face-model'] | ['computer-vision'] | [ 1.38268456e-01 1.50741339e-01 -2.55652238e-02 -4.38854784e-01
-3.20053846e-01 -7.25335598e-01 5.83642483e-01 -1.06318724e+00
4.62386727e-01 6.48660362e-01 2.80248314e-01 8.02533999e-02
1.18043266e-01 -8.28380823e-01 -8.77364099e-01 -7.79272974e-01
5.78924537e-01 4.24571902e-01 -5.22316933e-01 -2.06774890e-01
-1.56056315e-01 7.01331198e-01 -1.39102626e+00 2.11433563e-02
8.53032172e-01 9.37768102e-01 8.67213756e-02 6.63260341e-01
-2.43948802e-01 4.28923994e-01 -2.29904160e-01 -5.39268792e-01
6.33199334e-01 -1.04650927e+00 -2.18210489e-01 4.29630965e-01
7.98644364e-01 -7.19430745e-01 -5.18933356e-01 6.93071485e-01
4.83614028e-01 -1.49988577e-01 6.50523901e-01 -1.38772178e+00
-1.05662310e+00 7.22142169e-03 -8.08241069e-01 -4.99445915e-01
4.53966200e-01 2.67590284e-01 6.48515582e-01 -8.62539172e-01
8.69430780e-01 1.54407024e+00 3.46308708e-01 9.91004050e-01
-1.21951199e+00 -8.94321084e-01 6.31358922e-02 -2.33602047e-01
-1.09564054e+00 -5.25382876e-01 1.12747967e+00 -4.75544602e-01
1.99453443e-01 3.20372939e-01 1.12716877e+00 1.27522862e+00
3.26169170e-02 5.01752675e-01 1.17299581e+00 -1.11425251e-01
-1.64288700e-01 -1.61179289e-01 -8.19665313e-01 1.06022573e+00
-2.62437128e-02 1.53088227e-01 -8.15906346e-01 2.76692271e-01
1.60554802e+00 2.08567560e-01 -2.93695718e-01 -4.78995591e-01
-1.26142597e+00 7.11113989e-01 2.49021485e-01 -1.52663633e-01
-3.07301521e-01 1.73822910e-01 1.30340993e-01 3.73712540e-01
6.80953860e-01 2.77311385e-01 -1.90999150e-01 -1.26455829e-01
-7.39335716e-01 2.38517299e-01 4.67929065e-01 1.52267301e+00
9.21316564e-01 4.62596864e-01 -4.62900281e-01 6.63397968e-01
2.49926671e-01 9.06840682e-01 -4.06252816e-02 -1.28437853e+00
3.35713983e-01 5.83062172e-01 -2.52152354e-01 -9.43338752e-01
1.25081167e-01 -1.83826029e-01 -1.25785530e+00 3.47634494e-01
1.50322855e-01 5.46985641e-02 -1.03042316e+00 1.93493867e+00
4.98020083e-01 2.68262684e-01 -2.45307073e-01 8.40565443e-01
6.55431330e-01 6.11922145e-01 -4.09086496e-01 -2.20923945e-01
1.25376570e+00 -9.49224949e-01 -1.01685321e+00 1.77879542e-01
2.11423799e-01 -8.68969917e-01 1.07156587e+00 5.30506074e-02
-1.42106640e+00 -5.07150531e-01 -9.53015745e-01 -5.76449096e-01
2.73831546e-01 1.04036964e-01 5.69364786e-01 5.52274525e-01
-8.72869849e-01 1.34481713e-01 -8.08126748e-01 -3.88152637e-02
7.88220167e-01 2.01426908e-01 -7.39788473e-01 -2.74889439e-01
-8.67474556e-01 6.51583731e-01 -1.57416105e-01 3.52049381e-01
-1.13230920e+00 -9.65737462e-01 -9.23319638e-01 -1.48906335e-01
1.82593599e-01 -1.33949733e+00 8.05269837e-01 -7.58497834e-01
-2.11939406e+00 1.11628544e+00 -2.45566100e-01 3.74703765e-01
8.53009403e-01 -1.10316031e-01 -4.85477857e-02 1.62964202e-02
8.49678516e-02 6.57386005e-01 1.29773295e+00 -1.23728406e+00
-1.13621980e-01 -5.16310275e-01 2.74884403e-01 4.19422537e-01
-3.25505018e-01 -4.60132927e-01 -5.19423723e-01 -1.01220930e+00
1.41247138e-01 -9.32754517e-01 2.22015113e-01 6.75309360e-01
-5.26577950e-01 4.07327443e-01 1.11828160e+00 -8.20822299e-01
7.53768444e-01 -2.11401892e+00 8.29825699e-01 1.28370330e-01
3.66725564e-01 7.94578493e-02 -2.34336078e-01 7.10460171e-02
-7.04137888e-03 1.42617017e-01 -2.34218583e-01 -6.89113081e-01
-6.51868135e-02 4.25132841e-01 -3.37189376e-01 3.33023876e-01
3.24195236e-01 1.03240311e+00 -7.18184948e-01 -3.90457094e-01
3.15751553e-01 8.95164311e-01 -1.08604646e+00 7.77604938e-01
-1.83076933e-01 1.32181227e+00 -6.74200356e-01 5.54511666e-01
1.03532875e+00 5.57974866e-03 4.44121398e-02 -6.03135943e-01
3.20445821e-02 -7.19549134e-02 -7.96167672e-01 2.22659636e+00
-7.98812747e-01 3.64351451e-01 3.52067620e-01 -6.97511852e-01
1.02956665e+00 2.51803011e-01 4.43044931e-01 -4.60986555e-01
-1.92777906e-02 2.63157964e-01 -3.01678866e-01 -4.10866678e-01
1.38579383e-01 -3.76777887e-01 1.80035964e-01 4.31331962e-01
1.82963744e-01 -7.17367828e-01 -3.47868234e-01 9.46144685e-02
5.44339180e-01 6.26569629e-01 -1.19224221e-01 -1.78947881e-01
6.26070678e-01 -7.63806224e-01 6.16554737e-01 5.26485592e-02
4.47472781e-01 9.92199481e-01 6.81523204e-01 -4.05616134e-01
-1.39262307e+00 -1.10501134e+00 9.23762247e-02 5.90204656e-01
1.46656051e-01 -4.05019104e-01 -8.75433207e-01 -5.88738441e-01
-9.04381499e-02 3.71562183e-01 -7.32543170e-01 -3.54031712e-01
-7.43334115e-01 -1.05808750e-01 5.02300560e-01 3.29087555e-01
9.21642601e-01 -5.68285108e-01 -3.05010408e-01 -8.83388445e-02
-1.74195662e-01 -1.01064038e+00 -1.10495830e+00 -5.47134876e-01
-7.26419449e-01 -9.09217715e-01 -9.31723118e-01 -6.17342710e-01
9.50967491e-01 2.41226166e-01 8.02288473e-01 1.56406820e-01
-1.02286860e-01 4.22079057e-01 -2.23173246e-01 -2.07068920e-01
-5.21660805e-01 4.02501933e-02 1.07236296e-01 2.92337060e-01
-3.34386647e-01 -1.14428914e+00 -8.26454937e-01 4.44423914e-01
-8.98117244e-01 6.84647977e-01 4.30864364e-01 1.01919663e+00
6.09473169e-01 -4.37485576e-01 4.03271466e-01 -7.81681061e-01
2.03647524e-01 -2.89206039e-02 -5.45009375e-01 2.04891235e-01
-2.29099512e-01 1.70742348e-01 6.83745861e-01 -3.48488480e-01
-1.36134553e+00 -1.43309208e-02 -3.88057500e-01 -9.09484446e-01
5.66427112e-02 -1.54336303e-01 -8.75316560e-01 -2.64908254e-01
2.19044968e-01 2.98789620e-01 5.29972613e-01 -5.07975399e-01
6.61759555e-01 3.79300416e-01 3.33967894e-01 -7.66951501e-01
1.09948385e+00 6.41570807e-01 3.92574221e-01 -7.01975107e-01
-8.27495098e-01 4.72439647e-01 -6.67020619e-01 -2.61139959e-01
1.08406305e+00 -1.11346304e+00 -7.41860807e-01 8.86947155e-01
-1.19787967e+00 -3.03171456e-01 -4.31856394e-01 2.09907889e-01
-9.21393454e-01 2.16046691e-01 -4.62765187e-01 -4.63722438e-01
-2.67678171e-01 -1.41117895e+00 1.45869827e+00 1.85482308e-01
1.10337019e-01 -7.91787267e-01 -3.18396211e-01 3.52337539e-01
5.38350403e-01 6.30245268e-01 8.65380406e-01 4.49138165e-01
-9.40069854e-01 1.98927432e-01 -7.24916011e-02 3.23119402e-01
5.84276974e-01 -1.14788815e-01 -8.87319326e-01 -2.43125513e-01
1.97916105e-01 -1.90962195e-01 5.24463713e-01 1.89003438e-01
1.55099714e+00 -3.88835847e-01 -6.78128749e-02 1.28035617e+00
8.81667733e-01 8.96865726e-02 6.06260121e-01 -3.20490628e-01
1.39837277e+00 4.11880761e-01 2.75072455e-01 4.30150241e-01
4.00631458e-01 9.95152950e-01 4.04156178e-01 -2.48434842e-02
-4.71673906e-01 -8.17893684e-01 3.68133903e-01 1.05533004e+00
-5.28210104e-01 -1.86587870e-01 -2.77135819e-01 -1.85614657e-02
-1.37370193e+00 -7.51103342e-01 2.57639349e-01 1.98088944e+00
8.10361445e-01 -3.78500938e-01 -1.78515539e-01 -2.33689085e-01
7.39162624e-01 2.47243747e-01 -8.63766611e-01 -1.11413054e-01
-1.71434507e-01 2.05358744e-01 1.08288758e-01 4.94062632e-01
-5.32747865e-01 8.31473231e-01 5.40338469e+00 7.71964788e-01
-1.12217939e+00 2.27957085e-01 6.73040748e-01 -2.01896951e-01
-1.05860877e+00 1.24975583e-02 -5.62981427e-01 5.43736458e-01
1.70461014e-01 -3.97603214e-02 6.18339896e-01 4.93187845e-01
1.68103471e-01 5.36970556e-01 -1.20620883e+00 1.36198330e+00
2.03830466e-01 -1.56404042e+00 6.10461771e-01 4.17985588e-01
1.04009521e+00 -8.50725651e-01 2.57732034e-01 -1.06184006e-01
-8.85898694e-02 -1.07704508e+00 8.79330695e-01 7.02109218e-01
1.76930332e+00 -7.14413464e-01 3.62158865e-02 3.06584090e-02
-1.12100554e+00 1.33442163e-01 -8.86891112e-02 2.59197056e-01
5.24200022e-01 5.65635920e-01 -1.53248310e-01 7.44218171e-01
4.63214546e-01 1.10200226e+00 -1.28502920e-01 2.49784335e-01
-3.59927803e-01 1.07383512e-01 -1.38256922e-01 5.34011185e-01
-1.03935361e-01 -7.25500703e-01 7.26536095e-01 3.82968932e-01
8.54357302e-01 3.25145513e-01 -1.84408665e-01 1.23617852e+00
-4.84325230e-01 -2.84224898e-01 -8.96208942e-01 -6.28724545e-02
4.79363769e-01 1.14175904e+00 -2.35463202e-01 -5.49041964e-02
-1.64172053e-01 1.38842118e+00 1.95520326e-01 3.08479488e-01
-9.41082597e-01 -1.33296326e-01 1.15412438e+00 2.00715333e-01
2.60410726e-01 -4.84066993e-01 -1.35973498e-01 -1.59339738e+00
1.48918912e-01 -9.15441394e-01 -1.55033126e-01 -9.13963139e-01
-1.27136123e+00 5.14666796e-01 -3.03764567e-02 -1.22037923e+00
-1.26703039e-01 -4.68740493e-01 -6.09301925e-01 1.01646125e+00
-1.33022714e+00 -1.72106063e+00 -6.86341763e-01 7.55053282e-01
4.34174627e-01 -2.78369516e-01 7.62449563e-01 3.54678661e-01
-4.34009403e-01 8.31829846e-01 -1.40903324e-01 -7.24280775e-02
8.06079805e-01 -9.12190795e-01 3.94643873e-01 5.43188035e-01
-1.64817110e-01 5.62296271e-01 1.58400074e-01 -5.47264695e-01
-2.05145526e+00 -1.16959107e+00 1.00067876e-01 -5.31688094e-01
-8.21048990e-02 -5.81090152e-01 -7.64655054e-01 8.29627514e-01
1.94027916e-01 7.78553784e-02 5.18419683e-01 -3.94244432e-01
-4.99201000e-01 -4.40197647e-01 -1.32452285e+00 6.21952474e-01
1.88116896e+00 -7.53978252e-01 -9.55339149e-02 2.03896895e-01
8.25458467e-01 -7.99010992e-01 -1.06305313e+00 3.61692458e-01
7.05943167e-01 -9.61981595e-01 9.36783552e-01 -3.89215350e-01
8.50420058e-01 -3.91066313e-01 -9.09866765e-02 -1.47434402e+00
-1.59769535e-01 -1.09454954e+00 -3.05083185e-01 1.51680541e+00
-1.46193177e-01 -6.10899210e-01 6.44293427e-01 3.26986074e-01
-2.73615509e-01 -6.36210501e-01 -9.00130987e-01 -5.47959447e-01
1.04775846e-01 1.07218921e-01 1.15972126e+00 9.84037817e-01
-6.75185144e-01 1.93193451e-01 -8.93650711e-01 -1.87393710e-01
6.19109035e-01 3.18843305e-01 1.12890971e+00 -9.53035653e-01
-1.91194415e-01 -2.75859773e-01 -4.25966144e-01 -1.51604295e+00
3.25996488e-01 -9.49434757e-01 -3.38282526e-01 -1.09037566e+00
8.92435685e-02 -5.41559935e-01 5.07132053e-01 2.47439459e-01
1.01676889e-01 3.34869415e-01 2.19311684e-01 7.31590390e-02
1.90022290e-01 1.18863165e+00 2.20449686e+00 9.45186242e-03
4.61186916e-02 -4.42848742e-01 -8.18817079e-01 3.41213405e-01
4.18481797e-01 2.39320993e-02 -7.13805497e-01 -9.06215906e-01
2.15029001e-01 2.65916288e-01 3.89761448e-01 -6.66418850e-01
-1.21572174e-01 -2.59523243e-01 6.25465572e-01 -2.12838441e-01
6.16658270e-01 -8.67081404e-01 6.67349100e-01 6.07793368e-02
-2.08762795e-01 7.29817105e-03 2.28272751e-03 6.09379947e-01
-1.59403831e-01 3.70040089e-01 8.75873625e-01 -2.18182251e-01
-1.09920822e-01 1.13874316e+00 2.11938605e-01 7.19828457e-02
9.98959601e-01 -2.43624270e-01 3.51665914e-02 -6.14937663e-01
-5.13744295e-01 8.47051963e-02 9.03891742e-01 5.76844215e-01
7.69069254e-01 -1.88923144e+00 -7.10420191e-01 9.26250160e-01
-1.21265642e-01 6.46739244e-01 6.33147657e-01 6.71000183e-01
-7.20394194e-01 -9.67868641e-02 -6.58816040e-01 -7.20928431e-01
-1.04006827e+00 1.82245031e-01 4.64991361e-01 1.57418028e-01
-7.36853778e-01 6.99322402e-01 6.39847517e-01 -6.41447306e-01
-2.95317143e-01 8.90111625e-02 2.29395911e-01 -8.68900493e-02
4.10875112e-01 2.43791625e-01 -2.38518953e-01 -6.89493418e-01
2.57461928e-02 1.15553319e+00 1.36934012e-01 -5.74487075e-02
1.42315686e+00 -3.27588379e-01 -4.18192089e-01 1.22677386e-01
1.30434906e+00 1.69121072e-01 -1.70136082e+00 2.18977839e-01
-9.64571536e-01 -1.10611975e+00 -1.80083200e-01 -1.41223237e-01
-1.64342070e+00 1.11113405e+00 2.79819578e-01 -2.98408091e-01
1.26099765e+00 -1.77617908e-01 9.52397227e-01 -5.25305457e-02
3.59515250e-01 -5.26319742e-01 2.18416721e-01 3.89374048e-01
1.33018756e+00 -8.24645877e-01 -1.88161418e-01 -7.60718167e-01
-5.17791450e-01 1.24325228e+00 9.43603516e-01 -3.23545448e-02
7.20785141e-01 3.34768407e-02 -2.72440314e-01 -1.99250534e-01
-4.19944435e-01 5.23000777e-01 2.32568026e-01 7.01579511e-01
2.57883757e-01 7.38292746e-03 1.23176370e-02 2.39263177e-01
-3.28507185e-01 3.74245341e-03 5.25697708e-01 5.18197119e-01
4.12195534e-01 -1.25528312e+00 -1.01679832e-01 2.18289033e-01
-1.69607438e-02 1.16352148e-01 -9.96188447e-02 4.94454592e-01
2.83725619e-01 3.34553897e-01 1.64941207e-01 -3.85074645e-01
3.96480739e-01 -5.45909330e-02 1.00926745e+00 -4.56132174e-01
8.11999068e-02 9.45469830e-03 -2.31862515e-01 -6.52571201e-01
-6.14034534e-01 -3.45468491e-01 -7.02561259e-01 -6.62355244e-01
-1.10350646e-01 -1.02575041e-01 4.18419182e-01 6.31061673e-01
7.38736928e-01 4.31923151e-01 9.96396124e-01 -1.03051090e+00
-1.82632402e-01 -6.47684872e-01 -6.61928773e-01 5.30906618e-01
3.60988140e-01 -9.53795433e-01 -2.56105542e-01 2.82226562e-01] | [12.588520050048828, -0.4032139480113983] |
97c2c0f8-d3b9-40a3-b77d-dc56e84d0dda | combining-deep-learning-and-argumentative | null | null | https://aclanthology.org/J18-4011 | https://aclanthology.org/J18-4011.pdf | Combining Deep Learning and Argumentative Reasoning for the Analysis of Social Media Textual Content Using Small Data Sets | The use of social media has become a regular habit for many and has changed the way people interact with each other. In this article, we focus on analyzing whether news headlines support tweets and whether reviews are deceptive by analyzing the interaction or the influence that these texts have on the others, thus exploiting contextual information. Concretely, we define a deep learning method for relation{--}based argument mining to extract argumentative relations of attack and support. We then use this method for determining whether news articles support tweets, a useful task in fact-checking settings, where determining agreement toward a statement is a useful step toward determining its truthfulness. Furthermore, we use our method for extracting bipolar argumentation frameworks from reviews to help detect whether they are deceptive. We show experimentally that our method performs well in both settings. In particular, in the case of deception detection, our method contributes a novel argumentative feature that, when used in combination with other features in standard supervised classifiers, outperforms the latter even on small data sets. | ['Oana Cocarascu', 'Francesca Toni'] | 2018-12-01 | null | null | null | cl-2018-12 | ['deception-detection'] | ['miscellaneous'] | [ 6.16677739e-02 4.92671132e-01 -4.39505130e-01 -3.47914398e-01
-5.00335813e-01 -6.84924960e-01 1.17036009e+00 1.06357348e+00
-5.03416479e-01 7.47108996e-01 5.47406435e-01 -6.49547577e-01
1.09525174e-01 -9.49854255e-01 -4.80925262e-01 -5.88468552e-01
2.11333379e-01 3.74593675e-01 -1.14164490e-03 -5.35472333e-01
5.19973814e-01 2.57126242e-01 -1.25693083e+00 6.40764177e-01
7.75321126e-01 9.27753031e-01 -9.12758410e-01 3.89123291e-01
-1.61415301e-02 1.57632565e+00 -9.79877591e-01 -1.11301208e+00
-2.26889789e-01 -3.39519441e-01 -1.23696458e+00 6.12050411e-04
2.79537022e-01 -2.57918715e-01 6.08451962e-02 8.79477084e-01
1.42222062e-01 -2.04323649e-01 8.48489642e-01 -1.13856959e+00
-9.72318053e-02 1.06779349e+00 -7.15172589e-01 5.13116002e-01
5.95189571e-01 -1.76025048e-01 1.13631761e+00 -3.52285922e-01
5.91275036e-01 1.23811269e+00 6.69710994e-01 1.72683731e-01
-1.04303360e+00 -4.59401786e-01 -3.08770575e-02 3.61561686e-01
-5.14090538e-01 -5.88701248e-01 1.04634726e+00 -6.65509343e-01
7.03789771e-01 3.06264013e-01 7.35190928e-01 1.40934479e+00
2.61874162e-02 8.28625083e-01 1.16209424e+00 -5.97461045e-01
1.75172716e-01 2.64596313e-01 9.03834105e-01 6.69928968e-01
5.21989107e-01 -4.87879068e-01 -4.61572170e-01 -6.30281806e-01
-3.58809829e-01 -3.76382619e-01 -1.73690900e-01 1.81229860e-01
-7.47257233e-01 1.47068238e+00 3.43360782e-01 5.68956852e-01
-3.53187382e-01 -1.53518021e-01 8.19439590e-01 5.09832501e-01
8.86834919e-01 4.39074606e-01 -1.44447729e-01 -1.78904578e-01
-7.69986272e-01 3.71087790e-01 1.32773495e+00 -1.65565640e-01
3.30774307e-01 -5.06057858e-01 -8.15342516e-02 8.39714706e-01
-2.65159104e-02 1.58868834e-01 4.51791555e-01 -5.92531860e-01
6.50823891e-01 1.02062047e+00 2.14372426e-01 -1.50609529e+00
-6.75083160e-01 -3.62648219e-01 -6.79258943e-01 -1.94900483e-02
6.98312342e-01 -3.60111922e-01 -4.48654853e-02 1.37431180e+00
5.00003040e-01 -2.93447971e-01 1.22228287e-01 6.34427845e-01
7.92104602e-01 4.20165807e-01 -1.36399910e-01 -3.34343076e-01
1.53048909e+00 -5.16547382e-01 -7.41015315e-01 -2.89975735e-03
9.41381633e-01 -6.46404147e-01 7.37808883e-01 6.63368762e-01
-1.03474021e+00 1.39286727e-01 -1.24438977e+00 -1.57228410e-01
-4.56812233e-01 4.49248925e-02 6.04590237e-01 7.67540991e-01
-3.55334014e-01 8.22803020e-01 -3.55783075e-01 -1.03764050e-02
7.58117378e-01 1.24279134e-01 -2.90309340e-01 2.86548227e-01
-1.53615749e+00 1.16924977e+00 4.99118641e-02 -1.06363468e-01
-1.11769304e-01 -3.34835015e-02 -8.51231575e-01 8.65052938e-02
5.55098414e-01 -3.62173915e-01 1.11705852e+00 -1.04347217e+00
-1.23558140e+00 1.29861557e+00 -2.51863953e-02 -1.07566428e+00
8.22344303e-01 -2.67773092e-01 -4.31169957e-01 2.04268813e-01
1.44645035e-01 -1.55996114e-01 1.06934881e+00 -8.67668688e-01
-3.88858974e-01 -8.46147299e-01 4.58464921e-01 -1.53568178e-01
-5.42438984e-01 2.70760715e-01 4.74113733e-01 -3.28331023e-01
-1.43370956e-01 -7.98438311e-01 1.79091260e-01 -5.35893857e-01
-9.18929219e-01 -7.76483238e-01 8.10154736e-01 -4.87983882e-01
1.26372349e+00 -1.78078735e+00 -1.07351884e-01 4.82975602e-01
6.12056851e-01 5.46483040e-01 5.17853141e-01 6.92339599e-01
1.74421728e-01 3.98238242e-01 -1.64681390e-01 -1.32519394e-01
6.56295195e-02 -3.07741016e-03 -4.44014937e-01 7.08423972e-01
1.69192001e-01 8.33211184e-01 -8.04172158e-01 -4.40774977e-01
-1.12614915e-01 3.36172968e-01 -2.15647981e-01 -1.29421234e-01
-8.99742171e-02 2.65554667e-01 -4.77355987e-01 1.02939479e-01
4.32551295e-01 -3.77523154e-01 4.25099909e-01 -1.23477049e-01
5.48215210e-02 8.89140427e-01 -6.12752676e-01 7.14876413e-01
-6.49086475e-01 1.00381279e+00 6.72472566e-02 -1.37375808e+00
7.91115463e-01 1.40036359e-01 -3.06404456e-02 -5.54448247e-01
6.89869881e-01 2.56030411e-01 3.20091814e-01 -6.33637309e-01
2.24847004e-01 -4.52741869e-02 -1.49694547e-01 1.03082228e+00
-4.06132579e-01 1.04726635e-01 4.65489119e-01 4.70369637e-01
1.09448481e+00 -6.27594709e-01 6.68168783e-01 -1.91797987e-01
9.09931302e-01 -1.41499057e-01 1.95391968e-01 6.40325487e-01
-6.90662786e-02 5.76281891e-05 1.45035827e+00 -7.12358236e-01
-7.91392028e-01 -2.59200156e-01 -4.02468480e-02 8.71511698e-01
-3.07114013e-02 -4.71064329e-01 -7.73258090e-01 -1.23815775e+00
9.38177407e-02 8.07413638e-01 -9.05400515e-01 -1.67545080e-01
-6.66425943e-01 -8.73390436e-01 5.90964854e-01 -9.48625500e-04
6.51222467e-01 -8.11993420e-01 -9.06712949e-01 9.78718549e-02
-4.79485333e-01 -1.01079750e+00 2.19473168e-01 2.29701713e-01
-4.57164973e-01 -1.72158420e+00 -4.53271717e-02 -2.14875281e-01
2.45177820e-01 8.23948383e-02 1.11856449e+00 5.15876174e-01
1.40105218e-01 -6.79754764e-02 -5.01283526e-01 -6.95478678e-01
-8.51824105e-01 2.34790564e-01 -2.56572068e-01 3.92685443e-01
3.67937475e-01 -5.37629783e-01 -3.06221426e-01 1.43399984e-02
-8.83873701e-01 -2.74963051e-01 1.84084754e-02 7.69617200e-01
-4.73570168e-01 -1.38227344e-01 6.90116346e-01 -1.60285044e+00
1.28043389e+00 -6.76768601e-01 -2.88814485e-01 -1.12844512e-01
-4.50191706e-01 1.08718805e-01 5.84020853e-01 3.35803069e-02
-8.34357560e-01 -6.38896823e-01 -3.40690672e-01 4.24143523e-01
-9.26980935e-03 7.13305056e-01 1.58345506e-01 3.00666422e-01
9.97552156e-01 -3.95185411e-01 2.98442870e-01 -1.23305388e-01
1.07072197e-01 8.62477779e-01 -3.90795693e-02 -2.82176793e-01
6.65390074e-01 7.85769284e-01 1.25781074e-01 -7.81500161e-01
-1.54696965e+00 -4.58244205e-01 -4.25730199e-01 -8.65708590e-02
4.25658286e-01 -3.99520814e-01 -1.22783518e+00 3.52515399e-01
-1.48701608e+00 1.55039066e-02 1.86861366e-01 7.41703436e-02
-1.31486237e-01 5.26278615e-01 -7.61562467e-01 -1.02534366e+00
-4.07606483e-01 -9.81623352e-01 6.25331581e-01 -1.82169691e-01
-6.50856197e-01 -1.16342986e+00 2.25936860e-01 9.11427677e-01
8.43997449e-02 5.09256124e-01 9.01085794e-01 -1.46078825e+00
4.98133630e-01 -4.75319535e-01 -1.88253611e-01 3.73528153e-01
-4.25730981e-02 6.91116601e-03 -9.31695044e-01 1.33785173e-01
3.02014410e-01 -6.11804485e-01 1.07079101e+00 5.36028668e-02
7.82772183e-01 -9.91160810e-01 -2.11646095e-01 -1.92568421e-01
6.96799576e-01 -4.46792454e-01 4.24673319e-01 5.76513767e-01
4.23410743e-01 9.37539637e-01 4.46174353e-01 5.14212430e-01
2.97170460e-01 7.22191691e-01 5.55106044e-01 4.69773710e-02
3.27310115e-01 5.35409749e-02 3.09198052e-01 1.52939081e-01
-4.94604744e-02 -3.73285413e-01 -9.05239105e-01 3.30625385e-01
-1.97374237e+00 -1.26241577e+00 -8.09175789e-01 1.90560734e+00
8.97567511e-01 4.65848505e-01 4.87176061e-01 8.79259169e-01
6.81051493e-01 3.83878142e-01 -8.96458477e-02 -9.29103494e-01
-2.16475666e-01 1.43888578e-01 1.04579382e-01 6.31431162e-01
-1.22018850e+00 5.37240863e-01 4.75835514e+00 5.60082793e-01
-1.11001003e+00 2.42257968e-01 9.94139135e-01 2.14564670e-02
-2.95234710e-01 -2.05357358e-01 -4.24460948e-01 4.89048511e-01
8.71590316e-01 1.59748614e-01 -9.25665423e-02 6.81633532e-01
3.30321461e-01 -4.30257529e-01 -1.00892293e+00 6.37837112e-01
3.18169475e-01 -1.53274310e+00 -2.19827369e-01 2.31252536e-01
4.57718104e-01 -1.97493955e-01 -4.35549989e-02 3.15567367e-02
3.07658643e-01 -8.88097227e-01 7.10581183e-01 -3.34103219e-03
-4.39420864e-02 -8.50053728e-01 1.03810656e+00 5.82892060e-01
-1.05051234e-01 -2.78876841e-01 1.99882574e-02 -5.12888193e-01
1.66834235e-01 1.26070786e+00 -9.76701498e-01 3.16769153e-01
1.64663270e-01 7.27479696e-01 -5.22398949e-01 3.72792184e-01
-8.03524733e-01 9.98241484e-01 -2.09833682e-01 -3.59977722e-01
4.57354546e-01 3.22777443e-02 7.33809173e-01 1.15863824e+00
-3.31603706e-01 -6.65997434e-03 -1.71911061e-01 7.16071665e-01
-3.45342815e-01 2.07886130e-01 -7.52543151e-01 -8.85596499e-03
4.61505353e-02 1.50632918e+00 -7.80490875e-01 -4.64373678e-01
-7.17000142e-02 6.24139905e-01 6.71649873e-01 -1.84637740e-01
-7.02995360e-01 -6.10029399e-02 1.75776273e-01 3.67787004e-01
7.97128752e-02 1.64018929e-01 -1.81466833e-01 -1.26681328e+00
1.45787090e-01 -9.17035222e-01 6.18150532e-01 -3.48407924e-01
-1.44822896e+00 5.07848620e-01 -3.34033072e-01 -7.54781306e-01
-4.59196657e-01 -5.43302476e-01 -9.85989928e-01 4.22086537e-01
-1.48567235e+00 -1.13872147e+00 -1.94861181e-02 5.02538919e-01
3.33121210e-01 1.09457761e-01 4.03325260e-01 -9.69584733e-02
-5.35435975e-01 3.53999674e-01 -2.91668773e-01 5.24718046e-01
4.92073804e-01 -1.04003978e+00 1.42386213e-01 6.46254957e-01
4.58034188e-01 5.10123432e-01 9.18258607e-01 -5.38511634e-01
-8.87572050e-01 -5.30318797e-01 1.14754748e+00 -5.67633450e-01
1.04699659e+00 -3.72211039e-01 -8.36102486e-01 3.42397898e-01
4.16424215e-01 -2.47008324e-01 9.16703939e-01 6.68045104e-01
-7.35753894e-01 1.56827480e-01 -1.12133181e+00 3.76535386e-01
6.62092268e-01 -3.89600068e-01 -7.82278061e-01 6.02784395e-01
1.66556358e-01 -5.45239858e-02 -3.55850339e-01 1.35509357e-01
5.27962267e-01 -1.43219757e+00 5.93778372e-01 -9.39178824e-01
9.37004983e-01 2.18501762e-01 3.38687152e-01 -1.37655032e+00
9.38976929e-02 -5.46921015e-01 -2.06377938e-01 1.13871098e+00
6.05585635e-01 -8.91554832e-01 5.53614378e-01 3.00188124e-01
2.53588587e-01 -7.42948830e-01 -8.55352998e-01 -2.74442375e-01
1.53375819e-01 -3.23800266e-01 3.22526902e-01 1.31652248e+00
7.37568736e-01 8.60768020e-01 -1.49535805e-01 -2.95737624e-01
3.32001328e-01 4.12619472e-01 7.29040742e-01 -1.56102812e+00
-1.31487370e-01 -5.94484568e-01 -3.09227914e-01 -4.47192967e-01
5.97980440e-01 -8.77117455e-01 -5.59651077e-01 -1.29378963e+00
1.99589759e-01 -1.09181412e-01 1.38847098e-01 2.36325577e-01
-6.41566142e-02 3.36215287e-01 5.89933358e-02 2.78755456e-01
-5.37332416e-01 1.49503350e-01 8.47544968e-01 -3.83083880e-01
-4.36405875e-02 3.41488719e-01 -8.55633259e-01 1.10233414e+00
8.97327960e-01 -3.95822704e-01 -1.43866679e-02 7.69196004e-02
1.05182719e+00 -1.92111477e-01 5.90357363e-01 -5.25231242e-01
2.18059458e-02 2.34969482e-01 1.11321099e-02 -3.84623021e-01
1.16525732e-01 -4.50084805e-01 -6.28365338e-01 5.25807559e-01
-6.24439836e-01 -1.00743711e-01 -1.73177823e-01 5.36975980e-01
-3.05209786e-01 -5.51611960e-01 6.08901560e-01 -1.02630727e-01
2.74442136e-01 -3.03640455e-01 -6.50243580e-01 2.63646245e-01
6.76163018e-01 1.83640644e-01 -6.27823532e-01 -7.06419706e-01
-7.27733374e-01 -7.35921413e-02 1.49333641e-01 8.67192745e-02
2.48941913e-01 -8.05058420e-01 -1.09140384e+00 -2.80970514e-01
-4.86047342e-02 -4.44600940e-01 -1.38105169e-01 1.38448167e+00
-3.11038315e-01 2.33782604e-01 1.81236520e-01 -3.31728578e-01
-1.58288097e+00 4.38605487e-01 1.20675646e-01 -7.17002451e-01
-3.51466238e-01 5.33045471e-01 -2.71287352e-01 8.79670307e-02
-1.24444596e-01 -1.98179349e-01 -8.49354029e-01 9.08538163e-01
7.42118418e-01 5.75422823e-01 5.86320281e-01 -7.11028755e-01
-2.76517868e-01 -1.50528952e-01 -4.31717098e-01 -1.60529941e-01
1.44266641e+00 1.18291587e-01 -5.59532762e-01 6.61428452e-01
1.11451852e+00 5.44358253e-01 -4.17203605e-01 -6.42175004e-02
2.91416049e-01 -2.87425786e-01 6.35298993e-03 -8.52927685e-01
-8.04951906e-01 7.76875675e-01 -3.71567070e-01 1.21502352e+00
4.73526597e-01 1.98510230e-01 5.83363295e-01 4.71984655e-01
-7.63178542e-02 -1.01517022e+00 5.30203097e-02 6.43988550e-01
9.40147519e-01 -1.37833214e+00 2.74946213e-01 -6.01901889e-01
-5.40020704e-01 1.24005449e+00 7.56311268e-02 -2.33441189e-01
5.34678340e-01 4.90539074e-02 -6.59440383e-02 -6.67310059e-01
-6.88086629e-01 -3.65839829e-03 1.81691602e-01 1.58662707e-01
6.16729558e-01 -8.30809921e-02 -1.00226510e+00 5.11719525e-01
-3.77510339e-01 -4.00674164e-01 9.55421865e-01 6.40836358e-01
-3.00187111e-01 -1.04837728e+00 -3.53438824e-01 6.59139454e-01
-9.82150018e-01 -8.67928043e-02 -1.23052037e+00 7.27568865e-01
-4.27768007e-02 1.39197731e+00 -1.78333953e-01 -1.41494066e-01
9.52598304e-02 1.06427200e-01 3.37935299e-01 -4.83041674e-01
-1.27072215e+00 -4.75012898e-01 1.06790137e+00 -3.14186573e-01
-7.64105022e-01 -8.07878554e-01 -9.45657611e-01 -5.62390745e-01
-5.06138086e-01 4.44070101e-01 6.40565097e-01 1.51026273e+00
2.05111220e-01 8.96719024e-02 6.20784402e-01 -3.38780046e-01
-5.32013655e-01 -1.00404382e+00 -5.88831231e-02 7.61339545e-01
4.71173942e-01 -7.19401658e-01 -5.80326140e-01 -3.47687364e-01] | [8.528620719909668, 10.275663375854492] |
7d4ca4f6-c4bf-4fdb-9d75-d54b158a4132 | learning-deep-multiresolution-representations | 2102.08423 | null | https://arxiv.org/abs/2102.08423v1 | https://arxiv.org/pdf/2102.08423v1.pdf | Learning deep multiresolution representations for pansharpening | Retaining spatial characteristics of panchromatic image and spectral information of multispectral bands is a critical issue in pansharpening. This paper proposes a pyramid based deep fusion framework that preserves spectral and spatial characteristics at different scales. The spectral information is preserved by passing the corresponding low resolution multispectral image as residual component of the network at each scale. The spatial information is preserved by training the network at each scale with the high frequencies of panchromatic image alongside the corresponding low resolution multispectral image. The parameters of different networks are shared across the pyramid in order to add spatial details consistently across scales. The parameters are also shared across fusion layers within a network at a specific scale. Experiments suggest that the proposed architecture outperforms state of the art pansharpening models. The proposed model, code and dataset is publicly available at https://github.com/sohaibali01/deep_pyramid_fusion. | ['Junaid Imtiaz', 'Muhammad Imran Qureshi', 'Syed Abdul Mannan Kirmani', 'Muhammad Mohsin Riaz', 'Syed Sohaib Ali', 'Hannan Adeel'] | 2021-02-16 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 2.53953904e-01 -7.13838220e-01 -1.15885846e-01 -1.59360662e-01
-6.94339931e-01 -8.19160700e-01 2.41963312e-01 -2.46779561e-01
-3.05910349e-01 6.16831779e-01 3.97583187e-01 1.25956059e-01
-3.87109071e-01 -1.29246736e+00 -5.91892064e-01 -9.77496564e-01
1.22608103e-01 -4.67402399e-01 2.66816705e-01 -5.53563654e-01
-1.51994973e-01 7.96835542e-01 -1.49180949e+00 7.41005838e-01
9.55173612e-01 9.96813238e-01 5.08040786e-01 8.79146278e-01
4.13950264e-01 5.63894391e-01 -1.14192575e-01 -1.10710777e-01
9.08358395e-01 -6.55883029e-02 -5.49764335e-01 1.30983174e-01
7.36388683e-01 -4.63674396e-01 -7.69757569e-01 1.72926724e+00
4.35730219e-01 2.93084055e-01 6.47499710e-02 -9.17109489e-01
-1.05845940e+00 7.50387192e-01 -1.32209778e+00 5.78522444e-01
-3.88342172e-01 -7.47549683e-02 9.45973217e-01 -6.22212052e-01
1.81024045e-01 9.77315128e-01 9.45707321e-01 -2.68524975e-01
-1.02753115e+00 -7.64775455e-01 -1.47634953e-01 2.62921989e-01
-1.50270784e+00 -1.07165471e-01 9.81188953e-01 -1.71307087e-01
6.70251548e-01 3.45087290e-01 8.20226848e-01 2.97846109e-01
6.79622531e-01 3.32305342e-01 1.13351226e+00 -2.38400415e-01
-2.92791545e-01 -3.19642693e-01 2.85927057e-01 4.42619085e-01
2.95982718e-01 3.98371458e-01 -4.50576603e-01 1.12332761e-01
1.02963483e+00 4.06305224e-01 -4.60672021e-01 -3.49205658e-02
-9.43372607e-01 7.96505511e-01 1.18626571e+00 7.51032174e-01
-9.38740075e-01 -3.63667160e-02 6.65594637e-02 5.93765855e-01
6.44985497e-01 2.54081041e-02 -5.72231889e-01 7.68932939e-01
-1.19928646e+00 3.05664808e-01 4.35899310e-02 5.69094241e-01
1.23444772e+00 1.33795977e-01 1.32670760e-01 1.14449084e+00
9.77808759e-02 4.68709350e-01 4.56007510e-01 -1.02868521e+00
2.42939487e-01 5.19839823e-01 -5.79110309e-02 -1.20514405e+00
-5.81422865e-01 -7.25334823e-01 -1.35609257e+00 3.51395816e-01
-1.41982213e-01 -2.61687577e-01 -1.10467100e+00 1.52288473e+00
2.36649424e-01 2.45429412e-01 2.09293012e-02 1.01012969e+00
7.51895308e-01 1.19146562e+00 -6.00420684e-02 2.19523837e-03
1.45537674e+00 -1.12402678e+00 -6.40678525e-01 -2.28960350e-01
-2.04304103e-02 -1.07275772e+00 5.51203847e-01 9.41674411e-02
-9.62633193e-01 -9.74935234e-01 -1.16720223e+00 -1.70874164e-01
-8.62685680e-01 3.41654897e-01 6.59463465e-01 5.76736391e-01
-1.36765540e+00 6.18692040e-01 -5.81872821e-01 -2.99422383e-01
2.75564939e-01 1.51597932e-01 -4.31132644e-01 -5.71793728e-02
-1.40686297e+00 6.74412608e-01 9.02546525e-01 4.24568683e-01
-5.25734723e-01 -9.55667734e-01 -7.15804040e-01 1.19370900e-01
-1.81373715e-01 -3.81185800e-01 9.92188096e-01 -1.16626811e+00
-9.79513764e-01 5.49650073e-01 1.48658037e-01 -3.12169462e-01
1.04493953e-01 -2.88988590e-01 -9.78237629e-01 2.62637228e-01
2.11707458e-01 5.83480477e-01 1.00013745e+00 -1.26364076e+00
-1.09927118e+00 -3.99893850e-01 9.10056606e-02 4.47169393e-01
-1.48840532e-01 1.42206863e-01 -1.29098088e-01 -1.01815319e+00
3.60373050e-01 -6.34855330e-01 -9.80931371e-02 -1.82957187e-01
-7.24151134e-02 3.66530478e-01 1.12085915e+00 -1.35420704e+00
8.66481006e-01 -2.29259706e+00 -2.54189838e-02 1.60461903e-01
-2.88806502e-02 3.23409677e-01 -5.26822925e-01 3.74070406e-01
-5.21434605e-01 -1.65593594e-01 -5.22490680e-01 3.95981908e-01
-2.74380803e-01 -3.56539451e-02 -3.33674729e-01 7.38661051e-01
-2.23994516e-02 8.30714405e-01 -5.08970678e-01 -1.39556557e-01
4.90065455e-01 9.53991711e-01 -2.29161546e-01 -5.37529029e-02
2.29377910e-01 2.10347414e-01 -1.70770213e-01 5.81622362e-01
1.60478079e+00 -5.03933318e-02 -1.23402543e-01 -9.08922732e-01
-6.47218943e-01 -4.20802444e-01 -1.33724713e+00 1.71746087e+00
-2.77189672e-01 3.47913772e-01 5.81929922e-01 -8.26501310e-01
6.97603762e-01 4.52746153e-01 5.11778235e-01 -8.46816182e-01
4.55506034e-02 1.28484935e-01 -1.29524812e-01 -4.32306714e-03
9.32513237e-01 -2.53268778e-01 2.61993885e-01 1.61868691e-01
7.86114186e-02 -4.38483626e-01 1.86471105e-01 -1.51989415e-01
2.50366032e-01 7.70077389e-03 2.68669397e-01 -3.78065407e-01
6.96541190e-01 7.97969401e-02 5.97422600e-01 7.00240254e-01
-2.63852686e-01 5.73603570e-01 -2.98400432e-01 -6.48630142e-01
-1.31952393e+00 -1.03271818e+00 -1.96463719e-01 1.46396792e+00
1.26900867e-01 1.80900887e-01 -4.67435926e-01 -4.78464812e-02
-1.44350119e-02 3.39330256e-01 -7.62978971e-01 1.05638094e-01
-4.72302675e-01 -1.17636490e+00 3.03673327e-01 3.65538985e-01
1.34310532e+00 -8.51446331e-01 -4.25364822e-01 2.12580532e-01
-3.15815955e-01 -8.09056163e-01 -4.33658808e-01 1.80794179e-01
-7.85281658e-01 -9.67781007e-01 -8.10516357e-01 -8.13374698e-01
1.17232822e-01 8.99897277e-01 8.19073915e-01 -2.94837445e-01
-1.85405552e-01 -9.88343954e-02 -3.62233847e-01 -2.36978218e-01
-2.76005805e-01 1.16369963e-01 -3.94355536e-01 2.26360530e-01
1.45465672e-01 -9.36409414e-01 -7.61408150e-01 -1.07053682e-01
-1.46122670e+00 1.61085173e-01 6.88673437e-01 6.20959044e-01
6.71491742e-01 9.63521659e-01 3.59599769e-01 -3.74855280e-01
5.48049033e-01 -3.68455201e-01 -8.71083915e-01 2.51733184e-01
1.20234728e-01 -5.47098696e-01 6.29301071e-01 -3.23077291e-02
-1.44729650e+00 4.05827314e-02 -1.37135431e-01 -4.71926620e-03
-3.01374644e-01 5.63091993e-01 1.28059134e-01 -4.57368702e-01
6.87164426e-01 4.40328121e-01 -3.77046973e-01 -7.25332797e-01
5.82214117e-01 6.12727821e-01 8.92423987e-01 -8.19071084e-02
1.01586723e+00 8.64406288e-01 -1.45566717e-01 -9.72180426e-01
-1.20607066e+00 -5.86000204e-01 -9.53582406e-01 -2.18258113e-01
8.78108025e-01 -1.46753752e+00 -6.64030761e-02 9.94508803e-01
-8.51543546e-01 -1.41373128e-01 -1.55189231e-01 2.68158048e-01
-1.69397533e-01 3.11599970e-01 -7.90046632e-01 -3.43137026e-01
-5.40937364e-01 -6.93035960e-01 9.13085461e-01 7.50451028e-01
4.94729519e-01 -8.96123469e-01 2.74856389e-01 2.73357481e-01
7.58784294e-01 3.69974524e-01 6.17298424e-01 1.09550068e-02
-2.95484841e-01 -1.29100353e-01 -7.51071095e-01 5.80455184e-01
6.04573548e-01 4.38327380e-02 -8.89820278e-01 -4.99608964e-01
2.88831025e-01 -2.00932592e-01 1.35375988e+00 9.49739039e-01
1.04058588e+00 -2.48102739e-01 1.30054489e-01 9.70605731e-01
1.88752174e+00 4.72341850e-02 5.22678912e-01 5.76536119e-01
6.89286172e-01 3.83084089e-01 1.00833438e-01 1.79366097e-02
1.11170873e-01 5.08597344e-02 4.23161656e-01 -5.91000021e-01
-2.93613523e-01 8.90277773e-02 3.67857218e-02 4.60993707e-01
-1.99220806e-01 2.19108090e-01 -8.06615829e-01 7.27996349e-01
-1.79105866e+00 -1.38554120e+00 -2.21662521e-01 1.68678951e+00
7.89926648e-01 -4.82648760e-01 -8.64895359e-02 7.16651008e-02
1.04154694e+00 8.37112248e-01 -3.09762329e-01 -9.03499573e-02
-8.52572799e-01 3.46811861e-01 9.70734119e-01 7.98910558e-01
-1.70334041e+00 8.34751725e-01 5.87966299e+00 6.35969937e-01
-1.24601948e+00 2.90142357e-01 4.43072945e-01 8.71075168e-02
5.05301505e-02 -1.76174924e-01 -1.76262483e-01 9.78348553e-02
6.96538806e-01 -2.17816934e-01 6.13540471e-01 4.09045219e-01
1.79423317e-01 -2.47517124e-01 -1.46393642e-01 6.77209854e-01
-2.32417092e-01 -1.46013129e+00 2.32128024e-01 -2.41064385e-01
1.26332104e+00 7.25670576e-01 2.83185691e-01 -8.81056339e-02
5.73212922e-01 -8.29647422e-01 4.93239939e-01 5.03201902e-01
3.71023566e-01 -1.05707741e+00 8.11443925e-01 1.55486777e-01
-1.74104774e+00 -4.64801490e-01 -7.79017270e-01 5.97374737e-02
4.78413999e-02 6.40131652e-01 2.05776125e-01 1.14003980e+00
1.01575756e+00 9.16579843e-01 -5.92192590e-01 1.10298300e+00
3.48139834e-03 3.71921897e-01 -2.48362526e-01 1.01747119e+00
6.12242818e-01 -5.25923014e-01 3.65120798e-01 1.37031162e+00
5.63564897e-01 2.36430913e-01 1.55652449e-01 5.89095354e-01
-8.01909640e-02 -1.07675597e-01 -4.84451056e-01 1.88666552e-01
2.52944171e-01 1.49104536e+00 -4.65796322e-01 -3.68889362e-01
-5.80999136e-01 7.83154130e-01 -1.18243106e-01 6.15415752e-01
-6.99436009e-01 -2.97467828e-01 6.64792061e-01 -4.63459134e-01
5.90473831e-01 -1.06830010e-02 -3.46129686e-01 -1.11494553e+00
-2.30606690e-01 -9.12915885e-01 7.52317011e-01 -1.09547901e+00
-1.17228055e+00 5.88019729e-01 -9.20551494e-02 -9.85760272e-01
5.38704753e-01 -5.22163808e-01 -4.85034168e-01 1.45951021e+00
-2.06830859e+00 -1.61908817e+00 -5.28709471e-01 9.87673700e-01
2.53289878e-01 -8.25017542e-02 6.02160573e-01 2.48627320e-01
-3.74387801e-01 -8.19725543e-02 4.71659511e-01 8.86375681e-02
4.93343890e-01 -9.51222301e-01 3.05099517e-01 1.50865161e+00
-1.69497028e-01 2.20765397e-01 6.49270236e-01 -6.44065619e-01
-9.19122517e-01 -1.50087249e+00 6.39917731e-01 4.00834203e-01
8.96856844e-01 3.36346507e-01 -1.21862650e+00 7.08476067e-01
7.83991814e-01 5.03346734e-02 6.49143636e-01 -4.51904625e-01
-6.01497531e-01 -4.71641511e-01 -1.25304258e+00 1.90722436e-01
4.52751458e-01 -7.28074431e-01 -4.97194141e-01 4.53240335e-01
5.91493905e-01 -2.34258100e-01 -1.02005351e+00 5.82184672e-01
4.91630465e-01 -1.27982712e+00 1.10991812e+00 -1.86679047e-02
5.45953333e-01 -5.31233847e-01 -5.40677786e-01 -1.60599887e+00
-1.23846865e+00 -1.95760742e-01 3.19123656e-01 7.48235583e-01
1.98002905e-01 -7.21963882e-01 4.30081934e-01 -8.24996755e-02
-2.01518219e-02 4.74732257e-02 -7.18449771e-01 -4.55100596e-01
4.10623044e-01 8.81738774e-03 8.75522494e-01 1.19485521e+00
-6.64424777e-01 -1.78450048e-01 -4.86567765e-01 1.04436159e+00
9.10702109e-01 5.63904941e-01 1.37475312e-01 -1.15301502e+00
-5.49589731e-02 -5.65974057e-01 -2.19551995e-02 -4.54654008e-01
-5.75798750e-02 -8.62492979e-01 -3.35814357e-01 -1.69585884e+00
3.49388361e-01 2.33281121e-01 -6.04473412e-01 7.14864492e-01
-1.37795299e-01 7.86819637e-01 2.27690175e-01 1.54133484e-01
2.11738199e-01 3.80462617e-01 1.34541357e+00 -3.11797768e-01
-1.91953510e-01 -3.07409674e-01 -8.04045916e-01 7.25393653e-01
1.47784638e+00 -2.68982112e-01 -1.38100401e-01 -7.79164612e-01
-6.13803156e-02 6.68383986e-02 7.02237189e-01 -1.22972846e+00
2.90832937e-01 -2.77029753e-01 8.33251297e-01 -8.79319727e-01
3.41896743e-01 -1.10963798e+00 6.29093647e-01 4.62765187e-01
-1.09591313e-01 2.06852723e-02 6.71684802e-01 1.87960505e-01
-3.70708048e-01 1.35949314e-01 1.36125779e+00 -2.70789742e-01
-1.13937807e+00 4.09346461e-01 -6.52063712e-02 -6.48463488e-01
7.08963931e-01 -1.88161999e-01 -5.72052002e-01 -1.46091565e-01
-6.52363062e-01 2.68539451e-02 3.81417394e-01 5.15403569e-01
3.75616103e-01 -1.32621098e+00 -1.10096741e+00 4.29123133e-01
-1.03610441e-01 -3.91189247e-01 9.75111246e-01 5.15299082e-01
-7.30054200e-01 2.60640681e-01 -1.02462387e+00 -1.18099593e-01
-1.21562648e+00 3.28323990e-01 9.38645899e-01 -1.82929084e-01
-5.85445046e-01 8.52769613e-01 2.19888046e-01 -5.51604092e-01
-3.50764543e-01 -1.38507485e-01 -2.61200845e-01 1.52341992e-01
8.11399162e-01 2.41910562e-01 1.25495195e-01 -1.24643683e+00
-1.52505174e-01 6.89386904e-01 -5.76132387e-02 -1.59213804e-02
1.89571977e+00 -2.38719404e-01 -7.91870356e-01 3.52362543e-02
1.20827579e+00 -1.05415925e-01 -1.33415890e+00 -6.34036303e-01
-5.01459420e-01 -6.85212851e-01 6.60956264e-01 -9.72558618e-01
-1.60344112e+00 6.34315431e-01 9.25394833e-01 3.22868675e-01
1.78436232e+00 -3.94393831e-01 5.46213210e-01 2.24557053e-02
-1.68033484e-02 -1.22852921e+00 -5.28637946e-01 5.55774450e-01
1.07062376e+00 -1.05500185e+00 1.97287574e-01 -2.82625914e-01
-3.45040768e-01 1.08723676e+00 3.22166353e-01 -4.28140461e-01
8.93422425e-01 2.52149731e-01 3.16380084e-01 -1.82527438e-01
-1.60806224e-01 -1.79737672e-01 4.09729570e-01 6.16948307e-01
5.87411940e-01 4.08362657e-01 -3.62360962e-02 4.46774475e-02
-4.55823064e-01 -1.22531271e-03 3.45576733e-01 7.61147439e-01
-7.94310033e-01 -7.06556499e-01 -7.38651872e-01 3.60361606e-01
-4.91112947e-01 -2.80452222e-01 -1.20550096e-01 6.04891002e-01
4.19504404e-01 9.82604086e-01 2.01457348e-02 -1.57218471e-01
3.38263899e-01 -4.39622924e-02 9.80623811e-02 -1.00182049e-01
-8.60286891e-01 2.20244557e-01 -3.37285399e-01 -5.91410339e-01
-9.43071783e-01 -4.25870746e-01 -6.33808553e-01 -6.27043545e-01
2.90654805e-02 9.60124955e-02 3.56282234e-01 3.79900128e-01
1.46052212e-01 5.17307758e-01 6.89042628e-01 -1.09338582e+00
-1.87794313e-01 -9.56715882e-01 -1.11811709e+00 2.55025953e-01
7.56776810e-01 -1.17218882e-01 -1.22042522e-01 2.17384011e-01] | [10.152706146240234, -1.9261870384216309] |
545123bc-af56-4b9f-bf26-9dee4f369caf | unnoticeable-backdoor-attacks-on-graph-neural | 2303.01263 | null | https://arxiv.org/abs/2303.01263v1 | https://arxiv.org/pdf/2303.01263v1.pdf | Unnoticeable Backdoor Attacks on Graph Neural Networks | Graph Neural Networks (GNNs) have achieved promising results in various tasks such as node classification and graph classification. Recent studies find that GNNs are vulnerable to adversarial attacks. However, effective backdoor attacks on graphs are still an open problem. In particular, backdoor attack poisons the graph by attaching triggers and the target class label to a set of nodes in the training graph. The backdoored GNNs trained on the poisoned graph will then be misled to predict test nodes to target class once attached with triggers. Though there are some initial efforts in graph backdoor attacks, our empirical analysis shows that they may require a large attack budget for effective backdoor attacks and the injected triggers can be easily detected and pruned. Therefore, in this paper, we study a novel problem of unnoticeable graph backdoor attacks with limited attack budget. To fully utilize the attack budget, we propose to deliberately select the nodes to inject triggers and target class labels in the poisoning phase. An adaptive trigger generator is deployed to obtain effective triggers that are difficult to be noticed. Extensive experiments on real-world datasets against various defense strategies demonstrate the effectiveness of our proposed method in conducting effective unnoticeable backdoor attacks. | ['Suhang Wang', 'Xiang Zhang', 'Minhua Lin', 'Enyan Dai'] | 2023-02-11 | null | null | null | null | ['graph-classification'] | ['graphs'] | [ 2.97927320e-01 2.57965386e-01 -4.50480282e-01 1.89997420e-01
-2.02242926e-01 -1.37612879e+00 5.78603625e-01 1.59217983e-01
-3.33822444e-02 6.77025378e-01 -3.87133390e-01 -5.88644862e-01
1.28537016e-02 -1.41508365e+00 -7.40959942e-01 -7.91587651e-01
-4.44333166e-01 2.66304404e-01 6.61241531e-01 -2.24835873e-01
4.11688685e-02 7.92891085e-01 -6.40368760e-01 -1.59257382e-01
6.71550035e-01 2.42056251e-01 -4.64161187e-01 6.68996751e-01
8.78504887e-02 7.58750498e-01 -1.16324043e+00 -6.21453285e-01
5.90739429e-01 -4.96196926e-01 -5.37955821e-01 -2.22478881e-01
3.70783135e-02 -2.86645234e-01 -8.77913535e-01 1.54360187e+00
4.47218060e-01 -1.50430888e-01 2.43087575e-01 -1.78049207e+00
-3.43247920e-01 1.14505816e+00 -6.22692168e-01 4.57821488e-01
2.64251500e-01 4.67586666e-01 7.69044340e-01 -2.28572771e-01
4.26797599e-01 1.12403476e+00 3.46537679e-01 8.06838810e-01
-1.01241195e+00 -1.28216302e+00 2.12703452e-01 -1.09355740e-01
-1.29531920e+00 -1.48951858e-01 1.08604360e+00 7.11283684e-02
4.58287179e-01 7.01549709e-01 4.25119758e-01 1.37997568e+00
2.74777383e-01 3.70866120e-01 9.16586339e-01 -1.45999223e-01
2.77680933e-01 1.86846957e-01 1.03100307e-01 8.35216820e-01
7.08928704e-01 5.16912937e-01 -2.13144273e-01 -7.94966340e-01
7.27592945e-01 1.46713957e-01 -3.46888304e-01 2.93987896e-02
-7.63235986e-01 1.16477895e+00 9.40537870e-01 1.58214852e-01
-2.46413872e-01 3.09407473e-01 4.06346560e-01 3.55563760e-01
3.13024670e-01 5.98598599e-01 -1.70131087e-01 6.08497918e-01
-4.25065607e-01 7.44991824e-02 1.01549017e+00 6.52484596e-01
5.30979514e-01 3.18874538e-01 -1.93592668e-01 2.09247589e-01
1.61845267e-01 6.61601961e-01 2.97135059e-02 -1.83806077e-01
4.41361517e-01 7.69375563e-01 -2.44197309e-01 -1.58774555e+00
-2.25508377e-01 -6.96105421e-01 -7.70946145e-01 -9.87170264e-02
1.93026096e-01 -4.21112806e-01 -1.19382095e+00 1.82070947e+00
7.46653855e-01 7.48603582e-01 -7.39546716e-02 7.07278371e-01
5.56007206e-01 6.56136811e-01 1.52652249e-01 -1.76367596e-01
1.16771007e+00 -6.43947601e-01 -5.52918911e-01 -2.42332280e-01
5.17107427e-01 -1.71819717e-01 9.03283477e-01 7.98203349e-02
-5.23878694e-01 2.32583851e-01 -1.24347687e+00 9.74743187e-01
-5.10435641e-01 -6.70029879e-01 9.07590032e-01 1.17874420e+00
-6.42021358e-01 4.56322104e-01 -8.52575779e-01 -1.89444542e-01
4.93494928e-01 6.17533088e-01 -2.20673829e-01 -2.36765772e-01
-1.53357232e+00 2.67717153e-01 6.46131575e-01 -3.98268923e-02
-1.81009376e+00 -2.53411204e-01 -6.39742792e-01 -1.65679097e-01
1.00188971e+00 -2.24371612e-01 7.52367258e-01 -3.94009858e-01
-9.96878684e-01 3.91462624e-01 5.55034459e-01 -7.18249500e-01
2.51845658e-01 3.88879955e-01 -5.42192876e-01 4.72852439e-01
7.97701702e-02 1.01018563e-01 1.02403057e+00 -1.19647264e+00
-2.41053522e-01 -3.00265104e-01 5.70922971e-01 -9.41412598e-02
-8.78370464e-01 1.08613685e-01 -1.15139283e-01 -7.32890010e-01
4.37937789e-02 -1.02235210e+00 -5.08529067e-01 -6.52662873e-01
-1.18702817e+00 -1.20570682e-01 1.08551598e+00 -8.35569799e-02
1.18101466e+00 -1.79914713e+00 -2.45542675e-01 5.69544375e-01
6.49707794e-01 4.95599627e-01 -2.06832021e-01 6.44054949e-01
-6.35593897e-03 6.97276056e-01 -8.64923447e-02 3.86341721e-01
-2.81210154e-01 2.68189967e-01 -8.58796120e-01 6.79052889e-01
-2.32679173e-01 7.93953598e-01 -1.04187381e+00 -2.36689791e-01
4.24531139e-02 1.45810768e-01 -3.98562908e-01 2.95635492e-01
-3.65036815e-01 2.15951145e-01 -8.98255646e-01 9.73687649e-01
6.12425387e-01 -2.41378054e-01 1.79753572e-01 6.02287147e-03
6.71728969e-01 1.13340184e-01 -8.94325316e-01 4.76177365e-01
1.94011703e-01 -1.98291056e-02 -9.44964290e-02 -6.52893066e-01
9.46496308e-01 4.12322134e-01 1.92292929e-01 -1.12746328e-01
3.37503523e-01 1.42380431e-01 1.94889665e-01 -1.02227323e-01
-1.73563473e-02 -3.60174328e-02 -3.12705785e-01 6.50242746e-01
-1.60706803e-01 2.64294475e-01 2.98213344e-02 6.82825804e-01
1.77008975e+00 -7.92316973e-01 1.08192258e-01 4.15421091e-02
5.12049556e-01 2.06795111e-01 5.79729438e-01 1.16070807e+00
-2.77545363e-01 -7.39205405e-02 7.44779289e-01 -5.25167882e-01
-5.03302932e-01 -1.10385740e+00 4.02013868e-01 9.00023818e-01
2.96919137e-01 -5.47005236e-01 -9.04756367e-01 -1.54572928e+00
-7.40039721e-02 4.88349319e-01 -5.99458218e-01 -7.88753390e-01
-5.73015332e-01 -9.05138910e-01 1.25721049e+00 1.63863212e-01
5.81239164e-01 -1.23766911e+00 -9.40192789e-02 2.19240397e-01
2.75035352e-01 -1.13945556e+00 -5.29379427e-01 1.83108211e-01
-6.95915043e-01 -1.65494335e+00 -7.48701245e-02 -4.83310461e-01
1.06435549e+00 5.19011199e-01 6.72707498e-01 5.15778780e-01
-2.27369457e-01 1.12279825e-01 -4.35373396e-01 -3.86811912e-01
-6.90681934e-01 2.10478261e-01 3.10990810e-01 -5.97953685e-02
2.43040845e-01 -7.33960032e-01 -4.09588963e-01 4.27531183e-01
-1.15161204e+00 -6.72389865e-01 4.16300356e-01 5.65161228e-01
3.80644321e-01 7.05403566e-01 6.76759541e-01 -1.38763702e+00
9.35197055e-01 -8.03615630e-01 -8.42313707e-01 1.92154661e-01
-4.02249575e-01 -1.44366831e-01 1.10232651e+00 -1.02557588e+00
-1.55758023e-01 -1.30515635e-01 1.81755006e-01 -8.31978261e-01
1.41973421e-01 3.28051180e-01 -4.76279944e-01 -6.44360483e-01
9.89977896e-01 6.99901655e-02 -5.19723952e-01 1.35454703e-02
3.79931420e-01 1.13530874e-01 3.56709599e-01 -4.55881894e-01
1.70924997e+00 4.08234805e-01 3.41665596e-01 -5.15805662e-01
-6.27066612e-01 1.65442616e-01 2.63306469e-01 -2.30988115e-01
4.45569873e-01 -3.65101069e-01 -7.97918737e-01 2.75532395e-01
-9.10799980e-01 -8.41591209e-02 4.18494083e-02 4.74824198e-03
3.64057899e-01 3.17186266e-01 -7.53463686e-01 -8.24713707e-01
-6.87089086e-01 -9.84552026e-01 2.99792141e-01 3.93176824e-01
2.22587824e-01 -1.13211727e+00 -1.29344538e-02 3.57008469e-03
2.49982625e-01 8.92114997e-01 9.05809879e-01 -1.39672935e+00
-9.25750613e-01 -8.36673141e-01 1.91726372e-01 -1.57738142e-02
3.54784876e-01 -1.25452846e-01 -7.13974655e-01 -6.20476902e-01
1.64472088e-01 -3.22475970e-01 6.90783918e-01 2.14064354e-03
1.10362411e+00 -8.62975538e-01 -8.31017494e-01 7.35429406e-01
1.35447490e+00 2.84688860e-01 3.92976165e-01 1.64647460e-01
1.04211843e+00 2.66658634e-01 3.67965966e-01 1.57885075e-01
-2.27130726e-01 1.81163922e-01 9.94268537e-01 1.59040932e-02
3.28083068e-01 -8.26180160e-01 4.20228004e-01 2.66874492e-01
3.91577035e-01 -7.81862557e-01 -7.83977687e-01 2.66926497e-01
-1.36016846e+00 -7.14908123e-01 2.69421078e-02 2.17782784e+00
6.71950221e-01 6.04517043e-01 2.36056954e-01 3.66525531e-01
1.02785230e+00 4.59275991e-01 -6.49305999e-01 -1.58531234e-01
4.54879552e-02 2.40948945e-01 9.52715278e-01 4.19977546e-01
-1.07423949e+00 1.33879626e+00 5.83509350e+00 9.37847376e-01
-1.07715166e+00 9.80526209e-02 5.49891651e-01 1.15505308e-01
-3.98271561e-01 4.11666602e-01 -8.58347833e-01 3.71664166e-01
8.06784093e-01 -3.95429164e-01 6.55505002e-01 8.83967757e-01
-1.89332411e-01 4.36778128e-01 -7.96237111e-01 3.94689888e-01
-1.24365211e-01 -1.23668396e+00 2.68717378e-01 2.84749597e-01
5.96509576e-01 -9.11251977e-02 -6.70329183e-02 3.38706642e-01
8.60264421e-01 -1.16754842e+00 9.54436809e-02 -2.59232491e-01
6.00545406e-01 -1.17039442e+00 4.05287325e-01 5.06704986e-01
-1.17680383e+00 -3.56565833e-01 -5.09845793e-01 2.75235116e-01
-1.76223293e-01 6.80266380e-01 -1.19153881e+00 4.58260775e-01
2.72673488e-01 8.61192048e-02 -6.45709574e-01 7.21258521e-01
-8.08738470e-01 1.13744748e+00 -5.16624153e-01 -2.03418598e-01
3.76194626e-01 6.27809614e-02 1.08774555e+00 6.25600338e-01
-1.72645196e-01 1.78392202e-01 6.91325068e-01 9.34955239e-01
-4.84577328e-01 -1.53110042e-01 -1.36021364e+00 -5.33517241e-01
1.00938141e+00 1.44428790e+00 -1.17835307e+00 -5.36812209e-02
2.99094290e-01 7.20483840e-01 2.40887582e-01 4.55270439e-01
-1.15530825e+00 -7.05614507e-01 3.51135612e-01 2.21847922e-01
-2.39224061e-01 -8.43858048e-02 1.09671511e-01 -8.61646056e-01
-3.17314655e-01 -1.17482543e+00 7.85495043e-01 -1.81176007e-01
-1.26352942e+00 8.17583501e-01 -6.12357557e-02 -8.22976947e-01
9.19193178e-02 -1.70583382e-01 -1.14468062e+00 4.55044717e-01
-8.84703100e-01 -1.11451972e+00 -1.20161638e-01 1.10511243e+00
4.84834947e-02 -3.06202501e-01 6.72178686e-01 5.32233939e-02
-8.92380953e-01 9.64824200e-01 -6.77564144e-01 4.93439972e-01
1.76416248e-01 -9.43945944e-01 5.00387192e-01 1.37893367e+00
3.74242306e-01 7.63562024e-01 8.74822319e-01 -1.32629800e+00
-1.63903880e+00 -1.31486058e+00 1.43798858e-01 -2.98136026e-01
7.37886786e-01 -6.97056830e-01 -7.97626257e-01 8.59040022e-01
-4.51416820e-02 3.89589697e-01 4.54660207e-01 -3.84249181e-01
-5.17473817e-01 -2.05860864e-02 -1.60717523e+00 9.17439461e-01
9.72954571e-01 -4.35806096e-01 7.90455844e-03 9.35633004e-01
1.26263726e+00 -2.72672772e-01 -3.26805115e-01 4.22806233e-01
-1.68297738e-01 -4.72036272e-01 9.37587082e-01 -7.97356725e-01
-1.65776700e-01 -3.16953242e-01 9.87793431e-02 -1.03032923e+00
-6.02527745e-02 -1.14250624e+00 -4.24667984e-01 1.29176831e+00
2.48198479e-01 -1.05199635e+00 1.14227068e+00 1.23422161e-01
4.44234252e-01 -6.96142256e-01 -6.82826638e-01 -8.32827568e-01
-2.67045856e-01 -1.22153699e-01 7.98504055e-01 1.17712903e+00
-1.93683669e-01 2.83643067e-01 -5.63189447e-01 8.69983673e-01
1.09416068e+00 -3.85955542e-01 8.18683863e-01 -1.01784515e+00
-3.36102486e-01 1.08447656e-01 -4.02616620e-01 -2.54490703e-01
1.88689038e-01 -1.14761627e+00 -3.70283991e-01 -9.69968319e-01
-1.36582516e-02 -5.05373359e-01 -3.60711098e-01 8.78062487e-01
-3.87314409e-01 3.65634054e-01 1.79178081e-02 6.27690852e-02
-2.95209318e-01 1.33342892e-01 1.07094896e+00 -3.27764899e-01
-3.71068418e-01 2.92320907e-01 -8.27543914e-01 5.74980080e-01
1.08293664e+00 -1.16592515e+00 -8.80489647e-01 2.01428562e-01
7.93308169e-02 1.43315271e-01 4.80577260e-01 -7.76378155e-01
2.67807066e-01 -2.62609154e-01 1.19828902e-01 -5.26057184e-01
-7.85473883e-02 -6.62568212e-01 1.69662803e-01 1.00422657e+00
-9.42812636e-02 1.15410298e-01 7.40991309e-02 9.75220859e-01
3.15968871e-01 -9.15207639e-02 6.99220717e-01 -1.89037591e-01
-1.10923111e-01 7.70526588e-01 -8.33257064e-02 5.50571978e-02
1.33498240e+00 -9.27001014e-02 -6.97402179e-01 -3.78839403e-01
-3.82142752e-01 3.47134233e-01 4.07841772e-01 2.20657170e-01
7.99472153e-01 -1.10709023e+00 -5.04408717e-01 3.23880285e-01
-1.85537532e-01 -3.02387327e-01 -1.12913959e-01 3.29204351e-01
-3.71641964e-01 1.14603356e-01 7.70244077e-02 -2.41936415e-01
-1.26590168e+00 1.12131441e+00 4.48796183e-01 -6.08908892e-01
-6.31458580e-01 1.00155568e+00 8.13196227e-02 -4.23141539e-01
3.21404010e-01 3.83645773e-01 -9.61143970e-02 -5.03730714e-01
3.67358416e-01 3.12713623e-01 -3.34405303e-01 -1.54879868e-01
-5.61553359e-01 1.92466192e-02 -3.93339306e-01 9.39801931e-02
9.24167037e-01 5.33991039e-01 -2.68830091e-01 -2.69409180e-01
9.30183351e-01 3.90953362e-01 -7.54800141e-01 -3.31533072e-03
-3.41857262e-02 -5.35962820e-01 -1.97256699e-01 -6.87008500e-01
-1.46242869e+00 4.31195080e-01 2.86967576e-01 7.87162364e-01
1.12884986e+00 -9.73168015e-02 9.48729098e-01 5.74369013e-01
6.01578534e-01 -3.36772323e-01 3.56386006e-01 1.10811815e-01
3.77497315e-01 -5.83836377e-01 -1.04675665e-01 -8.71615231e-01
-2.54956156e-01 6.98420048e-01 9.35923457e-01 -5.89636326e-01
4.79821205e-01 4.25347775e-01 -1.74511001e-01 -5.81037879e-01
-6.18057966e-01 3.00920010e-01 -2.88363546e-01 6.45820737e-01
-4.30750608e-01 1.69102073e-01 -1.15921222e-01 5.65218866e-01
-1.06512882e-01 -5.37529111e-01 8.17111850e-01 8.91076863e-01
-6.10285282e-01 -1.28635108e+00 -6.49274468e-01 4.06665683e-01
-7.77142704e-01 -3.45408730e-02 -9.51118886e-01 7.00069726e-01
-2.74823129e-01 1.15463662e+00 -7.23214924e-01 -9.57585394e-01
1.29978448e-01 -2.62399584e-01 1.94539353e-01 -7.22409248e-01
-8.41350257e-01 -1.30257860e-01 -3.20794918e-02 -4.43281651e-01
3.03942978e-01 2.85303593e-01 -1.20850623e+00 -4.63241696e-01
-7.24975467e-01 5.37780523e-01 3.28077704e-01 5.45005083e-01
1.14768296e-01 4.61114585e-01 1.22566080e+00 -2.34630540e-01
-7.67036498e-01 -6.91322267e-01 -6.89051926e-01 7.67432004e-02
4.08900566e-02 -6.60008609e-01 -8.27400625e-01 -6.67226911e-01] | [6.091660976409912, 7.340043067932129] |
5accf2d0-972b-4361-a605-f738d85c0aaa | re-evaluating-lidar-scene-flow-for-autonomous | 2304.0215 | null | https://arxiv.org/abs/2304.02150v1 | https://arxiv.org/pdf/2304.02150v1.pdf | Re-Evaluating LiDAR Scene Flow for Autonomous Driving | Current methods for self-supervised LiDAR scene flow estimation work poorly on real data. A variety of flaws in common evaluation protocols have caused leading approaches to focus on problems that do not exist in real data. We analyze a suite of recent works and find that despite their focus on deep learning, the main challenges of the LiDAR scene flow problem -- removing the dominant rigid motion and robustly estimating the simple motions that remain -- can be more effectively solved with classical techniques such as ICP motion compensation and enforcing piecewise rigid assumptions. We combine these steps with a test-time optimization method to form a state-of-the-art system that does not require any training data. Because our final approach is dataless, it can be applied on different datasets with diverse LiDAR rigs without retraining. Our proposed approach outperforms all existing methods on Argoverse 2.0, halves the error rate on NuScenes, and even rivals the performance of supervised networks on Waymo and lidarKITTI. | ['Simon Lucey', 'Deva Ramanan', 'Nathaniel Chodosh'] | 2023-04-04 | null | null | null | null | ['motion-compensation', 'scene-flow-estimation'] | ['computer-vision', 'computer-vision'] | [-6.50512725e-02 -3.82498831e-01 -3.84440541e-01 -5.20243466e-01
-5.88805676e-01 -6.65501058e-01 5.22194862e-01 -5.68627179e-01
-5.85930228e-01 7.87396669e-01 7.63716102e-02 -2.56872416e-01
-2.18214784e-02 -5.74596286e-01 -6.15977228e-01 -5.14583051e-01
7.46810064e-02 9.13615704e-01 6.75297081e-01 -3.21661204e-01
3.29141766e-01 8.28293085e-01 -1.72300363e+00 -2.31191933e-01
7.91630626e-01 4.73236531e-01 7.10061863e-02 8.42003882e-01
-2.66605943e-01 9.18299317e-01 -3.35141122e-01 -1.71103477e-01
7.21291542e-01 6.71772379e-03 -1.03048778e+00 -1.25811910e-02
1.57016349e+00 -6.09253705e-01 -5.52831173e-01 5.81182182e-01
7.52976418e-01 2.33078375e-01 6.09535098e-01 -1.27695000e+00
-1.52870163e-01 -1.64558999e-02 -6.24823749e-01 2.11008236e-01
9.80769321e-02 4.70846176e-01 8.02962303e-01 -9.35844362e-01
9.83928502e-01 1.46606910e+00 1.08309197e+00 4.57506865e-01
-1.45447528e+00 -7.82857835e-01 -7.86681771e-02 1.64381698e-01
-1.22799265e+00 -7.42869139e-01 7.84754872e-01 -7.09325731e-01
1.16034818e+00 -2.87348893e-03 5.82984209e-01 1.02732050e+00
-3.64696085e-02 6.90961659e-01 8.22482884e-01 8.91288891e-02
-6.68574572e-02 -3.48325998e-01 8.09077825e-03 8.59423280e-01
2.69315749e-01 3.09222430e-01 -6.24577880e-01 1.65823236e-01
6.13597870e-01 -3.67262751e-01 -2.18492553e-01 -8.95057440e-01
-1.43435287e+00 8.00991654e-01 5.68459332e-01 -1.26901746e-01
3.76053713e-02 3.94433647e-01 5.71638048e-01 4.65342887e-02
4.20872509e-01 1.16713680e-01 -5.33445656e-01 -2.30303138e-01
-1.40607798e+00 6.05826676e-01 8.91127050e-01 1.07030654e+00
1.14552641e+00 3.50916713e-01 3.02831590e-01 5.07107794e-01
3.80316436e-01 6.74717188e-01 1.68477282e-01 -1.53674626e+00
6.65533185e-01 3.91234517e-01 1.10708132e-01 -8.68115425e-01
-4.34553623e-01 2.00314354e-03 -6.33145094e-01 5.51157773e-01
6.21304870e-01 -1.26395941e-01 -9.10964131e-01 1.31312108e+00
5.30145943e-01 5.42712986e-01 -9.92018357e-03 8.68171930e-01
1.00709426e+00 4.13827151e-01 -8.72518346e-02 1.64933041e-01
6.90458834e-01 -1.24347055e+00 -5.40174484e-01 -5.73526680e-01
6.42924964e-01 -9.48471248e-01 1.21357620e+00 1.61940962e-01
-9.79140282e-01 -8.15015972e-01 -1.21296823e+00 -6.16870761e-01
-3.24221462e-01 -3.90987191e-03 1.00365913e+00 5.52962065e-01
-1.11633897e+00 9.78721857e-01 -8.79568160e-01 -5.16693354e-01
6.54154420e-01 4.93674964e-01 -5.96456885e-01 -7.00875893e-02
-4.56498891e-01 9.80811954e-01 2.96907216e-01 2.85432160e-01
-9.56373870e-01 -1.13223398e+00 -9.71413434e-01 -4.28504616e-01
2.76828587e-01 -9.77401912e-01 1.37207067e+00 -2.81832844e-01
-1.58447886e+00 1.04976559e+00 -4.21116114e-01 -3.25605333e-01
1.11780572e+00 -6.62016094e-01 4.61668149e-02 -5.77419251e-02
3.37033063e-01 1.29284477e+00 6.92766607e-01 -1.09957421e+00
-7.24386573e-01 -6.78606555e-02 -1.39951259e-01 1.92555830e-01
1.09791741e-01 -3.03991288e-01 -5.10223985e-01 -1.37156233e-01
7.26917237e-02 -1.10730076e+00 -2.47058883e-01 3.65772873e-01
-3.85830134e-01 2.45611090e-02 1.44772220e+00 -2.21995652e-01
6.40034616e-01 -1.77123213e+00 1.56565905e-01 -3.55629385e-01
1.93687771e-02 4.80205745e-01 -5.59526533e-02 1.80825666e-01
-1.70684773e-02 2.05859598e-02 -5.16273379e-01 -7.52272606e-01
-2.14066863e-01 5.99191904e-01 -4.72080916e-01 8.77521992e-01
3.54503751e-01 9.97637808e-01 -9.72333014e-01 -6.93902135e-01
8.54353368e-01 4.91559386e-01 -6.83089137e-01 1.76730186e-01
-1.78342655e-01 7.90606558e-01 4.18022126e-02 5.83705068e-01
1.02533710e+00 5.85937798e-02 -2.70859867e-01 -2.97639638e-01
-4.08853471e-01 5.99646330e-01 -1.59873891e+00 2.18063259e+00
-3.79759610e-01 1.09036851e+00 3.65952961e-02 -8.58844161e-01
6.82447910e-01 -9.38861817e-02 8.20142746e-01 -3.05753052e-01
-1.78559288e-01 3.38629574e-01 -1.66532278e-01 -6.96521640e-01
6.20059431e-01 4.47875373e-02 3.29260796e-01 1.34040505e-01
2.80963272e-01 -6.78020895e-01 3.69645625e-01 -1.13668246e-02
9.05971706e-01 1.03819823e+00 -7.22942874e-02 -3.31608564e-01
5.98331034e-01 6.09903514e-01 6.74495161e-01 5.90084374e-01
-4.94675875e-01 9.69200909e-01 -1.71117182e-03 -7.65572786e-01
-1.13809538e+00 -9.82569218e-01 -3.03767115e-01 7.51092851e-01
2.55954862e-01 -3.79573077e-01 -3.19612622e-01 -7.32898891e-01
3.40303332e-01 3.96624744e-01 -3.07014048e-01 2.55876899e-01
-1.03817880e+00 -6.95444345e-01 7.72573829e-01 6.36034310e-01
7.99516499e-01 -9.36330855e-01 -9.43131268e-01 1.99967995e-01
-2.62453496e-01 -1.54435921e+00 -2.69032061e-01 9.00535658e-02
-1.20088530e+00 -1.21294439e+00 -2.53667265e-01 -4.83007073e-01
1.82967201e-01 5.51075816e-01 1.45367503e+00 -4.77605965e-03
-3.84164512e-01 2.46365845e-01 2.55850971e-01 -3.58020425e-01
-1.01061374e-01 4.36136603e-01 -4.20026015e-03 -4.74391848e-01
2.73201436e-01 -6.84900522e-01 -6.56049430e-01 4.31849390e-01
-5.44065893e-01 -1.30759686e-01 1.10444367e-01 4.86866593e-01
3.78571838e-01 -3.62223119e-01 1.74599569e-02 -8.51439953e-01
4.74516191e-02 -6.22559823e-02 -8.53572249e-01 -2.92480379e-01
-4.47757334e-01 1.14913322e-01 1.66944236e-01 -2.82058537e-01
-1.22013938e+00 5.09333432e-01 -7.79262483e-02 -5.88023603e-01
-3.40570897e-01 -1.37233697e-02 3.37536745e-02 -5.21168232e-01
7.78674662e-01 -2.41103113e-01 -1.22404240e-01 -4.30215031e-01
6.26588702e-01 2.30618939e-01 1.02312505e+00 -5.15059054e-01
1.29835355e+00 1.23130155e+00 4.31572199e-01 -1.13319385e+00
-9.35671568e-01 -8.03597152e-01 -1.29125094e+00 -2.27301478e-01
9.34015214e-01 -1.11598241e+00 -7.61504531e-01 4.81285632e-01
-1.30191267e+00 -4.88689244e-01 -4.46433067e-01 5.06903291e-01
-6.72095716e-01 5.67763805e-01 -3.85772854e-01 -5.46669900e-01
-1.98380947e-01 -1.30145907e+00 1.43659174e+00 9.83860344e-02
-3.24227631e-01 -1.10105944e+00 5.51956534e-01 4.66532141e-01
5.10132074e-01 4.57533628e-01 2.04730406e-01 4.55972329e-02
-9.83305156e-01 5.68448752e-02 -1.70095474e-01 2.53459096e-01
1.39306620e-01 6.08902276e-01 -1.32723653e+00 -2.22366840e-01
-1.81196615e-01 -5.20390868e-01 1.01207137e+00 4.15495485e-01
8.62212658e-01 2.17044204e-01 -1.37945443e-01 1.33734918e+00
1.40599513e+00 -3.75952423e-01 7.24920034e-01 5.96265614e-01
1.17082298e+00 6.74602926e-01 7.13076830e-01 -3.24342325e-02
6.22026443e-01 6.66689992e-01 7.85768092e-01 -2.54156977e-01
-4.35472310e-01 -3.22834462e-01 3.16220969e-01 6.07387722e-01
-3.63896728e-01 1.96531434e-02 -1.00801528e+00 6.94637656e-01
-2.03849435e+00 -9.77353275e-01 -7.35912323e-01 1.99247944e+00
4.23181683e-01 7.45505467e-02 -4.94445767e-03 -4.32656780e-02
1.88972309e-01 5.36536157e-01 -5.40787697e-01 -8.94197375e-02
-1.89632043e-01 3.31527412e-01 8.19221675e-01 8.20387721e-01
-1.33921504e+00 1.59126842e+00 7.27678442e+00 3.35212588e-01
-1.23718822e+00 5.08902185e-02 -1.30689606e-01 -1.67306527e-01
1.18425235e-01 4.50614959e-01 -1.11285746e+00 -2.66454685e-02
7.85203278e-01 2.40271956e-01 1.35417104e-01 8.54387403e-01
4.60907966e-01 -1.94602892e-01 -1.27230179e+00 9.33043242e-01
-1.00008056e-01 -1.53698003e+00 -2.16847420e-01 8.15862417e-03
6.67639077e-01 6.68450773e-01 -3.17878991e-01 2.88482845e-01
5.86803734e-01 -9.75305736e-01 6.56496346e-01 3.29418927e-01
6.30864561e-01 -3.10081214e-01 5.65272689e-01 2.94312119e-01
-1.37763071e+00 3.06017011e-01 -6.10269189e-01 -1.96631089e-01
4.60605085e-01 3.24339360e-01 -7.57455468e-01 7.04009712e-01
8.78784359e-01 1.14228380e+00 -6.92571461e-01 1.22213030e+00
-2.80464321e-01 5.55920780e-01 -6.96480453e-01 5.31898856e-01
3.37487817e-01 -3.45086604e-01 6.59617782e-01 1.43304574e+00
1.36669040e-01 -5.45429528e-01 2.70384490e-01 7.23483443e-01
-4.02032807e-02 -7.20865726e-02 -9.66322780e-01 5.64556181e-01
2.07314879e-01 1.33371735e+00 -5.76092362e-01 -2.82518566e-01
-2.89239883e-01 8.39526772e-01 4.34732735e-01 2.34008566e-01
-8.32274318e-01 7.69929681e-03 8.89144540e-01 2.17909604e-01
2.33223528e-01 -6.21546984e-01 -2.93464333e-01 -1.30253589e+00
7.04609901e-02 -4.83848155e-01 2.80912966e-01 -9.34427619e-01
-9.91509736e-01 2.25225702e-01 3.12330306e-01 -1.38614130e+00
-3.63606334e-01 -6.22896373e-01 -7.02636182e-01 7.13789999e-01
-2.04314804e+00 -1.28400624e+00 -7.45048285e-01 6.35140836e-01
7.14851558e-01 3.68618853e-02 4.71966535e-01 6.28283858e-01
-4.93659943e-01 2.80300565e-02 -3.26653838e-01 2.55505983e-02
1.01532519e+00 -1.27700150e+00 7.05427885e-01 1.03456903e+00
2.45800257e-01 3.32654089e-01 7.12344408e-01 -6.08846962e-01
-1.27356112e+00 -1.14526033e+00 9.06382322e-01 -8.89739335e-01
6.97332501e-01 -3.19662005e-01 -1.01274753e+00 8.95464242e-01
2.17547163e-01 4.63432729e-01 1.41985506e-01 -2.34574705e-01
-2.61723876e-01 -2.41327509e-01 -9.97683346e-01 5.95148623e-01
1.67054081e+00 -3.40929449e-01 -5.96247852e-01 5.04026651e-01
5.70330441e-01 -8.83906007e-01 -5.48315406e-01 6.73877120e-01
3.80624622e-01 -1.17591298e+00 1.24894440e+00 -8.71679068e-01
3.69497597e-01 -6.72256231e-01 -7.44081661e-02 -9.43388224e-01
-1.12923197e-01 -9.03868556e-01 -1.76870659e-01 1.11063957e+00
-1.58264153e-02 -4.21276003e-01 1.05747330e+00 3.67385983e-01
-2.48104721e-01 -1.04822703e-01 -1.06874573e+00 -7.39093244e-01
3.07572559e-02 -6.35512531e-01 2.92317331e-01 1.03998983e+00
-8.66181970e-01 5.10495484e-01 -5.05414963e-01 1.91159874e-01
8.34116697e-01 5.04575931e-02 1.57323349e+00 -1.46951628e+00
1.64307639e-01 -2.29055360e-01 -5.33823907e-01 -1.20483696e+00
4.47510481e-01 -8.76006305e-01 1.07781999e-01 -1.45292270e+00
-2.97338337e-01 -3.85558784e-01 4.50294375e-01 6.27430797e-01
-5.75051196e-02 4.18575108e-01 1.57385498e-01 5.27182758e-01
-3.48218709e-01 5.90954304e-01 1.23529410e+00 -1.34374201e-01
-3.30658644e-01 -2.59736538e-01 -2.14823663e-01 1.11389208e+00
5.85586250e-01 -6.59655511e-01 -4.81130779e-01 -7.45453119e-01
1.04282692e-01 -4.22645718e-01 5.38889647e-01 -1.28483331e+00
2.46129498e-01 -2.75148362e-01 1.30166948e-01 -1.07013738e+00
3.29811782e-01 -8.32234025e-01 2.52045870e-01 3.52279991e-01
1.82296693e-01 4.27012324e-01 4.24446791e-01 3.58711720e-01
-3.40007246e-02 -9.91584733e-02 8.93878281e-01 -6.29722700e-02
-9.91768122e-01 4.28810984e-01 3.79343703e-02 2.71440536e-01
7.67139971e-01 -4.17701870e-01 -4.62980300e-01 -1.69510260e-01
-2.48122960e-01 3.81273508e-01 5.42379558e-01 6.26439810e-01
3.07186007e-01 -1.18043160e+00 -8.09707820e-01 1.38588697e-01
1.39045613e-02 7.63929069e-01 -1.54167181e-02 7.66040981e-01
-1.16320395e+00 3.50482196e-01 -2.10892096e-01 -1.33200026e+00
-1.07640994e+00 -6.39127791e-02 6.62234902e-01 -1.43139973e-01
-9.32335913e-01 6.56401217e-01 -1.64987162e-01 -9.30552602e-01
8.51982012e-02 -3.16391736e-01 -1.44437850e-02 4.37511094e-02
4.65983488e-02 5.56843638e-01 2.97008008e-01 -9.88855720e-01
-5.03848970e-01 1.13615263e+00 2.31840879e-01 -6.85959682e-02
1.50544846e+00 1.24890767e-02 1.23532638e-02 4.73230302e-01
1.14922476e+00 1.47628948e-01 -1.57191086e+00 5.01113106e-03
-5.35005853e-02 -7.51219332e-01 1.94278330e-01 -1.69207752e-01
-1.27395475e+00 1.19471478e+00 6.62689388e-01 -2.95631409e-01
5.28546393e-01 -3.80066186e-01 5.69844306e-01 6.11327946e-01
1.86075553e-01 -9.57700968e-01 -3.51891778e-02 8.72170687e-01
6.13075078e-01 -1.60707879e+00 3.82619351e-01 -5.72672069e-01
-3.80652666e-01 1.13338304e+00 7.75664091e-01 -4.06727910e-01
4.82734114e-01 4.03956741e-01 3.36936772e-01 -2.52310753e-01
-4.22070801e-01 -3.62631470e-01 2.70569682e-01 7.71908939e-01
3.83426428e-01 -5.11378706e-01 3.03110015e-02 -6.28425956e-01
-5.43439090e-01 2.33211651e-01 4.81947273e-01 1.08184636e+00
-1.33380041e-01 -1.06784987e+00 -2.87407041e-01 -1.01590797e-03
-9.78350341e-02 1.75972581e-01 -3.89189094e-01 1.27562940e+00
1.48814306e-01 6.73450351e-01 1.95631921e-01 -2.02919170e-01
6.45229518e-01 5.53793162e-02 4.99599367e-01 -3.66936058e-01
-4.47066933e-01 -7.27977306e-02 1.83676854e-01 -1.07708466e+00
-1.01217234e+00 -8.15634131e-01 -1.05549526e+00 -6.94936991e-01
-2.83947229e-01 -4.19580251e-01 5.00315487e-01 8.96007001e-01
3.66314679e-01 3.60967815e-01 2.72225916e-01 -1.59604919e+00
-3.19833577e-01 -7.95335114e-01 1.35758379e-02 4.37748194e-01
6.45599663e-01 -8.62444580e-01 -5.40412545e-01 1.72358662e-01] | [8.513090133666992, -2.0340614318847656] |
4309305b-7ddd-46b6-94f8-f8b4b6f9da00 | context-aware-stand-alone-neural-spelling | 2011.06642 | null | https://arxiv.org/abs/2011.06642v1 | https://arxiv.org/pdf/2011.06642v1.pdf | Context-aware Stand-alone Neural Spelling Correction | Existing natural language processing systems are vulnerable to noisy inputs resulting from misspellings. On the contrary, humans can easily infer the corresponding correct words from their misspellings and surrounding context. Inspired by this, we address the stand-alone spelling correction problem, which only corrects the spelling of each token without additional token insertion or deletion, by utilizing both spelling information and global context representations. We present a simple yet powerful solution that jointly detects and corrects misspellings as a sequence labeling task by fine-turning a pre-trained language model. Our solution outperforms the previous state-of-the-art result by 12.8% absolute F0.5 score. | ['Liang Huang', 'Hairong Liu', 'Xiangci Li'] | 2020-11-12 | null | https://aclanthology.org/2020.findings-emnlp.37 | https://aclanthology.org/2020.findings-emnlp.37.pdf | findings-of-the-association-for-computational | ['spelling-correction'] | ['natural-language-processing'] | [ 5.18963516e-01 -2.29301378e-01 1.84552968e-01 -3.59842747e-01
-9.52488840e-01 -8.32466602e-01 3.53844613e-01 8.46181393e-01
-8.03273082e-01 7.15907335e-01 2.38555580e-01 -3.03484648e-01
6.52282834e-01 -5.72683215e-01 -7.38547206e-01 -4.51819450e-01
5.40586948e-01 1.20001964e-01 2.55960226e-01 -1.40571922e-01
5.28739810e-01 6.58461303e-02 -1.61558461e+00 6.22144818e-01
1.26572776e+00 2.59599745e-01 6.98862672e-01 8.11326325e-01
-5.22254527e-01 3.62175107e-01 -9.26231503e-01 -3.24138135e-01
-1.22393794e-01 -3.23146015e-01 -6.48881853e-01 -2.29897156e-01
6.59057200e-01 4.78643849e-02 1.36483535e-01 1.47060430e+00
4.12863731e-01 8.71798769e-02 3.39397550e-01 -4.54591036e-01
-1.18205965e+00 6.06620550e-01 -3.87865275e-01 1.87469095e-01
5.70893586e-01 3.82520892e-02 9.71564412e-01 -1.39591420e+00
2.25277603e-01 1.07932377e+00 5.69974244e-01 8.37460399e-01
-9.15652633e-01 -4.48688239e-01 3.46071929e-01 8.70318785e-02
-1.45884323e+00 -3.60946804e-01 2.12858334e-01 -1.70790434e-01
1.56942606e+00 2.90338218e-01 2.04946443e-01 8.56284261e-01
3.26315075e-01 8.43920231e-01 8.11369121e-01 -1.10977340e+00
1.58071533e-01 -2.23414570e-01 3.96664858e-01 6.78446412e-01
5.44248581e-01 -6.29076138e-02 -6.44285500e-01 2.62271110e-02
2.08703130e-01 -1.58688471e-01 -1.79578573e-01 2.71534175e-01
-1.24589193e+00 3.29387516e-01 -1.76938891e-01 4.59669203e-01
-3.31957549e-01 9.51741040e-02 4.51684117e-01 2.88503230e-01
5.23466527e-01 7.05730975e-01 -9.78678942e-01 -1.11154117e-01
-1.28262722e+00 1.90465659e-01 5.68017542e-01 9.57900882e-01
6.25640750e-01 9.83240083e-02 -5.39535105e-01 8.01903784e-01
1.89919919e-01 6.23262525e-01 8.35037351e-01 -1.61301866e-01
4.50847775e-01 4.44867641e-01 3.56037557e-01 -7.87965894e-01
-8.89056697e-02 -5.50430715e-01 -4.79361951e-01 -2.52444506e-01
4.39690620e-01 -4.86676320e-02 -1.22923434e+00 1.57765436e+00
2.81832386e-02 1.33398309e-01 -1.87701117e-02 6.87246859e-01
5.60805440e-01 7.54902005e-01 3.19447547e-01 -2.77060628e-01
1.37290657e+00 -1.14435911e+00 -8.44122112e-01 -7.79779971e-01
7.63918400e-01 -1.11049426e+00 1.26643312e+00 5.80303490e-01
-1.04356146e+00 -4.50794756e-01 -9.27621782e-01 -1.41664043e-01
-3.74498874e-01 5.31678617e-01 1.46644026e-01 5.59657991e-01
-1.01421010e+00 5.83816826e-01 -5.54493964e-01 -3.23936552e-01
-1.90837666e-01 6.83263317e-02 -2.27265567e-01 -2.77183950e-01
-1.24921715e+00 1.03035283e+00 5.77419043e-01 6.92220181e-02
-6.21976137e-01 -8.34173501e-01 -9.33118343e-01 1.95923463e-01
4.34764981e-01 -4.38175887e-01 1.64763117e+00 -1.11066246e+00
-1.11221921e+00 1.01585805e+00 -9.15151238e-01 -3.82211119e-01
3.27400684e-01 -6.56017244e-01 -6.79218948e-01 -1.93668574e-01
3.40837121e-01 2.35411853e-01 9.14012969e-01 -9.20947731e-01
-6.91326439e-01 -2.25848824e-01 -6.37589037e-01 3.02269962e-02
-1.16531804e-01 4.35341239e-01 -2.81455845e-01 -1.18377900e+00
9.68518630e-02 -5.20114481e-01 -1.51541293e-01 -4.76021141e-01
-3.14654976e-01 -2.69369304e-01 4.88383710e-01 -9.94297922e-01
1.53134406e+00 -2.02696300e+00 -2.09578559e-01 -1.90074872e-02
-1.15698293e-01 1.03833640e+00 -5.44169426e-01 4.40051347e-01
-1.37312263e-01 3.75843525e-01 -4.41715688e-01 -4.80181724e-01
-9.92152393e-02 1.78376809e-01 -5.08275330e-01 1.64344788e-01
3.98020387e-01 8.65422666e-01 -1.54519737e+00 -2.20977038e-01
1.50207371e-01 2.76825607e-01 -3.34244817e-01 1.45146742e-01
-2.65899539e-01 1.61523223e-02 2.69138187e-01 6.15249574e-01
7.91404188e-01 2.48240486e-01 2.88123816e-01 3.01007807e-01
-2.96285778e-01 8.56099010e-01 -1.20983672e+00 1.61167395e+00
-5.09833455e-01 4.41765845e-01 -7.55014420e-02 -6.93730831e-01
1.13379121e+00 3.30524027e-01 -3.51545602e-01 -5.58527350e-01
-2.07076222e-01 4.73716617e-01 -3.54691416e-01 -4.46267039e-01
9.30902302e-01 -2.71440297e-01 -2.99170911e-01 4.47569579e-01
1.09564550e-01 2.26262249e-02 2.10072547e-01 1.79230079e-01
1.15216208e+00 9.48198363e-02 7.87733436e-01 -1.55857965e-01
5.69654465e-01 -2.86954254e-01 8.03432763e-01 1.19353795e+00
-3.94167572e-01 7.71319270e-01 -1.23891979e-02 -4.09247965e-01
-7.96730757e-01 -1.12950706e+00 5.09742260e-01 1.40488100e+00
-1.34788185e-01 -7.50183463e-01 -8.65163326e-01 -8.62366021e-01
-1.07484646e-01 1.13110995e+00 -2.76560456e-01 -3.63251925e-01
-6.67991340e-01 -4.93116558e-01 7.89184809e-01 5.91986299e-01
2.47713089e-01 -1.23227608e+00 -3.67252767e-01 4.60372120e-01
-3.61131936e-01 -9.53707397e-01 -8.87420535e-01 2.03461796e-01
-5.63608229e-01 -6.33489490e-01 -5.35362542e-01 -1.14813936e+00
7.81956613e-01 4.62022275e-01 1.16339934e+00 6.43054545e-01
-4.68004271e-02 -3.71810459e-02 -6.85609698e-01 -4.32770550e-01
-6.53674245e-01 4.58570533e-02 3.41436356e-01 -2.45996878e-01
9.40012336e-01 -1.37589306e-01 -9.30717885e-02 -1.57516718e-01
-9.62005913e-01 -3.14079858e-02 7.24110126e-01 7.93425441e-01
5.64973176e-01 -7.21167624e-02 5.83982646e-01 -9.57917988e-01
6.37823105e-01 -1.59962773e-01 -3.93488616e-01 7.18550801e-01
-6.01129055e-01 2.97781885e-01 9.15792108e-01 -2.08063200e-01
-1.14145195e+00 4.48962361e-01 -2.79713064e-01 1.18289731e-01
-6.20033145e-01 3.47994626e-01 -2.35610634e-01 2.44628981e-01
6.10510230e-01 6.32419288e-01 -6.90916777e-01 -8.20918679e-01
4.14383799e-01 7.31279314e-01 1.11389005e+00 -4.03016955e-01
7.42052555e-01 -1.22538902e-01 -6.63989723e-01 -9.00375664e-01
-1.12672448e+00 -6.55518651e-01 -9.32830393e-01 6.04980513e-02
3.73105586e-01 -7.51039982e-01 -2.06748605e-01 6.86078668e-01
-1.79181528e+00 9.00573954e-02 -3.04656439e-02 2.86755562e-01
6.83235973e-02 8.78869057e-01 -6.43716991e-01 -8.01567078e-01
-5.80924690e-01 -6.00819826e-01 1.04001737e+00 6.05136901e-02
-6.15453720e-01 -6.62478328e-01 5.30039817e-02 1.07751057e-01
2.86652267e-01 -2.54220724e-01 8.17400217e-01 -8.38061571e-01
-9.81505811e-02 -3.52090001e-01 -1.22293904e-01 6.10822737e-01
1.78817019e-01 1.20594673e-01 -9.49542820e-01 -1.33487120e-01
-1.59056455e-01 1.65178910e-01 1.31792283e+00 3.96723039e-02
8.83993745e-01 -4.43557978e-01 -8.38072971e-02 2.25231782e-01
1.32714558e+00 -4.63253222e-02 4.27332938e-01 1.90619245e-01
7.57714808e-01 4.25867528e-01 7.93007851e-01 4.16130126e-01
1.92455202e-01 1.92870617e-01 1.46769240e-01 3.23904961e-01
-1.81541309e-01 -8.17298353e-01 6.20348573e-01 9.68316495e-01
3.84476691e-01 -4.25843865e-01 -1.14918280e+00 9.57600117e-01
-1.82282352e+00 -8.81102085e-01 -2.95237124e-01 2.31415319e+00
1.19402194e+00 1.45324767e-01 -4.48120892e-01 3.44328105e-01
1.17385340e+00 7.89914727e-02 -1.60994485e-01 -8.79818976e-01
-1.73004702e-01 3.90408874e-01 5.26078463e-01 7.95749724e-01
-1.01181221e+00 1.70179725e+00 6.77881956e+00 9.62498009e-01
-9.43018317e-01 -5.59451394e-02 1.04166903e-01 -6.98371977e-02
-4.48182762e-01 -1.54593969e-02 -1.02805257e+00 4.99761552e-01
9.25410330e-01 -6.84356764e-02 4.64720070e-01 7.28963315e-01
4.23276484e-01 -2.64211208e-01 -9.68253255e-01 7.83884883e-01
4.02271122e-01 -1.07872653e+00 3.06888223e-01 -6.79068148e-01
7.59018302e-01 -1.78565472e-01 -2.32047230e-01 3.07001710e-01
3.82177949e-01 -8.80205691e-01 1.02321875e+00 3.73300284e-01
7.53871262e-01 -7.29723692e-01 6.80838525e-01 6.28535628e-01
-1.09107697e+00 1.64098203e-01 -3.18372816e-01 -6.47398949e-01
8.94071758e-02 1.10171938e+00 -1.00462723e+00 2.82083720e-01
4.23089296e-01 5.85450828e-01 -6.51781082e-01 1.20289838e+00
-1.21851468e+00 9.72425222e-01 1.19330592e-01 -2.22239748e-01
5.96465580e-02 3.88180465e-01 5.97519159e-01 1.93940413e+00
4.77061957e-01 8.41807351e-02 1.40623152e-02 6.61420345e-01
-1.32162899e-01 1.60000876e-01 -2.97046334e-01 -1.01240888e-01
7.55040586e-01 1.08456588e+00 -6.19672120e-01 -6.04920387e-01
-2.03373820e-01 1.46329248e+00 5.71971834e-01 3.02508980e-01
-4.16164309e-01 -8.34752023e-01 8.05791676e-01 -3.27891499e-01
4.79230165e-01 -3.49331290e-01 -5.93527079e-01 -1.49467421e+00
1.66827515e-01 -7.77666569e-01 1.78737551e-01 -6.59859478e-01
-1.24026954e+00 2.74396002e-01 -7.38737226e-01 -8.92756641e-01
-6.95599541e-02 -6.41582489e-01 -7.21443772e-01 9.33800936e-01
-1.80994487e+00 -7.51452565e-01 1.16555363e-01 3.17405984e-02
7.83059597e-01 2.19220459e-01 9.23453867e-01 1.58815697e-01
-5.56733966e-01 7.79015899e-01 6.42309934e-02 2.88840085e-01
1.10560668e+00 -1.36980927e+00 9.74212110e-01 1.78596914e+00
2.95590401e-01 9.01012897e-01 1.00311685e+00 -9.17428255e-01
-9.36397910e-01 -1.39908326e+00 2.26557493e+00 -3.83445740e-01
4.32466149e-01 -4.60548401e-01 -1.22609842e+00 5.09682119e-01
8.59610066e-02 -1.19724087e-01 5.22476315e-01 -3.75012937e-03
-6.37374640e-01 1.25445023e-01 -1.18980360e+00 5.80743790e-01
1.00564635e+00 -6.94187462e-01 -1.28446770e+00 6.47236183e-02
9.37295735e-01 -3.13552409e-01 3.96427028e-02 2.96941679e-02
2.43580833e-01 -5.09356439e-01 6.04889512e-01 -6.27127707e-01
3.02032858e-01 -5.57941794e-01 -1.19294688e-01 -1.68806124e+00
-4.75046605e-01 -7.52116978e-01 1.97464198e-01 1.43445194e+00
4.75467950e-01 -1.86279967e-01 2.95104772e-01 5.88346899e-01
-3.04376334e-01 -1.15404233e-01 -7.40031064e-01 -8.59872997e-01
-1.42034842e-02 -7.39411533e-01 7.41021931e-01 8.27205956e-01
4.16498005e-01 2.19521746e-01 -4.99987274e-01 3.77541631e-01
2.99310207e-01 -1.79463834e-01 1.16786040e-01 -8.21517050e-01
-1.22909583e-01 -1.85603350e-01 -1.13454461e-02 -1.19972289e+00
4.09568101e-01 -8.56638908e-01 8.72755408e-01 -1.42468286e+00
-6.10542558e-02 -3.06731407e-02 -4.46679115e-01 7.91722894e-01
-7.00572908e-01 2.17532009e-01 2.41753966e-01 -3.52675587e-01
-8.22457492e-01 3.51561219e-01 6.71612501e-01 -5.02172671e-02
-1.30874366e-01 -1.48684487e-01 -7.89314210e-01 9.36381996e-01
9.29563403e-01 -8.00453722e-01 2.03176171e-01 -1.03685033e+00
6.07310951e-01 -3.80294949e-01 2.37516329e-01 -8.13129306e-01
4.16790932e-01 -3.04476410e-01 1.84395880e-01 -5.31647146e-01
-2.45899528e-01 -4.25670832e-01 -6.93988383e-01 4.60971951e-01
-6.26434028e-01 4.20493782e-01 3.31247687e-01 5.93913496e-01
-1.36910588e-01 -6.93570137e-01 5.48292935e-01 -5.10415435e-01
-1.08118176e+00 -2.27200463e-01 -9.75731552e-01 1.78761646e-01
6.81491137e-01 -2.57118661e-02 -4.15850192e-01 7.03966692e-02
-4.89030480e-01 -4.14509475e-02 4.32921946e-01 5.97146273e-01
8.05667400e-01 -9.84344602e-01 -8.83473337e-01 5.83820641e-01
1.84641287e-01 -2.55138218e-01 -3.84198911e-02 1.88646570e-01
-4.76052284e-01 5.08095384e-01 1.27828360e-01 -1.01456800e-02
-1.60404742e+00 5.80095649e-01 1.50953919e-01 -7.86520615e-02
-2.14869514e-01 1.14649737e+00 -3.02711040e-01 -6.79162920e-01
1.88408896e-01 -7.86613226e-01 6.56372830e-02 -1.29184395e-01
1.16950774e+00 2.72516936e-01 5.80862045e-01 -5.59777677e-01
-7.27615893e-01 2.76824564e-01 -3.59322488e-01 -6.94033504e-02
1.02529657e+00 -2.90634423e-01 -3.32190067e-01 3.42761964e-01
7.69837260e-01 4.23204392e-01 -6.74599469e-01 -4.50404584e-01
4.85900342e-01 -6.90083027e-01 -6.59186319e-02 -1.25723612e+00
-4.35771167e-01 9.62841094e-01 1.14674866e-02 -8.48625824e-02
9.41937208e-01 -2.59320617e-01 1.06097901e+00 5.59035420e-01
2.07277700e-01 -1.35270572e+00 -1.79280192e-01 1.22264445e+00
5.93203604e-01 -1.12635696e+00 -4.04282153e-01 -3.38676602e-01
-5.70602417e-01 1.13089740e+00 6.38979971e-01 -7.55078197e-02
1.16841786e-01 3.12138915e-01 1.43681303e-01 4.20175731e-01
-9.20762956e-01 -7.26639181e-02 2.82268494e-01 6.74047053e-01
7.48246193e-01 3.45966607e-01 -5.83418310e-01 9.60942447e-01
-2.26934299e-01 -4.04972360e-02 6.66389227e-01 9.05054569e-01
-1.00190294e+00 -1.25556779e+00 -4.66789454e-01 1.81437150e-01
-6.12684250e-01 -8.65871370e-01 -6.05940759e-01 -8.12429935e-03
3.06375653e-01 1.14492285e+00 -5.49136661e-02 -2.85384417e-01
3.34170014e-01 3.90860915e-01 2.25690439e-01 -1.12893045e+00
-9.24692690e-01 -2.72367001e-01 -3.87347266e-02 -2.97351539e-01
3.50313336e-02 -8.28192472e-01 -1.47847521e+00 -1.78969041e-01
-2.96975881e-01 1.21756621e-01 3.70088488e-01 1.22138989e+00
4.17627484e-01 2.71530569e-01 1.93427578e-01 -5.15595555e-01
-6.86573923e-01 -8.64355683e-01 -3.86782616e-01 1.93664044e-01
6.12451494e-01 1.19059332e-01 -2.84345299e-01 5.82631603e-02] | [10.973601341247559, 10.712369918823242] |
1ffc39df-3fa8-4f19-87e8-d538f2d9a39f | rates-of-convergence-of-spectral-methods-for | 1709.03183 | null | http://arxiv.org/abs/1709.03183v1 | http://arxiv.org/pdf/1709.03183v1.pdf | Rates of Convergence of Spectral Methods for Graphon Estimation | This paper studies the problem of estimating the grahpon model - the
underlying generating mechanism of a network. Graphon estimation arises in many
applications such as predicting missing links in networks and learning user
preferences in recommender systems. The graphon model deals with a random graph
of $n$ vertices such that each pair of two vertices $i$ and $j$ are connected
independently with probability $\rho \times f(x_i,x_j)$, where $x_i$ is the
unknown $d$-dimensional label of vertex $i$, $f$ is an unknown symmetric
function, and $\rho$ is a scaling parameter characterizing the graph sparsity.
Recent studies have identified the minimax error rate of estimating the graphon
from a single realization of the random graph. However, there exists a wide gap
between the known error rates of computationally efficient estimation
procedures and the minimax optimal error rate.
