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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
51574bde-a381-44a8-99b6-4a9bd1112bb7 | learning-from-synthetic-data-generated-with | 2305.04282 | null | https://arxiv.org/abs/2305.04282v2 | https://arxiv.org/pdf/2305.04282v2.pdf | Learning from synthetic data generated with GRADE | Recently, synthetic data generation and realistic rendering has advanced tasks like target tracking and human pose estimation. Simulations for most robotics applications are obtained in (semi)static environments, with specific sensors and low visual fidelity. To solve this, we present a fully customizable framework for generating realistic animated dynamic environments (GRADE) for robotics research, first introduced in [1]. GRADE supports full simulation control, ROS integration, realistic physics, while being in an engine that produces high visual fidelity images and ground truth data. We use GRADE to generate a dataset focused on indoor dynamic scenes with people and flying objects. Using this, we evaluate the performance of YOLO and Mask R-CNN on the tasks of segmenting and detecting people. Our results provide evidence that using data generated with GRADE can improve the model performance when used for a pre-training step. We also show that, even training using only synthetic data, can generalize well to real-world images in the same application domain such as the ones from the TUM-RGBD dataset. The code, results, trained models, and the generated data are provided as open-source at https://eliabntt.github.io/grade-rr. | ['Aamir Ahmad', 'Chenghao Xu', 'Elia Bonetto'] | 2023-05-07 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [-1.43070385e-01 -7.11942017e-02 3.91570240e-01 -2.54456788e-01
-2.39467293e-01 -5.07838488e-01 5.54938614e-01 -2.19583750e-01
-4.90442067e-01 6.28113747e-01 -1.21693425e-01 -3.40334396e-03
2.96881646e-01 -7.91627824e-01 -9.82785821e-01 -3.41105491e-01
-4.67449099e-01 8.26844633e-01 5.43513298e-01 -4.54169631e-01
-3.55496287e-01 7.45655239e-01 -1.99028397e+00 2.25062206e-01
6.23535395e-01 8.22206736e-01 3.98357600e-01 1.17736733e+00
5.34425795e-01 7.13938534e-01 -6.94040656e-01 -1.04504593e-01
6.64818347e-01 -1.72247693e-01 -3.61462712e-01 1.46113738e-01
5.05866110e-01 -4.84453410e-01 -4.35782105e-01 5.67151487e-01
6.69508100e-01 3.17761064e-01 3.31897795e-01 -1.64748383e+00
-1.95387274e-01 1.19903378e-01 4.01589982e-02 -8.85543227e-02
7.67192602e-01 8.77573073e-01 1.04788005e-01 -6.82047665e-01
9.54854488e-01 1.43789494e+00 8.80200684e-01 9.06506777e-01
-1.05717993e+00 -5.18491805e-01 5.82043966e-03 3.19872983e-02
-1.36930752e+00 -2.45101377e-01 4.37977195e-01 -5.60781479e-01
7.46547401e-01 3.47352207e-01 1.18981779e+00 1.72603011e+00
2.09654793e-02 4.46807235e-01 1.19345880e+00 -2.36270428e-01
3.31578195e-01 3.16329718e-01 -1.82758957e-01 7.08966672e-01
2.80502588e-01 6.10342443e-01 -3.24443221e-01 1.39702648e-01
8.32107008e-01 -2.37136349e-01 -3.67339075e-01 -8.17500770e-01
-1.52355409e+00 6.02283657e-01 8.24380577e-01 -1.41994029e-01
-3.05634677e-01 4.88539755e-01 1.75469190e-01 1.08502787e-02
1.69910029e-01 3.06845278e-01 -3.63236547e-01 1.85710210e-02
-6.33751452e-01 8.83958340e-01 8.36496055e-01 1.03051603e+00
4.54913318e-01 1.82679057e-01 -1.01158343e-01 2.20060095e-01
3.08697939e-01 9.25857067e-01 2.62657344e-01 -1.39978921e+00
1.80862799e-01 4.03524429e-01 5.82916379e-01 -8.45567107e-01
-7.44520426e-01 -4.57016051e-01 -5.05516291e-01 8.71797383e-01
7.41828561e-01 -3.61209989e-01 -1.09164047e+00 1.54240394e+00
7.83922255e-01 2.65326440e-01 2.46350504e-02 1.30462444e+00
1.01537132e+00 5.74279189e-01 1.52450517e-01 4.86518443e-01
1.17169011e+00 -9.79411483e-01 -2.51847923e-01 -4.56983894e-01
4.45475936e-01 -5.54741442e-01 1.26641250e+00 5.18929482e-01
-8.71493816e-01 -8.17255139e-01 -9.70088899e-01 5.18678920e-03
-5.97763836e-01 2.11620480e-01 6.55848742e-01 6.77089155e-01
-1.17796278e+00 6.29065454e-01 -9.70466077e-01 -7.31137156e-01
2.46377930e-01 2.13979200e-01 -3.23225319e-01 3.10706161e-03
-1.14517796e+00 1.15687478e+00 2.76681334e-01 1.03155926e-01
-1.48745835e+00 -6.71062946e-01 -7.56967247e-01 -4.57295775e-01
1.54527172e-01 -1.00266039e+00 1.31071246e+00 -7.92988479e-01
-1.47068644e+00 8.96626115e-01 3.64125460e-01 -8.44772279e-01
1.19317472e+00 -4.39679950e-01 -1.62655145e-01 6.74228743e-02
-6.69642240e-02 1.18116939e+00 4.95132208e-01 -1.58405244e+00
-2.99670666e-01 -1.95152208e-01 3.66988748e-01 1.88987970e-01
3.42850178e-01 -2.35518236e-02 -2.79146045e-01 -3.92144918e-01
-4.87589449e-01 -1.22065413e+00 -6.68414176e-01 4.21739101e-01
-3.61025751e-01 4.24571037e-01 6.21310949e-01 -7.40212083e-01
2.46565789e-01 -1.76565826e+00 2.52498150e-01 -6.95644170e-02
-3.05851009e-02 3.38044524e-01 -2.38115601e-02 2.57902443e-01
-8.01628176e-03 -2.72146106e-01 -2.24785060e-01 -5.26988626e-01
9.41627845e-02 6.63845763e-02 -1.41972378e-01 5.82349122e-01
5.81029570e-03 9.63015854e-01 -7.96160877e-01 -3.79869848e-01
7.78507411e-01 8.49887609e-01 -4.83122915e-01 2.67839581e-01
-6.03185654e-01 9.91148412e-01 -1.32126838e-01 4.95236665e-01
7.25865185e-01 -6.46826997e-02 -1.55798525e-01 -3.66718508e-02
-1.49313569e-01 -7.22783580e-02 -1.44751954e+00 1.84190500e+00
-5.46625495e-01 7.06539333e-01 1.70374572e-01 -3.98068577e-01
6.43795252e-01 -4.14288007e-02 3.43100548e-01 -6.89560413e-01
4.11178380e-01 -1.20107889e-01 -9.74588320e-02 -5.53192556e-01
6.10581517e-01 1.92105040e-01 -9.07929763e-02 1.20695764e-02
-7.34749287e-02 -4.91769522e-01 3.06792289e-01 2.73450881e-01
9.52446520e-01 8.32154095e-01 -1.88789979e-01 -7.40958750e-02
1.11678600e-01 6.74894214e-01 2.24334627e-01 7.56778419e-01
-2.23906472e-01 8.46152067e-01 2.40496770e-02 -4.87587601e-01
-1.35557687e+00 -1.30670393e+00 8.65384489e-02 7.86942065e-01
3.71356875e-01 -2.52745032e-01 -9.91959929e-01 -3.65936100e-01
4.21014018e-02 6.71268523e-01 -7.42628098e-01 -6.31208811e-03
-7.09400594e-01 -5.83849192e-01 5.56850255e-01 3.37371141e-01
4.93014008e-01 -1.28136885e+00 -1.49245989e+00 1.87283382e-02
-1.65491790e-01 -1.34168339e+00 2.80495703e-01 -6.61456287e-02
-5.44080019e-01 -1.19798017e+00 -7.25848138e-01 -2.68846422e-01
4.21338081e-01 2.37866178e-01 1.36227524e+00 2.62203999e-03
-7.46955097e-01 6.99572146e-01 -4.95459139e-01 -5.95921934e-01
-4.71302897e-01 -1.53963044e-01 6.34061173e-02 -4.95518625e-01
-2.48908058e-01 -3.05548161e-01 -7.54054844e-01 4.09860879e-01
-7.59551108e-01 3.93478274e-01 2.17580959e-01 2.21045196e-01
4.24716532e-01 -5.49043715e-01 6.12116419e-02 -5.10842502e-01
9.94360819e-02 -1.89700261e-01 -8.18150282e-01 -8.44979435e-02
5.37107140e-03 -2.00959057e-01 5.76610684e-01 -5.95620513e-01
-9.01182353e-01 4.51128393e-01 -2.17734709e-01 -6.12759590e-01
-5.86171687e-01 -9.14999396e-02 -9.71762091e-02 -2.14160174e-01
1.30323434e+00 -7.45789856e-02 -1.09957166e-01 -3.12099099e-01
3.59922022e-01 2.85253823e-01 6.01110041e-01 -4.43922758e-01
9.31826949e-01 8.21883738e-01 3.56213897e-02 -6.51859045e-01
-5.23799956e-01 -6.22808672e-02 -5.27428687e-01 -6.37326121e-01
9.02346969e-01 -1.07701683e+00 -7.79160500e-01 5.74338973e-01
-1.04322290e+00 -1.10360587e+00 -5.59750915e-01 6.11001909e-01
-7.70207167e-01 1.82806756e-02 -2.70782888e-01 -8.37152243e-01
-4.66224141e-02 -1.08277714e+00 1.28361440e+00 3.07051301e-01
-4.37461175e-02 -6.97845459e-01 1.29921257e-01 2.18513638e-01
3.83742720e-01 1.01188350e+00 -1.90676168e-01 -1.17180757e-01
-8.79796207e-01 -1.13747001e-01 7.53183737e-02 6.03473932e-02
-4.06873316e-01 9.43191871e-02 -9.78503287e-01 -3.41588497e-01
-3.27410877e-01 -4.51882809e-01 6.27888918e-01 2.67184824e-01
9.12524223e-01 6.06625117e-02 -3.55407417e-01 6.27211571e-01
1.21209753e+00 -1.97922662e-01 5.70820689e-01 4.68868971e-01
8.04160357e-01 7.38783717e-01 9.81104136e-01 4.65836138e-01
4.40030783e-01 1.04645324e+00 8.36073697e-01 -1.74809262e-01
-4.57527876e-01 -1.41277209e-01 4.59723890e-01 -5.77466860e-02
-3.13766599e-01 -2.34165296e-01 -1.07981849e+00 4.18267578e-01
-1.77571452e+00 -9.72185194e-01 -5.30941069e-01 2.26976776e+00
3.94005239e-01 8.90424941e-03 5.48861504e-01 -2.12604143e-02
6.48213506e-01 -2.01513425e-01 -4.16020423e-01 -1.74579155e-02
-1.19851306e-02 2.08043069e-01 6.83066189e-01 5.24109662e-01
-1.04402494e+00 1.02362561e+00 5.73919821e+00 3.01353186e-01
-1.17146003e+00 2.54541367e-01 3.17449272e-01 -3.30490738e-01
2.35988110e-01 -6.32220358e-02 -7.53469527e-01 4.28296387e-01
1.19184422e+00 7.27085844e-02 5.05127192e-01 1.15627658e+00
5.35126984e-01 -3.01195115e-01 -9.14526641e-01 8.48657310e-01
-2.38194093e-02 -1.19417953e+00 -2.12795466e-01 5.67979598e-03
4.85032767e-01 3.76315862e-01 4.85113598e-02 3.98719639e-01
7.58045912e-01 -1.20466018e+00 1.12220442e+00 7.07031071e-01
5.92870831e-01 -6.59581542e-01 5.03090382e-01 5.48186064e-01
-9.89078104e-01 1.89342111e-01 -4.46221530e-01 4.21262858e-03
2.45184913e-01 3.41550469e-01 -1.12959063e+00 6.01079166e-01
9.95358229e-01 4.67993826e-01 -1.02419412e+00 1.32361042e+00
-1.85050368e-01 1.03096887e-01 -5.87983251e-01 -5.81264645e-02
-1.03159018e-01 4.12154123e-02 5.85964024e-01 1.25616300e+00
2.97957122e-01 -3.10758322e-01 3.52730930e-01 6.87990189e-01
4.39970732e-01 -2.62636125e-01 -7.82252014e-01 4.79483664e-01
6.57168925e-02 1.28075469e+00 -7.33120561e-01 -3.16382885e-01
1.31276891e-01 9.60269094e-01 1.24346204e-01 2.28066429e-01
-1.41705644e+00 -7.55302832e-02 6.37985408e-01 5.07528782e-01
1.33721873e-01 -5.35885513e-01 -2.88204383e-02 -1.13852692e+00
-1.22701414e-02 -7.85578847e-01 3.95428650e-02 -1.29317677e+00
-6.62184119e-01 8.51161182e-01 3.96582216e-01 -1.55179203e+00
-5.57769418e-01 -8.82868946e-01 -1.43327147e-01 5.98910332e-01
-1.13575280e+00 -1.51593292e+00 -1.10716784e+00 6.86266363e-01
3.80003333e-01 2.26921514e-01 6.91287220e-01 2.44278327e-01
-2.30881810e-01 1.25188172e-01 -3.20462704e-01 2.21807826e-02
5.00452578e-01 -1.21447778e+00 6.42971098e-01 8.78314197e-01
-1.33247390e-01 2.16719761e-01 1.25148213e+00 -6.95944667e-01
-1.32548201e+00 -1.35149682e+00 1.06044980e-02 -7.59514749e-01
3.08300793e-01 -6.91081822e-01 -5.07629871e-01 7.67233968e-01
1.02072738e-01 3.16681355e-01 9.37307477e-02 -5.48849165e-01
4.50837202e-02 3.89030688e-02 -1.49482608e+00 6.88134849e-01
1.42025661e+00 1.35425016e-01 -1.76382557e-01 6.02905035e-01
6.91293895e-01 -1.07193398e+00 -6.60831273e-01 5.12446046e-01
5.88740885e-01 -1.41393673e+00 1.32479727e+00 -2.85405368e-01
2.53520995e-01 -6.47159457e-01 -1.57483459e-01 -1.44416428e+00
6.52944818e-02 -3.46961826e-01 -1.57460183e-01 6.77898526e-01
4.29858938e-02 -3.72570276e-01 8.03836286e-01 3.49508792e-01
-1.73563510e-01 -3.01814079e-01 -7.93750644e-01 -7.97343135e-01
-1.57799527e-01 -5.72178662e-01 4.79926974e-01 3.90226036e-01
-7.14794993e-01 -9.74041075e-02 -3.92853945e-01 4.12742853e-01
7.67914474e-01 -9.46906880e-02 1.58079016e+00 -1.17430496e+00
-5.03707945e-01 1.07016377e-01 -6.11034334e-01 -6.50942802e-01
-6.99536726e-02 -5.33134639e-01 2.22498789e-01 -1.77308154e+00
-3.84705573e-01 -6.06864095e-01 5.66341639e-01 1.70863688e-01
1.40896425e-01 5.95240116e-01 4.35484201e-01 1.33485692e-02
-6.51684701e-01 5.01439393e-01 1.30432153e+00 1.60141304e-01
-7.71674514e-02 9.28768367e-02 -7.12695718e-02 7.70012319e-01
1.01704049e+00 -3.64195019e-01 -1.87012464e-01 -2.19013199e-01
2.80340295e-02 -7.41554648e-02 1.22450912e+00 -1.95404422e+00
-7.23001286e-02 -1.01815619e-01 8.10644329e-01 -4.96023536e-01
9.16089773e-01 -1.00404787e+00 7.11055219e-01 8.89970899e-01
6.00573495e-02 1.08651608e-01 4.63706970e-01 3.24951470e-01
4.10653383e-01 -4.35165539e-02 8.86074245e-01 -5.19484997e-01
-8.57528150e-01 1.87274471e-01 -2.32443050e-01 7.45244604e-03
1.36719692e+00 -2.16425896e-01 -4.11453396e-01 -4.71627116e-01
-1.03977025e+00 1.05232231e-01 9.68291759e-01 5.41267335e-01
5.64492047e-01 -1.04665887e+00 -7.77995706e-01 5.78656904e-02
1.25925273e-01 3.45052600e-01 2.70741493e-01 4.36023980e-01
-1.17869925e+00 1.55326381e-01 -5.49113989e-01 -8.14010322e-01
-1.13523769e+00 6.24597073e-01 7.15214491e-01 -1.57562599e-01
-6.40663564e-01 5.35681129e-01 1.04113735e-01 -7.80411541e-01
1.93639383e-01 -5.07503033e-01 5.34812473e-02 -4.26644742e-01
4.76307839e-01 3.90811116e-01 8.54222849e-02 -8.43785763e-01
-4.62081730e-01 4.21260357e-01 6.18886888e-01 -4.22446549e-01
1.30512691e+00 1.09130748e-01 4.40762997e-01 4.01061952e-01
5.43775678e-01 1.57259181e-02 -1.76250327e+00 3.65021557e-01
-4.50053126e-01 -5.34399807e-01 -3.10008198e-01 -9.36141551e-01
-9.26188290e-01 7.44641900e-01 1.15263379e+00 -2.14445386e-02
7.52450883e-01 -1.44763058e-02 2.91516215e-01 3.00846279e-01
9.26320612e-01 -6.88151836e-01 1.40168935e-01 3.68255109e-01
1.12590933e+00 -1.02121925e+00 -1.05546884e-01 -5.79925060e-01
-7.56105244e-01 7.37868726e-01 1.13593745e+00 -5.19971132e-01
2.25431114e-01 5.42375863e-01 1.77047819e-01 -6.48881420e-02
-4.46259737e-01 -5.38699329e-01 -1.42160505e-01 1.26911128e+00
2.89047994e-02 1.43800721e-01 2.67961115e-01 1.75113663e-01
-6.25756323e-01 1.48347586e-01 7.38074481e-01 9.57620025e-01
-2.91763753e-01 -8.25168788e-01 -7.46196866e-01 4.62043993e-02
-1.15141638e-01 3.33941519e-01 -4.52046365e-01 1.12711883e+00
4.31067020e-01 8.62206399e-01 7.25671351e-02 -3.79513621e-01
6.37627423e-01 -3.22125703e-01 7.69058287e-01 -5.15010595e-01
-8.50510836e-01 -4.53336328e-01 2.19109073e-01 -7.73958445e-01
-4.54100609e-01 -5.99830568e-01 -1.21608734e+00 -4.04766411e-01
2.11340412e-02 -4.25713919e-02 9.50609088e-01 5.64168632e-01
4.13223565e-01 8.39834690e-01 1.90877393e-01 -1.66635454e+00
-1.55767679e-01 -1.06763017e+00 -5.12924679e-02 4.78355080e-01
2.13328496e-01 -8.64476144e-01 -2.30520532e-01 1.33108795e-01] | [7.611990451812744, -1.0849730968475342] |
cde7671f-5d31-47c2-9193-37d266aa486b | pevl-position-enhanced-pre-training-and | 2205.11169 | null | https://arxiv.org/abs/2205.11169v2 | https://arxiv.org/pdf/2205.11169v2.pdf | PEVL: Position-enhanced Pre-training and Prompt Tuning for Vision-language Models | Vision-language pre-training (VLP) has shown impressive performance on a wide range of cross-modal tasks, where VLP models without reliance on object detectors are becoming the mainstream due to their superior computation efficiency and competitive performance. However, the removal of object detectors also deprives the capability of VLP models in explicit object modeling, which is essential to various position-sensitive vision-language (VL) tasks, such as referring expression comprehension and visual commonsense reasoning. To address the challenge, we introduce PEVL that enhances the pre-training and prompt tuning of VLP models with explicit object position modeling. Specifically, PEVL reformulates discretized object positions and language in a unified language modeling framework, which facilitates explicit VL alignment during pre-training, and also enables flexible prompt tuning for various downstream tasks. We show that PEVL enables state-of-the-art performance of detector-free VLP models on position-sensitive tasks such as referring expression comprehension and phrase grounding, and also improves the performance on position-insensitive tasks with grounded inputs. We make the data and code for this paper publicly available at https://github.com/thunlp/PEVL. | ['Maosong Sun', 'Tat-Seng Chua', 'Zhiyuan Liu', 'Wei Ji', 'Ao Zhang', 'Qianyu Chen', 'Yuan YAO'] | 2022-05-23 | null | null | null | null | ['visual-relationship-detection', 'phrase-grounding', 'visual-commonsense-reasoning'] | ['computer-vision', 'natural-language-processing', 'reasoning'] | [ 1.44605428e-01 -9.99974310e-02 -1.88915253e-01 -4.18253332e-01
-9.20409262e-01 -6.84561133e-01 6.15114272e-01 1.31770015e-01
-5.67091525e-01 1.94155008e-01 2.63570607e-01 -4.12750185e-01
1.56553984e-01 -6.26492977e-01 -8.19547057e-01 -4.76163000e-01
4.15875763e-01 4.21044946e-01 3.93370062e-01 -2.90541470e-01
2.99697787e-01 5.51298559e-01 -1.51636958e+00 5.87005258e-01
7.49887586e-01 8.42035890e-01 6.41899765e-01 6.85202897e-01
-3.77969831e-01 1.00367522e+00 -2.70582110e-01 -3.84369105e-01
1.41729284e-02 -7.83765167e-02 -6.82886183e-01 -2.56204098e-01
7.66299605e-01 -2.71258831e-01 -4.10678327e-01 1.08815658e+00
5.03322303e-01 1.95497781e-01 4.78813887e-01 -1.31031752e+00
-1.32825613e+00 4.83958751e-01 -7.39299655e-01 3.28371912e-01
2.62700915e-01 5.57935894e-01 1.32129228e+00 -1.28002226e+00
3.90521705e-01 1.65386057e+00 5.16674161e-01 6.71745420e-01
-1.38804150e+00 -5.18221140e-01 5.26971817e-01 4.14048940e-01
-1.34995997e+00 -5.83814561e-01 7.90846109e-01 -4.38457608e-01
1.48229277e+00 8.91534165e-02 5.84905684e-01 1.08567929e+00
-1.55580357e-01 1.28274870e+00 1.08469105e+00 -5.46965837e-01
-1.32108778e-02 -7.99504668e-02 4.11914140e-01 9.05278206e-01
-2.64983270e-02 1.25571549e-01 -7.16010988e-01 2.20488086e-01
8.10978770e-01 -2.05826968e-01 -3.33842903e-01 -3.22273910e-01
-1.37445199e+00 8.46556127e-01 7.69699275e-01 1.67122513e-01
-3.14032912e-01 4.79295015e-01 3.45164746e-01 -1.57465916e-02
9.34872255e-02 4.34406340e-01 -3.04976642e-01 7.19068721e-02
-7.23997533e-01 1.92575067e-01 4.02265757e-01 1.19071698e+00
5.36133230e-01 1.81031957e-01 -6.44979417e-01 8.04812312e-01
4.37057227e-01 7.38991320e-01 1.85676422e-02 -1.19919133e+00
6.83987796e-01 6.10373795e-01 -1.00536883e-01 -8.06353688e-01
-4.90459800e-01 -3.84076416e-01 -6.33877575e-01 1.14437915e-01
5.17698824e-01 2.73371607e-01 -1.05936801e+00 2.03742194e+00
4.66970913e-02 -2.16885135e-01 7.58714080e-02 1.01106465e+00
1.35976076e+00 7.70724177e-01 4.18300420e-01 -1.53612476e-02
1.65692890e+00 -1.17127550e+00 -4.81320292e-01 -6.34880722e-01
6.15686715e-01 -5.39883435e-01 1.78170335e+00 2.12761134e-01
-1.14260781e+00 -6.45615280e-01 -7.22729385e-01 -9.51886058e-01
-2.78018415e-01 7.03305081e-02 7.94613957e-01 1.24365881e-01
-1.23898423e+00 -5.95986694e-02 -7.91851044e-01 -3.59637827e-01
8.78795087e-01 2.66514242e-01 -1.18433133e-01 -2.69453466e-01
-1.06319046e+00 9.74862695e-01 4.29676801e-01 1.46676660e-01
-7.18620956e-01 -7.75577128e-01 -1.05488861e+00 1.30441174e-01
5.11300623e-01 -9.72536027e-01 1.42562532e+00 -6.55419886e-01
-1.21429241e+00 1.34079552e+00 -4.41193134e-01 -3.94874722e-01
4.63026643e-01 -4.01962936e-01 -1.50171407e-02 1.80951387e-01
1.97208881e-01 1.34188175e+00 8.68952096e-01 -1.28544104e+00
-4.89204317e-01 -2.98975319e-01 3.57172459e-01 4.30081218e-01
1.77712310e-02 5.83374761e-02 -7.49174118e-01 -3.84553641e-01
1.78058192e-01 -7.04886734e-01 7.96949118e-02 4.16135997e-01
-4.53773558e-01 -5.11485517e-01 5.87485611e-01 -5.69311559e-01
7.67282486e-01 -2.21229339e+00 2.23235711e-01 -3.89429659e-01
4.18669522e-01 1.26995131e-01 -4.76941705e-01 1.28423527e-01
1.32560089e-01 -2.76928451e-02 -2.24252626e-01 -5.82452476e-01
2.65387207e-01 4.32873249e-01 -7.03033507e-01 2.26131573e-01
3.64817917e-01 1.51416469e+00 -9.26182270e-01 -6.59736156e-01
2.67233133e-01 4.08726186e-01 -6.39350176e-01 -2.81742085e-02
-7.51740634e-01 4.77505445e-01 -3.73928934e-01 7.27336884e-01
4.56462473e-01 -4.86762226e-01 -3.07470828e-01 -3.84168804e-01
-1.80117428e-01 2.95511514e-01 -7.05424130e-01 1.90760183e+00
-4.62921232e-01 8.47189307e-01 4.35981452e-02 -7.00447679e-01
5.51122725e-01 -2.28002723e-02 -6.70982003e-02 -1.11512911e+00
2.45183371e-02 -7.56846443e-02 1.59424424e-01 -4.60709721e-01
5.15348911e-01 -1.08081415e-01 -9.95710865e-02 2.76675701e-01
1.45985465e-02 -3.73367220e-01 1.41253784e-01 4.58963871e-01
7.12518156e-01 3.16986084e-01 4.13707376e-01 -1.92863837e-01
4.74058926e-01 2.09402978e-01 2.82547504e-01 9.40513849e-01
-2.84379989e-01 4.90674496e-01 4.71059412e-01 -4.60351780e-02
-7.52655506e-01 -1.38630879e+00 -1.33401155e-01 1.63223302e+00
2.91861892e-01 -2.15507805e-01 -3.38158846e-01 -2.65742183e-01
1.13037258e-01 1.19267619e+00 -4.42759395e-01 -9.97434556e-02
-5.57063222e-01 -5.65694690e-01 5.67905247e-01 8.86376262e-01
6.17626071e-01 -1.45552123e+00 -7.03892112e-01 -8.44123438e-02
-1.83725163e-01 -1.39464402e+00 -2.54943013e-01 2.32099012e-01
-6.65062964e-01 -8.42397034e-01 -5.76663494e-01 -9.26125526e-01
5.97674966e-01 4.17876542e-01 1.35823739e+00 9.27403420e-02
-2.09759250e-01 5.01267076e-01 -1.09331943e-01 -4.56522316e-01
-2.26871356e-01 -7.61717930e-02 -8.91144872e-02 -4.63205278e-01
4.58061308e-01 -3.21030706e-01 -4.54860300e-01 -9.90632251e-02
-5.64754665e-01 5.08521974e-01 5.05391419e-01 7.45635033e-01
9.57184315e-01 -6.21040642e-01 4.26483959e-01 -6.17513061e-01
5.36100924e-01 -1.64422132e-02 -8.22575808e-01 3.78650755e-01
-1.75131813e-01 1.75143406e-01 3.07182580e-01 -4.69279945e-01
-9.13249433e-01 -1.76219523e-01 -3.40633452e-01 -5.83749771e-01
-9.24597587e-03 3.72888565e-01 -2.21120492e-01 -5.48667908e-02
6.86008155e-01 4.42739278e-01 -2.86522090e-01 -3.44937146e-01
8.60860586e-01 2.33643398e-01 8.90201390e-01 -7.66118288e-01
7.12182343e-01 4.82105762e-01 -1.10773019e-01 -7.44281054e-01
-1.34962022e+00 -5.09164929e-01 -5.69344878e-01 2.69606471e-01
1.10606706e+00 -1.08925927e+00 -9.17325675e-01 5.18850982e-01
-1.45790505e+00 -6.74221039e-01 -2.88077801e-01 1.71010181e-01
-7.04312503e-01 1.61850289e-01 -6.10481322e-01 -6.64675713e-01
-3.80269110e-01 -1.15185058e+00 1.14266348e+00 2.86219418e-01
-1.52388126e-01 -9.38475788e-01 -3.09332222e-01 4.46051329e-01
3.07361215e-01 -1.30808190e-01 1.30553234e+00 -3.78228813e-01
-8.55601013e-01 2.20527619e-01 -7.98902154e-01 1.70016631e-01
-2.48468086e-01 -2.51747161e-01 -1.28565228e+00 -1.67358905e-01
-1.18355647e-01 -6.62833571e-01 1.26483881e+00 5.32163799e-01
1.30272508e+00 -2.99907755e-02 -3.31791967e-01 9.82286870e-01
1.36636627e+00 -1.00370601e-01 4.31928724e-01 1.71067506e-01
1.08523262e+00 5.16415596e-01 3.99398714e-01 1.97675545e-02
7.51640141e-01 7.27131307e-01 5.42678654e-01 -2.30330259e-01
-4.95717168e-01 -3.97761822e-01 1.93028510e-01 4.40791130e-01
1.43521249e-01 -2.65383691e-01 -1.23840845e+00 6.43975377e-01
-1.89519131e+00 -9.56120133e-01 -1.66130573e-01 1.78399885e+00
9.43275332e-01 1.63324162e-01 -1.28667697e-01 -3.15935344e-01
3.76529217e-01 3.07237744e-01 -8.04948568e-01 -2.98789978e-01
-3.45155478e-01 1.26593262e-01 2.40679398e-01 7.47985303e-01
-9.90226269e-01 1.58397508e+00 5.65312576e+00 6.42865419e-01
-1.03403258e+00 4.26334351e-01 2.90240228e-01 -1.30588442e-01
-2.65366971e-01 -1.64735854e-01 -9.46190894e-01 -6.54776618e-02
3.89325261e-01 -1.00419767e-01 4.41006243e-01 8.21584344e-01
1.10997036e-01 -2.06781358e-01 -1.49090099e+00 1.40041387e+00
1.40237138e-01 -1.44669724e+00 2.42827013e-01 -3.03427041e-01
4.00455117e-01 4.85505134e-01 3.21920067e-01 6.09739602e-01
2.10270762e-01 -1.15291035e+00 8.99355650e-01 2.93665886e-01
9.26995039e-01 -3.88467789e-01 2.52453804e-01 3.80169898e-01
-1.05433345e+00 -1.07394449e-01 -4.74381894e-01 -1.56938031e-01
2.37139031e-01 1.89557463e-01 -5.19376159e-01 7.98035115e-02
5.73481441e-01 6.39501870e-01 -6.08139455e-01 7.83379316e-01
-7.28881955e-01 4.79542911e-01 -3.15716773e-01 3.83790880e-02
3.52083504e-01 1.68997973e-01 5.99559605e-01 1.27509296e+00
-8.38243514e-02 2.42121667e-01 2.23466754e-01 1.43099260e+00
-1.66914389e-01 -1.14450477e-01 -4.99529421e-01 -1.58344693e-02
6.24427080e-01 1.01287138e+00 -5.18961430e-01 -2.74364322e-01
-6.12942874e-01 8.17112148e-01 5.64405143e-01 6.04635596e-01
-9.46936190e-01 -1.99193526e-02 6.67352021e-01 6.78793564e-02
3.53647351e-01 -4.40547764e-01 -4.57505196e-01 -1.18356752e+00
4.55158344e-03 -7.06459701e-01 3.47374976e-01 -1.16998816e+00
-1.25863862e+00 3.79686028e-01 6.59416988e-02 -7.70469010e-01
-1.54804587e-01 -1.03498423e+00 -4.62508589e-01 1.03419125e+00
-1.76001382e+00 -1.59157085e+00 -3.85205209e-01 7.50408292e-01
6.58609152e-01 4.40244414e-02 6.74864352e-01 -2.02817217e-01
-3.88903707e-01 5.55389762e-01 -4.43916887e-01 2.04316825e-01
5.59913933e-01 -1.09995925e+00 4.41665649e-01 9.49481547e-01
5.22609353e-01 8.36646914e-01 5.49870253e-01 -3.44640344e-01
-1.57819152e+00 -1.03490520e+00 7.50408292e-01 -7.95486927e-01
7.50102282e-01 -5.51862001e-01 -1.06241858e+00 8.92180681e-01
2.89867260e-02 2.08604157e-01 2.88395941e-01 3.02502602e-01
-7.74706960e-01 1.57333910e-01 -8.12193573e-01 8.60145628e-01
1.35362911e+00 -9.42831397e-01 -9.80001748e-01 6.61999226e-01
9.11585808e-01 -6.67800903e-01 -2.55182564e-01 1.28114328e-01
2.85842419e-01 -7.64627099e-01 1.26818120e+00 -5.57667315e-01
3.85456979e-01 -3.75814319e-01 -5.40089607e-01 -8.53091896e-01
-5.60931802e-01 -2.13989884e-01 -3.56572837e-01 1.07030654e+00
2.71083087e-01 -5.67224741e-01 2.85464048e-01 5.62295675e-01
-1.72654763e-01 -6.42190397e-01 -8.53011250e-01 -6.32382989e-01
6.06925786e-02 -8.05205524e-01 8.17239881e-02 6.28961027e-01
-3.75897348e-01 6.55240655e-01 1.36159047e-01 3.78248751e-01
5.79067886e-01 4.82897848e-01 5.71048796e-01 -9.13428426e-01
-4.56506312e-01 -7.12301910e-01 -1.50298640e-01 -1.42906177e+00
5.01850903e-01 -1.18092299e+00 1.46547377e-01 -1.97882342e+00
3.59331310e-01 -2.98512787e-01 -2.23842636e-01 9.37182903e-01
-2.21184283e-01 2.54173279e-01 6.48987293e-01 3.82675916e-01
-7.57369280e-01 5.12766242e-01 1.46422911e+00 -2.27282286e-01
-3.13743740e-01 -2.68058240e-01 -9.65015411e-01 9.46759820e-01
5.67943692e-01 -2.41337925e-01 -6.19049549e-01 -9.92750347e-01
2.34641969e-01 -1.36232927e-01 9.74533260e-01 -7.48303711e-01
1.04473047e-01 -1.57530233e-01 3.85146141e-01 -7.43560135e-01
5.47934473e-01 -3.77576977e-01 -5.74669898e-01 3.12555075e-01
-4.15780038e-01 1.15637317e-01 6.44626856e-01 3.38610500e-01
-9.57292318e-02 -4.08384055e-02 8.05309057e-01 -2.32680589e-01
-1.41335917e+00 2.39636332e-01 2.40984619e-01 4.03508276e-01
6.47928894e-01 -2.27365330e-01 -7.41829336e-01 -2.49276161e-01
-6.65055573e-01 5.00355005e-01 5.77929318e-01 4.55453068e-01
8.02999675e-01 -9.86796856e-01 -7.04060733e-01 -4.68974337e-02
3.45628977e-01 4.03204113e-01 2.10496664e-01 1.05153263e+00
-3.67872298e-01 5.90657949e-01 -1.78987071e-01 -9.15519655e-01
-1.09999311e+00 8.22717190e-01 3.63841474e-01 3.06995939e-02
-1.00900626e+00 1.38468611e+00 9.11099553e-01 -2.93365479e-01
2.50655949e-01 -5.54211497e-01 -2.08753608e-02 -1.03697315e-01
6.29445910e-01 -1.82381645e-01 -2.81834781e-01 -6.05793774e-01
-6.32922828e-01 6.18037701e-01 -1.95093066e-01 -1.41952679e-01
1.05816519e+00 -1.44687623e-01 -1.28880396e-01 6.31272912e-01
8.86886239e-01 -9.93309990e-02 -1.55720377e+00 -4.55887854e-01
-6.49953559e-02 -1.84671909e-01 2.57155806e-01 -1.03964126e+00
-7.89340436e-01 1.36326540e+00 2.23399222e-01 -3.29677403e-01
1.09441531e+00 4.07867223e-01 5.62700510e-01 6.39006555e-01
2.99973339e-01 -6.74074888e-01 2.69188762e-01 8.71457636e-01
1.29193294e+00 -1.36095512e+00 -9.88878533e-02 -4.19516414e-01
-7.65999734e-01 7.71636426e-01 8.13452840e-01 -1.47624677e-02
1.35648206e-01 2.87936777e-01 8.86195377e-02 -2.16066569e-01
-7.93923438e-01 -5.86593151e-01 4.38979656e-01 7.67606556e-01
4.09622520e-01 9.76158248e-04 1.58755511e-01 5.31623125e-01
-2.22030863e-01 -1.17899083e-01 4.26215343e-02 7.15401351e-01
-4.41630304e-01 -6.03489518e-01 -2.20865726e-01 2.37657323e-01
2.61902921e-02 -7.66744494e-01 -2.18116358e-01 9.84256029e-01
5.32489754e-02 6.47764087e-01 1.28756940e-01 1.42147020e-01
3.19101095e-01 2.44078770e-01 8.68809819e-01 -7.50359595e-01
-3.96075010e-01 -5.81697188e-02 6.75617810e-03 -5.75697184e-01
-2.71762133e-01 -4.11897063e-01 -1.78525317e+00 -1.32556170e-01
3.92310321e-03 -4.82264578e-01 2.88855910e-01 1.01547682e+00
3.63514274e-01 6.01596773e-01 -2.26154804e-01 -8.23153436e-01
-6.74478114e-01 -7.07634866e-01 -1.15000166e-01 3.75344813e-01
5.08584976e-01 -8.32948685e-01 -1.47853792e-02 2.28095148e-02] | [10.55911922454834, 1.7133899927139282] |
dcc4a176-d367-43e4-a2db-3a8dddc3da4f | hqp-a-human-annotated-dataset-for-detecting | 2304.14931 | null | https://arxiv.org/abs/2304.14931v2 | https://arxiv.org/pdf/2304.14931v2.pdf | HQP: A Human-Annotated Dataset for Detecting Online Propaganda | Online propaganda poses a severe threat to the integrity of societies. However, existing datasets for detecting online propaganda have a key limitation: they were annotated using weak labels that can be noisy and even incorrect. To address this limitation, our work makes the following contributions: (1) We present HQP: a novel dataset (N=30,000) for detecting online propaganda with high-quality labels. To the best of our knowledge, HQP is the first dataset for detecting online propaganda that was created through human annotation. (2) We show empirically that state-of-the-art language models fail in detecting online propaganda when trained with weak labels (AUC: 64.03). In contrast, state-of-the-art language models can accurately detect online propaganda when trained with our high-quality labels (AUC: 92.25), which is an improvement of ~44%. (3) To address the cost of labeling, we extend our work to few-shot learning. Specifically, we show that prompt-based learning using a small sample of high-quality labels can still achieve a reasonable performance (AUC: 80.27). Finally, we discuss implications for the NLP community to balance the cost and quality of labeling. Crucially, our work highlights the importance of high-quality labels for sensitive NLP tasks such as propaganda detection. | ['Stefan Feuerriegel', 'Dominique Geissler', 'Dominik Bär', 'Abdurahman Maarouf'] | 2023-04-28 | null | null | null | null | ['propaganda-detection'] | ['natural-language-processing'] | [ 4.07001153e-02 4.36250716e-02 -4.54097331e-01 8.56959596e-02
-1.05345678e+00 -7.57299721e-01 8.64565611e-01 5.11814177e-01
-5.67621589e-01 7.33731985e-01 4.82204795e-01 -5.23511708e-01
9.17907581e-02 -9.49722946e-01 -3.95235181e-01 -3.80482227e-01
-2.08874509e-01 1.03584476e-01 2.90832072e-01 -2.60769248e-01
4.50949967e-01 1.13907106e-01 -1.04125082e+00 1.21816091e-01
1.31425273e+00 3.11393142e-01 -5.78993857e-01 5.61672091e-01
-1.10107921e-01 1.19177604e+00 -9.95254099e-01 -6.12754166e-01
2.15576336e-01 -5.77320635e-01 -9.74378884e-01 -3.97972077e-01
6.53297544e-01 -7.21325517e-01 -2.67332315e-01 1.25896716e+00
5.44949532e-01 -1.99360728e-01 7.63239563e-01 -8.29617143e-01
-1.09417260e+00 6.61392629e-01 -5.39068222e-01 3.41952056e-01
5.70389986e-01 1.70384318e-01 9.80221927e-01 -6.45271003e-01
1.05285251e+00 1.24228501e+00 8.08976412e-01 4.73068029e-01
-1.09167886e+00 -8.81231666e-01 -5.72731122e-02 1.82097435e-01
-1.15955341e+00 -4.35369760e-01 5.77051818e-01 -8.33117008e-01
7.02657044e-01 -5.35232387e-02 6.72250092e-01 1.37821841e+00
6.62215352e-02 6.23409152e-01 1.51212847e+00 -6.14285290e-01
3.12401921e-01 -5.48654310e-02 5.12914240e-01 1.22592103e+00
4.69157487e-01 -7.85839111e-02 -5.88133276e-01 -7.96805501e-01
2.53575236e-01 -3.48402619e-01 2.31988654e-01 4.72633690e-01
-7.96360016e-01 1.44047058e+00 2.69657046e-01 5.11459231e-01
-5.09400740e-02 6.34342730e-02 5.95816553e-01 2.93987364e-01
1.00053227e+00 7.45030046e-01 -4.59836144e-03 -4.12904352e-01
-9.50992465e-01 2.67939027e-02 1.10125875e+00 3.03877681e-01
3.86597872e-01 1.88210811e-02 -1.95814565e-01 6.57002330e-01
2.95850390e-04 8.30667675e-01 -2.68958975e-02 -7.38459647e-01
3.15780818e-01 3.05769861e-01 3.39001715e-01 -1.35515320e+00
-6.31143451e-01 -8.08874369e-02 -4.18776721e-01 -6.01411909e-02
7.80667484e-01 -5.67027330e-01 -8.59794915e-01 1.64291906e+00
3.65848362e-01 2.31947564e-02 -4.36126500e-01 6.55843616e-01
8.07854593e-01 8.35833848e-01 4.20985907e-01 -3.78078520e-01
1.30852342e+00 -7.13168204e-01 -8.90083313e-01 -1.72440708e-01
9.89120781e-01 -7.88853288e-01 8.58019650e-01 1.44440949e-01
-5.01912117e-01 9.63275284e-02 -6.59724236e-01 1.93102971e-01
-6.43745005e-01 -2.94127077e-01 7.87712514e-01 9.83104289e-01
-6.29302979e-01 2.40015179e-01 -5.62302291e-01 -8.25562060e-01
6.17047489e-01 -4.23947185e-01 -2.59250347e-02 2.02001721e-01
-1.47255445e+00 1.13972270e+00 2.68962234e-02 -4.41933870e-01
-9.45718765e-01 -6.43343568e-01 -8.39584649e-01 -3.22355807e-01
5.13807237e-01 -1.14943944e-01 9.39713717e-01 -8.60063374e-01
-1.05184650e+00 9.98969734e-01 -4.60920744e-02 -5.37642360e-01
4.48943704e-01 -4.23105478e-01 -6.09787405e-01 3.60376418e-01
6.38614655e-01 2.37471446e-01 6.53196871e-01 -1.01196361e+00
-4.96281981e-01 -9.63555872e-02 3.21327075e-02 -4.12108809e-01
-6.03814006e-01 8.09837043e-01 4.64811563e-01 -5.95948875e-01
-4.06602293e-01 -6.51486635e-01 1.47394687e-01 -1.19255610e-01
-4.20202792e-01 -4.56084222e-01 8.03966999e-01 -1.14603531e+00
1.41072023e+00 -1.97566736e+00 -5.09210706e-01 -3.67810354e-02
3.72883141e-01 7.32340813e-01 -2.63745278e-01 7.32686937e-01
5.25345862e-01 7.37024307e-01 7.06072375e-02 1.96568191e-01
4.52766009e-03 -1.70757502e-01 -5.07299483e-01 9.05957520e-01
1.47404820e-01 1.03675270e+00 -1.43353462e+00 -6.17153764e-01
-5.77272587e-02 1.92278549e-01 -2.46064201e-01 -1.97906181e-01
-2.19355226e-02 9.45226029e-02 -3.66210938e-01 8.04990053e-01
3.79457235e-01 -4.17083412e-01 2.35344157e-01 3.18260610e-01
-4.12816972e-01 3.82706434e-01 -5.24456203e-01 1.13266301e+00
-5.14341965e-02 8.72480869e-01 1.26159415e-01 -5.49077034e-01
6.84974313e-01 9.30903032e-02 4.35526371e-01 -7.69846082e-01
3.71731162e-01 2.12177143e-01 -9.56593901e-02 -6.82275057e-01
2.18834355e-01 -1.96620718e-01 -4.19121504e-01 1.06629145e+00
-7.12959692e-02 2.48528525e-01 2.83335149e-01 5.84215522e-01
1.52448320e+00 -3.47398311e-01 7.13350117e-01 -7.60573000e-02
8.87765139e-02 3.30893725e-01 4.89694268e-01 1.39606011e+00
-9.06830251e-01 -1.52329445e-01 5.39825618e-01 -7.49948621e-01
-1.11449611e+00 -9.79443371e-01 -2.15825990e-01 1.66360009e+00
-1.41954869e-01 -5.20492315e-01 -5.64860225e-01 -1.03339410e+00
-1.21536404e-01 5.78148067e-01 -6.04846299e-01 2.12305322e-01
-5.81080019e-01 -9.13440228e-01 1.14520085e+00 1.22961707e-01
4.60090637e-01 -9.64232326e-01 -6.92881823e-01 3.11384499e-01
-4.18949783e-01 -1.03145790e+00 -1.12201288e-01 -1.62599608e-01
-2.02312708e-01 -1.48274636e+00 -3.79583001e-01 -8.21435690e-01
5.06027341e-01 3.91192943e-01 7.10038185e-01 3.42308789e-01
-4.18994009e-01 3.34407181e-01 -7.93270350e-01 -4.53499049e-01
-6.97300434e-01 3.63918841e-02 1.27236813e-01 -5.14025569e-01
4.71487105e-01 -2.14242861e-01 -1.47725210e-01 -2.93215010e-02
-5.77526748e-01 -1.70731291e-01 2.21998692e-01 9.02883351e-01
-2.76785403e-01 -7.39474520e-02 1.12138700e+00 -1.15951705e+00
8.95705879e-01 -5.72941244e-01 -4.56720710e-01 2.61215776e-01
-2.64998883e-01 -4.29395527e-01 7.27648675e-01 -4.82724279e-01
-1.07241178e+00 -3.46415430e-01 -1.05940506e-01 5.37820578e-01
-2.24895015e-01 3.40993106e-01 8.51079762e-01 -2.73650914e-01
1.27007842e+00 -1.20137572e-01 -1.25395566e-01 -4.31525856e-01
6.69065416e-01 9.59388673e-01 1.33231416e-01 -3.04357201e-01
9.05355513e-01 7.05502689e-01 -2.98260570e-01 -9.85607386e-01
-1.51712275e+00 -7.74437129e-01 -6.18883491e-01 -2.67491043e-01
7.18558252e-01 -8.00634384e-01 -5.01463830e-01 6.24943852e-01
-1.20490217e+00 -5.05745709e-01 1.27354696e-01 2.61338681e-01
-1.23400874e-01 8.15810323e-01 -1.35795510e+00 -1.34864080e+00
-6.66609526e-01 -2.27270201e-01 8.64543378e-01 -3.57633419e-02
-3.28116953e-01 -1.08383298e+00 3.32382709e-01 5.02979636e-01
3.33981663e-01 5.23123324e-01 8.48010361e-01 -9.13665533e-01
2.06438512e-01 -2.40585953e-01 -5.37844956e-01 -1.03137389e-01
2.74473310e-01 -2.72633135e-03 -9.09477413e-01 -2.11236820e-01
-2.82828301e-01 -1.00335562e+00 8.93801093e-01 1.70384556e-01
3.91201973e-01 -8.00409436e-01 -1.83887780e-01 -1.43055826e-01
1.08913672e+00 -5.95350191e-02 3.80903840e-01 5.24787962e-01
3.72935563e-01 5.51226914e-01 7.03112006e-01 6.51784897e-01
1.22991495e-01 3.36789876e-01 4.21299897e-02 4.43775021e-02
-1.63842931e-01 -6.42028689e-01 5.61519861e-01 6.63450599e-01
1.74316019e-01 -3.26666310e-02 -1.16496754e+00 6.64816797e-01
-1.75652909e+00 -1.44847107e+00 -2.74924874e-01 1.86097920e+00
9.94633257e-01 -6.87841652e-03 4.34287399e-01 1.66635141e-02
9.72353458e-01 5.67826807e-01 2.61574872e-02 -5.28252363e-01
-7.01897666e-02 1.77605569e-01 3.62270325e-01 7.71452248e-01
-1.55543649e+00 1.23202085e+00 7.31061077e+00 1.06142426e+00
-8.27763557e-01 6.57809019e-01 2.39565849e-01 -1.49185240e-01
6.42897859e-02 5.37350699e-02 -7.36228287e-01 6.59182727e-01
8.46091807e-01 -1.00389905e-01 4.02529448e-01 7.12654889e-01
3.45559359e-01 -8.01203325e-02 -3.38331670e-01 6.14502251e-01
6.07472777e-01 -1.46936381e+00 -2.07694516e-01 2.15743780e-01
1.03570545e+00 2.27978945e-01 -3.12048256e-01 4.55756992e-01
7.20243752e-01 -9.41129446e-01 7.51678169e-01 1.55273005e-02
5.98425925e-01 -3.86491328e-01 7.73908138e-01 7.32188642e-01
-6.88937604e-01 -2.87242115e-01 -3.46786231e-01 -6.25760317e-01
6.70827448e-01 8.04603696e-01 -9.27756906e-01 1.90795764e-01
2.94310033e-01 6.22487664e-01 -6.83376014e-01 1.00736284e+00
-6.73204243e-01 1.20801210e+00 -3.18533361e-01 -5.44525087e-01
5.82349062e-01 3.01599205e-01 4.12301302e-01 1.44759643e+00
-6.81699195e-04 2.10732386e-01 5.49084365e-01 5.95487356e-01
-8.68573710e-02 2.61021107e-01 -8.32502425e-01 -2.81991541e-01
7.74755359e-01 1.19000685e+00 -4.92261618e-01 -3.13721716e-01
-2.14122027e-01 5.28585315e-01 7.08768606e-01 1.10082395e-01
-8.55426729e-01 -2.69201100e-01 6.44523576e-02 1.53414950e-01
-2.91610181e-01 -2.95334965e-01 -1.68579727e-01 -1.02467334e+00
-3.94535750e-01 -6.69236183e-01 7.09902823e-01 -1.72534943e-01
-1.79443491e+00 -6.18303381e-02 -2.53355086e-01 -6.42506242e-01
3.23440507e-02 -5.70784450e-01 -5.15122414e-01 1.42315671e-01
-1.17333937e+00 -1.39983881e+00 -9.81230661e-02 6.83286926e-03
4.36264873e-01 2.40889061e-02 8.63873124e-01 1.94027379e-01
-3.48609924e-01 1.71292320e-01 2.39749271e-02 6.98214352e-01
9.56069708e-01 -7.84787476e-01 2.04647362e-01 9.54213142e-01
4.86385413e-02 4.49552178e-01 5.07884383e-01 -1.12000489e+00
-1.11537325e+00 -1.06755984e+00 1.36021006e+00 -7.14600503e-01
1.39112449e+00 -4.04924214e-01 -7.16709137e-01 4.66615111e-01
5.84370233e-02 -3.06530237e-01 1.00152504e+00 5.32246530e-01
-9.89137232e-01 5.59448779e-01 -1.27505171e+00 4.37942773e-01
1.19010425e+00 -7.22824633e-01 -7.74306893e-01 9.94946122e-01
4.33281481e-01 1.36100307e-01 -6.35824740e-01 -1.26906127e-01
6.88094735e-01 -5.26020050e-01 6.39282465e-01 -7.86730766e-01
4.20941412e-01 1.48120403e-01 2.27840230e-01 -9.57641363e-01
-6.05806351e-01 -5.96373618e-01 -7.88199231e-02 1.29178286e+00
1.86609447e-01 -6.92485213e-01 4.66913849e-01 1.89099103e-01
3.42276067e-01 -2.91968673e-01 -1.15204632e+00 -1.05740917e+00
2.39111662e-01 -2.22806573e-01 -1.78460121e-01 1.41490960e+00
6.45240963e-01 4.64647770e-01 -1.05089819e+00 -7.79537633e-02
7.04222620e-01 -5.59442453e-02 6.90141022e-01 -1.20024443e+00
1.41216680e-01 -2.07548663e-01 -1.87929243e-01 -6.08738482e-01
5.09588122e-01 -7.16400564e-01 1.82828810e-02 -1.69717062e+00
6.84892237e-01 -3.68526697e-01 2.99643815e-01 8.17198515e-01
-1.09413974e-01 4.69414741e-01 2.55172819e-01 2.04454407e-01
-8.87211442e-01 2.50458300e-01 9.13862169e-01 -2.50139773e-01
-2.15513721e-01 -6.57075465e-01 -7.01742530e-01 9.10206258e-01
9.51770127e-01 -6.83656514e-01 3.25705290e-01 -3.62815440e-01
4.54020232e-01 -4.42786098e-01 5.19999206e-01 -6.12995744e-01
3.77696157e-01 -5.49190879e-01 3.57837975e-02 -3.18684667e-01
5.14590293e-02 -5.14934063e-02 -3.94999683e-01 9.54712152e-01
-1.86383948e-01 -5.95341504e-01 -2.72479415e-01 7.21992433e-01
3.31002593e-01 -4.54296887e-01 7.73841679e-01 -2.09738731e-01
-4.98514920e-01 -1.61628097e-01 -9.15655911e-01 4.56606597e-01
9.13321733e-01 3.84939462e-01 -1.50185454e+00 -4.53214228e-01
-1.43351138e-01 -2.33318564e-02 4.13780272e-01 2.36823469e-01
2.08290756e-01 -1.14792168e+00 -8.76696646e-01 -3.27638388e-01
2.35313294e-03 -9.17711854e-01 -8.16788524e-02 8.53633285e-01
-6.26833677e-01 3.64534438e-01 1.99660417e-02 -1.01323833e-03
-1.20262480e+00 4.69246447e-01 -1.99483201e-01 -2.04645619e-01
-6.63031340e-01 4.22828525e-01 -4.74784583e-01 -2.62294412e-01
5.78289777e-02 5.09921908e-01 -1.56078681e-01 4.03301328e-01
8.84209275e-01 8.06980431e-01 -5.37412107e-01 -7.52535582e-01
-5.21057069e-01 2.67420977e-01 1.18477829e-02 7.09115714e-03
1.30752861e+00 1.86251208e-01 -2.41765320e-01 2.32461214e-01
7.86190093e-01 7.26844013e-01 -4.73594397e-01 -1.74363449e-01
6.20211475e-02 -5.97051501e-01 -1.96079716e-01 -9.97234702e-01
-2.52420843e-01 5.63891828e-01 2.94661552e-01 5.99259198e-01
4.30308670e-01 1.68508053e-01 8.99658263e-01 5.31551063e-01
6.35365307e-01 -1.59234095e+00 4.72185552e-01 8.54465783e-01
4.46599334e-01 -1.24238551e+00 2.52809912e-01 -6.85469270e-01
-3.14125270e-01 7.59739459e-01 3.50333720e-01 -1.44590229e-01
4.34209973e-01 5.17217852e-02 7.09412396e-02 -5.81264198e-01
-4.52778429e-01 -4.32437479e-01 -6.75505549e-02 6.80899680e-01
3.76388997e-01 2.99303055e-01 -1.11229622e+00 3.89413625e-01
7.08723068e-02 -1.33416250e-01 5.76425016e-01 1.02448428e+00
-1.12038851e+00 -7.42818117e-01 -4.94016379e-01 6.69202864e-01
-6.07721686e-01 -3.05222750e-01 -1.01980937e+00 5.92598796e-01
4.81797159e-02 1.58715045e+00 -2.68144906e-01 -2.41356224e-01
-2.02890784e-01 1.97724357e-01 2.69579351e-01 -8.69493723e-01
-8.57626975e-01 -1.46905214e-01 7.98007369e-01 -3.01303416e-01
-5.99021018e-01 -3.06649655e-01 -9.54399943e-01 -8.11567068e-01
-5.40438533e-01 2.12536305e-01 2.87575811e-01 9.87479270e-01
9.20758843e-02 -3.20206404e-01 5.31492770e-01 -3.59954387e-01
-8.02116692e-01 -1.09386790e+00 -5.79443991e-01 6.40415192e-01
3.94789595e-03 -5.78781486e-01 -7.12827563e-01 -1.44653976e-01] | [8.52349853515625, 10.626734733581543] |
1fd7da8d-1328-4aa1-9434-a44632a72d23 | diff-ttsg-denoising-probabilistic-integrated | 2306.09417 | null | https://arxiv.org/abs/2306.09417v2 | https://arxiv.org/pdf/2306.09417v2.pdf | Diff-TTSG: Denoising probabilistic integrated speech and gesture synthesis | With read-aloud speech synthesis achieving high naturalness scores, there is a growing research interest in synthesising spontaneous speech. However, human spontaneous face-to-face conversation has both spoken and non-verbal aspects (here, co-speech gestures). Only recently has research begun to explore the benefits of jointly synthesising these two modalities in a single system. The previous state of the art used non-probabilistic methods, which fail to capture the variability of human speech and motion, and risk producing oversmoothing artefacts and sub-optimal synthesis quality. We present the first diffusion-based probabilistic model, called Diff-TTSG, that jointly learns to synthesise speech and gestures together. Our method can be trained on small datasets from scratch. Furthermore, we describe a set of careful uni- and multi-modal subjective tests for evaluating integrated speech and gesture synthesis systems, and use them to validate our proposed approach. For synthesised examples please see https://shivammehta25.github.io/Diff-TTSG | ['Gustav Eje Henter', 'Éva Székely', 'Jonas Beskow', 'Simon Alexanderson', 'Siyang Wang', 'Shivam Mehta'] | 2023-06-15 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 7.73621127e-02 3.36689144e-01 1.15284912e-01 -4.74521220e-01
-1.13499391e+00 -5.55473983e-01 1.17079222e+00 -9.00707304e-01
-6.65037930e-02 4.82506424e-01 7.89965332e-01 -1.08030364e-02
2.69177437e-01 -1.55402064e-01 -4.40480858e-01 -8.53355527e-01
2.47732669e-01 4.84069884e-01 3.78241509e-01 1.91565938e-02
-1.53111801e-01 3.99739176e-01 -1.44446707e+00 4.34618026e-01
6.01856232e-01 5.51780522e-01 3.39942515e-01 1.27422893e+00
-1.43138453e-01 8.32405508e-01 -8.32452238e-01 -3.85609686e-01
1.74430534e-02 -8.20300996e-01 -7.51776814e-01 2.36727223e-01
3.28825235e-01 -5.99473953e-01 -5.48146904e-01 8.53632927e-01
8.69604766e-01 3.56606960e-01 7.06171274e-01 -1.33306265e+00
-6.78106368e-01 6.80053353e-01 -5.44983856e-02 -1.92486674e-01
7.36112058e-01 7.05673814e-01 7.70105124e-01 -9.27046299e-01
7.54765630e-01 1.74970615e+00 2.46879265e-01 1.11678159e+00
-8.99291635e-01 -5.70279658e-01 1.42363459e-01 -1.57595634e-01
-1.29792583e+00 -1.09729075e+00 5.05687237e-01 -3.35035145e-01
9.13648069e-01 4.91051793e-01 4.04533535e-01 1.93256295e+00
-2.44131058e-01 1.25845456e+00 1.18847883e+00 -3.27802837e-01
2.49387845e-01 1.52051598e-01 -3.82568359e-01 4.68690574e-01
-5.30711055e-01 3.54854733e-01 -9.85776007e-01 1.15900300e-02
6.49680376e-01 -3.60175341e-01 -2.61604071e-01 5.00407852e-02
-1.64176750e+00 3.69314730e-01 -1.85802653e-01 3.69264126e-01
-3.53703916e-01 2.57376581e-01 1.18796833e-01 6.91014305e-02
5.07547140e-01 -8.18605497e-02 -5.20215929e-02 -7.28200257e-01
-1.31045413e+00 3.94121557e-01 9.06183898e-01 1.19110608e+00
7.42268115e-02 9.13830847e-02 -4.58013177e-01 9.32269812e-01
7.03540444e-01 9.55454707e-01 4.26873595e-01 -1.32395017e+00
4.66950744e-01 -1.75580293e-01 1.55949563e-01 -6.06323898e-01
-2.80715436e-01 3.38916093e-01 -6.98404133e-01 1.61121801e-01
6.58626497e-01 -4.61252183e-01 -8.95131886e-01 1.68350339e+00
3.35668743e-01 3.25920641e-01 1.76122487e-01 1.07531941e+00
1.11934960e+00 9.09157097e-01 8.39748513e-03 -2.21772596e-01
1.05849135e+00 -1.39341235e+00 -1.19433761e+00 -8.69012848e-02
2.48377725e-01 -1.06222427e+00 1.28394639e+00 4.58684623e-01
-1.41043389e+00 -3.44079494e-01 -6.38125300e-01 4.81969863e-03
-4.67816703e-02 8.71173516e-02 2.41271377e-01 6.97023690e-01
-1.22429442e+00 5.24396241e-01 -1.16395199e+00 -5.55458426e-01
1.96396485e-01 1.31688908e-01 -3.34513605e-01 2.11735498e-02
-9.99955058e-01 1.00420558e+00 -1.12763345e-01 1.13848478e-01
-8.84034932e-01 -2.29582861e-01 -6.04347825e-01 -3.33119422e-01
3.97354424e-01 -5.69312632e-01 1.82772124e+00 -8.05758059e-01
-2.62027836e+00 4.48639721e-01 -5.77124953e-01 9.15185660e-02
9.51954722e-01 -3.90744656e-01 -5.82596242e-01 1.92187890e-01
-3.32924694e-01 9.81723845e-01 1.04783964e+00 -1.42623305e+00
-4.12247598e-01 -4.82785553e-02 -3.55634719e-01 4.85218257e-01
-1.33587234e-02 3.61806244e-01 -6.37093365e-01 -8.82051229e-01
-1.31602719e-01 -1.14696634e+00 1.05079345e-01 3.81053388e-02
-5.59597254e-01 -3.32907975e-01 7.97577381e-01 -7.95604706e-01
1.04108179e+00 -2.09064651e+00 3.86811852e-01 -1.16553985e-01
-1.26807913e-01 3.13032836e-01 -2.59098887e-01 5.37549555e-01
3.89248878e-01 5.09954318e-02 -3.62310171e-01 -7.60814309e-01
2.90450901e-01 2.41874009e-01 -1.61422014e-01 4.44050938e-01
1.19084440e-01 9.70095158e-01 -1.12279046e+00 -5.18455684e-01
3.34700614e-01 6.84454799e-01 -3.50521415e-01 5.07827163e-01
-3.20197821e-01 8.46973181e-01 -2.41578728e-01 7.61875510e-01
4.87454176e-01 -1.32583767e-01 1.50193453e-01 9.78434738e-03
3.07404958e-02 3.44010830e-01 -1.28894186e+00 2.00509453e+00
-4.10926431e-01 9.01547253e-01 2.27270156e-01 -3.93757343e-01
5.33250391e-01 9.12877798e-01 2.47808129e-01 -4.13237453e-01
9.95021164e-02 2.72357881e-01 -4.56302203e-02 -7.81082809e-01
4.14968699e-01 -2.09347740e-01 9.06877071e-02 6.69051290e-01
2.87997246e-01 -6.58336282e-01 -4.56365719e-02 6.71174154e-02
9.93767619e-01 5.03012836e-01 1.98588252e-01 1.37382463e-01
3.15557480e-01 -4.70819950e-01 9.06321704e-02 7.37753689e-01
-4.87361193e-01 1.03312933e+00 1.09233804e-01 2.50733465e-01
-8.67998719e-01 -1.28154111e+00 2.68764019e-01 1.09418941e+00
4.44337316e-02 -1.48882508e-01 -1.05055249e+00 -6.53191626e-01
-4.04779255e-01 9.28191543e-01 -1.47741243e-01 2.61018008e-01
-4.54644978e-01 -3.71767163e-01 8.78627002e-01 3.99333835e-01
1.82556078e-01 -1.41093159e+00 -2.77693212e-01 9.61399227e-02
-3.75077128e-01 -1.32614338e+00 -8.09485614e-01 -3.64051670e-01
-5.95628560e-01 -3.89201820e-01 -1.40053773e+00 -4.68169749e-01
2.39466175e-01 2.43397080e-03 9.03047323e-01 -1.80509508e-01
1.08749814e-01 5.66064954e-01 -2.89831340e-01 -2.51519561e-01
-9.23448324e-01 -1.29600540e-01 3.26850951e-01 1.25801578e-01
1.72011107e-01 -4.55375701e-01 -4.12603319e-01 5.27948439e-01
-7.34601557e-01 1.77396521e-01 5.50903022e-01 6.51919961e-01
-1.43980039e-02 -5.04082024e-01 4.03689116e-01 -4.20761108e-01
8.59434426e-01 -2.25288495e-01 -1.38747633e-01 1.75322995e-01
-1.65268525e-01 6.66314736e-02 1.05083913e-01 -8.05065989e-01
-1.47742796e+00 1.48136303e-01 -4.46954042e-01 -4.18387353e-01
-6.66488647e-01 6.64768592e-02 -2.69218117e-01 3.59435350e-01
4.92724210e-01 2.81956881e-01 2.49880478e-01 -3.73814136e-01
8.06114197e-01 1.09120107e+00 4.96164501e-01 -4.57027733e-01
6.23165786e-01 4.61826682e-01 -4.69046652e-01 -1.28257179e+00
-1.91478416e-01 -2.40274012e-01 -5.08515894e-01 -4.83900487e-01
8.80728543e-01 -8.10865283e-01 -7.80336678e-01 8.99951875e-01
-1.42377114e+00 -8.07655454e-01 -1.46458626e-01 6.44232988e-01
-8.47661436e-01 4.33402807e-01 -6.64956510e-01 -1.31447935e+00
1.91380996e-02 -1.46234918e+00 1.37636757e+00 8.28827173e-02
-8.06853652e-01 -7.55239367e-01 2.07172766e-01 4.48172808e-01
4.28582400e-01 -9.10278931e-02 3.18219699e-02 -6.43782616e-01
-5.60922503e-01 3.84084024e-02 4.70813084e-03 2.87368059e-01
3.25688928e-01 3.06678295e-01 -1.31639123e+00 8.98248479e-02
-2.88015664e-01 -3.11536103e-01 4.74997938e-01 4.76139784e-01
6.11255944e-01 -4.85669911e-01 -1.35348246e-01 2.09571987e-01
4.25667495e-01 2.28368968e-01 5.25988936e-01 -2.59078801e-01
7.85967290e-01 9.26957548e-01 4.17099237e-01 3.92543226e-01
4.24290627e-01 9.27082121e-01 -6.25382811e-02 1.73576936e-01
-5.64063430e-01 -2.41016492e-01 7.66883373e-01 1.22015953e+00
-1.99642599e-01 -8.13675940e-01 -7.79527962e-01 5.06760240e-01
-1.77642941e+00 -1.10296023e+00 -1.56906305e-03 1.75551558e+00
9.96821105e-01 -1.04488157e-01 3.62373888e-01 -6.88552782e-02
5.56024671e-01 3.74035805e-01 -2.84804225e-01 -2.91839927e-01
-1.41094372e-01 2.03600917e-02 1.00119896e-01 9.67320740e-01
-8.51811111e-01 1.22198927e+00 6.43140221e+00 9.74775136e-01
-1.19711900e+00 1.45818770e-01 3.18897665e-01 -4.46499586e-01
-4.14127767e-01 -3.20666254e-01 -6.13172054e-01 4.50946391e-01
1.27272248e+00 2.64302999e-01 7.84433305e-01 4.18500781e-01
6.55485928e-01 5.44854626e-03 -1.15368342e+00 9.12863672e-01
1.40769377e-01 -1.14666080e+00 -2.21251268e-02 -1.34645104e-01
7.42093921e-01 9.67342686e-03 1.40809759e-01 1.14160717e-01
5.88746846e-01 -1.03059030e+00 1.15446544e+00 7.85413802e-01
7.24570096e-01 -2.21179679e-01 4.75504309e-01 4.99401569e-01
-8.87719572e-01 3.38197827e-01 4.30826366e-01 2.21117541e-01
7.28325903e-01 9.96967778e-02 -8.59531820e-01 3.13193083e-01
5.53198159e-01 4.84292597e-01 -8.67858231e-02 5.19429326e-01
-3.59147400e-01 6.59953117e-01 -5.06454766e-01 -3.98441613e-01
2.54702181e-01 7.76038468e-02 8.71758640e-01 1.47496057e+00
6.08474731e-01 2.59856671e-01 -1.87304944e-01 8.07711482e-01
1.23210125e-01 -1.15896136e-01 -5.49871325e-01 -4.83682752e-01
4.64802951e-01 9.03622627e-01 -5.55134177e-01 -4.70940292e-01
-4.96478707e-01 1.27508700e+00 -1.77235320e-01 6.84891999e-01
-7.32875764e-01 9.59575996e-02 8.10386062e-01 -5.48910648e-02
1.16988190e-01 -4.79589641e-01 -3.78168225e-01 -1.18797278e+00
-1.11027323e-01 -1.22246838e+00 -1.35256276e-01 -9.38785613e-01
-1.15109158e+00 7.94733286e-01 2.40705237e-01 -1.16862845e+00
-8.30187261e-01 -4.36780900e-01 -4.26125437e-01 9.65256810e-01
-1.06401896e+00 -1.16615641e+00 -3.05491179e-01 3.97542626e-01
1.04749942e+00 -2.26124346e-01 7.93096602e-01 1.22070611e-01
-4.77548122e-01 5.61714232e-01 -7.51465037e-02 -1.87105983e-01
9.35138226e-01 -1.03134382e+00 6.38184190e-01 7.81437337e-01
2.39027262e-01 5.11141300e-01 1.01990569e+00 -6.34044707e-01
-1.26051414e+00 -6.46333277e-01 1.01117802e+00 -7.45962203e-01
6.07829690e-01 -5.44288695e-01 -7.39744723e-01 5.22216678e-01
6.71041250e-01 -8.70544165e-02 3.86278689e-01 -3.79576772e-01
2.85674632e-02 3.54878098e-01 -1.05895722e+00 1.05648041e+00
1.22572100e+00 -6.31103814e-01 -5.09888470e-01 1.50632456e-01
7.01852262e-01 -5.36851704e-01 -6.11423254e-01 1.07734539e-01
8.58686745e-01 -1.01140952e+00 8.80812883e-01 -1.61962420e-01
4.05354887e-01 -2.44225666e-01 -2.95785487e-01 -1.43978179e+00
1.83942795e-01 -1.21025074e+00 -4.64310229e-01 1.27194273e+00
4.80144680e-01 -4.16129082e-01 7.61203647e-01 7.72924364e-01
-3.92148532e-02 -3.70973110e-01 -8.65925193e-01 -7.73202300e-01
1.57343131e-02 -6.84784353e-01 5.19479990e-01 6.38065398e-01
9.54614952e-02 1.84574738e-01 -7.86807418e-01 3.10639918e-01
4.39962357e-01 -3.57194960e-01 1.08959997e+00 -7.51043737e-01
-3.83807600e-01 -6.28362119e-01 1.34454235e-01 -1.45156217e+00
1.65648833e-01 -4.80805337e-01 6.71514273e-01 -1.50105631e+00
-2.17150703e-01 2.80574206e-02 4.27646846e-01 2.04917654e-01
-1.02091573e-01 3.31845991e-02 3.56885582e-01 2.67984569e-01
-4.30621177e-01 8.69469166e-01 1.58234322e+00 -5.67717850e-02
-4.27032322e-01 1.92002673e-02 -1.43054277e-01 7.17047215e-01
5.26034474e-01 -4.39069539e-01 -4.34093833e-01 -3.49535733e-01
-4.92384136e-01 4.80741382e-01 2.91915059e-01 -7.46466279e-01
3.13973546e-01 -4.09910738e-01 -1.13628186e-01 -4.62112933e-01
8.21127772e-01 -6.53146207e-01 3.11322778e-01 3.77763994e-02
-5.37337959e-01 -1.82581961e-01 -2.68724952e-02 4.91762042e-01
-2.39716947e-01 2.42380220e-02 6.39590740e-01 -1.47563249e-01
-4.89238083e-01 8.99409205e-02 -8.20561826e-01 -5.33133186e-02
7.38667786e-01 -1.25809535e-01 -1.65722653e-01 -9.26880658e-01
-1.02666199e+00 -8.20740908e-02 2.47934520e-01 7.85778105e-01
6.20859385e-01 -1.30928755e+00 -7.64124036e-01 6.67276978e-02
-3.17680873e-02 -1.38510942e-01 4.14241433e-01 8.22035611e-01
-5.13693988e-01 4.27493691e-01 2.06474885e-01 -7.68565476e-01
-1.19291747e+00 8.63934830e-02 2.63321757e-01 1.02029853e-01
-6.61378086e-01 1.02303922e+00 1.87058467e-02 -6.70752168e-01
4.80277359e-01 -4.65737909e-01 1.56171679e-01 -1.12779327e-01
5.87607920e-01 5.34750104e-01 -2.04994991e-01 -8.37904274e-01
-4.93627936e-01 2.66856045e-01 3.86435419e-01 -1.14406693e+00
1.01269472e+00 -4.28713083e-01 3.58854681e-01 7.99527228e-01
9.50532794e-01 6.38390481e-02 -1.62596703e+00 -1.92769498e-01
-2.29521058e-02 -3.61601740e-01 8.45736638e-03 -8.99077058e-01
-6.96635842e-01 9.03037608e-01 5.19265175e-01 1.50549576e-01
7.83099651e-01 1.06585935e-01 8.55975628e-01 2.54270166e-01
1.86022386e-01 -1.00593114e+00 2.58559018e-01 5.13160408e-01
1.19432914e+00 -1.32082736e+00 -4.58453476e-01 -2.89019078e-01
-1.18713808e+00 9.85494912e-01 2.69151330e-01 1.66749239e-01
5.75398564e-01 6.45119965e-01 5.18683434e-01 1.66089475e-01
-7.92877793e-01 -1.81491882e-01 5.34117639e-01 7.34868050e-01
6.35341644e-01 2.25132018e-01 1.75877690e-01 4.17246193e-01
-3.06518763e-01 1.90716371e-01 2.83049971e-01 9.17131126e-01
-5.23561761e-02 -1.05798471e+00 -4.81172085e-01 -7.65621755e-03
-1.49878114e-01 -7.60338828e-02 -5.65158546e-01 5.94238281e-01
-3.96123171e-01 1.26530409e+00 -1.80772170e-01 -4.63764668e-01
3.30231518e-01 3.55919242e-01 5.49221396e-01 -4.10765767e-01
-3.69027913e-01 4.16729033e-01 3.48251492e-01 -7.89578736e-01
-7.49134541e-01 -9.69982028e-01 -1.14480507e+00 -4.30148780e-01
-1.06014282e-01 -1.76685080e-01 6.54538214e-01 1.20319307e+00
1.49961635e-01 5.09743750e-01 2.91788846e-01 -1.40956962e+00
-5.32983065e-01 -1.31710041e+00 -4.43801016e-01 2.70297974e-01
6.29815578e-01 -5.56260645e-01 -6.16268277e-01 2.99701124e-01] | [5.622539520263672, -0.11584766954183578] |
2f184c2f-0571-49b6-be29-8ab11a0ba6a5 | heterogeneous-graph-contrastive-multi-view | 2210.00248 | null | https://arxiv.org/abs/2210.00248v2 | https://arxiv.org/pdf/2210.00248v2.pdf | Heterogeneous Graph Contrastive Multi-view Learning | Inspired by the success of contrastive learning (CL) in computer vision and natural language processing, graph contrastive learning (GCL) has been developed to learn discriminative node representations on graph datasets. However, the development of GCL on Heterogeneous Information Networks (HINs) is still in the infant stage. For example, it is unclear how to augment the HINs without substantially altering the underlying semantics, and how to design the contrastive objective to fully capture the rich semantics. Moreover, early investigations demonstrate that CL suffers from sampling bias, whereas conventional debiasing techniques are empirically shown to be inadequate for GCL. How to mitigate the sampling bias for heterogeneous GCL is another important problem. To address the aforementioned challenges, we propose a novel Heterogeneous Graph Contrastive Multi-view Learning (HGCML) model. In particular, we use metapaths as the augmentation to generate multiple subgraphs as multi-views, and propose a contrastive objective to maximize the mutual information between any pairs of metapath-induced views. To alleviate the sampling bias, we further propose a positive sampling strategy to explicitly select positives for each node via jointly considering semantic and structural information preserved on each metapath view. Extensive experiments demonstrate HGCML consistently outperforms state-of-the-art baselines on five real-world benchmark datasets. | ['Shigen Shen', 'Xiao-Zhi Gao', 'Xiaolong Han', 'Donghua Yu', 'Qi Li', 'Zehong Wang'] | 2022-10-01 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [ 4.25208807e-01 3.70914519e-01 -5.15882254e-01 -3.18301320e-01
-3.85861099e-01 -4.44936723e-01 6.54646456e-01 4.33385409e-02
1.73780024e-01 4.91729051e-01 3.24458450e-01 -2.41299979e-02
-6.53138906e-02 -1.04695165e+00 -6.23671651e-01 -6.02254272e-01
2.90275291e-02 2.61445373e-01 3.09266150e-01 -1.29823193e-01
1.67332813e-01 3.63415897e-01 -1.19348478e+00 3.89304698e-01
8.94044936e-01 4.65142757e-01 3.32898974e-01 2.58953393e-01
-2.35623896e-01 7.25855887e-01 -3.35005850e-01 -4.45434302e-01
1.78941399e-01 -4.84485954e-01 -7.54263341e-01 2.19904825e-01
7.65166938e-01 -1.30265296e-01 -4.98224467e-01 1.26976049e+00
3.54123741e-01 -4.36584465e-02 6.34719014e-01 -1.65263164e+00
-7.51782894e-01 8.09274435e-01 -1.00005555e+00 1.41129866e-01
1.81468368e-01 8.02570358e-02 1.49504673e+00 -9.41535830e-01
1.00994110e+00 1.45860112e+00 4.38549817e-01 3.62604558e-01
-1.40480316e+00 -6.07505620e-01 7.19991207e-01 2.45216310e-01
-1.29141080e+00 -8.70373845e-02 1.22653067e+00 -2.18720779e-01
4.82956260e-01 2.72148877e-01 6.86973691e-01 1.33496559e+00
2.41067290e-01 1.04923141e+00 1.05273950e+00 -1.86909512e-01
1.10150553e-01 1.41625032e-01 1.98437452e-01 1.06051576e+00
5.98038316e-01 -7.86255151e-02 -5.15246212e-01 -2.00760826e-01
5.95302284e-01 1.38651028e-01 -2.70259142e-01 -8.86861324e-01
-1.08428156e+00 9.69796836e-01 7.93364227e-01 1.94587365e-01
-4.01101232e-01 -1.17826842e-01 4.88016158e-01 2.73174047e-01
5.78551114e-01 4.57401037e-01 -2.03434393e-01 5.78122556e-01
-6.66280746e-01 5.25222532e-02 6.34510994e-01 1.05820954e+00
1.01069105e+00 1.94186214e-02 -2.74386495e-01 7.50629842e-01
2.92790264e-01 2.45039016e-01 1.72286369e-02 -7.05401361e-01
5.84862769e-01 1.09886479e+00 -6.12035632e-01 -1.52537131e+00
-3.06801200e-01 -7.05527961e-01 -1.30337954e+00 -6.80672154e-02
-1.31413769e-02 1.68512061e-01 -9.52296734e-01 1.93023217e+00
4.34702516e-01 1.91367358e-01 -2.06027508e-01 7.71049917e-01
1.05207336e+00 6.11490965e-01 7.17297047e-02 -8.22646916e-02
9.99031961e-01 -1.15154600e+00 -4.55888748e-01 -4.94186968e-01
5.78794420e-01 -3.36780816e-01 1.11493480e+00 1.29997477e-01
-8.06210935e-01 -3.37339193e-01 -1.15886343e+00 1.78732067e-01
-2.60794938e-01 -9.53820869e-02 6.36618078e-01 3.84510368e-01
-1.08798552e+00 4.03933197e-01 -5.60055256e-01 -3.77908438e-01
4.49019790e-01 3.19072008e-02 -5.65126717e-01 -4.68601674e-01
-1.21055090e+00 6.19912446e-01 4.62943077e-01 7.68876970e-02
-9.97981668e-01 -7.52646148e-01 -1.01440609e+00 1.70622140e-01
8.40407908e-01 -9.11342978e-01 4.19395417e-01 -1.09828162e+00
-8.79110456e-01 8.13724935e-01 -1.01567708e-01 -1.12027124e-01
3.52387309e-01 2.03870833e-01 -3.65245372e-01 3.44200224e-01
3.80507380e-01 6.87885463e-01 1.00235593e+00 -1.82601810e+00
-2.52808243e-01 -5.18617809e-01 2.70850301e-01 3.61737877e-01
-5.52474439e-01 -5.06360173e-01 -8.14044237e-01 -8.30400825e-01
3.49388957e-01 -1.05277491e+00 -4.18518037e-01 2.41352571e-03
-8.39144826e-01 -1.84961557e-01 9.24774408e-01 -6.28619611e-01
1.25354421e+00 -2.03662181e+00 5.47548115e-01 4.73770320e-01
8.38674426e-01 9.64939669e-02 -5.75971127e-01 4.38977301e-01
-6.92101270e-02 2.40758330e-01 -1.94579095e-01 -1.82394609e-01
-2.47405887e-01 1.48404106e-01 4.00352338e-03 4.13565606e-01
1.51335403e-01 8.60896826e-01 -1.21954548e+00 -6.22501016e-01
1.13558687e-01 3.71234208e-01 -6.16780996e-01 1.96202055e-01
-1.28648922e-01 2.44356096e-01 -4.91434246e-01 6.76141918e-01
7.71346211e-01 -7.59470105e-01 6.59025431e-01 -6.28955185e-01
3.44419301e-01 -1.49770072e-02 -1.05638862e+00 1.51431203e+00
-2.59065747e-01 3.26885998e-01 1.55301705e-01 -1.09293044e+00
7.57055104e-01 -1.06847242e-01 4.57049459e-01 -5.65745354e-01
-1.57937482e-01 -4.21590991e-02 -1.08202405e-01 -3.34672421e-01
4.21093583e-01 6.82253167e-02 9.35701653e-02 4.00853127e-01
3.45031247e-02 1.58867076e-01 9.58716646e-02 8.32466602e-01
1.21011090e+00 -1.93027481e-02 4.34390634e-01 -2.90944159e-01
5.82328498e-01 -1.36613488e-01 7.62462795e-01 6.00560367e-01
-2.73088485e-01 5.75192750e-01 8.38449717e-01 -2.65085876e-01
-6.97930932e-01 -1.15240729e+00 4.41755354e-01 1.04506969e+00
5.77311039e-01 -7.06192434e-01 -6.05608225e-01 -1.24504626e+00
-8.27821121e-02 5.30052304e-01 -6.29796445e-01 -4.99775022e-01
-5.54031253e-01 -6.87486470e-01 2.28852302e-01 3.91992807e-01
6.22236371e-01 -9.01902676e-01 7.85735622e-02 -5.43378368e-02
-2.52787560e-01 -1.05258954e+00 -6.75773919e-01 -2.39379942e-01
-7.08400190e-01 -1.30672944e+00 -5.03980875e-01 -7.87957311e-01
1.03495908e+00 8.84187639e-01 1.45320678e+00 3.26272517e-01
-1.47855192e-01 5.08080840e-01 -3.44740242e-01 1.34056494e-01
-4.33837950e-01 4.94107008e-01 -3.82049739e-01 1.28727317e-01
1.00820102e-01 -6.67119801e-01 -6.77113295e-01 1.89959779e-01
-9.53619957e-01 4.18927968e-01 8.08584452e-01 9.26815629e-01
5.81885457e-01 -5.07900305e-02 6.41367555e-01 -1.51874483e+00
4.72105294e-01 -7.56957471e-01 -2.12789163e-01 5.50225914e-01
-8.95560563e-01 1.00477986e-01 7.15573490e-01 -1.34363562e-01
-8.62336814e-01 -2.33832300e-01 1.88792393e-01 -6.10779226e-01
1.33334473e-01 7.18538463e-01 -4.49211508e-01 -4.76918332e-02
3.32914978e-01 1.78257495e-01 4.33866605e-02 -1.57194808e-01
4.20971274e-01 1.65124804e-01 1.28895581e-01 -4.40249294e-01
8.35584939e-01 5.31572521e-01 1.85411990e-01 -9.72879171e-01
-9.69435453e-01 -4.00312662e-01 -4.93767202e-01 -3.58022720e-01
6.11697435e-01 -9.73031342e-01 -2.26516575e-01 3.06734383e-01
-7.86717117e-01 -7.36443512e-03 1.38476029e-01 1.31574338e-02
-1.43269956e-01 6.66419029e-01 -3.50574732e-01 -2.88966686e-01
-3.80949140e-01 -1.09083915e+00 9.91175234e-01 3.44619118e-02
-9.78145674e-02 -1.17554152e+00 2.12887279e-03 4.66667563e-01
2.93903440e-01 3.16391498e-01 1.26394153e+00 -5.35993099e-01
-8.19496036e-01 1.03295082e-02 -5.30740321e-01 1.54368013e-01
2.99420476e-01 -4.69026379e-02 -6.90989852e-01 -7.58886397e-01
-3.89687628e-01 -2.94924080e-01 1.04786241e+00 2.68868595e-01
1.27788031e+00 -2.89557636e-01 -6.09406888e-01 6.23769641e-01
1.51061571e+00 -3.02407354e-01 4.52822298e-01 1.83044657e-01
1.29011345e+00 8.33418310e-01 3.59714389e-01 1.45471826e-01
7.18360543e-01 4.15429860e-01 8.02470982e-01 -1.51155055e-01
-3.78579766e-01 -6.98490262e-01 2.61365831e-01 8.71627629e-01
1.86280325e-01 -4.15792733e-01 -6.87633216e-01 5.42225301e-01
-1.75079739e+00 -8.85362387e-01 -8.41883644e-02 2.02697206e+00
4.14377421e-01 2.31380463e-01 4.09359969e-02 -3.24602395e-01
1.06080163e+00 8.31961572e-01 -7.90871322e-01 1.80924937e-01
-1.98538899e-01 -3.69767725e-01 3.44411820e-01 3.56872916e-01
-1.11363673e+00 9.89901543e-01 5.51440239e+00 6.98384941e-01
-9.77969527e-01 1.44131156e-02 7.31008768e-01 1.45046905e-01
-1.04137909e+00 1.93659112e-01 -6.37509882e-01 3.80251229e-01
3.33540529e-01 -1.22629561e-01 3.07988644e-01 8.50641549e-01
-3.90799083e-02 2.59222150e-01 -9.45829034e-01 8.77631426e-01
2.93701053e-01 -1.47640908e+00 5.60773194e-01 9.94381011e-02
9.55105066e-01 8.71613398e-02 -2.52953991e-02 4.54365224e-01
2.31850535e-01 -7.38247037e-01 3.47842634e-01 1.97560623e-01
5.63826740e-01 -6.25604033e-01 3.77385527e-01 2.94866472e-01
-1.57403731e+00 9.12028402e-02 -3.55968773e-01 4.69273031e-01
-9.40469876e-02 6.33644998e-01 -8.14306676e-01 9.04366434e-01
4.73197520e-01 1.15854156e+00 -9.84211028e-01 5.97162545e-01
-2.23975375e-01 5.23869991e-01 6.01856001e-02 1.04201838e-01
3.57165068e-01 -3.68533194e-01 8.23619902e-01 9.12943363e-01
-2.63479091e-02 -3.63514036e-01 5.56474686e-01 9.61652100e-01
-3.78362447e-01 1.00604586e-01 -1.00141275e+00 -1.01623468e-01
5.86166024e-01 1.54152596e+00 -7.86890030e-01 -2.97220469e-01
-5.89355409e-01 9.92847502e-01 7.72616088e-01 4.61371124e-01
-6.17757797e-01 1.76935922e-02 4.24412906e-01 2.20878810e-01
1.36103615e-01 8.02729949e-02 -2.32857987e-01 -1.37895489e+00
1.02331214e-01 -9.85185504e-01 5.84680974e-01 -4.26138014e-01
-1.55283761e+00 4.87695307e-01 8.29987004e-02 -1.23390329e+00
9.87494439e-02 -1.61858678e-01 -7.08096802e-01 5.23433864e-01
-1.66810751e+00 -1.55907166e+00 -4.58247066e-01 5.05660117e-01
6.47561014e-01 -1.98421493e-01 3.93969774e-01 2.09089831e-01
-6.15749478e-01 6.06310010e-01 -2.65186250e-01 3.40309516e-02
6.10644639e-01 -1.27595723e+00 4.79988843e-01 9.75603223e-01
1.19511895e-01 6.82604730e-01 5.39364219e-01 -9.10930693e-01
-1.66084611e+00 -1.39981163e+00 5.02257586e-01 -7.75357112e-02
5.37135422e-01 -3.50856811e-01 -1.14036155e+00 7.59114861e-01
2.57277578e-01 1.58935264e-01 4.20263112e-01 1.98877707e-01
-7.01965809e-01 -1.70770094e-01 -9.94154453e-01 8.26243162e-01
1.39244962e+00 -5.60904086e-01 -3.24042618e-01 2.73764223e-01
7.69757450e-01 4.67827842e-02 -6.57293141e-01 6.76473141e-01
3.74765635e-01 -1.13113415e+00 9.90440726e-01 -4.83729631e-01
5.04672229e-01 -2.14992613e-01 -7.16548041e-02 -1.57054019e+00
-5.98270774e-01 -3.40812832e-01 -1.73853695e-01 1.34348929e+00
3.55065018e-01 -6.87733829e-01 1.05593884e+00 9.85560641e-02
-1.38030723e-01 -8.24546039e-01 -5.40596426e-01 -6.08883798e-01
-1.97577670e-01 1.35422006e-01 4.11515594e-01 1.20218587e+00
-3.18922639e-01 6.53944492e-01 -5.25226355e-01 3.87913108e-01
9.82649326e-01 4.73721415e-01 8.50957811e-01 -1.29973853e+00
-3.05844992e-01 -3.07194710e-01 -3.19277376e-01 -9.32202160e-01
4.60009396e-01 -1.19984901e+00 -2.54069030e-01 -1.72115421e+00
7.00796366e-01 -3.91296893e-01 -3.16304088e-01 2.70913601e-01
-5.95261276e-01 1.87098666e-03 2.54077435e-01 1.48130029e-01
-6.92837119e-01 7.35821187e-01 1.56220770e+00 -5.04676461e-01
6.70505920e-03 -2.05751076e-01 -1.01226461e+00 7.07764685e-01
6.37726605e-01 -2.42454842e-01 -9.68462884e-01 -2.65720218e-01
4.99252379e-01 -5.86199127e-02 3.82372350e-01 -5.00343382e-01
1.81860089e-01 -2.19995916e-01 1.08960338e-01 -6.99093938e-01
1.07642330e-01 -7.26862252e-01 1.18110858e-01 4.58054870e-01
-3.59751552e-01 5.65976575e-02 -2.80365556e-01 1.09194863e+00
-2.64133722e-01 1.22812353e-01 7.29587734e-01 -3.42005253e-01
-1.01634407e+00 6.33375049e-01 -2.57944763e-02 2.50027299e-01
1.03925228e+00 -1.17762852e-02 -5.59196711e-01 -3.01756591e-01
-4.51375753e-01 5.14857829e-01 6.15029156e-01 6.15482748e-01
7.72162914e-01 -1.40115011e+00 -6.77271307e-01 3.80967349e-01
3.88057828e-01 -1.05076106e-02 5.14548600e-01 6.49135590e-01
-2.64314651e-01 -3.49503309e-02 -5.08861579e-02 -5.99337637e-01
-1.35967183e+00 6.32497311e-01 1.10827357e-01 -6.34740114e-01
-7.36668408e-01 7.71891415e-01 5.67281604e-01 -6.19321883e-01
7.29695633e-02 3.84363346e-02 -3.06566566e-01 1.19501084e-01
1.02365173e-01 4.39172894e-01 -1.26533940e-01 -4.44504052e-01
-3.15917760e-01 3.46095204e-01 -4.48497385e-01 3.26003760e-01
1.29062474e+00 -2.98090696e-01 -3.04595500e-01 2.88667917e-01
1.30755854e+00 -2.02824902e-02 -1.17312729e+00 -3.68482590e-01
1.33684814e-01 -4.81839597e-01 7.49206841e-02 -4.18319255e-01
-1.45259500e+00 6.48300231e-01 1.46340936e-01 3.83037269e-01
1.08637059e+00 6.68361410e-02 7.47393906e-01 2.03416616e-01
3.75837415e-01 -8.33778143e-01 4.39142883e-01 2.27503702e-01
9.03301537e-01 -1.47995901e+00 2.82865465e-01 -9.36263204e-01
-6.99421346e-01 8.45533371e-01 1.01451910e+00 -1.95881099e-01
7.04145670e-01 1.23979114e-02 -2.73052722e-01 -6.72971904e-01
-8.29560339e-01 8.41531437e-03 3.03361416e-01 4.12719935e-01
2.49412283e-01 7.67303482e-02 -1.26590326e-01 1.05995871e-01
2.63683200e-01 -4.84515160e-01 4.01620179e-01 8.78646553e-01
-1.86403185e-01 -1.07825959e+00 -1.55838551e-02 6.51948512e-01
-1.29779547e-01 -2.40633972e-02 -7.73152232e-01 7.29225695e-01
-1.92477852e-01 6.83501482e-01 -2.56903797e-01 -6.39755905e-01
1.19323790e-01 -2.61798292e-01 3.58506531e-01 -8.18953097e-01
-4.93256181e-01 3.95150185e-02 9.89550725e-02 -5.89411557e-01
-4.69640076e-01 -4.58190292e-01 -8.83969545e-01 -2.65031993e-01
-2.94433087e-01 -1.61201358e-01 7.23235011e-02 9.18249309e-01
5.85203826e-01 6.82868898e-01 9.13778305e-01 -7.11776793e-01
-5.46232343e-01 -7.36710668e-01 -6.38360500e-01 5.04263222e-01
1.93248153e-01 -6.69938326e-01 -5.55672407e-01 -3.95533502e-01] | [7.387381553649902, 6.157910346984863] |
b69505cf-254b-4262-81e9-26de64268d71 | informed-down-sampled-lexicase-selection | 2301.01488 | null | https://arxiv.org/abs/2301.01488v1 | https://arxiv.org/pdf/2301.01488v1.pdf | Informed Down-Sampled Lexicase Selection: Identifying productive training cases for efficient problem solving | Genetic Programming (GP) often uses large training sets and requires all individuals to be evaluated on all training cases during selection. Random down-sampled lexicase selection evaluates individuals on only a random subset of the training cases allowing for more individuals to be explored with the same amount of program executions. However, creating a down-sample randomly might exclude important cases from the current down-sample for a number of generations, while cases that measure the same behavior (synonymous cases) may be overused despite their redundancy. In this work, we introduce Informed Down-Sampled Lexicase Selection. This method leverages population statistics to build down-samples that contain more distinct and therefore informative training cases. Through an empirical investigation across two different GP systems (PushGP and Grammar-Guided GP), we find that informed down-sampling significantly outperforms random down-sampling on a set of contemporary program synthesis benchmark problems. Through an analysis of the created down-samples, we find that important training cases are included in the down-sample consistently across independent evolutionary runs and systems. We hypothesize that this improvement can be attributed to the ability of Informed Down-Sampled Lexicase Selection to maintain more specialist individuals over the course of evolution, while also benefiting from reduced per-evaluation costs. | ['Lee Spector', 'Charles Ofria', 'Franz Rothlauf', 'Thomas Helmuth', 'Alexander Lalejini', 'Dominik Sobania', 'Martin Briesch', 'Ryan Boldi'] | 2023-01-04 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 5.40973306e-01 1.98117688e-01 -2.19177246e-01 -3.95749271e-01
-7.41791964e-01 -6.44372106e-01 2.93817639e-01 4.10354614e-01
-2.14003459e-01 1.00826311e+00 -5.90090975e-02 -2.16119319e-01
-2.56617248e-01 -1.05912924e+00 -6.72114491e-01 -7.26682961e-01
-1.96847022e-01 7.61264086e-01 3.02636415e-01 -2.55870402e-01
5.63017488e-01 1.45707592e-01 -2.26087117e+00 1.80250436e-01
1.32946289e+00 1.63388878e-01 2.57145166e-02 5.77049911e-01
2.74650082e-02 3.74886431e-02 -1.12980163e+00 -2.28520364e-01
4.38377291e-01 -9.20012355e-01 -4.25031692e-01 3.36250588e-02
3.99573147e-01 9.54689682e-02 3.82454008e-01 9.42633629e-01
7.67901778e-01 2.83148140e-01 1.68661803e-01 -1.18010962e+00
-4.37197685e-01 9.79457974e-01 -6.33978248e-01 3.64918970e-02
3.39901924e-01 5.58434129e-01 1.15373254e+00 -4.19131160e-01
7.94123352e-01 1.16956258e+00 7.37253845e-01 4.21026260e-01
-1.56552720e+00 -4.93169487e-01 5.97671643e-02 -4.83779818e-01
-1.33837771e+00 -3.34878176e-01 3.26551884e-01 -6.60099164e-02
1.24453986e+00 5.13551891e-01 1.20925784e+00 5.48562229e-01
3.69647592e-01 4.76398021e-01 8.27761829e-01 -5.63112438e-01
6.85335994e-01 -6.15672320e-02 9.13453102e-02 5.90887189e-01
1.04213977e+00 4.75270301e-01 -5.53075254e-01 -6.78863049e-01
1.77323446e-01 -5.62774181e-01 -4.22815204e-01 -4.69646811e-01
-7.98491657e-01 9.01944816e-01 -1.04474388e-01 1.92978606e-01
-4.97298270e-01 8.32428262e-02 3.08070898e-01 2.53613949e-01
1.79209396e-01 1.46260691e+00 -6.03870273e-01 -7.62048125e-01
-1.05904186e+00 5.68088710e-01 9.64738905e-01 8.81179869e-01
8.18530321e-01 1.87468484e-01 -3.89242291e-01 8.98338914e-01
-8.74984860e-02 3.45681161e-01 7.91283488e-01 -9.22967732e-01
2.46895209e-01 1.13224280e+00 -3.45181882e-01 -7.21020222e-01
-5.30879200e-02 -7.62762368e-01 -7.78316483e-02 2.80126899e-01
1.18314840e-01 -5.48268259e-01 -8.71338129e-01 1.86443281e+00
2.14241758e-01 -5.21544278e-01 5.07420860e-02 2.29886666e-01
2.38019928e-01 6.72361374e-01 -1.64925501e-01 -3.39708179e-01
1.10702157e+00 -8.85544181e-01 -3.00853420e-02 -3.92038912e-01
8.36466372e-01 -2.57985413e-01 1.21179259e+00 4.91532296e-01
-1.09947634e+00 -2.68661797e-01 -1.18587315e+00 5.93991578e-01
-6.77194446e-02 1.23502621e-02 9.12243009e-01 1.18165123e+00
-1.06728375e+00 9.14559662e-01 -6.70950234e-01 -3.90031546e-01
3.82315904e-01 5.71060061e-01 -6.32132515e-02 -1.29188673e-04
-6.87279344e-01 4.95329589e-01 6.77104473e-01 -2.71400779e-01
-4.49743956e-01 -8.74276459e-01 -5.74672639e-01 2.56884545e-01
5.62291622e-01 -7.48757303e-01 8.75962496e-01 -1.15791881e+00
-1.56024730e+00 4.92119282e-01 -1.86820656e-01 -2.87281811e-01
3.83144766e-01 1.15049630e-01 -1.10673442e-01 -2.46861860e-01
-4.49559987e-02 7.01125622e-01 6.56369209e-01 -1.18286395e+00
-7.79670060e-01 -1.79556221e-01 -4.52632308e-02 3.08674783e-01
-2.55569309e-01 -5.67802079e-02 -4.44076210e-01 -5.88033259e-01
-2.96989866e-02 -1.37117136e+00 -1.30479842e-01 -6.79651976e-01
-3.44424665e-01 6.84283003e-02 3.74880433e-01 -3.37278634e-01
1.61747468e+00 -1.89480507e+00 3.47788811e-01 3.83240610e-01
-8.59322250e-02 2.17194632e-01 -3.76659304e-01 4.15168285e-01
-1.14768207e-01 5.34128129e-01 -6.83663428e-01 1.05563521e-01
-1.19532712e-01 1.62138253e-01 5.62871955e-02 1.10638425e-01
3.94200087e-01 6.91858768e-01 -9.57146406e-01 -3.21268201e-01
-3.88830453e-01 4.16863039e-02 -1.26682138e+00 -1.84579298e-01
-5.44618547e-01 1.21661246e-01 -4.40386802e-01 8.33487928e-01
3.35853606e-01 -4.09455188e-02 2.38655522e-01 4.63055551e-01
-6.82811365e-02 1.66507035e-01 -1.06040406e+00 1.32482755e+00
-2.91417062e-01 6.01308882e-01 -4.88404185e-01 -4.08399612e-01
1.06651294e+00 -1.33413196e-01 1.66571006e-01 -5.17169714e-01
1.85073301e-01 3.99310708e-01 8.29045892e-01 -1.56667396e-01
8.49847794e-01 2.07701877e-01 -4.60498184e-02 9.90864277e-01
-3.01614523e-01 -5.97714484e-01 1.02862525e+00 -1.40085265e-01
1.22159600e+00 3.18527579e-01 2.85021007e-01 -3.17961901e-01
2.90665209e-01 3.26197565e-01 1.05107522e+00 1.10057175e+00
-6.13259561e-02 5.37678897e-01 8.66500437e-01 -1.95354745e-01
-1.13735247e+00 -4.72238123e-01 -2.04812989e-01 1.08901596e+00
-2.89241433e-01 -6.00027621e-01 -7.16874599e-01 -7.34769642e-01
1.36379108e-01 1.37614477e+00 -6.47210717e-01 -5.90636849e-01
-7.18847036e-01 -1.35612381e+00 7.93053508e-01 1.71163604e-01
2.02897847e-01 -1.12537229e+00 -1.23400688e+00 1.10156782e-01
3.80124032e-01 -8.25644583e-02 -2.11090878e-01 4.59032863e-01
-9.16428030e-01 -8.76901567e-01 -5.97559869e-01 -5.52773714e-01
9.04032290e-01 -3.18414807e-01 1.52652502e+00 5.45276165e-01
-3.29457581e-01 -1.87260434e-01 -4.25555348e-01 -3.06485653e-01
-8.45135629e-01 3.86152953e-01 -3.24918717e-01 -5.70434928e-01
4.14144665e-01 -5.40125012e-01 1.58803836e-02 3.05462986e-01
-5.76085329e-01 -1.84869871e-01 6.07143819e-01 1.35062802e+00
5.99148631e-01 4.52571720e-01 5.51755786e-01 -8.89911532e-01
9.45892870e-01 -4.85836923e-01 -8.67229581e-01 5.52640975e-01
-8.11723232e-01 6.37135923e-01 4.75564480e-01 -7.19426394e-01
-9.42821026e-01 -4.23466951e-01 2.05719262e-01 2.42663436e-02
2.01796621e-01 5.64207017e-01 -1.30181327e-01 -1.26799196e-01
7.86924601e-01 2.35518157e-01 7.90431127e-02 -9.70242247e-02
1.26706082e-02 2.38516241e-01 -2.16427241e-02 -1.04681969e+00
4.24321324e-01 -1.60696939e-01 -9.12043825e-02 -6.99815154e-01
-4.43615578e-02 3.57043236e-01 -5.25711812e-02 8.02506506e-02
2.75859475e-01 -3.50811958e-01 -3.80992442e-01 2.04156220e-01
-5.18303454e-01 -4.81392235e-01 -8.43734264e-01 2.26739764e-01
-2.42086709e-01 5.28230108e-02 -9.80189890e-02 -7.48865604e-01
-3.07398945e-01 -1.55638969e+00 9.55676496e-01 5.37165225e-01
-8.14140558e-01 -6.43892467e-01 2.03978866e-01 -1.55052498e-01
2.82451302e-01 3.61830801e-01 1.32529569e+00 -9.80732203e-01
-3.46540481e-01 -3.33278388e-01 5.55178225e-01 6.89989999e-02
1.93411857e-01 7.26960540e-01 -4.68533576e-01 -6.08692825e-01
-2.18727410e-01 -1.96037874e-01 5.87840617e-01 3.86924058e-01
8.36690426e-01 2.70141326e-02 -5.63756943e-01 8.46406400e-01
1.17085731e+00 5.41334689e-01 5.52031994e-01 6.20427966e-01
2.30354190e-01 7.37107575e-01 7.81773210e-01 3.92573297e-01
-1.87011629e-01 5.23960054e-01 -2.01295868e-01 3.03756952e-01
1.59429789e-01 -2.57211447e-01 3.25284481e-01 2.41640925e-01
6.87644780e-02 -5.73704839e-01 -1.09873259e+00 5.65293074e-01
-1.41236150e+00 -8.23223889e-01 2.78431624e-01 2.42378640e+00
1.07767618e+00 3.41276675e-01 3.13473701e-01 1.37361154e-01
1.03740001e+00 -2.58210689e-01 -8.72228146e-01 -6.73332453e-01
-3.46400946e-01 5.44612229e-01 5.92374504e-01 2.45670646e-01
-4.70365256e-01 7.44146466e-01 6.88897896e+00 7.00130403e-01
-1.10036910e+00 -5.38696051e-01 8.08212638e-01 -6.14946425e-01
-6.11259878e-01 3.30008090e-01 -8.79134476e-01 6.89032137e-01
9.79890883e-01 -7.56968319e-01 3.91729742e-01 8.50532174e-01
-2.14082390e-01 -5.61499178e-01 -1.10175848e+00 3.14781517e-01
-6.18138574e-02 -1.19300127e+00 8.39414522e-02 3.44578832e-01
1.27516150e+00 -2.94106871e-01 9.71969888e-02 3.00712913e-01
5.71535945e-01 -9.03255343e-01 7.54948318e-01 3.91615033e-02
4.59846765e-01 -1.15071750e+00 6.80388451e-01 1.37627661e-01
-6.41297817e-01 -3.97728533e-01 -2.34407470e-01 5.27097993e-02
1.03990316e-01 5.37683964e-01 -8.35621595e-01 5.41527510e-01
5.09336770e-01 2.13671401e-01 -1.05363309e+00 1.13319516e+00
-2.59226799e-01 7.36995935e-01 -4.32999730e-01 -4.69708085e-01
1.53975943e-02 -2.67363161e-01 9.72913086e-01 6.90529644e-01
7.13873148e-01 -2.95050651e-01 -2.05040455e-01 1.10408366e+00
2.49347091e-01 -1.30304679e-01 -3.09469253e-01 -5.80725193e-01
9.65506256e-01 7.40120173e-01 -9.15876150e-01 -3.05923104e-01
1.32115826e-01 5.93102753e-01 -1.31914662e-02 8.66563320e-02
-7.09857643e-01 -7.80933738e-01 5.90817869e-01 -1.03180796e-01
4.14333135e-01 1.78781554e-01 -4.11686033e-01 -6.41958833e-01
-1.19504660e-01 -1.58181310e+00 3.00178081e-01 -3.64814967e-01
-9.13536906e-01 7.71278322e-01 1.26838520e-01 -8.18449974e-01
-5.96422136e-01 -1.05587415e-01 -6.88017070e-01 1.05636060e+00
-8.34130526e-01 -3.42651397e-01 -1.03283282e-02 -3.51464212e-01
5.53390205e-01 -3.40184063e-01 6.63898885e-01 -5.10565676e-02
-9.10258710e-01 8.56431663e-01 1.37265086e-01 -5.89846015e-01
4.84128296e-01 -1.19255006e+00 7.43211031e-01 9.17808235e-01
-3.50270957e-01 1.19783914e+00 7.90927470e-01 -1.21476102e+00
-1.37657452e+00 -7.04854965e-01 5.33436179e-01 -2.83425242e-01
2.01049075e-01 -6.22791648e-02 -9.15879428e-01 4.58982676e-01
4.18647267e-02 -4.45415735e-01 7.14581549e-01 1.00903660e-01
1.43005311e-01 1.91873014e-01 -1.45424473e+00 8.35146129e-01
1.07419097e+00 1.46966234e-01 -5.09148180e-01 -1.22997098e-01
6.15100205e-01 -3.49758744e-01 -7.63166904e-01 2.54272193e-01
5.00358105e-01 -1.05829442e+00 6.10212564e-01 -3.63361567e-01
4.90551054e-01 -3.53920251e-01 9.48123261e-02 -1.66095412e+00
-2.80783415e-01 -9.26624179e-01 2.06459165e-01 1.32679188e+00
7.60748982e-01 -1.18131232e+00 9.09436345e-01 6.18538201e-01
-1.63916111e-01 -9.92424846e-01 -4.27850395e-01 -9.77827311e-01
2.72571385e-01 1.89306408e-01 1.21106005e+00 6.88066721e-01
-1.15975432e-01 -2.78902985e-02 1.72617614e-01 -1.14236645e-01
4.62118417e-01 2.03172415e-01 7.25891829e-01 -1.13535249e+00
-9.07745540e-01 -6.93594873e-01 -1.84463322e-01 -4.31923300e-01
-8.55562016e-02 -6.60824418e-01 2.02849016e-01 -1.02788925e+00
3.03133160e-01 -7.12535620e-01 3.77588391e-01 4.96043175e-01
-6.25165522e-01 -4.11075875e-02 4.76934239e-02 -1.09345606e-02
4.15506363e-02 3.41729850e-01 9.71052468e-01 1.59591913e-01
-8.07720423e-01 -2.23778501e-01 -9.26969826e-01 4.71827507e-01
7.32098699e-01 -5.54698229e-01 -5.34320712e-01 -2.79093206e-01
4.26467776e-01 -2.78497040e-01 -3.25824767e-01 -1.03128374e+00
-3.08633149e-01 -2.55613893e-01 1.91374645e-01 -2.70627022e-01
-3.56658064e-02 -2.37731963e-01 6.36046648e-01 8.63076866e-01
-5.87166369e-01 4.78205353e-01 6.03173971e-01 2.56119698e-01
8.15010816e-02 -7.24591732e-01 5.90845048e-01 -1.34493858e-01
-5.27731478e-01 -2.67873049e-01 -4.11578298e-01 2.96224535e-01
1.23406029e+00 -8.28675032e-01 -3.60239565e-01 1.72236741e-01
-1.03732057e-01 2.58975685e-01 1.24427903e+00 1.84798419e-01
1.48905158e-01 -6.84291959e-01 -6.98815465e-01 5.04485607e-01
1.21944353e-01 -4.92118448e-02 -1.62103102e-01 4.51394498e-01
-6.53275907e-01 2.25912690e-01 -1.99965045e-01 -6.47178829e-01
-1.35020030e+00 1.52846530e-01 2.96657592e-01 -1.23423062e-01
-4.51240063e-01 1.13982356e+00 -3.02415282e-01 -5.26131928e-01
-9.05790192e-04 -1.16916135e-01 1.85448945e-01 6.36392981e-02
2.54738897e-01 4.31452245e-01 5.70120960e-02 -2.83099174e-01
-4.02954489e-01 3.53237450e-01 -7.46184438e-02 -2.56265968e-01
1.55016148e+00 3.89437139e-01 -2.81292200e-01 2.76787639e-01
7.99052596e-01 3.67064893e-01 -1.10279524e+00 4.86788571e-01
7.75790662e-02 -8.12520206e-01 -3.14611532e-02 -9.30422366e-01
-1.01125097e+00 2.89891273e-01 2.28990391e-01 1.14466824e-01
1.22134602e+00 -3.94381464e-01 4.02656645e-01 4.31396246e-01
5.81498861e-01 -1.00883245e+00 -1.89406067e-01 3.42791349e-01
6.52823448e-01 -6.15959823e-01 2.71539450e-01 -4.39092398e-01
-6.54381514e-01 9.97392893e-01 8.14809382e-01 3.39385718e-02
-2.01034829e-01 3.20774287e-01 -5.66579163e-01 -2.21715569e-02
-1.07175696e+00 1.67868286e-01 -2.36551780e-02 6.15404487e-01
4.37790364e-01 -5.11694290e-02 -6.94255829e-01 3.67680937e-01
-6.51992023e-01 -1.33905292e-01 5.98222494e-01 1.54994106e+00
-4.12706017e-01 -1.40973067e+00 -4.40650642e-01 7.59378314e-01
-5.35032488e-02 -3.39335740e-01 -7.60902345e-01 8.18184435e-01
3.35413337e-01 7.01560080e-01 2.37356275e-01 -1.71846226e-01
2.06097990e-01 2.19387010e-01 7.39664197e-01 -8.26533675e-01
-1.14975417e+00 -1.94802612e-01 6.10836506e-01 -2.43665412e-01
2.99300432e-01 -1.03003645e+00 -1.21252251e+00 -3.73796016e-01
-7.22434819e-01 4.26570833e-01 4.52104807e-01 3.62165838e-01
7.15814948e-01 9.57731307e-01 2.36033633e-01 -5.78789234e-01
-6.80656672e-01 -5.38686574e-01 -3.39197189e-01 7.85235539e-02
-2.04275519e-01 -6.69830441e-01 -4.41818655e-01 -3.51010025e-01] | [8.008892059326172, 7.175429821014404] |
7e3df778-c5f3-4c31-9269-a7f6b3767a64 | fake-news-or-truth-using-satirical-cues-to | null | null | https://aclanthology.org/W16-0802 | https://aclanthology.org/W16-0802.pdf | Fake News or Truth? Using Satirical Cues to Detect Potentially Misleading News | null | ['Sarah Cornwell', 'Victoria Rubin', 'Niall Conroy', 'Yimin Chen'] | 2016-06-01 | null | null | null | ws-2016-6 | ['deception-detection'] | ['miscellaneous'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.340583801269531, 3.738247871398926] |
b8348b72-399c-40ec-907e-fc573c25d5e2 | towards-human-centered-explainable-ai-user | 2210.11584 | null | https://arxiv.org/abs/2210.11584v2 | https://arxiv.org/pdf/2210.11584v2.pdf | Towards Human-centered Explainable AI: User Studies for Model Explanations | Explainable AI (XAI) is widely viewed as a sine qua non for ever-expanding AI research. A better understanding of the needs of XAI users, as well as human-centered evaluations of explainable models are both a necessity and a challenge. In this paper, we explore how HCI and AI researchers conduct user studies in XAI applications based on a systematic literature review. After identifying and thoroughly analyzing 85 core papers with human-based XAI evaluations over the past five years, we categorize them along the measured characteristics of explanatory methods, namely trust, understanding, fairness, usability, and human-AI team performance. Our research shows that XAI is spreading more rapidly in certain application domains, such as recommender systems than in others, but that user evaluations are still rather sparse and incorporate hardly any insights from cognitive or social sciences. Based on a comprehensive discussion of best practices, i.e., common models, design choices, and measures in user studies, we propose practical guidelines on designing and conducting user studies for XAI researchers and practitioners. Lastly, this survey also highlights several open research directions, particularly linking psychological science and human-centered XAI. | ['Enkelejda Kasneci', 'Gjergji Kasneci', 'Tina Seidel', 'Vaibhav Unhelkar', 'Peizhu Qian', 'Lisa Fiedler', 'Thai-trang Nguyen', 'Tobias Leemann', 'Yao Rong'] | 2022-10-20 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [-6.16735555e-02 3.78734499e-01 -4.26297843e-01 -5.33783317e-01
5.27572632e-02 -4.29729193e-01 2.61688441e-01 1.62848681e-01
-6.10313341e-02 4.67042625e-01 3.52929413e-01 -5.30924261e-01
-6.30207777e-01 -9.95823517e-02 -3.25837612e-01 -2.63990425e-02
1.98905960e-01 4.40366715e-01 -6.97392225e-01 -2.73371577e-01
8.39475870e-01 -3.45356874e-02 -1.70189083e+00 2.48619437e-01
1.37874198e+00 7.76618421e-01 -2.14302436e-01 4.63492662e-01
5.57030179e-03 7.58673906e-01 -6.78686738e-01 -8.31169784e-01
-2.27172554e-01 -4.12493050e-01 -9.61040020e-01 -3.54371190e-01
2.88427413e-01 -4.30581808e-01 9.19007137e-02 5.45280933e-01
3.97786856e-01 7.64005119e-03 5.34226954e-01 -2.07189512e+00
-1.65418923e+00 9.91419435e-01 3.51664773e-03 -1.71710476e-01
5.73223293e-01 3.46304357e-01 1.10369229e+00 -7.92066932e-01
3.84005606e-01 1.18147016e+00 6.94227457e-01 5.85422039e-01
-1.11108673e+00 -8.08447778e-01 2.35628694e-01 1.79648593e-01
-1.09353733e+00 -5.50241530e-01 4.57219660e-01 -5.39760649e-01
7.79738724e-01 8.79692614e-01 7.09054351e-01 1.28871489e+00
1.87402710e-01 7.71009147e-01 1.13393772e+00 -6.20210469e-01
4.09741998e-01 7.32074976e-01 1.08863056e+00 1.17214888e-01
8.96508515e-01 -3.40399183e-02 -7.88037717e-01 -5.29285967e-01
5.55606544e-01 -3.07383039e-03 -1.94432095e-01 -1.46945164e-01
-1.47775388e+00 8.61990690e-01 2.25034460e-01 4.37020004e-01
-8.33214879e-01 7.07282871e-02 1.34678945e-01 2.79315323e-01
4.53895509e-01 1.20415342e+00 -3.16334546e-01 -6.60609245e-01
-4.75159407e-01 4.68330443e-01 9.56122935e-01 1.18898273e+00
2.14612022e-01 -3.10991742e-02 -3.22805732e-01 6.68886304e-01
6.89671040e-01 5.74087501e-01 2.81927377e-01 -1.11271143e+00
-3.70192587e-01 8.48294258e-01 3.10217142e-01 -9.91020441e-01
-3.37496996e-01 -4.00678515e-01 -4.25723642e-01 4.50422823e-01
3.45979422e-01 -2.03137383e-01 -3.87268484e-01 1.56823421e+00
-2.20929459e-01 -4.64919150e-01 -3.08811337e-01 1.11501157e+00
1.02921009e+00 2.04772800e-01 5.26579082e-01 -2.64867336e-01
1.38220108e+00 -8.37034345e-01 -1.11256516e+00 -4.29870009e-01
3.86630803e-01 -5.38998485e-01 1.62929058e+00 4.10836965e-01
-1.59259140e+00 -3.73378754e-01 -9.27768409e-01 -2.63409436e-01
-2.60188043e-01 1.94000572e-01 9.25348222e-01 1.18835938e+00
-1.08390236e+00 2.44507372e-01 -3.22332621e-01 -8.23811293e-01
2.35537976e-01 1.53109387e-01 1.16080783e-01 1.83788165e-01
-1.06024516e+00 1.17959797e+00 -3.14495355e-01 -7.71813616e-02
-5.49589217e-01 -8.41773808e-01 -2.97879755e-01 3.53102714e-01
3.82466286e-01 -1.01125312e+00 1.45239615e+00 -1.18316948e+00
-1.40454459e+00 2.92278677e-01 7.98644051e-02 -8.40861052e-02
-4.03035618e-02 -5.08191407e-01 -3.32947880e-01 -4.61902499e-01
3.85640562e-02 4.81803060e-01 2.86597848e-01 -1.52067268e+00
-5.55299282e-01 -4.37362224e-01 5.20823598e-01 4.33142304e-01
-5.75647771e-01 3.46672565e-01 5.38184941e-02 -1.85954019e-01
-4.12515104e-01 -9.90945220e-01 -1.45779148e-01 1.24896817e-01
-7.65731111e-02 -3.67315590e-01 3.64924848e-01 -4.05519724e-01
1.89523959e+00 -2.05262947e+00 -8.29339772e-02 2.93961167e-01
9.21507955e-01 1.04901977e-01 1.30728669e-02 7.49990582e-01
1.49165615e-01 7.70149469e-01 2.80613363e-01 6.57873526e-02
4.85888839e-01 -2.91766256e-01 5.11634164e-02 -3.09051257e-02
-4.39839631e-01 1.01088595e+00 -9.68660712e-01 4.22209594e-03
-1.90973412e-02 7.61538208e-01 -3.90371472e-01 2.25263208e-01
1.91312321e-02 5.47177456e-02 -2.94651568e-01 6.23334646e-01
2.33596310e-01 -6.21561885e-01 -1.37424856e-01 -7.37429690e-03
-5.85221887e-01 4.46636498e-01 -4.84399498e-01 1.55099523e+00
-3.71143937e-01 6.99297011e-01 -1.14537060e-01 -2.00345572e-02
7.87542820e-01 5.72476268e-01 2.79755473e-01 -6.62310839e-01
4.04238999e-01 -6.38697818e-02 5.33998966e-01 -3.27808380e-01
3.20028782e-01 2.74815857e-01 9.07678828e-02 1.04463327e+00
-3.51653725e-01 1.85636654e-01 -2.80759454e-01 4.30380613e-01
9.27867711e-01 -8.29412863e-02 4.99753952e-01 -6.16839468e-01
6.86283410e-02 1.52073085e-01 2.66037434e-01 1.11850619e+00
-5.35103083e-01 2.52809137e-01 2.77861506e-01 -4.91758198e-01
-8.69278967e-01 -4.10717398e-01 3.31701934e-02 1.39164472e+00
2.98066467e-01 -9.13555622e-01 -1.16146767e+00 -4.38035458e-01
-7.30996653e-02 1.70718980e+00 -7.62124538e-01 -4.72138673e-01
4.31137919e-01 -1.78764090e-01 2.26557076e-01 3.93261850e-01
3.57585251e-01 -1.33282650e+00 -1.26610625e+00 -1.45819262e-01
-4.05833304e-01 -2.90986627e-01 -2.01793060e-01 -3.85577768e-01
-6.83000207e-01 -6.80362046e-01 -1.99345037e-01 -6.99979886e-02
2.93041825e-01 8.60603452e-01 1.38402152e+00 7.33544409e-01
-1.93215102e-01 9.97583508e-01 -4.79236871e-01 -9.94943142e-01
9.21847969e-02 -1.56369522e-01 1.96661741e-01 -5.39695024e-01
8.88843477e-01 -2.42279321e-01 -7.14527190e-01 5.78803003e-01
-4.36567843e-01 2.37723723e-01 6.56184971e-01 4.17464167e-01
-2.34337121e-01 -3.19726795e-01 4.49706703e-01 -1.09381878e+00
1.34333980e+00 -6.61247373e-01 3.06592643e-01 5.54181099e-01
-1.49751067e+00 -4.84031171e-01 -1.38162166e-01 -2.79107600e-01
-1.29307449e+00 -8.21005821e-01 3.72573137e-01 4.40253913e-01
-2.88901389e-01 8.87003124e-01 -2.32967123e-01 -1.48147285e-01
1.25134015e+00 -7.79474616e-01 3.10680449e-01 -1.06386282e-01
5.53596795e-01 9.28985596e-01 5.74653409e-02 -5.06707132e-01
5.39584517e-01 -1.34160608e-01 -9.81351316e-01 -7.86796749e-01
-5.06732166e-01 1.45988867e-01 -1.83127567e-01 -6.87317789e-01
7.37926245e-01 -3.39587718e-01 -1.45517075e+00 -1.96402758e-01
-9.18331027e-01 -3.20033580e-01 -1.48491398e-01 7.55213499e-01
-5.13718247e-01 -2.09385946e-01 -2.39566058e-01 -1.24420035e+00
-8.16439390e-01 -1.07268345e+00 3.91987354e-01 4.14220065e-01
-1.10525358e+00 -6.88617110e-01 3.44528630e-02 7.20822334e-01
1.08322406e+00 -2.66818553e-01 1.28596485e+00 -7.22249329e-01
-3.35089564e-01 -3.00627381e-01 -2.47181356e-01 -2.58702993e-01
-4.53797691e-02 3.46039355e-01 -8.10023248e-01 6.60020560e-02
2.44763330e-01 -2.77342200e-01 -2.36985222e-01 5.50146461e-01
6.62360728e-01 -5.52680135e-01 -3.24771613e-01 1.95507240e-02
9.16434586e-01 7.79296458e-01 7.44358778e-01 5.73810518e-01
5.30471325e-01 6.62697494e-01 1.13712716e+00 6.08831346e-01
5.70014298e-01 4.09038931e-01 1.13558352e-01 1.50754988e-01
2.84163952e-01 -3.27765271e-02 9.29408371e-02 3.52337003e-01
-6.03921533e-01 -4.80330229e-01 -1.26604652e+00 2.03321576e-01
-1.97622919e+00 -8.82183075e-01 -7.83584639e-02 2.48792076e+00
3.66110206e-01 -5.57278134e-02 2.81680375e-01 -6.95478618e-02
5.28962493e-01 -5.75814962e-01 -5.76999843e-01 -5.78070939e-01
3.53243232e-01 -2.98246771e-01 -3.11416872e-02 4.57911223e-01
-4.66343582e-01 7.75056183e-01 7.36162281e+00 -2.07514420e-01
-7.93661714e-01 1.22033782e-01 8.39912832e-01 -2.85039730e-02
-8.54773819e-01 2.01771587e-01 -8.89699310e-02 2.49365076e-01
1.20284140e+00 -8.66111994e-01 6.25798345e-01 1.08854556e+00
2.34824061e-01 1.34306014e-01 -1.23047602e+00 9.30828154e-01
-7.03399396e-03 -1.00972712e+00 -2.57651269e-01 3.00457686e-01
2.66315073e-01 -6.22544348e-01 5.82041025e-01 2.91956365e-01
1.09365374e-01 -1.15214288e+00 7.53372431e-01 7.22992182e-01
5.32935321e-01 -7.07845449e-01 6.88202381e-01 4.65773754e-02
-1.82757393e-01 -8.45955238e-02 -3.52683693e-01 -7.50631094e-01
-2.15060741e-01 4.69764359e-02 -3.59977514e-01 7.95914456e-02
1.02862418e+00 5.28919339e-01 -4.74274337e-01 1.14867234e+00
5.66848367e-02 9.02613640e-01 1.25521004e-01 -4.11389679e-01
-1.75697893e-01 -4.73642856e-01 2.36244917e-01 6.87010527e-01
1.52460679e-01 6.75961375e-01 -4.46206272e-01 1.11048746e+00
4.72650290e-01 1.90060377e-01 -7.70133138e-01 -5.37268102e-01
9.69759166e-01 1.32227540e+00 -4.67130035e-01 -3.22104990e-01
-7.18307734e-01 7.74329960e-01 8.59663188e-02 5.64050853e-01
-8.60243261e-01 -1.46476984e-01 1.05755496e+00 2.31432483e-01
-6.94492459e-01 -1.66949540e-01 -1.20908177e+00 -8.53168964e-01
-4.20333058e-01 -1.50133705e+00 -3.11509296e-02 -1.17199051e+00
-1.21344292e+00 7.49146461e-01 1.20563030e-01 -9.83644307e-01
-2.00135425e-01 -2.32197642e-01 -5.30947685e-01 5.87627888e-01
-6.87199831e-01 -1.20304322e+00 -8.52926016e-01 7.25798085e-02
2.12811440e-01 5.02304640e-03 1.02364349e+00 5.99353388e-02
-7.46827483e-01 5.31395078e-01 -1.11139461e-01 -8.50275218e-01
1.18038607e+00 -9.91018951e-01 7.05349803e-01 2.99135059e-01
-5.97190782e-02 1.42757869e+00 7.36814439e-01 -8.29257786e-01
-1.75775862e+00 -1.54063821e-01 1.01835489e+00 -9.75127220e-01
2.35477582e-01 -2.89796680e-01 -8.92350614e-01 9.86782670e-01
8.65189552e-01 -1.08422303e+00 1.32885301e+00 7.37477601e-01
6.82116374e-02 1.24313630e-01 -9.48891222e-01 9.13883626e-01
1.35533047e+00 -1.86258808e-01 -1.19361751e-01 3.09996665e-01
5.86763084e-01 4.89858016e-02 -8.33682954e-01 -2.15719827e-03
9.41419065e-01 -1.17321336e+00 8.89565349e-01 -5.73958158e-01
3.85871828e-01 9.33115184e-02 3.04394424e-01 -1.23793519e+00
-1.16209495e+00 -9.40303743e-01 2.70199537e-01 1.40604460e+00
6.12563610e-01 -5.77235222e-01 4.67961431e-01 2.29415989e+00
-2.55496144e-01 -3.91457647e-01 9.45605934e-02 -1.32006228e-01
-3.49378914e-01 -4.49891210e-01 6.46475732e-01 1.03206396e+00
7.47626066e-01 5.89678288e-01 -5.58630824e-01 -1.10346846e-01
4.59061176e-01 -2.70064861e-01 9.84814525e-01 -1.77906179e+00
2.76156664e-01 -5.29132843e-01 -1.71015605e-01 -5.31721234e-01
-2.69729137e-01 -3.30973178e-01 -1.22481637e-01 -1.71494877e+00
4.13374692e-01 -3.67310226e-01 -1.22556053e-01 5.50570846e-01
-3.83157223e-01 -3.35240066e-02 3.23514193e-01 4.74417716e-01
-3.45713019e-01 4.26506042e-01 7.70271301e-01 3.63065451e-01
-3.96481097e-01 -6.68159351e-02 -1.83717191e+00 6.49735808e-01
7.43183374e-01 2.41367370e-02 -1.10015619e+00 -6.03823304e-01
6.17953956e-01 -5.29498532e-02 1.75044075e-01 -9.94573414e-01
3.79083693e-01 -4.92040575e-01 1.98416963e-01 -2.14096233e-01
2.53258944e-01 -1.16827118e+00 5.05828381e-01 -3.01103778e-02
-7.08247185e-01 4.70835745e-01 1.59487620e-01 5.17891236e-02
3.13198537e-01 -3.35184395e-01 2.84756254e-02 2.04056427e-01
-2.95068771e-01 4.49825376e-02 -5.86880445e-01 -3.38782072e-01
1.02258205e+00 -4.24791038e-01 -3.81113440e-01 -1.09658837e+00
-2.30220914e-01 1.95258483e-01 5.86484849e-01 5.18117249e-01
8.66784394e-01 -1.17890131e+00 -4.73994672e-01 -4.47600381e-03
3.69795591e-01 -7.81531692e-01 1.91248104e-01 6.85142994e-01
-1.75749764e-01 5.64038932e-01 -6.52069330e-01 3.17547582e-02
-1.24336827e+00 6.50639832e-01 -3.18393290e-01 7.38740325e-01
-3.96745920e-01 6.55990362e-01 4.40551817e-01 -3.95630999e-03
5.59455752e-01 -1.18353385e-02 -3.55407000e-01 -2.90650964e-01
7.42302954e-01 6.48288965e-01 -5.33149242e-01 -2.28090450e-01
-3.99752259e-01 3.76576092e-03 -3.09721857e-01 -4.35746908e-01
1.08550870e+00 -5.32570899e-01 -2.79255241e-01 8.18245828e-01
4.20164078e-01 -2.48611480e-01 -5.01594543e-01 1.71919927e-01
5.38980104e-02 -9.07024086e-01 5.43552116e-02 -1.56387377e+00
-5.27531087e-01 6.85745358e-01 4.28776979e-01 3.59413862e-01
8.11914325e-01 -1.12128191e-01 1.99169412e-01 3.41153085e-01
8.78633708e-02 -9.29780841e-01 -1.04004920e-01 -4.12633494e-02
1.30371344e+00 -1.01089299e+00 7.09677860e-02 -4.01447386e-01
-1.59988844e+00 9.28083658e-01 1.10518360e+00 5.33683479e-01
8.13504338e-01 -2.60602355e-01 3.38554233e-01 -6.30593777e-01
-1.03702199e+00 2.04820395e-01 4.05244976e-01 6.51397169e-01
1.39291537e+00 2.04440460e-01 -8.99426758e-01 1.15405405e+00
-3.37148190e-01 2.29386598e-01 5.55464923e-01 7.04276502e-01
-2.56151199e-01 -9.31135297e-01 -1.95430636e-01 4.45281863e-01
2.57098600e-02 -5.21827862e-02 -1.16983211e+00 1.00072122e+00
-4.59105611e-01 1.56502378e+00 -2.24757075e-01 -7.40348935e-01
3.54881257e-01 1.42374068e-01 -8.88872743e-02 -3.30543876e-01
-1.05898452e+00 -3.82108837e-01 3.09258372e-01 -7.72687256e-01
-8.89646336e-02 -6.27282023e-01 -9.86446440e-01 -8.91799510e-01
-6.41966701e-01 3.31305265e-01 1.10394478e+00 5.77089190e-01
7.89792061e-01 2.10607529e-01 2.53169507e-01 -2.55592316e-01
-1.37352914e-01 -9.79784250e-01 -1.88144654e-01 3.69522870e-01
-2.20035121e-01 -6.73124075e-01 -2.33133212e-01 -3.27200204e-01] | [9.040423393249512, 6.188049793243408] |
7acba4e6-39fa-42f7-ae8b-d13f4da192fd | xsemplr-cross-lingual-semantic-parsing-in | 2306.04085 | null | https://arxiv.org/abs/2306.04085v1 | https://arxiv.org/pdf/2306.04085v1.pdf | XSemPLR: Cross-Lingual Semantic Parsing in Multiple Natural Languages and Meaning Representations | Cross-Lingual Semantic Parsing (CLSP) aims to translate queries in multiple natural languages (NLs) into meaning representations (MRs) such as SQL, lambda calculus, and logic forms. However, existing CLSP models are separately proposed and evaluated on datasets of limited tasks and applications, impeding a comprehensive and unified evaluation of CLSP on a diverse range of NLs and MRs. To this end, we present XSemPLR, a unified benchmark for cross-lingual semantic parsing featured with 22 natural languages and 8 meaning representations by examining and selecting 9 existing datasets to cover 5 tasks and 164 domains. We use XSemPLR to conduct a comprehensive benchmark study on a wide range of multilingual language models including encoder-based models (mBERT, XLM-R), encoder-decoder models (mBART, mT5), and decoder-based models (Codex, BLOOM). We design 6 experiment settings covering various lingual combinations (monolingual, multilingual, cross-lingual) and numbers of learning samples (full dataset, few-shot, and zero-shot). Our experiments show that encoder-decoder models (mT5) achieve the highest performance compared with other popular models, and multilingual training can further improve the average performance. Notably, multilingual large language models (e.g., BLOOM) are still inadequate to perform CLSP tasks. We also find that the performance gap between monolingual training and cross-lingual transfer learning is still significant for multilingual models, though it can be mitigated by cross-lingual few-shot training. Our dataset and code are available at https://github.com/psunlpgroup/XSemPLR. | ['Rui Zhang', 'Zhiguo Wang', 'Jun Wang', 'Yusen Zhang'] | 2023-06-07 | null | null | null | null | ['semantic-parsing', 'cross-lingual-transfer', 'xlm-r'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.33761808e-01 -1.74445771e-02 -5.55241108e-01 -5.49374938e-01
-1.45512247e+00 -8.46914411e-01 4.41188604e-01 -1.01209998e-01
-2.79326051e-01 7.41818964e-01 3.68416846e-01 -7.47997344e-01
2.59191662e-01 -8.16399455e-01 -1.01204455e+00 -8.66964608e-02
4.03923005e-01 3.68824065e-01 2.68193394e-01 -4.64870602e-01
-1.72414511e-01 -1.63460582e-01 -1.17095423e+00 6.54842436e-01
1.14413917e+00 8.24381113e-01 4.72823411e-01 3.82430851e-01
-8.04147959e-01 9.75621700e-01 -3.31481904e-01 -8.47709537e-01
1.22983456e-02 -3.25028658e-01 -1.03593767e+00 -6.88648403e-01
2.24999756e-01 6.62214085e-02 7.60784447e-02 1.01869619e+00
4.82196569e-01 -2.20590144e-01 2.16067195e-01 -1.18407619e+00
-1.24435055e+00 9.48792100e-01 -3.74009490e-01 -1.62451088e-01
6.69884861e-01 1.44663438e-01 1.28880513e+00 -9.26612258e-01
7.11574435e-01 1.76867914e+00 7.81514108e-01 4.73532379e-01
-1.04662192e+00 -8.17443490e-01 1.16311938e-01 9.57053900e-02
-1.18921793e+00 -4.31506813e-01 3.52147788e-01 -8.39009434e-02
1.32790565e+00 3.46486494e-02 1.35006130e-01 1.36259830e+00
2.17317730e-01 1.02549124e+00 1.19644403e+00 -5.18256068e-01
3.80446613e-02 1.87237531e-01 3.14290106e-01 7.45435297e-01
1.89799234e-01 -1.71119273e-01 -6.26816630e-01 -1.27178714e-01
4.35951114e-01 -3.31079006e-01 4.24262546e-02 -7.40688890e-02
-1.17345464e+00 9.71636951e-01 2.46859804e-01 3.01263481e-01
7.63048902e-02 -6.38910715e-05 6.37231350e-01 7.32961178e-01
3.95781070e-01 3.96619350e-01 -1.19270372e+00 -2.22054422e-01
-3.33042979e-01 7.27889538e-02 9.95119810e-01 1.53752673e+00
1.07149005e+00 -1.87335238e-02 -1.10598758e-01 1.26407623e+00
1.33894965e-01 7.18792081e-01 6.88450933e-01 -8.70073497e-01
8.50194633e-01 6.76630974e-01 -1.91830620e-01 -2.97332466e-01
-1.36529714e-01 4.14596573e-02 -4.22704995e-01 -5.70818126e-01
3.75448734e-01 -4.74005312e-01 -5.82588136e-01 1.95600688e+00
1.64430980e-02 8.17427635e-02 7.96859384e-01 5.00346243e-01
1.18367779e+00 7.91994452e-01 4.25758660e-01 1.48651883e-01
1.79348743e+00 -1.05638468e+00 -5.40027082e-01 -6.60015523e-01
1.19031513e+00 -9.08061504e-01 1.86134100e+00 5.75216562e-02
-9.18181658e-01 -5.40686488e-01 -8.44969749e-01 -5.68586469e-01
-7.34788299e-01 1.44688770e-01 8.41020584e-01 4.66876775e-01
-9.13979650e-01 5.76311536e-02 -7.82721639e-01 -6.66379750e-01
1.00135021e-01 -1.88444346e-01 -2.80412793e-01 -4.94197488e-01
-1.76361322e+00 8.38439524e-01 5.87673366e-01 -4.35656399e-01
-5.63749492e-01 -7.80891120e-01 -1.25428176e+00 -1.13780148e-01
4.82561648e-01 -6.93748891e-01 1.38761926e+00 -6.79072857e-01
-1.30783272e+00 1.02639735e+00 -3.58284354e-01 -3.98936778e-01
2.66709495e-02 -4.02392626e-01 -5.19826651e-01 -3.49011391e-01
6.77250385e-01 7.82749891e-01 1.02139100e-01 -1.04892099e+00
-7.82390654e-01 -3.34840626e-01 3.81413311e-01 2.72894770e-01
-1.58832327e-01 2.63374925e-01 -7.98768461e-01 -4.97791797e-01
-1.00887775e-01 -7.40026593e-01 -1.31124659e-02 -5.69071889e-01
-3.79279882e-01 -5.47996044e-01 4.73172098e-01 -7.66738832e-01
1.14030695e+00 -2.15958428e+00 -1.88136205e-01 -5.51302075e-01
-4.42799956e-01 5.95993586e-02 -3.77557933e-01 5.96394300e-01
8.78585130e-02 2.72003829e-01 -2.73237467e-01 -3.77814412e-01
1.44189849e-01 7.37092793e-01 -3.47698838e-01 -2.23758936e-01
2.34461740e-01 1.26569998e+00 -1.01499236e+00 -5.86760640e-01
-2.49870890e-03 2.50903785e-01 -6.63641870e-01 3.33223313e-01
-5.61837316e-01 3.74091893e-01 -6.59515023e-01 1.01764977e+00
4.13295805e-01 -2.69613862e-01 3.30405891e-01 -1.57331675e-01
1.03653476e-01 7.51950502e-01 -7.91656137e-01 2.24098063e+00
-1.17653298e+00 2.18325332e-01 -1.99171007e-01 -7.36796379e-01
9.06947374e-01 3.75584513e-01 1.94973007e-01 -1.03712773e+00
-1.71945468e-01 5.28065205e-01 -3.82269651e-01 -6.25906408e-01
4.96840745e-01 -2.43330553e-01 -7.61924148e-01 3.64384681e-01
3.54372770e-01 1.16382260e-03 3.34232211e-01 6.55890033e-02
9.15746570e-01 3.36435556e-01 4.45678174e-01 -3.01625073e-01
4.44195509e-01 1.45616993e-01 7.29436040e-01 5.66484213e-01
-4.75391485e-02 2.41984040e-01 7.42606223e-01 -2.58788466e-01
-6.25645339e-01 -1.26246238e+00 -2.17202649e-01 1.64073658e+00
2.27680266e-01 -4.99546498e-01 -5.98338127e-01 -8.88610840e-01
3.99674326e-02 1.25356901e+00 -3.37411910e-02 -6.60534650e-02
-5.34984589e-01 -8.13589752e-01 9.01524425e-01 5.39881825e-01
5.33380747e-01 -1.26701033e+00 -2.08562270e-01 2.05146566e-01
-4.87297654e-01 -1.74491000e+00 -2.84456491e-01 1.43449113e-01
-6.32806361e-01 -1.13273549e+00 -2.20130265e-01 -1.02971935e+00
1.87905863e-01 2.55240221e-02 1.55744421e+00 -4.28356230e-01
4.53592837e-02 3.16220045e-01 -6.47984087e-01 -3.12086880e-01
-5.58398247e-01 3.37059677e-01 -2.05752388e-01 -4.09330845e-01
8.27591360e-01 -3.11094224e-01 -1.64337352e-01 1.92277849e-01
-7.47385263e-01 1.46473736e-01 6.49031341e-01 7.54347265e-01
6.18444741e-01 -3.66865307e-01 9.39862907e-01 -1.30272424e+00
7.37008214e-01 -7.91528165e-01 -5.44122815e-01 7.14440584e-01
-6.24685168e-01 2.06807837e-01 8.08615208e-01 -7.84499645e-02
-1.43610346e+00 -3.52109462e-01 -4.78523940e-01 -6.41869605e-02
-1.70433104e-01 8.16969514e-01 -3.33416998e-01 3.70588243e-01
5.63023210e-01 2.47148529e-01 -2.51055241e-01 -6.91003144e-01
7.37959146e-01 7.59874701e-01 4.36555475e-01 -1.21352351e+00
1.66944847e-01 3.24526522e-03 -7.39989102e-01 -4.90900785e-01
-1.24193287e+00 -2.87109196e-01 -4.67737079e-01 3.84923011e-01
9.30094242e-01 -1.23564517e+00 -2.10016429e-01 8.68228450e-02
-1.24606526e+00 -4.00040060e-01 -1.88990638e-01 4.50835586e-01
-4.95507121e-01 1.33251876e-01 -8.68463218e-01 -3.59875083e-01
-5.16605079e-01 -1.26115549e+00 1.33554006e+00 -8.96826833e-02
-1.47014752e-01 -1.35083950e+00 -3.48477922e-02 5.64670563e-01
2.94589907e-01 1.06630065e-01 1.45600009e+00 -7.85868347e-01
-3.43505830e-01 4.82454151e-02 -4.34842259e-01 4.76827830e-01
3.07646364e-01 -3.39488000e-01 -8.12297702e-01 -1.82426631e-01
-1.06962785e-01 -9.15970564e-01 6.51666462e-01 1.58571631e-01
9.59637105e-01 -2.53887594e-01 -1.56337589e-01 6.42326653e-01
1.70879173e+00 1.93700016e-01 4.88106787e-01 3.43309671e-01
7.21411288e-01 4.76326853e-01 8.33282828e-01 5.73417731e-02
1.18009543e+00 3.83607119e-01 -1.39998309e-02 1.69684570e-02
-2.39956409e-01 -5.35002351e-01 8.70925307e-01 1.27302074e+00
4.26306069e-01 -1.48720786e-01 -1.13906741e+00 6.43205404e-01
-1.74728382e+00 -3.31680864e-01 7.99445435e-02 2.19389725e+00
1.17559969e+00 -3.26467515e-03 -2.89841086e-01 -6.78716242e-01
4.98024017e-01 1.88560724e-01 -7.11474955e-01 -6.65658891e-01
-2.71077096e-01 4.01168048e-01 4.94463891e-01 5.40702343e-01
-8.89870763e-01 1.73649871e+00 5.65694380e+00 9.75736678e-01
-9.74486411e-01 5.49780250e-01 4.42657828e-01 2.28819638e-01
-5.85750282e-01 2.65318900e-01 -1.17579663e+00 3.89819324e-01
1.09049892e+00 -2.83426374e-01 4.48363811e-01 1.02681065e+00
-2.41629645e-01 1.56330824e-01 -1.24560726e+00 8.33212733e-01
-3.12862061e-02 -1.02506280e+00 2.03460306e-01 -5.37448466e-01
6.93795145e-01 6.65873587e-01 -2.79647231e-01 1.05156767e+00
8.76981139e-01 -8.72060537e-01 5.41600347e-01 -1.39945513e-03
1.26267350e+00 -5.54764032e-01 7.17382133e-01 3.16792369e-01
-1.39103580e+00 3.36149819e-02 -5.32685518e-01 1.83393970e-01
2.83884257e-01 2.14617491e-01 -4.07699496e-01 1.04107475e+00
7.00326741e-01 1.01381242e+00 -5.30630112e-01 -1.26646683e-02
-4.26624417e-01 5.49951971e-01 -6.87747225e-02 1.00766994e-01
4.30622399e-01 -1.20414734e-01 6.86706454e-02 1.39821720e+00
4.80735213e-01 -2.51373857e-01 4.19864416e-01 9.40190077e-01
-2.02101246e-01 4.03956473e-01 -6.63282394e-01 -1.28098130e-01
7.03918219e-01 1.04252720e+00 -2.04323381e-01 -5.37837625e-01
-1.08522403e+00 6.82448685e-01 4.85916913e-01 4.56622779e-01
-8.14697623e-01 -2.57536083e-01 8.14860821e-01 -1.46782428e-01
-7.99342692e-02 -7.03919306e-02 -3.59855920e-01 -1.63752508e+00
4.78405915e-02 -1.13821888e+00 8.17042053e-01 -7.49963582e-01
-1.59483886e+00 6.24902368e-01 1.41270399e-01 -9.25646126e-01
-4.29759443e-01 -7.54262686e-01 -4.11488652e-01 8.52657914e-01
-2.01976299e+00 -1.52766287e+00 3.37603390e-02 6.50192738e-01
9.64219451e-01 -3.90959740e-01 1.02127266e+00 5.22461176e-01
-5.59931815e-01 7.50464559e-01 -1.17185876e-01 2.43861675e-01
9.51823533e-01 -1.19337046e+00 6.30070865e-01 6.55315042e-01
9.76117924e-02 7.84639776e-01 1.49900854e-01 -5.36878943e-01
-1.71530032e+00 -1.37315941e+00 1.23995662e+00 -5.19506574e-01
9.08799708e-01 -4.17428076e-01 -1.00756347e+00 1.21349812e+00
3.11822116e-01 -1.89139575e-01 7.52600610e-01 5.35488784e-01
-5.41170955e-01 -4.58326563e-02 -9.31821048e-01 6.08165979e-01
1.08221900e+00 -7.58585572e-01 -7.23425984e-01 3.87448519e-01
1.29903936e+00 -4.01304394e-01 -1.18274379e+00 5.12149811e-01
4.79202002e-01 -8.16767395e-01 9.21593964e-01 -6.49115920e-01
6.33448601e-01 1.08806618e-01 -6.28620446e-01 -1.14242852e+00
1.73218757e-01 -2.48227581e-01 2.02627867e-01 1.35133541e+00
7.65467703e-01 -9.98245299e-01 3.82368773e-01 3.62954766e-01
-4.00783777e-01 -8.70052576e-01 -8.12636018e-01 -1.01208055e+00
5.75097978e-01 -9.13679481e-01 7.91417778e-01 1.20526838e+00
4.25073272e-03 7.96250105e-01 -3.96035761e-02 6.67064339e-02
3.67824852e-01 3.74766767e-01 7.15515733e-01 -7.52178669e-01
-2.86337674e-01 -2.49334544e-01 2.05012009e-01 -1.25357044e+00
6.62553966e-01 -1.45840526e+00 -2.05717430e-01 -1.62652767e+00
1.03240073e-01 -7.05487788e-01 -2.98707396e-01 9.51556325e-01
-2.60091722e-01 -2.49732748e-01 3.42830457e-02 8.85874853e-02
-6.03770912e-01 5.90783179e-01 1.05405986e+00 8.76626000e-02
1.62786618e-02 -2.80171931e-01 -1.02154458e+00 7.05750406e-01
8.08121860e-01 -4.50794369e-01 -6.00519001e-01 -1.00699139e+00
2.66274422e-01 1.87980875e-01 6.18679728e-03 -5.61412454e-01
-2.10571028e-02 -1.73178777e-01 -2.77294844e-01 -2.97446311e-01
7.81904086e-02 -3.74922842e-01 -2.79846728e-01 2.49006569e-01
-3.12447727e-01 2.63408124e-01 2.61157483e-01 1.40635356e-01
-5.58791220e-01 -1.71399906e-01 5.51624238e-01 -4.99618262e-01
-1.14329422e+00 1.65491253e-01 1.41728476e-01 7.46040463e-01
5.68843722e-01 9.77544263e-02 -4.37637597e-01 -8.87471586e-02
-4.84336108e-01 5.77457130e-01 2.43079677e-01 9.66493428e-01
3.48022252e-01 -1.37119210e+00 -6.90580666e-01 2.67357647e-01
5.51073015e-01 1.40412852e-01 6.05049580e-02 6.05150402e-01
-3.44760954e-01 7.16418624e-01 2.13619471e-02 -5.35444796e-01
-8.49790633e-01 3.00106913e-01 1.07223466e-01 -5.00202239e-01
-3.90596867e-01 8.07085156e-01 4.70215410e-01 -1.30451465e+00
5.15372679e-02 -6.34209692e-01 -6.94640204e-02 -1.92256749e-01
1.41381025e-01 5.49398474e-02 -5.43326996e-02 -4.47882593e-01
-2.71023035e-01 6.35689437e-01 -3.89244966e-02 2.72276014e-01
1.12135267e+00 -1.23162873e-01 -2.75595516e-01 6.57124460e-01
1.41639125e+00 -7.61940777e-02 -8.07782531e-01 -6.82380080e-01
4.27473575e-01 -8.07587355e-02 -3.48881751e-01 -8.91867518e-01
-8.47249925e-01 7.92070031e-01 2.10258543e-01 -1.83429703e-01
1.08930039e+00 3.75295490e-01 1.31508172e+00 4.98360246e-01
7.71244586e-01 -1.05691099e+00 -1.08353794e-01 9.64957535e-01
7.56220341e-01 -1.49580956e+00 -5.18650174e-01 -5.14844596e-01
-1.01336145e+00 7.20881224e-01 8.17025125e-01 2.00627282e-01
5.85353374e-01 3.47149193e-01 2.36810222e-01 -1.99381620e-01
-1.04476976e+00 -3.49446833e-01 -7.07174977e-03 3.20626259e-01
9.54198360e-01 2.66617775e-01 -2.91705400e-01 1.07805550e+00
-5.24277806e-01 1.87998191e-02 4.62493412e-02 9.60275114e-01
-1.72337845e-01 -1.46619630e+00 3.51013690e-02 2.62339622e-01
-6.19619787e-01 -5.45167208e-01 -1.42640322e-01 9.63145673e-01
1.85678661e-01 8.81052256e-01 -1.02359578e-01 -1.89094409e-01
4.47134048e-01 4.43735868e-01 2.41539106e-01 -9.81854498e-01
-3.98754895e-01 -5.67365140e-02 4.47481871e-01 -5.96446991e-01
-1.32254273e-01 -5.16949832e-01 -1.48928452e+00 -1.72874406e-01
6.27834722e-03 5.42896762e-02 5.42296469e-01 9.48430598e-01
4.98930126e-01 3.86848927e-01 3.19163650e-01 1.15089223e-01
-5.16309381e-01 -9.31066215e-01 -2.88349479e-01 4.96360660e-01
-1.81642681e-01 -4.10500497e-01 -6.70615360e-02 1.24677323e-01] | [10.879834175109863, 9.375723838806152] |
29a09de9-a1a9-4f7f-b6c4-12aa3767b204 | unsupervised-extractive-opinion-summarization | 2203.07921 | null | https://arxiv.org/abs/2203.07921v3 | https://arxiv.org/pdf/2203.07921v3.pdf | Unsupervised Extractive Opinion Summarization Using Sparse Coding | Opinion summarization is the task of automatically generating summaries that encapsulate information from multiple user reviews. We present Semantic Autoencoder (SemAE) to perform extractive opinion summarization in an unsupervised manner. SemAE uses dictionary learning to implicitly capture semantic information from the review and learns a latent representation of each sentence over semantic units. A semantic unit is supposed to capture an abstract semantic concept. Our extractive summarization algorithm leverages the representations to identify representative opinions among hundreds of reviews. SemAE is also able to perform controllable summarization to generate aspect-specific summaries. We report strong performance on SPACE and AMAZON datasets, and perform experiments to investigate the functioning of our model. Our code is publicly available at https://github.com/brcsomnath/SemAE. | ['Snigdha Chaturvedi', 'Chao Zhao', 'Somnath Basu Roy Chowdhury'] | 2022-03-15 | null | https://aclanthology.org/2022.acl-long.86 | https://aclanthology.org/2022.acl-long.86.pdf | acl-2022-5 | ['extractive-summarization'] | ['natural-language-processing'] | [ 2.66456544e-01 6.47527635e-01 -2.34008268e-01 -4.63007510e-01
-1.06010866e+00 -4.35023844e-01 6.48370326e-01 5.44237792e-01
5.73342061e-03 8.39761436e-01 1.25346577e+00 2.00710580e-01
4.84418839e-01 -8.72217655e-01 -6.95815921e-01 -3.56478602e-01
5.12173593e-01 4.62897748e-01 -4.79626447e-01 -3.39410633e-01
5.89914918e-01 -1.75130144e-01 -1.46670318e+00 6.17424786e-01
1.12751245e+00 6.65177941e-01 1.21098109e-01 9.05626297e-01
-4.39430416e-01 9.42523897e-01 -1.15240550e+00 -6.75151289e-01
-1.19629309e-01 -7.12150097e-01 -7.51560092e-01 3.70900899e-01
4.41419899e-01 -3.21556360e-01 -8.26606974e-02 1.13919175e+00
6.10620916e-01 3.26602608e-01 9.23812389e-01 -8.20915341e-01
-1.32936335e+00 1.14921176e+00 -1.93169281e-01 4.58531454e-02
5.19926071e-01 -1.06227793e-01 1.65354693e+00 -1.15001571e+00
7.21963942e-01 1.16188228e+00 3.02249998e-01 6.44277394e-01
-8.58500719e-01 -1.60244390e-01 1.95519581e-01 -1.74224973e-01
-7.53342032e-01 -6.41074121e-01 7.90227234e-01 2.84894817e-02
1.26569688e+00 1.45947605e-01 7.45695114e-01 1.24977589e+00
3.39349687e-01 1.46014595e+00 7.48734355e-01 -2.51769274e-01
7.51386404e-01 1.51503816e-01 5.77545524e-01 5.37252963e-01
7.81701863e-01 -7.23739982e-01 -7.64169037e-01 -2.79828399e-01
1.78166360e-01 1.81754515e-01 -9.65506509e-02 6.75882995e-02
-8.42854738e-01 1.20934296e+00 2.81984180e-01 1.21318266e-01
-8.98687959e-01 1.53543279e-01 6.17286026e-01 2.96853304e-01
9.51548994e-01 1.02154338e+00 -5.32008708e-01 -5.99921905e-02
-9.57370877e-01 3.01975965e-01 1.29244912e+00 1.06753635e+00
7.89553940e-01 5.39387107e-01 -3.43166381e-01 8.16896200e-01
-1.73166674e-02 8.71775150e-01 9.37280416e-01 -1.18047810e+00
1.79413944e-01 7.76903391e-01 -2.41541695e-02 -1.14429057e+00
4.22430150e-02 -3.33377004e-01 -7.50702262e-01 -5.21328032e-01
-7.45022893e-01 -5.47915041e-01 -1.03113306e+00 1.18806076e+00
-1.33671224e-01 -5.92385121e-02 7.71736681e-01 5.19202530e-01
1.57258809e+00 1.11428988e+00 9.95266289e-02 -2.21154243e-01
1.31105113e+00 -1.17349672e+00 -8.11912417e-01 -5.44025660e-01
5.21125734e-01 -4.00774181e-01 9.03334260e-01 2.32719466e-01
-1.16757631e+00 -2.47391060e-01 -1.18936682e+00 -1.95138469e-01
-4.31913644e-01 2.91691124e-01 6.03907287e-01 9.22176242e-02
-1.21048868e+00 5.29121280e-01 -7.17844665e-01 -4.92348611e-01
5.30110419e-01 3.62299988e-03 -3.09938788e-01 6.42886683e-02
-1.04265261e+00 7.43936121e-01 4.99419779e-01 -3.40871811e-01
-7.48666525e-01 -4.89745706e-01 -1.45505071e+00 2.45826304e-01
1.93494201e-01 -1.33942187e+00 1.45990026e+00 -1.32329559e+00
-1.50260794e+00 6.55728579e-01 -5.94839752e-01 -7.85728753e-01
-4.21870232e-01 -4.99323189e-01 -3.46048176e-01 5.05662024e-01
4.77168977e-01 6.96274340e-01 9.49740469e-01 -1.41516256e+00
-3.69388968e-01 -3.08316946e-01 1.34181092e-03 4.58399355e-01
-4.17173207e-01 -1.34280995e-01 -1.45655498e-01 -8.34966600e-01
-3.18087697e-01 -5.85099757e-01 -4.59017813e-01 -1.19258833e+00
-7.04139829e-01 -3.80401701e-01 3.24888676e-01 -8.85193825e-01
1.28566003e+00 -1.64414787e+00 3.08931708e-01 -1.82531029e-01
1.70845062e-01 1.29562974e-01 -3.66661817e-01 7.89497256e-01
1.63085371e-01 8.53693783e-02 -5.90227246e-01 -7.96298444e-01
2.47703418e-01 1.09831847e-01 -8.66332531e-01 -2.15052143e-01
2.85094827e-01 1.19980729e+00 -1.02155733e+00 -2.51271695e-01
-1.22748248e-01 3.31388742e-01 -6.15539849e-01 2.21762761e-01
-6.77512765e-01 -1.03818774e-01 -8.02302063e-01 6.56659544e-01
1.97071806e-01 -4.11216259e-01 -5.31499460e-02 -1.73052698e-01
2.59841233e-01 5.62083185e-01 -4.75613773e-01 1.93969476e+00
-5.05474865e-01 5.08124173e-01 -4.19960350e-01 -1.13530040e+00
9.67737257e-01 3.15735012e-01 1.75423041e-01 -4.26105231e-01
3.63643110e-01 1.37483969e-01 -8.70554984e-01 -2.40127116e-01
1.32435310e+00 -1.29958987e-01 -6.21273100e-01 9.81098115e-01
4.24240947e-01 -5.99866033e-01 3.53330791e-01 8.25146437e-01
1.06150734e+00 -2.26953387e-01 7.42829144e-01 -1.44643947e-01
3.02110225e-01 3.65012616e-01 3.21078718e-01 8.12980354e-01
2.83837318e-01 7.89487123e-01 3.62979025e-01 -3.04900616e-01
-9.67474639e-01 -1.07225847e+00 5.38668513e-01 7.52925754e-01
-6.92429385e-05 -1.04408038e+00 -7.51321256e-01 -7.84798980e-01
-3.84096205e-02 1.50694180e+00 -6.11083031e-01 -4.34730053e-01
-9.03773829e-02 -4.08963531e-01 1.15253240e-01 6.43407524e-01
2.27245852e-01 -1.54557550e+00 -4.43874866e-01 1.23789057e-01
-2.65210807e-01 -8.40526640e-01 -4.81586844e-01 -1.22653700e-01
-8.17701161e-01 -6.67154968e-01 -6.05411410e-01 -7.39577234e-01
9.03720975e-01 3.25687081e-01 1.47556424e+00 -2.34024748e-01
6.15092814e-02 9.50880706e-01 -7.44676828e-01 -8.45951974e-01
-6.37283862e-01 3.29348952e-01 1.14501208e-01 -7.93201700e-02
7.43386745e-01 -6.13547564e-01 -6.61128283e-01 -6.65650845e-01
-1.01597273e+00 -1.28519610e-02 5.80881238e-01 5.87744176e-01
7.12681830e-01 -1.44807607e-01 9.45192397e-01 -1.21527147e+00
1.35912597e+00 -7.42929280e-01 8.74490216e-02 8.99911895e-02
-4.40439939e-01 2.96204656e-01 8.01669478e-01 2.96741351e-02
-1.09156346e+00 -4.00672287e-01 -6.92721754e-02 -2.02304363e-01
-1.64317071e-01 8.33202958e-01 1.19433440e-01 9.78583813e-01
6.06302917e-01 6.63087070e-01 -7.24573731e-02 -2.19311938e-01
8.68463993e-01 8.72644782e-01 4.70517844e-01 -5.08578181e-01
4.55799043e-01 6.42644346e-01 -8.40208292e-01 -1.01166034e+00
-1.63636220e+00 -3.68621796e-01 -1.14658028e-01 3.97415049e-02
9.48508203e-01 -1.26072717e+00 7.70947838e-04 -1.44856353e-03
-1.28163183e+00 -2.10314952e-02 -9.34301198e-01 1.24011897e-01
-5.84115267e-01 2.27506518e-01 -5.84193826e-01 -4.28003043e-01
-1.46636677e+00 -5.53955257e-01 1.37002718e+00 6.49233580e-01
-8.31385374e-01 -1.13907695e+00 3.29313397e-01 4.99672174e-01
3.80851895e-01 1.69651374e-01 3.40231985e-01 -1.21945453e+00
-9.47055370e-02 -5.52059591e-01 2.10563049e-01 8.55435967e-01
3.37789178e-01 -8.19598660e-02 -7.72773206e-01 -2.94988066e-01
2.74769008e-01 -5.18797338e-01 1.32476771e+00 6.29410326e-01
9.13613617e-01 -9.33909357e-01 2.34377887e-02 -2.82422230e-02
1.26980257e+00 -2.21925005e-01 3.73650670e-01 -6.50774408e-03
7.43200302e-01 5.58414757e-01 4.26106215e-01 7.18497634e-01
7.70621955e-01 -2.40786567e-01 -3.83229740e-02 2.16664299e-01
-1.15312539e-01 -4.77830052e-01 8.66084158e-01 1.62113822e+00
2.38289893e-01 -3.21358562e-01 -4.85808372e-01 9.00767744e-01
-1.80439508e+00 -1.11850810e+00 3.33985209e-01 1.42415369e+00
8.87705326e-01 5.71560785e-02 -1.83292657e-01 -3.83705705e-01
5.42094350e-01 6.04936719e-01 -6.49043083e-01 -8.80373478e-01
-2.38993555e-01 5.68805575e-01 1.17799722e-01 5.16644716e-01
-7.55601346e-01 1.28293812e+00 5.93771744e+00 6.30688488e-01
-7.04118252e-01 5.85269518e-02 4.33810622e-01 -2.38945991e-01
-1.02548027e+00 1.14709206e-01 -6.34138525e-01 2.80906856e-01
1.08040154e+00 -8.24545026e-01 -1.05510065e-02 9.99154449e-01
8.10604766e-02 -3.17391120e-02 -7.71936178e-01 6.29163206e-01
8.54963541e-01 -1.72670734e+00 8.69497597e-01 -3.94841939e-01
1.41028225e+00 1.96862593e-01 -4.98737092e-04 4.09269065e-01
8.77785265e-01 -8.88157248e-01 3.98349196e-01 7.09461033e-01
3.72469336e-01 -9.17544067e-01 7.78340697e-01 3.91157940e-02
-7.48471975e-01 1.16164818e-01 -6.24936283e-01 -1.80892780e-01
2.22547412e-01 7.59259939e-01 -7.53939152e-01 7.13422239e-01
3.56910139e-01 1.52932680e+00 -5.89579463e-01 3.16721350e-01
-8.31519604e-01 7.11238980e-01 1.35387108e-01 -3.04462701e-01
2.63315052e-01 -3.46225649e-01 8.46635818e-01 1.30232131e+00
3.92750353e-01 2.34820426e-01 4.40089442e-02 8.57578397e-01
-4.24700618e-01 2.49616876e-01 -7.67285585e-01 -5.23347437e-01
4.52326417e-01 1.32809830e+00 -5.15492201e-01 -9.87570047e-01
-2.17580542e-01 1.40153980e+00 1.25230223e-01 4.66230780e-01
-3.54463190e-01 -5.49716890e-01 6.61045969e-01 -3.86411309e-01
5.67465723e-01 8.38859454e-02 -5.26392758e-01 -1.77325058e+00
-2.18701120e-02 -1.00583673e+00 4.14848953e-01 -1.14628196e+00
-1.34960198e+00 5.97741365e-01 -3.37355494e-01 -8.73287737e-01
-3.54078859e-01 -1.30391210e-01 -1.14740872e+00 3.63545567e-01
-1.20492935e+00 -1.01349485e+00 -3.35463285e-02 2.09586054e-01
1.28265631e+00 -4.68972474e-01 1.15912163e+00 -5.28696597e-01
-5.73020756e-01 2.15954304e-01 1.93911344e-02 -1.48740653e-02
5.86605668e-01 -1.65685773e+00 7.65751243e-01 9.11616564e-01
2.65226960e-01 9.32627201e-01 1.04214752e+00 -8.44529808e-01
-1.23154390e+00 -1.26770353e+00 1.19150615e+00 -8.44677031e-01
7.05180585e-01 6.87855110e-03 -6.35373414e-01 9.55771148e-01
9.40171778e-01 -8.12779546e-01 1.26702166e+00 -1.56510379e-02
-2.94641048e-01 8.92793015e-02 -7.06088006e-01 8.39716494e-01
7.07444012e-01 -5.70330203e-01 -1.44651234e+00 3.36230338e-01
1.31595421e+00 1.59098636e-02 -7.85648108e-01 3.21458466e-02
1.32249802e-01 -7.20953405e-01 6.34715199e-01 -6.89535379e-01
1.29936349e+00 -1.40214145e-01 -1.32039383e-01 -1.99348748e+00
1.03861682e-01 -4.49449599e-01 -6.85231268e-01 1.18533862e+00
5.53025365e-01 -5.79840302e-01 5.73255539e-01 3.81105632e-01
-4.30228472e-01 -7.04740524e-01 -1.72533125e-01 -3.20394725e-01
1.39262294e-02 -1.27212137e-01 6.07173264e-01 7.35065401e-01
2.10788310e-01 1.00880527e+00 -1.76143929e-01 -5.18831909e-02
6.48790121e-01 4.86660331e-01 8.02586138e-01 -9.12987232e-01
-8.06827918e-02 -3.47347081e-01 -2.94073075e-02 -9.91263568e-01
7.42604375e-01 -1.12088609e+00 -6.10796325e-02 -2.33707237e+00
3.40002894e-01 6.27381325e-01 -2.48453185e-01 3.77313733e-01
-3.87207121e-01 5.27711809e-02 3.24098542e-02 5.40978946e-02
-1.23319328e+00 1.11648118e+00 1.04476023e+00 -4.42238688e-01
-1.69521585e-01 -4.27020937e-02 -1.78057790e+00 7.65784681e-01
1.16114950e+00 -4.24444884e-01 -5.37875056e-01 -5.19748092e-01
4.23528075e-01 -2.53008962e-01 -1.38504103e-01 -6.58225894e-01
2.43044212e-01 2.36457631e-01 2.99175948e-01 -7.05331802e-01
1.00504719e-01 -2.10300654e-01 -4.90692616e-01 2.03542709e-01
-6.84938312e-01 1.02369353e-01 6.40649348e-02 5.85452855e-01
-5.66967607e-01 -1.56593516e-01 1.70659557e-01 -4.55824077e-01
-6.41818464e-01 1.80910364e-01 -5.73099136e-01 2.05427095e-01
4.89564121e-01 1.05165504e-01 -3.17364395e-01 -8.62532616e-01
-5.87826967e-01 4.26152468e-01 6.88250542e-01 5.03502846e-01
1.00884557e+00 -1.24610746e+00 -1.13306212e+00 -1.24646492e-01
3.72136831e-01 8.19272399e-02 2.03606904e-01 1.56585798e-01
-3.70445818e-01 4.05890852e-01 1.01503842e-01 2.44350843e-02
-8.40692222e-01 3.60835403e-01 -1.10257007e-01 -2.85488695e-01
-7.82816827e-01 7.69338608e-01 2.52860393e-02 -4.16379303e-01
-2.11427167e-01 -1.68062747e-01 -5.32246530e-01 4.43571210e-01
8.47576916e-01 2.27415293e-01 -1.22035250e-01 -5.00105321e-01
1.79962311e-02 2.43365839e-01 -3.74895632e-01 -1.34834260e-01
1.75635862e+00 -3.04698855e-01 -4.87481117e-01 5.54165959e-01
1.24399769e+00 1.69928357e-01 -7.95407891e-01 -2.99767941e-01
5.48058748e-02 2.60406703e-01 -5.79606928e-02 -5.98193765e-01
-8.01256180e-01 3.99064571e-01 -5.18398345e-01 2.49211773e-01
1.03543711e+00 2.45181561e-01 1.23212302e+00 7.05885708e-01
-1.13905057e-01 -1.30718911e+00 3.85873616e-01 7.11861789e-01
1.13498461e+00 -1.26029515e+00 2.39854217e-01 1.85316578e-01
-1.36361635e+00 9.08595145e-01 4.30666178e-01 -6.96661055e-01
3.30543607e-01 -1.53599441e-01 3.25827338e-02 -7.16377795e-01
-1.14772046e+00 -1.61778226e-01 3.31659436e-01 2.18010277e-01
3.86710852e-01 3.35453033e-01 -3.16489995e-01 1.19628441e+00
-7.71534979e-01 -3.32714409e-01 9.47866380e-01 8.87199700e-01
-7.29333460e-01 -7.35475361e-01 8.98460820e-02 8.13416183e-01
-3.39785814e-01 -4.50450838e-01 -9.42752123e-01 -4.56157289e-02
-7.23807335e-01 1.17230511e+00 1.35213882e-01 -2.90068865e-01
3.53833258e-01 1.49291292e-01 2.73854993e-02 -1.36699235e+00
-5.19295633e-01 -2.44469941e-01 2.79332697e-01 -5.10265827e-01
-5.52306592e-01 -5.87904871e-01 -1.44651794e+00 -1.47798881e-01
1.27258226e-01 7.06456423e-01 5.78415275e-01 8.46162319e-01
7.94493616e-01 6.92532659e-01 6.98266029e-01 -6.63677096e-01
-5.20788193e-01 -1.06422400e+00 -4.18101221e-01 5.92279315e-01
3.47131222e-01 1.06617033e-01 -5.52189231e-01 2.27391750e-01] | [12.42184829711914, 9.362201690673828] |
56d9e2cb-8484-4d2c-b3bf-939cf7cffa74 | practical-auto-calibration-for-spatial-scene | 2012.08375 | null | https://arxiv.org/abs/2012.08375v1 | https://arxiv.org/pdf/2012.08375v1.pdf | Practical Auto-Calibration for Spatial Scene-Understanding from Crowdsourced Dashcamera Videos | Spatial scene-understanding, including dense depth and ego-motion estimation, is an important problem in computer vision for autonomous vehicles and advanced driver assistance systems. Thus, it is beneficial to design perception modules that can utilize crowdsourced videos collected from arbitrary vehicular onboard or dashboard cameras. However, the intrinsic parameters corresponding to such cameras are often unknown or change over time. Typical manual calibration approaches require objects such as a chessboard or additional scene-specific information. On the other hand, automatic camera calibration does not have such requirements. Yet, the automatic calibration of dashboard cameras is challenging as forward and planar navigation results in critical motion sequences with reconstruction ambiguities. Structure reconstruction of complete visual-sequences that may contain tens of thousands of images is also computationally untenable. Here, we propose a system for practical monocular onboard camera auto-calibration from crowdsourced videos. We show the effectiveness of our proposed system on the KITTI raw, Oxford RobotCar, and the crowdsourced D$^2$-City datasets in varying conditions. Finally, we demonstrate its application for accurate monocular dense depth and ego-motion estimation on uncalibrated videos. | ['Bahram Zonooz', 'Elahe Arani', 'Shabbir Marzban', 'Matti Jukola', 'Hemang Chawla'] | 2020-12-15 | null | null | null | null | ['camera-auto-calibration'] | ['computer-vision'] | [-3.09016913e-01 -2.33838961e-01 1.78184122e-01 -5.98455608e-01
-6.24674857e-01 -7.14507937e-01 2.33770177e-01 -4.81425464e-01
-6.06075346e-01 7.88348377e-01 -3.59819293e-01 -2.76847720e-01
3.33332777e-01 -4.66376632e-01 -1.09607112e+00 -5.48736691e-01
4.56603229e-01 5.82251370e-01 6.42992556e-01 -2.78805017e-01
2.62427926e-01 4.81770784e-01 -1.61524820e+00 -1.45146564e-01
7.13437319e-01 8.59614074e-01 7.01860487e-01 9.95103657e-01
-3.51430774e-02 9.04412687e-01 -2.05409840e-01 -5.48933446e-01
6.08432293e-01 -1.01623945e-02 -4.42603528e-02 5.31020522e-01
7.76086032e-01 -8.73984098e-01 -7.97579467e-01 1.24251986e+00
9.82765481e-02 8.34283233e-02 2.61365026e-01 -1.53259575e+00
-7.14177564e-02 -2.76698261e-01 -6.75444603e-01 1.06292345e-01
2.97273338e-01 4.49059039e-01 1.73299432e-01 -9.51189816e-01
6.68886244e-01 1.04896510e+00 5.50359130e-01 3.17383379e-01
-4.04575169e-01 -7.21313477e-01 5.25886901e-02 7.06554711e-01
-1.51140261e+00 -7.67176926e-01 7.00599074e-01 -5.72062671e-01
5.32545686e-01 -3.26244414e-01 7.49268830e-01 8.86236489e-01
1.75675943e-01 4.68275756e-01 7.99385548e-01 -1.44674197e-01
2.83706218e-01 3.62503707e-01 -6.94060251e-02 7.86325395e-01
7.72691131e-01 1.58457398e-01 -5.80939949e-01 2.70556569e-01
8.41232359e-01 1.33497089e-01 -2.42791697e-01 -8.00426126e-01
-1.13893616e+00 6.80768371e-01 1.81119859e-01 -3.76971483e-01
-1.22383870e-01 3.72096717e-01 -3.27353110e-03 9.12415907e-02
-1.35101229e-01 -4.60302532e-02 -2.09548727e-01 -2.95729965e-01
-7.29518592e-01 2.01146588e-01 4.96186078e-01 1.83729744e+00
1.32264864e+00 3.32724065e-01 6.14887953e-01 4.50546980e-01
1.57540500e-01 1.19794297e+00 2.09791213e-01 -1.51533163e+00
6.92567825e-01 2.72335023e-01 6.44062877e-01 -1.24211240e+00
-3.16233367e-01 7.31917471e-02 -5.75458884e-01 3.39008212e-01
6.02895081e-01 -3.66068691e-01 -7.28089333e-01 1.25676036e+00
5.73802948e-01 3.69077802e-01 1.01808667e-01 1.37507153e+00
7.49935925e-01 2.58542717e-01 -5.81983626e-01 -2.12099046e-01
1.02606726e+00 -9.70641732e-01 -8.54621649e-01 -7.53925741e-01
4.92595315e-01 -7.91222930e-01 6.29408002e-01 2.33973861e-01
-8.74070406e-01 -6.50746644e-01 -1.08698630e+00 -2.47323439e-01
-6.98478073e-02 3.18272114e-01 3.86007577e-01 6.78559601e-01
-8.72003019e-01 -2.55533874e-01 -8.12443316e-01 -4.39741582e-01
-7.54827410e-02 2.53428161e-01 -7.80643463e-01 -6.97440445e-01
-9.82361853e-01 1.11543739e+00 1.41604051e-01 3.46470833e-01
-9.59808111e-01 -4.60671276e-01 -1.18114924e+00 -4.64098930e-01
6.10258579e-01 -5.47681808e-01 1.20790911e+00 -9.27347422e-01
-1.47281790e+00 7.00324953e-01 -4.07018781e-01 -4.96844530e-01
7.21038461e-01 -1.39156178e-01 -3.20943981e-01 2.91040838e-01
2.09253192e-01 6.65227234e-01 8.02790105e-01 -1.33396482e+00
-8.84511411e-01 -4.46096778e-01 4.04321909e-01 5.54087758e-01
1.39920086e-01 -3.06296498e-01 -8.89812410e-01 2.05601349e-01
3.66732866e-01 -1.26798058e+00 -3.59019130e-01 2.41837934e-01
-1.49540082e-01 6.87500715e-01 8.35069597e-01 -6.21685088e-01
3.77427548e-01 -2.05319452e+00 -1.99119702e-01 -1.14815623e-01
1.54170752e-01 1.60995349e-02 -3.68934050e-02 7.64273033e-02
4.75146830e-01 -7.46134102e-01 2.35714056e-02 -2.45662481e-01
-3.78360510e-01 3.85046661e-01 -9.35975239e-02 8.96690011e-01
-1.52792320e-01 7.21489727e-01 -9.04419661e-01 -4.83123511e-01
7.23780811e-01 2.50287354e-01 -5.56533575e-01 2.35966057e-01
6.22869767e-02 6.82960153e-01 -2.61118978e-01 6.10064030e-01
1.12377620e+00 1.49521306e-01 -4.20531072e-03 -7.40586743e-02
-3.24436963e-01 -3.03191721e-01 -1.44719565e+00 1.62077916e+00
-3.45428824e-01 1.30560195e+00 3.33395153e-01 -7.02318132e-01
8.46632361e-01 -1.18510514e-01 3.04875672e-01 -7.32865214e-01
2.80900091e-01 3.16933572e-01 -2.23026916e-01 -8.20997655e-01
1.07956314e+00 1.52994646e-02 2.94085424e-02 -5.00035621e-02
-2.85118073e-01 -5.73715508e-01 4.65893969e-02 2.73886979e-01
9.31827962e-01 1.44537224e-03 3.29296619e-01 1.76632896e-01
4.20574248e-01 5.69550335e-01 7.12424278e-01 3.88557851e-01
-3.56665134e-01 7.96925187e-01 9.13055316e-02 -2.90023237e-01
-1.43099427e+00 -7.85891593e-01 1.22679383e-01 3.21644425e-01
7.81780422e-01 7.33377188e-02 -6.90513730e-01 -1.48230001e-01
2.27163862e-02 4.83595848e-01 -3.08293998e-01 1.85267553e-01
-5.78995943e-01 -6.87353238e-02 2.48841584e-01 5.07842600e-01
7.78044760e-01 -2.41508976e-01 -8.27791512e-01 3.54037508e-02
-5.08458078e-01 -1.89598489e+00 -6.11475229e-01 -3.64240319e-01
-5.61317921e-01 -1.37988329e+00 -6.31742597e-01 -6.79841220e-01
7.58952141e-01 1.45447874e+00 7.36120224e-01 -2.07306340e-01
1.91653222e-02 6.72790587e-01 -1.30635932e-01 -4.54790831e-01
-1.81534678e-01 -4.27139193e-01 3.72047961e-01 3.19478810e-01
5.34302831e-01 -1.94770396e-01 -6.72146618e-01 7.11447597e-01
-4.12660480e-01 8.97991955e-02 3.27095956e-01 5.50689042e-01
5.34376800e-01 -1.88149601e-01 1.03145257e-01 -5.80983698e-01
-1.76590890e-01 -4.26923543e-01 -1.36898267e+00 -2.09540367e-01
-4.01870124e-02 -3.96445870e-01 4.25576240e-01 -2.20978260e-01
-1.10756922e+00 5.90907991e-01 3.25743914e-01 -1.00294220e+00
-1.34676665e-01 1.30659208e-01 -3.73567194e-01 -3.86458337e-01
7.44701803e-01 1.69261307e-01 1.88859999e-01 1.70302570e-01
3.71425122e-01 8.38807583e-01 8.69791985e-01 -2.01524403e-02
1.01204884e+00 1.00729609e+00 -3.38954777e-02 -1.43017089e+00
-3.83680314e-01 -9.29882228e-01 -8.35700154e-01 -4.60073560e-01
9.75115359e-01 -1.50573123e+00 -6.63359523e-01 4.94409651e-01
-1.42092192e+00 -2.38683388e-01 2.32718527e-01 9.02976573e-01
-6.35412693e-01 7.74988770e-01 -9.80638191e-02 -8.27685773e-01
4.62504745e-01 -1.32968104e+00 1.03573847e+00 2.46991307e-01
3.46794993e-01 -6.97860301e-01 -7.46599063e-02 7.55989194e-01
7.60060921e-02 -1.09210797e-02 1.72675867e-02 2.19087526e-01
-1.30593574e+00 -3.17354798e-01 -2.75090396e-01 1.71715483e-01
-1.17791034e-01 -9.69872400e-02 -9.67060149e-01 -1.33616537e-01
7.15301409e-02 -1.18232809e-01 4.70434666e-01 5.38375139e-01
4.77858126e-01 1.05388016e-01 -1.53193370e-01 8.22566986e-01
1.32318258e+00 2.84570098e-01 7.13195980e-01 4.40453023e-01
8.77691031e-01 7.44633317e-01 1.25452268e+00 4.12368000e-01
9.28972483e-01 8.49265635e-01 6.89190269e-01 1.87995017e-01
3.95362079e-02 -3.36789280e-01 4.28757787e-01 7.59144306e-01
-2.02786624e-02 6.24045357e-03 -8.85190547e-01 6.46612167e-01
-1.89224279e+00 -9.42105711e-01 -6.93756640e-01 2.44217348e+00
2.10666060e-01 -1.49746850e-01 -2.01057345e-01 -2.18426213e-01
7.95949876e-01 -2.48700410e-01 -7.94759572e-01 1.82604700e-01
-2.91954011e-01 -7.31734455e-01 1.21227252e+00 8.62184644e-01
-7.80497909e-01 1.13499713e+00 5.34897470e+00 1.97684124e-01
-8.78935874e-01 2.59231925e-01 1.28363267e-01 -1.12204462e-01
-1.48236036e-01 2.16347024e-01 -1.08242297e+00 4.37370807e-01
7.16066897e-01 1.77045185e-02 4.42074001e-01 1.04208422e+00
4.45302278e-01 -7.63417602e-01 -9.58906174e-01 1.69632578e+00
3.27850521e-01 -1.32435393e+00 -4.26323056e-01 1.58577368e-01
9.85188007e-01 2.13880733e-01 -1.88589290e-01 -4.15923931e-02
2.25452945e-01 -6.50760829e-01 9.70439434e-01 3.45159292e-01
8.35331023e-01 -6.46977365e-01 7.86260366e-01 7.50989854e-01
-1.10634089e+00 -1.49657890e-01 -9.39161718e-01 -7.66484737e-02
4.83398438e-01 4.21157062e-01 -8.67799103e-01 2.21039683e-01
5.23780406e-01 6.93190455e-01 -5.07112265e-01 1.16841376e+00
-1.09037817e-01 -3.92472520e-02 -2.14813858e-01 5.40785976e-02
1.50185212e-01 -5.41711986e-01 4.36009169e-01 6.91891074e-01
6.32077336e-01 5.51130474e-01 -6.82643279e-02 3.23660880e-01
2.33537614e-01 -2.96611756e-01 -1.19276440e+00 5.44722795e-01
5.45623004e-01 1.01668262e+00 -3.80996406e-01 -3.28935921e-01
-7.18884051e-01 9.00796533e-01 5.40549643e-02 5.46853662e-01
-1.14583218e+00 -1.09403312e-01 9.16964173e-01 3.12246799e-01
2.43581861e-01 -9.79385972e-01 -1.52360708e-01 -1.44785678e+00
1.51103571e-01 -5.37818789e-01 -3.05257320e-01 -1.36255121e+00
-5.23693621e-01 4.57126051e-01 -3.14848982e-02 -1.71210539e+00
-2.24610865e-01 -6.09097421e-01 2.77352557e-02 5.74853599e-01
-1.73784614e+00 -9.75586951e-01 -1.09929192e+00 9.16841388e-01
8.72904241e-01 -3.50536108e-01 1.21311424e-02 3.58376563e-01
-2.00446248e-01 3.04759622e-01 3.67296159e-01 -8.26758072e-02
8.18676829e-01 -7.24454284e-01 3.88266861e-01 1.11692512e+00
-1.46939531e-01 1.80053949e-01 9.00410414e-01 -5.78263044e-01
-1.90377891e+00 -1.16794860e+00 5.80112875e-01 -7.66181827e-01
3.95147353e-01 -3.21120203e-01 -7.15682447e-01 8.00705791e-01
2.53240834e-03 2.10324645e-01 5.36127761e-02 -6.41873837e-01
-7.80298840e-03 -3.41321409e-01 -8.79542291e-01 6.40658319e-01
1.11988485e+00 -4.41173792e-01 -2.68306792e-01 4.97155100e-01
5.70695579e-01 -8.96458209e-01 -9.25914645e-02 -1.71043247e-01
5.47173262e-01 -1.20918155e+00 9.55173135e-01 -1.10669166e-01
5.84653299e-03 -7.03971982e-01 -5.86121380e-01 -1.05964625e+00
1.56842545e-01 -5.27681351e-01 1.85600534e-01 6.74920380e-01
1.86293051e-02 -6.47241950e-01 1.00191212e+00 8.64367127e-01
-2.66628504e-01 2.67730981e-01 -1.06320703e+00 -7.53336966e-01
-4.60118055e-01 -7.77952015e-01 2.81325340e-01 7.10664272e-01
-2.76657432e-01 2.92545676e-01 -7.75753081e-01 6.99579239e-01
9.14853334e-01 -1.82948887e-01 1.73621011e+00 -9.59264040e-01
-1.02074012e-01 2.27250174e-01 -9.19673324e-01 -1.51810610e+00
2.07791075e-01 -2.30778426e-01 4.34291363e-01 -1.38079083e+00
-1.00949563e-01 -4.25611287e-01 6.55212700e-01 -3.32887590e-01
9.03717950e-02 1.28918976e-01 6.95954189e-02 2.97769517e-01
-5.24758399e-01 5.82827628e-01 1.16955698e+00 -1.12820767e-01
-4.58839862e-03 6.11307472e-03 -1.93564773e-01 8.36964190e-01
5.55902600e-01 -2.08791643e-01 -7.87257373e-01 -9.73181725e-01
2.54116327e-01 6.51765049e-01 4.67481554e-01 -1.10778773e+00
7.96584547e-01 -4.97858047e-01 3.80528606e-02 -8.22380006e-01
9.00227606e-01 -1.13571656e+00 4.58981395e-01 2.62730181e-01
3.54856014e-01 2.40519613e-01 5.43960277e-03 9.55912173e-01
-2.12158114e-01 -2.27697298e-01 9.36505318e-01 -2.55509049e-01
-1.25014937e+00 3.73961985e-01 -4.21770394e-01 -1.87271442e-02
1.40228295e+00 -6.94727778e-01 -4.65521276e-01 -7.91830420e-01
-8.32957104e-02 3.21252465e-01 9.13263142e-01 4.28721786e-01
1.00749135e+00 -1.24339497e+00 -4.86313254e-01 5.27459323e-01
3.77949059e-01 3.29573154e-01 5.36599934e-01 9.00041461e-01
-1.15447462e+00 6.30029798e-01 -3.22886139e-01 -9.87161815e-01
-1.26261485e+00 6.66363835e-01 4.18604881e-01 7.36733794e-01
-6.12816334e-01 4.40164715e-01 4.37389106e-01 -2.62168169e-01
-4.31435220e-02 -4.20294672e-01 5.61360903e-02 -2.95395017e-01
5.34921050e-01 5.70813596e-01 1.19133510e-01 -1.18092835e+00
-2.96330988e-01 9.96473610e-01 2.75937378e-01 -3.58680874e-01
6.70371413e-01 -8.52277756e-01 3.35795164e-01 2.56163478e-01
1.00318336e+00 6.42390773e-02 -1.77986515e+00 -2.01749191e-01
-4.44162101e-01 -9.95527685e-01 -7.21413344e-02 9.20925885e-02
-1.01107979e+00 1.05555201e+00 4.64393318e-01 -5.41457415e-01
6.60605848e-01 -2.92100549e-01 6.75874770e-01 9.42067802e-01
1.09564018e+00 -1.28058827e+00 1.77979964e-04 6.37006819e-01
6.00512624e-01 -1.74330282e+00 -1.65923208e-01 -4.51919377e-01
-1.00847363e+00 1.13230073e+00 8.62155616e-01 1.38957817e-02
5.16265869e-01 9.72794220e-02 3.83972019e-01 5.12548722e-02
-4.81248111e-01 -2.89965957e-01 -2.18255252e-01 8.84626806e-01
-4.57228959e-01 4.45962287e-02 2.75906682e-01 1.17149055e-01
-2.68033922e-01 -1.24772929e-01 1.31581128e+00 8.16930175e-01
-6.94768488e-01 -4.97403145e-01 -8.36979866e-01 -2.09542108e-03
3.84791344e-01 2.66212225e-01 6.59820139e-02 9.32420611e-01
3.96926850e-02 1.22969711e+00 9.63338464e-02 -3.98138136e-01
4.39942241e-01 -4.11804825e-01 4.73799586e-01 -4.18044001e-01
3.20663810e-01 -7.21737370e-02 1.66290060e-01 -6.20191574e-01
-3.18629682e-01 -8.91385555e-01 -1.13492525e+00 -5.44716179e-01
-2.56067455e-01 -1.04350753e-01 9.68464077e-01 9.30368066e-01
2.05968961e-01 -1.32265180e-01 7.02738643e-01 -1.13687348e+00
-1.24332793e-01 -6.39884174e-01 -6.56756938e-01 1.78842302e-02
6.45828962e-01 -9.23160732e-01 -2.22306222e-01 3.73900473e-01] | [8.104170799255371, -2.0753567218780518] |
cadfa467-c46d-4038-a31c-73ad3d6df9e9 | a-robust-visual-system-for-small-target | 1904.04363 | null | http://arxiv.org/abs/1904.04363v1 | http://arxiv.org/pdf/1904.04363v1.pdf | A Robust Visual System for Small Target Motion Detection Against Cluttered Moving Backgrounds | Monitoring small objects against cluttered moving backgrounds is a huge
challenge to future robotic vision systems. As a source of inspiration, insects
are quite apt at searching for mates and tracking prey -- which always appear
as small dim speckles in the visual field. The exquisite sensitivity of insects
for small target motion, as revealed recently, is coming from a class of
specific neurons called small target motion detectors (STMDs). Although a few
STMD-based models have been proposed, these existing models only use motion
information for small target detection and cannot discriminate small targets
from small-target-like background features (named as fake features). To address
this problem, this paper proposes a novel visual system model (STMD+) for small
target motion detection, which is composed of four subsystems -- ommatidia,
motion pathway, contrast pathway and mushroom body. Compared to existing
STMD-based models, the additional contrast pathway extracts directional
contrast from luminance signals to eliminate false positive background motion.
The directional contrast and the extracted motion information by the motion
pathway are integrated in the mushroom body for small target discrimination.