Here we analyze a spectral method, namely universal singular value
thresholding (USVT) algorithm, in the relatively sparse regime with the average
vertex degree $n\rho=\Omega(\log n)$. When $f$ belongs to H\"{o}lder or Sobolev
space with smoothness index $\alpha$, we show the error rate of USVT is at most
$(n\rho)^{ -2 \alpha / (2\alpha+d)}$, approaching the minimax optimal error
rate $\log (n\rho)/(n\rho)$ for $d=1$ as $\alpha$ increases. Furthermore, when
$f$ is analytic, we show the error rate of USVT is at most $\log^d
(n\rho)/(n\rho)$. In the special case of stochastic block model with $k$
blocks, the error rate of USVT is at most $k/(n\rho)$, which is larger than the
minimax optimal error rate by at most a multiplicative factor $k/\log k$. This
coincides with the computational gap observed for community detection. A key
step of our analysis is to derive the eigenvalue decaying rate of the edge
probability matrix using piecewise polynomial approximations of the graphon
function $f$. | ['Jiaming Xu'] | 2017-09-10 | rates-of-convergence-of-spectral-methods-for-1 | https://icml.cc/Conferences/2018/Schedule?showEvent=2335 | http://proceedings.mlr.press/v80/xu18a/xu18a.pdf | icml-2018-7 | ['graphon-estimation'] | ['graphs'] | [ 2.17787310e-01 5.40806174e-01 -1.01466380e-01 1.20795280e-01
-6.73404276e-01 -4.58704323e-01 -3.50548595e-01 2.15359464e-01
-2.56779939e-01 6.26643717e-01 -4.50992912e-01 -4.40071672e-01
-6.75912559e-01 -1.02754283e+00 -8.96429420e-01 -8.84118974e-01
-9.93187487e-01 3.58823597e-01 -4.78996150e-02 -1.29191294e-01
1.12494633e-01 2.57276416e-01 -1.29915357e+00 -5.49681842e-01
8.52252841e-01 1.33466446e+00 -9.94521379e-02 9.44308817e-01
-8.49266499e-02 4.39424336e-01 -1.78858131e-01 -5.64006150e-01
7.00441360e-01 -7.48721182e-01 -6.67889237e-01 2.46957868e-01
1.19794093e-01 3.03395361e-01 -5.08090317e-01 1.71897745e+00
2.72050709e-01 2.09028989e-01 7.17928052e-01 -1.18519330e+00
-2.31256783e-01 6.79799378e-01 -9.39500272e-01 2.27959022e-01
2.22830787e-01 -1.61955729e-01 1.13817382e+00 -4.74166483e-01
6.23285711e-01 7.84000814e-01 8.25568557e-01 3.90885398e-02
-1.36326253e+00 -1.00264239e+00 5.56202326e-03 -2.63671935e-01
-1.93937421e+00 -4.58157621e-02 9.03317988e-01 -6.14940703e-01
4.42940474e-01 2.26699844e-01 6.50778711e-01 7.84842521e-02
-1.08223610e-01 1.43017873e-01 9.01744783e-01 -5.05574942e-01
2.21053749e-01 5.61021268e-02 1.44839287e-01 1.30723774e+00
5.43790102e-01 -2.13103797e-02 -1.13823250e-01 -3.04118246e-01
9.07895982e-01 -1.71816975e-01 -4.14258182e-01 -1.73182651e-01
-7.53952861e-01 9.74885583e-01 3.19340646e-01 2.57759303e-01
-3.17411274e-01 5.48957467e-01 7.00499341e-02 6.19216084e-01
4.11579728e-01 1.03479490e-01 -2.64098227e-01 -3.41814868e-02
-7.33369112e-01 -1.65975973e-01 1.00579071e+00 1.23399270e+00
9.34034765e-01 6.67949170e-02 4.66715902e-01 7.41374910e-01
2.57941395e-01 7.73321152e-01 -3.94400239e-01 -1.18889880e+00
6.45951688e-01 6.45087957e-01 6.76536784e-02 -1.24867916e+00
-2.87875354e-01 -5.72041214e-01 -1.31050634e+00 6.28657341e-02
8.42108965e-01 -2.45276824e-01 -4.21960056e-01 2.04225469e+00
1.88809410e-01 -1.43341854e-01 -3.72864902e-01 6.31018579e-01
2.84013271e-01 5.73598981e-01 -3.06850672e-01 -9.39385891e-01
1.15735602e+00 -2.21916780e-01 -3.48432869e-01 -5.55514311e-03
6.63403034e-01 -7.24443853e-01 8.02518189e-01 3.72869521e-01
-1.39666247e+00 -6.05175272e-02 -8.60697985e-01 5.44869363e-01
6.93880096e-02 -1.88278884e-01 4.69283164e-01 8.64287198e-01
-1.12093222e+00 6.56578898e-01 -5.11550367e-01 -1.91107050e-01
1.22114338e-01 5.65821826e-01 -3.13768297e-01 -3.13614517e-01
-9.01282251e-01 1.72237754e-01 -2.08264008e-01 6.40929043e-02
-5.56167424e-01 -5.01921475e-01 -7.20834136e-01 1.53104246e-01
5.62828720e-01 -1.79825619e-01 3.25864524e-01 -8.22180808e-01
-9.19547200e-01 6.12547636e-01 -5.24981543e-02 -3.26046824e-01
3.51338685e-01 5.63425303e-01 -3.49714577e-01 4.05917227e-01
2.53357708e-01 -1.12266149e-02 8.12589169e-01 -1.15495884e+00
-3.04653734e-01 -6.07019603e-01 2.22853795e-02 -5.87230660e-02
-1.48315541e-02 -1.34431198e-01 -4.00265872e-01 -4.98462856e-01
5.64775884e-01 -1.06479657e+00 -3.66105139e-01 -2.80361846e-02
-3.62864792e-01 6.32858500e-02 4.07061040e-01 -5.64441323e-01
1.46243596e+00 -2.18298507e+00 1.92234907e-02 1.03850675e+00
6.80554807e-01 -1.49504066e-01 1.23394154e-01 5.33974290e-01
-2.27667332e-01 2.57150471e-01 -4.70236629e-01 1.07946955e-01
-2.26101145e-01 -5.01002930e-02 1.83263808e-01 9.64165330e-01
-5.17264187e-01 6.46872446e-02 -7.89267957e-01 -1.68471724e-01
-2.87296981e-01 4.17533308e-01 -5.74534774e-01 -2.28498489e-01
2.06714366e-02 -4.11236323e-02 -3.80342156e-01 4.97790813e-01
8.86008382e-01 -8.70309174e-01 5.52188754e-01 -1.89784989e-02
-4.01078165e-02 -2.08304018e-01 -1.93243539e+00 1.14202845e+00
-1.19469255e-01 4.65274364e-01 7.89083004e-01 -1.50616097e+00
6.98441625e-01 3.31176132e-01 8.72487128e-01 -2.48709425e-01
2.66065598e-01 3.37022156e-01 7.07103014e-02 -2.54995525e-01
-1.00384265e-01 -4.13514495e-01 5.65830385e-03 6.84747875e-01
-2.22806484e-01 1.41587004e-01 3.63217652e-01 5.16678870e-01
1.81716120e+00 -7.84669042e-01 2.02635989e-01 -6.96627617e-01
2.97059596e-01 -2.06376448e-01 6.12188995e-01 7.60293961e-01
-3.35358024e-01 1.53788000e-01 1.20009625e+00 1.64728146e-02
-9.32716608e-01 -9.74784315e-01 -2.24205002e-01 7.98807323e-01
3.73876154e-01 -4.07384127e-01 -9.52830374e-01 -5.37836075e-01
4.86005582e-02 2.26703897e-01 -7.04442799e-01 7.17344284e-02
-3.90407532e-01 -1.05626094e+00 2.46889129e-01 2.48858824e-01
6.11204207e-01 -7.04481781e-01 1.06886990e-01 2.07396105e-01
-8.33015963e-02 -9.51626062e-01 -8.46844137e-01 2.16178253e-01
-8.13758850e-01 -1.42629349e+00 -5.27527213e-01 -8.39687049e-01
1.16822135e+00 3.96065712e-01 8.80966723e-01 2.34877586e-01
-4.38120931e-01 5.89000165e-01 -1.73394844e-01 2.38024160e-01
-1.72090992e-01 -1.15704939e-01 2.06327453e-01 -4.34888639e-02
-4.97479811e-02 -7.50857234e-01 -9.49070752e-01 3.97025019e-01
-6.98378563e-01 -4.55050528e-01 3.90629709e-01 6.60305738e-01
7.56134987e-01 6.13901556e-01 3.61495197e-01 -9.24580753e-01
4.30144608e-01 -5.59662700e-01 -9.90601540e-01 -5.86922280e-02
-1.01894140e+00 7.88901895e-02 8.17232549e-01 -2.71955699e-01
-1.99009061e-01 -3.35653387e-02 1.03126958e-01 -3.41000527e-01
4.45881993e-01 6.85564697e-01 8.34557265e-02 -5.02988994e-01
6.24388218e-01 3.21903937e-02 8.31846073e-02 -3.65337372e-01
2.29402065e-01 2.61936188e-01 1.25044361e-01 -4.04194534e-01
6.38854206e-01 5.79161644e-01 5.22357881e-01 -1.03842103e+00
-4.94093388e-01 -4.24465954e-01 -5.95311709e-02 -9.99929905e-02
4.38193172e-01 -7.07781971e-01 -1.41650712e+00 1.63762867e-02
-6.20729089e-01 -3.21274072e-01 -3.29197884e-01 5.50108552e-01
-4.92720395e-01 5.60066164e-01 -7.02646911e-01 -1.26652515e+00
-1.53056517e-01 -9.23031926e-01 3.94481003e-01 -2.08745465e-01
1.20581061e-01 -1.02643216e+00 -1.89784676e-01 2.90918767e-01
2.26929277e-01 2.75281698e-01 1.23430669e+00 -2.18920603e-01
-6.46936119e-01 -6.61481798e-01 -5.58432519e-01 4.44382817e-01
-1.91277459e-01 -2.49428779e-01 -1.30644456e-01 -6.52747989e-01
7.45857060e-02 3.39502364e-01 5.21513462e-01 8.15965235e-01
1.05146742e+00 -7.64436245e-01 -4.61965621e-01 4.88932252e-01
1.86785483e+00 2.05249175e-01 4.54222858e-01 -4.87886101e-01
5.17715514e-01 3.00788134e-01 1.46068662e-01 6.10933721e-01
1.59770802e-01 2.79268682e-01 4.66780305e-01 1.89259276e-01
6.95710108e-02 4.74633835e-02 2.80391634e-01 1.04945481e+00
-3.17027777e-01 -2.90676534e-01 -7.96320260e-01 5.97318232e-01
-1.53151047e+00 -8.77395451e-01 -3.16089481e-01 2.77652025e+00
5.97758234e-01 4.01176214e-01 3.58358860e-01 2.15536714e-01
1.09011590e+00 -7.60654807e-02 -3.99821430e-01 -1.62321910e-01
-1.01572881e-02 5.45452118e-01 1.02198935e+00 9.19187784e-01
-5.80137432e-01 3.50990534e-01 5.18920565e+00 1.02018118e+00
-7.32179761e-01 3.38786505e-02 5.19018352e-01 -9.48586464e-02
-2.55542606e-01 3.51821959e-01 -5.85728526e-01 8.08350623e-01
8.98352146e-01 -1.44120425e-01 8.31065297e-01 8.40314686e-01
1.61670744e-01 -3.62706423e-01 -6.69164956e-01 1.11973250e+00
-5.66075854e-02 -1.15977037e+00 -5.28267443e-01 7.66563177e-01
7.71165669e-01 -1.17112607e-01 -2.45298501e-02 -1.01164930e-01
6.62771523e-01 -1.03433692e+00 1.78317819e-02 6.49111792e-02
1.17897356e+00 -9.29089189e-01 4.34318215e-01 3.99911433e-01
-1.70118296e+00 -1.36289895e-01 -2.46701255e-01 -2.64025219e-02
1.93792552e-01 1.01740599e+00 -4.59174395e-01 3.11266452e-01
6.96799219e-01 3.50379199e-01 1.91662833e-01 7.35866904e-01
2.22901255e-01 5.02913952e-01 -7.12919235e-01 1.75102521e-03
1.03506871e-01 -8.48027349e-01 8.31851006e-01 7.83618450e-01
5.86104691e-01 8.33737731e-01 2.08287537e-01 4.72412705e-01
-5.17464936e-01 3.61810923e-01 -6.79347038e-01 -2.15944543e-01
4.93124038e-01 8.71747673e-01 -1.02852249e+00 -1.01024427e-01
-3.47404748e-01 7.44292855e-01 2.37587526e-01 4.50114340e-01
-6.41845942e-01 -8.61025274e-01 6.68912292e-01 7.95943797e-01
5.11896074e-01 -4.30846393e-01 -2.57738233e-01 -9.08444643e-01
1.70270279e-01 -4.82261002e-01 4.65564460e-01 -1.30362466e-01
-1.15090859e+00 5.05815566e-01 -1.35204211e-01 -1.08585393e+00
1.46711573e-01 -4.74423796e-01 -2.00137883e-01 7.18011856e-01
-7.63203204e-01 -3.20583344e-01 6.54884204e-02 6.38350785e-01
-3.08246732e-01 -2.16330495e-02 6.12985075e-01 4.23333615e-01
-5.61495245e-01 6.35031343e-01 6.46945775e-01 3.18710089e-01
2.95745395e-02 -9.96952236e-01 -2.83356220e-01 8.09026301e-01
-1.81575671e-01 5.04490376e-01 8.91707480e-01 -7.09593415e-01
-1.41047502e+00 -9.10277545e-01 8.30424607e-01 4.92653400e-02
6.74649000e-01 -2.96890020e-01 -7.47997642e-01 6.01298392e-01
-2.69766688e-01 4.17208731e-01 8.12691092e-01 6.29868582e-02
-2.07147598e-01 -3.54235530e-01 -1.53792548e+00 4.57434863e-01
1.26127648e+00 -4.35151547e-01 5.08742750e-01 5.36614001e-01
3.04603010e-01 -1.10218190e-01 -1.30934417e+00 3.13569695e-01
3.16686183e-01 -9.13645685e-01 7.78734565e-01 -2.55693257e-01
-6.16250969e-02 -1.99443698e-01 -4.97559220e-01 -8.12273443e-01
-2.89776683e-01 -1.03586745e+00 -1.12037852e-01 6.84708416e-01
5.24173260e-01 -6.56362653e-01 1.15719879e+00 4.79569077e-01
2.54579067e-01 -7.81885147e-01 -1.31572044e+00 -7.46711016e-01
3.67447361e-02 -4.72315162e-01 3.74554396e-02 9.68798220e-01
1.86355054e-01 1.23160146e-01 -1.36940718e-01 2.74942398e-01
1.13346386e+00 -9.95142311e-02 4.09133017e-01 -1.21343160e+00
-6.37429297e-01 -4.81635243e-01 -4.32804465e-01 -1.05142856e+00
-9.51263532e-02 -9.49125290e-01 -1.71377152e-01 -1.16909814e+00
3.13678324e-01 -9.31151271e-01 -3.21859680e-02 -8.55158344e-02
3.41791511e-01 1.06566206e-01 -9.03439000e-02 -8.71364959e-03
-4.87612963e-01 1.85032904e-01 1.19562078e+00 1.11381449e-01
-2.86918581e-01 1.72440454e-01 -7.10625052e-01 8.45473170e-01
3.56713295e-01 -5.69953859e-01 -3.39394182e-01 -1.11765988e-01
8.43650043e-01 8.59265625e-01 1.59308478e-01 -7.96049416e-01
1.61402956e-01 3.23590487e-02 -1.50037512e-01 -1.57014921e-01
3.32095593e-01 -7.77920365e-01 3.78149807e-01 5.32349110e-01
-1.40743136e-01 5.68775348e-02 -4.00857776e-01 1.11689949e+00
9.03839171e-02 -2.20408544e-01 9.84793305e-01 -1.99295536e-01
1.26759857e-01 5.48210025e-01 -1.67169467e-01 2.38429442e-01
9.69598532e-01 -1.78601518e-01 -2.40237057e-01 -8.52288842e-01
-8.73015881e-01 1.25594407e-01 3.01612914e-01 -5.14594793e-01
3.53434265e-01 -1.17802775e+00 -4.95278925e-01 1.18264981e-01
-1.91277757e-01 5.86448312e-02 4.81286228e-01 1.15103996e+00
-6.02704644e-01 -3.48851234e-02 5.58387220e-01 -5.17300248e-01
-8.47222626e-01 5.45319319e-01 3.59765232e-01 -3.43725473e-01
-2.59710073e-01 1.19879794e+00 -1.04014032e-01 1.27885491e-01
1.76935688e-01 -8.40400904e-02 3.70600611e-01 -1.48201182e-01
1.67659238e-01 7.78144598e-01 -8.70253295e-02 -6.34229422e-01
-2.25696385e-01 6.86791956e-01 -3.78051260e-03 -2.27258995e-01
1.02258074e+00 -2.74538368e-01 -4.86106038e-01 7.94513151e-02
1.54036820e+00 3.06254774e-01 -1.06626332e+00 -1.97178543e-01
-3.68680537e-01 -4.09334928e-01 -2.42921233e-01 -2.39675835e-01
-1.36872673e+00 5.18136203e-01 5.83860457e-01 9.62431729e-01
9.89862859e-01 3.35838675e-01 6.86001658e-01 -7.26034045e-02
5.87237835e-01 -1.28098452e+00 -1.07661694e-01 4.13857877e-01
5.95057964e-01 -1.21273625e+00 3.32786553e-02 -9.09150362e-01
-2.38238454e-01 7.78057992e-01 2.36149251e-01 -5.82306325e-01
1.44188392e+00 -1.34989973e-02 -5.10140181e-01 -4.95304793e-01
-3.29094201e-01 -1.45643681e-01 1.13045767e-01 2.10098758e-01
4.71641064e-01 2.37088278e-01 -5.58336377e-01 6.23751104e-01
-2.43587330e-01 -2.18310282e-01 4.67318982e-01 5.96007526e-01
-5.58809280e-01 -1.02611697e+00 -1.98634014e-01 8.59106541e-01
-4.62180406e-01 -1.65252075e-01 7.16349334e-02 7.10832596e-01
6.60279095e-02 9.97583747e-01 -2.75498000e-03 -2.39303097e-01
1.30762339e-01 -9.78722870e-02 5.22498012e-01 -4.23614889e-01
-1.68147221e-01 2.48208240e-01 -8.08564126e-02 -3.31532151e-01
-2.83366054e-01 -5.92313051e-01 -1.36768723e+00 -8.79058778e-01
-3.35697144e-01 5.71287215e-01 5.43903768e-01 7.48868823e-01
3.20123583e-01 -5.54728881e-02 8.64459634e-01 -2.72180408e-01
-1.72987089e-01 -6.36418343e-01 -1.43851280e+00 3.25755745e-01
8.10274109e-02 -5.72020054e-01 -1.01816201e+00 -1.19937874e-01] | [6.700377941131592, 4.877086162567139] |
2ba5343c-2ec2-4d5b-ace5-02e0531a36a9 | based-benchmarking-analysis-and-structural | 2305.17477 | null | https://arxiv.org/abs/2305.17477v1 | https://arxiv.org/pdf/2305.17477v1.pdf | BASED: Benchmarking, Analysis, and Structural Estimation of Deblurring | This paper discusses the challenges of evaluating deblurring-methods quality and proposes a reduced-reference metric based on machine learning. Traditional quality-assessment metrics such as PSNR and SSIM are common for this task, but not only do they correlate poorly with subjective assessments, they also require ground-truth (GT) frames, which can be difficult to obtain in the case of deblurring. To develop and evaluate our metric, we created a new motion-blur dataset using a beam splitter. The setup captured various motion types using a static camera, as most scenes in existing datasets include blur due to camera motion. We also conducted two large subjective comparisons to aid in metric development. Our resulting metric requires no GT frames, and it correlates well with subjective human perception of blur. | ['Dmitriy Vatolin', 'Mikhail Dremin', 'Egor Chistov', 'Nikita Alutis'] | 2023-05-27 | null | null | null | null | ['deblurring'] | ['computer-vision'] | [ 1.35350049e-01 -8.19274127e-01 1.86359718e-01 -3.19913507e-01
-8.87579203e-01 -7.74231911e-01 4.45808232e-01 -2.43370771e-01
-3.90794933e-01 9.14668024e-01 6.25106633e-01 -1.73989236e-01
-3.32700387e-02 -1.38231680e-01 -3.24951619e-01 -6.36624098e-01
-1.07151449e-01 -4.91982132e-01 1.77363619e-01 1.54420093e-01
4.79390264e-01 1.90170527e-01 -1.25413299e+00 1.58480883e-01
1.00915289e+00 1.00033057e+00 3.72178853e-01 1.07951868e+00
6.32226229e-01 1.07776725e+00 -8.84787500e-01 -3.23518008e-01
4.42663044e-01 -6.58993840e-01 -6.54000282e-01 1.94810465e-01
9.11265850e-01 -7.39968717e-01 -6.03545308e-01 1.03406823e+00
6.02458239e-01 7.59113654e-02 5.18907189e-01 -1.06489170e+00
-8.22870016e-01 1.00930989e-01 -4.31707144e-01 6.40044689e-01
7.14179933e-01 4.28215176e-01 9.55377936e-01 -8.47261429e-01
4.09319907e-01 9.98711348e-01 7.86735952e-01 2.79027164e-01
-1.27469051e+00 -4.70754474e-01 -7.04055727e-01 6.70044422e-01
-1.20214272e+00 -8.33277762e-01 4.22432035e-01 -5.08009434e-01
5.08906841e-01 4.51067120e-01 5.25302351e-01 1.03070414e+00
4.32584316e-01 4.57717806e-01 1.44373536e+00 -3.55996370e-01
3.56162608e-01 -2.54326671e-01 3.90197709e-02 1.66603461e-01
3.45712960e-01 3.31333518e-01 -4.48708594e-01 -1.27850957e-02
9.56292033e-01 -3.54217857e-01 -9.78302777e-01 -1.76693156e-01
-1.53483653e+00 3.64838928e-01 3.47382307e-01 2.65044600e-01
-2.65089571e-01 3.00813913e-01 1.22812860e-01 5.21119297e-01
2.50658572e-01 7.11099982e-01 -2.73778409e-01 -5.73790550e-01
-1.48348534e+00 2.79354844e-02 6.93268001e-01 8.13575149e-01
6.13596261e-01 6.54568477e-03 -3.20681214e-01 9.20234323e-01
2.85228211e-02 4.55723792e-01 4.25394982e-01 -1.67005110e+00
2.57195443e-01 -1.69300511e-01 7.75467038e-01 -1.15991771e+00
-1.31169990e-01 -3.93853277e-01 -8.86621237e-01 4.36657786e-01
6.55681670e-01 -1.54570818e-01 -6.83845162e-01 1.32060254e+00
-4.03930366e-01 3.05263400e-01 -1.30051434e-01 1.35675025e+00
4.30624336e-01 5.46161771e-01 -4.24596637e-01 -5.02404809e-01
1.06210101e+00 -1.10201859e+00 -9.97746587e-01 -2.29676440e-01
2.48758957e-01 -1.26787972e+00 9.98704076e-01 5.31788945e-01
-1.30293071e+00 -7.21279800e-01 -1.22027469e+00 1.26341239e-01
2.06125960e-01 -2.03721840e-02 2.48518556e-01 8.39781582e-01
-1.54187524e+00 6.81139886e-01 -5.62833786e-01 -2.81037331e-01
1.27785847e-01 -1.82374358e-01 -2.67811805e-01 -4.06603783e-01
-1.05936170e+00 1.33616972e+00 -1.96872547e-01 -1.92886088e-02
-9.83241796e-01 -5.45152485e-01 -7.45978773e-01 -2.31224094e-02
1.07405275e-01 -7.53064930e-01 1.50661540e+00 -8.52433324e-01
-1.46871448e+00 4.39402878e-01 -3.64365250e-01 -1.61792129e-01
6.06696069e-01 -5.40820658e-01 -5.51646411e-01 2.66795158e-01
-6.11438975e-02 4.23763931e-01 1.10336208e+00 -1.36476386e+00
-4.28689003e-01 1.69628516e-01 2.32736215e-01 1.34071290e-01
-3.92658487e-02 2.08693475e-01 -5.80689788e-01 -9.69235599e-01
5.87965511e-02 -7.69907176e-01 -1.10352170e-02 2.49927747e-03
-3.93683314e-02 5.99440932e-01 7.09675908e-01 -1.03986096e+00
1.46237791e+00 -1.90760481e+00 -5.20664603e-02 -2.47425556e-01
2.85291612e-01 1.64062455e-01 -1.29782170e-01 3.92572939e-01
-6.11125305e-02 -6.59860298e-02 -1.65197536e-01 -1.90068424e-01
-2.44598418e-01 -1.44972026e-01 -9.08096954e-02 6.38065875e-01
-1.73664555e-01 5.82868636e-01 -1.05561173e+00 -3.02849293e-01
4.39891458e-01 1.85676783e-01 -5.29101670e-01 4.01027948e-01
2.63582140e-01 5.36789775e-01 3.22854996e-01 6.26508296e-01
7.69820273e-01 -4.43431735e-01 -3.07706475e-01 -6.19210422e-01
-1.05625838e-01 8.28863867e-03 -1.05549955e+00 1.67213356e+00
-5.80141246e-01 1.43317425e+00 4.73948829e-02 -3.76202792e-01
4.22061741e-01 3.95978212e-01 2.97648489e-01 -5.64881325e-01
-8.41053054e-02 1.81551665e-01 4.40082625e-02 -6.61094785e-01
8.55860412e-01 2.03131720e-01 4.06128407e-01 4.92755860e-01
-2.04747275e-01 -4.63752747e-01 1.92774758e-01 1.14718981e-01
1.40712714e+00 -7.04459995e-02 2.61147708e-01 -3.18456411e-01
4.36007619e-01 -7.87893906e-02 4.64403838e-01 8.29433203e-01
-4.66161489e-01 1.33714592e+00 4.99200821e-02 -6.39344677e-02
-1.28292179e+00 -1.02498567e+00 -1.61781088e-01 5.00874937e-01
5.42184174e-01 -3.44041526e-01 -9.38218236e-01 -1.69013456e-01
-5.36821663e-01 5.94372272e-01 -1.37248084e-01 1.17172003e-01
-3.04569602e-01 -7.46192515e-01 3.78921241e-01 2.32803375e-01
6.24772549e-01 -3.37347597e-01 -7.27211475e-01 1.76090837e-01
-5.39133906e-01 -1.35966420e+00 -9.67433035e-01 -4.32202220e-01
-7.23884225e-01 -1.24857438e+00 -1.21536517e+00 -4.52720106e-01
4.48010415e-01 8.31455648e-01 1.21662581e+00 -2.98961848e-02
-1.49233609e-01 4.34869200e-01 -4.53756213e-01 3.46012056e-01
-5.05302906e-01 -5.61592698e-01 1.05313331e-01 -1.45180389e-01
-6.09654970e-02 -4.91929263e-01 -1.21741343e+00 7.82591641e-01
-8.71610045e-01 2.66599655e-02 5.97223639e-01 8.64220977e-01
-1.28204703e-01 2.68725187e-01 4.86999489e-02 5.20260893e-02
8.16676021e-01 -2.21276432e-01 -6.38921142e-01 -1.12719659e-03
-7.22645938e-01 -3.36734027e-01 2.99276114e-01 -4.20201421e-01
-1.08726037e+00 -5.88984847e-01 2.57226199e-01 -4.91325051e-01
-1.02245569e-01 4.63042170e-01 3.32287811e-02 -2.07490310e-01
8.51694167e-01 3.96745987e-02 -9.32801664e-02 -5.06180108e-01
2.07222462e-01 1.02815199e+00 9.39408898e-01 -1.74629897e-01
8.03334177e-01 2.25758120e-01 -1.17723897e-01 -9.00420189e-01
-5.50253808e-01 -5.67190230e-01 -3.70146036e-01 -3.05956066e-01
6.86188579e-01 -1.20070374e+00 -5.68497300e-01 1.03996992e+00
-1.23763442e+00 -4.45283890e-01 3.33975822e-01 8.98514867e-01
-6.01016819e-01 8.43743086e-01 -1.00276005e+00 -5.63244224e-01
-1.06847860e-01 -1.44646823e+00 6.98521376e-01 2.14750782e-01
-4.08430338e-01 -1.01606107e+00 8.58419016e-02 6.05167985e-01
1.00575888e+00 -5.76645769e-02 4.33455735e-01 4.04459000e-01
-7.66362309e-01 -9.39679593e-02 -5.88224947e-01 6.99989200e-01
5.83080053e-01 -1.52505979e-01 -1.09120607e+00 -5.05985796e-01
3.25790137e-01 -1.49082378e-01 4.78406519e-01 7.13825643e-01
9.30546582e-01 -2.72201449e-01 1.77714512e-01 6.33925796e-01
1.42370176e+00 1.35251656e-01 9.99435604e-01 4.56661791e-01
5.68547249e-01 1.92674115e-01 5.68277299e-01 2.72987455e-01
2.96601623e-01 1.06452179e+00 8.99051800e-02 -1.87877789e-01
-6.07704103e-01 2.49836519e-01 5.76404810e-01 8.42584372e-01
-1.88145578e-01 -3.94174218e-01 -8.00418258e-01 4.66473073e-01
-1.49832833e+00 -1.02187943e+00 -3.30849856e-01 2.22096705e+00
1.10904539e+00 -1.51013941e-01 -1.42771244e-01 2.12508410e-01
7.17088521e-01 2.84733981e-01 -1.82805762e-01 2.82720954e-04
-3.96497279e-01 9.66770761e-03 5.42454958e-01 7.26430058e-01
-9.84339118e-01 5.77534258e-01 7.59167767e+00 5.12547314e-01
-1.05778396e+00 2.07010299e-01 4.84019488e-01 -1.70025304e-01
5.42624258e-02 8.22353140e-02 -9.87027586e-02 7.38844752e-01
9.61763382e-01 -1.53429553e-01 6.98137581e-01 4.15980488e-01
9.53237116e-01 -6.16231203e-01 -1.18019915e+00 1.40423715e+00
1.07082821e-01 -1.07136762e+00 -4.06095624e-01 -4.50627357e-02
9.55324769e-01 -1.09704718e-01 3.10576223e-02 -4.42243844e-01
2.44847476e-01 -1.09141684e+00 7.74690866e-01 6.37926519e-01
9.49012518e-01 -1.78893656e-01 9.18063283e-01 -9.67047215e-02
-7.51192033e-01 9.50502977e-02 -3.09122056e-01 -1.44429520e-01
3.76750976e-01 9.67034280e-01 -3.35636318e-01 4.55390543e-01
8.14341009e-01 7.50597894e-01 -6.78825796e-01 1.81098950e+00
-2.08269805e-01 6.37664318e-01 1.48249984e-01 5.13752937e-01
-8.78381655e-02 -2.48422727e-01 6.91695333e-01 1.29530382e+00
8.18669081e-01 -2.78532971e-02 -3.54475290e-01 5.79019010e-01
1.35495722e-01 -3.78196180e-01 -1.99848652e-01 2.76594520e-01
4.40823197e-01 1.05526125e+00 -3.23228836e-01 -2.42264941e-01
-6.14008009e-01 1.40352976e+00 -3.44281584e-01 7.71931946e-01
-8.54045928e-01 -3.28335851e-01 9.91148472e-01 -1.53647438e-01
1.54759437e-01 -5.56697249e-01 -3.58129680e-01 -1.45757806e+00
6.39547706e-02 -1.13540339e+00 -3.67943645e-02 -1.45178294e+00
-1.34082949e+00 5.19450665e-01 1.34632155e-01 -1.61106539e+00
-2.90873080e-01 -3.41490209e-01 -5.56050837e-01 1.04371727e+00
-1.40581799e+00 -5.92860878e-01 -9.25272644e-01 4.64441061e-01
6.18854761e-01 1.00653261e-01 4.43334043e-01 5.08643210e-01
-4.13148612e-01 3.99257779e-01 3.04877847e-01 -9.35742334e-02
1.33462894e+00 -1.16967654e+00 3.85127515e-01 1.20471692e+00
-1.70076177e-01 5.88362277e-01 1.25204837e+00 -3.33833456e-01
-1.38089263e+00 -7.68765509e-01 3.72064352e-01 -4.37999994e-01
6.64108157e-01 2.07692564e-01 -8.57099056e-01 2.42056847e-01
6.46413386e-01 -8.23304057e-03 4.31713372e-01 -2.29205772e-01
-1.88025907e-01 -1.74665973e-01 -8.81377816e-01 5.26243091e-01
8.97544026e-01 -6.71682775e-01 -6.37535453e-01 -3.11475303e-02
4.24565643e-01 -4.89060700e-01 -9.05334473e-01 2.55659670e-01
5.58448911e-01 -1.41417432e+00 1.02941275e+00 1.05331033e-01
6.16612732e-01 -6.72026634e-01 -2.44439796e-01 -1.83575344e+00
-6.73264384e-01 -8.03579211e-01 -3.02571595e-01 1.15686035e+00
1.94967136e-01 -3.61245841e-01 4.81200218e-01 5.74087381e-01
9.44655947e-03 -4.21993881e-02 -5.02459526e-01 -1.06689298e+00
-3.22091252e-01 -3.66174012e-01 3.57698679e-01 1.04194868e+00
-4.84532788e-02 1.11515217e-01 -7.34130085e-01 1.49555042e-01
9.35997903e-01 -1.64550662e-01 7.04079211e-01 -6.55970395e-01
-4.20729905e-01 -4.10249442e-01 -5.12139559e-01 -1.39447784e+00
-3.75556588e-01 -2.20960602e-01 2.54351437e-01 -1.46878731e+00
3.25688571e-01 -2.11168990e-01 -1.56868115e-01 -2.55000759e-02
-4.19524759e-01 5.19556463e-01 3.00088953e-02 6.00370228e-01
-4.08394665e-01 3.30478370e-01 1.23487735e+00 -1.07083425e-01
5.44671528e-02 -1.12276740e-01 -4.88509089e-01 5.79903781e-01
6.80121362e-01 -5.87666109e-02 -3.77137154e-01 -7.89408028e-01
-2.02238001e-02 2.60674149e-01 5.70076406e-01 -1.37311482e+00
2.75268644e-01 -2.57934958e-01 5.15932083e-01 -3.34860712e-01
3.03061813e-01 -7.82754302e-01 3.80817086e-01 1.81865454e-01
-1.24353111e-01 2.44486183e-01 4.99459729e-02 4.31156039e-01
-4.02960539e-01 -2.08705932e-01 1.06965685e+00 5.60203902e-02
-7.87350416e-01 -1.84863836e-01 -4.52411205e-01 -2.03053597e-02
6.24515533e-01 -4.13751960e-01 -7.35757232e-01 -1.01129687e+00
-2.16234848e-01 -2.30501816e-01 1.12166512e+00 2.30226293e-01
5.84085345e-01 -1.17713010e+00 -7.32985795e-01 -3.20009381e-01
-8.64984766e-02 -5.25574565e-01 8.67753625e-02 1.11406851e+00
-9.07457650e-01 2.85908431e-01 -2.91440755e-01 -7.46349454e-01
-1.22216678e+00 4.43007737e-01 3.61223847e-01 1.77639410e-01
-2.10736185e-01 6.06641114e-01 1.06730647e-01 5.75790703e-01
1.44789934e-01 -3.55100274e-01 4.14794832e-02 -2.37040520e-01
8.28973353e-01 7.39325464e-01 1.34292245e-01 -6.07212484e-01
-1.59113064e-01 5.11298776e-01 2.99431533e-01 -3.98945123e-01
9.53421175e-01 -7.77569354e-01 -1.38723314e-01 1.89858690e-01
1.19505382e+00 2.18328461e-01 -1.26163471e+00 -1.56681269e-01
1.78929642e-02 -1.14516318e+00 4.27809119e-01 -7.76628494e-01
-9.88653123e-01 6.17247581e-01 9.65546787e-01 2.52671331e-01
1.44405055e+00 -4.37816203e-01 8.15964520e-01 1.07920147e-01
4.61048335e-01 -8.10586512e-01 3.59676570e-01 4.59256738e-01
9.41542149e-01 -1.33775187e+00 1.81951776e-01 -2.92331994e-01
-4.22315478e-01 1.05917943e+00 3.30921978e-01 3.11003119e-01
3.23483825e-01 2.38378629e-01 4.19419497e-01 2.77242035e-01
-6.15986824e-01 -2.35269573e-02 4.85709429e-01 5.37590146e-01
5.92373967e-01 -3.17828327e-01 -2.20062658e-01 -4.37802747e-02
-2.26454198e-01 3.13162416e-01 9.27389085e-01 7.81839788e-01
-3.78763258e-01 -8.99259269e-01 -8.13488066e-01 2.09603697e-01
-5.87108374e-01 -3.21637958e-01 6.71254769e-02 2.31425449e-01
-1.57371998e-01 1.61175001e+00 -1.43114537e-01 -5.78472853e-01
6.38137385e-02 -5.32626152e-01 6.29787207e-01 -2.48067945e-01
-1.09646417e-01 1.50724843e-01 2.16973096e-01 -7.37561285e-01
-6.27854228e-01 -6.15084946e-01 -3.41756761e-01 -6.01456761e-01
-4.12579745e-01 -2.67869905e-02 5.15434861e-01 7.51720428e-01
1.80246577e-01 3.11094910e-01 6.85269654e-01 -1.10622311e+00
-3.49067986e-01 -1.24438763e+00 -4.10154611e-01 5.99850953e-01
8.11207712e-01 -4.22153413e-01 -7.69048393e-01 6.56321764e-01] | [11.614188194274902, -2.720612049102783] |
cf8dfd73-ed2b-4429-babf-dc8bcca10241 | training-of-a-skull-stripping-neural-network | 1810.10853 | null | http://arxiv.org/abs/1810.10853v1 | http://arxiv.org/pdf/1810.10853v1.pdf | Training of a Skull-Stripping Neural Network with efficient data augmentation | Skull-stripping methods aim to remove the non-brain tissue from acquisition
of brain scans in magnetic resonance (MR) imaging. Although several methods
sharing this common purpose have been presented in literature, they all suffer
from the great variability of the MR images. In this work we propose a novel
approach based on Convolutional Neural Networks to automatically perform the
brain extraction obtaining cutting-edge performance in the NFBS public
database. Additionally, we focus on the efficient training of the neural
network designing an effective data augmentation pipeline. Obtained results are
evaluated through Dice metric, obtaining a value of 96.5%, and processing time,
with 4.5s per volume. | ['Gabriele Valvano', 'Dante Chiappino', 'Emiliano Ricciardi', 'Nicola Martini', 'Gianmarco Santini', 'Daniele Della Latta', 'Andrea Leo'] | 2018-10-25 | null | null | null | null | ['skull-stripping'] | ['medical'] | [ 3.34698856e-01 2.59330004e-01 3.74804497e-01 -6.08968019e-01
-5.83051801e-01 -2.71143485e-02 3.36020470e-01 7.27742091e-02
-9.62013304e-01 7.44014323e-01 -8.96346010e-03 -7.60062933e-02
-2.00035289e-01 -5.67225754e-01 -5.29381037e-01 -6.12432361e-01
-3.90099347e-01 4.05331552e-01 2.90761203e-01 2.90803872e-02
2.36779556e-01 1.00572073e+00 -1.11640656e+00 7.15097412e-02
8.77925873e-01 1.07058787e+00 3.14308673e-01 3.71049374e-01
3.21079716e-02 6.62780583e-01 -5.68725705e-01 -1.96310043e-01
3.67798328e-01 2.37369556e-02 -8.64209235e-01 -1.10333852e-01
3.01420450e-01 -4.66407955e-01 -3.28084350e-01 1.07620835e+00
7.85555363e-01 1.04825236e-02 7.08311141e-01 -6.91827059e-01
-3.79226625e-01 6.88388884e-01 -6.00926459e-01 7.77630389e-01
-2.45694816e-01 6.53834790e-02 2.09415987e-01 -1.05245793e+00
7.05966353e-01 5.21418393e-01 5.55067956e-01 4.20311898e-01
-9.99959826e-01 -7.71294594e-01 -5.92426121e-01 4.31210577e-01
-1.49098074e+00 -5.29941618e-01 8.48543644e-01 -6.99560761e-01
7.94943452e-01 7.54710585e-02 7.35213757e-01 6.12541616e-01
5.58409572e-01 5.47707915e-01 1.38592184e+00 -2.28364348e-01
3.88920940e-02 -2.45474681e-01 3.61166984e-01 5.36522746e-01
2.92568117e-01 -2.31691808e-01 -2.21238226e-01 1.68458372e-01
7.12349117e-01 -9.15199295e-02 -4.20299113e-01 -1.69590026e-01
-1.07472789e+00 4.98853058e-01 5.75353086e-01 6.24733746e-01
-8.06688905e-01 -1.10366605e-01 7.56340086e-01 -1.30424425e-01
4.52415198e-01 2.55665898e-01 -1.45396233e-01 3.71278137e-01
-1.22691953e+00 -5.67842163e-02 2.18328610e-01 6.14176154e-01
2.43407950e-01 1.46390334e-01 -1.18862078e-01 6.81082904e-01
-9.35741737e-02 1.32370129e-01 8.55728507e-01 -3.55619282e-01
2.41045445e-01 4.50010419e-01 -3.47354710e-01 -9.67151165e-01
-1.13196111e+00 -6.44858837e-01 -1.16913998e+00 3.29689711e-01
2.79257059e-01 -3.07654083e-01 -1.17532492e+00 1.26815021e+00
3.81153613e-01 -6.22235052e-02 -2.95703083e-01 1.00140631e+00
9.28324461e-01 1.68959856e-01 1.59231022e-01 -2.69221812e-01
1.33097041e+00 -9.23732579e-01 -9.25341010e-01 -8.25547986e-03
4.45882589e-01 -6.38900518e-01 6.95696354e-01 6.32453561e-01
-1.22569215e+00 -2.53013611e-01 -1.22035718e+00 -1.70434471e-02
-3.25179994e-01 1.86861455e-01 5.49688339e-01 5.96849442e-01
-1.10304821e+00 6.81819320e-01 -9.62251842e-01 1.59353629e-01
9.55863059e-01 6.92173421e-01 -7.93007612e-01 1.87345043e-01
-8.74338746e-01 1.18928564e+00 6.52182341e-01 3.06293100e-01
-7.59233415e-01 -9.33428168e-01 -3.43521863e-01 1.62791845e-03
2.81917095e-01 -5.11946380e-01 9.34783697e-01 -8.78578484e-01
-1.30176795e+00 9.40832734e-01 3.30151141e-01 -8.06814909e-01
6.76489294e-01 -2.75217175e-01 -6.17691755e-01 5.07595062e-01
-1.90398633e-01 5.00056803e-01 8.11060190e-01 -8.90100718e-01
-1.47669718e-01 -8.06766272e-01 -3.85098785e-01 -7.75067881e-02
-2.01438785e-01 3.78807008e-01 -1.94154382e-01 -8.37287486e-01
3.69811624e-01 -6.22664928e-01 -1.80243865e-01 -2.39932507e-01
-3.54660332e-01 1.50476381e-01 6.32903874e-01 -1.27171814e+00
1.04873407e+00 -1.77042079e+00 -7.81842172e-02 2.38891035e-01
6.33856058e-01 4.45373327e-01 2.19166681e-01 -2.75987655e-01
-4.32441264e-01 -9.88814682e-02 -5.59606731e-01 -2.90078342e-01
-4.53486115e-01 -2.97681928e-01 1.83812246e-01 9.40339744e-01
1.32368729e-01 7.60164142e-01 -3.85400176e-01 -6.53065562e-01
2.05205992e-01 7.42443800e-01 -4.56795871e-01 1.14998266e-01
5.11553884e-01 6.79289639e-01 -2.76742697e-01 4.72081482e-01
1.06536794e+00 2.01199993e-01 4.69064340e-02 -4.43226010e-01
-8.17509294e-02 -3.67932655e-02 -8.93348038e-01 1.83000028e+00
-2.32065007e-01 5.20594001e-01 4.44094151e-01 -1.05414975e+00
7.84839690e-01 3.85544002e-01 8.23146641e-01 -7.66882777e-01
6.10311508e-01 3.02530587e-01 4.37699437e-01 -6.04751706e-01
3.11311781e-01 -2.32406363e-01 4.97605741e-01 4.42558140e-01
2.89564967e-01 -4.78885695e-02 1.17714792e-01 -1.59649611e-01
9.24899995e-01 -1.79948565e-02 2.95816064e-01 -5.81815898e-01
6.46545172e-01 -1.62505731e-01 3.85973036e-01 3.87850225e-01
-5.07933259e-01 7.69161224e-01 1.10849962e-01 -6.10363066e-01
-1.20804882e+00 -7.15385616e-01 -5.15181780e-01 4.31267917e-01
-3.86709452e-01 7.74507895e-02 -1.30905068e+00 -5.13222039e-01
-5.68248987e-01 4.01413798e-01 -6.99554682e-01 -3.86239477e-02
-1.02975285e+00 -1.01021242e+00 5.15801787e-01 4.15188432e-01
5.84511042e-01 -1.10087597e+00 -1.02306235e+00 3.14775705e-01
-4.82187495e-02 -1.11861014e+00 -3.04514021e-01 1.92936659e-01
-1.10122359e+00 -1.01871157e+00 -8.27989399e-01 -6.16210043e-01
7.74297655e-01 2.76786573e-02 8.80965710e-01 2.51851082e-01
-8.02671134e-01 -2.39510208e-01 -1.05010957e-01 -4.15637136e-01
-6.10106923e-02 3.36218953e-01 4.37616035e-02 -1.29934058e-01
3.77359152e-01 -7.78481305e-01 -8.83379161e-01 -2.86075054e-03
-1.24112105e+00 1.52471989e-01 8.75112176e-01 7.32033849e-01
6.04043424e-01 -7.90817216e-02 6.27491593e-01 -7.55286038e-01
6.02798581e-01 -4.67347324e-01 -5.60400963e-01 -2.50611547e-02
-5.29384732e-01 -1.26004547e-01 6.48259759e-01 -1.85787722e-01
-7.80138016e-01 2.57514805e-01 -5.00864327e-01 -1.82832852e-01
-4.06042844e-01 2.80159086e-01 7.46577233e-02 -2.97463566e-01
6.08921945e-01 2.05685034e-01 4.42988537e-02 -4.75122392e-01
-4.13690358e-02 5.29500723e-01 6.92571461e-01 -2.08741650e-01
4.46668595e-01 6.46644235e-01 4.56934571e-02 -7.85114825e-01
-5.44397354e-01 -2.91020960e-01 -1.11557674e+00 -3.66236061e-01
1.17730653e+00 -4.31183547e-01 -4.75147903e-01 4.49636698e-01
-1.09961021e+00 4.86327894e-02 5.66270575e-02 9.00730968e-01
-4.15812135e-01 3.68214279e-01 -4.77909237e-01 -5.11588275e-01
-1.11561823e+00 -1.58951497e+00 4.39282954e-01 9.99768823e-02
-2.90031210e-02 -5.21651626e-01 -1.67750254e-01 4.23883408e-01
6.57906711e-01 3.28475326e-01 8.91161144e-01 -8.77308309e-01
-2.52028108e-01 -1.78778946e-01 -3.09356391e-01 4.94805008e-01
1.17146401e-02 -2.50490934e-01 -1.02967882e+00 -1.04028434e-01
2.71409988e-01 3.56036164e-02 7.22667933e-01 7.92305648e-01
1.52365685e+00 5.86920716e-02 -2.49782875e-02 7.78432906e-01
1.41420984e+00 3.54060233e-01 6.63241267e-01 4.65193212e-01
4.77358103e-01 6.81065857e-01 1.72136113e-01 3.58585835e-01
-2.73448583e-02 4.52921480e-01 5.92912614e-01 -3.08658808e-01
-4.53299135e-01 4.67273086e-01 -1.64721474e-01 1.06746542e+00
-2.73835599e-01 3.77390236e-01 -1.30563128e+00 5.89109123e-01
-1.24272883e+00 -6.51472509e-01 -3.50372165e-01 2.14314675e+00
7.21925080e-01 1.87299848e-01 8.37052613e-02 9.11826417e-02
7.35519648e-01 -1.05146348e-01 -4.60010648e-01 -1.04171827e-01
-5.54038770e-02 6.43661380e-01 6.70424044e-01 4.65214729e-01
-1.27082789e+00 6.30140483e-01 6.65708637e+00 6.03473067e-01
-1.35529637e+00 4.26067382e-01 7.18506098e-01 -3.04657847e-01
2.31818572e-01 -5.48530102e-01 -4.77138400e-01 4.31418240e-01
9.05056357e-01 -5.62446797e-03 1.81574881e-01 6.92083120e-01
2.41937444e-01 -1.02246851e-01 -5.56996167e-01 1.08359206e+00
2.38331288e-01 -1.23391759e+00 -1.36957929e-01 -1.50390998e-01
4.29368168e-01 2.43303105e-01 -9.72540602e-02 -7.60262832e-02
-4.12718028e-01 -1.24998307e+00 6.97734356e-01 7.62524486e-01
6.49052441e-01 -1.13249969e+00 1.01437724e+00 1.13865569e-01
-7.59915471e-01 8.10519978e-02 -2.65704453e-01 3.09831500e-01
1.53196067e-01 8.80022407e-01 -9.35899794e-01 5.63216448e-01
8.18422973e-01 -5.72260693e-02 -5.62373817e-01 1.40181601e+00
1.11936647e-02 4.10111547e-01 -2.51476556e-01 2.74520367e-01
1.17647171e-01 -3.41599435e-01 4.57201809e-01 1.32059216e+00
2.72872180e-01 2.65206516e-01 -1.86077565e-01 9.10084665e-01
-2.25167811e-01 6.82916403e-01 -3.87787312e-01 2.49955893e-01
9.21134800e-02 1.57071090e+00 -1.12707222e+00 -3.16601276e-01
-2.00671196e-01 7.12727368e-01 2.91252553e-01 6.20062929e-03
-8.52219403e-01 -4.72176522e-01 2.26798207e-01 3.95758480e-01
-2.02185777e-03 -3.67999017e-01 -5.09312809e-01 -8.26984286e-01
2.00225636e-01 -6.26085222e-01 5.82772568e-02 -6.17004752e-01
-9.48265910e-01 1.14184427e+00 -5.55064157e-03 -7.55816221e-01
9.31201652e-02 -6.11098409e-01 -4.47704613e-01 8.95958245e-01
-1.49983382e+00 -8.25831711e-01 -3.69816422e-01 5.44196546e-01
3.70811105e-01 -1.88584715e-01 5.40083945e-01 7.98901141e-01
-7.06942439e-01 3.46883386e-01 -9.91081148e-02 2.16420874e-01
6.19822025e-01 -9.62375641e-01 2.23061189e-01 9.65693176e-01
-4.14274454e-01 7.69572258e-01 6.26438677e-01 -6.21423244e-01
-1.21916413e+00 -7.76837945e-01 5.75923324e-01 2.05269486e-01
5.95480561e-01 -1.49921492e-01 -1.12482405e+00 5.96796036e-01
3.58498275e-01 1.84354857e-01 5.99066019e-01 -3.70628029e-01
-1.89147647e-02 7.67486691e-02 -1.53101218e+00 3.72139364e-01
6.71788990e-01 -1.56748921e-01 -8.94164026e-01 2.39470512e-01
3.31969291e-01 -4.08148110e-01 -1.10165155e+00 5.84750056e-01
4.43442881e-01 -9.71555531e-01 1.03299272e+00 -5.15843391e-01
5.75969636e-01 -1.54920042e-01 3.14838886e-01 -1.17874539e+00
-2.18487740e-01 -2.29869068e-01 7.30152577e-02 7.82061875e-01
1.66180730e-01 -4.77339387e-01 7.75296986e-01 6.11035526e-01
-4.52127486e-01 -7.91149139e-01 -1.13514519e+00 -5.09398043e-01
3.24144691e-01 -4.23355341e-01 6.03996396e-01 9.48128641e-01
-4.10345644e-01 2.28607506e-02 -2.30300814e-01 6.38927817e-02
7.31661856e-01 -3.84309411e-01 2.40413263e-01 -1.37490022e+00
1.92275807e-01 -6.29875004e-01 -5.06995201e-01 -2.83314019e-01
1.96897954e-01 -1.12223673e+00 -1.10256784e-01 -1.45555329e+00
3.54370296e-01 -2.48533040e-01 -4.49948549e-01 4.21994269e-01
1.11710154e-01 4.96835649e-01 -1.74245164e-01 9.49994028e-02
-4.25718799e-02 3.73649240e-01 1.29027379e+00 -6.34419024e-02
1.46291135e-02 -4.26183581e-01 -2.70100385e-01 9.61716592e-01
1.13119733e+00 -6.16131008e-01 -2.36953512e-01 -6.19822204e-01
-2.18840182e-01 -2.39203215e-01 3.21546137e-01 -1.36853385e+00
2.56938249e-01 3.14111561e-01 6.05057418e-01 -8.42964053e-01
1.44312292e-01 -9.66793418e-01 1.50934324e-01 5.87149799e-01
-1.36089683e-01 2.24316359e-01 2.93085665e-01 -1.04110591e-01
-1.71775848e-01 -6.07653499e-01 1.27568197e+00 -2.16615722e-01
-3.34061265e-01 4.91437018e-01 -4.07596648e-01 -2.17047989e-01
1.05513597e+00 -2.49713033e-01 9.72464755e-02 2.05909565e-01
-1.01035607e+00 -2.72496939e-01 3.24756727e-02 7.56846517e-02
7.28560984e-01 -1.15955436e+00 -7.72190630e-01 1.14418328e-01
-2.58255452e-01 -1.51190534e-01 7.20985055e-01 1.41628385e+00
-9.59257185e-01 4.47918087e-01 -9.32452559e-01 -3.58917981e-01
-1.20801759e+00 5.09930909e-01 6.72562003e-01 -1.97298095e-01
-1.07387161e+00 6.97219729e-01 -1.77825056e-02 -3.11610818e-01
6.08377941e-02 -3.15489858e-01 -5.30410588e-01 9.74963233e-02
8.64874840e-01 4.20597106e-01 7.51243889e-01 -8.41009736e-01
-4.84281272e-01 2.75905997e-01 -2.29506627e-01 -9.87097342e-03
1.84325194e+00 1.52959257e-01 -3.43796581e-01 -1.45771131e-01
1.28578663e+00 -1.25028268e-01 -7.68104970e-01 -1.08890399e-01
1.76300853e-01 -3.02092403e-01 6.13964319e-01 -7.06885874e-01
-1.47013938e+00 9.53004658e-01 1.13970900e+00 -9.43338796e-02
1.16799080e+00 -5.03766537e-01 9.17259514e-01 9.04996395e-02
4.03766423e-01 -1.17960262e+00 -4.60634589e-01 3.43631566e-01
1.04710662e+00 -1.15669203e+00 8.60027075e-02 -2.96615839e-01
-3.27312261e-01 1.33384824e+00 4.00346339e-01 -3.00146997e-01
7.18510807e-01 6.14207685e-01 -3.95090617e-02 -5.29924989e-01
-7.82354102e-02 -3.73738296e-02 2.69464910e-01 5.92459679e-01
6.29710615e-01 -2.86553428e-02 -7.38571882e-01 7.18970537e-01
-2.29098499e-01 2.73874968e-01 5.45240700e-01 8.15438688e-01
-4.70460117e-01 -7.08251178e-01 -4.34349626e-01 6.07254386e-01
-9.43361223e-01 -3.04117024e-01 -1.06537722e-01 7.61482775e-01
9.13223028e-02 5.06731153e-01 -6.60546124e-02 -7.27564022e-02
3.22140515e-01 1.43382996e-01 5.39339364e-01 -4.02381331e-01
-9.22618985e-01 6.46327063e-02 -1.06018350e-01 -4.60973084e-01
-3.94382447e-01 -3.84662598e-01 -1.49052691e+00 -1.97374299e-01
-2.55469650e-01 -8.52101892e-02 1.00578141e+00 8.94654810e-01
2.58643687e-01 8.55550647e-01 2.76195318e-01 -8.76298368e-01
-2.67704576e-01 -1.21939933e+00 -7.04234421e-01 4.42125261e-01
3.35567817e-02 -8.06340218e-01 -4.33041938e-02 1.97243579e-02] | [14.184041023254395, -2.261899709701538] |
126f392b-9e1e-48d5-a9e1-a8d6b5f1d0ab | blind-image-deblurring-by-spectral-properties | 1209.2082 | null | http://arxiv.org/abs/1209.2082v3 | http://arxiv.org/pdf/1209.2082v3.pdf | Blind Image Deblurring by Spectral Properties of Convolution Operators | In this paper, we study the problem of recovering a sharp version of a given
blurry image when the blur kernel is unknown. Previous methods often introduce
an image-independent regularizer (such as Gaussian or sparse priors) on the
desired blur kernel. We shall show that the blurry image itself encodes rich
information about the blur kernel. Such information can be found through
analyzing and comparing how the spectrum of an image as a convolution operator
changes before and after blurring. Our analysis leads to an effective convex
regularizer on the blur kernel which depends only on the given blurry image. We
show that the minimizer of this regularizer guarantees to give good
approximation to the blur kernel if the original image is sharp enough. By
combining this powerful regularizer with conventional image deblurring
techniques, we show how we could significantly improve the deblurring results
through simulations and experiments on real images. In addition, our analysis
and experiments help explaining a widely accepted doctrine; that is, the edges
are good features for deblurring. | ['Shiyu Chang', 'Yi Ma', 'Guangcan Liu'] | 2012-09-10 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 1.49140030e-01 -2.79028207e-01 1.76966101e-01 -1.49291977e-01
-2.53294080e-01 -6.58112168e-01 8.81185234e-02 -7.25776315e-01
-5.57311438e-02 8.27120304e-01 6.45906746e-01 -1.48893250e-02
-1.54589713e-01 -2.74272829e-01 -7.96532452e-01 -9.77725506e-01
-9.80248675e-02 -2.98700362e-01 -2.10720628e-01 -1.18682850e-02
3.62292677e-01 2.66199827e-01 -9.82670844e-01 -5.05423099e-02
1.29078901e+00 6.28729880e-01 7.30862200e-01 8.97928178e-01
3.07654351e-01 9.69534576e-01 -4.88324910e-01 6.49751201e-02
4.90442961e-01 -6.73173964e-01 -1.02089238e+00 5.36989450e-01
2.92722374e-01 -4.72634375e-01 -9.04624104e-01 1.72538424e+00
1.47767797e-01 2.04261348e-01 4.41987157e-01 -7.35420048e-01
-1.37860811e+00 5.35322547e-01 -8.99039984e-01 4.76022720e-01
1.63968772e-01 5.21547273e-02 4.75438237e-01 -9.95647728e-01
4.42043751e-01 9.98481810e-01 8.17532778e-01 4.47862059e-01
-1.18698514e+00 -8.46690834e-02 -3.36692750e-01 3.23612928e-01
-1.52696180e+00 -7.67759860e-01 8.99886131e-01 -3.67081344e-01
1.07840098e-01 5.37987351e-01 2.56969184e-01 4.79386955e-01
3.51660341e-01 6.50875092e-01 1.42191124e+00 -5.35271764e-01
1.45454407e-01 1.45030066e-01 3.94661933e-01 6.39079869e-01
5.23273170e-01 6.79827631e-02 -3.79697382e-02 -1.42234340e-01
1.16264725e+00 2.36727118e-01 -1.46521759e+00 -2.72858709e-01
-1.20235014e+00 4.99165356e-01 7.26110220e-01 5.23129284e-01
-4.32481438e-01 2.16254264e-01 -2.29118645e-01 2.84869641e-01
6.20321929e-01 6.36753619e-01 -7.06622526e-02 -2.46446719e-03
-1.08223903e+00 -2.01800108e-01 8.55668783e-01 7.59135246e-01
8.47795248e-01 -1.12656914e-01 -1.68869004e-01 8.95286143e-01
1.62286535e-01 7.00793862e-01 3.81217986e-01 -1.26328874e+00
8.80643278e-02 -2.68459380e-01 7.06521094e-01 -8.33547592e-01
1.88161731e-01 -4.08834100e-01 -1.17855144e+00 1.41523600e-01
5.77493966e-01 -4.45222676e-01 -9.36255932e-01 1.69240153e+00
-1.93506986e-01 6.80130959e-01 9.61805061e-02 1.42582917e+00
2.47621924e-01 7.15177119e-01 -6.61196649e-01 -4.17428166e-01
1.16145897e+00 -9.63520229e-01 -1.18105733e+00 -3.28023732e-01
2.32306886e-02 -1.06595027e+00 8.26841712e-01 2.37853274e-01
-1.18568385e+00 -5.01354754e-01 -9.30496573e-01 -6.99496595e-03
3.13266814e-01 6.87601194e-02 3.52076709e-01 6.14934742e-01
-1.34447253e+00 5.65696180e-01 -6.40524805e-01 -2.34220698e-01
2.64076769e-01 -1.33225232e-01 -2.38160208e-01 -4.72942442e-01
-1.08058476e+00 1.25236988e+00 7.44041428e-03 4.50381994e-01
-7.47510850e-01 -7.34499276e-01 -7.77956009e-01 1.62549719e-01
8.68066922e-02 -8.38557780e-01 1.25560379e+00 -1.19713509e+00
-1.37290394e+00 6.12219691e-01 -5.11331320e-01 -3.66285920e-01
4.63974357e-01 -6.99250281e-01 -1.55852944e-01 4.07593161e-01
-2.11058762e-02 -1.13192141e-01 1.54745615e+00 -1.42192912e+00
-6.19016998e-02 -1.34556070e-01 -3.54372524e-02 2.27123648e-01
8.53602588e-02 -1.43919662e-01 -2.51947492e-01 -7.65373707e-01
1.30772928e-03 -7.73897886e-01 -3.05014491e-01 -4.50021178e-02
-4.27088618e-01 6.44367576e-01 8.46434653e-01 -1.01354349e+00
1.17069340e+00 -2.28212047e+00 3.56471419e-01 -6.26230016e-02
4.50311989e-01 7.27954060e-02 1.70418993e-02 2.84843475e-01
-4.55391318e-01 -1.30083457e-01 -6.67776167e-01 7.25952387e-02
-3.84060144e-01 4.49255258e-02 -5.68523109e-01 1.22170055e+00
-1.07147962e-01 8.92326295e-01 -9.59430695e-01 8.09811130e-02
2.94046789e-01 5.68039477e-01 -3.26834112e-01 4.83970284e-01
3.70860875e-01 4.82228488e-01 -1.44470125e-01 1.04687072e-01
1.33692312e+00 -5.79771399e-01 -2.08498910e-01 -4.64098841e-01
-2.11064100e-01 -4.66997236e-01 -1.02154684e+00 1.48044562e+00
-5.60054481e-01 1.09201527e+00 7.11649537e-01 -9.42331314e-01
4.68000025e-01 2.78340816e-01 2.50843018e-01 1.39954180e-01
-1.60887372e-02 1.83890581e-01 -1.68852076e-01 -7.83994257e-01
4.20452148e-01 -4.32157367e-01 4.23322111e-01 4.31860566e-01
-3.09535623e-01 -3.28367949e-01 -2.95757979e-01 2.18176723e-01
1.01393557e+00 -3.26209486e-01 2.66593456e-01 -7.94640362e-01
5.35073638e-01 -3.30835193e-01 1.51013777e-01 9.17495012e-01
-2.54117370e-01 1.01072216e+00 -5.62034510e-02 -1.30995542e-01
-1.12046206e+00 -8.64146650e-01 -3.52401376e-01 2.55607545e-01
7.78672993e-01 1.01088263e-01 -1.20421124e+00 -1.65187687e-01
-2.41285369e-01 5.10190904e-01 -6.03771567e-01 -1.32065937e-01
-4.06842530e-01 -9.95756149e-01 5.29636815e-02 1.26183599e-01
9.77588296e-01 -3.43487620e-01 -4.46759760e-01 -1.44202104e-02
-4.65273529e-01 -1.04989541e+00 -1.25869751e+00 -1.25125751e-01
-8.13401937e-01 -1.27374029e+00 -1.60832393e+00 -7.65767097e-01
1.00757444e+00 1.13481641e+00 7.15757668e-01 2.21023373e-02
-1.11808360e-01 6.15818679e-01 1.50868995e-03 3.01832199e-01
-4.18984592e-01 -7.54290402e-01 -1.58718880e-02 4.56487834e-01
-2.68838853e-01 -5.07691681e-01 -8.39614749e-01 3.28088373e-01
-1.00015354e+00 1.29740685e-01 5.09953916e-01 8.53778839e-01
-7.99023546e-03 4.78136539e-01 8.50571245e-02 -5.89393854e-01
9.34573591e-01 -3.67686838e-01 -5.64098597e-01 3.16093385e-01
-4.53601301e-01 3.60897362e-01 4.87765431e-01 -5.28192759e-01
-1.47238410e+00 -1.24252349e-01 4.17911083e-01 -6.73698545e-01
3.00249625e-02 4.76588637e-01 1.11240551e-01 -3.96142006e-01
8.11323643e-01 4.47947621e-01 1.47461742e-01 -7.70236075e-01
7.93552518e-01 8.37100863e-01 9.50291812e-01 -3.54334414e-01
1.05326784e+00 6.89831734e-01 -1.52394056e-01 -1.25731599e+00
-8.78053129e-01 -7.35804081e-01 -3.04207563e-01 -4.62240949e-02
6.65266931e-01 -9.22447741e-01 -5.19536853e-01 9.02330041e-01
-1.37048018e+00 -3.85092407e-01 -2.67688781e-01 7.00127125e-01
-6.07003987e-01 9.93645072e-01 -9.40541744e-01 -8.83225679e-01
-7.49945864e-02 -1.02138841e+00 5.78243256e-01 4.36508477e-01
1.76167265e-01 -1.29598749e+00 1.93288341e-01 1.48543388e-01
7.87901163e-01 -1.40845492e-01 4.07739371e-01 2.81121880e-01
-6.84137821e-01 1.42463148e-01 -7.26357281e-01 6.21367335e-01
7.31212199e-01 -3.92402202e-01 -1.04863179e+00 -4.16863561e-01
1.04018831e+00 3.29655141e-01 1.17437863e+00 9.79756713e-01
1.04869664e+00 -7.23342776e-01 -2.16840640e-01 1.02375603e+00
1.59290349e+00 -1.51434183e-01 8.08804691e-01 1.20085673e-02
8.66602063e-01 2.14834735e-01 8.16303715e-02 2.02545330e-01
-2.17160955e-01 6.12055063e-01 2.31045276e-01 -2.37594023e-01
-4.88319546e-01 1.83644056e-01 4.41894501e-01 6.79611862e-01
-3.23252797e-01 -1.91104077e-02 -4.82941121e-01 6.86501980e-01
-1.87463629e+00 -1.23294663e+00 -2.79952347e-01 2.23869038e+00
1.17757070e+00 -5.11949778e-01 -4.29365277e-01 -2.28442833e-01
1.26195180e+00 2.75739998e-01 -4.52583790e-01 6.04689941e-02
-2.85006970e-01 1.78481620e-02 8.27172816e-01 1.22946990e+00
-1.06033146e+00 5.73254406e-01 7.09062195e+00 5.70494652e-01
-1.05273509e+00 4.93305363e-02 4.15841341e-01 4.25592512e-01
-2.72396028e-01 5.82542419e-02 -7.88478926e-02 6.54632151e-01
6.00384593e-01 -6.80001318e-01 1.11675298e+00 6.62692547e-01
5.11240304e-01 -3.17560375e-01 -8.74101162e-01 1.37022257e+00
3.43984142e-02 -1.29561293e+00 -4.20346558e-01 1.09461129e-01
1.02989829e+00 -1.55598000e-01 1.09806187e-01 -4.78224903e-01
4.19197261e-01 -1.05217099e+00 5.69679141e-01 8.53061974e-01
7.12521076e-01 -3.29232097e-01 7.41120279e-01 2.48924688e-01
-6.26896441e-01 1.47856310e-01 -5.48195779e-01 -2.11123347e-01
2.72372931e-01 1.22792149e+00 -4.96126592e-01 5.44014037e-01
4.33997333e-01 9.98518705e-01 -2.30781853e-01 1.52143347e+00
-2.56965160e-01 5.82035065e-01 -2.65381411e-02 6.02047145e-01
1.31169800e-02 -5.41804790e-01 9.53478098e-01 1.39772534e+00
5.88401377e-01 4.58586812e-01 -3.02456707e-01 1.12769938e+00
-1.20316654e-01 -5.19104242e-01 -5.82152665e-01 3.27956438e-01
8.53260085e-02 9.58322942e-01 -4.85037267e-01 -3.31138879e-01
-2.61897653e-01 1.57535136e+00 -9.53523815e-02 1.03973603e+00
-6.84091449e-01 -5.13028324e-01 8.52729321e-01 -1.22736424e-01
3.60557020e-01 -1.67533860e-01 -4.49442893e-01 -1.78346729e+00
-1.65296778e-01 -6.17030799e-01 -1.92480907e-01 -1.19496298e+00
-1.43328249e+00 5.26920378e-01 -3.58002968e-02 -1.10391223e+00
2.34287456e-01 -4.16179150e-01 -7.46726871e-01 1.36555588e+00
-1.59087026e+00 -4.85226512e-01 -5.10748446e-01 7.01106250e-01
5.36715746e-01 3.35521579e-01 2.51293153e-01 -4.64410558e-02
-3.06640893e-01 -9.89244729e-02 5.22124350e-01 4.67277784e-03
7.85044372e-01 -1.42501223e+00 2.09439233e-01 1.36605489e+00
-1.44006804e-01 8.91334474e-01 1.29789138e+00 -5.34667253e-01
-1.49685514e+00 -9.04592931e-01 4.07272547e-01 -3.64076853e-01
7.48413682e-01 2.23445266e-01 -1.19244421e+00 7.43750036e-01
7.30946362e-01 2.51189679e-01 6.80962354e-02 -3.33716661e-01
-1.53296888e-01 8.83349925e-02 -1.13904512e+00 4.22469139e-01
7.26524174e-01 -6.45264268e-01 -8.62889588e-01 3.18513066e-01
5.68112671e-01 -5.50786674e-01 -4.76685911e-01 3.72214206e-02
2.60647714e-01 -1.01756644e+00 1.08220029e+00 -3.58468741e-01
4.95146930e-01 -5.68906009e-01 -1.63868498e-02 -1.93202806e+00
-7.55394757e-01 -1.13116276e+00 -3.71157914e-01 6.67693079e-01
-1.44247845e-01 -7.05869079e-01 2.81309634e-01 5.98043799e-01
-1.27963588e-01 -1.60421729e-01 -5.34509659e-01 -8.62157524e-01
-2.73099571e-01 6.42475933e-02 2.21303850e-01 1.15940559e+00
1.94435850e-01 2.27037594e-02 -8.44095170e-01 5.48674464e-01
1.11698222e+00 1.84093639e-01 2.94887483e-01 -6.67726636e-01
-4.08203393e-01 -3.15590173e-01 -7.98251703e-02 -1.73621392e+00
4.53550927e-03 -6.04889452e-01 5.14245093e-01 -1.34234309e+00
6.23614132e-01 -2.10555315e-01 -7.08554015e-02 -8.60628411e-02
-6.31216943e-01 1.03635922e-01 8.38987157e-02 6.20205700e-01
-9.19741392e-02 3.81533742e-01 1.75369406e+00 -1.93398520e-01
-1.83693156e-01 2.07099915e-01 -9.63477314e-01 8.47006023e-01
5.67197084e-01 -1.61512643e-01 -4.07714128e-01 -6.34912789e-01
-2.16593578e-01 1.55076429e-01 6.28272653e-01 -6.17130339e-01
2.67987788e-01 -2.44548991e-01 2.00469136e-01 1.46903634e-01
1.83713779e-01 -7.68013418e-01 3.84248763e-01 4.22938287e-01
-2.72571832e-01 -5.02223969e-01 8.62121210e-02 6.99390411e-01
-2.66773283e-01 -5.63578308e-01 1.13800359e+00 -1.93720683e-01
-5.69169700e-01 -6.21076822e-02 -2.95285612e-01 7.97710866e-02
5.70986569e-01 -9.25516561e-02 -6.15826190e-01 -9.80078876e-01
-7.28062153e-01 -2.51548082e-01 8.18268061e-01 -8.16559494e-02
6.76283121e-01 -1.16613185e+00 -7.99407840e-01 1.46043137e-01
-4.97591883e-01 -3.73906732e-01 2.63348460e-01 1.05637479e+00
-7.54650116e-01 2.64145702e-01 -1.29559711e-01 -3.40966195e-01
-1.13768363e+00 8.65345180e-01 6.03974581e-01 2.04002231e-01
-6.44938946e-01 8.33407104e-01 6.92338884e-01 5.65257847e-01
-9.31673348e-02 -4.72221881e-01 2.32847765e-01 -6.46013379e-01
8.85805368e-01 5.81950784e-01 -3.44717115e-01 -6.33943677e-01
-2.10865051e-01 8.34965289e-01 1.17711410e-01 -2.03457475e-01
1.20809233e+00 -7.83596039e-01 -5.68457842e-01 4.71137874e-02
1.30146575e+00 5.89760900e-01 -1.59196758e+00 -2.02151760e-01
-2.65213341e-01 -9.41252232e-01 4.28361654e-01 -5.00511706e-01
-1.11165202e+00 6.25653505e-01 4.77589428e-01 6.14338398e-01
1.36702836e+00 3.51618566e-02 6.42033339e-01 4.50121425e-02
1.74285129e-01 -4.31892931e-01 -9.77997109e-02 2.70502120e-01
1.26745808e+00 -1.04785216e+00 5.95740080e-02 -6.16760373e-01
-3.08883160e-01 1.16996038e+00 -7.28216618e-02 -3.53286862e-01
8.31664145e-01 1.37437090e-01 6.70013502e-02 -4.75767553e-02
-1.83037236e-01 -9.01468098e-02 4.13746923e-01 6.13515019e-01
2.44443282e-01 4.05147709e-02 -1.01943396e-01 2.13664770e-01
1.04086563e-01 2.42147312e-01 1.06180251e+00 4.73844171e-01
-8.32624495e-01 -5.69201231e-01 -1.00214970e+00 -8.93789977e-02
-5.04495442e-01 -3.50083590e-01 -8.32018815e-03 1.95159912e-01
-2.22100124e-01 1.06489098e+00 -2.43605733e-01 1.43655753e-02
-7.09740147e-02 -2.95526057e-01 8.36118758e-01 -4.21741873e-01
2.44706735e-01 1.38570547e-01 -3.92177314e-01 -4.32926863e-01
-5.44785738e-01 -3.92710209e-01 -6.77097023e-01 -5.00305474e-01
-4.60693955e-01 3.61332208e-01 2.08013877e-01 8.00024569e-01
1.32766813e-01 1.65715069e-01 8.75735998e-01 -9.03220713e-01
-6.22608542e-01 -9.63004827e-01 -9.84933138e-01 4.31676686e-01
1.24662685e+00 -1.42500997e-01 -9.09435511e-01 7.18703866e-01] | [11.623647689819336, -2.7639243602752686] |
c68fa6fb-ef76-40e1-8c3e-96368d91704d | cypur-nn-crop-yield-prediction-using | 2011.13265 | null | https://arxiv.org/abs/2011.13265v1 | https://arxiv.org/pdf/2011.13265v1.pdf | CYPUR-NN: Crop Yield Prediction Using Regression and Neural Networks | Our recent study using historic data of paddy yield and associated conditions include humidity, luminescence, and temperature. By incorporating regression models and neural networks (NN), one can produce highly satisfactory forecasting of paddy yield. Simulations indicate that our model can predict paddy yield with high accuracy while concurrently detecting diseases that may exist and are oblivious to the human eye. Crop Yield Prediction Using Regression and Neural Networks (CYPUR-NN) is developed here as a system that will facilitate agriculturists and farmers to predict yield from a picture or by entering values via a web interface. CYPUR-NN has been tested on stock images and the experimental results are promising. | ['A Balachandra', 'Thulasiram Gunta', 'Varun Yadav', 'Anirudh Hebbar', 'Sandesh Ramesh'] | 2020-11-26 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction'] | ['computer-vision', 'miscellaneous'] | [-2.50869036e-01 -3.12314898e-01 -1.57031238e-01 -1.30123794e-01
1.60674796e-01 -5.38514733e-01 5.83240688e-02 1.94443375e-01
2.15537623e-01 7.91311145e-01 -3.37914795e-01 -8.36848557e-01
1.19714625e-01 -1.73358643e+00 -3.28306645e-01 -7.34598279e-01
-4.18923140e-01 -7.85299316e-02 2.37967014e-01 -7.26562023e-01
1.41728476e-01 1.09121144e+00 -1.57183504e+00 -1.80295348e-01
6.17727518e-01 8.16953361e-01 7.66476572e-01 8.60910356e-01
1.08038403e-01 4.70192581e-01 -6.03842974e-01 2.30888233e-01
3.60003471e-01 -8.45806114e-03 -2.00255752e-01 -1.83832675e-01
-9.52406526e-02 -7.84202993e-01 5.86672872e-02 7.28287160e-01
3.27716410e-01 -4.62537408e-01 7.97766626e-01 -9.84177530e-01
-1.19102275e+00 5.97573578e-01 -8.93472910e-01 -2.27962360e-01
-1.51224449e-01 1.90068692e-01 2.42874488e-01 -6.83610559e-01
3.92377526e-01 1.15980625e+00 8.10563922e-01 -1.82699915e-02
-1.35437906e+00 -6.46629512e-01 -3.76684427e-01 4.03967872e-02
-1.38902318e+00 5.18953912e-02 2.53101230e-01 -1.96209878e-01
8.52006495e-01 3.51892591e-01 8.74151230e-01 1.47383571e-01
7.76094794e-01 4.33150172e-01 9.02450383e-01 -5.47625542e-01
1.07341200e-01 6.37765676e-02 -6.95265532e-02 6.83825016e-01
4.13852572e-01 5.43229401e-01 2.15344131e-04 4.88650389e-02
1.46456015e+00 1.48213103e-01 -4.60396931e-02 -2.64584441e-02
-5.74054837e-01 8.94727170e-01 7.27889657e-01 -3.40151601e-02
-9.06726182e-01 -1.23016454e-01 1.23038307e-01 3.79165679e-01
6.49683356e-01 4.27894831e-01 -9.01195884e-01 3.89289826e-01
-8.65055203e-01 1.18456140e-01 9.67379153e-01 8.43561590e-01
5.36288261e-01 6.27358198e-01 2.97843099e-01 9.14995849e-01
1.93524271e-01 1.25932837e+00 -6.89926967e-02 -8.99664104e-01
-3.41804355e-01 5.15637696e-01 4.05446738e-01 -1.43604755e+00
-2.39584193e-01 2.94680774e-01 -9.14248884e-01 8.64349008e-01
2.90659666e-01 -4.44070697e-01 -9.76490676e-01 1.16580021e+00
-2.05150634e-01 -2.24341318e-01 4.35505241e-01 7.28838265e-01
1.02174211e+00 1.37108994e+00 3.87645662e-01 -3.01264614e-01
1.27393103e+00 -3.06557089e-01 -9.60694909e-01 1.09482937e-01
6.66859806e-01 -8.14262450e-01 5.23533642e-01 5.32806516e-01
-9.95288849e-01 -4.39073294e-01 -9.30762172e-01 5.62504113e-01
-1.08992183e+00 7.71534860e-01 8.34753990e-01 5.03437281e-01
-1.08153176e+00 9.08172607e-01 -9.29374635e-01 -9.69631195e-01
-1.62479520e-01 4.10312831e-01 -3.76662612e-01 2.68596798e-01
-9.78719831e-01 1.60183203e+00 5.41281939e-01 8.23713064e-01
-5.73801816e-01 -4.54811096e-01 -7.86113501e-01 3.34792644e-01
1.23337328e-01 2.41640538e-01 7.51325727e-01 -4.46753502e-01
-1.54256022e+00 5.94190061e-01 3.43820304e-02 -3.62734705e-01
-8.70067328e-02 -4.13637638e-01 -6.88309968e-01 -3.16806361e-02
-2.73271710e-01 7.36021221e-01 9.60107818e-02 -1.18580902e+00
-6.34114444e-01 -4.54051286e-01 -5.21568537e-01 -1.20624736e-01
-1.53605223e-01 2.16819912e-01 3.32438052e-01 -4.44597185e-01
5.42005539e-01 -6.88340485e-01 -4.16123211e-01 3.74155045e-01
1.29226148e-01 2.16616035e-01 1.26059031e+00 -1.20256269e+00
8.50614011e-01 -1.49570715e+00 -6.40273750e-01 4.41847295e-01
-3.32827806e-01 7.00341821e-01 -2.95962572e-01 4.59278554e-01
-5.08910492e-02 8.27306733e-02 8.99067745e-02 1.11972082e+00
-5.09165108e-01 4.71958160e-01 -5.52494787e-02 1.24941096e-01
8.17736685e-01 5.34380078e-01 -5.88962615e-01 -1.92456529e-01
5.73484659e-01 3.72518629e-01 1.31213605e-01 4.69331920e-01
1.87998209e-02 -4.55628812e-01 -5.24467468e-01 1.29024553e+00
1.40702224e+00 1.28518939e-01 2.92646438e-01 -1.58602506e-01
-6.89086020e-01 -5.82119226e-01 -9.46161389e-01 5.07865846e-01
-4.95661914e-01 5.81950605e-01 1.04883455e-01 -1.05004132e+00
1.95162249e+00 4.28577423e-01 7.62344971e-02 -3.09065431e-01
-4.69052568e-02 1.34784818e-01 -2.99416363e-01 -7.14867651e-01
8.14713061e-01 4.72877681e-01 2.70554096e-01 1.48293883e-01
2.49282271e-02 -2.45845988e-01 1.54876843e-01 -3.40729535e-01
6.80713832e-01 7.55714893e-01 4.15752083e-01 -5.27732849e-01
3.31357926e-01 5.14178455e-01 4.69153553e-01 4.80001003e-01
-1.56052291e-01 1.90026000e-01 5.42926490e-01 -7.22804129e-01
-1.08200312e+00 -1.09689760e+00 -2.24886358e-01 1.25284266e+00
-6.17926121e-02 5.22960305e-01 -2.44468302e-01 -9.83308302e-04
3.25606763e-01 6.90207839e-01 -5.37085056e-01 1.23493195e-01
-1.92481101e-01 -1.40432072e+00 2.25594893e-01 9.12325442e-01
5.10975063e-01 -1.41463625e+00 -5.77938139e-01 5.00740945e-01
2.38032401e-01 -7.30173349e-01 5.84683895e-01 6.67522788e-01
-1.12399530e+00 -5.11271834e-01 -1.03365350e+00 -8.26077521e-01
6.98697269e-01 5.02190888e-01 1.19318771e+00 7.42335469e-02
-5.05539119e-01 -3.27525824e-01 -4.13694054e-01 -7.28375852e-01
-7.04907238e-01 -3.05534843e-02 -3.06045771e-01 -1.08125913e+00
9.81578171e-01 -5.49344182e-01 -4.74923849e-01 3.56141299e-01
-5.78007281e-01 5.40881278e-03 5.99579155e-01 6.77652895e-01
2.51988083e-01 2.13484969e-02 9.37409759e-01 -5.87711930e-01
3.95804793e-01 -6.13547206e-01 -1.32365561e+00 6.68126285e-01
-6.90999508e-01 -1.95674852e-01 4.08491492e-01 -3.24594975e-01
-1.08937228e+00 3.58423144e-01 1.32720530e-01 3.19916844e-01
-5.47453403e-01 7.06153333e-01 -1.25059327e-02 -3.28577533e-02
5.93769550e-01 1.80133302e-02 3.74519259e-01 -3.89048278e-01
9.11130607e-02 7.31385529e-01 8.85690331e-01 -1.44572347e-01
6.29251301e-01 -1.40659615e-01 2.79594421e-01 -1.11977625e+00
-7.72994012e-02 -1.20024681e-01 -8.31414461e-01 -5.71460247e-01
5.36751151e-01 -7.85470963e-01 -9.47891593e-01 9.14228797e-01
-9.96517539e-01 -1.49690047e-01 3.55789304e-01 6.57127976e-01
-7.54342079e-02 -2.05935791e-01 -6.44686162e-01 -1.12791932e+00
-5.14258087e-01 -5.49525797e-01 3.76700670e-01 7.16656387e-01
4.03126106e-02 -8.76467228e-01 -2.48436034e-01 -3.38855773e-01
7.06507623e-01 5.29985249e-01 8.39963794e-01 -1.97757185e-01
-1.10646993e-01 -3.68982822e-01 -7.77674437e-01 1.54947430e-01
5.57877958e-01 1.04267442e+00 -9.29419756e-01 2.11744234e-01
-6.99528515e-01 -3.61384779e-01 4.72278208e-01 1.05911970e+00
8.68234217e-01 6.90964535e-02 -3.40855479e-01 2.56202310e-01
1.79859996e+00 5.59386075e-01 6.67884052e-01 3.52816880e-01
-7.60865808e-02 8.73430610e-01 9.61054087e-01 5.29811382e-01
4.94836159e-02 -9.94406417e-02 7.50547409e-01 -8.04842889e-01
4.69673842e-01 -2.15767026e-01 2.04511046e-01 4.05586004e-01
-6.12483144e-01 -3.86922896e-01 -8.02959442e-01 6.40900016e-01
-1.67061365e+00 -1.06582296e+00 -3.53898764e-01 1.92473149e+00
6.33869171e-01 -9.37320590e-02 -4.89970297e-02 1.27197877e-01
7.35708416e-01 -9.41614807e-02 -4.46812302e-01 -1.14876783e+00
-3.24216455e-01 4.30654228e-01 1.43422043e+00 1.09348521e-01
-9.90555286e-01 1.24710333e+00 7.72654104e+00 2.89516568e-01
-1.28069520e+00 -7.14653790e-01 9.14349854e-01 5.27585804e-01
3.55846494e-01 -5.92329837e-02 -8.13996017e-01 -3.12992811e-01
8.65286112e-01 -1.20919012e-02 5.87770045e-02 8.51251602e-01
8.72755289e-01 -8.77897382e-01 -3.21349323e-01 -1.79397941e-01
-4.24714655e-01 -1.17122555e+00 -2.47980118e-01 -2.75635839e-01
4.15624499e-01 -3.44405383e-01 -2.65111029e-01 1.94955543e-01
1.13942587e+00 -8.83633673e-01 -2.67977148e-01 6.15323663e-01
5.18283010e-01 -7.65304208e-01 1.13348591e+00 5.20118661e-02
-1.33765543e+00 1.92013532e-01 -9.78812516e-01 -3.30349654e-01
-2.11036697e-01 1.79715380e-01 -8.16413283e-01 4.75124359e-01
9.61439669e-01 6.77142560e-01 -4.56835508e-01 8.69949818e-01
1.12296052e-01 8.16491425e-01 -6.25028610e-01 -4.50347215e-01
2.50036567e-01 -4.84473348e-01 -1.92919344e-01 1.19015574e+00
8.79976511e-01 3.34706992e-01 -2.50525504e-01 1.10810900e+00
6.73402309e-01 3.44130576e-01 -7.61945307e-01 5.13722599e-02
5.45438826e-01 1.23192883e+00 -9.38209057e-01 9.24167931e-02
-1.27483800e-01 4.59654272e-01 -4.36416954e-01 3.30469579e-01
-2.48172626e-01 -5.66896915e-01 3.81754547e-01 -2.25735202e-01
4.94556874e-01 -1.53672010e-01 -4.26625490e-01 -4.14796174e-01
-5.02602398e-01 -2.41212234e-01 -1.27862394e-01 -1.37835336e+00
-9.51353967e-01 2.05037862e-01 -2.49406649e-03 -1.09238970e+00
-2.57247746e-01 -1.22361624e+00 -8.62976968e-01 1.36873877e+00
-1.44839752e+00 -1.31472838e+00 -4.10754919e-01 -1.29119202e-01
2.97896892e-01 -5.18447220e-01 1.53474104e+00 -3.26352626e-01
-6.10286832e-01 -6.37152642e-02 5.74260354e-01 -2.08041184e-02
4.41279650e-01 -1.12614715e+00 2.76902914e-01 6.35855377e-01
-7.80246079e-01 -2.46208802e-01 9.09541845e-01 -8.61415505e-01
-1.12697303e+00 -1.03431511e+00 8.86178851e-01 4.21671629e-01
6.49768114e-01 1.90142989e-01 -9.99933481e-01 3.83845776e-01
1.55091569e-01 -2.72150934e-02 2.70538896e-01 1.04442853e-02
2.85653710e-01 -3.28254372e-01 -1.37771738e+00 4.56685185e-01
9.40342247e-02 2.36801654e-01 3.39647271e-02 6.57894164e-02
2.71117896e-01 1.97299551e-02 -1.29298735e+00 5.86700201e-01
1.11438358e+00 -4.65774268e-01 8.46947908e-01 -2.95612007e-01
7.45404601e-01 -1.13891298e-02 -2.24613443e-01 -1.44776607e+00
-4.20017272e-01 -1.43847749e-01 2.90370047e-01 1.40263855e+00
5.60635209e-01 -3.31363142e-01 7.88495660e-01 6.29832029e-01
7.53095746e-01 -2.43258819e-01 1.88283578e-01 -7.28020489e-01
2.53705978e-01 6.41459450e-02 9.14154708e-01 6.51614368e-01
-2.06014574e-01 -4.64912981e-01 -6.56728745e-01 8.36894751e-01
2.67161191e-01 -4.62889634e-02 5.66410124e-01 -1.37994730e+00
4.56550807e-01 -2.27643296e-01 -4.87680942e-01 -4.29260194e-01
-2.67951399e-01 -5.92897348e-02 1.59471855e-01 -1.31667233e+00
-1.27204910e-01 -4.71230000e-01 -2.23999858e-01 9.43679333e-01
-2.23430902e-01 8.52280110e-02 -1.96295530e-02 -1.55953288e-01
7.34453499e-01 1.61690533e-01 1.05930996e+00 1.67606503e-01
-6.31274521e-01 3.52598697e-01 -4.06780243e-01 7.65890837e-01
1.13611066e+00 -3.73544961e-01 -4.16206103e-03 -3.57986331e-01
-1.16309702e-01 7.86428213e-01 2.63771921e-01 -6.60467863e-01
-4.02490288e-01 -7.94898987e-01 9.09560502e-01 -1.34196186e+00
3.46362554e-02 -1.04327118e+00 1.79943025e-01 7.29832828e-01
-1.55891895e-01 4.20617670e-01 7.07585812e-01 -1.51390582e-01
9.51179340e-02 -4.91273105e-01 4.09632593e-01 -8.03929493e-02
-1.00042140e+00 -7.67720044e-02 -9.32130456e-01 -1.12672710e+00
9.37853694e-01 -3.77447009e-01 -4.34514880e-01 -4.85749692e-01
-9.39761758e-01 5.57363331e-01 6.07989822e-03 4.88280922e-01
5.61125457e-01 -1.32389319e+00 -1.11872387e+00 5.63073575e-01
8.10575858e-03 -5.55897355e-01 -2.74250269e-01 7.70041347e-02
-1.65400338e+00 3.02291065e-01 -1.24146748e+00 -5.38492322e-01
-1.37296367e+00 2.24380583e-01 2.90500492e-01 -8.49847868e-02
-1.45572439e-01 3.56194615e-01 -2.70066202e-01 -5.44675589e-01
8.89722034e-02 -3.44860703e-01 -8.13975930e-01 -1.43733323e-01
4.07014519e-01 4.60025728e-01 -2.14240670e-01 -2.99463540e-01
4.21696492e-02 3.29203963e-01 1.33510426e-01 2.89280444e-01
1.89612627e+00 -2.04317346e-02 -5.24130799e-02 4.74485576e-01
4.65777904e-01 -8.09295654e-01 -1.24135637e+00 -3.46504264e-02
9.83089805e-02 -3.11117828e-01 4.31281805e-01 -1.14618552e+00
-1.51198447e+00 6.90853298e-01 1.35453498e+00 5.38653076e-01
1.51675189e+00 -7.60507226e-01 4.93727863e-01 8.67025614e-01
-1.02986671e-01 -1.17929935e+00 -7.05363870e-01 2.35792130e-01
8.24027359e-01 -1.63492405e+00 2.24524051e-01 -3.11730951e-01
-4.92377728e-01 1.70541835e+00 8.01896632e-01 -3.21819782e-01
1.02859008e+00 9.10063684e-01 6.24116957e-01 2.62530178e-01
-9.11473334e-01 -1.63290218e-01 -4.39203143e-01 1.13894236e+00
7.31277466e-01 5.63174725e-01 -3.26068550e-01 9.91788805e-02
5.94222881e-02 6.68643475e-01 5.73059618e-01 1.15612257e+00
-8.97796273e-01 -8.97278309e-01 -9.97911572e-01 6.11309528e-01
-4.28301424e-01 -2.68506780e-02 -2.07606032e-01 1.01264679e+00
-1.12759545e-01 8.43541741e-01 2.80967712e-01 -7.73744732e-02
1.35099977e-01 -3.27551007e-01 6.93344176e-02 -5.28798029e-02
-5.01374722e-01 3.30707073e-01 -4.23990861e-02 -7.32040852e-02
-5.46011508e-01 -4.74691063e-01 -7.58286774e-01 -9.86285686e-01
-8.31385672e-01 -1.64118424e-01 9.15140629e-01 4.51436281e-01
-2.75133580e-01 2.96486795e-01 1.09026039e+00 -7.99953580e-01
-2.58863062e-01 -1.21481574e+00 -1.30032218e+00 -5.39013803e-01
-2.65921000e-02 -3.30308914e-01 6.41842112e-02 4.59726483e-01] | [9.346327781677246, -1.6144925355911255] |
a2c58f8e-e93f-4c14-9c0a-18276408ff5b | revisiting-self-training-for-few-shot | 2110.01256 | null | https://arxiv.org/abs/2110.01256v1 | https://arxiv.org/pdf/2110.01256v1.pdf | Revisiting Self-Training for Few-Shot Learning of Language Model | As unlabeled data carry rich task-relevant information, they are proven useful for few-shot learning of language model. The question is how to effectively make use of such data. In this work, we revisit the self-training technique for language model fine-tuning and present a state-of-the-art prompt-based few-shot learner, SFLM. Given two views of a text sample via weak and strong augmentation techniques, SFLM generates a pseudo label on the weakly augmented version. Then, the model predicts the same pseudo label when fine-tuned with the strongly augmented version. This simple approach is shown to outperform other state-of-the-art supervised and semi-supervised counterparts on six sentence classification and six sentence-pair classification benchmarking tasks. In addition, SFLM only relies on a few in-domain unlabeled data. We conduct a comprehensive analysis to demonstrate the robustness of our proposed approach under various settings, including augmentation techniques, model scale, and few-shot knowledge transfer across tasks. | ['Haizhou Li', 'Ran Cheng', 'Grandee Lee', 'Chen Zhang', 'Yan Zhang', 'Yiming Chen'] | 2021-10-04 | null | https://aclanthology.org/2021.emnlp-main.718 | https://aclanthology.org/2021.emnlp-main.718.pdf | emnlp-2021-11 | ['sentence-pair-classification', 'sentence-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.87696093e-01 1.49715513e-01 -6.23492718e-01 -5.75142086e-01
-1.26331615e+00 -3.99880797e-01 7.43693173e-01 1.97198004e-01
-6.54016733e-01 9.04478312e-01 4.10834223e-01 -8.00860524e-02
3.03684324e-01 -5.17242789e-01 -4.30768281e-01 -4.83378977e-01
4.06944573e-01 6.00711763e-01 3.04572821e-01 -4.16855901e-01
1.19741589e-01 -1.12646252e-01 -1.36251652e+00 3.88968676e-01
6.70048416e-01 7.42444515e-01 2.25812614e-01 7.29631782e-01
-4.52445000e-01 9.50714767e-01 -3.75357687e-01 -3.75991583e-01
1.21993842e-02 -4.55370367e-01 -9.19662714e-01 2.38409266e-01
2.83727765e-01 -2.64311433e-01 -3.44912499e-01 7.75884748e-01
8.01434755e-01 7.70709932e-01 7.25368798e-01 -8.98861825e-01
-9.17740464e-01 7.59815812e-01 -4.51846957e-01 4.36125189e-01
1.05906017e-01 1.70035496e-01 1.08729386e+00 -1.37153900e+00
7.90981650e-01 1.02790618e+00 4.34451520e-01 1.18516266e+00
-1.19927502e+00 -4.79705781e-01 1.89867958e-01 1.51399776e-01
-1.00143957e+00 -8.20114732e-01 7.89526403e-01 -4.71114367e-01
1.12772250e+00 -1.35374695e-01 -5.72880358e-03 1.34939051e+00
-1.41162261e-01 8.53777647e-01 1.20536613e+00 -9.97494280e-01
6.38248146e-01 4.81951296e-01 8.12387049e-01 5.77682555e-01
-1.42424762e-01 -2.16763213e-01 -8.15874875e-01 -2.96050221e-01
1.48108020e-01 1.16448008e-01 -1.22431353e-01 -5.65731704e-01
-1.00389647e+00 9.52823579e-01 4.07504179e-02 4.78863508e-01
1.21097490e-02 -1.37219220e-01 6.89919710e-01 4.99908447e-01
1.04693758e+00 4.33101267e-01 -6.49724543e-01 -3.57874066e-01
-9.09530461e-01 -2.13369250e-01 7.12971270e-01 1.00720274e+00
8.76893938e-01 1.18325781e-02 -7.56710708e-01 1.19704533e+00
-3.58309457e-03 1.31559327e-01 9.21759188e-01 -4.91416782e-01
6.38101637e-01 3.40261370e-01 6.29636794e-02 -1.30454926e-02
-1.99296162e-01 -3.44904780e-01 -7.36957252e-01 -2.44728580e-01
-1.39855007e-02 -3.48496616e-01 -1.03057444e+00 1.76005554e+00
3.11207503e-01 3.69458288e-01 3.32631171e-01 4.32993352e-01
8.76869380e-01 5.65641761e-01 3.17668378e-01 -4.78534669e-01
1.17941391e+00 -1.37021911e+00 -8.64764035e-01 -4.24113661e-01
1.11745465e+00 -4.81220007e-01 1.46706522e+00 -1.78583652e-01
-8.78517330e-01 -7.15211451e-01 -9.47613120e-01 -1.39892161e-01
-3.50483418e-01 -6.43680096e-02 2.22419158e-01 5.43713927e-01
-7.65205681e-01 6.16650879e-01 -6.02960646e-01 -6.58968031e-01
4.71140325e-01 -9.12977457e-02 -3.72487754e-01 -3.93988103e-01
-1.38825297e+00 1.04892719e+00 4.66833591e-01 -6.11172557e-01
-1.05643773e+00 -8.96239340e-01 -1.03574669e+00 3.00472021e-01
4.27533358e-01 -6.03282273e-01 1.72364581e+00 -4.69654202e-01
-1.48111665e+00 1.00714397e+00 -3.91483486e-01 -5.59743762e-01
2.74832785e-01 -1.89227715e-01 -3.88804793e-01 2.78905705e-02
1.48806572e-01 4.90261316e-01 8.94522727e-01 -1.05080605e+00
-4.67818677e-01 -3.15952569e-01 -7.22663198e-03 2.86665678e-01
-9.19425666e-01 7.12689534e-02 -2.91569471e-01 -5.88123202e-01
-5.19268394e-01 -5.41615486e-01 -2.25614324e-01 -1.56554744e-01
-1.96641967e-01 -4.19834256e-01 7.10543692e-01 -2.83657908e-01
1.43201518e+00 -2.06943893e+00 -6.73760921e-02 -5.32254636e-01
6.56612962e-02 6.41202569e-01 -3.65707695e-01 6.73552275e-01
-9.23478603e-03 4.31688353e-02 -3.41443390e-01 -8.14690709e-01
2.41160914e-02 1.39474258e-01 -3.38539511e-01 3.01477700e-01
1.37565523e-01 1.08085644e+00 -1.21525466e+00 -5.55056751e-01
3.00888658e-01 1.68376088e-01 -1.53232381e-01 4.71028388e-01
-1.92172050e-01 1.63280025e-01 -4.46595728e-01 3.70061010e-01
2.91187614e-01 -3.89235616e-01 -1.06256850e-01 1.79374926e-02
2.72857964e-01 8.80839005e-02 -6.88843012e-01 1.97876990e+00
-6.74401820e-01 4.04777050e-01 -3.97678941e-01 -9.89233732e-01
1.05159104e+00 5.40288568e-01 5.59892096e-02 -5.62292576e-01
6.67503923e-02 -1.10672787e-02 -3.19436461e-01 -6.54641747e-01
3.31438690e-01 -5.31084001e-01 -1.35295704e-01 9.31044579e-01
7.37961292e-01 5.10286242e-02 2.53766596e-01 4.57501620e-01
1.11262524e+00 2.81130690e-02 8.45242977e-01 -1.33549854e-01
4.99737769e-01 -2.76705056e-01 2.16753215e-01 1.03083277e+00
-5.85200191e-01 4.96143192e-01 1.45925432e-01 -3.65091890e-01
-1.14573455e+00 -8.94617736e-01 -8.48715529e-02 1.81542814e+00
-3.05447549e-01 -2.57380545e-01 -7.06473529e-01 -9.92493153e-01
-2.66577214e-01 1.22028887e+00 -9.06284273e-01 -4.51219440e-01
-8.32472667e-02 -5.53912461e-01 2.95127541e-01 6.44753575e-01
1.98085070e-01 -1.08890808e+00 -3.30477059e-01 2.55864322e-01
1.08308182e-03 -1.14298344e+00 -6.76273704e-01 4.33559030e-01
-8.34862232e-01 -8.34234953e-01 -8.61591876e-01 -8.58263791e-01
6.10218048e-01 6.61055446e-01 1.17149770e+00 -3.24935585e-01
-9.53833461e-02 5.62431037e-01 -6.42309606e-01 -3.07051688e-01
-5.56256175e-01 2.51810312e-01 2.56049722e-01 7.17388093e-02
7.43532002e-01 -4.44915235e-01 -1.38388842e-01 2.12009475e-02
-7.90419281e-01 2.58045290e-02 3.23120803e-01 1.24684334e+00
3.73886406e-01 -4.66836900e-01 8.80257547e-01 -1.49292803e+00
9.65364337e-01 -5.25274813e-01 -4.93355729e-02 7.51962185e-01
-5.88966131e-01 2.85122156e-01 7.53620744e-01 -5.43765783e-01
-1.45968378e+00 -5.24884760e-02 1.73456252e-01 -5.90476155e-01
-3.41014385e-01 4.67254341e-01 1.32107213e-01 1.81308180e-01
1.09676790e+00 2.18122363e-01 -2.01355800e-01 -6.89256310e-01
7.04332888e-01 1.06298530e+00 3.22546184e-01 -4.57286447e-01
6.96426153e-01 4.18627501e-01 -5.17752767e-01 -8.26977253e-01
-1.64679015e+00 -1.01533723e+00 -1.05025065e+00 -1.53399808e-02
6.19545519e-01 -9.59248245e-01 7.42281079e-02 2.72675335e-01
-1.07437527e+00 -4.26302433e-01 -6.57132387e-01 2.84351200e-01
-5.06355524e-01 2.41328806e-01 -6.13956213e-01 -1.00858474e+00
-6.73419058e-01 -6.14631355e-01 1.05957150e+00 1.04027227e-01
-2.26201370e-01 -1.25271499e+00 6.38678312e-01 4.12037194e-01
3.26681972e-01 -3.72569472e-01 8.46733034e-01 -1.38399923e+00
2.53905028e-01 -3.85228068e-01 -5.98793589e-02 2.92595297e-01
2.76931912e-01 -5.23161769e-01 -1.55959547e+00 -4.81377989e-01
1.60506338e-01 -1.08627236e+00 1.10148406e+00 1.70005843e-01
8.91064942e-01 -5.04546762e-02 -1.58540934e-01 2.92114794e-01
1.22241771e+00 -1.88727543e-01 1.19963765e-01 9.44726616e-02
6.60518467e-01 6.18491948e-01 8.38669896e-01 5.74625134e-01
2.31513932e-01 4.98823255e-01 -7.34507143e-02 1.16591163e-01
-6.21284731e-02 -3.81962985e-01 2.50295103e-01 1.21338916e+00
2.30958939e-01 -1.99831173e-01 -8.80821586e-01 4.75111276e-01
-2.02747989e+00 -1.06088531e+00 5.42610407e-01 2.19964147e+00
1.08145189e+00 3.55117381e-01 -7.10296631e-02 -1.90278292e-01
9.12205875e-01 4.21223223e-01 -6.82456493e-01 -2.79087663e-01
5.11686951e-02 3.54959369e-01 1.75685540e-01 5.83432853e-01
-1.25134242e+00 1.21648145e+00 6.17619658e+00 1.07315993e+00
-6.96965098e-01 6.81776762e-01 7.85026729e-01 -2.45285675e-01
-2.95776397e-01 -1.00551136e-01 -1.11601710e+00 1.04918592e-01
1.25024354e+00 -5.80454767e-01 2.09259257e-01 1.12798560e+00
1.24916106e-01 5.19755557e-02 -1.18159604e+00 8.29153538e-01
5.08405030e-01 -1.49062991e+00 6.25045523e-02 -3.63226771e-01
1.15667152e+00 3.15932989e-01 -2.22877651e-01 1.01211607e+00
3.55584621e-01 -7.38128603e-01 1.90553203e-01 4.20515031e-01
9.83262300e-01 -5.28101504e-01 7.96482742e-01 8.02115917e-01
-9.84650254e-01 -6.89714625e-02 -6.50569439e-01 -1.77107796e-01
3.31256479e-01 2.75365859e-01 -9.22690153e-01 2.59850562e-01
2.78946042e-01 7.01788783e-01 -6.79894745e-01 6.64633036e-01
-2.57781208e-01 7.04975665e-01 2.87550896e-01 -1.83845013e-01
2.51956075e-01 2.24298790e-01 2.77645409e-01 1.34103942e+00
-7.98559375e-03 2.19725996e-01 3.48172694e-01 5.31762362e-01
-3.95221710e-01 2.73915708e-01 -7.77393878e-01 -2.02522293e-01
6.92234933e-01 1.40527344e+00 -3.54563773e-01 -8.35228801e-01
-7.00144768e-01 9.50553715e-01 9.60204303e-01 4.28571969e-01
-3.08194965e-01 -2.04949334e-01 2.37408116e-01 -1.03090286e-01
2.62179703e-01 1.09099969e-01 -7.28121400e-02 -1.52413082e+00
-2.08219379e-01 -6.59671426e-01 5.06711006e-01 -6.86735511e-01
-1.67976654e+00 5.99199057e-01 -9.09716636e-02 -1.25144327e+00
-4.65099573e-01 -4.18667406e-01 -9.13640857e-01 7.80209720e-01
-1.58770001e+00 -1.23007774e+00 -1.08238086e-01 5.08213520e-01
1.27001595e+00 -6.06766999e-01 1.29412067e+00 -9.91448611e-02
-5.44651568e-01 6.66620433e-01 4.55920100e-01 1.11473896e-01
1.10951078e+00 -1.15086019e+00 7.12482512e-01 7.43875146e-01
3.65445822e-01 5.05330086e-01 6.54767811e-01 -5.77426136e-01
-1.04240751e+00 -1.23077881e+00 1.05373776e+00 -5.15211046e-01
9.55281138e-01 -5.73153436e-01 -1.16107607e+00 7.16066420e-01
3.03241968e-01 4.63251680e-01 1.06682825e+00 3.22583824e-01
-5.24655521e-01 1.37921795e-01 -9.75631893e-01 4.95592594e-01
7.55688190e-01 -8.47698212e-01 -1.04685462e+00 5.45319736e-01
1.02673388e+00 -2.47442964e-02 -6.69722855e-01 1.62340775e-01
1.98830172e-01 -7.28785336e-01 7.88031340e-01 -1.27459300e+00
4.87410665e-01 3.78464609e-01 -1.28902182e-01 -1.57399166e+00
-4.05540764e-01 -5.11692822e-01 -5.93086720e-01 1.33958340e+00
4.69348937e-01 -3.09224069e-01 6.77853703e-01 5.86769938e-01
-1.54717714e-01 -8.64908457e-01 -9.20355499e-01 -8.66462648e-01
2.16472253e-01 -3.22634608e-01 3.34545635e-02 1.14781380e+00
4.89541441e-01 1.03839564e+00 -7.09807515e-01 -5.37712336e-01
7.17284501e-01 -1.35676071e-01 7.35286951e-01 -1.16338444e+00
-4.98194307e-01 6.15745550e-03 -7.51946196e-02 -7.69163787e-01
5.55012107e-01 -8.37943852e-01 4.40876707e-02 -1.43514538e+00
7.28370309e-01 5.67018054e-03 -6.37375832e-01 5.21138787e-01
-6.37585282e-01 9.88161117e-02 2.00103611e-01 2.14911178e-01
-1.04159176e+00 8.85776997e-01 1.01056564e+00 -2.21964404e-01
-1.95775464e-01 1.14840582e-01 -5.56745887e-01 6.43214703e-01
9.06103969e-01 -4.54525083e-01 -6.73832297e-01 -1.71904176e-01
-1.95694193e-01 -1.13755814e-03 -1.50737450e-01 -7.18096256e-01
2.63131797e-01 -1.29700646e-01 1.27503231e-01 -3.64731371e-01
5.22340298e-01 -3.88012975e-01 -7.32512057e-01 2.98913032e-01
-9.50268030e-01 -3.78070593e-01 7.81942084e-02 8.22105765e-01
-1.06063113e-01 -7.81949639e-01 1.00649691e+00 -3.66135955e-01
-9.53676283e-01 4.07681257e-01 -1.25634491e-01 6.45346522e-01
1.12019753e+00 3.64214070e-02 -4.12180543e-01 -4.00270581e-01
-7.88547754e-01 2.64159232e-01 2.95662105e-01 5.88839650e-01
5.41475058e-01 -1.33913970e+00 -6.08522594e-01 6.62238300e-02
8.31450939e-01 -4.25296754e-01 5.17111838e-01 5.75218379e-01
2.84697503e-01 5.70556104e-01 -1.80255193e-02 -3.51687163e-01
-1.20370591e+00 1.02617991e+00 -4.73145656e-02 -6.30022824e-01
-5.57123482e-01 9.26540196e-01 2.58202642e-01 -7.60509789e-01
2.92066336e-01 2.12757722e-01 -3.21553320e-01 2.31008977e-01
9.68631983e-01 2.55833447e-01 1.66806672e-02 -3.34368467e-01
-1.02416709e-01 2.32545555e-01 -4.34123456e-01 -3.00064415e-01
1.27460718e+00 -3.30261230e-01 4.20990229e-01 1.26430738e+00
1.27243996e+00 -4.24502343e-01 -1.25673485e+00 -9.53200102e-01
3.29983830e-02 -2.53857374e-01 -1.44645842e-02 -6.66830122e-01
-3.36886019e-01 1.34775376e+00 3.31462413e-01 8.02508183e-03
6.43858731e-01 8.72441307e-02 6.51486695e-01 7.19323397e-01
3.12142640e-01 -1.41870022e+00 3.06338698e-01 7.41500854e-01
4.29461151e-01 -1.73906159e+00 -1.45527616e-01 -7.34003186e-02
-1.11916637e+00 9.22978401e-01 8.59583259e-01 1.14617668e-01
6.33896053e-01 4.15569134e-02 6.22911416e-02 4.84813526e-02
-1.30807078e+00 -2.85515815e-01 3.92951071e-01 6.33347750e-01
6.86438322e-01 -5.58483712e-02 -1.08540289e-01 6.32779598e-01
3.17387909e-01 1.28035992e-01 3.33172232e-01 1.09262204e+00
-9.11879063e-01 -1.03590751e+00 2.29945444e-02 6.62439525e-01
-1.88921332e-01 -4.84028339e-01 -2.47503117e-01 4.46740925e-01
-4.75338131e-01 9.04047072e-01 -1.18557855e-01 -1.72128588e-01
1.54392481e-01 6.97777867e-01 2.85115033e-01 -1.38456619e+00
-4.45830226e-01 -5.66438697e-02 1.44548580e-01 -1.94195315e-01
-3.47300440e-01 -3.14104855e-01 -9.98643219e-01 9.87766981e-02
-4.87245440e-01 2.53195435e-01 2.29964465e-01 1.35668027e+00
2.75499761e-01 5.42323589e-01 6.97479308e-01 -7.57557869e-01
-1.31420565e+00 -1.55370677e+00 -6.55183434e-01 6.12017691e-01
2.31958300e-01 -6.62062645e-01 -4.08062071e-01 1.58452794e-01] | [10.884246826171875, 7.79044246673584] |
9058c272-dec3-4686-b331-02301f764671 | fault-prognosis-in-particle-accelerator-power | 2209.1557 | null | https://arxiv.org/abs/2209.15570v1 | https://arxiv.org/pdf/2209.15570v1.pdf | Fault Prognosis in Particle Accelerator Power Electronics Using Ensemble Learning | Early fault detection and fault prognosis are crucial to ensure efficient and safe operations of complex engineering systems such as the Spallation Neutron Source (SNS) and its power electronics (high voltage converter modulators). Following an advanced experimental facility setup that mimics SNS operating conditions, the authors successfully conducted 21 fault prognosis experiments, where fault precursors are introduced in the system to a degree enough to cause degradation in the waveform signals, but not enough to reach a real fault. Nine different machine learning techniques based on ensemble trees, convolutional neural networks, support vector machines, and hierarchical voting ensembles are proposed to detect the fault precursors. Although all 9 models have shown a perfect and identical performance during the training and testing phase, the performance of most models has decreased in the prognosis phase once they got exposed to real-world data from the 21 experiments. The hierarchical voting ensemble, which features multiple layers of diverse models, maintains a distinguished performance in early detection of the fault precursors with 95% success rate (20/21 tests), followed by adaboost and extremely randomized trees with 52% and 48% success rates, respectively. The support vector machine models were the worst with only 24% success rate (5/21 tests). The study concluded that a successful implementation of machine learning in the SNS or particle accelerator power systems would require a major upgrade in the controller and the data acquisition system to facilitate streaming and handling big data for the machine learning models. In addition, this study shows that the best performing models were diverse and based on the ensemble concept to reduce the bias and hyperparameter sensitivity of individual models. | ['Sarah Cousineau', 'Pradeep Ramuhalli', 'Mark Wezensky', 'Chris Pappas', 'Majdi I. Radaideh'] | 2022-09-30 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-2.80547142e-01 -6.34775460e-02 3.58705640e-01 -1.70126170e-01
-3.06741714e-01 -8.94801617e-02 4.83255416e-01 1.14163600e-01
-2.21640453e-01 8.31579626e-01 -3.70628804e-01 -4.32254583e-01
-5.74724674e-01 -6.51174128e-01 -2.09431216e-01 -1.02483726e+00
-2.95572191e-01 8.65668535e-01 4.29460168e-01 -2.85532802e-01
3.91957194e-01 9.20753539e-01 -1.83039021e+00 5.08954525e-01
6.62656605e-01 1.09591937e+00 6.19162656e-02 7.62694538e-01
3.68106544e-01 8.94350171e-01 -1.14130020e+00 4.72990632e-01
4.30761697e-03 -2.50521034e-01 -6.81135774e-01 -4.88763839e-01
-2.44818985e-01 -4.60566804e-02 -3.86916965e-01 6.71144664e-01
8.46842945e-01 -2.05750808e-01 7.11164355e-01 -1.42664444e+00
2.26386324e-01 5.33500493e-01 3.75476591e-02 4.84412253e-01
2.95386195e-01 3.23851585e-01 3.74013245e-01 -8.81620765e-01
1.74017221e-01 5.32337070e-01 8.48162889e-01 2.87932128e-01
-7.37122476e-01 -7.60521233e-01 -7.11923420e-01 7.50050902e-01
-1.24201238e+00 -2.86539018e-01 3.32604378e-01 -5.90778232e-01
1.40700567e+00 4.06230122e-01 7.05932736e-01 7.73789883e-01
7.87045240e-01 -7.29326205e-03 1.11922610e+00 -5.70627332e-01
5.85724056e-01 1.43170848e-01 2.99224228e-01 3.57969254e-01
4.44414645e-01 5.07921457e-01 -2.70947248e-01 -1.90834060e-01
3.68288517e-01 -1.94264859e-01 -5.17883420e-01 1.78550944e-01
-6.53946519e-01 6.36501968e-01 3.69498730e-01 1.06519806e+00
-5.46503484e-01 -3.85449499e-01 6.10267222e-01 5.52417576e-01
2.15788454e-01 6.85857415e-01 -6.85675502e-01 -2.23040536e-01
-9.13391888e-01 5.61270863e-02 1.04009724e+00 2.90298074e-01
2.01366201e-01 4.41679358e-01 -1.35249957e-01 4.65558320e-01
2.01888904e-01 3.94325376e-01 6.68774247e-01 -1.36090294e-01
9.49772224e-02 6.30021751e-01 1.50097925e-02 -5.60102403e-01
-9.77196515e-01 -6.22935534e-01 -7.01194227e-01 7.13193953e-01
2.54226595e-01 -3.19536924e-01 -9.77640748e-01 7.53830075e-01
8.29291046e-02 -6.84966147e-02 9.53164548e-02 7.26595879e-01
6.67785943e-01 7.74474978e-01 -1.10809229e-01 -3.83437932e-01
1.20297742e+00 -5.51544666e-01 -7.97644079e-01 -5.75248734e-04
7.31918752e-01 -6.96118176e-01 4.62998003e-01 8.59448195e-01
-8.49385381e-01 -6.70828819e-01 -1.59561086e+00 7.69903719e-01
-2.24209234e-01 6.72282651e-02 3.54791969e-01 8.67775917e-01
-8.90430748e-01 1.20759404e+00 -7.11987436e-01 -1.73746169e-01
2.14067087e-01 5.79767942e-01 -2.29643986e-01 -7.19725043e-02
-1.37208593e+00 1.53252447e+00 3.81090999e-01 3.86521995e-01
-1.02113354e+00 -5.72818816e-01 -4.57802951e-01 4.38811071e-02
-1.48872808e-01 -4.61939394e-01 1.22068906e+00 -4.73816216e-01
-1.43800497e+00 1.55559525e-01 3.82717282e-01 -5.87687850e-01
4.29864734e-01 -1.41017258e-01 -9.10638869e-01 1.86292648e-01
-3.29225123e-01 -2.78749317e-01 5.46975493e-01 -8.47446501e-01
-8.63293767e-01 -2.10456699e-01 -4.39165264e-01 -2.26891503e-01
5.16573191e-02 8.91277790e-02 5.97097874e-01 -1.59141775e-02
2.18223020e-01 -6.19815946e-01 -1.83746308e-01 -7.09984124e-01
-5.66720814e-02 -4.33177918e-01 1.21394646e+00 -6.70995355e-01
1.04418504e+00 -1.90277386e+00 -1.42479226e-01 3.86480391e-01
-1.09370895e-01 3.91529679e-01 6.74896896e-01 6.72068238e-01
-5.41821003e-01 -3.82755518e-01 -1.25745729e-01 2.62856185e-01
-3.32341015e-01 3.52501661e-01 4.83697280e-02 6.37976348e-01
3.76833789e-02 1.62280858e-01 -6.60546780e-01 1.47024095e-01
5.13036609e-01 2.14645624e-01 -1.27212241e-01 2.22005323e-01
1.36999786e-01 3.66239905e-01 -1.31382078e-01 6.80104315e-01
5.78449070e-01 2.22475559e-01 -9.78972018e-02 -5.19913673e-01
-1.57262594e-01 -5.93251958e-02 -1.17271221e+00 9.82441127e-01
-4.56904858e-01 4.96445149e-01 1.20244265e-01 -1.30350208e+00
1.13687599e+00 8.67450535e-01 6.47648096e-01 -7.90105402e-01
3.74586761e-01 6.15361452e-01 2.91972667e-01 -9.68872070e-01
3.38861436e-01 -4.04338896e-01 3.50078568e-02 1.62407011e-01
2.14034066e-01 -4.60601479e-01 2.50493316e-03 -1.57603249e-01
1.37394547e+00 -3.99288714e-01 -6.10388033e-02 -6.45287335e-01
6.12548947e-01 1.44770429e-01 5.29795527e-01 4.98231173e-01
4.39779721e-02 4.04182017e-01 4.14732873e-01 -5.76967716e-01
-9.80767012e-01 -6.19277894e-01 -5.26052713e-01 3.64074886e-01
-1.46972343e-01 -1.93369046e-01 -4.72933173e-01 -6.98974252e-01
6.93387389e-02 1.07503104e+00 -6.55639693e-02 -5.29916525e-01
-3.98001581e-01 -1.04724991e+00 3.77768010e-01 5.05713642e-01
2.70146042e-01 -1.15784276e+00 -7.04400659e-01 2.83964992e-01
2.48979807e-01 -7.10162818e-01 6.66326821e-01 9.16052282e-01
-8.49402487e-01 -1.46468294e+00 -5.59290983e-02 -4.52788830e-01
6.99614584e-01 -4.46171135e-01 9.16000009e-01 3.98490459e-01
-7.16817498e-01 5.72518958e-03 -5.43611348e-01 -5.27926326e-01
-7.41947889e-01 -3.71614873e-01 3.21287721e-01 -4.65964347e-01
3.15014794e-02 -4.02168483e-01 -3.01695764e-01 4.83154535e-01
-6.71567142e-01 -5.04788816e-01 6.37813032e-01 1.20332789e+00
-1.40045760e-02 8.87220681e-01 7.09483027e-01 -6.37257457e-01
3.28557491e-01 -6.90704644e-01 -4.81241316e-01 -1.77333862e-01
-9.93091881e-01 -1.54858977e-01 7.55105078e-01 -1.11473262e-01
-8.01373601e-01 -3.10639977e-01 -6.09630883e-01 -2.06026331e-01
-2.89079934e-01 2.83440232e-01 -3.50554138e-02 -3.70112568e-01
8.81000102e-01 -2.48049334e-01 1.20973580e-01 -4.19279039e-01
-4.59994346e-01 9.18675184e-01 3.79799545e-01 -2.23959312e-01
6.55675650e-01 3.14522646e-02 -1.47978002e-02 -9.03050542e-01
-2.38487169e-01 -1.97019055e-01 -5.67620337e-01 -6.35756791e-01
6.28004193e-01 -7.66575634e-01 -5.72161257e-01 9.27037954e-01
-8.43052030e-01 -1.39323503e-01 -3.50573868e-01 8.22462201e-01
-1.20733194e-01 1.67527094e-01 -6.24877810e-01 -8.71864021e-01
-5.55879235e-01 -1.12729859e+00 5.56555688e-01 4.25287873e-01
-2.85268635e-01 -7.18221784e-01 -2.12453604e-01 8.74433592e-02
7.62254417e-01 2.80160367e-01 1.05802262e+00 -1.05858600e+00
6.13227524e-02 -6.26506448e-01 4.15111840e-01 6.77804351e-01
9.32440981e-02 2.28498638e-01 -1.04112768e+00 -6.28885627e-01
5.47678232e-01 -1.20066233e-01 5.30882359e-01 1.52930439e-01
9.09244001e-01 2.81379133e-01 -5.31770766e-01 1.74118116e-01
1.43165493e+00 9.07374024e-01 6.87670171e-01 3.48577261e-01
1.18957229e-01 3.05863112e-01 5.89560151e-01 4.43296701e-01
-2.88523644e-01 5.08899987e-01 5.60965002e-01 -1.15199849e-01
1.43253654e-01 4.92264181e-01 4.73984390e-01 7.85002530e-01
-8.69730935e-02 4.93301195e-04 -1.02137649e+00 2.02767387e-01
-1.32755291e+00 -9.59447503e-01 -7.29063869e-01 2.27317190e+00
3.47149760e-01 6.54922068e-01 -3.05021077e-01 1.01391280e+00
6.87729597e-01 -4.10073370e-01 -1.96693927e-01 -6.86160505e-01
6.60624355e-02 5.69490194e-01 4.32763278e-01 8.76591057e-02
-9.50182498e-01 6.26052078e-03 6.85186577e+00 5.85568666e-01
-1.28307378e+00 1.22675225e-01 2.75700420e-01 -1.25309810e-01
2.21260101e-01 3.89709547e-02 -5.82798421e-01 7.17060268e-01
1.33662832e+00 -1.22350231e-02 -8.73300731e-02 7.48733222e-01
2.75336932e-02 -6.69877589e-01 -1.06471014e+00 6.41842127e-01
-7.20223710e-02 -1.13104272e+00 -7.29218662e-01 -1.53519422e-01
4.49851692e-01 1.84091628e-01 -6.56984210e-01 5.42704761e-01
5.64282155e-03 -1.01215744e+00 7.07727849e-01 7.21686840e-01
4.72791791e-01 -9.54568267e-01 1.51976323e+00 4.12291318e-01
-6.30764723e-01 -6.71969533e-01 -2.01677695e-01 -5.02676129e-01
4.12583113e-01 1.11882520e+00 -1.04537261e+00 1.21062768e+00
1.07403374e+00 6.75534979e-02 -6.12603962e-01 1.16342700e+00
-1.69391885e-01 8.01940799e-01 -5.15776515e-01 -1.29940435e-01
-8.06977451e-02 1.64104804e-01 4.54008847e-01 6.78058326e-01
5.85045397e-01 -1.24381781e-01 -4.22195531e-02 1.39796376e-01
6.82770669e-01 -4.10075396e-01 -4.86466885e-01 4.75369424e-01
4.36421722e-01 1.37077379e+00 -6.61095798e-01 -2.34434560e-01
-1.79533586e-01 4.30995375e-01 -4.06967491e-01 -7.64742121e-02
-8.34605634e-01 -5.02101362e-01 3.30308795e-01 1.85718402e-01
1.13979906e-01 6.75742179e-02 -4.70445782e-01 -2.94152528e-01
-1.75300524e-01 -6.45008206e-01 5.34145594e-01 -8.64883363e-01
-1.26114225e+00 8.51581931e-01 1.23080403e-01 -1.12329543e+00
-3.56670558e-01 -7.80540586e-01 -1.05676591e+00 7.66836703e-01
-8.52866113e-01 -5.53674877e-01 -3.94324452e-01 4.38863128e-01
2.92739689e-01 -3.90750378e-01 8.37584674e-01 4.24497306e-01
-7.60403395e-01 1.20185070e-01 3.32524180e-01 -3.62253159e-01
5.61801672e-01 -1.15430462e+00 -1.22852400e-01 6.55971289e-01
-3.78045321e-01 -2.69371241e-01 1.02879894e+00 -7.06926405e-01
-1.36102355e+00 -7.22716749e-01 6.90107763e-01 -7.41531849e-02
4.99333382e-01 -9.42544714e-02 -9.34365571e-01 2.56104290e-01
4.16242152e-01 1.08867943e-01 2.94360697e-01 -9.60163027e-03
5.49900353e-01 -2.49399900e-01 -1.55406988e+00 -9.76353809e-02
2.88126230e-01 -2.91844219e-01 -6.39458001e-01 6.03923738e-01
3.31738144e-02 -4.84530240e-01 -1.19482052e+00 1.05310500e+00
3.40303153e-01 -1.18890715e+00 5.73665619e-01 -2.01420546e-01
7.89383892e-03 -3.41058195e-01 2.05040768e-01 -1.52510607e+00
-4.60781455e-01 -4.28810157e-02 -9.90178715e-03 1.13344646e+00
5.72648466e-01 -9.03903902e-01 4.83526468e-01 2.18141437e-01
-6.45334899e-01 -9.23223138e-01 -1.24643922e+00 -6.02203012e-01
3.72043392e-03 -4.14106220e-01 3.77094269e-01 7.37949669e-01
1.04888953e-01 1.31362885e-01 6.64970875e-02 5.32734334e-01
1.64902344e-01 -2.32880458e-01 1.85916260e-01 -1.44393122e+00
-1.23682231e-01 -2.25581348e-01 -7.47476697e-01 3.50088954e-01
-3.04194659e-01 -7.26613224e-01 1.24511696e-01 -1.55717969e+00
-4.25528556e-01 -7.78663695e-01 -4.93003100e-01 4.50500339e-01
7.85032958e-02 -1.55910300e-02 -2.12518260e-01 1.33736208e-01
5.21177426e-02 2.25584209e-01 6.13004744e-01 7.64037995e-03
2.16469511e-01 3.67862344e-01 -1.71276974e-03 6.57434106e-01
7.53411293e-01 -5.05261064e-01 -9.40501243e-02 -1.13000711e-02
8.28139335e-02 3.07594031e-01 4.04775620e-01 -1.79615486e+00
5.50030351e-01 5.35194397e-01 9.19248879e-01 -7.60735154e-01
-1.14352986e-01 -1.24463272e+00 7.25826919e-01 1.14160979e+00
6.07184231e-01 2.94876575e-01 3.08137655e-01 7.03447033e-03
-3.35613251e-01 -7.81698406e-01 9.08245504e-01 1.45314187e-01
-7.39933848e-01 -2.56483585e-01 -7.00554907e-01 -6.07413292e-01
1.61698413e+00 -1.37473270e-01 -3.51412475e-01 2.27529585e-01
-9.11276042e-01 3.01499724e-01 1.75146997e-01 3.32260907e-01
3.87035191e-01 -1.20991921e+00 -6.77229285e-01 7.14256525e-01
-2.44588599e-01 -2.46732160e-02 5.75962782e-01 1.14459908e+00
-7.81478941e-01 1.76617265e-01 -4.12769884e-01 -8.02127481e-01
-1.20787454e+00 1.66838944e-01 6.67345285e-01 -2.36298040e-01
-7.31638074e-01 8.66916239e-01 -8.12680244e-01 -1.22668393e-01
1.30172104e-01 -2.87140638e-01 -4.84507501e-01 3.27254176e-01
5.22817254e-01 7.15495586e-01 1.22463667e+00 -1.65407732e-01
-5.50520658e-01 2.49548495e-01 5.99894039e-02 3.97188216e-01
1.45509946e+00 6.26658916e-01 -1.18122019e-01 5.10581732e-01
6.80500984e-01 -2.90878981e-01 -7.18757570e-01 3.06800932e-01
1.88957214e-01 -4.85510305e-02 3.06378007e-01 -1.23301601e+00
-9.26994920e-01 5.59325099e-01 9.02243614e-01 5.31988561e-01
1.33305359e+00 -2.08188266e-01 3.36786687e-01 2.00498089e-01
6.76632285e-01 -1.21991789e+00 -3.82733017e-01 5.89870811e-01
8.97154868e-01 -5.99791467e-01 2.28104144e-01 6.98111504e-02
-4.63001728e-01 1.48861241e+00 6.37501836e-01 -1.70046166e-01
8.97037148e-01 9.28809524e-01 -1.99479595e-01 -5.21789193e-01
-1.05137789e+00 4.17458504e-01 -1.48382187e-01 5.81890404e-01
1.34385914e-01 -3.98857184e-02 -4.03266847e-01 8.38354290e-01
-1.29272178e-01 -8.39968100e-02 4.13051724e-01 1.14510512e+00
-7.25842953e-01 -9.92226124e-01 -7.49365568e-01 8.58499885e-01
-3.87572318e-01 3.10392350e-01 1.60124078e-01 9.37061608e-01
4.40409601e-01 1.00536382e+00 2.64901280e-01 -6.71926379e-01
8.34710836e-01 4.15184617e-01 3.93338740e-01 -3.43723476e-01
-1.10283649e+00 -4.62631166e-01 3.85219038e-01 -5.10223627e-01
4.45356928e-02 -6.09919608e-01 -1.43755329e+00 -2.01612815e-01
-8.50687981e-01 7.02868640e-01 1.03314400e+00 1.26507449e+00
-1.09773055e-01 1.16404140e+00 1.02765799e+00 -9.92602050e-01
-8.46988797e-01 -1.52007627e+00 -8.63325536e-01 -7.85725862e-02
-4.02361713e-02 -9.64372098e-01 -7.97256351e-01 -5.02439082e-01] | [6.660383224487305, 2.4353444576263428] |
859ab39b-0878-43cf-aded-d803ecf64b1a | classifier-stacking-for-native-language | null | null | https://aclanthology.org/W17-5044 | https://aclanthology.org/W17-5044.pdf | Classifier Stacking for Native Language Identification | This paper reports our contribution (team WLZ) to the NLI Shared Task 2017 (essay track). We first extract lexical and syntactic features from the essays, perform feature weighting and selection, and train linear support vector machine (SVM) classifiers each on an individual feature type. The output of base classifiers, as probabilities for each class, are then fed into a multilayer perceptron to predict the native language of the author. We also report the performance of each feature type, as well as the best features of a type. Our system achieves an accuracy of 86.55{\%}, which is among the best performing systems of this shared task. | ['Wen Li', 'Liang Zou'] | 2017-09-01 | null | null | null | ws-2017-9 | ['native-language-identification'] | ['natural-language-processing'] | [ 7.93702453e-02 8.86071026e-02 -8.03384900e-01 -5.65124452e-01
-9.36483979e-01 -6.93288624e-01 7.78433383e-01 3.97587150e-01
-7.70515561e-01 8.87357891e-01 4.26130027e-01 -5.04492104e-01
-1.73976030e-02 -5.76112628e-01 -4.49154764e-01 -2.20511302e-01
2.53801495e-01 4.93363798e-01 -2.89439917e-01 -3.29629853e-02
6.54653966e-01 2.91481197e-01 -1.13149440e+00 7.15265751e-01
7.54287064e-01 1.02715158e+00 -8.30375552e-02 9.80422735e-01
-3.24685216e-01 9.14719284e-01 -8.62965882e-01 -4.38136756e-01
-1.27124920e-01 -7.64434636e-02 -1.06172812e+00 -4.41711217e-01
7.22638845e-01 -2.50939250e-01 -5.30642748e-01 7.11936474e-01
7.50926882e-02 2.05055431e-01 1.20901787e+00 -8.75513017e-01
-9.61898148e-01 1.04862344e+00 -2.36561105e-01 4.54025537e-01
5.85774004e-01 -4.28147726e-02 1.60866046e+00 -1.23272455e+00
7.75632262e-01 8.04400086e-01 5.49516141e-01 3.21910471e-01
-8.95859420e-01 -8.77620280e-01 5.46982884e-02 2.53496945e-01
-9.56062317e-01 -5.99344313e-01 6.77151799e-01 -7.78255284e-01
1.30712712e+00 -3.06971893e-02 4.71375883e-01 1.14312518e+00
3.28783542e-01 1.08872485e+00 1.29654682e+00 -6.61900163e-01
1.56167727e-02 3.85353476e-01 1.05414200e+00 9.10866737e-01
-3.09235267e-02 -4.55253750e-01 -9.17489767e-01 -2.85537750e-01
6.02307962e-03 -2.14660168e-01 -1.50545970e-01 5.04062593e-01
-1.32451093e+00 1.02886724e+00 1.44626617e-01 3.27945173e-01
-3.01915139e-01 -1.36906177e-01 1.23840049e-01 5.54558694e-01
6.93369150e-01 8.81172419e-01 -9.97654676e-01 -3.43821764e-01
-1.06694078e+00 4.13328648e-01 1.38855934e+00 6.64391518e-01
6.76676512e-01 -2.86188126e-01 -5.19695103e-01 9.61121261e-01
2.74186313e-01 4.05458957e-01 7.48125672e-01 -6.90033734e-01
5.87091327e-01 7.48758256e-01 1.16389893e-01 -8.18057299e-01
-6.91224456e-01 -4.79320228e-01 -3.89671654e-01 -1.36808828e-01
6.00341260e-01 -4.48566377e-01 -4.83257681e-01 1.40065050e+00
-3.23596835e-01 -2.92001635e-01 9.68155041e-02 4.08676714e-01
1.24325001e+00 7.79326618e-01 2.06369683e-01 -1.83648262e-02
1.22481120e+00 -9.45671499e-01 -6.27746522e-01 -3.09774399e-01
9.07378793e-01 -9.55051422e-01 7.37459540e-01 3.18046600e-01
-1.02879643e+00 -3.72068763e-01 -1.07627571e+00 4.01778333e-02
-5.87754428e-01 4.95960295e-01 5.87303340e-01 4.97798979e-01
-8.19629610e-01 6.57767117e-01 -2.96262532e-01 9.07292683e-03
4.06690061e-01 3.68458927e-01 -2.86851078e-01 2.82011986e-01
-1.33370495e+00 1.21668577e+00 2.68880427e-01 -4.36366528e-01
-1.81243703e-01 -1.01615608e+00 -7.58534074e-01 -5.33926953e-03
-1.07809015e-01 -2.15061665e-01 1.25113165e+00 -9.31561470e-01
-1.68367600e+00 1.17561412e+00 -3.28037113e-01 -4.15236354e-01
2.73872130e-02 -1.73870489e-01 -3.38007867e-01 -2.68191636e-01
2.44273737e-01 2.80435920e-01 5.42918980e-01 -5.56267202e-01
-9.21546936e-01 -1.67171523e-01 -6.90327659e-02 3.26663181e-02
-7.55026400e-01 4.61395830e-01 1.84496328e-01 -3.69704813e-01
-1.10209793e-01 -7.72225440e-01 2.03155488e-01 -4.58313406e-01
-6.08839214e-01 -9.82788563e-01 2.21181169e-01 -1.05963302e+00
1.28567839e+00 -2.08871460e+00 2.98948407e-01 4.62544151e-02
2.60848373e-01 -8.27924088e-02 -2.65838392e-02 2.13496029e-01
1.36695072e-01 1.06977016e-01 3.11886579e-01 -2.94254541e-01
1.26545131e-01 -2.94077098e-01 -4.02141809e-01 4.08373475e-01
4.28997576e-02 1.00048852e+00 -6.19201303e-01 -3.18793088e-01
-2.39881039e-01 -6.18547238e-02 -3.74041915e-01 2.53658146e-01
-2.88850516e-01 -4.74343263e-02 -1.61823586e-01 4.47439641e-01
2.41091084e-02 -3.49589974e-01 -1.74630284e-02 -8.83876756e-02
-2.93961197e-01 9.23610866e-01 -7.46643424e-01 1.08942139e+00
-6.46380246e-01 8.58816564e-01 -1.14000946e-01 -6.22568607e-01
1.19796598e+00 1.13490276e-01 8.70693550e-02 -2.13427022e-01
-2.55441964e-02 4.62919623e-01 1.32621661e-01 -3.83446068e-01
4.68222737e-01 2.72208005e-01 -5.84289551e-01 9.60602283e-01
4.59477782e-01 2.09329635e-01 1.51578590e-01 2.51701206e-01
1.10506153e+00 -4.41514030e-02 6.25755608e-01 -4.20024931e-01
4.71524358e-01 8.90818313e-02 2.76663035e-01 1.04746902e+00
-1.80638626e-01 2.20960617e-01 5.90140820e-01 -7.20246732e-01
-1.04328239e+00 -6.83857977e-01 -2.44257972e-01 1.67878199e+00
-5.37420511e-01 -3.83737177e-01 -3.63711387e-01 -9.17909324e-01
2.81502128e-01 9.97968674e-01 -6.93642914e-01 -8.21300671e-02
-4.87145662e-01 -5.77338040e-01 7.37062991e-01 5.02675533e-01
1.32429793e-01 -1.25820184e+00 -3.69694829e-01 2.07030684e-01
-1.49396032e-01 -8.13565433e-01 -3.08300197e-01 6.31823301e-01
-3.22403014e-01 -9.70443070e-01 -5.32394886e-01 -1.07437122e+00
2.24620044e-01 -1.97201893e-01 9.53115463e-01 2.45352108e-02
-7.30688497e-02 7.92824849e-02 -3.01062405e-01 -5.33630967e-01
-3.32755983e-01 8.30905855e-01 8.33026990e-02 -1.71473324e-01
6.54231846e-01 -1.71575155e-02 -4.38180938e-02 -3.86736512e-01
-5.74946105e-02 6.71346113e-02 4.45027739e-01 9.99640763e-01
3.69944982e-02 -4.99154568e-01 6.49749696e-01 -9.27735388e-01
8.62112343e-01 -6.33510232e-01 -1.35812685e-01 4.02154833e-01
-6.06112957e-01 -1.11987755e-01 7.36536741e-01 -3.69282484e-01
-9.25734580e-01 -1.48525953e-01 7.25459978e-02 3.27033877e-01
-3.03513080e-01 9.54417229e-01 1.10721260e-01 1.28223270e-01
6.40725076e-01 9.33304876e-02 -1.90910503e-01 -5.70692599e-01
2.10542932e-01 1.20145941e+00 3.73713076e-01 -4.68077511e-01
5.38914204e-01 -3.69770467e-01 -2.91631937e-01 -9.44612563e-01
-1.58336878e+00 -3.50948215e-01 -1.03627968e+00 -6.55956566e-02
6.42442763e-01 -7.17036545e-01 -5.68584740e-01 6.17796183e-01
-1.33197606e+00 -3.21038216e-01 9.05027315e-02 5.85215271e-01
-2.37343490e-01 -1.57041535e-01 -8.64128709e-01 -6.38787746e-01
-5.70573092e-01 -9.27838862e-01 6.15688980e-01 5.23955405e-01
-8.74216318e-01 -1.18941545e+00 1.47651300e-01 5.90106547e-01
5.07768154e-01 -2.79955715e-01 1.36756623e+00 -1.54583001e+00
8.68143588e-02 -5.06936371e-01 -2.69889534e-01 5.49645051e-02
-2.89024234e-01 2.36944959e-01 -9.67120290e-01 1.84844725e-03
-3.53942245e-01 -6.29055977e-01 9.71166372e-01 3.80867153e-01
9.95490909e-01 -5.43527246e-01 -3.16109031e-01 5.04373670e-01
1.03112650e+00 -1.80819944e-01 -1.88985601e-01 6.97285354e-01
5.40480018e-01 4.80168641e-01 1.65574461e-01 5.85912347e-01
5.56103945e-01 4.98309016e-01 -2.12216571e-01 3.36690575e-01
4.02854607e-02 -7.35914856e-02 5.64763427e-01 7.60204256e-01
2.55811065e-01 -6.32107332e-02 -1.09394932e+00 4.03285295e-01
-1.53830457e+00 -7.13713944e-01 -2.43696585e-01 1.64983678e+00
1.22626066e+00 2.62892246e-01 -4.44240198e-02 8.16524625e-02
4.33041573e-01 2.67640680e-01 -1.12291791e-01 -9.06754196e-01
-2.70280987e-01 5.35264969e-01 5.95147669e-01 6.92934155e-01
-1.49637306e+00 1.11869931e+00 7.31358624e+00 5.84574342e-01
-1.00346220e+00 5.26255369e-02 4.65224087e-01 -2.21372858e-01
-2.02794850e-01 -2.06807405e-01 -1.25176775e+00 3.05780113e-01
1.32032418e+00 -5.98020434e-01 4.96975631e-01 8.15007210e-01
-3.95571947e-01 -4.69943322e-02 -1.08107841e+00 5.52527189e-01
4.09301460e-01 -1.33233953e+00 -5.62959500e-02 2.51037572e-02
7.05229521e-01 1.94713041e-01 1.57426983e-01 6.25348687e-01
4.45002615e-01 -1.22436476e+00 6.06206357e-01 7.47285783e-01
4.26193088e-01 -8.31202030e-01 9.08795357e-01 6.23781800e-01
-5.01133263e-01 -1.59369096e-01 -2.91985154e-01 -5.82357824e-01
-4.15582538e-01 6.92366362e-01 -8.68161678e-01 4.86057773e-02
3.53528649e-01 9.14918363e-01 -7.63572991e-01 7.07866728e-01
-5.28997183e-01 1.12691522e+00 -5.87501973e-02 -6.68181300e-01
2.25911066e-01 2.11265191e-01 5.14250815e-01 1.61418152e+00
-7.51061067e-02 2.06379928e-02 4.36971635e-01 5.63587308e-01
-5.66522598e-01 4.67413098e-01 -5.02594233e-01 -1.94483265e-01
6.86767399e-01 1.43402624e+00 -2.60189176e-01 -7.03963637e-01
-5.82306266e-01 8.52897048e-01 8.44838977e-01 9.34288353e-02
-2.89047539e-01 -8.18196416e-01 4.68457013e-01 -4.49298084e-01
5.38961291e-02 -2.09745511e-01 -9.69852865e-01 -1.20059729e+00
-2.51388878e-01 -6.88373506e-01 5.82950413e-01 -5.29570222e-01
-1.45435715e+00 4.49601799e-01 -4.06966984e-01 -6.11792326e-01
-3.13302457e-01 -1.01817775e+00 -6.35037780e-01 1.22850132e+00
-1.29243386e+00 -1.05154705e+00 1.25562862e-01 2.15481967e-01
5.21966815e-01 -8.75713885e-01 1.25698400e+00 -7.54811242e-02
-7.38895416e-01 9.41609859e-01 3.94253522e-01 5.76204717e-01
1.06296527e+00 -1.32146740e+00 2.95782059e-01 3.19434166e-01
2.58133203e-01 7.53842771e-01 4.49082553e-01 -6.44081473e-01
-1.47370923e+00 -9.98260260e-01 1.87330508e+00 -5.97821832e-01
1.12671793e+00 -1.48566410e-01 -8.31291437e-01 9.27038908e-01
3.08591485e-01 -3.97536308e-01 1.11781096e+00 6.93677843e-01
-5.46242535e-01 3.86021048e-01 -9.59869385e-01 3.55401725e-01
5.49717426e-01 -6.33139610e-01 -9.68553364e-01 6.26905560e-01
5.39553881e-01 -3.68046880e-01 -1.14966166e+00 -7.15125650e-02
9.47378337e-01 -5.26482940e-01 5.58751643e-01 -1.04350448e+00
1.02430499e+00 5.61459005e-01 -1.71472579e-01 -1.52765441e+00
-6.37647510e-01 -1.39059037e-01 -3.22337359e-01 1.21640515e+00
1.04340112e+00 -5.81046462e-01 6.08680606e-01 6.73429608e-01
-4.71246280e-02 -1.03916037e+00 -7.59437442e-01 -5.29457450e-01
4.56258655e-01 -3.35086554e-01 3.93731296e-01 1.19288766e+00
5.31664491e-01 5.17174542e-01 -1.64159521e-01 -3.18425983e-01
5.40717185e-01 4.07908320e-01 4.53428268e-01 -1.39885330e+00
-2.81108856e-01 -9.61294055e-01 -1.04166262e-01 -6.74518287e-01
8.59170914e-01 -1.58490169e+00 -1.09372936e-01 -1.42533171e+00
4.54800367e-01 -2.71216333e-01 -4.17048037e-01 8.10071468e-01
-3.65319252e-01 1.86631933e-01 1.57384083e-01 4.69661564e-01
-5.52427173e-01 3.40345013e-03 6.66288257e-01 -3.83597761e-01
-2.37926289e-01 1.42603919e-01 -9.21766281e-01 8.11262190e-01
1.00729191e+00 -3.88213187e-01 5.16769886e-01 -4.32989210e-01
2.21611023e-01 -2.53308237e-01 3.32816988e-02 -8.40822339e-01
3.09000611e-01 -3.09260875e-01 9.05677438e-01 -5.78430533e-01
1.08456425e-01 2.80428510e-02 -6.65326059e-01 5.84773600e-01
-1.03345358e+00 -1.44492626e-01 -6.14074171e-02 -1.54974744e-01
7.93357641e-02 -6.12118483e-01 5.75226367e-01 2.25542560e-02
-3.81510675e-01 5.13238721e-02 -6.98096991e-01 7.67576694e-02
6.40689909e-01 2.26918578e-01 -2.60229528e-01 -2.07434356e-01
-5.55281937e-01 1.20203428e-01 2.42218986e-01 5.58111489e-01
4.13278997e-01 -9.81795609e-01 -1.07297051e+00 2.68214613e-01
2.21073851e-02 -7.87773609e-01 -2.51224756e-01 5.29929578e-01
-7.50168934e-02 7.60730386e-01 -1.66293234e-01 -2.64362581e-02
-1.15067673e+00 5.97301312e-02 7.44328275e-02 -4.70386058e-01
-4.81916964e-01 1.01547456e+00 -4.77682084e-01 -5.52733481e-01
2.38814339e-01 2.00171515e-01 -7.19468057e-01 3.45439732e-01
8.04947555e-01 3.67458493e-01 6.87999725e-02 -6.93413138e-01
-4.34413522e-01 2.31967837e-01 -3.84861618e-01 -3.99148107e-01
1.44501305e+00 3.65269542e-01 -2.51457930e-01 8.96120071e-01
1.32196271e+00 7.38803223e-02 -4.66252416e-01 -5.79844773e-01
2.63921767e-01 -2.16147248e-02 1.50140166e-01 -1.06324041e+00
-3.88545990e-01 5.86986303e-01 -1.10505186e-01 5.94109036e-02
9.88322943e-02 -9.23012197e-03 7.79479086e-01 6.62018299e-01
3.54261547e-02 -1.16024685e+00 -3.00960362e-01 1.27477205e+00
7.35803545e-01 -1.07414448e+00 2.25299001e-01 2.36702897e-02
-5.52370548e-01 1.53530073e+00 4.20990169e-01 -5.11291325e-01
8.90139997e-01 3.94370228e-01 1.08498082e-01 -5.76983355e-02
-1.12557817e+00 2.65745223e-01 6.35591745e-01 2.86617607e-01
9.60959494e-01 3.27024221e-01 -5.72372794e-01 1.17416131e+00
-5.86943150e-01 -1.35589078e-01 5.16031146e-01 7.56183326e-01
-7.61954129e-01 -7.92459309e-01 -2.42517561e-01 1.23543715e+00
-6.52661800e-01 -2.71306068e-01 -7.52095699e-01 3.79906058e-01
-9.70979705e-02 9.42550004e-01 9.50705409e-02 -5.54760695e-01
5.98951951e-02 8.10475111e-01 4.89629418e-01 -8.75102162e-01
-1.06278288e+00 -5.42573392e-01 4.93793428e-01 -1.56004697e-01
2.46992022e-01 -8.50368798e-01 -1.05551231e+00 -1.59727082e-01
-1.57850146e-01 1.12031400e-01 9.27992463e-01 1.39002693e+00
-6.79126903e-02 3.34632397e-01 6.44910991e-01 -7.43815899e-01
-9.74180281e-01 -1.36289644e+00 -4.53330636e-01 2.65206426e-01
1.93790153e-01 -4.82761145e-01 -5.93901753e-01 -6.07594587e-02] | [10.502076148986816, 10.336982727050781] |
8f362d93-6218-4ade-924d-48816831df5a | mimic-extract-a-data-extraction-preprocessing | 1907.08322 | null | https://arxiv.org/abs/1907.08322v2 | https://arxiv.org/pdf/1907.08322v2.pdf | MIMIC-Extract: A Data Extraction, Preprocessing, and Representation Pipeline for MIMIC-III | Robust machine learning relies on access to data that can be used with standardized frameworks in important tasks and the ability to develop models whose performance can be reasonably reproduced. In machine learning for healthcare, the community faces reproducibility challenges due to a lack of publicly accessible data and a lack of standardized data processing frameworks. We present MIMIC-Extract, an open-source pipeline for transforming raw electronic health record (EHR) data for critical care patients contained in the publicly-available MIMIC-III database into dataframes that are directly usable in common machine learning pipelines. MIMIC-Extract addresses three primary challenges in making complex health records data accessible to the broader machine learning community. First, it provides standardized data processing functions, including unit conversion, outlier detection, and aggregating semantically equivalent features, thus accounting for duplication and reducing missingness. Second, it preserves the time series nature of clinical data and can be easily integrated into clinically actionable prediction tasks in machine learning for health. Finally, it is highly extensible so that other researchers with related questions can easily use the same pipeline. We demonstrate the utility of this pipeline by showcasing several benchmark tasks and baseline results. | ['Shirly Wang', 'Michael C. Hughes', 'Geeticka Chauhan', 'Tristan Naumann', 'Matthew B. A. McDermott', 'Marzyeh Ghassemi'] | 2019-07-19 | null | null | null | null | ['length-of-stay-prediction'] | ['medical'] | [ 0.29481182 -0.02438877 -0.09211102 -0.47305465 -1.3084937 -0.55477935
0.03642758 1.0110091 -0.35221294 0.4835894 0.59197766 -0.56166375
-0.347326 -0.4546411 -0.47688097 -0.47519153 -0.17897427 0.5632105
-0.26462656 0.19950482 -0.24068178 0.40327457 -1.0798764 0.7866175
0.577542 0.9124393 -0.1661455 0.5711742 0.35629895 0.7938116
-0.2527485 -0.12807041 0.24932374 -0.28023988 -0.7728959 -0.44683048
0.19106466 -0.19939539 0.00892641 0.39656344 0.78936714 -0.4112227
0.39157718 -1.05395 -0.4133694 0.349127 0.08567519 0.15097067
0.34393322 0.48542267 0.77550095 -0.41326222 0.9674775 0.9058795
1.3476475 0.4174534 -1.5727842 -0.39323172 -0.38828835 0.01750664
-0.99521905 -0.51943374 0.02850838 -0.8719137 1.2568693 0.6094599
0.63632554 1.1495072 0.6446837 0.30703774 0.8591745 -0.03502163
0.11164061 -0.21375594 0.20610309 0.57074267 0.21944226 -0.03240443
-0.21626493 -0.6641858 0.3140773 0.6666529 -0.15679815 0.02201535
-1.7336776 0.52646947 0.14004749 0.13371861 -0.68642896 -0.1035962
0.8611621 0.34479138 0.34401157 0.708853 -1.057847 -0.26583737
-0.7731083 0.25895232 0.7329847 0.7918522 0.06783796 -0.4743573
-0.32471812 0.54780257 -0.02218629 0.19808942 0.79877335 -0.9476052
0.22479145 0.76086295 -0.06389426 -0.6634122 -0.9845529 -0.36882764
-0.84683293 -0.14546664 0.3884234 -0.18097872 -0.7054543 1.642115
0.4185896 0.1168641 0.2144017 0.47508407 1.0242583 0.09558855
0.3867561 -0.18468675 1.6326213 -0.45374066 -0.7371327 0.27241892
1.145042 -0.7490401 1.1082807 0.33874035 -0.9224985 -0.10907609
-0.7328117 -0.44412178 -0.49191013 -0.22136147 0.5684005 0.18150376
-0.94899374 0.94948214 -1.4715577 -0.59734046 0.6399734 0.29222843
-0.73020697 -0.13872974 -0.85471517 0.8686808 0.25570032 -0.26817167
-0.6260149 -1.5514846 -0.9523836 -0.11802537 -0.05566197 -1.1949489
1.3854185 -0.55176467 -0.7669829 0.8506071 -0.06224397 -0.4531429
0.63774306 -0.13251452 -0.450862 0.03431623 0.23086254 0.28476608
0.2183224 -0.3711497 -0.72096133 -0.60243887 -0.8399056 -0.29379413
-0.14047404 0.21071106 -0.20833395 -0.84030294 -0.05364781 -0.90036935
-0.50023025 0.10445108 -0.45362934 0.06821144 0.6153607 -1.0736525
1.3568217 -2.0206084 -0.2098183 0.01447097 0.50212455 0.11377124
-0.13416813 0.67349225 -0.42110893 0.17031425 -0.24537218 -0.28659728
-0.3277282 0.05669519 -0.16011262 0.5623343 0.4199025 1.2740302
-0.9180112 -0.32305253 0.10175309 0.5301845 -0.7966285 0.32155722
0.06484777 0.6992516 -0.30974144 0.7557538 0.20564078 -0.5591317
0.24093889 -0.24746165 -0.0597768 0.5600787 -0.7551872 1.738558
-0.42553034 0.22390017 -0.08683835 -0.6334364 0.5356299 0.5414403
1.2042643 -0.40407652 -0.040924 0.20345017 0.01213184 -0.85249704
-0.24217029 -0.1153302 -0.05584523 0.5167263 -0.11116656 0.09311576
0.01112617 -0.17407186 1.8629692 -0.03222608 0.6877788 -0.31775647
-0.05115111 0.29561284 0.92883486 0.50957996 -0.26874214 0.8174857
0.393669 -0.7547183 -1.1723936 -1.2725583 -0.7202129 0.804867
-0.926138 -0.7557023 -0.24464911 -0.3890602 0.4432578 0.31177798
-0.78425294 -0.20720795 -0.3292247 -0.90551853 0.7803814 0.6195684
-0.14378573 -1.0017444 -0.7960412 0.6259503 -0.31559482 -0.9342407
-0.47588962 0.15202644 -1.0536611 -1.4930223 -0.3270492 -0.6488362
0.3922977 -0.18749423 1.2638576 0.12532353 -1.0243353 0.35822475
-0.1707017 -1.0060526 -0.5851398 0.1337613 -0.02161859 -0.32127076
0.55507433 -0.32008007 -0.84371245 0.02541552 -1.0211191 0.05483767
0.42978987 0.7542584 0.9962853 -0.44461244 1.0563751 -1.1686112
0.58652014 -0.903052 -0.25534046 0.13781242 -1.0604177 -0.06726791
0.78150904 0.0299437 -0.5884174 0.23807228 -0.29667923 -0.11912624
-0.27569786 0.77634823 0.10370344 0.5519896 0.72352177 -0.1424885
0.5778034 -0.7599896 0.14454332 0.77670664 0.7089938 -0.21228965
0.3019489 0.565208 0.10144489 -0.47090247 -0.71792257 -0.6321564
-0.6448283 0.5171687 1.029256 -1.0751194 -0.6652416 0.40899903
-0.65968126 -0.27638888 -0.5370476 0.67639434 -0.55683446 0.1316541
-0.9327018 -0.16566987 -0.84744895 -1.1061256 1.0564713 -0.35029477
-0.8766599 -1.1593736 0.4511097 0.17990509 0.33138442 0.86085784
1.2742957 -1.2402533 -0.09819349 -0.40274325 0.0135562 0.19835106
0.6845989 0.20205867 -0.88031816 -0.37531313 -0.0291251 -0.17402649
0.7268772 0.500762 1.4376446 -0.6186465 -0.50312644 0.9711432
1.1816337 0.04936663 0.4546709 0.3295374 0.5277011 0.5360899
0.35516003 0.5593624 0.653805 0.3671025 -0.07963865 -0.5360508
0.15955919 -0.02669121 -0.11859176 0.8805582 0.19144356 0.36803988
-1.1982183 0.5401105 -2.0700853 -0.84802485 -0.26759696 2.255778
1.2112569 -0.2551147 0.16348928 -0.04180747 0.19151306 -0.5448458
-0.72755146 -0.36611775 0.06661102 0.44667345 0.4445072 0.00971378
-1.1796854 0.29760683 7.1920805 0.1000944 -1.0918254 0.17858443
0.80446047 -0.41870058 -0.09665745 -0.31594822 -0.29100817 0.47189412
1.4531794 -0.48903042 0.10509199 0.73725057 0.8770769 0.3962378
-1.6526226 0.8785813 -0.4380802 -1.6562147 -0.24678534 -0.04360893
0.44597957 0.38417935 0.02983652 -0.01585221 0.28489402 -1.294679
0.23782332 0.77401245 1.0589057 -0.28589195 0.8480486 -0.04814736
-0.81792766 -0.11180039 0.06864588 0.07536346 -0.03639737 0.8115456
-1.3328204 0.65757006 0.82839036 1.060969 -0.59428644 1.1409899
0.56569797 0.66354847 -0.21288161 0.63513756 -0.2518185 0.1303726
0.1241938 1.4877214 0.37689266 0.05154605 0.30305403 0.5717781
0.03362279 0.4047063 -0.6861938 -0.18679468 0.46901232 1.3806888
-0.3207422 -0.26456752 -0.7313217 0.45623356 0.14678752 0.01430985
-0.6065862 -0.28106394 0.95253897 0.37409732 -0.27325538 0.1537099
-0.52003926 -1.1176037 -0.05615531 -1.3323174 1.0967617 -0.41162634
-1.7036389 0.5071253 -0.24952072 -1.2582186 -0.5160043 -0.611594
-0.38836962 0.9554485 -1.2892945 -1.0914034 -0.34159654 0.6956627
0.11408945 -0.14375722 1.3010414 0.46046215 -0.76303166 0.60564506
0.33136854 0.11642705 1.1716167 -1.10623 0.5242062 0.44388217
-0.3871684 0.9560102 0.41562268 -0.75526124 -1.400446 -1.7149721
0.9512906 -0.94183034 0.71392053 -0.02750557 -1.1569246 1.0435144
-0.26339298 0.12816305 1.509565 0.09814043 -0.48366818 -0.23629858
-1.2610798 0.27288488 0.70937246 -0.5073015 -0.682385 0.5154091
0.5588834 -0.3048032 -1.6270667 0.57849574 0.68402565 -0.5312044
0.8383436 -1.4090122 0.47464246 -0.07458027 0.09729291 -1.2513218
-0.32904875 -0.71579206 0.04641849 0.9060274 0.61643225 -0.88450754
0.18096459 1.0614004 -0.22411764 -0.9391303 -0.8790883 -0.4403633
0.2280783 -0.4426814 0.75602084 1.3128918 0.34189433 -0.06499692
0.01799956 -0.02172267 0.3189861 0.07431472 0.66147363 -1.4847649
-0.42056444 -0.23306517 -0.36622205 0.01478144 -0.38381028 -1.3367256
-0.27129984 -1.6751441 0.48866835 -0.6334427 -0.700805 1.0966918
-0.29929936 0.1346565 -0.07039239 0.5311023 -0.3774846 -0.01126185
0.8555956 0.19018358 -0.4433762 -0.05726061 -0.8752541 0.6092645
0.7161644 -0.6646337 -0.15410289 -0.3154108 -0.07433039 -0.05992433
0.4280856 -0.8759052 -0.11498442 -0.10021535 0.5893348 -0.34337834
-0.27942434 -0.8407959 0.6620709 0.8421501 -0.5094515 0.57836574
0.45073095 0.3945338 0.0290546 0.31285194 0.73014444 -0.17342176
-0.14868917 0.42893168 -0.33046588 0.21224613 1.076555 0.28413504
-0.47472802 -0.01990939 -0.8050106 0.31107914 0.47993833 0.6207148
0.23699555 -1.2290263 -1.0409034 0.3123203 0.5029608 -0.06156515
0.26299122 1.1819247 -0.61206746 0.58734304 -0.13382247 -0.6417798
-1.2730035 0.81176007 0.25942853 -0.35709414 -1.0238469 0.0257824
-0.02295619 -0.59128934 0.05973409 -0.84362334 0.25576308 -0.04807442
0.7447436 0.1878432 0.6206907 -0.13553338 -0.6360848 -0.07061782
-0.07643531 0.3680666 1.7317173 0.16856241 -0.28456122 0.5086513
1.3899919 -0.18836613 -1.0029291 -0.03183911 0.19232643 0.00751482
-0.23048803 -0.964112 -0.72668225 0.48294446 0.6854035 0.01776546
1.2127603 -0.14226317 0.8572466 0.38075587 -0.03179841 -0.86659056
-0.39349893 0.21280196 0.7065343 -1.1851335 -0.02676211 -0.04592801
-0.55161333 0.98006105 0.16207696 0.35118303 0.9414194 0.58962226
0.60327995 -0.20773299 -1.2858794 0.19401544 0.32400426 0.6909364
0.86111116 0.03580124 -0.47011343 1.1704314 -0.19160895 0.44933996
0.39293808 1.0281138 0.13025491 -1.174037 -0.2286308 1.0104072
-0.97117895 -0.3046798 -0.07792605 0.604129 0.04796493 0.9408389
0.07949954 -0.11176104 0.5892976 0.36914176 -0.04611774 -0.8612549
-0.79669046 0.04353591 0.0792988 -0.96066684 0.02409699 -0.87148744
-1.3764756 -0.07341871 0.42576396 -0.02607399 0.59883946 0.6946443
1.2191474 0.7772428 0.27021223 -0.22993472 -0.722974 -0.7371001
-0.26039952 0.98219204 0.5801491 -0.13799447 -0.04431919 0.46549648] | [7.904660701751709, 6.271363735198975] |
413fcfab-3516-4b0f-9e9d-3c9d4f08fd44 | abrupt-motion-tracking-via-nearest-neighbor | 1410.7484 | null | http://arxiv.org/abs/1410.7484v2 | http://arxiv.org/pdf/1410.7484v2.pdf | Abrupt Motion Tracking via Nearest Neighbor Field Driven Stochastic Sampling | Stochastic sampling based trackers have shown good performance for abrupt
motion tracking so that they have gained popularity in recent years. However,
conventional methods tend to use a two-stage sampling paradigm, in which the
search space needs to be uniformly explored with an inefficient preliminary
sampling phase. In this paper, we propose a novel sampling-based method in the
Bayesian filtering framework to address the problem. Within the framework,
nearest neighbor field estimation is utilized to compute the importance
proposal probabilities, which guide the Markov chain search towards promising
regions and thus enhance the sampling efficiency; given the motion priors, a
smoothing stochastic sampling Monte Carlo algorithm is proposed to approximate
the posterior distribution through a smoothing weight-updating scheme.
Moreover, to track the abrupt and the smooth motions simultaneously, we develop
an abrupt-motion detection scheme which can discover the presence of abrupt
motions during online tracking. Extensive experiments on challenging image
sequences demonstrate the effectiveness and the robustness of our algorithm in
handling the abrupt motions. | ['Jian Zhang', 'Yao Lu', 'Tianfei Zhou', 'Huijun Di', 'Feng Lv', 'Qingjie Zhao'] | 2014-10-28 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 6.30691126e-02 -6.26061440e-01 -4.55685109e-01 -2.03008369e-01
-6.69765592e-01 -2.77579516e-01 5.55384517e-01 -1.02050871e-01
-3.58655840e-01 6.51353955e-01 1.12148002e-01 1.84465405e-02
-1.02413237e-01 -5.50026536e-01 -3.90068263e-01 -8.93090069e-01
1.11564010e-01 4.00657393e-02 8.12492669e-01 3.14122409e-01
3.59196693e-01 3.55616510e-01 -1.23037624e+00 -4.19225484e-01
9.96446073e-01 8.88298452e-01 6.07519865e-01 5.43357015e-01
1.70707721e-02 6.00323439e-01 -4.80276495e-01 1.93176586e-02
1.24406561e-01 -4.32156295e-01 -1.71849668e-01 3.12547684e-01
3.00463051e-01 -5.49096644e-01 -2.43429095e-01 1.28041852e+00
4.80794191e-01 4.70739007e-01 3.57906163e-01 -8.51984143e-01
-1.08020590e-03 1.98139936e-01 -9.50422823e-01 3.94146293e-01
3.76758367e-01 4.15683597e-01 6.24328196e-01 -9.17954981e-01
4.51899827e-01 1.32208049e+00 8.34849715e-01 3.90645564e-01
-9.54785883e-01 -7.27289021e-01 4.66008306e-01 1.81937933e-01
-1.41704166e+00 -4.58298683e-01 8.42359304e-01 -4.36580867e-01
1.36409417e-01 -5.18953288e-03 8.84468913e-01 7.16994286e-01
2.00671270e-01 9.59204733e-01 7.91839838e-01 -2.15963900e-01
3.88147056e-01 -1.08738445e-01 -9.06346515e-02 6.21083975e-01
3.87472898e-01 3.90853405e-01 -5.36423981e-01 -4.39954996e-01
1.04891455e+00 2.82899678e-01 -3.15766692e-01 -4.68167484e-01
-1.22618067e+00 6.88358009e-01 2.53864795e-01 1.40578449e-01
-5.46404064e-01 2.41655275e-01 3.27035606e-01 -2.92805284e-01
4.65885371e-01 -1.63739249e-01 1.09765418e-01 -1.94734260e-01
-1.37912333e+00 5.05915165e-01 4.57857907e-01 8.66734266e-01
6.36426806e-01 -2.30845083e-02 -5.23512125e-01 4.60970908e-01
7.41334915e-01 6.23047650e-01 2.33454883e-01 -1.00099719e+00
2.86890209e-01 1.66776940e-01 7.32810318e-01 -1.10731161e+00
7.51160458e-02 -4.52048242e-01 -7.68926919e-01 2.22766697e-01
5.31561732e-01 -1.21903807e-01 -7.01442719e-01 1.43497860e+00
1.02098751e+00 6.13931775e-01 -3.61568123e-01 9.94575143e-01
6.37509972e-02 6.48749650e-01 1.80235252e-01 -5.66473067e-01
1.20005643e+00 -7.01676905e-01 -1.08783638e+00 6.53972402e-02
1.36665031e-01 -8.44101191e-01 6.25916541e-01 3.37244004e-01
-1.07976246e+00 -8.24356496e-01 -8.38236392e-01 4.74357724e-01
3.22800040e-01 2.83665806e-01 4.39916730e-01 5.16519487e-01
-6.42315149e-01 6.00296617e-01 -1.19195402e+00 -2.33186096e-01
4.77189779e-01 2.84428671e-02 4.24747765e-01 -7.67958388e-02
-9.22311664e-01 4.88563567e-01 5.12175679e-01 3.64158720e-01
-1.07507026e+00 -4.50617820e-01 -5.77146590e-01 -1.26533151e-01
5.65489829e-01 -6.19642258e-01 1.17235661e+00 -5.43904901e-01
-1.63776636e+00 4.17057648e-02 -6.54485881e-01 -4.19803709e-01
8.29864562e-01 -4.21175569e-01 -2.16102600e-01 1.39897928e-01
1.84895515e-01 4.12056118e-01 1.15181935e+00 -1.02066779e+00
-9.39930618e-01 -1.38675079e-01 -2.74634689e-01 3.45776737e-01
-2.36520305e-01 -2.32370989e-03 -6.75919592e-01 -9.01187778e-01
2.76384413e-01 -8.10814798e-01 -6.69595540e-01 3.69132072e-01
-1.11631051e-01 -3.64107579e-01 1.03929603e+00 -4.41100806e-01
1.43931854e+00 -2.24283934e+00 -1.67456523e-01 3.09570909e-01
1.75586760e-01 3.14485282e-01 4.02196765e-01 1.23492278e-01
6.12181664e-01 -5.04634976e-01 -1.70152172e-01 -2.35196471e-01
-1.49761036e-01 -1.38914271e-03 -4.50274497e-01 7.35390723e-01
-5.26478849e-02 4.50836509e-01 -1.21093380e+00 -8.41524363e-01
5.13096333e-01 4.01693404e-01 -4.70748693e-01 2.40766272e-01
-1.57435611e-01 7.27981150e-01 -9.97110486e-01 5.50145149e-01
1.01684892e+00 -2.31594786e-01 1.85305104e-02 -9.49573219e-02
-4.91276294e-01 -9.34539586e-02 -1.56814802e+00 1.57129407e+00
-1.29638791e-01 3.73404145e-01 3.00250947e-01 -4.85069722e-01
8.29126716e-01 2.30347916e-01 5.62338650e-01 -5.22746854e-02
3.54183950e-02 7.44360611e-02 -3.85626346e-01 -2.29633614e-01
7.59381175e-01 -6.53394312e-02 3.08082342e-01 4.30820882e-02
-5.40840030e-01 1.69825256e-01 4.34170850e-02 3.17003541e-02
8.20443213e-01 6.19575322e-01 4.21118170e-01 -9.21918452e-02
6.81993902e-01 1.56093091e-01 1.14029956e+00 9.51072991e-01
-5.47026157e-01 2.93197393e-01 -2.80997694e-01 -4.23102409e-01
-6.17540538e-01 -1.00343597e+00 -1.42686129e-01 7.15530574e-01
7.10622668e-01 -2.81721354e-01 -6.72915161e-01 -4.25779164e-01
-1.95634887e-01 2.92996436e-01 -2.27120847e-01 8.11867267e-02
-7.77839124e-01 -6.03212774e-01 1.38409287e-01 4.91045922e-01
7.70841777e-01 -7.66288221e-01 -9.47547615e-01 6.32064164e-01
-1.71608955e-01 -9.14985240e-01 -7.78901875e-01 -4.12998289e-01
-1.16858065e+00 -9.40576851e-01 -1.05825484e+00 -6.00049734e-01
6.12541020e-01 6.57491684e-01 5.25259554e-01 1.09458581e-01
-1.42854676e-01 2.06457302e-01 -1.42819211e-01 -3.54924239e-02
-1.59559563e-01 -2.80740082e-01 1.42277658e-01 1.32395342e-01
2.87560016e-01 -2.65721142e-01 -9.07792509e-01 5.71840882e-01
-6.75030112e-01 -4.19796258e-02 6.54426157e-01 8.25157225e-01
7.34305680e-01 1.91793174e-01 2.77473509e-01 -4.30704325e-01
2.77658731e-01 -4.14770663e-01 -1.12344646e+00 1.64571591e-02
-3.01625192e-01 -1.48956582e-01 4.73052979e-01 -8.14420044e-01
-1.40560222e+00 4.39822525e-01 -1.69171281e-02 -7.05071867e-01
-9.50721800e-02 3.81372094e-01 -1.07747734e-01 -2.56897569e-01
1.71158582e-01 5.32549500e-01 -2.23195016e-01 -5.51916659e-01
2.38594890e-01 4.26129490e-01 6.53334796e-01 -5.44850230e-01
1.06103885e+00 9.07052457e-01 2.29881871e-02 -8.03343952e-01
-6.73662126e-01 -8.07392776e-01 -3.50273103e-01 -5.29514194e-01
7.51002371e-01 -1.08728731e+00 -8.20431709e-01 4.35480326e-01
-1.08524144e+00 8.97716433e-02 -4.74662371e-02 9.23822880e-01
-4.04034674e-01 9.62543607e-01 -4.82130527e-01 -1.42364657e+00
-1.28394574e-01 -1.08988988e+00 1.03348231e+00 6.89671695e-01
-6.89747706e-02 -9.08355892e-01 1.82666838e-01 -9.75314900e-02
2.78999448e-01 2.12804824e-01 5.79531789e-02 -4.21167202e-02
-1.05758572e+00 -2.33857512e-01 -1.15349017e-01 -2.57190377e-01
1.62016168e-01 -9.93677899e-02 -6.20166659e-01 -5.48571944e-01
1.45511210e-01 2.00541764e-01 7.89233625e-01 8.38118315e-01
9.80301142e-01 -9.06560421e-02 -7.67711878e-01 5.52333415e-01
1.25751197e+00 2.56363124e-01 4.05048072e-01 2.46907860e-01
5.56104839e-01 1.64698556e-01 1.37019074e+00 8.74250829e-01
1.24798201e-01 8.20634782e-01 1.70180261e-01 1.93196565e-01
-2.03326531e-02 -4.18175817e-01 3.83190572e-01 6.21315777e-01
5.76004907e-02 2.68677864e-02 -4.58746582e-01 6.76769555e-01
-1.98715162e+00 -1.12021542e+00 -5.15906252e-02 2.46424079e+00
8.65021527e-01 3.84834319e-01 1.38605952e-01 -2.30900392e-01
1.09236515e+00 4.02915359e-01 -6.52274728e-01 5.61151505e-01
2.72556275e-01 -4.22079861e-01 5.29656708e-01 4.63593006e-01
-1.05165029e+00 8.95319939e-01 6.30418777e+00 1.21572602e+00
-8.28064442e-01 -7.43694836e-03 9.07639042e-02 1.49148270e-01
-4.24103811e-02 2.96330094e-01 -1.34978127e+00 8.34019423e-01
4.19550180e-01 -4.11575586e-02 5.32155074e-02 1.00096500e+00
7.17786610e-01 -4.98807669e-01 -4.39379960e-01 8.96464586e-01
-3.87367219e-01 -1.20370281e+00 -1.88471377e-01 2.79117171e-02
7.34308720e-01 -1.94474101e-01 -5.19519411e-02 -1.07084429e-02
4.59934413e-01 -2.35445350e-01 6.89913511e-01 7.18824804e-01
3.64438206e-01 -6.95779860e-01 3.59445214e-01 5.96527874e-01
-1.73998332e+00 -7.51858484e-03 -5.40247977e-01 5.55779114e-02
6.15955949e-01 7.94406652e-01 -5.55551410e-01 4.69442070e-01
7.45845318e-01 8.46999586e-01 -1.89358383e-01 1.71456754e+00
-1.84431106e-01 7.50480592e-01 -5.18251121e-01 -1.57927126e-01
2.35499233e-01 -3.25716019e-01 1.09738016e+00 1.02924061e+00
3.87592942e-01 4.07525450e-02 6.74869955e-01 7.08395302e-01
4.52755660e-01 -1.18934631e-01 -2.20864102e-01 3.04893732e-01
8.34065557e-01 1.11122143e+00 -9.18380499e-01 -4.93529141e-01
-2.82793105e-01 7.42836177e-01 -2.85378043e-02 4.34431851e-01
-1.01216424e+00 -4.36107844e-01 4.62233871e-01 -4.35214080e-02
7.73778856e-01 -4.42173660e-01 1.00160530e-02 -1.16411078e+00
7.90307894e-02 -5.63262939e-01 3.59369516e-01 -3.71731460e-01
-1.02022374e+00 1.58501595e-01 1.31823197e-01 -1.69537640e+00
-2.04582855e-01 2.92634424e-02 -8.74455631e-01 8.38672280e-01
-1.63373005e+00 -7.71936476e-01 -4.23342228e-01 4.85549837e-01
8.57369244e-01 2.30719388e-01 4.82480275e-03 1.97990850e-01
-4.41665977e-01 2.21936256e-01 3.11975718e-01 -3.91379111e-02
6.15497828e-01 -8.60354125e-01 2.69343942e-01 1.18117833e+00
-2.17589617e-01 8.44402373e-01 9.75382566e-01 -1.11961758e+00
-1.24519825e+00 -1.15412223e+00 5.69848955e-01 5.79661056e-02
6.41369104e-01 -7.50526637e-02 -1.09027004e+00 2.81671077e-01
-1.12797759e-01 9.75399315e-02 2.71181583e-01 -2.51335353e-01
2.01007485e-01 1.42943943e-02 -9.37146008e-01 6.33737266e-01
9.45141315e-01 -9.72495377e-02 -5.05565763e-01 1.24266990e-01
3.46238196e-01 -4.85190481e-01 -7.93987334e-01 3.49560648e-01
7.16507494e-01 -6.50891423e-01 8.87886107e-01 1.51137203e-01
-4.37145442e-01 -1.04618871e+00 2.12079391e-01 -1.00956869e+00
-4.53846246e-01 -1.11566865e+00 -5.62755585e-01 1.23138809e+00
-3.25208962e-01 -4.67905134e-01 9.64024425e-01 1.08365200e-01
1.40468091e-01 -5.01264691e-01 -9.87067103e-01 -9.29184854e-01
-6.26462340e-01 -1.66038632e-01 3.89000148e-01 5.35197318e-01
-3.67840111e-01 -1.37768447e-01 -6.31162167e-01 4.42001253e-01
1.17315865e+00 2.16581196e-01 9.01413858e-01 -1.13405430e+00
-4.08588111e-01 -2.07236648e-01 -2.12851837e-01 -1.90029335e+00
-1.36625484e-01 -6.51291236e-02 6.22474670e-01 -1.29283309e+00
3.09727043e-01 -4.10826355e-01 -2.50967771e-01 -2.53416359e-01
-7.72563994e-01 -4.67011482e-02 4.79466841e-02 7.07889557e-01
-9.46575284e-01 8.03764284e-01 1.16142654e+00 1.59018323e-01
-2.10547566e-01 4.86176938e-01 -1.60788044e-01 8.41970384e-01
3.98202628e-01 -6.33971393e-01 -2.94989049e-01 -2.04516947e-01
-2.35609964e-01 2.13105485e-01 4.52311575e-01 -1.26221359e+00
5.65058947e-01 -3.18444312e-01 5.59426844e-01 -1.26101279e+00
2.89091527e-01 -8.80637527e-01 1.44178852e-01 7.52257168e-01
-1.78797975e-01 -1.41358584e-01 -9.71109569e-02 1.34566259e+00
-1.62937522e-01 -2.08640546e-01 8.95200908e-01 6.21984303e-02
-7.90441871e-01 5.22459090e-01 -5.61014593e-01 -2.83512175e-01
1.01801395e+00 -4.02277112e-01 3.17602634e-01 -3.88555467e-01
-5.34384787e-01 4.57864881e-01 5.84534943e-01 9.54941958e-02
6.60030782e-01 -1.45441628e+00 -4.70439345e-01 -4.03237306e-02
-8.57026428e-02 5.47552370e-02 2.20593959e-01 1.08211184e+00
-2.71221012e-01 1.01074509e-01 2.40174904e-01 -9.64660406e-01
-1.21606767e+00 4.85015303e-01 1.07240655e-01 -2.51496851e-01
-7.98412085e-01 5.66031933e-01 1.60508096e-01 2.20242128e-01
4.13736075e-01 -3.08111191e-01 -2.27180243e-01 -2.62963295e-01
8.76833141e-01 6.44384325e-01 -6.56850934e-01 -5.38760304e-01
-3.25397491e-01 8.04255009e-01 -7.71026313e-02 -1.71261013e-01
8.04522157e-01 -6.34055018e-01 3.54941010e-01 3.01444858e-01
6.65293217e-01 9.11578462e-02 -1.99923015e+00 -4.36521888e-01
2.86316782e-01 -1.00618351e+00 2.13395998e-01 -5.45299798e-02
-8.47048521e-01 5.22276282e-01 6.09952271e-01 -1.51602970e-02
1.07707441e+00 -2.74582177e-01 1.05877495e+00 1.22933529e-01
3.91628593e-01 -9.95994747e-01 -1.77223906e-01 3.64980817e-01
1.15602463e-01 -1.09989965e+00 3.09540719e-01 -5.66100776e-01
-3.92568141e-01 1.04548001e+00 4.85852629e-01 -2.20134452e-01
4.61246997e-01 1.02700301e-01 -1.16268881e-01 1.10932328e-01
-5.06270349e-01 -2.11906716e-01 1.35706842e-01 5.18795133e-01
7.45883584e-02 -2.76079714e-01 -4.57566977e-01 6.12416212e-03
4.36431974e-01 3.57882649e-01 1.43301293e-01 1.19222677e+00
-8.67636025e-01 -8.13024223e-01 -7.44830370e-01 1.32222828e-02
-4.49073881e-01 1.28538638e-01 2.25454882e-01 5.17331481e-01
-8.32840726e-02 9.98283029e-01 -1.22900039e-01 -2.42375433e-02
5.09923883e-02 -4.16302800e-01 3.38611454e-01 -2.08447739e-01
-2.38689721e-01 7.75417864e-01 -2.90800154e-01 -6.31296873e-01
-6.57639325e-01 -1.04398298e+00 -1.08331466e+00 -1.13093937e-02
-8.27810407e-01 5.21113157e-01 4.99135911e-01 9.02007282e-01
1.55722007e-01 2.49747396e-01 7.67265379e-01 -1.02801669e+00
-8.47703218e-01 -8.79077375e-01 -4.90547776e-01 4.78108227e-02
4.21467960e-01 -9.18237925e-01 -2.32022434e-01 5.77626042e-02] | [6.410825729370117, -2.0736398696899414] |
2a6706ea-76d0-4a00-a827-061d54ed9b01 | a-deep-learning-approach-for-diabetic | null | null | https://ieeexplore.ieee.org/abstract/document/9298201/ | https://ieeexplore.ieee.org/abstract/document/9298201/ | A Deep Learning Approach for Diabetic Retinopathy detection using Transfer Learning | Diabetic Retinopathy is a primary complication of diabetes which more often than not, affects both eyes and anyone with type-1 or type-2 diabetes can develop it. A Diabetic patient should undergo eye tests periodically as the pace of development of this condition is slow. A Dataset of Fundus Photographs of retina is considered. Thus, there is a notable value in automatically categorizing the Fundus Photographs. Therefore, to get a consolidated and objective medical diagnosis, this paper proposes a transfer learning based approach for Diabetic Retinopathy categorization. The resizing of the dataset is performed, which converts the varied images into 224x224 format. The images are augmented using AUGMIX and pooled using GeM. Then, we have used pretrained models, namely SEResNeXt32x4d and EfficientNetb3. The pretraining of the aforementioned neural networks has been done on the ImageNet dataset. Then, the Diabetic Retinopathy images are migrated to these models. Based on the dataset already available, the output is ultimately split up into 5 levels according to the seriousness of the degree of DR. The experimental results show that the training accuracy of this method can reach as high as 0.91. Hence, the retina images of the Diabetic and Healthy patients can be easily classified using our proposed methodology, consequently reducing the number of reviews by medical professionals. | ['Himangi Pande', 'Karan Javali', 'Shardul Pharande', 'Bhargav Patil', 'Satwik Ramchandre'] | 2021-01-01 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [ 9.41039845e-02 -4.17473242e-02 -2.97284015e-02 -4.90536302e-01
-2.51474082e-01 6.30452260e-02 1.18936360e-01 -1.23374403e-01
-4.35734302e-01 8.67602944e-01 5.17775938e-02 -4.04684544e-01
-6.75836951e-02 -7.80692816e-01 -2.17037946e-01 -8.06902468e-01
3.07801723e-01 5.78912832e-02 2.03811973e-02 1.26514360e-01
3.82225722e-01 5.89518130e-01 -1.88870776e+00 3.78580958e-01
1.30727637e+00 1.07621932e+00 2.58440733e-01 6.26344442e-01
-7.12496266e-02 9.13917720e-01 -5.07070005e-01 -1.46269813e-01
3.61661524e-01 -5.49317002e-01 -5.18567741e-01 5.51598430e-01
6.30563021e-01 -7.05209196e-01 -1.75998360e-01 1.22525990e+00
7.46571541e-01 -1.49153394e-03 6.71261311e-01 -6.03065729e-01
-6.10655129e-01 1.99810024e-02 -6.44400418e-01 3.48089755e-01
-1.59597114e-01 1.63736895e-01 1.43376902e-01 -6.26824439e-01
2.60971099e-01 1.15246367e+00 3.11190397e-01 4.48549092e-01
-8.23967814e-01 -5.44519007e-01 -1.37529671e-01 5.45006454e-01
-1.12494826e+00 -3.44851822e-01 3.04101378e-01 -7.58691847e-01
5.04181623e-01 -1.76558532e-02 7.66060710e-01 4.08121049e-01
2.81889588e-01 1.82895228e-01 1.40827012e+00 -5.07097185e-01
1.37946039e-01 1.81452021e-01 3.71613979e-01 5.01249433e-01
5.79365671e-01 4.81819771e-02 1.55594736e-01 3.65290731e-01
7.92391360e-01 1.54650912e-01 -3.18285406e-01 2.09830347e-02
-8.78737807e-01 4.86725956e-01 6.85995877e-01 1.50167778e-01
-6.26849532e-01 -4.26031798e-01 4.02450264e-01 2.01782733e-01
2.02377155e-01 -7.02473894e-03 -2.13665336e-01 3.26878458e-01
-5.27199328e-01 -2.78650939e-01 1.74024642e-01 6.51601493e-01
7.91376770e-01 -8.97290334e-02 -1.89010408e-02 1.01200700e+00
3.64118069e-01 2.63642132e-01 6.16173148e-01 -6.58018351e-01
1.53346494e-01 1.16078901e+00 -4.88584638e-02 -8.12280893e-01
-3.18033755e-01 -3.92282546e-01 -1.09098768e+00 7.53933966e-01
3.08367580e-01 -2.81412929e-01 -1.48318160e+00 9.71823275e-01
2.99916804e-01 1.12223327e-01 3.98714125e-01 1.06295466e+00
9.53672588e-01 5.24481058e-01 2.36275032e-01 -2.09888667e-01
1.51341712e+00 -8.60276103e-01 -6.88768446e-01 -3.19456458e-02
4.93020505e-01 -9.08582389e-01 6.31373703e-01 4.74738657e-01
-8.43172491e-01 -8.30467343e-01 -9.54275906e-01 -2.91536808e-01
-2.17132375e-01 8.57044399e-01 5.77883065e-01 4.88332927e-01
-1.04383206e+00 2.41313413e-01 -5.75784028e-01 -6.96152508e-01
6.69081509e-01 3.23044717e-01 -1.98629811e-01 -4.97337580e-01
-7.44697928e-01 1.16350949e+00 5.29648066e-01 4.41464543e-01
-3.69543374e-01 -4.14305806e-01 -4.71092165e-01 -2.14114189e-01
-2.49939457e-01 -9.53321397e-01 8.39403093e-01 -1.02285409e+00
-1.25532079e+00 1.00840950e+00 -5.44542447e-02 -3.99411917e-01
5.43147445e-01 9.80848819e-02 -7.27631330e-01 3.10153872e-01
-4.49318662e-02 6.21505976e-01 5.37220836e-01 -1.00200725e+00
-1.31085622e+00 -5.72459340e-01 1.40220210e-01 3.18315208e-01
-1.26826167e-01 1.77403659e-01 -4.15018648e-01 -2.07497105e-01
-4.09403257e-03 -7.48572886e-01 -1.47207379e-01 1.76985949e-01
-2.92551100e-01 -2.12149158e-01 4.27065015e-01 -7.90824711e-01
9.91901040e-01 -2.17514873e+00 -2.67116457e-01 1.67664751e-01
3.22486490e-01 7.42917180e-01 -1.80353802e-02 -2.47920871e-01
-2.32855394e-01 -8.98756087e-02 2.68534068e-02 1.58459544e-01
-7.06107616e-01 9.03882086e-02 2.58173913e-01 2.81598747e-01
1.77016318e-01 2.95972466e-01 -4.60334837e-01 -4.54865843e-01
4.44624722e-01 6.06822670e-01 -3.02932262e-01 2.61044890e-01
1.76941648e-01 4.71824586e-01 -4.29029107e-01 6.97660625e-01
9.58852112e-01 -2.75081664e-01 7.33084455e-02 -4.64746684e-01
-3.80526870e-01 -8.54417831e-02 -8.99019420e-01 1.00361753e+00
-2.67895967e-01 6.70627654e-01 -3.78687918e-01 -9.48114395e-01
1.06096089e+00 2.40277007e-01 2.83738583e-01 -6.39266074e-01
4.07578766e-01 2.48014703e-01 4.04014617e-01 -9.64083970e-01
-7.97025207e-03 8.23294893e-02 8.23863626e-01 4.08900492e-02
-3.57409388e-01 5.02163351e-01 5.65278351e-01 -3.67020696e-01
5.48345864e-01 -2.04437878e-02 5.39128482e-01 3.99744064e-02
8.06132436e-01 2.68146992e-01 5.21003008e-01 1.28451958e-01
-1.23435423e-01 4.40201044e-01 3.96428198e-01 -6.47209704e-01
-9.47223246e-01 -7.43558824e-01 -5.74907005e-01 2.78675705e-01
1.15067817e-01 3.00253600e-01 -7.28216887e-01 -3.15536618e-01
-1.91015229e-02 4.28998530e-01 -5.40328979e-01 -1.66064024e-01
-2.37829491e-01 -1.07405603e+00 5.65633103e-02 3.96346748e-01
1.02077258e+00 -9.38127935e-01 -3.48838001e-01 -3.80534418e-02
8.25905800e-02 -7.37712860e-01 -5.22563979e-02 -4.98938829e-01
-1.21470940e+00 -1.56357431e+00 -8.91415179e-01 -1.29422557e+00
1.15033388e+00 6.03710949e-01 4.61267352e-01 1.07239105e-01
-6.09110355e-01 -7.03420639e-02 -1.34785339e-01 -3.89529675e-01
-5.08413374e-01 -4.10551041e-01 -1.22889258e-01 4.28840786e-01
8.45620871e-01 -3.69765192e-01 -1.01404548e+00 1.27978146e-01
-6.57165527e-01 1.06864609e-01 1.18412089e+00 4.15204644e-01
5.53610265e-01 2.43198916e-01 5.43008626e-01 -7.49379098e-01
4.12442148e-01 -2.23261088e-01 -7.20615566e-01 1.55659184e-01
-7.97345161e-01 -4.68284130e-01 2.86522210e-01 -3.20175856e-01
-1.09454095e+00 4.41696011e-02 2.77628303e-01 -2.69641578e-01
-5.19037426e-01 6.00655377e-01 -1.38538331e-01 -1.44637346e-01
6.11730456e-01 -3.27297486e-02 5.07713437e-01 -5.34893334e-01
3.97753157e-02 1.23130751e+00 6.05424106e-01 2.73641497e-01
3.32845956e-01 1.73468769e-01 -2.07269490e-01 -8.71498883e-01
-6.22973502e-01 -4.21461135e-01 -3.67047817e-01 -3.02459836e-01
1.19078743e+00 -1.11339152e+00 -6.06291831e-01 8.76477301e-01
-9.13952351e-01 5.56401461e-02 1.55738130e-01 1.03441226e+00
-1.77038953e-01 3.57105821e-01 -4.65794027e-01 -4.70022649e-01
-3.13510418e-01 -1.15115285e+00 2.35625476e-01 7.99500287e-01
2.34230250e-01 -8.07781577e-01 -1.94399655e-01 6.31536901e-01
3.06404740e-01 1.35173768e-01 1.26155722e+00 -3.36442620e-01
-8.08596730e-01 -2.43289828e-01 -6.57119036e-01 8.81373644e-01
5.84689140e-01 3.81174535e-01 -7.53169954e-01 -5.09813838e-02
-1.17304958e-01 3.09201870e-02 9.40303504e-01 7.40693390e-01
8.90947044e-01 -1.19823299e-01 -2.22496554e-01 5.08126915e-01
1.60778177e+00 7.52369523e-01 1.13836598e+00 6.39287293e-01
4.87593740e-01 5.20886660e-01 5.80843270e-01 1.32407233e-01
5.85877478e-01 1.74635664e-01 4.53143388e-01 -6.25970721e-01
-5.17232955e-01 3.32591444e-01 7.29875863e-02 6.80851460e-01
-6.25133812e-01 -5.30580953e-02 -7.33131230e-01 5.86711824e-01
-1.46555054e+00 -7.85366833e-01 -4.64325368e-01 2.33517313e+00
8.63753676e-01 -1.66833103e-01 -1.41862810e-01 -1.55193387e-02
1.35288715e+00 -4.94142950e-01 -6.34006381e-01 -3.08821648e-01
-2.83058118e-02 5.34022637e-02 3.76931161e-01 1.25957742e-01
-1.02579093e+00 5.38411796e-01 5.84792614e+00 1.05282098e-01
-1.55575871e+00 -4.32752311e-01 7.08539367e-01 4.72281910e-02
4.30971026e-01 -1.46391585e-01 -7.59213269e-01 7.07438648e-01
6.80113494e-01 -2.27875829e-01 2.92873144e-01 6.46432459e-01
5.53364158e-01 -4.51429397e-01 -7.96908796e-01 1.03977120e+00
1.52095079e-01 -9.86334383e-01 3.17616373e-01 9.07876119e-02
7.39023626e-01 -1.23347424e-01 1.13812588e-01 1.55061498e-01
-4.41352129e-02 -9.39103961e-01 -1.68375060e-01 1.02629483e+00
9.04058874e-01 -7.26825833e-01 1.14731634e+00 7.42084384e-02
-5.38527131e-01 -1.80701271e-01 -6.21932089e-01 1.25626191e-01
-1.25197440e-01 5.72542608e-01 -8.46932232e-01 3.77918184e-01
6.35197043e-01 8.81636083e-01 -7.46854484e-01 1.67596245e+00
-1.24328122e-01 4.36710119e-01 1.02419607e-01 2.71609217e-01
-1.12990178e-01 -6.68014050e-01 1.33224398e-01 6.41197562e-01
6.64875507e-01 4.24971491e-01 -6.13287650e-02 5.35186231e-01
5.60276657e-02 3.23649138e-01 -3.00350785e-01 1.78197011e-01
4.34885807e-02 1.27329206e+00 -2.85294831e-01 -5.18523455e-01
-6.28951609e-01 4.49146718e-01 -2.05050066e-01 4.26981390e-01
-5.10071397e-01 -4.23305899e-01 5.04363179e-01 1.11525200e-01
2.81079888e-01 4.08867449e-01 -2.67365098e-01 -9.01340961e-01
-3.06836553e-02 -9.04722631e-01 2.68278986e-01 -1.11187470e+00
-1.20908499e+00 8.20214629e-01 -4.72146124e-01 -1.48516846e+00
1.05098218e-01 -7.28881836e-01 -5.54225862e-01 1.13622451e+00
-1.87381232e+00 -9.97442186e-01 -8.57627273e-01 4.91066217e-01
2.23502964e-01 -5.68474114e-01 6.78776741e-01 5.83461404e-01
-9.60646093e-01 1.52459577e-01 3.82927395e-02 1.43121541e-01
1.04510152e+00 -1.03318274e+00 -3.57172430e-01 8.72651279e-01
-8.42788696e-01 5.00765026e-01 2.86565632e-01 -6.66858554e-01
-6.11993849e-01 -1.52783942e+00 8.39406729e-01 1.60213232e-01
2.99039155e-01 6.49077535e-01 -9.59302962e-01 4.72218901e-01
2.30706602e-01 5.09078205e-02 7.44762838e-01 -3.07082653e-01
8.71170163e-02 -4.44415361e-01 -1.19841993e+00 5.83303988e-01
5.26831627e-01 -1.61130503e-01 -6.66978061e-01 3.87664348e-01
1.13277696e-01 -3.59650403e-01 -1.09310222e+00 4.16082859e-01
3.82898152e-01 -1.08207834e+00 5.81352234e-01 -7.06137896e-01
7.65016556e-01 -6.06859922e-01 1.39618263e-01 -1.14601803e+00
-2.27639809e-01 -6.69748858e-02 3.71847987e-01 1.15334451e+00
2.13248596e-01 -1.02384317e+00 4.53843057e-01 5.74755907e-01
-2.51917928e-01 -4.51651543e-01 -4.25277829e-01 -3.13444138e-01
-1.78328469e-01 4.45491135e-01 3.42979848e-01 8.43411684e-01
-4.25601393e-01 2.85084069e-01 9.50846225e-02 4.76516545e-01
5.85876822e-01 1.08158402e-01 6.94375515e-01 -1.39220238e+00
3.46177787e-01 -2.86117971e-01 -8.14621925e-01 -6.14691436e-01
-3.82368982e-01 -8.08119774e-01 -3.08980197e-01 -2.00960231e+00
1.54228017e-01 -3.48837733e-01 -5.02182126e-01 5.13354003e-01
-2.80225098e-01 4.33949083e-01 -2.61388302e-01 3.60223770e-01
4.92263921e-02 1.15121782e-01 1.56781650e+00 -2.81052087e-02
-4.10484314e-01 3.07291597e-01 -8.95225167e-01 7.66846418e-01
1.05001259e+00 -3.36526823e-03 -3.99291575e-01 -3.45331371e-01
-2.39135340e-01 -4.54241112e-02 4.25859362e-01 -1.11029327e+00
2.42495015e-01 -1.02852441e-01 5.17271459e-01 -6.66481018e-01
2.26795450e-02 -7.39427030e-01 3.11079957e-02 3.41815382e-01
-1.16693728e-01 -3.94785613e-01 1.47661760e-01 3.20831776e-01
-4.34575796e-01 -1.09527037e-01 1.14332211e+00 -1.24380169e-02
-8.15567613e-01 2.62114018e-01 -4.86426085e-01 -3.68866056e-01
1.20361030e+00 -5.88548243e-01 -6.56345963e-01 8.55264589e-02
-8.39017272e-01 1.42337710e-01 4.51996028e-01 1.80835932e-01
5.97273946e-01 -1.11819184e+00 -9.24356341e-01 1.94794208e-01
2.68683374e-01 1.13566257e-01 6.74027741e-01 1.22011817e+00
-9.03319359e-01 2.53068179e-01 -7.58622169e-01 -5.62640667e-01
-1.52685201e+00 4.92224962e-01 5.52580774e-01 3.84359062e-01
-8.04696739e-01 2.32554615e-01 1.52716264e-01 2.32806295e-01
1.08499333e-01 -5.72484434e-01 -9.95571256e-01 2.34608322e-01
8.54907453e-01 5.65898895e-01 1.81828409e-01 -5.03502667e-01
-9.75037962e-02 8.80695760e-01 -5.01201987e-01 5.41798770e-01
1.27153635e+00 -1.87322214e-01 -4.84290153e-01 -8.32624063e-02
9.40657377e-01 -5.33660054e-02 -9.69575703e-01 -6.69569075e-02
-4.46417719e-01 -4.44947720e-01 2.82154262e-01 -9.47715402e-01
-1.28240013e+00 8.24904799e-01 1.22185767e+00 -4.45889607e-02
1.49431491e+00 -4.47188318e-01 5.42255998e-01 1.93389848e-01
1.62942670e-02 -7.95295596e-01 -4.92202193e-01 5.61965890e-02
6.24673903e-01 -1.21126318e+00 9.34151188e-02 -4.63442743e-01
-6.92908764e-01 1.31655979e+00 5.49606323e-01 -2.52944708e-01
3.86966646e-01 -3.05504173e-01 7.28596985e-01 1.27353057e-01
-5.20051956e-01 -4.32899266e-01 2.93611526e-01 7.36617029e-01
2.66997844e-01 -1.60862520e-01 -5.52629232e-01 1.58291861e-01
1.03642708e-02 5.84429204e-01 9.29747939e-01 5.05310178e-01
-6.68664277e-01 -9.28226352e-01 -3.28684241e-01 6.76367998e-01
-3.13281834e-01 -2.61728112e-02 -4.83978391e-02 7.98996210e-01
3.90410036e-01 9.93987083e-01 3.18810582e-01 -8.89859796e-02
4.05568719e-01 -2.16324672e-01 3.57890964e-01 -6.01776838e-01
-1.15590416e-01 1.19866803e-01 1.48525357e-01 -1.05185099e-01
-8.16808999e-01 -4.06687409e-01 -1.22798944e+00 -1.42872065e-01
-1.93357900e-01 -1.59483999e-01 6.39529109e-01 6.70935094e-01
4.82960790e-01 6.89870834e-01 7.30787933e-01 -3.10190350e-01
-4.52971399e-01 -1.11776519e+00 -5.84789991e-01 4.45363969e-02
3.33567470e-01 -5.18805087e-01 -4.40358698e-01 5.27265608e-01] | [15.84024429321289, -3.993350028991699] |
c06de952-30d6-49b7-8638-b0a897f5f331 | customer-profiling-segmentation-and-sales | 2302.01786 | null | https://arxiv.org/abs/2302.01786v1 | https://arxiv.org/pdf/2302.01786v1.pdf | Customer Profiling, Segmentation, and Sales Prediction using AI in Direct Marketing | In an increasingly customer-centric business environment, effective communication between marketing and senior management is crucial for success. With the rise of globalization and increased competition, utilizing new data mining techniques to identify potential customers is essential for direct marketing efforts. This paper proposes a data mining preprocessing method for developing a customer profiling system to improve sales performance, including customer equity estimation and customer action prediction. The RFM-analysis methodology is used to evaluate client capital and a boosting tree for prediction. The study highlights the importance of customer segmentation methods and algorithms to increase the accuracy of the prediction. The main result of this study is the creation of a customer profile and forecast for the sale of goods. | ['Islam Taj-Eddin', 'Mohamed Hamada', 'Mahmoud SalahEldin Kasem'] | 2023-02-03 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [-1.02039874e-01 -1.71431854e-01 -6.41027212e-01 -9.29756343e-01
-2.14441016e-01 -1.37214556e-01 1.17289415e-02 6.84886873e-01
-4.16645288e-01 5.03031909e-01 1.48075074e-01 -6.51986063e-01
-4.23741490e-01 -1.01384044e+00 6.80783167e-02 -3.81833971e-01
3.71309310e-01 9.29047465e-01 -1.60325453e-01 -3.33089828e-01
1.12909067e+00 9.10669684e-01 -1.48741961e+00 4.57348377e-01
9.47828412e-01 1.20941591e+00 2.94750512e-01 4.42732483e-01
-5.44449329e-01 7.67547667e-01 -4.04715270e-01 -6.05494022e-01
2.64652967e-01 -2.80718982e-01 -7.70483971e-01 4.26629812e-01
-4.05323654e-01 -2.87337124e-01 7.21400738e-01 5.13671875e-01
3.83543283e-01 3.18835564e-02 5.68628430e-01 -1.03086162e+00
-2.12056428e-01 7.27405667e-01 -7.36660361e-01 3.34787697e-01
1.67957187e-01 -5.71566343e-01 9.93710458e-01 -6.33373618e-01
3.33658129e-01 8.75551403e-01 2.41437748e-01 -1.75518274e-01
-9.77193058e-01 -7.53836691e-01 -3.34829316e-02 3.88456285e-01
-8.39046299e-01 1.73231572e-01 5.48307061e-01 -5.77490628e-01
8.58223736e-01 8.92704800e-02 1.04437637e+00 8.38028267e-02
1.71118855e-01 4.79880393e-01 1.21616781e+00 -6.50118649e-01
1.26021996e-01 8.96829724e-01 5.94510138e-01 -1.48035303e-01
4.76136714e-01 1.39890928e-02 -2.92222917e-01 3.27671170e-01
4.24965024e-01 3.35521430e-01 5.43709576e-01 -7.79403327e-03
-6.95677549e-02 1.42429829e+00 1.10381961e-01 3.66827130e-01
-6.77529037e-01 -6.35161459e-01 2.91727066e-01 2.03662038e-01
3.97927612e-01 4.36697960e-01 -4.59113717e-01 -4.94288385e-01
-1.07868397e+00 1.70581654e-01 9.05700922e-01 5.17092347e-01
3.97581130e-01 -5.46070114e-02 5.68120480e-01 6.40548348e-01
4.68852401e-01 9.75387618e-02 5.75053155e-01 -7.01584220e-01
1.35693744e-01 1.23978794e+00 -9.91557613e-02 -1.08035195e+00
-4.73796070e-01 -1.11437511e+00 -1.33878037e-01 -6.02576099e-02
1.88751984e-02 2.30637044e-01 -7.44337380e-01 5.93645453e-01
1.63740292e-01 -5.50943017e-01 -9.34189372e-03 5.71074784e-01
3.76985788e-01 4.57539529e-01 2.86145657e-01 -2.28531957e-01
1.06402099e+00 -6.45438373e-01 -5.01196980e-01 7.89734162e-03
7.02389896e-01 -9.29389477e-01 7.18741119e-01 8.59729528e-01
-7.56038666e-01 -5.90640366e-01 -5.83805621e-01 4.28472519e-01
-4.63324755e-01 -1.27893731e-01 9.41527665e-01 8.85099173e-01
-9.07396600e-02 3.59506160e-01 -4.96260017e-01 -4.60093886e-01
3.72657269e-01 6.96175098e-01 7.57320842e-04 -1.03624046e-01
-6.59419417e-01 1.04791367e+00 4.99195486e-01 -1.10487975e-01
1.17512311e-04 -4.26055580e-01 -3.89882535e-01 8.90881270e-02
-2.08343595e-01 -1.68213591e-01 1.12792432e+00 -1.18425822e+00
-7.75840878e-01 8.26930046e-01 6.57295734e-02 -6.29044354e-01
3.00783038e-01 -1.75707191e-01 -5.83050251e-01 -1.58257961e-01
2.00594708e-01 6.16397075e-02 2.05386072e-01 -1.29574680e+00
-1.76861668e+00 -9.13727164e-01 -4.92707968e-01 1.07632652e-01
4.90688393e-03 2.23622188e-01 -1.64682735e-02 -4.10420410e-02
5.20858109e-01 -4.44202036e-01 -3.27064902e-01 -1.59464693e+00
7.69722387e-02 -1.26555385e-02 7.07864523e-01 -1.03964508e+00
1.37174809e+00 -1.99956381e+00 -4.74248171e-01 9.44265246e-01
1.37168050e-01 -3.35076004e-02 6.85001612e-01 5.65656364e-01
-3.29832695e-02 1.05051719e-01 3.60094786e-01 1.40006170e-01
-1.76550224e-01 3.59840333e-01 -7.24417642e-02 4.10905154e-03
-3.03337425e-02 4.60284919e-01 -2.78718978e-01 -5.53132772e-01
3.35487127e-01 1.73489794e-01 -4.13494200e-01 2.77849376e-01
2.70545721e-01 3.05459976e-01 -4.50976819e-01 1.09855628e+00
6.91370547e-01 3.68304327e-02 2.37421155e-01 1.62626803e-01
-5.20666122e-01 2.04121828e-01 -1.01981020e+00 3.82467180e-01
-4.11002994e-01 1.93507135e-01 1.22992262e-01 -1.31519186e+00
1.46455979e+00 -1.56385258e-01 6.03869557e-01 -1.24920499e+00
6.55861378e-01 4.58780646e-01 2.28526846e-01 -4.72647160e-01
8.51197839e-01 -2.46006265e-01 -4.18440402e-02 1.05492234e-01
-4.16425437e-01 2.76315689e-01 3.26857328e-01 1.05131254e-01
1.55596420e-01 -1.43402234e-01 2.38272250e-01 -1.49690628e-01
3.13343495e-01 5.04852653e-01 7.27567554e-01 6.89677298e-02
-9.29633602e-02 -1.07771926e-01 5.16287982e-01 -3.60599756e-01
-9.00237918e-01 -5.01952112e-01 -3.65995675e-01 1.21876585e+00
-2.72757798e-01 2.30788231e-01 -5.50055027e-01 -1.35921597e-01
4.18040842e-01 9.71979678e-01 -1.64110273e-01 1.50341660e-01
-3.28205138e-01 -8.73216152e-01 -4.37286526e-01 6.12903237e-01
4.26466256e-01 -7.37451196e-01 -7.60818601e-01 5.98701537e-01
1.66400626e-01 -7.75758147e-01 3.33366394e-01 5.04539013e-01
-1.25912845e+00 -1.02787912e+00 -1.91776797e-01 -9.77134943e-01
6.24468505e-01 5.07246144e-02 1.00587022e+00 -9.07004904e-03
-3.43742937e-01 -5.55058159e-02 -6.43823326e-01 -8.27062964e-01
-4.36442077e-01 5.06581366e-01 -4.30977881e-01 -1.70852780e-01
1.09698200e+00 -2.89715260e-01 -5.72206795e-01 3.92405808e-01
-3.80506516e-01 -1.37129366e-01 7.63084114e-01 4.33916032e-01
4.68430281e-01 7.36534536e-01 1.11291862e+00 -1.22017598e+00
8.81307185e-01 -6.30887389e-01 -5.90515494e-01 1.64059959e-02
-1.60052085e+00 -4.76477504e-01 7.24657625e-02 -3.80160101e-02
-1.09822488e+00 -1.07612565e-01 -2.43546769e-01 5.83072126e-01
-1.05087407e-01 8.32619607e-01 4.78103086e-02 2.71788985e-01
2.38303021e-01 -1.66051969e-01 2.61303842e-01 -6.07584655e-01
-1.68262109e-01 1.06742787e+00 3.19186330e-01 -8.62370953e-02
1.12073243e-01 -4.92213257e-02 -1.06015742e-01 -5.97306848e-01
-6.90568388e-01 -1.10044920e+00 -8.32823277e-01 -5.22805572e-01
5.41844189e-01 -6.00622833e-01 -1.04422855e+00 1.67502373e-01
-3.34980249e-01 2.13705942e-01 -1.60150856e-01 9.27583158e-01
-1.45547345e-01 -2.64463603e-01 -3.98739487e-01 -1.07134402e+00
-2.97528327e-01 -8.73272300e-01 1.51326567e-01 5.95367074e-01
-4.29102182e-01 -8.69578838e-01 -1.69328883e-01 1.35555315e+00
3.74367148e-01 2.88043290e-01 1.03278458e+00 -1.02358866e+00
-4.36395079e-01 -7.21846402e-01 -2.18103409e-01 6.01054788e-01
-1.02569081e-01 2.05258563e-01 -3.46658379e-01 1.16970904e-01
1.41416565e-01 4.05488573e-02 4.75297093e-01 6.42977655e-01
5.86852014e-01 1.03421748e-01 -3.26173067e-01 2.28336602e-01
1.57262325e+00 9.93516028e-01 5.86655915e-01 9.92334485e-01
2.93877035e-01 1.27299881e+00 1.32463050e+00 7.15841174e-01
6.47275925e-01 7.91946799e-02 3.99119437e-01 -8.83634910e-02
6.48840249e-01 5.08590676e-02 -1.80749938e-01 7.16421366e-01
-2.31095076e-01 3.38370383e-01 -1.04982233e+00 4.98144716e-01
-1.77366078e+00 -8.86381090e-01 -3.44827414e-01 1.95624268e+00
3.53146017e-01 5.87327182e-01 7.98385978e-01 6.55755997e-01
6.02224350e-01 -8.91813934e-01 -1.75003663e-01 -1.32828999e+00
1.41527995e-01 4.97770309e-01 9.10109997e-01 5.03282487e-01
-5.31776905e-01 8.02911282e-01 6.48980284e+00 4.11810249e-01
-9.88435864e-01 -3.03868502e-01 9.85728741e-01 -3.43121588e-02
-2.16406152e-01 8.64655674e-02 -1.34937334e+00 2.90131062e-01
9.36523557e-01 -1.29070431e-01 2.32670344e-02 1.24759340e+00
5.52483141e-01 -6.81710780e-01 -3.42849582e-01 5.48779845e-01
-5.58236949e-02 -1.09514296e+00 -2.66657829e-01 7.01724946e-01
4.78554457e-01 -4.77378845e-01 -8.66600499e-03 3.78613591e-01
1.81455046e-01 -1.09621859e+00 3.99846345e-01 4.93291736e-01
-1.60375476e-01 -1.52320528e+00 1.33042729e+00 3.04139823e-01
-7.70968914e-01 -4.83621150e-01 -3.05414964e-02 -5.50337076e-01
4.06989664e-01 6.47169292e-01 -1.26036870e+00 2.31965348e-01
6.13226295e-01 1.70344710e-01 -2.74554193e-01 8.44799757e-01
3.66404653e-01 6.90571308e-01 5.60413748e-02 -1.73960015e-01
1.70076713e-01 -8.01177979e-01 -4.26098824e-01 9.47932720e-01
2.09185377e-01 -5.87698705e-02 5.19354418e-02 3.02489102e-01
3.39752734e-01 7.46115506e-01 9.78188030e-03 -4.18971390e-01
3.72334093e-01 1.25047219e+00 -9.01665330e-01 1.09413564e-02
-5.28986037e-01 3.18137884e-01 -1.63463086e-01 -2.12279081e-01
-4.31567013e-01 -4.27906603e-01 2.91584253e-01 9.50429797e-01
4.75292146e-01 -1.34123027e-01 -9.64143872e-01 -1.40504584e-01
-3.30901235e-01 -8.49565029e-01 4.15476412e-01 2.79257968e-02
-9.47385252e-01 1.19812712e-01 -2.10678741e-01 -4.13698912e-01
-2.77892828e-01 -6.06739461e-01 -5.56291580e-01 1.05124235e+00
-1.37293828e+00 -1.19679916e+00 -1.79206014e-01 1.70109153e-01
1.50651172e-01 -5.30914307e-01 4.77990270e-01 5.23035824e-01
-5.49844384e-01 1.80899978e-01 2.37270966e-01 -1.04846016e-01
8.67806152e-02 -1.03422487e+00 -3.55697572e-01 4.09092009e-01
-5.38286865e-01 3.76911163e-01 5.32526672e-01 -1.01886845e+00
-8.03640366e-01 -5.04312038e-01 1.41778433e+00 -5.17046526e-02
3.83137882e-01 -3.04653794e-02 -6.36617005e-01 3.98891181e-01
6.34066239e-02 -1.36462355e+00 1.42779434e+00 5.22614300e-01
3.13891530e-01 -6.19147122e-01 -1.34442830e+00 -2.09202096e-01
4.86936234e-02 -1.71583872e-02 -4.77176666e-01 -2.89671093e-01
9.82751846e-02 1.38432039e-02 -1.26860690e+00 3.32109332e-01
8.05444479e-01 -1.26249337e+00 5.87041259e-01 -4.69144464e-01
2.20265299e-01 3.73199552e-01 -2.00863808e-01 -4.47180331e-01
-3.66947383e-01 3.23146246e-02 5.20130396e-01 1.42077279e+00
7.53474295e-01 -6.99473500e-01 1.38424194e+00 1.04451728e+00
2.06991434e-01 -1.16060579e+00 -2.86165565e-01 -2.30837911e-01
-2.14592800e-01 -4.14576352e-01 5.81753552e-01 6.95058584e-01
-3.82865081e-03 4.99518126e-01 2.61709355e-02 -2.87617773e-01
4.20086741e-01 4.88792300e-01 7.89107740e-01 -1.52693844e+00
-1.23139076e-01 -5.77473998e-01 -5.72434485e-01 -3.22228014e-01
-4.43246603e-01 -4.78494972e-01 -7.32897699e-01 -1.73839366e+00
1.92485213e-01 -4.65449959e-01 -1.48085177e-01 -2.54000336e-01
3.02285761e-01 -8.48245621e-02 1.02871105e-01 2.94268668e-01
2.22799316e-01 -2.54969537e-01 9.39838946e-01 5.53101242e-01
-6.43374383e-01 7.49328136e-01 -1.26233721e+00 5.62597454e-01
1.19590223e+00 -4.17576104e-01 -4.65552151e-01 3.18720311e-01
6.75270617e-01 -1.25130951e-01 -1.84444785e-01 -4.75118697e-01
-6.63037971e-02 -3.31270099e-01 5.57998240e-01 -1.37008667e+00
-6.77908957e-02 -9.05903339e-01 2.71193206e-01 6.72592938e-01
-4.10271361e-02 3.32921028e-01 -1.50063068e-01 9.71858972e-04
-5.73910773e-01 -5.34941852e-01 7.38895357e-01 -9.90828150e-04
-6.68361187e-01 -1.00172900e-01 -2.80541748e-01 -3.80037516e-01
1.31127775e+00 -8.67300451e-01 6.19024374e-02 -3.95242602e-01
-8.95887971e-01 3.23292851e-01 3.25569898e-01 -2.62343083e-02
4.32833374e-01 -7.73099124e-01 -4.92686003e-01 1.70139313e-01
-1.92122459e-01 -1.34957597e-01 -5.49162962e-02 7.71627426e-01
-9.40675795e-01 5.19779682e-01 -6.53072894e-01 -1.78795412e-01
-1.32327855e+00 2.51121253e-01 -1.34509310e-01 -4.94369626e-01
-1.03405500e-02 7.65604138e-01 -3.29346627e-01 2.31852960e-02
2.35781863e-01 1.39447292e-02 -9.00069058e-01 7.31231153e-01
1.77696168e-01 9.07629251e-01 1.14992999e-01 -7.30131269e-01
-3.39870125e-01 3.14071715e-01 -4.28910524e-01 -1.41372606e-01
1.68293738e+00 -3.90705794e-01 -1.67300239e-01 4.39634770e-01
9.51687753e-01 7.31711537e-02 -6.62951827e-01 2.27977008e-01
6.26175523e-01 -9.68438447e-01 1.76175669e-01 -9.45519626e-01
-1.10015416e+00 7.33832419e-01 5.34133017e-01 5.90395629e-01
1.43181157e+00 -7.47971907e-02 7.01027811e-01 -3.05372775e-01
-2.20487751e-02 -1.86110389e+00 -7.94531628e-02 6.15826696e-02
2.98891038e-01 -1.39816761e+00 9.00634006e-02 -3.75155002e-01
-1.11872649e+00 1.11503851e+00 4.70739871e-01 1.79798380e-01
9.06188309e-01 9.20664296e-02 8.18513110e-02 -2.09565476e-01
-1.77658871e-01 -3.32627147e-01 -7.77421743e-02 5.57042599e-01
7.89883971e-01 3.68196607e-01 -8.05586159e-01 1.13477385e+00
-6.75295413e-01 -2.32962072e-02 2.86341846e-01 1.12548923e+00
-1.03825712e+00 -1.58321118e+00 -3.20925683e-01 8.90285432e-01
-9.60229874e-01 1.28496274e-01 -4.99843538e-01 1.16482186e+00
2.97104686e-01 1.07882464e+00 2.19348729e-01 -4.96233553e-01
4.61989045e-01 9.08854008e-02 1.95463121e-01 -4.09624159e-01
-8.24088037e-01 4.63214546e-01 3.83439392e-01 2.02699620e-02
-5.13492703e-01 -9.45320129e-01 -1.48106194e+00 -8.60031664e-01
-6.01299405e-01 5.00303149e-01 1.52860761e+00 6.79163992e-01
1.55658424e-01 4.39796776e-01 1.02670717e+00 5.22474945e-02
-2.56211519e-01 -8.11694026e-01 -1.04876864e+00 2.24339724e-01
-4.09326613e-01 -5.10741532e-01 -8.61470252e-02 -9.47518647e-02] | [9.25742244720459, 5.911329746246338] |
1e21ea6a-efa6-4b07-b985-5cf836c6cb8f | compressive-sensing-of-ecg-signals-using-plug | 2210.08204 | null | https://arxiv.org/abs/2210.08204v1 | https://arxiv.org/pdf/2210.08204v1.pdf | Compressive Sensing of ECG Signals using Plug-and-Play Regularization | Compressive Sensing (CS) has recently attracted attention for ECG data compression. In CS, an ECG signal is projected onto a small set of random vectors. Recovering the original signal from such compressed measurements remains a challenging problem. Traditional recovery methods are based on solving a regularized minimization problem, where a sparsity-promoting prior is used. In this paper, we propose an alternative iterative recovery algorithm based on the Plug-and-Play (PnP) method, which has recently become popular for imaging problems. In PnP, a powerful denoiser is used to implicitly perform regularization, instead of using hand-crafted regularizers; this has been found to be more successful than traditional methods. In this work, we use a PnP version of the Proximal Gradient Descent (PGD) algorithm for ECG recovery. To ensure mathematical convergence of the PnP algorithm, the signal denoiser in question needs to satisfy some technical conditions. We use a high-quality ECG signal denoiser fulfilling this condition by learning a Bayesian prior for small-sized signal patches. This guarantees that the proposed algorithm converges to a fixed point irrespective of the initialization. Importantly, through extensive experiments, we show that the reconstruction quality of the proposed method is superior to that of state-of-the-art methods. | ['Kunal Narayan Chaudhury', 'Ruturaj Gavaskar', 'Unni VS'] | 2022-10-15 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 7.15288520e-01 -1.64467424e-01 1.21098138e-01 -1.01131037e-01
-9.05631721e-01 -1.24678724e-01 1.30857050e-01 -5.84664904e-02
-3.53450388e-01 7.89257765e-01 6.23732433e-02 -8.32916424e-02
-4.48214889e-01 -4.23504263e-01 -6.41353726e-01 -1.06149006e+00
-9.07095969e-02 1.36145297e-02 -2.26414323e-01 -9.65688974e-02
8.60598236e-02 2.22862050e-01 -9.19880509e-01 -1.26140311e-01
1.05702651e+00 8.98232579e-01 4.32751656e-01 3.95767808e-01
3.04083943e-01 5.82618833e-01 -3.29069734e-01 -1.90578386e-01
4.52576995e-01 -9.82193470e-01 -3.93464744e-01 3.58110309e-01
-9.03956443e-02 -9.79128554e-02 -1.46369278e-01 1.13461757e+00
7.25362837e-01 2.03550845e-01 4.12586778e-01 -5.93041122e-01
9.93747823e-03 3.69355947e-01 -8.18570495e-01 1.99101284e-01
2.78672308e-01 -2.90610403e-01 5.47256351e-01 -1.09390032e+00
4.44235206e-01 5.38341522e-01 9.42575634e-01 3.34841371e-01
-1.41413713e+00 -4.27002251e-01 -3.26779813e-01 2.09086929e-02
-1.63431358e+00 -6.41036630e-01 1.25901544e+00 -2.57186711e-01
1.33948445e-01 3.71745706e-01 7.67883241e-01 6.38483584e-01
1.16244458e-01 6.52414262e-01 1.10881615e+00 -7.72367001e-01
4.22098666e-01 -1.54861987e-01 -2.99525499e-01 4.22590315e-01
3.77037674e-01 -2.53923357e-01 -4.28992748e-01 -4.83762801e-01
9.56762910e-01 1.94833651e-01 -7.69607306e-01 -4.89763111e-01
-1.15960872e+00 8.69781613e-01 2.94378459e-01 6.78319275e-01
-9.23489213e-01 1.71807483e-01 1.25851914e-01 1.66917637e-01
6.29031777e-01 2.28787556e-01 1.10924371e-01 1.61463648e-01
-1.43446672e+00 1.38558924e-01 7.07484305e-01 3.75079453e-01
4.56125170e-01 3.04084718e-01 -1.17887244e-01 9.51469004e-01
4.60828483e-01 3.72641981e-01 5.64598262e-01 -1.05274081e+00
3.82214010e-01 -8.46787244e-02 1.85998127e-01 -1.38706100e+00
1.10077681e-02 -9.33613002e-01 -1.38551462e+00 -1.11019447e-01
4.44215357e-01 -2.17961833e-01 -4.61089551e-01 1.62899923e+00
4.64684725e-01 7.84403801e-01 8.38616118e-02 1.08574820e+00
3.89921576e-01 6.63935423e-01 -2.97512263e-01 -9.05744553e-01
9.14765835e-01 -4.08468843e-01 -1.05543351e+00 -2.55445808e-01
1.60004497e-01 -7.55323946e-01 4.73086804e-01 7.44287193e-01
-1.27039981e+00 -4.46896881e-01 -1.13722908e+00 4.34333861e-01
6.48044765e-01 2.04226494e-01 1.62056491e-01 6.87171817e-01
-9.03372288e-01 7.91617572e-01 -8.53810728e-01 -4.22905162e-02
2.50242442e-01 1.49391174e-01 -3.21033567e-01 -4.04976845e-01
-9.47735429e-01 7.19347954e-01 6.99035674e-02 5.66806018e-01
-7.04925060e-01 -4.95697021e-01 -7.22550631e-01 -1.48885578e-01
3.86432320e-01 -6.70078278e-01 7.69346833e-01 -1.07479906e+00
-1.69231844e+00 8.14638734e-01 -3.19008827e-01 -3.87111157e-01
6.11207247e-01 -2.77384251e-01 -1.96697742e-01 5.23731053e-01
8.47441927e-02 -3.37131441e-01 1.59469712e+00 -1.30232203e+00
-5.28732054e-02 -2.68191844e-01 -4.54522699e-01 1.30926028e-01
-2.50734657e-01 -2.69392639e-01 -5.66108406e-01 -1.02651834e+00
7.12045252e-01 -8.40771317e-01 -5.06924272e-01 -1.40598580e-01
-2.21597522e-01 2.63747245e-01 3.59406561e-01 -1.08084643e+00
1.40915906e+00 -2.45972300e+00 4.64824736e-01 6.63166761e-01
2.81154364e-01 2.55243003e-01 1.17749192e-01 5.40475011e-01
-1.93418607e-01 -4.35584366e-01 -7.54816711e-01 -5.33965945e-01
-4.35304463e-01 1.86406225e-01 -1.59954026e-01 1.16674066e+00
-2.24090755e-01 4.66720641e-01 -9.05243695e-01 -3.78621608e-01
1.65173367e-01 7.20536411e-01 -5.66189468e-01 3.31412077e-01
3.90770793e-01 9.49575961e-01 -5.26826501e-01 2.14262530e-01
7.65855730e-01 -3.46981198e-01 4.80922699e-01 -2.42004707e-01
-1.00555286e-01 -1.06083192e-01 -1.57853007e+00 2.04715967e+00
-3.05572569e-01 1.99459642e-01 5.94018638e-01 -1.66409326e+00
9.36943412e-01 6.84105933e-01 8.25372458e-01 -3.19752604e-01
6.41653091e-02 6.59910083e-01 -1.76296666e-01 -5.78334570e-01
-7.05263689e-02 -5.58674812e-01 3.91842037e-01 2.96766967e-01
-2.99435765e-01 -4.37399633e-02 -1.86001882e-02 8.86513889e-02
1.11545575e+00 2.67919526e-02 6.41881645e-01 -3.86060327e-01
6.89810336e-01 -3.51012707e-01 7.96157479e-01 7.68690109e-01
1.24490030e-01 8.93183410e-01 1.48096750e-03 -1.64722711e-01
-7.92351902e-01 -5.56618273e-01 -4.40911472e-01 3.75389516e-01
6.21108413e-02 -2.69061714e-01 -8.18854511e-01 -2.12191343e-01
-2.88693041e-01 3.43999654e-01 -2.88968414e-01 -4.79057655e-02
-8.49272013e-01 -1.04054570e+00 1.85035825e-01 8.64360183e-02
4.79703516e-01 -6.69603467e-01 -8.42579126e-01 6.71951652e-01
-5.31857610e-01 -9.06100869e-01 -3.17509323e-01 5.37153035e-02
-1.31831384e+00 -7.84120202e-01 -1.37717998e+00 -7.06345081e-01
8.96176755e-01 3.82573843e-01 8.18955898e-01 2.23095536e-01
8.43993574e-03 3.45845222e-01 -5.24123669e-01 2.36033201e-02
-3.01765651e-01 -3.93381387e-01 7.87580684e-02 8.02524149e-01
-3.01313192e-01 -9.97683883e-01 -9.02610004e-01 1.63133457e-01
-1.10832322e+00 9.20513198e-02 8.05827677e-01 1.07437503e+00
8.06534350e-01 1.52104169e-01 6.14720583e-01 -8.07130098e-01
6.23295367e-01 -4.90366280e-01 -3.30667943e-01 -5.60623705e-02
-6.26753449e-01 -1.39061302e-01 6.73737049e-01 -3.41048479e-01
-7.24085152e-01 3.39949876e-01 -2.60385960e-01 -7.41948485e-01
3.09498519e-01 1.01761043e+00 -7.43737444e-02 -3.15168381e-01
5.11397243e-01 6.50626421e-01 2.82864183e-01 -6.19918823e-01
1.45698547e-01 4.94347602e-01 5.96874833e-01 -4.72312629e-01
7.79024839e-01 6.50575459e-01 1.19619191e-01 -1.03201985e+00
-7.59470642e-01 -6.81094468e-01 -3.45029175e-01 -2.67261952e-01
5.26431322e-01 -7.70060182e-01 -4.51567650e-01 2.93111682e-01
-9.06966507e-01 -2.11985394e-01 -3.85843843e-01 8.11720371e-01
-5.75377226e-01 8.30371380e-01 -4.34482068e-01 -9.79327559e-01
-4.24238205e-01 -8.24264109e-01 7.12112844e-01 -1.45432934e-01
-2.35398449e-02 -8.39331031e-01 3.75343502e-01 1.08640067e-01
5.43039799e-01 5.35592973e-01 2.43706390e-01 -1.53118417e-01
-2.66205788e-01 -3.34182143e-01 1.55496880e-01 4.88577515e-01
2.21583426e-01 -7.99026966e-01 -6.63833022e-01 -7.82459915e-01
1.03032076e+00 4.74104695e-02 7.25547850e-01 8.41425121e-01
1.05825782e+00 -3.38647962e-01 -3.13158840e-01 9.46687579e-01
1.63069880e+00 -5.56461997e-02 8.17857146e-01 1.38643250e-01
4.37233359e-01 3.97527307e-01 4.15452063e-01 7.89585173e-01
-6.15946250e-03 6.34146333e-01 2.72468865e-01 -1.37596115e-01
1.12281501e-01 2.21858472e-02 3.14031094e-02 1.08798122e+00
-3.96349013e-01 -9.18565411e-03 -6.43212497e-01 5.02671838e-01
-2.05221438e+00 -7.89723039e-01 -2.34117970e-01 2.44397664e+00
1.00415194e+00 -1.91088781e-01 -1.33144408e-01 7.17745245e-01
6.38845444e-01 2.04534426e-01 -4.52537954e-01 2.74603069e-01
1.93229113e-02 4.01514620e-01 2.93061376e-01 4.05405909e-01
-9.30374026e-01 7.57803097e-02 5.47240782e+00 8.72060299e-01
-1.20948136e+00 4.05150801e-01 2.56506383e-01 2.36008346e-01
-1.01016596e-01 1.25593945e-01 -2.22919971e-01 5.36024213e-01
6.84808314e-01 6.46611582e-03 5.97990811e-01 4.59081024e-01
4.30346638e-01 -2.07722168e-02 -6.26956761e-01 1.36337590e+00
2.10165545e-01 -1.06020689e+00 -5.32290101e-01 2.42175739e-02
6.15422130e-01 -4.31764394e-01 -1.36430115e-01 -2.37100646e-01
-4.59223598e-01 -8.93604279e-01 4.80592489e-01 6.77933812e-01
8.97832811e-01 -5.79440951e-01 6.13875866e-01 5.75259447e-01
-9.87378240e-01 3.00333593e-02 -3.38362038e-01 3.90275903e-02
5.41091919e-01 1.46865702e+00 -4.22620773e-01 7.78042018e-01
3.94202888e-01 1.04276073e+00 1.79615632e-01 1.14727044e+00
-2.55719781e-01 9.34572935e-01 -4.21856910e-01 5.16453803e-01
-7.20660836e-02 -6.35403275e-01 8.63290310e-01 9.48857963e-01
6.24246657e-01 4.80324119e-01 1.28559858e-01 6.84823453e-01
5.56317344e-02 2.89024830e-01 -4.79030401e-01 2.38642141e-01
1.86470866e-01 1.19709814e+00 -6.45234764e-01 -2.03860983e-01
-2.77186900e-01 1.02096200e+00 -1.31850317e-01 4.41031784e-01
-4.49537396e-01 -2.10275933e-01 1.32416310e-02 3.73306483e-01
5.35221636e-01 -1.65978163e-01 -4.59526069e-02 -1.35869610e+00
3.58013064e-01 -1.07339144e+00 3.08289647e-01 -3.41057420e-01
-1.27097368e+00 5.50973117e-01 -2.30366439e-01 -1.61446786e+00
-2.11567849e-01 8.40052031e-03 -5.41406453e-01 9.18293118e-01
-1.65639317e+00 -8.02174568e-01 -2.41616338e-01 8.80944788e-01
3.54117572e-01 1.37360513e-01 7.67455220e-01 6.06503308e-01
-2.90901393e-01 2.80554712e-01 3.84359479e-01 -1.06456608e-01
5.43033600e-01 -9.41988707e-01 -4.11576688e-01 1.16169143e+00
-3.75453979e-02 7.72892475e-01 1.01917458e+00 -6.82432115e-01
-1.60257077e+00 -8.47850144e-01 6.52355790e-01 2.55531222e-01
2.17262119e-01 9.47836637e-02 -9.79395568e-01 5.34120560e-01
1.23750746e-01 3.76508087e-01 6.95141673e-01 -2.91971087e-01
3.29170525e-01 -4.23407793e-01 -1.14044535e+00 9.25870463e-02
6.32813871e-01 -2.06111550e-01 -5.85300624e-01 3.57158601e-01
-8.61490220e-02 -4.89312887e-01 -7.97201574e-01 5.47038257e-01
4.79145378e-01 -8.46659303e-01 1.08420384e+00 -2.18969844e-02
2.42208898e-01 -5.26508451e-01 -1.62332609e-01 -1.31024909e+00
-2.81786531e-01 -1.13117766e+00 -2.97055155e-01 8.92202258e-01
-1.00419007e-01 -7.53377020e-01 5.88611901e-01 1.58630818e-01
-1.69250369e-01 -7.23373592e-01 -9.77737486e-01 -6.03485703e-01
-3.92232716e-01 -2.85685390e-01 -5.64741604e-02 1.07915521e+00
7.18190596e-02 1.06338561e-01 -8.48731995e-01 2.15097263e-01
1.09153175e+00 1.63213357e-01 4.97909784e-01 -1.00621545e+00
-7.95906723e-01 8.62362012e-02 -1.30093381e-01 -1.27095044e+00
-8.39063376e-02 -7.56297469e-01 2.84503371e-01 -1.49527335e+00
8.80634487e-02 -6.57764256e-01 -3.93097758e-01 4.55147633e-03
-4.08383518e-01 5.70025206e-01 1.38992697e-01 6.57359898e-01
-2.01454163e-01 5.18684626e-01 1.26896369e+00 9.85294133e-02
-2.48940229e-01 3.90769839e-01 -6.42529905e-01 6.25082195e-01
5.74876070e-01 -7.01524436e-01 -3.74779999e-01 -2.93820560e-01
3.53851020e-01 6.18596673e-01 2.03808591e-01 -9.99805331e-01
3.64119649e-01 2.05492184e-01 2.06184357e-01 -3.20554525e-01
4.67819244e-01 -8.05195808e-01 5.82988799e-01 6.72737062e-01
-1.32160455e-01 -3.21225911e-01 -2.37624854e-01 7.23559797e-01
-4.65406984e-01 -5.41680157e-01 8.06762457e-01 -1.94004431e-01
-6.04698211e-02 8.87198746e-02 -3.03079754e-01 -9.71728042e-02
6.69533610e-01 -1.66436866e-01 5.86442232e-01 -5.37260175e-01
-8.44259858e-01 -1.78628385e-01 1.42592430e-01 -5.54906428e-01
9.60894108e-01 -1.22181165e+00 -1.07208705e+00 2.21139088e-01
-3.36294413e-01 9.70444679e-02 3.40641230e-01 1.58287835e+00
-5.25427401e-01 -5.61515167e-02 2.38448903e-01 -7.38625348e-01
-9.16159689e-01 4.36727583e-01 1.85685173e-01 -4.28118408e-01
-1.10811830e+00 7.19549954e-01 -1.01830654e-01 1.24117350e-02
-5.02419733e-02 4.39625792e-02 -2.71689713e-01 -1.95301294e-01
6.18362904e-01 3.76627743e-01 1.97376385e-01 -4.37750280e-01
-2.67834127e-01 7.91332543e-01 4.80908275e-01 -2.72249877e-01
1.60829186e+00 -2.67578125e-01 -3.62008005e-01 2.19888538e-01
1.02479243e+00 5.72779953e-01 -1.02292585e+00 -4.17772204e-01
-2.40612566e-01 -5.93378842e-01 4.05147403e-01 -2.04945058e-01
-1.25613737e+00 5.36671579e-01 5.48781097e-01 3.23651582e-01
1.58430386e+00 -4.00821984e-01 1.00429368e+00 8.60149879e-03
3.86338115e-01 -6.31206751e-01 -5.10026999e-02 -1.75649822e-01
9.62358952e-01 -9.11295295e-01 3.54804218e-01 -4.15127546e-01
-2.38043964e-01 9.77859855e-01 -3.88599366e-01 -5.40203333e-01
7.60774016e-01 8.03644806e-02 -1.75387412e-02 -1.34645060e-01
2.09803078e-02 2.27452051e-02 2.65560061e-01 1.77205458e-01
2.82586873e-01 -5.76428622e-02 -9.40734744e-01 5.36135316e-01
2.17943892e-01 4.40919787e-01 2.87150413e-01 1.07970810e+00
-2.52710134e-01 -1.14323783e+00 -7.55293369e-01 3.07257116e-01
-7.90479064e-01 -1.24458656e-01 2.84660965e-01 2.98667401e-01
-1.10951088e-01 1.00024474e+00 -5.68491161e-01 1.11788981e-01
2.05540270e-01 -1.81201845e-01 6.25761509e-01 -4.69410747e-01
-2.52295166e-01 7.17033446e-01 -2.64417440e-01 -5.18065810e-01
-9.01849687e-01 -8.69030118e-01 -9.63382125e-01 1.77272320e-01
-4.32018667e-01 4.67327923e-01 7.30595827e-01 9.72037494e-01
1.15668692e-01 4.66114014e-01 8.77640247e-01 -9.63938773e-01
-7.02971637e-01 -8.57487977e-01 -9.42622364e-01 4.45955724e-01
4.85795468e-01 -3.98232162e-01 -5.12893677e-01 2.45235592e-01] | [11.79118824005127, -2.395260810852051] |
7312f22e-f114-4de9-a880-018c290d7062 | continuously-indexed-domain-adaptation | 2007.01807 | null | https://arxiv.org/abs/2007.01807v2 | https://arxiv.org/pdf/2007.01807v2.pdf | Continuously Indexed Domain Adaptation | Existing domain adaptation focuses on transferring knowledge between domains with categorical indices (e.g., between datasets A and B). However, many tasks involve continuously indexed domains. For example, in medical applications, one often needs to transfer disease analysis and prediction across patients of different ages, where age acts as a continuous domain index. Such tasks are challenging for prior domain adaptation methods since they ignore the underlying relation among domains. In this paper, we propose the first method for continuously indexed domain adaptation. Our approach combines traditional adversarial adaptation with a novel discriminator that models the encoding-conditioned domain index distribution. Our theoretical analysis demonstrates the value of leveraging the domain index to generate invariant features across a continuous range of domains. Our empirical results show that our approach outperforms the state-of-the-art domain adaption methods on both synthetic and real-world medical datasets. | ['Hao Wang', 'Dina Katabi', 'Hao He'] | 2020-07-03 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/1417-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/1417-Paper.pdf | icml-2020-1 | ['continuously-indexed-domain-adaptation'] | ['methodology'] | [ 6.76463425e-01 7.48980232e-03 -4.38361198e-01 -4.46925193e-01
-8.09865892e-01 -6.87347770e-01 3.89945060e-01 2.20011026e-01
-3.27262431e-01 1.21299338e+00 1.49736419e-01 -1.82799697e-01
-1.26878634e-01 -9.19099748e-01 -8.48575592e-01 -5.93791127e-01
-2.53831856e-02 8.89527082e-01 1.76017776e-01 -2.67085999e-01
-1.29467234e-01 2.05860078e-01 -8.68740916e-01 2.94179499e-01
1.17602098e+00 8.32757115e-01 -2.55096406e-01 4.47838306e-01
-2.97292834e-03 3.82149100e-01 -8.26988161e-01 -4.75656211e-01
3.18572432e-01 -7.37158358e-01 -7.51134217e-01 -2.31609583e-01
2.39005625e-01 -4.47589576e-01 -3.74975443e-01 9.72486079e-01
6.33920968e-01 1.10626221e-02 1.10564208e+00 -1.25888276e+00
-9.61165607e-01 3.27091604e-01 -3.97360146e-01 8.09459984e-02
4.12447155e-01 -5.07857390e-02 5.85928679e-01 -2.83604741e-01
7.45005846e-01 9.73153055e-01 7.62167633e-01 8.17962229e-01
-1.53406358e+00 -9.18205857e-01 6.17875904e-02 -4.21387404e-02
-1.07658470e+00 -6.75992444e-02 8.98293257e-01 -6.43674076e-01
4.17370856e-01 -7.48049766e-02 2.64297456e-01 1.54141283e+00
1.61109611e-01 5.99006295e-01 1.20170808e+00 -1.85159460e-01
5.47806025e-01 1.53994441e-01 -2.68870950e-01 2.76686519e-01
1.86920449e-01 2.38265038e-01 -3.01912904e-01 -5.55282593e-01
8.68056297e-01 2.02024296e-01 -8.29572454e-02 -6.77238524e-01
-1.42664933e+00 8.57570708e-01 2.23872706e-01 1.53226890e-02
-3.77963603e-01 -2.23038375e-01 6.35336637e-01 6.40583873e-01
5.87722063e-01 3.85756105e-01 -6.68039322e-01 3.38507444e-02
-7.01124847e-01 4.37001258e-01 8.75703037e-01 1.11975849e+00
4.56984788e-01 -2.79607356e-01 -1.78415328e-01 9.58081245e-01
-1.97047889e-01 5.51696718e-01 6.54461384e-01 -8.52667391e-01
4.73462880e-01 4.06811953e-01 2.55690932e-01 -6.16787136e-01
-2.16796398e-01 -1.58711612e-01 -1.15498006e+00 -2.41678245e-02
6.81550622e-01 -2.77173787e-01 -1.00922465e+00 2.13329816e+00
4.47699934e-01 4.57131088e-01 3.23274404e-01 5.37470281e-01
5.21632254e-01 1.94764599e-01 3.09399933e-01 3.54242027e-02
1.28551745e+00 -5.07677794e-01 -6.50990605e-01 -1.60785496e-01
4.21354204e-01 -4.23720479e-01 9.98315930e-01 2.31986478e-01
-8.47811103e-01 -3.89968753e-01 -9.40670848e-01 5.76652847e-02
-4.34671879e-01 -2.39960000e-01 4.69044626e-01 5.07730484e-01
-5.12395918e-01 5.28223455e-01 -8.79264235e-01 -5.64312458e-01
6.39152348e-01 3.91492933e-01 -5.23248136e-01 -2.11168259e-01
-1.70018053e+00 6.57502353e-01 5.24045825e-01 -8.56526494e-01
-6.83946490e-01 -1.11094308e+00 -7.95247257e-01 -1.37801692e-01
1.30786374e-01 -1.12568498e+00 1.08387804e+00 -1.15713811e+00
-1.44155657e+00 1.09551597e+00 6.39750808e-02 -6.35265410e-01
7.01226532e-01 -1.34966120e-01 -6.04708850e-01 1.56327099e-01
1.20366916e-01 4.14342821e-01 8.35124791e-01 -1.05549896e+00
-5.03821731e-01 -4.25317943e-01 1.48265921e-02 1.59687951e-01
-5.62621713e-01 -1.96592584e-01 -8.27742964e-02 -9.97602880e-01
-2.83428490e-01 -8.76740217e-01 -1.56093225e-01 4.28511918e-01
-2.55806655e-01 1.10487826e-01 5.67155302e-01 -7.72080958e-01
1.05969799e+00 -2.33339572e+00 2.00280547e-01 1.29695430e-01
1.52547732e-01 1.76959246e-01 -2.74388455e-02 3.52972537e-01
-1.66987255e-01 -7.24833533e-02 -7.89465129e-01 -4.66842428e-02
-5.40372450e-03 4.85637307e-01 -3.96098733e-01 3.72379929e-01
1.94526255e-01 7.16981649e-01 -1.12226045e+00 -6.20733619e-01
-1.89698055e-01 4.06684846e-01 -8.25656831e-01 3.39810103e-01
-1.38605729e-01 8.97474706e-01 -6.21005595e-01 4.70966488e-01
7.20575154e-01 -2.83979803e-01 2.53693372e-01 1.84351787e-01
6.45116687e-01 1.43870145e-01 -7.87495315e-01 2.11259270e+00
-3.01467299e-01 9.68085155e-02 -1.22095987e-01 -1.34182775e+00
9.50999081e-01 3.79542232e-01 7.80642867e-01 -4.64511007e-01
-1.94782928e-01 2.96480983e-01 1.21946335e-02 -2.09199920e-01
3.02224513e-02 -6.81367993e-01 -5.68556905e-01 3.41020823e-01
-8.58224090e-03 -7.44948536e-02 -1.63294002e-01 -1.75721217e-02
1.31808352e+00 4.54822630e-02 6.66099608e-01 -4.31282818e-02
5.52915275e-01 9.90825221e-02 8.23134422e-01 6.43666267e-01
-4.67706859e-01 7.68758893e-01 5.87966561e-01 -3.40148151e-01
-1.13776410e+00 -1.46877074e+00 -3.49063098e-01 9.87973273e-01
2.45746560e-02 3.30072820e-01 -6.65788889e-01 -1.09841204e+00
4.75444108e-01 5.11978388e-01 -9.24969196e-01 -6.48711264e-01
-4.93464887e-01 -5.82924426e-01 8.44321907e-01 7.57073045e-01
3.47152591e-01 -8.45194161e-01 -2.67125428e-01 2.97220379e-01
-2.26902068e-01 -9.95501518e-01 -5.98836422e-01 6.40460551e-02
-1.05852282e+00 -9.92166936e-01 -1.11960053e+00 -9.38125193e-01
7.08056211e-01 -4.74351197e-01 1.40980375e+00 -4.45706964e-01
-5.40070375e-03 3.38863581e-01 -2.74188936e-01 -5.09315014e-01
-8.49274278e-01 3.32459927e-01 2.05098465e-01 -3.58107798e-02
5.30976892e-01 -8.52654934e-01 -8.02359521e-01 3.68406892e-01
-1.14068198e+00 -2.32253149e-01 4.84927267e-01 1.31927049e+00
7.11541116e-01 -2.13277191e-01 1.03454471e+00 -1.55597043e+00
7.41794705e-01 -8.68869007e-01 -2.68846840e-01 3.38365883e-01
-4.95328099e-01 1.65339351e-01 9.52020109e-01 -8.82981837e-01
-9.70612466e-01 5.16872779e-02 2.97128167e-02 -4.79973942e-01
-4.93779570e-01 3.73488605e-01 -3.19419593e-01 3.16235662e-01
9.01554644e-01 3.29829931e-01 2.83000350e-01 -4.50194150e-01
2.56411493e-01 7.04777241e-01 8.40562642e-01 -8.23899627e-01
8.11787307e-01 6.01638734e-01 -9.79037508e-02 -2.71813095e-01
-6.03112817e-01 -3.01084638e-01 -7.48194754e-01 5.25143862e-01
7.14066684e-01 -9.46653605e-01 -2.13581324e-01 5.26594520e-01
-8.06562245e-01 -5.22529125e-01 -5.85547388e-01 4.41651046e-01
-9.87264931e-01 3.13224822e-01 -4.53309625e-01 -1.05448388e-01
-3.00346971e-01 -8.19785237e-01 9.94828463e-01 -1.50164291e-01
-3.67880613e-01 -1.46211183e+00 4.42159235e-01 3.99596617e-02
4.18120086e-01 6.59353375e-01 1.11302066e+00 -1.03478479e+00
1.42426789e-02 -1.83845669e-01 -1.18634617e-02 3.49111885e-01
6.81745410e-01 -6.20791256e-01 -6.59490943e-01 -4.38341409e-01
-2.06013441e-01 -4.78104979e-01 6.61604345e-01 1.79213077e-01
1.47513795e+00 -3.42127919e-01 -4.74424958e-01 8.86889219e-01
1.22302616e+00 3.44784379e-01 5.54232240e-01 2.31087431e-01
3.61886144e-01 3.65746707e-01 7.45253205e-01 5.60724676e-01
2.65876204e-01 6.57525957e-01 1.53915130e-03 -3.35716158e-01
1.62649562e-03 -3.98029089e-01 9.92801785e-02 3.62054348e-01
1.99031964e-01 -1.06599838e-01 -9.81735170e-01 7.83089936e-01
-1.63606012e+00 -7.99963355e-01 8.46088409e-01 2.33676267e+00
1.47118318e+00 3.75075415e-02 2.90242106e-01 -1.09313421e-01
7.13030279e-01 -2.51513332e-01 -1.22067773e+00 -3.52456540e-01
6.17860854e-02 5.13437152e-01 4.94194955e-01 1.43186510e-01
-1.32471883e+00 5.32020211e-01 6.50799465e+00 6.02929115e-01
-9.35132563e-01 9.37853977e-02 5.24759412e-01 8.21837559e-02
-2.37191901e-01 -4.05683845e-01 -3.50833178e-01 7.20241904e-01
9.38716590e-01 -6.34689748e-01 3.26161414e-01 8.55502784e-01
-3.99571747e-01 4.07095373e-01 -1.45546639e+00 6.50099158e-01
-8.46658424e-02 -8.82548213e-01 7.46817961e-02 1.59466445e-01
9.26876307e-01 -3.24389398e-01 4.22467947e-01 3.62915158e-01
6.72880292e-01 -9.84708071e-01 1.03617251e-01 4.71290559e-01
1.43446660e+00 -7.66946077e-01 5.82917631e-01 3.00233573e-01
-7.99247801e-01 -2.07895190e-02 -2.08106250e-01 2.30220690e-01
2.59976462e-02 4.97091144e-01 -9.30850565e-01 6.21328771e-01
4.09438789e-01 8.87139916e-01 -1.22504339e-01 7.19971895e-01
5.89277446e-02 6.31548047e-01 -2.16827378e-01 5.52548528e-01
-2.97139704e-01 1.21199880e-02 4.25997496e-01 1.11087775e+00
4.18475598e-01 1.62535638e-01 1.44938633e-01 7.29834080e-01
-3.08148623e-01 -1.74921572e-01 -1.01799417e+00 -3.31334770e-02
6.64375126e-01 4.92870539e-01 -2.35636726e-01 -4.78164583e-01
-5.25622725e-01 1.28393638e+00 1.64519474e-01 3.36360157e-01
-9.53534305e-01 -5.77378690e-01 1.15939486e+00 1.87425360e-01
2.96236038e-01 9.85899121e-02 -4.10763294e-01 -1.16359150e+00
-5.48348352e-02 -1.11760950e+00 8.19340825e-01 -2.23924175e-01
-1.97569811e+00 4.05906230e-01 1.50912598e-01 -1.59820676e+00
-4.36423004e-01 -4.36703622e-01 -2.80993164e-01 1.02462554e+00
-1.57965887e+00 -1.14163029e+00 -1.02588177e-01 1.08896172e+00
1.57017142e-01 -3.61278236e-01 1.24322379e+00 3.57735664e-01
-1.32166654e-01 1.19479358e+00 6.83491647e-01 4.47768688e-01
1.33216500e+00 -1.39678872e+00 4.81537223e-01 3.40472281e-01
-4.23678070e-01 6.78272069e-01 4.72260207e-01 -6.98420823e-01
-8.84603679e-01 -1.35377467e+00 5.95602691e-01 -4.72761393e-01
5.32737970e-01 -3.36566061e-01 -1.26957357e+00 7.36808896e-01
-6.46252632e-02 2.22268030e-01 1.10580206e+00 1.86493620e-02
-6.07453585e-01 -2.54683971e-01 -1.71184599e+00 3.66218537e-01
1.06268394e+00 -5.48092842e-01 -7.75332212e-01 3.78397137e-01
7.34007478e-01 -5.58793962e-01 -1.48849416e+00 4.82022703e-01
5.99778771e-01 -4.90807235e-01 1.39267182e+00 -1.05001640e+00
6.75043523e-01 4.52354699e-02 -7.83418268e-02 -1.63267422e+00
-2.31122613e-01 -3.17819476e-01 -1.99391350e-01 1.06633723e+00
2.23883644e-01 -1.07740569e+00 7.66112208e-01 6.52080953e-01
3.40577602e-01 -4.23906177e-01 -9.88843560e-01 -1.02260077e+00
7.84488678e-01 1.00353993e-01 9.93631244e-01 1.30843592e+00
-5.00166640e-02 8.55296291e-03 -2.81661183e-01 1.27043486e-01
6.08862936e-01 1.04335345e-01 7.18640089e-01 -1.28896272e+00
-6.03107631e-01 -1.23646684e-01 -5.22139728e-01 -7.67391026e-01
3.47886056e-01 -8.28255594e-01 -1.57953560e-01 -1.12040830e+00
2.39526644e-01 -7.92086720e-01 -7.78282881e-01 4.94351119e-01
-4.10607725e-01 7.02641308e-02 -1.29626647e-01 1.76851228e-01
-2.72362530e-01 3.73923004e-01 1.41428792e+00 -3.34864289e-01
-1.49784684e-01 1.34893253e-01 -9.26854789e-01 4.98129129e-01
1.06042707e+00 -6.63088381e-01 -6.83981657e-01 -2.34543681e-01
-2.13550240e-01 2.14806274e-01 2.68428206e-01 -9.39246833e-01
-1.85754858e-02 -5.29934406e-01 4.47428554e-01 -1.08556300e-01
1.94334060e-01 -8.71302664e-01 1.50793567e-01 4.88913387e-01
-4.85621065e-01 -1.77150562e-01 3.36994976e-01 8.23319077e-01
-3.21348667e-01 2.99285591e-01 9.60011184e-01 -9.55150351e-02
-4.90449816e-01 4.13200110e-01 1.05636828e-01 6.67608202e-01
1.12639940e+00 -4.06740680e-02 -2.37075239e-01 -2.71164387e-01
-9.50664043e-01 1.83174685e-01 7.94586718e-01 3.19939017e-01
5.38659573e-01 -1.44742012e+00 -8.90279114e-01 4.21491742e-01
3.67961764e-01 1.79172799e-01 1.17368624e-01 4.69941288e-01
-2.65662134e-01 8.46781954e-02 -5.57353973e-01 -5.13498664e-01
-1.10986423e+00 8.68285894e-01 3.11474085e-01 -6.50083482e-01
-5.03479779e-01 6.14669144e-01 6.34487808e-01 -7.35744476e-01
1.34080544e-01 -2.08492517e-01 4.83344495e-02 -1.62694380e-01
3.66208017e-01 1.20846689e-01 -1.27856836e-01 -2.32498512e-01
-4.24642980e-01 4.80689973e-01 -2.21047387e-01 -9.72699523e-02
1.19888628e+00 1.40607759e-01 2.78102666e-01 3.93000782e-01
1.09984720e+00 -2.96190351e-01 -1.33987546e+00 -6.61382079e-01
-2.44478896e-01 -4.76095140e-01 -6.88131452e-01 -1.18111444e+00
-8.48008573e-01 9.24832225e-01 6.76985383e-01 -1.46246180e-01
1.58862031e+00 2.26962790e-02 9.49904978e-01 1.67897180e-01
4.31005031e-01 -9.28374469e-01 3.97856943e-02 3.66818070e-01
7.50252008e-01 -1.27271950e+00 -1.24682225e-02 -2.84660459e-01
-8.93573344e-01 5.87725282e-01 4.87562835e-01 -1.62431568e-01
6.71190083e-01 2.37971976e-01 1.20034978e-01 3.81616920e-01
-5.47415316e-01 6.22831173e-02 2.05885321e-01 1.20729721e+00
2.89187133e-01 3.86852294e-01 -3.11888397e-01 8.61692429e-01
-1.65845200e-01 4.42155153e-01 1.67514250e-01 1.03605545e+00
1.32910252e-01 -1.69383442e+00 -4.33874041e-01 5.42553127e-01
-7.10349441e-01 -7.86992488e-04 -3.20865601e-01 7.61507571e-01
1.51127279e-01 3.89437854e-01 9.38734114e-02 -1.98087186e-01
4.75588292e-01 3.64299268e-01 4.27457184e-01 -6.16173685e-01
-3.44614744e-01 -4.14147586e-01 -3.20155680e-01 -2.79309154e-01
-3.37532818e-01 -9.02090013e-01 -1.18816996e+00 -2.17014775e-01
3.77426356e-01 -1.20541900e-02 2.34985918e-01 5.66914439e-01
5.84571242e-01 5.48160374e-01 5.17544448e-01 -1.79753393e-01
-9.30639446e-01 -7.90759563e-01 -7.10651577e-01 1.12402892e+00
6.32898211e-01 -8.45884323e-01 1.44975595e-02 4.73677814e-01] | [10.362872123718262, 3.1112613677978516] |
084c836e-b99d-4c2e-ada7-7615ad89cf22 | wantwords-an-open-source-online-reverse | null | null | https://aclanthology.org/2020.emnlp-demos.23 | https://aclanthology.org/2020.emnlp-demos.23.pdf | WantWords: An Open-source Online Reverse Dictionary System | A reverse dictionary takes descriptions of words as input and outputs words semantically matching the input descriptions. Reverse dictionaries have great practical value such as solving the tip-of-the-tongue problem and helping new language learners. There have been some online reverse dictionary systems, but they support English reverse dictionary queries only and their performance is far from perfect. In this paper, we present a new open-source online reverse dictionary system named WantWords (https://wantwords.thunlp.org/). It not only significantly outperforms other reverse dictionary systems on English reverse dictionary performance, but also supports Chinese and English-Chinese as well as Chinese-English cross-lingual reverse dictionary queries for the first time. Moreover, it has user-friendly front-end design which can help users find the words they need quickly and easily. All the code and data are available at https://github.com/thunlp/WantWords. | ['Maosong Sun', 'Zhiyuan Liu', 'Yanhui Yang', 'Lei Zhang', 'Fanchao Qi'] | 2020-10-01 | null | null | null | emnlp-2020-11 | ['reverse-dictionary'] | ['natural-language-processing'] | [-6.37661517e-01 -5.54198086e-01 -7.08854735e-01 -2.43586734e-01
-6.03529096e-01 -1.02039170e+00 3.96048993e-01 1.84060752e-01
-7.89297938e-01 5.56464851e-01 4.28394675e-01 -7.87121654e-01
2.51888275e-01 -8.42896163e-01 -1.46134287e-01 -1.91019416e-01
7.46590137e-01 6.03294909e-01 2.46981978e-01 -1.06829381e+00
3.33850354e-01 1.17012084e-01 -1.44716334e+00 1.96080610e-01
1.21476448e+00 3.22249055e-01 7.94667065e-01 2.83876956e-01
-5.39129674e-01 4.59088027e-01 -1.81507304e-01 -5.34635901e-01
-2.16776580e-02 -2.63473421e-01 -3.96330595e-01 -7.45090187e-01
1.00304022e-01 -1.59956411e-01 -3.66396725e-01 1.25626755e+00
9.32695687e-01 5.28976656e-02 2.34281674e-01 -8.45883369e-01
-1.66977191e+00 8.91594946e-01 -1.07990347e-01 2.82548428e-01
8.77933323e-01 -6.02005832e-02 1.11467922e+00 -1.43401015e+00
5.12676120e-01 7.62473524e-01 5.33665776e-01 7.86504030e-01
-4.95424241e-01 -1.28812754e+00 1.05383964e-02 3.56259048e-01
-1.84944105e+00 -5.02316713e-01 3.08776468e-01 -1.48054421e-01
1.47446406e+00 4.02273983e-01 8.64398420e-01 1.03847587e+00
-7.45462626e-03 8.52499962e-01 1.03520310e+00 -5.56067109e-01
-3.41150373e-01 4.00272012e-01 3.70378420e-02 7.32376277e-01
2.37033650e-01 9.46478322e-02 -3.62385541e-01 2.16499314e-01
5.92295647e-01 2.26242438e-01 -4.91653413e-01 3.81055921e-01
-1.32833230e+00 7.41741717e-01 1.43291458e-01 6.99606180e-01
3.81400296e-03 -1.65597051e-01 5.02322018e-01 6.34313762e-01
1.41100690e-01 2.40703151e-01 -6.54540718e-01 -3.58703285e-01
-4.30128187e-01 6.10156022e-02 7.75972009e-01 1.84842098e+00
5.13118982e-01 -1.39830699e-02 3.30087721e-01 1.17379951e+00
3.61065894e-01 1.00475597e+00 1.18294311e+00 -1.86271787e-01
3.98987770e-01 4.80223238e-01 -3.05759221e-01 -6.67481899e-01
-2.30970457e-01 -5.88499904e-01 -5.72808087e-01 -7.24038839e-01
5.35556078e-02 1.06284410e-01 -7.23371029e-01 1.43811882e+00
2.05218554e-01 -2.09286526e-01 1.68035343e-01 7.92307496e-01
1.56835759e+00 8.31194520e-01 2.61113364e-02 -1.16746157e-01
1.66014576e+00 -1.17008424e+00 -1.04233146e+00 -3.51594955e-01
1.04700935e+00 -1.40591192e+00 1.65636063e+00 2.25188181e-01
-9.66719031e-01 -5.74092209e-01 -6.70832038e-01 -7.00728118e-01
-1.01939130e+00 2.03434587e-01 4.32173491e-01 6.68710291e-01
-9.72510278e-01 -9.11343694e-02 -5.21208882e-01 -4.05903727e-01
-2.07953349e-01 1.58689484e-01 -3.31781536e-01 -4.64690924e-01
-1.58214903e+00 9.85652983e-01 4.41306949e-01 -3.14050674e-01
-4.40211385e-01 -5.31907856e-01 -1.04642963e+00 -1.67251259e-01
2.19025835e-01 -3.80344659e-01 1.38107204e+00 -6.22995436e-01
-1.20912588e+00 1.05740738e+00 -2.87795871e-01 2.98830688e-01
1.37110114e-01 -2.02255264e-01 -1.06501341e+00 -4.68114763e-01
5.02616763e-01 3.59572172e-01 2.92869180e-01 -7.13009715e-01
-7.76436746e-01 -1.49940446e-01 5.41044623e-02 3.16963166e-01
-6.81768358e-01 2.44529679e-01 -1.08414102e+00 -1.07124138e+00
-1.08879559e-01 -9.78680313e-01 1.99665874e-02 -1.14067569e-01
-8.53786469e-02 -5.50058246e-01 4.78854060e-01 -6.89002573e-01
1.92662251e+00 -2.31718087e+00 -2.33287692e-01 -4.83138785e-02
-1.87463462e-02 5.09593308e-01 -2.34011576e-01 8.73586416e-01
-3.00439186e-02 1.84601650e-01 1.33800626e-01 -5.75488880e-02
8.87216404e-02 3.75595540e-01 -2.96999782e-01 2.08099008e-01
-6.76991880e-01 1.13383937e+00 -1.14979303e+00 -4.24317539e-01
1.74912125e-01 4.96265531e-01 -5.82345366e-01 2.04706445e-01
3.59835289e-02 -2.54517738e-02 -4.19640899e-01 8.48627090e-01
3.08345348e-01 1.51011627e-02 -7.15498172e-04 1.21090733e-01
-4.55848455e-01 6.28781796e-01 -8.75734448e-01 1.65765727e+00
-1.08144104e+00 4.10954177e-01 -1.96985468e-01 -4.34272647e-01
9.91034508e-01 5.56007206e-01 8.07665139e-02 -1.06929970e+00
1.39711052e-01 8.92341614e-01 -1.64155021e-01 -6.58226669e-01
8.48253071e-01 -2.75611073e-01 -2.01892033e-01 5.45460403e-01
3.01723871e-02 -2.48474911e-01 4.65866864e-01 3.24530192e-02
5.91930509e-01 -1.25573836e-02 7.43808448e-01 -3.77380043e-01
6.24614775e-01 2.23685279e-01 4.58717823e-01 3.33860368e-01
1.22850090e-01 3.97562504e-01 -4.36583042e-01 -1.75419360e-01
-6.63242877e-01 -9.67006981e-01 -1.08128801e-01 1.46385741e+00
2.86770940e-01 -9.23110545e-01 -2.64972538e-01 -3.82202923e-01
1.04435101e-01 1.02231932e+00 -1.47760555e-01 -2.70972222e-01
-5.61204016e-01 -1.05688073e-01 5.63689768e-01 5.64841151e-01
3.14047068e-01 -1.24720776e+00 7.58923367e-02 4.19351965e-01
-4.98985261e-01 -1.10844898e+00 -1.45172179e+00 1.01316147e-01
-5.16067505e-01 -7.97506690e-01 -8.44204247e-01 -1.60343003e+00
6.13247812e-01 7.41674781e-01 1.21307445e+00 5.57985187e-01
-1.16450489e-01 3.33946884e-01 -8.41078460e-01 -4.27330166e-01
-1.87410519e-01 4.18188989e-01 3.48849416e-01 -6.84587121e-01
1.03140748e+00 -3.61144125e-01 -4.46561426e-01 3.70360970e-01
-8.64775896e-01 -8.08202138e-04 3.55738580e-01 6.22036934e-01
6.08381867e-01 -9.66295749e-02 7.17161357e-01 -9.38379526e-01
1.15659106e+00 -6.82698965e-01 -5.92235088e-01 3.76896799e-01
-1.09127808e+00 -1.53406039e-01 1.05256176e+00 -5.92330694e-01
-5.33351719e-01 -2.71973431e-01 -1.01392627e+00 -1.32733181e-01
1.79915160e-01 9.04154062e-01 9.26952157e-03 9.58792120e-02
5.19529343e-01 7.47037113e-01 -3.50295633e-01 -7.69331634e-01
6.12556458e-01 1.25914133e+00 3.23308229e-01 -4.62074369e-01
6.70125663e-01 -9.50136483e-02 -9.77812648e-01 -7.86264777e-01
-7.70226717e-01 -8.93933654e-01 -4.82702464e-01 1.38020888e-02
6.62148774e-01 -1.28889120e+00 -5.39660513e-01 2.32468396e-01
-1.02780890e+00 -2.40157604e-01 -2.64719844e-01 4.67951655e-01
-1.32378474e-01 1.93452582e-01 -5.62782228e-01 -1.55273497e-01
-5.64763069e-01 -1.14708030e+00 7.51768112e-01 3.55992973e-01
-1.76297843e-01 -1.37890470e+00 -1.00475505e-01 1.16290152e-01
7.12657392e-01 -7.01369941e-01 8.01264822e-01 -7.20270395e-01
-5.92114367e-02 -1.78991944e-01 1.16359517e-01 2.26170734e-01
4.39213395e-01 -2.32547030e-01 -3.09499592e-01 -4.65202183e-01
-1.78991973e-01 -1.05353504e-01 4.15744305e-01 -1.71032459e-01
1.13384628e+00 -4.28913951e-01 -2.97044605e-01 7.61844456e-01
1.56041849e+00 4.61235404e-01 3.13438028e-01 4.59205091e-01
6.49577439e-01 1.48352951e-01 6.18355334e-01 3.51134002e-01
1.06827128e+00 6.09674215e-01 -4.11998108e-02 1.21707618e-02
-3.38591576e-01 -7.02312171e-01 6.16697252e-01 2.00822020e+00
3.10231686e-01 -3.16094369e-01 -1.03346300e+00 1.02902794e+00
-1.37480843e+00 -6.81257129e-01 -3.00952822e-01 1.84248197e+00
1.35777855e+00 -6.53816238e-02 -9.07769129e-02 -2.27897093e-01
4.34192419e-01 -9.57423896e-02 -3.60987544e-01 -5.47876596e-01
-1.37217194e-01 6.41716301e-01 5.30306876e-01 8.48016381e-01
-4.14414704e-01 1.53405797e+00 5.46061659e+00 1.11142838e+00
-1.42773056e+00 4.49118793e-01 -2.39737868e-01 1.07480764e-01
-8.54414821e-01 1.04791157e-01 -1.22904587e+00 4.76044416e-01
5.83749413e-01 -8.92825246e-01 6.87913537e-01 8.93543661e-01
-7.57072046e-02 4.51521218e-01 -6.40603900e-01 1.34820127e+00
1.58103839e-01 -1.16861737e+00 1.47716016e-01 -1.67506918e-01
4.89648670e-01 4.04600590e-01 1.20856829e-01 7.17754424e-01
3.69312614e-01 -8.80612016e-01 7.66178727e-01 9.97509807e-02
1.43120074e+00 -6.50486112e-01 4.51260000e-01 6.19391322e-01
-1.58000767e+00 9.14079845e-02 -4.26987588e-01 1.04456069e-02
1.24650598e-01 4.12742943e-01 -3.87379915e-01 1.02879472e-01
6.62207305e-01 1.06973720e+00 -5.99523067e-01 7.40325749e-01
-6.98067904e-01 3.67170036e-01 -2.32256502e-01 -5.62583029e-01
9.09389630e-02 -3.03689927e-01 2.04672739e-01 1.51326442e+00
8.12738001e-01 4.53603208e-01 1.98131084e-01 3.83289695e-01
-1.20344989e-01 7.51708567e-01 -7.66238630e-01 -2.73087919e-01
9.40740824e-01 1.11370623e+00 -3.19144994e-01 -4.55850035e-01
-7.48022676e-01 9.78667974e-01 2.42562592e-01 1.40544131e-01
-6.48510516e-01 -7.57321239e-01 8.81249964e-01 2.44085059e-01
1.80615023e-01 -5.77960670e-01 -3.86581011e-02 -1.56302559e+00
-4.48617525e-02 -1.29557419e+00 4.30286437e-01 -8.53115261e-01
-1.30348766e+00 8.23756814e-01 -2.74501532e-01 -1.45851743e+00
-4.96903844e-02 -7.69809067e-01 -2.48252317e-01 8.93056035e-01
-1.66464889e+00 -9.03175771e-01 -2.89547950e-01 9.54314709e-01
8.43175530e-01 -3.26464087e-01 1.23208833e+00 7.95114398e-01
-3.73774827e-01 8.96925330e-01 3.59295219e-01 2.42878944e-01
8.76613319e-01 -1.06875420e+00 6.28897369e-01 6.51809871e-01
2.55492330e-01 1.24705541e+00 3.58053893e-01 -6.94736600e-01
-1.64508498e+00 -9.29209471e-01 1.69299161e+00 -3.74279857e-01
8.38504612e-01 -4.03401077e-01 -7.37100422e-01 8.57271791e-01
5.64353049e-01 2.72612479e-02 1.14159286e+00 -2.15653293e-02
-4.68853205e-01 3.50129344e-02 -8.37377667e-01 9.04388130e-01
1.37491000e+00 -8.79042923e-01 -5.93965352e-01 6.04085922e-01
9.03013229e-01 -6.83514953e-01 -9.15458977e-01 -4.86710249e-03
6.11981571e-01 -4.89026964e-01 8.12990010e-01 -2.36601204e-01
1.88498229e-01 -3.42387080e-01 -3.36490273e-01 -1.46504354e+00
-4.20649350e-01 -4.65068460e-01 5.31278774e-02 1.10899413e+00
6.26187146e-01 -1.05599284e+00 2.32581720e-01 9.69419703e-02
-3.56865168e-01 -8.78260374e-01 -6.83815062e-01 -9.34763551e-01
2.66383022e-01 -5.61718404e-01 8.78411055e-01 1.34955728e+00
3.48109305e-01 5.29020905e-01 3.74052813e-03 -1.81453288e-01
1.45719409e-01 2.05687180e-01 5.06174028e-01 -5.57358205e-01
-1.32148281e-01 -4.90635365e-01 1.24375746e-01 -1.80615389e+00
1.58299580e-01 -1.62681544e+00 -2.98038363e-01 -1.61945856e+00
-1.71872169e-01 -9.70322669e-01 -3.62110361e-02 8.19530368e-01
-1.28296822e-01 2.09725022e-01 1.87854171e-01 2.39880398e-01
-2.74994701e-01 5.75110078e-01 1.41058445e+00 -1.49437729e-02
-8.95728096e-02 -1.82656929e-01 -1.29928446e+00 3.16113174e-01
9.96602297e-01 -7.10480392e-01 -4.74580437e-01 -1.11300302e+00
4.30481970e-01 -2.03619570e-01 -3.69922489e-01 -6.93020523e-01
5.48000395e-01 -2.95363784e-01 -1.98050141e-01 -3.09691370e-01
-9.92731098e-03 -9.78833437e-01 5.09003084e-03 6.20024800e-01
-7.98842013e-02 1.08500123e+00 2.50565767e-01 -1.52101696e-01
-3.28557879e-01 -3.99048239e-01 5.77532411e-01 -3.60454053e-01
-9.40052986e-01 3.45076293e-01 -4.21509624e-01 5.31235456e-01
9.74904776e-01 -8.25502574e-02 -3.21552217e-01 -5.34475207e-01
-4.40782070e-01 3.95279378e-01 4.18912590e-01 1.13217819e+00
8.10662508e-01 -1.55075133e+00 -7.55021334e-01 5.22732973e-01
6.84577405e-01 -3.58757496e-01 -2.38315403e-01 6.06097519e-01
-7.39585161e-01 6.72649562e-01 4.96085063e-02 -5.19936457e-02
-1.35196865e+00 6.38861656e-01 2.11368755e-01 2.03254193e-01
-6.14603043e-01 1.09024310e+00 -2.50207245e-01 -9.00965273e-01
1.74392745e-01 -3.02780360e-01 -2.61595428e-01 -2.03075092e-02
7.04236865e-01 -1.05855472e-01 4.42643417e-03 -9.27814603e-01
-4.49854672e-01 8.19666028e-01 -2.83482939e-01 -5.51725626e-02
1.19082606e+00 -2.55343527e-01 -3.58303100e-01 3.29518735e-01
1.40380859e+00 6.62354410e-01 9.37833115e-02 -6.83801889e-01
-1.82736501e-01 -6.03495717e-01 -1.07788324e-01 -7.70376623e-01
-1.15359628e+00 7.71091878e-01 4.97673154e-01 -9.17519331e-02
1.23248088e+00 -3.56203690e-03 1.43734252e+00 3.06432694e-01
6.57520175e-01 -1.14953291e+00 -2.80700654e-01 1.01429069e+00
8.12287152e-01 -1.11839104e+00 -3.42092128e-03 -2.52497613e-01
-5.15661240e-01 1.10090196e+00 5.64399660e-01 1.02177508e-01
1.08593524e+00 3.46560687e-01 6.40266001e-01 -1.90799087e-01
-7.48380661e-01 -6.09280288e-01 2.36462072e-01 5.14985740e-01
9.83411968e-01 1.98763505e-01 -9.03464258e-01 7.84564614e-01
-8.84863913e-01 -2.16794554e-02 3.52008939e-01 1.00916588e+00
-4.35493082e-01 -1.72332132e+00 3.32338326e-02 3.20542693e-01
-4.11696851e-01 -8.06443036e-01 -2.23320231e-01 8.25071216e-01
1.80268794e-01 8.94112051e-01 -2.70989448e-01 -6.19891763e-01
6.66541755e-01 -1.01906367e-01 2.33133018e-01 -1.10897887e+00
-8.59199762e-01 -1.80172279e-01 -1.12967700e-01 -2.74536848e-01
1.85896665e-01 -5.01315653e-01 -1.53161120e+00 -5.81158340e-01
-5.57062864e-01 4.32222933e-01 7.38602877e-01 6.23206317e-01
2.59713918e-01 3.43685858e-02 4.71485734e-01 2.08538044e-02
-3.37020487e-01 -8.12084496e-01 -3.38639319e-01 1.13548100e-01
1.77670673e-01 -3.74459028e-01 -1.73066929e-01 -1.19643852e-01] | [11.00733470916748, 9.867774963378906] |
c582835d-5b45-4176-967c-110ad0464db8 | eventclip-adapting-clip-for-event-based | 2306.06354 | null | https://arxiv.org/abs/2306.06354v1 | https://arxiv.org/pdf/2306.06354v1.pdf | EventCLIP: Adapting CLIP for Event-based Object Recognition | Recent advances in 2D zero-shot and few-shot recognition often leverage large pre-trained vision-language models (VLMs) such as CLIP. Due to a shortage of suitable datasets, it is currently infeasible to train such models for event camera data. Thus, leveraging existing models across modalities is an important research challenge. In this work, we propose EventCLIP, a new method that utilizes CLIP for zero-shot and few-shot recognition on event camera data. First, we demonstrate the suitability of CLIP's image embeddings for zero-shot event classification by converting raw events to 2D grid-based representations. Second, we propose a feature adapter that aggregates temporal information over event frames and refines text embeddings to better align with the visual inputs. We evaluate our work on N-Caltech, N-Cars, and N-ImageNet datasets under the few-shot learning setting, where EventCLIP achieves state-of-the-art performance. Finally, we show that the robustness of existing event-based classifiers against data variations can be further boosted by ensembling with EventCLIP. | ['Igor Gilitschenski', 'Xudong Liu', 'Ziyi Wu'] | 2023-06-10 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 1.47829205e-01 -5.02650499e-01 -2.62873739e-01 -5.81521034e-01
-8.99280667e-01 -2.46481657e-01 1.01305747e+00 1.21824756e-01
-4.53103274e-01 1.04243286e-01 3.81956071e-01 1.82572767e-01
2.34351069e-01 -5.65255523e-01 -8.88356149e-01 -3.68549585e-01
1.12839654e-01 1.20918527e-02 4.46701497e-01 8.99598897e-02
5.15614357e-03 3.92945677e-01 -1.82947552e+00 5.12700677e-01
1.21675089e-01 1.26227832e+00 1.47169292e-01 8.13762903e-01
-2.45144919e-01 1.24795628e+00 -2.20361933e-01 -3.00314069e-01
2.72006959e-01 -3.37846220e-01 -2.58355707e-01 2.19981298e-01
7.80778825e-01 -6.20922029e-01 -8.02199006e-01 7.94926524e-01
3.80003601e-01 4.30000752e-01 7.24533498e-01 -1.75660217e+00
-8.92234743e-01 2.70089805e-01 -5.49935222e-01 4.79665071e-01
2.31625870e-01 3.19584996e-01 9.15201485e-01 -1.38510799e+00
7.34970629e-01 1.27355683e+00 6.22238815e-01 5.34591258e-01
-8.72582376e-01 -6.56653881e-01 2.54237950e-01 7.30670333e-01
-1.24169922e+00 -6.19962692e-01 6.00668430e-01 -5.78486800e-01
1.42666304e+00 -2.96250254e-01 6.65542722e-01 1.67245507e+00
2.82687485e-01 9.90310550e-01 5.22719741e-01 -4.77284670e-01
4.03443843e-01 -1.87458679e-01 1.68840453e-01 5.67579508e-01
-8.87837857e-02 2.50694245e-01 -1.12198651e+00 4.68317233e-03
4.68490839e-01 8.70178938e-01 2.70749480e-02 -5.76142550e-01
-1.32930517e+00 9.23178673e-01 2.37936482e-01 8.12264010e-02
-3.55323255e-01 4.67531174e-01 5.58533609e-01 1.37988538e-01
5.14768839e-01 -2.78363395e-02 -2.85804808e-01 -3.61760646e-01
-9.83170390e-01 1.08921535e-01 6.11392140e-01 1.13876128e+00
6.30199730e-01 6.15107529e-02 -4.58421111e-01 7.17520118e-01
6.85048848e-02 4.87958997e-01 4.52598244e-01 -7.38548100e-01
3.37892979e-01 4.28015798e-01 -9.19460598e-03 -7.43640721e-01
-1.59850970e-01 2.33332589e-01 -6.07713521e-01 7.55077451e-02
1.89722344e-01 6.01306856e-02 -1.21408677e+00 1.63011050e+00
1.46684095e-01 9.07394946e-01 9.64648128e-02 8.39116454e-01
8.01132917e-01 8.43338966e-01 3.38438064e-01 -1.40484795e-01
1.35436094e+00 -1.18713343e+00 -6.57786846e-01 -3.98646832e-01
3.19512606e-01 -4.96118784e-01 1.08164191e+00 1.54255986e-01
-8.21605861e-01 -6.44290090e-01 -9.88654196e-01 -1.71917543e-01
-6.86374784e-01 1.01591893e-01 5.39883375e-01 1.35208741e-01
-6.12819970e-01 3.25348228e-01 -1.21920180e+00 -9.61461544e-01
5.00365794e-01 -2.39881888e-01 -5.80723286e-01 -3.94743711e-01
-1.00120580e+00 8.18000436e-01 4.47296381e-01 -3.36922944e-01
-1.24993098e+00 -8.93353105e-01 -1.32056987e+00 9.81559381e-02
4.26713198e-01 -2.63711125e-01 1.37436914e+00 -5.68632126e-01
-1.13459706e+00 8.19327950e-01 -2.58746564e-01 -6.23267591e-01
1.18773170e-01 -3.06475103e-01 -5.59756577e-01 4.40826595e-01
1.26623675e-01 7.77030766e-01 1.15416789e+00 -7.02586055e-01
-7.47141898e-01 -3.58837217e-01 2.79806964e-02 -4.50559426e-04
-6.01142406e-01 3.40359837e-01 -6.40655220e-01 -6.12518668e-01
-3.91924083e-01 -7.65864432e-01 -1.19978018e-01 5.04296303e-01
1.61194541e-02 -3.42694134e-01 1.09676862e+00 -1.91404238e-01
8.20061922e-01 -2.54854679e+00 -9.40686464e-02 -5.70455074e-01
3.70553099e-02 9.75211337e-02 -4.22063679e-01 5.54282963e-01
-8.50989148e-02 -3.19618642e-01 1.48518249e-01 -7.38507807e-01
2.35385895e-01 2.78659731e-01 -6.92928731e-01 4.59925175e-01
6.52538240e-01 1.07281864e+00 -9.36076045e-01 -6.54107273e-01
7.81996667e-01 5.41217744e-01 -4.76800174e-01 3.92801225e-01
-1.56776965e-01 -1.21055052e-01 -8.66584852e-02 7.49728501e-01
3.17563862e-01 -4.74262804e-01 -3.46744031e-01 -3.04492205e-01
-1.85085341e-01 -1.66277543e-01 -9.97355521e-01 2.10456538e+00
-3.67742747e-01 8.55329871e-01 -5.76813817e-01 -9.91340697e-01
6.69669390e-01 3.49383801e-01 5.75674951e-01 -6.41545951e-01
1.58872515e-01 -3.26249033e-01 -7.77834475e-01 -5.86376607e-01
6.61885023e-01 -1.12351581e-01 -3.68980467e-01 3.84183496e-01
7.39974976e-01 1.60484821e-01 2.15794414e-01 3.61137837e-01
1.29129219e+00 1.26175046e-01 4.63952452e-01 4.27059352e-01
-9.85103101e-02 2.11106930e-02 4.23213512e-01 1.04877329e+00
-5.25487781e-01 9.62826073e-01 1.33076921e-01 -4.49294746e-01
-1.01053262e+00 -1.28558481e+00 -2.65722480e-02 1.43788779e+00
2.04283595e-01 -4.40015435e-01 -3.05042773e-01 -7.08950400e-01
1.57696512e-02 9.65879440e-01 -7.23112106e-01 -4.03572291e-01
-2.16760382e-01 -5.98223209e-01 6.04175389e-01 1.02564561e+00
3.83081377e-01 -8.35635304e-01 -9.89814758e-01 2.87935495e-01
-1.99256931e-02 -1.62014425e+00 -5.29044151e-01 3.11346591e-01