Extensive experiments showed the significant and consistent improvements of the
proposed visual system model over existing STMD-based models against fake
features. | ['Shigang Yue', 'Jigen Peng', 'Hongxin Wang', 'Xuqiang Zheng'] | 2019-04-08 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 1.98123246e-01 -3.75893950e-01 -2.33627968e-02 6.94710910e-02
3.43505055e-01 -7.46670961e-01 6.07653737e-01 -6.50043607e-01
-6.15496576e-01 6.59553707e-01 -3.72944236e-01 1.38817713e-01
5.71248889e-01 -6.18898749e-01 -4.53376472e-01 -1.04347253e+00
2.75205523e-02 -2.35216826e-01 1.21476972e+00 1.57046821e-02
5.86436629e-01 5.41116536e-01 -1.72745800e+00 3.73210102e-01
2.80918241e-01 9.42567289e-01 5.06212592e-01 7.77751446e-01
2.07341146e-02 5.38906574e-01 -9.32319462e-01 9.73736942e-02
5.38968921e-01 -7.04055667e-01 -9.74954478e-03 -2.21402012e-02
3.96925926e-01 -6.08464837e-01 -2.26190075e-01 1.35690749e+00
4.99879360e-01 -2.60993361e-01 6.68577671e-01 -1.27206469e+00
-8.20504427e-01 -9.63097159e-03 -9.16635036e-01 8.41503084e-01
9.80684608e-02 7.95881510e-01 1.86320126e-01 -7.44724452e-01
5.52206635e-01 1.61376953e+00 4.81742948e-01 1.23045492e+00
-1.09816623e+00 -6.97285712e-01 -1.10407010e-01 2.68013120e-01
-1.01625347e+00 -5.94888330e-01 7.06960499e-01 -4.01469558e-01
9.58448410e-01 3.70218545e-01 7.04327106e-01 1.24791133e+00
5.78939915e-01 6.06265724e-01 1.27931440e+00 -1.43106282e-01
1.36198580e-01 2.93929219e-01 2.04080567e-01 6.21421099e-01
9.73298073e-01 6.22283816e-01 -4.29714143e-01 5.21428213e-02
1.08171272e+00 -2.10714892e-01 -5.50083220e-01 -2.24828735e-01
-1.25269115e+00 4.68207091e-01 5.01109004e-01 3.37535530e-01
7.69756287e-02 5.79144061e-02 -7.11500421e-02 1.63309887e-01
-1.51316136e-01 5.78446463e-02 -6.53894022e-02 8.33323002e-02
-3.71897817e-01 -2.60642264e-02 8.68538678e-01 6.03807151e-01
4.23129648e-01 4.80161875e-01 -2.72137910e-01 4.83534634e-01
6.12769783e-01 1.13278186e+00 9.20722008e-01 -8.78793836e-01
2.60374218e-01 7.02899694e-01 5.77124596e-01 -1.37351751e+00
-3.37284893e-01 -1.07803404e-01 -7.55147278e-01 8.75986040e-01
7.14581132e-01 3.68684977e-02 -7.73775876e-01 1.67401552e+00
6.18885040e-01 1.18732825e-01 1.54663265e-01 1.31179464e+00
1.10710275e+00 6.99353337e-01 -2.28522331e-01 -4.37285423e-01
1.20640492e+00 -8.04460108e-01 -7.08499491e-01 -6.01491570e-01
3.09535116e-01 -6.42280519e-01 9.02891338e-01 2.87541211e-01
-7.11240530e-01 -6.85824573e-01 -1.23848295e+00 3.31779987e-01
-4.87440944e-01 4.59811598e-01 4.47825968e-01 8.98002803e-01
-7.65760124e-01 3.99292648e-01 -7.19694138e-01 -5.86913586e-01
4.05087680e-01 4.34744030e-01 -2.49658942e-01 2.62975752e-01
-5.78214884e-01 9.58261430e-01 2.27362603e-01 4.67791736e-01
-1.17087317e+00 3.45591336e-01 -5.52885532e-01 -2.35433877e-01
8.37491155e-02 -6.13045633e-01 1.02352703e+00 -1.11603963e+00
-1.63808608e+00 1.08565474e+00 -2.46197358e-01 -3.06765705e-01
3.31204206e-01 9.42114890e-02 -3.56274694e-01 2.04231218e-01
7.64451921e-03 7.08737791e-01 1.34528756e+00 -1.35610723e+00
-7.99900830e-01 -5.31058788e-01 -3.57386202e-01 2.37466961e-01
-2.15966269e-01 2.38120064e-01 -6.51581362e-02 -6.28018558e-01
4.36889790e-02 -7.18490362e-01 -2.48533804e-02 5.27823448e-01
-1.94209442e-01 4.51579969e-03 1.44400108e+00 -8.57177824e-02
1.00619853e+00 -2.16890240e+00 -2.32683033e-01 -5.58632255e-01
1.33457586e-01 8.21453154e-01 -2.55975723e-01 -1.12620667e-01
4.96304661e-01 -3.09785098e-01 -1.08667806e-01 2.10050493e-01
-3.07739139e-01 9.40007195e-02 -3.69251668e-01 6.75981879e-01
4.25510675e-01 8.29601526e-01 -8.36899161e-01 -3.59515280e-01
2.93436736e-01 2.15725258e-01 -9.09829661e-02 8.88721570e-02
4.21353355e-02 4.90650088e-01 -5.26720464e-01 1.13275754e+00
1.12331414e+00 -1.26108034e-02 -3.72121692e-01 -1.06655903e-01
-3.68899256e-01 -2.67182291e-01 -1.18223691e+00 8.64552617e-01
4.43063229e-01 7.78018773e-01 3.82212192e-01 -8.82457674e-01
1.05738318e+00 -2.35361561e-01 -1.65028110e-01 -4.88424718e-01
3.97676677e-01 2.99302697e-01 3.21727157e-01 -7.21215963e-01
1.94065645e-01 2.60293260e-02 1.49830878e-01 -1.19921258e-02
-2.02959836e-01 -7.61151388e-02 -8.30580071e-02 -2.82396883e-01
1.00991905e+00 2.99330324e-01 3.74359846e-01 -1.86816156e-01
5.65230191e-01 3.56841415e-01 8.83889914e-01 6.73467219e-01
-8.07216704e-01 3.01834553e-01 -6.32983595e-02 -6.45187080e-01
-5.71924865e-01 -9.96527672e-01 -2.46334802e-02 5.58869123e-01
9.02625382e-01 3.03180546e-01 -6.05318308e-01 -4.29491729e-01
8.80731866e-02 2.05060661e-01 -7.62495875e-01 -3.73186320e-01
-6.04610622e-01 -9.84510362e-01 5.83108306e-01 2.52010494e-01
9.01441932e-01 -1.45274770e+00 -1.61963034e+00 -2.38927118e-02
1.35871157e-01 -9.16768193e-01 -2.79306710e-01 9.59500521e-02
-7.47105598e-01 -1.13767719e+00 -8.41975331e-01 -8.43418300e-01
6.07507527e-01 1.06994748e+00 5.41006923e-01 -2.19718426e-01
-6.57058299e-01 1.63538113e-01 -2.24890724e-01 -6.40661001e-01
-1.48244739e-01 -1.05223989e+00 2.68887073e-01 1.60059839e-01
7.41633236e-01 -1.97878271e-01 -6.26722693e-01 6.58267617e-01
-6.97128236e-01 -1.22165062e-01 7.96035409e-01 8.94442677e-01
3.11823428e-01 -1.87581465e-01 6.92704925e-03 -1.45908907e-01
2.42132887e-01 -1.47942126e-01 -1.06682670e+00 3.70590314e-02
1.06533244e-01 -9.72309783e-02 6.69421673e-01 -1.19917178e+00
-1.01065612e+00 1.18896022e-01 6.30219698e-01 -4.39348757e-01
-4.41959113e-01 -4.55665469e-01 -2.67615765e-01 -7.08514333e-01
7.03132451e-01 5.66743255e-01 -6.99639786e-04 -2.00149298e-01
-1.41679466e-01 7.63428330e-01 6.52243018e-01 1.47478312e-01
8.80951583e-01 9.68812108e-01 3.80289227e-01 -1.23221493e+00
-3.68050843e-01 -3.78323704e-01 -4.36335385e-01 -4.42915976e-01
8.40012252e-01 -5.86092293e-01 -9.28667903e-01 1.07271683e+00
-1.28572631e+00 -2.27282867e-01 -2.01567680e-01 7.52179801e-01
-2.37596720e-01 5.89676976e-01 -5.91492653e-01 -1.12448132e+00
-2.19724059e-01 -1.03457093e+00 1.01088250e+00 9.38464940e-01
3.45851302e-01 -4.73662376e-01 2.33030505e-02 -1.03558689e-01
5.84286273e-01 2.21055225e-01 2.49669164e-01 -2.37632021e-01
-8.55040550e-01 -2.08931819e-01 -4.50480402e-01 7.42755011e-02
2.51005590e-01 1.57733053e-01 -1.21542358e+00 -2.36003101e-01
5.62834978e-01 -2.42388248e-01 1.17617476e+00 7.39606798e-01
3.09076607e-01 -1.41506910e-01 -7.36766517e-01 6.99683011e-01
1.27279103e+00 7.03957438e-01 3.88035208e-01 5.90404868e-01
3.77614349e-01 5.57261586e-01 8.88339758e-01 4.29565609e-01
-3.06411803e-01 6.47100210e-01 7.70423114e-01 2.14116618e-01
-3.20021421e-01 1.47054821e-01 7.39588559e-01 2.51266837e-01
-8.74321833e-02 -2.91991651e-01 -2.44933933e-01 3.29815358e-01
-1.56902242e+00 -1.20361483e+00 -5.54711521e-01 2.49217391e+00
2.70464659e-01 2.71533430e-01 5.30965291e-02 -1.44247130e-01
8.91687810e-01 -8.12143534e-02 -9.44593728e-01 2.30591418e-03
-7.46120632e-01 -4.13731247e-01 6.74214363e-01 1.47790834e-01
-1.23199880e+00 1.12309599e+00 6.54245138e+00 5.58274209e-01
-1.32059979e+00 -2.04410926e-01 1.06806353e-01 1.54583797e-01
3.56761575e-01 -8.08907673e-03 -1.43386161e+00 5.90300798e-01
3.77447486e-01 1.41545296e-01 1.63383096e-01 8.57916832e-01
1.62308961e-01 -6.24764681e-01 -7.19437540e-01 1.15114009e+00
1.14100948e-01 -7.66470969e-01 2.69632459e-01 1.52174279e-01
1.99033782e-01 -3.17198187e-01 2.43758187e-01 -6.88148588e-02
1.31076097e-01 -5.98082542e-01 5.38641751e-01 3.72998923e-01
3.40559959e-01 -2.93378569e-02 7.15445995e-01 5.36584258e-01
-1.12629998e+00 -4.07820553e-01 -1.24066913e+00 -3.32495600e-01
-7.34646916e-02 1.25611067e-01 -5.41976631e-01 -1.06243432e-01
9.80688572e-01 6.71413004e-01 -7.75857329e-01 1.35501564e+00
9.69826505e-02 3.66154015e-01 -5.75146258e-01 -6.43101633e-01
1.33433759e-01 -2.12146506e-01 1.09141374e+00 1.03100777e+00
3.83548170e-01 5.82207404e-02 -3.41741025e-01 1.07029808e+00
5.00437140e-01 -3.52255672e-01 -1.03435719e+00 -1.40988469e-01
3.58519912e-01 1.35279596e+00 -9.50506508e-01 -3.36655855e-01
-3.68935853e-01 1.31262350e+00 -1.14399269e-01 1.79934919e-01
-7.02261567e-01 -4.65735346e-01 4.57326621e-01 -1.89448327e-01
5.48862636e-01 -1.54440552e-01 4.58669886e-02 -1.32946813e+00
-1.15084976e-01 -5.03654063e-01 2.24425644e-01 -9.42066252e-01
-1.10219026e+00 4.31097478e-01 -2.15349257e-01 -1.72112334e+00
1.50806755e-01 -1.17721140e+00 -6.27439022e-01 6.06243253e-01
-1.17866015e+00 -1.13191819e+00 -5.91464877e-01 7.35590160e-01
6.33918881e-01 -3.13719839e-01 5.55056870e-01 -3.22724849e-01
-5.63437760e-01 3.18525434e-01 2.40206003e-01 -1.06734904e-02
7.49776721e-01 -7.79555976e-01 2.35752270e-01 1.15276825e+00
-1.49213830e-02 3.51499885e-01 7.41505861e-01 -7.35714972e-01
-1.63668525e+00 -6.55500293e-01 2.61436850e-01 -5.81113338e-01
4.44935203e-01 -2.22059354e-01 -6.77260578e-01 2.19832987e-01
-3.10141053e-02 2.98793107e-01 2.79895097e-01 -1.09933078e+00
-1.00062147e-01 -1.76476091e-01 -1.30276084e+00 5.19847751e-01
9.69879746e-01 3.18480372e-01 -8.47951174e-01 -1.12372123e-01
3.06421906e-01 -5.82039310e-03 2.71538869e-02 3.84240687e-01
8.05989742e-01 -1.20529187e+00 9.58102226e-01 -3.59062850e-01
-2.38528103e-01 -9.37700212e-01 -2.05963373e-01 -9.48822856e-01
-4.71042067e-01 -4.86716419e-01 -1.93953272e-02 1.07428801e+00
-1.51597306e-01 -6.96283281e-01 7.12585449e-01 5.20386957e-02
4.45097499e-03 2.07241431e-01 -1.03513062e+00 -1.04893374e+00
-4.84722704e-01 3.20485651e-01 -2.33534008e-01 7.30236292e-01
-1.39019296e-01 3.30420226e-01 -5.00100970e-01 3.75112653e-01
9.85549271e-01 3.80568683e-01 8.88051152e-01 -1.21959007e+00
-3.66793513e-01 -5.68986833e-01 -7.86356568e-01 -1.28481460e+00
-5.56821942e-01 -1.43356502e-01 2.20478982e-01 -1.27534580e+00
1.77239254e-01 2.71661222e-01 -2.12559447e-01 3.11388046e-01
-1.94431677e-01 4.91479933e-01 9.14615393e-02 4.71718818e-01
-5.57297885e-01 3.65241796e-01 1.48754704e+00 -1.35989383e-01
-4.00999397e-01 2.24371597e-01 -1.94735318e-01 8.81073833e-01
6.14100635e-01 -5.39004982e-01 4.37080413e-02 -3.42103750e-01
-5.24446607e-01 -1.38847083e-01 6.25416279e-01 -1.34918094e+00
2.47325748e-01 -4.38782454e-01 8.43956530e-01 -6.38578355e-01
6.13133848e-01 -7.94555545e-01 -1.16341598e-01 1.24284208e+00
3.63818705e-01 -1.82030529e-01 4.42096859e-01 8.15820277e-01
9.59316734e-04 -2.02513829e-01 1.05987060e+00 -4.92604136e-01
-1.09337604e+00 -2.75173366e-01 -9.95202601e-01 -2.54844904e-01
1.36311054e+00 -8.48730266e-01 -9.86237705e-01 -7.89857004e-03
-3.52849990e-01 -2.00634524e-01 5.42372048e-01 2.76558280e-01
9.11479354e-01 -1.04860783e+00 -2.39926428e-01 6.71448290e-01
-6.95039034e-02 -3.38925570e-01 -1.72747634e-02 8.51651847e-01
-6.66828513e-01 5.82577527e-01 -9.16140854e-01 -7.49135494e-01
-1.51550031e+00 9.39526320e-01 3.02813411e-01 4.27089959e-01
-2.12400749e-01 1.26495075e+00 7.33196378e-01 2.54520744e-01
1.27572164e-01 -5.35479069e-01 -4.33152437e-01 -5.10866225e-01
8.58941257e-01 3.23835462e-01 -7.57780790e-01 -8.49500179e-01
-5.03592968e-01 1.08798623e+00 3.43111992e-01 1.03325672e-01
9.42992330e-01 -2.81205922e-01 -9.13196132e-02 6.44426048e-01
4.30461287e-01 1.11382473e-02 -1.46499550e+00 2.33780146e-01
-1.42325744e-01 -8.69052351e-01 -3.53207618e-01 -6.15878522e-01
-6.70792162e-01 1.27205861e+00 9.46059227e-01 3.10640216e-01
1.16063678e+00 -1.82273448e-01 3.09711844e-01 5.93698502e-01
6.40748143e-01 -7.66945541e-01 4.41247582e-01 1.57048076e-01
5.71922600e-01 -1.33039081e+00 -2.92274088e-01 -3.23992878e-01
-6.67012095e-01 1.20625854e+00 1.23778594e+00 -4.12083030e-01
2.25840434e-01 4.42797750e-01 1.31360129e-01 2.15385512e-01
-7.85757244e-01 -3.44471782e-01 1.17175050e-01 1.22248185e+00
-1.49565652e-01 -3.65698814e-01 -3.74166965e-01 5.39559960e-01
2.97618359e-01 3.80105935e-02 5.68859935e-01 1.05992472e+00
-1.23032606e+00 -6.12016261e-01 -8.07460010e-01 8.03580880e-02
-3.07391524e-01 2.98713833e-01 -9.89012539e-01 7.97267377e-01
2.21384376e-01 1.00385725e+00 -1.33003644e-03 -7.11878419e-01
1.40232310e-01 -5.47837734e-01 7.13209510e-01 -3.17173839e-01
-1.90138116e-01 3.81395251e-01 -3.03703368e-01 -6.58105195e-01
-6.75723493e-01 -3.00326169e-01 -1.00160456e+00 1.24410041e-01
-5.42799115e-01 -4.04283628e-02 5.00495791e-01 4.41796124e-01
8.67947116e-02 -1.12958603e-01 4.23591346e-01 -1.04350615e+00
-6.46011710e-01 -1.09108734e+00 -8.18975151e-01 2.46789873e-01
4.57085878e-01 -9.09144044e-01 -8.46272469e-01 8.70318189e-02] | [8.371395111083984, -0.8705784678459167] |
c12137fe-63d5-4fec-a719-9014fccc5f82 | leveraging-automated-unit-tests-for-1 | 2110.06773 | null | https://arxiv.org/abs/2110.06773v2 | https://arxiv.org/pdf/2110.06773v2.pdf | Leveraging Automated Unit Tests for Unsupervised Code Translation | With little to no parallel data available for programming languages, unsupervised methods are well-suited to source code translation. However, the majority of unsupervised machine translation approaches rely on back-translation, a method developed in the context of natural language translation and one that inherently involves training on noisy inputs. Unfortunately, source code is highly sensitive to small changes; a single token can result in compilation failures or erroneous programs, unlike natural languages where small inaccuracies may not change the meaning of a sentence. To address this issue, we propose to leverage an automated unit-testing system to filter out invalid translations, thereby creating a fully tested parallel corpus. We found that fine-tuning an unsupervised model with this filtered data set significantly reduces the noise in the translations so-generated, comfortably outperforming the state-of-the-art for all language pairs studied. In particular, for Java $\to$ Python and Python $\to$ C++ we outperform the best previous methods by more than 16% and 24% respectively, reducing the error rate by more than 35%. | ['Guillaume Lample', 'Gabriel Synnaeve', 'Mark Harman', 'Francois Charton', 'Jie M. Zhang', 'Baptiste Roziere'] | 2021-10-13 | leveraging-automated-unit-tests-for | https://openreview.net/forum?id=cmt-6KtR4c4 | https://openreview.net/pdf?id=cmt-6KtR4c4 | iclr-2022-4 | ['code-translation', 'unsupervised-machine-translation'] | ['computer-code', 'natural-language-processing'] | [ 0.45856282 0.13274637 -0.03413518 -0.4566904 -1.2525196 -0.9712853
0.425884 0.5536622 -0.39570454 0.85381716 0.05556704 -0.6187658
0.2997037 -0.8370475 -1.0326649 -0.15900418 0.14223619 0.36781517
0.15271646 -0.35426217 0.44076338 -0.18916741 -1.5086565 0.4781607
1.244189 0.11044304 -0.07211912 0.47143745 -0.47748885 0.4764696
-0.7586156 -0.68125427 0.310288 -0.72423375 -0.92331034 -0.28467947
0.28520763 -0.07302513 0.5581583 1.1663568 0.4012799 -0.35041568
0.15177605 -0.8745797 -0.496925 0.98054665 -0.32568178 -0.08609191
0.80426335 0.1192425 0.7973725 -0.9317373 0.86541235 0.8576316
0.84393847 0.5224911 -1.6297668 -0.39085913 -0.43761152 -0.4526573
-1.2438651 -0.4956886 0.47826928 -0.6813069 1.3467311 0.33746004
0.23885366 1.0026067 0.23081167 0.24052282 1.2709314 -0.8872622
0.14301717 0.34992597 0.02018832 0.6349845 0.3028451 -0.04980332
-0.42953357 -0.5625929 0.04663954 -0.2800393 0.02242416 -0.01872972
-1.4793892 0.6865406 -0.23948547 0.4047215 -0.02594188 0.01262431
0.8100585 0.80208856 0.46592593 0.7473185 -0.6335988 -0.7339019
-1.1702001 0.21686105 1.0694383 1.0851558 1.0943848 -0.1018111
0.03485069 0.7890852 -0.03529462 0.52206033 0.5411178 -0.60315233
0.7916497 0.89616483 0.08327634 -0.73132706 0.01759602 -0.24799272
-0.2391191 0.20348404 0.4271774 -0.10063852 -0.70737153 1.7254629
0.21449079 -0.68130964 0.10848897 0.48349908 0.3563831 0.27804556
-0.20483191 -0.40280226 0.83449614 -0.8206059 -0.32281563 -0.29973316
0.9869578 -1.3896841 1.5714529 0.2007166 -0.99560535 -0.38508514
-1.1366428 0.1272665 -0.24272929 -0.16232373 0.49823922 0.9517145
-1.0424688 0.87851053 -0.9897454 -0.62976974 0.05920199 0.4743721
-0.50202036 -0.09318823 -0.76992995 0.8596601 0.33815005 -0.31648427
-0.55565923 -0.7405053 -0.7223243 -0.21296637 0.31563064 -0.46032348
1.3340209 -1.3637247 -1.683631 0.9186092 -0.24963371 -0.30699748
0.6488684 -0.29585084 -0.31751937 -0.3000517 0.5656092 0.32681006
0.53482485 -1.0477235 -0.24697106 -0.16089788 -0.08522113 -0.273634
-0.29602996 0.4499968 -0.2680275 -0.5576696 -0.13929927 -0.98564863
-0.19342032 -0.44717684 -0.20970672 0.10484711 0.41612294 -0.83416516
1.2393734 -2.0781205 0.07921921 0.1380683 -0.15339738 0.17479202
-0.30937237 0.8591116 0.06385227 0.3822693 -0.81540316 -0.17431769
-0.04637375 0.32345036 -0.3156035 0.26547796 0.5273095 0.6858836
-1.0243624 -0.39177445 -0.27932736 0.10776804 -0.8358109 0.15687546
-0.39760548 0.34828007 -0.1886163 0.67882437 0.5630292 0.08814298
0.36827493 0.5453583 -0.40841207 0.8742398 -1.0477244 1.9893005
-0.67299324 0.51119846 -0.34655997 -0.8106361 1.1731769 0.28678843
0.22107732 -0.7435169 -0.21927364 0.81528986 0.17264462 -0.5185762
0.5602431 -0.02587593 -0.25074506 0.7174437 -0.01902981 -0.3375789
0.63800406 0.11081225 1.3476906 0.49986812 0.18006237 -0.4180435
0.4668804 0.46261692 0.7441761 0.7323213 0.16719913 0.7496426
0.68197376 -0.30005616 -1.3420271 -0.8055834 0.10381368 0.9212938
-0.30612764 -0.80093324 -1.0266588 -0.66365016 -0.21574573 0.884545
-0.37263855 -0.09794105 -0.7714866 -0.758671 0.8279003 0.25459063
0.16134371 -1.0311083 -0.61995107 0.5103574 -0.30235434 -0.88957137
-0.33736917 0.29416054 -1.0414369 -0.96340084 -0.26141724 -0.70274985
0.7914244 -0.23002808 1.5072751 0.2995138 -0.12648658 -0.08417854
-0.5017174 -0.17283496 -1.2188476 0.32627577 -0.04729924 -0.41087848
0.47812963 -0.6868956 -0.09839682 0.18018976 -0.9843476 -0.04949525
0.63095105 0.95210016 0.36923838 -0.41186583 0.4648925 -1.3119166
0.70721906 -0.29683807 -0.6151466 0.16027279 -0.94548297 0.2906843
1.0578377 -0.43437508 -0.7807133 0.11959163 -0.18618068 0.23196888
-0.12449877 0.80085343 0.0779754 -0.03841391 1.0863746 0.30381954
-0.02786782 -0.51225615 0.23236473 0.7081822 0.37767792 -0.76733124
0.905452 0.12559767 -0.41743606 -0.3977379 -0.13285807 -0.23169495
-0.64235514 0.13583283 0.3608497 -0.67275745 0.05560122 0.10574117
-1.149489 -0.3780693 -0.22415103 0.29505163 -0.49484622 0.5778711
-0.46321708 -0.39288068 -0.45724446 -1.4484292 1.0054609 -0.19135195
-0.63787544 -0.5576905 0.55717736 0.2243307 0.6108462 0.33693904
1.1781858 -0.5786541 -0.2894375 -0.3098446 0.07345457 0.39497617
0.1770709 0.44320005 -0.5750651 -0.27936965 -0.14963965 -0.27641675
0.2537246 -0.50973815 0.37983495 -0.23674795 -0.05176542 0.18889117
1.5448512 -0.18354726 0.7476174 0.65424764 0.35848886 0.38243532
0.4965989 0.2719956 0.11858153 0.59352547 0.11566447 0.07146242
0.01083954 -0.30295005 0.7190868 1.1959308 0.11869384 0.19974613
-1.2189598 0.89529145 -1.638764 -0.6368643 -0.37220877 2.4592671
1.380077 0.2981868 0.01237899 0.10295913 0.4544045 -0.2598578
-0.02942749 -0.7658343 0.01481845 0.6581745 0.48323968 0.4135114
-0.63732874 0.95239913 6.3024135 0.3382658 -1.1802864 0.22929281
0.11871542 0.16756533 -0.7342519 0.6317436 -0.49495047 0.5723349
1.2460738 -0.242591 0.5270869 0.95910484 0.33637002 -0.20131655
-1.2720249 0.54926944 0.02523464 -1.091842 -0.16597418 -0.2044175
0.9964495 0.35761264 -0.5082145 0.35724187 0.32904983 -0.84499896
0.86731565 0.2443292 0.7309826 -0.64375675 0.80441374 0.47542596
-0.6904931 0.4263723 -0.36693123 -0.274564 -0.10992712 0.92168707
-0.82856935 0.60619235 0.6764901 0.4308376 -0.8186888 0.772155
-0.3513708 0.8834939 -0.4453127 -0.08359285 -0.07220036 -0.16411878
0.5026779 1.4819021 0.4301886 -0.506919 0.2273992 1.0130687
-0.13759185 0.50904083 -0.66645604 -0.30405262 0.13655972 0.89186215
-0.711102 -0.34300536 -0.6092098 1.109926 0.36846745 0.21038626
-0.75354344 -0.79889256 0.4943343 0.09958868 0.26863232 -0.37662807
-0.47375938 -1.3267621 0.64779955 -1.4328188 -0.05570583 -0.24564634
-0.9538134 0.7970619 -0.35885507 -1.2620186 -0.50844514 -0.20655704
-0.50037855 0.8852643 -1.2316704 -0.7988682 0.03652224 0.07917451
0.3688404 -0.03199012 1.1628498 0.5103971 -0.26118597 0.7901149
0.38269702 0.12946682 1.0461551 -1.0872127 0.77527225 1.3231047
0.19610326 1.152122 0.8557989 -0.69182867 -1.6602131 -1.0796894
1.3716291 -0.47070277 0.7802183 -0.6092291 -0.9664206 0.44136864
0.33438468 -0.12541102 0.73217994 0.0533209 -0.6879075 0.01374696
-1.0543938 0.6425326 0.85722756 -0.75722325 -0.61350024 0.43496618
0.68395656 -0.4449128 -0.976776 0.2010254 0.4088336 -1.0344105
0.38234267 -0.5477463 0.8292632 -0.52654713 -0.304721 -1.151002
0.18096995 -0.86697316 0.41693875 1.6356854 0.89251417 -0.691528
0.57116574 0.56213874 -0.1105064 -0.2710188 -0.7200769 -0.85595775
0.32800058 -0.5416781 0.5339062 1.0432236 0.38601705 0.13649316
-0.04934313 -0.09299526 0.39281133 0.32882556 1.1695831 -0.6866315
-0.67713773 -0.27770606 -0.25545195 -0.54984313 0.09279146 -1.0818498
0.2756751 -1.1787126 0.3981366 -0.34837416 0.23037583 0.616985
-0.30399597 0.52723086 0.01725986 0.328256 -0.21604373 0.12089888
0.53221023 -0.16659968 -0.24243353 -0.13594909 -0.66487396 0.41741067
0.809439 -1.0168463 -0.06620485 -0.7752389 0.62235653 -0.21242836
-0.05304514 -0.84918255 -0.0637245 -0.09964047 -0.3076219 -0.1288719
-0.52397513 -0.5401305 0.3786832 0.48698878 -0.4062476 0.41962254
0.3354763 0.2126353 -0.36761916 -0.42407316 0.60070795 -0.2950507
-0.38916272 -0.397075 -0.6089398 0.2638852 0.6198794 -0.11784557
-0.20456806 0.13790786 -0.23756185 -0.29053333 1.0422072 0.5152901
0.09775628 -1.0213882 -0.7554527 0.32457578 0.45590037 -0.15609929
-0.4848395 0.8317211 -0.7878503 0.19437984 -0.0550892 -0.64580715
-1.1172427 0.3752909 0.06761861 -0.22116923 -0.37772444 0.4937661
-0.60687345 -1.0648384 -0.26867744 -0.439833 0.45695668 -0.44741985
0.2355285 0.10625357 0.669278 -0.46316573 -0.5594206 0.38845748
-0.09304703 -0.26018685 1.3415303 0.18398453 -0.7524626 0.45711288
1.2058089 0.5322412 -0.6526671 -0.08986297 0.4247984 -0.502266
-0.49060634 -0.93621004 -0.5052803 0.58063066 0.3372232 0.08036341
0.8347512 -0.2804502 0.60098636 0.5165157 0.7290169 -1.0401882
-0.33863425 0.6035795 0.5314317 -1.4016721 -0.06705674 -0.4092999
-0.19125015 1.249678 0.29734406 -0.07585149 0.08844554 0.62814665
0.3462427 0.1998511 -0.6213914 0.18154258 -0.10076658 0.48354712
0.88800174 0.03789685 -0.93054026 0.30986947 -0.349067 0.21504338
0.81141615 1.4578309 -0.11710497 -1.8964612 -0.50172526 0.34731397
-0.6399146 -0.50967705 -0.6739729 0.7412877 0.03574399 1.0448648
-0.25117853 -0.2921321 0.3880809 0.3981306 0.48278865 -0.9488541
-1.2003082 -0.06310995 0.25258112 -0.54054385 -0.41391432 -0.7936495
-1.2802488 -0.42346263 -0.41150403 0.31472337 0.74994683 0.98604393
0.56930196 0.21603043 0.4852259 -0.35781097 -0.7256188 -0.94141746
0.09994888 0.58864814 -0.01141058 -0.19344713 -0.25764468 0.45908636] | [7.7479047775268555, 7.854860305786133] |
e6f4b2b4-4a65-4299-9a47-f88a18d09e85 | deep-supervised-and-convolutional-generative | 1403.1347 | null | http://arxiv.org/abs/1403.1347v1 | http://arxiv.org/pdf/1403.1347v1.pdf | Deep Supervised and Convolutional Generative Stochastic Network for Protein Secondary Structure Prediction | Predicting protein secondary structure is a fundamental problem in protein
structure prediction. Here we present a new supervised generative stochastic
network (GSN) based method to predict local secondary structure with deep
hierarchical representations. GSN is a recently proposed deep learning
technique (Bengio & Thibodeau-Laufer, 2013) to globally train deep generative
model. We present the supervised extension of GSN, which learns a Markov chain
to sample from a conditional distribution, and applied it to protein structure
prediction. To scale the model to full-sized, high-dimensional data, like
protein sequences with hundreds of amino acids, we introduce a convolutional
architecture, which allows efficient learning across multiple layers of
hierarchical representations. Our architecture uniquely focuses on predicting
structured low-level labels informed with both low and high-level
representations learned by the model. In our application this corresponds to
labeling the secondary structure state of each amino-acid residue. We trained
and tested the model on separate sets of non-homologous proteins sharing less
than 30% sequence identity. Our model achieves 66.4% Q8 accuracy on the CB513
dataset, better than the previously reported best performance 64.9% (Wang et
al., 2011) for this challenging secondary structure prediction problem. | ['Olga G. Troyanskaya', 'Jian Zhou'] | 2014-03-06 | null | null | null | null | ['protein-secondary-structure-prediction'] | ['medical'] | [ 5.08171737e-01 4.34198439e-01 1.12574033e-01 -5.21470666e-01
-1.05965865e+00 -5.47715664e-01 3.38185906e-01 1.33204788e-01
-2.26478055e-01 1.16164517e+00 2.05293536e-01 -5.37473500e-01
2.82129079e-01 -6.43757582e-01 -1.17341256e+00 -1.02549088e+00
-1.80717900e-01 9.33429956e-01 1.23450682e-01 -3.95459570e-02
3.50335948e-02 5.08564830e-01 -1.19496906e+00 8.57626855e-01
8.97621572e-01 6.27155960e-01 7.41329074e-01 8.64759505e-01
3.34115811e-02 1.03545177e+00 -5.12934327e-01 -1.12615444e-01
-2.00930536e-01 -8.03345323e-01 -1.20584273e+00 -5.37910983e-02
2.25496724e-01 -8.63916948e-02 1.17996663e-01 5.98895192e-01
4.86737251e-01 -1.46938637e-01 9.86495733e-01 -5.64082265e-01
-7.92684078e-01 5.04469812e-01 -3.38471562e-01 6.14009015e-02
1.29132763e-01 2.61617690e-01 1.28042424e+00 -7.06402123e-01
1.04987037e+00 1.21431828e+00 7.77701914e-01 6.82731986e-01
-1.93003619e+00 -2.52576053e-01 -1.77797601e-01 1.59662262e-01
-1.09554148e+00 7.23477677e-02 2.30330855e-01 -6.20807290e-01
1.60508657e+00 -3.34094256e-01 4.62746322e-01 1.37192369e+00
7.65979886e-01 7.29209483e-01 1.12704396e+00 -1.93663597e-01
3.49291593e-01 -7.92295992e-01 -1.00191766e-02 8.24725389e-01
-2.18625501e-01 -9.33163837e-02 -3.60349774e-01 -7.27886260e-01
5.30630291e-01 1.53584182e-01 -2.75737252e-02 -5.60910642e-01
-1.05266750e+00 1.14759231e+00 6.35660172e-01 -6.45182729e-02
-6.96322024e-01 8.61483887e-02 1.74104884e-01 2.35748757e-02
3.23116124e-01 2.95482755e-01 -9.49456275e-01 -1.25498861e-01
-8.42496514e-01 2.73638040e-01 9.24414158e-01 6.32454455e-01
9.29759562e-01 -2.35546902e-01 -1.40979849e-02 5.52862465e-01
4.56273794e-01 6.09977096e-02 5.59416234e-01 -6.14801109e-01
-1.06450945e-01 3.96227330e-01 2.25484297e-02 -1.16845660e-01
-5.62230706e-01 -3.00305367e-01 -8.95085990e-01 2.40809157e-01
3.84990335e-01 -2.86648303e-01 -1.14138377e+00 1.86309969e+00
1.52967781e-01 1.34050071e-01 3.36975098e-01 6.12489402e-01
6.76154256e-01 9.84446883e-01 5.87762296e-01 -2.15239838e-01
1.19099331e+00 -6.77621424e-01 -1.51873961e-01 1.13314487e-01
6.77326322e-01 -7.66524911e-01 8.15639615e-01 3.87008756e-01
-7.28940070e-01 -5.27244210e-01 -9.46196795e-01 -3.29120576e-01
-1.01108551e-01 1.19251035e-01 6.18666291e-01 2.05078483e-01
-1.08841610e+00 9.78242040e-01 -1.11656082e+00 -4.77616966e-01
4.42022324e-01 5.38107753e-01 -2.68169761e-01 1.53733715e-01
-1.19366527e+00 8.61455739e-01 7.43153512e-01 -3.09057653e-01
-1.32659304e+00 -6.04323328e-01 -5.05656481e-01 2.20420137e-01
-1.61077350e-01 -7.85934031e-01 1.33590651e+00 -8.47692728e-01
-1.53290248e+00 9.74105656e-01 -4.93310869e-01 -8.98689628e-01
-1.00675195e-01 -2.33581111e-01 9.68585610e-02 7.52406269e-02
8.34207144e-03 8.89057875e-01 3.82001579e-01 -1.02698123e+00
-2.90405631e-01 -3.19876075e-01 -2.85724789e-01 -1.77915357e-02
5.38770795e-01 -1.03623442e-01 4.24371272e-01 -5.03123522e-01
9.90270451e-02 -9.57474113e-01 -6.55401886e-01 -4.04114127e-01
-3.20781708e-01 -4.31657284e-01 5.60078263e-01 -8.07723224e-01
6.16562605e-01 -1.66267383e+00 4.64789480e-01 2.76160315e-02
2.98847258e-01 3.67430955e-01 -1.69001058e-01 9.35059309e-01
-4.25006568e-01 -2.62889981e-01 -6.28679574e-01 3.75807062e-02
-7.62652010e-02 3.98778707e-01 -2.71277159e-01 3.13231260e-01
5.23785949e-01 1.20466530e+00 -6.85969710e-01 -5.11361882e-02
-1.16395719e-01 8.80459428e-01 -6.56965971e-01 4.36619222e-01
-7.80498207e-01 5.06142676e-01 -2.37778544e-01 3.54016960e-01
5.86672008e-01 -9.22940135e-01 8.82865429e-01 -3.84287685e-02
1.89350560e-01 7.76121140e-01 -5.51875830e-01 1.72031641e+00
-2.63543278e-01 2.31363699e-01 -3.08201939e-01 -9.35748935e-01
1.07777953e+00 2.53223240e-01 5.06589890e-01 -1.26426026e-01
-1.27417296e-01 7.32179210e-02 2.76417971e-01 -6.86930642e-02
-6.86793402e-02 -3.66977751e-01 1.37484640e-01 5.48264682e-01
7.22139418e-01 2.82550186e-01 4.61610891e-02 2.85760760e-01
1.30348587e+00 7.81730831e-01 7.39357829e-01 -4.95597571e-01
3.12826097e-01 -6.56422973e-02 6.60651684e-01 4.94783968e-01
-3.18744369e-02 6.97341144e-01 9.15325999e-01 -7.50706196e-01
-1.36075103e+00 -1.19727218e+00 -9.68242586e-02 1.56572235e+00
-4.84496325e-01 -7.03812540e-01 -1.00677407e+00 -8.10208261e-01
-2.37159476e-01 3.71399969e-01 -6.15044892e-01 -2.91315705e-01
-5.71487844e-01 -1.29167867e+00 2.71496922e-01 4.70779240e-01
8.49244073e-02 -1.66810787e+00 -5.31182110e-01 7.29243398e-01
-1.68618843e-01 -6.14570141e-01 -2.02057406e-01 1.04513705e+00
-9.81594801e-01 -1.20551884e+00 -6.39078915e-01 -1.11410427e+00
4.26769435e-01 -3.86501700e-01 1.53186572e+00 -2.51792014e-01
-3.85558426e-01 -4.09663498e-01 -2.31320724e-01 -2.11205691e-01
-8.66921127e-01 4.94292825e-01 -1.66024104e-01 -2.00198472e-01
6.93944752e-01 -8.16410899e-01 -6.36043608e-01 -3.00866440e-02
-9.03194249e-01 2.23822132e-01 9.21421766e-01 1.16936862e+00
1.07126057e+00 -5.77865601e-01 5.24882257e-01 -1.17469096e+00
3.73388052e-01 -5.04386663e-01 -7.30595648e-01 2.81746626e-01
-3.93395096e-01 7.44195759e-01 8.17368984e-01 -2.19519794e-01
-9.73579884e-01 8.47531438e-01 -7.10090458e-01 1.89443976e-01
-4.92063642e-01 3.76682192e-01 -2.23687455e-01 2.16306046e-01
8.59153628e-01 5.97953618e-01 3.04712821e-02 -7.19458640e-01
4.11055773e-01 4.64686543e-01 2.23121852e-01 -5.68171859e-01
2.24966660e-01 2.70266950e-01 2.63831168e-01 -5.26749372e-01
-9.72999632e-01 -2.58879483e-01 -9.70529616e-01 5.37302434e-01
1.03476429e+00 -8.77709270e-01 -1.03324103e+00 4.18859988e-01
-1.16343570e+00 -6.68956280e-01 -1.25979763e-02 2.27619752e-01
-1.19450450e+00 5.07846117e-01 -1.04749763e+00 -3.51532161e-01
-5.71952164e-01 -1.16337538e+00 1.36352956e+00 -1.37745157e-01
-3.06664914e-01 -7.44371295e-01 6.21424019e-01 8.13681781e-02
9.79163870e-02 3.74953598e-01 1.20325792e+00 -7.89910078e-01
-5.25346696e-01 5.38579464e-01 1.86563935e-02 2.14242101e-01
1.45194456e-01 -8.87642503e-02 -8.20250928e-01 -3.80145401e-01
-1.39118269e-01 -8.00150156e-01 1.32452118e+00 5.06295681e-01
1.13455880e+00 -2.27460206e-01 -2.99843460e-01 6.79530680e-01
1.42142260e+00 2.13242382e-01 6.96159124e-01 3.72520804e-01
6.44739449e-01 3.82117540e-01 2.79677480e-01 3.95082980e-01
-2.53270892e-03 4.55990881e-01 5.20298183e-01 -1.54979602e-01
8.54540765e-02 -4.67967927e-01 4.46096659e-01 3.93234551e-01
-5.05685359e-02 -3.30212265e-01 -9.74771440e-01 8.74383301e-02
-1.87895370e+00 -1.04845703e+00 -7.36976936e-02 1.68112695e+00
1.25668955e+00 5.55915292e-03 2.52870619e-01 -3.93146664e-01
5.78877389e-01 -3.00174039e-02 -8.64893854e-01 -4.44100708e-01
-2.57225394e-01 7.12965190e-01 3.02239746e-01 5.30156791e-01
-1.20107913e+00 1.17981601e+00 6.84995508e+00 8.03154469e-01
-8.88910592e-01 -1.81796432e-01 8.71764660e-01 1.24647781e-01
-1.10915877e-01 1.19682208e-01 -9.87263978e-01 5.12500167e-01
1.53076708e+00 2.64263004e-01 2.35356823e-01 9.94931519e-01
-2.15410385e-02 2.87203670e-01 -9.51158881e-01 5.30991375e-01
-2.05267265e-01 -1.65735292e+00 1.48151428e-01 1.61741227e-01
8.15227270e-01 5.67302704e-01 -9.17073116e-02 1.49118707e-01
1.06301582e+00 -1.38310444e+00 2.02058395e-03 4.70320314e-01
5.79160094e-01 -1.08049357e+00 6.13290131e-01 5.89694619e-01
-8.66178155e-01 3.37847859e-01 -7.90579259e-01 1.12403728e-01
1.51255146e-01 6.57503843e-01 -1.22548890e+00 1.54841080e-01
5.73427200e-01 7.57941723e-01 -3.94697011e-01 3.61646712e-01
-3.85956734e-01 9.32336986e-01 -1.40315041e-01 -1.03909843e-01
3.13772827e-01 -4.12350506e-01 -1.46607682e-01 1.22131121e+00
-2.02114023e-02 2.05241218e-01 2.72706628e-01 8.92433107e-01
-2.00420335e-01 7.29653761e-02 -1.99224323e-01 -5.46108596e-02
-1.68760613e-01 1.07028103e+00 -6.44986451e-01 -3.23009044e-01
-2.43501112e-01 1.09801686e+00 9.50945079e-01 3.50896478e-01
-7.62989521e-01 -3.12170327e-01 9.49246049e-01 3.41552906e-02
8.20879042e-01 -1.87866747e-01 3.69734675e-01 -9.50477839e-01
-4.96305287e-01 -1.04745936e+00 3.11110318e-01 -5.82517743e-01
-1.50229716e+00 7.44712830e-01 -5.35384417e-01 -6.24515414e-01
-5.61561108e-01 -8.18873227e-01 -2.25240454e-01 1.12805283e+00
-1.18266249e+00 -1.20942938e+00 3.63315880e-01 1.46146089e-01
3.68776500e-01 -2.31252894e-01 1.20977902e+00 -2.77850300e-01
-2.25831479e-01 1.60782859e-01 7.55626023e-01 -3.26405913e-02
6.06102347e-01 -1.58570027e+00 1.08364248e+00 5.04135966e-01
2.57008523e-01 8.25420201e-01 7.06989825e-01 -7.12634742e-01
-8.46433699e-01 -1.33892012e+00 1.20856357e+00 -6.13147855e-01
3.88571560e-01 -7.38049269e-01 -1.31914234e+00 9.36221361e-01
2.91435272e-01 -2.43267670e-01 1.29656982e+00 3.95891853e-02
-3.17500472e-01 5.00448704e-01 -1.06601667e+00 2.01898739e-01
1.05330455e+00 -5.26896775e-01 -6.35192037e-01 5.87451160e-01
7.82524109e-01 -2.58458585e-01 -1.01660275e+00 4.26990330e-01
3.72364402e-01 -1.25050282e+00 8.84470105e-01 -1.23258746e+00
5.95958591e-01 -2.55728692e-01 -1.03820615e-01 -1.09456861e+00
-8.30499113e-01 -6.18927360e-01 -2.43541673e-01 6.24666810e-01
6.45110369e-01 -3.30855459e-01 1.13962853e+00 -1.48545206e-01
-2.14007393e-01 -1.08407104e+00 -6.18718624e-01 -4.78382707e-01
4.71140891e-01 2.38397196e-01 5.01686573e-01 4.18985248e-01
-1.13684453e-01 7.46125638e-01 -3.88774663e-01 -4.96409200e-02
4.35114831e-01 5.74318171e-01 5.15894234e-01 -1.24407959e+00
-7.29662120e-01 5.85091971e-02 -3.26485604e-01 -1.47034180e+00
5.28714001e-01 -1.10345328e+00 2.07275271e-01 -1.54683459e+00
6.42750859e-01 3.68526042e-01 -4.93809789e-01 4.97969657e-01
-2.34576002e-01 1.79280534e-01 -2.40899831e-01 1.80233791e-01
-7.06787825e-01 7.15265691e-01 9.77021635e-01 1.53951366e-02
9.83969495e-02 9.43360031e-02 -5.00821650e-01 4.39467460e-01
9.61029708e-01 -5.93247175e-01 -2.92183876e-01 1.97195172e-01
1.64954334e-01 -1.44615307e-01 2.13793635e-01 -7.72780180e-01
-3.89397621e-01 -9.53730494e-02 8.33923697e-01 -8.41319025e-01
1.42487645e-01 -3.48470271e-01 2.85044521e-01 9.53674138e-01
-5.74474216e-01 1.62438080e-02 -7.97804892e-02 5.95369935e-01
-6.35624900e-02 8.41382667e-02 1.10391283e+00 -3.30701023e-01
-5.10571241e-01 4.41636950e-01 -8.30973387e-01 -1.58134863e-01
8.45223069e-01 2.35095203e-01 -7.31382817e-02 -1.04828171e-01
-1.24127293e+00 -2.23163515e-01 5.93641698e-01 5.91416378e-03
5.33157945e-01 -9.52309549e-01 -7.60511994e-01 2.49883771e-01
1.04304008e-01 -1.25120997e-01 8.14683512e-02 2.54980177e-01
-9.15947795e-01 7.05666661e-01 -5.00490189e-01 -6.93726718e-01
-1.15918231e+00 7.35826433e-01 4.08090174e-01 -5.44978499e-01
-4.50952888e-01 1.00100625e+00 5.39148510e-01 -6.12855136e-01
-2.86955863e-01 -3.38845104e-01 -1.20473899e-01 -2.95364052e-01
2.65469402e-01 -1.19935513e-01 9.83585715e-02 -7.53138125e-01
-2.90593684e-01 2.37996876e-01 -4.78936255e-01 3.85047644e-01
1.49136496e+00 1.90559447e-01 -2.70377129e-01 3.62465709e-01
1.42920661e+00 -6.80525541e-01 -1.77086735e+00 -1.91085145e-01
1.65740594e-01 3.74838710e-01 -5.94841897e-01 -8.84661376e-01
-4.58489150e-01 9.02002990e-01 7.47764945e-01 -2.94860482e-01
7.64471829e-01 4.61161643e-01 8.76969874e-01 6.88962221e-01
4.51398134e-01 -6.07831478e-01 -7.51440600e-02 8.71986628e-01
5.68267345e-01 -1.38634193e+00 -2.41770014e-01 4.11729924e-02
-5.53147376e-01 1.00837576e+00 4.09350753e-01 -2.29828671e-01
3.91928166e-01 2.25760639e-01 1.26797855e-01 -1.82044372e-01
-1.27379465e+00 -2.61262208e-01 7.21454099e-02 8.27040792e-01
1.09046519e+00 1.83602706e-01 -3.08413714e-01 4.19144601e-01
-1.95934609e-01 -1.32512793e-01 1.73041210e-01 8.77498746e-01
-7.40666330e-01 -1.44946063e+00 6.94559291e-02 4.22350287e-01
-6.91908121e-01 -4.28060710e-01 -7.59969890e-01 2.45469525e-01
5.87375984e-02 5.25639296e-01 -1.42678380e-01 -5.87711297e-02
-1.58638835e-01 4.35304821e-01 4.96024042e-01 -8.98725927e-01
-5.62276304e-01 1.72765687e-01 -2.95694858e-01 -6.24317467e-01
-4.49049711e-01 -6.53561711e-01 -1.57784498e+00 -9.97091979e-02
1.18498288e-01 4.05345887e-01 2.88062215e-01 7.18487084e-01
7.31488347e-01 4.11211431e-01 2.95057535e-01 -7.67362118e-01
-7.40546823e-01 -1.16820598e+00 -6.04230583e-01 4.81624275e-01
3.20105255e-01 -3.72447222e-01 -5.47074107e-03 4.21929926e-01] | [4.689384460449219, 5.639500617980957] |
9d5d42c2-9027-4e64-9904-5b00599e1629 | temporal-action-segmentation-from-timestamp | 2103.06669 | null | https://arxiv.org/abs/2103.06669v3 | https://arxiv.org/pdf/2103.06669v3.pdf | Temporal Action Segmentation from Timestamp Supervision | Temporal action segmentation approaches have been very successful recently. However, annotating videos with frame-wise labels to train such models is very expensive and time consuming. While weakly supervised methods trained using only ordered action lists require less annotation effort, the performance is still worse than fully supervised approaches. In this paper, we propose to use timestamp supervision for the temporal action segmentation task. Timestamps require a comparable annotation effort to weakly supervised approaches, and yet provide a more supervisory signal. To demonstrate the effectiveness of timestamp supervision, we propose an approach to train a segmentation model using only timestamps annotations. Our approach uses the model output and the annotated timestamps to generate frame-wise labels by detecting the action changes. We further introduce a confidence loss that forces the predicted probabilities to monotonically decrease as the distance to the timestamps increases. This ensures that all and not only the most distinctive frames of an action are learned during training. The evaluation on four datasets shows that models trained with timestamps annotations achieve comparable performance to the fully supervised approaches. | ['Juergen Gall', 'Yazan Abu Farha', 'Zhe Li'] | 2021-03-11 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Li_Temporal_Action_Segmentation_From_Timestamp_Supervision_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Li_Temporal_Action_Segmentation_From_Timestamp_Supervision_CVPR_2021_paper.pdf | cvpr-2021-1 | ['weakly-supervised-action-localization'] | ['computer-vision'] | [ 5.86333334e-01 1.90733105e-01 -5.52100480e-01 -9.01324391e-01
-8.09475183e-01 -6.09057307e-01 7.27521122e-01 1.49218202e-01
-6.60929203e-01 6.80428147e-01 1.48154169e-01 6.55866712e-02
3.15196544e-01 -2.67490536e-01 -8.60375106e-01 -5.77779174e-01
-2.94456512e-01 4.46608841e-01 8.41068983e-01 3.71032119e-01
1.60669193e-01 5.20830154e-02 -1.51780891e+00 5.37115991e-01
4.59387481e-01 1.02242315e+00 -1.11315146e-01 7.18860686e-01
1.53767973e-01 1.30634284e+00 -6.75881386e-01 -2.49503423e-02
4.34957772e-01 -7.61327922e-01 -1.06367719e+00 2.69491732e-01
5.01659214e-01 -5.46557307e-01 -1.43044278e-01 6.26476109e-01
-1.43749183e-02 2.06292868e-01 2.73973376e-01 -1.48092532e+00
9.92183611e-02 7.21559525e-01 -4.86277163e-01 2.46186629e-01
5.89752674e-01 2.55724583e-02 1.04446292e+00 -3.29113483e-01
9.59642291e-01 9.16855276e-01 7.64204800e-01 5.57417035e-01
-1.06093323e+00 -5.15998304e-01 4.67385292e-01 2.20823810e-01
-8.93638551e-01 -2.84008801e-01 5.74015558e-01 -3.95631164e-01
7.86982238e-01 5.40571921e-02 6.30132735e-01 1.05410957e+00
-1.89868256e-01 1.13581228e+00 1.03313291e+00 -4.27297950e-01
3.18637818e-01 2.77699088e-03 4.96354848e-02 6.62290931e-01
-2.56411642e-01 -1.17766909e-01 -7.53733218e-01 -8.34216326e-02
6.35157883e-01 4.25443277e-02 3.16813029e-02 -2.40726367e-01
-1.49015450e+00 6.05228901e-01 2.15081945e-02 4.22334105e-01
-2.16986343e-01 5.51642954e-01 7.21189797e-01 8.36265162e-02
5.76586008e-01 2.53064454e-01 -5.36000073e-01 -6.64823949e-01
-1.41275680e+00 6.73126476e-03 6.02729857e-01 9.16247308e-01
7.04943657e-01 -2.45050251e-01 -2.40139112e-01 3.13875765e-01
1.13012262e-01 -9.65444520e-02 3.95157158e-01 -1.60266054e+00
3.05060744e-01 5.74835181e-01 3.77082467e-01 -5.25635958e-01
-2.53676414e-01 1.86276525e-01 -1.75521281e-02 6.11597709e-02
7.52935112e-01 -1.70793280e-01 -1.07225871e+00 1.69265068e+00
4.46100414e-01 6.44116461e-01 -1.88094869e-01 7.75254309e-01
1.99288547e-01 6.22121692e-01 4.44813073e-01 -3.78030866e-01
9.73187149e-01 -1.10947764e+00 -8.42112541e-01 -7.54947141e-02
1.01299787e+00 -5.71690917e-01 1.02315819e+00 1.79410920e-01
-1.03833902e+00 -5.83645761e-01 -9.54998255e-01 -2.07719649e-03
-1.18047759e-01 2.60220081e-01 6.47656024e-01 3.94058228e-01
-8.74816597e-01 9.93016779e-01 -1.46329653e+00 -5.00981331e-01
4.46345419e-01 3.43051642e-01 -2.99299419e-01 2.86939263e-01
-1.06794417e+00 6.24516189e-01 5.98807096e-01 -1.70939162e-01
-1.01021898e+00 -4.92125601e-01 -1.03655088e+00 -2.51215309e-01
4.52456385e-01 -5.45921251e-02 1.79775381e+00 -1.36092365e+00
-1.40438187e+00 7.63018310e-01 -2.94137716e-01 -9.99936283e-01
8.04882586e-01 -3.81820828e-01 6.58729896e-02 5.72523117e-01
4.20459479e-01 1.05668104e+00 7.41043389e-01 -9.72321510e-01
-1.23398256e+00 5.58398068e-02 3.02373141e-01 1.13511190e-01
-4.05332834e-01 3.65578353e-01 -4.95351464e-01 -4.75457579e-01
-7.74816573e-02 -1.12751639e+00 -3.23854059e-01 -5.69594242e-02
-1.92904830e-01 -4.21676844e-01 1.17620468e+00 -5.58775246e-01
1.23251915e+00 -2.27420950e+00 -2.15567142e-01 -1.38330132e-01
-2.90261120e-01 5.89128509e-02 5.49451783e-02 2.37560704e-01
-1.14828460e-02 1.10078296e-02 -3.63528043e-01 -6.95824921e-01
-1.26592726e-01 6.10043585e-01 -2.38801673e-01 4.08921838e-01
2.45727569e-01 6.04283214e-01 -1.26542974e+00 -8.66933286e-01
2.54443198e-01 1.76231176e-01 -4.33181882e-01 2.75994599e-01
-5.88316262e-01 8.02412331e-01 -3.25219125e-01 4.65529293e-01
-2.18936279e-02 -2.65284508e-01 2.65444666e-01 -1.72485858e-02
-2.84656435e-01 4.08833504e-01 -9.99655664e-01 1.79584169e+00
-9.58054364e-02 7.31637657e-01 -4.46435720e-01 -1.02751637e+00
5.22870839e-01 6.20201111e-01 1.04752481e+00 -4.69588637e-01
-6.07532673e-02 2.99696270e-02 -3.19670439e-01 -6.20419323e-01
3.18178445e-01 -1.53237477e-01 -1.56887770e-01 6.83089316e-01
1.39604986e-01 2.86478810e-02 6.56823218e-01 3.24127495e-01
1.15575433e+00 1.00682199e+00 3.22990529e-02 2.96808690e-01
3.46853077e-01 3.51793498e-01 9.17857766e-01 5.60735524e-01
-4.02070463e-01 5.94952643e-01 7.37208068e-01 -5.43361902e-01
-1.01136613e+00 -4.85086858e-01 1.39601260e-01 1.33796918e+00
1.81389332e-01 -6.90423608e-01 -9.70968246e-01 -1.19504201e+00
-3.14664751e-01 6.08258903e-01 -6.30055130e-01 -3.15564014e-02
-9.54617321e-01 -1.51707098e-01 6.65726066e-01 1.05488944e+00
4.47638392e-01 -1.06512141e+00 -1.12366688e+00 2.43491426e-01
-3.70955974e-01 -1.60918391e+00 -5.27384400e-01 2.62974709e-01
-1.08466804e+00 -1.10369313e+00 -6.02212012e-01 -7.77217686e-01
8.10495198e-01 -1.47588953e-01 8.80966485e-01 -8.24156255e-02
-7.90047571e-02 5.45999885e-01 -5.44370532e-01 -3.77580971e-01
-3.27413052e-01 -1.61822438e-01 -4.41815816e-02 9.78861600e-02
4.32791412e-01 -2.06032574e-01 -5.00119150e-01 5.29471815e-01
-8.98366988e-01 9.88245383e-02 2.15356145e-02 5.64264715e-01
7.37896681e-01 -1.00846570e-02 3.02834928e-01 -9.51177597e-01
-1.25569862e-03 -1.77810751e-02 -5.42592168e-01 8.95428061e-02
-3.45901847e-01 -4.36977744e-02 5.94560742e-01 -6.04588687e-01
-1.24400961e+00 7.84852386e-01 2.20959380e-01 -4.92208511e-01
-4.51933086e-01 2.46562585e-01 2.57621437e-01 2.79333413e-01
4.83102292e-01 -1.15932778e-01 -2.60104060e-01 -4.35303628e-01
2.26602867e-01 2.91827291e-01 4.99749035e-01 -5.13404667e-01
5.19578636e-01 8.70628297e-01 -1.49575740e-01 -4.47107702e-01
-1.20750475e+00 -6.73671186e-01 -1.04032314e+00 -5.29012561e-01
1.09171879e+00 -8.36378217e-01 -4.12996620e-01 3.26928616e-01
-1.09459066e+00 -7.68265367e-01 -6.60658240e-01 7.41520047e-01
-8.78760159e-01 5.62724531e-01 -7.87958801e-01 -8.22240889e-01
2.61037916e-01 -8.93520117e-01 1.17675531e+00 1.17636561e-01
-6.79646313e-01 -9.70358849e-01 8.22382271e-02 4.45028663e-01
-1.39397308e-01 6.32075489e-01 4.10855860e-01 -7.83138454e-01
-4.68836665e-01 -4.09095973e-01 7.22062960e-02 4.89451110e-01
2.13877439e-01 2.27016345e-01 -8.89594078e-01 -3.37053798e-02
-5.91747975e-03 -5.17903566e-01 7.34465599e-01 3.81217986e-01
1.02610099e+00 -1.96430653e-01 -3.76787931e-01 2.16120705e-01
1.00154960e+00 3.61883521e-01 4.12209809e-01 4.37639117e-01
7.22420454e-01 7.01499701e-01 1.38809288e+00 5.01503825e-01
3.08202863e-01 6.18878663e-01 2.01423883e-01 -1.90267637e-01
6.49737045e-02 -4.00904596e-01 7.35116363e-01 2.78414994e-01
-3.55953097e-01 -2.78131980e-02 -8.20661068e-01 7.00998724e-01
-2.17010546e+00 -1.18934286e+00 -2.31102750e-01 2.16208601e+00
1.05906212e+00 4.99720573e-01 4.52695340e-01 3.36310595e-01
6.68920934e-01 3.23821723e-01 -4.20841068e-01 -2.45806247e-01
2.95129538e-01 7.57940412e-02 5.44033289e-01 4.54137474e-01
-1.51120067e+00 1.10812700e+00 6.97517824e+00 3.11434805e-01
-9.92624402e-01 1.32504091e-01 5.35539448e-01 -3.59948903e-01
2.27061778e-01 4.31721300e-01 -6.58548474e-01 4.84970421e-01
1.15617418e+00 1.44056246e-01 -1.78354353e-01 9.17608023e-01
6.71711683e-01 -5.49988806e-01 -1.67180014e+00 6.13175571e-01
6.07479140e-02 -1.06815386e+00 -2.16591462e-01 -2.13832796e-01
8.89763832e-01 -1.56281561e-01 -3.75203520e-01 1.93960741e-01
2.16110259e-01 -6.91392243e-01 9.38789129e-01 2.89486110e-01
5.57621837e-01 -4.40827161e-01 6.59616113e-01 3.46964449e-01
-1.23475957e+00 -4.20517661e-02 9.67809260e-02 -2.85133421e-01
5.62463343e-01 4.30379301e-01 -8.83171737e-01 2.30824187e-01
6.54323339e-01 1.10573328e+00 -5.06311357e-01 9.31236506e-01
-6.22118771e-01 9.96238649e-01 -3.69041085e-01 2.99834549e-01
5.71919560e-01 -6.24274835e-02 1.58206031e-01 1.33648181e+00
-5.29489890e-02 4.80701402e-02 6.09092116e-01 3.17464441e-01
1.13568194e-01 -1.89235464e-01 -4.56620991e-01 -2.20006928e-01
1.16577424e-01 9.58574176e-01 -1.17369652e+00 -7.36752093e-01
-5.26351869e-01 1.12718320e+00 -1.27281711e-01 2.74724990e-01
-1.30602968e+00 -2.81172078e-02 2.38126919e-01 2.35384136e-01
2.94824034e-01 -3.93050700e-01 -1.26845270e-01 -8.68739665e-01
1.70851305e-01 -6.56260371e-01 7.56095231e-01 -6.45491600e-01
-6.89371049e-01 2.38134295e-01 2.84662127e-01 -1.42616320e+00
-4.33314502e-01 -2.46539161e-01 -5.32188833e-01 1.08766265e-01
-1.43201673e+00 -1.06135738e+00 -2.37279385e-01 4.96387959e-01
7.94386506e-01 5.50054371e-01 5.54220140e-01 3.89372051e-01
-3.77672374e-01 1.86771423e-01 -5.15122890e-01 4.73968148e-01
1.00214684e+00 -1.39464664e+00 2.89702505e-01 9.45595622e-01
3.24205667e-01 4.39577132e-01 7.40239024e-01 -8.53582263e-01
-7.30877936e-01 -1.04128826e+00 1.00178719e+00 -5.94066381e-01
7.05122113e-01 -2.30322748e-01 -7.94513881e-01 1.15680182e+00
1.60668463e-01 8.02097172e-02 7.11478829e-01 -1.78948149e-01
-2.58096188e-01 3.54236625e-02 -8.91836822e-01 3.07327121e-01
1.03915977e+00 -5.88226914e-01 -6.60031796e-01 6.25837922e-01
7.12832034e-01 -3.97506863e-01 -7.96741664e-01 5.05120158e-01
4.93333846e-01 -1.03245306e+00 5.44296205e-01 -6.46043658e-01
3.41807753e-01 -5.80729187e-01 2.85198718e-01 -5.83414197e-01
2.18379766e-01 -6.65991783e-01 -1.66975379e-01 1.30502617e+00
3.76851916e-01 -2.34579757e-01 1.09706545e+00 7.88220227e-01
4.00845148e-02 -5.26617467e-01 -8.65287185e-01 -1.02396357e+00
-5.60647845e-01 -4.68842506e-01 8.67154002e-02 9.97444510e-01
2.38249287e-01 -1.17653655e-02 -3.92298013e-01 4.93298210e-02
5.45427322e-01 6.68507144e-02 5.94449818e-01 -1.05713785e+00
-1.14125222e-01 7.33733699e-02 -4.93798941e-01 -1.07806313e+00
2.41291836e-01 -4.62255448e-01 5.05017996e-01 -1.46882153e+00
1.25560403e-01 -2.96083122e-01 -3.58336210e-01 9.53178346e-01
1.01070153e-02 4.72507805e-01 -1.74015891e-02 2.70590305e-01
-1.21746588e+00 1.35019317e-01 9.24252987e-01 1.42901167e-02
-2.61822551e-01 2.72157211e-02 1.08910687e-01 1.13977396e+00
8.05061460e-01 -7.05988586e-01 -5.63724339e-01 -3.01005363e-01
-9.80693474e-02 4.97555584e-02 2.71625727e-01 -1.22969413e+00
3.55478466e-01 -2.35220701e-01 2.26983160e-01 -5.84240317e-01
3.63342494e-01 -8.40746403e-01 -1.61451530e-02 4.75568533e-01
-7.12320209e-01 -2.14567706e-01 -4.78050746e-02 6.59556270e-01
-4.72758025e-01 -2.80901343e-01 7.52603948e-01 -3.08143914e-01
-8.70610535e-01 2.46893242e-01 -4.82761711e-01 -3.93775031e-02
1.46050370e+00 -4.64803159e-01 2.56747186e-01 -4.15538818e-01
-9.25465107e-01 2.70553499e-01 5.26287973e-01 3.77917916e-01
2.01356098e-01 -1.06231332e+00 -2.55646676e-01 -1.96658120e-01
-7.02082529e-04 -1.03717961e-03 -1.99482143e-01 9.63509738e-01
-2.84637868e-01 2.93511301e-01 -3.54391616e-03 -9.05299187e-01
-1.57630920e+00 4.71252114e-01 4.76202592e-02 -2.87206322e-01
-6.76213324e-01 7.72831917e-01 -1.86499774e-01 -8.21292177e-02
6.64093971e-01 -5.55671334e-01 -2.09430516e-01 3.79140317e-01
4.44979042e-01 4.12094116e-01 -2.06606820e-01 -5.68882942e-01
-4.10596758e-01 4.62289095e-01 -8.87935795e-03 -4.28291440e-01
1.29972434e+00 6.32483140e-02 8.79672691e-02 7.69283414e-01
1.02422559e+00 -2.78579205e-01 -1.88946354e+00 -8.21746066e-02
7.63295174e-01 -4.32270974e-01 -4.04291183e-01 -4.54373270e-01
-7.65231669e-01 6.79366529e-01 3.75273913e-01 2.31311888e-01
1.02487361e+00 1.02100614e-02 9.33925569e-01 9.83652994e-02
5.80708802e-01 -1.54970801e+00 4.45670426e-01 4.97303873e-01
1.94504306e-01 -1.20143712e+00 1.26764011e-02 -3.85970473e-01
-9.25471008e-01 9.11364973e-01 7.33796358e-01 -8.41521192e-03
2.18388453e-01 3.69593382e-01 2.71690965e-01 -1.65810939e-02
-6.01589322e-01 -3.03887069e-01 -5.47656380e-02 4.64400679e-01
5.54063201e-01 -3.08938235e-01 -5.32472908e-01 6.22953176e-02
2.21403256e-01 4.05407876e-01 4.47837085e-01 1.47689819e+00
-4.26240057e-01 -1.34820580e+00 -2.17200086e-01 1.74307555e-01
-6.86178923e-01 2.75611311e-01 -5.16856730e-01 6.19050086e-01
2.46983245e-01 1.04940820e+00 1.39719710e-01 -1.09704956e-01
1.89203769e-01 4.57545757e-01 3.86246026e-01 -8.51983607e-01
-5.10917306e-01 1.13663375e-01 3.78839076e-01 -9.65599120e-01
-1.13850832e+00 -9.18871045e-01 -1.83654308e+00 2.62002975e-01
-3.17005724e-01 3.25852096e-01 4.83089417e-01 1.07938862e+00
8.25205892e-02 3.63430619e-01 5.94788551e-01 -9.12568688e-01
-2.41341799e-01 -7.10809648e-01 -1.67064846e-01 7.28789151e-01
2.05129176e-01 -4.86900389e-01 -4.26726669e-01 8.17344606e-01] | [8.379790306091309, 0.516192615032196] |
7ad8bf55-4b91-456c-8369-30a24d304287 | adc-net-an-open-source-deep-learning-network | 2201.12625 | null | https://arxiv.org/abs/2201.12625v1 | https://arxiv.org/pdf/2201.12625v1.pdf | ADC-Net: An Open-Source Deep Learning Network for Automated Dispersion Compensation in Optical Coherence Tomography | Chromatic dispersion is a common problem to degrade the system resolution in optical coherence tomography (OCT). This study is to develop a deep learning network for automated dispersion compensation (ADC-Net) in OCT. The ADC-Net is based on a redesigned UNet architecture which employs an encoder-decoder pipeline. The input section encompasses partially compensated OCT B-scans with individual retinal layers optimized. Corresponding output is a fully compensated OCT B-scans with all retinal layers optimized. Two numeric parameters, i.e., peak signal to noise ratio (PSNR) and structural similarity index metric computed at multiple scales (MS-SSIM), were used for objective assessment of the ADC-Net performance. Comparative analysis of training models, including single, three, five, seven and nine input channels were implemented. The five-input channels implementation was observed as the optimal mode for ADC-Net training to achieve robust dispersion compensation in OCT | ['Visual Science', 'Department of Ophthalmology', 'University of Illinois at Chicago', 'Department of Biomedical Engineering', 'Xincheng Yao', 'Tobiloba Adejumo', 'Taeyoon Son', 'David Le', 'Shaiban Ahmed'] | 2022-01-29 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 2.11783290e-01 -1.98683456e-01 2.98797786e-01 -2.58297175e-01
-3.35192949e-01 -1.88029274e-01 -4.52723540e-02 -1.20131604e-01
-7.22112298e-01 7.98953474e-01 2.07080051e-01 -3.60062510e-01
-2.50494242e-01 -1.28150672e-01 -2.50251561e-01 -6.88203335e-01
-2.73483038e-01 -2.27256924e-01 3.96685869e-01 3.11974645e-01
6.41209126e-01 6.17288768e-01 -1.48581016e+00 6.24947846e-01
1.11747134e+00 1.25577676e+00 1.91323712e-01 1.11220682e+00
3.61053884e-01 4.72643971e-01 -3.91113907e-01 -2.69451737e-01
5.97190917e-01 -3.42404038e-01 -7.00011730e-01 -2.37781212e-01
9.69009995e-01 -7.14765251e-01 -1.39170084e-02 1.13148046e+00
1.28049481e+00 -4.30136353e-01 6.03331566e-01 -5.70644259e-01
-4.26670313e-01 7.45127574e-02 -4.22248811e-01 7.56276369e-01
-2.88529634e-01 6.78682625e-01 5.27604818e-01 -7.41621912e-01
5.00388384e-01 8.20670307e-01 7.00624764e-01 1.17406718e-01
-1.45648396e+00 -5.48250377e-01 -9.55078602e-01 3.50741029e-01
-1.07467043e+00 -3.14298838e-01 8.67086127e-02 -1.03059268e+00
1.00619864e+00 -1.73870176e-02 1.20150292e+00 2.88813621e-01
9.64669228e-01 -5.58006652e-02 1.60198379e+00 -3.14380020e-01
4.64546829e-02 -7.64337331e-02 -3.09936814e-02 4.17445511e-01
6.27663791e-01 5.64129829e-01 -1.79155275e-01 1.86339915e-01
1.23170686e+00 -8.51417840e-01 -2.37539515e-01 1.23539105e-01
-9.78207052e-01 4.87941176e-01 8.25888693e-01 -5.74562177e-02
-5.01296103e-01 1.92450896e-01 4.22135025e-01 4.56733167e-01
5.99244982e-02 1.10027826e+00 -1.46127820e-01 -3.58395614e-02
-1.17249489e+00 -1.71990097e-01 9.46656540e-02 4.75717902e-01
5.02155483e-01 1.82670146e-01 -6.14128470e-01 7.43631721e-01
2.77791262e-01 1.62737712e-01 5.72103977e-01 -1.30564296e+00
3.46572325e-03 5.46613276e-01 1.22910269e-01 -4.18474674e-01
-6.27537966e-01 -8.49996030e-01 -8.48098814e-01 8.23446572e-01
1.14627272e-01 -5.54709733e-01 -1.03318381e+00 8.32733154e-01
-8.16737488e-02 2.90793657e-01 -2.06984654e-01 1.19883049e+00
8.27836394e-01 3.29094119e-02 -2.61714846e-01 -3.10320139e-01
1.09614015e+00 -6.87830567e-01 -5.92882752e-01 -1.79812200e-02
3.01334441e-01 -1.31036532e+00 7.33635008e-01 4.29191589e-01
-1.49661255e+00 -7.51399875e-01 -1.30586553e+00 -1.79553971e-01
1.27611116e-01 7.02611506e-01 2.16724351e-01 6.36289001e-01
-1.62297928e+00 6.85927868e-01 -7.10872412e-01 -3.38965170e-02
5.84203005e-01 8.05176258e-01 -1.14810728e-01 2.26186886e-01
-6.85484350e-01 9.12280023e-01 2.18611926e-01 1.69435680e-01
-4.80114371e-01 -7.47507811e-01 -1.87940896e-01 -2.16055378e-01
-7.13388979e-01 -1.20776892e+00 1.16469419e+00 -7.09907174e-01
-1.70011222e+00 1.06729233e+00 -9.69348177e-02 -7.21994102e-01
5.07396817e-01 -1.07444204e-01 -4.73703414e-01 5.86568534e-01
-2.28193309e-02 8.70800793e-01 1.09152615e+00 -8.61981452e-01
-4.91848260e-01 -2.42776945e-01 -2.84331650e-01 -6.94023520e-02
1.83906734e-01 2.42759749e-01 3.23899090e-02 -3.20965588e-01
3.21547210e-01 -7.03951120e-01 -1.95019990e-01 3.74148399e-01
-4.64495778e-01 6.84941530e-01 1.54133633e-01 -6.46915913e-01
1.34432507e+00 -1.99733925e+00 -5.55339813e-01 1.52405620e-01
6.12116635e-01 9.59502995e-01 -3.83330882e-01 1.08145149e-02
-4.98397499e-01 2.00975128e-02 1.49810433e-01 1.06735677e-01
-6.00303173e-01 -4.26817745e-01 2.91463435e-01 4.45943385e-01
3.26589465e-01 6.72312796e-01 -6.36754572e-01 -3.77070814e-01
3.12439173e-01 3.75591129e-01 -8.38012755e-01 3.28077488e-02
1.18335914e-02 7.63954282e-01 7.97846243e-02 5.93031466e-01
1.11469293e+00 -5.32713473e-01 -2.13191405e-01 -7.37677991e-01
-6.41579747e-01 -3.18920501e-02 -1.07344973e+00 1.04912734e+00
-8.37157443e-02 1.33632970e+00 -2.09710091e-01 4.74592522e-02
7.40371287e-01 2.54217267e-01 5.59036016e-01 -7.28029549e-01
4.41273987e-01 6.86536491e-01 8.05924535e-01 -1.09501493e+00
2.45768920e-01 -8.24747533e-02 1.09601474e+00 1.12875067e-01
-1.44789323e-01 -9.35369655e-02 6.51635751e-02 -3.41071039e-01
7.97373533e-01 -3.44196111e-01 4.58705753e-01 -9.56946686e-02
4.99011636e-01 -1.81237116e-01 2.05181673e-01 4.10332263e-01
-4.30914879e-01 8.31427991e-01 7.95028269e-01 -4.01251316e-01
-1.36622810e+00 -9.92997289e-01 -6.46162868e-01 -2.04936694e-02
-3.62437926e-02 -2.15616137e-01 -7.28762567e-01 2.60477006e-01
-1.55731127e-01 1.43971816e-01 -1.78350836e-01 2.58964181e-01
-2.36099899e-01 -6.55704916e-01 5.85047126e-01 8.08862299e-02
9.00609851e-01 -4.52752054e-01 -9.86689389e-01 3.27129483e-01
2.29084507e-01 -9.29626644e-01 -3.74340147e-01 -3.53326321e-01
-1.27020574e+00 -1.32883477e+00 -8.21355462e-01 -8.01681757e-01
5.69262028e-01 5.00833765e-02 6.80241108e-01 -3.37191641e-01
-8.60968828e-01 -1.39456689e-01 2.56639600e-01 -3.91984433e-01
-1.60035729e-01 -4.88462567e-01 -8.49952474e-02 1.19390920e-01
2.77131200e-01 -7.69320786e-01 -1.55252993e+00 2.31814533e-01
-5.35428286e-01 9.60038230e-02 1.36670148e+00 8.48554790e-01
8.81590784e-01 -1.37147754e-01 1.77583043e-02 -4.28698391e-01
8.90957832e-01 2.44417816e-01 -8.77221406e-01 -2.11278766e-01
-8.77015650e-01 -1.77771658e-01 1.42939895e-01 -1.89564109e-01
-5.26533782e-01 -1.73258185e-01 1.34248435e-01 -4.43383515e-01
8.21384043e-02 4.94134277e-01 5.57573915e-01 -8.11334133e-01
1.14259434e+00 4.07969058e-02 3.58329743e-01 -1.48640633e-01
-2.09376037e-01 9.81140614e-01 4.89318758e-01 2.31026217e-01
8.11003819e-02 2.85273552e-01 5.07065356e-01 -9.04321790e-01
-3.79530877e-01 -5.26019812e-01 -6.47614241e-01 -3.66711587e-01
8.39719474e-01 -1.06458616e+00 -8.48651886e-01 8.52979779e-01
-1.29618299e+00 -1.29995391e-01 1.04504600e-01 1.26494801e+00
-4.31699485e-01 1.83186203e-01 -6.20289326e-01 -1.89811721e-01
-5.71632981e-01 -1.47824490e+00 6.90879762e-01 4.03124094e-01
-6.42463341e-02 -6.72429204e-01 1.46931663e-01 3.06354165e-01
6.11474752e-01 1.39517412e-01 1.13746953e+00 3.13776881e-02
-8.81676078e-01 -8.74214061e-03 -1.02010083e+00 1.03547490e+00
4.09037247e-02 3.93164545e-01 -1.12716818e+00 -3.15994054e-01
-1.06321946e-01 2.22248659e-01 6.71623766e-01 1.32590604e+00
1.05116808e+00 -3.18198860e-01 3.14038433e-03 1.10529351e+00
1.90405917e+00 2.46339709e-01 1.31701696e+00 5.13490558e-01
1.75910592e-01 2.29583651e-01 2.48039529e-01 5.10510683e-01
-1.49272844e-01 4.23370779e-01 6.50124133e-01 -3.28864694e-01
-8.04796398e-01 4.97455209e-01 -8.44559520e-02 4.04697716e-01
-5.92010558e-01 1.12907015e-01 -9.13915992e-01 4.11215544e-01
-1.01203895e+00 -7.49478400e-01 -5.83453894e-01 2.25807190e+00
8.79264653e-01 -3.95533722e-03 -1.08393975e-01 -2.57262856e-01
7.74255931e-01 -2.64546216e-01 -6.95504665e-01 -3.22476864e-01
-1.24708280e-01 6.36776030e-01 8.04247379e-01 4.50726479e-01
-8.87511075e-01 4.85688239e-01 7.30198479e+00 2.79813200e-01
-1.73812056e+00 -1.34660572e-01 2.68754125e-01 -5.03361702e-01
3.50717872e-01 -1.81691021e-01 -5.94643831e-01 7.65379727e-01
9.88741994e-01 -1.60904750e-02 1.75627545e-01 2.15189993e-01
9.84328091e-01 -3.39908361e-01 -8.38306487e-01 1.23971760e+00
-4.31152433e-01 -1.62952721e+00 8.40932652e-02 3.71841848e-01
8.93412769e-01 4.56077218e-01 5.23414135e-01 -7.08279848e-01
-3.35901201e-01 -1.20803666e+00 -9.06260088e-02 9.65640366e-01
1.39853597e+00 -3.56799096e-01 1.04555070e+00 -4.66294080e-01
-5.09129643e-01 -1.50823951e-01 -4.19978946e-01 1.82551116e-01
-2.19834626e-01 6.72345400e-01 -1.27361381e+00 2.14446366e-01
5.47382534e-01 7.87966192e-01 -6.95098400e-01 2.40219331e+00
-5.36807701e-02 3.72093529e-01 2.22530752e-01 1.51941702e-01
2.19752878e-01 -2.77815342e-01 8.73158872e-01 1.15414822e+00
6.22861326e-01 -9.41738114e-02 -7.74034142e-01 7.32820392e-01
3.14326316e-01 -1.39310025e-02 -1.05757572e-01 1.92413516e-02
4.58816826e-01 9.49713767e-01 -8.08196738e-02 1.22556366e-01
-2.93282926e-01 4.32386667e-01 -4.08735365e-01 5.32200396e-01
-4.11193669e-01 -5.75754046e-01 9.23505604e-01 8.19488585e-01
1.96563274e-01 -1.42508268e-01 -6.09768331e-01 -5.17809391e-01
-1.77609414e-01 -6.71304405e-01 -1.41844526e-01 -1.44175005e+00
-1.08607364e+00 5.20190120e-01 -5.43983757e-01 -1.94361413e+00
1.53023481e-01 -9.84173179e-01 -5.79762340e-01 1.44271469e+00
-1.85278189e+00 -9.42573011e-01 -5.23132145e-01 5.33243954e-01
-7.83269629e-02 -5.71631253e-01 4.46255207e-01 3.67989302e-01
-3.55611891e-01 4.74712133e-01 3.62573445e-01 -3.21226269e-02
1.00348568e+00 -1.18369901e+00 -1.69138640e-01 8.32589805e-01
-8.84160459e-01 7.36580014e-01 4.80635107e-01 -6.52743638e-01
-6.86324179e-01 -1.20254016e+00 8.30061376e-01 1.70866862e-01
5.63482225e-01 7.17267692e-01 -3.64059359e-01 5.39695442e-01
3.48965198e-01 2.95710452e-02 7.80204594e-01 -4.91243273e-01
1.29413813e-01 -3.87187511e-01 -1.18255329e+00 5.32534719e-01
3.49912405e-01 -3.45534772e-01 -2.03114003e-01 2.02143759e-01
3.12452227e-01 -8.28361034e-01 -1.15086651e+00 2.38306493e-01
8.10388327e-01 -1.53439856e+00 8.10644686e-01 -3.41319531e-01
7.68476188e-01 -4.06093359e-01 1.50959253e-01 -1.31065726e+00
-5.35991192e-01 -8.36027563e-01 2.32944265e-01 3.24892908e-01
3.81355792e-01 -1.06296432e+00 6.11563087e-01 9.40808505e-02
-6.34243608e-01 -7.38655031e-01 -9.70166981e-01 -5.92266977e-01
-2.57352948e-01 1.19901195e-01 3.76137674e-01 3.42783004e-01
-5.27673304e-01 2.14782387e-01 -1.19093023e-02 5.08700609e-01
7.20594764e-01 -3.21458101e-01 4.24132794e-01 -1.31305611e+00
1.84328668e-02 -6.36855364e-01 -1.12251389e+00 -7.78288126e-01
-7.34635055e-01 -6.13032103e-01 -4.68947470e-01 -1.62298512e+00
-1.52946264e-01 -2.18573138e-01 -1.56640187e-01 -1.18835844e-01
2.87312210e-01 3.21231663e-01 -2.24192023e-01 4.66318041e-01
3.67260575e-01 -1.15891330e-01 2.16386652e+00 -1.41758807e-02
-5.19579172e-01 4.40651059e-01 -5.47512293e-01 5.35971522e-01
7.06479132e-01 -9.82561484e-02 -5.32974124e-01 -6.94803596e-01
3.25541377e-01 1.61257282e-01 7.43997276e-01 -1.52350116e+00
7.24518001e-01 1.80768818e-01 7.07455933e-01 -6.22416377e-01
3.57284218e-01 -4.97981042e-01 3.05253774e-01 7.10178256e-01
-2.84874797e-01 -2.99755573e-01 2.88472593e-01 4.76494059e-02
-4.98293489e-01 -6.04796223e-02 1.53912866e+00 -5.75863868e-02
-6.69779420e-01 4.00477648e-01 -1.14174053e-01 -2.71544665e-01
7.10859716e-01 -1.01483595e+00 -6.78113461e-01 1.55738229e-02
-1.07794559e+00 -8.17832947e-02 2.74846166e-01 -6.22990489e-01
1.01644647e+00 -1.01540291e+00 -9.80261505e-01 5.63157260e-01
2.13754885e-02 -3.22990537e-01 5.25997579e-01 1.61567593e+00
-1.34037530e+00 6.58126950e-01 -9.90703344e-01 -7.58549213e-01
-1.57994783e+00 -3.89124364e-01 1.16090238e+00 1.08467825e-01
-3.69099259e-01 1.05317330e+00 -3.13907027e-01 4.87870991e-01
7.85084069e-02 -6.70955896e-01 -3.20625007e-01 -7.07679288e-03
6.19121134e-01 8.25294018e-01 4.09921169e-01 -1.30298123e-01
3.25541794e-02 9.58134770e-01 -1.64514914e-01 6.37331083e-02
1.26671720e+00 -3.49003464e-01 -6.00222588e-01 -8.89185444e-03
1.13585901e+00 -2.11872041e-01 -1.24321938e+00 -7.87553340e-02
-5.11359870e-01 -8.59565020e-01 5.74010074e-01 -1.21394908e+00
-8.64409924e-01 1.10493159e+00 1.49924254e+00 -1.59750402e-01
1.40615630e+00 -7.57135451e-01 6.64680064e-01 1.45119373e-02
-1.24438785e-01 -9.60839450e-01 2.05865502e-01 2.93606073e-01
8.95106971e-01 -1.10572314e+00 1.32759362e-01 -2.77264744e-01
-4.54346985e-01 1.47153187e+00 5.50293565e-01 -1.06922992e-01
7.32804894e-01 -9.39511880e-02 2.99376160e-01 -2.36139551e-01
-4.54740256e-01 -2.34284401e-01 6.60069585e-01 7.32700646e-01
6.11373842e-01 -1.61441922e-01 -4.05944616e-01 7.19082579e-02
-4.01862562e-01 6.01450145e-01 9.25149620e-01 3.37974995e-01
-7.06328034e-01 -7.57371962e-01 -6.06947839e-02 1.03380775e+00
-5.01923203e-01 -4.94125336e-01 -7.12907240e-02 5.84018350e-01
5.29760718e-01 7.27894247e-01 1.44617468e-01 -2.65744776e-01
2.98135817e-01 -5.61695516e-01 7.02774942e-01 -6.49039567e-01
-6.99042737e-01 4.36156780e-01 5.07973731e-02 -5.47986984e-01
-3.97765696e-01 -3.70612234e-01 -8.97899568e-01 -3.34091961e-01
7.33882859e-02 -4.59413022e-01 7.63364375e-01 2.94359744e-01
6.23274982e-01 5.32140315e-01 6.14392221e-01 -4.25441682e-01
-2.36475542e-01 -1.18181479e+00 -1.05014396e+00 -5.34055293e-01
1.05782652e+00 -3.35748792e-01 -3.02745521e-01 2.69626588e-01] | [15.820513725280762, -3.998412609100342] |
0fcbe66a-f998-441e-99fa-7e7b3fda79d2 | cursive-caption-text-detection-in-videos | 2301.03164 | null | https://arxiv.org/abs/2301.03164v1 | https://arxiv.org/pdf/2301.03164v1.pdf | Cursive Caption Text Detection in Videos | Textual content appearing in videos represents an interesting index for semantic retrieval of videos (from archives), generation of alerts (live streams) as well as high level applications like opinion mining and content summarization. One of the key components of such systems is the detection of textual content in video frames and the same makes the subject of our present study. This paper presents a robust technique for detection of textual content appearing in video frames. More specifically we target text in cursive script taking Urdu text as a case study. Detection of textual regions in video frames is carried out by fine-tuning object detectors based on deep convolutional neural networks for the specific case of text detection. Since it is common to have videos with caption text in multiple-scripts, cursive text is distinguished from Latin text using a script-identification module. Finally, detection and script identification are combined in a single end-to-end trainable system. Experiments on a comprehensive dataset of around 11,000 video frames report an F-measure of 0.91. | ['Imran Siddiqi', 'Ali Mirza'] | 2023-01-09 | null | null | null | null | ['semantic-retrieval'] | ['natural-language-processing'] | [ 4.84251887e-01 -2.68980801e-01 5.15603982e-02 -1.63561702e-01
-8.47639680e-01 -9.34020162e-01 8.95713449e-01 3.42308730e-01
-4.71669853e-01 3.09584916e-01 3.53927255e-01 1.48976266e-01
3.07677239e-01 -3.21532100e-01 -9.17237580e-01 -6.70445681e-01
-1.11281484e-01 3.16080526e-02 6.91466391e-01 1.53839469e-01
7.32985139e-01 6.33647978e-01 -1.77550149e+00 9.34368134e-01
5.14095128e-01 1.06579077e+00 4.21711802e-01 1.11590517e+00
-1.08512245e-01 1.22642279e+00 -1.00035310e+00 -5.18209457e-01
6.37491196e-02 -5.06561577e-01 -6.77145720e-01 5.45413613e-01
8.45816493e-01 -6.33638859e-01 -4.05836284e-01 1.11963534e+00
3.75833273e-01 1.44957572e-01 6.59415305e-01 -1.06536412e+00
-1.02489106e-01 6.74609959e-01 -4.47354317e-01 6.13695741e-01
6.64200842e-01 1.50552979e-02 6.28681421e-01 -9.59814548e-01
8.59691024e-01 1.12162030e+00 5.92035949e-01 1.10528789e-01
-5.54318547e-01 -2.00733170e-01 -2.55607247e-01 3.09429526e-01
-1.30684698e+00 -5.49728334e-01 5.84463835e-01 -6.65110290e-01
8.41659129e-01 4.42580074e-01 2.92761177e-01 1.01370740e+00
1.21001191e-01 1.02540505e+00 5.86261749e-01 -4.90779042e-01
2.22048327e-01 3.43505651e-01 -3.40049528e-02 6.39253795e-01
1.45401403e-01 -9.32853460e-01 -6.44200325e-01 4.31590416e-02
5.39246500e-01 -3.19152623e-02 -1.65341273e-01 1.29183680e-01
-1.19734395e+00 6.57155216e-01 -1.68110728e-01 4.67730612e-01
-5.26311338e-01 4.48671319e-02 1.01213217e+00 2.60181576e-01
4.65026557e-01 1.02672279e-01 -1.20454662e-01 -2.56284922e-01
-1.45848405e+00 2.68973380e-01 7.36952364e-01 9.97994900e-01
2.84791589e-01 1.56017482e-01 -4.72282082e-01 7.74019599e-01
-1.04099885e-02 4.03243989e-01 5.18826544e-01 -7.45071948e-01
6.50874019e-01 6.77322328e-01 9.65578556e-02 -1.14654684e+00
-1.93924546e-01 -3.75309214e-02 -2.89670110e-01 -2.48669356e-01
3.07850271e-01 -3.46343845e-01 -6.62890315e-01 8.67319465e-01
3.10590804e-01 -4.77175787e-02 -1.79771751e-01 9.17470455e-01
8.15961480e-01 1.06789351e+00 -1.68257296e-01 -3.67732316e-01
1.84662247e+00 -7.09977031e-01 -7.93128550e-01 5.77836260e-02
3.15732419e-01 -1.13658249e+00 5.35679698e-01 5.99215984e-01
-8.54241312e-01 -5.03443539e-01 -8.24058354e-01 8.48495588e-02
-5.63909471e-01 6.58891022e-01 4.11266301e-05 6.57815099e-01
-8.48333359e-01 2.95466691e-01 -4.87069607e-01 -1.05204415e+00
4.02697951e-01 1.56932116e-01 -2.50710487e-01 6.35887533e-02
-7.29274750e-01 4.99602467e-01 5.43468177e-01 -3.37138660e-02
-1.01256573e+00 -2.66016692e-01 -5.80033958e-01 -1.58911943e-01
6.67801738e-01 2.84255166e-02 1.10109639e+00 -1.61458564e+00
-1.09864450e+00 1.21761346e+00 -1.00667395e-01 -4.89806324e-01
8.21864247e-01 -4.64622766e-01 -5.66259623e-01 9.07068610e-01
2.24588916e-01 3.25920224e-01 1.38836539e+00 -7.53649652e-01
-9.03994143e-01 -4.84373942e-02 -2.03140154e-02 4.55408059e-02
-5.86302340e-01 9.83485103e-01 -6.31466627e-01 -9.14186060e-01
-3.78455639e-01 -5.79123437e-01 3.89707565e-01 -2.34124660e-01
-4.89290714e-01 -3.71783644e-01 1.38679743e+00 -1.18213427e+00
1.31539607e+00 -1.92631423e+00 -1.09191693e-01 -1.44956056e-02
4.93425652e-02 3.56466144e-01 1.89265162e-01 7.05989063e-01
-2.30749790e-02 -1.09161228e-01 9.23346430e-02 2.25790188e-01
-2.99205501e-02 -3.87392282e-01 -3.66770387e-01 6.45799935e-01
4.39852595e-01 3.44090819e-01 -6.69247031e-01 -7.93600202e-01
3.32503676e-01 2.74766266e-01 -4.02216911e-01 2.69279748e-01
-4.38226670e-01 5.81207825e-03 -5.33329785e-01 8.16526830e-01
3.44383508e-01 -1.04061030e-02 2.14260910e-02 -4.23947811e-01
-4.52550471e-01 4.41753909e-05 -1.26744926e+00 1.24701786e+00
1.03128225e-01 1.30471659e+00 8.97328407e-02 -1.01019526e+00
7.32122600e-01 5.51827371e-01 5.58952451e-01 -3.04253578e-01
3.39725286e-01 9.16983113e-02 -4.58148301e-01 -1.10459101e+00
1.02598977e+00 5.47405839e-01 -7.29016811e-02 3.61882538e-01
1.15259781e-01 3.39954019e-01 8.90126824e-01 3.72491926e-01
1.04763317e+00 1.85430095e-01 1.14561230e-01 -2.17856020e-01
8.25856090e-01 2.18760535e-01 3.41235995e-02 7.56894767e-01
-1.81260630e-01 8.40022922e-01 8.42841923e-01 -4.47676659e-01
-1.27811289e+00 -3.76240522e-01 -7.40233213e-02 1.32295322e+00
-3.09691936e-01 -4.45588499e-01 -1.24565673e+00 -6.03352070e-01
-3.15037102e-01 5.21347642e-01 -5.37481844e-01 2.92375952e-01
-7.91370273e-01 -7.03292429e-01 5.45710385e-01 2.10935891e-01
4.58756298e-01 -1.13743639e+00 -8.77039373e-01 1.71983048e-01
-3.11378598e-01 -1.53291428e+00 -4.71506745e-01 3.22098993e-02
-5.01995921e-01 -1.17903125e+00 -9.58005428e-01 -8.01761985e-01
4.75834429e-01 2.27451190e-01 8.23724449e-01 -1.96256518e-01
-6.33184731e-01 6.47826314e-01 -7.56595492e-01 -2.27435365e-01
-7.21985459e-01 -1.96173694e-02 -7.92963356e-02 5.19165218e-01
4.59476352e-01 2.53732771e-01 -3.88942599e-01 1.66571066e-01
-1.23582065e+00 -2.83147126e-01 1.29046082e-01 1.86513633e-01
9.45760533e-02 1.11663667e-02 2.68490404e-01 -5.70216000e-01
4.50255305e-01 -6.43035591e-01 -7.46400714e-01 2.18506306e-01
3.08690161e-01 -4.85072017e-01 7.81333029e-01 -3.63917768e-01
-8.99820685e-01 3.38616103e-01 1.89676732e-01 -5.24477780e-01
-4.77436900e-01 1.97337568e-01 1.06854774e-01 1.58314288e-01
6.59185708e-01 5.43853164e-01 -3.14892083e-01 -4.77560967e-01
-7.20565440e-05 1.14611578e+00 4.74446535e-01 -2.00143635e-01
4.95278031e-01 4.13399667e-01 -3.44830692e-01 -1.55278361e+00
-5.48451304e-01 -8.14841151e-01 -6.54241323e-01 -9.06366765e-01
1.21246421e+00 -1.06631279e+00 -4.38531697e-01 5.96409798e-01
-1.34867966e+00 2.25729465e-01 1.45057186e-01 3.33311081e-01
-4.51372564e-01 5.56391656e-01 -5.80018222e-01 -8.96892428e-01
-3.74121934e-01 -1.10620010e+00 1.32735133e+00 3.31381083e-01
-2.05024630e-01 -5.81164956e-01 -3.17270756e-01 5.28229356e-01
1.07311651e-01 3.81799072e-01 3.09336483e-01 -9.79783893e-01
-6.29579723e-01 -6.65556967e-01 -1.65067345e-01 3.06331724e-01
-1.19041920e-01 5.41694820e-01 -1.09586477e+00 -7.19262809e-02
-2.26624936e-01 -1.81183726e-01 8.35951269e-01 3.59896243e-01
1.02443194e+00 -2.05467120e-01 -9.95457452e-03 1.86815843e-01
1.36601472e+00 2.01400161e-01 5.83671331e-01 5.98006666e-01
7.86031961e-01 6.56413198e-01 6.67295158e-01 8.01663578e-01
-1.85594231e-01 5.45889437e-01 4.22880113e-01 2.98641860e-01
-3.80545259e-02 2.78286636e-01 8.86805356e-01 4.71461713e-01
-8.12983587e-02 -5.14373422e-01 -8.54146004e-01 6.06798887e-01
-1.70831406e+00 -1.38678396e+00 -5.05061984e-01 2.02784467e+00
5.39799333e-01 -6.52896613e-03 5.15629172e-01 1.31556094e-01
1.26521373e+00 1.10830665e-01 -1.80828065e-01 -4.74184126e-01
-1.69068381e-01 -2.97077030e-01 6.77966416e-01 -1.28882587e-01
-1.78833675e+00 9.13253725e-01 5.43408537e+00 8.71583164e-01
-1.18794084e+00 2.15578941e-03 4.95091230e-01 -3.39612365e-01
6.57144785e-01 -5.20214736e-01 -1.02104783e+00 7.26413429e-01
9.69390512e-01 1.33569598e-01 4.34264876e-02 8.38444948e-01
6.27621412e-01 -3.64507914e-01 -8.90115440e-01 9.44953680e-01
7.10605621e-01 -1.49723995e+00 1.47780553e-01 -2.87959784e-01
7.59516597e-01 4.71332967e-02 -3.65436465e-01 -1.26719654e-01
-3.29880863e-01 -4.97822791e-01 1.22486877e+00 3.37608308e-01
5.77574015e-01 -8.19959044e-01 7.10293591e-01 -1.97052881e-02
-1.04423237e+00 -2.04153255e-01 -3.83052856e-01 4.10351187e-01
3.76371294e-02 4.91454571e-01 -8.71825278e-01 3.48279059e-01
8.40260684e-01 9.12138581e-01 -8.59981537e-01 1.35795259e+00
1.88514635e-01 7.60044992e-01 -2.63142794e-01 -1.97830871e-01
4.48707402e-01 2.22788732e-02 7.79729009e-01 1.96893203e+00
4.13438886e-01 -3.49099189e-01 1.56282440e-01 4.13441420e-01
-2.09880039e-01 3.77934933e-01 -5.23050308e-01 -4.73812163e-01
1.38184234e-01 1.45422554e+00 -1.45046794e+00 -6.51300192e-01
-4.67856377e-01 1.13577807e+00 -3.12931240e-01 1.47196591e-01
-9.18035328e-01 -6.74842298e-01 2.22274050e-01 1.90223083e-01
7.44642913e-01 1.09094433e-01 3.41170639e-01 -1.15689659e+00
2.04433754e-01 -1.13258111e+00 2.76213855e-01 -8.87061954e-01
-8.04340720e-01 4.69211996e-01 -9.30566341e-02 -1.42498755e+00
-3.98614407e-01 -7.69952595e-01 -5.92231095e-01 3.44127297e-01
-8.92466843e-01 -9.76135969e-01 -4.13044035e-01 6.22976720e-01
1.16948533e+00 -4.51986194e-01 1.99882537e-01 4.98520494e-01
-8.84617448e-01 2.05851525e-01 4.91935700e-01 5.52671850e-01
8.67677093e-01 -1.12532604e+00 1.43161073e-01 1.39122391e+00
-7.76846111e-02 1.81326941e-01 9.72664654e-01 -6.80544198e-01
-1.79475176e+00 -1.27049994e+00 5.82302034e-01 -4.43116337e-01
8.51796627e-01 -3.72811735e-01 -7.59888411e-01 4.42119390e-01
6.91687524e-01 -4.10586238e-01 2.87215710e-01 -7.94137537e-01
-6.36189580e-02 -4.59024459e-02 -8.15625787e-01 4.35633719e-01
3.53886873e-01 -4.59618837e-01 -3.87511611e-01 9.62667644e-01
1.46023437e-01 -2.33093470e-01 -5.82380235e-01 -2.77252674e-01
4.49913591e-01 -9.15722668e-01 7.70477176e-01 -3.82249743e-01
7.10902691e-01 -4.31939691e-01 -9.66193900e-02 -5.76930165e-01
2.88829744e-01 -7.36532331e-01 -6.34897649e-02 1.56707025e+00
-1.66817173e-01 1.50887221e-01 6.19062543e-01 1.57283366e-01
-2.90938169e-02 2.63591200e-01 -6.79442108e-01 -5.22927523e-01
-6.13452077e-01 -4.32987958e-01 -7.53586516e-02 9.63479817e-01
-6.60552159e-02 2.21265525e-01 -6.07333660e-01 2.28111371e-02
5.62627971e-01 -1.03595138e-01 7.51321673e-01 -1.01036918e+00
4.99555431e-02 -5.80901265e-01 -6.64471984e-01 -5.74137449e-01
-7.35651106e-02 -6.29502714e-01 1.21406011e-01 -1.35807729e+00
3.20728570e-01 5.91627896e-01 2.20063344e-01 -2.19684169e-02
7.27900788e-02 4.69844699e-01 2.75484025e-01 1.33421898e-01
-1.10686147e+00 -1.86546311e-01 5.37416816e-01 4.15584370e-02
1.10134386e-01 -3.13527472e-02 -4.55052555e-02 8.15355062e-01
7.75016963e-01 -4.65850383e-01 5.01178056e-02 -1.50269061e-01
3.91108185e-01 -2.85914056e-02 3.46905112e-01 -1.13658977e+00
1.42409474e-01 1.82734020e-02 5.72016418e-01 -1.11607671e+00
6.27579093e-02 -7.91759670e-01 -2.32179910e-01 3.40838879e-01
-3.61801773e-01 9.49953943e-02 2.05422431e-01 4.92744863e-01
-1.89419955e-01 -6.53146148e-01 6.63052917e-01 -2.20337823e-01
-1.13040698e+00 -3.28962356e-01 -1.23986840e+00 -1.00076543e-02
1.18917966e+00 -5.13871133e-01 -3.60065997e-01 -2.60679722e-01
-2.63631076e-01 -1.05823362e-02 3.53460610e-01 5.77725291e-01
4.51255530e-01 -7.46854603e-01 -8.87240112e-01 -6.52043745e-02
1.27074882e-01 -5.30026257e-01 1.14897020e-01 8.88228118e-01
-1.15488994e+00 4.89598304e-01 -2.40300268e-01 -7.04857707e-01
-1.75743628e+00 4.59502816e-01 1.74827740e-01 4.04121250e-01
-5.16188323e-01 5.70184886e-01 -1.03996001e-01 3.44321221e-01
4.93381947e-01 -1.91722557e-01 -6.11587763e-01 7.85292745e-01
1.06343079e+00 6.47910953e-01 2.27210850e-01 -9.90150034e-01
-4.25528228e-01 4.46971059e-01 -8.87287408e-02 2.68871814e-01
1.23928905e+00 -1.86063349e-01 -2.53204793e-01 4.47733849e-01
1.32147193e+00 -3.23149003e-02 -1.12853110e+00 -2.90519688e-02
2.06727996e-01 -2.64157772e-01 8.54770392e-02 -4.98649389e-01
-8.91945660e-01 7.52073288e-01 6.36595368e-01 6.70582414e-01
1.05090082e+00 -5.22932969e-02 5.04587352e-01 5.97540021e-01
-2.62624174e-01 -1.57841444e+00 1.70107707e-01 4.49536383e-01
7.99796641e-01 -1.21729553e+00 1.04105026e-01 -1.03074923e-01
-6.21011734e-01 1.64440715e+00 2.11615562e-01 -2.06279844e-01
4.57839705e-02 1.98464841e-01 -1.64368913e-01 -1.54444641e-02
-6.18082821e-01 -2.43271828e-01 3.86150688e-01 3.20167750e-01
6.01105630e-01 -2.93766826e-01 -3.16073686e-01 1.03507482e-01
2.25792080e-01 -8.60831589e-02 1.03123808e+00 1.20768917e+00
-9.44154024e-01 -4.30505276e-01 -8.88732016e-01 6.88648522e-01
-1.25493586e+00 -1.70663074e-02 -5.29583335e-01 5.38466394e-01
-1.44769147e-01 9.68494654e-01 3.68120134e-01 -1.41160917e-02
6.56988472e-02 3.19275796e-01 2.61122227e-01 -3.18920165e-01
-1.07950842e+00 4.80837256e-01 2.39004746e-01 -4.69208509e-01
-6.73289955e-01 -1.10913265e+00 -9.25045431e-01 -1.05508737e-01
-8.46368149e-02 -9.63507742e-02 9.42719698e-01 8.55866849e-01
1.03820294e-01 5.39979696e-01 5.51387787e-01 -9.60487247e-01
-1.40867263e-01 -9.12128389e-01 -4.55381393e-01 4.64541018e-01
3.07409972e-01 -1.87047839e-01 -7.88289681e-02 9.98466909e-01] | [10.408671379089355, 0.6204200983047485] |
25601c27-4d3b-46ba-b5f4-d3ee214dc206 | addressing-time-bias-in-bipartite-graph | 1911.12558 | null | https://arxiv.org/abs/1911.12558v1 | https://arxiv.org/pdf/1911.12558v1.pdf | Addressing Time Bias in Bipartite Graph Ranking for Important Node Identification | The goal of the ranking problem in networks is to rank nodes from best to worst, according to a chosen criterion. In this work, we focus on ranking the nodes according to their quality. The problem of ranking the nodes in bipartite networks is valuable for many real-world applications. For instance, high-quality products can be promoted on an online shop or highly reputed restaurants attract more people on venues review platforms. However, many classical ranking algorithms share a common drawback: they tend to rank older movies higher than newer movies, though some newer movies may have a high quality. This time bias originates from the fact that older nodes in a network tend to have more connections than newer ones. In the study, we develop a ranking method using a rebalance approach to diminish the time bias of the rankings in bipartite graphs. | ['Mingyang Zhou', 'Jiao Wu', 'Hao Liao', 'Alexandre Vidmer'] | 2019-11-28 | null | null | null | null | ['graph-ranking'] | ['graphs'] | [-2.64960468e-01 2.42845073e-01 -5.13288140e-01 -2.99659997e-01
-5.74892648e-02 -4.87955421e-01 3.79842609e-01 6.43266499e-01
-3.60013932e-01 9.05315459e-01 6.59357086e-02 9.72610414e-02
-9.34508026e-01 -1.32849419e+00 -1.94704473e-01 -5.17426610e-01
-2.68247843e-01 6.72037065e-01 5.13607740e-01 -6.86082423e-01
3.40550601e-01 4.93473858e-01 -1.41801250e+00 1.11158796e-01
7.92721510e-01 5.46340346e-01 1.11027263e-01 1.40272498e-01
-2.79847793e-02 4.08942401e-01 -4.50321347e-01 -1.01692498e+00
2.95807302e-01 -3.22743446e-01 -6.79092765e-01 -2.81956255e-01
2.30072930e-01 4.34979126e-02 -1.26442090e-01 1.44695246e+00
1.41644493e-01 1.93748474e-01 5.72848499e-01 -1.58293927e+00
-3.33982944e-01 1.11805546e+00 -5.83768129e-01 1.21630706e-01
4.32370812e-01 -6.83837950e-01 1.70948470e+00 -3.57946008e-01
8.78421068e-01 1.07650983e+00 2.06944376e-01 1.45985380e-01
-1.22860408e+00 -2.80746788e-01 3.64194781e-01 3.28432113e-01
-1.18750250e+00 3.45417075e-02 8.98622990e-01 -3.73510152e-01
-1.96354482e-02 5.50037861e-01 7.70815551e-01 5.19539595e-01
2.53523350e-01 1.91829845e-01 1.17578363e+00 -4.63426597e-02
1.43850297e-01 4.57626790e-01 2.84969985e-01 1.99331433e-01
7.77485549e-01 -3.33459646e-01 -5.63907266e-01 -1.36188313e-01
3.06993127e-01 3.03159505e-01 -9.62050781e-02 -7.51848340e-01
-9.51881528e-01 8.82155895e-01 8.11840236e-01 7.48913407e-01
-4.47866201e-01 -2.55524188e-01 1.68956637e-01 7.10392952e-01
5.15639782e-01 8.92858088e-01 -2.98449159e-01 5.06044142e-02
-6.93863630e-01 9.22022983e-02 8.34823430e-01 7.38785267e-01
8.72141242e-01 -6.86531544e-01 -4.27102670e-02 8.51090670e-01
5.07674098e-01 -3.96321565e-02 -9.38366726e-02 -7.32627928e-01
2.18428403e-01 8.72539222e-01 9.68933254e-02 -1.73299444e+00
-4.14225996e-01 -6.68830514e-01 -9.92759645e-01 1.08211562e-01
5.41887283e-01 3.01746666e-01 -1.18318498e-01 1.58600605e+00
7.15383962e-02 -5.02325714e-01 -2.96943069e-01 1.02961552e+00
7.33478308e-01 6.13511980e-01 -1.76659986e-01 -4.40839380e-01
1.09606826e+00 -8.15797746e-01 -5.73556304e-01 2.50049084e-01
2.10466444e-01 -7.49876380e-01 7.24198639e-01 7.11722195e-01
-1.19583166e+00 -5.59485495e-01 -6.76026881e-01 4.34361219e-01
-3.57416153e-01 -2.32166260e-01 6.42555892e-01 5.46100438e-01
-1.34010923e+00 1.27087557e+00 -3.67538333e-02 -6.42008781e-01
-1.57494649e-01 5.20571709e-01 -2.62101829e-01 2.61140335e-02
-1.54683292e+00 1.00805426e+00 1.17815256e-01 1.17223635e-01
-5.01454413e-01 -2.78392166e-01 -2.96883047e-01 2.81391084e-01
4.42656934e-01 -5.05298853e-01 7.51904547e-01 -1.13885033e+00
-1.18930030e+00 8.52920711e-01 2.95060575e-01 3.47575027e-04
5.94855130e-01 2.18397424e-01 -7.92453587e-01 1.12231091e-01
1.74844027e-01 2.71305144e-01 3.43079627e-01 -1.26190138e+00
-9.40476656e-01 -4.28552508e-01 7.72426486e-01 2.22321212e-01
-6.66224480e-01 -2.07099572e-01 -5.31407475e-01 -7.71406963e-02
3.76439184e-01 -7.75542557e-01 -5.83984256e-01 -3.69431853e-01
-4.03558016e-01 -4.88108873e-01 2.97403336e-01 -3.22033875e-02
1.45439065e+00 -1.84761894e+00 4.07776922e-01 9.10136998e-01
6.77303970e-01 -1.73373610e-01 -8.38881060e-02 6.88162386e-01
1.64409086e-01 3.19802761e-01 5.77503860e-01 3.03384751e-01
-5.01875952e-02 2.08663233e-02 4.03848529e-01 3.77936095e-01
-2.20650226e-01 1.21404409e-01 -1.35130620e+00 -6.18713498e-01
-2.02345937e-01 4.97162305e-02 -3.82052541e-01 -1.66007802e-01
2.49738201e-01 2.05613211e-01 -6.20840728e-01 4.21328694e-01
5.88242829e-01 -3.80199879e-01 7.37934411e-01 -3.08092296e-01
-2.41793558e-01 2.88567752e-01 -1.20023382e+00 1.03278542e+00
-2.32713938e-01 3.87004524e-01 2.13967368e-01 -8.52643251e-01
1.03129792e+00 1.48667306e-01 7.83394754e-01 -5.87646008e-01
2.36817017e-01 4.84161466e-01 2.75057077e-01 -8.91056284e-02
9.95966554e-01 -3.46233733e-02 9.00251940e-02 2.92330474e-01
-3.07875216e-01 2.96492148e-02 1.06732905e+00 7.97909677e-01
1.03754115e+00 -4.05749470e-01 1.54689699e-01 -4.71945852e-01
5.34499228e-01 -4.49430570e-02 5.35368323e-01 5.97664833e-01
-2.53949344e-01 1.71321318e-01 1.18368161e+00 -1.13645703e-01
-1.15824223e+00 -9.55282807e-01 -1.03011981e-01 9.80650663e-01
6.21380448e-01 -4.63795841e-01 -2.80721813e-01 -7.05117166e-01
8.35168213e-02 1.00419082e-01 -4.58291709e-01 -1.67508125e-01
-8.78854319e-02 -4.29455787e-01 -3.34692955e-01 9.70741082e-03
6.57872930e-02 -9.05426919e-01 -4.55810782e-03 2.02589795e-01
-3.09060276e-01 -4.79679823e-01 -2.31778339e-01 7.09178150e-02
-1.06772327e+00 -9.37683821e-01 -9.41016853e-01 -6.04262829e-01
9.98301446e-01 6.26990974e-01 1.57816505e+00 5.03965795e-01
3.99598598e-01 4.56607975e-02 -7.75145948e-01 5.87065294e-02
-2.19379172e-01 5.90160966e-01 2.95276493e-01 1.28113657e-01
1.77729383e-01 -4.84077185e-01 -5.86878061e-01 7.86816418e-01
-8.25426459e-01 -2.00726658e-01 5.08772612e-01 5.31651556e-01
1.56987578e-01 6.18102670e-01 7.13819444e-01 -1.18361270e+00
8.83721054e-01 -7.08359420e-01 -6.34950280e-01 2.76457906e-01
-1.07593691e+00 -3.59080359e-02 6.92787051e-01 -1.16326600e-01
-7.62788892e-01 -2.61951447e-01 1.09447114e-01 3.84027362e-01
5.81043422e-01 8.12430561e-01 -1.25977784e-01 1.33364554e-02
4.03664500e-01 -6.50030077e-01 -2.36058131e-01 -3.93309444e-01
1.65713891e-01 3.34944755e-01 -2.61062682e-01 -3.24010640e-01
9.11291897e-01 2.65542209e-01 3.79817903e-01 -3.71600956e-01
-5.89915276e-01 -7.32641280e-01 -6.06583893e-01 -8.09205770e-01
3.72330666e-01 -4.32549298e-01 -1.01129293e+00 3.67928445e-02
-9.08267200e-01 4.20997381e-01 -5.53273484e-02 5.91098011e-01
8.92681330e-02 2.94981360e-01 -5.32045960e-01 -6.98358476e-01
1.66430458e-01 -8.77949893e-01 3.00422013e-01 4.79267865e-01
-2.13783175e-01 -8.60476553e-01 1.53430074e-01 1.32548556e-01
3.93536925e-01 -1.01955840e-02 8.41791213e-01 -5.39596856e-01
-6.71901643e-01 -1.74985856e-01 -3.55888635e-01 9.70721692e-02
7.16653839e-02 3.97148222e-01 -2.32801020e-01 -4.06685799e-01
-6.21766567e-01 8.35984945e-02 7.64441073e-01 3.85349125e-01
4.42159235e-01 -1.17512457e-01 -5.00780821e-01 -2.25078285e-01
1.52539504e+00 3.08208704e-01 4.86882925e-01 4.14447248e-01
3.68487507e-01 1.18058991e+00 1.00076008e+00 5.44474423e-01
9.81928110e-02 7.74588645e-01 7.11668015e-01 -2.16693267e-01
4.68879014e-01 -2.52943903e-01 1.39554217e-01 1.07237148e+00
-3.63225937e-01 -4.31640804e-01 -5.03860593e-01 6.06656432e-01
-1.93329954e+00 -9.33076501e-01 -7.71515250e-01 2.49662828e+00
5.99451005e-01 6.26376033e-01 3.95664692e-01 2.86254048e-01
1.19412327e+00 1.50150001e-01 -4.62586395e-02 -3.32966000e-01
2.97527593e-02 -3.45004231e-01 5.52110791e-01 4.03888851e-01
-6.00367486e-01 4.46989417e-01 6.04659510e+00 7.16814458e-01
-6.73504591e-01 1.48388641e-02 7.62540102e-01 -1.36171281e-01
-7.77482808e-01 2.02733979e-01 -6.82596564e-01 4.31558132e-01
5.85049272e-01 -3.86074692e-01 2.51569450e-01 8.65595937e-01
2.59668589e-01 -3.93818825e-01 -8.31492543e-01 6.75209165e-01
-2.04059869e-01 -9.16522563e-01 -1.46678314e-01 1.77267730e-01
8.91905010e-01 -3.88550580e-01 -8.45401883e-02 2.49223653e-02
4.38908637e-01 -6.01287961e-01 5.01162112e-01 6.36861861e-01
2.82333881e-01 -1.03914177e+00 8.34237814e-01 5.97437322e-02
-1.14189124e+00 -1.35290520e-02 -6.55190051e-01 -3.09124738e-01
4.03483808e-01 1.31263912e+00 -3.38508666e-01 7.87920117e-01
7.87069619e-01 7.55644560e-01 -5.48819721e-01 1.32070732e+00
-2.01835170e-01 1.19591400e-01 -1.74435461e-03 -7.06392407e-01
1.13211557e-01 -8.93599451e-01 5.67354381e-01 3.56958300e-01
3.53694022e-01 -1.19536318e-01 -3.12338527e-02 4.33163702e-01
-2.54900306e-01 6.29584610e-01 -7.82902181e-01 -1.37256876e-01
2.20227823e-01 1.75062025e+00 -1.25232816e+00 -2.29527280e-01
-2.68734515e-01 5.84730327e-01 2.00837731e-01 4.43202034e-02
-7.08966494e-01 -4.78990763e-01 2.30660632e-01 5.79403102e-01
-3.78992379e-01 -7.51715079e-02 1.92801133e-01 -8.42236936e-01
-1.62397310e-01 -6.37794077e-01 3.40388030e-01 -6.64712846e-01
-1.56130016e+00 8.56988966e-01 -1.11191571e-01 -1.37810040e+00
3.67950767e-01 -3.91024381e-01 -2.78119743e-01 4.12354141e-01
-1.37757671e+00 -4.70609277e-01 -2.63260573e-01 3.57580543e-01
-3.58464941e-02 -3.80630866e-02 2.20520109e-01 8.68421733e-01
-2.48709723e-01 7.86615461e-02 4.25297230e-01 -2.98161834e-01
7.68874049e-01 -1.25578284e+00 -3.29940259e-01 4.90945131e-01
1.28674120e-01 7.26781905e-01 8.21758628e-01 -6.71751976e-01
-7.67904341e-01 -4.26069885e-01 1.58845496e+00 -1.54028893e-01
6.44380510e-01 -1.67363025e-02 -3.63658756e-01 -7.20090494e-02
2.10767180e-01 -4.77873564e-01 5.09149134e-01 9.70277548e-01
-1.49715051e-01 -7.09395528e-01 -9.89415169e-01 7.77292013e-01
1.11118388e+00 -1.80716559e-01 -1.26723871e-01 4.62547213e-01
1.00814909e-01 3.41979474e-01 -9.51341629e-01 3.21123362e-01
7.52949059e-01 -1.30349147e+00 6.15520000e-01 -4.91626501e-01
6.69081628e-01 -4.32992131e-01 3.66121143e-01 -1.62859774e+00
-7.83562005e-01 -5.11350095e-01 5.45441687e-01 1.44961190e+00
7.73114145e-01 -5.61650693e-01 1.04907250e+00 4.49504465e-01
3.16062689e-01 -5.50617695e-01 -6.89137697e-01 -7.89971054e-01
-4.25244629e-01 4.24026489e-01 4.75665897e-01 1.02120984e+00
3.42067569e-01 5.98614872e-01 -5.34603655e-01 -2.44305089e-01
3.31011891e-01 2.77455807e-01 4.86722261e-01 -2.06637692e+00
-4.37650457e-02 -6.36322737e-01 -4.62301344e-01 -6.98980749e-01
-3.48366797e-01 -7.97991514e-01 -2.11043715e-01 -1.89165378e+00
6.24362886e-01 -8.27614069e-01 -8.41766357e-01 -7.25036412e-02
6.59872293e-02 4.48634952e-01 2.66761392e-01 2.19037563e-01
-1.05322433e+00 1.05261430e-02 1.40201640e+00 -1.42187193e-01
-1.70748398e-01 5.76972425e-01 -8.47530365e-01 5.31161487e-01
7.23683119e-01 -8.09076726e-01 -3.56570393e-01 -4.27663960e-02
1.47253203e+00 2.30813444e-01 -1.52863845e-01 -6.63432121e-01
2.48431012e-01 -2.59686798e-01 -1.48795709e-01 -5.63439608e-01
8.59121457e-02 -1.17889118e+00 5.61932683e-01 5.80934286e-01
-5.16499162e-01 2.23074570e-01 -8.29996347e-01 4.65282232e-01
-3.86806726e-01 -7.02713490e-01 6.71248496e-01 -1.19733796e-01
-4.27250743e-01 2.54063785e-01 -4.27503675e-01 -4.01409388e-01
1.00090456e+00 -2.00538978e-01 -4.71335679e-01 -7.83514857e-01
-7.18414068e-01 2.33291164e-01 5.46919227e-01 5.61603069e-01
2.87061155e-01 -1.49365687e+00 -5.44396341e-01 -6.17232025e-01
2.35345960e-01 -6.75993025e-01 1.43751979e-01 1.02172339e+00
-3.52708519e-01 2.33442023e-01 -5.62144458e-01 -1.73966989e-01
-1.55054772e+00 4.79051054e-01 -5.01916036e-02 -7.48145938e-01
1.26314655e-01 8.35790217e-01 -1.08427539e-01 -2.11086199e-01
2.62858599e-01 1.71608731e-01 -9.76542473e-01 8.34435582e-01
-2.43310840e-03 9.28954780e-01 -9.22035724e-02 -6.19681001e-01
-4.08205092e-01 3.91426772e-01 -4.04887982e-02 -1.49074242e-01
1.27220547e+00 -2.69344956e-01 -5.73192477e-01 3.24975878e-01
9.00096595e-01 5.28314650e-01 -4.27392185e-01 -9.64949057e-02
2.23799199e-01 -7.59150863e-01 -1.38578072e-01 -3.89413387e-01
-1.51683176e+00 3.28381598e-01 1.93144217e-01 1.23320949e+00
1.06928420e+00 -1.18135093e-02 4.08466548e-01 3.03035855e-01
6.37500703e-01 -1.43498588e+00 2.23279238e-01 2.80853987e-01
4.69214350e-01 -9.10424232e-01 3.69127929e-01 -6.82228744e-01
-6.09191000e-01 1.07780135e+00 5.85572839e-01 -1.25055658e-02
8.60820115e-01 -3.51520807e-01 -1.78894266e-01 -4.15011197e-01
-5.68984389e-01 -3.84338737e-01 3.35464716e-01 2.85602361e-01
7.36800671e-01 1.35773763e-01 -1.48489654e+00 2.58968115e-01
9.39404890e-02 -2.94754893e-01 6.68023050e-01 3.82113934e-01
-5.49276471e-01 -1.69053972e+00 -1.28353491e-01 7.08223462e-01
-5.39262056e-01 -1.48120925e-01 -6.32694662e-01 4.20322359e-01
2.55819857e-01 1.31208205e+00 -1.41984224e-01 -5.09644866e-01
4.82983202e-01 -6.05113149e-01 2.34587207e-01 -4.07539546e-01
-9.05858874e-01 6.05976246e-02 4.48089749e-01 -2.91002274e-01
-8.48936617e-01 -5.69913030e-01 -7.07250655e-01 -6.93639398e-01
-7.61553764e-01 7.64757514e-01 5.22731662e-01 4.35378760e-01
7.85044506e-02 6.49857581e-01 1.13781095e+00 -2.62351453e-01
-2.53087342e-01 -6.53624296e-01 -1.36097550e+00 5.92254937e-01
-2.41743237e-01 -7.85194635e-01 -4.76507872e-01 -5.27510643e-01] | [9.498974800109863, 5.843824863433838] |
06c869ed-2674-4288-9d78-1ee24311f4e3 | spatio-temporal-point-process-for-multiple | 2302.02444 | null | https://arxiv.org/abs/2302.02444v1 | https://arxiv.org/pdf/2302.02444v1.pdf | Spatio-Temporal Point Process for Multiple Object Tracking | Multiple Object Tracking (MOT) focuses on modeling the relationship of detected objects among consecutive frames and merge them into different trajectories. MOT remains a challenging task as noisy and confusing detection results often hinder the final performance. Furthermore, most existing research are focusing on improving detection algorithms and association strategies. As such, we propose a novel framework that can effectively predict and mask-out the noisy and confusing detection results before associating the objects into trajectories. In particular, we formulate such "bad" detection results as a sequence of events and adopt the spatio-temporal point process}to model such events. Traditionally, the occurrence rate in a point process is characterized by an explicitly defined intensity function, which depends on the prior knowledge of some specific tasks. Thus, designing a proper model is expensive and time-consuming, with also limited ability to generalize well. To tackle this problem, we adopt the convolutional recurrent neural network (conv-RNN) to instantiate the point process, where its intensity function is automatically modeled by the training data. Furthermore, we show that our method captures both temporal and spatial evolution, which is essential in modeling events for MOT. Experimental results demonstrate notable improvements in addressing noisy and confusing detection results in MOT datasets. An improved state-of-the-art performance is achieved by incorporating our baseline MOT algorithm with the spatio-temporal point process model. | ['Xia Jia', 'Qian Xu', 'Zenghui Zhang', 'John See', 'Weiyao Lin', 'Kean Chen', 'Tao Wang'] | 2023-02-05 | null | null | null | null | ['multiple-object-tracking'] | ['computer-vision'] | [ 1.65022418e-01 -6.72635972e-01 -8.79178010e-03 -7.49296397e-02
-4.72752303e-01 -3.45789611e-01 6.03549123e-01 7.86499307e-02
-4.61279452e-01 3.26643407e-01 -1.96227189e-02 1.06376171e-01
-1.02412358e-01 -8.02684367e-01 -8.27127516e-01 -9.01193917e-01
-2.79659741e-02 4.97450709e-01 7.00132966e-01 1.04328863e-01
-8.71651694e-02 2.62477130e-01 -1.63203228e+00 1.82913616e-01
6.85465097e-01 9.18885171e-01 5.73268712e-01 7.25107551e-01
-1.39981464e-01 9.72307742e-01 -8.84894729e-01 -1.54991761e-01
1.12050563e-01 -4.32741880e-01 -2.40808114e-01 6.98417351e-02
2.62302995e-01 -2.19691738e-01 -6.48278654e-01 1.09549999e+00
3.12285870e-01 4.11559403e-01 5.45204580e-01 -1.31166613e+00
-4.59348589e-01 5.01370370e-01 -8.11422288e-01 7.31312752e-01
-9.37945247e-02 4.06414598e-01 7.83521056e-01 -7.08270788e-01
1.98839143e-01 1.30645537e+00 7.18597054e-01 5.04921138e-01
-1.14742124e+00 -8.79391909e-01 6.26863062e-01 3.67397279e-01
-1.53717434e+00 -6.92519769e-02 6.21025026e-01 -5.04257321e-01
5.46186745e-01 3.53829712e-01 9.23932374e-01 1.33789551e+00
9.55090225e-02 1.28240609e+00 4.62455541e-01 2.14876488e-01
-1.32368699e-01 -1.93736151e-01 1.23780496e-01 3.11555266e-01
2.21503407e-01 4.21885476e-02 -2.51457840e-01 9.93724242e-02
7.26103604e-01 5.41931570e-01 -1.50919631e-01 -5.49781881e-02
-1.29548466e+00 3.99540782e-01 4.74292338e-01 4.02601868e-01
-4.70326215e-01 4.99352574e-01 2.38829285e-01 -1.80566952e-01
3.86919081e-01 5.35601226e-04 -1.47638947e-01 -7.00628459e-02
-1.13714123e+00 3.40962440e-01 3.41048896e-01 1.01903522e+00
6.03069067e-01 -1.23783693e-01 -7.49789178e-01 5.66829145e-01
2.09166080e-01 6.25211477e-01 2.97471017e-01 -7.30194509e-01
3.69471163e-01 7.29963422e-01 4.85926390e-01 -1.25340855e+00
-3.85988295e-01 -6.45170212e-01 -9.31794107e-01 -2.45906577e-01
3.91833395e-01 -8.89611617e-02 -9.04136896e-01 1.82727718e+00
3.46106261e-01 1.07236874e+00 -2.09883496e-01 9.03775930e-01
7.42981374e-01 9.51719403e-01 4.21731651e-01 -4.64702010e-01
1.32027090e+00 -9.16321933e-01 -1.01768315e+00 -1.82494760e-01
3.36407930e-01 -5.15820265e-01 6.50553703e-01 1.31203920e-01
-9.03722823e-01 -7.22478092e-01 -7.40523398e-01 2.26950839e-01
-2.94414937e-01 2.76796371e-01 2.71240562e-01 1.09806776e-01
-7.37661541e-01 4.58456278e-01 -1.20562959e+00 -2.61454552e-01
5.33237457e-01 4.28585112e-01 2.01334238e-01 -1.51593285e-02
-9.36502278e-01 6.34123743e-01 4.56890732e-01 2.88293093e-01
-1.12082684e+00 -5.46899140e-01 -4.63196844e-01 3.06886509e-02
6.94196463e-01 -8.06778491e-01 1.00154972e+00 -5.21307290e-01
-8.31119597e-01 4.61306870e-01 -5.08312047e-01 -6.27526700e-01
6.69922233e-01 -3.18147361e-01 -6.04219437e-01 -1.77481323e-01
4.08488125e-01 5.92684627e-01 9.26888764e-01 -1.21707082e+00
-1.23263729e+00 -6.76857606e-02 -1.06411740e-01 1.35089025e-01
-3.64379704e-01 2.12725475e-01 -1.00089169e+00 -7.27102041e-01
6.06264174e-03 -9.16404247e-01 -3.00437778e-01 -2.05901653e-01
-3.46408218e-01 -6.21547163e-01 1.28269923e+00 -3.58977705e-01
1.50470746e+00 -2.26677966e+00 2.06369206e-01 -3.64980102e-01
3.54010820e-01 3.03086370e-01 6.22672886e-02 9.05817375e-02
4.27384585e-01 2.86925305e-02 -1.37312025e-01 -6.58427358e-01
-1.04309350e-01 3.70811284e-01 -3.52322042e-01 5.89800775e-01
3.99336249e-01 9.34200406e-01 -1.20201015e+00 -5.39242744e-01
5.88027596e-01 6.76552236e-01 -1.88611969e-01 2.87161499e-01
-3.75186503e-01 8.75224054e-01 -5.23568809e-01 6.27356768e-01
6.91921771e-01 -5.42316318e-01 -2.08427176e-01 -2.59428859e-01
-2.97448844e-01 2.10937038e-02 -1.25423765e+00 1.40512216e+00
-8.68157148e-02 5.34994483e-01 -1.02883138e-01 -8.57922852e-01
7.30717361e-01 2.29386747e-01 8.68977785e-01 -4.49737996e-01
1.91803917e-01 -1.08057112e-02 -9.24650952e-02 -5.58065832e-01
6.97183788e-01 6.07667640e-02 -1.99565347e-02 3.72043364e-02
-2.45314911e-01 5.80109298e-01 3.65006983e-01 -8.44998285e-03
1.17887604e+00 1.56619072e-01 -4.67109047e-02 1.48644105e-01
3.98306638e-01 1.08050313e-02 1.03304052e+00 1.09471965e+00
-4.62387532e-01 5.90170145e-01 1.86071321e-02 -3.88226748e-01
-6.17317975e-01 -1.02454436e+00 -1.03339911e-01 9.97050941e-01
7.40622997e-01 -2.86713272e-01 -3.73406082e-01 -6.01667762e-01
-2.12508202e-01 5.58704495e-01 -7.09863126e-01 -1.52885512e-01
-1.13462257e+00 -1.19793451e+00 5.42173386e-01 6.70604944e-01
5.37417531e-01 -1.07242656e+00 -7.18136907e-01 6.08557582e-01
-5.02812922e-01 -1.24312103e+00 -5.51584601e-01 8.53935853e-02
-4.57793444e-01 -1.17349374e+00 -7.05128491e-01 -5.40317416e-01
3.72302592e-01 6.39331818e-01 9.51903880e-01 1.85104921e-01
-1.81600466e-01 1.09378673e-01 -3.51929277e-01 -5.39838850e-01
-3.52874584e-02 -1.24168895e-01 -2.49866508e-02 3.68819803e-01
4.74322587e-01 -3.35297108e-01 -8.48461092e-01 5.55045426e-01
-9.74581063e-01 -3.52678038e-02 5.32879472e-01 6.14583731e-01
7.41817594e-01 3.53548646e-01 3.29253852e-01 -3.09591025e-01
1.49078071e-01 -7.21846282e-01 -4.97885376e-01 1.77861318e-01
-2.71876366e-03 -3.62111002e-01 5.50673127e-01 -7.93392777e-01
-1.03811383e+00 1.59125224e-01 1.74223766e-01 -9.32013869e-01
-2.08146229e-01 7.32246339e-02 -7.17239901e-02 2.83665299e-01
2.16086820e-01 2.96941727e-01 -5.18251240e-01 -4.04562116e-01
1.62996873e-01 2.59609520e-01 7.28463650e-01 -4.01762605e-01
7.07634211e-01 9.16007578e-01 -4.98260790e-03 -5.96540630e-01
-1.05855095e+00 -1.01045859e+00 -5.16138017e-01 -5.51498413e-01
1.01175284e+00 -1.07048571e+00 -7.99780846e-01 3.26615661e-01
-1.28477907e+00 -1.19889364e-01 -1.78102329e-01 5.24968863e-01
-2.50758171e-01 1.96017265e-01 -5.98992407e-01 -1.13333404e+00
-3.78801525e-02 -1.25255525e+00 1.42493415e+00 4.22942907e-01
-1.79959834e-02 -8.36761773e-01 -1.16116874e-01 1.52986050e-01
1.10816807e-01 4.04635757e-01 2.58403987e-01 -4.72463518e-01
-1.08776212e+00 -2.34910652e-01 -3.71295005e-01 -2.03023732e-01
2.26664439e-01 -3.53674851e-02 -7.93960452e-01 -2.12275594e-01
9.31502879e-02 2.86493927e-01 1.21014667e+00 7.45890677e-01
1.23661005e+00 -5.38734794e-02 -1.04673958e+00 5.51636755e-01
1.15743351e+00 2.54954368e-01 2.95648009e-01 2.41104022e-01
1.12990046e+00 4.92869347e-01 8.37417066e-01 4.21352416e-01
4.91842777e-01 9.08709586e-01 4.96899188e-01 5.65553233e-02
-1.96235970e-01 -2.95424461e-01 3.10756981e-01 5.91859341e-01
-2.03207120e-01 -6.58357441e-01 -7.31124103e-01 8.31305861e-01
-2.55550766e+00 -1.34271669e+00 -5.51866770e-01 2.03575468e+00
4.61443365e-01 3.81232537e-02 3.03228348e-01 3.00931837e-03
1.06637645e+00 2.22043306e-01 -6.34155691e-01 6.78272843e-01
-2.27789968e-01 -3.80381405e-01 4.78343755e-01 7.05446228e-02
-1.60052192e+00 9.91390467e-01 5.49477339e+00 9.35452104e-01
-1.03158712e+00 2.37851471e-01 1.71141297e-01 -4.42225665e-01
2.38700092e-01 -2.75964856e-01 -1.21898484e+00 8.38428617e-01
7.82869101e-01 7.58645162e-02 1.81995496e-01 4.54580277e-01
6.54714406e-01 1.78208947e-01 -8.43037128e-01 1.03170574e+00
-2.41279379e-01 -1.14304483e+00 1.67690795e-02 -1.55280565e-03
6.00000858e-01 2.81187836e-02 1.02309674e-01 5.53374112e-01
4.11187500e-01 -8.44975352e-01 9.08186853e-01 7.89769411e-01
1.87184513e-01 -6.51994109e-01 6.24334753e-01 7.06125200e-01
-1.78297246e+00 -2.80148178e-01 -3.99255306e-01 6.85620010e-02
4.93310511e-01 4.76313055e-01 -5.96600533e-01 5.92996955e-01
1.03993392e+00 1.31080449e+00 -4.71929401e-01 1.42559254e+00
8.04130808e-02 6.43084288e-01 -5.12850702e-01 1.22023234e-03
2.78593510e-01 -1.22768693e-01 9.60943878e-01 1.45064735e+00
4.91915017e-01 -1.02503143e-01 6.68462157e-01 1.04308426e+00
2.34205928e-02 -3.01586360e-01 -5.40209472e-01 1.87682852e-01
3.96261901e-01 1.19998848e+00 -1.09567213e+00 -4.94443744e-01
-4.37172920e-01 6.90532863e-01 1.71542585e-01 3.47017825e-01
-1.32964647e+00 2.86425918e-01 6.93727255e-01 3.24780904e-02
7.05746412e-01 -1.08188756e-01 -1.47582099e-01 -1.16223931e+00
1.13883063e-01 -3.44047785e-01 5.57942390e-01 -4.89694804e-01
-1.30563581e+00 4.48631078e-01 1.09312020e-01 -1.42899704e+00
6.70379028e-02 -1.44297555e-01 -5.87654412e-01 5.68863451e-01
-1.46879900e+00 -1.24296391e+00 -4.77923602e-01 4.66614038e-01
8.22641969e-01 3.47085118e-01 2.17494816e-01 8.21572602e-01
-1.01179385e+00 2.73142159e-01 -9.95974988e-02 1.86171860e-01
4.60647374e-01 -1.09152675e+00 3.14884871e-01 1.00763500e+00
2.40711913e-01 4.36015010e-01 8.76988411e-01 -9.60632741e-01
-1.11482537e+00 -1.82476473e+00 5.32801986e-01 -7.63476491e-01
7.02443600e-01 -2.38098040e-01 -1.10604310e+00 8.50949943e-01
-6.31795898e-02 3.96137327e-01 2.74567753e-01 -2.55940318e-01
1.52465776e-01 4.79336195e-02 -7.34656692e-01 6.70233548e-01
1.40010929e+00 -1.68206319e-01 -3.88917983e-01 3.54491740e-01
7.30511069e-01 -4.00229335e-01 -4.79467124e-01 6.76279724e-01
1.76081836e-01 -7.69053042e-01 1.19945657e+00 -3.65944862e-01
-2.33063027e-02 -9.02858615e-01 -5.10237217e-02 -9.04540539e-01
-4.91580606e-01 -3.91432166e-01 -6.52649462e-01 1.21004033e+00
6.01202890e-04 -1.00400500e-01 7.61327624e-01 1.64517730e-01
-2.21672997e-01 -6.78327501e-01 -8.34785521e-01 -8.39989305e-01
-3.66115719e-01 -6.35168076e-01 5.56556404e-01 7.59811819e-01
-6.55962527e-01 1.30478159e-01 -7.81497598e-01 7.27831304e-01
7.12759018e-01 1.41326800e-01 7.49288499e-01 -1.23693383e+00
-1.85878575e-01 -5.14504850e-01 -3.13604355e-01 -1.54141808e+00
-3.82995941e-02 -5.50568521e-01 4.92267579e-01 -1.47601032e+00
3.29689980e-01 -5.24557054e-01 -6.35836124e-01 1.90201059e-01
-6.21949494e-01 1.84018329e-01 2.51174092e-01 6.80141568e-01
-1.24962449e+00 6.87219203e-01 1.20877671e+00 -3.65771294e-01
-3.45107973e-01 2.98535168e-01 -1.87229380e-01 7.54465044e-01
4.56244379e-01 -7.61039019e-01 -2.06465855e-01 -4.39051539e-01
1.30806323e-02 9.27167162e-02 8.20913970e-01 -1.22620988e+00
6.15331173e-01 -3.01673681e-01 4.29963619e-01 -1.22953558e+00
7.11146355e-01 -9.41286623e-01 3.31919402e-01 3.50777626e-01
-1.61562279e-01 -1.47608593e-02 1.62773073e-01 1.26041281e+00
-1.12486631e-01 -2.43744371e-03 4.83913898e-01 -2.82450825e-01
-8.48607302e-01 6.85340464e-01 -4.44961131e-01 -1.06095791e-01
1.21388388e+00 -9.25181508e-02 -1.11569352e-01 -1.75490379e-01
-7.01116800e-01 6.25078321e-01 1.55274197e-01 5.76868415e-01
4.83020157e-01 -1.23647773e+00 -5.19952655e-01 -1.83065683e-01
2.81900447e-02 3.45524639e-01 2.92663068e-01 1.05450010e+00
1.02691054e-01 2.14760721e-01 3.80313575e-01 -1.10901594e+00
-1.20533657e+00 6.73211873e-01 3.97951633e-01 -3.03291321e-01
-9.15946841e-01 7.40588009e-01 5.22918403e-01 -3.78676727e-02
3.70155007e-01 -4.77157921e-01 -5.37997723e-01 3.20983827e-02
5.89044154e-01 4.40657705e-01 -3.87322128e-01 -8.48562241e-01
-2.98353314e-01 4.92723286e-01 1.25074955e-02 2.64229208e-01
1.11517382e+00 -3.16134155e-01 2.30459109e-01 7.26178467e-01
8.19091022e-01 -3.67082745e-01 -1.57955587e+00 -2.61805624e-01
2.35957075e-02 -3.51273984e-01 -3.31966095e-02 -4.35069114e-01
-1.00445104e+00 6.63123012e-01 6.11357093e-01 4.56327170e-01
8.52837026e-01 1.26755640e-01 1.05482459e+00 1.12273879e-01
2.88096696e-01 -8.10440063e-01 9.10128728e-02 4.71476823e-01
4.75583434e-01 -1.28051221e+00 -3.58829856e-01 -4.90803361e-01
-3.01883012e-01 7.19231904e-01 7.65045583e-01 -1.00488365e-01
5.57230115e-01 1.87944308e-01 -1.31653026e-01 -3.24428916e-01
-6.26434326e-01 -5.53730667e-01 3.07200670e-01 3.55858058e-01
7.52046853e-02 1.36274517e-01 2.84363586e-03 5.63075602e-01
1.56284198e-01 -3.42154317e-02 1.89781800e-01 8.83099854e-01
-4.58250910e-01 -7.59476483e-01 -5.91276765e-01 5.07475734e-01
-5.53335905e-01 1.67939544e-01 -3.55750397e-02 7.13614047e-01
6.61919236e-01 1.20815277e+00 3.93056482e-01 -4.90693212e-01
4.08257812e-01 -2.38722190e-01 2.04349130e-01 -5.05713582e-01
-5.84384322e-01 4.34040606e-01 -2.68630415e-01 -2.71204650e-01
-7.42524505e-01 -8.65462363e-01 -1.29402697e+00 -1.77426115e-01
-5.29155970e-01 -1.72114089e-01 1.73176169e-01 1.13386786e+00
3.41370195e-01 1.10752845e+00 1.74005613e-01 -8.84705663e-01
-1.51308134e-01 -9.81581748e-01 -2.37566561e-01 6.03862226e-01
6.54322863e-01 -1.07734895e+00 -1.63030580e-01 7.12924125e-03] | [6.268341541290283, -2.0895845890045166] |
a9c826c0-646e-4adb-ac83-d16d049e09b9 | beyond-sift-using-binary-features-for-loop | 1709.05833 | null | http://arxiv.org/abs/1709.05833v1 | http://arxiv.org/pdf/1709.05833v1.pdf | Beyond SIFT using Binary features for Loop Closure Detection | In this paper a binary feature based Loop Closure Detection (LCD) method is
proposed, which for the first time achieves higher precision-recall (PR)
performance compared with state-of-the-art SIFT feature based approaches. The
proposed system originates from our previous work Multi-Index hashing for Loop
closure Detection (MILD), which employs Multi-Index Hashing
(MIH)~\cite{greene1994multi} for Approximate Nearest Neighbor (ANN) search of
binary features. As the accuracy of MILD is limited by repeating textures and
inaccurate image similarity measurement, burstiness handling is introduced to
solve this problem and achieves considerable accuracy improvement.
Additionally, a comprehensive theoretical analysis on MIH used in MILD is
conducted to further explore the potentials of hashing methods for ANN search
of binary features from probabilistic perspective. This analysis provides more
freedom on best parameter choosing in MIH for different application scenarios.
Experiments on popular public datasets show that the proposed approach achieved
the highest accuracy compared with state-of-the-art while running at 30Hz for
databases containing thousands of images. | ['Lan Xu', 'Lu Fang', 'Guyue Zhou', 'Lei Han'] | 2017-09-18 | null | null | null | null | ['loop-closure-detection'] | ['computer-vision'] | [-6.82027952e-04 -6.64249718e-01 -4.63648200e-01 -1.52939364e-01
-1.09781420e+00 -3.71147484e-01 5.62592328e-01 8.55816960e-01
-4.83565658e-01 4.36025470e-01 -2.60235928e-02 -2.42388114e-01
-2.88530558e-01 -8.86127830e-01 -5.86973548e-01 -5.12608349e-01
-4.02175695e-01 1.05347224e-01 6.28819823e-01 -1.08774245e-01
9.73739743e-01 6.34954751e-01 -2.20410275e+00 1.18036658e-01
3.48054975e-01 1.22708273e+00 1.88094065e-01 6.63878024e-01
2.74382919e-01 4.26096618e-01 -4.76208419e-01 -2.50173301e-01
2.87066430e-01 1.45638511e-01 -6.22858942e-01 -4.64810610e-01
5.68764269e-01 -6.34497941e-01 -5.83940566e-01 8.43555629e-01
6.42714620e-01 1.50158012e-03 7.83685565e-01 -1.39507461e+00
-5.93534112e-01 3.14121872e-01 -6.95200145e-01 4.09643710e-01
8.88590515e-01 -2.22899854e-01 1.00464749e+00 -1.09012496e+00
5.56092024e-01 1.17919672e+00 1.03953910e+00 -2.70582944e-01
-1.06040847e+00 -5.96900403e-01 -5.39692223e-01 5.00463486e-01
-2.05488205e+00 -9.13996473e-02 5.07544518e-01 -7.64001384e-02
1.17957211e+00 2.93479294e-01 6.35416985e-01 3.16919327e-01
6.44836545e-01 7.93280184e-01 1.02814221e+00 -6.76577330e-01
1.43427625e-01 9.92238149e-02 3.94408226e-01 8.82479429e-01
6.72997177e-01 2.86418468e-01 -8.39931250e-01 -6.59786761e-01
5.44352949e-01 -7.10841129e-03 -3.93290110e-02 -4.93071884e-01
-1.23917866e+00 9.75716710e-01 5.37763596e-01 2.92907298e-01
-2.42242545e-01 2.26857975e-01 6.72873855e-01 3.52148503e-01
-6.49411231e-02 4.44385782e-02 -2.08056681e-02 -2.11233765e-01
-1.20931911e+00 5.89157879e-01 6.18398488e-01 1.04757667e+00
8.91685426e-01 -5.12603223e-01 -3.68852258e-01 4.30163503e-01
5.05555987e-01 5.60749352e-01 5.26361465e-01 -6.12999082e-01
2.15176016e-01 5.78259945e-01 1.69326439e-01 -1.47926164e+00
-2.25097910e-01 9.75200441e-03 -6.95179760e-01 7.34886006e-02
1.39888510e-01 7.19940305e-01 -5.72771430e-01 1.09999096e+00
4.90212649e-01 -4.63518538e-02 4.95136529e-02 6.15098119e-01
6.60606325e-01 6.45372629e-01 -1.63257375e-01 1.66512672e-02
1.55803704e+00 -5.02176046e-01 -5.71113229e-01 3.60963464e-01
7.72560894e-01 -1.43039715e+00 8.10584068e-01 4.27955031e-01
-5.71161091e-01 -6.51217222e-01 -1.46423316e+00 -6.10089228e-02
-4.40288395e-01 2.69938082e-01 7.28965878e-01 6.80873275e-01
-1.07347560e+00 5.51533282e-01 -8.02558005e-01 -7.24187493e-01
5.32040857e-02 5.89652121e-01 -3.48627836e-01 -1.96545050e-01
-1.07156968e+00 7.48984754e-01 3.66570830e-01 -3.23848993e-01
-4.91232038e-01 -3.39542687e-01 -9.20864940e-01 -3.30429077e-01
-1.32202208e-01 -2.33138964e-01 1.01705921e+00 1.41332686e-01
-9.14010882e-01 7.59860754e-01 -2.24545673e-01 -8.72306943e-01
2.50502259e-01 -3.15119803e-01 -1.77523375e-01 6.08164787e-01
1.29897311e-01 7.39648998e-01 8.94904554e-01 -8.14571381e-01
-7.13413119e-01 -2.84709394e-01 -2.43888035e-01 -3.50620262e-02
-3.16010147e-01 -4.26959023e-02 -1.88862398e-01 -7.53950298e-01
3.81815284e-01 -9.67791557e-01 2.07731828e-01 2.63133377e-01
-1.82106674e-01 -4.64816272e-01 1.34312057e+00 -2.59042650e-01
1.66044438e+00 -2.25133729e+00 -4.25509572e-01 4.15633082e-01
-1.36630863e-01 4.06549007e-01 2.38561869e-01 1.02410471e+00
4.14024323e-01 -1.88856006e-01 1.29447728e-01 -1.85256436e-01
1.41153634e-01 -5.67142898e-03 -3.86597097e-01 1.04135764e+00
-2.29872003e-01 3.69652718e-01 -6.04588330e-01 -1.11194289e+00
6.33951843e-01 7.52468169e-01 -4.57436413e-01 7.32378587e-02
3.19583446e-01 -2.87949711e-01 -4.75446492e-01 1.05479610e+00
9.21428502e-01 -1.03675514e-01 -3.08982044e-01 -4.73774612e-01
-3.08559686e-01 -5.98450787e-02 -1.47857308e+00 1.57915890e+00
-1.09019816e-01 3.66977781e-01 -5.10597646e-01 -7.72871733e-01
1.23472726e+00 1.96082324e-01 4.16908026e-01 -5.26234090e-01
1.41361922e-01 4.19192970e-01 -5.43015420e-01 -3.76671165e-01
1.06363046e+00 5.90686977e-01 -2.65366763e-01 2.71008164e-01
-1.05101652e-01 -1.05113164e-01 1.22405156e-01 2.28569612e-01
1.09487450e+00 7.25738630e-02 6.48792922e-01 -5.35360456e-01
7.73743331e-01 2.84087688e-01 1.92833260e-01 1.07074964e+00
-6.58560932e-01 5.52561283e-01 -1.55483931e-01 -6.02933228e-01
-8.97242069e-01 -9.53366041e-01 -5.58499634e-01 7.03119755e-01
7.15290070e-01 -7.82433391e-01 -5.62791526e-01 -1.33923888e-01
4.09336746e-01 8.38948563e-02 -3.96325320e-01 -2.34436728e-02
-4.19733226e-01 -6.21706188e-01 8.01443815e-01 2.55496174e-01
9.54789102e-01 -7.80359387e-01 -1.16168690e+00 2.07512751e-01
-2.35280052e-01 -8.47561002e-01 -2.87027895e-01 1.73944421e-02
-8.38235021e-01 -1.07878327e+00 -4.98967171e-01 -9.78636801e-01
3.24042678e-01 5.64823270e-01 7.97262371e-01 3.27791125e-01
-8.17615747e-01 4.16691214e-01 -7.02797353e-01 -8.75616372e-02
-1.80119693e-01 8.86321142e-02 2.18247354e-01 -2.98431545e-01
7.34669387e-01 -2.90659547e-01 -9.70164299e-01 2.84620643e-01
-9.29610193e-01 -6.15707636e-01 8.35767746e-01 8.14015448e-01
7.76242435e-01 2.19119668e-01 1.15871698e-01 -1.42993912e-01
5.14965773e-01 -2.93529719e-01 -8.10848951e-01 1.68360621e-01
-9.28334653e-01 4.47422177e-01 2.76725113e-01 -3.79368126e-01
-5.13081729e-01 2.31244251e-01 5.37117720e-02 -2.75425255e-01
9.12742093e-02 1.93310335e-01 4.43179041e-01 -7.96577215e-01
6.80759251e-01 5.75084627e-01 2.30819881e-02 -3.01249236e-01
3.97372097e-02 1.05191588e+00 6.71343863e-01 -4.27081198e-01
6.35691941e-01 5.20104468e-01 3.72865766e-01 -8.48721087e-01
-2.10378960e-01 -9.75038350e-01 -5.60683072e-01 1.01767220e-01
3.29721361e-01 -1.03173184e+00 -1.00579774e+00 5.76704204e-01
-8.80537570e-01 4.83560771e-01 2.30030224e-01 5.73302150e-01
-6.61341250e-01 6.50235295e-01 -5.82481503e-01 -9.08233345e-01
-5.95539689e-01 -1.23342621e+00 1.39457619e+00 7.93606490e-02
-1.90159604e-01 -4.35409307e-01 1.71555996e-01 2.29810141e-02
3.61931592e-01 2.35278338e-01 6.89572513e-01 -4.26666498e-01
-7.82779694e-01 -6.68655634e-01 -3.65318179e-01 -1.03593908e-01
-1.13026008e-01 1.00867391e-01 -6.18314385e-01 -6.23371542e-01
-3.10894489e-01 -3.34406674e-01 6.42677367e-01 3.18506777e-01
7.65570581e-01 -3.04477066e-01 -5.67451358e-01 3.52392554e-01
2.03709412e+00 -1.83392778e-01 7.49331892e-01 8.37069750e-01
3.08716390e-02 3.13641243e-02 1.46839666e+00 1.05323172e+00
5.66147625e-01 8.00167799e-01 4.25483674e-01 2.37733275e-01
-1.25562087e-01 -3.33776295e-01 1.61571205e-01 4.53561485e-01
3.87833834e-01 -7.78943971e-02 -5.99667549e-01 8.68461072e-01
-1.76999676e+00 -9.94773805e-01 -8.69401991e-02 2.56440496e+00
5.87429047e-01 9.20431614e-02 1.84298649e-01 8.44157398e-01
9.30098236e-01 1.29625395e-01 -1.12581447e-01 -4.48281229e-01
1.14140250e-01 1.96222708e-01 9.37601626e-01 3.35607141e-01
-1.40394711e+00 6.53769016e-01 6.13007116e+00 1.13254619e+00
-9.28633213e-01 -1.09923854e-01 8.08682144e-02 4.88254875e-01
1.31919429e-01 2.36427099e-01 -1.33579445e+00 3.73726487e-01
7.30021060e-01 -1.23932861e-01 8.84976313e-02 8.48865569e-01
-2.22271129e-01 -6.77491784e-01 -7.00801909e-01 1.26787889e+00
1.03857189e-01 -1.16488481e+00 2.64382869e-01 1.34057835e-01
4.61159796e-01 2.61349175e-02 1.53280362e-01 9.38979071e-03
-1.61244616e-01 -6.08994007e-01 6.37502551e-01 3.55274230e-01
7.25532293e-01 -1.00717080e+00 9.13139999e-01 4.04864214e-02
-1.75796199e+00 -4.76996116e-02 -5.98719478e-01 6.97349310e-02
-2.70212829e-01 5.56884587e-01 -8.85865629e-01 4.14359927e-01
1.02164984e+00 4.14799333e-01 -7.90083647e-01 1.35096061e+00
3.89645964e-01 1.76717252e-01 -8.04608345e-01 -4.47414130e-01
2.36134887e-01 3.30629349e-01 4.62635994e-01 1.29365230e+00
5.53221107e-01 -8.28942284e-02 1.96111158e-01 2.23309204e-01
4.23225790e-01 4.31799889e-01 -8.79203677e-01 3.31581861e-01
1.01688647e+00 1.00546002e+00 -7.61553824e-01 -1.54819980e-01
-3.26387346e-01 8.15758944e-01 1.09122591e-02 -2.61460453e-01
-7.64931023e-01 -1.02530193e+00 3.40557069e-01 2.68732101e-01
7.31276333e-01 -3.98305506e-01 9.49149132e-02 -7.76738167e-01
1.10954441e-01 -5.88689804e-01 5.77490926e-01 -2.95125782e-01
-7.39344120e-01 4.91132170e-01 2.99262822e-01 -1.76595199e+00
-3.37587982e-01 -2.87607700e-01 -1.04196399e-01 3.23121220e-01
-1.49715090e+00 -1.18433499e+00 -2.08029136e-01 7.34717250e-01
2.53488839e-01 7.86216110e-02 8.71824205e-01 2.73944408e-01
6.87054098e-02 1.05679572e+00 3.78937066e-01 -2.27109268e-01
8.44193578e-01 -7.58203983e-01 2.27556393e-01 7.68785179e-01
5.37911355e-02 6.60587072e-01 9.16627824e-01 -7.05572426e-01
-1.88960850e+00 -8.85570705e-01 1.23487508e+00 8.52608755e-02
5.24134338e-01 -6.49038106e-02 -6.33775532e-01 2.05264494e-01
-1.68660313e-01 3.62364024e-01 4.46235895e-01 -5.04975200e-01
-4.43693936e-01 -2.52817631e-01 -1.74378157e+00 2.25701779e-01
7.03381360e-01 -6.85625553e-01 -6.43974423e-01 1.82635546e-01
4.28178012e-01 -3.43830913e-01 -1.23215187e+00 6.84227765e-01
7.86946476e-01 -1.30985117e+00 1.17477500e+00 5.22790909e-01
-1.89627275e-01 -6.18041337e-01 -8.15001607e-01 -2.96232551e-01
-9.97143090e-02 -5.62564671e-01 -3.17812353e-01 1.17056525e+00
-1.64906099e-01 -4.02921945e-01 4.47092623e-01 4.91319597e-03
3.29308599e-01 -7.71333337e-01 -1.09312701e+00 -8.63840222e-01
-5.00035226e-01 -1.86284900e-01 5.92677534e-01 3.04558963e-01
1.12660661e-01 -2.47306466e-01 -2.65089899e-01 4.04079765e-01
1.06018054e+00 3.50766867e-01 7.49154925e-01 -1.11588144e+00
2.38465294e-02 5.29548749e-02 -1.26807821e+00 -8.43123913e-01
-2.86965877e-01 -4.93070245e-01 -9.77827236e-02 -1.00734055e+00
2.49890491e-01 -6.11598074e-01 -2.59501129e-01 2.58101851e-01
1.35670230e-01 7.36806631e-01 7.34706670e-02 5.29449224e-01
-7.16145158e-01 4.22556430e-01 6.38715863e-01 -6.83304518e-02
-8.84912163e-02 -7.91601837e-02 -1.32800892e-01 2.49389395e-01
4.12316829e-01 -6.80157185e-01 -1.83965281e-01 8.77994522e-02
-1.08186804e-01 3.67932357e-02 5.18265426e-01 -1.45331204e+00
4.70241606e-01 3.78817081e-01 2.87951648e-01 -1.22068727e+00
2.53528357e-01 -8.43306780e-01 4.23106030e-02 9.33699787e-01
-9.05499980e-02 5.64071238e-01 1.62602976e-01 7.89083064e-01
-3.11312497e-01 -2.97846407e-01 7.35190034e-01 9.30797458e-02
-9.19566751e-01 -3.44291283e-03 -3.17536473e-01 -4.06079561e-01
1.05647218e+00 -4.49170858e-01 -3.33104521e-01 -2.49007747e-01
-1.11431785e-01 -1.58219337e-01 7.89124131e-01 3.11103702e-01
9.00845408e-01 -1.47638834e+00 -3.87337416e-01 4.10066545e-01
5.70095241e-01 -4.91400063e-01 -1.31865125e-02 7.13654995e-01
-9.98301387e-01 7.19957530e-01 -1.31174281e-01 -1.02809763e+00
-1.46827984e+00 7.06283092e-01 -2.80632794e-01 -3.22841346e-01
-6.34237111e-01 4.65298712e-01 -6.73182786e-01 -4.84073907e-02
4.73149359e-01 -3.30657780e-01 1.05381399e-01 -2.37807214e-01
5.38593829e-01 5.28126776e-01 3.39894712e-01 -8.42716515e-01
-6.14655137e-01 9.61580515e-01 -2.66298503e-01 -9.86349955e-02
9.09842014e-01 -3.28111798e-01 -3.22315991e-02 2.35344961e-01
1.51159239e+00 -2.42242426e-01 -7.20511496e-01 -2.04605490e-01
-4.63048853e-02 -8.34591150e-01 2.06169561e-01 -1.73359722e-01
-4.10754919e-01 4.70430255e-01 1.21639371e+00 -5.60205132e-02
1.05680346e+00 -3.41717809e-01 1.06627727e+00 5.67705691e-01
9.78008211e-01 -1.02235818e+00 -8.30606669e-02 1.56167179e-01
4.88638759e-01 -1.33337379e+00 5.43213904e-01 -2.25736141e-01
-1.37627169e-01 1.13727605e+00 1.86708495e-01 -5.80872715e-01
8.84149194e-01 1.77886426e-01 -3.40935498e-01 -5.85658774e-02
-3.58148962e-01 -1.16780818e-01 1.17350698e-01 4.16110367e-01
1.64985806e-01 -2.27122143e-01 -8.05220783e-01 -1.30399108e-01
-1.61447048e-01 1.25281751e-01 1.77893177e-01 1.64776003e+00
-6.74583554e-01 -1.18288279e+00 -8.06599259e-01 3.91146004e-01
-5.82495272e-01 -8.69083554e-02 1.38269022e-01 1.05374420e+00
-1.54908925e-01 1.00520086e+00 -7.94961378e-02 -6.56224728e-01
-9.19868350e-02 -3.07821572e-01 6.70436919e-01 1.00224972e-01
-5.55175066e-01 -1.72603622e-01 -4.20803845e-01 -7.19194293e-01
-6.37744546e-01 -7.46575773e-01 -1.24817491e+00 -4.91157293e-01
-6.00909710e-01 3.24514925e-01 9.11028087e-01 5.82654476e-01
5.75692773e-01 -3.55448037e-01 7.98159063e-01 -7.80117393e-01
-8.18065226e-01 -7.04300940e-01 -7.01525390e-01 2.50597566e-01
5.58134675e-01 -9.43275392e-01 -4.65427607e-01 -7.37809837e-02] | [7.502941608428955, -1.9274152517318726] |
613735f6-2c38-439a-9ba0-9e0de41647a6 | spatial-reuse-in-dense-wireless-areas-a-cross | 2202.05655 | null | https://arxiv.org/abs/2202.05655v1 | https://arxiv.org/pdf/2202.05655v1.pdf | Spatial Reuse in Dense Wireless Areas: A Cross-layer Optimization Approach via ADMM | This paper introduces an efficient method for communication resource use in dense wireless areas where all nodes must communicate with a common destination node. The proposed method groups nodes based on their \newt{distance from the destination} and creates a structured multi-hop configuration in which each group can relay its neighbor's data. \newt{The large number of active radio nodes and the common direction of communication toward a single destination are exploited to reuse the limited spectrum resources in spatially separated groups}. Spectrum allocation constraints among groups are then embedded in a joint routing and resource allocation framework to optimize the route and amount of resources allocated to each node. \newt{The solution to this problem uses coordination among the lower-layers of the wireless-network protocol stack to outperform conventional approaches where these layers are decoupled. Furthermore, the structure of this problem is exploited to obtain} a semi-distributed optimization algorithm based on the alternating direction method of multipliers (ADMM) where each node can optimize its resources independently based on local channel information. | ['Ghadah Aldabbagh', 'John M. Cioffi', 'Golnaz Farhadi', 'Borja Peleato', 'Haleh Tabrizi'] | 2022-02-11 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [ 3.35244894e-01 7.25996494e-01 -7.01378047e-01 -1.90795928e-01
-2.48653412e-01 -4.49451715e-01 -2.49511991e-02 -6.67244643e-02
-6.77854538e-01 1.34961820e+00 -1.61740467e-01 -3.58030081e-01
-6.42218471e-01 -9.88609850e-01 5.69199547e-02 -1.05378914e+00
-5.02389073e-01 6.19208068e-02 -1.50528565e-01 -1.01292416e-01
3.08532745e-01 6.83947623e-01 -6.84484124e-01 -5.00375986e-01
4.31944638e-01 1.21520889e+00 6.41937673e-01 4.40459549e-01
-2.54030347e-01 1.76546350e-01 -7.01671124e-01 2.90498972e-01
6.37612402e-01 -4.89104152e-01 -7.77504444e-01 3.35913271e-01
-9.16412354e-01 -2.78339654e-01 -2.27322459e-01 8.27836573e-01
5.97154856e-01 1.69251159e-01 4.72296149e-01 -1.46789896e+00
-1.14049157e-02 9.88168716e-01 -9.90524530e-01 1.03866577e-01
3.50838974e-02 -4.33145970e-01 8.69047463e-01 -4.25423682e-01
5.28114796e-01 7.34834254e-01 3.59873652e-01 5.60479760e-01
-1.06534982e+00 -8.83012891e-01 3.47653419e-01 -2.34317392e-01
-1.60835421e+00 -6.13680422e-01 7.61233091e-01 2.20660850e-01
5.52798986e-01 6.31239116e-01 6.30502224e-01 2.06765562e-01
1.64730057e-01 -1.15071803e-01 5.74847698e-01 -7.36709654e-01
5.52372992e-01 1.36650562e-01 -4.15430427e-01 6.88927352e-01
4.42634165e-01 -1.55431405e-01 -3.85737121e-01 -1.21158838e-01
6.43734574e-01 -3.13797295e-01 -2.26489946e-01 -1.95321247e-01
-1.31067753e+00 8.74593973e-01 6.22062624e-01 7.16937482e-01
-8.02577376e-01 4.00996983e-01 -1.96851224e-01 4.97985184e-01
3.42818767e-01 3.89044493e-01 -2.63111830e-01 5.54964602e-01
-1.09536552e+00 -1.20056488e-01 9.61095095e-01 1.42179120e+00
9.52338994e-01 -6.93692267e-02 1.36935979e-01 7.28166103e-01
1.08329701e+00 5.23008466e-01 -3.13255489e-01 -1.37601113e+00
6.78840399e-01 2.60473877e-01 1.56541497e-01 -1.06493759e+00
-9.55534279e-01 -9.56745446e-01 -1.34138060e+00 2.12973997e-01
1.48957461e-01 -9.23861563e-01 -3.13458771e-01 1.96016395e+00
4.31531012e-01 -3.74319375e-01 3.67300481e-01 7.05439448e-01
1.66646466e-01 1.00067115e+00 -6.20669723e-02 -1.08197415e+00
8.65630209e-01 -7.28047967e-01 -9.69609559e-01 -2.75736839e-01
6.22870624e-01 -9.05911922e-01 -4.20020789e-01 -1.19356839e-02
-1.56608570e+00 -4.54396680e-02 -1.41822016e+00 4.95335430e-01
-1.91888049e-01 2.88096280e-03 3.79873127e-01 9.84311223e-01
-1.50014484e+00 1.61903203e-01 -3.23271066e-01 -3.86432230e-01
2.78354138e-01 9.07798946e-01 1.35116041e-01 7.44226128e-02
-9.81247842e-01 4.38584328e-01 5.71880817e-01 3.53389323e-01
-2.80385524e-01 -4.96421337e-01 -6.73102856e-01 -1.55891880e-01
3.17199498e-01 -6.58245087e-01 6.53470874e-01 -5.40579200e-01
-1.56688380e+00 2.36529499e-01 -1.63216442e-01 -9.44057852e-02
2.35676959e-01 7.17099369e-01 -3.01264405e-01 3.05773765e-01
4.30706501e-01 6.90549731e-01 2.78759778e-01 -1.19081700e+00
-9.18466210e-01 -1.45110920e-01 -2.82296930e-02 4.35913384e-01
-6.46526814e-01 -5.05199376e-03 -5.79278648e-01 -6.76849067e-01
6.79942727e-01 -9.88398790e-01 -9.81318355e-01 3.20470601e-01
-6.97539091e-01 9.11844298e-02 6.76441669e-01 -1.89991400e-01
1.67770863e+00 -1.91301835e+00 6.55603051e-01 1.26083970e+00
4.12427276e-01 -3.17584783e-01 -9.65374988e-03 7.61885762e-01
3.37479740e-01 3.38573128e-01 1.27656460e-02 -2.27909610e-01
-2.72638857e-01 4.11542147e-01 6.71978116e-01 6.66194856e-01
-5.62722564e-01 6.91930428e-02 -7.00929523e-01 -5.76690137e-01
2.02026349e-02 1.26732618e-01 -3.50167483e-01 -8.78930166e-02
1.72917962e-01 4.95475560e-01 -1.11891508e+00 3.13804060e-01
1.05732596e+00 -2.42500931e-01 6.87491059e-01 -9.10331458e-02
-5.76692998e-01 -1.83911622e-01 -1.68737876e+00 1.80699015e+00
-5.88341773e-01 1.64647251e-01 1.22205055e+00 -1.28579688e+00
6.60673797e-01 6.32524729e-01 1.26960766e+00 -6.68961883e-01
1.34721354e-01 6.40893996e-01 -2.97907025e-01 -1.88258663e-01
1.02568150e-01 -9.27758962e-02 -1.73397452e-01 6.97009146e-01
-2.94141620e-01 2.99403518e-01 3.12718689e-01 1.62335277e-01
1.21199477e+00 -5.85287333e-01 3.28738064e-01 -7.04435170e-01
6.38240755e-01 -7.62328207e-02 6.95724368e-01 8.28819752e-01
3.56533006e-02 -3.08937609e-01 9.93634015e-02 3.85914259e-02
-7.11869895e-01 -8.57619643e-01 1.49783328e-01 9.60175753e-01
5.97918034e-01 -1.05455585e-01 -5.49102843e-01 -1.83839664e-01
-3.74412090e-01 1.92304924e-01 -4.67801392e-01 1.63576841e-01
-2.24695832e-01 -9.43204880e-01 2.47124389e-01 -2.72480786e-01
8.46281111e-01 -4.74321991e-01 -6.23446345e-01 8.20371211e-01
-1.00779958e-01 -9.63768363e-01 -3.60413939e-01 4.72711772e-01
-6.39280975e-01 -7.01988995e-01 -9.22303557e-01 -8.04585397e-01
8.68640482e-01 4.70448554e-01 5.98562002e-01 2.63708502e-01
-9.95597690e-02 3.78702998e-01 -3.41383070e-01 -4.29805577e-01
-1.23237982e-01 4.62179244e-01 -1.51632890e-01 7.19181001e-02
-3.74947399e-01 -8.12575996e-01 -8.87169659e-01 6.40341282e-01
-5.79759598e-01 5.93131185e-02 7.78478682e-01 2.32684258e-02
5.37372470e-01 5.31959057e-01 9.57538068e-01 -2.78823316e-01
4.94437009e-01 -1.09356141e+00 -5.24590194e-01 6.23625554e-02
-2.77408510e-01 -9.13541391e-02 2.69429326e-01 1.46729127e-01
-7.87628889e-01 8.49492922e-02 2.06873454e-02 5.50381839e-01
4.67391521e-01 3.46773505e-01 -3.56194854e-01 -6.18316531e-01
1.00695506e-01 -2.43816018e-01 1.60126984e-01 -7.38657489e-02
2.03017786e-01 6.59902871e-01 -3.80168587e-01 -2.60743886e-01
9.80151951e-01 7.19328940e-01 8.15773249e-01 -1.24309504e+00
-5.26637316e-01 -3.03996652e-01 -4.40627277e-01 -4.12739605e-01
9.31290030e-01 -9.61565554e-01 -7.64206886e-01 -1.79637671e-02
-1.12731504e+00 -2.27745593e-01 -8.18498656e-02 8.58512998e-01
-2.79002130e-01 3.21111828e-01 -1.28999967e-02 -1.09763730e+00
-2.38805592e-01 -1.05718422e+00 4.51377302e-01 2.41861552e-01
-2.29741871e-01 -1.14738631e+00 -3.75357240e-01 -2.73147002e-02
8.10958862e-01 5.77986717e-01 7.62504637e-01 -6.50312081e-02
-6.97913110e-01 6.07159398e-02 -2.64653414e-01 -3.01079676e-02
3.77059430e-01 -4.81198013e-01 -1.74823165e-01 -5.61115682e-01
-2.35892624e-01 2.66896546e-01 3.74318719e-01 7.06342459e-01
8.70380521e-01 -4.10832763e-01 -8.14038217e-01 5.28554678e-01
1.71486759e+00 2.23074675e-01 2.81419456e-01 2.11707443e-01
2.18002945e-01 7.87523687e-01 4.22957808e-01 8.59563887e-01
5.93767345e-01 6.69578910e-01 8.77267540e-01 -3.84679019e-01
6.45307600e-02 8.24392676e-01 3.30846645e-02 4.45812106e-01
-2.76340753e-01 -8.31741393e-01 -2.95471519e-01 2.90880114e-01
-1.63700736e+00 -8.40024531e-01 1.18482430e-02 1.97874129e+00
3.70227247e-01 -4.83004227e-02 1.98038854e-02 5.10705352e-01
7.80229688e-01 6.55635148e-02 -2.64505655e-01 -3.16702038e-01
5.30807599e-02 9.87612456e-02 1.22122991e+00 6.26761079e-01
-5.91953933e-01 3.34715247e-01 5.96728468e+00 8.26508880e-01
-8.72469842e-01 1.59942266e-02 2.55268037e-01 -2.75954744e-03
-4.17592466e-01 -1.28788173e-01 -6.20455623e-01 4.19645071e-01
6.65632129e-01 -2.12163374e-01 3.93972099e-01 1.91840604e-01
9.50981617e-01 -3.69085014e-01 -5.53922057e-01 5.50281882e-01
-5.15722632e-01 -1.47042680e+00 -2.97028929e-01 8.37438226e-01
6.47168577e-01 -9.33315903e-02 -1.78657979e-01 -5.59824347e-01
2.85629451e-01 -6.82664514e-01 6.65889978e-01 3.22100163e-01
6.53154194e-01 -8.59433353e-01 2.17494071e-01 3.72987896e-01
-1.51947308e+00 -4.12346303e-01 -2.63758034e-01 -1.16988055e-01
4.70559955e-01 7.25527883e-01 -4.84446913e-01 9.71225023e-01
6.31164849e-01 3.17724049e-01 3.05767477e-01 1.01585591e+00
3.69798392e-01 1.09811381e-01 -6.42356634e-01 -3.05884987e-01
5.47930360e-01 -3.08809131e-01 6.89047992e-01 7.81713426e-01
7.81536877e-01 4.15245175e-01 6.70100093e-01 4.85943615e-01
-2.37499848e-01 1.70566842e-01 -1.98951483e-01 3.00982922e-01
9.15136993e-01 1.29685390e+00 -1.05914533e+00 1.19556084e-01
-5.49348533e-01 5.10639548e-01 -2.43488997e-01 6.85865760e-01
-5.59278250e-01 -8.68887722e-01 6.20130479e-01 1.08502887e-01
1.11645021e-01 -5.95810175e-01 -2.39233404e-01 -3.02607864e-01
1.67019516e-02 -2.21151158e-01 3.27835530e-01 -3.19802463e-02
-6.01635516e-01 3.91790092e-01 -5.62847964e-02 -1.18318295e+00
1.51863560e-01 2.57596746e-02 -5.34092546e-01 1.01783109e+00
-2.00874877e+00 -6.98627472e-01 -1.77004039e-01 9.24656808e-01
3.12665939e-01 -1.64230213e-01 9.04437900e-01 4.08817470e-01
-4.16348577e-01 5.09879887e-01 1.41196623e-01 -2.32419714e-01
8.55019689e-03 -7.16058195e-01 -8.56409013e-01 7.76547074e-01
-7.78987527e-01 4.62184042e-01 4.94254172e-01 -3.48919600e-01
-1.42190003e+00 -9.72498298e-01 7.91629493e-01 8.57698739e-01
3.89129907e-01 -1.56913817e-01 4.40655574e-02 2.15921149e-01
3.83148819e-01 -3.56023386e-02 1.17938817e+00 -4.65503961e-01
5.76557338e-01 -3.58602911e-01 -1.60036910e+00 5.53354442e-01
1.00536680e+00 4.17535782e-01 5.07016420e-01 4.43564951e-01
4.79924858e-01 1.19592525e-01 -9.69517827e-01 5.68583943e-02
4.53274876e-01 -5.93680561e-01 8.36382687e-01 1.87210292e-01
-4.28272873e-01 -5.04712522e-01 -5.55138350e-01 -1.16230905e+00
-4.52529401e-01 -1.42398262e+00 3.83722961e-01 1.12802041e+00
9.00280595e-01 -4.81690854e-01 8.44177306e-01 2.23551005e-01
-2.84517016e-02 -5.80998361e-01 -1.32599938e+00 -4.35252875e-01
-1.52850077e-01 -3.34767342e-01 4.04458612e-01 5.64389348e-01
6.54409006e-02 3.40171844e-01 -3.69954139e-01 8.23763788e-01
1.16256154e+00 -3.69915813e-01 3.83301824e-01 -1.46729958e+00
6.22344809e-03 -2.42983967e-01 -4.88484018e-02 -1.25420642e+00
6.25255033e-02 -7.64903605e-01 -2.67088085e-01 -1.89814126e+00
-4.23100650e-01 -1.42195570e+00 -1.38905227e-01 3.11088592e-01
7.90957212e-01 3.43171358e-01 8.53422135e-02 -4.62008081e-03
-5.84251344e-01 1.29385740e-01 1.26735592e+00 6.43656924e-02
-6.55023336e-01 4.28415507e-01 -9.55873072e-01 4.51083452e-01
1.00790739e+00 -3.70223135e-01 -7.76651382e-01 -5.61816275e-01
4.08039063e-01 8.05811584e-01 -5.01022302e-02 -1.01542652e+00
3.86175603e-01 -5.11614561e-01 3.69847655e-01 -2.34220669e-01
4.83642101e-01 -1.62427354e+00 1.85484320e-01 6.62323356e-01
-2.41480947e-01 -3.39297861e-01 -2.27367610e-01 7.43970752e-01
3.07294309e-01 -5.88164553e-02 9.30825233e-01 -4.87796441e-02
-5.06354153e-01 3.20222050e-01 -9.37907398e-01 -4.51499313e-01
1.68310869e+00 -5.26059151e-01 3.50318044e-01 -5.13871193e-01
-1.02776909e+00 8.75586689e-01 -2.01853648e-01 6.13747016e-02
4.00073618e-01 -1.04658020e+00 -6.74717009e-01 -1.80313513e-01
-3.88245583e-01 -1.02521084e-01 2.07431644e-01 1.00028431e+00
-3.41889381e-01 3.69301766e-01 -2.42923439e-01 -2.14654729e-01
-1.07521510e+00 -7.40561336e-02 6.59803987e-01 -1.95701569e-01
-1.71180889e-01 9.33455229e-01 -4.28345680e-01 1.07297271e-01
2.80615002e-01 1.60472780e-01 -4.79316682e-01 1.95942134e-01
2.69643039e-01 7.04475641e-01 -2.56952375e-01 -7.01967955e-01
-4.72755551e-01 7.34761775e-01 7.45304301e-02 -3.72649878e-01
1.44412756e+00 -1.16782641e+00 -6.52685642e-01 -3.86035413e-01
1.26163805e+00 4.31031585e-02 -6.97007537e-01 -2.95960814e-01
-1.38358533e-01 -2.06237257e-01 5.27139783e-01 -3.88429672e-01
-1.45928514e+00 -1.11981705e-01 2.94885725e-01 4.27942574e-01
1.38368273e+00 -9.03582573e-02 4.52966392e-01 4.37985390e-01
7.96281457e-01 -1.38539207e+00 4.71261851e-02 2.01814383e-01
4.28477705e-01 -5.49926877e-01 2.38158971e-01 -1.06985462e+00
1.67500079e-01 1.31823134e+00 1.20789930e-01 1.93963364e-01
1.28215492e+00 3.60855073e-01 -2.23537236e-01 -1.15129404e-01
-1.49320036e-01 -2.10843638e-01 -5.04193425e-01 6.03381991e-01
2.13252291e-01 -1.20302983e-01 -8.95698488e-01 1.54869512e-01
-3.29639856e-03 -2.93923229e-01 6.33201778e-01 1.34725034e+00
-9.53714967e-01 -1.60442555e+00 -3.86771590e-01 2.82816648e-01
-6.49714410e-01 2.02130571e-01 1.74280077e-01 7.25619256e-01
3.44888687e-01 1.71127725e+00 -1.64172813e-01 -4.07260247e-02
2.90963389e-02 -6.13205969e-01 1.52488559e-01 -6.65036380e-01
1.18763529e-01 3.37256044e-01 2.76024669e-01 -5.11336327e-01
-9.49118257e-01 -2.95292944e-01 -1.49661601e+00 -2.50580758e-01
-1.70282051e-01 7.05686688e-01 9.63043094e-01 8.66534054e-01
1.72826633e-01 8.09084177e-01 1.34296608e+00 -9.43286359e-01
-8.13744031e-03 -1.93200201e-01 -9.69011784e-01 -8.39815319e-01
5.31023800e-01 -4.04042870e-01 -3.99420738e-01 -3.47464055e-01] | [6.000926971435547, 1.5596858263015747] |
929b4c2d-f9ab-4d92-9f08-720842f98bc1 | mesaha-net-multi-encoders-based-self-adaptive | 2304.01576 | null | https://arxiv.org/abs/2304.01576v1 | https://arxiv.org/pdf/2304.01576v1.pdf | MESAHA-Net: Multi-Encoders based Self-Adaptive Hard Attention Network with Maximum Intensity Projections for Lung Nodule Segmentation in CT Scan | Accurate lung nodule segmentation is crucial for early-stage lung cancer diagnosis, as it can substantially enhance patient survival rates. Computed tomography (CT) images are widely employed for early diagnosis in lung nodule analysis. However, the heterogeneity of lung nodules, size diversity, and the complexity of the surrounding environment pose challenges for developing robust nodule segmentation methods. In this study, we propose an efficient end-to-end framework, the multi-encoder-based self-adaptive hard attention network (MESAHA-Net), for precise lung nodule segmentation in CT scans. MESAHA-Net comprises three encoding paths, an attention block, and a decoder block, facilitating the integration of three types of inputs: CT slice patches, forward and backward maximum intensity projection (MIP) images, and region of interest (ROI) masks encompassing the nodule. By employing a novel adaptive hard attention mechanism, MESAHA-Net iteratively performs slice-by-slice 2D segmentation of lung nodules, focusing on the nodule region in each slice to generate 3D volumetric segmentation of lung nodules. The proposed framework has been comprehensively evaluated on the LIDC-IDRI dataset, the largest publicly available dataset for lung nodule segmentation. The results demonstrate that our approach is highly robust for various lung nodule types, outperforming previous state-of-the-art techniques in terms of segmentation accuracy and computational complexity, rendering it suitable for real-time clinical implementation. | ['Yeong Gil Shin', 'Tariq Mahmood Khan', 'Sung Hyun Kim', 'Shi Sub Byon', 'Siddique Latif', 'Abdullah Shahid', 'Azka Rehman', 'Muhammad Usman'] | 2023-04-04 | null | null | null | null | ['lung-cancer-diagnosis', 'lung-nodule-segmentation', 'hard-attention'] | ['medical', 'medical', 'methodology'] | [ 3.42624694e-01 3.77227038e-01 -2.98910499e-01 -1.45115897e-01
-1.03932190e+00 -3.05823117e-01 2.04191625e-01 -2.16707855e-01
-3.17093939e-01 3.40868026e-01 3.28823626e-02 -6.68586195e-01
-1.91408712e-02 -6.33678496e-01 -3.45871478e-01 -7.70228148e-01
1.29619673e-01 9.19697881e-01 6.74549580e-01 3.49460334e-01
-3.32909137e-01 7.04763472e-01 -8.80008638e-01 4.21059042e-01
6.74809337e-01 9.25771892e-01 6.86712503e-01 8.74651730e-01
-6.37016743e-02 8.90757501e-01 -3.10159312e-03 -9.64179188e-02
2.16531113e-01 -6.04849339e-01 -1.06529689e+00 3.63280982e-01
1.06148385e-01 -5.48551023e-01 -4.56910044e-01 8.09511423e-01
5.87321460e-01 -3.18512768e-01 7.13005602e-01 -5.82017839e-01
-3.61355841e-01 5.60396910e-01 -3.99068117e-01 4.82267350e-01
-4.45973575e-01 4.04178053e-01 9.23550129e-01 -1.01330507e+00
4.16418821e-01 5.19476712e-01 6.89378977e-01 6.33740366e-01
-6.26265645e-01 -2.74940729e-01 -3.15668285e-01 -1.09080337e-01
-1.28353500e+00 1.66625246e-01 2.21249178e-01 -5.27526200e-01
7.59831965e-01 5.61409295e-01 9.05162930e-01 5.49006522e-01
5.70521176e-01 7.27166355e-01 6.74449027e-01 -3.67385037e-02
-8.56519490e-02 -8.24467838e-02 -4.20141041e-01 9.84527290e-01
3.65512878e-01 4.91257459e-02 5.46436429e-01 -1.58261403e-01
1.28720617e+00 2.92828679e-01 -3.53980303e-01 -3.35138291e-01
-1.49382806e+00 7.42636263e-01 1.11603642e+00 5.93030512e-01
-6.96407020e-01 3.87027383e-01 1.79577783e-01 -5.55534661e-01
3.27489197e-01 1.13058798e-01 2.96024829e-02 3.77193362e-01
-9.84653950e-01 -1.01621933e-01 5.46801686e-01 6.17802978e-01
5.59219420e-02 -1.81097627e-01 -9.34209704e-01 5.40812671e-01
5.75872064e-01 3.75542372e-01 9.76821959e-01 -5.05389094e-01
1.97580472e-01 5.59050798e-01 -2.03223944e-01 -2.20150560e-01
-5.12031555e-01 -7.01400757e-01 -1.04798222e+00 -2.26398483e-01
6.63145930e-02 -1.29001196e-02 -1.36796844e+00 1.24415755e+00
5.51482439e-01 3.27562422e-01 -3.57577980e-01 1.11767519e+00
1.05419219e+00 2.97986090e-01 2.64009297e-01 -5.59364213e-03
1.58180344e+00 -1.38601160e+00 -3.85261953e-01 -3.73400211e-01
6.74531162e-01 -4.64222521e-01 8.82980049e-01 -4.42314178e-01
-1.31439674e+00 -4.07697469e-01 -6.86306119e-01 -1.32443145e-01
2.38340989e-01 5.16249776e-01 3.72561842e-01 5.55632114e-01
-1.01637900e+00 3.00260663e-01 -1.17413104e+00 -1.15152664e-01
8.03474247e-01 7.66993403e-01 4.61513884e-02 -9.08013154e-03
-8.54395151e-01 5.85162938e-01 4.90886122e-01 1.49708241e-01
-1.06613445e+00 -9.67114866e-01 -4.11808312e-01 2.77320772e-01
5.41627705e-01 -1.13364840e+00 1.76849174e+00 -6.11424208e-01
-1.37626421e+00 9.21491742e-01 -8.49787332e-03 -4.29731548e-01
6.58163130e-01 3.38046759e-01 5.84548712e-02 3.29228997e-01
1.36014059e-01 7.93421984e-01 8.60778809e-01 -8.34708273e-01
-5.86562932e-01 -2.42548913e-01 -6.31541491e-01 2.99369425e-01
1.34231761e-01 -9.28176269e-02 -8.90571535e-01 -6.18053436e-01
1.06847800e-01 -1.07999229e+00 -8.32056761e-01 2.57459134e-01
-5.91253817e-01 7.00992271e-02 1.07837963e+00 -7.56818771e-01
1.21934867e+00 -1.95011938e+00 6.21903017e-02 3.69847834e-01
6.22298956e-01 3.40466380e-01 2.43213639e-01 -5.82611203e-01
-1.34576514e-01 3.83642733e-01 -5.24191678e-01 5.33285700e-02
-3.37901413e-01 7.43832588e-02 4.25240725e-01 2.77234524e-01
2.78684467e-01 1.66680491e+00 -7.44605541e-01 -9.76491570e-01
4.64164585e-01 4.80202407e-01 -5.63454449e-01 5.01225531e-01
-3.34950417e-01 6.26701415e-01 -7.57975101e-01 7.62182415e-01
3.04722667e-01 -9.33842003e-01 1.20292328e-01 -1.85213521e-01
1.13667198e-03 1.22331697e-02 -5.85857630e-01 1.27805197e+00
-4.79387790e-01 3.25954348e-01 1.69269219e-01 -2.50131875e-01
2.90025949e-01 8.10535073e-01 1.05716622e+00 -3.11856002e-01
5.76274991e-01 4.73638207e-01 6.16520643e-01 -6.41160667e-01
-8.60252902e-02 -2.26635635e-01 1.55988976e-01 6.30743325e-01
-2.29666695e-01 -5.61602831e-01 1.58997715e-01 -8.80331174e-02
1.31226337e+00 -5.35247326e-01 6.80249453e-01 -6.14376254e-02
6.95650518e-01 1.24504991e-01 1.98479414e-01 5.07309139e-01
-5.02918482e-01 1.00999486e+00 2.88899422e-01 -1.01702318e-01
-1.15054405e+00 -1.00496793e+00 -2.23633915e-01 5.68166077e-01
-2.06057325e-01 1.92040175e-01 -8.42536092e-01 -1.06160343e+00
-2.07075790e-01 2.65821844e-01 -7.81429350e-01 1.65096432e-01
-7.99594104e-01 -8.05467606e-01 3.92260611e-01 7.75639296e-01
5.96551120e-01 -1.32723689e+00 -9.51702416e-01 4.91195738e-01
-3.30603838e-01 -1.05281746e+00 -8.28585565e-01 3.10882628e-01
-1.17116714e+00 -1.14577281e+00 -1.27457714e+00 -8.62472177e-01
8.80964637e-01 3.18438858e-01 1.25405574e+00 3.54213029e-01
-8.61822426e-01 2.10065380e-01 1.66022614e-01 -2.23507002e-01
-5.71865022e-01 2.63398200e-01 -8.21804523e-01 -2.67344266e-01
-1.14155449e-01 -1.02129683e-01 -7.99564958e-01 4.30879831e-01
-1.00347364e+00 3.59642744e-01 1.38799679e+00 9.50320959e-01
1.18190813e+00 6.21179231e-02 1.38986170e-01 -1.20964968e+00
3.65944773e-01 -7.81707227e-01 -3.34597379e-01 2.52162337e-01
1.55270146e-02 -2.38981515e-01 4.48714942e-01 -1.15671530e-01
-1.02978754e+00 4.33649063e-01 -4.36329693e-01 -3.79491419e-01
-4.11984883e-02 4.95706558e-01 8.72182921e-02 -3.62364382e-01
4.82730031e-01 1.48285940e-01 -1.34680243e-02 9.62094590e-02
1.81317683e-02 6.27314270e-01 5.72779059e-01 1.13720588e-01
7.66297519e-01 4.46074605e-01 1.02189198e-01 -6.60046160e-01
-8.45800638e-01 -7.55450666e-01 -8.81303310e-01 -2.23168060e-01
1.22142994e+00 -6.50347829e-01 -2.20984086e-01 1.16673067e-01
-8.06019902e-01 -4.75695401e-01 -5.75961769e-01 5.84969163e-01
-5.55583894e-01 1.26525372e-01 -7.46574938e-01 -1.23970911e-01
-7.40086198e-01 -1.51252711e+00 1.15335500e+00 2.81810194e-01
-5.34982383e-02 -1.01265132e+00 -6.20324351e-02 4.94298607e-01
6.94910169e-01 1.07955121e-01 1.09066248e+00 -7.23338306e-01
-1.04540563e+00 -3.35202217e-01 -5.94808757e-01 1.14497624e-01
2.92254865e-01 -4.38371040e-02 -7.22685635e-01 -2.68341243e-01
-1.59992557e-02 -2.25214273e-01 7.67786145e-01 8.66030514e-01
1.68948865e+00 -7.51773342e-02 -6.53735697e-01 7.45094419e-01
1.33689189e+00 1.98512912e-01 2.83393443e-01 -1.32524505e-01
9.71364379e-01 -1.14510663e-01 2.64382988e-01 2.56089538e-01
4.96447124e-02 3.22186023e-01 7.66784549e-01 -5.52854478e-01
-5.29368997e-01 1.55635253e-01 -2.77725101e-01 5.98034263e-01
-1.38175562e-01 -4.45366114e-01 -1.12335503e+00 6.18140042e-01
-1.33020639e+00 -6.86719477e-01 -2.45335102e-01 1.87875783e+00
5.47302783e-01 -1.85813736e-02 -1.43056912e-02 -1.25194535e-01
6.04831576e-01 2.84232553e-02 -8.19878578e-01 9.92910936e-02
6.15177095e-01 4.68890458e-01 5.79105318e-01 2.40958899e-01
-1.31427836e+00 4.73941714e-01 6.01524544e+00 8.02691877e-01
-1.24308622e+00 4.22704399e-01 9.13211703e-01 4.97964248e-02
-2.39965245e-01 -4.93220687e-01 -4.63203460e-01 2.51719922e-01
6.56609297e-01 -6.11619055e-02 -1.30085886e-01 9.06721592e-01
5.62293455e-02 -7.17619853e-03 -1.02525222e+00 4.34625715e-01
-1.49871990e-01 -1.40562236e+00 -1.50237709e-01 1.79597497e-01
8.04923058e-01 5.20032167e-01 1.49150401e-01 1.22108594e-01
4.56347913e-02 -1.20997930e+00 2.19184712e-01 1.30738929e-01
9.96494055e-01 -4.42956537e-01 1.10076439e+00 3.40654522e-01
-1.41554439e+00 2.80910693e-02 -1.35215923e-01 9.59592342e-01
1.46729812e-01 6.47827208e-01 -1.67644942e+00 5.16712785e-01
4.90232706e-01 3.60863000e-01 -8.63821805e-01 1.39341235e+00
7.43691325e-02 7.62548685e-01 -4.30371612e-01 -1.65765081e-02
5.58009028e-01 1.33902580e-01 3.73720527e-01 1.07694244e+00
6.45870149e-01 3.22012007e-01 2.28531122e-01 1.00198746e+00
-2.70628691e-01 1.88723177e-01 -1.44709408e-01 -1.44071057e-01
1.68627709e-01 1.44172919e+00 -1.29964447e+00 -4.24058020e-01
-4.81713831e-01 8.15992653e-01 2.42082365e-02 -2.12133639e-02
-1.25470793e+00 1.46675751e-01 -2.03531384e-01 3.87880564e-01
5.39597988e-01 3.13532114e-01 -2.87757814e-01 -5.79223752e-01
-2.44839489e-01 -6.53386950e-01 4.43323672e-01 -5.66121161e-01
-8.76930416e-01 9.67267156e-01 -2.44317934e-01 -1.21250117e+00
-2.40496308e-01 -4.82496709e-01 -9.25101876e-01 6.21007562e-01
-1.34320748e+00 -1.19962931e+00 -8.17144513e-01 3.95885259e-01
6.44706607e-01 8.74523893e-02 4.50452656e-01 2.45778173e-01
-6.03529692e-01 3.52945358e-01 -1.15487978e-01 1.95822462e-01
2.50678420e-01 -1.31135356e+00 2.28974357e-01 6.03821158e-01
-1.96292438e-02 -1.74095318e-01 -6.17126375e-02 -7.40632713e-01
-9.80524719e-01 -1.73449945e+00 4.92055625e-01 -2.22936347e-01
3.46777588e-01 2.01194584e-01 -9.47502136e-01 7.54294872e-01
2.08718255e-02 6.32580459e-01 5.34357965e-01 -9.63613212e-01
4.86329198e-01 3.90245378e-01 -1.27097535e+00 5.68905592e-01
5.98585546e-01 -6.87079653e-02 -1.28834173e-01 6.11787498e-01
6.70098007e-01 -8.64661574e-01 -9.59101379e-01 7.47968912e-01
2.50584215e-01 -9.33321357e-01 1.25600946e+00 7.70924054e-03
5.05620778e-01 -1.40577257e-01 3.41303736e-01 -8.95011485e-01
-8.78933191e-01 -1.56132644e-02 -2.51069926e-02 3.86815190e-01
6.30483031e-01 -3.16484094e-01 1.20584238e+00 5.28922260e-01
-5.52604616e-01 -1.34933114e+00 -8.27097893e-01 -1.48537010e-01
8.66536424e-02 -2.13703975e-01 3.34332585e-01 2.83381522e-01
-7.45553374e-01 -1.41953424e-01 3.25157046e-01 7.29275793e-02
5.14957726e-01 -8.29921570e-03 1.37697652e-01 -9.13818121e-01
-5.13542593e-01 -8.89459908e-01 -1.11033723e-01 -9.63625550e-01
-2.68300533e-01 -1.27230167e+00 2.42827207e-01 -1.92157090e+00
5.93885541e-01 -3.19804907e-01 -2.29437247e-01 4.03018087e-01
-4.81804579e-01 4.80372250e-01 -6.18543988e-03 3.69154841e-01
-3.53116453e-01 2.39918113e-01 2.00551701e+00 -1.57179356e-01
-5.07397689e-02 6.28275216e-01 -4.05253798e-01 7.62681186e-01
8.29417646e-01 -6.08973980e-01 -3.70060295e-01 -1.93831906e-01
-5.89594841e-01 3.18444431e-01 5.64756274e-01 -1.22574222e+00
2.92039424e-01 -1.59932569e-01 6.87929392e-01 -1.02290845e+00
1.27830312e-01 -1.06562984e+00 1.32458985e-01 1.13424528e+00
-2.52496004e-01 -2.83249497e-01 1.01604566e-01 5.06768942e-01
-2.30636328e-01 -3.34699154e-01 1.12009287e+00 -5.43165028e-01
-3.61237764e-01 1.00645089e+00 -6.32025182e-01 1.27402201e-01
1.47501552e+00 -2.35149533e-01 2.85346150e-01 7.53054917e-02
-7.64995039e-01 2.77989835e-01 2.35809628e-02 -7.15703368e-02
6.49012089e-01 -1.33837473e+00 -9.35987055e-01 3.58670622e-01
-1.46823108e-01 6.50203645e-01 4.68049079e-01 1.32526994e+00
-9.75043952e-01 8.01380575e-01 1.26044467e-01 -8.97845924e-01
-1.28960407e+00 3.44964355e-01 8.74090672e-01 -9.65213656e-01
-7.23787010e-01 1.17552269e+00 5.10857701e-01 -3.52846384e-01
1.27943933e-01 -9.09419417e-01 1.52551517e-01 -5.58102667e-01
8.60225931e-02 8.95978287e-02 2.61842191e-01 -4.50245947e-01
-2.33336195e-01 4.69310910e-01 -1.97305262e-01 1.49135947e-01
7.79674590e-01 1.33557975e-01 -5.21713234e-02 -3.06079183e-02
9.44968522e-01 -2.88515925e-01 -1.12427151e+00 -1.74263775e-01
-5.54773621e-02 -2.26347119e-01 4.38560635e-01 -8.49759459e-01
-1.44766414e+00 8.02574933e-01 6.75871730e-01 1.84044503e-02
1.14056075e+00 2.92729259e-01 1.07761014e+00 1.49469957e-01
-1.66360036e-01 -3.76508325e-01 2.18712911e-01 2.72306740e-01
8.12746108e-01 -1.39213777e+00 4.26104665e-02 -8.51856351e-01
-7.07854569e-01 1.00076675e+00 9.44270372e-01 5.21298610e-02
7.80710459e-01 4.50581044e-01 7.70339146e-02 -2.73804247e-01
-7.83406436e-01 -3.73894542e-01 6.53427958e-01 1.61285043e-01
7.43556082e-01 1.35181248e-01 1.84100747e-01 6.18642449e-01
1.71716020e-01 1.07012056e-01 2.19586819e-01 8.74562860e-01
-6.38939798e-01 -7.90716410e-01 -4.37242985e-01 9.13119376e-01
-5.48828840e-01 -9.07215104e-02 -4.16210651e-01 1.06893480e+00
1.57435179e-01 2.26365373e-01 1.42174333e-01 2.31247976e-01
8.30557644e-02 -1.65538818e-01 4.42905337e-01 -9.94246960e-01
-9.12392020e-01 3.51144612e-01 -3.55288923e-01 -2.92619348e-01
-8.53763819e-02 -4.72300917e-01 -1.57594824e+00 1.70893773e-01
-3.45844716e-01 8.98169260e-03 4.02207822e-01 7.97785401e-01
1.79508850e-01 1.18682003e+00 7.18683779e-01 -8.90818894e-01
-5.61477840e-01 -8.50845277e-01 -1.80886090e-01 1.63845588e-02
1.56225905e-01 -2.82984465e-01 -1.41462356e-01 -1.24706589e-01] | [15.409163475036621, -2.128084897994995] |
043673de-beda-4532-99af-4850fb4819d5 | identifying-nuanced-dialect-for-arabic-tweets | null | null | https://aclanthology.org/2020.wanlp-1.30 | https://aclanthology.org/2020.wanlp-1.30.pdf | Identifying Nuanced Dialect for Arabic Tweets with Deep Learning and Reverse Translation Corpus Extension System | In this paper, we present our work for the NADI Shared Task (Abdul-Mageed and Habash, 2020): Nuanced Arabic Dialect Identification for Subtask-1: country-level dialect identification. We introduce a Reverse Translation Corpus Extension Systems (RTCES) to handle data imbalance along with reported results on several experimented approaches of word and document representations and different models architectures. The top scoring model was based on AraBERT (Antoun et al., 2020), with our modified extended corpus based on reverse translation of the given Arabic tweets. The selected system achieved a macro average F1 score of 20.34% on the test set, which places us as the 7th out of 18 teams in the final ranking Leaderboard. | ['Marwan Torki', 'Youssef Kishk', 'Rawan Tahssin'] | null | null | null | null | coling-wanlp-2020-12 | ['dialect-identification'] | ['natural-language-processing'] | [-2.52027482e-01 3.65867503e-02 1.50296822e-01 -5.59423923e-01
-1.14028966e+00 -6.77527666e-01 9.60765481e-01 4.34113406e-02
-5.76897442e-01 6.90974891e-01 4.81132179e-01 -3.48059386e-01
-6.87160119e-02 -4.28796262e-01 -8.67746174e-02 -3.39181602e-01
-3.41015160e-02 1.29547060e+00 -2.61718571e-01 -1.43836951e+00
6.45667195e-01 3.88491452e-01 -1.05704665e+00 7.52699256e-01
1.35839677e+00 5.12848437e-01 -9.75011364e-02 7.05324888e-01
5.96318394e-02 7.72430718e-01 -8.27536106e-01 -8.88097465e-01
9.82016921e-02 -5.39663494e-01 -1.32190990e+00 -3.77790421e-01
7.64756978e-01 -1.13847673e-01 3.28552239e-02 5.65821826e-01
8.45741630e-01 -2.99048126e-02 8.40892255e-01 -8.64592433e-01
-9.54084873e-01 1.11978602e+00 -5.87687492e-01 1.94971904e-01
5.05259216e-01 -3.73831898e-01 9.73448932e-01 -1.51452661e+00
6.68333292e-01 1.14723587e+00 7.07648814e-01 4.45392430e-01
-9.42343831e-01 -6.63105190e-01 9.28360671e-02 4.36158866e-01
-1.59582281e+00 -5.61200261e-01 3.90875250e-01 -3.54871869e-01
1.14967561e+00 4.19067144e-01 2.11725712e-01 8.96683931e-01
-1.21850930e-01 7.84750283e-01 1.53790498e+00 -8.48687649e-01
-3.82067949e-01 5.35196006e-01 5.48554242e-01 5.82637310e-01
9.06353593e-02 -5.21000862e-01 -7.16119826e-01 1.48231471e-02
2.15391994e-01 -9.09731627e-01 -9.03015435e-02 4.40627217e-01
-1.33527851e+00 9.56128180e-01 2.01312065e-01 4.52843100e-01
-2.86481023e-01 -7.60279000e-01 3.75236332e-01 8.49293828e-01
6.38819575e-01 5.54165840e-01 -5.32680809e-01 -1.39687479e-01
-7.71469414e-01 4.35000420e-01 7.15320945e-01 8.81485879e-01
4.30236608e-01 4.81054522e-02 -1.87115267e-01 1.51542914e+00
1.34709567e-01 6.12343848e-01 9.25582409e-01 -2.30881423e-01
7.08879411e-01 5.75556934e-01 5.53569682e-02 -7.85324275e-01
-6.20936513e-01 -3.42984706e-01 -6.86690748e-01 -9.70646665e-02
7.76467323e-01 -5.64649761e-01 -8.53199840e-01 1.37575948e+00
-1.67386562e-01 -5.20493567e-01 1.50618389e-01 1.01620436e+00
7.07899034e-01 9.34685409e-01 -4.48923230e-01 -7.11918697e-02
1.34143257e+00 -1.12236631e+00 -5.72844446e-01 -2.66065896e-01
7.64324129e-01 -1.33301020e+00 9.26716387e-01 7.67116070e-01
-1.10651255e+00 -4.54206139e-01 -1.04888117e+00 1.13001391e-01
-6.88248098e-01 4.10280079e-01 2.81827450e-01 1.09094274e+00
-1.36412799e+00 -3.29035260e-02 -2.49927372e-01 -7.41944611e-01
-3.61678839e-01 5.06485939e-01 -4.78245437e-01 -1.20845683e-01
-1.41253531e+00 1.48804939e+00 1.13492347e-01 1.50447458e-01
-5.00355721e-01 -1.91578209e-01 -4.37018633e-01 -6.07946038e-01
-3.12153488e-01 1.20815590e-01 8.75940979e-01 -1.27096951e+00
-1.75872064e+00 1.42859757e+00 1.33379921e-01 -4.74715024e-01
3.71619701e-01 -4.22039837e-01 -9.42638993e-01 -3.52685928e-01
3.00447289e-02 3.65964472e-01 3.64112079e-01 -1.07322466e+00
-7.55238235e-01 -5.74693203e-01 -1.10988736e-01 5.23606420e-01
-3.55399072e-01 6.86630905e-01 -4.02338170e-02 -8.33300769e-01
1.68291479e-01 -1.06955802e+00 3.97939868e-02 -1.25863659e+00
-8.39541107e-02 -8.18627849e-02 2.63493121e-01 -1.46571958e+00
1.42306805e+00 -1.57594097e+00 2.83217907e-01 4.95400608e-01
-2.65221536e-01 3.16359401e-01 -4.33679670e-01 8.09043527e-01
-2.63398349e-01 -1.98192492e-01 -6.28437325e-02 -2.35683903e-01
-3.90743129e-02 -2.87975073e-01 -3.26853186e-01 3.62365007e-01
1.40800506e-01 3.76312196e-01 -6.88614547e-01 -8.63736048e-02
-3.21276814e-01 1.35548174e-01 -1.70042276e-01 1.11646324e-01
1.75986335e-01 2.52346843e-01 3.81369174e-01 7.83184767e-01
1.10054672e+00 5.00172675e-01 4.99237061e-01 -6.79734796e-02
-4.57489789e-01 5.60082793e-01 -1.19506490e+00 1.47721016e+00
-5.06357253e-01 6.81986630e-01 2.06791967e-01 -6.31071806e-01
1.47678590e+00 2.48684734e-01 1.66843638e-01 -9.05156732e-01
7.36346394e-02 9.15246665e-01 2.34559655e-01 -1.58563152e-01
1.09976101e+00 2.23499611e-01 -2.99990118e-01 8.10663998e-01
4.03016768e-02 2.26951987e-01 5.71486533e-01 2.98455894e-01
4.35455382e-01 1.97337672e-01 1.64216056e-01 -8.34946573e-01
9.44722533e-01 2.53574967e-01 5.41852489e-02 6.03363276e-01
-2.88123816e-01 9.35840368e-01 2.39246413e-01 -6.01961732e-01
-8.04284394e-01 -9.71432030e-01 -1.11740522e-01 1.68946397e+00
-4.84774053e-01 -3.85709733e-01 -8.78075838e-01 -7.33002186e-01
-4.84931171e-01 7.32318282e-01 -6.12963855e-01 9.39725190e-02
-1.17008507e+00 -1.32807207e+00 1.09857035e+00 -4.09348421e-02
4.52134818e-01 -6.62161469e-01 -1.76079236e-02 3.11451077e-01
-9.53864813e-01 -5.32742679e-01 -5.31626165e-01 -8.18505958e-02
-2.77107298e-01 -8.62851858e-01 -1.04933083e+00 -1.43403721e+00
2.85632432e-01 5.64204790e-02 1.35263002e+00 -2.96217561e-01
2.42141187e-01 -1.29757121e-01 -8.08898568e-01 -4.14444447e-01
-6.21536016e-01 5.88286459e-01 8.13156143e-02 2.50507772e-01
5.64489841e-01 2.17014775e-01 -1.33887097e-01 3.93037230e-01
-3.65744382e-01 4.44823457e-03 3.12410802e-01 9.37747359e-01
-9.81209651e-02 -7.27421880e-01 8.58276784e-01 -8.10035765e-01
8.49945486e-01 -4.20575380e-01 -2.93344527e-01 5.70952415e-01
-6.37827754e-01 -2.35818490e-01 3.28667313e-01 -1.82709500e-01
-1.16129458e+00 -2.73735195e-01 -4.92082834e-01 7.32929885e-01
1.60506591e-01 7.15443790e-01 1.03653647e-01 4.01581489e-02
1.03676164e+00 4.34386224e-01 1.48308799e-01 -5.15578926e-01
1.43456921e-01 1.39809263e+00 2.81943291e-01 -6.50300801e-01
1.61015868e-01 -1.42958716e-01 -5.00514805e-01 -6.65482700e-01
-5.14674842e-01 -2.35615000e-01 -8.20691168e-01 -2.33674556e-01
4.19820607e-01 -1.19623566e+00 -1.62324041e-01 9.60923135e-01
-1.14413631e+00 -3.14771354e-01 4.88522470e-01 4.17297065e-01
-4.04214621e-01 -4.22317423e-02 -8.17084432e-01 -7.48845339e-01
-5.89771271e-01 -1.11665082e+00 7.02437401e-01 -3.48882824e-01
-5.73359013e-01 -8.68287504e-01 6.76837087e-01 5.79293609e-01
6.59883976e-01 1.42685086e-01 1.05386961e+00 -1.00518847e+00
2.17798412e-01 -1.47310551e-02 -6.19895346e-02 1.64552703e-01
1.15521841e-01 2.07580961e-02 -9.86061633e-01 -4.98065442e-01
-4.02312249e-01 -5.01813471e-01 6.12109601e-01 6.39965087e-02
1.94232076e-01 -2.54403263e-01 4.31252241e-01 6.34570941e-02
1.23293614e+00 4.11439016e-02 6.34035647e-01 7.99127877e-01
4.40609306e-01 1.01791477e+00 9.19143200e-01 3.90772432e-01
9.12467062e-01 8.81540477e-01 -2.22010911e-02 1.40407523e-02
-2.05100551e-01 1.86593994e-01 8.96177232e-01 1.34723449e+00
-3.54054928e-01 -1.52533591e-01 -1.40700436e+00 8.28899682e-01
-1.59320319e+00 -6.81281149e-01 -6.76030874e-01 2.44670105e+00
8.78432930e-01 -3.28240961e-01 7.45250404e-01 6.72532320e-02
7.73545623e-01 -2.50976115e-01 3.35285068e-01 -1.01276398e+00
-7.14177608e-01 3.30856025e-01 3.43194276e-01 1.02493048e+00
-9.15974617e-01 1.37523961e+00 6.11229610e+00 7.89858997e-01
-1.01732039e+00 2.10159466e-01 6.31989956e-01 -2.79372763e-02
6.21870495e-02 -3.12508225e-01 -9.16029513e-01 4.16072965e-01
1.60459876e+00 -2.21237034e-01 5.72020054e-01 1.51228085e-01
-2.52517462e-01 1.01143174e-01 -7.27428496e-01 7.97321022e-01
7.45697796e-01 -9.55509186e-01 3.82179111e-01 2.39551446e-04
8.80421340e-01 4.82640594e-01 2.60240167e-01 5.83187222e-01
3.81653547e-01 -1.17159796e+00 9.72205460e-01 2.45490879e-01
8.57721806e-01 -1.19733071e+00 1.17943788e+00 5.59201017e-02
-7.09893644e-01 -6.45566955e-02 -2.46914446e-01 -7.66109526e-02
-1.84910610e-01 1.24985604e-02 -1.10806954e+00 8.01982045e-01
4.32414263e-01 4.27564830e-01 -8.96127701e-01 5.18768072e-01
1.72643602e-01 8.31032991e-01 -2.66052604e-01 -1.99272200e-01
4.54827279e-01 -4.08910483e-01 4.55180407e-01 1.71390104e+00
4.51039404e-01 -4.02958572e-01 -1.58356354e-01 -2.49793157e-01
1.33051097e-01 8.11497808e-01 -4.26953018e-01 3.76538813e-01
3.65954369e-01 1.20519257e+00 -4.75881249e-01 -1.88355267e-01
-3.39966238e-01 1.14013243e+00 4.96504575e-01 9.05755162e-02
-5.30747831e-01 -7.37278402e-01 4.63284343e-01 -8.71525332e-02
-1.35640904e-01 -1.27628714e-01 -5.01096129e-01 -9.74213600e-01
-2.38085613e-01 -1.63819218e+00 8.02149892e-01 -4.66037810e-01
-1.18815947e+00 1.29823649e+00 -2.27822766e-01 -1.30402160e+00
-5.41578293e-01 -1.04387450e+00 -2.03086346e-01 1.11620235e+00
-1.00730288e+00 -1.64237034e+00 -4.37134579e-02 6.47451699e-01
5.11747658e-01 -9.66463447e-01 1.08190584e+00 7.33128667e-01
-4.72575933e-01 1.17210257e+00 5.07864058e-01 3.24441314e-01
1.46681178e+00 -1.33034647e+00 4.91846561e-01 7.55638838e-01
1.34892315e-01 6.01935029e-01 5.21460474e-01 -4.54401940e-01
-1.13034558e+00 -9.05356050e-01 1.73451960e+00 -8.84568691e-01
6.86712325e-01 -4.71577048e-01 -5.13292015e-01 7.43538976e-01
7.62949169e-01 -1.13882399e+00 7.84585357e-01 4.44587857e-01
-5.35692215e-01 -3.01293105e-01 -1.24435949e+00 6.46026731e-01
5.79217315e-01 -4.11394447e-01 -3.67282361e-01 6.37109816e-01
1.78934306e-01 -3.85685951e-01 -8.41943622e-01 6.02866746e-02
6.99621975e-01 -9.19896603e-01 7.19938517e-01 -6.41920030e-01
3.28706473e-01 -1.93667901e-03 -5.58269441e-01 -1.80579853e+00
-3.15731555e-01 -3.93775731e-01 5.17239809e-01 1.38976657e+00
9.01845336e-01 -6.92134917e-01 2.87336022e-01 1.18670933e-01
-2.26038426e-01 -3.60316247e-01 -6.58570647e-01 -6.26858950e-01
7.84896672e-01 -1.43632486e-01 7.44329572e-01 1.25655890e+00
2.97177881e-01 7.69841492e-01 -5.48606634e-01 -1.16802827e-01
5.98417744e-02 -2.11430322e-02 6.93695843e-01 -9.45965111e-01
7.53600225e-02 -7.33052433e-01 -9.11592320e-02 -4.72903490e-01
-6.49837032e-02 -1.14429736e+00 -1.71390250e-01 -1.39687133e+00
-7.38286301e-02 -2.31720656e-01 -2.95826316e-01 2.26979002e-01
-7.34555572e-02 7.48635769e-01 2.56981045e-01 1.80533931e-01
-3.16180795e-01 -3.98558006e-02 6.81401908e-01 -1.89808771e-01
-1.74179181e-01 -2.72887442e-02 -9.68266606e-01 2.47527510e-01
1.09724319e+00 -1.08760176e-02 -6.61316141e-02 -8.64424467e-01
4.65104133e-01 -3.34310114e-01 -3.68953407e-01 -7.32635379e-01
2.07178704e-02 -2.43664030e-02 5.43206334e-02 -6.49844587e-01
1.30464375e-01 -2.92117059e-01 -1.49009138e-01 4.92803305e-01
-3.67094100e-01 1.00216770e+00 1.99736476e-01 -5.94625473e-01
-3.70644987e-01 -2.16921404e-01 7.64799595e-01 3.80007438e-02
-5.55210173e-01 -1.02529086e-01 -7.82567620e-01 1.02937333e-02
5.49594402e-01 7.77526274e-02 -5.00846207e-01 -4.94168043e-01
-4.53976572e-01 1.34193107e-01 1.53261751e-01 6.07061327e-01
4.78011161e-01 -1.24005890e+00 -1.97641861e+00 3.58914465e-01
3.48805070e-01 -8.03984165e-01 -6.87190518e-02 8.45979512e-01
-1.00670040e+00 5.85553169e-01 -7.21717954e-01 1.44616524e-02
-1.26639998e+00 -2.40024343e-01 2.96402842e-01 -5.04507005e-01
2.46037647e-01 9.42594588e-01 -4.31694865e-01 -1.25372863e+00
-2.47042835e-01 5.51706791e-01 -7.53501654e-01 5.94962716e-01
8.25646877e-01 8.30080152e-01 8.25372994e-01 -1.27209651e+00
-6.14585996e-01 3.32893670e-01 -6.40008986e-01 -8.19210172e-01
1.11165881e+00 -1.97986916e-01 -6.72044694e-01 5.29345572e-01
1.00347567e+00 5.57151735e-01 -3.25715430e-02 -1.35507867e-01
1.94314435e-01 -2.53524631e-01 -3.27876240e-01 -1.56172299e+00
-5.57499230e-01 6.73363149e-01 9.09779847e-01 9.13790837e-02
9.32472765e-01 -5.42447031e-01 6.00357950e-01 5.07399201e-01
5.18795311e-01 -1.56819320e+00 -3.45280200e-01 1.38828075e+00
1.27828169e+00 -1.24448085e+00 -2.33045086e-01 3.98857594e-02
-1.08135247e+00 1.17933309e+00 8.72539759e-01 -5.25188819e-02
3.04010987e-01 -1.06694892e-01 9.08036649e-01 1.71101257e-01
-5.59220374e-01 -1.89391166e-01 3.52346808e-01 9.59714472e-01
1.14190257e+00 3.34539950e-01 -8.83341014e-01 6.10610366e-01
-6.51835442e-01 -6.10608697e-01 7.10916102e-01 4.52519506e-01
-2.97690868e-01 -1.21649885e+00 -6.59900188e-01 3.85045737e-01
-4.28076148e-01 -4.13931668e-01 -9.43585992e-01 8.59722078e-01
-9.52157825e-02 1.27325284e+00 2.58752882e-01 -7.22006559e-01
3.70440871e-01 2.04518512e-01 5.88706315e-01 -3.22785407e-01
-1.21426058e+00 -1.35584176e-01 7.11975217e-01 2.42699329e-02
-1.62443176e-01 -9.64518726e-01 -9.03775156e-01 -5.57216406e-01
-1.08024575e-01 3.92660409e-01 7.14231133e-01 6.93311751e-01
2.27398306e-01 -5.27150147e-02 9.36764181e-01 -4.89860982e-01
-5.33342957e-01 -1.68329775e+00 -4.90769416e-01 2.65082777e-01
1.33390546e-01 -2.44141012e-01 -2.83124764e-02 -1.24029100e-01] | [10.16650104522705, 10.767913818359375] |
e1cd970e-d493-4600-8ed9-33853f018ee0 | representation-learning-for-resource | null | null | https://openreview.net/forum?id=zlR44Dbobb7 | https://openreview.net/pdf?id=zlR44Dbobb7 | Representation Learning for Resource-Constrained Keyphrase Generation | State-of-the-art keyphrase generation methods generally depend on large annotated datasets, limiting their performance in domains with constrained resources. To overcome this challenge, we investigate pre-training strategies to learn an intermediate representation suitable for the keyphrase generation task. We introduce salient span recovery and salient span prediction as guided denoising language modeling objectives that condense the domain-specific knowledge essential for keyphrase generation. Through experiments on benchmarks spanning multiple domains, we show the effectiveness of the proposed approaches for facilitating low resource and zero-shot keyphrase generation. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 9.75590572e-02 -1.73893888e-02 -5.54388463e-01 2.75789976e-01
-1.21179390e+00 -6.55537844e-01 9.56863523e-01 6.59625411e-01
-4.38656956e-01 9.24985349e-01 7.33430624e-01 -1.63318276e-01
1.21050617e-02 -7.88469374e-01 -6.53519571e-01 -2.19610959e-01
1.85252614e-02 2.96016425e-01 3.70485902e-01 -4.73845840e-01
7.42885649e-01 3.24307323e-01 -1.56077600e+00 5.10527670e-01
8.84969294e-01 9.74688590e-01 5.29933155e-01 6.20025814e-01
-5.45064747e-01 9.98686433e-01 -7.80725539e-01 -1.53689235e-01
3.28567475e-01 -3.50845665e-01 -1.01531363e+00 8.07802752e-02
4.28737760e-01 -3.94901931e-01 -7.40088642e-01 6.99227095e-01
5.40336668e-01 5.25104642e-01 6.49738371e-01 -1.22392452e+00
-8.07073653e-01 1.04455626e+00 -4.93866384e-01 4.65777516e-01
5.41588485e-01 -1.68551132e-01 1.35252106e+00 -1.17001319e+00
8.42678010e-01 9.52499151e-01 2.10211709e-01 4.08166796e-01
-1.02643728e+00 -5.99507749e-01 3.44012737e-01 3.35665852e-01
-1.42100632e+00 -4.86838758e-01 1.08115137e+00 -9.92596820e-02
8.93045902e-01 5.76368831e-02 6.92226946e-01 1.05531967e+00
-2.18458138e-02 1.20348239e+00 9.98191833e-01 -5.52653432e-01
4.24390696e-02 1.74509920e-02 8.89505148e-02 5.12484550e-01
3.03580582e-01 -4.55559380e-02 -1.21430564e+00 -4.59845483e-01
6.87888980e-01 -1.20101541e-01 -1.85530916e-01 -1.09993853e-01
-1.45624197e+00 8.12056959e-01 -5.84054738e-02 9.79032665e-02
-6.79494202e-01 1.16958013e-02 7.09557533e-01 4.27318782e-01
6.69664681e-01 9.17886496e-01 -3.93624127e-01 -1.75924823e-01
-1.49852455e+00 9.09666777e-01 7.58980989e-01 1.13895905e+00
7.28092611e-01 -7.57628456e-02 -8.24933589e-01 9.09364581e-01
-2.43177220e-01 3.31688404e-01 5.26326597e-01 -5.51193535e-01
6.18763387e-01 6.74161077e-01 3.75337690e-01 -8.63434374e-01
4.10971604e-02 -5.96833080e-02 -5.88296771e-01 -5.78978300e-01
4.20788489e-02 5.02283461e-02 -1.08665073e+00 1.43927026e+00
2.00354919e-01 2.87021548e-01 1.38399050e-01 5.58812797e-01
9.67663407e-01 9.86400485e-01 1.31109983e-01 -3.14314246e-01
1.43152189e+00 -1.13987529e+00 -6.61783099e-01 -1.10438347e-01
3.58900964e-01 -1.02110720e+00 1.45723832e+00 2.77484804e-01
-1.26646698e+00 -4.96865034e-01 -9.15467918e-01 -2.63771921e-01
-4.81191963e-01 2.85876423e-01 5.13764918e-01 -2.86970497e-03
-5.91282845e-01 4.49570447e-01 -2.98940986e-01 -1.53411567e-01
2.33204722e-01 -2.14468583e-01 -9.13723558e-02 4.51715440e-02
-1.61246526e+00 7.24063158e-01 9.02734041e-01 -6.05849743e-01
-1.24765623e+00 -1.33693743e+00 -7.54269302e-01 -7.93160275e-02
7.75611222e-01 -6.93998575e-01 1.37276280e+00 -3.13885778e-01
-1.34140456e+00 1.00065088e+00 -9.44681689e-02 -7.36098528e-01
2.82437235e-01 -6.64352596e-01 -4.17126358e-01 5.84064662e-01
2.33596280e-01 8.55129421e-01 1.28338468e+00 -1.19476628e+00
-9.43180561e-01 2.25673601e-01 1.71153992e-01 3.79617602e-01
-8.22120130e-01 1.88015118e-01 -5.10304749e-01 -1.49199641e+00
-3.44006747e-01 -6.01749361e-01 -2.97405273e-01 -3.06713939e-01
-5.23612797e-01 -4.43303347e-01 8.84537995e-01 -8.40357006e-01
1.73375654e+00 -1.83930373e+00 1.56683728e-01 -2.92636622e-02
1.79530606e-01 9.90784690e-02 -3.32002163e-01 8.78682911e-01
1.98874921e-01 -4.73246239e-02 2.49881577e-02 -2.04953566e-01
8.98256451e-02 -1.35548934e-01 -1.22841370e+00 -1.01063974e-01
1.07839555e-01 8.37900758e-01 -1.02382636e+00 -8.48994732e-01
-6.21720627e-02 -2.26619486e-02 -2.68080294e-01 7.01779485e-01
-5.74180067e-01 8.16991627e-02 -5.90164185e-01 6.53206170e-01
3.13220263e-01 -2.46503606e-01 -2.30401278e-01 -3.50760162e-01
8.87972265e-02 5.36174953e-01 -1.06801593e+00 1.97514141e+00
-5.99823415e-01 3.62086535e-01 -3.38045269e-01 -7.99018621e-01
1.13938355e+00 1.89667061e-01 5.16206205e-01 -4.94780511e-01
-1.85439512e-01 8.95738378e-02 -6.80153489e-01 1.45814577e-02
1.45283532e+00 2.27672085e-02 -5.39656878e-01 7.21563041e-01
2.42642224e-01 -5.69680274e-01 6.28573537e-01 7.99337804e-01
9.54452634e-01 -6.64708391e-02 3.68148386e-01 -2.96459526e-01
4.27855194e-01 1.81590587e-01 6.56934977e-02 8.89900506e-01
1.44068539e-01 5.61701536e-01 3.71203095e-01 -4.04600561e-01
-1.28612792e+00 -9.34983313e-01 3.59928042e-01 1.50474727e+00
2.86527485e-01 -9.69050109e-01 -5.04816473e-01 -7.43967116e-01
-4.42981981e-02 8.18996966e-01 -4.46367025e-01 -2.33534187e-01
-5.80428064e-01 -2.09452853e-01 6.91861808e-01 3.88897300e-01
2.21015051e-01 -1.22456467e+00 -6.25512660e-01 4.02177602e-01
-5.20792425e-01 -1.40110421e+00 -8.08772027e-01 1.78953148e-02
-5.76378465e-01 -8.67121398e-01 -1.27447522e+00 -7.69505501e-01
5.72077036e-01 5.71326375e-01 1.54438436e+00 6.65173307e-03
-5.16644418e-01 5.71166754e-01 -7.59621501e-01 -4.27160501e-01
-4.09998536e-01 5.09084105e-01 1.56913444e-01 -3.65233958e-01
5.66782542e-02 -4.99215007e-01 -6.05367243e-01 -1.70100614e-01
-1.14555597e+00 3.27712983e-01 5.74927509e-01 7.67494261e-01
7.95152605e-01 9.51931030e-02 6.04579508e-01 -6.07551038e-01
1.43645155e+00 -4.35310602e-01 -5.03804088e-01 5.19012451e-01
-6.70161366e-01 3.05658638e-01 9.39019203e-01 -8.30423832e-01
-1.05851841e+00 -2.98944384e-01 5.35671711e-01 -5.31117141e-01
2.20676839e-01 6.46410823e-01 1.64126083e-01 2.74051219e-01
7.15607941e-01 8.20669651e-01 -5.66263139e-01 -4.84414726e-01
9.26300466e-01 3.73745620e-01 5.41286528e-01 -1.28811097e+00
9.59612370e-01 3.69806796e-01 -2.38420680e-01 -8.95043194e-01
-1.07691586e+00 -5.96962869e-01 -4.80620056e-01 -9.48929787e-03
3.64620417e-01 -1.25800276e+00 5.85730076e-02 3.77238125e-01
-1.12230670e+00 -1.81101397e-01 -6.95530236e-01 9.49301990e-04
-5.54387987e-01 7.36372888e-01 -5.23928821e-01 -4.83737648e-01
-1.14818668e+00 -7.19984412e-01 1.20597017e+00 2.25822344e-01
-4.16365176e-01 -6.35249496e-01 1.76410839e-01 1.43933827e-02
1.52858019e-01 4.06945683e-02 9.86296058e-01 -7.81537771e-01
-5.83238363e-01 -2.18174040e-01 -2.56172925e-01 1.09558329e-01
8.94417241e-02 -3.10802162e-01 -3.82529229e-01 -1.81084096e-01
-6.29486859e-01 -9.48777556e-01 1.24484897e+00 9.69576612e-02
1.36110163e+00 -5.40359795e-01 -2.32091919e-01 2.83383965e-01
1.15697849e+00 -2.63530076e-01 4.27285284e-01 4.40808237e-01
5.97336829e-01 5.01519322e-01 1.08796799e+00 1.07879412e+00
5.39717913e-01 5.05333483e-01 -2.37149134e-01 8.34054872e-02
-2.18270347e-01 -8.70172918e-01 3.46181363e-01 7.83019185e-01
1.33301124e-01 -1.47050709e-01 -7.17588127e-01 1.14990211e+00
-1.80345285e+00 -8.91102552e-01 4.82622683e-01 1.81494415e+00
1.53528225e+00 3.43233645e-01 3.29137862e-01 6.19718507e-02
3.99837345e-01 7.39909053e-01 -3.19451362e-01 -1.20248325e-01
-1.41237408e-01 4.69172657e-01 4.40313756e-01 1.47743672e-01
-1.03217518e+00 1.53751695e+00 6.58408833e+00 1.33783388e+00
-8.77906382e-01 -1.96513906e-01 3.81810457e-01 -4.50332835e-02
-6.70348108e-01 7.15699121e-02 -1.14260733e+00 2.20276430e-01
6.46953523e-01 -8.97708833e-01 4.41864729e-01 1.05076575e+00
7.04630744e-03 -8.40350688e-02 -9.51536059e-01 9.98641372e-01
1.31014278e-02 -1.72442973e+00 6.13871515e-01 -2.47939348e-01
8.75294983e-01 -4.95897174e-01 1.33893415e-01 3.34962308e-01
5.50652146e-01 -7.94405341e-01 8.21935952e-01 5.19039392e-01
8.95693004e-01 -9.35231090e-01 6.29696995e-02 2.50464976e-01
-1.48151696e+00 1.58447132e-01 -5.28574586e-01 1.76190332e-01
2.50203460e-01 7.25485325e-01 -9.90896642e-01 4.69374001e-01
3.83803666e-01 4.70022112e-01 -4.25514609e-01 6.95451975e-01
-4.72348630e-01 5.35397351e-01 -2.87873387e-01 -1.28701836e-01
2.93553978e-01 2.76996195e-01 6.80826247e-01 1.38786733e+00
4.22794223e-01 1.34236529e-01 5.97679019e-01 8.61044288e-01
-3.52733165e-01 2.61899441e-01 -4.10434753e-01 -6.67188048e-01
9.89107728e-01 1.30903566e+00 -9.06467140e-01 -6.88860774e-01
-1.65003225e-01 1.07652032e+00 2.33072087e-01 3.16129744e-01
-6.93740129e-01 -6.42226338e-01 6.64867282e-01 1.35206968e-01
3.10085684e-01 -3.72177392e-01 -6.04736619e-02 -1.35704994e+00
-7.15404302e-02 -1.15688074e+00 5.68935096e-01 -5.66957235e-01
-1.23923123e+00 5.24233818e-01 4.80080694e-01 -1.11219335e+00
-3.24586511e-01 -8.97695422e-02 -5.54129124e-01 6.70432925e-01
-1.76109564e+00 -1.65638411e+00 -1.80996045e-01 7.69274056e-01
1.02695799e+00 -4.96366143e-01 8.86402428e-01 -1.93217099e-01
-9.24958214e-02 6.91468298e-01 -3.44167292e-01 -1.07939281e-01
8.50136459e-01 -1.23112202e+00 8.23660791e-01 1.03419399e+00
4.78259653e-01 6.76637888e-01 8.34855139e-01 -7.62152433e-01
-1.46516728e+00 -9.69137669e-01 7.63552010e-01 -9.47903916e-02
9.46000099e-01 -2.75153071e-01 -7.77199030e-01 1.45048395e-01
3.15706462e-01 1.04382234e-02 4.75272685e-01 -5.09044677e-02
-5.97523928e-01 1.76416133e-02 -7.32339025e-01 9.57766294e-01
9.91974115e-01 -7.69644260e-01 -9.89486158e-01 2.64559478e-01
1.22461891e+00 -4.47200894e-01 -7.95652330e-01 2.88680434e-01
2.81789869e-01 -3.13795358e-01 1.30061710e+00 -7.16349244e-01
7.95597792e-01 -9.13213789e-02 -1.55345723e-01 -1.37232733e+00
-8.30161348e-02 -1.08582807e+00 -8.83748710e-01 1.18193316e+00
1.50037736e-01 1.18536450e-01 7.46982694e-01 9.20013189e-02
5.64326486e-03 -5.91603100e-01 -6.08168244e-01 -7.19507635e-01
-6.22165762e-02 -3.25106740e-01 6.26116216e-01 8.14157665e-01
1.30825624e-01 5.55813849e-01 -9.01298940e-01 -2.34384805e-01
5.57858586e-01 4.06065285e-01 1.04546356e+00 -9.03019249e-01
-2.38172501e-01 -3.33996117e-01 2.47535586e-01 -1.39183247e+00
4.70167547e-01 -6.09753370e-01 -1.29199058e-01 -1.44868207e+00
2.97982693e-01 -1.95332736e-01 -3.66963267e-01 5.02874255e-01
-5.76851606e-01 -2.32615396e-01 3.38563621e-01 3.10510397e-01
-7.81558812e-01 7.75945187e-01 1.31350374e+00 -3.12822700e-01
-4.47345734e-01 -3.02432805e-01 -9.01147366e-01 4.22319293e-01
6.15100145e-01 -3.58067036e-01 -8.08247328e-01 3.26194316e-02
3.17183793e-01 -4.91631217e-02 1.61386847e-01 -8.74902606e-01
1.89414784e-01 -7.19581604e-01 1.10568203e-01 -9.27967250e-01
2.52716333e-01 -4.02192980e-01 -3.42352718e-01 2.41221786e-01
-7.34211683e-01 5.34986630e-02 4.84129190e-01 5.29972613e-01
-3.05703968e-01 -1.37032568e-01 5.87124586e-01 -3.69731545e-01
-1.13612211e+00 4.94494647e-01 -2.77803898e-01 6.33464277e-01
9.08751547e-01 3.68130282e-02 -8.36893320e-02 -5.26150346e-01
-3.00316095e-01 -2.10207477e-02 2.36211166e-01 6.59397006e-01
1.12988257e+00 -1.33502221e+00 -9.07567739e-01 -8.67458582e-02
6.04871213e-01 -1.16689712e-01 5.44768274e-02 8.33338574e-02
-3.70473921e-01 4.28222656e-01 -1.06911533e-01 2.34177366e-01
-1.19616520e+00 7.76793420e-01 -3.09564173e-01 -9.80459094e-01
-7.28289962e-01 9.20301914e-01 -7.08967969e-02 -1.15376756e-01
3.71487081e-01 -4.11741704e-01 -1.51149035e-01 3.62200767e-01
8.62073481e-01 3.58366817e-01 -1.31423637e-01 -1.82840213e-01
-2.17451155e-02 1.73155338e-01 -5.99817634e-01 -3.46412718e-01
1.19165218e+00 -1.06198050e-01 -4.49556187e-02 1.81949273e-01
8.33474219e-01 3.93213741e-02 -9.23501551e-01 -6.94179237e-01
2.12931201e-01 -4.12812561e-01 1.75737321e-01 -5.73716640e-01
-4.40000504e-01 5.80314517e-01 1.28485173e-01 9.68750715e-02
1.31449521e+00 5.94736962e-03 1.58828449e+00 6.18702233e-01
4.93194222e-01 -1.51263142e+00 7.72458076e-01 6.43942356e-01
1.12715507e+00 -9.36300218e-01 4.28158909e-01 -1.61644757e-01
-8.53339493e-01 1.16957593e+00 5.83722532e-01 -1.37128860e-01
6.47201717e-01 9.60494354e-02 -3.11144114e-01 -8.62175748e-02
-9.99820173e-01 -3.14544946e-01 6.31806195e-01 4.29072917e-01
2.23744631e-01 -7.23037645e-02 -4.83563364e-01 8.02163720e-01
-2.60204256e-01 1.47979736e-01 4.79857177e-01 1.07102346e+00
-7.37582326e-01 -1.17233610e+00 -2.95485169e-01 3.77175927e-01
-6.52264237e-01 -5.37684619e-01 -5.33037007e-01 4.39362258e-01
-2.65191466e-01 6.28986418e-01 -2.82069176e-01 -2.55825400e-01
2.76629359e-01 -3.46583035e-03 4.37002659e-01 -7.92226911e-01
-4.29704219e-01 5.20995334e-02 1.27137318e-01 -3.25287938e-01
-2.30281934e-01 -6.29281849e-02 -1.08732045e+00 -4.38781708e-01
-1.50054500e-01 3.11517090e-01 1.42099157e-01 6.74815536e-01
5.14185488e-01 4.14660305e-01 6.05050802e-01 -6.73977137e-01
-7.89956570e-01 -1.05903971e+00 -6.37497425e-01 5.97185552e-01
1.86130077e-01 -4.39710945e-01 5.23304678e-02 4.13131565e-01] | [12.284310340881348, 8.89174747467041] |
5ffb0272-4bb9-42a9-8e96-e293142cd21e | linguistic-analysis-processing-line-for | null | null | https://aclanthology.org/L12-1494 | https://aclanthology.org/L12-1494.pdf | Linguistic Analysis Processing Line for Bulgarian | This paper presents a linguistic processing pipeline for Bulgarian including morphological analysis, lemmatization and syntactic analysis of Bulgarian texts. The morphological analysis is performed by three modules ― two statistical-based and one rule-based. The combination of these modules achieves the best result for morphological tagging of Bulgarian over a rich tagset (680 tags). The lemmatization is based on rules, generated from a large morphological lexicon of Bulgarian. The syntactic analysis is implemented via MaltParser. The two statistical morphological taggers and MaltParser are trained on datasets constructed within BulTreeBank project. The processing pipeline includes also a sentence splitter and a tokenizer. All tools in the pipeline are packed in modules that can also perform separately. The whole pipeline is designed to be able to serve as a back-end of a web service oriented interface, but it also supports the user tasks with a command-line interface. The processing pipeline is compatible with the Text Corpus Format, which allows it to delegate the management of the components to the WebLicht platform. | ['ar', 'Laska Laskova', 'Aleks Savkov', 'Stanislava Kancheva', 'Petya Osenova', 'Kiril Simov'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['morphological-tagging'] | ['natural-language-processing'] | [-2.22002372e-01 2.79565215e-01 4.11681771e-01 -4.95153606e-01