-3.75364780e-01 -1.07011938e+00 -6.04934692e-01 -6.28406227e-01
1.52509838e-01 6.87519133e-01 9.26587582e-01 -5.12505412e-01
-7.34421670e-01 9.56004322e-01 -7.85999656e-01 -7.50347912e-01
1.24706089e-01 -2.67597288e-01 1.02392145e-01 1.95309147e-01
8.96442890e-01 -4.53143775e-01 -5.22805870e-01 6.58193976e-02
-9.72848177e-01 -5.40098473e-02 3.17158312e-01 7.04888880e-01
6.11595035e-01 -4.27413881e-01 5.14968395e-01 -2.84904033e-01
1.32838124e-02 -6.62325859e-01 -2.68187791e-01 3.73838454e-01
-1.31765708e-01 -2.98216224e-01 4.55672204e-01 -8.51653159e-01
-1.18446839e+00 3.00645888e-01 2.50303984e-01 -1.30033076e+00
-3.74600708e-01 2.55198091e-01 1.26499176e-01 1.44820288e-01
6.28507197e-01 1.62079021e-01 -3.49543691e-01 -3.00957471e-01
7.63635159e-01 6.67434692e-01 7.02091336e-01 -2.27063954e-01
5.88332534e-01 8.44900429e-01 -3.85249257e-01 -8.85765731e-01
-1.31648743e+00 -8.61302733e-01 -5.75166762e-01 -3.17710280e-01
1.27057683e+00 -1.42158473e+00 -2.50801951e-01 5.36880612e-01
-1.22889805e+00 -2.82429844e-01 -6.14922225e-01 7.56218672e-01
-6.91773415e-01 7.16845393e-02 -5.35716534e-01 -7.96960235e-01
-1.21641479e-01 -7.58635938e-01 1.52939832e+00 3.15369725e-01
1.76349841e-02 -6.99038327e-01 2.17236981e-01 -2.33903453e-02
2.45197281e-01 2.81128734e-01 4.96483177e-01 -8.41759741e-01
-4.83636945e-01 -4.20701951e-01 -4.67407256e-01 2.41391584e-01
-6.07086793e-02 3.03115882e-02 -1.08016384e+00 -1.95069328e-01
-2.15447377e-02 -8.95258486e-01 1.18306577e+00 2.99356759e-01
1.16479981e+00 2.47355223e-01 -3.59781593e-01 6.67980492e-01
1.45358706e+00 1.30143836e-02 6.06830537e-01 -2.45244503e-02
6.05687618e-01 2.13322937e-02 7.05827236e-01 8.03968191e-01
4.83514965e-01 5.15444398e-01 3.63943845e-01 2.32200831e-01
-3.42856616e-01 -5.28552949e-01 5.67374468e-01 5.06709337e-01
2.22798660e-01 -4.24683958e-01 -8.43407810e-01 9.58173215e-01
-2.08729982e+00 -1.31919539e+00 3.66392374e-01 1.72993457e+00
5.33365548e-01 7.31059536e-02 -1.70547515e-01 -3.45775902e-01
6.31827593e-01 6.00363493e-01 -6.84533060e-01 2.71622818e-02
-1.73382889e-02 2.49422818e-01 3.48505676e-01 -2.35708192e-01
-1.38516748e+00 9.65731323e-01 6.05261230e+00 6.96752310e-01
-1.16246557e+00 4.04815406e-01 1.43564373e-01 -6.71383321e-01
4.16398168e-01 3.61045487e-02 -1.16849864e+00 3.75057399e-01
1.12833452e+00 -2.83512175e-01 2.35055134e-01 1.10170424e+00
7.75936944e-03 1.38065182e-02 -1.50358510e+00 1.40878165e+00
8.33803117e-01 -1.47048712e+00 4.61335555e-02 -3.12057793e-01
8.45621109e-01 3.64725560e-01 -1.95550486e-01 8.35774779e-01
3.14219177e-01 -8.04284155e-01 8.54129612e-01 7.39633620e-01
9.41417754e-01 -4.31669384e-01 3.47465366e-01 2.82143623e-01
-1.49635136e+00 -1.34977013e-01 -4.72335607e-01 -1.08951651e-01
5.39056599e-01 3.54294986e-01 -6.17969096e-01 3.74388427e-01
8.44427943e-01 1.31128013e+00 -5.69337308e-01 9.82258201e-01
7.58775547e-02 5.85759580e-01 -3.35976839e-01 1.06402606e-01
1.43245459e-01 3.37936878e-01 3.93545181e-01 1.30131114e+00
5.24187922e-01 1.76709279e-01 5.32435775e-01 8.27075243e-01
-1.45705625e-01 -2.53428102e-01 -9.31020677e-01 -3.01747024e-01
3.78604919e-01 1.20819151e+00 -5.14021993e-01 -5.16639829e-01
-1.02406752e+00 1.20878029e+00 4.32230532e-01 3.71881336e-01
-1.15004265e+00 -4.67779398e-01 8.16778719e-01 -2.96150476e-01
8.26634765e-01 -3.77145290e-01 3.08617115e-01 -1.66820586e+00
-2.19835743e-01 -4.59361374e-01 5.43122709e-01 -1.04150689e+00
-1.62118959e+00 2.40399152e-01 2.29115114e-01 -1.33113575e+00
-2.16813624e-01 -6.52234077e-01 -7.34191179e-01 1.45921826e-01
-1.53101611e+00 -1.51969099e+00 -6.00120187e-01 8.97489727e-01
1.11743689e+00 -2.18474835e-01 7.74177253e-01 3.51995885e-01
-5.56588888e-01 5.10334313e-01 -9.76274982e-02 2.44059369e-01
1.03173566e+00 -8.74658644e-01 4.79393870e-01 8.73281062e-01
6.21256828e-01 1.62209198e-01 4.54073757e-01 -4.21755910e-01
-1.71114922e+00 -1.65203536e+00 7.09048390e-01 -6.01515412e-01
9.12038743e-01 -4.60509270e-01 -8.22687984e-01 9.90780473e-01
2.27418378e-01 7.18208969e-01 7.13667750e-01 -1.19996920e-01
-7.37762034e-01 -9.13020968e-02 -8.14976096e-01 4.75677073e-01
1.14648604e+00 -9.01460767e-01 -9.78835523e-01 3.13228846e-01
8.88991952e-01 -1.88387409e-01 -6.88811600e-01 2.21124977e-01
3.89062047e-01 -6.45337045e-01 9.53502595e-01 -9.94499505e-01
5.03540993e-01 -1.06899746e-01 -6.47731721e-01 -1.07358158e+00
-1.25877887e-01 -1.52289465e-01 -5.21245778e-01 1.16053939e+00
-1.39257669e-01 -1.78576827e-01 4.16593581e-01 4.23621595e-01
-8.72809812e-02 -2.05837086e-01 -9.59736049e-01 -9.71921921e-01
-2.53271341e-01 -8.71476710e-01 2.93389887e-01 8.38645041e-01
-4.95750383e-02 2.93347776e-01 -5.81443071e-01 2.32845306e-01
7.33375490e-01 5.07409871e-02 8.65296602e-01 -1.08447242e+00
-1.77984208e-01 8.01217556e-02 -7.62137175e-01 -8.93786192e-01
4.21210766e-01 -8.11181962e-01 2.53046781e-01 -1.36710083e+00
5.40276408e-01 3.18561733e-01 -5.92736900e-01 7.18882859e-01
4.83744740e-02 4.03571397e-01 4.80886489e-01 1.50598940e-02
-1.27807021e+00 8.35680425e-01 5.73006034e-01 -2.16457129e-01
1.02925353e-01 -6.82645380e-01 -1.83568358e-01 7.78836429e-01
4.24712926e-01 -4.28460926e-01 -3.56972605e-01 -3.94549519e-01
-6.86167628e-02 1.55150250e-01 9.17131543e-01 -1.23887098e+00
5.20221055e-01 -4.16824669e-01 5.16918838e-01 -8.47827613e-01
6.66481614e-01 -7.22726583e-01 -1.09411873e-01 -1.36835307e-01
-4.44284827e-01 -1.99623242e-01 2.42984027e-01 1.15675068e+00
-2.75530159e-01 -7.19868317e-02 6.41280115e-01 -1.22373134e-01
-1.21360838e+00 6.27090871e-01 -2.71136999e-01 2.44379267e-01
1.43270755e+00 8.94994009e-03 -4.52425927e-01 -2.58041471e-01
-5.33999562e-01 2.84918129e-01 3.91001999e-01 8.55763674e-01
8.84787142e-01 -1.62998748e+00 -4.86843884e-01 3.28741521e-01
8.46294463e-01 -3.73574525e-01 5.20984769e-01 7.03020871e-01
1.58805810e-02 1.93015814e-01 -1.85498223e-01 -7.96513259e-01
-1.18658590e+00 8.39650452e-01 -1.52713191e-02 8.93358216e-02
-9.80781198e-01 6.76123559e-01 2.83900112e-01 -1.11717485e-01
5.07991135e-01 -3.59808773e-01 6.92408085e-02 1.79860920e-01
1.00671732e+00 1.70063019e-01 -8.06091428e-02 -4.07309473e-01
-4.16684508e-01 4.18971509e-01 -2.12656289e-01 -1.15800411e-01
1.41289151e+00 -1.50554180e-01 5.03897130e-01 9.07936275e-01
1.27896452e+00 -7.21793294e-01 -1.67502546e+00 -4.30553406e-01
-2.05644384e-01 -4.50821459e-01 1.45388693e-01 -3.91781628e-01
-8.59511614e-01 1.26812100e+00 7.70487607e-01 -7.71465674e-02
1.01332057e+00 1.80387348e-01 9.92825925e-01 5.84458053e-01
4.31478888e-01 -1.07900202e+00 4.93215650e-01 6.98164403e-01
6.72935367e-01 -1.66342247e+00 -3.12462717e-01 1.31168738e-01
-8.50268245e-01 1.13582265e+00 5.51959574e-01 -2.33548224e-01
8.64497244e-01 2.30107039e-01 -2.33250812e-01 -1.97535485e-01
-1.25856018e+00 -4.91737604e-01 1.86640069e-01 5.20128191e-01
-1.08967565e-01 -1.73331678e-01 2.60887325e-01 8.40983152e-01
3.91417593e-01 5.01583755e-01 4.56336409e-01 1.15160620e+00
-5.27382731e-01 -5.19513249e-01 -1.83118865e-01 4.19043273e-01
-2.59951234e-01 -4.54457328e-02 -4.65850569e-02 6.60221338e-01
1.21246070e-01 9.32174563e-01 4.73237813e-01 -3.40606481e-01
4.45649058e-01 5.06568253e-01 4.75402892e-01 -7.60537863e-01
-3.63330171e-02 -2.09704757e-01 -1.72439918e-01 -8.74142349e-01
-5.03494740e-01 -7.66650558e-01 -1.07231724e+00 5.10019101e-02
-1.02440633e-01 -5.23792624e-01 5.48549354e-01 1.02430439e+00
5.81924736e-01 4.00701463e-01 4.02097732e-01 -1.04844260e+00
-6.69226944e-01 -8.55878413e-01 -6.01340771e-01 6.25806808e-01
3.81541371e-01 -9.52763259e-01 -5.04452884e-01 5.38507104e-01] | [8.898645401000977, 0.9232564568519592] |
3e5bc5a9-3fdf-43d2-ae1f-852afafa62ee | combo-pre-training-representations-of-binary | 2210.05102 | null | https://arxiv.org/abs/2210.05102v1 | https://arxiv.org/pdf/2210.05102v1.pdf | COMBO: Pre-Training Representations of Binary Code Using Contrastive Learning | Compiled software is delivered as executable binary code. Developers write source code to express the software semantics, but the compiler converts it to a binary format that the CPU can directly execute. Therefore, binary code analysis is critical to applications in reverse engineering and computer security tasks where source code is not available. However, unlike source code and natural language that contain rich semantic information, binary code is typically difficult for human engineers to understand and analyze. While existing work uses AI models to assist source code analysis, few studies have considered binary code. In this paper, we propose a COntrastive learning Model for Binary cOde Analysis, or COMBO, that incorporates source code and comment information into binary code during representation learning. Specifically, we present three components in COMBO: (1) a primary contrastive learning method for cold-start pre-training, (2) a simplex interpolation method to incorporate source code, comments, and binary code, and (3) an intermediate representation learning algorithm to provide binary code embeddings. Finally, we evaluate the effectiveness of the pre-trained representations produced by COMBO using three indicative downstream tasks relating to binary code: algorithmic functionality classification, binary code similarity, and vulnerability detection. Our experimental results show that COMBO facilitates representation learning of binary code visualized by distribution analysis, and improves the performance on all three downstream tasks by 5.45% on average compared to state-of-the-art large-scale language representation models. To the best of our knowledge, COMBO is the first language representation model that incorporates source code, binary code, and comments into contrastive code representation learning and unifies multiple tasks for binary code analysis. | ['Yu Huang', 'Kevin Leach', 'Huajie Shao', 'Scott Thomas Andersen', 'Kevin Cao', 'Yueke Zhang', 'Chen Huang', 'Yifan Zhang'] | 2022-10-11 | null | null | null | null | ['vulnerability-detection', 'computer-security'] | ['miscellaneous', 'miscellaneous'] | [ 1.28900200e-01 -2.89736420e-01 -7.73708642e-01 -2.94974864e-01
-6.81776464e-01 -8.75926971e-01 3.15429121e-01 8.30502391e-01
7.12237433e-02 2.02838052e-02 1.59440920e-01 -1.15887189e+00
3.73042405e-01 -8.80803704e-01 -6.14226162e-01 -3.85369137e-02
-1.38909951e-01 -6.64610136e-03 1.26240123e-02 -2.98242658e-01
5.89383006e-01 1.81796744e-01 -1.42712212e+00 4.23110485e-01
1.16709805e+00 7.27898240e-01 1.52023330e-01 7.69468725e-01
-6.42530680e-01 1.13721466e+00 -8.17373097e-01 -4.05600727e-01
7.97641929e-03 -1.76034525e-01 -6.14096880e-01 -5.14706075e-01
1.05719538e-02 -3.20049703e-01 -2.55247533e-01 1.27070475e+00
1.75808102e-01 -3.00813794e-01 8.44634473e-01 -1.40304303e+00
-1.37900853e+00 9.75912273e-01 -6.44615948e-01 1.22261465e-01
5.78866601e-01 1.45412579e-01 1.17508817e+00 -8.64395797e-01
3.10291857e-01 8.96502137e-01 8.82132947e-01 3.88725698e-01
-1.17931569e+00 -6.69482589e-01 -7.58872554e-02 -2.63321325e-02
-1.25071371e+00 2.15508528e-02 8.85335743e-01 -1.02223802e+00
1.38721371e+00 6.27503619e-02 3.74390215e-01 1.08381879e+00
5.78408360e-01 5.02111971e-01 6.37911856e-01 -4.03093755e-01
3.53865385e-01 3.48006964e-01 7.87570655e-01 1.03150177e+00
3.57900262e-01 1.61382854e-01 7.93106630e-02 -7.22786665e-01
3.16568129e-02 2.80257821e-01 -8.75232443e-02 -3.67498547e-01
-7.25643694e-01 1.13903415e+00 5.35167873e-01 1.43426195e-01
4.76883613e-02 4.85536665e-01 8.57104778e-01 4.93819863e-01
2.69571751e-01 5.31289518e-01 -3.83696854e-01 -5.36583841e-01
-9.52754974e-01 -3.75513434e-02 9.28009391e-01 9.40478861e-01
8.10767055e-01 3.49810094e-01 1.27829924e-01 9.03299391e-01
5.74238002e-01 5.49361467e-01 8.98179829e-01 -3.01480174e-01
7.85203397e-01 1.14116454e+00 -4.88631546e-01 -1.01886654e+00
2.22463403e-02 -2.06706718e-01 -4.63152349e-01 5.41269064e-01
-1.33053675e-01 2.30326243e-02 -7.81181753e-01 1.35358393e+00
-2.94071674e-01 -2.44173452e-01 2.91832209e-01 2.61261642e-01
9.64807510e-01 8.02149355e-01 -1.87360018e-01 3.45638126e-01
1.23931599e+00 -9.92218018e-01 -2.74281353e-01 -3.89519840e-01
1.03405237e+00 -7.17910647e-01 1.22167158e+00 1.52327389e-01
-4.76671398e-01 -4.98707443e-01 -1.34648275e+00 -1.55374765e-01
-6.98190570e-01 2.46716037e-01 7.40271866e-01 1.09035754e+00
-9.49157596e-01 3.10333669e-01 -6.42665982e-01 3.15700024e-02
5.48053443e-01 1.23416968e-01 -1.82056427e-01 -1.51508838e-01
-8.57501030e-01 6.36723399e-01 3.44024599e-01 -4.66736823e-01
-1.14594841e+00 -1.00526035e+00 -1.48371112e+00 3.57092947e-01
6.77658767e-02 -1.56323418e-01 1.11391807e+00 -1.06479979e+00
-1.17674208e+00 6.13214970e-01 -5.79981357e-02 -3.80035967e-01
-2.37639427e-01 4.65747435e-03 -4.94003475e-01 -1.71593085e-01
3.94564420e-02 1.25228956e-01 1.03714311e+00 -1.09506333e+00
-1.27349004e-01 6.03725649e-02 2.87387073e-01 -4.71081913e-01
-6.92460835e-01 2.77229995e-01 -4.02697809e-02 -9.79620636e-01
-6.29523993e-01 -6.69363618e-01 -1.09736966e-02 1.76809907e-01
-8.23917463e-02 -2.45067418e-01 9.48268175e-01 -9.50861812e-01
1.71716404e+00 -2.54626513e+00 1.19725890e-01 2.90386200e-01
4.66888696e-01 2.64079452e-01 -2.62409061e-01 3.05571437e-01
-5.10549307e-01 5.98180294e-01 -6.55898273e-01 -3.64980916e-03
2.59498000e-01 -1.23818763e-01 -8.56785417e-01 4.25515473e-01
3.49423915e-01 1.10919058e+00 -9.18975830e-01 -2.64074534e-01
-1.12215713e-01 3.36214781e-01 -9.92781103e-01 2.50814438e-01
-3.58938277e-01 -4.78313178e-01 -2.94188380e-01 1.03580558e+00
4.66929615e-01 -3.00979763e-01 5.99915944e-02 1.02878653e-01
-4.59128544e-02 5.37737429e-01 -6.36277199e-01 1.49484015e+00
-9.41572070e-01 9.24350977e-01 -2.74854153e-01 -9.26127195e-01
1.16297853e+00 5.62382601e-02 -4.00718972e-02 -4.11396295e-01
2.36216277e-01 3.41529250e-01 1.52466461e-01 -4.22679245e-01
5.68569124e-01 3.44442219e-01 -7.16953695e-01 9.35637236e-01
-7.33789131e-02 -6.25887394e-01 2.26829693e-01 4.43495125e-01
1.38963139e+00 1.25005797e-01 6.41642690e-01 -4.08143289e-02
5.61023533e-01 6.47400171e-02 1.60624921e-01 5.16878903e-01
4.96993326e-02 3.67898643e-01 9.25960660e-01 -2.45133027e-01
-9.80606616e-01 -1.01018322e+00 1.16088393e-03 1.06867337e+00
-4.59353775e-02 -1.10566425e+00 -6.29955351e-01 -1.03422999e+00
1.74829826e-01 8.02259803e-01 -6.42106891e-01 -6.21359050e-01
-5.96520007e-01 -5.66257656e-01 9.38235462e-01 9.86589253e-01
5.69779985e-02 -9.14525628e-01 -4.67502058e-01 1.02194563e-01
1.39330477e-01 -2.95851737e-01 -8.03835034e-01 4.04717684e-01
-7.16879606e-01 -1.17460763e+00 -2.44536981e-01 -8.52251649e-01
7.31503129e-01 -6.55635202e-04 1.07414973e+00 6.05049253e-01
-5.78907609e-01 2.31098935e-01 -6.30187690e-01 -1.21508330e-01
-9.93781626e-01 5.15065566e-02 -3.27059627e-01 -5.15865624e-01
3.07834834e-01 -4.67660397e-01 7.16392174e-02 -1.11066274e-01
-9.70022321e-01 -3.96619022e-01 6.06431782e-01 9.27231967e-01
2.63594836e-02 8.16637576e-02 1.12005360e-01 -7.25274444e-01
8.29215467e-01 -9.48442101e-01 -7.18886495e-01 2.92066664e-01
-6.33388996e-01 3.89206499e-01 1.20337784e+00 -4.33960974e-01
-7.54340351e-01 -1.17677703e-01 -3.41805488e-01 -2.59996533e-01
2.71676511e-01 9.74563897e-01 8.62083212e-02 -1.96256191e-01
1.08777618e+00 3.54345828e-01 2.04225063e-01 -3.98741186e-01
4.01796937e-01 1.16428483e+00 3.13221842e-01 -8.77673626e-01
1.16113544e+00 -3.10012996e-02 -4.90817636e-01 -4.07121867e-01
-1.83805779e-01 -2.13409349e-01 -3.19565654e-01 3.21467489e-01
5.95132589e-01 -7.39373803e-01 -2.90036649e-01 2.61010349e-01
-1.13504899e+00 -4.05377388e-01 -1.10266507e-01 -9.71066132e-02
-2.89538294e-01 7.52265394e-01 -6.36396885e-01 -5.06245017e-01
-3.17699552e-01 -1.57155800e+00 8.59503150e-01 -1.56948697e-02
-4.94567931e-01 -1.00609028e+00 1.32188931e-01 1.07326232e-01
5.12858987e-01 1.49577290e-01 1.63032389e+00 -6.27668142e-01
-3.13826233e-01 -1.49959490e-01 -3.58524770e-01 4.28910702e-01
3.48626435e-01 3.43271464e-01 -6.94338977e-01 -3.48520160e-01
-1.52148351e-01 -3.55901271e-01 7.34295905e-01 -2.97669291e-01
1.50672781e+00 -3.91668886e-01 -3.94556224e-01 8.54331076e-01
1.47184110e+00 4.55555856e-01 5.70211112e-01 3.29577625e-01
7.67037034e-01 1.27126202e-01 3.42969060e-01 4.34790909e-01
3.36035281e-01 4.66757298e-01 6.84440017e-01 1.49600014e-01
-1.95655644e-01 -2.80117512e-01 8.45077336e-01 9.31372881e-01
4.58279431e-01 -5.39401080e-04 -1.27987373e+00 5.95665455e-01
-1.43288696e+00 -8.09728980e-01 2.34407466e-02 1.90450323e+00
1.07982492e+00 1.01441450e-01 7.06774532e-04 2.36523733e-01
5.05078971e-01 8.02421793e-02 -5.31748772e-01 -9.20423269e-01
5.12501478e-01 4.19525981e-01 3.94408524e-01 3.56081843e-01
-1.01045942e+00 5.87456346e-01 5.83671570e+00 8.67548466e-01
-1.15610790e+00 1.87846735e-01 2.91424751e-01 4.07617480e-01
-8.39601099e-01 2.20319301e-01 -5.84979653e-01 8.32329571e-01
1.22386777e+00 -5.73311090e-01 6.31450236e-01 1.63210189e+00
-5.08067369e-01 3.23931873e-01 -1.36816704e+00 8.88028860e-01
4.03250724e-01 -1.39409351e+00 -3.65504138e-02 -4.13491391e-02
5.55383682e-01 -3.70877571e-02 2.12477297e-01 7.80534565e-01
4.01874989e-01 -1.21311951e+00 9.78914738e-01 -3.76254995e-03
1.11138189e+00 -8.07594538e-01 8.06959093e-01 -2.16974225e-02
-1.58496904e+00 -7.31542051e-01 -1.84305698e-01 -2.39039902e-02
-3.01045775e-01 3.74765664e-01 -7.23824143e-01 4.58179951e-01
5.19663572e-01 1.10640204e+00 -1.22355533e+00 8.36069942e-01
-3.74312133e-01 7.00062513e-01 2.28008896e-01 -3.14271092e-01
1.42417699e-01 2.64036596e-01 1.90941811e-01 1.38747454e+00
4.28812563e-01 -5.15838504e-01 2.54256994e-01 1.28075290e+00
-2.41527855e-01 -1.07412718e-01 -7.67219126e-01 -7.98630893e-01
5.55744886e-01 8.81928802e-01 -6.04462504e-01 -3.82380605e-01
-7.62045920e-01 7.40010142e-01 4.06678945e-01 1.80913895e-01
-1.10422063e+00 -1.27576005e+00 9.30376470e-01 -1.02463171e-01
3.87471318e-01 -3.08421612e-01 -5.41003048e-01 -1.19791186e+00
1.33698016e-01 -1.20889091e+00 2.20410436e-01 -6.08508646e-01
-1.13506639e+00 7.23666310e-01 9.44419357e-04 -1.26415503e+00
-5.08779883e-01 -8.96526754e-01 -7.52384901e-01 8.02301228e-01
-1.52548039e+00 -8.40753198e-01 -2.30648756e-01 1.15360908e-01
5.04311502e-01 -7.29255378e-01 9.03405964e-01 3.90431345e-01
-5.39046943e-01 9.90995586e-01 2.32070744e-01 5.29900849e-01
3.01525921e-01 -1.30747199e+00 8.75967383e-01 1.02255392e+00
-9.17942598e-02 1.33139503e+00 2.25384668e-01 -7.65828848e-01
-1.64177299e+00 -1.27212179e+00 5.91624618e-01 -6.09958410e-01
1.07555890e+00 -6.87893093e-01 -1.06147027e+00 7.41881013e-01
-8.08096603e-02 2.56008893e-01 9.38207448e-01 -1.54783472e-01
-1.15738857e+00 5.34815602e-02 -1.09632301e+00 4.31878120e-01
5.94746709e-01 -1.05856836e+00 -8.96812856e-01 1.60786435e-01
1.02897799e+00 -2.57822216e-01 -8.15024078e-01 -2.32111476e-02
4.95035022e-01 -5.53530514e-01 1.04158068e+00 -2.88337618e-01
9.38446581e-01 -5.09476662e-01 -3.18749160e-01 -1.16693747e+00
-1.99229687e-01 -3.15296978e-01 -4.91498530e-01 1.38667953e+00
4.49772328e-01 -6.61056042e-01 4.17258054e-01 2.34554455e-01
-2.57897615e-01 -7.35429227e-01 -4.93893713e-01 -7.90361047e-01
3.96659911e-01 -5.11643767e-01 9.22061741e-01 9.35483456e-01
3.01951647e-01 -3.68771306e-03 7.54150078e-02 -1.69738289e-02
2.40005285e-01 5.56794643e-01 5.63027680e-01 -1.14533043e+00
-7.02709615e-01 -7.89132297e-01 -5.26094675e-01 -9.01190758e-01
7.28855193e-01 -1.50266504e+00 3.79739255e-02 -1.30883551e+00
2.55246431e-01 -5.18561065e-01 7.77036790e-03 8.13933790e-01
-1.82022035e-01 3.10656279e-02 8.42894018e-02 1.15548544e-01
6.87844977e-02 3.74684960e-01 1.60921574e-01 -8.53643775e-01
2.84738187e-02 -2.74256110e-01 -1.04419720e+00 4.52310830e-01
7.90335894e-01 -7.34211385e-01 -5.09890437e-01 -4.57470149e-01
4.66258943e-01 -1.59876943e-01 3.25919449e-01 -7.65558779e-01
-4.55279723e-02 -1.01496913e-01 3.22384350e-02 -1.91820413e-01
-9.16123912e-02 -6.79294944e-01 -2.41639197e-01 7.30895698e-01
-2.14397311e-01 3.69113952e-01 4.18030560e-01 4.46447760e-01
-1.38498902e-01 -9.11625326e-01 6.96783900e-01 -2.34839302e-02
-9.75440741e-01 -7.23849796e-03 -7.76151538e-01 7.02734292e-02
1.00517893e+00 -2.07782239e-01 -7.01386690e-01 8.88730064e-02
8.18040688e-03 -1.05233394e-01 9.15562868e-01 7.64824986e-01
8.76333237e-01 -1.26904154e+00 -3.57261300e-01 5.30276775e-01
5.10662973e-01 -6.38051927e-01 -2.95525879e-01 3.50300282e-01
-8.38867188e-01 2.01231018e-01 -5.83640076e-02 -2.66088992e-01
-1.32698536e+00 9.57625329e-01 9.87668335e-03 -1.51466861e-01
-5.54540396e-01 7.34942436e-01 -5.09051643e-02 -8.04416895e-01
1.36321783e-01 -7.63885915e-01 -1.70824006e-01 -9.68729183e-02
7.99899280e-01 1.37592271e-01 9.98029634e-02 -3.92701954e-01
-6.42778158e-01 5.85726023e-01 -1.98136836e-01 4.28893954e-01
1.27452540e+00 5.30136645e-01 -5.72117269e-01 3.24810624e-01
1.62977707e+00 2.29088500e-01 -7.48131454e-01 1.53267533e-01
1.93205729e-01 -5.34595370e-01 -9.52528194e-02 -7.48335600e-01
-8.54982495e-01 1.10278082e+00 4.48948175e-01 1.36249870e-01
9.68803823e-01 -4.66889516e-02 7.12316990e-01 5.59817672e-01
3.66954863e-01 -6.17197216e-01 4.09929454e-01 7.04977393e-01
7.14516997e-01 -1.04282546e+00 5.44439396e-03 -2.40180001e-01
-3.25977176e-01 1.45085990e+00 6.86786890e-01 -1.11340784e-01
5.92514873e-01 8.76173437e-01 -1.33708149e-01 -3.96826603e-02
-6.28803849e-01 2.58605599e-01 2.65923053e-01 8.08916330e-01
6.97255373e-01 -1.67898044e-01 -5.31865433e-02 7.54029036e-01
-2.66176999e-01 -2.38703921e-01 9.41559553e-01 1.44367719e+00
-3.03408176e-01 -1.45719802e+00 -4.13381547e-01 7.18023062e-01
-3.20082456e-01 -6.53502345e-01 -4.52444553e-01 4.26284701e-01
-4.28782552e-02 8.31121266e-01 -3.53859812e-02 -6.05085909e-01
-6.74751867e-03 4.15125974e-02 7.93896839e-02 -1.30254614e+00
-9.16595638e-01 -8.30850661e-01 -2.22262636e-01 -3.81929845e-01
3.13847184e-01 -2.48699263e-01 -1.41062951e+00 -2.11586535e-01
-1.56555057e-01 2.50311404e-01 7.94360161e-01 5.72701931e-01
4.61846471e-01 9.38382447e-01 5.38073242e-01 -8.21100593e-01
-6.50485575e-01 -5.37453651e-01 -4.20639589e-02 1.90606061e-02
6.69854164e-01 -6.46217346e-01 -5.33924043e-01 1.67579308e-01] | [7.235850811004639, 7.818892002105713] |
897c1fbe-cf38-4a28-9ca8-cb29527d2e74 | crack-segmentation-for-low-resolution-images | null | null | https://ieeexplore.ieee.org/abstract/document/9511400 | http://www.mva-org.jp/Proceedings/2021/papers/O1-1-2.pdf | Crack Segmentation for Low-Resolution Images using Joint Learning with Super-Resolution | This paper proposes a method for crack segmentation on low-resolution images. Detailed cracks on their high-resolution images are estimated by super resolution from the low-resolution images. Our proposed method optimizes super-resolution images for the crack segmentation. For this method, we propose the Boundary Combo loss to express the local details of the crack. Experimental results demonstrate that our method outperforms the combinations of other previous approaches. | ['Norimichi Ukita', 'Yuki Kondo'] | 2021-07-25 | null | null | null | international-conference-on-machine-vision | ['crack-segmentation'] | ['computer-vision'] | [ 2.58956850e-01 -3.35944779e-02 1.22088991e-01 -1.51428431e-01
-1.34495854e+00 2.09099591e-01 -1.61738023e-01 -4.68814403e-01
-2.40606233e-01 6.58902049e-01 2.20383555e-02 6.64116800e-01
3.13689381e-01 -9.98448431e-01 -2.87336946e-01 -8.07335258e-01
2.66472369e-01 3.58051300e-01 1.11894238e+00 -1.25059396e-01
8.91847134e-01 4.98919636e-01 -1.41736925e+00 6.28817499e-01
8.96209419e-01 6.64611816e-01 6.24055207e-01 8.23082685e-01
1.30335256e-01 7.79720843e-02 -1.39098227e-01 2.87588358e-01
-1.76797420e-01 -1.33773416e-01 -8.62937212e-01 2.75314301e-02
5.12530446e-01 -5.77437878e-01 -1.39800668e-01 1.21497691e+00
3.50894630e-01 -2.43977934e-01 6.77281618e-01 -1.92107409e-02
-5.46080291e-01 6.81198299e-01 -1.13045609e+00 6.34991944e-01
7.27138042e-01 -4.59495902e-01 8.83683860e-01 -1.60429096e+00
7.93302536e-01 1.34455276e+00 9.62552190e-01 4.71657991e-01
-1.23569000e+00 -5.20981133e-01 -9.94136743e-03 9.86829400e-02
-1.56861043e+00 -1.50229782e-01 1.23861933e+00 -4.22764629e-01
7.17842460e-01 -1.53917596e-01 -3.22874486e-02 5.62111080e-01
1.18145928e-01 1.46121591e-01 1.28754854e+00 -4.71966892e-01
5.99865394e-04 -4.63222772e-01 5.86514950e-01 1.09559500e+00
3.77451271e-01 1.93739712e-01 -5.38069248e-01 -1.35456905e-01
1.75754762e+00 -2.26613000e-01 -3.04320186e-01 3.87163609e-01
-8.38998258e-01 5.72741210e-01 4.05164719e-01 6.26139700e-01
-8.68113264e-02 -6.14880025e-02 -1.93061993e-01 -2.02333495e-01
7.31248021e-01 2.16553777e-01 3.18963647e-01 4.24268484e-01
-1.01706374e+00 3.10667008e-02 5.10032102e-02 6.20226085e-01
1.00844538e+00 -4.34928574e-03 5.22715330e-01 9.45292354e-01
4.63746935e-01 1.64258882e-01 1.78731367e-01 -1.53041983e+00
3.25307995e-01 3.18480194e-01 1.47411570e-01 -8.51624370e-01
-5.00456154e-01 2.09951669e-01 -9.44847643e-01 6.15879297e-01
3.56744587e-01 9.38710347e-02 -1.05524766e+00 4.91105765e-01
3.66057396e-01 4.21647877e-01 -8.59319642e-02 1.00700092e+00
8.52196991e-01 5.90649724e-01 -5.21378994e-01 -4.44481254e-01
1.08388376e+00 -8.72864842e-01 -1.15327942e+00 -7.21937791e-02
-3.63787919e-01 -7.59488583e-01 1.15755105e+00 5.13339162e-01
-1.30377269e+00 -4.60855126e-01 -1.37844980e+00 -2.12398887e-01
2.26368353e-01 -1.83938514e-03 -3.09537929e-02 2.39834607e-01
-7.54498780e-01 1.07414949e+00 -9.56171870e-01 1.99741632e-01
4.17623162e-01 1.32867610e-02 -1.69774979e-01 -1.56571060e-01
-1.08285737e+00 3.99047524e-01 3.19243580e-01 1.83099017e-01
-6.75965965e-01 -3.30312788e-01 -7.81520844e-01 -3.02345246e-01
9.50580686e-02 -1.53401509e-01 7.94864476e-01 2.55209863e-01
-1.47857428e+00 8.21890593e-01 -3.61423671e-01 1.06716745e-01
6.99536363e-03 -2.89768368e-01 -5.47487512e-02 9.56880748e-01
1.82318568e-01 -7.96593800e-02 1.20694590e+00 -1.99178553e+00
-6.85315847e-01 -3.23561579e-01 -2.23144218e-01 -1.24599330e-01
3.24856013e-01 1.39320463e-01 -2.48464271e-01 -6.21364474e-01
9.50411558e-01 -9.80842710e-02 -6.36847794e-01 -3.37353885e-01
-5.96667945e-01 8.63483101e-02 1.29274881e+00 -8.87617111e-01
1.36918724e+00 -2.04652572e+00 8.64126310e-02 1.57518029e-01
3.06367069e-01 -2.23343298e-01 2.77793795e-01 1.61611978e-02
1.73301682e-01 4.30484086e-01 -1.06330967e+00 -5.18683612e-01
-7.90515900e-01 2.10055038e-01 4.87628207e-02 5.61431527e-01
6.20724708e-02 3.89553934e-01 -4.19830143e-01 -1.23446488e+00
-1.62908003e-01 6.72347903e-01 -2.49036655e-01 1.67061776e-01
2.04117000e-01 8.07888567e-01 -6.65769458e-01 9.54618454e-01
1.15794051e+00 -2.32976437e-01 -4.72587228e-01 -5.10731816e-01
-6.37850463e-01 -3.24770540e-01 -1.35346782e+00 1.57287312e+00
-1.06240474e-01 -1.60721000e-02 7.81036019e-01 -5.87628126e-01
8.44754159e-01 5.24583995e-01 4.20212060e-01 -4.16032970e-01
-5.25268950e-02 4.29524571e-01 -8.28884482e-01 -6.48573995e-01
4.20979589e-01 -5.66893280e-01 2.35638414e-02 3.93682927e-01
-6.97321415e-01 -6.26896739e-01 8.41209367e-02 -9.29643679e-03
5.35902500e-01 1.20642290e-01 -3.27263594e-01 -3.09870899e-01
6.65911436e-01 -4.72941287e-02 6.43309116e-01 3.67429405e-01
6.73671998e-03 1.39676178e+00 9.27130505e-02 -3.26290935e-01
-1.42849851e+00 -1.14703989e+00 -7.11997926e-01 2.30912238e-01
6.57034814e-01 1.90290902e-02 -1.20752811e+00 -3.06910962e-01
-3.32525045e-01 -1.03041954e-01 -5.45856237e-01 4.42902654e-01
-1.34146082e+00 -9.92303312e-01 4.16997105e-01 7.31931508e-01
6.70107961e-01 -7.27159977e-01 -4.74797159e-01 5.32494336e-02
-5.34685791e-01 -1.38765073e+00 -5.41229725e-01 -3.52207214e-01
-1.34969676e+00 -1.05167782e+00 -9.22218859e-01 -1.06538188e+00
8.99086654e-01 -2.15225411e-03 7.55581379e-01 4.57265317e-01
-6.61330283e-01 1.70247376e-01 -2.95810014e-01 5.62694073e-01
-2.44202614e-01 -2.47022033e-01 -1.39635444e-01 5.71376979e-02
-5.69291234e-01 -6.15677178e-01 -6.56612754e-01 4.23563331e-01
-7.25510299e-01 4.38162386e-02 4.90706205e-01 9.03940082e-01
1.45766902e+00 7.51083076e-01 3.31963778e-01 -1.02812791e+00
6.25071108e-01 -2.62024570e-02 -5.05960822e-01 -1.93229541e-01
-1.01292217e+00 1.53130457e-01 2.71638691e-01 -1.53725594e-01
-1.77794862e+00 1.23203687e-01 -3.83532763e-01 -2.62252301e-01
-1.60168290e-01 3.50301802e-01 4.56639715e-02 -3.21204454e-01
4.92733896e-01 2.37185471e-02 -2.46352628e-01 -1.04423499e+00
2.19089212e-03 6.26037419e-01 8.88082206e-01 -8.41745079e-01
1.03040087e+00 1.22085965e+00 -1.84137654e-02 -1.26179266e+00
-1.05445421e+00 -5.54005384e-01 -1.34698653e+00 -2.21195832e-01
1.30928493e+00 -6.64640307e-01 -4.10617381e-01 7.59115338e-01
-1.21291995e+00 -1.82701945e-01 -1.64176941e-01 5.46980083e-01
-4.29631382e-01 1.01915550e+00 -1.47633338e+00 -7.61542618e-01
-1.54506952e-01 -1.18256664e+00 1.30374491e+00 5.51310658e-01
4.76852864e-01 -7.17317462e-01 2.61824667e-01 7.38266647e-01
2.11760700e-01 5.05678654e-01 6.99105740e-01 5.98283172e-01
-9.95806575e-01 2.43934058e-02 -6.70726657e-01 2.07394868e-01
5.12054935e-02 3.52187306e-01 -9.50171530e-01 8.04796722e-03
2.81291634e-01 -1.95908338e-01 1.09888744e+00 6.88306987e-01
8.32604527e-01 1.02080263e-01 -5.28199613e-01 6.52412057e-01
1.98071873e+00 -2.25615680e-01 7.22263873e-01 1.46148607e-01
1.06306028e+00 3.42147559e-01 1.03952575e+00 1.73182309e-01
8.96667875e-03 7.20160604e-01 4.84328866e-01 -1.50913924e-01
-2.47400418e-01 2.61688739e-01 -4.47212979e-02 8.38640332e-01
-9.41960752e-01 6.79656267e-01 -8.35001290e-01 5.75774431e-01
-1.54258275e+00 -1.09649563e+00 -9.50005472e-01 1.64491975e+00
9.15914178e-01 2.34266400e-01 -2.68709332e-01 2.43070915e-01
1.10702515e+00 2.68085808e-01 -5.44161379e-01 -1.00907952e-01
-3.11773151e-01 2.07191437e-01 3.11704814e-01 1.33673334e+00
-1.18696201e+00 1.06547379e+00 7.92230463e+00 9.02689099e-01
-5.86380839e-01 4.34791297e-01 3.20636749e-01 5.73000968e-01
-4.25315768e-01 -8.32783803e-02 -1.11997771e+00 1.30025432e-01
4.87099558e-01 1.04406528e-01 1.59282312e-01 3.93576294e-01
1.39738008e-01 -3.63034785e-01 -3.88216287e-01 7.68154502e-01
3.91247740e-04 -1.27141821e+00 -2.56705433e-01 -1.14104383e-01
9.11236823e-01 -2.51722515e-01 5.70095098e-03 -5.43839872e-01
4.05681878e-02 -1.09406805e+00 4.78653133e-01 8.49332809e-01
1.33850980e+00 -8.21101606e-01 5.07366538e-01 2.71944314e-01
-1.79706252e+00 -7.61451572e-02 -5.84587634e-01 4.46965592e-03
8.31846535e-01 9.33504105e-01 -1.18895277e-01 5.22279620e-01
1.01394510e+00 8.88243914e-01 -2.19463006e-01 4.30352300e-01
-2.82984525e-01 5.01577795e-01 -3.72422576e-01 9.68214512e-01
-1.36391714e-01 -3.56527179e-01 5.61455309e-01 1.23053944e+00
2.14135259e-01 7.72281766e-01 2.79758900e-01 1.15728843e+00
1.45958200e-01 -1.27404481e-01 -3.71848047e-01 4.88833845e-01
3.60100329e-01 1.04550850e+00 -9.55655396e-01 -6.97872043e-02
-3.33828121e-01 8.06256235e-01 1.18883029e-01 2.55687922e-01
-5.70141435e-01 -2.46207014e-01 -1.61286048e-03 3.93606901e-01
2.89522499e-01 -4.08422470e-01 -9.98949528e-01 -1.04999053e+00
-5.95437586e-02 -2.48508051e-01 4.28928018e-01 -6.91046417e-01
-1.22379279e+00 5.21301866e-01 2.07539126e-01 -8.19925785e-01
3.36173952e-01 -3.03225964e-01 -5.69240391e-01 8.48881245e-01
-1.75583029e+00 -1.22523999e+00 -2.08458304e-01 4.92161483e-01
8.63904297e-01 6.79068327e-01 3.91803205e-01 1.60537094e-01
-6.99369133e-01 9.93005186e-02 -8.96092132e-02 3.01900983e-01
4.63004619e-01 -1.18825126e+00 2.48835832e-01 1.11448431e+00
-4.66302872e-01 1.97336376e-01 6.98012233e-01 -1.18721604e+00
-5.05424261e-01 -6.26510203e-01 5.00124156e-01 -3.57482642e-01
2.36042798e-01 1.96318030e-01 -1.71565068e+00 3.78949642e-01
6.57534674e-02 1.17252007e-01 6.55465275e-02 -3.52645427e-01
-1.48847058e-01 4.17237282e-01 -1.61162603e+00 -4.96291891e-02
7.73105621e-01 -6.59110963e-01 -6.73272014e-01 -5.36633730e-02
6.42973840e-01 -6.25331998e-01 -1.52523911e+00 7.67603934e-01
5.14042020e-01 -1.17732179e+00 1.44822097e+00 4.35479939e-01
7.24253416e-01 -4.18508202e-01 -1.98781211e-03 -7.59006619e-01
-5.93845129e-01 -1.84462637e-01 -1.20771870e-01 8.91209185e-01
4.12175417e-01 -4.09296036e-01 1.01364625e+00 4.62893993e-01
-5.01349509e-01 -8.05521369e-01 -1.13746238e+00 -2.43426234e-01
3.16232771e-01 1.03356376e-01 3.18680346e-01 8.61473978e-01
-7.14380592e-02 1.08448423e-01 1.28768086e-02 1.02896297e+00
1.52640629e+00 4.39215481e-01 -2.43261933e-01 -1.55897808e+00
2.27273151e-01 -2.88694322e-01 2.36234054e-01 -8.22353840e-01
-6.79148957e-02 -1.38107985e-01 3.08530301e-01 -1.69435215e+00
4.21124697e-01 -4.83481050e-01 -4.54744324e-02 -1.70000121e-01
-4.09701496e-01 7.11360931e-01 -5.98946273e-01 5.79895020e-01
-8.03887993e-02 1.78462550e-01 1.82050836e+00 5.08645996e-02
-4.52670828e-02 -8.59645531e-02 -1.18018128e-01 1.24423003e+00
6.53505802e-01 -5.41250527e-01 2.42418990e-01 -1.68122157e-01
1.39716059e-01 4.15736616e-01 2.15359241e-01 -1.01672566e+00
2.99156427e-01 -2.26021588e-01 1.05468251e-01 -1.07844555e+00
5.52431047e-01 -7.04590499e-01 -1.68002993e-01 4.08681154e-01
-1.38688996e-01 -4.22269642e-01 -3.47303033e-01 7.73068845e-01
-4.56841618e-01 -6.91260397e-01 1.64918375e+00 -5.55165410e-01
-4.61751759e-01 1.82174549e-01 -3.06335688e-01 1.09221235e-01
6.87559664e-01 -5.26231766e-01 -4.21030134e-01 1.11323811e-01
-1.25563073e+00 -4.98427264e-02 9.71860588e-01 -2.96012461e-01
1.53755665e+00 -9.56847072e-01 -9.22040641e-01 2.23926753e-01
-2.96659976e-01 7.50266790e-01 6.15275145e-01 6.94274247e-01
-8.79756391e-01 -1.03743985e-01 -1.34516120e-01 -6.63199127e-01
-1.37443960e+00 2.95118183e-01 5.11590481e-01 -5.26640892e-01
-1.30280042e+00 9.92904663e-01 2.65406333e-02 -8.67740903e-03
-3.20080519e-01 -3.53577197e-01 -7.66829908e-01 -2.26189390e-01
7.13583112e-01 8.78938794e-01 -3.90526950e-01 -1.01286674e+00
1.67355016e-02 1.80115986e+00 -2.51477689e-01 -2.21777737e-01
1.49497509e+00 -6.65032506e-01 -5.01169264e-01 7.21352220e-01
8.42738092e-01 2.05755010e-01 -1.48420143e+00 -2.14977831e-01
-9.16326195e-02 -7.53686786e-01 2.76610821e-01 3.87084205e-03
-1.27368641e+00 9.15490985e-01 5.74974298e-01 1.44028455e-01
9.94011998e-01 3.80868763e-01 1.21635771e+00 -1.29849359e-01
5.31820953e-01 -1.47337282e+00 4.36779261e-01 4.58045691e-01
7.72111416e-01 -1.17715633e+00 5.80254674e-01 -1.10171437e+00
-3.00980896e-01 1.23168695e+00 7.23446429e-01 -5.48548818e-01
8.94451499e-01 7.34061837e-01 2.45032478e-02 -7.68138170e-01
-2.99962640e-01 -5.61398081e-02 -2.29194820e-01 5.87989569e-01
1.92694306e-01 -4.76035208e-01 -3.82214487e-01 5.08221507e-01
5.39574862e-01 -3.17775279e-01 1.06792271e+00 8.20833564e-01
-1.10930645e+00 -9.38320279e-01 -1.01270556e+00 3.66738960e-02
-8.36282253e-01 7.31684789e-02 3.93546261e-02 3.77436250e-01
1.67588554e-02 8.87190878e-01 1.92172118e-02 -1.05014868e-01
2.97158837e-01 -2.29307085e-01 4.75498289e-01 -9.74430978e-01
8.70961323e-02 4.54031765e-01 -1.02501854e-01 -6.08427942e-01
-6.42194927e-01 -5.47124982e-01 -2.02179098e+00 5.46820402e-01
-6.93117678e-01 2.02998251e-01 4.36177492e-01 6.30600452e-01
-2.08064944e-01 4.31784779e-01 8.84482980e-01 -1.48540938e+00
1.06254444e-01 -8.20282280e-01 -1.10428500e+00 1.80387810e-01
5.59264839e-01 -8.00219297e-01 -7.58709073e-01 6.09527230e-01] | [10.969758033752441, -2.2942800521850586] |
1bc7def2-948b-4fd3-b11d-fc8f0dccec1c | joint-matrix-decomposition-for-deep | 2107.04386 | null | https://arxiv.org/abs/2107.04386v3 | https://arxiv.org/pdf/2107.04386v3.pdf | Joint Matrix Decomposition for Deep Convolutional Neural Networks Compression | Deep convolutional neural networks (CNNs) with a large number of parameters require intensive computational resources, and thus are hard to be deployed in resource-constrained platforms. Decomposition-based methods, therefore, have been utilized to compress CNNs in recent years. However, since the compression factor and performance are negatively correlated, the state-of-the-art works either suffer from severe performance degradation or have relatively low compression factors. To overcome this problem, we propose to compress CNNs and alleviate performance degradation via joint matrix decomposition, which is different from existing works that compressed layers separately. The idea is inspired by the fact that there are lots of repeated modules in CNNs. By projecting weights with the same structures into the same subspace, networks can be jointly compressed with larger ranks. In particular, three joint matrix decomposition schemes are developed, and the corresponding optimization approaches based on Singular Value Decomposition are proposed. Extensive experiments are conducted across three challenging compact CNNs for different benchmark data sets to demonstrate the superior performance of our proposed algorithms. As a result, our methods can compress the size of ResNet-34 by 22X with slighter accuracy degradation compared with several state-of-the-art methods. | ['Jiahao Zhou', 'Lei Huang', 'Weize Sun', 'Shaowu Chen'] | 2021-07-09 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 1.09795727e-01 -3.06204230e-01 1.44932221e-03 -3.81105840e-01
-1.01922676e-01 -1.44149363e-02 1.36122644e-01 -2.12610781e-01
-7.64039457e-01 4.12387758e-01 1.87404633e-01 -2.01240540e-01
-2.15015382e-01 -6.70281410e-01 -7.70914435e-01 -7.89504886e-01
3.55502963e-01 -2.25234568e-01 3.29743892e-01 -6.85833022e-02
6.56561106e-02 2.06332445e-01 -1.50545645e+00 1.16678670e-01
7.52043009e-01 1.18086553e+00 6.11346960e-01 9.09796730e-02
-5.55210523e-02 7.21216261e-01 -2.90799499e-01 -5.52206218e-01
3.16449225e-01 -1.11821972e-01 -4.19118106e-01 9.79560688e-02
3.07612926e-01 -5.59418261e-01 -1.10562825e+00 1.46424758e+00
4.00442988e-01 -1.71757340e-02 1.10065416e-01 -9.09859359e-01
-4.90127176e-01 8.18627417e-01 -6.87463045e-01 7.10997507e-02
-4.12055939e-01 -9.80221331e-02 9.06417906e-01 -8.87650669e-01
1.87749341e-01 1.35552621e+00 7.28064239e-01 5.76804876e-01
-1.04100251e+00 -1.08321047e+00 3.58915508e-01 2.96797067e-01
-1.48774147e+00 -6.25144422e-01 6.73636198e-01 -1.93745151e-01
8.54781270e-01 9.65078175e-02 7.27033496e-01 7.25521088e-01
-7.52519593e-02 7.01186061e-01 6.49204493e-01 1.11916266e-01
1.30342647e-01 -1.99187145e-01 -1.71520174e-01 7.60036886e-01
6.92121446e-01 -3.49639326e-01 -3.28940064e-01 1.30266204e-01
8.16423953e-01 6.40126228e-01 -4.15558547e-01 -9.11339298e-02
-1.22045374e+00 6.39908731e-01 7.16059625e-01 2.29460001e-01
-3.16927105e-01 3.27291310e-01 7.24552691e-01 1.92570046e-01
3.15090686e-01 -1.42516807e-01 -4.49130744e-01 -6.21496625e-02
-1.02853143e+00 3.21202427e-01 5.92370391e-01 1.13104820e+00
6.27915680e-01 1.74541101e-01 7.32518733e-02 1.09499872e+00
5.09403408e-01 6.47690669e-02 7.53287315e-01 -6.14860713e-01
1.02764249e+00 7.86405027e-01 -4.63701040e-01 -1.40815222e+00
-2.88381457e-01 -8.25523734e-01 -1.63568246e+00 -2.52925813e-01
-8.94167349e-02 -2.99460262e-01 -7.00380564e-01 1.73022866e+00
-3.69791053e-02 3.75880182e-01 2.37353265e-01 8.98397982e-01
9.20863569e-01 6.29351914e-01 -5.09525314e-02 -1.65939838e-01
1.42146194e+00 -1.26441944e+00 -7.19010413e-01 -3.23570877e-01
5.21178246e-01 -7.39082336e-01 7.22283781e-01 1.99743003e-01
-1.23584139e+00 -8.05037737e-01 -1.46576834e+00 -6.12127930e-02
1.57772064e-01 4.22300398e-01 6.24749005e-01 3.41535300e-01
-9.23471808e-01 7.47945249e-01 -1.15341437e+00 3.22485380e-02
8.26223314e-01 3.95799458e-01 -2.36661747e-01 -3.40534985e-01
-1.03874803e+00 5.26219666e-01 6.40987694e-01 4.36935365e-01
-6.88708484e-01 -7.04232097e-01 -6.13460183e-01 4.47153091e-01
1.25612408e-01 -4.63028580e-01 1.08351445e+00 -5.72394073e-01
-1.28877997e+00 3.23781192e-01 -8.02206546e-02 -5.70975125e-01
5.22996128e-01 -2.41509020e-01 -3.25384974e-01 1.36347502e-01
-2.93425173e-01 6.71706796e-01 6.30565584e-01 -6.91689312e-01
-7.11560547e-01 -3.55622321e-01 6.71792999e-02 9.30994377e-02
-1.33315277e+00 -4.24442366e-02 -8.95451903e-01 -7.13538408e-01
7.01285958e-01 -9.06771839e-01 -6.13298953e-01 1.11855790e-01
-3.87990952e-01 -1.19088568e-01 7.99021542e-01 -5.42369843e-01
1.44484663e+00 -2.23801684e+00 2.99471021e-01 -1.08906016e-01
6.97683334e-01 6.09051108e-01 -2.45414764e-01 8.47569779e-02
-1.40233904e-01 1.72800437e-01 -1.29115671e-01 -6.58954799e-01
-1.92312852e-01 2.68867552e-01 -2.38619804e-01 4.74911332e-01
1.26635939e-01 5.62416613e-01 -6.37323737e-01 -4.06434447e-01
-1.42916530e-01 6.16346121e-01 -7.90701151e-01 1.78564891e-01
9.71405953e-02 -5.44451624e-02 -5.34509420e-01 4.45748329e-01
1.22135496e+00 -2.91610301e-01 2.10053891e-01 -5.01931429e-01
-1.34103164e-01 1.60501882e-01 -1.07068276e+00 1.75300264e+00
-3.10190588e-01 6.01277411e-01 1.73525617e-01 -1.27867234e+00
8.35125506e-01 3.31077188e-01 3.36590260e-01 -4.49223250e-01
2.90842533e-01 3.58108133e-01 1.77459210e-01 -3.97054940e-01
4.90321845e-01 1.08091928e-01 4.10367638e-01 9.26616639e-02
-3.05832527e-03 4.57213134e-01 3.29704672e-01 1.09114744e-01
1.21613169e+00 -3.59493643e-01 -1.16822980e-01 -1.53697401e-01
5.20893276e-01 -4.15128887e-01 1.00495672e+00 2.18071237e-01
-2.35701911e-02 5.60087919e-01 7.65194297e-01 -3.75085890e-01
-1.15843022e+00 -4.45090443e-01 -2.54125863e-01 6.56503797e-01
1.48411751e-01 -6.24996126e-01 -8.69917512e-01 -1.43874332e-01
-3.10192257e-01 -8.99058133e-02 -2.70673811e-01 -2.45408237e-01
-5.28878391e-01 -9.19654548e-01 7.56368399e-01 7.05609679e-01
1.17642081e+00 -6.12734795e-01 -6.08794272e-01 2.86091447e-01
-1.48010522e-01 -1.57897460e+00 -1.90775722e-01 -6.49516061e-02
-1.43168795e+00 -7.79160142e-01 -1.04118466e+00 -1.08281863e+00
9.72704470e-01 8.36522460e-01 7.04449296e-01 4.98216778e-01
5.38947899e-03 -3.65169466e-01 -4.52871203e-01 -2.08810166e-01
3.25701125e-02 3.79042894e-01 1.81413263e-01 9.44004729e-02
7.52506033e-02 -9.49958146e-01 -8.98498654e-01 2.67199904e-01
-1.29618835e+00 4.98991668e-01 1.08141840e+00 8.13919663e-01
6.70377910e-01 4.16125566e-01 3.53762239e-01 -7.88387835e-01
5.73538423e-01 -4.48973179e-01 -7.27077842e-01 2.67913379e-02
-6.12541378e-01 6.70747086e-02 9.59127188e-01 -5.21130145e-01
-8.22069705e-01 1.16404459e-01 -1.30430654e-01 -8.62678647e-01
1.51138142e-01 8.87253821e-01 -3.58961731e-01 -1.52995616e-01
2.32089907e-01 3.64147544e-01 1.44906696e-02 -7.58269489e-01
-1.11746088e-01 4.76904303e-01 3.74273926e-01 -2.81872153e-01
9.30077195e-01 3.54708195e-01 1.14870153e-01 -4.87757832e-01
-6.57827258e-01 -2.72943497e-01 -2.54113317e-01 8.22309181e-02
6.09608114e-01 -1.41085565e+00 -6.40325427e-01 8.10144961e-01
-1.26946390e+00 1.36802331e-01 3.31524163e-01 7.35065818e-01
1.01169571e-01 4.74017829e-01 -8.23787034e-01 -4.29799020e-01
-3.49341393e-01 -1.45323110e+00 4.16925997e-01 4.66404766e-01
6.25959575e-01 -5.35754025e-01 -3.91317010e-01 2.16103971e-01
5.74572384e-01 -6.03787787e-03 8.35442781e-01 -3.39367092e-01
-7.55543411e-01 -2.38853619e-01 -7.59475350e-01 6.34706676e-01
-1.19563609e-01 -1.53039888e-01 -7.12489843e-01 -6.82147861e-01
1.07409120e-01 -2.95852661e-01 1.24000120e+00 9.46421735e-03
1.67890394e+00 -5.10996103e-01 -1.84494808e-01 1.08739114e+00
1.60175252e+00 -1.28159493e-01 7.37222731e-01 3.15457284e-01
8.95767987e-01 3.44082564e-01 2.93108165e-01 7.05194354e-01
1.99806422e-01 3.60937804e-01 8.31419885e-01 2.60309018e-02
8.12717676e-02 -2.17445850e-01 1.59269631e-01 1.69096375e+00
-2.87030905e-01 8.15421715e-02 -6.55609488e-01 4.43836808e-01
-1.94472229e+00 -6.91834271e-01 -1.38148218e-01 1.80552447e+00
8.70827556e-01 2.89627522e-01 -3.46208125e-01 2.83105314e-01
7.05392718e-01 4.93619293e-01 -8.00913632e-01 2.79050440e-01
-7.77596757e-02 -9.03806370e-03 8.05386603e-01 -2.37448558e-01
-1.00345445e+00 6.33869469e-01 5.27452087e+00 1.02177393e+00
-1.03162241e+00 1.04375795e-01 7.73940682e-01 -3.83676410e-01
-1.50205776e-01 1.00010552e-01 -9.31637883e-01 5.10106027e-01
7.50182927e-01 -5.45057468e-02 4.59797800e-01 1.00743282e+00
-1.76017918e-02 4.64062214e-01 -7.36830711e-01 1.31313026e+00
-9.45938528e-02 -1.49738395e+00 1.31062165e-01 1.40456930e-01
9.67966735e-01 2.41375700e-01 1.36088163e-01 2.32481733e-01
-8.82511884e-02 -9.29945707e-01 6.36082828e-01 3.69549870e-01
8.35032701e-01 -1.00722325e+00 8.51465762e-01 3.75072539e-01
-1.58698201e+00 -2.25705042e-01 -1.32318461e+00 -1.12163067e-01
-1.22624591e-01 9.25679505e-01 -1.07434124e-01 4.31541175e-01
8.54601681e-01 9.66649771e-01 -5.90726316e-01 1.10009468e+00
-1.02257188e-02 6.78341925e-01 -7.54192621e-02 -1.98013440e-01
3.91043812e-01 -2.58066684e-01 1.89710200e-01 1.05744863e+00
5.38114011e-01 2.52580613e-01 -8.95404369e-02 7.82933295e-01
-5.88319540e-01 1.37851372e-01 -1.34571701e-01 -1.60478622e-01
5.66319227e-01 1.44681191e+00 -6.22869551e-01 -2.55671829e-01
-6.87030733e-01 7.68437445e-01 4.94494230e-01 3.13736230e-01
-8.48748624e-01 -6.37658179e-01 8.67642999e-01 -8.02330896e-02
5.17713666e-01 -4.40948933e-01 -2.02543423e-01 -1.34458160e+00
3.60985637e-01 -8.83130789e-01 -1.16259940e-01 -3.23539108e-01
-8.24018717e-01 7.38485217e-01 -2.38999590e-01 -1.63350141e+00
4.36600685e-01 -5.59014261e-01 -5.06611824e-01 6.14771724e-01
-1.77567160e+00 -7.53854990e-01 -6.23954177e-01 5.18852592e-01
5.95146060e-01 -4.94849116e-01 5.38926959e-01 7.85796106e-01
-1.09710276e+00 8.18405271e-01 3.89436781e-01 3.59503239e-01
3.16785276e-01 -5.61178625e-01 3.97022277e-01 9.23116922e-01
-2.69473910e-01 8.77043903e-01 3.75181377e-01 -2.13336572e-01
-1.57318008e+00 -1.45925653e+00 6.05361938e-01 6.94857359e-01
6.27381742e-01 -2.11070776e-01 -9.94445920e-01 3.52909952e-01
1.48508444e-01 2.39728644e-01 5.61123550e-01 -1.59527883e-01
-5.07522762e-01 -4.56628889e-01 -7.04458594e-01 7.56520867e-01
1.34547412e+00 -2.93466061e-01 -1.57063752e-01 3.87258649e-01
1.10017943e+00 -5.09487808e-01 -8.07641804e-01 5.91489434e-01
4.07655865e-01 -9.67759013e-01 8.63174558e-01 -5.74116111e-01
9.69817460e-01 -2.51154482e-01 -3.26651901e-01 -1.09266007e+00
-3.85614097e-01 -3.83521587e-01 -2.64983714e-01 1.19161344e+00
1.57216683e-01 -6.66633725e-01 8.59239817e-01 5.32711327e-01
-2.57692963e-01 -1.22348201e+00 -9.35573578e-01 -7.37842679e-01
-2.02417821e-01 -3.89053851e-01 7.50173390e-01 6.76457703e-01
-3.41558248e-01 1.85300559e-01 -4.34603184e-01 1.50619447e-01
5.86629331e-01 -3.67395133e-02 5.83195150e-01 -1.16077697e+00
-3.16294968e-01 -7.10188031e-01 -6.44614577e-01 -1.50691438e+00
5.57818003e-02 -7.94771731e-01 -4.41346364e-03 -1.33505249e+00
4.95315939e-01 -5.41182458e-01 -4.43602681e-01 4.97475445e-01
-1.58496916e-01 1.67271122e-01 2.93866307e-01 6.19024217e-01
-5.76921344e-01 9.14556682e-01 1.31347764e+00 -1.29020631e-01
1.74761087e-01 -1.81390092e-01 -8.88619781e-01 8.64225924e-01
8.26494336e-01 -4.35907394e-01 -5.82305551e-01 -1.08434880e+00
4.04283643e-01 -1.19004352e-02 1.58532962e-01 -1.57849574e+00
6.55051649e-01 1.21661328e-01 3.22136939e-01 -6.09228194e-01
2.50749260e-01 -8.82482469e-01 2.34287623e-02 7.26564765e-01
-2.96175063e-01 2.84596652e-01 1.98316619e-01 7.34575391e-01
-5.13457596e-01 -3.51406276e-01 7.93807864e-01 -8.82144794e-02
-4.16089147e-01 8.15624654e-01 -9.78149027e-02 -1.76211074e-01
6.29243672e-01 1.47453947e-02 -2.35502020e-01 -3.23026278e-03
-1.50498793e-01 1.65651128e-01 5.71510457e-02 3.39719087e-01
1.05951130e+00 -1.58548307e+00 -8.19430649e-01 1.58301696e-01
-1.88576698e-01 5.80652773e-01 6.09261096e-01 8.97989094e-01
-7.32818425e-01 4.71228302e-01 -3.31749320e-01 -4.45698798e-01
-1.26508439e+00 2.97495067e-01 5.51488390e-03 -3.80270749e-01
-5.73681295e-01 1.01668179e+00 1.35027379e-01 -2.08897576e-01
5.23195207e-01 -3.64633918e-01 -3.56302470e-01 -2.62876283e-02
7.19142914e-01 2.45117083e-01 1.50335044e-01 -4.98438954e-01
-1.26182243e-01 5.13617396e-01 -3.92681241e-01 2.50705808e-01
1.67623103e+00 7.52163827e-02 -3.69327009e-01 -1.02687791e-01
1.62067378e+00 -4.33122486e-01 -1.44946253e+00 -7.16066718e-01
-3.99670392e-01 -5.39391577e-01 2.93292046e-01 2.10709855e-01
-1.97179794e+00 1.03230906e+00 5.77518702e-01 -7.63751045e-02
1.45302987e+00 -5.24595141e-01 1.27464616e+00 7.51012087e-01
8.55934918e-02 -1.05706167e+00 1.17787428e-01 6.40777230e-01
7.33893037e-01 -9.44124281e-01 3.35909516e-01 -3.29001993e-01
-2.81112283e-01 1.15387189e+00 9.09478903e-01 -2.37824023e-01
7.88474143e-01 1.83305711e-01 -4.13129777e-01 7.16923401e-02
-8.49154055e-01 2.53128499e-01 2.16096062e-02 1.97018702e-02
3.50828797e-01 -7.92736858e-02 -3.68562132e-01 8.96947503e-01
-2.17662394e-01 -1.01179980e-01 3.42857420e-01 8.92959237e-01
-5.12025416e-01 -8.70254159e-01 -5.79608753e-02 6.73453510e-01
-5.05732596e-01 -2.56224990e-01 2.50719547e-01 3.43988687e-01
1.39020249e-01 7.89377987e-01 -2.73591280e-02 -9.60507035e-01
2.93554276e-01 -4.89617735e-01 2.96445321e-02 -3.43934238e-01
-2.02308431e-01 4.27483320e-02 -4.13528830e-01 -5.80602825e-01
-4.91531432e-01 -4.89829242e-01 -1.16510439e+00 -6.15289271e-01
-3.66668880e-01 -1.74100205e-01 7.57704437e-01 7.17212677e-01
3.79648238e-01 9.19494450e-01 6.90478265e-01 -8.36409390e-01
-8.29474151e-01 -1.01452005e+00 -4.42445576e-01 2.62726307e-01
7.07054436e-02 -4.10714537e-01 -3.18308264e-01 -1.59579813e-01] | [8.49351692199707, 3.086530923843384] |
06b01333-c3ce-45d4-bcca-bde6aca1346d | fast-road-segmentation-via-uncertainty-aware | 2203.04537 | null | https://arxiv.org/abs/2203.04537v1 | https://arxiv.org/pdf/2203.04537v1.pdf | Fast Road Segmentation via Uncertainty-aware Symmetric Network | The high performance of RGB-D based road segmentation methods contrasts with their rare application in commercial autonomous driving, which is owing to two reasons: 1) the prior methods cannot achieve high inference speed and high accuracy in both ways; 2) the different properties of RGB and depth data are not well-exploited, limiting the reliability of predicted road. In this paper, based on the evidence theory, an uncertainty-aware symmetric network (USNet) is proposed to achieve a trade-off between speed and accuracy by fully fusing RGB and depth data. Firstly, cross-modal feature fusion operations, which are indispensable in the prior RGB-D based methods, are abandoned. We instead separately adopt two light-weight subnetworks to learn road representations from RGB and depth inputs. The light-weight structure guarantees the real-time inference of our method. Moreover, a multiscale evidence collection (MEC) module is designed to collect evidence in multiple scales for each modality, which provides sufficient evidence for pixel class determination. Finally, in uncertainty-aware fusion (UAF) module, the uncertainty of each modality is perceived to guide the fusion of the two subnetworks. Experimental results demonstrate that our method achieves a state-of-the-art accuracy with real-time inference speed of 43+ FPS. The source code is available at https://github.com/morancyc/USNet. | ['Anlong Ming', 'Wenteng Liang', 'Fei Sheng', 'Feng Xue', 'Yicong Chang'] | 2022-03-09 | null | null | null | null | ['road-segementation'] | ['computer-vision'] | [ 1.01354785e-01 5.55408821e-02 -1.67608544e-01 -4.75766331e-01
-6.86740518e-01 1.33454371e-02 4.54549640e-01 -5.03705628e-02
-4.66469377e-01 7.50065684e-01 -3.72418612e-01 -2.66958982e-01
-1.84048846e-01 -1.23245120e+00 -7.45403886e-01 -8.23263228e-01
3.53161484e-01 2.09415272e-01 7.69338906e-01 -1.74924508e-01
1.23541825e-01 5.57844877e-01 -2.19291067e+00 -1.58540711e-01
1.30942535e+00 1.56368244e+00 3.66854250e-01 4.99601215e-01
-2.85806835e-01 3.39695424e-01 -6.01103082e-02 -3.39482695e-01
2.31341049e-01 -1.43946633e-01 -3.35982889e-01 1.61153392e-03
3.38468939e-01 -5.05346358e-01 -5.83268225e-01 1.13293040e+00
3.31807196e-01 2.09574044e-01 4.59468901e-01 -1.35896742e+00
-8.67687389e-02 9.39023346e-02 -6.57078624e-01 -1.16588892e-02
8.76032263e-02 3.28468949e-01 7.66197085e-01 -8.39886963e-01
2.80347019e-01 1.01097178e+00 3.91428173e-01 2.82514632e-01
-9.25916672e-01 -7.87953317e-01 3.28641564e-01 4.96237427e-01
-1.51209497e+00 -4.36616242e-01 1.03851235e+00 -1.86395422e-01
4.87611502e-01 1.44603625e-01 9.69044328e-01 8.16815555e-01
3.24120402e-01 7.51558721e-01 1.25341046e+00 7.84918144e-02
3.02339494e-01 -3.00827399e-02 -1.43416360e-01 7.66435266e-01
3.43220443e-01 2.73483753e-01 -7.38065302e-01 4.13735807e-01
9.12739038e-01 -4.04537059e-02 -2.61263043e-01 -2.74494350e-01
-1.12856448e+00 4.78837132e-01 8.51964831e-01 3.09021734e-02
-4.41639751e-01 2.73820877e-01 2.47870162e-02 -1.18831053e-01
4.40890431e-01 -3.49739909e-01 -1.96120948e-01 -1.43797055e-01
-8.78528655e-01 2.21368968e-02 3.46150815e-01 8.30090940e-01
1.11427414e+00 -4.28559370e-02 1.12627111e-01 5.91674209e-01
6.98320031e-01 8.87817502e-01 9.68778729e-02 -1.22621191e+00
4.49371338e-01 6.20289564e-01 1.04189374e-01 -9.85095799e-01
-3.81315529e-01 -2.02590033e-01 -9.82905567e-01 6.70035124e-01
4.48219448e-01 2.24055022e-01 -1.06349063e+00 1.36658978e+00
4.67428446e-01 5.07724285e-01 1.70244947e-02 1.06324124e+00
9.54989910e-01 3.86824489e-01 -8.28634501e-02 9.82247014e-03
1.37503994e+00 -6.35235012e-01 -6.93607390e-01 -3.50686908e-01
-7.90985953e-03 -4.31064487e-01 1.03351700e+00 3.42824906e-01
-1.00644052e+00 -7.29215264e-01 -1.31094670e+00 -3.04805458e-01
-5.40501237e-01 6.12757690e-02 6.50843680e-01 3.94883990e-01
-7.58451760e-01 3.84164542e-01 -9.77206945e-01 4.69392072e-03
5.52227020e-01 1.28856435e-01 -1.78253561e-01 -2.22187728e-01
-1.32785463e+00 8.89364362e-01 4.40943480e-01 6.05299771e-01
-5.86745322e-01 -4.55266446e-01 -9.43821728e-01 -2.12317258e-01
4.23216850e-01 -6.86499894e-01 9.10320878e-01 -3.58044684e-01
-1.71744061e+00 5.77976763e-01 -2.86969393e-01 -2.07767069e-01
5.95413983e-01 -4.52557206e-02 -3.36563468e-01 3.22095186e-01
1.14356734e-01 9.23649132e-01 6.79256737e-01 -1.47724974e+00
-1.12959051e+00 -3.99968117e-01 1.38434321e-01 2.64993131e-01
-9.02565010e-03 -8.28459263e-01 -8.41003299e-01 -2.88501233e-01
6.96138740e-01 -6.17619991e-01 -1.95332721e-01 3.50297868e-01
-3.23645920e-01 -1.31517783e-01 7.44710803e-01 -3.78798902e-01
9.00597155e-01 -2.02297235e+00 4.71866988e-02 3.62436265e-01
2.74828285e-01 4.53213565e-02 2.09171101e-01 -2.37558931e-01
4.94995594e-01 -1.73543226e-02 -5.71821570e-01 -4.84023839e-01
-2.01058183e-02 6.10794008e-01 -1.97364539e-01 4.77106988e-01
3.89905661e-01 8.63101482e-01 -1.06812549e+00 -8.26581180e-01
8.52136493e-01 7.91094065e-01 -1.24222398e-01 -2.85488348e-02
-1.42637908e-01 5.96403062e-01 -7.49160707e-01 8.66325378e-01
9.82943952e-01 2.55267676e-02 -2.18192592e-01 -3.89472008e-01
-3.24579895e-01 2.81862020e-01 -1.39112389e+00 1.91117585e+00
-5.03636956e-01 3.65856826e-01 1.36729285e-01 -5.73553562e-01
9.35721517e-01 3.69082727e-02 3.45364392e-01 -9.30720627e-01
2.46138990e-01 4.28991675e-01 -2.56803751e-01 -3.18316400e-01
5.67123294e-01 8.64712819e-02 1.13666221e-01 -9.67143252e-02
-2.01818123e-01 -6.99078202e-01 1.69568866e-01 5.27358167e-02
6.71817183e-01 4.98209238e-01 -1.33312345e-01 5.40291779e-02
6.51269794e-01 -9.62361768e-02 8.82460177e-01 5.26921570e-01
-3.49957556e-01 6.20032430e-01 1.98083907e-01 -4.15056646e-02
-7.03599811e-01 -1.26504636e+00 -5.06815195e-01 3.53341460e-01
9.33506310e-01 -3.63707952e-02 -4.08270866e-01 -3.78728300e-01
1.56440511e-01 5.92209697e-01 -5.32888532e-01 -1.23107262e-01
-3.14375162e-01 -5.50104678e-01 3.54287297e-01 6.93595052e-01
1.07230401e+00 -6.70049787e-01 -8.71632755e-01 2.07348213e-01
-3.44669312e-01 -1.28852844e+00 1.08400784e-01 3.90195847e-03
-1.07057357e+00 -1.07089031e+00 -3.56504470e-01 -1.34076610e-01
5.81570923e-01 4.23514366e-01 7.47024357e-01 1.95147425e-01
4.20267582e-02 2.23461404e-01 -9.96725410e-02 -3.17876786e-01
5.76032791e-03 -7.88468048e-02 -5.46271093e-02 -5.59681021e-02
1.96349636e-01 -8.10201287e-01 -7.98535228e-01 4.13828254e-01
-7.92901218e-01 3.92238945e-01 8.10754955e-01 5.05850911e-01
1.00981843e+00 2.40951642e-01 2.35308960e-01 -2.29887560e-01
-1.84482299e-02 -3.65365148e-01 -8.10776591e-01 5.31596206e-02
-5.90284646e-01 -1.00968972e-01 1.39875397e-01 -6.55753240e-02
-1.25080752e+00 1.59816414e-01 -3.84706557e-01 -4.34295714e-01
-2.72274792e-01 3.17295074e-01 -3.01268816e-01 -6.37326241e-02
2.09563315e-01 1.92034930e-01 1.32357806e-01 -1.37414739e-01
4.69840914e-01 5.47434866e-01 7.75059521e-01 -6.01758063e-01
8.16506267e-01 9.30386901e-01 2.85948873e-01 -6.42001450e-01
-7.34846890e-01 -3.20678771e-01 -6.43535554e-01 -7.28480458e-01
9.06383455e-01 -1.02288020e+00 -9.37057555e-01 6.90286994e-01
-9.30487692e-01 -2.35820562e-01 -2.66654402e-01 6.26021206e-01
-5.76778114e-01 2.45566294e-01 -4.80248183e-01 -9.51083243e-01
-4.86091524e-02 -1.31721127e+00 1.17889726e+00 6.82113290e-01
4.79066461e-01 -6.19194031e-01 -3.98007542e-01 5.75160265e-01
2.56552786e-01 3.01795691e-01 2.97942072e-01 3.63191694e-01
-9.11443293e-01 -1.00506835e-01 -6.64646268e-01 2.33743086e-01
1.02271356e-01 2.77360678e-01 -1.27589154e+00 2.47755215e-01
-2.25446299e-01 -2.70546466e-01 1.19751644e+00 4.53637719e-01
1.25639725e+00 3.65743548e-01 -3.48454893e-01 6.37966514e-01
1.33882654e+00 -1.68384507e-01 8.54122937e-01 3.23889852e-01
7.97526956e-01 6.88495159e-01 9.52176988e-01 3.48623365e-01
9.25711989e-01 4.64256436e-01 9.07604158e-01 -1.27087578e-01
-3.54355395e-01 -2.21879810e-01 1.40009150e-01 7.06903100e-01
-4.89206344e-01 -1.01684228e-01 -8.17576587e-01 3.45798194e-01
-2.14954495e+00 -7.15923846e-01 -3.75166863e-01 2.30581617e+00
8.39592457e-01 6.06977880e-01 -1.85184330e-01 3.25605512e-01
5.57969630e-01 1.35285422e-01 -8.13850522e-01 7.22426996e-02
-2.45157510e-01 -6.60766661e-02 7.71779716e-01 6.10782564e-01
-9.83515918e-01 7.72716343e-01 4.91295671e+00 1.12795949e+00
-1.08101130e+00 1.18080944e-01 6.40929580e-01 7.56238922e-02
-5.53396761e-01 8.69408436e-03 -7.68669784e-01 5.45120776e-01
6.45524561e-01 3.14723939e-01 1.55565247e-01 5.94444513e-01
2.92623430e-01 -6.95852876e-01 -5.04808784e-01 8.77507269e-01
-3.40888888e-01 -1.25351751e+00 -2.78999031e-01 1.40987784e-01
5.77980042e-01 2.39234701e-01 -8.63082856e-02 -4.78729233e-03
2.67439216e-01 -6.96352303e-01 9.28487420e-01 8.91947865e-01
7.37279117e-01 -9.22824502e-01 8.16313744e-01 3.79983693e-01
-1.50420046e+00 3.88555303e-02 -3.87188017e-01 -1.65310558e-02
4.75612342e-01 1.21890855e+00 -9.09695774e-02 9.25738871e-01
9.51112270e-01 8.41192365e-01 -3.91699106e-01 9.18044984e-01
-6.99667037e-01 2.90435195e-01 -6.98680222e-01 2.79242992e-01
2.79032886e-01 -4.26723391e-01 5.72760403e-01 8.73149812e-01
2.44011447e-01 1.16046101e-01 1.18451267e-01 1.00224292e+00
1.30620822e-01 -3.99888694e-01 -3.76905680e-01 4.38387871e-01
5.87218583e-01 1.47190511e+00 -9.25328612e-01 -3.26638699e-01
-3.23020697e-01 6.53350472e-01 1.07008167e-01 3.73869807e-01
-1.02325559e+00 -4.16555613e-01 7.11925030e-01 5.69910072e-02
2.40725473e-01 -4.98437196e-01 -7.16653407e-01 -9.90308225e-01
2.29164988e-01 -7.88058490e-02 8.25603667e-04 -8.37549746e-01
-1.09402525e+00 4.16306347e-01 2.08745543e-02 -1.39327013e+00
2.52022237e-01 -5.54360926e-01 -4.53778714e-01 9.18112993e-01
-2.20565677e+00 -1.25011599e+00 -7.49165833e-01 5.36050975e-01
3.25134784e-01 4.74266082e-01 3.59011889e-01 4.53426689e-01
-6.89791083e-01 8.66132006e-02 -2.10185379e-01 -1.75398260e-01
3.85679781e-01 -1.21425176e+00 7.28010833e-02 1.08657372e+00
-2.63112396e-01 1.15967177e-01 3.87168974e-01 -6.13492250e-01
-1.34737742e+00 -9.25548196e-01 4.51921910e-01 -1.94541126e-01
5.72440386e-01 -7.66320527e-02 -8.95680666e-01 1.27730379e-02
-2.54206151e-01 3.70771438e-01 1.60662577e-01 -2.46734515e-01
-1.64409399e-01 -5.09646773e-01 -1.18624806e+00 5.03640413e-01
1.09276223e+00 -5.04469275e-01 -2.89440662e-01 -1.54206678e-01
7.70209491e-01 -6.40602708e-01 -9.96164560e-01 8.47155452e-01
6.93312764e-01 -1.27201355e+00 1.06808424e+00 3.58506322e-01
4.99724418e-01 -9.11179960e-01 -3.07913601e-01 -8.63323450e-01
2.44957972e-02 6.87293038e-02 -4.16067421e-01 1.16222250e+00
1.77146584e-01 -9.15108681e-01 6.04818940e-01 5.43348253e-01
-3.43063205e-01 -9.67522144e-01 -1.34469247e+00 -5.53372622e-01
-3.16570312e-01 -9.74334180e-01 7.74298966e-01 4.66320485e-01
-4.43307996e-01 -6.82428703e-02 7.02195466e-02 4.54436332e-01
8.82902682e-01 2.73008615e-01 5.40884078e-01 -1.27654731e+00
1.30915895e-01 -7.29450941e-01 -4.44302499e-01 -1.29745054e+00
-8.05756524e-02 -5.36335349e-01 5.15574634e-01 -1.81768537e+00
-3.33779246e-01 -9.11546826e-01 -4.69977885e-01 3.97481829e-01
-2.20765769e-01 5.40777445e-01 -2.21173000e-02 2.38024890e-01
-4.20838535e-01 8.65558982e-01 1.45501697e+00 -1.22705005e-01
-4.92955558e-02 5.85183576e-02 -4.11369711e-01 8.79991949e-01
8.66489112e-01 -2.03853235e-01 -4.23326552e-01 -4.00470942e-01
3.03741127e-01 -5.81012405e-02 7.40763545e-01 -1.26341546e+00
3.39185923e-01 -2.47347683e-01 4.20739293e-01 -9.61190760e-01
6.86439037e-01 -8.26057434e-01 -1.61360025e-01 3.55083227e-01
3.18964094e-01 -4.21945274e-01 2.29091987e-01 7.27477729e-01
-3.10183167e-01 9.13843960e-02 7.25793302e-01 1.07745610e-01
-1.04570973e+00 4.32939798e-01 -1.78386435e-01 -3.20365995e-01
8.43022168e-01 -7.01591730e-01 -3.01226586e-01 -1.51878893e-01
-2.52877980e-01 5.34497917e-01 4.95581299e-01 2.93431520e-01
9.02423978e-01 -1.34364367e+00 -4.83278543e-01 1.51153162e-01
2.30830401e-01 8.42817068e-01 4.86497611e-01 1.22633362e+00
-3.95977348e-01 4.04904671e-02 -1.72027901e-01 -1.00035954e+00
-6.41297698e-01 1.05709704e-02 2.32684419e-01 8.96408111e-02
-7.32632577e-01 7.33436108e-01 -1.33652031e-01 -3.58540863e-01
1.41919181e-01 -5.15561759e-01 -1.87885821e-01 6.04063012e-02
3.30608487e-01 5.46224654e-01 7.30760098e-02 -6.78106010e-01
-4.80811924e-01 8.02338541e-01 4.83054042e-01 -2.13217080e-01
1.06128728e+00 -5.38635552e-01 -3.07606254e-02 6.03361607e-01
7.23354459e-01 -1.98758408e-01 -1.76841021e+00 -1.83013588e-01
-5.04077315e-01 -5.82532048e-01 6.08490586e-01 -4.85962957e-01
-1.29850149e+00 1.01802790e+00 6.13716125e-01 -5.53707592e-02
1.19039047e+00 -1.21266536e-01 9.74294960e-01 1.56047031e-01
5.16437232e-01 -1.25622869e+00 -3.30484569e-01 3.25078756e-01
6.06036782e-01 -1.53344917e+00 2.01652721e-01 -6.43919945e-01
-4.46199298e-01 1.01316464e+00 7.86614478e-01 6.29609600e-02
7.75009215e-01 2.38938898e-01 -8.41485858e-02 -2.22093031e-01
-3.13145816e-01 -6.46909773e-01 3.81763369e-01 4.14835334e-01
-8.07146206e-02 1.40105411e-01 -6.98563159e-02 4.07495171e-01
6.66036233e-02 1.06054649e-01 2.13552460e-01 9.15454268e-01
-7.24136651e-01 -7.25095570e-01 -3.35443407e-01 4.05026913e-01
2.43502527e-01 1.44974768e-01 1.59262195e-01 6.92174733e-01
5.37562668e-01 1.08485901e+00 2.83038855e-01 -5.97650230e-01
3.53754133e-01 -3.33737910e-01 4.29366261e-01 -1.73013180e-01
1.07742906e-01 -6.29105344e-02 -2.31546350e-02 -9.18638706e-01
-6.74516320e-01 -4.71881747e-01 -1.66444182e+00 -3.96820843e-01
-4.54264194e-01 -1.74576759e-01 9.88571227e-01 1.01466429e+00
4.05414551e-01 7.16752827e-01 5.40415287e-01 -1.19284332e+00
1.00466229e-01 -7.07614422e-01 -5.39912641e-01 -1.25802353e-01
1.87702924e-01 -1.15698302e+00 -4.22670066e-01 -2.29418173e-01] | [8.409512519836426, -2.365255832672119] |
ba14d6d2-9c13-4f70-ad8d-e362835ccbdc | the-neural-data-router-adaptive-control-flow | 2110.07732 | null | https://arxiv.org/abs/2110.07732v4 | https://arxiv.org/pdf/2110.07732v4.pdf | The Neural Data Router: Adaptive Control Flow in Transformers Improves Systematic Generalization | Despite progress across a broad range of applications, Transformers have limited success in systematic generalization. The situation is especially frustrating in the case of algorithmic tasks, where they often fail to find intuitive solutions that route relevant information to the right node/operation at the right time in the grid represented by Transformer columns. To facilitate the learning of useful control flow, we propose two modifications to the Transformer architecture, copy gate and geometric attention. Our novel Neural Data Router (NDR) achieves 100% length generalization accuracy on the classic compositional table lookup task, as well as near-perfect accuracy on the simple arithmetic task and a new variant of ListOps testing for generalization across computational depths. NDR's attention and gating patterns tend to be interpretable as an intuitive form of neural routing. Our code is public. | ['Jürgen Schmidhuber', 'Kazuki Irie', 'Róbert Csordás'] | 2021-10-14 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [ 1.21092230e-01 1.30823568e-01 -2.70631760e-01 -2.92800277e-01
-5.52996516e-01 -8.78254414e-01 1.91600531e-01 4.13189471e-01
-4.47486416e-02 8.06474030e-01 3.51650298e-01 -1.10427403e+00
-3.39383662e-01 -1.07454836e+00 -7.05423772e-01 -1.90468833e-01
-4.16012228e-01 6.22366548e-01 2.20108986e-01 -6.39992833e-01
7.42585123e-01 7.42991269e-01 -1.27276421e+00 5.16046941e-01
6.75872922e-01 1.19296753e+00 9.56645235e-03 4.80296165e-01
-2.72841990e-01 1.03297865e+00 -4.74531949e-01 -5.56281149e-01
2.45764926e-01 -1.87158167e-01 -1.17473531e+00 -5.01924694e-01
5.61442256e-01 -1.76101372e-01 -1.81472749e-01 8.26862693e-01
2.22650275e-01 1.53654516e-01 4.92036343e-01 -1.19529653e+00
-8.47910047e-01 1.04979575e+00 -5.35230398e-01 6.50216341e-01
2.72391438e-01 2.41872638e-01 1.35595608e+00 -6.89290881e-01
4.95721370e-01 1.34016860e+00 7.69219935e-01 5.60378373e-01
-1.52527213e+00 -6.80872083e-01 3.16244423e-01 1.12607308e-01
-1.26925313e+00 -2.69658267e-01 3.05740058e-01 -2.27695912e-01
1.43112051e+00 2.02637285e-01 6.26792729e-01 6.62522316e-01
4.98676747e-01 6.19056761e-01 9.33930218e-01 1.72417611e-03
3.63959551e-01 -2.35182911e-01 1.51524261e-01 8.01194847e-01
3.84735048e-01 2.23342273e-02 -6.97674155e-01 1.93893127e-02
8.75499845e-01 -1.34726346e-01 -1.93023890e-01 -3.31120938e-01
-8.99191618e-01 7.23718226e-01 9.82900500e-01 2.13081732e-01
6.57815021e-03 3.43621165e-01 5.30646145e-01 6.84942842e-01
-3.10146175e-02 1.21518421e+00 -7.27047145e-01 -2.32314974e-01
-6.27036154e-01 4.77483064e-01 9.20307219e-01 1.18167913e+00
7.75171936e-01 1.08282514e-01 -3.35700661e-02 2.21889377e-01
-1.64887950e-01 4.52837422e-02 2.85583317e-01 -7.79180348e-01
7.72091389e-01 8.64149988e-01 -2.78922588e-01 -9.10797596e-01
-5.99026442e-01 -8.68522584e-01 -7.39470601e-01 1.82130784e-01
6.99374557e-01 1.98284358e-01 -6.99958265e-01 1.62342656e+00
-1.60113364e-01 -4.86939818e-01 -3.05663869e-02 5.97664237e-01
2.96297669e-01 7.70534158e-01 1.71831548e-01 2.32136339e-01
1.11324620e+00 -8.09221268e-01 -3.38317573e-01 -4.12191302e-01
9.65683639e-01 -3.01245362e-01 1.42986035e+00 6.47602558e-01
-1.42379141e+00 -3.22207242e-01 -1.20030737e+00 -6.70871377e-01
-6.89740121e-01 -5.28435886e-01 1.19275200e+00 5.17843425e-01
-1.22664404e+00 6.75602853e-01 -6.09289706e-01 -1.18169881e-01
7.02006578e-01 5.68143666e-01 -2.20091537e-01 -2.31492102e-01
-9.58077729e-01 9.90552604e-01 1.60340160e-01 1.82231814e-01
-8.75325561e-01 -1.14307129e+00 -7.61128128e-01 4.90622282e-01
4.81249452e-01 -6.42583311e-01 1.37366366e+00 -7.84471333e-01
-9.95510101e-01 6.18911505e-01 -8.83612409e-02 -6.68812215e-01
1.69778034e-01 -2.13569164e-01 -1.42588109e-01 -3.04724306e-01
4.03545722e-02 7.45862424e-01 3.07393283e-01 -8.16297472e-01
-5.43328285e-01 -5.61481535e-01 3.57789129e-01 1.34706646e-01
-1.47412270e-01 -3.36585999e-01 -1.01478398e-01 -7.04342842e-01
2.38928109e-01 -4.12141800e-01 -3.48927468e-01 5.47853447e-02
-4.74159867e-01 -3.86556506e-01 3.78202111e-01 -4.21718150e-01
1.58166623e+00 -2.04505587e+00 1.57429665e-01 4.57466990e-01
4.52768922e-01 -2.64223903e-01 1.14929587e-01 4.26362514e-01
-2.86422431e-01 4.83535051e-01 -1.92553580e-01 3.65954608e-01
9.66978818e-02 9.69805568e-02 -7.91750968e-01 2.12561280e-01
3.46624821e-01 1.07725859e+00 -1.01822948e+00 -6.82735667e-02
-2.33955011e-01 -1.38214752e-01 -1.10414720e+00 -1.43152341e-01
-4.49713916e-01 4.07907888e-02 -3.95262688e-01 7.17724800e-01
1.57289296e-01 -4.77034301e-01 1.50490284e-01 -7.92749226e-03
-2.21601173e-01 9.46984529e-01 -8.34173083e-01 1.60114360e+00
-6.54646516e-01 8.03337455e-01 -1.80303514e-01 -7.90171206e-01
7.09068894e-01 -2.47404158e-01 -2.55872846e-01 -1.14890933e+00
-5.26348539e-02 2.81306952e-01 5.47866940e-01 -8.85986835e-02
7.26346433e-01 -1.35380477e-01 -3.31082612e-01 5.19009233e-01
-3.51219594e-01 -1.82716474e-01 2.26566806e-01 3.92059386e-01
1.31851733e+00 -4.12289575e-02 3.39415491e-01 -7.88089216e-01
2.74233371e-01 2.62514591e-01 3.67616177e-01 8.57530177e-01
2.50527948e-01 1.82762176e-01 1.14278114e+00 -7.43389070e-01
-1.01671970e+00 -1.05134380e+00 -2.02372268e-01 1.42015183e+00
-2.71948874e-02 -6.59361005e-01 -5.26216567e-01 -2.92091399e-01
1.45698056e-01 9.33957756e-01 -8.20699871e-01 -3.21400732e-01
-8.24243009e-01 -3.56489331e-01 5.59981406e-01 1.14116073e+00
4.78165537e-01 -1.01222348e+00 -7.98558414e-01 2.89027125e-01
2.83728987e-01 -5.73240578e-01 -5.10979235e-01 8.07547033e-01
-1.20477235e+00 -9.93574619e-01 -6.54235780e-02 -7.63268113e-01
6.73188984e-01 -1.83730304e-01 1.51121509e+00 2.28971913e-01
-3.24305326e-01 -1.51258245e-01 2.16993485e-02 -1.26184776e-01
-1.87975511e-01 5.76445043e-01 -3.04816753e-01 -6.10027134e-01
2.75457412e-01 -6.05639637e-01 -4.68752414e-01 4.16102618e-01
-6.38445258e-01 -7.06670880e-02 4.82128441e-01 8.12748313e-01
3.59013587e-01 1.08141489e-01 4.51075584e-01 -1.03960598e+00
8.64056587e-01 -5.05872190e-01 -6.97270632e-01 5.48586696e-02
-7.06220329e-01 5.74334919e-01 1.05781305e+00 -4.26658355e-02
-4.98546153e-01 -3.66743147e-01 -1.19977169e-01 -1.46710873e-01
2.05086604e-01 4.29790229e-01 -1.55349508e-01 -4.95975930e-03
9.58325028e-01 5.36163040e-02 -2.39834949e-01 -3.58568192e-01
2.58721352e-01 -9.12566483e-02 4.74583030e-01 -1.00996196e+00
6.19021773e-01 1.27466256e-02 2.40365893e-01 -4.04035002e-01
-6.42516136e-01 1.51309758e-01 -2.23650590e-01 4.07027096e-01
6.62620425e-01 -5.32539010e-01 -1.11638641e+00 -4.08116207e-02
-8.21170807e-01 -7.88788438e-01 -4.64840919e-01 -3.24793935e-01
-5.06427646e-01 -2.63674051e-01 -6.12766087e-01 -4.66752142e-01
-1.70267612e-01 -1.35245454e+00 8.33548009e-01 5.88776432e-02
-5.45215786e-01 -1.01329446e+00 -4.54591542e-01 -3.89793068e-01
7.08943605e-01 -5.63371852e-02 1.87292063e+00 -7.84491360e-01
-1.16504180e+00 -4.13417295e-02 -3.51973712e-01 -7.99597800e-02
-1.81801289e-01 -2.05961853e-01 -6.11134887e-01 -1.40714273e-01
-4.39611584e-01 -4.15451437e-01 8.90119553e-01 1.61361784e-01
1.72705841e+00 -3.32631588e-01 -4.56162870e-01 7.78838456e-01
1.35765517e+00 3.33007008e-01 6.61687553e-01 3.59647512e-01
4.52444226e-01 3.55227649e-01 1.44754335e-01 1.67524755e-01
4.11502063e-01 4.15869206e-01 4.95520592e-01 8.01515430e-02
1.07207283e-01 -5.99817812e-01 8.65195245e-02 3.00426006e-01
1.23891182e-01 -2.72756696e-01 -1.20128572e+00 3.32814246e-01
-1.27718854e+00 -7.87543118e-01 9.47363973e-02 2.27830100e+00
8.40481520e-01 6.36371434e-01 3.22884321e-02 3.29645693e-01
2.09610552e-01 -2.62585431e-02 -6.82445943e-01 -9.43509102e-01
7.50816688e-02 6.81169331e-01 6.33009553e-01 7.07158625e-01
-5.61216295e-01 9.78772640e-01 7.60534668e+00 7.30958581e-01
-9.73908067e-01 -2.57491410e-01 9.87308383e-01 -5.48254810e-02
-5.68786025e-01 2.03414727e-02 -1.04182827e+00 5.24726957e-02
9.90384221e-01 -1.09383792e-01 8.66297185e-01 6.06193066e-01
-2.70288974e-01 -1.97253808e-01 -1.68668509e+00 7.65029490e-01
-3.02450091e-01 -1.61679304e+00 4.30231094e-01 -1.11926161e-02
4.34760869e-01 -3.65714610e-01 3.96969527e-01 4.34244961e-01
4.58460450e-01 -1.53204417e+00 6.97119474e-01 1.91663891e-01
8.78410578e-01 -9.13893998e-01 2.29841217e-01 -5.88756753e-03
-1.12086880e+00 -5.61688960e-01 -2.98719764e-01 -5.55992186e-01
-3.58682960e-01 -1.20411729e-02 -7.67556190e-01 -8.81286263e-02
6.67765439e-01 6.05398774e-01 -7.49958456e-01 9.15226698e-01
-3.21889758e-01 3.17654520e-01 -2.40640208e-01 -3.50000471e-01
5.06668389e-01 1.24983773e-01 1.22097112e-01 9.39290106e-01
1.04966208e-01 1.04871877e-01 -2.42782846e-01 1.00622499e+00
-2.84644008e-01 -1.01866968e-01 -8.21281552e-01 -1.32584602e-01
6.34880483e-01 8.26786518e-01 -9.62048292e-01 -1.71960413e-01
-1.13771282e-01 4.21340644e-01 6.90735579e-01 3.08479816e-01
-4.68212754e-01 -4.54743057e-01 6.30081296e-01 4.41318691e-01
2.85748661e-01 -1.60414413e-01 -8.90439630e-01 -6.90519989e-01
-5.41722439e-02 -1.04400837e+00 5.43444932e-01 -8.50782871e-01
-9.96255994e-01 5.37217796e-01 9.60904360e-03 -7.03034699e-01
-1.89354196e-01 -1.09819901e+00 -6.56882524e-01 9.93245006e-01
-1.09450948e+00 -5.04541397e-01 9.85663906e-02 5.61711133e-01
4.26745057e-01 -7.88372681e-02 7.50308514e-01 1.69968426e-01
-3.83619159e-01 9.63646054e-01 -3.22981656e-01 -2.43934877e-02
1.27177849e-01 -1.57110095e+00 8.89223099e-01 4.98522967e-01
-9.90534201e-02 1.22787118e+00 7.46193767e-01 -3.16686213e-01
-1.83105469e+00 -7.35613644e-01 6.67253733e-01 -6.48901403e-01
8.86361659e-01 -6.59954727e-01 -8.71783912e-01 1.10478961e+00
1.00914724e-01 -2.86629528e-01 5.09618938e-01 5.26444972e-01
-7.99215198e-01 -3.71776640e-01 -9.21253204e-01 9.01495874e-01
1.26276088e+00 -5.63972771e-01 -6.55006707e-01 2.58172482e-01
7.33208179e-01 -4.80072767e-01 -6.79679990e-01 1.40544727e-01
5.30260623e-01 -8.39265406e-01 8.16695154e-01 -8.96887362e-01
7.97092319e-01 -1.40128165e-01 -3.69608730e-01 -1.18969834e+00
-3.89949143e-01 -8.85483265e-01 1.03751853e-01 7.67103255e-01
8.96838844e-01 -8.17393780e-01 6.76259935e-01 4.18807685e-01
-4.18095380e-01 -1.16774774e+00 -6.12330258e-01 -5.00636160e-01
4.47224587e-01 -4.91407037e-01 8.20133269e-01 7.04579830e-01
2.54578024e-01 6.98248684e-01 1.73328310e-01 -3.15092392e-02
2.35267445e-01 2.55581766e-01 5.87815344e-01 -1.04861844e+00
-3.21207076e-01 -1.00353420e+00 -3.73121709e-01 -1.41841984e+00
-1.48456246e-01 -1.20196009e+00 -1.92376480e-01 -1.38089156e+00
-8.12549517e-02 -6.39866710e-01 -1.41649351e-01 5.52177191e-01
2.38917097e-01 4.36216742e-02 -3.38816978e-02 -1.91576213e-01
-4.12020415e-01 1.03946924e-01 1.39952040e+00 -6.50556460e-02
-4.96841259e-02 -1.65198386e-01 -1.38141966e+00 4.64109898e-01
6.64134502e-01 -2.05855325e-01 -6.46612585e-01 -6.65315211e-01
1.07361627e+00 1.97460771e-01 3.13061863e-01 -1.07449937e+00
3.34365785e-01 6.48886561e-02 5.40744662e-01 -7.04480290e-01
-5.39148822e-02 -6.96751475e-01 -1.36177421e-01 6.09676361e-01
-7.99094617e-01 9.70923007e-01 7.55787015e-01 2.66077161e-01
9.98628885e-02 1.53276296e-02 5.63025415e-01 -2.57021546e-01
-8.04662287e-01 2.27084160e-01 -3.32020730e-01 5.24570167e-01
6.38840497e-01 -3.65573674e-01 -7.15436459e-01 -2.68409640e-01
-6.87476456e-01 4.50563908e-01 4.37033623e-01 3.23349625e-01
4.58602816e-01 -9.18584287e-01 -2.13105664e-01 4.95696366e-01
2.30126344e-02 1.66642711e-01 -1.38990492e-01 6.48043931e-01
-8.21287930e-01 6.50004745e-01 -4.09758657e-01 -2.85700530e-01
-3.61388743e-01 8.26889396e-01 5.65566361e-01 -2.90018827e-01
-5.93505740e-01 1.05933571e+00 5.28272152e-01 -5.50674647e-02
3.38370770e-01 -9.08130646e-01 2.70163149e-01 -6.74326867e-02
6.28787577e-01 3.40782136e-01 3.03096265e-01 1.49915710e-01
-3.04128468e-01 3.80395561e-01 -3.79570246e-01 2.33379111e-01
1.13289893e+00 2.14027852e-01 -4.28782970e-01 5.57947099e-01
1.02906787e+00 -4.77278866e-02 -9.90420401e-01 5.75101860e-02
1.51305422e-01 -2.05581337e-01 -1.11968227e-01 -1.16464543e+00
-8.46924007e-01 1.28752029e+00 1.33303404e-01 4.83703107e-01
1.14609754e+00 -9.81844813e-02 5.07167459e-01 7.41381526e-01
4.95707840e-01 -5.42326808e-01 -4.91064750e-02 8.25014114e-01
9.25098658e-01 -5.85017025e-01 -1.61466151e-02 -2.42630333e-01
-2.46733859e-01 1.29338050e+00 9.04335618e-01 -2.41700977e-01
3.77354115e-01 7.03720868e-01 -4.51859891e-01 -2.63743430e-01
-1.15328121e+00 1.10326283e-01 4.14108299e-02 4.43359673e-01
5.90117157e-01 -2.83532768e-01 1.87911153e-01 4.53860164e-01
-7.39099741e-01 -2.69702137e-01 4.83082503e-01 8.42991114e-01
-5.25541961e-01 -5.95856965e-01 2.63251420e-02 5.47396004e-01
-3.60854864e-01 -5.61204135e-01 -2.67958850e-01 1.21697128e+00
-3.51378649e-01 3.98844749e-01 4.94099677e-01 -2.90805161e-01
4.45398986e-01 1.62888005e-01 5.68703234e-01 -5.70105374e-01
-8.88926268e-01 -2.94157028e-01 -2.45526750e-02 -8.29753876e-01
4.03048068e-01 -4.12942499e-01 -1.51004541e+00 -7.74885356e-01
2.56471187e-01 1.65544569e-01 1.13427952e-01 7.28572369e-01
2.75754869e-01 8.07645440e-01 1.60086110e-01 -5.53488731e-01
-7.25441337e-01 -5.98470271e-01 -3.92270595e-01 -3.42668556e-02
4.83327061e-01 -5.38935900e-01 -6.72960579e-02 -3.85613918e-01] | [9.466938972473145, 7.117514610290527] |
92cf9d9b-ff3a-44e2-a659-1268aac90014 | discriminant-projection-representation-based | 1712.01643 | null | http://arxiv.org/abs/1712.01643v1 | http://arxiv.org/pdf/1712.01643v1.pdf | Discriminant Projection Representation-based Classification for Vision Recognition | Representation-based classification methods such as sparse
representation-based classification (SRC) and linear regression classification
(LRC) have attracted a lot of attentions. In order to obtain the better
representation, a novel method called projection representation-based
classification (PRC) is proposed for image recognition in this paper. PRC is
based on a new mathematical model. This model denotes that the 'ideal
projection' of a sample point $x$ on the hyper-space $H$ may be gained by
iteratively computing the projection of $x$ on a line of hyper-space $H$ with
the proper strategy. Therefore, PRC is able to iteratively approximate the
'ideal representation' of each subject for classification. Moreover, the
discriminant PRC (DPRC) is further proposed, which obtains the discriminant
information by maximizing the ratio of the between-class reconstruction error
over the within-class reconstruction error. Experimental results on five
typical databases show that the proposed PRC and DPRC are effective and
outperform other state-of-the-art methods on several vision recognition tasks. | ['Qingxiang Feng', 'Yicong Zhou'] | 2017-11-19 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 3.42732817e-01 -3.66625458e-01 -1.80109799e-01 -3.59187722e-01
-7.43851304e-01 4.46077824e-01 3.52140814e-01 -3.46499950e-01
-1.72910109e-01 4.11177278e-01 4.97477278e-02 1.23013102e-01
-4.41620111e-01 -7.98237741e-01 -3.38709384e-01 -9.25458670e-01
1.39702857e-01 -1.28082493e-02 -6.60464540e-02 9.08326171e-03
4.62008804e-01 5.02724528e-01 -1.74833250e+00 2.67092258e-01
9.01411235e-01 1.27808475e+00 3.99275988e-01 -1.60338327e-01
-1.94647044e-01 5.61319888e-01 -3.96109492e-01 -8.47082119e-03
3.11133355e-01 -6.13877773e-01 -3.51397157e-01 3.15745085e-01
3.71408351e-02 2.08949119e-01 -4.88718271e-01 1.29374349e+00
2.08244175e-01 3.48820895e-01 9.03131008e-01 -9.56355453e-01
-8.23831379e-01 1.52219459e-01 -8.94443214e-01 3.20374593e-02
4.33741570e-01 -3.18312287e-01 6.35708630e-01 -1.24348950e+00
4.84364301e-01 1.29969096e+00 4.56835210e-01 2.78768003e-01
-1.21120441e+00 -7.41859734e-01 2.10687462e-02 6.51906073e-01
-1.93455446e+00 -1.10909700e-01 1.15897703e+00 -4.24540669e-01
5.39053082e-01 4.42162186e-01 5.17904401e-01 4.98008907e-01
4.79524164e-03 8.23711872e-01 1.43259549e+00 -6.57658160e-01
1.37133792e-01 3.31360370e-01 3.87004256e-01 7.79705048e-01
1.22082248e-01 1.91087481e-02 -3.86771470e-01 -7.18317702e-02
7.83301055e-01 3.89331579e-01 -7.95910418e-01 -6.04965806e-01
-1.05101311e+00 8.21696520e-01 7.40259528e-01 3.85750800e-01
-4.63115960e-01 -4.07679439e-01 6.09970801e-02 -1.70027897e-01
1.80962980e-01 8.78518075e-02 1.95279077e-01 2.41447479e-01
-8.61751735e-01 -3.90777849e-02 4.69753742e-01 7.48557448e-01
9.87026691e-01 2.90559173e-01 7.81177431e-02 1.33336771e+00
4.54010069e-01 5.99799335e-01 9.47459161e-01 -6.57152712e-01
4.79354292e-01 1.05411530e+00 -1.38382599e-01 -1.51336813e+00
-1.33228987e-01 -6.37209296e-01 -1.28069460e+00 1.22733273e-01
3.71674597e-02 3.20440799e-01 -5.48565209e-01 1.25524569e+00
1.51943371e-01 5.06150544e-01 2.83640236e-01 1.04017782e+00
7.68479884e-01 9.88140166e-01 -2.22346440e-01 -5.70048988e-01
1.19693542e+00 -5.13548076e-01 -4.61501867e-01 -7.38316551e-02
1.29359499e-01 -6.86534643e-01 7.49594927e-01 4.73609954e-01
-5.88721633e-01 -8.63562405e-01 -1.21597433e+00 4.63281572e-01
-1.18485615e-01 7.29553819e-01 3.67064565e-01 4.48722959e-01
-4.76803213e-01 3.35966557e-01 -6.80199087e-01 -1.04175009e-01
3.42375994e-01 2.03192666e-01 -5.94710350e-01 -5.96368730e-01
-6.93689406e-01 8.06477189e-01 2.37554908e-01 4.11696583e-01
-2.97869384e-01 -4.14133698e-01 -8.75023425e-01 -4.73507494e-02
2.26556718e-01 -6.59367368e-02 3.90984833e-01 -9.27609801e-01
-1.36751533e+00 7.34103501e-01 -3.77194524e-01 -1.62504330e-01
-8.34802687e-02 1.19783446e-01 -4.75406080e-01 3.08318079e-01
6.70063496e-02 1.96942031e-01 1.10754216e+00 -1.21273148e+00
-2.95437247e-01 -7.41250157e-01 -6.01156712e-01 3.37210119e-01
-1.81617498e-01 -4.23319004e-02 -3.74275923e-01 -5.76013386e-01
1.02733505e+00 -8.57394814e-01 -1.94585860e-01 -1.23558544e-01
-2.86489844e-01 -3.61351639e-01 1.05459881e+00 -7.12466896e-01
1.09869599e+00 -2.37717795e+00 4.82156008e-01 5.49481809e-01
-1.00716181e-01 4.90959197e-01 2.96992157e-03 1.20755084e-01
-4.90944535e-01 -3.17233324e-01 -5.08120298e-01 -6.03858307e-02
-5.57287395e-01 2.59186849e-02 -1.77563831e-01 7.99295723e-01
7.09664822e-02 3.83888632e-01 -4.06961083e-01 -3.98232043e-01
4.28353548e-01 6.41098082e-01 -2.25839287e-01 1.92966118e-01
3.70637476e-01 3.47166091e-01 -5.63542247e-01 7.35061944e-01
9.59160089e-01 -9.02685225e-02 1.19709522e-01 -3.79584938e-01
-2.17185140e-01 -2.64520735e-01 -1.52495074e+00 1.48156929e+00
-2.68041819e-01 4.71278012e-01 -9.82338786e-02 -1.80260324e+00
1.53581548e+00 7.01376721e-02 5.29882371e-01 -5.99910855e-01
2.03520525e-02 2.49894604e-01 -1.46181166e-01 -2.97251672e-01
1.33734792e-01 -8.94812942e-02 2.69608557e-01 1.44280836e-01
-1.95487410e-01 2.68911928e-01 -1.98316455e-01 -1.06665321e-01
4.87864792e-01 -3.38093452e-02 5.76607287e-01 -2.97722429e-01
1.26812470e+00 -2.43304506e-01 7.86594748e-01 1.51455954e-01
-5.17323725e-02 6.83936119e-01 1.64422125e-01 -3.36639047e-01
-5.85115314e-01 -8.28710854e-01 -4.85455334e-01 2.17906132e-01
3.17140847e-01 -2.27567360e-01 -5.68144441e-01 -5.13160110e-01
-1.20685235e-01 5.69298327e-01 -4.03408796e-01 -2.50682026e-01
-6.25049114e-01 -7.26117373e-01 -1.50788933e-01 3.49405140e-01
9.68059957e-01 -1.01598978e+00 -2.84258634e-01 1.13550238e-01
-3.12372029e-01 -7.42689669e-01 -2.20616519e-01 -3.84587139e-01
-8.44888091e-01 -1.28087676e+00 -9.52299893e-01 -9.43449676e-01
1.15046632e+00 8.11758637e-01 2.34309986e-01 9.46100056e-03
-4.56183910e-01 3.76361132e-01 -4.35676545e-01 5.38335592e-02
5.35533838e-02 -3.74372572e-01 1.48205236e-01 7.67139256e-01
5.72827458e-01 -3.75695765e-01 -6.79852724e-01 5.59966803e-01
-5.33302128e-01 -1.07845888e-01 7.67312944e-01 8.84626627e-01
1.08351433e+00 2.63647407e-01 4.92441922e-01 -4.73759234e-01
4.32120293e-01 -4.48348641e-01 -7.34810412e-01 2.81306148e-01
-6.70448601e-01 -5.58627248e-02 6.46745265e-01 -4.33031708e-01
-9.63376999e-01 1.46628603e-01 7.11575747e-02 -9.11079466e-01
7.68446699e-02 7.29460359e-01 -3.01242352e-01 -2.63321251e-01
5.18706381e-01 1.03111780e+00 3.51162016e-01 -7.14942098e-01
2.86726337e-02 9.21379387e-01 4.50048178e-01 -3.70791167e-01
6.39451325e-01 3.28612119e-01 3.86428326e-01 -1.27908111e+00
-4.77428168e-01 -8.20621312e-01 -4.32729483e-01 -2.42232621e-01
6.23592496e-01 -8.30830336e-01 -6.60442770e-01 3.60423476e-01
-8.01951647e-01 4.87349391e-01 -1.50674239e-01 1.04807544e+00
-5.17634451e-01 6.36652231e-01 2.30405685e-02 -1.00773716e+00
-2.78014481e-01 -1.34256518e+00 6.76965654e-01 4.09486175e-01
2.19043836e-01 -3.94396871e-01 -6.26498833e-02 3.77238989e-01
1.22044414e-01 9.54335108e-02 8.36511672e-01 -4.10266459e-01
-6.42785668e-01 -5.67245603e-01 -5.07570922e-01 7.59176910e-01
1.56050948e-02 -2.43871152e-01 -5.63335657e-01 -3.79923880e-01
3.59040022e-01 2.33044121e-02 7.98061550e-01 4.21921462e-01
1.48631370e+00 -2.24580318e-01 -3.95649254e-01 7.88257420e-01
1.75726438e+00 5.08118629e-01 1.07731533e+00 1.76075324e-01
6.25687242e-01 5.35650849e-01 6.82991266e-01 3.28754842e-01
5.53755788e-03 8.89846444e-01 3.52635980e-02 2.35930562e-01
-9.09388363e-02 -1.44400448e-01 2.13642105e-01 7.76340127e-01
-3.38816881e-01 3.93408239e-01 -6.79503083e-01 8.96316841e-02
-1.81754482e+00 -1.14052117e+00 7.41003007e-02 2.50301123e+00
2.25755364e-01 -2.34319314e-01 -7.82388821e-02 5.65041363e-01
8.11613560e-01 1.76590979e-01 -3.56458217e-01 -5.06555997e-02
1.62447076e-02 1.98843613e-01 9.93777290e-02 2.23410577e-01
-8.58974814e-01 5.07104516e-01 5.38266516e+00 1.09720552e+00
-1.26653707e+00 -3.09380084e-01 4.33514416e-01 4.55256373e-01
1.06838956e-01 -1.23112544e-01 -1.00682044e+00 5.53826153e-01
4.66767430e-01 -8.30613598e-02 5.96301615e-01 1.03048360e+00
-2.03679856e-02 9.50521044e-03 -7.59743810e-01 1.62302291e+00
5.24704933e-01 -1.20478797e+00 3.85304034e-01 5.02259731e-02
4.56657082e-01 -4.83983904e-01 1.58130914e-01 4.48200315e-01
-5.27339458e-01 -9.04884517e-01 1.29232734e-01 9.76748228e-01
9.18303728e-01 -7.09336281e-01 6.08319700e-01 5.29768825e-01
-1.37947404e+00 -8.60042274e-02 -7.93214619e-01 2.49075860e-01
-3.92188519e-01 4.30288076e-01 -4.47878480e-01 6.57476723e-01
5.05230308e-01 1.12389052e+00 -4.88130569e-01 1.03442466e+00
-8.26818347e-02 3.45915020e-01 -1.28767550e-01 -1.13043226e-01
2.34075211e-04 -8.73095274e-01 6.65566802e-01 9.33004975e-01
5.17089844e-01 6.04975820e-01 2.10799828e-01 8.51438284e-01
2.16960132e-01 4.57660556e-01 -5.42243242e-01 2.13505074e-01
4.19364870e-01 1.12823427e+00 -4.92347270e-01 -1.90602034e-01
-3.71928960e-01 7.00772703e-01 3.86003047e-01 3.90020221e-01
-3.84860635e-01 -5.11981964e-01 3.75133902e-01 1.53250219e-02
4.70494568e-01 -1.72219187e-01 -7.71687105e-02 -1.30360103e+00
1.25175044e-01 -8.63544941e-01 3.44869971e-01 -7.08171248e-01
-1.10685301e+00 5.39229751e-01 -6.74056762e-04 -1.82728732e+00
-7.89981522e-03 -5.85943520e-01 -5.59485137e-01 1.21239567e+00
-1.39498544e+00 -9.08400714e-01 -3.20880324e-01 7.99774289e-01
6.70338988e-01 -5.75257778e-01 9.76077676e-01 -1.79828536e-02
-8.20185065e-01 4.52405870e-01 2.56624132e-01 6.71050698e-02
6.51776642e-02 -6.81197524e-01 -6.88958704e-01 8.07214379e-01
6.07658029e-02 7.85894036e-01 2.37973168e-01 -3.52024466e-01
-1.49120402e+00 -9.22266781e-01 6.18116021e-01 3.01790565e-01
1.28360718e-01 3.19697946e-01 -1.08017707e+00 5.06409824e-01
-4.62418944e-01 9.95112136e-02 7.35655427e-01 -2.07159109e-02
-5.65531194e-01 -7.13399291e-01 -1.11493111e+00 3.47569108e-01
5.30707181e-01 -5.17269671e-01 -6.97458208e-01 2.41214529e-01
1.33881912e-01 -1.86674640e-01 -8.39953244e-01 4.88065392e-01
5.08589923e-01 -8.18829775e-01 1.39472663e+00 -9.09896940e-02
2.48784363e-01 -5.34377933e-01 -5.78321517e-01 -1.09507501e+00
-3.94598782e-01 7.12128133e-02 9.29945111e-02 9.90957677e-01