-8.98875296e-01 -8.16911817e-01 5.41359544e-01 7.55431592e-01
-9.02289093e-01 4.53558892e-01 1.89473480e-01 -6.63867772e-01
1.06011488e-01 -1.02673781e+00 -5.90181649e-02 -7.23639190e-01
3.03379837e-02 9.37721133e-01 5.72345138e-01 -2.10646510e-01
2.27289423e-02 6.41194820e-01 -1.10288429e+00 4.74603951e-01
6.90836489e-01 4.95387465e-01 5.53218842e-01 6.97224796e-01
-5.25990605e-01 2.44512379e-01 -4.25897628e-01 -7.40149558e-01
1.78678468e-01 -1.91462561e-01 -1.04352820e+00 -1.17206074e-01
-1.64097503e-01 2.62179226e-01 2.51097064e-02 1.07326150e+00
5.07078767e-01 8.53510052e-02 2.48511061e-01 -4.01344717e-01
2.77498603e-01 1.23808897e+00 2.12105185e-01 7.79602900e-02
2.90085316e-01 1.25450892e-02 9.12927270e-01 -6.94196880e-01
7.75944650e-01 1.47706854e+00 5.45218468e-01 1.17321163e-02
-8.61240029e-01 -3.23486984e-01 -4.14348602e-01 4.09599319e-02
-8.95273924e-01 -4.54443365e-01 3.25369120e-01 -4.90693331e-01
1.11825764e+00 1.91824108e-01 6.13585889e-01 3.06210041e-01
1.94571808e-01 4.05311972e-01 1.19497681e+00 -8.46177459e-01
3.20432484e-01 3.65230501e-01 3.34584981e-01 5.10087252e-01
4.54215646e-01 -3.40803951e-01 -4.03328501e-02 -1.88820586e-01
5.58694720e-01 -4.75114673e-01 3.30436885e-01 -4.25267639e-03
-1.00837708e+00 7.52048969e-01 -1.95550576e-01 7.90161192e-01
-4.98071283e-01 -1.91591859e-01 9.47364092e-01 2.64961243e-01
5.01438856e-01 1.83396578e-01 -6.50262058e-01 -2.13921472e-01
-9.66414034e-01 2.54711181e-01 1.19134247e+00 8.57007086e-01
6.84282482e-01 4.45242226e-02 2.64746726e-01 9.01134431e-01
4.05144811e-01 5.13866305e-01 7.44929492e-01 -4.13720012e-01
4.91121531e-01 7.46569157e-01 3.34970132e-02 -3.54580492e-01
-5.00415325e-01 1.11470558e-01 -2.35417366e-01 2.20087156e-01
7.06192195e-01 -4.34578001e-01 -8.37724745e-01 1.14936352e+00
6.51148319e-01 -6.01751626e-01 1.44321471e-01 4.70566839e-01
9.23100412e-01 4.20627236e-01 3.95201802e-01 -4.11456704e-01
2.04715657e+00 -3.52675766e-01 -7.83233345e-01 5.77748865e-02
9.09538567e-01 -1.49185205e+00 8.78863275e-01 4.38294977e-01
-1.51922536e+00 -3.30767095e-01 -7.97840297e-01 -2.07936719e-01
-8.33833516e-01 3.02936137e-01 4.96071339e-01 8.76595557e-01
-1.26002610e+00 6.38178349e-01 -1.07021666e+00 -7.98228741e-01
-3.52149129e-01 5.39196491e-01 -5.80932200e-01 3.98875505e-01
-1.02106786e+00 1.11711073e+00 1.13560319e+00 -8.59290585e-02
-8.21182504e-02 6.68469369e-02 -1.38612103e+00 -1.39551805e-02
1.64917439e-01 -2.54446507e-01 1.27601767e+00 -7.47383118e-01
-1.56795132e+00 1.38176847e+00 -5.28525234e-05 -4.74316210e-01
4.93157864e-01 1.84345528e-01 -2.94405699e-01 2.39750773e-01
1.86518133e-01 6.59094974e-02 3.23027939e-01 -8.11695993e-01
-9.42603111e-01 -5.43563485e-01 -3.46354038e-01 3.92513759e-02
6.88617676e-02 9.43046033e-01 -5.79957187e-01 -6.88360691e-01
1.20710857e-01 -7.25065172e-01 -2.73935884e-01 -1.33633614e+00
-6.46107197e-02 -2.24480242e-01 4.17954475e-01 -1.42160380e+00
1.25503981e+00 -1.96390498e+00 -2.43857190e-01 1.39216915e-01
-4.26177084e-01 5.43496072e-01 2.06539050e-01 7.69165933e-01
-2.64642119e-01 -6.94545582e-02 -1.72661275e-01 -3.95522386e-01
2.46986106e-01 4.38128948e-01 7.39626810e-02 5.23115754e-01
6.31863400e-02 5.37136555e-01 -1.02906013e+00 -7.35333920e-01
5.20150781e-01 1.73069090e-02 -1.06550075e-01 1.00504301e-01
-9.47122499e-02 1.21096134e-01 -3.12301576e-01 4.11059707e-01
9.76025522e-01 7.06472337e-01 9.97568011e-01 2.51685262e-01
-7.95897007e-01 9.65007305e-01 -1.37839675e+00 1.67842984e+00
-4.29627895e-01 -7.32234418e-02 5.42989194e-01 -9.14103031e-01
1.16046393e+00 6.42754972e-01 2.13871807e-01 1.48170069e-01
4.94632244e-01 5.39610386e-01 1.95758566e-02 -6.24017656e-01
9.04086649e-01 -3.11158955e-01 -5.67884564e-01 4.57381964e-01
5.17027855e-01 -2.78472513e-01 7.77755201e-01 5.76995425e-02
9.76241350e-01 5.44325411e-01 8.18944633e-01 -6.42166018e-01
6.87439263e-01 2.18777150e-01 8.29104006e-01 1.53349027e-01
1.32164761e-01 -4.23895158e-02 6.65084720e-01 -3.25060189e-01
-1.27011049e+00 -8.28172863e-01 -2.86195129e-01 1.22087204e+00
-6.30788326e-01 -6.51642323e-01 -1.03526187e+00 -6.23156071e-01
-2.99087703e-01 8.03664565e-01 -3.71020794e-01 4.69289273e-01
-8.57328534e-01 -7.28022635e-01 6.64511740e-01 1.99396551e-01
4.12676126e-01 -1.61827326e+00 -7.23568559e-01 5.98545253e-01
-1.42144218e-01 -1.00134647e+00 1.26118273e-01 5.19966185e-01
-8.91967356e-01 -7.92204976e-01 -1.11267097e-01 -1.14602757e+00
2.27072462e-01 -2.94441223e-01 9.74836886e-01 -1.71075374e-01
-5.32875396e-02 -4.97178286e-02 -6.81760609e-01 -7.01775193e-01
-9.87878859e-01 3.14104110e-01 -1.60278574e-01 -4.20851409e-01
6.72985911e-01 -5.89061737e-01 -6.16260581e-02 -4.68818322e-02
-1.01417267e+00 -2.34653398e-01 5.42335987e-01 2.77111620e-01
4.13993984e-01 -7.32041430e-03 2.88201243e-01 -1.19327748e+00
3.59358698e-01 -1.75218761e-01 -8.74473512e-01 9.54530686e-02
-2.68855363e-01 -1.50753036e-01 6.98640049e-01 5.09111658e-02
-1.31282365e+00 4.57629949e-01 -7.93788970e-01 3.36886466e-01
-8.55712295e-01 6.30088925e-01 -8.44566464e-01 3.02948684e-01
3.73657793e-02 2.19788719e-02 -2.60247350e-01 -1.14286268e+00
5.09539664e-01 9.13620114e-01 5.99390566e-01 -5.49116969e-01
6.39219165e-01 3.14989895e-01 -2.89711475e-01 -9.05280113e-01
-7.26751983e-02 -9.52590942e-01 -1.15044463e+00 -1.05396330e-01
9.89487112e-01 -6.35832369e-01 -3.50768477e-01 4.72383678e-01
-1.13290417e+00 -4.67850983e-01 -4.85158682e-01 5.13036609e-01
-4.94945467e-01 6.46672070e-01 -9.79440749e-01 -8.69087636e-01
-4.77099806e-01 -8.67044210e-01 6.43225670e-01 5.08350953e-02
-3.75122130e-01 -1.14357364e+00 3.27669829e-01 2.05680117e-01
-3.40995938e-01 1.01242669e-01 1.02921391e+00 -1.31784475e+00
2.73415923e-01 -2.84221023e-01 1.22546889e-01 3.01498026e-01
-1.62142009e-01 2.86210626e-01 -8.49214852e-01 -1.45286262e-01
1.49082392e-01 6.41659722e-02 7.05653667e-01 1.94649816e-01
1.06733844e-01 -3.34677100e-01 -8.84967446e-02 3.69999528e-01
1.46864653e+00 2.85220534e-01 8.37512851e-01 6.21273637e-01
2.05607459e-01 1.00657046e+00 8.77785623e-01 3.02592456e-01
4.40800428e-01 5.46666563e-01 9.08324867e-02 1.82295814e-01
1.57079458e-01 9.91034061e-02 6.83061957e-01 1.15380979e+00
-9.83498842e-02 2.33945459e-01 -1.30135500e+00 7.69230068e-01
-1.93654978e+00 -8.80963206e-01 -7.85177112e-01 2.31512332e+00
8.82984102e-01 1.33489847e-01 4.48530108e-01 4.26167309e-01
7.30723798e-01 -8.17485973e-02 5.66396117e-01 -1.15065670e+00
1.24928951e-01 6.87722683e-01 5.10280907e-01 8.72597694e-01
-1.27698171e+00 1.46852696e+00 5.39990187e+00 8.22322845e-01
-9.26015794e-01 4.15943921e-01 1.75813530e-02 2.97044158e-01
1.41083151e-01 2.97660977e-01 -1.08096373e+00 3.31673503e-01
1.34545732e+00 -7.48932958e-02 7.35582560e-02 7.30314016e-01
7.09645689e-01 -5.02727866e-01 -5.08915901e-01 4.21226561e-01
-1.75063536e-01 -1.11548221e+00 -1.81480303e-01 3.42652619e-01
7.72320405e-02 3.88217092e-01 -6.31726325e-01 3.38550299e-01
7.41415203e-01 -4.38857675e-01 1.03928804e+00 1.59130201e-01
7.65968621e-01 -8.34011674e-01 1.03070986e+00 2.13554576e-01
-1.28390241e+00 2.07846522e-01 -4.98333752e-01 -5.21985628e-02
5.36466241e-01 6.36238873e-01 -1.06694651e+00 8.94930959e-01
4.75954831e-01 1.28610909e-01 -4.87138450e-01 1.20543826e+00
-4.04397935e-01 8.53830397e-01 -2.91050434e-01 1.55583233e-01
3.97519767e-01 -7.62299061e-01 6.65760636e-01 2.18585849e+00
1.86433867e-01 1.37908846e-01 2.12988213e-01 -3.45581286e-02
3.98384035e-01 9.04266655e-01 -1.94582850e-01 9.75520387e-02
2.35784635e-01 1.80056119e+00 -1.34456623e+00 -5.89907169e-01
-2.11740419e-01 7.82184720e-01 2.00280800e-01 -1.87275440e-01
-3.03003788e-01 -8.76599193e-01 3.56634945e-01 2.44605169e-01
3.81381750e-01 -4.49893087e-01 -5.19087851e-01 -9.42781627e-01
-1.37354597e-01 -9.51453626e-01 8.26911986e-01 -5.17155707e-01
-5.62205076e-01 7.63371289e-01 1.07171327e-01 -4.76244599e-01
-4.77484256e-01 -8.70838761e-01 -6.35858297e-01 1.14345193e+00
-8.41616809e-01 -1.33635175e+00 1.47621647e-01 2.61054605e-01
4.05352712e-01 -2.56215967e-02 1.28900838e+00 2.37548113e-01
-5.11350453e-01 -1.56739876e-01 5.26089631e-02 6.04910553e-01
7.23109305e-01 -1.77294278e+00 4.91235793e-01 1.11994874e+00
-5.12601770e-02 7.13235259e-01 7.47317672e-01 -8.76735330e-01
-8.27153683e-01 -1.03228652e+00 1.84767258e+00 -2.05280066e-01
1.00439215e+00 -4.63911533e-01 -8.64458263e-01 6.99647009e-01
6.37458086e-01 -6.29317105e-01 9.52010751e-01 1.76500843e-03
5.58458269e-02 9.69369039e-02 -1.10992920e+00 1.30670398e-01
3.46437037e-01 -1.21037066e-01 -1.04178166e+00 2.38412723e-01
7.12365136e-02 -2.78207630e-01 -1.21352232e+00 -1.26097128e-01
2.67789304e-01 -6.90523744e-01 3.05622339e-01 -4.75331455e-01
-6.48304820e-02 -3.37858617e-01 2.18357787e-01 -1.06893599e+00
-2.72578925e-01 -9.95793939e-01 6.76742911e-01 1.69439006e+00
4.10356969e-01 -7.05464602e-01 3.21963042e-01 2.37969592e-01
-3.56955916e-01 1.51887551e-01 -1.07151520e+00 -6.10908926e-01
1.30160123e-01 -7.29970872e-01 5.42497039e-01 7.28087425e-01
8.51506054e-01 3.27205867e-01 3.15295279e-01 4.98280823e-02
2.62168348e-01 -1.94262937e-01 5.49748659e-01 -1.23985040e+00
-3.75971138e-01 -4.95815217e-01 -4.40080911e-01 -1.92979082e-01
1.31269276e-01 -1.07409155e+00 5.99079318e-02 -1.63842440e+00
-1.79175422e-01 -4.31948423e-01 1.38802558e-01 6.91183984e-01
4.89873886e-02 2.39533544e-01 7.89062753e-02 1.48777351e-01
-2.17894867e-01 -1.60125762e-01 4.85703230e-01 3.53098184e-01
-3.80987823e-01 -5.47822937e-02 -2.29135230e-01 1.10686171e+00
9.86287177e-01 -6.51337683e-01 2.47694954e-01 -1.47700757e-01
3.87635976e-01 -2.55371690e-01 -4.70943786e-02 -8.36241007e-01
-4.83213626e-02 -6.61288798e-02 -4.75885049e-02 -6.80138052e-01
4.16381732e-02 -5.67350388e-01 1.77061468e-01 4.86280471e-01
1.77785978e-01 3.73802215e-01 3.36912721e-01 -6.18681833e-02
-1.18326612e-01 -9.18113351e-01 1.09231472e+00 -5.03630221e-01
-4.48970854e-01 -8.42979923e-02 -1.06800115e+00 -1.25927687e-01
1.01602197e+00 -1.51068673e-01 1.28889799e-01 7.23869652e-02
-1.19484198e+00 -1.89121380e-01 7.59339809e-01 -2.72231430e-01
-1.00323558e-01 -8.24460745e-01 -8.27103198e-01 1.60161451e-01
-1.47211358e-01 -1.38122737e-01 -8.85703787e-02 8.51366103e-01
-1.22885096e+00 6.53259397e-01 -2.19499975e-01 1.58200160e-01
-1.55356610e+00 5.93156874e-01 4.54626158e-02 -7.44548976e-01
-5.41215360e-01 4.49167490e-01 -4.80700374e-01 -4.27717179e-01
-2.02491239e-01 -4.70998853e-01 -5.32919645e-01 5.83171725e-01
5.44350982e-01 4.43929583e-01 5.30543864e-01 -1.18563044e+00
-5.47063172e-01 8.18474740e-02 2.53607988e-01 -7.07935154e-01
1.55540156e+00 -4.17092890e-02 -7.08216250e-01 4.01499927e-01
5.64185619e-01 6.53721869e-01 -5.55360675e-01 3.60274091e-02
6.42979026e-01 1.18336126e-01 -3.80664393e-02 -6.36598885e-01
-4.04406250e-01 7.12323248e-01 -9.10469983e-03 5.21723807e-01
8.27435434e-01 -6.38073683e-02 5.91426909e-01 3.54406834e-02
3.61645311e-01 -1.42650807e+00 -1.05361295e+00 8.70658040e-01
4.68553126e-01 -3.93719226e-01 -2.62147307e-01 -7.36339629e-01
-6.21384203e-01 1.34741223e+00 -2.64477462e-01 -1.48786560e-01
6.01434529e-01 7.38387585e-01 4.45837468e-01 9.10795818e-04
-5.43843448e-01 -8.46531868e-01 -6.32224604e-02 6.12278044e-01
8.09895992e-01 2.64058262e-01 -1.35713863e+00 9.22690392e-01
-7.10181892e-01 -1.22357965e-01 5.94739258e-01 6.50955737e-01
-6.04622781e-01 -1.76949072e+00 -6.86215878e-01 4.45954800e-02
-1.12456739e+00 -4.85552430e-01 -5.19982457e-01 9.22214329e-01
4.62080449e-01 9.71370280e-01 1.35026410e-01 -1.12795174e-01
2.79227406e-01 6.41022742e-01 3.58516902e-01 -1.08876050e+00
-1.41428995e+00 6.54463053e-01 8.34297776e-01 -2.28648171e-01
-2.12731063e-01 -1.17452562e+00 -1.26628375e+00 -3.10484707e-01
-1.31440505e-01 7.66116917e-01 9.22679782e-01 9.10774946e-01
-2.56702483e-01 5.46072610e-02 3.00106220e-02 -8.14356029e-01
-2.40259126e-01 -1.35982871e+00 -1.00566244e+00 1.99173823e-01
-6.14875317e-01 -6.48551956e-02 2.19265129e-02 5.73198199e-01] | [10.297926902770996, 10.203086853027344] |
a212e04f-e3e9-4fd1-be12-20114b4309e1 | disentangled-latent-transformer-for | 2201.06357 | null | https://arxiv.org/abs/2201.06357v2 | https://arxiv.org/pdf/2201.06357v2.pdf | Disentangled Latent Transformer for Interpretable Monocular Height Estimation | Monocular height estimation (MHE) from remote sensing imagery has high potential in generating 3D city models efficiently for a quick response to natural disasters. Most existing works pursue higher performance. However, there is little research exploring the interpretability of MHE networks. In this paper, we target at exploring how deep neural networks predict height from a single monocular image. Towards a comprehensive understanding of MHE networks, we propose to interpret them from multiple levels: 1) Neurons: unit-level dissection. Exploring the semantic and height selectivity of the learned internal deep representations; 2) Instances: object-level interpretation. Studying the effects of different semantic classes, scales, and spatial contexts on height estimation; 3) Attribution: pixel-level analysis. Understanding which input pixels are important for the height estimation. Based on the multi-level interpretation, a disentangled latent Transformer network is proposed towards a more compact, reliable, and explainable deep model for monocular height estimation. Furthermore, a novel unsupervised semantic segmentation task based on height estimation is first introduced in this work. Additionally, we also construct a new dataset for joint semantic segmentation and height estimation. Our work provides novel insights for both understanding and designing MHE models. | ['Sining Chen', 'Zhitong Xiong', 'Xiao Xiang Zhu', 'Yilei Shi'] | 2022-01-17 | null | null | null | null | ['unsupervised-semantic-segmentation'] | ['computer-vision'] | [ 4.44823235e-01 5.26597738e-01 -3.19945157e-01 -5.78575015e-01
-4.64508623e-01 -1.29557803e-01 4.80624557e-01 9.60893929e-02
6.10907786e-02 5.85473657e-01 5.16406953e-01 -2.51463532e-01
-2.51005203e-01 -1.50644708e+00 -8.37798595e-01 -8.07773888e-01
5.84309958e-02 5.62613964e-01 -9.53642651e-02 -8.28483105e-02
1.10850818e-01 5.49810648e-01 -1.73096824e+00 9.39212516e-02
1.05228961e+00 1.07983577e+00 3.77923757e-01 5.14378130e-01
6.69154748e-02 4.63671058e-01 -1.67596534e-01 -5.65552786e-02
3.87698740e-01 -2.17017338e-01 -9.01924968e-01 4.48966503e-01
4.69935089e-01 -5.29227734e-01 -7.35058635e-02 1.01599038e+00
1.91500351e-01 -1.48369312e-01 7.29959071e-01 -1.37106788e+00
-9.50658917e-01 1.03048229e+00 -6.09325588e-01 -2.23871693e-01
-1.81436837e-01 1.43394485e-01 1.13922560e+00 -6.47938371e-01
2.29865029e-01 1.30089629e+00 7.07373440e-01 4.29857112e-02
-1.20012498e+00 -5.45870483e-01 4.31630045e-01 1.03177987e-01
-1.45571959e+00 -1.02474108e-01 7.60110438e-01 -5.55658937e-01
5.77221334e-01 2.49071345e-01 9.38064158e-01 8.10693622e-01
-1.33118927e-01 1.02063501e+00 1.12834418e+00 -2.04541519e-01
7.01527521e-02 -2.06713900e-01 -2.11279970e-02 6.68988407e-01
4.81969953e-01 4.78292145e-02 -2.93649346e-01 3.68640453e-01
1.21629822e+00 7.26302192e-02 -1.57259777e-01 -3.85594517e-01
-1.10236168e+00 1.09234846e+00 9.69713509e-01 1.63872749e-01
-4.15610135e-01 2.45825812e-01 -1.02582380e-01 -2.69995511e-01
7.76535988e-01 4.80431676e-01 -5.63891888e-01 5.20201445e-01
-1.27237844e+00 2.96519727e-01 1.00674622e-01 6.69288576e-01
1.09616399e+00 1.89069763e-01 -3.79815623e-02 7.14606822e-01
4.68039632e-01 6.64107621e-01 7.45947734e-02 -9.16946352e-01
6.35079384e-01 8.92768562e-01 -1.88824028e-01 -1.16567004e+00
-8.15384746e-01 -7.18514085e-01 -1.11451232e+00 1.02019049e-01
2.64613181e-01 1.40552178e-01 -9.47284460e-01 1.87628329e+00
3.56198214e-02 -6.69453368e-02 -8.85055214e-02 1.13287294e+00
8.24918747e-01 4.85566050e-01 3.04903001e-01 3.80697548e-01
1.24823725e+00 -7.59030938e-01 -2.31325954e-01 -6.87476933e-01
6.58239543e-01 -2.56186128e-01 9.02887762e-01 -4.24415916e-02
-1.08615649e+00 -5.95032275e-01 -1.09290099e+00 -5.01225948e-01
-6.33982360e-01 5.46219349e-01 8.94253910e-01 4.04765785e-01
-1.02017546e+00 5.40582895e-01 -7.11316049e-01 -3.82614374e-01
5.73115766e-01 1.32045820e-01 5.68324700e-02 1.37959659e-01
-1.38925898e+00 6.19469941e-01 5.09212673e-01 2.76329249e-01
-4.73732829e-01 -6.12476766e-01 -1.04840183e+00 2.68128097e-01
-8.08399543e-03 -1.12105882e+00 8.25699151e-01 -8.29262018e-01
-9.36232209e-01 1.20539105e+00 -1.37373403e-01 -5.78716695e-01
2.42339760e-01 -9.45839882e-02 1.51711658e-01 9.06330943e-02
4.90945280e-01 1.34062552e+00 5.09816825e-01 -1.43425739e+00
-6.06935263e-01 -1.06805670e+00 1.00240171e-01 4.57871914e-01
-2.41128393e-02 -6.85892880e-01 1.73603356e-01 -5.35679460e-01
8.54386508e-01 -8.22685897e-01 -2.47376427e-01 4.52463664e-02
-6.31502271e-01 4.86102998e-02 4.60360169e-01 -8.93752992e-01
9.66233432e-01 -1.88821995e+00 2.29945377e-01 7.16239363e-02
6.11208320e-01 -1.15082599e-01 1.84347257e-01 1.66908607e-01
-8.70771823e-04 3.73238921e-01 -5.01542985e-01 -2.93552667e-01
2.11312518e-01 8.95253271e-02 -4.32630867e-01 2.89161205e-01
5.85064925e-02 1.30288732e+00 -5.89771926e-01 -4.04005587e-01
5.49381196e-01 4.68780756e-01 -6.40604794e-01 8.06937218e-02
-1.37106970e-01 5.53101897e-01 -4.46228325e-01 6.13737881e-01
9.01782453e-01 -4.11405891e-01 7.68569633e-02 -3.97929221e-01
-3.62663418e-01 8.43890533e-02 -8.32190037e-01 1.46666586e+00
-5.68131328e-01 8.48678529e-01 -7.76630044e-02 -1.07147455e+00
1.08964992e+00 4.62251343e-03 2.97155648e-01 -7.20644236e-01
7.88866505e-02 1.16036162e-01 -3.55394900e-01 -2.05681100e-01
7.06920862e-01 -1.44008338e-01 -1.70443416e-01 1.01552442e-01
-4.08663094e-01 -3.09469163e-01 -2.36449122e-01 -4.22768993e-03
3.42893481e-01 1.64132684e-01 2.54703313e-01 -2.83925086e-01
1.44392148e-01 -1.08452998e-01 3.36887389e-01 6.01953506e-01
-2.46566370e-01 7.95753598e-01 2.91594267e-01 -7.74108827e-01
-1.08963370e+00 -1.28917754e+00 -2.84883082e-01 9.98764873e-01
3.00723732e-01 1.00367345e-01 -8.57200503e-01 1.02008723e-01
2.95040328e-02 7.19392121e-01 -8.12801301e-01 -1.64553765e-02
-1.54228076e-01 -9.77595747e-01 7.27177858e-01 9.51925159e-01
1.23170841e+00 -9.91869509e-01 -8.62999260e-01 -5.49007691e-02
-6.79695308e-01 -1.38007092e+00 1.11942865e-01 -4.00628634e-02
-1.16381466e+00 -8.32994819e-01 -5.40644765e-01 -5.25751770e-01
5.78280985e-01 4.94770169e-01 1.13848281e+00 3.42280343e-02
-1.74513221e-01 1.42469052e-02 -8.85542780e-02 -3.95768851e-01
4.95645702e-02 5.83797991e-01 -3.20722818e-01 -1.07822016e-01
3.51132065e-01 -7.85652995e-01 -9.02493894e-01 1.53002545e-01
-8.11107218e-01 8.68922710e-01 8.18198562e-01 2.80477226e-01
5.41969597e-01 2.99434625e-02 1.95658028e-01 -6.90599859e-01
1.98705524e-01 -5.90247095e-01 -5.83297312e-01 8.78454149e-02
-6.90707564e-01 1.08605675e-01 2.60413378e-01 2.78939188e-01
-1.22090364e+00 -1.69419814e-02 -2.49966741e-01 -7.14446232e-02
-6.33721054e-01 4.61367577e-01 -3.33125323e-01 3.69031370e-01
4.85518664e-01 2.87979573e-01 -3.69903684e-01 -3.44074547e-01
5.83170950e-01 5.61404884e-01 4.81954515e-01 -4.39942122e-01
7.94280946e-01 7.84679651e-01 2.18412310e-01 -1.04936934e+00
-1.27075815e+00 -3.80662829e-01 -1.03462982e+00 -9.06948075e-02
1.41491532e+00 -1.22316444e+00 -5.02547443e-01 6.86754465e-01
-9.37241495e-01 -4.99848664e-01 -7.83187896e-02 2.94746548e-01
-6.25289321e-01 1.55263156e-01 -4.96922344e-01 -7.36949325e-01
-2.20916286e-01 -1.14839661e+00 1.85321212e+00 2.01738253e-01
-7.90811256e-02 -1.18760884e+00 -1.62993446e-01 8.31534863e-01
3.63380462e-01 5.04574239e-01 1.07201707e+00 1.25405654e-01
-1.04561269e+00 3.31861734e-01 -8.62344921e-01 -1.83294371e-01
-1.97442353e-01 -5.50179221e-02 -1.21203816e+00 1.85478836e-01
-3.39236885e-01 -2.82370985e-01 1.15471411e+00 1.10823083e+00
1.50595081e+00 -2.72359043e-01 -1.73527241e-01 1.18290615e+00
1.47526085e+00 -2.33713642e-01 9.27117288e-01 5.80519795e-01
1.07396853e+00 9.55451369e-01 2.74597138e-01 4.29130524e-01
8.64483118e-01 5.76226115e-01 7.58135617e-01 -7.32986033e-01
-1.73223503e-02 -7.42874444e-01 -1.89210296e-01 5.34416437e-01
-2.26673111e-01 -5.81182502e-02 -1.10146236e+00 5.54119825e-01
-1.79978156e+00 -9.03547883e-01 -4.39187974e-01 1.86984062e+00
1.58203408e-01 -1.04889125e-01 -8.87684971e-02 2.98128664e-01
6.47803009e-01 5.38019538e-01 -6.14766896e-01 -1.41552240e-01
-5.38901031e-01 -5.42150214e-02 7.30468690e-01 5.98151982e-01
-1.35530603e+00 1.17553258e+00 5.54293346e+00 5.21745324e-01
-1.05764699e+00 -8.68344083e-02 9.63567138e-01 3.45059961e-01
-6.56836271e-01 8.03839937e-02 -7.81823218e-01 -8.04949477e-02
5.03535032e-01 2.46732786e-01 2.11022258e-01 8.23789597e-01
3.82500023e-01 -1.38202950e-01 -6.38838649e-01 9.37579632e-01
-1.54959008e-01 -1.37605023e+00 2.54436493e-01 4.62566465e-01
8.35228086e-01 1.16861865e-01 2.41491556e-01 1.13049947e-01
2.42676780e-01 -1.19931006e+00 9.02296007e-01 3.95747602e-01
7.84143209e-01 -5.57686448e-01 5.41528940e-01 4.78466451e-01
-1.70560825e+00 -6.39493540e-02 -4.47512388e-01 -5.00405967e-01
1.01606563e-01 7.57488191e-01 -5.38800180e-01 5.40050626e-01
5.05350351e-01 7.20404446e-01 -6.51100397e-01 4.91225123e-01
-5.00184000e-01 4.34990406e-01 -1.01084244e-02 4.25788075e-01
3.92608881e-01 -3.26346636e-01 1.00229271e-01 8.73796225e-01
2.67369598e-01 2.52386063e-01 2.23159701e-01 1.48407328e+00
1.20451078e-01 1.24072827e-01 -7.52191901e-01 4.03221697e-02
3.14177394e-01 9.93078947e-01 -1.05486882e+00 -1.87303632e-01
-4.91488874e-02 7.91227698e-01 4.12010103e-01 3.57778192e-01
-8.68628144e-01 1.22596532e-01 6.14260674e-01 3.53684038e-01
4.13041078e-02 -3.62506479e-01 -9.27423954e-01 -1.20953655e+00
-3.31817240e-01 -2.64395595e-01 5.76671511e-02 -1.11512649e+00
-8.53261530e-01 2.37970188e-01 2.47156069e-01 -6.80501997e-01
2.08569355e-02 -4.99897420e-01 -4.65063244e-01 8.11941206e-01
-1.59857416e+00 -1.56789505e+00 -7.93784618e-01 1.74219117e-01
6.15904868e-01 3.30577195e-01 6.83051765e-01 1.81025695e-02
-6.00759208e-01 6.86833188e-02 -3.36945653e-01 2.32322097e-01
-2.32732315e-02 -1.25676334e+00 7.95981526e-01 7.14093328e-01
1.40445843e-01 2.94118077e-01 4.49488372e-01 -5.87258995e-01
-6.90462887e-01 -1.29953933e+00 1.07022285e+00 -3.45910460e-01
2.26441428e-01 -2.12006941e-01 -7.93114901e-01 7.84645498e-01
-3.11022073e-01 -5.83112478e-01 5.17978311e-01 3.05321276e-01
-1.64662838e-01 1.13051564e-01 -8.66612375e-01 5.59592009e-01
1.35219562e+00 -6.07290983e-01 -3.76570970e-01 1.97085306e-01
7.60101557e-01 -1.22520886e-01 -7.05217719e-01 5.91305554e-01
5.99457145e-01 -1.38132846e+00 1.17665184e+00 -2.25236878e-01
9.90485847e-01 -4.56850976e-03 -4.21301097e-01 -1.12741566e+00
-6.02445245e-01 3.93631846e-01 3.48367125e-01 7.69023418e-01
3.93998325e-01 -5.11039734e-01 1.03490591e+00 4.83844310e-01
-2.55227774e-01 -8.32754493e-01 -4.97074366e-01 -3.77886802e-01
2.72114545e-01 -6.20827317e-01 9.99999404e-01 8.45650315e-01
-5.80296993e-01 4.00752068e-01 -3.49440604e-01 6.64407969e-01
9.30303812e-01 5.72963655e-01 7.52457201e-01 -1.54157364e+00
-5.17752953e-02 -6.18236661e-01 -3.65815490e-01 -1.25342715e+00
1.83316067e-01 -9.39437807e-01 -1.65629148e-01 -1.97255659e+00
4.30095494e-01 -3.04366082e-01 2.53428578e-01 4.46491033e-01
-1.53876290e-01 3.32483828e-01 2.39053205e-01 3.18379760e-01
-1.94460481e-01 8.94480824e-01 1.18244314e+00 -1.50992960e-01
9.99315754e-02 -4.20171395e-02 -6.81219220e-01 9.54843640e-01
1.08730769e+00 -3.65034826e-02 -5.83541989e-01 -7.97740579e-01
3.94790202e-01 1.56783983e-01 8.32702577e-01 -9.50555027e-01
-1.96485877e-01 -1.99362636e-01 6.26079082e-01 -9.15830374e-01
3.73528302e-01 -4.70014274e-01 4.05502245e-02 3.79822224e-01
-2.87314475e-01 -2.56726384e-01 -4.18994091e-02 2.15895295e-01
-1.49768591e-01 6.68963045e-02 7.05745518e-01 -4.57098156e-01
-8.88556123e-01 5.38349569e-01 -2.12431610e-01 1.30111307e-01
5.75175881e-01 -6.11599267e-01 -2.19051719e-01 -5.74897885e-01
-9.25615072e-01 1.50689140e-01 4.71428454e-01 3.15271109e-01
6.86906755e-01 -1.18295622e+00 -6.45759046e-01 3.55013371e-01
9.10143331e-02 1.19788870e-01 4.18114692e-01 4.92677450e-01
-5.84520400e-01 6.93480313e-01 -3.09751332e-01 -7.48760700e-01
-8.07366669e-01 6.94654956e-02 4.31483686e-01 -3.14519584e-01
-4.73113090e-01 7.42974102e-01 8.74061704e-01 -7.86111534e-01
-1.97110265e-01 -5.27381420e-01 -2.37650946e-01 1.40546426e-01
-1.88326333e-02 4.14085835e-01 -2.68209457e-01 -9.43052948e-01
-2.23764572e-02 8.74041855e-01 6.86161518e-01 -1.29764304e-01
1.31450689e+00 -5.58806896e-01 4.88373861e-02 5.47727764e-01
1.05557597e+00 -5.68555474e-01 -1.33994317e+00 -6.19530007e-02
-2.28756413e-01 -4.92882788e-01 3.78987342e-01 -4.89444941e-01
-1.31471193e+00 1.30993950e+00 5.83962023e-01 1.73050594e-02
1.13475096e+00 6.51864707e-02 7.87448406e-01 5.98070733e-02
3.84365976e-01 -9.11827028e-01 -2.01296419e-01 4.47553188e-01
8.53954315e-01 -1.60406673e+00 1.50230890e-02 -6.12665892e-01
-4.51513052e-01 9.74425673e-01 6.18562639e-01 2.00793654e-01
5.52934587e-01 -2.32731789e-01 -1.10913537e-01 -6.00210786e-01
-1.96857110e-01 -7.42026687e-01 4.02743638e-01 5.70334554e-01
4.52673405e-01 7.57494748e-01 1.24254778e-01 4.30233419e-01
-7.21373618e-01 -3.06256056e-01 2.64294237e-01 2.99799651e-01
-7.94240296e-01 -5.66785812e-01 -4.03264880e-01 2.75778919e-01
1.91118896e-01 -3.01249743e-01 -4.83950108e-01 7.86544144e-01
4.58824813e-01 7.48759568e-01 1.34468675e-01 -2.85859257e-01
1.01427399e-01 -1.16013370e-01 3.97725016e-01 -5.90627730e-01
3.88329327e-02 -1.29787698e-01 -1.57289878e-01 -4.62650210e-01
-5.19569933e-01 -4.96255547e-01 -1.16794753e+00 -3.33855540e-01
-5.61730713e-02 -2.12019756e-01 7.66197264e-01 1.11614144e+00
9.60451663e-02 4.39450204e-01 4.24586445e-01 -1.13946354e+00
-5.80069162e-02 -7.62651443e-01 -7.26986408e-01 3.96354496e-01
1.14183836e-02 -7.86511838e-01 -3.80459785e-01 -1.47076443e-01] | [9.154725074768066, -1.3824307918548584] |
6f945a0a-1cb4-4425-a160-5033bd7b3cdf | evd-surgical-guidance-with-retro-reflective | 2306.15490 | null | https://arxiv.org/abs/2306.15490v2 | https://arxiv.org/pdf/2306.15490v2.pdf | EVD Surgical Guidance with Retro-Reflective Tool Tracking and Spatial Reconstruction using Head-Mounted Augmented Reality Device | Augmented Reality (AR) has been used to facilitate surgical guidance during External Ventricular Drain (EVD) surgery, reducing the risks of misplacement in manual operations. During this procedure, the key challenge is accurately estimating the spatial relationship between pre-operative images and actual patient anatomy in AR environment. This research proposes a novel framework utilizing Time of Flight (ToF) depth sensors integrated in commercially available AR Head Mounted Devices (HMD) for precise EVD surgical guidance. As previous studies have proven depth errors for ToF sensors, we first assessed their properties on AR-HMDs. Subsequently, a depth error model and patient-specific parameter identification method are introduced for accurate surface information. A tracking pipeline combining retro-reflective markers and point clouds is then proposed for accurate head tracking. The head surface is reconstructed using depth data for spatial registration, avoiding fixing tracking targets rigidly on the patient's skull. Firstly, $7.580\pm 1.488 mm$ depth value error was revealed on human skin, indicating the significance of depth correction. Our results showed that the error was reduced by over $85\%$ using proposed depth correction method on head phantoms in different materials. Meanwhile, the head surface reconstructed with corrected depth data achieved sub-millimetre accuracy. An experiment on sheep head revealed $0.79 mm$ reconstruction error. Furthermore, a user study was conducted for the performance in simulated EVD surgery, where five surgeons performed nine k-wire injections on a head phantom with virtual guidance. Results of this study revealed $2.09 \pm 0.16 mm$ translational accuracy and $2.97\pm 0.91$ degree orientational accuracy. | ['Guangzhi Wang', 'Hui Ding', 'Zhe Zhao', 'Yihao Liu', 'Yuxing Yang', 'Long Qian', 'Du Liu', 'Wenqing Yan', 'Haowei Li'] | 2023-06-27 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [-2.92602360e-01 4.46867257e-01 4.57390875e-01 1.80219412e-02
-8.27485204e-01 -2.25027636e-01 -4.66968827e-02 1.61850959e-01
-5.34123600e-01 5.05055606e-01 2.85666399e-02 -4.94271487e-01
8.28452185e-02 -3.65131795e-01 -5.31766295e-01 -5.74556589e-01
-3.29701990e-01 4.47469801e-01 1.33685768e-01 -4.27239621e-03
1.73721299e-01 6.30145729e-01 -9.02767658e-01 -3.13110650e-01
5.53839147e-01 9.48459268e-01 6.33663893e-01 5.01915753e-01
3.61089528e-01 2.20630482e-01 -5.60162902e-01 -2.16371879e-01
5.43433845e-01 3.07483226e-03 -1.40000165e-01 -2.47307777e-01
1.00954495e-01 -5.80784023e-01 -3.34683895e-01 8.13047886e-01
9.67392087e-01 -2.20314816e-01 3.65678132e-01 -5.19372523e-01
1.16965659e-01 3.86954658e-03 -6.07647657e-01 8.59672353e-02
5.70015788e-01 8.22287332e-03 -2.48461768e-01 -9.77603197e-01
7.99834311e-01 2.99873859e-01 1.00617397e+00 8.26287985e-01
-6.03548646e-01 -1.07971215e+00 -4.44848359e-01 -6.95165575e-01
-1.40176380e+00 -2.71237880e-01 4.82970655e-01 -7.39604950e-01
9.94283795e-01 3.90622169e-01 1.03825617e+00 4.09646213e-01
1.30950272e+00 -1.20630547e-01 1.37675238e+00 -3.98309469e-01
1.42901808e-01 5.64346313e-01 -4.25339155e-02 9.23695505e-01
6.10425591e-01 6.54936373e-01 -2.31934652e-01 -2.73041606e-01
1.31625056e+00 7.36093074e-02 -8.42108190e-01 -5.71577609e-01
-9.33059573e-01 4.09493804e-01 5.13197541e-01 3.60736489e-01
-3.41928720e-01 -1.49988711e-01 3.29519004e-01 -3.22174191e-01
1.97737232e-01 2.85801172e-01 -4.00866091e-01 -3.33603293e-01
-3.68237227e-01 -2.92828172e-01 4.56077069e-01 1.21518302e+00
-1.72466099e-01 4.20238152e-02 2.94213712e-01 3.63631815e-01
8.22033405e-01 7.03612387e-01 6.54503345e-01 -4.68249798e-01
2.89889485e-01 3.99074227e-01 3.12688082e-01 -7.77341425e-01
-8.53061497e-01 -5.91969490e-01 -8.27197671e-01 3.51870894e-01
-4.43220437e-02 -4.95896190e-02 -1.16808307e+00 1.12405610e+00
6.49796069e-01 3.16866755e-01 -4.38208971e-03 1.00793493e+00
1.12028360e+00 -4.57387306e-02 -1.56558212e-02 -6.60198390e-01
1.54970431e+00 -1.80530578e-01 -9.85412061e-01 -9.30753946e-02
1.15609491e+00 -8.39213431e-01 8.78790319e-01 3.18649888e-01
-1.28366816e+00 -2.29989409e-01 -9.91077244e-01 5.04811049e-01
3.45926762e-01 -1.91978812e-02 5.31889081e-01 1.00358784e+00
-1.03953826e+00 3.03947538e-01 -1.22039521e+00 -8.98246691e-02
2.32072815e-01 7.64321148e-01 -6.20937109e-01 1.41605914e-01
-6.66893482e-01 1.18672347e+00 -2.38893136e-01 3.58513564e-01
-5.40349543e-01 -1.04488826e+00 -7.98994720e-01 -8.38178635e-01
-3.84867102e-01 -9.31503952e-01 1.15209627e+00 1.25826240e-01
-1.79720891e+00 1.36623693e+00 -6.22120239e-02 -3.66763592e-01
5.62489152e-01 -3.20862800e-01 -5.43535590e-01 6.90875724e-02
-2.16495618e-02 -1.39473574e-02 3.67745429e-01 -1.31376731e+00
-9.57369134e-02 -1.04428267e+00 -3.60512584e-01 3.46379131e-01
4.85817403e-01 1.47529244e-01 -2.97278374e-01 -5.40391743e-01
1.01695526e+00 -1.20291162e+00 -2.87732691e-01 7.60206953e-02
-6.71489388e-02 5.78182876e-01 4.72636759e-01 -9.14389193e-01
1.12559748e+00 -2.05959201e+00 -4.33437735e-01 2.77612090e-01
6.18919969e-01 7.44035169e-02 8.06261599e-01 -1.61299095e-01
-1.66178823e-01 -1.53328285e-01 -1.19350821e-01 -3.51086438e-01
-5.21314621e-01 -2.18035087e-01 2.43700653e-01 1.03719962e+00
-1.10600102e+00 9.03485358e-01 -5.02822995e-01 -3.38774323e-01
3.22681755e-01 8.27230036e-01 -3.74681205e-01 2.75308102e-01
7.37064183e-01 1.11217892e+00 -2.52350926e-01 7.75768459e-01
9.69127774e-01 7.42518306e-02 -1.16470948e-01 -2.27840677e-01
-2.61451125e-01 -3.30638029e-02 -8.38326156e-01 2.06108212e+00
-7.44592965e-01 1.09453440e-01 3.79494607e-01 2.31439136e-02
9.89540160e-01 6.97832167e-01 8.24359059e-01 -1.09135389e+00
6.38638437e-01 5.75374424e-01 -9.17983055e-02 -4.17646378e-01
3.68516833e-01 -7.18779027e-01 2.28972346e-01 9.10511166e-02
-5.46125174e-01 -6.64085031e-01 -1.19016600e+00 -4.06011082e-02
9.37854469e-01 -1.14687141e-02 2.90750027e-01 -3.72969151e-01
2.41437718e-01 5.70643768e-02 2.64230281e-01 3.21067959e-01
-3.49805444e-01 7.31436312e-01 -3.80245417e-01 -3.61094773e-01
-5.07374704e-01 -8.70889723e-01 -8.45332503e-01 -2.57222682e-01
4.80003774e-01 -2.92942107e-01 -7.03303576e-01 -3.12349230e-01
-1.23493746e-01 5.05849779e-01 -5.23957133e-01 -4.98316003e-05
-6.63068950e-01 -3.99563491e-01 2.16738567e-01 4.73396242e-01
1.49103150e-01 -5.84541738e-01 -1.31848717e+00 2.06416205e-01
3.19527119e-01 -1.10141218e+00 -7.27995560e-02 6.63985461e-02
-1.45302534e+00 -1.04437852e+00 -8.49171638e-01 -5.08074582e-01
9.06792462e-01 1.23398840e-01 9.05576110e-01 1.33775985e-02
-4.01794255e-01 7.02410400e-01 -1.29066497e-01 -7.77188063e-01
-7.12340400e-02 -6.65180027e-01 4.14427400e-01 -8.78470898e-01
8.92774016e-02 -1.38883263e-01 -1.06976962e+00 4.49346155e-01
-2.53211826e-01 5.62082380e-02 2.21644729e-01 6.74501836e-01
8.38720024e-01 -4.20451820e-01 -6.10230565e-02 -8.28587294e-01
4.75932240e-01 -4.10790630e-02 -7.92037308e-01 -2.91436881e-01
-5.80140293e-01 -5.15541673e-01 3.27536836e-02 -1.69987530e-02
-7.83624589e-01 -1.53240234e-01 -2.15460032e-01 -6.58206761e-01
1.05786254e-03 3.28716755e-01 1.09178372e-01 -5.16654670e-01
7.03641117e-01 -1.15840040e-01 2.97943056e-01 -1.76969275e-01
-4.35727090e-01 5.23973525e-01 5.08742034e-01 -1.52249113e-01
5.32170594e-01 5.25345504e-01 1.25958905e-01 -6.58846736e-01
1.43210907e-02 -3.49807143e-01 -5.80481887e-01 -2.66958594e-01
8.25445592e-01 -1.07722855e+00 -8.39948118e-01 3.07895094e-01
-1.02253580e+00 -3.59266013e-01 -2.89937496e-01 1.34784997e+00
-3.98045659e-01 1.50309846e-01 -4.28264737e-01 -7.59149432e-01
-6.08528256e-01 -1.55151343e+00 1.20438516e+00 1.20560870e-01
-3.53071243e-01 -1.02127790e+00 3.12672779e-02 2.87708163e-01
4.41685230e-01 4.70204920e-01 3.92652035e-01 9.31212120e-03
-5.67449808e-01 -7.58967936e-01 2.73144156e-01 -3.51611704e-01
1.65544435e-01 -5.00158548e-01 -9.06880140e-01 -6.21724725e-01
8.60920012e-01 5.48190713e-01 -2.45763540e-01 1.02631986e+00
8.19921434e-01 8.46596584e-02 -6.90379918e-01 1.03570092e+00
1.48655605e+00 8.05894852e-01 7.78511465e-01 2.80124962e-01
6.05308235e-01 1.16643444e-01 1.10346556e+00 5.23653746e-01
2.41715670e-01 7.91712821e-01 5.49892664e-01 -1.27799854e-01
-2.39661634e-02 -9.47184802e-04 -7.88306743e-02 1.06403995e+00
-2.51392514e-01 1.32289037e-01 -1.16860926e+00 -4.19338159e-02
-8.64807427e-01 2.72170659e-02 -4.18121338e-01 2.79689336e+00
3.39484096e-01 8.15185532e-02 -3.71489227e-01 4.06166576e-02
3.18513036e-01 -5.62625468e-01 -3.59864086e-01 -3.56620073e-01
5.19020319e-01 8.21079254e-01 7.90870607e-01 6.93519711e-01
-5.08732498e-01 5.64348221e-01 5.63568354e+00 1.76794920e-02
-1.53478205e+00 4.88096744e-01 4.94739302e-02 2.86449287e-02
-3.42884868e-01 -1.76607385e-01 -8.12986493e-01 3.92787337e-01
6.62551701e-01 1.77865336e-03 -4.14417475e-01 5.92491686e-01
3.05750638e-01 -7.66758323e-01 -7.40453064e-01 1.50437438e+00
-9.14212093e-02 -1.08781481e+00 -5.78544676e-01 4.55399901e-01
2.89541036e-01 -3.00862826e-02 8.03578049e-02 -1.79591671e-01
-4.35362220e-01 -1.06911397e+00 3.64816338e-01 6.47617161e-01
1.42280519e+00 -3.11208457e-01 9.22070742e-01 5.02310812e-01
-1.03936827e+00 4.67606843e-01 -9.59877297e-02 1.25881329e-01
3.23859900e-01 4.03462976e-01 -1.47100019e+00 4.79487032e-01
6.35339320e-01 1.41099855e-01 -8.31304118e-02 1.10533249e+00
1.17758565e-01 -5.37181497e-02 -4.84183758e-01 2.68200248e-01
-4.22336638e-01 -2.42324442e-01 6.81696773e-01 5.28960288e-01
6.20724380e-01 8.55371237e-01 -5.26279151e-01 2.50336856e-01
4.17995125e-01 2.45596930e-01 -9.52025473e-01 7.42723703e-01
3.32006812e-01 8.96590352e-01 -6.38635576e-01 1.55465901e-01
-1.74795300e-01 7.95741677e-01 -5.37782133e-01 1.69881105e-01
-1.08753526e+00 2.77659744e-02 2.52705127e-01 7.93355286e-01
-5.10566473e-01 -5.35619676e-01 -6.81132674e-01 -9.28815484e-01
2.17158824e-01 -2.37233967e-01 7.41299288e-03 -1.07967675e+00
-2.51091540e-01 9.26623762e-01 -1.58476442e-01 -1.74008489e+00
4.60991710e-02 -5.29870868e-01 -1.89304799e-01 1.21517205e+00
-1.03157175e+00 -7.57355750e-01 -7.50766039e-01 3.91462684e-01
1.93754658e-01 1.24420010e-01 1.22474694e+00 9.48078260e-02
-3.30430061e-01 6.37547612e-01 -2.58701473e-01 -2.69205004e-01
5.17305374e-01 -7.54659474e-01 -1.87563468e-02 4.06334609e-01
-6.13527000e-01 1.03160942e+00 6.37606025e-01 -1.05616987e+00
-1.73435223e+00 -6.32492959e-01 3.49089324e-01 -6.57898426e-01
8.26254338e-02 -1.14978947e-01 -7.94263661e-01 7.37183630e-01
-1.18773602e-01 5.49553394e-01 6.78725719e-01 -4.97187048e-01
4.28571880e-01 2.18074456e-01 -1.86363423e+00 2.21004441e-01
1.08235788e+00 -4.52372909e-01 -3.01299542e-01 1.67434812e-01
4.82927412e-01 -1.79489696e+00 -1.24788880e+00 8.84275615e-01
7.80622303e-01 -8.90599191e-01 7.63806581e-01 9.91621539e-02
-1.86353475e-01 -3.31692368e-01 -1.05242804e-01 -1.04390645e+00
1.84111148e-01 -5.94060123e-01 9.58918706e-02 3.02954286e-01
1.89965397e-01 -1.24348462e+00 9.51005459e-01 1.16198826e+00
-7.25005984e-01 -9.28282499e-01 -1.27805734e+00 -3.05089623e-01
-9.54749286e-02 -5.05499005e-01 1.88675985e-01 7.44347095e-01
1.40241563e-01 -3.62448573e-01 4.30256218e-01 7.73885310e-01
6.13815188e-01 -4.57509160e-01 7.19809949e-01 -1.05094290e+00
-7.90468045e-03 2.14597806e-01 -7.70071089e-01 -8.13509047e-01
-3.20395708e-01 -6.98245108e-01 -1.82814613e-01 -1.58957505e+00
-2.52053976e-01 -1.13264847e+00 6.59696683e-02 -2.28315860e-01
2.30061859e-01 7.13307932e-02 -2.17737779e-01 2.90904760e-01
3.91803622e-01 2.20548421e-01 1.68737030e+00 6.20317280e-01
-4.23168987e-01 2.29507938e-01 -6.77700266e-02 8.28311741e-01
7.22610176e-01 -5.80058873e-01 -4.16178852e-01 -4.89562184e-01
1.07168257e-01 8.96719277e-01 3.67604613e-01 -1.14061141e+00
4.28611577e-01 4.95827436e-01 3.68267864e-01 -8.28445137e-01
6.25314593e-01 -1.39993942e+00 8.35625648e-01 8.56181502e-01
6.70518577e-01 4.13387328e-01 6.73373997e-01 2.85294682e-01
7.16638416e-02 3.75934429e-02 9.59091663e-01 -2.14156568e-01
-1.60171375e-01 4.13931668e-01 2.72767395e-02 -2.05569610e-01
1.29982376e+00 -9.93206680e-01 1.31418809e-01 6.26166835e-02
-1.13518083e+00 -5.10876417e-01 7.56227314e-01 -1.46492556e-01
1.14948130e+00 -9.45982993e-01 -1.76178232e-01 6.89441323e-01
-8.28078482e-03 5.35904646e-01 6.70279205e-01 1.57604694e+00
-1.12326884e+00 5.86594403e-01 2.01796293e-01 -1.11649096e+00
-1.42153990e+00 2.40582123e-01 9.59437251e-01 1.18425675e-01
-1.03403723e+00 1.08878767e+00 3.49420816e-01 -2.58441418e-01
8.06335732e-02 -7.55544960e-01 3.54348794e-02 -5.29432952e-01
1.46316901e-01 -4.11358252e-02 5.77149451e-01 -6.57770455e-01
-6.09785676e-01 1.27942693e+00 9.50539708e-02 -1.25205308e-01
1.05763519e+00 -3.38308990e-01 2.75062412e-01 1.30923972e-01
8.76952946e-01 4.20651376e-01 -8.57508719e-01 1.81073800e-01
-5.25586307e-01 -8.46171737e-01 2.87203938e-01 -8.90123010e-01
-1.13688827e+00 7.17224538e-01 1.22869825e+00 -5.34294784e-01
1.02788019e+00 -1.09087974e-02 6.41483307e-01 -6.15025401e-01
1.23863542e+00 -4.35058266e-01 -4.89939094e-01 1.25607893e-01
9.20275092e-01 -1.00195122e+00 8.18739086e-02 -9.39049006e-01
-5.94232500e-01 8.59667003e-01 5.65985858e-01 2.31080893e-02
1.01195300e+00 8.24790835e-01 3.60897005e-01 -7.54285991e-01
-2.78228689e-02 5.86572528e-01 1.34294638e-02 4.78463888e-01
7.84168065e-01 3.33649814e-01 -4.40871775e-01 4.00345653e-01
-4.67373699e-01 3.42430174e-01 4.49410439e-01 1.38635647e+00
-1.43665701e-01 -4.55082834e-01 -4.84438866e-01 2.68112838e-01
-5.44328511e-01 -4.66668010e-02 4.63523358e-01 1.01865530e+00
-1.30275950e-01 4.99899566e-01 -5.37784174e-02 -3.34247500e-01
8.66415322e-01 -3.29268187e-01 9.33561265e-01 -7.12744892e-01
-8.37928355e-01 3.10463905e-01 -3.77922207e-01 -6.40682340e-01
4.97489572e-02 -4.64643657e-01 -1.72709858e+00 7.92763978e-02
-5.82771063e-01 2.48222798e-01 1.38624775e+00 2.47576162e-01
4.08653706e-01 6.14439547e-01 5.74066520e-01 -5.03840625e-01
-1.03794290e-02 -6.85457289e-01 -8.16379607e-01 -1.30023420e-01
2.42920369e-01 -8.60078454e-01 -5.57310045e-01 -5.48569262e-01] | [13.75894832611084, -2.984117031097412] |
59d1e902-3f86-4f37-8c34-f742d4200107 | online-segmentation-of-lidar-sequences | 2206.08194 | null | https://arxiv.org/abs/2206.08194v2 | https://arxiv.org/pdf/2206.08194v2.pdf | Online Segmentation of LiDAR Sequences: Dataset and Algorithm | Roof-mounted spinning LiDAR sensors are widely used by autonomous vehicles. However, most semantic datasets and algorithms used for LiDAR sequence segmentation operate on $360^\circ$ frames, causing an acquisition latency incompatible with real-time applications. To address this issue, we first introduce HelixNet, a $10$ billion point dataset with fine-grained labels, timestamps, and sensor rotation information necessary to accurately assess the real-time readiness of segmentation algorithms. Second, we propose Helix4D, a compact and efficient spatio-temporal transformer architecture specifically designed for rotating LiDAR sequences. Helix4D operates on acquisition slices corresponding to a fraction of a full sensor rotation, significantly reducing the total latency. Helix4D reaches accuracy on par with the best segmentation algorithms on HelixNet and SemanticKITTI with a reduction of over $5\times$ in terms of latency and $50\times$ in model size. The code and data are available at: https://romainloiseau.fr/helixnet | ['Loïc Landrieu', 'Mathieu Aubry', 'Romain Loiseau'] | 2022-06-16 | null | null | null | null | ['lidar-semantic-segmentation'] | ['computer-vision'] | [-5.52405370e-03 -3.77137184e-01 -4.13333327e-02 -5.22800922e-01
-7.19201446e-01 -8.00034225e-01 4.44691062e-01 2.86094189e-01
-7.34686673e-01 2.74027765e-01 -7.11307347e-01 -5.64838767e-01
2.00815976e-01 -1.04711890e+00 -7.55770922e-01 -1.65822104e-01
-1.19818173e-01 8.51896405e-01 9.23505902e-01 1.49860233e-01
1.65394604e-01 9.90146339e-01 -1.78683722e+00 -1.31991863e-01
4.94782448e-01 1.41864979e+00 2.49874651e-01 6.85604036e-01
-3.12513024e-01 3.06948304e-01 -3.45030487e-01 -1.01541564e-01
5.83653569e-01 2.80247211e-01 -6.66821301e-01 -1.56357214e-01
5.38325787e-01 -4.57909554e-01 -3.24104488e-01 6.20196283e-01
2.39688292e-01 -1.25380278e-01 3.02211702e-01 -1.45458126e+00
4.18593109e-01 2.91868091e-01 -6.01384223e-01 4.00604844e-01
2.43817657e-01 3.77073646e-01 7.54746616e-01 -9.53097999e-01
8.70033622e-01 1.03392339e+00 8.24452996e-01 5.92910610e-02
-1.07086205e+00 -8.55627835e-01 -1.89471349e-01 4.32319015e-01
-1.59381902e+00 -6.03001058e-01 3.89497787e-01 -5.25572181e-01
1.03005993e+00 2.50552416e-01 7.42823899e-01 5.47622740e-01
4.28052098e-02 3.75375092e-01 8.16390753e-01 2.65873581e-01
3.67628485e-01 -4.11244154e-01 2.28764862e-01 6.33242488e-01
2.07426980e-01 9.38562378e-02 -6.71416640e-01 -1.07553534e-01
6.97617233e-01 -1.15429498e-01 3.89104694e-01 -5.55702090e-01
-1.39204252e+00 4.50839072e-01 3.50352466e-01 -2.30962440e-01
-3.00952554e-01 5.97214520e-01 6.55156553e-01 1.05459780e-01
2.53302693e-01 3.59425992e-02 -4.45017785e-01 -4.96033937e-01
-1.12192535e+00 2.34615237e-01 2.06898794e-01 1.33294320e+00
1.13246191e+00 1.23763874e-01 4.56234068e-01 3.41130674e-01
2.34127253e-01 1.01961184e+00 -1.00216620e-01 -1.45373762e+00
4.73338872e-01 6.61701798e-01 -4.79254499e-02 -4.92890060e-01
-7.16545463e-01 5.10658324e-02 -5.19983232e-01 2.99491018e-01
3.87594163e-01 3.64827067e-02 -9.50595379e-01 1.13125741e+00
4.76350427e-01 4.02827263e-01 -2.98465759e-01 7.35912919e-01
4.55314010e-01 4.89590108e-01 7.03113675e-02 -1.40151661e-02
1.59736764e+00 -2.94497967e-01 -1.14585139e-01 -2.28402168e-01
5.63601255e-01 -8.89840186e-01 1.09900999e+00 2.33822927e-01
-9.33399022e-01 -4.53332484e-01 -1.06378853e+00 -2.27174148e-01
-3.25874478e-01 -1.53557748e-01 4.45139438e-01 5.56574285e-01
-9.90505099e-01 3.65428001e-01 -1.19173670e+00 -3.87813956e-01
4.87007469e-01 3.62752765e-01 -1.62852466e-01 -1.21325278e-03
-6.90270722e-01 6.43983662e-01 1.74410522e-01 -1.04853943e-01
-7.20568419e-01 -8.47012103e-01 -8.03457320e-01 -3.02169204e-01
4.43755239e-01 -3.58115524e-01 1.46155226e+00 1.33156136e-01
-1.06165576e+00 8.92918587e-01 -3.34041953e-01 -1.00082576e+00
6.73844278e-01 -3.28814209e-01 -3.01862270e-01 3.21416497e-01
3.65184575e-01 9.66386616e-01 6.34572387e-01 -9.37648714e-01
-8.64483297e-01 -5.97787976e-01 -2.40259513e-01 -4.04402092e-02
3.43498886e-01 -1.39675751e-01 -7.43552446e-01 -1.44318983e-01
3.40457588e-01 -1.26359642e+00 -4.23580706e-01 2.65434235e-01
-9.68534425e-02 -9.69360173e-02 1.42306173e+00 -2.01594204e-01
6.81872845e-01 -2.14638543e+00 -5.15283644e-01 3.10593367e-01
1.10762924e-01 2.42042720e-01 5.93786463e-02 1.97159857e-01
3.81556392e-01 -5.09888865e-02 -4.92370754e-01 -3.14064622e-01
-1.29006505e-01 4.12785590e-01 -4.24723357e-01 6.66920424e-01
-1.58693969e-01 6.77221060e-01 -7.18854964e-01 -6.35071933e-01
8.18551660e-01 3.23433518e-01 -2.35201731e-01 -2.41073668e-01
-3.01339447e-01 2.95481265e-01 -4.06797409e-01 8.80804360e-01
9.28309441e-01 -1.43522099e-01 1.89408571e-01 -1.68440551e-01
-6.14376247e-01 3.77496183e-01 -1.24284220e+00 1.86744714e+00
-2.27626115e-01 9.16069686e-01 1.75452188e-01 -5.20885944e-01
9.24255073e-01 -2.00005360e-02 1.16005552e+00 -9.78908420e-01
1.18201151e-01 3.43810588e-01 -5.75542033e-01 -5.62452935e-02
9.52712178e-01 3.00742745e-01 -3.26074004e-01 4.21491295e-01
-4.76461381e-01 -6.57850742e-01 4.58681762e-01 1.81425005e-01
1.34306371e+00 1.75750673e-01 -1.05808347e-01 -2.38091812e-01
1.96522072e-01 5.55075884e-01 6.48383617e-01 4.84007686e-01
-2.46498927e-01 3.99129063e-01 1.81479156e-01 -3.97196263e-01
-1.21904218e+00 -1.25869942e+00 -3.38247836e-01 7.47839212e-01
6.69374347e-01 -5.40311694e-01 -6.91841543e-01 -3.17521095e-01
2.55346805e-01 6.48551226e-01 -6.57548457e-02 4.21883851e-01
-9.63671565e-01 -1.56819567e-01 7.61519551e-01 7.03873098e-01
6.96088254e-01 -8.43023539e-01 -1.65751278e+00 3.04161221e-01
1.35493269e-02 -1.55189121e+00 -1.10893138e-01 9.90665033e-02
-1.07925355e+00 -1.17888761e+00 2.97484603e-02 -1.11438558e-01
1.98458001e-01 5.37699699e-01 1.14672625e+00 -2.18271732e-01
-4.67910916e-01 2.76046932e-01 -2.74820179e-01 -5.36639512e-01
5.23159616e-02 3.08286637e-01 6.85267746e-02 -7.02618122e-01
4.89683390e-01 -5.42881489e-01 -6.94436848e-01 5.41734934e-01
-7.78531432e-01 6.96200356e-02 2.09648058e-01 -1.01482093e-01
8.93979549e-01 -1.79516792e-01 -5.27172312e-02 -4.86550897e-01
-7.77472258e-02 -3.63429159e-01 -1.28145754e+00 -4.31566268e-01
-6.57440424e-01 -1.39664084e-01 3.38382095e-01 1.94287643e-01
-4.36169297e-01 4.92635876e-01 -1.99416816e-01 -6.52146041e-01
-2.00810149e-01 3.15522100e-03 2.48449638e-01 1.32739335e-01
4.09028143e-01 5.49737662e-02 1.32392064e-01 -4.33175892e-01
4.80690777e-01 3.47636759e-01 7.42104530e-01 -5.01327395e-01
8.43838215e-01 1.02165926e+00 1.87281877e-01 -9.92023468e-01
-1.59217447e-01 -6.93982780e-01 -7.50695705e-01 -4.31204975e-01
8.05177093e-01 -8.51698577e-01 -1.03696144e+00 4.83919293e-01
-1.02294898e+00 -5.10782480e-01 -6.17068172e-01 3.65184218e-01
-7.44888842e-01 2.72364140e-01 -3.88047606e-01 -5.49372375e-01
-2.41182014e-01 -1.35906482e+00 1.31681418e+00 -3.42036225e-02
-1.75117120e-01 -1.95290521e-01 -8.26492831e-02 3.75356883e-01
1.10961750e-01 2.89852798e-01 4.73611534e-01 1.86007319e-03
-1.15950871e+00 4.40421030e-02 -5.73373437e-01 -1.84143454e-01
-3.54301572e-01 2.96232074e-01 -7.53164291e-01 -2.09757879e-01
-3.15718502e-01 6.83021471e-02 5.06378949e-01 3.30329925e-01
1.09685135e+00 3.13711703e-01 -4.87770855e-01 5.04725039e-01
1.51229095e+00 3.38491529e-01 5.78356564e-01 3.09243619e-01
4.80910599e-01 2.41102636e-01 1.12115252e+00 5.75636506e-01
6.76206708e-01 8.48887682e-01 7.01925159e-01 4.11722586e-02
-9.75144804e-02 -8.37488770e-02 1.91012934e-01 7.21914291e-01
7.27398768e-02 1.52369710e-02 -1.46657956e+00 7.92224407e-01
-1.59890616e+00 -8.07582259e-01 -5.64621210e-01 2.16812754e+00
2.75597483e-01 3.64798188e-01 3.33667964e-01 1.80574313e-01
4.92466778e-01 3.13161850e-01 -8.14247072e-01 -4.56836909e-01
6.77106977e-02 4.18291926e-01 1.30518377e+00 4.21727628e-01
-8.70714903e-01 1.09142268e+00 6.18890524e+00 7.71959722e-01
-1.33119905e+00 1.39684051e-01 9.16194841e-02 -1.95191488e-01
-1.72553933e-03 1.84544578e-01 -8.37754548e-01 6.19022965e-01
1.35633755e+00 -1.02493398e-01 1.54309690e-01 9.51147139e-01
4.35868949e-01 -5.07630825e-01 -8.32938194e-01 1.09720647e+00
-4.95006651e-01 -1.59990573e+00 -3.69988710e-01 4.42490965e-01
7.44073018e-02 8.91555786e-01 -2.36762404e-01 -8.58091563e-02
4.15558010e-01 -7.25254834e-01 1.06041086e+00 1.90383673e-01
1.24069941e+00 -8.36318314e-01 3.56062323e-01 2.89040327e-01
-1.83042324e+00 2.54901707e-01 -2.72539288e-01 -1.42658725e-02
6.43190980e-01 6.14921093e-01 -1.07970881e+00 5.08238196e-01
1.06841922e+00 5.71894407e-01 -6.17407143e-01 7.64157712e-01
2.42760003e-01 5.41554570e-01 -9.97719824e-01 3.51210773e-01
2.48139024e-01 -3.50823194e-01 5.44370174e-01 1.13395274e+00
6.08803153e-01 2.81356931e-01 1.14277534e-01 2.84548432e-01
-3.51344310e-02 -2.33304843e-01 -7.23275304e-01 3.02597076e-01
1.01125836e+00 9.64301288e-01 -1.09140134e+00 -4.21633840e-01
-2.69709110e-01 6.02127016e-01 -7.35671446e-02 -1.20268844e-01
-1.18959129e+00 -3.40988457e-01 9.87055242e-01 5.39290130e-01
3.44980836e-01 -9.88352418e-01 -7.92774916e-01 -7.54416168e-01
9.22761932e-02 -3.72410774e-01 2.16813639e-01 -6.52051330e-01
-4.63343501e-01 3.69778901e-01 6.06219694e-02 -1.47970700e+00
-2.03786716e-01 -4.54598874e-01 -1.49346991e-02 2.68533915e-01
-1.33267808e+00 -9.38351512e-01 -6.01041377e-01 7.08180666e-01
6.58071220e-01 2.34373420e-01 4.62848783e-01 5.52716136e-01
-3.61937284e-01 1.07417747e-01 1.14071406e-02 -3.36182751e-02
3.03308427e-01 -9.85525072e-01 1.01252282e+00 8.92236531e-01
1.08583972e-01 2.11602256e-01 7.64329255e-01 -6.48363888e-01
-1.67312789e+00 -1.10948527e+00 5.64935505e-01 -4.72334713e-01
5.59281111e-01 -3.23989838e-01 -6.94149017e-01 7.29026139e-01
-2.82388013e-02 1.59855813e-01 2.49899417e-01 -3.10566276e-01
-2.82246262e-01 -4.22161341e-01 -1.30960917e+00 3.62263888e-01
1.36128318e+00 -3.90092313e-01 -2.09384233e-01 1.39209360e-01
6.31162226e-01 -6.53182924e-01 -9.08342957e-01 6.74392998e-01
5.66544116e-01 -1.15642750e+00 1.07052791e+00 2.43867978e-01
-1.78042337e-01 -7.51019120e-01 -3.90619934e-01 -4.16415393e-01
1.64025024e-01 -4.03245240e-01 -4.96799983e-02 8.34356785e-01
1.74428657e-01 -7.10446119e-01 1.11261177e+00 3.68817806e-01
-2.45189384e-01 -3.60536426e-01 -1.46670663e+00 -8.28802347e-01
-2.93706000e-01 -1.23063850e+00 8.77784669e-01 6.93324327e-01
-5.16413987e-01 6.83836192e-02 2.00971022e-01 2.29883298e-01
9.67288435e-01 2.84312785e-01 1.10116982e+00 -1.27600658e+00
5.52660286e-01 -2.61599481e-01 -7.25196302e-01 -1.17778182e+00
-1.62633345e-01 -5.63776910e-01 8.57586712e-02 -1.36678636e+00
-4.55904007e-01 -9.64436471e-01 2.07410261e-01 3.63926262e-01
6.61548734e-01 6.98375583e-01 3.24402720e-01 4.33610082e-01
-5.27244985e-01 1.88614503e-01 5.09301543e-01 5.06583527e-02
4.17335555e-02 -8.72582421e-02 1.57311901e-01 8.79282653e-01
8.97368371e-01 -4.36036646e-01 -4.12649512e-01 -7.90094912e-01
1.23956859e-01 1.69578150e-01 5.61778903e-01 -1.55966115e+00
3.23443681e-01 -4.20589238e-01 6.93196654e-02 -1.59529757e+00
8.74501586e-01 -1.04764080e+00 6.30295038e-01 5.27206481e-01
3.62552166e-01 5.84048986e-01 4.07980651e-01 3.29026818e-01
3.53595018e-02 2.40055099e-01 9.46004033e-01 1.20489895e-02
-1.09623194e+00 5.19402623e-01 -5.29847980e-01 -1.64051265e-01
1.38754463e+00 -6.26536071e-01 -3.00233930e-01 1.93319485e-01
-4.80133086e-01 4.10789520e-01 9.62734640e-01 3.21950376e-01
5.60103655e-01 -9.99809980e-01 -4.00674164e-01 1.64225250e-01
6.43926412e-02 4.42488074e-01 3.60964119e-01 6.03000641e-01
-1.27817476e+00 6.42881572e-01 -1.84329927e-01 -1.23907340e+00
-1.34549797e+00 2.82798946e-01 7.62781277e-02 1.44590214e-01
-8.91380668e-01 6.29543304e-01 -4.39760387e-01 -4.54058588e-01
-2.54864305e-01 -5.52181959e-01 4.53455269e-01 2.72992253e-01
5.77171892e-02 9.53428447e-01 3.02635670e-01 -7.60059834e-01
-8.57209980e-01 9.08029675e-01 2.93442726e-01 -5.26807487e-01
1.04821610e+00 -3.12364042e-01 2.03999691e-02 3.11150432e-01
8.65036964e-01 -2.34496146e-02 -1.39913952e+00 -1.83066398e-01
2.26617172e-01 -4.69109684e-01 -1.25353873e-01 -1.32574648e-01
-9.94252264e-01 8.63249004e-01 7.68035769e-01 2.07305580e-01
8.23037028e-01 2.14783605e-02 1.28765750e+00 1.42279580e-01
1.07879496e+00 -1.21278369e+00 -5.31993747e-01 6.63280904e-01
3.01583499e-01 -9.15062308e-01 2.13709056e-01 -4.58147585e-01
-3.29713881e-01 7.70936549e-01 5.15188575e-01 -2.95185596e-02
4.94872153e-01 6.33129239e-01 3.01272959e-01 -4.40143853e-01
-4.46621865e-01 -2.00321749e-01 -4.39025015e-01 5.81823051e-01
-2.03870416e-01 3.08900088e-01 -1.26353756e-01 -3.92886430e-01
-6.07760668e-01 9.11819637e-02 3.31225127e-01 1.19853926e+00
-6.49387360e-01 -9.79283035e-01 -4.59056228e-01 4.76198852e-01
8.31962898e-02 2.45253995e-01 -3.65844578e-03 7.82247782e-01
2.28071317e-01 8.37771356e-01 6.74285948e-01 -3.78629059e-01
4.95709896e-01 -8.86036828e-02 1.80384934e-01 -4.77838278e-01
-2.72018164e-01 -1.90926701e-01 9.88768786e-02 -1.11248720e+00
-2.41194278e-01 -9.76093888e-01 -1.70775700e+00 -8.41596186e-01
1.44147173e-01 -1.80067848e-02 1.12523472e+00 5.53230762e-01
7.46762037e-01 1.17730871e-01 4.47207779e-01 -1.04414570e+00
2.04369947e-02 -4.90720391e-01 -3.34159523e-01 -1.20331571e-01
1.70088679e-01 -5.73122740e-01 1.32580608e-01 1.25066742e-01] | [8.072729110717773, -2.61867094039917] |
8e77eb08-5220-4016-8dcf-06691dce6cac | fitannotator-a-flexible-and-intelligent-text | null | null | https://aclanthology.org/2021.naacl-demos.5 | https://aclanthology.org/2021.naacl-demos.5.pdf | FITAnnotator: A Flexible and Intelligent Text Annotation System | In this paper, we introduce FITAnnotator, a generic web-based tool for efficient text annotation. Benefiting from the fully modular architecture design, FITAnnotator provides a systematic solution for the annotation of a variety of natural language processing tasks, including classification, sequence tagging and semantic role annotation, regardless of the language. Three kinds of interfaces are developed to annotate instances, evaluate annotation quality and manage the annotation task for annotators, reviewers and managers, respectively. FITAnnotator also gives intelligent annotations by introducing task-specific assistant to support and guide the annotators based on active learning and incremental learning strategies. This assistant is able to effectively update from the annotator feedbacks and easily handle the incremental labeling scenarios. | ['Tingwen Liu', 'Li Quangang', 'Bowen Yu', 'Yanzeng Li'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['text-annotation'] | ['natural-language-processing'] | [-5.48880026e-02 6.87629104e-01 -2.50377417e-01 -5.58500111e-01
-4.50970739e-01 -1.10203063e+00 2.69019216e-01 8.68001759e-01
-5.72476268e-01 8.94610226e-01 9.87866223e-02 4.75260578e-02
-3.17749918e-01 -2.79679865e-01 8.54671970e-02 -2.56914228e-01
3.25081915e-01 1.10346210e+00 5.67690253e-01 -1.14510268e-01
2.07765698e-01 3.76615167e-01 -1.80349422e+00 5.09306967e-01
9.14365590e-01 7.24970639e-01 6.78779066e-01 6.87197983e-01
-8.78344655e-01 9.94775534e-01 -6.43501282e-01 -4.44458663e-01
-1.13110133e-01 1.39393747e-01 -1.42877388e+00 -2.96977758e-02
1.32832885e-01 1.69529662e-01 5.57724595e-01 8.45998943e-01
6.10731065e-01 8.71452242e-02 2.95657367e-01 -1.23993909e+00
-1.98791325e-01 1.15855598e+00 4.22695577e-01 -7.72567019e-02
7.52968669e-01 1.11962706e-02 9.59424198e-01 -9.57606971e-01
1.09131587e+00 1.07545161e+00 7.50374258e-01 7.30827510e-01
-5.85799813e-01 -1.82896078e-01 8.07998329e-02 4.68794167e-01
-1.05593050e+00 -5.29652894e-01 3.21516722e-01 -6.98114574e-01
9.86870944e-01 4.50515747e-01 3.65977734e-01 7.20489323e-01
-6.72879636e-01 7.13075638e-01 6.72533989e-01 -9.62613881e-01
4.07785237e-01 6.74816787e-01 7.61352777e-01 7.73669600e-01
8.60432684e-02 -8.98975492e-01 -6.15763664e-01 -5.91889322e-01
1.88831374e-01 -4.41458791e-01 -2.40311578e-01 -4.76822942e-01
-9.23104346e-01 1.85913607e-01 -2.85761923e-01 5.49372613e-01
-3.93970191e-01 -1.81074530e-01 1.05229807e+00 9.65679437e-02
4.52565581e-01 9.26070333e-01 -1.13229644e+00 -4.17141438e-01
-2.58002758e-01 2.66945790e-02 1.18419278e+00 1.29739273e+00
8.18532825e-01 -4.30930614e-01 -3.86861622e-01 1.13345790e+00
4.27387297e-01 -3.62936035e-02 6.11517549e-01 -9.91365075e-01
2.38585860e-01 1.38124049e+00 7.49154449e-01 -4.07278955e-01
-6.39134884e-01 -1.32409250e-02 -2.05643699e-01 7.94282705e-02
5.45476019e-01 -4.20935974e-02 -4.45242196e-01 1.22140419e+00
6.65359199e-01 -4.65199351e-01 -1.63145009e-02 6.60130739e-01
1.02657390e+00 1.65855899e-01 6.79649949e-01 -6.92566335e-01
1.77503312e+00 -1.05111337e+00 -1.40284765e+00 1.26731936e-02
1.19450510e+00 -8.07714403e-01 1.31251776e+00 4.15459931e-01
-9.51939583e-01 -5.79554439e-01 -5.26623070e-01 -2.11415961e-01
-9.84016716e-01 5.42791784e-01 4.32074249e-01 4.52018648e-01
-9.53747213e-01 4.13342267e-01 -7.33658731e-01 -7.67213166e-01
3.79429124e-02 1.79409534e-01 -4.79324549e-01 3.52361262e-01
-1.32276273e+00 1.11200297e+00 1.02144432e+00 -9.68670025e-02
-4.67700124e-01 -5.48154950e-01 -6.54583156e-01 1.81932971e-01
8.02403033e-01 -5.18847704e-01 1.81093967e+00 -1.00782704e+00
-1.46482050e+00 1.13082004e+00 -4.01013233e-02 -2.11709738e-01
5.12256026e-01 -2.85457611e-01 -3.22768778e-01 1.56983912e-01
9.64662880e-02 2.06158608e-01 2.29136929e-01 -8.73818755e-01
-6.35093987e-01 -3.25540274e-01 1.10401437e-01 4.65257376e-01
-7.67260551e-01 5.47795236e-01 -4.62299854e-01 -3.22628796e-01
-5.30124366e-01 -4.61257011e-01 -2.51167536e-01 2.61601150e-01
-1.08041307e-02 -7.65778542e-01 1.01182306e+00 -6.05991542e-01
1.53422856e+00 -1.85221672e+00 -1.50222152e-01 -1.77793175e-01
3.23868573e-01 7.00569689e-01 1.19670793e-01 6.40042067e-01
1.62826423e-02 1.44573852e-01 -2.05261767e-01 -3.28839540e-01
1.60027623e-01 3.63329023e-01 2.26227194e-01 -3.82282078e-01
-1.81346878e-01 6.49355292e-01 -1.34405720e+00 -8.84171665e-01
1.04130007e-01 -6.81514964e-02 5.78366928e-02 6.34582937e-01
-7.40720391e-01 3.20707589e-01 -3.57593179e-01 6.44392133e-01
3.64542782e-01 -3.00288171e-01 7.21250594e-01 -3.93222123e-02
-3.34373713e-01 1.85219422e-01 -1.07478082e+00 1.88418639e+00
-5.07419944e-01 3.09581995e-01 4.38536882e-01 -8.65288615e-01
1.09343624e+00 1.00286353e+00 3.32418352e-01 -1.38732702e-01
-1.92501750e-02 4.80383784e-01 -4.89311516e-01 -1.20149648e+00
6.87280059e-01 6.30826294e-01 -1.46948949e-01 9.13346887e-01
4.52763110e-01 4.07547981e-01 8.41495156e-01 2.64351040e-01
1.07946002e+00 3.68639857e-01 1.17268336e+00 -2.55175352e-01
9.49452460e-01 3.23369354e-01 4.20165569e-01 7.34044373e-01
-1.94417819e-01 -3.68668169e-01 1.89083979e-01 -8.75439405e-01
-1.08483911e+00 -3.35928649e-01 4.50797155e-02 1.75178874e+00
-3.22305471e-01 -6.74537361e-01 -9.24722314e-01 -1.27146506e+00
-3.11401099e-01 6.77029669e-01 -2.25967452e-01 1.04467675e-01
-3.08273375e-01 -3.61228257e-01 4.85871285e-01 3.25232804e-01
3.58414829e-01 -1.37241793e+00 -8.96922767e-01 4.00156587e-01
-3.51243794e-01 -1.01914501e+00 -2.42563009e-01 4.07238275e-01
-4.74579483e-01 -1.31211627e+00 -5.93282543e-02 -8.17719758e-01
5.58325171e-01 -3.18113983e-01 1.32557952e+00 3.77618879e-01
-3.79570991e-01 6.01004541e-01 -8.78212690e-01 -6.08031869e-01
-9.40626740e-01 5.21071792e-01 -2.01155812e-01 -2.76204437e-01
4.36161876e-01 -1.90231249e-01 5.10679148e-02 7.41589725e-01
-8.20711792e-01 9.91842002e-02 1.51143475e-02 5.75585783e-01
4.06913966e-01 -1.62763298e-01 6.48995578e-01 -1.37715340e+00
8.02796066e-01 -3.69738162e-01 -6.31691635e-01 9.02119994e-01
-8.53258014e-01 7.98908412e-04 7.96847761e-01 -3.04020286e-01
-1.34362733e+00 6.53056264e-01 -1.46041706e-01 7.68242031e-02
-5.16227305e-01 5.21296680e-01 -5.58234870e-01 2.37609163e-01
1.05746865e+00 -1.87132701e-01 -2.78480768e-01 -9.31913555e-01
4.28420246e-01 1.19312263e+00 5.09566247e-01 -4.78739053e-01
3.89542282e-01 -1.08870067e-01 -5.82379401e-01 -2.22502738e-01
-1.17417943e+00 -8.60307455e-01 -1.15844774e+00 -5.94998121e-01
6.20338798e-01 -5.67062199e-01 -9.08082068e-01 2.66251028e-01
-1.43095148e+00 -5.29240608e-01 -5.49723446e-01 -1.72110144e-02
-4.69803005e-01 4.04794425e-01 -2.93968409e-01 -1.16993690e+00
-7.31576562e-01 -7.28118479e-01 8.41446459e-01 2.37103134e-01
-6.64105773e-01 -1.16453898e+00 3.62558030e-02 4.80641186e-01
3.69942129e-01 -5.83480000e-02 8.21884274e-01 -1.43729746e+00
-2.29357854e-01 -3.55581969e-01 1.01984732e-01 7.72591159e-02
6.17088377e-02 -1.88317876e-02 -9.89557445e-01 2.03993767e-01
-4.76560891e-01 -4.50808853e-01 -2.79093310e-02 -4.25855070e-01
1.09609795e+00 -6.50082290e-01 -4.46789622e-01 -8.99710581e-02
9.53780174e-01 3.03762019e-01 3.09064299e-01 5.00836730e-01
4.39841568e-01 7.85623074e-01 1.02080703e+00 5.01959026e-01
1.67025074e-01 9.97007132e-01 4.04931396e-01 3.07554573e-01
1.22758970e-01 -1.66945308e-01 -1.20431580e-01 5.69392145e-01
-1.50922472e-02 -5.59943020e-01 -1.24504387e+00 2.31214345e-01
-2.53307819e+00 -8.43615651e-01 -4.15379107e-01 1.88379502e+00
1.17796469e+00 -1.37104481e-01 1.03260666e-01 3.34737360e-01
1.03873944e+00 -4.38591421e-01 -3.51966232e-01 -5.34684479e-01
1.82760835e-01 -1.56678095e-01 1.88194066e-01 4.07898307e-01
-9.76983130e-01 1.12030840e+00 6.97392321e+00 7.78940558e-01
-5.12082398e-01 5.30569792e-01 5.59618920e-02 4.85388070e-01
2.57859260e-01 3.67683768e-02 -9.58829403e-01 4.66638297e-01
8.01105022e-01 -3.95744354e-01 1.89363569e-01 1.41774464e+00
3.47151011e-01 4.95059276e-03 -9.77680504e-01 4.45891827e-01
-4.62441817e-02 -1.37297261e+00 -9.08284262e-03 -4.44242418e-01
1.62484437e-01 -3.95917684e-01 -8.87189507e-01 2.32655063e-01
4.59904224e-01 -3.64191622e-01 8.49539816e-01 9.24746156e-01
6.82680309e-01 -3.29867840e-01 1.08621204e+00 7.10809529e-01
-1.06090760e+00 -3.18051726e-01 -4.21174280e-02 -2.25875229e-01
3.39344919e-01 5.81310332e-01 -1.48410416e+00 4.88670826e-01
6.96080267e-01 3.22624564e-01 -7.40241766e-01 1.12810302e+00
-4.18708652e-01 3.78655255e-01 -5.42904548e-02 -3.32690716e-01
-4.07569885e-01 2.94189364e-01 5.24840891e-01 1.52522016e+00
-1.41907170e-01 -1.39830308e-02 4.64567959e-01 3.21876317e-01
7.79252034e-03 7.35726833e-01 -1.20650254e-01 -2.16967762e-01
1.02588081e+00 1.85510504e+00 -8.78415465e-01 -8.76336813e-01
1.91066995e-01 8.75535786e-01 5.30018687e-01 3.81490099e-03
-3.01033169e-01 -7.24022329e-01 1.54551655e-01 3.21121871e-01
-9.12372097e-02 -6.63827732e-02 -1.04861714e-01 -8.19135189e-01
1.74867101e-02 -8.20053220e-01 8.35583627e-01 -1.34794259e+00
-9.94445384e-01 7.85404325e-01 1.95194289e-01 -9.93604481e-01
-4.89929706e-01 -6.31856561e-01 -2.44432867e-01 5.05573571e-01
-8.36550355e-01 -1.07779121e+00 -7.26733148e-01 3.13351333e-01
5.58573008e-01 -2.31005207e-01 1.34246778e+00 4.23225313e-01
-4.73208249e-01 3.86882514e-01 -3.49893123e-01 1.61978498e-01
8.34764004e-01 -1.55835080e+00 9.77701843e-02 5.13108909e-01
-7.84967914e-02 6.23148859e-01 6.11860335e-01 -6.75843775e-01
-8.48135233e-01 -9.61195052e-01 1.38884938e+00 -6.59167111e-01
7.08369255e-01 -6.59851134e-01 -1.06889331e+00 4.94476527e-01
1.95035502e-01 7.44889528e-02 8.96625102e-01 7.51290843e-02
-3.06169033e-01 8.61483812e-02 -1.28237975e+00 1.21946119e-01
1.38266242e+00 -4.28788692e-01 -6.25633419e-01 7.34412432e-01
8.11947644e-01 -4.35029328e-01 -9.34915066e-01 1.57028511e-01
3.47541988e-01 -4.44749683e-01 3.70564550e-01 -6.81976676e-01
-1.69748962e-01 -5.17858744e-01 4.65028554e-01 -8.73644829e-01
-2.94814289e-01 -8.26429188e-01 -2.66857803e-01 1.63473570e+00
7.42568791e-01 -4.05919105e-01 4.19828951e-01 7.97139108e-01
-4.79847610e-01 -1.84815273e-01 -6.23159945e-01 -4.69502985e-01
-1.06426406e+00 -4.37752336e-01 7.47339547e-01 1.19742048e+00
6.89696550e-01 5.31506896e-01 -1.94121882e-01 -3.42719220e-02
9.65598598e-02 -5.07518768e-01 6.97032630e-01 -1.82253993e+00
-1.75707147e-01 -1.63477719e-01 -3.12619328e-01 -4.45869714e-01
2.14788467e-01 -8.65575552e-01 2.21020371e-01 -1.73047352e+00
-1.90159678e-02 -4.82333720e-01 7.55596906e-02 1.26786196e+00
-2.04765588e-01 6.42392933e-02 -1.11211091e-01 4.66772854e-01
-1.27204418e+00 -2.02985138e-01 7.23392665e-01 1.75956473e-01
-1.33795023e-01 -4.03134245e-03 -4.92171586e-01 8.51457000e-01
6.08808339e-01 -6.90643132e-01 -3.47134560e-01 -3.03762525e-01
6.93315327e-01 -2.45563298e-01 -6.43667802e-02 -6.70692265e-01
4.64757055e-01 -1.53796673e-01 -7.61825591e-02 -2.35362545e-01
-2.06151381e-01 -1.03655958e+00 4.08237994e-01 2.53762394e-01
-8.90157342e-01 6.51959330e-02 7.88873062e-03 2.54247874e-01
-5.96839935e-02 -1.14011109e+00 4.72538620e-01 -5.55239320e-01
-8.95698965e-01 6.45254850e-02 -8.60165238e-01 -1.03535287e-01
1.00212216e+00 -1.92847788e-01 -4.12071645e-01 1.15001328e-01
-1.23945642e+00 3.36180508e-01 4.93196636e-01 2.83746749e-01
-1.99128360e-01 -1.03715074e+00 -2.91656226e-01 -2.18915015e-01
7.29546010e-01 2.72791442e-02 -5.54174557e-02 4.32815343e-01
-6.69413924e-01 2.77190506e-01 -1.62337542e-01 -3.12201828e-01
-1.57151067e+00 5.37017882e-01 2.16949567e-01 -7.08441436e-01
-2.37525016e-01 4.59291607e-01 -5.38978398e-01 -8.17744076e-01
5.08436978e-01 1.09198697e-01 -8.51277649e-01 4.60011601e-01
8.56256843e-01 3.82089525e-01 5.27586758e-01 -3.01321954e-01
-3.15199405e-01 2.28212513e-02 -1.54297352e-02 6.92040175e-02
1.12696493e+00 -3.28272939e-01 -3.43920141e-01 7.28365541e-01
3.68240952e-01 -1.32480726e-01 -6.89541698e-01 -3.22752774e-01
9.07446980e-01 -2.10356951e-01 -4.59241718e-01 -1.50647378e+00
-4.21856791e-01 4.31316763e-01 4.86343682e-01 6.67187274e-01
8.09196591e-01 1.59063563e-01 4.93994467e-02 7.97662139e-01
3.60245436e-01 -1.48682547e+00 -4.59895171e-02 6.32608473e-01
8.37668419e-01 -8.25314403e-01 -7.16553181e-02 -5.84939003e-01
-6.43335640e-01 1.35109115e+00 8.38885546e-01 8.30669224e-01
3.17042708e-01 4.88462389e-01 4.31555659e-01 -3.29318196e-01
-1.18265533e+00 -2.06095621e-01 7.79161230e-02 9.00738597e-01
7.83239543e-01 -2.12320194e-01 -6.82600260e-01 7.84788728e-01
1.95651308e-01 2.13071257e-01 3.38503748e-01 1.00145912e+00
-7.79810905e-01 -1.44689918e+00 -4.27673072e-01 1.06415808e-01
-8.33131820e-02 1.92656562e-01 -9.26593065e-01 3.76212507e-01
5.43537080e-01 7.67876565e-01 6.63167238e-02 1.70068398e-01
4.98298734e-01 7.62748718e-01 4.36846763e-02 -1.12630653e+00
-1.12370217e+00 -2.74710596e-01 7.93910444e-01 -1.70914680e-01
-5.82230270e-01 -4.83869195e-01 -1.32682812e+00 4.59926814e-01
-6.77063346e-01 8.32007408e-01 8.09533000e-01 1.01973271e+00
4.08342570e-01 1.23406731e-01 2.20606625e-01 -5.58749318e-01
-2.59085536e-01 -1.36451721e+00 -1.78833812e-01 3.99413168e-01
-4.71476108e-01 -5.74868500e-01 -1.06443368e-01 5.28140724e-01] | [9.285938262939453, 8.758474349975586] |
aebc2895-d06f-49c7-aaa4-bf928de89113 | deep-reinforcement-learning-for-automatic-run | 2210.15498 | null | https://arxiv.org/abs/2210.15498v1 | https://arxiv.org/pdf/2210.15498v1.pdf | Deep reinforcement learning for automatic run-time adaptation of UWB PHY radio settings | Ultra-wideband technology has become increasingly popular for indoor localization and location-based services. This has led recent advances to be focused on reducing the ranging errors, whilst research focusing on enabling more reliable and energy efficient communication has been largely unexplored. The IEEE 802.15.4 UWB physical layer allows for several settings to be selected that influence the energy consumption, range, and reliability. Combined with the available link state diagnostics reported by UWB devices, there is an opportunity to dynamically select PHY settings based on the environment. To address this, we propose a deep Q-learning approach for enabling reliable UWB communication, maximizing packet reception rate (PRR) and minimizing energy consumption. Deep Q-learning is a good fit for this problem, as it is an inherently adaptive algorithm that responds to the environment. Validation in a realistic office environment showed that the algorithm outperforms traditional Q-learning, linear search and using a fixed PHY layer. We found that deep Q-learning achieves a higher average PRR and reduces the ranging error while using only 14% of the energy compared to a fixed PHY setting in a dynamic office environment. | ['Eli de Poorter', 'Adnan Shahid', 'Dieter Coppens'] | 2022-10-13 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [ 9.63606387e-02 -1.75586477e-01 -4.07437027e-01 -4.34330404e-01
-9.76502001e-01 -2.52271891e-01 -4.41292338e-02 1.21336371e-01
-6.80055976e-01 1.25036144e+00 1.87176801e-02 -4.69678879e-01
-7.74796307e-01 -9.49104905e-01 -6.66928664e-02 -8.09967399e-01
-3.92454952e-01 5.18218726e-02 -5.10186367e-02 4.77755405e-02
3.19815874e-01 4.17947203e-01 -1.06446707e+00 -7.35658288e-01
7.95462787e-01 1.28049517e+00 4.51919407e-01 9.71284986e-01
5.83046451e-02 2.69715548e-01 -1.14429796e+00 2.97818959e-01
-6.51742741e-02 -6.18568301e-01 -3.31936270e-01 -5.78221679e-01
-8.37880373e-02 -2.61409372e-01 -3.12307864e-01 4.25403893e-01
1.40923309e+00 4.50485945e-01 1.44322619e-01 -1.06152558e+00
-1.83869451e-01 5.61047375e-01 -2.68013000e-01 4.50766295e-01
6.22563660e-01 -2.77444482e-01 5.51530600e-01 1.82016436e-02
7.49534145e-02 9.17869270e-01 7.14224458e-01 5.29363275e-01
-1.13771737e+00 -9.62294102e-01 -3.24411035e-01 2.99593121e-01
-1.58212125e+00 -6.13847554e-01 6.74749494e-01 4.57432866e-01
8.13798726e-01 4.61642832e-01 9.60693479e-01 1.05875182e+00
9.44411874e-01 9.36570242e-02 8.53282988e-01 -8.27591062e-01
1.03729069e+00 -1.24652900e-01 -1.99838132e-01 4.94893849e-01
2.46601596e-01 2.65400678e-01 -5.02103984e-01 -3.03246289e-01
4.49541569e-01 -2.83353955e-01 7.86800217e-03 -1.71955302e-01
-5.24148226e-01 6.14283860e-01 4.93597984e-01 4.73844439e-01
-2.89829999e-01 9.52632010e-01 -1.28861144e-01 2.44181439e-01
5.18582463e-02 8.61546695e-01 -4.41670567e-01 -7.68082142e-01
-1.02504599e+00 1.66290984e-01 9.13185418e-01 7.08989024e-01
4.70231235e-01 1.58066288e-01 -3.46928895e-01 1.03018081e+00
7.41941154e-01 6.74407303e-01 1.43687353e-01 -1.54370236e+00
-2.22471673e-02 -2.75156438e-01 2.79474944e-01 -5.29332340e-01
-7.59787381e-01 -8.39420199e-01 -3.81007552e-01 1.20042145e-01
1.83892250e-01 -7.16255128e-01 -8.36331427e-01 1.89179516e+00
2.58803368e-01 -4.08440679e-02 -7.41744041e-03 4.27130759e-01
2.39800915e-01 7.68673360e-01 2.00208932e-01 -4.40038741e-01
1.09158015e+00 -2.65052944e-01 -1.10601377e+00 -4.04646844e-01
1.46975815e-01 -7.07731485e-01 4.92568463e-01 4.41417783e-01
-1.05475628e+00 -4.15843248e-01 -1.43224704e+00 4.25083578e-01
-2.14031190e-01 -5.25073290e-01 6.47711098e-01 1.71495485e+00
-1.20305741e+00 3.94679457e-01 -8.26482952e-01 -6.48018718e-01
3.60391855e-01 7.07784295e-01 5.87915540e-01 -3.61121118e-01
-1.32870591e+00 8.53230834e-01 1.92226365e-01 -1.19309261e-01
-2.92456895e-01 -5.40282905e-01 -8.11638892e-01 1.80019543e-01
3.06763560e-01 -6.57483339e-01 1.44037962e+00 -9.31356996e-02
-1.92889869e+00 -3.00836056e-01 -5.47979884e-02 -5.19068062e-01
2.70294756e-01 1.01855777e-01 -6.73670173e-01 9.13973246e-03
-6.23004027e-02 4.85562801e-01 5.74007154e-01 -1.17961633e+00
-7.43424475e-01 -2.35307157e-01 -7.95090199e-02 9.60058719e-02
-2.58049637e-01 -3.07017028e-01 -1.67879045e-01 -1.78279176e-01
1.58991829e-01 -8.90990615e-01 -5.67851961e-01 1.87968969e-01
8.27440992e-02 -1.47769466e-01 7.76572049e-01 -1.09492742e-01
1.21053243e+00 -1.75087368e+00 -6.41005993e-01 6.00752950e-01
-3.01839918e-01 -1.81900278e-01 9.80819538e-02 4.53553259e-01
8.02693307e-01 1.31016463e-01 8.85610729e-02 -2.80647010e-01
-1.40487447e-01 4.80271131e-01 2.09870756e-01 3.93073827e-01
-5.29189944e-01 3.49599183e-01 -1.15999389e+00 -6.65903807e-01
4.69594151e-01 6.14828944e-01 -4.41223681e-01 1.15804493e-01
2.35533550e-01 5.26986599e-01 -7.27618277e-01 6.64192080e-01
6.90949321e-01 7.32174963e-02 3.40862334e-01 1.50235236e-01
-3.00508231e-01 1.74508184e-01 -1.14280236e+00 1.81576800e+00
-1.20582008e+00 6.71151042e-01 2.45607182e-01 -7.36870527e-01
9.61170852e-01 1.26297191e-01 8.85749996e-01 -1.49224186e+00
3.04335922e-01 3.56919080e-01 4.40865681e-02 -4.48361933e-01
5.33605039e-01 -1.85210422e-01 -2.90174007e-01 3.93531293e-01
-1.92064568e-02 -3.23321074e-01 -1.21047914e-01 -1.49339169e-01
1.61663735e+00 1.76420878e-03 2.97292113e-01 -2.11352289e-01
1.17574856e-01 -4.52345192e-01 5.03871024e-01 1.30194151e+00
-5.06233215e-01 -3.55499834e-02 -3.91140968e-01 1.28128618e-01
-4.83305305e-01 -1.22171962e+00 -5.26243865e-01 1.13975036e+00
7.17148781e-01 -1.17912889e-01 -3.99891734e-01 5.38706668e-02
1.62299648e-02 1.16489732e+00 -1.95842043e-01 -4.36184883e-01
-3.46768022e-01 -6.39892161e-01 6.05273545e-01 1.07688926e-01
7.03661859e-01 -6.70948088e-01 -1.17809069e+00 7.57443428e-01
-1.13189816e-01 -1.10958302e+00 -3.13079692e-02 8.10540974e-01
-6.53202057e-01 -3.98274779e-01 -3.94973814e-01 -2.27246508e-01
1.58767626e-01 1.75048709e-01 7.61751175e-01 -7.77818561e-02
-5.13914108e-01 8.51298571e-01 -4.28891480e-01 -5.99960387e-01
-1.08860917e-01 2.80170381e-01 6.35659769e-02 -7.11627245e-01
-5.98939173e-02 -5.83536029e-01 -9.83746886e-01 2.20351800e-01
-3.75494361e-01 -7.59046257e-01 6.02462828e-01 4.65967566e-01
3.40776324e-01 4.77710426e-01 1.01791406e+00 -1.93691641e-01
7.78443336e-01 -5.78043044e-01 -3.43374133e-01 1.21826068e-01
-9.20898139e-01 2.31921762e-01 4.38149899e-01 -1.49393603e-01
-1.16457260e+00 -9.02666152e-02 -6.92026794e-01 7.07983017e-01
1.54402956e-01 1.04420885e-01 -6.28205314e-02 -6.43707573e-01
6.20434821e-01 -4.89001460e-02 -4.27494109e-01 -8.24315473e-02
9.84813869e-02 8.83578479e-01 2.94606745e-01 -6.96512818e-01
5.03925383e-01 2.47358754e-01 3.48809123e-01 -1.23838842e+00
-6.98637605e-01 -2.12245032e-01 5.86452149e-02 -3.65064204e-01
5.16861081e-01 -5.25311828e-01 -7.61399090e-01 -3.18520993e-01
-7.28036284e-01 -5.21722972e-01 -1.81707814e-01 7.09693074e-01
-5.36987960e-01 1.82474524e-01 -7.77212530e-02 -1.38239980e+00
-3.71186197e-01 -7.57904887e-01 6.80016875e-01 6.36919796e-01
-4.90514874e-01 -8.83215666e-01 2.56354153e-01 1.18135966e-01
9.44818854e-01 2.25325033e-01 6.08654737e-01 1.19679779e-01
-3.59894931e-01 -2.52802372e-01 1.76963702e-01 -3.07752639e-01
3.22506368e-01 -3.92572403e-01 -9.68170166e-01 -4.06094670e-01
-1.86054364e-01 -1.05210833e-01 5.02286911e-01 9.40716803e-01
1.41865408e+00 3.24264318e-02 -6.28099382e-01 6.50950134e-01
1.48638070e+00 7.19287395e-01 6.55936062e-01 4.34789866e-01
-1.45810008e-01 -4.04080115e-02 7.74695277e-01 7.73780167e-01
2.58836359e-01 7.77008533e-01 6.26528084e-01 1.63874943e-02
2.05458328e-02 -1.21663250e-02 -4.73814011e-02 5.96964248e-02
1.04814626e-01 -7.69377410e-01 -5.97740054e-01 7.37126097e-02
-1.44908988e+00 -8.89192700e-01 4.72957253e-01 2.24913335e+00
6.33964419e-01 4.06541765e-01 -1.78887457e-01 3.39415550e-01
2.78285742e-01 3.08369964e-01 -6.16956174e-01 -7.05934942e-01
5.76835394e-01 6.62264824e-01 1.03702438e+00 5.33290923e-01
-6.55886650e-01 3.14706475e-01 6.82137346e+00 6.70604467e-01
-9.98249769e-01 1.48420393e-01 1.79930031e-01 -9.10339803e-02
-3.00191760e-01 -3.52920979e-01 -4.93488342e-01 6.93523407e-01
1.38712668e+00 -9.96220484e-02 4.53256100e-01 7.02243984e-01
6.10142827e-01 -7.88068414e-01 -7.96786666e-01 1.07829976e+00
-3.84697646e-01 -1.07733357e+00 -7.36036301e-01 3.69883507e-01
4.35900688e-01 -1.98475569e-01 -2.04157773e-02 4.07124609e-01
2.83471495e-01 -8.78713727e-01 4.47985470e-01 4.81164038e-01
6.13734663e-01 -1.11459160e+00 7.37221181e-01 3.47165406e-01
-1.16091514e+00 -6.46435678e-01 -4.50048536e-01 -1.09417722e-01
3.64504755e-01 7.32518613e-01 -8.09226692e-01 4.13215250e-01
8.71567369e-01 -3.00275356e-01 -1.14222594e-01 1.66468382e+00
-1.07514761e-01 6.99499071e-01 -5.26769757e-01 -5.45452595e-01
1.38337478e-01 2.82591611e-01 2.31071264e-01 1.29225004e+00
8.88363123e-01 1.71248108e-01 1.34189546e-01 4.67585415e-01
-8.12263861e-02 -3.77256781e-01 -2.00152859e-01 5.14846861e-01
1.29826534e+00 1.06904662e+00 -6.63471460e-01 2.24228486e-01
8.76622573e-02 7.52701998e-01 -4.58433002e-01 4.20989335e-01
-7.54482388e-01 -7.83198357e-01 5.68060100e-01 -6.84514493e-02
8.49938095e-02 -7.72341728e-01 -2.99173236e-01 -6.58164173e-02
-5.35508633e-01 -4.43697944e-02 2.41205022e-01 -4.28453773e-01
-6.21881723e-01 3.48408408e-02 -2.20271945e-02 -7.76838779e-01
-1.89811572e-01 -1.29856290e-02 -4.97572333e-01 6.39193296e-01
-1.77205181e+00 -4.94145811e-01 -7.54070342e-01 1.36660010e-01
5.44794798e-01 3.54542077e-01 1.04754770e+00 4.47844774e-01
-2.53281742e-01 6.83831155e-01 4.30630058e-01 -5.57694733e-01
5.80547810e-01 -1.22918403e+00 -1.36445016e-01 3.89843553e-01
-3.10770363e-01 4.70221221e-01 1.05115676e+00 -2.96976894e-01
-1.75448859e+00 -8.54213893e-01 3.93101096e-01 -5.90461167e-03
1.26262426e-01 -2.45105132e-01 -2.96672553e-01 -1.46361262e-01
1.94263831e-01 -1.39139732e-02 1.06085384e+00 4.46237475e-02
5.30211031e-01 -6.13502681e-01 -1.66740763e+00 4.30383712e-01
1.03047597e+00 1.48581574e-02 4.20741662e-02 -2.30527250e-03
4.46455985e-01 -2.40259141e-01 -8.91703188e-01 4.39455695e-02
1.06676722e+00 -6.04361892e-01 1.12659454e+00 4.64336783e-01
-5.79855740e-01 -1.08552687e-01 -2.38824546e-01 -1.37743890e+00
-4.14493918e-01 -8.12230408e-01 -4.44879740e-01 1.04403698e+00
2.20113739e-01 -4.73544955e-01 8.14560294e-01 6.22768581e-01
1.86424702e-02 -8.08919609e-01 -1.50355887e+00 -1.01822007e+00
-4.94648337e-01 -6.46044612e-01 5.44349372e-01 1.51510224e-01
-2.73452014e-01 1.90853104e-01 -3.27769667e-01 2.95350820e-01
1.04565096e+00 -4.23004419e-01 3.86799157e-01 -1.18910170e+00
-1.56276822e-01 -1.58700466e-01 -2.04568118e-01 -1.26516163e+00
-2.28504643e-01 -2.59096891e-01 6.20731831e-01 -2.16408539e+00
-4.42851812e-01 -1.10098636e+00 -6.61756873e-01 2.40343824e-01
2.10321590e-01 2.95605659e-01 -3.09076160e-01 -3.74242455e-01
-7.79294550e-01 6.97074592e-01 8.55673552e-01 -6.60893694e-02
-6.28265440e-01 4.67795223e-01 -6.00608110e-01 2.18141705e-01
1.08291280e+00 -6.05181217e-01 -5.84147453e-01 -1.72169104e-01
1.46028042e-01 3.80819291e-01 -9.64984968e-02 -1.59376395e+00
3.97107035e-01 -3.49911332e-01 8.34101677e-01 -3.36149544e-01
5.87167799e-01 -1.12805212e+00 -2.02205312e-02 8.80730212e-01
-2.24232972e-01 -5.03638089e-01 1.77566901e-01 8.95685852e-01
5.44126987e-01 -3.66654664e-01 9.28067744e-01 1.64387107e-01
-8.67905021e-01 1.62801426e-02 -8.18370342e-01 -3.55651706e-01
1.02419055e+00 -6.64789557e-01 1.27200400e-02 -7.99203753e-01
-4.09554690e-01 4.20849264e-01 1.12790037e-02 2.07492933e-01
6.14114165e-01 -1.17453730e+00 -1.26666166e-02 -1.67216122e-01
-1.79757118e-01 -4.49900061e-01 1.94982305e-01 3.69907469e-01
-2.73635268e-01 3.78806382e-01 -2.89840162e-01 -4.62115198e-01
-1.12735057e+00 -2.41703749e-01 6.77682102e-01 -3.47214639e-02
-1.78603038e-01 8.55540574e-01 -1.10355794e+00 -1.25554368e-01
5.07469118e-01 -1.20939776e-01 -2.28649452e-02 -2.05705941e-01
3.23359936e-01 7.73595929e-01 -1.33137340e-02 1.28831565e-01
-7.69752383e-01 5.12115002e-01 5.62471628e-01 -3.49376619e-01
1.18337774e+00 -5.61543345e-01 3.80074918e-01 1.52747601e-01
8.99827242e-01 2.01785773e-01 -1.25765216e+00 1.58772767e-01
1.00695565e-01 -5.63178897e-01 6.67100787e-01 -9.47374225e-01
-4.38503325e-01 4.18858349e-01 1.44326675e+00 5.56609631e-01
1.08994627e+00 -1.61533982e-01 1.03778899e+00 6.68242991e-01
1.07944024e+00 -1.52732790e+00 3.36209655e-01 2.43545651e-01
1.82761982e-01 -9.92583871e-01 4.70749617e-01 1.17058218e-01
1.99867353e-01 1.22844720e+00 3.03159356e-01 2.34282449e-01
6.95328236e-01 4.84448433e-01 -5.86749688e-02 1.87341601e-01
-1.96903899e-01 -1.94274083e-01 -3.33176345e-01 8.61911118e-01
5.12682259e-01 4.50627208e-02 -4.59444225e-01 1.86148994e-02
-3.72676790e-01 8.95700976e-02 2.28184998e-01 1.38163364e+00
-1.07709754e+00 -1.36585784e+00 -5.90187907e-01 5.31238794e-01
-5.09421051e-01 5.09315729e-01 2.76619494e-01 6.23087585e-01
3.37315977e-01 1.67090821e+00 -7.95247182e-02 -2.31030762e-01
2.30472207e-01 -3.92686486e-01 8.88781488e-01 -5.10151684e-01
-4.79941592e-02 -1.34514049e-01 2.98516661e-01 -5.41621208e-01
-3.52360338e-01 -3.64842921e-01 -1.52100837e+00 -1.54149964e-01
-4.69785839e-01 5.18853247e-01 1.37458539e+00 9.98101711e-01
2.61128098e-01 1.16377151e+00 8.17198455e-01 -8.23849738e-01
-4.69234645e-01 -5.65060675e-01 -6.62210226e-01 -8.99830282e-01
6.51888371e-01 -7.73048282e-01 -1.18491352e-01 -8.31498742e-01] | [6.244296550750732, 1.0978450775146484] |
c28c0ed7-5a0f-41e9-907b-319517e64540 | multilingual-word-embeddings-using | 1612.04732 | null | http://arxiv.org/abs/1612.04732v1 | http://arxiv.org/pdf/1612.04732v1.pdf | Multilingual Word Embeddings using Multigraphs | We present a family of neural-network--inspired models for computing
continuous word representations, specifically designed to exploit both
monolingual and multilingual text. This framework allows us to perform
unsupervised training of embeddings that exhibit higher accuracy on syntactic
and semantic compositionality, as well as multilingual semantic similarity,
compared to previous models trained in an unsupervised fashion. We also show
that such multilingual embeddings, optimized for semantic similarity, can
improve the performance of statistical machine translation with respect to how
it handles words not present in the parallel data. | ['Radu Soricut', 'Nan Ding'] | 2016-12-14 | null | null | null | null | ['multilingual-word-embeddings'] | ['methodology'] | [-2.05710188e-01 -7.98759982e-02 -6.04173601e-01 -4.22083765e-01
-6.79757953e-01 -6.66599810e-01 9.14004147e-01 5.04074574e-01
-7.86365092e-01 6.10748053e-01 6.68293774e-01 -6.54564559e-01
1.35333806e-01 -6.94250286e-01 -5.14259219e-01 -3.12552333e-01
7.79059995e-03 8.26956153e-01 -3.74155164e-01 -5.52836835e-01
-4.61075716e-02 3.54242772e-01 -1.19237006e+00 -6.41194433e-02
1.02294910e+00 3.85520071e-01 3.58140379e-01 4.46722448e-01
-4.23285604e-01 3.47383946e-01 -2.78709739e-01 -5.41872084e-01
-2.47495770e-02 -2.27665976e-01 -7.84682095e-01 -5.23643017e-01
5.24974644e-01 1.93370566e-01 -3.95972043e-01 1.22788191e+00
5.03413200e-01 7.92641714e-02 8.78675640e-01 -6.25706196e-01
-1.57168412e+00 9.14877355e-01 8.73626992e-02 4.44421202e-01
2.91953832e-01 -3.30597788e-01 1.50117183e+00 -1.28748155e+00
9.85263884e-01 1.25470340e+00 7.07622051e-01 3.98708463e-01
-1.27676368e+00 -2.58266509e-01 -2.23250873e-02 3.31858516e-01
-1.35084951e+00 -4.95535821e-01 6.09732091e-01 -2.24061087e-01
1.44809341e+00 -6.79632872e-02 3.77703100e-01 1.24356616e+00
3.81847620e-01 5.78201950e-01 9.45609391e-01 -9.47917163e-01
-7.72704929e-02 2.88110435e-01 2.43813232e-01 6.13323629e-01
2.84676194e-01 9.92214903e-02 -2.91091084e-01 -1.56764343e-01
5.65531015e-01 -1.19271688e-01 -2.57359128e-02 -4.00833189e-01
-1.53541350e+00 1.27900624e+00 2.23724917e-01 9.62863266e-01
-2.65737623e-01 3.27461101e-02 6.26038551e-01 5.55402875e-01
8.64380002e-01 8.60222638e-01 -7.97548234e-01 -5.85635006e-03
-7.38168478e-01 -1.02641962e-01 6.80453420e-01 9.12407994e-01
9.78670359e-01 2.91859090e-01 1.22506633e-01 1.26553571e+00
3.57206874e-02 7.89435983e-01 1.23460948e+00 -4.79647189e-01
2.58792639e-01 3.52410346e-01 -4.45956469e-01 -9.39557433e-01
-3.88943970e-01 -3.71482998e-01 -5.55592358e-01 -5.20653546e-01
-4.92436253e-02 -4.15687682e-03 -6.34366453e-01 1.99381423e+00
3.19487252e-03 -2.49721095e-01 6.42268658e-01 6.18628383e-01
6.42099977e-01 6.07800066e-01 1.64062172e-01 -3.90256047e-02
1.33274269e+00 -1.11841989e+00 -9.58749354e-01 -2.14235544e-01
1.01841879e+00 -1.13876438e+00 1.07405043e+00 -2.38830850e-01
-1.00722563e+00 -5.86875200e-01 -8.89172733e-01 -3.95521641e-01
-9.02380228e-01 -4.36378382e-02 6.81904793e-01 4.79409903e-01
-1.24406230e+00 5.72783649e-01 -5.71209431e-01 -8.52955759e-01
-3.65971252e-02 1.27163559e-01 -5.73253572e-01 -1.64303616e-01
-1.62294245e+00 1.52081645e+00 6.05602920e-01 -2.92483687e-01
-6.02593839e-01 -5.55519164e-01 -1.30008399e+00 -1.49443492e-01
-2.68147588e-01 -5.33103287e-01 1.01804960e+00 -1.13478184e+00
-1.35185027e+00 9.58840013e-01 -4.11458671e-01 -4.89966273e-01
-2.61024952e-01 -1.72720045e-01 -7.46795714e-01 1.04783066e-01
2.43048474e-01 7.38248765e-01 5.37934244e-01 -8.27664077e-01
-3.99617344e-01 -3.59548688e-01 -7.22483471e-02 3.58604133e-01
-9.28794086e-01 3.08529139e-01 -1.40965670e-01 -9.94480610e-01
-9.18068588e-02 -7.71555841e-01 -3.61024588e-01 -3.66731524e-01
-2.95772925e-02 -4.54771429e-01 4.28236991e-01 -9.15171981e-01
8.87104094e-01 -1.95581973e+00 6.61080003e-01 -1.94635186e-02
-1.92942247e-01 2.77647823e-01 -5.83413184e-01 7.42245257e-01
-1.65519744e-01 2.09555477e-01 -1.87442660e-01 -3.88545305e-01
9.69219804e-02 7.34176159e-01 -2.44794950e-01 4.15703326e-01
5.48950553e-01 1.37026095e+00 -1.03250849e+00 -4.80010122e-01
3.07241753e-02 4.38730001e-01 -4.63698596e-01 1.71662733e-01
-6.12405082e-03 1.94975019e-01 -7.36510679e-02 6.00552738e-01
3.76224130e-01 2.04333439e-01 7.60109961e-01 7.37200491e-03
-1.54390678e-01 6.35326505e-01 -4.00525659e-01 2.11037517e+00
-1.06216657e+00 7.42509663e-01 -2.76177615e-01 -1.22658455e+00
9.43266273e-01 5.17581403e-01 1.00345567e-01 -9.21272933e-01
2.46630013e-01 6.17946684e-01 -7.86929950e-03 -3.57847691e-01
9.81159687e-01 -4.31733817e-01 -2.75622129e-01 6.12215698e-01
6.98972642e-01 7.74172768e-02 1.22196622e-01 8.08414295e-02
6.21850848e-01 2.40863822e-02 4.43660051e-01 -7.38808751e-01
4.65446979e-01 1.48327261e-01 2.92480439e-01 2.44342834e-01
6.10503480e-02 2.91641474e-01 6.23512194e-02 -5.60614407e-01
-1.40218675e+00 -1.19693708e+00 -3.53731781e-01 1.50779498e+00
-2.96234637e-02 -4.02659774e-01 -5.75989068e-01 -5.43570459e-01
1.27103347e-02 8.72238040e-01 -6.19732559e-01 -2.79554963e-01
-7.14431882e-01 -7.18729019e-01 6.20500982e-01 6.26954138e-01
-4.40363437e-01 -1.07502794e+00 2.57491529e-01 2.90251732e-01
-1.17743522e-01 -1.26299560e+00 -5.33418655e-01 5.49597025e-01
-1.04573786e+00 -8.34951162e-01 -5.29261887e-01 -1.34644043e+00
5.64690113e-01 1.14706323e-01 1.31275785e+00 -1.22852147e-01
-1.76977545e-01 3.56525451e-01 -5.38179874e-01 -4.99118194e-02
-7.22492754e-01 3.87037843e-01 5.06944060e-01 -2.79517591e-01
8.69652152e-01 -6.98954046e-01 4.80388775e-02 -9.11173895e-02
-1.08956528e+00 -5.52505851e-01 4.87962037e-01 1.03796780e+00
3.45709264e-01 -5.92804551e-01 6.60080075e-01 -7.52226472e-01
9.84217703e-01 -7.23295689e-01 -1.29340887e-01 3.35192114e-01
-7.46742606e-01 5.54916263e-01 9.25168931e-01 -7.27714479e-01
-5.14202535e-01 -5.16066790e-01 -2.14735836e-01 -2.97653735e-01
-3.87695245e-02 6.50034845e-01 1.86029240e-01 3.77851911e-02
6.96059346e-01 4.55779910e-01 4.84466553e-02 -7.03821182e-01
1.02399850e+00 7.49170840e-01 4.47857976e-01 -8.42630684e-01
7.88327813e-01 2.81718802e-02 -4.29700971e-01 -8.07044983e-01
-5.08398831e-01 -4.80028570e-01 -1.05915332e+00 4.88332629e-01
1.11019623e+00 -9.27886367e-01 2.01193318e-01 -4.36121225e-02
-1.51444876e+00 2.27156058e-01 -3.59433264e-01 8.51684809e-01
-4.26227182e-01 3.91073197e-01 -6.99782550e-01 -2.15106428e-01
-3.39341432e-01 -1.06400526e+00 8.62143636e-01 -3.44731122e-01
-5.30866027e-01 -1.94590425e+00 5.75046241e-01 -9.55243856e-02
6.23974323e-01 -2.98854589e-01 1.56723511e+00 -1.26697171e+00
6.57363608e-02 -4.79411148e-02 -1.30641654e-01 7.13260353e-01
1.98898077e-01 -2.63796270e-01 -7.26558924e-01 -2.57675022e-01
-2.40756094e-01 -4.14978236e-01 6.98396504e-01 5.21986149e-02
3.39242131e-01 -3.18521708e-01 1.74131691e-02 5.60921252e-01
1.51237702e+00 -2.94946998e-01 1.42942578e-01 5.13332248e-01
7.42426813e-01 5.45669556e-01 -2.76269838e-02 -1.76232770e-01
5.19959152e-01 5.72797060e-01 1.13185622e-01 -1.71980739e-01
-2.18373060e-01 -4.51896727e-01 5.82207203e-01 1.96216881e+00
2.23350093e-01 2.06935912e-01 -8.01425099e-01 1.03338778e+00
-1.73338616e+00 -4.99947160e-01 -4.41548526e-02 1.93909824e+00
1.00160849e+00 -3.98647249e-01 -2.03642547e-01 -2.75869101e-01
7.23113418e-01 3.74638110e-01 -2.02283654e-02 -1.25696766e+00
-5.11151552e-01 9.32185233e-01 5.83298683e-01 6.50200963e-01
-7.66301811e-01 1.51150799e+00 8.00032711e+00 8.16214144e-01
-8.94993126e-01 8.03780079e-01 2.55909413e-02 3.31515282e-01
-7.96797156e-01 1.91938430e-02 -5.42275131e-01 7.10807890e-02
1.41599870e+00 -1.74250945e-01 5.92891753e-01 7.57938623e-01
-2.19129890e-01 5.25900841e-01 -1.33534992e+00 6.14369452e-01
5.82159281e-01 -1.31693769e+00 4.80963588e-01 -1.41097337e-01
9.89299476e-01 3.21242779e-01 3.44174653e-02 4.13182437e-01
5.75480580e-01 -1.05548918e+00 5.25519013e-01 6.37533218e-02
7.19532490e-01 -8.82855773e-01 8.48034084e-01 1.38180524e-01
-9.49146152e-01 1.72511443e-01 -7.24858820e-01 -9.94498730e-02
5.47269322e-02 3.35817754e-01 -6.66986406e-01 8.00014734e-01
3.09331864e-01 9.23891306e-01 -4.85166162e-01 2.18282774e-01
-5.44274688e-01 2.15214208e-01 2.44667511e-02 -1.55273229e-01
6.03916407e-01 -3.20323825e-01 4.47594464e-01 1.50522137e+00
3.08888137e-01 -7.90852010e-01 2.58718252e-01 6.14104152e-01
-2.98787117e-01 7.77551532e-01 -9.36402380e-01 -3.96257877e-01
4.30396080e-01 9.50541675e-01 -1.38681307e-01 -5.04267454e-01
-4.68564749e-01 1.21831346e+00 9.06343758e-01 4.36215103e-01
-4.77366179e-01 -3.00004691e-01 1.05304742e+00 -2.43603244e-01
2.34212130e-01 -8.12772393e-01 -4.73622158e-02 -1.44808960e+00
-9.41993967e-02 -9.53963935e-01 1.23469561e-01 -3.99381727e-01
-1.58644462e+00 9.87514079e-01 -2.24523902e-01 -8.32520247e-01
-5.23286223e-01 -1.00373232e+00 -4.60633844e-01 1.00840759e+00
-1.86281121e+00 -1.28998911e+00 6.99649870e-01 4.48710233e-01
5.23203433e-01 -7.00609565e-01 1.50790048e+00 3.02183241e-01
-3.28893393e-01 6.63552105e-01 6.20211720e-01 2.02880532e-01
9.09086823e-01 -1.19509399e+00 6.27580106e-01 6.35960996e-01
8.42071950e-01 9.12821949e-01 3.91278267e-01 -2.77840316e-01
-1.57119286e+00 -1.09250021e+00 1.81977069e+00 -5.64757705e-01
1.30868268e+00 -4.33870763e-01 -7.46241391e-01 9.29805815e-01
7.37088919e-01 -4.64312099e-02 1.03800297e+00 6.59805536e-01
-9.02320743e-01 2.22687662e-01 -6.76464200e-01 7.55924761e-01
1.00871503e+00 -1.24864244e+00 -1.31407726e+00 6.00621760e-01
1.09036493e+00 1.99647844e-01 -1.05366611e+00 1.61479488e-01
3.82401615e-01 -4.34277445e-01 9.54924524e-01 -1.21686935e+00
5.54507613e-01 1.97065637e-01 -6.47464931e-01 -1.71690893e+00
-4.44339573e-01 -2.67080814e-01 2.17247829e-01 8.48026156e-01
7.87286341e-01 -8.55135918e-01 -3.56846265e-02 -3.35473180e-01
-1.94263071e-01 -4.55055594e-01 -1.09033561e+00 -1.19844770e+00
8.25591743e-01 -4.28892642e-01 4.72214043e-01 1.42857826e+00
2.18827710e-01 4.95097786e-01 -1.64599702e-01 6.49922639e-02
2.91824937e-01 2.21108552e-02 2.66547680e-01 -1.11864626e+00
2.78396420e-02 -6.07958972e-01 -6.42574131e-01 -7.89532185e-01
1.00343585e+00 -1.73788655e+00 -1.12536989e-01 -1.44110596e+00
1.01527490e-01 -3.08241487e-01 -6.71691298e-01 3.65020186e-01
-1.65920988e-01 6.05009139e-01 -7.39566088e-02 3.05993408e-01
-4.13728446e-01 7.30894208e-01 7.53058195e-01 -1.84437007e-01
1.69572398e-01 -9.54683661e-01 -6.41756117e-01 5.25853157e-01
7.02450097e-01 -5.97335458e-01 1.35227479e-02 -1.00959146e+00
1.79375127e-01 -4.71096724e-01 -1.09675676e-01 -6.26902044e-01
-2.25192308e-03 -6.30741240e-03 9.67731029e-02 -2.59668492e-02
1.69241711e-01 -6.58793867e-01 -2.77015805e-01 3.09765995e-01
-5.31129360e-01 6.82229996e-01 1.38947785e-01 3.62984031e-01
-5.98747194e-01 -4.45673287e-01 4.91726428e-01 -6.12251014e-02
-6.64867103e-01 6.44341260e-02 -5.07007539e-01 1.19599976e-01
6.45144522e-01 2.41791263e-01 -1.25056729e-01 -3.10590509e-02
-7.16531932e-01 -5.41036986e-02 5.82352042e-01 9.77038741e-01
3.07807028e-01 -1.97546458e+00 -8.65749061e-01 3.51969689e-01
5.58919013e-01 -9.00122225e-01 -3.08271974e-01 7.57585347e-01
-3.22290808e-01 6.80682480e-01 -3.07042122e-01 -3.24429601e-01
-8.12254429e-01 7.85453856e-01 -1.95904002e-02 -3.61521930e-01
-3.79679948e-01 5.42761981e-01 -2.03080192e-01 -1.34236848e+00
-9.09301341e-02 -2.52339303e-01 -1.85405284e-01 2.25495741e-01
2.35739753e-01 -5.71677275e-02 1.44837096e-01 -1.06430113e+00
-4.38937187e-01 7.66947567e-01 -1.20686315e-01 -1.88391253e-01
1.27117348e+00 -9.09760222e-02 -5.21005034e-01 7.01019406e-01
1.63576174e+00 1.66859657e-01 -1.60650238e-01 -7.24105537e-01
3.24128687e-01 -2.01060042e-01 6.41949824e-04 -4.34694022e-01
-6.33411705e-01 1.03939021e+00 3.77866238e-01 -7.41292834e-02
4.78060901e-01 3.64198163e-02 1.00172400e+00 5.88638365e-01
2.21575037e-01 -1.26957405e+00 -6.31186366e-02 1.04189801e+00
4.67290789e-01 -1.18889689e+00 -3.60034049e-01 3.97468731e-02
-4.67312038e-01 1.42048919e+00 -3.87802124e-02 -3.71815622e-01
6.06204152e-01 8.17431659e-02 3.26636404e-01 9.65805817e-03
-7.16330290e-01 -3.04809332e-01 3.97455961e-01 7.98520386e-01
8.40064704e-01 3.23780388e-01 -6.74035549e-01 2.49974459e-01
-2.61919379e-01 -5.60959280e-01 1.31681070e-01 8.09292674e-01
-3.87762636e-01 -1.65280318e+00 -1.60470337e-01 8.11047666e-03
-4.80014473e-01 -8.44562292e-01 -3.74879956e-01 6.27672076e-01
6.11285493e-02 6.96667552e-01 3.71282339e-01 -1.80369616e-01
1.92217156e-02 6.14482284e-01 4.53152031e-01 -8.14857423e-01
-6.25937700e-01 -1.53441653e-01 3.38466436e-01 -3.43502015e-01
-4.45343584e-01 -2.17697263e-01 -9.96599376e-01 -2.25640714e-01
-1.10921875e-01 3.45323265e-01 1.06819654e+00 1.14302731e+00
3.28159988e-01 3.08760822e-01 7.73868024e-01 -7.58085728e-01
-5.71860850e-01 -1.22867358e+00 -4.45522517e-01 5.23606479e-01
7.83695132e-02 -3.98566991e-01 -3.50855529e-01 -1.08588189e-01] | [11.007097244262695, 9.90283203125] |
923d9055-51ad-43cc-ac3a-a67e8bcca57c | fact-based-text-editing-1 | 2007.00916 | null | https://arxiv.org/abs/2007.00916v1 | https://arxiv.org/pdf/2007.00916v1.pdf | Fact-based Text Editing | We propose a novel text editing task, referred to as \textit{fact-based text editing}, in which the goal is to revise a given document to better describe the facts in a knowledge base (e.g., several triples). The task is important in practice because reflecting the truth is a common requirement in text editing. First, we propose a method for automatically generating a dataset for research on fact-based text editing, where each instance consists of a draft text, a revised text, and several facts represented in triples. We apply the method into two public table-to-text datasets, obtaining two new datasets consisting of 233k and 37k instances, respectively. Next, we propose a new neural network architecture for fact-based text editing, called \textsc{FactEditor}, which edits a draft text by referring to given facts using a buffer, a stream, and a memory. A straightforward approach to address the problem would be to employ an encoder-decoder model. Our experimental results on the two datasets show that \textsc{FactEditor} outperforms the encoder-decoder approach in terms of fidelity and fluency. The results also show that \textsc{FactEditor} conducts inference faster than the encoder-decoder approach. | ['chao qiao', 'Hayate Iso', 'Hang Li'] | 2020-07-02 | fact-based-text-editing | https://aclanthology.org/2020.acl-main.17 | https://aclanthology.org/2020.acl-main.17.pdf | acl-2020-6 | ['fact-based-text-editing'] | ['natural-language-processing'] | [ 5.69678783e-01 6.62210226e-01 -1.48644656e-01 -6.28888130e-01
-8.35969925e-01 -2.94647813e-01 8.31115603e-01 4.40250605e-01
-2.31880665e-01 1.24937224e+00 4.98864561e-01 -3.39546740e-01
-4.12736495e-04 -1.11967421e+00 -1.38679457e+00 1.73253253e-01
5.79216361e-01 5.96890867e-01 -6.80540800e-02 -3.45968336e-01
4.70011622e-01 -2.21218675e-01 -1.48378527e+00 8.08350086e-01
1.31674147e+00 7.66835809e-01 2.68602312e-01 5.06692171e-01
-7.08035529e-01 1.12642992e+00 -8.66716802e-01 -1.40999770e+00
2.78903339e-02 -6.08330965e-01 -1.07441378e+00 -1.31286472e-01
4.74183828e-01 -6.14953279e-01 -3.77329320e-01 1.12367463e+00
2.61381596e-01 9.49300528e-02 5.76580107e-01 -1.03162348e+00
-1.19529557e+00 1.52427649e+00 -4.02303994e-01 -1.45205617e-01
7.99717724e-01 -3.06160957e-01 7.54305899e-01 -9.99738514e-01
8.50605607e-01 1.24219656e+00 5.50677180e-01 4.87336993e-01
-7.24280059e-01 -3.98862571e-01 1.51533917e-01 2.65316755e-01
-1.28552997e+00 -6.66367412e-01 5.13972104e-01 -1.89996019e-01
1.17750204e+00 3.61656278e-01 5.91221809e-01 9.51758444e-01
4.68386084e-01 9.35273588e-01 4.59729999e-01 -5.94931245e-01
2.72251721e-02 3.53224665e-01 -1.36839464e-01 7.09064662e-01
7.92024791e-01 -2.38629773e-01 -7.96735942e-01 1.35295168e-01
4.65325266e-01 -1.62932068e-01 -2.08713040e-01 1.56007782e-01
-1.42166746e+00 6.14465177e-01 -1.21689789e-01 7.66066415e-03
-4.31314647e-01 2.51208156e-01 6.54813111e-01 4.25726235e-01
4.21415120e-01 5.48704386e-01 -4.28251326e-01 -3.41263205e-01
-9.54879045e-01 5.21468759e-01 1.07771111e+00 1.53543818e+00
6.26420975e-01 -6.73481822e-02 -4.00999814e-01 7.20843732e-01
6.00318378e-03 6.69926465e-01 3.80027533e-01 -9.49543297e-01
1.15720069e+00 7.18972802e-01 3.08078945e-01 -1.03223085e+00
1.69040973e-03 -4.43462133e-02 -9.73107457e-01 -5.11050999e-01
-2.12050695e-02 -3.20844024e-01 -7.73718596e-01 1.32271945e+00
1.63133517e-01 -3.33617091e-01 2.63450205e-01 5.37675440e-01
9.92727757e-01 8.48555744e-01 -2.42168918e-01 -2.57468641e-01
1.21765399e+00 -7.51982093e-01 -1.18673575e+00 -3.70379584e-03
6.97117031e-01 -7.15757191e-01 9.65713084e-01 4.00444090e-01
-1.35099649e+00 -3.61233860e-01 -9.92365062e-01 -6.59001231e-01
-9.16064084e-01 4.78458703e-01 6.53945506e-01 6.77128732e-01
-7.93176293e-01 4.84362543e-01 -4.18163180e-01 -3.76253009e-01
3.10235023e-01 -8.11508745e-02 -2.90148914e-01 -1.27490386e-01
-1.43767822e+00 8.67847621e-01 1.10732412e+00 1.37142882e-01
-5.90383351e-01 -8.42447877e-01 -9.97263491e-01 2.64213324e-01
9.30841327e-01 -1.05411887e+00 1.43857515e+00 -6.54273212e-01
-1.51872742e+00 5.69631815e-01 -3.87932181e-01 -5.36929667e-01
4.74909097e-01 -1.98248670e-01 -8.02063107e-01 -1.12831391e-01
2.11378276e-01 4.28073674e-01 8.10690045e-01 -1.35518944e+00
-1.09530747e+00 1.51531966e-02 4.57425058e-01 1.34337142e-01
-1.90519005e-01 -8.62653628e-02 -8.71410906e-01 -1.00354600e+00
-3.07011485e-01 -4.46029663e-01 2.90531754e-01 -2.05265567e-01
-9.38949347e-01 -1.52456120e-01 4.84549791e-01 -9.34751213e-01
1.73794317e+00 -1.57105625e+00 3.41640040e-02 1.31385580e-01
2.38741830e-01 4.81833577e-01 8.17744248e-03 9.04881120e-01
-1.61302816e-02 2.98062325e-01 -5.53195655e-01 -1.71797827e-01
2.47823507e-01 1.72552660e-01 -6.72683477e-01 -4.68166769e-01
2.73883879e-01 9.64164495e-01 -9.55194533e-01 -6.17044270e-01
-1.52143598e-01 9.35659930e-02 -7.81755745e-01 -1.43478826e-01
-7.43322194e-01 -1.89159542e-01 -4.20886368e-01 4.38539296e-01
6.34932637e-01 -9.04151201e-02 3.42917204e-01 -1.96934655e-01
-3.31089199e-02 5.22206247e-01 -1.27852738e+00 1.87391448e+00
-4.44615841e-01 5.46502113e-01 -4.86802727e-01 -7.41197467e-01
8.79017413e-01 3.81975114e-01 1.19647056e-01 -7.22494185e-01
-1.60127208e-02 2.03442976e-01 -5.33005118e-01 -7.25317597e-01
1.46506369e+00 2.30528452e-02 -1.43859744e-01 8.56130362e-01
-1.39548451e-01 -3.85335237e-01 8.23695838e-01 8.31102133e-01
8.44642401e-01 1.64100260e-01 2.89054453e-01 1.48014769e-01
5.24673522e-01 1.47762775e-01 4.76310670e-01 9.62615669e-01
5.17234802e-01 1.26096159e-01 7.57210016e-01 -3.80587995e-01
-1.10974479e+00 -6.62623584e-01 6.74382001e-02 6.81305289e-01
-7.64732063e-02 -8.89701605e-01 -7.52595663e-01 -7.86296666e-01
2.42852345e-01 1.40991020e+00 -7.12597668e-01 -4.51344639e-01
-4.08475876e-01 -2.89979368e-01 7.43541896e-01 4.94827449e-01
7.56862342e-01 -9.27142799e-01 -4.06458080e-01 3.75619829e-01
-6.79672837e-01 -1.10604858e+00 -7.96078682e-01 -1.95598692e-01
-4.85718787e-01 -1.01180053e+00 -2.60035127e-01 -5.92414200e-01
5.91961443e-01 5.60905412e-02 1.15270388e+00 1.19298130e-01
1.44969791e-01 2.37039506e-01 -5.81065238e-01 -8.38697195e-01
-7.67260492e-01 1.45103112e-01 -8.65331069e-02 -2.74272989e-02
3.44426244e-01 -2.32019514e-01 -5.19950353e-02 -1.31846070e-01
-1.49488366e+00 5.47493637e-01 6.61942899e-01 7.00573802e-01
6.30946815e-01 4.95508999e-01 7.18779266e-01 -1.65937221e+00
1.07239318e+00 -3.85788262e-01 -3.71621966e-01 8.49275649e-01
-6.41771734e-01 2.54897684e-01 9.95976746e-01 1.02773391e-01
-1.71937597e+00 -2.02840313e-01 -2.72042230e-02 2.23854147e-02
3.15976292e-01 1.20699418e+00 -2.58320183e-01 4.19941992e-01
5.96455216e-01 5.18528640e-01 -4.60993052e-01 -2.38613188e-01
7.17065215e-01 8.88727903e-01 8.62360895e-01 -8.16334009e-01
6.45045578e-01 3.02097112e-01 -3.81735116e-01 -3.99182469e-01
-1.15761971e+00 1.37251958e-01 -6.30059481e-01 -2.42838293e-01
4.45049345e-01 -8.69002759e-01 -6.37154579e-01 2.97197431e-01
-1.41656637e+00 -1.76561207e-01 -5.76531947e-01 2.58740634e-01
-6.62452102e-01 3.19305718e-01 -2.70963639e-01 -5.26183307e-01
-4.99685794e-01 -7.85691381e-01 1.10925925e+00 5.33391722e-02
-3.33128497e-02 -9.62343991e-01 -1.87208027e-01 4.68764782e-01
1.63809523e-01 2.00035810e-01 1.30044043e+00 -5.13738692e-01
-5.97784340e-01 -2.41440400e-01 -3.50195080e-01 1.11915834e-01
2.97116280e-01 3.15347582e-01 -4.85465318e-01 4.20433022e-02
-3.75783175e-01 -4.39003497e-01 8.02651048e-01 -1.11478485e-01
1.63367236e+00 -5.30780852e-01 -2.31558606e-01 4.55134809e-01
1.32451487e+00 6.06015265e-01 9.88162458e-01 3.70647423e-02
6.99414432e-01 2.79513717e-01 6.49477959e-01 7.59186089e-01
1.01537097e+00 4.18897033e-01 1.11070603e-01 3.70283574e-01
-1.31588802e-01 -6.95177078e-01 2.75963813e-01 1.12771368e+00
-1.52565548e-02 -6.89442098e-01 -7.39080429e-01 6.35176659e-01
-1.91342330e+00 -1.20764983e+00 -1.83398589e-01 2.01608849e+00
1.47234440e+00 1.18167631e-01 -4.37645346e-01 1.94021285e-01
7.27859259e-01 -5.90932928e-02 -6.13731325e-01 -7.06694961e-01
-5.30392453e-02 5.67341894e-02 2.50672042e-01 4.78523374e-01
-6.68969810e-01 1.05466580e+00 5.81028366e+00 7.42355227e-01
-7.93080628e-01 -2.88139045e-01 1.72163948e-01 -1.14483893e-01
-8.29889178e-01 2.55380142e-02 -9.19256032e-01 6.08868182e-01
7.95401812e-01 -8.42658997e-01 6.78459525e-01 3.75025153e-01
1.92003638e-01 -4.18375462e-01 -1.33473742e+00 8.66976261e-01
5.01466393e-01 -1.83009553e+00 9.92980719e-01 -4.18501168e-01
9.48918462e-01 -7.56149352e-01 -2.38217592e-01 6.11005008e-01
5.35511076e-01 -7.95094311e-01 1.06624281e+00 1.13587534e+00
1.32801330e+00 -6.85444653e-01 5.72602510e-01 3.37038070e-01
-1.05381215e+00 1.79664567e-01 -4.98358697e-01 1.34516001e-01
1.05272748e-01 1.03540349e+00 -7.77524948e-01 1.28130519e+00
5.47844052e-01 8.63138199e-01 -4.06006813e-01 6.82759464e-01
-2.26068035e-01 1.50052637e-01 -4.74065542e-02 -1.40769020e-01
-1.62672222e-01 9.82500962e-04 3.67489487e-01 1.32190979e+00
3.43731523e-01 3.59394789e-01 -9.55549255e-02 9.97398615e-01
-8.82549405e-01 -9.13553238e-02 -5.36900222e-01 -3.24033767e-01
7.30493784e-01 7.58299708e-01 -3.30534399e-01 -9.74903047e-01
-3.46498400e-01 1.03538656e+00 4.10681397e-01 3.57270241e-01
-9.52839732e-01 -1.13556015e+00 7.03691393e-02 -2.01433301e-01
4.34443951e-01 1.09937161e-01 -4.72870469e-01 -1.54983306e+00
6.27502024e-01 -1.16411197e+00 2.94943392e-01 -1.18647301e+00
-8.27069700e-01 4.40047503e-01 1.59563854e-01 -9.05958951e-01
-1.65869251e-01 -2.00311750e-01 -3.51240665e-01 8.57290387e-01
-1.52261293e+00 -9.15926397e-01 -3.21280539e-01 7.43091762e-01
5.68553925e-01 -1.90302487e-02 5.56117952e-01 5.13161778e-01
-6.04956627e-01 7.82003880e-01 1.93163007e-01 1.62629724e-01
8.39021444e-01 -1.30586481e+00 7.25490868e-01 1.05753839e+00
-5.50077744e-02 9.34738696e-01 5.42109072e-01 -1.26603818e+00
-1.73480582e+00 -1.52619696e+00 1.55519187e+00 -4.72809792e-01
4.46700811e-01 -3.28061968e-01 -8.86219680e-01 1.22903752e+00
4.45327282e-01 -6.74401402e-01 4.08494681e-01 -1.16673514e-01
-1.31577894e-01 -1.42018795e-01 -1.03540766e+00 7.26791561e-01
1.18816006e+00 -7.16579020e-01 -1.00571847e+00 5.27941108e-01
9.74379659e-01 -9.49295819e-01 -9.06188130e-01 -1.66517019e-01
3.93317401e-01 -6.40017927e-01 4.04322624e-01 -7.83995867e-01
1.03798425e+00 -3.47422063e-01 1.82871018e-02 -1.64152741e+00
-8.98033157e-02 -7.87456095e-01 -4.63005662e-01 1.47442937e+00
7.30968773e-01 -4.69565511e-01 4.20948595e-01 7.49782860e-01
-3.05978060e-01 -5.30127585e-01 -6.73687816e-01 -7.05048978e-01
-6.24899641e-02 -5.59600651e-01 1.33257663e+00 1.04888988e+00
1.62606150e-01 2.89369762e-01 -5.59648812e-01 -1.29623085e-01
2.09284887e-01 1.26942262e-01 8.47162008e-01 -9.50088739e-01
1.66813046e-01 -7.10291043e-02 3.33413601e-01 -1.14177883e+00
-1.28278941e-01 -1.17889071e+00 2.45891139e-02 -2.05689311e+00
2.15624347e-01 -2.11360350e-01 3.61347854e-01 6.30104661e-01
-3.51892978e-01 -4.99829233e-01 1.45092502e-01 -8.87564123e-02
-5.11457503e-01 8.41793835e-01 1.37466407e+00 -3.71725708e-01
-2.70861443e-02 -3.12083334e-01 -1.14391482e+00 3.99475843e-01
4.44082052e-01 -5.76249242e-01 -4.44020718e-01 -1.05842316e+00
8.66573870e-01 4.14696842e-01 1.31513327e-01 -7.64983594e-01
6.70828581e-01 -2.77895987e-01 2.14374870e-01 -8.42638433e-01
6.55001998e-02 -8.41816247e-01 2.29796365e-01 8.21102634e-02
-7.02730656e-01 1.48455054e-01 1.32436991e-01 5.92836261e-01
-2.53548086e-01 -3.42637002e-01 1.54992506e-01 -1.77044347e-01
-6.70980036e-01 1.33440539e-01 -3.85149002e-01 4.00364220e-01
7.81733990e-01 -1.30501106e-01 -7.41968453e-01 -5.33182263e-01
-3.64441812e-01 3.87464613e-01 -3.15944739e-02 4.17709529e-01
7.30317533e-01 -1.38695145e+00 -8.13739240e-01 8.84395763e-02
3.84601355e-01 4.72208485e-02 3.31707932e-02 4.07445103e-01
-5.11379778e-01 8.78497183e-01 -4.37561944e-02 1.22987263e-01
-6.06043935e-01 4.71920371e-01 2.44235128e-01 -3.98076892e-01
-5.91456115e-01 5.69910944e-01 -2.99464405e-01 -6.69577241e-01
1.64548218e-01 -8.82300377e-01 7.90378302e-02 1.62922721e-02
5.90786755e-01 3.70674878e-01 5.10084093e-01 -2.06107497e-02
1.91117655e-02 7.95883909e-02 -4.34465975e-01 8.85442421e-02
1.06398213e+00 -2.21976951e-01 -2.60436594e-01 2.98759282e-01
7.41315901e-01 8.30001831e-02 -7.46374726e-01 -4.79640603e-01
-6.34356262e-03 -4.56393838e-01 -1.32746175e-01 -1.38821435e+00
-9.77261245e-01 4.35182422e-01 -3.20653439e-01 1.91643015e-01
1.15482700e+00 -4.72829551e-01 1.14874351e+00 1.04053271e+00
4.98846918e-01 -1.44372547e+00 -3.20747077e-01 8.73265803e-01
1.26860559e+00 -1.06769800e+00 1.76743224e-01 -4.99236226e-01
-8.87820363e-01 1.20492852e+00 6.09821379e-01 4.98260260e-01
3.71337086e-01 2.29034442e-02 -3.00057530e-01 -3.40209574e-01
-1.17031276e+00 8.16045478e-02 5.46373427e-02 4.88648057e-01
3.51508975e-01 6.91458583e-02 -3.55329901e-01 7.79609680e-01
-5.74969769e-01 7.02519953e-01 1.02109528e+00 9.64251757e-01
-2.36014813e-01 -9.12411928e-01 -2.61904836e-01 1.12158906e+00
-2.19942048e-01 -3.65138113e-01 -6.42933190e-01 6.82157576e-01
1.72943145e-01 9.65973079e-01 -5.28063402e-02 -4.52756941e-01
7.65980303e-01 1.82770774e-01 3.25874895e-01 -6.91313028e-01
-8.15815866e-01 -6.91291928e-01 5.54476142e-01 -3.15545082e-01
-2.66058296e-01 -5.27888119e-01 -1.35691547e+00 -8.30925941e-01
-1.76722214e-01 3.03238273e-01 6.54123366e-01 9.85048294e-01
5.76595366e-01 8.09192777e-01 3.13278735e-01 8.58074278e-02
-2.46865988e-01 -7.87423730e-01 -5.48157215e-01 2.45441005e-01
5.14430702e-02 -3.82673681e-01 2.83357054e-01 7.50181556e-01] | [11.610897064208984, 8.715604782104492] |
eba2f3e0-9250-43d9-8ed0-a0937d058f92 | pulse-shape-aided-multipath-delay-estimation | 2306.15320 | null | https://arxiv.org/abs/2306.15320v1 | https://arxiv.org/pdf/2306.15320v1.pdf | Pulse Shape-Aided Multipath Delay Estimation for Fine-Grained WiFi Sensing | Due to the finite bandwidth of practical wireless systems, one multipath component can manifest itself as a discrete pulse consisting of multiple taps in the digital delay domain. This effect is called channel leakage, which complicates the multipath delay estimation problem. In this paper, we develop a new algorithm to estimate multipath delays of leaked channels by leveraging the knowledge of pulse-shaping functions, which can be used to support fine-grained WiFi sensing applications. Specifically, we express the channel impulse response (CIR) as a linear combination of overcomplete basis vectors corresponding to different delays. Considering the limited number of paths in physical environments, we formulate the multipath delay estimation as a sparse recovery problem. We then propose a sparse Bayesian learning (SBL) method to estimate the sparse vector and determine the number of physical paths and their associated delay parameters from the positions of the nonzero entries in the sparse vector. Simulation results show that our algorithm can accurately determine the number of paths, and achieve superior accuracy in path delay estimation and channel reconstruction compared to two benchmarking schemes. | ['Chenshu Wu', 'He Chen', 'Ke Xu'] | 2023-06-27 | null | null | null | null | ['benchmarking', 'benchmarking'] | ['miscellaneous', 'robots'] | [ 1.57142967e-01 -5.05220473e-01 -3.36508863e-02 -7.15480074e-02
-9.29287136e-01 -5.53265989e-01 -7.56944492e-02 -3.44507918e-02
-2.95708198e-02 7.22135544e-01 3.61176789e-01 -4.14942026e-01
-8.51717964e-02 -5.74464202e-01 -5.18329620e-01 -1.01926756e+00
-4.98569131e-01 -1.20145984e-01 4.25158665e-02 2.26510391e-01
1.04118675e-01 2.54177123e-01 -5.83921373e-01 -8.96743909e-02
5.17788708e-01 1.11902320e+00 -4.48166020e-02 5.39632082e-01
-6.93809167e-02 4.07828569e-01 -6.22036338e-01 2.54208118e-01
9.38259885e-02 -1.88423768e-01 2.25279629e-02 -3.38588923e-01
-2.77213514e-01 -6.74643338e-01 -9.44042921e-01 7.51299977e-01
5.11388838e-01 2.22740807e-02 5.58833361e-01 -1.09795856e+00
2.27766007e-01 7.39653289e-01 -6.42299831e-01 8.84461626e-02
4.44007367e-01 -5.73841691e-01 5.63912034e-01 -8.21933687e-01
-2.51189843e-02 1.20729709e+00 1.12434530e+00 -3.04066360e-01
-8.25961888e-01 -1.08857822e+00 -1.22603275e-01 1.79186195e-01
-1.90610826e+00 -6.96442664e-01 9.88923430e-01 -2.43244007e-01
2.51077265e-01 3.92366201e-02 3.37886006e-01 1.10587728e+00
2.00676903e-01 5.08391559e-01 8.85481477e-01 -4.01845783e-01
4.37606514e-01 -1.67196259e-01 1.78481191e-02 4.41609323e-01
4.77149636e-01 2.90410310e-01 -6.14176035e-01 -6.94496691e-01
9.49838817e-01 -2.18953900e-02 -5.79088449e-01 8.88692737e-02
-1.07579267e+00 5.12273908e-01 1.56662241e-01 1.06982552e-01
-6.20176792e-01 4.91950631e-01 -1.19358078e-01 1.87048584e-01
-7.51730502e-02 -1.50820345e-01 -2.52721220e-01 -7.84220397e-02
-1.11399388e+00 6.00869693e-02 1.05549634e+00 9.27561045e-01
9.64274406e-01 2.35116720e-01 -1.70132264e-01 6.05160594e-01
7.29089618e-01 1.20522976e+00 -2.13353813e-01 -8.70201647e-01
4.05502379e-01 -5.65888166e-01 4.99053657e-01 -1.39284015e+00
-4.66494590e-01 -6.17466211e-01 -1.11107469e+00 -6.21827781e-01
4.48990434e-01 -6.86153114e-01 -4.85555381e-01 1.48551476e+00
2.43872002e-01 1.05532980e+00 8.67016315e-02 6.68241799e-01
3.70725006e-01 9.80565012e-01 -1.65968299e-01 -3.75917554e-01
8.78048122e-01 -3.93211544e-01 -6.86886907e-01 -2.32329816e-01
3.60491127e-01 -9.03041184e-01 -1.94599666e-02 5.59093595e-01
-6.17443204e-01 -1.45148143e-01 -1.11645389e+00 7.87100434e-01
4.09304291e-01 1.79463357e-01 5.69068611e-01 8.70260179e-01
-4.97875482e-01 1.41591206e-01 -7.57471323e-01 3.34390044e-01
5.70212081e-02 -1.03279538e-01 5.97656853e-02 -6.81375921e-01
-1.28799891e+00 -4.42859381e-02 1.78558901e-02 4.57226068e-01
-8.56590688e-01 -9.06676769e-01 -6.66048288e-01 9.11504030e-02
1.98484719e-01 -1.75583839e-01 1.11533654e+00 -2.08218917e-01
-1.04962075e+00 -3.86004746e-01 -4.75476027e-01 -2.73512095e-01
-4.04333882e-02 -1.33444831e-01 -1.04200923e+00 2.79026896e-01
2.88838462e-04 -3.39516163e-01 1.22248518e+00 -1.37020111e+00
-4.23899531e-01 -1.21449344e-02 -1.51130930e-01 -2.27774739e-01
-5.78108057e-02 -4.77274448e-01 -4.66430813e-01 -7.70144045e-01
6.62669122e-01 -1.09103000e+00 -5.52627146e-01 -1.25946015e-01
-5.58574378e-01 7.90699601e-01 4.74330276e-01 -9.30997789e-01
1.64183033e+00 -2.49625516e+00 -3.25240105e-01 9.23170328e-01
-5.46478443e-02 -4.91746627e-02 -5.73667772e-02 9.14272189e-01
4.16713655e-01 -3.99030864e-01 -3.85042161e-01 -1.29464008e-02
-1.83263972e-01 2.37676516e-01 -5.65716326e-01 7.02179313e-01
-4.69240099e-01 1.61482953e-02 -1.04724574e+00 -1.54408440e-01
-1.37785405e-01 8.24447274e-01 -4.00584459e-01 1.66901171e-01
3.68081301e-01 8.08968365e-01 -9.71937001e-01 5.41280508e-01
1.23503375e+00 4.20526266e-02 2.33375892e-01 -5.14269054e-01
-3.44782472e-01 -1.79366712e-02 -1.68710518e+00 1.71929574e+00
-7.25993633e-01 3.52506071e-01 4.87741917e-01 -7.71119416e-01
9.55366254e-01 5.95627308e-01 6.65052056e-01 -7.78512120e-01
-2.78602149e-02 4.57057923e-01 -3.34394991e-01 -3.37109953e-01
1.22119620e-01 -4.23148014e-02 -2.61545777e-01 5.31743765e-01
-3.09368491e-01 4.00946587e-01 -3.96726489e-01 1.38423592e-01
1.38305473e+00 -4.49607730e-01 1.17835350e-01 1.67750433e-01
5.74542463e-01 -6.10309064e-01 8.49013925e-01 1.00527394e+00
1.89888790e-01 1.85952619e-01 1.61196917e-01 -4.20541093e-02
-5.60108721e-01 -9.34691370e-01 -5.69363199e-02 6.10950947e-01
3.16353053e-01 -3.81492108e-01 -2.48974800e-01 -3.87566015e-02
1.31695867e-01 3.36707324e-01 -8.34639464e-03 -1.80584729e-01
-5.77035964e-01 -5.88836312e-01 8.36007476e-01 1.30708367e-01
4.06474262e-01 -2.40053669e-01 -1.67246744e-01 8.14147592e-01
-4.57111001e-01 -1.40436089e+00 -3.77002776e-01 -4.92910556e-02
-5.30152082e-01 -7.97714531e-01 -7.32961476e-01 -4.46436524e-01
4.01840597e-01 9.25981998e-01 4.77336943e-01 -2.19182789e-01
-1.64283216e-01 6.34444535e-01 -5.82665563e-01 -2.36428499e-01
6.81359097e-02 -3.63620102e-01 1.73708573e-02 5.29490590e-01
-2.61651844e-01 -9.66671348e-01 -7.69003332e-01 3.96907717e-01
-6.05536878e-01 -4.02246416e-01 8.09703827e-01 5.50699413e-01
5.33852756e-01 6.62166059e-01 5.47785163e-01 -4.88672823e-01
4.51396167e-01 -9.90039945e-01 -6.09723687e-01 4.78735380e-02
-2.43692547e-01 8.64104629e-02 3.39659244e-01 -2.52228349e-01
-1.09483600e+00 1.58992022e-01 -2.28736222e-01 -2.35412735e-02
2.12523788e-01 9.75110650e-01 -2.25302279e-01 -7.01535702e-01
2.77710766e-01 4.19055462e-01 -6.12610877e-01 -6.43793464e-01
3.37496310e-01 6.56566918e-01 2.98065811e-01 -7.57572830e-01
1.08874965e+00 6.92455828e-01 3.13744485e-01 -1.32498515e+00
-6.29064620e-01 -7.71475673e-01 -1.92634925e-01 3.96635830e-02
-1.03709087e-01 -1.25161350e+00 -3.74240249e-01 5.42321742e-01
-1.06147170e+00 -3.22247744e-02 5.74512422e-01 8.43206882e-01
4.40873113e-03 7.03938901e-01 -5.32094419e-01 -1.01830161e+00
-2.30171889e-01 -6.50925040e-01 6.36441171e-01 -1.81859955e-02
1.40910186e-02 -8.93922269e-01 3.94178063e-01 -3.53104860e-01
6.46184266e-01 1.86094642e-01 7.74435163e-01 6.24604970e-02
-7.05873013e-01 -4.32995558e-01 -2.48679131e-01 -8.45966041e-02
-5.00356928e-02 -5.61383188e-01 -6.29091501e-01 -4.47079629e-01
-1.26576826e-01 4.26282525e-01 5.60978651e-01 7.45055139e-01
8.31787407e-01 -4.03823346e-01 -6.87081635e-01 8.69541347e-01
1.56080174e+00 1.07405499e-01 6.67202652e-01 -2.86205590e-01
7.54752517e-01 -2.41933968e-02 8.41207325e-01 1.27929902e+00
4.29037303e-01 6.27385318e-01 8.86265337e-02 3.07193547e-01
6.37755767e-02 -2.90448695e-01 2.07238913e-01 9.28608656e-01
2.57901788e-01 -1.71855599e-01 -6.60281658e-01 3.45121086e-01
-1.62948024e+00 -7.85541117e-01 -3.48715663e-01 2.40330029e+00
4.74554837e-01 -1.75215930e-01 -1.54295176e-01 2.78676480e-01
4.64789867e-01 3.27452660e-01 -4.03360724e-01 1.73036262e-01
1.49529621e-01 2.45337352e-01 1.06813717e+00 5.67690134e-01
-7.03663349e-01 3.37219805e-01 6.36852980e+00 6.92283452e-01
-1.18864429e+00 6.04101978e-02 -1.47567019e-01 4.12362397e-01
-6.60444379e-01 2.20252410e-01 -7.69466221e-01 6.20511889e-01
9.15095627e-01 -5.01239151e-02 5.04578710e-01 1.45860791e-01
4.86858398e-01 -1.48523554e-01 -6.68533444e-01 1.21910632e+00
-4.04508799e-01 -1.05393314e+00 -4.06205237e-01 -3.10076941e-02
6.19998872e-01 -1.26528859e-01 2.78051775e-02 1.03247255e-01
-1.32130474e-01 -7.71142244e-01 6.17665231e-01 7.16199517e-01
6.56205595e-01 -7.21875906e-01 4.40115482e-01 2.34060079e-01
-1.55319190e+00 -3.43787163e-01 -3.76200229e-01 -7.49442801e-02
5.58016539e-01 1.44312394e+00 -9.02053118e-01 5.98471761e-01
3.28803480e-01 5.70753396e-01 3.91123056e-01 1.74814177e+00
-2.99764067e-01 1.13445520e+00 -5.31269073e-01 3.28653753e-01
2.12821618e-01 -3.79524603e-02 6.85304821e-01 1.35764635e+00
1.13304400e+00 4.55841422e-01 3.30772698e-01 2.97932923e-01
1.27941400e-01 -3.33689749e-02 -1.85997441e-01 9.54504684e-02
1.40908587e+00 9.69973683e-01 -2.12284759e-01 2.73713201e-01
-5.90513051e-01 7.07177818e-01 -5.21758974e-01 1.04666328e+00
-6.80834591e-01 -4.69607890e-01 5.80013573e-01 -5.78023270e-02
6.12566888e-01 -8.19383204e-01 9.11828876e-02 -8.11469793e-01
-7.38748834e-02 -6.38362348e-01 -9.47186947e-02 -2.04010293e-01
-8.97651434e-01 3.57193500e-02 -4.64576542e-01 -1.43699288e+00
1.25603667e-02 1.10207342e-01 -6.68579102e-01 8.93965304e-01
-1.92568612e+00 -8.48027408e-01 -3.20359141e-01 8.35425138e-01
-2.17607215e-01 1.04586266e-01 1.08719599e+00 7.54732847e-01
-3.79808486e-01 5.88588357e-01 7.53285468e-01 8.00307393e-02
5.97079277e-01 -4.40882474e-01 9.71214399e-02 7.10286915e-01
2.12860312e-02 4.92216349e-01 8.06236386e-01 -5.98820806e-01
-1.73828626e+00 -1.08735204e+00 5.81912100e-01 4.86088097e-01
5.74769557e-01 8.37694332e-02 -7.91101873e-01 4.09501612e-01
-5.75366557e-01 1.52850732e-01 1.14640021e+00 -1.95804480e-02
-5.67690909e-01 -2.61418134e-01 -9.21344995e-01 5.16729318e-02
4.41563785e-01 -5.29369116e-01 1.38694033e-01 1.05921298e-01
4.15235400e-01 -2.23214790e-01 -7.03300595e-01 8.10329840e-02
9.81527448e-01 -5.95924675e-01 1.38755810e+00 2.46829540e-01
-2.55256146e-01 -3.83488864e-01 -5.39846361e-01 -1.18213546e+00
-6.04568362e-01 -8.11004221e-01 -5.03474712e-01 1.24759567e+00
1.24558762e-01 -4.40901637e-01 5.10894358e-01 3.17314230e-02
1.17883816e-01 -2.58941621e-01 -9.07110035e-01 -6.79282248e-01
-5.32959223e-01 -6.16078436e-01 7.11512148e-01 5.80236375e-01
-2.32030645e-01 5.49817868e-02 -9.77159977e-01 1.11744368e+00
1.38106835e+00 -6.51761889e-02 4.70489562e-01 -1.18104136e+00
-7.87836015e-01 2.27988034e-01 -7.28132874e-02 -1.55536675e+00
-3.14609036e-02 -4.40648526e-01 3.73517215e-01 -1.23654568e+00
-3.58039618e-01 -1.31485522e+00 -6.03349328e-01 -8.56871624e-03
2.33643875e-01 6.59345835e-02 -3.13285112e-01 1.73928380e-01
-4.34628189e-01 4.89567161e-01 8.92931461e-01 7.77906738e-03
-1.92176849e-01 6.61400080e-01 -6.57414973e-01 6.80004656e-01
5.43938577e-01 -6.88129067e-01 -3.65640432e-01 -5.64777970e-01
2.73823500e-01 8.98607016e-01 2.42336005e-01 -1.32156444e+00
3.58197540e-01 -1.48363590e-01 1.58190101e-01 -6.30884528e-01
5.69418609e-01 -1.05070055e+00 4.52049971e-01 4.06694531e-01
-5.11712581e-03 -7.48613775e-01 3.94841656e-02 1.15269923e+00
-1.35480583e-01 -2.52907518e-02 2.56441951e-01 1.99188828e-01
-7.73868263e-01 5.82700729e-01 -6.70026779e-01 -2.06265241e-01
4.67971355e-01 1.61796764e-01 1.08786315e-01 -8.30738366e-01
-3.85224015e-01 5.11804558e-02 -1.46639869e-01 -1.30856723e-01
7.45734274e-01 -1.42826271e+00 -7.49903738e-01 1.75025001e-01
2.57230233e-02 -4.66845602e-01 7.56524384e-01 8.17114651e-01
-2.08459020e-01 6.45188808e-01 2.97984537e-02 -2.50976413e-01
-8.84356380e-01 -1.57834832e-02 -1.24897889e-03 7.91809186e-02
-6.26527011e-01 9.78900075e-01 -9.59007591e-02 1.74048692e-01
5.18324316e-01 1.06143460e-01 1.09897228e-02 -2.38901004e-01
1.02760220e+00 4.95762914e-01 -1.86395703e-03 -5.45880675e-01
-4.78068143e-01 7.90564775e-01 2.55129457e-01 -2.58700460e-01
1.04992521e+00 -4.92879272e-01 -3.72592248e-02 1.58746079e-01
1.53398824e+00 5.83785295e-01 -1.28416836e+00 -9.03464377e-01
-3.03873867e-01 -7.16272295e-01 2.61041045e-01 -5.65279007e-01
-8.37491810e-01 6.27789676e-01 5.42816103e-01 -2.50889629e-01
1.02257347e+00 -3.53165388e-01 1.34755468e+00 4.14093703e-01
1.06434882e+00 -6.33851230e-01 -2.70913616e-02 5.63082814e-01
3.20527345e-01 -6.98377728e-01 2.17613503e-02 -6.11658335e-01
7.15296865e-02 1.10319006e+00 -3.96909922e-01 -6.21133931e-02
1.11901164e+00 4.54327792e-01 1.07576653e-01 1.90963492e-01
-8.98866430e-02 2.20635295e-01 4.16476205e-02 6.99727297e-01
1.35996312e-01 4.01421636e-01 -1.35790870e-01 8.73169422e-01
7.22626746e-02 -7.30796531e-02 5.57788014e-01 1.01774144e+00
-8.94546032e-01 -1.35887599e+00 -7.96929598e-01 2.27331176e-01
-2.64674395e-01 -1.49032533e-01 5.08162200e-01 -5.09884069e-03
-2.19925076e-01 1.20477653e+00 -4.08368260e-01 -3.87807012e-01
1.03479184e-01 -4.09009755e-01 5.20148456e-01 -1.85452044e-01
4.98157084e-01 3.07549715e-01 1.60422444e-01 -5.81186831e-01
8.32872316e-02 -6.38063431e-01 -1.48252451e+00 -3.45075786e-01
-1.17667899e-01 4.27024364e-01 7.83122838e-01 9.88328695e-01
1.35156602e-01 5.76773882e-01 1.11246109e+00 -7.04540908e-01
-5.68127453e-01 -2.60668963e-01 -1.12052476e+00 -2.30200842e-01
7.01256216e-01 -5.97359419e-01 -3.12476993e-01 -3.57072175e-01] | [6.410588264465332, 1.2915033102035522] |
c2dccd80-b7c3-4287-a660-505f7e5c6d99 | classification-of-breast-tumours-based-on | 2209.01380 | null | https://arxiv.org/abs/2209.01380v1 | https://arxiv.org/pdf/2209.01380v1.pdf | Classification of Breast Tumours Based on Histopathology Images Using Deep Features and Ensemble of Gradient Boosting Methods | Breast cancer is the most common cancer among women worldwide. Early-stage diagnosis of breast cancer can significantly improve the efficiency of treatment. Computer-aided diagnosis (CAD) systems are widely adopted in this issue due to their reliability, accuracy and affordability. There are different imaging techniques for a breast cancer diagnosis; one of the most accurate ones is histopathology which is used in this paper. Deep feature transfer learning is used as the main idea of the proposed CAD system's feature extractor. Although 16 different pre-trained networks have been tested in this study, our main focus is on the classification phase. The Inception-ResNet-v2 which has both residual and inception networks profits together has shown the best feature extraction capability in the case of breast cancer histopathology images among all tested CNNs. In the classification phase, the ensemble of CatBoost, XGBoost and LightGBM has provided the best average accuracy. The BreakHis dataset was used to evaluate the proposed method. BreakHis contains 7909 histopathology images (2,480 benign and 5,429 malignant) in four magnification factors. The proposed method's accuracy (IRv2-CXL) using 70% of BreakHis dataset as training data in 40x, 100x, 200x and 400x magnification is 96.82%, 95.84%, 97.01% and 96.15%, respectively. Most studies on automated breast cancer detection have focused on feature extraction, which made us attend to the classification phase. IRv2-CXL has shown better or comparable results in all magnifications due to using the soft voting ensemble method which could combine the advantages of CatBoost, XGBoost and LightGBM together. | ['Samaneh Emami', 'Hamid Nasiri', 'Sayed Ali Sheikholeslamzadeh', 'Mohammad Reza Abbasniya'] | 2022-09-03 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [-1.24524474e-01 1.90315526e-02 -1.56259164e-01 -3.45267236e-01
-5.11656284e-01 -8.63490626e-03 5.66308975e-01 2.89000690e-01
-6.97309375e-01 7.73449540e-01 -1.49385676e-01 -5.40951729e-01
-3.87499988e-01 -8.45740080e-01 -2.01366037e-01 -9.74383712e-01
4.16229852e-02 2.80067563e-01 1.66268826e-01 -3.53874445e-01
2.23169863e-01 7.74540424e-01 -1.16758430e+00 5.40267050e-01
4.18600321e-01 1.08280683e+00 2.40476593e-01 9.41893578e-01
2.79872995e-02 7.04803646e-01 -2.67715454e-01 -3.38424891e-01
2.18168572e-01 -1.74388736e-01 -8.87167156e-01 -2.66540647e-01
1.10778198e-01 -2.05535889e-01 -4.59734015e-02 7.34566391e-01
8.12960029e-01 -3.18417758e-01 8.33607674e-01 -1.03786790e+00
-9.93375257e-02 2.80676186e-01 -8.59800696e-01 5.90234458e-01
-1.84342474e-01 -4.18146513e-02 4.35994655e-01 -6.69257760e-01
6.77693427e-01 9.64919627e-01 9.04258132e-01 5.62570870e-01
-8.78569603e-01 -8.70289803e-01 -8.47079992e-01 5.02304137e-01
-1.29579008e+00 -1.73524037e-01 4.48793560e-01 -4.44438040e-01
9.24348652e-01 5.49495518e-01 7.49971926e-01 5.58737278e-01
9.76260960e-01 4.77751642e-01 1.55648613e+00 -6.90642834e-01
4.86887340e-03 4.66142446e-01 4.19443876e-01 7.73047030e-01
5.14627934e-01 2.71667302e-01 2.35357620e-02 -1.43801287e-01
5.23784041e-01 2.82365113e-01 -4.33085077e-02 1.83586687e-01
-6.52484715e-01 8.57470334e-01 8.38827610e-01 9.26649868e-01
-3.23059916e-01 9.73693207e-02 5.96130729e-01 3.58051360e-01
2.14507401e-01 2.78786182e-01 -3.62332046e-01 3.68236482e-01
-7.74074018e-01 8.93683266e-03 3.91921043e-01 4.30864483e-01
4.85638857e-01 -2.99079806e-01 1.30905900e-02 8.01007807e-01
3.69165301e-01 4.35151726e-01 8.24504733e-01 -2.61503190e-01
-7.92668015e-02 9.26139712e-01 -4.10589039e-01 -1.25833666e+00
-7.82356977e-01 -6.11312449e-01 -1.30339456e+00 5.39396644e-01
1.83368832e-01 -7.85362348e-03 -1.29423642e+00 9.72607136e-01
3.26075524e-01 -2.50593722e-01 2.84201235e-01 6.70225024e-01
1.24416077e+00 5.24868011e-01 2.34310955e-01 5.21309972e-02
1.73875940e+00 -5.94564319e-01 -5.13700843e-01 2.08582953e-01
7.85310328e-01 -8.80307794e-01 2.06280798e-01 2.60126650e-01
-6.46339774e-01 -6.06780767e-01 -1.21635878e+00 8.61893669e-02
-6.38660967e-01 4.11181331e-01 7.15577662e-01 6.60289466e-01
-1.12730801e+00 3.76802325e-01 -9.63357270e-01 -8.77913892e-01
7.12072551e-01 6.46377563e-01 -8.35624099e-01 -2.74400681e-01
-8.51875246e-01 1.11639774e+00 3.45438600e-01 1.18223384e-01
-6.38219178e-01 -5.12145519e-01 -3.04873616e-01 -9.46434215e-02
-1.77295804e-01 -5.34325480e-01 1.02795851e+00 -1.03805244e+00
-1.06928742e+00 1.12253141e+00 1.58069760e-01 -4.80630755e-01
5.76859713e-01 3.55341792e-01 -3.22877228e-01 3.45622957e-01
7.54996240e-02 7.86098182e-01 7.67104700e-02 -9.88503575e-01
-1.02254486e+00 -5.39180875e-01 -2.01504081e-01 -2.28288956e-02
-2.55029202e-01 -3.29325423e-02 -6.28984049e-02 -2.94908434e-01
2.33092040e-01 -8.67952704e-01 -6.24511503e-02 2.27798857e-02
-3.37209731e-01 -2.10354522e-01 9.68405247e-01 -7.09532797e-01
8.84969890e-01 -2.04562020e+00 -3.53451997e-01 3.83218646e-01
2.35256895e-01 4.61713314e-01 2.85710067e-01 3.39232951e-01
-2.14930460e-01 2.76131123e-01 2.22243920e-01 -5.46055585e-02
-5.97362697e-01 1.15206361e-01 5.79223871e-01 6.68177307e-01
1.43313512e-01 7.61148691e-01 -4.04244900e-01 -8.68953943e-01
3.01466405e-01 5.87319314e-01 -1.29569501e-01 -2.77424343e-02
5.18722177e-01 1.27601564e-01 -4.11422074e-01 9.24049079e-01
8.24000180e-01 -1.26298636e-01 1.52109088e-02 -5.27663529e-01
-3.43433172e-02 -7.07633257e-01 -6.96744859e-01 1.14216793e+00
-3.00068140e-01 8.18152130e-01 -1.36926800e-01 -1.03876746e+00
8.78784716e-01 4.76758242e-01 5.79311132e-01 -7.63590932e-01
5.53744435e-01 5.15990317e-01 4.94096160e-01 -8.02219093e-01
2.95904409e-02 -2.89515525e-01 4.35015261e-01 -5.79232238e-02
1.86483368e-01 2.95102417e-01 2.08583370e-01 -1.27686011e-02
1.24198711e+00 -4.60496664e-01 8.49315166e-01 -5.58526814e-01
6.50992095e-01 3.65630418e-01 2.75957853e-01 4.50736552e-01
-3.03511798e-01 4.26793516e-01 4.60243344e-01 -7.25690842e-01
-8.41956437e-01 -4.08450723e-01 -5.83004415e-01 5.71673512e-01
-1.44745752e-01 2.79497147e-01 -3.90013218e-01 -6.21152937e-01
-3.11506242e-02 6.71077818e-02 -1.14874768e+00 -4.92562167e-02
-4.90379304e-01 -1.09875655e+00 6.49996519e-01 3.65010768e-01
9.90330815e-01 -1.11275423e+00 -8.76261175e-01 -4.32166457e-03
2.86693037e-01 -4.21340972e-01 4.75179106e-01 4.21322644e-01
-1.07229424e+00 -1.55170763e+00 -8.98489654e-01 -8.44416142e-01
1.00466669e+00 1.79085582e-01 6.90402567e-01 5.59337139e-01
-8.88109028e-01 -1.38243005e-01 -4.48104501e-01 -6.78337574e-01
-5.86393058e-01 2.35625997e-01 -6.35065258e-01 -1.15784541e-01
4.74397749e-01 -4.59204651e-02 -7.34738231e-01 -3.89344990e-02
-7.59888232e-01 4.04188922e-03 1.11786842e+00 1.10151541e+00
4.30970639e-01 1.41051859e-01 3.17970306e-01 -1.06255603e+00
2.09291250e-01 -4.72113520e-01 -2.22348198e-01 2.65306905e-02
-6.27082229e-01 -4.51528460e-01 3.95430028e-01 -6.30197823e-02
-7.77788460e-01 -1.87455118e-01 -2.99684435e-01 2.52438430e-02
-2.40922064e-01 5.83390057e-01 2.18922183e-01 -3.43503267e-01
8.66198897e-01 -1.47663206e-01 2.67401963e-01 -2.44501278e-01
-6.95798099e-01 7.51956761e-01 1.53262973e-01 1.34356141e-01
3.36083978e-01 5.29900193e-01 4.51134026e-01 -7.85854340e-01
-2.40738660e-01 -6.59414053e-01 -2.62496918e-01 -2.69214004e-01
1.04355156e+00 -6.95901573e-01 -7.02128112e-01 8.12524259e-01
-7.73111045e-01 -9.15565901e-03 2.46039107e-01 4.87899244e-01
7.52704442e-02 -1.39710039e-01 -6.20723128e-01 -6.52061522e-01
-8.55529785e-01 -1.15899873e+00 7.63823569e-01 5.47523975e-01
-7.75181279e-02 -9.60615218e-01 -1.51763588e-01 1.21502608e-01
7.35913634e-01 7.60798752e-01 1.07746840e+00 -7.65566289e-01
-1.90435704e-02 -7.41560578e-01 -5.43185174e-01 2.08511457e-01
3.27283531e-01 2.79222518e-01 -9.77842033e-01 -3.88203830e-01
-2.37766817e-01 -2.89240759e-02 9.64728057e-01 7.15092778e-01
8.19589734e-01 1.46043561e-02 -9.15844381e-01 4.73575294e-01
2.03451109e+00 6.99226141e-01 8.04247439e-01 7.82385409e-01
3.26361120e-01 4.41255569e-01 5.25288224e-01 3.60300094e-02
8.61736760e-02 2.63777167e-01 6.41386628e-01 -6.23473585e-01
-3.26967984e-01 3.81668061e-01 -2.47734547e-01 3.67377877e-01
-3.59528989e-01 -2.16142446e-01 -1.23757660e+00 4.45837349e-01
-1.32790267e+00 -8.70484114e-01 -4.26909268e-01 1.68006384e+00
5.35812616e-01 1.68254659e-01 -2.17387781e-01 8.21518004e-01
6.17034853e-01 -4.47454363e-01 -6.42481968e-02 -4.30425733e-01
8.73415694e-02 4.90513384e-01 6.31179929e-01 1.13480337e-01
-1.22184956e+00 2.06018746e-01 5.01367712e+00 7.62179673e-01
-1.66727638e+00 2.50503063e-01 1.16276944e+00 2.11225361e-01
4.47369546e-01 -3.46596211e-01 -7.67689049e-01 4.38718081e-01
7.71241546e-01 3.28410566e-01 -3.36876214e-01 7.14299560e-01
1.17604390e-01 -7.50189543e-01 -7.70560324e-01 7.63808310e-01
-5.98046780e-02 -1.28497171e+00 -2.39422426e-01 2.50236481e-01
6.72099113e-01 -8.47794637e-02 -1.34632289e-01 2.08342984e-01
-1.85127392e-01 -1.24655044e+00 1.10785574e-01 6.29086673e-01
8.63114178e-01 -8.61439407e-01 1.73471332e+00 1.76399499e-01
-7.46125460e-01 -2.27060214e-01 -3.11409473e-01 2.16836497e-01
-5.10252416e-01 4.93408114e-01 -1.20765567e+00 6.83904886e-01
1.00847888e+00 4.27140445e-01 -8.24256957e-01 1.14574635e+00
4.42281753e-01 6.42959177e-01 -3.31670076e-01 -4.03543949e-01
2.70782143e-01 3.96732688e-01 4.76333126e-02 1.32649398e+00
5.49238622e-01 1.79348841e-01 -1.93660215e-01 2.08801493e-01
1.85642749e-01 4.07364726e-01 -3.49591285e-01 2.35952675e-01
1.41973421e-01 1.69522703e+00 -1.10843110e+00 -4.32262212e-01
-1.84277520e-01 3.41716170e-01 -1.69366583e-01 -1.58705294e-01
-6.19349301e-01 -5.85698724e-01 1.93378702e-02 4.17277604e-01
7.86235258e-02 5.96304655e-01 -2.64628023e-01 -3.20189655e-01
-5.00073075e-01 -8.10700536e-01 5.02497137e-01 -6.64014518e-01
-9.86816347e-01 8.12552989e-01 6.17928058e-02 -1.06663358e+00
1.35046035e-01 -9.90989804e-01 -7.08576024e-01 8.51207733e-01
-1.39287233e+00 -1.46384072e+00 -8.55475962e-01 3.50551337e-01
3.89096379e-01 -3.96878541e-01 8.74730229e-01 3.47070307e-01
-5.02420604e-01 5.69489121e-01 1.71515107e-01 2.18105212e-01
6.08118117e-01 -1.21797776e+00 -6.08308017e-01 4.04771298e-01
-6.81700885e-01 2.89803684e-01 4.84647870e-01 -4.01957303e-01
-1.21915340e+00 -1.06752801e+00 7.38469064e-01 1.06505670e-01
1.49084762e-01 2.04745695e-01 -5.43873489e-01 2.88183779e-01
3.47778678e-01 3.96614850e-01 7.02331066e-01 -3.96411568e-01
2.37381548e-01 -5.41938126e-01 -1.60491335e+00 9.51363891e-02
1.37272581e-01 1.81969598e-01 1.01796035e-02 1.70419857e-01
-2.20180571e-01 -4.11974043e-01 -9.78049994e-01 8.08070064e-01
7.77917981e-01 -1.14899254e+00 5.18657207e-01 -2.57650405e-01
6.79184139e-01 -1.26203999e-01 -3.21327373e-02 -9.62003529e-01
-4.64060009e-01 3.06376994e-01 4.95277971e-01 1.11851656e+00
4.19921190e-01 -8.27254057e-01 8.65449548e-01 1.28034893e-02
-4.69686799e-02 -1.29074240e+00 -8.68949592e-01 -2.26721674e-01
2.43884716e-02 6.03043139e-02 3.21882576e-01 8.92834127e-01
-3.99533689e-01 -3.99789661e-02 1.38762206e-01 -1.07600711e-01
2.98999488e-01 -2.28042305e-01 6.34850144e-01 -1.31580222e+00
-6.84516802e-02 -4.30323750e-01 -9.98093784e-01 1.97448298e-01
-5.82056165e-01 -9.78334248e-01 -2.83702701e-01 -1.65415740e+00
7.11412013e-01 -6.19388759e-01 -4.32495803e-01 7.15458214e-01
-2.87938118e-03 6.26934469e-01 -4.52309139e-02 2.14798853e-01
1.62553474e-01 -4.20255125e-01 1.33832538e+00 -3.21769178e-01
1.41578451e-01 -3.89378630e-02 -6.86221540e-01 5.08742034e-01
9.65466142e-01 -4.76038903e-01 -1.45226881e-01 -1.80529892e-01
-1.24740995e-01 6.77328631e-02 4.84158903e-01 -1.28698921e+00
1.99392468e-01 8.73061866e-02 1.01538873e+00 -5.98859072e-01
2.19795138e-01 -9.59249437e-01 6.00078285e-01 1.16743839e+00
-7.76000619e-02 1.14040878e-02 2.85774708e-01 2.33727545e-01
-3.47732127e-01 -3.27258050e-01 1.07099175e+00 -3.23004276e-01
-6.13822222e-01 1.00462578e-01 -3.73105556e-01 -8.52250099e-01
1.47636724e+00 -6.73737586e-01 -3.91566694e-01 1.08052120e-01
-6.33533001e-01 -5.28160147e-02 8.10083598e-02 -2.46847160e-02
3.73551071e-01 -1.25517333e+00 -8.85558665e-01 9.51889381e-02
1.96172208e-01 -6.64848462e-02 4.45041150e-01 1.23099768e+00
-1.05019212e+00 5.46203256e-01 -7.15838611e-01 -7.48048723e-01
-1.79349732e+00 9.97027159e-02 6.52845740e-01 -6.46886885e-01
-1.99497566e-01 9.42753911e-01 -2.08446726e-01 6.32668734e-02
-4.51606959e-02 -2.54392445e-01 -7.78115690e-01 7.97818527e-02
2.61965126e-01 6.07683897e-01 5.41025460e-01 -6.09551668e-01
-4.90662277e-01 5.72972715e-01 -5.32347441e-01 3.01568389e-01
1.32952797e+00 4.08679456e-01 -3.07607919e-01 1.09739445e-01
1.57434583e+00 -3.19287628e-01 -2.89417475e-01 3.47118139e-01
-9.01089609e-02 -2.09475160e-01 2.76602119e-01 -1.13398302e+00
-1.21140385e+00 7.96938121e-01 1.40281808e+00 2.32687861e-01
1.28935647e+00 -2.45185837e-01 3.75165194e-01 9.50944200e-02
2.35396639e-01 -6.75232947e-01 -1.88596278e-01 -3.70339565e-02
8.02269280e-01 -1.48917568e+00 2.72288978e-01 -2.27564916e-01
-1.37486592e-01 1.52692020e+00 7.63626873e-01 -4.11899447e-01
8.30128014e-01 2.46961325e-01 2.57880777e-01 -4.33047622e-01
-7.40085781e-01 1.31431445e-01 1.16022773e-01 3.65878463e-01
7.51645863e-01 1.21458195e-01 -6.92335665e-01 3.89623761e-01
-1.45838141e-01 2.96039701e-01 3.31527025e-01 1.21309555e+00
-4.75574166e-01 -7.63835967e-01 -5.14094770e-01 9.53479707e-01
-9.86208320e-01 3.60186934e-01 -2.41145372e-01 1.35828376e+00
5.83712399e-01 8.44526649e-01 -5.45163080e-02 -4.16156173e-01
1.26079202e-01 -3.21643621e-01 3.31848621e-01 -2.12014690e-01
-9.28320646e-01 -4.72169481e-02 -8.18351060e-02 -1.52675226e-01
-5.92658758e-01 -4.38342124e-01 -1.21043253e+00 -4.38295901e-01
-6.62945092e-01 9.90171507e-02 1.13460815e+00 7.68209934e-01
5.59786782e-02 7.64385104e-01 4.31096792e-01 -4.93154168e-01
-2.55798459e-01 -1.47479260e+00 -5.16551018e-01 4.50051054e-02
3.16646427e-01 -5.90598047e-01 -1.53529271e-01 2.66167279e-02] | [15.26440715789795, -2.786406993865967] |
4cb33dcb-e00a-491e-9d75-180a7b7e4248 | double-permutation-equivariance-for-knowledge | 2302.01313 | null | https://arxiv.org/abs/2302.01313v5 | https://arxiv.org/pdf/2302.01313v5.pdf | Inductive Link Prediction for Both New Nodes and New Relation Types via Double Equivariance | Despite recent advances in relational learning, the task of inductive link prediction in discrete attributed multigraphs with both new nodes and new relation types in test remains an open problem. In this work we tackle this task by defining the concept of double exchangeability and its associated double-permutation equivariant graph neural network that are equivariant to permutations of both node identities and edge relations. Our neural architecture imposes a structural representation of relations that can inductively generalize from training nodes and relations to arbitrarily new test nodes and relations, without the need for adaptation or retraining, thus enabling a new direction in relational learning. Finally, we introduce a general blueprint for such double equivariant representations and empirically showcase its capability on two proposed real-world benchmarks that no existing works can perform accurately. | ['Bruno Ribeiro', 'Jincheng Zhou', 'Yangze Zhou', 'Jianfei Gao'] | 2023-02-02 | null | null | null | null | ['inductive-link-prediction', 'relational-reasoning', 'logical-reasoning'] | ['graphs', 'natural-language-processing', 'reasoning'] | [ 4.60096985e-01 7.71523237e-01 -5.74335694e-01 -4.89661932e-01
-7.41878152e-02 -6.72289491e-01 7.31969774e-01 2.14625373e-01
4.03928086e-02 1.18643034e+00 -1.54564261e-01 -6.73787355e-01
-7.40643859e-01 -1.41696143e+00 -1.03087771e+00 -3.72156084e-01
-8.92385483e-01 1.02218568e+00 3.39111537e-01 -4.25562471e-01
-6.01172447e-01 6.79169118e-01 -1.27426064e+00 2.76432727e-02
5.89360774e-01 5.49348891e-01 -6.19892001e-01 7.11679220e-01
2.50168681e-01 1.00426364e+00 -3.39923620e-01 -8.77405345e-01
4.25777972e-01 -2.43177637e-01 -1.36216497e+00 -1.67854413e-01
6.21388495e-01 3.66596915e-02 -8.69499087e-01 8.33216488e-01
1.37681112e-01 2.76883125e-01 8.68081450e-01 -1.57930815e+00
-1.09071040e+00 1.19204950e+00 -2.83153623e-01 2.29824558e-01
3.43678087e-01 -4.43217427e-01 1.64845133e+00 -5.27378678e-01
9.83457685e-01 1.10461092e+00 7.76270628e-01 3.13552827e-01
-1.75200391e+00 -6.31394744e-01 4.22909819e-02 5.07074356e-01
-1.43560457e+00 -7.15385750e-02 6.69525087e-01 -2.22187966e-01
8.93649042e-01 3.98787349e-01 5.50049603e-01 1.13831139e+00
3.75244580e-02 3.05305511e-01 9.28195298e-01 -1.74811125e-01
-1.80081919e-01 1.71118658e-02 3.33042830e-01 8.78633499e-01
5.68969965e-01 -5.00905775e-02 -2.26017445e-01 9.93726179e-02
6.63916588e-01 -2.04449192e-01 -1.58935472e-01 -9.03669477e-01
-9.39369082e-01 6.40926301e-01 9.80553865e-01 2.77366579e-01
2.90174987e-02 1.21438190e-01 3.98884922e-01 8.72253001e-01
4.07285303e-01 5.71815312e-01 -7.36973643e-01 6.25977337e-01
-2.33959094e-01 -1.25905961e-01 1.26794219e+00 1.18390739e+00
9.18127060e-01 -7.69484974e-03 -4.47082855e-02 6.88439429e-01
-2.01382507e-02 5.53012565e-02 9.21104662e-03 -4.88859057e-01
4.32625502e-01 1.04823983e+00 -7.12055683e-01 -1.09245801e+00
-6.52197182e-01 -8.16406190e-01 -1.10253942e+00 -1.02134414e-01
1.87785625e-01 1.68695077e-01 -7.61132479e-01 1.90526664e+00
2.85330147e-01 5.24419725e-01 2.12425083e-01 1.79095447e-01
1.10007787e+00 1.57208532e-01 -1.26681745e-01 -2.26693541e-01
8.49523544e-01 -7.45345294e-01 -5.11434853e-01 1.84865445e-01
6.52102053e-01 -7.14950338e-02 6.98983133e-01 -2.06891764e-02
-1.09890687e+00 -4.26330209e-01 -1.27924430e+00 -4.22429740e-02
-7.89329648e-01 -6.61111712e-01 1.22891951e+00 5.29558301e-01
-1.33532059e+00 7.50918686e-01 -3.91143739e-01 -3.41594756e-01
3.97679031e-01 9.76419449e-01 -9.38152194e-01 2.87859850e-02
-1.57155585e+00 9.04553413e-01 7.52059281e-01 2.19762802e-01
-7.54008114e-01 -8.36521089e-01 -1.02345610e+00 1.19990110e-01
7.95184076e-01 -7.02740848e-01 7.62812316e-01 -5.67730725e-01
-1.15130234e+00 1.09119904e+00 3.49442661e-01 -4.95628417e-01
3.54020387e-01 2.72286624e-01 -8.16471994e-01 -9.65537280e-02
-9.64898169e-02 3.44886363e-01 4.65727866e-01 -1.41320503e+00
-4.15465236e-01 -2.35300735e-01 6.05752349e-01 -2.27343966e-03
-5.70855796e-01 -5.78431368e-01 -2.72545755e-01 -5.34784257e-01
2.32940868e-01 -1.01322842e+00 -2.18192358e-02 -4.59901631e-01
-9.32558119e-01 -6.50320649e-01 6.55529559e-01 -5.61232828e-02
9.29394186e-01 -1.62789118e+00 5.99614918e-01 6.90783024e-01
6.99806988e-01 3.01392050e-04 -3.69464487e-01 3.17407280e-01
-7.48833477e-01 3.09553474e-01 -8.63830671e-02 1.34781107e-01
3.60754877e-03 5.61228693e-01 -1.01679482e-01 4.08267647e-01
3.80089015e-01 1.08748555e+00 -8.94138575e-01 -5.61273336e-01
-1.57817662e-01 2.09036544e-01 -4.83388633e-01 1.53362289e-01
-1.56290829e-01 2.94911444e-01 -6.21738695e-02 5.92570543e-01
6.09928191e-01 -4.58687425e-01 9.05153275e-01 -3.30005735e-01
6.08505785e-01 2.75927991e-01 -1.13963056e+00 1.25776386e+00
-4.27214295e-01 5.39744914e-01 -2.42043972e-01 -1.35172963e+00
1.02300251e+00 1.28503531e-01 5.66779077e-01 -4.48985457e-01
4.26411554e-02 -1.10219486e-01 4.03399229e-01 -3.68774623e-01
3.53974730e-01 9.27297957e-03 -6.61201626e-02 2.33438566e-01
4.64189321e-01 8.76261368e-02 4.00125176e-01 4.44248736e-01
1.48320317e+00 4.07014936e-02 2.13262305e-01 -1.90640211e-01
7.56137967e-01 -3.27400148e-01 4.36981440e-01 6.72409952e-01
3.68483752e-01 -2.29537953e-02 1.04258883e+00 -4.98510778e-01
-6.86076701e-01 -1.59795058e+00 -3.25009942e-01 1.27845049e+00
1.00673795e-01 -3.73656780e-01 -1.06613904e-01 -1.25993383e+00
2.20884100e-01 2.50243247e-01 -9.93432164e-01 -5.19305050e-01
-7.57543981e-01 -6.87136889e-01 6.24432445e-01 6.25153005e-01
1.36463195e-01 -1.05280387e+00 4.92098033e-01 -5.91465607e-02
2.17204422e-01 -1.45497465e+00 -1.71202924e-02 5.75684667e-01
-9.16262031e-01 -1.59535956e+00 -6.33126795e-02 -1.13458705e+00
8.80461335e-01 -2.58089960e-01 1.62171423e+00 2.59759933e-01
-1.56152621e-01 3.30070794e-01 -1.01844467e-01 1.44895554e-01
-5.39258420e-01 7.25218117e-01 1.05810985e-01 2.03331606e-03
-3.30610014e-02 -1.19575095e+00 5.59584657e-03 2.93655843e-01
-8.83972168e-01 -9.56952497e-02 6.18065476e-01 8.56791615e-01
4.99892682e-01 1.20401233e-01 7.67971992e-01 -1.74106526e+00
5.23822904e-01 -7.59692430e-01 -5.34246445e-01 5.18560410e-01
-7.51384258e-01 4.50811803e-01 6.23127043e-01 -3.04270893e-01
-6.57274842e-01 -1.05111532e-01 1.26115903e-01 -1.01083197e-01
9.13359001e-02 8.56729686e-01 -3.26974452e-01 -2.48932034e-01
7.05432832e-01 -2.73903340e-01 -3.41904342e-01 -1.19368836e-01
6.47065818e-01 -1.49980441e-01 7.89768338e-01 -6.94473982e-01
1.31133664e+00 1.91932186e-01 1.12474322e+00 -3.36254925e-01
-7.63351858e-01 1.21634126e-01 -1.21791708e+00 1.06938846e-01
5.28246999e-01 -5.87480247e-01 -1.14084041e+00 -1.19240746e-01
-7.79256821e-01 -3.98449987e-01 -3.58855903e-01 1.30172893e-01
-3.56201261e-01 2.24987060e-01 -7.10436523e-01 -1.29240081e-01
5.45025170e-02 -7.78422773e-01 2.48045936e-01 -8.40452239e-02
-1.40936837e-01 -1.47204912e+00 1.28945187e-01 3.21707278e-02
6.06119111e-02 5.42307258e-01 1.58024776e+00 -1.06375086e+00
-8.79721880e-01 -3.23860407e-01 -4.01851743e-01 1.06648237e-01
1.75857469e-01 -9.45684463e-02 -5.68581164e-01 -2.68327683e-01
-6.98315084e-01 -4.93237048e-01 8.64897490e-01 -2.05404133e-01
1.09561074e+00 -2.26209015e-01 -5.66439867e-01 8.61670136e-01
1.26589692e+00 -1.64496407e-01 3.54200006e-01 1.76117241e-01
1.16496241e+00 6.76807463e-01 8.57071427e-04 -1.77889451e-01
7.37203062e-01 6.15582585e-01 7.03254580e-01 -8.58398154e-02
-3.10670495e-01 -2.41273329e-01 -6.55913875e-02 7.70159960e-01
-5.65863848e-01 -3.25828403e-01 -8.22896898e-01 3.58796328e-01
-1.68362665e+00 -5.76663971e-01 -4.82834250e-01 1.96972895e+00
9.72305417e-01 2.52277613e-01 -8.46188962e-02 1.89095780e-01
7.50560880e-01 1.36904776e-01 -4.41824883e-01 -4.08805370e-01
-2.26597011e-01 8.38056505e-01 4.54323709e-01 5.97257495e-01
-1.11155605e+00 9.03643310e-01 6.35101795e+00 3.22802544e-01
-5.70787609e-01 3.53429727e-02 4.02647585e-01 4.06025648e-01
-6.06038690e-01 3.21879089e-02 -5.37464023e-01 -3.55607867e-01
8.90149891e-01 -5.51764444e-02 5.79820931e-01 5.68108201e-01
-7.65632212e-01 4.79207575e-01 -1.82120168e+00 5.29621661e-01
2.72751655e-02 -1.35577416e+00 2.48581946e-01 2.00151894e-02
9.98289049e-01 -2.20759422e-01 1.48799857e-02 7.90218890e-01
8.55347097e-01 -1.37599719e+00 -5.35291769e-02 5.94295919e-01
1.00813949e+00 -7.90688276e-01 6.80520713e-01 -2.67681718e-01
-1.37947333e+00 2.07706913e-02 -1.57328501e-01 -8.02007318e-02
-3.22643161e-01 1.92772686e-01 -1.17556620e+00 1.15991831e+00
4.30769444e-01 1.04566050e+00 -1.00033569e+00 6.13006592e-01
-4.79269236e-01 4.19568509e-01 -5.50006814e-02 1.84381366e-01
-1.74425736e-01 -4.39200662e-02 3.64458472e-01 9.62323248e-01
8.19188543e-03 -1.69293195e-01 1.91202328e-01 6.59050405e-01
-8.90208185e-01 -1.82543248e-01 -9.12776232e-01 1.59863144e-01
4.12891626e-01 1.48007965e+00 -8.27323318e-01 -4.15605679e-02
-3.41076970e-01 6.56697452e-01 9.89753425e-01 4.58356291e-01
-7.31588662e-01 -1.33536667e-01 4.17165786e-01 1.12922423e-01
1.35313734e-01 -7.68427700e-02 1.52000049e-02 -1.14038980e+00
-8.88995603e-02 -4.65733528e-01 8.75283003e-01 -3.95516813e-01
-1.50089645e+00 7.90838003e-01 1.05607465e-01 -8.31126332e-01
-2.49568835e-01 -8.05438936e-01 -4.19080555e-01 4.57898319e-01
-1.40104413e+00 -1.45013905e+00 -2.78200358e-01 7.84968674e-01
-2.57918775e-01 -4.72861141e-01 1.14902580e+00 4.58343804e-01
-5.26830494e-01 9.69842017e-01 -2.38590837e-01 4.41050887e-01
6.99940145e-01 -1.64551437e+00 4.03806418e-01 4.84981984e-01
4.39849764e-01 5.26621938e-01 5.05210161e-01 -5.50171375e-01
-1.53861487e+00 -1.21556747e+00 7.00755060e-01 -6.78133786e-01
1.05060601e+00 -5.80956280e-01 -9.51347232e-01 1.40066040e+00
1.21187888e-01 6.31547213e-01 5.43536484e-01 9.20214772e-01
-7.03237593e-01 -5.43521643e-01 -8.23383391e-01 6.05392039e-01
1.77018893e+00 -5.54746747e-01 -3.83015841e-01 4.62311745e-01
8.64690185e-01 -3.70267242e-01 -1.37462306e+00 1.12011158e+00
3.94955814e-01 -7.16449857e-01 1.23970425e+00 -1.10120261e+00
3.99471432e-01 -1.20965250e-01 7.68312961e-02 -1.32637227e+00
-5.66300333e-01 -3.98703039e-01 -4.55722570e-01 1.35533524e+00
7.67756879e-01 -8.43945861e-01 1.02034831e+00 1.50573507e-01
-1.91201478e-01 -8.72449815e-01 -8.33773017e-01 -7.51528084e-01
3.46354991e-02 -1.96321428e-01 5.99096119e-01 1.51678455e+00
-4.46778126e-02 8.88080657e-01 -2.75772244e-01 2.51043528e-01
4.89314586e-01 4.84755486e-02 8.04416060e-01 -1.71646440e+00
-3.89773518e-01 -4.58423227e-01 -8.41261625e-01 -5.09319127e-01
9.45055664e-01 -1.86139870e+00 -4.92519140e-01 -1.50442076e+00
2.44640946e-01 -7.19731987e-01 -4.61278975e-01 6.33550644e-01
-6.88668266e-02 4.32330042e-01 -2.79439211e-01 -2.06408530e-01
-7.23791897e-01 4.07125324e-01 1.27418149e+00 -4.89550352e-01
-3.80612351e-02 1.58687562e-01 -7.05364466e-01 5.23358285e-01
5.14529169e-01 -5.31318843e-01 -8.65698457e-01 -5.74552193e-02
7.93042183e-01 1.88290793e-02 4.32315081e-01 -7.19543219e-01
2.56172657e-01 5.15485331e-02 3.95709157e-01 -2.35249832e-01
1.96789056e-01 -1.03560293e+00 3.40863585e-01 3.23824495e-01
-5.65340638e-01 1.70055494e-01 5.19417925e-03 7.69420683e-01
2.46366896e-02 -1.30590558e-01 3.66776645e-01 2.13467196e-01
-6.93345189e-01 5.95010936e-01 2.08096474e-01 2.50579268e-01
1.20590270e+00 5.57911806e-02 -5.39665103e-01 -2.69118726e-01
-1.29348898e+00 3.19596320e-01 -2.32452992e-02 5.36880314e-01
5.07854044e-01 -1.41994870e+00 -8.48839581e-01 1.81270033e-01
1.31372184e-01 2.72684786e-02 -2.84351930e-02 7.16223955e-01
-2.64842510e-01 3.44154201e-02 -3.52416456e-01 -3.83181334e-01
-1.21083045e+00 7.67910063e-01 4.11757469e-01 -7.40714550e-01
-4.44473475e-01 8.57119501e-01 5.37946075e-02 -9.79038358e-01
2.15096414e-01 -1.33334547e-01 -4.68516648e-01 -7.75611540e-03
-3.32089275e-01 3.22863430e-01 1.85230106e-01 -5.79734683e-01
-2.34936208e-01 2.45906428e-01 -2.41660208e-01 4.90006775e-01
1.36695528e+00 2.70791113e-01 -6.39461160e-01 5.38999200e-01
1.17200637e+00 1.04446020e-02 -5.86679161e-01 -4.34110284e-01
3.89410764e-01 -2.44545788e-02 -6.84532106e-01 -6.15393698e-01
-1.13252616e+00 3.84435445e-01 7.62040392e-02 7.87351966e-01
7.98489630e-01 4.78440225e-01 1.74392074e-01 9.42466438e-01
1.64297044e-01 -4.48785663e-01 2.03422278e-01 8.23902786e-01
8.46162260e-01 -1.07125580e+00 2.11618021e-01 -8.81859720e-01
-8.09142962e-02 1.13880444e+00 9.33064461e-01 -3.18686008e-01
9.02378201e-01 2.34786510e-01 -5.28106809e-01 -2.93985307e-01
-7.21392691e-01 -3.50132525e-01 6.87684298e-01 9.64740336e-01
5.11799276e-01 2.61492640e-01 -1.56180114e-02 4.14151192e-01
-3.45534772e-01 -4.54773158e-01 4.45179135e-01 3.43654186e-01
2.56207108e-01 -1.46545768e+00 2.64822334e-01 7.70164371e-01
-2.53528029e-01 -1.08761415e-01 -6.38141751e-01 1.33149803e+00
1.44536600e-01 5.49318671e-01 5.92307933e-02 -6.80177689e-01
5.23911953e-01 6.94118673e-03 8.58085155e-01 -9.09656286e-01
-3.26097637e-01 -7.57072091e-01 4.66779083e-01 -3.00640725e-02
-5.15885413e-01 -3.68034422e-01 -1.23232758e+00 -4.06095237e-01
-3.49123389e-01 8.25911108e-03 -1.57064553e-02 9.24891472e-01
1.06936589e-01 1.17544878e+00 5.67451656e-01 -3.32051665e-01
-1.42589569e-01 -9.72569227e-01 -6.33758128e-01 5.69414496e-01
2.76093781e-01 -7.74804294e-01 -7.15734735e-02 -1.26379088e-01] | [7.116056442260742, 6.351359844207764] |
1541ed69-768a-4633-ac6b-252300683426 | tcgan-semantic-aware-and-structure-preserved | 2302.08047 | null | https://arxiv.org/abs/2302.08047v1 | https://arxiv.org/pdf/2302.08047v1.pdf | TcGAN: Semantic-Aware and Structure-Preserved GANs with Individual Vision Transformer for Fast Arbitrary One-Shot Image Generation | One-shot image generation (OSG) with generative adversarial networks that learn from the internal patches of a given image has attracted world wide attention. In recent studies, scholars have primarily focused on extracting features of images from probabilistically distributed inputs with pure convolutional neural networks (CNNs). However, it is quite difficult for CNNs with limited receptive domain to extract and maintain the global structural information. Therefore, in this paper, we propose a novel structure-preserved method TcGAN with individual vision transformer to overcome the shortcomings of the existing one-shot image generation methods. Specifically, TcGAN preserves global structure of an image during training to be compatible with local details while maintaining the integrity of semantic-aware information by exploiting the powerful long-range dependencies modeling capability of the transformer. We also propose a new scaling formula having scale-invariance during the calculation period, which effectively improves the generated image quality of the OSG model on image super-resolution tasks. We present the design of the TcGAN converter framework, comprehensive experimental as well as ablation studies demonstrating the ability of TcGAN to achieve arbitrary image generation with the fastest running time. Lastly, TcGAN achieves the most excellent performance in terms of applying it to other image processing tasks, e.g., super-resolution as well as image harmonization, the results further prove its superiority. | ['Danfeng Sun', 'Yong liu', 'Xiongtao Zhang', 'Lili Yan', 'Yunliang Jiang'] | 2023-02-16 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 4.56260085e-01 1.65681452e-01 1.47393290e-02 -1.53837904e-01
-6.47045434e-01 -2.68382430e-01 4.74842280e-01 -8.87916446e-01
-3.72640118e-02 9.09463346e-01 1.85487956e-01 2.37774730e-01
-2.12967992e-01 -1.15708685e+00 -9.36143458e-01 -1.09428000e+00
6.06579423e-01 -2.78715730e-01 2.37906843e-01 -5.46659768e-01
-6.66414201e-02 4.68916476e-01 -1.53342628e+00 1.76822037e-01
1.03204715e+00 9.39208984e-01 3.50551575e-01 3.20179731e-01
-6.20819209e-03 8.46679926e-01 -6.88395798e-01 -4.16039854e-01
3.97023201e-01 -7.41098762e-01 -4.19819176e-01 -1.11492798e-01
4.73042190e-01 -3.78154606e-01 -6.74865127e-01 1.25517046e+00
7.70685554e-01 -1.49844391e-02 3.73192817e-01 -1.15165818e+00
-1.27075911e+00 6.32069588e-01 -6.61342263e-01 1.07547261e-01
-2.05577716e-01 2.69892037e-01 6.27622783e-01 -6.72215104e-01
6.76486254e-01 1.15686738e+00 4.53709424e-01 8.88653159e-01
-1.13295281e+00 -9.49964583e-01 -7.55839199e-02 1.90152958e-01
-1.30020106e+00 -4.10416484e-01 1.07558191e+00 3.12310960e-02
6.10150635e-01 1.49815515e-01 4.28933799e-01 1.10975957e+00
4.19947684e-01 3.22817862e-01 1.32893157e+00 -2.74543256e-01
-5.45342863e-02 1.66042503e-02 -5.23549497e-01 6.87512875e-01
1.80857837e-01 4.51039433e-01 -7.12026715e-01 3.33657235e-01
1.18022835e+00 7.62423649e-02 -3.75205785e-01 -1.46719173e-01
-1.10939968e+00 6.09051108e-01 9.89763319e-01 4.88575935e-01
-3.72627050e-01 2.46126711e-01 7.69052804e-02 1.40017509e-01
3.43260139e-01 2.51626253e-01 -2.46759737e-03 4.23343271e-01
-8.54622900e-01 7.03582913e-02 1.99018389e-01 1.14998019e+00
7.20186532e-01 6.91095054e-01 -4.34864938e-01 6.57826722e-01
-1.73026815e-01 6.41381502e-01 5.73813081e-01 -7.94454813e-01
1.37333691e-01 4.12927032e-01 -1.20150529e-01 -1.12840629e+00
8.80594086e-03 -6.79653823e-01 -1.57815194e+00 4.06596869e-01
-2.08558198e-02 -1.28877655e-01 -1.01134133e+00 2.05255795e+00
1.28041610e-01 3.25018972e-01 3.39786112e-01 8.73859465e-01
1.11344934e+00 8.30513000e-01 -3.43948454e-02 -3.00130963e-01
1.34340167e+00 -8.08253825e-01 -7.98032284e-01 -2.01374382e-01
-3.46694499e-01 -7.27223098e-01 8.19119275e-01 1.58074182e-02
-1.10864556e+00 -9.81633186e-01 -1.29409111e+00 -1.48861095e-01
-1.51543140e-01 -1.88959041e-03 4.61265713e-01 5.07155478e-01
-1.05625677e+00 4.59146827e-01 -5.70525885e-01 3.25154588e-02
6.24077916e-01 2.31474444e-01 -4.01154786e-01 -1.11790717e-01
-1.45498872e+00 5.61521828e-01 5.60850680e-01 1.54612521e-02
-8.26468170e-01 -7.35951841e-01 -7.77636170e-01 1.34413168e-01
1.17850840e-01 -8.40717673e-01 7.81941950e-01 -1.17612636e+00
-1.63886034e+00 4.86964673e-01 -5.64474165e-02 -4.15914387e-01
2.40046218e-01 1.83370620e-01 -4.39206481e-01 1.57621369e-01
1.73351526e-01 7.15865254e-01 1.22280419e+00 -1.23911107e+00
-4.77093786e-01 -3.65539402e-01 -3.26693021e-02 2.30747983e-01
-4.61969584e-01 -9.54934582e-02 -2.26870224e-01 -9.69066322e-01
1.10099621e-01 -6.83956146e-01 -1.60097823e-01 -1.21679060e-01
-3.17574263e-01 1.63873568e-01 1.02219391e+00 -6.88098729e-01
8.13570976e-01 -2.22098374e+00 1.67475328e-01 -1.63128227e-01
2.01631054e-01 4.07485694e-01 -1.94055855e-01 1.54651999e-01
-1.34409472e-01 1.31568834e-02 -2.58066624e-01 2.58584060e-02
-2.98501849e-01 9.84337702e-02 -6.58063293e-01 1.75755441e-01
4.65248644e-01 1.35842407e+00 -6.02028430e-01 -5.27658641e-01
2.08689973e-01 7.93629229e-01 -3.11724901e-01 2.81458974e-01
-2.38129105e-02 6.77376926e-01 -4.89803493e-01 4.10714328e-01
8.73206496e-01 -2.72654980e-01 1.69678088e-02 -6.43145323e-01
-4.01241779e-02 -3.47481966e-01 -8.65367830e-01 1.66812479e+00
-5.44798195e-01 3.51410776e-01 -1.53483674e-01 -8.74923766e-01
1.07820857e+00 2.67594546e-01 2.71610409e-01 -1.07097280e+00
-9.36824922e-03 7.01842755e-02 -2.10030496e-01 -1.89701214e-01
3.99677843e-01 -4.59730059e-01 -6.22584261e-02 2.30560362e-01
2.62654573e-01 -1.46231115e-01 -2.42395818e-01 1.83765858e-01
7.44520545e-01 2.39686310e-01 2.58091897e-01 -1.61465332e-01
4.07692462e-01 -2.16510907e-01 6.27534151e-01 5.84218085e-01
6.22492805e-02 1.00668097e+00 1.55523330e-01 -2.86123097e-01
-1.38812375e+00 -1.04702246e+00 9.97584090e-02 6.73647046e-01
4.46123838e-01 -9.06775333e-03 -1.00479841e+00 -2.31188968e-01
-5.93396068e-01 4.70359623e-01 -5.45996249e-01 -4.93314832e-01
-6.68728173e-01 -7.95180678e-01 7.21512735e-01 5.70573092e-01
1.34319830e+00 -1.14954603e+00 -5.13114095e-01 2.42735222e-01
-3.83132309e-01 -1.27828598e+00 -4.40749645e-01 -2.01978996e-01
-6.14786386e-01 -7.47113228e-01 -8.49557400e-01 -1.04166794e+00
5.87565243e-01 4.22680646e-01 8.03741455e-01 -1.32768482e-01
-4.49017197e-01 -7.79695213e-02 -1.96990073e-01 -1.77026376e-01
-3.65333706e-01 4.64428104e-02 -2.54762590e-01 2.55596459e-01
-1.65522546e-01 -9.01562631e-01 -7.41553724e-01 2.48523816e-01
-1.15237534e+00 4.10190523e-01 9.23236609e-01 1.08788490e+00
8.03455293e-01 6.75714970e-01 7.52240598e-01 -8.04955840e-01
3.25860411e-01 -1.62814140e-01 -4.63264376e-01 2.52004504e-01
-5.16881704e-01 3.13766040e-02 9.53278184e-01 -4.37517673e-01
-1.55495250e+00 -8.35587531e-02 3.41142453e-02 -6.24178469e-01
1.17492780e-01 1.40062636e-02 -5.68693876e-01 -3.09227884e-01
6.82740390e-01 8.36301386e-01 2.11088017e-01 -1.00133352e-01
5.45898855e-01 2.99536198e-01 9.07204628e-01 -4.03955489e-01
1.21142459e+00 6.81661367e-01 2.15067193e-01 -6.48855090e-01
-8.82107973e-01 2.17930391e-01 -2.40672857e-01 5.94576471e-04
9.86996651e-01 -1.14277887e+00 -4.16205078e-01 9.12561238e-01
-1.02755725e+00 -3.37837003e-02 -4.24589485e-01 1.94385961e-01
-6.09979093e-01 8.22103247e-02 -6.12581968e-01 -3.58401299e-01
-6.56333625e-01 -9.92193282e-01 8.22119594e-01 7.07008421e-01
5.27471840e-01 -6.96308911e-01 -2.77157366e-01 2.30905235e-01
9.23310935e-01 6.27060235e-01 8.89143705e-01 6.91271350e-02
-1.01059628e+00 1.85469598e-01 -5.42945325e-01 6.25037849e-01
2.89409041e-01 -1.62531465e-01 -1.05479717e+00 -3.69150549e-01
2.84833759e-01 -2.13452429e-01 8.63938928e-01 4.31582481e-01
1.11434710e+00 -4.80892032e-01 -1.12423308e-01 9.70054448e-01
1.78558004e+00 1.97996989e-01 1.18156910e+00 1.25180528e-01
8.37979972e-01 3.91879350e-01 4.29449379e-01 1.12993166e-01
-9.06849951e-02 6.03443563e-01 5.38528621e-01 -3.76045555e-01
-6.33950770e-01 -5.59918284e-01 2.37552464e-01 6.23670697e-01
-2.73987442e-01 -3.27368408e-01 -2.15660945e-01 3.62997562e-01
-1.58484411e+00 -1.22050714e+00 3.07619035e-01 1.87744796e+00
8.34586918e-01 -2.18100920e-02 -5.18179655e-01 -8.30012336e-02
1.02054369e+00 5.43474138e-01 -7.25164413e-01 1.24283716e-01
-5.27381778e-01 6.08804584e-01 5.61415315e-01 1.42953783e-01
-9.45214629e-01 1.10460365e+00 5.91242361e+00 1.27404654e+00
-1.24951804e+00 1.98562905e-01 8.39975595e-01 5.15952110e-02
-3.09893817e-01 -5.87795153e-02 -6.33911252e-01 5.21039546e-01
5.54001331e-01 -3.55615616e-01 5.86076379e-01 7.22948790e-01
-1.05743343e-02 3.51650834e-01 -3.94026637e-01 1.02152383e+00
2.74233907e-01 -1.50924993e+00 4.29102778e-01 -7.03182966e-02
1.10428774e+00 -3.39261115e-01 4.69053924e-01 9.82523896e-03
1.30970642e-01 -1.25212240e+00 5.73469162e-01 6.24646246e-01
1.47303748e+00 -1.01589453e+00 5.80494046e-01 2.12570146e-01
-1.10065901e+00 -4.49882150e-02 -6.23883784e-01 2.45394453e-01
2.39438843e-02 5.75595200e-01 -3.56370538e-01 8.97129893e-01
7.09886432e-01 6.81855917e-01 -4.56493825e-01 4.29533541e-01
-4.00104940e-01 3.96341383e-01 3.35227847e-02 3.59149992e-01
1.14089608e-01 -1.64649934e-01 5.37989020e-01 5.72243273e-01
6.09923542e-01 3.47956449e-01 -3.33624393e-01 1.33304024e+00
-2.47217402e-01 -1.61288410e-01 -7.39950180e-01 1.18635401e-01
4.50792938e-01 1.36992657e+00 -6.03056431e-01 -2.07268208e-01
-1.79620087e-01 1.02429426e+00 1.52125895e-01 4.10953343e-01
-1.04477596e+00 -6.23640478e-01 4.92245138e-01 1.12028271e-01
4.28348541e-01 3.36272418e-02 -1.37856007e-01 -1.07459581e+00
1.29766643e-01 -9.11095500e-01 -2.45988555e-02 -9.89182234e-01
-1.08223641e+00 9.67218876e-01 -1.60398155e-01 -1.35036218e+00
-2.10130036e-01 -2.56212831e-01 -7.04312027e-01 9.88494992e-01
-1.57117343e+00 -1.56957567e+00 -7.85523295e-01 8.96322966e-01
4.16360557e-01 -3.98902744e-01 6.25603020e-01 1.08262226e-01
-3.58625025e-01 6.59368217e-01 8.52554068e-02 2.22890556e-01
6.56924605e-01 -8.18692386e-01 4.41569328e-01 1.16069961e+00
-6.41920231e-03 5.77592134e-01 4.44520503e-01 -5.15871704e-01
-1.32891834e+00 -1.47097540e+00 3.23505402e-01 1.13639280e-01
1.48113146e-01 -1.36476621e-01 -8.80091012e-01 4.73186970e-01
4.07284588e-01 1.46582246e-01 1.18244871e-01 -5.62947214e-01
-5.35600305e-01 -5.07315397e-01 -1.30478263e+00 4.96970206e-01
1.18591750e+00 -5.40633976e-01 -3.13727349e-01 -6.56422460e-03
1.03276181e+00 -4.72448379e-01 -7.99457073e-01 5.62550902e-01
3.59784544e-01 -1.12330759e+00 1.11879849e+00 -3.40728015e-01
8.21111023e-01 -4.82063770e-01 -6.64906427e-02 -1.28862095e+00
-6.11348689e-01 -6.60644531e-01 2.70212263e-01 1.48636353e+00
-3.73985842e-02 -8.44448745e-01 4.60924685e-01 1.82828575e-01
-8.72855932e-02 -4.33002025e-01 -9.01648402e-01 -6.98866367e-01
3.77231985e-02 2.71799833e-01 9.15719211e-01 6.72459841e-01
-7.26349652e-01 3.46566141e-01 -6.81164682e-01 2.48561591e-01
9.57725346e-01 3.02279115e-01 6.94812179e-01 -7.74386585e-01
-2.53185987e-01 -1.71292081e-01 -5.15962064e-01 -6.26698852e-01
1.32084697e-01 -6.27614737e-01 -7.03018112e-03 -1.40840077e+00
4.10391867e-01 -2.66782790e-01 -3.84786844e-01 3.90901715e-01
-1.73815340e-01 7.64172316e-01 8.19073468e-02 2.31136531e-01
-2.46435970e-01 7.70204246e-01 1.75115168e+00 -1.76435769e-01
1.57476917e-01 -2.31238753e-01 -1.04615784e+00 3.69043916e-01
8.90711248e-01 -2.63971001e-01 -7.42793620e-01 -3.22729558e-01
6.30857307e-04 5.18780835e-02 6.91969752e-01 -1.15064454e+00
1.75730392e-01 -1.84808582e-01 6.21058583e-01 -2.07734719e-01
1.97322398e-01 -5.45932531e-01 6.53779447e-01 3.70698035e-01
-1.31995916e-01 -3.65791857e-01 1.15314253e-01 6.28737450e-01
-4.72436488e-01 3.67182344e-02 1.29553246e+00 -3.15597773e-01
-8.39644194e-01 5.45925677e-01 2.80616254e-01 4.57825735e-02
1.08637583e+00 -1.33766070e-01 -6.23214126e-01 -2.85408109e-01
-3.78704607e-01 -3.75838697e-01 5.95351219e-01 5.58202922e-01
8.42632413e-01 -1.54017496e+00 -8.29230130e-01 4.29596424e-01
-4.76944298e-02 2.22759172e-01 8.04717302e-01 3.33468020e-01
-4.10983920e-01 2.51538128e-01 -7.19284534e-01 -3.20284635e-01
-9.83341694e-01 8.06738198e-01 4.55521703e-01 -1.16487533e-01
-8.55228782e-01 6.59902453e-01 7.39652753e-01 2.55194660e-02
-3.31897497e-01 1.07969984e-01 -1.61964402e-01 -4.05583829e-01
7.02186048e-01 5.97877353e-02 -1.19484797e-01 -6.56392395e-01
-2.87065078e-02 7.49761879e-01 -6.99549392e-02 7.47042373e-02
1.28452659e+00 -1.34245262e-01 -2.81401128e-01 -1.04895480e-01
1.02670121e+00 -3.26549001e-02 -1.51347995e+00 -3.95929962e-01
-7.33685553e-01 -5.08485377e-01 1.29193082e-01 -6.06891215e-01
-1.67736948e+00 6.83111668e-01 6.64861858e-01 -8.51909965e-02
1.45602679e+00 -1.14737622e-01 1.01543975e+00 -1.69623598e-01
5.90297639e-01 -7.99783945e-01 3.25459927e-01 1.42354770e-02
9.23868537e-01 -9.97166991e-01 -1.80436060e-01 -4.26794857e-01
-5.80947161e-01 9.02304411e-01 7.59802401e-01 -3.75729471e-01
4.12533969e-01 2.83401400e-01 -9.36383158e-02 -3.80523759e-03
-6.01084590e-01 -1.22418828e-01 1.98621914e-01 7.95384705e-01
-4.61097062e-02 -7.16026127e-03 -9.87423584e-02 5.60370326e-01
-3.22826862e-01 2.99966130e-02 4.35662895e-01 5.16218305e-01
-2.80977219e-01 -8.56899500e-01 -1.70103729e-01 9.86381695e-02
-5.44695377e-01 -2.45138168e-01 2.11916286e-02 7.31023610e-01
2.79215217e-01 7.34247029e-01 -1.55343693e-02 -4.66232806e-01
1.59214288e-01 -3.34236354e-01 5.76567829e-01 -3.81216347e-01
-2.43379563e-01 -2.87599340e-02 -4.42449033e-01 -6.09986484e-01
-6.53967917e-01 -1.00190699e-01 -1.07971537e+00 -3.22248429e-01
-2.42302254e-01 -1.06078081e-01 4.19743657e-01 6.18095398e-01
6.08242929e-01 8.91702652e-01 7.90745914e-01 -6.26325965e-01
-4.58943307e-01 -8.67846727e-01 -6.80266559e-01 3.58610600e-01
1.49048284e-01 -4.87945616e-01 -2.79096127e-01 2.52819180e-01] | [11.53748893737793, -1.1051793098449707] |
c37c6a0d-a1e2-4ff8-945c-7b0c5b8cd340 | surrogate-assisted-multi-objective-neural | 2208.06820 | null | https://arxiv.org/abs/2208.06820v1 | https://arxiv.org/pdf/2208.06820v1.pdf | Surrogate-assisted Multi-objective Neural Architecture Search for Real-time Semantic Segmentation | The architectural advancements in deep neural networks have led to remarkable leap-forwards across a broad array of computer vision tasks. Instead of relying on human expertise, neural architecture search (NAS) has emerged as a promising avenue toward automating the design of architectures. While recent achievements in image classification have suggested opportunities, the promises of NAS have yet to be thoroughly assessed on more challenging tasks of semantic segmentation. The main challenges of applying NAS to semantic segmentation arise from two aspects: (i) high-resolution images to be processed; (ii) additional requirement of real-time inference speed (i.e., real-time semantic segmentation) for applications such as autonomous driving. To meet such challenges, we propose a surrogate-assisted multi-objective method in this paper. Through a series of customized prediction models, our method effectively transforms the original NAS task into an ordinary multi-objective optimization problem. Followed by a hierarchical pre-screening criterion for in-fill selection, our method progressively achieves a set of efficient architectures trading-off between segmentation accuracy and inference speed. Empirical evaluations on three benchmark datasets together with an application using Huawei Atlas 200 DK suggest that our method can identify architectures significantly outperforming existing state-of-the-art architectures designed both manually by human experts and automatically by other NAS methods. | ['Fan Yang', 'Changxiao Qiu', 'Haoming Zhang', 'Shihua Huang', 'Ran Cheng', 'Zhichao Lu'] | 2022-08-14 | null | null | null | null | ['real-time-semantic-segmentation'] | ['computer-vision'] | [ 4.48618531e-01 -7.23624751e-02 8.13477300e-03 -4.88218248e-01
-9.90958214e-01 -3.98982793e-01 4.50878114e-01 -2.28814945e-01
-6.22322321e-01 3.59019995e-01 -3.38952899e-01 -3.90726566e-01
-1.73902884e-01 -5.38074493e-01 -6.64983332e-01 -7.06947327e-01
2.40254313e-01 6.03145361e-01 4.07561123e-01 -1.84008345e-01
3.75255138e-01 4.16351557e-01 -1.65415668e+00 -9.64937790e-04
1.15248632e+00 1.46918166e+00 4.96310830e-01 2.52980679e-01
-3.10213149e-01 4.41696644e-01 -4.34931844e-01 -4.13908869e-01
3.78523976e-01 -2.57498562e-01 -9.33459520e-01 1.50769472e-01
2.68507510e-01 5.96395545e-02 1.33822694e-01 1.15552151e+00
4.23360646e-01 1.38219789e-01 4.45759445e-01 -1.06236863e+00
-5.18154502e-01 5.05940855e-01 -6.84831142e-01 2.99667984e-01
-3.88775468e-01 2.47598186e-01 1.20356178e+00 -8.64335775e-01
1.57139927e-01 1.02398241e+00 6.02653205e-01 4.05780286e-01
-1.24683106e+00 -5.56076646e-01 3.77010763e-01 3.61120671e-01
-1.32786381e+00 -4.61040854e-01 9.22735155e-01 -5.51256955e-01
8.79624546e-01 2.19421133e-01 5.81128180e-01 6.90307200e-01
-1.51791153e-02 9.23763871e-01 1.01269341e+00 -1.31814525e-01
4.86244559e-01 4.82087508e-02 1.05496570e-01 7.76761353e-01
8.70735124e-02 -1.71300173e-01 -4.94626075e-01 2.00059175e-01
5.19811511e-01 -4.40827310e-01 -5.67711145e-02 -2.79232264e-01
-1.08492386e+00 8.58817756e-01 5.52028775e-01 2.20629528e-01
-5.19002259e-01 9.30995867e-02 3.64943027e-01 -8.10093433e-02
3.84739697e-01 7.44624317e-01 -5.57560682e-01 1.27073824e-01
-1.13477468e+00 2.40935266e-01 3.80267173e-01 6.63665533e-01
7.81920016e-01 2.19660908e-01 -1.14093766e-01 1.01796305e+00
3.50444525e-01 1.30219147e-01 5.33079505e-01 -8.94620597e-01
3.85124475e-01 9.49094951e-01 -1.20828770e-01 -7.93015420e-01
-4.05066103e-01 -8.15818250e-01 -8.97746801e-01 2.22291455e-01
3.63990307e-01 1.08947091e-01 -1.07241249e+00 1.64696920e+00
2.72118032e-01 4.65026088e-02 -1.68398157e-01 1.10621893e+00
5.65239608e-01 4.82079983e-01 8.98987204e-02 4.05203067e-02
1.49495184e+00 -1.20584273e+00 -3.51131797e-01 -6.86368763e-01
3.60693783e-01 -5.20481765e-01 1.02363217e+00 4.87481087e-01
-9.76471186e-01 -6.46042347e-01 -1.18397415e+00 1.49267623e-02
-2.76792109e-01 2.97816992e-01 7.38846123e-01 7.26944149e-01
-1.07756793e+00 3.82145494e-01 -9.31269705e-01 1.83720211e-03
8.48331213e-01 6.66527867e-01 8.38205144e-02 3.58341843e-01
-9.06931758e-01 6.04591966e-01 5.91379464e-01 4.63501900e-01
-7.12643683e-01 -6.13335669e-01 -6.04408145e-01 5.07341214e-02
7.51286983e-01 -7.33682930e-01 1.29438293e+00 -1.33321035e+00
-1.64574158e+00 9.12427664e-01 -1.01565704e-01 -6.68152809e-01
4.93724138e-01 -2.22908586e-01 -2.19559491e-01 7.54768476e-02
1.30477041e-01 9.43164408e-01 8.25665653e-01 -1.25778091e+00
-8.71401608e-01 -4.76319134e-01 -1.72523111e-02 2.04056948e-01
-5.59910357e-01 -3.33735347e-02 -8.46667647e-01 -6.11885071e-01
1.86027348e-01 -1.05137074e+00 -6.04185343e-01 -6.34344816e-02
-4.23552781e-01 -3.16974938e-01 7.33923912e-01 -4.38282728e-01
1.21190834e+00 -1.88289344e+00 2.99890220e-01 1.19545341e-01
2.25640818e-01 5.80883980e-01 -1.42178282e-01 -3.27678382e-01
1.78295657e-01 8.39698389e-02 -4.96799767e-01 -5.07283628e-01
-2.42087357e-02 1.06876880e-01 -2.80727506e-01 2.41315395e-01
4.55544770e-01 1.04650605e+00 -6.30150020e-01 -6.24785125e-01
2.05348626e-01 3.31395686e-01 -5.47008574e-01 1.88133389e-01
-3.66863489e-01 5.88576674e-01 -6.51303649e-01 7.51409352e-01
3.63107145e-01 -6.34277344e-01 -7.79866502e-02 -3.49461257e-01
-1.72796071e-01 8.03591013e-02 -8.53748679e-01 1.76562750e+00
-3.75515640e-01 4.46151167e-01 1.65525675e-01 -1.44997692e+00
1.10708106e+00 -1.09216077e-02 4.32856292e-01 -1.00234377e+00
3.04048449e-01 3.76296073e-01 9.22389179e-02 -3.18059444e-01
4.99236077e-01 7.84608945e-02 -1.32995531e-01 1.18385278e-01
-1.40151009e-01 4.75408249e-02 2.02336952e-01 -2.71668553e-01
7.86057413e-01 1.40538529e-01 1.28764203e-02 -5.54875016e-01
6.80124223e-01 3.95361304e-01 8.86339366e-01 6.06259048e-01
-3.80584657e-01 5.33307076e-01 2.39268258e-01 -6.18927956e-01
-1.05891490e+00 -7.70586550e-01 -7.36074001e-02 1.05004168e+00
2.54791290e-01 9.14585032e-03 -1.16288054e+00 -6.03270233e-01
-1.71347216e-01 4.14111733e-01 -5.15295923e-01 3.93269956e-02
-7.76750505e-01 -1.03845584e+00 5.72291195e-01 7.34049976e-01
8.18963408e-01 -1.07236290e+00 -1.01442289e+00 3.45733851e-01
-2.65136231e-02 -1.42681122e+00 -2.97865957e-01 1.68400332e-01
-9.31815147e-01 -7.73339152e-01 -6.84230328e-01 -8.24040055e-01
5.85338175e-01 2.95873761e-01 1.08837438e+00 1.65498495e-01
-4.43750054e-01 -8.98859575e-02 -5.10076880e-02 -3.57851088e-01
-1.77880391e-01 3.47339153e-01 -1.59482479e-01 2.32449010e-01
2.89734870e-01 -4.08376992e-01 -9.55192745e-01 5.37686765e-01
-9.46379364e-01 3.93998176e-01 9.32212114e-01 8.28053653e-01
9.23729420e-01 -3.41152325e-02 8.30042660e-01 -7.55077362e-01
3.59861314e-01 -3.32915813e-01 -1.02522349e+00 3.19842786e-01
-9.09935057e-01 2.65579343e-01 6.84408844e-01 -1.55302331e-01
-9.49459732e-01 4.16862607e-01 -3.26781541e-01 -4.19926971e-01
-6.03826270e-02 6.77015781e-01 -1.35297000e-01 -1.21262752e-01
3.35023254e-01 1.97501257e-01 9.05252546e-02 -4.49306130e-01
2.36582071e-01 5.04092753e-01 6.46872103e-01 -5.66784620e-01
6.31055534e-01 4.08226460e-01 7.16482103e-02 -8.32788289e-01
-1.04882550e+00 -3.32543135e-01 -6.29966080e-01 -2.19696760e-01
1.26042533e+00 -8.92585456e-01 -7.18666792e-01 6.47812486e-01
-8.72827411e-01 -3.29466671e-01 1.78724006e-01 4.91024330e-02
-4.74747866e-01 2.71161675e-01 -2.44557992e-01 -6.56655848e-01
-5.92440367e-01 -1.76325250e+00 1.08536005e+00 3.96026760e-01
-2.00980410e-01 -7.89154530e-01 -2.72033930e-01 1.00617349e+00
5.43954909e-01 5.13799563e-02 1.04211819e+00 -5.82105458e-01
-9.32154894e-01 -3.04393806e-02 -4.67968404e-01 5.16137540e-01
-1.50810391e-01 -1.75229445e-01 -9.96216536e-01 -1.03455953e-01
-1.24523193e-01 -3.74693662e-01 8.32077146e-01 4.18021351e-01
1.77374625e+00 -1.27029195e-01 -2.63461649e-01 8.58666599e-01
1.36722934e+00 3.14439505e-01 3.87193620e-01 6.57364249e-01
8.15246046e-01 5.00893116e-01 5.77402055e-01 2.25709796e-01
4.78182226e-01 8.73250008e-01 5.31034172e-01 -1.84256896e-01
-2.22283415e-03 4.39926125e-02 -6.09792210e-02 6.60368681e-01
-3.77378223e-04 -4.14616242e-02 -1.24377096e+00 6.28475010e-01
-1.95134485e+00 -5.06943524e-01 1.11418515e-01 1.99761319e+00
6.29156947e-01 3.92526001e-01 1.28267109e-01 8.43361095e-02
5.79488814e-01 1.51504353e-01 -1.00927877e+00 -1.88051194e-01
1.57622024e-02 2.13913456e-01 5.27656555e-01 1.52392536e-01
-1.22131991e+00 1.01089597e+00 5.61445570e+00 9.81989324e-01
-1.26156819e+00 6.00637943e-02 1.09827030e+00 -2.66079679e-02
-2.38599911e-01 -1.16330482e-01 -1.01030171e+00 4.43277359e-01
7.52575696e-01 1.73970371e-01 4.35657978e-01 8.43159020e-01
1.33126587e-01 7.17349797e-02 -8.09703350e-01 1.04076087e+00
-1.02615975e-01 -1.61531711e+00 4.58241068e-02 1.15193648e-03
7.39274681e-01 2.61977166e-01 1.61209956e-01 2.45463625e-01
1.05780549e-03 -1.11232901e+00 1.00824714e+00 1.33007720e-01
5.32001734e-01 -6.67795956e-01 5.89206576e-01 3.24827224e-01
-1.28422558e+00 -3.25120181e-01 -1.02246597e-01 2.98565328e-01
1.81362763e-01 6.45322144e-01 -5.13755918e-01 4.34568256e-01
7.73964524e-01 3.74878705e-01 -5.77561140e-01 1.08767140e+00
-1.80636928e-03 5.70267081e-01 -2.21639916e-01 -4.78244312e-02
6.92193270e-01 -2.96671331e-01 4.74496543e-01 9.82668579e-01
1.07094504e-01 4.11513168e-03 1.17210127e-01 1.11015761e+00
2.18370575e-02 7.94348121e-03 -6.26684353e-02 -1.97286889e-01
4.26436901e-01 1.37728739e+00 -1.13890123e+00 -8.14987570e-02
-3.66401196e-01 7.72889316e-01 4.13532525e-01 2.76617467e-01
-9.12192166e-01 -6.15816563e-02 6.50608540e-01 -3.54550704e-02
4.32101727e-01 -2.26292983e-01 -9.31799591e-01 -7.19881177e-01
1.76119670e-01 -8.82209957e-01 1.37930632e-01 -3.97135019e-01
-8.51385653e-01 9.34116900e-01 -2.66098112e-01 -9.50558603e-01
3.05793434e-02 -6.76001966e-01 -5.03205121e-01 7.35174477e-01
-1.69336772e+00 -1.15922570e+00 -4.08212721e-01 2.42458925e-01
9.31884706e-01 -3.37389082e-01 6.25358164e-01 5.14250875e-01
-1.09144914e+00 6.53297603e-01 -6.50046691e-02 -2.99909860e-02
1.37721077e-01 -1.14607334e+00 6.67096496e-01 8.18621337e-01
8.98684040e-02 1.57037675e-01 5.77476740e-01 -2.87192613e-01
-1.40960526e+00 -1.18630183e+00 3.82982463e-01 -1.81085855e-01
5.15910387e-01 -2.61134475e-01 -9.73942637e-01 3.38185787e-01
-2.13609282e-02 -7.97188878e-02 4.08363819e-01 4.98983078e-02
-1.66852579e-01 -2.71862775e-01 -7.12708294e-01 5.85950077e-01
9.68126297e-01 -1.66882768e-01 -2.95366973e-01 1.07896358e-01