1.08855134e-02 -7.53717780e-01 6.97843850e-01 3.07071298e-01
-2.56476793e-02 -1.07857072e+00 -1.12896752e+00 -7.26216257e-01
-2.91068673e-01 -2.94340909e-01 4.08604711e-01 9.23080504e-01
2.17221186e-01 2.87894517e-01 -1.89677954e-01 2.43739814e-01
9.36153054e-01 3.74684095e-01 5.30697405e-01 -1.18692601e+00
-1.81233734e-01 -3.08961213e-01 -8.54671001e-01 -9.51806366e-01
2.56199360e-01 -9.82542872e-01 -8.06033090e-02 -1.45065153e+00
1.96959615e-01 -6.05636716e-01 -5.45795321e-01 1.13549307e-01
-6.73046932e-02 3.50999869e-02 2.86285371e-01 9.65444565e-01
-2.96287417e-01 7.91112661e-01 1.21788681e+00 -2.24761471e-01
-2.50460505e-01 3.69035989e-01 -6.45635366e-01 6.92346752e-01
5.69755733e-01 -1.68647066e-01 -2.51860261e-01 3.72882597e-02
-5.30200839e-01 3.72774541e-01 2.20433041e-01 -1.22110808e+00
1.40804961e-01 -3.29059642e-03 4.91317302e-01 -7.02522218e-01
6.20702863e-01 -8.88896585e-01 1.88816234e-01 4.84116465e-01
-2.02684730e-01 -4.81599778e-01 -2.04331353e-01 8.14410388e-01
-4.61758137e-01 -3.15025866e-01 1.04419839e+00 -1.11088574e-01
-6.76406920e-01 2.21979514e-01 -1.75785884e-01 -4.91044641e-01
1.14072859e+00 -4.47117388e-01 -9.23008285e-03 -6.25662506e-02
-5.23362815e-01 -1.85128376e-01 -5.84012307e-02 3.99971493e-02
1.29413044e+00 -1.42546475e+00 -7.22644329e-01 7.11629987e-01
2.09473610e-01 -2.12759048e-01 3.21032226e-01 7.22884417e-01
-4.53395456e-01 5.65300047e-01 -1.16140999e-01 -7.33359694e-01
-1.38200521e+00 6.20465457e-01 3.07911843e-01 -1.07541673e-01
-8.34378958e-01 4.66216058e-01 4.40111756e-02 -2.69967675e-01
1.05357997e-01 2.07284372e-02 -7.39978135e-01 -1.61198109e-01
7.32162952e-01 6.59560859e-01 -2.63227880e-01 -1.08482385e+00
-3.62427264e-01 1.03775287e+00 -7.76510462e-02 5.71134835e-02
1.13211095e+00 1.41847938e-01 -2.54805982e-01 1.99565634e-01
1.64993417e+00 -2.84584582e-01 -8.61269057e-01 -5.73660374e-01
-2.91241944e-01 -8.55474651e-01 2.42150471e-01 -4.78682131e-01
-1.16379881e+00 7.86775589e-01 7.63522744e-01 -7.18041286e-02
1.26179898e+00 -2.34203145e-01 4.33186829e-01 3.02387536e-01
5.25529981e-01 -7.32158363e-01 -1.63908470e-02 -8.09331536e-02
1.25138903e+00 -1.05813527e+00 2.93546230e-01 -7.53015876e-01
-5.89403868e-01 1.14137685e+00 5.58776438e-01 -4.85337347e-01
9.58302677e-01 -6.21134222e-01 -3.42625707e-01 -7.27797076e-02
-1.14408545e-01 1.00308172e-01 5.82985640e-01 5.56995571e-01
1.82000250e-01 2.39941999e-01 -6.16701663e-01 6.53987885e-01
3.94304795e-03 -5.03154062e-02 7.81600028e-02 7.66875625e-01
-5.59699535e-01 -8.74652743e-01 -7.99102187e-01 4.74067509e-01
5.28841428e-02 2.52886564e-01 8.69407505e-02 5.54125011e-01
3.53676383e-03 7.94387817e-01 1.83684807e-02 -4.88325298e-01
2.61763304e-01 -1.23365559e-01 3.46825033e-01 -5.44457614e-01
2.04331830e-01 1.13546528e-01 -4.67542022e-01 -6.21240914e-01
-4.77242827e-01 -8.71019244e-01 -1.20860040e+00 8.09768029e-03
-4.18464035e-01 3.48302901e-01 7.48076081e-01 7.92426944e-01
1.53271824e-01 5.92292286e-02 1.08837485e+00 -7.00158477e-01
-6.37175441e-01 -8.63090336e-01 -1.00094187e+00 2.83571243e-01
-1.44195780e-01 -7.50945747e-01 -5.97474456e-01 -1.31510839e-01] | [12.47513198852539, 0.4217372536659241] |
38faaf79-0268-4de2-a9a6-9e1b1bf05f61 | cooperative-decision-making-in-shared-spaces | 2306.14617 | null | https://arxiv.org/abs/2306.14617v1 | https://arxiv.org/pdf/2306.14617v1.pdf | Cooperative Decision-Making in Shared Spaces: Making Urban Traffic Safer through Human-Machine Cooperation | In this paper, a cooperative decision-making is presented, which is suitable for intention-aware automated vehicle functions. With an increasing number of highly automated and autonomous vehicles on public roads, trust is a very important issue regarding their acceptance in our society. The most challenging scenarios arise at low driving speeds of these highly automated and autonomous vehicles, where interactions with vulnerable road users likely occur. Such interactions must be addressed by the automation of the vehicle. The novelties of this paper are the adaptation of a general cooperative and shared control framework to this novel use case and the application of an explicit prediction model of the pedestrian. An extensive comparison with state-of-the-art algorithms is provided in a simplified test environment. The results show the superiority of the proposed model-based algorithm compared to state-of-the-art solutions and its suitability for real-world applications due to its real-time capability. | ['Sören Hohmann', 'Dongxu Yang', 'Balint Varga'] | 2023-06-26 | null | null | null | null | ['autonomous-vehicles', 'decision-making'] | ['computer-vision', 'reasoning'] | [-1.79766223e-01 5.61758041e-01 -2.22284794e-01 -7.17433691e-01
1.43553400e-02 1.74268886e-01 9.37772155e-01 1.18313715e-01
-6.89176142e-01 1.10060394e+00 -4.02208090e-01 -5.39997339e-01
-1.59538835e-01 -8.62423599e-01 -4.14636850e-01 -5.51537514e-01
-1.26537710e-01 7.59557188e-01 8.39289248e-01 -6.77370369e-01
8.32085535e-02 5.90524495e-01 -2.33211565e+00 -2.96609938e-01
1.14505899e+00 7.81307995e-01 3.71140033e-01 5.09738624e-01
2.48736620e-01 5.58407724e-01 -4.67640795e-02 -6.53466523e-01
5.67365587e-02 1.92417100e-01 -5.72806954e-01 1.66925803e-01
-7.65789971e-02 -5.07099390e-01 1.76502988e-01 8.83337855e-01
2.02485621e-01 1.76033169e-01 6.42647028e-01 -2.00648046e+00
2.03928664e-01 1.57546312e-01 -3.00305355e-02 -2.83528894e-01
2.36427978e-01 2.40755066e-01 5.74697375e-01 -5.84123492e-01
7.06528127e-01 1.05202889e+00 3.60927373e-01 2.88803309e-01
-9.73480701e-01 -6.66698098e-01 3.81618410e-01 1.18100643e+00
-1.65757823e+00 -4.59081799e-01 6.50072515e-01 -4.62375522e-01
7.38100350e-01 2.14213654e-01 7.64935672e-01 7.67192602e-01
5.13738632e-01 8.29806507e-01 1.07144690e+00 -3.10772240e-01
7.21635580e-01 8.32694530e-01 2.98132479e-01 2.38689929e-01
8.47925901e-01 2.89797992e-01 -7.60687562e-03 6.72642887e-03
-1.93318307e-01 -1.02104232e-01 1.63613260e-01 -8.56911123e-01
-7.30575621e-01 9.51898277e-01 8.00714716e-02 2.53825575e-01
-9.59301531e-01 -1.15015805e-01 4.22385871e-01 9.88598913e-02
3.05381358e-01 -5.07130325e-01 -1.75831422e-01 -3.57876748e-01
-7.12734342e-01 4.44838732e-01 1.06832004e+00 8.75087678e-01
9.58268225e-01 -1.70478423e-03 2.18797460e-01 2.53236294e-01
9.77783263e-01 4.11754608e-01 -1.68645695e-01 -8.77109170e-01
4.18622531e-02 5.56497574e-01 4.67756629e-01 -1.09486985e+00
-3.57222378e-01 -2.95003831e-01 -6.74173117e-01 8.90536129e-01
1.31164506e-01 -2.90927500e-01 -3.23234051e-01 1.21389008e+00
7.42437899e-01 2.44669214e-01 4.78142023e-01 6.04382217e-01
2.70763814e-01 4.29363966e-01 4.23285425e-01 -4.64012146e-01
1.03392839e+00 -7.72176743e-01 -1.12546277e+00 -2.22484127e-01
4.58316505e-01 -6.91452503e-01 2.64476687e-01 3.67399812e-01
-6.71757400e-01 -5.71706057e-01 -1.18330801e+00 4.87732112e-01
-4.71270680e-01 -1.49801448e-01 3.97909045e-01 1.05845070e+00
-9.91783381e-01 1.62442446e-01 -4.21909720e-01 -9.24918652e-01
2.21250132e-01 5.80839276e-01 -2.49248445e-01 -4.68151160e-02
-1.29807615e+00 1.62637699e+00 8.65985081e-02 1.85629785e-01
-7.71510422e-01 -1.89106613e-01 -6.69731736e-01 -2.57978261e-01
3.93268704e-01 -2.71053702e-01 1.29362190e+00 -9.34872627e-01
-1.60330069e+00 6.44565403e-01 -1.89101294e-01 -9.56085384e-01
1.08924735e+00 -1.75174311e-01 -6.79096043e-01 -8.41274187e-02
1.37127697e-01 3.59483987e-01 6.12493992e-01 -1.64951563e+00
-1.26302695e+00 -1.65227111e-02 -8.63617286e-03 9.69172865e-02
-1.09883714e-02 -8.64357501e-02 -9.80852544e-02 3.56833041e-01
-6.53561532e-01 -9.85681891e-01 -5.39630115e-01 -3.45235504e-02
3.31156440e-02 -4.64631706e-01 1.38254797e+00 -3.34072202e-01
1.11582792e+00 -1.81515801e+00 -2.76581496e-01 5.15120089e-01
-8.66713002e-02 5.65271199e-01 2.53356367e-01 7.14835525e-01
3.73586088e-01 -2.05340520e-01 2.55641699e-01 -3.57075125e-01
2.70830899e-01 5.37763178e-01 1.76460430e-01 6.39735401e-01
-4.04552259e-02 3.49708050e-01 -6.31951749e-01 -7.55680144e-01
8.51148248e-01 5.47733665e-01 -1.03716232e-01 2.23298371e-01
5.83033524e-02 2.47873768e-01 -7.71514714e-01 2.24801242e-01
8.96650016e-01 4.90314275e-01 2.75450975e-01 3.36584859e-02
-6.93872988e-01 -3.80338728e-01 -1.37395537e+00 6.35555744e-01
-4.67914760e-01 5.46814859e-01 4.10064876e-01 -8.92459512e-01
9.34060574e-01 5.74006319e-01 3.23288530e-01 -6.18989885e-01
5.18122315e-01 3.90098333e-01 3.17238688e-01 -5.70459366e-01
5.81776142e-01 3.18645984e-01 1.29449740e-01 2.13066101e-01
-4.48094726e-01 -1.07894167e-01 2.11647227e-01 1.62761316e-01
6.18095517e-01 2.11007446e-02 6.18798971e-01 -4.43564355e-01
1.19872499e+00 1.02631859e-01 6.11670613e-01 4.06982422e-01
-9.83617783e-01 -3.65212291e-01 4.29894701e-02 -4.56545055e-01
-8.94774020e-01 -3.85209888e-01 -1.24776460e-01 5.35566211e-01
5.14541209e-01 2.30420064e-02 -7.69957483e-01 -5.85154772e-01
8.55262503e-02 1.34321642e+00 -4.54159498e-01 -3.66141051e-02
-2.46360555e-01 -1.29361480e-01 -8.42439756e-02 -2.96331439e-02
8.02426457e-01 -6.76609337e-01 -1.00756884e+00 4.42839682e-01
-2.41719317e-02 -1.45721674e+00 4.39301908e-01 -3.09155375e-01
-2.71462321e-01 -1.03223097e+00 -1.36504993e-01 -5.51648855e-01
4.39523220e-01 4.86130178e-01 6.19079649e-01 3.51870328e-01
1.47934064e-01 5.22554100e-01 -4.70888823e-01 -9.11979258e-01
-9.62635577e-01 -8.51357635e-03 1.47071287e-01 7.16973364e-01
6.61353230e-01 -2.87785143e-01 -4.84393358e-01 8.71633828e-01
-1.94342747e-01 1.20371580e-01 5.41536331e-01 2.99198389e-01
1.51764065e-01 4.18975860e-01 8.38896036e-01 -6.31205380e-01
4.47682798e-01 -7.11784363e-01 -8.09620142e-01 6.28792942e-02
-8.97367179e-01 -4.11470443e-01 3.95863533e-01 -7.58225843e-02
-1.26350403e+00 2.40236789e-01 -1.58480659e-01 2.22014293e-01
-5.59467256e-01 7.00050443e-02 -4.55280542e-01 -3.09229076e-01
2.51603723e-01 -1.63306534e-01 6.72422230e-01 1.47617787e-01
2.88618267e-01 1.15056920e+00 -6.30479446e-03 -9.89083648e-02
8.04847360e-01 4.19195056e-01 2.99229562e-01 -1.01561069e+00
1.30834505e-01 -5.37635624e-01 -5.82103670e-01 -1.02038240e+00
8.33874822e-01 -8.90375316e-01 -1.11528707e+00 5.60530245e-01
-1.14499199e+00 -2.67481744e-01 -6.20709620e-02 5.73583484e-01
-7.69917250e-01 3.79979491e-01 2.17068344e-01 -1.44898248e+00
-6.14554323e-02 -1.45082390e+00 3.79926652e-01 2.86355168e-01
-3.27260733e-01 -9.44760025e-01 -3.54290128e-01 6.54751539e-01
8.09261262e-01 1.29169347e-02 4.18833584e-01 -8.73956859e-01
-6.39411867e-01 -6.12003326e-01 4.96410392e-02 3.32551897e-01
-1.59590259e-01 3.87168646e-01 -9.08146620e-01 -1.43753784e-02
-3.24755877e-01 3.25068012e-02 8.36740658e-02 1.84944004e-01
1.92654625e-01 -5.04852459e-03 -6.98704660e-01 -5.93291581e-01
1.25236416e+00 3.98326665e-01 7.32121885e-01 5.84228277e-01
7.93558657e-02 1.13907063e+00 1.49549699e+00 5.93166411e-01
1.06470037e+00 9.57970560e-01 9.94421661e-01 -2.45815814e-02
1.21294416e-01 1.73210993e-01 3.75978619e-01 2.92111665e-01
-2.66852409e-01 -2.68895417e-01 -9.45025027e-01 7.00693429e-01
-2.61579990e+00 -1.06322372e+00 -7.16401994e-01 2.32920742e+00
-2.58684088e-03 2.85282314e-01 2.31689110e-01 4.85599697e-01
9.62239385e-01 -2.66915321e-01 -2.01071084e-01 -8.42047453e-01
2.51993418e-01 -4.97307241e-01 7.09417760e-01 7.84824848e-01
-1.10481846e+00 8.19055319e-01 5.95594740e+00 7.25228786e-01
-8.28647494e-01 4.67525661e-01 4.10486460e-01 5.56572318e-01
7.88810775e-02 3.03170625e-02 -8.13115895e-01 3.75791192e-01
1.36244297e+00 -4.19849068e-01 3.90377045e-02 9.52053189e-01
9.34225380e-01 -6.07276261e-01 -7.86873877e-01 4.62812811e-01
-9.90489945e-02 -9.60197449e-01 -3.81417334e-01 2.94180930e-01
5.36066294e-01 1.06378561e-02 -2.75137842e-01 1.05863109e-01
3.22306693e-01 -5.96395195e-01 9.49953258e-01 6.39749169e-01
-3.86415944e-02 -1.11189842e+00 1.28663051e+00 5.54530323e-01
-1.00170410e+00 -2.10689247e-01 -7.33531192e-02 -3.02448273e-01
6.43339097e-01 3.93955261e-01 -9.07262206e-01 5.74324906e-01
4.88960803e-01 5.52088678e-01 -3.81093115e-01 1.14906156e+00
-2.66103715e-01 4.77255195e-01 -3.48249882e-01 -7.23460793e-01
2.38765597e-01 -2.62885869e-01 5.91013551e-01 9.60913837e-01
1.09778941e-01 -1.90503970e-01 1.24122255e-01 3.63777697e-01
8.10302734e-01 4.03381675e-01 -1.03293884e+00 5.43959498e-01
4.83916789e-01 1.43467295e+00 -3.34970415e-01 -3.75043184e-01
-7.36394882e-01 4.34683591e-01 -7.62869269e-02 1.48730680e-01
-8.77815485e-01 -1.43624231e-01 9.48382378e-01 4.83210385e-01
4.02814895e-01 -1.60798043e-01 -1.96332872e-01 -2.54525870e-01
-6.39572442e-02 -4.73990440e-01 -2.02757195e-01 -4.94059682e-01
-7.49265313e-01 7.15397239e-01 3.24965239e-01 -1.60968685e+00
-4.83195812e-01 -4.14512634e-01 -6.83789253e-01 4.52121705e-01
-1.76627088e+00 -1.51123881e+00 -2.38965765e-01 4.48147953e-01
3.64636183e-01 -5.79861403e-01 6.45994067e-01 4.08975661e-01
-4.53594923e-01 2.11624160e-01 2.30223369e-02 -7.28261352e-01
3.95591617e-01 -5.69038808e-01 -4.45148535e-02 9.89786088e-01
-9.30270374e-01 2.15402041e-02 1.31479740e+00 -7.59990811e-01
-1.42246723e+00 -1.05114710e+00 1.34218478e+00 -7.83795863e-02
6.80573821e-01 -1.71449885e-01 -5.35956502e-01 4.13769066e-01
6.18568659e-01 -5.18866368e-02 4.99599010e-01 -9.91288945e-02
4.03776854e-01 -5.27055621e-01 -1.59585893e+00 6.94837570e-01
5.72266102e-01 1.05160579e-01 -4.01965052e-01 9.24267396e-02
1.71230465e-01 3.43335569e-01 -4.32013363e-01 2.88183779e-01
5.75012088e-01 -1.21715498e+00 4.42293197e-01 -7.76704866e-03
-4.01372343e-01 -5.21802604e-01 -2.21572250e-01 -1.00878668e+00
-1.00653037e-01 -6.98451459e-01 2.16495052e-01 1.17592061e+00
4.19041395e-01 -1.17742062e+00 7.03379273e-01 1.10410547e+00
6.87213615e-02 -3.80683899e-01 -1.45846486e+00 -5.02273619e-01
-2.91217774e-01 -6.54788733e-01 3.54267627e-01 2.72706300e-01
4.33271118e-02 1.35219455e-01 -4.82462525e-01 3.92790586e-01
9.09506321e-01 -4.48956847e-01 1.30853903e+00 -1.68458915e+00
4.06023085e-01 -1.03530064e-01 -1.14431012e+00 -3.42429012e-01
3.79915684e-01 -1.63524717e-01 1.42471254e-01 -1.58289742e+00
-1.92917913e-01 -4.88156676e-01 1.77058950e-01 1.39738098e-01
2.82394797e-01 -1.95861012e-02 2.10815206e-01 -1.05640307e-01
-6.90710962e-01 5.75072169e-01 7.13034809e-01 -1.65144317e-02
4.66420799e-02 4.83275205e-01 -3.87811840e-01 8.60499620e-01
9.50508177e-01 -2.86448628e-01 -5.04254997e-01 3.34304422e-01
1.24856399e-03 -1.22623667e-01 3.47949177e-01 -1.22504604e+00
6.61272645e-01 -3.36030573e-01 -6.16553426e-01 -7.71434128e-01
5.14244318e-01 -1.74581099e+00 5.97304285e-01 8.50422621e-01
3.12365979e-01 -2.27631494e-01 -7.70789431e-03 6.64222658e-01
-2.06111357e-01 -2.41719410e-01 8.44601631e-01 4.48000222e-01
-1.10674393e+00 -1.24611005e-01 -1.32219172e+00 -9.47398245e-01
2.30679727e+00 -6.93704605e-01 -1.45618320e-01 -7.29347646e-01
-3.99746746e-01 6.69985175e-01 4.08671618e-01 6.17227197e-01
4.65488136e-01 -1.01119018e+00 -7.78595626e-01 1.08071566e-01
2.19686404e-01 -5.23654103e-01 3.92154157e-01 9.52604175e-01
-4.62029725e-01 3.74103963e-01 -5.95925450e-01 -5.09580016e-01
-1.59204018e+00 4.73464280e-01 1.31759569e-01 1.48913339e-01
-2.47002169e-01 -1.89134344e-01 -2.78749287e-01 -1.94384336e-01
1.91236123e-01 1.20772004e-01 -5.95326781e-01 -1.40965236e-02
4.33820575e-01 8.20578098e-01 -4.27509286e-03 -1.43213892e+00
-5.65669954e-01 4.82453287e-01 -3.19445995e-03 -7.58671165e-02
1.00486171e+00 -8.00185502e-01 8.90625939e-02 3.82954985e-01
5.52558184e-01 -1.86720550e-01 -1.11322510e+00 4.51899432e-02
-1.06794819e-01 -2.98783690e-01 3.33644778e-01 -5.87529123e-01
-7.06574261e-01 5.58608949e-01 9.01172161e-01 2.99864769e-01
5.57453632e-01 -3.17926675e-01 3.58154744e-01 3.81758541e-01
1.06772697e+00 -1.43067610e+00 -6.51877522e-01 2.97671676e-01
8.28614771e-01 -1.53350389e+00 1.21217348e-01 -7.63084829e-01
-1.02404368e+00 8.99661243e-01 6.49275959e-01 1.63855717e-01
1.09045577e+00 7.36703947e-02 2.53489614e-01 2.41263866e-01
-8.66233766e-01 -5.25726497e-01 -2.09428012e-01 1.02558160e+00
-1.48496106e-01 3.10564458e-01 -9.79882538e-01 3.55541676e-01
2.57627308e-01 2.41141662e-01 7.88623869e-01 1.07691908e+00
-7.70998955e-01 -1.25151956e+00 -3.91796559e-01 1.58255752e-02
-7.23344535e-02 8.41056168e-01 -5.60903661e-02 1.07949340e+00
4.18067247e-01 1.81526756e+00 -1.16298243e-01 -4.17648524e-01
5.13199747e-01 -2.04674341e-02 -3.00440848e-01 -1.47895306e-01
-5.03248274e-01 -3.10061723e-01 6.62211239e-01 -4.88422364e-01
-8.78569722e-01 -1.07612085e+00 -1.30998099e+00 -5.83013833e-01
-3.72716784e-01 3.67610395e-01 1.16060591e+00 9.64252710e-01
5.82617939e-01 1.51187450e-01 8.64795327e-01 -9.87993360e-01
-2.77059406e-01 -6.41399741e-01 -4.43797261e-01 9.26930755e-02
1.20846525e-01 -9.29368138e-01 -3.09967220e-01 -2.79506236e-01] | [5.6112260818481445, 1.315854549407959] |
1aa624a3-1e8b-4f04-be93-ba9c63e31ee6 | plug-and-play-vqa-zero-shot-vqa-by-conjoining | 2210.08773 | null | https://arxiv.org/abs/2210.08773v3 | https://arxiv.org/pdf/2210.08773v3.pdf | Plug-and-Play VQA: Zero-shot VQA by Conjoining Large Pretrained Models with Zero Training | Visual question answering (VQA) is a hallmark of vision and language reasoning and a challenging task under the zero-shot setting. We propose Plug-and-Play VQA (PNP-VQA), a modular framework for zero-shot VQA. In contrast to most existing works, which require substantial adaptation of pretrained language models (PLMs) for the vision modality, PNP-VQA requires no additional training of the PLMs. Instead, we propose to use natural language and network interpretation as an intermediate representation that glues pretrained models together. We first generate question-guided informative image captions, and pass the captions to a PLM as context for question answering. Surpassing end-to-end trained baselines, PNP-VQA achieves state-of-the-art results on zero-shot VQAv2 and GQA. With 11B parameters, it outperforms the 80B-parameter Flamingo model by 8.5% on VQAv2. With 738M PLM parameters, PNP-VQA achieves an improvement of 9.1% on GQA over FewVLM with 740M PLM parameters. Code is released at https://github.com/salesforce/LAVIS/tree/main/projects/pnp-vqa | ['Steven C. H. Hoi', 'Silvio Savarese', 'Boyang Li', 'Junnan Li', 'Anthony Meng Huat Tiong'] | 2022-10-17 | null | null | null | null | ['network-interpretation'] | ['computer-vision'] | [-2.41986457e-02 3.89793605e-01 -1.25877529e-01 -3.76703173e-01
-1.22629535e+00 -5.93855381e-01 7.13549495e-01 -3.37271780e-01
-3.89515817e-01 4.12098289e-01 3.19744498e-01 -5.55809021e-01
4.17133123e-01 -6.76162899e-01 -9.28829253e-01 -2.52239525e-01
6.94386125e-01 5.73536992e-01 5.03936768e-01 -4.13609326e-01
-1.81424484e-01 -5.61809689e-02 -1.34428799e+00 6.46432579e-01
7.63382554e-01 1.01082504e+00 4.64197457e-01 1.13622403e+00
-4.61211860e-01 1.27336788e+00 -3.23862582e-01 -8.66485715e-01
2.36598060e-01 -5.21206856e-01 -1.27137387e+00 1.32838815e-01
1.07901084e+00 -7.61224985e-01 -7.03926921e-01 8.63289356e-01
3.86800259e-01 1.44268334e-01 5.17931581e-01 -1.55664921e+00
-1.28761733e+00 1.15507826e-01 -5.57894230e-01 2.68403709e-01
2.56912291e-01 1.00149214e+00 1.37920761e+00 -1.13663661e+00
8.44944596e-01 1.52444005e+00 2.93303251e-01 1.01440728e+00
-1.15928066e+00 -3.62242937e-01 1.72804192e-01 4.88351613e-01
-1.01635492e+00 -5.20024240e-01 4.66009289e-01 -3.99381936e-01
1.14859056e+00 3.52082625e-02 5.61833739e-01 1.24304748e+00
-1.14759684e-01 1.29519534e+00 6.85886681e-01 -1.31089926e-01
4.63744551e-02 -2.18862221e-01 3.04283142e-01 9.16882753e-01
-6.11966243e-03 -1.94031626e-01 -4.31901038e-01 -6.78772340e-03
6.82681859e-01 -1.31029740e-01 -2.78248399e-01 -2.87320346e-01
-9.79397774e-01 9.80016232e-01 7.63218343e-01 -1.58814117e-01
-4.19365883e-01 7.77245939e-01 3.55324537e-01 2.09901214e-01
1.01311840e-01 3.31010073e-01 -3.71872962e-01 -1.87375955e-02
-7.01711595e-01 3.49982083e-01 4.12007123e-01 1.15415072e+00
6.90108001e-01 1.47565663e-01 -9.76176322e-01 7.44463444e-01
5.08578360e-01 8.14064443e-01 -1.05830543e-01 -1.54171228e+00
5.95153391e-01 5.55197239e-01 2.01938778e-01 -3.03707987e-01
1.91592351e-02 3.58840427e-03 -6.03804052e-01 2.28779584e-01
5.07672906e-01 5.21848071e-03 -1.57516468e+00 1.65221930e+00
1.94108754e-01 -2.95865014e-02 2.81457156e-01 1.11040545e+00
1.50916636e+00 1.09634805e+00 5.51932514e-01 3.13593537e-01
1.70382416e+00 -1.71037197e+00 -6.29998982e-01 -5.89005351e-01
3.20061207e-01 -5.20851910e-01 1.82638657e+00 -2.94403564e-02
-1.42425060e+00 -6.02448642e-01 -6.58823073e-01 -8.85880232e-01
-4.30705808e-02 -1.50130779e-01 3.33498269e-01 2.35359713e-01
-1.52320600e+00 -1.53231695e-01 -4.62727189e-01 -3.46726120e-01
9.18576717e-01 -1.48472875e-01 -2.32813001e-01 -6.10001802e-01
-1.09481514e+00 7.06534863e-01 9.34161618e-02 -8.03445429e-02
-1.73038936e+00 -9.63356018e-01 -1.07979012e+00 8.61149207e-02
5.67820191e-01 -1.33910406e+00 1.78372264e+00 -9.67858016e-01
-1.25523055e+00 1.10925829e+00 -4.22400564e-01 -6.19231939e-01
5.69635034e-01 -2.21894994e-01 2.58736257e-02 6.66132331e-01
2.58908480e-01 1.37716758e+00 8.38048220e-01 -1.24151003e+00
-4.38659430e-01 -6.33365735e-02 5.13575435e-01 1.73343390e-01
4.03249413e-01 -2.66490001e-02 -9.56124127e-01 -1.11930393e-01
-3.91894490e-01 -5.53953528e-01 -1.78942204e-01 6.24728143e-01
-1.65879518e-01 -5.20012438e-01 6.92040443e-01 -8.89041722e-01
4.33086753e-01 -1.99252665e+00 -2.49832459e-02 -4.59433734e-01
4.21088815e-01 5.32317698e-01 -8.78774524e-01 2.92575657e-01
4.03898329e-01 -1.56072438e-01 -5.25982082e-01 -6.04986370e-01
1.45993367e-01 5.03120661e-01 -6.38145328e-01 1.12892412e-01
4.91715521e-01 1.94775522e+00 -1.04285955e+00 -5.43995857e-01
2.90468931e-01 5.95250249e-01 -5.45693219e-01 3.74735773e-01
-8.97985995e-01 1.75459400e-01 -2.16025546e-01 8.14093173e-01
5.01289845e-01 -7.22808778e-01 -2.17826962e-01 -6.80744499e-02
3.20174366e-01 8.23754445e-02 -4.02584463e-01 1.95916593e+00
-4.10514593e-01 8.62453640e-01 -6.67979717e-02 -5.62321663e-01
6.35191500e-01 2.81263322e-01 2.90727131e-02 -9.54352379e-01
2.95940429e-01 -2.21686155e-01 -2.59065270e-01 -6.99331880e-01
3.73720080e-01 2.41142526e-01 3.55412364e-02 1.07583135e-01
6.31487072e-01 -2.18311295e-01 2.29200915e-01 8.09600294e-01
1.03600550e+00 2.62696564e-01 2.77035475e-01 1.57724172e-01
4.33931023e-01 3.00588198e-02 1.89985976e-01 9.76089180e-01
-6.31691456e-01 9.16135669e-01 6.80465996e-01 -2.64199495e-01
-1.24602485e+00 -1.45721865e+00 3.72455359e-01 1.31862056e+00
2.14587867e-01 -4.05185163e-01 -6.81623101e-01 -7.67437994e-01
-8.80393460e-02 1.12616134e+00 -6.70089483e-01 6.10064566e-02
-1.95132211e-01 -7.53195360e-02 6.87348008e-01 6.34242296e-01
6.82875812e-01 -1.48635590e+00 -5.60973763e-01 -4.93222326e-02
-4.51132268e-01 -1.34841180e+00 -6.02501571e-01 -4.38342869e-01
-5.26227057e-01 -1.11370134e+00 -8.95307779e-01 -7.50402570e-01
3.63951951e-01 6.28899813e-01 1.69987833e+00 2.01461539e-01
-2.63461590e-01 8.81304681e-01 -4.87286925e-01 -3.66536140e-01
-3.49720895e-01 -1.96757346e-01 -7.86440551e-01 -1.35117710e-01
2.76024014e-01 -2.75656492e-01 -9.91832495e-01 -6.32277876e-02
-8.09356332e-01 2.41382837e-01 4.65320051e-01 7.37840116e-01
7.78867543e-01 -1.17140841e+00 4.98945713e-01 -5.99379241e-01
2.75339425e-01 -3.67284745e-01 -6.38562262e-01 6.02142215e-01
-2.65322834e-01 1.89957172e-02 2.80479908e-01 -1.68906271e-01
-9.58538949e-01 -8.11688788e-03 -4.17196184e-01 -9.36874568e-01
-7.41573721e-02 -2.08141908e-01 -4.87721294e-01 1.69247136e-01
7.14366794e-01 3.93169045e-01 -1.87135173e-03 -1.71205416e-01
1.25006390e+00 4.11632031e-01 8.78573239e-01 -4.09046173e-01
7.11923420e-01 5.45068562e-01 -4.83878344e-01 -8.19124639e-01
-1.09764075e+00 -5.99131167e-01 -4.08252850e-02 -1.46989107e-01
1.55635738e+00 -1.05333543e+00 -9.19072688e-01 3.45715195e-01
-1.58650100e+00 -8.52161527e-01 -4.50376511e-01 -2.36107677e-01
-8.18264008e-01 4.25568968e-01 -6.23841286e-01 -8.00059438e-01
-9.33178842e-01 -1.07394385e+00 1.26612854e+00 3.60578507e-01
3.37801650e-02 -6.84565425e-01 -1.44907594e-01 1.02358603e+00
2.49246553e-01 -1.39310181e-01 7.23412752e-01 -2.75616616e-01
-9.79316533e-01 3.22585821e-01 -7.67471313e-01 4.77915972e-01
-3.48912835e-01 -2.98910320e-01 -1.17062807e+00 -2.58688688e-01
-2.31330365e-01 -8.85893345e-01 1.18095124e+00 4.30307627e-01
1.19823420e+00 -3.04212779e-01 1.51843235e-01 8.06617200e-01
1.42930448e+00 -6.38140142e-02 9.96545851e-01 1.31801352e-01
9.57981408e-01 3.97433132e-01 6.12594903e-01 8.71459991e-02
8.44953418e-01 3.35157365e-01 9.63070393e-01 -1.10737085e-01
-7.21983433e-01 -5.37532508e-01 3.99271250e-01 1.20509230e-01
4.66524549e-02 -4.87424195e-01 -1.05255604e+00 1.04825950e+00
-1.86763251e+00 -9.59655404e-01 -2.33217224e-01 1.52029026e+00
7.02796638e-01 -2.37933263e-01 9.47255492e-02 -6.72015429e-01
3.02302748e-01 6.38142109e-01 -8.04081678e-01 -3.17625910e-01
-1.89272732e-01 1.42011851e-01 2.84784794e-01 9.09053147e-01
-7.94629872e-01 1.44891119e+00 5.47694397e+00 6.17379844e-01
-7.33374238e-01 6.12143815e-01 6.21581495e-01 -3.79412174e-01
-7.03450024e-01 1.74767807e-01 -5.94277680e-01 2.27809280e-01
7.18650460e-01 1.17384143e-01 5.52617252e-01 8.38220358e-01
-3.01333759e-02 2.59594452e-02 -8.66241574e-01 1.22451615e+00
3.42093438e-01 -1.58634055e+00 4.98325348e-01 -1.95088416e-01
7.15576410e-01 6.68931544e-01 1.86700579e-02 5.42586982e-01
4.52344120e-01 -1.14599478e+00 7.00042844e-01 4.26125318e-01
9.20120835e-01 -3.45723003e-01 4.34671700e-01 1.87240373e-02
-9.02747452e-01 1.15572043e-01 -6.57431066e-01 2.32434142e-02
5.62248647e-01 -3.39865610e-02 -7.05602467e-01 4.10819978e-01
7.73609936e-01 4.27185029e-01 -6.84581757e-01 7.96581805e-01
-8.50693762e-01 7.22546279e-01 3.16594914e-02 1.70629576e-01
5.66388965e-01 4.19932827e-02 4.41460013e-01 7.91837037e-01
-4.12756503e-02 3.18806857e-01 -2.78086681e-02 1.01353800e+00
-4.28942055e-01 -2.26829364e-03 -4.41821277e-01 -3.90839651e-02
2.89800495e-01 1.16074395e+00 -2.08477899e-01 -6.70902729e-01
-8.05797279e-01 1.31932878e+00 4.85325307e-01 7.55830765e-01
-9.70258594e-01 -1.52927965e-01 8.09272110e-01 1.01505019e-01
6.47356868e-01 -4.38137054e-02 1.05716981e-01 -1.23850214e+00
-5.55417649e-02 -9.58768845e-01 5.71103454e-01 -1.47041059e+00
-1.38977659e+00 7.73703635e-01 -1.98530152e-01 -7.47571588e-01
-2.64510065e-01 -6.81879044e-01 -5.79191089e-01 9.24119115e-01
-1.82313466e+00 -1.72882879e+00 -6.44640625e-01 8.41957211e-01
9.59200799e-01 -6.31313920e-02 6.51313245e-01 -1.04319073e-01
-4.11178432e-02 5.50571144e-01 -3.85807395e-01 2.54061729e-01
6.72010124e-01 -1.15133631e+00 1.10081875e+00 9.17145789e-01
5.06494224e-01 1.29798383e-01 6.55507445e-01 -5.08185923e-01
-1.27674675e+00 -1.35608447e+00 8.20776761e-01 -9.83352065e-01
6.81541502e-01 -3.33389252e-01 -1.16895378e+00 9.02822793e-01
8.13851774e-01 4.59647000e-01 3.08251500e-01 -2.69627690e-01
-9.01058912e-01 4.14392687e-02 -9.09454226e-01 7.55364299e-01
1.19337308e+00 -8.48333001e-01 -9.53624487e-01 3.61774623e-01
1.30691385e+00 -2.36627027e-01 -5.02965689e-01 1.07599534e-01
3.32494944e-01 -7.40046024e-01 1.10600650e+00 -7.95446634e-01
6.38825476e-01 -3.85954350e-01 -3.26029718e-01 -7.80711114e-01
-2.90381789e-01 -4.86890316e-01 -2.69413978e-01 1.01958942e+00
4.04060781e-01 -3.94042343e-01 6.43167675e-01 4.98305231e-01
-9.18156430e-02 -7.65580535e-01 -7.59721279e-01 -5.36672533e-01
1.22138850e-01 -5.88881075e-01 2.65914112e-01 5.98603189e-01
-8.00249636e-01 7.61140823e-01 -4.80594158e-01 6.67094812e-02
7.96967566e-01 3.27186622e-02 1.12711954e+00 -6.22561932e-01
-5.22486687e-01 -2.84844995e-01 -2.26776376e-01 -1.24057615e+00
1.78972274e-01 -9.98582184e-01 1.71838626e-01 -2.40292931e+00
3.44709903e-01 3.39823872e-01 -1.35626808e-01 7.35036314e-01
-4.37012315e-01 5.59116066e-01 8.25459659e-01 1.01241060e-01
-1.02219534e+00 8.80557775e-01 1.53391290e+00 -5.71470141e-01
-3.23801674e-02 -4.95273173e-01 -7.23277867e-01 5.38895071e-01
7.20586538e-01 -1.17590554e-01 -8.93587887e-01 -9.04677272e-01
-1.97066483e-03 1.34844840e-01 9.42771673e-01 -5.79572439e-01
8.13635737e-02 -1.51382282e-01 2.20398113e-01 -6.57569945e-01
6.28462374e-01 -2.37468094e-01 -3.38247001e-01 4.44775969e-01
-2.56353080e-01 -3.04321460e-02 2.24536076e-01 5.05822718e-01
-1.09126695e-01 -1.31904304e-01 6.61469281e-01 -2.76437521e-01
-1.33627176e+00 6.81529105e-01 -6.83035403e-02 6.12113655e-01
9.38350260e-01 5.64498343e-02 -1.03055310e+00 -6.24011338e-01
-4.94902074e-01 9.27749097e-01 5.08984089e-01 5.47122240e-01
9.22860980e-01 -1.07072711e+00 -1.00777459e+00 -2.52875030e-01
5.39087236e-01 2.69067168e-01 8.80287349e-01 5.93268275e-01
-6.97036922e-01 4.38718319e-01 -2.09460527e-01 -6.60929620e-01
-1.22067773e+00 7.62919843e-01 4.40007031e-01 4.22724932e-02
-8.49178910e-01 1.15247369e+00 7.64926314e-01 -1.79985523e-01
9.43601429e-02 -2.03161001e-01 -6.59445859e-03 -1.06074847e-01
8.11522126e-01 1.44286151e-03 -4.75023180e-01 -5.04774034e-01
-4.31957126e-01 3.36879671e-01 -1.88673422e-01 -3.58818710e-01
1.06372762e+00 -1.42824426e-01 6.59046769e-02 2.70830959e-01
9.73639369e-01 -4.33696389e-01 -1.78662288e+00 -2.92309850e-01
-4.31593925e-01 -2.70798922e-01 -2.09598809e-01 -1.03229415e+00
-9.44190025e-01 1.16569304e+00 3.04920048e-01 -3.78424138e-01
9.82253313e-01 7.52212584e-01 9.26885843e-01 4.20188308e-01
1.05061203e-01 -5.15388131e-01 5.12328804e-01 6.00898385e-01
1.27945733e+00 -1.48214626e+00 -4.24344659e-01 -2.62133062e-01
-1.18810964e+00 4.61225986e-01 9.14498508e-01 -2.57559419e-01
2.96311881e-02 -4.60469037e-01 5.54161549e-01 -2.31180802e-01
-1.10328174e+00 -6.03443682e-01 4.96360034e-01 7.39170671e-01
-1.79775842e-02 -1.36336416e-01 1.92060228e-02 6.30012274e-01
7.40455762e-02 1.38295954e-02 4.96297151e-01 5.11305511e-01
-4.22088146e-01 -7.25282967e-01 -1.36598989e-01 1.48028925e-01
-1.33129880e-01 -3.59231055e-01 -4.01889861e-01 7.42618322e-01
-2.20472485e-01 9.92645800e-01 1.65442124e-01 1.96619462e-02
3.58257562e-01 2.12728098e-01 6.19100034e-01 -5.84819078e-01
-3.11569422e-01 -3.45000207e-01 7.78737739e-02 -9.58779514e-01
-1.75442815e-01 -2.63421118e-01 -1.29806161e+00 -8.04081187e-02
3.04313689e-01 -1.94839805e-01 3.67988437e-01 8.64921212e-01
5.22198915e-01 6.05240762e-01 -1.74644411e-01 -4.59856480e-01
-3.88043642e-01 -8.93318176e-01 -1.16299214e-02 4.95676935e-01
5.47361195e-01 -3.43356192e-01 -8.81453380e-02 8.71108025e-02] | [10.885479927062988, 1.692568063735962] |
2a74472c-e802-4ee7-95ef-b3705bb98109 | efficient-column-generation-for-cell | 1709.07337 | null | http://arxiv.org/abs/1709.07337v1 | http://arxiv.org/pdf/1709.07337v1.pdf | Efficient Column Generation for Cell Detection and Segmentation | We study the problem of instance segmentation in biological images with
crowded and compact cells. We formulate this task as an integer program where
variables correspond to cells and constraints enforce that cells do not
overlap. To solve this integer program, we propose a column generation
formulation where the pricing program is solved via exact optimization of very
small scale integer programs. Column generation is tightened using odd set
inequalities which fit elegantly into pricing problem optimization. Our column
generation approach achieves fast stable anytime inference for our instance
segmentation problems. We demonstrate on three distinct light microscopy
datasets, with several hundred cells each, that our proposed algorithm rapidly
achieves or exceeds state of the art accuracy. | ['Julian Yarkony', 'Miguel A. Gonzalez-Ballester', 'Shaofei Wang', 'Chong Zhang'] | 2017-09-21 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 5.24150252e-01 4.65784848e-01 -1.07726216e-01 -2.44525701e-01
-9.55397427e-01 -7.99254060e-01 1.62691757e-01 4.50998724e-01
-7.28791296e-01 1.42906260e+00 -5.84894598e-01 -9.85452458e-02
-1.50970832e-01 -5.50235808e-01 -1.11449301e+00 -1.01355338e+00
-7.23784044e-02 1.25647116e+00 1.45536557e-01 3.47156912e-01
5.80149651e-01 6.56841099e-01 -1.23093903e+00 3.67007226e-01
8.93498182e-01 1.05612671e+00 1.74061675e-02 1.19002390e+00
-2.04087928e-01 5.76409578e-01 -7.05723822e-01 -2.70110577e-01
4.53310758e-01 9.97916311e-02 -9.96274292e-01 4.77259248e-01
5.35543501e-01 2.28745006e-02 3.88879418e-01 9.53043818e-01
4.07380730e-01 -9.37729888e-03 6.33295953e-01 -1.53666532e+00
-6.27923012e-01 3.95376235e-01 -9.80255842e-01 3.54841411e-01
-7.55262226e-02 1.99032351e-01 1.12134039e+00 -6.21686161e-01
1.01102030e+00 1.16133869e+00 3.29905957e-01 5.91248751e-01
-1.97765374e+00 -3.76169384e-02 3.81035835e-01 -4.74237382e-01
-1.17418683e+00 -2.89900124e-01 1.20135687e-01 -6.79899633e-01
1.15975583e+00 4.30962533e-01 8.55614185e-01 3.20811003e-01
2.83755362e-01 7.17261136e-01 1.38032138e+00 -7.36108720e-02
6.07604623e-01 -9.26716402e-02 4.28926408e-01 7.81324685e-01
5.56346476e-01 -3.01204175e-01 -2.20826924e-01 -4.80563253e-01
7.06874013e-01 -8.64233524e-02 -2.45412976e-01 -3.97432894e-02
-1.20532167e+00 8.15288246e-01 -9.27928463e-02 -3.25071812e-01
-6.12999611e-02 4.94402975e-01 2.30162039e-01 -2.08510265e-01
6.09576881e-01 6.16235852e-01 -8.17554593e-01 2.53495961e-01
-1.00520420e+00 7.18486249e-01 1.20753884e+00 1.08739507e+00
5.43414235e-01 -4.51701492e-01 -2.33245462e-01 7.77728438e-01
1.36559065e-02 5.45500338e-01 -1.75672323e-01 -1.61970282e+00
2.72761881e-01 4.60817277e-01 4.35512543e-01 -6.91886067e-01
-5.47278047e-01 -1.63467780e-01 -5.64052582e-01 2.66582489e-01
8.44353378e-01 -2.54657328e-01 -9.97761369e-01 1.44660127e+00
3.77264470e-01 3.24386060e-01 -1.90760955e-01 5.97005725e-01
3.93135071e-01 8.41345727e-01 -1.13908418e-01 -7.95400798e-01
1.29080725e+00 -9.32342827e-01 -5.70610046e-01 -2.32721809e-02
3.62145633e-01 -3.50397289e-01 6.42239928e-01 6.57493591e-01
-1.48887634e+00 1.89448312e-01 -7.07202613e-01 -4.17762190e-01
-5.70667446e-01 -2.78949857e-01 8.91215026e-01 4.18073177e-01
-1.14610863e+00 5.30875325e-01 -6.17088914e-01 -6.96045980e-02
9.13590610e-01 9.25074637e-01 -2.30376184e-01 2.23481610e-01
-2.42355898e-01 3.77096534e-01 7.97507688e-02 2.44795099e-01
-5.90488911e-01 -9.24280941e-01 -7.20406592e-01 -3.07099074e-01
4.44988549e-01 -9.43355620e-01 9.38042164e-01 -3.47896129e-01
-1.27265370e+00 1.61876774e+00 -5.34587562e-01 -8.69830906e-01
6.11271143e-01 3.10065866e-01 5.74263036e-01 3.02494913e-01
3.04078132e-01 9.18470442e-01 3.12869310e-01 -1.62941825e+00
-7.72345781e-01 -4.77417618e-01 3.16500217e-01 -2.28799909e-01
3.79323959e-01 -2.60703802e-01 -5.54861546e-01 -2.90675223e-01
-1.63486481e-01 -1.08518410e+00 -7.66563416e-01 1.78843409e-01
-6.97624445e-01 1.05362214e-01 4.82130796e-01 -4.00478184e-01
8.32282662e-01 -1.45271003e+00 4.19388533e-01 2.88730741e-01
5.84349632e-01 -4.62430984e-01 1.60361105e-03 -1.66950077e-01
5.10227442e-01 4.95822340e-01 -7.27168441e-01 -5.91227770e-01
1.37374625e-01 8.00901175e-01 -1.87565863e-01 6.56735301e-01
1.51301295e-01 1.10670662e+00 -7.50603616e-01 -9.37290132e-01
-1.82527706e-01 8.59274492e-02 -1.09634876e+00 2.77326647e-02
-7.86233723e-01 3.39665413e-01 -3.50055605e-01 9.58007812e-01
9.57190990e-01 -5.34438729e-01 2.35308915e-01 6.42803386e-02
-1.98707819e-01 -7.13556468e-01 -1.08234596e+00 1.28878045e+00
-3.45210172e-02 4.66312617e-01 5.61462462e-01 -1.13877773e+00
3.84212792e-01 -3.14789534e-01 1.01590014e+00 -2.30194777e-01
2.11121216e-01 2.68715143e-01 -3.59732389e-01 -3.13600302e-01
4.45348233e-01 -2.82550067e-01 -6.73838034e-02 1.54071093e-01
-1.88870177e-01 -8.73175621e-01 6.15210652e-01 1.74343601e-01
9.50464904e-01 -1.59135938e-01 8.14011246e-02 -8.08267653e-01
3.95232290e-01 3.30106944e-01 8.41298461e-01 8.76584649e-01
-2.30320752e-01 5.35975397e-01 1.11341023e+00 -4.01606828e-01
-1.04658306e+00 -8.88603926e-01 -4.52772021e-01 8.25700462e-01
2.96094596e-01 2.73079604e-01 -1.18171585e+00 -2.27380097e-01
4.51099098e-01 -1.38534173e-01 -1.01090670e+00 8.82835925e-01
-6.80482686e-01 -1.34997320e+00 2.48641789e-01 3.10504138e-01
9.30570811e-03 -5.47253847e-01 -6.96323752e-01 3.06756675e-01
-1.21030686e-02 -1.39935958e+00 -4.27218050e-01 5.03030956e-01
-9.90838826e-01 -1.12701786e+00 -8.74990880e-01 -8.93497825e-01
1.20136702e+00 -3.84724960e-02 1.17228603e+00 5.43639004e-01
-8.93975914e-01 1.89184144e-01 2.52259493e-01 -6.38611078e-01
4.05569077e-01 -4.16504741e-02 -2.33829305e-01 -2.63219833e-01
-1.34974003e-01 -1.92232877e-01 -5.33421159e-01 1.25213251e-01
-9.61385012e-01 -2.54236367e-02 -2.05826059e-01 1.31662786e+00
1.71963263e+00 -4.29769792e-02 2.45647147e-01 -1.54745293e+00
4.59206939e-01 -3.17829192e-01 -1.22736430e+00 2.37889931e-01
-2.67797500e-01 1.53835267e-01 5.45374990e-01 -2.45651484e-01
-5.25789142e-01 2.69318610e-01 1.60245404e-01 -2.21395850e-01
1.79791734e-01 1.79109603e-01 1.41774669e-01 -6.50598943e-01
1.72444373e-01 -3.50220501e-02 -5.83042391e-02 2.09898412e-01
2.28578478e-01 3.14636491e-02 5.41383862e-01 -1.18727982e+00
4.20748442e-01 1.25416946e+00 5.82361639e-01 -6.53463900e-01
-8.80500853e-01 -3.04815382e-01 -5.44893682e-01 -5.41574135e-02
1.14235222e+00 -5.55491328e-01 -1.40972328e+00 3.30598563e-01
-1.27924216e+00 -8.63591254e-01 -4.85476643e-01 -2.21379355e-01
-1.09472513e+00 1.51606455e-01 -9.36042547e-01 -1.06605208e+00
5.26400730e-02 -1.35162604e+00 1.51139450e+00 1.77656993e-01
1.08845957e-01 -1.14470506e+00 1.20502345e-01 6.49253368e-01
-3.85070257e-02 8.92106056e-01 8.93969476e-01 -2.62439728e-01
-1.09415460e+00 2.18460947e-01 -4.60572660e-01 -3.34492922e-01
-3.94564480e-01 5.08404255e-01 -8.96275818e-01 -1.72584906e-01
-2.46578738e-01 -2.60792702e-01 9.66368735e-01 1.22483218e+00
1.68235934e+00 -5.45625865e-01 -7.91562974e-01 1.01597035e+00
1.84414470e+00 1.02102838e-01 4.61208045e-01 2.05421984e-01
5.18600285e-01 6.87890351e-01 6.20600700e-01 6.45424604e-01
3.53660077e-01 3.07150573e-01 4.48584408e-01 -2.29760647e-01
5.06870389e-01 5.07035136e-01 -3.79478663e-01 2.20612288e-01
-1.92922875e-01 -6.11371875e-01 -7.78922558e-01 5.87404609e-01
-2.11068869e+00 -9.16720212e-01 -2.92135954e-01 1.83005106e+00
9.35774267e-01 1.27105266e-01 1.96170896e-01 -1.00635991e-01
4.84745055e-01 -1.61614969e-01 -8.20479095e-01 -6.53610528e-01
-4.56358403e-01 2.64153570e-01 9.83488142e-01 1.09679723e+00
-1.08840406e+00 8.34936380e-01 7.66258574e+00 5.81174910e-01
-6.64415836e-01 -2.26796623e-02 1.50939846e+00 -6.37354434e-01
-4.08966541e-01 -1.46181017e-01 -1.11370039e+00 5.26007593e-01
3.96382838e-01 -2.94828236e-01 7.40866601e-01 4.08500016e-01
2.80639589e-01 -4.32914704e-01 -1.21536565e+00 9.21797216e-01
-1.36792228e-01 -1.91972482e+00 -1.28181025e-01 6.32234693e-01
1.39371145e+00 -1.81599185e-01 2.46015027e-01 -2.19186068e-01
6.49882019e-01 -1.29197598e+00 6.48655713e-01 5.39550662e-01
5.94414473e-01 -7.11856484e-01 7.05228686e-01 2.08980322e-01
-8.69710505e-01 -6.45234510e-02 -4.92671728e-01 -3.26753664e-03
4.68785673e-01 1.00564206e+00 -4.33341980e-01 -4.43471558e-02
3.74719888e-01 2.62407660e-01 -2.59612426e-02 1.10527289e+00
6.47222638e-01 -6.24503382e-02 -6.03303671e-01 5.10313138e-02
5.56314364e-02 -4.70183104e-01 4.43556637e-01 1.50436401e+00
2.94892434e-02 7.60297000e-01 4.11451131e-01 1.08025181e+00
-3.81399512e-01 -1.68035273e-02 -5.28057754e-01 -3.63672152e-02
1.95458993e-01 1.20043015e+00 -1.42892909e+00 -2.91562200e-01
-2.93630138e-02 8.55983675e-01 6.42073750e-01 5.55615187e-01
-9.28026378e-01 -1.37036920e-01 8.79200101e-01 3.07130814e-02
3.00091773e-01 -1.91785976e-01 -8.97159219e-01 -8.73503208e-01
-2.93952078e-01 -4.84455407e-01 2.55878389e-01 -4.90782291e-01
-1.44480014e+00 1.94697499e-01 -1.85450554e-01 -3.13392729e-01
1.50244057e-01 -1.18338931e+00 -3.66561741e-01 5.76906323e-01
-1.55502295e+00 -8.82595003e-01 -5.59243299e-02 2.10586399e-01
4.28398430e-01 4.60057557e-01 3.18976760e-01 1.14283174e-01
-7.51392603e-01 3.09941560e-01 3.75133753e-02 -2.34094128e-01
1.83739793e-02 -1.67391920e+00 -1.74627498e-01 4.32492018e-01
-3.17116290e-01 6.02795184e-01 9.31442499e-01 -6.36979043e-01
-1.58857453e+00 -9.37722623e-01 6.44168377e-01 -5.46684504e-01
6.17174923e-01 -3.63278180e-01 -5.86391568e-01 6.18063271e-01
-4.38949689e-02 5.43910146e-01 6.75256670e-01 -3.12812805e-01
1.54214516e-01 1.85541317e-01 -1.65251255e+00 3.48148406e-01
1.00583994e+00 -2.13867769e-01 8.11003521e-02 9.27392423e-01
8.49072635e-01 -8.69645655e-01 -9.31898296e-01 1.71979070e-01
4.59834158e-01 -5.31417668e-01 1.04537535e+00 -5.47753990e-01
5.55112958e-01 -3.65431130e-01 -3.03146422e-01 -8.16741407e-01
-6.11426532e-02 -6.99719369e-01 -2.41341010e-01 8.69752467e-01
6.86461151e-01 -6.31024778e-01 1.10186708e+00 1.02466321e+00
2.12882295e-01 -1.30741942e+00 -1.13260221e+00 -7.38061666e-01
3.40615749e-01 5.12455106e-02 5.18200636e-01 8.36785674e-01
1.34756401e-01 -1.06559634e-01 8.51262175e-03 9.44985449e-02
1.19009960e+00 4.11609650e-01 5.46374023e-01 -1.16618705e+00
-5.43023229e-01 -5.32430053e-01 -9.67952237e-02 -9.06546891e-01
4.46753532e-01 -7.05623627e-01 3.37879270e-01 -1.56958520e+00
6.79127753e-01 -4.47934836e-01 8.03733468e-02 1.58607855e-01
-7.37213623e-03 5.11289418e-01 1.00146219e-01 -1.36931777e-01
-9.02484834e-01 -1.17434576e-01 1.49468648e+00 -5.31793952e-01
-1.23089530e-01 -3.73096049e-01 -6.92943811e-01 6.05349720e-01
6.10024869e-01 -3.57803822e-01 -3.53168659e-02 -6.00812554e-01
4.65988159e-01 2.83754706e-01 3.34363490e-01 -7.25221395e-01
3.51115882e-01 -7.28278160e-01 2.34453022e-01 -5.20544589e-01
4.40814972e-01 -7.20028698e-01 1.74924061e-01 4.43557292e-01
-3.78469825e-01 -7.79236332e-02 3.19545805e-01 7.41271317e-01
7.89574534e-02 -1.91284791e-01 1.07483649e+00 -3.48617464e-01
-1.62985280e-01 6.16193771e-01 -2.18809947e-01 4.57042068e-01
1.41475332e+00 -3.96459579e-01 -6.82480812e-01 1.66515529e-01
-1.09229231e+00 7.70450234e-01 7.38956690e-01 -5.80695927e-01
3.27104092e-01 -7.92366624e-01 -5.12466908e-01 -2.75803119e-01
-1.37547761e-01 5.09508967e-01 1.12821206e-01 9.68396842e-01
-9.09103394e-01 6.06274903e-01 -1.19857475e-01 -8.07891607e-01
-1.21724820e+00 5.16939700e-01 4.13328111e-01 -6.08546197e-01
-1.76407897e-03 1.30080485e+00 2.82494813e-01 -3.46388191e-01
-1.71762123e-03 -8.70174468e-01 7.63123333e-02 1.16374172e-01
2.90683508e-01 6.00623667e-01 -4.34860170e-01 -3.14322084e-01
-3.19417000e-01 7.64999390e-01 -3.85209545e-03 -2.69533724e-01
1.49701643e+00 -1.22693665e-01 -7.21988916e-01 4.08148438e-01
1.01182485e+00 1.21000737e-01 -1.57565534e+00 3.30092102e-01
-2.30088949e-01 -3.29987377e-01 -3.07035446e-01 -5.81022382e-01
-1.17870164e+00 6.11920416e-01 7.06106499e-02 2.91691989e-01
8.11752439e-01 9.63133723e-02 7.94161260e-01 4.29208279e-01
7.17487037e-01 -1.38619936e+00 -1.67877600e-01 5.95677316e-01
4.97935325e-01 -1.21525645e+00 3.45509112e-01 -7.66926050e-01
-2.51374722e-01 8.22511613e-01 5.60222745e-01 -3.53790164e-01
5.99813163e-01 1.02231848e+00 -3.60686898e-01 -3.39369565e-01
-1.02293789e+00 -9.91454124e-02 -2.92468101e-01 5.03785193e-01
1.53678492e-01 2.81899869e-01 -4.81121868e-01 5.69813013e-01
-5.07127158e-02 2.90480167e-01 5.84128618e-01 7.66584694e-01
-6.66425049e-01 -8.08131814e-01 -3.03981870e-01 8.31969619e-01
-7.77705848e-01 -4.29446762e-03 -3.87598723e-01 4.17965531e-01
4.04128283e-01 7.22977519e-01 2.69271612e-01 3.08089733e-01
-2.46731262e-03 -4.42076385e-01 8.18887293e-01 -6.08744979e-01
-5.22390127e-01 2.42016390e-02 -1.55272141e-01 -5.64715981e-01
-4.93729323e-01 -6.22426569e-01 -1.74166322e+00 -4.17156905e-01
-3.69409978e-01 1.14482239e-01 2.59702057e-01 9.60431874e-01
1.90666869e-01 4.24991429e-01 1.82801694e-01 -9.46952462e-01
2.92819813e-02 -4.60765772e-02 -9.39011335e-01 1.19652562e-01
4.03446257e-01 -5.09646356e-01 -3.71316582e-01 5.04538238e-01] | [14.492613792419434, -3.189276933670044] |
3a21c84b-5cc7-4d8d-a927-85278a8ec473 | towards-streaming-egocentric-action | 2110.05386 | null | https://arxiv.org/abs/2110.05386v2 | https://arxiv.org/pdf/2110.05386v2.pdf | Towards Streaming Egocentric Action Anticipation | Egocentric action anticipation is the task of predicting the future actions a camera wearer will likely perform based on past video observations. While in a real-world system it is fundamental to output such predictions before the action begins, past works have not generally paid attention to model runtime during evaluation. Indeed, current evaluation schemes assume that predictions can be made offline, and hence that computational resources are not limited. In contrast, in this paper, we propose a "streaming" egocentric action anticipation evaluation protocol which explicitly considers model runtime for performance assessment, assuming that predictions will be available only after the current video segment is processed, which depends on the processing time of a method. Following the proposed evaluation scheme, we benchmark different state-of-the-art approaches for egocentric action anticipation on two popular datasets. Our analysis shows that models with a smaller runtime tend to outperform heavier models in the considered streaming scenario, thus changing the rankings generally observed in standard offline evaluations. Based on this observation, we propose a lightweight action anticipation model consisting in a simple feed-forward 3D CNN, which we propose to optimize using knowledge distillation techniques and a custom loss. The results show that the proposed approach outperforms prior art in the streaming scenario, also in combination with other lightweight models. | ['Giovanni Maria Farinella', 'Antonino Furnari'] | 2021-10-11 | null | null | null | null | ['action-anticipation'] | ['computer-vision'] | [ 3.12250674e-01 3.17163378e-01 -1.99087888e-01 -4.09142643e-01
-1.78960606e-01 -8.61034095e-02 8.09196472e-01 4.15234864e-01
-9.06045616e-01 5.28001666e-01 5.26604116e-01 6.92077279e-02
-1.77819222e-01 -5.84068716e-01 -9.02612507e-01 -4.13682580e-01
-2.95555413e-01 4.52877969e-01 3.66173446e-01 -9.29237008e-02
2.30290994e-01 5.00707209e-01 -2.07592607e+00 4.96648192e-01
3.53567839e-01 1.10560393e+00 2.01577589e-01 9.53449547e-01
4.45831239e-01 1.22394693e+00 -3.86294842e-01 -4.19088215e-01
4.04146463e-01 -1.05021916e-01 -7.60882616e-01 6.45397455e-02
3.45353097e-01 -7.39238322e-01 -2.54671425e-01 6.08218074e-01
5.81284702e-01 3.45863760e-01 3.08648676e-01 -1.05221510e+00
3.85684639e-01 7.59569228e-01 9.43424776e-02 1.16955899e-01
6.84008718e-01 5.15216053e-01 9.24080789e-01 -3.82712483e-01
1.00084805e+00 8.45822334e-01 6.11486375e-01 4.29747730e-01
-9.00968552e-01 -9.66615677e-02 5.26256442e-01 9.31639194e-01
-1.04433334e+00 -6.82198405e-01 7.58164167e-01 -3.82037848e-01
1.25473869e+00 2.49068201e-01 9.53218043e-01 1.27314746e+00
2.90307879e-01 1.12190163e+00 7.20207512e-01 -2.93543965e-01
6.27964318e-01 1.96160197e-01 -2.53246933e-01 1.15772046e-01
-4.47181650e-02 -1.90104321e-01 -6.96850240e-01 1.88349068e-01
5.60172610e-02 1.30005153e-02 -1.52000308e-01 -5.33172190e-01
-1.17820585e+00 4.16574568e-01 8.68667588e-02 2.35708132e-01
-1.09599090e+00 3.75124544e-01 9.77314055e-01 2.22193003e-01
7.21321642e-01 5.19820601e-02 -3.36879849e-01 -8.75476897e-01
-1.23995972e+00 6.31127059e-01 9.96451259e-01 7.69101977e-01
5.45114577e-01 -2.15723172e-01 -3.12882036e-01 2.73801178e-01
5.39216027e-02 4.86655310e-02 6.58313513e-01 -9.29486454e-01
4.27192926e-01 6.38257384e-01 3.50590646e-01 -1.00258255e+00
-4.54015791e-01 -3.55765462e-01 -3.20596576e-01 3.96203935e-01
2.79359609e-01 -5.16405329e-02 -3.43788803e-01 1.38837337e+00
4.77688879e-01 4.56444025e-01 1.52377844e-01 9.36817586e-01
2.24483877e-01 1.27928689e-01 1.78818852e-01 -4.00365561e-01
9.87823427e-01 -1.08091271e+00 -6.89033687e-01 1.01772241e-01
8.94984901e-01 -6.47911727e-01 9.65234399e-01 9.19596136e-01
-1.24590111e+00 -6.14519954e-01 -1.11399937e+00 1.78712472e-01
-2.17975318e-01 3.50032270e-01 4.46337730e-01 6.19095922e-01
-1.01221359e+00 1.19444525e+00 -1.16555166e+00 -5.76648176e-01
2.97605187e-01 3.54696363e-01 -3.47034544e-01 1.10769480e-01
-8.59158337e-01 9.14903343e-01 4.98859882e-01 3.67565721e-01
-1.13748050e+00 -5.40036678e-01 -5.58682978e-01 2.36173764e-01
5.91325283e-01 -6.71868801e-01 1.57697749e+00 -1.25946939e+00
-1.76305747e+00 5.14241159e-01 -1.21756241e-01 -1.40959871e+00
1.31030750e+00 -8.71634066e-01 -1.90216541e-01 3.54269475e-01
-4.37939554e-01 5.26148260e-01 8.13259482e-01 -7.40159512e-01
-6.70986116e-01 -1.79059818e-01 8.52194071e-01 3.55402619e-01
-5.92397034e-01 -1.28975496e-01 -2.31090561e-01 -8.33042413e-02
-5.33296824e-01 -9.51919198e-01 -3.59740973e-01 -6.58471212e-02
-1.42577529e-01 -2.50696957e-01 7.87490427e-01 -4.53515232e-01
1.37081087e+00 -1.91432071e+00 1.82279095e-01 -1.27329916e-01
-7.58487955e-02 5.01793742e-01 2.06392065e-01 7.55923510e-01
-7.48226643e-02 -4.77289319e-01 4.26666401e-02 -9.09039438e-01
2.84733832e-01 8.69811326e-02 -2.34786808e-01 6.41772151e-01
-2.79203832e-01 4.24627990e-01 -9.80144680e-01 -5.77952921e-01
8.48800659e-01 5.45000970e-01 -8.26887429e-01 3.12219083e-01
-6.35105908e-01 2.62603879e-01 -2.27902323e-01 2.60764509e-01
5.01078546e-01 6.10684976e-02 2.95753062e-01 -1.39121026e-01
-3.76846075e-01 2.39208445e-01 -1.31999075e+00 1.93838477e+00
-8.54914844e-01 6.55615032e-01 -2.99683750e-01 -1.09815705e+00
6.51161551e-01 5.33520103e-01 7.94600844e-01 -4.80255067e-01
1.78373143e-01 3.13683189e-02 -9.20262039e-02 -7.86560476e-01
7.88985431e-01 3.49619180e-01 5.06243467e-01 3.81079078e-01
-1.88068837e-01 3.96056652e-01 4.94663328e-01 2.04934720e-02
1.37089241e+00 8.50107372e-01 3.15114349e-01 1.18343517e-01
8.59906197e-01 5.81382960e-03 4.10268724e-01 5.32729208e-01
-3.18978965e-01 4.63378638e-01 6.78948760e-01 -7.07585335e-01
-9.91380930e-01 -5.73416173e-01 3.98405820e-01 1.01292455e+00
1.00435326e-02 -7.22182631e-01 -1.10007536e+00 -8.20046484e-01
-4.23142433e-01 8.97482574e-01 -7.62840211e-01 -1.18029870e-01
-6.60962760e-01 -3.32670748e-01 2.06907853e-01 3.91014963e-01
4.49218750e-01 -1.35625255e+00 -1.54842401e+00 3.15584242e-01
2.81436611e-02 -1.21523273e+00 4.78748558e-03 -3.07848781e-01
-1.06083596e+00 -1.24190617e+00 -5.94374120e-01 1.13448285e-01
2.18775511e-01 -9.34807491e-03 1.15209651e+00 -4.25013490e-02
-4.52468134e-02 8.86458397e-01 -7.95996249e-01 -4.69589323e-01
-3.81473601e-01 1.29283831e-01 1.59172475e-01 5.11833668e-01
4.43317801e-01 -7.86207676e-01 -1.19727039e+00 -4.40159030e-02
-9.55878735e-01 1.94540367e-01 6.22081101e-01 3.65768909e-01
5.09142995e-01 -1.35800213e-01 2.03796834e-01 -1.02161181e+00
3.09096545e-01 -4.47305888e-01 -3.95870864e-01 1.45706087e-01
-5.89308560e-01 -3.61821800e-02 1.01895404e+00 -3.26783001e-01
-1.17534554e+00 3.95560712e-01 -2.70114511e-01 -7.35539258e-01
-4.62122798e-01 2.25852981e-01 -4.54043895e-02 1.91823408e-01
6.21463716e-01 2.61753917e-01 -1.79148853e-01 -4.84364301e-01
1.94413066e-01 4.58376020e-01 1.26510888e-01 -1.59546793e-01
2.41378531e-01 7.83914208e-01 2.05397177e-02 -8.53473186e-01
-6.09548330e-01 -5.67543089e-01 -6.89876258e-01 -7.65649617e-01
8.01104665e-01 -8.55110526e-01 -1.14145720e+00 3.84223878e-01
-1.25511134e+00 -4.48559672e-01 -7.49156237e-01 7.69819796e-01
-1.21917605e+00 5.05852997e-01 -1.41311526e-01 -1.07550156e+00
-3.36114585e-01 -1.18820536e+00 1.01259923e+00 -7.68720955e-02
-3.71732086e-01 -9.52257931e-01 2.66280174e-01 3.01466793e-01
3.32082063e-01 2.99334884e-01 9.19851959e-02 -7.82838941e-01
-6.38285220e-01 -5.37360489e-01 2.72969276e-01 6.37532771e-01
-3.43995273e-01 9.56170168e-03 -1.21629393e+00 -7.95897096e-02
6.44817296e-03 -1.11704536e-01 5.97859681e-01 3.39499354e-01
1.17733145e+00 -2.96324998e-01 -1.08548716e-01 4.05165076e-01
1.40075779e+00 -1.58264250e-01 9.07311380e-01 5.26865780e-01
3.52634639e-01 7.14823544e-01 1.09751403e+00 1.05965662e+00
3.68027389e-01 9.27242041e-01 1.02512813e+00 2.86685169e-01
-7.74677144e-03 -1.40349925e-01 6.85133338e-01 3.01341355e-01
-6.00983143e-01 -3.68226558e-01 -4.84402806e-01 7.94878602e-01
-2.21654177e+00 -1.09868288e+00 -1.11238703e-01 2.50739098e+00
1.31179601e-01 4.45231408e-01 4.12687212e-01 2.67703682e-01
3.30285132e-01 5.09646177e-01 -4.21793491e-01 -7.01275051e-01
2.97963232e-01 9.18683037e-02 5.88013291e-01 3.41821998e-01
-1.12361073e+00 6.88967586e-01 5.22319603e+00 5.98373652e-01
-1.23995924e+00 2.10648075e-01 2.45063275e-01 -5.60901046e-01
7.99973905e-02 7.99392834e-02 -6.54313207e-01 5.06803513e-01
1.29825795e+00 -1.94933921e-01 1.41151190e-01 1.22817099e+00
8.39292765e-01 -3.38608801e-01 -1.44238806e+00 7.78962374e-01
1.01544537e-01 -1.07671857e+00 -1.04704730e-01 -2.04339474e-02
4.59714472e-01 1.08429663e-01 -9.20354500e-02 1.05520912e-01
-2.58688599e-01 -6.49867892e-01 8.21679235e-01 9.70028102e-01
9.39888582e-02 -6.35410488e-01 8.91431570e-01 4.71263915e-01
-1.04927766e+00 -3.37924629e-01 -3.02324742e-01 -4.36563611e-01
5.02603650e-01 5.75482965e-01 -1.06934857e+00 7.19192982e-01
5.93937576e-01 9.68362927e-01 -4.19913888e-01 1.24729860e+00
-1.81988642e-01 7.22779870e-01 -2.70825475e-01 -4.36466970e-02
3.29082876e-01 9.06007439e-02 7.73147404e-01 1.37662590e+00
4.05116796e-01 -3.90051961e-01 -1.78735837e-01 3.71717334e-01
2.24873424e-01 4.02212560e-01 -6.73280716e-01 2.30076835e-01
-1.32411763e-01 1.11259282e+00 -5.49311221e-01 -5.74821770e-01
-4.11974460e-01 1.14559090e+00 1.11402802e-01 1.28689175e-02
-9.13498461e-01 -9.60224122e-02 6.72303259e-01 3.99950773e-01
6.08179450e-01 -2.02640027e-01 1.24996677e-01 -1.03503227e+00
4.62492883e-01 -4.91116464e-01 3.29202622e-01 -7.20207930e-01
-4.18355763e-01 5.99552929e-01 1.98302299e-01 -1.72015858e+00
-7.74894893e-01 -5.00393927e-01 -6.18718147e-01 2.84723222e-01
-1.46753895e+00 -1.05853379e+00 -3.82386684e-01 4.14333701e-01
9.20619547e-01 8.92735422e-02 6.71186686e-01 3.75054300e-01
-3.16601783e-01 3.11075032e-01 -1.33116990e-01 -4.78320032e-01
6.01597726e-01 -1.16062939e+00 3.24393392e-01 8.64756882e-01
1.40495777e-01 2.40702912e-01 1.26246607e+00 -4.50793296e-01
-1.28778684e+00 -8.42148185e-01 1.01138628e+00 -3.28072131e-01
6.06144488e-01 -1.48284137e-01 -6.82217002e-01 7.06485152e-01
2.64894068e-01 2.55528465e-02 4.09409553e-01 -1.77163631e-02
1.25051230e-01 -3.66273940e-01 -8.96577001e-01 5.96637368e-01
1.11463058e+00 -2.48149008e-01 -4.16223288e-01 6.25755250e-01
5.34104705e-01 -2.60105103e-01 -9.85454977e-01 3.37939799e-01
8.69900107e-01 -1.61705351e+00 8.09422433e-01 -5.04846454e-01
5.40200114e-01 -1.43955141e-01 9.11948308e-02 -9.71623361e-01
3.86072099e-01 -6.32835567e-01 -6.01922452e-01 8.20124865e-01
7.55944103e-02 -2.51398027e-01 1.23359048e+00 6.09108508e-01
-2.19911277e-01 -8.64740908e-01 -1.04622638e+00 -7.14562297e-01
-6.65432096e-01 -9.14536655e-01 4.37880218e-01 2.76194006e-01
-1.14365838e-01 -3.58237267e-01 -6.57654405e-01 -9.13576111e-02
3.58069539e-01 -1.30793855e-01 1.11378801e+00 -8.65330040e-01
-5.85261583e-01 -3.00862044e-01 -8.95838082e-01 -1.04928851e+00
1.60836041e-01 -2.25123465e-01 -4.27790694e-02 -1.30156636e+00
-1.03895003e-02 -6.51419312e-02 -4.26895261e-01 3.33501697e-02
1.74411744e-01 2.23447278e-01 3.56638223e-01 -3.00149457e-03
-1.20391107e+00 7.09919035e-01 7.79205382e-01 1.20118164e-01
-2.81877249e-01 4.60494310e-01 6.37024865e-02 1.03631353e+00
6.72031462e-01 -2.34244511e-01 -6.88222647e-01 -2.29658723e-01
4.85979944e-01 -1.61850303e-01 3.98026139e-01 -1.65397811e+00
4.51483905e-01 -1.91258285e-02 -6.95234537e-02 -5.20201206e-01
7.04107821e-01 -1.12408447e+00 1.64180517e-01 6.65922523e-01
-3.48246217e-01 -2.55041067e-02 -2.28120554e-02 6.87117696e-01
-1.79669157e-01 -4.26077187e-01 1.96072370e-01 -1.24536455e-01
-9.90612149e-01 1.58992395e-01 -4.46713984e-01 -4.63208348e-01
1.24920297e+00 -5.45250416e-01 1.75599113e-01 -6.03330255e-01
-1.01335084e+00 -4.66761738e-03 5.42493880e-01 2.43471608e-01
5.25259078e-01 -9.00648236e-01 -6.63752377e-01 -2.77494669e-01
1.94731295e-01 -3.00899237e-01 8.17066848e-01 1.28160584e+00
-7.03076184e-01 3.61926734e-01 -2.65120625e-01 -5.60217917e-01
-1.38775074e+00 7.15909362e-01 1.95918232e-01 -5.22474408e-01
-8.19062650e-01 4.22159076e-01 -3.61913294e-02 1.04861245e-01
4.51693147e-01 -2.33698502e-01 -5.59004784e-01 1.68574721e-01
5.16040683e-01 8.25916708e-01 3.16472441e-01 -5.59799850e-01
-1.82887912e-01 2.23606080e-01 -1.17697842e-01 -5.98716177e-02
1.47085309e+00 -1.18710831e-01 2.62948334e-01 6.11921966e-01
8.77825022e-01 -1.29675746e-01 -1.62670732e+00 1.94822147e-01
-1.80418715e-02 -5.98183453e-01 -1.38045013e-01 -3.80678803e-01
-7.48160362e-01 9.21578407e-01 8.20004940e-01 1.16876431e-01
1.36900795e+00 -4.58447993e-01 6.77527368e-01 5.17545700e-01
3.98456246e-01 -1.50540400e+00 -1.62604958e-01 1.48268878e-01
7.96160221e-01 -1.03329492e+00 3.87070000e-01 -1.97022691e-01
-6.84754252e-01 1.16858530e+00 5.37839055e-01 -2.88914531e-01
4.92573947e-01 -1.82603911e-01 -2.57627428e-01 -2.74519399e-02
-1.10551381e+00 -1.94441423e-01 -4.75651473e-02 3.29614580e-01
3.24930012e-01 2.75103785e-02 -7.49095857e-01 2.45379835e-01
-9.52133238e-02 4.84183013e-01 6.74525678e-01 9.22256052e-01
-8.32807720e-02 -1.08006692e+00 -1.57443397e-02 2.37216920e-01
-5.82844257e-01 1.49707377e-01 -1.91000909e-01 8.08816969e-01
2.30976969e-01 7.97985017e-01 4.38304059e-02 -4.56626713e-01
6.73791349e-01 -4.95745018e-02 4.65864599e-01 -3.75225872e-01
-8.53573203e-01 -3.82417470e-01 3.80455971e-01 -1.24521494e+00
-8.93456042e-01 -9.65653241e-01 -7.60984898e-01 -2.77236164e-01
-2.75629330e-02 3.32732365e-04 9.30085242e-01 1.03693545e+00
3.78896236e-01 6.37199998e-01 4.31916535e-01 -1.28406537e+00
-5.90565026e-01 -1.06246698e+00 -2.49324977e-01 4.93260056e-01
1.57981887e-01 -5.95712543e-01 -2.89094776e-01 2.41132841e-01] | [8.141510009765625, 0.40105077624320984] |
3b36938b-b060-4477-8ad2-c4ab8ff2bc4f | 3d-deeply-supervised-network-for-automatic | 1607.00582 | null | http://arxiv.org/abs/1607.00582v1 | http://arxiv.org/pdf/1607.00582v1.pdf | 3D Deeply Supervised Network for Automatic Liver Segmentation from CT Volumes | Automatic liver segmentation from CT volumes is a crucial prerequisite yet
challenging task for computer-aided hepatic disease diagnosis and treatment. In
this paper, we present a novel 3D deeply supervised network (3D DSN) to address
this challenging task. The proposed 3D DSN takes advantage of a fully
convolutional architecture which performs efficient end-to-end learning and
inference. More importantly, we introduce a deep supervision mechanism during
the learning process to combat potential optimization difficulties, and thus
the model can acquire a much faster convergence rate and more powerful
discrimination capability. On top of the high-quality score map produced by the
3D DSN, a conditional random field model is further employed to obtain refined
segmentation results. We evaluated our framework on the public MICCAI-SLiver07
dataset. Extensive experiments demonstrated that our method achieves
competitive segmentation results to state-of-the-art approaches with a much
faster processing speed. | ['Pheng-Ann Heng', 'Yueming Jin', 'Qi Dou', 'Lequan Yu', 'Hao Chen', 'Jing Qin'] | 2016-07-03 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [-2.00704560e-01 8.73718485e-02 -2.93872714e-01 -5.97205579e-01
-1.06870663e+00 -1.76416874e-01 3.60265106e-01 1.93758056e-01
-4.89203960e-01 4.21172768e-01 1.94322467e-01 -4.83805448e-01
1.37972534e-01 -5.23672044e-01 -5.11605203e-01 -8.67834508e-01
-3.08796465e-01 6.18602753e-01 1.30215123e-01 2.79300392e-01
-2.49293908e-01 5.99849522e-01 -4.57818955e-01 -3.43135633e-02
1.02629781e+00 1.10368061e+00 3.61279368e-01 6.01630330e-01
1.04547955e-01 5.45221567e-01 -1.56587988e-01 5.92628978e-02
4.06269789e-01 -4.33012754e-01 -8.35905671e-01 1.55892581e-01
1.29430979e-01 -4.77712065e-01 -3.73980492e-01 1.16342819e+00
7.81159103e-01 -1.63253978e-01 6.58895850e-01 -5.00681996e-01
-1.58892065e-01 7.01550424e-01 -5.36046565e-01 3.67990255e-01
-6.18194826e-02 3.42715770e-01 6.27453923e-01 -8.32845092e-01
2.69959062e-01 6.31413817e-01 8.00520718e-01 4.00872111e-01
-1.17359293e+00 -4.21789736e-01 -1.43148333e-01 -4.22277004e-01
-1.40605855e+00 -4.85443845e-02 4.94642913e-01 -3.33427995e-01
5.58334768e-01 2.95389406e-02 9.06399250e-01 7.72611439e-01
3.20926040e-01 1.10905766e+00 9.16319907e-01 -2.05356807e-01
4.70919348e-02 -1.83168396e-01 -5.95462695e-02 1.44523549e+00
2.44167924e-01 6.33437932e-03 9.68339518e-02 -4.88483347e-02
1.15903783e+00 5.43706417e-02 -6.26574039e-01 -6.32877290e-01
-1.55546999e+00 8.58496785e-01 1.09454048e+00 3.12694252e-01
-6.08359039e-01 5.52060828e-03 3.50470513e-01 -2.19648615e-01
4.21560913e-01 2.98000783e-01 -4.97173697e-01 4.10602003e-01
-1.12004089e+00 -2.04619795e-01 8.80970895e-01 5.99475086e-01
2.23448485e-01 -5.78840673e-02 -4.55734968e-01 5.28203368e-01
5.98317623e-01 3.63434017e-01 4.43767965e-01 -5.11126697e-01
8.84024128e-02 4.99902964e-01 -4.31642354e-01 -3.26344311e-01
-8.87434900e-01 -1.06004596e+00 -1.37935615e+00 1.50402710e-01
6.02638066e-01 -1.84650213e-01 -1.39159942e+00 1.44925725e+00
5.52622199e-01 2.83111125e-01 -8.97656903e-02 1.31399536e+00
1.04071820e+00 1.48378327e-01 3.27391446e-01 3.60493292e-03
1.48944044e+00 -1.19712281e+00 -3.66682351e-01 1.14275031e-01
8.75521541e-01 -4.75902140e-01 6.69904530e-01 8.35839808e-02
-8.79869878e-01 -2.16051891e-01 -9.29313302e-01 8.64882991e-02
2.67751455e-01 3.78789246e-01 9.49886322e-01 5.54166019e-01
-1.09642279e+00 4.11617190e-01 -1.44314313e+00 -1.87806547e-01
1.04289663e+00 5.80400467e-01 -1.96785569e-01 -2.45405927e-01
-8.77634406e-01 9.24187839e-01 3.37379485e-01 1.92958981e-01
-1.32483959e+00 -1.06398773e+00 -1.18661296e+00 4.56417259e-03
3.46088380e-01 -1.09296155e+00 1.42747808e+00 -5.21650910e-01
-1.63796568e+00 1.08124423e+00 1.58815712e-01 -6.81955695e-01
7.73601353e-01 1.98753268e-01 2.45481372e-01 4.34487790e-01
1.48716629e-01 6.96327865e-01 4.55067933e-01 -7.48423278e-01
-2.03971297e-01 -4.07335699e-01 -3.24727088e-01 2.90367156e-01
7.47553930e-02 -2.66454071e-01 -7.12192595e-01 -8.88122380e-01
2.79168516e-01 -8.84262562e-01 -8.26595902e-01 3.95674288e-01
-5.26019394e-01 -1.94691494e-01 2.80555874e-01 -8.62425447e-01
7.07145035e-01 -1.91887510e+00 2.22854763e-01 3.67900729e-01
5.45156777e-01 3.55797201e-01 1.52153119e-01 -5.35644293e-01
3.10678463e-02 -2.07793135e-02 -6.59625828e-01 -3.75507563e-01
-1.00818180e-01 1.07671872e-01 3.88123900e-01 8.11795115e-01
9.71600413e-02 1.23469591e+00 -9.79744554e-01 -7.58999705e-01
3.65010560e-01 6.17935777e-01 -6.18082941e-01 3.58040988e-01
-7.18282536e-02 9.27992523e-01 -6.84099138e-01 7.45045304e-01
5.71315765e-01 -7.67715931e-01 2.16868564e-01 -2.41719708e-01
1.32967785e-01 1.97553903e-01 -5.23195744e-01 2.41426277e+00
-4.38228428e-01 1.86155647e-01 3.52800548e-01 -1.12613595e+00
6.06176257e-01 3.77696067e-01 8.20103526e-01 -7.00602472e-01
4.62278545e-01 2.54362017e-01 1.08218558e-01 -4.21368062e-01
-2.43736833e-01 -2.75537610e-01 2.43058279e-02 3.60888690e-01
1.96244806e-01 -4.40473557e-01 -1.78067997e-01 1.90875053e-01
8.90899658e-01 7.67980516e-02 4.20717657e-01 -8.59462917e-01
7.01623976e-01 3.07239145e-02 7.81633973e-01 5.98376453e-01
-5.62862039e-01 5.78056276e-01 4.13619369e-01 -6.25722647e-01
-8.18967879e-01 -1.06568265e+00 -4.81465548e-01 4.48415518e-01
1.94206610e-01 -1.00957021e-01 -8.55264902e-01 -1.30817747e+00
6.28939345e-02 2.09716544e-01 -6.55244052e-01 1.19706772e-01
-6.64839149e-01 -1.23289657e+00 5.53442836e-01 5.86912930e-01
6.29908025e-01 -7.23331451e-01 -5.14970124e-01 2.77382344e-01
-4.73102033e-02 -1.03838193e+00 -5.75447142e-01 3.33886862e-01
-1.15365672e+00 -1.18965435e+00 -1.16310632e+00 -1.13602519e+00
9.68737662e-01 6.25703558e-02 1.37826204e+00 3.55852693e-01
-5.84927499e-01 -2.16598399e-02 -4.60530631e-02 -3.51155162e-01
-4.49202925e-01 2.44251236e-01 -1.93371832e-01 -4.11847353e-01
8.79362449e-02 -2.65237719e-01 -1.02834404e+00 1.97217196e-01
-6.76034510e-01 3.35188836e-01 8.25062573e-01 1.08989978e+00
8.96548212e-01 -1.19862087e-01 4.28899050e-01 -8.91852200e-01
8.83486494e-02 -3.36319059e-01 -8.41864765e-01 2.09566325e-01
-4.77236360e-01 1.83224320e-01 4.76772964e-01 -7.96484724e-02
-8.55006039e-01 5.66676617e-01 -5.38093805e-01 -3.78329575e-01
-7.86551312e-02 6.54495895e-01 1.33877292e-01 -4.49413806e-01
4.24002349e-01 1.40027270e-01 4.07265425e-01 -4.93726283e-01
6.97967410e-02 2.90036291e-01 5.31556726e-01 -3.08121651e-01
6.75961375e-01 5.16390204e-01 2.07406715e-01 -3.51262510e-01
-1.03291559e+00 -4.03118521e-01 -7.01257467e-01 2.88774781e-02
1.13183081e+00 -1.24300992e+00 -4.56777394e-01 6.53288066e-01
-7.34669089e-01 -7.95812130e-01 -2.28987902e-01 8.49015415e-01
-3.82493109e-01 3.48759025e-01 -1.17252028e+00 -1.14609078e-01
-8.60718250e-01 -1.63101041e+00 1.01615810e+00 4.85455006e-01
2.19589338e-01 -1.26513577e+00 -1.32379636e-01 2.92690337e-01
6.67892814e-01 5.05902886e-01 7.70614982e-01 -7.52314508e-01
-7.51467526e-01 -1.39626205e-01 -4.75299746e-01 1.88190177e-01
7.17699081e-02 -3.96770716e-01 -6.65944874e-01 -5.63400567e-01
-4.44859564e-02 -2.57822156e-01 1.10731268e+00 8.49295557e-01
1.45533597e+00 9.80554074e-02 -3.90023798e-01 1.26793420e+00
1.31368709e+00 -3.39935541e-01 1.32059574e-01 3.77459787e-02
7.07734168e-01 -6.03318997e-02 3.46326411e-01 2.78810292e-01
5.86736083e-01 3.83760035e-01 6.26535058e-01 -8.61209214e-01
-4.22875613e-01 -3.00414767e-02 -2.30123475e-01 8.63281667e-01
2.52320021e-01 1.32285133e-01 -1.24087250e+00 4.48714703e-01
-1.41514289e+00 -2.39128053e-01 -1.09659441e-01 1.93459952e+00
8.86387527e-01 9.22170803e-02 -1.41768172e-01 -2.95254409e-01
4.55012739e-01 5.74889816e-02 -5.57067513e-01 3.33880663e-01
2.59350121e-01 4.03244555e-01 5.29175758e-01 5.03916204e-01
-1.53364933e+00 6.57661617e-01 6.08415508e+00 7.33096480e-01
-1.30700636e+00 1.29984722e-01 1.01553726e+00 2.07787260e-01
6.64245859e-02 -5.30494332e-01 -4.39402372e-01 2.90036917e-01
3.62962097e-01 5.70638850e-02 7.04832226e-02 8.38560522e-01
8.88165906e-02 -1.62901029e-01 -1.03544629e+00 9.47149634e-01
4.93653901e-02 -1.47709453e+00 -2.09942609e-01 9.72780958e-02
7.45807469e-01 5.68320453e-01 -7.95846805e-02 3.18460554e-01
3.32555175e-01 -1.11569548e+00 1.33947402e-01 2.89657086e-01
1.14009166e+00 -6.34691596e-01 1.14367580e+00 4.27443713e-01
-1.04523933e+00 3.00172895e-01 -9.89028066e-02 4.16622549e-01
1.42293990e-01 9.41780090e-01 -1.26104009e+00 7.66329706e-01
4.37059075e-01 8.72070968e-01 -4.77695882e-01 1.47480440e+00
-5.05947769e-01 6.25802934e-01 -3.50611567e-01 2.26064473e-01
3.08425277e-01 -7.36961095e-03 4.43330765e-01 1.37675846e+00
2.00230211e-01 3.86956215e-01 5.74923635e-01 7.50967264e-01
-5.34124494e-01 -6.48569837e-02 -8.70611742e-02 3.15325916e-01
-3.73013802e-02 1.57429838e+00 -1.10673618e+00 -5.34115314e-01
-2.56558985e-01 9.54216003e-01 1.51514053e-01 3.20160538e-02
-8.60708296e-01 4.84817438e-02 3.31067890e-01 -2.09347844e-01
2.40332887e-01 -1.53249621e-01 -3.56406301e-01 -1.54282463e+00
-3.48704547e-01 -5.83847642e-01 6.55221879e-01 -2.33810216e-01
-1.40039456e+00 7.15421438e-01 -5.30033112e-01 -1.00434959e+00
7.97266513e-03 -6.29240215e-01 -6.19048238e-01 1.01674283e+00
-2.08894324e+00 -1.07385159e+00 -6.00292802e-01 5.80828607e-01
4.19957399e-01 7.75035024e-02 9.45412755e-01 4.00618583e-01
-5.29319406e-01 5.59420109e-01 -1.40564740e-01 8.00071955e-01
5.43296218e-01 -1.52363455e+00 1.93343371e-01 7.67982960e-01
-1.70623288e-01 1.41966969e-01 4.09271345e-02 -5.00722826e-01
-1.43800855e+00 -1.32730687e+00 3.49180877e-01 -2.70210683e-01
3.35384965e-01 -5.66374175e-02 -8.68376613e-01 5.27504981e-01
1.11661099e-01 7.31894433e-01 8.57812643e-01 -2.70386189e-01
-1.78590696e-02 9.70175192e-02 -1.40469396e+00 1.11590870e-01
8.34099829e-01 -2.15628132e-01 -5.26683033e-01 5.03946245e-01
6.49902284e-01 -1.11510801e+00 -1.07024360e+00 8.10344338e-01
2.25039646e-01 -7.13988006e-01 1.04719830e+00 -2.97936559e-01
2.22727925e-01 -2.49355480e-01 2.71414578e-01 -1.34133399e+00
-3.51660758e-01 -4.40523893e-01 -1.12520717e-01 4.23110783e-01
4.22387898e-01 -3.38132590e-01 9.68607724e-01 4.56984907e-01
-5.00628114e-01 -1.18701410e+00 -8.58007073e-01 -3.24394464e-01
3.32107216e-01 -1.08895056e-01 4.15698260e-01 1.04204106e+00
-1.75497547e-01 3.92870419e-02 1.92541108e-01 4.10152107e-01
1.20091665e+00 3.04883391e-01 3.64401519e-01 -1.08774877e+00
-2.18637049e-01 -5.68301439e-01 -3.21381092e-01 -1.32463324e+00
5.12656197e-03 -1.46468449e+00 1.71232924e-01 -1.64020836e+00
5.93968868e-01 -6.34467602e-01 -5.90737402e-01 4.64113444e-01
-5.30623317e-01 5.37303448e-01 1.12058349e-01 1.01210296e-01
-8.09051871e-01 5.46843708e-01 1.80497968e+00 -2.16511413e-01
-1.05399497e-01 1.86567143e-01 -5.11849582e-01 8.07194531e-01
6.78744972e-01 -4.59481031e-01 1.00131467e-01 -6.22992814e-01
-4.52001899e-01 5.19536972e-01 4.10454959e-01 -8.66208375e-01
3.26195687e-01 3.28849286e-01 8.99522424e-01 -5.77749074e-01
-1.63489923e-01 -7.59119868e-01 -3.63477200e-01 9.89339054e-01
-1.76100314e-01 -4.28671509e-01 1.02429144e-01 3.63598615e-01
-1.59035847e-01 1.86831683e-01 1.18093610e+00 -2.84833670e-01
-2.34535575e-01 9.44481373e-01 -1.86427161e-01 1.81756262e-02
1.07699203e+00 3.77649724e-01 2.51976550e-01 1.42465727e-02
-8.60286713e-01 7.30928361e-01 2.55404711e-01 -8.93165991e-02
5.87805629e-01 -1.27048159e+00 -1.01567090e+00 4.34753746e-01
-2.02187419e-01 8.57098937e-01 1.79906636e-01 1.40625536e+00
-9.36572552e-01 3.94015878e-01 1.26472861e-01 -1.09563518e+00
-9.02880967e-01 3.12175095e-01 8.79479289e-01 -7.78668523e-01
-1.06638563e+00 1.03397548e+00 2.96247661e-01 -6.37915075e-01
3.67476970e-01 -5.49921632e-01 1.53213842e-02 -3.91165555e-01
3.92172694e-01 -2.86718488e-01 1.93254411e-01 -3.42833012e-01
-4.58737463e-01 3.28904808e-01 -8.68040770e-02 2.66473740e-01
1.53134668e+00 9.10843462e-02 -4.07250039e-02 -4.11687106e-01
1.08576071e+00 -3.60408247e-01 -1.49379957e+00 -6.25727296e-01
2.86639994e-03 -3.28720838e-01 6.66786134e-01 -1.21583080e+00
-1.58621478e+00 8.72377217e-01 5.49018681e-01 -2.48514563e-01
1.23718309e+00 3.41399992e-03 9.78128850e-01 7.26980418e-02
3.29276711e-01 -3.96320343e-01 -1.82498500e-01 3.97048384e-01
4.71769899e-01 -1.69173050e+00 1.56298995e-01 -4.77618217e-01
-4.14436519e-01 1.17320263e+00 4.79019463e-01 -1.89796329e-01
6.56689703e-01 4.67918485e-01 3.12510580e-01 -2.04602033e-01
-2.34829918e-01 -2.06229746e-01 4.10691977e-01 2.65077829e-01
3.80818903e-01 2.05738530e-01 -3.34096253e-02 7.05178738e-01
2.08449483e-01 9.57975835e-02 1.40676975e-01 6.73312724e-01
-2.49588504e-01 -8.31384897e-01 -4.77219857e-02 4.17315304e-01
-7.63677537e-01 -1.83265209e-01 2.80144691e-01 7.82740951e-01
-2.75761008e-01 2.78396308e-01 -1.49177566e-01 2.03049749e-01
4.38930281e-02 -1.28869846e-01 5.32319665e-01 -5.31164408e-01
-8.40769231e-01 3.02575588e-01 -4.18788075e-01 -6.75116360e-01
-2.94404328e-01 -5.13391972e-01 -1.62338924e+00 -1.52155915e-02
-3.27686518e-01 2.49743357e-01 8.50596786e-01 7.93436289e-01
1.77980483e-01 8.42188299e-01 8.42278004e-01 -9.81477499e-01
-8.15472245e-01 -8.48865747e-01 -3.38324547e-01 3.37703615e-01
5.57797551e-01 -2.71887064e-01 -1.79828316e-01 -1.42575458e-01] | [14.510231018066406, -2.6239445209503174] |
759ac6f0-c53f-4eab-a5b1-601c99b77695 | cross-document-language-modeling | 2101.00406 | null | https://arxiv.org/abs/2101.00406v2 | https://arxiv.org/pdf/2101.00406v2.pdf | CDLM: Cross-Document Language Modeling | We introduce a new pretraining approach geared for multi-document language modeling, incorporating two key ideas into the masked language modeling self-supervised objective. First, instead of considering documents in isolation, we pretrain over sets of multiple related documents, encouraging the model to learn cross-document relationships. Second, we improve over recent long-range transformers by introducing dynamic global attention that has access to the entire input to predict masked tokens. We release CDLM (Cross-Document Language Model), a new general language model for multi-document setting that can be easily applied to downstream tasks. Our extensive analysis shows that both ideas are essential for the success of CDLM, and work in synergy to set new state-of-the-art results for several multi-text tasks. Code and models are available at https://github.com/aviclu/CDLM. | ['Ido Dagan', 'Arie Cattan', 'Matthew E. Peters', 'Iz Beltagy', 'Arman Cohan', 'Avi Caciularu'] | 2021-01-02 | null | https://aclanthology.org/2021.findings-emnlp.225 | https://aclanthology.org/2021.findings-emnlp.225.pdf | findings-emnlp-2021-11 | ['cross-document-language-modeling'] | ['natural-language-processing'] | [-1.29737318e-01 6.60477951e-02 -3.78894120e-01 -5.84331155e-01
-1.32396889e+00 -6.63814783e-01 9.87320125e-01 1.89716741e-01
-4.20879364e-01 5.34296334e-01 4.42278892e-01 -4.49515283e-01
2.28520334e-01 -2.09472597e-01 -8.88215125e-01 -3.97567481e-01
-3.52830768e-01 6.22661412e-01 1.67350426e-01 -2.69771338e-01
-1.04038067e-01 2.81626552e-01 -9.19709802e-01 8.82227480e-01
3.45340550e-01 2.29218245e-01 2.96902895e-01 8.31289947e-01
-3.12400281e-01 1.01561606e+00 -4.26232576e-01 -4.07061249e-01
-6.51258752e-02 -1.87456682e-01 -9.31598067e-01 -4.25720364e-02
4.72927779e-01 -2.36907452e-01 -2.68331409e-01 6.61586702e-01
4.21633154e-01 5.18738553e-02 4.68427449e-01 -9.12299454e-01
-1.05199480e+00 1.41129410e+00 -6.07906640e-01 3.67575496e-01
-8.89271945e-02 -1.22669145e-01 1.39681482e+00 -1.25510240e+00
2.42201567e-01 1.49789119e+00 5.14049470e-01 5.87490499e-01
-1.46116090e+00 -6.20689154e-01 9.04451311e-01 -1.22676022e-01
-1.37146032e+00 -6.86038971e-01 5.75888813e-01 -4.31867808e-01
1.51956010e+00 -1.32381199e-02 -4.46985429e-03 1.32153559e+00
2.25486159e-01 1.23232830e+00 8.97489071e-01 -8.67715597e-01
-4.36765194e-01 3.34428281e-01 4.73300010e-01 7.26088226e-01
-2.76656728e-02 -1.81698844e-01 -6.53041422e-01 -1.71663553e-01
2.01298669e-01 -9.05733928e-02 6.48659095e-02 -1.09485192e-02
-1.18350875e+00 8.69411409e-01 1.46595269e-01 7.05570221e-01
8.25374275e-02 4.07126188e-01 3.28617781e-01 3.59623492e-01
8.91041040e-01 1.93104848e-01 -5.76641500e-01 8.66023973e-02
-1.13330090e+00 1.75099775e-01 4.58463341e-01 8.52067232e-01
7.33558893e-01 -5.92085831e-02 -3.55541378e-01 8.67218256e-01
6.41059577e-01 2.10508212e-01 6.01620138e-01 -3.98588151e-01
6.18440986e-01 1.88552156e-01 -1.71417162e-01 -4.92416412e-01
-4.37711746e-01 -6.38695717e-01 -6.38222158e-01 -1.22582793e-01
1.13357514e-01 -1.77684098e-01 -9.45887864e-01 1.83109844e+00
-2.28643715e-01 1.39790988e-02 -5.75857842e-03 2.94212073e-01
5.37798524e-01 8.34772587e-01 1.35603547e-01 1.22641914e-01
1.17908037e+00 -1.66038871e+00 -5.05742848e-01 -7.77474105e-01
1.03927100e+00 -8.20137978e-01 1.16340888e+00 4.10789013e-01
-1.29411507e+00 -6.76299572e-01 -9.55443323e-01 -5.09956121e-01
-6.60449028e-01 1.97687492e-01 5.91028750e-01 2.57139206e-01
-1.35614336e+00 2.68863857e-01 -1.05304670e+00 -2.00699702e-01
1.59712017e-01 3.75765055e-01 -1.17104687e-01 -1.25000477e-01
-1.42093873e+00 9.56478715e-01 3.33476424e-01 1.52148921e-02
-9.22894895e-01 -5.58362007e-01 -7.21178889e-01 3.19678113e-02
5.62248714e-02 -4.98389602e-01 1.50006270e+00 -7.31600702e-01
-1.08634460e+00 1.20565748e+00 -5.31853318e-01 -5.40542662e-01
4.53514725e-01 -5.42412341e-01 -4.58127081e-01 -3.02572578e-01
1.34569198e-01 6.81074083e-01 5.41182935e-01 -1.43828475e+00
-4.47757035e-01 -9.76322517e-02 -1.81737304e-01 1.15951709e-01
-5.67368329e-01 4.76283133e-01 -7.66857386e-01 -7.67789185e-01
-3.92744482e-01 -7.24574268e-01 -1.85220376e-01 -4.48406518e-01
-6.06800973e-01 -3.22879523e-01 5.56264281e-01 -5.60157180e-01
1.35462999e+00 -2.12353420e+00 1.99516654e-01 1.25051010e-02
1.55339554e-01 2.39470243e-01 -5.69820046e-01 9.25316811e-01
-3.00467461e-01 3.18448216e-01 -2.21926540e-01 -1.37782538e+00
1.16682932e-01 1.35160208e-01 -6.81525171e-01 1.50636911e-01
4.28627700e-01 1.22193873e+00 -5.44510663e-01 -3.21766049e-01
-1.40131950e-01 4.94628876e-01 -3.43948781e-01 5.91590889e-02
-5.31663477e-01 2.30016172e-01 -2.05764323e-01 3.42999548e-01
4.83604759e-01 -5.23184597e-01 1.49382055e-01 3.08638334e-01
-9.37723368e-02 7.54186749e-01 -9.22703981e-01 1.74745655e+00
-5.30748904e-01 4.93045896e-01 1.56888753e-01 -8.50688159e-01
6.04141772e-01 3.16166103e-01 2.29016602e-01 -4.28566664e-01
-1.28691405e-01 4.31528548e-03 -1.80688247e-01 -1.01902366e-01
5.87597549e-01 -1.21895544e-01 -1.20895833e-01 8.95008862e-01
1.93584651e-01 9.46355835e-02 2.29305565e-01 7.32806563e-01
7.26854801e-01 -4.05038409e-02 1.09667271e-01 -2.00906157e-01
4.98747170e-01 -1.80356428e-01 -1.43128932e-01 1.02555895e+00
3.55872273e-01 4.50386047e-01 3.33519518e-01 -1.93280112e-02
-7.72923946e-01 -8.37901831e-01 -4.69684191e-02 1.64612091e+00
-5.09430826e-01 -7.13185430e-01 -3.13459486e-01 -5.71137607e-01
2.46368453e-01 6.91277683e-01 -5.96401811e-01 -1.10562481e-01
-6.63198233e-01 -9.49738979e-01 8.38812709e-01 7.73060620e-01
-1.76932022e-01 -7.29985118e-01 4.02978420e-01 2.15029985e-01
6.33966131e-03 -1.07135785e+00 -7.38299549e-01 7.55720377e-01
-5.46837866e-01 -3.79580319e-01 -5.26171267e-01 -1.09894633e+00
2.90809959e-01 2.90952861e-01 1.36628711e+00 3.33789915e-01
-9.10995528e-03 3.06510150e-01 -5.33090867e-02 -3.36898297e-01
-7.67413259e-01 4.60585028e-01 3.91044915e-02 -1.81582943e-01
4.94166493e-01 -4.47925389e-01 3.24387327e-02 -1.17505446e-01
-8.12796474e-01 7.18102753e-02 6.34207845e-01 8.00946355e-01
3.93772423e-01 -1.21768251e-01 4.46453094e-01 -1.04342782e+00
8.98625374e-01 -6.56204581e-01 -3.63983989e-01 3.61239910e-01
-6.18261397e-01 2.23625779e-01 6.15447998e-01 -4.69047219e-01
-7.01251328e-01 -2.67005742e-01 -1.90802187e-01 -2.04589456e-01
-7.67331719e-02 7.71265388e-01 6.55684471e-02 3.73591602e-01
3.60059470e-01 3.24762017e-01 -1.92931339e-01 -7.89179206e-01
6.96215749e-01 7.41638839e-01 4.00687844e-01 -6.28525376e-01
7.42648005e-01 4.27008778e-01 -3.55678260e-01 -5.22554815e-01
-1.03201866e+00 -6.48405731e-01 -8.63353312e-01 4.30988312e-01
6.50896728e-01 -1.51880956e+00 -3.17918479e-01 4.68863845e-01
-1.33857512e+00 -7.37936854e-01 1.40795693e-01 3.00999910e-01
-1.20279208e-01 3.62090737e-01 -9.70440567e-01 -7.62185097e-01
-2.92967975e-01 -7.57582426e-01 1.18524826e+00 -4.12484646e-01
-2.76968420e-01 -1.56251132e+00 2.87137568e-01 3.18222582e-01
3.46242100e-01 -5.59613645e-01 1.05660748e+00 -1.03823185e+00
-3.91977698e-01 -7.42993131e-02 -7.77618065e-02 4.94809389e-01
1.24361821e-01 -6.44867793e-02 -1.14722478e+00 -6.10261202e-01
-3.49659652e-01 -5.81326544e-01 1.43892384e+00 1.81407735e-01
9.13656950e-01 -2.44084999e-01 -6.79168105e-01 3.78987968e-01
1.14572442e+00 -3.88875365e-01 -5.71308360e-02 2.37479419e-01
8.85696173e-01 4.13957119e-01 9.23144519e-02 1.63661793e-01
8.48284483e-01 6.68951869e-01 -2.94434894e-02 -4.25132036e-01
-2.03672409e-01 -4.04431164e-01 7.05728889e-01 1.01395524e+00
5.12635589e-01 -7.06363142e-01 -1.16931343e+00 7.55478919e-01
-1.90764427e+00 -7.92274654e-01 -1.71490461e-01 1.74635911e+00
1.15138471e+00 4.71926630e-01 1.95467189e-01 -3.31863433e-01
3.81649345e-01 3.22215140e-01 -3.13255340e-01 -5.64825892e-01
-3.35693657e-01 1.84729710e-01 4.15684164e-01 1.14663017e+00
-1.17928588e+00 1.38464546e+00 7.05041361e+00 6.31001115e-01
-1.21226132e+00 3.69545400e-01 6.07766271e-01 -4.61745620e-01
-4.78184521e-01 -6.34789541e-02 -1.43103111e+00 3.36780250e-01
1.30255616e+00 -2.11934149e-01 4.23620015e-01 5.31189024e-01
2.75685251e-01 1.87741935e-01 -1.27165115e+00 4.63561475e-01
2.80360103e-01 -1.27274740e+00 2.86461681e-01 -7.95058236e-02
6.85324728e-01 7.41840363e-01 2.81068146e-01 4.04148698e-01
7.60209501e-01 -1.21180046e+00 7.79443681e-01 3.01297635e-01
5.41649818e-01 -6.05175257e-01 3.29748571e-01 5.03017366e-01
-9.85222638e-01 1.15436986e-01 -2.38755465e-01 -7.08145574e-02
2.61364013e-01 5.65693915e-01 -8.84742558e-01 4.98904794e-01
5.73593378e-01 9.23506856e-01 -8.49288583e-01 3.16802561e-01
-2.26473704e-01 7.96258390e-01 -1.54413819e-01 2.03381896e-01
6.45245969e-01 3.14774543e-01 3.52593392e-01 1.86339355e+00
-1.30264848e-01 -5.41818082e-01 4.71739292e-01 9.32186067e-01
-1.89624697e-01 4.75330576e-02 -4.23316121e-01 -3.58597189e-01
2.47340694e-01 1.15095961e+00 -3.02451164e-01 -4.92868334e-01
-7.58459628e-01 1.04233420e+00 8.19176376e-01 3.66923451e-01
-6.41996086e-01 -8.12713429e-02 6.87541425e-01 4.52006981e-02
4.84979331e-01 -6.03247762e-01 -1.73605829e-01 -1.26212239e+00
-3.71412262e-02 -8.69699597e-01 4.34053540e-01 -6.09928191e-01
-1.44020271e+00 7.90989876e-01 -3.44839245e-02 -7.04966128e-01
-5.44670641e-01 -7.63435960e-01 -6.62908792e-01 1.19526112e+00
-1.84757042e+00 -1.64404166e+00 4.31139529e-01 5.38095176e-01
6.45351589e-01 -1.11602314e-01 1.00053668e+00 2.93546945e-01
-4.77224976e-01 8.21525574e-01 4.66984987e-01 2.60837346e-01
1.03705394e+00 -1.29956520e+00 9.74093020e-01 9.94801104e-01
5.72110295e-01 1.04057348e+00 3.33079368e-01 -5.07422209e-01
-1.12802601e+00 -1.18916154e+00 1.47920263e+00 -9.83350098e-01
1.17206430e+00 -1.07361889e+00 -1.05640924e+00 1.41161895e+00
5.67139626e-01 -9.33287963e-02 8.98117006e-01 4.78577554e-01
-5.46686649e-01 1.95207939e-01 -3.84814948e-01 5.80972672e-01
7.85345137e-01 -8.21774602e-01 -4.77747560e-01 6.12913966e-01
1.01565957e+00 -2.43839964e-01 -6.22663319e-01 -1.03325143e-01
2.96884745e-01 -5.42952120e-01 9.35191453e-01 -1.06734300e+00
4.11504537e-01 -8.81166533e-02 -2.39096344e-01 -1.37591922e+00
-5.57350159e-01 -7.69535124e-01 -3.44404429e-01 1.39213657e+00
9.36388373e-01 -6.49592340e-01 3.40997577e-01 2.31377512e-01
-1.94331914e-01 -7.78078914e-01 -5.61724961e-01 -9.17276680e-01
8.70122254e-01 -4.99150157e-01 2.53849506e-01 9.94467497e-01
1.87261626e-01 7.13927984e-01 -6.41114652e-01 3.80196393e-01
4.07398701e-01 6.22357577e-02 4.99871880e-01 -9.66780603e-01
-8.47434044e-01 -5.54374099e-01 2.06216261e-01 -1.38947940e+00
5.83558977e-01 -1.44982314e+00 -3.00604999e-02 -1.57269216e+00
4.03092742e-01 -2.61076808e-01 -5.96350193e-01 8.91908765e-01
-2.73193598e-01 3.18713605e-01 2.45552406e-01 2.95773655e-01
-7.18876064e-01 3.11255783e-01 7.35014021e-01 -2.36009285e-01
-8.62668455e-02 -5.67291416e-02 -1.14617264e+00 4.02142853e-01
8.43708992e-01 -6.26107931e-01 -2.35999122e-01 -1.11160064e+00
1.70097902e-01 -4.69301343e-01 2.23963290e-01 -5.72345853e-01
4.08365548e-01 1.51490211e-01 3.46417606e-01 -6.92875504e-01
4.81466174e-01 -3.33353490e-01 -3.94058943e-01 2.57810324e-01
-7.71622360e-01 3.08957517e-01 6.74712479e-01 3.58434886e-01
-2.25827530e-01 -7.30792657e-02 6.73262596e-01 -3.45426440e-01
-3.93164605e-01 1.31257564e-01 -5.42451322e-01 1.18353836e-01
5.27363181e-01 3.04396749e-01 -4.05758232e-01 -4.19706464e-01
-7.52709031e-01 5.21266937e-01 2.29246691e-01 8.42120171e-01
1.83055967e-01 -9.61225867e-01 -9.57963407e-01 3.53935778e-01
1.82870105e-01 -5.77476062e-02 -1.39764860e-01 6.97997212e-01
2.24037260e-01 1.00452638e+00 6.10341191e-01 -4.82941866e-01
-1.29260409e+00 6.47308648e-01 3.63406032e-01 -6.68743193e-01
-4.15196300e-01 1.14908600e+00 8.24408606e-02 -5.94237566e-01
3.66534680e-01 -4.75899339e-01 1.45076111e-01 7.82131031e-02
6.91887975e-01 -2.45848760e-01 3.25869799e-01 -4.49684948e-01
-5.79449356e-01 4.76073921e-01 -8.66502047e-01 -3.62759918e-01
1.40314043e+00 -4.00636852e-01 -2.73761123e-01 8.69291008e-01
1.33546424e+00 4.75571394e-01 -9.38890755e-01 -6.79024100e-01
2.49648958e-01 1.07836120e-01 1.82199776e-01 -1.06506574e+00
-5.48357964e-01 9.76995945e-01 3.22380327e-02 8.40892419e-02
6.17198527e-01 4.03943211e-01 4.61849302e-01 4.38420177e-01
-1.33582205e-01 -7.03193486e-01 1.47804901e-01 8.65166903e-01
7.86918402e-01 -1.28127158e+00 -3.98055054e-02 3.88050303e-02
-7.73789108e-01 9.15357828e-01 5.87116063e-01 1.44605100e-01
8.57529342e-01 8.15334082e-01 3.25796455e-01 -1.09312586e-01
-1.37827313e+00 -3.37001383e-02 3.36716413e-01 1.16590068e-01
1.12130857e+00 -1.49905026e-01 1.91658869e-01 7.11901426e-01
-9.21689868e-02 -3.08699131e-01 2.91061312e-01 8.70974064e-01
-3.88530999e-01 -1.52340901e+00 -1.17895395e-01 1.29679471e-01
-5.60742259e-01 -7.73881555e-01 -5.41156173e-01 6.17649913e-01
-1.49706289e-01 1.03698182e+00 2.27927104e-01 -1.16762921e-01
-1.66900307e-02 5.30100167e-01 2.77435720e-01 -1.12043750e+00
-6.77744448e-01 2.20458478e-01 7.77797624e-02 -2.28672534e-01
-2.33057365e-01 -5.21639585e-01 -1.29135561e+00 -2.11829796e-01
-2.28499025e-01 2.06951559e-01 5.38734257e-01 9.95230258e-01
6.05999649e-01 5.22213757e-01 3.89552653e-01 -6.83831096e-01
-5.87774336e-01 -1.22946906e+00 -3.79153311e-01 -6.50322139e-02
8.10598433e-01 -1.92021936e-01 -3.83889526e-01 6.68441430e-02] | [10.790103912353516, 8.753933906555176] |
f4f88971-b1c0-451e-8edc-6ea3b61f867e | intelligent-robotic-sonographer-mutual | 2307.03705 | null | https://arxiv.org/abs/2307.03705v1 | https://arxiv.org/pdf/2307.03705v1.pdf | Intelligent Robotic Sonographer: Mutual Information-based Disentangled Reward Learning from Few Demonstrations | Ultrasound (US) imaging is widely used for biometric measurement and diagnosis of internal organs due to the advantages of being real-time and radiation-free. However, due to high inter-operator variability, resulting images highly depend on operators' experience. In this work, an intelligent robotic sonographer is proposed to autonomously "explore" target anatomies and navigate a US probe to a relevant 2D plane by learning from expert. The underlying high-level physiological knowledge from experts is inferred by a neural reward function, using a ranked pairwise image comparisons approach in a self-supervised fashion. This process can be referred to as understanding the "language of sonography". Considering the generalization capability to overcome inter-patient variations, mutual information is estimated by a network to explicitly extract the task-related and domain features in latent space. Besides, a Gaussian distribution-based filter is developed to automatically evaluate and take the quality of the expert's demonstrations into account. The robotic localization is carried out in coarse-to-fine mode based on the predicted reward associated to B-mode images. To demonstrate the performance of the proposed approach, representative experiments for the "line" target and "point" target are performed on vascular phantom and two ex-vivo animal organ phantoms (chicken heart and lamb kidney), respectively. The results demonstrated that the proposed advanced framework can robustly work on different kinds of known and unseen phantoms. | ['and Nassir Navab', 'Michael Burke', 'Ying Hu', 'Mingchuan Zhou', 'Yuan Bi', 'Zhongliang Jiang'] | 2023-07-07 | null | null | null | null | ['navigate'] | ['reasoning'] | [ 2.20549807e-01 2.17089698e-01 1.65862128e-01 -4.82411057e-01