6.55564308e-01 -4.01080847e-01 -5.79624712e-01 6.55405104e-01
5.02390444e-01 -9.46101665e-01 1.05308807e+00 -3.63244027e-01
4.61448967e-01 -3.53879452e-01 -8.39320421e-02 -1.04653215e+00
-2.30491579e-01 -6.40706062e-01 9.80364457e-02 1.00676334e+00
6.11150980e-01 -6.27167702e-01 1.02625108e+00 7.27556169e-01
-5.18034160e-01 -1.42026031e+00 -1.02237034e+00 -6.07378483e-01
-1.54598325e-01 -3.82949263e-01 5.91856420e-01 5.50218642e-01
-7.01907992e-01 4.22610760e-01 -1.84046831e-02 1.98371515e-01
7.10866868e-01 2.27856919e-01 6.05992734e-01 -1.34759295e+00
-3.56015801e-01 -1.06313121e+00 -2.59940416e-01 -1.23721576e+00
2.22550720e-01 -6.53100073e-01 3.17877680e-01 -1.45553029e+00
1.46086603e-01 -6.82545960e-01 -3.36347878e-01 5.65180600e-01
-2.79051811e-01 2.61310816e-01 6.28708750e-02 3.48113626e-01
-7.30300188e-01 6.74670398e-01 1.10355532e+00 -1.79202750e-01
-2.06620187e-01 1.46761760e-01 -8.88990104e-01 9.00004625e-01
8.18935037e-01 -2.14821234e-01 -5.60382187e-01 -6.87893391e-01
2.41050318e-01 -3.26146483e-02 4.71387923e-01 -1.02342868e+00
1.89420626e-01 -1.47106603e-01 3.25881764e-02 -7.24987924e-01
3.72077584e-01 -6.00988686e-01 -1.05369119e-02 3.41403008e-01
-2.20638648e-01 2.01893345e-01 2.13025883e-01 4.64891613e-01
-3.81581098e-01 -2.55925149e-01 9.35858071e-01 -7.77107328e-02
-1.22062409e+00 3.39156628e-01 -9.81314704e-02 4.06882353e-02
1.13237810e+00 -3.90023351e-01 -2.30027974e-01 2.11019754e-01
-4.36043262e-01 4.88326848e-01 2.69140393e-01 5.78756034e-01
5.27403653e-01 -9.31365669e-01 -4.71102417e-01 1.65334731e-01
1.73350815e-02 4.43983555e-01 2.30585709e-01 8.71474862e-01
-5.02896488e-01 5.89718759e-01 -1.26894787e-01 -9.34270442e-01
-1.04969931e+00 1.06926590e-01 3.89996618e-01 -3.37372392e-01
-6.38305008e-01 1.04342377e+00 3.24225962e-01 -2.45523795e-01
3.55228543e-01 -5.61559126e-02 -1.72350422e-01 2.67780926e-02
3.03783029e-01 1.69464573e-01 4.40733641e-01 -5.54839849e-01
-4.44356024e-01 5.53818524e-01 -2.44406909e-01 7.67656416e-02
1.44305646e+00 -1.35100946e-01 6.21901713e-02 1.82511076e-01
1.09627175e+00 -7.33241856e-01 -1.65899801e+00 -2.09993243e-01
2.86065102e-01 -2.15873599e-01 4.88627404e-01 -8.38269234e-01
-1.48997366e+00 8.21507931e-01 6.00641727e-01 1.02641590e-01
1.37028658e+00 -2.36589983e-02 1.11336446e+00 2.91512102e-01
3.88268858e-01 -1.30157673e+00 2.41404533e-01 3.90900642e-01
6.62966430e-01 -1.40923953e+00 -1.55209303e-01 -3.14563632e-01
-7.99503386e-01 8.33583891e-01 7.73739457e-01 1.25254793e-02
4.52898562e-01 1.41849354e-01 2.00595304e-01 -2.89597750e-01
-5.65186739e-01 -1.96780145e-01 5.62413335e-01 1.68260232e-01
1.69659093e-01 1.94180477e-02 -1.35968089e-01 7.62673974e-01
-1.49153769e-01 -1.68376803e-01 -3.74549907e-03 6.76132143e-01
-5.47540486e-01 -9.14935291e-01 -1.78962141e-01 5.75897694e-01
-4.37615991e-01 8.20846409e-02 -1.39449045e-01 4.19838250e-01
1.64200366e-02 7.59665549e-01 9.36300382e-02 -3.85088831e-01
2.57655114e-01 -3.56590152e-02 1.41860336e-01 -4.13128912e-01
-5.75630009e-01 -3.88102999e-05 3.88537049e-02 -6.64005995e-01
-4.08376575e-01 -5.73767126e-01 -1.27497661e+00 1.69554025e-01
-3.71875763e-01 -7.60101750e-02 9.34550405e-01 1.27198589e+00
5.20768940e-01 7.34749675e-01 4.42623436e-01 -9.28145051e-01
-7.71417022e-01 -5.81655800e-01 -1.07755572e-01 2.52999038e-01
1.11012891e-01 -6.85063958e-01 -7.61430487e-02 -4.41858768e-02] | [9.4392728805542, -0.16261568665504456] |
e65fd484-a0e9-4870-83da-3869c8370efc | role-of-data-augmentation-in-unsupervised | 2208.07734 | null | https://arxiv.org/abs/2208.07734v4 | https://arxiv.org/pdf/2208.07734v4.pdf | Self-supervision is not magic: Understanding Data Augmentation in Image Anomaly Detection | Self-supervised learning (SSL) has emerged as a promising alternative to create supervisory signals to real-world tasks, avoiding the extensive cost of labeling. SSL is particularly attractive for unsupervised tasks such as anomaly detection (AD), where labeled anomalies are costly to secure, difficult to simulate, or even nonexistent. A large catalog of augmentation functions have been used for SSL-based AD (SSAD) on image data, and recent works have observed that the type of augmentation has a significant impact on performance. Motivated by those, this work sets out to put image-based SSAD under a larger lens and carefully investigate the role of data augmentation in AD through extensive experiments on three different models across 420 different tasks. Our main finding is that self-supervision acts as a yet-another model hyperparameter and should be chosen carefully in regard to the nature of true anomalies. That is, the alignment between data augmentation and the underlying anomaly-generating mechanism in given data is the key to the success of SSAD, and in the lack thereof, SSL even impairs (!) the accuracy. Moving beyond proposing another SSAD method, our study contributes to a better understanding of this growing area and lays out new directions for future research. | ['Leman Akoglu', 'Tiancheng Zhao', 'Jaemin Yoo'] | 2022-08-16 | null | null | null | null | ['self-supervised-anomaly-detection'] | ['computer-vision'] | [ 4.69609648e-01 1.56940117e-01 -2.66793966e-01 -5.00026584e-01
-3.61402243e-01 -2.37400770e-01 9.47344184e-01 3.82541299e-01
-2.91372985e-01 3.05563599e-01 1.11312382e-01 -4.43865031e-01
-5.57203963e-02 -4.38554674e-01 -7.24177897e-01 -7.61530578e-01
5.05297258e-02 1.89531639e-01 7.35530853e-02 -1.89358518e-01
1.92650571e-01 5.31990528e-01 -1.55497766e+00 -2.70917505e-01
7.38232672e-01 1.02753329e+00 -1.72258750e-01 1.66718349e-01
-4.10290845e-02 6.77648783e-01 -6.40809238e-01 -4.29305166e-01
4.78818923e-01 -5.64292908e-01 -5.90421975e-01 5.89131653e-01
7.64357448e-01 -1.69055924e-01 -1.60611451e-01 9.88254607e-01
2.64638096e-01 -2.52859388e-02 6.94611073e-01 -1.59467328e+00
-4.37721729e-01 4.12007600e-01 -5.69439590e-01 5.77456415e-01
-1.80870324e-01 4.10064876e-01 1.12810481e+00 -7.56289542e-01
3.01901400e-01 1.11878622e+00 5.51745355e-01 4.62444007e-01
-1.51722395e+00 -5.73017418e-01 4.89245087e-01 1.31002339e-02
-9.71908450e-01 -6.93726659e-01 9.22618032e-01 -4.48851109e-01
5.23585021e-01 -7.20963953e-03 4.36839879e-01 1.53446698e+00
2.53332946e-02 7.21343517e-01 1.27395105e+00 -6.03180647e-01
4.14541394e-01 2.73370922e-01 2.32846975e-01 5.23581207e-01
5.48768878e-01 9.49374139e-02 -5.56145608e-01 -2.68521756e-01
6.79902077e-01 -2.17323989e-01 2.46531777e-02 -4.81518954e-01
-9.57605004e-01 8.39419842e-01 1.56169564e-01 2.80587018e-01
-2.72019178e-01 -1.37548283e-01 5.57551444e-01 5.78798711e-01
6.20380223e-01 8.48825037e-01 -2.84844965e-01 -1.39690144e-02
-5.53697228e-01 1.16086446e-01 3.14872652e-01 3.93271118e-01
5.78655899e-01 5.00687718e-01 3.01955845e-02 8.33294928e-01
1.98764384e-01 2.59308696e-01 5.78481495e-01 -7.77584493e-01
3.01355541e-01 8.22031081e-01 -5.18395863e-02 -1.04386055e+00
-3.38174999e-01 -7.79732347e-01 -7.97067940e-01 3.00070792e-01
6.82143927e-01 -4.56214622e-02 -7.99873710e-01 2.10414410e+00
2.24326596e-01 3.47109944e-01 -1.46326706e-01 9.07646775e-01
2.22369134e-01 2.70655394e-01 1.54117167e-01 -1.23427331e-01
1.17701244e+00 -5.76985180e-01 -7.29706585e-01 -6.84485972e-01
8.10961902e-01 -5.55002630e-01 1.53743732e+00 3.93403918e-01
-6.52677298e-01 -4.10623997e-01 -1.12296498e+00 3.62436593e-01
-2.84893066e-01 2.11219732e-02 7.41945148e-01 6.37084961e-01
-8.03312540e-01 3.26759815e-01 -8.83576632e-01 -6.23361886e-01
4.72107232e-01 1.40634671e-01 -3.51871520e-01 1.91228464e-01
-1.10476005e+00 9.22299445e-01 1.70113385e-01 5.06169461e-02
-8.74665856e-01 -5.37710428e-01 -9.15725768e-01 -1.36389330e-01
5.36421418e-01 -4.85812336e-01 9.95859325e-01 -1.15500557e+00
-1.05553079e+00 9.99207616e-01 -9.29056779e-02 -8.41483593e-01
3.95715803e-01 -2.59916395e-01 -5.03178596e-01 -6.70569018e-02
2.33378008e-01 4.72720802e-01 1.22179806e+00 -1.29064894e+00
-3.34428221e-01 -7.36117423e-01 -2.12463439e-01 1.73764244e-01
-6.01889610e-01 4.59457980e-03 5.40925190e-02 -8.88317227e-01
2.18175769e-01 -9.25210595e-01 -1.87354773e-01 -9.20107439e-02
-4.50670838e-01 -2.66574085e-01 9.52138782e-01 -2.16613322e-01
1.09608436e+00 -2.45610237e+00 -1.25989333e-01 3.00976455e-01
3.08750570e-01 3.44867051e-01 -8.54773968e-02 2.54418641e-01
-3.62269461e-01 4.08757813e-02 -5.17390907e-01 -5.89801788e-01
-2.15878412e-01 2.44512931e-01 -5.66338181e-01 5.55185616e-01
6.40924156e-01 6.36661410e-01 -6.76408172e-01 -2.30400234e-01
1.52152836e-01 1.15133472e-01 -2.75849104e-01 1.74550802e-01
-2.71222025e-01 6.00997329e-01 -3.48534226e-01 6.34836495e-01
3.01604450e-01 -4.46767479e-01 -1.29787520e-01 2.53183365e-01
5.38566522e-02 2.78231919e-01 -1.15423930e+00 1.21454024e+00
-1.44284159e-01 7.65164554e-01 -1.66724175e-01 -1.22545874e+00
9.75896001e-01 2.15071052e-01 4.24595416e-01 -7.87858248e-01
6.33608773e-02 2.61362225e-01 3.03045958e-01 -3.94518048e-01
1.88343495e-01 -6.04865365e-02 1.86934277e-01 6.39114380e-01
6.94438443e-03 6.82497248e-02 1.16964370e-01 2.79213637e-01
1.28828096e+00 -1.79007590e-01 3.88551891e-01 -1.32201642e-01
4.24439311e-01 9.60877258e-03 5.86217225e-01 9.34921026e-01
-3.40905368e-01 5.25693476e-01 6.64161444e-01 -3.19680542e-01
-1.06067181e+00 -8.43621016e-01 -4.14725482e-01 9.13158834e-01
-1.31116718e-01 -6.13049082e-02 -7.59718418e-01 -8.01249206e-01
2.85456493e-03 7.71831095e-01 -6.78079486e-01 -5.19816697e-01
-3.05186242e-01 -9.95426714e-01 6.47889376e-01 4.54689771e-01
4.93779391e-01 -1.14569712e+00 -2.98095554e-01 -4.20799032e-02
-6.47841841e-02 -1.33123195e+00 -1.71867043e-01 3.56171995e-01
-1.22088969e+00 -1.08497643e+00 -2.55215049e-01 -5.14051378e-01
1.02064025e+00 3.77065331e-01 9.70417738e-01 1.20128646e-01
4.69616018e-02 5.06049395e-01 -4.12740558e-01 -7.90382862e-01
-5.96325219e-01 8.66812915e-02 3.67699802e-01 4.68793184e-01
5.45244098e-01 -7.04349220e-01 -2.81822503e-01 4.80707496e-01
-1.15468943e+00 -4.05326426e-01 5.29871762e-01 8.95597577e-01
3.36516052e-01 1.11763224e-01 8.80428910e-01 -1.06330633e+00
7.26976514e-01 -5.49564421e-01 -6.87606692e-01 -1.85539886e-01
-1.01618719e+00 9.50919390e-02 6.15192413e-01 -3.05840373e-01
-9.23554718e-01 -5.41074611e-02 -3.38501632e-02 -5.37159979e-01
-6.04984701e-01 4.12257969e-01 -8.98618251e-02 -5.55034839e-02
9.07423496e-01 2.02584490e-01 4.61219996e-01 -4.68109012e-01
9.09885466e-02 3.54010224e-01 4.55115825e-01 -4.39855546e-01
1.11290467e+00 5.94726324e-01 6.97086332e-03 -1.05329216e+00
-1.14290869e+00 -3.21256548e-01 -5.82470000e-01 -1.30785704e-02
6.25794411e-01 -8.38975191e-01 -8.28330591e-02 6.45919979e-01
-6.96349025e-01 -3.31372589e-01 -5.00828743e-01 3.76583397e-01
-2.29315326e-01 4.55443829e-01 -3.38850170e-01 -9.01044130e-01
3.76522467e-02 -1.16452730e+00 7.34027088e-01 -5.84444031e-02
-3.77015263e-01 -1.03886783e+00 6.38899654e-02 5.50914586e-01
3.60615462e-01 9.68593732e-02 1.02677143e+00 -1.17969871e+00
-4.43320841e-01 -3.12284648e-01 -1.30114004e-01 7.57352531e-01
3.00702244e-01 -9.54713374e-02 -1.24905157e+00 -1.01830825e-01
4.55825835e-01 -3.84717077e-01 8.30839515e-01 3.40905517e-01
1.20371723e+00 -2.19373703e-01 1.75764412e-01 2.49394670e-01
9.98549819e-01 -1.41017428e-02 4.92367834e-01 7.01819062e-01
5.33655703e-01 6.65742278e-01 7.40453839e-01 2.75725901e-01
1.68061391e-01 6.45583689e-01 6.90450370e-01 -4.05914605e-01
-1.42696658e-02 -1.36493266e-01 4.90404248e-01 3.96331728e-01
1.99549764e-01 -1.17655210e-01 -8.05258512e-01 4.83677268e-01
-1.58612502e+00 -8.01713288e-01 -2.24132389e-01 2.43337393e+00
7.37985253e-01 5.06432831e-01 4.31177109e-01 5.94259024e-01
4.86404687e-01 2.57780731e-01 -5.67224383e-01 -1.67943269e-01
-3.74608427e-01 -2.79336199e-02 2.19094664e-01 1.81013569e-01
-1.24986780e+00 7.56822467e-01 5.84008598e+00 4.64387685e-01
-1.13235915e+00 -6.69197217e-02 7.65715301e-01 2.37979561e-01
-8.06755945e-02 3.14274840e-02 -5.46340108e-01 5.67371070e-01
8.02302539e-01 2.49189317e-01 1.64376527e-01 8.30332458e-01
4.23125446e-01 -2.62960523e-01 -1.18479288e+00 5.33090055e-01
1.45173281e-01 -8.76245558e-01 1.89043526e-02 2.60099024e-01
5.43887079e-01 7.47584924e-02 3.15176100e-01 5.84411025e-02
2.66828872e-02 -7.86634088e-01 5.02977073e-01 1.51706384e-02
3.21093023e-01 -4.34125215e-01 8.44587088e-01 2.95391738e-01
-5.22572458e-01 -1.67476922e-01 -5.21290340e-02 -2.11607814e-01
-4.13508620e-03 7.84371734e-01 -9.82509434e-01 1.03882454e-01
5.18422604e-01 5.30475795e-01 -9.86748457e-01 8.03821445e-01
-2.72567391e-01 1.25621176e+00 -2.74432719e-01 3.78605366e-01
3.27789426e-01 -1.14807375e-01 8.70860159e-01 8.67960393e-01
-5.89276217e-02 -2.63785481e-01 6.98702857e-02 7.41641700e-01
2.47519538e-02 9.38858390e-02 -9.32925403e-01 -3.17582101e-01
4.02148813e-01 1.03279233e+00 -7.48095691e-01 -5.05318493e-02
-5.21692812e-01 6.63561463e-01 2.02742651e-01 3.25025469e-01
-5.85452259e-01 7.42498711e-02 6.62212312e-01 4.10588950e-01
-1.06861375e-01 -2.05861434e-01 -6.98212743e-01 -1.04752505e+00
1.77460730e-01 -1.20100498e+00 5.26944935e-01 -4.05654430e-01
-1.56063998e+00 4.17797148e-01 -1.06242150e-01 -1.37031984e+00
-1.74521968e-01 -4.95989978e-01 -5.96115828e-01 3.42350930e-01
-1.52894056e+00 -8.89595568e-01 -3.30511987e-01 4.58571672e-01
5.81196249e-01 -4.20533657e-01 7.42245257e-01 2.36251667e-01
-1.04777527e+00 5.81496418e-01 -2.18012735e-01 1.93909600e-01
9.64185417e-01 -1.18308079e+00 5.10159552e-01 1.21710515e+00
5.15361488e-01 6.15916729e-01 8.62901211e-01 -7.05939889e-01
-9.47140217e-01 -1.03018522e+00 6.47777736e-01 -7.82466114e-01
7.44130731e-01 -2.91626781e-01 -1.25454593e+00 7.76575685e-01
2.03103628e-02 6.35684095e-03 6.65026128e-01 2.86548913e-01
-3.60590398e-01 -2.00256959e-01 -9.89353955e-01 5.76429725e-01
9.24270511e-01 -4.16413069e-01 -5.02369761e-01 1.64780080e-01
6.25972092e-01 -5.04326932e-02 -4.87882793e-01 4.38139617e-01
-4.89655845e-02 -1.10063756e+00 8.01797509e-01 -6.85819924e-01
5.61059237e-01 -2.10216343e-01 -3.87152582e-02 -1.34831774e+00
-1.00491881e-01 -4.29081500e-01 -1.93496317e-01 1.40473092e+00
4.42700803e-01 -9.71562743e-01 7.18344331e-01 6.48854792e-01
-2.12695897e-01 -6.60006702e-01 -7.33388841e-01 -9.22833741e-01
-2.63476610e-01 -7.21652865e-01 4.70800132e-01 1.27785075e+00
-3.62065434e-01 4.07025725e-01 -4.44137841e-01 3.39940935e-01
6.80040538e-01 -4.00867820e-01 9.61621583e-01 -1.56230199e+00
-1.84282269e-02 -2.90018380e-01 -5.48675239e-01 -6.43097103e-01
2.41184354e-01 -7.41035223e-01 -2.12895289e-01 -6.93902493e-01
-7.93224499e-02 -7.12683201e-01 -4.90053624e-01 6.34776235e-01
-2.40363896e-01 2.61723608e-01 -4.77563180e-02 3.48563492e-01
-2.89198846e-01 5.75414896e-01 9.05818284e-01 1.07742757e-01
-3.00125897e-01 3.76824677e-01 -1.02099180e+00 9.55189764e-01
8.86794686e-01 -5.39351821e-01 -6.08778179e-01 -2.13927075e-01
8.31704214e-02 -4.22247320e-01 5.28200746e-01 -7.98941016e-01
-4.42635603e-02 -2.23661929e-01 1.10792466e-01 -1.09473348e-01
2.60051638e-01 -8.77517104e-01 -4.94184107e-01 1.44126713e-01
-5.29468834e-01 1.53889433e-01 1.64738089e-01 7.41496146e-01
-2.90317744e-01 -2.63937086e-01 8.56722414e-01 -2.15065144e-02
-7.96488345e-01 6.84857294e-02 -3.30367386e-01 3.75808537e-01
8.65929663e-01 -2.22258031e-01 -1.36536866e-01 -4.53621387e-01
-4.97762829e-01 9.24043953e-02 4.32427436e-01 6.16710126e-01
3.18388790e-01 -1.11834300e+00 -5.58364451e-01 5.61087012e-01
3.93651217e-01 9.14875567e-02 -1.37444541e-01 9.94383991e-01
5.07121608e-02 2.46691808e-01 -5.65689467e-02 -8.40726972e-01
-9.95620966e-01 2.16317579e-01 2.00722501e-01 -8.63117352e-02
-8.10099721e-01 6.77881002e-01 1.60793692e-01 -2.26510793e-01
5.45952559e-01 5.41376323e-03 -1.52059659e-01 1.71435788e-01
2.84964383e-01 2.37972319e-01 3.28827053e-01 -4.92691606e-01
-1.30141482e-01 4.78496440e-02 -3.15367073e-01 5.59303947e-02
1.37921596e+00 -1.44628599e-01 2.68977713e-02 6.78583801e-01
4.78386253e-01 -9.29314271e-02 -1.20407879e+00 -5.19856155e-01
2.13920251e-01 -5.76051593e-01 1.04235187e-01 -5.80277264e-01
-1.06799293e+00 7.31182218e-01 5.65943182e-01 4.78227794e-01
1.11331367e+00 1.03511021e-01 4.15384740e-01 3.64573151e-01
1.17250696e-01 -1.07978868e+00 5.00792682e-01 2.72524774e-01
6.71163023e-01 -1.70163763e+00 -7.74820596e-02 -1.85639068e-01
-9.35466230e-01 7.01111853e-01 9.58923101e-01 -1.09267477e-02
4.46626574e-01 3.56698707e-02 2.39138767e-01 -3.26877177e-01
-6.10148072e-01 -2.00741425e-01 2.16227949e-01 5.39447248e-01
2.76609987e-01 -2.25872159e-01 4.72375337e-04 2.01835230e-01
-1.40878275e-01 -3.40243608e-01 5.37472010e-01 8.93332541e-01
-1.80541292e-01 -1.19139779e+00 -4.63746279e-01 7.52396226e-01
-3.93356800e-01 7.20389336e-02 -5.75193226e-01 7.80423582e-01
1.22519769e-02 9.48605597e-01 2.36243576e-01 -1.98884696e-01
2.64730424e-01 4.40319926e-01 1.93307661e-02 -6.94477975e-01
-3.54751885e-01 -3.42653096e-02 -3.22342687e-03 -3.90476167e-01
-4.04861927e-01 -8.94140601e-01 -9.16076124e-01 6.92425221e-02
-4.10307020e-01 -1.04766622e-01 5.59764087e-01 1.18948913e+00
4.56175238e-01 2.82705486e-01 8.51080537e-01 -3.38256657e-01
-7.76649475e-01 -9.02476370e-01 -4.99044180e-01 6.55100405e-01
4.98641789e-01 -9.77423728e-01 -8.32464814e-01 -1.10867679e-01] | [7.693565368652344, 2.427936315536499] |
6cbb465a-0a3d-453f-8319-89684a1e8abe | geometry-complete-diffusion-for-3d-molecule | 2302.04313 | null | https://arxiv.org/abs/2302.04313v4 | https://arxiv.org/pdf/2302.04313v4.pdf | Geometry-Complete Diffusion for 3D Molecule Generation and Optimization | Denoising diffusion probabilistic models (DDPMs) have recently taken the field of generative modeling by storm, pioneering new state-of-the-art results in disciplines such as computer vision and computational biology for diverse tasks ranging from text-guided image generation to structure-guided protein design. Along this latter line of research, methods have recently been proposed for generating 3D molecules using equivariant graph neural networks (GNNs) within a DDPM framework. However, such methods are unable to learn important geometric and physical properties of 3D molecules during molecular graph generation, as they adopt molecule-agnostic and non-geometric GNNs as their 3D graph denoising networks, which negatively impacts their ability to effectively scale to datasets of large 3D molecules. In this work, we address these gaps by introducing the Geometry-Complete Diffusion Model (GCDM) for 3D molecule generation, which outperforms existing 3D molecular diffusion models by significant margins across conditional and unconditional settings for the QM9 dataset as well as for the larger GEOM-Drugs dataset. Importantly, we demonstrate that the geometry-complete denoising process GCDM learns for 3D molecule generation allows the model to generate realistic and stable large molecules at the scale of GEOM-Drugs, whereas previous methods fail to do so with the features they learn. Additionally, we show that GCDM's geometric features can effectively be repurposed to directly optimize the geometry and chemical composition of existing 3D molecules for specific molecular properties, demonstrating new, real-world versatility of molecular diffusion models. Our source code, data, and reproducibility instructions are freely available at https://github.com/BioinfoMachineLearning/Bio-Diffusion. | ['Jianlin Cheng', 'Alex Morehead'] | 2023-02-08 | null | null | null | null | ['protein-design', '3d-molecule-generation'] | ['medical', 'medical'] | [ 3.14202458e-01 9.90739539e-02 -8.49095955e-02 3.29223312e-02
-7.65489399e-01 -9.53184307e-01 8.51553202e-01 2.72528738e-01
-8.69270042e-02 9.53162074e-01 2.02804536e-01 -7.43633866e-01
-1.69418335e-01 -1.00360167e+00 -9.32206511e-01 -1.04921949e+00
-1.57380745e-01 6.96981192e-01 -1.09286688e-01 -3.39326203e-01
3.15333158e-01 8.89041483e-01 -7.17082739e-01 4.61625531e-02
1.13604176e+00 2.74290204e-01 1.71325043e-01 7.50355363e-01
1.39759824e-01 3.42663616e-01 -1.86592847e-01 -3.38456780e-01
2.19774619e-01 -7.10104108e-01 -5.06518304e-01 -1.91417813e-01
3.72421712e-01 -4.96622473e-02 -6.30659580e-01 7.81741858e-01
9.59117889e-01 1.74050897e-01 1.14305210e+00 -7.99621046e-01
-1.14708412e+00 2.81626850e-01 -4.00056064e-01 -3.66954803e-02
3.45763564e-01 7.37549543e-01 7.88026333e-01 -9.22969878e-01
1.08881021e+00 1.27010119e+00 6.94223285e-01 8.38880718e-01
-1.66748619e+00 -5.19404411e-01 -2.06375699e-02 -7.66602233e-02
-1.24035716e+00 -2.77202845e-01 7.28919506e-01 -7.36613274e-01
1.19876409e+00 1.95920421e-03 7.72271216e-01 1.43174458e+00
7.33999312e-01 3.45543504e-01 9.99323130e-01 -1.67154625e-01
5.39813459e-01 -4.47579205e-01 -2.74724871e-01 7.42666364e-01
3.66668731e-01 2.22721711e-01 -6.01579964e-01 -5.44244289e-01
8.78942430e-01 2.20967066e-02 -2.59155333e-01 -5.27969837e-01
-1.06797791e+00 1.20070505e+00 5.30040145e-01 7.74387568e-02
-5.90651453e-01 2.93440521e-01 -1.01876512e-01 -1.01834573e-01
5.38593829e-01 7.56702960e-01 -3.19091856e-01 1.05098542e-02
-6.57565296e-01 7.16539681e-01 8.35735500e-01 8.28219056e-01
5.64064801e-01 1.09117001e-01 -5.30286767e-02 2.82914162e-01
3.01119059e-01 4.57490355e-01 3.09938006e-03 -8.14672232e-01
1.74801350e-01 4.22982365e-01 -2.22265329e-02 -7.54127145e-01
-5.55914402e-01 -4.07785088e-01 -9.48718667e-01 1.59572095e-01
3.17126870e-01 -2.79579848e-01 -1.20024276e+00 1.83010340e+00
6.40979707e-01 -8.42072591e-02 1.03631854e-01 6.36263490e-01
9.63138700e-01 8.07754040e-01 3.58090639e-01 -1.03479937e-01
8.87383759e-01 -4.92852807e-01 -1.31837487e-01 1.33514460e-02
8.19866121e-01 -6.81308985e-01 6.48450196e-01 3.62577796e-01
-1.13649285e+00 -1.43685102e-01 -9.34970796e-01 -1.03198074e-01
-4.39417958e-01 -4.50583220e-01 1.21903634e+00 5.86383164e-01
-1.08822036e+00 1.00420320e+00 -8.91267240e-01 -2.89546788e-01
8.70659530e-01 5.78963101e-01 -4.19709831e-01 -2.92556405e-01
-1.10783696e+00 8.51410449e-01 1.62785247e-01 -4.84589487e-02
-1.38836193e+00 -1.14790666e+00 -7.09853649e-01 -3.63419950e-01
-1.54643105e-02 -1.46437871e+00 7.55782247e-01 -4.11100864e-01
-1.55323660e+00 5.87131202e-01 -9.80119035e-02 -5.09007096e-01
4.52770770e-01 2.34634697e-01 5.84976561e-02 9.19512212e-02
-1.35553494e-01 1.11151028e+00 6.27319634e-01 -1.09016097e+00
1.25930712e-01 -4.59393680e-01 4.24423218e-02 1.46367341e-01
9.04743671e-02 -5.42995453e-01 1.58706278e-01 -6.40526891e-01
-2.92117428e-02 -1.03514886e+00 -8.55557859e-01 -3.78104523e-02
-7.39600360e-01 -2.55093724e-02 4.98303413e-01 -4.30296600e-01
6.70136154e-01 -1.52811480e+00 7.49855757e-01 2.56121963e-01
5.15197277e-01 1.77543014e-01 -4.39570636e-01 8.49105179e-01
-2.69351184e-01 4.76590633e-01 -5.40672421e-01 -8.21302757e-02
-2.01127604e-01 -1.01760417e-01 -3.05335075e-02 4.27811205e-01
2.75505334e-01 1.24726474e+00 -9.61090982e-01 1.41380191e-01
6.30835667e-02 1.10621464e+00 -9.28684473e-01 -7.76024014e-02
-7.72099137e-01 9.64865744e-01 -7.04441130e-01 5.02039135e-01
6.89864576e-01 -4.46820736e-01 3.03559422e-01 -1.51239306e-01
6.61234334e-02 1.70968115e-01 -6.00678205e-01 1.90875947e+00
-1.13494001e-01 1.06698036e-01 -1.78998232e-01 -6.98193848e-01
7.12836325e-01 1.97909161e-01 6.91188693e-01 -4.09335554e-01
-1.87688783e-01 2.26352081e-01 2.03076318e-01 -1.96300656e-01
2.56834894e-01 -6.49407804e-01 2.74242848e-01 3.51269215e-01
7.64807612e-02 -6.75175905e-01 8.88043046e-02 5.81259608e-01
1.25116313e+00 3.57928067e-01 -4.81380522e-02 -1.72036380e-01
7.95749128e-02 3.19439918e-01 2.27504745e-01 6.58532441e-01
1.54322162e-01 6.56807363e-01 5.21210909e-01 -2.78632015e-01
-1.39754796e+00 -1.13516033e+00 -4.07567248e-02 5.55025816e-01
-6.77390769e-02 -4.95006114e-01 -8.39230537e-01 -5.79986930e-01
1.77992359e-01 5.28675258e-01 -6.82961047e-01 -3.13139379e-01
-2.74673820e-01 -1.52621114e+00 4.67213392e-01 -2.77087521e-02
-1.76408544e-01 -8.13645005e-01 3.64150077e-01 5.42565167e-01
4.12890583e-01 -7.27076888e-01 -4.79082048e-01 1.93440050e-01
-8.68044436e-01 -1.07511914e+00 -8.93666625e-01 -5.20361066e-01
8.11598599e-01 1.98758349e-01 8.58030081e-01 -1.92653894e-01
-5.82539558e-01 1.98680773e-01 -5.26677119e-04 -3.32087994e-01
-7.52290368e-01 2.09173650e-01 1.20993920e-01 -3.77578378e-01
7.48405084e-02 -1.05704761e+00 -1.10977399e+00 -2.51437798e-02
-9.96278405e-01 2.60564953e-01 5.15283465e-01 6.41783655e-01
8.53846490e-01 -1.76715761e-01 8.50103378e-01 -1.06036651e+00
8.41576874e-01 -6.54039860e-01 -5.20126760e-01 -1.18772246e-01
-6.83576047e-01 2.86630452e-01 5.73619783e-01 -4.24085855e-01
-7.48349726e-01 1.23430021e-01 -5.61271787e-01 2.60097184e-03
-9.92940888e-02 4.72588450e-01 -2.51097023e-01 -4.69737172e-01
9.08297300e-01 3.47983092e-01 1.33773729e-01 -3.43602300e-01
7.80608237e-01 1.88405097e-01 -7.19317496e-02 -7.82516181e-01
8.15955222e-01 5.69226921e-01 5.89814126e-01 -7.89426744e-01
-5.17478585e-01 2.14834046e-02 -5.22983074e-01 1.94125310e-01
8.30342829e-01 -9.32076991e-01 -8.20310712e-01 5.91748178e-01
-1.14054739e+00 -5.97264111e-01 -1.30296007e-01 3.17258149e-01
-6.29178107e-01 5.65201223e-01 -8.66286814e-01 -5.16245365e-01
-4.67737645e-01 -1.45920384e+00 1.11894608e+00 7.43327737e-02
-2.54344851e-01 -1.33528209e+00 2.60870248e-01 3.96489203e-01
4.95198280e-01 6.40812695e-01 1.34680152e+00 -4.18095291e-01
-9.99226391e-01 4.08988930e-02 1.88115746e-01 5.30486852e-02
3.08104753e-01 2.09706184e-02 -6.74274325e-01 -3.11536252e-01
-3.72951359e-01 -1.68631181e-01 1.08296788e+00 7.84708798e-01
7.90484190e-01 -3.40686768e-01 -5.28440833e-01 8.09335291e-01
1.26479256e+00 2.66723901e-01 5.69525003e-01 -5.12243211e-02
1.13400745e+00 3.71263117e-01 -9.11109000e-02 3.59249800e-01
1.93677485e-01 5.41591525e-01 4.48041528e-01 -1.74835369e-01
-2.66622245e-01 -6.38208091e-01 4.10388410e-01 5.21503150e-01
-1.64844766e-01 -6.03723824e-01 -6.94123030e-01 9.44495648e-02
-1.56410801e+00 -9.95926678e-01 -3.20876181e-01 2.06243348e+00
1.10581744e+00 8.24368522e-02 1.50875866e-01 -4.18829560e-01
4.78388041e-01 1.47279441e-01 -1.06622410e+00 -1.45275131e-01
-4.37100768e-01 5.10608554e-01 5.31506062e-01 8.56599092e-01
-8.10412288e-01 1.04262161e+00 6.23050594e+00 1.04769671e+00
-1.03347003e+00 -7.49682561e-02 1.03944516e+00 -1.38236269e-01
-9.87404227e-01 2.14339361e-01 -9.79996383e-01 3.38157237e-01
9.90158737e-01 -1.98832363e-01 4.98405516e-01 4.37845975e-01
7.16617882e-01 1.70086190e-01 -9.73140776e-01 7.79182076e-01
-2.64263511e-01 -2.06793141e+00 5.87355435e-01 4.10271466e-01
1.00347674e+00 5.03713675e-02 3.27776402e-01 -2.23814413e-01
4.50415820e-01 -1.53442419e+00 2.93941826e-01 6.34687841e-01
6.52541339e-01 -8.35542023e-01 1.08786762e-01 1.56257778e-01
-6.74508929e-01 5.57344317e-01 -3.05946052e-01 2.38564238e-01
2.54023761e-01 8.66747439e-01 -8.94185960e-01 4.96934682e-01
1.84634421e-02 8.59611511e-01 -2.45959908e-01 8.63037825e-01
-1.09187730e-01 5.44058621e-01 -3.29180151e-01 -8.40926096e-02
3.71102333e-01 -7.05537677e-01 7.41150379e-01 9.62658465e-01
3.28255087e-01 1.18438452e-01 -5.31692095e-02 1.33394372e+00
-3.91985595e-01 1.18783517e-02 -8.04666460e-01 -5.04089355e-01
2.81830460e-01 1.07686520e+00 -6.06895268e-01 -3.76868732e-02
6.51362613e-02 8.84843051e-01 1.47346184e-01 6.78394556e-01
-7.62682080e-01 -1.09677695e-01 8.55727732e-01 3.43312711e-01
2.83424348e-01 -5.80026865e-01 -8.59086812e-02 -9.18598950e-01
-3.29209536e-01 -9.01133955e-01 -7.58421374e-03 -7.10751891e-01
-1.58526087e+00 3.14301044e-01 -3.31281900e-01 -6.17701173e-01
7.57358447e-02 -8.36583853e-01 -5.42189062e-01 1.02120626e+00
-1.33289599e+00 -1.19629002e+00 2.24300489e-01 1.95769101e-01
1.98045164e-01 2.79184021e-02 7.71839023e-01 1.22817144e-01
-5.34556389e-01 1.02182955e-01 4.70788509e-01 -4.43955779e-01
6.42112315e-01 -1.17755222e+00 8.94523084e-01 4.95153248e-01
-5.14263473e-03 8.78581583e-01 8.01439464e-01 -1.03308070e+00
-1.89329374e+00 -1.20924163e+00 4.36313212e-01 -8.59490454e-01
5.95863402e-01 -5.58045268e-01 -7.34373450e-01 3.13971013e-01
-5.54187298e-02 -1.80758357e-01 8.12736094e-01 -1.18132651e-01
-2.89435923e-01 4.44756389e-01 -1.14447236e+00 7.44019032e-01
1.44716561e+00 -3.32556546e-01 1.91093892e-01 1.17034054e+00
8.92953277e-01 -4.89101022e-01 -1.22972775e+00 2.49369696e-01
2.03699559e-01 -8.04237723e-01 1.32877171e+00 -9.56096292e-01
4.97333735e-01 -2.96004564e-01 -7.06050126e-03 -1.42209816e+00
-3.35916966e-01 -1.18805635e+00 -2.13944599e-01 8.02587271e-01
8.48794878e-01 -7.19133019e-01 1.00343180e+00 4.08491433e-01
-2.11853370e-01 -1.09037471e+00 -8.79369915e-01 -5.87203085e-01
6.85760617e-01 -3.14061105e-01 4.38892096e-01 6.72353566e-01
-2.84206301e-01 4.85709906e-01 -2.26815596e-01 -4.88165133e-02
5.68630695e-01 1.06606921e-02 6.32781506e-01 -9.49461818e-01
-4.96879607e-01 -5.73230863e-01 -2.99688935e-01 -1.19218373e+00
-1.04035586e-02 -1.19622397e+00 -4.24644232e-01 -1.86698627e+00
3.67186636e-01 -6.07155681e-01 1.77175000e-01 2.13887811e-01
-1.24519035e-01 6.87487423e-02 -4.80905697e-02 1.34674937e-01
-3.25214475e-01 9.85071421e-01 1.75031412e+00 -3.72214615e-01
-4.12141293e-01 -4.08941880e-02 -1.00640368e+00 2.67339468e-01
6.74552739e-01 -4.91122395e-01 -5.63846290e-01 -3.03978007e-02
3.70700330e-01 -5.10478802e-02 4.63951528e-01 -7.17396677e-01
-4.45865877e-02 -2.95598090e-01 6.48164570e-01 -1.07609667e-01
4.70553666e-01 -1.75272614e-01 5.30551136e-01 5.26053607e-01
-1.07830733e-01 -9.51090083e-02 3.07958335e-01 8.50759387e-01
3.64983380e-01 8.35776255e-02 8.42931211e-01 -3.08793962e-01
-8.73577520e-02 1.08829117e+00 -4.20672476e-01 1.15358390e-01
9.16831613e-01 -2.21122697e-01 -4.53028381e-01 -4.69797760e-01
-7.41062999e-01 -2.19716311e-01 8.79442334e-01 5.34385741e-02
6.54260993e-01 -1.03216600e+00 -7.59147644e-01 -6.81507811e-02
-1.17030784e-01 2.56144941e-01 4.33644086e-01 7.25282371e-01
-6.55100644e-01 5.05404830e-01 1.82306930e-01 -5.20086348e-01
-8.37048650e-01 5.78044713e-01 5.89491367e-01 -1.85694024e-01
-3.72148335e-01 7.78137505e-01 5.41013777e-01 -5.86666405e-01
-4.03906763e-01 -7.14645162e-02 4.01373178e-01 -2.37441435e-01
1.97522134e-01 -5.63025698e-02 7.50709102e-02 -4.06417936e-01
-1.65298611e-01 6.41803741e-01 -2.69015253e-01 1.76144555e-01
1.75691366e+00 2.67472625e-01 4.93618809e-02 -2.37328231e-01
1.02600098e+00 1.23831436e-01 -1.59690988e+00 2.19971597e-01
-4.25699681e-01 7.60705024e-02 -6.77676275e-02 -7.70869434e-01
-7.81361163e-01 7.76349664e-01 3.86140972e-01 -2.53447980e-01
5.02415836e-01 1.39931962e-01 9.31377172e-01 2.73219168e-01
2.49986485e-01 -5.67624569e-01 1.24315746e-01 4.73090112e-01
7.69800246e-01 -9.16575253e-01 1.84853151e-01 -4.14107412e-01
-3.88080627e-01 1.01126039e+00 2.29917705e-01 -4.39672032e-03
5.43374419e-01 -2.64499579e-02 -3.56499881e-01 -4.70601380e-01
-6.42893255e-01 -1.18806008e-02 2.22650051e-01 1.07243693e+00
6.16569996e-01 3.04424260e-02 4.21707891e-02 1.89517945e-01
-5.33547578e-03 -1.72825903e-01 4.82605159e-01 9.29708898e-01
-3.86233151e-01 -1.59249139e+00 1.00822104e-02 2.58221209e-01
-1.95702478e-01 -5.31614542e-01 -6.35502040e-01 5.99122882e-01
-1.05981678e-01 7.31975853e-01 -4.64457214e-01 -1.27732366e-01
1.73909768e-01 -8.65717977e-02 8.89775455e-01 -6.06381357e-01
-2.52342641e-01 -4.18269224e-02 -1.12509482e-01 -2.29930282e-01
-3.26177269e-01 -4.70276028e-01 -1.28408408e+00 -7.30921507e-01
-1.49085477e-01 2.85311878e-01 6.39541626e-01 6.97724462e-01
1.03108335e+00 2.97969371e-01 4.02587086e-01 -1.18579137e+00
-2.49139085e-01 -6.34207428e-01 -3.99427444e-01 2.91980296e-01
1.92094937e-01 -5.00547349e-01 -2.55997837e-01 -2.25181542e-02] | [5.057145595550537, 5.699237823486328] |
c7b3b676-e417-4db5-b681-0b9c6b7eb73c | semantic-scene-segmentation-for-robotics | 2108.11128 | null | https://arxiv.org/abs/2108.11128v1 | https://arxiv.org/pdf/2108.11128v1.pdf | Semantic Scene Segmentation for Robotics Applications | Semantic scene segmentation plays a critical role in a wide range of robotics applications, e.g., autonomous navigation. These applications are accompanied by specific computational restrictions, e.g., operation on low-power GPUs, at sufficient speed, and also for high-resolution input. Existing state-of-the-art segmentation models provide evaluation results under different setups and mainly considering high-power GPUs. In this paper, we investigate the behavior of the most successful semantic scene segmentation models, in terms of deployment (inference) speed, under various setups (GPUs, input sizes, etc.) in the context of robotics applications. The target of this work is to provide a comparative study of current state-of-the-art segmentation models so as to select the most compliant with the robotics applications requirements. | ['Anastasios Tefas', 'Maria Tzelepi'] | 2021-08-25 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 3.51185769e-01 -1.32096574e-01 5.23874424e-02 -2.87332922e-01
2.08637454e-02 -4.98963147e-01 5.22053540e-01 2.86680013e-01
-8.73019099e-01 2.85331100e-01 -5.75057566e-01 -5.14710307e-01
-2.56241351e-01 -9.19344962e-01 -5.97404599e-01 -4.15424973e-01
3.87099124e-02 8.93386960e-01 9.45639789e-01 -3.05642486e-01
6.55974686e-01 7.70495951e-01 -2.22136950e+00 -2.18305424e-01
9.57359612e-01 7.49879420e-01 8.87041628e-01 6.83679640e-01
-2.83433914e-01 1.10707236e-02 -4.33953255e-01 -1.40226647e-01
7.49962032e-02 -9.32997540e-02 -8.29684794e-01 2.60703832e-01
1.67993307e-01 -6.67023659e-02 6.98822364e-02 1.18326950e+00
2.15491146e-01 3.06223899e-01 4.98511255e-01 -1.20301723e+00
2.66335994e-01 2.32559577e-01 -2.91069478e-01 2.31829703e-01
2.63642490e-01 4.94391434e-02 4.92623150e-01 -4.82846558e-01
7.95984328e-01 1.06022036e+00 3.77684534e-01 3.63421828e-01
-9.98861611e-01 -3.43899190e-01 2.71125436e-01 2.02038199e-01
-1.04024732e+00 -7.90522695e-02 4.21204567e-01 -3.80901545e-01
1.00766599e+00 1.33117884e-01 7.05274343e-01 7.13971913e-01
3.07865858e-01 5.61148465e-01 9.16182995e-01 -3.33397329e-01
7.71430731e-01 1.11200511e-02 4.04189706e-01 5.36050856e-01
6.55808747e-01 -3.42462331e-01 -3.79580796e-01 7.20223458e-03
8.84652078e-01 -3.11101586e-01 1.00454926e-01 -5.22245169e-01
-1.13993192e+00 8.00755143e-01 3.13051730e-01 5.06482959e-01
-2.35322639e-01 2.69780338e-01 4.69269335e-01 -7.91130215e-02
1.45436838e-01 5.40176272e-01 -6.68874741e-01 -2.96614587e-01
-7.25919425e-01 3.05878818e-01 8.75949860e-01 1.21180665e+00
9.66702878e-01 -2.37315565e-01 2.89812177e-01 5.73293805e-01
3.91220540e-01 3.89774859e-01 4.76642728e-01 -9.32907581e-01
3.44720125e-01 5.43357551e-01 -3.81443240e-02 -6.70602262e-01
-1.00440550e+00 -3.18758965e-01 -4.64550078e-01 3.19181055e-01
5.16101301e-01 -1.41134411e-01 -1.19746268e+00 1.24262786e+00
3.94732505e-01 3.32286060e-02 1.23435289e-01 1.08712018e+00
8.37229669e-01 4.66800719e-01 2.73005188e-01 2.78157115e-01
1.74294114e+00 -9.42113459e-01 -3.86947304e-01 -8.06547821e-01
6.95101738e-01 -8.51785779e-01 1.15003085e+00 4.49939013e-01
-6.60529912e-01 -5.01624465e-01 -1.00911796e+00 -1.76338762e-01
-5.40131688e-01 3.41215998e-01 1.01444757e+00 9.26770210e-01
-9.47465837e-01 4.99769926e-01 -1.26945674e+00 -9.69003081e-01
7.21866935e-02 4.98795480e-01 -1.45277441e-01 -2.45837942e-02
-7.41067469e-01 1.02195215e+00 6.81551456e-01 3.03723998e-02
-5.44215262e-01 -2.78173417e-01 -7.86652148e-01 -1.08760424e-01
5.68868697e-01 -7.35728860e-01 1.26502275e+00 -5.92040658e-01
-1.48012221e+00 9.29748416e-01 1.52428269e-01 -6.47843361e-01
8.07125151e-01 -3.54018003e-01 1.16037861e-01 6.91923127e-02
1.56739220e-01 8.59644890e-01 4.06690687e-01 -1.19256723e+00
-9.08465564e-01 -4.91161644e-01 2.90814549e-01 3.93415987e-01
2.62142476e-02 -1.12984963e-01 -8.12525749e-01 2.80983690e-02
2.97591418e-01 -1.27085888e+00 -8.07775617e-01 -2.78938919e-01
-2.63640434e-01 1.04779061e-02 1.08725011e+00 -2.10152671e-01
5.51262140e-01 -2.03325462e+00 6.34841397e-02 1.12142175e-01
-4.07516122e-01 3.05433959e-01 2.65295386e-01 2.45306090e-01
5.43900907e-01 -1.98484346e-01 -4.71072704e-01 -4.66103554e-01
-6.68224394e-02 6.04336500e-01 9.39333066e-02 5.70525944e-01
-1.21703655e-01 4.94962901e-01 -8.05850089e-01 -6.51343882e-01
8.65342021e-01 2.84895629e-01 -4.94261861e-01 -2.00028121e-02
-3.28123808e-01 7.61753500e-01 -7.57682621e-01 3.72493207e-01
5.14477372e-01 3.56233679e-02 1.71005085e-01 2.62209848e-02
-5.11707723e-01 7.64462277e-02 -1.24239349e+00 2.22323465e+00
-5.57184458e-01 5.35169244e-01 2.23872423e-01 -9.20097232e-01
9.14818227e-01 -2.07434714e-01 4.07081604e-01 -5.66695631e-01
5.19507885e-01 5.37418962e-01 1.81606501e-01 -2.81128705e-01
1.07780039e+00 3.76781315e-01 -2.25884944e-01 5.93258925e-02
-2.97446586e-02 -5.78193843e-01 7.71056235e-01 -9.33388323e-02
7.31252968e-01 4.20867503e-01 2.49326289e-01 -7.73238480e-01
3.19642425e-01 5.59948504e-01 1.06248729e-01 5.97254336e-01
-7.87477046e-02 6.22216523e-01 4.77315009e-01 -1.91875488e-01
-1.00045383e+00 -4.87354487e-01 -3.43547493e-01 9.51075196e-01
9.13605094e-01 -8.35172608e-02 -1.28269422e+00 -8.00251216e-02
-2.39065915e-01 6.06719136e-01 -2.00759336e-01 2.63345718e-01
-6.75583541e-01 -9.84992206e-01 3.13782483e-01 2.98346788e-01
6.43321574e-01 -1.31610835e+00 -2.02024579e+00 3.86624902e-01
1.67677239e-01 -1.73270655e+00 4.44300771e-01 3.84564608e-01
-1.22530723e+00 -1.24601507e+00 -4.66532588e-01 -7.36595392e-01
6.11395597e-01 4.46489125e-01 1.00426626e+00 5.68057820e-02
-4.93550867e-01 3.34770769e-01 -4.91805702e-01 -3.46435934e-01
-1.40994996e-01 5.72018266e-01 -2.74108112e-01 -5.04841983e-01
2.51860395e-02 -3.28229576e-01 -5.07197738e-01 5.61660349e-01
-1.13843417e+00 3.26782137e-01 6.08824253e-01 3.84376377e-01
7.75611937e-01 1.01138532e-01 -8.51811767e-02 -1.07569635e+00
3.40164006e-01 -2.34185413e-01 -1.00531220e+00 4.66423109e-02
-3.20353150e-01 -1.00332245e-01 6.47698522e-01 -1.84983194e-01
-8.88812602e-01 3.00648481e-01 -4.20473099e-01 7.52864545e-03
-6.74834132e-01 4.28602010e-01 -1.52600333e-01 -1.62936702e-01
6.09698832e-01 2.53675058e-02 -1.94502205e-01 -4.33759123e-01
3.23178411e-01 4.39846843e-01 5.56906700e-01 -7.11377561e-01
6.01889119e-02 6.15556955e-01 2.25727260e-01 -1.21400297e+00
-3.17131698e-01 -7.24298358e-01 -7.95619369e-01 -2.68178165e-01
8.06821525e-01 -6.17620409e-01 -4.24828351e-01 6.74016118e-01
-1.05692410e+00 -6.79942310e-01 -1.63147505e-02 6.00329161e-01
-9.11459744e-01 2.57659674e-01 -3.96070570e-01 -8.06675553e-01
-1.86886936e-01 -1.64340353e+00 1.31242299e+00 5.18164635e-01
-9.78853852e-02 -7.54661977e-01 -3.10991704e-01 3.56917679e-01
1.28277168e-01 2.36217499e-01 5.42489648e-01 -3.76585424e-01
-6.88578963e-01 2.24470273e-02 -2.83894122e-01 -1.91755012e-01
-2.46074021e-01 7.61454776e-02 -7.91644454e-01 -6.47933111e-02
-2.12440446e-01 1.77777097e-01 6.77114964e-01 6.07303798e-01
1.07096565e+00 4.91649002e-01 -5.22121489e-01 6.25796854e-01
1.62483811e+00 2.48290330e-01 7.75394917e-01 6.86739445e-01
5.30985117e-01 8.05042565e-01 1.15239418e+00 2.60720611e-01
1.50959715e-01 7.90004909e-01 8.09946775e-01 8.90445560e-02
-4.54600975e-02 1.61003843e-01 -1.17478251e-01 4.71153080e-01
-1.23092733e-01 -2.10351571e-01 -1.26650822e+00 5.65303922e-01
-2.21000290e+00 -1.27255321e-01 -6.47300243e-01 2.12540579e+00
-3.09931673e-02 4.36292082e-01 1.08864650e-01 1.73058033e-01
7.38341987e-01 -2.33316254e-02 -5.75800836e-01 -5.25156617e-01
2.02809915e-01 1.83156192e-01 1.09465241e+00 3.07610065e-01
-1.24680257e+00 1.19293571e+00 5.85558224e+00 8.56167078e-01
-1.27215004e+00 2.48759195e-01 2.45918021e-01 2.37319306e-01
1.59032509e-01 7.20319338e-03 -8.33696008e-01 4.55695599e-01
8.24147105e-01 2.34211177e-01 1.99930131e-01 1.15829992e+00
5.09788655e-02 -7.63311803e-01 -6.42681003e-01 9.04593945e-01
-4.49771844e-02 -9.54417706e-01 -3.21700424e-01 6.12429827e-02
4.44115281e-01 4.21109945e-01 -3.66431326e-01 -7.61812031e-02
1.00484394e-01 -7.17611909e-01 9.77826357e-01 -1.58239529e-01
5.93228698e-01 -6.74099982e-01 1.00957108e+00 4.67537880e-01
-1.02649927e+00 2.03131258e-01 -4.57039416e-01 -2.10583121e-01
4.41393286e-01 6.14895821e-01 -4.76905227e-01 6.37272060e-01
9.18407857e-01 3.33938688e-01 -4.21843380e-01 1.11622417e+00
-3.08297426e-01 1.08968258e-01 -6.87188864e-01 -4.56575722e-01
5.86393774e-01 -3.77487808e-01 2.30788678e-01 1.48131156e+00
4.61905569e-01 6.32861909e-03 3.52436483e-01 5.43855608e-01
4.15328890e-01 3.04633349e-01 -3.96954387e-01 3.08911383e-01
2.54965335e-01 1.25644565e+00 -1.84436214e+00 -2.65858322e-01
-2.45681167e-01 8.70739341e-01 1.29452854e-01 1.43613601e-02
-7.10072994e-01 -3.55798721e-01 7.79803813e-01 1.53778926e-01
1.26390263e-01 -7.95172036e-01 -9.24929321e-01 -7.66685426e-01
-1.90554410e-01 -2.55427480e-01 1.34029120e-01 -6.93023860e-01
-4.91081595e-01 4.23937947e-01 1.92101300e-01 -9.98108804e-01
-9.26918834e-02 -9.39947248e-01 -2.25962192e-01 5.18722951e-01
-1.60483479e+00 -1.12224078e+00 -5.70393860e-01 4.86669749e-01
8.09066474e-01 2.68994629e-01 7.44025111e-01 2.34743685e-01
-3.87017429e-01 -1.10719495e-01 -2.63311416e-02 -1.58601359e-01
3.20525840e-02 -1.09116459e+00 6.84995472e-01 9.74402130e-01
3.20348069e-02 3.94432008e-01 1.16242218e+00 -5.37061334e-01
-1.57643831e+00 -9.61919367e-01 2.31104732e-01 1.39937019e-02
4.11864102e-01 -3.80503267e-01 -6.42051220e-01 4.11189467e-01
-3.14688802e-01 -7.98930973e-02 1.51942939e-01 9.06337723e-02
2.74790406e-01 3.43585402e-01 -1.26651216e+00 7.48366237e-01
1.14557815e+00 -4.61268127e-02 -3.91164094e-01 1.90242797e-01
3.73533875e-01 -1.02257299e+00 -5.43360114e-01 5.23159981e-01
5.23475409e-01 -1.19090056e+00 1.00011408e+00 4.27615410e-03
2.50860423e-01 -4.65046495e-01 -1.45566622e-02 -9.45492685e-01
7.30186850e-02 -2.00281456e-01 2.94211000e-01 5.55429816e-01
2.29504824e-01 -8.08288753e-01 1.02032089e+00 4.60123986e-01
-4.94370371e-01 -5.77472627e-01 -9.93498206e-01 -7.38115191e-01
-3.44999105e-01 -6.45222127e-01 4.06121820e-01 5.06066024e-01
-1.89218655e-01 2.47635692e-01 1.03179120e-01 2.46045217e-01
4.28719074e-01 3.56538177e-01 1.06520450e+00 -1.23365319e+00
1.03480227e-01 -6.23306274e-01 -8.96161556e-01 -9.95437086e-01
-9.36254933e-02 -3.54310989e-01 3.58857632e-01 -1.95670712e+00
-2.05407262e-01 -6.67330682e-01 2.17240378e-01 2.23839998e-01
2.84453761e-02 3.41139644e-01 2.65296191e-01 2.48228490e-01
-5.62143683e-01 8.34759921e-02 1.03003716e+00 2.76169747e-01
-3.22626531e-01 7.22384974e-02 -3.28581184e-01 9.88408446e-01
1.03663278e+00 -1.86090246e-01 -4.14939553e-01 -5.54089367e-01
6.42082393e-02 -2.17570052e-01 1.94940850e-01 -1.45097971e+00
3.95270079e-01 -1.65570438e-01 -2.62573481e-01 -5.53814888e-01
4.73496497e-01 -8.39552939e-01 1.76851913e-01 8.75938356e-01
2.98462421e-01 -1.53723344e-01 3.88014019e-01 2.80188888e-01
-1.78280726e-01 -7.54784703e-01 8.50352049e-01 -3.76462191e-01
-1.46743619e+00 -5.26235774e-02 -4.70514953e-01 -1.77858517e-01
1.27823448e+00 -6.34320498e-01 -3.11875314e-01 2.12293357e-01
-5.81553996e-01 1.14195406e-01 7.91029513e-01 4.92859125e-01
2.30048925e-01 -5.29611409e-01 -2.95343220e-01 3.89875881e-02
1.90443352e-01 5.36914289e-01 3.00725251e-01 5.93625665e-01
-1.33997750e+00 5.42092144e-01 -5.20460129e-01 -9.11318958e-01
-1.30082488e+00 3.57254505e-01 1.66616030e-02 -1.69641331e-01
-8.36200178e-01 5.14703155e-01 1.15465939e-01 -4.66453552e-01
6.02199696e-02 -6.66145265e-01 -2.07051724e-01 -5.83519265e-02
9.01822671e-02 4.77644265e-01 4.89299566e-01 -5.27319372e-01
-5.29140353e-01 1.00397682e+00 5.47006369e-01 -8.70455503e-02
9.18313026e-01 -4.39104103e-02 -5.20706587e-02 2.73392588e-01
6.40531063e-01 -5.01789510e-01 -1.15555143e+00 2.48796925e-01
2.21799836e-01 -4.15207177e-01 -2.97564920e-02 -4.02093112e-01
-8.99873078e-01 9.25260901e-01 5.31541467e-01 2.24680960e-01
9.54648972e-01 2.04658844e-02 6.69603705e-01 2.81083673e-01
1.33144033e+00 -1.31379485e+00 -5.13236523e-01 7.13884652e-01
3.08506578e-01 -1.03009009e+00 2.59027213e-01 -1.12630641e+00
-3.78635556e-01 1.07384813e+00 5.79855561e-01 -2.62420118e-01
5.10306180e-01 3.78716201e-01 2.19360851e-02 -2.85035312e-01
2.66758073e-02 -6.80062950e-01 -2.59994745e-01 5.87240160e-01
1.22825257e-01 3.29691380e-01 -7.32187033e-01 1.17356531e-01
-5.08033991e-01 -2.49368083e-02 5.57031691e-01 1.24162900e+00
-7.46236265e-01 -8.87078762e-01 -4.99320537e-01 3.03462982e-01
-3.64546895e-01 3.76545876e-01 -1.60672545e-01 8.33748460e-01
2.63434708e-01 1.11713028e+00 2.45514825e-01 -1.03600830e-01
5.25254309e-01 -3.80235583e-01 5.62483490e-01 -6.77842677e-01
-6.38532758e-01 -5.70139810e-02 1.84395432e-01 -5.08917749e-01
-6.34381056e-01 -7.13091850e-01 -1.56436837e+00 -1.39669314e-01
-3.06105345e-01 -3.07916142e-02 1.41991186e+00 8.79236639e-01
2.69285679e-01 6.55405819e-01 -2.60256201e-01 -1.24658835e+00
2.43389741e-01 -9.27123487e-01 -6.08581722e-01 2.48801380e-01
-4.59244519e-01 -8.43448639e-01 6.96935281e-02 -5.02040647e-02] | [8.504948616027832, -1.996851921081543] |
3ba1364d-e5de-4e97-8475-6234ff177324 | a-content-adaptive-learnable-time-frequency | 2303.10446 | null | https://arxiv.org/abs/2303.10446v2 | https://arxiv.org/pdf/2303.10446v2.pdf | Content Adaptive Front End For Audio Signal Processing | We propose a learnable content adaptive front end for audio signal processing. Before the modern advent of deep learning, we used fixed representation non-learnable front-ends like spectrogram or mel-spectrogram with/without neural architectures. With convolutional architectures supporting various applications such as ASR and acoustic scene understanding, a shift to a learnable front ends occurred in which both the type of basis functions and the weight were learned from scratch and optimized for the particular task of interest. With the shift to transformer-based architectures with no convolutional blocks present, a linear layer projects small waveform patches onto a small latent dimension before feeding them to a transformer architecture. In this work, we propose a way of computing a content-adaptive learnable time-frequency representation. We pass each audio signal through a bank of convolutional filters, each giving a fixed-dimensional vector. It is akin to learning a bank of finite impulse-response filterbanks and passing the input signal through the optimum filter bank depending on the content of the input signal. A content-adaptive learnable time-frequency representation may be more broadly applicable, beyond the experiments in this paper. | ['Chris Chafe', 'Prateek Verma'] | 2023-03-18 | null | null | null | null | ['audio-signal-processing'] | ['audio'] | [ 2.72481024e-01 1.03848368e-01 2.13868305e-01 -4.77239996e-01
-7.14545906e-01 -6.79011583e-01 2.93627143e-01 -1.47377580e-01
-2.71520734e-01 2.82497227e-01 6.53953969e-01 -1.03779808e-01
-3.76319826e-01 -7.63129771e-01 -5.79448104e-01 -5.75853050e-01
-3.94702613e-01 9.92157981e-02 -1.75023302e-01 -3.19605321e-01
-7.47277588e-02 3.38782668e-01 -1.59962082e+00 7.02165484e-01
1.77636698e-01 1.16813636e+00 1.70603856e-01 1.27396095e+00
6.27055168e-02 5.76331198e-01 -6.40228152e-01 6.68880939e-02
3.56832385e-01 -4.93991435e-01 -7.25829422e-01 5.73551022e-02
3.51564467e-01 -1.14666291e-01 -3.76047552e-01 7.51829565e-01
3.81567538e-01 4.75223035e-01 5.64895153e-01 -9.23160732e-01
-6.05677724e-01 9.96737421e-01 1.99394852e-01 2.53790617e-01
2.43223667e-01 4.51123416e-02 1.30793047e+00 -9.70106363e-01
3.11857343e-01 1.28226840e+00 8.80945623e-01 4.15412396e-01
-1.31698227e+00 -4.11120653e-01 1.08758137e-01 2.98836917e-01
-1.02781022e+00 -6.67014182e-01 1.00438070e+00 -4.96689528e-01
1.05813587e+00 3.01552117e-01 8.04093957e-01 9.41806316e-01
1.11194015e-01 4.94643688e-01 7.59106100e-01 -6.47735715e-01
3.48824441e-01 5.44323474e-02 6.62731100e-03 4.11193907e-01
-4.38093901e-01 -2.68988125e-02 -5.87945402e-01 -9.66230631e-02
7.88997293e-01 -6.94181100e-02 -3.34940940e-01 -8.03893507e-02
-1.03013515e+00 9.24312115e-01 3.77823681e-01 5.84308803e-01
-5.17965198e-01 4.27211702e-01 6.17677391e-01 8.70249093e-01
4.48853463e-01 5.05334139e-01 -7.40643919e-01 -8.88958424e-02
-9.73224103e-01 1.12267941e-01 8.21189404e-01 3.29645813e-01
5.66099524e-01 6.98936105e-01 -1.30882159e-01 7.80861020e-01
6.71821833e-02 3.35411765e-02 8.05311441e-01 -8.99509430e-01
2.06564561e-01 3.99198793e-02 -1.40881464e-01 -8.37350130e-01
-5.31684935e-01 -7.67681658e-01 -7.87039459e-01 1.88109517e-01
5.29385865e-01 -3.75392735e-01 -9.33855474e-01 1.75691378e+00
5.89215420e-02 5.28095067e-01 2.03816569e-03 9.83090520e-01
2.90444762e-01 1.06397200e+00 -2.09748507e-01 -6.57747015e-02
1.23124647e+00 -6.69700682e-01 -6.39347136e-01 -1.83291301e-01
2.55610168e-01 -8.28078032e-01 1.32966959e+00 7.31234848e-01
-1.04571009e+00 -7.97707856e-01 -1.19718063e+00 -1.70707002e-01
-4.67310399e-01 8.66845399e-02 3.70511025e-01 5.78876019e-01
-1.11490393e+00 7.85967112e-01 -5.93252897e-01 6.76389262e-02
9.60863680e-02 2.87222117e-01 -2.18141884e-01 4.16398883e-01
-1.27821636e+00 7.22614527e-01 4.62079912e-01 5.85505087e-03
-1.12696266e+00 -9.65789258e-01 -7.82369435e-01 5.32033086e-01
8.34308490e-02 -7.02877522e-01 1.41861641e+00 -1.37224138e+00
-2.02850819e+00 4.07168686e-01 1.53600633e-01 -8.97913635e-01
-1.38865728e-02 -4.27661419e-01 -4.35097009e-01 7.29558021e-02
-2.85360664e-01 2.93410778e-01 1.64545476e+00 -7.12730885e-01
-5.62902570e-01 9.51519087e-02 2.73159534e-01 -7.22166002e-02
-7.01779902e-01 -6.93206713e-02 2.74412960e-01 -1.05919552e+00
8.20715800e-02 -8.69238734e-01 -1.57013133e-01 -1.32363200e-01
9.20479670e-02 -1.65190130e-01 8.68134201e-01 -5.59702575e-01
1.20979941e+00 -2.39623022e+00 4.89867151e-01 1.42756522e-01
6.12983555e-02 9.03804880e-03 -3.67604375e-01 6.37762487e-01
-5.01213491e-01 -2.13669762e-01 -2.67714858e-01 -1.84287295e-01
1.73710436e-01 -3.42431525e-03 -8.48219931e-01 4.11419064e-01
4.30516481e-01 5.77873826e-01 -8.54932368e-01 1.53134346e-01
3.21401536e-01 8.06906223e-01 -9.12485361e-01 2.61637598e-01
-2.93859124e-01 3.30165267e-01 -1.27049208e-01 9.12267193e-02
1.76097751e-01 1.55653534e-02 3.46523598e-02 -4.11370695e-01
-2.35048652e-01 7.94962347e-01 -1.38122809e+00 1.85304201e+00
-1.05010760e+00 1.06013536e+00 2.98310578e-01 -1.18811893e+00
8.65191996e-01 7.19171464e-01 5.70291340e-01 -3.54196727e-01
-5.49618490e-02 1.39119089e-01 1.99041843e-01 -4.22753125e-01
4.40576941e-01 -4.70301777e-01 -1.05061159e-01 3.72172326e-01
6.13328397e-01 -3.02487463e-01 -2.79412121e-01 -2.26988077e-01
1.03721666e+00 4.41690460e-02 1.76864207e-01 -3.07218730e-01
5.83121240e-01 -3.80776554e-01 2.48771697e-01 4.93120879e-01
3.82294804e-01 8.17291915e-01 1.32495195e-01 -6.58699155e-01
-1.07965326e+00 -9.20967281e-01 -1.67784348e-01 1.56334138e+00
-7.12424994e-01 -6.79106772e-01 -7.15023100e-01 2.37415992e-02
-1.22792058e-01 5.32147586e-01 -5.04574239e-01 -3.42482984e-01
-8.83667707e-01 -1.55147836e-01 6.48758888e-01 3.16287786e-01
3.28188576e-02 -1.07023740e+00 -7.10846603e-01 7.38947749e-01
4.09475230e-02 -7.29998767e-01 -5.26700199e-01 8.79457176e-01
-9.45102751e-01 -5.36707044e-01 -7.01680541e-01 -7.39261031e-01
1.86883256e-01 -1.46306053e-01 9.60927665e-01 -4.70485896e-01
-2.75312096e-01 5.75308204e-01 -3.54280740e-01 -6.12654567e-01
-2.73916155e-01 1.53226510e-01 1.61381349e-01 5.41791737e-01
-1.69722348e-01 -8.84970605e-01 -6.88125134e-01 -1.91453546e-01
-1.02147436e+00 -3.47506493e-01 3.43760490e-01 7.87911177e-01
4.70008761e-01 2.20921189e-01 7.11853981e-01 -5.90202749e-01
9.74035263e-01 -5.07803082e-01 -2.88668782e-01 -1.38559014e-01
3.49925943e-02 3.06021452e-01 1.12396991e+00 -8.88514698e-01
-6.77718580e-01 1.23564079e-01 -4.14630145e-01 -6.45436645e-01
5.59345298e-02 6.86612487e-01 1.37852967e-01 2.53443867e-01
9.91859198e-01 1.64679602e-01 -3.10673743e-01 -7.48662770e-01
5.94133317e-01 6.90312207e-01 5.38594663e-01 -4.14029717e-01
7.63520896e-01 1.77712947e-01 -3.30691636e-01 -1.03677702e+00
-8.10981512e-01 -3.57381254e-01 -5.01691580e-01 -1.72167256e-01
6.83955967e-01 -7.16428876e-01 -4.97403264e-01 2.85429478e-01
-1.07819152e+00 -4.64293182e-01 -9.00954247e-01 6.93631947e-01
-7.93005288e-01 -1.66360781e-01 -5.72267771e-01 -7.74367511e-01
-3.13240975e-01 -7.77847111e-01 9.15485740e-01 -9.43353847e-02
-5.00044048e-01 -1.11415219e+00 3.07350367e-01 -3.71525317e-01
8.14435244e-01 -2.36797635e-03 9.52789903e-01 -4.96677250e-01
-1.60594225e-01 -2.11076796e-01 4.02372599e-01 6.08837426e-01
2.75413722e-01 -2.35607401e-01 -1.56612325e+00 -4.06784505e-01
5.41361034e-01 -2.04594031e-01 8.58775318e-01 5.54207087e-01
1.40892148e+00 -7.30192780e-01 2.26861700e-01 8.23789656e-01
1.23587155e+00 2.16987237e-01 4.30572122e-01 1.51321560e-01
3.84219050e-01 5.08614004e-01 2.67846715e-02 4.39304650e-01
-9.58057470e-04 6.87766910e-01 2.07820341e-01 3.39946784e-02
-2.34749645e-01 -2.60564953e-01 6.03425920e-01 1.20737934e+00
-6.38958737e-02 -1.68114543e-01 -6.33346677e-01 5.39561570e-01
-1.57804155e+00 -1.05025280e+00 5.62846303e-01 2.12758350e+00
9.20715272e-01 2.59972274e-01 2.05748215e-01 6.22191370e-01
3.13942492e-01 1.92428291e-01 -5.04944146e-01 -6.65253460e-01
1.83032572e-01 7.77307630e-01 9.73263681e-02 4.98255074e-01
-9.74789858e-01 6.32375598e-01 6.44783115e+00 7.73065984e-01
-1.78766739e+00 1.54339656e-01 1.68999106e-01 -1.35130554e-01
-3.15997332e-01 -8.17587599e-02 -3.23104382e-01 2.39293516e-01
1.61371887e+00 -3.48668724e-01 8.19626033e-01 7.46588290e-01
1.67574972e-01 5.92049837e-01 -1.46726811e+00 1.12346935e+00
-3.16687852e-01 -1.36305606e+00 -5.25575839e-02 -1.44575924e-01
2.12817803e-01 -9.97530818e-02 4.06675905e-01 5.07848740e-01
8.16285238e-02 -1.30031288e+00 1.13691664e+00 6.55307829e-01
7.29383111e-01 -6.12511635e-01 1.84434503e-01 1.69784412e-01
-1.42140782e+00 -3.50488931e-01 -4.00464654e-01 -3.58751595e-01
-6.16392791e-02 5.41650772e-01 -9.60748196e-01 2.23815382e-01
6.58750057e-01 6.43477440e-01 -5.51772192e-02 1.00711274e+00
1.21594697e-01 1.17018044e+00 -3.71193022e-01 2.22055092e-01
4.69847202e-01 7.46279582e-02 7.04201221e-01 1.59411323e+00
4.26014900e-01 6.09684810e-02 7.61368871e-02 5.98025680e-01
-1.23924352e-01 3.02236676e-02 -5.29279172e-01 -1.36137515e-01
3.12362194e-01 1.29857981e+00 -5.38328290e-01 -2.43213773e-01
-3.47752601e-01 6.66723788e-01 1.35519058e-01 3.89571190e-01
-5.71802437e-01 -7.11363673e-01 7.63701975e-01 2.64935762e-01
5.38270891e-01 -4.52510089e-01 6.79296404e-02 -1.13746798e+00
-9.95609164e-02 -1.04693055e+00 3.59484345e-01 -6.80490553e-01
-1.24641705e+00 7.71366358e-01 -2.83074733e-02 -1.49696529e+00
-6.39045596e-01 -6.10650957e-01 -6.71622157e-01 9.14972723e-01
-1.22687936e+00 -8.28609645e-01 2.13703707e-01 7.33107924e-01
9.31868374e-01 -2.99154818e-01 1.15906668e+00 2.41089106e-01
7.33076259e-02 4.03717071e-01 6.69719884e-04 1.04618058e-01
5.47350645e-01 -1.24395490e+00 4.02511597e-01 5.49818397e-01
6.92290187e-01 6.34498000e-01 8.87944937e-01 4.32339460e-02
-1.52652550e+00 -9.96653378e-01 5.45008242e-01 -9.26723704e-02
9.41199958e-01 -6.83297634e-01 -1.15326881e+00 6.04771793e-01
3.99732351e-01 4.53139506e-02 6.93867385e-01 1.75156623e-01
-4.30856168e-01 -4.61193442e-01 -8.13785255e-01 4.07926500e-01
7.28062510e-01 -1.10279548e+00 -7.92839408e-01 2.37838849e-01
9.25358951e-01 -3.76416922e-01 -8.61554921e-01 8.85696933e-02
6.13668442e-01 -6.62211478e-01 1.00309443e+00 -7.94134140e-01
1.14059053e-01 -1.75393611e-01 -2.62780845e-01 -1.65430868e+00
-7.51101971e-01 -8.54905307e-01 -2.21031621e-01 8.37883830e-01
3.15366775e-01 -4.93565261e-01 5.80108464e-01 3.27049159e-02
-4.87319887e-01 -4.79167491e-01 -9.44818914e-01 -5.00515878e-01
4.65529338e-02 -7.91718185e-01 5.85451365e-01 7.80688167e-01
-7.87305366e-03 4.08545822e-01 -5.05304396e-01 2.47132942e-01
2.17695117e-01 1.19589351e-01 4.32053745e-01 -1.11497235e+00
-7.10432947e-01 -4.10218894e-01 -4.48757708e-01 -1.13230550e+00
-2.92392005e-03 -9.74318087e-01 6.77515715e-02 -1.07690096e+00
-5.90401769e-01 -2.61859953e-01 -6.58507168e-01 5.74996471e-01
3.72322261e-01 9.42841768e-02 3.21898818e-01 -1.27273530e-01
-7.79940113e-02 4.12636012e-01 9.03181314e-01 -2.76473790e-01
-4.65772957e-01 2.39568353e-01 -4.32462156e-01 7.96385467e-01
8.54692340e-01 -5.66268027e-01 -6.74089789e-01 -5.53415418e-01
2.94904381e-01 3.58753651e-01 3.69508028e-01 -1.30300641e+00
2.88949579e-01 -4.62371744e-02 3.83519858e-01 -1.79528281e-01
7.10578740e-01 -8.63565922e-01 4.07504082e-01 4.13943559e-01
-7.37449467e-01 -1.91802651e-01 3.74981165e-01 4.01792705e-01
-5.07850766e-01 -3.37967575e-01 6.73627436e-01 -2.80697495e-01
-4.86001492e-01 8.43953267e-02 -6.47506297e-01 -6.49084747e-02
3.20167780e-01 -2.07376003e-01 2.56045192e-01 -5.54127336e-01
-1.15552473e+00 -3.94857943e-01 -2.43152276e-01 6.26576602e-01
7.83515096e-01 -1.32534826e+00 -8.84205401e-01 5.48101068e-01
-2.57106990e-01 -3.29698384e-01 1.94644913e-01 2.01474085e-01
-9.30232033e-02 4.88911331e-01 -4.30344164e-01 -4.80187684e-01
-9.42751348e-01 2.35341504e-01 6.04415536e-01 1.71152614e-02
-8.34840894e-01 7.89008796e-01 2.81863123e-01 -2.18341857e-01
3.58734876e-01 -7.69520640e-01 -1.90474629e-01 2.74267882e-01
5.77340186e-01 1.64921790e-01 2.72270262e-01 -4.64569837e-01
-1.61706179e-01 6.60814703e-01 1.83831543e-01 -4.10701156e-01
1.80169642e+00 1.68779060e-01 8.09221417e-02 9.93945539e-01
1.43539155e+00 1.21205039e-01 -1.19440591e+00 -2.73254395e-01
6.21327683e-02 -1.65345713e-01 2.43199021e-01 -4.43384826e-01
-1.01642621e+00 1.03113270e+00 6.39637351e-01 7.15783954e-01
1.31367886e+00 -2.38248393e-01 6.36484742e-01 5.05850017e-01
1.94022968e-01 -9.86419678e-01 3.75055492e-01 7.95719445e-01
1.16844845e+00 -5.70396006e-01 -2.99807578e-01 2.33680278e-01
-2.36121997e-01 1.67354774e+00 -6.53133243e-02 -5.48232675e-01
9.30557966e-01 4.99135494e-01 8.35189819e-02 -1.04808636e-01
-9.99994516e-01 -1.40532210e-01 7.39281118e-01 5.67977965e-01
6.29005134e-01 -9.24725235e-02 1.92749649e-01 6.62421584e-01
-7.43362308e-01 -1.10793700e-02 4.68626261e-01 6.97753310e-01
-6.80728555e-01 -1.11065197e+00 -4.99772578e-01 3.18051934e-01
-5.15688956e-01 -9.87654999e-02 -9.78173390e-02 3.28048259e-01
1.97847143e-01 7.11516201e-01 2.37884983e-01 -4.36959624e-01
4.11421716e-01 4.79720592e-01 4.95729059e-01 -1.00023532e+00
-9.14059997e-01 1.65634677e-01 -1.70772940e-01 -2.72837073e-01
-2.76448041e-01 -3.11174780e-01 -1.18082213e+00 1.75309643e-01
-6.88130334e-02 1.71391889e-01 7.69592941e-01 7.05206573e-01
1.23592570e-01 9.20607865e-01 7.67091036e-01 -1.17369580e+00
-7.00211763e-01 -8.31655204e-01 -3.17522615e-01 1.45007864e-01
9.84642208e-01 -1.93572760e-01 -3.24302912e-01 5.10177076e-01] | [15.362954139709473, 5.54480504989624] |
e9d7417a-8a0c-49ab-8d2d-5b283afab89f | additive-feature-hashing | 2102.03943 | null | https://arxiv.org/abs/2102.03943v1 | https://arxiv.org/pdf/2102.03943v1.pdf | Additive Feature Hashing | The hashing trick is a machine learning technique used to encode categorical features into a numerical vector representation of pre-defined fixed length. It works by using the categorical hash values as vector indices, and updating the vector values at those indices. Here we discuss a different approach based on additive-hashing and the "almost orthogonal" property of high-dimensional random vectors. That is, we show that additive feature hashing can be performed directly by adding the hash values and converting them into high-dimensional numerical vectors. We show that the performance of additive feature hashing is similar to the hashing trick, and we illustrate the results numerically using synthetic, language recognition, and SMS spam detection data. | ['M. Andrecut'] | 2021-02-07 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [-7.33548477e-02 -3.85441095e-01 -3.79422575e-01 -3.03055823e-01
-8.53604376e-01 -9.18317556e-01 7.18397439e-01 5.70014238e-01
-4.74570841e-01 6.95433676e-01 2.67546952e-01 -3.49919528e-01
6.72018752e-02 -9.68841672e-01 -5.28886735e-01 -9.09082413e-01
-6.08192146e-01 5.17579436e-01 1.31171554e-01 -3.55815679e-01
7.33612239e-01 6.10219896e-01 -1.75851798e+00 4.15732324e-01
8.84307083e-03 9.18014646e-01 -5.24416685e-01 1.07816279e+00
-1.20603088e-02 5.14646113e-01 -6.44632280e-01 -3.91393661e-01
3.52945000e-01 -1.93162516e-01 -7.79034734e-01 -3.47630352e-01
1.71599075e-01 -6.34282231e-01 -6.63993835e-01 7.29039311e-01
2.92392433e-01 1.52663395e-01 9.23141181e-01 -1.60817575e+00
-8.44485044e-01 5.56270003e-01 -1.83918163e-01 -4.43084426e-02
8.97634089e-01 -1.56349465e-01 1.16930771e+00 -1.06671548e+00
3.93782347e-01 1.22946298e+00 1.18456781e+00 1.01467289e-01
-1.35526752e+00 -4.05331463e-01 -8.88333976e-01 3.43480319e-01
-1.72912371e+00 -7.46165812e-02 4.84885097e-01 -3.77516240e-01
9.78101015e-01 5.80630541e-01 5.63308418e-01 4.84883010e-01
3.38172466e-01 3.69684190e-01 8.92633855e-01 -5.61528563e-01
2.86397338e-01 1.73231974e-01 4.22199726e-01 7.93972790e-01
2.92702287e-01 1.63821012e-01 -3.38014662e-01 -1.06421173e+00
4.16102618e-01 3.26984823e-01 1.03279941e-01 -4.90431607e-01
-1.42614198e+00 1.62261760e+00 2.45260060e-01 3.14993113e-01
-1.71930239e-01 3.57652396e-01 6.76598907e-01 6.21345520e-01
-1.44967362e-01 6.39427900e-02 -2.01555550e-01 -3.59299511e-01
-7.83377588e-01 4.92911667e-01 9.27738428e-01 7.35332906e-01
1.04023516e+00 -1.83019623e-01 -1.00326054e-01 6.35554016e-01
1.28531650e-01 6.04038835e-01 7.92186737e-01 -9.11235750e-01
1.19809523e-01 1.86177015e-01 1.94466889e-01 -1.28084111e+00
-3.09563130e-01 3.30140233e-01 -8.82107973e-01 -3.34772728e-02
4.05822307e-01 3.23710978e-01 -4.48242933e-01 1.57212651e+00
2.27347851e-01 -5.07176779e-02 -4.34225760e-02 4.18921560e-01
6.53257370e-01 1.00307095e+00 -3.85410845e-01 -1.24799602e-01
1.31439340e+00 -4.09846395e-01 -6.38432741e-01 6.03775918e-01
8.84293377e-01 -7.52581179e-01 1.00049520e+00 1.17293701e-01
-9.90549743e-01 -3.97767544e-01 -1.17113221e+00 -2.74940431e-01
-8.35096300e-01 -2.87480682e-01 6.74571157e-01 9.14078772e-01
-1.31117463e+00 7.23075449e-01 -4.24270898e-01 -1.99201386e-02
-2.73262680e-01 7.37946212e-01 -6.95345104e-01 1.16748244e-01
-1.57118547e+00 5.65034032e-01 2.90771484e-01 -5.01049280e-01
-2.32410565e-01 -3.91452134e-01 -1.25151801e+00 1.28013387e-01
-5.15324235e-01 -2.80289114e-01 1.14077437e+00 1.07448764e-01
-1.00331485e+00 6.35873020e-01 -3.83504301e-01 -6.77811325e-01
-1.90940857e-01 3.53583336e-01 -1.95329145e-01 4.47225511e-01
2.09410414e-02 5.28235435e-01 1.09973323e+00 -9.06536639e-01
-3.12942982e-01 -1.92023352e-01 -1.28887177e-01 -1.33769855e-01
-7.14578927e-01 1.68384537e-02 3.44692379e-01 -7.07331419e-01
2.46252015e-01 -9.31447685e-01 -1.91647768e-01 1.10729486e-02
-1.56109175e-02 -2.02147797e-01 8.78005683e-01 -2.75652170e-01
1.34878409e+00 -2.46057391e+00 -1.18312344e-01 5.75425088e-01
3.20832372e-01 1.24699119e-02 5.71309179e-02 8.71964157e-01
-1.20376177e-01 8.70935768e-02 -3.50093186e-01 -1.05195649e-01
3.00463676e-01 2.69846380e-01 -5.59490561e-01 8.07195961e-01
-6.52163401e-02 6.33015037e-01 -9.53830719e-01 -6.74051642e-01
2.56914139e-01 6.58308327e-01 -7.82072127e-01 6.90592642e-05
6.00100338e-01 -4.74874765e-01 1.07318453e-01 5.64815223e-01
9.00406539e-01 -1.70796856e-01 1.05390847e-01 -2.15288088e-01
-8.77972040e-03 4.37430531e-01 -1.36307800e+00 7.39684403e-01
-9.16555822e-02 6.05126381e-01 -2.87878841e-01 -8.43759239e-01
9.72708702e-01 3.05416375e-01 6.35691941e-01 -3.94778430e-01
8.08842406e-02 2.40405843e-01 -4.59069669e-01 1.12600058e-01
9.92452323e-01 -2.50423014e-01 -7.48200953e-01 8.65028799e-01
-5.10592498e-02 -2.65076011e-01 8.57780725e-02 4.68459904e-01
1.07036877e+00 -1.02331364e+00 6.40824795e-01 -2.11371537e-02
6.70005023e-01 -9.22876745e-02 4.56003956e-02 9.13018584e-01
-1.60967842e-01 5.27649879e-01 5.44240415e-01 -4.84165341e-01
-1.36876488e+00 -1.39736152e+00 -4.92285311e-01 1.12568152e+00
-3.95540856e-02 -8.96533668e-01 -5.98781586e-01 -3.96721184e-01
7.26415277e-01 2.78354436e-01 -9.06441867e-01 -3.58523726e-01
-5.78253746e-01 -5.43836713e-01 7.35960782e-01 5.38231075e-01
3.89122628e-02 -8.89711320e-01 -2.55151361e-01 1.82445392e-01
-1.01520196e-01 -5.86135745e-01 -5.70896745e-01 5.63717961e-01
-8.01261961e-01 -7.15197504e-01 -4.95487809e-01 -1.07851768e+00
4.34845328e-01 5.57922840e-01 8.28470409e-01 3.27213079e-01
-4.61922586e-01 2.84137696e-01 -4.63116676e-01 4.52652484e-01
-5.80589235e-01 -3.43322009e-02 4.57160294e-01 -1.43827662e-01
4.00761187e-01 -4.27035421e-01 -3.69893610e-01 2.40752488e-01
-1.16396332e+00 -7.37389028e-01 2.27785185e-01 1.32793427e+00
3.28183472e-01 5.59546985e-02 2.74699956e-01 -4.60569829e-01
6.16765797e-01 -5.67380309e-01 -3.74826372e-01 -2.16363743e-01
-1.95287943e-01 3.64585817e-01 9.95042801e-01 -4.93204981e-01
-1.72092915e-02 3.48316580e-01 -3.20317119e-01 -3.52395512e-02
2.91443765e-01 2.36259878e-01 2.45734096e-01 -6.28814042e-01
6.53199553e-01 7.28290915e-01 4.46547568e-01 -2.55061299e-01
7.54069328e-01 1.25864601e+00 6.58652186e-01 -3.65477413e-01
1.03997469e+00 5.98671019e-01 1.11872621e-01 -8.80500376e-01
-6.63824901e-02 -6.27316952e-01 -7.72119164e-01 4.06444043e-01
2.82001525e-01 -4.69277740e-01 -1.08346832e+00 2.96408355e-01
-9.04197693e-01 1.24227807e-01 -4.11190629e-01 3.53493392e-01
-9.52230930e-01 9.06386018e-01 -9.32248652e-01 -7.93272257e-01
-3.38256627e-01 -8.52569222e-01 1.10549569e+00 -3.48145485e-01
-4.31312889e-01 -7.93174207e-01 3.29774320e-01 -3.96796793e-01
4.48420644e-01 -2.56032526e-01 1.29932618e+00 -8.31752777e-01
-4.64210734e-02 -6.32116735e-01 -2.23142475e-01 1.01956844e-01
1.94491241e-02 -5.69790378e-02 -5.71407676e-01 -4.49616283e-01
1.30544007e-01 -3.42379183e-01 6.63835645e-01 -3.58020291e-02
1.13192928e+00 -8.01983416e-01 -2.05037042e-01 6.28853381e-01
1.30743158e+00 6.20553121e-02 7.85666883e-01 3.66890699e-01
1.35689065e-01 2.03901410e-01 6.01249337e-01 1.07931662e+00
4.15587693e-01 6.12865806e-01 1.55128345e-01 1.42605558e-01
3.07353109e-01 -2.11746946e-01 2.18523830e-01 9.55777109e-01
7.06234872e-01 1.68350682e-01 -4.91448343e-01 7.20256925e-01
-1.48394012e+00 -1.32487178e+00 -9.24555734e-02 2.49505401e+00
1.08069670e+00 -5.03362017e-03 5.40602267e-01 7.95213282e-01
8.98871303e-01 9.00708959e-02 -4.33985405e-02 -9.95434105e-01
-1.35609806e-01 4.45367724e-01 6.75668538e-01 6.74972653e-01
-1.20608497e+00 6.29760265e-01 8.57681751e+00 7.99607813e-01
-8.96132410e-01 -3.62011865e-02 1.09178431e-01 2.17516974e-01
-3.26230317e-01 -7.28300884e-02 -7.84627378e-01 7.68359005e-01
1.20055878e+00 -3.83444309e-01 4.83629525e-01 8.11789930e-01
-4.57674116e-01 1.83098406e-01 -9.48996544e-01 1.22786593e+00
-3.44398022e-02 -1.34692132e+00 4.12572294e-01 2.05042794e-01
2.49268308e-01 -4.40949023e-01 4.83397126e-01 4.28857982e-01
1.87434524e-01 -1.08985770e+00 3.89762700e-01 1.18583158e-01
8.59519541e-01 -8.88690054e-01 7.25906253e-01 -1.22955136e-01
-1.53720319e+00 -4.21791971e-01 -7.59047925e-01 -1.53515369e-01
2.79841740e-02 3.83346111e-01 -1.05549943e+00 3.59961987e-02
3.74716789e-01 3.96865696e-01 -6.50667429e-01 1.21201789e+00
6.15821302e-01 3.91736120e-01 -5.52965283e-01 -3.58102202e-01
3.77983153e-01 -7.77945593e-02 2.32260719e-01 1.44095469e+00
4.99293923e-01 1.26407593e-01 -8.45139995e-02 2.36651465e-01
1.56969726e-01 1.98049337e-01 -1.05994797e+00 -7.70978332e-02
8.76802325e-01 1.03597009e+00 -5.30956209e-01 -6.75333440e-01
-3.30417335e-01 9.82123494e-01 3.52807753e-02 -2.50244558e-01
-7.17186987e-01 -1.24323976e+00 9.00677800e-01 1.28751129e-01
7.63829648e-01 -6.19403958e-01 -2.96670139e-01 -7.22960949e-01
-1.66452885e-01 -6.79531157e-01 5.97675562e-01 -3.66263539e-01
-1.13172901e+00 3.70025456e-01 -1.58231944e-01 -1.55846477e+00
-8.14530790e-01 -6.17510676e-01 -1.64755359e-01 5.43109715e-01
-9.05024827e-01 -5.24241567e-01 2.29365453e-01 8.73985827e-01
-1.96880937e-01 -9.81986299e-02 1.26516354e+00 1.30679801e-01
3.06379586e-01 1.21456647e+00 8.54061425e-01 1.52050778e-01
6.26793742e-01 -1.23848653e+00 7.96231508e-01 -8.42398126e-03
-2.04601008e-02 9.16001022e-01 7.07961917e-01 -3.49963486e-01
-1.76156592e+00 -6.58122659e-01 1.46391046e+00 -2.90013939e-01
9.11369324e-01 -6.32516861e-01 -1.09734404e+00 4.87150222e-01
-1.40681535e-01 1.55040205e-01 9.14543927e-01 -2.20548704e-01
-9.46881890e-01 1.22541539e-01 -1.73642755e+00 5.01209736e-01
5.96932530e-01 -1.02565980e+00 -8.60981703e-01 6.07935727e-01
8.31547797e-01 -1.19198889e-01 -1.14304793e+00 2.75280774e-02
8.45918179e-01 -7.27557480e-01 1.60521579e+00 -5.67976177e-01
-6.63301125e-02 -3.00100386e-01 -6.63534939e-01 -8.72078121e-01
-7.75096595e-01 -5.95015109e-01 -2.64704108e-01 7.02339590e-01
1.33213833e-01 -9.07893300e-01 6.40799880e-01 4.36202474e-02
5.98313749e-01 -5.18442452e-01 -1.14034247e+00 -1.01492643e+00
4.87290025e-02 -9.42751095e-02 9.57760632e-01 8.84591341e-01
6.85860217e-01 -2.68490259e-02 -4.61128503e-01 -2.05854416e-01
6.61151350e-01 1.76051646e-01 5.40668130e-01 -1.13924384e+00
-1.99744985e-01 -2.90179402e-01 -1.20673108e+00 -9.77670193e-01
8.12460557e-02 -9.77699459e-01 -8.67364556e-02 -7.29654491e-01
1.80810779e-01 -5.48632681e-01 -3.16481471e-01 5.84683537e-01
1.73810303e-01 8.66254091e-01 7.83169866e-02 2.04140589e-01
-6.15506530e-01 6.28199995e-01 5.27171969e-01 -1.16078652e-01
-3.69181521e-02 -2.61972696e-01 -5.64270556e-01 1.88593745e-01
7.13205695e-01 -6.69847786e-01 1.25845090e-01 3.21095914e-01
2.07732588e-01 1.46898344e-01 3.98938268e-01 -9.25952613e-01
2.44034693e-01 2.32226357e-01 3.79436970e-01 -1.03221607e+00
7.10183442e-01 -6.83162570e-01 -1.15522519e-01 8.81147921e-01
-3.75788540e-01 6.36704206e-01 -7.47189969e-02 3.71662855e-01
-1.01587340e-01 -6.45689666e-01 7.12195992e-01 2.12118760e-01
-2.69246399e-01 -1.06961288e-01 -8.09884369e-01 -2.31738806e-01
8.35206151e-01 -3.98016483e-01 -2.68786103e-01 -7.27787733e-01
-6.69452190e-01 -3.07765692e-01 7.56700456e-01 2.80612726e-02
9.77248669e-01 -1.92086542e+00 -5.53077817e-01 7.42223799e-01
9.55161974e-02 -9.72626626e-01 -1.95135385e-01 4.53760415e-01
-5.47479808e-01 6.72529697e-01 -3.54301423e-01 -5.60794532e-01
-1.38517046e+00 9.75767553e-01 -2.25447282e-01 5.58858588e-02
-5.04625916e-01 4.64976460e-01 -4.11262661e-01 -5.86874366e-01
3.02451015e-01 -1.30405622e-02 -3.65295634e-02 2.39837646e-01
1.01060247e+00 4.72093433e-01 2.06492022e-01 -9.14816260e-01
-5.63595831e-01 7.22293556e-01 -1.24150440e-01 -4.73092496e-01
9.80847955e-01 1.21565945e-02 -4.92404163e-01 3.15923333e-01
2.19267249e+00 4.41365503e-02 -2.94188887e-01 -2.44872421e-01
-1.34956285e-01 -7.40723133e-01 -5.08507371e-01 8.50514695e-02
-3.84338170e-01 6.97502553e-01 5.93283713e-01 7.20148265e-01
6.84427440e-01 -7.77019113e-02 1.21765113e+00 8.07805836e-01
6.27951741e-01 -7.34158754e-01 -3.92423682e-02 7.83817649e-01
4.40075517e-01 -8.63772273e-01 -9.72153991e-02 -3.69906902e-01
-3.24955821e-01 1.21006155e+00 -2.94943213e-01 -3.94476771e-01
8.92727911e-01 4.04822201e-01 -3.18164557e-01 3.10778506e-02
-6.16043329e-01 -8.19047689e-02 -7.89079145e-02 7.47665107e-01
3.41989279e-01 1.75309002e-01 -5.36589742e-01 3.80514085e-01
-7.47257054e-01 -3.28976125e-01 6.69862926e-01 1.21675932e+00
-7.63864875e-01 -1.30699408e+00 -9.98404026e-01 5.84119678e-01
-1.96938917e-01 -3.49276781e-01 -4.01562542e-01 5.56986392e-01
-4.80386972e-01 7.73666263e-01 2.85037518e-01 -9.65468466e-01
-1.51957080e-01 3.29421550e-01 4.12983865e-01 -3.16261023e-01
-6.91532135e-01 -6.23870730e-01 -3.27762961e-01 -4.70568806e-01
2.95072466e-01 -8.34353507e-01 -1.37439489e+00 -8.25442672e-01
-6.09194711e-02 6.56372607e-01 8.30212235e-01 4.46266055e-01
2.91946799e-01 -3.48201096e-01 1.41801810e+00 -7.85111487e-01
-1.14441860e+00 -4.51189518e-01 -7.05218196e-01 4.73521411e-01
7.16103613e-01 -7.60785580e-01 -9.31020916e-01 -2.36722574e-01] | [8.289899826049805, 4.005582332611084] |
18c25a95-0df9-4dd4-ae71-759dd2ac159a | using-semantic-similarity-and-text-embedding | 2303.16694 | null | https://arxiv.org/abs/2303.16694v1 | https://arxiv.org/pdf/2303.16694v1.pdf | Using Semantic Similarity and Text Embedding to Measure the Social Media Echo of Strategic Communications | Online discourse covers a wide range of topics and many actors tailor their content to impact online discussions through carefully crafted messages and targeted campaigns. Yet the scale and diversity of online media content make it difficult to evaluate the impact of a particular message. In this paper, we present a new technique that leverages semantic similarity to quantify the change in the discussion after a particular message has been published. We use a set of press releases from environmental organisations and tweets from the climate change debate to show that our novel approach reveals a heavy-tailed distribution of response in online discourse to strategic communications. | ['Hywel T. P. Williams', "Saffron O'Neill", 'Travis Coan', 'Ben Dennes', 'Tristan J. B. Cann'] | 2023-03-29 | null | null | null | null | ['semantic-textual-similarity', 'semantic-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.23275477e-01 3.28728408e-01 -2.52488196e-01 -3.07908386e-01
-8.66016150e-01 -1.19593763e+00 1.40222478e+00 9.73068178e-01
-5.16306520e-01 7.77404547e-01 1.36916220e+00 -5.53850353e-01
-8.40572491e-02 -8.69966209e-01 -4.45811689e-01 -4.81341362e-01
6.08040802e-02 2.13500842e-01 6.07198000e-01 -7.28692412e-01
8.04729462e-01 3.28115076e-02 -1.19257843e+00 6.62875354e-01
3.40340674e-01 6.49971902e-01 1.84294358e-01 6.26781106e-01
-8.84018302e-01 1.04927063e+00 -9.87604499e-01 -5.88922501e-01
-2.71289106e-02 -4.93797481e-01 -8.70556533e-01 -3.44543815e-01
2.86542565e-01 2.56115139e-01 -6.18268661e-02 1.06062734e+00
4.78965014e-01 -1.12727098e-01 4.00662214e-01 -6.80520594e-01
-6.08400116e-03 1.12148237e+00 -3.03673744e-01 9.65895414e-01
6.08166575e-01 -3.20412546e-01 1.28290451e+00 -2.22785413e-01
1.42291915e+00 1.50316823e+00 8.24850678e-01 -3.25423241e-01
-1.03080547e+00 -5.88247895e-01 3.08810353e-01 -5.40662482e-02
-6.27112389e-01 -2.32550293e-01 7.44902611e-01 -8.64266217e-01
3.95524919e-01 6.43700659e-01 6.70629263e-01 1.28881788e+00
3.86833370e-01 1.45677000e-01 1.71088624e+00 -8.11072662e-02
3.78048003e-01 -1.68216433e-02 4.11517695e-02 2.25115102e-02
8.92972946e-02 -3.37162673e-01 -5.99634528e-01 -7.67319679e-01
-3.17442507e-01 -1.37214452e-01 1.28529504e-01 6.01368845e-01
-1.27124941e+00 1.32469118e+00 2.73409218e-01 8.61617684e-01
-3.59201640e-01 -4.03005034e-02 8.76901925e-01 4.98421282e-01
1.08557773e+00 6.77260578e-01 -1.03632681e-01 -4.41793382e-01
-7.08579659e-01 5.36922336e-01 1.30295789e+00 1.82910502e-01
9.52866733e-01 -8.03898275e-01 -2.88529515e-01 5.29192030e-01
1.73136771e-01 6.24504328e-01 9.52622369e-02 -8.99596810e-01
6.29528940e-01 6.68798208e-01 4.11333591e-01 -2.12860036e+00
-4.59507287e-01 -4.00778681e-01 -3.61138433e-02 -1.44557774e-01
5.50628662e-01 -5.25168300e-01 -1.54025272e-01 1.43391669e+00
7.26855874e-01 -3.16110164e-01 -2.53627747e-01 7.88271785e-01
8.15637231e-01 7.57334590e-01 4.27277744e-01 -2.73414791e-01
1.17995560e+00 -1.48859039e-01 -1.10084641e+00 -1.82593957e-01
4.00586903e-01 -1.09688759e+00 7.78743029e-01 -1.99126661e-01
-7.50239789e-01 2.01364458e-01 -7.81358302e-01 7.34588429e-02
-5.45164168e-01 -1.18793559e+00 1.64857954e-01 3.86948317e-01
-5.31653941e-01 6.45443738e-01 -1.36346921e-01 -5.49383581e-01
2.29233265e-01 -3.06763917e-01 1.52798574e-02 3.18358690e-01
-1.50155652e+00 9.76118505e-01 2.27649599e-01 -5.64574182e-01
-2.84524411e-01 -7.49840438e-01 -2.14814603e-01 -3.93592060e-01
6.39109910e-01 -1.97858915e-01 1.39108968e+00 -8.97872269e-01
-1.20864677e+00 1.03143132e+00 1.23347551e-01 -4.30875331e-01
7.03574359e-01 6.21985532e-02 -4.34862316e-01 2.46996358e-01
3.63452405e-01 -3.14561069e-01 6.19287133e-01 -9.88625884e-01
-9.87795115e-01 -2.98710644e-01 3.68667483e-01 5.39592700e-03
-3.34191382e-01 7.23363161e-01 5.18727362e-01 -6.53668702e-01
-1.38363406e-01 -7.73725629e-01 -2.97145694e-01 -6.67206526e-01
-2.12584227e-01 -1.71887234e-01 8.50421607e-01 -7.52367973e-01
1.65826869e+00 -2.06705284e+00 -1.83647871e-02 2.42143810e-01
3.20276260e-01 -1.76677197e-01 1.98709399e-01 1.39775288e+00
5.76300621e-01 7.29176581e-01 -3.14346366e-02 3.02877247e-01
2.04857692e-01 3.06078851e-01 -6.67746723e-01 7.13518381e-01
-3.65840644e-01 6.17044210e-01 -1.34617102e+00 -4.00982678e-01
-2.55325645e-01 3.89972255e-02 -2.13758886e-01 -1.62834972e-01
-6.03500545e-01 5.99556029e-01 -9.41836119e-01 9.43750814e-02
2.84202039e-01 -3.70741338e-01 5.86322904e-01 -1.84321970e-01
-8.59367013e-01 8.23826849e-01 -6.28460526e-01 1.16788328e+00
-3.10132980e-01 1.16159272e+00 2.97042370e-01 -5.90776145e-01
8.42485547e-01 9.03590024e-02 3.60611737e-01 -8.26491594e-01
6.44859076e-01 2.59557247e-01 -3.30195688e-02 -5.26950121e-01
6.50183976e-01 -4.28533942e-01 -9.02916014e-01 8.24943185e-01
-6.19640708e-01 -1.85510769e-01 1.97429523e-01 1.74979746e-01
1.18207192e+00 -5.98502815e-01 6.15078211e-01 -6.14382327e-01
1.97997481e-01 5.29481888e-01 1.97467789e-01 8.82601380e-01
-2.17044264e-01 1.75468698e-01 7.25987256e-01 -5.55406630e-01
-1.09742558e+00 -4.89449948e-01 -1.33988038e-01 1.63796031e+00
2.03961775e-01 -5.38192689e-01 -5.90275884e-01 -5.12489557e-01
2.80841380e-01 7.90696979e-01 -8.70542526e-01 4.18748111e-01
-7.78088927e-01 -6.34777606e-01 3.50477904e-01 -2.84273922e-01
4.27938402e-01 -7.14574218e-01 -6.84463143e-01 5.52699506e-01
-6.13243759e-01 -1.05049837e+00 -3.70041549e-01 -3.18196565e-01
-3.56262028e-01 -1.21627355e+00 -2.22833082e-01 -1.01209603e-01
1.06935054e-01 1.92512766e-01 1.07460022e+00 -3.61864984e-01
2.64843613e-01 2.36237019e-01 -6.33935630e-01 -7.17251480e-01
-1.12099969e+00 3.55000079e-01 -4.84345436e-01 1.61488578e-01
1.79207131e-01 -3.79761189e-01 -5.03439784e-01 2.67112076e-01
-8.84717822e-01 -3.43759537e-01 -7.35222921e-02 3.35763335e-01
-6.31522983e-02 -3.13365459e-01 5.48001170e-01 -1.26314139e+00
1.07664692e+00 -1.13403165e+00 -3.21800858e-01 -6.43574968e-02
-3.29180181e-01 -2.74667621e-01 1.75511703e-01 -2.55445629e-01
-1.02827346e+00 -7.75076687e-01 -1.43591985e-01 6.39416099e-01
1.57366917e-01 1.07190812e+00 7.06070483e-01 6.13108873e-02
9.85641599e-01 -3.66121978e-01 2.34981664e-02 -4.69079107e-01
4.72279578e-01 8.54086578e-01 2.02971622e-01 -9.67215002e-02
8.60119998e-01 1.01768136e+00 -4.02695030e-01 -6.86863124e-01
-1.44617915e+00 -5.88239253e-01 5.51115833e-02 -6.93449140e-01
8.74257267e-01 -7.50691175e-01 -8.71430337e-01 7.12142140e-02
-1.14594722e+00 -3.06218099e-02 -2.12348759e-01 2.54506052e-01
-1.08438954e-01 8.80638584e-02 -1.39261588e-01 -5.75226068e-01
-1.82484061e-01 -3.89351308e-01 5.09891331e-01 8.16531759e-03
-9.25807416e-01 -1.44150388e+00 5.64480603e-01 4.29282695e-01
9.29644167e-01 9.26588237e-01 3.32688332e-01 -1.07592988e+00
-1.58612654e-01 -2.05275252e-01 -2.09694847e-01 -1.95890933e-01
3.99922520e-01 2.77815498e-02 -7.18787372e-01 -1.03721589e-01
1.28060609e-01 2.17720512e-02 5.64039230e-01 3.92687395e-02
3.62508357e-01 -1.03436077e+00 -4.00343776e-01 -1.32725254e-01
1.23619914e+00 -2.24168785e-02 4.60163206e-01 1.08343279e+00
1.29974842e-01 9.46738660e-01 5.72679877e-01 7.94256568e-01
3.70663732e-01 3.80182415e-01 4.80710208e-01 3.19106996e-01
1.04582705e-01 -3.27244669e-01 2.31445879e-01 9.10677969e-01
7.88192824e-02 -1.77567959e-01 -1.14353740e+00 7.07702756e-01
-1.62911761e+00 -1.49763000e+00 -3.26219201e-01 1.60937500e+00
1.06991613e+00 3.15532118e-01 3.14238876e-01 -3.23285870e-02
9.33411598e-01 9.43584442e-01 1.71731606e-01 -4.96019721e-01
-3.44016969e-01 -9.18031186e-02 8.61999273e-01 9.06717539e-01
-8.37838769e-01 5.20694792e-01 7.25102425e+00 6.68173850e-01
-1.22939301e+00 5.88173687e-01 4.27347600e-01 1.49643555e-01
-9.71665561e-01 1.38591379e-01 -5.66899896e-01 6.00904286e-01
1.32373786e+00 -8.27518404e-01 -1.68426603e-01 2.76848286e-01
5.69037735e-01 -3.83373082e-01 -3.00459266e-01 2.17848390e-01
-8.56800601e-02 -2.11464858e+00 -3.83506805e-01 -1.37755588e-01
8.97438288e-01 3.97986531e-01 -2.38507614e-01 -2.72387713e-01
5.07049382e-01 -5.53116262e-01 1.03381109e+00 3.67116898e-01
1.73940420e-01 -3.71539563e-01 4.52699274e-01 2.59805411e-01
-5.56466281e-01 -2.20314980e-01 7.69061148e-02 -6.26323402e-01
4.71544296e-01 8.71275187e-01 -1.20785117e+00 2.63578594e-01
5.56274533e-01 4.80145186e-01 -2.94419855e-01 6.41157269e-01
-1.13162920e-01 1.14467025e+00 -2.39044353e-01 -6.54651940e-01
7.04269648e-01 -1.19986003e-02 1.12944949e+00 1.57433891e+00
1.38152555e-01 1.22681046e-02 1.95208743e-01 3.34730595e-01
-6.31246716e-02 3.26455444e-01 -5.78537881e-01 -6.66274190e-01
6.21945679e-01 1.08707011e+00 -7.58330345e-01 -1.98653176e-01
-3.68496358e-01 9.22313035e-02 2.04648077e-01 1.22761153e-01
-5.20650208e-01 -1.07389279e-01 5.76061249e-01 5.40912867e-01
2.69592285e-01 -2.12793410e-01 -3.33467215e-01 -4.37927037e-01
-3.15008551e-01 -7.54997969e-01 4.32767034e-01 -1.48570374e-01
-1.56315267e+00 2.61364758e-01 9.44944546e-02 -6.90554321e-01
8.14281255e-02 3.44938040e-02 -7.99264371e-01 5.61029136e-01
-1.44488907e+00 -4.95026559e-01 -2.50303924e-01 -6.38793185e-02
1.34885341e-01 2.33870611e-01 3.60292375e-01 2.15890020e-01
1.21059388e-01 -1.13008864e-01 3.75807136e-01 -8.14795196e-02
7.33874559e-01 -9.15584028e-01 3.92152011e-01 1.12006582e-01
-3.53430122e-01 2.53218293e-01 1.53317177e+00 -8.69012713e-01
-1.17383146e+00 -9.63350296e-01 1.51083148e+00 -5.53473413e-01
1.51078582e+00 -1.81336492e-01 -6.60141706e-01 4.04133677e-01
7.35311210e-01 -6.49224997e-01 9.78564322e-01 1.48694381e-01
-3.65565807e-01 1.39663428e-01 -1.06523693e+00 6.45664215e-01
8.93863559e-01 -6.60809278e-01 -1.23338020e+00 7.93519437e-01
8.98248434e-01 -3.97583514e-01 -1.05237830e+00 -5.17574474e-02
4.59564865e-01 -8.12255800e-01 5.61222434e-01 -6.51681006e-01
5.26582539e-01 -2.11667761e-01 -4.65533376e-01 -1.29249084e+00
-1.90273359e-01 -1.01083839e+00 4.65379298e-01 1.07953191e+00
4.03206795e-01 -9.41560447e-01 7.82046616e-02 1.75879255e-01
-3.79792117e-02 -5.23267351e-02 -8.26439679e-01 -3.07000667e-01
5.54807708e-02 -1.04535215e-01 4.39013422e-01 1.23281038e+00
3.19416881e-01 2.79364139e-01 -2.93762125e-02 -2.24257752e-01
2.67131984e-01 4.10715520e-01 7.53526568e-01 -1.31129003e+00
1.19292244e-01 -3.86212856e-01 -2.92845041e-01 -2.69399136e-01
-1.31588325e-01 -8.87402713e-01 -2.71885604e-01 -1.48862886e+00
2.03568488e-01 -2.83262819e-01 1.93698153e-01 -2.51606822e-01
6.38128966e-02 2.04648077e-01 1.59954101e-01 3.77636462e-01
-5.11383057e-01 1.24080941e-01 1.25404489e+00 -2.72273481e-01
-7.26386085e-02 -2.97472626e-01 -1.01516223e+00 7.72139728e-01
7.20012486e-01 -7.72907317e-01 1.01919301e-01 1.45934578e-02
1.27458537e+00 -2.12250769e-01 3.15390855e-01 -3.53267938e-01
1.27176315e-01 -6.41583145e-01 -3.86307895e-01 -5.49290717e-01
-3.28231543e-01 -4.60963875e-01 6.43378198e-01 6.15402639e-01
-8.23726356e-01 1.37567520e-01 1.84639506e-02 8.74694109e-01
-5.04813790e-01 -4.30393331e-02 4.08528775e-01 -2.07767025e-01
-4.38691914e-01 -1.71003953e-01 -7.39273608e-01 7.47276545e-01
9.01201487e-01 1.15249090e-01 -7.40418494e-01 -6.21701419e-01
-7.14783728e-01 -8.79082680e-02 4.42614377e-01 4.91336852e-01
-2.08829656e-01 -1.09196830e+00 -1.29284036e+00 -8.74903262e-01
-2.95202527e-03 -5.86591601e-01 8.15084111e-03 1.08008778e+00
-4.56884027e-01 8.19993466e-02 6.05189390e-02 -2.92115211e-01
-1.08358133e+00 -8.43520556e-03 -7.65130147e-02 -1.56667531e-01
-7.56609738e-01 4.59157467e-01 -2.11130321e-01 -1.83349699e-01
-2.95036435e-01 -6.70093149e-02 -2.95347422e-01 1.20321095e+00
7.92727649e-01 6.74274981e-01 -2.13026688e-01 -8.36618960e-01
-3.34036440e-01 1.23009063e-01 2.57189065e-01 -3.47526670e-01
1.43772948e+00 -5.92992902e-01 -3.70535582e-01 9.58465517e-01
1.48813272e+00 5.39726853e-01 -7.72314191e-01 -3.81457239e-01
5.90084612e-01 -8.08361351e-01 3.41928713e-02 -7.14772642e-01
-3.47443759e-01 1.11637883e-01 -5.92008829e-02 1.05104709e+00
4.04975921e-01 4.31716949e-01 7.85485506e-01 1.76339120e-01
5.83497360e-02 -1.53009498e+00 1.83664318e-02 7.33773291e-01
1.22076714e+00 -9.01435018e-01 1.43219441e-01 -3.88800561e-01
-5.03172934e-01 9.17027414e-01 -4.59755272e-01 -1.33722261e-01
7.22278357e-01 3.45127642e-01 3.28583091e-01 -5.92028260e-01
-5.58835864e-01 -6.19299971e-02 -5.15617989e-02 1.05466433e-01
4.46403980e-01 2.23861977e-01 -1.14600444e+00 2.31046766e-01
-4.38950986e-01 -3.32442135e-01 8.83573949e-01 9.90598023e-01
-1.01193202e+00 -8.60438049e-01 -5.66350281e-01 3.79571915e-01
-1.07099712e+00 1.02986977e-01 -8.66437793e-01 6.32131815e-01
-2.01740935e-01 1.16720653e+00 5.54440618e-02 -3.15406770e-01
1.57403067e-01 -4.08422910e-02 -1.05972305e-01 -4.59702164e-01
-1.26960707e+00 -1.62355393e-01 1.14254248e+00 -5.29546797e-01
-1.03456438e+00 -1.18118536e+00 -1.18303311e+00 -6.61903024e-01
-1.03766605e-01 3.00101012e-01 1.02839339e+00 1.17719269e+00
3.85082245e-01 3.83029819e-01 8.55967581e-01 -3.78010839e-01
-6.49411500e-01 -9.70513523e-01 -2.49303073e-01 8.24126303e-01
4.72671628e-01 -6.54617906e-01 -7.00271606e-01 -1.03881918e-01] | [8.623408317565918, 9.88591480255127] |
9f17276e-a5a8-4272-a7c9-dc382333fa36 | pac-gan-an-effective-pose-augmentation-scheme | 1906.01792 | null | https://arxiv.org/abs/1906.01792v1 | https://arxiv.org/pdf/1906.01792v1.pdf | PAC-GAN: An Effective Pose Augmentation Scheme for Unsupervised Cross-View Person Re-identification | Person re-identification (person Re-Id) aims to retrieve the pedestrian images of a same person that captured by disjoint and non-overlapping cameras. Lots of researchers recently focuse on this hot issue and propose deep learning based methods to enhance the recognition rate in a supervised or unsupervised manner. However, two limitations that cannot be ignored: firstly, compared with other image retrieval benchmarks, the size of existing person Re-Id datasets are far from meeting the requirement, which cannot provide sufficient pedestrian samples for the training of deep model; secondly, the samples in existing datasets do not have sufficient human motions or postures coverage to provide more priori knowledges for learning. In this paper, we introduce a novel unsupervised pose augmentation cross-view person Re-Id scheme called PAC-GAN to overcome these limitations. We firstly present the formal definition of cross-view pose augmentation and then propose the framework of PAC-GAN that is a novel conditional generative adversarial network (CGAN) based approach to improve the performance of unsupervised corss-view person Re-Id. Specifically, The pose generation model in PAC-GAN called CPG-Net is to generate enough quantity of pose-rich samples from original image and skeleton samples. The pose augmentation dataset is produced by combining the synthesized pose-rich samples with the original samples, which is fed into the corss-view person Re-Id model named Cross-GAN. Besides, we use weight-sharing strategy in the CPG-Net to improve the quality of new generated samples. To the best of our knowledge, we are the first try to enhance the unsupervised cross-view person Re-Id by pose augmentation, and the results of extensive experiments show that the proposed scheme can combat the state-of-the-arts. | ['Chengyuan Zhang', 'Shichao Zhang', 'Lei Zhu'] | 2019-06-05 | null | null | null | null | ['cross-view-person-re-identification'] | ['computer-vision'] | [ 1.24886774e-01 -2.07787648e-01 9.87155437e-02 -3.09799284e-01
-6.36244178e-01 -2.61312485e-01 6.98746622e-01 -5.95573485e-01
-4.71993715e-01 7.55192995e-01 5.26405096e-01 5.00165761e-01
1.97637901e-01 -9.18876827e-01 -6.58229113e-01 -7.30227232e-01
5.74853480e-01 6.03318572e-01 1.20102994e-01 -3.15492749e-01
-2.28861317e-01 4.81301039e-01 -1.63112748e+00 6.62981197e-02
6.65440440e-01 5.80642462e-01 -4.46061455e-02 4.35147375e-01
2.86875904e-01 2.61768430e-01 -7.19469845e-01 -7.48591185e-01
4.27657247e-01 -6.36600137e-01 -4.81649846e-01 4.04070675e-01
5.21467924e-01 -7.39956796e-01 -6.70242012e-01 9.64839041e-01
9.62421775e-01 3.17055523e-01 6.86100006e-01 -1.43127441e+00
-5.34929633e-01 9.03411880e-02 -7.76795745e-01 4.09106649e-02
6.00838602e-01 2.41329104e-01 3.66672069e-01 -7.39695489e-01
5.41011631e-01 1.47384226e+00 7.42602766e-01 1.11953402e+00
-7.61444092e-01 -7.34379172e-01 2.80557811e-01 1.71697825e-01
-1.48207092e+00 -2.69331485e-01 8.52725089e-01 -3.52508545e-01
3.41061562e-01 2.81074345e-01 9.53826725e-01 1.51932704e+00
-3.60531837e-01 9.21860158e-01 9.96800959e-01 -3.14873904e-01
-1.34443030e-01 -1.27319545e-02 -1.97830290e-01 5.20048916e-01
4.25175399e-01 2.65246779e-01 -3.50609750e-01 1.29961213e-02
9.71086860e-01 4.01082039e-01 -3.52042258e-01 -4.28814888e-01
-1.19696212e+00 4.85421866e-01 4.44384128e-01 -4.17108200e-02
-1.49464309e-01 -3.07491850e-02 5.69222867e-01 -1.32308424e-01
2.53432989e-01 -1.03298552e-01 -1.22598737e-01 1.20681383e-01
-7.69214451e-01 4.39521968e-01 2.60729045e-01 9.66359913e-01
4.89838243e-01 9.46947038e-02 -2.19999358e-01 8.73336732e-01
3.40179563e-01 9.20822501e-01 6.62724018e-01 -4.93723571e-01
6.10613465e-01 6.20715976e-01 1.49436876e-01 -1.00361264e+00
-1.11960605e-01 -5.85316837e-01 -1.09430742e+00 -3.51286419e-02
4.37480152e-01 -2.07805112e-01 -1.03737855e+00 1.84871256e+00
3.54169965e-01 1.93677619e-01 1.59313858e-01 1.21155858e+00
1.02320206e+00 6.06927156e-01 1.20016858e-01 2.81196125e-02
1.42352080e+00 -1.20854938e+00 -4.45713341e-01 -3.05819094e-01
5.05585335e-02 -6.26475096e-01 8.78915429e-01 2.11186320e-01
-8.29809368e-01 -1.02891362e+00 -1.05145574e+00 1.62567988e-01
-2.68321842e-01 3.85941893e-01 2.93680787e-01 9.36945379e-01
-4.01134044e-01 1.27982676e-01 -5.11581481e-01 -5.12068689e-01
3.90593261e-01 2.22236201e-01 -7.28457689e-01 -4.18919921e-01
-1.07664871e+00 6.59894586e-01 4.07194227e-01 2.91190118e-01
-1.08860767e+00 -4.33244109e-01 -7.83791780e-01 -1.83474362e-01
3.79248500e-01 -1.01188040e+00 7.45526731e-01 -1.07586443e+00
-1.30240762e+00 8.81580532e-01 -1.62883058e-01 -1.59888938e-01
8.42416048e-01 -3.02389503e-01 -5.62824368e-01 -1.28325587e-02
2.77066499e-01 6.83785021e-01 8.58614743e-01 -1.58651435e+00
-5.44422328e-01 -6.46890402e-01 -6.54813945e-02 3.53521049e-01
-3.51505011e-01 -8.39771554e-02 -9.56574261e-01 -8.79345179e-01
-8.28618929e-02 -1.13905370e+00 -2.65030652e-01 -2.85885990e-01
-5.41674197e-01 5.34594059e-02 7.23589003e-01 -9.55985665e-01
7.32546210e-01 -1.94214785e+00 2.41594940e-01 1.24572448e-01
3.01265065e-02 4.76996928e-01 -1.40430644e-01 2.29734197e-01
-2.27232993e-01 -9.59145501e-02 -1.39301047e-01 -5.60503602e-01
-1.17890239e-01 1.82207599e-01 -5.43482192e-02 4.74960595e-01
-8.80557522e-02 1.02509844e+00 -8.05358171e-01 -4.54107553e-01
4.03633505e-01 6.23067081e-01 -3.10877711e-01 4.07076180e-01
1.56603679e-01 8.10991049e-01 -4.48474675e-01 5.95380604e-01
8.48655164e-01 1.87733516e-01 -1.91254050e-01 -5.30429304e-01
3.31359029e-01 -4.14603323e-01 -1.39289367e+00 1.62347376e+00
5.11898734e-02 -1.17999166e-02 -3.96945566e-01 -9.15215671e-01
8.65869164e-01 2.90658355e-01 5.43150485e-01 -5.36202848e-01
1.64373755e-01 2.86746547e-02 -2.88042307e-01 -4.33527380e-01
4.07949358e-01 -7.70192593e-02 -2.70602465e-01 1.92119822e-01
5.48392534e-02 3.87095541e-01 7.61584416e-02 -7.51416758e-02
4.64944661e-01 3.45932394e-01 -1.11750066e-02 2.59632051e-01
9.93765235e-01 -1.87736511e-01 8.21592271e-01 4.37013447e-01
-2.77922988e-01 1.04785812e+00 -8.15081503e-03 -4.98639494e-01
-1.40913963e+00 -1.11360431e+00 1.83368161e-01 7.07426786e-01
2.85045803e-01 -1.95688725e-01 -9.78015482e-01 -8.81860375e-01
-2.02619851e-01 1.61262289e-01 -5.19206882e-01 -1.53714791e-01
-6.62718534e-01 -9.14305270e-01 8.30857813e-01 8.50453436e-01
1.20497346e+00 -9.40135062e-01 -2.01859906e-01 -6.96017817e-02
-5.42001367e-01 -1.23098135e+00 -8.31514120e-01 -6.42865241e-01
-5.22723556e-01 -1.19052517e+00 -1.49624777e+00 -8.88151109e-01
9.39429820e-01 3.29161525e-01 8.04202616e-01 -1.51023790e-01
-9.53593254e-02 5.45441568e-01 -4.22376931e-01 -2.32683837e-01
-5.78971058e-02 -4.95874770e-02 3.83808494e-01 3.31081718e-01
4.52324778e-01 -4.76564795e-01 -8.91464651e-01 5.62509954e-01
-6.69805586e-01 1.67023942e-01 6.32329762e-01 9.72630262e-01
6.68880880e-01 9.96639431e-02 5.64881861e-01 -5.46201944e-01
3.68395030e-01 -9.84658226e-02 -1.26184031e-01 2.50177205e-01
-2.51870424e-01 -2.90721536e-01 5.51861286e-01 -5.23604214e-01
-1.16296852e+00 1.60248369e-01 -4.16220665e-01 -5.31757414e-01
-4.84792799e-01 3.48913134e-03 -9.32239473e-01 6.89167902e-02
3.06771189e-01 5.84535956e-01 -3.89012434e-02 -5.04635155e-01
2.91329890e-01 6.20812833e-01 8.57576609e-01 -6.31453872e-01
9.93282735e-01 6.12324536e-01 -1.38989732e-01 -4.71656144e-01
-5.68003953e-01 -4.61613774e-01 -5.85585237e-01 -3.66785049e-01
1.07679856e+00 -1.22599721e+00 -6.91605568e-01 8.95682096e-01
-1.00626266e+00 1.25171319e-01 -2.89773494e-01 4.60464567e-01
-4.15430903e-01 6.23527944e-01 -3.92838746e-01 -7.66237438e-01
-5.41574895e-01 -1.09760058e+00 1.13064086e+00 5.38767517e-01
1.26906529e-01 -5.05511105e-01 -2.79863942e-02 7.88226724e-01
3.48045677e-02 5.04831612e-01 2.42137969e-01 -5.17123342e-01
-3.53053361e-01 -4.15052503e-01 -1.29033387e-01 5.00012398e-01
1.51212230e-01 -5.10570884e-01 -8.28126907e-01 -5.97427368e-01
-3.13369751e-01 -1.98149383e-01 7.25762308e-01 9.53239128e-02
1.05765951e+00 -2.75754958e-01 -5.17185926e-01 6.08656168e-01
1.27602124e+00 1.16963863e-01 9.19269383e-01 4.25077796e-01
1.12771106e+00 5.55024147e-01 6.03217781e-01 3.42699170e-01
4.68065172e-01 8.91424358e-01 1.64852321e-01 -2.07612261e-01
-3.10096771e-01 -7.57244825e-01 2.93911487e-01 5.90345144e-01
-6.18501902e-01 -2.84010708e-01 -5.62092543e-01 5.94646215e-01
-1.84315801e+00 -1.19864285e+00 -9.19013377e-03 2.22266483e+00
4.46648508e-01 -9.16820019e-02 4.67183471e-01 2.67614633e-01
1.03409648e+00 -4.22002934e-02 -5.62491596e-01 3.39156657e-01
-1.17083699e-01 -1.57531813e-01 4.68369275e-01 2.53392339e-01
-1.17011356e+00 7.58753240e-01 4.97411537e+00 9.30396497e-01
-7.46953249e-01 1.81348145e-01 5.44512510e-01 7.55868182e-02
-5.12838699e-02 -3.14984351e-01 -1.04085588e+00 6.81096196e-01
3.87953490e-01 2.09769592e-01 1.81021199e-01 8.94775569e-01
9.16315019e-02 1.96440518e-01 -9.11161959e-01 1.49588084e+00
4.94351089e-01 -8.19710970e-01 3.43477577e-01 1.45001262e-01
8.13999891e-01 -5.32734752e-01 -3.21597122e-02 2.47077465e-01
2.81271804e-02 -9.77714360e-01 6.40591025e-01 7.93957651e-01
9.48001504e-01 -1.16928661e+00 1.07199788e+00 4.02551144e-01
-1.25602424e+00 2.76777968e-02 -4.14123267e-01 2.94762224e-01
3.99671286e-01 2.81996071e-01 -3.71265233e-01 1.02059627e+00
9.29850519e-01 6.16102278e-01 -7.53383040e-01 9.28695261e-01
-1.52917773e-01 2.39820302e-01 -1.47958323e-01 2.95029581e-01
-1.56502575e-01 -1.14274688e-01 5.72996080e-01 9.97671485e-01
2.82587171e-01 8.09641331e-02 2.88104534e-01 6.01190627e-01
1.41185671e-01 -1.14079021e-01 -4.35606480e-01 3.26001763e-01
2.48466119e-01 9.90553260e-01 -3.42407405e-01 -3.57247144e-01
-3.78423095e-01 1.39573789e+00 -1.09707043e-01 3.42548341e-01
-8.49241614e-01 -6.19886033e-02 4.45407867e-01 3.61183047e-01
2.85425305e-01 -1.29731715e-01 8.82019997e-02 -1.39359260e+00
5.65783754e-02 -1.17820179e+00 4.17250633e-01 -7.96006024e-01
-1.61927366e+00 5.19897461e-01 2.78793782e-01 -1.46056974e+00
-2.54419833e-01 -3.77077013e-01 -4.28450763e-01 9.75534499e-01
-1.14195514e+00 -1.90394092e+00 -8.33044231e-01 9.62918043e-01
5.71189106e-01 -6.68857157e-01 6.96390390e-01 5.83200276e-01
-6.46945715e-01 9.61001933e-01 -1.16119340e-01 6.18948221e-01
8.29117179e-01 -8.47010195e-01 4.26164716e-01 1.12392235e+00
-1.57985017e-01 7.13820279e-01 4.84350652e-01 -8.15495729e-01
-1.10723686e+00 -1.33599889e+00 4.37148690e-01 -3.05203617e-01
-2.46205896e-01 -2.36963183e-01 -4.28907305e-01 7.39120841e-01
-4.33576070e-02 1.14049546e-01 7.18536079e-01 -2.45307431e-01
-1.58356130e-01 -3.27953607e-01 -1.29667830e+00 6.65935814e-01
1.24248588e+00 -2.55259246e-01 -6.27615988e-01 2.94767153e-02
3.83401096e-01 -4.15927708e-01 -7.62198091e-01 6.74023986e-01
6.09930992e-01 -8.60690534e-01 1.43090618e+00 -4.67223287e-01
3.30641448e-01 -6.72403336e-01 -2.03126729e-01 -1.11438978e+00
-2.51495957e-01 -1.17140360e-01 8.10151268e-03 1.70084524e+00
-3.81795436e-01 -6.41363144e-01 1.09545946e+00 3.83694917e-01
1.00172885e-01 -5.14841318e-01 -7.53611982e-01 -6.90570176e-01
-4.52497788e-02 1.47441577e-04 9.29540038e-01 6.28296435e-01
-6.50723100e-01 2.44146407e-01 -7.63181865e-01 2.47119233e-01
9.60998356e-01 -2.20049843e-01 1.38230479e+00 -9.81727958e-01
-4.42828089e-01 5.82194664e-02 -6.40846431e-01 -1.16298938e+00
-2.51303930e-02 -4.89364356e-01 -2.13575751e-01 -1.57521486e+00
6.06820703e-01 -4.07897949e-01 -1.16308788e-02 2.69637883e-01
-3.90540123e-01 5.46862364e-01 3.89347345e-01 3.99957210e-01
-3.20728153e-01 8.14419925e-01 1.49500096e+00 -3.73610497e-01
1.96157470e-02 1.21542715e-01 -5.90570390e-01 6.65774941e-01
6.36621714e-01 -1.46762893e-01 -5.37380636e-01 -2.15383545e-01
-2.55870074e-01 -8.65079239e-02 9.39204276e-01 -1.41665304e+00
1.34494498e-01 4.19529602e-02 9.92878556e-01 -8.45753253e-01
5.32149434e-01 -7.58922577e-01 4.61605668e-01 4.11983460e-01
9.36047360e-02 4.40066829e-02 1.04869930e-02 7.08984196e-01
-1.48669884e-01 -1.07137851e-01 6.49874628e-01 -4.89346594e-01
-8.02672207e-01 5.43302119e-01 -5.18535972e-02 3.87247913e-02
1.05899966e+00 -4.56329793e-01 -1.32117659e-01 -4.66331840e-01
-5.98078549e-01 4.33353372e-02 5.83203793e-01 5.21370113e-01
6.69893205e-01 -1.71645844e+00 -8.03925157e-01 2.15422258e-01
1.63434848e-01 1.63311347e-01 5.51200151e-01 4.06374574e-01
-3.96013260e-01 2.77717024e-01 -4.66586620e-01 -4.19209987e-01
-1.33674204e+00 7.67544687e-01 4.40227151e-01 -3.75809759e-01
-6.96458578e-01 5.67472100e-01 3.03369880e-01 -6.25482082e-01
1.00307040e-01 4.96389836e-01 -3.73798013e-01 -2.47049943e-01
5.62783301e-01 4.30324286e-01 -2.39733472e-01 -1.07968259e+00
-3.85835737e-01 7.60521829e-01 -2.23009959e-02 -4.00133729e-01
1.13194358e+00 -1.13511078e-01 1.65585458e-01 -1.79400146e-02
1.03471601e+00 1.01444982e-01 -1.30690956e+00 -9.71416160e-02
-7.50461578e-01 -5.94905853e-01 -4.69301224e-01 -6.54563487e-01
-1.24116790e+00 7.27875650e-01 1.01892281e+00 -6.18515432e-01
1.09416091e+00 -2.22738996e-01 1.17572534e+00 -1.73773058e-02
5.36913872e-01 -1.07756948e+00 4.05378699e-01 1.18843839e-01
7.92531967e-01 -1.02909517e+00 1.28380423e-02 -4.69439328e-01
-6.71898127e-01 8.32153916e-01 1.07552195e+00 -1.85260162e-01
2.76848972e-01 -1.00229330e-01 -1.45010784e-01 2.05479622e-01
2.43645921e-01 -4.18530822e-01 3.07875007e-01 1.06814289e+00
-1.52397146e-02 7.87456578e-04 -2.51069516e-01 9.66386080e-01
-2.58334368e-01 -7.04892650e-02 4.93413918e-02 7.07967699e-01
-1.11803286e-01 -1.36399221e+00 -8.47597122e-01 1.07917599e-01
-2.21202046e-01 3.11474860e-01 -2.56246477e-01 7.29786456e-01
6.94797277e-01 8.17011774e-01 -1.81788906e-01 -6.81980371e-01
5.27384877e-01 -1.44048989e-01 7.00691402e-01 -2.21218303e-01
-2.61011809e-01 -5.79477996e-02 -5.65053932e-02 -1.51419595e-01
-5.87091506e-01 -5.98399401e-01 -8.10703635e-01 -3.72890383e-01
-1.40638143e-01 4.15850617e-02 1.55527577e-01 9.59692538e-01
6.78022802e-02 5.92056811e-01 4.51243669e-01 -1.05102110e+00
-3.10132504e-01 -1.03235102e+00 -2.46253014e-01 8.64103556e-01
-2.47240476e-02 -7.50805438e-01 -5.72506785e-02 3.03606004e-01] | [14.589067459106445, 0.8653116822242737] |
52acc169-c60d-4780-bea5-6d5e2d1c6bdf | semi-supervised-junction-tree-variational | 2208.05119 | null | https://arxiv.org/abs/2208.05119v5 | https://arxiv.org/pdf/2208.05119v5.pdf | Semi-Supervised Junction Tree Variational Autoencoder for Molecular Property Prediction | Molecular Representation Learning is essential to solving many drug discovery and computational chemistry problems. It is a challenging problem due to the complex structure of molecules and the vast chemical space. Graph representations of molecules are more expressive than traditional representations, such as molecular fingerprints. Therefore, they can improve the performance of machine learning models. We propose SeMole, a method that augments the Junction Tree Variational Autoencoders, a state-of-the-art generative model for molecular graphs, with semi-supervised learning. SeMole aims to improve the accuracy of molecular property prediction when having limited labeled data by exploiting unlabeled data. We enforce that the model generates molecular graphs conditioned on target properties by incorporating the property into the latent representation. We propose an additional pre-training phase to improve the training process for our semi-supervised generative model. We perform an experimental evaluation on the ZINC dataset using three different molecular properties and demonstrate the benefits of semi-supervision. | ['Tony Shen', 'Martin Ester', 'Atia Hamidizadeh'] | 2022-08-10 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 5.05408823e-01 1.18025936e-01 -7.12029576e-01 -4.21946466e-01
-5.89637160e-01 -5.50847173e-01 7.01814115e-01 2.27029935e-01
-1.59340259e-02 1.18596137e+00 2.33370334e-01 -5.51201940e-01
6.19411580e-02 -1.03081262e+00 -1.04752779e+00 -9.79341745e-01
1.32111967e-01 6.36468410e-01 -4.50052693e-02 7.52703324e-02
-4.26606350e-02 4.44870085e-01 -1.07006741e+00 4.58619624e-01
1.16255975e+00 4.97882664e-01 9.50683355e-02 1.97683409e-01
-1.04565583e-01 9.86433864e-01 -1.86129689e-01 -4.17648166e-01
-2.38714069e-01 -6.75531447e-01 -5.73947132e-01 9.62278694e-02
1.69932038e-01 5.07359058e-02 -2.85087734e-01 1.00866091e+00
3.68600845e-01 3.53975743e-01 1.21900380e+00 -9.74955022e-01
-9.34357107e-01 6.76160991e-01 -3.08284789e-01 -1.63755849e-01
2.38537431e-01 2.89203264e-02 1.17838633e+00 -7.78348386e-01
8.63291800e-01 1.28091216e+00 4.09950227e-01 6.51842833e-01
-1.74497867e+00 -7.81875849e-01 1.27750024e-01 1.99545100e-01
-1.32085562e+00 -3.19073409e-01 9.53356743e-01 -6.77581191e-01
9.93907273e-01 -1.66943535e-01 4.73574132e-01 1.30695462e+00
2.22383872e-01 5.12732446e-01 7.87543654e-01 -1.41585067e-01
4.48498160e-01 -1.87761690e-02 -7.60256778e-03 9.02263761e-01
2.89159894e-01 1.40713990e-01 -5.07899046e-01 -6.74203634e-01
6.93603575e-01 4.64949816e-01 -3.40092331e-01 -7.90270865e-01
-6.96201503e-01 1.38810980e+00 7.15626836e-01 9.84113365e-02
-5.13975561e-01 2.88842022e-01 2.97034848e-02 -7.58129433e-02
5.16934574e-01 4.39035594e-01 -2.71064043e-01 2.88526267e-01
-8.32622349e-01 -8.13089982e-02 6.75754666e-01 7.54919112e-01
9.00382161e-01 1.54514700e-01 -2.44719889e-02 7.12140143e-01
6.76929653e-01 2.55035162e-01 3.31133306e-01 -3.80011737e-01
2.19070390e-01 7.87181199e-01 -1.78912446e-01 -3.68850738e-01
-1.44993916e-01 -4.79773849e-01 -8.78739774e-01 -3.23158577e-02
4.89283986e-02 4.58385237e-02 -1.17552900e+00 1.75403357e+00
4.09726799e-01 2.60744154e-01 2.50277281e-01 4.24846262e-01
9.91055727e-01 9.29312408e-01 6.93197787e-01 -4.44714993e-01
8.97200644e-01 -8.37985098e-01 -7.15173304e-01 1.34522155e-01
7.69963801e-01 -4.56021547e-01 4.76876646e-01 8.93917978e-02
-5.28809845e-01 -4.21387941e-01 -1.01944029e+00 -6.20709173e-02
-4.29072976e-01 4.95556556e-02 1.14397740e+00 5.85765600e-01
-1.95044592e-01 1.19622266e+00 -1.06610966e+00 1.40893623e-01
7.02254295e-01 5.65013289e-01 -5.49977362e-01 -1.72579348e-01
-1.03719425e+00 5.01881182e-01 5.98455250e-01 -1.64665401e-01
-1.05650663e+00 -7.57500648e-01 -1.11333823e+00 2.60065317e-01
3.27432960e-01 -6.56408846e-01 8.30552220e-01 -5.98233044e-01
-1.61324930e+00 4.46318120e-01 -2.25549996e-01 -4.33850795e-01
1.71442047e-01 1.27063245e-01 -2.08130151e-01 1.70361567e-02
-5.78279495e-02 3.16497266e-01 7.18636394e-01 -1.12417459e+00
1.34464964e-01 -3.36279362e-01 -2.45884821e-01 -2.22201467e-01
5.11410926e-03 -5.51127732e-01 -5.27502298e-02 -4.87547845e-01
-1.32835269e-01 -9.30709243e-01 -6.33238673e-01 -3.50621760e-01
-6.38433874e-01 -3.07319105e-01 5.68807662e-01 -3.21981639e-01
1.01316524e+00 -2.01640081e+00 4.90005046e-01 4.79515105e-01
3.34011525e-01 2.85602719e-01 -2.59092927e-01 7.02060938e-01
-4.41402912e-01 2.46369556e-01 -4.32072937e-01 -1.89830735e-01
-1.91538334e-01 2.79977649e-01 -3.18587840e-01 4.02955353e-01
2.88417578e-01 1.02790201e+00 -1.00324166e+00 -3.58856887e-01
1.67107344e-01 9.04740751e-01 -6.61885798e-01 3.67752761e-01
-8.34534824e-01 6.81395531e-01 -7.59453833e-01 5.45765638e-01
5.48864901e-01 -8.56586993e-01 6.25110686e-01 -2.32375816e-01
2.31582984e-01 2.78568298e-01 -7.89665699e-01 1.74115217e+00
-2.07113177e-01 -9.09676775e-02 -6.32530153e-01 -9.83996034e-01
9.17113662e-01 4.58144069e-01 5.26301801e-01 -2.95025110e-01
5.98467253e-02 1.39035597e-01 1.96853861e-01 -1.48083374e-01
-5.35157286e-02 -3.51642519e-01 3.59860420e-01 3.36357623e-01
5.23876965e-01 -1.35195469e-02 2.82416523e-01 2.69418657e-01
6.65201247e-01 5.39871275e-01 4.68360692e-01 6.53814003e-02
4.99176592e-01 -2.47879148e-01 5.71982145e-01 3.00198138e-01
4.52457041e-01 2.08281964e-01 5.79387069e-01 -3.62224638e-01
-8.64093542e-01 -8.65791023e-01 -2.42008597e-01 8.96206677e-01
-1.80508047e-01 -8.66773784e-01 -6.23086214e-01 -8.26319754e-01
7.94398934e-02 6.61670804e-01 -7.77703345e-01 -5.00640690e-01
-2.63633858e-02 -1.08178937e+00 1.13592789e-01 4.63611871e-01
8.72891545e-02 -9.48212743e-01 1.75245330e-01 4.61191297e-01
2.76964843e-01 -9.83260155e-01 -2.00947016e-01 5.23900628e-01
-9.46464360e-01 -1.33037031e+00 -6.04756057e-01 -5.22332847e-01
8.71184230e-01 -7.11733475e-02 9.35278237e-01 -5.81658147e-02
-1.78159341e-01 -2.31816113e-01 -2.38618538e-01 -2.86660373e-01
-8.34190667e-01 1.01950437e-01 -2.61190515e-02 1.55664146e-01
2.85617679e-01 -8.67835522e-01 -5.27635396e-01 2.23044143e-03
-9.30710435e-01 1.24033190e-01 6.72777057e-01 1.11691833e+00
1.14000726e+00 -1.21643290e-01 4.47288722e-01 -1.59883761e+00
4.30721134e-01 -4.45348978e-01 -6.71995699e-01 2.99006611e-01
-6.72727644e-01 8.53118658e-01 7.74331152e-01 -6.20887935e-01
-8.88983965e-01 3.90242875e-01 -2.79412299e-01 -4.44888890e-01
2.68727564e-03 9.00267422e-01 -3.92364353e-01 -1.51442200e-01
6.77765071e-01 1.88736722e-01 -1.78933278e-01 -6.13280833e-01
4.81808335e-01 4.73009646e-01 1.15121931e-01 -6.90453351e-01
6.70029998e-01 2.00357810e-01 5.65473855e-01 -7.33921051e-01
-9.01020169e-01 -4.21508998e-01 -7.02303290e-01 3.49806607e-01
8.22909594e-01 -1.00593102e+00 -7.98599303e-01 -3.73837054e-02
-9.99425411e-01 -3.04204285e-01 -2.37263609e-02 6.25835001e-01
-5.67061782e-01 5.86966097e-01 -4.74174351e-01 -6.51873887e-01
-5.31253576e-01 -1.14983237e+00 1.02488410e+00 6.56287521e-02
-2.07100459e-03 -1.15174365e+00 5.17486632e-01 1.73416346e-01
-6.34097904e-02 5.68286836e-01 1.36064696e+00 -7.73277640e-01
-5.73402703e-01 -1.38500869e-01 9.45707709e-02 1.38999894e-01
5.05285144e-01 6.80909008e-02 -1.04014289e+00 -3.19126874e-01
-5.08136749e-01 -4.32808071e-01 1.29953253e+00 2.71238208e-01
1.30275691e+00 -2.42160454e-01 -5.91166258e-01 7.66096771e-01
1.35313535e+00 3.81440848e-01 5.06441414e-01 -2.72698671e-01
1.08321774e+00 2.79027462e-01 2.32638329e-01 3.93669993e-01
-1.58596322e-01 5.45772552e-01 4.12744462e-01 -1.97027519e-01
1.24096572e-01 -9.14340258e-01 3.57574224e-01 6.60211921e-01
-3.93118143e-01 -3.94707531e-01 -5.90017498e-01 -5.59190921e-02
-2.03848934e+00 -1.15225637e+00 -5.98969981e-02 2.22020912e+00
1.12518036e+00 -1.01085462e-01 3.43934484e-02 -4.33623828e-02
4.73276913e-01 1.03900075e-01 -7.76433945e-01 -1.04719490e-01
1.24677181e-01 6.16300464e-01 3.48195612e-01 6.57585204e-01
-9.83341455e-01 1.18723297e+00 6.21372032e+00 8.02750230e-01
-1.18881285e+00 -8.90262797e-02 5.43158770e-01 4.19204265e-01
-5.79649389e-01 1.08821318e-01 -7.75795043e-01 4.89693314e-01
9.78195608e-01 -1.51955001e-02 2.88954049e-01 1.03195274e+00
-3.48014571e-02 3.21251154e-01 -1.50468957e+00 9.55175579e-01
-1.44952061e-02 -1.82313907e+00 6.17019892e-01 2.32533321e-01
8.11062813e-01 -7.86098242e-02 3.17393728e-02 1.28827527e-01
5.03178000e-01 -1.39746130e+00 1.45506054e-01 3.77623767e-01
8.08823228e-01 -8.70557308e-01 4.26588446e-01 3.64550501e-01
-1.00638759e+00 4.02413279e-01 -5.62598944e-01 2.69083083e-01
-1.06611699e-02 4.73595291e-01 -1.02381790e+00 6.75377011e-01
-1.71749294e-01 1.16758573e+00 -4.31820720e-01 8.34199786e-01
-6.25526249e-01 7.38957763e-01 -6.20028339e-02 -1.69216588e-01
1.50378615e-01 -7.18223870e-01 8.61916170e-02 8.93797696e-01
2.88617387e-02 -8.67796093e-02 3.71771783e-01 1.31593442e+00
-3.99208099e-01 1.91072881e-01 -8.17511499e-01 -7.61601567e-01
1.33360773e-01 9.36397970e-01 -4.29945648e-01 -3.07268977e-01
-3.57292056e-01 8.11138451e-01 4.15788263e-01 7.08825350e-01
-8.02474976e-01 -4.46121134e-02 4.33278322e-01 3.11091207e-02
3.56257796e-01 -2.07808390e-01 2.95007318e-01 -1.23118937e+00
-4.76660997e-01 -8.06543767e-01 2.42731825e-01 -4.99370486e-01
-1.24334657e+00 5.60274601e-01 -2.79060245e-01 -1.07809806e+00
-3.68142933e-01 -9.02523816e-01 -4.75725710e-01 8.26729774e-01
-1.59057271e+00 -1.27060652e+00 5.87658510e-02 4.43672270e-01
1.86003089e-01 -2.41521910e-01 1.27064168e+00 1.34990573e-01
-7.26785004e-01 3.67798746e-01 4.29261655e-01 -1.82928145e-02
5.58357418e-01 -1.24416757e+00 1.31820053e-01 4.30048883e-01
5.93102276e-01 9.25497293e-01 5.51824331e-01 -8.34513724e-01
-1.40008378e+00 -1.27499712e+00 4.97591466e-01 -3.42014223e-01
6.05791867e-01 -5.76291263e-01 -9.79244351e-01 7.22148240e-01
-1.69791505e-01 1.89295188e-01 1.58815241e+00 3.36051196e-01
-5.97088516e-01 3.97827476e-01 -8.44549835e-01 2.74977893e-01
9.21175301e-01 -8.43281329e-01 -2.50840485e-01 6.76642895e-01
7.15128958e-01 -2.35493198e-01 -1.18309259e+00 2.77196914e-01
3.56757581e-01 -4.66659337e-01 1.01775289e+00 -1.19076967e+00
5.36555469e-01 -3.35127026e-01 -9.07743722e-02 -1.40066385e+00
-4.58688080e-01 -6.35746777e-01 -5.12958944e-01 9.62800860e-01
7.59963036e-01 -5.08310199e-01 1.04352903e+00 4.91870344e-01
1.31042793e-01 -7.81280577e-01 -6.70067847e-01 -6.28672719e-01
9.97015610e-02 5.94149902e-02 5.80918133e-01 9.39470351e-01
-4.99671150e-04 1.00909328e+00 -5.16007960e-01 4.87245061e-02
6.19345605e-01 5.35722971e-01 7.22406626e-01 -1.45782220e+00
-7.91570961e-01 -5.36981001e-02 -5.15458703e-01 -9.06688452e-01
4.64064568e-01 -1.18144131e+00 -4.28339511e-01 -1.55291450e+00
5.35505176e-01 -1.98121220e-01 -4.10510838e-01 6.69671237e-01
-3.17909986e-01 9.66230175e-04 -2.23597080e-01 1.65551886e-01
-5.09261429e-01 1.04094613e+00 1.48673451e+00 -5.39558768e-01
-2.49405116e-01 5.41117750e-02 -5.75908005e-01 4.09385771e-01
6.35190904e-01 -5.49336672e-01 -6.75809622e-01 2.32690468e-01
2.42296636e-01 1.04166456e-02 7.86602050e-02 -6.40779614e-01
-1.50614470e-01 -3.33329499e-01 5.78324735e-01 -3.51429731e-01
4.21286136e-01 -6.38056874e-01 4.06666636e-01 6.01890981e-01
-3.12414557e-01 -5.38695157e-01 -2.89350022e-02 9.93583381e-01
-3.44726712e-01 -1.49380490e-01 7.77735770e-01 -4.49083261e-02
-3.55790585e-01 9.10075784e-01 -6.99704960e-02 -3.58874470e-01
8.19368422e-01 1.05761349e-01 -2.18110189e-01 -2.21832648e-01
-8.91628504e-01 -3.20913829e-03 5.31865060e-01 5.25607541e-02
6.57271564e-01 -1.33847010e+00 -2.80185938e-01 2.65203685e-01
3.64914149e-01 4.05864343e-02 -9.11565796e-02 4.35299098e-01
-3.61110270e-01 4.71356064e-01 -9.72884148e-02 -3.78416896e-01
-1.28779149e+00 1.00526297e+00 3.50017160e-01 -4.40274417e-01
-2.52759397e-01 5.59208035e-01 5.04832745e-01 -2.95282990e-01
3.16917710e-02 -1.72360793e-01 -3.61335874e-01 -1.87089577e-01
2.76939183e-01 -3.08830570e-02 -8.05352554e-02 -5.42975307e-01
-2.26251900e-01 4.53960180e-01 -3.66885722e-01 4.12344366e-01
1.64322579e+00 7.31892645e-01 -1.91319082e-02 2.25556821e-01
1.18311584e+00 -1.03840679e-02 -1.24646294e+00 -3.28515768e-01
-2.09672898e-01 -9.61471871e-02 7.89984763e-02 -6.09706044e-01
-7.30114281e-01 1.05726659e+00 3.18159938e-01 -2.33913019e-01
3.55497032e-01 5.51786162e-02 3.86822253e-01 7.52204776e-01
2.77507067e-01 -6.89329743e-01 2.14576825e-01 2.06782907e-01
6.44800186e-01 -1.21888006e+00 3.86672497e-01 -8.66989613e-01
-4.35892195e-01 1.03532517e+00 3.26195776e-01 -3.12592797e-02
5.07164299e-01 -1.62159413e-01 -2.98691660e-01 -4.22934234e-01
-6.22056782e-01 -1.82807967e-01 5.65004230e-01 5.88448048e-01
9.49399233e-01 1.72482580e-01 -9.65099856e-02 4.81995493e-01
2.03307942e-01 3.74491587e-02 6.64068535e-02 8.95488083e-01
-1.98492453e-01 -1.83458912e+00 1.34776220e-01 4.00796115e-01
-3.73041987e-01 -3.38977098e-01 -7.93482065e-01 3.28557193e-01
-1.52189946e-02 6.94049776e-01 -4.25187498e-01 -2.74136186e-01
-8.11634436e-02 2.25643143e-01 7.69155502e-01 -9.54938829e-01
-3.58632989e-02 2.54759341e-01 -3.13341357e-02 -3.91138405e-01
-6.54259086e-01 -2.22974896e-01 -1.36520970e+00 -1.78785637e-01
-6.67898536e-01 7.43128717e-01 4.84897465e-01 8.08697939e-01
6.08146429e-01 6.69131219e-01 6.77866578e-01 -6.14700675e-01
-4.90619957e-01 -8.36673021e-01 -6.14849091e-01 6.05707288e-01
1.86163068e-01 -9.58287120e-01 1.99635699e-02 2.18258172e-01] | [5.134910583496094, 5.9044976234436035] |
3c3019be-2956-417a-88f6-0e45cedd8c37 | few-shot-learning-of-new-sound-classes-for | 2106.07144 | null | https://arxiv.org/abs/2106.07144v1 | https://arxiv.org/pdf/2106.07144v1.pdf | Few-shot learning of new sound classes for target sound extraction | Target sound extraction consists of extracting the sound of a target acoustic event (AE) class from a mixture of AE sounds. It can be realized using a neural network that extracts the target sound conditioned on a 1-hot vector that represents the desired AE class. With this approach, embedding vectors associated with the AE classes are directly optimized for the extraction of sound classes seen during training. However, it is not easy to extend this framework to new AE classes, i.e. unseen during training. Recently, speech, music, or AE sound extraction based on enrollment audio of the desired sound offers the potential of extracting any target sound in a mixture given only a short audio signal of a similar sound. In this work, we propose combining 1-hot- and enrollment-based target sound extraction, allowing optimal performance for seen AE classes and simple extension to new classes. In experiments with synthesized sound mixtures generated with the Freesound Dataset (FSD) datasets, we demonstrate the benefit of the combined framework for both seen and new AE classes. Besides, we also propose adapting the embedding vectors obtained from a few enrollment audio samples (few-shot) to further improve performance on new classes. | ['Shoko Araki', 'Keisuke Kinoshita', 'Tsubasa Ochiai', 'Jorge Bennasar Vázquez', 'Marc Delcroix'] | 2021-06-14 | null | null | null | null | ['target-sound-extraction'] | ['audio'] | [ 4.22216147e-01 4.01920918e-03 3.13331127e-01 -1.19973995e-01
-1.21368611e+00 -5.87042272e-01 3.52882057e-01 1.58790886e-01
-3.79140556e-01 2.78279990e-01 2.42781758e-01 5.15798032e-02
-2.80902889e-02 -6.95763767e-01 -5.39403856e-01 -7.24997282e-01
-2.45140582e-01 6.55668750e-02 2.13062793e-01 6.86527416e-02
-2.04970092e-01 6.00831509e-01 -2.06751990e+00 4.43167686e-01
4.25942034e-01 1.05846417e+00 4.98542011e-01 1.19616640e+00
-9.26158652e-02 3.91696572e-01 -1.13221395e+00 1.92926154e-02
3.22648466e-01 -6.23504758e-01 -3.24395537e-01 -9.45017338e-02
7.10875154e-01 -3.62781167e-01 -3.07549536e-01 6.68819606e-01
8.45024705e-01 3.42232883e-01 7.82210588e-01 -1.17956948e+00
4.90024649e-02 9.40753222e-01 -5.78722209e-02 2.84096926e-01
2.37574816e-01 -4.69261855e-02 1.24204385e+00 -1.23388791e+00
2.12351978e-01 1.00003624e+00 4.69823927e-01 5.53486407e-01
-1.08573782e+00 -1.06715167e+00 -8.99771526e-02 2.32074991e-01
-1.31498015e+00 -9.22676980e-01 1.28229499e+00 -2.83439517e-01
8.66366744e-01 4.66347784e-01 6.69602752e-01 1.16785634e+00
-3.70556355e-01 1.00165069e+00 6.37904704e-01 -6.34139776e-01
4.39099729e-01 3.02499592e-01 1.19859263e-01 1.18576236e-01
-1.28591523e-01 2.08219007e-01 -7.98237145e-01 -3.14264983e-01
4.71827239e-02 -3.25844049e-01 -6.69832110e-01 -4.49834466e-02
-8.87776971e-01 6.81141913e-01 5.80017194e-02 4.08567786e-01
-4.97507393e-01 1.17459394e-01 1.26450598e-01 2.92337745e-01
3.17788959e-01 4.42071676e-01 -4.36189860e-01 -3.42375159e-01
-1.32495987e+00 3.62306297e-01 1.02618289e+00 6.12833977e-01
7.21902013e-01 6.41311407e-01 -1.22206934e-01 1.03934169e+00
2.60177970e-01 6.85211301e-01 4.01075870e-01 -7.76705980e-01
4.76726919e-01 -3.34431268e-02 7.71699622e-02 -8.18928182e-01
-3.05259198e-01 -8.17091882e-01 -4.74518389e-01 3.38955253e-01
3.62636238e-01 -2.44024038e-01 -9.84155715e-01 2.01248932e+00
4.97763872e-01 6.79257631e-01 3.70395899e-01 5.47005653e-01
8.70742977e-01 1.06114113e+00 -2.80110657e-01 -3.45022470e-01
1.11405551e+00 -8.66738021e-01 -6.73125207e-01 -3.55327457e-01
1.29391193e-01 -8.46140504e-01 9.92536664e-01 6.83109701e-01
-8.37971270e-01 -9.01332319e-01 -1.18620420e+00 3.94228309e-01
-2.48105094e-01 2.88030565e-01 8.45371559e-02 9.34676886e-01
-5.73523879e-01 4.95948344e-01 -5.91571450e-01 3.13381813e-02
1.59896523e-01 3.11819822e-01 -3.24493311e-02 1.59186795e-01
-1.17551434e+00 2.85200894e-01 2.35297829e-01 -9.35421437e-02
-1.17893100e+00 -1.06742108e+00 -7.49622405e-01 2.88485676e-01
2.54857540e-01 -1.94776312e-01 1.28870952e+00 -7.99861491e-01
-1.59315944e+00 2.26169139e-01 -9.47487131e-02 -4.45448548e-01
6.47764876e-02 -4.44614053e-01 -8.11143756e-01 4.78463560e-01
-1.12774886e-01 4.49123472e-01 1.41188693e+00 -1.43074894e+00
-8.67556334e-01 -2.65622903e-02 -2.62661308e-01 1.97316110e-01
-8.24757218e-01 -1.26204282e-01 -9.70074013e-02 -8.31641674e-01
-7.34554231e-02 -9.57730353e-01 1.48486599e-01 -2.66849846e-01
-4.40474272e-01 -8.38220492e-02 1.05791605e+00 -6.89870775e-01
1.35398030e+00 -2.77533150e+00 -8.58776495e-02 2.71567434e-01
1.07371379e-02 3.18972111e-01 -4.93937701e-01 3.36617917e-01
-2.13852763e-01 -2.66309649e-01 -1.56696543e-01 -5.16691983e-01
2.23342344e-01 -7.03473762e-02 -7.08112955e-01 -5.90844126e-03
4.86837059e-01 1.33026853e-01 -8.81097853e-01 -2.63934314e-01
6.41988963e-02 6.73961222e-01 -6.73929811e-01 5.84871769e-01
5.79028688e-02 2.08759218e-01 -1.63630888e-01 3.46311092e-01
6.10634863e-01 4.29948628e-01 -2.01454967e-01 -3.80331427e-01
1.47115439e-01 5.40895045e-01 -1.76897418e+00 1.39540446e+00
-7.48290002e-01 5.29234469e-01 2.05236733e-01 -5.72468996e-01
1.11497712e+00 7.37511754e-01 4.37879622e-01 -4.93418984e-02
-1.31880149e-01 4.44032967e-01 3.70267779e-01 -2.97491282e-01
4.22981262e-01 -2.47534677e-01 -4.99021495e-03 5.56320071e-01
5.28798223e-01 -3.84344906e-01 1.71122536e-01 -1.74367279e-01
1.11519039e+00 -1.91802397e-01 1.10568956e-01 2.67226815e-01
4.23721582e-01 -6.28134549e-01 3.93489242e-01 8.88591647e-01
-5.46823144e-02 8.22375119e-01 -1.58559635e-01 2.01097295e-01
-7.44428992e-01 -1.37108696e+00 -1.28680512e-01 1.12718141e+00
-4.03925627e-01 -6.51124060e-01 -6.62594378e-01 -6.42025232e-01
-2.05413878e-01 9.62305665e-01 -2.03168854e-01 -3.48254114e-01
-7.91713059e-01 -6.13314152e-01 7.97683835e-01 3.11616927e-01
2.71271653e-02 -1.17071486e+00 -6.97946131e-01 6.53355420e-01
-2.49154165e-01 -1.11967969e+00 -4.61644202e-01 5.97097456e-01
-5.01623511e-01 -6.77600801e-01 -8.05236280e-01 -8.50655437e-01
5.22625148e-02 1.04363531e-01 8.25043857e-01 -4.70423907e-01
-1.81732759e-01 5.28526485e-01 -4.54739124e-01 -7.60901332e-01
-8.22751105e-01 1.01452656e-01 3.84453684e-01 4.61082578e-01
1.14367351e-01 -9.47898149e-01 -3.10460657e-01 -5.30963652e-02
-8.58944416e-01 -2.84484535e-01 4.46701616e-01 6.56026721e-01
4.44775075e-01 4.71131027e-01 8.96777093e-01 -4.21085656e-01
5.63697219e-01 -4.59563851e-01 -1.09285273e-01 -1.11923814e-01
-2.17234373e-01 -2.04585522e-01 8.63211393e-01 -1.04273498e+00
-9.47388768e-01 1.72101483e-01 -5.23947239e-01 -7.08274245e-01
-4.80071962e-01 3.33852172e-01 -3.36392283e-01 4.41764712e-01
7.31965482e-01 3.01218659e-01 -4.51499283e-01 -8.32161188e-01
2.06891879e-01 1.07211435e+00 5.27883291e-01 -4.54267889e-01
1.04779518e+00 9.14502367e-02 -2.22414196e-01 -1.16805243e+00
-7.21825838e-01 -5.61764956e-01 -3.59968871e-01 -1.45632058e-01
5.72827876e-01 -9.16687787e-01 -2.31275707e-01 5.52390039e-01
-9.75382328e-01 -7.79536441e-02 -7.19066918e-01 9.47737277e-01
-3.69443536e-01 1.26744732e-01 -3.87918860e-01 -1.14316154e+00
-4.67746437e-01 -1.08182430e+00 1.10021937e+00 1.67463943e-01
-4.60915029e-01 -5.03562212e-01 3.88707340e-01 -1.54629663e-01
2.93804675e-01 -1.88787624e-01 7.66784906e-01 -1.15270281e+00
-2.87204742e-01 -4.11223590e-01 6.20006740e-01 6.85997605e-01
6.32960379e-01 4.79096845e-02 -1.67569554e+00 -4.04467523e-01
2.33680561e-01 -2.03583017e-01 8.42496097e-01 1.64380848e-01
8.09055686e-01 -3.71323466e-01 -1.29654467e-01 3.60386401e-01
1.02145207e+00 6.17876589e-01 1.80298477e-01 -1.35261133e-01
4.91803795e-01 4.02043045e-01 5.12174249e-01 4.16200101e-01
-1.81624636e-01 8.85162354e-01 5.12353592e-02 6.24538586e-02
-5.39989412e-01 -2.88983613e-01 6.68443799e-01 1.33493578e+00
3.85025293e-01 -3.99592817e-01 -7.19613612e-01 7.01215386e-01
-1.09392762e+00 -9.85078216e-01 3.47448140e-01 2.26958680e+00
1.11554229e+00 2.19665542e-01 1.29685357e-01 8.34888101e-01
6.41718924e-01 1.81898713e-01 -4.26891655e-01 -8.83308724e-02
4.55178991e-02 9.65240598e-01 -3.42727751e-01 4.39735472e-01
-1.15272832e+00 4.06236321e-01 5.59188271e+00 1.03728116e+00
-1.22190046e+00 1.59207717e-01 1.22515187e-01 -2.68601000e-01
-2.29341745e-01 -1.77119315e-01 -1.06909001e+00 3.95726800e-01
1.28022730e+00 -1.41750500e-01 3.94186646e-01 7.21059442e-01
9.45354719e-03 2.10051432e-01 -1.36473823e+00 1.09224081e+00
6.87506348e-02 -8.83624911e-01 1.57332078e-01 -1.77710339e-01
3.52800518e-01 -2.61552274e-01 2.86555827e-01 5.60176194e-01
-2.14191601e-01 -5.96946478e-01 8.95820558e-01 2.03487977e-01
6.73095047e-01 -8.95633101e-01 4.09215182e-01 3.28139365e-01
-1.38344932e+00 -2.49564558e-01 -1.56107932e-01 3.20230573e-01
-4.22944501e-02 6.20249927e-01 -1.39969492e+00 5.81007302e-01
7.61331618e-01 4.10468131e-01 -4.70979750e-01 1.12623405e+00
-4.91000116e-02 1.25165653e+00 -7.05066800e-01 1.03851967e-01
-1.96307730e-02 5.57620600e-02 1.08861816e+00 1.35549819e+00
6.79230630e-01 -2.08739668e-01 -1.22741591e-02 7.37688959e-01
-4.52511907e-02 2.17973694e-01 -5.96584439e-01 -2.03309879e-01
7.20861375e-01 1.25783253e+00 -4.97443765e-01 -2.60498166e-01
-1.85952291e-01 7.07765758e-01 -1.52551487e-01 4.85780776e-01
-7.10135698e-01 -8.69965494e-01 7.85755634e-01 -2.08002478e-01
7.03330398e-01 -1.36981010e-02 2.70514518e-01 -9.35469866e-01
-2.00849827e-02 -1.02364528e+00 3.26739699e-01 -6.32378340e-01
-1.08928180e+00 8.85890186e-01 -1.32822543e-01 -1.64404070e+00
-6.06063604e-01 -3.49058777e-01 -7.13556111e-01 9.13094044e-01
-1.28549457e+00 -8.53577614e-01 -9.40890461e-02 4.44621116e-01
6.26035690e-01 -3.07986796e-01 1.15756357e+00 4.89273727e-01
-2.37308547e-01 8.14220190e-01 5.03920838e-02 7.21845552e-02
7.36132205e-01 -1.25027633e+00 1.82909414e-01 7.16219366e-01
8.78171623e-01 2.30287492e-01 7.61592805e-01 -2.60899365e-01
-1.00654733e+00 -1.22411239e+00 7.59356380e-01 -2.24998429e-01
4.14592266e-01 -7.21943915e-01 -9.45346117e-01 2.73487896e-01
2.02854332e-02 -1.57517031e-01 1.00554204e+00 -1.09232806e-01
-1.46545365e-01 -5.16756415e-01 -7.57213295e-01 5.97635448e-01
6.03514075e-01 -7.46113181e-01 -8.33315432e-01 -5.53717315e-02
8.66614819e-01 -1.15126587e-01 -7.84430623e-01 4.25749540e-01
3.77326399e-01 -6.39330029e-01 1.12581551e+00 -4.20180380e-01
5.93448132e-02 -4.81363326e-01 -4.48403865e-01 -1.53095996e+00
-2.04351079e-03 -6.25127673e-01 -5.46009362e-01 1.68543696e+00
4.79779333e-01 -2.17376992e-01 4.93631631e-01 -5.02791069e-02
-2.69851953e-01 -4.48094338e-01 -9.52419758e-01 -8.98829103e-01
-1.49042755e-01 -8.82351935e-01 7.07872152e-01 7.49279380e-01
-2.94024110e-01 4.98675466e-01 -4.91855741e-01 6.01323128e-01
5.62517583e-01 2.76604861e-01 9.53696132e-01 -1.09927034e+00
-8.92560482e-01 -1.67323679e-01 -3.46979767e-01 -9.44171429e-01
1.00595400e-01 -8.79682183e-01 3.08401078e-01 -1.05477059e+00
-2.39137083e-01 -4.23045516e-01 -6.31533265e-01 4.02852029e-01
-3.40029448e-01 2.80476511e-01 4.33828473e-01 -9.11511779e-02
7.26964101e-02 7.36478269e-01 9.07167137e-01 -3.44902098e-01
-5.34995019e-01 5.48382699e-01 -2.83068925e-01 6.81190729e-01
6.74681544e-01 -9.01065350e-01 -5.84646642e-01 9.71518829e-02
-1.28737360e-01 1.99439034e-01 2.35744417e-01 -1.40532303e+00
7.41258636e-02 3.24318886e-01 2.45044321e-01 -7.29531288e-01
8.74476373e-01 -9.30188298e-01 3.18131596e-01 2.73264468e-01
-3.73233914e-01 -6.74829423e-01 4.19574320e-01 5.39441109e-01
-4.22142357e-01 -6.06557131e-01 6.19635701e-01 2.61722505e-01
-3.05572987e-01 6.45575821e-02 -4.04277325e-01 9.00843590e-02
6.36533439e-01 -1.36015758e-01 2.52314121e-01 -6.00357413e-01
-1.05868936e+00 -4.93596226e-01 -2.86619365e-01 5.20811856e-01
8.14330161e-01 -1.35789192e+00 -8.00413728e-01 4.71764714e-01
-1.08520379e-02 -9.60903838e-02 1.60917684e-01 4.07230526e-01
1.52638286e-01 -2.74726655e-02 1.43203977e-02 -5.31682134e-01
-1.41311991e+00 4.09606099e-01 3.07580262e-01 -1.99171621e-02
-5.42906284e-01 8.59628558e-01 4.92189735e-01 -4.08998877e-01
4.68201131e-01 -3.87827188e-01 -4.11110610e-01 2.19113484e-01
6.57842755e-01 2.50430882e-01 2.34508038e-01 -4.79357213e-01
-2.30637804e-01 3.33448201e-01 1.29541352e-01 -4.25982445e-01
1.49117184e+00 4.73585576e-02 3.91229004e-01 1.12507915e+00
1.43075383e+00 7.25746870e-01 -1.01309860e+00 -1.98808149e-01
-1.45813540e-01 -2.79165000e-01 -1.47570102e-02 -5.95332980e-01
-9.75215614e-01 1.15383172e+00 9.17859077e-01 2.28305086e-01
1.34684515e+00 -2.83134934e-02 9.21282709e-01 3.85366082e-01
1.60583168e-01 -7.36718714e-01 2.63331622e-01 2.05766961e-01
8.72942924e-01 -5.53418636e-01 -4.26881284e-01 -2.56512225e-01
-2.27694780e-01 1.42370713e+00 4.42229450e-01 1.02490626e-01
1.00601232e+00 4.94024247e-01 2.07459211e-01 1.04693882e-01
-6.02965951e-01 -1.93890065e-01 5.51194668e-01 4.51351762e-01
3.32911722e-02 -1.19660608e-01 4.70788985e-01 9.21297967e-01
-5.68601489e-01 -5.07197976e-01 6.08095288e-01 9.15586948e-01
-5.39530933e-01 -1.08283544e+00 -5.33950925e-01 5.67361534e-01
-5.21459460e-01 -3.23131949e-01 -2.49908686e-01 5.24111688e-01
1.37592703e-01 1.17628872e+00 1.01904437e-01 -5.25138199e-01
5.11349857e-01 5.75552940e-01 2.84706384e-01 -8.21822584e-01
-5.43258727e-01 5.25033593e-01 1.32196903e-01 9.57544334e-03
-2.30736017e-01 -5.97581983e-01 -1.08657253e+00 4.78547305e-01
-6.15378678e-01 4.65815604e-01 4.83699560e-01 6.37510538e-01
1.07165091e-01 7.88450301e-01 1.02584767e+00 -9.35118973e-01
-6.85892880e-01 -1.04223597e+00 -8.14737618e-01 4.48107749e-01
7.06394494e-01 -4.56770718e-01 -8.45892787e-01 3.68689239e-01] | [15.217419624328613, 5.395630836486816] |
06fb6197-2117-43af-bf36-17b0b034866c | deep-crisp-boundaries-from-boundaries-to | 1801.02439 | null | http://arxiv.org/abs/1801.02439v3 | http://arxiv.org/pdf/1801.02439v3.pdf | Deep Crisp Boundaries: From Boundaries to Higher-level Tasks | Edge detection has made significant progress with the help of deep
Convolutional Networks (ConvNet). These ConvNet based edge detectors have
approached human level performance on standard benchmarks. We provide a
systematical study of these detectors' outputs. We show that the detection
results did not accurately localize edge pixels, which can be adversarial for
tasks that require crisp edge inputs. As a remedy, we propose a novel
refinement architecture to address the challenging problem of learning a crisp
edge detector using ConvNet. Our method leverages a top-down backward
refinement pathway, and progressively increases the resolution of feature maps
to generate crisp edges. Our results achieve superior performance, surpassing
human accuracy when using standard criteria on BSDS500, and largely
outperforming state-of-the-art methods when using more strict criteria. More
importantly, we demonstrate the benefit of crisp edge maps for several
important applications in computer vision, including optical flow estimation,
object proposal generation and semantic segmentation. | ['Yupei Wang', 'Xin Zhao', 'Kaiqi Huang', 'Yin Li'] | 2018-01-08 | null | null | null | null | ['object-proposal-generation'] | ['computer-vision'] | [ 1.42318025e-01 5.07344902e-02 9.35831442e-02 -3.53413634e-02
-3.01027268e-01 -6.16113484e-01 5.22011280e-01 -2.28137061e-01
-6.70029104e-01 6.01529360e-01 -8.52982923e-02 -4.32687372e-01
3.81251812e-01 -7.19011605e-01 -7.36283720e-01 -2.49424219e-01
-2.15332825e-02 3.60643528e-02 5.72206020e-01 -2.32236564e-01
3.76232177e-01 5.79267502e-01 -1.30108035e+00 1.46762565e-01
7.48164833e-01 9.97006893e-01 -2.95755178e-01 9.04005229e-01
4.44823802e-02 6.26357615e-01 -4.76426840e-01 -4.66707230e-01
6.65678978e-01 -4.19922143e-01 -8.11896086e-01 -1.15784824e-01
9.64757919e-01 -5.49667418e-01 -5.55209458e-01 1.26247561e+00
4.09504652e-01 2.43975595e-01 5.62255383e-01 -1.40490472e+00
-9.79662001e-01 2.87914902e-01 -6.73898637e-01 4.43462461e-01
2.46620774e-01 4.56567377e-01 1.05588186e+00 -1.01165211e+00
8.93376827e-01 1.10196316e+00 1.02156222e+00 7.80803382e-01
-1.22823870e+00 -6.15593970e-01 1.86684653e-01 1.93376318e-01
-1.27801657e+00 -4.35686499e-01 4.27027225e-01 -4.32861477e-01
1.01357365e+00 -1.07572936e-01 6.75655186e-01 9.50443745e-01
1.81886673e-01 8.45401764e-01 8.40244830e-01 -1.63985312e-01
1.69612527e-01 -1.88510776e-01 -4.86635603e-02 1.03812122e+00
6.27299368e-01 5.09492218e-01 -3.80585641e-01 3.24024886e-01
1.15657282e+00 -1.66356489e-01 -3.97851855e-01 -1.76073566e-01
-1.06465971e+00 7.25062549e-01 1.02840960e+00 4.16725554e-04
-2.85266608e-01 5.29345810e-01 2.34805465e-01 1.83410764e-01
1.67929813e-01 7.04572499e-01 -3.68196338e-01 4.16119657e-02
-1.13304663e+00 1.79718927e-01 7.69202888e-01 1.09480059e+00
6.17238760e-01 2.05460504e-01 -5.35348952e-01 3.92308570e-02
6.10201173e-02 1.46158159e-01 1.43779665e-02 -1.33081913e+00
1.37846902e-01 4.71116573e-01 3.16935211e-01 -8.86703193e-01
-4.37418610e-01 -5.25635719e-01 -7.60711849e-01 9.74838018e-01
7.90092647e-01 -3.37319016e-01 -1.36550665e+00 1.48814642e+00
2.02085733e-01 6.49866343e-01 -9.29222405e-02 1.21163929e+00
1.00391996e+00 3.06262404e-01 2.10470915e-01 4.19248492e-01
1.20713592e+00 -1.16882658e+00 -2.90338427e-01 -3.76028657e-01
3.89566690e-01 -5.96715271e-01 7.66130269e-01 4.24355954e-01
-1.15659893e+00 -6.01858437e-01 -1.12771559e+00 -3.58316928e-01
-2.84510761e-01 -4.49959934e-02 8.94295990e-01 7.40633488e-01
-1.37816560e+00 7.75982916e-01 -8.33108842e-01 -8.66673365e-02
9.58093166e-01 3.24213475e-01 -4.41773146e-01 -1.37879089e-01
-9.77618694e-01 7.59741485e-01 3.48434776e-01 1.43375754e-01
-7.40459263e-01 -9.23761845e-01 -1.02838027e+00 2.36760750e-01
2.80245751e-01 -1.13518512e+00 1.07249928e+00 -8.44744205e-01
-1.38209212e+00 8.93995464e-01 -6.37604520e-02 -8.31241906e-01
1.02428496e+00 -2.50275761e-01 -1.08695149e-01 3.99619073e-01
5.98859228e-02 1.36642325e+00 9.19177353e-01 -8.77004921e-01
-1.00277984e+00 1.68232009e-01 8.35308954e-02 -1.06696837e-01
-5.53568713e-02 -1.36580780e-01 -5.23194075e-01 -7.77553737e-01
-1.03310347e-01 -9.05731916e-01 -6.14821255e-01 6.15730107e-01
-6.00104511e-01 -1.11841828e-01 7.76152790e-01 -2.12381065e-01
9.30364788e-01 -2.10202551e+00 -2.70977318e-01 8.69029164e-02
6.44408286e-01 5.08716285e-01 -1.83154479e-01 -2.88968295e-01
1.21889301e-01 3.11948597e-01 -4.37411636e-01 -4.78961349e-01
-2.20362153e-02 -1.48403004e-01 -1.36281207e-01 4.18448448e-01
6.56385839e-01 1.36611044e+00 -1.08187854e+00 -4.62316751e-01
4.62409258e-01 5.84774017e-01 -8.38060200e-01 -7.21628815e-02
-9.53572392e-02 3.72430682e-01 -2.39603519e-01 5.44973135e-01
6.52017534e-01 -4.04510647e-01 -5.11393964e-01 -3.48957330e-01
-1.13156319e-01 -1.78682040e-02 -1.25783217e+00 1.75591528e+00
-1.12362474e-01 1.17711461e+00 5.79418056e-02 -6.60253704e-01
7.62780964e-01 9.06718448e-02 4.73582357e-01 -4.29895431e-01
1.99563757e-01 1.39929280e-01 2.18103960e-01 -1.91282511e-01
6.14837170e-01 1.49657443e-01 1.65387422e-01 2.46181890e-01
1.53914690e-01 -1.71582654e-01 3.42589110e-01 2.72159368e-01
1.33876646e+00 2.06621170e-01 3.46547775e-02 -2.38262787e-01
3.01618665e-01 1.36645079e-01 5.42172492e-01 1.10589993e+00
-6.80810928e-01 9.70013320e-01 4.86249447e-01 -6.59346640e-01
-9.83894527e-01 -1.11133802e+00 -1.96173608e-01 8.43154430e-01
6.05255127e-01 -1.38205066e-01 -8.57718229e-01 -7.90290415e-01
2.95839697e-01 4.24315691e-01 -9.30679500e-01 -1.22439899e-01
-5.26274145e-01 -5.96872151e-01 8.90905082e-01 9.39983189e-01
8.50628376e-01 -1.29684675e+00 -9.13798332e-01 3.54155183e-01
2.20527470e-01 -1.62691593e+00 -6.86055660e-01 8.71877745e-03
-6.61922932e-01 -1.23630965e+00 -8.13782394e-01 -8.41789305e-01
8.04938078e-01 1.07457541e-01 1.18644547e+00 4.13626224e-01
-7.67401934e-01 2.21815541e-01 -1.36648476e-01 -1.92039922e-01
-1.92164421e-01 1.24025919e-01 -3.14537495e-01 -1.88429162e-01
3.39385360e-01 -2.32728004e-01 -1.10615373e+00 1.15535833e-01
-7.52007306e-01 -5.08875363e-02 5.62463701e-01 7.68192351e-01
3.31789941e-01 -2.97974735e-01 5.43532193e-01 -8.93189907e-01
6.82260334e-01 -1.00558214e-01 -6.17945850e-01 -4.95828427e-02
-6.84437037e-01 2.80986816e-01 5.68174362e-01 -1.43421158e-01
-9.21465099e-01 3.12189549e-01 -2.03427061e-01 -4.67542440e-01
-2.50174254e-01 -1.42276987e-01 4.52397138e-01 -6.04154527e-01
8.56585324e-01 -2.26245955e-01 -2.47882202e-01 -1.44159244e-02
5.73217511e-01 1.54914051e-01 1.08796799e+00 -4.46063280e-01
7.18148708e-01 7.25459635e-01 2.78955549e-01 -3.72129440e-01
-7.54443109e-01 -3.11514586e-01 -4.94647592e-01 -8.50195810e-02
9.91386414e-01 -7.18354702e-01 -8.81844342e-01 5.10409772e-01
-1.22246754e+00 -5.33203542e-01 -4.23633188e-01 3.06915730e-01
-3.97866249e-01 1.79377422e-01 -9.86300766e-01 -3.32035899e-01
-5.35521090e-01 -1.26778007e+00 1.00179446e+00 6.55326903e-01
-1.53233215e-01 -1.13562524e+00 -3.66539091e-01 -1.04731120e-01
5.88862419e-01 5.68408072e-01 2.67818391e-01 -3.63130093e-01
-8.49569023e-01 -1.36642292e-01 -7.73721516e-01 1.98293254e-01
1.44034103e-02 2.98274755e-01 -9.56222534e-01 -2.04989836e-01
-6.02942050e-01 -1.86204121e-01 1.34738398e+00 6.48905277e-01
9.78013337e-01 1.39098316e-01 -3.73127908e-01 1.19442177e+00
1.30048132e+00 -1.84545323e-01 6.43202782e-01 3.62870455e-01
7.98962116e-01 1.51166499e-01 1.13566257e-01 2.20564649e-01
1.47749245e-01 2.48293504e-01 4.97642189e-01 -6.26121759e-01
-6.79834127e-01 -1.76454365e-01 -8.96978006e-02 -1.43810004e-01
-7.95686468e-02 -3.45289707e-01 -7.60521710e-01 7.06182182e-01
-1.71965981e+00 -9.15729702e-01 -2.49593019e-01 1.78688097e+00
5.92584670e-01 6.95353448e-01 -4.44459319e-02 -1.45618364e-01
8.57396364e-01 -1.29708918e-02 -7.58011818e-01 -4.26862210e-01
-9.38659683e-02 5.93360722e-01 8.27184379e-01 5.96165121e-01
-1.34059429e+00 1.37641478e+00 7.04527664e+00 2.01955393e-01
-1.09796274e+00 -2.60414809e-01 7.93821573e-01 2.41360664e-01
-1.02177776e-01 -7.57369921e-02 -7.40623713e-01 3.14798594e-01
2.35874727e-01 -7.84817263e-02 3.20765585e-01 7.21267879e-01
5.06245764e-03 -9.99356434e-02 -1.24972641e+00 9.57007706e-01
-2.24944249e-01 -1.73878098e+00 -1.48124292e-01 -1.30752087e-01
1.18311822e+00 2.68335491e-01 2.02900484e-01 9.87292603e-02
7.93609142e-01 -1.28393853e+00 5.36898375e-01 4.38970089e-01
8.77043724e-01 -4.56785917e-01 5.73519409e-01 -6.67192414e-02
-1.23438609e+00 9.37562957e-02 -2.34541133e-01 -2.69778725e-02
4.33024406e-01 4.24201578e-01 -8.15526187e-01 1.18911602e-01
8.22169721e-01 8.91138136e-01 -5.44654608e-01 1.52453554e+00
-4.87465531e-01 3.62584203e-01 -4.34183449e-01 2.05175102e-01
6.50763810e-01 1.26181647e-01 5.20262778e-01 1.57458866e+00
1.55589864e-01 -1.67471077e-02 1.20359242e-01 1.31786394e+00
-6.06147587e-01 -4.80836272e-01 -4.55562323e-01 3.50696892e-01
5.64147711e-01 1.35030329e+00 -1.08976471e+00 -4.68021482e-01
-3.44723135e-01 1.23783910e+00 2.93239802e-01 5.83722651e-01
-7.86245346e-01 -7.36714184e-01 9.17911053e-01 -8.72148499e-02
5.11702240e-01 -2.33953848e-01 -6.04991972e-01 -9.77894962e-01
-1.51577353e-01 -2.82298207e-01 2.46229365e-01 -7.34628022e-01
-1.18984365e+00 5.32570899e-01 -4.02114481e-01 -9.75348473e-01
-2.01280877e-01 -8.61053169e-01 -9.19196308e-01 6.45873070e-01
-1.73513925e+00 -8.12055051e-01 -6.50305092e-01 6.02524161e-01
3.68603081e-01 1.63241357e-01 4.03929710e-01 3.48235458e-01
-4.54928160e-01 8.07984293e-01 -2.82232553e-01 7.20421553e-01
6.81770802e-01 -1.37725079e+00 1.26108563e+00 1.33664227e+00
3.10430527e-01 4.53175455e-01 5.33583462e-01 -4.76071030e-01
-9.45849478e-01 -1.19332731e+00 3.51271182e-01 -5.08162022e-01
5.02143621e-01 -1.18661322e-01 -7.69568741e-01 8.07278335e-01
1.24371804e-01 8.50958884e-01 -9.87321064e-02 -4.98010248e-01
-3.23021680e-01 2.42238119e-01 -1.28207028e+00 7.31371880e-01
1.49376714e+00 -2.67928094e-01 -3.98759693e-01 -9.81780067e-02
7.71668613e-01 -6.93810940e-01 -5.58030427e-01 4.80870456e-01
5.71885645e-01 -1.07835948e+00 1.17252862e+00 -8.15210164e-01
4.23390001e-01 -3.91273111e-01 4.73994166e-01 -1.28423584e+00
-5.20457804e-01 -9.07290459e-01 -1.16066895e-01 7.48944759e-01
4.68367904e-01 -6.32404923e-01 1.12802029e+00 7.80251980e-01
-4.26260859e-01 -5.45478046e-01 -8.23438466e-01 -6.61230266e-01
-3.58549096e-02 -3.27932626e-01 3.52757454e-01 8.19010913e-01
-3.36384803e-01 -1.60334874e-02 -6.59190789e-02 1.14750070e-02
7.91876972e-01 3.12268194e-02 6.46337450e-01 -1.05302429e+00
-8.76838863e-02 -1.00265944e+00 -9.12793696e-01 -1.32126582e+00
1.02034904e-01 -8.01412642e-01 2.15268031e-01 -1.71218646e+00
-3.05245906e-01 -4.61870462e-01 -4.21724766e-01 4.96518046e-01
-5.41325867e-01 8.63041878e-01 2.89076954e-01 -1.94615602e-01
-6.53794706e-01 1.05972327e-01 1.57147419e+00 -1.62377164e-01
-2.25246787e-01 -8.40538964e-02 -7.93636501e-01 7.60941327e-01
7.93263495e-01 -4.46800470e-01 -1.17917813e-01 -6.02234781e-01
6.01474680e-02 -3.81351590e-01 7.42287874e-01 -1.24963915e+00
5.88603079e-01 2.25507468e-02 5.80646396e-01 -1.15613855e-01
-1.58513542e-02 -6.37697279e-01 -2.47434795e-01 6.79834366e-01
-2.94074923e-01 3.23602036e-02 3.13169211e-01 4.40245956e-01
-5.82932793e-02 -4.76167239e-02 8.85044515e-01 -1.21176206e-01
-1.28902924e+00 5.77153683e-01 -1.33996800e-01 5.29673040e-01
1.03361046e+00 -5.01437545e-01 -5.02131879e-01 -2.13159546e-01
-7.71453857e-01 2.58148521e-01 5.97794473e-01 4.07431036e-01
7.25276768e-01 -1.09272873e+00 -8.74431193e-01 3.40540022e-01
-1.07708476e-01 2.05699310e-01 -2.70306736e-01 7.18898773e-01
-1.04642165e+00 9.31676403e-02 -4.91703898e-01 -6.74894631e-01
-6.35680497e-01 3.63975137e-01 6.37546122e-01 1.68258384e-01
-1.04611623e+00 1.21735215e+00 3.12069505e-01 9.91794989e-02
2.01873824e-01 -6.27586305e-01 1.63151070e-01 -4.26704884e-01
3.80526394e-01 4.66557503e-01 6.00511730e-02 -2.36612037e-01
-4.49264735e-01 5.86875141e-01 -5.79380281e-02 -2.00569746e-03
1.04960072e+00 5.20368330e-02 4.17553753e-01 -4.21581507e-01
1.06704724e+00 -1.36403561e-01 -2.01882815e+00 -2.22793370e-02
-7.90353045e-02 -4.27458048e-01 2.77739137e-01 -6.14110231e-01
-1.47517645e+00 9.28866923e-01 5.37846625e-01 7.59603083e-02
9.17159259e-01 -1.32001087e-01 1.12553799e+00 3.46649736e-01
9.55493525e-02 -9.68488991e-01 2.51319427e-02 5.17529309e-01
3.59388113e-01 -1.55051994e+00 -1.62328854e-01 -4.63870257e-01
-3.91113013e-01 1.24641478e+00 8.25589657e-01 -5.64823568e-01
4.51317340e-01 5.67691088e-01 4.46179919e-02 -1.12013124e-01
-4.35561806e-01 -5.65670013e-01 4.67186719e-01 6.49893463e-01
3.25102121e-01 -1.11763172e-01 -8.14419147e-03 1.19040944e-01
-5.20103872e-02 2.09296227e-01 5.71213186e-01 7.13846028e-01
-5.22221327e-01 -6.90266788e-01 -1.19969361e-01 3.91334027e-01
-5.26440144e-01 -3.79363060e-01 -2.66426414e-01 9.28643525e-01
-1.97107177e-02 7.90031970e-01 3.14198285e-01 -2.13601395e-01
3.28784585e-01 -2.93980271e-01 4.31449383e-01 -3.55317354e-01
-5.91679573e-01 -5.26993454e-01 -1.94809452e-01 -7.68072307e-01
-1.28799438e-01 -2.70872384e-01 -1.50646973e+00 -3.58632684e-01
-2.68832594e-01 -2.31789991e-01 3.39265347e-01 7.89547384e-01
4.95306730e-01 5.92396200e-01 2.70845413e-01 -1.06455624e+00
-1.79275274e-01 -5.42913616e-01 -7.61052668e-02 7.22797871e-01
5.60014546e-01 -5.49784482e-01 -4.04409379e-01 2.60936528e-01] | [9.450489044189453, 0.12095154821872711] |
3461e50d-adff-4dd9-951a-e87f7ffa1325 | promil-probabilistic-multiple-instance | 2306.10535 | null | https://arxiv.org/abs/2306.10535v1 | https://arxiv.org/pdf/2306.10535v1.pdf | ProMIL: Probabilistic Multiple Instance Learning for Medical Imaging | Multiple Instance Learning (MIL) is a weakly-supervised problem in which one label is assigned to the whole bag of instances. An important class of MIL models is instance-based, where we first classify instances and then aggregate those predictions to obtain a bag label. The most common MIL model is when we consider a bag as positive if at least one of its instances has a positive label. However, this reasoning does not hold in many real-life scenarios, where the positive bag label is often a consequence of a certain percentage of positive instances. To address this issue, we introduce a dedicated instance-based method called ProMIL, based on deep neural networks and Bernstein polynomial estimation. An important advantage of ProMIL is that it can automatically detect the optimal percentage level for decision-making. We show that ProMIL outperforms standard instance-based MIL in real-world medical applications. We make the code available. | ['Bartosz Zieliński', 'Jacek Tabor', 'Robert Sabiniewicz', 'Arkadiusz Lewicki', 'Dawid Rymarczyk', 'Łukasz Struski'] | 2023-06-18 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 4.19398546e-01 4.74955708e-01 -7.99939990e-01 -4.79194045e-01
-9.64046597e-01 -4.06241566e-01 1.68663695e-01 8.42277050e-01
-6.13974072e-02 1.22414744e+00 -1.68594301e-01 -2.68780798e-01
-4.39594775e-01 -1.24487543e+00 -9.72536683e-01 -9.24706995e-01
-1.69749022e-01 9.61958826e-01 -1.72717407e-01 2.55714595e-01
7.99551532e-02 4.46193248e-01 -1.30637372e+00 9.00862157e-01
8.36164951e-01 1.17293561e+00 -4.45300519e-01 4.30228800e-01
-2.22313479e-01 9.95326281e-01 -6.67414486e-01 -5.32155871e-01
2.58386761e-01 -1.22144870e-01 -9.49566245e-01 3.25690247e-02
2.43885338e-01 3.98996882e-02 4.36109990e-01 8.23206127e-01
1.44692257e-01 -3.64754573e-02 9.70051944e-01 -1.57400346e+00
-5.23108393e-02 6.89062297e-01 -5.85462868e-01 1.11078098e-03
1.22840255e-01 -2.08664924e-01 1.33621049e+00 -7.20718205e-01
2.25059167e-01 9.37614679e-01 7.37963974e-01 1.52876198e-01
-1.35898578e+00 -5.83936930e-01 2.50981897e-01 1.61903292e-01
-1.04258871e+00 9.86810401e-03 7.70354390e-01 -4.93085831e-01
7.38625467e-01 5.09063244e-01 8.33772123e-01 4.35948133e-01
3.73661548e-01 1.08645904e+00 1.46847463e+00 -5.48326075e-01
3.61207068e-01 3.45608771e-01 5.74312627e-01 5.63588560e-01
3.79221529e-01 -7.02101812e-02 -2.10868210e-01 -5.64407527e-01
3.74200583e-01 2.66988784e-01 -1.73155032e-02 -3.42044234e-01
-1.03438079e+00 8.80515993e-01 4.83413011e-01 2.06207365e-01
-5.37475944e-01 1.82314485e-01 3.58729511e-01 4.41870838e-02
8.16984594e-01 5.19557476e-01 -5.45155704e-01 3.00789118e-01
-1.04424810e+00 5.09767056e-01 1.03217363e+00 4.76554483e-01
7.96744466e-01 -6.58882558e-01 -4.10306752e-01 7.66797066e-01
6.36871383e-02 -1.98894516e-02 1.65417492e-01 -7.06336498e-01
2.67652959e-01 9.67875242e-01 5.77899180e-02 -9.51346457e-01
-5.99373221e-01 -7.68581033e-01 -8.73103857e-01 7.21065179e-02
5.98525584e-01 1.94592804e-01 -6.83581531e-01 1.37795556e+00
3.11072916e-01 4.42602515e-01 -1.66917685e-02 5.24906456e-01
7.34672487e-01 5.36559641e-01 9.76129472e-02 -5.89208663e-01
1.12250221e+00 -7.87537098e-01 -5.96889734e-01 -2.45825291e-01
6.55184507e-01 -2.26990923e-01 4.73410398e-01 7.65511572e-01
-8.91189098e-01 -1.61370829e-01 -8.44612777e-01 4.62321550e-01
-4.75224674e-01 -2.76023000e-02 8.51130247e-01 5.43386936e-01
-6.46170199e-01 6.00806415e-01 -5.27658224e-01 1.60690516e-01
5.47251582e-01 5.39178610e-01 -4.54678953e-01 -1.31797627e-01
-1.19380462e+00 7.77299821e-01 5.98195255e-01 2.10380927e-02
-4.94259596e-01 -8.90742898e-01 -6.53530240e-01 1.43132374e-01
4.48931009e-01 -5.21944344e-01 1.14627242e+00 -1.18228006e+00
-6.51017725e-01 1.13204598e+00 -1.48719802e-01 -7.25826025e-01
6.41606867e-01 1.08294956e-01 -1.76085711e-01 -1.22364692e-01
-1.95999127e-02 4.52845186e-01 7.49854028e-01 -1.36954820e+00
-9.25033748e-01 -4.73330140e-01 3.42138916e-01 -7.18347207e-02
-1.02633424e-01 -4.02891129e-01 -3.62885781e-02 -3.96157146e-01
1.18568800e-01 -8.14190149e-01 -4.67766136e-01 -4.10236269e-01
-6.65224373e-01 -5.43713570e-01 3.89376789e-01 -3.65579784e-01
1.31108332e+00 -1.69454741e+00 -1.13469407e-01 5.04407287e-01
4.07261610e-01 5.80577850e-02 2.98769355e-01 2.40167350e-01
-3.18355799e-01 2.25043908e-01 -3.09180766e-01 -6.32027015e-02
-6.58443868e-02 3.45896751e-01 -4.67643410e-01 3.59799385e-01
3.48831147e-01 7.31329083e-01 -1.00309026e+00 -6.50032103e-01
2.18879342e-01 7.31153041e-02 -7.40526795e-01 3.61448824e-02
-4.72049743e-01 3.66626769e-01 -1.98990524e-01 6.81493998e-01
5.97719729e-01 -5.38563669e-01 3.68707985e-01 -3.12185079e-01
3.09425980e-01 1.36978880e-01 -1.23464000e+00 7.30431020e-01
-3.60027850e-01 2.18104392e-01 -5.52360415e-01 -1.68022072e+00
7.92162359e-01 3.40253055e-01 1.05163383e+00 -2.32570440e-01
-7.03763887e-02 3.34160179e-01 -2.54069030e-01 -3.10070962e-01
1.57940462e-01 -5.49978733e-01 -1.31053805e-01 3.48448128e-01
-2.78078109e-01 -2.04524621e-02 2.96770841e-01 -5.35034761e-02
1.07071102e+00 -1.74709380e-01 8.76754403e-01 -1.81365192e-01
4.72791612e-01 1.64387926e-01 9.62863803e-01 8.67101550e-01
-9.09778923e-02 4.72184539e-01 1.03881598e+00 -6.91516161e-01
-7.75099039e-01 -9.93527949e-01 -6.02952302e-01 8.13530803e-01
-5.17438538e-02 -1.94529563e-01 -5.97939968e-01 -9.91990626e-01
2.74996758e-01 6.60070717e-01 -7.59184599e-01 -7.20224231e-02
-2.67915189e-01 -1.24817419e+00 2.33134013e-02 4.86467391e-01
8.22516605e-02 -1.07957363e+00 -3.56630147e-01 3.70322347e-01
-3.36924762e-01 -9.28772986e-01 1.24488190e-01 5.59818566e-01
-9.15553570e-01 -1.34280682e+00 -4.93563443e-01 -5.59709549e-01
9.31295812e-01 -3.77813399e-01 1.49737513e+00 9.99958664e-02
-1.85636625e-01 2.50358403e-01 -2.62155980e-01 -7.39560127e-01
-4.92405981e-01 -1.38823818e-02 5.54446317e-02 2.66009837e-01
5.24935484e-01 -2.63064295e-01 -4.17839438e-01 6.19587749e-02
-7.11698949e-01 3.41577083e-01 5.23846209e-01 1.15109336e+00
1.23105669e+00 4.09744084e-01 9.74163949e-01 -1.43300498e+00
4.87006724e-01 -6.81404889e-01 -4.23950016e-01 6.12718761e-01
-5.31141043e-01 -7.49822631e-02 6.10336840e-01 -3.42527539e-01
-3.98294091e-01 2.44008034e-01 -8.36731344e-02 -1.73290089e-01
-2.56047308e-01 8.80921900e-01 1.09465234e-01 2.39670768e-01
4.65149850e-01 -5.83736487e-02 -1.57657221e-01 -1.38293356e-01
-1.40926972e-01 6.71865106e-01 1.42759472e-01 -6.28885686e-01
2.13899717e-01 3.44238609e-01 3.12683851e-01 -5.71209848e-01
-1.55798841e+00 -6.20822668e-01 -5.41922927e-01 -3.63796920e-01
6.93934500e-01 -6.82343423e-01 -9.22658384e-01 4.28131111e-02
-9.51279461e-01 -3.23647231e-01 -3.72193694e-01 3.68480861e-01
-8.50609839e-01 -5.91548756e-02 -4.50630993e-01 -9.45880234e-01
-4.06231165e-01 -1.24229836e+00 8.86169612e-01 3.91292349e-02
-2.62444884e-01 -1.03822637e+00 4.24668659e-03 5.66421151e-01
-1.32272720e-01 6.20645225e-01 1.15551853e+00 -1.09110212e+00
-3.91357809e-01 -6.40938222e-01 -1.24863386e-01 3.90884489e-01
1.48445487e-01 -1.82365045e-01 -9.81680036e-01 -3.05095494e-01
-2.77136743e-01 -3.32560897e-01 8.12337995e-01 7.91361034e-01
1.64048052e+00 -3.53077650e-01 -5.24039745e-01 1.89014181e-01
1.52531219e+00 9.65866446e-02 3.35530132e-01 2.28576601e-01
4.32869643e-01 5.15761435e-01 9.25751090e-01 4.40272033e-01
2.79905498e-01 6.65722787e-01 6.91542327e-01 -4.13543522e-01
4.29742783e-01 3.62869464e-02 -6.89834580e-02 3.88343811e-01
-1.71461686e-01 -1.11586183e-01 -9.48734999e-01 4.04268295e-01
-2.09456182e+00 -9.45560098e-01 -1.09414592e-01 2.42665863e+00
9.29628134e-01 4.31909323e-01 3.26540947e-01 5.97620428e-01
6.33395910e-01 -2.79918283e-01 -4.49322701e-01 -3.71712476e-01
1.81315407e-01 2.17540085e-01 4.76404220e-01 3.95721108e-01
-1.40318918e+00 3.84833127e-01 6.55524302e+00 8.43054116e-01
-8.96282256e-01 1.13309629e-01 1.39778185e+00 1.91086866e-02
-1.47267222e-01 -1.96986750e-01 -1.00887549e+00 4.37483907e-01
7.73873568e-01 -1.60576269e-01 3.44797708e-02 9.42878366e-01
-2.68982165e-02 -1.68292463e-01 -1.43797588e+00 6.84722900e-01
-1.87962875e-02 -1.55759573e+00 -9.02341381e-02 2.22288117e-01
6.92864716e-01 -3.98229569e-01 -1.12067396e-02 5.07520258e-01
2.58221358e-01 -1.25097382e+00 4.27219182e-01 5.45703650e-01
6.42402649e-01 -1.03338933e+00 8.99798214e-01 6.31984472e-01
-8.56863022e-01 -2.76044607e-01 -3.10839295e-01 -6.02050349e-02
-8.01751763e-02 1.24498808e+00 -1.16412508e+00 5.38914919e-01
4.77818340e-01 7.46997297e-01 -2.62522310e-01 1.32871854e+00
-6.29413128e-02 7.49923706e-01 -3.09379250e-01 5.57171740e-02
6.76754341e-02 -9.91408080e-02 2.93338686e-01 1.18973100e+00
9.75303948e-02 -2.23631486e-02 5.80819547e-01 7.91161954e-01
9.91901755e-03 2.49066368e-01 -5.40624082e-01 2.37718925e-01
1.66842774e-01 1.34152150e+00 -8.10142517e-01 -3.71527791e-01
-3.02526504e-01 4.21290964e-01 4.17837411e-01 1.43573299e-01
-8.34762573e-01 1.05323277e-01 2.12593749e-01 1.91722348e-01
-4.75814007e-02 5.37728012e-01 -5.98702312e-01 -8.91752899e-01
-2.64706872e-02 -7.65142918e-01 8.16130579e-01 -4.17908400e-01
-1.61290550e+00 2.40942940e-01 -1.36514427e-02 -1.35715210e+00
-3.20266187e-01 -8.12796295e-01 -4.08601284e-01 6.34088814e-01
-1.48563492e+00 -1.21436644e+00 -1.73652872e-01 3.58261526e-01
1.87743887e-01 7.40579441e-02 9.81271088e-01 2.21853554e-01
-5.13566434e-01 5.43580472e-01 -1.12886176e-01 1.41462594e-01
5.22135556e-01 -1.59437525e+00 -3.42755318e-01 3.56779128e-01
1.35713995e-01 2.35400230e-01 8.49781752e-01 -4.53741014e-01
-8.94140065e-01 -1.37427282e+00 1.05768275e+00 -4.64572608e-01
3.02531511e-01 -3.89217958e-02 -8.93518150e-01 8.66947711e-01
-5.16451001e-01 4.66729432e-01 1.06608498e+00 4.11941856e-01
-6.85162097e-02 -5.56343555e-01 -1.36791182e+00 2.27902547e-01
4.93829668e-01 -7.25443754e-03 -4.59160447e-01 8.74084413e-01
3.96553010e-01 -5.07730365e-01 -1.16656768e+00 8.45887423e-01
5.14307022e-01 -1.03178787e+00 1.13475323e+00 -7.30696976e-01
7.43294537e-01 -1.66721717e-01 -1.43318594e-01 -1.39173472e+00
-2.33203933e-01 1.85019314e-01 -4.19476926e-01 8.36555243e-01
5.70844531e-01 -8.63199592e-01 9.23080921e-01 6.27158761e-01
1.02268890e-01 -1.37060416e+00 -9.10908639e-01 -6.63437784e-01
2.40452215e-01 -6.19353831e-01 7.34403849e-01 1.06502771e+00
3.20594698e-01 -1.09703273e-01 -2.80166179e-01 1.78705171e-01
8.25571895e-01 5.43200135e-01 5.30692875e-01 -1.51397824e+00
-5.88255167e-01 -3.06540608e-01 -5.16003847e-01 -2.51014739e-01
3.08420449e-01 -1.22360075e+00 -1.46137215e-02 -1.79261565e+00
5.57248354e-01 -1.03670824e+00 -8.18811595e-01 8.45440686e-01
-2.40383357e-01 4.16450620e-01 -1.15253113e-01 1.56469010e-02
-5.61376393e-01 -2.07264006e-01 1.01043522e+00 -2.99018890e-01
-1.54342279e-01 5.99268794e-01 -6.46733224e-01 7.62890637e-01
8.01351547e-01 -5.89873016e-01 -6.97399527e-02 4.23605144e-01
5.79092503e-01 3.95523250e-01 3.70712101e-01 -8.95218730e-01
4.82547097e-02 -5.69023669e-01 4.68853593e-01 -7.96333134e-01
2.58479625e-01 -9.66511011e-01 3.58519405e-01 6.92622483e-01
-6.14064455e-01 -4.72499549e-01 -3.24650742e-02 5.10211170e-01
-2.81113714e-01 -3.77542168e-01 8.77584457e-01 -1.07291780e-01
-4.41223055e-01 4.47141290e-01 7.48654231e-02 -1.57011807e-01
1.38395226e+00 -9.61582884e-02 -2.51477629e-01 -1.39365703e-01
-9.35581446e-01 2.63108760e-01 3.22802388e-03 -8.08872208e-02
4.63926882e-01 -1.30863285e+00 -8.79540861e-01 1.20360918e-01
3.96007776e-01 2.62586474e-01 1.15594856e-01 1.08049846e+00
-3.50234032e-01 4.74851251e-01 1.42089516e-01 -8.66621554e-01
-1.28854382e+00 6.38791025e-01 6.13498449e-01 -9.11450565e-01
-5.07682621e-01 6.47958338e-01 3.73481840e-01 -5.38129628e-01
3.17622334e-01 -1.95660412e-01 -3.53890300e-01 1.38186008e-01
6.13024473e-01 2.54129112e-01 3.11414391e-01 -3.63153219e-01
-4.16389495e-01 1.99622706e-01 -1.33901192e-02 3.10797483e-01
1.36812913e+00 6.18030906e-01 -4.16622788e-01 7.64703453e-01
8.64542961e-01 -2.57361144e-01 -8.59120131e-01 -1.66735634e-01
-5.96466176e-02 -2.82721341e-01 -3.59553732e-02 -9.69501734e-01
-1.11622214e+00 7.15198040e-01 3.23842347e-01 4.79268283e-01
1.24445295e+00 -1.62376706e-02 3.44064176e-01 2.88153678e-01
5.17648458e-01 -1.19027984e+00 -1.21646903e-01 2.39590049e-01
8.83153081e-01 -1.52389431e+00 1.19915619e-01 -6.19414926e-01
-4.03468877e-01 1.20433211e+00 6.28630400e-01 -1.77118674e-01
8.10445786e-01 3.34121138e-01 -2.91162461e-01 -1.35896847e-01
-7.90826917e-01 1.17520131e-01 4.63997215e-01 2.89262533e-01
4.74721462e-01 5.86541295e-01 -3.88933450e-01 8.65194380e-01
-9.22190100e-02 2.57628441e-01 2.93969870e-01 6.38977408e-01
-2.84620255e-01 -8.97338390e-01 -4.87831861e-01 1.22954130e+00
-7.48867273e-01 -5.03934957e-02 -5.54395877e-02 3.85491788e-01
3.14166337e-01 8.58577430e-01 2.41381809e-01 -1.72138020e-01
1.90696761e-01 1.63352966e-01 5.74774086e-01 -7.38432229e-01
-6.21831596e-01 -1.99547321e-01 5.06525934e-02 -2.91986883e-01
-4.39534783e-01 -4.64901924e-01 -1.35129273e+00 -1.96323484e-01
-4.35523570e-01 1.32864818e-01 3.94460827e-01 1.20483065e+00
-1.30068868e-01 6.61429882e-01 6.04478836e-01 -5.75943470e-01
-3.27295512e-01 -6.49901092e-01 -7.96419561e-01 4.71286982e-01
4.44237977e-01 -7.36964047e-01 -5.04047036e-01 -1.16816215e-01] | [9.230507850646973, 3.9543490409851074] |
026f8a25-f27f-4ad7-ab0c-6941e31735fc | multi-gat-a-graphical-attention-based | null | null | https://ieeexplore.ieee.org/abstract/document/9354900/ | https://ieeexplore.ieee.org/abstract/document/9354900/ | Multi-GAT: A Graphical Attention-based Hierarchical Multimodal Representation Learning Approach for Human Activity Recognition | Recognizing human activities is one of the crucial capabilities that a robot needs to have to be useful around people. Although modern robots are equipped with various types of sensors, human activity recognition (HAR) still remains a challenging problem, particularly in the presence of noisy sensor data. In this work, we introduce a multimodal graphical attention-based HAR approach, called Multi-GAT, which hierarchically learns complementary multimodal features. We develop a multimodal mixture-of-experts model to disentangle and extract salient modality-specific features that enable feature interactions. Additionally, we introduce a novel message-passing based graphical attention approach to capture cross-modal relation for extracting complementary multimodal features. The experimental results on two multimodal human activity datasets suggest that Multi-GAT outperformed state-of-the-art HAR algorithms across all datasets and metrics tested. Finally, the experimental results with noisy sensor data indicate that Multi-GAT consistently outperforms all the evaluated baselines. The robust performance suggests that Multi-GAT can enable seamless human-robot collaboration in noisy human environments. | ['Tariq Iqbal', 'Md Mofijul Islam'] | 2021-04-01 | null | null | null | ieee-robotics-and-automation-letters-2021-4 | ['multimodal-activity-recognition'] | ['computer-vision'] | [ 5.58098368e-02 -7.73677751e-02 1.03275761e-01 -3.84984195e-01
-1.01360655e+00 -2.05585778e-01 7.76081920e-01 1.35417596e-01
-4.96679246e-01 5.72418928e-01 6.70583010e-01 4.17242557e-01
-2.77687401e-01 -1.85077652e-01 -5.48986077e-01 -7.15770721e-01
-2.71543294e-01 5.02050400e-01 -9.51458514e-02 -3.49706948e-01
3.76741365e-02 2.46114835e-01 -1.62882793e+00 3.08624148e-01
6.65036201e-01 8.32606196e-01 2.91350335e-01 6.99721098e-01
1.53615639e-01 1.11643660e+00 -4.77974087e-01 -1.71982110e-01
-1.91742450e-01 -2.46331692e-01 -7.63492167e-01 3.06075305e-01
-8.47503729e-03 -1.06720209e-01 -3.50884706e-01 7.72134125e-01
5.81328988e-01 4.11318749e-01 6.79103553e-01 -1.77549469e+00
-4.81092304e-01 4.08184409e-01 -4.59872037e-01 -6.96955994e-02
1.18737733e+00 2.24334463e-01 8.73073459e-01 -1.10061586e+00
2.83244640e-01 1.56300426e+00 4.39414084e-01 2.77163059e-01
-1.10942078e+00 -3.16097796e-01 2.16797695e-01 4.42655146e-01
-1.28564334e+00 -4.24660891e-01 8.28798592e-01 -5.62946618e-01
9.54213500e-01 2.03383550e-01 5.56541085e-01 1.54927039e+00
1.22825904e-02 1.15754652e+00 9.18130636e-01 -3.33870232e-01
6.78568408e-02 -1.20768495e-01 2.36644432e-01 7.91923225e-01
1.47636041e-01 -5.44619799e-01 -1.19736910e+00 -2.39294201e-01
7.34030962e-01 3.02039206e-01 -2.65339434e-01 -6.17466211e-01
-2.05351496e+00 5.87155461e-01 2.87658334e-01 2.81280458e-01
-7.13672757e-01 1.92507446e-01 3.82652879e-01 -9.71163511e-02
-7.81912506e-02 3.66742879e-01 -1.49945855e-01 -5.55928767e-01
-1.22204117e-01 1.72959501e-03 7.03176498e-01 9.68221784e-01
5.62215269e-01 -3.19801599e-01 -3.15661728e-01 9.43720758e-01
6.72688544e-01 6.98781013e-01 3.84441227e-01 -1.06420469e+00
6.67632163e-01 8.82810950e-01 4.18891758e-01 -1.08805597e+00
-6.88436806e-01 1.17236048e-01 -1.02254665e+00 -1.59605235e-01
2.64830738e-01 -1.56439066e-01 -6.09407127e-01 1.62503731e+00
2.78234780e-01 -1.96269765e-01 2.18584120e-01 1.07480001e+00
9.62268591e-01 5.45046687e-01 3.63893181e-01 3.02120037e-02
1.56540298e+00 -1.23635924e+00 -9.95028436e-01 -5.76280177e-01
3.34667265e-01 -3.33159596e-01 9.92534578e-01 5.87632179e-01
-8.10251236e-01 -6.95033014e-01 -1.02339959e+00 -2.45286617e-02
-4.40093935e-01 2.16395348e-01 7.60309100e-01 2.83915043e-01
-5.72440207e-01 -9.16997269e-02 -9.07095850e-01 -7.15184748e-01
1.24072313e-01 3.49660099e-01 -1.14925551e+00 -1.87289774e-01
-9.07707453e-01 9.77860212e-01 3.28697205e-01 3.13747853e-01
-1.00575459e+00 1.43958747e-01 -1.37254572e+00 -1.78156748e-01
2.56166130e-01 -7.79363930e-01 1.30539083e+00 -6.05153978e-01
-1.43843722e+00 4.14488584e-01 -4.22892869e-01 -1.20818801e-01
3.09352726e-01 -7.62039185e-01 -2.32921302e-01 2.42164642e-01
3.38765830e-01 7.42453516e-01 5.72189450e-01 -1.43318665e+00
-5.90101659e-01 -4.35787946e-01 -1.44239530e-01 5.78227401e-01
-1.94604769e-01 3.90495248e-02 -5.14215410e-01 -3.82473975e-01
6.35321289e-02 -9.11647260e-01 -2.97346741e-01 -4.57532823e-01
-4.03189838e-01 -5.15543222e-01 7.44121671e-01 -4.95777160e-01
8.47863436e-01 -2.04213166e+00 7.61538088e-01 1.32628217e-01
1.62319407e-01 -2.69646287e-01 -8.73242542e-02 8.20234835e-01
2.42421001e-01 -3.97814482e-01 -3.56843442e-01 -8.29925656e-01
4.25865114e-01 2.65116006e-01 2.32520252e-01 5.48816800e-01
2.97191113e-01 9.70542789e-01 -1.07515454e+00 -6.62923694e-01
5.75959861e-01 7.23653495e-01 1.28533747e-02 5.48131585e-01
9.00564194e-02 9.08630013e-01 -5.07825792e-01 8.89722407e-01
2.25852594e-01 -4.01079237e-01 2.81663269e-01 -1.12296268e-01
1.78422749e-01 -4.13937628e-01 -1.17334342e+00 2.05496836e+00
-1.95341334e-01 6.44114137e-01 3.73016931e-02 -7.90470004e-01
7.26531148e-01 4.03479546e-01 4.75119889e-01 -5.88090837e-01
3.14735264e-01 -1.57443702e-01 -3.30080092e-01 -1.05599797e+00
6.69932783e-01 3.16750675e-01 -9.05066550e-01 1.99494392e-01
2.41737545e-01 3.86181027e-01 -8.10275823e-02 2.42267579e-01
1.23992109e+00 1.77085876e-01 5.84674895e-01 2.68464833e-01
5.80729783e-01 -2.31127158e-01 3.84977639e-01 8.18550706e-01
-5.07150114e-01 4.74589556e-01 1.91273987e-01 -1.56432688e-01
-4.40962344e-01 -1.05518138e+00 5.23973525e-01 1.59642529e+00
3.71209770e-01 -4.24708068e-01 -4.82244849e-01 -6.23599112e-01
-2.78801471e-01 3.82696748e-01 -6.62888885e-01 1.11152686e-01
-2.83966959e-01 -5.39672554e-01 6.67663455e-01 7.81127512e-01
7.15084493e-01 -1.18101692e+00 -8.69479060e-01 1.19464666e-01
-8.62385213e-01 -1.29275489e+00 -1.76300898e-01 1.91296428e-01
-2.27659434e-01 -1.11708570e+00 -7.81789839e-01 -8.01320732e-01
5.81506729e-01 5.44391334e-01 8.39403987e-01 -2.48533964e-01
-1.15818009e-01 1.15547740e+00 -6.87845469e-01 -4.33612168e-01
2.14702472e-01 3.62368510e-03 8.18344429e-02 4.16708827e-01
5.75100303e-01 -4.87947136e-01 -4.61736143e-01 4.09910649e-01
-5.86012840e-01 2.05943119e-02 9.62704420e-01 8.77030432e-01
2.02774972e-01 -2.30125830e-01 5.21362185e-01 -5.62245660e-02
7.52548575e-01 -8.44994903e-01 1.15967698e-01 3.94228429e-01
2.99783349e-01 1.98805198e-01 5.73030449e-02 -6.32487357e-01
-1.19928944e+00 5.91985345e-01 3.87284100e-01 -1.47349462e-01
-6.98874235e-01 5.81311047e-01 -7.19694018e-01 1.29424781e-01
3.02998245e-01 1.14861302e-01 -2.54830956e-01 -4.25521314e-01
5.51372886e-01 9.40992236e-01 8.24424744e-01 -6.41981065e-01
3.14087689e-01 5.02763689e-01 -5.73639907e-02 -1.10354722e+00
-4.42038983e-01 -8.62187624e-01 -7.64765382e-01 -5.43152690e-01
1.24057388e+00 -1.13371599e+00 -1.05991685e+00 7.89156079e-01
-1.35769594e+00 -2.64589012e-01 2.64580578e-01 5.84843040e-01
-6.47380233e-01 6.29629433e-01 -5.37092626e-01 -1.26530600e+00
-1.91196222e-02 -1.25530624e+00 1.50830579e+00 2.41530821e-01
-5.08812249e-01 -7.25956261e-01 1.33325949e-01 8.51506770e-01
-8.23024660e-02 4.30302322e-01 2.91792274e-01 -7.02938080e-01
-4.64870125e-01 -1.66673556e-01 -1.24269247e-01 -1.99606240e-01
1.43057674e-01 -5.65199733e-01 -8.96967292e-01 -2.64555253e-02
-5.01420319e-01 -6.39610171e-01 6.27688527e-01 2.92991307e-02
5.59338152e-01 1.83077506e-03 -4.84916240e-01 -1.55598342e-01
6.47326827e-01 4.26538056e-03 6.09764278e-01 4.78150010e-01
9.14369285e-01 7.88144708e-01 9.10070539e-01 7.44405985e-01
1.08250940e+00 6.17649376e-01 4.76955771e-01 3.93592939e-02
4.80605721e-01 -1.58527777e-01 7.00081289e-01 7.10459590e-01
-1.43451244e-01 -2.42284954e-01 -9.87104774e-01 7.09090889e-01
-2.63148737e+00 -9.35853064e-01 -2.43582144e-01 1.75810301e+00
3.38320911e-01 -4.05718535e-01 3.82124484e-01 2.09925592e-01
5.79018354e-01 8.46358687e-02 -2.20884576e-01 -7.29710562e-03
-1.57681361e-01 -4.18684006e-01 5.98378778e-02 2.18166277e-01
-1.50607622e+00 6.08830988e-01 6.01751423e+00 2.87601501e-01
-3.03379238e-01 1.57612175e-01 -9.00861397e-02 6.78161606e-02
2.35467836e-01 -5.10083020e-01 -4.19273227e-01 3.15443873e-01
6.91290915e-01 4.80794311e-01 2.73316175e-01 8.56550813e-01
8.35014582e-02 -5.49425483e-01 -1.29620576e+00 1.52234435e+00
5.13700902e-01 -4.26114440e-01 -6.70127124e-02 8.48380441e-05
4.17580962e-01 -5.45429513e-02 -2.51269192e-01 3.07319850e-01
5.17758131e-01 -1.22096479e+00 8.27770770e-01 1.06526351e+00
9.67743397e-02 -7.46920884e-01 1.03076732e+00 4.68857914e-01
-1.48725402e+00 -2.74554431e-01 1.90490872e-01 -1.48809657e-01
4.93875176e-01 9.25095454e-02 -6.36352420e-01 6.86461568e-01
1.00840163e+00 7.12371409e-01 -7.92173684e-01 9.53486145e-01
-4.32896942e-01 4.48000096e-02 -1.27416313e-01 -2.81799287e-01
-2.76590288e-02 9.32991952e-02 5.77080667e-01 1.53013253e+00
3.11162114e-01 1.28661543e-01 4.87540901e-01 3.80407393e-01
7.14216530e-02 -1.96416050e-01 -7.55260408e-01 -7.85350949e-02
4.97609943e-01 1.31125760e+00 -3.98011953e-01 -1.62501782e-01
-6.17301047e-01 1.39764285e+00 5.30392587e-01 3.62641543e-01
-8.23303223e-01 -4.38846946e-01 5.22446096e-01 -4.95049357e-01
1.42435312e-01 -6.92527652e-01 -1.88317092e-03 -1.03945422e+00
1.58367306e-01 -8.96134615e-01 5.62245607e-01 -1.25838149e+00
-1.39843571e+00 3.42401057e-01 1.42126545e-01 -1.28195000e+00
-2.50890136e-01 -4.85669941e-01 -2.05900133e-01 3.73255193e-01
-9.42904532e-01 -1.68466771e+00 -7.55382299e-01 9.70631957e-01
6.09080732e-01 -5.82957715e-02 9.69502091e-01 1.03483707e-01
-5.75546563e-01 3.57215524e-01 -3.81811351e-01 3.75483572e-01
7.30421185e-01 -1.20075798e+00 -9.17228609e-02 7.74791837e-01
1.08383782e-01 7.89772451e-01 6.49420738e-01 -6.15126014e-01
-1.85589278e+00 -7.56594718e-01 7.38879859e-01 -7.07580507e-01
4.84701335e-01 -4.34136897e-01 -6.87254786e-01 9.96490479e-01
6.25397444e-01 -2.68104076e-01 1.06317186e+00 1.38046965e-01
-2.86964983e-01 2.44345814e-01 -9.15951669e-01 5.46155632e-01
1.07052231e+00 -5.61298907e-01 -1.00084221e+00 -9.61858928e-02
4.94811654e-01 -3.93947996e-02 -9.60762739e-01 4.54396904e-01
7.02603877e-01 -7.60907531e-01 8.87123525e-01 -5.02887189e-01
2.64505185e-02 -6.24061584e-01 -5.40715992e-01 -1.41139734e+00
-4.49839473e-01 -6.36297107e-01 -3.48844230e-01 1.12962925e+00
1.49702922e-01 -4.56436604e-01 1.78791851e-01 7.38884866e-01
-4.86336984e-02 -3.38754281e-02 -8.77064645e-01 -4.50064808e-01
-7.25878179e-01 -5.60577035e-01 3.75750721e-01 8.86487544e-01
7.68378913e-01 6.07981265e-01 -7.49372184e-01 4.45980966e-01
7.21209586e-01 -4.32963192e-01 1.20900238e+00 -1.25065315e+00
-1.21984035e-01 -1.13554187e-01 -7.68906176e-01 -1.03066802e+00
2.29136467e-01 -3.29623818e-01 5.31571627e-01 -1.85224962e+00
5.00379324e-01 3.62627178e-01 -3.70743245e-01 6.75921738e-01
-2.50749856e-01 6.78621815e-04 1.73115283e-01 1.16152875e-01
-1.59378612e+00 9.76206422e-01 7.12373018e-01 -3.47375125e-01
-1.65191621e-01 -4.09368157e-01 -5.94842970e-01 7.88249016e-01
7.53168702e-01 -1.04435056e-01 -1.86268717e-01 -2.82389611e-01
2.16304600e-01 4.06337902e-02 7.56541133e-01 -1.18852186e+00
6.13368869e-01 -2.74504542e-01 5.33085763e-01 -7.43046820e-01
9.22288358e-01 -8.69708061e-01 8.76395330e-02 2.50813793e-02
-1.73241839e-01 1.03054039e-01 1.97469313e-02 8.75333726e-01
-2.84192175e-01 3.23252559e-01 1.15835138e-01 2.41874587e-02
-1.03979063e+00 -2.95810401e-01 -9.48690951e-01 -2.66494483e-01
1.05194068e+00 -4.42544855e-02 -3.02400023e-01 -6.87616706e-01
-7.70426989e-01 6.13216698e-01 1.01224624e-01 7.91415215e-01
8.16721499e-01 -1.58724022e+00 -5.06044328e-01 -4.19880599e-02
7.97084093e-01 -5.48006669e-02 4.55187500e-01 9.63447571e-01
5.12788743e-02 4.63162869e-01 -2.59492755e-01 -6.99551642e-01
-1.48541665e+00 3.18263590e-01 2.83175111e-02 -5.49862124e-02
-2.62939066e-01 7.06439376e-01 -7.50011951e-02 -6.15529001e-01
7.45576203e-01 -2.39148647e-01 -3.95285815e-01 6.20018691e-02
6.80961311e-01 7.55299687e-01 -3.00435841e-01 -1.27373362e+00
-6.53190196e-01 4.45423126e-01 4.31653082e-01 -4.90617484e-01
1.28953159e+00 -4.46073174e-01 -1.08479939e-01 8.63117516e-01
7.28518546e-01 -3.63537639e-01 -1.27107453e+00 -3.14554572e-01
2.10769355e-01 -1.72741517e-01 -3.49220484e-01 -7.59410858e-01
-4.46964234e-01 8.20033431e-01 5.57076573e-01 -2.04774388e-03
9.81067300e-01 3.09215248e-01 4.74924088e-01 8.70989382e-01
7.40001023e-01 -1.08783853e+00 5.55758059e-01 5.13807297e-01
1.20205069e+00 -1.61619961e+00 -2.15203419e-01 -2.60865062e-01
-1.08811080e+00 7.56335199e-01 6.95099592e-01 3.76673698e-01
2.30614394e-01 -3.54266278e-02 -1.92605350e-02 -4.18850541e-01
-4.77362484e-01 -5.12512147e-01 4.11228299e-01 9.29065645e-01
2.51887918e-01 2.05631465e-01 1.99060425e-01 1.15236843e+00
9.12440196e-02 -1.61084384e-01 1.46400943e-01 1.46303856e+00
-4.71549481e-01 -7.50990033e-01 -9.13258612e-01 -6.82011545e-02
-1.92933008e-02 4.37529057e-01 -7.85834849e-01 6.83861077e-01
1.02165148e-01 1.57208323e+00 -3.73230606e-01 -6.77874506e-01
6.13927841e-01 2.54756749e-01 5.14334381e-01 -2.34144241e-01
-5.02396762e-01 -3.81304584e-02 3.58244598e-01 -8.82237971e-01
-9.82881844e-01 -8.15062463e-01 -1.27675653e+00 8.11129510e-02
1.46836136e-02 4.13970882e-03 5.61180234e-01 1.08345902e+00
4.63988513e-01 4.58338529e-01 2.47992918e-01 -1.48878908e+00
7.84478560e-02 -1.33666801e+00 -2.92923629e-01 5.68713903e-01
2.97731370e-01 -1.17104352e+00 -8.93235579e-02 2.09124163e-01] | [13.087021827697754, 4.878861427307129] |
84567dab-3de1-4c46-a4cc-078d17ac0595 | a-comprehensive-multi-scale-approach-for | 2307.03270 | null | https://arxiv.org/abs/2307.03270v1 | https://arxiv.org/pdf/2307.03270v1.pdf | A Comprehensive Multi-scale Approach for Speech and Dynamics Synchrony in Talking Head Generation | Animating still face images with deep generative models using a speech input signal is an active research topic and has seen important recent progress. However, much of the effort has been put into lip syncing and rendering quality while the generation of natural head motion, let alone the audio-visual correlation between head motion and speech, has often been neglected. In this work, we propose a multi-scale audio-visual synchrony loss and a multi-scale autoregressive GAN to better handle short and long-term correlation between speech and the dynamics of the head and lips. In particular, we train a stack of syncer models on multimodal input pyramids and use these models as guidance in a multi-scale generator network to produce audio-aligned motion unfolding over diverse time scales. Our generator operates in the facial landmark domain, which is a standard low-dimensional head representation. The experiments show significant improvements over the state of the art in head motion dynamics quality and in multi-scale audio-visual synchrony both in the landmark domain and in the image domain. | ['Xavier Alameda-Pineda', 'Dominique Vaufreydaz', 'Louis Airale'] | 2023-07-04 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [-1.71928734e-01 1.34619534e-01 2.49953210e-01 -2.52916634e-01
-9.25361693e-01 -2.24397525e-01 7.35753536e-01 -6.83935761e-01
1.15813702e-01 3.80450517e-01 6.71921492e-01 4.34950978e-01
3.24800402e-01 -3.40552926e-01 -6.99029088e-01 -9.61507797e-01
6.59594759e-02 2.48846650e-01 1.12863831e-01 -3.09283704e-01
-3.71741802e-01 4.42636847e-01 -1.83608484e+00 2.82516629e-01
3.50035489e-01 7.78482914e-01 2.38020299e-03 1.00559914e+00
1.81835979e-01 7.91487873e-01 -7.16580868e-01 -4.48845863e-01
-5.91062270e-02 -9.91220176e-01 -3.96830320e-01 1.92766756e-01
7.00518072e-01 -2.15849638e-01 -4.83146638e-01 6.97661042e-01
1.12681437e+00 -3.30148824e-02 4.86209989e-01 -1.59798205e+00
-3.06300640e-01 2.42240652e-01 -7.45350838e-01 -8.45661536e-02
5.12646556e-01 3.45452130e-01 6.88692927e-01 -7.94897199e-01
9.20173705e-01 1.79321170e+00 4.31085646e-01 8.53433013e-01
-1.21305835e+00 -8.39300811e-01 1.62111223e-02 2.93176383e-01
-1.35891962e+00 -7.55634427e-01 9.72912788e-01 -3.72597694e-01
6.00763798e-01 1.18508384e-01 8.91411543e-01 1.60866320e+00
2.42655545e-01 8.13660562e-01 7.64358461e-01 -2.83811510e-01
1.00412406e-02 -6.72874078e-02 -6.36731088e-01 4.55689013e-01
-5.28785765e-01 2.02527151e-01 -9.22571182e-01 6.86279237e-02
9.93980646e-01 -5.09598196e-01 -2.81164765e-01 -1.71861514e-01
-9.65450346e-01 8.45146596e-01 1.58621147e-01 3.19572747e-01
-3.19235593e-01 3.85531843e-01 2.33160496e-01 -2.52002086e-02
5.41399002e-01 -3.26530784e-02 1.07997760e-01 -3.29593927e-01
-1.33692074e+00 5.06343007e-01 6.32377267e-01 5.79135180e-01
2.33344346e-01 5.52413166e-01 -1.85277805e-01 8.65677178e-01
5.26968122e-01 7.98961520e-01 5.15896142e-01 -1.23789728e+00
1.72653198e-01 7.42001832e-02 -2.36708462e-01 -9.67704415e-01
-4.35302585e-01 -1.88583702e-01 -1.03420734e+00 5.11930168e-01
4.04749990e-01 -3.46324891e-01 -8.53556335e-01 2.37493658e+00
6.27040386e-01 7.24464476e-01 -2.36080196e-02 1.04700994e+00
8.76960695e-01 7.79062331e-01 6.05653133e-03 -3.43640685e-01
1.34469986e+00 -7.54554391e-01 -1.17717111e+00 -8.56159069e-03
-5.97528964e-02 -1.02552009e+00 8.75911534e-01 2.62117572e-02
-1.64339042e+00 -7.57374644e-01 -7.66710103e-01 -2.36795783e-01
1.05313003e-01 -1.68198664e-02 2.24038586e-01 3.48081321e-01
-1.47895086e+00 3.14200193e-01 -9.36385751e-01 -2.00676262e-01
3.60354558e-02 1.44332856e-01 -4.15550083e-01 2.49043673e-01
-1.30165243e+00 6.06589496e-01 -3.09801489e-01 -1.78092979e-02
-8.98039162e-01 -7.02627420e-01 -1.09908426e+00 -1.12507746e-01
-4.49760146e-02 -1.01802456e+00 1.31240392e+00 -8.49899709e-01
-1.91667247e+00 7.27757454e-01 -4.24529344e-01 -2.45958492e-01
7.78970361e-01 -5.37348688e-02 -4.40764278e-01 2.81300485e-01
-1.10195920e-01 1.15284550e+00 1.29420280e+00 -1.16615295e+00
-4.10112232e-01 -2.97391117e-01 -3.41651946e-01 3.43243599e-01
-7.02431099e-03 1.66380540e-01 -5.14329374e-01 -9.95308280e-01
-6.43780977e-02 -1.32310808e+00 2.37455562e-01 -1.44953635e-02
-2.76898563e-01 -4.34130095e-02 9.84610438e-01 -6.72088087e-01
1.13915038e+00 -2.08367085e+00 6.28309727e-01 -2.12924048e-01
1.90730300e-02 8.25104024e-03 -2.53217518e-01 1.52005881e-01
-3.27683628e-01 -1.41660020e-01 1.41373247e-01 -9.51557577e-01
-9.89646092e-02 -4.51027453e-02 -4.49633569e-01 4.28634435e-01
2.82892019e-01 1.02234328e+00 -6.26199067e-01 -4.72919285e-01
1.37231022e-01 1.20261705e+00 -6.98272586e-01 4.36845243e-01
-1.35846615e-01 9.04834867e-01 2.13429213e-01 5.50353825e-01
5.83614945e-01 -7.42789581e-02 -1.33810475e-01 -1.74991310e-01
1.15770161e-01 -7.72189721e-02 -1.18439734e+00 1.83400559e+00
-4.98276412e-01 8.91134381e-01 3.18252921e-01 -2.88708061e-01
6.47555351e-01 8.04331005e-01 7.51286089e-01 -6.68163478e-01
7.62913749e-02 3.44290771e-02 2.33891029e-02 -5.24977565e-01
2.65553802e-01 -4.54048365e-01 3.06118466e-02 3.62951577e-01
3.00540775e-01 -5.77094555e-01 -2.28807498e-02 4.61521931e-02
6.22866809e-01 3.19831550e-01 -1.68329477e-01 2.46063590e-01
4.44156468e-01 -7.81990588e-01 2.66093433e-01 -1.17133960e-01
-1.38556396e-04 1.18813431e+00 6.24461293e-01 2.93514859e-02
-1.25633419e+00 -1.00963485e+00 1.34543896e-01 1.13341796e+00
-1.36857465e-01 -5.39421737e-01 -9.97393668e-01 -8.29325020e-02
-1.25856861e-01 4.52867776e-01 -7.80464828e-01 -2.26436749e-01
-7.35610187e-01 -5.77827811e-01 6.87215686e-01 5.39345562e-01
3.02687794e-01 -1.37251699e+00 -5.10503471e-01 1.54629946e-01
-5.11034369e-01 -1.32714391e+00 -7.92129636e-01 -3.85937691e-01
-4.38109905e-01 -6.95007622e-01 -1.31588852e+00 -6.19645774e-01
1.41307756e-01 -6.34733662e-02 9.30528164e-01 -5.58586359e-01
-5.06876349e-01 3.74669254e-01 3.71986106e-02 -4.81120378e-01
-5.75539827e-01 -2.54611701e-01 1.43264890e-01 3.54143530e-01
-2.16843799e-01 -8.87510300e-01 -6.60338342e-01 4.02213901e-01
-8.86064649e-01 2.68286020e-01 1.24554083e-01 7.52100289e-01
3.18439603e-01 -3.86352569e-01 6.12959146e-01 -6.92735910e-02
6.56716347e-01 -3.99077773e-01 -3.80843014e-01 -1.51147306e-01
-7.53609315e-02 8.04036707e-02 1.85926676e-01 -7.03496814e-01
-1.07683289e+00 3.85443717e-02 -4.56378728e-01 -8.25523734e-01
-9.89492983e-02 -1.20408267e-01 -3.65194201e-01 2.16029793e-01
4.73018289e-01 9.76974145e-02 2.80100554e-01 -1.88842967e-01
5.50010741e-01 3.75475794e-01 8.74227405e-01 -2.28879482e-01