-8.03784132e-01 -4.69512314e-01 -6.50101379e-02 1.09303690e-01
-5.67514658e-01 6.27808154e-01 -2.03371152e-01 1.21674500e-01
-5.89739561e-01 -2.76464880e-01 -6.53533161e-01 -9.62245524e-01
-3.30476612e-01 3.02045822e-01 -7.17175901e-02 1.80203617e-01
1.61550000e-01 6.00769281e-01 -1.00189376e+00 -1.26150213e-02
9.38750148e-01 1.08197546e+00 5.69624364e-01 5.44709861e-01
3.05839092e-01 4.65794355e-01 -1.86041966e-01 -8.32513869e-02
2.40745232e-01 -2.44496182e-01 -7.53753781e-01 3.41448039e-02
-7.61749074e-02 -1.38160765e-01 -2.04896629e-01 1.24080718e+00
9.18754697e-01 1.74641058e-01 7.91869402e-01 -6.92147970e-01
-4.22625750e-01 4.80131507e-01 -4.20336127e-01 8.03558379e-02
4.27431285e-01 2.02311575e-01 6.05344355e-01 -5.81366599e-01
7.57192671e-01 9.86089706e-01 4.29937840e-01 4.99970376e-01
-1.21352303e+00 -4.13049906e-01 -1.51779339e-01 -1.97863905e-03
-1.22059977e+00 -6.66451231e-02 7.88998902e-01 -8.13214064e-01
1.30220935e-01 2.00539678e-01 4.84779030e-01 1.12404037e+00
8.67755890e-01 5.20536423e-01 1.35309339e+00 -1.28970504e-01
3.11947227e-01 5.38117588e-01 -1.32424653e-01 7.58414507e-01
-1.02355346e-01 2.81493813e-01 -3.29574436e-01 -2.54744947e-01
1.04645860e+00 -3.72181952e-01 -6.40079439e-01 -8.39769602e-01
-1.39114463e+00 5.26556969e-01 6.57849431e-01 5.41208684e-01
-7.23094463e-01 -5.06785959e-02 3.15614492e-01 4.73416299e-02
4.13767062e-02 5.15591681e-01 -4.06646281e-01 1.38320159e-02
-3.39842111e-01 -1.92966864e-01 9.22622561e-01 7.38932908e-01
3.11156392e-01 -1.83355838e-01 -3.59515548e-01 5.65764070e-01
5.25863528e-01 5.23725092e-01 6.93094909e-01 -9.53181148e-01
1.13395199e-01 2.61983037e-01 4.26291823e-01 -1.16563988e+00
-6.96200609e-01 -8.18884015e-01 -8.78794670e-01 1.83570519e-01
3.60010892e-01 -1.70921475e-01 -6.77853644e-01 1.72368324e+00
5.84579229e-01 3.78661007e-01 1.98177874e-01 1.31703985e+00
7.45707572e-01 1.49194628e-01 3.23248595e-01 -3.52111459e-01
1.33941817e+00 -4.58427101e-01 -4.78380084e-01 2.57101115e-02
4.53801274e-01 -4.65662479e-01 1.00397253e+00 3.92466396e-01
-7.40065336e-01 -7.09724545e-01 -7.85133958e-01 6.94014966e-01
1.42102987e-01 4.26629394e-01 3.97561461e-01 5.93327224e-01
-5.71225166e-01 8.23094487e-01 -1.10704505e+00 -3.49734306e-01
3.17864716e-01 3.91897321e-01 -5.08112252e-01 7.73005933e-02
-1.26436186e+00 9.99515653e-01 3.57445598e-01 6.42965853e-01
-9.79597330e-01 -7.93576598e-01 -7.92846143e-01 -6.22245483e-02
8.95372033e-02 -7.82335699e-01 7.68954098e-01 -6.74688756e-01
-1.89640486e+00 7.67267227e-01 2.04328462e-01 -1.78826660e-01
6.64139032e-01 -1.46941915e-01 -2.03701347e-01 4.67840672e-01
5.60514070e-02 4.73498136e-01 6.23160720e-01 -1.32667029e+00
-2.26182044e-02 -4.95288849e-01 -2.25965530e-01 4.35788006e-01
-1.45168379e-01 -2.91157603e-01 -1.19157977e-01 -3.81996989e-01
5.76340497e-01 -1.13864708e+00 -3.81520897e-01 2.33924180e-01
-2.39461794e-01 1.13354921e-01 1.27524674e-01 -7.88069427e-01
6.85673952e-01 -2.12718129e+00 3.46381932e-01 6.38167143e-01
5.03078401e-02 -1.68632746e-01 -1.64962009e-01 -5.33690006e-02
7.94956163e-02 -3.55075270e-01 -2.24089012e-01 1.99884340e-01
-2.02389091e-01 -1.01419650e-02 4.07340944e-01 8.54778707e-01
-1.62970647e-01 7.01413393e-01 -1.15168548e+00 -7.05621064e-01
2.88672507e-01 3.29728276e-01 -2.23110735e-01 3.18418652e-01
8.45920742e-02 1.46053112e+00 -9.01730657e-01 4.41852152e-01
6.55252457e-01 -3.06483388e-01 2.44319230e-01 -6.84385955e-01
1.34227350e-01 -4.39113826e-01 -1.12914515e+00 2.22296786e+00
-5.38434982e-01 9.46046934e-02 2.75364816e-01 -1.05266178e+00
8.69118273e-01 5.76998472e-01 7.17481077e-01 -5.26607513e-01
4.51555043e-01 3.41874063e-01 3.52019578e-01 -1.02422726e+00
-2.85487980e-01 -5.03035374e-02 5.04626222e-02 7.11793453e-02
1.42086089e-01 -1.68834314e-01 -2.82951027e-01 -8.18260536e-02
8.90873790e-01 2.77932882e-01 1.41137213e-01 -5.76404035e-01
8.00930917e-01 -2.74934560e-01 3.94712806e-01 7.83595085e-01
-4.92443621e-01 2.97706872e-01 2.01421350e-01 -1.63185149e-01
-4.77367133e-01 -1.20246065e+00 -4.56539571e-01 4.74752009e-01
6.31785631e-01 5.03812075e-01 -5.63269436e-01 -7.40424156e-01
8.89464095e-02 5.36324501e-01 -8.45587075e-01 -2.00121731e-01
-1.98704436e-01 -4.29946840e-01 3.00627917e-01 2.12606296e-01
2.90695667e-01 -1.15145016e+00 -8.03090751e-01 3.08687627e-01
-2.32828617e-01 -1.08676064e+00 -6.97905496e-02 1.52437001e-01
-8.29385519e-01 -8.89749825e-01 -1.04420412e+00 -6.74821258e-01
7.75295317e-01 -4.00603771e-01 5.63262522e-01 -5.34093738e-01
-6.51509166e-01 7.76029706e-01 -1.74285620e-01 -1.70537636e-01
-5.57620049e-01 -2.82938629e-01 1.20755143e-01 1.61289155e-01
-1.24097943e-01 -5.15066683e-01 -1.02006364e+00 5.63423216e-01
-5.19715607e-01 -3.05140138e-01 8.99262667e-01 1.15760553e+00
5.91724336e-01 8.56281444e-03 5.49723625e-01 -5.98735094e-01
5.06372392e-01 -3.56359094e-01 -7.07615018e-01 4.06511575e-01
-2.39286631e-01 1.55566663e-01 1.29990578e-01 -6.28922760e-01
-1.29928505e+00 4.43264484e-01 7.57419243e-02 -4.48680341e-01
-2.96051919e-01 7.91722119e-01 1.56198636e-01 -5.29998243e-01
9.55949366e-01 3.32514912e-01 2.53652245e-01 -5.43393977e-02
5.68446591e-02 4.88258362e-01 5.26525259e-01 -7.70260274e-01
5.07883430e-01 2.26922065e-01 2.87665367e-01 -5.54833055e-01
-5.32987833e-01 -5.10640681e-01 -7.12593317e-01 -5.67215502e-01
9.04007614e-01 -7.79865682e-01 -1.24170196e+00 4.36475664e-01
-9.60571051e-01 -1.77797109e-01 -5.32502010e-02 1.24067140e+00
-7.33702898e-01 6.49969816e-01 -7.66883194e-01 -8.41352344e-01
-2.87067741e-01 -1.51234829e+00 7.97867239e-01 3.14477295e-01
-2.95127295e-02 -1.01803982e+00 -8.97646844e-02 2.60071009e-01
4.04232621e-01 2.53384173e-01 6.81500137e-01 -5.56273282e-01
-4.97676909e-01 -3.10052276e-01 -1.72612488e-01 4.54142451e-01
1.25279069e-01 -5.60671508e-01 -9.76544678e-01 -2.54039526e-01
4.12723422e-01 -4.27324355e-01 2.44377315e-01 9.04545486e-01
1.13819981e+00 1.68917924e-01 -3.64266247e-01 3.13949734e-01
1.27179313e+00 1.94835797e-01 4.69350666e-01 -9.40028578e-02
4.10673976e-01 7.88998902e-01 9.51266706e-01 4.94769692e-01
7.97048062e-02 3.53953719e-01 4.46114540e-01 3.15609737e-03
3.84085566e-01 8.30306411e-02 -9.76405516e-02 5.63655615e-01
-9.76948068e-02 -6.54005483e-02 -6.31841421e-01 2.00725719e-01
-1.62955809e+00 -5.90504646e-01 2.14895904e-01 2.21341276e+00
6.35672629e-01 1.16895130e-02 -5.63675761e-01 -3.71839046e-01
7.29296625e-01 -4.81411248e-01 -6.71531975e-01 1.69090346e-01
2.19363570e-01 -2.80902125e-02 5.26898980e-01 2.94026196e-01
-1.09895527e+00 5.10964453e-01 5.33074188e+00 7.65606046e-01
-1.30137873e+00 -6.69134557e-02 6.93799317e-01 5.40026128e-01
-4.19845507e-02 -3.18618715e-01 -1.52524054e-01 3.40156019e-01
4.31720316e-01 2.77994365e-01 2.03942090e-01 8.47837031e-01
3.05521280e-01 -4.12987322e-01 -1.21662843e+00 9.41964269e-01
-1.91147327e-01 -9.18384433e-01 -3.60051304e-01 -1.64179683e-01
5.38603604e-01 -2.02342942e-01 1.62917167e-01 -6.90105464e-03
-2.68002510e-01 -8.60211849e-01 3.49497527e-01 1.02829278e+00
7.63986051e-01 -1.04454733e-01 9.68035996e-01 5.79610884e-01
-8.44861329e-01 -6.98716715e-02 -2.57786334e-01 5.80680430e-01
7.27505311e-02 3.58136922e-01 -9.47143853e-01 7.80808210e-01
6.16523981e-01 3.41745228e-01 -1.47495985e-01 1.09227180e+00
-3.87513228e-02 2.66107589e-01 -3.02848399e-01 -1.26735181e-01
1.90373138e-02 -9.43543091e-02 6.81902945e-01 8.45908105e-01
6.06129229e-01 2.96701163e-01 1.00422166e-01 9.43331718e-01
4.36845660e-01 3.43340188e-01 -2.52275139e-01 2.09252194e-01
3.07509542e-01 1.53139734e+00 -8.34858298e-01 3.94355357e-02
-2.70218104e-02 1.11569679e+00 -1.34626865e-01 4.07400399e-01
-8.33263516e-01 -2.55798697e-01 -2.10195199e-01 -1.12258404e-01
-1.44068897e-02 5.17650545e-02 -5.19064367e-02 -9.69126582e-01
-8.43088776e-02 -4.92768914e-01 2.16134965e-01 -9.38885808e-01
-1.32747948e+00 6.77983046e-01 -4.64439131e-02 -1.47138000e+00
-4.63611893e-02 -7.18541861e-01 -2.26519316e-01 1.06897342e+00
-1.26390266e+00 -9.51217890e-01 -5.39831042e-01 5.61087072e-01
7.19148219e-02 -1.32198676e-01 1.03017485e+00 2.54627943e-01
-1.19918294e-01 5.96916139e-01 9.18157995e-02 9.10069197e-02
7.92249441e-01 -1.05758786e+00 -6.53062582e-01 2.44697034e-01
-2.86974162e-01 7.19867468e-01 7.71851778e-01 -5.99047124e-01
-1.30094063e+00 -5.38153231e-01 1.47145510e-01 -4.87023741e-02
5.92734218e-01 1.15169846e-01 -7.16920316e-01 2.78702796e-01
-2.01480463e-01 5.08478701e-01 4.66813564e-01 -1.95063531e-01
2.96795398e-01 -1.57784536e-01 -1.45960319e+00 2.79288322e-01
6.88946187e-01 -2.54254192e-01 -2.63417840e-01 3.83047521e-01
3.47862840e-01 -8.72637510e-01 -1.35942221e+00 6.75383747e-01
8.39305341e-01 -6.77027524e-01 8.56076896e-01 -3.29706490e-01
1.37098014e-01 -1.63505465e-01 8.31719935e-02 -1.28196120e+00
-2.30068371e-01 -4.38441575e-01 3.59453142e-01 5.91937661e-01
2.57422179e-01 -6.41414046e-01 7.95923054e-01 6.51672363e-01
-1.25306770e-01 -7.20017374e-01 -1.17919588e+00 -4.88001972e-01
-2.59654254e-01 -9.78723392e-02 -1.05620436e-01 6.55590713e-01
1.36690095e-01 6.20757528e-02 -2.57349849e-01 7.08697319e-01
8.47726882e-01 3.32874469e-02 3.30375403e-01 -1.30803156e+00
-6.16544068e-01 -3.29622626e-02 -7.46192098e-01 -9.57315683e-01
1.26652345e-01 -8.37306917e-01 4.84709054e-01 -1.13412619e+00
2.21135333e-01 -7.36327112e-01 -7.09060073e-01 -1.82050258e-01
-9.78736058e-02 -1.23605408e-01 -3.06431890e-01 1.62483454e-01
-4.43368077e-01 4.01288956e-01 1.81051850e+00 7.84691572e-02
-2.18950167e-01 4.25273448e-01 -1.30874380e-01 5.86130917e-01
4.69148576e-01 -4.17745978e-01 -4.43382412e-01 -3.79961655e-02
-7.40157068e-02 9.89916265e-01 6.11010313e-01 -1.18752158e+00
3.06036055e-01 -1.28698591e-02 4.41655964e-01 -2.61589587e-01
4.06584591e-01 -1.03241205e+00 2.30127975e-01 8.64359617e-01
-6.33431733e-01 -5.79721689e-01 6.78573875e-03 8.86865795e-01
-1.69808596e-01 -3.70372027e-01 9.06533241e-01 -2.55220622e-01
-4.29600298e-01 3.02436084e-01 -2.26608276e-01 -3.73922467e-01
1.09662640e+00 -7.97040462e-02 3.20628434e-01 -3.91039997e-01
-1.37067032e+00 1.58798069e-01 -1.44170970e-01 -5.72923943e-02
5.85801363e-01 -9.60392296e-01 -6.53330922e-01 1.61983654e-01
2.22589433e-01 -1.62296057e-01 1.06072736e+00 1.34278190e+00
-5.05718112e-01 2.15642288e-01 -2.51502752e-01 -1.10890150e+00
-8.45221043e-01 3.87893349e-01 6.70580149e-01 -3.24049503e-01
-2.74329245e-01 9.93149638e-01 3.81315827e-01 -5.96559346e-01
3.16072851e-01 -3.43542904e-01 -3.14623207e-01 -4.45716172e-01
-7.77182356e-02 1.48945868e-01 -1.14378691e-01 -5.34697175e-01
-3.33886355e-01 7.26359069e-01 2.44261742e-01 -7.61404559e-02
9.83898282e-01 -3.26136589e-01 1.50415972e-01 1.94257632e-01
1.04792440e+00 -1.28158972e-01 -1.26768005e+00 -3.14715087e-01
-3.20050493e-02 -2.68811226e-01 -4.95483018e-02 -1.11841881e+00
-8.70912969e-01 9.26657438e-01 1.18435478e+00 -2.44841710e-01
1.00793862e+00 3.22611723e-03 2.80228257e-01 3.30995351e-01
4.94784713e-01 -8.30116689e-01 1.95007294e-01 -2.56890487e-02
9.80567694e-01 -1.51106858e+00 -1.39472410e-01 -6.24950171e-01
-9.33546484e-01 1.14519203e+00 6.23855174e-01 -6.41707852e-02
9.10925984e-01 1.44961193e-01 2.64556050e-01 -1.69268236e-01
-2.56286934e-02 2.31458932e-01 5.55513084e-01 5.92463255e-01
2.88746417e-01 -2.80468259e-03 -2.65631825e-01 7.72807598e-01
3.11442018e-01 2.93494165e-01 4.04300503e-02 7.44087100e-01
-3.56339604e-01 -5.90673327e-01 -2.77425408e-01 1.39634147e-01
-3.26011062e-01 2.93673694e-01 6.49093747e-01 5.61787724e-01
-2.36281008e-03 5.07727921e-01 -4.49701428e-01 -1.94469959e-01
2.35209569e-01 -2.69424051e-01 7.70213544e-01 -4.10108149e-01
-3.98011416e-01 3.51598293e-01 -3.84415418e-01 -6.56057894e-01
-4.60295081e-01 -5.20858884e-01 -1.27240586e+00 5.05450487e-01
-5.12164950e-01 2.03925788e-01 8.63969624e-01 8.76955211e-01
2.09746629e-01 5.68534613e-01 7.83420861e-01 -9.19608831e-01
-1.11792624e+00 -1.05900168e+00 -7.88574874e-01 4.26993996e-01
-8.95868167e-02 -8.04038227e-01 -3.07183534e-01 -6.45151362e-02] | [14.423158645629883, -2.133993625640869] |
599aa5ed-9efb-4585-a7b2-c0dd1072f507 | general-multi-label-image-classification-with | 2011.14027 | null | https://arxiv.org/abs/2011.14027v1 | https://arxiv.org/pdf/2011.14027v1.pdf | General Multi-label Image Classification with Transformers | Multi-label image classification is the task of predicting a set of labels corresponding to objects, attributes or other entities present in an image. In this work we propose the Classification Transformer (C-Tran), a general framework for multi-label image classification that leverages Transformers to exploit the complex dependencies among visual features and labels. Our approach consists of a Transformer encoder trained to predict a set of target labels given an input set of masked labels, and visual features from a convolutional neural network. A key ingredient of our method is a label mask training objective that uses a ternary encoding scheme to represent the state of the labels as positive, negative, or unknown during training. Our model shows state-of-the-art performance on challenging datasets such as COCO and Visual Genome. Moreover, because our model explicitly represents the uncertainty of labels during training, it is more general by allowing us to produce improved results for images with partial or extra label annotations during inference. We demonstrate this additional capability in the COCO, Visual Genome, News500, and CUB image datasets. | ['Yanjun Qi', 'Vicente Ordonez', 'Tianlu Wang', 'Jack Lanchantin'] | 2020-11-27 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Lanchantin_General_Multi-Label_Image_Classification_With_Transformers_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Lanchantin_General_Multi-Label_Image_Classification_With_Transformers_CVPR_2021_paper.pdf | cvpr-2021-1 | ['multi-label-image-classification'] | ['computer-vision'] | [ 6.61323249e-01 2.85332412e-01 -2.74152935e-01 -7.03628540e-01
-8.73960197e-01 -8.03617954e-01 8.03075969e-01 2.64195532e-01
-1.82513550e-01 5.30088961e-01 -2.00605273e-01 -2.52336651e-01
1.59225643e-01 -6.32127345e-01 -1.00058877e+00 -6.59139454e-01
1.31733015e-01 5.36962867e-01 1.71575770e-01 2.86115944e-01
-4.50163186e-02 1.99771181e-01 -1.75151217e+00 9.89223778e-01
1.74281120e-01 1.57345927e+00 -1.85268864e-01 5.91681302e-01
1.81991369e-01 1.38980460e+00 -3.00062239e-01 -7.20887363e-01
1.37368426e-01 -1.44009262e-01 -9.18715417e-01 4.21480119e-01
7.57462323e-01 1.27934630e-03 9.41422954e-02 1.00770092e+00
1.04736827e-01 -4.23875332e-01 1.06560028e+00 -1.58899963e+00
-7.18555331e-01 4.36072201e-01 -5.40184140e-01 -2.78080672e-01
7.20069185e-03 2.78585833e-02 1.36045742e+00 -9.10506427e-01
6.99580729e-01 1.16215110e+00 7.73848534e-01 4.81627852e-01
-1.46611619e+00 -5.63557506e-01 1.70739904e-01 1.45654246e-01
-1.31941140e+00 -2.94368118e-01 2.79741347e-01 -8.89814794e-01
5.29037833e-01 1.92525432e-01 2.02807054e-01 1.06709015e+00
1.28942251e-03 8.46020460e-01 1.38102615e+00 -5.22246718e-01
-2.60841977e-02 5.03589988e-01 2.02585518e-01 1.23538280e+00
-2.73594767e-01 5.93876727e-02 -5.19555986e-01 -1.45771354e-01
8.80404562e-02 -9.35450569e-02 2.54273620e-02 -5.97619951e-01
-1.18272519e+00 8.23833466e-01 4.00279880e-01 -1.91587418e-01
-1.10829905e-01 4.01300430e-01 5.18476546e-01 1.16509072e-01
6.05378866e-01 3.30053270e-01 -7.15027869e-01 4.62102830e-01
-6.81944966e-01 -5.68364374e-02 4.51884598e-01 8.17611396e-01
9.81963038e-01 -5.47622383e-01 -6.01959825e-01 8.54719579e-01
3.87137055e-01 1.78045705e-01 1.55523181e-01 -9.87046719e-01
1.88781366e-01 8.24435830e-01 1.92160904e-02 -5.36109507e-01
-5.18433452e-01 -3.97881240e-01 -5.38847029e-01 4.56489801e-01
3.65202993e-01 2.05802724e-01 -1.38596785e+00 1.87398136e+00
3.13471049e-01 1.65504307e-01 2.08297130e-02 5.19741178e-01
9.89309847e-01 3.27995628e-01 2.61309803e-01 1.92731962e-01
1.49715030e+00 -1.19569838e+00 -4.61223125e-01 -3.21138829e-01
7.86739469e-01 -3.63017261e-01 7.70945787e-01 2.85138130e-01
-7.41721869e-01 -4.06794161e-01 -9.96376336e-01 -2.45225430e-01
-7.35206544e-01 6.10565603e-01 5.17039716e-01 5.68284094e-01
-1.07218111e+00 4.75492239e-01 -4.61504310e-01 -1.20785207e-01
7.32777834e-01 4.02521014e-01 -6.03703439e-01 -2.32374668e-01
-8.41864347e-01 9.42530215e-01 4.23135400e-01 -2.23368540e-01
-1.18857050e+00 -6.80028200e-01 -1.04623330e+00 1.43266812e-01
3.83315176e-01 -3.56675267e-01 1.26337671e+00 -1.09339464e+00
-9.93426263e-01 1.57337284e+00 -8.18518773e-02 -3.76404822e-01
5.20576775e-01 3.30274314e-01 -2.34594315e-01 1.59449786e-01
3.85668427e-01 1.20057642e+00 8.40882778e-01 -1.38110209e+00
-9.19761360e-01 -2.61526406e-01 2.58735657e-01 -1.77104577e-01
-7.08714724e-02 1.77475587e-02 -3.93808395e-01 -3.91585588e-01
-5.39440550e-02 -1.10202646e+00 1.07730165e-01 3.16308647e-01
-7.36247718e-01 -4.02721733e-01 6.21852815e-01 -2.94600427e-01
6.82036638e-01 -2.29354906e+00 2.25266248e-01 1.44840613e-01
3.56938988e-01 -2.06939906e-01 -2.64768332e-01 1.27663702e-01
-3.07071894e-01 3.12346905e-01 -8.41066241e-02 -7.04935849e-01
1.28275931e-01 3.12594712e-01 -1.36912569e-01 4.83339489e-01
7.09763050e-01 8.83137047e-01 -6.30393207e-01 -6.56722248e-01
3.08612864e-02 4.63598400e-01 -3.17815632e-01 1.66030094e-01
-4.96356279e-01 4.57146287e-01 -9.47632566e-02 9.81490433e-01
3.79601479e-01 -1.03051877e+00 2.64370382e-01 -4.11871582e-01
6.81723729e-02 8.97606835e-02 -9.71532285e-01 1.33040881e+00
-3.37326556e-01 4.87755477e-01 -5.28418496e-02 -9.45371270e-01
5.03876328e-01 3.72400045e-01 4.58806425e-01 -4.72402513e-01
2.29914486e-01 5.55469319e-02 -4.92630869e-01 -3.43335062e-01
1.28664091e-01 -6.40676245e-02 -1.88386247e-01 4.80117410e-01
4.14939553e-01 2.03964800e-01 4.27793086e-01 1.97078422e-01
1.03491175e+00 1.81139857e-01 2.94337064e-01 -1.59470551e-02
4.17917460e-01 -1.67654067e-01 6.35528684e-01 6.73233151e-01
-3.45454402e-02 5.80925167e-01 9.87610102e-01 -4.54847664e-01
-9.73887861e-01 -7.94979632e-01 -3.41335446e-01 1.41270685e+00
-1.49744019e-01 -3.99282575e-01 -2.45993748e-01 -1.18993104e+00
1.55310795e-01 3.23409855e-01 -1.16870129e+00 5.96047491e-02
-1.83362395e-01 -7.32913017e-01 4.90084082e-01 3.96857798e-01
1.29906401e-01 -8.78782094e-01 -4.50512528e-01 -1.39371261e-01
-1.27693534e-01 -1.33251131e+00 -2.99328685e-01 7.43769407e-01
-1.84007719e-01 -1.46287513e+00 -3.45675051e-01 -8.83565843e-01
8.30946505e-01 -2.21832275e-01 1.19467354e+00 2.91245822e-02
-2.48447895e-01 2.04071149e-01 -2.56692708e-01 -2.28505775e-01
-5.82207501e-01 -1.27199024e-01 -4.65398580e-01 5.18205583e-01
-1.45039260e-01 -1.77284062e-01 -2.43846476e-01 3.75003546e-01
-9.13585961e-01 4.69365001e-01 4.94518548e-01 1.14481783e+00
7.91756690e-01 -2.31531456e-01 4.59435433e-01 -1.38548052e+00
-5.40756844e-02 -5.69368482e-01 -6.39698803e-01 8.17153156e-01
-7.24040449e-01 2.68061042e-01 4.09100413e-01 -4.42149431e-01
-7.06050158e-01 6.33778691e-01 9.38343704e-02 -4.62066233e-01
-1.20055988e-01 3.86981189e-01 -1.12705648e-01 -1.87815174e-01
4.14198756e-01 -1.18690230e-01 -1.74709931e-01 -3.70229393e-01
6.96300387e-01 6.30934179e-01 4.27912116e-01 -3.80192041e-01
2.74149925e-01 5.75563312e-01 3.62378985e-01 1.34267369e-02
-1.50803769e+00 -4.24648851e-01 -7.32147336e-01 -2.94516325e-01
1.02839541e+00 -1.03350365e+00 -8.79651308e-01 5.02323806e-01
-9.93314087e-01 -3.56051058e-01 -1.14506863e-01 -4.13286849e-04
-4.11597729e-01 6.07097638e-04 -7.34117091e-01 -4.74341959e-01
-1.48395598e-01 -1.52853727e+00 1.46249390e+00 3.50868478e-02
1.34799898e-01 -8.60337794e-01 -2.50412017e-01 5.63651621e-01
7.97816738e-02 4.92810965e-01 1.23839903e+00 -7.44029820e-01
-6.94557846e-01 -1.46916643e-01 -6.33842230e-01 4.30965841e-01
-2.79329363e-02 4.38486375e-02 -1.29454339e+00 -3.20671767e-01
-4.98349905e-01 -1.05616391e+00 1.16707551e+00 -3.45605575e-02
1.38157809e+00 -3.12767744e-01 -6.40000880e-01 6.03628874e-01
1.48244059e+00 -9.84638557e-02 2.00922474e-01 1.37878433e-01
7.16553152e-01 6.69218600e-01 5.64136803e-01 2.77288198e-01
6.66551769e-01 7.76383638e-01 8.32968354e-01 -1.73489407e-01
-2.00788453e-01 -4.64025587e-02 3.62079479e-02 1.74441487e-01
5.64246774e-01 -4.84823734e-01 -8.26025784e-01 5.29428065e-01
-1.78668785e+00 -7.48163998e-01 -1.84093658e-02 1.86998844e+00
9.23962533e-01 3.54522392e-02 -7.14572668e-02 -2.30575856e-02
7.93510377e-01 4.64045033e-02 -7.28408158e-01 -6.34709597e-02
-2.72909492e-01 1.40423208e-01 5.63179433e-01 2.54035026e-01
-1.59646845e+00 7.30086803e-01 6.55333853e+00 5.71581066e-01
-1.19755197e+00 2.85282105e-01 1.08568692e+00 7.06762867e-03
-3.28715742e-02 -7.19180107e-02 -9.61367667e-01 2.53921807e-01
8.03621411e-01 3.78371358e-01 2.77105629e-01 6.93713963e-01
-5.35758495e-01 -3.57213151e-03 -1.43976593e+00 7.08247781e-01
2.87822455e-01 -1.43267941e+00 -1.10375613e-01 6.28650859e-02
7.72419155e-01 1.60428524e-01 3.76403630e-01 2.58374035e-01
6.01115644e-01 -1.07020032e+00 1.14803588e+00 3.58636945e-01
1.38748801e+00 -3.09276611e-01 4.68424916e-01 1.38991207e-01
-9.17593896e-01 -3.34176928e-01 7.38111436e-02 2.52196014e-01
-6.77773282e-02 4.86778855e-01 -1.05749571e+00 3.22159410e-01
5.17501175e-01 9.87313569e-01 -1.11139953e+00 7.03395545e-01
-3.33553940e-01 4.45541501e-01 -1.41065717e-01 3.76868665e-01
1.73855945e-01 3.50301117e-01 -3.05727780e-01 1.12794793e+00
4.71231267e-02 -3.94116193e-01 4.68056619e-01 6.30666375e-01
-4.20984983e-01 -9.39734876e-02 -4.22005057e-01 -7.04218224e-02
2.36549169e-01 1.62565768e+00 -7.82171845e-01 -4.03426766e-01
-6.49664819e-01 8.95227432e-01 8.31291080e-01 1.62750483e-01
-9.79473412e-01 1.01997748e-01 3.16847265e-01 -2.36388251e-01
5.66766024e-01 4.80984569e-01 -2.39709422e-01 -1.03340101e+00
-1.41800210e-01 -6.97206616e-01 6.67891502e-01 -1.01590931e+00
-1.29486656e+00 4.88671243e-01 -3.67054671e-01 -1.07125330e+00
-2.04648733e-01 -1.03982961e+00 -7.08938316e-02 5.16601086e-01
-1.61217570e+00 -1.78372395e+00 -2.08004624e-01 2.91479737e-01
1.56038985e-01 -1.53406039e-02 1.01056123e+00 4.56832021e-01
-4.61523682e-01 5.53528130e-01 1.43269777e-01 2.11248085e-01
9.35894191e-01 -1.41261387e+00 1.73906293e-02 4.66479242e-01
3.64907205e-01 -3.55642475e-02 3.16933155e-01 -2.95767844e-01
-6.85069144e-01 -1.43786025e+00 8.93346488e-01 -7.70806670e-01
5.79900026e-01 -5.59305370e-01 -5.18978834e-01 1.14544773e+00
1.51459426e-01 5.93993306e-01 8.85139585e-01 6.92291651e-03
-1.01118016e+00 3.43047343e-02 -1.15173900e+00 1.14417240e-01
8.65169346e-01 -7.01900125e-01 2.86632869e-02 7.15948701e-01
7.14424789e-01 -5.01301348e-01 -9.51810896e-01 5.83356321e-01
7.23797202e-01 -8.32670331e-01 9.42053139e-01 -7.27438509e-01
6.34920657e-01 -3.63764852e-01 -3.77868414e-01 -1.24962556e+00
-3.77540708e-01 2.11816445e-01 1.48607805e-01 1.29116750e+00
7.89654016e-01 -3.55669826e-01 5.42946815e-01 3.96752805e-01
-6.62483126e-02 -9.52502191e-01 -8.80354166e-01 -2.61915565e-01
-1.55573666e-01 -2.41838947e-01 4.83172417e-01 1.10159039e+00
-2.57535607e-01 5.55359960e-01 -5.36700070e-01 2.67365277e-01
6.42293513e-01 2.24865139e-01 3.41368407e-01 -1.35347378e+00
-3.40283900e-01 -1.53958246e-01 -4.31185305e-01 -5.93397260e-01
5.40289998e-01 -1.31927896e+00 2.85023823e-02 -1.41243529e+00
5.85804164e-01 -7.50270128e-01 -6.58286393e-01 1.21875930e+00
-8.67988691e-02 7.82780766e-01 2.10792631e-01 1.66271836e-01
-1.08168232e+00 1.68316185e-01 1.04570162e+00 -4.89419699e-01
5.42580009e-01 -9.07616764e-02 -6.21843338e-01 5.57691455e-01
4.69140708e-01 -7.05197453e-01 -1.64374352e-01 -5.35747349e-01
5.67011952e-01 6.78788424e-02 5.85118830e-01 -6.63640976e-01
6.99573606e-02 6.92112148e-02 3.67768347e-01 -5.41087627e-01
5.72668076e-01 -8.48434091e-01 3.18576574e-01 2.95016557e-01
-8.18201959e-01 -1.42822647e-02 3.70048685e-03 5.64153671e-01
-8.92206505e-02 -2.02124491e-01 9.07087147e-01 -1.39965713e-01
-7.16374338e-01 2.13102371e-01 -1.01893932e-01 1.95631199e-02
1.12568903e+00 2.53346950e-01 -7.78746009e-01 -1.64250862e-02
-1.00975025e+00 4.99888808e-01 4.64960605e-01 3.99442583e-01
2.91111797e-01 -1.39850056e+00 -6.09239459e-01 2.70699769e-01
7.58983135e-01 -2.13096812e-01 -1.11890282e-03 6.82503581e-01
-4.86356169e-02 2.74423957e-01 -9.24487412e-02 -8.40717316e-01
-1.53152478e+00 6.93645656e-01 4.91476774e-01 -5.00693798e-01
-1.43835455e-01 1.04482234e+00 4.23672885e-01 -6.53952241e-01
2.69335359e-01 -2.25056291e-01 -2.65296906e-01 2.70123482e-01
3.52746725e-01 -7.01063275e-02 3.14310454e-02 -8.48018765e-01
-3.27947199e-01 5.03934920e-01 -1.59626782e-01 9.61404014e-03
1.14526963e+00 1.33213447e-02 -3.05723548e-01 6.41958058e-01
1.41569912e+00 -2.59141922e-01 -1.29360068e+00 -3.58192742e-01
3.23947221e-02 -1.75391689e-01 -1.08609855e-01 -1.30076492e+00
-1.11795628e+00 7.60415792e-01 8.12364399e-01 1.09912261e-01
9.93433833e-01 4.82304543e-01 1.46592364e-01 2.29230896e-01
2.19102249e-01 -7.33149707e-01 3.05769145e-01 2.38240600e-01
5.07380486e-01 -1.65707016e+00 -2.25335836e-01 -5.22465110e-01
-6.56467080e-01 8.16993594e-01 5.11392057e-01 4.63191867e-01
5.98368227e-01 2.55763173e-01 2.37715110e-01 -4.04639989e-01
-1.07006335e+00 -3.44690740e-01 3.18840414e-01 3.04942161e-01
3.40647817e-01 5.91090582e-02 9.33790952e-02 2.56882668e-01
3.84048313e-01 3.70521806e-02 4.26251262e-01 6.80075586e-01
-2.98426569e-01 -1.09426320e+00 -8.81277844e-02 6.45611823e-01
-8.44417930e-01 -1.04523227e-01 -3.16151351e-01 4.21985686e-01
6.59224093e-01 8.50830793e-01 4.57228348e-02 -3.78605932e-01
1.24710472e-02 3.71863484e-01 4.81338203e-01 -8.24121773e-01
-5.63185632e-01 -2.81239510e-01 2.46418089e-01 -4.79380727e-01
-6.28873169e-01 -6.31304622e-01 -1.05435419e+00 2.49802858e-01
-4.01546150e-01 -4.69159484e-02 7.89233804e-01 9.55792665e-01
3.83116752e-01 5.95479965e-01 5.62992990e-01 -4.05267924e-01
-3.08269083e-01 -8.74073565e-01 -5.35633028e-01 7.42066264e-01
4.96271849e-01 -9.33541477e-01 -3.26759398e-01 2.87822843e-01] | [9.643339157104492, 3.996398687362671] |
b485ea6c-9224-46c6-8fb9-9004b5e70ff4 | examining-the-mapping-functions-of-denoising | 1904.06157 | null | https://arxiv.org/abs/1904.06157v2 | https://arxiv.org/pdf/1904.06157v2.pdf | Examining the Mapping Functions of Denoising Autoencoders in Singing Voice Separation | The goal of this work is to investigate what singing voice separation approaches based on neural networks learn from the data. We examine the mapping functions of neural networks based on the denoising autoencoder (DAE) model that are conditioned on the mixture magnitude spectra. To approximate the mapping functions, we propose an algorithm inspired by the knowledge distillation, denoted the neural couplings algorithm (NCA). The NCA yields a matrix that expresses the mapping of the mixture to the target source magnitude information. Using the NCA, we examine the mapping functions of three fundamental DAE-based models in music source separation; one with single-layer encoder and decoder, one with multi-layer encoder and single-layer decoder, and one using skip-filtering connections (SF) with a single-layer encoding and decoding. We first train these models with realistic data to estimate the singing voice magnitude spectra from the corresponding mixture. We then use the optimized models and test spectral data as input to the NCA. Our experimental findings show that approaches based on the DAE model learn scalar filtering operators, exhibiting a predominant diagonal structure in their corresponding mapping functions, limiting the exploitation of inter-frequency structure of music data. In contrast, skip-filtering connections are shown to assist the DAE model in learning filtering operators that exploit richer inter-frequency structures. | ['Gerald Schuller', 'Estefanía Cano', 'Konstantinos Drossos', 'Stylianos Ioannis Mimilakis'] | 2019-04-12 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 3.19456697e-01 2.83445138e-02 2.08874449e-01 -5.48605658e-02
-4.84208047e-01 -4.87028658e-01 2.64796913e-01 -4.68395859e-01
-2.28983954e-01 3.37994605e-01 4.26187068e-01 -1.85687672e-02
-4.57745641e-01 -5.67867458e-01 -7.91340828e-01 -7.66485155e-01
-1.46741122e-01 2.66972184e-02 -2.25737140e-01 -2.99164444e-01
-1.95955992e-01 2.70397604e-01 -1.83781338e+00 6.65324569e-01
8.35519433e-01 1.02254856e+00 2.11881489e-01 1.11785281e+00
-3.43007818e-02 7.90840983e-01 -7.11982191e-01 -2.15559617e-01
4.40567464e-01 -9.46883619e-01 -5.66497922e-01 -1.77555680e-01
3.62335533e-01 -2.24022314e-01 -4.40253317e-01 1.23333716e+00
5.82845449e-01 3.50301981e-01 7.33818352e-01 -9.53309119e-01
-6.75519347e-01 1.26183832e+00 1.60421804e-01 4.00650389e-02
1.58593521e-01 2.71758661e-02 1.18086350e+00 -9.73993242e-01
2.46063679e-01 1.26751113e+00 1.05035722e+00 6.42103851e-01
-1.35901868e+00 -5.51736355e-01 -4.32631969e-01 4.44890529e-01
-1.34168100e+00 -6.78735316e-01 1.12605393e+00 -3.94439101e-01
9.85355616e-01 3.96164745e-01 6.95663631e-01 9.52215672e-01
-1.59252062e-01 7.59106636e-01 9.34984922e-01 -7.42882729e-01
1.77673012e-01 1.23236775e-01 -3.65336090e-02 5.35425901e-01
-2.93004453e-01 5.28227746e-01 -1.03695691e+00 -8.93097147e-02
7.99004734e-01 -6.23726189e-01 -7.27705419e-01 -4.83660065e-02
-1.09651196e+00 8.25416088e-01 2.94302076e-01 5.75153828e-01
-6.47683203e-01 1.74598679e-01 2.69359857e-01 7.31532931e-01
2.20806271e-01 6.10675156e-01 -3.35566193e-01 -3.42896916e-02
-1.37250364e+00 1.57703593e-01 1.24739015e+00 4.26110953e-01
5.35788596e-01 8.35476995e-01 -9.24225242e-05 1.05711198e+00
2.28015915e-01 4.80313957e-01 8.59173000e-01 -1.25507605e+00
9.21602771e-02 -7.79827014e-02 -2.41308004e-01 -7.08442628e-01
-2.87711412e-01 -9.34823155e-01 -8.97433817e-01 3.27962518e-01
3.41817349e-01 -3.56865108e-01 -6.70113146e-01 1.97359824e+00
-8.73153806e-02 4.98882204e-01 3.92806768e-01 9.26782668e-01
5.56099772e-01 8.14611614e-01 -4.75293875e-01 -3.99598777e-01
8.47037971e-01 -1.03697908e+00 -1.16461313e+00 1.07388884e-01
1.85815930e-01 -8.85062754e-01 8.85284185e-01 7.10212171e-01
-1.34067500e+00 -1.12852871e+00 -1.32197654e+00 8.47813487e-02
-4.47890550e-01 5.13526857e-01 1.21334016e-01 6.48849010e-01
-1.19898593e+00 1.33545101e+00 -5.83615065e-01 -2.77434811e-02
-1.57486990e-01 4.88117576e-01 -6.14432022e-02 6.58311725e-01
-1.53159773e+00 8.88961971e-01 7.45927930e-01 2.68557906e-01
-9.08482730e-01 -8.54131997e-01 -6.60973191e-01 4.03447986e-01
-4.65616845e-02 -6.76886141e-01 1.34484029e+00 -1.27213895e+00
-2.17391396e+00 1.48165435e-01 1.08204201e-01 -7.59684443e-01
-1.43351465e-01 -4.73091125e-01 -8.30822051e-01 2.22039148e-01
-3.32846642e-01 5.46687007e-01 1.40498781e+00 -1.22002494e+00
-4.34843004e-01 1.46190479e-01 -2.70282835e-01 6.30447939e-02
-5.48255861e-01 -2.76289195e-01 2.88859785e-01 -1.02728832e+00
7.40650445e-02 -5.70315540e-01 2.02809647e-01 -5.01972735e-01
-3.54768068e-01 1.22618243e-01 6.07074738e-01 -9.63826954e-01
1.64606810e+00 -2.35142779e+00 6.56640828e-01 3.92663181e-01
-1.34534985e-01 3.39867890e-01 -5.15986443e-01 6.74454272e-01
-4.05049354e-01 -2.85250038e-01 -4.11850542e-01 -1.71006575e-01
2.74774820e-01 1.47418812e-01 -6.35133147e-01 8.73570815e-02
2.97073126e-01 4.73258734e-01 -7.53989577e-01 6.64263498e-03
9.99976397e-02 8.03616226e-01 -8.40968728e-01 5.89073002e-01
-3.16794589e-02 3.04784536e-01 3.25428545e-01 9.82379913e-02
5.03297806e-01 2.79398024e-01 1.39840901e-01 -8.29356313e-01
-2.43618488e-01 5.66987455e-01 -1.59873605e+00 1.71898043e+00
-6.01972997e-01 6.16378427e-01 6.80132508e-01 -8.99444461e-01
8.26423883e-01 7.63570905e-01 4.30777699e-01 -8.83471146e-02
2.05150828e-01 6.34804845e-01 6.62325919e-01 -3.66288871e-01
2.97925770e-01 -5.42432666e-01 5.54231346e-01 3.02074164e-01
9.92849290e-01 -2.74579108e-01 2.20874324e-01 -2.05773294e-01
7.37165093e-01 2.58433133e-01 9.28433761e-02 -3.58637989e-01
8.49871159e-01 -4.56025183e-01 4.90040630e-01 7.23290801e-01
2.42888525e-01 6.37899041e-01 1.14002131e-01 -9.54967812e-02
-8.72492909e-01 -1.23254669e+00 -9.25246254e-02 1.10338068e+00
-4.95940357e-01 -5.79098821e-01 -1.15255940e+00 6.45381957e-03
-5.70550747e-02 9.94964361e-01 -4.20513928e-01 -4.39669222e-01
-6.18219435e-01 -4.73535031e-01 8.87938738e-01 4.01687503e-01
4.10802841e-01 -1.07461190e+00 -3.62035394e-01 5.23077726e-01
-1.55229583e-01 -7.11096704e-01 -7.15573490e-01 7.89067209e-01
-7.10026443e-01 -6.92662001e-01 -6.82261825e-01 -7.28618860e-01
-1.10757440e-01 -2.28238240e-01 8.46011817e-01 -4.03451115e-01
1.32483333e-01 5.78165233e-01 -2.37211615e-01 -4.32359993e-01
-8.27514648e-01 5.94523624e-02 5.18920660e-01 5.04785478e-01
2.34270632e-01 -1.21765971e+00 -2.95387536e-01 -1.89796556e-02
-1.13230181e+00 -8.26582536e-02 5.56433797e-01 9.61706102e-01
3.81646097e-01 3.06034625e-01 8.03415000e-01 -4.77014780e-01
1.01887715e+00 -3.41868758e-01 -3.37814897e-01 6.07124604e-02
-5.01631200e-01 3.29600841e-01 8.60668778e-01 -7.29102194e-01
-9.42826807e-01 2.13848516e-01 -5.58135450e-01 -9.82198536e-01
-6.50405437e-02 4.11060601e-01 -1.32499039e-01 -3.12493648e-02
8.24790180e-01 2.99410552e-01 -7.00128684e-03 -8.71614158e-01
6.32557154e-01 8.62528503e-01 1.10768211e+00 -4.57642853e-01
8.83157074e-01 1.12885544e-02 -3.21203679e-01 -1.01032662e+00
-8.80020797e-01 -3.82852465e-01 -7.74773061e-01 -2.69543171e-01
9.78382230e-01 -7.13432431e-01 -6.55262351e-01 4.61875856e-01
-1.30914819e+00 -2.31780857e-01 -8.38524818e-01 1.07389510e+00
-8.98852229e-01 9.52330902e-02 -9.51621056e-01 -9.57394063e-01
-4.15538311e-01 -9.11677718e-01 6.89846039e-01 7.64256790e-02
-1.04658857e-01 -9.37093377e-01 3.52271765e-01 -8.74467343e-02
7.39488900e-01 -3.19471866e-01 1.01899374e+00 -6.58662140e-01
-1.10201292e-01 2.13640109e-01 4.84360188e-01 1.13170314e+00
5.75138852e-02 -1.54108375e-01 -1.61889493e+00 -6.70589432e-02
6.92912102e-01 -1.01366341e-01 1.04543090e+00 6.42953932e-01
9.21406507e-01 -4.02354389e-01 3.97024930e-01 8.36916983e-01
1.06656516e+00 2.99443424e-01 4.23672318e-01 -1.63060755e-01
5.85218847e-01 5.76328278e-01 -2.12550163e-01 1.26246840e-01
-2.13556215e-01 3.44144702e-01 1.58361897e-01 2.51718424e-02
-5.41093409e-01 -4.77097869e-01 8.60292792e-01 1.88926482e+00
-3.54537159e-01 6.90977201e-02 -3.36861402e-01 3.31830740e-01
-1.63702285e+00 -1.05643964e+00 4.20446470e-02 2.18236136e+00
1.05198348e+00 -1.05618052e-02 1.60274729e-01 6.11125588e-01
5.57688713e-01 2.66176194e-01 -3.88325214e-01 -5.94110966e-01
-2.53465444e-01 7.39301801e-01 6.85270801e-02 8.23011577e-01
-1.10548735e+00 7.19953179e-01 6.69954443e+00 1.06021428e+00
-1.15922177e+00 1.92574248e-01 -1.57893360e-01 2.51571346e-05
-1.73495024e-01 -2.24982172e-01 -3.97031218e-01 1.55525804e-01
1.57107389e+00 4.14416157e-02 1.22415388e+00 5.73830903e-01
5.51455095e-02 5.26825130e-01 -1.36397433e+00 1.00621247e+00
7.90189654e-02 -1.21187747e+00 1.78383276e-01 -1.85906872e-01
5.56529999e-01 -1.82589442e-01 6.45440817e-02 3.71167749e-01
-8.97480100e-02 -1.04331028e+00 9.79534566e-01 8.21356177e-01
6.56765282e-01 -7.09982693e-01 5.96037388e-01 3.20863008e-01
-1.17686951e+00 -2.79191822e-01 -3.38660568e-01 -1.15315184e-01
8.98623168e-02 5.43567657e-01 -6.22378349e-01 5.08616030e-01
4.58199561e-01 4.91132945e-01 2.11456586e-02 7.62244046e-01
-1.61502957e-01 1.04624426e+00 -2.02907145e-01 2.80426353e-01
9.81359161e-04 -4.97624874e-01 9.24118578e-01 1.36774361e+00
5.52515090e-01 -1.48683563e-01 -2.35018477e-01 1.22847176e+00
-1.60439223e-01 -4.14605588e-02 -3.88863564e-01 -4.15763438e-01
2.26972714e-01 1.11105192e+00 -1.93720907e-01 -1.99440986e-01
-2.69181997e-01 9.85231757e-01 3.59900743e-02 6.25132084e-01
-4.81693059e-01 -4.72545743e-01 6.80039525e-01 -8.30559507e-02
5.16141415e-01 -1.81621134e-01 -1.17916372e-02 -9.62688506e-01
-3.81377935e-01 -1.25123656e+00 7.14145824e-02 -7.86334932e-01
-1.46350420e+00 8.77172053e-01 -1.65940404e-01 -1.31163788e+00
-5.84042251e-01 -9.19278264e-01 -6.82187736e-01 1.06622207e+00
-1.33832264e+00 -7.49460042e-01 3.12628895e-01 6.87804639e-01
3.99563164e-01 -4.23030257e-01 1.20036125e+00 4.43163723e-01
-2.25629255e-01 3.52775931e-01 2.17051879e-01 8.77043530e-02
4.57260847e-01 -1.54372942e+00 -5.10156900e-02 6.21049464e-01
7.75314689e-01 6.83602095e-01 7.00763345e-01 -2.13267609e-01
-1.26947474e+00 -7.45725930e-01 7.37065434e-01 -6.39481619e-02
7.46550679e-01 -3.72352928e-01 -1.16440749e+00 4.07863289e-01
6.46758914e-01 -3.13419551e-01 8.66798818e-01 -3.76033522e-02
-2.68840164e-01 -2.48994842e-01 -6.50620401e-01 2.79317200e-01
7.98747361e-01 -1.14488208e+00 -1.03991926e+00 1.07610375e-01
6.71337605e-01 -2.70110697e-01 -9.27621365e-01 3.02403599e-01
5.26805103e-01 -1.14196301e+00 1.06199419e+00 -6.77955031e-01
4.27054822e-01 -4.73524779e-01 -3.68177295e-01 -1.84388435e+00
-5.85651219e-01 -8.73890340e-01 -5.09218693e-01 1.01844382e+00
5.04825592e-01 -3.11805308e-01 1.72279611e-01 -2.03115806e-01
-3.94791037e-01 -4.90587652e-01 -9.10385370e-01 -9.20116603e-01
2.28906050e-01 -6.39811218e-01 4.74615037e-01 8.30849111e-01
-7.07090646e-02 6.00634396e-01 -5.79439163e-01 3.76700729e-01
4.50221539e-01 1.25053495e-01 2.19173893e-01 -1.15138090e+00
-9.15038526e-01 -5.75652599e-01 -2.70704343e-03 -9.27548766e-01
2.68114060e-01 -1.05262744e+00 2.15003774e-01 -9.77991939e-01
-3.75377059e-01 2.57896423e-01 -6.39505506e-01 1.68994799e-01
6.99463561e-02 1.41648859e-01 4.64728832e-01 6.63079843e-02
1.24248192e-01 6.87693894e-01 9.84876752e-01 -4.77269366e-02
-4.27192301e-01 1.96119562e-01 -1.42955408e-01 1.09544921e+00
6.94381535e-01 -3.78271520e-01 -4.29157168e-01 -3.00735027e-01
1.18955567e-01 2.33859703e-01 3.87086540e-01 -1.57513773e+00
4.58948106e-01 3.38022918e-01 3.39689344e-01 -3.75561625e-01
6.23088956e-01 -8.20247650e-01 2.67556190e-01 5.06214499e-01
-6.21865928e-01 -3.39520425e-01 2.35754311e-01 4.17732865e-01
-7.29583502e-01 -6.46218896e-01 6.08790100e-01 7.38883838e-02
-3.05859417e-01 -2.49895304e-01 -4.79651660e-01 -1.23297088e-01
4.19393554e-02 -4.68796445e-03 2.74719238e-01 -7.35052824e-01
-1.24657142e+00 -5.04787743e-01 -3.52664411e-01 9.96327400e-02
4.76411164e-01 -1.59686375e+00 -8.85804474e-01 6.28904343e-01
-5.67325830e-01 -5.52116454e-01 2.79866993e-01 7.98316419e-01
-1.29978344e-01 3.07214439e-01 -2.84443617e-01 -2.49867529e-01
-8.31614971e-01 5.05571961e-01 8.61188412e-01 -2.30458826e-02
-3.10111135e-01 9.70254838e-01 1.38667583e-01 -6.28686309e-01
3.81811708e-01 -4.74340260e-01 -1.98610634e-01 5.82274087e-02
4.64488208e-01 6.16989255e-01 6.81872070e-02 -7.18260527e-01
-6.24370587e-04 5.59689999e-01 5.75722694e-01 -5.58055580e-01
1.28372169e+00 2.18700305e-01 -2.50824749e-01 7.88478851e-01
1.17683649e+00 4.82737154e-01 -1.03276443e+00 -1.97222501e-01
3.91723402e-02 1.53920648e-03 3.59108895e-01 -8.73982072e-01
-1.07122576e+00 1.19464314e+00 7.59434700e-01 5.45079112e-01
1.59247124e+00 -4.62875515e-01 7.31330037e-01 3.38836908e-01
-1.85868889e-01 -1.24977827e+00 -3.74713540e-02 6.70594394e-01
1.21585953e+00 -6.11683011e-01 -6.40335500e-01 -2.48074844e-01
-4.11137164e-01 1.63079870e+00 1.53825328e-01 -3.18516910e-01
1.02122056e+00 5.77053726e-01 2.31221437e-01 2.84074284e-02
-5.85174799e-01 -4.42437917e-01 7.75401890e-01 5.05240917e-01
4.78563368e-01 4.85203341e-02 -1.54031739e-01 1.03448606e+00
-7.08442569e-01 -5.03601655e-02 2.76544780e-01 1.40297011e-01
-4.93060768e-01 -1.20891201e+00 -6.55212939e-01 2.04648510e-01
-4.00259614e-01 -4.20890898e-01 -4.85288888e-01 4.35071349e-01
5.30183911e-01 1.00305498e+00 -5.27446531e-02 -8.17099512e-01
4.28577632e-01 7.53350377e-01 4.43312109e-01 -4.48764741e-01
-1.02930367e+00 5.77527881e-01 2.54815780e-02 -3.05573791e-01
-6.17400587e-01 -3.09852034e-01 -1.10904944e+00 1.35049179e-01
-4.06444788e-01 4.42348540e-01 6.17357850e-01 7.28358030e-01
1.83592871e-01 1.00389397e+00 6.24087751e-01 -9.69298959e-01
-9.48825836e-01 -1.30442250e+00 -9.50509191e-01 2.01645240e-01
6.03998959e-01 -3.14566821e-01 -5.70180357e-01 4.00110781e-01] | [15.439133644104004, 5.8208794593811035] |
89e20914-78d1-4e08-a203-904912c67607 | on-building-machine-learning-pipelines-for | 2306.07118 | null | https://arxiv.org/abs/2306.07118v1 | https://arxiv.org/pdf/2306.07118v1.pdf | On building machine learning pipelines for Android malware detection: a procedural survey of practices, challenges and opportunities | As the smartphone market leader, Android has been a prominent target for malware attacks. The number of malicious applications (apps) identified for it has increased continually over the past decade, creating an immense challenge for all parties involved. For market holders and researchers, in particular, the large number of samples has made manual malware detection unfeasible, leading to an influx of research that investigate Machine Learning (ML) approaches to automate this process. However, while some of the proposed approaches achieve high performance, rapidly evolving Android malware has made them unable to maintain their accuracy over time. This has created a need in the community to conduct further research, and build more flexible ML pipelines. Doing so, however, is currently hindered by a lack of systematic overview of the existing literature, to learn from and improve upon the existing solutions. Existing survey papers often focus only on parts of the ML process (e.g., data collection or model deployment), while omitting other important stages, such as model evaluation and explanation. In this paper, we address this problem with a review of 42 highly-cited papers, spanning a decade of research (from 2011 to 2021). We introduce a novel procedural taxonomy of the published literature, covering how they have used ML algorithms, what features they have engineered, which dimensionality reduction techniques they have employed, what datasets they have employed for training, and what their evaluation and explanation strategies are. Drawing from this taxonomy, we also identify gaps in knowledge and provide ideas for improvement and future work. | ['Huang Shengqiang', 'Ronnie Salvador Giagone', 'Yang Zhou', 'Anandharaju Durai Raju', 'Ibrahim Abualhaol', 'Masoud Mehrabi Koushki'] | 2023-06-12 | null | null | null | null | ['dimensionality-reduction', 'android-malware-detection'] | ['methodology', 'miscellaneous'] | [ 2.17132911e-01 -2.47283638e-01 -8.54843616e-01 4.48225923e-02
-4.74555910e-01 -9.30673361e-01 6.17469192e-01 -1.49145350e-01
1.10290952e-01 4.71759260e-01 -3.69387902e-02 -9.60067868e-01
-8.68592337e-02 -3.42154384e-01 -3.70515883e-01 -3.23514104e-01
-8.85791052e-03 1.50657058e-01 8.72142091e-02 1.68970376e-02
6.90430880e-01 4.19952899e-01 -1.78451657e+00 1.25970155e-01
8.10290039e-01 9.62894201e-01 -3.74126583e-02 5.78563273e-01
-7.28632286e-02 8.24076355e-01 -8.28158259e-01 -7.58888125e-01
1.41552374e-01 -2.83637941e-01 -5.36218822e-01 4.04255727e-04
3.65796909e-02 -4.09201056e-01 -2.51421239e-02 9.14227784e-01
2.63765633e-01 -4.54545856e-01 4.00942087e-01 -1.81860745e+00
-6.05919003e-01 3.15235883e-01 -5.48408031e-01 3.19843560e-01
3.96055549e-01 2.21676067e-01 6.64460123e-01 -5.80815196e-01
4.68782157e-01 8.68706822e-01 9.54852819e-01 8.08361411e-01
-9.83833432e-01 -8.14751446e-01 4.26441804e-02 2.29466975e-01
-1.16706002e+00 -5.12591183e-01 9.46148157e-01 -6.54722810e-01
1.23193300e+00 5.53459346e-01 9.07375216e-01 1.59563875e+00
2.85659045e-01 6.30258739e-01 1.26990068e+00 -1.15838408e-01
3.29241633e-01 5.99868655e-01 2.91024595e-01 3.46193105e-01
7.85201728e-01 -1.19652869e-02 -1.87117711e-01 -7.44463265e-01
2.02387825e-01 2.14252621e-01 1.12469912e-01 -3.42154801e-01
-4.50389504e-01 1.00446033e+00 -2.43848473e-01 3.75451237e-01
-1.96661308e-01 -2.53480524e-01 5.57130277e-01 -4.30831406e-03
3.13566834e-01 5.68982363e-01 -7.37812281e-01 -1.05887008e+00
-1.15841818e+00 2.46149480e-01 1.07906663e+00 6.11694932e-01
4.51370597e-01 9.29114074e-02 6.56940162e-01 5.60334802e-01
5.63340247e-01 2.45225459e-01 6.88780546e-01 -9.09851193e-01
3.30619007e-01 9.09844041e-01 -2.24654064e-01 -1.44259679e+00
-6.25268295e-02 -1.21082656e-01 -4.23195422e-01 7.81958736e-03
6.75313696e-02 -3.05808961e-01 -5.75404286e-01 1.44429219e+00
5.50718009e-02 2.45924324e-01 -1.32029861e-01 1.06778592e-01
5.54677963e-01 4.80634123e-01 4.56770994e-02 -4.59994525e-01
1.26082051e+00 -7.05434680e-01 -5.63532770e-01 -4.95660275e-01
6.34140313e-01 -8.42915893e-01 1.27093923e+00 6.36855960e-01
-7.84517467e-01 -1.74388543e-01 -1.06275630e+00 2.02722594e-01
-5.88480175e-01 -2.35854045e-01 8.81563067e-01 1.55250657e+00
-8.60795081e-01 3.57673705e-01 -1.19272947e+00 -5.15598357e-01
8.48790824e-01 5.28078198e-01 5.26934862e-02 8.03537220e-02
-8.36508811e-01 7.92245865e-01 -1.71664625e-01 -2.34116793e-01
-7.68552601e-01 -6.27204299e-01 -6.06948078e-01 -3.62943441e-01
5.08180320e-01 -3.16268772e-01 1.20976996e+00 -8.58385563e-01
-1.34743845e+00 5.17989218e-01 -3.20404351e-01 -6.20032966e-01
1.73476160e-01 -1.99785605e-01 -5.62587976e-01 -2.80007958e-01
7.30699534e-03 2.88223416e-01 9.15548623e-01 -1.39717782e+00
-6.16376519e-01 -4.39231783e-01 2.91935474e-01 -1.13204062e-01
-7.64010072e-01 4.90657657e-01 -5.39231539e-01 -5.56524515e-01
-4.67446178e-01 -1.21140122e+00 -1.93058625e-01 -6.72201395e-01
-3.75688195e-01 -7.31925620e-03 1.67510056e+00 -9.21382964e-01
1.94183505e+00 -2.02090096e+00 -2.71581441e-01 -3.63024622e-02
5.13003349e-01 6.06915593e-01 2.77345687e-01 5.01688600e-01
1.27279341e-01 8.78224373e-01 -3.24856341e-01 -4.92122859e-01
-2.75166094e-01 2.27487087e-01 -4.23317939e-01 2.66733855e-01
1.51398152e-01 8.51662457e-01 -6.78538978e-01 -2.59269506e-01
2.37896711e-01 9.57692325e-01 -5.48997641e-01 -1.99649334e-02
-1.71069995e-01 2.61924237e-01 -4.19822037e-01 1.05899727e+00
6.46462858e-01 -3.19722235e-01 1.51333585e-01 7.58345202e-02
-2.83963960e-02 6.11217082e-01 -8.64244163e-01 8.67983758e-01
-4.09640163e-01 8.83518159e-01 6.75561428e-02 -7.01303184e-01
6.04156196e-01 1.51011318e-01 9.14036751e-01 -3.18249553e-01
1.37341678e-01 3.27194422e-01 3.08041126e-01 -6.51870131e-01
4.09290642e-01 4.19551104e-01 4.89604771e-02 6.39896512e-01
-7.18577087e-01 1.01750992e-01 1.71600655e-02 1.04021765e-01
1.38641131e+00 1.40599953e-02 4.46691155e-01 1.78850964e-01
5.17573059e-01 2.15535223e-01 3.33721548e-01 4.38752770e-01
-5.24699509e-01 2.02585623e-01 7.20542610e-01 -1.64870203e-01
-5.86552799e-01 -6.58398986e-01 -7.44509771e-02 7.72436261e-01
-1.15566202e-01 -9.59926486e-01 -1.15437400e+00 -1.08276629e+00
-6.14178218e-02 6.03997886e-01 -6.33888662e-01 -8.27635825e-02
-6.29364908e-01 -9.50184822e-01 6.99285626e-01 3.16863120e-01
5.06634474e-01 -9.25064266e-01 -7.94619679e-01 -1.66079849e-02
-7.66367614e-02 -1.07681811e+00 -1.63015544e-01 -7.98979998e-02
-1.01165104e+00 -1.43430996e+00 -5.22931144e-02 -3.35480571e-01
4.39820617e-01 5.91320574e-01 1.00365400e+00 4.09965068e-01
-7.64445290e-02 6.50280058e-01 -3.46908987e-01 -6.17532790e-01
-6.94526792e-01 2.60038942e-01 8.06587338e-02 -2.80224442e-01
8.29083145e-01 -3.62122774e-01 -2.75004923e-01 4.53475744e-01
-7.78346479e-01 -4.40149486e-01 6.01947367e-01 4.15749431e-01
2.67235249e-01 4.86345112e-01 5.47381759e-01 -1.01215112e+00
1.11025786e+00 -9.61455703e-01 -4.02936786e-01 -1.50598928e-01
-1.07662725e+00 -5.84377706e-01 4.47599053e-01 -6.68990374e-01
-6.32841885e-01 -1.53688058e-01 -3.70722830e-01 -1.75282761e-01
-1.07297242e-01 3.11462909e-01 -3.96689206e-01 -2.76234508e-01
6.36161506e-01 1.23759255e-01 2.97824323e-01 -3.93530935e-01
3.58334966e-02 1.14064431e+00 -1.38315946e-01 -4.92134728e-02
8.28089654e-01 3.44462812e-01 -2.93783098e-01 -1.05497539e+00
-2.10513964e-01 -1.26569107e-01 -2.54166722e-01 5.98502457e-02
5.36056221e-01 -3.98667872e-01 -5.54327846e-01 4.46512163e-01
-8.33807826e-01 -2.16802403e-01 2.39933282e-02 9.49875787e-02
-1.93487689e-01 5.25005698e-01 -4.12185878e-01 -1.00604546e+00
-2.03981280e-01 -1.86401248e+00 7.31966972e-01 1.25958726e-01
-7.57586420e-01 -1.20208836e+00 -1.81295916e-01 7.78515220e-01
7.75607467e-01 1.51550204e-01 8.38285744e-01 -8.78230035e-01
-3.71060938e-01 -5.68167388e-01 8.90380703e-03 5.31525552e-01
5.38079858e-01 5.00141680e-01 -1.07511699e+00 -2.28211418e-01
6.06935799e-01 2.04327554e-01 3.07608545e-01 4.15436924e-01
1.45453620e+00 -7.25199342e-01 -7.67697096e-01 3.79468709e-01
9.43982422e-01 5.82700431e-01 6.49857163e-01 4.91404742e-01
8.65960956e-01 6.08603120e-01 6.09311879e-01 1.66380495e-01
6.98486507e-01 7.58081377e-01 6.02901697e-01 3.04890245e-01
2.73123290e-02 -1.63459435e-01 7.53421903e-01 7.91895866e-01
-6.15992434e-02 1.39926316e-03 -1.05606711e+00 1.98896244e-01
-1.32275569e+00 -9.18842435e-01 -1.10677630e-01 2.24955797e+00
5.89680552e-01 3.61611247e-01 7.06577539e-01 4.99650180e-01
4.60370839e-01 1.43402070e-01 -6.23090565e-01 -6.30320191e-01
1.63137153e-01 -1.35546163e-01 3.29550982e-01 4.10630196e-01
-1.04320276e+00 8.07533741e-01 6.73120928e+00 8.12668979e-01
-1.44078898e+00 1.96287826e-01 7.90374994e-01 9.51407626e-02
-1.98755637e-01 2.31213257e-01 -9.18458700e-01 8.56473446e-01
1.37881684e+00 -1.45455971e-01 7.43105531e-01 1.23340440e+00
2.74497598e-01 -5.19804098e-02 -8.72647583e-01 1.23107815e+00
3.81045043e-01 -1.31507492e+00 -2.55681336e-01 7.37520576e-01
6.43913269e-01 4.62111123e-02 5.44373631e-01 4.30400014e-01
-9.23827291e-02 -1.04749620e+00 1.68175474e-01 -1.73011363e-01
4.23681378e-01 -5.30479372e-01 4.48371798e-01 3.25125366e-01
-1.06206524e+00 -6.97813272e-01 1.13669194e-01 -3.77741724e-01
-6.81721419e-02 4.04687852e-01 -1.07335877e+00 1.38923258e-01
6.39598727e-01 8.76604438e-01 -9.67307925e-01 8.96164536e-01
-3.58467475e-02 1.28172278e+00 -1.15562566e-01 -2.60844886e-01
-7.54360557e-02 1.32270962e-01 4.58989143e-01 1.03690648e+00
2.19368100e-01 -4.60779548e-01 -9.12683532e-02 4.66869593e-01
-4.91992272e-02 -7.49342963e-02 -9.49942231e-01 -6.57031834e-01
8.78740370e-01 1.37801480e+00 -7.61946917e-01 -1.20474830e-01
-6.99195564e-01 5.58817089e-01 -3.06053221e-01 -1.58800576e-02
-9.30403650e-01 2.79789530e-02 1.07104492e+00 5.69497764e-01
-1.31910399e-01 -4.03812379e-01 -8.07253957e-01 -8.18239510e-01
5.57429083e-02 -1.74515128e+00 1.88649282e-01 -2.13811159e-01
-1.07315886e+00 4.64630485e-01 1.77618280e-01 -1.04427075e+00
-6.23226523e-01 -6.23496652e-01 -5.07042944e-01 3.23741466e-01
-1.05864286e+00 -9.17769372e-01 -1.66100383e-01 6.66819364e-02
6.08173788e-01 -2.58546919e-01 7.64987826e-01 4.61822420e-01
-8.06125700e-01 5.75909853e-01 -2.07297713e-01 -2.49582931e-01
2.82937288e-01 -7.26325631e-01 6.80648565e-01 6.47956729e-01
1.57881752e-02 1.21604085e+00 4.94811922e-01 -1.10757136e+00
-1.91785502e+00 -8.94101799e-01 5.81837595e-01 -1.22939777e+00
8.89739096e-01 -6.13530219e-01 -8.13095927e-01 7.62769520e-01
2.81106494e-02 -5.06150663e-01 9.72365916e-01 6.47549145e-03
-2.29104474e-01 -2.98876353e-02 -1.34242320e+00 7.10005105e-01
8.34837496e-01 -4.09641683e-01 -2.26326995e-02 1.21629775e-01
4.74149078e-01 -2.67572761e-01 -6.99713469e-01 4.70513135e-01
7.15256751e-01 -1.15151584e+00 8.72582376e-01 -3.63606781e-01
1.07660197e-01 1.08135212e-03 -1.31728426e-01 -6.21357977e-01
1.82353258e-01 -9.19321239e-01 -8.99218678e-01 1.32430351e+00
5.39575398e-01 -9.02072847e-01 1.11509478e+00 7.72846937e-01
2.72909440e-02 -1.58643603e+00 -6.35341883e-01 -7.76910961e-01
-2.62083076e-02 -9.49838698e-01 7.26379097e-01 9.79510009e-01
-2.17753544e-01 2.18732655e-01 -4.04304057e-01 -1.03690103e-01
3.34798664e-01 -4.99253452e-01 1.11588490e+00 -1.24798071e+00
-3.01704258e-01 -8.07949007e-01 -5.65080225e-01 -8.64773452e-01
-1.27847075e-01 -5.07780850e-01 -6.61131799e-01 -1.40335536e+00
4.35456872e-01 -4.01358604e-01 3.08631688e-01 4.12297398e-01
-1.22195184e-01 3.80547464e-01 3.04530747e-02 4.98754770e-01
-3.29259306e-01 -1.58866316e-01 6.59697473e-01 1.46702737e-01
-5.08495569e-01 3.48941654e-01 -1.39135873e+00 1.04124975e+00
1.05211532e+00 -4.19544965e-01 -7.15861320e-01 -1.53262153e-01
3.04958940e-01 -6.47263169e-01 2.78022707e-01 -1.03838229e+00
-1.39514133e-01 -1.35580689e-01 2.82581776e-01 -4.41918403e-01
4.51542139e-01 -9.70540643e-01 2.94558823e-01 6.47186697e-01
2.87349969e-01 5.53879321e-01 4.10704166e-01 2.06777990e-01
1.20634355e-01 -2.95747757e-01 4.71127808e-01 1.66816562e-01
-4.33382303e-01 2.13406801e-01 -5.35773635e-01 2.81663463e-02
1.40465403e+00 -6.62271798e-01 -3.65323365e-01 -4.42429870e-01
-1.94788367e-01 -2.77420133e-01 9.85887587e-01 8.03829610e-01
4.24448490e-01 -9.36905503e-01 -1.16423694e-02 2.87939429e-01
-1.24752097e-01 -5.41690409e-01 -4.49067028e-03 1.01619232e+00
-2.07510754e-01 4.11482900e-01 1.50611222e-01 -4.20341879e-01
-1.56005621e+00 7.51936495e-01 1.00176118e-01 -3.48675460e-01
-2.60576487e-01 3.71981740e-01 -4.10021007e-01 -2.85802931e-01
3.76400799e-01 -3.57691236e-02 -2.78880328e-01 2.71147620e-02
6.59525752e-01 8.82054031e-01 4.31618374e-03 -9.69313204e-01
-6.65675521e-01 4.80817437e-01 -6.65214360e-02 2.41061524e-01
1.12779057e+00 -5.48774227e-02 -1.76648170e-01 3.66133749e-01
1.26288867e+00 5.98154604e-01 -9.34526205e-01 5.99998772e-01
2.08190963e-01 -6.33740425e-01 -3.13488603e-01 -6.98088825e-01
-9.68750596e-01 7.91471124e-01 6.94526434e-01 9.10797715e-01
1.09024107e+00 -2.20927857e-02 8.00367534e-01 -5.03682159e-02
2.23808587e-01 -7.39176273e-01 1.55708924e-01 3.32439631e-01
5.23700356e-01 -1.28211415e+00 1.20982617e-01 -6.04311943e-01
-5.66151857e-01 7.79624760e-01 5.37811995e-01 2.64009982e-01
9.19140875e-01 5.38334906e-01 -4.11061980e-02 -1.80046380e-01
-4.44491684e-01 2.77683169e-01 3.46905887e-02 1.07585466e+00
5.08014023e-01 3.65453996e-02 -3.30625594e-01 7.94447541e-01
-3.66908848e-01 -7.93802924e-03 4.61043775e-01 1.13575947e+00
-1.94572911e-01 -1.69627702e+00 -4.49142963e-01 1.05695772e+00
-8.46875489e-01 -3.78226563e-02 -8.82413805e-01 1.01098835e+00
3.91345799e-01 1.49932098e+00 -4.26945925e-01 -1.01470649e+00
1.38741255e-01 -1.93516314e-01 1.63796008e-01 -7.61121094e-01
-6.67297423e-01 -1.92460716e-01 1.50682032e-01 -5.06527841e-01
-2.40331888e-02 -1.06346977e+00 -9.57862377e-01 -5.69672406e-01
-4.03054386e-01 -5.84086180e-02 1.10863209e+00 8.61858785e-01
7.82552660e-01 1.93939656e-01 6.47343576e-01 -8.00439835e-01
-3.64000618e-01 -7.53584206e-01 6.98331892e-02 1.09476291e-01
1.00933909e-01 -8.01076591e-01 -4.78876382e-01 -1.02898188e-01] | [14.419124603271484, 9.678455352783203] |
887700cc-bf94-4110-956c-881bbcacf12a | fine-grained-ecg-classification-based-on-deep | 1901.06469 | null | https://arxiv.org/abs/1901.06469v3 | https://arxiv.org/pdf/1901.06469v3.pdf | Deep Time-Frequency Representation and Progressive Decision Fusion for ECG Classification | Early recognition of abnormal rhythms in ECG signals is crucial for monitoring and diagnosing patients' cardiac conditions, increasing the success rate of the treatment. Classifying abnormal rhythms into exact categories is very challenging due to the broad taxonomy of rhythms, noises and lack of large-scale real-world annotated data. Different from previous methods that utilize hand-crafted features or learn features from the original signal domain, we propose a novel ECG classification method by learning deep time-frequency representation and progressive decision fusion at different temporal scales in an end-to-end manner. First, the ECG wave signal is transformed into the time-frequency domain by using the Short-Time Fourier Transform. Next, several scale-specific deep convolutional neural networks are trained on ECG samples of a specific length. Finally, a progressive online decision fusion method is proposed to fuse decisions from the scale-specific models into a more accurate and stable one. Extensive experiments on both synthetic and real-world ECG datasets demonstrate the effectiveness and efficiency of the proposed method. | ['Xiaobin Xu', 'Jing Tian', 'Jing Zhang', 'Yuxiang Yang', 'Yang Cao'] | 2019-01-19 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 3.52025628e-01 -5.42067111e-01 1.46198466e-01 -6.49662137e-01
-8.61560702e-01 -4.97535050e-01 -1.37819290e-01 3.26300532e-01
-2.54577488e-01 7.12474942e-01 8.31906423e-02 -1.02025077e-01
-4.16848212e-01 -4.19490904e-01 -1.56492069e-02 -6.58617973e-01
-5.98557651e-01 6.34628683e-02 -1.13131572e-02 -8.72035790e-03
3.33608352e-02 4.11228895e-01 -8.89556408e-01 2.74470121e-01
9.89337802e-01 1.41376293e+00 -9.65663195e-02 8.67373347e-01
2.60731906e-01 5.61872363e-01 -8.36426675e-01 7.88586512e-02
2.17344612e-01 -1.00032902e+00 -4.77157354e-01 1.27149507e-01
-2.17119128e-01 -5.06688692e-02 -3.04714650e-01 8.99076998e-01
1.09055138e+00 -1.07324317e-01 2.69461632e-01 -7.63690770e-01
-1.62339751e-02 4.84961987e-01 -5.56930482e-01 7.61451125e-01
2.63689250e-01 4.56027947e-02 6.55727804e-01 -6.48668170e-01
2.83338577e-01 6.45148695e-01 1.06707776e+00 4.90726717e-02
-1.09275138e+00 -7.59829402e-01 -1.10378735e-01 3.61570239e-01
-1.42946875e+00 -2.12565929e-01 1.28235412e+00 -3.87376249e-01
5.25806665e-01 1.46698073e-01 9.13833320e-01 7.29017138e-01
4.46675092e-01 3.76523525e-01 1.04080260e+00 -1.69662595e-01
9.98689532e-02 -5.55956721e-01 3.97029473e-03 4.71071839e-01
-5.38843535e-02 3.95063870e-02 -4.68119323e-01 -3.53116870e-01
7.43014872e-01 3.87179941e-01 -5.17374873e-01 -7.95415714e-02
-1.70748842e+00 3.92723620e-01 4.85873073e-01 6.02119207e-01
-6.92778349e-01 -3.90999205e-02 8.27399969e-01 6.85403585e-01
5.15761197e-01 4.45514202e-01 -6.94230437e-01 -3.77527356e-01
-1.12869000e+00 1.46341339e-01 5.45912027e-01 1.63657174e-01
4.16015178e-01 1.45673603e-01 -4.04780746e-01 7.25790977e-01
8.70935991e-02 2.78467238e-01 8.27879071e-01 -6.54245734e-01
2.31224462e-01 4.64874983e-01 1.34543657e-01 -1.16318607e+00
-8.42781425e-01 -1.11512578e+00 -1.32990396e+00 -9.42491218e-02
4.61692214e-01 -3.91708702e-01 -6.85770750e-01 1.55118763e+00
3.08492959e-01 6.93659544e-01 1.41546622e-01 1.00429654e+00
5.73401451e-01 5.52199900e-01 2.13281028e-02 -7.57081985e-01
1.41831338e+00 -3.75945151e-01 -9.40353930e-01 5.16561531e-02
5.57787955e-01 -6.12138927e-01 4.69261467e-01 6.09710932e-01
-8.14010799e-01 -8.12160075e-01 -1.23958528e+00 2.89322883e-01
1.23383306e-01 3.75096768e-01 4.38500196e-01 3.18702042e-01
-5.87297857e-01 8.72762382e-01 -8.87996614e-01 -1.91391520e-02
6.29023552e-01 3.09416413e-01 -1.55385360e-01 2.04354271e-01
-1.47760367e+00 3.98747802e-01 3.86379212e-01 5.32561004e-01
-5.96391439e-01 -5.65975606e-01 -7.74387598e-01 7.47713074e-02
-3.86274084e-02 -5.87743938e-01 1.10150325e+00 -1.12838829e+00
-1.26727819e+00 5.23663282e-01 -1.03840783e-01 -6.26174808e-01
5.83556414e-01 -2.35851333e-01 -7.20218182e-01 2.60861874e-01
3.81305106e-02 -1.25454366e-01 1.19203615e+00 -6.20050728e-01
-7.78110921e-01 -4.81880188e-01 -2.16828406e-01 1.45209014e-01
-2.61962235e-01 -1.54525205e-01 -1.79196745e-02 -1.01389778e+00
5.39482594e-01 -6.51617825e-01 -3.04102331e-01 -1.91564798e-01
-1.51807055e-01 -6.36878610e-02 7.18586564e-01 -8.87161791e-01
1.60470188e+00 -2.44632268e+00 2.38342658e-01 2.96610743e-01
6.52210832e-01 2.25550175e-01 8.02573562e-02 1.72155023e-01
-2.92996913e-01 -2.70181358e-01 -3.20085913e-01 1.40892968e-01
-4.52301025e-01 4.19484032e-03 -2.32090607e-01 6.01847112e-01
1.69987723e-01 6.61120474e-01 -9.71184254e-01 -4.51155156e-01
2.46480510e-01 4.91268963e-01 -5.18607423e-02 1.43023893e-01
3.76732528e-01 9.04520869e-01 -4.99373496e-01 5.05995333e-01
4.47179198e-01 -2.01139674e-01 1.91316515e-01 -5.47313392e-01
1.14726879e-01 1.80551320e-01 -1.12426746e+00 1.90872824e+00
-2.77356595e-01 3.93124789e-01 -3.26710314e-01 -1.33618426e+00
1.00507832e+00 7.51775742e-01 9.85147297e-01 -5.52999735e-01
1.71867892e-01 5.25074780e-01 3.53533298e-01 -5.92637181e-01
-2.46361226e-01 -2.60388106e-01 -1.63801476e-01 2.81521291e-01
-7.85988197e-03 1.67728186e-01 1.70279935e-01 -2.98883975e-01
1.17857885e+00 -4.33265567e-02 5.52698553e-01 -2.02008992e-01
6.12190604e-01 -3.71896982e-01 1.01397288e+00 5.28866887e-01
-4.38091159e-01 7.73812532e-01 4.19883072e-01 -1.08570898e+00
-5.95749736e-01 -8.95079970e-01 -2.90998340e-01 6.05178237e-01
7.09558278e-02 -3.59061360e-01 -4.39665318e-01 -7.30483174e-01
-1.41746163e-01 -8.17816108e-02 -6.47369683e-01 -3.57532203e-01
-8.50240171e-01 -9.61730421e-01 6.67642474e-01 7.20269024e-01
5.06779194e-01 -1.08083463e+00 -1.17533886e+00 7.35916913e-01
-4.90263164e-01 -8.51565957e-01 -5.26064277e-01 1.84995607e-01
-1.12700391e+00 -1.02803540e+00 -9.89803791e-01 -7.83962607e-01
2.65847623e-01 -1.81162536e-01 9.50698376e-01 5.34114353e-02
-6.00807369e-01 -1.32752478e-01 -4.22365248e-01 -4.49987113e-01
-1.24949038e-01 2.34028287e-02 3.53898406e-02 5.46911538e-01
1.68735385e-01 -8.42080891e-01 -1.07965040e+00 5.86066097e-02
-5.47615528e-01 -1.61407694e-01 6.09177709e-01 1.08554626e+00
5.94426215e-01 3.26919526e-01 1.20669878e+00 -6.31598592e-01
8.83101106e-01 -2.10809782e-01 -2.21112877e-01 -8.72677565e-02
-6.10888779e-01 -2.10682899e-01 7.44254231e-01 -5.64103305e-01
-6.03312671e-01 2.05888599e-01 -2.12929457e-01 -3.88735831e-01
-6.57182336e-02 6.44636154e-01 1.94943115e-01 1.83838233e-01
7.18976438e-01 4.06909108e-01 -3.27928886e-02 -1.96747273e-01
-1.39363140e-01 5.86102009e-01 7.54583657e-01 -3.43123257e-01
5.73532581e-01 3.03776085e-01 -8.14990401e-02 -3.76096070e-01
-8.51800144e-01 -4.26851898e-01 -8.08188438e-01 -1.77041456e-01
8.75481963e-01 -8.91833186e-01 -4.73928481e-01 6.76197290e-01
-1.02480459e+00 1.05256759e-01 -4.94994700e-01 7.42166817e-01
-5.16236544e-01 4.79887694e-01 -7.05058396e-01 -6.61006987e-01
-8.17810595e-01 -8.18435848e-01 1.06323957e+00 2.53677070e-01
-4.00193721e-01 -7.01641858e-01 1.72811583e-01 -2.07572669e-01
4.07294333e-01 7.22578585e-01 7.73657799e-01 -6.52712226e-01
3.57407145e-02 -5.37684262e-01 -5.78023531e-02 3.18541825e-01
5.32895327e-01 -3.74872804e-01 -7.66032159e-01 -3.76489967e-01
4.19145107e-01 -8.19518790e-02 6.77515745e-01 5.55371583e-01
1.18667102e+00 1.12321703e-02 -3.33093554e-01 8.13539565e-01
1.21480405e+00 5.72927952e-01 3.80303562e-01 4.97962274e-02
5.51086307e-01 9.09703895e-02 5.89434922e-01 7.67481744e-01
2.32565045e-01 2.81279504e-01 -3.62161510e-02 -3.66786152e-01
1.24939859e-01 2.56445527e-01 -1.24999695e-02 1.07851541e+00
-3.68824810e-01 4.06212449e-01 -8.49226117e-01 6.24412298e-01
-1.92291594e+00 -9.51748252e-01 1.63168117e-01 2.22799158e+00
9.29662585e-01 2.10513875e-01 1.44871414e-01 8.45525861e-01
6.86581731e-01 1.26526356e-02 -8.93068969e-01 1.75736174e-02
-1.68856960e-02 3.21083546e-01 -1.41851395e-01 -7.19557703e-02
-1.33946443e+00 5.77702112e-02 5.86331511e+00 5.36802828e-01
-1.70028615e+00 1.22995526e-01 7.75229514e-01 6.09372221e-02
3.70190918e-01 -2.33422101e-01 -1.50581915e-02 4.08858806e-01
8.68762016e-01 -3.45902532e-01 2.24777371e-01 4.45069641e-01
2.85771251e-01 4.20792520e-01 -1.03730738e+00 1.45232272e+00
-2.05274791e-01 -1.16975355e+00 -5.78509808e-01 -4.20493811e-01
3.46448570e-01 -9.50935259e-02 -2.72342801e-01 2.32542247e-01
-4.27606642e-01 -7.57333219e-01 5.84319472e-01 7.61876047e-01
1.08384633e+00 -8.77687037e-01 9.28533375e-01 1.60912544e-01
-1.66664314e+00 -4.64099169e-01 2.21556961e-03 -1.34812102e-01
3.55299525e-02 8.45890284e-01 -5.67641258e-01 8.25784087e-01
7.81390548e-01 1.30190766e+00 -5.12476921e-01 1.07605135e+00
1.22113444e-01 8.66895556e-01 -2.79030204e-01 4.09745157e-01
-2.21253008e-01 -1.98652013e-03 5.23635924e-01 1.15510631e+00
5.12673914e-01 2.99689710e-01 4.58736897e-01 3.35025489e-01
1.62279353e-01 1.74717858e-01 -3.79387140e-01 -4.19494100e-02
3.09151024e-01 1.38591528e+00 -7.79939950e-01 -4.81038988e-01
-3.15189689e-01 9.18633819e-01 -1.09254286e-01 3.76488298e-01
-7.82718301e-01 -1.00814176e+00 1.77449778e-01 5.80863133e-02
1.73071548e-02 2.48078126e-02 -4.05562967e-01 -1.20611477e+00
2.30804861e-01 -1.05779099e+00 5.81451833e-01 -3.98186862e-01
-1.38090861e+00 1.08733582e+00 -3.98187667e-01 -1.79207039e+00
-4.42634553e-01 -8.69560242e-02 -7.18295336e-01 8.63416910e-01
-1.27403867e+00 -7.27428854e-01 -4.11704332e-01 7.18757272e-01
5.59630573e-01 -8.88784230e-02 1.02248943e+00 5.91697693e-01
-2.86692828e-01 4.83673662e-01 -1.52403459e-01 4.58390445e-01
7.55407453e-01 -1.23919010e+00 1.09944522e-01 7.58638442e-01
-2.40284763e-02 4.61703539e-01 5.61981678e-01 -3.64161581e-01
-9.68369305e-01 -1.08758700e+00 8.02086830e-01 -8.72916821e-03
4.47619289e-01 -5.53653985e-02 -9.68271375e-01 1.52742416e-01
3.02711148e-02 5.47767282e-01 7.04489052e-01 -2.01996714e-02
-4.33373675e-02 -7.87353158e-01 -9.05778468e-01 1.91910088e-01
7.87145615e-01 -6.25589490e-01 -7.00596809e-01 1.55757815e-01
3.45077425e-01 -5.46485841e-01 -1.32071996e+00 8.48316610e-01
9.15048003e-01 -6.93669975e-01 8.07287097e-01 -3.70790243e-01
1.66618645e-01 -4.68926013e-01 1.82954997e-01 -1.43864775e+00
-4.28944647e-01 -8.98867965e-01 -1.42731488e-01 7.40657389e-01
1.39564425e-01 -7.19791532e-01 3.53009671e-01 -1.64173782e-01
-2.03103930e-01 -1.02900243e+00 -9.53647256e-01 -3.37828368e-01
-3.76534194e-01 -3.07769954e-01 5.81839740e-01 1.09264362e+00
8.57491642e-02 3.31770331e-01 -4.83895093e-01 1.20288447e-01
5.79528987e-01 6.32326245e-01 2.45170891e-01 -1.57809770e+00
-3.59452903e-01 -2.40140855e-01 -7.57241607e-01 -4.93388712e-01
-3.09221387e-01 -7.59438694e-01 1.03170224e-01 -1.33707952e+00
1.02052195e-02 -3.49021167e-01 -1.08423626e+00 4.52458888e-01
-4.24647123e-01 4.98566210e-01 -3.55741195e-02 2.66917527e-01
-4.76461291e-01 2.83696026e-01 1.23003018e+00 -6.67197630e-02
-4.43718255e-01 2.37019941e-01 -5.21424353e-01 8.61521184e-01
7.39330769e-01 -5.03449142e-01 -6.07939065e-01 9.21018142e-03
7.09042475e-02 6.88291013e-01 9.64776203e-02 -1.45800090e+00
-7.71939605e-02 2.46008441e-01 8.01789820e-01 -6.00799799e-01
1.18266538e-01 -6.81344688e-01 8.55299905e-02 7.72846162e-01
-4.09635723e-01 2.30139896e-01 7.52761513e-02 6.31294429e-01
-5.94589055e-01 3.63055527e-01 7.85846055e-01 3.23945098e-02
-3.20592374e-01 4.76193339e-01 -2.39328891e-01 1.42938629e-01
9.37177479e-01 -1.98732853e-01 3.11552435e-01 -2.31589288e-01
-1.18053055e+00 7.62580261e-02 -2.20603704e-01 3.58972460e-01
7.25169182e-01 -1.46462464e+00 -9.93689299e-01 4.75032985e-01
2.75639415e-01 2.55111419e-02 5.25458515e-01 1.25418782e+00
-4.83035922e-01 -5.65032773e-02 -3.89836609e-01 -1.04148972e+00
-1.14872015e+00 2.39139512e-01 6.89862669e-01 -4.38770533e-01
-1.01339376e+00 4.73219246e-01 -6.34369031e-02 1.44043729e-01
9.77098197e-02 -6.86677098e-01 -2.83154607e-01 1.12348862e-01
7.20042765e-01 7.79570937e-02 2.09303141e-01 -5.00278652e-01
-4.98845875e-01 7.02308834e-01 6.38542548e-02 1.44063219e-01
1.30894732e+00 -5.96686527e-02 -6.37428612e-02 6.48305476e-01
1.03157365e+00 -2.42471829e-01 -1.10280240e+00 -3.88453573e-01
-1.00097455e-01 -2.43921831e-01 -1.46839410e-01 -7.03958333e-01
-1.26637065e+00 1.05034590e+00 1.18634391e+00 2.60487556e-01
1.82916415e+00 -5.60484231e-01 1.05143702e+00 2.71500230e-01
2.51269549e-01 -9.04114246e-01 4.76153754e-02 4.90323193e-02
7.89648533e-01 -9.03680623e-01 -7.82846510e-02 3.19934301e-02
-5.55430532e-01 1.37316477e+00 8.77963305e-02 -2.59822309e-01
9.80525851e-01 2.20109433e-01 4.89476800e-01 6.71609654e-04
-5.94593346e-01 1.19658157e-01 2.78281033e-01 4.94931281e-01
6.22504115e-01 5.99868782e-02 -5.05381644e-01 9.95170772e-01
-7.60545582e-03 5.52812040e-01 1.88339233e-01 9.09540713e-01
-2.33026862e-01 -9.91748333e-01 -2.68384606e-01 7.46867418e-01
-8.83298755e-01 1.93205282e-01 2.12049067e-01 1.31762177e-01
3.11857015e-01 9.09232020e-01 -1.43459037e-01 -5.45743823e-01
2.58545876e-01 3.45809132e-01 4.33104604e-01 -3.69467795e-01
-6.69243336e-01 5.29375970e-01 -2.40809262e-01 -4.76999342e-01
-3.14532161e-01 -5.59462130e-01 -1.40270030e+00 4.90506262e-01
-1.51480898e-01 2.01645479e-01 3.23366493e-01 1.02204621e+00
5.39662600e-01 1.01737201e+00 9.20475721e-01 -7.17996418e-01
-6.95746899e-01 -1.16030920e+00 -7.19412148e-01 7.93482661e-01
8.88856471e-01 -3.51965815e-01 -5.26982956e-02 4.48448867e-01] | [14.2920560836792, 3.2956340312957764] |
8a4faaf7-3bd9-4605-a175-ab78c902c3c8 | arch-animation-ready-clothed-human | 2108.07845 | null | https://arxiv.org/abs/2108.07845v4 | https://arxiv.org/pdf/2108.07845v4.pdf | ARCH++: Animation-Ready Clothed Human Reconstruction Revisited | We present ARCH++, an image-based method to reconstruct 3D avatars with arbitrary clothing styles. Our reconstructed avatars are animation-ready and highly realistic, in both the visible regions from input views and the unseen regions. While prior work shows great promise of reconstructing animatable clothed humans with various topologies, we observe that there exist fundamental limitations resulting in sub-optimal reconstruction quality. In this paper, we revisit the major steps of image-based avatar reconstruction and address the limitations with ARCH++. First, we introduce an end-to-end point based geometry encoder to better describe the semantics of the underlying 3D human body, in replacement of previous hand-crafted features. Second, in order to address the occupancy ambiguity caused by topological changes of clothed humans in the canonical pose, we propose a co-supervising framework with cross-space consistency to jointly estimate the occupancy in both the posed and canonical spaces. Last, we use image-to-image translation networks to further refine detailed geometry and texture on the reconstructed surface, which improves the fidelity and consistency across arbitrary viewpoints. In the experiments, we demonstrate improvements over the state of the art on both public benchmarks and user studies in reconstruction quality and realism. | ['Tony Tung', 'Stefano Soatto', 'Shunsuke Saito', 'Yuanlu Xu', 'Tong He'] | 2021-08-17 | null | http://openaccess.thecvf.com//content/ICCV2021/html/He_ARCH_Animation-Ready_Clothed_Human_Reconstruction_Revisited_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/He_ARCH_Animation-Ready_Clothed_Human_Reconstruction_Revisited_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-object-reconstruction-from-a-single-image'] | ['computer-vision'] | [ 1.42117947e-01 6.88088834e-02 4.59702492e-01 -1.60396099e-01
-4.04519618e-01 -5.57105541e-01 4.13675845e-01 -5.47369838e-01
6.57365024e-02 4.15727109e-01 3.71743858e-01 3.01165760e-01
3.42980742e-01 -8.04022849e-01 -1.10098457e+00 -3.62607867e-01
2.06313044e-01 7.60021687e-01 3.16145301e-01 -6.06832206e-01
-2.54854739e-01 5.83606720e-01 -1.40697467e+00 2.88059682e-01
4.40021634e-01 9.34442163e-01 -3.09051350e-02 6.82489276e-01
3.38004738e-01 6.09761357e-01 -3.21589470e-01 -6.91560090e-01
6.02136612e-01 -6.13793314e-01 -6.68824971e-01 5.16929865e-01
9.70750093e-01 -6.57415211e-01 -5.99956036e-01 8.67794514e-01
5.06542325e-01 1.26518756e-01 4.45765018e-01 -1.05641723e+00
-8.04831862e-01 1.78886861e-01 -6.73979938e-01 -5.54296792e-01
8.39696288e-01 3.77817124e-01 8.62796605e-01 -1.03890336e+00
1.25524914e+00 1.56856561e+00 9.23660636e-01 6.92844987e-01
-1.44439733e+00 -3.61953706e-01 1.39704332e-01 1.39119290e-02
-1.47059107e+00 -4.80811030e-01 1.08819187e+00 -3.40947390e-01
6.08262181e-01 4.97629344e-01 1.31928134e+00 1.50941133e+00
1.05613247e-01 6.87437654e-01 9.67123270e-01 -2.86870956e-01
-4.50351238e-02 -2.01305553e-01 -6.58821523e-01 1.03093243e+00
-1.19214378e-01 2.37570792e-01 -5.10847092e-01 -1.45478159e-01
1.72216988e+00 -1.24838382e-01 -1.62270516e-01 -1.05065489e+00
-1.52214968e+00 5.36963761e-01 3.82990122e-01 -2.27565438e-01
-4.10891235e-01 4.40488696e-01 3.18845391e-01 3.65997478e-02
4.99204367e-01 2.47114912e-01 -3.00912797e-01 -7.60117471e-02
-6.45420015e-01 5.75853765e-01 6.36431217e-01 1.24206161e+00
2.83911854e-01 2.51413107e-01 5.15023544e-02 7.38142729e-01
2.35277399e-01 6.80869162e-01 -4.16977644e-01 -1.55656219e+00
2.27451831e-01 2.97778130e-01 1.48273140e-01 -1.33753276e+00
-2.83920765e-01 -2.04292297e-01 -1.13022113e+00 4.28285122e-01
4.33309972e-01 2.90300846e-01 -8.64387035e-01 1.74797320e+00
6.45622909e-01 1.60863578e-01 -2.64091671e-01 1.45465660e+00
9.85826850e-01 5.68373084e-01 -1.69950023e-01 1.53820738e-01
1.41248381e+00 -1.21234083e+00 -5.96224070e-01 1.36211574e-01
-1.93058848e-01 -7.68868446e-01 1.21304846e+00 3.99894655e-01
-1.60369825e+00 -7.13931441e-01 -8.78877819e-01 -3.49741936e-01
3.45036089e-01 -2.52423882e-01 4.40477878e-01 2.61739314e-01
-1.06646872e+00 6.47161722e-01 -8.65892470e-01 -2.83713549e-01
1.96724951e-01 -4.81132884e-03 -6.63663745e-01 -6.24029227e-02
-1.01562870e+00 8.00724030e-01 -1.99191958e-01 1.04732633e-01
-1.10888231e+00 -8.57222497e-01 -1.06586933e+00 -3.04649055e-01
3.79873574e-01 -1.30909514e+00 9.85963643e-01 -9.58120942e-01
-1.75604618e+00 8.81054580e-01 1.49385452e-01 3.83525118e-02
1.11118901e+00 -3.74782048e-02 -1.19887352e-01 1.21080890e-01
-1.34997338e-01 8.18738103e-01 8.34896028e-01 -1.70366335e+00
-1.66680217e-01 -3.08458209e-01 2.24641740e-01 5.13901472e-01
2.70202249e-01 -3.73782098e-01 -8.53120506e-01 -1.02071881e+00
3.69116962e-01 -1.12620926e+00 -2.74662852e-01 7.58813679e-01
-3.21002245e-01 1.94880649e-01 6.25419617e-01 -9.56096113e-01
6.23250067e-01 -1.89353108e+00 8.50037634e-01 1.16545148e-01
4.03233290e-01 -3.23229998e-01 -1.35159194e-01 1.72911897e-01
1.83377221e-01 -2.87131846e-01 -1.68557510e-01 -7.45856643e-01
6.96039051e-02 4.37873781e-01 -5.04164547e-02 7.07479119e-01
-6.12555519e-02 9.34565485e-01 -8.28098714e-01 -6.15043044e-01
5.22296190e-01 9.22681689e-01 -8.51190507e-01 2.42154032e-01
-2.52006352e-01 1.09897780e+00 -8.84739310e-02 7.98511267e-01
7.70948172e-01 -1.65032089e-01 2.14238942e-01 -5.97862303e-01
1.11271910e-01 -2.43644379e-02 -1.32104969e+00 2.42455673e+00
-4.02483076e-01 2.48899236e-01 3.59668255e-01 -4.51101601e-01
6.62319839e-01 3.02995920e-01 7.15972424e-01 -7.30655611e-01
1.05054513e-01 -6.22387417e-03 -4.23717320e-01 -3.30278307e-01
6.27144814e-01 -2.53585875e-01 -1.30006239e-01 1.30960718e-01
-1.70279983e-02 -3.56548965e-01 -4.52100158e-01 -5.28575704e-02
6.77260160e-01 8.61183941e-01 4.45672497e-02 -5.77152967e-02
1.91808879e-01 3.87954675e-02 4.48772311e-01 2.45493710e-01
4.97973114e-02 1.28600252e+00 1.50616050e-01 -9.37359154e-01
-1.72577977e+00 -1.35165167e+00 -5.88358939e-02 9.15751278e-01
5.88602126e-01 -4.18884337e-01 -6.47438467e-01 -3.51348162e-01
-7.53118843e-03 4.21444267e-01 -8.09278011e-01 -1.28303301e-02
-8.88128698e-01 -1.16782114e-01 4.80279535e-01 5.57360590e-01
4.75231856e-01 -9.48814750e-01 -7.58979261e-01 9.02473032e-02
-5.12823462e-01 -1.22885823e+00 -6.88516140e-01 -3.75743836e-01
-7.19496787e-01 -7.10063815e-01 -8.71377587e-01 -6.76940322e-01
7.13433504e-01 1.64313450e-01 1.48692298e+00 1.16410062e-01
-2.82863975e-01 5.39970219e-01 -1.75381467e-01 1.98332369e-01
-5.51136076e-01 -3.15905839e-01 1.58061907e-01 -1.42962396e-01
-6.00119710e-01 -8.08311343e-01 -8.50761950e-01 6.08418226e-01
-6.61624372e-01 7.77208507e-01 1.66140914e-01 7.37991810e-01
8.83751750e-01 -3.42412710e-01 -2.70820975e-01 -4.65751559e-01
1.33746937e-01 -6.53567240e-02 -2.89247781e-01 1.12518445e-01
-1.32046983e-01 -8.92578140e-02 4.96309906e-01 -6.03928447e-01
-7.92025208e-01 4.32653010e-01 -4.50902611e-01 -8.83838594e-01
3.75485560e-03 -1.44744664e-01 -2.87574649e-01 -1.31774142e-01
5.61626434e-01 1.03524864e-01 1.90527707e-01 -5.96627712e-01
5.31281710e-01 -1.72778033e-02 7.52034724e-01 -9.45855856e-01
7.59801984e-01 7.03446388e-01 1.26955569e-01 -5.97406387e-01
-5.53913236e-01 6.03013560e-02 -8.31076801e-01 -5.12116790e-01
9.71611857e-01 -1.04763043e+00 -9.83764410e-01 3.20776045e-01
-1.28399825e+00 -4.29329306e-01 -5.93377948e-01 8.25643316e-02
-1.03662729e+00 5.80364823e-01 -7.89970517e-01 -4.68491048e-01
-2.75131881e-01 -1.38512373e+00 1.59500647e+00 -3.09705794e-01
-4.39400375e-01 -9.12748218e-01 7.08123595e-02 4.06763494e-01
1.03673063e-01 8.94636571e-01 5.94891310e-01 3.71548891e-01
-6.24683857e-01 1.57245487e-01 3.67495567e-02 9.38606486e-02
-7.46280923e-02 -1.21845178e-01 -6.53147221e-01 -4.07194108e-01
-2.36758873e-01 -1.96338505e-01 3.84045899e-01 3.26014131e-01
9.74516273e-01 -4.02716041e-01 1.83393881e-01 1.00658822e+00
1.23617470e+00 -4.17940617e-01 7.30339646e-01 1.13699406e-01
1.25597310e+00 5.94894886e-01 4.09373134e-01 6.37652099e-01
6.65615439e-01 1.19933712e+00 6.72099710e-01 -3.84553432e-01
-5.81756413e-01 -7.33076692e-01 3.23672414e-01 8.57529461e-01
-7.70681500e-01 -1.49140973e-03 -4.63154554e-01 2.49580175e-01
-1.82708621e+00 -8.78741860e-01 -1.91256717e-01 2.09932661e+00
7.20912397e-01 -1.93906009e-01 4.79905427e-01 -2.09949598e-01
4.21366662e-01 2.17689231e-01 -6.15810990e-01 -2.23877504e-01
-2.53112853e-01 1.49611741e-01 3.56942356e-01 6.65655494e-01
-7.55461812e-01 9.61542249e-01 6.57428408e+00 5.05739689e-01
-8.00040483e-01 1.65561438e-01 4.00096416e-01 -2.12425262e-01
-5.86254597e-01 -1.15156867e-01 -1.95620060e-01 7.99966007e-02
2.66040862e-01 3.95970732e-01 6.80524707e-01 8.86514902e-01
8.87377784e-02 3.24439913e-01 -1.17020345e+00 1.22123301e+00
3.37474704e-01 -1.38245201e+00 2.94118762e-01 1.85849518e-01
7.69703329e-01 -3.73075664e-01 3.41915637e-02 -1.37559727e-01
3.16462934e-01 -1.07187021e+00 1.55109692e+00 7.06841826e-01
1.36103630e+00 -7.06250787e-01 2.88690984e-01 1.78456187e-01
-1.34037042e+00 5.22197366e-01 -3.49583983e-01 4.26019840e-02
4.40122068e-01 4.28111404e-02 -3.75735313e-01 6.24516785e-01
8.20268095e-01 7.44419396e-01 -2.67212570e-01 4.06915009e-01
2.40040980e-02 -3.81175019e-02 -3.45266551e-01 4.20938641e-01
-7.92733878e-02 -3.77577424e-01 6.45716190e-01 8.19160283e-01
2.35735044e-01 2.14933008e-01 2.56499112e-01 1.07965672e+00
3.24180815e-04 -1.22312188e-01 -5.31565070e-01 5.30356884e-01
-4.79282625e-02 1.09158981e+00 -6.37771428e-01 -3.03577363e-01
-1.15514845e-01 1.72729802e+00 2.29913265e-01 2.34076813e-01
-1.20818436e+00 3.93146694e-01 7.44272649e-01 7.01761067e-01
2.51026571e-01 -5.65987587e-01 -3.25484216e-01 -1.36857486e+00
1.98792160e-01 -1.06795120e+00 3.10868528e-02 -9.41432476e-01
-1.24554932e+00 6.19338870e-01 -3.05643231e-02 -1.34822655e+00
-3.77039537e-02 -3.84584486e-01 -2.85698101e-02 6.78779602e-01
-7.93729663e-01 -1.66629517e+00 -6.18451238e-01 7.49745965e-01
7.55977154e-01 3.29116762e-01 9.98986065e-01 3.21345717e-01
1.03392508e-02 5.76805413e-01 -3.23562235e-01 8.54106396e-02
6.59121811e-01 -1.11579692e+00 8.00718188e-01 4.68580991e-01
9.03423876e-02 2.98820674e-01 8.07638288e-01 -7.21491635e-01
-1.90882468e+00 -9.02468324e-01 2.04629838e-01 -7.89802790e-01
8.48227888e-02 -6.32298172e-01 -5.81744790e-01 7.49144912e-01
-1.70184411e-02 3.36228788e-01 2.75141031e-01 -9.33374241e-02
-4.85846639e-01 1.29266232e-01 -1.27429581e+00 1.00713301e+00
1.73293746e+00 -3.37940842e-01 -3.51707816e-01 9.59354118e-02
6.25521243e-01 -1.05683196e+00 -1.22051668e+00 3.68639320e-01
1.18474388e+00 -1.06259179e+00 1.55224252e+00 -5.10399759e-01
7.20933795e-01 -4.91978347e-01 -4.95280147e-01 -1.17311525e+00
-6.17491603e-01 -7.26949155e-01 -1.97132558e-01 7.13740826e-01
-5.99516258e-02 9.26790833e-02 8.61062765e-01 6.24112368e-01
-1.13937274e-01 -7.18109190e-01 -9.81419265e-01 -4.79873300e-01
-7.83056617e-02 -3.48616093e-01 7.17945337e-01 9.38293695e-01
-3.18357468e-01 1.75875783e-01 -1.18162525e+00 -4.58897091e-03
8.90569210e-01 2.03944474e-01 1.16482484e+00 -1.11597979e+00
-4.38409597e-01 -1.78579614e-02 -5.40993035e-01 -1.52704382e+00
-1.51795939e-01 -7.06343591e-01 8.32836032e-02 -1.50616014e+00
1.70832202e-01 -4.46414590e-01 2.64462739e-01 1.72383428e-01
1.51070729e-01 7.70160556e-01 3.60302418e-01 2.39411831e-01
-6.93960428e-01 8.13035786e-01 2.03033948e+00 6.47724494e-02
-1.00948557e-01 -3.55708599e-01 -3.84935260e-01 9.14173663e-01
2.50890076e-01 -1.61319911e-01 -2.17257947e-01 -8.45129907e-01
2.85034031e-01 4.82362956e-01 8.44186664e-01 -8.15836191e-01
-1.55878693e-01 -2.53227264e-01 7.52228618e-01 -5.53460240e-01
8.79788637e-01 -1.11077356e+00 9.70235169e-01 3.21771294e-01
-2.80934483e-01 4.68962073e-01 -1.06896840e-01 6.71406150e-01
3.31314921e-01 4.12202358e-01 9.31935191e-01 -4.00022268e-01
-4.04398024e-01 6.23157740e-01 1.61202773e-01 1.61558688e-02
8.14458013e-01 -4.06430632e-01 2.11751252e-01 -6.16945326e-01
-9.89526153e-01 -1.79484680e-01 1.25111628e+00 5.91220498e-01
9.70033646e-01 -1.73453939e+00 -8.77956629e-01 3.21319938e-01
1.21329322e-01 2.80279100e-01 5.26516736e-01 7.84986973e-01
-1.13667536e+00 -2.33250335e-01 -5.58227777e-01 -8.96148205e-01
-1.25972772e+00 5.05239010e-01 4.06554937e-01 4.83520143e-02
-1.33789277e+00 5.97463012e-01 3.76345336e-01 -6.20844245e-01
3.16077888e-01 -2.82972574e-01 4.02707011e-01 -6.08531237e-01
2.53541470e-01 3.21812242e-01 -1.35394678e-01 -1.05300951e+00
-1.43826157e-01 8.79436970e-01 4.79451776e-01 -3.00093949e-01
1.34627676e+00 -2.05997050e-01 -4.51879352e-02 3.19131732e-01
1.01605737e+00 2.57493228e-01 -1.69404721e+00 -1.70704052e-01
-8.18999469e-01 -7.54447341e-01 -3.81433636e-01 -6.80939078e-01
-1.21821952e+00 6.62775278e-01 4.54886615e-01 -2.20673829e-01
8.78024876e-01 9.13628191e-02 1.06869948e+00 -6.16107211e-02
7.36241996e-01 -8.47707689e-01 9.19458568e-02 2.31261119e-01
1.39207423e+00 -1.04818380e+00 1.70554072e-01 -6.92618132e-01
-8.33122492e-01 1.07552552e+00 6.10545754e-01 -4.26151395e-01
4.42922175e-01 2.60674030e-01 -8.05994123e-02 -3.44847530e-01
-4.35048372e-01 3.61919999e-01 4.61473525e-01 7.29986787e-01
1.54030427e-01 2.42360324e-01 1.15453541e-01 4.16558564e-01
-4.62270856e-01 -3.14562351e-01 2.95723587e-01 6.68289185e-01
9.70306545e-02 -9.12859201e-01 -6.26897037e-01 -1.32726789e-01
-4.59775031e-02 2.04694480e-01 -4.61586088e-01 8.07303905e-01
2.92586684e-01 5.48533559e-01 1.07247151e-01 -6.73644364e-01
8.50748003e-01 -3.78405869e-01 1.10304391e+00 -2.88922161e-01
-4.80805635e-01 2.37287119e-01 2.22095072e-01 -8.54882061e-01
-4.14035499e-01 -5.49840033e-01 -9.84173477e-01 -7.73072124e-01
9.88443047e-02 -3.23527843e-01 4.89271671e-01 3.88742805e-01
3.51782084e-01 5.36348164e-01 4.13093984e-01 -1.34157598e+00
-3.19792718e-01 -6.42323792e-01 -4.37037319e-01 8.30437243e-01
2.34632939e-01 -8.35368276e-01 1.04671139e-02 2.84922719e-01] | [7.277676582336426, -1.2697633504867554] |
b7e47a4f-bcfe-4c7a-85e0-efdd8c10ac97 | non-rigid-point-cloud-registration-with | 2205.12796 | null | https://arxiv.org/abs/2205.12796v3 | https://arxiv.org/pdf/2205.12796v3.pdf | Non-rigid Point Cloud Registration with Neural Deformation Pyramid | Non-rigid point cloud registration is a key component in many computer vision and computer graphics applications. The high complexity of the unknown non-rigid motion make this task a challenging problem. In this paper, we break down this problem via hierarchical motion decomposition. Our method called Neural Deformation Pyramid (NDP) represents non-rigid motion using a pyramid architecture. Each pyramid level, denoted by a Multi-Layer Perception (MLP), takes as input a sinusoidally encoded 3D point and outputs its motion increments from the previous level. The sinusoidal function starts with a low input frequency and gradually increases when the pyramid level goes down. This allows a multi-level rigid to nonrigid motion decomposition and also speeds up the solving by 50 times compared to the existing MLP-based approach. Our method achieves advanced partialto-partial non-rigid point cloud registration results on the 4DMatch/4DLoMatch benchmark under both no-learned and supervised settings. | ['Tatsuya Harada', 'Yang Li'] | 2022-05-25 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [ 1.29315212e-01 -1.93289116e-01 -1.81784853e-01 5.95785072e-03
-9.62002635e-01 -4.14284676e-01 4.91299868e-01 -3.51418614e-01
-3.18345368e-01 3.03246409e-01 2.19784915e-01 -3.26966569e-02
1.28459156e-01 -6.69924736e-01 -1.17942691e+00 -1.02005553e+00
-1.61297649e-01 8.82905483e-01 4.36007708e-01 -1.70392804e-02
1.33506790e-01 8.34909558e-01 -1.25008631e+00 3.35323751e-01
5.67034245e-01 5.54222703e-01 3.76231633e-02 6.77522123e-01
2.46505663e-01 2.59710878e-01 9.13224816e-02 1.38869330e-01
4.96314973e-01 1.24502979e-01 -1.02080929e+00 -1.55599127e-02
8.40127409e-01 -2.58234739e-01 -2.27543175e-01 1.00467014e+00
4.85788345e-01 3.99477303e-01 5.30978739e-01 -8.26525509e-01
-6.44071758e-01 3.13349426e-01 -1.10655200e+00 -8.92356783e-02
3.39392692e-01 3.07436083e-02 6.85706735e-01 -1.09196448e+00
7.82782674e-01 1.58796751e+00 1.12994063e+00 4.86826807e-01
-1.46282041e+00 -3.00127774e-01 -2.02393699e-02 -5.12451418e-02
-1.28830206e+00 -7.47511759e-02 7.76573002e-01 -7.17371106e-01
1.07063687e+00 2.20714554e-01 5.29225528e-01 5.48632026e-01
4.08354700e-01 3.49368632e-01 7.90371537e-01 -1.93810344e-01
3.98875512e-02 -7.14023709e-01 3.89247127e-02 7.31452167e-01
2.81108841e-02 3.83794978e-02 -1.55936927e-01 -3.56374860e-01
1.33848774e+00 3.52907665e-02 -3.31470758e-01 -3.38192791e-01
-1.54327881e+00 4.35600191e-01 6.43755078e-01 4.41587389e-01
-5.87448239e-01 5.93138337e-01 -7.06558153e-02 1.15821086e-01
3.65080923e-01 -4.14561331e-02 -5.30954957e-01 -3.95808257e-02
-1.10991073e+00 3.43024254e-01 7.83042789e-01 6.04870915e-01
1.08196068e+00 1.44615486e-01 -6.76868996e-03 5.55416465e-01
4.47169095e-01 5.33194602e-01 6.35744035e-01 -1.41204929e+00
3.43881011e-01 4.44502920e-01 2.33891100e-01 -1.16453159e+00
-5.89657426e-01 -1.16562080e-02 -1.16835690e+00 6.38734639e-01
4.00833994e-01 1.09271601e-01 -1.11538005e+00 1.45122457e+00
7.08370209e-01 7.21764743e-01 -1.11948892e-01 9.95857537e-01
7.34509706e-01 9.10090864e-01 -8.33763182e-02 -2.32983530e-01
1.35035729e+00 -1.03246915e+00 -3.16377610e-01 5.85621744e-02
2.88009554e-01 -8.13708186e-01 1.10486293e+00 1.28513843e-01
-1.23995650e+00 -7.39011049e-01 -7.06203163e-01 -3.74666959e-01
1.71287134e-01 -1.17751829e-01 1.60989329e-01 9.34749097e-02
-1.41975391e+00 1.09774411e+00 -1.49497664e+00 5.59071265e-02
1.54378071e-01 8.06589961e-01 -5.91286600e-01 2.73104347e-02
-7.89266467e-01 7.14486539e-01 1.91984653e-01 2.10379168e-01
-2.42909491e-01 -8.75200868e-01 -8.55822384e-01 -5.34869581e-02
-3.43248136e-02 -1.04388618e+00 1.15197551e+00 -6.14617944e-01
-1.77167380e+00 1.00764620e+00 -4.30329323e-01 -3.25366914e-01
5.71936429e-01 -2.37645596e-01 2.41057441e-01 -7.55547881e-02
-7.71909952e-02 9.33374763e-01 1.02277112e+00 -1.29162979e+00
-1.61124572e-01 -3.61597091e-01 -3.33847135e-01 3.88135254e-01
3.58670950e-01 -1.00201108e-01 -7.81404078e-01 -7.51321435e-01
5.94869494e-01 -1.53369594e+00 -3.15648764e-01 -1.27043322e-01
-2.14454651e-01 -2.29454964e-01 1.00223649e+00 -7.69884169e-01
5.08618891e-01 -1.94798243e+00 5.79686821e-01 4.46832031e-02
5.83887212e-02 -4.76793572e-02 -5.67830540e-02 1.80904903e-02
-3.25429261e-01 -4.13569994e-02 -5.98188043e-01 -5.67160308e-01