6.67485833e-01 5.51832378e-01 2.76437372e-01 -1.10910475e+00
-2.95075595e-01 -1.14030346e-01 -4.42889541e-01 -4.47028160e-01
1.12189114e+00 -1.09172511e+00 -8.39505851e-01 8.50823224e-01
-1.31704271e+00 -5.27065396e-01 -2.76187032e-01 3.42820376e-01
-8.88536990e-01 8.99414420e-02 -7.24201024e-01 -7.91990161e-01
-2.83922225e-01 -1.44201648e+00 1.61861074e+00 3.70552421e-01
-3.94859672e-01 -8.08730185e-01 3.25461388e-01 1.73965007e-01
4.17604655e-01 5.49196899e-01 5.71334839e-01 8.29170942e-02
-5.02168477e-01 -6.17192648e-02 2.06037119e-01 1.56888515e-01
-4.52409908e-02 3.33379328e-01 -1.22788930e+00 -2.79652447e-01
-3.52849402e-02 -4.48483855e-01 5.85934162e-01 9.34279680e-01
6.13669932e-01 -2.64498949e-01 -2.60042623e-02 8.01017404e-01
7.73648024e-01 1.62792176e-01 7.09166586e-01 -3.06752950e-01
6.27768815e-01 9.57187772e-01 2.44195044e-01 5.97463369e-01
4.00378823e-01 1.18954062e+00 4.44992423e-01 -2.70381629e-01
-7.38363087e-01 -4.07124937e-01 5.65319479e-01 7.65699685e-01
-9.07597169e-02 -1.68259025e-01 -6.22422040e-01 4.46051896e-01
-1.76700747e+00 -1.20917726e+00 3.21665257e-01 1.83184612e+00
8.47435653e-01 -2.61123538e-01 5.35338998e-01 1.54380798e-01
6.14291191e-01 3.06304246e-01 -4.98763084e-01 -1.99558511e-01
-4.41642314e-01 2.73855150e-01 -3.25904042e-01 6.75944567e-01
-9.10424352e-01 1.00674367e+00 5.97593880e+00 7.23732293e-01
-1.73115563e+00 9.64114591e-02 5.89968324e-01 -5.25623739e-01
-2.61285335e-01 -5.38269460e-01 -8.52452278e-01 5.07922173e-01
1.12142551e+00 -2.29059801e-01 2.49293014e-01 5.82230747e-01
5.37368894e-01 4.85353693e-02 -1.13046241e+00 1.29297137e+00
3.45100552e-01 -1.08868349e+00 -1.13268629e-01 1.74224287e-01
7.51454651e-01 -3.21223825e-01 5.49290419e-01 3.90296355e-02
6.21681772e-02 -1.32365370e+00 1.06590569e+00 5.42795777e-01
1.02417433e+00 -8.20520461e-01 3.12632203e-01 1.96334735e-01
-1.43180168e+00 2.04539999e-01 1.47228986e-01 5.81856489e-01
6.16265893e-01 1.69404913e-02 -7.13797450e-01 1.93498299e-01
7.53243864e-01 4.46922839e-01 -3.21485162e-01 8.58352065e-01
-2.37514898e-01 3.93428117e-01 -3.39963853e-01 4.19420391e-01
-2.02975949e-04 7.93185905e-02 8.46633911e-01 1.20750678e+00
4.55674022e-01 3.06493752e-02 -2.22297803e-01 7.19219387e-01
-4.72513842e-04 -4.79118526e-02 -6.52954459e-01 6.61803037e-02
1.27182856e-01 1.21765316e+00 -3.17278653e-01 -8.12657475e-02
-1.12658143e-01 1.00626016e+00 -2.12201625e-01 4.42800105e-01
-1.12513518e+00 2.18112711e-02 1.07521176e+00 2.87880987e-01
1.28890574e-01 -1.71593562e-01 4.08424139e-02 -1.00179589e+00
-1.13329083e-01 -9.45308447e-01 9.14728716e-02 -1.12093174e+00
-6.62132382e-01 7.87808955e-01 5.26999496e-03 -1.06398082e+00
-1.17699385e+00 -5.02643622e-02 -5.51688373e-01 9.68042791e-01
-1.12475610e+00 -1.32431030e+00 -4.34403449e-01 9.40562129e-01
6.99778795e-01 4.39278968e-02 8.07466567e-01 3.32766294e-01
-3.24227005e-01 8.57541502e-01 -4.49246645e-01 -2.68604141e-02
9.49868619e-01 -8.28163207e-01 4.44078475e-01 6.27061963e-01
1.55694827e-01 3.52695107e-01 1.02337110e+00 -3.70479405e-01
-1.07767177e+00 -9.53399003e-01 7.38367558e-01 -2.82329082e-01
3.12271416e-01 -5.09955227e-01 -8.84141445e-01 4.48544741e-01
4.85123038e-01 1.37288272e-01 5.22233307e-01 -5.76897383e-01
-1.93776891e-01 -1.48665950e-01 -9.20549929e-01 8.05925429e-01
8.00576448e-01 -6.49063528e-01 -1.30498528e-01 -4.12101969e-02
6.62579119e-01 -6.69188023e-01 -6.27642095e-01 3.42420876e-01
8.88588607e-01 -1.04176176e+00 1.04768264e+00 -4.34678674e-01
4.77394372e-01 -2.41677716e-01 8.74600187e-03 -1.38364637e+00
-3.58722731e-02 -1.12479174e+00 -2.72670418e-01 1.34333062e+00
1.27349883e-01 -4.18756716e-02 7.71013975e-01 3.43142629e-01
1.62627831e-01 -5.50323248e-01 -1.06594229e+00 -3.44726175e-01
9.56606790e-02 -3.33500296e-01 3.93644065e-01 6.71186626e-01
-1.58585563e-01 5.39165497e-01 -7.87491441e-01 -3.08314655e-02
5.60968339e-01 -3.22815590e-02 1.08473253e+00 -1.04780889e+00
-2.22038046e-01 -4.95209038e-01 -4.61227119e-01 -7.72933066e-01
3.07605952e-01 -5.53974628e-01 -3.76474559e-02 -1.26389718e+00
-9.61614400e-02 3.70312333e-01 3.69314432e-01 2.30274171e-01
3.39703970e-02 3.25704098e-01 4.70975310e-01 -3.74095840e-03
-1.08146124e-01 8.53182971e-01 1.43964684e+00 1.88011318e-01
-3.41301113e-01 7.20720217e-02 -2.43353114e-01 7.39003956e-01
3.49810958e-01 -1.89210236e-01 -5.96314490e-01 -1.06927142e-01
-2.79746763e-02 7.99191892e-01 4.17563856e-01 -1.14282024e+00
3.41461539e-01 5.33274002e-02 5.29950202e-01 -4.39873606e-01
1.08172560e+00 -5.78195632e-01 5.78774750e-01 1.70544103e-01
-4.23566848e-01 2.87950397e-01 2.88671762e-01 4.10613239e-01
-4.57476884e-01 6.80838943e-01 1.29729283e+00 1.02011234e-01
-1.60850778e-01 5.02896547e-01 -3.72967646e-02 1.79722518e-01
9.44882870e-01 5.20663597e-02 -5.41623980e-02 -1.16088712e+00
-1.13117981e+00 -1.05665572e-01 3.43976408e-01 7.44944692e-01
6.44578457e-01 -1.59337270e+00 -9.43498611e-01 5.29837132e-01
-2.82143265e-01 -2.23235950e-01 5.24705946e-01 9.61114228e-01
-1.39782220e-01 3.95757765e-01 -3.57251048e-01 -9.40658689e-01
-1.55772650e+00 1.70090660e-01 5.00216246e-01 -3.02456226e-02
-6.53231084e-01 9.51478481e-01 4.31367010e-01 2.32000843e-01
4.78795230e-01 -1.34527892e-01 -3.11430514e-01 4.29374188e-01
5.75628519e-01 2.55438536e-01 -1.97551697e-01 -1.24192953e+00
-1.95102662e-01 9.43478763e-01 4.49568897e-01 -7.55426049e-01
1.19612694e+00 -2.85246372e-01 1.55715585e-01 5.36195755e-01
1.40127218e+00 4.99060452e-02 -1.58317816e+00 1.36673301e-01
-4.72738355e-01 -2.20917806e-01 -2.05633938e-01 -5.46555996e-01
-1.33953059e+00 1.28771794e+00 6.88496232e-01 -1.50432242e-02
1.22716379e+00 1.10941723e-01 8.37477446e-01 -5.26286662e-01
7.68076032e-02 -8.56721282e-01 6.86637521e-01 3.61856014e-01
1.34915912e+00 -9.46496248e-01 -6.15017533e-01 -2.49777153e-01
-8.76001418e-01 9.95679438e-01 5.41370928e-01 1.39015866e-02
7.60548353e-01 6.41286194e-01 2.72047609e-01 1.33145764e-01
-9.69686627e-01 -2.17484713e-01 4.67706770e-01 6.68167233e-01
6.17197573e-01 -3.12536120e-01 2.14176431e-01 3.58500004e-01
-5.90437710e-01 -6.23409264e-02 2.78630614e-01 3.49155962e-01
-4.45661917e-02 -9.37570632e-01 -4.89022017e-01 -4.63046551e-01
-7.06137598e-01 1.54213756e-01 -2.90692389e-01 7.70212948e-01
7.17471167e-03 9.60840404e-01 1.05601616e-01 -3.49123597e-01
4.35860068e-01 2.64428228e-01 5.99131823e-01 -2.56784141e-01
-3.88137788e-01 7.68369019e-01 -2.48129666e-01 -8.28918755e-01
-3.57965231e-01 -6.70042217e-01 -1.21323478e+00 -3.05279553e-01
6.73405677e-02 -2.37233058e-01 8.62385988e-01 5.82977831e-01
4.21805829e-01 6.88482881e-01 5.81363380e-01 -1.42846298e+00
-3.49409044e-01 -1.08509839e+00 -5.44300139e-01 5.31356514e-01
7.35647917e-01 -6.64331615e-01 -5.43781161e-01 1.91654816e-01] | [13.217887878417969, -0.4503404498100281] |
aeb525e5-7270-48d9-b754-8d929a1898ef | semi-supervised-clustering-with-inaccurate | 2104.02146 | null | https://arxiv.org/abs/2104.02146v1 | https://arxiv.org/pdf/2104.02146v1.pdf | Semi-Supervised Clustering with Inaccurate Pairwise Annotations | Pairwise relational information is a useful way of providing partial supervision in domains where class labels are difficult to acquire. This work presents a clustering model that incorporates pairwise annotations in the form of must-link and cannot-link relations and considers possible annotation inaccuracies (i.e., a common setting when experts provide pairwise supervision). We propose a generative model that assumes Gaussian-distributed data samples along with must-link and cannot-link relations generated by stochastic block models. We adopt a maximum-likelihood approach and demonstrate that, even when supervision is weak and inaccurate, accounting for relational information significantly improves clustering performance. Relational information also helps to detect meaningful groups in real-world datasets that do not fit the original data-distribution assumptions. Additionally, we extend the model to integrate prior knowledge of experts' accuracy and discuss circumstances in which the use of this knowledge is beneficial. | ['Thibaut Vidal', 'Michel Gendreau', 'Daniel Gribel'] | 2021-04-05 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 5.04655614e-02 5.82700193e-01 -4.38995987e-01 -8.26010585e-01
-7.90167451e-01 -5.66276610e-01 5.06425202e-01 4.09942269e-01
-1.85516536e-01 9.54653084e-01 1.41957283e-01 -3.40186894e-01
-7.09228098e-01 -6.35670960e-01 -8.36024582e-01 -6.88785255e-01
3.11856121e-01 9.71449554e-01 2.28869230e-01 1.18286900e-01
8.32132548e-02 2.67438382e-01 -1.38269269e+00 1.67899117e-01
1.09873748e+00 2.13482648e-01 -5.62443351e-03 3.41920823e-01
1.31887197e-01 9.34984565e-01 -8.12435627e-01 -6.62653625e-01
4.60363895e-01 -1.89266399e-01 -8.69129598e-01 2.58947968e-01
4.23501909e-01 -2.52261788e-01 -7.57028535e-02 9.96848643e-01
1.87578052e-01 3.15018259e-02 9.57517445e-01 -1.53569543e+00
-6.02738321e-01 8.61177266e-01 -5.75029910e-01 4.51760255e-02
4.10906643e-01 -1.28555328e-01 1.19917047e+00 -3.04437131e-01
6.57973886e-01 1.22660363e+00 9.85089004e-01 8.85785446e-02
-1.65295637e+00 -5.94395101e-01 3.65912348e-01 9.23578292e-02
-1.65338349e+00 -1.32287830e-01 5.25006831e-01 -6.12954140e-01
6.33909702e-01 4.13429923e-02 1.08817361e-01 1.06350851e+00
-1.03267185e-01 5.42756855e-01 1.16848695e+00 -4.84559417e-01
1.66489795e-01 3.35054189e-01 4.76287276e-01 3.59161258e-01
6.68974876e-01 -2.51033485e-01 -6.92580938e-01 -4.68468338e-01
7.11474061e-01 -7.03137415e-03 -1.74583748e-01 -6.40186727e-01
-1.13015270e+00 6.13920689e-01 2.69301891e-01 3.19324911e-01
-2.96714544e-01 7.02735316e-03 -6.15061028e-03 3.00778270e-01
4.84740466e-01 3.11115921e-01 -3.64327341e-01 1.68027394e-02
-1.22101843e+00 1.89577192e-01 8.87740016e-01 1.31244779e+00
9.44201350e-01 -4.07231867e-01 3.07844691e-02 7.22245514e-01
3.32816452e-01 1.97306097e-01 6.05315156e-03 -1.43324411e+00
4.95325238e-01 5.60263455e-01 6.30617440e-01 -1.00094497e+00
-2.73663491e-01 -3.43117416e-01 -4.94326502e-01 -5.31589016e-02
9.40085113e-01 2.12077349e-02 -7.11185455e-01 1.70301366e+00
5.09614289e-01 2.14576900e-01 -9.52547193e-02 7.61370718e-01
2.43388608e-01 -4.62709740e-02 -2.49368162e-03 -3.63239735e-01
1.05177116e+00 -6.10386312e-01 -8.84982169e-01 1.59221604e-01
8.51481140e-01 -5.34748852e-01 8.14684391e-01 5.49144804e-01
-8.38442862e-01 -4.53045547e-01 -8.39557350e-01 -1.14440367e-01
-1.27777085e-01 8.46209005e-02 5.60733676e-01 6.11274004e-01
-1.12283552e+00 5.97651601e-01 -1.03603613e+00 -4.26331848e-01
3.01299751e-01 3.46987486e-01 -5.32983780e-01 -4.15623188e-01
-9.70565975e-01 8.85084152e-01 1.67526796e-01 6.80595031e-03
-6.55657709e-01 -5.29581368e-01 -6.64424896e-01 7.16685131e-02
7.80919135e-01 -6.08947515e-01 1.14460301e+00 -9.19532418e-01
-6.48133576e-01 6.47607028e-01 -4.30920035e-01 -3.74952108e-01
6.84107721e-01 -2.12204397e-01 -1.21249147e-01 2.00563058e-01
4.06712949e-01 4.11061436e-01 4.22598481e-01 -1.70641148e+00
-3.91538382e-01 -4.05377805e-01 8.24920535e-02 3.94240886e-01
-2.17676580e-01 -1.03195116e-01 -2.00225189e-01 -4.54210043e-01
3.80633652e-01 -1.02129996e+00 -2.85947561e-01 -4.72703483e-03
-7.02218652e-01 -4.48002070e-01 5.99429548e-01 -6.86322391e-01
1.16026795e+00 -1.97039938e+00 -5.26625328e-02 8.35959494e-01
2.70846039e-01 -1.30982503e-01 2.92926997e-01 5.85170507e-01
-1.78375952e-02 3.80861491e-01 -4.13050234e-01 -5.29451489e-01
7.32950047e-02 7.09823072e-01 1.31296083e-01 4.72878635e-01
4.57282141e-02 3.96454394e-01 -9.20790374e-01 -7.14574873e-01
-4.75379117e-02 2.60304272e-01 -2.52933323e-01 1.65712491e-01
-1.65717930e-01 4.54800129e-01 -2.53751546e-01 5.13935745e-01
5.86905658e-01 -5.73220670e-01 7.42371678e-01 -1.36397704e-01
1.69304222e-01 3.86454254e-01 -1.48825300e+00 1.51862013e+00
-1.51998205e-02 3.62449020e-01 3.67834777e-01 -9.92393553e-01
7.87816525e-01 3.83568674e-01 5.01700282e-01 1.19572707e-01
2.58340966e-02 5.34250326e-02 1.45792559e-01 -3.06624919e-01
3.09169084e-01 -1.62637472e-01 1.55731052e-01 6.62802219e-01
8.64321440e-02 -1.03965454e-01 4.44706827e-01 6.11726999e-01
1.32357240e+00 9.69118550e-02 3.38906586e-01 -2.74281472e-01
3.86277102e-02 8.77522528e-02 8.26502323e-01 9.80331182e-01
-1.37259141e-01 6.28713250e-01 8.36037934e-01 1.40423506e-01
-1.01706254e+00 -1.02185905e+00 -3.20852399e-01 8.78194153e-01
3.26143727e-02 -4.58628893e-01 -6.50752425e-01 -9.47922468e-01
1.38965622e-01 8.00107300e-01 -7.33509123e-01 2.61571296e-02
9.09532309e-02 -6.58643067e-01 4.57997829e-01 6.52357340e-01
5.62782772e-02 -3.75310123e-01 -8.37549195e-02 2.44234931e-02
-2.87432253e-01 -1.00063920e+00 -1.05433144e-01 4.05575722e-01
-7.94926345e-01 -1.47459292e+00 -1.89669400e-01 -3.32474530e-01
1.04347277e+00 2.55883783e-01 1.11222434e+00 3.28533232e-01
1.32464573e-01 6.01068497e-01 -5.69940865e-01 -1.35895699e-01
-3.05832535e-01 -4.47036959e-02 7.97148570e-02 -1.53031260e-01
8.01232159e-01 -5.76896429e-01 -9.61020887e-02 6.93457305e-01
-8.45313609e-01 -2.11849019e-01 3.89475316e-01 9.58705962e-01
4.95772749e-01 3.39240402e-01 5.08796632e-01 -1.50376368e+00
5.77939808e-01 -6.10589802e-01 -4.45984930e-01 6.15978420e-01
-9.34934795e-01 1.37070124e-03 1.43948555e-01 -3.52784038e-01
-1.40853727e+00 -2.46933172e-03 3.17427307e-01 -6.11561358e-01
-5.86521089e-01 6.15465343e-01 -3.56934994e-01 1.25640139e-01
9.14469838e-01 -5.01986861e-01 1.98204014e-02 -5.19140601e-01
2.12177426e-01 7.74585307e-01 3.34976315e-01 -9.70723271e-01
7.87916064e-01 6.04639947e-01 -2.25290898e-02 -3.93890381e-01
-1.14054525e+00 -7.07230389e-01 -1.32395828e+00 -1.01779707e-01
6.41700804e-01 -1.06168652e+00 -5.42672098e-01 -4.49322164e-02
-8.71076286e-01 -3.65357161e-01 -3.46274585e-01 6.75622165e-01
-3.58010620e-01 5.82352221e-01 -6.81226194e-01 -1.08005869e+00
5.82342505e-01 -1.08528972e+00 9.60162520e-01 -5.78389615e-02
-5.55077791e-01 -1.19812799e+00 -1.48405999e-01 7.23471999e-01
-1.82708710e-01 1.60845459e-01 5.94885230e-01 -1.17575097e+00
-6.81205511e-01 -1.00546487e-01 -1.51508227e-01 3.39643896e-01
2.85166055e-01 2.43377954e-01 -9.71935689e-01 -1.92118481e-01
-1.63284197e-01 -3.47382724e-01 4.49956477e-01 3.14673930e-01
8.35618138e-01 -2.48199776e-01 -3.80584508e-01 2.22625509e-02
1.12197363e+00 1.28977988e-02 2.72091120e-01 4.62064855e-02
8.66011858e-01 1.19165111e+00 8.96800756e-01 4.70729470e-01
7.47580171e-01 6.42450213e-01 2.59262949e-01 1.25657201e-01
2.90523082e-01 -3.88990372e-01 -5.26459031e-02 5.49820006e-01
2.61097439e-02 -1.90716252e-01 -1.20960617e+00 6.78457499e-01
-2.22567558e+00 -9.80605304e-01 -7.46136546e-01 2.17627215e+00
1.06910098e+00 2.19972134e-01 1.26855209e-01 2.89106816e-01
9.61735189e-01 -4.06697780e-01 -3.13918918e-01 2.03258991e-02
-1.03319138e-01 -8.44909549e-02 5.90980470e-01 7.17721701e-01
-8.55630517e-01 6.79276824e-01 6.95815372e+00 6.21105134e-01
-9.10791010e-02 2.10361987e-01 5.53271472e-01 1.55264035e-01
-4.32075143e-01 3.83147091e-01 -6.54267073e-01 4.67714369e-01
8.39642346e-01 2.32352823e-01 1.41415462e-01 6.80497289e-01
1.52404144e-01 -4.37813580e-01 -1.45178473e+00 5.16492188e-01
6.67097047e-02 -7.63219297e-01 -3.42476338e-01 3.91362160e-01
9.56690848e-01 -2.83796340e-01 -2.20209315e-01 -8.20253938e-02
1.34577441e+00 -8.30021143e-01 5.41782618e-01 7.03459203e-01
5.01935780e-01 -6.82560027e-01 1.13070667e+00 7.20874369e-01
-7.38961756e-01 -1.83888394e-02 -3.28041732e-01 -1.48268342e-01
1.56268761e-01 1.04310656e+00 -9.78809237e-01 9.38582659e-01
6.86631918e-01 9.77397621e-01 -7.28586614e-01 9.41531479e-01
-6.37146890e-01 8.04833055e-01 -4.10297424e-01 4.97387022e-01
-1.65864632e-01 -3.79703492e-01 2.35098526e-01 9.60096359e-01
2.55133778e-01 5.07925972e-02 3.95836055e-01 7.26913154e-01
-1.89718790e-02 -2.50538290e-01 -7.60966003e-01 1.64237976e-01
9.81292188e-01 1.01289403e+00 -8.78621697e-01 -4.37776864e-01
-5.24083138e-01 7.29198039e-01 4.62897062e-01 5.68931282e-01
-3.69851023e-01 6.44840896e-02 2.76752114e-01 2.48717949e-01
1.43947288e-01 -1.86908454e-01 -4.52191889e-01 -9.57570195e-01
5.41249895e-03 -8.94984007e-01 5.74829638e-01 -1.03231502e+00
-1.71658540e+00 -1.09589040e-01 4.48056966e-01 -9.39391732e-01
-3.70031506e-01 -4.47462380e-01 -2.44126290e-01 8.50272357e-01
-1.02703750e+00 -1.21933675e+00 -2.03654483e-01 5.24186909e-01
-8.42459351e-02 2.25066200e-01 4.89757925e-01 3.00541967e-01
-4.28724080e-01 5.43114305e-01 1.44723640e-03 1.20510466e-01
1.24837697e+00 -1.62232459e+00 -2.30688199e-01 7.78811038e-01
3.66239786e-01 1.13779700e+00 9.16202307e-01 -1.03817487e+00
-4.80087459e-01 -8.49013567e-01 9.32211876e-01 -9.34969664e-01
6.54718161e-01 -3.55573028e-01 -1.27583230e+00 1.14708161e+00
3.71783040e-02 -1.95533574e-01 1.16880834e+00 5.43649673e-01
-5.45453370e-01 -1.45580303e-02 -1.32412505e+00 1.10819697e-01
1.09971774e+00 -7.02968538e-01 -7.33883440e-01 4.97387856e-01
5.11670053e-01 -1.73473090e-01 -1.10446429e+00 2.29922861e-01
1.75642073e-01 -9.68870223e-01 6.72562599e-01 -3.90443653e-01
-3.17600556e-02 -6.58724189e-01 -1.83317408e-01 -1.30389225e+00
-3.43698621e-01 -2.89951116e-01 1.86817124e-01 1.59830320e+00
5.55633605e-01 -4.46846217e-01 8.01298141e-01 1.08307564e+00
-2.24968791e-02 -1.26317769e-01 -7.94662654e-01 -8.24928045e-01
-4.39068452e-02 -3.67479742e-01 4.39474523e-01 1.53559613e+00
1.20377071e-01 1.53667390e-01 -3.53043407e-01 4.10678297e-01
7.21696496e-01 -2.04407051e-01 8.86896431e-01 -1.75244319e+00
-3.18643540e-01 4.30813394e-02 -3.77466738e-01 -8.62803817e-01
4.95436221e-01 -6.43474936e-01 2.59919055e-02 -1.68348944e+00
2.66950160e-01 -7.56446600e-01 -5.52479811e-02 6.88312769e-01
-3.02298605e-01 3.46562192e-02 -1.18299790e-01 5.62254190e-01
-7.27355182e-01 5.15154041e-02 7.86349654e-01 1.62489921e-01
-1.23540256e-02 1.16308980e-01 -9.21335936e-01 6.53977573e-01
5.30971289e-01 -7.25350678e-01 -6.58195913e-01 -2.00845078e-01
4.56131101e-01 -1.27788275e-01 3.41926128e-01 -6.34378970e-01
6.17046297e-01 -1.28486767e-01 2.48708040e-01 -4.98381704e-01
2.90307671e-01 -1.30100882e+00 5.77378035e-01 -4.78898995e-02
-6.68435872e-01 -2.23136088e-03 -3.77072841e-01 1.05481172e+00
-3.60142559e-01 -4.62184370e-01 2.36913323e-01 -9.47961137e-02
5.18950447e-02 -2.29399711e-01 -1.81335822e-01 -8.31974000e-02
1.13299310e+00 -1.83821544e-01 -3.35926831e-01 -5.52773952e-01
-1.09490275e+00 4.26132113e-01 9.13404942e-01 -1.78892627e-01
9.50773731e-02 -1.12379730e+00 -4.93058354e-01 -2.20281139e-01
9.43749771e-02 3.37708294e-01 -4.93965596e-02 9.93489504e-01
-1.83465272e-01 1.60276011e-01 2.40408540e-01 -7.81937361e-01
-1.38888669e+00 5.57726085e-01 4.98612896e-02 -3.59516621e-01
-1.01057939e-01 6.94253147e-01 -5.47991693e-02 -6.39635384e-01
3.24716032e-01 -2.41988122e-01 -4.38385271e-02 2.48495698e-01
3.73865068e-02 4.76242244e-01 6.66988045e-02 -7.26964653e-01
-1.80845946e-01 8.34222734e-02 -8.01359341e-02 -1.69278622e-01
1.11206102e+00 -3.23599070e-01 -1.90511093e-01 9.68545675e-01
4.63338524e-01 6.66374117e-02 -1.43564725e+00 -5.33258915e-01
4.37709093e-01 -7.01529801e-01 -3.38237166e-01 -7.87465870e-01
-6.11213028e-01 8.71163130e-01 5.81026562e-02 4.06687945e-01
5.83246708e-01 1.56393945e-01 -1.92933843e-01 4.55094576e-01
5.28323293e-01 -1.34787428e+00 -3.02762222e-02 5.94819710e-02
3.70253742e-01 -1.35590088e+00 3.27402234e-01 -8.83471072e-01
-7.10858464e-01 6.03567421e-01 7.79186189e-01 2.09450185e-01
7.09430099e-01 2.41239697e-01 4.09709364e-02 -2.27070332e-01
-7.78352559e-01 -2.30682984e-01 1.67396367e-01 1.01942337e+00
5.73291779e-01 1.50661573e-01 -3.07340026e-01 4.68573153e-01
-1.38225421e-01 -2.48491615e-01 6.17244422e-01 9.44146335e-01
-1.18856519e-01 -1.29128802e+00 -5.71480930e-01 7.80856192e-01
-4.80136037e-01 2.27521863e-02 -7.25049138e-01 8.98284733e-01
2.97106415e-01 1.32385290e+00 7.95707405e-02 -1.51577428e-01
1.63474083e-01 2.46413305e-01 2.39401460e-01 -9.27079499e-01
-5.90785205e-01 1.89254344e-01 4.16323006e-01 -3.08885634e-01
-1.03477693e+00 -9.19258475e-01 -1.06268430e+00 -2.33480230e-01
-6.49771869e-01 5.23369193e-01 3.62843305e-01 1.07499218e+00
1.78115427e-01 4.70549583e-01 2.20512331e-01 -4.12257046e-01
-6.13039374e-01 -1.10966420e+00 -8.45923245e-01 7.43115366e-01
6.39452785e-02 -1.08253312e+00 -6.22299612e-01 3.92490923e-01] | [9.533527374267578, 4.321042060852051] |
8d74f0b0-e66e-4060-9dfe-67f1a0fdd406 | optimistic-bounds-for-multi-output-prediction | 2002.09769 | null | https://arxiv.org/abs/2002.09769v1 | https://arxiv.org/pdf/2002.09769v1.pdf | Optimistic bounds for multi-output prediction | We investigate the challenge of multi-output learning, where the goal is to learn a vector-valued function based on a supervised data set. This includes a range of important problems in Machine Learning including multi-target regression, multi-class classification and multi-label classification. We begin our analysis by introducing the self-bounding Lipschitz condition for multi-output loss functions, which interpolates continuously between a classical Lipschitz condition and a multi-dimensional analogue of a smoothness condition. We then show that the self-bounding Lipschitz condition gives rise to optimistic bounds for multi-output learning, which are minimax optimal up to logarithmic factors. The proof exploits local Rademacher complexity combined with a powerful minoration inequality due to Srebro, Sridharan and Tewari. As an application we derive a state-of-the-art generalization bound for multi-class gradient boosting. | ['Ata Kaban', 'Henry WJ Reeve'] | 2020-02-22 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 2.10648358e-01 -2.66562682e-02 -5.17844260e-01 -7.00091362e-01
-1.45243394e+00 -8.11015368e-01 1.38340980e-01 6.24776661e-01
-5.48511326e-01 8.49714518e-01 -2.01669350e-01 -3.75134379e-01
-2.42577448e-01 -5.06524742e-01 -8.68975699e-01 -1.01167774e+00
-8.35537240e-02 2.47426882e-01 1.17053322e-01 -2.03059435e-01
1.70903862e-01 5.79924822e-01 -1.28606045e+00 3.79077882e-01
5.57412863e-01 1.41799080e+00 -5.23005962e-01 1.18476093e+00
1.55884802e-01 4.89347935e-01 -3.36768061e-01 -4.76826102e-01
3.42984796e-01 -3.17049921e-01 -1.03893101e+00 -2.54796147e-01
7.67497301e-01 7.25699065e-04 4.47404116e-01 1.17337930e+00
5.79863310e-01 1.91527545e-01 9.23690975e-01 -1.70968390e+00
-3.51863325e-01 2.15465635e-01 -8.10370266e-01 2.79867768e-01
-1.46742105e-01 -6.19353712e-01 1.17788279e+00 -7.78764486e-01
1.65314957e-01 1.21674907e+00 1.11617112e+00 5.56000888e-01
-1.23147571e+00 -5.80800831e-01 1.55916616e-01 2.47032326e-02
-9.94456351e-01 -6.48251697e-02 4.80774015e-01 -5.03045321e-01
5.36998332e-01 4.57832038e-01 -7.42864162e-02 6.97644949e-01
3.39433223e-01 7.79882252e-01 1.36632276e+00 -7.42711782e-01
1.01242669e-01 4.07480747e-01 4.53499585e-01 1.08889604e+00
4.66296673e-02 2.45196875e-02 -3.41587335e-01 -6.17057264e-01
5.05580127e-01 1.83229689e-02 -8.10499638e-02 -3.31248254e-01
-6.46030188e-01 1.48372984e+00 3.48188609e-01 2.45999351e-01
2.70008612e-02 3.15918505e-01 6.31502688e-01 5.80656111e-01
8.86468112e-01 4.07729640e-05 -7.87498653e-01 1.75640628e-01
-6.24288023e-01 2.97695249e-01 8.90972316e-01 7.09447205e-01
6.15771294e-01 -1.97345823e-01 -4.18755785e-02 9.55402374e-01
2.29042143e-01 5.55181682e-01 3.59034777e-01 -8.83512318e-01
5.27880728e-01 5.61091788e-02 1.19589113e-01 -5.18449783e-01
-6.11988664e-01 -5.00098407e-01 -7.23473728e-01 5.31847894e-01
9.74108160e-01 -1.68492943e-01 -3.97909611e-01 2.01632476e+00
5.84034145e-01 9.35836956e-02 -2.01481670e-01 6.99451327e-01
1.14859268e-01 4.80868906e-01 -4.79698181e-02 -4.31797951e-01
1.15544021e+00 -1.15437603e+00 -2.42855981e-01 4.79889624e-02
9.98462319e-01 -4.15759683e-01 1.02226901e+00 6.61140621e-01
-8.93682480e-01 -2.11097777e-01 -1.19431162e+00 -2.06499144e-01
-4.20918971e-01 -3.69274735e-01 6.84417963e-01 9.35969114e-01
-7.97838569e-01 6.43391848e-01 -5.49652278e-01 3.93014215e-02
3.30424070e-01 3.87406170e-01 -5.57540894e-01 3.32733095e-02
-9.52913642e-01 8.72815728e-01 2.06478253e-01 -1.15319639e-01
-6.72869146e-01 -8.00926566e-01 -7.30890632e-01 -1.75581470e-01
8.67266431e-02 -4.70198125e-01 1.49688792e+00 -1.18355107e+00
-1.40155590e+00 1.03292215e+00 -5.12457341e-02 -4.18525785e-01
9.34800565e-01 -1.90269053e-01 -9.86488312e-02 7.11511970e-02
-7.34670386e-02 -1.00999638e-01 8.31040442e-01 -9.68169272e-01
-1.01194394e+00 -8.88696253e-01 1.82072356e-01 1.45685315e-01
-4.37274486e-01 1.46926120e-01 2.09787458e-01 -6.35932028e-01
-2.25431189e-01 -6.97019815e-01 -3.93924028e-01 8.16649646e-02
3.41150053e-02 -3.09178233e-01 6.58563673e-01 -5.84606767e-01
9.23222780e-01 -1.88130534e+00 8.10474455e-02 2.55390793e-01
2.63516366e-01 -7.59665519e-02 -4.19620909e-02 1.27952009e-01
-1.27286747e-01 1.77951485e-01 -2.85137147e-01 -4.57108855e-01
1.06896445e-01 -1.16211630e-01 -3.91479522e-01 1.06087089e+00
-7.78285041e-02 6.71549380e-01 -9.26998496e-01 -3.61970127e-01
-1.40170440e-01 3.64159882e-01 -3.62751454e-01 -2.22951367e-01
-1.16005607e-01 3.29173118e-01 -3.38780999e-01 5.09754360e-01
7.36406565e-01 -4.40557986e-01 -1.47891626e-01 1.45594388e-01
1.27979577e-01 -4.22543883e-02 -1.30331278e+00 1.47059965e+00
-9.78315115e-01 2.69488215e-01 4.04929519e-01 -1.32316923e+00
5.08410633e-01 1.94313049e-01 4.31013554e-01 -2.07896188e-01
1.86567649e-01 5.98130584e-01 -6.84172094e-01 -8.40795599e-03
2.59604994e-02 -7.75645494e-01 -3.59221756e-01 4.13225859e-01
1.84255317e-02 1.12675905e-01 -5.16031012e-02 -1.30720019e-01
9.28069830e-01 -4.41830009e-02 4.97289181e-01 -6.17526472e-01
6.80853605e-01 -4.02345479e-01 4.40649241e-01 8.63046408e-01
-2.36874402e-01 2.86640048e-01 5.78143656e-01 -2.05071539e-01
-8.07947397e-01 -9.57366765e-01 -4.77230847e-01 1.82871234e+00
-1.02871306e-01 -8.68863538e-02 -4.86493111e-01 -1.17105782e+00
3.74058813e-01 3.55362117e-01 -9.54334855e-01 -1.19617134e-01
-4.62219507e-01 -9.68359888e-01 4.80511487e-01 6.11399829e-01
1.75480142e-01 -3.26950848e-01 -4.73810703e-01 9.78426263e-02
1.58475757e-01 -6.66512787e-01 -6.52230799e-01 6.29618108e-01
-9.46221352e-01 -1.05192482e+00 -9.95226443e-01 -8.74050856e-01
3.94550979e-01 -8.29009190e-02 1.03475237e+00 -1.27436817e-01
-1.46852568e-01 2.62919635e-01 -3.49786691e-02 -5.13995826e-01
-4.56585407e-01 1.98292598e-01 4.78719808e-02 8.20624828e-03
1.20840266e-01 -3.61761004e-01 -2.43114173e-01 3.46427441e-01
-7.50472844e-01 -9.30395573e-02 3.79343152e-01 1.09657300e+00
7.70584643e-01 -1.57354087e-01 9.41592753e-01 -1.08152616e+00
4.70098853e-01 -5.86138487e-01 -9.89988923e-01 6.85282469e-01
-7.64530182e-01 4.41338271e-01 9.77362394e-01 -6.66412532e-01
-6.82775736e-01 8.05693716e-02 -1.52004376e-01 -9.96513292e-02
2.43246302e-01 1.70237169e-01 -4.17652205e-02 -5.31873286e-01
7.63531446e-01 -2.07759947e-01 -2.84146994e-01 -5.82167447e-01
6.82264507e-01 9.71719503e-01 4.52642530e-01 -8.19346845e-01
3.87065858e-01 5.11736512e-01 4.42961425e-01 -6.04455888e-01
-1.37207592e+00 -7.81900048e-01 -5.31004012e-01 -4.73774821e-02
6.11524343e-01 -6.01183176e-01 -1.03510678e+00 4.15986449e-01
-1.00053072e+00 -5.72179437e-01 -1.50637060e-01 4.43838865e-01
-7.90329695e-01 2.83707500e-01 -9.16564167e-01 -1.15172529e+00
-4.55363274e-01 -8.22027147e-01 1.00128627e+00 4.64054570e-02
3.68517011e-01 -1.43321621e+00 2.29770243e-01 4.09760982e-01
2.55286455e-01 5.33009887e-01 1.00430524e+00 -7.71473110e-01
1.55295655e-01 -1.52001753e-01 -3.36322576e-01 6.31568372e-01
-2.26948395e-01 -6.72992527e-01 -9.83392537e-01 -5.31128526e-01
3.32565516e-01 -7.02521086e-01 1.08262050e+00 3.39992434e-01
1.20204735e+00 -5.31651437e-01 -2.06413835e-01 7.49926507e-01
1.65179050e+00 -3.80485922e-01 -1.56075060e-01 3.33524108e-01
6.52955234e-01 4.82989043e-01 5.99783421e-01 5.56637704e-01
2.41320789e-01 5.46769381e-01 3.84113401e-01 -1.62471756e-01
2.87616313e-01 2.54084449e-02 1.64779320e-01 5.94521046e-01
2.34578282e-01 1.43599257e-01 -6.62054300e-01 3.73108089e-01
-1.85654914e+00 -6.61405504e-01 -2.92042762e-01 2.59034371e+00
1.15253615e+00 1.66115388e-01 6.05918527e-01 3.42212558e-01
8.09076071e-01 -1.91398397e-01 -7.53510416e-01 -8.05517793e-01
-1.21183760e-01 4.66590196e-01 1.25033176e+00 1.03445518e+00
-1.35831189e+00 5.64964116e-01 6.67688179e+00 9.75956142e-01
-1.04068613e+00 8.21526289e-01 7.04398394e-01 -5.96930832e-02
3.72377783e-02 -2.35351905e-01 -1.14096344e+00 3.16955686e-01
1.21715164e+00 -2.81475872e-01 3.92323971e-01 1.15393591e+00
-3.51974279e-01 2.68615596e-02 -1.12379599e+00 9.69236195e-01
2.91789398e-02 -9.73833621e-01 -6.19405746e-01 1.19456939e-01
7.72470653e-01 -1.42947910e-02 1.23326689e-01 3.28766555e-01
3.27661633e-01 -1.09325397e+00 6.98177814e-01 -1.55630484e-01
9.58859682e-01 -1.09720671e+00 5.55319607e-01 3.62326562e-01
-1.26720226e+00 -3.93790901e-01 -1.39902696e-01 -1.08824223e-01
-4.62066568e-02 7.18250215e-01 -2.21238717e-01 5.86493075e-01
2.48137951e-01 3.34056407e-01 -3.05297762e-01 9.36836183e-01
-4.14426140e-02 4.34973598e-01 -6.28733218e-01 -7.33045116e-02
2.81791121e-01 -1.49555817e-01 3.24042477e-02 1.60360789e+00
-4.67559062e-02 -1.71254501e-01 2.47880980e-01 1.39899313e-01
-3.56934994e-01 4.05986309e-01 -3.27797532e-01 3.87879848e-01
8.09307843e-02 1.19517553e+00 -5.95025480e-01 -3.02772433e-01
-6.22467995e-01 1.01988018e+00 6.32607698e-01 2.24041805e-01
-8.88341486e-01 -6.07417226e-01 4.59876567e-01 -2.76457340e-01
2.97689885e-01 -7.26442738e-03 -6.38599932e-01 -9.93881464e-01
2.59690970e-01 -6.66674852e-01 9.09080863e-01 1.04585104e-01
-1.49916029e+00 1.78154573e-01 1.47420257e-01 -7.38534212e-01
-3.08976203e-01 -9.42232490e-01 -2.91789800e-01 1.02628982e+00
-1.72458422e+00 -1.31108737e+00 1.29617780e-01 5.40895522e-01
1.97398528e-01 1.62201434e-01 9.57341611e-01 5.76791048e-01
-2.36934602e-01 9.51701939e-01 5.76825559e-01 -1.37439817e-01
6.73087060e-01 -1.54189229e+00 -8.78587663e-02 6.86567068e-01
-1.98680349e-02 1.65943932e-02 7.03751147e-01 -2.01433226e-01
-1.03139102e+00 -9.60841715e-01 8.13481092e-01 -3.73484790e-01
8.40415061e-01 -4.10886973e-01 -8.02057981e-01 8.72434795e-01
-5.39937913e-01 5.36480963e-01 8.50434244e-01 2.83982277e-01
-7.34687388e-01 -2.54721910e-01 -1.36344099e+00 -5.71186543e-02
7.05420732e-01 -6.41469419e-01 -1.64457843e-01 7.80364990e-01
3.76970857e-01 -3.54673952e-01 -1.08709526e+00 2.28762716e-01
8.35103035e-01 -6.74459100e-01 9.34880257e-01 -1.15593779e+00
2.11441994e-01 2.02076897e-01 -4.48221296e-01 -1.21857035e+00
-1.07927084e-01 -6.82752013e-01 -3.71401966e-01 9.25081253e-01
5.08179903e-01 -8.71561110e-01 5.17446935e-01 4.78757769e-01
8.28789100e-02 -1.24859321e+00 -1.13106620e+00 -1.08893895e+00
8.12227786e-01 -3.64974797e-01 4.22083773e-02 9.54983175e-01
8.95758942e-02 3.67861867e-01 -5.94574869e-01 7.44836554e-02
9.19442773e-01 3.63196507e-02 3.03029180e-01 -1.35012007e+00
-8.19845736e-01 -6.80294096e-01 -3.55115443e-01 -9.98743951e-01
3.94029260e-01 -1.18612182e+00 -1.04891889e-01 -1.17353296e+00
1.81447804e-01 -7.75447488e-01 -7.92979658e-01 5.25132120e-01
-9.94441286e-02 3.31004113e-01 1.21595770e-01 -6.69444278e-02
-5.91344059e-01 2.21613690e-01 8.40022862e-01 7.01925484e-03
-9.59641039e-02 4.60671246e-01 -7.61559725e-01 6.97654903e-01
8.01531732e-01 -8.73926520e-01 -1.63286313e-01 -2.08682775e-01
3.75037789e-01 1.16016313e-01 2.43627936e-01 -4.86293137e-01
-9.40050632e-02 -2.81525314e-01 1.58838511e-01 -2.76460480e-02
4.22975749e-01 -6.06270432e-01 -3.18370283e-01 6.72793329e-01
-6.97888732e-01 5.00953458e-02 1.15804859e-02 5.95126927e-01
2.08721146e-01 -5.42870283e-01 1.41674244e+00 1.64993897e-01
-1.69856846e-01 3.23423982e-01 1.88409895e-01 5.58230162e-01
9.51544821e-01 3.21441293e-01 -3.07583541e-01 -2.94838309e-01
-5.32942176e-01 1.91758856e-01 1.34757176e-01 2.61534303e-01
2.04557121e-01 -1.45556080e+00 -9.41376746e-01 8.94530490e-02
-3.87343355e-02 -3.75815272e-01 -8.28155801e-02 1.06981003e+00
-2.22049624e-01 2.66676068e-01 2.43777316e-02 -3.75654906e-01
-1.44276226e+00 5.45967042e-01 3.62843603e-01 -5.65757930e-01
-3.34683239e-01 1.04705918e+00 6.24845289e-02 -1.87436581e-01
4.60394830e-01 -3.53391618e-01 3.80892545e-01 -1.41643390e-01
8.00034761e-01 7.17527628e-01 3.74238014e-01 -4.06455547e-01
-5.34034789e-01 6.31272554e-01 -6.74449280e-02 -4.56981272e-01
1.18367147e+00 1.17690161e-01 1.53874069e-01 9.74773228e-01
2.02099323e+00 -3.40610631e-02 -1.04803479e+00 -3.11133981e-01
2.22457364e-01 -2.77026176e-01 5.21668196e-02 -8.68281305e-01
-6.57769263e-01 9.39980924e-01 9.02005613e-01 2.55989790e-01
1.04875374e+00 1.59570631e-02 7.87811637e-01 4.17267978e-01
3.40814948e-01 -1.37274599e+00 -5.87783679e-02 3.19006950e-01
8.15047026e-01 -1.42261589e+00 -3.01287547e-02 -1.76552862e-01
-2.81471699e-01 1.16608489e+00 1.85139179e-01 -6.51500523e-02
1.00558639e+00 3.34077418e-01 -4.75962423e-02 4.50757384e-01
-5.15580237e-01 -1.56630147e-02 3.03719312e-01 2.30089232e-01
4.23048019e-01 1.53802097e-01 -4.55986202e-01 5.75676382e-01
1.12087131e-02 -1.19861491e-01 2.21752725e-03 7.56844580e-01
-5.13378918e-01 -1.18900847e+00 -3.88059348e-01 4.71683502e-01
-9.02882755e-01 1.24547079e-01 2.11649723e-02 5.96869051e-01
-1.09923951e-01 1.00389814e+00 -3.47152323e-01 -7.28477985e-02
-1.04420295e-03 2.41506830e-01 8.97860467e-01 -3.05849910e-01
-5.58425307e-01 -1.19803458e-01 -1.55645773e-01 -2.50823200e-01
-3.42588931e-01 -3.98096472e-01 -1.22930849e+00 -3.59294146e-01
-7.00854897e-01 1.64866403e-01 1.17725646e+00 1.06191874e+00
-1.10205799e-01 5.64848669e-02 8.47037911e-01 -5.65230787e-01
-1.43251169e+00 -9.68016803e-01 -6.83302224e-01 3.79705697e-01
8.50410283e-01 -6.21086419e-01 -8.09340537e-01 -1.34458482e-01] | [8.272382736206055, 4.145859718322754] |
71cd0c9d-bb7a-487f-8839-116ed49550c4 | a-one-class-classifier-for-the-detection-of | 2305.11795 | null | https://arxiv.org/abs/2305.11795v1 | https://arxiv.org/pdf/2305.11795v1.pdf | A One-Class Classifier for the Detection of GAN Manipulated Multi-Spectral Satellite Images | The highly realistic image quality achieved by current image generative models has many academic and industrial applications. To limit the use of such models to benign applications, though, it is necessary that tools to conclusively detect whether an image has been generated synthetically or not are developed. For this reason, several detectors have been developed providing excellent performance in computer vision applications, however, they can not be applied as they are to multispectral satellite images, and hence new models must be trained. In general, two-class classifiers can achieve very good detection accuracies, however they are not able to generalise to image domains and generative models architectures different than those used during training. For this reason, in this paper, we propose a one-class classifier based on Vector Quantized Variational Autoencoder 2 (VQ-VAE 2) features to overcome the limitations of two-class classifiers. First, we emphasize the generalization problem that binary classifiers suffer from by training and testing an EfficientNet-B4 architecture on multiple multispectral datasets. Then we show that, since the VQ-VAE 2 based classifier is trained only on pristine images, it is able to detect images belonging to different domains and generated by architectures that have not been used during training. Last, we compare the two classifiers head-to-head on the same generated datasets, highlighting the superiori generalization capabilities of the VQ-VAE 2-based detector. | ['Mauro Barni', 'Giovanna Maria Dimitri', 'Lydia Abady'] | 2023-05-19 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 3.98029149e-01 -8.26670378e-02 1.64638400e-01 -1.58833325e-01
-5.17076373e-01 -4.91511017e-01 8.82305562e-01 -2.37756193e-01
-2.97949076e-01 8.69187176e-01 -5.97000599e-01 -2.16104895e-01
-1.55343011e-01 -1.13480675e+00 -6.26031697e-01 -1.04426908e+00
1.11125901e-01 4.78519648e-01 4.86916304e-01 -3.54649484e-01
-1.03107020e-01 6.97080731e-01 -2.24833941e+00 3.57982725e-01
9.14940953e-01 8.94996643e-01 4.48698997e-01 6.01318598e-01
-1.85726061e-02 5.28622627e-01 -8.37941170e-01 -1.70089021e-01
2.69156486e-01 -7.82383323e-01 -4.86299962e-01 3.18468392e-01
4.32770312e-01 -2.33079121e-01 -1.17511516e-02 1.25449014e+00
2.65539765e-01 -1.80281147e-01 1.13222265e+00 -1.35566616e+00
-5.88988721e-01 3.85366142e-01 -1.55366093e-01 9.98433232e-02
-7.08209584e-03 1.69904992e-01 7.20345676e-01 -7.79407680e-01
6.71149552e-01 9.57603276e-01 5.68947673e-01 6.13281369e-01
-1.54743862e+00 -4.06243503e-01 -3.12785417e-01 1.48250401e-01
-1.35449040e+00 -5.15864640e-02 7.90360689e-01 -5.81706464e-01
8.16751420e-01 4.27005678e-01 7.70421684e-01 1.43199313e+00
1.39665410e-01 6.70734048e-01 1.65304935e+00 -3.92727077e-01
4.29924995e-01 6.59338653e-01 -2.27682710e-01 4.03084964e-01
2.97757983e-01 3.13860953e-01 6.99213892e-02 1.62737846e-01
6.64071560e-01 -3.21474046e-01 -4.58829641e-01 -4.29148167e-01
-7.93330610e-01 1.08769107e+00 7.01708496e-01 9.42845285e-01
-4.47315961e-01 -1.25806898e-01 1.60637841e-01 2.71661699e-01
2.52180517e-01 4.09016371e-01 -8.35034922e-02 4.34760481e-01
-1.31806254e+00 2.00213403e-01 4.82546985e-01 4.93349314e-01
8.88310790e-01 4.83068794e-01 7.44809434e-02 7.34622538e-01
8.60513225e-02 7.63918281e-01 8.14303160e-01 -4.84623432e-01
-2.12390080e-01 6.94458425e-01 -1.73258573e-01 -9.07733798e-01
-3.25774223e-01 -7.71236718e-01 -1.01036191e+00 6.60535812e-01
1.41696274e-01 1.80668831e-01 -1.00300992e+00 1.40389895e+00
-6.42047438e-04 -1.87007681e-01 5.76835692e-01 8.17341566e-01
8.83673966e-01 7.79458046e-01 -3.89938541e-02 -1.13397792e-01
9.74542797e-01 -4.06606764e-01 -4.21305686e-01 -3.42137218e-01
3.47017437e-01 -5.43222547e-01 8.26511502e-01 5.15402019e-01
-5.28452098e-01 -8.42812836e-01 -1.25825596e+00 6.19903684e-01
-6.89088285e-01 4.44293052e-01 4.53760207e-01 9.33626831e-01
-1.04522514e+00 5.97461343e-01 -5.34784734e-01 -5.72668791e-01
3.09660196e-01 1.06379941e-01 -2.89704829e-01 2.23455295e-01
-1.20176816e+00 1.00776005e+00 8.08483541e-01 1.46727249e-01
-1.11822355e+00 -6.16418105e-03 -5.57623446e-01 1.63661744e-02
-1.75864965e-01 -2.85991222e-01 7.98532903e-01 -1.34786081e+00
-1.29444158e+00 9.27661240e-01 2.38448501e-01 -5.96691906e-01
6.81882441e-01 4.65696543e-01 -5.98516643e-01 2.00914025e-01
2.03572854e-01 6.52974367e-01 9.87574756e-01 -1.69104242e+00
-6.76691473e-01 -1.63559034e-01 -1.90902092e-02 -2.43590713e-01
-3.94816875e-01 -3.58316183e-01 2.68736839e-01 -4.49105561e-01
2.25852773e-01 -1.00548565e+00 1.09117189e-02 -1.72616020e-01
-3.11789483e-01 -1.06437681e-02 1.15611088e+00 -3.83934647e-01
5.90564311e-01 -2.11142898e+00 -1.87153555e-02 2.46484235e-01
-1.86533451e-01 7.86883116e-01 -2.06818476e-01 2.65787989e-01
-1.29265472e-01 1.28055528e-01 -6.62941933e-01 1.55101091e-01
-2.06255481e-01 5.30435681e-01 -4.59397376e-01 3.76495361e-01
4.46269214e-01 7.07067311e-01 -7.51490295e-01 -4.15536404e-01
5.35819530e-01 5.58970571e-01 -1.95279330e-01 8.47040787e-02
-2.68917590e-01 3.95967722e-01 -3.31629932e-01 5.32067716e-01
8.61993670e-01 -1.66149125e-01 2.23366365e-01 -1.70851514e-01
-6.40545264e-02 -4.42420840e-01 -1.15816629e+00 1.08485591e+00
-4.80133921e-01 7.74832845e-01 -8.89949575e-02 -1.35370898e+00
1.15208960e+00 3.05780053e-01 2.52163738e-01 -6.56692624e-01
2.54242003e-01 6.12265050e-01 2.38474943e-02 -4.13248509e-01
3.19382042e-01 -3.82573873e-01 1.90597430e-01 6.23062551e-02
4.55662996e-01 -4.31032360e-01 5.15737355e-01 -1.54570058e-01
5.04378796e-01 2.20444605e-01 2.30450675e-01 -2.44537592e-01
7.78586388e-01 2.54406631e-01 2.81980872e-01 8.01772296e-01
2.48637889e-02 5.60980201e-01 3.27823073e-01 -3.19146931e-01
-1.10877061e+00 -9.61279511e-01 -6.46271944e-01 6.59066677e-01
2.21585333e-02 2.04033673e-01 -8.67568195e-01 -5.63181102e-01
-2.06624463e-01 8.95139039e-01 -5.59421778e-01 -4.57554683e-02
-1.72560945e-01 -1.35492945e+00 7.31631637e-01 2.02944592e-01
7.82015204e-01 -1.02755368e+00 -1.01765203e+00 3.59877765e-01
-1.15286656e-01 -1.07729197e+00 6.67674899e-01 2.96671659e-01
-8.94512951e-01 -1.14320362e+00 -7.96765685e-01 -6.07846260e-01
3.27005714e-01 2.31116921e-01 9.78499591e-01 1.96566861e-02
-2.08862111e-01 2.77317196e-01 -6.22666538e-01 -3.46894056e-01
-1.11079812e+00 2.31063366e-02 -6.53583035e-02 2.62364119e-01
1.57438934e-01 -6.07345700e-01 -2.69552350e-01 3.96930695e-01
-1.28264523e+00 -2.38501161e-01 8.53157878e-01 1.08188820e+00
4.23004478e-01 2.82934725e-01 5.49813509e-01 -5.80869198e-01
3.29517126e-01 -2.74843276e-01 -6.86065376e-01 3.43972534e-01
-6.68220937e-01 1.00955263e-01 7.15125263e-01 -3.65436226e-01
-9.87022638e-01 3.74119252e-01 -4.99656588e-01 -4.03212100e-01
-4.81639028e-01 4.35467035e-01 -9.65582505e-02 -1.98847920e-01
1.20515954e+00 7.33358622e-01 1.79665547e-03 -1.55158624e-01
1.52508751e-01 9.28769410e-01 4.20278400e-01 -8.62827376e-02
9.35197771e-01 4.82533872e-01 1.09172471e-01 -1.24735403e+00
-5.16693354e-01 -1.43806592e-01 -5.04457176e-01 -4.59462911e-01
9.95428801e-01 -8.54618192e-01 -9.87747759e-02 6.87044084e-01
-1.06385815e+00 -9.36386883e-02 -4.18095887e-01 3.35152060e-01
-5.39393961e-01 3.37043494e-01 -3.21223646e-01 -1.17304754e+00
-8.98784474e-02 -1.34029567e+00 9.02102828e-01 2.75878578e-01
3.14173520e-01 -8.39082301e-01 -7.07988488e-03 4.00310867e-02
4.82760340e-01 5.08395195e-01 7.96956003e-01 -6.10350490e-01
-4.85327154e-01 -3.68405193e-01 -5.09181656e-02 8.20458472e-01
1.23042436e-02 3.08067441e-01 -1.35474062e+00 -2.87008196e-01
1.27205789e-01 -4.65642601e-01 1.10951161e+00 2.94579387e-01
7.93233216e-01 -1.13369860e-02 -2.43963733e-01 4.81319934e-01
1.66595232e+00 2.46669516e-01 9.12672400e-01 3.84437680e-01
2.26787522e-01 5.58854580e-01 3.24544281e-01 -1.08484581e-01
-1.56188130e-01 6.64160132e-01 9.07277942e-01 -2.27229655e-01
-2.15398565e-01 -9.27257836e-02 4.45640504e-01 3.32556635e-01
-3.39320838e-01 -4.33442086e-01 -6.46530330e-01 5.02077997e-01
-1.44954360e+00 -1.23070669e+00 -7.54928112e-01 2.11949182e+00
4.22042102e-01 1.91390857e-01 2.20469665e-02 6.07584894e-01
7.70870864e-01 1.80353075e-01 -3.19077373e-01 -1.80284277e-01
-7.35846460e-01 2.60091692e-01 4.12487805e-01 1.77250817e-01
-1.32562912e+00 7.24517405e-01 6.03061867e+00 9.52244699e-01
-1.61864936e+00 2.03347087e-01 4.49400783e-01 5.20482659e-01
-2.14056611e-01 -1.15223214e-01 -5.77737153e-01 3.75773847e-01
9.36351240e-01 3.07569116e-01 6.47007152e-02 1.09053850e+00
-2.16796249e-01 -1.00384824e-01 -7.02731311e-01 9.55807865e-01
1.90069720e-01 -1.16305387e+00 1.26728550e-01 1.97348744e-01
7.03350425e-01 2.01355256e-02 1.35262311e-01 3.41982126e-01
1.66842481e-03 -9.33862925e-01 8.04309726e-01 2.93328434e-01
7.34297454e-01 -4.26641256e-01 9.02017355e-01 5.38419366e-01
-9.32095706e-01 -2.06710547e-01 -6.74627960e-01 2.24338636e-01
-4.71459106e-02 6.27151728e-01 -8.98391008e-01 7.38901794e-01
5.85116208e-01 3.16569388e-01 -8.44743311e-01 9.65762675e-01
-4.16530579e-01 5.65966487e-01 -1.23326436e-01 -1.24248706e-01
3.36250633e-01 -1.73997983e-01 6.17058277e-01 1.00915039e+00
7.26590335e-01 -3.06917369e-01 -1.80220250e-02 1.17386997e+00
5.18454909e-01 8.57170671e-02 -8.16785574e-01 -6.88567236e-02
-2.34201439e-02 1.38748932e+00 -9.26612496e-01 -3.15681309e-01
-2.23084599e-01 1.07541120e+00 -1.55160859e-01 2.39738271e-01
-8.07062447e-01 -1.74188688e-01 2.59668887e-01 1.45108029e-01
3.77984285e-01 2.59115044e-02 9.32852626e-02 -1.23437905e+00
-2.24273324e-01 -8.16403806e-01 1.41943157e-01 -9.45745945e-01
-1.23054290e+00 1.13829434e+00 1.19441748e-02 -1.39664328e+00
-7.31726170e-01 -1.03502917e+00 -3.45737517e-01 8.03530633e-01
-1.41908979e+00 -1.32880557e+00 -5.26563048e-01 4.15311217e-01
3.16178918e-01 -3.11310351e-01 1.05357194e+00 1.34965032e-01
-3.32535118e-01 1.23965427e-01 3.25822949e-01 2.04217911e-01
2.43179455e-01 -1.22650993e+00 -2.36717779e-02 1.03813171e+00
5.62485099e-01 2.44146306e-02 8.82806599e-01 -3.13615978e-01
-8.14789116e-01 -1.16210330e+00 5.02240121e-01 -2.39634573e-01
2.89092153e-01 -1.66909665e-01 -1.08307385e+00 4.02311593e-01
1.57038808e-01 -6.80839818e-04 3.66350472e-01 -5.33981383e-01
-2.49216154e-01 -1.20569542e-01 -1.29038346e+00 1.09442912e-01
5.56704879e-01 -4.19921875e-01 -6.74607635e-01 4.47332025e-01
-5.84101230e-02 -6.78054094e-02 -7.67734528e-01 4.39474285e-01
3.09873998e-01 -1.55165398e+00 9.03255343e-01 -1.84412047e-01
4.52150673e-01 -3.70131761e-01 -2.70602912e-01 -1.44761395e+00
-1.39240518e-01 2.08663866e-01 3.16691667e-01 1.07070875e+00
4.73653913e-01 -8.44605684e-01 5.69060862e-01 -1.68434307e-01
-1.22869387e-02 -3.12124312e-01 -1.04692268e+00 -1.13592029e+00
1.93773836e-01 -2.11363360e-01 5.76653779e-01 9.37749505e-01
-7.18480766e-01 2.86284775e-01 -2.91016459e-01 1.27313212e-01
6.39286757e-01 3.72249246e-01 6.60337448e-01 -1.75051820e+00
-3.12461227e-01 -5.96807778e-01 -7.72371829e-01 -5.08700192e-01
1.21340074e-01 -9.42208707e-01 7.94086531e-02 -1.38266039e+00
2.22372338e-02 -3.98932785e-01 -4.25482616e-02 3.44067484e-01
1.58446833e-01 7.67306805e-01 2.06917867e-01 4.17981476e-01
2.52495348e-01 5.81446946e-01 9.60311115e-01 -4.61747557e-01
-5.81617700e-03 -1.74487438e-02 -1.60622820e-01 6.24655426e-01
8.04955840e-01 -4.24086839e-01 -3.25442821e-01 6.23807423e-02
9.55797061e-02 -1.54491737e-01 8.63642931e-01 -1.54501855e+00
-1.62718490e-01 -5.45805786e-04 6.16059124e-01 -3.55146229e-01
4.22134876e-01 -7.86112130e-01 5.27258396e-01 7.91054904e-01
1.64587110e-01 -4.04094577e-01 -3.37005630e-02 3.73035789e-01
-4.80590403e-01 -7.62759864e-01 1.07582450e+00 -4.43698972e-01
-9.91270483e-01 -5.27794138e-02 -7.30049968e-01 -4.60389793e-01
1.21333158e+00 -4.88422275e-01 -1.55146092e-01 -2.93260485e-01
-7.78675795e-01 -4.16290611e-01 5.84444046e-01 2.16827154e-01
5.36668658e-01 -1.09092081e+00 -8.54935706e-01 3.96041483e-01
3.36156368e-01 -2.58506238e-01 3.72734457e-01 5.75085580e-01
-7.01465786e-01 3.92432392e-01 -5.87394416e-01 -1.02713919e+00
-1.09257317e+00 6.85054123e-01 7.68458307e-01 -9.21674371e-02
-3.77355605e-01 5.36496639e-01 8.93779695e-02 -4.22977537e-01
-4.12263155e-01 -7.87365958e-02 -4.19776410e-01 3.92224282e-01
3.04637611e-01 1.35754481e-01 3.64719748e-01 -9.80627239e-01
-2.54446745e-01 4.05015230e-01 4.58332330e-01 -1.27181590e-01
1.29972970e+00 2.93298960e-01 -1.77356172e-02 3.91601026e-01
8.79755378e-01 -3.07383478e-01 -1.00909770e+00 1.57939181e-01
-2.39232093e-01 -3.16474736e-01 1.19896874e-01 -6.84732020e-01
-1.16069806e+00 1.01998508e+00 1.11040127e+00 8.66641283e-01
1.26982534e+00 -1.33479208e-01 1.48779646e-01 2.82076597e-01
5.06099224e-01 -1.02109003e+00 -1.21114805e-01 1.61869705e-01
9.88393903e-01 -1.20448577e+00 -2.21062616e-01 -3.07446331e-01
-5.48313916e-01 1.25270760e+00 3.56130242e-01 -1.66886404e-01
3.10374171e-01 2.13161707e-02 2.17291549e-01 -3.78551781e-02
-2.07785323e-01 -6.86434984e-01 1.65654823e-01 8.92172813e-01
2.44407848e-01 2.09608600e-01 -4.00001347e-01 5.21950573e-02
-1.98858187e-01 -5.07363155e-02 8.20109069e-01 6.34192407e-01
-5.43398917e-01 -1.20986235e+00 -7.14146972e-01 4.70750570e-01
-1.29843414e-01 1.88602358e-01 -4.74312931e-01 1.13247001e+00
4.23479259e-01 9.49298322e-01 2.53920704e-02 -6.26107395e-01
1.94149584e-01 2.18799785e-01 5.17915308e-01 -2.35506132e-01
-4.96470869e-01 -1.10271424e-01 -1.57150418e-01 -5.72676808e-02
-8.94452393e-01 -6.19531453e-01 -5.52160621e-01 8.81401449e-02
-4.26332027e-01 1.03262760e-01 8.71390462e-01 7.84364104e-01
-6.26130253e-02 5.58821321e-01 6.10770285e-01 -9.77518737e-01
-7.77930319e-01 -1.12352741e+00 -8.38469386e-01 3.51747990e-01
8.64903480e-02 -8.31681192e-01 -5.89921296e-01 -7.55530819e-02] | [10.105962753295898, -1.5087966918945312] |
a5979ddf-da90-4b6c-953c-94d2e904dc3e | generalized-gloves-of-neural-additive-models | 2209.10082 | null | https://arxiv.org/abs/2209.10082v1 | https://arxiv.org/pdf/2209.10082v1.pdf | Generalized Gloves of Neural Additive Models: Pursuing transparent and accurate machine learning models in finance | For many years, machine learning methods have been used in a wide range of fields, including computer vision and natural language processing. While machine learning methods have significantly improved model performance over traditional methods, their black-box structure makes it difficult for researchers to interpret results. For highly regulated financial industries, transparency, explainability, and fairness are equally, if not more, important than accuracy. Without meeting regulated requirements, even highly accurate machine learning methods are unlikely to be accepted. We address this issue by introducing a novel class of transparent and interpretable machine learning algorithms known as generalized gloves of neural additive models. The generalized gloves of neural additive models separate features into three categories: linear features, individual nonlinear features, and interacted nonlinear features. Additionally, interactions in the last category are only local. The linear and nonlinear components are distinguished by a stepwise selection algorithm, and interacted groups are carefully verified by applying additive separation criteria. Empirical results demonstrate that generalized gloves of neural additive models provide optimal accuracy with the simplest architecture, allowing for a highly accurate, transparent, and explainable approach to machine learning. | ['Weicheng Ye', 'Dangxing Chen'] | 2022-09-21 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 8.36529136e-02 2.92275012e-01 -2.45999545e-01 -7.79408336e-01
-2.48949438e-01 -5.75317085e-01 5.10549247e-01 -3.62442993e-02
-1.57429636e-01 7.28769660e-01 -1.62056103e-01 -5.68766534e-01
-4.16539520e-01 -5.51100373e-01 -5.15004873e-01 -4.41185743e-01
-5.93442330e-03 1.76318526e-01 -2.93199211e-01 1.17767394e-01
1.90335318e-01 5.68028271e-01 -1.56874347e+00 3.22950184e-01
1.36725914e+00 1.04949570e+00 -2.11688980e-01 2.26957202e-01
-3.13927799e-01 8.92844558e-01 -4.24089760e-01 -6.92131579e-01
3.51029754e-01 -4.08986956e-01 -3.93590212e-01 6.37914240e-02
3.13420832e-01 -3.99374098e-01 1.38472646e-01 8.95950794e-01
-1.25374317e-01 -5.42074703e-02 9.96015310e-01 -1.40981221e+00
-1.31595111e+00 6.52270317e-01 -4.63022590e-01 -2.34327286e-01
3.04018632e-02 6.82030693e-02 1.27700138e+00 -1.12301326e+00
8.65230262e-02 1.40111887e+00 5.87273419e-01 5.64954817e-01
-1.27999854e+00 -7.04076648e-01 4.56190497e-01 -4.90340404e-02
-1.15128398e+00 -3.13767761e-01 6.69973075e-01 -6.46731079e-01
8.04266810e-01 7.31804252e-01 4.88450676e-01 4.33073431e-01
4.56605703e-01 6.59476638e-01 1.26166165e+00 -5.81171095e-01
2.45623440e-01 8.76230061e-01 4.29929852e-01 9.07150030e-01
6.53165936e-01 1.29745603e-01 -5.16102850e-01 -1.10870220e-01
7.10115850e-01 2.14499578e-01 -2.52954453e-01 -4.36068267e-01
-8.89055490e-01 1.07643318e+00 3.67561340e-01 2.53302127e-01
-3.04886729e-01 3.60363256e-03 -2.69014657e-01 3.96617562e-01
7.56719232e-01 8.62885356e-01 -6.33105397e-01 3.04294884e-01
-7.41499364e-01 5.47019020e-02 6.34730995e-01 7.19192088e-01
8.78231287e-01 3.38206887e-01 2.43179962e-01 7.84812629e-01
5.58479011e-01 4.34828699e-01 2.65264690e-01 -9.91367579e-01
2.37922981e-01 1.02915633e+00 8.41844007e-02 -1.34300590e+00
-4.82676953e-01 -7.41072655e-01 -9.85329747e-01 6.71620965e-01
5.37192106e-01 5.02594523e-02 -7.58960307e-01 1.52198207e+00
2.06476808e-01 -5.91124117e-01 4.11722884e-02 8.75322044e-01
8.47185373e-01 6.20912731e-01 -2.38424465e-02 -2.63868511e-01
1.26873755e+00 -8.64923179e-01 -7.96651304e-01 -3.21674913e-01
4.88419056e-01 -6.08075440e-01 1.33011186e+00 5.63513815e-01
-1.18067169e+00 -3.78857881e-01 -1.02834904e+00 -2.53941447e-01
-3.88322234e-01 2.78677940e-01 1.08908308e+00 8.17877650e-01
-7.72915900e-01 6.82674885e-01 -5.59612870e-01 1.02665454e-01
3.60178202e-01 4.83633101e-01 -5.52687347e-01 2.66111434e-01
-9.21912491e-01 9.78556454e-01 1.75572872e-01 3.87182891e-01
-6.06059358e-02 -6.10707998e-01 -8.36756289e-01 2.20093593e-01
1.75581917e-01 -6.84788525e-01 1.12488961e+00 -1.48000300e+00
-1.30678248e+00 7.70222008e-01 -3.93938869e-01 -4.71180916e-01
5.81458390e-01 -4.34653699e-01 -3.24449360e-01 -1.79600731e-01
-2.84555018e-01 3.95546615e-01 1.04990137e+00 -1.59340119e+00
-5.84153116e-01 -4.08514857e-01 -3.77043411e-02 2.19857872e-01
-7.20185399e-01 1.54931292e-01 8.11881647e-02 -5.39759815e-01
3.96565825e-01 -6.66580141e-01 -1.71216846e-01 4.27121878e-01
-2.13410899e-01 -8.33211467e-02 6.79989815e-01 -6.59117758e-01
1.40338826e+00 -2.14916158e+00 -2.21801415e-01 2.02631876e-01
8.25264454e-01 -4.00660969e-02 1.39648870e-01 -1.83234453e-01
-2.17341870e-01 6.31400883e-01 -2.09371448e-01 -2.66571105e-01
2.06530884e-01 8.38518888e-02 -3.62425774e-01 1.86998725e-01
4.82551008e-01 8.07883918e-01 -5.29343367e-01 -3.14754248e-01
2.45218784e-01 3.67727071e-01 -3.92233402e-01 8.19191784e-02
1.58138439e-01 1.96485054e-02 -4.77804780e-01 6.13947630e-01
5.23777008e-01 -4.11735296e-01 -1.12550221e-01 -1.38541102e-01
-1.08133435e-01 1.21238448e-01 -1.25157714e+00 5.30208409e-01
-2.92467505e-01 6.55545473e-01 -3.65589038e-02 -9.06760693e-01
1.12466216e+00 1.38737589e-01 2.33492807e-01 -6.00357831e-01
-2.86926590e-02 3.96264285e-01 2.22793549e-01 -5.01941741e-01
5.87950408e-01 -2.56779879e-01 1.78906485e-01 5.92585027e-01
-2.73596108e-01 -2.34433696e-01 -7.81243518e-02 6.79099411e-02
5.18082380e-01 -1.79332048e-01 7.22395062e-01 -2.91541904e-01
1.78735152e-01 -4.65246662e-02 7.36106575e-01 7.39908457e-01
-1.71097413e-01 6.19811952e-01 3.45814556e-01 -8.06192696e-01
-8.86429727e-01 -1.05541492e+00 -3.23420800e-02 1.00835347e+00
2.48758867e-02 -6.78701997e-02 -6.86962485e-01 -5.61190784e-01
2.31464550e-01 8.60701025e-01 -6.93394601e-01 -2.14763150e-01
-1.98467031e-01 -6.79174960e-01 1.08189151e-01 5.31793177e-01
2.43922785e-01 -6.53254807e-01 -5.74531078e-01 -5.20810746e-02
1.62902728e-01 -8.22459579e-01 -2.44078457e-01 2.28617042e-01
-1.10696423e+00 -9.49467182e-01 -3.34015787e-01 -3.34751457e-01
1.00214612e+00 5.16714394e-01 1.10286379e+00 4.33507204e-01
-1.82148144e-01 8.16089660e-02 -2.11021006e-01 -8.56154859e-01
-3.92723978e-01 -2.90003419e-01 2.69968957e-01 3.08679402e-01
3.07194620e-01 -1.38272285e-01 -2.99247712e-01 3.47905308e-01
-6.87564671e-01 3.42314929e-01 8.57310832e-01 8.48621547e-01
4.62600708e-01 -4.94999439e-03 4.77833152e-01 -8.97748411e-01
6.29168630e-01 -2.43852168e-01 -4.44169909e-01 3.24569851e-01
-1.12221479e+00 2.14533493e-01 7.80728817e-01 -3.98064941e-01
-1.11136794e+00 8.83967951e-02 3.99874181e-01 -2.02392548e-01
-4.39066216e-02 6.82392955e-01 -3.09671909e-01 -1.60335243e-01
9.56268311e-01 -6.99362457e-02 9.17445868e-02 -3.67464840e-01
4.67427850e-01 6.91313088e-01 2.49949902e-01 -3.11231494e-01
1.10505724e+00 4.00863498e-01 -1.80050179e-01 -8.03598464e-01
-7.16924369e-01 -1.30522810e-02 -6.31803453e-01 -3.07236761e-01
5.68428516e-01 -5.62852442e-01 -3.23480844e-01 2.72170037e-01
-1.10141635e+00 4.15287241e-02 -3.73565674e-01 7.74389148e-01
-2.56361485e-01 2.53414214e-01 -3.03694189e-01 -1.25594938e+00
-2.69529045e-01 -9.62853611e-01 6.77208304e-01 4.12983954e-01
-6.40240610e-01 -1.16918027e+00 -5.95308781e-01 4.72476780e-01
3.75681639e-01 2.33569518e-01 1.28016198e+00 -8.21105421e-01
-5.32125056e-01 -4.28354710e-01 -2.74932235e-01 4.77000624e-01
4.62522775e-01 5.99523008e-01 -1.04730797e+00 7.13859871e-02
1.29473850e-01 -1.30222514e-01 8.72286022e-01 5.00721216e-01
1.05491614e+00 -7.87416279e-01 -1.60915688e-01 4.94302213e-01
9.90894318e-01 4.10420805e-01 3.79676610e-01 2.13874504e-01
6.03873312e-01 8.69149208e-01 3.48531097e-01 1.99256971e-01
3.20754260e-01 3.39768648e-01 4.45741862e-01 -5.59374094e-01
4.13436890e-01 2.77078710e-02 2.77273297e-01 6.61578715e-01
-3.43618095e-01 -2.55384948e-03 -8.01230907e-01 2.03611001e-01
-1.80670059e+00 -1.06492341e+00 -4.97727841e-01 2.39182806e+00
6.55188203e-01 2.60791451e-01 -1.13443680e-01 3.59420389e-01
5.60593963e-01 -2.07032084e-01 -5.38065851e-01 -7.20340133e-01
-3.01707536e-01 -2.18294725e-01 1.48751453e-01 4.76271003e-01
-1.08930397e+00 5.69061399e-01 6.91811752e+00 6.18163347e-01
-9.39037561e-01 -2.07223609e-01 1.04737151e+00 -1.26448736e-01
-8.09578538e-01 1.01698674e-02 -4.80816841e-01 2.63301671e-01
5.09965479e-01 -3.98669511e-01 3.48106712e-01 1.07557213e+00
5.24118066e-01 -1.29835373e-02 -1.22259378e+00 8.84106636e-01
-1.44910008e-01 -1.28462684e+00 2.05762848e-01 1.35615826e-01
7.20689476e-01 -5.73456228e-01 4.66649979e-01 8.70740861e-02
3.14130425e-01 -1.45231783e+00 1.01470768e+00 7.27040112e-01
4.95084316e-01 -5.61621130e-01 5.26265383e-01 3.02871525e-01
-1.03139472e+00 -3.52227479e-01 -3.75740290e-01 -5.32568574e-01
-2.61651963e-01 9.39534664e-01 -3.38114589e-01 4.77629095e-01
6.07712626e-01 4.38522011e-01 -4.67300981e-01 9.75331604e-01
-7.37133399e-02 6.73217773e-01 -1.76109686e-01 -2.05469042e-01
4.76042293e-02 -2.26847395e-01 2.66875446e-01 9.05276239e-01
3.47611755e-01 1.80264995e-01 -2.54256338e-01 1.32502306e+00
4.75523397e-02 3.26116592e-01 -6.41980469e-01 -2.30932504e-01
3.21812242e-01 1.29576695e+00 -7.39718616e-01 -2.71816611e-01
-4.68694329e-01 5.65566778e-01 2.21466273e-01 4.09741908e-01
-7.08540201e-01 -2.24236339e-01 5.53439856e-01 3.01803406e-02
-4.29202706e-01 -1.95115879e-01 -1.13623106e+00 -1.10963833e+00
1.91133752e-01 -1.00905025e+00 2.58890033e-01 -6.60503507e-01
-1.38172925e+00 5.16463459e-01 -1.77697852e-01 -1.34401917e+00
-3.12553674e-01 -8.32385004e-01 -4.79698569e-01 9.05629337e-01
-1.22543883e+00 -1.09558737e+00 -3.18621218e-01 2.61362672e-01
4.61898953e-01 -2.94729710e-01 8.78912807e-01 -1.02633901e-01
-5.77325821e-01 6.32187605e-01 4.11276102e-01 -8.22780505e-02
5.62548101e-01 -1.29559660e+00 1.64897725e-01 5.92777252e-01
3.00275624e-01 9.71802294e-01 7.58269072e-01 -3.59053165e-01
-1.21355474e+00 -9.86325562e-01 1.08683789e+00 -3.87410551e-01
4.58603531e-01 -4.06932175e-01 -1.15827668e+00 5.54977834e-01
-3.33457328e-02 -2.14241579e-01 8.50298166e-01 2.02948764e-01
-6.48626387e-01 -3.96475017e-01 -1.05094838e+00 6.46392107e-01
5.85102677e-01 -3.77736628e-01 -5.12039244e-01 3.07563454e-01
5.78142405e-01 -9.40919816e-02 -6.23062968e-01 3.32250834e-01
8.26160073e-01 -1.13242674e+00 6.32641375e-01 -7.35702217e-01
5.97856224e-01 -2.30793089e-01 5.22064343e-02 -1.15538967e+00
-5.17228067e-01 -6.02940619e-01 -2.44529724e-01 1.04325628e+00
1.03353834e+00 -1.02242148e+00 6.81210041e-01 1.38365245e+00
-4.48737480e-02 -8.79662216e-01 -5.30386746e-01 -9.90692377e-01
2.45368555e-01 -4.79652435e-01 6.61762178e-01 1.18060482e+00
7.18639567e-02 9.41127986e-02 -5.15847683e-01 -1.18104249e-01
6.67465568e-01 1.06879488e-01 6.42764568e-01 -1.63237226e+00
-4.39526290e-01 -9.07458186e-01 -2.63545603e-01 -7.25328982e-01
-9.40447077e-02 -6.15426481e-01 -1.38546601e-01 -1.64104486e+00
4.49615806e-01 -4.42617536e-01 -2.79424042e-01 6.80699289e-01
-3.12075645e-01 2.04574451e-01 2.16651157e-01 4.83544886e-01
-6.54868484e-02 2.19540983e-01 9.79799926e-01 -1.23660557e-01
-3.34816217e-01 8.86179805e-02 -1.19917500e+00 1.05289614e+00
8.51171792e-01 -2.29573950e-01 -4.78325784e-01 -6.00617409e-01
3.29221427e-01 -3.88450623e-01 4.79324192e-01 -5.84516644e-01
-4.38978560e-02 -5.93222439e-01 5.52233458e-01 -4.48840559e-02
2.51432747e-01 -9.60054040e-01 2.99064219e-01 4.34352010e-01
-5.38049936e-01 -2.46449243e-02 3.23311351e-02 1.28500938e-01
-1.63954288e-01 -1.54524848e-01 7.77900815e-01 1.85410362e-02
-3.57668161e-01 6.56536818e-02 -3.77780437e-01 -3.95146579e-01
1.10941553e+00 -7.44368315e-01 -3.52427572e-01 -7.11032093e-01
-5.28727055e-01 -1.77462716e-02 2.38483340e-01 4.44330364e-01
7.72296906e-01 -1.08119881e+00 -6.73951030e-01 2.25907937e-01
8.96626525e-03 -2.16784135e-01 1.06299601e-01 6.60013974e-01
-3.52632135e-01 4.22779232e-01 -1.34193366e-02 -2.38226011e-01
-1.37582624e+00 4.42361325e-01 3.61508012e-01 -1.59153417e-01
-1.69149324e-01 7.70732403e-01 6.39960587e-01 -3.48447204e-01
1.85193554e-01 -5.25417149e-01 -1.97927460e-01 -9.71657038e-02
7.37000883e-01 3.79490763e-01 -1.17897682e-01 -6.08209074e-01
-1.43524855e-01 5.20964921e-01 -2.01514348e-01 1.40542269e-01
1.26719463e+00 -4.84528318e-02 -1.75871402e-01 6.79754674e-01
7.87672639e-01 -6.00884259e-02 -1.09850967e+00 -2.28037219e-02
-3.88823599e-02 -4.56300527e-01 -5.85955530e-02 -9.30989385e-01
-8.82461250e-01 1.15188873e+00 3.45858753e-01 7.59326696e-01
1.21404719e+00 -2.55434781e-01 8.08887184e-02 4.73012418e-01
1.63082153e-01 -1.09657145e+00 -7.44395554e-02 2.87520677e-01
9.17091250e-01 -1.38953066e+00 1.88642696e-01 -5.68258464e-01
-7.08395183e-01 1.28007472e+00 6.01039529e-01 2.44232491e-01
3.55732113e-01 3.45114827e-01 2.35805422e-01 9.91524830e-02
-6.80132329e-01 2.21349597e-01 6.95096433e-01 5.29661417e-01
6.35380268e-01 9.47093815e-02 -4.31795329e-01 1.04165924e+00
-3.61857176e-01 -2.24400580e-01 4.51214105e-01 5.09805202e-01
-7.50010669e-01 -6.29587531e-01 -6.40771985e-01 7.52896011e-01
-4.61136281e-01 -2.03712359e-01 -8.24311972e-01 1.04784024e+00
-4.79846122e-03 1.18744171e+00 9.73770246e-02 -5.26809216e-01
1.71621904e-01 -2.61538364e-02 7.53302202e-02 -3.83172393e-01
-5.57120323e-01 -9.43612903e-02 -2.26021255e-03 -3.34975809e-01
-1.81651980e-01 -4.89329815e-01 -1.36578107e+00 -4.76936013e-01
-6.69634283e-01 2.78542340e-01 7.63567448e-01 1.19562590e+00
4.38111186e-01 2.90929586e-01 7.09254861e-01 -5.61160505e-01
-7.59513259e-01 -5.97421467e-01 -5.68792760e-01 2.44922370e-01
2.62338459e-01 -5.45785844e-01 -6.61965370e-01 3.19651157e-01] | [8.786056518554688, 5.507346153259277] |
3efbf1e1-974d-4c82-bd54-709864455d0c | graph-similarity-drives-zeolite-diffusionless | 1812.02685 | null | https://arxiv.org/abs/1812.02685v2 | https://arxiv.org/pdf/1812.02685v2.pdf | Graph similarity drives zeolite diffusionless transformations and intergrowth | Predicting and directing polymorphic transformations is a critical challenge in zeolite synthesis. Although interzeolite transformations enable selective crystallization, their design lacks predictions to connect framework similarity and experimental observations. Here, computational and theoretical tools are combined to data-mine, analyze and explain interzeolite relations. It is observed that building units are weak predictors of topology interconversion and insufficient to explain intergrowth. By introducing a supercell-invariant metric that compares crystal structures using graph theory, we show that topotactic and reconstructive (diffusionless) transformations occur only between graph-similar pairs. Furthermore, all known instances of intergrowth occur between either structurally-similar or graph-similar frameworks. Backed with exhaustive literature results, we identify promising pairs for realizing novel diffusionless transformations and intergrowth. Hundreds of low-distance pairs are identified among known zeolites, and thousands of hypothetical frameworks are connected to known zeolites counterparts. The theory opens a venue to understand and control zeolite polymorphism. | ['Rafael Gomez-Bombarelli', 'Elsa Olivetti', 'Zach Jensen', 'Daniel Schwalbe-Koda'] | 2018-12-06 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 2.72564471e-01 2.44269907e-01 -5.06069660e-01 -1.21566072e-01
3.77127938e-02 -7.10266173e-01 8.26073408e-01 1.69991538e-01
3.50582123e-01 9.66127157e-01 2.68672943e-01 -4.12943572e-01
-3.98359120e-01 -1.37180126e+00 -7.58436203e-01 -8.07063162e-01
-7.27459714e-02 1.20806110e+00 2.90190160e-01 -6.24729276e-01
4.81832206e-01 3.32495958e-01 -1.51176620e+00 4.81221914e-01
1.05231524e+00 1.82445839e-01 6.65195525e-01 4.80547041e-01
-3.74992728e-01 1.84434876e-01 2.69926280e-01 -1.38436809e-01
4.50733066e-01 -4.89882946e-01 -1.02280509e+00 5.59994094e-02
3.47771257e-01 3.48977208e-01 -2.49652088e-01 7.91882277e-01
1.95481524e-01 -2.27824926e-01 1.20948076e+00 -4.82690364e-01
-8.92788351e-01 7.75096714e-01 -3.35896015e-01 -8.78358409e-02
6.40042067e-01 5.79150356e-02 1.45362782e+00 -1.02333474e+00
1.26604474e+00 1.09966815e+00 3.83814573e-01 6.13068998e-01
-1.38520432e+00 -5.01463175e-01 -1.71822980e-01 -1.78152859e-01
-1.11740053e+00 -5.63414693e-01 4.90779519e-01 -7.75963664e-01
1.46676373e+00 3.78977537e-01 1.20893431e+00 7.07701027e-01
4.64116722e-01 2.87772208e-01 9.44454014e-01 -4.95055407e-01
1.93204746e-01 -1.97205469e-01 -1.21952094e-01 6.94450021e-01
8.09056461e-01 4.21354115e-01 -8.83790433e-01 -1.30878046e-01
7.95954823e-01 9.54056978e-02 -2.40510851e-01 -8.36829782e-01
-1.13371718e+00 6.71704531e-01 3.82619917e-01 2.32680306e-01
-2.16834359e-02 -6.22829720e-02 7.40401745e-02 3.92746449e-01
-3.77504639e-02 1.20258176e+00 -3.86839747e-01 -2.02478349e-01
-9.06882703e-01 4.61789608e-01 9.91355062e-01 1.15046978e+00
1.28440678e+00 -1.13387301e-03 6.90528154e-01 4.86522377e-01
3.56606543e-01 4.23426569e-01 3.16484153e-01 -3.42512429e-01
5.18210590e-01 7.98569500e-01 -1.49815276e-01 -5.45694470e-01
-3.84470314e-01 1.46919802e-01 -6.14492774e-01 -1.21802628e-01
3.04293931e-01 6.64903641e-01 -6.64692760e-01 1.14875066e+00
2.19477996e-01 -2.34494418e-01 -8.64344183e-03 3.24016154e-01
7.68453717e-01 3.25926840e-01 -2.90256262e-01 -4.00071889e-01
1.05277538e+00 -7.68253684e-01 2.41499096e-02 2.31521785e-01
6.91261351e-01 -6.63518965e-01 1.06715167e+00 3.75783741e-01
-1.19841564e+00 -2.18111306e-01 -1.21799457e+00 5.81258088e-02
-3.45036864e-01 -6.29223764e-01 1.13775671e+00 1.03395498e+00
-6.93506002e-01 9.43316579e-01 -9.08246517e-01 -6.94611549e-01
-8.03147536e-03 6.79646194e-01 -4.20866221e-01 1.46715343e-01
-9.43573058e-01 7.20474243e-01 6.96485639e-01 -5.23040295e-01
-8.56460512e-01 -7.06131816e-01 -2.28953540e-01 -3.73098969e-01
2.23233432e-01 -1.20632458e+00 6.64052248e-01 -1.02935396e-01
-1.58547390e+00 9.36969936e-01 -4.10874307e-01 -2.74147719e-01
2.77688622e-01 3.41073632e-01 -4.54102486e-01 -2.39850134e-01
1.64931387e-01 5.29344738e-01 4.36954916e-01 -1.00045288e+00
-4.25941199e-01 -2.22530365e-01 -6.43023029e-02 2.34343216e-01
-5.19591160e-02 -7.71266878e-01 2.45960176e-01 -2.42034644e-01
7.82794535e-01 -9.37690914e-01 -4.06871378e-01 -6.45117760e-01
-8.62957716e-01 -2.98743993e-01 6.24431014e-01 1.25829786e-01
1.08098698e+00 -1.43690169e+00 3.97072852e-01 7.00227737e-01
7.44774699e-01 -3.98614585e-01 1.67799905e-01 8.76765311e-01
-3.74638885e-01 5.64139068e-01 -2.93797344e-01 8.48035395e-01
3.51020582e-02 1.53340995e-01 -2.68691391e-01 4.29960310e-01
-1.96267948e-01 1.00775623e+00 -8.58863115e-01 1.85638949e-01
6.44835591e-01 4.07770537e-02 -1.10188365e+00 -2.22264171e-01
-5.82618415e-01 6.51759267e-01 -6.25190377e-01 8.99325907e-01
7.92790890e-01 -2.83287972e-01 7.63872027e-01 -9.29425731e-02
-5.91028690e-01 7.98422515e-01 -8.43664169e-01 1.63796198e+00
-3.13838989e-01 1.92521214e-01 -3.42114598e-01 -7.10562646e-01
1.27347219e+00 -5.95691651e-02 4.64037120e-01 -6.58958733e-01
-4.14878607e-01 6.20236695e-01 3.20245057e-01 -2.27572843e-01
7.13644922e-01 -4.32092845e-01 1.74890712e-01 4.97061461e-01
5.21023106e-03 -9.13688898e-01 5.38946807e-01 1.17229760e-01
1.01982486e+00 -8.19649324e-02 6.34203494e-01 -1.07485902e+00
5.05063117e-01 -3.60344024e-03 2.73755938e-01 8.70712876e-01
4.75225866e-01 5.22126734e-01 3.03580582e-01 -1.02999401e+00
-1.80743921e+00 -1.38197887e+00 -4.31599975e-01 8.24396133e-01
5.78325391e-01 -1.28119695e+00 -4.13529396e-01 7.22570792e-02
9.21743736e-02 4.21919763e-01 -3.65107656e-01 -1.95717365e-01
-4.16701645e-01 -1.01490271e+00 4.93281007e-01 2.13048339e-01
1.89399645e-01 -6.14223003e-01 -1.00839518e-01 4.71211314e-01
2.02248484e-01 -2.63839096e-01 -3.41339931e-02 4.97378886e-01
-9.89478588e-01 -1.26784635e+00 -1.06524475e-01 -4.49692070e-01
4.15810794e-01 3.98344666e-01 1.21132588e+00 1.64726764e-01
-4.28312093e-01 -1.43606305e-01 3.31729800e-02 5.03518134e-02
-6.62723482e-01 5.87217569e-01 4.44570601e-01 -9.50473726e-01
3.58975857e-01 -1.17351997e+00 -5.51349223e-01 4.60302562e-01
-5.19565105e-01 3.51502359e-01 3.20991784e-01 8.91926169e-01
8.82551432e-01 1.96026847e-01 1.27564102e-01 -1.19657278e+00
6.01472259e-01 -4.92001772e-01 -4.72620428e-01 6.27956927e-01
-1.10335433e+00 6.54407799e-01 5.96120059e-01 5.71460724e-02
-1.09644186e+00 2.56944031e-01 -4.60353829e-02 3.25421125e-01
-2.19395071e-01 3.64842772e-01 -4.60360020e-01 -2.92055774e-02
8.61895859e-01 4.25352693e-01 -4.15536910e-01 -2.82277852e-01
5.56983948e-01 2.20357910e-01 2.85202682e-01 -1.30211699e+00
8.15252662e-01 7.89800644e-01 1.34106323e-01 -1.48047149e+00
-3.06435704e-01 -3.34605604e-01 -9.71996427e-01 1.15027092e-01
7.87082195e-01 -7.34049916e-01 -8.27788830e-01 -2.07146898e-01
-7.64110208e-01 -2.21321225e-01 -4.31888223e-01 1.86834514e-01
-9.11232233e-01 3.25402141e-01 -3.54098529e-01 -3.68875682e-01
7.63319880e-02 -1.53155088e+00 1.15438533e+00 -6.94423914e-02
-5.88386595e-01 -1.13989818e+00 5.96133709e-01 3.95142436e-01
2.92342722e-01 -1.31875247e-01 1.31819689e+00 -5.26012778e-01
-1.13541234e+00 5.70944130e-01 4.80192453e-02 -8.71151447e-01
6.73931062e-01 3.16284269e-01 -5.53830445e-01 -1.49679899e-01
-3.35403174e-01 -1.19165726e-01 1.26380002e+00 3.14806342e-01
7.63538718e-01 -1.18293554e-01 -8.56033564e-01 9.65321064e-01
1.19148552e+00 4.42031384e-01 7.29328513e-01 7.17209995e-01
1.02601087e+00 3.15574199e-01 1.06446452e-01 4.03785795e-01
1.14955999e-01 5.23399830e-01 2.26081997e-01 2.77083725e-01
7.06942752e-02 -7.49183595e-01 2.69715935e-01 1.01303315e+00
-7.68724561e-01 -3.22197348e-01 -1.20794570e+00 2.75997072e-02
-1.41667032e+00 -1.16168594e+00 -6.66926682e-01 2.12195158e+00
6.45869613e-01 3.11463028e-01 1.35113627e-01 -7.41013288e-02
7.06825793e-01 1.14009060e-01 -7.58497059e-01 -2.88707525e-01
-6.72181010e-01 5.50809562e-01 7.33580470e-01 7.17377782e-01
-7.39844143e-01 1.36910272e+00 7.74380875e+00 7.11309850e-01
-1.07952499e+00 -2.52774358e-01 1.41122892e-01 2.97637563e-02
-1.11006904e+00 7.24279523e-01 -8.02762866e-01 -3.60727347e-02
6.16920233e-01 -4.26095992e-01 8.26571524e-01 6.66954637e-01
-4.73147295e-02 2.49045059e-01 -1.27316141e+00 8.85553002e-01
-3.95363241e-01 -2.05383277e+00 4.89555895e-01 5.26950181e-01
1.10839903e+00 2.01242521e-01 1.68820769e-02 -2.82582015e-01
8.10203254e-01 -1.21154189e+00 5.95620811e-01 3.99047166e-01
9.23012793e-01 -5.56633472e-01 -1.32344604e-01 -1.83743641e-01
-1.80993390e+00 2.00603023e-01 -8.54200721e-01 -4.00223613e-01
1.74084261e-01 7.69808590e-01 -1.31163502e+00 8.61104786e-01
3.74850333e-01 1.23228765e+00 -2.72652119e-01 5.75162113e-01
-2.62095839e-01 2.50081003e-01 -2.46151507e-01 3.36579718e-02
6.53365534e-03 -8.60173106e-01 6.30064130e-01 8.03024769e-01
4.08999115e-01 -7.57058263e-02 1.29262477e-01 1.17558992e+00
4.91813496e-02 -1.93749201e-02 -1.00769806e+00 -3.59139293e-01
5.13156533e-01 6.28007412e-01 -1.07942152e+00 -2.61561781e-01
-3.08295190e-01 6.61315322e-01 1.33903578e-01 1.21805951e-01
-3.08411717e-01 -9.60134994e-03 7.68100619e-01 4.74005252e-01
2.20835865e-01 -2.58860141e-01 -2.87736386e-01 -1.41914904e+00
-3.31580043e-01 -5.92392147e-01 1.20350793e-01 -3.86177301e-01
-1.29110885e+00 2.69858390e-01 -3.77767324e-01 -7.73000419e-01
-1.35510817e-01 -5.90515196e-01 -7.05555558e-01 2.85425544e-01
-1.03691065e+00 -9.52346265e-01 -7.94463083e-02 3.40107203e-01
5.63363194e-01 -4.77212757e-01 7.00715482e-01 -1.00561894e-01
-3.12152386e-01 1.39831519e-02 6.57565951e-01 -8.45359981e-01
6.10819519e-01 -1.49795687e+00 6.49508238e-01 1.94255054e-01
2.76212186e-01 9.91896212e-01 8.99249315e-01 -1.12560296e+00
-1.69829023e+00 -1.01590097e+00 5.89377522e-01 -6.73404813e-01
9.37310636e-01 -6.61623478e-01 -6.16299808e-01 4.70062762e-01
-9.65929478e-02 -6.06486797e-01 7.18261719e-01 4.74752188e-01
-3.87404889e-01 2.17734680e-01 -7.72026598e-01 5.87260604e-01
1.70454144e+00 -9.48442996e-01 -6.80253923e-01 6.35485232e-01
4.96931046e-01 -1.69053361e-01 -1.00032258e+00 1.58955738e-01
8.59443963e-01 -1.24364686e+00 1.23714435e+00 -6.96630299e-01
5.89110255e-01 -4.85151887e-01 -3.67280900e-01 -1.07746220e+00
-5.06663263e-01 -7.17525899e-01 3.62254530e-01 8.10841560e-01
8.20445538e-01 -5.12012422e-01 1.11482859e+00 5.13278127e-01
-3.80937129e-01 -2.29947954e-01 -6.88881993e-01 -1.00162148e+00
3.86071086e-01 1.41638383e-01 1.00886655e+00 1.33359289e+00
6.97880805e-01 5.19591033e-01 -1.80847958e-01 -4.03392524e-01
4.93359894e-01 4.38810110e-01 7.50253439e-01 -1.53139150e+00
-3.04435521e-01 -5.72224796e-01 -4.56542850e-01 -1.06382561e+00
2.17874318e-01 -1.56810582e+00 -4.93720382e-01 -1.30457222e+00
2.87724763e-01 -8.02267671e-01 1.57942042e-01 9.73790884e-02
5.60323954e-01 -2.18774468e-01 -3.03927690e-01 6.80472970e-01
-5.00580549e-01 6.62368834e-01 1.36052990e+00 -5.08327842e-01
-5.76313734e-01 -3.28852713e-01 -6.25522971e-01 5.32631457e-01
8.44531536e-01 -4.33185697e-01 -5.67208409e-01 -2.32585315e-02
9.80998814e-01 1.72469363e-01 -2.62745231e-01 -8.90236855e-01
1.42190650e-01 -5.08899629e-01 8.33586305e-02 -6.33147240e-01
-5.28489314e-02 -4.95630056e-01 8.86253536e-01 7.45659709e-01
-1.71632707e-01 -8.09563547e-02 -3.97620082e-01 7.64938533e-01
-6.09542057e-02 -1.44858703e-01 4.92058963e-01 -7.15607285e-01
-6.09850049e-01 4.60419208e-01 -5.71682394e-01 -1.99920103e-01
9.59951222e-01 -8.22434306e-01 -4.04291630e-01 1.47709370e-01
-9.19925570e-01 -3.08793545e-01 1.24594223e+00 1.56708539e-01
4.00353938e-01 -1.07105017e+00 -2.62446940e-01 3.40385050e-01
3.33292961e-01 1.96695000e-01 7.04303011e-02 3.98195356e-01
-1.07499993e+00 8.36051583e-01 -4.32493806e-01 -9.38678145e-01
-8.76859009e-01 4.06187862e-01 2.80719608e-01 -1.65847123e-01
-3.64363074e-01 7.55712509e-01 8.06826830e-01 -9.45198774e-01
-7.82445431e-01 -4.31736350e-01 2.87145853e-01 -2.49756053e-01
-1.36747966e-02 3.43251169e-01 3.39088082e-01 -2.35611036e-01
-2.33881637e-01 6.13550603e-01 -3.10359269e-01 3.89557272e-01
1.69877696e+00 -1.63791537e-01 -5.50853312e-01 3.19539398e-01
5.65695107e-01 1.97548106e-01 -8.95948529e-01 2.22993895e-01
2.24705592e-01 -4.86651868e-01 -7.17794120e-01 -7.42262080e-02
-6.98693812e-01 7.58160770e-01 1.82159841e-01 2.31074095e-01
4.14672822e-01 5.11021852e-01 3.79763603e-01 1.01906621e+00
6.03886902e-01 -8.23928356e-01 -2.82704718e-02 5.86491883e-01
5.16212642e-01 -8.89432967e-01 2.17786610e-01 -6.22772694e-01
2.23555658e-02 1.41170943e+00 7.33047783e-01 -3.15734744e-02
5.68766713e-01 2.59955168e-01 -8.35734189e-01 -6.71962678e-01
-1.12774134e+00 1.71860129e-01 -1.66879371e-01 6.87295139e-01
8.33916426e-01 4.15493131e-01 -3.11446309e-01 7.23050535e-02
-9.95666206e-01 -5.48038006e-01 7.80148745e-01 7.19673872e-01
-8.63736987e-01 -1.85918295e+00 -2.22068280e-01 8.00702095e-01
2.61348158e-01 -4.27669346e-01 -7.44418502e-01 8.72603595e-01
1.14784807e-01 4.18309987e-01 1.06922746e-01 -6.07469320e-01
5.80466799e-02 -1.17803156e-01 7.49479711e-01 -5.59182525e-01
-1.39515609e-01 7.97569379e-02 1.77641168e-01 -1.85853124e-01
-5.52259624e-01 -6.79050803e-01 -1.22637248e+00 -7.16974497e-01
-8.42445970e-01 5.96851528e-01 3.88663560e-02 8.12590003e-01
2.06525728e-01 2.69717097e-01 3.68402272e-01 -6.89012945e-01
2.40619615e-01 -1.98726386e-01 -1.21280968e+00 2.75281101e-01
-2.49682799e-01 -9.31576967e-01 4.82352190e-02 -9.04379471e-04] | [5.11847448348999, 5.533304691314697] |
3eaf39ca-ea84-4d09-8e24-4f76b57010b4 | dpll-mapf-an-integration-of-multi-agent-path | 2111.06494 | null | https://arxiv.org/abs/2111.06494v1 | https://arxiv.org/pdf/2111.06494v1.pdf | DPLL(MAPF): an Integration of Multi-Agent Path Finding and SAT Solving Technologies | In multi-agent path finding (MAPF), the task is to find non-conflicting paths for multiple agents from their initial positions to given individual goal positions. MAPF represents a classical artificial intelligence problem often addressed by heuristic-search. An important alternative to search-based techniques is compilation of MAPF to a different formalism such as Boolean satisfiability (SAT). Contemporary SAT-based approaches to MAPF regard the SAT solver as an external tool whose task is to return an assignment of all decision variables of a Boolean model of input MAPF. We present in this short paper a novel compilation scheme called DPLL(MAPF) in which the consistency checking of partial assignments of decision variables with respect to the MAPF rules is integrated directly into the SAT solver. This scheme allows for far more automated compilation where the SAT solver and the consistency checking procedure work together simultaneously to create the Boolean model and to search for its satisfying assignment. | ['Pavel Surynek', 'Martin Čapek'] | 2021-11-11 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 3.08478743e-01 6.14295483e-01 -2.69289494e-01 -3.33238959e-01
-4.95889872e-01 -8.87531102e-01 5.32583952e-01 4.23644453e-01
1.50847197e-01 1.42275381e+00 -3.63767743e-01 -5.41172385e-01
-5.31082034e-01 -1.36710346e+00 -3.85987788e-01 -4.90668565e-01
-2.66612113e-01 1.29270232e+00 6.72498226e-01 -4.05768216e-01
8.61079916e-02 3.59671086e-01 -1.26524127e+00 1.00422911e-01
7.95015395e-01 6.74627423e-01 -3.51245664e-02 5.92673302e-01
-3.31917018e-01 6.99672639e-01 -6.44000113e-01 -3.74914587e-01
4.44115937e-01 -7.84778655e-01 -1.50063896e+00 8.57360736e-02
-8.37044120e-02 5.05493442e-03 4.95672114e-02 1.36776388e+00
-4.30566192e-01 6.91725640e-03 2.74685830e-01 -2.16195440e+00
2.13016585e-01 6.66076481e-01 -2.47807965e-01 -1.32434368e-02
9.72488821e-01 9.74620730e-02 1.19100773e+00 1.13643460e-01
9.93375421e-01 1.27675772e+00 1.99801594e-01 4.09200072e-01
-1.34244859e+00 -1.53327480e-01 2.43530765e-01 7.99697161e-01
-1.41853249e+00 -1.02242127e-01 4.07005221e-01 -1.35484979e-01
1.49044931e+00 7.93627381e-01 9.16123331e-01 2.56243229e-01
4.49957639e-01 1.90612063e-01 1.38541961e+00 -6.90346837e-01
7.22415030e-01 1.44324452e-01 2.95752347e-01 7.13161170e-01
5.01113057e-01 3.58175278e-01 -3.66048455e-01 -4.66054708e-01
6.02453530e-01 -5.60167015e-01 2.57728919e-02 -3.97595823e-01
-1.11867082e+00 1.10278463e+00 1.40740678e-01 1.64567560e-01
-3.65378857e-01 3.63876045e-01 2.69351751e-01 6.35403395e-01
-2.93263078e-01 8.79131556e-01 -5.48412383e-01 5.66092767e-02
-4.75047320e-01 8.19125891e-01 1.23653042e+00 9.19511914e-01
8.03715646e-01 -3.28483164e-01 7.84257650e-02 -2.08282456e-01
3.93026799e-01 4.86442894e-01 -3.03953379e-01 -1.28014088e+00
2.91124701e-01 8.95526052e-01 3.81054908e-01 -9.02704954e-01
-2.81181395e-01 -1.68582112e-01 -1.55969292e-01 7.11539090e-01
5.06006420e-01 -1.16283283e-01 -4.85498518e-01 1.75623786e+00
6.33914173e-01 -8.60591456e-02 3.18229258e-01 8.45808804e-01
4.54242051e-01 9.48035955e-01 -4.39139992e-01 -7.75806665e-01
1.11327982e+00 -1.03139353e+00 -6.47033632e-01 -3.33209604e-01
5.36802828e-01 -2.22691104e-01 9.01778191e-02 7.54793048e-01
-1.56206751e+00 3.84039700e-01 -1.17070961e+00 3.25161904e-01
-3.45264584e-01 -7.84131885e-01 9.56964672e-01 3.29817593e-01
-1.40162802e+00 8.57942179e-02 -6.02212727e-01 -2.30108708e-01
-9.36535299e-02 9.61811543e-01 -5.90328932e-01 -3.59576970e-01
-1.02737439e+00 1.18066549e+00 6.12256467e-01 -1.57646202e-02
-7.54331052e-01 1.22791715e-01 -9.13834989e-01 1.02039725e-01
1.22024167e+00 -7.75966704e-01 1.55898166e+00 -9.65385258e-01
-1.52871919e+00 6.17309868e-01 -4.66210157e-01 -5.63834429e-01
2.11220667e-01 7.35567689e-01 -2.59237289e-01 9.34414640e-02
2.81169623e-01 3.51855338e-01 3.18815768e-01 -1.20679915e+00
-7.44651496e-01 -3.12857181e-01 6.93169773e-01 2.25671660e-03
7.96101987e-01 2.29855537e-01 -1.58554539e-01 3.26062173e-01
4.26448971e-01 -8.32330346e-01 -7.40016222e-01 -4.54221696e-01
-5.36715984e-01 -2.47185528e-01 3.67771626e-01 -8.54772106e-02
1.35391951e+00 -1.39642954e+00 7.96433330e-01 8.14605236e-01
1.99829310e-01 -2.80025035e-01 -2.70333916e-01 8.35433304e-01
-1.89330846e-01 -2.32904211e-01 -1.62622228e-01 1.43329456e-01
3.67932588e-01 7.63607442e-01 -1.26884342e-03 4.48695838e-01
2.06141621e-01 9.51954603e-01 -9.46828485e-01 -6.15593791e-01
1.09301463e-01 -5.19561291e-01 -7.92203188e-01 -2.83183847e-02
-1.07110107e+00 1.70782670e-01 -4.44045722e-01 6.22312605e-01
8.04453075e-01 1.78591385e-02 7.27713883e-01 5.73990345e-01
-3.74935150e-01 5.04816413e-01 -1.74502754e+00 1.34839439e+00
6.01944365e-02 1.98255330e-01 2.71083504e-01 -8.53840768e-01
8.49351287e-01 2.03654960e-01 3.57629418e-01 -9.57294703e-01
3.04666162e-01 2.96747893e-01 1.56406194e-01 -2.45582566e-01
3.68627578e-01 -3.35693389e-01 -5.38486481e-01 6.04949951e-01
-2.11046949e-01 -2.81399667e-01 6.49728835e-01 1.70513928e-01
1.41226637e+00 -2.18013644e-01 7.71615386e-01 -3.97316724e-01
9.87945318e-01 9.06916678e-01 7.34894872e-01 7.44310558e-01
1.90354124e-01 -2.58689046e-01 9.60948348e-01 -8.34574461e-01
-6.88091040e-01 -6.01222277e-01 2.61044204e-01 3.43915612e-01
3.81044149e-01 -7.56312251e-01 -7.28624642e-01 -5.35110414e-01
-2.15978220e-01 7.25196123e-01 -4.96055901e-01 -2.43180450e-02
-7.73227870e-01 -2.37342268e-01 -4.27202620e-02 -4.03571539e-02
3.02430630e-01 -9.52533960e-01 -1.16895974e+00 5.44384062e-01
-2.16234446e-01 -8.09060931e-01 3.15817863e-01 5.59990525e-01
-5.18895388e-01 -1.29000437e+00 3.40280890e-01 -6.84943676e-01
8.72145295e-01 9.23186988e-02 9.84581411e-01 3.96830082e-01
1.17343664e-02 -1.56467065e-01 -4.48705971e-01 -2.82467306e-01
-5.59386790e-01 5.94409071e-02 -3.34328562e-01 -6.04364991e-01
1.35913178e-01 -3.14135760e-01 2.88186282e-01 4.79846805e-01
-6.85175776e-01 -5.84970601e-02 -1.65562611e-02 7.77573466e-01
8.30975771e-01 9.77549016e-01 2.85701334e-01 -6.73661351e-01
4.62153375e-01 -3.28548729e-01 -1.33048820e+00 4.78294909e-01
-4.50794965e-01 1.05135515e-01 6.85286999e-01 6.89293891e-02
-5.10304570e-01 1.47317052e-01 3.15714926e-01 1.14347532e-01
-3.54639515e-02 9.76847708e-01 -4.55241054e-01 -1.40901268e-01
2.87820399e-01 9.77267697e-02 5.29169627e-02 1.63423181e-01
-1.86254214e-02 -5.71254548e-03 6.83849096e-01 -8.18284035e-01
8.64998162e-01 9.63587537e-02 8.54683518e-01 2.38729879e-01
-2.48589724e-01 -2.59692855e-02 -2.76339352e-01 -2.16634676e-01
4.96126503e-01 -1.56500712e-01 -1.02660477e+00 -8.56168743e-04
-1.29707098e+00 -5.41589737e-01 -4.63274419e-01 -1.27419336e-02
-7.77811110e-01 -1.75766442e-02 9.51750949e-02 -1.11350691e+00
1.22460082e-01 -1.28210807e+00 7.74750054e-01 1.19474694e-01
-5.43365538e-01 -8.17251801e-01 4.60800856e-01 1.49755850e-01
9.52727571e-02 5.61848819e-01 1.17030168e+00 -7.44388282e-01
-8.12064469e-01 -3.53375047e-01 2.38539558e-02 -6.19813502e-01
-1.90504462e-01 -5.14544025e-02 -2.22011641e-01 -2.89957732e-01
-1.02682523e-02 2.43363157e-01 -1.92871943e-01 4.39978093e-01
3.78657430e-01 -7.25936413e-01 -5.92885971e-01 -7.95401912e-03
1.83263767e+00 7.34804988e-01 6.76229060e-01 8.41176569e-01
-4.07774091e-01 6.79679394e-01 1.04876471e+00 3.89748037e-01
4.75483388e-01 9.21289504e-01 6.94332600e-01 2.83767670e-01
4.92789179e-01 2.24862620e-01 1.27955005e-01 -7.65093714e-02
-3.17769498e-02 -5.93573213e-01 -1.15346730e+00 3.79703671e-01
-2.21087313e+00 -1.13255072e+00 -5.04126906e-01 2.12373185e+00
7.33279228e-01 2.76542574e-01 3.55016410e-01 5.13625264e-01
6.16980016e-01 -4.82693613e-01 -5.88211827e-02 -1.02461922e+00
1.44484475e-01 2.97483534e-01 3.84503305e-01 1.32994032e+00
-6.82301164e-01 8.73245299e-01 6.81327009e+00 1.66079983e-01
-8.04091930e-01 -2.47971881e-02 -1.61759824e-01 -9.80637521e-02
-3.57577801e-01 5.34375787e-01 -5.55274129e-01 -5.54285645e-02
8.79685044e-01 -4.30503100e-01 9.95792329e-01 5.03088057e-01
8.79145563e-02 -7.40585744e-01 -1.31349969e+00 3.24614614e-01
-3.38062942e-02 -1.32395995e+00 -3.28063488e-01 2.82450020e-01
7.27937877e-01 -5.89073420e-01 -3.71538371e-01 6.56162053e-02
6.43469512e-01 -1.15007651e+00 1.00897408e+00 1.38941184e-01
3.56956691e-01 -1.08792567e+00 9.93993044e-01 3.31409335e-01
-9.92058635e-01 -2.36618280e-01 -8.21865946e-02 -6.66324794e-01
5.49453676e-01 5.47858812e-02 -1.08250237e+00 9.17637587e-01
4.16707784e-01 9.73529592e-02 2.29074303e-02 1.34725225e+00
-2.94220597e-01 -2.54763782e-01 -4.56911683e-01 -9.40233767e-02
5.39481819e-01 -3.68748933e-01 8.88020635e-01 6.69347346e-01
4.47822213e-02 4.07280266e-01 2.80615509e-01 1.00980055e+00
8.50707412e-01 -3.62704813e-01 -4.32936013e-01 7.28077516e-02
3.47863466e-01 8.47187638e-01 -1.06446981e+00 -2.32278898e-01
-2.23005503e-01 5.13501585e-01 1.69952810e-02 1.68961734e-01
-8.94580245e-01 -3.23807210e-01 5.59521914e-01 7.95619264e-02
2.03337610e-01 -2.60110855e-01 -3.61801445e-01 -6.25617146e-01
8.09838548e-02 -1.19580638e+00 5.18419683e-01 -8.14976037e-01
-6.04297042e-01 7.79304028e-01 5.42144537e-01 -6.75926566e-01
-6.54751658e-01 -3.75955909e-01 -7.93124497e-01 8.46815884e-01
-1.53314805e+00 -7.65306890e-01 -1.39780357e-01 7.25563288e-01
2.95164078e-01 6.10442273e-02 1.05981624e+00 -2.50975132e-01
-6.08436167e-01 5.17235212e-02 -5.80328047e-01 -6.35627091e-01
-1.62454590e-01 -1.17197299e+00 -7.92793278e-03 1.12773550e+00
-3.01699281e-01 6.20457768e-01 1.13528121e+00 -8.09678435e-01
-1.84912050e+00 -7.04713881e-01 1.34552217e+00 7.26922415e-03
6.82730138e-01 5.26465848e-02 -3.98024291e-01 1.01797271e+00
3.89411390e-01 -3.55078727e-01 2.90313303e-01 -1.24075912e-01
-3.88948526e-03 3.98294739e-02 -1.36349070e+00 4.02873963e-01
6.39951766e-01 6.38779178e-02 -5.45306087e-01 2.91988760e-01
4.92697656e-01 -6.60180211e-01 -4.84222829e-01 9.61414129e-02
7.11492449e-02 -6.82087123e-01 5.54982960e-01 -8.61370146e-01
2.88670093e-01 -1.09300172e+00 -3.67835432e-01 -1.13910139e+00
-5.40759206e-01 -8.99613559e-01 3.00037730e-02 7.31830657e-01
6.02346420e-01 -9.79057074e-01 7.19399333e-01 8.32883120e-01
-2.70248443e-01 -6.81119084e-01 -1.49305105e+00 -7.85035491e-01
-3.04469466e-01 -1.54824913e-01 1.12871921e+00 8.79075289e-01
9.06312406e-01 2.73458630e-01 6.79740831e-02 4.54850942e-01
3.97660524e-01 6.15595281e-01 5.96921742e-01 -1.50361323e+00
-6.85184419e-01 -5.58708847e-01 -3.51534426e-01 -1.73975661e-01
3.17747176e-01 -9.22312140e-01 2.25646973e-01 -1.94075823e+00
1.03556596e-01 -4.78108793e-01 3.52680773e-01 9.47490752e-01
6.33013368e-01 -2.71772087e-01 8.67684036e-02 -2.33130723e-01
-9.32457447e-01 -1.77151471e-01 1.39975655e+00 -2.81779945e-01
-2.16832429e-01 4.24723029e-02 -6.14622295e-01 3.26419741e-01
6.74439669e-01 -7.89695740e-01 -4.48672831e-01 -1.99346468e-01
9.02918577e-01 8.64299297e-01 3.83478194e-01 -6.72711611e-01
5.80046713e-01 -1.05400670e+00 -5.34517050e-01 -2.35708296e-01
1.53077886e-01 -1.22353089e+00 1.12248194e+00 8.72089982e-01
-9.09949541e-02 4.69498128e-01 2.95627892e-01 -7.36682042e-02
-2.84454644e-01 -7.34364152e-01 3.66727978e-01 -2.78671175e-01
-6.91129982e-01 -2.87750244e-01 -6.97394907e-01 -3.93084109e-01
1.58881998e+00 -3.18012834e-01 -5.61984599e-01 -2.75138438e-01
-8.13190937e-01 5.68646669e-01 6.08964205e-01 -3.43429208e-01
4.74477232e-01 -9.66880798e-01 -3.89059633e-01 1.93272412e-01
-9.71712545e-02 3.10980082e-01 -2.35406429e-01 9.69182432e-01
-7.44519830e-01 7.72086203e-01 -4.40927416e-01 -1.90316662e-01
-1.30006957e+00 7.98041344e-01 6.45219624e-01 -4.84889388e-01
-3.27018738e-01 6.17753386e-01 -4.13967222e-01 -1.10734016e-01
1.09687895e-01 -4.49966758e-01 4.57586311e-02 -3.46121877e-01
7.13672876e-01 3.76476258e-01 1.40263990e-01 -3.94902080e-01
-9.13355231e-01 2.39102706e-01 1.42545015e-01 -4.18707192e-01
1.43123043e+00 -1.18044019e-03 -8.09906721e-01 -3.29569727e-01
5.16851425e-01 -1.73745770e-02 -4.47749168e-01 -7.11847246e-02
-5.17250039e-02 -4.00548190e-01 6.23913221e-02 -1.21776319e+00
-8.84917080e-01 -7.86468685e-02 -3.74570996e-01 6.36872053e-01
1.23609400e+00 2.39624247e-01 3.16740155e-01 5.55884421e-01
1.02264094e+00 -7.65727282e-01 -4.98046070e-01 7.07263231e-01
7.43030488e-01 -6.82035625e-01 1.42311111e-01 -9.38061416e-01
-4.28089142e-01 1.21154082e+00 6.77549064e-01 -3.11391558e-02
-1.38402060e-01 9.06356812e-01 -3.05756778e-01 -4.07234937e-01
-1.03421462e+00 -2.73847133e-01 -3.09014618e-01 6.67689979e-01
-5.32625318e-01 1.24507537e-02 -6.01759970e-01 6.22236192e-01
-5.75988710e-01 1.45808265e-01 7.95858681e-01 1.34102619e+00
-5.97573936e-01 -1.69932973e+00 -7.47617841e-01 -2.31324113e-03
8.00929144e-02 3.38302374e-01 -7.91056216e-01 7.17697442e-01
3.18286180e-01 1.19603729e+00 -3.76154110e-02 -2.82101929e-01
2.55156726e-01 -1.69410482e-01 1.02212930e+00 -6.25736952e-01
-8.30546618e-01 -4.90784459e-02 7.14864850e-01 -7.36988783e-01
-3.64192277e-01 -1.05678344e+00 -1.62078989e+00 -5.13262510e-01
-4.22797412e-01 6.43836737e-01 3.57403785e-01 1.14328086e+00
-3.41532677e-01 4.94321167e-01 3.60469818e-01 -4.22355294e-01
-2.62631536e-01 6.84926137e-02 -4.16530460e-01 -4.35639858e-01
2.56763935e-01 -6.99544549e-01 -3.42997499e-02 -4.80539620e-01] | [4.986104965209961, 1.9085493087768555] |
00fe571d-210b-4755-865c-8ed8e066ff07 | omni-aggregation-networks-for-lightweight | 2304.10244 | null | https://arxiv.org/abs/2304.10244v2 | https://arxiv.org/pdf/2304.10244v2.pdf | Omni Aggregation Networks for Lightweight Image Super-Resolution | While lightweight ViT framework has made tremendous progress in image super-resolution, its uni-dimensional self-attention modeling, as well as homogeneous aggregation scheme, limit its effective receptive field (ERF) to include more comprehensive interactions from both spatial and channel dimensions. To tackle these drawbacks, this work proposes two enhanced components under a new Omni-SR architecture. First, an Omni Self-Attention (OSA) block is proposed based on dense interaction principle, which can simultaneously model pixel-interaction from both spatial and channel dimensions, mining the potential correlations across omni-axis (i.e., spatial and channel). Coupling with mainstream window partitioning strategies, OSA can achieve superior performance with compelling computational budgets. Second, a multi-scale interaction scheme is proposed to mitigate sub-optimal ERF (i.e., premature saturation) in shallow models, which facilitates local propagation and meso-/global-scale interactions, rendering an omni-scale aggregation building block. Extensive experiments demonstrate that Omni-SR achieves record-high performance on lightweight super-resolution benchmarks (e.g., 26.95 dB@Urban100 $\times 4$ with only 792K parameters). Our code is available at \url{https://github.com/Francis0625/Omni-SR}. | ['Jinfan Liu', 'Yutian Liu', 'Bingbing Ni', 'Xuanhong Chen', 'Hang Wang'] | 2023-04-20 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Omni_Aggregation_Networks_for_Lightweight_Image_Super-Resolution_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Omni_Aggregation_Networks_for_Lightweight_Image_Super-Resolution_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-super-resolution'] | ['computer-vision'] | [ 1.07793301e-01 2.89745796e-02 4.59841080e-02 -1.67705685e-01
-8.42188120e-01 4.67991866e-02 2.52119511e-01 -4.14406896e-01
-2.27330193e-01 6.61906004e-01 4.58392262e-01 -1.85203984e-01
-1.15145385e-01 -7.96391964e-01 -6.15691006e-01 -8.02738369e-01
-4.66801226e-01 -3.63582939e-01 6.81676567e-01 -2.77319878e-01
1.13869779e-01 2.94656456e-01 -1.77369297e+00 3.94842654e-01
1.16681397e+00 1.16530550e+00 5.53031862e-01 7.74995148e-01
-4.13737409e-02 6.47175252e-01 -3.27512145e-01 1.56800434e-01
2.37479210e-01 -1.32980257e-01 -5.66014171e-01 -2.97212541e-01
5.54653287e-01 -5.49642861e-01 -4.36011940e-01 1.11357093e+00
7.89331436e-01 -1.21599603e-02 2.10021347e-01 -5.93638241e-01
-7.39305139e-01 7.23770142e-01 -1.30070269e+00 8.07093561e-01
-2.58880913e-01 3.07514727e-01 1.06781530e+00 -1.04812920e+00
3.64249647e-01 1.18786323e+00 4.66111690e-01 4.61605102e-01
-1.37969422e+00 -1.16138577e+00 4.63031650e-01 8.78626332e-02
-1.52000558e+00 -4.99365181e-01 4.89514828e-01 1.62882585e-04
1.04764843e+00 3.60238075e-01 5.31043053e-01 7.37880886e-01
1.39624327e-01 6.95551634e-01 1.28174460e+00 2.24067234e-02
2.31173307e-01 -2.22962089e-02 3.69488567e-01 2.49254882e-01
-1.84590963e-03 9.21640471e-02 -7.15039968e-01 1.02450222e-01
1.34053838e+00 -2.20015228e-01 -3.56557399e-01 1.78953141e-01
-9.66153502e-01 5.32208741e-01 7.21898377e-01 2.35262737e-01
-2.88986713e-01 1.32313430e-01 1.84584856e-01 -7.99302086e-02
7.52784371e-01 2.43237853e-01 -3.82806510e-01 4.35363278e-02
-1.00692022e+00 2.54498065e-01 -5.43819880e-03 1.10475659e+00
8.79188418e-01 1.64373338e-01 -4.19809341e-01 1.12658560e+00
1.47815689e-01 5.68440318e-01 2.30096489e-01 -1.08753073e+00
3.58098537e-01 4.52890426e-01 1.28495088e-02 -7.40793467e-01
-5.69903612e-01 -1.07873309e+00 -1.32620621e+00 3.07778239e-01
3.42652127e-02 -6.88734725e-02 -8.43775868e-01 1.76537633e+00
4.09510821e-01 6.03854597e-01 -2.52340227e-01 1.20271420e+00
1.07926869e+00 7.12843120e-01 2.03364730e-01 -1.51938275e-01
1.53981793e+00 -9.39867437e-01 -3.98273647e-01 -1.88674733e-01
2.35160291e-01 -5.38884044e-01 1.16398644e+00 3.74702543e-01
-1.60319877e+00 -8.01071703e-01 -1.05957985e+00 -3.27500761e-01
-1.11813016e-01 -5.29402234e-02 7.41324008e-01 2.89042503e-01
-1.39535654e+00 2.34670416e-01 -6.55881405e-01 -2.63192989e-02
6.50391102e-01 3.84860575e-01 3.21934164e-01 2.80562285e-02
-1.28890157e+00 1.10122070e-01 -6.53750896e-02 6.26439601e-02
-5.23929894e-01 -1.23826671e+00 -4.90114421e-01 3.05868983e-01
1.93045959e-01 -6.02419615e-01 7.68056989e-01 -6.92538917e-01
-1.16063571e+00 5.07393181e-01 -4.14872587e-01 -4.88889158e-01
1.36772290e-01 -1.67327344e-01 -5.65117538e-01 2.02801436e-01
8.19909275e-02 8.07054639e-01 8.64275217e-01 -1.38822317e+00
-8.53275359e-01 -7.21828222e-01 1.62054092e-01 5.38378894e-01
-5.42486727e-01 2.99152941e-01 -7.69960999e-01 -6.43959999e-01
2.84948260e-01 -4.93979394e-01 -4.09629315e-01 -1.76699594e-01
-2.43202612e-01 1.66200697e-01 6.36957169e-01 -5.91162741e-01
1.77187788e+00 -2.24923468e+00 -1.88360717e-02 3.88960540e-02
8.54991317e-01 2.28660032e-01 -7.29157478e-02 -2.18701810e-01
-5.78236729e-02 1.88608050e-01 -1.26317605e-01 -5.45472741e-01
-4.48357940e-01 -8.70735645e-02 -3.12468141e-01 2.64686465e-01
1.10775366e-01 6.52134359e-01 -5.95239341e-01 -5.01780629e-01
1.61733299e-01 9.29161727e-01 -7.02122808e-01 -6.22329637e-02
1.61563709e-01 5.31271160e-01 -5.26910543e-01 6.84912920e-01
1.31226492e+00 -6.87111557e-01 -1.09926373e-01 -3.54759663e-01
-6.85046196e-01 1.57882884e-01 -1.37840295e+00 1.57829165e+00
-6.46894097e-01 4.34068352e-01 5.69069266e-01 -5.60838163e-01
6.75808907e-01 -2.54469439e-02 4.86832738e-01 -1.26846766e+00
-1.24429762e-01 8.67269710e-02 -1.52727187e-01 -3.04012615e-02
6.65309548e-01 4.48517978e-01 3.33565116e-01 1.25492737e-01
-1.44221261e-01 4.92170423e-01 1.10970633e-02 3.18350196e-01
1.02537489e+00 -9.28862777e-04 -1.00473642e-01 -7.46411204e-01
3.94779295e-01 -3.78345042e-01 6.17907941e-01 8.82592499e-01
-8.34018141e-02 7.32025623e-01 1.89744264e-01 -3.28232259e-01
-9.90759075e-01 -1.10139120e+00 -5.11398554e-01 1.22977340e+00
6.50521755e-01 -4.57655787e-01 -7.41159379e-01 1.03523143e-01
-2.92836159e-01 3.31704140e-01 -6.95744216e-01 2.28209525e-01
-6.09778106e-01 -1.16536760e+00 2.00917631e-01 5.98085225e-01
8.58637929e-01 -7.61439085e-01 -8.39030683e-01 2.28428662e-01
-1.37010574e-01 -9.78997171e-01 -3.31090599e-01 -1.99570041e-03
-8.30128014e-01 -3.61411393e-01 -7.07214117e-01 -3.14559400e-01
3.60673457e-01 8.70494187e-01 1.18521476e+00 -4.24598577e-04
-5.81927419e-01 -4.22169715e-02 -2.11875811e-01 -2.30393574e-01
3.94434690e-01 1.44767970e-01 -2.25980535e-01 3.54298428e-02
2.03167886e-01 -1.02416229e+00 -1.37360084e+00 4.75592732e-01
-7.37761796e-01 5.85896432e-01 7.14250445e-01 7.85976112e-01
8.97270560e-01 9.99259800e-02 3.80498379e-01 -5.65845907e-01
1.76538467e-01 -6.63012385e-01 -6.13039553e-01 -7.42903352e-02
-6.00487649e-01 -2.82407761e-01 4.81230617e-01 -2.80706733e-01
-1.43622863e+00 -5.35739243e-01 -1.58506200e-01 -2.08527297e-01
-1.50461048e-01 1.38771743e-01 -2.14219138e-01 -5.16776294e-02
6.22274578e-01 3.77528310e-01 -3.29182118e-01 -5.58872581e-01
1.28094196e-01 5.64781308e-01 2.76909113e-01 -4.14219230e-01
6.40162051e-01 9.09231603e-01 -2.21770152e-01 -1.14042044e+00
-7.74527252e-01 -5.80322027e-01 -2.92530656e-01 -1.35329977e-01
8.38489354e-01 -1.43217719e+00 -6.42959952e-01 4.75475073e-01
-8.58185232e-01 -5.90297639e-01 -7.84962475e-02 2.93166280e-01
-1.85742855e-01 5.49791530e-02 -7.35421538e-01 -8.91534507e-01
-5.50743341e-01 -1.07030416e+00 1.21121824e+00 6.37194514e-01
9.31655988e-02 -5.86666584e-01 -4.08385634e-01 4.09972191e-01
8.84837508e-01 -2.89460480e-01 5.98723769e-01 9.83688235e-02
-1.02658129e+00 4.85553563e-01 -1.12134957e+00 5.89699335e-02
-2.55176872e-01 -2.68622577e-01 -1.21433890e+00 -3.03729951e-01
-6.72805011e-02 -3.98935452e-02 1.02493989e+00 8.96218956e-01
1.61012077e+00 -4.42797691e-02 -2.75100648e-01 1.11721241e+00
1.63206327e+00 2.99325283e-03 8.55798900e-01 2.87298858e-01
6.97087526e-01 3.08950394e-01 4.73689586e-01 7.94248462e-01
3.81991893e-01 8.18446159e-01 4.89716142e-01 -6.89855695e-01
-5.18282771e-01 1.81561336e-01 -2.06817519e-02 4.09181923e-01
-4.83961374e-01 -3.35573517e-02 -7.09161460e-01 3.20586622e-01
-1.66505086e+00 -9.27588105e-01 -1.72327906e-01 2.18335843e+00
7.65161455e-01 1.53374463e-01 1.41364649e-01 -2.27385417e-01
4.27237988e-01 4.14970100e-01 -8.82445991e-01 -1.10282041e-01
-5.01812816e-01 2.72700667e-01 8.10451627e-01 6.13599956e-01
-1.06401229e+00 8.25219214e-01 5.26179934e+00 1.19358599e+00
-1.19521666e+00 3.71539921e-01 1.00403607e+00 -4.82485771e-01
-5.05363643e-01 -3.24208856e-01 -1.27950323e+00 5.55963576e-01
7.66451359e-01 1.32313117e-01 5.41662633e-01 5.24042606e-01
4.20415968e-01 -2.83858269e-01 -3.48536998e-01 1.04659688e+00
-1.70250341e-01 -1.47805858e+00 -2.07674578e-01 4.17448759e-01
6.79297507e-01 4.74049538e-01 3.69617999e-01 2.51568735e-01
2.24974364e-01 -1.01160789e+00 5.75894058e-01 1.61464930e-01
1.21284926e+00 -7.41524935e-01 2.97301888e-01 2.86106765e-02
-1.60795271e+00 -3.64951044e-01 -3.84006053e-01 7.47283967e-03
-9.96942911e-03 7.21555114e-01 9.34852362e-02 5.88595867e-01
1.38392413e+00 6.70807421e-01 -5.28991878e-01 7.27554440e-01
3.68582249e-01 6.45721793e-01 -4.33651924e-01 3.34383547e-01
2.02307209e-01 -2.42214084e-01 7.23230183e-01 1.32708025e+00
2.39864439e-01 4.28753227e-01 -1.79983929e-01 1.10351956e+00
2.61389613e-01 -4.21145447e-02 -4.42895815e-02 7.45197535e-01
6.39479578e-01 1.29478300e+00 -7.41607964e-01 -3.35253030e-01
-7.12833345e-01 7.97110379e-01 3.24801534e-01 8.27633142e-01
-1.15502274e+00 -1.06432542e-01 1.02353883e+00 5.52563131e-01
3.88957739e-01 -1.81224033e-01 -8.45557928e-01 -1.09950531e+00
1.47752166e-01 -6.94204330e-01 2.51340687e-01 -6.16948068e-01
-8.20637286e-01 8.90214026e-01 -1.23338357e-01 -1.08405101e+00
6.01610839e-01 -1.65116176e-01 -5.62091112e-01 1.15716434e+00
-1.91820335e+00 -1.21088481e+00 -6.04944348e-01 6.61164224e-01
6.76396430e-01 2.68885553e-01 4.70061630e-01 5.93604445e-01
-7.71376431e-01 7.65943170e-01 1.99865505e-01 -2.65356898e-01
5.10534108e-01 -1.04148269e+00 4.32019591e-01 8.89995992e-01
-3.59907895e-01 6.48868740e-01 5.65451324e-01 -4.51786667e-01
-1.26410532e+00 -9.90112185e-01 3.20938796e-01 -2.01572310e-02
6.95559382e-01 -5.16309738e-01 -1.08740592e+00 2.07419813e-01
-3.60107236e-02 1.96799219e-01 3.86133730e-01 3.77753645e-01
-4.01450187e-01 -3.75078857e-01 -8.62011492e-01 8.64341378e-01
1.48058534e+00 -2.78107822e-01 3.77880037e-01 3.46493721e-02
8.55016708e-01 -3.28302711e-01 -9.72859859e-01 4.33538169e-01
5.09999871e-01 -1.45264375e+00 1.34910965e+00 4.98552471e-02
6.06857300e-01 -2.97778457e-01 -1.22656226e-01 -7.77509630e-01
-7.71565616e-01 -7.79502630e-01 -3.57591301e-01 1.13985384e+00
4.83150661e-01 -7.20972836e-01 7.60355532e-01 5.57034850e-01
-2.44696259e-01 -1.19609368e+00 -1.03284216e+00 -5.91853440e-01
-1.37153596e-01 -3.45119536e-01 6.74262822e-01 5.31206131e-01
-3.06720227e-01 3.21753323e-01 -4.20204848e-01 5.29417038e-01
1.08678639e+00 1.08735472e-01 4.67301846e-01 -9.60188210e-01
-3.76392782e-01 -6.95037127e-01 -5.47767058e-03 -1.53256357e+00
-4.58281606e-01 -3.48710865e-01 -3.22395504e-01 -1.32464862e+00
3.77975315e-01 -8.02056015e-01 -5.93214929e-01 6.11497238e-02
-3.12974572e-01 4.08623457e-01 -4.24243994e-02 4.58707422e-01
-6.87904716e-01 4.66449916e-01 1.38484478e+00 2.54346520e-01
-4.52930719e-01 -2.23751783e-01 -9.79760349e-01 5.99165320e-01
9.28035975e-01 1.47076026e-01 -5.98567903e-01 -9.32303786e-01
1.25639409e-01 9.52912197e-02 4.84507561e-01 -1.18520284e+00
2.71170050e-01 -4.36777808e-02 4.29292053e-01 -6.64716125e-01
4.76641119e-01 -4.49990004e-01 -1.46424174e-01 9.98862274e-03
-2.80989766e-01 -2.92539269e-01 3.64659905e-01 5.50997913e-01
-1.86251268e-01 6.27189517e-01 1.16909087e+00 -1.13117255e-01
-8.05570781e-01 6.28743887e-01 9.02508758e-03 9.73702073e-02
6.62147045e-01 -5.82179427e-01 -4.76698071e-01 -9.45831239e-02
-5.71389258e-01 4.71995980e-01 4.15585369e-01 2.14992404e-01
6.24395251e-01 -8.81600797e-01 -8.88429523e-01 3.44646841e-01
-2.23312024e-02 3.04864615e-01 1.33590925e+00 1.03457475e+00
-1.16734058e-01 2.74954438e-01 -1.61692634e-01 -6.23102069e-01
-1.27024972e+00 2.67707646e-01 1.76905066e-01 -3.46301854e-01
-1.01537907e+00 1.32658887e+00 9.11077380e-01 3.04808378e-01
2.93305546e-01 -2.22699955e-01 -3.41492057e-01 -9.63874683e-02
1.14890528e+00 4.25477773e-01 -1.79123595e-01 -5.23136795e-01
-2.07507074e-01 6.62081242e-01 -3.01368773e-01 3.39796469e-02
1.34015322e+00 -7.78244257e-01 -1.01371855e-01 6.65124804e-02
9.44954693e-01 2.59301588e-02 -1.69391370e+00 -4.75284427e-01
-5.94686508e-01 -5.57536423e-01 6.22982979e-01 -7.16931701e-01
-1.07363522e+00 8.87803733e-01 8.62831295e-01 7.30704069e-02
1.54719698e+00 7.31770471e-02 1.02357233e+00 -4.53812957e-01
3.51438284e-01 -1.13707888e+00 -3.03362161e-01 3.49276245e-01
8.54292095e-01 -1.26981068e+00 3.50821540e-02 -7.16074467e-01
-4.81827706e-01 5.33007681e-01 9.79829371e-01 9.03850701e-03
8.96912277e-01 7.03979611e-01 -2.60950834e-01 -1.82315364e-01
-7.79625058e-01 -2.00967133e-01 1.21869355e-01 4.61854428e-01
4.63078558e-01 1.68930441e-01 -4.05889452e-02 6.44825399e-01
2.25299999e-01 -8.90501067e-02 2.46499419e-01 5.01817524e-01
-5.25490224e-01 -4.94261771e-01 -1.61288619e-01 5.52322328e-01
-4.21276689e-01 -6.64352655e-01 4.19991612e-01 3.99374098e-01
2.30974734e-01 8.26389551e-01 3.56710583e-01 -3.10353875e-01
5.90103269e-02 -6.74443841e-01 1.34421378e-01 -4.63346899e-01
-4.21837419e-01 5.08417130e-01 -7.46747628e-02 -9.69940662e-01
-1.90799683e-01 -4.00326967e-01 -1.20716226e+00 -3.78128171e-01
2.85178516e-03 -1.40945315e-01 5.31418979e-01 3.45928878e-01
8.43065977e-01 7.72851825e-01 5.51544309e-01 -1.02571893e+00
-2.61081576e-01 -9.13045764e-01 -7.92351127e-01 -1.56545751e-02
2.86435306e-01 -6.01630390e-01 -4.42975163e-01 -2.76809335e-01] | [10.869632720947266, -1.9115865230560303] |
9008af95-4efc-4a8a-aeec-eadea8289404 | robust-model-predictive-techno-economic | 2305.03272 | null | https://arxiv.org/abs/2305.03272v1 | https://arxiv.org/pdf/2305.03272v1.pdf | Robust Model Predictive Techno-Economic Control of Active Distribution Networks | Stochastic controllers are perceived as a promising solution for techno-economic operation of distribution networks having higher generation uncertainties at large penetration of renewables. These controllers are supported by forecasters capable of predicting generation uncertainty by means of lower/upper bounds rather than by probability density function (PDF). Hence, the stochastic controller assumes a suitable PDF for scenario creation and optimization, requiring validation of the assumption. To effectively bridge the forecaster's capability and resolve the assumption issues, the paper proposes a robust model prediction-based techno-economic controller, which essentially utilizes only the lower/upper bounds of the forecast, eliminating the necessity of PDF. Both discrete and continuous control resources such as tap-changers and DERs are utilized for regulating the lower/upper bounds of the network states and robustly minimizing the cost of energy import. The proposed controller is implemented for UKGDS network and validated by comparing performance at various confidence levels of lower/upper bound forecast. | ['Zhaoyu Wang', 'Rui Cheng', 'Prashant Tiwari', 'Salish Maharjan'] | 2023-05-05 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-3.93475533e-01 1.04108982e-01 -1.73606411e-01 1.28036201e-01
-1.75346240e-01 -9.72610056e-01 4.98745680e-01 2.19500467e-01
2.60942847e-01 1.37845802e+00 -1.87886998e-01 -4.91418451e-01
-7.10191429e-01 -1.08531141e+00 -1.96599424e-01 -7.40186214e-01
6.81035295e-02 4.65143710e-01 -2.70931218e-02 -2.12046757e-01
3.14459205e-02 6.30463064e-01 -1.42706299e+00 -3.98624033e-01
9.69974041e-01 1.19087422e+00 1.74169376e-01 3.65550518e-01
-6.10995404e-02 2.13433430e-01 -8.81005943e-01 2.58310467e-01
1.64224684e-01 -2.12033629e-01 1.54604495e-01 -3.15817744e-01
-7.17054367e-01 -2.83127159e-01 2.96664894e-01 1.11236787e+00
5.27777076e-01 2.82478899e-01 1.05169606e+00 -1.56814420e+00
-5.27285635e-01 5.03464937e-01 -4.11661029e-01 3.19199950e-01
7.60407373e-02 9.02363136e-02 3.98494333e-01 -5.22418201e-01
7.32509196e-02 9.20029342e-01 3.23124230e-01 2.03031693e-02
-9.78083432e-01 -6.90446556e-01 1.93767011e-01 -8.82340446e-02
-1.59334767e+00 -3.95991467e-02 7.45371401e-01 -4.74683762e-01
1.17752826e+00 2.81339318e-01 6.57073796e-01 2.22517014e-01
4.99531001e-01 -9.85562801e-03 1.05107081e+00 -5.38572729e-01
4.94525909e-01 3.06621879e-01 -1.82848811e-01 7.28320405e-02
6.11529350e-01 1.48352981e-01 8.98615643e-02 -2.29547173e-02
6.93650723e-01 -3.91234040e-01 -4.79896158e-01 -4.04795706e-02
-5.68959057e-01 6.82815254e-01 -1.73965678e-01 3.31492871e-01
-8.62284720e-01 -3.51092778e-02 1.09034322e-01 4.92388792e-02
3.53899002e-01 -7.86570385e-02 -5.71415365e-01 1.27906129e-02
-1.04628336e+00 -3.25791053e-02 6.89754963e-01 1.24157953e+00
-2.23061264e-01 1.22007287e+00 -2.13575497e-01 1.46826789e-01
5.77099144e-01 1.13219225e+00 3.70049566e-01 -5.96876562e-01
2.16161549e-01 2.17736289e-01 8.22102547e-01 -7.66534090e-01
-3.37581724e-01 -9.82501328e-01 -8.07006299e-01 3.71188790e-01
2.24753156e-01 -7.12652206e-01 -6.01401508e-01 1.57130909e+00
2.60519177e-01 -1.77976619e-02 -1.55070022e-01 4.72508550e-01
-4.16097015e-01 9.44269657e-01 -1.89354852e-01 -9.59847927e-01
9.90157783e-01 -1.80412993e-01 -1.01126492e+00 7.58681953e-01
-2.09064960e-01 -6.39114559e-01 4.30391908e-01 4.76033300e-01
-1.16123950e+00 -3.93698066e-01 -1.03473151e+00 9.64290917e-01
-7.63455391e-01 3.02684575e-01 -9.34334695e-02 1.01983190e+00
-9.10220861e-01 5.01990378e-01 -6.81064129e-01 -8.18744674e-02
-2.48223826e-01 8.79935026e-02 3.53611082e-01 6.26292765e-01
-1.45916402e+00 1.31447971e+00 7.17239380e-01 7.23102152e-01
-4.80165243e-01 -7.24979222e-01 -4.85057682e-01 5.41185915e-01
6.89974949e-02 -4.69587088e-01 8.51809680e-01 -5.47796547e-01
-1.77366483e+00 -5.93437195e-01 3.08637828e-01 -7.91475236e-01
5.05685031e-01 2.88479775e-01 -9.69488859e-01 1.42031804e-01
-3.41680020e-01 -2.11585596e-01 8.81209612e-01 -1.00383818e+00
-7.28971958e-01 4.58256416e-02 -3.43255311e-01 1.44957140e-01
-8.13770071e-02 -2.26872757e-01 7.91017532e-01 -5.76962471e-01
-1.07402228e-01 -6.21413887e-01 -1.23089552e-01 -4.54528481e-01
-3.33861917e-01 -1.79263592e-01 1.00029349e+00 -7.96581686e-01
1.27332973e+00 -1.43130648e+00 -3.68070483e-01 7.83863842e-01
-4.86596346e-01 4.36435848e-01 4.52161163e-01 8.29920053e-01
-3.44706684e-01 2.80759186e-01 4.01431732e-02 4.33679491e-01
4.24301952e-01 1.68973580e-01 -5.06033480e-01 4.22655702e-01
3.21818382e-01 6.62654862e-02 -7.26833522e-01 2.32777372e-01
8.32874477e-01 4.72736031e-01 1.45017773e-01 -2.47190639e-01
-4.32719380e-01 1.85114995e-01 -6.36282265e-01 2.54326969e-01
8.47487926e-01 4.19281155e-01 2.13865340e-01 -1.39332071e-01
-5.25046945e-01 -3.41215342e-01 -1.71883786e+00 5.86106956e-01
-8.22771490e-01 2.03327596e-01 3.98538828e-01 -1.02299035e+00
1.04914105e+00 5.35549760e-01 4.37083572e-01 -2.66395122e-01
2.25722585e-02 2.09342763e-01 1.90376732e-02 3.40776443e-02
1.15104161e-01 -3.86117131e-01 4.35910821e-01 3.37595731e-01
5.35713583e-02 -2.93852001e-01 3.65302175e-01 -2.63173491e-01
3.07889640e-01 2.61757616e-02 4.23284769e-01 -8.94209683e-01
8.98226023e-01 -4.59707260e-01 8.05229485e-01 6.48763254e-02
-1.59037352e-01 -7.50295296e-02 3.62646252e-01 2.12512314e-01
-7.74615467e-01 -1.09547675e+00 -9.50839221e-02 1.42001912e-01
-1.64867505e-01 2.65089780e-01 -1.47366211e-01 -1.43834606e-01
2.85725653e-01 1.78529894e+00 -1.70274779e-01 -2.14447424e-01
-3.07754073e-02 -7.83166885e-01 4.77456637e-02 3.46517771e-01
3.23099136e-01 -2.00533092e-01 -6.35691106e-01 5.55859864e-01
4.12929654e-01 -6.29040301e-01 -4.87457477e-02 2.69011766e-01
-5.13959229e-01 -9.80622292e-01 -6.34831727e-01 1.16407573e-01
6.40797555e-01 -1.52408779e-01 8.21982861e-01 -1.32701933e-01
2.09375933e-01 5.04126549e-01 -1.66049544e-02 -7.65233517e-01
-2.91944563e-01 -3.19352597e-01 4.00426537e-01 -2.15490341e-01
-4.90358233e-01 -7.87096262e-01 -4.10742611e-01 2.20729113e-01
-5.06415129e-01 -2.10778832e-01 1.14859559e-01 5.35301864e-01
5.03207743e-01 1.17729461e+00 1.61844254e+00 -4.12279725e-01
9.12560046e-01 -5.30020595e-01 -1.42579556e+00 6.07412875e-01
-1.13070810e+00 -3.90599035e-02 1.13942659e+00 1.98598090e-03
-1.47819078e+00 -2.32171655e-01 3.12961906e-01 -3.93333822e-01
-3.50316428e-02 5.83663821e-01 -3.63803715e-01 1.43801883e-01
-1.97689712e-01 1.95219159e-01 -1.97333023e-01 -3.69223505e-01
1.60873190e-01 4.53365713e-01 2.64043421e-01 -5.84706843e-01
1.05956948e+00 -7.91411474e-02 6.40230954e-01 -5.91875792e-01
-1.78572893e-01 -2.99316701e-02 -3.04237157e-01 -6.53140247e-01
3.04396272e-01 -7.43212223e-01 -1.01603854e+00 3.44832271e-01
-8.57085466e-01 2.42790744e-01 -2.48405278e-01 7.73394585e-01
-5.10318995e-01 -8.85736570e-02 -2.52107769e-01 -1.65478837e+00
-4.74802107e-01 -7.61616707e-01 1.12944983e-01 5.72661877e-01
2.04357505e-01 -1.20367670e+00 -3.09252650e-01 -3.61678004e-01
9.20231283e-01 8.83635700e-01 1.10581231e+00 -4.35760885e-01
-3.23752940e-01 -3.05245429e-01 4.14349176e-02 6.45291924e-01
1.77226126e-01 6.84908569e-01 -4.14693713e-01 -4.05683756e-01
1.72821999e-01 4.80854839e-01 -1.67753920e-01 4.70167011e-01
5.79576015e-01 -3.94689083e-01 -1.35608450e-01 -1.63253069e-01
2.06126046e+00 7.88945436e-01 3.26419860e-01 -6.98157176e-02
-4.43155132e-02 4.06874985e-01 7.77866125e-01 9.32356834e-01
2.50618607e-01 1.64616078e-01 5.90643346e-01 5.80742776e-01
4.15415764e-01 -1.87430501e-01 1.82806075e-01 6.07132494e-01
-1.11806057e-01 -6.49006307e-01 -7.31248796e-01 4.33116913e-01
-1.57878435e+00 -8.51036608e-01 -1.30574927e-01 2.35485148e+00
2.55297869e-01 4.52836484e-01 -3.67217474e-02 4.23187405e-01
1.24469674e+00 -1.49353266e-01 -4.68557924e-01 -8.29519212e-01
-5.73613830e-02 9.80093852e-02 9.30031955e-01 4.38716561e-01
-3.43497992e-01 -5.69541864e-02 5.23558617e+00 1.08067977e+00
-1.09583676e+00 -3.07006806e-01 5.02306163e-01 -5.78413457e-02
-4.12376970e-01 -1.07482970e-01 -6.30868971e-01 9.92881298e-01
1.51027358e+00 -9.27254498e-01 3.82773638e-01 7.02556133e-01
1.29014981e+00 -5.95977902e-01 -4.68351126e-01 2.01855421e-01
-5.94433844e-01 -1.21416044e+00 2.67409980e-02 -2.85893958e-02
1.00687730e+00 -4.28944558e-01 4.85389829e-02 3.43824387e-01
1.71465516e-01 -6.07848465e-01 8.51747990e-01 1.06174457e+00
3.71074587e-01 -1.39791679e+00 1.15231550e+00 6.16066992e-01
-1.32721698e+00 -2.96158046e-01 -1.37947530e-01 -2.63995510e-02
6.44538820e-01 9.00114238e-01 -6.03135467e-01 1.08669055e+00
2.54535466e-01 -1.20480089e-02 1.11819305e-01 8.51521015e-01
-1.12829603e-01 6.86977088e-01 -6.75077915e-01 -1.12081570e-05
7.12861866e-02 -5.75252056e-01 5.32614648e-01 7.70982087e-01
8.10750186e-01 2.68421054e-01 1.31035179e-01 7.99810410e-01
4.95404392e-01 -1.59150138e-01 -2.82956123e-01 -2.45476171e-01
1.11938918e+00 8.95711899e-01 -7.04635322e-01 -2.09816456e-01
-1.20466709e-01 -6.38162196e-02 -5.17477930e-01 6.66523159e-01
-1.09842527e+00 -5.20445406e-01 1.86579227e-01 -7.58332759e-03
2.16281906e-01 -3.25782597e-01 -5.83725810e-01 -4.88576978e-01
-6.15883432e-02 -2.88176924e-01 1.95750073e-01 -7.34277248e-01
-1.06620705e+00 1.41699538e-01 2.90942430e-01 -1.33548486e+00
-7.48116791e-01 -3.40661854e-01 -1.09525263e+00 1.51645076e+00
-1.63557780e+00 -9.14354324e-01 3.31951588e-01 3.96142483e-01
3.11578840e-01 -2.30026171e-01 7.36416161e-01 1.40064552e-01
-9.63499486e-01 -4.03963439e-02 7.75225580e-01 -4.95414495e-01
-1.59673148e-03 -1.20383894e+00 -6.24001920e-01 1.02382386e+00
-7.59764194e-01 4.35082197e-01 1.18607461e+00 -5.49105167e-01
-1.23535287e+00 -8.75581324e-01 4.73250389e-01 5.01095653e-01
9.23768044e-01 7.87409395e-02 -4.79255170e-01 2.53656477e-01
6.04432404e-01 -2.32389018e-01 3.43225598e-01 -5.81775367e-01
3.81381005e-01 -1.83030531e-01 -1.76324081e+00 2.36952364e-01
-3.43177095e-02 -2.36719534e-01 -2.20861033e-01 1.50558233e-01
3.17535013e-01 -1.78864021e-02 -1.30484819e+00 4.01926130e-01
3.64853054e-01 -6.24409854e-01 6.11154735e-01 -1.85273573e-01
-4.28780764e-01 -5.80310524e-01 -3.14737707e-01 -1.90516460e+00
7.63188004e-02 -8.73814404e-01 -4.84593153e-01 1.75177801e+00
4.68088776e-01 -1.05684578e+00 2.09401488e-01 7.65866518e-01
-1.55668512e-01 -6.72057629e-01 -1.31962454e+00 -1.10988081e+00
1.87908456e-01 -1.31651178e-01 8.46851647e-01 7.66829014e-01
1.22550152e-01 -2.39832819e-01 6.84888288e-02 6.56148016e-01
8.49227607e-01 5.33343144e-02 -6.48827627e-02 -9.18570697e-01
4.98527624e-02 -4.00464803e-01 7.32755139e-02 2.22565994e-01
3.56566869e-02 -1.91980109e-01 -1.91894993e-01 -1.69779789e+00
-8.25916588e-01 -2.08853245e-01 -5.91645896e-01 -1.30573781e-02
3.54523629e-01 -5.26503980e-01 2.00939372e-01 -1.71662450e-01
3.10743332e-01 9.13074195e-01 7.97682047e-01 1.35799468e-01
2.14474872e-02 4.92255926e-01 2.49193478e-02 8.27724636e-01
1.21356833e+00 8.30650609e-03 -1.15129578e+00 1.74277946e-01
9.39957276e-02 6.68897688e-01 1.06397159e-02 -1.14889967e+00
1.98802546e-01 -5.30607581e-01 4.35418367e-01 -1.11731064e+00
-4.38774116e-02 -1.46346855e+00 8.51792276e-01 6.21098936e-01
4.97688800e-02 3.63398403e-01 2.02368259e-01 8.50510418e-01
-1.10076845e-01 -1.74899206e-01 8.39662850e-01 7.24261105e-02
-3.48069340e-01 -6.46570250e-02 -7.55168498e-01 -3.90402049e-01
1.37664306e+00 -2.06747249e-01 -5.67411005e-01 -6.05190217e-01
-9.52936172e-01 7.85404384e-01 7.14532435e-02 7.92038143e-02
2.40473017e-01 -9.57308292e-01 -3.21174920e-01 -9.51963738e-02
-7.31247663e-01 -4.54826832e-01 3.70246738e-01 4.65185016e-01
-3.38435680e-01 1.01756036e+00 -2.88438112e-01 -8.02601948e-02
-3.01241964e-01 5.84923506e-01 7.34524071e-01 -2.06824899e-01
1.78137615e-01 1.66535378e-01 -7.01600730e-01 -3.06154490e-02
-6.24684095e-02 -3.99754256e-01 -1.81922004e-01 4.53649640e-01
2.51983613e-01 9.45657730e-01 1.74335435e-01 -3.65363240e-01
-2.78919458e-01 5.72445579e-02 5.80425620e-01 -1.62688568e-01
1.20923638e+00 -5.99620640e-01 2.13237833e-02 3.69425774e-01
4.94863123e-01 -1.18555330e-01 -1.26876450e+00 4.01559949e-01
1.40000880e-01 -5.68639934e-02 3.19409490e-01 -1.23757398e+00
-1.24392068e+00 3.70837212e-01 5.14360130e-01 6.96405709e-01
1.11499059e+00 -1.17086852e+00 3.74799818e-01 -1.84323743e-01
7.57174551e-01 -1.53015327e+00 -9.53727305e-01 3.25506657e-01
9.30073440e-01 -5.21346986e-01 -1.38823790e-02 -3.45202595e-01
-3.54156047e-01 1.18331563e+00 3.89900416e-01 -4.03800756e-01
1.14192474e+00 5.80098808e-01 -5.74158132e-01 4.52648431e-01
-1.01946032e+00 -1.92396492e-02 3.52484062e-02 8.50562394e-01
2.83639878e-01 4.68021631e-01 -8.44533443e-01 9.22951221e-01
7.62028918e-02 1.32296398e-01 8.74549329e-01 7.84269035e-01
-4.98774618e-01 -6.12660944e-01 -5.43907344e-01 4.07457799e-01
-4.02183205e-01 1.31542668e-01 5.79795301e-01 1.01956952e+00
3.74807000e-01 1.19606256e+00 5.86127341e-02 2.49082014e-01
5.76630533e-01 2.18378678e-01 1.42254606e-01 -2.41203293e-01
-3.36735487e-01 1.77858666e-01 4.16651636e-01 7.35185901e-03
-5.67647442e-03 -6.67991459e-01 -1.44443071e+00 -3.72052103e-01
-7.62348592e-01 7.35114098e-01 9.63194788e-01 1.00757062e+00
2.19515443e-01 6.86887562e-01 1.10037601e+00 -5.67123890e-01
-1.13340461e+00 -8.79407346e-01 -1.11439574e+00 -6.76294088e-01
-2.12223485e-01 -9.11718845e-01 -5.70760131e-01 -3.86175364e-01] | [5.665369510650635, 2.5491013526916504] |
39a7b7e0-513d-4053-a9b7-cdb910ba1284 | qpp-real-time-quantization-parameter | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Kryzhanovskiy_QPP_Real-Time_Quantization_Parameter_Prediction_for_Deep_Neural_Networks_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Kryzhanovskiy_QPP_Real-Time_Quantization_Parameter_Prediction_for_Deep_Neural_Networks_CVPR_2021_paper.pdf | QPP: Real-Time Quantization Parameter Prediction for Deep Neural Networks | Modern deep neural networks (DNNs) cannot be effectively used in mobile and embedded devices due to strict requirements for computational complexity, memory, and power consumption. The quantization of weights and feature maps (activations) is a popular approach to solve this problem. Training-aware quantization often shows excellent results but requires a full dataset, which is not always available. Post-training quantization methods, in turn, are applied without fine-tuning but still work well for many classes of tasks like classification, segmentation, and so on. However, they either imply a big overhead for quantization parameters (QPs) calculation at runtime (dynamic methods) or lead to an accuracy drop if pre-computed static QPs are used (static methods). Moreover, most inference frameworks don't support dynamic quantization. Thus we propose a novel quantization approach called QPP: quantization parameter prediction. With a small subset of a training dataset or unlabeled data from the same domain, we find the predictor that can accurately estimate QPs of activations given only the NN's input data. Such a predictor allows us to avoid complex calculation of precise values of QPs while maintaining the quality of the model. To illustrate our method's efficiency, we added QPP into two dynamic approaches: 1) Dense+Sparse quantization, where the predetermined percentage of activations are not quantized, 2) standard quantization with equal quantization steps. We provide experiments on a wide set of tasks including super-resolution, facial landmark, segmentation, and classification. | ['Aleksandr Zuruev', 'Nikolay Kozyrskiy', 'Gleb Balitskiy', 'Vladimir Kryzhanovskiy'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['parameter-prediction'] | ['miscellaneous'] | [ 2.64700532e-01 -9.53946635e-02 -3.91630143e-01 -5.56113780e-01
-7.51961708e-01 -2.41912484e-01 2.64970094e-01 7.72779360e-02
-6.79954112e-01 6.81728542e-01 -2.97080368e-01 -1.88270286e-01
9.35883820e-02 -1.00899506e+00 -6.57482028e-01 -7.67093897e-01
3.47220838e-01 3.66149247e-01 5.78667402e-01 6.06625602e-02
2.04304948e-01 3.79415870e-01 -1.80064881e+00 1.72591850e-01
1.01253390e+00 1.49937725e+00 3.83075327e-01 3.96463305e-01
-4.06529337e-01 2.86319226e-01 -7.23359287e-01 -4.18193102e-01
3.17675114e-01 -1.66678369e-01 -5.63680887e-01 1.34605737e-02
4.22322243e-01 -5.38301885e-01 -7.62389600e-02 1.50223315e+00
5.70141852e-01 1.52569413e-01 4.89239246e-01 -1.26874363e+00
-3.79202336e-01 6.26408160e-01 -4.04758960e-01 2.08480552e-01
-1.23490043e-01 -8.40211287e-03 7.75475323e-01 -7.73760259e-01
3.48490655e-01 1.20955193e+00 6.00661159e-01 6.95101440e-01
-1.17599261e+00 -9.79632616e-01 2.10478783e-01 2.15443641e-01
-1.70393026e+00 -6.51974976e-01 5.85663974e-01 -7.70778209e-02
8.64918351e-01 8.53562355e-02 6.37122929e-01 7.72242248e-01
-1.15525909e-01 6.60087526e-01 6.85522377e-01 -2.49070942e-01
7.77762532e-01 1.42686382e-01 6.35370146e-03 5.94454110e-01
6.74324334e-02 -2.61556506e-01 -6.24463916e-01 -1.31912500e-01
9.21437740e-01 1.21821180e-01 -3.51745456e-01 -3.79664190e-02
-9.20622051e-01 7.99017310e-01 3.76822174e-01 2.42763206e-01
-4.32151675e-01 3.49269927e-01 3.21050107e-01 3.55515808e-01
2.03763440e-01 7.91669190e-02 -5.31673253e-01 -3.30635339e-01
-1.16262805e+00 1.56674668e-01 5.14484704e-01 9.25199986e-01
1.13786066e+00 1.76706627e-01 6.59282412e-03 1.08882725e+00
1.46191031e-01 3.72100055e-01 9.69650328e-01 -1.04740798e+00
6.52255952e-01 5.84482849e-01 2.72735860e-02 -1.05596614e+00
-2.88862646e-01 -1.73074707e-01 -1.02533805e+00 2.17355136e-02
4.54580605e-01 -2.25334495e-01 -1.15759265e+00 1.85052717e+00
3.70888174e-01 2.56715953e-01 -1.32554412e-01 8.95306110e-01
6.71402991e-01 7.97373056e-01 -3.19396630e-02 -2.56530046e-01
1.18068409e+00 -7.10150123e-01 -8.17748129e-01 -2.33248040e-01
6.39299393e-01 -3.22708607e-01 1.18373036e+00 5.77174246e-01
-9.40539300e-01 -6.26988709e-01 -1.13995540e+00 -5.39551042e-02
-4.09522235e-01 2.44343936e-01 5.69745719e-01 8.38849425e-01
-1.18315125e+00 8.70136857e-01 -1.03024650e+00 1.05782468e-02
5.63398302e-01 8.95144045e-01 -1.06432825e-01 3.97157781e-02
-1.19574153e+00 5.06374121e-01 5.25771618e-01 1.57461524e-01
-5.55550754e-01 -5.22036135e-01 -6.75447822e-01 2.52800375e-01
2.90626198e-01 -2.95843124e-01 1.25667822e+00 -1.30031753e+00
-1.65841341e+00 3.50836098e-01 -3.34066957e-01 -4.11905080e-01
2.07249999e-01 -1.54167712e-01 -2.38120213e-01 8.71003568e-02
-8.95496905e-02 1.02978671e+00 1.03913760e+00 -8.07770550e-01
-7.64689565e-01 -4.28861648e-01 1.23519987e-01 1.81125030e-01
-9.40751791e-01 -4.07168955e-01 -7.70457804e-01 -4.82583076e-01
4.15150046e-01 -7.27774143e-01 -3.04790378e-01 3.87741238e-01
-2.39122659e-01 -3.35242957e-01 7.09635615e-01 -3.85752887e-01
1.35809505e+00 -2.26623082e+00 -1.76483199e-01 3.76357049e-01
1.19095795e-01 3.93805832e-01 -2.36977041e-02 -2.64059722e-01
1.71867952e-01 2.15648487e-01 -2.16862097e-01 -4.15569931e-01
2.00244822e-02 4.67937469e-01 -1.59924492e-01 1.29637823e-01
1.00836143e-01 7.07216382e-01 -7.66251564e-01 -6.27752423e-01
2.29597270e-01 5.77726960e-01 -7.99932063e-01 -9.94978920e-02
-2.20934004e-01 4.15065326e-02 -3.44871968e-01 6.84883237e-01
6.04337752e-01 -3.63098741e-01 2.00507954e-01 -4.70388263e-01
1.14018902e-01 2.72742420e-01 -1.58109438e+00 1.59249282e+00
-4.76875752e-01 4.44926739e-01 -2.85374909e-03 -9.65565324e-01
9.06291246e-01 2.99885273e-01 2.68844336e-01 -8.43313038e-01
1.81485713e-01 3.13251495e-01 -8.43619332e-02 -1.83650032e-01
4.45975572e-01 7.04105720e-02 2.14659393e-01 2.22102702e-01
1.20792545e-01 2.00159654e-01 1.76730171e-01 -2.43265569e-01
1.02543294e+00 -2.51133233e-01 1.77260235e-01 9.65660717e-03
3.20418596e-01 -3.04081887e-01 9.36891496e-01 4.96197343e-01
-2.91093498e-01 6.09873354e-01 5.17842770e-01 -3.93539220e-01
-9.46611404e-01 -7.18876064e-01 -2.81489015e-01 1.19426823e+00
7.66162649e-02 -4.62863564e-01 -1.00048423e+00 -4.58920181e-01
-1.14036843e-01 3.69680136e-01 -4.05098081e-01 -7.54843131e-02
-4.44471389e-01 -9.29263532e-01 5.22589982e-01 5.84912062e-01
7.47263432e-01 -9.33779895e-01 -7.65057206e-01 2.11712182e-01
-8.75365064e-02 -1.22859335e+00 -3.99904966e-01 4.19179738e-01
-1.21927297e+00 -6.86793625e-01 -6.22546971e-01 -6.72493935e-01
7.66551018e-01 -2.56762393e-02 6.44113719e-01 1.23373449e-01
1.35249034e-01 -2.63052374e-01 -1.44979298e-01 -1.53823644e-01
-5.21572791e-02 2.49615982e-01 1.93311572e-01 4.69971076e-02
3.48021567e-01 -7.49680221e-01 -6.55659676e-01 2.76062012e-01
-9.32738245e-01 -6.49408102e-02 6.35624528e-01 8.83110344e-01
9.80554879e-01 2.22755432e-01 6.87051415e-01 -8.56271386e-01
5.60409307e-01 -2.22963154e-01 -8.30959737e-01 9.98897478e-02
-6.99145973e-01 1.97909474e-01 9.23064470e-01 -7.93839633e-01
-5.16644001e-01 1.97975621e-01 -4.36682492e-01 -8.23370159e-01
4.43464523e-05 3.76269996e-01 -3.59637737e-01 -1.98466495e-01
6.66061640e-01 1.57287821e-01 -1.46697432e-01 -4.49811339e-01
9.49045494e-02 8.38971198e-01 4.57612306e-01 -2.72402197e-01
4.34108973e-01 1.76642478e-01 -2.25371301e-01 -7.40911365e-01
-5.83926916e-01 -1.77028671e-01 -5.25644004e-01 1.23562507e-01
5.06602645e-01 -7.46733785e-01 -7.44962215e-01 2.46732444e-01
-1.00975752e+00 -5.01697540e-01 -2.01504067e-01 3.01504165e-01
-2.46221796e-01 2.38150686e-01 -4.33453351e-01 -7.60811806e-01
-5.32445669e-01 -1.41871631e+00 1.01716149e+00 2.87312746e-01
-4.45959643e-02 -7.98614800e-01 -4.62620676e-01 -2.44076829e-03
4.16285992e-01 1.11101694e-01 8.86720061e-01 -3.35907817e-01
-4.95589435e-01 -8.57385173e-02 -2.60413140e-01 4.63535488e-01
2.72410303e-01 -1.04012720e-01 -1.11155927e+00 -1.85857773e-01
-2.49152347e-01 -3.63478035e-01 7.17492044e-01 3.74105692e-01
1.79213095e+00 -4.78848159e-01 -2.55557656e-01 8.42553616e-01
1.35517764e+00 2.28679493e-01 5.42641759e-01 9.38685238e-02
7.25068867e-01 6.86662123e-02 5.46542168e-01 5.44619918e-01
1.77571997e-01 7.82240450e-01 4.82782841e-01 1.56095922e-01
-9.26365424e-03 -2.01991394e-01 2.99510032e-01 8.09296131e-01
-1.72669701e-02 -1.06298797e-01 -8.47438574e-01 4.75954592e-01
-1.65731859e+00 -6.40233338e-01 3.40647250e-01 2.39465833e+00
1.14661372e+00 3.82784069e-01 -8.29363465e-02 6.51550412e-01
6.06193006e-01 -1.33190647e-01 -9.12441015e-01 -4.06642199e-01
2.93486625e-01 3.72702032e-01 6.93089843e-01 3.38210285e-01
-9.80313003e-01 9.88437831e-01 6.21260929e+00 1.20354581e+00
-1.49741304e+00 1.34855255e-01 7.63962150e-01 -1.22612283e-01
-1.84178203e-01 -3.42175424e-01 -1.23776770e+00 7.57262826e-01
1.20738935e+00 1.43301308e-01 6.13985837e-01 1.12677252e+00
4.98183817e-02 -1.46886066e-01 -1.13978457e+00 1.53156888e+00
-2.56510139e-01 -1.36425519e+00 1.56824201e-01 -4.89446111e-02
5.80459893e-01 3.29962112e-02 9.86781046e-02 3.27212572e-01
-1.71013232e-02 -9.31539357e-01 7.35222280e-01 1.07473925e-01
1.02568340e+00 -8.68234277e-01 8.34125936e-01 5.78910410e-01
-1.07462108e+00 -2.34366864e-01 -7.02897251e-01 2.01173872e-01
-1.50709584e-01 7.56733119e-01 -7.34260798e-01 -5.07202186e-02
7.80673265e-01 4.51090932e-01 -4.33853149e-01 8.26042712e-01
-7.41864890e-02 5.82805812e-01 -6.41673625e-01 -1.61624894e-01
2.10064769e-01 -2.36825589e-02 -7.55878463e-02 9.57069814e-01
3.56528193e-01 2.01734632e-01 -2.04824433e-02 7.52133369e-01
-3.00882876e-01 6.48420230e-02 -1.38999715e-01 5.14170304e-02
9.93247151e-01 1.08519554e+00 -9.23654079e-01 -4.34187263e-01
-3.35354000e-01 1.01027536e+00 3.16082805e-01 3.20446104e-01
-7.72165895e-01 -4.78877664e-01 7.03230917e-01 8.84679481e-02
4.43156153e-01 -7.70306140e-02 -4.03902501e-01 -9.74982738e-01
4.29132469e-02 -8.49107862e-01 4.95212488e-02 -4.31190103e-01
-8.60868216e-01 5.43483675e-01 -1.98455781e-01 -1.12390029e+00
-3.84699225e-01 -5.80230832e-01 -1.73914135e-01 7.21258581e-01
-1.62409139e+00 -4.91571546e-01 -3.83533031e-01 6.99946404e-01
5.41549683e-01 4.71953787e-02 8.92781019e-01 6.28762960e-01
-6.68273211e-01 1.08421075e+00 1.39354005e-01 2.45395243e-01
4.95414048e-01 -1.09051478e+00 2.37134904e-01 5.17431617e-01
8.93240422e-02 5.30659676e-01 4.56666887e-01 -3.03769618e-01
-1.34545696e+00 -1.21299756e+00 7.71776676e-01 8.19121301e-02
3.12850028e-01 -3.78925860e-01 -1.01979053e+00 4.31804031e-01
-4.30646449e-01 4.36005563e-01 5.93865633e-01 -4.34329128e-03
-4.69062589e-02 -5.69764137e-01 -1.25913179e+00 5.79797447e-01
8.15549254e-01 -4.15572643e-01 -5.37823252e-02 1.59978941e-01
7.11630285e-01 -5.19260883e-01 -9.73648727e-01 2.95969486e-01
6.14508867e-01 -8.73172641e-01 8.28068972e-01 9.04316455e-03
1.56139001e-01 -2.47048005e-01 -2.86543101e-01 -1.05333471e+00
-1.10471666e-01 -3.41728866e-01 -3.58362257e-01 1.05772376e+00
4.06742901e-01 -6.32550895e-01 1.03285122e+00 8.48889649e-01
4.33607660e-02 -1.05891287e+00 -1.10596383e+00 -6.24651730e-01
-2.85412550e-01 -5.73091626e-01 8.29845726e-01 7.75889754e-01
-1.70094103e-01 1.32008746e-01 -7.47105107e-02 2.63021976e-01
4.67632383e-01 -8.49647671e-02 5.29046953e-01 -1.15085161e+00
-2.13920459e-01 -3.54867727e-01 -6.70818865e-01 -1.26993632e+00
-2.05983757e-03 -5.81063211e-01 1.15329884e-01 -1.34536433e+00
-2.82362878e-01 -7.91543841e-01 -2.88201720e-01 9.00370121e-01
7.75348535e-03 3.50410968e-01 -6.59208670e-02 2.95409769e-01
-5.45858502e-01 4.95346427e-01 1.05980074e+00 -7.14154989e-02
-4.66775656e-01 5.63376173e-02 -4.52340573e-01 8.10712755e-01
7.94424176e-01 -4.97651309e-01 -7.07156122e-01 -6.20367050e-01
1.44942462e-01 8.10583085e-02 1.62920266e-01 -1.20506012e+00
4.45331573e-01 -1.24821030e-01 5.40348232e-01 -5.37832141e-01
4.88608867e-01 -7.90305257e-01 -3.65085863e-02 3.32533985e-01
-2.79321104e-01 2.29110382e-02 1.36713997e-01 4.37852025e-01
-2.92242289e-01 -3.48747760e-01 9.14820433e-01 8.61907844e-03
-8.23211670e-01 5.53907216e-01 -1.81243926e-01 -1.20221011e-01
6.73440516e-01 -5.13953447e-01 9.41250771e-02 -1.85566083e-01
-7.59247065e-01 1.38652295e-01 3.46620739e-01 1.55453026e-01
8.25454354e-01 -1.36395860e+00 -8.96023512e-02 3.58569741e-01
-2.71823674e-01 4.00435418e-01 -7.50972256e-02 5.60767949e-01
-3.92521501e-01 2.64848083e-01 -1.00581139e-01 -6.60977244e-01
-1.04838204e+00 3.87903810e-01 3.67738158e-01 -6.16512224e-02
-5.38363278e-01 8.32652807e-01 1.76729858e-02 -2.03794435e-01
6.29945636e-01 -7.62504041e-01 -2.45529965e-01 8.48823190e-02
7.50543892e-01 1.54858023e-01 3.23530465e-01 -3.57836217e-01
-2.26366818e-01 5.66903770e-01 -1.33238686e-02 -1.48607671e-01
1.08873940e+00 -1.10896468e-01 1.31761223e-01 5.83587885e-01
1.16889000e+00 -5.23076952e-01 -1.39016390e+00 -3.38798851e-01
-1.16370991e-01 -3.70476365e-01 4.25157428e-01 -3.28928649e-01
-1.54863441e+00 1.12837422e+00 9.02949095e-01 2.43635792e-02
1.38416278e+00 -4.24419791e-01 1.03450096e+00 6.52413845e-01
5.63564241e-01 -1.39362943e+00 -1.36118203e-01 3.25260580e-01
4.23757643e-01 -1.10752606e+00 -1.60035133e-01 -3.49278450e-01
-4.82012272e-01 1.22627449e+00 7.00113356e-01 3.11154388e-02
8.47942531e-01 3.89905781e-01 -1.26924649e-01 2.47498319e-01
-8.34240913e-01 -6.63453117e-02 7.71174580e-02 4.75402385e-01
1.98541552e-01 -4.61938865e-02 -8.95342007e-02 6.72982514e-01
-3.86810571e-01 2.61298925e-01 1.18710406e-01 8.69430780e-01
-7.65319586e-01 -9.52572644e-01 -2.62602091e-01 6.34081841e-01
-4.07099724e-01 -1.71058938e-01 1.11835696e-01 3.77608269e-01
4.73111987e-01 6.62625432e-01 2.81220615e-01 -5.92661083e-01
1.80295497e-01 1.10820487e-01 2.60829449e-01 -5.61978102e-01
-3.42973918e-01 5.12609072e-02 -3.24751884e-01 -6.80193365e-01
-2.64691710e-01 -4.72694844e-01 -1.58469856e+00 -3.66099298e-01
-4.23240274e-01 4.33505885e-02 8.02893162e-01 9.49676394e-01
2.80173302e-01 4.43814993e-01 4.40407395e-01 -7.32532799e-01
-6.93453610e-01 -8.72842968e-01 -5.72301567e-01 6.58728927e-02
3.06261599e-01 -6.92434728e-01 -1.83608577e-01 2.23170206e-01] | [8.601255416870117, 3.112410545349121] |
5655ba17-5669-49d4-8844-01c90926f3a6 | modern-hopfield-networks-for-few-and-zero | 2104.03279 | null | https://arxiv.org/abs/2104.03279v3 | https://arxiv.org/pdf/2104.03279v3.pdf | Modern Hopfield Networks for Few- and Zero-Shot Reaction Template Prediction | Finding synthesis routes for molecules of interest is an essential step in the discovery of new drugs and materials. To find such routes, computer-assisted synthesis planning (CASP) methods are employed which rely on a model of chemical reactivity. In this study, we model single-step retrosynthesis in a template-based approach using modern Hopfield networks (MHNs). We adapt MHNs to associate different modalities, reaction templates and molecules, which allows the model to leverage structural information about reaction templates. This approach significantly improves the performance of template relevance prediction, especially for templates with few or zero training examples. With inference speed several times faster than that of baseline methods, we improve predictive performance for top-k exact match accuracy for $\mathrm{k}\geq5$ in the retrosynthesis benchmark USPTO-50k. | ['Günter Klambauer', 'Sepp Hochreiter', 'Jörg K. Wegner', 'Marwin Segler', 'Jonas Verhoeven', 'Paulo Neves', 'Natalia Dyubankova', 'Philipp Renz', 'Philipp Seidl'] | 2021-04-07 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 4.04129803e-01 2.18807664e-02 -4.94469881e-01 -5.65856993e-02
-5.68336606e-01 -9.76685226e-01 8.30304027e-01 4.33811098e-01
-4.57238376e-01 1.06380558e+00 2.14194730e-01 -5.26606321e-01
-2.18940288e-01 -8.73460293e-01 -7.77300894e-01 -7.78460562e-01
9.48041081e-02 3.84280622e-01 6.67620972e-02 -3.36781710e-01
7.43695974e-01 9.39838052e-01 -8.68870139e-01 4.59823787e-01
5.95101833e-01 9.93392944e-01 2.64550924e-01 2.67863959e-01
5.94063364e-02 5.28235674e-01 -1.08846843e-01 -6.26088977e-01
2.12475434e-01 -1.28960088e-01 -1.06211519e+00 -8.20821762e-01
1.39307886e-01 -2.27787867e-02 -4.53455508e-01 8.63732636e-01
6.61231339e-01 5.35600185e-01 9.87319231e-01 -6.86516881e-01
-1.40621498e-01 7.77915776e-01 2.12304473e-01 2.19896972e-01
3.49831760e-01 4.10551965e-01 9.85535145e-01 -1.28722668e+00
8.74903619e-01 9.02538836e-01 5.69357574e-01 5.26734054e-01
-1.24457490e+00 -1.04534376e+00 -3.11459862e-02 3.34490508e-01
-1.42189133e+00 -6.70570910e-01 3.34090680e-01 -3.25901926e-01
1.77742279e+00 -4.83810976e-02 6.69824958e-01 1.21641088e+00
7.21384346e-01 2.39206403e-01 6.47498250e-01 -1.52580604e-01
5.86837530e-01 -3.57662886e-01 -2.09463477e-01 6.88073456e-01
4.53547621e-03 5.56944609e-01 -7.53755271e-01 4.85646650e-02
6.39143467e-01 4.64256592e-02 -2.50252753e-01 6.82714805e-02
-1.35248327e+00 9.49617743e-01 8.38880062e-01 2.60538161e-01
-5.22581935e-01 2.17305794e-01 1.78840905e-01 -1.58251449e-01
1.37373298e-01 1.42587745e+00 -5.15203238e-01 3.13092947e-01
-9.66088772e-01 3.94341648e-01 8.28922451e-01 1.07120359e+00
6.04861319e-01 3.46926153e-02 -1.96465045e-01 1.87194183e-01
1.70746949e-02 5.80669828e-02 8.62145275e-02 -9.42440033e-01
4.49680954e-01 4.41058218e-01 7.10043758e-02 -5.47306716e-01
-8.50753903e-01 -3.43107849e-01 -7.56219923e-01 -8.58834386e-02
3.31140548e-01 1.58693150e-01 -1.16278279e+00 1.51843929e+00
2.78064787e-01 -2.53467232e-01 2.47315913e-01 5.28582752e-01
9.13253069e-01 9.76347089e-01 4.15930152e-01 -4.76447791e-01
1.01815963e+00 -1.06067574e+00 -5.50167561e-01 2.17604950e-01
6.31233692e-01 -7.84996867e-01 3.48354071e-01 8.35315883e-01
-1.09334135e+00 -1.96752191e-01 -1.20828521e+00 -6.31599575e-02
-1.02219796e+00 -2.06373304e-01 7.17904508e-01 4.29306746e-01
-7.98152924e-01 1.54297602e+00 -4.26765889e-01 -1.00812037e-02
6.18784428e-01 1.00022686e+00 -4.55351651e-01 -7.03725740e-02
-1.37626457e+00 1.18845510e+00 1.02362561e+00 5.87303452e-02
-1.38084722e+00 -1.02748346e+00 -6.48299873e-01 2.07852319e-01
8.35213244e-01 -5.40479839e-01 1.25031102e+00 -4.59252506e-01
-1.80815411e+00 1.31352216e-01 -4.22467962e-02 -5.31566620e-01
-2.89054029e-02 2.85906851e-01 -4.33202565e-01 3.06205630e-01
-1.06114082e-01 1.16872001e+00 6.65169299e-01 -6.41257167e-01
-7.99679235e-02 -3.66527513e-02 1.62263028e-02 1.20376088e-01
3.07912648e-01 6.35911003e-02 1.32671967e-02 -4.51537281e-01
-5.26507646e-02 -8.11643302e-01 -6.73807502e-01 -2.99369127e-01
-5.57338595e-01 -2.92922616e-01 -2.06648916e-01 -5.64584374e-01
9.71532643e-01 -1.44235098e+00 3.73490870e-01 6.65635049e-01
2.88350731e-01 3.00040156e-01 -1.91290632e-01 8.86889696e-01
-5.03140628e-01 2.66706407e-01 -1.41604794e-02 4.33200300e-01
-1.45699427e-01 -2.57728815e-01 -1.92694589e-01 1.46664396e-01
1.98772207e-01 1.12813628e+00 -8.51131141e-01 -8.47342163e-02
-1.73999164e-02 4.77360964e-01 -6.86753511e-01 -1.04415826e-01
-7.72075295e-01 3.86676550e-01 -4.24409091e-01 7.39860833e-01
1.00699276e-01 -4.90075976e-01 3.33523661e-01 -5.96993566e-01
-4.79048729e-01 6.21872485e-01 -8.19720328e-01 1.91821361e+00
-2.82373965e-01 -1.13439545e-01 -7.46149242e-01 -6.72109306e-01
7.01115489e-01 6.50158465e-01 2.96607345e-01 -8.16763401e-01
2.35473856e-01 2.63171524e-01 3.60320181e-01 6.58618510e-02
3.60243469e-01 -3.02349061e-01 2.15832531e-01 2.55038887e-01
3.74914229e-01 -1.42473638e-01 4.32555795e-01 1.84881777e-01
1.04029942e+00 3.18988383e-01 5.95759511e-01 -3.72411668e-01
4.27524656e-01 -1.77099034e-02 4.34861094e-01 6.24932349e-01
2.10491180e-01 1.44597918e-01 3.21985573e-01 -8.52668047e-01
-1.27348328e+00 -6.88448370e-01 6.74052387e-02 8.59636426e-01
-2.39926502e-01 -8.12699437e-01 -7.78325737e-01 -5.33761680e-01
-4.00882900e-01 7.99437046e-01 -5.88858962e-01 -3.53188187e-01
-4.19327825e-01 -6.41961694e-01 5.62235951e-01 5.70968091e-01
1.21731482e-01 -1.36850607e+00 1.43027548e-02 8.05534542e-01
1.61552504e-01 -9.04646814e-01 -5.50245166e-01 7.26501882e-01
-7.35354543e-01 -1.15907741e+00 -5.78352273e-01 -3.75930071e-01
5.00369847e-01 -2.85724401e-02 7.71109283e-01 -2.55878299e-01
-3.40638667e-01 -3.56651157e-01 4.66199256e-02 -5.55401146e-01
-6.12522662e-01 3.34591091e-01 1.84168369e-01 -5.90172648e-01
3.46824666e-03 -4.98997271e-01 -8.86022568e-01 2.26584345e-01
-6.83307290e-01 3.86904590e-02 5.74779809e-01 8.04704964e-01
7.39281178e-01 -8.10912694e-04 7.92145133e-01 -5.28720319e-01
2.46809185e-01 -2.81738788e-01 -1.08596551e+00 4.12839383e-01
-9.42408085e-01 5.23064375e-01 9.54633057e-01 -3.63762349e-01
-8.06646824e-01 4.73349065e-01 -1.66180059e-01 -2.78440535e-01
6.57674298e-02 6.13452256e-01 -3.73594582e-01 -4.13293660e-01
7.25234449e-01 2.54428208e-01 -2.88273394e-01 -2.86507368e-01
4.88027573e-01 -4.93133292e-02 7.39452168e-02 -7.53546357e-01
3.60140860e-01 2.41241623e-02 5.95263898e-01 -4.17098552e-01
-4.71719921e-01 -1.86987579e-01 -5.29071987e-01 1.12543695e-01
8.46567631e-01 -7.51690447e-01 -1.52777767e+00 3.07255350e-02
-1.30215657e+00 -4.95902598e-01 2.39953965e-01 6.24989450e-01
-7.12688327e-01 1.61164865e-01 -7.27509081e-01 -2.88184226e-01
-7.63754189e-01 -1.63353550e+00 6.95675433e-01 7.59077370e-02
-1.87241122e-01 -6.72406793e-01 3.37299071e-02 2.30288908e-01
2.81883001e-01 -7.89494738e-02 1.33347332e+00 -1.06240010e+00
-1.07926726e+00 -4.50969823e-02 -1.26338139e-01 -1.54408440e-01
6.04389496e-02 -2.20265940e-01 -8.89562130e-01 4.90270406e-02
-6.25024140e-01 -3.08231920e-01 1.04539943e+00 3.67282242e-01
1.35512292e+00 -6.21110916e-01 -5.44732988e-01 3.31032932e-01
1.29851556e+00 8.19232225e-01 7.86384702e-01 2.12577432e-01
8.00639987e-01 5.92500389e-01 4.06057298e-01 2.10838631e-01
-1.59115404e-01 5.51665425e-01 5.15792310e-01 2.38138855e-01
1.77439764e-01 -3.43373388e-01 1.82609305e-01 2.63990343e-01
-5.00395536e-01 -4.54535037e-01 -9.97429430e-01 7.42109632e-03
-1.47244203e+00 -1.16974664e+00 3.14014703e-01 2.24551678e+00
1.15052557e+00 1.70083404e-01 1.11776508e-01 2.09701601e-02
5.30111015e-01 -2.32777998e-01 -7.17387497e-01 -4.05809730e-01
1.29993707e-01 1.13136983e+00 7.46595740e-01 5.41247547e-01
-9.14001822e-01 1.45089877e+00 6.86013699e+00 1.22553205e+00
-1.11747336e+00 -3.57804686e-01 6.78557634e-01 -9.21774879e-02
-3.80460583e-02 2.01144829e-01 -1.01697755e+00 2.17260048e-01
1.29903579e+00 2.24568211e-02 9.13571119e-01 4.98893917e-01
3.28018248e-01 1.55308679e-01 -1.57785690e+00 6.57829881e-01
-3.25229675e-01 -2.40780091e+00 2.56960005e-01 8.97068605e-02
6.74011707e-01 -2.11972371e-01 -2.05436964e-02 1.07468683e-02
2.89434910e-01 -1.50678992e+00 5.30416369e-01 4.39086556e-01
7.19108284e-01 -1.13836157e+00 2.01073512e-01 -2.81490088e-02
-1.04860961e+00 -8.12126324e-02 -2.73262084e-01 3.30848157e-01
-9.49657783e-02 2.63143003e-01 -1.30652237e+00 7.51979887e-01
1.30438596e-01 7.19883800e-01 -8.10470954e-02 9.45584059e-01
-2.01427892e-01 2.54218906e-01 -3.36906433e-01 -2.40909263e-01
4.08249348e-01 -1.69786170e-01 1.11816019e-01 9.28374171e-01
3.70737672e-01 4.08813477e-01 1.94887340e-01 9.83856022e-01
-4.17721331e-01 2.27834299e-01 -6.15390599e-01 -4.94726330e-01
5.62978148e-01 1.11247003e+00 -1.05396104e+00 -3.09848160e-01
6.08983159e-04 6.31934285e-01 -3.20321023e-02 3.91754389e-01
-7.57890522e-01 -3.30160856e-01 2.92531610e-01 -1.10311553e-01
5.27964771e-01 -3.78534868e-02 1.83921099e-01 -5.89403033e-01
-4.96117234e-01 -1.15037262e+00 1.73723578e-01 -7.26947665e-01
-7.61193752e-01 4.81305629e-01 -6.93395287e-02 -8.20143104e-01
8.04115459e-02 -9.17972207e-01 -4.50696439e-01 7.43895650e-01
-1.29818320e+00 -1.04229009e+00 4.77890968e-01 4.34923887e-01
5.34246564e-01 -2.88915634e-01 1.06315672e+00 2.07113713e-01
-6.39394760e-01 4.69123483e-01 6.33621663e-02 -3.59277576e-01
6.41093731e-01 -9.65136051e-01 5.61109424e-01 3.83592546e-01
2.95205344e-03 8.80631208e-01 6.99653983e-01 -7.21281290e-01
-1.36919272e+00 -1.13545823e+00 7.82541394e-01 -2.01084733e-01
5.52408516e-01 -3.76753122e-01 -4.57170129e-01 2.24317685e-01
1.89733490e-01 -2.87094742e-01 7.80529737e-01 -2.56524265e-01
-6.34194136e-01 9.49935764e-02 -1.04125059e+00 7.83115387e-01
1.15053356e+00 -4.36478108e-01 -1.37284756e-01 7.22625852e-01
9.37883496e-01 -4.02390242e-01 -1.39667106e+00 2.90380955e-01
6.59527898e-01 -4.12958860e-01 1.32135522e+00 -8.89832795e-01
5.92846215e-01 -2.22934857e-01 -1.15204379e-01 -1.25081348e+00
-4.28668171e-01 -9.51361418e-01 4.84817252e-02 3.96795034e-01
9.03598905e-01 -4.14893657e-01 6.66091502e-01 5.51132083e-01
-4.71375048e-01 -8.02494764e-01 -7.75750756e-01 -7.10905254e-01
1.54326767e-01 -3.68216783e-02 8.68964612e-01 8.55375111e-01
2.88240314e-01 6.04987800e-01 -1.39543116e-01 -2.96445005e-02
1.65568233e-01 4.09684982e-03 -6.31181970e-02 -1.15788543e+00
-3.37120056e-01 -6.05401635e-01 2.70440262e-02 -5.09334087e-01
2.03015462e-01 -1.21993494e+00 -1.19031727e-01 -1.18399787e+00
2.27693945e-01 -2.97838837e-01 -4.02227253e-01 8.08452427e-01
2.61097968e-01 -2.37875450e-02 -2.13619098e-02 -5.63280806e-02
-6.28712177e-01 5.43393970e-01 1.16905248e+00 -3.96046042e-01
-2.62976646e-01 -1.96095645e-01 -5.91660917e-01 3.92113328e-01
9.29160893e-01 -6.50941193e-01 -4.22976047e-01 2.57537127e-01
8.37499738e-01 4.13818002e-01 8.80256295e-02 -8.37714255e-01
3.78233939e-01 -4.31192219e-01 6.40511632e-01 -7.55837977e-01
4.93242770e-01 -4.47116852e-01 3.91085088e-01 7.46437490e-01
-6.74945474e-01 -9.65892226e-02 1.71580806e-01 6.90698564e-01
3.87376368e-01 -2.46530846e-01 4.22068745e-01 -4.78591561e-01
-2.66558290e-01 8.44438672e-01 -6.41889274e-01 -6.20612025e-01
6.50585651e-01 -1.34982586e-01 -3.91011924e-01 -2.69178599e-02
-8.99535418e-01 -4.07103859e-02 1.67944670e-01 5.69894947e-02
6.45403564e-01 -1.04898477e+00 -1.25557899e-01 -3.04167509e-01
3.02611589e-01 1.07416287e-01 6.53865794e-03 7.81981707e-01
-5.74767828e-01 1.04010427e+00 -1.84617653e-01 5.94391711e-02
-9.00521159e-01 1.08518767e+00 5.34763157e-01 -4.00950372e-01
-9.48685706e-02 7.50060678e-01 2.99694866e-01 -1.75552025e-01
2.90125519e-01 -4.00682896e-01 -2.73299217e-01 2.57240087e-02
5.41014016e-01 2.90738881e-01 6.00033462e-01 -1.88244030e-01
-3.28073412e-01 2.99163252e-01 -6.43558502e-01 1.77998245e-01
1.32705975e+00 7.19808757e-01 1.16885766e-01 -3.00454766e-01
9.13796008e-01 -4.18509156e-01 -1.12852001e+00 -5.69674186e-02
1.21242933e-01 1.40374959e-01 1.76827252e-01 -1.31951070e+00
-7.97642112e-01 8.73169661e-01 4.04252738e-01 -5.47223210e-01
6.05350018e-01 -2.84362465e-01 6.31203771e-01 1.17370498e+00
2.94481099e-01 -1.11108840e+00 2.67620951e-01 4.94008213e-01
8.70904803e-01 -9.93722618e-01 2.50380278e-01 -3.10930312e-01
-2.28942424e-01 1.39456308e+00 1.65123567e-01 1.79579407e-01
3.53876621e-01 -2.49460101e-01 -7.02455699e-01 -4.42304254e-01
-7.48943806e-01 -2.30441131e-02 2.72687584e-01 2.84102827e-01
4.35470998e-01 -9.56958681e-02 -3.35171700e-01 5.25863647e-01
1.21705517e-01 -6.48790300e-02 2.93845713e-01 8.75637770e-01
-3.52685899e-01 -1.55046880e+00 -3.16086970e-02 1.43801481e-01
-6.35870099e-01 -9.20727968e-01 -6.93502247e-01 5.59168577e-01
3.65578569e-03 7.45109856e-01 -5.81467569e-01 -2.98585206e-01
4.40802462e-02 3.71157497e-01 9.85149503e-01 -3.84099722e-01
-9.24200356e-01 1.64751232e-01 2.78015643e-01 -7.39497125e-01
-2.73393810e-01 -1.09579630e-01 -1.49050486e+00 -3.35881650e-01
-5.47515213e-01 2.24677354e-01 7.78155565e-01 1.07942927e+00
5.76225638e-01 3.14001858e-01 5.51611602e-01 -1.03158331e+00
-3.74079376e-01 -4.70760435e-01 -4.17358242e-02 -3.84271860e-01
-3.27487215e-02 -6.34897530e-01 2.64138401e-01 -3.42794359e-02] | [4.50202751159668, 6.0998029708862305] |
9c2b46dd-b5b6-4818-8fee-38f52a2594e9 | opa-3d-occlusion-aware-pixel-wise-aggregation | 2211.01142 | null | https://arxiv.org/abs/2211.01142v1 | https://arxiv.org/pdf/2211.01142v1.pdf | OPA-3D: Occlusion-Aware Pixel-Wise Aggregation for Monocular 3D Object Detection | Despite monocular 3D object detection having recently made a significant leap forward thanks to the use of pre-trained depth estimators for pseudo-LiDAR recovery, such two-stage methods typically suffer from overfitting and are incapable of explicitly encapsulating the geometric relation between depth and object bounding box. To overcome this limitation, we instead propose OPA-3D, a single-stage, end-to-end, Occlusion-Aware Pixel-Wise Aggregation network that to jointly estimate dense scene depth with depth-bounding box residuals and object bounding boxes, allowing a two-stream detection of 3D objects, leading to significantly more robust detections. Thereby, the geometry stream denoted as the Geometry Stream, combines visible depth and depth-bounding box residuals to recover the object bounding box via explicit occlusion-aware optimization. In addition, a bounding box based geometry projection scheme is employed in an effort to enhance distance perception. The second stream, named as the Context Stream, directly regresses 3D object location and size. This novel two-stream representation further enables us to enforce cross-stream consistency terms which aligns the outputs of both streams, improving the overall performance. Extensive experiments on the public benchmark demonstrate that OPA-3D outperforms state-of-the-art methods on the main Car category, whilst keeping a real-time inference speed. We plan to release all codes and trained models soon. | ['Federico Tombari', 'Didier Stricker', 'Benjamin Busam', 'Jason Rambach', 'Guangyao Zhai', 'Fabian Manhardt', 'Yan Di', 'Yongzhi Su'] | 2022-11-02 | null | null | null | null | ['monocular-3d-object-detection'] | ['computer-vision'] | [ 1.63715601e-01 -1.55411422e-01 -8.52706358e-02 -4.53132749e-01
-9.46856856e-01 -4.27753896e-01 7.05796361e-01 2.41082534e-01
-5.18935442e-01 2.38100260e-01 -1.66556507e-01 -3.33502918e-01
2.51081854e-01 -8.85160863e-01 -9.81616080e-01 -6.71079934e-01
5.41152805e-02 5.73185802e-01 6.78837657e-01 2.00303048e-01
3.41574758e-01 8.75221729e-01 -1.86949778e+00 5.60911074e-02
7.23871231e-01 1.41340411e+00 2.44527668e-01 6.06990755e-01
-1.71658292e-01 4.76034284e-01 -1.64953172e-01 -3.42831075e-01
6.46445632e-01 1.12553172e-01 -9.53873992e-02 3.08761388e-01
1.07683873e+00 -8.97462666e-01 -2.15039298e-01 8.97815824e-01
4.77956027e-01 1.62943408e-01 5.72007775e-01 -1.05660295e+00
-1.44729733e-01 -1.64552942e-01 -8.06070626e-01 -2.97737792e-02
2.07119793e-01 3.96062702e-01 9.81162965e-01 -1.29796672e+00
5.16482532e-01 1.45963502e+00 7.03664601e-01 2.57480711e-01
-1.39108324e+00 -7.05831349e-01 4.35504913e-01 -1.47895087e-02
-1.50913239e+00 -3.52861077e-01 8.37370932e-01 -3.23311210e-01
1.19767725e+00 9.47586894e-02 5.43198586e-01 6.02870643e-01
-5.40652983e-02 8.71797383e-01 8.64900649e-01 -7.66382441e-02
2.27315322e-01 1.47058994e-01 -2.18494609e-01 6.77688181e-01
3.91354620e-01 3.50193530e-01 -5.06243229e-01 3.37684639e-02
7.25551248e-01 1.14412196e-01 1.83297340e-02 -1.05760336e+00
-9.69783604e-01 7.36466467e-01 8.78356278e-01 -3.20995331e-01
-3.48521739e-01 1.32614464e-01 2.70135552e-01 -1.83510989e-01
7.44609058e-01 -4.17070128e-02 -3.07767242e-01 1.90697446e-01
-9.70361590e-01 7.04046965e-01 3.91128182e-01 9.90645826e-01
9.95026708e-01 -2.89475679e-01 -1.43725157e-01 7.31344283e-01
5.46408713e-01 7.59485006e-01 -2.17784449e-01 -1.06386757e+00
7.63658762e-01 9.14632499e-01 2.13557482e-01 -7.45855093e-01
-3.28093559e-01 -4.09399569e-01 -4.61056441e-01 6.48990273e-01
4.85944271e-01 3.43479335e-01 -1.04325366e+00 1.36213255e+00
9.17929113e-01 2.78669477e-01 -1.81107819e-01 1.09166980e+00
6.65662587e-01 4.59897965e-01 3.54051739e-02 1.93558961e-01
1.41157830e+00 -8.97889733e-01 3.86734307e-03 -6.29791141e-01
6.26031518e-01 -6.90909445e-01 7.45394230e-01 3.23913127e-01
-1.16898704e+00 -6.11054242e-01 -1.20992899e+00 -5.53628623e-01
-2.73200542e-01 2.29145020e-01 5.38019836e-01 5.03668249e-01
-6.93450153e-01 2.20508948e-01 -8.54520202e-01 -1.72226771e-03
8.35411906e-01 2.34092057e-01 -2.54887283e-01 -3.33133548e-01
-3.88005674e-01 8.95357966e-01 3.73780489e-01 1.04442544e-01
-8.64582241e-01 -9.88878846e-01 -1.06796539e+00 -1.24321930e-01
6.27467573e-01 -8.33938658e-01 1.10741591e+00 -2.22083583e-01
-1.20620573e+00 1.07894397e+00 -4.34238791e-01 -6.32901549e-01
7.24619627e-01 -2.62386441e-01 1.40284389e-01 -2.79576294e-02
2.79205233e-01 1.10477448e+00 9.63816524e-01 -1.28940582e+00
-1.19002330e+00 -7.74130940e-01 -1.60457537e-01 3.09754163e-01
1.05995506e-01 -4.43084151e-01 -8.76001239e-01 -2.13891640e-01
4.02313620e-01 -7.98722267e-01 -2.36273423e-01 5.45976877e-01
-3.23383749e-01 -2.39476576e-01 9.29216325e-01 -2.31754959e-01
8.54344487e-01 -2.32347798e+00 -5.77041879e-02 6.09119907e-02
2.70378292e-01 1.94642216e-01 -1.08300205e-02 -5.81911765e-03
2.46006578e-01 -3.42048079e-01 -4.84271437e-01 -1.09077561e+00
3.49563621e-02 2.80716598e-01 -3.59476298e-01 7.01923966e-01
7.97026753e-01 7.60877073e-01 -8.64041865e-01 -3.93477470e-01
7.62658060e-01 7.23731160e-01 -7.30066478e-01 1.31082997e-01
-4.04395401e-01 2.92390376e-01 -3.42917651e-01 9.64235127e-01
1.34395003e+00 -4.19480689e-02 -2.96187907e-01 1.78572163e-02
-3.27584922e-01 5.22759795e-01 -1.25289071e+00 1.73970342e+00
-4.65164036e-01 6.11088514e-01 2.43083715e-01 -5.56405306e-01
1.05067503e+00 -2.92632341e-01 3.47969025e-01 -8.61224115e-01
1.89952314e-01 3.20133954e-01 -2.92641997e-01 -1.30489647e-01
7.64304996e-01 -1.85478069e-02 1.28059104e-01 -3.21843550e-02
-2.26751730e-01 -6.33440495e-01 1.01546131e-01 2.61550486e-01
8.60424757e-01 5.21554470e-01 1.95130825e-01 2.43512958e-01
6.26438737e-01 1.44188879e-02 3.47297937e-01 5.51805854e-01
-2.65463531e-01 8.71570885e-01 2.41431862e-01 -3.08672339e-01
-1.14539397e+00 -1.17168653e+00 -4.14717436e-01 6.11816883e-01
4.08431351e-01 -2.10328624e-01 -4.83528405e-01 -6.30150020e-01
6.33093059e-01 7.25542009e-01 -4.63409692e-01 9.78564918e-02
-7.19520032e-01 -3.82343739e-01 1.87213823e-01 7.27759123e-01
4.19814080e-01 -5.97554147e-01 -9.32256937e-01 2.27292493e-01
1.12694778e-01 -1.46213865e+00 -3.21236521e-01 4.85857606e-01
-9.81084168e-01 -9.84531283e-01 -5.05043089e-01 -3.12200010e-01
5.75789630e-01 7.05209851e-01 9.71233964e-01 -5.39105898e-03
-4.22147393e-01 3.30189266e-03 -6.10637181e-02 -4.98633087e-01
-1.12960018e-01 -1.01961121e-01 -3.54606777e-01 -7.28413090e-02
5.29211819e-01 -4.96451408e-01 -8.43330324e-01 2.76997477e-01
-6.69793129e-01 2.12518107e-02 6.51553333e-01 4.71533418e-01
9.10117507e-01 -2.90266246e-01 2.66257357e-02 -6.27716959e-01
-1.37939826e-01 -2.70223826e-01 -1.21050465e+00 -4.33279514e-01
-6.29269660e-01 -4.23701592e-02 2.35010788e-01 -2.33581945e-01
-8.56574059e-01 6.67810559e-01 -2.27945641e-01 -7.02576101e-01
-2.19245881e-01 -1.68754086e-01 -3.17969352e-01 -5.57151400e-02
5.63837886e-01 1.38329521e-01 -1.40632335e-02 -4.15474951e-01
6.09955013e-01 4.75585788e-01 9.02104497e-01 -2.38384515e-01
1.18258142e+00 9.51753199e-01 1.82173178e-01 -7.12146997e-01
-9.43930089e-01 -1.00800014e+00 -6.73595726e-01 -1.77897573e-01
8.75270963e-01 -1.41624141e+00 -6.41694009e-01 4.38482732e-01
-1.29018056e+00 -1.96046814e-01 -3.24235529e-01 3.85538250e-01
-3.89471829e-01 2.43056625e-01 -2.29304507e-01 -1.08330762e+00
-5.91382496e-02 -1.04776669e+00 1.78989410e+00 -1.90767029e-03
2.31295377e-01 -4.77411628e-01 -2.15847582e-01 4.72622275e-01
2.59187240e-02 3.21599692e-01 5.35376132e-01 -1.05886921e-01
-1.32526028e+00 -4.35286015e-01 -7.53132164e-01 3.98207337e-01
-2.58747071e-01 -1.73622191e-01 -1.39242303e+00 -8.29482898e-02
-1.30002737e-01 -8.16634372e-02 1.08149815e+00 4.06842321e-01
1.00434327e+00 2.33710110e-01 -4.58820850e-01 8.10451329e-01
1.41260290e+00 -2.60655135e-01 5.02680480e-01 3.15858722e-01
8.54672909e-01 5.42404890e-01 9.59660769e-01 4.35352355e-01
6.71484053e-01 7.62306571e-01 9.43163693e-01 -6.70608729e-02
-4.54745740e-01 -3.73456806e-01 1.90350473e-01 2.07254410e-01
1.41308933e-01 -1.62009314e-01 -6.38102651e-01 4.68673497e-01
-1.60145664e+00 -7.28802860e-01 -5.39115012e-01 2.36709571e+00
5.49975693e-01 5.43740809e-01 2.93191671e-01 1.01994455e-01
2.29950413e-01 2.59743005e-01 -8.62337530e-01 -1.01030832e-02
-1.16063058e-01 -2.04518642e-02 9.74158347e-01 6.23914123e-01
-1.12320900e+00 9.19511735e-01 4.85203600e+00 7.18451619e-01
-1.02180731e+00 2.82479562e-02 3.46612036e-01 -2.58381903e-01
-2.18511000e-01 1.67076930e-01 -1.39058554e+00 2.03777254e-01
5.38113773e-01 3.75393391e-01 1.81535482e-01 9.84382272e-01
2.33868018e-01 -5.13006449e-01 -1.23248398e+00 1.00552976e+00
-3.44596580e-02 -1.09810448e+00 -1.56630874e-01 3.84007782e-01
5.51675260e-01 4.15111899e-01 -1.45749703e-01 2.34509677e-01
-2.51935702e-02 -5.66315651e-01 1.22573817e+00 1.39228716e-01
8.05444837e-01 -7.82659829e-01 4.76609856e-01 6.14616454e-01
-1.51616287e+00 -2.18009263e-01 -4.61989880e-01 -6.46618903e-02
3.27878743e-01 7.72270262e-01 -1.13089871e+00 4.10824716e-01
6.89317465e-01 7.29041994e-01 -5.87429166e-01 1.16240203e+00
-3.05468440e-01 1.46574620e-02 -6.12961531e-01 1.72708809e-01
3.76490176e-01 -1.77167077e-02 6.86785221e-01 1.10680282e+00
1.81308717e-01 -6.17030784e-02 1.56861514e-01 1.24875379e+00
-1.59044921e-01 -2.13112250e-01 -5.18069744e-01 4.67127144e-01
6.37508631e-01 1.19067490e+00 -7.72290707e-01 -2.46093720e-01
-5.36797702e-01 7.20078468e-01 3.38555992e-01 1.29459545e-01
-9.28440809e-01 -1.57190323e-01 8.97523820e-01 5.47029078e-01
7.85208583e-01 -4.41593319e-01 -6.37938440e-01 -8.27246249e-01
3.20975184e-01 -3.54117334e-01 -2.96902526e-02 -6.10519290e-01
-1.07808363e+00 2.39056036e-01 -4.97035384e-02 -1.38564706e+00
-1.83946595e-01 -7.55519509e-01 -1.40982002e-01 9.49762881e-01
-2.12168384e+00 -1.18266714e+00 -7.00197756e-01 3.84664387e-01
6.59517646e-01 4.64707941e-01 3.98271561e-01 5.06256998e-01
-4.04888064e-01 3.17272484e-01 -2.80445158e-01 -2.00254485e-01
6.80883765e-01 -1.20011544e+00 6.74742818e-01 8.46327245e-01
5.11629656e-02 2.43693143e-01 5.25802910e-01 -6.84757292e-01
-1.47769034e+00 -1.24146247e+00 7.80112386e-01 -8.00697386e-01
3.62164587e-01 -8.09152186e-01 -9.23428476e-01 2.75734454e-01
-4.68554556e-01 4.29822296e-01 2.20849668e-03 -1.96549028e-01
-5.52868426e-01 -3.40050936e-01 -9.56575990e-01 3.90626729e-01
1.24687874e+00 -6.34247780e-01 -2.75004625e-01 1.60074353e-01
7.80476689e-01 -7.79377043e-01 -4.16122705e-01 6.03953123e-01
4.51362610e-01 -1.23969495e+00 1.43969452e+00 1.54477283e-01
3.38153094e-01 -6.70638323e-01 -2.33402744e-01 -6.18979394e-01
-1.86815839e-02 -1.84027061e-01 -3.27811658e-01 1.07899261e+00
2.49510825e-01 -4.93740976e-01 1.01791775e+00 4.61281955e-01
-4.56745565e-01 -8.10224235e-01 -1.05570173e+00 -7.19427705e-01
-1.97285131e-01 -9.81329858e-01 4.63236272e-01 4.13176596e-01
-7.45606542e-01 3.10463279e-01 -9.00808126e-02 3.68506104e-01
8.19028080e-01 2.17850864e-01 1.33256936e+00 -1.35025454e+00
-1.81773782e-01 -6.30329967e-01 -4.50825691e-01 -1.82513857e+00
-8.22117552e-02 -7.87759781e-01 4.71080601e-01 -1.28522205e+00
-1.64740495e-02 -6.50822699e-01 1.75632745e-01 8.21562186e-02
-2.29563877e-01 4.74803746e-01 9.27642137e-02 1.83773771e-01
-4.53742504e-01 5.52530766e-01 1.21975708e+00 2.37967763e-02
-3.65870416e-01 2.99947485e-02 -3.97737384e-01 7.04763830e-01
3.26874971e-01 -5.48699141e-01 -2.36663327e-01 -5.05155981e-01
-6.50583953e-02 -2.09861860e-01 8.56514573e-01 -1.02256060e+00
3.11605394e-01 1.01153962e-01 5.22107542e-01 -1.40521920e+00
7.96205878e-01 -9.24915195e-01 -3.01089019e-01 2.72930801e-01
-3.98853701e-03 -4.08759952e-01 2.35699758e-01 7.88701296e-01
-3.77443209e-02 5.27436472e-02 8.36301327e-01 4.42083292e-02
-7.58095086e-01 5.21025717e-01 1.73871875e-01 -1.82860985e-01
1.04174268e+00 -6.70475960e-01 -1.04321279e-02 1.21247187e-01
-1.37713552e-01 3.94294798e-01 6.83577180e-01 4.60545003e-01
6.65026128e-01 -1.19690311e+00 -7.57552326e-01 4.89186496e-01
3.42791200e-01 8.85202110e-01 2.18194112e-01 8.89632106e-01
-5.19706905e-01 4.26852375e-01 2.76370049e-01 -1.14051294e+00
-1.21501970e+00 4.21709806e-01 2.70480752e-01 -1.06294140e-01
-8.93592477e-01 1.11921883e+00 3.27346802e-01 -3.33279699e-01
5.60238063e-01 -6.76224947e-01 2.95867324e-01 -2.02553850e-02
5.71691751e-01 2.27803066e-01 3.05519849e-01 -5.71430445e-01
-4.86420870e-01 6.91108644e-01 4.89035156e-03 1.94711480e-02
1.23657835e+00 -3.51296723e-01 1.62431493e-01 2.67673343e-01
1.25457048e+00 5.43056801e-02 -1.94131529e+00 -4.07195598e-01
-1.13424972e-01 -8.95653069e-01 3.96623045e-01 -5.05520761e-01
-8.59329641e-01 1.11236203e+00 6.50682867e-01 -1.72587126e-01
9.32602704e-01 5.58858849e-02 8.65679145e-01 1.98867634e-01
2.68683136e-01 -8.02881062e-01 -6.91530555e-02 4.78658140e-01
6.09468579e-01 -1.56744230e+00 2.52465308e-01 -7.29876280e-01
-1.42000124e-01 9.42104161e-01 6.02494359e-01 -2.57777363e-01
3.34544390e-01 4.07522917e-01 -1.98903620e-01 -1.42750487e-01
-5.76565385e-01 -5.22600889e-01 4.58152235e-01 5.73759556e-01
2.43038591e-02 -2.55165640e-02 2.75756389e-01 8.39525685e-02
-3.06874812e-02 -2.17343301e-01 -1.28187360e-02 8.39051723e-01
-6.64312899e-01 -9.45873082e-01 -5.28824806e-01 1.54751435e-01
4.68526743e-02 3.53185274e-02 -1.91246167e-01 9.04818296e-01
4.95530248e-01 6.66985512e-01 3.77421021e-01 -9.46246982e-02
6.93617344e-01 -2.17848092e-01 4.86693203e-01 -6.42060578e-01
-2.35648081e-01 2.43559331e-01 -6.98910980e-03 -8.34815085e-01
-1.59262419e-01 -7.78563201e-01 -1.34887505e+00 -3.00013751e-01
-5.46533883e-01 -5.10574162e-01 9.37145174e-01 6.53001904e-01
2.83701569e-01 2.95892924e-01 4.62176859e-01 -1.43806851e+00
-5.92566133e-01 -7.16081560e-01 -2.83944756e-01 3.14898401e-01
6.74830377e-01 -7.49924898e-01 -3.89907598e-01 -2.78755844e-01] | [7.8098320960998535, -2.637403726577759] |
986c6807-3705-4c74-83ee-b30e094ec091 | user-centric-conversational-recommendation | 2204.09263 | null | https://arxiv.org/abs/2204.09263v2 | https://arxiv.org/pdf/2204.09263v2.pdf | User-Centric Conversational Recommendation with Multi-Aspect User Modeling | Conversational recommender systems (CRS) aim to provide highquality recommendations in conversations. However, most conventional CRS models mainly focus on the dialogue understanding of the current session, ignoring other rich multi-aspect information of the central subjects (i.e., users) in recommendation. In this work, we highlight that the user's historical dialogue sessions and look-alike users are essential sources of user preferences besides the current dialogue session in CRS. To systematically model the multi-aspect information, we propose a User-Centric Conversational Recommendation (UCCR) model, which returns to the essence of user preference learning in CRS tasks. Specifically, we propose a historical session learner to capture users' multi-view preferences from knowledge, semantic, and consuming views as supplements to the current preference signals. A multi-view preference mapper is conducted to learn the intrinsic correlations among different views in current and historical sessions via self-supervised objectives. We also design a temporal look-alike user selector to understand users via their similar users. The learned multi-aspect multi-view user preferences are then used for the recommendation and dialogue generation. In experiments, we conduct comprehensive evaluations on both Chinese and English CRS datasets. The significant improvements over competitive models in both recommendation and dialogue generation verify the superiority of UCCR. | ['Qing He', 'Fuzhen Zhuang', 'Xiang Ao', 'Yongchun Zhu', 'Ruobing Xie', 'Shuokai Li'] | 2022-04-20 | null | null | null | null | ['dialogue-understanding'] | ['natural-language-processing'] | [-9.17679667e-02 -1.23506702e-01 -5.92556715e-01 -7.53480554e-01
-8.57111812e-01 -5.64437330e-01 6.98957980e-01 -3.05305034e-01
5.03327288e-02 3.88515890e-01 1.29552865e+00 -3.51830497e-02
-3.38245034e-01 -5.69798589e-01 -1.41958714e-01 -4.68403071e-01
4.17254716e-01 4.71893013e-01 -1.84982806e-01 -8.16068649e-01
4.36543673e-01 -2.32455060e-01 -1.46020222e+00 1.10673964e+00
1.12000895e+00 8.58109832e-01 6.15407765e-01 5.92046380e-01
-1.40664265e-01 4.15526032e-01 -1.07447796e-01 -4.41038400e-01
-2.70100802e-01 -4.96591747e-01 -7.91192472e-01 2.81374753e-01
-1.25404567e-01 -3.26879799e-01 -2.47185364e-01 3.65246862e-01
6.92984045e-01 7.45432436e-01 5.85978687e-01 -8.50902200e-01
-9.79864597e-01 1.09489214e+00 -2.34425694e-01 8.08816850e-02
7.88415909e-01 -2.80999422e-01 1.61272299e+00 -1.08540559e+00
4.77981776e-01 1.40672302e+00 2.10294232e-01 8.44168842e-01
-8.37865770e-01 -2.22140342e-01 8.61967027e-01 1.84963599e-01
-7.04961717e-01 -3.87015820e-01 1.02470100e+00 -2.60746449e-01
3.89320999e-01 8.54200363e-01 5.02721786e-01 1.44722962e+00
-3.88900518e-01 1.10999775e+00 8.56376231e-01 1.02336057e-01
-1.18390106e-01 6.68046296e-01 3.65008593e-01 1.99486479e-01
-5.52961230e-01 -1.52859017e-01 -7.62983024e-01 -2.43541643e-01
3.72744560e-01 6.16918683e-01 -5.36281884e-01 -2.08412021e-01
-1.33148420e+00 8.65049541e-01 2.02193469e-01 1.94773510e-01
-2.22274959e-01 -7.39304960e-01 3.24319959e-01 3.88329715e-01
5.68555534e-01 6.20037258e-01 -8.63892794e-01 -2.26326659e-01
-1.75450742e-01 1.16276346e-01 9.53885734e-01 1.24844456e+00
5.49310327e-01 -3.76127303e-01 -6.93800390e-01 1.40074468e+00
6.93913937e-01 4.65341270e-01 7.00344265e-01 -7.39515245e-01
5.82498729e-01 7.35612512e-01 3.06615502e-01 -1.03858924e+00
-2.58366048e-01 -7.27398217e-01 -7.54273534e-01 -9.83236372e-01
-2.01884240e-01 -3.99676859e-01 -2.84306407e-01 1.57082856e+00
2.18037143e-01 -1.09185599e-01 4.20690477e-01 1.07301223e+00
1.38137639e+00 8.83078158e-01 -2.19526857e-01 -5.73593318e-01
1.38416672e+00 -1.35434234e+00 -6.06170714e-01 -3.85468714e-02
3.47211719e-01 -7.88842201e-01 1.44666088e+00 2.75729358e-01
-7.67431319e-01 -7.14234710e-01 -5.41564703e-01 3.27371992e-02
7.08890986e-03 7.03019559e-01 6.56962991e-01 4.43214327e-01
-5.78303456e-01 2.35973433e-01 -1.46494105e-01 -3.33213300e-01
-1.70535415e-01 2.78544724e-01 3.22018415e-02 -8.55452865e-02
-1.37930083e+00 5.96053600e-01 -2.21217066e-01 1.97231248e-01
-5.84315181e-01 -8.71551692e-01 -5.60492873e-01 6.51852414e-02
4.66175020e-01 -7.49540985e-01 1.42398179e+00 -8.68010998e-01
-1.83162081e+00 2.18715608e-01 -4.39240396e-01 1.70122504e-01
1.13890558e-01 -2.11790413e-01 -8.58882844e-01 -3.06409150e-01
-1.14038229e-01 3.08844987e-02 5.15185833e-01 -1.67125285e+00
-1.21069372e+00 -2.72605807e-01 5.76368332e-01 9.48743939e-01
-5.59146106e-01 -2.09845901e-01 -7.75537968e-01 -5.27877569e-01
-4.88599204e-02 -1.00323975e+00 -2.48069376e-01 -7.65678048e-01
-2.27477938e-01 -4.98810053e-01 5.92707813e-01 -5.38047850e-01
1.38216424e+00 -1.97864139e+00 4.43504244e-01 1.36661366e-01
5.00849672e-02 -2.54830550e-02 -3.09484780e-01 6.75779760e-01
3.05489361e-01 -1.75891832e-01 3.80713344e-01 -4.69053596e-01
1.02914460e-01 8.28498453e-02 -5.48197448e-01 1.82111114e-02
-5.79301953e-01 6.91964567e-01 -1.19727552e+00 -3.53176817e-02
-1.11457435e-02 3.41485023e-01 -9.46895897e-01 8.24910939e-01
-1.99920893e-01 8.25056434e-01 -9.72724319e-01 2.24388868e-01
2.74125338e-01 -4.48167771e-01 4.11297143e-01 -6.55942082e-01
-6.45758957e-02 7.66826689e-01 -8.62220943e-01 1.81791127e+00
-1.21382523e+00 -6.57151490e-02 -9.41928551e-02 -5.80532908e-01
1.01882780e+00 6.01236463e-01 5.40042043e-01 -6.30409896e-01
-9.21123698e-02 -4.77960594e-02 -6.49308562e-02 -6.09745681e-01
8.09958458e-01 1.01549022e-01 -1.49392322e-01 9.11144257e-01
5.85408919e-02 3.73602688e-01 6.42654533e-03 4.75183725e-01
3.28243285e-01 4.21732552e-02 1.57215863e-01 -2.44368270e-01
9.77240503e-01 -4.11433220e-01 5.99292755e-01 5.46884537e-01
4.97122467e-01 6.65973544e-01 1.43640444e-01 -1.51496917e-01
-3.78672749e-01 -6.64507568e-01 3.12699646e-01 1.78912938e+00
4.74981755e-01 -8.38510931e-01 -1.90792531e-01 -9.15237188e-01
-3.87491465e-01 8.26470137e-01 -6.82692409e-01 -3.11823547e-01
-3.69483292e-01 -7.87518978e-01 -3.31225038e-01 4.15487468e-01
1.14374839e-01 -1.00524664e+00 1.44012049e-01 1.28253102e-01
-8.40558469e-01 -8.52691054e-01 -1.25467026e+00 -5.16799808e-01
-7.75362551e-01 -8.96434903e-01 -6.87540531e-01 -6.06195152e-01
6.19529188e-01 9.21612859e-01 1.26188838e+00 1.59303948e-01
5.86967766e-01 8.89116406e-01 -9.49030578e-01 5.76798096e-02
-6.10913560e-02 2.25060031e-01 2.65435755e-01 5.98351061e-01
2.77627051e-01 -7.67602086e-01 -8.83108079e-01 8.36839259e-01
-3.57573479e-01 3.41874570e-01 4.38120365e-01 9.37524438e-01
4.53894317e-01 -2.75060773e-01 8.08410525e-01 -1.49696863e+00
1.09703958e+00 -9.15456712e-01 1.24798320e-01 5.10717452e-01
-7.90139258e-01 -2.49316797e-01 8.10298741e-01 -4.87679541e-01
-1.81758630e+00 -2.85341829e-01 -3.09954226e-01 1.04192026e-01
-1.20209351e-01 7.43492246e-01 -5.31468630e-01 4.87775534e-01
2.93058634e-01 3.36704195e-01 -4.17775303e-01 -8.17625105e-01
7.52229571e-01 9.25529540e-01 5.76256551e-02 -9.17919338e-01
3.70604903e-01 2.41124555e-01 -9.18209493e-01 -6.77916169e-01
-1.35749066e+00 -1.01039767e+00 -3.21900696e-01 -3.31320822e-01
5.35842359e-01 -1.08864713e+00 -8.17599058e-01 -2.36452408e-02
-9.95041311e-01 1.16538823e-01 -1.41055556e-02 5.22984624e-01
-4.60344970e-01 2.94512004e-01 -4.31550831e-01 -9.33170199e-01
-5.18507242e-01 -1.13218534e+00 9.06716049e-01 6.50939465e-01
-4.96663116e-02 -1.25179625e+00 1.06688753e-01 7.69606650e-01
3.70318532e-01 -7.42636740e-01 7.27434397e-01 -9.98361409e-01
-2.10001722e-01 8.26448426e-02 -7.09642097e-02 -9.30605829e-03
5.73409140e-01 -2.86199003e-01 -1.00246966e+00 -2.66556501e-01
2.02111211e-02 -1.13013081e-01 5.39626479e-01 1.61490902e-01
1.17886782e+00 -5.66435456e-01 -3.75976324e-01 2.91515589e-01
8.67624879e-01 3.36253136e-01 2.72125453e-01 -2.26182178e-01
8.63667786e-01 9.49812710e-01 9.92192507e-01 6.60033941e-01
1.05244577e+00 8.01729679e-01 6.98290616e-02 6.58416674e-02
2.88530648e-01 -6.83257580e-01 4.91866648e-01 1.52047551e+00
-5.00172436e-01 -4.95201945e-01 -6.69247359e-02 4.10889566e-01
-2.11571670e+00 -1.10227954e+00 3.91926896e-03 2.03639674e+00
7.66759098e-01 -3.41468304e-01 2.80874610e-01 -5.70679307e-01
7.07000971e-01 3.14628899e-01 -4.63101864e-01 -1.79485977e-01
1.91181049e-01 -5.11228979e-01 -2.13558629e-01 4.86930162e-01
-7.65342414e-01 7.29709923e-01 4.66304684e+00 6.69189274e-01
-7.74554789e-01 1.23620071e-01 6.52590692e-01 -7.49982893e-02
-1.14667332e+00 -1.26045167e-01 -9.46771502e-01 4.58406687e-01
4.94246453e-01 -4.06878889e-01 7.56119192e-01 7.04048038e-01
4.06106532e-01 5.38888037e-01 -1.20816517e+00 9.90239978e-01
2.93083340e-01 -1.23349977e+00 3.06937218e-01 2.68834326e-02
8.53315592e-01 -3.07779610e-01 2.29558423e-01 6.62641883e-01
4.29015964e-01 -5.86342096e-01 1.75395072e-01 7.54418254e-01
3.94686997e-01 -8.11368644e-01 6.51412070e-01 5.23476839e-01
-1.37988257e+00 -3.24791402e-01 -3.93995970e-01 2.78975666e-01
2.88963497e-01 3.56319278e-01 -4.95555669e-01 9.38701987e-01
5.13597965e-01 1.30140114e+00 -2.29607046e-01 5.54998040e-01
1.62040759e-02 4.11153376e-01 2.81699926e-01 -7.22590163e-02
-6.38352185e-02 -7.18782246e-01 4.60692942e-01 1.17601013e+00
6.12017214e-01 6.38939917e-01 3.14287454e-01 4.93478775e-01
-1.14673465e-01 4.98311490e-01 -2.73578227e-01 1.01420887e-01
5.76648235e-01 1.56479740e+00 -8.93464684e-02 -3.41560654e-02
-6.91505671e-01 8.91488850e-01 6.10421859e-02 3.91719222e-01
-3.59211653e-01 1.87535763e-01 8.70887756e-01 -2.96287090e-01
2.90577143e-01 3.26360077e-01 8.41009170e-02 -1.51928258e+00
-2.72387594e-01 -1.19103634e+00 6.59563899e-01 -4.65204686e-01
-1.67418754e+00 6.97031915e-01 -3.73636037e-01 -1.69679642e+00
-1.87943101e-01 -1.10390529e-01 -9.99485850e-01 8.46067011e-01
-1.34930468e+00 -1.41416085e+00 -1.50739387e-01 8.28271627e-01
1.17602360e+00 -5.66004455e-01 9.54158783e-01 3.30899298e-01
-4.19967026e-01 7.38750577e-01 9.22581479e-02 -3.07909340e-01
8.43728840e-01 -1.26423609e+00 3.39565426e-02 3.52021188e-01
2.10167170e-01 1.06813633e+00 6.57960176e-01 -3.69116008e-01
-1.77102160e+00 -9.54127073e-01 8.65160167e-01 -4.83723968e-01
2.36110374e-01 -1.62250429e-01 -6.11745298e-01 5.70952773e-01
2.02090323e-01 -6.83726370e-01 1.26414335e+00 9.49903846e-01
-1.50665209e-01 -2.72834986e-01 -6.82132721e-01 8.03245127e-01
1.08327436e+00 -5.28200448e-01 -8.07631135e-01 2.30021566e-01
9.39692616e-01 -4.60079193e-01 -9.70077634e-01 2.92960852e-01
8.43825996e-01 -7.96363115e-01 1.22144127e+00 -8.16320062e-01
5.78764141e-01 -9.68751013e-02 -3.08105648e-01 -1.86791468e+00
-3.55444014e-01 -6.42912149e-01 -1.77365273e-01 1.61583745e+00
7.25380957e-01 -3.43635529e-01 4.94358808e-01 5.88225126e-01
-4.02362853e-01 -6.93907142e-01 -1.68620795e-01 2.79955447e-01
-5.11078238e-01 -1.14026085e-01 8.37287664e-01 1.15629792e+00
5.83920479e-01 9.98447359e-01 -1.10790849e+00 7.68602714e-02
1.62326440e-01 9.49425399e-01 6.84072495e-01 -1.35436463e+00
-4.35599446e-01 -2.70629257e-01 8.34501326e-01 -1.79494190e+00
6.51274621e-02 -7.96875000e-01 5.52220643e-02 -1.62413144e+00
3.65873486e-01 -7.30461955e-01 -5.99437535e-01 -2.16748506e-01
-4.75097537e-01 -2.82654196e-01 -5.56735024e-02 1.46264777e-01
-9.38310981e-01 9.25517559e-01 1.74014127e+00 1.03278376e-01
-5.50006151e-01 9.75326419e-01 -1.58878052e+00 5.22924542e-01
6.42410278e-01 -1.35126770e-01 -1.11877632e+00 -2.94352263e-01
4.99823391e-01 5.20009995e-01 -4.22428399e-01 -1.61223218e-01
2.70366520e-01 -7.02902138e-01 -6.26563886e-03 -5.83466649e-01
5.71835339e-01 -7.95755923e-01 -1.05567172e-01 -3.40132326e-01
-1.01266718e+00 -2.39253137e-02 -4.46138710e-01 9.63938951e-01
-8.18178579e-02 4.02798355e-02 1.67596057e-01 -3.27468872e-01
-5.63386440e-01 7.69570529e-01 -3.49492669e-01 -8.46849233e-02
3.16640437e-01 3.12806278e-01 -3.73834848e-01 -9.28498209e-01
-9.42142963e-01 7.25435197e-01 3.64815369e-02 1.19134295e+00
7.64270604e-01 -1.44124794e+00 -6.11506283e-01 -2.72950809e-02
4.31666791e-01 -3.31566781e-01 8.49872768e-01 7.13531375e-01
5.95512688e-01 4.73771393e-01 -1.76910087e-02 -6.25272766e-02
-1.44315910e+00 4.07905430e-01 1.93974525e-01 -2.05707967e-01
-3.56106222e-01 8.22742283e-01 4.84624416e-01 -1.09016204e+00
2.70373821e-01 1.29698426e-01 -1.34415138e+00 4.73837972e-01
7.08470702e-01 3.60689253e-01 -3.50186229e-01 -6.86455548e-01
1.06757797e-01 2.83751667e-01 -3.55921656e-01 -1.34046108e-01
1.45098948e+00 -8.75126660e-01 -1.50773585e-01 6.67625129e-01
9.11482215e-01 5.09341776e-01 -8.70176435e-01 -7.34538794e-01
-1.86507791e-01 -5.83576679e-01 -1.17505491e-01 -9.77288902e-01
-1.21321404e+00 8.59181702e-01 2.33916759e-01 1.40523717e-01
9.71637726e-01 9.93406773e-02 9.60107207e-01 4.62725103e-01
3.01784307e-01 -1.15099192e+00 7.60864019e-02 5.39710701e-01
1.10943794e+00 -1.15525687e+00 -1.30226016e-01 -4.81127679e-01
-1.46391284e+00 1.01151586e+00 1.05902779e+00 3.67571294e-01
8.86197865e-01 -6.10222697e-01 2.73740917e-01 -3.63221020e-02
-1.34546697e+00 -1.87036052e-01 8.29299390e-01 5.38051546e-01
7.95961022e-01 2.65699118e-01 -4.72327888e-01 1.61603177e+00
-2.22893149e-01 -3.20932120e-01 4.54611123e-01 3.72833461e-01
-2.71915197e-01 -1.37985170e+00 3.78186941e-01 6.98408425e-01
-4.36581448e-02 -1.38862118e-01 -2.82674640e-01 -6.96659312e-02
-1.03478223e-01 1.43621385e+00 -2.85093009e-01 -9.78703201e-01
5.71982682e-01 -2.69359827e-01 -6.51149228e-02 -9.04630423e-01
-8.10011446e-01 2.90503919e-01 5.43143213e-01 -4.56268340e-01
-6.75100327e-01 -5.93726635e-01 -8.07261705e-01 -1.94384396e-01
-3.57106775e-01 8.51140499e-01 2.87790686e-01 1.17536807e+00
4.61887181e-01 5.58401644e-01 1.33464241e+00 -6.87059045e-01
-5.16676724e-01 -1.05718231e+00 -6.28964722e-01 3.72254074e-01
2.53115267e-01 -4.58038956e-01 -1.59383729e-01 2.75137601e-03] | [12.346664428710938, 7.499161243438721] |
c4010e06-1a0b-45eb-a6af-d1c8dbefa6a5 | learning-to-explore-intrinsic-saliency-for | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Zhang_Learning_to_Explore_Intrinsic_Saliency_for_Stereoscopic_Video_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Zhang_Learning_to_Explore_Intrinsic_Saliency_for_Stereoscopic_Video_CVPR_2019_paper.pdf | Learning to Explore Intrinsic Saliency for Stereoscopic Video | The human visual system excels at biasing the stereoscopic visual signals by the attention mechanisms. Traditional methods relying on the low-level features and depth relevant information for stereoscopic video saliency prediction have fundamental limitations. For example, it is cumbersome to model the interactions between multiple visual cues including spatial, temporal, and depth information as a result of the sophistication. In this paper, we argue that the high-level features are crucial and resort to the deep learning framework to learn the saliency map of stereoscopic videos. Driven by spatio-temporal coherence from consecutive frames, the model first imitates the mechanism of saliency by taking advantage of the 3D convolutional neural network. Subsequently, the saliency originated from the intrinsic depth is derived based on the correlations between left and right views in a data-driven manner. Finally, a Convolutional Long Short-Term Memory (Conv-LSTM) based fusion network is developed to model the instantaneous interactions between spatio-temporal and depth attributes, such that the ultimate stereoscopic saliency maps over time are produced. Moreover, we establish a new large-scale stereoscopic video saliency dataset (SVS) including 175 stereoscopic video sequences and their fixation density annotations, aiming to comprehensively study the intrinsic attributes for stereoscopic video saliency detection. Extensive experiments show that our proposed model can achieve superior performance compared to the state-of-the-art methods on the newly built dataset for stereoscopic videos.
| [' Jianmin Jiang', ' Sam Kwong', ' Shikai Li', ' Shiqi Wang', ' Xu Wang', 'Qiudan Zhang'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['video-saliency-detection'] | ['computer-vision'] | [ 3.08114439e-01 -3.73362660e-01 -2.43969128e-01 -2.14054182e-01
-4.06729251e-01 3.56173702e-02 3.68977368e-01 -2.16611564e-01
-2.12270781e-01 7.28804111e-01 5.34604609e-01 2.12151483e-01
3.70513871e-02 -4.13207710e-01 -8.60845506e-01 -8.21162164e-01
1.27889961e-02 -3.96813899e-01 8.59323680e-01 -3.71409625e-01
8.26533616e-01 -1.37802335e-02 -2.08073068e+00 3.66090059e-01
1.17906964e+00 1.33617425e+00 1.05819440e+00 3.64538193e-01
9.23917964e-02 1.00970960e+00 5.29417489e-03 1.44120395e-01
-3.92701626e-02 -5.79069793e-01 -7.55977929e-01 -1.22801468e-01
3.00672740e-01 -4.48295772e-01 -2.95545727e-01 1.29690254e+00
5.67154288e-01 1.30094364e-01 4.06528085e-01 -1.14571404e+00
-6.80940688e-01 2.45746762e-01 -6.40468955e-01 9.04297888e-01
5.26359737e-01 2.91455626e-01 8.64441931e-01 -1.00212717e+00
6.96030676e-01 9.87166405e-01 8.76851380e-02 4.88768131e-01
-7.95108676e-01 -4.11689252e-01 3.71844649e-01 7.36011028e-01
-9.90774155e-01 -2.99336165e-01 1.17699635e+00 -4.66392398e-01
4.68550652e-01 7.94587061e-02 9.30126727e-01 1.16967571e+00
3.48013192e-01 1.20993137e+00 1.08180654e+00 -1.39554247e-01
-3.70057151e-02 3.77812125e-02 -8.19057226e-02 6.09745741e-01
-2.82654874e-02 3.97293299e-01 -9.40707266e-01 3.50827873e-01
1.05495727e+00 3.59094530e-01 -4.92490411e-01 -5.99090278e-01
-1.42487168e+00 5.69855392e-01 9.23169553e-01 4.51601088e-01
-4.35500115e-01 -2.23908529e-01 2.89459437e-01 -7.35278726e-02
4.72381175e-01 4.05520439e-01 -3.33613962e-01 3.85707580e-02
-1.01343524e+00 -4.96095046e-02 -3.68763953e-02 7.83334136e-01
8.96278799e-01 3.02203745e-01 -4.11506683e-01 6.52438760e-01
2.09889948e-01 5.06733537e-01 8.28754306e-01 -8.97195458e-01
2.20913127e-01 5.41511953e-01 1.62727043e-01 -1.21217418e+00
-3.21837336e-01 -3.06251228e-01 -6.69353127e-01 1.49900064e-01
5.88025488e-02 8.49837363e-02 -6.86897039e-01 1.73856175e+00
6.10042252e-02 3.02142620e-01 -1.41509861e-01 1.61869180e+00
1.07377613e+00 3.18459332e-01 1.71320155e-01 -2.83421874e-01
1.11700082e+00 -9.23217773e-01 -6.47324562e-01 -1.93277702e-01
3.76629196e-02 -6.20128870e-01 1.08880746e+00 -1.53015614e-01
-1.19946063e+00 -8.14753413e-01 -9.92914140e-01 -3.41971576e-01
-4.46722478e-01 -4.86722738e-02 5.28922379e-01 -8.14281553e-02
-1.20061052e+00 3.64748150e-01 -5.73239803e-01 -2.06487522e-01
4.64587510e-01 3.86666328e-01 7.79686794e-02 4.01192576e-01
-1.61656547e+00 9.07226562e-01 3.50594521e-01 4.18654934e-04
-1.08358133e+00 -6.20867074e-01 -1.01592422e+00 1.66481614e-01
3.44880342e-01 -8.06344032e-01 9.77230370e-01 -1.55032063e+00
-1.20397675e+00 7.76469529e-01 -6.82539821e-01 -3.01625639e-01
1.73323423e-01 -3.87941152e-02 -2.95584083e-01 5.56426823e-01
3.24105382e-01 1.03956139e+00 1.02605104e+00 -1.19124484e+00
-1.11964417e+00 -2.88117439e-01 2.68842399e-01 5.70926785e-01
-3.76953840e-01 1.44496053e-01 -2.29113594e-01 -4.92660612e-01
5.15294448e-02 -4.48167294e-01 2.43773088e-02 -2.65141785e-01
-2.60184109e-01 1.08000753e-03 8.51639748e-01 -5.50415695e-01
1.10289931e+00 -1.95740330e+00 4.76445913e-01 -4.19486254e-01
4.58906800e-01 2.41661355e-01 1.59021869e-01 -1.82639346e-01
-6.86966106e-02 -2.81448513e-01 6.96605742e-02 4.34887819e-02
-4.65177268e-01 -4.64998364e-01 -3.80769223e-01 2.49109045e-01
2.62101352e-01 1.09070802e+00 -1.35719013e+00 -9.54286933e-01
6.26722932e-01 4.19016451e-01 -4.07487035e-01 2.70087630e-01
-1.58783756e-02 7.20942974e-01 -7.06541598e-01 8.76820505e-01
4.98623341e-01 -3.18459690e-01 -4.59980905e-01 -2.58735985e-01
-4.70733583e-01 2.17350230e-01 -5.15923440e-01 1.99130034e+00
-1.60720065e-01 9.05817568e-01 -2.28735685e-01 -8.16584051e-01
7.78451085e-01 1.25665873e-01 5.54310739e-01 -1.02643251e+00
3.42771709e-01 4.37704831e-01 -2.15763733e-01 -7.36657441e-01
4.91303533e-01 7.37705305e-02 2.57590294e-01 1.37256861e-01
1.94119781e-01 7.21076280e-02 -4.91092205e-02 9.04302076e-02
3.68041575e-01 1.95927456e-01 2.01667711e-01 -4.54482675e-01
8.59129548e-01 -1.58472866e-01 4.91077065e-01 3.65459919e-01
-5.83197951e-01 9.27602053e-01 3.02433670e-01 -6.00066066e-01
-9.69315469e-01 -1.02962768e+00 4.75311019e-02 1.11253762e+00
9.47258174e-01 1.62966236e-01 -6.48828208e-01 -4.83420998e-01
-3.67359489e-01 4.51167017e-01 -8.65182161e-01 -3.98445010e-01
-3.79104763e-01 -4.16288435e-01 -2.80690849e-01 4.67690378e-01
7.50173211e-01 -1.56496501e+00 -9.56818342e-01 -7.31827915e-02
-5.02042890e-01 -9.31768596e-01 -6.85859382e-01 -1.66258007e-01
-8.65104556e-01 -1.15157402e+00 -1.07560265e+00 -1.16772437e+00
5.06663978e-01 9.74230826e-01 8.47589910e-01 -9.42986757e-02
-2.27576625e-02 2.81115733e-02 -1.97160468e-01 -3.75963241e-01
3.10186058e-01 1.18333980e-01 1.12585686e-02 2.30265960e-01
5.61019838e-01 -6.97915077e-01 -1.10045171e+00 3.00197214e-01
-8.53006184e-01 6.76590681e-01 5.94624043e-01 8.47864747e-01
5.35524249e-01 -4.05236870e-01 6.02531850e-01 -2.33031243e-01
1.94565266e-01 -6.08872533e-01 -5.20392597e-01 1.18515223e-01
-1.69674665e-01 -8.03675279e-02 4.53102559e-01 -3.06777894e-01
-1.26764607e+00 -3.33862772e-05 2.82214135e-01 -6.84464097e-01
-1.90085489e-02 2.46625796e-01 -1.86588224e-02 -1.35743722e-01
3.88773978e-01 8.37131321e-01 -2.40312126e-02 -3.39695555e-03
-8.80996287e-02 5.65327287e-01 5.93935728e-01 -9.23427492e-02
3.76209766e-01 6.85608208e-01 -4.79046442e-02 -5.33344686e-01
-1.18005633e+00 -4.50987518e-01 -7.94969380e-01 -6.04742110e-01
1.00577235e+00 -1.10303104e+00 -7.50811279e-01 5.89159906e-01
-1.18223917e+00 -4.39472264e-03 6.47734478e-03 5.19458711e-01
-1.00019038e+00 3.15683514e-01 -4.59459096e-01 -6.99464262e-01
-3.17409426e-01 -1.37792933e+00 1.31493402e+00 6.09558821e-01
1.96072698e-01 -9.10018027e-01 -1.59364164e-01 1.85331479e-01
4.75624561e-01 2.69077688e-01 6.14474237e-01 -1.28058657e-01
-1.06053114e+00 2.98036277e-01 -6.05010808e-01 1.25950009e-01
2.84879625e-01 -2.20556587e-01 -9.44979727e-01 -2.31974609e-02
3.52833837e-01 -2.89976150e-01 1.03757048e+00 9.37756360e-01
1.37900972e+00 2.98609659e-02 -2.86550492e-01 7.32947350e-01
1.33870804e+00 1.02254756e-01 5.96426725e-01 5.34485698e-01
9.15367305e-01 7.18340039e-01 9.80059683e-01 2.13456750e-01
4.81357187e-01 6.18216515e-01 7.29325712e-01 -2.64736831e-01
-2.00036261e-02 -4.47480917e-01 2.84497678e-01 5.79050779e-01
-2.98972934e-01 2.91377246e-01 -5.59597254e-01 8.90240371e-01
-1.98981190e+00 -1.31748807e+00 -4.66931798e-02 2.23432446e+00
6.88923776e-01 1.20304428e-01 2.66651288e-02 -1.09129414e-01
1.02808046e+00 5.12682676e-01 -8.50003004e-01 1.21187214e-02
-4.65456486e-01 -4.65655923e-01 2.27209583e-01 2.60038108e-01
-1.12700677e+00 1.06373453e+00 6.21296883e+00 7.23358750e-01
-1.54648042e+00 2.38345116e-02 7.03241169e-01 -3.38185519e-01
-5.00462949e-01 -1.25224620e-01 -5.15567303e-01 1.07957959e+00
4.86132443e-01 -2.43907347e-01 4.21789020e-01 7.97710359e-01
5.16789317e-01 -3.38597715e-01 -7.54195929e-01 1.09900570e+00
4.21624571e-01 -1.38234115e+00 2.11007982e-01 -2.23687544e-01
1.08139551e+00 2.30086401e-01 5.65859318e-01 -1.41053528e-01
-1.67329147e-01 -8.60278904e-01 8.56527328e-01 8.53188217e-01
7.33129859e-01 -5.85862160e-01 7.63709545e-01 2.34614640e-01
-1.10770559e+00 -3.92604440e-01 -4.43526328e-01 -1.54899687e-01
2.02420503e-01 3.42424959e-01 -3.26147616e-01 3.92531514e-01
1.02113807e+00 1.34124529e+00 -6.81402922e-01 1.11155558e+00
-1.92028545e-02 -2.26408709e-02 1.59194604e-01 -1.94782868e-01
2.41327912e-01 8.81779846e-03 7.64707565e-01 6.71084881e-01
3.75381052e-01 -2.19068844e-02 -2.63727367e-01 1.02981377e+00
3.96491677e-01 -7.85001069e-02 -6.85416818e-01 3.41695637e-01
2.28246301e-01 1.10034406e+00 -4.37724322e-01 -5.08221388e-01
-4.62343156e-01 8.42603505e-01 2.65491188e-01 6.02442205e-01
-9.42186177e-01 -2.05302507e-01 5.11173487e-01 1.63343474e-01
5.09451866e-01 2.50552714e-01 -4.01489735e-01 -1.47596169e+00
5.29043116e-02 -5.96569717e-01 2.26305112e-01 -1.39239264e+00
-8.54098797e-01 7.79209852e-01 -2.23105162e-01 -1.59894919e+00
-2.80582905e-01 -3.47048998e-01 -7.04215467e-01 9.75299180e-01
-2.08290029e+00 -1.01115286e+00 -6.49669766e-01 9.67493594e-01
8.51384163e-01 -1.82285503e-01 2.04700142e-01 -3.93620953e-02
-3.22456867e-01 1.79189071e-02 -4.29375544e-02 -2.20047981e-01
6.29889488e-01 -1.02412903e+00 -1.08546205e-01 9.41256225e-01
-3.21511686e-01 4.47614908e-01 8.28384638e-01 -5.28600812e-01
-9.22295809e-01 -8.75809312e-01 9.42260861e-01 -4.63613480e-01
6.04045868e-01 -1.08443968e-01 -6.87194228e-01 3.11253458e-01
2.57582873e-01 6.91265007e-03 1.49347037e-01 -2.71719277e-01
-2.11084988e-02 -1.03450969e-01 -8.50370526e-01 8.11898649e-01
1.20105183e+00 -8.18211555e-01 -8.39877605e-01 4.58021095e-04
9.43889856e-01 -2.96003997e-01 -3.01754326e-01 5.47945261e-01
5.77813208e-01 -1.53637230e+00 1.05171669e+00 -3.97718847e-01
9.88403618e-01 -4.40826833e-01 2.81774681e-02 -1.14056754e+00
-3.82391095e-01 -3.72553825e-01 -1.79873258e-02 7.85856724e-01
3.12363971e-02 -3.92777681e-01 5.71882010e-01 2.42079705e-01
-3.62248302e-01 -7.81592011e-01 -8.42118800e-01 -3.11939567e-01
-5.23381472e-01 1.63312871e-02 5.36955357e-01 6.77519381e-01
2.34675631e-01 3.27254981e-01 -5.24047792e-01 -1.21146835e-01
5.09966910e-01 6.00704849e-01 2.90427327e-01 -1.17258573e+00
2.91919470e-01 -6.44622445e-01 -5.02583325e-01 -1.19455624e+00
2.36905083e-01 -5.53509474e-01 -3.37424269e-03 -1.22140539e+00
6.84237003e-01 3.11346889e-01 -8.94121826e-01 -6.86604530e-02
-6.78545237e-01 2.19021872e-01 1.15590863e-01 3.38210016e-01
-1.08035135e+00 9.19353604e-01 1.76603627e+00 -4.97450829e-02
-3.01753610e-01 -1.64268777e-01 -7.83514678e-01 7.76513875e-01
6.87616467e-01 -5.86325191e-02 -4.77213413e-01 -5.14279544e-01
2.01179311e-02 3.02981615e-01 6.89261556e-01 -1.02820671e+00
3.59367251e-01 -3.83700579e-01 5.58450162e-01 -7.68644869e-01
3.84318709e-01 -6.15419149e-01 -5.45646489e-01 3.53416532e-01
-4.22736257e-01 -1.79879308e-01 -1.54428795e-01 5.44397593e-01
-6.24493837e-01 1.36463821e-01 7.76252329e-01 -1.90071985e-01
-1.12604916e+00 5.12187421e-01 -1.61650836e-01 -4.28447276e-02
9.98505235e-01 -5.61446428e-01 -3.81729335e-01 -3.77536982e-01
-3.76040429e-01 2.67621696e-01 7.00627387e-01 7.30330408e-01
8.40960503e-01 -1.39453650e+00 -2.96816021e-01 4.05431330e-01
1.74110159e-01 -2.32353806e-01 7.54016519e-01 1.22266829e+00
-1.03730053e-01 7.43497789e-01 -9.59142148e-01 -8.64770591e-01
-1.00229990e+00 9.72006202e-01 2.19180092e-01 3.08050424e-01
-2.18637556e-01 5.96643388e-01 9.12474751e-01 2.82830298e-01
-1.09729117e-05 -3.63341808e-01 -9.07203615e-01 -4.64823246e-02
6.04150772e-01 1.72659636e-01 -4.96993452e-01 -9.65363801e-01
-3.07315856e-01 8.60810935e-01 1.62056148e-01 7.73826838e-02
1.22142971e+00 -6.85864151e-01 -9.17619988e-02 6.13286912e-01
1.16012251e+00 -3.42014223e-01 -1.69861758e+00 -2.72000581e-01
-3.81579757e-01 -7.32744455e-01 1.05428040e-01 -3.59071314e-01
-1.21986151e+00 1.07981682e+00 5.40617168e-01 3.92751768e-02
1.40062153e+00 -4.20589074e-02 9.55186963e-01 -1.63746700e-01
5.96099257e-01 -1.04483306e+00 3.16653341e-01 3.97085249e-01
8.36827695e-01 -1.80100739e+00 -2.52349555e-01 -3.15111101e-01
-9.56677377e-01 9.29490924e-01 8.70899379e-01 -3.11950177e-01
4.59255815e-01 -5.60961306e-01 -1.34765923e-01 -1.88972518e-01
-6.32149577e-01 -5.67557514e-01 4.97005850e-01 7.47817039e-01
1.28966138e-01 -1.67462394e-01 -1.89715654e-01 6.41795218e-01
-4.37386706e-02 1.08893625e-01 4.34986591e-01 4.37502027e-01
-7.62428939e-01 -3.44348907e-01 -7.86615312e-02 4.08927828e-01
-2.50536084e-01 -3.24501693e-01 -2.22182497e-01 3.74781579e-01
1.51381627e-01 8.62395048e-01 1.45856991e-01 -4.70539063e-01
-1.06349133e-01 -3.30602258e-01 1.96476325e-01 -4.00335819e-01
-1.24424249e-01 3.13510969e-02 -4.13176596e-01 -8.38157177e-01
-9.53784168e-01 -7.13306725e-01 -1.12016165e+00 -1.45252119e-03
-1.07176535e-01 -8.20493251e-02 3.08940768e-01 9.31342900e-01
4.88575667e-01 4.82561529e-01 8.24422479e-01 -1.26722050e+00
3.43218148e-02 -8.56832922e-01 -5.48520982e-01 5.22059321e-01
7.77258396e-01 -1.05937958e+00 -5.55529475e-01 6.57380372e-02] | [9.771821022033691, -0.29674339294433594] |
58c8e2c6-4961-44f5-b885-9de160cb0285 | identity-preserving-pose-robust-face | 2111.10634 | null | https://arxiv.org/abs/2111.10634v1 | https://arxiv.org/pdf/2111.10634v1.pdf | Identity-Preserving Pose-Robust Face Hallucination Through Face Subspace Prior | Over the past few decades, numerous attempts have been made to address the problem of recovering a high-resolution (HR) facial image from its corresponding low-resolution (LR) counterpart, a task commonly referred to as face hallucination. Despite the impressive performance achieved by position-patch and deep learning-based methods, most of these techniques are still unable to recover identity-specific features of faces. The former group of algorithms often produces blurry and oversmoothed outputs particularly in the presence of higher levels of degradation, whereas the latter generates faces which sometimes by no means resemble the individuals in the input images. In this paper, a novel face super-resolution approach will be introduced, in which the hallucinated face is forced to lie in a subspace spanned by the available training faces. Therefore, in contrast to the majority of existing face hallucination techniques and thanks to this face subspace prior, the reconstruction is performed in favor of recovering person-specific facial features, rather than merely increasing image quantitative scores. Furthermore, inspired by recent advances in the area of 3D face reconstruction, an efficient 3D dictionary alignment scheme is also presented, through which the algorithm becomes capable of dealing with low-resolution faces taken in uncontrolled conditions. In extensive experiments carried out on several well-known face datasets, the proposed algorithm shows remarkable performance by generating detailed and close to ground truth results which outperform the state-of-the-art face hallucination algorithms by significant margins both in quantitative and qualitative evaluations. | ['Mohammad Rahmati', 'Ali Abbasi'] | 2021-11-20 | null | null | null | null | ['3d-face-reconstruction', 'face-hallucination', 'face-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.62680525e-01 2.54150301e-01 1.73787236e-01 -1.60445839e-01
-5.85788071e-01 -1.24156527e-01 7.76655436e-01 -4.33151096e-01
-9.66827050e-02 8.36334407e-01 3.35278481e-01 5.02849579e-01
-1.10953465e-01 -5.84765136e-01 -5.41718006e-01 -8.49592745e-01
1.14967883e-01 3.06132138e-01 -5.12184501e-01 -1.87698081e-01
1.07443873e-02 9.16496396e-01 -2.02732062e+00 1.28235295e-01
5.66722751e-01 9.11604047e-01 -2.48862877e-02 4.85944748e-02
7.98306093e-02 3.98110688e-01 -3.51115108e-01 -5.09360015e-01
3.25602740e-01 -5.19428313e-01 -5.44309855e-01 6.02471352e-01
8.13070714e-01 -2.57640600e-01 -3.78860086e-01 1.17513728e+00
5.42502761e-01 7.25694001e-02 6.20349944e-01 -8.51206899e-01
-8.83581698e-01 -1.31123289e-01 -7.37889171e-01 -6.93167076e-02
9.20794249e-01 -1.78246185e-01 4.46477711e-01 -1.46386576e+00
8.14485669e-01 1.51858139e+00 7.64498413e-01 6.42049730e-01
-1.72575450e+00 -5.87484837e-01 -3.25259238e-01 -6.19450510e-02
-1.65137279e+00 -9.29467320e-01 9.08648551e-01 -3.40622216e-01
6.22596681e-01 1.08812533e-01 3.34501982e-01 1.23469281e+00
-7.77506977e-02 2.42525935e-01 1.51104522e+00 -4.52113509e-01
2.48707179e-03 3.90662551e-01 -5.64338565e-01 6.11555874e-01
1.73736528e-01 3.30548435e-01 -6.12809300e-01 -2.24030823e-01
1.03193259e+00 -2.91014519e-02 -6.62893057e-01 -6.36438310e-01
-1.00269043e+00 5.99138141e-01 2.59334058e-01 4.76943702e-01
-7.71800876e-01 -5.78570962e-01 7.12642446e-02 2.02408582e-01
6.82985425e-01 2.92861074e-01 1.40527517e-01 4.16239738e-01
-1.20057833e+00 3.97371173e-01 6.12729490e-01 6.50879741e-01
8.12333226e-01 3.87311816e-01 -7.02894405e-02 9.89994526e-01
1.41017854e-01 3.07454258e-01 2.90643007e-01 -9.21726644e-01
4.16357368e-02 3.36625189e-01 2.55313098e-01 -1.30892682e+00
-5.93295060e-02 -5.17830849e-01 -1.24595296e+00 5.07605016e-01
2.83887118e-01 2.57477254e-01 -8.45912457e-01 1.76087809e+00
3.93617362e-01 1.70834944e-01 2.85443127e-01 1.01823878e+00
7.77275324e-01 4.05008376e-01 -9.07256901e-02 -7.67857075e-01
1.14879453e+00 -4.11581486e-01 -9.97736156e-01 9.19042677e-02
-4.04880553e-01 -9.06551600e-01 7.70826757e-01 5.02745569e-01
-1.33473694e+00 -9.12199497e-01 -8.58990312e-01 9.83049646e-02
-7.27704912e-02 -2.34610680e-02 1.31692976e-01 5.68166137e-01
-1.28615212e+00 6.43071413e-01 -2.86666155e-01 -4.97751564e-01
6.55965507e-01 2.29583040e-01 -1.07042301e+00 -3.60485435e-01
-8.49925458e-01 1.06483829e+00 3.24303806e-02 2.32007802e-01
-8.83551896e-01 -7.02949822e-01 -9.22753096e-01 4.83042262e-02
1.73747391e-01 -5.22353351e-01 6.44405067e-01 -1.05519342e+00
-1.31600249e+00 1.20154524e+00 -3.63822937e-01 -9.08189416e-02
6.74168646e-01 -1.13369927e-01 -7.17912376e-01 3.10435206e-01
2.20913347e-02 6.73421800e-01 1.47686315e+00 -1.75105655e+00
-1.26219705e-01 -8.11469495e-01 -4.09531116e-01 2.31518954e-01
-2.17731699e-01 1.59925804e-01 -3.20893914e-01 -6.79988027e-01
1.84449732e-01 -5.72650433e-01 9.21597481e-02 2.44658887e-01
-6.37860447e-02 1.34766236e-01 9.55599070e-01 -8.21376443e-01
6.44583941e-01 -2.21757650e+00 4.37184125e-01 1.22000007e-02
2.04842746e-01 4.96034920e-01 -1.63455307e-01 3.77159506e-01
-5.62556028e-01 -2.52373844e-01 -3.73372614e-01 -6.09663844e-01
-3.05886924e-01 1.03494301e-01 -3.67745429e-01 8.49241316e-01
3.83826673e-01 5.79247773e-01 -7.19693601e-01 -3.71364713e-01
3.65760326e-01 1.11200607e+00 -3.44157070e-01 3.65624636e-01
2.44490445e-01 8.09807479e-01 1.22008212e-02 6.73011422e-01
1.10733402e+00 2.50537544e-01 2.85681933e-02 -2.80270278e-01
-2.91078221e-02 -4.58273441e-01 -1.16250658e+00 1.62217271e+00
-2.67930061e-01 3.57919246e-01 4.40433592e-01 -9.96063948e-01
1.22057021e+00 8.18764865e-01 5.21172225e-01 -7.23666430e-01
-5.33451512e-02 1.41169637e-01 -4.69305933e-01 -5.01745820e-01
3.74143779e-01 -6.58285737e-01 4.53534216e-01 1.01157174e-01
2.60173529e-01 4.35067602e-02 -2.48283774e-01 -2.64001012e-01
5.40412843e-01 2.57117838e-01 4.22794104e-01 -1.26273438e-01
8.64764810e-01 -4.76191401e-01 3.77986550e-01 1.93796903e-01
-1.40004247e-01 9.69698071e-01 1.77745476e-01 -3.10391784e-01
-1.15304685e+00 -1.14928818e+00 -4.38266128e-01 4.78524357e-01
-1.15871906e-01 8.13604593e-02 -8.38340878e-01 -2.89378434e-01
9.38507030e-04 3.62812161e-01 -7.19386578e-01 -4.40579765e-02
-4.19403285e-01 -6.00268722e-01 4.24018770e-01 1.02896243e-01
6.67618871e-01 -1.23678398e+00 -2.94063896e-01 6.83760718e-02
-1.84319109e-01 -1.29174352e+00 -1.14514388e-01 -4.35087562e-01
-8.23533833e-01 -9.53576982e-01 -1.12545002e+00 -7.62052000e-01
9.03311789e-01 3.60618383e-01 8.78715932e-01 -6.07267767e-02
-4.95738804e-01 3.46342415e-01 -2.38371775e-01 8.82150233e-02
-4.88089293e-01 -7.06353724e-01 4.00452018e-01 6.08134568e-01
2.48845935e-01 -7.87230492e-01 -2.99254090e-01 1.40742093e-01
-1.00635397e+00 -1.73569649e-01 7.41637170e-01 1.07036626e+00
5.70145130e-01 3.50673527e-01 7.76425421e-01 -7.25848079e-01
5.00197768e-01 -3.62002492e-01 -3.26427668e-01 -7.69223422e-02
-6.65590823e-01 -1.50831208e-01 7.41046667e-01 -4.56845582e-01
-1.34839892e+00 1.72086626e-01 -1.70205131e-01 -8.87511373e-01
-5.25316834e-01 1.29887372e-01 -4.24373835e-01 -3.81845355e-01
7.34301031e-01 4.79382604e-01 5.92905462e-01 -7.00819552e-01
1.56819105e-01 5.82353830e-01 9.04148281e-01 -2.43711278e-01
1.08874226e+00 6.55363441e-01 9.52052698e-02 -1.09231722e+00
-6.49810016e-01 -2.75884926e-01 -8.19389582e-01 -3.11871469e-01
7.00239778e-01 -9.46738064e-01 -4.45809096e-01 4.61825579e-01
-9.41633523e-01 3.73496771e-01 -2.60215223e-01 1.82306021e-01
-7.15231299e-01 6.33173466e-01 -2.71798790e-01 -9.67026412e-01
-2.72199754e-02 -9.94826198e-01 1.07566977e+00 2.27458999e-01
-6.64711148e-02 -7.38533795e-01 -8.93970728e-02 5.65275073e-01
4.14774269e-01 5.63092768e-01 8.81112158e-01 -7.31977820e-02
-2.47591451e-01 -3.71328145e-01 -2.21785337e-01 6.61944866e-01
3.21169496e-01 -4.09972072e-01 -1.30804133e+00 -6.25796556e-01
2.85651624e-01 -2.18137264e-01 5.14196455e-01 1.45241946e-01
8.70870769e-01 -2.91539520e-01 -7.99112320e-02 5.80091774e-01
1.53193450e+00 -1.79627433e-01 9.37714159e-01 -3.69171873e-02
4.83914614e-01 9.89928603e-01 5.17767191e-01 4.34636652e-01
-8.39536935e-02 9.73204553e-01 5.28975546e-01 -3.17994416e-01
-4.83983815e-01 -3.27208072e-01 2.80892998e-01 5.41645959e-02
-4.52509314e-01 2.02387959e-01 -3.54789317e-01 5.06966829e-01
-1.53880012e+00 -1.30946457e+00 1.94761530e-01 2.45808434e+00
6.02026463e-01 -4.11482751e-01 1.10830702e-01 4.20856059e-01
7.79691219e-01 2.55326509e-01 -3.79178494e-01 -1.68327078e-01
-4.57247347e-01 3.19883853e-01 -1.21742979e-01 4.13729161e-01
-9.20250058e-01 7.80173898e-01 6.36194324e+00 6.29162848e-01
-1.04215026e+00 8.46228674e-02 4.33194041e-01 7.11930841e-02
-3.76061946e-02 -3.69124234e-01 -5.07519603e-01 1.66004479e-01
5.64538836e-01 -1.30701408e-01 6.31737471e-01 5.78278065e-01
2.97787279e-01 3.50536890e-02 -1.03547704e+00 1.34271097e+00
5.71962595e-01 -9.75518763e-01 2.41958052e-01 2.85606980e-01
7.20911324e-01 -6.83468699e-01 3.89345556e-01 6.17380626e-02
-4.30317014e-01 -1.57134283e+00 4.84043419e-01 8.29862058e-01
1.24508059e+00 -9.83199120e-01 6.65992439e-01 2.28354663e-01
-8.68116200e-01 -1.27435237e-01 -5.44472933e-01 5.14939614e-02
1.03147828e-03 4.84707296e-01 -6.48603261e-01 8.09477448e-01
7.44555652e-01 5.88647187e-01 -3.13779056e-01 7.14993954e-01
5.60583621e-02 5.23330942e-02 1.86516225e-01 8.20212185e-01
-2.33222365e-01 -3.77276450e-01 8.30742300e-01 8.63412976e-01
4.08921480e-01 2.05587298e-01 -2.06605449e-01 1.16463435e+00
-1.08866937e-01 6.51484430e-02 -9.52724159e-01 2.07225665e-01
1.92245334e-01 1.51654017e+00 -1.76636875e-01 2.58824159e-03
-4.43521410e-01 1.13749599e+00 2.77237236e-01 5.37927032e-01
-5.10513365e-01 9.51546580e-02 7.14104414e-01 5.29710531e-01
3.57467562e-01 7.46377325e-03 8.28579962e-02 -1.07491994e+00
1.85137406e-01 -1.06090438e+00 -1.56478658e-02 -7.69870877e-01
-1.28445506e+00 9.58307624e-01 3.16796191e-02 -1.33938360e+00
-4.63863581e-01 -3.41272622e-01 -2.36463174e-01 1.20115459e+00
-1.56478381e+00 -1.16820502e+00 -5.17657459e-01 8.38976026e-01
4.58875358e-01 -4.36216623e-01 1.15345705e+00 3.88818502e-01
-3.05597484e-01 4.98195648e-01 1.11577390e-02 -2.18929484e-01
7.70451486e-01 -7.73161292e-01 -1.44497588e-01 7.60124445e-01
3.10393833e-02 4.85821396e-01 1.03405643e+00 -4.24551517e-01
-1.36915183e+00 -1.09844744e+00 9.20952499e-01 -1.11032881e-01
2.00592428e-02 -2.20162719e-01 -1.22542477e+00 3.79326820e-01
1.16361409e-01 2.95570046e-01 5.19661009e-01 -1.70225665e-01
-4.72817361e-01 -9.78195295e-02 -1.65459001e+00 4.49464083e-01
9.65023994e-01 -5.92971265e-01 -7.37237394e-01 1.27972260e-01
-2.70601176e-02 -9.30339545e-02 -9.99415755e-01 5.59454739e-01
4.64262903e-01 -1.30261040e+00 1.23399198e+00 -4.00921315e-01
3.97805214e-01 -3.34105730e-01 -2.26723775e-01 -1.28719199e+00
-4.40225154e-01 -4.85146463e-01 -1.28277794e-01 1.27901936e+00
-1.74836278e-01 -4.32907045e-01 6.67707264e-01 2.35612735e-01
1.04114383e-01 -5.17593741e-01 -1.04238224e+00 -6.56712592e-01
-1.80327252e-01 2.59338945e-01 4.27150935e-01 1.00380421e+00
-3.60684633e-01 1.99017063e-01 -6.73877776e-01 1.90339074e-01
1.08008111e+00 1.74273878e-01 6.50761783e-01 -1.40864372e+00
2.28238292e-02 -2.76108444e-01 -5.46060085e-01 -4.50782955e-01
5.47189355e-01 -7.37616301e-01 -2.87901554e-02 -1.13518131e+00
2.51824319e-01 -4.50388575e-03 6.19111061e-02 2.05923051e-01
5.77059127e-02 8.18920314e-01 -9.94301308e-03 3.31159294e-01
-2.14495938e-02 7.87338376e-01 1.32569838e+00 2.02852219e-01
1.27246063e-02 -1.27569190e-03 -7.69010186e-01 6.51625633e-01
3.41739476e-01 -6.69390112e-02 -2.47556850e-01 -1.19815990e-02
-3.59432399e-01 4.57726687e-01 7.19255567e-01 -1.03932369e+00
-3.19399275e-02 1.94743723e-01 9.28276718e-01 -3.09294105e-01
8.56705666e-01 -8.80929887e-01 6.45259559e-01 1.11130863e-01
-1.56061530e-01 -1.95383117e-01 1.04163438e-01 6.27524197e-01
-5.18647373e-01 9.96118262e-02 1.33202577e+00 -2.62224257e-01
-5.45284271e-01 3.64102513e-01 -1.52936742e-01 -3.83279920e-01
1.01343143e+00 -6.27281547e-01 6.59156069e-02 -5.16577423e-01
-1.01221645e+00 -5.00602067e-01 7.57521093e-01 4.62601840e-01
8.95262241e-01 -1.40321779e+00 -1.04836583e+00 6.69686913e-01
7.85805956e-02 -2.24584147e-01 5.38896859e-01 8.46274853e-01
-1.10522337e-01 2.90069669e-01 -7.15748489e-01 -4.10018474e-01
-1.40570188e+00 8.17409515e-01 4.23254550e-01 1.13647074e-01
-8.20040345e-01 4.96515453e-01 3.37774426e-01 -3.09637815e-01
2.26465240e-01 5.83586037e-01 -5.21704435e-01 1.38484165e-01
7.37138748e-01 2.62486666e-01 1.67463809e-01 -1.34972572e+00
-1.62354603e-01 6.51726246e-01 4.44386788e-02 -5.43408059e-02
1.48742759e+00 -2.40819067e-01 -2.94238776e-01 6.12308495e-02
1.17188549e+00 3.78524140e-02 -1.26826048e+00 -3.64223838e-01
-2.25225627e-01 -9.83513117e-01 1.08826429e-01 -5.79370022e-01
-1.11269450e+00 6.84814990e-01 7.45236695e-01 2.58798944e-04
1.52858651e+00 -1.13583006e-01 3.27999741e-01 -1.83062434e-01
5.18704593e-01 -6.44962013e-01 1.09735675e-01 9.56162885e-02
1.27663743e+00 -1.27738893e+00 1.98976714e-02 -5.04998684e-01
-4.78488296e-01 1.02987826e+00 3.77829045e-01 -2.51872927e-01
4.10995185e-01 1.05110422e-01 -1.53990477e-01 -1.24921881e-01
-4.24573809e-01 -1.58923909e-01 2.53292233e-01 8.91215026e-01
3.23973566e-01 -2.17318699e-01 -1.22445971e-02 1.48847505e-01
-1.29750460e-01 1.01012975e-01 4.54973012e-01 4.42053020e-01
-2.40891740e-01 -9.21129823e-01 -8.76608908e-01 1.10110581e-01
-4.70445514e-01 2.39333034e-01 -2.15510100e-01 8.60491812e-01
3.95279974e-02 9.05080557e-01 -2.49615833e-02 -1.30049676e-01
4.99441594e-01 1.42964095e-01 7.66791224e-01 -4.68092173e-01
-2.27547020e-01 1.99467897e-01 -1.41698182e-01 -6.28810108e-01
-6.58577681e-01 -8.01494896e-01 -7.74375260e-01 -3.48165393e-01
-4.77355942e-02 -8.39917064e-02 4.34726238e-01 6.79113448e-01
2.55100042e-01 1.68486521e-01 8.18165600e-01 -1.40124035e+00
-5.72995305e-01 -9.11694884e-01 -9.71834123e-01 6.68789744e-01
5.47450423e-01 -9.60503697e-01 -4.29620951e-01 1.30270526e-01] | [12.845039367675781, -0.025765951722860336] |
1a13e854-fd43-4deb-9df2-b2b9e516cfc2 | discrete-attacks-and-submodular-optimization | 1812.00151 | null | http://arxiv.org/abs/1812.00151v2 | http://arxiv.org/pdf/1812.00151v2.pdf | Discrete Adversarial Attacks and Submodular Optimization with Applications to Text Classification | Adversarial examples are carefully constructed modifications to an input that
completely change the output of a classifier but are imperceptible to humans.
Despite these successful attacks for continuous data (such as image and audio
samples), generating adversarial examples for discrete structures such as text
has proven significantly more challenging. In this paper we formulate the
attacks with discrete input on a set function as an optimization task. We prove
that this set function is submodular for some popular neural network text
classifiers under simplifying assumption. This finding guarantees a $1-1/e$
approximation factor for attacks that use the greedy algorithm. Meanwhile, we
show how to use the gradient of the attacked classifier to guide the greedy
search. Empirical studies with our proposed optimization scheme show
significantly improved attack ability and efficiency, on three different text
classification tasks over various baselines. We also use a joint sentence and
word paraphrasing technique to maintain the original semantics and syntax of
the text. This is validated by a human subject evaluation in subjective metrics
on the quality and semantic coherence of our generated adversarial text. | ['Pin-Yu Chen', 'Qi Lei', 'Michael Witbrock', 'Lingfei Wu', 'Inderjit S. Dhillon', 'Alexandros G. Dimakis'] | 2018-12-01 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 6.95151269e-01 3.51528049e-01 2.40949303e-01 -5.01629472e-01
-1.03997791e+00 -1.17613459e+00 6.28424585e-01 1.11476935e-01
-4.80397493e-01 7.50606716e-01 1.05820917e-01 -2.90893763e-01
2.67512649e-01 -6.88003480e-01 -1.11642516e+00 -6.98625803e-01
8.95310715e-02 7.54814520e-02 -1.31452590e-01 -4.11904752e-01
2.58854419e-01 4.32207853e-01 -9.23531890e-01 4.07914132e-01
6.05830073e-01 9.25763369e-01 -3.17888647e-01 9.88916039e-01
8.02896693e-02 7.39625454e-01 -1.19542396e+00 -9.62714374e-01
7.44375885e-01 -3.82137239e-01 -7.81742990e-01 -8.86720270e-02
8.17218125e-01 -3.28343004e-01 -7.89517462e-01 1.59408402e+00
5.27108371e-01 8.92517194e-02 6.93943381e-01 -1.55396831e+00
-8.37416530e-01 1.01163268e+00 -1.91706493e-01 -5.07595092e-02
3.44473541e-01 4.49992716e-02 1.11378288e+00 -6.42432272e-01
3.32132787e-01 1.40362608e+00 6.03549123e-01 7.39487708e-01
-1.16122103e+00 -1.05478358e+00 1.51796848e-01 7.66182542e-02
-1.31835818e+00 -3.23255718e-01 9.10482049e-01 -1.48459310e-02
7.74311185e-01 7.90026546e-01 6.52795807e-02 1.41691482e+00
4.30017114e-01 6.31243825e-01 1.00127316e+00 -3.96519631e-01
2.71095425e-01 5.81380069e-01 -1.81033742e-02 7.00394869e-01
3.33501548e-01 -6.09770343e-02 -4.22446728e-01 -5.28378069e-01
8.82477760e-02 -2.49689799e-02 -5.46097219e-01 -1.08339205e-01
-8.14965785e-01 1.10254347e+00 4.18025851e-01 2.75446791e-02
-7.18990387e-03 5.32893479e-01 5.49009204e-01 8.11168551e-01
3.56097311e-01 6.29680336e-01 -2.41694093e-01 2.15897858e-01
-6.70518875e-01 3.45254809e-01 1.04769957e+00 1.02715051e+00
2.12223336e-01 3.90829086e-01 -4.59896177e-02 7.20014572e-01
-1.38316760e-02 9.24602389e-01 5.68061769e-01 -6.36393666e-01
8.07569742e-01 1.48774266e-01 3.28098647e-02 -1.40482771e+00
-7.15710670e-02 -3.68224025e-01 -9.29460704e-01 2.93978721e-01
4.56250519e-01 -3.20421398e-01 -6.16730034e-01 1.99166453e+00
1.55379295e-01 -1.55546844e-01 2.32931003e-01 6.54717505e-01
3.72893065e-01 7.45970249e-01 -1.92003638e-01 -6.88213632e-02
1.13664091e+00 -7.08863318e-01 -7.10556805e-01 -3.07989717e-01
6.84209049e-01 -8.21544349e-01 1.32931042e+00 4.58523959e-01
-9.85699356e-01 -2.11884663e-01 -1.46184015e+00 1.06530644e-01
-5.26365817e-01 -2.84632236e-01 1.89172179e-01 1.25741220e+00
-6.66436970e-01 5.23459733e-01 -4.57064122e-01 2.94421259e-02
4.18727964e-01 5.94832361e-01 -3.95149827e-01 7.01285005e-02
-1.59258318e+00 6.98044956e-01 3.37690175e-01 -1.29978821e-01
-9.06843007e-01 -6.53670490e-01 -8.55619907e-01 1.81606077e-02
1.27638057e-01 -4.65588778e-01 1.26122928e+00 -1.22531688e+00
-1.54323840e+00 8.38670135e-01 2.67977536e-01 -9.56109703e-01
9.25448537e-01 -3.22669953e-01 -2.14469805e-01 2.74819165e-01
-3.36421937e-01 4.69886124e-01 1.43472123e+00 -1.08203197e+00
-3.92138481e-01 -2.70995766e-01 3.16551447e-01 1.85780451e-01
-8.85195732e-01 4.74389121e-02 2.11519673e-01 -1.31397080e+00
-2.08928630e-01 -1.10599422e+00 -2.03210101e-01 2.69826889e-01
-6.80766761e-01 2.18996242e-01 1.12351823e+00 -7.57036805e-01
1.11482286e+00 -2.16735125e+00 1.10805921e-01 3.66627574e-01
1.52317524e-01 2.00076282e-01 -3.82866673e-02 5.21340311e-01
-3.11730355e-01 4.26097572e-01 -4.57750857e-01 -3.95519197e-01
2.60991424e-01 -6.11930713e-02 -1.02819157e+00 8.78618181e-01
-5.58520742e-02 7.03331113e-01 -5.02391100e-01 -2.02869698e-01
-3.13501656e-02 1.95872232e-01 -6.79015279e-01 2.49891758e-01
-3.27939540e-01 -2.61074394e-01 -3.75699669e-01 2.12518439e-01
6.93118095e-01 -9.29783285e-03 -7.36541227e-02 4.13251901e-03
8.04505706e-01 1.81135107e-02 -1.08381248e+00 1.40198934e+00
-4.97864157e-01 8.97843778e-01 8.15651119e-02 -1.11228704e+00
7.74690986e-01 3.35587680e-01 6.44828975e-02 -1.58987731e-01
2.64222145e-01 -1.41662993e-02 -8.07995498e-02 -1.55427605e-01
3.89814287e-01 -1.39432609e-01 -4.83997613e-01 6.78098738e-01
-2.03202561e-01 -3.94426942e-01 -4.66806740e-01 4.54382837e-01
1.19280553e+00 -5.27453482e-01 2.56199032e-01 -1.65029779e-01
5.55773318e-01 -1.49938896e-01 -9.65835352e-04 1.16047716e+00
-2.24557221e-01 4.87301677e-01 5.81399858e-01 -1.98794395e-01
-1.15682685e+00 -9.13694918e-01 -2.11323481e-02 1.03832984e+00
1.59218500e-03 -3.00345600e-01 -1.19037604e+00 -1.02282524e+00
3.59859318e-02 9.26926851e-01 -6.99481249e-01 -6.25275373e-01
-5.41214406e-01 -6.83203161e-01 1.23028135e+00 3.25481206e-01
6.76476717e-01 -7.84409046e-01 -1.21821478e-01 -2.37396522e-03
-4.64464873e-02 -1.28314126e+00 -9.73122478e-01 1.27341717e-01
-4.58445638e-01 -8.18204165e-01 -5.19712806e-01 -9.45666194e-01
7.15107679e-01 -1.27033964e-02 6.07292354e-01 -3.80231999e-02
-1.54194310e-01 3.46378028e-01 -2.65490919e-01 -5.89356244e-01
-9.52709913e-01 -6.14036508e-02 1.80332884e-01 1.65286750e-01
7.24773854e-02 -5.32858193e-01 -3.51623744e-01 2.70683616e-01
-1.32865310e+00 -2.75784284e-01 1.88725725e-01 1.06542528e+00
1.28564090e-01 2.49171749e-01 5.53844094e-01 -9.07815695e-01
9.55091655e-01 -2.36887217e-01 -6.36226177e-01 1.62391722e-01
-3.68405759e-01 1.86514556e-01 1.30680501e+00 -8.56405318e-01
-6.14134431e-01 -8.18918049e-02 -7.75389150e-02 -5.39597511e-01
1.81976214e-01 8.63059461e-02 -3.51345092e-01 -3.21559817e-01
1.08533013e+00 3.71008873e-01 6.49497891e-03 6.20826147e-02
4.05695885e-01 9.99670804e-01 4.76931304e-01 -5.14495492e-01
1.41756678e+00 5.86651325e-01 -3.45386937e-02 -8.58021498e-01
-6.79429173e-01 1.48833007e-01 -1.01192355e-01 1.68253541e-01
4.74232882e-01 -5.97382545e-01 -8.44159067e-01 5.99518955e-01
-1.27117860e+00 -3.53514887e-02 -1.74179539e-01 1.41774863e-01
-5.52459478e-01 6.57050788e-01 -6.38243139e-01 -7.40426481e-01
-6.96302831e-01 -1.02869618e+00 8.43329370e-01 -2.81404555e-01
-2.15142950e-01 -1.08203769e+00 -2.85276771e-01 2.47621700e-01
2.09924415e-01 4.32944149e-01 9.61912036e-01 -1.28625059e+00
-3.22355866e-01 -7.37163842e-01 1.43327653e-01 8.51345360e-01
7.09187537e-02 -2.66136140e-01 -1.09623921e+00 -6.37564600e-01
5.21953881e-01 -5.12534559e-01 5.96093535e-01 -1.64616108e-01
1.40063620e+00 -1.02838850e+00 1.37062687e-02 7.65335441e-01
1.19230878e+00 2.04305798e-01 4.66822773e-01 1.74969941e-01
7.20080018e-01 4.88072783e-01 3.37254673e-01 3.65578443e-01
-4.46533769e-01 4.96573180e-01 4.16153312e-01 1.05841614e-01
2.65784919e-01 -3.45374316e-01 8.54709268e-01 5.40678442e-01
8.11435759e-01 -6.82383060e-01 -6.53031409e-01 1.39673442e-01
-1.51174664e+00 -9.79596376e-01 3.74399602e-01 2.13215208e+00
1.09039497e+00 6.33884847e-01 -2.11784214e-01 5.65274000e-01
8.50238621e-01 2.37532333e-01 -6.05853140e-01 -7.32337654e-01
-3.72766048e-01 3.99439931e-01 8.31623793e-01 7.19889581e-01
-1.13713634e+00 1.03310668e+00 6.21005726e+00 1.12261260e+00
-1.12253857e+00 -5.45730218e-02 6.96521878e-01 -2.75004834e-01
-2.94269353e-01 -2.91311949e-01 -5.42537808e-01 4.48492497e-01
8.81793499e-01 -6.05284929e-01 6.19848788e-01 9.13408160e-01
-4.86441106e-02 6.23192012e-01 -1.30123568e+00 8.81500363e-01
2.99368531e-01 -1.23585570e+00 4.43374842e-01 -1.60641581e-01
6.13696694e-01 -3.79300356e-01 5.06867409e-01 1.22080550e-01
5.01672685e-01 -1.31992245e+00 7.36927867e-01 -1.09043509e-01
8.13237369e-01 -9.46625471e-01 4.80436683e-01 2.59941876e-01
-7.32751846e-01 -1.38787940e-01 -4.44745481e-01 2.13339090e-01
-3.36494386e-01 9.56528634e-02 -1.02366590e+00 1.21669009e-01
3.29413176e-01 2.12457329e-01 -4.99335706e-01 3.58027577e-01
-1.31797194e-01 7.03226984e-01 -4.48092312e-01 -4.41855699e-01
3.64097565e-01 1.61197647e-01 9.31177139e-01 1.32967949e+00
-1.41004354e-01 -4.53054570e-02 1.66989826e-02 7.24888742e-01
-5.39165378e-01 2.15149298e-01 -9.19393182e-01 -1.66433662e-01
5.58350742e-01 8.28745306e-01 -3.20330977e-01 -2.12090254e-01
-4.02667820e-02 1.35792339e+00 1.57313377e-01 5.19208193e-01
-1.14724803e+00 -9.86145854e-01 7.59091020e-01 -2.35973597e-01
1.73503980e-01 -7.74317384e-02 -4.29076374e-01 -1.02012074e+00
3.95180851e-01 -1.44380820e+00 4.32974100e-01 -5.40438712e-01
-1.23784471e+00 6.56949639e-01 -1.69648051e-01 -9.70694244e-01
-2.44390994e-01 -6.53254688e-01 -3.84711355e-01 8.13879192e-01
-7.90075719e-01 -9.46468949e-01 6.73959926e-02 9.06117916e-01
6.66141033e-01 -5.08203864e-01 9.74082530e-01 -1.15305401e-01
-4.40968543e-01 1.50665736e+00 2.98111498e-01 4.68444169e-01
6.27468646e-01 -1.22979784e+00 6.56269968e-01 9.88750696e-01
2.55110472e-01 4.57564533e-01 1.09284854e+00 -4.71863449e-01
-1.49063635e+00 -1.19262576e+00 4.19837266e-01 -4.03013915e-01
8.41733217e-01 -8.11516345e-01 -8.22102964e-01 8.60099375e-01
2.19762936e-01 -1.09417021e-01 4.97815639e-01 -5.01742840e-01
-8.45608115e-01 -2.09963787e-02 -1.55521786e+00 9.48337197e-01
9.59040165e-01 -8.37545335e-01 -5.42279840e-01 7.30088234e-01
1.13265538e+00 -4.85486031e-01 -6.73507273e-01 1.28424510e-01
3.95186812e-01 -5.02597928e-01 9.97880042e-01 -1.18868601e+00
4.51107234e-01 3.39044519e-02 -7.16581941e-01 -1.35467100e+00
1.91979200e-01 -1.27515078e+00 -4.36726287e-02 8.67568851e-01
4.64930266e-01 -9.15784359e-01 8.62973154e-01 5.20600021e-01
1.40686721e-01 -4.74927574e-01 -1.05482388e+00 -8.26870143e-01
5.26924074e-01 -3.87853563e-01 3.54299545e-01 9.85322118e-01
8.88915434e-02 1.91561088e-01 -5.05258620e-01 5.04188418e-01
7.55804837e-01 -3.35896224e-01 8.05641115e-01 -5.62616169e-01
-4.33625996e-01 -3.33837062e-01 -6.76009297e-01 -9.26017821e-01
5.80295026e-01 -1.08716667e+00 -1.84891000e-01 -6.59044266e-01
-1.56708062e-01 -1.21245146e-01 -2.66226474e-02 3.81585062e-01
-2.16849104e-01 1.30164638e-01 2.51863033e-01 5.59919849e-02
-9.45669711e-02 4.98815268e-01 8.79723847e-01 -5.42206287e-01
2.08309263e-01 6.22151680e-02 -7.96390712e-01 6.35562658e-01
9.71030295e-01 -8.83819461e-01 -5.06597817e-01 -3.44139218e-01
1.64331034e-01 -2.22280119e-02 3.67048800e-01 -9.12063241e-01
1.39863938e-01 -1.05324529e-01 1.61951259e-01 -6.77214786e-02
4.80581939e-01 -1.02794445e+00 -2.11446568e-01 7.16261804e-01
-1.02119279e+00 9.13048163e-02 2.73457885e-01 8.40747118e-01
-5.07916920e-02 -4.43100303e-01 1.13130677e+00 2.23922223e-01
1.84078574e-01 2.22850531e-01 -2.31059283e-01 4.10895973e-01
1.11407495e+00 -2.29256507e-02 -3.64418864e-01 -8.49313796e-01
-5.84105968e-01 -8.21958557e-02 2.71601617e-01 4.33022112e-01
6.69539452e-01 -1.16354477e+00 -8.57773185e-01 1.50874987e-01
-1.88267037e-01 -4.14046317e-01 -1.28952712e-01 8.31691101e-02
-5.90689421e-01 3.43639642e-01 7.55205676e-02 -2.72034049e-01
-1.54582524e+00 6.75891578e-01 6.86278462e-01 -2.08246842e-01
-4.86802876e-01 8.95670235e-01 2.30165541e-01 -1.42238811e-01
7.31105924e-01 -1.67956188e-01 3.61327738e-01 -3.46454591e-01
4.88796830e-01 9.83092263e-02 6.20390736e-02 -2.54819989e-01
-9.84942839e-02 1.78599417e-01 -3.76883715e-01 -3.59407961e-01
8.60273480e-01 3.20926458e-02 1.71126157e-01 1.23897694e-01
1.74206924e+00 2.88027376e-01 -8.38570416e-01 -3.98365825e-01
-3.22250158e-01 -4.77438390e-01 -1.74316734e-01 -5.72303474e-01
-8.12857032e-01 7.45466709e-01 6.04076326e-01 4.99859840e-01
9.19610381e-01 -5.17406642e-01 1.00582874e+00 1.07506526e+00
7.15392306e-02 -9.01523948e-01 2.63263315e-01 3.45404893e-01
1.18716764e+00 -9.73782122e-01 -1.73105478e-01 -5.55746615e-01
-5.41726708e-01 9.32119906e-01 3.94189030e-01 -3.28990251e-01
6.79792941e-01 4.01368320e-01 -5.54292183e-03 2.64983296e-01
-6.85776353e-01 6.03127778e-01 9.95541960e-02 5.22623122e-01
-3.61530222e-02 -1.30457252e-01 -1.11712299e-01 4.29810584e-01
-8.35690141e-01 -6.12756371e-01 7.03904212e-01 9.16562915e-01
-4.32244390e-01 -9.49824214e-01 -6.64768517e-01 2.78993338e-01
-1.01392031e+00 -3.73719275e-01 -7.37078369e-01 6.12486362e-01
-5.84347129e-01 1.00731063e+00 -1.51720375e-01 -5.45618594e-01
2.33049631e-01 1.30927786e-01 4.00895804e-01 -4.77330655e-01
-9.08701837e-01 -5.05660355e-01 1.23279288e-01 -1.87923804e-01
2.59135455e-01 -4.19810027e-01 -1.06701076e+00 -5.02817154e-01
-3.10711384e-01 2.24296421e-01 8.75274181e-01 6.51952028e-01
9.09025446e-02 1.59602240e-01 1.21807206e+00 -2.98470825e-01
-1.57411134e+00 -9.26004350e-01 -4.66576219e-01 7.80539930e-01
4.47598636e-01 -1.13771178e-01 -9.34775352e-01 1.63024873e-01] | [5.958797454833984, 8.00013542175293] |
58a8b229-d3a2-4b5a-8ea3-977ff875dfa7 | personalized-multi-faceted-trust-modeling-to | 2111.06440 | null | https://arxiv.org/abs/2111.06440v1 | https://arxiv.org/pdf/2111.06440v1.pdf | Personalized multi-faceted trust modeling to determine trust links in social media and its potential for misinformation management | In this paper, we present an approach for predicting trust links between peers in social media, one that is grounded in the artificial intelligence area of multiagent trust modeling. In particular, we propose a data-driven multi-faceted trust modeling which incorporates many distinct features for a comprehensive analysis. We focus on demonstrating how clustering of similar users enables a critical new functionality: supporting more personalized, and thus more accurate predictions for users. Illustrated in a trust-aware item recommendation task, we evaluate the proposed framework in the context of a large Yelp dataset. We then discuss how improving the detection of trusted relationships in social media can assist in supporting online users in their battle against the spread of misinformation and rumours, within a social networking environment which has recently exploded in popularity. We conclude with a reflection on a particularly vulnerable user base, older adults, in order to illustrate the value of reasoning about groups of users, looking to some future directions for integrating known preferences with insights gained through data analysis. | ['Queenie Chen', 'Gaurav Sahu', 'Xueguang Ma', 'Robin Cohen', 'Alexandre Parmentier'] | 2021-11-11 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-4.62631464e-01 6.60129249e-01 -3.14132690e-01 -7.88904607e-01
4.22970019e-02 6.46011829e-02 3.18089008e-01 9.70067978e-01
-3.91481698e-01 8.02712083e-01 6.52705431e-01 7.48972148e-02
-6.18662059e-01 -6.82127416e-01 -2.53302246e-01 -1.21534623e-01
-5.88859022e-01 6.46912932e-01 1.14008702e-01 -6.35083079e-01
2.92750120e-01 -3.04372851e-02 -1.57371652e+00 4.80324447e-01
8.82262588e-01 8.35099876e-01 -2.79068798e-01 1.32442802e-01
5.57651281e-01 1.03655696e+00 -7.11987913e-02 -1.33624840e+00
-2.09252626e-01 1.13699034e-01 -8.95545363e-01 -3.87747377e-01
-1.22231096e-01 -6.85956299e-01 4.57825005e-01 6.34401023e-01
3.65762144e-01 4.19170707e-02 8.03972661e-01 -1.53366685e+00
-5.27571499e-01 1.08055687e+00 -4.00932990e-02 -2.36044809e-01
9.02752697e-01 -7.90568173e-01 1.39607561e+00 -5.67274570e-01
5.01720548e-01 1.11824858e+00 9.98481691e-01 4.57112104e-01
-9.11472023e-01 -3.93734783e-01 4.71612304e-01 3.99489969e-01
-9.18881774e-01 -2.94126242e-01 6.50341451e-01 -3.31999660e-01
7.19079912e-01 3.92333597e-01 9.94078815e-01 1.31494689e+00
8.75221565e-02 8.44336987e-01 9.63649511e-01 -2.79447645e-01
5.43271363e-01 7.64807999e-01 5.06542981e-01 6.24469459e-01
6.27360165e-01 -5.07159717e-02 -9.91221666e-01 -9.84359384e-01
1.61596075e-01 1.03425175e-01 3.72784287e-02 -5.44149876e-01
-7.70786822e-01 1.24306214e+00 4.93541598e-01 3.71530414e-01
-5.21743894e-01 -4.90456790e-01 2.49207377e-01 6.91684842e-01
1.01496077e+00 4.22416538e-01 -5.14853716e-01 -1.37418106e-01
-8.48750412e-01 2.01335490e-01 1.57417786e+00 6.69130385e-01
4.40016478e-01 -6.34261429e-01 4.81010854e-01 8.37726057e-01
1.14310634e+00 -8.65734369e-02 2.04289734e-01 -1.01177430e+00
-1.67731747e-01 5.05005240e-01 2.54306346e-01 -1.19163966e+00
-7.13001788e-01 -3.42324942e-01 -4.56668258e-01 9.87289026e-02
1.91580147e-01 -3.52640718e-01 3.99956614e-01 1.42486441e+00
5.95742524e-01 7.97424763e-02 8.40931907e-02 4.83974397e-01
5.78718126e-01 -3.05023193e-01 -3.54235694e-02 -5.45944154e-01
1.22232378e+00 -7.64926493e-01 -5.95614672e-01 1.64362527e-02
6.83627248e-01 -3.92622411e-01 2.90762156e-01 8.47706497e-01
-1.23116970e+00 2.03094278e-02 -1.14515746e+00 2.62351662e-01
-3.00155461e-01 -5.44174790e-01 9.50949550e-01 1.02698970e+00
-1.23913038e+00 9.28354979e-01 -6.67480469e-01 -9.02210355e-01
2.78463095e-01 5.05455792e-01 -3.22993040e-01 -3.24818306e-02
-1.56059778e+00 1.25042105e+00 -1.47915840e-01 -1.56542405e-01
-5.71036518e-01 -4.28365827e-01 -6.54864132e-01 -6.84702843e-02
1.36203557e-01 -8.19886565e-01 9.67115581e-01 -1.11038172e+00
-1.23442793e+00 6.48963571e-01 2.77217060e-01 -5.54217458e-01
6.11589789e-01 -4.44302976e-01 -7.11265326e-01 4.25325520e-02
1.66121319e-01 6.83172867e-02 7.39693046e-01 -1.30506611e+00
-6.26041830e-01 -6.80986524e-01 2.97022939e-01 1.78252637e-01
-8.32257032e-01 4.12269026e-01 3.36623162e-01 -3.22682709e-01
-1.20391056e-01 -6.63522243e-01 -2.37078771e-01 1.79730486e-02
1.46009186e-02 -2.28624254e-01 1.32389009e-01 -2.92317659e-01
1.27511156e+00 -1.59644067e+00 7.78107941e-02 7.80731499e-01
4.61001873e-01 -1.19074352e-01 1.73853725e-01 1.10679591e+00
3.35773677e-01 2.13219225e-01 3.52306396e-01 -8.51198435e-01
-7.20588025e-03 1.88063860e-01 2.18504071e-01 3.72691661e-01
-3.16402048e-01 3.14914256e-01 -1.04374862e+00 -4.04304653e-01
-4.12684739e-01 4.74398345e-01 -7.42133379e-01 2.18797430e-01
-1.01062912e-03 8.80623236e-02 -7.92683780e-01 5.54089546e-01
2.88132221e-01 -4.10369039e-01 6.36525452e-01 6.79654032e-02
-2.85189543e-02 9.37761888e-02 -9.83047545e-01 1.25367534e+00
-2.05868512e-01 -2.09175512e-01 3.15908521e-01 -6.19748533e-01
8.80986094e-01 4.95874614e-01 7.10232735e-01 -5.77329248e-02
5.74455857e-01 2.69946635e-01 -8.83969963e-02 -4.58225071e-01
5.55662453e-01 -8.85708444e-03 8.63780156e-02 1.18670857e+00
-1.09314740e-01 2.90693551e-01 -2.85363913e-01 7.16256917e-01
8.60685885e-01 -1.43594071e-01 6.05887175e-01 -4.88851547e-01
2.47142226e-01 -2.95880318e-01 1.21119589e-01 7.25278199e-01
-4.70944524e-01 1.65190399e-01 4.87426013e-01 -4.32668954e-01
-6.87158406e-01 -6.49690688e-01 -9.22074690e-02 1.30116343e+00
-2.80891992e-02 -7.13214695e-01 -4.74315524e-01 -7.19005525e-01
2.53797948e-01 6.57019317e-01 -1.02927136e+00 -2.00524345e-01
1.16364568e-01 -6.51659429e-01 2.94149309e-01 5.11609353e-02
1.15986206e-01 -8.57247710e-01 -4.29645449e-01 3.63085061e-01
-3.15915823e-01 -8.53936672e-01 9.86006409e-02 -1.19171634e-01
-9.78658199e-01 -1.15182781e+00 -5.44509947e-01 -5.60397685e-01
4.61173356e-01 1.58828169e-01 1.29087591e+00 8.54896665e-01
3.49041730e-01 1.10361493e+00 -9.82765377e-01 -5.86677074e-01
-4.02703494e-01 6.52725995e-03 7.29889452e-01 2.02990860e-01
4.33780342e-01 -1.00785339e+00 -6.35682225e-01 6.82461739e-01
-5.47790289e-01 -1.89569563e-01 -4.76894267e-02 4.74986523e-01
-4.68218058e-01 -2.93962210e-01 1.07505047e+00 -1.13426161e+00
1.01517713e+00 -1.35074568e+00 3.17048341e-01 3.90218258e-01
-1.27530515e+00 -2.91421831e-01 -3.56416404e-02 -9.43985730e-02
-9.30095136e-01 -4.52674478e-01 -1.75489858e-01 5.22952080e-01
2.66091675e-01 8.88284564e-01 3.53311181e-01 -1.63255006e-01
8.60879779e-01 -4.93798643e-01 4.59023654e-01 -3.49193126e-01
3.42060298e-01 8.87824655e-01 -2.90445805e-01 -6.20386899e-01
2.85990208e-01 6.17306471e-01 -4.85596657e-01 -5.96410930e-01
-9.17148113e-01 -5.09856343e-01 -7.58051395e-01 -6.00313544e-01
4.87123787e-01 -8.84959221e-01 -1.16990030e+00 3.48713964e-01
-8.38289917e-01 -4.98255268e-02 2.10259438e-01 5.83928049e-01
-5.44167221e-01 4.15886790e-01 -6.56280160e-01 -1.13642418e+00
-4.66594338e-01 -4.30755019e-01 8.91135111e-02 1.37899136e-02
-6.32317603e-01 -1.37938511e+00 4.18934226e-01 8.88356507e-01
4.78864461e-01 -1.13542028e-01 5.59424460e-01 -1.32213318e+00
9.55927074e-02 -1.92010865e-01 1.65965185e-01 2.64501929e-01
1.03036184e-02 6.28778040e-02 -8.06487441e-01 -4.30832177e-01
1.34370878e-01 -7.72243798e-01 2.86948621e-01 -4.10526432e-02
3.20321545e-02 -2.61648536e-01 -3.66436869e-01 -3.50170612e-01
9.63180780e-01 -3.98362786e-01 2.98888713e-01 4.14107829e-01
2.18355015e-01 1.03314674e+00 9.78998363e-01 1.19178891e+00
1.31341064e+00 4.58696246e-01 8.09719801e-01 2.58575141e-01
7.37328053e-01 -7.25010037e-02 3.64137560e-01 6.13435209e-01
-5.88732958e-01 -9.84654762e-03 -5.41836262e-01 3.37157875e-01
-2.33800507e+00 -7.61807978e-01 -1.04256019e-01 2.02842879e+00
5.77519834e-01 8.97488967e-02 6.79858983e-01 2.15642199e-01
5.46508133e-01 -1.27379611e-01 -3.38952959e-01 -6.36227429e-01
2.85144269e-01 -1.02243699e-01 -2.79045254e-02 6.14425063e-01
-5.83797574e-01 6.40781462e-01 6.41666889e+00 -1.22218154e-01
-3.06503862e-01 2.66473502e-01 6.50788784e-01 -3.09329294e-02
-6.29300594e-01 -2.18591422e-01 -5.46739578e-01 4.73193638e-03
1.01423323e+00 -9.82776880e-02 3.25051725e-01 7.93312252e-01
3.30748022e-01 -3.29403162e-01 -1.24767077e+00 2.88246959e-01
3.29668254e-01 -8.90618026e-01 -2.94634223e-01 -4.73085642e-02
4.96038854e-01 1.51142418e-01 -1.83149308e-01 1.04751319e-01
6.08779013e-01 -4.65245873e-01 6.31521106e-01 7.25181758e-01
-2.50252575e-01 -7.02366948e-01 7.43787944e-01 4.68899637e-01
-5.02757311e-01 -3.42105687e-01 -3.13989967e-01 -7.11602032e-01
3.72293115e-01 6.26277387e-01 -8.18485677e-01 6.36372566e-01
8.93139422e-01 1.14115548e+00 -2.84919888e-01 9.65016365e-01
1.16722241e-01 1.88687503e-01 -2.46121913e-01 -4.20412123e-01
-1.16715752e-01 -1.06128253e-01 1.12044431e-01 7.14396596e-01
1.91040114e-01 3.60417962e-01 -1.10507995e-01 2.99834639e-01
2.92246759e-01 4.64181155e-01 -6.28127754e-01 3.86842489e-01
4.90605295e-01 1.46538484e+00 -4.52195078e-01 -1.14581689e-01
-7.27656484e-01 8.59556615e-01 6.43607855e-01 -2.87745409e-02
-5.49426377e-01 5.32495439e-01 6.57248199e-01 3.80520105e-01
8.45881030e-02 3.99649888e-02 -6.38227984e-02 -1.28750157e+00
-3.04125577e-01 -9.26318884e-01 6.18162632e-01 -7.18602717e-01
-1.70087957e+00 5.12870789e-01 -5.52496761e-02 -1.04703772e+00
-3.07079434e-01 -1.51655346e-01 -6.60669744e-01 1.54526383e-01
-1.42092621e+00 -1.36074328e+00 -6.86624274e-02 9.20924425e-01
-2.95717269e-01 -1.11984782e-01 1.16768789e+00 3.39715600e-01
-1.99221745e-01 6.51207924e-01 -3.28354761e-02 -4.13084179e-01
1.03872955e+00 -9.05637920e-01 9.49685276e-02 2.18734797e-02
-1.04845613e-02 8.35860670e-01 7.73955643e-01 -9.20501530e-01
-9.18374538e-01 -3.25153887e-01 1.19355619e+00 -5.86081207e-01
1.05469680e+00 -2.04849198e-01 -7.19676316e-01 7.63991177e-01
1.39963791e-01 -3.90131414e-01 1.48519886e+00 8.65707219e-01
-4.29728031e-01 3.87048684e-02 -1.70462990e+00 3.59729528e-01
1.06020772e+00 -3.20312768e-01 -4.72423613e-01 4.38271999e-01
3.37736964e-01 3.77698421e-01 -1.50284648e+00 1.08224973e-01
1.14081585e+00 -1.66368639e+00 8.54810596e-01 -7.72466242e-01
3.39936227e-01 4.41648632e-01 -1.58151194e-01 -1.28136086e+00
-5.29986918e-01 -5.64706743e-01 -5.72660565e-02 1.01551962e+00
5.87615967e-01 -8.41116130e-01 9.86859500e-01 1.43848598e+00
2.35675201e-01 -9.18956459e-01 -7.03505039e-01 -7.53429607e-02
-2.52448648e-01 -3.05544674e-01 3.15613925e-01 1.18932271e+00
9.53051388e-01 1.51644364e-01 -9.05931473e-01 2.03115985e-01
6.95467651e-01 -4.41380680e-01 2.54558742e-01 -2.00806141e+00
-4.15914804e-01 2.34048605e-01 -2.43292004e-01 -4.07910764e-01
-1.00187220e-01 -3.76264215e-01 -5.73383868e-01 -1.44741154e+00
3.04530770e-01 -7.14878201e-01 -5.01103163e-01 2.63431251e-01
1.83434516e-01 2.86889374e-01 -3.99664510e-03 2.44464293e-01
-9.63120699e-01 3.16100061e-01 9.55046833e-01 2.44474307e-01
-1.15524847e-02 8.23540688e-01 -1.23817599e+00 8.68709743e-01
7.61913598e-01 -5.86047888e-01 -5.50339103e-01 -3.34103890e-02
1.26217043e+00 2.54065543e-01 1.08568303e-01 -6.97971880e-01
5.06588042e-01 2.00387776e-01 6.56580180e-02 -2.02703178e-01
5.31852067e-01 -1.15826309e+00 1.50406033e-01 4.57132459e-01
-3.45786810e-01 1.82120483e-02 -4.61148858e-01 6.35954082e-01
5.72193973e-02 -4.98717397e-01 2.80298680e-01 -2.17054561e-01
-1.96050972e-01 1.85012326e-01 -5.06253362e-01 -6.08626485e-01
1.08525217e+00 -5.57012521e-02 -2.10130170e-01 -9.77848172e-01
-1.50014520e+00 4.46426660e-01 4.87029374e-01 5.29026449e-01
7.75703311e-01 -1.09631741e+00 -7.55387485e-01 5.24819866e-02
2.36311764e-01 -9.51771140e-01 1.49162546e-01 9.00193095e-01
1.34940267e-01 -2.51596302e-01 -5.20551741e-01 1.86519340e-01
-1.63945055e+00 3.77897114e-01 9.97864455e-02 -1.06637508e-01
9.95747223e-02 8.87601316e-01 -6.48529828e-01 -3.15131277e-01
4.32971239e-01 -1.29722372e-01 -8.56222451e-01 6.22139931e-01
6.25244498e-01 6.36071563e-01 5.72431134e-03 -3.61676902e-01
-4.15027410e-01 1.86483364e-03 -5.23557961e-01 -2.68490817e-02
1.73991799e+00 -5.42392850e-01 -5.82201779e-01 5.86836100e-01
8.17936778e-01 6.09059744e-02 -6.95050240e-01 -1.87717721e-01
2.42641464e-01 -1.09380201e-01 -2.08969310e-01 -9.50602233e-01
-7.52093315e-01 1.59980416e-01 2.79219627e-01 8.36566746e-01
6.27113998e-01 1.71097517e-01 3.73798281e-01 5.91699779e-01
8.70531380e-01 -1.01152623e+00 2.66734183e-01 1.94662049e-01
7.58014441e-01 -1.60594833e+00 6.49801731e-01 -4.11441416e-01
-9.39334393e-01 1.04665780e+00 3.30201268e-01 -4.53400016e-02
1.47418225e+00 -2.31743768e-01 7.31306570e-03 -4.80530918e-01
-9.26333010e-01 1.68102473e-01 -4.92945239e-02 6.22548103e-01
5.11188865e-01 1.63823143e-01 -4.40491676e-01 9.24308181e-01
5.82497828e-02 1.08579397e-01 8.52837861e-01 9.37512100e-01
-6.87881529e-01 -1.61855209e+00 1.66324243e-01 7.14390695e-01
-5.29207587e-01 -1.96652547e-01 -5.97934425e-01 5.13473451e-01
-9.53970253e-02 1.46584785e+00 -4.54250574e-01 -6.40976429e-01
2.12965101e-01 -1.94548607e-01 2.38786891e-01 -6.83605969e-01
-1.11117887e+00 -4.39117134e-01 7.74933338e-01 -5.00150740e-01
-8.48817766e-01 -1.03034794e+00 -7.65068114e-01 -4.95424867e-01
-5.41986763e-01 4.81338412e-01 6.06948316e-01 9.27509964e-01
3.66018802e-01 -2.62529850e-01 6.92226350e-01 -6.34431183e-01
-4.07281369e-01 -8.64696681e-01 -8.73420537e-01 7.41446018e-02
2.36476153e-01 -7.07099020e-01 -4.05078560e-01 -5.50479472e-01] | [9.937392234802246, 6.021758079528809] |
22b95013-d7b7-4e71-b3df-35a0e627303f | extracting-event-temporal-relations-via | 2109.05527 | null | https://arxiv.org/abs/2109.05527v1 | https://arxiv.org/pdf/2109.05527v1.pdf | Extracting Event Temporal Relations via Hyperbolic Geometry | Detecting events and their evolution through time is a crucial task in natural language understanding. Recent neural approaches to event temporal relation extraction typically map events to embeddings in the Euclidean space and train a classifier to detect temporal relations between event pairs. However, embeddings in the Euclidean space cannot capture richer asymmetric relations such as event temporal relations. We thus propose to embed events into hyperbolic spaces, which are intrinsically oriented at modeling hierarchical structures. We introduce two approaches to encode events and their temporal relations in hyperbolic spaces. One approach leverages hyperbolic embeddings to directly infer event relations through simple geometrical operations. In the second one, we devise an end-to-end architecture composed of hyperbolic neural units tailored for the temporal relation extraction task. Thorough experimental assessments on widely used datasets have shown the benefits of revisiting the tasks on a different geometrical space, resulting in state-of-the-art performance on several standard metrics. Finally, the ablation study and several qualitative analyses highlighted the rich event semantics implicitly encoded into hyperbolic spaces. | ['Yulan He', 'Gabriele Pergola', 'Xingwei Tan'] | 2021-09-12 | null | https://aclanthology.org/2021.emnlp-main.636 | https://aclanthology.org/2021.emnlp-main.636.pdf | emnlp-2021-11 | ['temporal-relation-extraction'] | ['natural-language-processing'] | [-1.01665370e-01 4.48433518e-01 -5.30584753e-02 -4.85780269e-01
-3.74751657e-01 -5.84884882e-01 1.09335399e+00 1.03511369e+00
-7.46396124e-01 2.87004769e-01 8.30496967e-01 -2.15026870e-01
-4.42814589e-01 -1.14430273e+00 -4.16918486e-01 -3.57091039e-01
-7.60608077e-01 3.76166642e-01 2.86411703e-01 -2.34037831e-01
1.21113472e-01 5.42293251e-01 -1.22170794e+00 3.72098684e-01
2.35942885e-01 1.05532193e+00 -5.91947734e-01 6.72674716e-01
7.36024305e-02 1.31977844e+00 -2.74523765e-01 -5.38609028e-01
-4.13042009e-02 -4.29654360e-01 -1.10029662e+00 -5.02692044e-01
-1.98009714e-01 -2.68784076e-01 -9.84372556e-01 4.61081326e-01
3.15887153e-01 4.77106601e-01 8.20007682e-01 -1.35654688e+00
-1.02854741e+00 1.03884304e+00 -1.34956345e-01 8.10487688e-01
5.55047274e-01 -2.90241688e-01 1.64265203e+00 -1.00636423e+00
7.98598707e-01 1.05881381e+00 7.92503655e-01 -1.85058899e-02
-1.28501499e+00 -2.25578591e-01 1.48980945e-01 5.60486972e-01
-1.39508998e+00 -1.48740828e-01 7.74723530e-01 -4.33949560e-01
1.50046301e+00 2.14820057e-01 7.52292395e-01 1.20532107e+00
2.61720628e-01 7.80141056e-01 6.38629317e-01 -1.40669763e-01
2.89422780e-01 -3.34038168e-01 3.06757629e-01 3.95811230e-01
-7.92334452e-02 3.88966829e-01 -7.69699693e-01 1.05520055e-01
6.53527737e-01 6.87191859e-02 2.29736809e-02 -3.72404456e-01
-1.77244198e+00 1.05592966e+00 8.66649985e-01 5.30328929e-01
-4.72776800e-01 4.07458067e-01 8.22881162e-01 1.21211469e-01
6.34793103e-01 7.94429600e-01 -2.72025198e-01 -2.78428912e-01
-7.54109263e-01 4.34524566e-01 6.65858984e-01 6.23523295e-01
3.13875318e-01 -5.39211094e-01 -4.21317607e-01 3.00190926e-01
1.70205459e-01 -3.72990340e-01 1.90067291e-01 -5.61547458e-01
5.18329084e-01 9.15260017e-01 -1.14402086e-01 -1.30911469e+00
-1.08130157e+00 -9.65954363e-02 -6.11404479e-01 -1.58513963e-01
4.39643860e-01 -1.93766914e-02 -5.16927361e-01 1.78625059e+00
3.57967764e-01 2.79274344e-01 1.22084342e-01 8.00041974e-01
9.48660612e-01 7.21360266e-01 3.80146176e-01 -1.43187139e-02
1.66711915e+00 -5.78889847e-01 -6.76417708e-01 4.43862043e-02
8.23597312e-01 -2.13383019e-01 9.42008376e-01 -3.25182057e-03
-1.22314489e+00 -2.27622375e-01 -1.23172998e+00 -7.25091398e-01
-1.00205445e+00 -7.90840611e-02 8.28194857e-01 4.84042466e-02
-8.37063730e-01 9.04653788e-01 -1.14179516e+00 -5.68775237e-01
4.66074228e-01 1.32355213e-01 -4.15733933e-01 5.98479092e-01
-1.66546357e+00 1.12179172e+00 8.26584339e-01 6.91067427e-02
-4.49672103e-01 -8.42285156e-01 -1.25043547e+00 2.96448320e-01
1.16176769e-01 -3.36316258e-01 1.18389177e+00 -1.27278571e-03
-9.67567742e-01 1.05051851e+00 1.78195179e-01 -8.13528597e-01
3.85597765e-01 -3.16953391e-01 -5.44988096e-01 3.61319631e-01
1.69701964e-01 5.64063132e-01 2.29050636e-01 -6.17685437e-01
-6.52336359e-01 -3.54887843e-01 3.10093045e-01 1.82945058e-01
-4.66358423e-01 2.25919902e-01 -5.82173616e-02 -9.14035559e-01
4.23325121e-01 -7.29758859e-01 -5.35640232e-02 9.98671725e-02
-3.56871843e-01 -7.62119770e-01 6.85538650e-01 -3.99876475e-01
1.46045625e+00 -2.13318229e+00 2.81755060e-01 -5.01348302e-02
4.08187956e-01 -5.28447628e-01 2.78002441e-01 7.88018167e-01
-4.83555794e-01 1.28306616e-02 -2.39857465e-01 -3.45400453e-01
5.22125781e-01 1.76170114e-02 -7.63543367e-01 6.68487072e-01
7.80881047e-01 1.20116246e+00 -1.27276492e+00 -4.62916642e-01
3.37374300e-01 5.35975695e-01 -4.57042098e-01 1.61943227e-01
3.07632014e-02 8.82250220e-02 -2.07475871e-01 2.55364776e-01
9.73126516e-02 -3.12620968e-01 1.97444912e-02 -3.96069944e-01
-2.53189206e-01 8.74457240e-01 -1.05761564e+00 1.81060398e+00
-3.48323286e-01 1.01611662e+00 -9.62740719e-01 -9.84545231e-01
7.08030701e-01 7.10505843e-01 8.62016737e-01 -6.80767238e-01
3.14192891e-01 -1.52266279e-01 -1.96018085e-01 -4.61496353e-01
7.21382201e-01 -1.19211085e-01 -7.49638677e-01 7.60736942e-01
3.07439238e-01 8.29236060e-02 3.39596301e-01 3.97636950e-01
1.34072387e+00 2.34842956e-01 7.37279117e-01 -1.45018771e-01
1.65389672e-01 -3.14242989e-02 2.28014827e-01 4.94851470e-01
-4.18692231e-01 6.85444474e-01 9.61759388e-01 -1.01158178e+00
-1.08604395e+00 -1.67492640e+00 -3.01177859e-01 1.30161166e+00
7.17610419e-02 -8.91090989e-01 -1.51911467e-01 -7.52914727e-01
-2.51264989e-01 7.98228860e-01 -1.34779942e+00 -3.96341473e-01
-8.23067069e-01 -8.41374576e-01 9.63856220e-01 1.07966578e+00
3.81893396e-01 -1.22234285e+00 -1.12890506e+00 4.70321089e-01
-2.06583679e-01 -1.37772512e+00 -2.67706275e-01 5.42503595e-01
-5.21354616e-01 -1.03361225e+00 -2.98735231e-01 -7.97726870e-01
1.27061412e-01 -4.55046624e-01 1.30856049e+00 -6.13795757e-01
-3.65504682e-01 2.39863530e-01 -6.27680898e-01 -3.24646175e-01
9.79531184e-02 1.40775591e-01 2.75841728e-02 -4.10595499e-02
6.84843540e-01 -8.90964270e-01 -9.00077224e-01 1.55168185e-02
-8.97872090e-01 -1.00018479e-01 1.57830939e-01 4.25383300e-01
3.28573465e-01 2.95604374e-02 4.99311686e-01 -6.09352171e-01
5.63664198e-01 -7.65096128e-01 -4.84013371e-02 -3.21149006e-02
-1.63449436e-01 2.66653180e-01 5.91756642e-01 -4.39164937e-01
-9.07978952e-01 -1.43570274e-01 4.58782278e-02 -1.35859102e-01
-2.09679037e-01 6.60628080e-01 2.74954885e-01 6.37980580e-01
9.42166984e-01 -6.50270209e-02 -7.08980560e-01 -2.30349209e-02
8.41369629e-01 2.93638706e-02 7.03420460e-01 -5.58070242e-01
4.26636875e-01 8.21837008e-01 -2.72669084e-02 -4.28102136e-01
-9.36715364e-01 -3.80068064e-01 -8.67227018e-01 6.18563555e-02
1.31235194e+00 -6.74802721e-01 -8.31885993e-01 -2.54830066e-02
-1.25623560e+00 -2.69794047e-01 -7.65728474e-01 7.34240949e-01
-7.07680047e-01 -6.01881631e-02 -8.86070192e-01 -5.18069863e-01
-6.08030520e-02 -5.82688510e-01 1.37520552e+00 3.67523655e-02
-9.45497274e-01 -1.18164945e+00 5.34985065e-01 -3.69546652e-01
-9.89756268e-03 7.21567631e-01 1.07827926e+00 -8.06743801e-01
-2.62723535e-01 -1.19892269e-01 -4.88405585e-01 -6.10884070e-01
-2.39424780e-02 -7.57650286e-02 -9.33595300e-01 2.19009101e-01
-7.87604749e-02 -2.28865325e-01 1.02000916e+00 1.26234546e-01
9.40352917e-01 -2.24232495e-01 -4.16562587e-01 5.66610336e-01
1.13666379e+00 1.63751259e-01 6.19766951e-01 3.78348082e-01
7.37049997e-01 8.84996891e-01 3.59090567e-01 6.62239850e-01
7.43943453e-01 5.28261304e-01 2.88813829e-01 9.20122936e-02
1.75585933e-02 -3.42500925e-01 3.47882837e-01 3.94556820e-01
9.26294550e-02 -2.10096404e-01 -1.11116123e+00 9.55485582e-01
-1.98164427e+00 -1.18706393e+00 -1.48246720e-01 1.68810809e+00
9.50221539e-01 3.26401204e-01 2.89315939e-01 5.03519893e-01
2.66755849e-01 5.30618072e-01 -1.95636094e-01 -5.33146143e-01
-2.56401300e-03 3.60540539e-01 8.46940055e-02 2.14097038e-01
-1.57512021e+00 9.09039617e-01 6.24178600e+00 1.78950980e-01
-9.84572470e-01 -5.57221621e-02 3.88865650e-01 -2.23439515e-01
-9.91425514e-02 -9.68002528e-02 -4.41640943e-01 6.46226183e-02
1.09857285e+00 -2.12014720e-01 -2.93486528e-02 3.64394605e-01
1.45334259e-01 1.11418590e-01 -1.93400729e+00 9.99643028e-01
-1.77744448e-01 -1.55848730e+00 -9.36955288e-02 -1.21676542e-01
4.04943585e-01 -1.52004942e-01 -9.03978385e-03 4.91082877e-01
4.13108915e-01 -1.20819461e+00 9.45321202e-01 4.66757476e-01
5.92539370e-01 -6.26839221e-01 4.24518287e-01 -2.65784681e-01
-1.75190723e+00 7.60512725e-02 8.43586996e-02 -3.79084408e-01
5.89153886e-01 3.15551817e-01 -9.54317987e-01 4.56272036e-01
8.68019581e-01 1.09656894e+00 -7.03458250e-01 6.69226050e-01
-2.81529039e-01 4.02921408e-01 -4.83563781e-01 7.47261122e-02
5.03429651e-01 -1.23176025e-02 4.92700040e-01 1.44797730e+00
5.07559776e-02 3.23768467e-01 -1.37068555e-01 1.27927256e+00
2.41642166e-03 -3.06403060e-02 -9.24978077e-01 -3.53328615e-01
2.62873292e-01 1.08920956e+00 -1.12414920e+00 -2.18330935e-01
-4.02947068e-01 9.75619495e-01 6.81888402e-01 1.88193604e-01
-1.29771268e+00 -6.38535380e-01 5.93163371e-01 -6.72509447e-02
2.21003532e-01 -4.85969692e-01 -5.60776770e-01 -9.87816453e-01
1.43980667e-01 -8.16729516e-02 9.82237220e-01 -6.48775637e-01
-1.22621775e+00 6.78286433e-01 4.01395500e-01 -9.87193584e-01
-4.28042293e-01 -5.16084611e-01 -7.26264656e-01 3.45630348e-01
-8.23466659e-01 -1.13987267e+00 -6.44802675e-02 6.12265110e-01
2.88857430e-01 4.29430425e-01 8.86945009e-01 2.50078171e-01
-5.28505802e-01 4.24952328e-01 -3.88217419e-01 6.18843973e-01
3.15198064e-01 -1.72127080e+00 6.83870256e-01 8.81372273e-01
6.95057988e-01 6.22243583e-01 6.39458060e-01 -2.81396985e-01
-9.30440962e-01 -1.13644266e+00 1.45341015e+00 -7.82647192e-01
1.02233255e+00 -6.98739290e-01 -8.77625108e-01 1.12206960e+00
1.53348297e-01 3.37671429e-01 8.45158458e-01 4.32727367e-01
-6.68601394e-01 1.53797284e-01 -7.58648932e-01 1.06533718e+00
1.38593519e+00 -1.04660642e+00 -1.02797389e+00 1.64843082e-01
1.01308429e+00 -2.84946054e-01 -1.31028914e+00 4.58058476e-01
3.86119306e-01 -8.01744819e-01 1.18099523e+00 -1.02438796e+00
8.03892136e-01 -2.58408189e-01 -2.16203734e-01 -1.02360392e+00
-3.41923535e-01 -6.35864437e-01 -4.81961966e-01 1.14497662e+00
4.58666682e-01 -3.23705673e-01 6.23762012e-01 4.08425838e-01
6.55058865e-03 -8.94528031e-01 -8.97303462e-01 -5.27771294e-01
2.64977694e-01 -6.88388586e-01 3.79193455e-01 1.14985108e+00
5.95372558e-01 6.87974393e-01 1.45777166e-01 2.76695281e-01
2.88319319e-01 1.31063238e-01 1.03531167e-01 -1.14092731e+00
5.09481989e-02 -6.22822285e-01 -6.98580086e-01 -6.28940582e-01
1.80272922e-01 -1.06317222e+00 4.62309970e-03 -1.58654904e+00
-7.88283125e-02 -1.88890442e-01 -4.45007205e-01 4.08825219e-01
-9.69752967e-02 2.76636302e-01 -1.30019724e-01 -1.80876955e-01
-7.89579451e-01 8.68729174e-01 6.38171971e-01 -1.82047170e-02
-2.79705882e-01 -5.55159748e-01 -4.61087942e-01 8.05958569e-01
6.77686751e-01 -4.96408850e-01 -5.75986385e-01 -4.90085363e-01
7.87868679e-01 7.41564184e-02 7.22465456e-01 -1.04418874e+00
4.08453286e-01 1.17799871e-01 2.78108627e-01 -6.37683868e-01
4.77596253e-01 -6.79140806e-01 -2.17595920e-01 -3.90749909e-02
-7.60572374e-01 4.47339088e-01 2.04546303e-01 6.01915419e-01
-3.91173542e-01 3.80694509e-01 4.55809414e-01 2.85279155e-01
-9.06187952e-01 1.90090030e-01 -2.39578903e-01 4.66664284e-01
1.27219450e+00 -2.51329988e-02 -1.77913308e-01 -2.82895088e-01
-1.20137191e+00 -4.39374968e-02 -9.29104686e-02 6.61316991e-01
6.22071624e-01 -1.67196655e+00 -7.24484682e-01 -1.81403756e-01
3.22442710e-01 9.44288971e-04 -5.07912077e-02 9.70553160e-01
-6.79997265e-01 4.42486346e-01 -4.93347161e-02 -4.91634756e-01
-7.91181743e-01 8.52453530e-01 2.24883080e-01 -5.93071461e-01
-9.58339930e-01 9.53311443e-01 3.43275487e-01 -3.52116704e-01
2.27993429e-01 -1.05125332e+00 -3.03029269e-01 6.24827445e-01
4.45260137e-01 3.61774057e-01 8.79399478e-02 -5.18790603e-01
-6.39644384e-01 2.08011582e-01 -4.05566730e-02 -6.10475898e-01
1.39989746e+00 -2.69479863e-03 -2.99049802e-02 8.64268064e-01
1.50581241e+00 -3.92169863e-01 -1.17503035e+00 -3.00170362e-01
5.63239992e-01 -3.30386753e-03 -1.80225790e-01 -3.10620248e-01
-5.18087626e-01 9.52118158e-01 1.44954145e-01 8.68930995e-01
9.07251060e-01 3.32367897e-01 6.97542250e-01 2.90945649e-01
3.32414061e-02 -8.30215633e-01 2.45536894e-01 7.29051352e-01
1.02507329e+00 -1.08352280e+00 -7.78394714e-02 -2.02157706e-01
-5.24022758e-01 1.23613489e+00 2.35029787e-01 -4.31343913e-01
8.92075896e-01 1.35069922e-01 -4.37215567e-01 -6.36650860e-01
-7.54392564e-01 -4.01207954e-01 4.31533158e-01 2.12447122e-01
6.90026760e-01 6.95641758e-03 -1.24899536e-01 6.32202566e-01
-6.36061728e-01 -3.85919183e-01 2.57512212e-01 9.00874078e-01
1.15127526e-01 -5.65670252e-01 -4.34883796e-02 3.01843453e-02
-4.41537023e-01 -1.90278053e-01 -4.49234843e-01 9.31650221e-01
1.63623556e-01 9.49183822e-01 6.78966820e-01 -2.50936955e-01
5.70284426e-01 4.15586792e-02 5.31469703e-01 -5.31041503e-01
-6.66961491e-01 -4.73387122e-01 1.30280510e-01 -8.41955543e-01
-2.96021044e-01 -9.35874343e-01 -1.72068846e+00 -1.39624894e-01
1.99525759e-01 -2.23525949e-02 1.03069231e-01 8.24111104e-01
2.32804596e-01 9.04754639e-01 2.61738509e-01 -6.82669520e-01
-1.56430125e-01 -7.05025077e-01 -4.02385801e-01 7.48700261e-01
4.28814530e-01 -7.65058756e-01 -3.38890105e-02 1.46805406e-01] | [9.18298053741455, 9.140591621398926] |
f6298fed-3dc7-4dbb-beda-47b0126fb072 | deep-learning-methods-allow-fully-automated | null | null | https://www.nature.com/articles/s41598-021-89111-9 | https://www.nature.com/articles/s41598-021-89111-9.pdf | Deep learning methods allow fully automated segmentation of metacarpal bones to quantify volumetric bone mineral density | Arthritis patients develop hand bone loss, which leads to destruction and functional impairment of the affected joints. High resolution peripheral quantitative computed tomography (HR-pQCT) allows the quantification of volumetric bone mineral density (vBMD) and bone microstructure in vivo with an isotropic voxel size of 82 micrometres. However, image-processing to obtain bone characteristics is a time-consuming process as it requires semi-automatic segmentation of the bone. In this work, a fully automatic vBMD measurement pipeline for the metacarpal (MC) bone using deep learning methods is introduced. Based on a dataset of HR-pQCT volumes with MC measurements for 541 patients with arthritis, a segmentation network is trained. The best network achieves an intersection over union as high as 0.94 and a Dice similarity coefficient of 0.97 while taking only 33 s to process a whole patient yielding a speedup between 2.5 and 4.0 for the whole workflow. Strong correlation between the vBMD measurements of the expert and of the automatic pipeline are achieved for the average bone density with 0.999 (Pearson) and 0.996 (Spearman’s rank) with p<0.001 for all correlations. A qualitative assessment of the network predictions and the manual annotations yields a 65.9% probability that the expert favors the network predictions. Further, the steps to integrate the pipeline into the clinical workflow are shown. In order to make these workflow improvements available to others, we openly share the code of this work. | ['Andreas Maier', 'Arnd Kleyer', 'Georg Schett', 'Gerhard Krönke', 'Anna-Maria Liphardt', 'David Simon', 'Timo Meinderink', 'Lukas Folle'] | 2021-05-06 | null | null | null | null | ['3d-instance-segmentation-1'] | ['computer-vision'] | [-6.33865222e-02 2.83429861e-01 5.29916398e-02 -3.38490486e-01
-8.34659576e-01 -9.46295541e-03 -3.20788193e-03 4.75082755e-01
-5.96658409e-01 7.62950480e-01 4.80057932e-02 -3.48962061e-02
-3.06510776e-01 -1.06930137e+00 -4.08575416e-01 -7.03592360e-01
-5.02642572e-01 1.46533108e+00 6.02245748e-01 2.44385868e-01
7.05900937e-02 7.09582865e-01 -9.88565326e-01 4.48531717e-01
3.88794959e-01 8.80352795e-01 1.61535561e-01 7.71799803e-01
1.59041047e-01 4.74403471e-01 -2.13987857e-01 -1.95298165e-01
2.27959275e-01 -7.05225766e-02 -8.35319638e-01 -3.04718584e-01
1.02034114e-01 -6.88070595e-01 1.38197979e-02 5.61960220e-01
9.21657562e-01 -2.69812137e-01 7.32731998e-01 -6.74171209e-01
-1.09308243e-01 6.00665033e-01 -5.14979839e-01 3.28762114e-01
8.70852079e-03 3.42386335e-01 7.66525388e-01 -7.84834206e-01
7.68529356e-01 8.20053101e-01 8.53053153e-01 2.82862425e-01
-1.27378619e+00 -6.25634432e-01 -8.10647070e-01 1.81355283e-01
-1.53916395e+00 -9.69873890e-02 5.41557278e-03 -6.75369382e-01
1.06259823e+00 1.57982647e-01 1.33419371e+00 5.60414910e-01
7.43780315e-01 9.92710590e-02 1.20933211e+00 -1.66995332e-01
4.46209460e-01 -5.82769096e-01 -2.40578473e-01 8.14545989e-01
3.44224811e-01 -1.10355280e-01 -1.51575267e-01 -2.10313275e-01
1.16716921e+00 -1.02699444e-01 1.84815973e-01 -2.44642302e-01
-1.08202016e+00 7.64723122e-01 5.96871138e-01 3.27096730e-01
-7.42627978e-01 3.75763446e-01 5.66401780e-01 -1.81241944e-01
2.45850608e-01 4.96099517e-03 -1.84760854e-01 -1.69059440e-01
-1.12838078e+00 3.21088731e-01 4.95403200e-01 2.74285495e-01
2.53749907e-01 -2.86139846e-01 -2.89235208e-02 8.98807704e-01
5.25964558e-01 4.57776964e-01 3.53314340e-01 -1.25260401e+00
-2.13060632e-01 6.09227657e-01 -2.86676466e-01 -9.20871139e-01
-7.64605999e-01 -4.46760833e-01 -9.91805255e-01 5.28353214e-01
7.59907126e-01 7.48303384e-02 -9.43695664e-01 1.14931810e+00
4.64400649e-01 -4.44929034e-01 -8.81700814e-01 1.18922055e+00
4.87353086e-01 -6.51689172e-02 1.14227310e-01 -7.03728572e-02
1.52943051e+00 -6.86745703e-01 -3.44707787e-01 1.48811638e-01
4.25767601e-01 -6.82803690e-01 1.00487924e+00 4.55976546e-01
-1.37542379e+00 -1.60319954e-01 -8.30219865e-01 6.54849112e-02
9.42060351e-02 -2.07587242e-01 2.55045444e-01 5.53216875e-01
-1.20046461e+00 1.07806897e+00 -1.46567261e+00 -3.79025251e-01
7.35059798e-01 7.67383158e-01 -4.76350397e-01 -1.77251354e-01
-1.02232325e+00 9.85174298e-01 4.25631441e-02 6.37582317e-02
-7.52073944e-01 -9.94717121e-01 -1.48334876e-01 -1.76919788e-01
-2.02711880e-01 -1.17236829e+00 9.33297217e-01 -4.22388166e-01
-1.46018124e+00 1.04394186e+00 1.78768411e-01 -2.69348234e-01
1.03649056e+00 -2.28412330e-01 1.65771082e-01 6.66745543e-01
4.08487022e-01 4.58868533e-01 1.17036805e-01 -8.47894967e-01
-4.29239631e-01 -8.17729354e-01 -4.28110212e-01 -1.48149624e-01
2.45657265e-01 1.52785495e-01 -4.56457347e-01 -2.45584309e-01
4.69041973e-01 -8.98326695e-01 -5.52090883e-01 4.07995582e-01
-3.66384745e-01 1.79014653e-01 1.64798945e-02 -1.19199598e+00
8.69519472e-01 -1.55584180e+00 -3.23413163e-01 8.26658189e-01
6.27929688e-01 -1.26420960e-01 3.55446309e-01 2.26488516e-01
-1.68701246e-01 3.18449573e-03 -3.65328014e-01 4.20029089e-02
-3.48793626e-01 3.24329846e-02 7.49378920e-01 8.01443696e-01
-2.68120825e-01 9.10872400e-01 -6.95960701e-01 -7.76564777e-01
1.79710373e-01 7.11563706e-01 -5.56214333e-01 -1.98424365e-02
2.01448634e-01 6.35569334e-01 -2.49363750e-01 6.12060249e-01
4.77787465e-01 -5.13088822e-01 4.95256126e-01 -2.45497257e-01
1.28240034e-01 9.89438221e-02 -9.99284267e-01 1.62060511e+00
-2.65949786e-01 8.99194032e-02 4.09868747e-01 -5.67702532e-01
9.01939809e-01 3.61109853e-01 1.15125692e+00 -6.91081643e-01
2.67614275e-01 5.11744440e-01 4.82750535e-01 -3.09276491e-01
-4.42496896e-01 -7.73613036e-01 6.10082269e-01 7.69523203e-01
-2.07819864e-01 7.44251236e-02 1.61540061e-01 1.46590531e-01
1.49889946e+00 8.31149891e-02 3.55875582e-01 -2.82982916e-01
2.42595717e-01 1.16215765e-01 1.55829132e-01 4.22039241e-01
-4.29414749e-01 8.10543656e-01 5.40141344e-01 -3.47102553e-01
-1.49656975e+00 -1.52122748e+00 -3.71218145e-01 5.96731246e-01
-5.61990440e-01 -2.33066857e-01 -7.78902590e-01 -6.15091249e-02
1.92481816e-01 5.86316399e-02 -6.73184335e-01 2.60717511e-01
-7.10404456e-01 -8.12171876e-01 6.14818275e-01 6.76605046e-01
1.88759401e-01 -7.72988379e-01 -6.19104683e-01 3.84194463e-01
-1.65522024e-01 -7.83627689e-01 3.04345518e-01 8.18723142e-02
-1.35861778e+00 -1.14776289e+00 -8.84863377e-01 -5.30259609e-01
4.03023541e-01 -4.87184584e-01 1.01460290e+00 3.91067803e-01
-8.45488071e-01 1.03935219e-01 8.15154985e-03 1.39167756e-01
-3.97360742e-01 -7.10904300e-02 5.70211150e-02 -7.49269009e-01
9.28427577e-02 -9.94474530e-01 -1.25633788e+00 1.45717591e-01
-3.87125254e-01 1.43421829e-01 8.05477917e-01 4.42552000e-01
1.19017696e+00 -5.27800880e-02 2.08391577e-01 -7.12103724e-01
3.25543106e-01 -4.71669048e-01 -1.52634755e-01 -1.89570710e-01
-9.36491311e-01 -3.25581193e-01 3.29559505e-01 1.39163136e-01
-7.11884916e-01 9.98932943e-02 -4.26659644e-01 1.25274241e-01
-1.30563453e-01 4.25021678e-01 2.95036972e-01 1.64835274e-01
5.70983708e-01 -2.01174647e-01 6.47826910e-01 -4.53225791e-01
1.54663339e-01 5.18557906e-01 3.94933134e-01 -4.79002714e-01
2.29671925e-01 8.16741824e-01 4.72503573e-01 -6.05518818e-01
-2.99983978e-01 -5.35230815e-01 -9.60787117e-01 -5.69382727e-01
1.01636469e+00 -6.63369656e-01 -9.31733966e-01 2.09365457e-01
-8.49971473e-01 -5.39751053e-01 -1.21147688e-02 7.19984770e-01
-6.45751536e-01 4.11049575e-01 -1.06313848e+00 -4.09202665e-01
-1.02270734e+00 -1.20486426e+00 7.58673310e-01 -3.47812712e-01
-8.61123323e-01 -8.76442373e-01 4.26798254e-01 6.85811818e-01
5.27056038e-01 6.01175129e-01 1.01325321e+00 -5.89075387e-01
-2.57593304e-01 -3.24395269e-01 -3.41645002e-01 2.08399519e-01
3.46983932e-02 1.79372266e-01 -5.65207005e-01 1.59980491e-01
-1.23186164e-01 -1.86537448e-02 4.34159487e-01 8.08115184e-01
8.60942423e-01 2.12535977e-01 -2.85576582e-01 1.47310793e-01
1.57986927e+00 4.27920168e-04 7.64317572e-01 6.67806625e-01
5.98135769e-01 7.85846829e-01 3.69306415e-01 5.64391553e-01
-7.80529389e-03 5.67923725e-01 6.31488025e-01 -2.33873352e-02
-3.76295179e-01 2.32318431e-01 -9.27817672e-02 7.98590899e-01
-7.17936337e-01 4.89883125e-01 -1.54253745e+00 5.84906161e-01
-1.30316913e+00 -5.92970788e-01 -9.22975838e-01 2.05004930e+00
1.04805946e+00 3.65863711e-01 4.21648115e-01 1.48349911e-01
7.52408683e-01 -4.72057432e-01 -3.34697574e-01 -1.40407100e-01
3.99077058e-01 9.68204081e-01 4.49093074e-01 6.57608628e-01
-4.33332890e-01 4.87850904e-01 6.37320328e+00 5.80590487e-01
-8.96898746e-01 3.76143545e-01 4.76867676e-01 -5.31352758e-01
1.44118413e-01 -1.49766877e-01 -8.33647326e-02 4.26641852e-01
1.31024075e+00 1.20737001e-01 2.22261131e-01 6.27999842e-01
6.91013515e-01 -5.97966313e-01 -7.19678402e-01 5.19506335e-01
-5.75164437e-01 -1.23459363e+00 -2.05374256e-01 3.18839967e-01
6.74286783e-02 4.85264540e-01 -3.18153501e-01 -3.42480540e-01
3.50244820e-01 -1.19690406e+00 1.69751391e-01 7.03901231e-01
7.79273808e-01 -6.66323900e-01 9.59695995e-01 -1.06593058e-01
-9.04626548e-01 3.84436518e-01 -2.59737521e-01 -1.64405376e-01
3.80464375e-01 1.26189625e+00 -1.26996171e+00 3.05085033e-01
9.92528617e-01 1.38369709e-01 -4.43099111e-01 1.13852215e+00
-2.27265731e-01 5.79700053e-01 -6.06191039e-01 2.43966699e-01
-7.90021718e-02 -2.21238792e-01 1.14894368e-01 1.05710638e+00
1.68909520e-01 7.80292377e-02 -2.67090499e-01 7.97785461e-01
3.39665264e-01 5.08529961e-01 1.69322848e-01 2.71637946e-01
2.37073898e-01 1.30577648e+00 -1.16195381e+00 -1.51823848e-01
-1.85132623e-01 5.66390872e-01 1.74551412e-01 -2.69150168e-01
-6.95739925e-01 6.41905516e-02 3.52069885e-01 9.12671626e-01
-1.67122513e-01 -1.70491174e-01 -8.28741074e-01 -1.98155358e-01
-1.67185172e-01 -6.05244815e-01 3.46121550e-01 -8.18040788e-01
-1.32209313e+00 2.87875086e-01 -2.91474551e-01 -6.54538870e-01
-5.36448322e-02 -5.56361854e-01 -2.90093750e-01 9.32397723e-01
-7.17485249e-01 -9.22577381e-01 -4.28130955e-01 2.71181881e-01
-1.32469580e-01 2.49363601e-01 8.70473385e-01 4.96965498e-01
-3.70447159e-01 -3.07484232e-02 8.54031891e-02 3.86712700e-01
6.72962427e-01 -1.36968327e+00 4.19630259e-02 2.43321225e-01
-5.55316627e-01 7.45596886e-01 6.97431624e-01 -1.08965838e+00
-7.67631292e-01 -9.53912020e-01 8.18797290e-01 -2.73659855e-01
9.08348680e-01 2.56510347e-01 -7.87367940e-01 5.22241592e-01
-1.17122829e-01 -5.08984439e-02 1.36280298e+00 -1.73751011e-01
5.53143285e-02 1.57855362e-01 -1.26026535e+00 2.25268066e-01
8.13721359e-01 -9.15367082e-02 -3.65460992e-01 4.49274272e-01
1.02502048e-01 -3.29184204e-01 -1.84628940e+00 5.16027927e-01
1.18237340e+00 -1.22679019e+00 1.22431111e+00 -3.15774173e-01
8.27713609e-01 -4.66034673e-02 -1.42324597e-01 -5.05625486e-01
-4.42377031e-01 3.70796770e-01 5.37549071e-02 7.08752275e-01
2.64691442e-01 -3.55027944e-01 1.17137134e+00 4.95076835e-01
-1.90714985e-01 -1.22632754e+00 -1.12653673e+00 -5.69921374e-01
4.99406010e-01 -6.23867631e-01 -1.99144762e-02 5.41560769e-01
-1.28138393e-01 -1.28372684e-02 2.69898802e-01 6.97852904e-03
7.50437200e-01 -1.82565719e-01 4.41276312e-01 -1.43999016e+00
-3.44373524e-01 -3.65155607e-01 -5.36278903e-01 -1.34893462e-01
-3.19439948e-01 -1.14710081e+00 -2.21234426e-01 -2.15577388e+00
5.39470971e-01 -4.35140401e-01 -1.87824935e-01 4.29775596e-01
2.49558598e-01 7.10416436e-01 -3.22705716e-01 6.21492863e-01
-1.30706072e-01 1.27282247e-01 1.44629228e+00 1.30817086e-01
-6.58133999e-02 -5.34559228e-03 -1.65136933e-01 7.17533886e-01
1.15144992e+00 -6.90063834e-01 -6.24866877e-03 -9.63028222e-02
3.18145394e-01 -1.98866706e-02 6.03695631e-01 -1.17079866e+00
-1.00201152e-01 1.01786226e-01 7.85554171e-01 -6.16583228e-01
3.30867201e-01 -7.37243354e-01 5.83836377e-01 1.39653194e+00
3.71752046e-02 -2.00223595e-01 -1.13101318e-01 2.42548034e-01
9.17332172e-02 -9.73260403e-02 1.12092936e+00 -6.12729013e-01
2.84417924e-02 2.41502032e-01 -6.27626538e-01 -2.20883518e-01
8.52345943e-01 -2.73849636e-01 -8.48619416e-02 -1.58781633e-01
-1.46443141e+00 -1.06464803e-01 4.12183493e-01 -4.25088286e-01
5.14051437e-01 -1.17245960e+00 -8.90453219e-01 -4.14160281e-01
-2.96361685e-01 7.75953084e-02 4.73932415e-01 1.59013522e+00
-1.40831578e+00 3.92497778e-01 -7.70492256e-01 -9.23779726e-01
-1.40417409e+00 -4.98812124e-02 6.24043643e-01 -6.41001523e-01
-7.77760148e-01 6.81451559e-01 -3.13559681e-01 -3.08051854e-01
-2.64973730e-01 -7.34214857e-02 1.01871431e-01 -4.49312851e-02
3.34175915e-01 1.04506218e+00 3.48866373e-01 -5.49426615e-01
-6.51767373e-01 5.49474895e-01 2.31694477e-03 -3.28078508e-01
1.57799554e+00 -1.65075719e-01 -7.69504547e-01 3.63955498e-01
1.12052274e+00 -1.35124251e-01 -7.95653701e-01 6.35058954e-02
-1.03840202e-01 -1.42921031e-01 1.21763483e-01 -9.46991146e-01
-1.14509809e+00 9.34522510e-01 1.19056010e+00 -1.08221829e-01
6.39819503e-01 1.85608000e-01 9.08603251e-01 -6.74921200e-02
3.87266546e-01 -1.05457354e+00 6.22823313e-02 1.15036540e-01
8.89502466e-01 -7.49902129e-01 5.11045277e-01 -6.27266347e-01
-1.90663517e-01 1.18161035e+00 4.04062063e-01 -3.10440481e-01
8.10186982e-01 3.37731749e-01 1.87628448e-01 -6.23767972e-01
-2.47158647e-01 1.27871409e-01 -6.92038015e-02 7.96827793e-01
8.74858081e-01 2.34552532e-01 -8.76041889e-01 5.85224807e-01
-5.00681639e-01 2.36063659e-01 2.07860440e-01 8.79118383e-01
-6.34429812e-01 -9.94239092e-01 -4.28972244e-01 7.43315876e-01
-8.33017945e-01 4.38953787e-02 -4.16416585e-01 1.02131450e+00
6.39440045e-02 5.97578585e-01 1.72305703e-01 -8.94300938e-02
1.61570832e-01 3.81321274e-02 6.03265285e-01 -7.27697849e-01
-7.01402485e-01 4.26864535e-01 1.42973617e-01 -9.11033034e-01
-4.80347067e-01 -6.53681219e-01 -2.02952218e+00 -4.64592367e-01
1.08705960e-01 -2.88459867e-01 6.43669128e-01 8.81217360e-01
5.60387745e-02 6.25990391e-01 -1.49978772e-02 -6.54336751e-01
-7.51453936e-02 -9.98315930e-01 -1.01325583e+00 8.82704780e-02
-3.79076123e-01 -7.41386294e-01 -1.08663887e-01 -1.07445061e-01] | [14.174847602844238, -2.1623783111572266] |
37540365-46c6-4457-820d-571907fb5bc5 | epicardial-adipose-tissue-segmentation-from | 2204.12904 | null | https://arxiv.org/abs/2204.12904v1 | https://arxiv.org/pdf/2204.12904v1.pdf | Epicardial Adipose Tissue Segmentation from CT Images with A Semi-3D Neural Network | Epicardial adipose tissue is a type of adipose tissue located between the heart wall and a protective layer around the heart called the pericardium. The volume and thickness of epicardial adipose tissue are linked to various cardiovascular diseases. It is shown to be an independent cardiovascular disease risk factor. Fully automatic and reliable measurements of epicardial adipose tissue from CT scans could provide better disease risk assessment and enable the processing of large CT image data sets for a systemic epicardial adipose tissue study. This paper proposes a method for fully automatic semantic segmentation of epicardial adipose tissue from CT images using a deep neural network. The proposed network uses a U-Net-based architecture with slice depth information embedded in the input image to segment a pericardium region of interest, which is used to obtain an epicardial adipose tissue segmentation. Image augmentation is used to increase model robustness. Cross-validation of the proposed method yields a Dice score of 0.86 on the CT scans of 20 patients. | ['Irena Galić', 'Marija Habijan', 'Marin Benčević'] | 2022-04-27 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 2.68846005e-01 3.10727954e-01 -2.53062159e-01 -3.34423125e-01
-4.12741452e-01 -5.05743861e-01 -3.25707011e-02 3.29261273e-01
-4.71244454e-01 3.75901371e-01 1.04837850e-01 -3.50985080e-01
2.71011561e-01 -8.56145859e-01 -8.34563002e-02 -6.20322108e-01
-5.72060049e-01 7.22969651e-01 2.23938972e-01 5.88390648e-01
-6.21629357e-02 7.61995852e-01 -4.84877884e-01 1.79933831e-01
7.94640481e-01 1.45347571e+00 -1.60723552e-02 7.47496128e-01
-2.05287620e-01 1.02937020e-01 -1.18651837e-01 -8.50579962e-02
7.43463278e-01 -4.65430617e-01 -7.54988968e-01 -2.35780761e-01
4.44520444e-01 -4.56794769e-01 2.62671322e-01 9.37366843e-01
6.17959619e-01 -3.86909515e-01 6.94178939e-01 -7.09019125e-01
-8.14061705e-03 4.72170800e-01 -5.46627879e-01 3.19599807e-01
-6.09966040e-01 -1.77482530e-01 5.77253103e-01 -6.31616771e-01
3.17388475e-01 6.69920385e-01 8.54990304e-01 3.61603379e-01
-1.09649265e+00 -7.69618094e-01 -6.11894608e-01 -3.22780758e-01
-1.12259734e+00 4.94905263e-02 6.79916203e-01 -6.92026138e-01
8.24818574e-03 4.48368400e-01 1.20340073e+00 2.02853099e-01
5.75212836e-01 5.20143390e-01 1.08921552e+00 -2.73338228e-01
3.59255821e-01 -9.19225067e-02 -2.21932903e-01 1.08714604e+00
5.43820918e-01 -2.54060596e-01 2.68365920e-01 -2.79299527e-01
1.34744120e+00 -9.23763365e-02 1.59703746e-01 -6.55780911e-01
-9.28374708e-01 7.29424238e-01 5.64812243e-01 2.30254605e-01
-4.08159763e-01 2.26922899e-01 9.36613560e-01 -2.76066154e-01
5.15081823e-01 1.06682763e-01 -4.57279772e-01 2.11512566e-01
-1.10693097e+00 -1.85640380e-01 4.55083042e-01 5.29878065e-02
6.19120002e-02 1.96933135e-01 2.67721504e-01 9.76887226e-01
5.49580872e-01 3.42338443e-01 5.89593768e-01 -1.35081160e+00
3.88626873e-01 8.39138091e-01 -2.51614600e-01 -8.17701340e-01
-5.84112346e-01 -4.67526942e-01 -1.19718826e+00 6.64654136e-01
7.35096931e-01 -3.10592175e-01 -6.39376342e-01 1.39966357e+00
7.20205009e-01 -1.95731685e-01 -2.64683247e-01 1.33867848e+00
5.77268183e-01 1.87456146e-01 4.60936368e-01 1.16256304e-01
1.53041947e+00 -6.16216958e-01 -2.12440237e-01 2.07878016e-02
4.59599733e-01 -4.09015447e-01 4.78199065e-01 2.56054521e-01
-1.17722476e+00 -4.04795885e-01 -9.12626147e-01 1.49599925e-01
-2.89994907e-02 4.89104062e-01 5.26818752e-01 1.03756678e+00
-6.92284763e-01 6.30824625e-01 -1.08186483e+00 -1.72025010e-01
8.49735975e-01 2.71194965e-01 -4.53610450e-01 3.79663408e-01
-9.09801602e-01 6.85246944e-01 4.54400331e-01 1.70029894e-01
-5.54569006e-01 -1.09279001e+00 -8.67439806e-01 6.84883818e-02
4.94583212e-02 -8.60300064e-01 6.18461370e-01 -1.36137962e+00
-1.26134622e+00 1.03038514e+00 4.58432823e-01 -7.82206059e-01
7.20827997e-01 -8.69603902e-02 1.30585670e-01 8.91066015e-01
1.04818270e-02 4.91813779e-01 5.85323215e-01 -1.04245090e+00
-1.41910255e-01 -5.86460888e-01 -8.13203156e-01 -1.17440689e-02
1.29481912e-01 -5.25689460e-02 -1.12040322e-02 -1.08017361e+00
4.35684294e-01 -8.70448589e-01 -4.85873610e-01 5.78154087e-01
-1.68775976e-01 5.31576395e-01 6.03002667e-01 -1.29950917e+00
8.35474312e-01 -1.69747066e+00 -3.31926584e-01 5.95025420e-01
5.04755259e-01 3.45700234e-01 9.47430208e-02 -4.03850108e-01
-3.53860617e-01 5.52760601e-01 -5.29413700e-01 2.67881691e-01
-2.30537936e-01 -1.08251877e-01 4.70475078e-01 6.36116207e-01
2.36602828e-01 8.48607719e-01 -7.76350379e-01 -1.01942527e+00
4.42077518e-01 4.36654538e-01 -2.55701572e-01 -9.14015844e-02
2.15065703e-01 3.99997532e-01 -6.37608826e-01 6.69469416e-01
9.07080531e-01 -2.53632199e-02 5.05948126e-01 -2.46495783e-01
1.45345554e-01 -1.48933649e-01 -7.54866362e-01 1.66820228e+00
-4.69021916e-01 5.81626356e-01 6.15370870e-01 -8.19871664e-01
9.00496602e-01 4.61494952e-01 1.26670825e+00 -6.80933774e-01
5.97583532e-01 2.38883533e-02 4.26017493e-01 -5.14816046e-01
-8.89239982e-02 -7.14869320e-01 1.72430530e-01 6.06902719e-01
-3.98968905e-01 9.17626917e-02 4.55666892e-02 7.98270404e-02
6.46369517e-01 8.99652913e-02 4.14849699e-01 -5.80979526e-01
6.54150844e-01 -7.10766017e-02 7.79971600e-01 5.33800162e-02
-6.06067121e-01 7.66189992e-01 4.44623023e-01 -9.35701787e-01
-1.35849905e+00 -1.47521091e+00 -3.48797023e-01 3.84064734e-01
-1.30273640e-01 8.68613645e-02 -9.55922902e-01 -7.49476612e-01
-1.10136054e-01 4.89799790e-02 -6.89548731e-01 2.55271077e-01
-5.84918022e-01 -9.47909772e-01 7.17230320e-01 6.45134032e-01
8.87078404e-01 -6.67964697e-01 -1.22735727e+00 1.02009878e-01
-4.59218591e-01 -8.42969775e-01 -2.04983085e-01 -6.22731857e-02
-1.53086030e+00 -1.24406195e+00 -1.15139580e+00 -4.67156798e-01
7.40347743e-01 -4.63516772e-01 1.18762100e+00 2.05394924e-01
-1.04520214e+00 -1.06209017e-01 2.22720318e-02 -3.98591340e-01
-6.85674131e-01 -2.68424094e-01 -1.68553516e-01 -1.94614053e-01
-2.25744382e-01 -5.06457448e-01 -1.05449593e+00 1.23622388e-01
-6.91713214e-01 2.22220272e-01 6.78120315e-01 5.51464558e-01
7.96339631e-01 -8.43219161e-02 2.55428106e-01 -6.52601242e-01
2.36475170e-01 -5.55941343e-01 -6.50865853e-01 -9.47307274e-02
-3.84135306e-01 -3.84729087e-01 4.39262748e-01 -1.42098710e-01
-7.69418180e-01 6.33686110e-02 5.06773628e-02 -7.31075928e-02
-1.56545311e-01 2.97272354e-01 -9.81118181e-04 -5.19181974e-02
3.23013455e-01 -2.95877010e-02 4.58413064e-01 -2.85967618e-01
-4.03921679e-02 2.10912690e-01 3.26549530e-01 -5.24402976e-01
2.18470976e-01 5.67300141e-01 5.82000673e-01 -6.94904804e-01
-3.01311016e-01 -2.69806862e-01 -8.97690058e-01 -3.27059478e-01
1.24698758e+00 -7.13672638e-01 -6.10128224e-01 -4.98207435e-02
-7.71309316e-01 -7.13449895e-01 -1.70280218e-01 4.82277453e-01
-4.05025005e-01 6.85429096e-01 -7.39125550e-01 -5.64284623e-01
-8.19068313e-01 -7.21118629e-01 6.38923109e-01 8.98290873e-02
-3.82083297e-01 -1.41058981e+00 2.08036959e-01 5.42185366e-01
3.03648561e-01 8.75310421e-01 1.17807829e+00 -4.46561426e-01
-2.00889781e-01 -3.13718110e-01 -4.91719007e-01 6.32007003e-01
2.72724740e-02 -3.68322134e-02 -5.04813015e-01 2.96062797e-01
-2.14019880e-01 -3.31667662e-02 5.90985179e-01 1.06428957e+00
1.26904929e+00 -3.47398013e-01 -9.15039983e-03 6.01579666e-01
1.81836975e+00 7.30706230e-02 4.26905692e-01 2.32116744e-01
4.57641721e-01 6.38176382e-01 4.49769676e-01 4.77194756e-01
1.66551247e-01 8.03982690e-02 6.98187232e-01 -7.72027731e-01
-3.91933978e-01 4.23965782e-01 -1.96366221e-01 6.55711234e-01
-2.59292483e-01 2.55996823e-01 -1.19814956e+00 7.04556048e-01
-9.13640022e-01 -7.26812661e-01 -3.94338846e-01 2.19535470e+00
7.26444185e-01 -7.45676160e-02 3.13000768e-01 -1.46140814e-01
5.65335691e-01 -3.95015568e-01 -4.76831108e-01 -5.74458659e-01
4.15567875e-01 6.13678217e-01 7.72879720e-01 4.03318733e-01
-1.27929044e+00 2.77094513e-01 6.43256903e+00 1.52269632e-01
-1.25758147e+00 1.77165687e-01 1.16343725e+00 -7.45011643e-02
1.36095375e-01 -4.50373173e-01 1.61008194e-01 5.74948668e-01
7.63809204e-01 1.80004925e-01 -1.82690486e-01 6.96580291e-01
4.78994429e-01 -6.97008789e-01 -5.76409876e-01 6.19226992e-01
-2.87015527e-01 -1.26433969e+00 -2.50865351e-02 9.14538726e-02
5.04859209e-01 5.74449822e-02 -2.86817141e-02 -3.78812611e-01
-5.91353141e-03 -9.81523097e-01 2.80962259e-01 4.58917856e-01
9.14915442e-01 -6.96743548e-01 1.05822814e+00 3.73302810e-02
-1.23410904e+00 8.26072022e-02 -1.50530979e-01 2.31776759e-01
-6.70988783e-02 9.21440840e-01 -1.13791978e+00 3.63612562e-01
6.36581659e-01 -1.24635966e-02 -4.72293645e-01 1.20046580e+00
4.89205748e-01 7.96092808e-01 -3.42273027e-01 3.50862056e-01
1.04606599e-01 -6.39153063e-01 4.57967371e-01 1.00784886e+00
4.39800590e-01 4.28473838e-02 8.74274820e-02 1.14078248e+00
-6.74560145e-02 5.62670290e-01 -4.92627591e-01 1.02452673e-01
-4.03933339e-02 1.74329901e+00 -1.41364181e+00 -5.03904641e-01
-3.65382768e-02 5.36995351e-01 -3.68794292e-01 2.64604688e-02
-6.62645161e-01 3.09609193e-02 3.86278629e-01 4.69740182e-01
-2.60764033e-01 -1.31051540e-01 -9.31372881e-01 -5.80919325e-01
-2.57028967e-01 -4.43646163e-01 4.75002587e-01 -4.68138367e-01
-8.86474311e-01 3.06107968e-01 -2.34065562e-01 -1.20819843e+00
6.06943183e-02 -5.57305992e-01 -7.91553438e-01 1.13298178e+00
-1.21886671e+00 -9.53102946e-01 -4.55003798e-01 -2.32854679e-01
3.08893263e-01 -8.82959515e-02 9.58064616e-01 2.77398825e-01
-2.86481023e-01 6.01426102e-02 -2.37738013e-01 7.06120431e-01
4.44877535e-01 -1.62788224e+00 5.33715859e-02 6.36096120e-01
-6.28426373e-01 6.15008771e-01 3.32742363e-01 -1.03617871e+00
-7.41235018e-01 -1.40249944e+00 3.53986740e-01 -1.53510720e-01
2.91282892e-01 9.53097045e-02 -5.94830453e-01 2.65835345e-01
-5.13604060e-02 3.71657163e-01 1.47728240e+00 -5.07100165e-01
6.62392378e-02 -1.14834107e-01 -1.51823938e+00 4.45452422e-01
-1.07620154e-02 1.45569026e-01 -3.75808835e-01 -8.88159722e-02
-7.26311952e-02 -2.07798511e-01 -1.38309681e+00 2.56119490e-01
9.78563726e-01 -9.69594777e-01 1.42740262e+00 -2.75196016e-01
7.91938365e-01 -1.01704113e-01 1.51067570e-01 -8.41716051e-01
-1.91529632e-01 -1.42962128e-01 1.57492667e-01 6.81329191e-01
7.84400329e-02 -1.96091682e-01 1.24352133e+00 6.37471676e-01
7.70877525e-02 -8.34562182e-01 -9.43079233e-01 -4.62053031e-01
4.79294360e-01 -2.48028010e-01 1.86936166e-02 7.95245528e-01
-3.80794048e-01 -3.07626128e-01 3.50842118e-01 -1.24093905e-01
9.45243001e-01 -3.89930010e-02 3.59102219e-01 -1.61185741e+00
2.37109587e-01 -6.63967192e-01 -1.63255453e-01 -1.22930501e-02
-1.54780298e-01 -1.23527789e+00 -2.14412019e-01 -1.65105510e+00
5.32505214e-01 -8.47234666e-01 -5.04707515e-01 3.17439437e-01
3.60421874e-02 8.67206812e-01 2.42530704e-01 -3.81901890e-01
7.31553808e-02 1.05407178e-01 1.45949268e+00 -4.56022993e-02
-1.42110035e-01 -1.35573536e-01 -4.15514439e-01 1.22896683e+00
1.35543382e+00 -4.02374029e-01 -2.67498970e-01 -2.82149874e-02
-4.63920385e-02 3.13189268e-01 7.09777176e-01 -9.52683449e-01
-5.30760586e-01 3.77871282e-02 1.22431302e+00 -6.15872324e-01
2.18569279e-01 -1.17391002e+00 2.79786233e-02 1.07764828e+00
-4.13142353e-01 -1.50582358e-01 2.45223209e-01 7.90245160e-02
1.23886101e-01 -1.79558545e-01 1.26773143e+00 -6.57634795e-01
-1.97760183e-02 3.01085263e-01 -7.46741176e-01 8.54015946e-02
1.09323704e+00 -3.64458144e-01 4.16322082e-01 4.77015264e-02
-9.52534735e-01 -6.31636977e-02 2.33932927e-01 -1.41256467e-01
6.31173790e-01 -1.28907382e+00 -7.32876718e-01 -4.69033718e-02
-3.99077624e-01 -6.30215257e-02 2.77842522e-01 1.30622447e+00
-1.28921068e+00 7.27847964e-02 -6.70538485e-01 -6.08111680e-01
-1.63821971e+00 5.18385470e-02 8.34791124e-01 -5.94014049e-01
-8.46708536e-01 6.86650038e-01 7.49836043e-02 -2.56006628e-01
-2.57918805e-01 -6.21118248e-01 -6.66432828e-02 -5.67275025e-02
2.37707600e-01 5.19860446e-01 -3.52835089e-01 -5.26194215e-01
-3.98054063e-01 6.22066855e-01 4.16843951e-01 -5.59297986e-02
1.08783817e+00 -2.14949697e-01 -4.33094203e-01 2.27816984e-01
1.22930777e+00 1.22594021e-01 -9.32040930e-01 9.68555734e-02
-6.19479679e-02 -4.37471211e-01 4.24072295e-01 -1.25627124e+00
-1.53972983e+00 1.17569053e+00 1.25481153e+00 1.12153940e-01
1.11137652e+00 -3.35265249e-01 1.00437868e+00 -3.97351354e-01
-5.34301028e-02 -1.16197085e+00 1.03615299e-01 -1.45725399e-01
6.71424925e-01 -1.29985321e+00 3.20530891e-01 -5.51424921e-01
-5.63406706e-01 1.43398154e+00 3.77157629e-01 -3.31884533e-01
5.95985353e-01 4.97327834e-01 5.46537995e-01 1.34073928e-01
-6.55921623e-02 2.34046623e-01 4.75490540e-01 5.35276890e-01
6.29818976e-01 2.42860898e-01 -3.44362468e-01 6.27056003e-01
8.71581510e-02 1.60451189e-01 8.77316520e-02 6.01401210e-01
-5.20208597e-01 -8.35134208e-01 -5.82753360e-01 5.29962718e-01
-1.20198989e+00 -2.87250318e-02 -4.24670577e-01 8.52876306e-01
6.06521547e-01 1.56994984e-01 3.03710580e-01 4.64021444e-01
-5.35911024e-01 2.93014556e-01 5.03047764e-01 -3.29309583e-01
-7.98113525e-01 3.94659102e-01 -7.26555288e-02 -3.84651214e-01
-5.55940792e-02 -4.67403233e-01 -1.63885593e+00 1.65229246e-01
2.44531676e-01 -1.82413287e-03 1.03851330e+00 6.17438972e-01
3.95687819e-02 8.08842897e-01 2.71882951e-01 -5.17001092e-01
1.30492613e-01 -8.44362080e-01 -7.16825902e-01 2.45989755e-01
1.75884619e-01 -3.42502624e-01 2.74052292e-01 3.25256288e-01] | [14.278718948364258, -2.5439538955688477] |
5d9ee25d-adc6-4a57-affc-ec880f9a7d5a | babyslm-language-acquisition-friendly | 2306.01506 | null | https://arxiv.org/abs/2306.01506v2 | https://arxiv.org/pdf/2306.01506v2.pdf | BabySLM: language-acquisition-friendly benchmark of self-supervised spoken language models | Self-supervised techniques for learning speech representations have been shown to develop linguistic competence from exposure to speech without the need for human labels. In order to fully realize the potential of these approaches and further our understanding of how infants learn language, simulations must closely emulate real-life situations by training on developmentally plausible corpora and benchmarking against appropriate test sets. To this end, we propose a language-acquisition-friendly benchmark to probe spoken language models at the lexical and syntactic levels, both of which are compatible with the vocabulary typical of children's language experiences. This paper introduces the benchmark and summarizes a range of experiments showing its usefulness. In addition, we highlight two exciting challenges that need to be addressed for further progress: bridging the gap between text and speech and between clean speech and in-the-wild speech. | ['Alejandrina Cristia', 'Emmanuel Dupoux', 'Hervé Bredin', 'Okko Räsänen', 'María Andrea Cruz Blandón', 'Hadrien Titeux', 'Yaya Sy', 'Marvin Lavechin'] | 2023-06-02 | null | null | null | null | ['language-acquisition'] | ['natural-language-processing'] | [ 2.35418916e-01 5.19035637e-01 2.51539331e-02 -8.59019995e-01
-6.51826143e-01 -6.30308807e-01 6.02278411e-01 3.14137131e-01
-4.94358271e-01 3.64494532e-01 4.02709037e-01 -3.54307055e-01
1.92791268e-01 -4.96788621e-01 -6.24381602e-01 -2.41169766e-01
-2.05507800e-01 3.91054541e-01 2.74273425e-01 -3.14053178e-01
-1.59135416e-01 3.01488429e-01 -1.96774352e+00 5.69418669e-01
6.73757017e-01 2.68153131e-01 3.82266492e-01 8.48476768e-01
-4.29940522e-01 8.34584713e-01 -5.41810930e-01 -2.20272467e-01
-1.30739555e-01 -5.17434061e-01 -9.95926738e-01 -6.49444088e-02
6.28580928e-01 -4.31462198e-01 -2.16317043e-01 1.01384294e+00
4.56527203e-01 2.33506814e-01 4.53369617e-01 -5.35615385e-01
-7.00529575e-01 1.06123829e+00 8.19148570e-02 5.02511680e-01
6.85687006e-01 -6.90493733e-02 7.17733383e-01 -7.34163165e-01
6.22298956e-01 1.23419011e+00 2.97344983e-01 1.20350111e+00
-1.16792703e+00 -4.21752661e-01 5.01196563e-01 -1.25205830e-01
-1.04283559e+00 -1.07424462e+00 6.04307950e-01 -6.20458364e-01
1.23435140e+00 5.25221340e-02 5.93414426e-01 1.39707971e+00
-5.97890854e-01 8.18392813e-01 1.16031909e+00 -8.83252561e-01
1.47302404e-01 3.97537738e-01 2.15166762e-01 5.97424448e-01
-1.73696816e-01 4.73567933e-01 -9.12833750e-01 2.80341923e-01
5.87524831e-01 -6.44790947e-01 -4.32576001e-01 -3.50200623e-01
-1.04010379e+00 5.53601325e-01 -9.04467702e-02 7.65711784e-01
9.95666906e-02 7.04487711e-02 5.16424000e-01 5.05794764e-01
3.65906894e-01 3.70535940e-01 -5.83827734e-01 -5.17167568e-01
-9.40010607e-01 -1.00724809e-01 8.09400856e-01 1.04038882e+00
2.70782501e-01 2.34841287e-01 2.35164806e-01 1.36555982e+00
5.72294712e-01 3.17308635e-01 7.22571611e-01 -8.36662889e-01
4.45181161e-01 1.99969128e-01 -3.31641167e-01 -1.50364354e-01
-2.26495177e-01 -1.48956388e-01 -2.06148997e-01 1.05623104e-01
7.90894091e-01 -5.23876250e-02 -7.20132113e-01 2.05767512e+00
2.78379202e-01 2.60203063e-01 4.54534322e-01 4.36874986e-01
1.17846096e+00 6.44979060e-01 3.74842912e-01 -4.15375322e-01
1.08353448e+00 -6.58876061e-01 -5.24560571e-01 -3.53599846e-01
7.41993546e-01 -5.48697650e-01 1.40083289e+00 4.12191808e-01
-1.56006050e+00 -4.66134161e-01 -1.09178305e+00 -1.30439460e-01
-3.95920038e-01 -2.32209772e-01 6.58140004e-01 9.87585127e-01
-1.28445756e+00 5.68112135e-01 -9.73806739e-01 -5.82319021e-01
1.46462113e-01 2.28613421e-01 -5.00658751e-01 1.02708638e-01
-1.24158132e+00 1.02062297e+00 4.82522696e-01 -1.36599764e-01
-1.07747746e+00 -5.67330837e-01 -1.11639440e+00 -1.20427907e-02
-4.79465835e-02 -1.14074007e-01 1.87094343e+00 -9.54860568e-01
-1.74245667e+00 1.42697668e+00 -5.28238043e-02 -3.60500425e-01
2.34635666e-01 -2.96177655e-01 -4.28718328e-01 1.77524522e-01
-1.96467444e-01 7.00502992e-01 3.53236884e-01 -1.04688358e+00
-4.56163883e-01 -2.86482304e-01 1.06687300e-01 1.84192359e-01
-3.25797737e-01 5.64649820e-01 -7.44619891e-02 -7.06371844e-01
8.64455327e-02 -5.66282272e-01 2.74613444e-02 -1.38771310e-01
1.60810068e-01 -3.93644989e-01 -2.74490435e-02 -6.66884124e-01
9.34108496e-01 -2.27615047e+00 -3.40482779e-02 -9.80353057e-02
-1.21737197e-01 5.11208951e-01 -2.64802545e-01 7.82661498e-01
-2.47616872e-01 5.40593751e-02 5.07939011e-02 -5.72138846e-01
-1.12782277e-01 2.53573447e-01 -2.87634790e-01 3.54007185e-01
1.93250015e-01 7.34452903e-01 -1.15327883e+00 -3.00288767e-01
1.86495304e-01 4.44434732e-01 -3.58690590e-01 4.80813950e-01
-3.77125084e-01 6.79259598e-01 -6.38992116e-02 2.40868956e-01
-1.58942807e-02 2.59256572e-01 1.97476581e-01 5.38404524e-01
-3.67016554e-01 1.14456856e+00 -7.58212745e-01 1.73563612e+00
-6.56693339e-01 7.94328570e-01 2.26202562e-01 -1.14842117e+00
8.07312012e-01 7.58142054e-01 -2.74726301e-01 -6.83822513e-01
1.45564139e-01 2.86682159e-01 2.05947176e-01 -7.05558360e-01
-8.20523649e-02 -3.39091986e-01 2.47491032e-01 5.37345588e-01
5.15475690e-01 -3.69247854e-01 8.80967975e-02 -7.31412247e-02
7.32070386e-01 6.89776242e-02 3.88508618e-01 -5.48274636e-01
5.97705007e-01 -4.49357808e-01 1.01268627e-01 5.53841591e-01
-5.37695782e-03 4.84639913e-01 2.30398804e-01 -2.79263049e-01
-7.46293545e-01 -1.21565855e+00 -7.04919100e-02 1.60549033e+00
-4.75968063e-01 -1.83718324e-01 -1.17446017e+00 -4.41853613e-01
-5.22317708e-01 1.18034816e+00 -5.19840896e-01 -3.57261300e-02
-7.25868464e-01 5.88079467e-02 6.29383743e-01 6.52055502e-01
-1.28741756e-01 -1.44099426e+00 -5.59774578e-01 2.19628081e-01
3.95584196e-01 -1.24762809e+00 -7.82711357e-02 2.78617799e-01
-5.96690893e-01 -7.39790261e-01 -6.86485887e-01 -1.29464018e+00
5.90029180e-01 -5.85394315e-02 1.07745290e+00 3.07740420e-01
1.22627877e-02 8.44363511e-01 -5.41917861e-01 -5.62996686e-01
-1.11935842e+00 -5.60025759e-02 1.95273727e-01 -5.12589395e-01
4.25915211e-01 -8.96880746e-01 -1.09611534e-01 -1.42815664e-01
-6.27422214e-01 7.52689466e-02 5.17828725e-02 6.46376371e-01
1.90879047e-01 -3.72384131e-01 6.87372923e-01 -7.39818990e-01
6.07488990e-01 -4.78316993e-01 -6.28638148e-01 4.80126590e-01
-1.61395788e-01 8.61058012e-02 6.04358435e-01 -6.39623940e-01
-1.08176231e+00 9.19328108e-02 -5.55485249e-01 5.55282794e-02
-7.16581225e-01 3.27548623e-01 -1.11907408e-01 2.90975600e-01
7.37381339e-01 4.10450578e-01 -9.11237265e-04 -6.17321193e-01
5.47604442e-01 8.06158364e-01 6.96660817e-01 -1.06992817e+00
4.86320645e-01 -9.77082774e-02 -6.80331290e-01 -1.34683061e+00
-8.54369164e-01 -1.59940347e-01 -7.27166831e-01 -6.56054318e-02
7.42359102e-01 -9.55126166e-01 -2.87372917e-01 4.26827937e-01
-1.28358579e+00 -8.36738527e-01 -5.14487445e-01 7.65498102e-01
-7.03001142e-01 2.52835184e-01 -6.33483410e-01 -9.61491525e-01
-3.23353522e-02 -1.00121224e+00 6.92758918e-01 2.44055316e-01
-5.63691497e-01 -1.15572655e+00 4.13361341e-01 3.10707062e-01
2.54234552e-01 -1.77276611e-01 9.40463364e-01 -1.24209750e+00
-1.27803504e-01 2.51501679e-01 3.19082439e-01 7.79532492e-01
5.71454577e-02 1.20353982e-01 -1.53399301e+00 -2.84309059e-01
2.21029431e-01 -7.28315711e-01 4.15176868e-01 -2.63823178e-02
7.12707520e-01 7.36038527e-03 1.39793962e-01 3.39634567e-01
8.97467136e-01 4.34819788e-01 1.52613983e-01 2.30033975e-02
2.96542674e-01 1.34672785e+00 1.97221950e-01 -1.48195431e-01
5.41508615e-01 3.15783739e-01 -1.92680880e-01 2.18628719e-01
-4.40666705e-01 -5.52820086e-01 4.01866853e-01 1.38811100e+00
1.85445175e-01 -1.77436367e-01 -1.33501863e+00 1.03628731e+00
-1.24172807e+00 -8.38727117e-01 2.69022763e-01 2.19987249e+00
1.22276902e+00 2.71559358e-01 2.02387989e-01 3.00719410e-01
5.88222504e-01 8.19970295e-02 -1.99377850e-01 -8.73286724e-01
1.61155865e-01 5.96260369e-01 -3.66696447e-01 8.24304104e-01
-7.58286595e-01 1.18414199e+00 7.04502201e+00 2.81672060e-01
-1.17899323e+00 9.23639312e-02 5.00565529e-01 4.21042927e-02
-4.90040958e-01 -3.50932479e-01 -7.56252110e-01 2.12790877e-01
1.45371151e+00 -1.03763156e-01 6.87652290e-01 8.18201184e-01
3.44282351e-02 3.33681703e-02 -1.86017966e+00 8.08669746e-01
2.43875608e-02 -9.05625403e-01 -2.89436609e-01 -5.73437274e-01
3.16890568e-01 3.69847834e-01 -1.16632313e-01 4.09251302e-01
3.73566866e-01 -1.25677729e+00 8.33248198e-01 1.69259876e-01
9.24971342e-01 -2.55509377e-01 1.81425169e-01 7.26993144e-01
-8.79139721e-01 2.72450209e-01 -3.52084897e-02 -3.62916917e-01
3.95331681e-02 -6.16077632e-02 -1.09974170e+00 -1.70856118e-01
6.14683211e-01 3.12521636e-01 -3.93902212e-01 6.60470843e-01
-6.37287617e-01 1.00759220e+00 -5.29506981e-01 -4.57820207e-01
1.25482172e-01 3.39741439e-01 2.01591626e-01 1.51460636e+00
1.09940298e-01 3.81667852e-01 -2.26695642e-01 6.83443904e-01
3.01855266e-01 4.90829617e-01 -8.26001704e-01 -4.06572014e-01
7.58135021e-01 4.85542357e-01 -6.08310938e-01 -1.63912311e-01
-8.59802246e-01 4.07591224e-01 6.11009002e-01 1.05741300e-01
-6.11788221e-02 2.25184336e-02 6.80171669e-01 2.21579060e-01
1.30505174e-01 -4.55323488e-01 1.00904668e-03 -1.04540193e+00
-3.78781296e-02 -1.08668804e+00 -1.46054983e-01 -4.97829497e-01
-1.14446294e+00 6.63190007e-01 3.62577319e-01 -6.00170553e-01
-8.23799729e-01 -7.25168407e-01 -7.41052568e-01 8.09084535e-01
-1.22998834e+00 -7.99497843e-01 8.83591846e-02 3.45122755e-01
1.01001191e+00 -2.49083281e-01 1.15138721e+00 4.37624566e-02
-3.40896189e-01 7.43362904e-01 -3.04893762e-01 3.02622229e-01
2.05723420e-01 -1.16790473e+00 1.03644681e+00 6.64907157e-01
7.57382214e-01 8.91612232e-01 8.52333665e-01 -2.88468748e-01
-8.75963449e-01 -4.10157293e-01 9.82234299e-01 -4.15176243e-01
8.05563509e-01 -7.38477826e-01 -1.24274361e+00 7.18756557e-01
4.89830762e-01 -8.89840350e-02 8.66694331e-01 2.03120455e-01
-4.42611277e-01 1.14822246e-01 -9.64823127e-01 4.56382155e-01
1.33926392e+00 -9.23157036e-01 -1.30298471e+00 2.35990345e-01
8.48062217e-01 -4.11970019e-01 -8.94808233e-01 3.92355084e-01
5.93351364e-01 -1.04599929e+00 6.89195931e-01 -5.96149087e-01
1.91778243e-01 1.88541487e-01 -2.14375824e-01 -1.28881180e+00
2.47525081e-01 -8.02495122e-01 1.41356692e-01 1.75614417e+00
5.21959841e-01 -5.15741944e-01 5.08296371e-01 6.53841317e-01
-2.34433413e-01 -8.78926098e-01 -9.83472943e-01 -6.63700223e-01
6.79867923e-01 -7.97093689e-01 5.47705054e-01 8.84854913e-01
3.72514039e-01 3.54844391e-01 3.25947255e-01 1.27007812e-01
3.37069601e-01 -4.39461827e-01 4.27220523e-01 -1.09062052e+00
-3.29530537e-01 -4.11188781e-01 -3.39054167e-01 -9.65089858e-01
4.73933458e-01 -7.78380573e-01 2.54460722e-01 -1.20072329e+00
-2.98241735e-01 -3.72077674e-01 -6.99308515e-02 2.21709222e-01
1.56354293e-01 -3.63371491e-01 7.28015900e-02 -4.07646000e-01
-3.19235504e-01 2.88915753e-01 7.82239556e-01 2.13081822e-01
-9.66579244e-02 4.86445576e-02 -6.18367970e-01 1.10637724e+00
1.03680444e+00 -3.19010943e-01 -1.00798547e+00 -8.51860344e-01
6.52620792e-02 1.99150413e-01 -1.69751361e-01 -8.63984287e-01
2.17324689e-01 -1.42114297e-01 -1.28909022e-01 -3.00327241e-01
2.13246629e-01 -4.28516120e-01 -4.34402555e-01 9.01931003e-02
-8.50859463e-01 -1.24508135e-01 2.88058162e-01 -2.44838558e-02
-2.73776978e-01 -6.40001595e-01 9.12880719e-01 -2.66339868e-01
-7.79593170e-01 -1.68606445e-01 -5.72766244e-01 5.17878175e-01
9.04297411e-01 -2.82785930e-02 -2.69120961e-01 -4.02401328e-01
-1.09904563e+00 2.09113002e-01 5.09908199e-01 6.79806948e-01
6.01436615e-01 -8.51502240e-01 -7.15472937e-01 6.09428287e-01
1.79141030e-01 6.87166154e-02 -5.53067699e-02 1.98363960e-01
-5.59922576e-01 3.51817280e-01 1.03209592e-01 -4.08022374e-01
-1.43403816e+00 5.41814804e-01 4.33791876e-01 3.01982343e-01
-3.77925158e-01 1.29105282e+00 3.29376459e-01 -5.60436487e-01
8.22069585e-01 -6.48323655e-01 -5.10649860e-01 -1.98289409e-01
9.15498137e-01 1.68385208e-01 -8.08560476e-02 -6.63808107e-01
-3.60489696e-01 4.20861185e-01 -1.15167148e-01 -3.56250167e-01
1.34553254e+00 -1.09910227e-01 1.92615371e-02 1.05670917e+00
1.05990648e+00 2.79980481e-01 -1.11295891e+00 -2.76906431e-01
4.05076563e-01 -2.66225805e-04 -3.09385657e-01 -6.46901786e-01
-6.80770993e-01 1.35739946e+00 5.33130109e-01 4.48374748e-01
7.61186481e-01 5.83298922e-01 4.93523479e-01 4.98427629e-01
3.34028095e-01 -9.32622969e-01 -1.08754516e-01 4.88764107e-01
1.10012007e+00 -9.99385536e-01 -3.78771365e-01 -5.58124959e-01
-4.21249568e-01 1.04224157e+00 5.95961511e-01 -2.00655088e-01
6.64146781e-01 5.01434445e-01 4.44635689e-01 -5.29915281e-03
-1.05339336e+00 -2.49690548e-01 2.60158867e-01 8.97643685e-01
9.53880370e-01 1.95248779e-02 -9.12787467e-02 5.70240855e-01
-7.30838358e-01 -2.86631197e-01 3.50156039e-01 8.18095207e-01
-6.00287199e-01 -1.33812392e+00 -1.18358076e-01 1.20947799e-02
-5.72172523e-01 -3.07972431e-01 -4.81393129e-01 7.15809524e-01
1.74590707e-01 1.00667584e+00 -1.61434084e-01 -1.33380331e-02
5.05866826e-01 4.38730896e-01 8.64813149e-01 -1.40936565e+00
-6.31898880e-01 -1.90775786e-02 5.06322742e-01 -3.82609397e-01
-4.60467279e-01 -9.90806520e-01 -1.27223706e+00 3.63065183e-01
-1.57288641e-01 2.03523695e-01 7.17592955e-01 9.96642530e-01
-3.94068420e-01 2.67415822e-01 2.83476055e-01 -7.85919368e-01
-6.39422297e-01 -9.03676152e-01 -2.92212039e-01 -1.14075246e-03
6.49172068e-01 -4.16054279e-01 -5.58070779e-01 1.31823838e-01] | [10.773660659790039, 9.091550827026367] |
eeacc466-44d8-4efd-9e64-1fde43b0897f | small-footprint-keyword-spotting-on-raw-audio | 1911.02086 | null | https://arxiv.org/abs/1911.02086v2 | https://arxiv.org/pdf/1911.02086v2.pdf | Small-Footprint Keyword Spotting on Raw Audio Data with Sinc-Convolutions | Keyword Spotting (KWS) enables speech-based user interaction on smart devices. Always-on and battery-powered application scenarios for smart devices put constraints on hardware resources and power consumption, while also demanding high accuracy as well as real-time capability. Previous architectures first extracted acoustic features and then applied a neural network to classify keyword probabilities, optimizing towards memory footprint and execution time. Compared to previous publications, we took additional steps to reduce power and memory consumption without reducing classification accuracy. Power-consuming audio preprocessing and data transfer steps are eliminated by directly classifying from raw audio. For this, our end-to-end architecture extracts spectral features using parametrized Sinc-convolutions. Its memory footprint is further reduced by grouping depthwise separable convolutions. Our network achieves the competitive accuracy of 96.4% on Google's Speech Commands test set with only 62k parameters. | ['Ludwig Kürzinger', 'Simon Mittermaier', 'Gerhard Rigoll', 'Bernd Waschneck'] | 2019-11-05 | null | null | null | null | ['small-footprint-keyword-spotting'] | ['speech'] | [ 4.10982668e-01 -1.30353510e-01 -7.90188462e-03 -4.99076635e-01
-9.62845027e-01 -5.49873888e-01 1.49526805e-01 -7.36845424e-04
-8.36661398e-01 1.96642190e-01 -2.49774922e-02 -8.80524874e-01
7.57282302e-02 -6.35106921e-01 -3.71220142e-01 -4.78424430e-01
1.09725595e-02 6.42383993e-02 3.16968352e-01 3.03778827e-01
8.23627412e-02 5.55122197e-01 -1.88895738e+00 3.40941876e-01
6.88804150e-01 1.49232936e+00 4.02763486e-01 1.25146103e+00
-2.16732314e-03 3.06105465e-01 -8.71897399e-01 -2.58006632e-01
2.90514193e-02 1.61932305e-01 -4.61810172e-01 -5.01129687e-01
3.46289128e-01 -6.52185559e-01 -6.05525196e-01 8.58989298e-01
1.12132823e+00 2.11146474e-01 2.58087993e-01 -1.27103484e+00
-3.31973642e-01 7.42268264e-01 1.45601586e-01 3.78082782e-01
4.21107829e-01 2.25256458e-01 9.60997522e-01 -9.09521222e-01
-2.97352970e-01 7.92661488e-01 5.94713390e-01 4.80675489e-01
-1.08893561e+00 -9.20004606e-01 -4.02544215e-02 4.88024265e-01
-1.68917429e+00 -1.10638309e+00 5.46325922e-01 -2.15566158e-03
2.02895164e+00 8.02262306e-01 5.80701649e-01 9.92016137e-01
-2.22787052e-01 7.89760292e-01 5.46565235e-01 -3.55007678e-01
4.24697369e-01 1.38271734e-01 1.60343409e-01 3.48599613e-01
-1.55050829e-01 -2.49655262e-01 -9.10584331e-01 -3.61810595e-01
9.27207842e-02 -6.23669289e-02 -4.44955677e-01 4.54605132e-01
-1.01482546e+00 3.87645751e-01 -1.60057202e-01 1.50243789e-01
-2.58573145e-01 2.42630735e-01 4.94102508e-01 2.51612455e-01
2.38110811e-01 1.61737442e-01 -8.21802318e-01 -8.94125581e-01
-1.31258810e+00 -3.23996961e-01 9.04048085e-01 9.77333546e-01
3.40042710e-01 4.64810818e-01 -1.04016043e-01 9.23995376e-01
1.72089696e-01 8.47939432e-01 8.79860759e-01 -5.68994045e-01
5.81673086e-01 3.03441435e-01 -3.33873242e-01 -3.95772010e-01
-6.41108096e-01 -2.72822052e-01 -6.34901464e-01 -2.20609773e-02
2.08624199e-01 -1.09825805e-01 -1.04019892e+00 1.40138543e+00
5.99117167e-02 1.26564667e-01 -6.89507499e-02 7.69501328e-01
6.79501176e-01 8.28460157e-01 4.22823214e-04 -1.11499012e-01
1.67739487e+00 -7.43194044e-01 -8.22676122e-01 -1.06692493e-01
5.80186546e-01 -8.81814241e-01 1.56102133e+00 5.94009757e-01
-9.95725155e-01 -3.91705394e-01 -1.30744147e+00 -3.15925092e-01
-5.03167689e-01 4.52486157e-01 5.48614621e-01 1.31333280e+00
-1.22939706e+00 5.44927835e-01 -1.00970733e+00 -1.30006596e-02
3.66523832e-01 1.02870822e+00 -4.74749971e-03 6.38308644e-01
-1.11419594e+00 3.79005224e-01 2.44770885e-01 -8.45694318e-02
-3.41276437e-01 -8.85242701e-01 -6.47305369e-01 6.64332271e-01
2.47548848e-01 -2.89646119e-01 1.56058431e+00 -5.40936589e-01
-2.32517910e+00 2.45619282e-01 -2.80254185e-01 -7.68854856e-01
2.08100658e-02 -2.05811799e-01 -9.71838653e-01 3.22360285e-02
-5.49120307e-01 4.16646272e-01 1.08008635e+00 -2.18437374e-01
-9.35085118e-01 -4.82004769e-02 -3.03198963e-01 3.68649289e-02
-1.18011713e+00 2.98921838e-02 -6.36013389e-01 -5.28187037e-01
-1.39299862e-03 -8.90418172e-01 2.45986938e-01 -4.27897662e-01
-2.93788016e-01 -3.65314633e-01 1.11813998e+00 -1.05631721e+00
1.63008690e+00 -2.26107693e+00 -5.24302304e-01 3.18878531e-01
3.00791161e-03 5.93580544e-01 -6.97715729e-02 -1.57820620e-02
8.08238685e-02 3.58666517e-02 2.19377145e-01 -5.61640680e-01
3.07866037e-01 -1.11897722e-01 -4.25294608e-01 2.65423626e-01
1.43336311e-01 7.62552977e-01 -3.09445798e-01 -1.72160581e-01
5.25817275e-01 8.04784894e-01 -8.24646533e-01 1.76610142e-01
7.47811049e-03 -1.73062518e-01 1.34743452e-01 7.41346300e-01
4.37440783e-01 7.14756399e-02 6.85777366e-02 -3.10051173e-01
-5.47481589e-02 1.36799145e+00 -1.12665057e+00 1.52827632e+00
-1.10934126e+00 9.31359529e-01 1.79390714e-01 -6.90282345e-01
5.46460748e-01 4.32095468e-01 2.24978834e-01 -9.29489374e-01
3.12315077e-01 3.75901431e-01 1.08916657e-02 -3.71510565e-01
5.24999559e-01 6.00642979e-01 -8.67000744e-02 2.85815060e-01
7.04872012e-02 -1.91551119e-01 -3.14249754e-01 -1.67132057e-02
1.16273141e+00 -2.85703510e-01 8.02166909e-02 -1.43946633e-01
4.78915423e-01 -4.55319673e-01 1.94230035e-01 4.69983637e-01
1.97931044e-02 2.81690240e-01 1.38068339e-03 -1.45674616e-01
-8.80424082e-01 -9.36870098e-01 -1.68804243e-01 1.48210669e+00
-4.77121949e-01 -8.16289067e-01 -9.44937348e-01 -3.09568167e-01
-2.31381759e-01 9.15677011e-01 8.32819417e-02 -1.20892905e-01
-6.29516244e-01 -5.26484251e-01 9.57771242e-01 6.14736199e-01
2.25006595e-01 -7.37479746e-01 -8.81620824e-01 3.65314454e-01
1.62997708e-01 -1.33646679e+00 -7.22993433e-01 7.28321493e-01
-4.41046774e-01 -4.11117911e-01 -3.62249851e-01 -5.26770115e-01
2.53269345e-01 8.92691836e-02 7.97826409e-01 -1.60084292e-01
-4.52122360e-01 3.09995532e-01 -1.76739171e-01 -5.46611965e-01
-2.10184127e-01 5.64997971e-01 5.96452713e-01 -1.16086505e-01
5.33608317e-01 -9.15795624e-01 -7.06385970e-01 2.02271327e-01
-5.16951501e-01 -1.06058173e-01 5.89643419e-01 6.15974963e-01
3.72850746e-01 2.00831696e-01 4.40928757e-01 3.27422842e-02
6.51841581e-01 7.70386215e-03 -8.08991849e-01 2.70990115e-02
-8.24645817e-01 -2.19768398e-02 8.85989070e-01 -7.73294389e-01
-6.33304834e-01 4.00736749e-01 -6.25808120e-01 -2.80573457e-01
-1.25504896e-01 -1.10168047e-01 -4.38399494e-01 -7.74937421e-02
2.88887680e-01 5.10219693e-01 -2.26590082e-01 -5.98547697e-01
1.44177616e-01 1.70950639e+00 5.33346593e-01 -8.46026540e-02
4.68391061e-01 6.43908530e-02 -5.15883446e-01 -1.33733332e+00
-1.05061591e-01 -3.14520180e-01 -3.17624956e-01 1.69018015e-01
5.78465521e-01 -9.39102113e-01 -1.38299990e+00 4.91266727e-01
-1.03911769e+00 -2.96319246e-01 -6.98112696e-02 6.87690020e-01
-1.35766730e-01 1.55551434e-01 -3.50309253e-01 -1.15796530e+00
-1.05920613e+00 -1.19562602e+00 1.19812238e+00 1.94890469e-01
-5.42161584e-01 -3.65293264e-01 -5.25292456e-01 3.08937162e-01
7.49819458e-01 -8.85926604e-01 6.17926836e-01 -8.99358392e-01
-4.26271349e-01 -3.40815425e-01 -4.51925322e-02 5.16143620e-01
1.01301938e-01 -2.93287426e-01 -1.61147022e+00 -1.69035479e-01
-5.03318757e-02 -3.42747313e-03 5.96957624e-01 2.80198485e-01
1.78371179e+00 -6.01038694e-01 -2.91397363e-01 7.44626343e-01
7.77544081e-01 5.26875019e-01 4.04722750e-01 -2.44083971e-01
5.62269330e-01 1.03263177e-01 5.81799746e-02 4.18512881e-01
2.16321230e-01 9.94364679e-01 5.82731850e-02 1.03066504e-01
-1.91544786e-01 -9.27684680e-02 5.27734339e-01 1.23447418e+00
3.40653241e-01 -4.32069153e-01 -9.75498617e-01 5.01130819e-01
-1.40459585e+00 -6.61910594e-01 1.89820871e-01 2.31492734e+00
9.87622261e-01 3.98717552e-01 1.07739531e-01 7.41752863e-01
2.90926397e-01 -4.75138314e-02 -5.85306466e-01 -6.19451225e-01
1.08223751e-01 8.25899601e-01 7.78802752e-01 5.66908300e-01
-1.02237236e+00 7.32988596e-01 6.00613642e+00 1.32170582e+00
-1.52733898e+00 3.36231112e-01 5.31331539e-01 -9.40925479e-01
-4.62986790e-02 -5.66197813e-01 -1.02734685e+00 6.86677516e-01
1.89522469e+00 1.48863137e-01 8.83782685e-01 1.06869829e+00
1.91309482e-01 -3.31921950e-02 -1.13125336e+00 1.49791968e+00
-7.61994645e-02 -1.18996465e+00 -4.25863415e-01 9.95600447e-02
-8.17957744e-02 2.80033767e-01 2.01712653e-01 2.73471624e-01
-1.60200939e-01 -1.10206592e+00 7.59659171e-01 3.69771123e-02
1.27881396e+00 -9.54014301e-01 4.73236471e-01 1.08325131e-01
-1.38251114e+00 -2.06865832e-01 3.83702666e-02 -2.90662944e-01
-1.90192536e-02 8.51593614e-01 -1.45819700e+00 -1.00694731e-01
1.13087654e+00 -2.37886295e-01 -3.84967089e-01 5.14614522e-01
5.75850382e-02 9.65346336e-01 -9.59305644e-01 -7.18361318e-01
-1.49915799e-01 2.24136397e-01 1.76227242e-01 1.58439231e+00
4.21977937e-01 8.93219411e-02 -1.60560474e-01 3.36864740e-01
-1.13840513e-01 -7.26264268e-02 -1.79896086e-01 -2.36383989e-01
1.02187121e+00 1.39691317e+00 -6.76284730e-01 -3.30363035e-01
-3.06530505e-01 1.14562523e+00 -4.71061021e-02 1.27536014e-01
-9.69953656e-01 -9.34593141e-01 1.18184257e+00 9.00111422e-02
3.19482177e-01 -5.37288904e-01 -5.67328990e-01 -8.13025534e-01
2.04736829e-01 -6.88432872e-01 -9.72111449e-02 -3.44322950e-01
-5.74008942e-01 6.82434618e-01 -5.15932441e-01 -9.19030488e-01
-3.67290080e-01 -6.67275965e-01 -3.14233035e-01 1.03115451e+00
-1.17063189e+00 -8.02148998e-01 -8.07510540e-02 4.95159119e-01
7.57922530e-01 -2.11434007e-01 1.21170449e+00 7.29311228e-01
-5.00117302e-01 1.20086753e+00 -4.31196503e-02 9.91950408e-02
4.09146488e-01 -1.03803194e+00 9.36654806e-01 7.45808721e-01
3.26427579e-01 6.97285950e-01 3.46260458e-01 -2.46511206e-01
-1.86344862e+00 -7.92389512e-01 9.01361942e-01 -1.81431308e-01
6.76600218e-01 -1.03965461e+00 -7.21524119e-01 7.45153800e-02
1.48770928e-01 5.17733246e-02 9.86195862e-01 8.53081420e-02
-3.57383281e-01 -2.53991514e-01 -8.92686546e-01 8.04682136e-01
1.17587972e+00 -1.09280598e+00 1.47070102e-02 2.72760510e-01
9.63823020e-01 -4.66624320e-01 -7.44641483e-01 1.22159071e-01
7.25342631e-01 -5.44415832e-01 9.52596545e-01 -1.73129007e-01
-4.69641566e-01 -2.96701729e-01 -3.30723554e-01 -9.75669324e-01
2.00884487e-03 -1.13552928e+00 -5.36076605e-01 1.19473493e+00
6.88980281e-01 -7.55151153e-01 8.45882177e-01 6.56186640e-01
-2.19045267e-01 -7.37757802e-01 -1.50310969e+00 -6.80827022e-01
-5.14958978e-01 -1.26418245e+00 8.62803817e-01 3.78397077e-01
2.72388577e-01 2.44839489e-01 -1.24354154e-01 4.76088285e-01
5.02608567e-02 -3.35713089e-01 4.80242252e-01 -7.87196934e-01
-4.00150090e-01 -5.67724466e-01 -3.58182877e-01 -1.06448412e+00
1.25875726e-01 -6.82056487e-01 -2.93591395e-02 -9.15608227e-01
-4.52927768e-01 -3.87564361e-01 -7.16539770e-02 6.88658059e-01
1.61990404e-01 3.70260477e-01 5.76512306e-04 -1.43226251e-01
-3.50810021e-01 2.50646204e-01 1.76747203e-01 -2.94449836e-01
-4.51837778e-01 8.12757760e-02 -2.69172639e-01 5.52098155e-01
1.06248486e+00 -1.53640360e-01 -5.07793367e-01 -3.91578555e-01
1.13222115e-01 -2.52072424e-01 1.64577425e-01 -1.42615712e+00
4.84404236e-01 2.66613185e-01 2.96357810e-01 -6.28489375e-01
8.18501770e-01 -1.01813209e+00 -3.86293530e-02 5.15652478e-01
-3.06116361e-02 -1.09023608e-01 4.04830843e-01 1.39882088e-01
7.85401016e-02 -8.84847902e-03 5.66683352e-01 6.27735794e-01
-5.24978101e-01 -9.42075476e-02 -7.16188490e-01 -4.88434643e-01
7.22637951e-01 -2.60885060e-01 -1.06044114e-01 -3.88192832e-01
-6.42767668e-01 -2.48288751e-01 -1.12627320e-01 5.16588509e-01
5.93609035e-01 -1.00984454e+00 -8.84333700e-02 8.14862370e-01
-2.50385821e-01 -9.74396542e-02 2.80498832e-01 6.49928093e-01
-4.49449092e-01 8.39556694e-01 3.25787812e-01 -6.40095055e-01
-1.78172052e+00 1.91168725e-01 3.21874827e-01 2.50697196e-01
-5.69416881e-01 1.13519275e+00 -6.65445685e-01 -1.73737377e-01
8.90941560e-01 -7.25148678e-01 1.48065478e-01 2.78038383e-02
8.86709511e-01 5.52675724e-01 8.79421711e-01 -1.46083608e-01
-7.75605738e-01 3.37659776e-01 9.59230214e-02 -2.93675572e-01
1.07210791e+00 4.65708710e-02 3.32262605e-01 1.71686634e-01
1.41891599e+00 2.68421382e-01 -8.02351296e-01 -9.50555131e-02
2.05272827e-02 -8.83135572e-02 5.32930255e-01 -8.52236271e-01
-9.70682085e-01 9.39515650e-01 1.01646864e+00 3.20737183e-01
1.43426442e+00 -2.04674929e-01 1.16088760e+00 6.01046443e-01
2.27468595e-01 -1.49269974e+00 -3.17267627e-01 5.36097825e-01
5.49521446e-01 -9.28192139e-01 -2.31537774e-01 -3.65517229e-01
-1.66208833e-01 1.11313903e+00 3.41263175e-01 3.26663882e-01
8.79897118e-01 1.24326360e+00 -1.37725338e-01 3.20783764e-01
-7.49338329e-01 -8.12891200e-02 4.78881359e-01 7.03923225e-01
2.89950609e-01 3.19030792e-01 8.86769891e-02 9.78222072e-01
-8.40505242e-01 -1.61352560e-01 1.69205423e-02 7.59908617e-01
-3.65364343e-01 -7.94460118e-01 -3.70484769e-01 6.77165389e-01
-5.87879956e-01 -5.85385799e-01 -1.68655798e-01 -2.77423915e-02
-9.33932979e-03 1.35339510e+00 5.58788121e-01 -1.06902432e+00
3.00272286e-01 4.45532411e-01 5.18506318e-02 -4.28216457e-01
-8.41343939e-01 2.25913957e-01 2.18346119e-01 -6.93945348e-01
2.37501889e-01 -3.57550830e-01 -1.38660610e+00 -3.61370564e-01
-5.56925356e-01 -8.10507610e-02 1.47373939e+00 6.86462164e-01
9.75580573e-01 1.00681520e+00 5.61272681e-01 -9.29554999e-01
-5.91681361e-01 -1.12023044e+00 -1.18672818e-01 -3.72997999e-01
3.57740313e-01 -1.36891484e-01 -4.39043820e-01 -9.10013244e-02] | [14.435965538024902, 5.864406108856201] |
cc37530d-ea37-46ea-9c77-23c433efbf9d | rule-based-classification-of-hyperspectral | 2107.10638 | null | https://arxiv.org/abs/2107.10638v1 | https://arxiv.org/pdf/2107.10638v1.pdf | Rule-Based Classification of Hyperspectral Imaging Data | Due to its high spatial and spectral information content, hyperspectral imaging opens up new possibilities for a better understanding of data and scenes in a wide variety of applications. An essential part of this process of understanding is the classification part. In this article we present a general classification approach based on the shape of spectral signatures. In contrast to classical classification approaches (e.g. SVM, KNN), not only reflectance values are considered, but also parameters such as curvature points, curvature values, and the curvature behavior of spectral signatures are used to develop shape-describing rules in order to use them for classification by a rule-based procedure using IF-THEN queries. The flexibility and efficiency of the methodology is demonstrated using datasets from two different application fields and leads to convincing results with good performance. | ['Frank Boochs', 'Alain Tremeau', 'Songuel Polat'] | 2021-07-21 | null | null | null | null | ['classification'] | ['methodology'] | [ 5.69539189e-01 -5.13828158e-01 -9.88488793e-02 -4.64161843e-01
-7.05427006e-02 -9.30347502e-01 6.09598339e-01 3.89359355e-01
-1.95154727e-01 7.70303369e-01 -4.58128154e-01 -4.17094857e-01
-8.33786964e-01 -9.00347829e-01 -8.98668319e-02 -9.67424035e-01
-1.08706839e-01 2.33997315e-01 2.35285789e-01 -3.88560802e-01
5.84368825e-01 1.13143480e+00 -1.77089417e+00 1.77380472e-01
1.16566408e+00 1.03407216e+00 2.34581158e-01 2.54555196e-01
-1.32308990e-01 1.05914153e-01 -1.17987983e-01 7.88384303e-02
3.08670372e-01 -2.27917686e-01 -8.47783506e-01 3.99919540e-01
-1.49334772e-02 -2.27817427e-02 4.22243446e-01 1.19272828e+00
-3.11994255e-02 3.29688966e-01 1.00792360e+00 -7.35634446e-01
-2.73821414e-01 7.43777454e-02 -3.54125053e-01 -2.65239358e-01
4.84334737e-01 -3.16518955e-02 1.11323440e+00 -6.17860019e-01
5.33583879e-01 6.12042785e-01 5.09641051e-01 -2.36020200e-02
-1.39135313e+00 -5.48214559e-03 -1.62769124e-01 4.69686568e-01
-1.17835367e+00 6.68914467e-02 7.03438222e-01 -7.81861424e-01
4.47895199e-01 5.42486429e-01 7.78110623e-01 2.57620543e-01
-2.28857264e-01 3.43273759e-01 1.54591715e+00 -6.09862685e-01
4.06734467e-01 4.38627571e-01 5.07475495e-01 3.87142360e-01
3.65544856e-01 7.79287070e-02 1.08626731e-01 -2.39123985e-01
4.99088526e-01 -4.87830974e-02 -3.40795875e-01 -7.42900729e-01
-8.59226346e-01 9.78245020e-01 5.38393557e-01 4.49385673e-01
-3.64870131e-01 -5.65895915e-01 1.31646439e-01 1.73287824e-01
1.71746284e-01 8.00864279e-01 -3.90077889e-01 2.67094105e-01
-6.82838559e-01 3.58200446e-02 7.48775840e-01 2.13357419e-01
1.07431710e+00 -1.38942942e-01 2.06623152e-01 9.82116222e-01
2.01335236e-01 5.28047502e-01 1.91372782e-01 -5.45680106e-01
-1.53238559e-03 8.87020230e-01 1.41808257e-01 -1.07773638e+00
-5.63342929e-01 -4.18898284e-01 -6.72690570e-01 5.93860984e-01
5.12830973e-01 1.11418866e-01 -7.77025759e-01 1.07603812e+00
3.59919667e-01 -3.61409485e-01 2.63304710e-01 8.44082713e-01
4.92995054e-01 6.05783284e-01 -8.47208425e-02 -3.21801692e-01
1.06813276e+00 -3.00544918e-01 -2.18412161e-01 2.23093644e-01
5.56348026e-01 -7.59328783e-01 9.53175306e-01 7.30148733e-01
-3.90232861e-01 -3.40109766e-01 -8.25687349e-01 4.66105878e-01
-7.78812289e-01 5.19814014e-01 7.17023432e-01 7.07652390e-01
-6.52195871e-01 8.83536398e-01 -4.10767674e-01 -6.62111044e-01
1.60311282e-01 4.81245846e-01 -3.39807838e-01 8.36350396e-02
-7.65947223e-01 9.55227435e-01 7.25214601e-01 2.22709879e-01
7.28230998e-02 -3.59535217e-01 -5.94970584e-01 1.56688653e-02
2.98377275e-01 1.09391641e-02 8.29481006e-01 -9.27847862e-01
-1.49315035e+00 5.74448287e-01 -1.04261227e-01 -1.05380796e-01
4.33476210e-01 2.33433306e-01 -3.43855590e-01 2.94129908e-01
-2.96120435e-01 7.96077922e-02 5.70545912e-01 -1.33063328e+00
-4.25982356e-01 -4.30546731e-01 -1.91776417e-02 -7.79170766e-02
-2.94011027e-01 8.36778358e-02 2.70065635e-01 -1.33784175e-01
5.16431034e-01 -8.22284341e-01 -1.59782633e-01 -3.43588647e-03
-3.38109702e-01 3.46705355e-02 9.09525812e-01 -5.26502013e-01
7.44372904e-01 -1.91946137e+00 6.99406117e-02 9.67216492e-01
-2.06200615e-01 5.49495339e-01 9.39531773e-02 6.68552279e-01
-2.56041259e-01 4.85736541e-02 -5.92465580e-01 6.61825180e-01
-3.46058339e-01 1.17284060e-01 -2.16211513e-01 4.67040330e-01
2.06054464e-01 3.85875016e-01 -7.84424782e-01 1.50900781e-01
6.39684558e-01 4.04383391e-01 -6.39481843e-02 -2.32917611e-02
-2.36635819e-01 3.85202736e-01 -5.47643363e-01 6.46583617e-01
6.74648643e-01 -2.30305687e-01 2.13257506e-01 -3.40716451e-01
-5.04992306e-01 -1.53426707e-01 -1.32643628e+00 7.66317904e-01
-2.77066052e-01 5.36469340e-01 -1.11346908e-01 -1.30878925e+00
1.17560279e+00 1.46472469e-01 5.60613871e-01 -1.87515169e-01
-1.13854986e-02 4.70317364e-01 -1.34040797e-02 -6.11637890e-01
3.41638923e-01 -1.84046611e-01 5.69013596e-01 1.53633207e-01
-2.79595375e-01 -3.67854625e-01 3.03646803e-01 -4.76264119e-01
3.63642097e-01 1.79720461e-01 6.98656261e-01 -4.50528502e-01
1.03596723e+00 3.14999014e-01 -3.38258520e-02 2.22316578e-01
2.31146604e-01 3.54103148e-01 3.93479347e-01 -4.25223678e-01
-7.95659482e-01 -5.51877558e-01 -7.25885391e-01 4.45976257e-01
1.34794340e-01 1.21459261e-01 -5.01689970e-01 -2.35336587e-01
2.40278751e-01 5.19630373e-01 -4.86094803e-01 1.12063780e-01
-1.37684622e-03 -9.87547576e-01 -9.23522115e-02 1.81485713e-01
3.90276134e-01 -9.64368224e-01 -5.34397185e-01 7.65379295e-02
2.66488701e-01 -9.58415151e-01 4.05749440e-01 1.99718103e-01
-1.23102748e+00 -1.38520098e+00 -5.06483912e-01 -1.96378186e-01
7.28524029e-01 3.40713829e-01 4.52068657e-01 -1.00651778e-01
-4.72033590e-01 9.21407193e-02 -6.08095586e-01 -3.52838576e-01
-6.69537723e-01 7.33875558e-02 -2.68202573e-01 3.31011951e-01
1.57660767e-01 -3.62909079e-01 -3.36457223e-01 5.13444424e-01
-1.00619471e+00 -2.54400343e-01 6.71579778e-01 6.77031100e-01
4.96388197e-01 2.10044667e-01 3.13421935e-01 -9.09993231e-01
4.21258420e-01 -2.04044148e-01 -1.09904456e+00 5.37534952e-01
-9.40118730e-01 -1.06773533e-01 7.99544215e-01 -2.41337672e-01
-7.63774991e-01 4.00046974e-01 1.13616161e-01 3.52582112e-02
-6.15097463e-01 7.92854726e-01 -5.98775521e-02 -7.06999540e-01
8.09385717e-01 3.02940190e-01 3.34271044e-01 -5.83481491e-01
2.56739147e-02 7.32922792e-01 2.17745110e-01 -3.36176664e-01
9.23351645e-01 4.31913018e-01 5.38963020e-01 -1.43720019e+00
-4.55105066e-01 -7.73180485e-01 -8.11058819e-01 -4.30053025e-01
4.81295913e-01 -2.60466486e-01 -8.96663308e-01 3.93224061e-01
-8.86488736e-01 -2.32437570e-02 -1.87608793e-01 5.82404315e-01
-3.84338200e-01 5.13973236e-01 -3.69416736e-02 -1.26721263e+00
-7.75197074e-02 -1.22576559e+00 6.77503884e-01 4.14730161e-01
1.43675864e-01 -1.20381021e+00 -2.43334487e-01 3.40332508e-01
3.38409573e-01 3.84894520e-01 1.20254946e+00 -6.86407804e-01
-4.92195904e-01 -3.13165754e-01 -4.59714562e-01 5.69108963e-01
4.06970441e-01 3.84644926e-01 -9.06725705e-01 -9.81063843e-02
-2.79491037e-01 2.33167149e-02 7.77986169e-01 3.04685473e-01
1.10496700e+00 -1.99018121e-02 -1.25777557e-01 6.13296330e-01
1.94455576e+00 3.82959664e-01 6.81287944e-01 4.71784800e-01
3.45053107e-01 8.67531717e-01 5.96007586e-01 3.80198359e-01
-3.19748491e-01 6.92953467e-01 6.22588992e-01 -2.10825309e-01
3.87792170e-01 2.52337486e-01 -1.24715865e-02 5.29104434e-02
-5.30167103e-01 5.38393036e-02 -1.04179966e+00 1.24133416e-01
-1.61997354e+00 -1.07744801e+00 -6.88957393e-01 2.42041755e+00
3.74621570e-01 -2.58305281e-01 1.82683602e-01 6.14649892e-01
6.53526366e-01 -2.93649863e-02 -3.02184582e-01 -4.54870671e-01
-2.47580439e-01 9.99504030e-02 5.10627925e-01 5.37264287e-01
-1.10212648e+00 6.23804331e-01 6.62529802e+00 5.04468381e-01
-1.60031331e+00 -5.20058811e-01 2.47480303e-01 6.06077433e-01
-2.03250930e-01 1.39808565e-01 -5.78782082e-01 1.20053940e-01
2.20867187e-01 1.79494351e-01 6.28480673e-01 6.61991537e-01
3.77687275e-01 -5.04433930e-01 -6.11365914e-01 6.56088173e-01
-1.49809107e-01 -1.26619267e+00 1.17390744e-01 1.40807509e-01
6.14989519e-01 -1.80785805e-01 -1.75068796e-01 -4.85908061e-01
-1.46180108e-01 -8.15448225e-01 3.01019609e-01 6.59892499e-01
7.08474874e-01 -5.42900562e-01 7.54747629e-01 2.30712458e-01
-8.62887025e-01 -4.00313854e-01 -2.84592062e-01 -7.71643966e-02
-3.06145340e-01 8.41860235e-01 -8.96385074e-01 1.05736685e+00
2.35132560e-01 6.55751646e-01 -4.71817106e-01 1.50325906e+00
-1.82543665e-01 5.71137369e-01 -4.35188919e-01 -4.05608177e-01
1.51064530e-01 -9.44701076e-01 5.36277294e-01 1.15697706e+00
3.31447154e-01 2.31028393e-01 5.50409255e-04 8.14828217e-01
6.16863370e-01 6.56400144e-01 -5.83346486e-01 -1.34145275e-01
2.26467233e-02 1.40413928e+00 -9.42946911e-01 8.76057670e-02
-2.53631622e-01 4.32096660e-01 -1.86941102e-01 4.82835025e-01
-1.68443650e-01 -3.17147940e-01 4.99633402e-01 2.80486166e-01
2.99259514e-01 -3.95723701e-01 -4.21313077e-01 -9.39913392e-01
6.84425533e-02 -6.28174126e-01 3.32631737e-01 -8.11517179e-01
-1.04523242e+00 4.14946288e-01 2.72261739e-01 -1.33375585e+00
-5.25316149e-02 -1.09593308e+00 -6.14025235e-01 1.01740265e+00
-1.83285928e+00 -9.94130015e-01 -5.73290586e-01 3.84641677e-01
4.06927764e-02 -1.74746424e-01 9.70320523e-01 -1.71471313e-01
-4.17179674e-01 -1.54375419e-01 7.57950068e-01 -1.84558518e-02
2.74162859e-01 -1.21021199e+00 -3.92420888e-01 5.06980240e-01
9.23789516e-02 3.37319642e-01 7.76991725e-01 -3.23722690e-01
-1.21684396e+00 -7.04268038e-01 5.20841658e-01 1.28247589e-01
7.97304928e-01 1.43491194e-01 -9.85947013e-01 -3.56836803e-02
-2.31452778e-01 -1.60847887e-01 7.59774029e-01 1.40484124e-01
-1.47725493e-01 -4.52655256e-01 -1.10307240e+00 2.73350388e-01
4.05898184e-01 -3.13959688e-01 -4.00771707e-01 4.48500514e-01
-1.85226738e-01 5.11519164e-02 -7.51833856e-01 6.52674437e-01
6.49737716e-01 -1.08734405e+00 8.16052198e-01 -4.20707226e-01
8.53940472e-02 -4.50313032e-01 -6.68091178e-02 -1.23423421e+00
-2.72872597e-01 -4.85415787e-01 5.14455497e-01 6.98837101e-01
5.98810792e-01 -9.91328776e-01 6.00288749e-01 3.71063292e-01
2.30427906e-01 -8.46389949e-01 -5.57349741e-01 -8.65165889e-01
-2.70302773e-01 -1.67812735e-01 5.39000988e-01 9.07274485e-01
-2.85248756e-02 -9.86417904e-02 2.07559705e-01 4.33134526e-01
6.04353130e-01 6.02019370e-01 4.13196802e-01 -1.95610011e+00
-8.49026144e-02 -6.29777968e-01 -6.83057845e-01 -3.58257383e-01
4.91233403e-03 -8.43018711e-01 -2.60819167e-01 -1.48011911e+00
-3.49041745e-02 -6.36531949e-01 -1.67659111e-02 5.34423232e-01
1.04342066e-01 2.09528819e-01 -7.68331811e-02 3.39653254e-01
2.47403055e-01 1.47715807e-01 1.02055013e+00 4.40493338e-02
-5.35890520e-01 3.01358193e-01 -2.87437767e-01 6.84194982e-01
9.70104277e-01 -4.40996513e-02 -3.69495928e-01 1.43418759e-01
2.13373348e-01 -9.27962139e-02 6.20907962e-01 -9.52799261e-01
-2.92645339e-02 -6.40364230e-01 3.11757356e-01 -3.14312547e-01
2.22197279e-01 -1.01850379e+00 3.22041720e-01 3.51220965e-01
-2.20434934e-01 -4.72316891e-01 9.40211304e-03 2.66013622e-01
-2.58639961e-01 -7.42883801e-01 9.95283902e-01 1.20306641e-01
-7.31893957e-01 3.31763215e-02 -2.49333799e-01 -6.55261099e-01
1.18826628e+00 -6.39110684e-01 -2.20763221e-01 -1.85753211e-01
-8.48371267e-01 -2.44262457e-01 5.72033226e-01 -6.07754067e-02
3.57157946e-01 -7.13818908e-01 -6.03880048e-01 3.66110533e-01
3.98843467e-01 -4.17827368e-01 -2.98435874e-02 8.66423428e-01
-7.40162671e-01 4.41461235e-01 -2.56712645e-01 -8.09053302e-01
-1.44057488e+00 2.24654451e-01 6.37617171e-01 1.68331131e-01
-1.64632723e-01 9.84669626e-02 -1.90300941e-01 -2.62932390e-01
-1.81762025e-01 -4.08885807e-01 -5.82032144e-01 2.02771723e-01
2.40891010e-01 4.16633219e-01 3.47170621e-01 -6.66709542e-01
-3.04325789e-01 1.00624049e+00 4.90757912e-01 9.34824795e-02
1.47312927e+00 2.74947137e-01 -4.50147033e-01 3.48519742e-01
9.33586001e-01 -1.25213945e-02 -9.03683424e-01 8.38482566e-03
1.41532928e-01 -6.34429932e-01 1.04505695e-01 -9.20104980e-01
-7.75350451e-01 7.88246751e-01 5.63771248e-01 7.15973973e-01
1.26302409e+00 -1.75359607e-01 -5.85234575e-02 6.74467087e-01
2.97729671e-01 -1.07805824e+00 -4.69756126e-01 3.82442862e-01
8.01878572e-01 -1.24352431e+00 2.56605536e-01 -9.73007798e-01
-4.28280115e-01 1.77223957e+00 9.57217067e-02 2.71330416e-01
6.44498587e-01 -1.35061353e-01 1.23333022e-01 -8.13429244e-03
-1.02891780e-01 -5.74590981e-01 6.14162207e-01 8.22991073e-01
3.61221075e-01 2.60579616e-01 -6.40266716e-01 4.04227637e-02
2.08743382e-02 -4.28522117e-02 4.76418614e-01 5.89589834e-01
-7.94076443e-01 -1.34810269e+00 -5.49272835e-01 5.93598664e-01
4.82826754e-02 1.82794809e-01 -6.00521803e-01 7.61878729e-01
3.48520763e-02 9.38755393e-01 -4.96123701e-01 -4.14943635e-01
3.43295813e-01 1.44412056e-01 3.35812956e-01 -5.24914801e-01
-3.23811144e-01 -1.58911675e-01 3.88360061e-02 -1.73117027e-01
-6.42045736e-01 -8.90167832e-01 -1.14946616e+00 1.17142498e-02
-7.13786423e-01 3.70540380e-01 1.19029295e+00 9.59750414e-01
5.83605953e-02 -1.20553799e-01 1.03996336e+00 -6.11242115e-01
-8.57481360e-01 -6.98875546e-01 -8.62742245e-01 3.22913557e-01
1.70280531e-01 -6.48588300e-01 -4.38214302e-01 -1.93301365e-02] | [9.763328552246094, -1.7420289516448975] |
789d5160-e214-4d53-996c-16382cb04c3d | daer-to-reject-seeds-with-dual-loss | 2009.07414 | null | https://arxiv.org/abs/2009.07414v3 | https://arxiv.org/pdf/2009.07414v3.pdf | Ground-truth or DAER: Selective Re-query of Secondary Information | Many vision tasks use secondary information at inference time -- a seed -- to assist a computer vision model in solving a problem. For example, an initial bounding box is needed to initialize visual object tracking. To date, all such work makes the assumption that the seed is a good one. However, in practice, from crowdsourcing to noisy automated seeds, this is often not the case. We hence propose the problem of seed rejection -- determining whether to reject a seed based on the expected performance degradation when it is provided in place of a gold-standard seed. We provide a formal definition to this problem, and focus on two meaningful subgoals: understanding causes of error and understanding the model's response to noisy seeds conditioned on the primary input. With these goals in mind, we propose a novel training method and evaluation metrics for the seed rejection problem. We then use seeded versions of the viewpoint estimation and fine-grained classification tasks to evaluate these contributions. In these experiments, we show our method can reduce the number of seeds that need to be reviewed for a target performance by over 23% compared to strong baselines. | ['Stephan J. Lemmer', 'Jason J. Corso'] | 2020-09-16 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Lemmer_Ground-Truth_or_DAER_Selective_Re-Query_of_Secondary_Information_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Lemmer_Ground-Truth_or_DAER_Selective_Re-Query_of_Secondary_Information_ICCV_2021_paper.pdf | iccv-2021-1 | ['viewpoint-estimation'] | ['computer-vision'] | [ 4.58169937e-01 1.21845156e-01 1.14561133e-01 -4.30641294e-01
-8.81191552e-01 -7.74530649e-01 6.04977548e-01 8.26456696e-02
-6.85775161e-01 6.64507747e-01 -1.12104610e-01 -1.72300220e-01
5.25047898e-01 -3.17084789e-01 -9.26008880e-01 -5.78802645e-01
3.66352946e-01 5.53548455e-01 9.89047408e-01 6.59234598e-02
2.70306021e-01 4.15498972e-01 -1.62050605e+00 2.30316222e-01
8.38535786e-01 7.73238003e-01 5.70948124e-01 6.37117863e-01
2.58757293e-01 6.61654830e-01 -9.70940351e-01 -3.89839143e-01
4.14499283e-01 -3.66202801e-01 -6.48646355e-01 4.50554729e-01
8.06273341e-01 -4.78716731e-01 1.21013828e-01 1.31593823e+00
3.75909060e-01 1.50191456e-01 7.17405319e-01 -1.50137961e+00
-2.63090402e-01 3.33035529e-01 -4.51019108e-01 3.40608746e-01
8.11899602e-02 5.91377556e-01 9.31942046e-01 -1.01843834e+00
6.86218739e-01 1.17684340e+00 6.23479307e-01 6.87550902e-01
-1.26078165e+00 -3.34502369e-01 6.22501016e-01 -1.11159883e-01
-1.12411273e+00 -6.64848447e-01 3.06650966e-01 -6.77485585e-01
6.03155792e-01 8.77173096e-02 4.85425860e-01 1.14128017e+00
-1.51075333e-01 6.65027320e-01 1.15303588e+00 -4.18814570e-01
4.43719953e-01 3.58392626e-01 8.47563297e-02 6.11525476e-01
6.65091574e-01 2.84904242e-01 -5.05744636e-01 -2.72486001e-01
6.87772155e-01 -5.94394863e-01 -2.53056943e-01 -2.73268700e-01
-1.16576195e+00 6.09990299e-01 4.82955009e-01 -5.29286899e-02
-2.38563314e-01 2.54421711e-01 3.11223418e-02 -7.77607635e-02
3.92388493e-01 5.74465394e-01 -2.84443051e-01 1.96573004e-01
-1.17725408e+00 4.41547185e-01 7.01022506e-01 1.00712931e+00
4.63759720e-01 -2.34421417e-02 -5.24640560e-01 4.17971492e-01
4.39148039e-01 6.27008319e-01 7.91301429e-02 -1.06679499e+00
1.26148030e-01 3.22944611e-01 8.59564781e-01 -7.29731739e-01
-1.52718157e-01 -2.75227189e-01 -1.43387586e-01 7.13621438e-01
9.84703660e-01 -1.10481046e-01 -1.27150512e+00 1.79918838e+00
5.81902444e-01 4.29515958e-01 -3.22387964e-01 1.19734108e+00
7.88806915e-01 1.32288009e-01 1.38842285e-01 -1.07594498e-01
1.20732951e+00 -1.07027698e+00 -4.28993762e-01 -7.17811942e-01
2.10626855e-01 -8.60870242e-01 1.12358391e+00 3.58033031e-01
-1.02164078e+00 -6.73969626e-01 -1.11215854e+00 1.21485598e-01
-9.83918011e-02 3.62244964e-01 3.13816547e-01 5.30571878e-01
-1.18659091e+00 6.64387643e-01 -7.67972827e-01 -4.36913908e-01
2.97911286e-01 2.25739896e-01 -8.56305361e-02 -1.10070352e-02
-7.03500271e-01 1.11044717e+00 4.83040251e-02 1.64421976e-01
-1.34218645e+00 -3.20060909e-01 -4.75346893e-01 -1.77250057e-01
7.63694882e-01 -6.15912437e-01 1.73383152e+00 -1.23688340e+00
-9.97763395e-01 9.15967107e-01 -4.84904647e-01 -4.22589242e-01
8.70524704e-01 -2.59588480e-01 1.38749033e-01 4.84706350e-02
5.11334777e-01 8.51498008e-01 1.12516356e+00 -1.63611102e+00
-8.41895700e-01 -1.31905988e-01 2.73675948e-01 1.87225074e-01
2.81140059e-01 3.14075112e-01 -8.52107286e-01 -6.17981911e-01
-9.02106240e-03 -1.09681165e+00 -5.51141918e-01 3.05507302e-01
-4.16763514e-01 -1.93939343e-01 4.26798791e-01 -3.66684079e-01
7.46694922e-01 -1.92175734e+00 -1.79131120e-01 7.96420947e-02
4.51239228e-01 3.77431244e-01 -2.05307543e-01 -8.29965845e-02
2.20751494e-01 1.10169508e-01 3.62841152e-02 -5.62084198e-01
-1.93863526e-01 -9.93598159e-03 -4.23878521e-01 5.68272591e-01
5.69059968e-01 8.66511106e-01 -1.13204598e+00 -5.45948386e-01
2.61061847e-01 2.79663533e-01 -2.46108770e-01 1.83677271e-01
-6.65137529e-01 3.21046293e-01 -3.90833497e-01 5.24070680e-01
6.10886633e-01 -6.22058928e-01 -1.22911446e-01 -1.16694503e-01
-1.62273534e-02 1.00366198e-01 -1.52520084e+00 1.11251235e+00
2.17869669e-01 5.38016498e-01 2.23101884e-01 -3.70723069e-01
5.61706126e-01 -1.32542048e-02 1.32805482e-01 -3.00365120e-01
2.85067290e-01 2.09184483e-01 9.83047336e-02 -1.87928185e-01
5.23392141e-01 1.52530819e-01 3.46592009e-01 3.22352976e-01
-1.06626682e-01 -1.86147526e-01 3.38066667e-01 3.12398374e-01
1.16119313e+00 3.52234274e-01 2.96440363e-01 -1.08698480e-01
2.68309504e-01 3.54481995e-01 6.78007007e-01 1.39113998e+00
-5.80003321e-01 9.39724743e-01 3.90236557e-01 -3.09274495e-01
-1.07194364e+00 -9.04254019e-01 1.54669717e-01 1.05934012e+00
4.48426366e-01 -3.49593908e-01 -1.05578804e+00 -9.28528726e-01
4.68920805e-02 5.89186251e-01 -7.16578662e-01 1.19620584e-01
-3.32751036e-01 -4.68836635e-01 4.02383983e-01 6.42395556e-01
1.97273418e-01 -1.08549225e+00 -1.09132481e+00 5.93254976e-02
-1.19027942e-01 -1.36887324e+00 -5.59876621e-01 2.62096316e-01
-6.02056503e-01 -1.34645951e+00 -6.14408553e-01 -5.27434766e-01
9.02343094e-01 7.22290516e-01 1.48204565e+00 5.58305442e-01
-1.89783186e-01 3.47315550e-01 -3.95482689e-01 -6.69573903e-01
-5.33198833e-01 -3.24952781e-01 1.40720636e-01 -2.12810054e-01
3.50291610e-01 8.30825269e-02 -5.00261545e-01 6.70245290e-01
-7.70955443e-01 2.08170377e-02 4.90665019e-01 5.68351924e-01
5.78219593e-01 -3.21906090e-01 3.23398292e-01 -7.32903242e-01
3.48068684e-01 -1.21184006e-01 -1.06415415e+00 3.53824705e-01
-3.78637791e-01 1.29072860e-01 3.74341130e-01 -6.22275829e-01
-7.88057864e-01 4.25664872e-01 2.07567722e-01 -6.09035969e-01
-2.39140168e-01 5.70611469e-02 -8.12223405e-02 -1.81704581e-01
1.11528444e+00 -7.75141120e-02 -1.75666645e-01 -2.24163935e-01
3.52189362e-01 4.09793377e-01 5.11675835e-01 -5.83953202e-01
9.23137307e-01 6.62029326e-01 -1.97952807e-01 -6.17226958e-01
-1.27384734e+00 -6.96476042e-01 -5.44523954e-01 -3.45846385e-01
7.46469140e-01 -9.96291995e-01 -4.12633806e-01 3.62950861e-01
-1.47382486e+00 -5.19107997e-01 -3.25554132e-01 1.71550840e-01
-4.70901936e-01 3.18249971e-01 -1.37747303e-02 -1.01850080e+00
8.97076204e-02 -1.44989049e+00 1.26744974e+00 2.97922939e-01
-1.20334677e-01 -4.73968834e-01 -2.84039080e-01 9.62842777e-02
2.14357182e-01 5.90935610e-02 2.05382496e-01 -8.01410794e-01
-7.59537637e-01 -1.78340264e-02 -3.45144570e-01 3.52547169e-01
-1.26251593e-01 4.13485691e-02 -1.25755286e+00 -3.00262094e-01
8.52590203e-02 -4.29190964e-01 8.75766754e-01 6.23025835e-01
7.93593347e-01 9.82016996e-02 -5.28089821e-01 2.70440251e-01
1.26261723e+00 -3.88377942e-02 3.83653551e-01 3.40901673e-01
4.39131171e-01 5.00034273e-01 8.97531450e-01 3.41712773e-01
2.72406787e-01 7.77715683e-01 5.97933888e-01 -7.94958100e-02
-2.62651354e-01 -8.56413245e-02 3.50539535e-01 -1.27221510e-01
4.04923186e-02 -3.79348248e-01 -8.06623816e-01 6.96816385e-01
-1.97741771e+00 -7.76785076e-01 -3.24365526e-01 2.52950907e+00
8.81966293e-01 5.31781375e-01 2.26683035e-01 -2.74391383e-01
8.79057109e-01 -9.00427923e-02 -7.46989727e-01 3.58369976e-01
-1.24722149e-03 -2.03995869e-01 5.90899348e-01 7.92664349e-01
-1.31792521e+00 1.28929186e+00 6.47426844e+00 5.50148845e-01
-9.05524909e-01 1.32183567e-01 6.53012633e-01 -1.45587102e-01
5.90434261e-02 1.90687060e-01 -1.29263711e+00 4.28107291e-01
6.66465402e-01 9.10163596e-02 4.12233233e-01 9.57194090e-01
3.40885609e-01 -6.52209461e-01 -1.41691935e+00 8.20906341e-01
2.58429021e-01 -1.07873118e+00 -2.75196016e-01 -1.43282935e-01
6.70175076e-01 3.30233604e-01 -1.75000653e-01 2.93438546e-02
8.29909265e-01 -9.91357267e-01 1.18729126e+00 3.08909118e-01
6.63939118e-01 -1.30354002e-01 4.44494277e-01 4.28127497e-01
-1.01546836e+00 2.60438472e-01 -4.03839916e-01 6.96439818e-02
1.34488165e-01 6.68322682e-01 -1.13761914e+00 9.70952585e-02
6.38866603e-01 1.68039873e-01 -9.46688533e-01 1.54383421e+00
-6.82662785e-01 5.85685015e-01 -5.57120144e-01 -1.37401775e-01
1.35974437e-01 3.52438450e-01 7.17475593e-01 1.00449646e+00
-6.36870191e-02 -4.80389548e-03 5.17679095e-01 8.92157137e-01
1.77491933e-01 -3.51711690e-01 -4.45490211e-01 2.73150623e-01
6.56659484e-01 1.22530258e+00 -1.20059168e+00 -3.93555939e-01
-3.32409084e-01 9.99414921e-01 3.61134768e-01 4.58066195e-01
-9.94532049e-01 -4.07343283e-02 5.18804967e-01 2.09571168e-01
4.91176635e-01 -1.55717209e-01 -2.23745361e-01 -1.07405901e+00
3.13621014e-01 -1.03910983e+00 -2.21070852e-02 -1.04967141e+00
-1.22342873e+00 6.44207895e-01 -4.73030731e-02 -1.13167620e+00
-2.82348663e-01 -6.99158430e-01 -4.01623100e-01 1.06731868e+00
-1.44296622e+00 -1.00125170e+00 -5.89288056e-01 3.41021270e-01
6.15153611e-01 2.10549816e-01 4.13150489e-01 -1.53281793e-01
-9.75630656e-02 4.22168911e-01 -2.38577336e-01 1.70873821e-01
8.78398836e-01 -1.40507972e+00 9.19491410e-01 1.41755378e+00
3.75356138e-01 5.57852030e-01 1.14916229e+00 -1.03710067e+00
-9.24853981e-01 -1.07964516e+00 7.21510589e-01 -1.01092100e+00
4.42205459e-01 -3.37507635e-01 -7.18236506e-01 5.71408153e-01
-1.35886416e-01 4.26904976e-01 5.43452390e-02 -7.47079477e-02
-3.92675817e-01 2.48148009e-01 -1.12201178e+00 6.67024612e-01
1.06465423e+00 -1.94762677e-01 -5.11325419e-01 4.68331575e-01
7.64175177e-01 -6.17106497e-01 -9.07302126e-02 2.62409329e-01
3.46925467e-01 -9.22705114e-01 8.51602614e-01 -7.26330340e-01
2.52877086e-01 -8.47917438e-01 -7.56839663e-02 -1.25318444e+00
-3.29213768e-01 -4.83075142e-01 3.66297811e-02 8.90264511e-01
3.91511440e-01 -7.07263872e-02 8.93591642e-01 8.12945664e-01
2.63765864e-02 -4.65128481e-01 -8.26717615e-01 -7.75688410e-01
-3.76719505e-01 -5.04806399e-01 1.45540282e-01 4.33003128e-01
-7.74270833e-01 3.75344187e-01 -2.18214482e-01 5.07395327e-01
8.14834237e-01 -2.12985307e-01 1.11795902e+00 -1.29280365e+00
-3.16499978e-01 -2.42082834e-01 -2.80880958e-01 -1.09941256e+00
-9.23936591e-02 -3.77316564e-01 7.68705487e-01 -1.71971929e+00
2.67409027e-01 -5.58899403e-01 -6.11467324e-02 4.73359495e-01
-6.52463675e-01 3.04743379e-01 5.10554314e-01 2.55755126e-01
-9.45697963e-01 8.06754529e-02 1.13491893e+00 -3.59196402e-02
-1.07660480e-01 2.27773547e-01 -8.07463288e-01 9.80686367e-01
4.80145454e-01 -5.71494877e-01 -4.21487302e-01 -5.05801320e-01
2.32773334e-01 -3.03829879e-01 8.25302064e-01 -1.06210363e+00
3.36719841e-01 -3.78214598e-01 4.67879772e-01 -4.15383846e-01
3.33354890e-01 -8.38916779e-01 -1.52009428e-01 1.83537886e-01
-2.00530738e-01 2.09633000e-02 2.07268581e-01 7.59422958e-01
2.43550345e-01 -6.79087818e-01 8.29344392e-01 -3.13780218e-01
-8.15605223e-01 1.55629158e-01 -2.11361378e-01 3.77564937e-01
1.02886188e+00 -3.51967424e-01 -4.58029360e-01 -2.87758440e-01
-8.52501273e-01 2.77803123e-01 7.80636609e-01 2.62311041e-01
5.34092605e-01 -1.01647449e+00 -8.47733438e-01 -3.10786208e-03
2.18646958e-01 1.89869910e-01 -3.69441003e-01 7.07816005e-01
-1.80787459e-01 5.90145960e-02 2.16717288e-01 -8.18915606e-01
-1.49620903e+00 4.15741146e-01 4.30985093e-01 -5.90315014e-02
-4.47901815e-01 1.21834290e+00 2.24906251e-01 8.45376551e-02
6.58505082e-01 -5.52639484e-01 5.59771527e-03 -1.13573134e-01
7.31974423e-01 4.21001613e-02 1.99293464e-01 -5.12622356e-01
-4.20653194e-01 1.78855136e-01 -6.58096224e-02 -4.62183118e-01
1.00009072e+00 4.02464382e-02 2.51858801e-01 4.30050343e-01
3.53310078e-01 1.28624350e-01 -1.79958081e+00 -3.71067017e-01
5.65755181e-02 -7.60320485e-01 -7.59275258e-02 -1.09046769e+00
-6.84913814e-01 5.02171397e-01 5.91509283e-01 2.21189007e-01
6.87783718e-01 2.16182873e-01 8.23381916e-02 4.60291088e-01
5.53554356e-01 -1.06958723e+00 6.18357733e-02 4.66713339e-01
8.99357557e-01 -1.62853897e+00 5.32818586e-02 -5.02617359e-01
-8.62070382e-01 7.13048518e-01 9.78050888e-01 -1.99437305e-01
2.75092393e-01 4.62205082e-01 2.45289728e-01 -1.37185678e-01
-8.05817187e-01 -4.66525108e-01 3.99671823e-01 7.32675135e-01
6.15179166e-02 -1.36315644e-01 1.09557621e-01 3.92328590e-01
-1.65550259e-03 1.44748971e-01 6.10470295e-01 8.92795086e-01
-8.21113825e-01 -8.38568628e-01 -6.53641999e-01 6.23848677e-01
-2.77967840e-01 -2.03952700e-01 -7.69425750e-01 5.96732795e-01
2.51595974e-01 1.25995219e+00 -1.44187525e-01 -2.39313781e-01
3.28053713e-01 -2.09869191e-01 6.65998757e-01 -9.72738326e-01
-7.18643844e-01 1.83072045e-01 1.66663110e-01 -7.02188075e-01
-6.78273022e-01 -7.69761384e-01 -9.73055005e-01 1.62720233e-01
-8.64098489e-01 -1.57684058e-01 4.51381177e-01 1.13194013e+00
8.18831399e-02 3.19282711e-01 1.42860472e-01 -1.11435401e+00
-9.04100835e-01 -9.22500551e-01 -9.22374874e-02 3.56348068e-01
3.81513834e-01 -9.53533053e-01 -5.19038737e-01 3.44560713e-01] | [9.140669822692871, 1.0341920852661133] |
846090fd-832b-4ea6-83bf-bc32e413601b | face2exp-combating-data-biases-for-facial | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zeng_Face2Exp_Combating_Data_Biases_for_Facial_Expression_Recognition_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zeng_Face2Exp_Combating_Data_Biases_for_Facial_Expression_Recognition_CVPR_2022_paper.pdf | Face2Exp: Combating Data Biases for Facial Expression Recognition | Facial expression recognition (FER) is challenging due to the class imbalance caused by data collection. Existing studies tackle the data bias problem using only labeled facial expression dataset. Orthogonal to existing FER methods, we propose to utilize large unlabeled face recognition (FR) datasets to enhance FER. However, this raises another data bias problem---the distribution mismatch between FR and FER data. To combat the mismatch, we propose the Meta-Face2Exp framework, which consists of a base network and an adaptation network. The base network learns prior expression knowledge on class-balanced FER data while the adaptation network is trained to fit the pseudo labels of FR data generated by the base model. To combat the mismatch between FR and FER data, Meta-Face2Exp uses a circuit feedback mechanism, which improves the base network with the feedback from the adaptation network. Experiments show that our Meta-Face2Exp achieves comparable accuracy to state-of-the-art FER methods with 10% of the labeled FER data utilized by the baselines. We also demonstrate that the circuit feedback mechanism successfully eliminates data bias. | ['Bo Tang', 'Fei Wang', 'YuTing Liu', 'Xiao Yan', 'Zhiyuan Lin', 'Dan Zeng'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['facial-expression-recognition'] | ['computer-vision'] | [ 2.24584490e-01 2.41426006e-01 -3.60187918e-01 -1.15265143e+00
-7.00449765e-01 6.81704059e-02 -7.31981397e-02 -1.00834644e+00
-2.07029969e-01 8.17039728e-01 -9.08903033e-02 2.66281337e-01
4.01876748e-01 -5.14547765e-01 -6.92278385e-01 -6.80250823e-01
2.45020494e-01 8.50752145e-02 -4.58740234e-01 -4.52504396e-01
-2.15528727e-01 5.18263400e-01 -1.85444891e+00 7.28703082e-01
7.04156220e-01 1.67920256e+00 -4.66543972e-01 1.52810320e-01
-1.42234057e-01 1.42133105e+00 -7.32868433e-01 -5.57647586e-01
3.82202208e-01 -6.39215708e-01 -5.83623767e-01 1.51889980e-01
7.81275988e-01 -5.78966081e-01 -3.79476696e-01 1.03266513e+00
6.97723866e-01 -3.93511467e-02 5.85211813e-01 -1.90820730e+00
-7.11153507e-01 1.70942634e-01 -9.72191572e-01 -5.75903989e-02
1.72094434e-01 -1.21650130e-01 4.52665627e-01 -1.37057579e+00
6.06621325e-01 1.52535605e+00 8.05823207e-01 1.29244804e+00
-1.12916923e+00 -1.41449428e+00 2.48752654e-01 3.19227651e-02
-1.60958159e+00 -1.08606887e+00 8.42584431e-01 -2.41918504e-01
5.79782307e-01 2.26690006e-02 4.45875078e-01 1.21820092e+00
-2.18677193e-01 8.18685710e-01 1.17677867e+00 -5.52844107e-01
3.85735510e-03 2.37472042e-01 2.07144126e-01 7.86867321e-01
-3.97815168e-01 1.28852934e-01 -6.95257485e-01 -2.22562909e-01
5.01896381e-01 -7.93722868e-02 -3.93053323e-01 7.60568902e-02
-1.21523663e-01 6.94179893e-01 4.01089162e-01 4.73069698e-02
-2.53437430e-01 2.73963809e-02 3.87102872e-01 5.82022607e-01
8.81287992e-01 2.09627882e-01 -5.43572783e-01 1.20159969e-01
-1.03698218e+00 8.37634951e-02 6.05449855e-01 9.81162965e-01
1.13448453e+00 4.01445240e-01 -2.93399870e-01 1.28975654e+00
3.96554798e-01 6.70146585e-01 3.10510069e-01 -1.11772931e+00
1.99413791e-01 7.26093411e-01 -2.13300318e-01 -9.03985143e-01
-4.50192057e-02 -1.72155164e-02 -6.64932013e-01 1.80809751e-01
2.47033656e-01 -4.11642790e-01 -1.02330947e+00 2.35143948e+00
2.87534803e-01 2.25174338e-01 -8.11752155e-02 9.09869015e-01
1.02351534e+00 2.36844406e-01 2.60590881e-01 -3.49745899e-01
9.49757159e-01 -1.04651380e+00 -9.60318327e-01 -1.00930184e-01
6.00504458e-01 -4.98051673e-01 9.70374584e-01 3.11400026e-01
-8.63372087e-01 -5.05580366e-01 -8.65350008e-01 -1.44639956e-02
-7.53040984e-02 5.90181351e-01 5.43731928e-01 7.47182369e-01
-1.14856768e+00 1.76766783e-01 -2.93338925e-01 -3.32097826e-03
1.20314288e+00 5.79030812e-01 -6.49771392e-01 -4.09786463e-01
-1.12653446e+00 4.96020466e-01 -1.45941526e-01 4.19147283e-01
-8.04591596e-01 -9.33929026e-01 -8.00862193e-01 -1.95080146e-01
2.37732559e-01 -2.13379815e-01 1.43544221e+00 -2.19142699e+00
-1.60002518e+00 1.10403693e+00 -3.71015072e-01 1.47184744e-01
4.10101593e-01 -2.51938552e-01 -7.08575547e-01 1.11621358e-01
-1.86204895e-01 9.02502477e-01 1.03166568e+00 -1.35698640e+00
-2.75940955e-01 -6.02915645e-01 -2.42670968e-01 -1.38069242e-01
-3.91902328e-01 1.95362285e-01 -2.98541695e-01 -4.38271374e-01
-4.38321531e-01 -7.86792874e-01 1.53678760e-01 3.42187941e-01
-1.35907501e-01 -1.42080829e-01 1.22662401e+00 -4.58676249e-01
1.19721889e+00 -2.42092800e+00 -4.08030391e-01 2.24612802e-01
2.44135335e-01 3.67505491e-01 -4.48537141e-01 -1.98432714e-01
-5.97521842e-01 -5.03284223e-02 7.86999539e-02 -4.67943460e-01
-2.29104310e-01 2.46943653e-01 -4.25661117e-01 4.96123493e-01
6.74686790e-01 6.63699508e-01 -7.08292067e-01 -4.70453948e-01
-3.83951038e-01 6.10333860e-01 -7.67534792e-01 6.67670310e-01
-1.37601838e-01 4.17195857e-01 -2.63311803e-01 9.72102046e-01
1.17300916e+00 -1.84256345e-01 2.52681732e-01 -6.06439412e-01
4.07253444e-01 -3.60902488e-01 -7.87100434e-01 1.45442331e+00
-2.35727847e-01 5.63148201e-01 3.23887646e-01 -1.06575108e+00
1.34578276e+00 3.47122401e-01 6.14228010e-01 -8.13903093e-01
6.00530505e-01 3.15341324e-01 -1.78326026e-01 -5.81681073e-01
3.21674466e-01 -4.42815036e-01 3.13655913e-01 3.31578463e-01
5.83619416e-01 4.42639232e-01 2.82609044e-03 -1.03684574e-01
9.17952538e-01 4.01318014e-01 -1.91650599e-01 -5.02858907e-02
5.37638307e-01 -3.30643535e-01 1.00180280e+00 2.03115955e-01
-3.95563096e-01 4.43471372e-01 5.95553875e-01 -5.41177273e-01
-6.16990507e-01 -6.27034247e-01 -1.33159801e-01 1.42653060e+00
-1.93263337e-01 -3.25725824e-01 -1.00063145e+00 -1.18142307e+00
4.51977924e-02 2.21130133e-01 -1.06360567e+00 -4.57113355e-01
-3.10981601e-01 -7.82835305e-01 8.62876236e-01 5.85052550e-01
7.13581622e-01 -8.13462377e-01 -5.24245650e-02 -2.00044125e-01
-4.25159901e-01 -9.81906176e-01 -4.35489893e-01 -2.04693973e-02
-4.38259721e-01 -1.16608930e+00 -4.91583228e-01 -5.20418167e-01
1.08181572e+00 6.18702509e-02 1.26313150e+00 2.84241825e-01
-3.05739492e-01 3.17914665e-01 -2.95056790e-01 -6.93291426e-01
-3.04034144e-01 -3.03425431e-01 -5.82109503e-02 5.95188737e-01
1.07657301e+00 -2.70283997e-01 -4.63495880e-01 5.73427200e-01
-7.25890994e-01 -1.90909430e-01 2.23256975e-01 1.08501613e+00
6.51963472e-01 -5.43239295e-01 8.73331726e-01 -1.05457604e+00
3.20444375e-01 -6.65239394e-01 -2.54072636e-01 3.34732085e-01
-5.28798103e-01 -2.52352595e-01 4.29753959e-01 -6.22649431e-01
-1.42464280e+00 6.95129931e-02 -2.54489452e-01 -9.72337186e-01
2.17475314e-02 8.38542357e-02 -5.15361130e-01 -3.26177895e-01
8.17204118e-01 -5.35633147e-01 4.33048069e-01 -2.68298835e-01
1.17155567e-01 1.19144714e+00 4.54382390e-01 -7.31154144e-01
1.22338295e-01 4.35183197e-01 -1.37273163e-01 -5.30811548e-01
-1.26931345e+00 -6.66423514e-02 -3.45286489e-01 -6.26981974e-01
4.08956409e-01 -1.36081350e+00 -5.10646880e-01 6.80310786e-01
-1.00109971e+00 -3.90781373e-01 -4.02964264e-01 1.83372706e-01
-4.91924673e-01 -3.73725951e-01 -6.57822073e-01 -1.02694821e+00
-3.42346251e-01 -1.01519775e+00 1.35548007e+00 3.01165402e-01
-1.47800758e-01 -4.78792369e-01 -1.83598790e-02 3.66977036e-01
4.44995701e-01 4.39778179e-01 4.22474951e-01 -4.14940238e-01
8.23355317e-02 -5.07975928e-03 -4.46823567e-01 7.61149883e-01
4.11634862e-01 2.20232725e-01 -1.73620069e+00 -2.48119950e-01
-5.36827650e-03 -1.12163365e+00 7.55955815e-01 -1.74434539e-02
1.40054595e+00 -1.23587422e-01 -9.76421162e-02 7.36977875e-01
1.21776700e+00 1.30699545e-01 8.98535192e-01 -1.41070798e-01
5.78872740e-01 7.80128717e-01 7.70912409e-01 4.95127767e-01
2.55010575e-01 5.56324482e-01 2.25564316e-01 -5.90694547e-01
-3.11364979e-01 -2.43415937e-01 5.07138431e-01 2.25198627e-01
5.63591532e-03 -1.02483258e-01 -4.99631643e-01 8.36355984e-02
-1.59870446e+00 -9.14384842e-01 3.05841923e-01 1.80864251e+00
9.78403330e-01 -6.56162381e-01 1.45910844e-01 5.33751622e-02
6.55231118e-01 -1.24969654e-01 -6.85014665e-01 -3.41554821e-01
-2.85885066e-01 3.70818466e-01 4.78507839e-02 1.78150997e-01
-9.34345365e-01 7.95700729e-01 6.55686235e+00 9.04291332e-01
-1.72909498e+00 1.19562708e-01 1.26037812e+00 -3.19865793e-01
-8.42647702e-02 -5.01943171e-01 -7.94004977e-01 3.53991300e-01
9.04789627e-01 2.69297421e-01 3.30510050e-01 1.11881101e+00
8.20924267e-02 7.15377182e-02 -1.14587581e+00 1.32470620e+00
4.73309338e-01 -7.15205610e-01 1.41233848e-02 -9.31804702e-02
6.48151755e-01 -1.40588850e-01 8.71458128e-02 5.84974468e-01
1.73782036e-01 -1.28212488e+00 6.27376854e-01 7.75174618e-01
1.54454362e+00 -7.62797952e-01 8.73944342e-01 -2.04689518e-01
-8.22959960e-01 -1.49747625e-01 -3.16892087e-01 1.18998781e-01
-1.53465033e-01 6.31574631e-01 -4.85934556e-01 4.61606562e-01
9.24453795e-01 7.84004450e-01 -4.33522850e-01 2.12856367e-01
-2.36033529e-01 6.16200864e-01 -1.39075980e-01 5.46052873e-01
-3.94651294e-01 -1.45775691e-01 -1.00012019e-01 1.01778805e+00
1.98122159e-01 2.59337462e-02 1.01006269e-01 8.31114352e-01
-7.00799286e-01 1.96791172e-01 -5.22675216e-01 -1.70512617e-01
3.68221313e-01 1.42915809e+00 1.15997821e-01 -2.40213662e-01
-5.45768201e-01 7.66602755e-01 7.35353053e-01 5.36806285e-01
-7.47465968e-01 -6.38222620e-02 8.40734899e-01 1.25236318e-01
-2.05059629e-02 5.89536846e-01 2.28430629e-01 -1.15059686e+00
-9.11640748e-02 -1.21684527e+00 4.54364538e-01 -9.42689836e-01
-1.65604258e+00 9.55243051e-01 -2.83881783e-01 -1.06004798e+00
-2.40080655e-01 -6.37746334e-01 -1.42957315e-01 8.11076283e-01
-1.79202104e+00 -1.34538162e+00 -6.61699951e-01 8.93823028e-01
1.17992692e-01 -4.11204696e-01 9.34208453e-01 7.96636343e-01
-9.07784343e-01 1.24014926e+00 -3.66635233e-01 3.72635216e-01
1.29645145e+00 -7.68899202e-01 -5.68762362e-01 4.32143092e-01
-2.19069257e-01 5.40897667e-01 2.42716759e-01 -4.03964400e-01
-1.21077538e+00 -1.35921359e+00 5.39798617e-01 -7.58648813e-02
6.36587515e-02 -3.92206877e-01 -9.10616815e-01 6.87909722e-01
-8.90741795e-02 7.92033136e-01 1.31722677e+00 5.50629161e-02
-8.70903075e-01 -5.92475057e-01 -1.61082041e+00 1.97101280e-01
1.11119175e+00 -5.45229316e-01 -9.74117965e-02 -9.38753933e-02
1.62720695e-01 -4.65910047e-01 -9.42641079e-01 7.99608588e-01
8.50566924e-01 -8.88763011e-01 4.14366633e-01 -9.03443754e-01
5.71846366e-01 -1.89010695e-01 -4.70870227e-01 -1.32585835e+00
-7.79571682e-02 -4.06806082e-01 -1.33617342e-01 1.49109066e+00
2.90439427e-01 -4.31464583e-01 8.50667119e-01 8.39376450e-01
1.11939043e-01 -9.21118021e-01 -7.76364803e-01 -4.73525882e-01
-1.53561801e-01 -1.41964555e-01 9.03646350e-01 1.09516776e+00
-3.26706290e-01 3.37994754e-01 -5.77724636e-01 -3.49767834e-01
2.55790025e-01 -6.63963705e-02 1.02592850e+00 -1.03537869e+00
-3.66956629e-02 -3.38618159e-02 -1.55945495e-01 -8.59009683e-01
7.21648276e-01 -8.09879005e-01 2.27363408e-01 -5.17030776e-01
2.66374320e-01 -4.88854825e-01 -3.58727783e-01 8.69937360e-01
-2.77388990e-01 7.24980712e-01 1.50655672e-01 9.79424566e-02
-6.06300116e-01 9.94844794e-01 1.17376757e+00 6.06461801e-02
3.59073952e-02 -4.69029218e-01 -1.05403924e+00 7.58516908e-01
5.54571390e-01 -4.20584947e-01 -5.43493569e-01 -3.76962155e-01
-4.47540358e-02 -2.49612108e-01 1.17166556e-01 -6.00883603e-01
-1.83975741e-01 -8.25793967e-02 7.87886620e-01 5.18299229e-02
2.45432019e-01 -8.68703425e-01 2.75897402e-02 -2.17026189e-01
-2.42456064e-01 -1.40705779e-01 3.80200654e-01 2.38759398e-01
-5.05132675e-01 1.10441729e-01 1.03497171e+00 8.62512290e-02
-5.00307739e-01 7.03694284e-01 -1.56751320e-01 2.26543814e-01
8.63255262e-01 -9.71857980e-02 -5.23169160e-01 -4.38520133e-01
-4.20734555e-01 2.35858783e-01 2.94730157e-01 4.74676400e-01
5.01545310e-01 -1.62042809e+00 -5.77245951e-01 5.65522313e-01
3.01909208e-01 -1.28343239e-01 2.62399644e-01 8.60759974e-01
8.23260546e-02 -2.52228230e-01 -3.57520670e-01 -4.58903819e-01
-1.50379133e+00 2.47511134e-01 8.26498449e-01 -9.66166612e-03
2.18723670e-01 1.05893838e+00 3.85806710e-01 -6.50488198e-01
1.48971662e-01 3.32253277e-01 -1.80333823e-01 1.85098365e-01
9.57861662e-01 1.96412325e-01 1.19460590e-01 -8.91122401e-01
-2.40560398e-01 4.24309820e-01 -1.42770067e-01 2.21057475e-01
1.34171200e+00 -4.46298812e-03 -2.88800031e-01 2.98981547e-01
1.33425236e+00 -3.21186744e-02 -1.45852804e+00 -2.89775789e-01
-3.35890681e-01 -6.82373464e-01 -2.15089675e-02 -8.93330336e-01
-1.59784627e+00 7.45837271e-01 8.51576388e-01 -6.59613073e-01
1.77914417e+00 -2.92307973e-01 4.05421615e-01 7.06360117e-02
3.72144461e-01 -1.17230606e+00 2.88045019e-01 2.48639792e-01
9.40409124e-01 -1.29242587e+00 -2.66479254e-01 -7.06482768e-01
-6.75357699e-01 1.15447485e+00 1.25238848e+00 -4.18294929e-02
8.53028059e-01 5.24274290e-01 6.02531910e-01 -2.95891374e-01
-7.83526659e-01 9.19131190e-02 4.38642427e-02 6.49036646e-01
5.55211663e-01 -2.94622093e-01 -1.40331592e-02 9.37542975e-01
-2.37328410e-02 8.68745089e-01 1.30548626e-02 1.03380036e+00
8.90203416e-02 -1.22323298e+00 -2.33902201e-01 5.86470306e-01
-7.24858999e-01 2.20889166e-01 -6.17469907e-01 6.86720848e-01
4.00448978e-01 9.67190325e-01 1.78112134e-01 -6.57410324e-01
5.37949324e-01 3.03491145e-01 7.22222984e-01 -3.56493741e-01
-5.12785256e-01 5.45940641e-03 2.02985540e-01 -9.04135466e-01
-9.55574811e-01 -3.26008469e-01 -1.07165956e+00 -2.42365018e-01
-5.19891143e-01 1.36867911e-01 2.58614331e-01 8.51897240e-01
7.04405189e-01 3.62457633e-01 1.08534026e+00 -4.76403207e-01
-4.84854639e-01 -1.14262438e+00 -7.66819715e-01 6.73171043e-01
3.69517684e-01 -9.23088193e-01 -3.92505080e-01 2.44140819e-01] | [13.59283447265625, 1.6457409858703613] |
f2eb9cb2-c241-410d-a91f-be1a82fba5dd | transformer-based-multi-grained-features-for | 2211.12280 | null | https://arxiv.org/abs/2211.12280v1 | https://arxiv.org/pdf/2211.12280v1.pdf | Transformer Based Multi-Grained Features for Unsupervised Person Re-Identification | Multi-grained features extracted from convolutional neural networks (CNNs) have demonstrated their strong discrimination ability in supervised person re-identification (Re-ID) tasks. Inspired by them, this work investigates the way of extracting multi-grained features from a pure transformer network to address the unsupervised Re-ID problem that is label-free but much more challenging. To this end, we build a dual-branch network architecture based upon a modified Vision Transformer (ViT). The local tokens output in each branch are reshaped and then uniformly partitioned into multiple stripes to generate part-level features, while the global tokens of two branches are averaged to produce a global feature. Further, based upon offline-online associated camera-aware proxies (O2CAP) that is a top-performing unsupervised Re-ID method, we define offline and online contrastive learning losses with respect to both global and part-level features to conduct unsupervised learning. Extensive experiments on three person Re-ID datasets show that the proposed method outperforms state-of-the-art unsupervised methods by a considerable margin, greatly mitigating the gap to supervised counterparts. Code will be available soon at https://github.com/RikoLi/WACV23-workshop-TMGF. | ['Xiaojin Gong', 'Menglin Wang', 'Jiachen Li'] | 2022-11-22 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 5.84234782e-02 -5.54763973e-02 -3.26598100e-02 -4.13698971e-01
-5.33967435e-01 -3.74109358e-01 7.52911508e-01 -2.22228184e-01
-6.42538249e-01 6.27639592e-01 2.78439462e-01 2.71935254e-01
-1.23493150e-01 -7.11999238e-01 -4.92801130e-01 -6.90854490e-01
1.64330915e-01 4.12219763e-01 2.68891305e-02 1.32770285e-01
-2.09089026e-01 4.89296943e-01 -1.69660938e+00 2.04266794e-02
7.22881734e-01 1.03787065e+00 -2.82597870e-01 4.24196362e-01
1.90967962e-01 6.91810310e-01 -3.65902722e-01 -9.06143069e-01
5.12030959e-01 -3.18702042e-01 -8.75574827e-01 1.47804379e-01
8.04844975e-01 -4.66541976e-01 -6.41834676e-01 1.14106798e+00
6.82523370e-01 3.02647501e-01 6.04123533e-01 -1.33669949e+00
-7.36818969e-01 4.55859214e-01 -5.10765612e-01 1.93081602e-01
2.76130647e-03 1.50626525e-01 9.04873669e-01 -9.39776003e-01
2.47234628e-01 1.17593062e+00 9.74919856e-01 7.85400987e-01
-1.23366797e+00 -8.37918818e-01 1.82756037e-01 2.73816228e-01
-1.57959700e+00 -5.24526894e-01 8.28611493e-01 -4.79786485e-01
7.03112721e-01 9.27576944e-02 5.64166129e-01 1.24643409e+00
-3.41713756e-01 8.83119762e-01 1.23188543e+00 -3.76606256e-01
-1.53463870e-01 2.08763555e-01 3.42900485e-01 9.06599343e-01
2.21225053e-01 2.48725697e-01 -5.88491201e-01 3.90577763e-02
7.13058650e-01 3.88118654e-01 -1.02940795e-03 -2.65315652e-01
-1.25062728e+00 5.93704402e-01 6.55813098e-01 1.61101088e-01
-1.90449119e-01 7.09981918e-02 4.04236674e-01 2.85444111e-01
4.62312192e-01 6.55294955e-02 -1.95583001e-01 2.41444353e-02
-9.87865925e-01 3.33996654e-01 4.83877659e-01 7.60567129e-01
9.81193960e-01 -1.78762823e-01 -5.61982691e-01 1.11000550e+00
1.59658957e-02 4.35847551e-01 6.46596491e-01 -6.57445192e-01
4.28421319e-01 6.43302023e-01 8.59773718e-03 -8.41880143e-01
-4.41582859e-01 -8.28755498e-01 -1.30860138e+00 7.26422369e-02
6.88518822e-01 -1.24634445e-01 -9.49260652e-01 1.69772780e+00
1.82604238e-01 3.06331426e-01 -3.32815975e-01 9.83056247e-01
7.55512118e-01 1.59754783e-01 1.39803037e-01 3.54613245e-01
1.43374681e+00 -1.31031835e+00 -1.07469387e-01 -4.42825332e-02
3.91551346e-01 -1.24007702e-01 7.78930545e-01 2.89651722e-01
-8.66354406e-01 -9.78671432e-01 -9.52532351e-01 -8.51004645e-02
-5.38800240e-01 4.91731882e-01 3.54868799e-01 7.78309464e-01
-1.26320899e+00 6.28638268e-01 -4.88680542e-01 -4.17445987e-01
6.55899107e-01 4.22451138e-01 -5.30878782e-01 -9.91978347e-02
-1.08905518e+00 4.14679021e-01 2.31067672e-01 2.86685586e-01
-1.01431334e+00 -6.17725194e-01 -5.76538086e-01 3.03632654e-02
1.51923612e-01 -7.91064441e-01 9.48541820e-01 -1.14616787e+00
-1.44893765e+00 1.20737660e+00 -8.03662166e-02 -4.67168719e-01
7.24888206e-01 -1.80294752e-01 -4.80483532e-01 2.11075187e-01
3.36427927e-01 7.26617754e-01 1.07974184e+00 -1.13392043e+00
-8.05246770e-01 -5.30721188e-01 1.75245151e-01 2.98990086e-02
-7.54087567e-01 -8.22700188e-03 -6.18939996e-01 -9.21667814e-01
-3.55871290e-01 -1.09885466e+00 1.05071785e-02 -5.31638637e-02
-7.90950418e-01 -3.63958627e-01 5.09480059e-01 -7.83169985e-01
9.80151355e-01 -2.10507560e+00 5.14980815e-02 2.52282053e-01
7.76144803e-01 4.72094297e-01 -2.77548999e-01 1.28071502e-01
-1.86740264e-01 -1.23013761e-02 -6.00235425e-02 -8.13813090e-01
8.33009258e-02 -2.69288510e-01 3.68724624e-03 4.78796035e-01
4.07969542e-02 1.07214117e+00 -8.30440164e-01 -3.44559848e-01
3.52424681e-01 4.61306155e-01 -1.74382880e-01 2.15213895e-01
4.55704689e-01 5.47852457e-01 -2.68529683e-01 7.39792109e-01
6.44110620e-01 -3.08855265e-01 -1.22094974e-01 -4.18439090e-01
-5.36048226e-02 -6.02122918e-02 -9.56180453e-01 1.39553654e+00
-2.52355903e-01 4.50741261e-01 -2.59890258e-01 -1.24140429e+00
1.01872659e+00 3.63031439e-02 4.70474333e-01 -8.56361568e-01
1.09262116e-01 -5.05312644e-02 -3.30519646e-01 -1.48120716e-01
3.67958874e-01 1.76428288e-01 -2.11036995e-01 4.94511425e-01
3.71341288e-01 8.68088841e-01 1.06687970e-01 9.90614370e-02
1.00648284e+00 6.91176653e-02 1.09820636e-02 -1.31917119e-01
7.47118771e-01 -3.28997880e-01 6.38516784e-01 1.14517868e+00
-7.04107642e-01 8.72150660e-01 1.52677283e-01 -6.64730191e-01
-1.11290920e+00 -1.19293642e+00 -1.33342566e-02 1.32918942e+00
2.05621228e-01 -3.23676318e-01 -9.80326355e-01 -9.27646875e-01
1.28266662e-01 2.04332545e-01 -8.10710669e-01 -1.21261500e-01
-4.18931961e-01 -7.70485699e-01 8.92751753e-01 6.98184490e-01
1.18334138e+00 -8.75058651e-01 -2.48296186e-01 6.47004470e-02
-4.09076333e-01 -1.06813776e+00 -5.83542407e-01 -5.75548746e-02
-4.35467958e-01 -9.68581855e-01 -1.30865026e+00 -8.27756405e-01
7.54374623e-01 5.07245958e-01 8.85719478e-01 4.12825346e-02
-3.46193552e-01 6.25888884e-01 -4.83262956e-01 -1.07336864e-01
2.13598683e-01 2.09882975e-01 3.62255722e-01 6.37298405e-01
7.06512690e-01 -6.35241449e-01 -7.91666448e-01 4.56827432e-01
-4.49749589e-01 -5.09542711e-02 4.98348594e-01 1.01198614e+00
5.85450053e-01 9.48553756e-02 4.69632447e-01 -6.68037176e-01
4.68757719e-01 -2.64149070e-01 -3.33206117e-01 3.76616299e-01
-5.58975160e-01 -1.24576613e-01 7.96147525e-01 -4.90213364e-01
-1.07582223e+00 6.60226792e-02 2.25391433e-01 -5.42205691e-01
-3.80884796e-01 4.01158817e-02 -2.41696820e-01 -1.10568866e-01
4.99059767e-01 3.75626385e-01 -1.89593658e-01 -6.44590080e-01
3.69013518e-01 8.45762968e-01 8.90037000e-01 -5.00286043e-01
1.13318610e+00 5.41896105e-01 -3.57207835e-01 -5.94856918e-01
-8.96359444e-01 -6.01194620e-01 -9.10887957e-01 -3.82518023e-01
9.69654262e-01 -1.15655863e+00 -6.63911343e-01 1.14897776e+00
-7.28226066e-01 -3.39153409e-01 -4.28730011e-01 3.32404882e-01
-2.26154923e-01 3.85088921e-01 -5.77333033e-01 -6.16972029e-01
-4.55677897e-01 -8.99553239e-01 9.22866583e-01 6.20012224e-01
7.91022182e-02 -7.59766340e-01 2.55254228e-02 6.45019770e-01
4.12533581e-01 -1.52842356e-02 2.99848408e-01 -7.76486039e-01
-2.99681902e-01 -2.76672155e-01 -6.31304562e-01 5.54397762e-01
1.58654988e-01 -4.85746145e-01 -1.38445294e+00 -5.94203472e-01
-4.64708269e-01 -4.61822152e-01 1.24790394e+00 1.40119940e-01
1.35702395e+00 -2.89550841e-01 -4.57837999e-01 9.07229781e-01
1.16023862e+00 -4.12715197e-01 5.14698684e-01 5.88473737e-01
1.10635006e+00 6.11901045e-01 1.79793224e-01 5.43672562e-01
7.30974317e-01 7.75198162e-01 1.43774137e-01 -2.67600864e-01
-5.20558417e-01 -4.04916257e-01 2.99283475e-01 4.38159585e-01
-4.87238586e-01 -1.25327855e-01 -7.31222451e-01 7.71515727e-01
-1.83494687e+00 -1.02647352e+00 2.01095864e-01 2.22570539e+00
5.83704829e-01 -9.92996842e-02 6.64349258e-01 4.78498004e-02
1.12575078e+00 1.69577152e-01 -6.17160320e-01 1.50413081e-01
-1.43644765e-01 8.85762796e-02 7.11257756e-01 8.44335854e-02
-1.50970984e+00 8.26747417e-01 5.16212654e+00 9.56206083e-01
-8.71339321e-01 3.26839805e-01 7.79111266e-01 -5.53667024e-02