-2.67808847e-02 2.28170529e-01 -1.83999762e-01 7.09237695e-01
1.39535395e-02 9.94956911e-01 -6.88093126e-01 -4.27357256e-01
2.96860009e-01 1.02313614e+00 -1.00506949e+00 -8.34244043e-02
-2.84467451e-02 9.58185375e-01 -3.54294688e-01 6.08327448e-01
9.16655540e-01 -3.72115403e-01 -2.86185861e-01 -5.48203945e-01
-2.79740930e-01 -1.68051586e-01 -1.41699219e+00 1.97849786e+00
-3.97179127e-01 3.39159489e-01 8.41584653e-02 -5.70395589e-01
6.21756196e-01 3.80158573e-01 8.56428921e-01 -1.60574183e-01
-1.99973285e-01 1.75813586e-01 -1.18985765e-01 -3.38872164e-01
5.30518234e-01 -2.86769181e-01 3.28478999e-02 1.86397955e-01
-3.36737335e-01 -3.72786105e-01 -3.96059990e-01 -5.32224357e-01
1.00915229e+00 7.77391434e-01 1.32602736e-01 -2.88429707e-01
7.55896986e-01 1.28691513e-02 7.22850740e-01 2.86569297e-01
6.33560941e-02 9.41239357e-01 1.20156161e-01 -8.24237108e-01
-9.56152856e-01 -9.41328049e-01 -1.41797364e-01 9.13500667e-01
1.77604049e-01 6.49283230e-02 -8.39524567e-01 -8.77000913e-02
2.44984642e-01 2.11483970e-01 -3.56010705e-01 1.99704930e-01
-1.38375902e+00 -8.03052068e-01 1.67075038e-01 5.57364643e-01
5.09445548e-01 -1.23781347e+00 -5.83013356e-01 2.71244884e-01
-3.00152451e-01 -1.26320910e+00 -8.57837439e-01 -2.49599859e-01
-1.30220008e+00 -7.49477863e-01 -8.83620679e-01 -1.07812750e+00
7.62198091e-01 2.81886067e-02 9.60823894e-01 -1.64390415e-01
8.35684873e-03 2.59655207e-01 2.03115985e-01 2.38894045e-01
-4.74323541e-01 3.87780368e-02 2.82374829e-01 1.22044377e-01
-2.66608205e-02 -9.18858826e-01 -6.71793401e-01 3.37232769e-01
-7.22233772e-01 5.00083007e-02 5.30573845e-01 5.36374986e-01
1.42613494e+00 -1.94247290e-01 -1.15556709e-01 -5.71921170e-01
2.68703282e-01 -2.17230290e-01 -7.73801565e-01 -4.03536558e-02
2.61079762e-02 2.21283510e-01 4.20065910e-01 -6.04573011e-01
-7.35248685e-01 7.66562164e-01 -3.02818537e-01 -9.29062128e-01
-1.81665942e-01 1.08112000e-01 3.54896747e-02 -7.11019695e-01
5.48371613e-01 1.64777949e-01 -2.88183659e-01 -6.92542672e-01
4.23945010e-01 5.92575073e-02 9.39918399e-01 -7.03875184e-01
1.11883152e+00 7.13467479e-01 3.95404398e-01 -8.74902725e-01
-2.12060884e-01 -3.81689668e-01 -1.16229939e+00 -1.04713947e-01
1.24579835e+00 -8.82687092e-01 -1.12735188e+00 6.66263878e-01
-1.50875115e+00 -6.29209936e-01 -2.98827052e-01 4.82751787e-01
-7.88927674e-01 6.32180035e-01 -7.75335848e-01 -3.67722660e-01
-6.09945655e-01 -1.53459549e+00 1.39765131e+00 9.93740037e-02
-1.21617481e-01 -8.75981688e-01 3.95653129e-01 1.32603675e-01
2.58408725e-01 7.64664590e-01 7.60439873e-01 2.99233198e-02
-8.14846873e-01 -8.11130255e-02 1.23852782e-01 3.41718495e-02
2.49313071e-01 -3.10122035e-02 -7.33613014e-01 -4.75945503e-01
3.92050326e-01 -6.09642155e-02 6.53579116e-01 6.79715753e-01
1.18240964e+00 -4.62956727e-01 -2.40785271e-01 1.14916492e+00
1.39353061e+00 -2.07714066e-01 4.54621792e-01 2.24210411e-01
1.36439097e+00 3.79472256e-01 8.02474320e-02 -9.38542280e-03
5.81086695e-01 1.06622612e+00 4.87658262e-01 -1.52927458e-01
-6.65188655e-02 1.22669242e-01 3.15965116e-01 1.09169281e+00
-7.52490759e-01 5.04136443e-01 -1.08159184e+00 2.05678284e-01
-1.77639091e+00 -8.76422465e-01 -4.68064576e-01 2.14626384e+00
1.01107371e+00 -1.11127645e-01 -7.33773708e-02 9.96339973e-03
6.30937934e-01 1.39125541e-01 -6.08460248e-01 -2.01958388e-01
5.11669107e-02 1.07480668e-01 4.82200176e-01 1.01838458e+00
-1.18568456e+00 1.13333571e+00 5.66426992e+00 3.74715447e-01
-1.25859797e+00 9.77765620e-02 3.54200304e-01 2.40185007e-01
-9.62824449e-02 -1.90141052e-01 -8.11151862e-01 3.38171840e-01
7.78191268e-01 8.73219804e-04 5.74314177e-01 6.82009041e-01
2.89410561e-01 2.77084112e-01 -1.27981555e+00 1.05115926e+00
-1.13170087e-01 -1.42054081e+00 1.95702329e-01 3.40060033e-02
8.98007989e-01 5.38787186e-01 -3.35492343e-02 -9.58106946e-03
2.67789900e-01 -9.89839792e-01 4.65782851e-01 6.54313743e-01
7.28393614e-01 -5.06352425e-01 3.46622437e-01 5.53568780e-01
-1.50555670e+00 5.01654744e-01 -3.21997106e-01 2.23200008e-01
4.38374579e-01 2.27021173e-01 -4.15914387e-01 5.35548270e-01
8.66595149e-01 5.16030252e-01 -2.31976137e-02 9.06800449e-01
-5.75733893e-02 2.90130943e-01 -5.58921456e-01 7.79703140e-01
1.77683860e-01 -5.10638356e-01 7.77426183e-01 1.15305889e+00
4.59202558e-01 2.61221647e-01 4.25819546e-01 6.92123711e-01
-2.95531541e-01 -2.46457115e-01 -3.82244438e-01 6.88950777e-01
2.28504673e-01 1.18176520e+00 -6.82046533e-01 -1.69744223e-01
-3.90901327e-01 1.18586898e+00 1.48540020e-01 4.14436072e-01
-6.49729252e-01 8.55590329e-02 7.15540767e-01 2.60896742e-01
3.27289462e-01 -6.76603019e-01 -2.21323833e-01 -1.13237977e+00
-3.74940187e-02 -7.38981724e-01 1.81777403e-01 -7.23537624e-01
-9.47515786e-01 7.35928893e-01 8.92472565e-02 -1.32818544e+00
-4.18741971e-01 -5.12564123e-01 -5.34374595e-01 1.12735569e+00
-1.54865074e+00 -1.37788713e+00 -4.45383668e-01 7.44888484e-01
5.92375338e-01 3.76839042e-01 9.05939519e-01 3.09860379e-01
-1.78327933e-01 2.21038640e-01 -1.41585087e-02 1.26247823e-01
5.94786108e-01 -1.26368344e+00 8.74763131e-01 7.27477074e-01
6.02635480e-02 4.59975868e-01 6.30973935e-01 -6.61524236e-01
-1.67983031e+00 -1.08455682e+00 8.39814186e-01 -7.08099365e-01
4.83387768e-01 1.03829578e-01 -1.55290568e+00 8.36265266e-01
-5.70262372e-02 5.16977012e-01 2.83361077e-01 -6.27823532e-01
-9.04042125e-02 -1.28136203e-02 -1.29778457e+00 5.98898470e-01
8.93205106e-01 -4.02434111e-01 -8.49748671e-01 3.03660095e-01
9.18602049e-01 -1.15063071e+00 -1.19528890e+00 5.15900433e-01
3.42656702e-01 -5.92773259e-01 1.39356673e+00 -3.86617631e-01
2.39515588e-01 -5.83112001e-01 -1.17600940e-01 -1.07843685e+00
-6.79188490e-01 -1.17395949e+00 -3.25372607e-01 6.19832635e-01
-8.04275572e-02 -4.38411176e-01 8.97908270e-01 6.36251092e-01
-2.90866077e-01 -7.20144808e-01 -1.18349290e+00 -5.98753691e-01
4.58148539e-01 -2.88078159e-01 4.11346823e-01 1.14921391e+00
-6.09594107e-01 -1.07578255e-01 -2.86133438e-01 6.89402997e-01
9.37176943e-01 2.86777884e-01 6.57311857e-01 -1.37736893e+00
-3.21885467e-01 -4.54598337e-01 -3.57031643e-01 -1.16227651e+00
2.17590287e-01 -1.08419394e+00 3.97557952e-02 -1.48533988e+00
-2.18304396e-01 -2.63187617e-01 1.11775570e-01 7.22507238e-01
-1.18295767e-03 5.58213651e-01 9.07429308e-02 8.46519172e-01
-4.20493186e-02 2.95873255e-01 1.43186688e+00 -3.90716679e-02
-6.77861333e-01 6.12932220e-02 -3.40694562e-02 1.28718352e+00
5.61968505e-01 -5.28282821e-01 5.56998886e-02 -8.29908013e-01
-5.45167364e-03 2.91197568e-01 4.10056800e-01 -9.98474836e-01
5.71281791e-01 -2.36361429e-01 2.48184711e-01 -8.38074982e-01
4.13651139e-01 -8.33874464e-01 6.23909771e-01 7.34073639e-01
1.82518363e-01 6.01338148e-01 3.58734518e-01 3.95569831e-01
-1.62332907e-01 2.30304837e-01 1.04171407e+00 -1.73910275e-01
-4.19829071e-01 6.88581645e-01 -2.20936071e-02 -1.04870848e-01
6.91387415e-01 -3.23757619e-01 1.31981149e-01 4.88012098e-02
-9.57030475e-01 8.11982527e-02 7.16245353e-01 3.62773120e-01
5.99249423e-01 -1.65260732e+00 -7.40059674e-01 2.96632349e-01
-4.45244431e-01 6.44881308e-01 3.09362590e-01 1.06910515e+00
-8.85789454e-01 1.62322268e-01 -1.92318067e-01 -1.06560779e+00
-1.32533157e+00 2.67290056e-01 5.50428808e-01 -2.57779062e-01
-1.16412771e+00 6.26217246e-01 1.96386173e-01 -5.06364048e-01
1.62439331e-01 -8.35621595e-01 -1.85344532e-01 -3.15937459e-01
2.34474912e-01 4.23220158e-01 1.45935073e-01 -1.18277347e+00
-3.94839942e-01 1.65245354e+00 2.65757948e-01 -1.55142173e-01
1.47660363e+00 2.60952771e-01 -4.57383066e-01 4.30587679e-01
1.33107007e+00 4.26171944e-02 -1.53611577e+00 -3.15024763e-01
9.31303948e-02 -1.95823744e-01 -8.76126066e-02 -1.01598561e-01
-9.84746337e-01 7.36826539e-01 5.41744947e-01 -1.84792392e-02
9.35697198e-01 -2.76442710e-02 8.37749124e-01 6.63273454e-01
3.77961308e-01 -6.12322271e-01 -9.87649933e-02 7.60357738e-01
1.24857378e+00 -1.04732227e+00 7.40660727e-02 -3.01779628e-01
-3.84203583e-01 1.13626230e+00 3.08190376e-01 -5.88229299e-01
7.60932803e-01 3.45810950e-01 -1.68550014e-01 4.57395576e-02
-2.19339564e-01 2.59364456e-01 7.35846817e-01 4.33476925e-01
3.28906536e-01 -5.40901572e-02 -4.05501463e-02 2.14343488e-01
-3.17578316e-01 1.46249816e-01 2.52096146e-01 7.07970023e-01
-3.26003402e-01 -1.24144828e+00 -7.82874703e-01 -8.86400938e-02
-5.07919788e-01 9.90985036e-02 4.57230695e-02 7.22319305e-01
9.24011394e-02 1.09686293e-01 3.12033385e-01 -2.77132392e-01
4.98400360e-01 -6.24175481e-02 5.88714778e-01 -4.53253388e-01
-6.85466409e-01 4.12736177e-01 -3.67912412e-01 -9.10810530e-01
-6.77397430e-01 -7.37390518e-01 -1.84573030e+00 -3.32626581e-01
3.13157231e-01 -2.27372363e-01 5.47317684e-01 6.51314437e-01
4.55628872e-01 3.64026546e-01 4.96069759e-01 -1.90119362e+00
-5.72495997e-01 -7.22971141e-01 3.79386805e-02 4.53176141e-01
6.43955708e-01 -3.97096425e-01 -3.24396312e-01 5.35656631e-01] | [8.166386604309082, -2.4280550479888916] |
ec62b3ca-4261-4b70-889b-7c585421915d | interpretable-deep-learning-for-forecasting | 2302.05762 | null | https://arxiv.org/abs/2302.05762v1 | https://arxiv.org/pdf/2302.05762v1.pdf | Interpretable Deep Learning for Forecasting Online Advertising Costs: Insights from the Competitive Bidding Landscape | As advertisers increasingly shift their budgets toward digital advertising, forecasting advertising costs is essential for making budget plans to optimize marketing campaign returns. In this paper, we perform a comprehensive study using a variety of time-series forecasting methods to predict daily average cost-per-click (CPC) in the online advertising market. We show that forecasting advertising costs would benefit from multivariate models using covariates from competitors' CPC development identified through time-series clustering. We further interpret the results by analyzing feature importance and temporal attention. Finally, we show that our approach has several advantages over models that individual advertisers might build based solely on their collected data. | ['Maximilian Kaiser', 'Qiwei Han', 'Fynn Oldenburg'] | 2023-02-11 | null | null | null | null | ['marketing', 'time-series-clustering'] | ['miscellaneous', 'time-series'] | [-8.88671279e-02 -2.50789315e-01 -9.85277057e-01 -1.04282236e+00
-7.24981427e-01 -4.73565638e-01 7.52401888e-01 2.61498958e-01
-2.46236876e-01 2.39154458e-01 6.04456127e-01 -8.28270495e-01
-4.47153568e-01 -6.42058849e-01 -5.21829963e-01 -1.98861491e-02
-4.87269104e-01 4.01100755e-01 -2.19790339e-01 -1.34123340e-01
4.12160665e-01 3.33702564e-01 -1.59999669e+00 3.75339121e-01
6.61697149e-01 1.42120862e+00 -9.38685536e-02 3.56202126e-01
-1.71493739e-01 5.27060032e-01 -5.89033604e-01 -8.50763857e-01
4.89761770e-01 -2.34064013e-02 9.99909043e-02 2.38586470e-01
2.85668999e-01 -2.99202263e-01 -4.65537488e-01 8.13542843e-01
-2.38330409e-01 -3.52586098e-02 3.06766123e-01 -9.00467396e-01
-8.37483227e-01 8.49336505e-01 -3.52216721e-01 6.63507223e-01
-9.63808447e-02 6.32343888e-02 1.45497882e+00 -2.60933757e-01
6.88601017e-01 1.15239787e+00 4.32042867e-01 -9.17580947e-02
-1.58179486e+00 -4.98264700e-01 1.03165638e+00 2.76495785e-01
-7.62944520e-01 -2.16425523e-01 1.08402503e+00 -5.02316475e-01
5.55025578e-01 5.72178543e-01 9.28397477e-01 1.09228516e+00
7.60533094e-01 7.59154141e-01 1.45977259e+00 -1.37261271e-01
2.12069973e-01 2.47683212e-01 5.96273303e-01 -1.98984370e-01
-1.90544829e-01 6.86516821e-01 -2.24737957e-01 -4.77584064e-01
4.74109918e-01 3.29835534e-01 5.41779995e-01 3.12891901e-02
-5.09571075e-01 1.44895458e+00 3.65293413e-01 5.69888316e-02
-7.76106596e-01 2.64696747e-01 -1.64810121e-01 5.75543404e-01
1.16355038e+00 3.34673941e-01 -6.75950646e-01 -1.05452172e-01
-8.62780452e-01 4.68576431e-01 8.17331493e-01 7.56684184e-01
3.02931696e-01 -8.10560659e-02 3.98030989e-02 6.23421609e-01
4.05206203e-01 3.11687261e-01 4.96935576e-01 -9.99650478e-01
4.07449752e-02 2.10019574e-01 3.52144063e-01 -1.03197932e+00
-3.71447891e-01 -1.01489532e+00 1.08928621e-01 -3.12882483e-01
1.52572244e-01 1.22036330e-01 -8.39042246e-01 1.47474146e+00
-1.03465192e-01 -1.00249916e-01 -5.18658757e-01 7.12603092e-01
-8.77841339e-02 4.90497977e-01 5.35734892e-01 -6.55130208e-01
1.44383085e+00 -7.28022099e-01 -9.16625142e-01 -4.87895936e-01
3.71435702e-01 -9.36583519e-01 7.41262019e-01 5.75271189e-01
-9.50831711e-01 -3.89318466e-01 -6.21887207e-01 4.09605592e-01
-3.09415013e-01 -2.61323094e-01 1.55524445e+00 6.64116561e-01
-4.66482580e-01 6.98808908e-01 -1.00445724e+00 -3.83725874e-02
3.35379690e-01 9.09119323e-02 8.21260333e-01 1.91873372e-01
-9.74320531e-01 7.65034378e-01 -5.63572407e-01 -1.14814013e-01
-7.82240212e-01 -7.06642807e-01 -2.19131812e-01 2.04293076e-02
2.32704729e-01 -2.08201826e-01 1.86704695e+00 -1.11205065e+00
-1.14782381e+00 1.36959046e-01 -3.05560529e-01 -1.06202352e+00
2.19028428e-01 1.10763788e-01 -1.32737625e+00 -4.60771948e-01
4.94799800e-02 2.24047095e-01 7.29535699e-01 -8.35915506e-01
-9.51366782e-01 -8.42345119e-01 -2.13449314e-01 -3.43758136e-01
-1.31008834e-01 2.32282802e-01 -1.48768187e-01 -8.10493588e-01
2.63401389e-01 -1.15685415e+00 -8.50343704e-01 -6.17848933e-01
1.07444050e-02 -3.15201342e-01 4.46724236e-01 -7.86629856e-01
1.65435314e+00 -1.85386312e+00 -1.45164728e-01 4.11884099e-01
1.23739406e-01 -4.83096749e-01 2.85792705e-02 3.79653454e-01
1.07517717e-02 1.16998173e-01 4.42828596e-01 -9.84560400e-02
2.58445561e-01 2.87017733e-01 -6.62900329e-01 2.88991719e-01
6.57317862e-02 1.08178926e+00 -4.14280146e-01 -1.03278436e-01
-9.48178470e-02 9.48797092e-02 -6.88107491e-01 -4.92893308e-01
-8.25063944e-01 4.13818300e-01 -7.63052464e-01 1.00557244e+00
4.90704864e-01 -3.99787486e-01 5.29672444e-01 7.36199459e-03
-4.94483083e-01 5.66481352e-01 -4.39991921e-01 1.15835834e+00
-5.46525776e-01 6.29304647e-01 4.95594367e-02 -1.03019917e+00
5.67114294e-01 -2.49540970e-01 6.22630060e-01 -1.38709331e+00
2.90312797e-01 2.38694958e-02 4.18593645e-01 -1.14423208e-01
4.05483335e-01 -1.12783417e-01 -3.15984964e-01 5.61028838e-01
-3.79258782e-01 4.53125834e-01 1.08989730e-01 9.89817977e-02
8.73828888e-01 -4.91905719e-01 -2.43761107e-01 -2.78479397e-01
-2.14576602e-01 3.42172205e-01 6.31423116e-01 5.33199430e-01
-2.63320237e-01 -5.01522005e-01 4.16233063e-01 -7.88681388e-01
-9.26759005e-01 -9.21632528e-01 -5.68151057e-01 1.46290672e+00
-3.93560231e-01 -3.99092346e-01 1.57543644e-02 -3.88791472e-01
5.45194864e-01 1.10355771e+00 -8.11433077e-01 2.13050246e-01
-4.70487028e-01 -8.45038354e-01 -6.92575634e-01 5.08418739e-01
-1.40809864e-01 -3.69913787e-01 -1.57336414e-01 6.57775402e-01
5.23409881e-02 -7.87072361e-01 -5.75925410e-01 1.29484877e-01
-1.15325058e+00 -5.55931985e-01 -3.08243454e-01 -3.36693197e-01
2.14629009e-01 3.92606646e-01 1.30878282e+00 -4.94452119e-01
-3.72567385e-01 6.03620410e-01 -3.12159862e-02 -8.19730639e-01
-1.03428356e-01 -4.31514941e-02 -9.89717469e-02 1.31563887e-01
1.15989149e+00 -6.59445763e-01 -9.63230669e-01 7.24373519e-01
-5.07658720e-01 -4.78893459e-01 7.02115059e-01 3.11251253e-01
7.44626701e-01 2.34291315e-01 6.50986075e-01 -7.67380297e-01
9.87907052e-01 -8.73690426e-01 -1.23703122e+00 3.07345483e-02
-1.20987177e+00 -2.63911039e-02 4.75321971e-02 -1.03393757e+00
-9.14753497e-01 -1.35774836e-01 3.54347527e-01 -6.88766986e-02
4.13248926e-01 8.96424711e-01 5.00087440e-01 1.34319991e-01
3.76412362e-01 5.66852354e-02 1.37462676e-01 -9.15597141e-01
6.79177165e-01 6.01499319e-01 3.22107784e-02 -1.77261963e-01
4.06986445e-01 5.61854303e-01 -3.78982127e-01 -4.78265673e-01
-1.00123942e+00 -3.13311487e-01 -6.59430921e-02 -2.69242406e-01
6.27462089e-01 -1.03455091e+00 -1.05907702e+00 -3.54810953e-01
-6.74783051e-01 3.54806632e-02 -1.46661565e-01 1.00694406e+00
-3.63018900e-01 -2.19960272e-01 -7.29425490e-01 -1.07117319e+00
3.04657072e-01 -1.01231682e+00 6.88035250e-01 1.35791274e-02
-1.82942241e-01 -1.05266058e+00 8.20736885e-02 5.34600198e-01
7.59355187e-01 -4.73895110e-02 8.51769805e-01 -6.67826295e-01
-9.18334484e-01 -5.50094604e-01 -1.25984792e-02 1.23502254e-01
-3.61860469e-02 -7.34383017e-02 -5.05352855e-01 3.49443518e-02
4.14659202e-01 3.86403173e-01 8.73842835e-01 1.36976528e+00
1.43368101e+00 -6.09182656e-01 -5.29830337e-01 2.75310010e-01
1.35187101e+00 7.13202894e-01 2.93539077e-01 3.97314668e-01
-1.79173380e-01 1.00726533e+00 9.84531939e-01 6.80329025e-01
5.00355005e-01 8.27673316e-01 4.16244686e-01 3.58713657e-01
2.97513962e-01 -4.32798505e-01 1.84208110e-01 5.60195327e-01
-9.93047208e-02 1.58307090e-01 -3.14717799e-01 4.84658957e-01
-1.80968642e+00 -1.01318002e+00 -2.33276412e-01 2.09153891e+00
3.55638325e-01 3.70194793e-01 7.15553820e-01 -5.93693435e-01
7.01259196e-01 2.58027881e-01 -8.76422346e-01 -5.81466615e-01
-8.59773010e-02 1.47382453e-01 1.22089124e+00 5.01709819e-01
-9.27076876e-01 6.71164989e-01 8.09752846e+00 4.18662459e-01
-6.83936656e-01 1.27910048e-01 1.02876782e+00 -6.99922919e-01
-7.04721391e-01 2.24633887e-01 -1.09849536e+00 7.79825509e-01
1.67458701e+00 -5.10822952e-01 6.52208865e-01 1.21434736e+00
5.33729494e-01 1.95582837e-01 -9.30509746e-01 5.58463573e-01
-2.43306264e-01 -1.39358485e+00 -3.56244951e-01 9.51083302e-01
6.47781909e-01 6.01473078e-02 7.28753388e-01 4.55478072e-01
9.98791456e-01 -6.58374071e-01 5.36183238e-01 5.05063534e-01
2.52426025e-02 -5.68309724e-01 1.95151508e-01 4.84649986e-02
-7.92157412e-01 -8.05275738e-01 -3.76741320e-01 -3.83261621e-01
7.28778839e-01 8.48249614e-01 -5.05164564e-01 -9.24732313e-02
3.76445949e-01 5.86809695e-01 -5.32008708e-01 7.12698281e-01
3.36023897e-01 6.90013289e-01 -1.70191303e-01 -1.36897594e-01
4.94929314e-01 -5.37328422e-01 2.44287342e-01 5.10839999e-01
4.20861036e-01 1.07425377e-01 5.08913733e-02 6.49008214e-01
2.18062729e-01 9.28363502e-02 -2.11758971e-01 -7.08907068e-01
2.41299763e-01 9.62964833e-01 -3.36147279e-01 -2.29195610e-01
-8.87541234e-01 3.27392578e-01 -2.79421121e-01 3.45402241e-01
-1.11786604e+00 1.15016833e-01 9.92162347e-01 6.54303491e-01
5.03187597e-01 -5.72156489e-01 -9.19142663e-02 -9.67507899e-01
1.32766468e-02 -7.42742598e-01 4.35924888e-01 -3.20598930e-01
-1.68863606e+00 4.10542130e-01 -2.51331250e-03 -1.17416263e+00
-1.75462887e-01 -6.23730659e-01 -4.18590367e-01 5.76153278e-01
-1.56590617e+00 -1.02431846e+00 2.33034745e-01 3.02843243e-01
9.39164877e-01 -8.26706737e-02 2.92861432e-01 3.68807673e-01
-4.17003632e-01 1.59977585e-01 1.61132917e-01 -5.32109857e-01
3.11113089e-01 -8.03398550e-01 5.31180084e-01 2.09549069e-01
6.13096654e-02 8.19992721e-01 1.02810168e+00 -9.81202126e-01
-1.45590925e+00 -7.82946348e-01 9.67045009e-01 -3.99967164e-01
1.56410229e+00 -3.86444390e-01 -3.42801869e-01 1.19339216e+00
7.37471730e-02 -5.62769949e-01 1.13000119e+00 8.58147860e-01
-2.74940431e-01 -2.16278479e-01 -7.88898051e-01 4.27861840e-01
9.66435254e-01 -2.06683055e-01 -3.26592594e-01 9.36147273e-01
1.00533545e+00 4.19792861e-01 -1.37054873e+00 -1.09428400e-02
9.19494390e-01 -5.99006951e-01 9.63035882e-01 -1.18258238e+00
2.00840339e-01 4.97282207e-01 -5.46672404e-01 -1.16418815e+00
-8.26288104e-01 -5.26503861e-01 1.08169848e-02 7.07537770e-01
7.94684470e-01 -8.96181583e-01 9.85955298e-01 1.01739824e+00
8.63476545e-02 -3.54768038e-01 -8.65242302e-01 -1.01507187e+00
-2.14995574e-02 -7.98516691e-01 8.71445298e-01 7.38496780e-01
-1.23160794e-01 2.10499898e-01 -4.09356952e-01 4.57637273e-02
8.50515425e-01 6.93413436e-01 3.73214394e-01 -1.74242246e+00
-6.60617769e-01 -6.45424426e-01 -1.86549366e-01 -1.40350878e+00
4.39777523e-02 -5.50777197e-01 -5.90197265e-01 -8.95987570e-01
-4.16148081e-02 -4.87135172e-01 -3.98059279e-01 -2.66326934e-01
5.19480884e-01 -2.24058881e-01 9.20475051e-02 5.43798029e-01
-3.43237519e-01 2.99408674e-01 1.08359456e+00 -2.50435621e-01
-4.32510853e-01 7.92122364e-01 -1.00032580e+00 4.71424490e-01
7.50972688e-01 -3.47124130e-01 -3.33348542e-01 -2.15161920e-01
3.30167681e-01 3.68704438e-01 1.86222777e-01 -1.86759114e-01
1.07399531e-01 -5.39913952e-01 2.94584811e-01 -9.31623578e-01
4.98998940e-01 -8.06462944e-01 4.65869159e-01 5.21132588e-01
-6.40322983e-01 3.79520297e-01 6.66192994e-02 1.23815799e+00
-1.38925299e-01 1.90318078e-01 2.68464357e-01 -2.42033646e-01
-3.85101527e-01 2.65655160e-01 -5.79468191e-01 -7.25527167e-01
8.44908655e-01 1.90898865e-01 -5.37712753e-01 -7.60267138e-01
-1.49510086e+00 2.30380014e-01 1.18072197e-01 6.05377257e-01
1.36829056e-02 -1.50625539e+00 -4.74079698e-01 -6.30993843e-02
2.07050264e-01 -1.29320669e+00 4.11294281e-01 8.78501117e-01
2.20928952e-01 1.11465359e+00 1.61474869e-01 -3.73885602e-01
-8.74423265e-01 1.00116920e+00 -1.91072658e-01 -3.87798935e-01
-1.49063423e-01 5.40917337e-01 2.90434808e-02 2.10230157e-01
1.96067631e-01 -5.14949620e-01 4.69902239e-04 3.99009228e-01
5.06611466e-01 2.00839505e-01 -3.74127477e-02 -4.14888591e-01
-1.31292626e-01 1.62645698e-01 -5.86829484e-01 -4.21659440e-01
1.74551725e+00 -2.66832829e-01 2.94975162e-01 3.54081661e-01
1.12323201e+00 -2.74156239e-02 -1.39392281e+00 -3.34321439e-01
4.06770915e-01 -8.18600595e-01 6.92139149e-01 -1.04280317e+00
-1.14521825e+00 3.06461722e-01 8.51787508e-01 7.35467970e-01
9.24930930e-01 2.74545044e-01 8.39515388e-01 -2.23498240e-01
3.33881795e-01 -1.39355648e+00 -1.05069153e-01 -3.61714453e-01
6.82835758e-01 -1.33512950e+00 1.31428972e-01 -3.60924661e-01
-7.43162394e-01 5.94323575e-01 -1.39961600e-01 -2.31147349e-01
1.32313597e+00 -3.15208137e-01 -1.91043198e-01 -5.71012080e-01
-1.23595619e+00 -2.77000934e-01 3.07094902e-01 3.22133005e-01
3.88462931e-01 6.49664819e-01 -9.57822442e-01 9.87149954e-01
-3.94375056e-01 -2.50186056e-01 2.93266803e-01 8.96686554e-01
-6.20944798e-01 -1.46903861e+00 -8.55501592e-02 1.01465642e+00
-6.70481861e-01 -1.15409836e-01 -3.41524810e-01 7.45347261e-01
-4.38949883e-01 1.20239270e+00 5.93670845e-01 -4.17377144e-01
4.26216185e-01 -1.37349800e-03 2.12787628e-01 -3.47974658e-01
-3.23879272e-01 9.15998220e-01 4.30142015e-01 -7.53408372e-01
-3.02458584e-01 -9.67107117e-01 -9.24724936e-02 -4.58475947e-01
-5.11041999e-01 1.23843245e-01 1.50360155e+00 3.52206081e-01
7.08008945e-01 5.21350503e-01 1.14075375e+00 -3.46586585e-01
-9.21512365e-01 -9.10800934e-01 -7.57704496e-01 8.55821595e-02
-9.98439938e-02 -8.35879803e-01 -7.94325054e-01 -1.91195995e-01] | [9.636994361877441, 5.623012065887451] |
38ea765d-d7fd-47e6-9c5a-cbf31e806b71 | multi-angle-point-cloud-vae-unsupervised | 1907.12704 | null | https://arxiv.org/abs/1907.12704v1 | https://arxiv.org/pdf/1907.12704v1.pdf | Multi-Angle Point Cloud-VAE: Unsupervised Feature Learning for 3D Point Clouds from Multiple Angles by Joint Self-Reconstruction and Half-to-Half Prediction | Unsupervised feature learning for point clouds has been vital for large-scale point cloud understanding. Recent deep learning based methods depend on learning global geometry from self-reconstruction. However, these methods are still suffering from ineffective learning of local geometry, which significantly limits the discriminability of learned features. To resolve this issue, we propose MAP-VAE to enable the learning of global and local geometry by jointly leveraging global and local self-supervision. To enable effective local self-supervision, we introduce multi-angle analysis for point clouds. In a multi-angle scenario, we first split a point cloud into a front half and a back half from each angle, and then, train MAP-VAE to learn to predict a back half sequence from the corresponding front half sequence. MAP-VAE performs this half-to-half prediction using RNN to simultaneously learn each local geometry and the spatial relationship among them. In addition, MAP-VAE also learns global geometry via self-reconstruction, where we employ a variational constraint to facilitate novel shape generation. The outperforming results in four shape analysis tasks show that MAP-VAE can learn more discriminative global or local features than the state-of-the-art methods. | ['Matthias Zwicker', 'Yu-Shen Liu', 'Zhizhong Han', 'Xiyang Wang'] | 2019-07-30 | multi-angle-point-cloud-vae-unsupervised-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Han_Multi-Angle_Point_Cloud-VAE_Unsupervised_Feature_Learning_for_3D_Point_Clouds_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Han_Multi-Angle_Point_Cloud-VAE_Unsupervised_Feature_Learning_for_3D_Point_Clouds_ICCV_2019_paper.pdf | iccv-2019-10 | ['3d-point-cloud-linear-classification', 'unsupervised-3d-point-cloud-linear-evaluation'] | ['computer-vision', 'computer-vision'] | [-2.28890732e-01 -1.43499345e-01 -2.43756716e-04 -6.68664575e-01
-1.01291740e+00 -6.41493499e-01 4.99836832e-01 1.49446642e-02
-6.51797578e-02 2.54563808e-01 -1.19559921e-01 -3.18787992e-02
-8.93779099e-02 -1.00865889e+00 -1.14521575e+00 -7.23380327e-01
1.51552439e-01 8.54809821e-01 1.73176765e-01 -1.24214716e-01
1.54663354e-01 9.13907588e-01 -1.58114612e+00 2.37713277e-01
9.40438449e-01 8.24959517e-01 5.79156756e-01 4.03568923e-01
-3.28818440e-01 1.57837570e-01 -9.51556042e-02 -2.44723763e-02
3.92712682e-01 4.82871607e-02 -4.96434540e-01 5.53981140e-02
6.95409536e-01 -4.56854701e-01 -3.32343638e-01 6.85727298e-01
4.17084754e-01 1.52489483e-01 6.92690909e-01 -1.09629500e+00
-5.51213503e-01 2.12209914e-02 -4.71108943e-01 -5.76574095e-02
1.48127511e-01 2.41446525e-01 1.21456873e+00 -1.23438227e+00
7.20934808e-01 1.15928507e+00 5.87715209e-01 2.64307708e-01
-1.07852507e+00 -9.09587145e-01 1.56414911e-01 4.79732715e-02
-1.37276936e+00 -2.37213746e-01 1.25373960e+00 -6.04240119e-01
9.06870067e-01 2.10745377e-03 8.71347666e-01 8.81806731e-01
-2.16644742e-02 9.67863023e-01 6.75135732e-01 -8.12020227e-02
1.89628914e-01 -3.28875929e-01 -3.93988103e-01 8.50385547e-01
-3.03062648e-01 2.28583977e-01 -4.61704969e-01 -1.09977573e-01
1.20303190e+00 3.93961728e-01 -6.85516745e-02 -8.67510080e-01
-1.30165637e+00 9.13147628e-01 9.38382268e-01 8.36440548e-02
-3.96659672e-01 4.99829277e-02 1.48640960e-01 1.86991155e-01
5.58083057e-01 4.66317058e-01 -6.56854987e-01 1.97001435e-02
-9.46622133e-01 4.12690461e-01 3.30790609e-01 1.11786079e+00
1.19230795e+00 8.64596069e-02 1.12820983e-01 7.35048234e-01
5.40343642e-01 7.40156889e-01 1.85302228e-01 -8.00715864e-01
6.88673913e-01 7.55144656e-01 -1.86868049e-02 -9.61279809e-01
-4.69145715e-01 -6.58200562e-01 -7.97599316e-01 3.01679224e-01
6.55081794e-02 1.60372734e-01 -1.10180748e+00 1.43410754e+00
5.17379105e-01 4.16460603e-01 -2.31300414e-01 9.75865126e-01
7.76287198e-01 7.90350735e-01 -4.27547395e-01 -1.63902305e-02
7.11606622e-01 -9.82373893e-01 -5.84461428e-02 -4.87906411e-02
5.91581643e-01 -5.73289335e-01 1.00778568e+00 3.18137348e-01
-7.62867391e-01 -7.20237136e-01 -9.07074094e-01 -2.37113640e-01
-1.01226337e-01 1.80499867e-01 3.60810131e-01 -1.09152503e-01
-7.81428099e-01 6.01533234e-01 -1.06198835e+00 8.38544890e-02
7.33474672e-01 3.57869536e-01 -5.32010436e-01 -2.96177715e-01
-5.94626307e-01 3.65898430e-01 -4.32225615e-02 -3.46238501e-02
-6.70924783e-01 -9.31712270e-01 -1.05992734e+00 -1.87343825e-02
9.42544192e-02 -9.21054363e-01 1.00301480e+00 -6.08959377e-01
-1.52100587e+00 7.81358719e-01 -3.76336485e-01 -1.07918136e-01
3.93029273e-01 -3.02795649e-01 6.33331239e-02 5.92664853e-02
3.67166251e-01 8.73021305e-01 1.06956708e+00 -1.55317676e+00
-3.89331341e-01 -6.01303279e-01 -3.13035011e-01 3.07216674e-01
1.04028508e-01 -4.86918807e-01 -4.75625485e-01 -6.01637840e-01
6.53730869e-01 -9.67214108e-01 -1.72518268e-01 2.82418638e-01
-1.65876448e-01 -4.04312164e-01 1.10558319e+00 -4.05880243e-01
6.44013047e-01 -2.28004527e+00 3.09505612e-01 1.98233783e-01
3.14003468e-01 1.18919119e-01 -3.46427441e-01 2.98986375e-01
-7.13554546e-02 -1.34493411e-01 -2.71859676e-01 -6.89935565e-01
-1.35409653e-01 4.60901409e-01 -6.01781487e-01 5.28393805e-01
4.13481504e-01 1.11300755e+00 -9.55854774e-01 -3.69475693e-01
5.77487767e-01 5.90338171e-01 -7.62496769e-01 3.19735885e-01
-4.05923963e-01 9.02386427e-01 -6.29532099e-01 6.19751573e-01
9.07285035e-01 -3.89385819e-01 -3.44825536e-01 -3.63707334e-01
-2.55469978e-01 7.19116032e-02 -8.97287130e-01 2.07345462e+00
-5.39862156e-01 4.95841205e-01 -1.36186853e-01 -9.84799981e-01
1.30213463e+00 8.61733258e-02 7.58679926e-01 -7.26909518e-01
-1.35251746e-01 2.34862700e-01 -1.78018406e-01 -2.91207761e-01
1.23234190e-01 -3.30433905e-01 2.82146215e-01 1.23222351e-01
6.48012012e-02 -6.48817003e-01 -5.06320477e-01 3.23181264e-02
7.86787450e-01 5.06830394e-01 -4.34595812e-03 1.16093524e-01
3.24557692e-01 -1.91476509e-01 7.27462888e-01 3.32668871e-01
6.19024709e-02 1.07575810e+00 -9.37646180e-02 -6.96508169e-01
-1.35196805e+00 -1.22058487e+00 -9.54869092e-02 8.35253060e-01
2.21406102e-01 -4.74776953e-01 -2.63079882e-01 -7.66894758e-01
3.79849315e-01 5.85874498e-01 -3.60512555e-01 -5.75955622e-02
-7.20882416e-01 -1.36596590e-01 -1.66538954e-02 7.35631287e-01
3.58912379e-01 -1.04001355e+00 -4.05045569e-01 2.38212019e-01
2.55236821e-03 -1.08946598e+00 -4.80091840e-01 1.91227496e-02
-1.18026698e+00 -7.33694077e-01 -5.45320868e-01 -6.96146131e-01
6.59076869e-01 5.44032097e-01 1.02780640e+00 -2.08901148e-02
-8.21567848e-02 2.49322161e-01 -3.72922063e-01 -3.52914006e-01
4.22044247e-02 9.70146358e-02 5.54882083e-03 1.09340744e-02
1.11229569e-01 -1.07773435e+00 -6.03611648e-01 3.29843700e-01
-6.07245564e-01 1.15289636e-01 7.21209824e-01 7.66392648e-01
1.15346217e+00 -2.63607085e-01 4.22632396e-01 -5.50076008e-01
-5.84649108e-02 -4.87439305e-01 -5.59330523e-01 -1.20079555e-02
-2.56659985e-01 2.07008585e-01 6.77603841e-01 -1.14324965e-01
-6.88215494e-01 5.35507798e-01 -4.28090125e-01 -1.15544462e+00
-1.56334430e-01 3.70100021e-01 -2.92722255e-01 -2.12156802e-01
3.79745364e-01 5.51441789e-01 4.64201607e-02 -6.07236922e-01
4.78394479e-01 3.73796493e-01 4.07311231e-01 -6.58413112e-01
1.07432425e+00 6.01603746e-01 6.10586861e-03 -1.03508186e+00
-8.57992768e-01 -6.26605749e-01 -1.06636941e+00 -2.83060852e-03
8.64491463e-01 -1.03762865e+00 -6.04382515e-01 4.92902428e-01
-1.19953859e+00 -3.08272064e-01 -3.92826557e-01 3.36785436e-01
-7.10043788e-01 3.26370806e-01 -2.02258825e-01 -4.63811815e-01
-4.03222382e-01 -1.13637805e+00 1.73383164e+00 2.59069391e-02
2.32099727e-01 -8.32490742e-01 1.70575082e-01 3.79471451e-01
6.65464550e-02 2.78410554e-01 7.56992996e-01 -4.78113472e-01
-8.57162416e-01 -1.43553883e-01 -2.81110018e-01 3.04310262e-01
3.17763120e-01 1.27638131e-01 -7.80124545e-01 -5.08228362e-01
-2.46342849e-02 -3.99315596e-01 7.23331690e-01 2.71686852e-01
1.27662921e+00 -7.37236962e-02 -4.38889414e-01 1.20303059e+00
1.28224468e+00 -1.52373239e-01 3.89751941e-01 2.99365342e-01
1.18159676e+00 2.61317402e-01 6.45723581e-01 5.07252634e-01
6.71585441e-01 6.82144761e-01 7.92165458e-01 -4.69973013e-02
1.02765970e-01 -6.32892609e-01 9.93458927e-02 1.11884582e+00
-1.21545397e-01 1.48871571e-01 -1.04938829e+00 5.35925686e-01
-1.79942000e+00 -6.64918721e-01 1.10457866e-02 2.08480930e+00
5.67977011e-01 -2.95330957e-02 2.31701769e-02 -2.53241211e-01
3.69228184e-01 3.37320238e-01 -8.95552516e-01 1.56783462e-01
-1.80099487e-01 2.40773901e-01 2.02833891e-01 3.80386412e-01
-1.09372699e+00 1.16477847e+00 5.52850914e+00 8.08371782e-01
-1.50947404e+00 -4.54677865e-02 2.03209534e-01 2.57205646e-02
-6.80003405e-01 1.08897768e-01 -6.22154415e-01 3.05148125e-01
3.16718638e-01 2.93278486e-01 1.03413701e-01 1.00239170e+00
2.55815852e-02 3.82688224e-01 -1.12940919e+00 1.28508317e+00
-9.11263153e-02 -1.47153568e+00 2.94054627e-01 2.57815242e-01
8.08741987e-01 5.46530187e-01 -5.38431816e-02 2.87484437e-01
1.64189026e-01 -9.51179922e-01 6.29502296e-01 7.13726938e-01
8.64858687e-01 -7.89481997e-01 4.31669354e-01 7.59599447e-01
-1.46324539e+00 1.35095626e-01 -5.73789358e-01 9.89133716e-02
9.95710865e-02 6.40166879e-01 -7.42778182e-01 7.88996875e-01
5.51514506e-01 1.08928359e+00 -3.28394622e-01 9.27118599e-01
-2.87664920e-01 3.73642087e-01 -6.33416533e-01 3.69115770e-01
3.17840129e-01 -5.32077372e-01 7.59367883e-01 9.31103289e-01
5.27923942e-01 1.41150936e-01 4.35277969e-01 1.02931833e+00
-2.80693583e-02 -5.25695179e-03 -8.04421306e-01 2.28614718e-01
6.26721084e-01 1.13291717e+00 -4.24507946e-01 -1.48196474e-01
-4.67822075e-01 1.05471396e+00 6.07899129e-01 2.29439273e-01
-5.14199495e-01 -7.04638585e-02 7.70144939e-01 2.07099482e-01
5.57847261e-01 -6.99027181e-01 -2.86454082e-01 -1.29594564e+00
1.99439928e-01 -5.60894549e-01 3.89893949e-02 -9.96677756e-01
-1.29821181e+00 5.15791476e-01 -1.39545366e-01 -1.53954780e+00
-2.08088592e-01 -4.77279633e-01 -7.95547783e-01 7.53812253e-01
-1.46627283e+00 -1.66909707e+00 -4.08881873e-01 6.36236668e-01
6.87423766e-01 -2.89793640e-01 5.80413699e-01 3.12895281e-03
-1.69392303e-01 5.81261933e-01 2.51523703e-01 1.38241112e-01
6.39978707e-01 -1.30005336e+00 7.29066432e-01 4.26387548e-01
4.87053335e-01 5.50147951e-01 3.80935192e-01 -7.07995117e-01
-1.57293928e+00 -1.24121678e+00 4.04605418e-01 -4.84218001e-01
2.58910745e-01 -5.35683632e-01 -1.18098068e+00 7.36491024e-01
-4.03209776e-01 6.11218274e-01 3.17490697e-01 -1.70046203e-02
-5.51058650e-01 -1.33361042e-01 -1.01252151e+00 3.32061023e-01
1.28297269e+00 -7.62852609e-01 -5.95626056e-01 3.72058481e-01
9.49922800e-01 -5.51382065e-01 -8.79449010e-01 6.55638754e-01
4.19858694e-01 -9.74499881e-01 1.11303365e+00 -5.17330468e-01
5.98336697e-01 -3.34233195e-01 -3.35027546e-01 -1.44961488e+00
-5.34632742e-01 -3.41528028e-01 -1.33389086e-01 1.03723061e+00
1.77490443e-01 -5.86078405e-01 1.08085692e+00 9.41347182e-02
-5.27415514e-01 -1.22869098e+00 -9.27823782e-01 -7.47592866e-01
4.57515836e-01 -5.44288814e-01 7.94045269e-01 1.04388738e+00
-5.72337508e-01 3.59534353e-01 -1.56574249e-01 4.17658001e-01
6.53413355e-01 6.75704956e-01 1.08730721e+00 -1.39536703e+00
-1.65530041e-01 -3.15449238e-01 -6.68812156e-01 -1.66263032e+00
2.72632748e-01 -1.09419835e+00 -1.33479312e-02 -1.55886436e+00
8.62262473e-02 -6.79148257e-01 -1.09905517e-02 4.59188223e-01
-6.16404265e-02 1.03573404e-01 2.76858211e-01 5.04854441e-01
-4.14494097e-01 1.11961424e+00 1.61556411e+00 -1.07472390e-01
-2.68187344e-01 1.17952943e-01 -2.81045288e-01 7.02929437e-01
5.40433764e-01 -3.84638935e-01 -3.58637154e-01 -6.01882756e-01
3.51171434e-01 4.46824767e-02 4.53155428e-01 -9.30921972e-01
3.86822075e-01 -3.71776164e-01 5.23105621e-01 -1.13783157e+00
5.18894851e-01 -8.10662746e-01 -8.83754045e-02 7.92005006e-03
1.64265916e-01 -5.43843322e-02 1.02617964e-01 7.21921086e-01
-3.76418263e-01 1.34179756e-01 6.09204650e-01 -1.12751693e-01
-5.31593144e-01 9.05777216e-01 3.72006088e-01 -1.08860746e-01
8.03645909e-01 -3.44011962e-01 1.41062424e-01 -3.04148138e-01
-7.81761467e-01 3.56940031e-01 7.12277055e-01 5.01588643e-01
1.06688952e+00 -1.49469101e+00 -7.73639858e-01 6.98460817e-01
3.31636131e-01 1.09897327e+00 4.19997990e-01 5.46787024e-01
-5.04519701e-01 2.05893561e-01 -1.45295084e-01 -1.23352885e+00
-1.09749150e+00 4.88944501e-01 2.84072191e-01 -4.90493551e-02
-1.05976760e+00 1.07569945e+00 6.35206938e-01 -9.82977092e-01
-1.74766079e-01 -2.83382773e-01 6.79577589e-02 -2.23311841e-01
2.95622572e-02 1.02564454e-01 4.55659032e-02 -8.27321351e-01
-2.96912730e-01 1.24214888e+00 -2.27965377e-02 4.44390550e-02
1.72105122e+00 5.56890368e-02 5.48426341e-03 5.35062194e-01
1.56961930e+00 2.27730677e-01 -1.59944403e+00 -5.28902888e-01
-2.83882618e-01 -6.54386103e-01 8.51891041e-02 -3.18148226e-01
-1.20757186e+00 1.15847743e+00 3.42415929e-01 -4.50820118e-01
8.00848365e-01 2.17653662e-01 9.59941626e-01 4.29949522e-01
4.56929684e-01 -6.56691849e-01 3.21845114e-01 8.21591496e-01
1.24717414e+00 -1.47929299e+00 -2.09380221e-02 -4.82682556e-01
-5.68143010e-01 1.29380131e+00 6.66618764e-01 -4.42977399e-01
8.16414952e-01 3.25191841e-02 -6.15287051e-02 -4.59988356e-01
-5.96251249e-01 -2.30745710e-02 6.61872268e-01 5.18489599e-01
1.71012536e-01 1.29482582e-01 4.89305586e-01 4.24032003e-01
-5.60461044e-01 -2.75210917e-01 -2.23662015e-02 7.94536114e-01
-3.55483353e-01 -1.10552573e+00 -2.04360336e-01 4.86625791e-01
2.22887337e-01 1.32015646e-01 -3.37697893e-01 6.30902946e-01
2.28491500e-01 2.93799579e-01 3.70905638e-01 -6.58392310e-01
3.90942872e-01 -5.34226857e-02 4.76545691e-01 -7.55866945e-01
-1.53575897e-01 1.57420725e-01 -4.81802791e-01 -6.39402807e-01
-1.10830784e-01 -8.19865048e-01 -1.46221733e+00 -6.13667071e-02
-2.31727228e-01 -3.80717292e-02 7.04134762e-01 1.10140288e+00
6.06233597e-01 3.66203398e-01 8.96191418e-01 -1.31156337e+00
-5.57289183e-01 -8.45328987e-01 -4.98209268e-01 5.32446563e-01
6.97010815e-01 -8.37739527e-01 -2.96899945e-01 -1.95190400e-01] | [8.198442459106445, -3.4305944442749023] |
660bb7ca-d6ed-485d-9ae5-3d3509659f4d | how-does-beam-search-improve-span-level | 2212.10767 | null | https://arxiv.org/abs/2212.10767v1 | https://arxiv.org/pdf/2212.10767v1.pdf | How Does Beam Search improve Span-Level Confidence Estimation in Generative Sequence Labeling? | Text-to-text generation models have increasingly become the go-to solution for a wide variety of sequence labeling tasks (e.g., entity extraction and dialog slot filling). While most research has focused on the labeling accuracy, a key aspect -- of vital practical importance -- has slipped through the cracks: understanding model confidence. More specifically, we lack a principled understanding of how to reliably gauge the confidence of a model in its predictions for each labeled span. This paper aims to provide some empirical insights on estimating model confidence for generative sequence labeling. Most notably, we find that simply using the decoder's output probabilities is not the best in realizing well-calibrated confidence estimates. As verified over six public datasets of different tasks, we show that our proposed approach -- which leverages statistics from top-$k$ predictions by a beam search -- significantly reduces calibration errors of the predictions of a generative sequence labeling model. | ['Karthik Raman', 'Iftekhar Naim', 'Kazuma Hashimoto'] | 2022-12-21 | null | null | null | null | ['slot-filling'] | ['natural-language-processing'] | [ 3.14928204e-01 5.94826937e-01 -4.69502330e-01 -6.50074184e-01
-1.43571138e+00 -7.85559237e-01 3.16498995e-01 2.21570238e-01
-2.16425329e-01 1.03923619e+00 3.50625455e-01 -7.58979321e-01
3.48653406e-01 -4.22101766e-01 -6.62240386e-01 -2.27835640e-01
4.43639308e-01 8.58451307e-01 -4.46882695e-02 -5.15953861e-02
3.89007479e-01 -2.06494346e-01 -8.55323315e-01 1.25855565e-01
9.46272373e-01 8.01601827e-01 1.20061636e-01 6.01479888e-01
-3.58601689e-01 8.27699900e-01 -5.97561181e-01 -1.03410625e+00
-1.42062709e-01 -7.05490172e-01 -9.55271900e-01 -8.53537321e-02
8.33972916e-02 -1.37963459e-01 4.67696786e-02 1.08014345e+00
1.24235734e-01 -2.03756794e-01 8.84024858e-01 -1.10851347e+00
-2.90964514e-01 1.15892613e+00 -4.15691346e-01 2.93613765e-02
2.86079913e-01 4.48827334e-02 1.51316881e+00 -5.31407595e-01
7.12761998e-01 1.08455086e+00 7.21492410e-01 5.66471457e-01
-1.25648415e+00 -7.14734316e-01 1.12201288e-01 -2.03260556e-01
-1.26250839e+00 -6.45864248e-01 2.66652197e-01 -7.35795438e-01
9.93248463e-01 1.47916958e-01 1.39313474e-01 1.01636147e+00
1.54928593e-02 7.51017869e-01 9.84011114e-01 -6.81079745e-01
1.69856295e-01 3.52596879e-01 3.18609804e-01 8.57128084e-01
4.93412226e-01 -9.76518542e-02 -8.32698643e-01 -2.61120081e-01
4.18090969e-01 -7.54034817e-01 -1.68168098e-01 3.61098498e-02
-9.37041759e-01 9.74843383e-01 -3.35759252e-01 1.53619573e-01
2.42610648e-02 2.19770357e-01 3.83984447e-01 -8.07481185e-02
6.06370628e-01 6.22924149e-01 -6.71216249e-01 -7.09579945e-01
-1.14128315e+00 3.36479396e-01 1.13832021e+00 1.29544497e+00
7.16255426e-01 -8.11675191e-02 -2.17322975e-01 6.86174035e-01
4.01726246e-01 3.26925486e-01 2.14781225e-01 -7.81421602e-01
7.06074595e-01 4.03115273e-01 4.94559437e-01 -5.15537202e-01
-3.22409332e-01 -2.06641227e-01 -3.41371983e-01 -2.55138755e-01
8.46010804e-01 -6.16807699e-01 -8.92989218e-01 2.12454152e+00
1.77150831e-01 1.62630409e-01 -7.76082501e-02 6.32841825e-01
2.90802896e-01 3.92874211e-01 4.84027475e-01 -2.85748124e-01
1.43435121e+00 -7.52652287e-01 -6.55103207e-01 -6.32826149e-01
8.25424671e-01 -8.80833089e-01 9.30912793e-01 1.62483945e-01
-8.74389350e-01 -2.97680497e-01 -1.02435327e+00 -3.40906344e-02
-1.35827707e-02 1.17876135e-01 8.21957409e-01 1.07543671e+00
-8.00647140e-01 4.89437431e-01 -7.43873179e-01 -4.65352297e-01
-1.64838936e-02 2.71716323e-02 -1.40587837e-01 1.51907373e-02
-1.14363837e+00 9.72817183e-01 5.16707301e-01 -2.00054526e-01
-6.05119348e-01 -4.45174575e-01 -7.62234747e-01 8.79845098e-02
5.05120635e-01 -6.14333391e-01 1.89600539e+00 -3.94914567e-01
-1.38538373e+00 5.94660223e-01 -6.26260161e-01 -5.63867152e-01
5.48693955e-01 -4.45587128e-01 -2.74330229e-01 -4.54348385e-01
2.94338554e-01 5.59491932e-01 6.56358302e-01 -1.01458335e+00
-8.03811789e-01 -7.74512962e-02 -2.50013530e-01 1.77949280e-01
3.54006849e-02 5.06485440e-02 -5.17724872e-01 -3.94779474e-01
2.62954712e-01 -1.04161882e+00 -3.18916768e-01 -6.66162550e-01
-8.78180563e-01 -2.69882977e-01 1.59662962e-01 -7.96025336e-01
1.63639748e+00 -1.77227652e+00 -1.62988931e-01 1.50157899e-01
1.09530352e-01 -1.60698611e-02 2.75098085e-01 6.36723101e-01
1.70752436e-01 4.55904990e-01 -1.78431347e-01 -5.43174922e-01
2.87544966e-01 -7.63908252e-02 -7.08491266e-01 8.30484405e-02
2.63012290e-01 1.02110469e+00 -7.69997180e-01 -6.15944982e-01
-2.77669001e-02 1.64353281e-01 -4.83317256e-01 1.69100314e-01
-6.93837225e-01 3.98901522e-01 -4.30264771e-01 4.76689786e-01
2.05780447e-01 -7.24683166e-01 7.20582962e-01 6.85546175e-02
3.93631086e-02 7.38830149e-01 -9.72558439e-01 1.57534599e+00
-1.65115550e-01 5.49652398e-01 -3.32697809e-01 -6.45325184e-01
7.31419861e-01 3.38205278e-01 1.01611316e-01 -3.24521869e-01
1.64217666e-01 3.22084606e-01 -2.72576581e-04 -2.77653426e-01
9.12294745e-01 -2.01483771e-01 -4.35510725e-01 4.65449572e-01
1.29162863e-01 -9.76879522e-02 2.12202758e-01 2.73534149e-01
1.00932312e+00 2.71710008e-01 4.61971402e-01 -8.28777105e-02
6.02823794e-02 3.38815868e-01 7.65788376e-01 1.02626276e+00
-8.27172697e-02 5.07943094e-01 6.71862245e-01 7.63490573e-02
-1.27778077e+00 -9.18357193e-01 6.35828590e-03 1.13344657e+00
-1.49100631e-01 -6.65408552e-01 -1.06539452e+00 -7.02545822e-01
-2.00055796e-03 1.16045094e+00 -3.88022900e-01 1.45268813e-01
-5.30092657e-01 -7.50048578e-01 6.64323568e-01 5.37471950e-01
7.45735094e-02 -6.93500698e-01 -2.49794707e-01 4.52079296e-01
-5.83205521e-01 -1.37480760e+00 -6.13252580e-01 3.66164833e-01
-7.03004360e-01 -9.65040743e-01 -2.55858660e-01 -4.90667790e-01
4.34714258e-01 -9.34506804e-02 1.41032267e+00 -1.65081024e-01
9.81501341e-02 1.51177883e-01 -3.58489841e-01 -3.59726042e-01
-8.30442667e-01 4.81137812e-01 -5.97399101e-02 -5.33102632e-01
6.91532433e-01 -2.44351983e-01 -2.48987213e-01 3.07009876e-01
-3.59979242e-01 2.17424840e-01 4.83914435e-01 7.33515620e-01
3.02261353e-01 -3.38775545e-01 8.03567350e-01 -1.48062229e+00
7.90227532e-01 -4.29147899e-01 -5.37530541e-01 5.38390875e-01
-1.04293323e+00 3.12579870e-01 3.46283406e-01 -6.01358972e-02
-1.01924717e+00 -3.07054594e-02 -4.49470252e-01 6.52625784e-02
-1.03932813e-01 6.11252069e-01 -1.02547184e-02 5.24572432e-01
6.13304377e-01 1.54369354e-01 -2.99457759e-01 -5.15662849e-01
5.62294781e-01 6.68083608e-01 5.98319769e-01 -8.07682395e-01
5.43783665e-01 -3.00226770e-02 -3.66366506e-01 -5.40510476e-01
-1.14583051e+00 -4.80056554e-01 -3.39018047e-01 5.01524769e-02
9.54688609e-01 -1.00545168e+00 -6.86810136e-01 1.82795510e-01
-1.06453347e+00 -4.92587686e-01 2.58387662e-02 3.57586831e-01
-5.82748234e-01 5.01738906e-01 -8.78385603e-01 -1.14290643e+00
-3.39096338e-01 -1.09427118e+00 9.51473057e-01 3.99339080e-01
-7.61961043e-01 -7.96570957e-01 1.94295496e-01 4.73913640e-01
2.77422965e-01 -2.65837938e-01 1.13962197e+00 -9.47456777e-01
-5.98352194e-01 -1.11237898e-01 -1.11595578e-01 -1.80300162e-03
-8.34720060e-02 2.51134466e-02 -1.16840839e+00 -4.28001992e-02
-2.01835468e-01 -3.39466304e-01 6.89553082e-01 4.67345715e-01
8.27497303e-01 -2.38988116e-01 -5.46027899e-01 1.50913119e-01
1.16820741e+00 2.27673143e-01 4.77389008e-01 -3.15835401e-02
2.52681702e-01 4.37550157e-01 7.93723524e-01 5.11802256e-01
5.26440024e-01 6.79165781e-01 -1.74397945e-01 4.32063222e-01
5.80685884e-02 -8.53335023e-01 8.59435089e-03 7.46142745e-01
2.25978732e-01 -6.48461938e-01 -9.07856643e-01 4.13911909e-01
-1.86861718e+00 -9.06154335e-01 1.33778661e-01 2.17365909e+00
1.04509294e+00 5.67771256e-01 7.68925622e-02 -1.05054587e-01
7.90860415e-01 8.70252103e-02 -5.08594692e-01 -1.26561105e-01
1.12087131e-01 1.63917810e-01 5.31720042e-01 7.72854924e-01
-9.11628246e-01 1.19016874e+00 6.90752411e+00 7.99005151e-01
-7.59831965e-01 5.08399755e-02 1.00209939e+00 4.44237709e-01
-4.99550998e-01 2.23770589e-01 -1.40802288e+00 5.81684411e-01
1.26569617e+00 -1.42858043e-01 1.40889913e-01 1.08957040e+00
1.34694234e-01 -4.21917081e-01 -1.28151548e+00 7.07532704e-01
-2.23093852e-01 -1.46570873e+00 -2.46692017e-01 -9.55166668e-02
7.06352174e-01 -1.27929106e-01 -1.63092818e-02 5.90501547e-01
8.43480647e-01 -1.07913029e+00 9.72189724e-01 3.95148844e-01
9.60557818e-01 -5.40068448e-01 5.26304305e-01 4.85560566e-01
-7.88855433e-01 3.18241835e-01 -2.11750656e-01 8.25174451e-02
5.54363906e-01 6.92977905e-01 -1.57236052e+00 1.18460543e-01
2.36810282e-01 5.18183559e-02 -2.82565534e-01 7.72647083e-01
-2.75608838e-01 1.13920557e+00 -2.14352697e-01 -2.20640242e-01
-5.45005836e-02 -1.00924961e-01 1.07211925e-01 1.53900230e+00
2.70343900e-01 1.18930586e-01 -4.15765829e-02 7.86653459e-01
-1.57500893e-01 -6.35698214e-02 -4.21788990e-01 -5.87989807e-01
8.55111122e-01 1.14089835e+00 -7.58894444e-01 -3.69163990e-01
-3.87767583e-01 9.30799305e-01 5.06516576e-01 2.09996998e-01
-8.66821766e-01 -7.50865638e-02 7.25235879e-01 -1.60217136e-01
3.02014232e-01 -4.52220798e-01 -6.92157447e-01 -1.06297207e+00
-1.74772084e-01 -7.99359918e-01 2.76201785e-01 -6.23655081e-01
-1.17210805e+00 5.39907217e-01 -1.03798337e-01 -6.72765613e-01
-9.40722585e-01 -4.24677163e-01 -2.52527088e-01 1.09786844e+00
-1.31239748e+00 -8.23098600e-01 -4.37784642e-02 -7.62000754e-02
6.18293941e-01 1.61019266e-01 9.54602540e-01 1.75275907e-01
-5.62091053e-01 8.85283113e-01 5.97267561e-02 3.10066611e-01
7.25880265e-01 -1.18188834e+00 1.00293362e+00 9.66674566e-01
3.83398235e-01 7.36230731e-01 1.12887168e+00 -8.93605292e-01
-1.00939572e+00 -8.00264537e-01 1.33339107e+00 -8.84482443e-01
6.31052613e-01 -4.50236827e-01 -7.56374002e-01 9.40144181e-01
3.52611877e-02 -6.17233455e-01 7.87985146e-01 4.38548923e-01
-4.75582093e-01 4.17277932e-01 -9.00724411e-01 5.84617078e-01
9.86886263e-01 -7.84334064e-01 -3.47827375e-01 1.82871521e-01
9.73154664e-01 -7.32134461e-01 -7.50803351e-01 3.18078250e-01
6.62547410e-01 -9.44101453e-01 5.76698720e-01 -7.99705446e-01
3.92054588e-01 -1.35480776e-01 -4.16574508e-01 -1.16054940e+00
-1.02118745e-01 -7.02441096e-01 -5.04060909e-02 1.36821318e+00
1.08425200e+00 -4.69536155e-01 1.18808675e+00 1.21060467e+00
-1.83872953e-01 -5.82244575e-01 -7.76800275e-01 -4.68026131e-01
-5.94918989e-02 -7.85875320e-01 5.04217267e-01 8.62848699e-01
3.10006946e-01 6.31204486e-01 -4.35830921e-01 8.68435204e-02
5.03478289e-01 -9.60838515e-03 7.97620893e-01 -9.56294715e-01
-5.80363631e-01 -1.44574285e-01 2.18601022e-02 -1.38294327e+00
1.68870226e-01 -5.74704587e-01 4.41436529e-01 -1.35116863e+00
3.24655801e-01 -7.04036772e-01 -2.69391923e-03 3.44239175e-01
-5.45368731e-01 1.87075010e-03 1.45271480e-01 7.87640139e-02
-6.84232354e-01 7.03056827e-02 7.17659056e-01 3.11398506e-01
1.23363696e-01 1.70950040e-01 -1.06658936e+00 5.06330013e-01
9.00379539e-01 -5.52147388e-01 -4.19510245e-01 -2.97883332e-01
5.13323128e-01 6.36915028e-01 -7.52174258e-02 -8.82222414e-01
2.61125773e-01 -2.85068125e-01 4.19241041e-02 -5.44193208e-01
2.30087474e-01 -3.31345111e-01 2.88404495e-01 2.08031029e-01
-6.85781777e-01 1.14309199e-01 1.00570716e-01 8.43200624e-01
1.00254923e-01 -5.78859687e-01 4.78928268e-01 -1.06801689e-01
-7.15135694e-01 -6.73525631e-02 -2.12166682e-01 6.45070970e-01
5.64247549e-01 4.46442794e-03 -6.18531823e-01 -5.95126331e-01
-5.45162499e-01 1.13165844e-02 5.47650039e-01 2.62060583e-01
-2.66611651e-02 -8.12029719e-01 -5.54883838e-01 -2.20090412e-02
1.07521318e-01 -1.63253263e-01 -4.44570445e-02 5.09571373e-01
-1.46054447e-01 9.36369061e-01 4.39722240e-01 -4.92296070e-01
-9.45690811e-01 7.04579279e-02 2.17975974e-01 -6.02901161e-01
-2.60322213e-01 1.15511453e+00 -6.63714763e-03 -2.89046526e-01
1.58887178e-01 -4.13914412e-01 3.77129734e-01 -2.73195475e-01
3.42440158e-01 7.64226988e-02 7.41860867e-02 -3.15899223e-01
-1.37542039e-01 1.93532147e-02 -2.80654997e-01 -5.51619411e-01
9.14033294e-01 -2.61914581e-01 3.37449998e-01 4.98898119e-01
7.87380397e-01 1.87995747e-01 -1.35282564e+00 -2.59773284e-01
5.09889603e-01 -3.18658441e-01 -4.36521262e-01 -1.10045969e+00
-3.03283602e-01 6.97023273e-01 9.20949802e-02 2.16442376e-01
4.31536466e-01 1.04385316e-01 7.60324836e-01 3.64826441e-01
6.18808508e-01 -1.04047000e+00 -1.75353944e-01 6.65328085e-01
1.80776253e-01 -1.45271158e+00 -2.09808156e-01 -4.12860870e-01
-9.75711048e-01 7.50976384e-01 6.28957331e-01 4.66768444e-01
3.29454631e-01 3.18836004e-01 -1.72635168e-02 1.86670776e-02
-9.14977372e-01 -9.83973872e-03 1.41632920e-02 4.07295942e-01
8.42836738e-01 3.51349682e-01 -2.51800448e-01 7.25352168e-01
-5.47561347e-01 -1.27784595e-01 3.79014760e-01 6.01989508e-01
-7.88540661e-01 -1.26035249e+00 1.05251810e-02 7.48706639e-01
-6.62951052e-01 -3.64032865e-01 -1.70662552e-01 5.76875746e-01
-2.31899530e-01 1.25174677e+00 -1.86537668e-01 -5.54671288e-01
-7.12194815e-02 6.98949218e-01 3.92341167e-01 -8.61522079e-01
-3.48195195e-01 -9.90044791e-03 6.91472590e-01 -2.44018674e-01
-1.06283471e-01 -7.75920331e-01 -1.16393447e+00 -4.97635216e-01
-5.85317969e-01 4.78502065e-01 7.51973271e-01 9.96299088e-01
2.46712431e-01 1.92218676e-01 1.75129458e-01 -2.76342362e-01
-9.14672673e-01 -1.15990007e+00 -4.95361805e-01 1.61700413e-01
-1.16044618e-01 -4.38763916e-01 -1.79186195e-01 6.05806261e-02] | [10.838382720947266, 8.777430534362793] |
2bac0234-6755-4d81-8b02-6741206abb3d | integration-of-automatic-sentence | null | null | https://aclanthology.org/2020.lt4hala-1.8 | https://aclanthology.org/2020.lt4hala-1.8.pdf | Integration of Automatic Sentence Segmentation and Lexical Analysis of Ancient Chinese based on BiLSTM-CRF Model | The basic tasks of ancient Chinese information processing include automatic sentence segmentation, word segmentation, part-of-speech tagging and named entity recognition. Tasks such as lexical analysis need to be based on sentence segmentation because of the reason that a plenty of ancient books are not punctuated. However, step-by-step processing is prone to cause multi-level diffusion of errors. This paper designs and implements an integrated annotation system of sentence segmentation and lexical analysis. The BiLSTM-CRF neural network model is used to verify the generalization ability and the effect of sentence segmentation and lexical analysis on different label levels on four cross-age test sets. Research shows that the integration method adopted in ancient Chinese improves the F1-score of sentence segmentation, word segmentation and part of speech tagging. Based on the experimental results of each test set, the F1-score of sentence segmentation reached 78.95, with an average increase of 3.5{\%}; the F1-score of word segmentation reached 85.73{\%}, with an average increase of 0.18{\%}; and the F1-score of part-of-speech tagging reached 72.65, with an average increase of 0.35{\%}. | ['Changwei Xu', 'Liming Xiao', 'Ning Cheng', 'Xingyue Hao', 'Sijia Ge', 'Minxuan Feng', 'Bin Li'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['lexical-analysis'] | ['natural-language-processing'] | [-1.50969341e-01 1.18229106e-01 1.32156178e-01 -4.88931447e-01
-5.80148995e-01 -4.83823150e-01 3.69674573e-03 3.59586328e-01
-1.03227043e+00 8.84847462e-01 4.81956750e-02 -5.10562181e-01
2.93752968e-01 -8.08134198e-01 -3.89744520e-01 -4.76463854e-01
2.65405774e-01 2.72364229e-01 4.77624774e-01 -1.58756375e-01
5.58281600e-01 6.53095618e-02 -1.16519070e+00 3.07832599e-01
1.09874439e+00 4.90474671e-01 5.36481559e-01 5.49622595e-01
-8.41714740e-01 3.98750097e-01 -1.18825352e+00 -4.11103487e-01
-2.39467710e-01 -5.71352482e-01 -1.01697338e+00 7.11205378e-02
-2.34256268e-01 2.52010617e-02 2.65832782e-01 1.37179804e+00
4.63158488e-01 1.57605305e-01 5.07176340e-01 -4.68372852e-01
-6.87584519e-01 1.12635338e+00 -3.58636707e-01 3.24967623e-01
2.22000733e-01 -3.15703213e-01 8.00601840e-01 -5.69306612e-01
4.80355322e-01 8.94240499e-01 4.83516544e-01 6.15243137e-01
-8.31868529e-01 -7.57593334e-01 1.42468676e-01 -2.78096884e-01
-1.27751291e+00 -3.34272231e-03 4.06466067e-01 -3.80174577e-01
1.13311172e+00 -9.32359621e-02 5.44980049e-01 4.65181321e-01
4.70474541e-01 7.27120697e-01 1.21296978e+00 -8.45182300e-01
9.15284380e-02 2.18403205e-01 6.65756881e-01 4.18304086e-01
2.59821981e-01 -5.13866365e-01 -1.21883288e-01 4.34230328e-01
7.44218111e-01 -3.08985680e-01 1.50600821e-01 9.83468056e-01
-8.87287199e-01 8.37970316e-01 -8.20630565e-02 1.13831365e+00
-2.32791498e-01 -2.37933934e-01 5.09194493e-01 2.15391372e-03
5.40649772e-01 3.31905663e-01 -8.33049893e-01 -3.61514986e-01
-1.00244462e+00 -2.75954306e-01 7.02203155e-01 9.03745770e-01
5.81544638e-01 3.11492562e-01 -1.28066659e-01 1.09866858e+00
2.19705075e-01 6.41379118e-01 8.63865554e-01 -6.76609457e-01
2.98459291e-01 6.89210355e-01 -2.52340078e-01 -5.44114470e-01
-5.04136324e-01 -3.54948878e-01 -6.19193494e-01 -2.94187397e-01
4.65869695e-01 -4.31584626e-01 -1.27604258e+00 1.58365536e+00
5.83496615e-02 -6.82838142e-01 7.21369535e-02 3.94600034e-01
8.31152320e-01 1.00834453e+00 6.12532437e-01 -5.73763371e-01
1.53534198e+00 -7.55775750e-01 -1.19277060e+00 -2.09154919e-01
8.56681406e-01 -1.15462995e+00 1.05990696e+00 2.55479872e-01
-1.06591535e+00 -8.19217086e-01 -7.60712028e-01 2.68836141e-01
-3.32866967e-01 2.43836492e-01 2.20836967e-01 1.12488532e+00
-8.05713117e-01 4.31069553e-01 -7.54838943e-01 -4.10250902e-01
-4.46915068e-02 4.05380845e-01 -1.06041372e-01 2.43295997e-01
-1.30824196e+00 8.04628611e-01 9.91927028e-01 3.21702994e-02
-1.77518025e-01 -2.50556052e-01 -6.91789627e-01 -5.46620376e-02
9.63095427e-02 1.60455972e-01 1.34952390e+00 -9.33917642e-01
-1.37258542e+00 8.01163614e-01 -1.56768143e-01 -3.48518454e-02
1.31016225e-02 -3.22737753e-01 -7.56892145e-01 1.12395324e-01
5.84006786e-01 5.03974259e-01 -3.90420738e-03 -7.71642804e-01
-1.00148904e+00 -4.85345662e-01 -4.66989338e-01 9.32229534e-02
-2.43882731e-01 6.03252113e-01 -2.72420436e-01 -5.59819162e-01
1.73344135e-01 -7.13691533e-01 -2.79471815e-01 -9.69593525e-01
-2.85036296e-01 -5.27040958e-01 4.57533717e-01 -1.33055770e+00
1.66080940e+00 -2.08035898e+00 -5.73683500e-01 4.34226505e-02
-1.70266911e-01 4.58321333e-01 3.20911437e-01 2.37005517e-01
4.22250219e-02 6.04061723e-01 -4.34730142e-01 2.08632974e-03
-8.44407976e-02 3.49954873e-01 4.33728844e-01 4.53597754e-02
1.72341689e-01 6.27573907e-01 -6.92015588e-01 -8.52796078e-01
3.10999956e-02 2.72430182e-01 -6.64206268e-03 -1.12789951e-01
-3.26419100e-02 3.13434511e-01 -5.05494952e-01 5.90991795e-01
5.58300912e-01 2.06732064e-01 8.44330937e-02 3.82671744e-01
-6.21066213e-01 4.01142746e-01 -1.23602915e+00 1.56085479e+00
-1.95742399e-01 4.03741121e-01 -1.75141200e-01 -7.50716090e-01
1.32428455e+00 5.37341356e-01 1.73773050e-01 -9.53411341e-01
3.95413905e-01 4.95871603e-01 2.46502399e-01 -5.71877420e-01
7.94570982e-01 -3.52890134e-01 -5.20763516e-01 2.50571072e-01
8.23464096e-02 9.12408754e-02 4.50412363e-01 8.74282867e-02
8.02414775e-01 9.64809358e-02 3.55762154e-01 -5.21656275e-01
5.80343306e-01 1.00103922e-01 7.95393169e-01 4.45076913e-01
-2.56189972e-01 2.98865736e-01 3.81237596e-01 -3.99735980e-02
-1.10729623e+00 -7.11285472e-01 -2.97062963e-01 1.16034758e+00
-2.46499866e-01 -7.52594620e-02 -1.25876570e+00 -6.54130459e-01
-6.42689586e-01 1.06762528e+00 -1.42318666e-01 1.35411724e-01
-7.46522784e-01 -7.80307412e-01 7.61397123e-01 5.91765761e-01
8.24023843e-01 -1.61682582e+00 -5.34696579e-01 6.21587336e-01
-3.26180458e-01 -1.11790025e+00 -2.38435537e-01 2.55530804e-01
-8.18801403e-01 -5.93122482e-01 -7.36725152e-01 -1.29110515e+00
5.53490579e-01 -3.04165632e-01 8.77719402e-01 3.27589750e-01
1.32509977e-01 -1.53951868e-02 -8.94549608e-01 -4.55674112e-01
-6.84130311e-01 2.74396062e-01 -1.46561444e-01 -4.74014759e-01
6.67707562e-01 -6.78665414e-02 -1.58657134e-01 1.84820041e-01
-9.22854066e-01 -2.34212652e-01 6.84819162e-01 5.10102212e-01
5.13995707e-01 1.67127952e-01 7.61534452e-01 -9.96349812e-01
6.18276954e-01 -5.75533509e-02 -3.76348108e-01 3.11858833e-01
-5.65916240e-01 -1.40885815e-01 5.39242029e-01 -2.84620255e-01
-1.41718352e+00 -7.73595795e-02 -7.34284818e-01 6.03550196e-01
-5.43064237e-01 5.42936385e-01 -4.61427629e-01 5.81288874e-01
3.46079856e-01 3.46908122e-01 -3.80356967e-01 -5.45176446e-01
3.88399623e-02 9.89926875e-01 5.71688056e-01 -4.08054054e-01
2.23051488e-01 -3.68358046e-01 -4.66251194e-01 -1.13861799e+00
-8.46324146e-01 -4.93046850e-01 -1.02602196e+00 -2.14065120e-01
1.49333680e+00 -6.82246745e-01 -5.11384308e-01 6.93543196e-01
-1.21645951e+00 -3.28986466e-01 -1.16469808e-01 7.30567932e-01
1.55320451e-01 6.67635083e-01 -8.18596303e-01 -9.17596102e-01
-4.99213874e-01 -9.64806914e-01 5.70859849e-01 7.08818555e-01
-4.31905925e-01 -8.88746738e-01 -1.09042086e-01 3.45571548e-01
-3.44799794e-02 -2.02990696e-02 8.78825366e-01 -1.12625217e+00
5.78913353e-02 -2.27788880e-01 1.50429625e-02 8.10786605e-01
9.06418338e-02 2.31426835e-01 -7.30000913e-01 1.21924810e-01
9.36878100e-02 1.62977710e-01 5.13256192e-01 5.46803415e-01
5.52395105e-01 2.96051055e-02 -5.71699999e-02 -2.48231634e-01
1.27966237e+00 1.06436980e+00 7.98513770e-01 4.19031650e-01
5.07965684e-01 6.01108193e-01 7.45891094e-01 1.74260363e-01
2.90342808e-01 1.55756474e-01 -1.17347486e-01 1.17781974e-01
-2.90088113e-02 -2.10946593e-02 4.58132178e-01 1.41876662e+00
1.60029635e-01 -1.07468165e-01 -1.20002913e+00 8.17369580e-01
-1.35624063e+00 -7.70965219e-01 -8.02816212e-01 1.98029220e+00
1.03415012e+00 4.49539930e-01 7.95359388e-02 4.18036968e-01
1.22119331e+00 -1.72584236e-01 2.25233003e-01 -8.29808772e-01
-1.40890494e-01 5.06785810e-01 3.95363331e-01 5.49140751e-01
-9.26469386e-01 1.33899069e+00 5.47939777e+00 9.36036110e-01
-8.98388743e-01 1.02018520e-01 4.96725291e-01 5.50290644e-01
-5.18669002e-02 1.20516695e-01 -1.23806572e+00 8.84576857e-01
1.24037707e+00 5.28445505e-02 -2.82470047e-01 6.26673400e-01
9.75742787e-02 -6.87978566e-01 -3.87866437e-01 3.65617633e-01
-1.29823923e-01 -7.57903814e-01 -3.33434612e-01 -9.96705517e-02
7.06134558e-01 -1.70786530e-01 -6.08310163e-01 4.43720043e-01
3.08323145e-01 -6.49294734e-01 7.46647179e-01 3.58319849e-01
6.17269456e-01 -8.28365564e-01 1.16124940e+00 4.41117704e-01
-1.21217477e+00 3.37012827e-01 -3.33085746e-01 -3.65402907e-01
5.17121971e-01 7.42559433e-01 -6.15433335e-01 5.99427760e-01
8.52854669e-01 8.53526667e-02 -5.36973953e-01 1.00075948e+00
-8.07036996e-01 1.04687619e+00 -3.31372947e-01 -5.81293762e-01
2.48886973e-01 -4.00397271e-01 1.24854065e-01 1.56587148e+00
2.72952616e-01 3.78084540e-01 1.81290179e-01 3.79163980e-01
1.66741207e-01 5.57585359e-01 -5.47643891e-03 -9.73389223e-02
7.30610549e-01 1.00756073e+00 -1.43758416e+00 -4.79004920e-01
-2.98853219e-01 7.73883045e-01 -3.55046876e-02 7.88493734e-03
-7.37269878e-01 -9.75879908e-01 -7.85710365e-02 -8.98062512e-02
4.18456584e-01 -3.07354689e-01 -7.24344313e-01 -6.41528010e-01
-3.24933976e-02 -4.18396473e-01 2.47072607e-01 -4.95055139e-01
-9.89002287e-01 6.74946547e-01 -4.05474901e-02 -5.36086202e-01
8.15960616e-02 -4.56254333e-01 -6.66315377e-01 1.07910347e+00
-8.66416037e-01 -8.62109900e-01 1.10922288e-02 1.38077587e-01
7.27300584e-01 -1.45759106e-01 8.45707953e-01 5.20438015e-01
-9.13836241e-01 4.63875800e-01 4.58294302e-02 7.05446243e-01
4.54348117e-01 -1.11431682e+00 2.80335546e-01 1.20457852e+00
-1.09185845e-01 4.69355047e-01 4.51980442e-01 -1.04812968e+00
-3.89483720e-01 -8.55174243e-01 1.65156710e+00 -1.10067382e-01
5.57917655e-01 -2.07124054e-01 -1.05529571e+00 6.33749425e-01
3.67839575e-01 -8.39960873e-01 8.77073050e-01 8.11147913e-02
3.50308627e-01 8.83314759e-02 -1.12024188e+00 4.30730045e-01
5.02657950e-01 -3.36350173e-01 -8.72118235e-01 1.82907701e-01
1.04325211e+00 -1.13553166e-01 -1.24061406e+00 1.18273482e-01
3.82878095e-01 -6.78954005e-01 2.89066494e-01 -2.00860202e-01
3.05258453e-01 -2.53829479e-01 -1.37508169e-01 -8.75467718e-01
-4.58899438e-01 -2.36931756e-01 9.07645166e-01 1.92434204e+00
9.17288482e-01 -6.06950760e-01 5.31237423e-01 7.23023355e-01
-6.06025100e-01 -2.48352662e-01 -9.04702008e-01 -5.80513775e-01
2.15311900e-01 -4.78209913e-01 2.77047724e-01 8.45073164e-01
2.87626516e-02 6.42599225e-01 8.39988217e-02 -1.02128595e-01
2.28587508e-01 -4.37816381e-01 1.13713272e-01 -1.23092604e+00
1.12291709e-01 -3.65248412e-01 -4.97287475e-02 -6.98439598e-01
1.04984581e-01 -4.84858185e-01 2.24473536e-01 -1.89158571e+00
-1.65559892e-02 -2.08754957e-01 -9.45495963e-02 6.52231514e-01
-3.25086892e-01 4.42111865e-02 1.81110695e-01 -1.07314792e-02
-4.15369302e-01 1.90919802e-01 1.10844445e+00 2.86409587e-01
-3.71401370e-01 1.65385738e-01 -5.56711376e-01 6.21477723e-01
1.20034099e+00 -7.24182367e-01 4.84050401e-02 -4.00557876e-01
2.40178719e-01 -1.89942401e-03 -3.83252680e-01 -7.98985422e-01
1.60549134e-01 -2.10250765e-02 5.38401604e-01 -8.62951815e-01
-1.88072860e-01 -5.52828252e-01 -5.21883704e-02 7.11136043e-01
8.12147278e-03 3.28914583e-01 3.84658933e-01 -2.64180414e-02
-1.40809581e-01 -9.60219502e-01 6.79660678e-01 -4.50356960e-01
-8.68053913e-01 -4.23892170e-01 -9.73347127e-01 1.57222316e-01
1.12289834e+00 -5.99706352e-01 -1.19544581e-01 1.04230776e-01
-8.45081925e-01 1.05705291e-01 1.48388967e-01 6.89550415e-02
1.04791909e-01 -8.84771287e-01 -6.54535770e-01 1.62035540e-01
-4.20982778e-01 1.50155231e-01 2.11476266e-01 6.26843691e-01
-7.84441829e-01 3.78482312e-01 -2.56804347e-01 -3.33959371e-01
-1.24704337e+00 9.93899405e-02 -1.42215282e-01 -4.77446914e-01
-2.03135818e-01 9.66417074e-01 -7.11193308e-02 -3.95204753e-01
-1.11972792e-02 -3.82359326e-01 -6.08050942e-01 2.04814062e-01
2.35487640e-01 5.52235484e-01 5.22260144e-02 -7.81236291e-01
-3.45915318e-01 6.22666836e-01 -1.31045088e-01 -3.47155988e-01
1.02675235e+00 -1.58265442e-01 -5.13247311e-01 7.27624536e-01
8.00388753e-01 1.30716726e-01 -5.83403111e-01 1.73359096e-01
4.53426689e-01 2.43558753e-02 -2.65780628e-01 -1.12909210e+00
-6.34962201e-01 8.00627232e-01 2.91215599e-01 4.33599591e-01
1.03569639e+00 -8.07267725e-02 8.66071463e-01 1.00370064e-01
2.19041541e-01 -1.72997499e+00 -2.79110342e-01 9.98350799e-01
1.82917327e-01 -8.95154119e-01 -2.13568240e-01 -2.89662987e-01
-7.28291273e-01 9.47363913e-01 5.26802540e-01 4.16151546e-02
5.70297003e-01 2.45343298e-01 2.27108836e-01 -1.21520590e-02
3.31840925e-02 -1.02920391e-01 1.10577196e-01 3.41716796e-01
9.59775448e-01 2.89782375e-01 -1.19905949e+00 1.03486073e+00
-5.55467963e-01 -2.96364486e-01 5.47491074e-01 1.11169958e+00
-9.92110074e-01 -1.31571519e+00 -4.14614022e-01 3.06500018e-01
-1.08061576e+00 -1.84686080e-01 -3.21549743e-01 9.24754739e-01
3.48527521e-01 1.29064023e+00 2.34747455e-01 -3.03808331e-01
3.20147485e-01 4.61766362e-01 2.12111801e-01 -8.10688317e-01
-9.14423287e-01 5.47245741e-01 2.74434537e-01 1.09283105e-01
-5.76484501e-01 -6.70022905e-01 -1.89889479e+00 -1.27672240e-01
-6.26708746e-01 7.24080265e-01 8.68806601e-01 1.21950340e+00
-3.07273448e-01 7.45106578e-01 2.89329588e-01 -1.56365484e-01
-1.35790527e-01 -1.45885992e+00 -8.83279979e-01 -2.76288427e-02
-4.43568379e-01 -5.14132567e-02 -8.12551305e-02 2.74400294e-01] | [10.115629196166992, 10.130635261535645] |
90e0eaab-bc71-4281-a1e8-72e7a0ffa4d3 | towards-globally-consistent-stochastic-human | 2305.12554 | null | https://arxiv.org/abs/2305.12554v1 | https://arxiv.org/pdf/2305.12554v1.pdf | Towards Globally Consistent Stochastic Human Motion Prediction via Motion Diffusion | Stochastic human motion prediction aims to predict multiple possible upcoming pose sequences based on past human motion trajectories. Prior works focused heavily on generating diverse motion samples, leading to inconsistent, abnormal predictions from the immediate past observations. To address this issue, in this work, we propose DiffMotion, a diffusion-based stochastic human motion prediction framework that considers both the kinematic structure of the human body and the globally temporally consistent nature of motion. Specifically, DiffMotion consists of two modules: 1) a transformer-based network for generating an initial motion reconstruction from corrupted motion, and 2) a multi-stage graph convolutional network to iteratively refine the generated motion based on past observations. Facilitated by the proposed direct target prediction objective and the variance scheduler, our method is capable of predicting accurate, realistic and consistent motion with an appropriate level of diversity. Our results on benchmark datasets demonstrate that DiffMotion outperforms previous methods by large margins in terms of accuracy and fidelity while demonstrating superior robustness. | ['Girish Chowdhary', 'Jiarui Sun'] | 2023-05-21 | null | null | null | null | ['motion-prediction', 'stochastic-human-motion-prediction', 'human-motion-prediction'] | ['computer-vision', 'computer-vision', 'time-series'] | [ 1.40857711e-01 3.37904580e-02 -1.10456839e-01 1.21368589e-02
-3.84930342e-01 -1.68871835e-01 6.15136564e-01 -4.41403538e-01
-1.08889617e-01 6.67181551e-01 6.40990555e-01 5.92517555e-02
1.31700769e-01 -5.32696426e-01 -5.42423487e-01 -4.80821848e-01
-3.14779043e-01 3.63114536e-01 4.10879314e-01 -7.38647506e-02
-4.44195680e-02 4.37372625e-01 -1.42467749e+00 9.47191343e-02
5.98100543e-01 6.88141286e-01 2.67204374e-01 1.03062403e+00
6.17300212e-01 1.29923081e+00 -2.61390686e-01 -2.23248705e-01
1.74169570e-01 -9.33133185e-01 -7.48205721e-01 2.51399100e-01
8.29611048e-02 -5.84037125e-01 -6.14907324e-01 5.57681680e-01
6.13514662e-01 5.96898973e-01 4.73646402e-01 -1.24934161e+00
-2.48229235e-01 4.79838192e-01 -3.03818136e-01 1.36328965e-01
6.50027573e-01 8.42978299e-01 7.53562808e-01 -7.30524182e-01
1.03780866e+00 1.18802798e+00 7.62801468e-01 8.51054668e-01
-1.06652808e+00 -3.82259697e-01 2.37319022e-01 2.93766886e-01
-1.16182435e+00 -4.48689729e-01 8.70039880e-01 -5.21638811e-01
8.53669703e-01 1.05439074e-01 9.92259741e-01 1.51952338e+00
7.93826222e-01 8.57674778e-01 3.78937721e-01 1.38859957e-01
3.66885602e-01 -7.81261027e-01 -4.34184223e-01 7.19876587e-01
-2.74972524e-02 2.80602574e-01 -6.87129974e-01 -1.60095617e-01
8.01412106e-01 -1.66512251e-01 -3.17223430e-01 -3.68752122e-01
-1.64999986e+00 4.28216666e-01 2.32497856e-01 5.58927543e-02
-9.36901927e-01 5.98607838e-01 1.68411136e-01 -2.24168554e-01
3.65922451e-01 2.12648734e-01 -4.77822721e-02 -3.65032524e-01
-1.20732331e+00 7.39606619e-01 6.86961710e-01 7.71570802e-01
3.81521970e-01 4.06880051e-01 -5.39813995e-01 1.52627870e-01
4.79463756e-01 5.39838552e-01 5.04484892e-01 -1.47214973e+00
5.21797597e-01 1.61531240e-01 5.01426756e-01 -1.39059126e+00
-5.57817400e-01 -4.49908674e-01 -9.57560420e-01 1.00809984e-01
3.09096247e-01 -5.09005368e-01 -9.22623158e-01 1.97706103e+00
6.37742162e-01 6.43141448e-01 -1.31022111e-01 1.29766178e+00
3.15950364e-01 6.97303653e-01 3.02028745e-01 -1.11532643e-01
6.70373380e-01 -1.15513444e+00 -7.23728180e-01 -3.80788505e-01
5.51841855e-01 -4.44206268e-01 5.74161530e-01 1.88019499e-01
-1.40410101e+00 -7.94616520e-01 -9.64927852e-01 1.09815486e-01
5.63807249e-01 1.35721400e-01 1.77758038e-01 2.12699771e-01
-1.20219100e+00 1.04567719e+00 -1.32633400e+00 -1.99705109e-01
1.47974610e-01 1.92485064e-01 -2.32076749e-01 5.24230786e-02
-1.06884253e+00 5.17074466e-01 3.02055240e-01 3.76599967e-01
-1.01828420e+00 -5.56121528e-01 -1.02150369e+00 -1.31240740e-01
1.91290647e-01 -1.51634490e+00 1.17367327e+00 -7.89844275e-01
-1.68617833e+00 6.15459979e-02 -2.09963232e-01 -6.91898763e-01
1.05517673e+00 -4.57637370e-01 -2.30702549e-01 2.67044634e-01
1.20798998e-01 9.42275763e-01 9.71317172e-01 -8.87998879e-01
-5.76898038e-01 -1.34307057e-01 -5.45813441e-01 5.14944255e-01
2.50144541e-01 -5.50318360e-01 -6.12241328e-01 -9.53395605e-01
1.30748088e-02 -1.39331472e+00 -7.10853755e-01 -9.03001577e-02
-4.80110645e-01 2.10812435e-01 6.87744379e-01 -8.33184361e-01
1.35466480e+00 -1.77153361e+00 5.38225889e-01 1.86404660e-01
3.02135319e-01 1.02563068e-01 -1.04311042e-01 4.89566356e-01
2.35390052e-01 -2.39328757e-01 -1.87890559e-01 -6.08982027e-01
-1.31911829e-01 8.12276639e-03 -3.10424089e-01 5.61413884e-01
2.84382731e-01 1.21546268e+00 -1.22736084e+00 -3.70853692e-01
2.96670258e-01 5.72546422e-01 -6.01104319e-01 4.39541191e-01
-4.03726071e-01 1.02498293e+00 -5.27429879e-01 3.42062414e-01
2.75319338e-01 -4.62258548e-01 4.07141328e-01 -3.86594199e-02
2.34306037e-01 -5.14836647e-02 -1.17187965e+00 1.85377133e+00
-7.40289465e-02 4.66749072e-01 -2.52701551e-01 -2.62704760e-01
7.35720992e-01 3.47938746e-01 9.75407302e-01 -3.40865135e-01
1.33252114e-01 3.61486850e-03 -9.52588618e-02 -4.26908255e-01
8.87278318e-01 5.53851798e-02 -4.61270753e-03 4.95292306e-01
-3.38949859e-01 2.45310292e-01 -5.11718765e-02 1.30837038e-01
1.41565728e+00 6.89369261e-01 1.19966120e-01 8.67079496e-02
4.90070432e-01 1.40669316e-01 8.26686859e-01 5.59610426e-01
-4.81200218e-01 9.65738833e-01 1.00875096e-02 -5.27266622e-01
-1.14065421e+00 -1.12029338e+00 9.54121053e-01 5.79907417e-01
3.89656395e-01 -3.22985917e-01 -6.39349043e-01 -4.35818136e-01
-1.30341798e-01 6.26116753e-01 -4.93895352e-01 -2.91561544e-01
-1.07265806e+00 -4.26430702e-01 3.29902738e-01 7.10359275e-01
5.29707849e-01 -1.22014308e+00 -9.42669153e-01 6.23254240e-01
-4.57686067e-01 -1.20570219e+00 -8.70421827e-01 -6.81803942e-01
-8.64351749e-01 -9.01172280e-01 -9.90412354e-01 -4.98162866e-01
4.36832517e-01 1.87067524e-01 1.02135801e+00 1.08070880e-01
-1.36864349e-01 5.94012201e-01 -2.09249824e-01 1.81902468e-01
-5.99588096e-01 -5.36417998e-02 6.51129410e-02 3.33808511e-01
-2.76892304e-01 -5.99353671e-01 -1.10852158e+00 1.96966961e-01
-7.33335555e-01 4.41666722e-01 4.32706594e-01 7.94169188e-01
5.23771286e-01 -1.12095140e-01 3.47993344e-01 -4.89257246e-01
4.96643275e-01 -8.10061097e-01 -1.77749321e-01 -1.14108801e-01
-4.67031837e-01 -1.53175471e-02 5.85085154e-01 -5.29060185e-01
-1.21035302e+00 6.04318857e-01 -1.08910508e-01 -7.99340367e-01
-1.00366622e-01 3.02799284e-01 1.08924687e-01 3.22016060e-01
6.46377385e-01 4.09236759e-01 1.09897137e-01 -5.70506565e-02
5.22456169e-01 -5.05415685e-02 8.41375291e-01 -2.20701084e-01
7.00463474e-01 7.02515900e-01 2.38428921e-01 -6.17233932e-01
-3.41918141e-01 -3.57226133e-01 -7.26202309e-01 -7.64267743e-01
1.19557130e+00 -1.00081813e+00 -5.66621542e-01 7.65387833e-01
-1.10695386e+00 -7.14223802e-01 -2.33304769e-01 6.88981593e-01
-8.72457147e-01 6.34851158e-01 -8.70096684e-01 -9.85078096e-01
-4.29598689e-01 -1.05728233e+00 1.24164093e+00 -1.06153272e-01
-8.22181106e-01 -9.53780115e-01 2.93576807e-01 2.25607887e-01
1.30465567e-01 7.41384864e-01 3.20028454e-01 -1.04119629e-01
-9.32583928e-01 -1.10204734e-01 4.08479571e-01 -9.72808823e-02
3.14477459e-02 -2.67778523e-02 -3.95282924e-01 -3.93345892e-01
-1.66743606e-01 -2.44000014e-02 7.30400443e-01 6.48917496e-01
6.08437598e-01 -3.96211863e-01 -4.77815807e-01 5.50391614e-01
1.07918036e+00 3.07647940e-02 7.32150793e-01 2.01984823e-01
1.03441691e+00 6.52168751e-01 8.02161753e-01 7.61780441e-01
5.73473096e-01 6.88166618e-01 3.70553702e-01 3.63964051e-01
-3.99939805e-01 -7.28231370e-01 5.81188917e-01 7.36666203e-01
-5.17133884e-02 -5.92834055e-01 -9.21009064e-01 7.28052914e-01
-2.26175976e+00 -1.29949093e+00 -1.65258616e-01 1.95641613e+00
3.02687705e-01 1.21110089e-01 4.39374834e-01 -9.03865546e-02
7.44833112e-01 4.71009463e-01 -7.60106564e-01 -1.49505772e-02
1.56465277e-01 -3.13552409e-01 2.12440610e-01 4.57003474e-01
-9.99360085e-01 9.03184652e-01 6.45264578e+00 4.72647816e-01
-1.01837397e+00 -1.79338947e-01 6.49496973e-01 -3.06638420e-01
-4.06301856e-01 -1.29620239e-01 -6.64604068e-01 6.31039798e-01
9.19117391e-01 -1.31523877e-01 1.91320643e-01 6.44985080e-01
6.75324142e-01 5.47353253e-02 -8.16030681e-01 8.11242878e-01
-8.48970637e-02 -1.48704255e+00 2.26341113e-01 1.37224961e-02
9.31529105e-01 -3.45833898e-02 1.06180660e-01 -7.44127035e-02
4.62691784e-01 -9.60272849e-01 1.10959554e+00 1.02224350e+00
3.14524829e-01 -7.52655327e-01 4.53645885e-01 6.86084688e-01
-1.46306026e+00 -2.87774429e-02 1.89685464e-01 -1.85669482e-01
7.81197131e-01 4.48336989e-01 -7.86472797e-01 6.61957741e-01
4.13469672e-01 1.07882297e+00 -3.02139372e-01 9.43376958e-01
-2.50311553e-01 5.72630048e-01 -1.16537750e-01 7.28668645e-02
2.18100727e-01 -2.49581877e-02 8.57901156e-01 1.05170190e+00
6.13053977e-01 1.45773530e-01 3.86593252e-01 8.80717516e-01
4.20505792e-01 -1.90326795e-01 -5.76085508e-01 1.68941438e-01
3.71763289e-01 9.45371151e-01 -5.81695378e-01 -2.84389555e-01
4.92631691e-03 1.32830513e+00 1.68972477e-01 3.74027640e-01
-9.76651192e-01 2.23488212e-01 6.53166711e-01 1.26266673e-01
4.42914844e-01 -7.56881952e-01 -2.23601863e-01 -1.02199566e+00
1.14642002e-01 -6.04517102e-01 4.06616688e-01 -8.12631547e-01
-9.99187708e-01 6.67795777e-01 -1.64297968e-01 -1.55727553e+00
-1.12531042e+00 3.82405333e-02 -6.40696526e-01 8.36467803e-01
-1.02850211e+00 -1.25429487e+00 -3.58477294e-01 5.10364056e-01
6.53496087e-01 8.96564275e-02 2.97796398e-01 -1.03971608e-01
-4.74874049e-01 2.10777387e-01 -3.32905263e-01 -8.29151943e-02
3.52124274e-01 -9.73171294e-01 1.11484194e+00 1.10146940e+00
-4.87830825e-02 1.52838260e-01 9.82702374e-01 -1.29564047e+00
-1.43140674e+00 -1.50116956e+00 8.75651121e-01 -5.38742125e-01
5.47359228e-01 2.01608226e-01 -8.89779150e-01 5.91687202e-01
-1.76266596e-01 -4.24684659e-02 1.76299408e-01 -7.73484290e-01
2.72721708e-01 3.41869742e-01 -7.97908425e-01 1.05997956e+00
1.51972842e+00 -9.57844406e-02 -2.99167871e-01 4.69781272e-02
5.91173887e-01 -7.12615073e-01 -7.27904797e-01 4.33320999e-01
7.03393579e-01 -1.08442867e+00 9.64950502e-01 -3.87475222e-01
6.25554860e-01 -3.78799438e-01 1.63562179e-01 -1.11471486e+00
-6.17768288e-01 -1.06910431e+00 -6.42370582e-01 6.83473945e-01
1.74149320e-01 -2.01745722e-02 1.30556297e+00 7.50230014e-01
-5.08457869e-02 -7.83431232e-01 -8.18866909e-01 -7.76466191e-01
-3.40173930e-01 -7.02149093e-01 5.31773806e-01 5.01340389e-01
-3.69064659e-01 1.23960055e-01 -1.06752527e+00 2.92848885e-01
7.14675248e-01 6.28078952e-02 1.12353337e+00 -6.21226788e-01
-5.57868123e-01 -1.62349105e-01 -5.16839206e-01 -1.45308638e+00
1.98219359e-01 -4.92506534e-01 5.23177385e-01 -1.62954772e+00
-1.10134587e-01 1.63658008e-01 1.82150751e-01 4.21130732e-02
-4.79042053e-01 7.88611099e-02 3.15455288e-01 5.60166299e-01
-7.23258078e-01 8.76825929e-01 1.45869899e+00 6.61645755e-02
-5.10311365e-01 1.73301533e-01 -3.14672031e-02 7.04609156e-01
6.28954291e-01 -2.68267214e-01 -7.58048177e-01 -3.44072759e-01
-5.23660518e-02 5.88897526e-01 6.07852221e-01 -1.45958769e+00
4.47490513e-01 -3.88167620e-01 5.30305982e-01 -7.61474133e-01
5.47118425e-01 -4.11565244e-01 7.53576458e-01 8.98040950e-01
-3.54946792e-01 4.70862269e-01 -8.04179609e-02 1.13496327e+00
1.30797671e-02 4.34303761e-01 4.61463004e-01 -1.67992353e-01
-1.17736733e+00 6.36048615e-01 -6.89349651e-01 3.06754932e-02
1.07137454e+00 -3.86167824e-01 7.43509829e-02 -8.44488978e-01
-9.78535533e-01 2.16724619e-01 5.62682927e-01 5.75830042e-01
8.87267292e-01 -1.56228304e+00 -6.99154735e-01 4.36233990e-02
-2.32010528e-01 -5.75266592e-02 5.29013455e-01 7.38184273e-01
-7.08873630e-01 2.70420581e-01 -1.50796860e-01 -7.14218915e-01
-1.04005909e+00 4.06293571e-01 3.30145657e-01 -3.63344431e-01
-1.05826473e+00 5.76644421e-01 1.52704967e-02 -2.05544140e-02
-3.53067406e-02 -2.38709271e-01 -1.02098972e-01 -5.42685807e-01
4.03105319e-01 7.23255157e-01 -4.31620330e-01 -1.09024000e+00
-2.93057501e-01 3.27156693e-01 4.29487824e-01 -5.13899386e-01
1.02131593e+00 -5.17307758e-01 3.91730577e-01 4.00203079e-01
8.89254272e-01 -2.26812392e-01 -2.00325871e+00 1.12142533e-01
1.69413775e-01 -4.00032669e-01 -2.63114363e-01 -4.50052023e-01
-1.01760769e+00 3.61208200e-01 2.69931585e-01 -2.62278944e-01
9.78348672e-01 -3.81661773e-01 1.46046007e+00 -2.72414591e-02
6.31323993e-01 -9.82471585e-01 3.87283474e-01 5.19312084e-01
7.77199805e-01 -9.06490803e-01 -1.10537922e-02 -3.90573233e-01
-1.00122905e+00 9.57490385e-01 6.44807577e-01 -2.35344216e-01
5.51581562e-01 -6.85576275e-02 3.49287651e-02 -3.80537286e-02
-1.02654505e+00 -5.10764010e-02 6.88427985e-01 7.15177417e-01
1.09958351e-01 -1.46491110e-01 -1.27232432e-01 3.70889068e-01
-2.40191594e-01 3.04663181e-01 3.31121981e-01 9.33647096e-01
-3.08891267e-01 -7.50357985e-01 -3.74237895e-01 8.01099017e-02
-2.09014907e-01 2.89694458e-01 -2.65302539e-01 6.01593018e-01
-1.32617444e-01 9.52780843e-01 -4.24758419e-02 -7.84256518e-01
1.54887810e-01 -2.72466034e-01 2.96439379e-01 -3.29569846e-01
-4.20758873e-01 -7.51257269e-03 2.32910320e-01 -1.13195813e+00
-3.73583406e-01 -7.73188651e-01 -1.34399641e+00 -4.35203582e-01
2.63328612e-01 -1.73928067e-01 1.28725737e-01 9.83198583e-01
7.79560804e-01 4.70170617e-01 4.88744438e-01 -1.36396360e+00
-5.29858232e-01 -6.26032114e-01 -3.94557148e-01 7.95846820e-01
3.19560051e-01 -4.19827133e-01 -1.59799576e-01 2.76302516e-01] | [7.330674171447754, -0.14473344385623932] |
64067071-ac55-4a45-b559-6b81bcafa116 | hidanet-rgb-d-salient-object-detection-via | 2301.07405 | null | https://arxiv.org/abs/2301.07405v1 | https://arxiv.org/pdf/2301.07405v1.pdf | HiDAnet: RGB-D Salient Object Detection via Hierarchical Depth Awareness | RGB-D saliency detection aims to fuse multi-modal cues to accurately localize salient regions. Existing works often adopt attention modules for feature modeling, with few methods explicitly leveraging fine-grained details to merge with semantic cues. Thus, despite the auxiliary depth information, it is still challenging for existing models to distinguish objects with similar appearances but at distinct camera distances. In this paper, from a new perspective, we propose a novel Hierarchical Depth Awareness network (HiDAnet) for RGB-D saliency detection. Our motivation comes from the observation that the multi-granularity properties of geometric priors correlate well with the neural network hierarchies. To realize multi-modal and multi-level fusion, we first use a granularity-based attention scheme to strengthen the discriminatory power of RGB and depth features separately. Then we introduce a unified cross dual-attention module for multi-modal and multi-level fusion in a coarse-to-fine manner. The encoded multi-modal features are gradually aggregated into a shared decoder. Further, we exploit a multi-scale loss to take full advantage of the hierarchical information. Extensive experiments on challenging benchmark datasets demonstrate that our HiDAnet performs favorably over the state-of-the-art methods by large margins. | ['Cédric Demonceaux', 'Chao Ma', 'Fabrice Meriaudeau', 'Guillaume Allibert', 'Zongwei Wu'] | 2023-01-18 | null | null | null | null | ['rgb-d-salient-object-detection', 'saliency-detection'] | ['computer-vision', 'computer-vision'] | [ 6.55519888e-02 -1.75600290e-01 -1.96386158e-01 -4.12043452e-01
-9.35290456e-01 -2.60211021e-01 4.48345274e-01 3.34213257e-01
-1.54769763e-01 3.78344893e-01 3.65005404e-01 1.86689958e-01
-8.21972862e-02 -6.30643010e-01 -7.65767992e-01 -7.19792247e-01
3.55590999e-01 -2.47267067e-01 8.31176460e-01 -2.60422230e-01
3.02209795e-01 4.07953292e-01 -1.78024125e+00 3.77038360e-01
1.05101526e+00 1.30663216e+00 5.18173158e-01 2.46131286e-01
1.25075594e-01 7.02372372e-01 -6.67355582e-02 -1.68446675e-01
3.39014083e-01 -2.13411972e-01 -5.87814212e-01 1.70851290e-01
5.25596321e-01 -5.03531456e-01 -3.79434705e-01 1.16403008e+00
3.49134117e-01 4.94946875e-02 2.72118360e-01 -1.26595318e+00
-8.35533500e-01 2.69961059e-01 -9.50256884e-01 3.49021018e-01
3.19861919e-01 2.10752189e-01 1.21082509e+00 -9.84765947e-01
2.36350983e-01 1.33390379e+00 4.04883265e-01 6.65642172e-02
-1.05602908e+00 -5.14578879e-01 6.94929123e-01 3.06776971e-01
-1.32291853e+00 -1.60647914e-01 1.24916828e+00 -1.52317062e-01
6.20404363e-01 1.34510294e-01 6.83168948e-01 8.34021986e-01
1.18390366e-01 1.16001701e+00 1.18606138e+00 -1.78242490e-01
6.47562891e-02 -5.42445965e-02 1.28234640e-01 9.47695911e-01
3.25022072e-01 7.11018443e-02 -8.30349863e-01 1.19714580e-01
1.07958269e+00 4.37107444e-01 -2.40167886e-01 -7.44781077e-01
-1.30340064e+00 8.57320130e-01 1.09224391e+00 3.46873522e-01
-4.80597943e-01 1.06764220e-01 1.51285782e-01 -2.51457959e-01
5.68571746e-01 2.39249855e-01 -3.89190972e-01 3.22725117e-01
-9.81787741e-01 5.90780005e-02 3.58685739e-02 9.26506102e-01
1.18316412e+00 -9.23113599e-02 -4.40000653e-01 5.81330299e-01
5.47132134e-01 4.52963948e-01 4.15837377e-01 -8.25631738e-01
3.23309243e-01 8.89743567e-01 3.94421071e-02 -1.27193570e+00
-4.59697545e-01 -6.71928704e-01 -7.79000044e-01 9.03724134e-02
9.39455628e-02 3.06345582e-01 -9.78216827e-01 1.76420474e+00
4.20948565e-01 3.87725085e-01 -2.19979271e-01 1.36097956e+00
8.13491344e-01 3.70505035e-01 1.72633782e-01 1.44484892e-01
1.42151892e+00 -1.22779238e+00 -2.25590438e-01 -3.83836299e-01
2.14921027e-01 -5.30989826e-01 1.10474753e+00 -1.19801044e-01
-1.11560428e+00 -6.72645152e-01 -1.12652838e+00 -5.72921574e-01
-3.41067433e-01 4.15764563e-02 8.55657935e-01 3.20571154e-01
-1.02469635e+00 1.93161532e-01 -8.74572992e-01 -1.43764317e-01
5.31820774e-01 2.80928582e-01 -1.90439805e-01 -7.46908560e-02
-1.13960290e+00 6.64432764e-01 3.82498652e-01 7.33974054e-02
-9.26749647e-01 -6.74884498e-01 -1.16343570e+00 1.08748242e-01
3.95960242e-01 -8.92797887e-01 1.04061782e+00 -8.85217845e-01
-1.18799317e+00 7.37432778e-01 -3.67146701e-01 -2.31707901e-01
1.60016432e-01 -2.37702027e-01 1.66874733e-02 4.20805156e-01
3.42655987e-01 8.30740333e-01 8.54480028e-01 -1.59688652e+00
-9.88609135e-01 -5.72501659e-01 5.25930166e-01 4.62145448e-01
-3.35200965e-01 -1.91937909e-01 -6.05496585e-01 -6.58095419e-01
3.89962703e-01 -5.49790680e-01 -2.46560678e-01 4.40748148e-02
-5.41155696e-01 -1.03175960e-01 6.26077533e-01 -3.61749887e-01
1.02294016e+00 -2.04375648e+00 3.77841651e-01 -1.21381693e-01
4.74477112e-01 3.01054567e-02 -3.08974218e-02 -1.29375622e-01
1.85789600e-01 -1.78331271e-01 -2.07487881e-01 -5.60922444e-01
9.86649990e-02 -3.76116857e-02 -2.01262817e-01 4.25289482e-01
4.74957943e-01 1.12015498e+00 -1.17416060e+00 -5.96876740e-01
4.45036054e-01 6.13570273e-01 -6.47370875e-01 3.39811705e-02
-3.53888944e-02 4.21955496e-01 -8.94579470e-01 1.03955305e+00
6.98558390e-01 -5.34034789e-01 -4.20794219e-01 -5.33481598e-01
-2.61603534e-01 1.68048725e-01 -8.41915309e-01 2.10975146e+00
-4.60868686e-01 3.49084824e-01 4.09460478e-02 -7.97978878e-01
7.05930054e-01 -1.98667780e-01 3.15704376e-01 -7.62097955e-01
2.22687230e-01 3.74034494e-01 -3.24032545e-01 -6.17178716e-02
7.19207406e-01 -5.90936691e-02 -2.43015215e-01 1.06837876e-01
2.27811430e-02 5.25449440e-02 -2.06471786e-01 9.68057290e-02
7.15314209e-01 8.69378820e-02 2.64563322e-01 -2.23537847e-01
6.99652255e-01 -2.55571783e-01 6.21188641e-01 6.55938506e-01
-3.91581953e-01 8.69115949e-01 2.51373857e-01 -2.64751196e-01
-6.80023074e-01 -1.02286649e+00 4.32989113e-02 1.30041540e+00
9.14744556e-01 -1.19338118e-01 -5.35710931e-01 -7.41034985e-01
1.40256792e-01 3.97635013e-01 -8.47685516e-01 -3.79110545e-01
-3.47379148e-01 -6.48234069e-01 1.03739396e-01 6.49922967e-01
7.68913686e-01 -7.98109591e-01 -9.37277913e-01 1.21767618e-01
-1.64938807e-01 -1.21879411e+00 -5.43240726e-01 1.78880677e-01
-8.01304638e-01 -7.75630951e-01 -9.34418857e-01 -8.53569925e-01
5.68169415e-01 9.34822559e-01 9.46858525e-01 9.04737636e-02
-1.51025161e-01 1.88893631e-01 -4.71345723e-01 -2.22965732e-01
2.82551289e-01 3.10411841e-01 -2.03306824e-01 2.11894989e-01
2.96707749e-01 -6.03961587e-01 -9.39345896e-01 2.27814585e-01
-1.00764263e+00 4.47950125e-01 9.64053333e-01 7.03845859e-01
6.85167611e-01 -1.61187425e-01 4.37056720e-01 -2.95421958e-01
1.00119613e-01 -4.52563763e-01 -4.07977134e-01 2.59984702e-01
-1.94361731e-01 1.55069813e-01 2.74749070e-01 -3.95452604e-02
-9.60463822e-01 6.78026602e-02 -2.53638299e-03 -6.74288273e-01
-1.88800856e-01 2.36676320e-01 -5.49734950e-01 -2.39089623e-01
7.08248466e-02 3.88546675e-01 -4.39411521e-01 -4.60367233e-01
5.10174096e-01 3.96358371e-01 4.86835092e-01 -5.07036388e-01
8.40406895e-01 7.72900343e-01 -1.02702081e-01 -4.89897221e-01
-1.44174445e+00 -4.45141852e-01 -6.95015311e-01 -8.11539474e-04
9.52579558e-01 -1.29341030e+00 -4.97639596e-01 6.77761734e-01
-1.09464037e+00 -1.73967317e-01 -1.35553747e-01 2.81431556e-01
-4.46919471e-01 3.14264774e-01 -5.55889904e-01 -4.96379614e-01
-1.44637302e-01 -1.34421480e+00 1.74151218e+00 5.49953699e-01
4.14439231e-01 -8.34538758e-01 -2.73209244e-01 3.12574238e-01
3.61568093e-01 3.55657130e-01 7.13081479e-01 -1.97480440e-01
-1.02211618e+00 5.68283349e-02 -9.49981928e-01 4.76091467e-02
2.21896335e-01 -3.65903527e-01 -1.00804770e+00 -1.33869037e-01
-1.90462582e-02 -2.71698892e-01 1.21369016e+00 3.91409874e-01
1.39166164e+00 1.01296633e-01 -3.21695954e-01 8.37437630e-01
1.52489936e+00 -3.47841650e-01 3.13267827e-01 5.71699202e-01
1.20124936e+00 3.52439851e-01 7.58823872e-01 5.28090775e-01
8.96169186e-01 6.80293977e-01 7.61980712e-01 -3.91746253e-01
-1.24730676e-01 -2.85231739e-01 1.37619019e-01 5.59471905e-01
1.53672710e-01 6.85174838e-02 -6.48148239e-01 7.81700015e-01
-1.88330042e+00 -7.14175403e-01 5.48787974e-02 2.13192606e+00
7.74374366e-01 2.62599349e-01 1.30663306e-01 -5.43721840e-02
7.61449456e-01 5.01138091e-01 -7.53636658e-01 1.71354800e-01
-4.15101439e-01 -1.36716813e-01 5.69722712e-01 4.30294991e-01
-1.29466987e+00 9.21118677e-01 4.86533403e+00 9.72689927e-01
-1.27767372e+00 2.02456400e-01 7.67909050e-01 -2.34281510e-01
-6.26437724e-01 1.40311774e-02 -8.72251809e-01 5.47640443e-01
2.87870586e-01 1.07099919e-03 2.43824758e-02 6.80750072e-01
-2.22178339e-03 -2.34684914e-01 -9.74397540e-01 1.01940775e+00
2.25442395e-01 -1.19198608e+00 8.29126984e-02 -2.41919514e-02
9.59536314e-01 2.05885127e-01 2.66785711e-01 1.10738166e-01
1.06570527e-01 -6.66874945e-01 1.11911607e+00 5.14155090e-01
4.35787767e-01 -8.26629698e-01 5.70600629e-01 8.34965780e-02
-1.55939496e+00 -1.84518039e-01 -4.17822272e-01 5.81781380e-02
2.55924016e-01 7.73692191e-01 -6.28223941e-02 9.09520864e-01