5.54102771e-02 -1.91470310e-01 -1.05040073e+00 7.00104415e-01
7.27615356e-01 2.18461808e-02 5.37746012e-01 6.84394121e-01
-4.16484028e-02 1.78272441e-01 -1.01844645e+00 1.41733706e+00
1.75698414e-01 -1.04269850e+00 8.22456330e-02 9.26209241e-02
7.45648861e-01 8.42193663e-02 2.16373399e-01 3.07788670e-01
4.45517570e-01 -1.02986169e+00 7.76339173e-01 7.22072124e-01
1.10772693e+00 -7.52846658e-01 8.03004503e-01 3.74151357e-02
-1.48909414e+00 -3.81276906e-01 -4.03096765e-01 1.40133932e-01
-7.38437921e-02 7.18533993e-01 -3.66371989e-01 6.50798023e-01
1.28838205e+00 9.34636533e-01 -9.36561465e-01 1.05016744e+00
-1.12405114e-01 6.40957117e-01 -2.61729449e-01 3.72388422e-01
4.24816683e-02 -2.17988398e-02 3.39772373e-01 1.25803316e+00
1.19248107e-01 -1.67376146e-01 1.25568867e-01 8.88083875e-01
-3.05927962e-01 -2.49802709e-01 -2.62353867e-01 2.11747944e-01
3.73283923e-01 1.43623650e+00 -4.38148052e-01 -3.75479907e-01
-5.53029537e-01 1.43168366e+00 6.68060303e-01 5.41329801e-01
-8.49843323e-01 -5.87374806e-01 5.75355768e-01 1.43423259e-01
4.39960122e-01 5.85727543e-02 7.45688230e-02 -1.36227643e+00
1.67870209e-01 -6.12941861e-01 5.31777561e-01 -4.56123918e-01
-1.88208532e+00 5.99011958e-01 -8.30019265e-02 -1.51940250e+00
-9.85787809e-02 -6.27807856e-01 -5.45545638e-01 8.32183719e-01
-1.70359159e+00 -1.73325992e+00 -6.77940965e-01 1.01489604e+00
2.26335332e-01 -6.10988855e-01 6.12375379e-01 5.21443903e-01
-8.82329345e-01 1.47355843e+00 1.76539227e-01 8.51561546e-01
8.11869800e-01 -1.14538252e+00 3.87116075e-01 1.02092457e+00
5.81954457e-02 5.24508655e-01 5.57378940e-02 -5.20239592e-01
-9.28102851e-01 -1.47844028e+00 7.48302758e-01 -2.80246943e-01
3.88036132e-01 -4.23423052e-01 -5.52297413e-01 8.00940394e-01
-1.79421484e-01 3.16906303e-01 7.10583389e-01 9.79787037e-02
-8.26902390e-01 -5.38361073e-01 -1.25498164e+00 4.53594565e-01
1.38944352e+00 -6.85467303e-01 -3.87401551e-01 7.95768276e-02
3.07903290e-01 1.17396317e-01 -8.91887665e-01 2.58136690e-01
7.61598587e-01 -1.13606393e+00 1.11230850e+00 -3.09010953e-01
9.74422544e-02 -3.23681355e-01 8.38259980e-02 -1.10753334e+00
-7.99355507e-01 -5.39436698e-01 7.66438339e-03 1.61018717e+00
2.73276083e-02 -9.03230131e-01 8.59198689e-01 5.67845225e-01
2.02063859e-01 -3.12190831e-01 -9.37082827e-01 -9.62292194e-01
-7.61190755e-03 -1.05569869e-01 6.08090341e-01 8.98484230e-01
-2.64345706e-01 7.82338455e-02 -6.64199352e-01 7.71609023e-02
1.17601728e+00 -1.79826781e-01 8.93135548e-01 -1.34654510e+00
-2.61044025e-01 -5.14903963e-01 -5.48241615e-01 -1.16113138e+00
1.94333568e-01 -1.15667880e+00 -1.87019944e-01 -1.14494336e+00
6.90499961e-01 -6.28990114e-01 -6.58075690e-01 7.93264031e-01
-1.57662600e-01 7.33547807e-01 3.53485554e-01 5.39132535e-01
-8.18321645e-01 6.05024695e-01 8.40822577e-01 -5.30704021e-01
-2.05136612e-02 2.46709641e-02 -7.90350437e-01 5.67546785e-01
7.89180338e-01 -3.70033890e-01 -1.48810238e-01 -3.75646383e-01
-3.27902824e-01 -5.08992493e-01 1.00510633e+00 -1.45452511e+00
3.77064139e-01 3.18855613e-01 8.00998151e-01 -4.85033691e-01
2.36674652e-01 -6.02746367e-01 -5.09404019e-03 1.89882562e-01
-5.35870254e-01 -1.29847214e-01 -1.92978770e-01 5.74473381e-01
-1.36650771e-01 -1.10419422e-01 7.52460897e-01 -1.41278073e-01
-7.68655777e-01 7.64127135e-01 -1.85286865e-01 -3.23400088e-02
8.19126546e-01 -4.87008959e-01 -2.94239700e-01 -2.32073322e-01
-7.72453547e-01 9.35331360e-02 4.09832984e-01 4.72910613e-01
4.44078386e-01 -1.46666145e+00 -8.89070392e-01 3.02612752e-01
3.51566344e-01 -2.60596216e-01 4.71329093e-01 7.85217941e-01
-5.41768819e-02 3.19381595e-01 -4.11893606e-01 -5.10562122e-01
-1.09440887e+00 4.66105431e-01 6.03110254e-01 -5.29124022e-01
-7.81928599e-01 9.92203295e-01 4.01546538e-01 -5.54323673e-01
3.05508196e-01 3.15783441e-01 -2.16937944e-01 -2.82054748e-02
7.74548233e-01 6.78521931e-01 -1.65788621e-01 -9.04529870e-01
-4.38810736e-01 6.89047694e-01 -3.80344093e-01 -1.63452886e-02
1.24774873e+00 -3.93013209e-01 -5.36973774e-02 4.30545723e-03
1.34695148e+00 -1.91828579e-01 -1.51885676e+00 -6.49989843e-01
-2.37415463e-01 -3.02289784e-01 -1.53638810e-01 -7.96469033e-01
-1.36529505e+00 7.01697588e-01 1.02060783e+00 -2.87899282e-02
1.16122437e+00 3.07243820e-02 9.82184410e-01 3.28613669e-01
3.45188320e-01 -1.29032660e+00 1.62992895e-01 3.36971015e-01
4.69827920e-01 -1.22024238e+00 -1.50141135e-01 -7.12958276e-02
-5.03569424e-01 9.56165254e-01 6.63274169e-01 -2.69482911e-01
6.42144084e-01 -1.56238945e-02 -1.62493601e-01 5.06803840e-02
-9.91838612e-03 -5.14845312e-01 2.44271278e-01 1.10970259e+00
-1.09731324e-01 2.31745273e-01 2.33288601e-01 8.78521383e-01
-1.64403871e-01 1.17530569e-01 1.54697552e-01 4.19667453e-01
-6.74505308e-02 -1.01913106e+00 -3.41581255e-01 6.30399287e-01
-3.82421255e-01 -7.54637271e-02 -2.95006067e-01 4.46393043e-01
4.63741541e-01 9.81122851e-01 8.24159086e-02 -7.91327298e-01
1.91650301e-01 -2.21819177e-01 3.76866013e-01 -2.16167912e-01
-6.80509686e-01 -3.90194207e-01 1.08395718e-01 -5.53005874e-01
-5.70457935e-01 -6.03429079e-01 -7.41753221e-01 -4.05768126e-01
-7.98742548e-02 -1.60366580e-01 1.08708821e-01 9.21855390e-01
4.41808254e-01 2.80117720e-01 7.79746711e-01 -1.08352149e+00
-5.00641584e-01 -8.75625372e-01 -5.84532976e-01 6.92957878e-01
4.80944067e-01 -7.78281510e-01 -4.10442889e-01 1.48349911e-01] | [14.751924514770508, 1.03270423412323] |
bd124f03-f2c2-4a24-893d-e53d11ff0e64 | annotation-inspired-implicit-discourse | 2306.06480 | null | https://arxiv.org/abs/2306.06480v1 | https://arxiv.org/pdf/2306.06480v1.pdf | Annotation-Inspired Implicit Discourse Relation Classification with Auxiliary Discourse Connective Generation | Implicit discourse relation classification is a challenging task due to the absence of discourse connectives. To overcome this issue, we design an end-to-end neural model to explicitly generate discourse connectives for the task, inspired by the annotation process of PDTB. Specifically, our model jointly learns to generate discourse connectives between arguments and predict discourse relations based on the arguments and the generated connectives. To prevent our relation classifier from being misled by poor connectives generated at the early stage of training while alleviating the discrepancy between training and inference, we adopt Scheduled Sampling to the joint learning. We evaluate our method on three benchmarks, PDTB 2.0, PDTB 3.0, and PCC. Results show that our joint model significantly outperforms various baselines on three datasets, demonstrating its superiority for the task. | ['Michael Strube', 'Wei Liu'] | 2023-06-10 | null | null | null | null | ['relation-classification', 'implicit-discourse-relation-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.79392332e-01 1.00142395e+00 -3.91713411e-01 -5.48638761e-01
-7.45901883e-01 -5.03088415e-01 9.58700597e-01 1.48272365e-01
-1.07628122e-01 1.07240105e+00 7.66351819e-01 -5.44275880e-01
2.73770481e-01 -8.14488292e-01 -6.43135607e-01 -2.85650820e-01
1.96725056e-01 5.10981023e-01 2.53560424e-01 -3.20275426e-01
3.07980762e-03 -3.67024064e-01 -9.64794040e-01 8.99787724e-01
1.01160896e+00 9.23751175e-01 4.25172597e-02 5.65070450e-01
-6.45837337e-02 1.65035188e+00 -9.47513580e-01 -6.55345201e-01
-2.64012307e-01 -5.44407189e-01 -1.38728762e+00 -1.61601946e-01
1.15405833e-02 -4.41022962e-01 -2.81608880e-01 8.15487862e-01
1.22409679e-01 6.38848171e-02 8.86441469e-01 -9.49852705e-01
-7.24701762e-01 1.30612624e+00 -3.88237327e-01 8.75014067e-02
4.79326993e-01 -3.12692137e-03 1.54484582e+00 -5.95459282e-01
7.89142251e-01 1.58951104e+00 4.42109704e-01 5.26241958e-01
-1.42365932e+00 -4.10237372e-01 5.23132741e-01 3.03365797e-01
-8.44238281e-01 -5.93186259e-01 8.54239464e-01 -5.81180632e-01
1.02045333e+00 4.58896369e-01 2.89696664e-01 1.53459430e+00
-3.07035804e-01 1.03020561e+00 9.04650033e-01 -3.60460669e-01
4.18832339e-02 -2.94919997e-01 4.00214404e-01 4.39828068e-01
-1.25644103e-01 -8.60653296e-02 -4.89672959e-01 -3.26765701e-02
2.64410675e-01 -8.52748990e-01 -4.71615255e-01 4.08347607e-01
-1.07011044e+00 8.27525437e-01 6.42915964e-01 1.21092960e-01
-2.40127102e-01 3.16326134e-02 6.13520265e-01 2.24190995e-01
7.45150506e-01 7.10516930e-01 -2.66477704e-01 -3.13804090e-01
-3.96869183e-01 4.52478051e-01 9.31328177e-01 8.16709757e-01
2.76641369e-01 -3.67983341e-01 -9.12784874e-01 9.20018613e-01
3.24482560e-01 -2.00078130e-01 8.02187771e-02 -1.11142766e+00
9.79646742e-01 7.52957344e-01 6.16284572e-02 -8.80344093e-01
-2.80130446e-01 -6.29120469e-02 -6.62409246e-01 -6.62069917e-02
7.03747630e-01 -3.62201571e-01 -3.76939833e-01 1.83019447e+00
2.46072561e-01 -2.49187853e-02 3.00602883e-01 8.85850668e-01
1.28337586e+00 7.05030322e-01 1.95111543e-01 -2.05218524e-01
1.27460349e+00 -1.38926303e+00 -9.93756950e-01 -3.24601114e-01
8.39508414e-01 -5.35478711e-01 1.01581538e+00 1.75899416e-01
-1.12617302e+00 -3.66615146e-01 -1.11773646e+00 -4.73686486e-01
3.52292925e-01 9.59289446e-02 6.10780001e-01 -1.34814650e-01
-5.26157200e-01 5.17172456e-01 -8.52837324e-01 3.11074942e-01
6.56037748e-01 -1.46702528e-01 7.85773769e-02 2.55205989e-01
-1.46302319e+00 1.15398872e+00 7.35555947e-01 2.21273139e-01
-6.42936587e-01 -5.30657768e-01 -8.92615080e-01 2.14821935e-01
4.80333596e-01 -6.52578950e-01 1.64121473e+00 -8.40052545e-01
-1.73983097e+00 8.33480179e-01 -1.85499683e-01 -7.47099221e-01
7.47201145e-01 -4.95578915e-01 1.57207370e-01 -1.48881406e-01
3.10959592e-02 5.61696291e-01 3.72150242e-01 -1.07374799e+00
-5.78123987e-01 2.59851396e-01 5.29799283e-01 2.69334853e-01
-1.10277109e-01 -9.93952304e-02 -6.41502440e-02 -5.87850153e-01
-8.88801813e-02 -6.96201444e-01 -8.95457864e-02 -3.32024455e-01
-9.97500360e-01 -9.33670878e-01 6.67453349e-01 -7.90088773e-01
1.17693090e+00 -1.95556438e+00 3.98847431e-01 -2.37846375e-01
3.48463207e-01 3.28224480e-01 -1.33383304e-01 1.67044878e-01
2.62199286e-02 9.84363407e-02 -2.23860160e-01 -5.84460795e-01
1.23025827e-01 3.68245989e-01 -6.07390165e-01 8.88129920e-02
6.94446266e-01 8.71272564e-01 -1.22223890e+00 -5.42537451e-01
-2.16631681e-01 1.15454115e-01 -5.65930605e-01 6.90517962e-01
-9.72865582e-01 6.87726140e-01 -4.22047973e-01 3.65639955e-01
3.33113283e-01 -4.42406207e-01 7.44148374e-01 -2.23956347e-01
2.24916875e-01 1.62892008e+00 -4.80552763e-01 1.33287704e+00
-5.33155501e-01 1.06991386e+00 4.92241196e-02 -1.20318770e+00
8.62116754e-01 7.21859097e-01 -7.94500634e-02 -6.09278917e-01
3.18382621e-01 1.46888807e-01 4.65971500e-01 -7.46929884e-01
3.83805543e-01 7.03087226e-02 -2.73140781e-02 5.68630040e-01
-1.92351509e-02 8.23170692e-02 4.49880332e-01 2.71958441e-01
1.11768854e+00 2.66969413e-01 2.63184816e-01 -1.20931298e-01
4.59561974e-01 -8.74137953e-02 7.00272679e-01 5.12321651e-01
7.33288601e-02 4.28609908e-01 1.28252387e+00 -2.02332616e-01
-5.69057167e-01 -8.84214282e-01 -9.19330567e-02 9.94865596e-01
-1.78668480e-02 -6.50872767e-01 -6.36464298e-01 -1.07936037e+00
-8.54609385e-02 8.22870016e-01 -7.42888987e-01 -7.27935834e-03
-1.01109326e+00 -5.61344385e-01 6.14768982e-01 6.20889902e-01
5.57532609e-01 -1.24004364e+00 -2.82199293e-01 3.44673663e-01
-8.02912652e-01 -1.30295765e+00 2.55918913e-02 5.46326935e-02
-4.71656799e-01 -1.26967359e+00 -1.78121939e-01 -8.74701262e-01
5.56961060e-01 -2.88163841e-01 1.56282842e+00 3.55393469e-01
5.62772989e-01 -3.82982612e-01 -4.63718504e-01 -1.65824071e-01
-9.66065168e-01 4.57765222e-01 -6.64859891e-01 -3.62149179e-01
8.39220881e-02 -5.46738207e-01 -3.38094056e-01 2.61736382e-02
-2.97642380e-01 6.19642496e-01 6.03420213e-02 1.19335878e+00
1.18245892e-01 -2.11289734e-01 8.37230384e-01 -1.33724928e+00
1.04850185e+00 -5.23752153e-01 -5.16816020e-01 2.31386423e-01
-2.00942039e-01 3.53300460e-02 4.50014621e-01 -4.09432769e-01
-1.28552699e+00 -4.10291404e-01 -2.21996352e-01 2.30319366e-01
1.48638025e-01 7.88033187e-01 -3.23510975e-01 8.27304006e-01
7.54422426e-01 -5.52344739e-01 -7.00999796e-02 -4.19083059e-01
5.28644681e-01 6.86579883e-01 6.46601081e-01 -1.02505469e+00
3.70402783e-01 -9.50872749e-02 -3.13827097e-01 -4.21279728e-01
-1.70269775e+00 1.64688081e-01 -2.46825144e-01 -1.65747907e-02
1.02084255e+00 -9.53513145e-01 -8.76575768e-01 1.36569336e-01
-1.87766695e+00 -8.58785331e-01 -1.46661714e-01 1.69616699e-01
-3.37354243e-01 1.59841865e-01 -1.17556274e+00 -7.59211481e-01
-4.11340773e-01 -1.13646531e+00 6.86448932e-01 4.98655885e-02
-1.00010860e+00 -1.07259905e+00 1.11760218e-02 5.51187754e-01
5.21837510e-02 6.05591178e-01 1.21402216e+00 -6.26835883e-01
-4.93420005e-01 3.66801649e-01 -4.72982258e-01 4.42144811e-01
2.78434336e-01 1.37680724e-01 -1.21042478e+00 1.24635592e-01
-5.65451570e-02 -8.45221281e-01 9.57064390e-01 -8.17899033e-03
1.14014626e+00 -7.88638413e-01 -4.33532625e-01 2.90167212e-01
5.27081728e-01 -3.55293192e-02 5.46087205e-01 4.01682794e-01
6.66417360e-01 8.98915350e-01 6.07771814e-01 1.29627660e-01
7.37057745e-01 6.54073179e-01 4.43782538e-01 1.38064940e-02
-3.85559797e-01 -1.69095457e-01 2.87208736e-01 7.88957000e-01
5.06366007e-02 -3.21827292e-01 -8.93870652e-01 5.75983584e-01
-2.14619803e+00 -8.21848452e-01 -4.29049700e-01 1.46365261e+00
1.76499200e+00 7.12690890e-01 4.24587540e-02 1.84634611e-01
5.01033723e-01 3.08833569e-01 -2.25558147e-01 -4.08406526e-01
-1.54716983e-01 2.17095584e-01 -4.42476034e-01 9.15967226e-01
-1.21753395e+00 9.26591635e-01 5.63085651e+00 5.84769368e-01
-1.05882835e+00 2.24143490e-02 9.65002298e-01 -1.64412167e-02
-3.70507419e-01 1.47388369e-01 -5.49834967e-01 6.03656232e-01
5.74669242e-01 -9.53153800e-03 2.54354537e-01 6.51427567e-01
-1.04985677e-01 -7.21164420e-02 -1.61736655e+00 1.92421347e-01
-4.15774286e-01 -1.63720298e+00 -2.10765004e-01 -2.97812849e-01
7.45018840e-01 -6.63183704e-02 -1.24371141e-01 6.59833372e-01
7.37925529e-01 -1.27459836e+00 8.06469440e-01 1.00540943e-01
5.69999099e-01 -2.55969644e-01 6.71073020e-01 4.88348663e-01
-6.98033571e-01 7.73384273e-02 -8.22243746e-03 -6.79231763e-01
3.95505220e-01 7.24898934e-01 -1.33932853e+00 3.19786310e-01
1.65873125e-01 7.32171595e-01 -3.61348420e-01 4.27308977e-01
-1.09683216e+00 1.05038726e+00 -2.08741009e-01 -1.10465489e-01
1.13828100e-01 4.92463224e-02 4.76713777e-01 1.25039434e+00
-2.37573802e-01 9.21487659e-02 1.97285682e-01 1.30072582e+00
-4.44689691e-01 -4.03458476e-01 -2.55431771e-01 2.56424975e-02
8.46821308e-01 1.13457000e+00 -2.88441986e-01 -3.82241577e-01
-1.71124414e-01 5.09157896e-01 9.61719811e-01 3.63607496e-01
-9.63541746e-01 -7.25708455e-02 3.94522041e-01 -7.81485811e-02
1.30755827e-01 1.17350677e-02 -6.76782668e-01 -8.56561720e-01
1.78276226e-01 -9.23744202e-01 3.82883579e-01 -5.80013216e-01
-1.48469448e+00 7.69886315e-01 5.75413406e-02 -8.26617897e-01
-5.81789613e-01 -4.94432926e-01 -9.75795865e-01 9.83198464e-01
-1.48580694e+00 -1.05067039e+00 -2.18666360e-01 -8.89922082e-02
4.52665150e-01 1.33299291e-01 9.55705225e-01 9.11719799e-02
-5.55685282e-01 6.07154727e-01 -5.64768374e-01 6.04825914e-01
7.13460028e-01 -1.43344092e+00 5.14457524e-01 6.16015196e-01
-1.25800863e-01 6.75365567e-01 7.11678684e-01 -4.48315412e-01
-4.17179286e-01 -1.03512609e+00 1.15744722e+00 -6.00722313e-01
7.32575297e-01 -5.54713190e-01 -1.13408434e+00 8.62936795e-01
7.28577137e-01 -1.65004075e-01 5.81988573e-01 7.73870528e-01
-5.27918398e-01 2.24376976e-01 -6.71284735e-01 7.52216578e-01
1.03880978e+00 -5.02729177e-01 -1.09121299e+00 4.51539993e-01
1.07863295e+00 -1.03978527e+00 -8.18103671e-01 5.06507695e-01
1.27122432e-01 -7.56244183e-01 7.63449907e-01 -6.93836212e-01
1.29735625e+00 -5.66117577e-02 5.57467528e-02 -1.32665336e+00
-8.13958570e-02 -6.53133392e-01 -7.03715026e-01 1.78887105e+00
9.31156337e-01 -3.78776401e-01 5.99293351e-01 5.37046075e-01
-2.77035445e-01 -1.13507473e+00 -8.98677170e-01 -3.89620781e-01
3.19376588e-01 -1.28264269e-02 5.19030988e-01 1.08331645e+00
5.09307325e-01 1.25640070e+00 -3.26025151e-02 -3.07392806e-01
1.48355618e-01 3.88880730e-01 6.94425821e-01 -1.16191721e+00
-7.46415854e-01 -6.93968832e-01 5.00333250e-01 -1.33428276e+00
7.35618472e-01 -9.89022017e-01 2.37284735e-01 -1.58820963e+00
-4.63626720e-02 -7.84347057e-01 1.32324830e-01 5.83753526e-01
-7.68129110e-01 -2.35888436e-02 1.70077696e-01 8.21821019e-02
-5.79462409e-01 7.52839029e-01 1.53818130e+00 -3.63828629e-01
-1.37922496e-01 1.54340211e-02 -8.08676302e-01 8.88934135e-01
7.14170992e-01 -2.84468532e-01 -5.99095762e-01 -8.07243645e-01
4.32739437e-01 2.08183169e-01 2.36983225e-01 -4.14449960e-01
-1.28112614e-01 -1.26622379e-01 -1.18571140e-01 -5.03104508e-01
4.74895298e-01 -1.05294719e-01 -4.88746881e-01 1.86507359e-01
-9.67776239e-01 -2.95675635e-01 -3.03997118e-02 2.96863198e-01
-3.48648161e-01 -2.27372050e-01 4.34090912e-01 8.20213780e-02
-2.41928138e-02 -3.57025504e-01 -3.70347917e-01 6.34313226e-01
5.96146345e-01 4.08648700e-01 -1.06916595e+00 -3.34356785e-01
-6.54601455e-01 5.64443469e-01 6.61296695e-02 2.72568077e-01
2.07332984e-01 -1.32212162e+00 -9.50505435e-01 -3.11732143e-01
-1.21788211e-01 9.49639678e-01 -3.48687410e-01 9.08652782e-01
-4.18629944e-01 2.26862565e-01 2.47678146e-01 -4.72417980e-01
-1.20298362e+00 2.15351179e-01 2.31295511e-01 -8.47435832e-01
-6.97179735e-01 1.21758974e+00 2.83357501e-01 -3.18554133e-01
4.83591348e-01 -6.67194486e-01 -3.63121033e-01 -5.33671156e-02
3.83204818e-01 -2.63523757e-02 -7.40329921e-02 -2.78280616e-01
-1.19145222e-01 -3.79229069e-01 -2.89182663e-01 -2.85713039e-02
1.26607108e+00 8.98561105e-02 -3.59329283e-01 3.94988656e-01
9.39929485e-01 3.23138051e-02 -1.59948325e+00 -2.80669242e-01
4.76375759e-01 -4.63722646e-02 -3.28259408e-01 -9.58008111e-01
-6.11434937e-01 8.13012362e-01 -5.42578697e-01 5.39636314e-01
7.14287519e-01 2.25751773e-02 8.38899553e-01 4.24122959e-01
-1.87678829e-01 -9.23880696e-01 3.55141342e-01 1.09032702e+00
1.08784211e+00 -1.13818824e+00 5.50054871e-02 -8.42469454e-01
-6.36313379e-01 8.76971483e-01 1.03305149e+00 -2.57662475e-01
2.44854644e-01 4.42660272e-01 1.13419391e-01 -9.74848717e-02
-1.27221859e+00 1.93026036e-01 2.51025259e-01 6.14316642e-01
1.12697744e+00 1.55331686e-01 -5.38712263e-01 9.08543289e-01
-4.53146398e-01 -2.42203221e-01 3.79165500e-01 7.08406746e-01
1.63315870e-02 -1.25169694e+00 -7.24941492e-02 1.88926071e-01
-3.98928016e-01 -1.49559185e-01 -7.55483508e-01 5.85471392e-01
-3.85546759e-02 1.00569451e+00 1.17366463e-01 -2.28069440e-01
1.18736662e-01 -5.99560216e-02 6.26036465e-01 -6.86439812e-01
-7.57871985e-01 -2.18954057e-01 1.19644225e+00 -2.43744344e-01
-4.53094900e-01 -5.12961388e-01 -1.37700546e+00 -1.90057799e-01
-4.06627655e-01 3.40474218e-01 -5.46540208e-02 1.13884270e+00
2.26096407e-01 1.14544153e+00 3.62616062e-01 -5.80780387e-01
-7.71118462e-01 -1.47183144e+00 2.02209726e-01 7.28073597e-01
3.94779444e-01 -7.65171230e-01 1.50615126e-02 9.74639580e-02] | [10.808899879455566, 9.259331703186035] |
99f3d754-573b-4654-a7be-3d18b9118a4a | computational-ergonomics-for-task-delegation | 2203.11007 | null | https://arxiv.org/abs/2203.11007v2 | https://arxiv.org/pdf/2203.11007v2.pdf | Computational ergonomics for task delegation in Human-Robot Collaboration: spatiotemporal adaptation of the robot to the human through contactless gesture recognition | The high prevalence of work-related musculoskeletal disorders (WMSDs) could be addressed by optimizing Human-Robot Collaboration (HRC) frameworks for manufacturing applications. In this context, this paper proposes two hypotheses for ergonomically effective task delegation and HRC. The first hypothesis states that it is possible to quantify ergonomically professional tasks using motion data from a reduced set of sensors. Then, the most dangerous tasks can be delegated to a collaborative robot. The second hypothesis is that by including gesture recognition and spatial adaptation, the ergonomics of an HRC scenario can be improved by avoiding needless motions that could expose operators to ergonomic risks and by lowering the physical effort required of operators. An HRC scenario for a television manufacturing process is optimized to test both hypotheses. For the ergonomic evaluation, motion primitives with known ergonomic risks were modeled for their detection in professional tasks and to estimate a risk score based on the European Assembly Worksheet (EAWS). A Deep Learning gesture recognition module trained with egocentric television assembly data was used to complement the collaboration between the human operator and the robot. Additionally, a skeleton-tracking algorithm provided the robot with information about the operator's pose, allowing it to spatially adapt its motion to the operator's anthropometrics. Three experiments were conducted to determine the effect of gesture recognition and spatial adaptation on the operator's range of motion. The rate of spatial adaptation was used as a key performance indicator (KPI), and a new KPI for measuring the reduction in the operator's motion is presented in this paper. | ['Alina Glushkova', 'Sotiris Manitsaris', 'Gavriela Senteri', 'Dimitris Papanagiotou', 'Brenda Elizabeth Olivas-Padilla'] | 2022-03-21 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 1.23877026e-01 5.00495017e-01 1.37775391e-01 8.50607753e-02
-3.07949215e-01 -1.55988544e-01 -7.98356682e-02 2.25838367e-02
-8.46937954e-01 1.36494562e-01 8.75099972e-02 -1.89351320e-01
-1.06297469e+00 -3.30029815e-01 -5.60301661e-01 -4.71245944e-01
-3.99221443e-02 5.36756992e-01 -7.00866953e-02 -2.31924340e-01
3.50844324e-01 1.10418189e+00 -1.58246553e+00 1.48322523e-01
5.60194612e-01 6.26885891e-01 1.08450079e+00 6.09546125e-01
4.26447153e-01 2.79854715e-01 -7.62891412e-01 1.17457129e-01
5.16933739e-01 -3.73317182e-01 -7.60130703e-01 2.45546907e-01
-2.51261055e-01 -2.90867031e-01 5.30835278e-02 5.70362687e-01
7.73230851e-01 5.50708890e-01 5.63083529e-01 -1.01333964e+00
-8.06900784e-02 1.98899791e-01 7.99069852e-02 -2.70247877e-01
5.01520991e-01 4.54515398e-01 3.46813537e-02 -5.10134101e-01
7.38935888e-01 1.19246280e+00 4.00413990e-01 3.13590199e-01
-8.47336769e-01 -2.37640858e-01 -2.37726897e-01 4.57406372e-01
-1.36819863e+00 1.61122158e-01 8.49546611e-01 -6.59554899e-01
9.33192074e-01 4.98810470e-01 8.19623172e-01 9.35083568e-01
7.31018364e-01 8.47182721e-02 8.16384017e-01 -6.31702125e-01
3.54438066e-01 -1.26521721e-01 -2.29991838e-01 6.35547638e-01
3.69042158e-01 1.32461309e-01 -2.70786583e-01 3.26471806e-01
8.48113239e-01 1.04613379e-01 9.48553681e-02 -3.94457519e-01
-1.15013385e+00 4.00132507e-01 4.58831936e-01 4.82649624e-01
-8.30167413e-01 2.64910281e-01 4.43651915e-01 1.59472898e-01
-1.71311289e-01 9.53779221e-01 -5.43144822e-01 -3.41038048e-01
-1.67430505e-01 2.84449190e-01 4.51695025e-01 7.08795309e-01
3.73833448e-01 -2.40157753e-01 -3.72518212e-01 5.87628722e-01
1.13872133e-01 4.84128952e-01 2.39731908e-01 -1.42876315e+00
4.06258106e-01 8.10767531e-01 2.86340535e-01 -1.10836935e+00
-1.10553908e+00 -1.39247000e-01 -2.53727943e-01 6.18674099e-01
4.59047407e-01 -2.67328352e-01 -7.87554860e-01 1.13335860e+00
4.40922558e-01 -6.22510493e-01 -2.74899542e-01 1.41481507e+00
-2.07423661e-02 1.35922119e-01 4.66119573e-02 -1.44244909e-01
1.60991514e+00 -5.44519544e-01 -9.84961689e-01 -1.75511643e-01
9.92301762e-01 -7.47666717e-01 1.18852079e+00 5.41297793e-01
-7.99600244e-01 -1.04273677e+00 -9.08091903e-01 1.61636323e-01
-2.76402950e-01 3.45388561e-01 2.93213636e-01 8.51927400e-01
-3.91135871e-01 9.64576602e-01 -1.19480681e+00 -4.75605816e-01
-3.20623308e-01 5.36093116e-01 -3.45124781e-01 2.43427809e-02
-9.98871505e-01 1.68660605e+00 2.91444927e-01 6.48913085e-01
-6.81589425e-01 -2.95583755e-01 -9.13805366e-01 -1.89500660e-01
6.21197820e-01 -6.53567314e-01 8.75232220e-01 -3.90002906e-01
-1.62140501e+00 3.28660071e-01 4.24870104e-01 -1.26227349e-01
6.83547854e-01 -7.99389660e-01 -3.79829943e-01 2.26358697e-01
2.82542765e-01 3.01839262e-01 5.36295652e-01 -1.15874183e+00
-2.96477497e-01 -6.12697899e-01 -1.87103406e-01 5.48798442e-01
7.91462362e-02 9.52926278e-02 -4.01494563e-01 -3.63980621e-01
2.55830437e-02 -1.44898474e+00 -3.26950133e-01 -2.28043631e-01
-2.39604309e-01 -1.75128967e-01 6.49896026e-01 -9.08519566e-01
8.37583542e-01 -2.10627055e+00 5.16766191e-01 3.89095366e-01
-2.70425618e-01 1.91701293e-01 -7.02404091e-03 3.93141180e-01
-9.12545174e-02 -2.76893467e-01 -6.21483549e-02 2.92725451e-02
-4.85980436e-02 3.53879660e-01 6.23473167e-01 5.28406262e-01
-1.84456594e-02 6.02076650e-01 -7.23356128e-01 -1.82915121e-01
6.18135452e-01 8.56155008e-02 -1.58118695e-01 3.42540503e-01
2.87459522e-01 7.88396478e-01 -1.48687214e-01 4.08390522e-01
2.76294500e-01 6.06989801e-01 3.60243022e-01 -2.01643586e-01
-1.68588072e-01 -2.26779625e-01 -1.23745441e+00 1.58879066e+00
-6.34135365e-01 1.96895152e-01 2.92412460e-01 -7.57308483e-01
1.06376910e+00 3.65859449e-01 8.80355835e-01 -8.66680145e-01
2.55780578e-01 2.53208578e-01 1.59538478e-01 -1.20208240e+00
2.35210612e-01 1.01333812e-01 -1.79219484e-01 2.59917319e-01
-2.67933756e-01 -2.96870053e-01 -8.72171894e-02 -4.71918285e-01
1.18286109e+00 4.71558750e-01 6.02355599e-02 -1.84980229e-01
2.22556680e-01 7.15377107e-02 2.63509423e-01 4.66108739e-01
-2.93590158e-01 2.26034671e-01 2.05613896e-01 -3.54642540e-01
-7.27195084e-01 -9.88526225e-01 5.16218305e-01 8.85366321e-01
1.85383797e-01 4.59811501e-02 -1.08546388e+00 -4.35163677e-01
-6.80997819e-02 1.07864738e+00 -3.49037766e-01 -7.09281743e-01
-9.55509543e-01 -1.72874093e-01 2.23777220e-01 6.17003560e-01
1.19978823e-01 -1.19952571e+00 -1.55854404e+00 1.10990390e-01
-1.94174498e-01 -8.37320149e-01 -2.24015087e-01 3.19443405e-01
-7.45425105e-01 -1.34345758e+00 -7.42522061e-01 -5.64765632e-01
6.90213382e-01 7.05923066e-02 3.12386513e-01 -7.05571547e-02
-6.56091332e-01 7.55030990e-01 -5.09577155e-01 -7.04039693e-01
-6.25478506e-01 -1.08777381e-01 4.92945671e-01 -3.68316412e-01
1.20401699e-02 8.04973543e-02 -7.31879354e-01 7.97898531e-01
-5.08486390e-01 -1.45412773e-01 9.29797947e-01 3.53806287e-01
5.41930318e-01 1.37996033e-01 3.06128174e-01 -3.52313042e-01
8.57253253e-01 -2.19849534e-02 -2.83794075e-01 1.84461087e-01
-7.80259490e-01 -1.25896141e-01 2.02855751e-01 -5.61219513e-01
-1.35039437e+00 4.59440082e-01 4.55745794e-02 -3.46428752e-01
-6.27551675e-01 5.14058113e-01 -3.80490959e-01 1.13874257e-01
6.85444474e-01 -3.34493011e-01 3.51737827e-01 -5.66514969e-01
3.00256640e-01 4.25264299e-01 6.33818984e-01 -6.36426270e-01
4.61634219e-01 6.77270740e-02 3.29220474e-01 -8.38423729e-01
-1.81947872e-01 -6.91478789e-01 -1.16904306e+00 -9.10585105e-01
1.28085566e+00 -2.58721262e-01 -1.10632432e+00 2.11036965e-01
-1.29453695e+00 -5.24045289e-01 -3.33093196e-01 1.13623691e+00
-1.00578940e+00 3.73774827e-01 -1.39454678e-01 -1.07884729e+00
4.95115295e-02 -1.34107471e+00 1.09813845e+00 -2.21898124e-01
-7.90923595e-01 -3.97699863e-01 -2.10563540e-01 7.41612732e-01
1.25314400e-01 5.90630889e-01 9.55136776e-01 -3.84720385e-01
-7.04238564e-02 -5.49460411e-01 2.03270584e-01 4.25263315e-01
3.13757271e-01 -6.43011391e-01 -3.92046154e-01 -1.69225041e-05
4.36138809e-01 1.78810447e-01 -2.10578181e-02 5.04709601e-01
1.01597047e+00 -3.98820750e-02 -3.05366933e-01 8.00300837e-02
9.42740262e-01 5.15262604e-01 7.93738186e-01 4.80394304e-01
7.68720031e-01 1.48517740e+00 1.41090107e+00 1.31282806e-01
-1.39112517e-01 1.05057168e+00 7.07359433e-01 -1.69223875e-01
-1.16392121e-01 2.36102298e-01 3.77742052e-01 6.21326566e-01
-8.47250640e-01 2.17179969e-01 -1.02158213e+00 1.46419272e-01
-1.75646245e+00 -5.83951652e-01 -7.02691436e-01 2.06961060e+00
1.19535312e-01 1.71595544e-01 1.31878525e-01 1.65990382e-01
6.44764125e-01 -6.94682121e-01 -3.43572825e-01 -5.93850613e-01
5.99560142e-01 1.26583770e-01 9.50646520e-01 1.67941958e-01
-6.19407237e-01 3.22259367e-01 5.67170954e+00 3.84078264e-01
-6.06456637e-01 2.09839627e-01 7.37964362e-02 -3.53203475e-01
4.38846290e-01 -2.61850297e-01 -3.03593248e-01 3.96596760e-01
1.09780419e+00 4.70310628e-01 2.17301935e-01 9.43247318e-01
7.24330127e-01 -5.23322463e-01 -1.07743168e+00 6.44687831e-01
-1.33302644e-01 -7.18355656e-01 -6.27854526e-01 5.59328385e-02
4.95731607e-02 -6.14387333e-01 -3.42662707e-02 2.45547727e-01
-3.36412340e-01 -6.84316993e-01 8.13919246e-01 1.03127229e+00
5.48560619e-01 -7.94003844e-01 1.11491561e+00 4.65952486e-01
-9.44960117e-01 -4.76566136e-01 -4.09179330e-02 -2.16398865e-01
5.54796338e-01 5.17933965e-02 -1.15389645e+00 8.12313676e-01
7.20782638e-01 -5.02425015e-01 -5.73071539e-01 6.26551151e-01
-1.78881079e-01 -4.73606773e-02 -2.19601780e-01 -1.48247659e-01
8.03867429e-02 -1.96816891e-01 5.90329587e-01 9.12533164e-01
5.28973937e-01 -9.27129909e-02 8.48590657e-02 7.40063310e-01
7.18164802e-01 7.53659150e-03 -6.67596877e-01 2.69184291e-01
1.68998495e-01 9.69661355e-01 -9.58061039e-01 2.95708746e-01
1.18274450e-01 1.34295535e+00 -1.71324521e-01 3.68280828e-01
-7.47238457e-01 -6.53487504e-01 4.28526789e-01 3.52473229e-01
-3.46282423e-01 -5.37956953e-01 -3.65442127e-01 1.64629906e-01
3.23511451e-01 -6.13769829e-01 1.30165040e-01 -9.13279712e-01
-5.19199848e-01 -1.36141017e-01 5.62885880e-01 -1.13182282e+00
-4.30311501e-01 -1.10289562e+00 -1.72056153e-01 1.01906526e+00
-3.71309012e-01 -8.98311973e-01 -2.66196907e-01 1.07273653e-01
7.14984834e-01 2.13461574e-02 7.98118949e-01 4.87973820e-03
-5.42864919e-01 1.74971759e-01 -3.84906352e-01 -3.03702623e-01
6.17435634e-01 -1.03933752e+00 -1.22834548e-01 5.88543773e-01
-6.16885960e-01 7.54600108e-01 8.01611781e-01 -1.19084728e+00
-1.55549204e+00 -8.45550179e-01 5.62605679e-01 -7.12664962e-01
2.83078760e-01 -1.15524292e-01 -4.82131481e-01 3.56551588e-01
-4.14523989e-01 -6.25313699e-01 4.59868729e-01 -1.38274834e-01
5.86768150e-01 -2.23898347e-02 -1.05188441e+00 7.77606905e-01
1.16761684e+00 -3.40443254e-01 -7.24534392e-01 5.99999547e-01
7.44177938e-01 -4.90959167e-01 -1.06589246e+00 4.89348948e-01
7.72734642e-01 -5.71261585e-01 1.07937682e+00 -5.21564364e-01
7.70476460e-02 -1.90329164e-01 1.26228973e-01 -1.17603898e+00
-3.27813923e-01 -4.21694577e-01 1.31504540e-03 6.04375422e-01
4.88293357e-02 -1.71092302e-01 5.39356053e-01 8.52567673e-01
-7.12538242e-01 -6.09875560e-01 -1.08390784e+00 -1.25179386e+00
-5.35176218e-01 -7.13592649e-01 2.58434922e-01 4.40721601e-01
6.43959343e-02 -1.23195685e-01 -2.53259689e-01 4.35998052e-01
9.77982953e-02 -6.31442726e-01 9.66502666e-01 -1.13194156e+00
-1.66329771e-01 -3.35279405e-01 -5.55411696e-01 -9.73822623e-02
-1.31032974e-01 -5.21090746e-01 4.87331688e-01 -1.78476310e+00
-3.84868711e-01 3.04230992e-02 -8.47354680e-02 2.29096502e-01
1.23364344e-01 -4.98817533e-01 4.39276785e-01 3.77106965e-02
-1.34166509e-01 4.91468981e-02 1.43699276e+00 1.27454683e-01
-5.51076591e-01 4.27929223e-01 -1.03265285e-01 6.84544861e-01
8.19728553e-01 -5.13234377e-01 -4.38324541e-01 -2.44919151e-01
1.48765177e-01 2.76462615e-01 5.66312909e-01 -1.07580888e+00
2.61656374e-01 -2.83626735e-01 4.32441056e-01 -5.56002975e-01
4.87705320e-01 -1.27835441e+00 5.96742272e-01 1.31531739e+00
-2.10214064e-01 1.16427794e-01 1.54067367e-01 4.74962145e-01
3.83124918e-01 -5.74838758e-01 4.20731544e-01 -6.91552386e-02
-4.61166769e-01 -5.65511227e-01 -9.27556515e-01 -1.00614643e+00
1.31367743e+00 -5.79950571e-01 2.08704114e-01 2.62899958e-02
-1.18320584e+00 2.88868360e-02 2.40024209e-01 5.69797635e-01
6.65325820e-01 -1.04913139e+00 -2.23460391e-01 -5.71503416e-02
1.30846873e-01 -5.51999100e-02 6.03044569e-01 1.26828063e+00
-7.85743833e-01 5.81236541e-01 -5.63352883e-01 -4.20109898e-01
-1.24347329e+00 7.85710692e-01 1.85718060e-01 -1.43669978e-01
-5.35593808e-01 3.81708294e-01 -1.62342221e-01 -6.44388199e-01
2.53936678e-01 -4.28920269e-01 -7.87371472e-02 -2.54285455e-01
6.32328019e-02 1.37807715e+00 4.92254168e-01 -4.83158588e-01
-4.73778903e-01 4.87785190e-01 5.54053962e-01 -1.46185592e-01
9.25742805e-01 -2.50778794e-01 1.56857699e-01 4.82007563e-01
8.17391992e-01 -1.76008940e-01 -1.22347224e+00 4.75086272e-01
4.31680679e-01 -2.54507005e-01 -3.76367480e-01 -1.12808394e+00
-4.82434124e-01 5.52453399e-01 1.40055943e+00 -1.36316523e-01
1.10194361e+00 4.20078300e-02 4.44435775e-01 7.25121975e-01
4.58416700e-01 -1.90740275e+00 3.15933406e-01 1.31495088e-01
1.34635985e+00 -8.63284886e-01 -5.43548949e-02 -5.77618599e-01
-8.17633331e-01 1.09637237e+00 5.86389542e-01 1.63919330e-01
3.22941959e-01 7.26618171e-02 1.79153323e-01 -7.36657143e-01
3.42722267e-01 -1.86240524e-01 6.86923862e-01 1.01394618e+00
2.00336784e-01 4.81595904e-01 -7.05802321e-01 7.68162668e-01
-1.30489796e-01 -1.13196895e-01 2.00090498e-01 1.06835330e+00
-5.61524093e-01 -7.49186039e-01 -8.91446948e-01 2.32283115e-01
2.04427540e-03 9.47370648e-01 -4.37434822e-01 1.24064541e+00
6.27976120e-01 1.05109382e+00 -2.76627298e-02 -5.48240662e-01
1.09387994e+00 1.83866918e-01 4.62287813e-01 -9.19439673e-01
-6.38913631e-01 2.68854678e-01 2.85982817e-01 -1.13413620e+00
-1.37959614e-01 -7.42155194e-01 -1.58702493e+00 2.70652592e-01
-1.83805630e-01 -2.98731565e-01 1.18795705e+00 1.12852645e+00
2.39469692e-01 1.11526370e+00 1.75308228e-01 -1.20203912e+00
-5.64202428e-01 -1.09393382e+00 -7.01207817e-01 4.94566977e-01
-3.59384358e-01 -1.08703136e+00 1.81101292e-01 -1.00453950e-01] | [4.991964817047119, 0.9143423438072205] |
6ec97456-9d9e-4312-a6c1-8a2ee29171c7 | an-end-to-end-neural-network-for-image | 1907.01432 | null | https://arxiv.org/abs/1907.01432v3 | https://arxiv.org/pdf/1907.01432v3.pdf | An End-to-End Neural Network for Image Cropping by Learning Composition from Aesthetic Photos | As one of the fundamental techniques for image editing, image cropping discards unrelevant contents and remains the pleasing portions of the image to enhance the overall composition and achieve better visual/aesthetic perception. In this paper, we primarily focus on improving the accuracy of automatic image cropping, and on further exploring its potential in public datasets with high efficiency. From this respect, we propose a deep learning based framework to learn the objects composition from photos with high aesthetic qualities, where an anchor region is detected through a convolutional neural network (CNN) with the Gaussian kernel to maintain the interested objects' integrity. This initial detected anchor area is then fed into a light weighted regression network to obtain the final cropping result. Unlike the conventional methods that multiple candidates are proposed and evaluated iteratively, only a single anchor region is produced in our model, which is mapped to the final output directly. Thus, low computational resources are required for the proposed approach. The experimental results on the public datasets show that both cropping accuracy and efficiency achieve the state-ofthe-art performances. | ['Xujun Peng', 'Peng Lu', 'Hao Zhang', 'Xiaofu Jin'] | 2019-07-02 | null | null | null | null | ['image-cropping'] | ['computer-vision'] | [ 6.04311824e-01 5.39532974e-02 3.24589685e-02 -1.89707175e-01
-2.64447182e-01 -2.74382502e-01 1.99363187e-01 1.32387534e-01
-2.80892372e-01 4.11520243e-01 -1.90282241e-01 7.20085800e-02
2.02082381e-01 -1.09293318e+00 -9.77702260e-01 -8.29454660e-01
2.53185898e-01 -3.79499912e-01 2.47545943e-01 -7.41290301e-02
3.08749169e-01 4.52181756e-01 -1.75190866e+00 2.76735723e-01
1.19491827e+00 1.25780189e+00 5.16318440e-01 3.60996515e-01
-1.83927953e-01 6.60241187e-01 -5.25415361e-01 -5.38740933e-01
4.85082686e-01 -1.98563978e-01 -2.67566621e-01 5.65406621e-01
4.09772545e-01 -6.33074403e-01 9.29938927e-02 1.29731596e+00
4.31484193e-01 -1.61343198e-02 4.56987262e-01 -1.42156637e+00
-1.21857142e+00 5.36364079e-01 -9.92318034e-01 -4.66802210e-01
-7.31462613e-02 1.25922814e-01 7.81037390e-01 -1.27275431e+00
5.09772301e-01 1.12989116e+00 3.11155438e-01 2.23717257e-01
-1.00923455e+00 -9.43356097e-01 3.52665395e-01 1.45857766e-01
-1.36596894e+00 -3.06398630e-01 1.02503526e+00 -1.89229086e-01
3.06115419e-01 2.36520931e-01 8.14827621e-01 5.80677927e-01
5.27785718e-02 9.03311729e-01 9.27865565e-01 -6.14330471e-01
1.91878267e-02 3.66074085e-01 -3.16300541e-01 8.56026590e-01
2.55595714e-01 -1.12014227e-01 -3.65853399e-01 3.20057124e-01
8.56130242e-01 2.47981429e-01 -2.68039644e-01 -6.43136203e-01
-1.24927461e+00 5.79219818e-01 7.83963203e-01 1.36804521e-01
-6.20954633e-01 2.07003519e-01 5.17272688e-02 1.52585786e-02
4.08495545e-01 1.67638257e-01 -2.96915144e-01 4.57519829e-01
-1.00367427e+00 2.02494979e-01 1.35134354e-01 1.05497777e+00
9.37488317e-01 1.52630741e-02 -2.58885592e-01 9.43982899e-01
1.91704601e-01 4.96808678e-01 1.09339999e-02 -9.14953828e-01
3.79368275e-01 9.95013893e-01 3.12797308e-01 -1.28412974e+00
-1.10365443e-01 -6.05357111e-01 -1.09205759e+00 6.94241166e-01
9.08088684e-02 -1.40091136e-01 -8.64907563e-01 1.32678962e+00
4.04480487e-01 -1.54644430e-01 -3.21771763e-02 8.41693401e-01
7.54897416e-01 8.00723672e-01 2.95281738e-01 -1.15187310e-01
1.32750571e+00 -1.30557537e+00 -7.51589894e-01 -2.04567730e-01
1.36928812e-01 -1.18725109e+00 1.17534637e+00 6.26609266e-01
-1.26203454e+00 -7.50786066e-01 -1.27887177e+00 -2.15593979e-01
-4.60644633e-01 7.82678545e-01 5.44062018e-01 3.96427423e-01
-1.03393197e+00 4.42450583e-01 -2.48867989e-01 -1.01022489e-01
6.65483952e-01 1.68484766e-02 -2.79611617e-01 -1.35908693e-01
-9.22239065e-01 6.30475521e-01 6.46270275e-01 2.81855375e-01
-7.60284007e-01 -7.75517881e-01 -6.58050418e-01 2.25387752e-01
2.23470300e-01 -2.90070772e-01 8.94083023e-01 -1.47446752e+00
-1.23614013e+00 6.76737547e-01 1.12881819e-02 -2.08738342e-01
5.23502171e-01 -1.95871383e-01 -3.82362962e-01 8.53908360e-02
-2.33302936e-02 1.16346931e+00 1.11957455e+00 -1.82031906e+00
-1.04431057e+00 -1.33948429e-02 2.06521913e-01 3.36062938e-01
-7.76179612e-01 4.20161802e-03 -7.56592274e-01 -9.13441479e-01
1.84873089e-01 -7.01375604e-01 -1.21854737e-01 7.07735419e-01
-3.08217913e-01 6.04560114e-02 9.10810709e-01 -9.52667832e-01
1.32498789e+00 -2.13578796e+00 -8.84966459e-03 3.23475391e-01
1.04517773e-01 3.81158739e-01 -1.94483802e-01 1.72323272e-01
-1.43255785e-01 2.55085770e-02 -5.37465513e-01 -1.98453397e-01
-1.98266834e-01 -1.57453746e-01 -1.65411219e-01 2.44116843e-01
6.76566958e-01 1.02758646e+00 -7.95316041e-01 -5.70653617e-01
5.37152767e-01 4.96068329e-01 -4.08474982e-01 1.86359152e-01
-1.62500277e-01 1.17082983e-01 -2.49381125e-01 1.03847718e+00
1.27627003e+00 -4.76449504e-02 -2.30105314e-02 -6.38996601e-01
-1.40535548e-01 -5.62085867e-01 -1.28516746e+00 1.59704220e+00
-5.92320204e-01 5.46229124e-01 1.70526966e-01 -7.10063636e-01
1.17925358e+00 -1.60461634e-01 4.17492777e-01 -9.82735217e-01
1.97867140e-01 3.89018878e-02 -2.82540232e-01 -6.09917700e-01
8.36453438e-01 2.65625149e-01 1.39556184e-01 3.49461168e-01
-3.14464241e-01 9.17971060e-02 1.27834335e-01 8.23502168e-02
4.00584757e-01 3.79568338e-01 1.91015214e-01 -1.79126054e-01
5.45196533e-01 1.18235692e-01 5.06986499e-01 3.00750345e-01
-9.99409035e-02 7.75538623e-01 2.16180146e-01 -3.81198853e-01
-1.40584505e+00 -9.55558658e-01 -2.06867717e-02 1.12255394e+00
6.65023446e-01 6.10473901e-02 -1.04007435e+00 -3.92009169e-01
-2.00911880e-01 8.77857327e-01 -7.00710297e-01 -1.91654548e-01
-5.30841231e-01 -6.43880069e-01 1.00060299e-01 4.50948358e-01
8.04336071e-01 -1.50416303e+00 -6.30652726e-01 2.07792565e-01
-4.21760589e-01 -8.47236574e-01 -5.78578293e-01 -1.33011386e-01
-5.32173097e-01 -1.00073326e+00 -9.68923151e-01 -1.10051620e+00
1.14512658e+00 7.77980447e-01 8.04437339e-01 4.42777902e-01
-3.86035681e-01 -1.65101871e-01 -5.29772639e-01 -5.99873185e-01
-3.22973698e-01 -1.28475308e-01 -3.68467093e-01 3.89186442e-01
1.65419698e-01 -3.83317053e-01 -9.25101280e-01 1.11092068e-01
-1.27822077e+00 4.87799674e-01 1.10719097e+00 6.74747527e-01
5.79961300e-01 4.26585227e-01 5.34085155e-01 -5.21679103e-01
5.18911779e-01 -1.70430213e-01 -5.06643951e-01 5.26829302e-01
-6.75885797e-01 -2.67201841e-01 5.07541835e-01 -6.03440046e-01
-1.37095761e+00 1.82973042e-01 1.65139928e-01 -3.34221125e-01
1.57649755e-01 1.48159727e-01 -3.13410431e-01 -3.04115534e-01
4.31292355e-01 3.91253650e-01 4.71707210e-02 -3.24998081e-01
5.95628262e-01 6.17033780e-01 5.44481158e-01 -2.01677188e-01
1.01671672e+00 4.86182332e-01 -2.88071901e-01 -3.61882806e-01
-4.96488899e-01 1.44437738e-02 -7.14484155e-01 -5.72878182e-01
7.64073789e-01 -9.95353580e-01 -5.88013887e-01 6.40596151e-01
-1.10361755e+00 -5.23988903e-03 -1.19079642e-01 8.42230096e-02
-2.12172240e-01 4.65669811e-01 -2.39351571e-01 -9.25458789e-01
-7.98960149e-01 -1.06750488e+00 1.15500462e+00 6.16459548e-01
2.06543893e-01 -3.79881173e-01 -5.13888657e-01 1.66316256e-01
3.71536463e-01 4.10256296e-01 9.91408825e-01 1.97135061e-01
-7.72269249e-01 -1.68844834e-01 -8.79177034e-01 4.73399282e-01
2.39715934e-01 4.40286398e-01 -9.75213349e-01 -1.69255748e-01
-4.73271728e-01 -1.14331797e-01 8.22283864e-01 2.25105032e-01
1.53619814e+00 -1.61812082e-01 -1.75793961e-01 6.43449962e-01
1.65737212e+00 3.45806926e-01 8.56066227e-01 4.57869291e-01
6.68440282e-01 7.95416474e-01 1.13910723e+00 5.11679471e-01
1.97231114e-01 1.71627566e-01 9.50943410e-01 -7.92532742e-01
-2.69110829e-01 -4.74577069e-01 3.51690203e-01 4.33498532e-01
7.52789527e-02 -3.47410679e-01 -3.75939459e-01 7.90287137e-01
-1.79806495e+00 -6.59931302e-01 -2.46855784e-02 2.13753581e+00
6.78057611e-01 -3.29803419e-03 -3.18161309e-01 1.50552362e-01
1.10662496e+00 1.09206252e-01 -6.95152462e-01 -2.33923361e-01
-3.46180260e-01 1.38278365e-01 6.77588105e-01 1.11952007e-01
-1.12535000e+00 1.00161576e+00 5.87218142e+00 1.03496587e+00
-1.01317728e+00 -1.56990692e-01 9.17626321e-01 -6.12044632e-02
-3.33543956e-01 -1.81227461e-01 -5.10283053e-01 4.21303689e-01
2.39802338e-02 -2.21207850e-02 5.57896316e-01 8.96663010e-01
4.50643480e-01 -2.82979727e-01 -4.44274127e-01 8.89518499e-01
1.71028912e-01 -1.23087251e+00 2.40799949e-01 -1.85980767e-01
9.49345171e-01 -6.77195072e-01 2.77329743e-01 -6.73511848e-02
2.67630398e-01 -8.53027999e-01 1.03320158e+00 5.89646697e-01
9.93411362e-01 -1.08022487e+00 5.29721737e-01 6.61983415e-02
-1.30939782e+00 -4.53086346e-01 -6.60645366e-01 3.24687541e-01
1.49404883e-01 6.66641116e-01 -6.68186724e-01 6.29307687e-01
8.71894717e-01 5.37962675e-01 -7.70904064e-01 9.53675628e-01
-1.48971125e-01 1.34805977e-01 -6.44687712e-02 3.36134806e-02
2.06281200e-01 -3.67118239e-01 1.24588385e-01 9.86352503e-01
6.44001007e-01 -4.21755351e-02 6.38710782e-02 9.75423098e-01
-3.13181609e-01 4.76155251e-01 -2.32262194e-01 5.13401031e-02
3.85898858e-01 1.72707605e+00 -8.93669307e-01 -4.43750024e-01
-3.26634198e-01 1.07905948e+00 3.14393222e-01 1.73033327e-01
-1.06285655e+00 -5.38124382e-01 3.04560930e-01 8.95005316e-02
4.79381293e-01 1.25494733e-01 -4.71053332e-01 -9.59738016e-01
3.46621364e-01 -7.88456440e-01 -1.60462863e-03 -1.31357312e+00
-1.06763220e+00 4.88940597e-01 -3.17141980e-01 -1.61555016e+00
3.93638074e-01 -4.69125450e-01 -7.08368957e-01 8.80021274e-01
-1.77150977e+00 -1.66115344e+00 -9.28018332e-01 4.60051209e-01
6.79024398e-01 1.20873466e-01 5.16848743e-01 4.26292241e-01
-4.46079731e-01 5.36289692e-01 6.78864792e-02 4.53220233e-02
8.56640220e-01 -8.87667537e-01 2.93579787e-01 1.18459070e+00
-2.47136071e-01 2.76235044e-01 5.61794758e-01 -7.55881011e-01
-1.06382203e+00 -1.49654078e+00 6.01410329e-01 2.00635776e-01
2.34144941e-01 -1.75141349e-01 -8.50725114e-01 2.18988970e-01
5.27013063e-01 -2.12156698e-01 1.56862631e-01 -5.60175598e-01
-5.99121042e-02 -3.55327755e-01 -1.20911503e+00 8.00020397e-01
1.03652012e+00 6.64579421e-02 -1.36016786e-01 1.86395213e-01
9.43424404e-01 -2.07247794e-01 -7.33660221e-01 4.84389067e-01
7.59332240e-01 -9.84625459e-01 1.00903535e+00 -7.64998794e-02
7.51024127e-01 -6.41397536e-01 2.13428941e-02 -1.16186810e+00
-5.08082688e-01 -2.96509892e-01 1.73705518e-01 1.47801149e+00
3.20284843e-01 -1.52599856e-01 5.66106975e-01 5.05458832e-01
6.44471198e-02 -7.62377381e-01 -3.28544974e-01 -1.75869048e-01
-2.56261587e-01 -8.82263258e-02 9.64340568e-01 7.97916174e-01
-5.83562732e-01 -1.94617525e-01 -7.22612500e-01 2.01306000e-01
7.58339107e-01 4.68509525e-01 8.47529888e-01 -8.71479928e-01
2.07981259e-01 -3.73297572e-01 -3.10838725e-02 -7.98538327e-01
-2.40865469e-01 -5.50140142e-01 1.48360297e-01 -1.56572258e+00
2.77231574e-01 -4.67119098e-01 -4.56895202e-01 4.75110918e-01
-4.15695757e-01 5.96641362e-01 2.92324364e-01 -5.96810766e-02
-4.34231132e-01 6.35174394e-01 1.55157328e+00 -4.51009810e-01
-1.69246644e-01 -1.78862184e-01 -1.07022083e+00 6.70596004e-01
9.13499594e-01 -2.12810770e-01 -3.61900210e-01 -5.12367487e-01
-2.61322875e-03 -4.65959638e-01 3.67107511e-01 -8.49410653e-01
1.02595530e-01 -3.74246806e-01 8.89491379e-01 -9.22931015e-01
1.62780792e-01 -1.07685053e+00 2.05910996e-01 5.00653863e-01
-3.60787123e-01 -1.37703821e-01 3.40525843e-02 4.62960422e-01
5.43561880e-04 -2.85270035e-01 1.02927768e+00 -8.20964947e-02
-9.59288776e-01 4.32811767e-01 7.29812384e-02 -6.23587430e-01
1.37175310e+00 -2.90655524e-01 -6.55922741e-02 -2.24504501e-01
-4.18930292e-01 2.17433035e-01 6.84292853e-01 6.76880896e-01
7.55946755e-01 -1.48523510e+00 -6.67563200e-01 2.72991598e-01
3.11835855e-01 3.26386578e-02 5.52172780e-01 4.00849968e-01
-7.64549017e-01 -6.81064427e-02 -6.44650459e-01 -3.44839245e-01
-1.31774461e+00 1.00745857e+00 1.27248466e-01 1.05341360e-01
-5.65271199e-01 6.15590394e-01 5.54383457e-01 -1.55752510e-01
2.30104595e-01 -2.61376083e-01 -4.15722519e-01 9.44918767e-02
7.41305768e-01 2.50620306e-01 -1.07481495e-01 -5.78396022e-01
-6.69315532e-02 5.84626198e-01 -8.01063478e-02 2.57786810e-01
1.40592134e+00 -4.50959474e-01 -2.94584274e-01 -1.54524550e-01
9.91246164e-01 8.10557082e-02 -1.52234435e+00 -3.99196476e-01
-5.36832988e-01 -7.75968611e-01 1.65480569e-01 -7.91737854e-01
-1.54662263e+00 7.83540547e-01 8.99035931e-01 1.43665135e-01
1.63548625e+00 -4.46597666e-01 8.56671274e-01 4.51447293e-02
2.01385349e-01 -1.24680936e+00 8.12249184e-02 -3.26429993e-01
1.20696604e+00 -1.23950791e+00 1.75689995e-01 -6.34334862e-01
-8.36444676e-01 1.19202518e+00 9.48782027e-01 -3.24865282e-01
2.79306710e-01 2.53915280e-01 6.87960386e-02 1.16445214e-01
-3.14743221e-01 -2.31827602e-01 1.47616953e-01 7.82340884e-01
6.34657368e-02 1.71271234e-03 -2.76192069e-01 4.88384038e-01
2.12664716e-02 3.60893011e-02 4.35556054e-01 7.05495477e-01
-6.90635562e-01 -9.27904427e-01 -4.92390007e-01 4.43035573e-01
-2.52927542e-01 -2.47478679e-01 -3.32533121e-01 5.82775891e-01
5.12387335e-01 8.62818062e-01 9.87406299e-02 -3.10747743e-01
3.55714083e-01 -2.42828295e-01 1.93346247e-01 -1.42986268e-01
-4.91679996e-01 2.12659672e-01 -2.48112649e-01 -3.56105417e-01
-4.73961890e-01 -2.11348489e-01 -9.27215993e-01 -3.91441107e-01
-5.40357769e-01 -3.34351510e-01 7.50754774e-01 4.50833380e-01
4.29379344e-01 7.95220435e-01 8.93767118e-01 -9.16182518e-01
-1.35551929e-01 -8.02356720e-01 -5.65764248e-01 3.98551524e-01
2.69281603e-02 -5.95058739e-01 -6.29792586e-02 4.85823095e-01] | [11.2971830368042, -1.091201663017273] |
7c629e42-6f4c-476e-8a1d-4d0b2755bda6 | stratified-transfer-learning-for-cross-domain | 1801.00820 | null | http://arxiv.org/abs/1801.00820v1 | http://arxiv.org/pdf/1801.00820v1.pdf | Stratified Transfer Learning for Cross-domain Activity Recognition | In activity recognition, it is often expensive and time-consuming to acquire
sufficient activity labels. To solve this problem, transfer learning leverages
the labeled samples from the source domain to annotate the target domain which
has few or none labels. Existing approaches typically consider learning a
global domain shift while ignoring the intra-affinity between classes, which
will hinder the performance of the algorithms. In this paper, we propose a
novel and general cross-domain learning framework that can exploit the
intra-affinity of classes to perform intra-class knowledge transfer. The
proposed framework, referred to as Stratified Transfer Learning (STL), can
dramatically improve the classification accuracy for cross-domain activity
recognition. Specifically, STL first obtains pseudo labels for the target
domain via majority voting technique. Then, it performs intra-class knowledge
transfer iteratively to transform both domains into the same subspaces.
Finally, the labels of target domain are obtained via the second annotation. To
evaluate the performance of STL, we conduct comprehensive experiments on three
large public activity recognition datasets~(i.e. OPPORTUNITY, PAMAP2, and UCI
DSADS), which demonstrates that STL significantly outperforms other
state-of-the-art methods w.r.t. classification accuracy (improvement of 7.68%).
Furthermore, we extensively investigate the performance of STL across different
degrees of similarities and activity levels between domains. And we also
discuss the potential of STL in other pervasive computing applications to
provide empirical experience for future research. | ['Philip S. Yu', 'Lisha Hu', 'Yiqiang Chen', 'Jindong Wang', 'Xiaohui Peng'] | 2017-12-25 | null | null | null | null | ['cross-domain-activity-recognition'] | ['computer-vision'] | [ 4.63830978e-01 -2.86608428e-01 -6.78706765e-01 -3.92244071e-01
-8.33096921e-01 -7.56194174e-01 4.66494143e-01 -5.55846393e-02
-2.62634873e-01 1.03222167e+00 2.34422326e-01 4.74102609e-02
-1.56657279e-01 -7.16853738e-01 -6.09171152e-01 -8.37205589e-01
6.97880238e-02 1.85364604e-01 3.36835891e-01 3.56890410e-01
7.84747675e-02 9.91418883e-02 -1.21955061e+00 4.68824506e-01
1.18522716e+00 1.17161965e+00 -1.11406773e-01 -1.07484765e-01
-1.25171527e-01 9.21503484e-01 -5.66357195e-01 1.80318803e-02
1.13795787e-01 -5.41648686e-01 -9.72221494e-01 2.39269286e-01
1.17063396e-01 -1.59456462e-01 -2.58263081e-01 8.30012798e-01
3.62471402e-01 1.52032033e-01 9.01091635e-01 -1.34137213e+00
-6.33270085e-01 2.72920668e-01 -8.41138482e-01 3.43968511e-01
5.45850635e-01 -6.38578311e-02 7.70178914e-01 -8.37270856e-01
3.61650795e-01 8.64462912e-01 6.50078356e-01 4.02492285e-01
-1.32172155e+00 -1.18955386e+00 4.35387373e-01 4.56917942e-01
-1.57859945e+00 -2.28697836e-01 9.95190144e-01 -6.41311884e-01
5.83547473e-01 -1.02833025e-01 4.34962451e-01 1.41813779e+00
-2.26228490e-01 1.11866331e+00 1.41651058e+00 -2.56466657e-01
4.23695743e-01 1.26132056e-01 3.60684037e-01 4.77487892e-01
3.28364670e-01 -2.91204065e-01 -8.51481855e-01 -2.47357070e-01
6.11590445e-01 1.16671398e-01 -3.34703624e-01 -7.38342285e-01
-1.32559490e+00 4.91980374e-01 3.16247970e-01 3.57171953e-01
-2.52193719e-01 -2.74816126e-01 3.77808571e-01 1.49338588e-01
5.54722250e-01 5.84882796e-02 -5.86177230e-01 -3.74931782e-01
-6.49844348e-01 -1.30298391e-01 7.25731730e-01 8.97282839e-01
9.08855379e-01 -3.30924898e-01 -2.63800025e-01 1.10068500e+00
2.50550091e-01 3.21715057e-01 7.00995862e-01 -6.98182464e-01
7.41117477e-01 9.82940555e-01 1.17951676e-01 -6.69769824e-01
-7.07251653e-02 -5.18606424e-01 -7.62263417e-01 -2.42398143e-01
4.45726722e-01 -2.53205508e-01 -6.80283844e-01 1.87481368e+00
4.50448602e-01 6.85869932e-01 2.90291935e-01 6.15742266e-01
4.31949258e-01 6.50641501e-01 2.80181229e-01 -1.65905386e-01
1.04920089e+00 -1.15631878e+00 -5.61949372e-01 -4.09215122e-01
8.41278970e-01 -1.29482880e-01 1.10111976e+00 4.28368837e-01
-5.50404787e-01 -5.57660043e-01 -1.12260258e+00 3.01308453e-01
-1.94566339e-01 3.48322332e-01 5.50553799e-01 6.52807534e-01
-4.32655841e-01 4.61241037e-01 -9.89170253e-01 -5.14019251e-01
8.89609277e-01 3.34224284e-01 -4.66355145e-01 -2.53578782e-01
-1.15505636e+00 4.45667326e-01 6.80472910e-01 -4.60909337e-01
-9.86441135e-01 -7.92960227e-01 -6.71983242e-01 5.08597605e-02
3.99576396e-01 -2.31250733e-01 1.22435224e+00 -1.21936321e+00
-1.35825145e+00 6.70002878e-01 -1.50119632e-01 -1.64121523e-01
4.12900656e-01 -2.53545731e-01 -6.74226582e-01 -9.19099823e-02
3.32444042e-01 3.65003854e-01 5.22199690e-01 -1.05075073e+00
-9.52056468e-01 -4.69729215e-01 -1.25714421e-01 4.44010735e-01
-7.94845879e-01 -3.57530147e-01 -4.83724326e-01 -7.03154206e-01
3.68624292e-02 -9.59349155e-01 3.30342501e-01 -9.30881575e-02
3.01535334e-03 -4.92878735e-01 9.31184292e-01 -5.44680119e-01
1.39532495e+00 -2.37308431e+00 6.72778115e-02 1.93809032e-01
-8.04421455e-02 2.95918345e-01 6.06875233e-02 2.07062975e-01
-1.41472593e-01 -3.02017361e-01 -4.66215968e-01 7.37730190e-02
-1.30059838e-01 3.93804073e-01 -9.37992632e-02 5.55073261e-01
2.09880635e-01 7.37272859e-01 -9.30062652e-01 -5.36159277e-01
-4.20743637e-02 1.81111068e-01 -5.01801431e-01 1.78671286e-01
1.21369958e-01 8.51082385e-01 -6.36870623e-01 6.81690454e-01
5.08535028e-01 -5.09473860e-01 4.84163374e-01 -1.06149852e-01
1.88395634e-01 4.60344762e-01 -1.25378764e+00 1.92718911e+00
-2.11401388e-01 3.92566323e-01 -3.55347902e-01 -1.38823140e+00
8.89367282e-01 1.84516266e-01 7.93887675e-01 -7.14617789e-01
-1.29086882e-01 1.98507741e-01 -1.05645381e-01 -2.08344594e-01
-2.32487231e-01 5.27510978e-02 -3.27246010e-01 5.14590621e-01
1.38521418e-01 7.00775325e-01 -6.49568811e-02 -9.53172892e-02
1.25694549e+00 3.65242928e-01 6.36736631e-01 -2.71817565e-01
7.50914991e-01 -8.03169385e-02 9.64441121e-01 4.74253833e-01
-5.52133501e-01 1.45869985e-01 3.31451774e-01 -1.80801436e-01
-4.75935668e-01 -1.27781546e+00 -5.72695732e-02 1.46137309e+00
4.19494480e-01 -1.21112570e-01 -7.86649168e-01 -1.22286415e+00
2.31681801e-02 4.44488734e-01 -7.09958673e-01 -4.73414421e-01
-4.50804532e-01 -8.09101224e-01 7.48803318e-01 9.09771144e-01
1.08134735e+00 -9.53339517e-01 -1.91932723e-01 8.65629315e-02
-4.62745875e-01 -1.02309704e+00 -4.58567202e-01 2.42194057e-01
-9.84729588e-01 -1.08537745e+00 -6.21303022e-01 -9.09074306e-01
5.78396857e-01 3.31669122e-01 7.69850910e-01 -4.60386544e-01
3.16535443e-01 3.18818390e-01 -4.99886692e-01 -8.83709490e-02
5.49998358e-02 4.70148921e-01 1.94079906e-01 4.00186539e-01
7.67635584e-01 -8.34407866e-01 -6.76822126e-01 7.79387712e-01
-6.42111421e-01 -1.06887005e-01 7.19716787e-01 7.51430154e-01
5.87854385e-01 1.98240355e-01 8.10285151e-01 -9.42096949e-01
5.25367200e-01 -9.46667254e-01 -1.71696752e-01 4.00163740e-01
-6.99435830e-01 6.48564473e-02 3.96272838e-01 -8.19517493e-01
-1.40781641e+00 2.20768586e-01 3.54025573e-01 -1.84836030e-01
-2.70045936e-01 4.30274665e-01 -6.31718993e-01 3.39647755e-02
7.33038306e-01 5.61862171e-01 -4.12833035e-01 -5.90991497e-01
-1.09405043e-02 1.02037132e+00 4.82091516e-01 -8.52019548e-01
7.40401685e-01 5.33642054e-01 -4.83515233e-01 -4.53383178e-01
-1.11219740e+00 -7.16176033e-01 -7.06127346e-01 -1.08553618e-01
7.23633409e-01 -1.20766521e+00 -2.98142254e-01 7.15050757e-01
-6.44226491e-01 -5.11204064e-01 -1.27085924e-01 5.13597369e-01
-4.29377377e-01 4.54585075e-01 -3.13910455e-01 -5.32865286e-01
-9.49617922e-02 -8.01358938e-01 8.74305725e-01 2.50232100e-01
-2.99506485e-01 -1.01917756e+00 2.87649781e-01 6.73078299e-01
-1.58652011e-02 9.36580822e-03 7.08458483e-01 -1.02353561e+00
-2.17779696e-01 -1.13228895e-02 -2.13203028e-01 3.81142974e-01
5.96836984e-01 -6.50393784e-01 -1.02025437e+00 -3.82935703e-01
-3.11618805e-01 -5.14909387e-01 5.93914688e-01 -8.32097456e-02
1.34308922e+00 -6.46985024e-02 -7.14769602e-01 3.75875741e-01
9.14030135e-01 5.99178135e-01 6.39913797e-01 4.32134807e-01
7.30518997e-01 4.47211295e-01 1.09332454e+00 6.69775367e-01
4.10438955e-01 7.44546831e-01 -1.08657941e-01 6.94604591e-02
-5.32473139e-02 -3.95324200e-01 5.88395715e-01 7.42793858e-01
5.30222319e-02 -4.31551188e-02 -1.00534391e+00 5.96321225e-01
-1.98199749e+00 -8.38139236e-01 2.66697943e-01 2.14197302e+00
1.04622126e+00 9.98106152e-02 2.27992192e-01 3.26609641e-01
7.25385964e-01 4.54779193e-02 -8.51899207e-01 2.33332798e-01
1.61898986e-01 2.08061367e-01 3.91023755e-01 1.20734252e-01
-1.45846319e+00 7.42833376e-01 5.82123518e+00 1.06910336e+00
-9.03933942e-01 3.99853617e-01 4.81461227e-01 2.58072298e-02
1.94254979e-01 -1.60828099e-01 -8.07129800e-01 8.03475380e-01
7.36966729e-01 -1.25003994e-01 3.47492754e-01 1.11470723e+00
-1.04717351e-01 -1.35265633e-01 -1.30293465e+00 1.18686342e+00
1.05371118e-01 -8.48586917e-01 -1.20207027e-01 9.50230137e-02
8.60902011e-01 -8.09645578e-02 -2.20013425e-01 8.34426403e-01
3.31004530e-01 -6.14692092e-01 3.80538821e-01 1.97551429e-01
7.39207089e-01 -6.67890847e-01 4.49886024e-01 6.22949004e-01
-1.45676374e+00 -2.72584230e-01 -1.51538011e-02 -2.05582500e-01
-2.25339592e-01 3.15195084e-01 -7.02604592e-01 5.61337054e-01
9.00194705e-01 1.19462299e+00 -5.45063198e-01 8.52570474e-01
-2.05625385e-01 7.99966037e-01 -1.07627220e-01 2.14623004e-01
8.37122872e-02 -2.71414906e-01 -4.24176222e-03 1.15944350e+00
3.36843878e-01 7.67548531e-02 5.02922654e-01 4.53347892e-01
-3.40585023e-01 -5.33216894e-02 -3.37291092e-01 -2.81645525e-02
6.24053955e-01 8.81380200e-01 -4.54170525e-01 -4.75196570e-01
-6.56786084e-01 1.22227192e+00 5.29271722e-01 3.64134103e-01
-1.15087175e+00 -2.89358139e-01 8.08212280e-01 1.12430826e-01
2.20348895e-01 -9.08396691e-02 -2.21440271e-01 -1.36977029e+00
2.28471428e-01 -8.06597173e-01 7.89282739e-01 -4.49814051e-01
-1.42847264e+00 1.56038940e-01 1.65370986e-01 -1.62684941e+00
7.01188222e-02 -4.60527748e-01 -3.24057966e-01 5.07320642e-01
-1.48533380e+00 -1.01793659e+00 -6.99705184e-01 9.25975740e-01
5.45014977e-01 -9.14307013e-02 6.52940154e-01 7.76995480e-01
-7.35540569e-01 7.08659947e-01 2.55969852e-01 4.35317069e-01
1.03082943e+00 -1.06216514e+00 -6.24024076e-04 6.38240159e-01
5.59031405e-02 3.92760664e-01 -9.76930931e-03 -6.10744596e-01
-9.21239138e-01 -1.40763199e+00 4.67510313e-01 -4.53243822e-01
5.80894768e-01 -3.38624924e-01 -1.14189768e+00 9.40262675e-01
-1.21652320e-01 2.53276259e-01 1.21812201e+00 4.94777597e-02
-5.56127548e-01 -4.17082459e-01 -1.07303548e+00 3.13205123e-01
1.36868274e+00 -5.66532314e-01 -8.04760456e-01 1.45960853e-01
3.37473869e-01 -1.71268314e-01 -1.10552073e+00 4.86821383e-01
6.65473104e-01 -6.96623325e-01 8.97688985e-01 -5.84590375e-01
1.81696698e-01 -4.41569120e-01 -2.71552086e-01 -1.23091137e+00
-3.98204893e-01 -1.09855449e-02 -3.04640353e-01 1.35167456e+00
2.32755989e-01 -7.64428616e-01 1.00500262e+00 4.19848740e-01
-4.55372483e-02 -4.81337130e-01 -9.61122930e-01 -1.08080673e+00
-2.53726859e-02 -3.11861306e-01 5.16935408e-01 1.41330850e+00
2.49274701e-01 4.65433657e-01 -3.39887649e-01 1.60512999e-01
5.17302513e-01 2.14265645e-01 7.75238156e-01 -1.44982648e+00
-4.31813806e-01 -6.30446002e-02 -2.92789310e-01 -1.26616299e+00
3.85061264e-01 -9.06326711e-01 -9.19573754e-02 -1.24432266e+00
5.02555192e-01 -6.08972192e-01 -7.67741978e-01 8.05373311e-01
-1.79766744e-01 3.01514804e-01 -2.53401041e-01 5.27932525e-01
-1.07185054e+00 6.65464044e-01 1.04079819e+00 -2.40655676e-01
-5.00569940e-01 -4.48469780e-02 -7.79788911e-01 7.17142463e-01
9.47784007e-01 -5.51681161e-01 -7.08340883e-01 -2.46863499e-01
-3.50330442e-01 -2.18150437e-01 1.65701061e-02 -1.37353063e+00
1.00576170e-01 -3.70084882e-01 6.46046996e-01 -4.01993841e-01
2.60474145e-01 -9.50333655e-01 1.87390372e-01 4.32182908e-01
-2.82249629e-01 -6.13550365e-01 6.04567267e-02 9.13378716e-01
-2.47888997e-01 7.54367039e-02 7.75827229e-01 1.94190994e-01
-9.37062800e-01 3.14806044e-01 -3.18256289e-01 1.50065809e-01
1.46803582e+00 -3.90583605e-01 -1.86088249e-01 -6.40604049e-02
-6.99544668e-01 1.81810766e-01 3.63683581e-01 4.78654385e-01
1.57521054e-01 -1.63877559e+00 -2.78348744e-01 1.39627323e-01
6.21617913e-01 -1.19621214e-02 1.55218422e-01 9.11195695e-01
-4.15245891e-02 2.81155705e-01 -3.60346437e-01 -7.89885163e-01
-1.21063828e+00 4.46890920e-01 3.43271434e-01 -5.25501788e-01
-2.47319385e-01 5.24685025e-01 4.64837343e-01 -6.16466522e-01
1.81320697e-01 -1.94740519e-01 -1.98853686e-01 9.69765782e-02
4.27865386e-01 5.40350735e-01 4.57139611e-02 -4.52909648e-01
-7.69283891e-01 5.06679356e-01 -1.09305128e-01 5.83717227e-02
1.10379791e+00 -1.67493280e-02 2.63055712e-01 4.20604050e-01
1.11548316e+00 -1.54270098e-01 -1.57398212e+00 -6.41956329e-01
3.83219719e-01 -4.98162448e-01 -3.03075522e-01 -8.98058653e-01
-8.97261381e-01 7.14175999e-01 7.68586338e-01 -3.37975413e-01
1.44624019e+00 1.38563231e-01 5.78747809e-01 3.87186736e-01
4.85992581e-01 -1.17957199e+00 5.42645752e-01 2.59480298e-01
4.10594225e-01 -1.24530494e+00 -2.95735616e-02 -5.58682263e-01
-8.24094653e-01 5.94258726e-01 9.91928935e-01 5.65333106e-02
4.88151610e-01 -3.73626724e-02 -2.52177924e-01 1.06498465e-01
-4.70155835e-01 4.83585987e-03 1.90305382e-01 8.05064201e-01
3.31253856e-01 1.85585588e-01 -2.18957245e-01 9.16842759e-01
3.98067176e-01 3.14398289e-01 1.52485436e-02 1.22597885e+00
-3.97315174e-01 -1.33191633e+00 -3.39977682e-01 4.23342079e-01
-2.56656379e-01 2.75749326e-01 -4.97635096e-01 6.19219363e-01
3.46449435e-01 9.88161027e-01 -7.78793618e-02 -4.65582073e-01
4.41850841e-01 3.88697982e-01 4.62715358e-01 -6.68560684e-01
-2.93364614e-01 -5.73216826e-02 8.12881514e-02 -5.15413165e-01
-7.29468763e-01 -8.53938401e-01 -1.26426530e+00 1.20139541e-02
-8.05391222e-02 1.95070326e-01 3.69546227e-02 9.71478879e-01
5.15056968e-01 4.59660590e-01 6.02190971e-01 -4.01746333e-01
-4.69450593e-01 -1.01849771e+00 -6.32078290e-01 6.97528958e-01
-3.96595784e-02 -1.14313197e+00 -1.14697821e-01 2.23770365e-01] | [8.068449020385742, 1.044541835784912] |
7974a011-580b-4ed5-821c-97d3ba3e4e96 | cif-pt-bridging-speech-and-text | 2305.17499 | null | https://arxiv.org/abs/2305.17499v1 | https://arxiv.org/pdf/2305.17499v1.pdf | CIF-PT: Bridging Speech and Text Representations for Spoken Language Understanding via Continuous Integrate-and-Fire Pre-Training | Speech or text representation generated by pre-trained models contains modal-specific information that could be combined for benefiting spoken language understanding (SLU) tasks. In this work, we propose a novel pre-training paradigm termed Continuous Integrate-and-Fire Pre-Training (CIF-PT). It relies on a simple but effective frame-to-token alignment: continuous integrate-and-fire (CIF) to bridge the representations between speech and text. It jointly performs speech-to-text training and language model distillation through CIF as the pre-training (PT). Evaluated on SLU benchmark SLURP dataset, CIF-PT outperforms the state-of-the-art model by 1.94% of accuracy and 2.71% of SLU-F1 on the tasks of intent classification and slot filling, respectively. We also observe the cross-modal representation extracted by CIF-PT obtains better performance than other neural interfaces for the tasks of SLU, including the dominant speech representation learned from self-supervised pre-training. | ['Zejun Ma', 'Lu Lu', 'Jun Zhang', 'Peihao Wu', 'Zhecheng An', 'Linhao Dong'] | 2023-05-27 | null | null | null | null | ['spoken-language-understanding', 'intent-classification', 'slot-filling', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 4.51295376e-01 6.55398905e-01 -3.06014091e-01 -5.59655070e-01
-1.18877137e+00 -2.04933971e-01 7.71466494e-01 5.50709590e-02
-3.87087703e-01 6.03283703e-01 5.60580611e-01 -4.56248552e-01
5.12937725e-01 -4.97960865e-01 -7.98031926e-01 -3.84200126e-01
3.62999231e-01 6.65333152e-01 3.49678695e-02 -4.09265935e-01
-2.67942864e-02 -4.22572792e-01 -1.69883299e+00 8.94359827e-01
1.02751637e+00 1.01058388e+00 6.23687685e-01 5.89448154e-01
-7.01998830e-01 8.41511488e-01 -6.03200197e-01 3.30795288e-01
-3.93919110e-01 -5.45822740e-01 -1.12947702e+00 -1.88660026e-01
2.24865541e-01 -2.54362792e-01 -2.55210221e-01 5.50511360e-01
4.94401932e-01 3.41441691e-01 7.00543940e-01 -9.89441633e-01
-5.99120080e-01 9.84863997e-01 -2.56583780e-01 5.71661927e-02
6.13278091e-01 -2.04503551e-01 1.01468575e+00 -1.18289864e+00
8.17157686e-01 1.49866498e+00 2.73027629e-01 9.55094576e-01
-1.17130661e+00 -5.67985594e-01 2.84843355e-01 1.95212901e-01
-1.32147670e+00 -1.09205687e+00 5.39294481e-01 -1.82429060e-01
1.85340929e+00 2.54675299e-01 9.41895023e-02 1.13198948e+00
-1.63147628e-01 1.47649038e+00 9.30801332e-01 -9.09321725e-01
2.72189289e-01 -2.02450887e-04 3.39288831e-01 4.66620237e-01
-6.00207090e-01 7.65143558e-02 -1.11000097e+00 1.11257523e-01
3.19084704e-01 -5.16479373e-01 -3.44934851e-01 3.99472982e-01
-1.00781119e+00 6.53323352e-01 1.62160099e-02 5.10565639e-01
-1.97095886e-01 -1.38960063e-01 6.53992176e-01 2.83329368e-01
7.11853743e-01 1.46333575e-01 -6.98850453e-01 -3.27840120e-01
-1.01709676e+00 -5.19024543e-02 6.94889307e-01 8.72090995e-01
6.71114564e-01 3.25978637e-01 -5.44490993e-01 1.29680920e+00
4.91432399e-01 4.83304292e-01 7.47042894e-01 -6.22514009e-01
6.16766036e-01 2.82912076e-01 -2.89167494e-01 1.86419897e-02
-3.26913446e-01 -2.59823382e-01 -6.51119053e-01 -2.24189758e-01
3.58575694e-02 -7.26261660e-02 -1.14462900e+00 1.88651121e+00
1.47021741e-01 3.29639256e-01 6.57825947e-01 3.93564522e-01
1.30482769e+00 1.25881624e+00 3.16855013e-01 -2.63081759e-01
1.47241521e+00 -1.24793422e+00 -1.17918789e+00 -7.09724069e-01
9.52871084e-01 -7.21175134e-01 1.25419009e+00 -4.48047853e-04
-9.82626796e-01 -7.73344696e-01 -9.75899756e-01 -1.65600777e-01
-3.51741761e-01 4.95438069e-01 7.87606910e-02 3.45507562e-01
-1.11670601e+00 1.86889425e-01 -6.43441558e-01 -5.22726238e-01
1.46179110e-01 1.53266937e-01 -4.38220441e-01 8.52397084e-02
-1.47912729e+00 8.42204034e-01 6.22911394e-01 -1.78690612e-01
-8.95512104e-01 -7.14609325e-01 -1.29293180e+00 1.97898686e-01
4.20338303e-01 -5.02550423e-01 1.72841311e+00 -5.88930964e-01
-2.02842307e+00 9.96015370e-01 -7.59005010e-01 -6.87369347e-01
-4.26507555e-02 -4.32819843e-01 -3.04983884e-01 -1.93462726e-02
1.54248744e-01 1.09084046e+00 8.07654619e-01 -1.17755508e+00
-6.85935438e-01 -1.97337911e-01 -2.78392822e-01 4.01644439e-01
-3.53987440e-02 -7.98555743e-03 -1.02057248e-01 -5.41448891e-01
1.41055182e-01 -5.79933405e-01 3.55804056e-01 -5.60720623e-01
-3.77708644e-01 -8.91116858e-01 1.18416059e+00 -9.86843109e-01
1.55025852e+00 -2.22772980e+00 9.69856977e-02 -4.22574371e-01
-2.84144521e-01 6.70342863e-01 -4.33455020e-01 7.43847132e-01
-5.35815246e-02 -7.12717324e-02 -2.05510050e-01 -1.16323507e+00
2.28132784e-01 4.97544140e-01 -5.80826402e-01 -1.11420773e-01
5.19940078e-01 9.94662225e-01 -7.13169158e-01 -3.77161533e-01
4.52161789e-01 4.25705671e-01 -4.32201624e-01 5.04875481e-01
-6.03449941e-01 5.66762745e-01 1.05098430e-02 6.03779614e-01
4.57457513e-01 9.61748138e-02 1.90924585e-01 -1.62195623e-01
-1.39986590e-01 1.17903626e+00 -6.69220209e-01 2.15116835e+00
-7.92035460e-01 6.60672426e-01 1.24124423e-01 -1.03990185e+00
1.07519388e+00 7.63155699e-01 8.91273189e-03 -1.06435227e+00
1.79783195e-01 3.39447856e-01 -1.25622123e-01 -4.16932702e-01
5.61276674e-01 -1.52791351e-01 -1.68640763e-01 5.77240109e-01
7.77791083e-01 -9.37311873e-02 1.39064774e-01 1.86654046e-01
8.29249620e-01 1.54369295e-01 4.80208009e-01 -2.05262989e-01
7.02463925e-01 -3.72681439e-01 2.77820706e-01 7.07083166e-01
-2.23942310e-01 5.09093523e-01 1.42047703e-01 1.47342801e-01
-7.57074893e-01 -8.98972452e-01 -2.80875117e-01 1.66172850e+00
-2.92339444e-01 -6.71557844e-01 -9.77044284e-01 -6.27568662e-01
-3.20811987e-01 1.31378031e+00 -4.15966958e-01 -1.40046373e-01
-4.76160705e-01 -2.12866977e-01 7.13749349e-01 5.02422273e-01
5.62621057e-01 -1.39027739e+00 -9.93766412e-02 3.50067139e-01
-6.27487123e-01 -1.46747327e+00 -4.54310417e-01 4.16944921e-01
-5.52305996e-01 -5.54280639e-01 -6.82679951e-01 -1.08284557e+00
3.84532869e-01 1.21109053e-01 1.05740464e+00 -2.24987835e-01
3.10150355e-01 1.83482081e-01 -7.15823948e-01 -3.32355171e-01
-7.93803632e-01 3.24833572e-01 -2.66060717e-02 -3.19105051e-02
3.96418780e-01 -4.04711962e-01 5.55867366e-02 1.05846658e-01
-6.80216014e-01 5.05682051e-01 4.17737395e-01 1.05205309e+00
5.82814217e-01 -3.68980855e-01 7.64040709e-01 -5.63145578e-01
5.92990696e-01 -2.30202019e-01 -5.41612245e-02 3.34595531e-01
-1.44803345e-01 2.20953524e-01 3.90361458e-01 -3.07349592e-01
-1.26435649e+00 -6.99778423e-02 -6.75564229e-01 -2.97757030e-01
-4.32933271e-01 6.42501771e-01 -3.26726645e-01 4.00384009e-01
3.98849636e-01 5.60641706e-01 2.81443191e-03 -6.12056315e-01
4.85669971e-01 1.15070808e+00 6.84328258e-01 -5.75109780e-01
8.87803640e-03 6.70125559e-02 -7.51659036e-01 -1.32084870e+00
-1.05730665e+00 -5.31672239e-01 -5.78575671e-01 -6.40242454e-03
9.82486844e-01 -9.65361714e-01 -3.99078012e-01 6.08671963e-01
-1.62872684e+00 -6.62612915e-01 -3.49704921e-01 3.22649777e-01
-5.12537003e-01 1.83902636e-01 -6.43767953e-01 -1.24749815e+00
-7.42922604e-01 -9.81176138e-01 1.36414981e+00 2.71276087e-01
-4.75438625e-01 -8.41641605e-01 1.78404555e-01 5.18972874e-01
4.43785220e-01 -5.51011324e-01 8.41035426e-01 -1.02149189e+00
-1.13750637e-01 2.70471334e-01 -1.48010045e-01 4.65276599e-01
6.42177314e-02 -2.37058491e-01 -1.54843068e+00 -2.36526355e-01
-1.42138153e-01 -6.83236003e-01 9.93674755e-01 5.33577323e-01
6.58066571e-01 -3.66602808e-01 -2.63373733e-01 3.58594537e-01
7.77126968e-01 4.59841967e-01 7.53641248e-01 1.95537508e-01
3.31459165e-01 7.08353162e-01 6.37684345e-01 1.87169671e-01
7.37370849e-01 7.36778557e-01 -2.58389954e-02 -7.34658539e-02
-4.23151314e-01 -5.97479284e-01 7.29600608e-01 1.17037404e+00
3.57577652e-01 -1.85345903e-01 -9.18932855e-01 5.74272096e-01
-2.09066820e+00 -6.21798873e-01 2.29800314e-01 1.93677616e+00
1.05663133e+00 2.11907327e-01 -9.44631919e-02 2.22166523e-01
5.01709878e-01 3.50404650e-01 -2.11345375e-01 -4.13582861e-01
-1.83966413e-01 3.95726770e-01 -1.44435361e-01 8.63009274e-01
-1.13786221e+00 1.49917138e+00 5.52678633e+00 1.22053659e+00
-1.19079363e+00 4.25413787e-01 6.37117028e-01 3.83834571e-01
-2.59759307e-01 -1.03625126e-01 -1.22774947e+00 1.76507443e-01
1.36120069e+00 5.42728566e-02 2.08824471e-01 7.91947842e-01
1.52568385e-01 -3.21520120e-01 -1.01863289e+00 9.09316897e-01
1.00349225e-01 -1.37168288e+00 1.91043586e-01 -3.34467649e-01
5.81614256e-01 1.23502091e-01 -1.20876767e-01 8.96965265e-01
4.36914414e-02 -9.87351358e-01 8.23199153e-01 1.78598806e-01
1.02716243e+00 -4.60442960e-01 8.69109869e-01 5.79392493e-01
-1.40720379e+00 3.52635652e-01 -1.72339842e-01 -9.65399519e-02
5.22610664e-01 2.23281637e-01 -1.18674898e+00 5.66724956e-01
3.97119790e-01 5.14244735e-01 -1.72214583e-02 4.38762814e-01
-3.87554407e-01 8.98006618e-01 -4.17862147e-01 5.30888289e-02
3.67791384e-01 1.70176744e-01 4.82616991e-01 1.42284679e+00
2.88878292e-01 -8.17742646e-02 2.25692675e-01 8.41082692e-01
7.76446089e-02 2.51332492e-01 -4.99602705e-01 -4.35347766e-01
6.98307455e-01 8.05322587e-01 -4.61179018e-01 -6.10143125e-01
-3.63901645e-01 9.66834724e-01 3.94318223e-01 3.86256754e-01
-4.92993951e-01 -1.99715674e-01 6.78990662e-01 -1.58887506e-01
4.25816536e-01 -3.12322021e-01 -1.67169735e-01 -1.03402042e+00
-1.57858148e-01 -7.04538465e-01 2.55987465e-01 -8.39106798e-01
-9.82724071e-01 1.15820765e+00 -3.76386428e-03 -9.52545047e-01
-7.86139131e-01 -5.48066616e-01 -6.50776505e-01 9.90477502e-01
-1.67664587e+00 -1.58843184e+00 1.05929159e-01 4.19400126e-01
1.07346177e+00 -3.93011004e-01 1.20701265e+00 1.23556830e-01
-5.93003988e-01 6.53885186e-01 -1.21627115e-01 2.09385633e-01
5.18188834e-01 -8.94213974e-01 6.63756073e-01 7.34402716e-01
4.21512455e-01 4.00123656e-01 5.37870228e-01 -5.26456237e-01
-8.51442277e-01 -9.00008082e-01 1.40280414e+00 -1.28599033e-01
4.81100798e-01 -5.67702770e-01 -1.06819880e+00 6.83510125e-01
5.92544258e-01 -3.52176100e-01 6.98662400e-01 1.53812513e-01
-3.03594172e-01 1.67557627e-01 -7.92361200e-01 3.12390715e-01
8.80740941e-01 -1.15835798e+00 -9.49129879e-01 1.11058816e-01
1.16991270e+00 -5.90953767e-01 -5.14494956e-01 4.77159709e-01
2.19929218e-01 -7.83537984e-01 7.94497907e-01 -5.61277986e-01
3.31666082e-01 -1.39312431e-01 -4.32192326e-01 -1.29530632e+00
1.51798800e-01 -7.45108843e-01 -2.64178544e-01 1.55802023e+00
6.01215065e-01 -4.77325112e-01 3.64209861e-01 -8.41240659e-02
-7.05426991e-01 -7.60203004e-01 -1.29521716e+00 -5.56138813e-01
4.04868759e-02 -7.84074783e-01 3.11956137e-01 7.60164440e-01
3.72456312e-01 1.00356114e+00 -2.27381065e-01 -7.54582658e-02
1.90121308e-01 -2.27264807e-01 5.95586777e-01 -8.61617804e-01
-2.79919684e-01 -2.94363737e-01 1.89186722e-01 -1.45398903e+00
6.80220425e-01 -1.08889878e+00 3.87069166e-01 -1.74557173e+00
-3.06513548e-01 -3.44960868e-01 -1.48133337e-01 9.15855467e-01
-1.33285508e-01 -2.77659386e-01 3.17937136e-01 1.41334757e-01
-5.50608754e-01 1.07543933e+00 1.13374066e+00 -2.65784770e-01
-5.16333461e-01 -9.11822021e-02 -2.67651290e-01 3.47578347e-01
6.44360542e-01 -2.04632878e-01 -3.72814655e-01 -2.81311572e-01
-4.30933237e-01 3.26720625e-01 -2.03463748e-01 -8.96352291e-01
1.83425650e-01 1.99461043e-01 -2.89103925e-01 -1.00986028e+00
5.93589425e-01 -2.78644294e-01 -5.03371298e-01 2.58324772e-01
-3.91196042e-01 -4.47399646e-01 6.42336249e-01 2.79105492e-02
-6.66220248e-01 -4.21947926e-01 6.33057773e-01 5.30845411e-02
-1.07543314e+00 -3.25729698e-01 -7.15371013e-01 1.32529020e-01
5.84102392e-01 6.85201138e-02 -6.01188838e-01 -6.76685393e-01
-8.60758066e-01 1.32071540e-01 -2.69551158e-01 5.01213729e-01
5.72938561e-01 -1.17169678e+00 -6.45284176e-01 6.67822063e-01
2.63843030e-01 1.33089736e-01 4.79375243e-01 8.19990933e-01
-1.03017529e-02 1.07179511e+00 -2.55164993e-03 -7.78365314e-01
-1.25671101e+00 8.87752548e-02 2.63699174e-01 -5.51822722e-01
-3.64117354e-01 1.11761045e+00 4.32226002e-01 -9.71673846e-01
4.96324897e-01 -5.82703531e-01 -1.79448798e-01 8.17996487e-02
7.41721511e-01 1.88972279e-02 2.25356296e-01 -8.26854587e-01
-4.52808917e-01 1.77010059e-01 -1.86912894e-01 -4.80763614e-01
1.19000804e+00 -3.50538820e-01 1.14245199e-01 6.96336389e-01
1.04970133e+00 -2.58785665e-01 -1.14534032e+00 -5.28855801e-01
4.53256350e-03 2.58756518e-01 2.14295909e-01 -9.34145749e-01
-5.39914966e-01 1.20235288e+00 3.52123886e-01 6.18936568e-02
8.24128747e-01 2.46564716e-01 1.07903647e+00 4.08111215e-01
2.98010349e-01 -1.20507824e+00 1.54324234e-01 1.29145074e+00
1.11553967e+00 -1.13941181e+00 -7.28433788e-01 -4.53289658e-01
-8.42902064e-01 8.86442542e-01 7.87438869e-01 1.83707729e-01
6.40720308e-01 4.74429160e-01 3.06744903e-01 3.15549940e-01
-1.03367996e+00 -6.15930021e-01 5.50520658e-01 6.79053545e-01
6.95544362e-01 5.42494059e-02 -1.73302099e-01 8.45040441e-01
-3.62617701e-01 3.10562714e-03 -1.17872059e-02 1.06771898e+00
-6.27548933e-01 -1.25110281e+00 -2.66009927e-01 1.64281711e-01
1.83075637e-01 -4.22060281e-01 -2.73058206e-01 6.01735771e-01
-3.81900370e-02 1.03801632e+00 4.02807266e-01 -5.17330527e-01
1.51280448e-01 7.83138275e-01 3.26133132e-01 -9.37309146e-01
-4.71735865e-01 6.01843834e-01 5.05110204e-01 -4.04396117e-01
-3.52941483e-01 -4.05616939e-01 -1.64382243e+00 1.06811590e-01
-3.23682159e-01 1.78765252e-01 5.94152510e-01 1.45849550e+00
5.12849152e-01 9.42566752e-01 2.47043982e-01 -1.08640087e+00
-2.76569039e-01 -1.61706126e+00 -3.71091753e-01 -9.12513025e-03
4.18818086e-01 -6.98601723e-01 -2.78500348e-01 3.32129374e-02] | [14.025487899780273, 6.988629341125488] |
82dbe740-6a55-4e98-a919-a94368188e46 | blended-grammar-network-for-human-parsing | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4548_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123690188.pdf | Blended Grammar Network for Human Parsing | Although human parsing has made great progress, it still faces a challenge, i.e., how to extract the whole foreground from similar or cluttered scenes effectively. In this paper, we propose a Blended Grammar Network (BGNet), to deal with the challenge. BGNet exploits the inherent hierarchical structure of a human body and the relationship of different human parts by means of grammar rules in both cascaded and paralleled manner. In this way, conspicuous parts, which are easily distinguished from the background, can amend the segmentation of inconspicuous ones, improving the foreground extraction. We also design a Part-aware Convolutional Recurrent Neural Network (PCRNN) to pass messages which are generated by grammar rules. To train PCRNNs effectively, we present a blended grammar loss to supervise the training of PCRNNs. We conduct extensive experiments to evaluate BGNet on PASCAL-Person-Part, LIP, and PPSS datasets. BGNet obtains state-of-the-art performance on these human parsing datasets. | ['Ming Tang', 'Yingying Chen', 'Xiaomei Zhang', 'Bingke Zhu', 'Jinqiao Wang'] | null | null | null | null | eccv-2020-8 | ['human-parsing'] | ['computer-vision'] | [ 4.22752082e-01 3.95379514e-01 1.84334502e-01 -5.56043267e-01
-4.19227600e-01 -3.04182500e-01 2.78491676e-01 -3.80948663e-01
-3.11264783e-01 3.90941769e-01 -2.48263660e-03 -1.44829288e-01
4.83338058e-01 -7.45339394e-01 -7.76180863e-01 -6.65527642e-01
3.29958111e-01 2.47596845e-01 8.61320913e-01 -3.88641171e-02
-1.97744071e-01 2.08215609e-01 -1.29512274e+00 5.47921896e-01
9.01226759e-01 8.04091215e-01 3.83895665e-01 5.65947473e-01
-3.42535853e-01 9.88168478e-01 -6.91619515e-01 -7.52925575e-01
2.11176649e-01 -6.33818686e-01 -7.65343368e-01 2.90688634e-01
5.55413723e-01 -3.62526596e-01 -3.66722941e-01 1.18604779e+00
3.67673248e-01 -1.11693308e-01 2.70188808e-01 -9.17631865e-01
-3.96389812e-01 7.26698220e-01 -8.71420205e-01 2.05861956e-01
1.26771480e-02 1.51385024e-01 8.95047605e-01 -4.19917285e-01
5.79980135e-01 1.66858053e+00 4.93201315e-01 8.60673070e-01
-8.32910478e-01 -6.52558863e-01 7.01635063e-01 -8.49285871e-02
-9.18610215e-01 -1.91623822e-01 1.01016486e+00 -1.67473238e-02
5.30551195e-01 -1.62199140e-05 7.10615277e-01 1.17177689e+00
2.17657939e-01 1.47825444e+00 8.25300336e-01 -2.20957339e-01
-3.17197181e-02 -4.21439946e-01 3.51183355e-01 9.89216089e-01
2.32069507e-01 -1.23727649e-01 -2.82598346e-01 2.36277819e-01
7.80207574e-01 -1.00588478e-01 -1.09889857e-01 -3.60392362e-01
-7.74269223e-01 6.73264444e-01 4.13659155e-01 4.65950407e-02
-3.25666070e-01 1.23181380e-01 4.34218377e-01 -3.01091790e-01
2.66851932e-01 -7.42243603e-02 -5.45622587e-01 2.26328760e-01
-7.14844167e-01 1.54464662e-01 7.15670228e-01 1.00820374e+00
7.02328324e-01 -7.18495473e-02 -1.82676792e-01 9.91669118e-01
5.64382493e-01 4.44928110e-01 1.97691545e-01 -1.01090300e+00
5.71423769e-01 7.91916132e-01 -4.36284661e-01 -8.69060397e-01
-4.64657575e-01 -4.35775608e-01 -8.31354737e-01 8.14354867e-02
6.51705086e-01 -4.00925994e-01 -1.28556752e+00 1.75844300e+00
6.90455735e-01 2.60424893e-02 5.42540736e-02 8.86338770e-01
1.07371604e+00 6.29839838e-01 4.61697727e-01 1.17105193e-01
1.65882564e+00 -1.23691356e+00 -7.30067194e-01 -6.60815835e-01
4.20915008e-01 -5.76876402e-01 9.71090674e-01 3.65184635e-01
-1.03597987e+00 -6.76580787e-01 -8.27937782e-01 -2.75647134e-01
-8.68576095e-02 2.28772059e-01 7.59039223e-01 5.55813193e-01
-7.05052614e-01 5.15807569e-01 -1.04692233e+00 -2.40924597e-01
7.01264739e-01 3.05384040e-01 -1.57665029e-01 -5.06288372e-02
-1.06513047e+00 3.10629368e-01 5.92061102e-01 5.22068024e-01
-6.14099860e-01 -2.40488276e-01 -1.09619725e+00 -3.42479311e-02
7.64154673e-01 -6.42247558e-01 1.26574385e+00 -1.27860212e+00
-1.46776628e+00 9.26546812e-01 4.98049408e-02 -3.73095930e-01
5.57257116e-01 -5.15508473e-01 -2.62389064e-01 2.83135384e-01
2.11706489e-01 1.02635074e+00 8.89755905e-01 -1.14628816e+00
-9.13168490e-01 -4.55903649e-01 3.77320983e-02 8.31252187e-02
2.86688387e-01 3.03646505e-01 -1.03210557e+00 -9.30940330e-01
4.03876364e-01 -8.67752194e-01 -2.85888404e-01 -6.13324828e-02
-7.73363471e-01 -4.64884967e-01 1.05236471e+00 -9.64063942e-01
1.02103341e+00 -2.39847612e+00 7.71125406e-02 -1.76794827e-01
2.62592137e-01 5.06633759e-01 -2.73112953e-01 -2.37495273e-01
2.77775288e-01 4.48826654e-03 -5.02914011e-01 -4.00951833e-01
-1.31493732e-01 7.08369493e-01 1.95085257e-03 8.55288506e-02
6.46188498e-01 8.79805863e-01 -8.74696970e-01 -9.59595561e-01
3.75612639e-02 3.61167431e-01 -4.60763872e-01 3.10265392e-01
-4.44362283e-01 3.94739270e-01 -6.46495104e-01 9.53404486e-01
8.04705679e-01 -1.62960991e-01 3.31816226e-01 -1.02090612e-01
2.07108527e-01 1.44444525e-01 -1.08704281e+00 1.58366799e+00
1.96171962e-02 2.10809916e-01 3.67523640e-01 -7.65524268e-01
7.93159485e-01 -1.10158958e-01 1.40337076e-03 -7.17053950e-01
3.18436891e-01 2.57990905e-03 2.52655834e-01 -4.07246798e-01
2.18855560e-01 1.20495029e-01 -8.04496706e-02 1.21651217e-01
9.98391286e-02 3.48593503e-01 3.36071163e-01 1.67095557e-01
1.06301570e+00 5.46319246e-01 1.63280427e-01 -2.25542918e-01
5.63050985e-01 -3.50053340e-01 1.19733429e+00 6.83661103e-01
-5.92599034e-01 7.67133296e-01 8.56825590e-01 -5.81244349e-01
-5.38701892e-01 -1.08229554e+00 2.14225680e-01 1.20325255e+00
3.37703109e-01 -2.05405653e-01 -1.14622569e+00 -1.31710768e+00
-3.47803921e-01 3.46012414e-01 -6.59254909e-01 5.26218787e-02
-1.13531506e+00 -8.62901270e-01 5.26239753e-01 8.28972161e-01
1.10356688e+00 -1.59468782e+00 -8.28938246e-01 4.53275353e-01
-3.40591490e-01 -1.51077163e+00 -5.21607220e-01 2.09579870e-01
-7.88423300e-01 -1.29718733e+00 -6.34409487e-01 -9.84629095e-01
5.50723970e-01 1.40684634e-01 1.20369864e+00 2.15506330e-01
-2.93233395e-01 -1.41238615e-01 -4.69322145e-01 -5.32666326e-01
-4.17213082e-01 2.09722504e-01 -5.76627374e-01 4.84710298e-02
3.07888240e-01 -3.54442626e-01 -5.20165980e-01 3.57066512e-01
-8.22603226e-01 3.23528796e-01 8.07447433e-01 7.52082229e-01
7.25721598e-01 2.61995662e-02 4.14970279e-01 -1.15540111e+00
8.14645812e-02 -5.62394448e-02 -6.52438700e-01 2.98040450e-01
8.12735111e-02 1.18386321e-01 4.99468505e-01 -4.08333153e-01
-1.32299542e+00 4.02346253e-01 -4.59392041e-01 -1.32418081e-01
-3.38880241e-01 -1.65003926e-01 -8.52329135e-01 1.24943703e-01
6.27696663e-02 8.25221390e-02 -2.13045046e-01 -5.76670051e-01
4.26422060e-01 3.66859853e-01 9.33587909e-01 -6.62225366e-01
4.74428326e-01 3.74979705e-01 -1.65958151e-01 -6.76706195e-01
-1.07352209e+00 -3.71723235e-01 -7.52157092e-01 -1.44797295e-01
1.30831349e+00 -7.94118822e-01 -3.34124267e-01 8.04264367e-01
-1.35865724e+00 -6.11372232e-01 2.72161677e-04 2.68242322e-02
-3.09306294e-01 6.01573050e-01 -9.52132463e-01 -7.41631866e-01
-4.14550453e-01 -1.08600569e+00 1.24976456e+00 5.82096159e-01
-1.38761541e-02 -5.95621347e-01 -3.47679645e-01 4.85973239e-01
-2.15067133e-01 4.63529378e-01 8.55496943e-01 -6.13890290e-01
-6.99665308e-01 6.54139444e-02 -5.38751960e-01 4.95731503e-01
1.94787607e-01 1.62539482e-01 -9.75564539e-01 1.23388171e-01
-1.36951596e-01 -1.25436038e-01 1.17875504e+00 3.83882850e-01
1.24721169e+00 -1.18722253e-01 -3.75611365e-01 7.16541171e-01
8.51416528e-01 2.48696357e-01 8.15763354e-01 2.16888979e-01
9.86551881e-01 7.79120207e-01 5.40381849e-01 4.67269085e-02
3.61865968e-01 3.29415619e-01 4.52423692e-01 -3.85985553e-01
-4.93527353e-01 -3.83031398e-01 2.55685449e-01 5.57904601e-01
-1.04174241e-01 -2.47531414e-01 -6.40362084e-01 2.88335621e-01
-1.82493341e+00 -6.74369693e-01 -2.58646697e-01 1.58697081e+00
7.77302980e-01 5.11603951e-01 1.63793772e-01 -3.67741771e-02
1.00899661e+00 2.62848794e-01 -5.38287580e-01 -2.29695424e-01
-3.25760901e-01 1.48821443e-01 2.63806134e-01 1.85035959e-01
-1.65005577e+00 1.36100686e+00 5.61392689e+00 6.79283023e-01
-9.47875440e-01 -5.79561144e-02 8.85137618e-01 3.88607979e-01
2.44730085e-01 -1.15886897e-01 -1.02422905e+00 6.16497636e-01
3.86072606e-01 6.48291230e-01 5.91729805e-02 9.31015432e-01
-6.73012286e-02 -4.10339572e-02 -8.35210681e-01 7.12523103e-01
-6.18526749e-02 -8.04735601e-01 5.46444915e-02 -1.83507115e-01
4.62782174e-01 -1.21159539e-01 -2.92309940e-01 5.83957553e-01
4.32566643e-01 -5.95477283e-01 7.89038539e-01 2.48288810e-02
1.76812828e-01 -6.37414396e-01 9.29528177e-01 3.22944939e-01
-1.31858492e+00 2.40717307e-02 -3.80406171e-01 1.46820083e-01
2.90707946e-01 5.45614898e-01 -2.65929371e-01 6.21633947e-01
8.27534139e-01 5.25590360e-01 -6.58406019e-01 8.31587255e-01
-8.18219543e-01 6.34289682e-01 -3.17947507e-01 1.16475359e-01
3.54887456e-01 -1.59771949e-01 3.64434093e-01 1.44330835e+00
-2.17497125e-01 3.43208164e-01 3.28747958e-01 8.97111773e-01
-1.87981397e-01 4.88281175e-02 -1.31572321e-01 1.65474817e-01
9.34446007e-02 1.50229597e+00 -1.19453490e+00 -4.75756794e-01
-4.53084320e-01 1.14508617e+00 4.46996957e-01 3.08603585e-01
-9.88312364e-01 -3.85739982e-01 3.38143498e-01 -3.11178192e-02
6.20036185e-01 1.81702562e-02 -8.73957127e-02 -1.10914946e+00
2.52422243e-01 -1.02340150e+00 6.02466226e-01 -5.52974343e-01
-1.33304214e+00 6.85640037e-01 -1.86707184e-01 -6.89448059e-01
1.40964925e-01 -6.68324053e-01 -7.44923055e-01 5.27116179e-01
-1.51333117e+00 -1.32566273e+00 -3.64161551e-01 4.17427123e-01
7.09298134e-01 2.62304217e-01 2.65729219e-01 3.04379821e-01
-9.58548784e-01 5.42344093e-01 -7.43140101e-01 7.98960149e-01
5.21502256e-01 -1.39154637e+00 7.35387683e-01 1.18636954e+00
3.09926756e-02 3.53658170e-01 3.34516317e-01 -9.92712140e-01
-8.86640429e-01 -1.40899861e+00 6.60722435e-01 -1.77925095e-01
3.48769039e-01 -6.24495625e-01 -1.10110092e+00 6.84676886e-01
1.31441548e-01 1.54682547e-01 2.75208771e-01 -1.38118062e-02
-4.30346638e-01 -4.82601710e-02 -9.53221977e-01 6.66088045e-01
1.34045291e+00 -1.20997921e-01 -8.26670706e-01 1.21248171e-01
9.94824946e-01 -5.95455766e-01 -3.76344830e-01 6.94475293e-01
4.80295122e-01 -1.12701476e+00 8.24871302e-01 -4.29181844e-01
4.26511168e-01 -2.85144389e-01 -5.22733890e-02 -7.72983849e-01
-2.36512393e-01 -6.55008554e-01 -1.41179442e-01 1.44113052e+00
2.53398567e-01 -4.75142032e-01 1.00520790e+00 4.41712528e-01
-3.52824360e-01 -7.90069640e-01 -7.05029547e-01 -6.38419986e-01
5.82588352e-02 -1.42095953e-01 4.84126270e-01 4.92428184e-01
-5.67143857e-01 4.85223502e-01 -2.33517081e-01 2.66168118e-01
7.85498202e-01 -2.74243224e-02 8.34275603e-01 -1.13595140e+00
-3.36525112e-01 -4.03186142e-01 -3.05689752e-01 -1.22655690e+00
2.45505095e-01 -4.63557124e-01 4.20410305e-01 -1.54167378e+00
2.58997381e-01 -2.76291072e-01 -1.97022587e-01 7.14664698e-01
-7.87369788e-01 8.79371092e-02 4.97900933e-01 -1.31644771e-01
-9.28512990e-01 3.55495423e-01 1.53284097e+00 -2.02049509e-01
-1.48928627e-01 5.57216033e-02 -6.96460009e-01 1.19284654e+00
7.73226798e-01 -4.00472015e-01 -2.24678114e-01 -4.63661939e-01
-2.54445612e-01 -9.23790783e-02 4.67994243e-01 -9.49382663e-01
4.50675301e-02 8.60400572e-02 5.55812716e-01 -9.29890692e-01
5.91611415e-02 -5.04429877e-01 -2.56443709e-01 4.41739231e-01
-1.22232996e-02 -3.16021480e-02 2.14455351e-01 5.80189586e-01
-1.49376556e-01 -3.91213559e-02 9.13562894e-01 -4.25284863e-01
-9.01250422e-01 3.15234840e-01 -1.86626464e-01 3.33755344e-01
7.46340692e-01 -1.60554737e-01 -3.08172554e-01 1.18833974e-01
-7.15610087e-01 3.90921593e-01 1.44264609e-01 3.26737702e-01
4.35989827e-01 -8.88308287e-01 -5.93763649e-01 3.48932624e-01
-1.65765271e-01 4.65886176e-01 3.86504829e-01 5.39484143e-01
-5.83841622e-01 -1.00193888e-01 -1.15726523e-01 -6.40073061e-01
-1.39712894e+00 5.61960876e-01 4.07761246e-01 -4.58593369e-01
-1.07088470e+00 9.75561142e-01 8.66701365e-01 -4.82869148e-01
3.67170602e-01 -6.89770997e-01 -1.91087544e-01 -2.33627945e-01
6.07887805e-01 8.26540366e-02 -2.96172023e-01 -3.96938264e-01
-2.87921697e-01 6.21082246e-01 -2.48199955e-01 3.11664373e-01
9.21610892e-01 -8.01309273e-02 -2.18177170e-01 -2.47157980e-02
8.62197518e-01 -3.29400934e-02 -1.69409740e+00 -1.39903054e-01
1.30756482e-01 -6.64173663e-02 -2.26018727e-01 -8.77426863e-01
-1.44224167e+00 9.66271281e-01 4.11018968e-01 -1.11228228e-03
1.24501157e+00 2.09399149e-01 1.14445496e+00 1.27693042e-01
1.75282925e-01 -1.06684399e+00 4.83160019e-02 5.80032706e-01
5.27112603e-01 -1.20056748e+00 -2.80940741e-01 -1.00049508e+00
-5.47752023e-01 1.08952081e+00 8.92843962e-01 -1.45558357e-01
3.36945504e-01 3.92388374e-01 2.92813271e-01 -1.14371434e-01
-3.75806779e-01 -4.41981375e-01 2.13908598e-01 6.36260569e-01
2.81465501e-01 3.54209878e-02 -2.55022079e-01 9.63406205e-01
-5.25405910e-03 -2.15433046e-01 2.25696966e-01 1.00531876e+00
-5.29957771e-01 -1.18983901e+00 -3.08262467e-01 2.28155449e-01
-8.07141244e-01 -2.73468290e-02 -6.01603866e-01 1.02058852e+00
5.22672415e-01 9.66500819e-01 -1.30012948e-02 -1.68468699e-01
3.81185830e-01 2.36504339e-02 4.57794666e-01 -5.63063323e-01
-5.61086595e-01 5.93645155e-01 1.55440196e-01 -7.19850659e-01
-4.14178401e-01 -7.04561591e-01 -1.53491533e+00 1.36628494e-01
-1.99354246e-01 -1.34770378e-01 2.91872472e-01 1.03487539e+00
9.23130661e-02 9.93299723e-01 9.05221999e-02 -5.42270601e-01
-3.25252533e-01 -8.18604231e-01 -5.40835857e-01 4.66201484e-01
7.66240358e-02 -5.19436777e-01 -6.19228259e-02 7.25674033e-02] | [8.691500663757324, 0.05304848402738571] |
91c4937e-1812-4ee9-adec-561e6dbb56d9 | dvg-face-dual-variational-generation-for | 2009.09399 | null | https://arxiv.org/abs/2009.09399v2 | https://arxiv.org/pdf/2009.09399v2.pdf | DVG-Face: Dual Variational Generation for Heterogeneous Face Recognition | Heterogeneous Face Recognition (HFR) refers to matching cross-domain faces and plays a crucial role in public security. Nevertheless, HFR is confronted with challenges from large domain discrepancy and insufficient heterogeneous data. In this paper, we formulate HFR as a dual generation problem, and tackle it via a novel Dual Variational Generation (DVG-Face) framework. Specifically, a dual variational generator is elaborately designed to learn the joint distribution of paired heterogeneous images. However, the small-scale paired heterogeneous training data may limit the identity diversity of sampling. In order to break through the limitation, we propose to integrate abundant identity information of large-scale visible data into the joint distribution. Furthermore, a pairwise identity preserving loss is imposed on the generated paired heterogeneous images to ensure their identity consistency. As a consequence, massive new diverse paired heterogeneous images with the same identity can be generated from noises. The identity consistency and identity diversity properties allow us to employ these generated images to train the HFR network via a contrastive learning mechanism, yielding both domain-invariant and discriminative embedding features. Concretely, the generated paired heterogeneous images are regarded as positive pairs, and the images obtained from different samplings are considered as negative pairs. Our method achieves superior performances over state-of-the-art methods on seven challenging databases belonging to five HFR tasks, including NIR-VIS, Sketch-Photo, Profile-Frontal Photo, Thermal-VIS, and ID-Camera. The related code will be released at https://github.com/BradyFU. | ['Yibo Hu', 'Xiang Wu', 'Huaibo Huang', 'Ran He', 'Chaoyou Fu'] | 2020-09-20 | null | null | null | null | ['heterogeneous-face-recognition'] | ['computer-vision'] | [ 8.88946131e-02 -2.48760864e-01 -7.71591961e-02 -3.22910905e-01
-1.11573553e+00 -5.43645978e-01 5.45684516e-01 -6.68729842e-01
3.92201217e-03 7.83598006e-01 5.02087362e-02 2.67711818e-01
-3.73493247e-02 -7.82079339e-01 -5.09850502e-01 -1.18497109e+00
6.33943379e-01 2.87970185e-01 -3.62886369e-01 -3.36600751e-01
-1.85581669e-01 4.29044515e-01 -1.59052062e+00 2.36446261e-01
8.31662714e-01 1.19810641e+00 1.13552488e-01 -1.10775284e-01
7.60728493e-02 2.58689255e-01 -5.35659909e-01 -6.89007521e-01
5.74801087e-01 -4.99399513e-01 -1.90635934e-01 1.42062381e-01
4.56316918e-01 -2.88423359e-01 -5.61536610e-01 1.25332618e+00
9.59097147e-01 5.99374883e-02 6.51692212e-01 -1.66981697e+00
-1.14622962e+00 4.84562479e-02 -8.20850074e-01 -2.90638864e-01
4.98584419e-01 3.08284104e-01 7.28566349e-01 -1.19676328e+00
8.21866751e-01 1.34551597e+00 4.44005042e-01 9.55033123e-01
-1.27839959e+00 -1.20114648e+00 1.48877308e-01 2.15165168e-01
-1.82337558e+00 -6.99188888e-01 1.17069709e+00 -4.12546009e-01
-2.58984268e-02 3.40884089e-01 3.66853118e-01 1.55477238e+00
-3.75817865e-01 6.44726872e-01 1.23768651e+00 5.54024577e-02
-1.99734438e-02 2.86282301e-01 -4.99420792e-01 4.47453618e-01
1.54508784e-01 1.74341649e-01 -5.52072942e-01 -3.00121993e-01
6.59037650e-01 1.74587905e-01 -6.08736753e-01 -3.19274634e-01
-9.87743258e-01 7.60591984e-01 2.57678896e-01 1.45300806e-01
-3.28070402e-01 -5.67907453e-01 3.11407298e-01 1.79191366e-01
5.30691028e-01 -1.85936823e-01 -2.29510795e-02 4.93999183e-01
-6.62108481e-01 2.26484984e-01 2.67364562e-01 9.85334754e-01
9.32689786e-01 7.01825880e-03 -4.00014967e-01 1.21502018e+00
4.61822122e-01 8.28273654e-01 3.33481759e-01 -9.22574162e-01
7.18126714e-01 5.07202029e-01 3.94818634e-02 -1.22597826e+00
2.18648151e-01 -2.89564073e-01 -1.24516749e+00 1.11706957e-01
2.50873446e-01 3.36362943e-02 -6.79016769e-01 1.96632111e+00
6.55069292e-01 2.91669220e-01 3.03772122e-01 9.79900658e-01
1.07382941e+00 7.13224828e-01 -1.35444969e-01 -3.23669374e-01
1.22837055e+00 -6.20846689e-01 -6.99559093e-01 8.87331590e-02
-1.11475654e-01 -9.54166830e-01 9.05135155e-01 2.34695598e-01
-9.94563758e-01 -6.83655202e-01 -9.84360516e-01 -6.05775639e-02
-8.19329396e-02 3.26311946e-01 1.34026085e-03 5.58196425e-01
-7.57837355e-01 3.11450481e-01 -7.62385800e-02 1.02847919e-01
7.39894748e-01 2.43242800e-01 -8.14179063e-01 -6.17706418e-01
-1.28747594e+00 3.74049842e-01 7.46456161e-02 4.22802866e-01
-9.27195251e-01 -7.76081383e-01 -7.59261131e-01 -2.94463933e-01
3.17280591e-01 -7.29608655e-01 7.51664281e-01 -7.69597590e-01
-1.33542585e+00 1.18811345e+00 -1.24349177e-01 3.01112056e-01
9.33188796e-01 4.11244512e-01 -6.65874422e-01 1.05724871e-01
3.14396799e-01 3.46274704e-01 1.09687340e+00 -1.60531986e+00
-2.49174759e-01 -7.58396387e-01 -2.08217129e-01 1.10243641e-01
-4.47673112e-01 -1.45513043e-01 -5.06383479e-01 -6.67797565e-01
7.12597445e-02 -9.70730305e-01 1.48452416e-01 7.66844749e-02
-3.50481629e-01 -1.84144020e-01 8.45708013e-01 -8.74286830e-01
7.95518458e-01 -2.40968609e+00 2.74190903e-01 1.04337461e-01
3.19629282e-01 2.71228045e-01 -5.39255679e-01 1.65697247e-01
-1.86095610e-01 -5.93287162e-02 -2.52388537e-01 -4.23225164e-01
1.21329635e-01 -3.40386517e-02 -2.18162775e-01 6.41709864e-01
2.59158790e-01 7.47677684e-01 -7.98638761e-01 -5.88877261e-01
6.07819483e-02 7.51840293e-01 -2.87194133e-01 4.70973730e-01
5.30309184e-03 7.92528033e-01 -5.32401979e-01 9.62163866e-01
1.38434136e+00 6.73483089e-02 8.97196829e-02 -5.85909545e-01
1.46302149e-01 -5.55256069e-01 -1.20091224e+00 1.63481295e+00
-4.38116699e-01 1.77450716e-01 4.01760191e-01 -1.02681279e+00
1.16722107e+00 4.79959548e-01 5.12053072e-01 -8.72358084e-01
1.49667621e-01 5.41628361e-01 -3.21178645e-01 -5.21587193e-01
1.65638208e-01 -3.28498989e-01 1.10592350e-01 9.94820446e-02
1.83178615e-02 7.67686516e-02 -1.90594450e-01 -2.87339222e-02
5.40665507e-01 1.85685232e-01 -7.50773698e-02 -4.35488485e-03
9.03001726e-01 -4.54580367e-01 9.76965606e-01 2.23756835e-01
-2.64450848e-01 1.01958084e+00 3.28042895e-01 -2.45183215e-01
-1.11907709e+00 -1.32370949e+00 -2.62427330e-01 5.80627143e-01
3.76275808e-01 4.92676944e-02 -4.72936392e-01 -6.75328851e-01
7.78079033e-02 8.52100104e-02 -6.59426987e-01 -2.18155101e-01
-4.54558849e-01 -7.08839118e-01 4.61002737e-01 2.85869151e-01
7.64244914e-01 -8.81142437e-01 5.27761877e-01 -6.88002035e-02
-4.93484050e-01 -1.07715821e+00 -6.95174217e-01 -6.37399375e-01
-2.51180440e-01 -1.21765244e+00 -1.21638179e+00 -8.51810336e-01
7.17559516e-01 4.87873673e-01 8.96890104e-01 -1.79943964e-01
-4.08504933e-01 2.39583477e-01 -2.01808661e-01 4.71925326e-02
-3.69418234e-01 -3.41011733e-01 1.79019213e-01 7.03856349e-01
2.54227906e-01 -6.62643135e-01 -8.17047000e-01 5.62987030e-01
-1.00603724e+00 -3.95201109e-02 4.57309335e-01 1.25534523e+00
7.63749421e-01 -1.10325620e-01 7.15300322e-01 -5.97282112e-01
4.60176468e-01 -7.28459954e-01 -5.36566734e-01 5.11420012e-01
-1.92917511e-01 -3.02178293e-01 8.76314580e-01 -6.02632761e-01
-1.26780093e+00 -8.94709863e-03 -2.33567700e-01 -1.01715386e+00
-5.86394034e-02 6.82687610e-02 -8.84416223e-01 -2.35418051e-01
4.60236192e-01 4.39555049e-01 1.96676537e-01 -3.01178426e-01
2.72416651e-01 8.78389776e-01 5.29428363e-01 -8.80786121e-01
1.17396438e+00 6.53745532e-01 5.22857197e-02 -7.49491096e-01
-5.24846017e-01 -2.89660037e-01 -1.45761609e-01 -1.82646260e-01
6.19512022e-01 -1.25760436e+00 -7.25137711e-01 7.70579696e-01
-1.07688177e+00 9.39981788e-02 -3.85480970e-02 2.65002906e-01
-2.44374394e-01 3.27266127e-01 -3.68220538e-01 -7.62214541e-01
-2.78385401e-01 -1.18232012e+00 1.29799902e+00 3.41459513e-01
4.76803511e-01 -5.35906553e-01 -3.59968394e-02 7.68197596e-01
1.98894545e-01 4.68962938e-01 4.39932555e-01 -2.38834828e-01
-5.64717412e-01 -1.76691219e-01 -4.26467448e-01 7.10436702e-01
3.22569400e-01 2.44168248e-02 -1.23980618e+00 -4.39971924e-01
1.03923336e-01 -4.42104548e-01 6.55388772e-01 6.71358109e-02
1.25668800e+00 -2.41702691e-01 -1.50444105e-01 7.61953354e-01
1.36807191e+00 2.81121045e-01 6.48355126e-01 -1.67633355e-01
8.44807267e-01 9.55673993e-01 6.71604156e-01 5.62555611e-01
2.30942801e-01 7.72578120e-01 2.27163464e-01 -6.36485815e-02
-7.63697401e-02 -5.02972543e-01 2.34077230e-01 5.41792035e-01
-1.31427526e-01 -2.20487744e-01 -4.96253729e-01 3.68308932e-01
-1.72097504e+00 -1.23956871e+00 1.42621741e-01 2.25274205e+00
6.40581727e-01 -5.41526794e-01 1.54620409e-02 -1.03610761e-01
1.29595923e+00 3.65541905e-01 -7.95812309e-01 3.70389163e-01
-5.82140923e-01 -6.37341756e-03 4.70035151e-02 1.31591067e-01
-8.27441275e-01 4.95824605e-01 4.30039358e+00 1.30644941e+00
-1.04611313e+00 3.61035347e-01 9.31683660e-01 -2.23550528e-01
-7.32843995e-01 -3.25921088e-01 -7.54650891e-01 7.56652057e-01
1.62083283e-01 -2.29698569e-01 5.33256412e-01 6.87536657e-01
3.03096473e-02 4.85877693e-01 -7.89428234e-01 1.52445805e+00
2.36819834e-01 -1.25643671e+00 1.13705024e-01 2.19706833e-01
9.47125554e-01 -2.91420519e-01 5.37927151e-01 1.87526852e-01
2.31522657e-02 -9.54659104e-01 6.15984857e-01 3.47916245e-01
1.23481452e+00 -7.04673290e-01 5.94534874e-01 1.07014507e-01
-1.41026509e+00 -1.97075475e-02 -5.21367669e-01 4.61250991e-01
5.83530329e-02 5.09690106e-01 -1.29098579e-01 1.11626792e+00
5.55601120e-01 7.48734474e-01 -3.25892836e-01 5.40877998e-01
-7.49836937e-02 -5.70558570e-02 -7.60802254e-02 4.64551181e-01
-2.14022949e-01 -6.70332909e-01 4.23216701e-01 4.14849341e-01
5.90016067e-01 1.17082179e-01 1.35664567e-01 1.05419195e+00
-4.70419228e-01 1.19849473e-01 -7.45233357e-01 6.80511370e-02
7.07102776e-01 1.38154793e+00 3.87731828e-02 -1.54115885e-01
-4.78729546e-01 1.16940367e+00 1.27338514e-01 4.95804727e-01
-9.17812347e-01 -6.90746829e-02 8.88039291e-01 -7.56463036e-02
7.74077177e-02 2.62432188e-01 1.10165201e-01 -1.63774848e+00
5.08826673e-01 -1.08777499e+00 3.09635818e-01 -6.77130759e-01
-1.90589726e+00 8.22144806e-01 -2.03950897e-01 -1.62162173e+00
7.30699152e-02 -3.79006177e-01 -5.51358521e-01 1.15452278e+00
-1.54617727e+00 -1.50648320e+00 -6.99053526e-01 9.33769226e-01
4.23482694e-02 -5.34076512e-01 6.41789734e-01 8.47662628e-01
-8.56267869e-01 9.04647529e-01 2.50957340e-01 2.60163575e-01
8.48462641e-01 -7.11862147e-01 2.26257257e-02 6.54640853e-01
-1.69205904e-01 5.97134769e-01 2.92147845e-01 -4.21074003e-01
-1.45278621e+00 -1.16254294e+00 4.04264569e-01 -8.75489190e-02
3.31783801e-01 -4.86004353e-01 -8.77559543e-01 2.78932631e-01
3.68598253e-02 6.55543804e-01 7.07220197e-01 -2.68790781e-01
-8.33203733e-01 -5.10714114e-01 -1.49617517e+00 5.50611258e-01
1.27579319e+00 -8.65442812e-01 -1.55211523e-01 2.77869254e-01
6.04439855e-01 -2.38702625e-01 -9.94615197e-01 5.76678574e-01
7.42987037e-01 -1.03038788e+00 1.06641090e+00 -2.73212582e-01
4.00030613e-01 -5.30438960e-01 -4.21503901e-01 -1.16845238e+00
-1.51823074e-01 -5.98084450e-01 2.70967841e-01 1.83833253e+00
-2.63094045e-02 -9.85500932e-01 8.26280177e-01 2.78478652e-01
1.63897835e-02 -6.49897397e-01 -1.00182438e+00 -9.50450003e-01
1.13039069e-01 1.64920151e-01 9.20307815e-01 1.05757868e+00
-6.65615559e-01 1.04849793e-01 -5.49625993e-01 2.09164202e-01
1.00080132e+00 3.63854885e-01 7.47092903e-01 -1.18420708e+00
-1.73369989e-01 -1.50500640e-01 -2.79609472e-01 -7.22837865e-01
5.92278302e-01 -8.57833862e-01 -1.88213408e-01 -1.03897035e+00
4.31726784e-01 -7.01786160e-01 -2.45703012e-01 2.36210093e-01
-1.11762352e-01 6.64009511e-01 2.45137930e-01 2.44684294e-01
-1.85157627e-01 1.07082582e+00 1.58266747e+00 -3.89217824e-01
1.07501961e-01 -8.08708668e-02 -6.83505535e-01 2.79044122e-01
8.64300668e-01 -2.24929169e-01 -4.54907238e-01 -2.05929786e-01
-1.13368757e-01 2.83868581e-01 5.44847131e-01 -7.84194827e-01
6.64596781e-02 -2.72861063e-01 5.29386520e-01 -3.72981071e-01
6.19500160e-01 -7.72848547e-01 6.15428090e-01 1.65958032e-01
-1.83012914e-02 -1.84735268e-01 -1.53991461e-01 5.39302945e-01
-5.08640766e-01 1.16385728e-01 1.07397676e+00 -7.19470456e-02
-4.74380970e-01 9.82014954e-01 4.82036501e-01 2.67041028e-01
1.15498495e+00 -1.96547523e-01 -4.64824140e-01 -1.21829793e-01
-4.33677465e-01 1.15296803e-01 6.38931632e-01 6.46131873e-01
8.16327810e-01 -1.89150763e+00 -1.01997781e+00 6.08138084e-01
4.18667823e-01 4.99836132e-02 1.02826750e+00 6.82859898e-01
-1.10870361e-01 -3.66822518e-02 -3.58025610e-01 -3.83855790e-01
-1.37242436e+00 6.94549978e-01 3.30172330e-01 1.32299691e-01
-2.62443244e-01 7.89301157e-01 4.55116481e-01 -6.00664556e-01
-1.68078244e-02 6.03849471e-01 1.23752886e-02 3.84481370e-01
6.36325002e-01 3.10207188e-01 -1.34621471e-01 -1.04632592e+00
-3.85282218e-01 7.69052029e-01 1.19285777e-01 7.15044141e-02
1.19026148e+00 -1.65880188e-01 -2.48148918e-01 3.34234387e-02
1.77727330e+00 1.81855783e-01 -1.33208668e+00 -3.69370729e-01
-7.06745923e-01 -8.24621618e-01 -3.09464514e-01 -3.20782274e-01
-1.44885194e+00 9.01860774e-01 7.25393713e-01 -9.55502316e-02
1.28493965e+00 -8.32622126e-03 8.62867475e-01 -2.49725178e-01
3.87554049e-01 -8.75830233e-01 2.30367884e-01 -1.18868716e-01
1.04580784e+00 -1.53409636e+00 -2.38240361e-01 -5.77880442e-01
-6.54158771e-01 7.81383634e-01 8.01536798e-01 4.14660685e-02
4.88407135e-01 -1.30635872e-01 4.17775614e-03 9.74115208e-02
-3.99439573e-01 -4.84705754e-02 2.40745649e-01 8.14909279e-01
6.35520145e-02 1.31349295e-01 -2.18571827e-01 6.34649038e-01
1.78946540e-01 -3.63444239e-02 8.24908093e-02 4.08667058e-01
2.52758563e-01 -1.38743722e+00 -5.92834175e-01 1.65846452e-01
-2.74864852e-01 1.47873014e-01 -1.96774170e-01 5.37800848e-01
5.07818580e-01 9.76609826e-01 -2.71782458e-01 -6.57571077e-01
2.79059142e-01 -2.11449936e-01 6.05737567e-01 -3.14834028e-01
-2.83776015e-01 -1.12347580e-01 -2.75364250e-01 -4.60153043e-01
-3.31018507e-01 -6.29423022e-01 -6.93785548e-01 -5.66720307e-01
-1.62270993e-01 1.33914948e-01 3.11630189e-01 5.11479735e-01
4.73434150e-01 2.20156893e-01 1.25103855e+00 -8.46633017e-01
-7.79907584e-01 -7.10577667e-01 -8.01053762e-01 8.16154063e-01
3.88857663e-01 -8.04003716e-01 -6.17876053e-01 -2.08789408e-01] | [13.104403495788574, 0.314214289188385] |
a6629414-b9fd-4b44-8991-dffbfc44c72a | brain-inspired-spiking-neural-network-for | 2304.04697 | null | https://arxiv.org/abs/2304.04697v2 | https://arxiv.org/pdf/2304.04697v2.pdf | Brain-Inspired Spiking Neural Network for Online Unsupervised Time Series Prediction | Energy and data-efficient online time series prediction for predicting evolving dynamical systems are critical in several fields, especially edge AI applications that need to update continuously based on streaming data. However, current DNN-based supervised online learning models require a large amount of training data and cannot quickly adapt when the underlying system changes. Moreover, these models require continuous retraining with incoming data making them highly inefficient. To solve these issues, we present a novel Continuous Learning-based Unsupervised Recurrent Spiking Neural Network Model (CLURSNN), trained with spike timing dependent plasticity (STDP). CLURSNN makes online predictions by reconstructing the underlying dynamical system using Random Delay Embedding by measuring the membrane potential of neurons in the recurrent layer of the RSNN with the highest betweenness centrality. We also use topological data analysis to propose a novel methodology using the Wasserstein Distance between the persistence homologies of the predicted and observed time series as a loss function. We show that the proposed online time series prediction methodology outperforms state-of-the-art DNN models when predicting an evolving Lorenz63 dynamical system. | ['Saibal Mukhopadhyay', 'Biswadeep Chakraborty'] | 2023-04-10 | null | null | null | null | ['topological-data-analysis', 'time-series-prediction'] | ['graphs', 'time-series'] | [ 1.29935279e-01 -2.78776675e-01 3.67350131e-01 -5.97799569e-02
2.08754167e-01 -4.45263416e-01 5.38925350e-01 3.50060046e-01
-4.63475674e-01 9.07740474e-01 -2.18857363e-01 -1.08592689e-01
-6.43311083e-01 -8.14980030e-01 -8.81458461e-01 -1.06051314e+00
-7.29959846e-01 3.83980066e-01 6.76160991e-01 -3.12796921e-01
4.92226928e-01 7.56113946e-01 -1.62162471e+00 2.54249247e-03
6.44124091e-01 1.26084220e+00 2.96345651e-01 6.53741479e-01
2.24706307e-02 6.83319509e-01 -1.11541122e-01 1.80212051e-01
3.67864788e-01 -6.40485227e-01 -2.68195421e-01 -7.01069951e-01
-2.74481446e-01 1.77897722e-01 -9.12279189e-01 7.40077794e-01
6.23903513e-01 3.20467800e-01 6.65088117e-01 -1.19162476e+00
-5.97209930e-01 5.72887480e-01 8.43132809e-02 8.39748442e-01
-3.17428857e-01 1.92645639e-02 7.62410760e-01 -7.09296703e-01
8.81067693e-01 4.69271868e-01 9.75331187e-01 5.24405241e-01
-1.43393433e+00 -5.03746569e-01 -8.61780345e-02 6.70160174e-01
-1.14702058e+00 -2.86834061e-01 1.11103523e+00 -3.83321971e-01
1.22063327e+00 -2.20908672e-01 1.31257772e+00 9.03916597e-01
1.02005398e+00 1.48867965e-01 9.32651341e-01 -2.97189560e-02
8.79057705e-01 -5.31802893e-01 -8.17673579e-02 5.65548539e-01
7.39190802e-02 2.69725740e-01 -6.37709498e-01 -1.04348965e-01
7.26999640e-01 2.62746453e-01 -2.36382827e-01 -4.25042629e-01
-1.21912205e+00 3.73956800e-01 3.51142466e-01 4.90528047e-01
-5.02515495e-01 3.63313287e-01 4.41016078e-01 7.04374969e-01
6.35625899e-01 3.08631539e-01 -6.58829689e-01 -4.70759839e-01
-9.39815879e-01 1.09376565e-01 7.97829628e-01 3.90405446e-01
6.53452158e-01 2.66224504e-01 4.68248069e-01 4.69270051e-01
2.11065216e-03 4.08110261e-01 8.82419825e-01 -9.52666521e-01
2.31818985e-02 6.97006285e-01 -1.74500033e-01 -1.04252613e+00
-5.99185050e-01 -3.22643250e-01 -1.13031936e+00 2.29549780e-01
1.59912661e-01 -1.16345443e-01 -3.99683923e-01 1.54810488e+00
1.11098900e-01 7.76526332e-01 3.71992551e-02 4.35078621e-01
2.84440480e-02 1.11316025e+00 -4.33765143e-01 -6.58400297e-01
2.01428697e-01 -4.29465145e-01 -6.94411993e-01 3.09581816e-01
6.60128534e-01 4.44065519e-02 6.09471023e-01 1.43397763e-01
-1.04250562e+00 -2.33268186e-01 -1.03283644e+00 3.57233971e-01
-5.13435662e-01 -4.39515978e-01 1.17317699e-01 -7.86149874e-02
-1.47693014e+00 1.41025662e+00 -1.51811004e+00 -3.77468079e-01
1.20524056e-01 4.61284816e-01 -9.61936191e-02 5.30659974e-01
-1.03874552e+00 9.95566189e-01 3.71946335e-01 1.42711490e-01
-7.55832970e-01 -6.35428905e-01 -2.76148856e-01 -2.14719735e-02
-5.05718410e-01 -3.56278867e-01 6.05799437e-01 -8.24715257e-01
-1.91960990e+00 3.05214047e-01 -2.58821789e-02 -9.72032607e-01
4.46463227e-01 4.99668926e-01 -4.72872794e-01 1.16169900e-01
-5.36083877e-01 3.24999213e-01 7.38911569e-01 -5.23836374e-01
-8.57568309e-02 -3.03598195e-01 -7.87030339e-01 -1.31428048e-01
-5.13146400e-01 -5.84413826e-01 7.10041046e-01 -5.25291502e-01
2.95361727e-01 -8.98310959e-01 -1.02317899e-01 3.28197211e-01
3.18344086e-01 -3.87179941e-01 1.12425125e+00 -4.29991305e-01
1.16649187e+00 -2.07550097e+00 4.84820306e-01 2.82169491e-01
1.15817405e-01 1.01963185e-01 -1.05118297e-01 7.36575961e-01
-1.90087795e-01 -1.81166962e-01 -4.64947432e-01 5.83961606e-02
-1.43311396e-01 2.92365700e-01 -6.79721594e-01 6.12595439e-01
-1.69416666e-02 6.98298037e-01 -8.23650599e-01 8.23883340e-03
1.11007720e-01 6.05563283e-01 -1.71445683e-01 -4.63784002e-02
-2.09558874e-01 5.64353824e-01 2.20464356e-02 9.50528681e-02
3.14772189e-01 -3.71451378e-01 -9.71485451e-02 2.37160519e-01
-6.90345466e-01 1.01444364e-01 -6.15008891e-01 1.46309853e+00
-3.02918345e-01 1.08420432e+00 -7.04179049e-01 -1.56775796e+00
1.31547534e+00 5.75159080e-02 8.89813125e-01 -7.93250263e-01
9.38475132e-02 5.02257884e-01 2.08064452e-01 -3.71436626e-01
7.96382129e-02 1.74025804e-01 3.06398481e-01 6.39770389e-01
2.10819408e-01 2.87529290e-01 -1.71897441e-01 -1.11730419e-01
1.69042623e+00 9.12706554e-02 -4.28059213e-02 -4.24027890e-01
2.90671051e-01 -3.58680338e-01 5.42393327e-01 3.28591555e-01
-3.78233671e-01 2.24164724e-01 6.47755921e-01 -8.18045378e-01
-1.53076839e+00 -1.05571783e+00 -2.37224206e-01 6.24040604e-01
-2.05149390e-02 -5.12404516e-02 -4.78059322e-01 8.62096250e-02
-1.40673071e-01 4.89046931e-01 -7.83032954e-01 -4.02364016e-01
-6.24730766e-01 -6.77335739e-01 5.09216964e-01 3.02290887e-01
2.67522067e-01 -1.38276148e+00 -9.69117284e-01 7.27554381e-01
1.73692461e-02 -7.93540359e-01 2.88830195e-02 3.49931926e-01
-1.30575705e+00 -7.04147220e-01 -7.35959291e-01 -9.85322952e-01
6.56941891e-01 -3.46783459e-01 4.05647606e-01 -1.83139250e-01
-3.64597350e-01 3.42443764e-01 -3.62559706e-01 6.18050136e-02
-1.94369748e-01 1.12173217e-03 5.33585131e-01 4.31200713e-02
1.31308377e-01 -1.45210195e+00 -7.28893399e-01 2.40859643e-01
-7.81857550e-01 -3.38684162e-03 1.20306492e-01 8.06167364e-01
1.10185087e+00 1.31605551e-01 9.97231424e-01 -2.47695237e-01
4.89891618e-01 -7.00329244e-01 -8.19109917e-01 3.24871331e-01
-8.08291316e-01 4.06121522e-01 1.30624604e+00 -8.87638867e-01
-3.72921735e-01 6.83033094e-02 1.79905802e-01 -4.22467887e-01
3.80091399e-01 5.69909930e-01 6.12938166e-01 -4.84264851e-01
6.78827703e-01 9.94761050e-01 5.26570790e-02 -6.40949933e-03
-2.34819010e-01 2.85917044e-01 2.98495770e-01 -8.58838782e-02
4.79078978e-01 4.94213700e-01 3.84156436e-01 -8.83705616e-01
-9.18277949e-02 5.31777665e-02 -7.82419086e-01 -5.10177195e-01
3.42676580e-01 -3.84840876e-01 -8.77472997e-01 9.26844776e-01
-1.11182439e+00 -7.36033678e-01 -5.28765440e-01 3.82453471e-01
-9.87836182e-01 1.51541755e-01 -6.83458090e-01 -9.61655319e-01
-3.63345504e-01 -2.02435181e-01 2.58463532e-01 4.41164859e-02
1.44891486e-01 -1.25213242e+00 7.89379537e-01 -7.78014302e-01
8.28594863e-01 6.03496373e-01 1.00665939e+00 -6.57208562e-01
-3.78113061e-01 -2.05240399e-01 1.54908732e-01 1.65178254e-01
-5.65467067e-02 2.88414299e-01 -4.89366651e-01 -2.81943321e-01
2.39353240e-01 5.53702787e-02 7.88599908e-01 5.29603362e-01
1.34541523e+00 -3.47002774e-01 -2.67367244e-01 5.79328656e-01
1.53469992e+00 3.60870510e-01 6.34450972e-01 3.58485244e-02
2.62341142e-01 4.88329977e-01 -1.09087154e-01 6.82452500e-01
3.41363847e-01 -1.01305023e-02 3.37617368e-01 9.37123477e-01
3.00549835e-01 -4.01234865e-01 8.09829712e-01 1.78674841e+00
-2.55954295e-01 -1.00555360e-01 -1.10297489e+00 6.32185280e-01
-2.12243938e+00 -1.25901401e+00 2.56352037e-01 2.24336648e+00
6.63136601e-01 2.45463058e-01 -1.29324421e-02 3.50228518e-01
7.58288264e-01 -8.72671902e-02 -1.36477363e+00 -2.69411296e-01
-3.91610354e-01 3.04609627e-01 3.99769187e-01 1.95371509e-01
-4.05282080e-01 4.81308818e-01 5.90981960e+00 3.63358498e-01
-1.44940352e+00 2.60760589e-03 3.81696701e-01 -1.62816986e-01
-1.10376574e-01 -2.23465357e-02 -4.37643588e-01 8.57331216e-01
1.80020702e+00 -6.15033031e-01 9.48868573e-01 3.87258500e-01
4.78143990e-01 1.14448927e-01 -1.01937318e+00 1.11506832e+00
-1.42330155e-01 -1.58070076e+00 -1.58518016e-01 8.72816294e-02
6.01800203e-01 5.91664791e-01 8.26453641e-02 -5.90619184e-02
-2.09740371e-01 -6.68828189e-01 3.30917358e-01 1.36091125e+00
4.71806437e-01 -6.98320627e-01 5.35963774e-01 6.81234419e-01
-1.21451044e+00 -1.99661598e-01 -7.31846690e-01 -3.31025571e-01
4.38457541e-02 7.43266702e-01 -3.15486461e-01 -2.62035042e-01
7.51394868e-01 1.35795474e+00 -3.05236816e-01 1.33874929e+00
4.01120186e-01 6.81049407e-01 -7.54951537e-01 -6.59125507e-01
-2.21595585e-01 -3.88431579e-01 6.91131115e-01 5.29465020e-01
9.14474428e-01 1.95620850e-01 -6.05972409e-01 8.91043305e-01
-8.14361423e-02 -1.12535946e-01 -9.26633716e-01 -3.92878532e-01
6.14817321e-01 8.50307107e-01 -9.53400552e-01 1.10325433e-01
9.05121341e-02 9.24068868e-01 4.09852058e-01 2.41606757e-01
-7.02553332e-01 -6.53489470e-01 2.56195128e-01 2.72508293e-01
4.62291658e-01 -7.52810240e-01 -5.22682853e-02 -1.18271518e+00
6.86169416e-02 5.04481383e-02 -1.19645886e-01 -7.46701300e-01
-1.39074099e+00 7.94783890e-01 -4.06787276e-01 -1.60596931e+00
-4.03445125e-01 -4.97350484e-01 -6.77344203e-01 3.37331980e-01
-1.28887343e+00 -4.19265956e-01 1.11856647e-01 6.90390110e-01
2.43186504e-01 -3.68944943e-01 1.03876495e+00 1.40840709e-01
-3.90823305e-01 2.92039543e-01 9.05931473e-01 -2.79485673e-01
1.85326487e-01 -9.19175923e-01 5.52751303e-01 6.38045073e-01
-1.16934352e-01 3.11564893e-01 7.60359466e-01 -6.10870659e-01
-1.37808299e+00 -1.15669799e+00 8.48500013e-01 1.51669318e-02
1.30543411e+00 -4.45367187e-01 -1.12673271e+00 4.10079688e-01
-1.03559941e-01 5.23540199e-01 6.75047994e-01 -5.50547421e-01
1.41362563e-01 -3.42602879e-01 -1.11824095e+00 5.02634346e-01
1.18356097e+00 -5.98508954e-01 -4.25559133e-01 2.13997796e-01
4.73497570e-01 1.01964816e-01 -9.89119411e-01 8.35573971e-02
6.51129127e-01 -7.61159778e-01 3.53461236e-01 -3.83491516e-01
2.31203720e-01 -1.81843922e-01 -6.60672858e-02 -1.49526262e+00
-3.03438634e-01 -8.97416830e-01 -7.82345235e-01 5.33636928e-01
3.87137324e-01 -1.22660279e+00 6.47574067e-01 1.68513849e-01
7.10642114e-02 -8.85121763e-01 -1.53835571e+00 -9.25457120e-01
1.65068612e-01 -1.41980007e-01 1.01141281e-01 8.13376784e-01
4.61847723e-01 -2.32063919e-01 -8.08783844e-02 2.92259641e-02
5.55239677e-01 -1.52141348e-01 -2.26892494e-02 -1.65612280e+00
4.95342491e-03 -4.62231964e-01 -9.98568058e-01 -5.73022485e-01
3.30183417e-01 -9.06479001e-01 2.91788727e-02 -1.17808771e+00
-3.81173551e-01 -3.44345421e-01 -7.33692944e-01 2.54767388e-01
7.13821709e-01 -2.00419009e-01 -2.43025765e-01 5.76504529e-01
-4.56874728e-01 1.05517507e+00 7.23235309e-01 2.34245926e-01
-2.73435116e-01 1.81287020e-01 2.33389348e-01 3.82677108e-01
9.63634074e-01 -7.68138111e-01 -4.35916215e-01 -1.44890875e-01
9.15856183e-01 3.03764731e-01 3.99564296e-01 -1.48900473e+00
9.51115727e-01 7.09670410e-02 3.31320286e-01 -6.47148252e-01
1.84743419e-01 -8.55493903e-01 2.61003256e-01 8.74787450e-01
-5.03457844e-01 6.51152432e-01 1.29168317e-01 1.12922215e+00
-1.65624693e-02 4.90362756e-02 6.79997385e-01 3.28375340e-01
-3.61796021e-01 7.19083488e-01 -8.50511134e-01 -5.85974120e-02
9.55765247e-01 -2.90200889e-01 -4.13723469e-01 -2.63753921e-01
-9.85051334e-01 -6.75787851e-02 3.62954795e-01 1.41928330e-01
8.89769554e-01 -1.17104006e+00 -4.72997516e-01 2.93909639e-01
-2.70137280e-01 -4.65322077e-01 3.40180427e-01 8.69411051e-01
-7.67436802e-01 3.56830880e-02 -8.25976074e-01 -5.25767028e-01
-4.75916892e-01 2.53823847e-01 5.64794064e-01 2.10021716e-02
-8.84414434e-01 4.13446188e-01 -8.91194820e-01 -3.90109628e-01
7.31391087e-02 -5.25267124e-01 -8.52767900e-02 3.75719294e-02
3.05717647e-01 6.48745596e-01 3.42118293e-02 -3.76054138e-01
-3.39985698e-01 5.63501239e-01 2.98402190e-01 -1.97045118e-01
2.04232264e+00 -1.37992814e-01 -6.02625012e-01 1.36341870e+00
1.29056442e+00 -6.83523655e-01 -1.60020435e+00 -2.71326035e-01
1.97190374e-01 2.65379071e-01 1.10108294e-02 -3.26749891e-01
-8.10118735e-01 1.03890383e+00 1.07301354e+00 5.55827916e-01
1.18352640e+00 -4.77912366e-01 1.11775827e+00 9.31165576e-01
5.61306119e-01 -1.46451986e+00 7.91081116e-02 9.04153109e-01
7.02300668e-01 -7.10155487e-01 -3.81956995e-01 6.06992662e-01
-4.45126407e-02 1.73030484e+00 1.50966778e-01 -8.13131034e-01
1.38737607e+00 2.23869145e-01 -6.05517149e-01 2.25551739e-01
-1.30579364e+00 4.12798166e-01 -1.51811257e-01 4.35864508e-01
-6.75322488e-02 -2.11732499e-02 -1.90017357e-01 3.45851570e-01
-2.30407909e-01 1.68986008e-01 7.53099561e-01 8.73878777e-01
-7.05188632e-01 -7.18709111e-01 4.51565474e-01 6.55489147e-01
-1.01816142e-02 -4.63659018e-02 -2.22748324e-01 1.83130398e-01
-3.27954024e-01 2.55588204e-01 3.52340847e-01 -7.07197011e-01
-1.71001509e-01 2.97473133e-01 4.73224044e-01 -7.58760795e-02
-1.85057968e-01 -5.64516366e-01 -7.48435736e-01 -2.82330722e-01
-2.79368073e-01 -8.67054820e-01 -1.49248683e+00 -4.56237763e-01
-2.57674158e-01 -1.15180857e-01 8.86244535e-01 1.03764975e+00
8.00781786e-01 3.34178597e-01 1.09253967e+00 -1.00178957e+00
-3.33160728e-01 -7.55571127e-01 -6.44442141e-01 -2.31110975e-01
6.03168964e-01 -4.96666849e-01 -8.06486547e-01 5.51431216e-02] | [6.750194072723389, 3.4342970848083496] |
ab1973d9-24f3-4843-911d-3a9a89540420 | deepstory-video-story-qa-by-deep-embedded | 1707.00836 | null | http://arxiv.org/abs/1707.00836v1 | http://arxiv.org/pdf/1707.00836v1.pdf | DeepStory: Video Story QA by Deep Embedded Memory Networks | Question-answering (QA) on video contents is a significant challenge for
achieving human-level intelligence as it involves both vision and language in
real-world settings. Here we demonstrate the possibility of an AI agent
performing video story QA by learning from a large amount of cartoon videos. We
develop a video-story learning model, i.e. Deep Embedded Memory Networks
(DEMN), to reconstruct stories from a joint scene-dialogue video stream using a
latent embedding space of observed data. The video stories are stored in a
long-term memory component. For a given question, an LSTM-based attention model
uses the long-term memory to recall the best question-story-answer triplet by
focusing on specific words containing key information. We trained the DEMN on a
novel QA dataset of children's cartoon video series, Pororo. The dataset
contains 16,066 scene-dialogue pairs of 20.5-hour videos, 27,328 fine-grained
sentences for scene description, and 8,913 story-related QA pairs. Our
experimental results show that the DEMN outperforms other QA models. This is
mainly due to 1) the reconstruction of video stories in a scene-dialogue
combined form that utilize the latent embedding and 2) attention. DEMN also
achieved state-of-the-art results on the MovieQA benchmark. | ['Byoung-Tak Zhang', 'Seong-Ho Choi', 'Min-Oh Heo', 'Kyung-Min Kim'] | 2017-07-04 | null | null | null | null | ['video-story-qa'] | ['computer-vision'] | [ 1.93552047e-01 -6.13120198e-02 1.50008321e-01 -5.28525591e-01
-1.25456262e+00 -2.40492553e-01 5.36580265e-01 -2.63465375e-01
-2.94291645e-01 4.25982744e-01 7.78699458e-01 2.60653079e-01
2.12710053e-01 -7.75745451e-01 -1.44210362e+00 -4.65127319e-01
-6.17662966e-02 5.78025103e-01 3.03253651e-01 -1.87876225e-01
1.93382517e-01 -2.88737088e-01 -1.65079427e+00 1.29264212e+00
4.61926430e-01 1.07664883e+00 7.13412464e-01 1.31799757e+00
-2.05045640e-01 2.03894472e+00 -6.60575569e-01 -8.31088424e-01
-1.37891725e-01 -5.46639204e-01 -8.97144854e-01 2.54916549e-01
9.87011552e-01 -1.14320767e+00 -9.45501089e-01 7.73423553e-01
2.37638548e-01 3.26567858e-01 5.13897061e-01 -1.20994961e+00
-8.64065349e-01 5.82909346e-01 -1.32729799e-01 1.86535388e-01
1.00367260e+00 3.46273810e-01 1.05202210e+00 -1.16166103e+00
1.06750166e+00 1.39636540e+00 3.04045498e-01 5.35638750e-01
-7.78690100e-01 -3.76770735e-01 7.62160942e-02 1.16091430e+00
-1.12355185e+00 -3.83213431e-01 6.97902441e-01 -4.28139359e-01
1.25718093e+00 4.25121821e-02 9.09394443e-01 1.61432683e+00
4.35203820e-01 1.29387701e+00 4.78610754e-01 -4.53243814e-02
4.09269929e-02 -2.50339806e-01 -1.32945821e-01 9.22703087e-01
-4.68357176e-01 -3.25756103e-01 -1.08744264e+00 3.18266451e-01
5.91689885e-01 5.62413894e-02 -5.10403872e-01 -4.67999429e-01
-1.44074035e+00 1.12976313e+00 2.58846968e-01 1.22200929e-01
-5.35427928e-01 3.84023249e-01 5.50044775e-01 5.56404054e-01
3.99275273e-01 4.85636920e-01 -1.62084684e-01 -4.27154660e-01
-7.72431552e-01 3.71177167e-01 7.24900544e-01 1.10080516e+00
3.50164592e-01 1.94374084e-01 -3.90096754e-01 6.63084924e-01
1.40879661e-01 5.75280070e-01 4.05434281e-01 -1.41006494e+00
9.31897819e-01 4.13814306e-01 2.44340096e-02 -1.15658593e+00
-1.72223330e-01 2.70831585e-01 -4.98755842e-01 -4.12314981e-01
2.59609193e-01 8.90576690e-02 -8.79299521e-01 1.45041835e+00
1.08008012e-01 5.60534656e-01 4.00441051e-01 9.84156489e-01
1.63035727e+00 1.36198080e+00 -9.81354937e-02 -2.13973731e-01
1.25831652e+00 -1.65871835e+00 -1.02627337e+00 -7.07165524e-02
3.48236531e-01 -5.33175886e-01 1.11979818e+00 4.22857702e-01
-1.41531503e+00 -7.71338999e-01 -1.01508582e+00 -6.43871009e-01
-2.15086773e-01 -2.01414138e-01 1.94445312e-01 -5.27740680e-02
-1.09611678e+00 1.89011350e-01 -5.30020177e-01 -2.66442925e-01
3.09505761e-01 -1.49028182e-01 -6.45222485e-01 -8.52656722e-01
-1.27375388e+00 7.43196309e-01 4.03427094e-01 -1.18163424e-02
-1.74986410e+00 -6.51626348e-01 -1.40619648e+00 1.78406760e-01
5.17395020e-01 -8.36283565e-01 1.47539854e+00 -1.11130738e+00
-1.53399062e+00 6.34560883e-01 -1.64083004e-01 -8.35234404e-01
2.97223777e-01 -5.09972990e-01 -4.83038068e-01 1.28668249e+00
2.48916402e-01 1.11508155e+00 1.20035040e+00 -9.80072677e-01
-5.94329417e-01 -2.70887949e-02 6.09128773e-01 4.64487433e-01
-1.93223596e-01 1.04477182e-01 -6.00731730e-01 -5.70263326e-01
-2.24039227e-01 -5.66918612e-01 2.35298470e-01 -4.88811322e-02
2.59066429e-02 -1.74086124e-01 8.36799681e-01 -1.23093224e+00
8.60910475e-01 -1.84474123e+00 5.07201016e-01 -8.25160444e-01
2.28680626e-01 6.21354878e-02 -4.85822529e-01 7.37780154e-01
2.80685220e-02 -5.35651028e-01 -8.49902332e-02 -6.13407314e-01
-1.49842784e-01 1.96334094e-01 -7.78252840e-01 3.62857550e-01
3.74998987e-01 1.17880714e+00 -1.01396406e+00 -6.07633948e-01
3.86303514e-01 5.75990617e-01 -5.55012107e-01 7.69579530e-01
-7.76868165e-01 2.38859192e-01 -1.51338950e-01 6.18195057e-01
3.02864075e-01 -3.58320922e-01 -1.17901139e-01 -3.22047263e-01
6.70596212e-02 2.23475903e-01 -4.54183757e-01 2.55322075e+00
-2.77604938e-01 1.33730721e+00 -2.31420770e-01 -1.02099431e+00
3.74211907e-01 6.59828663e-01 3.81041825e-01 -8.66509676e-01
1.09295115e-01 -2.68055111e-01 -4.21374083e-01 -1.35658681e+00
7.38166988e-01 2.48445436e-01 -1.37055174e-01 3.28077853e-01
6.41506970e-01 -3.75852257e-01 4.37670350e-01 6.57357991e-01
1.22637701e+00 4.27715778e-01 -2.81096026e-02 3.97559494e-01
3.13647628e-01 3.68210047e-01 5.42102717e-02 7.52628326e-01
-1.09864049e-01 1.08232129e+00 5.38329661e-01 -6.10540569e-01
-1.05937362e+00 -1.07731414e+00 3.88920367e-01 1.27898967e+00
5.63434698e-02 -4.94286865e-01 -7.46512949e-01 -4.52436805e-01
-4.72402424e-01 1.05763638e+00 -7.20506132e-01 -1.24926075e-01
-7.94523716e-01 6.04281463e-02 3.48613918e-01 4.01531070e-01
7.95140564e-01 -1.45751655e+00 -8.63698065e-01 2.02518567e-01
-8.59941840e-01 -1.47930801e+00 -5.76559067e-01 -5.46135128e-01
-4.52173620e-01 -1.12304664e+00 -8.75005484e-01 -1.01564753e+00
3.29737097e-01 6.90871477e-01 1.39893878e+00 -3.20755124e-01
5.92302270e-02 9.97905731e-01 -7.70157874e-01 -5.16370572e-02
-3.26379389e-01 -5.47236323e-01 -2.59419680e-01 2.25638464e-01
2.30787680e-01 -3.38148952e-01 -4.70144778e-01 2.46258080e-02
-8.56382608e-01 6.70108676e-01 3.17048192e-01 9.80996907e-01
4.16114330e-01 -3.33091289e-01 3.44560713e-01 -5.07675350e-01
1.83521077e-01 -8.05853069e-01 -3.98619682e-01 4.45297867e-01
3.08176517e-01 -1.56268671e-01 5.47520101e-01 -5.47083676e-01
-1.24405348e+00 -3.45539711e-02 -1.45194307e-01 -9.91374195e-01
-8.75661820e-02 4.41186726e-01 2.26193778e-02 4.08867985e-01
2.68355638e-01 5.25412798e-01 -2.81305730e-01 -8.95558670e-02
6.14055216e-01 5.46840787e-01 9.13621247e-01 -3.11154395e-01
3.42556477e-01 4.24737781e-01 -4.50967342e-01 -9.09969747e-01
-1.08418620e+00 -4.36871350e-01 -5.43534875e-01 -6.47903621e-01
1.58364201e+00 -1.36880422e+00 -7.97592640e-01 4.13803846e-01
-1.56386733e+00 -3.21127951e-01 -1.22634329e-01 4.96850997e-01
-1.02195370e+00 3.10547262e-01 -9.27103162e-01 -4.58257735e-01
-2.61684656e-01 -1.16454363e+00 1.08750796e+00 8.35361555e-02
2.17338488e-01 -6.42685354e-01 1.36380538e-01 1.21442807e+00
-8.12842995e-02 3.69823217e-01 7.24442542e-01 -3.78431201e-01
-1.03540468e+00 4.80534844e-02 -2.76123554e-01 8.89752433e-02
-4.70115095e-01 -1.36458620e-01 -9.75909352e-01 -3.14785600e-01
5.00816464e-01 -9.73452270e-01 1.13985956e+00 5.30839443e-01
1.03841650e+00 -3.42829138e-01 2.92153627e-01 5.11274219e-01
1.27301049e+00 3.37889969e-01 8.33374560e-01 1.56530902e-01
9.67097163e-01 8.56643021e-01 9.35535610e-01 2.12764099e-01
1.06753623e+00 4.99285132e-01 7.33048499e-01 3.17502469e-01
-3.58427793e-01 -7.01362312e-01 8.60387802e-01 1.18983352e+00
2.30912030e-01 -4.88442421e-01 -8.32532108e-01 8.99185240e-01
-1.99370229e+00 -1.37336814e+00 -7.76856244e-02 1.60135889e+00
4.99922931e-01 -2.01229528e-02 -7.19286427e-02 -2.83835053e-01
6.05953455e-01 5.23224890e-01 -5.42260587e-01 -2.96869159e-01
-2.31103644e-01 -1.74965739e-01 -1.84978738e-01 3.23785096e-01
-1.08727133e+00 9.63219404e-01 5.42222738e+00 6.67610765e-01
-7.68415332e-01 3.86680394e-01 4.49330866e-01 -3.77638608e-01
-1.85131520e-01 -4.76989746e-01 -3.69741082e-01 3.17618579e-01
1.22534227e+00 1.17804728e-01 4.30852979e-01 7.30760574e-01
3.23799103e-02 -2.93577433e-01 -1.20877564e+00 1.20982659e+00
1.05528319e+00 -1.84764802e+00 4.38753963e-01 -5.98639011e-01
8.41572642e-01 4.50689979e-02 8.90178233e-02 6.77468598e-01
-6.87236786e-02 -1.11886001e+00 1.03525543e+00 6.72810733e-01
9.39362705e-01 -6.70982599e-01 6.47855759e-01 3.93928498e-01
-1.14513946e+00 -2.72443622e-01 -4.29357946e-01 -7.77615979e-02
4.14827824e-01 -2.61505842e-01 -8.19869399e-01 4.61386204e-01
1.04892099e+00 1.02494061e+00 -5.17764091e-01 7.42039800e-01
-2.04531789e-01 5.22085845e-01 1.97165281e-01 -7.89256208e-03
4.97173607e-01 2.00736672e-02 5.80686092e-01 1.01866984e+00
2.21170515e-01 4.62318957e-01 -2.78461665e-01 6.92124188e-01
-1.66460559e-01 -1.16423123e-01 -7.98587203e-01 -2.32765228e-01
6.88953251e-02 8.76505852e-01 -3.34317505e-01 -5.18124878e-01
-8.94631624e-01 1.15345407e+00 4.76169795e-01 4.74290729e-01
-1.07695258e+00 4.08401638e-02 3.62687230e-01 -2.26119563e-01
6.77236199e-01 -2.20039561e-01 4.18912947e-01 -1.55809689e+00
-6.21275045e-02 -1.10083783e+00 5.10032356e-01 -1.55572641e+00
-1.02872992e+00 7.17019141e-01 2.36391887e-01 -1.31258595e+00
-6.31433010e-01 -4.70165402e-01 -4.63346452e-01 -7.30089322e-02
-1.26485085e+00 -1.41014218e+00 -4.23915714e-01 9.44668174e-01
1.69633293e+00 -3.83938253e-01 7.60611415e-01 2.49149293e-01
-3.59235913e-01 2.55500585e-01 -2.45157275e-02 1.32692397e-01
6.72581136e-01 -1.16598690e+00 3.57934833e-01 8.85827363e-01
6.82138801e-01 2.36533079e-02 8.09963822e-01 -5.02242923e-01
-1.76585126e+00 -1.08776045e+00 7.37878144e-01 -7.39192009e-01
7.05584466e-01 -5.93166173e-01 -8.25678587e-01 9.97138143e-01
8.91291738e-01 -2.80787319e-01 4.49827164e-01 -7.03716695e-01
-3.19991648e-01 8.47817436e-02 -8.32339108e-01 5.81717193e-01
8.61789763e-01 -8.82581711e-01 -1.13886762e+00 6.11499071e-01
1.45774138e+00 -6.16864800e-01 -8.40366006e-01 1.78005114e-01
5.62278450e-01 -1.06804502e+00 1.02172244e+00 -6.08252764e-01
1.22101450e+00 -1.45671694e-02 -5.10681152e-01 -1.00402844e+00
5.74051328e-02 -3.72078061e-01 -4.67848837e-01 9.01721060e-01
4.15423624e-02 2.04834446e-01 5.36694825e-01 4.51503545e-01
-4.26783264e-01 -7.34293342e-01 -8.58713090e-01 -1.36919871e-01
-2.54994512e-01 -6.29654348e-01 2.60687292e-01 7.39686131e-01
-2.61644781e-01 6.53327525e-01 -9.80754077e-01 9.22220796e-02
4.07533675e-01 2.84202307e-01 7.75267780e-01 -4.03962791e-01
-3.15256923e-01 1.86493874e-01 -3.92080873e-01 -1.43931675e+00
3.69179636e-01 -4.20148015e-01 1.35981232e-01 -1.74472010e+00
4.22718257e-01 7.91533887e-01 3.66209708e-02 -9.78518575e-02
5.87006658e-02 2.20816672e-01 4.61461306e-01 3.21387639e-03
-1.44998896e+00 9.20557380e-01 1.36350107e+00 -5.05135834e-01
1.18366644e-01 -4.36235934e-01 4.16533127e-02 7.18382001e-01
3.07824045e-01 -3.82008195e-01 -5.80947936e-01 -1.11051393e+00
1.79972008e-01 9.76441443e-01 4.48897928e-01 -1.00852096e+00
4.38265562e-01 2.40416005e-02 4.60215479e-01 -1.20557928e+00
1.15296507e+00 -7.09522545e-01 7.65834972e-02 3.73961061e-01
-5.79907358e-01 2.53613383e-01 4.55922522e-02 7.68918216e-01
-5.80013931e-01 -1.42241776e-01 1.88025355e-01 -3.07881653e-01
-1.25339174e+00 2.56492347e-01 -5.44548213e-01 1.39605924e-01
1.12563086e+00 -2.36905292e-01 -7.04404473e-01 -9.89512861e-01
-5.64516068e-01 6.61315560e-01 9.84350964e-02 7.89737582e-01
1.39549220e+00 -1.34206080e+00 -1.03231275e+00 -1.38475979e-02
3.10362458e-01 1.14334887e-02 8.88327837e-01 4.52967018e-01
-9.53049242e-01 6.52738631e-01 -4.94550228e-01 -7.34153569e-01
-1.34985280e+00 7.88782477e-01 1.52938450e-02 2.98304874e-02
-8.54841828e-01 1.26313460e+00 4.97966290e-01 -3.29725519e-02
4.61393356e-01 -2.55673647e-01 -5.94314337e-01 2.34645262e-01
8.64805579e-01 3.04147363e-01 -5.48220694e-01 -9.98414218e-01
2.60795839e-02 4.55248207e-01 -1.36227906e-01 -2.38470480e-01
1.47504807e+00 -3.54839653e-01 1.17604416e-02 7.23121464e-01
1.37692213e+00 -5.59248924e-01 -1.59141040e+00 -1.70492291e-01
-4.80267823e-01 -6.19107425e-01 -2.27535456e-01 -4.50566322e-01
-8.59825194e-01 1.15166008e+00 3.85150939e-01 -3.99311632e-02
1.12881207e+00 1.76235229e-01 1.20391989e+00 6.86569631e-01
3.15042526e-01 -8.81236374e-01 9.16167676e-01 4.96407121e-01
1.42730296e+00 -1.41646028e+00 -2.16928408e-01 2.17371248e-02
-1.09296107e+00 1.22761440e+00 7.91831732e-01 -3.02541643e-01
1.15641683e-01 -4.06693459e-01 -1.07610330e-01 -3.61203849e-01
-1.48060107e+00 5.01750819e-02 3.67317766e-01 5.58283448e-01
-1.76197514e-01 -6.74265623e-02 2.90724456e-01 6.47260427e-01
-2.23432839e-01 -1.92740947e-01 8.47074091e-01 5.41032434e-01
-4.58381414e-01 -2.97293991e-01 -2.70321190e-01 3.48017067e-01
-2.73562849e-01 -1.20865256e-01 -1.32351279e-01 6.20281756e-01
1.58409066e-02 1.28226912e+00 2.70331740e-01 -4.98917848e-01
1.29362330e-01 -4.52025868e-02 6.54212356e-01 -6.56822324e-01
-3.17972809e-01 -1.98363781e-01 3.34931374e-01 -9.44149256e-01
-6.17019773e-01 -5.90332448e-01 -1.00719297e+00 -2.46392906e-01
9.90451798e-02 -1.08112797e-01 4.13146913e-01 1.06417871e+00
9.07559171e-02 5.47659576e-01 3.68542433e-01 -7.85494685e-01
-4.06194478e-02 -9.36662018e-01 -2.09914431e-01 4.36681420e-01
6.02312386e-01 -5.62071562e-01 -1.13526814e-01 4.18825030e-01] | [10.465916633605957, 0.9155983924865723] |
9351c643-6607-49fc-9dcc-a233eae9f0f5 | towards-adversarial-realism-and-robust | 2301.13122 | null | https://arxiv.org/abs/2301.13122v3 | https://arxiv.org/pdf/2301.13122v3.pdf | Towards Adversarial Realism and Robust Learning for IoT Intrusion Detection and Classification | The Internet of Things (IoT) faces tremendous security challenges. Machine learning models can be used to tackle the growing number of cyber-attack variations targeting IoT systems, but the increasing threat posed by adversarial attacks restates the need for reliable defense strategies. This work describes the types of constraints required for a realistic adversarial cyber-attack example and proposes a methodology for a trustworthy adversarial robustness analysis with a realistic adversarial evasion attack vector. The proposed methodology was used to evaluate three supervised algorithms, Random Forest (RF), Extreme Gradient Boosting (XGB), and Light Gradient Boosting Machine (LGBM), and one unsupervised algorithm, Isolation Forest (IFOR). Constrained adversarial examples were generated with the Adaptative Perturbation Pattern Method (A2PM), and evasion attacks were performed against models created with regular and adversarial training. Even though RF was the least affected in binary classification, XGB consistently achieved the highest accuracy in multi-class classification. The obtained results evidence the inherent susceptibility of tree-based algorithms and ensembles to adversarial evasion attacks and demonstrates the benefits of adversarial training and a security by design approach for a more robust IoT network intrusion detection and cyber-attack classification. | ['Eva Maia', 'Isabel Praça', 'João Vitorino'] | 2023-01-30 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [ 3.80605161e-01 -7.54832104e-02 1.81465298e-01 -2.36470625e-01
-2.00274527e-01 -8.48274469e-01 9.56229329e-01 1.34204645e-02
-1.78312197e-01 9.47534919e-01 -3.45629781e-01 -7.76720822e-01
-3.97247165e-01 -1.06315196e+00 -6.63150489e-01 -9.65894818e-01
-4.06316906e-01 4.02599275e-01 4.92749438e-02 -4.01032239e-01
2.19024736e-02 9.79589760e-01 -1.37989318e+00 2.40320355e-01
6.81227148e-01 1.34490347e+00 -8.05403113e-01 9.43865418e-01
3.78911495e-01 6.43496215e-01 -9.21157181e-01 -8.47965896e-01
8.06424379e-01 1.10360440e-02 -5.09894013e-01 -7.19194710e-01
1.85969576e-01 -4.01332378e-02 2.17286423e-02 1.00628984e+00
8.18641603e-01 -2.84697115e-01 7.59115279e-01 -1.89685369e+00
-2.01609299e-01 6.71025395e-01 6.16629384e-02 1.89309046e-01
5.02422452e-01 4.99009848e-01 1.65911734e-01 -4.23529558e-02
1.89091459e-01 1.31386828e+00 7.87172318e-01 7.10240304e-01
-1.26883829e+00 -1.28246939e+00 -1.83288068e-01 1.67682484e-01
-9.55940843e-01 -6.58113882e-02 9.36210811e-01 -4.39733088e-01
7.86872685e-01 6.82625175e-01 5.13802052e-01 1.78233123e+00
9.17235434e-01 -6.98608086e-02 1.76491714e+00 -3.80451173e-01
7.97009885e-01 2.44353130e-01 8.07486475e-02 9.03777778e-02
6.17430627e-01 1.04770780e+00 1.66484326e-01 -6.98091507e-01
-2.45922700e-01 1.91429183e-02 2.88885593e-01 -3.44059356e-02
-7.38776326e-01 8.68671536e-01 6.67316914e-01 3.82775456e-01
-3.73406589e-01 3.27672027e-02 7.61241317e-01 7.00770438e-01
1.39611691e-01 5.79407394e-01 -1.06798017e+00 3.10878634e-01
-3.63990366e-01 -8.08805004e-02 1.06797624e+00 4.78897303e-01
2.79196113e-01 5.74997127e-01 2.60148376e-01 1.55423179e-01
4.11383033e-01 9.28789556e-01 3.90342683e-01 -1.96256831e-01
3.42079192e-01 3.97835910e-01 -3.44022214e-02 -1.07183373e+00
-5.68369389e-01 -3.98259193e-01 -1.10544479e+00 9.72241163e-01
3.28164995e-01 -2.76125908e-01 -8.88959825e-01 1.34995270e+00
5.95975161e-01 -1.28852380e-02 4.75060225e-01 -2.13402603e-02
3.60357493e-01 3.85068983e-01 5.87252736e-01 -2.44920373e-01
1.10121250e+00 -1.92028955e-01 -4.04839367e-01 -7.48056248e-02
1.49322227e-01 -6.40738130e-01 6.28603697e-01 6.86894238e-01
-4.10417199e-01 -4.94325906e-01 -1.27545369e+00 1.21664977e+00
-9.15133595e-01 -1.04193449e+00 6.15534484e-01 1.85382509e+00
-5.07735252e-01 5.76038837e-01 -4.65383172e-01 -1.56440854e-01
5.71781874e-01 6.92638040e-01 -2.05844596e-01 4.29300033e-02
-1.36590123e+00 1.13556659e+00 3.50836068e-01 -2.23222286e-01
-1.01822257e+00 -4.03076172e-01 -4.48678672e-01 -4.14628774e-01
-4.70694721e-01 -6.53431296e-01 5.93118787e-01 -1.19913399e+00
-1.62350643e+00 4.82514799e-01 8.71740043e-01 -9.25893128e-01
7.16593027e-01 7.25661740e-02 -1.12933373e+00 -9.65788215e-02
-4.07569706e-01 5.70949242e-02 1.42956412e+00 -1.36311972e+00
-4.81566526e-02 -6.31720841e-01 8.41058642e-02 -6.80729747e-01
-3.11357498e-01 3.27483326e-01 1.37121654e+00 -8.19855511e-01
-6.13626465e-02 -1.22766757e+00 -2.76544213e-01 -5.57740390e-01
-5.04118025e-01 2.19382882e-01 1.25870287e+00 -5.65151215e-01
8.18035126e-01 -1.84203529e+00 -4.10263449e-01 6.62315547e-01
-3.84857655e-01 5.99251091e-01 1.41175807e-01 5.66638172e-01
-4.78211343e-01 5.50152719e-01 -1.60890758e-01 3.35437477e-01
1.50959969e-01 2.66736418e-01 -5.80382168e-01 4.44859862e-01
-1.69100046e-01 7.92409062e-01 -7.84211457e-01 -1.91585526e-01
4.41542387e-01 4.96307999e-01 -1.49997503e-01 5.17492183e-02
-2.64349192e-01 8.28257263e-01 -4.98577565e-01 9.79387701e-01
8.52827430e-01 4.85883832e-01 -3.50583829e-02 -3.05137485e-01
3.05292815e-01 -1.84185237e-01 -9.39618945e-01 4.38792467e-01
-7.32854366e-01 7.56526887e-02 -9.85407755e-02 -9.34673429e-01
1.14838886e+00 3.96194637e-01 3.64882737e-01 -6.06497884e-01
5.96368015e-01 2.72711694e-01 2.16047093e-01 -2.08552957e-01
-5.29910862e-01 -3.23880076e-01 -4.98511314e-01 4.14041162e-01
-1.99879929e-01 -3.63593102e-01 -4.30314392e-01 -1.07497975e-01
1.26515913e+00 -2.44964480e-01 6.41896605e-01 -3.06538105e-01
9.27938402e-01 -2.60696381e-01 2.51143098e-01 8.44201684e-01
-4.95689183e-01 -3.10655266e-01 -2.66451892e-02 -1.15481126e+00
-6.59827650e-01 -1.22492921e+00 -3.56474012e-01 8.85640025e-01
-1.39504656e-01 1.99280724e-01 -5.95106125e-01 -1.30347598e+00
2.47593775e-01 1.17398083e+00 -5.83444118e-01 -5.83895266e-01
-5.19912839e-01 -1.08821964e+00 1.05658984e+00 4.90311421e-02
8.76087070e-01 -1.10856688e+00 -5.08525550e-01 2.40764581e-02
4.17661458e-01 -9.14348304e-01 4.57943857e-01 5.87047219e-01
-8.41890454e-01 -1.12626827e+00 4.16438520e-01 -3.22285056e-01
2.93107510e-01 -5.26746392e-01 1.32686007e+00 -2.80513335e-02
-2.63995707e-01 5.11329770e-01 -6.11462355e-01 -1.12171304e+00
-1.42630768e+00 -3.99480194e-01 5.70611835e-01 -5.52559458e-02
2.52824128e-01 -1.14967024e+00 -3.99222165e-01 8.31551552e-01
-9.18348014e-01 -8.75082314e-01 5.01717687e-01 7.04343677e-01
1.56465709e-01 5.81912279e-01 8.94470036e-01 -5.70400178e-01
2.98875719e-01 -6.34145677e-01 -6.81763589e-01 1.10299960e-01
-1.08325672e+00 -3.44580263e-02 1.26156974e+00 -8.68646979e-01
-7.25742400e-01 3.11146164e-03 -4.93702233e-01 -6.52853921e-02
-6.35531783e-01 -2.16388911e-01 -6.25528872e-01 -8.89120162e-01
1.23508084e+00 -1.33397222e-01 -3.62170160e-01 -1.82386532e-01
2.48910651e-01 8.10757160e-01 -3.39244902e-02 -4.50750768e-01
1.43930733e+00 5.15937984e-01 4.60873514e-01 -7.19119906e-01
-5.30405998e-01 4.04992491e-01 -4.54540074e-01 -5.15674472e-01
7.12852240e-01 -3.67353112e-01 -8.53472114e-01 7.12318897e-01
-6.60865366e-01 -1.36376351e-01 -4.83112514e-01 1.63399965e-01
-3.49405527e-01 2.32798800e-01 -5.67338616e-02 -9.84332919e-01
-8.47875834e-01 -9.61644471e-01 3.71926278e-01 -1.52508482e-01
-1.80757493e-01 -9.20025885e-01 2.23751798e-01 4.58598346e-01
6.87426567e-01 1.24315703e+00 9.33788419e-01 -1.25472796e+00
-1.71323657e-01 -8.19178879e-01 5.46617508e-01 6.92549527e-01
8.17479659e-03 1.05744958e-01 -1.15916371e+00 -4.74203259e-01
5.53175092e-01 -4.26454633e-01 1.05544522e-01 -5.79418754e-03
7.32668579e-01 -8.85684252e-01 -1.57349691e-01 6.46384597e-01
1.44248044e+00 5.30653000e-01 7.91749418e-01 7.33906269e-01
3.44041288e-01 3.02322507e-01 3.83489192e-01 2.20335171e-01
-4.46979314e-01 6.43083036e-01 1.22747207e+00 3.60412627e-01
3.02263677e-01 -3.05453055e-02 5.66976368e-01 1.29021600e-01
-5.08889975e-03 -1.53559640e-01 -9.19962525e-01 -3.32085907e-01
-1.22024322e+00 -1.22867906e+00 -6.45674020e-02 2.30256009e+00
3.57264489e-01 6.66481316e-01 2.47529432e-01 9.37286019e-01
6.04697824e-01 -1.36582702e-02 -6.56198621e-01 -1.13442218e+00
-1.40267968e-01 5.49117506e-01 7.14366019e-01 2.74789184e-01
-1.42338479e+00 3.91644895e-01 6.39938736e+00 5.91101050e-01
-1.07202125e+00 2.24147096e-01 6.38655484e-01 2.63780862e-01
9.02293772e-02 -2.28224412e-01 -4.83227223e-01 7.17378974e-01
1.48900998e+00 1.59402981e-01 5.08932471e-01 1.25785422e+00
7.80051351e-02 4.86985922e-01 -7.64406979e-01 5.20603180e-01
-2.39901796e-01 -7.60998726e-01 1.71154812e-01 3.83122116e-02
8.03412497e-01 1.38797075e-01 1.24814764e-01 3.68564725e-01
7.96197951e-01 -9.91113305e-01 4.81484383e-01 3.68648231e-01
4.73243475e-01 -1.01511240e+00 9.85479653e-01 3.47796440e-01
-8.85123372e-01 -7.77617633e-01 -1.35110971e-02 -2.10487723e-01
-2.70130426e-01 6.26423836e-01 -7.54428923e-01 5.79603434e-01
8.85593951e-01 4.66791261e-03 -6.80195212e-01 4.36787903e-01
-3.01023960e-01 8.45312536e-01 -5.71421266e-01 -4.06570584e-01
-1.51655570e-01 8.37773532e-02 8.73731077e-01 1.07219315e+00
-3.58544998e-02 -2.65007436e-01 9.01455656e-02 2.11884961e-01
3.71898860e-01 -1.83159318e-02 -9.11898911e-01 2.64985472e-01
4.91042227e-01 1.09484363e+00 -7.47147441e-01 -2.72817984e-02
3.72252390e-02 6.32442594e-01 -4.61770117e-01 6.72607571e-02
-9.12209570e-01 -1.02262862e-01 7.02279091e-01 -8.94386619e-02
9.02420878e-02 1.15078561e-01 -6.02400899e-01 -8.12713861e-01
-2.55751640e-01 -1.36630034e+00 6.84787989e-01 -5.29219508e-01
-1.77361310e+00 8.22606325e-01 1.22866392e-01 -1.41572940e+00
-3.76909912e-01 -7.48531163e-01 -9.28449512e-01 5.22255063e-01
-8.37024271e-01 -1.56725407e+00 -2.23408043e-01 1.04302669e+00
3.93939130e-02 -6.12785518e-01 1.30889440e+00 -6.10949695e-02
-2.48741582e-01 7.37279296e-01 3.75167876e-02 -1.60399333e-01
-4.01140470e-03 -9.99349654e-01 4.89297301e-01 8.74066651e-01
-1.04468144e-01 3.88315886e-01 1.08570802e+00 -7.64818490e-01
-1.22699451e+00 -1.31404352e+00 2.68516928e-01 -6.47692621e-01
7.47462332e-01 -3.14254880e-01 -1.78278342e-01 5.23084641e-01
-3.95645611e-02 1.64861351e-01 9.11273062e-01 -3.61689299e-01
-8.19511175e-01 -5.36013842e-01 -2.31716633e+00 5.28617024e-01
6.41970813e-01 -1.82944819e-01 -5.01631856e-01 5.43584287e-01
7.18152761e-01 3.74600649e-01 -1.04741704e+00 7.60030389e-01
7.14390457e-01 -1.18750000e+00 1.40631902e+00 -7.61771142e-01
-2.43605286e-01 -3.18940639e-01 -5.05749166e-01 -1.11889696e+00
-2.35108137e-02 -9.79689121e-01 -4.45468277e-01 1.20187664e+00
1.87494531e-01 -1.27235651e+00 5.00689328e-01 3.66151601e-01
3.63594234e-01 -5.01178622e-01 -1.29276586e+00 -1.13260150e+00
3.79297316e-01 -7.46963561e-01 9.02617633e-01 9.29371595e-01
-2.99805671e-01 4.98940386e-02 -1.93629399e-01 4.61732835e-01
8.52379978e-01 -5.22870898e-01 9.06680524e-01 -1.38613677e+00
-3.39146644e-01 -1.17728576e-01 -1.12484586e+00 4.56775457e-01
7.22777396e-02 -6.94962740e-01 -5.01305640e-01 -7.66182423e-01
-6.61824226e-01 -7.09928930e-01 -4.94167209e-01 3.12888652e-01
4.67120633e-02 5.11211157e-01 1.90783754e-01 5.43991178e-02
1.51468769e-01 -5.85875707e-03 5.06214797e-01 -4.25596207e-01
-5.99309281e-02 8.94062757e-01 -5.13718188e-01 1.00452662e+00
1.31644762e+00 -8.82623017e-01 -2.70125002e-01 5.15969157e-01
3.37270796e-01 -4.00639713e-01 6.62305951e-01 -1.40086484e+00
-2.57347822e-01 -4.48315777e-03 5.94626486e-01 -1.94640175e-01
3.52372602e-02 -1.59791112e+00 6.68011189e-01 1.37657297e+00
2.00751990e-01 2.93514431e-01 3.16442817e-01 5.97487152e-01
4.34461504e-01 -1.16908543e-01 1.27369189e+00 6.31457195e-02
-2.79521868e-02 9.91740599e-02 -3.51199448e-01 -3.00188333e-01
1.44124842e+00 -3.30963314e-01 -2.42676765e-01 -1.79714829e-01
-7.83435822e-01 -4.95471835e-01 5.30472636e-01 1.37926057e-01
3.53447974e-01 -1.16196215e+00 -6.10646904e-01 5.09205461e-01
-1.04365990e-01 -7.46678054e-01 -1.29837140e-01 2.90255457e-01
-4.71279770e-01 -1.71733811e-01 -5.20743191e-01 -3.74371469e-01
-1.45150733e+00 1.28297019e+00 4.99689728e-01 -6.14368916e-01
-2.68536776e-01 5.95353067e-01 -6.09155059e-01 -6.12382710e-01
2.20036849e-01 2.46119440e-01 -2.00655937e-01 -3.37728381e-01
4.13987339e-01 8.67515624e-01 4.47015494e-01 -6.15890861e-01
-6.41705215e-01 4.71570581e-01 2.13918447e-01 2.19847932e-01
1.10701454e+00 3.73509526e-01 -7.80173466e-02 2.80432813e-02
8.89702439e-01 2.51864702e-01 -6.14990711e-01 5.28984845e-01
2.20348742e-02 -2.19518155e-01 -3.47933024e-01 -1.43984377e+00
-7.99111903e-01 4.11670715e-01 1.12283385e+00 1.06083226e+00
1.28004313e+00 -4.69122261e-01 4.62494165e-01 6.03134990e-01
8.22580993e-01 -5.62596202e-01 -8.32413137e-02 1.95160910e-01
8.84219229e-01 -1.03600812e+00 2.70060748e-01 -2.49332860e-01
-1.28969178e-01 1.03669167e+00 1.84687242e-01 -3.61939222e-01
1.28476942e+00 6.37359440e-01 1.91077441e-01 1.32370666e-01
-2.31472358e-01 3.98527414e-01 -4.52621952e-02 1.41536641e+00
-4.39956725e-01 2.54508644e-01 -2.13366106e-01 3.85886610e-01
-5.23773611e-01 -3.69661421e-01 1.72202513e-01 9.92674768e-01
-2.45821830e-02 -1.26160562e+00 -9.75112915e-01 4.77628499e-01
-6.90420985e-01 7.57443011e-02 -7.10957408e-01 9.52406287e-01
5.39143682e-01 1.45971107e+00 -6.29685640e-01 -1.08280981e+00
3.37623000e-01 2.48120859e-01 2.61986703e-01 9.40219834e-02
-1.41608870e+00 -8.59332800e-01 1.29296929e-01 -5.49097002e-01
-3.28908622e-01 -4.91491139e-01 -4.55584466e-01 -5.79325318e-01
-4.51911330e-01 3.59366350e-02 1.17106092e+00 9.11432087e-01
8.94981995e-02 4.46648479e-01 1.27126169e+00 -1.02481318e+00
-1.04645967e+00 -9.12684083e-01 -1.85818136e-01 5.39404809e-01
7.43817538e-02 -6.32649362e-01 -1.19570839e+00 -1.34046704e-01] | [5.541747570037842, 7.571475982666016] |
c8c73c90-8dc0-4fdf-90f4-e4b36a02344f | layout-and-task-aware-instruction-prompt-for | 2306.00526 | null | https://arxiv.org/abs/2306.00526v2 | https://arxiv.org/pdf/2306.00526v2.pdf | Layout and Task Aware Instruction Prompt for Zero-shot Document Image Question Answering | The pre-training-fine-tuning paradigm based on layout-aware multimodal pre-trained models has achieved significant progress on document image question answering. However, domain pre-training and task fine-tuning for additional visual, layout, and task modules prevent them from directly utilizing off-the-shelf instruction-tuning language foundation models, which have recently shown promising potential in zero-shot learning. Contrary to aligning language models to the domain of document image question answering, we align document image question answering to off-the-shell instruction-tuning language foundation models to utilize their zero-shot capability. Specifically, we propose layout and task aware instruction prompt called LATIN-Prompt, which consists of layout-aware document content and task-aware descriptions. The former recovers the layout information among text segments from OCR tools by appropriate spaces and line breaks. The latter ensures that the model generates answers that meet the requirements, especially format requirements, through a detailed description of task. Experimental results on three benchmarks show that LATIN-Prompt can improve the zero-shot performance of instruction-tuning language foundation models on document image question answering and help them achieve comparable levels to SOTAs based on the pre-training-fine-tuning paradigm. Quantitative analysis and qualitative analysis demonstrate the effectiveness of LATIN-Prompt. We provide the code in supplementary and will release the code to facilitate future research. | ['Yin Zhang', 'Yixin Ou', 'Yunhao Li', 'Wenjin Wang'] | 2023-06-01 | null | null | null | null | ['optical-character-recognition'] | ['computer-vision'] | [ 2.92990178e-01 1.17167115e-01 -3.73961300e-01 -3.79907936e-01
-1.17401814e+00 -7.92553067e-01 6.38178110e-01 1.47389635e-01
-4.49291378e-01 1.72740445e-02 4.13503617e-01 -7.44260252e-01
-4.48587835e-02 -5.76572239e-01 -8.86913300e-01 -6.59312829e-02
7.73620605e-01 3.88719440e-01 4.32332039e-01 -4.08101618e-01
5.77827573e-01 -1.41757458e-01 -1.79250908e+00 9.77832854e-01
1.03661036e+00 7.84687340e-01 6.85736835e-01 1.19629014e+00
-9.76047039e-01 9.77795541e-01 -7.76112676e-01 -1.35761797e-01
-3.07764858e-01 -4.62098241e-01 -9.15543497e-01 1.60648033e-01
1.02352905e+00 -7.20944464e-01 -2.70647049e-01 6.45271599e-01
4.74137127e-01 2.70063818e-01 7.95531034e-01 -1.08570075e+00
-1.40266466e+00 4.70563471e-01 -4.33453649e-01 1.82404563e-01
5.14478922e-01 5.15629411e-01 9.62878823e-01 -1.01639485e+00
6.59659982e-01 1.21082699e+00 1.97716206e-01 7.56461859e-01
-1.06232905e+00 -4.21290308e-01 1.50174692e-01 3.97783816e-01
-9.95725989e-01 -3.80306065e-01 6.46157980e-01 -6.58194959e-01
1.47223997e+00 1.78508654e-01 -2.62571834e-02 1.12517333e+00
1.25009030e-01 1.08724213e+00 6.70050323e-01 -7.58081019e-01
1.52242616e-01 3.17430705e-01 5.96634328e-01 8.86452913e-01
-1.60617083e-01 -3.34054649e-01 -5.65261006e-01 2.52409816e-01
3.37669969e-01 -3.31242420e-02 -2.83492684e-01 -5.64897120e-01
-8.59275162e-01 6.99948490e-01 1.75623849e-01 1.49893627e-01
1.70000821e-01 1.02058135e-01 6.15220547e-01 1.44154549e-01
5.09131476e-02 6.28957510e-01 -4.02940840e-01 -2.56562710e-01
-9.53588307e-01 -1.26859888e-01 5.79515696e-01 1.51698601e+00
9.69247162e-01 1.71632767e-02 -9.30921316e-01 8.25965464e-01
1.29058838e-01 6.05206668e-01 5.31886995e-01 -1.00711620e+00
7.22100675e-01 9.72084463e-01 -1.54924393e-01 -6.07156098e-01
-2.24107206e-01 -1.78337455e-01 -3.37270021e-01 -1.96816787e-01
4.09601659e-01 7.56355524e-02 -1.09887815e+00 1.32207656e+00
-4.34340015e-02 -3.84131223e-01 7.63926702e-03 7.50421345e-01
1.18223107e+00 9.74937022e-01 3.15631181e-01 3.51253957e-01
1.80293107e+00 -1.42025256e+00 -9.73163903e-01 -4.43209738e-01
1.09445775e+00 -9.03199375e-01 2.32503629e+00 5.45696244e-02
-8.74292076e-01 -1.07416236e+00 -1.14311314e+00 -6.10918820e-01
-6.88751340e-01 3.66163433e-01 2.49567240e-01 7.74102449e-01
-8.79702985e-01 -9.89323929e-02 -3.62258852e-01 -6.10070229e-01
2.49961719e-01 -1.44451275e-01 -1.05716651e-02 -4.97139603e-01
-7.99564302e-01 6.90457523e-01 2.04958200e-01 -5.78803003e-01
-1.01267731e+00 -1.13859594e+00 -1.12315893e+00 2.95755476e-01
7.10253537e-01 -7.58160174e-01 1.71432829e+00 -7.17179894e-01
-1.44255507e+00 8.22250783e-01 -2.65952766e-01 1.43291131e-01
2.88313888e-02 -3.34885001e-01 -1.61431581e-01 5.21988451e-01
7.92184696e-02 9.47036624e-01 9.89209890e-01 -1.34586978e+00
-4.60932702e-01 -2.23604426e-01 2.70577371e-01 -1.86102949e-02
-8.08708489e-01 -4.44409326e-02 -1.07109022e+00 -3.88781071e-01
-4.72125381e-01 -4.63428289e-01 2.80398786e-01 -4.33629639e-02
-2.43541405e-01 -2.28688538e-01 1.18625104e+00 -6.60932779e-01
1.55465114e+00 -2.11474276e+00 -2.97415346e-01 -2.55737245e-01
-5.85920177e-02 3.89375269e-01 -8.14029694e-01 5.58427036e-01
8.14549439e-03 1.48534060e-01 2.28740629e-02 -2.20114544e-01
2.81311899e-01 1.33208126e-01 -5.59651613e-01 -5.89856617e-02
2.84578741e-01 1.38744652e+00 -8.61120582e-01 -6.95983589e-01
4.01949078e-01 2.23625869e-01 -8.16112936e-01 5.19660652e-01
-7.96294510e-01 -2.14232147e-01 -3.12652797e-01 7.60185301e-01
4.37181026e-01 -4.97760862e-01 -6.19254000e-02 -3.63483608e-01
3.96872172e-03 1.97407454e-02 -6.82806194e-01 2.11955142e+00
-8.40397358e-01 8.42208445e-01 -1.48743972e-01 -5.19662082e-01
9.10049438e-01 2.29053557e-01 -2.60216370e-02 -1.29844141e+00
3.58367106e-03 -6.49171248e-02 -3.24359715e-01 -1.13362777e+00
9.06443715e-01 4.75752890e-01 -2.05493957e-01 5.82123578e-01
2.76222289e-01 -2.31320322e-01 5.22789598e-01 7.52253294e-01
9.95058596e-01 3.52638870e-01 -9.51462239e-02 -1.40144676e-01
5.23098111e-01 4.60263222e-01 -4.11831111e-01 8.26094270e-01
-2.14062795e-01 4.98606056e-01 2.95737833e-01 6.37117550e-02
-1.17233431e+00 -1.02010953e+00 2.55536705e-01 1.96669936e+00
-2.73699444e-02 -5.19296408e-01 -1.13004255e+00 -8.65113318e-01
-1.81338787e-01 1.16495895e+00 -6.16771579e-01 -2.30393305e-01
-4.42522883e-01 1.43937925e-02 4.68330383e-01 8.57474804e-01
4.28855449e-01 -8.47270966e-01 -8.13953817e-01 -2.11721137e-02
-3.64361167e-01 -1.18133223e+00 -1.01236629e+00 2.88768291e-01
-6.79334402e-01 -9.93647754e-01 -8.75272632e-01 -1.04000449e+00
8.52391005e-01 7.93665767e-01 1.36841869e+00 3.64507198e-01
-4.53384429e-01 1.14133525e+00 -6.56044602e-01 -2.27956444e-01
-2.78106302e-01 1.70940906e-01 -6.19724691e-01 -4.48852539e-01
6.65961146e-01 5.49823530e-02 -6.10890865e-01 3.99169415e-01
-1.34615850e+00 1.23008229e-01 6.34355307e-01 7.07863688e-01
3.57367188e-01 -4.41414833e-01 4.45331395e-01 -7.01287091e-01
9.12436783e-01 -1.72234833e-01 -3.99019778e-01 7.95181274e-01
-4.81378645e-01 4.44796324e-01 5.00944376e-01 -5.21213651e-01
-1.20652294e+00 -6.04244620e-02 4.78559323e-02 -4.65238094e-01
-3.20804626e-01 2.65132993e-01 -2.92563766e-01 3.18513632e-01
9.27314103e-01 2.90778786e-01 -3.60305786e-01 -5.09404063e-01
8.29736292e-01 8.92173469e-01 8.45880628e-01 -9.33924615e-01
6.31091237e-01 2.48038694e-01 -6.45751178e-01 -8.78506541e-01
-1.10581136e+00 -9.30069029e-01 -5.42209327e-01 -2.44019330e-01
1.24547386e+00 -7.47246981e-01 -6.07707322e-01 5.74686099e-03
-1.40788615e+00 -6.24027669e-01 -4.12224025e-01 -1.21821970e-01
-5.71615875e-01 3.63993406e-01 -5.99821448e-01 -6.62809610e-01
-2.61309803e-01 -1.28024113e+00 1.26480007e+00 3.71014506e-01
-2.92832047e-01 -9.47900355e-01 2.19000001e-02 7.92948008e-01
4.02114481e-01 -3.36792052e-01 1.53842330e+00 -5.00040233e-01
-5.91866493e-01 1.00775287e-01 -6.48505926e-01 2.15491399e-01
-1.88087523e-01 -7.53662214e-02 -1.20204270e+00 -1.25561759e-01
-2.50851840e-01 -6.63006485e-01 7.97590375e-01 2.95285821e-01
1.38444018e+00 -7.00660944e-02 -1.25689477e-01 5.44569790e-01
1.45504868e+00 1.32898062e-01 5.33580244e-01 5.98053873e-01
7.82964349e-01 8.63360047e-01 7.87289262e-01 3.44698608e-01
7.36222684e-01 5.50986648e-01 3.38101983e-01 -3.34292166e-02
-6.59547567e-01 -7.44546235e-01 3.57555240e-01 7.71449625e-01
5.80028117e-01 -2.89539367e-01 -1.16426659e+00 7.60824323e-01
-1.77647519e+00 -7.44767487e-01 -2.77611375e-01 1.80619216e+00
8.24357629e-01 -1.53004974e-01 -1.08754486e-01 -2.61488199e-01
3.28310877e-01 3.07157189e-02 -5.65869272e-01 -8.69059503e-01
1.14498362e-01 9.71863419e-03 2.93094784e-01 4.98762310e-01
-6.71012044e-01 1.20892060e+00 5.92344856e+00 9.74837482e-01
-7.74647355e-01 2.37688333e-01 4.18329358e-01 1.76352803e-02
-7.59051383e-01 -1.57545216e-03 -1.17481542e+00 8.21573138e-02
1.04364944e+00 -6.69909790e-02 2.31819436e-01 9.85322177e-01
1.84538499e-01 -2.52140254e-01 -1.35141706e+00 1.00410998e+00
5.19661129e-01 -1.58550179e+00 4.36198235e-01 -3.49044316e-02
8.06230545e-01 -4.44205642e-01 3.39839339e-01 8.73183310e-01
7.22314045e-02 -9.24691856e-01 7.19392300e-01 4.30104166e-01
1.14957273e+00 -6.62213206e-01 3.59325051e-01 4.88487422e-01
-1.08447075e+00 -2.25927576e-01 -4.36686426e-01 1.56494841e-01
1.16127491e-01 1.02260113e-01 -9.05339003e-01 2.99990594e-01
7.20008850e-01 1.43619403e-01 -1.00231743e+00 6.92690074e-01
-2.95703918e-01 4.98942465e-01 3.99977416e-01 -2.46709555e-01
4.63388592e-01 3.01469117e-01 1.05506517e-02 1.65161204e+00
1.80073529e-01 -1.13995105e-01 1.81148618e-01 9.02690351e-01
-9.34983194e-02 3.21939230e-01 -4.33717936e-01 -4.40256953e-01
2.54267573e-01 1.11306512e+00 -4.55981463e-01 -5.73052466e-01
-7.60984659e-01 1.08640862e+00 2.50408798e-01 6.26270056e-01
-7.44836152e-01 -6.49605036e-01 3.87743086e-01 3.89663428e-01
3.61807555e-01 -4.08014059e-01 -4.15338099e-01 -9.94647861e-01
-1.66181400e-01 -1.20003986e+00 3.93471479e-01 -1.20665431e+00
-9.27650034e-01 4.54144061e-01 1.01830751e-01 -8.90926957e-01
-2.62418658e-01 -1.02067292e+00 -4.96677369e-01 6.52169764e-01
-1.64044881e+00 -1.40005577e+00 -5.84451556e-01 7.48371303e-01
1.27153289e+00 -2.12659553e-01 8.89943004e-01 3.82471941e-02
-5.41382372e-01 8.65799069e-01 7.12746307e-02 1.83637992e-01
1.03166831e+00 -1.26263559e+00 2.68061042e-01 7.58280635e-01
1.75686866e-01 7.36848116e-01 6.13235176e-01 -6.44529760e-01
-1.86330914e+00 -7.97914028e-01 6.96503699e-01 -7.37948477e-01
8.03679168e-01 -5.05541742e-01 -1.03044832e+00 3.97171140e-01
6.78385794e-01 -3.60227168e-01 9.18480337e-01 8.35031644e-02
-8.78369510e-01 1.41759276e-01 -6.51130319e-01 6.10876739e-01
6.27798796e-01 -9.60932851e-01 -8.83929729e-01 2.23018751e-01
1.00426447e+00 -1.26225352e-01 -5.29838383e-01 9.57436934e-02
3.56232107e-01 -6.41815603e-01 8.57370317e-01 -5.98049283e-01
8.16745818e-01 -3.03750128e-01 -3.41469705e-01 -9.19071198e-01
-2.11736619e-01 -3.23207945e-01 -4.41026479e-01 1.26282632e+00
3.81301761e-01 2.51473427e-01 7.53859699e-01 6.43895149e-01
-4.49717849e-01 -6.47968233e-01 -2.85173029e-01 -8.86647105e-01
9.35301557e-02 -5.89635968e-01 3.64998639e-01 5.75790942e-01
8.05086270e-02 7.85963535e-01 3.74151953e-02 -2.25310549e-01
3.56895119e-01 1.04512684e-01 1.13414085e+00 -6.31842315e-01
-5.18934846e-01 -5.83172560e-01 4.04419154e-01 -1.65142179e+00
1.73858538e-01 -7.57434070e-01 2.93778926e-01 -1.96617985e+00
5.68754077e-01 1.82808444e-01 3.23942974e-02 5.18926561e-01
-3.98266703e-01 2.41596857e-03 3.87514949e-01 -1.49291800e-02
-1.03457117e+00 4.28691030e-01 1.57493806e+00 -5.88028491e-01
-2.06141591e-01 -6.04664087e-01 -8.75663579e-01 4.90156293e-01
4.32698637e-01 -8.04605260e-02 -8.24572682e-01 -9.28982019e-01
6.78663775e-02 -1.68502912e-01 1.05568491e-01 -9.00134623e-01
4.51456070e-01 -9.38518420e-02 2.95461476e-01 -7.97388077e-01
2.39803314e-01 -7.42819488e-01 -9.31940258e-01 3.26501012e-01
-6.51858747e-01 8.72057229e-02 6.48913145e-01 3.65711004e-01
-1.49389729e-01 -6.72381878e-01 5.61197519e-01 -2.45753229e-02
-1.36611605e+00 -4.26486284e-02 -4.82264936e-01 3.40439916e-01
9.19575751e-01 -4.38754231e-01 -9.64246750e-01 -5.37009835e-01
-1.01206832e-01 4.04215574e-01 5.09603620e-01 7.92558014e-01
8.57473791e-01 -9.57835734e-01 -5.10733664e-01 7.65588805e-02
8.95639300e-01 -3.95817429e-01 6.51865602e-01 4.05442446e-01
-3.36178422e-01 8.90409350e-01 -1.88255087e-01 -7.07180142e-01
-1.34704411e+00 1.06582785e+00 -6.55992851e-02 -3.39590997e-01
-3.49674761e-01 8.12218487e-01 6.82591558e-01 -4.66206014e-01
5.36869109e-01 -3.93794298e-01 -2.66169254e-02 -6.47114888e-02
9.49945807e-01 1.95915326e-01 3.98684293e-02 -9.07963142e-02
-6.64959848e-02 5.98880768e-01 -1.92240328e-01 -2.36863628e-01
8.28641355e-01 -3.16833615e-01 2.88638175e-01 3.71744156e-01
1.32145536e+00 -1.04238026e-01 -1.35988235e+00 -2.11594716e-01
2.86113828e-01 -2.03264341e-01 1.83996439e-01 -1.18659353e+00
-3.79480034e-01 1.46564102e+00 5.03231287e-01 -2.57082701e-01
1.14303195e+00 1.60320804e-01 8.37095737e-01 6.16523206e-01
-6.62808791e-02 -1.43918943e+00 9.59104240e-01 8.80802691e-01
7.54928291e-01 -1.23936927e+00 -3.45144987e-01 -1.52250648e-01
-6.31449819e-01 1.03436875e+00 1.24945581e+00 2.83174545e-01
1.08609714e-01 2.90771931e-01 1.96731195e-01 -8.97595212e-02
-9.67535734e-01 -4.35949236e-01 7.01883674e-01 9.24436450e-01
6.53840065e-01 -2.57486433e-01 1.91856354e-01 8.03896546e-01
-8.70849658e-03 -1.81190565e-01 3.54578435e-01 9.90661085e-01
-9.13828969e-01 -1.01320183e+00 -5.62174916e-01 1.78191602e-01
1.15345754e-01 -3.93328816e-01 -3.77201915e-01 7.79604495e-01
3.42068784e-02 1.18478537e+00 3.35009634e-01 -2.28375956e-01
6.28272653e-01 3.47754776e-01 5.59800446e-01 -9.13335919e-01
-6.06706798e-01 -8.47672597e-02 -1.86166316e-02 -6.26133800e-01
2.17133686e-01 -9.00741220e-02 -1.29470754e+00 -1.07414328e-01
-9.41745415e-02 -1.42936241e-02 8.55066240e-01 9.04871583e-01
5.25943160e-01 9.41065013e-01 3.84893827e-02 -4.26165849e-01
-6.29059076e-01 -1.02424037e+00 3.20890583e-02 3.38185877e-01
1.15930676e-01 -3.14962357e-01 -9.76662338e-03 3.86540383e-01] | [11.263021469116211, 1.9080158472061157] |
713d9cf9-6a85-493d-a5b8-7c7be4aba458 | it-takes-two-to-tango-navigating | 2305.09022 | null | https://arxiv.org/abs/2305.09022v1 | https://arxiv.org/pdf/2305.09022v1.pdf | It Takes Two to Tango: Navigating Conceptualizations of NLP Tasks and Measurements of Performance | Progress in NLP is increasingly measured through benchmarks; hence, contextualizing progress requires understanding when and why practitioners may disagree about the validity of benchmarks. We develop a taxonomy of disagreement, drawing on tools from measurement modeling, and distinguish between two types of disagreement: 1) how tasks are conceptualized and 2) how measurements of model performance are operationalized. To provide evidence for our taxonomy, we conduct a meta-analysis of relevant literature to understand how NLP tasks are conceptualized, as well as a survey of practitioners about their impressions of different factors that affect benchmark validity. Our meta-analysis and survey across eight tasks, ranging from coreference resolution to question answering, uncover that tasks are generally not clearly and consistently conceptualized and benchmarks suffer from operationalization disagreements. These findings support our proposed taxonomy of disagreement. Finally, based on our taxonomy, we present a framework for constructing benchmarks and documenting their limitations. | ['Su Lin Blodgett', 'Hal Daumé III', 'Xingdi Yuan', 'Arjun Subramonian'] | 2023-05-15 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [ 1.83820739e-01 4.37810451e-01 -7.98813105e-01 -5.11286974e-01
-1.10418952e+00 -9.86559749e-01 6.79903328e-01 4.76316094e-01
-5.99287450e-01 6.77686512e-01 9.11241710e-01 -5.57402968e-01
-6.19210720e-01 -2.07563117e-01 -4.70469713e-01 5.33292517e-02
9.43922222e-01 4.67223793e-01 -2.35284612e-01 1.06731176e-01
8.58863652e-01 1.79374948e-01 -1.36622989e+00 4.69539583e-01
1.31274998e+00 4.60470170e-01 1.96053132e-01 4.58233654e-01
-5.78359187e-01 1.03694212e+00 -8.79137039e-01 -7.22210169e-01
-5.41601330e-02 -1.49349213e-01 -1.43624222e+00 -3.92109722e-01
6.40554726e-01 -5.39663769e-02 5.53046763e-02 1.00697088e+00
4.34843987e-01 -2.17822373e-01 6.42606020e-01 -1.41361105e+00
-9.97767091e-01 7.70543993e-01 1.46085128e-01 4.39158589e-01
6.24355912e-01 1.74821809e-01 1.24953568e+00 -5.50413787e-01
8.54664743e-01 1.19631886e+00 1.01201105e+00 4.84339565e-01
-1.44994915e+00 -9.73416030e-01 1.20196037e-01 1.56843707e-01
-9.82975781e-01 -8.78650904e-01 2.00295389e-01 -1.00720203e+00
1.11390829e+00 3.56506914e-01 4.54888284e-01 1.11317205e+00
1.67151123e-01 2.01767460e-01 1.46389568e+00 -4.39035952e-01
-1.20343929e-02 3.16029072e-01 6.62363648e-01 1.25997029e-02
9.02271330e-01 -1.52839437e-01 -7.42666721e-01 -5.77192724e-01
3.91077608e-01 -4.14496601e-01 -3.22756976e-01 3.94497216e-02
-1.21477163e+00 6.26058638e-01 -3.81374061e-02 6.30169690e-01
-3.62685442e-01 -4.72322553e-02 3.07241082e-01 4.11813349e-01
2.17959553e-01 1.20709193e+00 -4.74796265e-01 -7.62326896e-01
-9.22160625e-01 4.36880827e-01 1.25766814e+00 7.09457338e-01
4.77929533e-01 -3.36757660e-01 -4.85749543e-01 8.60193372e-01
2.82913119e-01 3.53577614e-01 3.45731288e-01 -1.68728006e+00
7.90293097e-01 7.64808595e-01 8.23382258e-01 -1.04611135e+00
-6.34392798e-01 -1.97318241e-01 -2.39036027e-02 -1.34266883e-01
5.95372915e-01 -2.75707841e-01 -3.80880535e-01 2.01041746e+00
-3.65905762e-01 -3.99115980e-01 8.87740031e-02 8.14586222e-01
8.88392627e-01 2.29403879e-02 6.12276077e-01 -4.08448756e-01
1.35004175e+00 -8.16577971e-01 -9.44106996e-01 -5.13854802e-01
1.03820789e+00 -8.55626941e-01 1.55531991e+00 1.15304999e-01
-1.43729365e+00 -6.01503491e-01 -7.69672453e-01 -4.70229775e-01
-8.46025497e-02 -3.97514515e-02 6.97316706e-01 7.87043333e-01
-9.81078684e-01 4.57657844e-01 -5.86644828e-01 -6.18982911e-01
-1.02572754e-01 -1.85938001e-01 -2.00546816e-01 2.78475940e-01
-1.00068450e+00 1.29428113e+00 4.33686346e-01 -1.32068783e-01
-4.00750518e-01 -8.43445480e-01 -3.49873304e-01 5.99757582e-02
1.95571616e-01 -9.32736099e-01 1.50737500e+00 -9.05851543e-01
-1.07327926e+00 8.27863514e-01 -4.05079365e-01 -2.17828616e-01
9.40828547e-02 -3.09355825e-01 -4.70824689e-01 -3.80766429e-02
4.17497307e-01 3.74830008e-01 -2.23011654e-02 -1.15432823e+00
-1.06299961e+00 -2.81832099e-01 3.39581460e-01 5.70563316e-01
-3.24177414e-01 3.50216359e-01 7.14454204e-02 -2.34450430e-01
2.18826205e-01 -7.88781226e-01 2.41721138e-01 -4.48865414e-01
1.87365152e-02 -5.56982160e-01 2.59352416e-01 -4.82768208e-01
1.65990436e+00 -1.76273859e+00 -1.53719127e-01 -2.97893703e-01
6.63915753e-01 -1.05018422e-01 -1.88917760e-02 7.35976756e-01
1.85606733e-01 7.42554545e-01 6.52466118e-02 4.78828652e-03
2.43728548e-01 9.21993777e-02 -4.19943392e-01 3.19989473e-01
-1.45689785e-01 9.36959207e-01 -1.07886267e+00 -4.77259606e-01
-7.67783895e-02 4.39444892e-02 -5.14182925e-01 2.50623412e-02
-1.06684543e-01 3.62848401e-01 -3.88597757e-01 5.72015524e-01
4.98643458e-01 -5.54493845e-01 4.49495047e-01 -2.08332285e-01
-5.90413034e-01 1.21610928e+00 -5.33363402e-01 1.72471380e+00
-1.42711386e-01 7.69835532e-01 1.25632837e-01 -8.19259882e-01
8.85056615e-01 6.10971212e-01 2.44819194e-01 -6.88573003e-01
-3.38760674e-01 3.08154076e-01 5.50136566e-01 -7.00976253e-01
5.68966985e-01 -1.99086621e-01 -5.31678945e-02 7.47858346e-01
-1.06341943e-01 -4.36813354e-01 1.74142450e-01 1.26729339e-01
1.18137074e+00 -1.17557503e-01 2.89494187e-01 -6.19928420e-01
2.13232800e-01 5.36529839e-01 8.71709883e-01 1.14224541e+00
-5.96937776e-01 -2.09533237e-02 7.10092187e-01 -1.23747192e-01
-9.47691202e-01 -9.24520791e-01 -3.24397415e-01 1.06436729e+00
-7.88596570e-02 -6.83929265e-01 -6.21291041e-01 -6.37136698e-01
2.67778318e-02 1.18804574e+00 -5.37289560e-01 2.14307383e-03
-2.28816062e-01 -3.59321296e-01 8.86483252e-01 4.56032842e-01
2.92385191e-01 -9.59307551e-01 -8.70469868e-01 4.04641032e-02
-8.35712969e-01 -1.11635160e+00 -1.20786659e-01 -2.58369774e-01
-9.71434534e-01 -1.48510194e+00 -1.64991885e-01 -4.95693654e-01
1.81240603e-01 5.62312603e-01 1.59573781e+00 1.14984997e-01
3.82714063e-01 8.87223959e-01 -1.59464538e-01 -6.29415214e-01
-4.77777094e-01 1.13518350e-01 -4.93579656e-02 -1.03326190e+00
9.89365876e-01 -2.38871560e-01 -4.78526384e-01 3.51648033e-01
-6.17201626e-01 -6.64818957e-02 8.38660300e-01 5.61033189e-01
7.23209679e-02 -4.97919291e-01 7.92771935e-01 -1.08140171e+00
1.55428612e+00 -5.89182675e-01 -2.06226349e-01 6.07933819e-01
-1.24223065e+00 -3.14614326e-01 7.39057409e-03 -1.18206427e-01
-1.31551158e+00 -1.02771580e+00 3.21421206e-01 2.63979018e-01
-1.80073798e-01 8.22612286e-01 3.08713049e-01 -7.01614693e-02
1.01848412e+00 -2.88594723e-01 5.49318716e-02 -1.93836600e-01
2.68381506e-01 7.70578265e-01 3.27985644e-01 -1.15214908e+00
4.05206621e-01 1.84541464e-01 -6.52583957e-01 -6.62932158e-01
-1.15478230e+00 -5.71837723e-01 -1.95337966e-01 -1.87044159e-01
6.21882498e-01 -7.64240503e-01 -9.55564439e-01 -2.51008570e-01
-1.13116896e+00 -4.62210298e-01 -3.03299315e-02 9.84041452e-01
-6.52790725e-01 2.53323644e-01 -7.62041748e-01 -7.37626255e-01
-3.29519272e-01 -1.31338155e+00 5.44668317e-01 2.97380120e-01
-1.12510967e+00 -1.03624809e+00 3.93770337e-01 1.01071584e+00
5.49382210e-01 -7.83277974e-02 1.31077266e+00 -8.33812833e-01
4.72399592e-02 3.60184133e-01 -4.84234452e-01 7.81230032e-02
8.79219696e-02 -1.88090429e-01 -6.73244476e-01 -1.01031132e-01
4.65233475e-01 -6.34222806e-01 1.41563132e-01 4.39395636e-01
9.82881308e-01 -3.75184953e-01 -4.45500195e-01 2.92270422e-01
1.17702758e+00 2.12307081e-01 3.15560222e-01 8.24283242e-01
1.01626784e-01 1.09634602e+00 8.71822119e-01 2.60128155e-02
6.83683872e-01 4.88843977e-01 -5.13498858e-02 4.90915626e-01
1.85495391e-01 -2.64071435e-01 1.01991542e-01 9.16333318e-01
-1.07751049e-01 -2.47639474e-02 -1.40266967e+00 4.77967262e-01
-1.85218954e+00 -7.57035732e-01 -1.45016417e-01 2.23287082e+00
8.59629452e-01 2.99683452e-01 6.39485717e-02 -3.51494133e-01
6.43635333e-01 2.33672142e-01 -4.85777408e-01 -4.77299541e-01
1.20720975e-01 -1.85942814e-01 1.70427904e-01 5.26086569e-01
-3.73214751e-01 9.02242362e-01 8.30407429e+00 1.51024476e-01
-8.17568898e-01 2.43845925e-01 2.07744688e-01 -2.65773870e-02
-7.94499218e-01 3.39056283e-01 -5.94775915e-01 3.87889713e-01
1.14428163e+00 -7.30969071e-01 1.87103897e-01 4.63899225e-01
2.48146877e-01 2.62101173e-01 -1.37531936e+00 6.84745669e-01
-1.31595075e-01 -1.30614340e+00 -1.06908306e-01 -3.25541198e-02
5.11831105e-01 4.42539230e-02 4.32929024e-02 5.81355631e-01
6.57112837e-01 -1.18229663e+00 7.60834634e-01 3.93844604e-01
4.26284283e-01 -1.16469957e-01 7.22372413e-01 4.06156510e-01
-6.50773525e-01 -2.39649504e-01 -5.27159393e-01 -5.93801618e-01
1.16411336e-01 1.10811725e-01 -9.29724634e-01 2.88924485e-01
7.11278141e-01 2.13069201e-01 -1.81372046e-01 7.76489437e-01
-2.84007341e-01 9.53968942e-01 1.56899810e-01 2.25543708e-01
2.24504292e-01 -2.96294898e-01 4.10838455e-01 1.37862885e+00
8.95917043e-02 1.35283485e-01 -1.32804558e-01 1.12253547e+00
-5.00488579e-02 3.81987030e-03 -5.05687833e-01 -5.80709994e-01
1.43552542e+00 1.07603037e+00 -4.24564332e-01 -3.34652930e-01
-8.34196508e-01 2.70281553e-01 6.39278591e-01 5.80202699e-01
-6.31016970e-01 -3.16491090e-02 9.29094732e-01 -4.01641652e-02
-5.56643367e-01 -3.23709965e-01 -7.31266022e-01 -9.56359267e-01
-3.02297901e-02 -1.23880029e+00 5.13992965e-01 -8.65157068e-01
-1.05067277e+00 5.16055450e-02 1.50471419e-01 -5.98999560e-01
1.83092449e-02 -2.82811850e-01 -3.37120891e-01 1.04599714e+00
-1.28986764e+00 -6.44682825e-01 -6.01063669e-01 6.83556721e-02
2.99925059e-01 2.00620189e-01 9.08792555e-01 -1.33284345e-01
-4.35205519e-01 4.66455728e-01 -1.83445409e-01 -1.43283352e-01
1.20638537e+00 -1.17568469e+00 3.29084277e-01 3.66630584e-01
-2.20926907e-02 1.17144299e+00 7.64883161e-01 -8.40205073e-01
-1.18312311e+00 -3.39951128e-01 1.25469148e+00 -8.01115155e-01
7.44228721e-01 1.17287703e-01 -8.63353908e-01 1.06458545e+00
2.97519863e-01 -8.17192137e-01 1.24969184e+00 7.34058917e-01
-4.86266106e-01 1.77828088e-01 -1.28829730e+00 5.35189092e-01
1.37778878e+00 -8.28350306e-01 -1.39213896e+00 3.18641752e-01
6.85214996e-01 -3.18439871e-01 -1.25015140e+00 4.51682478e-01
8.06762218e-01 -1.20201850e+00 7.84942627e-01 -7.89050937e-01
2.33452380e-01 2.49050573e-01 -1.73234761e-01 -1.06745017e+00
-6.52547956e-01 -4.02191818e-01 1.88467383e-01 1.13183486e+00
5.49714684e-01 -1.03200507e+00 6.84443355e-01 1.16451049e+00
-4.10077602e-01 -4.82350528e-01 -7.11250901e-01 -4.80644584e-01
5.49020469e-01 -4.73984003e-01 5.54597080e-01 1.49673545e+00
3.46675307e-01 5.79961598e-01 3.62929553e-01 8.24600384e-02
3.14236611e-01 2.02419702e-03 9.02088404e-01 -1.66954505e+00
5.25067048e-03 -6.59950733e-01 1.31346956e-01 -9.75341976e-01
2.54711628e-01 -8.24282110e-01 -3.49575937e-01 -1.94691610e+00
5.64196110e-01 -5.91708541e-01 -2.11462736e-01 2.68812865e-01
-1.93278734e-02 -4.79381084e-01 4.18303251e-01 8.61350238e-01
-4.71498370e-01 -1.19088404e-02 9.16414499e-01 2.67782092e-01
-1.99770957e-01 -5.11477590e-01 -1.65057170e+00 9.69597399e-01
7.39474237e-01 -5.10755062e-01 -6.15857542e-01 -8.72999310e-01
6.23671591e-01 2.03860372e-01 -8.50918703e-03 -6.83539629e-01
4.61059928e-01 -6.82541490e-01 -6.64881477e-03 -3.38716447e-01
1.96944326e-01 -5.95791399e-01 2.79412895e-01 5.26703656e-01
-7.51087606e-01 3.75490516e-01 2.02298731e-01 1.56367779e-01
-2.69115478e-01 -5.36986887e-01 1.08517155e-01 -1.21755034e-01
-4.18991446e-01 -2.70575106e-01 -1.17687948e-01 6.20742619e-01
6.56551301e-01 -2.84188032e-01 -9.85024393e-01 -4.55741525e-01
-4.81750906e-01 4.03606594e-01 6.95016742e-01 3.82847667e-01
1.08799726e-01 -1.11653435e+00 -7.06688464e-01 -5.40252686e-01
2.36638069e-01 -2.07648367e-01 1.16921566e-01 1.16641378e+00
-3.31767887e-01 1.29604721e+00 -1.32328118e-04 -2.29285672e-01
-9.60204899e-01 1.37845546e-01 1.78584084e-01 -3.28690857e-01
-3.85227770e-01 5.20419717e-01 4.68048662e-01 -4.10659552e-01
2.41796374e-01 -6.59785569e-01 -3.13154310e-01 2.01742560e-01
3.65580469e-01 7.21560359e-01 -2.25515500e-01 -2.93905139e-01
-3.04433793e-01 3.91485333e-01 -1.19192302e-01 -3.92157316e-01
8.38461936e-01 -2.06705138e-01 -2.83658326e-01 7.37684727e-01
6.12699926e-01 1.07650325e-01 -7.66724765e-01 -1.31885856e-01
6.38675451e-01 -5.10067344e-01 -1.82674631e-01 -1.15096271e+00
-4.44108173e-02 4.60501522e-01 5.57126217e-02 4.04860705e-01
5.55314124e-01 -2.09147800e-02 2.40389973e-01 5.18511713e-01
3.98571372e-01 -1.32346964e+00 -8.09831694e-02 5.00569284e-01
1.17598701e+00 -9.24153686e-01 -8.71416032e-02 -4.21708673e-01
-7.67972350e-01 9.31700945e-01 7.88390219e-01 3.58774453e-01
5.13523638e-01 1.12110533e-01 2.14895174e-01 -4.50548500e-01
-1.00776041e+00 1.82998598e-01 5.88247515e-02 6.08824849e-01
1.13516188e+00 1.59245864e-01 -1.32047832e+00 9.78699744e-01
-6.12580121e-01 3.61447155e-01 4.17754054e-01 8.01536679e-01
-4.82076406e-01 -1.00771487e+00 -5.22649705e-01 6.94437563e-01
-7.27065623e-01 -1.16521167e-02 -8.61144781e-01 1.04743671e+00
-1.86917916e-01 1.29957044e+00 2.03829229e-01 -3.47736001e-01
3.71104687e-01 4.66906369e-01 3.78015965e-01 -7.81242549e-01
-8.25848579e-01 -4.50931609e-01 6.74606502e-01 -6.27144635e-01
-7.02242792e-01 -6.41185880e-01 -7.76268899e-01 -4.73779410e-01
-4.02559847e-01 7.90772974e-01 5.28128028e-01 1.07451487e+00
6.23150229e-01 1.52257577e-01 -1.75053358e-01 2.22584624e-02
-8.98081005e-01 -1.46490777e+00 -9.10202041e-02 4.62889373e-01
-1.13337852e-01 -8.83847117e-01 -5.26863575e-01 -5.06707251e-01] | [9.921162605285645, 8.379427909851074] |
7c2ff1c9-7d60-48ba-adc9-e3ac1f06675d | relative-facial-action-unit-detection | 1405.0085 | null | http://arxiv.org/abs/1405.0085v1 | http://arxiv.org/pdf/1405.0085v1.pdf | Relative Facial Action Unit Detection | This paper presents a subject-independent facial action unit (AU) detection
method by introducing the concept of relative AU detection, for scenarios where
the neutral face is not provided. We propose a new classification objective
function which analyzes the temporal neighborhood of the current frame to
decide if the expression recently increased, decreased or showed no change.