7.96650589e-01 1.09297359e+00 -8.76508594e-01 1.14850926e+00
-1.97571993e-01 -6.67458177e-02 -2.92348653e-01 7.33049735e-02
5.78566611e-01 2.20991626e-01 4.63272095e-01 1.01502252e+00
2.56604135e-01 2.71189306e-02 2.47609302e-01 8.98856640e-01
-1.65433194e-02 -1.80261478e-01 -1.93014026e-01 3.86050373e-01
3.63676131e-01 1.30698919e+00 -7.64679790e-01 -3.47248435e-01
-6.33429885e-01 1.18264592e+00 4.42244351e-01 2.87799805e-01
-1.05141830e+00 -1.88873380e-01 7.80775547e-01 8.08476470e-04
6.61670983e-01 -2.78361559e-01 -4.48458493e-01 -1.34652209e+00
1.39969185e-01 -5.77065825e-01 1.98706195e-01 -8.89785111e-01
-1.21607161e+00 6.65863335e-01 -1.16991878e-01 -1.34148979e+00
2.39080101e-01 -4.83804733e-01 -4.16858435e-01 8.71415854e-01
-2.33270311e+00 -1.67098689e+00 -6.06088519e-01 7.02096045e-01
5.99950075e-01 3.51974815e-01 3.47069204e-01 1.35751903e-01
-5.49393356e-01 5.00171781e-01 -9.43358690e-02 -4.78376895e-02
5.08402646e-01 -1.22240543e+00 8.91968533e-02 9.76858735e-01
-3.57088223e-02 5.83414137e-01 3.27640504e-01 -4.27159041e-01
-1.28150761e+00 -1.26532423e+00 5.73147595e-01 -4.58893627e-01
5.85781395e-01 -2.95305908e-01 -9.39206779e-01 2.98159868e-01
8.64903256e-02 2.76414454e-01 3.31681550e-01 -9.06281769e-02
-5.64624846e-01 -2.17000902e-01 -8.78182948e-01 4.45562780e-01
9.99759614e-01 -7.86624551e-01 -6.80469871e-01 -2.67250650e-02
1.16059399e+00 -4.27325755e-01 -7.25830555e-01 5.85399806e-01
3.88586283e-01 -1.43279827e+00 1.18172383e+00 -1.34368837e-01
6.90570772e-01 -6.80780590e-01 -4.51419473e-01 -9.97880161e-01
-3.98409754e-01 -2.92017728e-01 -2.59225011e-01 1.12485802e+00
8.81070942e-02 -4.26829994e-01 6.23094082e-01 3.59901577e-01
-3.51742834e-01 -9.86268580e-01 -8.74577343e-01 -5.11313260e-01
-1.64057419e-01 -3.47681493e-01 7.36325026e-01 7.26641595e-01
-2.06998751e-01 3.66689891e-01 -3.52688789e-01 5.65393627e-01
7.65026689e-01 7.39748359e-01 4.23827291e-01 -1.23625350e+00
-1.50725156e-01 -6.95064664e-01 -5.83402455e-01 -1.43785179e+00
4.13309410e-02 -6.74760759e-01 1.51424572e-01 -1.56257415e+00
5.28587520e-01 -4.58276838e-01 -8.62711489e-01 5.09151399e-01
-6.55359924e-01 5.15718162e-01 2.13818192e-01 1.54390663e-01
-1.19947135e+00 1.05328810e+00 1.29675138e+00 -7.06268251e-02
-1.62162498e-01 -4.57127154e-01 -1.13925421e+00 8.11532974e-01
5.74205399e-01 -2.13227198e-01 -2.99515009e-01 -7.82573044e-01
-1.03366695e-01 -1.78113177e-01 8.98859084e-01 -1.21002591e+00
3.01225424e-01 -1.33472919e-01 5.53628683e-01 -7.65534043e-01
4.60515916e-01 -6.94676638e-01 -5.63806236e-01 1.41955167e-01
-2.30763078e-01 -2.89740711e-01 2.03384057e-01 8.46481264e-01
-4.19895470e-01 3.24536055e-01 8.25871825e-01 -1.57191139e-02
-1.06000543e+00 6.28936052e-01 2.05090865e-01 3.18135768e-02
9.42063868e-01 -2.39344776e-01 -2.97821879e-01 -7.40703270e-02
-4.08809751e-01 3.55674148e-01 8.72305393e-01 7.05873668e-01
7.05431163e-01 -1.38659525e+00 -4.78105217e-01 3.13160777e-01
4.22198832e-01 2.65992969e-01 5.70641696e-01 1.12780738e+00
-8.05617347e-02 5.12357831e-01 -3.58459532e-01 -8.41249168e-01
-8.48301291e-01 5.78383982e-01 2.47327179e-01 -1.27325833e-01
-3.77769619e-01 1.11459410e+00 8.98553312e-01 -7.30734644e-03
1.18422709e-01 -6.51202321e-01 1.81809533e-02 -3.01869046e-02
5.19252002e-01 -3.80391255e-02 -5.61081572e-03 -1.00564933e+00
-5.30848145e-01 9.39524233e-01 -7.64496550e-02 1.99923635e-01
1.34057510e+00 -6.20684624e-01 -2.28436776e-02 4.34176773e-01
1.24977255e+00 -1.15278706e-01 -1.75613725e+00 -5.79403937e-01
-1.61624327e-01 -6.98016644e-01 3.93281043e-01 -4.94337201e-01
-1.26195312e+00 1.02529371e+00 5.28694928e-01 1.39894262e-01
1.50073135e+00 1.99786082e-01 9.78144884e-01 -6.50100037e-02
4.51574206e-01 -7.33990133e-01 2.98217356e-01 3.97626281e-01
7.76476979e-01 -1.56415033e+00 -3.42565365e-02 -5.73454618e-01
-6.40262961e-01 7.27212548e-01 7.24191189e-01 -2.08775342e-01
6.15711749e-01 -1.26463905e-01 -2.44503677e-01 -2.47567654e-01
-4.74920332e-01 -6.90050125e-01 5.60406446e-01 4.39812213e-01
2.13999718e-01 -7.89215267e-02 1.80067524e-01 8.53009939e-01
2.37650275e-01 -1.08386949e-01 1.91447422e-01 7.94172108e-01
-7.68209934e-01 -7.77531981e-01 -2.48854578e-01 2.33205900e-01
-2.46685445e-01 -2.80201942e-01 -1.94850489e-01 6.41840577e-01
3.16553205e-01 8.44164729e-01 -1.04229555e-01 -3.90478730e-01
2.30245903e-01 -3.92889440e-01 3.93409967e-01 -5.32370031e-01
-3.10215265e-01 2.20029101e-01 -4.45491731e-01 -7.61449099e-01
-6.96394086e-01 -6.55986071e-01 -1.19771171e+00 4.94140275e-02
-2.18359426e-01 -1.93984926e-01 3.40790480e-01 9.93882418e-01
5.73107779e-01 7.16788948e-01 7.42264450e-01 -1.31482601e+00
-1.85429007e-01 -6.17244124e-01 -5.26887953e-01 2.86711723e-01
7.56330609e-01 -1.02805459e+00 -4.02919471e-01 -2.88407713e-01] | [9.714252471923828, -0.7000154852867126] |
811e3161-4f73-42fc-ac3f-3708308d128c | image-to-image-retrieval-by-learning | 2012.147 | null | https://arxiv.org/abs/2012.14700v1 | https://arxiv.org/pdf/2012.14700v1.pdf | Image-to-Image Retrieval by Learning Similarity between Scene Graphs | As a scene graph compactly summarizes the high-level content of an image in a structured and symbolic manner, the similarity between scene graphs of two images reflects the relevance of their contents. Based on this idea, we propose a novel approach for image-to-image retrieval using scene graph similarity measured by graph neural networks. In our approach, graph neural networks are trained to predict the proxy image relevance measure, computed from human-annotated captions using a pre-trained sentence similarity model. We collect and publish the dataset for image relevance measured by human annotators to evaluate retrieval algorithms. The collected dataset shows that our method agrees well with the human perception of image similarity than other competitive baselines. | ['Eun-Sol Kim', 'Jonghun Park', 'Changjin Han', 'SeongEun Lee', 'Sungwook Jeon', 'Woo Young Kang', 'Sangwoong Yoon'] | 2020-12-29 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 5.40922523e-01 1.74057558e-01 -3.31434190e-01 -7.34902918e-01
-7.37581193e-01 -3.95623773e-01 8.22424054e-01 8.39576423e-01
-4.23544914e-01 1.88999679e-02 6.23090446e-01 1.16017476e-01
-1.82061821e-01 -7.79222190e-01 -7.79756486e-01 -1.01085111e-01
1.83820143e-01 3.50621015e-01 3.20740730e-01 -1.55390248e-01
6.87616646e-01 4.79522705e-01 -1.61149371e+00 5.19972563e-01
2.36229762e-01 1.20411634e+00 4.86672163e-01 7.64986336e-01
-1.80977166e-01 1.30187774e+00 -3.38559330e-01 -4.41524416e-01
1.73545241e-01 -4.34567571e-01 -1.15745354e+00 6.50665611e-02
1.26055384e+00 -2.69098133e-01 -7.03497231e-01 1.35546947e+00
1.79033965e-01 2.48405561e-01 8.66045237e-01 -1.18702066e+00
-1.26084268e+00 5.01515687e-01 -1.41993150e-01 3.36896896e-01
9.13027167e-01 -1.09139584e-01 1.26057196e+00 -7.26760328e-01
9.40044820e-01 1.47686410e+00 2.94715434e-01 1.15952127e-01
-1.05756617e+00 1.18789539e-01 -3.85539345e-02 4.67939854e-01
-1.26885509e+00 -1.07208200e-01 9.21607435e-01 -2.75095463e-01
9.92310643e-01 4.32526559e-01 6.53709412e-01 5.60025215e-01
2.44068712e-01 5.12014329e-01 7.95717835e-01 -4.20009106e-01
8.97639021e-02 1.59426592e-02 1.26579747e-01 7.82669604e-01
-3.97974066e-02 -2.36167744e-01 -3.35650623e-01 -2.42106779e-03
5.94544888e-01 1.30880639e-01 -2.05931559e-01 -5.90528667e-01
-1.06336117e+00 7.94099987e-01 1.14450526e+00 3.25588882e-01
-2.89796233e-01 4.69315112e-01 6.16508186e-01 3.93411487e-01
4.93843526e-01 6.92970395e-01 1.48258701e-01 5.38325369e-01
-8.07574689e-01 5.60252480e-02 5.75276375e-01 1.07995641e+00
1.01535773e+00 -2.81523347e-01 -6.17288291e-01 7.04145372e-01
4.11967099e-01 5.35976529e-01 4.10889775e-01 -1.11374283e+00
1.36962205e-01 8.76606643e-01 -3.45605314e-01 -1.87298560e+00
-1.11944959e-01 -1.02251887e-01 -7.06553996e-01 -2.57832140e-01
-1.78368427e-02 1.04332733e+00 -7.96897113e-01 1.58815372e+00
6.50972426e-02 -3.00766099e-02 5.63768148e-02 1.01968980e+00
1.44931924e+00 7.32894063e-01 3.55402529e-01 1.93257313e-02
1.27509356e+00 -1.13174415e+00 -4.54467118e-01 -2.54122168e-01
3.18871588e-01 -8.88794363e-01 1.39543092e+00 -9.63224247e-02
-1.30479646e+00 -8.06310594e-01 -9.38836873e-01 -3.35023612e-01
-5.96327960e-01 -1.81645945e-01 3.15744877e-01 2.29396656e-01
-1.66422689e+00 7.02635765e-01 -1.13281630e-01 -9.52278793e-01
2.69307315e-01 1.06921233e-01 -6.11060143e-01 -1.98674276e-01
-9.91321445e-01 8.90973747e-01 6.46831691e-01 -2.72147745e-01
-6.53619766e-01 -2.77680904e-01 -1.04275572e+00 7.40857869e-02
-4.49933819e-02 -9.19770181e-01 9.66939092e-01 -1.23641312e+00
-1.06585193e+00 1.66012239e+00 -2.61390917e-02 -4.64863807e-01
7.28093013e-02 8.04491118e-02 -1.41575754e-01 8.28600585e-01
2.70362735e-01 1.03313887e+00 8.25323164e-01 -1.32993507e+00
-3.25283587e-01 -1.91314593e-01 4.47266370e-01 3.86680841e-01
-3.08364391e-01 1.80841088e-01 -8.72062266e-01 -3.38356167e-01
2.68640995e-01 -7.22668648e-01 -2.43605971e-01 1.19161554e-01
-8.48287195e-02 -3.45435292e-01 6.82066977e-01 -4.17517185e-01
9.12520766e-01 -1.82642102e+00 -1.93345230e-02 1.70258656e-01
2.55839854e-01 8.27157348e-02 -7.05617547e-01 6.16927683e-01
-1.55474305e-01 1.54722184e-01 -9.58902910e-02 -1.22023471e-01
-1.05610549e-01 2.63469249e-01 -2.95259625e-01 3.66466582e-01
1.38383627e-01 1.31075084e+00 -1.40784717e+00 -9.95902121e-01
5.89577198e-01 2.36993074e-01 -4.18430537e-01 6.88293636e-01
-2.25792140e-01 6.86694533e-02 -4.61524367e-01 4.26042467e-01
5.46451330e-01 -5.51893055e-01 8.75205360e-03 -6.61392689e-01
4.54810500e-01 -7.60834548e-04 -6.14289641e-01 1.81655669e+00
-3.13083559e-01 9.82020259e-01 -5.35965264e-01 -1.17389894e+00
9.66567993e-01 -1.08121149e-01 3.14042836e-01 -1.22960508e+00
1.12645209e-01 -2.08731696e-01 -7.10611880e-01 -5.88769555e-01
9.01422381e-01 4.00755137e-01 -1.88323647e-01 4.25802380e-01
7.06784502e-02 -7.76426196e-01 1.77256003e-01 7.08381951e-01
1.09255028e+00 -6.16132207e-02 5.40097117e-01 -3.26718360e-01
8.80510747e-01 -1.75679494e-02 -3.48436326e-01 1.07572186e+00
-3.27976882e-01 8.28732729e-01 2.35712677e-01 -7.05097020e-01
-1.25869691e+00 -1.15072477e+00 3.48961085e-01 1.27072990e+00
7.58674085e-01 -4.91449505e-01 -9.92990553e-01 -4.96836632e-01
-2.32483894e-01 3.86950910e-01 -6.90177500e-01 -4.03472096e-01
-3.50952953e-01 1.38352126e-01 2.39269972e-01 2.43284255e-01
4.84659195e-01 -1.36449897e+00 -3.34464163e-01 -2.68421352e-01
-2.82809436e-01 -1.54997170e+00 -8.22389305e-01 -3.41818631e-01
-5.54013610e-01 -1.04973555e+00 -3.94037753e-01 -1.32602191e+00
8.30391645e-01 8.12105060e-01 1.83089006e+00 5.24464130e-01
-3.83136481e-01 1.03795862e+00 -5.15530169e-01 1.10401303e-01
-6.34177625e-01 -2.41063356e-01 -2.73661643e-01 -2.42833301e-01
1.09606609e-01 -2.93973535e-01 -7.52796412e-01 1.73190877e-01
-1.27429307e+00 1.43609457e-02 4.34773594e-01 4.09369588e-01
8.49174798e-01 -4.58693027e-01 -1.38533283e-02 -5.84560156e-01
8.57943654e-01 -1.68138146e-01 -4.76771176e-01 6.88126028e-01
-5.77382863e-01 3.10701847e-01 3.79245013e-01 -2.26919726e-01
-5.18886149e-01 1.06675707e-01 2.51558155e-01 -5.98314106e-01
-7.81174377e-02 5.38734853e-01 9.37873051e-02 -4.44245607e-01
8.61998558e-01 1.99249625e-01 -4.06507969e-01 2.04738021e-01
9.64721680e-01 5.21489799e-01 1.02490926e+00 -4.31919187e-01
6.51717842e-01 3.89175504e-01 2.19033748e-01 -6.89130664e-01
-1.01421225e+00 -8.17908347e-01 -7.98407435e-01 -6.16166592e-01
9.95760977e-01 -8.61364722e-01 -7.60125399e-01 7.19947144e-02
-1.52414572e+00 -1.18863486e-01 -2.51637548e-01 3.31784368e-01
-8.78216982e-01 7.85595000e-01 -4.84441280e-01 -3.33138943e-01
-6.83290660e-01 -9.63139772e-01 1.48392844e+00 1.95382386e-01
-1.38732016e-01 -9.83856618e-01 9.92560089e-02 4.34469789e-01
3.93979907e-01 4.11294580e-01 8.26497495e-01 -4.83296245e-01
-8.25761080e-01 -3.87914181e-01 -1.03828931e+00 2.88468838e-01
-2.44666692e-02 -1.66795254e-02 -7.77302086e-01 -7.29807466e-02
-1.03019424e-01 -6.14726007e-01 6.75523460e-01 4.62706655e-01
1.32755959e+00 -3.41956526e-01 -9.57710966e-02 3.96266788e-01
1.71757770e+00 -3.44840765e-01 7.49469101e-01 2.56378174e-01
8.45405579e-01 6.47452891e-01 5.65666318e-01 1.48572102e-01
5.46906948e-01 6.73559487e-01 6.90008461e-01 -2.09966004e-01
-3.85648102e-01 -4.25376087e-01 1.00537814e-01 8.84917378e-01
2.49368697e-01 -6.31443799e-01 -7.86709130e-01 4.38173562e-01
-1.91953099e+00 -1.09273899e+00 -3.27313602e-01 2.13719296e+00
5.90166032e-01 -5.47450781e-02 -1.92828581e-01 -5.21144211e-01
9.52567518e-01 4.34511274e-01 -1.87327087e-01 -7.28053689e-01
-2.16296583e-01 -8.01299661e-02 5.73536813e-01 5.45094848e-01
-1.08120227e+00 1.39182293e+00 7.47489691e+00 7.71899104e-01
-1.06008661e+00 -6.29003271e-02 6.86932921e-01 3.64416093e-01
-3.00383031e-01 1.51744317e-02 5.20058163e-02 7.76267890e-03
7.73969173e-01 -4.22283024e-01 4.41278338e-01 9.82130647e-01
-4.10304889e-02 -4.18184213e-02 -1.32195985e+00 1.39132082e+00
7.84335554e-01 -1.47133565e+00 6.68751001e-01 -4.14908797e-01
6.95968330e-01 2.63528109e-01 9.78056639e-02 -6.68065995e-03
2.29390278e-01 -9.22036409e-01 7.49006569e-01 7.37057686e-01
5.81774592e-01 -2.81612664e-01 6.03569627e-01 -1.98506355e-01
-1.39942348e+00 3.86033028e-01 -7.74506509e-01 -1.11175276e-01
-9.90101253e-05 2.75241971e-01 -1.02122974e+00 5.42837739e-01
5.96191704e-01 8.36114585e-01 -1.25427425e+00 8.27437162e-01
-2.75714278e-01 5.31394314e-03 6.43217713e-02 -2.23050907e-01
3.03564459e-01 8.71244911e-03 2.55712390e-01 1.21235263e+00
1.63065679e-02 -2.83930134e-02 3.40789914e-01 1.02128732e+00
-2.81475216e-01 4.83417362e-01 -1.09772921e+00 -1.86471954e-01
1.58111572e-01 1.55051315e+00 -1.00233674e+00 -5.37463188e-01
-2.51667112e-01 1.23808503e+00 5.17496049e-01 2.59761900e-01
-6.24612451e-01 -1.59597948e-01 -6.54510260e-02 -2.59620231e-02
-1.95501819e-01 3.48772928e-02 9.65764001e-02 -8.61490428e-01
4.97794859e-02 -5.53204417e-01 5.08827031e-01 -1.72061944e+00
-1.48743784e+00 1.03950560e+00 1.55839726e-01 -1.22324073e+00
-4.45382327e-01 -5.66003203e-01 -5.78593612e-01 5.52072585e-01
-1.30256689e+00 -1.47020698e+00 -8.41329873e-01 7.23339856e-01
3.52207392e-01 -4.67656218e-02 7.36899495e-01 -5.19533977e-02
1.63013831e-01 3.13301653e-01 -1.79845095e-01 1.00697830e-01
7.40662992e-01 -1.14863694e+00 6.96268022e-01 6.56284451e-01
4.76622999e-01 4.80892032e-01 8.71019304e-01 -5.55153131e-01
-1.33501232e+00 -1.10608602e+00 9.20177162e-01 -5.57943463e-01
6.24108613e-01 -3.04104295e-02 -8.45136166e-01 2.69336283e-01
6.67343795e-01 3.16533387e-01 4.13861573e-01 -3.92440259e-01
-7.66762733e-01 -1.89637896e-02 -1.01818454e+00 5.90973675e-01
1.21597815e+00 -1.13742006e+00 -7.97398508e-01 6.77946031e-01
1.06490946e+00 -1.98043287e-01 -8.59783351e-01 3.47093701e-01
4.00251776e-01 -9.52864528e-01 1.38186038e+00 -5.87252259e-01
7.93988228e-01 -1.41671643e-01 -6.85319424e-01 -9.50983644e-01
-4.14575964e-01 -1.28052786e-01 4.32959795e-01 7.87450314e-01
-4.36828146e-03 -7.31225461e-02 6.63858533e-01 6.73866451e-01
2.57714819e-02 -2.77394146e-01 -5.04202247e-01 -7.08978057e-01
-3.82465750e-01 -2.05319211e-01 3.58200401e-01 7.84371436e-01
-3.95660549e-02 5.36320210e-01 -1.03156991e-01 8.10938254e-02
6.45060539e-01 2.41052434e-01 8.16400707e-01 -1.20379496e+00
-9.83443037e-02 -6.70582354e-01 -1.14128363e+00 -7.71338224e-01
5.51191866e-01 -1.32695258e+00 1.04131833e-01 -1.90800035e+00
5.92633188e-01 1.47748068e-01 -3.46504360e-01 9.49740708e-02
-2.29424581e-01 6.73872292e-01 4.01698291e-01 2.83614784e-01
-1.43424141e+00 2.34403968e-01 1.20955551e+00 -6.13676369e-01
1.04001716e-01 -6.51759028e-01 -3.96335900e-01 6.16805673e-01
6.22301877e-01 -4.87381697e-01 -5.69548726e-01 -5.82405388e-01
4.90957528e-01 -2.09610406e-02 5.81870377e-01 -1.09579206e+00
4.21434224e-01 -3.20039600e-01 1.86797038e-01 -5.87886393e-01
2.17503220e-01 -7.96731770e-01 4.77508754e-02 4.30170208e-01
-8.70501339e-01 3.46876413e-01 -2.97877546e-02 6.63321555e-01
-4.58463401e-01 -3.97787541e-01 5.96109211e-01 -3.20287585e-01
-1.03921568e+00 4.92317855e-01 -2.10625622e-02 1.75569981e-01
6.74759984e-01 -3.38070154e-01 -4.29475725e-01 -8.82843673e-01
-4.11044359e-01 -5.88300824e-02 7.42141485e-01 6.79940104e-01
1.07117856e+00 -1.52866066e+00 -6.04788840e-01 -2.86487728e-01
8.21187794e-01 -6.16912842e-01 1.04631439e-01 1.25816658e-01
-7.75393248e-01 4.02717412e-01 -3.64986360e-01 -8.19493532e-01
-1.46790004e+00 1.04441023e+00 2.03661844e-01 -6.30332753e-02
-3.74105632e-01 6.93248093e-01 3.24607611e-01 -2.17866272e-01
1.97925463e-01 -4.41867501e-01 -3.86003971e-01 -2.48441339e-01
4.88032699e-01 -1.41962022e-01 -1.82598144e-01 -1.14197981e+00
-2.75059223e-01 9.32479084e-01 1.77908763e-01 6.76464438e-02
1.03097343e+00 -3.36995780e-01 -5.69065988e-01 2.03874543e-01
1.84852707e+00 -4.21119899e-01 -5.38514316e-01 -5.03112137e-01
1.08058244e-01 -5.20009637e-01 1.24695562e-01 -3.72584611e-01
-9.83129859e-01 7.47821331e-01 8.48976195e-01 3.31353515e-01
1.13942075e+00 4.91940856e-01 4.28441495e-01 8.14685285e-01
1.80866838e-01 -9.78635848e-01 7.76916385e-01 3.96378666e-01
1.13665736e+00 -1.50467944e+00 2.57147044e-01 -3.79642218e-01
-5.70004284e-01 1.09569263e+00 3.59258592e-01 -4.49479848e-01
3.18354517e-01 -5.85578024e-01 2.46021468e-02 -5.90088427e-01
-4.27722663e-01 -5.65864027e-01 9.12371278e-01 6.99641883e-01
3.58324170e-01 -5.37504554e-02 -1.87400863e-01 -9.69913602e-02
4.35761325e-02 -2.63289601e-01 3.71621102e-01 4.28064883e-01
-6.03562295e-01 -8.55346560e-01 -1.30145818e-01 3.32951784e-01
-1.24906160e-01 -3.79988581e-01 -8.55790854e-01 3.92412126e-01
-4.28353578e-01 9.99055922e-01 1.21430777e-01 -3.34714711e-01
3.69278431e-01 -5.39020538e-01 6.53961718e-01 -6.03301942e-01
-5.77896178e-01 -3.63523275e-01 -2.60621220e-01 -8.94444466e-01
-8.13631773e-01 6.84009269e-02 -1.06852150e+00 7.02700764e-02
-6.05150638e-03 -1.33768544e-02 6.35272682e-01 7.51426995e-01
2.98403203e-01 2.32944593e-01 6.16562128e-01 -9.13210332e-01
7.21045882e-02 -6.05164528e-01 -2.42256656e-01 1.17553926e+00
-4.20733728e-02 -1.21542618e-01 -1.60266802e-01 2.84309596e-01] | [10.729267120361328, 1.4149819612503052] |
3012b3a0-078c-4c77-8eb9-fbd0f55f3b2e | single-image-deep-defocus-estimation-and-its | 2107.14443 | null | https://arxiv.org/abs/2107.14443v2 | https://arxiv.org/pdf/2107.14443v2.pdf | Single image deep defocus estimation and its applications | Depth information is useful in many image processing applications. However, since taking a picture is a process of projection of a 3D scene onto a 2D imaging sensor, the depth information is embedded in the image. Extracting the depth information from the image is a challenging task. A guiding principle is that the level of blurriness due to defocus is related to the distance between the object and the focal plane. Based on this principle and the widely used assumption that Gaussian blur is a good model for defocus blur, we formulate the problem of estimating the spatially varying defocus blurriness as a Gaussian blur classification problem. We solved the problem by training a deep neural network to classify image patches into one of the 20 levels of blurriness. We have created a dataset of more than 500000 image patches of size $32\times32$ which are used to train and test several well-known network models. We find that MobileNetV2 is suitable for this application due to its low memory requirement and high accuracy. The trained model is used to determine the patch blurriness which is then refined by applying an iterative weighted guided filter. The result is a defocus map that carries the information of the degree of blurriness for each pixel. We compare the proposed method with state-of-the-art techniques and we demonstrate its successful applications in adaptive image enhancement, defocus magnification, and multi-focus image fusion. | ['Guang Deng', 'Fernando J. Galetto'] | 2021-07-30 | null | null | null | null | ['defocus-estimation'] | ['computer-vision'] | [ 5.25922418e-01 -5.79129100e-01 3.73419166e-01 -4.82527673e-01
-1.18756637e-01 -2.14578092e-01 2.20928416e-01 -3.69100839e-01
-4.56283033e-01 7.97524631e-01 2.64625460e-01 -7.10928589e-02
-1.77192569e-01 -5.71958005e-01 -6.05687022e-01 -9.58850265e-01
1.88997507e-01 -1.06753252e-01 2.41412893e-01 2.06965670e-01
6.53295517e-01 6.26270533e-01 -1.64706373e+00 1.77010074e-01
1.20608735e+00 1.15968275e+00 9.14146543e-01 8.73995364e-01
2.27215782e-01 8.21728885e-01 -8.10380578e-01 5.60563281e-02
1.35849267e-01 -3.00651133e-01 -5.69545925e-01 1.80923238e-01
5.54752111e-01 -7.87737668e-01 -5.03717840e-01 1.53481078e+00
5.72689712e-01 1.25696048e-01 6.31522655e-01 -6.88910067e-01
-5.99289715e-01 3.53779122e-02 -7.85087585e-01 6.76648676e-01
1.17869258e-01 -4.88023609e-02 1.02985814e-01 -8.37320268e-01
3.17535818e-01 8.82905543e-01 5.02573431e-01 4.04028445e-01
-7.65745580e-01 -3.15994948e-01 -5.66907525e-01 4.71851230e-01
-1.11428297e+00 -5.00659406e-01 7.65787542e-01 -4.60517228e-01
8.50965321e-01 2.06602529e-01 2.98853666e-01 5.01528263e-01
7.62063205e-01 3.04565221e-01 1.22562289e+00 -5.13873577e-01
2.92633414e-01 1.76759418e-02 2.63899028e-01 5.86046576e-01
3.73093456e-01 3.37740391e-01 -2.75904924e-01 1.37859166e-01
1.08588994e+00 3.30476820e-01 -9.96129870e-01 -1.14435166e-01
-1.05012846e+00 4.49862212e-01 7.29255557e-01 3.26241523e-01
-4.55261976e-01 1.39601037e-01 -2.20769599e-01 5.26839644e-02
5.31624675e-01 6.07934594e-01 -2.02818602e-01 -9.46856141e-02
-9.85392630e-01 3.85441445e-02 4.88146722e-01 4.38114524e-01
9.40546274e-01 -2.43230298e-01 -3.31131392e-03 9.83301103e-01
3.23545516e-01 6.00470126e-01 5.82242012e-01 -1.14718938e+00
3.16015258e-03 3.25944752e-01 3.29832762e-01 -9.68805730e-01
-1.92878082e-01 -2.12856591e-01 -1.10830557e+00 6.41473711e-01
3.71540755e-01 -2.61529684e-01 -1.12853706e+00 1.39521325e+00
1.20808661e-01 4.68716145e-01 2.03966741e-02 1.36195791e+00
8.45254779e-01 6.36523008e-01 -6.27145529e-01 -4.01080072e-01
1.24773574e+00 -9.08252358e-01 -9.15893316e-01 -3.34953010e-01
-5.62632792e-02 -1.09092510e+00 5.18264771e-01 3.59533250e-01
-1.16252649e+00 -7.79013693e-01 -1.22128665e+00 -6.57116175e-02
-1.09515443e-01 3.32057662e-02 5.06525695e-01 5.43304622e-01
-1.23989284e+00 4.80309576e-01 -6.68421030e-01 -8.29996020e-02
4.30715889e-01 2.47948453e-01 -1.53570339e-01 -5.04062176e-01
-1.07193446e+00 1.07042241e+00 1.70156136e-01 2.37339407e-01
-8.07711959e-01 -6.53817773e-01 -7.93610275e-01 1.08588301e-01
-7.38208964e-02 -7.31938839e-01 1.15304565e+00 -6.78584158e-01
-1.28186131e+00 7.14668751e-01 -4.79079098e-01 -4.61028636e-01
2.14978367e-01 -1.34427220e-01 -2.03685790e-01 3.76678199e-01
-8.77918825e-02 3.97481680e-01 1.28235090e+00 -1.19362617e+00
-7.94272006e-01 -4.99515414e-01 2.02825636e-01 4.31088150e-01
1.49317309e-01 -1.09964656e-02 -2.21697062e-01 -2.20369592e-01
1.88223317e-01 -4.63936895e-01 -8.12386796e-02 -1.79102466e-01
3.60334255e-02 2.05149621e-01 1.12816119e+00 -6.65534019e-01
1.08187640e+00 -2.23318863e+00 2.40715817e-02 -3.80301774e-01
5.42698026e-01 3.95442873e-01 1.14569351e-01 -2.29853615e-02
-1.69229507e-02 -3.78099561e-01 -4.27501559e-01 -1.75893784e-01
-4.71755505e-01 -8.78764465e-02 -1.47676930e-01 6.16543651e-01
-6.82001337e-02 7.46226847e-01 -8.20525169e-01 -1.81245476e-01
6.09056294e-01 7.22220659e-01 -3.06733131e-01 6.80175483e-01
4.00078902e-03 3.28437686e-01 -7.08930492e-02 4.27011132e-01
1.37312639e+00 -2.57851213e-01 -4.84578073e-01 -5.39518893e-01
-3.83984894e-01 -1.27203763e-01 -8.60429525e-01 1.61856043e+00
-3.75002205e-01 1.01378477e+00 2.13912144e-01 -6.74910367e-01
7.76712298e-01 7.32948631e-02 2.67348856e-01 -3.93143862e-01
3.31981033e-01 2.99439698e-01 5.26384264e-03 -9.74831343e-01
4.78824288e-01 -2.30592027e-01 4.10377294e-01 3.57360601e-01
-9.26612318e-02 -6.65055037e-01 2.48955376e-02 -2.00829223e-01
1.07778609e+00 -3.20235103e-01 7.94391930e-02 -2.68862724e-01
4.82518405e-01 -3.42736185e-01 7.82309920e-02 8.04594874e-01
-3.83047432e-01 8.55372787e-01 -1.07441191e-02 -4.62127149e-01
-9.56926286e-01 -9.21953976e-01 -3.29576701e-01 2.52179593e-01
7.03747451e-01 3.60390514e-01 -1.11137366e+00 -3.14572245e-01
-1.83214024e-01 6.09121859e-01 -4.93187726e-01 -1.79588154e-01
-3.74076068e-01 -8.83903861e-01 -5.77542409e-02 4.10697907e-02
1.16659558e+00 -1.05125117e+00 -1.01883519e+00 -1.04652643e-01
-3.44621539e-01 -9.42110777e-01 -6.37169361e-01 7.41045699e-02
-6.78259432e-01 -1.28740060e+00 -1.02935243e+00 -8.76389027e-01
8.37217033e-01 1.00707650e+00 8.07719409e-01 -1.18171930e-01
-3.73709381e-01 1.66399240e-01 -2.92365924e-02 -3.54613334e-01
-3.45658839e-01 -5.23234665e-01 -4.41356413e-02 9.25827622e-02
5.30265331e-01 -4.24328923e-01 -1.06916964e+00 1.65090069e-01
-1.11562288e+00 1.55729428e-01 6.44624412e-01 8.59225214e-01
1.31762013e-01 7.58283377e-01 1.16317615e-01 -4.11087841e-01
8.04922998e-01 -9.01333392e-02 -7.21941650e-01 -1.11501530e-01
-4.09098595e-01 -1.28620639e-01 1.24415360e-01 -3.10348153e-01
-1.51056480e+00 -2.67738372e-01 4.67692204e-02 -6.19164109e-01
-4.49220955e-01 4.80147213e-01 2.87542846e-02 -4.02414024e-01
6.71300054e-01 1.79138452e-01 2.10663736e-01 -3.92910510e-01
2.90628579e-02 1.11556423e+00 6.87866926e-01 1.57252401e-01
3.23752999e-01 5.04027188e-01 -3.13456468e-02 -1.11044741e+00
-8.48630309e-01 -4.75017071e-01 -4.87293541e-01 -4.08152223e-01
1.10236633e+00 -8.03595126e-01 -7.80154824e-01 1.19934845e+00
-1.48556566e+00 -2.67973572e-01 9.06156152e-02 6.35829389e-01
-3.79110783e-01 6.46157742e-01 -6.61092043e-01 -6.53730154e-01
-1.88827515e-01 -1.40707600e+00 7.70696640e-01 6.28929615e-01
3.19976300e-01 -1.00900388e+00 -6.86247796e-02 4.63867038e-01
7.39799917e-01 -2.79573891e-02 8.23033690e-01 3.33633095e-01
-8.98965299e-01 -3.83215700e-03 -7.22154558e-01 6.87615514e-01
6.75190568e-01 -4.47990239e-01 -1.21506321e+00 -2.64904499e-01
9.67730165e-01 6.72672987e-02 1.10578787e+00 1.07521069e+00
1.24178851e+00 -6.37005195e-02 -1.76106572e-01 8.17458093e-01
1.74156237e+00 5.14084458e-01 9.81916249e-01 1.85082033e-01
6.24422073e-01 2.26491347e-01 4.86576915e-01 7.84079209e-02
-2.96923425e-02 5.82969368e-01 6.46188915e-01 -1.92584142e-01
-4.20832902e-01 2.18125686e-01 2.46004164e-02 5.14843285e-01
7.24107400e-03 -3.69128257e-01 -6.69106066e-01 5.16433299e-01
-1.51969814e+00 -1.05983233e+00 -1.70259938e-01 2.24955106e+00
8.91209602e-01 -1.27912134e-01 -6.28832757e-01 -5.19507332e-03
8.88965845e-01 2.14790747e-01 -5.03655791e-01 -2.17225283e-01
-1.62762940e-01 7.29196891e-02 5.17299056e-01 9.76087391e-01
-1.06998146e+00 5.84641635e-01 6.42724991e+00 6.20980561e-01
-1.43828189e+00 -1.46847412e-01 6.81417704e-01 1.73629716e-01
3.18776146e-02 -4.17964399e-01 -5.29180646e-01 9.23598349e-01
6.25400186e-01 5.42672835e-02 9.05129492e-01 3.04968297e-01
4.16604161e-01 -7.73083866e-01 -9.30608630e-01 1.48450637e+00
3.92800152e-01 -1.18487024e+00 -2.39234716e-01 -5.18979281e-02
8.98862422e-01 1.49927050e-01 1.95453897e-01 -5.10474086e-01
-1.37564495e-01 -1.00813532e+00 2.82329261e-01 8.68721187e-01
9.26128864e-01 -4.42971319e-01 9.66422439e-01 4.50968593e-01
-5.00156522e-01 -8.15178379e-02 -6.57151163e-01 -2.32192039e-01
1.93608746e-01 1.07482064e+00 -7.44298220e-01 2.63319254e-01
8.85711908e-01 7.76583791e-01 -3.25091988e-01 1.45313323e+00
-1.80328935e-01 1.07076213e-01 -6.07010201e-02 1.91365615e-01
-8.49513933e-02 -4.13906455e-01 5.26877761e-01 1.01964426e+00
7.45978832e-01 1.88658506e-01 -6.64028943e-01 9.90290999e-01
2.02390435e-03 -7.47322261e-01 -5.93344271e-01 3.60867679e-01
2.72502571e-01 1.11408305e+00 -3.28816146e-01 -3.54493469e-01
-3.88492823e-01 1.19336438e+00 -3.05920877e-02 5.09802401e-01
-6.68650627e-01 -6.09293103e-01 7.98067391e-01 -9.18860957e-02
4.88012344e-01 -9.98531058e-02 -3.08305353e-01 -1.19836652e+00
-1.17054306e-01 -5.95327854e-01 -2.19054833e-01 -1.53598261e+00
-1.09366453e+00 7.85216987e-01 -1.24373823e-01 -1.02338326e+00
2.30035633e-02 -8.65279555e-01 -5.22834420e-01 1.40855694e+00
-1.82305396e+00 -5.55508912e-01 -9.90902603e-01 4.43982512e-01
7.06901968e-01 7.05979988e-02 4.70391929e-01 2.22781911e-01
-2.48226911e-01 -2.00626954e-01 2.26748168e-01 -3.14215630e-01
7.69034266e-01 -1.14968836e+00 3.66711318e-02 1.06186092e+00
-3.98987234e-01 4.10430282e-01 8.94289851e-01 -5.50289810e-01
-1.12455678e+00 -9.51719820e-01 8.72361243e-01 -4.36383039e-01
3.63511592e-01 5.08513825e-04 -9.59113419e-01 1.30380258e-01
5.68022907e-01 -7.44370595e-02 8.89556184e-02 -6.96187615e-01
2.06531361e-01 -1.92008376e-01 -1.36752546e+00 1.95668101e-01
7.57168055e-01 -4.98689651e-01 -7.04058826e-01 8.36109444e-02
7.84650385e-01 -9.10724044e-01 -4.93744850e-01 5.19807398e-01
2.79814333e-01 -1.46112812e+00 8.48200500e-01 1.69484913e-01
7.52106845e-01 -4.99128878e-01 -7.78176114e-02 -1.68354332e+00
-4.46976721e-01 -3.56838346e-01 -1.62125751e-01 6.71556830e-01
-4.16800827e-02 -7.29195952e-01 6.91017091e-01 4.14012492e-01
-2.98610687e-01 -6.00762069e-01 -7.08437860e-01 -1.69802666e-01
-4.98074114e-01 2.01781876e-02 5.46756446e-01 7.10712552e-01
-2.87798673e-01 2.86749303e-01 -2.90601432e-01 5.08818328e-01
9.32314575e-01 6.06002286e-02 3.52452368e-01 -1.04950714e+00
-7.76128620e-02 -2.68996984e-01 -3.35993439e-01 -1.51158237e+00
-2.04969600e-01 -3.13734829e-01 3.09331030e-01 -1.59216535e+00
4.19938177e-01 -2.37252101e-01 -2.89517701e-01 -2.42187664e-01
-3.00209403e-01 3.35178137e-01 -1.94060922e-01 3.51186723e-01
-1.34620950e-01 2.51112849e-01 1.50301039e+00 -2.15478942e-01
-2.48642519e-01 2.21980244e-01 -6.54608428e-01 6.18049681e-01
5.44389129e-01 4.94069643e-02 -5.35684764e-01 -9.07555401e-01
-1.23669341e-01 1.62383914e-01 5.71019411e-01 -1.07382739e+00
4.81558084e-01 -8.13746899e-02 5.92328846e-01 -7.44039476e-01
5.30039549e-01 -9.98183072e-01 1.52390718e-01 2.46876702e-01
-5.36648966e-02 -3.84272754e-01 1.16383985e-01 4.03806686e-01
-3.89438868e-01 -6.32041276e-01 1.05910254e+00 -2.46288627e-01
-7.09816694e-01 1.47153407e-01 -2.52423018e-01 -4.62073088e-01
7.24480867e-01 -3.66698742e-01 -7.18205452e-01 -3.84998530e-01
-2.57853061e-01 -2.93669641e-01 7.30043232e-01 1.51902169e-01
1.10829771e+00 -1.06547379e+00 -4.81797695e-01 6.93078458e-01
-1.93285197e-01 1.67199105e-01 5.16445756e-01 8.24234128e-01
-8.70165348e-01 3.33764672e-01 -2.32583389e-01 -8.02954495e-01
-1.32236612e+00 5.74961543e-01 8.08906674e-01 3.51859391e-01
-1.96128190e-01 1.06350505e+00 5.35004497e-01 2.97115564e-01
1.08073294e-01 -5.23307323e-01 -2.33709455e-01 -6.48103952e-01
9.03302550e-01 3.68630499e-01 4.82647866e-02 -4.11961377e-01
-1.26622349e-01 7.70674527e-01 1.62587110e-02 1.29588842e-01
1.31943464e+00 -5.07163048e-01 -5.93050778e-01 1.80539146e-01
1.23911345e+00 -4.94518019e-02 -1.53997147e+00 -1.08384974e-01
-6.49550259e-01 -9.42815959e-01 4.60260510e-01 -8.82481992e-01
-9.62955773e-01 9.80445743e-01 9.93785381e-01 4.61984694e-01
1.57794940e+00 -7.91021064e-02 5.54431438e-01 2.20749602e-02
2.59441286e-01 -7.06636131e-01 2.86403187e-02 4.47604895e-01
9.22106028e-01 -1.25217187e+00 -8.58967304e-02 -3.59617084e-01
-1.18612245e-01 1.02926099e+00 4.84225780e-01 -1.17885418e-01
7.75841236e-01 2.98418403e-01 2.52909213e-01 -2.83047259e-01
-3.26376319e-01 1.27777070e-01 2.84058183e-01 6.37842119e-01
1.91633314e-01 -2.59337723e-01 -4.27939836e-03 2.20578928e-02
-6.62990063e-02 2.63577282e-01 7.39723325e-01 5.34342706e-01
-8.99017215e-01 -5.30930698e-01 -5.86990118e-01 3.92279327e-01
-3.02767605e-01 -7.84640163e-02 4.69054542e-02 1.43836915e-01
4.60382700e-01 1.05606771e+00 2.11829513e-01 -1.45914644e-01
1.10557064e-01 -4.64630872e-01 8.51524472e-01 -2.43442893e-01
1.22708552e-01 1.09358869e-01 -3.97576034e-01 -2.83169597e-01
-7.27088928e-01 -4.08319771e-01 -7.53296971e-01 -2.49203637e-01
-4.41320002e-01 4.54343818e-02 8.25487733e-01 8.21440816e-01
3.15437615e-01 4.84513968e-01 7.67004550e-01 -1.16064644e+00
-2.06884950e-01 -1.23679340e+00 -7.39705622e-01 4.06648248e-01
1.01915193e+00 -6.41731739e-01 -8.20178807e-01 2.37017676e-01] | [11.392026901245117, -2.747540235519409] |
9771846e-f7a5-4e41-b446-d05ff2bf2c53 | a-research-platform-for-multi-robot-dialogue-1 | 1910.05624 | null | https://arxiv.org/abs/1910.05624v1 | https://arxiv.org/pdf/1910.05624v1.pdf | A Research Platform for Multi-Robot Dialogue with Humans | This paper presents a research platform that supports spoken dialogue interaction with multiple robots. The demonstration showcases our crafted MultiBot testing scenario in which users can verbally issue search, navigate, and follow instructions to two robotic teammates: a simulated ground robot and an aerial robot. This flexible language and robotic platform takes advantage of existing tools for speech recognition and dialogue management that are compatible with new domains, and implements an inter-agent communication protocol (tactical behavior specification), where verbal instructions are encoded for tasks assigned to the appropriate robot. | ['Jesse Bloecker', 'Stephanie M. Lukin', 'Matthew Marge', 'Clare Voss', 'Eric Holder', 'Stephen Nogar', 'Cory J. Hayes'] | 2019-10-12 | a-research-platform-for-multi-robot-dialogue | https://aclanthology.org/N19-4023 | https://aclanthology.org/N19-4023.pdf | naacl-2019-6 | ['dialogue-management'] | ['natural-language-processing'] | [-2.16442525e-01 6.03996158e-01 3.02349478e-01 -5.82757413e-01
-2.04703629e-01 -8.53193760e-01 7.88261890e-01 -3.31945205e-03
-4.71735865e-01 1.11945331e+00 -3.82300206e-02 -3.70914519e-01
-5.98411381e-01 -3.63755524e-01 1.72180101e-01 -2.25082323e-01
-1.34946495e-01 9.97379124e-01 3.78440320e-01 -1.07739627e+00
2.11843804e-01 4.99171793e-01 -1.62618494e+00 1.05581813e-01
4.04203475e-01 5.32281697e-01 9.34750915e-01 1.12686944e+00
-1.06213577e-01 1.32287705e+00 -9.26056743e-01 3.38934690e-01
4.96242717e-02 -1.15543552e-01 -1.67124188e+00 2.45432466e-01
-4.52622294e-01 -6.42093599e-01 -9.00053009e-02 5.72992325e-01
1.34901196e-01 4.48891371e-01 4.73390371e-01 -1.92290580e+00
-9.79460077e-04 8.58027220e-01 3.64482403e-01 -5.39321303e-01
1.17075491e+00 2.82441884e-01 6.04738116e-01 -8.89286250e-02
6.88357770e-01 1.57700276e+00 3.90561342e-01 6.02866948e-01
-5.12498558e-01 -1.29534021e-01 -2.73465604e-01 4.32507545e-02
-1.17144847e+00 -2.94630110e-01 6.22859038e-02 -4.86457884e-01
1.23540652e+00 2.18849733e-01 1.79597661e-01 1.11997771e+00
3.81836355e-01 2.32164115e-01 9.12503064e-01 -5.03769994e-01
1.23392165e-01 6.18194282e-01 3.77339900e-01 8.21742594e-01
-3.81333649e-01 -1.66609362e-01 -3.76047492e-01 -4.08869982e-01
7.79607773e-01 -5.91564715e-01 -1.67213142e-01 -5.14321387e-01
-1.38156402e+00 6.70806170e-01 -3.27775419e-01 6.43440723e-01
-6.38334095e-01 5.32223992e-02 7.04836071e-01 9.03840184e-01
-4.73257184e-01 5.92662752e-01 -5.47316968e-01 -8.17857623e-01
4.44248706e-01 1.98418081e-01 1.72185218e+00 1.21389961e+00
7.45540321e-01 -9.94281024e-02 1.35592163e-01 9.34347451e-01
5.99748611e-01 4.50642347e-01 5.62316835e-01 -1.73570359e+00
2.08813950e-01 4.47157860e-01 8.77717018e-01 -7.41456151e-01
-7.42494583e-01 7.63702512e-01 1.28646299e-01 7.01920927e-01
7.81753436e-02 -6.29734874e-01 1.62994899e-02 1.04233599e+00
4.21802551e-01 -7.40899146e-01 1.02471864e+00 8.94995928e-01
1.13466907e+00 6.75292671e-01 5.35916276e-02 4.25026827e-02
1.59401417e+00 -9.61589217e-01 -8.63935530e-01 -2.62207389e-01
1.10112572e+00 -5.81138134e-01 1.10545158e+00 5.16380012e-01
-8.17132115e-01 -3.11493397e-01 -9.52047050e-01 1.35141388e-01
-3.55099916e-01 1.17707439e-01 7.35186517e-01 2.94730514e-01
-1.44385540e+00 2.34228581e-01 -4.42351848e-01 -9.47398126e-01
-1.01844764e+00 6.89053118e-01 -6.18261933e-01 2.18061551e-01
-1.19031024e+00 1.56312680e+00 5.34732997e-01 -9.25145522e-02
-9.25167441e-01 5.49883127e-01 -1.18883371e+00 -1.59085557e-01
4.72004563e-01 -4.51787114e-01 1.96626925e+00 -5.25752783e-01
-2.58919382e+00 5.93522489e-01 8.29914749e-01 -2.13148206e-01
1.18220240e-01 -1.21306129e-01 -1.76265672e-01 3.73875469e-01
3.55634332e-01 6.79163337e-01 2.26683110e-01 -1.34416580e+00
-8.37768912e-01 4.57856320e-02 7.49061048e-01 6.72603905e-01
2.48289332e-01 5.49610317e-01 2.54212111e-01 3.05627882e-01
-3.14503580e-01 -9.44798112e-01 -1.76388726e-01 -4.98916149e-01
-1.86745018e-01 -1.78729430e-01 7.32076883e-01 -6.26632422e-02
1.57122537e-01 -2.14929533e+00 4.21691746e-01 -5.57405837e-02
-2.28760794e-01 -1.01685077e-01 -2.87556589e-01 1.10955191e+00
4.22300190e-01 -3.10752958e-01 2.46482044e-01 -3.48754257e-01
7.77574480e-01 6.22137368e-01 -1.88044265e-01 1.35823429e-01
-5.63127279e-01 -3.36012542e-02 -7.46092916e-01 -5.95656872e-01
5.96649170e-01 -6.02453668e-03 3.78521159e-02 4.93768185e-01
-2.39270478e-01 2.57397056e-01 -8.31285775e-01 1.85915634e-01
1.69177711e-01 3.15659970e-01 6.52014792e-01 6.12153351e-01
-8.23401213e-01 3.66385370e-01 -1.22994626e+00 1.72763431e+00
-8.00176144e-01 3.28809917e-01 1.11228645e+00 -4.85146642e-01
1.20191061e+00 7.50315487e-01 2.16754735e-01 1.42261848e-01
4.50153381e-01 3.30409497e-01 -8.67107064e-02 -8.41680110e-01
7.52171278e-01 4.02595133e-01 -7.41806030e-01 8.09352279e-01
4.42724973e-01 -9.29166377e-01 -6.62411228e-02 2.23052539e-02
1.47393072e+00 4.24151182e-01 5.69520772e-01 8.63187015e-02
5.95023334e-01 6.00898147e-01 -9.46051851e-02 9.40931976e-01
-4.42102998e-01 -3.58201236e-01 3.98430616e-01 -3.68152022e-01
-4.39133376e-01 -6.07283771e-01 6.01052046e-01 1.46631002e+00
3.49461585e-01 -3.93837243e-01 -5.34358799e-01 -4.48020190e-01
-4.46292341e-01 9.90684867e-01 -1.58533314e-03 2.30610251e-01
-2.68282205e-01 2.67585218e-01 7.23557949e-01 -4.15412486e-01
6.04768097e-01 -1.59083712e+00 -1.49612880e+00 3.19763333e-01
-1.96862742e-01 -1.42825055e+00 -7.90674537e-02 4.02271569e-01
5.72089776e-02 -8.96358848e-01 -4.62653115e-02 -1.18337631e+00
4.54066209e-02 3.76881629e-01 4.64598656e-01 1.04478158e-01
-8.41496214e-02 1.40631187e+00 -8.58126163e-01 -3.70561004e-01
-1.15443993e+00 -1.19024098e-01 3.76844049e-01 -3.47190857e-01
-2.56703824e-01 -4.48940247e-01 1.96574986e-01 6.96903288e-01
-4.41267610e-01 9.31266174e-02 3.67299855e-01 4.93504971e-01
-4.76232797e-01 -3.45847934e-01 3.02566975e-01 7.37769082e-02
1.38789284e+00 -2.03512892e-01 -4.29801732e-01 5.97145557e-01
5.41511411e-03 -1.48261949e-01 3.64262611e-01 -2.48201340e-01
-7.62290895e-01 4.08089280e-01 2.56440729e-01 2.54768997e-01
-9.34681535e-01 4.32045907e-01 -5.17784953e-02 -4.24720407e-01
4.05383378e-01 2.20555216e-01 6.17566228e-01 1.00206494e-01
2.17913270e-01 1.35426009e+00 5.92880130e-01 -6.42529309e-01
8.93361494e-02 -1.29489407e-01 -3.44633013e-01 -1.18991554e+00
2.87070662e-01 -3.84553730e-01 -6.80583656e-01 -6.86778188e-01
9.69100833e-01 -6.69506431e-01 -1.69996727e+00 4.02184635e-01
-1.82074475e+00 -9.75697398e-01 6.14527017e-02 6.37607396e-01
-1.18292570e+00 5.63653111e-01 -6.52682185e-01 -1.23358059e+00
-1.43925026e-01 -1.30150485e+00 8.63511026e-01 1.78458720e-01
-8.01422954e-01 -7.32889235e-01 1.48152551e-02 1.57663226e-01
3.99008304e-01 -1.44618392e-01 6.18355274e-01 -9.19910789e-01
4.19719033e-02 2.56667044e-02 1.47234365e-01 -7.64407068e-02
1.01579987e-01 -1.43524438e-01 -2.89235651e-01 1.38228208e-01
1.80000335e-01 -1.03792405e+00 -5.54177046e-01 -2.01529264e-01
-1.73788562e-01 -3.48212719e-01 -3.70549947e-01 -4.75536168e-01
6.52849019e-01 7.23537207e-01 3.64651084e-01 6.64908588e-01
-1.12112574e-01 1.33526945e+00 1.27531981e+00 6.85470283e-01
7.49940813e-01 1.04393363e+00 4.16145474e-01 2.74424225e-01
5.43596089e-01 3.67777020e-01 8.89805973e-01 4.30479854e-01
2.01348349e-01 -3.11153889e-01 -1.11145175e+00 -4.04345095e-02
-2.15737343e+00 -7.91412234e-01 -7.75304064e-02 1.53906238e+00
3.78178954e-01 -2.78705001e-01 1.42848521e-01 -2.06559539e-01
5.42523265e-01 -1.06859244e-01 5.45246676e-02 -9.97532904e-01
3.29989642e-01 -2.93133169e-01 2.70927072e-01 1.08757353e+00
-9.77217793e-01 1.13063884e+00 6.34930277e+00 1.17977820e-01
-7.37829387e-01 2.40652338e-01 -4.10582364e-01 4.59823012e-01
4.15664673e-01 -4.63710912e-02 -5.02758205e-01 -3.40178490e-01
1.15313196e+00 -3.36921997e-02 7.23466337e-01 1.04150498e+00
4.25902575e-01 -6.87120140e-01 -9.03091311e-01 7.59529889e-01
4.70546372e-02 -1.01943707e+00 -4.16640103e-01 -1.51438951e-01
-4.71395224e-01 1.64770391e-02 -4.89967376e-01 5.55546582e-01
1.04749155e+00 -6.60746992e-01 6.96195006e-01 3.28285933e-01
4.10279751e-01 -4.56236959e-01 8.92783642e-01 9.36823666e-01
-8.75219643e-01 -3.63135606e-01 -1.09382249e-01 -4.10178512e-01
4.76410270e-01 -8.71288419e-01 -1.52574718e+00 6.59906626e-01
6.54491723e-01 -1.60075724e-01 3.65186818e-02 5.84332168e-01
-1.68260679e-01 -4.03487891e-01 -2.58950442e-01 -8.91386986e-01
4.65500236e-01 -5.47292411e-01 6.87045872e-01 8.35248590e-01
4.20133233e-01 4.61650193e-01 6.89072907e-01 4.27244246e-01
9.93211508e-01 1.50792331e-01 -1.14991462e+00 1.94698691e-01
6.72437251e-01 1.30971992e+00 -8.48607957e-01 -4.10013460e-02
-1.61073267e-01 1.29440236e+00 -1.09639123e-01 -3.13045233e-02
-6.20377958e-01 -1.11941075e+00 6.14335716e-01 -5.64513683e-01
-6.70441836e-02 -7.18754113e-01 4.57477748e-01 -1.88670263e-01
-6.21354997e-01 -1.04721725e+00 -4.31673042e-02 -1.40957785e+00
-5.74313581e-01 1.06040227e+00 3.51458251e-01 -1.12943351e+00
-9.57789838e-01 -6.93561614e-01 -5.36664009e-01 4.58820879e-01
-4.06518072e-01 -1.19048274e+00 -4.63580787e-01 3.92074436e-01
6.16989851e-01 -5.97924888e-01 1.71048689e+00 -3.73392463e-01
-1.84833974e-01 -4.03208703e-01 -5.87977648e-01 -1.97263911e-01
6.39537096e-01 -1.00938785e+00 -2.51197964e-01 -2.18777224e-01
-5.19717336e-01 4.38445687e-01 9.64984715e-01 -4.76002574e-01
-1.76929343e+00 -2.05658957e-01 5.44724762e-01 -1.36633381e-01
8.49303663e-01 -2.64487296e-01 -3.60492617e-01 8.31570864e-01
7.82808721e-01 -6.51849747e-01 6.26439393e-01 -1.99977458e-01
4.34674442e-01 1.98610425e-01 -1.25675750e+00 6.75888777e-01
5.96516788e-01 -3.92420650e-01 -8.68444741e-01 6.71885848e-01
9.67671573e-01 -5.89631796e-01 -6.42524540e-01 -1.19879805e-02
3.62063408e-01 -8.02882850e-01 4.32554305e-01 -3.94034356e-01
-1.89176634e-01 -3.61668587e-01 -3.09637606e-01 -1.51058745e+00
2.72168308e-01 -1.28256202e+00 7.09875226e-01 1.07195890e+00
3.97534847e-01 -9.23148811e-01 8.29928219e-02 9.99845088e-01
-6.21975064e-01 3.98766249e-01 -1.15017307e+00 -6.27091706e-01
-5.80548525e-01 -2.49471530e-01 4.74514663e-01 6.72908127e-01
1.24537456e+00 7.86741316e-01 -2.70298988e-01 4.75950986e-01
-3.72817934e-01 -4.38147396e-01 1.19091380e+00 -1.26943338e+00
-2.83377111e-01 -2.06045106e-01 -4.09245282e-01 -1.06311643e+00
6.31975591e-01 -4.47681367e-01 8.57639015e-01 -1.83109629e+00
-7.63417602e-01 -3.69678497e-01 7.12269187e-01 6.92598939e-01
1.02071810e+00 -7.68646538e-01 3.12567770e-01 1.31924152e-01
-9.92413819e-01 4.58990693e-01 7.57683754e-01 1.25922263e-01
-6.11147404e-01 6.25439435e-02 -1.67957962e-01 7.46318936e-01
8.45282137e-01 -2.03646168e-01 -4.13178593e-01 -1.36131600e-01
-6.73603639e-02 1.08492708e+00 2.54006058e-01 -1.19000840e+00
8.75545323e-01 -4.39757228e-01 -6.13173723e-01 3.10372636e-02
7.82976449e-01 -1.12518167e+00 2.36509636e-01 8.71374190e-01
-4.73806709e-01 3.68109375e-01 2.11786419e-01 1.45126641e-01
-9.63491127e-02 -7.36752093e-01 2.06182078e-01 -2.82767087e-01
-1.06670809e+00 -6.94318056e-01 -1.77409756e+00 -9.94901717e-01
1.76946580e+00 -8.81705731e-02 -7.21274614e-01 -1.14172709e+00
-6.66675925e-01 6.14295483e-01 3.04757476e-01 3.84925842e-01
9.05224264e-01 -5.39777398e-01 -2.94169456e-01 -8.63038972e-02
2.56112162e-02 -5.39763212e-01 -5.27907163e-02 5.20562291e-01
-9.26064551e-01 4.30713385e-01 -6.50610447e-01 -3.63084078e-01
-1.50916934e+00 1.21335499e-02 5.19906700e-01 3.87468845e-01
-3.90794098e-01 3.70094001e-01 -5.04496455e-01 -1.20689678e+00
3.32957089e-01 -3.10633332e-01 -6.99799955e-01 -1.74022943e-01
5.50103664e-01 2.52170026e-01 -2.94820368e-01 -6.65134490e-01
-4.02853787e-01 2.52280891e-01 3.61868680e-01 -8.74765217e-01
1.14579391e+00 -4.49991047e-01 -5.11681795e-01 8.92989814e-01
2.91140676e-01 -1.92611843e-01 -5.87878644e-01 3.63776058e-01
2.36721009e-01 3.27597857e-01 -5.35189271e-01 -8.60032082e-01
2.98318326e-01 1.67950094e-01 3.42859626e-01 8.49081874e-01
2.10891858e-01 1.26065463e-01 1.67180970e-01 1.42616487e+00
1.33743143e+00 -1.25764287e+00 3.04834038e-01 1.27375519e+00
1.43234038e+00 -9.98347461e-01 -3.13236088e-01 -6.10884130e-01
-1.36389363e+00 1.50779939e+00 9.63848770e-01 3.42269540e-01
2.73246437e-01 6.28292799e-01 6.91154301e-01 -4.68604058e-01
-1.04307890e+00 -1.49270236e-01 -7.71477103e-01 1.11107969e+00
-6.37154207e-02 1.48102315e-02 -1.74619377e-01 4.88920391e-01
-5.21444142e-01 -1.40089631e-01 1.23844373e+00 1.43011475e+00
-1.01052177e+00 -1.05780041e+00 -7.30347991e-01 -2.43675455e-01
2.25449085e-01 5.12951076e-01 -9.98313904e-01 1.08957732e+00
-2.50580549e-01 1.61036384e+00 7.42612127e-03 -5.62718153e-01
7.26702690e-01 9.56996903e-02 1.51147351e-01 -1.03699183e+00
-1.03650701e+00 -5.56707561e-01 8.25492382e-01 -5.41051209e-01
-3.56673926e-01 -6.86089337e-01 -1.84634566e+00 -1.38429366e-03
2.10893720e-01 7.43435681e-01 9.97179806e-01 9.17448878e-01
2.35538110e-01 2.40556762e-01 5.83684742e-01 -1.24821711e+00
-7.89885521e-01 -1.32594013e+00 -5.96959293e-01 -3.61329526e-01
1.96085840e-01 -6.33091748e-01 -3.18868816e-01 -3.50362897e-01] | [4.441879749298096, 0.6466375589370728] |
55b7cf8e-b479-4b1e-9fdb-ad5bd47395cc | from-shallow-to-deep-compositional-reasoning | 2206.12533 | null | https://arxiv.org/abs/2206.12533v1 | https://arxiv.org/pdf/2206.12533v1.pdf | From Shallow to Deep: Compositional Reasoning over Graphs for Visual Question Answering | In order to achieve a general visual question answering (VQA) system, it is essential to learn to answer deeper questions that require compositional reasoning on the image and external knowledge. Meanwhile, the reasoning process should be explicit and explainable to understand the working mechanism of the model. It is effortless for human but challenging for machines. In this paper, we propose a Hierarchical Graph Neural Module Network (HGNMN) that reasons over multi-layer graphs with neural modules to address the above issues. Specifically, we first encode the image by multi-layer graphs from the visual, semantic and commonsense views since the clues that support the answer may exist in different modalities. Our model consists of several well-designed neural modules that perform specific functions over graphs, which can be used to conduct multi-step reasoning within and between different graphs. Compared to existing modular networks, we extend visual reasoning from one graph to more graphs. We can explicitly trace the reasoning process according to module weights and graph attentions. Experiments show that our model not only achieves state-of-the-art performance on the CRIC dataset but also obtains explicit and explainable reasoning procedures. | ['Zihao Zhu'] | 2022-06-25 | null | null | null | null | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [-4.09207791e-02 6.78397000e-01 4.68915440e-02 -4.42094922e-01
-8.06145668e-02 -5.95914125e-01 3.67371798e-01 1.34150982e-02
1.11659102e-01 2.70021498e-01 3.43678564e-01 -6.56087518e-01
1.04016095e-01 -1.05916572e+00 -9.10038233e-01 -1.31581664e-01
3.64896029e-01 4.49305862e-01 3.87193501e-01 -3.58678252e-01
1.58036590e-01 2.22429428e-02 -1.16342568e+00 1.00406635e+00
1.04382360e+00 7.35705197e-01 2.83309072e-01 6.72080338e-01
-7.24763691e-01 1.71700847e+00 -4.64984596e-01 -9.70405102e-01
-3.53020698e-01 -8.33025575e-01 -1.40637553e+00 1.60260558e-01
4.37584132e-01 -2.96870232e-01 -5.23572505e-01 1.27919376e+00
-4.58167046e-02 -9.23617110e-02 6.74227476e-01 -1.36666131e+00
-1.80671453e+00 8.18946779e-01 -3.95385057e-01 6.29800707e-02
5.16790807e-01 3.88268501e-01 1.45315075e+00 -7.11304903e-01
5.76681912e-01 1.54751837e+00 2.07493991e-01 7.56298244e-01
-1.01205707e+00 -3.62219840e-01 6.01596057e-01 9.17708814e-01
-8.35545421e-01 1.17580988e-01 9.46661592e-01 -4.03769672e-01
9.61206913e-01 2.30119184e-01 9.26948726e-01 7.71181464e-01
1.48104444e-01 9.69929814e-01 8.31542492e-01 -1.32189810e-01
1.64786115e-01 -1.67327553e-01 3.59861761e-01 1.27054369e+00
2.94317633e-01 -4.66534197e-01 -4.28550750e-01 2.75458544e-01
8.48794281e-01 6.46073893e-02 -4.69038188e-01 -6.13631845e-01
-1.18050992e+00 8.99812222e-01 1.28363943e+00 1.64737344e-01
-3.52647364e-01 6.14159048e-01 3.11298460e-01 2.69310057e-01
-1.01112790e-01 4.97521102e-01 -1.51727155e-01 6.00881696e-01
-3.34431708e-01 -2.65636802e-01 6.79868340e-01 8.94561052e-01
8.82174850e-01 2.71343123e-02 -3.35837394e-01 3.11612695e-01
5.81668913e-01 2.33572140e-01 1.05122797e-01 -1.07022369e+00
7.30644226e-01 1.43428528e+00 -3.58299643e-01 -1.27392864e+00
-3.77781570e-01 -2.98260659e-01 -1.01930392e+00 3.90730239e-02
2.52684623e-01 2.83165336e-01 -1.01702976e+00 1.64212465e+00
3.07753921e-01 -1.02568217e-01 2.18108296e-01 1.09791827e+00
1.61813557e+00 6.20388389e-01 3.29492390e-01 3.48691225e-01
1.72771275e+00 -1.30485249e+00 -8.87111425e-01 -6.03313982e-01
3.99391115e-01 -2.14078546e-01 1.42087686e+00 -1.38109595e-01
-1.13212621e+00 -6.84146941e-01 -1.03550756e+00 -7.72879481e-01
-5.53798199e-01 -7.53532425e-02 8.28942180e-01 5.87254837e-02
-1.23198342e+00 7.05199838e-02 -3.74706537e-01 -3.67829591e-01
7.19966233e-01 5.71417846e-02 -3.97978693e-01 -2.70130485e-01
-1.41620636e+00 9.61539507e-01 5.35170197e-01 3.20284307e-01
-8.44811201e-01 -3.05144012e-01 -1.27541959e+00 5.21684170e-01
4.97984022e-01 -1.36711490e+00 9.41596568e-01 -9.91621494e-01
-1.00906551e+00 9.50613797e-01 -2.53410250e-01 -2.52204478e-01
1.51306838e-01 2.98918597e-02 -2.32051581e-01 6.27304018e-01
5.23565337e-03 7.45762050e-01 6.68088078e-01 -1.43947470e+00
-8.93965214e-02 -3.91173273e-01 8.50212157e-01 1.42707065e-01
9.93504748e-03 -3.14590007e-01 -6.78753316e-01 -2.07405508e-01
2.60013193e-01 -4.56354678e-01 -4.35532071e-02 2.27786854e-01
-4.13358301e-01 -2.00430349e-01 6.41556323e-01 -9.94100749e-01
1.02412045e+00 -1.79363334e+00 5.77478886e-01 -7.91685134e-02
8.43735158e-01 -1.76080875e-02 -2.27284998e-01 3.47522020e-01
-6.63655400e-02 1.96006998e-01 -1.99705720e-01 1.62321821e-01
2.37353757e-01 4.82982010e-01 -3.06521505e-01 5.86931705e-02
4.65869993e-01 1.81866598e+00 -9.70740736e-01 -6.71245158e-01
1.74011905e-02 3.62652361e-01 -6.74812734e-01 4.25127149e-01
-5.16484976e-01 2.62865424e-01 -3.57231468e-01 7.08894551e-01
6.08727396e-01 -1.08883822e+00 2.91231871e-01 -6.63261414e-01
5.92354238e-01 1.46067962e-01 -7.58364141e-01 1.65701771e+00
-2.28601485e-01 6.50600195e-01 7.05502108e-02 -1.13447809e+00
7.63504267e-01 1.61478966e-01 -4.82973218e-01 -8.54498923e-01
1.08630575e-01 -2.25522354e-01 1.62537560e-01 -7.42648959e-01
1.30682603e-01 -2.34339491e-01 -1.48551632e-03 3.98876458e-01
4.36965048e-01 -1.03466108e-01 1.08633026e-01 7.65942156e-01
8.87394845e-01 3.81964557e-02 5.11671484e-01 2.41363216e-02
8.89420331e-01 -2.04636492e-02 2.71016330e-01 5.08780003e-01
-2.85065055e-01 3.56823295e-01 9.95267689e-01 -5.34879804e-01
-6.51378512e-01 -1.15588403e+00 6.88491821e-01 9.90190387e-01
5.35799146e-01 -5.17620385e-01 -6.94036603e-01 -8.99178743e-01
-1.17711022e-01 7.69245744e-01 -9.92442429e-01 -4.11771029e-01
-3.20923805e-01 -1.08145714e-01 2.54835993e-01 7.69652307e-01
8.46781492e-01 -1.41151917e+00 -4.03368443e-01 -1.64211586e-01
-5.07723927e-01 -1.17771184e+00 -3.82952869e-01 -1.10540181e-01
-6.39108837e-01 -1.52452040e+00 -2.46610895e-01 -1.04268789e+00
1.03679073e+00 4.19500738e-01 1.60718930e+00 8.42684448e-01
-9.44449976e-02 8.12825680e-01 -3.71810794e-01 -1.78282037e-01
-1.48559675e-01 -2.21140698e-01 -8.44084978e-01 -3.70588177e-03
2.25346074e-01 -4.11601931e-01 -7.26477802e-01 6.26986921e-02
-9.57219362e-01 5.91277719e-01 5.84088147e-01 7.67499864e-01
5.11175036e-01 -1.44987732e-01 4.54862535e-01 -1.03838849e+00
7.90756464e-01 -5.17819464e-01 -3.65503937e-01 8.91466081e-01
-2.39242092e-01 4.58900988e-01 6.84778929e-01 -5.64205572e-02
-1.16917050e+00 -2.37299740e-01 8.58171433e-02 -3.44689697e-01
4.45928238e-02 5.86088538e-01 -5.29739320e-01 -4.20791209e-02
3.76096696e-01 2.83006996e-01 -1.20089628e-01 6.88538849e-02
1.03053904e+00 -3.53440940e-02 6.20567799e-01 -5.67687213e-01
9.79313970e-01 4.99471068e-01 -6.79117581e-03 -2.65336961e-01
-1.18114698e+00 -3.18167955e-02 -6.49872601e-01 -3.88310850e-01
1.40811467e+00 -7.32765019e-01 -1.26293707e+00 -5.51277734e-02
-1.50367701e+00 -4.33949441e-01 -2.16472801e-02 -1.35294423e-01
-5.03137648e-01 5.10178030e-01 -8.16620409e-01 -5.69903910e-01
-4.90000904e-01 -1.03136134e+00 7.79924989e-01 4.66138005e-01
-1.09913787e-02 -1.26413584e+00 -2.77188659e-01 8.03645432e-01
3.32118332e-01 1.09473810e-01 1.43525958e+00 -2.73272961e-01
-1.04336703e+00 3.27293962e-01 -8.52444232e-01 -6.28125146e-02
-8.28784034e-02 -1.95395038e-01 -8.25478256e-01 8.12360197e-02
-1.32842362e-01 -5.77384293e-01 1.07713580e+00 -2.84772776e-02
1.55615389e+00 -4.06410396e-01 -5.67422956e-02 5.84921718e-01
1.31127381e+00 -4.81834151e-02 9.05374706e-01 3.11880708e-01
1.09317589e+00 7.52620041e-01 4.76635955e-02 -2.62431741e-01
1.16771066e+00 1.48633122e-02 1.01220369e+00 -3.85650009e-01
-3.25206041e-01 -6.22208536e-01 8.84056389e-02 7.85582125e-01
-6.79815263e-02 -6.07781261e-02 -8.45198393e-01 4.36676413e-01
-2.02175069e+00 -1.20104957e+00 -3.23257893e-01 1.42667234e+00
5.83411038e-01 -1.06217712e-01 -2.87807077e-01 -1.59873009e-01
7.11525619e-01 3.21218252e-01 -7.05004096e-01 -5.64739406e-01
-1.19679019e-01 -7.94073194e-02 -1.98390737e-01 5.65115988e-01
-8.14393401e-01 1.01015890e+00 6.06833076e+00 2.01932251e-01
-6.80870414e-01 3.27982679e-02 5.23679137e-01 3.71357352e-01
-8.68496478e-01 2.44647726e-01 -1.38976097e-01 -4.22741286e-02
4.66457307e-01 -8.96254107e-02 7.08524287e-01 6.65011823e-01
-4.06943560e-01 1.63817897e-01 -1.15297794e+00 1.07891810e+00
2.93851972e-01 -1.65764344e+00 7.10023940e-01 -3.08973193e-01
4.50939775e-01 -4.19491440e-01 -1.45490125e-01 7.29065359e-01
4.06602949e-01 -1.28662634e+00 7.75587618e-01 6.06352270e-01
4.31899101e-01 -5.55212379e-01 6.10530853e-01 3.05716217e-01
-1.65222692e+00 -2.08987035e-02 -5.87467849e-01 -2.17381135e-01
1.59928322e-01 1.95254654e-01 -4.77170825e-01 7.76247263e-01
5.85076034e-01 6.99037373e-01 -1.00892568e+00 6.55660987e-01
-1.11215031e+00 2.65343249e-01 5.59722364e-01 -1.52272895e-01
1.98537394e-01 -1.29735768e-01 -1.00839883e-01 8.48283172e-01
1.29201412e-02 4.43900347e-01 -8.65968838e-02 1.44300103e+00
-4.06719983e-01 -1.56591877e-01 -6.83240891e-01 -1.64058998e-01
6.40793815e-02 1.36580837e+00 -5.45546412e-01 -5.85294902e-01
-6.99579716e-01 1.33694065e+00 1.12438130e+00 7.74138808e-01
-9.65951979e-01 -3.46253961e-01 2.98624575e-01 -1.19452342e-01
3.63762259e-01 -1.41193077e-01 -2.36620471e-01 -1.53844666e+00
-4.13109921e-02 -8.39511275e-01 8.75918031e-01 -1.60746157e+00
-1.52656448e+00 7.18993068e-01 -4.33488786e-01 -7.74375737e-01
9.38765258e-02 -1.01176345e+00 -8.83160233e-01 9.26032960e-01
-1.69975305e+00 -1.52110183e+00 -6.48850739e-01 9.23945963e-01
3.26297551e-01 1.79963216e-01 6.97202623e-01 -1.30648196e-01
-2.53038615e-01 2.91689873e-01 -9.09402370e-01 5.02255321e-01
2.58720994e-01 -1.45810735e+00 4.71432447e-01 8.93030584e-01
4.37323838e-01 8.32392991e-01 3.93746763e-01 -5.91539323e-01
-1.57159472e+00 -9.87490714e-01 7.24937618e-01 -5.40807486e-01
8.51638317e-01 -3.30002367e-01 -1.41677201e+00 1.04956603e+00
7.99331248e-01 8.98187608e-02 6.91251814e-01 2.19940946e-01
-9.86441851e-01 1.29088730e-01 -7.86600053e-01 8.11325133e-01
1.16495609e+00 -1.06055415e+00 -1.19179714e+00 9.14564058e-02
1.07239509e+00 -3.03980708e-01 -5.51015079e-01 1.83898985e-01
1.73668176e-01 -9.81081188e-01 1.02981210e+00 -1.12134731e+00
7.47279108e-01 -6.38790250e-01 -5.26176915e-02 -1.17276394e+00
-6.70491576e-01 -3.79315093e-02 -4.80102330e-01 1.04967833e+00
4.94244069e-01 -5.31753123e-01 4.93644387e-01 7.78290689e-01
-2.29608789e-02 -6.49329603e-01 -4.99356538e-01 -2.17934251e-01
-2.21851781e-01 -3.32124114e-01 6.40004098e-01 1.07238865e+00
3.50282937e-01 9.55966890e-01 -1.94523379e-01 3.93702149e-01
6.37816846e-01 7.04891980e-01 6.67119086e-01 -1.15141141e+00
-3.78926873e-01 -4.78782713e-01 -3.11783612e-01 -1.11356783e+00
3.84007663e-01 -1.28247201e+00 -2.43770853e-01 -2.63912010e+00
7.25506365e-01 4.73922312e-01 -2.92423308e-01 6.68347061e-01
-6.86540782e-01 -2.51332410e-02 4.08976734e-01 -6.58278540e-02
-1.17865121e+00 6.05030477e-01 1.90572989e+00 -5.73839128e-01
4.63198900e-01 -7.03455329e-01 -1.19559383e+00 7.79887438e-01
6.50613606e-01 5.33188470e-02 -7.27891862e-01 -8.42816174e-01
7.70081997e-01 1.04687102e-01 9.26765442e-01 -5.92362881e-01
3.91411483e-01 -2.88849324e-01 6.92576468e-01 -6.12042665e-01
7.93271661e-02 -7.77735949e-01 -1.47359148e-01 5.10054111e-01
-3.86928648e-01 2.32235849e-01 8.48175958e-02 6.52585566e-01
-4.20114100e-01 5.61412945e-02 4.04249877e-01 -5.68447769e-01
-1.07983148e+00 3.00478190e-01 -1.15272626e-01 2.15014011e-01
8.69306445e-01 8.32543075e-02 -9.95107472e-01 -7.18608856e-01
-8.29337180e-01 7.09914446e-01 3.41579229e-01 2.31285736e-01
1.02099228e+00 -1.57786715e+00 -3.96462470e-01 -2.18627870e-01
4.49779272e-01 -6.96168765e-02 5.40259182e-01 5.13676047e-01
-6.92905188e-01 4.19796020e-01 -3.15946966e-01 -3.32225919e-01
-1.11105084e+00 1.13406241e+00 6.02749884e-01 -2.24004626e-01
-5.57792246e-01 8.83175135e-01 8.32919657e-01 -6.49202883e-01
1.78788044e-02 -4.46039915e-01 -5.18482447e-01 -1.25295311e-01
6.82416081e-01 -2.16330886e-01 -4.25201416e-01 -4.92869973e-01
-4.17037368e-01 6.56376004e-01 2.36790329e-01 3.02360803e-01
1.06048453e+00 -1.79996178e-01 -6.02751374e-01 3.57245207e-01
8.58526587e-01 -4.33891088e-01 -1.01639295e+00 -1.70276016e-01
-1.73374936e-01 -2.01046467e-01 -3.58082414e-01 -7.91165590e-01
-1.27004421e+00 1.42843747e+00 -5.51691130e-02 3.62491846e-01
1.10886753e+00 5.65487266e-01 4.86766279e-01 4.82674778e-01
1.40811607e-01 -7.06663191e-01 6.18272960e-01 4.79897201e-01
1.29326332e+00 -1.33601093e+00 -6.87758699e-02 -4.29068863e-01
-8.39296222e-01 1.21912694e+00 1.14024854e+00 -1.14366496e-02
1.82177678e-01 -4.44327742e-01 9.49270278e-02 -9.05988216e-01
-9.71484005e-01 -4.44428682e-01 7.29294360e-01 4.92458999e-01
4.77419496e-01 1.52605176e-01 -1.42942950e-01 7.75844693e-01
1.50389045e-01 -3.65860879e-01 2.63870627e-01 4.77473646e-01
-5.71696997e-01 -6.77498639e-01 -1.97033763e-01 1.59958795e-01
8.14317539e-02 -2.44043529e-01 -8.20583105e-01 8.33927929e-01
-1.31911831e-02 1.01509070e+00 -2.20458224e-01 -2.64810145e-01
3.51052523e-01 1.95538655e-01 4.98534620e-01 -4.43938613e-01
-4.25637245e-01 -5.92584789e-01 9.44705680e-02 -8.07084799e-01
-4.61042851e-01 7.23677501e-02 -1.78208470e+00 -2.19285309e-01
1.42324075e-01 2.04316378e-02 1.59682497e-01 1.05662131e+00
3.22064131e-01 1.08142602e+00 -4.68120314e-02 -2.57631749e-01
1.22744657e-01 -6.14216447e-01 -3.65248054e-01 7.34941840e-01
1.31872371e-01 -4.94234145e-01 -2.38791138e-01 9.41694900e-02] | [10.61657428741455, 1.8302829265594482] |
3b236bb4-7edf-4c48-ad00-872b08aef973 | td-gen-graph-generation-with-tree | 2106.10656 | null | https://arxiv.org/abs/2106.10656v2 | https://arxiv.org/pdf/2106.10656v2.pdf | TD-GEN: Graph Generation With Tree Decomposition | We propose TD-GEN, a graph generation framework based on tree decomposition, and introduce a reduced upper bound on the maximum number of decisions needed for graph generation. The framework includes a permutation invariant tree generation model which forms the backbone of graph generation. Tree nodes are supernodes, each representing a cluster of nodes in the graph. Graph nodes and edges are incrementally generated inside the clusters by traversing the tree supernodes, respecting the structure of the tree decomposition, and following node sharing decisions between the clusters. Finally, we discuss the shortcomings of standard evaluation criteria based on statistical properties of the generated graphs as performance measures. We propose to compare the performance of models based on likelihood. Empirical results on a variety of standard graph generation datasets demonstrate the superior performance of our method. | ['Greg Mori', 'Amir H. Abdi', 'Hossein Hajimirsadeghi', 'Hamed Shirzad'] | 2021-06-20 | null | null | null | null | ['tree-decomposition'] | ['graphs'] | [ 6.65415943e-01 8.32939744e-01 -2.62371391e-01 1.55116301e-02
-3.41930479e-01 -8.28742862e-01 6.72065318e-01 2.64328748e-01
1.30197734e-01 8.11571777e-01 -1.18011422e-01 -4.70944613e-01
-3.62660468e-01 -1.37353206e+00 -4.52546537e-01 -6.19729161e-01
-5.78012884e-01 7.25443900e-01 3.75121891e-01 7.56101087e-02
1.71084642e-01 4.24880981e-01 -1.32893538e+00 -5.11989892e-02
9.21011627e-01 3.65622193e-01 1.62028790e-01 8.48168969e-01
1.09148316e-01 4.18003470e-01 -6.98044240e-01 -5.55691242e-01
5.03326237e-01 -6.00018501e-01 -8.80623639e-01 6.28661215e-01
3.44907828e-02 7.86505565e-02 -4.12732154e-01 9.09994841e-01
2.24432483e-01 -4.71498109e-02 7.14240611e-01 -1.74847376e+00
-2.88199633e-01 1.26413226e+00 -5.74339330e-01 -9.99486819e-02
4.68035489e-01 -3.18177015e-01 1.39498591e+00 -4.66600329e-01
1.04353309e+00 1.19117713e+00 4.02605265e-01 3.21351498e-01
-1.49451852e+00 -1.71358898e-01 2.11538956e-01 -1.95283201e-02
-1.49539018e+00 -2.38969073e-01 6.29179657e-01 -2.40838394e-01
6.79351091e-01 5.11021554e-01 7.16637433e-01 6.15408003e-01
2.59054244e-01 3.81077915e-01 7.61040509e-01 -8.38645279e-01
4.14992362e-01 -1.97217822e-01 1.44219056e-01 9.19270515e-01
1.11196876e+00 1.10568237e-02 -3.62631083e-01 -5.22738457e-01
6.76130533e-01 -7.71657825e-01 -1.08272575e-01 -8.41805100e-01
-9.98824954e-01 8.57716680e-01 2.35936821e-01 6.39607832e-02
-3.18708718e-01 2.70314276e-01 2.03287289e-01 6.11962676e-02
2.56098986e-01 3.20893586e-01 -1.57109842e-01 3.58004510e-01
-8.68446767e-01 2.74426639e-01 9.73825872e-01 1.50895071e+00
7.88717985e-01 -9.56219286e-02 -4.69930559e-01 4.14150238e-01
2.78087556e-01 2.32522815e-01 -2.53450245e-01 -7.87103593e-01
5.12771845e-01 8.32677186e-01 -2.04309568e-01 -1.03085482e+00
-1.45092905e-01 -5.05655944e-01 -8.76617432e-01 -1.29250631e-01
1.69998780e-01 -2.79264390e-01 -1.14030731e+00 1.75501227e+00
4.89624232e-01 7.80887008e-02 -2.24172175e-01 7.27191567e-02
3.24369133e-01 3.14730257e-01 -2.17058972e-01 -5.28649807e-01
1.09084284e+00 -8.64311397e-01 -4.16665047e-01 1.96098298e-01
8.48423004e-01 -3.56115669e-01 3.97989601e-01 2.52004772e-01
-1.09571886e+00 -2.95770168e-01 -8.71118188e-01 5.24596214e-01
-1.69269592e-01 -1.48968190e-01 8.64275932e-01 1.27478242e+00
-1.54405677e+00 7.01893747e-01 -7.42353082e-01 -3.31735551e-01
-7.74179101e-02 3.66523445e-01 -2.37488016e-01 -1.37208268e-01
-7.59727716e-01 3.91162097e-01 7.98545241e-01 2.24888809e-02
-8.54959726e-01 -3.13404083e-01 -1.06034362e+00 1.27222016e-01
6.06737077e-01 -1.05556202e+00 1.01369548e+00 -3.76368523e-01
-9.60814834e-01 6.65059984e-01 -2.78543174e-01 -4.73326415e-01
2.89236456e-01 5.03856242e-01 -2.71849304e-01 1.79547682e-01
4.16364461e-01 6.17570043e-01 6.35332406e-01 -1.39369857e+00
-7.78075695e-01 -1.77499563e-01 -4.75687049e-02 8.85994360e-02
-8.49868804e-02 -2.00828999e-01 -8.93907189e-01 -5.63676238e-01
3.60982507e-01 -1.44863033e+00 -8.43187571e-01 -8.74862969e-01
-1.22337806e+00 -3.80655617e-01 4.53656703e-01 -2.24569276e-01
1.77160835e+00 -1.85680330e+00 2.67855316e-01 1.00971794e+00
5.33310771e-01 -2.97051966e-01 -4.72684801e-01 9.73064601e-01
-2.23938867e-01 6.23152733e-01 -2.09459603e-01 -6.20698463e-03
1.03794917e-01 1.99561849e-01 -1.16046369e-01 1.49450958e-01
-1.43386319e-01 7.57471502e-01 -9.64852929e-01 -4.33377892e-01
1.03435740e-01 -2.60468841e-01 -5.31061828e-01 -9.87177044e-02
-1.56661764e-01 -2.33562917e-01 -4.28934932e-01 3.42176348e-01
7.75749683e-01 -4.23506230e-01 9.77342188e-01 1.72274590e-01
1.42907232e-01 4.60488975e-01 -1.45510781e+00 1.17992699e+00
7.08091930e-02 3.22330929e-02 -1.82373747e-02 -7.46334553e-01
7.86335945e-01 7.86216408e-02 4.78797615e-01 9.90697555e-03
-2.01710060e-01 3.03892232e-02 2.92920381e-01 2.01351807e-01
5.30456603e-01 4.27650124e-01 -2.91607469e-01 8.17750514e-01
-1.58641860e-01 -2.44498551e-01 1.17445433e+00 1.00569212e+00
1.65730739e+00 -1.73290417e-01 4.68580782e-01 -3.16968590e-01
3.41361403e-01 -6.17852807e-03 4.68243331e-01 1.09654534e+00
3.00795227e-01 3.83319944e-01 8.86015654e-01 -1.70071855e-01
-9.61869836e-01 -1.12031591e+00 5.78991100e-02 8.20017755e-01
1.00589916e-02 -1.19463968e+00 -9.45158064e-01 -8.76793683e-01
2.24940386e-02 7.50061274e-01 -7.02882349e-01 -1.50856346e-01
-3.66678596e-01 -9.72544074e-01 2.80974507e-01 1.87390596e-01
3.28816357e-03 -9.19612765e-01 -3.63666832e-01 3.96604419e-01
-1.13229387e-01 -1.37566125e+00 -3.82074445e-01 1.33013725e-01
-9.44879889e-01 -1.35524106e+00 -2.06495777e-01 -8.51488352e-01
1.33154738e+00 5.33836961e-01 1.31310940e+00 4.59664464e-01
-2.36389622e-01 4.35939401e-01 -5.89483380e-01 6.60827309e-02
-7.07029819e-01 5.62703371e-01 -3.55337597e-02 -2.46711701e-01
-3.60083252e-01 -7.45212495e-01 -1.15571246e-01 2.29990914e-01
-8.17545652e-01 2.33559996e-01 4.68461245e-01 5.93860507e-01
5.62443256e-01 7.72841394e-01 2.74964899e-01 -1.24715543e+00
8.66719842e-01 -5.66603065e-01 -7.64926791e-01 4.73735422e-01
-7.94330895e-01 3.03084999e-01 2.92329043e-01 7.72663355e-02
-7.89224923e-01 1.93871334e-01 4.17563409e-01 9.63848680e-02
2.11892888e-01 6.54316545e-01 -2.04689011e-01 -8.59843418e-02
2.67332911e-01 7.71767572e-02 -3.55953008e-01 -1.46538779e-01
7.83178985e-01 2.98522711e-02 2.81450570e-01 -6.26299620e-01
1.20611274e+00 3.60410899e-01 5.74017227e-01 -5.33489108e-01
-2.29815647e-01 -1.93269774e-01 -6.90648019e-01 -1.44357979e-01
4.85003263e-01 -5.10198414e-01 -5.17879188e-01 2.15764090e-01
-1.02896690e+00 -2.08319515e-01 -2.72010982e-01 7.91189522e-02
-3.68294626e-01 8.03658485e-01 -5.44925988e-01 -9.57400382e-01
-2.25789994e-01 -8.10414672e-01 9.16663647e-01 -6.72899261e-02
-3.29482377e-01 -9.34426546e-01 2.15010256e-01 -3.69112231e-02
-1.37350470e-01 5.35603642e-01 1.41197789e+00 -6.17967308e-01
-1.06206810e+00 -1.80295229e-01 -5.07119633e-02 -1.46178782e-01
3.60855699e-01 3.69617850e-01 -2.04697877e-01 -5.03870487e-01
-7.04679430e-01 2.50588745e-01 7.54059494e-01 3.54910880e-01
1.01447797e+00 -3.12153101e-01 -1.06038296e+00 3.59716058e-01
1.34045899e+00 2.49780551e-01 6.26313210e-01 7.03264028e-02
6.93157434e-01 6.76153660e-01 2.30368227e-01 4.79957491e-01
5.08559108e-01 6.04507089e-01 3.75147343e-01 5.04963323e-02
-1.74046189e-01 -6.36195719e-01 2.55390078e-01 1.01166832e+00
-4.99797203e-02 -9.20910656e-01 -8.47672701e-01 7.31730461e-01
-1.81687462e+00 -6.97389781e-01 -7.61806548e-01 2.34782577e+00
4.24581438e-01 2.34657630e-01 2.23028839e-01 4.19641733e-01
1.08178318e+00 2.96118073e-02 -1.23052843e-01 -3.87913316e-01
-6.89789802e-02 5.70152998e-01 7.28125691e-01 5.53957343e-01
-6.52958035e-01 9.86365795e-01 8.19916153e+00 7.94821322e-01
-9.75900441e-02 -3.18637222e-01 4.79720294e-01 2.21417353e-01
-6.95525348e-01 6.68277264e-01 -7.75798798e-01 2.68785179e-01
7.44189978e-01 -7.94676423e-01 4.62160349e-01 7.40260422e-01
-2.95522455e-02 -1.77328631e-01 -1.21847916e+00 4.00387675e-01
-1.81930691e-01 -1.30385518e+00 4.09184366e-01 6.19912326e-01
1.03947687e+00 -3.24859887e-01 -2.36857727e-01 2.83559170e-02
1.11846566e+00 -8.24325562e-01 3.58418167e-01 1.59084219e-02
6.95724308e-01 -8.23877990e-01 2.71701962e-01 2.43067741e-01
-1.64593267e+00 1.05279544e-02 -1.69734508e-01 9.23579261e-02
3.48213851e-01 1.01601660e+00 -1.42065942e+00 1.11329436e+00
3.60741347e-01 2.05294460e-01 -6.93724334e-01 1.00814927e+00
-5.35608649e-01 6.38587832e-01 -3.58968198e-01 2.23391369e-01
-2.60727443e-02 -6.17382228e-01 7.22088397e-01 9.17430401e-01
4.41326439e-01 -5.53207472e-02 3.57321262e-01 7.03854680e-01
-1.05034098e-01 2.49105562e-02 -8.72906506e-01 -1.34821132e-01
9.81407940e-01 1.40524924e+00 -1.33030689e+00 -3.98797929e-01
-5.96703636e-03 6.37547374e-01 2.78239399e-01 2.83548355e-01
-6.32353604e-01 -6.55473828e-01 1.65189683e-01 2.13523477e-01
5.09612441e-01 -4.22381669e-01 -6.79263100e-02 -7.69608736e-01
1.07505247e-01 -9.83891785e-01 6.46523237e-01 -4.20362025e-01
-9.64049518e-01 6.04827583e-01 3.02578896e-01 -7.50150800e-01
-5.28318822e-01 -2.48213544e-01 -8.79064918e-01 5.10807514e-01
-8.21884573e-01 -9.62623000e-01 -1.33704215e-01 3.47845972e-01
-2.19885055e-02 7.33760968e-02 7.13996649e-01 -2.42042616e-01
-5.29243469e-01 6.08521342e-01 -1.47531703e-01 3.32311797e-03
-5.79288080e-02 -1.29477811e+00 1.05116141e+00 1.31965768e+00
3.94815505e-01 6.25230134e-01 3.89099598e-01 -1.09146523e+00
-1.31690872e+00 -9.82098222e-01 1.09101820e+00 -2.62844145e-01
4.20408696e-01 -6.50575578e-01 -4.16407555e-01 8.61745179e-01
2.14947388e-01 -5.20735204e-01 5.21242201e-01 3.81075263e-01
-1.73696205e-01 4.06883359e-02 -9.79468703e-01 8.44613910e-01
1.66233802e+00 -6.48764223e-02 -9.16977674e-02 5.13509035e-01
6.72299385e-01 -1.43497825e-01 -8.68233621e-01 5.95965862e-01
2.60428280e-01 -9.14596677e-01 6.83916450e-01 -4.56170261e-01
3.63202766e-02 -4.67705548e-01 1.23649061e-01 -1.35833156e+00
-7.80566216e-01 -1.24228573e+00 3.79307382e-02 1.25195980e+00
7.01019168e-01 -6.32734060e-01 1.05873024e+00 3.15741062e-01
4.41423282e-02 -3.92302573e-01 -9.43914294e-01 -8.90968740e-01
-4.26893175e-01 -1.72717735e-01 9.49420452e-01 7.36441255e-01
2.19290420e-01 7.10040092e-01 -2.10403860e-01 1.51487112e-01
1.06052518e+00 2.16617554e-01 1.11009228e+00 -1.42435825e+00
-3.55173290e-01 -3.97798985e-01 -3.57850105e-01 -7.33982682e-01
1.33390710e-01 -1.31436741e+00 -2.97107756e-01 -1.98605561e+00
5.19754171e-01 -6.68742776e-01 9.92305204e-02 4.96010900e-01
-2.53288507e-01 -1.68180689e-01 2.41917297e-01 -1.14538722e-01
-5.70078731e-01 4.10776466e-01 1.18760204e+00 1.54067352e-01
-3.15564424e-01 1.62514254e-01 -8.87718618e-01 4.92038816e-01
6.91269577e-01 -6.88994408e-01 -8.23613465e-01 -5.11603020e-02
4.01736349e-01 3.62381876e-01 -6.23540804e-02 -8.07680786e-01
1.35242343e-01 -7.22232461e-02 -1.51677132e-02 -7.87196517e-01
-1.18044779e-01 -5.13702273e-01 8.72084916e-01 7.50941753e-01
-3.10965866e-01 3.86131614e-01 -2.52998829e-01 7.64932156e-01
1.58908606e-01 -2.56041646e-01 4.29276437e-01 -1.58178080e-02
-3.35840136e-01 4.12436157e-01 -3.75882268e-01 -7.91556016e-02
1.28165853e+00 -5.17357767e-01 -1.87020555e-01 -3.50391895e-01
-7.66914189e-01 3.76048952e-01 4.30606395e-01 3.04048061e-01
4.49071437e-01 -1.38329816e+00 -8.72245669e-01 1.78922087e-01
1.28467605e-01 -6.41363189e-02 -1.50751129e-01 4.57345575e-01
-2.07298890e-01 3.19265068e-01 1.16255783e-01 -3.29462856e-01
-1.46772170e+00 5.61575413e-01 -2.50470042e-02 -7.57533550e-01
-4.55229551e-01 5.59666276e-01 2.86867559e-01 -3.82396847e-01
-1.95375308e-01 -1.80491403e-01 8.07457939e-02 -2.44593859e-01
-4.41333540e-02 6.01087987e-01 -3.35396938e-02 -3.85164499e-01
-3.90298992e-01 1.62628785e-01 -2.18584254e-01 -2.36052215e-01
9.15749073e-01 -1.22374937e-01 -6.25951827e-01 -1.69318646e-01
5.19815385e-01 3.04259241e-01 -4.63909060e-01 3.32417600e-02
3.99943560e-01 -4.98262346e-01 -4.03647780e-01 -3.81742477e-01
-9.18627322e-01 5.95180914e-02 -1.90140471e-01 8.21115732e-01
1.01772881e+00 1.83025263e-02 5.05643964e-01 2.30203733e-01
7.28512704e-01 -1.02059793e+00 -2.03244716e-01 2.71370590e-01
5.02967179e-01 -3.05776387e-01 2.23982498e-01 -1.30609667e+00
-3.62396687e-01 7.34010875e-01 4.89759207e-01 -5.95838651e-02
5.27013779e-01 1.75210580e-01 -6.98411167e-01 -7.31610730e-02
-1.08486581e+00 -3.04592341e-01 5.11118434e-02 9.89058495e-01
1.37353659e-01 5.08012652e-01 -6.55992210e-01 2.56493807e-01
-4.51844841e-01 -3.07170570e-01 8.41628075e-01 7.15558290e-01
-4.13844824e-01 -1.82635808e+00 -2.60004818e-01 5.70108414e-01
-1.33079454e-01 -2.41619006e-01 -8.11828852e-01 8.91951144e-01
1.45930871e-01 1.27266943e+00 -2.94948611e-02 -4.52593297e-01
1.61077812e-01 -2.70533621e-01 6.26574814e-01 -9.67830181e-01
-2.53504872e-01 -1.76503081e-02 7.14328945e-01 -2.43662611e-01
-1.14678577e-01 -6.43108547e-01 -9.77147102e-01 -5.42046547e-01
-6.27471745e-01 5.76007247e-01 1.71662703e-01 4.32103723e-01
6.33611441e-01 4.54403311e-01 1.09685874e+00 -4.70812380e-01
-3.73619378e-01 -5.96786141e-01 -9.44128156e-01 2.06467852e-01
-4.95108664e-01 -4.21099991e-01 -3.67207021e-01 -8.31049457e-02] | [7.006231784820557, 5.401900768280029] |
1ca69239-9932-43e5-952e-5c219e305d32 | pstr-end-to-end-one-step-person-search-with | 2204.0334 | null | https://arxiv.org/abs/2204.03340v1 | https://arxiv.org/pdf/2204.03340v1.pdf | PSTR: End-to-End One-Step Person Search With Transformers | We propose a novel one-step transformer-based person search framework, PSTR, that jointly performs person detection and re-identification (re-id) in a single architecture. PSTR comprises a person search-specialized (PSS) module that contains a detection encoder-decoder for person detection along with a discriminative re-id decoder for person re-id. The discriminative re-id decoder utilizes a multi-level supervision scheme with a shared decoder for discriminative re-id feature learning and also comprises a part attention block to encode relationship between different parts of a person. We further introduce a simple multi-scale scheme to support re-id across person instances at different scales. PSTR jointly achieves the diverse objectives of object-level recognition (detection) and instance-level matching (re-id). To the best of our knowledge, we are the first to propose an end-to-end one-step transformer-based person search framework. Experiments are performed on two popular benchmarks: CUHK-SYSU and PRW. Our extensive ablations reveal the merits of the proposed contributions. Further, the proposed PSTR sets a new state-of-the-art on both benchmarks. On the challenging PRW benchmark, PSTR achieves a mean average precision (mAP) score of 56.5%. The source code is available at \url{https://github.com/JialeCao001/PSTR}. | ['Fahad Shahbaz Khan', 'Mubarak Shah', 'Jin Xie', 'Hisham Cholakkal', 'Rao Muhammad Anwer', 'Yanwei Pang', 'Jiale Cao'] | 2022-04-07 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Cao_PSTR_End-to-End_One-Step_Person_Search_With_Transformers_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Cao_PSTR_End-to-End_One-Step_Person_Search_With_Transformers_CVPR_2022_paper.pdf | cvpr-2022-1 | ['person-search'] | ['computer-vision'] | [ 8.82904008e-02 -1.55703232e-01 -7.97902793e-02 -5.53283930e-01
-1.25719774e+00 -2.90195674e-01 6.71517134e-01 -1.52076229e-01
-6.64593101e-01 3.18111837e-01 4.46555346e-01 3.02748054e-01
7.24657401e-02 -4.31702495e-01 -6.19908631e-01 -2.88322508e-01
1.96502551e-01 9.17130888e-01 2.66322464e-01 -1.10558860e-01
-1.97689503e-01 1.62965283e-01 -1.54453146e+00 2.88102090e-01
6.31902814e-01 8.22404742e-01 1.49076089e-01 7.16441870e-01
7.34257400e-01 4.44520086e-01 -4.62582201e-01 -1.09331143e+00
4.89414990e-01 -1.62951663e-01 -7.39635408e-01 5.10401540e-02
8.41096938e-01 -5.83008170e-01 -1.02273726e+00 8.29621315e-01
1.15893281e+00 1.86877951e-01 4.40959930e-01 -1.15799415e+00
-8.93630266e-01 3.22539151e-01 -6.57397389e-01 3.45547467e-01
7.45334148e-01 5.20043194e-01 1.11663675e+00 -1.06635046e+00
3.13730270e-01 1.44060695e+00 8.83630753e-01 8.99420977e-01
-1.05760288e+00 -7.54033208e-01 1.15766212e-01 3.12147468e-01
-1.78416836e+00 -5.45473278e-01 3.15617770e-01 -4.29300159e-01
1.10714221e+00 2.43619397e-01 5.84587395e-01 1.22934353e+00
-3.70273143e-01 1.23290122e+00 7.98401415e-01 -1.55915171e-01
-3.16109359e-01 6.79566409e-04 4.19713974e-01 7.81540990e-01
3.66837651e-01 2.90282726e-01 -8.12264979e-01 -1.09525919e-01
6.65694773e-01 1.80379346e-01 -2.59683728e-01 -2.14772671e-02
-1.02100611e+00 5.32064140e-01 4.55224037e-01 -1.71375871e-02
-1.13422260e-01 2.35564917e-01 4.86491382e-01 1.15642510e-01
2.14832366e-01 -6.76997080e-02 -2.87587732e-01 -1.37749001e-01
-1.00819755e+00 5.45904517e-01 5.73721886e-01 1.16787171e+00
3.78074855e-01 -3.69160801e-01 -9.43528175e-01 1.11407936e+00
3.88818800e-01 8.72615755e-01 4.54894483e-01 -4.58974093e-01
5.99344552e-01 6.12973511e-01 1.71317399e-01 -5.86502612e-01
-1.83723241e-01 -8.15740168e-01 -7.18321621e-01 -2.69052595e-01
3.25816065e-01 1.62120983e-01 -1.02110684e+00 1.78475678e+00
2.91401297e-01 4.36878890e-01 -1.75187722e-01 1.33071399e+00
1.13114679e+00 2.66711354e-01 2.04193756e-01 5.02151191e-01
1.98997140e+00 -1.45670044e+00 -1.56735644e-01 -5.05151331e-01
3.02980393e-01 -4.32292193e-01 8.98486197e-01 -1.67351246e-01
-1.17618847e+00 -8.89281809e-01 -8.03338349e-01 -3.66568059e-01
-1.54720485e-01 6.38678730e-01 2.04685479e-01 7.11817503e-01
-1.05674279e+00 1.39796272e-01 -6.30346656e-01 -6.50570333e-01
5.33800125e-01 4.46194053e-01 -5.44900298e-01 -2.14431778e-01
-1.17071187e+00 6.91495717e-01 1.10450722e-02 8.49247873e-02
-9.20239508e-01 -7.54989862e-01 -8.74688566e-01 6.86130598e-02
3.08527201e-01 -1.13371575e+00 1.31303668e+00 -5.11882186e-01
-1.11041069e+00 1.57101226e+00 -4.95869160e-01 -5.61289251e-01
7.48748362e-01 -4.20341969e-01 -4.88442421e-01 1.00869089e-02
4.50080663e-01 5.21792173e-01 6.42807662e-01 -8.59802127e-01
-8.64834547e-01 -6.54585660e-01 -2.35340849e-01 1.95505202e-01
-2.31630877e-01 5.48574567e-01 -1.36967158e+00 -7.94816077e-01
-4.06713724e-01 -9.46710467e-01 1.10325359e-01 -4.82971184e-02
-5.71947098e-01 -6.39609993e-01 1.71545327e-01 -1.07941985e+00
1.23275805e+00 -1.92709160e+00 8.43397081e-02 2.97572836e-02
2.95266390e-01 4.68266219e-01 -3.31259936e-01 2.66357750e-01
-4.77756634e-02 -4.29655105e-01 -1.13527112e-01 -1.17581081e+00
3.64386648e-01 -3.55548799e-01 -1.81783754e-02 7.27540672e-01
-9.30421650e-02 1.37176132e+00 -6.60310745e-01 -3.74046326e-01
3.96933481e-02 5.61827481e-01 -4.59765196e-01 3.93963009e-01
3.86779755e-01 2.08410010e-01 -2.27655068e-01 1.01999032e+00
4.78958547e-01 -3.76389861e-01 -9.87210870e-02 -2.29994476e-01
6.84678480e-02 2.31859654e-01 -1.18188965e+00 1.72341335e+00
-3.03961639e-03 4.17411298e-01 -1.73816159e-01 -7.11896360e-01
7.47732937e-01 1.17574677e-01 2.23262027e-01 -8.97465408e-01
-1.80785581e-02 1.99709326e-01 -3.63495886e-01 -1.07089683e-01
5.56089699e-01 4.10333782e-01 -3.69288534e-01 2.69127697e-01
2.03892246e-01 9.95150924e-01 1.84324741e-01 3.18093866e-01
1.14957726e+00 6.70868754e-02 1.89578444e-01 -2.33295441e-01
8.30014288e-01 -3.42121273e-01 7.73752809e-01 1.04586995e+00
-4.59117740e-01 7.63882637e-01 -1.16654433e-01 -4.70930964e-01
-1.07403231e+00 -1.13322067e+00 -8.09525102e-02 1.28641164e+00
2.55944759e-01 -6.65155232e-01 -6.08634114e-01 -6.74905241e-01
4.22662228e-01 2.05182895e-01 -6.28654599e-01 -2.98370831e-02
-6.87926948e-01 -5.63067734e-01 9.75053966e-01 7.14539289e-01
8.49484086e-01 -7.79370904e-01 -3.03019434e-01 -5.87019213e-02
-3.66054684e-01 -1.35353446e+00 -1.35467720e+00 -4.34942365e-01
-1.02109924e-01 -1.09088194e+00 -1.32725143e+00 -1.00809383e+00
5.81422567e-01 3.68007243e-01 1.17221141e+00 1.27448991e-01
-6.02937818e-01 6.79415584e-01 -1.49101228e-01 -2.71831620e-02
1.45261377e-01 2.26147011e-01 1.83416456e-01 1.13443963e-01
7.36001551e-01 -2.22567126e-01 -1.05677330e+00 6.61065996e-01
-1.15812957e-01 -6.47636801e-02 5.28488278e-01 8.49900544e-01
6.52163148e-01 -4.77464139e-01 2.83779442e-01 -3.70132536e-01
3.27503890e-01 -1.62561551e-01 -5.06034255e-01 5.97431958e-01
-4.47111636e-01 -2.10299149e-01 1.28309742e-01 -2.94159412e-01
-8.48867416e-01 2.83714812e-02 -2.68948764e-01 -3.14124286e-01
-1.67823404e-01 -1.86368600e-01 -3.99425775e-01 2.80050673e-02
4.25960898e-01 7.67214835e-01 -3.57992709e-01 -7.82501876e-01
7.54086599e-02 9.44118679e-01 1.08136749e+00 -5.57108641e-01
8.78061056e-01 3.86953026e-01 -4.84259367e-01 -4.40906525e-01
-9.21897888e-01 -9.57336843e-01 -4.84006077e-01 -1.67868987e-01
7.63212621e-01 -1.68424904e+00 -1.05081451e+00 9.45594907e-01
-9.55762982e-01 -3.95764709e-01 -1.78548545e-01 1.13563679e-01
-2.54009396e-01 4.84963238e-01 -6.61729872e-01 -7.51428187e-01
-9.44521785e-01 -1.01378155e+00 1.64422607e+00 4.19099689e-01
-1.08462520e-01 -5.36536515e-01 1.26722530e-01 9.00391996e-01
1.59309030e-01 -1.59035668e-01 -2.07394660e-01 -7.51160741e-01
-5.94033301e-01 -4.73773479e-01 -4.95570242e-01 -6.68255538e-02
-2.47165889e-01 -9.51826751e-01 -8.89022171e-01 -8.43541503e-01
-7.54290164e-01 -3.42419237e-01 1.37960243e+00 1.01885973e-02
9.87601280e-01 -1.78363413e-01 -7.18621254e-01 8.30378890e-01
1.28089774e+00 -2.89484829e-01 6.54704452e-01 3.84599805e-01
7.81545341e-01 2.19275013e-01 4.78428662e-01 5.43196261e-01
1.06002426e+00 1.44377244e+00 -2.23790288e-01 -1.74902961e-01
-7.15486825e-01 -5.75782895e-01 4.12918866e-01 4.02653776e-02
-2.85034895e-01 -2.45173410e-01 -9.60162222e-01 7.72990823e-01
-2.16889882e+00 -1.18383801e+00 3.67595144e-02 2.31862211e+00
7.40947366e-01 -1.48755342e-01 7.22559750e-01 -1.59815043e-01
9.54355717e-01 4.10276651e-02 -6.86624706e-01 3.73652786e-01
-1.10502571e-01 8.83450583e-02 5.93942940e-01 5.46892583e-01
-1.42072129e+00 1.04904151e+00 4.96822071e+00 8.59807789e-01
-3.68327767e-01 4.08270806e-01 4.39231396e-01 -3.31831008e-01
1.74315795e-01 -3.81027371e-01 -1.65771842e+00 7.92108774e-01
5.38241804e-01 -5.42946272e-02 3.03037316e-01 8.21088076e-01
-1.09961160e-01 1.78572088e-01 -1.23630202e+00 1.65192318e+00
4.44742531e-01 -1.00494635e+00 -8.67207795e-02 -5.70334308e-03
4.10147458e-01 1.66429162e-01 3.00816018e-02 4.83810157e-01
1.57623455e-01 -9.07198787e-01 8.53237748e-01 6.33595645e-01
1.08810234e+00 -5.99668205e-01 6.17775917e-01 1.38132721e-01
-1.82038498e+00 -1.78717151e-01 -2.58649558e-01 2.52035052e-01
3.80757093e-01 3.15353423e-01 -3.19159567e-01 4.90077972e-01
1.07947636e+00 8.39416265e-01 -8.32450509e-01 1.20616364e+00
-2.99790800e-01 3.09700549e-01 -3.42725724e-01 2.41357982e-01
-3.48529100e-01 1.79379925e-01 6.84945941e-01 1.64292657e+00
1.46079183e-01 6.14630543e-02 2.67539561e-01 6.76024675e-01
-1.58244908e-01 -2.63454139e-01 8.16982538e-02 5.37655950e-01
6.40937150e-01 1.07245898e+00 -1.18444681e-01 -4.89161462e-01
-5.04837334e-01 1.59461093e+00 5.82597315e-01 2.16783896e-01
-1.01747775e+00 -1.70435101e-01 1.08721125e+00 2.91945010e-01
5.42912722e-01 2.24277582e-02 7.92588443e-02 -1.52204728e+00
3.30030262e-01 -9.58085775e-01 9.44789112e-01 -4.81557339e-01
-1.50096130e+00 4.50331330e-01 -1.75065920e-01 -9.25220013e-01
-4.41980325e-02 -3.07065398e-01 -3.64251673e-01 1.21710563e+00
-1.54725862e+00 -1.80471849e+00 -4.94248599e-01 9.37833786e-01
6.85824454e-01 -6.43789649e-01 6.68603063e-01 7.29868650e-01
-1.02288079e+00 1.58414972e+00 -3.13592941e-01 7.34519482e-01
9.53347504e-01 -1.20259619e+00 9.22110379e-01 1.17687047e+00
-4.23754081e-02 7.61642873e-01 3.47598433e-01 -8.02501440e-01
-1.32617366e+00 -1.28589702e+00 1.30807078e+00 -5.99945784e-01
2.09868968e-01 -6.49017751e-01 -4.74566877e-01 8.30327153e-01
-1.18070133e-01 1.67088117e-02 7.26725459e-01 2.78279930e-01
-7.10891366e-01 -1.76981807e-01 -1.05491054e+00 5.16275406e-01
1.64065492e+00 -7.10324407e-01 -4.58874434e-01 3.46990466e-01
4.60062116e-01 -5.39993942e-01 -7.17262566e-01 3.15406740e-01
8.20481777e-01 -9.01505232e-01 1.63900459e+00 -3.91238064e-01
2.59291176e-02 -3.93487632e-01 -1.29373446e-01 -8.29356134e-01
-7.81715274e-01 -5.72191834e-01 -4.46871608e-01 1.33214617e+00
1.16127200e-01 -7.62466967e-01 7.28590786e-01 7.65819073e-01
1.39156833e-01 -7.43443549e-01 -1.00257015e+00 -1.00377870e+00
-3.64957750e-01 -2.33242199e-01 5.49926102e-01 4.02732968e-01
-3.23926330e-01 2.39657253e-01 -8.32183301e-01 5.63089371e-01
1.29029715e+00 1.40382499e-01 9.87410128e-01 -9.62659419e-01
-8.28673601e-01 -3.03166151e-01 -5.28332829e-01 -1.47547591e+00
1.85282696e-02 -1.09019029e+00 -8.61070380e-02 -1.55649185e+00
9.31208491e-01 -2.34165981e-01 -2.81995237e-01 6.66430771e-01
-5.19903123e-01 4.19958860e-01 4.21879917e-01 5.38300991e-01
-1.02630138e+00 5.81053615e-01 6.96161866e-01 -3.18690538e-01
9.02533159e-03 1.39267981e-01 -8.75153065e-01 3.38022768e-01
6.55150712e-01 -2.96712160e-01 1.33618474e-01 -6.48994863e-01
-1.88116223e-01 -2.95710117e-01 9.65406954e-01 -1.15436113e+00
6.84443951e-01 5.57066083e-01 5.78616858e-01 -6.34760201e-01
5.24862707e-01 -2.93752611e-01 5.61578684e-02 6.89944148e-01
-3.86392444e-01 -3.52018094e-03 -9.07398686e-02 5.97180784e-01
2.95708179e-02 2.43842870e-01 8.64173889e-01 2.75723729e-02
-9.95359659e-01 7.22753227e-01 2.21403509e-01 1.51845351e-01
8.63279343e-01 -1.42908290e-01 -3.05349112e-01 -1.99757531e-01
-5.76170385e-01 7.07400620e-01 3.04226965e-01 6.83510721e-01
6.19363725e-01 -1.34932303e+00 -1.21661353e+00 1.16115987e-01
4.35287356e-01 -3.47171575e-01 4.53112990e-01 5.61979294e-01
4.55583930e-02 4.88193721e-01 2.24348933e-01 -4.62531835e-01
-1.71259880e+00 3.57133418e-01 6.69755936e-01 -4.58423972e-01
-8.54742885e-01 1.25953841e+00 2.41188914e-01 -4.40569520e-01
5.76248586e-01 4.10695881e-01 1.03937805e-01 -2.49346256e-01
1.01997125e+00 5.16645133e-01 -1.82302088e-01 -9.03202295e-01
-8.27818096e-01 7.15241790e-01 -2.98795283e-01 -2.03064047e-02
1.09349942e+00 -1.55345604e-01 2.43258744e-01 -2.45795786e-01
9.78281736e-01 -2.83722878e-01 -1.23685002e+00 -6.77151799e-01
-1.61314279e-01 -5.55776298e-01 -3.02280098e-01 -1.06903660e+00
-9.76755857e-01 3.92983049e-01 8.49140584e-01 -5.80632150e-01
1.02170849e+00 4.82103795e-01 1.19935060e+00 2.10428715e-01
5.21015286e-01 -1.12014627e+00 6.41476363e-02 4.14736778e-01
9.92635012e-01 -1.37453330e+00 -8.72186944e-02 -4.54637140e-01
-5.41501582e-01 4.34981674e-01 7.84700155e-01 -7.83225000e-02
4.07416224e-01 1.71347514e-01 -4.07369226e-01 -5.06514534e-02
-5.43563306e-01 -8.22115719e-01 6.79171503e-01 5.86448371e-01
1.98241875e-01 1.71528623e-01 -8.94404277e-02 1.01279795e+00
-2.03857154e-01 1.45489573e-01 -4.02377665e-01 5.16990125e-01
-2.61611670e-01 -1.16305959e+00 -3.38133514e-01 2.63700932e-01
-2.38308921e-01 -1.36635840e-01 -2.63639122e-01 4.36975539e-01
2.94257432e-01 8.84451330e-01 -4.70322259e-02 -5.20173371e-01
6.66449964e-01 -6.55475408e-02 5.50110698e-01 -3.66085440e-01
-8.10114026e-01 -2.66121238e-01 1.80063203e-01 -7.58150935e-01
-1.29068479e-01 -8.82985055e-01 -8.04581046e-01 -3.97936523e-01
1.99955299e-01 -1.35861278e-01 5.72257563e-02 7.50829935e-01
6.32280111e-01 2.80000031e-01 2.76205510e-01 -5.58558822e-01
-5.87172747e-01 -9.89482284e-01 -3.92725676e-01 6.59460902e-01
2.04559565e-01 -5.73144555e-01 8.68801847e-02 -1.36463448e-01] | [14.83632755279541, 0.818330705165863] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.