This approach is a significant change from the conventional absolute method
which decides about AU classification using the current frame, without an
explicit comparison with its neighboring frames. Our proposed method improves
robustness to individual differences such as face scale and shape, age-related
wrinkles, and transitions among expressions (e.g., lower intensity of
expressions). Our experiments on three publicly available datasets (Extended
Cohn-Kanade (CK+), Bosphorus, and DISFA databases) show significant improvement
of our approach over conventional absolute techniques. Keywords: facial action
coding system (FACS); relative facial action unit detection; temporal
information; | ['Louis-Philippe Morency', 'Mahmoud Khademi'] | 2014-05-01 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 4.75764543e-01 -1.86333299e-01 -2.06767157e-01 -5.85342884e-01
-3.97706896e-01 -3.98700684e-01 6.54655695e-01 -1.84441328e-01
-4.24023479e-01 7.61443198e-01 2.52672255e-01 3.85012597e-01
3.52377325e-01 -3.73316288e-01 -3.50570112e-01 -1.22232771e+00
-1.18158497e-01 -3.20451230e-01 2.47842610e-01 -2.68256098e-01
2.29949787e-01 7.88851738e-01 -1.92634642e+00 3.10228974e-01
1.87026083e-01 1.39080465e+00 -5.12046456e-01 5.63180983e-01
3.01194400e-01 9.19778764e-01 -5.75384319e-01 -1.19715489e-01
2.92132527e-01 -7.17659533e-01 -5.38338006e-01 4.18144435e-01
6.68701828e-01 -4.62292403e-01 1.72258958e-01 1.14724839e+00
6.00608885e-01 1.92964628e-01 6.13078654e-01 -1.52088547e+00
-3.01379621e-01 3.30936350e-02 -9.49650764e-01 4.67513531e-01
6.94117010e-01 -8.81862715e-02 4.69457537e-01 -9.03012276e-01
8.06261122e-01 1.48752248e+00 5.78587830e-01 8.67910445e-01
-1.03939474e+00 -1.16542363e+00 -1.35441003e-02 3.74215931e-01
-1.67493880e+00 -8.88466179e-01 7.76048899e-01 -4.07929987e-01
6.80087745e-01 2.14855030e-01 7.36137390e-01 8.92409503e-01
8.30416307e-02 5.10902584e-01 1.18728149e+00 -6.02973998e-01
2.46666089e-01 -2.63910651e-01 -9.48802829e-02 8.11846018e-01
-2.75964975e-01 -3.11823189e-02 -5.87933660e-01 -2.41584256e-01
5.80233753e-01 -2.78163522e-01 -3.33561331e-01 -6.82431757e-02
-8.91757905e-01 5.05484641e-01 -9.30735469e-03 4.82604116e-01
-3.75219345e-01 6.34082481e-02 4.69949812e-01 3.88267666e-01
6.91926241e-01 -3.42684895e-01 -3.28944117e-01 -3.36014241e-01
-1.03773272e+00 1.28883608e-02 4.07525450e-01 7.54298210e-01
9.20394897e-01 2.08109662e-01 -3.62227894e-02 7.26776063e-01
2.52534121e-01 4.53578800e-01 7.05008507e-01 -1.26212609e+00
-1.88098311e-01 5.62643528e-01 5.16946129e-05 -1.31721079e+00
-2.67403960e-01 4.84311074e-01 -4.72151846e-01 5.65906823e-01
3.53281826e-01 -3.56579781e-01 -7.79084027e-01 1.98417294e+00
6.34127438e-01 3.45672518e-01 -9.10349265e-02 7.43844509e-01
8.61553371e-01 4.65664536e-01 1.08623914e-01 -1.00419569e+00
1.17717326e+00 -5.06801188e-01 -1.18271887e+00 3.00843269e-01
6.20132923e-01 -8.50753844e-01 3.22517633e-01 3.93080235e-01
-8.57697785e-01 -6.51791155e-01 -1.07221770e+00 1.72249511e-01
-4.22997475e-01 2.45727018e-01 4.18480873e-01 9.41617012e-01
-1.34568691e+00 4.92575318e-01 -6.70092642e-01 -7.84926236e-01
1.61042020e-01 5.87193549e-01 -9.17254150e-01 3.31974149e-01
-1.07924378e+00 5.80494702e-01 1.11501664e-01 2.19987258e-01
-3.96055549e-01 -2.03960359e-01 -1.06283700e+00 -4.90963459e-01
2.05514625e-01 1.95860028e-01 1.14649153e+00 -2.05139661e+00
-1.81077957e+00 1.23363698e+00 -5.27076960e-01 -9.40770879e-02
3.64318371e-01 3.51382606e-02 -7.04798877e-01 3.41937631e-01
-2.64138788e-01 9.11339700e-01 1.22061598e+00 -7.97166049e-01
-4.70868498e-01 -6.80720627e-01 -1.37247816e-01 2.43384376e-01
-2.05013826e-01 6.86203480e-01 -3.46863508e-01 -5.55616617e-01
1.47279054e-01 -1.05792010e+00 3.90981585e-01 6.94967270e-01
1.37556568e-01 -4.76869017e-01 1.12529838e+00 -6.23361886e-01
1.28368819e+00 -2.38747478e+00 -1.61248505e-01 1.18147165e-01
-1.17550105e-01 2.85505354e-01 -4.64879014e-02 -6.04360811e-02
-5.21588027e-01 -1.05272420e-01 -6.47487864e-02 -1.20216109e-01
-5.23693621e-01 3.52552921e-01 3.23782325e-01 1.02558506e+00
2.50650346e-02 2.30666816e-01 -8.02550316e-01 -1.02084219e+00
1.20792747e-01 5.58549821e-01 -3.86806279e-01 -7.15984777e-02
4.39185619e-01 3.41700047e-01 -9.35385227e-02 1.16321003e+00
7.84554183e-01 4.41693157e-01 8.57990384e-02 -4.43385214e-01
-3.15352648e-01 -4.99369472e-01 -1.31204975e+00 1.20589483e+00
-2.49425247e-02 9.66849506e-01 3.91891569e-01 -8.01984131e-01
9.73184168e-01 7.83475816e-01 7.00274646e-01 -5.67194462e-01
5.57303786e-01 7.83051699e-02 -2.30657309e-02 -4.47456986e-01
2.65552700e-01 -1.42513648e-01 3.64285618e-01 1.52335733e-01
1.70870945e-01 4.04281527e-01 2.66909629e-01 -1.32328361e-01
7.48830795e-01 2.17955098e-01 7.86921978e-01 -3.89077991e-01
9.52144921e-01 -6.82913363e-01 9.19539452e-01 5.13275750e-02
-1.14481199e+00 6.70626938e-01 6.19089782e-01 -2.98905998e-01
-4.47758466e-01 -7.81160057e-01 -2.21676841e-01 1.15755403e+00
5.64125143e-02 -4.23506409e-01 -9.88761127e-01 -6.81950867e-01
-1.50820628e-01 5.04386239e-02 -1.23207510e+00 -1.06039427e-01
-5.01121223e-01 -6.69084728e-01 4.74908441e-01 4.07082707e-01
7.72137344e-01 -9.85994995e-01 -5.71782410e-01 -1.76899508e-01
-1.64943814e-01 -1.03489316e+00 -6.48520708e-01 -4.27504897e-01
-7.13503420e-01 -1.16814601e+00 -7.95312345e-01 -6.65951967e-01
8.17303836e-01 1.27714053e-01 5.73661983e-01 2.48332378e-02
-2.72962570e-01 5.02147615e-01 -3.88924748e-01 -2.38101169e-01
-2.58887380e-01 -8.98658514e-01 3.78664821e-01 8.57449472e-01
5.33375978e-01 -3.51123482e-01 -7.23173141e-01 5.21813869e-01
-6.46295130e-01 -4.46746975e-01 5.79645932e-02 4.81921345e-01
4.44431186e-01 -1.63900718e-01 2.22952440e-01 -5.94610453e-01
3.25452060e-01 -2.18243733e-01 -2.64685363e-01 2.29672834e-01
-4.50895071e-01 -3.69152218e-01 -8.70322660e-02 -6.25604331e-01
-1.25637519e+00 3.77471417e-01 9.14633498e-02 -4.50048447e-01
-2.51859277e-01 -1.74119785e-01 -1.38305441e-01 -3.65025610e-01
6.39115632e-01 1.04364641e-01 3.63590926e-01 -1.53877780e-01
2.25903783e-02 9.10226643e-01 6.73560679e-01 -1.19679213e-01
3.18759203e-01 8.85422647e-01 -8.68262444e-03 -1.09962451e+00
-3.00898820e-01 -6.19427741e-01 -1.19086099e+00 -7.88537741e-01
1.02430022e+00 -1.02256525e+00 -7.43209660e-01 8.42851937e-01
-1.05046570e+00 1.99233983e-02 2.18761384e-01 4.07179147e-01
-5.11851370e-01 6.38575375e-01 -3.78702521e-01 -1.03966570e+00
-3.27667952e-01 -9.54482019e-01 1.23644197e+00 3.07412297e-01
-5.33008277e-01 -7.53426731e-01 1.66116565e-01 1.56215310e-01
1.56705290e-01 7.88953066e-01 3.65402132e-01 -1.49290696e-01
2.49194726e-01 -1.27745852e-01 5.78813255e-02 5.57552874e-01
7.20499575e-01 9.44676638e-01 -1.00073874e+00 -2.07290843e-01
-5.91967888e-02 -2.35981390e-01 4.49258000e-01 3.56496662e-01
7.83771753e-01 -3.39777112e-01 -1.24313079e-01 4.88402843e-01
1.08985770e+00 6.70559645e-01 7.86129236e-01 -9.80802402e-02
2.06915125e-01 6.46721840e-01 9.28316772e-01 6.82613373e-01
-2.00886011e-01 8.92817795e-01 2.38025740e-01 -2.01167718e-01
-1.41153142e-01 2.33447492e-01 9.00835335e-01 3.43261868e-01
-6.37779236e-01 2.07957625e-02 -3.97853285e-01 3.13833147e-01
-1.55313921e+00 -1.24376822e+00 4.71229739e-02 2.05248976e+00
7.38733232e-01 -3.25259566e-01 3.22940320e-01 4.80040789e-01
1.01005852e+00 2.27918804e-01 -3.36554646e-01 -6.96617126e-01
-2.06241205e-01 2.04905197e-01 2.43431672e-01 4.37173754e-01
-1.25500703e+00 9.14238214e-01 6.65619421e+00 7.62552440e-01
-1.46250117e+00 1.99903429e-01 7.38691628e-01 -9.80967581e-02
5.73883832e-01 -3.91489238e-01 -6.22317791e-01 4.65466738e-01
7.82789171e-01 -1.73100889e-01 3.66381444e-02 6.69530869e-01
4.23102170e-01 -6.13078654e-01 -8.77227724e-01 1.42809057e+00
6.85253143e-01 -6.88456237e-01 -4.05227654e-02 -1.70515016e-01
5.31498849e-01 -4.95219350e-01 -3.61905135e-02 2.26179715e-02
-1.70481563e-01 -6.46844864e-01 6.54827654e-01 4.33895200e-01
9.39545929e-01 -7.74768054e-01 6.40264928e-01 -2.76645094e-01
-1.41686761e+00 8.80256072e-02 -4.43500727e-02 -1.81352541e-01
-2.00474337e-01 3.61956991e-02 -5.92371702e-01 5.95916919e-02
8.12494695e-01 8.83135736e-01 -6.12403095e-01 4.21052366e-01
-1.19608521e-01 4.61894035e-01 -2.33523324e-01 1.18192233e-01
-9.25042704e-02 -1.98959425e-01 4.67750221e-01 1.19566894e+00
3.93543392e-01 4.93381470e-01 -2.19596624e-01 1.71079382e-01
2.49721125e-01 5.31130254e-01 -5.71531594e-01 1.49890170e-01
1.35089412e-01 1.24647033e+00 -8.46200287e-01 -3.58684868e-01
-6.54460728e-01 1.31056547e+00 -2.04174757e-01 2.32226118e-01
-7.04779327e-01 -1.21157080e-01 9.30370033e-01 3.65599357e-02
5.83629161e-02 1.04031235e-01 3.19239169e-01 -9.01599407e-01
-8.87546167e-02 -7.52387464e-01 7.42006361e-01 -9.56251621e-01
-6.11864150e-01 4.89296287e-01 8.34731758e-03 -1.50734651e+00
-4.29429650e-01 -5.32621920e-01 -4.52965856e-01 3.58773500e-01
-1.06889820e+00 -1.04788840e+00 -3.78493577e-01 1.04647982e+00
4.69204098e-01 -2.08689515e-02 5.89319706e-01 4.27481383e-01
-5.58540940e-01 8.99818838e-01 -2.20337451e-01 4.31790233e-01
1.03419888e+00 -8.15201521e-01 -5.26534259e-01 8.66333008e-01
-1.31404161e-01 3.95439357e-01 7.93363094e-01 -4.74424839e-01
-9.35302556e-01 -7.68673360e-01 9.74449933e-01 -2.54586607e-01
4.20104593e-01 -2.47606158e-01 -4.77009416e-01 6.12220824e-01
3.09175793e-02 4.11067724e-01 8.46526921e-01 -3.06040764e-01
-3.02705854e-01 -5.14695108e-01 -1.52729952e+00 4.93408531e-01
9.12494838e-01 -3.63577247e-01 -2.57047325e-01 1.42860636e-01
1.12199113e-02 -1.12245359e-01 -8.09670269e-01 6.32604897e-01
1.11513758e+00 -1.30067015e+00 5.46790600e-01 -2.52736837e-01
-1.16835989e-01 -3.91983181e-01 -2.99162835e-01 -8.37329388e-01
-1.33582711e-01 -8.29773724e-01 1.39149189e-01 1.36796951e+00
-2.86859944e-02 -4.28455561e-01 6.19054854e-01 5.73067546e-01
3.05830508e-01 -3.49242657e-01 -1.32483006e+00 -6.21678770e-01
-4.85136360e-01 -1.49201870e-01 2.26928592e-01 1.03729713e+00
2.44799256e-01 -7.72268325e-02 -5.14182448e-01 -1.06213234e-01
2.38655940e-01 -1.96272492e-01 6.26095116e-01 -1.07964849e+00
2.44903684e-01 -3.15469593e-01 -1.32217109e+00 -3.39659452e-01
2.03374580e-01 -3.29515070e-01 -7.66118392e-02 -6.10799313e-01
1.87278628e-01 3.80800158e-01 -3.30218703e-01 7.55534172e-01
5.70101291e-03 8.66180480e-01 5.72602339e-02 4.67665493e-02
-4.69791889e-01 4.81491327e-01 9.77230489e-01 1.33707181e-01
-1.04289711e-01 -2.90624723e-02 -1.58926502e-01 9.45565522e-01
5.30729473e-01 -2.28889301e-01 -3.04482907e-01 1.93978548e-01
-1.78788662e-01 7.28745461e-02 2.73311939e-02 -1.32885849e+00
8.71755481e-02 -1.64832339e-01 4.54388052e-01 -3.50640774e-01
4.39411163e-01 -8.11544240e-01 8.36911574e-02 5.77147722e-01
-1.20280690e-01 3.42865855e-01 1.65924147e-01 5.06335199e-01
-4.87909108e-01 1.18567280e-01 1.33047783e+00 6.02767915e-02
-1.07148099e+00 2.19412282e-01 -7.56050110e-01 -5.02488077e-01
1.40526104e+00 -6.93803072e-01 1.32606298e-01 -7.27116168e-01
-8.64270985e-01 -1.87491298e-01 4.61975813e-01 4.67103750e-01
4.90388960e-01 -1.48055208e+00 -6.40350521e-01 2.92559654e-01
1.95937753e-01 -8.71675432e-01 2.18428269e-01 1.13480318e+00
-4.85431194e-01 4.50794660e-02 -5.13834715e-01 -6.18520021e-01
-2.42776990e+00 3.34300071e-01 5.65869570e-01 4.12406534e-01
9.71014891e-03 9.24165726e-01 1.82339072e-01 4.57768627e-02
1.86779857e-01 -2.96015352e-01 -6.88108206e-01 6.82129562e-01
7.98736870e-01 5.17507613e-01 1.04818940e-01 -1.49255145e+00
-5.08737206e-01 8.29175532e-01 2.09097505e-01 -1.66779518e-01
7.53940761e-01 -2.80346155e-01 -3.81751448e-01 6.81440413e-01
1.49079514e+00 -6.54013455e-02 -1.19516432e+00 -1.20861232e-01
-3.00525695e-01 -5.77827394e-01 -3.13140154e-02 -3.72744322e-01
-1.33380115e+00 6.74382985e-01 1.35143268e+00 -2.71651357e-01
1.51847517e+00 -1.03833117e-01 3.31898242e-01 4.90941517e-02
2.31457651e-01 -1.39986670e+00 1.93597302e-02 1.87786385e-01
9.02807295e-01 -1.27316678e+00 1.54286236e-01 -4.74634171e-01
-3.78308207e-01 1.43912899e+00 7.04580784e-01 2.53885388e-01
9.94952500e-01 2.13379785e-01 3.21758777e-01 -7.27568120e-02
-6.83507442e-01 -3.59284371e-01 4.71331738e-02 5.38376570e-01
6.04604483e-01 -1.59278989e-01 -5.86381495e-01 -6.51989654e-02
1.42382560e-02 1.37229443e-01 4.07115072e-01 1.15064526e+00
-4.96800095e-01 -7.42577851e-01 -4.70834821e-01 4.81136926e-02
-6.15395606e-01 2.21577361e-01 -7.18844652e-01 9.51879323e-01
4.77776617e-01 9.31117117e-01 4.36842680e-01 -2.76884526e-01
7.80209228e-02 3.24562281e-01 6.70155346e-01 -2.72933781e-01
-3.30787778e-01 2.33365357e-01 -2.42095185e-03 -8.28875542e-01
-1.29602659e+00 -1.08203042e+00 -1.19575870e+00 -2.37312630e-01
-1.73132122e-01 -1.08271383e-01 3.60831052e-01 8.12674046e-01
1.02651983e-01 -1.76546633e-01 7.92688847e-01 -8.47132981e-01
1.77412167e-01 -1.05170155e+00 -7.91246712e-01 6.38880491e-01
6.16274178e-01 -9.58220899e-01 -7.22374856e-01 6.56164944e-01] | [13.667009353637695, 1.7776947021484375] |
e119f68d-5b5d-4daa-beac-f496d1209bbb | a-spectral-approach-to-unsupervised-object | 1907.02731 | null | https://arxiv.org/abs/1907.02731v5 | https://arxiv.org/pdf/1907.02731v5.pdf | A 3D Convolutional Approach to Spectral Object Segmentation in Space and Time | We formulate object segmentation in video as a graph partitioning problem in space and time, in which nodes are pixels and their relations form local neighborhoods. We claim that the strongest cluster in this pixel-level graph represents the salient object segmentation. We compute the main cluster using a novel and fast 3D filtering technique that finds the spectral clustering solution, namely the principal eigenvector of the graph's adjacency matrix, without building the matrix explicitly - which would be intractable. Our method is based on the power iteration for finding the principal eigenvector of a matrix, which we prove is equivalent to performing a specific set of 3D convolutions in the space-time feature volume. This allows us to avoid creating the matrix and have a fast parallel implementation on GPU. We show that our method is much faster than classical power iteration applied directly on the adjacency matrix. Different from other works, ours is dedicated to preserving object consistency in space and time at the level of pixels. For that, it requires powerful pixel-wise features at the frame level. This makes it perfectly suitable for incorporating the output of a backbone network or other methods and fast-improving over their solution without supervision. In experiments, we obtain consistent improvement, with the same set of hyper-parameters, over the top state of the art methods on DAVIS-2016 dataset, both in unsupervised and semi-supervised tasks. We also achieve top results on the well-known SegTrackv2 dataset. | ['Marius Leordeanu', 'Elena Burceanu'] | 2019-07-05 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 2.16098085e-01 8.64452347e-02 6.89698085e-02 5.02725877e-02
-3.47393334e-01 -6.82276189e-01 1.86268687e-01 2.17157751e-01
-6.15814447e-01 -2.95431688e-02 -2.81237096e-01 -3.00319016e-01
-1.42742276e-01 -8.26504588e-01 -8.91011059e-01 -8.44510138e-01
-2.25688711e-01 4.51406240e-01 8.33407342e-01 6.54411316e-02
1.96184084e-01 5.86681545e-01 -1.47941339e+00 4.55294028e-02
6.88996553e-01 1.00763297e+00 2.18127668e-01 6.71355665e-01
-4.05803286e-02 5.27080178e-01 -7.23776445e-02 -2.28864297e-01
5.93992412e-01 -4.23654646e-01 -9.76746440e-01 6.40113533e-01
7.63931692e-01 6.47188798e-02 -4.17564362e-01 1.33335257e+00
2.92537034e-01 2.66056210e-01 5.76485276e-01 -1.21057212e+00
-1.48469672e-01 4.82821792e-01 -9.34850395e-01 9.50971097e-02
6.85101748e-02 -1.19288191e-01 9.23638880e-01 -7.02863157e-01
9.42213118e-01 1.01420546e+00 8.52225780e-01 2.01646939e-01
-1.44138920e+00 -2.30564877e-01 7.47993886e-02 3.21607262e-01
-1.50354600e+00 1.15466207e-01 8.67864728e-01 -5.06545305e-01
8.34927738e-01 3.14634204e-01 9.43662405e-01 4.52808887e-01
-1.33220002e-01 7.99327374e-01 1.00748432e+00 -4.25291538e-01
3.11896652e-01 -2.60399103e-01 2.74794430e-01 1.05035257e+00
1.38688892e-01 -3.83947283e-01 -2.99349278e-01 3.65888290e-02
9.28359210e-01 1.78076122e-02 -3.85221541e-01 -8.91323447e-01
-1.37323141e+00 6.66301548e-01 5.16465545e-01 4.13397402e-01
-2.21203849e-01 2.51690537e-01 1.95825711e-01 1.16295315e-01
4.51184094e-01 -1.66112166e-02 -3.81192744e-01 1.96763590e-01
-1.20102501e+00 4.12543043e-02 9.78470087e-01 8.65950465e-01
9.95189071e-01 -4.58363086e-01 -5.97546883e-02 4.56780940e-01
2.35723197e-01 3.08798224e-01 1.52106047e-01 -1.28568327e+00
8.65779817e-02 6.71303689e-01 -2.68904150e-01 -1.28175533e+00
-6.46712899e-01 -4.53847200e-01 -9.13534045e-01 3.34641248e-01
8.57390940e-01 -6.10759072e-02 -9.72753227e-01 1.53365636e+00
7.04212248e-01 7.52168894e-01 -3.20114642e-01 1.08431232e+00
5.85798621e-01 5.08279741e-01 -3.28641534e-01 -3.86536449e-01
1.41536379e+00 -9.90942657e-01 -4.18813825e-01 9.41772759e-02
5.86240113e-01 -6.41225100e-01 9.32455957e-01 3.73145282e-01
-9.57454681e-01 -5.39595902e-01 -9.32798147e-01 -1.03355952e-01
-1.60775110e-01 -9.74276336e-05 6.31941020e-01 5.19451261e-01
-1.19809544e+00 9.15788054e-01 -9.87008214e-01 -4.94572878e-01
3.78896475e-01 2.54263043e-01 -3.93489987e-01 1.71506494e-01
-7.18569219e-01 3.51936758e-01 3.41431022e-01 7.74826109e-02
-4.95048493e-01 -6.81457043e-01 -7.18360186e-01 -6.44077733e-02
6.59459472e-01 -6.81236446e-01 6.28255367e-01 -9.70561683e-01
-1.23014367e+00 1.18685031e+00 -1.66723743e-01 -6.73859775e-01
7.78082669e-01 1.80946272e-02 1.67131856e-01 4.52977300e-01
1.32931024e-01 6.10183179e-01 1.00246847e+00 -1.06354487e+00
-6.53545082e-01 -5.32729387e-01 -5.38298227e-02 9.33959931e-02
-2.32225373e-01 -1.91693455e-01 -1.22062039e+00 -5.93256772e-01
7.53488183e-01 -1.20836282e+00 -5.32328725e-01 9.58830565e-02
-4.70372647e-01 -3.66193235e-01 9.01873887e-01 -5.59450030e-01
1.06864500e+00 -2.29014230e+00 3.83547693e-01 6.70640767e-01
5.50352693e-01 7.22020492e-02 1.04253374e-01 4.98417243e-02
-1.67479977e-01 9.75612830e-03 -5.47801852e-01 -2.64246345e-01
-1.74312741e-01 1.16671242e-01 6.95967376e-02 9.41756606e-01
-3.30375992e-02 9.34417427e-01 -7.74328172e-01 -6.97195888e-01
2.80267864e-01 4.11445260e-01 -6.67092860e-01 -1.70681253e-01
-1.67605489e-01 2.66632408e-01 -3.18168402e-01 1.67328417e-01
9.05493975e-01 -4.94850814e-01 2.02238739e-01 -4.96437937e-01
-1.57938153e-01 -8.68961215e-02 -1.72438955e+00 2.06617570e+00
8.32714513e-02 5.64589441e-01 1.87129527e-01 -1.33390117e+00
5.23241818e-01 -3.59896533e-02 9.29714918e-01 -3.25132728e-01
2.75896788e-01 7.22827464e-02 -1.24601334e-01 -3.44353497e-01
1.19061902e-01 3.09561759e-01 2.25406125e-01 4.17662770e-01
1.60201490e-01 -8.02644715e-02 6.08020127e-01 5.44942379e-01
1.19378150e+00 2.05822170e-01 -1.70134440e-01 -6.76147044e-01
5.12670934e-01 -1.65530619e-05 4.48785186e-01 7.25827336e-01
-1.64894789e-01 8.25175226e-01 7.39152133e-01 -2.53210515e-01
-7.62238503e-01 -1.02264297e+00 -1.85373276e-01 6.33610010e-01
5.24118125e-01 -6.16284251e-01 -1.28981638e+00 -8.71439338e-01
-6.34005144e-02 1.75651252e-01 -6.50584400e-01 1.60002038e-01
-5.56952298e-01 -6.95394456e-01 6.05097264e-02 2.79841751e-01
5.90194762e-01 -6.97506905e-01 -7.49248564e-01 1.25269353e-01
-3.05010170e-01 -1.43941259e+00 -6.32480800e-01 2.48855665e-01
-1.04723787e+00 -1.31430101e+00 -6.29432559e-01 -1.01617134e+00
9.07410920e-01 4.81823325e-01 9.29022491e-01 2.46146977e-01
-3.35849881e-01 4.17374045e-01 -1.84718236e-01 -5.10828756e-02
8.05903971e-02 3.61026973e-02 -1.91627681e-01 4.14057761e-01
1.28794596e-01 -7.95264661e-01 -5.65119326e-01 3.06243211e-01
-8.32427263e-01 1.70494005e-01 2.28281051e-01 4.02770877e-01
1.01627827e+00 2.98614621e-01 -1.44394040e-01 -9.48868692e-01
2.84284074e-02 -1.22193381e-01 -8.11504722e-01 2.10961804e-01
-3.08953732e-01 8.98997262e-02 2.65008003e-01 -3.26548994e-01
-4.77193594e-01 7.41757810e-01 -2.02716775e-02 -5.40035665e-01
-7.05233216e-02 1.04006425e-01 4.83919829e-02 -3.41275215e-01
5.08738339e-01 1.12572595e-01 2.31496282e-02 -4.54013795e-01
5.01189470e-01 8.01851302e-02 6.46522105e-01 -3.79117787e-01
1.06159174e+00 9.55126822e-01 4.02790129e-01 -8.67507696e-01
-8.92412484e-01 -8.99168670e-01 -1.00556302e+00 -4.17769730e-01
1.18773139e+00 -7.51891255e-01 -8.14728379e-01 3.32123905e-01
-9.86827552e-01 -4.01333898e-01 -4.98596251e-01 4.40835655e-01
-7.42719352e-01 7.50324130e-01 -6.75268710e-01 -4.47562814e-01
1.76349226e-02 -1.04346001e+00 9.37105477e-01 1.01129629e-01
-3.70604582e-02 -8.37502420e-01 -3.18155922e-02 2.30438292e-01
-8.18247572e-02 4.01884615e-01 6.20942473e-01 -2.05105007e-01
-7.38001347e-01 1.45092800e-01 -4.23135936e-01 2.22540557e-01
-2.42327064e-01 1.94059387e-01 -7.15427756e-01 -2.03572527e-01
2.74871558e-01 1.67445600e-01 1.21010411e+00 6.74744129e-01
1.29136086e+00 -5.90703003e-02 -3.34675252e-01 1.01040173e+00
1.52541506e+00 -2.61031419e-01 4.51059759e-01 1.16508774e-01
1.10275483e+00 6.95780575e-01 3.87740105e-01 1.90081716e-01
2.36359552e-01 6.85935974e-01 3.83799344e-01 -3.60363036e-01
-3.27455729e-01 -2.01464705e-02 2.43628442e-01 6.64322913e-01
-3.21663618e-01 1.36328802e-01 -8.03589046e-01 5.92489421e-01
-2.25790668e+00 -8.62891138e-01 -7.32818186e-01 2.12633753e+00
6.78562462e-01 1.33415133e-01 3.82223338e-01 2.63546973e-01
6.84004843e-01 7.69400271e-03 -4.02033597e-01 3.77641656e-02
-2.74890453e-01 3.51487607e-01 8.56472194e-01 5.83568215e-01
-1.44243848e+00 8.79994214e-01 5.79208326e+00 9.08863902e-01
-9.05357242e-01 3.00525367e-01 6.43082261e-01 -7.51306191e-02
-2.29116920e-02 4.99212332e-02 -4.72353518e-01 3.09187979e-01
4.46305603e-01 1.02589853e-01 4.19014782e-01 7.63004899e-01
5.41236326e-02 -2.85232574e-01 -1.02978587e+00 1.12087250e+00
4.59287092e-02 -1.44132304e+00 -2.43697852e-01 6.81696385e-02
8.52428913e-01 1.91101685e-01 -2.30247945e-01 -3.76342773e-01
3.32853608e-02 -6.44991815e-01 8.92014802e-01 1.45524725e-01
3.75991076e-01 -6.86832488e-01 3.95353138e-01 3.31523567e-01
-1.36287653e+00 2.12514296e-01 -2.91236371e-01 7.51158446e-02
1.55176163e-01 9.18910563e-01 -3.30740601e-01 4.78899181e-01
9.61193204e-01 9.18067217e-01 -6.49444878e-01 1.06984830e+00
-3.12416498e-02 6.03294134e-01 -7.75766432e-01 1.92013383e-01
2.80795813e-01 -7.06474662e-01 5.84191442e-01 1.35800266e+00
2.28841245e-01 1.56644970e-01 3.38793993e-01 7.82959759e-01
-6.34313077e-02 2.07257494e-01 -3.55030894e-01 3.59413803e-01
-1.19422622e-01 1.47170472e+00 -1.53983974e+00 -4.10760224e-01
-3.51239860e-01 1.25729454e+00 3.26011300e-01 3.40543211e-01
-8.18785191e-01 -1.95488080e-01 4.06882048e-01 2.51013607e-01
6.90715015e-01 -5.16009510e-01 -3.63032520e-01 -1.14543700e+00
1.90678239e-01 -5.82911789e-01 4.05169964e-01 -3.96414071e-01
-1.06143177e+00 2.98948020e-01 -1.93312928e-01 -1.04570484e+00
1.49690554e-01 -5.89918435e-01 -4.15031135e-01 4.47390020e-01
-1.31555223e+00 -8.49922001e-01 -2.83307344e-01 1.02551901e+00
2.87575215e-01 4.19536233e-01 3.11321616e-01 3.65730077e-01
-5.23059189e-01 2.84040242e-01 -2.34352350e-02 3.58459592e-01
3.09383392e-01 -1.25341797e+00 3.27059954e-01 1.19292772e+00
6.29560173e-01 3.23108435e-01 5.32531023e-01 -5.62318802e-01
-1.52718663e+00 -1.05948865e+00 7.72614241e-01 -3.95541906e-01
7.76845217e-01 -5.61922848e-01 -7.95496106e-01 6.44185364e-01
1.22298367e-01 3.43208402e-01 2.15320155e-01 -1.62063897e-01
-1.16254374e-01 1.03770029e-02 -9.39404011e-01 4.69358385e-01
1.51723790e+00 -2.73252249e-01 -2.71951169e-01 6.39449596e-01
7.90785193e-01 -5.34980476e-01 -7.05808818e-01 3.59198421e-01
2.75973499e-01 -1.07975650e+00 1.06172788e+00 -3.88179809e-01
1.68251559e-01 -8.13756406e-01 7.84651563e-02 -8.04670274e-01
-4.65097666e-01 -8.55273783e-01 -6.05004020e-02 9.89267170e-01
3.10231745e-01 -3.48713964e-01 9.91443157e-01 3.44545692e-01
-2.67753359e-02 -7.11324155e-01 -9.39982891e-01 -7.40844727e-01
-3.51282477e-01 -7.32485414e-01 9.65912119e-02 1.01189780e+00
-3.05092722e-01 2.02213645e-01 5.05074626e-03 3.37115258e-01
1.04377222e+00 3.41782868e-01 7.04726934e-01 -1.26538539e+00
-3.80970180e-01 -5.43708920e-01 -6.65844023e-01 -1.18460834e+00
1.19929135e-01 -1.12069011e+00 1.43526541e-02 -1.50443864e+00
1.89396635e-01 -4.32668447e-01 -1.67882770e-01 4.09824938e-01
-5.39126955e-02 6.52934194e-01 3.77539992e-01 1.35209590e-01
-8.36119354e-01 8.09675455e-02 1.27285910e+00 -1.63164809e-01
-3.40793133e-01 -1.64920047e-01 -3.25431019e-01 1.01258075e+00
5.22527635e-01 -5.40769279e-01 -2.65117705e-01 -3.04492980e-01
1.26227006e-01 -3.00381690e-01 5.37003398e-01 -1.12480772e+00
4.01252359e-01 6.79386929e-02 1.61142603e-01 -6.66156292e-01
3.12935375e-02 -9.68675971e-01 2.00351223e-01 5.38117230e-01
-2.46878807e-02 -1.23858199e-01 3.90107483e-02 5.73016107e-01
-1.12440929e-01 -2.51831919e-01 7.80957401e-01 -3.30490321e-02
-9.07583952e-01 5.04159987e-01 -1.32929415e-01 9.41587314e-02
1.29285622e+00 -4.05153275e-01 1.53732300e-01 -3.57122570e-02
-1.05363059e+00 2.15556920e-01 7.17809916e-01 1.03913300e-01
4.47626293e-01 -1.25103891e+00 -6.01061344e-01 1.46599561e-01
-2.47224972e-01 2.36593738e-01 3.84240389e-01 1.31509423e+00
-8.05303574e-01 -7.82494247e-02 1.14071434e-02 -1.16542172e+00
-1.34196377e+00 5.90705872e-01 3.37788820e-01 -1.01094224e-01
-9.72236753e-01 9.43240583e-01 3.52944493e-01 -1.04096957e-01
2.84878254e-01 -4.07625258e-01 -6.54265508e-02 2.20213607e-01
2.88521171e-01 3.08279693e-01 1.31146312e-01 -7.39088535e-01
-4.76448625e-01 1.09776103e+00 3.83961916e-01 -2.05373056e-02
1.34551573e+00 -1.55644953e-01 -5.67124665e-01 2.86503166e-01
1.46804452e+00 5.69436215e-02 -1.36768115e+00 -2.93600798e-01
-6.67411759e-02 -4.49120909e-01 2.56522268e-01 -1.01411574e-01
-1.61597490e+00 7.44938672e-01 6.96768582e-01 3.84287506e-01
1.21274710e+00 2.91593611e-01 7.04171538e-01 2.94968635e-01
2.36485660e-01 -1.19247198e+00 -1.19838409e-01 4.37769234e-01
5.09250522e-01 -1.01856685e+00 8.46823156e-02 -9.36653376e-01
-2.42335066e-01 1.09637105e+00 3.35255176e-01 -5.75109601e-01
9.81795788e-01 3.35676461e-01 -1.49504349e-01 -2.70831048e-01
-2.24139199e-01 -5.38055182e-01 5.05476475e-01 4.84180152e-01
1.76959231e-01 6.39283378e-03 -4.46100116e-01 1.33087337e-01
-1.50860429e-01 -2.54421741e-01 3.74430150e-01 6.48095310e-01
-3.81397098e-01 -9.13916588e-01 -2.04151571e-01 3.55396450e-01
-3.12702209e-01 6.62658364e-02 -2.75048375e-01 6.90681994e-01
3.31385940e-01 8.21385384e-01 3.22181076e-01 -2.41134748e-01
3.46250534e-01 -1.14903122e-01 7.02472448e-01 -4.94880557e-01
-3.81980807e-01 3.65131706e-01 -2.40779921e-01 -1.07111776e+00
-8.00769210e-01 -9.17609155e-01 -1.65812778e+00 -1.50713414e-01
-2.88073450e-01 4.01167981e-02 5.35169661e-01 8.78840208e-01
2.87823558e-01 3.71336401e-01 5.36599874e-01 -1.02740216e+00
-2.89530940e-02 -4.07813162e-01 -7.35115707e-01 7.50245988e-01
4.95786183e-02 -4.10801917e-01 -3.19807857e-01 2.79954076e-01] | [9.148632049560547, -0.20750819146633148] |
ef069eac-b44f-4712-a430-41a154b754a0 | small-signal-stability-techniques-for-power | 2111.01694 | null | https://arxiv.org/abs/2111.01694v1 | https://arxiv.org/pdf/2111.01694v1.pdf | Small-Signal Stability Techniques for Power System Modal Analysis, Control, and Numerical Integration | This thesis proposes novel Small-Signal Stability Analysis (SSSA)-based techniques that contribute to electric power system modal analysis, automatic control, and numerical integration. Modal analysis is a fundamental tool for power system stability analysis and control. The thesis proposes a SSSA approach to determine the Participation Factors (PFs) of algebraic variables in power system dynamic modes. The thesis also explores SSSA techniques for the design of power system controllers. The contributions on this topic are twofold: i) Investigate a promising control approach, that is to synthesize automatic regulators for power systems based on the theory of fractional calculus. ii) Propose a novel perspective on the potential impact of time delays on power system stability. Through SSSA, the thesis systematically identifies the control parameter settings for which delays in PSSs improve the damping of a power system. Both analytical and simulation-based results are presented. Finally, SSSA is utilized in the thesis to systematically propose a delay-based method to reduce the coupling of the equations of power system models for transient stability analysis. The method consists in identifying the variables that, when subjected to a delay equal to the time step of the numerical integration, leave practically unchanged the system trajectories. Such an one-step-delay approximation increases the sparsity of the system Jacobian matrices and can be used in conjunction with state-of-the-art techniques for the integration of DAEs. The proposed approach is evaluated in terms of accuracy, convergence and computational burden. Throughout the thesis, the proposed techniques are duly validated through numerical tests based on real-world network models. | ['Georgios Tzounas'] | 2021-11-02 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [ 5.40495664e-02 4.06602696e-02 2.98032854e-02 6.68524981e-01
-1.10514656e-01 -1.06236041e+00 1.70961872e-01 1.94120795e-01
3.90161484e-01 9.80057955e-01 -5.58520138e-01 -5.18253207e-01
-8.84851336e-01 -5.34275711e-01 -5.80208004e-02 -1.10858059e+00
-7.63679966e-02 -1.59824956e-02 -1.26606971e-01 -6.88123882e-01
2.52581257e-02 9.89729702e-01 -1.21734214e+00 -4.48083699e-01
8.10599685e-01 6.37223959e-01 -4.78092849e-01 4.61665541e-01
5.04202485e-01 4.22267348e-01 -9.26318288e-01 3.94129843e-01
6.83717430e-02 -6.13264203e-01 -5.41139543e-01 1.50206953e-01
-6.82789445e-01 3.71155851e-02 -1.68215796e-01 1.12628651e+00
6.87783718e-01 6.44036591e-01 7.93918490e-01 -1.66013515e+00
3.01429689e-01 6.24433517e-01 -4.57179487e-01 3.59002560e-01
6.10596180e-01 -5.73373735e-02 4.86026496e-01 -5.27974963e-01
3.96899760e-01 1.14192057e+00 7.50038028e-01 6.19525127e-02
-1.52448988e+00 -4.54206824e-01 -2.25509405e-01 1.04549564e-01
-1.68283486e+00 -7.24226981e-03 1.16981149e+00 -3.71141225e-01
8.80059779e-01 7.35692799e-01 7.58472085e-01 -5.84510453e-02
2.81545877e-01 3.24297518e-01 8.38590086e-01 -6.85065985e-01
4.24387187e-01 1.89332262e-01 3.72802675e-01 2.95357317e-01
1.89170361e-01 -2.24612094e-02 3.36952209e-02 -5.39337695e-01
7.34879494e-01 -5.47888577e-01 -3.93581808e-01 -2.48538494e-01
-6.27240360e-01 9.82875347e-01 3.12579572e-02 8.22890997e-01
-4.04051781e-01 -1.64854318e-01 5.21813750e-01 4.44037676e-01
4.75875914e-01 5.14961600e-01 -7.93027505e-02 4.63457927e-02
-7.87024140e-01 2.81503707e-01 7.69672871e-01 5.62419713e-01
1.24338999e-01 9.61926520e-01 2.32368693e-01 5.02327204e-01
2.66526341e-01 7.47410119e-01 5.45792505e-02 -9.15238142e-01
-1.48048624e-01 6.01627588e-01 2.94299811e-01 -1.07257771e+00
-6.27032399e-01 -4.68813092e-01 -5.73951006e-01 4.90397692e-01
4.77830917e-01 -6.00507677e-01 -4.44025211e-02 1.44063818e+00
5.63058376e-01 -1.50780037e-01 1.37467474e-01 5.45650601e-01
3.44932601e-02 1.28750229e+00 -4.00590450e-01 -1.10725558e+00
1.14765298e+00 -2.38280818e-01 -1.13760459e+00 6.14009202e-01
5.53663135e-01 -8.92938495e-01 5.86533487e-01 4.33188289e-01
-1.24402368e+00 -1.02847286e-01 -1.20426011e+00 8.04011881e-01
-2.00368062e-01 4.59109068e-01 -6.35410324e-02 7.91179776e-01
-1.07330239e+00 6.59198761e-01 -9.49319363e-01 -2.41371036e-01
-4.39109147e-01 4.64280486e-01 1.95766121e-01 1.06914878e+00
-1.41302121e+00 1.23374569e+00 2.95479566e-01 4.88177657e-01
-5.85262738e-02 -1.16129112e+00 -5.31990767e-01 1.39124513e-01
3.35768729e-01 -3.56467932e-01 9.73076642e-01 -1.01002216e+00
-1.86055577e+00 1.63170472e-02 -2.92409539e-01 -2.98700809e-01
2.39965469e-01 4.35102373e-01 -6.07189834e-01 6.22372985e-01
-1.67668164e-01 -4.01576519e-01 7.98912168e-01 -1.00270033e+00
-4.10112768e-01 1.95345774e-01 -1.51560828e-01 -3.51085216e-02
-1.81684852e-01 5.71224727e-02 7.12754130e-01 -6.54404998e-01
8.48754868e-02 -8.96818578e-01 -3.45417619e-01 -3.37434888e-01
-2.94910043e-01 -2.79317677e-01 1.03111160e+00 -6.95513129e-01
1.55456007e+00 -2.02991414e+00 3.85878503e-01 1.07204318e+00
-3.38330120e-01 4.99672949e-01 4.72192973e-01 1.18353760e+00
-5.01645982e-01 -1.37597635e-01 -5.02217151e-02 5.91486156e-01
1.18705466e-01 -1.58778623e-01 -5.91055691e-01 7.75379896e-01
1.45499334e-01 2.69972235e-01 -5.56759953e-01 7.33436719e-02
7.36965537e-01 5.06936550e-01 -7.23519847e-02 -3.88629943e-01
2.41376519e-01 3.91501367e-01 -5.46160638e-01 5.92597872e-02
6.01704180e-01 1.05209008e-01 3.44554722e-01 -5.79840958e-01
-5.47351003e-01 -1.74664453e-01 -1.63731408e+00 6.17375553e-01
-4.73332912e-01 5.97447395e-01 5.80875993e-01 -1.19731355e+00
8.26364636e-01 8.44563961e-01 9.63537633e-01 -2.17266589e-01
6.05751693e-01 2.51712471e-01 -2.64007058e-02 -1.58427417e-01
4.98376675e-02 -6.20515458e-02 4.14212123e-02 4.82227743e-01
-1.37695283e-01 -4.07481760e-01 8.33270013e-01 1.02246709e-01
5.36781549e-01 -3.43843311e-01 4.01515424e-01 -1.02671039e+00
1.53291059e+00 1.80015445e-01 5.88637710e-01 -3.47301304e-01
-5.67433126e-02 -6.31767213e-01 1.00822735e+00 -4.71632481e-02
-9.35672581e-01 -5.73273778e-01 -3.03892642e-01 4.10237968e-01
5.80739155e-02 -1.95522472e-01 -8.06001365e-01 2.88893282e-01
2.10946053e-03 1.04257452e+00 -3.14020246e-01 -4.44743156e-01
-6.69741273e-01 -7.36766994e-01 3.02116871e-01 1.28237322e-01
-1.33129686e-01 -3.68200719e-01 -7.83674240e-01 3.30309629e-01
1.61350027e-01 -7.02868879e-01 5.23581579e-02 5.57751991e-02
-9.29548383e-01 -1.15621221e+00 -6.05109751e-01 -7.33707845e-01
1.00464213e+00 -1.85935289e-01 3.39014947e-01 1.35742918e-01
-3.34431708e-01 6.39410555e-01 -1.12922840e-01 -4.69341092e-02
-6.94754183e-01 -1.85124546e-01 3.88833374e-01 -2.27054730e-01
-2.17898756e-01 -6.20569170e-01 -1.26453146e-01 5.59141695e-01
-7.11270392e-01 -5.25771558e-01 -9.80802551e-02 7.02679634e-01
2.89359599e-01 8.68366063e-01 1.15687633e+00 -2.00541601e-01
1.06601989e+00 -2.54381180e-01 -1.46846545e+00 1.52879387e-01
-8.29335392e-01 -1.64646491e-01 1.01675177e+00 -4.17724520e-01
-9.38124776e-01 8.56914371e-02 1.56052664e-01 -3.37673455e-01
3.43311578e-01 5.82074702e-01 -3.31275677e-03 -6.81233346e-01
2.92579651e-01 7.48602375e-02 4.98721421e-01 -1.26792356e-01
8.22556317e-02 2.30333462e-01 2.68549383e-01 -4.58843827e-01
1.04278219e+00 1.35518432e-01 7.90757239e-01 -1.21434510e+00
5.05089983e-02 -2.36807972e-01 -4.48526263e-01 -4.08007175e-01
1.16192475e-01 -5.32853186e-01 -1.29014194e+00 6.27101719e-01
-9.92809236e-01 5.46598211e-02 -2.99466908e-01 4.34911609e-01
-4.53626692e-01 2.44038180e-01 -7.10831881e-01 -1.25197804e+00
-3.54181856e-01 -1.24593914e+00 1.38440549e-01 4.73686337e-01
-5.51020145e-01 -1.37511694e+00 -9.36767273e-03 -2.22105622e-01
2.92007118e-01 6.48020506e-01 9.57354188e-01 -1.88979641e-01
-9.23549533e-02 -8.49653408e-02 6.46482229e-01 1.95510566e-01
-4.40836735e-02 7.12522686e-01 -3.82878125e-01 -7.73363709e-01
1.96938008e-01 4.93536115e-01 -4.66600060e-01 6.65840268e-01
-7.49928430e-02 -2.93605953e-01 -4.09739047e-01 1.46362811e-01
1.83879733e+00 7.16905594e-01 2.98221856e-01 2.58538425e-01
8.35043713e-02 6.13597453e-01 7.74560452e-01 5.40772676e-01
-1.83913842e-01 4.78258997e-01 1.94535553e-02 -6.15056977e-02
5.64955711e-01 4.80605453e-01 4.78051186e-01 8.58596146e-01
8.87897238e-02 5.57684563e-02 -4.92601305e-01 5.23913801e-01
-1.54895580e+00 -6.85155690e-01 -5.27954161e-01 1.93850541e+00
4.90729868e-01 1.35467108e-02 4.48071182e-01 1.19709837e+00
1.01685095e+00 -2.95917541e-01 -2.18347371e-01 -9.71566796e-01
-9.12380666e-02 2.53702343e-01 6.05014324e-01 8.59102607e-01
-8.56998503e-01 8.04734752e-02 5.84977627e+00 1.05598342e+00
-1.52631652e+00 -3.45472515e-01 -2.60573067e-02 1.21397637e-01
4.55363616e-02 4.92720231e-02 -5.57296634e-01 4.06514943e-01
1.16637182e+00 -1.17739236e+00 3.21273744e-01 6.07599616e-01
9.21357274e-01 -3.75057220e-01 -5.99043846e-01 5.32511771e-01
-3.94426912e-01 -9.78712559e-01 -4.59158719e-01 -3.36730659e-01
1.01930404e+00 -1.05545485e+00 -7.10408315e-02 -2.88086951e-01
-5.31540632e-01 -4.07451004e-01 6.28742754e-01 4.27765578e-01
8.59227479e-02 -1.39490557e+00 7.76074588e-01 3.26805830e-01
-1.32839668e+00 -3.10999185e-01 1.21052116e-01 -1.35277718e-01
7.69940257e-01 5.85627377e-01 -6.25741184e-01 9.93524432e-01
9.20676067e-02 5.39760172e-01 -1.79160506e-01 9.36048388e-01
-1.81047082e-01 7.56104410e-01 -6.42343163e-01 -4.11857426e-01
-3.09043622e-04 -6.90826595e-01 9.66127813e-01 4.58916128e-01
1.82941467e-01 2.84725398e-01 -6.99549764e-02 1.12128222e+00
9.53079879e-01 5.57115711e-02 -1.97689638e-01 5.71142696e-02
8.12462926e-01 1.12452817e+00 -8.06387842e-01 -1.62351072e-01
1.26609996e-01 2.35118158e-02 -9.56805468e-01 6.90093994e-01
-8.51290464e-01 -1.03000331e+00 3.98110539e-01 1.63155556e-01
-5.86646982e-02 -2.04966053e-01 -4.25615549e-01 -3.66068810e-01
4.63828854e-02 -7.27211952e-01 2.84927994e-01 -5.59794128e-01
-7.63675690e-01 4.42785352e-01 4.34771299e-01 -1.64691377e+00
-1.04915571e+00 -1.47291660e-01 -8.52841914e-01 1.20258784e+00
-1.02650023e+00 -6.36920929e-01 6.31123483e-01 6.44690156e-01
1.18582316e-01 -3.39211673e-02 7.03018129e-01 5.69348335e-01
-8.86318624e-01 2.19047889e-01 6.79331481e-01 -3.73165816e-01
-4.86533828e-02 -8.41051757e-01 -3.71161520e-01 1.14844823e+00
-9.14018035e-01 6.47274196e-01 1.35493016e+00 -2.49677658e-01
-1.41180658e+00 -3.71876746e-01 6.62215292e-01 5.10917127e-01
1.04586494e+00 1.45625576e-01 -8.45907211e-01 1.93040505e-01
3.69688839e-01 -3.45993727e-01 3.16274375e-01 -7.50290632e-01
8.04111123e-01 -4.72618848e-01 -1.38820314e+00 3.67975354e-01
-4.47687745e-01 -5.00881910e-01 -4.62210059e-01 -1.31811082e-01
1.82400987e-01 -2.33592853e-01 -1.35600710e+00 2.15501875e-01
2.07829043e-01 -3.31693411e-01 9.06721890e-01 2.93851495e-02
-3.23166102e-01 -1.02782524e+00 4.56252903e-01 -1.61409736e+00
-7.61018917e-02 -1.47995615e+00 -8.45142826e-02 1.22528875e+00
2.89400309e-01 -1.11999428e+00 2.96989769e-01 4.55737412e-01
1.44389227e-01 -6.32468224e-01 -1.16240263e+00 -8.60159397e-01
2.22294569e-01 3.10043782e-01 3.04388106e-01 9.25090373e-01
9.43253517e-01 2.15172544e-01 2.90426880e-01 6.16367698e-01
6.01495028e-01 1.61809903e-02 2.21339822e-01 -1.07625318e+00
1.30693838e-01 -6.13429546e-01 -2.22396269e-01 -5.97009882e-02
1.96549922e-01 -2.26350293e-01 -4.14669156e-01 -1.21186030e+00
-8.71358156e-01 -9.45167914e-02 2.63431631e-02 1.95340246e-01
1.59774646e-01 -1.27261698e-01 -7.40482239e-03 -5.49519882e-02
3.57458085e-01 3.64909679e-01 8.63979340e-01 1.19489850e-02
-4.60663468e-01 4.06350762e-01 2.61403799e-01 5.58757782e-01
8.81797552e-01 -1.43833160e-01 -9.86765921e-01 4.31125730e-01
2.60713309e-01 7.51296580e-01 2.68951207e-01 -1.07547235e+00
2.99313724e-01 -1.61038078e-02 -1.38731226e-01 -9.93174613e-01
3.84792164e-02 -1.20843041e+00 7.38462210e-01 1.11032963e+00
-1.44909680e-01 2.32281834e-01 5.51711082e-01 9.99445170e-02
-3.75265747e-01 -3.34168106e-01 8.46225560e-01 6.30942762e-01
-2.76913971e-01 -7.58925200e-01 -1.19615924e+00 -4.81911808e-01
1.46366906e+00 -2.34534606e-01 -2.13472098e-01 -4.84691292e-01
-9.43271637e-01 3.81063432e-01 7.69310370e-02 -9.55532342e-02
-1.19004048e-01 -1.16366959e+00 -3.26766104e-01 1.24229848e-01
-8.38433087e-01 -5.32607019e-01 4.50362682e-01 1.27550566e+00
-6.50915325e-01 8.01514328e-01 -4.25835937e-01 -6.75497055e-01
-1.48575521e+00 3.53793383e-01 5.87248802e-01 -5.40551059e-02
-2.30752416e-02 1.61385357e-01 -5.92525303e-01 9.74296257e-02
-2.11839035e-01 -4.01377857e-01 -3.76559377e-01 4.07834858e-01
2.54593492e-01 1.12421083e+00 2.84100682e-01 -7.36139834e-01
-3.62402022e-01 8.92405272e-01 6.39678836e-01 -1.60753980e-01
1.13153493e+00 -4.94914770e-01 -5.05116642e-01 5.08746624e-01
8.90894115e-01 8.31875801e-02 -8.19343925e-01 4.77431625e-01
-2.83645570e-01 2.32834648e-02 5.31128161e-02 -4.57020342e-01
-1.27606153e+00 3.88196170e-01 4.51203614e-01 9.34317768e-01
1.37176538e+00 -8.82361293e-01 2.97593236e-01 2.36763000e-01
2.78363168e-01 -1.39780450e+00 -6.21947765e-01 4.15157437e-01
6.78658605e-01 -1.78609192e-01 4.30937260e-01 -9.67302680e-01
-1.18025087e-01 1.53482437e+00 9.68387052e-02 -5.52786231e-01
8.41402590e-01 5.29357135e-01 -1.93766691e-02 3.14681023e-01
-4.82132822e-01 3.51527214e-01 1.85225159e-01 3.59112322e-01
4.38526183e-01 -9.85259786e-02 -1.26209056e+00 7.08224535e-01
4.17928956e-02 4.49737720e-03 1.09908617e+00 1.08083332e+00
-3.25832367e-01 -1.13171899e+00 -1.00697434e+00 -1.34672120e-01
-5.58110058e-01 6.03106439e-01 -7.46399611e-02 1.02871251e+00
-3.04791719e-01 1.19251370e+00 -2.19486803e-01 1.39052480e-01
7.45334029e-01 -2.73321290e-02 3.13832730e-01 2.14618474e-01
-8.75743151e-01 3.52673024e-01 1.03773072e-01 -1.41961113e-01
-2.20864579e-01 -7.57964969e-01 -1.62844896e+00 -5.71283400e-01
-8.10697794e-01 8.47155213e-01 5.85325003e-01 7.48062074e-01
5.06545380e-02 9.45752740e-01 9.48148966e-01 -6.15433693e-01
-7.17560887e-01 -5.67345440e-01 -8.69538605e-01 -2.28391260e-01
2.76084542e-01 -9.02891695e-01 -6.98117316e-01 -1.29776537e-01] | [5.617125511169434, 2.676239490509033] |
0b3415e0-191d-462b-8c3c-750f3b22af56 | towards-robot-vision-module-development-with | null | null | https://openreview.net/forum?id=H1BHbmWCZ | https://openreview.net/pdf?id=H1BHbmWCZ | TOWARDS ROBOT VISION MODULE DEVELOPMENT WITH EXPERIENTIAL ROBOT LEARNING | n this paper we present a thrust in three directions of visual development us- ing supervised and semi-supervised techniques. The first is an implementation of semi-supervised object detection and recognition using the principles of Soft At- tention and Generative Adversarial Networks (GANs). The second and the third are supervised networks that learn basic concepts of spatial locality and quantity respectively using Convolutional Neural Networks (CNNs). The three thrusts to- gether are based on the approach of Experiential Robot Learning, introduced in previous publication. While the results are unripe for implementation, we believe they constitute a stepping stone towards autonomous development of robotic vi- sual modules. | ['Joanne Bechta Dugan', 'Ahmed A Aly'] | 2018-01-01 | null | null | null | iclr-2018-1 | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 2.52833784e-01 7.67540574e-01 2.30189916e-02 -3.58123571e-01
6.39228225e-02 -5.60077071e-01 1.18411255e+00 -5.38268924e-01
-2.67245263e-01 9.28779900e-01 -6.52620643e-02 -2.11267605e-01
9.82214604e-03 -9.35778558e-01 -1.14504898e+00 -6.66731775e-01
-2.52366990e-01 2.11881340e-01 2.27582976e-01 -7.06228912e-01
3.98586035e-01 8.82496715e-01 -1.63147783e+00 2.36762419e-01
5.79352379e-01 9.59009349e-01 6.39067218e-02 1.06277740e+00
-3.66258845e-02 1.54674423e+00 -4.98756796e-01 -8.98977816e-02
5.11311233e-01 -4.68742251e-01 -8.46156836e-01 9.53901932e-03
3.61142963e-01 -2.68729120e-01 -5.13934076e-01 9.85423982e-01
1.10561103e-01 -2.81130165e-01 1.20048225e+00 -1.43777621e+00
-1.10500371e+00 6.04540527e-01 -5.60617931e-02 -1.92700233e-02
1.47951841e-01 2.53820926e-01 3.81993502e-01 -8.75492156e-01
7.34170914e-01 1.28391659e+00 8.00810516e-01 8.40340853e-01
-9.39047933e-01 -4.20992166e-01 -2.35360891e-01 -1.22676156e-01
-1.00534117e+00 -3.04527014e-01 8.02891254e-01 -6.84928954e-01
8.48352849e-01 -1.02767602e-01 7.14912474e-01 1.06105113e+00
4.55657393e-01 8.45971406e-01 1.27931976e+00 -8.27689290e-01
5.22871494e-01 3.73061866e-01 -4.16810274e-01 9.02934849e-01
2.50028551e-01 9.50336993e-01 -3.92595768e-01 3.90798271e-01
1.53087509e+00 -1.99871480e-01 3.93290430e-01 -7.99676239e-01
-1.20920062e+00 9.24981952e-01 8.76999259e-01 6.22496665e-01
-2.62674332e-01 5.35276413e-01 2.07964256e-01 4.32412744e-01
1.48321703e-01 6.34214103e-01 -4.76620048e-02 -3.83343175e-02
-6.83813751e-01 6.71634749e-02 7.44305670e-01 1.16879630e+00
7.94074118e-01 8.01300943e-01 2.22482577e-01 4.03696954e-01
5.58123052e-01 4.51458514e-01 7.44603515e-01 -1.17768013e+00
-5.35058714e-02 4.48579729e-01 7.37140253e-02 -7.84946740e-01
-1.74647033e-01 1.42479047e-01 -5.59256911e-01 1.32231128e+00
2.82347292e-01 -5.58896184e-01 -1.43124998e+00 1.20858014e+00
-1.01755328e-01 -1.41118914e-01 2.38946393e-01 4.24852252e-01
1.00251067e+00 6.23674691e-01 7.89485350e-02 3.51665467e-01
6.12628102e-01 -1.14417171e+00 -5.42324781e-01 -6.25939369e-02
3.92974257e-01 -4.29884672e-01 4.65383202e-01 2.65657634e-01
-1.22449934e+00 -9.34728384e-01 -1.54225206e+00 9.95850116e-02
-1.18478703e+00 -6.90089464e-02 9.52742994e-01 6.53156698e-01
-1.48372066e+00 6.03807688e-01 -8.87245595e-01 -6.27420247e-01
6.77632451e-01 3.67163002e-01 -5.19797683e-01 3.11901450e-01
-7.91585386e-01 1.30409586e+00 4.89617735e-01 -4.24309857e-02
-1.40955544e+00 -2.19372600e-01 -1.00042677e+00 -4.16768253e-01
-2.76616037e-01 -3.23910445e-01 1.35185719e+00 -1.42645252e+00
-1.86329615e+00 1.10715055e+00 6.02320611e-01 -7.10841775e-01
4.90593791e-01 -1.09505601e-01 -2.68253922e-01 2.44618267e-01
-6.70914948e-02 1.09725070e+00 9.22093511e-01 -1.63024509e+00
-6.05208337e-01 -1.13606408e-01 6.10952498e-03 4.75709327e-02
-8.45117196e-02 9.46287438e-02 3.96750122e-01 -8.42273951e-01
3.11797559e-02 -6.70406938e-01 -2.82244951e-01 4.85307783e-01
-3.34082007e-01 1.80293899e-02 1.19387388e+00 -2.78087676e-01
1.63719952e-01 -1.92296290e+00 3.45321536e-01 -4.94908653e-02
1.38242498e-01 4.81807619e-01 -1.11598991e-01 5.15089631e-01
-3.46465439e-01 -1.40729219e-01 -1.58853233e-01 -7.11715147e-02
1.80577740e-01 4.90402997e-01 -2.81816602e-01 4.63091165e-01
4.61104423e-01 1.51967609e+00 -9.53470051e-01 -3.58841330e-01
5.86593807e-01 3.41463298e-01 7.54961371e-02 4.91089195e-01
-1.56348825e-01 3.42527807e-01 -1.66835845e-01 7.28877783e-01
3.54059547e-01 3.12247962e-01 -3.59694034e-01 2.99600214e-01
-5.38769305e-01 1.29528776e-01 -7.81165719e-01 1.68807542e+00
-3.80057871e-01 8.38453889e-01 6.21297508e-02 -1.50088298e+00
1.17973745e+00 1.82364911e-01 2.83942074e-01 -6.75748825e-01
2.43954793e-01 1.58574805e-01 -2.75678188e-01 -5.16170740e-01
2.37823457e-01 -1.19509771e-01 5.52331991e-02 1.53458759e-01
7.41350412e-01 -7.60133088e-01 6.73040971e-02 7.01695606e-02
9.68221068e-01 8.52539718e-01 2.60325193e-01 -4.56307024e-01
2.63543695e-01 3.45931739e-01 5.65877669e-02 7.13455319e-01
-1.86310679e-01 4.22321856e-01 5.05772114e-01 -6.64809346e-01
-1.38079906e+00 -1.40242243e+00 8.06738883e-02 9.80862677e-01
-1.00438997e-01 1.76138103e-01 -7.67690957e-01 -7.47911870e-01
-9.39252228e-03 5.58105826e-01 -1.13092661e+00 -1.27680168e-01
-4.09810692e-01 -5.08420654e-02 7.57194459e-01 9.88120854e-01
5.44409096e-01 -1.67275727e+00 -9.96039629e-01 7.58546963e-02
7.99513996e-01 -7.02801168e-01 5.71762562e-01 8.06061149e-01
-7.57585108e-01 -8.18605959e-01 -1.13594234e+00 -1.40608811e+00
8.74130666e-01 -3.50798033e-02 1.07049251e+00 -8.02040622e-02
-5.13174772e-01 6.89085245e-01 -5.62741578e-01 -9.92412925e-01
-8.30562949e-01 -4.03879136e-01 8.35092068e-02 -6.68646157e-01
3.08858871e-01 -8.86047959e-01 -3.99565250e-01 1.03710540e-01
-8.26610327e-01 7.75873140e-02 9.20811713e-01 7.73734868e-01
3.94862324e-01 -1.42493453e-02 4.06624019e-01 -7.83961356e-01
1.66368470e-01 -3.46277654e-01 -8.75998497e-01 1.02563158e-01
-5.06432652e-01 -4.99803238e-02 6.48816049e-01 -3.86982888e-01
-1.08717620e+00 4.48721379e-01 -1.80038586e-01 -2.21779764e-01
-7.49507666e-01 -2.30375715e-02 1.54967979e-01 -6.74607038e-01
1.06275845e+00 1.97417825e-01 1.81531757e-01 -5.06073274e-02
6.13375008e-01 5.12345612e-01 9.79294360e-01 -2.52167583e-01
1.26043999e+00 6.47006929e-01 6.94495812e-02 -8.18712711e-01
-2.85799563e-01 -3.72367166e-02 -1.01801813e+00 -2.88723797e-01
1.07228124e+00 -6.97336435e-01 -2.11097404e-01 8.58224034e-01
-1.08163297e+00 -8.98514986e-01 -9.93591845e-01 1.70444757e-01
-1.05901778e+00 1.85808897e-01 -6.70253038e-01 -8.34008634e-01
1.60435960e-02 -6.81849599e-01 8.06013107e-01 4.47322845e-01
1.71820983e-01 -1.23845673e+00 4.17742789e-01 -2.41831541e-01
5.79199791e-01 5.54847419e-01 4.53311682e-01 -4.94583905e-01
-4.82475102e-01 -2.70099878e-01 -1.46278501e-01 7.23460913e-01
7.79982731e-02 1.39987528e-01 -1.03345764e+00 3.14671025e-02
-1.80595756e-01 -8.34240079e-01 4.55926031e-01 3.73846769e-01
7.15738654e-01 -2.75832355e-01 -1.90769225e-01 6.81911469e-01
1.55476952e+00 4.72010702e-01 1.11091697e+00 4.95318562e-01
5.57194650e-01 4.86904293e-01 4.80079263e-01 4.73064370e-02
-3.90803069e-02 2.88052917e-01 6.98116064e-01 -4.77785528e-01
-4.20939982e-01 -5.01920700e-01 3.07253927e-01 4.56560969e-01
-4.21101183e-01 1.92990661e-01 -1.02223301e+00 5.91826081e-01
-1.68417168e+00 -9.51345384e-01 5.84366657e-02 1.66499650e+00
5.54934323e-01 2.87899226e-01 2.45149076e-01 9.20143723e-02
3.89655292e-01 -2.59721071e-01 -4.35586035e-01 -6.53054535e-01
-2.04138741e-01 6.01358593e-01 5.87700307e-01 1.76338047e-01
-1.20533264e+00 1.06373692e+00 8.30625057e+00 2.62036651e-01
-1.19165647e+00 -6.77800849e-02 3.26771855e-01 5.66196620e-01
2.67711952e-02 -8.23096707e-02 -5.03482819e-01 1.98637992e-01
8.15566540e-01 2.60292709e-01 3.90460938e-01 1.16126204e+00
-2.83040166e-01 -2.01737255e-01 -1.15759015e+00 5.69779217e-01
1.69985980e-01 -1.29208291e+00 -9.17991474e-02 -8.24100301e-02
9.88300383e-01 2.32002974e-01 2.04876855e-01 4.43853438e-01
1.04563844e+00 -1.53579283e+00 8.54815483e-01 6.08000100e-01
8.13354731e-01 -7.39067674e-01 6.62959218e-01 3.26023519e-01
-8.48588169e-01 -1.73812553e-01 -7.61467159e-01 -1.61425084e-01
-2.62509823e-01 -1.48798535e-02 -6.61038578e-01 2.27756202e-01
7.75524676e-01 6.98872268e-01 -6.32681906e-01 8.29791129e-01
-6.50790095e-01 4.24818069e-01 -4.01108861e-02 -4.19437945e-01
4.22772586e-01 2.84657516e-02 2.86646992e-01 1.39976203e+00
6.99456036e-02 -2.09367231e-01 -3.07631820e-01 1.14836884e+00
2.53849804e-01 -2.33447075e-01 -1.22937727e+00 -1.11085653e-01
9.90440398e-02 1.19862103e+00 -6.62172198e-01 -2.08495677e-01
-3.95694256e-01 1.00494027e+00 3.16423804e-01 3.54375958e-01
-6.25299692e-01 -7.69323528e-01 -4.88601252e-02 -8.34613144e-02
6.11555159e-01 -2.61331975e-01 -7.43403256e-01 -7.87169099e-01
-4.17492151e-01 -5.11467397e-01 -1.02456875e-01 -1.15846002e+00
-1.09959424e+00 3.89033645e-01 9.50164944e-02 -1.21738470e+00
-6.53611124e-01 -1.15613854e+00 -9.26597774e-01 5.73062241e-01
-1.23411810e+00 -1.60741591e+00 -3.74729842e-01 4.47857529e-01
4.42835182e-01 -7.10964918e-01 1.04427922e+00 -4.57921982e-01
6.91572875e-02 3.74635249e-01 2.25828350e-01 4.45147842e-01
1.48070797e-01 -1.68678916e+00 5.37881792e-01 7.96613872e-01
3.09449285e-01 3.59588861e-01 7.82522619e-01 -4.07098919e-01
-1.14623368e+00 -8.14103186e-01 2.35085845e-01 -3.58992130e-01
7.03211248e-01 -4.45757747e-01 -4.33337718e-01 8.96798551e-01
6.50948405e-01 1.79336265e-01 2.39197895e-01 -4.65704978e-01
-1.11954041e-01 1.54458344e-01 -1.35832846e+00 4.52097863e-01
8.29140306e-01 -5.01556635e-01 -9.96864259e-01 3.71584028e-01
4.91892397e-01 -2.41141990e-01 -8.60522509e-01 4.99517441e-01
5.38075566e-01 -1.16502500e+00 9.80140746e-01 -5.47016144e-01
7.84129918e-01 -1.63421094e-01 1.84657052e-02 -1.42630780e+00
-3.18878770e-01 -6.95368409e-01 -1.51493207e-01 9.40879881e-01
3.85861516e-01 -5.01626849e-01 9.92608607e-01 1.71952769e-01
-2.32203662e-01 -5.29403031e-01 -7.24090755e-01 -7.97593951e-01
5.33936739e-01 -9.23141167e-02 1.72395810e-01 9.20464575e-01
2.10593387e-01 8.96361172e-02 -9.36192423e-02 -4.02808934e-01
5.83199799e-01 -3.11745644e-01 9.07276034e-01 -9.98278558e-01
-1.56074688e-01 -3.23142380e-01 -1.06143606e+00 -7.78688669e-01
3.74222733e-02 -5.12441933e-01 4.38301653e-01 -1.68799484e+00
-2.15224117e-01 -1.00794360e-01 -2.27877259e-01 6.12258136e-01
6.75643265e-01 5.21908402e-01 -2.44452596e-01 1.31849095e-01
-5.73443949e-01 5.16555965e-01 1.03847170e+00 -1.03516608e-01
-1.05066877e-02 -8.97403061e-02 -4.19284582e-01 1.08502412e+00
9.80610192e-01 -1.58528790e-01 -3.40800792e-01 -1.67775527e-01
-5.89243770e-02 -3.60280782e-01 8.06518674e-01 -1.36129880e+00
2.63864517e-01 -7.01177642e-02 7.44082749e-01 -4.57711160e-01
1.17394105e-01 -8.44312727e-01 -2.38989368e-01 7.33282328e-01
-2.91996419e-01 -7.42340609e-02 2.55463094e-01 2.78289288e-01
-2.21839160e-01 -4.15890425e-01 1.20890510e+00 -6.51177764e-01
-1.36109841e+00 -1.88121989e-01 -7.39259243e-01 -4.17002171e-01
1.34938192e+00 -4.91152138e-01 -1.76335260e-01 -4.22921389e-01
-7.42368639e-01 -2.47719586e-01 7.68766224e-01 4.00008261e-01
7.37760007e-01 -1.30966210e+00 -5.99853635e-01 3.96389633e-01
7.29156882e-02 9.35553983e-02 -3.62639368e-01 1.72188073e-01
-1.11114788e+00 6.18105233e-01 -1.14626479e+00 -4.51502055e-01
-6.13837361e-01 7.47070074e-01 4.45088446e-01 2.36063048e-01
-2.79700726e-01 1.03675723e+00 6.29387200e-02 -6.41847312e-01
3.87425572e-01 -6.68449923e-02 -3.21035504e-01 -9.27845478e-01
3.61512989e-01 4.09356892e-01 -4.03338522e-01 -7.25962579e-01
7.44897127e-02 3.01497996e-01 2.96014845e-01 -2.49939248e-01
1.59433210e+00 3.27737838e-01 -1.75923124e-01 6.00740850e-01
8.40364337e-01 -3.74921173e-01 -1.42458510e+00 2.06860006e-01
-2.33613551e-02 5.74261844e-02 -1.02955043e-01 -8.95503640e-01
-9.70946372e-01 9.31257010e-01 8.28585565e-01 2.55965114e-01
8.88309240e-01 1.33082598e-01 1.47557870e-01 4.90629196e-01
5.44589281e-01 -1.24614930e+00 3.08576882e-01 5.10664582e-01
1.01677370e+00 -1.06469083e+00 -2.06833631e-02 -2.51745015e-01
-4.12477940e-01 1.23488557e+00 9.39786375e-01 -8.40334475e-01
7.38634348e-01 8.15095305e-01 1.92870229e-01 -4.02320892e-01
-2.33592242e-01 -2.31073707e-01 1.58165231e-01 1.38224828e+00
3.21733832e-01 -1.25974283e-01 5.54083884e-02 -1.31911322e-01
-7.80234337e-02 1.67526528e-01 2.64973611e-01 1.47152221e+00
-6.77809417e-01 -8.95561278e-01 -3.70400727e-01 1.45249605e-01
-2.66099781e-01 2.56994039e-01 -6.55816972e-01 1.13634372e+00
3.59686315e-01 7.16890633e-01 -3.04848533e-02 -5.17033160e-01
1.87433064e-01 -1.65258586e-01 8.33244681e-01 -6.41773283e-01
-2.42044106e-01 -4.81405169e-01 -1.29058093e-01 -3.54521096e-01
-7.90302992e-01 -4.03566182e-01 -1.15252376e+00 -6.92412481e-02
-1.48247764e-01 -1.91945404e-01 1.03222167e+00 8.56961131e-01
-2.03156963e-01 3.59543055e-01 5.28833330e-01 -1.38240254e+00
-4.02925909e-01 -1.23139966e+00 -8.25833082e-01 9.36204866e-02
1.21488944e-01 -6.98466122e-01 -2.75432110e-01 4.34098244e-01] | [9.768508911132812, 2.0912349224090576] |
5c229098-0836-4db7-bd31-0ac0ac4919f2 | sample-average-approximation-for-black-box-vi | 2304.06803 | null | https://arxiv.org/abs/2304.06803v2 | https://arxiv.org/pdf/2304.06803v2.pdf | Sample Average Approximation for Black-Box VI | We present a novel approach for black-box VI that bypasses the difficulties of stochastic gradient ascent, including the task of selecting step-sizes. Our approach involves using a sequence of sample average approximation (SAA) problems. SAA approximates the solution of stochastic optimization problems by transforming them into deterministic ones. We use quasi-Newton methods and line search to solve each deterministic optimization problem and present a heuristic policy to automate hyperparameter selection. Our experiments show that our method simplifies the VI problem and achieves faster performance than existing methods. | ['Daniel Sheldon', 'Justin Domke', 'Javier Burroni'] | 2023-04-13 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 6.42867237e-02 -2.15431988e-01 -2.26767197e-01 -1.25478357e-01
-1.39268839e+00 -6.53807521e-01 4.87284869e-01 -1.55900836e-01
-7.05366790e-01 1.33951044e+00 -1.74141470e-02 -7.85595834e-01
-1.61132142e-01 -6.16758764e-01 -6.28668785e-01 -1.02639079e+00
1.50467455e-01 7.62875497e-01 3.26570153e-01 -3.69106174e-01
6.78353012e-01 4.78357792e-01 -1.44494295e+00 -4.67773043e-02
1.15672433e+00 9.07448173e-01 -6.94977194e-02 1.01553774e+00
-4.18382764e-01 3.40704113e-01 -7.55742073e-01 -4.82177854e-01
4.91514236e-01 -7.23180532e-01 -7.68250823e-01 1.62076756e-01
3.99060063e-02 -2.12409988e-01 1.41294807e-01 9.01591301e-01
6.35472178e-01 6.19658113e-01 8.03448617e-01 -1.25843906e+00
-9.75630581e-02 5.46143591e-01 -8.01119030e-01 1.52839795e-01
4.11865383e-01 3.54174286e-01 9.03254449e-01 -9.93617892e-01
4.44920778e-01 1.40488863e+00 8.55462193e-01 4.00702745e-01
-1.35187626e+00 -8.89811069e-02 3.30758780e-01 -1.48843184e-01
-1.44400573e+00 -2.32887045e-01 4.81699526e-01 -1.04898386e-01
9.74007487e-01 5.54227352e-01 8.76792014e-01 5.90758562e-01
-3.31502669e-02 1.08848429e+00 1.10215187e+00 -7.06381857e-01
8.36477458e-01 3.24661434e-01 -4.79662511e-03 7.84288585e-01
2.98400402e-01 -9.78961214e-02 -2.10108191e-01 -9.26223040e-01
5.48817456e-01 -3.70999336e-01 -2.92548925e-01 -5.39070189e-01
-8.46441031e-01 1.20041513e+00 -1.88483417e-01 -2.48977736e-01
-5.30533612e-01 4.72038478e-01 3.30073774e-01 2.21565560e-01
3.61599147e-01 6.75618291e-01 -5.59686780e-01 -2.35205874e-01
-1.01608264e+00 8.70972931e-01 1.18425059e+00 7.22357929e-01
7.70495534e-01 3.28600615e-01 -3.64052176e-01 9.26412404e-01
1.05097741e-01 6.72825873e-01 4.39276248e-01 -1.25269699e+00
4.22838539e-01 4.60331403e-02 7.83397436e-01 -6.10862017e-01
-1.11585326e-01 -2.98215926e-01 -1.21230587e-01 4.54761893e-01
6.91668034e-01 -5.64848185e-01 -7.86056638e-01 1.23200917e+00
7.45598078e-01 5.12031689e-02 -1.27902880e-01 6.28503382e-01
8.52372646e-02 9.66328621e-01 -3.23376715e-01 -7.75568962e-01
5.86459517e-01 -1.13506651e+00 -6.14886999e-01 5.95807694e-02
6.29324198e-01 -5.92500031e-01 1.36969221e+00 6.18323743e-01
-1.55698323e+00 4.07967567e-02 -1.07097232e+00 4.11686808e-01
-2.17959151e-01 -6.05520196e-02 7.46474385e-01 9.16828275e-01
-1.19553089e+00 9.09308732e-01 -6.64211929e-01 1.37811050e-01
2.98919201e-01 5.81485212e-01 4.80757207e-01 3.60499054e-01
-7.73987949e-01 7.44199395e-01 5.43254197e-01 1.78270519e-01
-7.14074373e-01 -6.83260798e-01 -7.05563068e-01 -1.34461280e-02
7.95365870e-01 -7.89515436e-01 1.50408912e+00 -8.55769277e-01
-2.24976897e+00 4.29840684e-01 -4.85517353e-01 -5.72792530e-01
8.74225974e-01 -2.53748864e-01 3.02926034e-01 -2.03815430e-01
-4.55503762e-01 2.52616614e-01 1.10731614e+00 -1.48647773e+00
-6.99227035e-01 6.27589747e-02 -9.37347263e-02 4.11691695e-01
-2.21183106e-01 9.82376933e-02 -4.81397063e-01 -4.95367616e-01
-6.14076145e-02 -1.05822647e+00 -9.61334944e-01 -2.30510741e-01
-6.61674619e-01 -1.08935095e-01 3.50428164e-01 -2.19789267e-01
1.76061416e+00 -1.61895442e+00 3.29473674e-01 7.57560313e-01
4.99488935e-02 2.19400600e-01 3.06086428e-03 4.79139864e-01
2.68856466e-01 2.19793320e-01 -7.43095696e-01 -4.15808827e-01
1.20237954e-01 1.83237240e-01 -5.13611913e-01 3.77601147e-01
-2.55334169e-01 7.57677317e-01 -1.08393550e+00 -5.55357873e-01
-7.92669058e-02 4.86389548e-02 -8.33494723e-01 1.17705412e-01
-4.41757947e-01 -1.05919372e-02 -5.62803447e-01 4.61578578e-01
6.37840211e-01 -1.53700560e-01 1.96116164e-01 1.72717482e-01
-1.71300799e-01 2.16115028e-01 -1.75588179e+00 1.15584207e+00
-6.46601737e-01 2.77583122e-01 1.21680804e-01 -9.60451305e-01
7.58894324e-01 -1.88297346e-01 4.77217436e-01 1.19574815e-01
2.28429526e-01 4.47337717e-01 -3.11060578e-01 -2.13790327e-01
4.15650547e-01 5.12907729e-02 2.19257563e-01 6.44370079e-01
-2.83873826e-01 -7.59233475e-01 4.75506216e-01 1.10116452e-01
6.68470025e-01 2.83878684e-01 5.28162360e-01 -4.27467018e-01
7.27940202e-01 2.50079274e-01 3.48562121e-01 1.34588432e+00
7.46914605e-03 4.65749770e-01 5.79788327e-01 -7.59935737e-01
-9.21012104e-01 -9.68934655e-01 5.76943997e-03 1.11937344e+00
-1.04004622e-01 -3.69449735e-01 -9.36842501e-01 -7.92680919e-01
3.45004201e-01 1.06110454e+00 -6.17802620e-01 9.99888182e-02
-8.16311002e-01 -1.23106897e+00 -9.87455547e-02 2.19416246e-01
5.82943670e-02 -8.01177144e-01 -5.41245461e-01 4.78494167e-01
1.57281861e-01 -6.15256488e-01 -5.96761167e-01 2.10671604e-01
-1.14985752e+00 -5.73704064e-01 -8.60870659e-01 -4.13291872e-01
6.90462887e-01 -1.82178877e-02 1.26101363e+00 2.05178678e-01
-3.04078728e-01 9.79809538e-02 1.24953374e-01 -5.38738906e-01
-4.55204189e-01 1.11587137e-01 1.14427350e-01 -1.39067084e-01
2.82426234e-02 -4.45610255e-01 -5.08433282e-01 1.98811755e-01
-4.04666573e-01 -2.50660628e-01 1.42072186e-01 1.09836292e+00
1.02200973e+00 -1.25475496e-01 2.75625527e-01 -1.15460730e+00
1.24495006e+00 -4.13732290e-01 -1.18535423e+00 3.18372607e-01
-8.81052852e-01 5.69527864e-01 7.77728856e-01 -6.46797895e-01
-9.82503653e-01 2.70085931e-01 -2.05899477e-01 -3.10001194e-01
4.04687643e-01 2.41466179e-01 2.86906749e-01 -4.72948015e-01
8.71034384e-01 2.16755107e-01 4.88588586e-02 -3.77587825e-01
4.36710179e-01 5.13344049e-01 3.63365948e-01 -9.50402498e-01
8.51013303e-01 4.80776817e-01 -3.39955091e-02 -6.39955580e-01
-8.46557856e-01 -2.42573708e-01 -6.68166578e-02 -8.39902684e-02
2.08931968e-01 -2.33021796e-01 -8.03737998e-01 2.05602571e-01
-9.19102013e-01 -6.17959261e-01 -6.23146713e-01 2.49630570e-01
-9.18696523e-01 2.93343633e-01 -3.24169755e-01 -1.07754374e+00
-4.56304669e-01 -1.35707510e+00 9.04491246e-01 2.32105285e-01
-2.08047137e-01 -1.04776847e+00 3.39149892e-01 -1.81095630e-01
4.11838531e-01 8.47288594e-02 6.40774250e-01 -5.05524993e-01
-3.35693300e-01 -1.28749341e-01 1.97473407e-01 -8.70951042e-02
-2.78337628e-01 3.67432237e-01 -5.56318045e-01 -4.43264633e-01
1.03601485e-01 -1.87823176e-01 7.07594454e-01 9.68110859e-01
1.47239590e+00 -7.19038606e-01 -4.51460540e-01 1.14925194e+00
1.64903963e+00 2.51418918e-01 2.90539622e-01 7.77297556e-01
3.36568922e-01 2.47223184e-01 7.30030596e-01 9.90793824e-01
3.68187651e-02 6.37056410e-01 5.90263270e-02 -3.61423939e-02
4.17240530e-01 -2.33215228e-01 1.39363945e-01 2.38765404e-01
1.37207704e-02 -2.42176831e-01 -8.65711570e-01 4.87726569e-01
-1.99269092e+00 -7.19228745e-01 4.44084033e-02 2.29602504e+00
1.01649368e+00 2.20737740e-01 6.97915733e-01 5.65123856e-02
4.80390877e-01 -8.90415683e-02 -8.28652501e-01 -9.61398900e-01
8.20040405e-02 1.65012643e-01 7.97244370e-01 9.67706680e-01
-7.90935159e-01 1.08694851e+00 8.73163223e+00 1.13328099e+00
-6.09750867e-01 -2.07515821e-01 7.55890191e-01 -4.84556645e-01
-6.58940732e-01 7.60377273e-02 -1.07634687e+00 5.04599571e-01
9.03577745e-01 -4.35140699e-01 1.01036310e+00 9.75678742e-01
3.26274991e-01 -3.51183385e-01 -8.00667167e-01 8.26128185e-01
-3.10482413e-01 -1.60057998e+00 1.35070160e-01 -2.92233527e-01
1.22052085e+00 -2.79572368e-01 9.80078578e-02 2.30020463e-01
9.10053194e-01 -9.37429488e-01 4.81613755e-01 1.96442395e-01
3.64836514e-01 -1.14167106e+00 5.72330952e-01 2.84186929e-01
-8.79583657e-01 -3.73185396e-01 -4.06004608e-01 1.08850546e-01
3.20100695e-01 5.55354059e-01 -6.65893257e-01 -9.98705626e-02
6.21804178e-01 -1.81227967e-01 -8.68079960e-02 1.35712850e+00
-2.63323575e-01 8.09306562e-01 -8.12950671e-01 -7.04604030e-01
5.34457266e-01 -8.20204616e-01 8.34420562e-01 1.25665283e+00
1.81707650e-01 7.86396638e-02 3.28757674e-01 8.14716578e-01
3.10130686e-01 5.42451084e-01 -3.35958421e-01 2.42265418e-01
7.24480569e-01 7.34485507e-01 -8.25080991e-01 -5.67190826e-01
-1.30785316e-01 8.62228274e-01 3.62596482e-01 8.35391700e-01
-7.80031264e-01 -7.63618827e-01 5.00189483e-01 -2.75585413e-01
6.30111396e-01 -1.68140933e-01 -5.64438581e-01 -9.48458254e-01
-1.38372079e-01 -1.31059361e+00 5.56208611e-01 -3.83550882e-01
-8.23317945e-01 5.97425640e-01 3.40763539e-01 -7.57491410e-01
-6.93741679e-01 -4.78739023e-01 -8.15507889e-01 8.27551961e-01
-1.11948180e+00 -2.36200392e-01 1.30092949e-01 3.58580709e-01
7.45618165e-01 -3.61299701e-02 3.19624662e-01 -4.09068018e-01
-6.36327624e-01 8.35212350e-01 7.21922219e-01 -6.78460777e-01
2.15370968e-01 -1.42750537e+00 6.53603792e-01 8.02415907e-01
-1.78641289e-01 7.22674608e-01 1.24730778e+00 -3.93285036e-01
-1.48676980e+00 -3.26923132e-01 4.67822015e-01 -5.39895147e-02
8.38166654e-01 -1.94577172e-01 -7.30483890e-01 3.31304133e-01
-9.25894603e-02 -7.84621313e-02 3.19930464e-01 2.10786760e-01
2.59668857e-01 2.54881307e-02 -1.29472423e+00 9.72667754e-01
9.15364802e-01 -7.10339611e-03 -1.20931707e-01 5.91978371e-01
5.55173516e-01 -7.52786636e-01 -5.22965610e-01 9.89879370e-02
3.62795234e-01 -6.72797382e-01 1.25152087e+00 -6.76431477e-01
7.78570073e-03 -2.72779256e-01 2.12772533e-01 -1.65501535e+00
1.17494948e-02 -1.46638024e+00 -5.81903279e-01 6.16470635e-01
6.20758057e-01 -8.97708833e-01 9.29268956e-01 7.18681574e-01
3.25164557e-01 -1.48406982e+00 -7.10348785e-01 -8.88231814e-01
2.57219166e-01 -1.86836019e-01 7.85704613e-01 3.26745570e-01
-1.56201512e-01 -9.53793824e-02 -2.59887159e-01 -1.99467279e-02
8.83976519e-01 4.15645421e-01 7.49993682e-01 -7.16814518e-01
-7.89525747e-01 -6.61627173e-01 2.05372646e-01 -1.10955548e+00
1.06112845e-01 -3.45836103e-01 2.18738571e-01 -1.27546668e+00
-8.75633210e-02 -3.98494333e-01 -1.61226749e-01 6.32847175e-02
-6.39379680e-01 -2.39969473e-02 2.92831846e-02 -2.52186377e-02
-6.02049828e-01 5.73789001e-01 8.64658594e-01 1.32614866e-01
-9.16580975e-01 3.35658222e-01 -4.85019237e-01 8.60391200e-01
8.04918468e-01 -5.51545560e-01 -4.35682625e-01 -1.32455036e-01
3.15239549e-01 1.82968140e-01 -3.53771955e-01 -6.61148727e-01
3.69207584e-04 -5.67121744e-01 8.54343325e-02 -6.14663661e-01
4.26193327e-02 -2.24009082e-01 -1.27993494e-01 5.83237767e-01
-4.62192535e-01 3.84506553e-01 2.63967097e-01 4.19919342e-01
1.24452084e-01 -8.44095767e-01 9.33459938e-01 -1.95529938e-01
-2.18678743e-01 1.91543341e-01 -7.35532701e-01 3.90903503e-01
8.16946089e-01 -1.99937910e-01 2.02673644e-01 -5.78581274e-01
-6.80588722e-01 5.59265077e-01 6.04686260e-01 -3.66967380e-01
5.32987893e-01 -9.62459326e-01 -6.86981201e-01 -1.17272109e-01
-3.64541918e-01 -2.29855493e-01 -3.06242853e-01 5.89862108e-01
-8.37100446e-01 -2.57102903e-02 3.39650214e-01 -3.89392108e-01
-1.01996219e+00 7.27526724e-01 6.02624893e-01 -4.49053049e-01
-4.41616595e-01 1.25507045e+00 -3.55922043e-01 -3.19653362e-01
5.11589885e-01 6.74421759e-03 1.98709522e-03 -1.77661493e-01
7.27628767e-01 8.65119457e-01 -1.86811611e-01 1.61163345e-01
-1.79114193e-01 5.19004166e-01 -2.43029281e-01 -5.72709799e-01
1.31342757e+00 -1.45454898e-01 -3.35810371e-02 4.00149465e-01
1.18918467e+00 1.44874841e-01 -1.27060664e+00 -1.32832691e-01
-9.62018967e-05 -6.43255472e-01 1.98176995e-01 -5.63683927e-01
-6.59432530e-01 3.23752433e-01 3.44280422e-01 5.53625077e-02
1.02039933e+00 -5.37731707e-01 8.12940657e-01 9.90631521e-01
2.93551385e-01 -1.49851000e+00 -3.05883855e-01 5.05076706e-01
6.95921719e-01 -1.03101361e+00 5.04874647e-01 -3.40113163e-01
-6.20449364e-01 1.16459608e+00 4.04782414e-01 -2.89679855e-01
4.39333797e-01 5.67726851e-01 1.19154088e-01 9.78141204e-02
-9.10162210e-01 -2.40668088e-01 2.98803933e-02 3.15405309e-01
-6.57959282e-02 -1.93731114e-01 -9.13039744e-01 5.55977458e-03
-2.63143986e-01 -2.06780374e-01 3.25752318e-01 1.01976204e+00
-4.74675566e-01 -1.08372259e+00 -6.48860097e-01 4.26077753e-01
-4.78467643e-01 -2.15895832e-01 -2.03908414e-01 6.84446752e-01
-6.27176642e-01 7.25004733e-01 -3.01033139e-01 -1.03272475e-01
1.26694724e-01 1.94635876e-02 5.14631152e-01 -1.11517996e-01
-6.72092974e-01 7.58282617e-02 2.28644460e-01 -7.77628601e-01
1.66083500e-01 -8.74872863e-01 -1.11064208e+00 -3.38668466e-01
-3.88838142e-01 6.71803951e-01 7.86059678e-01 1.00436378e+00
1.71969067e-02 1.20328218e-01 1.01028955e+00 -9.64626074e-01
-1.26189733e+00 -2.90529668e-01 -3.48470122e-01 -3.57711542e-04
3.16190749e-01 -6.12729132e-01 -7.27824152e-01 -4.07112330e-01] | [6.650849342346191, 4.051882743835449] |
f3ea5920-0abe-427a-9880-b189e3b9a9b8 | zero-shot-single-image-restoration-through | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Kar_Zero-Shot_Single_Image_Restoration_Through_Controlled_Perturbation_of_Koschmieders_Model_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Kar_Zero-Shot_Single_Image_Restoration_Through_Controlled_Perturbation_of_Koschmieders_Model_CVPR_2021_paper.pdf | Zero-Shot Single Image Restoration Through Controlled Perturbation of Koschmieder's Model | Real-world image degradation due to light scattering can be described based on the Koschmieder's model. Training deep models to restore such degraded images is challenging as real-world paired data is scarcely available and synthetic paired data may suffer from domain-shift issues. In this paper, a zero-shot single real-world image restoration model is proposed leveraging a theoretically deduced property of degradation through the Koschmieder's model. Our zero-shot network estimates the parameters of the Koschmieder's model, which describes the degradation in the input image, to perform image restoration. We show that a suitable degradation of the input image amounts to a controlled perturbation of the Koschmieder's model that describes the image's formation. The optimization of the zero-shot network is achieved by seeking to maintain the relation between its estimates of Koschmieder's model parameters before and after the controlled perturbation, along with the use of a few no-reference losses. Image dehazing and underwater image restoration are carried out using the proposed zero-shot framework, which in general outperforms the state-of-the-art quantitatively and subjectively on multiple standard real-world image datasets. Additionally, the application of our zero-shot framework for low-light image enhancement is also demonstrated. | ['Prabir Kumar Biswas', 'Debashis Sen', 'Sobhan Kanti Dhara', 'Aupendu Kar'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['underwater-image-restoration', 'image-dehazing'] | ['computer-vision', 'computer-vision'] | [ 4.86298859e-01 -4.53912131e-02 6.11516416e-01 -1.12149812e-01
-6.81577981e-01 -2.27553807e-02 4.12153482e-01 -3.84928733e-01
-5.64193428e-01 8.13050926e-01 8.01424533e-02 2.12444738e-02
-3.81389081e-01 -5.68358779e-01 -1.01699603e+00 -1.44141042e+00
6.32223114e-02 -1.65317580e-01 8.98652151e-02 -6.55635118e-01
2.75539845e-01 2.68471986e-01 -1.93722618e+00 -1.33687273e-01
1.05888486e+00 1.13061702e+00 5.31999588e-01 9.80865836e-01
4.04824942e-01 7.12390006e-01 -7.81248510e-01 -1.29422486e-01
5.28923571e-01 -3.97240788e-01 -3.92058164e-01 1.85218900e-01
4.23433512e-01 -6.67675853e-01 -7.10394144e-01 1.55133653e+00
7.82928109e-01 4.32400256e-01 6.46773219e-01 -8.97498429e-01
-8.52280200e-01 -1.10682301e-01 -3.74037653e-01 3.82792145e-01
1.12287074e-01 4.89793688e-01 4.17208850e-01 -7.57039249e-01
6.20542228e-01 9.63578820e-01 6.43434107e-01 6.08546972e-01
-1.29790461e+00 -3.73268545e-01 -3.69950593e-01 5.91549456e-01
-1.31236267e+00 -6.25299454e-01 6.88483000e-01 -2.54673272e-01
6.77138329e-01 1.46447923e-02 5.07958710e-01 7.84905612e-01
4.35420811e-01 2.37440482e-01 1.19270325e+00 -6.60166681e-01
3.94000530e-01 -1.14283882e-01 2.82885530e-03 3.63658875e-01
1.58386663e-01 5.09294689e-01 -6.76478326e-01 1.69298247e-01
5.77626288e-01 -2.35392481e-01 -8.78314734e-01 -2.72182226e-01
-6.93499863e-01 3.61917675e-01 5.42147517e-01 1.67205527e-01
-4.27091897e-01 5.09172864e-02 9.88767445e-02 5.56506336e-01
5.49038529e-01 4.03742492e-01 -4.66251224e-02 2.32112750e-01
-9.53681409e-01 1.38135239e-01 4.77543771e-01 7.61494637e-01
8.80830705e-01 2.85728276e-01 -2.04403117e-01 9.84249949e-01
3.20648998e-01 5.54741383e-01 4.64078993e-01 -1.28062463e+00
1.39849365e-01 -1.82660867e-03 6.24107897e-01 -6.77626193e-01
-1.46284044e-01 -4.25862968e-01 -1.11232340e+00 7.89193392e-01
3.95060688e-01 1.58219054e-01 -1.27991021e+00 1.72673738e+00
8.09547231e-02 5.79896748e-01 7.14144349e-01 1.16416514e+00
9.24418151e-01 7.76592076e-01 -4.07820702e-01 -5.14678597e-01
9.93906558e-01 -4.82006639e-01 -8.69522631e-01 -2.07594514e-01
1.31879702e-01 -5.04260063e-01 1.16435063e+00 4.07550603e-01
-1.11802208e+00 -4.29525524e-01 -1.45224118e+00 -2.97023617e-02
1.73330158e-02 -2.77128041e-01 -1.01090766e-01 5.37822247e-01
-1.20145667e+00 8.36524665e-01 -5.14938056e-01 -2.38432065e-01
3.88618767e-01 1.51373103e-01 -4.30655688e-01 -4.77559805e-01
-1.28743672e+00 1.01299798e+00 1.41456425e-01 4.85272706e-01
-1.31989682e+00 -9.06338871e-01 -7.94526219e-01 -8.07870738e-03
6.81059882e-02 -6.99778616e-01 1.13406384e+00 -8.28114271e-01
-1.63123262e+00 8.93006444e-01 1.57941416e-01 -5.18781364e-01
5.40224493e-01 5.77298366e-02 -3.42176557e-01 5.39701939e-01
-1.48567647e-01 2.16393262e-01 1.30469728e+00 -1.86841714e+00
-1.67408541e-01 -2.19537571e-01 4.11246829e-02 2.64729410e-01
-1.73053220e-01 -1.19575031e-01 -1.70437559e-01 -5.22996068e-01
1.04805052e-01 -4.72723097e-01 -1.41841337e-01 4.84258860e-01
-1.37510166e-01 5.73421896e-01 8.25141490e-01 -7.89681375e-01
7.14239538e-01 -2.31082296e+00 2.63970848e-02 -2.44065389e-01
1.79495275e-01 6.02382898e-01 -4.11108464e-01 4.43695813e-01
5.27362432e-03 -2.39212081e-01 -8.54880989e-01 -5.27661264e-01
-2.36652985e-01 3.99713874e-01 -2.69593745e-01 1.09115255e+00
-2.15698555e-01 4.33528215e-01 -8.39409292e-01 -1.13431327e-01
4.13131833e-01 7.82821834e-01 -3.09184700e-01 5.26631415e-01
1.04165249e-01 5.67835271e-01 4.00315136e-01 5.25169194e-01
9.96227682e-01 1.09238155e-01 -1.82166263e-01 -5.82336187e-01
-2.36119524e-01 -3.81955504e-01 -7.62368321e-01 1.39558053e+00
-5.24701536e-01 8.69207919e-01 3.45967799e-01 -1.01747847e+00
8.41727734e-01 2.62581646e-01 2.20701829e-01 -1.03687692e+00
2.79761672e-01 2.60940731e-01 -1.22812405e-01 -9.57356751e-01
2.95832485e-01 -7.52541900e-01 4.70084250e-01 1.91415250e-01
2.44028494e-01 -3.55345935e-01 1.44431949e-01 1.47065327e-01
9.67983186e-01 -6.58469051e-02 1.70797974e-01 -5.39843380e-01
3.93122107e-01 -4.48160499e-01 4.17624295e-01 9.17391717e-01
-4.24125791e-01 1.03698361e+00 -1.00979783e-01 -2.04747111e-01
-1.20800352e+00 -1.04085219e+00 -3.58735859e-01 3.66669387e-01
6.62132800e-01 4.31806654e-01 -7.22529531e-01 1.46107361e-01
-4.07781661e-01 8.52159441e-01 -6.26262069e-01 -5.65579772e-01
-2.27787808e-01 -1.09775269e+00 5.05758762e-01 -8.08628500e-02
8.89423728e-01 -8.67051542e-01 -8.43202472e-01 -2.62748245e-02
-3.99976969e-01 -1.04347992e+00 -8.28148052e-02 2.67651349e-01
-5.87586820e-01 -1.10897851e+00 -9.73996520e-01 -6.94231391e-01
6.88236177e-01 5.24723828e-01 9.58245099e-01 2.24353984e-01
-4.02403682e-01 3.69185597e-01 -3.23253065e-01 -1.78867474e-01
-7.27479041e-01 -1.07452035e+00 6.39493242e-02 3.42036098e-01
-2.55529702e-01 -8.79173219e-01 -9.94194329e-01 2.41425872e-01
-1.40930557e+00 -1.92902580e-01 3.23560923e-01 1.23777807e+00
3.40846211e-01 4.10150230e-01 3.68552834e-01 -1.63638830e-01
3.42419058e-01 -1.61080971e-01 -4.47529644e-01 1.30217001e-01
-6.65899754e-01 5.14141768e-02 5.27934194e-01 -4.71055627e-01
-1.41631043e+00 -3.37095529e-01 -6.17894940e-02 -7.54466534e-01
-5.13654528e-03 2.66584992e-01 -2.05631703e-01 -5.78193724e-01
7.82085061e-01 4.58661437e-01 2.71867275e-01 -4.03326035e-01
1.54726610e-01 6.83342934e-01 1.13990951e+00 -1.72786206e-01
1.04913974e+00 7.81605780e-01 1.37889355e-01 -1.25460494e+00
-7.75210857e-01 -3.49358946e-01 -3.22974950e-01 -5.49342632e-01
6.54517353e-01 -9.46254134e-01 -7.09498286e-01 1.26205659e+00
-1.11044228e+00 -6.00286186e-01 -4.51582938e-01 3.45692754e-01
-7.86724150e-01 7.15729058e-01 -7.21417546e-01 -9.51090753e-01
-2.43311450e-01 -9.85012531e-01 8.17026377e-01 3.80348295e-01
6.44474506e-01 -8.57060790e-01 3.45926546e-02 2.45779037e-01
4.61330235e-01 -2.44185980e-02 8.20643187e-01 3.09622496e-01
-5.21795928e-01 -2.08284846e-03 -3.76121551e-01 8.72356832e-01
-2.10648961e-02 -1.20838352e-01 -1.10207450e+00 -6.85370982e-01
5.77759445e-01 -2.38289848e-01 1.12469876e+00 7.63429701e-01
7.77372360e-01 -3.62784177e-01 3.80505145e-01 1.14831913e+00
1.83213592e+00 1.79690301e-01 1.32569337e+00 6.56448066e-01
2.57883877e-01 5.89563131e-01 5.27274489e-01 5.07797420e-01
6.64414391e-02 5.53721130e-01 1.16967559e+00 -2.28594258e-01
-4.47366118e-01 5.88265099e-02 3.24092031e-01 4.63233739e-01
-1.02356128e-01 -4.92237329e-01 -4.72483665e-01 5.51158011e-01
-1.35595977e+00 -9.87777174e-01 -1.99666675e-02 2.38879228e+00
7.10393667e-01 -1.59137517e-01 -7.18946934e-01 2.19910204e-01
7.19719708e-01 3.70970637e-01 -6.73199832e-01 1.35189950e-01
-5.91197252e-01 -1.03336282e-01 7.19573617e-01 6.87661350e-01
-8.79504919e-01 5.07426560e-01 6.01132727e+00 5.55078387e-01
-1.02936316e+00 7.25769103e-02 3.83647621e-01 7.17132390e-02
-1.66745797e-01 -5.06184027e-02 -4.30084795e-01 6.32482648e-01
8.44830513e-01 -2.78828770e-01 5.88242531e-01 9.69568267e-02
5.01764476e-01 -3.76667023e-01 -5.08129001e-01 1.14093983e+00
4.18129921e-01 -1.12419736e+00 -1.65212452e-01 -6.43851236e-02
8.00063014e-01 -1.82155684e-01 8.44694450e-02 -2.02222392e-01
1.10556141e-01 -5.68301618e-01 8.50042820e-01 1.00572634e+00
9.42596436e-01 -4.36020255e-01 6.90056741e-01 3.31607878e-01
-5.58938205e-01 -3.06556612e-01 -6.40693903e-01 -7.72324130e-02
3.35824907e-01 7.45698631e-01 -3.63492399e-01 4.35522914e-01
9.67738688e-01 7.95300424e-01 -1.92102984e-01 1.28143847e+00
-2.25375235e-01 4.59289998e-01 -6.22409210e-02 5.84417820e-01
-1.04932748e-01 -4.50073332e-01 8.54769111e-01 7.82280982e-01
6.95993841e-01 2.82981783e-01 -3.89163107e-01 7.73073018e-01
1.03356671e-02 -4.30151045e-01 -3.94782394e-01 2.05251321e-01
1.98236704e-01 7.14373171e-01 -4.08826470e-01 -1.76774591e-01
-1.43771023e-01 1.16326380e+00 -2.82571852e-01 7.24819541e-01
-4.61990088e-01 -5.13306797e-01 8.89748693e-01 1.69376820e-01
8.56778398e-02 -1.13393283e-02 1.42938877e-02 -1.13418984e+00
5.03646582e-02 -6.88291311e-01 -4.33815643e-02 -1.31581807e+00
-1.32626009e+00 7.47066975e-01 2.24533621e-02 -1.68647850e+00
-1.13416347e-03 -4.68490332e-01 -6.42904878e-01 9.10483837e-01
-2.05449009e+00 -8.18404436e-01 -6.53841794e-01 6.91515625e-01
3.04582179e-01 -3.54867466e-02 5.98046482e-01 2.93701470e-01
-4.93570149e-01 2.52406090e-01 5.77340603e-01 -3.81689966e-01
8.14486504e-01 -8.93691719e-01 6.72473609e-02 1.27907634e+00
-4.96444762e-01 3.31748307e-01 1.55990791e+00 -4.18247283e-01
-1.31616998e+00 -8.14174712e-01 3.65893632e-01 3.64130527e-01
5.91804504e-01 8.06273967e-02 -1.23127210e+00 2.34508798e-01
2.75995195e-01 2.93534249e-01 1.44185185e-01 -8.62895191e-01
-2.76386708e-01 -4.71962303e-01 -1.48812830e+00 3.52559119e-01
8.09928179e-01 -5.82246363e-01 -3.96622956e-01 3.32979292e-01
6.47859454e-01 -2.99233705e-01 -5.65228879e-01 4.42651302e-01
4.49740410e-01 -1.24452770e+00 1.17526627e+00 -2.23315328e-01
6.82695448e-01 -5.16681015e-01 -5.56517065e-01 -1.75267088e+00
2.21371874e-02 -6.37856185e-01 5.85937090e-02 8.58296812e-01
-5.54112308e-02 -6.60509408e-01 3.15776974e-01 2.92698592e-01
-6.16311908e-01 -2.43374735e-01 -1.14763153e+00 -9.78537679e-01
-1.38211384e-01 -1.05662182e-01 7.00340569e-02 5.69118798e-01
-6.08365893e-01 -1.39140010e-01 -7.94856668e-01 7.19004512e-01
1.21903193e+00 -3.13113153e-01 5.42499661e-01 -1.08398509e+00
-2.08543643e-01 9.01892334e-02 -5.37832320e-01 -8.44505906e-01
7.09936842e-02 -2.56964833e-01 7.21287668e-01 -1.69747949e+00
8.28691944e-02 1.38110910e-02 -2.50924081e-01 -4.59311903e-02
3.08961188e-03 5.92290282e-01 1.63385823e-01 4.68442261e-01
-9.03143138e-02 9.62295592e-01 1.33196354e+00 -3.57185453e-01
-1.42744973e-01 -1.25404030e-01 -4.36875314e-01 4.58631486e-01
4.69209015e-01 -3.42785716e-01 -4.15371954e-01 -5.39484620e-01
1.98542103e-01 2.72580653e-01 7.83905506e-01 -1.25703931e+00
4.36641723e-01 -1.18346009e-02 -1.93040464e-02 -1.29443213e-01
8.07244599e-01 -6.81648493e-01 1.38216957e-01 4.01428789e-01
-1.21026732e-01 -6.04989588e-01 -1.77292489e-02 8.03211331e-01
-4.06578481e-01 -5.01601279e-01 1.64431262e+00 -2.15336949e-01
-7.09587097e-01 2.07980294e-02 -4.01324362e-01 -2.48559669e-01
7.62222350e-01 -4.11395580e-01 -6.42906249e-01 -5.58691144e-01
-7.16244340e-01 -1.33537874e-01 7.27201998e-01 -1.66084185e-01
9.06966388e-01 -7.94327497e-01 -7.41846144e-01 4.22229379e-01
5.27947955e-02 -2.41562456e-01 8.70196760e-01 8.46980870e-01
-7.28172064e-01 -3.47693980e-01 -4.22835499e-01 -5.02943933e-01
-1.15755820e+00 5.30978262e-01 8.74412477e-01 1.54486433e-01
-9.57312107e-01 8.55873168e-01 2.34225586e-01 -1.55889302e-01
2.06041783e-01 7.01969266e-02 -7.56866336e-02 -2.20136166e-01
8.32110405e-01 4.76783246e-01 2.20325977e-01 -7.79933512e-01
-1.40144760e-02 5.72735071e-01 2.67346382e-01 -7.47864619e-02
1.55843174e+00 -8.36444497e-01 -1.96452275e-01 3.07153642e-01
1.15540135e+00 -5.59171259e-01 -1.82300150e+00 -2.18173161e-01
-6.89756155e-01 -7.87430882e-01 5.20761430e-01 -7.58486509e-01
-1.11525512e+00 9.50917542e-01 1.18085599e+00 -2.94572860e-02
1.67611825e+00 -3.03446382e-01 6.34320796e-01 2.95403004e-01
3.88196975e-01 -9.45879638e-01 9.69609618e-02 1.74029797e-01
1.10314083e+00 -1.24662149e+00 -7.47102723e-02 -2.40025043e-01
-3.69024336e-01 9.38169718e-01 2.44799897e-01 -1.78144738e-01
7.82112181e-01 2.04551622e-01 4.21099305e-01 -9.48506035e-03
-5.33070982e-01 -1.67689413e-01 -1.40438214e-01 8.92950535e-01
-3.57101440e-01 -2.61409611e-01 -1.21098220e-01 1.40241101e-01
1.09289177e-01 -4.92856745e-03 1.13338292e+00 7.41762578e-01
-8.00193369e-01 -6.26469493e-01 -6.02383077e-01 2.02517942e-01
-2.78244764e-01 -8.79931524e-02 3.15277725e-01 3.57620627e-01
6.73529953e-02 1.20072019e+00 -1.42711252e-01 -1.56976685e-01
3.80256295e-01 -4.23254490e-01 5.29105425e-01 -2.27047130e-01
2.26798318e-02 1.81612857e-02 -3.40550244e-01 -3.61648947e-01
-8.56494665e-01 -4.61485922e-01 -7.97225475e-01 -1.20933361e-01
-1.46401063e-01 -4.47362252e-02 7.50034273e-01 1.00444078e+00
-1.18040983e-02 4.73269850e-01 7.44520962e-01 -1.32431281e+00
-6.98358417e-01 -1.10742736e+00 -1.13972211e+00 4.02281761e-01
9.41481709e-01 -8.46864343e-01 -1.03461945e+00 3.02884191e-01] | [11.272497177124023, -2.465681314468384] |
9882ee96-c81b-4661-a5dc-d8ac84b0a3e9 | backpropagating-through-structured-argmax | 1805.04658 | null | http://arxiv.org/abs/1805.04658v1 | http://arxiv.org/pdf/1805.04658v1.pdf | Backpropagating through Structured Argmax using a SPIGOT | We introduce the structured projection of intermediate gradients optimization
technique (SPIGOT), a new method for backpropagating through neural networks
that include hard-decision structured predictions (e.g., parsing) in
intermediate layers. SPIGOT requires no marginal inference, unlike structured
attention networks (Kim et al., 2017) and some reinforcement learning-inspired
solutions (Yogatama et al., 2017). Like so-called straight-through estimators
(Hinton, 2012), SPIGOT defines gradient-like quantities associated with
intermediate nondifferentiable operations, allowing backpropagation before and
after them; SPIGOT's proxy aims to ensure that, after a parameter update, the
intermediate structure will remain well-formed.
We experiment on two structured NLP pipelines: syntactic-then-semantic
dependency parsing, and semantic parsing followed by sentiment classification.
We show that training with SPIGOT leads to a larger improvement on the
downstream task than a modularly-trained pipeline, the straight-through
estimator, and structured attention, reaching a new state of the art on
semantic dependency parsing. | ['Sam Thomson', 'Noah A. Smith', 'Hao Peng'] | 2018-05-12 | backpropagating-through-structured-argmax-1 | https://aclanthology.org/P18-1173 | https://aclanthology.org/P18-1173.pdf | acl-2018-7 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [ 2.22741768e-01 9.19486225e-01 -2.63985664e-01 -8.54210258e-01
-1.05846810e+00 -5.22205770e-01 4.42591906e-01 2.20018327e-01
-5.11637032e-01 6.60775423e-01 4.41383183e-01 -6.18660688e-01
2.27187291e-01 -6.52525425e-01 -1.15090168e+00 -3.83428156e-01
-2.04256475e-01 5.67243814e-01 8.02820548e-02 -2.88558528e-02
2.33216390e-01 1.03689425e-01 -9.54453409e-01 5.03520310e-01
6.71342969e-01 1.06733084e+00 4.00915265e-01 1.02136040e+00
-4.12172407e-01 1.17147839e+00 -3.02742660e-01 -8.57581854e-01
4.89604659e-02 -5.36555707e-01 -1.15656471e+00 -2.79818416e-01
3.35563362e-01 -2.35429540e-01 2.46008143e-01 1.11967623e+00
-9.42740776e-03 -1.11681595e-01 4.68681723e-01 -9.74611402e-01
-9.40581322e-01 1.15239084e+00 -1.66486681e-01 -9.04156864e-02
1.41742691e-01 1.70593590e-01 1.67842066e+00 -8.93539310e-01
5.44933617e-01 1.70483339e+00 9.48188901e-01 6.53044701e-01
-1.20486164e+00 -2.64712036e-01 7.81354904e-01 -1.17062941e-01
-3.30831081e-01 -2.04190880e-01 4.83412445e-01 -2.93833584e-01
1.40954804e+00 -5.41342199e-02 4.85812068e-01 9.37192082e-01
2.56289482e-01 1.18026710e+00 8.57892632e-01 -4.57724541e-01
3.71763349e-01 1.77763514e-02 5.12551248e-01 1.06045222e+00
1.06695198e-01 -1.20581411e-01 -4.91050214e-01 1.83278043e-02
4.26762849e-01 -2.70481944e-01 1.33444801e-01 -6.96156323e-02
-9.85374868e-01 1.03914213e+00 6.43328309e-01 -1.94648150e-02
-3.90928745e-01 4.42672104e-01 4.86901313e-01 2.91296810e-01
5.79331040e-01 5.90868473e-01 -1.19513345e+00 -1.35603026e-01
-8.76117110e-01 -1.93418916e-02 1.21333838e+00 7.58228660e-01
9.39380825e-01 -1.19160630e-01 -2.80447662e-01 5.76026022e-01
4.73518550e-01 1.16034642e-01 5.34397304e-01 -1.31536400e+00
7.22110629e-01 3.54431540e-01 4.79508974e-02 -2.21642435e-01
-5.08332312e-01 -2.47536495e-01 -4.31580752e-01 2.31460139e-01
7.38078117e-01 -6.78277791e-01 -1.07066000e+00 1.93979514e+00
7.56585822e-02 -2.01334149e-01 3.06067705e-01 8.68389547e-01
4.86811072e-01 8.63177955e-01 5.19237518e-01 3.91571075e-02
1.28385043e+00 -1.53189540e+00 -4.26158816e-01 -9.23426330e-01
7.74957418e-01 -3.89804095e-01 1.24347472e+00 3.19199771e-01
-1.26530695e+00 -3.86637270e-01 -9.12710667e-01 -4.54949081e-01
-2.93327779e-01 -6.88581988e-02 1.11037362e+00 4.83496636e-01
-1.34860110e+00 1.12745678e+00 -1.20264804e+00 -1.46938100e-01
6.42801881e-01 2.84190983e-01 -3.62280488e-01 1.84664041e-01
-1.14244509e+00 1.11361289e+00 3.17559630e-01 2.49193296e-01
-8.16212177e-01 -6.86611831e-01 -1.13595605e+00 3.47724438e-01
1.85066998e-01 -7.42736518e-01 1.71519184e+00 -1.41921198e+00
-2.05172873e+00 9.05116856e-01 -4.80167031e-01 -7.10241199e-01
2.11473882e-01 -6.26739383e-01 4.08012360e-01 -1.06543556e-01
2.41186082e-01 1.01341879e+00 7.24676847e-01 -8.01084459e-01
-5.21357834e-01 -3.46053392e-01 2.89867848e-01 3.12860876e-01
-8.12619999e-02 2.15007052e-01 -2.52755165e-01 -1.93265557e-01
2.81547397e-01 -8.43014836e-01 -4.69206333e-01 -7.04624280e-02
-7.13606596e-01 -4.00011212e-01 1.90346256e-01 -1.07561362e+00
7.38853395e-01 -1.92802715e+00 3.73009801e-01 -3.26767653e-01
-4.51151207e-02 -8.12215954e-02 -2.43446141e-01 1.49390697e-01
-9.31149796e-02 3.23889583e-01 -8.13868403e-01 -8.56511891e-01
3.48268926e-01 4.15258616e-01 -1.53334230e-01 2.07197949e-01
8.08515191e-01 1.19542348e+00 -1.06305659e+00 -2.18935430e-01
-2.00831905e-01 1.84963241e-01 -8.45268607e-01 3.81063610e-01
-6.64839804e-01 3.23330879e-01 -2.90146351e-01 5.85517049e-01
4.37158555e-01 -3.05376917e-01 2.26857990e-01 1.24904901e-01
-1.35079369e-01 9.99357641e-01 -6.71936691e-01 1.64585912e+00
-8.47987771e-01 7.56476820e-01 3.47190768e-01 -1.17439985e+00
5.70723295e-01 8.55410099e-02 -1.07209794e-01 -4.90808368e-01
1.11361742e-01 3.04266959e-01 -1.29039973e-01 -1.78346276e-01
3.77836674e-01 -1.41255200e-01 -3.48384559e-01 3.57475936e-01
4.94840473e-01 -2.43694097e-01 2.12527081e-01 2.05058739e-01
1.08803916e+00 6.67371511e-01 1.72958508e-01 -2.98111498e-01
3.05716515e-01 2.12200046e-01 7.28082359e-01 9.16005075e-01
-1.32793844e-01 4.99693662e-01 1.16303635e+00 -3.28670949e-01
-8.93396497e-01 -1.01350713e+00 9.71009210e-02 1.58011818e+00
-3.06409389e-01 -2.45260462e-01 -1.20355046e+00 -1.09569192e+00
1.03847648e-03 8.28353941e-01 -8.25347781e-01 1.05042763e-01
-8.15347493e-01 -8.38705003e-01 3.31720114e-01 7.84467220e-01
4.67448145e-01 -1.41032565e+00 -5.51013708e-01 3.38712215e-01
2.01339915e-01 -9.68041122e-01 -3.65874290e-01 1.00066507e+00
-1.27898061e+00 -8.44429910e-01 -5.86364388e-01 -1.04201031e+00
7.41044819e-01 -5.12089372e-01 1.54448152e+00 -1.99889705e-01
2.04928920e-01 -2.19169147e-02 -9.64266062e-02 -1.25276595e-01
-5.58813393e-01 2.70296633e-01 -6.51198268e-01 -3.07908684e-01
1.24718241e-01 -3.96633595e-01 -3.78317952e-01 -2.25862801e-01
-3.91571939e-01 1.71945035e-01 8.99269819e-01 1.19691133e+00
5.57212293e-01 -6.09398425e-01 2.41314545e-01 -1.39707315e+00
6.43090904e-01 -5.00031292e-01 -7.45453238e-01 1.40484706e-01
-6.46196008e-01 8.05756032e-01 9.32765782e-01 -2.24209100e-01
-1.20883548e+00 7.78528377e-02 -5.10240018e-01 5.58682978e-02
-1.21911265e-01 4.91142452e-01 -1.81216180e-01 2.82336980e-01
3.91683161e-01 -7.23642632e-02 -3.06230098e-01 -5.25415778e-01
7.12355077e-01 2.65569568e-01 4.29227740e-01 -6.33269787e-01
4.33542967e-01 2.42540419e-01 -2.66454428e-01 -1.77523717e-01
-1.56131923e+00 -3.61258201e-02 -6.05800748e-01 3.32711875e-01
1.37109947e+00 -7.85238922e-01 -4.96463656e-01 4.78624701e-01
-1.45940340e+00 -1.06432343e+00 -3.70431155e-01 2.99464911e-01
-4.93654191e-01 1.32114112e-01 -1.26496935e+00 -7.09370017e-01
-4.79912311e-01 -1.11977243e+00 1.20336175e+00 2.39695162e-01
-2.73942977e-01 -1.22784698e+00 1.11739943e-02 1.14703514e-01
2.92073280e-01 -1.64687037e-01 1.01426363e+00 -7.46833265e-01
-4.91323203e-01 1.11062840e-01 -2.69306272e-01 7.16469884e-01
-3.79936457e-01 -1.89186767e-01 -1.06003749e+00 1.12476185e-01
1.81096271e-01 -5.28825164e-01 1.18582356e+00 5.37119865e-01
1.03133094e+00 -4.62937504e-01 1.95551515e-02 9.34639096e-01
1.22048378e+00 -6.17341436e-02 3.22470844e-01 5.52009344e-01
6.09632015e-01 7.63996184e-01 4.27697897e-01 -1.23261362e-01
6.46783113e-01 1.92699097e-02 5.02811968e-01 -1.29850386e-02
-2.08053254e-02 -5.41597843e-01 7.00682759e-01 7.63325512e-01
2.91629732e-01 -3.56840342e-02 -6.45477295e-01 4.42332357e-01
-1.87651038e+00 -5.94219744e-01 -8.62132534e-02 1.80560660e+00
8.67414474e-01 6.63390696e-01 -3.58503938e-01 -2.52768159e-01
6.86175287e-01 2.58987695e-01 -7.51158595e-01 -1.15261579e+00
-5.06912805e-02 6.46776438e-01 7.78231502e-01 9.07677531e-01
-1.06562972e+00 1.45675182e+00 6.22845268e+00 1.68921083e-01
-9.50388134e-01 4.14052099e-01 9.80540693e-01 3.61021936e-01
-3.76906872e-01 4.61117685e-01 -9.69639301e-01 3.30143601e-01
1.29568684e+00 4.71414328e-01 4.90262568e-01 1.21321607e+00
-2.66413212e-01 -2.11747110e-01 -1.18776333e+00 2.72896022e-01
-2.51677066e-01 -1.19198167e+00 -3.93174738e-01 -5.00704467e-01
6.13073945e-01 3.23174924e-01 -2.38520019e-02 7.18928874e-01
9.83071566e-01 -9.22499418e-01 1.01189840e+00 6.85681701e-02
2.35491484e-01 -4.05483425e-01 6.28000855e-01 2.91911632e-01
-7.72603393e-01 -2.46434599e-01 -3.97712469e-01 -3.80404353e-01
3.26079190e-01 7.87396908e-01 -6.65769577e-01 -1.41743096e-02
7.51871705e-01 6.75264597e-01 -3.45909506e-01 3.39661628e-01
-1.25345600e+00 1.05907047e+00 -2.98116416e-01 -4.24751252e-01
6.19396985e-01 -2.35072657e-01 2.92773426e-01 1.39541566e+00
-1.91385791e-01 -3.36224794e-01 -1.24232560e-01 9.64382172e-01
-3.11418682e-01 -1.06779017e-01 -2.84715015e-02 -8.20216909e-02
2.40870655e-01 1.24023676e+00 -8.52970541e-01 -5.31372547e-01
-5.24380922e-01 1.36645949e+00 8.53891373e-01 4.16165531e-01
-7.08856821e-01 -3.83266151e-01 6.30696476e-01 -3.53126377e-01
7.73509920e-01 -1.80216268e-01 -7.84184337e-01 -1.13752460e+00
1.00900501e-01 -4.54094619e-01 3.74224544e-01 -9.03484941e-01
-1.25711262e+00 5.97957313e-01 -3.01975399e-01 -2.68743306e-01
-4.13000822e-01 -1.08753109e+00 -9.25069034e-01 1.05154967e+00
-1.78550982e+00 -9.70171750e-01 1.86423659e-01 1.57004043e-01
7.07038224e-01 3.71486962e-01 9.66081858e-01 -2.38364622e-01
-7.31386065e-01 4.68178242e-01 -1.06544875e-01 3.70416135e-01
4.18271869e-01 -1.75040674e+00 1.04109800e+00 9.13283944e-01
-1.26207396e-02 5.74999869e-01 4.71535712e-01 -5.58449686e-01
-1.47569108e+00 -1.00046480e+00 1.06692147e+00 -3.81728351e-01
9.41032648e-01 -6.39585912e-01 -7.17427373e-01 1.19438350e+00
4.45253760e-01 4.82802317e-02 1.82925165e-01 4.75015819e-01
-5.10441840e-01 1.51815549e-01 -9.59837198e-01 4.27484423e-01
9.19114649e-01 -4.88417953e-01 -8.62359405e-01 4.05814886e-01
1.07082558e+00 -4.95937049e-01 -5.92637122e-01 1.24126248e-01
3.85217994e-01 -1.13013542e+00 6.52709544e-01 -8.44555676e-01
8.49534631e-01 3.16612154e-01 -4.48969640e-02 -1.52694082e+00
-2.40272641e-01 -7.92274117e-01 -1.55683130e-01 9.37437713e-01
1.13248503e+00 -9.02685344e-01 9.69585717e-01 6.60382330e-01
-6.54212177e-01 -9.83370960e-01 -6.58980966e-01 -4.33893740e-01
5.98167002e-01 -3.94117832e-01 2.60361493e-01 6.86022043e-01
-3.17362919e-02 5.31889141e-01 4.93677892e-02 3.27961326e-01
5.61635733e-01 5.14406823e-02 3.23840916e-01 -8.64945710e-01
-7.60631084e-01 -5.78905642e-01 -2.86506712e-02 -1.32499409e+00
5.08423924e-01 -1.02455056e+00 5.88771880e-01 -1.87998784e+00
6.74518868e-02 -2.90703684e-01 -2.53704756e-01 7.09392250e-01
-5.18813550e-01 -1.46254301e-01 1.14252113e-01 -1.07837968e-01
-6.86893344e-01 3.07192534e-01 1.11985934e+00 3.61337066e-02
-1.46939382e-01 -1.62429094e-01 -1.02687478e+00 9.53055203e-01
9.03367817e-01 -6.80722237e-01 -6.84650801e-03 -7.17336476e-01
4.50554788e-01 1.11554369e-01 1.54826790e-01 -6.35476351e-01
2.26585820e-01 6.04647659e-02 2.97692180e-01 -2.13538468e-01
2.38575414e-01 -2.10075870e-01 -5.76770306e-01 5.16573966e-01
-6.47039294e-01 1.02561280e-01 1.13272674e-01 2.08620280e-01
-2.84749389e-01 -6.00466073e-01 6.78217888e-01 -5.53869307e-01
-5.62971830e-01 5.76756801e-03 -3.47656786e-01 4.70155954e-01
3.74698818e-01 9.22004282e-02 -1.73482418e-01 -1.71526164e-01
-8.63152444e-01 4.12790149e-01 1.63032725e-01 -3.61025077e-03
2.56176203e-01 -7.19193935e-01 -5.93529940e-01 -7.78436149e-03
-5.09440482e-01 4.75426614e-01 -2.04696238e-01 7.45120704e-01
-5.23986638e-01 3.77430618e-01 1.36866778e-01 -4.39793438e-01
-6.61142826e-01 2.00836957e-01 3.12019438e-01 -7.47307122e-01
-3.81946415e-01 1.49398279e+00 2.25916401e-01 -9.22593892e-01
1.73616663e-01 -8.11920345e-01 1.68022886e-01 -6.03901558e-02
1.74468175e-01 -1.11613020e-01 1.25866011e-01 3.87650989e-02
-2.04956472e-01 2.50020593e-01 -5.37670702e-02 -3.23625356e-01
1.53712487e+00 9.28908512e-02 -3.56321186e-01 5.33715606e-01
1.30482304e+00 -6.67856634e-02 -1.78558147e+00 6.90034181e-02
4.51251328e-01 1.92511499e-01 -3.89986928e-03 -1.00991118e+00
-8.80771577e-01 1.21064162e+00 2.45940369e-02 2.77582496e-01
8.55177402e-01 2.83116907e-01 8.74219835e-01 6.43378019e-01
1.06621422e-01 -1.04728830e+00 -7.58099928e-02 1.00312781e+00
5.70181191e-01 -1.38404465e+00 -3.54189545e-01 -2.59426147e-01
-6.95671439e-01 1.15704823e+00 5.66794455e-01 -3.94857019e-01
6.13431573e-01 5.03661156e-01 -2.35262066e-02 5.99903949e-02
-9.98289645e-01 -2.13780969e-01 -1.06216721e-01 3.31272572e-01
6.13361835e-01 7.25511611e-02 -6.40044436e-02 8.41565728e-01
-3.67978454e-01 -3.56408834e-01 1.96953818e-01 1.03216612e+00
-5.59853494e-01 -1.12904513e+00 -2.64133602e-01 4.55326855e-01
-7.42320657e-01 -6.56636119e-01 -2.35174134e-01 7.34722078e-01
-1.69494838e-01 6.73758864e-01 1.66907296e-01 3.40112075e-02
1.69735506e-01 4.82864588e-01 4.74851042e-01 -9.71934080e-01
-8.23468029e-01 -3.09485853e-01 4.26763266e-01 -8.52089763e-01
-1.71330735e-01 -6.24325335e-01 -1.52298748e+00 2.27758840e-01
-1.54788598e-01 3.22272509e-01 1.08254194e+00 1.11553776e+00
1.53075010e-01 5.90381682e-01 4.85136926e-01 -1.08460319e+00
-7.05039144e-01 -1.09143424e+00 -4.61497791e-02 2.84314185e-01
4.09260362e-01 -2.11873755e-01 -6.17403984e-01 -4.66923825e-02] | [10.477815628051758, 9.320786476135254] |
0107907d-b0fd-4db6-9fea-5fa266828175 | deep-factorized-metric-learning | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Deep_Factorized_Metric_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Deep_Factorized_Metric_Learning_CVPR_2023_paper.pdf | Deep Factorized Metric Learning | Learning a generalizable and comprehensive similarity metric to depict the semantic discrepancies between images is the foundation of many computer vision tasks. While existing methods approach this goal by learning an ensemble of embeddings with diverse objectives, the backbone network still receives a mix of all the training signals. Differently, we propose a deep factorized metric learning method (DFML) to factorize the training signal and employ different samples to train various components of the backbone network. We factorize the network to different sub-blocks and devise a learnable router to adaptively allocate the training samples to each sub-block with the objective to capture the most information. The metric model trained by DFML captures different characteristics with different sub-blocks and constitutes a generalizable metric when using all the sub-blocks. The proposed DFML achieves state-of-the-art performance on all three benchmarks for deep metric learning including CUB-200-2011, Cars196, and Stanford Online Products. We also generalize DFML to the image classification task on ImageNet-1K and observe consistent improvement in accuracy/computation trade-off. Specifically, we improve the performance of ViT-B on ImageNet (+0.2% accuracy) with less computation load (-24% FLOPs). | ['Jiwen Lu', 'Jie zhou', 'Junlong Li', 'Wenzhao Zheng', 'Chengkun Wang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [-9.82734263e-02 -1.34042069e-01 -1.77830905e-01 -8.34087789e-01
-8.96944225e-01 -3.75314057e-01 3.68289351e-01 -8.46415311e-02
-6.27000749e-01 1.73787013e-01 6.72823712e-02 -3.21325883e-02
-2.11529836e-01 -6.68477476e-01 -7.02813447e-01 -6.01470351e-01
-6.03627227e-02 3.59424770e-01 2.12771043e-01 -1.71743810e-01
7.58050606e-02 4.73507792e-01 -1.27830601e+00 1.80171579e-01
6.64618194e-01 1.50469244e+00 2.65720546e-01 6.20593250e-01
6.69211969e-02 8.06143165e-01 -5.65396607e-01 -5.46048939e-01
6.74470782e-01 -2.56645560e-01 -7.06693351e-01 5.29478528e-02
8.75332534e-01 -6.55025661e-01 -7.99805641e-01 1.16315591e+00
4.15758222e-01 1.60263076e-01 5.58319330e-01 -1.54548156e+00
-7.50697672e-01 6.16645217e-01 -4.95091319e-01 2.28188932e-01
-2.56045967e-01 1.72851652e-01 1.21080840e+00 -8.77074718e-01
3.05049866e-01 1.30195308e+00 4.38188463e-01 5.02269685e-01
-1.07252169e+00 -6.58788919e-01 1.07106969e-01 6.85468376e-01
-1.41740525e+00 -3.61234874e-01 6.29730940e-01 -3.78514618e-01
7.42041647e-01 -3.63850966e-02 1.82068244e-01 8.59876454e-01
1.52779445e-01 7.36331403e-01 8.42622280e-01 3.73187177e-02
1.90442488e-01 -9.96100903e-03 3.64009857e-01 8.43328714e-01
1.98523566e-01 -4.31114212e-02 -5.24727702e-01 1.07319959e-01
5.23140013e-01 2.61064738e-01 -2.44859636e-01 -3.06079805e-01
-1.27954721e+00 8.43704998e-01 6.82853699e-01 -4.79432456e-02
-1.29697010e-01 5.67758679e-01 4.72802490e-01 6.01544082e-01
2.88249820e-01 3.60341579e-01 -5.78942001e-01 -1.17901593e-01
-8.94311845e-01 1.57764450e-01 7.16201067e-01 9.72279251e-01
1.21648276e+00 -2.20013723e-01 -3.90020162e-01 9.72031832e-01
2.00237975e-01 5.03619432e-01 4.60754871e-01 -1.28538036e+00
5.87615550e-01 4.67873216e-01 -7.89412111e-02 -1.12143648e+00
-2.66249090e-01 -5.70235491e-01 -6.89823449e-01 2.41539001e-01
4.24132943e-01 -4.13385825e-03 -9.50640976e-01 1.95374334e+00
2.28912875e-01 3.06637675e-01 -2.63445377e-01 1.01195920e+00
5.35833657e-01 5.24665892e-01 -2.51621842e-01 4.32968557e-01
1.20290065e+00 -1.41780996e+00 -1.78655759e-01 -2.84196913e-01
7.02441156e-01 -5.34239650e-01 1.14288867e+00 3.08174193e-01
-7.78369904e-01 -8.03891480e-01 -1.61401784e+00 -3.96316707e-01
-2.71659374e-01 3.32478553e-01 3.98559362e-01 3.94385993e-01
-1.11965966e+00 8.50737810e-01 -8.21425438e-01 -1.11170344e-01
6.04540765e-01 2.88558245e-01 -4.49851096e-01 -5.89226604e-01
-9.23701763e-01 8.59445214e-01 7.81579912e-02 -5.11543229e-02
-1.18322504e+00 -1.04737544e+00 -8.05650234e-01 8.93234387e-02
1.05634727e-01 -5.79602659e-01 1.03276968e+00 -8.79726231e-01
-1.28988338e+00 7.49807119e-01 4.28344272e-02 -5.06210625e-01
4.69043434e-01 -1.01517081e-01 -2.89241195e-01 1.52150556e-01
7.45259225e-02 9.77808416e-01 8.29499602e-01 -8.50726962e-01
-7.78590262e-01 -2.67697215e-01 3.36619645e-01 -7.40312114e-02
-7.21233130e-01 -1.58287898e-01 -6.44710720e-01 -5.44173896e-01
-7.90699720e-02 -9.87487435e-01 -1.36955827e-01 6.81798518e-01
-6.90522417e-02 -2.46415943e-01 8.32150936e-01 -5.74517846e-01
1.14478409e+00 -2.35051489e+00 2.17372596e-01 3.50685194e-02
6.92093730e-01 2.24947527e-01 -6.86454535e-01 -4.84677106e-02
1.51467830e-01 -1.50345787e-01 -2.23732769e-01 -6.55928731e-01
3.14325750e-01 4.00764346e-01 1.53356167e-02 7.47577846e-01
3.19621027e-01 9.07401443e-01 -9.33652759e-01 -2.46060833e-01
3.08673203e-01 4.94736761e-01 -7.27057278e-01 2.14736030e-01
-2.59893052e-02 -1.30336314e-01 -2.85275668e-01 4.48381662e-01
9.19223547e-01 -4.98456210e-01 -6.28302917e-02 -5.62897384e-01
3.11712354e-01 2.90759027e-01 -1.13249743e+00 2.17629838e+00
-5.16633093e-01 9.07631159e-01 -8.18522796e-02 -1.16244340e+00
9.70584214e-01 -2.17525572e-01 5.94157755e-01 -7.81307280e-01
1.13943689e-01 9.37892646e-02 6.19751140e-02 -2.86534399e-01
3.29975843e-01 4.08615887e-01 -1.02750763e-01 4.50755477e-01
4.66237932e-01 2.23469779e-01 2.51072407e-01 1.94309950e-01
1.46910143e+00 -2.13056803e-01 -2.62225829e-02 -2.90711492e-01
5.52695870e-01 -4.01492625e-01 5.62023103e-01 6.23717844e-01
-5.52351654e-01 6.54010832e-01 4.53165323e-01 -6.76763356e-01
-1.04204261e+00 -1.13608253e+00 1.20402896e-03 1.25689673e+00
3.50420564e-01 -5.44249535e-01 -7.29135573e-01 -9.75105822e-01
2.97003597e-01 3.69865447e-01 -8.49185407e-01 -4.97155130e-01
-4.65533197e-01 -6.50285482e-01 5.65631449e-01 5.40022194e-01
8.22049320e-01 -6.46197498e-01 -4.50706780e-01 9.68448296e-02
-1.83645606e-01 -1.33005834e+00 -8.72859299e-01 2.44585007e-01
-4.35347974e-01 -1.18387592e+00 -3.51266384e-01 -5.95439851e-01
6.14534438e-01 5.58360815e-01 1.16559923e+00 8.59910506e-04
-2.22919852e-01 1.79536223e-01 -3.22763115e-01 -1.44972578e-01
-1.98215649e-01 1.32827073e-01 -1.56104743e-01 3.58390689e-01
5.80473781e-01 -4.50700164e-01 -8.86824846e-01 6.70687318e-01
-8.48225653e-01 -2.51809597e-01 4.56334621e-01 7.96715796e-01
5.51592410e-01 -2.10140154e-01 3.38279247e-01 -4.26626861e-01
2.27857038e-01 -6.57756627e-01 -5.03378868e-01 2.04219207e-01
-7.02498317e-01 3.84055257e-01 6.21123195e-01 -5.82570732e-01
-3.93540263e-01 -1.52311504e-01 5.28344586e-02 -7.95277536e-01
1.60106152e-01 1.30232438e-01 -3.11780065e-01 -2.12496087e-01
7.11762846e-01 -3.18033174e-02 1.18165448e-01 -4.55633372e-01
5.23823321e-01 6.48449957e-01 5.60046017e-01 -6.87639654e-01
8.40887904e-01 4.38437700e-01 -3.91702503e-02 -3.31419826e-01
-1.02891493e+00 -3.91994655e-01 -3.25727642e-01 -1.47713006e-01
8.19482088e-01 -9.51100528e-01 -5.61378717e-01 6.19059384e-01
-1.03290105e+00 -4.11621183e-01 -3.67270529e-01 4.75813955e-01
-4.28154707e-01 1.42456383e-01 -5.06802320e-01 1.13254324e-01
-2.80879140e-01 -1.53965855e+00 1.09770536e+00 5.87045550e-02
-2.61829197e-02 -7.90097117e-01 -2.65890136e-02 2.93430984e-01
6.49929643e-01 2.48355359e-01 7.55371571e-01 -5.77227294e-01
-4.48222071e-01 3.28979753e-02 -6.28809333e-01 1.01390040e+00
2.67882198e-01 -2.02676982e-01 -1.05427361e+00 -4.53958303e-01
-5.71194515e-02 -5.61463714e-01 1.13834608e+00 1.62214532e-01
1.52607584e+00 -1.22490086e-01 -3.27355973e-02 1.14668167e+00
1.39387524e+00 -1.19538486e-01 5.33007503e-01 3.04408759e-01
8.19544494e-01 3.23566258e-01 5.03383994e-01 4.48886663e-01
7.28402197e-01 6.78946018e-01 6.71285987e-01 -6.24907203e-02
-4.12500352e-01 -6.99706227e-02 5.10691226e-01 7.64444590e-01
4.45135444e-01 -1.47408560e-01 -5.74994624e-01 4.92826670e-01
-1.84970963e+00 -5.89816630e-01 3.45235944e-01 1.97049558e+00
7.38975704e-01 1.88176021e-01 -2.73004454e-02 2.12387279e-01
5.55362642e-01 4.94668394e-01 -7.95611560e-01 -2.05787733e-01
-4.08642069e-02 2.62983114e-01 8.11836541e-01 4.38559920e-01
-1.11763620e+00 7.18472064e-01 5.61600161e+00 8.70151281e-01
-1.34875500e+00 2.34812841e-01 8.35079491e-01 -3.86332542e-01
-1.91874146e-01 -1.27783969e-01 -6.81046247e-01 6.37851417e-01
1.07398593e+00 -6.39317334e-02 6.42868161e-01 9.63753104e-01
-4.98549230e-02 2.94816703e-01 -1.56948292e+00 1.30941415e+00
2.02124879e-01 -1.25435233e+00 1.29381210e-01 6.55958876e-02
6.40601099e-01 5.87304473e-01 2.16925994e-01 2.80836642e-01
3.34735662e-01 -1.03363431e+00 9.73837554e-01 3.57704073e-01
9.78833556e-01 -6.47585094e-01 6.47865415e-01 -1.82663985e-02
-1.16984606e+00 -2.45811641e-01 -7.22172081e-01 4.42186855e-02
-1.56389505e-01 7.84265816e-01 -4.88817424e-01 3.45863223e-01
9.19501781e-01 1.05050147e+00 -7.65196025e-01 8.46087992e-01
1.50197502e-02 5.57463408e-01 -3.33136916e-01 1.78212628e-01
4.14579958e-01 -1.94117889e-01 1.75242230e-01 1.07436919e+00
3.02192330e-01 -3.20170909e-01 1.35268345e-01 7.33264446e-01
-5.91319799e-01 -1.82313725e-01 -1.43675119e-01 1.85311869e-01
5.87940216e-01 1.40708065e+00 -3.15775096e-01 -2.84811467e-01
-7.12713301e-01 1.00494421e+00 6.57247007e-01 2.25183666e-01
-1.06421018e+00 -6.12624526e-01 1.39305413e+00 -2.47967299e-02
5.03328323e-01 -3.26891959e-01 -9.62932259e-02 -9.31985021e-01
2.14491382e-01 -9.12528574e-01 3.17150980e-01 -4.28110093e-01
-1.37958658e+00 5.58399439e-01 -1.26921132e-01 -1.17183435e+00
-2.22571380e-02 -8.51622462e-01 -3.60765845e-01 7.80451477e-01
-1.75699198e+00 -9.49248552e-01 -7.82692671e-01 7.30603337e-01
5.20964146e-01 -2.31655240e-01 5.45722842e-01 5.99374175e-01
-5.27736902e-01 1.09875727e+00 9.76271182e-02 3.76047760e-01
1.00246990e+00 -1.29951572e+00 6.18669569e-01 7.73872852e-01
2.82378227e-01 3.07338864e-01 3.40713710e-01 7.31342882e-02
-1.32912350e+00 -1.47943664e+00 6.83323145e-01 -3.48479927e-01
7.08802581e-01 -4.52674806e-01 -6.91812158e-01 4.53307569e-01
3.30077782e-02 7.23458290e-01 4.54848111e-01 -2.58905739e-01
-1.07800221e+00 -7.92891741e-01 -1.25076580e+00 2.79175013e-01
1.17265224e+00 -6.91226184e-01 -2.70261019e-01 1.67997882e-01
9.51419294e-01 -2.50014931e-01 -9.58279431e-01 2.70331591e-01
5.54214478e-01 -9.71447051e-01 8.48686397e-01 -3.44253629e-01
4.28355813e-01 -4.50333536e-01 -6.46001041e-01 -1.40877080e+00
-4.62343723e-01 -5.52523553e-01 -1.81981102e-01 9.57780898e-01
2.45606124e-01 -6.52062654e-01 7.13992238e-01 3.70701998e-01
-3.00273150e-01 -9.63295341e-01 -8.90910864e-01 -6.14345610e-01
1.49722233e-01 -5.31215370e-01 8.08957040e-01 7.29470015e-01
-4.85426575e-01 1.42795667e-01 -2.66610771e-01 7.89129585e-02
7.94772685e-01 -2.26182699e-01 7.37431347e-01 -9.53135908e-01
-4.39489275e-01 -4.67503309e-01 -8.95104408e-01 -1.42373323e+00
2.57664323e-01 -1.13863957e+00 -1.03643432e-01 -1.26927507e+00
1.73730209e-01 -5.10711551e-01 -8.64390969e-01 6.12013936e-01
-1.73027635e-01 3.60663444e-01 3.89632106e-01 2.48995498e-02
-7.62241006e-01 4.62629169e-01 1.09557772e+00 -6.27011120e-01
2.58872122e-01 -3.46835166e-01 -9.20375526e-01 3.28333974e-01
6.16279721e-01 -5.43252468e-01 -7.45244563e-01 -9.36099231e-01
9.84751340e-03 -4.59969461e-01 3.87990028e-01 -1.14025557e+00
4.21882212e-01 -6.96287155e-02 3.22856873e-01 -2.93035448e-01
8.18272382e-02 -7.84843326e-01 -6.02880605e-02 2.86856562e-01
-4.58894551e-01 1.45016357e-01 -1.28429309e-01 4.60106820e-01
-2.35179469e-01 2.50147786e-02 1.03897071e+00 1.56282276e-01
-7.19757557e-01 7.89983571e-01 3.16197723e-01 2.92112082e-01
9.05223310e-01 -1.20740067e-02 -5.13427615e-01 -1.46156251e-01
-2.85413772e-01 4.09697145e-01 3.91982675e-01 6.47133887e-01
6.57185912e-01 -1.51784623e+00 -7.86231160e-01 2.12140635e-01
3.39699149e-01 1.00255147e-01 3.45092416e-02 5.27779698e-01
-5.97624421e-01 1.28790021e-01 -3.61769170e-01 -7.43295372e-01
-7.34614432e-01 3.84353846e-01 6.25432611e-01 -3.16330075e-01
-4.49123055e-01 1.01576984e+00 2.29749262e-01 -2.51048028e-01
4.13898587e-01 -4.98950422e-01 2.07240552e-01 3.95111293e-02
6.44298613e-01 6.50059521e-01 4.08082515e-01 -5.03254533e-01
-4.44952041e-01 5.82310259e-01 -2.93102443e-01 1.29086152e-01
1.32642496e+00 -7.06740320e-02 6.53452948e-02 8.21368471e-02
2.03352976e+00 -4.65749115e-01 -1.68718016e+00 -6.20204091e-01
-3.88624929e-02 -3.94392937e-01 3.13117415e-01 -6.15243196e-01
-1.69102991e+00 8.11957955e-01 7.06846893e-01 -4.01105195e-01
1.20676744e+00 -9.53899994e-02 9.51993704e-01 5.30452847e-01
4.32125032e-01 -1.04196501e+00 3.69556308e-01 3.56906354e-01
7.51552880e-01 -1.36501110e+00 -1.26994416e-01 1.55304596e-01
-3.42577785e-01 1.06784987e+00 6.21473134e-01 -4.77616817e-01
1.00829089e+00 3.36197078e-01 3.87205184e-02 -2.17903573e-02
-8.58558893e-01 5.05564697e-02 3.66111934e-01 4.50592935e-01
-2.43921019e-02 7.78343305e-02 -3.38436216e-02 5.20457625e-01
-1.24507941e-01 -2.05961838e-01 2.95618862e-01 5.98249018e-01
-3.88036817e-01 -9.63509440e-01 1.99948832e-01 6.11037612e-01
-2.04780281e-01 2.95250025e-02 -7.71519467e-02 5.05108595e-01
2.79014885e-01 1.02261770e+00 1.74586430e-01 -8.58196259e-01
3.04728627e-01 -2.52153575e-01 4.66880023e-01 -3.82600784e-01
-4.42344218e-01 -4.80030358e-01 -1.73238456e-01 -1.22955739e+00
-4.50835824e-01 -4.43659455e-01 -9.80161428e-01 -5.31326056e-01
-4.59567085e-02 -1.73694402e-01 7.55758464e-01 7.94109523e-01
4.56555873e-01 4.54926819e-01 1.06097579e+00 -8.03984284e-01
-9.78429556e-01 -9.75155652e-01 -5.04175842e-01 4.92999762e-01
6.13523185e-01 -7.13334978e-01 -6.90195858e-01 -1.17366277e-01] | [9.475468635559082, 3.019587993621826] |
5811f138-1784-4b17-b74b-82e93a0bc0e3 | annotating-columns-with-pre-trained-language | 2104.01785 | null | https://arxiv.org/abs/2104.01785v2 | https://arxiv.org/pdf/2104.01785v2.pdf | Annotating Columns with Pre-trained Language Models | Inferring meta information about tables, such as column headers or relationships between columns, is an active research topic in data management as we find many tables are missing some of this information. In this paper, we study the problem of annotating table columns (i.e., predicting column types and the relationships between columns) using only information from the table itself. We develop a multi-task learning framework (called Doduo) based on pre-trained language models, which takes the entire table as input and predicts column types/relations using a single model. Experimental results show that Doduo establishes new state-of-the-art performance on two benchmarks for the column type prediction and column relation prediction tasks with up to 4.0% and 11.9% improvements, respectively. We report that Doduo can already outperform the previous state-of-the-art performance with a minimal number of tokens, only 8 tokens per column. We release a toolbox (https://github.com/megagonlabs/doduo) and confirm the effectiveness of Doduo on a real-world data science problem through a case study. | ['Wang-Chiew Tan', 'Chen Chen', 'Çağatay Demiralp', 'Dan Zhang', 'Yuliang Li', 'Jinfeng Li', 'Yoshihiko Suhara'] | 2021-04-05 | null | null | null | null | ['type-prediction', 'table-annotation', 'table-annotation', 'column-type-annotation', 'columns-property-annotation'] | ['computer-code', 'knowledge-base', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.35638759e-01 2.06457108e-01 -5.57000339e-01 -4.01231378e-01
-1.00509477e+00 -6.76598012e-01 2.80376762e-01 9.45153713e-01
-1.01355486e-01 1.01908958e+00 1.67097196e-01 -5.49884975e-01
1.79703012e-01 -1.24360371e+00 -1.20882523e+00 -1.65793419e-01
-1.67987481e-01 8.25619161e-01 1.13591768e-01 -2.71132976e-01
1.30737290e-01 -1.82926178e-01 -1.43416977e+00 1.05017936e+00
7.30305374e-01 1.30983818e+00 -2.33300000e-01 5.66326559e-01
-7.18952775e-01 1.22716677e+00 -5.06899714e-01 -8.29907715e-01
6.89615980e-02 2.47386605e-01 -9.90771234e-01 -1.89177394e-01
4.34827089e-01 5.02385944e-02 -1.99393168e-01 7.35185266e-01
2.35551327e-01 -2.98373431e-01 3.28137308e-01 -1.34612298e+00
-5.77017844e-01 1.38808596e+00 -5.09723246e-01 -1.86185986e-02
4.57737625e-01 -5.76393068e-01 1.49824941e+00 -8.79277647e-01
5.39083660e-01 1.17531276e+00 6.52004957e-01 1.84609279e-01
-1.21770394e+00 -5.84591329e-01 2.48431310e-01 6.85744435e-02
-1.45281565e+00 -6.33179963e-01 3.81716490e-01 -3.88279468e-01
1.32652569e+00 5.28108537e-01 -1.66123196e-01 9.42375958e-01
6.66117609e-01 9.16586757e-01 8.54886949e-01 -4.03462172e-01
9.75003317e-02 4.59185243e-01 4.02279735e-01 7.24571705e-01
8.50493610e-01 -7.61896729e-01 -7.86270261e-01 -1.05402321e-01
2.50486016e-01 -6.37191832e-02 3.81510228e-01 -4.23081517e-01
-1.43870878e+00 4.50478643e-01 1.11471184e-01 4.90355566e-02
1.27423108e-02 -4.16296273e-02 7.71817327e-01 3.59463811e-01
6.82038605e-01 4.22968775e-01 -1.07090783e+00 -2.18846455e-01
-3.93762916e-01 2.64346331e-01 1.53876388e+00 1.62312818e+00
7.31522381e-01 -4.85365868e-01 -2.53082454e-01 8.89350712e-01
-1.17901647e-02 3.71416360e-01 9.46029574e-02 -5.36880493e-01
1.22428787e+00 9.04679477e-01 2.23392814e-01 -8.44678164e-01
-4.12323713e-01 -4.25672084e-01 -7.65094221e-01 -6.11730456e-01
6.28359020e-01 -2.61623502e-01 -5.17624021e-01 1.38406968e+00
1.63913757e-01 -4.35864359e-01 1.93317190e-01 2.18439102e-01
1.06663752e+00 6.32232428e-01 -9.90179032e-02 -9.30585191e-02
1.57467389e+00 -8.71092737e-01 -9.61827159e-01 -6.73448801e-01
1.22953534e+00 -7.37285852e-01 9.69315767e-01 4.48337287e-01
-8.62777472e-01 -4.39510942e-01 -9.64429379e-01 -4.95619416e-01
-1.02811861e+00 3.91500503e-01 1.04690313e+00 4.55612063e-01
-5.75241685e-01 3.68448377e-01 -8.03317249e-01 -2.23464042e-01
2.64465392e-01 2.84330815e-01 -5.28263032e-01 -1.12899356e-01
-1.15498972e+00 6.87404990e-01 5.79632998e-01 9.21531394e-02
-2.53415763e-01 -9.91153121e-01 -1.03235018e+00 2.54940897e-01
1.20671761e+00 -6.32844448e-01 1.09244168e+00 1.36011228e-01
-7.70782053e-01 9.09858704e-01 -5.23544669e-01 -5.14673233e-01
3.81831199e-01 -5.39605916e-01 -5.02120078e-01 -7.30803013e-01
2.73107827e-01 1.44828632e-01 -2.08507538e-01 -1.38879573e+00
-9.58733261e-01 -3.13132703e-01 1.17519848e-01 -1.31147534e-01
-4.31294680e-01 -1.40884012e-01 -9.24609125e-01 -5.33469379e-01
5.07534258e-02 -7.45714247e-01 8.48107040e-02 -2.98431307e-01
-1.26083112e+00 -5.06605566e-01 3.54450822e-01 -6.79313779e-01
1.76289725e+00 -1.76487732e+00 -1.55132920e-01 1.96679067e-02
5.64038038e-01 -2.30965510e-01 2.61393368e-01 7.69755304e-01
8.23872015e-02 4.07287061e-01 -1.76328465e-01 -5.23703754e-01
2.30527252e-01 2.63478428e-01 -4.55011159e-01 -1.14000022e-01
1.30203620e-01 8.44624460e-01 -6.32282794e-01 -6.37492657e-01
-1.40217930e-01 2.29574181e-02 -5.00775814e-01 1.35812983e-01
-6.05696857e-01 -2.48511255e-01 -3.57026190e-01 1.14148402e+00
7.39593565e-01 -4.99734938e-01 6.98707461e-01 -3.64421129e-01
-1.50717318e-01 8.64159822e-01 -1.34449422e+00 1.41754389e+00
-5.71085572e-01 2.84546763e-01 -3.88387412e-01 -8.67985308e-01
1.08389127e+00 -4.10655048e-03 5.70782483e-01 -5.89039266e-01
-8.03378671e-02 3.18725467e-01 9.76166576e-02 -5.03962338e-01
5.43061614e-01 4.06338006e-01 -7.07939565e-01 2.90052652e-01
5.24239875e-02 1.56824201e-01 7.88420558e-01 3.18490744e-01
1.15491760e+00 7.20158312e-03 5.77216089e-01 -4.30378824e-01
4.93265688e-01 4.07008566e-02 6.88569963e-01 8.64292979e-01
4.27755356e-01 2.26192370e-01 1.02087188e+00 -6.27435863e-01
-9.95194554e-01 -8.95871520e-01 -2.46837273e-01 1.17640340e+00
-6.37121946e-02 -1.18272090e+00 -4.01741564e-01 -6.55503511e-01
6.01931393e-01 7.78713346e-01 -9.53827024e-01 7.84966722e-02
-5.85664392e-01 -9.63575959e-01 2.44256541e-01 8.49516988e-01
3.30892205e-01 -7.24480093e-01 2.23720055e-02 3.71756852e-01
-3.42073202e-01 -1.50936043e+00 -1.22316524e-01 6.23035908e-01
-6.11472845e-01 -1.01815462e+00 2.51103729e-01 -6.73441291e-01
4.18946475e-01 -7.22684041e-02 1.72188413e+00 -1.27896801e-01
-7.84037784e-02 -2.15046272e-01 -2.79823303e-01 -7.04922616e-01
-2.57546216e-01 6.75082922e-01 -1.63225904e-01 -3.19899976e-01
7.23643363e-01 -2.75304079e-01 1.99250877e-01 1.90688625e-01
-5.28629422e-01 3.22302759e-01 7.00672686e-01 7.31373012e-01
8.19402575e-01 9.91081372e-02 3.95653844e-01 -1.98230159e+00
3.81043017e-01 -7.50676632e-01 -6.62608266e-01 7.88884640e-01
-9.21023011e-01 4.62800205e-01 8.48344088e-01 5.47955558e-02
-7.45437622e-01 -2.06080243e-01 1.41905308e-01 1.94869995e-01
6.28127530e-02 1.02839887e+00 -5.51777065e-01 5.71263134e-01
2.18514279e-01 2.39350628e-02 -4.16409850e-01 -7.02036917e-01
1.58424795e-01 6.03937984e-01 5.70102036e-01 -9.14097488e-01
7.53769577e-01 2.98248708e-01 4.55395095e-02 -1.19158514e-01
-1.44101536e+00 -3.28959554e-01 -8.52500260e-01 2.46946320e-01
3.64993513e-01 -1.04968107e+00 -1.11328208e+00 1.23934582e-01
-8.47973347e-01 -3.16070497e-01 1.70757338e-01 4.03651223e-02
-2.26470262e-01 -6.54229149e-02 -7.96875000e-01 -5.96369922e-01
-2.85970181e-01 -1.00119114e+00 1.04475963e+00 -3.19397688e-01
-1.47223830e-01 -1.23338211e+00 -2.13456884e-01 4.54619914e-01
1.21443272e-01 2.84261823e-01 1.38215423e+00 -9.91319060e-01
-7.98355877e-01 -2.73127377e-01 -2.80533999e-01 -1.65548548e-01
4.31425512e-01 5.48832491e-02 -8.06460798e-01 -4.92020324e-02
-4.00391668e-01 -4.23705965e-01 9.66929078e-01 -6.02373034e-02
1.70701051e+00 -4.88208681e-01 -7.07937717e-01 6.45325065e-01
1.54721248e+00 2.07043022e-01 4.16572571e-01 5.74263453e-01
1.07112134e+00 5.16615391e-01 8.91601026e-01 8.27484369e-01
9.65344369e-01 6.30309582e-01 2.74631590e-01 6.89678937e-02
3.81067283e-02 -5.51806331e-01 -6.31855354e-02 8.51539254e-01
2.90140986e-01 -6.90619886e-01 -1.15356624e+00 3.24765921e-01
-1.92537045e+00 -4.22826707e-01 -3.71607006e-01 2.15862727e+00
1.32067239e+00 7.58960664e-01 -1.01877280e-01 2.63393760e-01
4.06186074e-01 1.75063033e-02 -5.75648367e-01 -4.81082141e-01
-1.57594576e-01 -1.40035331e-01 9.15090740e-01 4.47835147e-01
-1.57950330e+00 1.04858387e+00 6.06556129e+00 5.46745121e-01
-7.69605279e-01 -3.03991526e-01 1.00640213e+00 2.57622488e-02
-3.65910619e-01 -3.73417921e-02 -1.48998487e+00 3.96693259e-01
1.05280304e+00 -4.80108529e-01 3.24530572e-01 8.73103380e-01
-2.28788108e-01 -1.00857303e-01 -1.62117398e+00 7.97834218e-01
6.41870201e-02 -1.62720215e+00 1.95803672e-01 1.45254076e-01
4.01539087e-01 -2.99350530e-01 -1.03425942e-02 6.23309970e-01
4.44335014e-01 -9.81520355e-01 7.08699286e-01 4.55985308e-01
8.77334833e-01 -7.12854862e-01 1.13141870e+00 8.83032233e-02
-1.30419457e+00 1.07077613e-01 -4.41567928e-01 -1.18784301e-01
-4.03861940e-01 9.82379198e-01 -9.04790699e-01 8.23551059e-01
8.75840962e-01 1.01930976e+00 -1.00035310e+00 6.81707025e-01
2.78947175e-01 4.77147281e-01 -1.79674476e-01 -1.07631795e-01
-2.42534623e-01 3.66762392e-02 1.02531454e-02 1.33178997e+00
5.80266230e-02 -2.26493746e-01 3.45486760e-01 6.54385388e-01
-7.55090356e-01 6.46257550e-02 -5.19795179e-01 -5.97877391e-02
7.66763747e-01 1.19450927e+00 -5.65013647e-01 -5.92389047e-01
-7.55402803e-01 2.76749641e-01 5.15610278e-01 2.05262043e-02
-7.04195619e-01 -5.51097214e-01 5.89564741e-01 -7.52831390e-03
4.64323312e-01 -3.40940878e-02 -9.94164586e-01 -1.28990352e+00
4.61400121e-01 -8.68849158e-01 7.18826830e-01 -4.49362695e-01
-1.29752851e+00 5.53152740e-01 3.13446261e-02 -1.01710844e+00
-3.27601463e-01 -9.46627021e-01 5.25443703e-02 6.92078531e-01
-1.34953415e+00 -9.87116992e-01 -2.64700800e-01 9.12226960e-02
3.98330688e-01 -1.62849292e-01 1.01659358e+00 4.19956833e-01
-9.35009062e-01 1.11325037e+00 2.27150485e-01 7.43466735e-01
1.01749957e+00 -1.75127804e+00 6.28885627e-01 7.11184561e-01
-1.72056407e-02 8.02330554e-01 8.06555212e-01 -6.46311641e-01
-2.12911034e+00 -1.16517639e+00 1.46029675e+00 -9.01683867e-01
9.21091676e-01 -1.11609495e+00 -9.37777817e-01 1.31005561e+00
9.22984257e-03 1.77208051e-01 9.67501283e-01 8.78698409e-01
-7.86312103e-01 -7.29785919e-01 -9.06239688e-01 4.72961813e-01
1.04691231e+00 -3.06443959e-01 -2.95112222e-01 5.62337995e-01
8.42040896e-01 -7.92185664e-01 -1.34829819e+00 4.84609693e-01
6.90700114e-01 -6.08152390e-01 8.80839765e-01 -5.93605280e-01
8.08104277e-01 -8.12240764e-02 -4.09667641e-01 -9.99402940e-01
-1.48918569e-01 -3.29424500e-01 -6.59152925e-01 1.57153630e+00
1.11673093e+00 -5.41015506e-01 8.13229799e-01 7.96848714e-01
-1.76727861e-01 -9.72470522e-01 -4.43590850e-01 -6.20370150e-01
3.11261937e-02 -4.72776234e-01 6.75738513e-01 9.37882662e-01
2.29381055e-01 4.98294950e-01 -3.48134786e-01 3.56190354e-02
4.70050305e-01 6.01401627e-01 9.69505668e-01 -1.25917661e+00
-1.68289438e-01 -1.60189271e-01 4.80033606e-02 -7.49951482e-01
1.76511437e-01 -1.08777905e+00 -1.50376782e-01 -1.78091896e+00
3.11823666e-01 -5.97426832e-01 -4.49361384e-01 9.58831072e-01
-3.42562556e-01 -1.37862265e-01 9.12420452e-02 1.17660508e-01
-8.30149293e-01 4.87904064e-02 8.67024660e-01 -3.24348420e-01
8.43902454e-02 1.06094100e-01 -1.27438855e+00 2.97506064e-01
7.38978028e-01 -6.35356247e-01 -1.98357612e-01 -6.61731601e-01
6.24482751e-01 6.74481615e-02 -4.44945246e-01 -8.81438136e-01
3.01964074e-01 -1.09572604e-01 5.08655548e-01 -8.19995761e-01
2.67892659e-01 -4.86189872e-01 -2.91526705e-01 3.10110480e-01
-7.00232625e-01 3.99444908e-01 4.94220287e-01 2.72763789e-01
-2.77911723e-01 2.90394515e-01 6.18536733e-02 -1.49963602e-01
-6.84322894e-01 2.04537258e-01 -7.94937760e-02 2.36820087e-01
7.16010153e-01 3.69297147e-01 -8.34465146e-01 5.08197024e-03
-3.79249781e-01 6.32609129e-01 4.93022539e-02 5.77577233e-01
3.79153550e-01 -1.09800291e+00 -7.96010256e-01 1.46629319e-01
6.61508083e-01 1.25865832e-01 -1.78879082e-01 4.77473885e-01
-4.53563154e-01 1.04165483e+00 5.58302328e-02 -3.59090090e-01
-1.15929282e+00 8.65592360e-01 -4.61278287e-05 -6.93129838e-01
-4.01418269e-01 9.13593054e-01 -2.64175564e-01 -6.95309699e-01
4.02951986e-01 -9.95512068e-01 -9.62031037e-02 1.60356775e-01
3.75146419e-01 -7.18415082e-02 6.35802150e-01 1.42621666e-01
-6.52496457e-01 -3.14266644e-02 -4.83739853e-01 4.10340816e-01
1.35504234e+00 9.16772038e-02 -5.09900630e-01 1.11312616e+00
1.08741200e+00 3.86011451e-01 -7.71270692e-01 -5.60254037e-01
5.32757282e-01 -4.83579367e-01 -3.90329361e-01 -1.27251983e+00
-8.85084748e-01 6.05204642e-01 -2.12571368e-01 4.27948326e-01
8.01860034e-01 1.95873573e-01 6.89344347e-01 8.37697506e-01
4.92706656e-01 -8.29511225e-01 -3.98833185e-01 5.76914668e-01
6.72915101e-01 -1.47034562e+00 2.15033308e-01 -8.91742289e-01
-6.15321279e-01 1.09904575e+00 8.81703079e-01 1.30723089e-01
8.37636471e-01 9.63013589e-01 -2.17322484e-02 9.05001685e-02
-1.67818666e+00 -5.75491972e-02 1.99237436e-01 2.21995980e-01
1.23931336e+00 3.00462157e-01 -1.28800645e-01 9.09184813e-01
-5.83889544e-01 -5.51864468e-02 5.98466337e-01 1.11562324e+00
-7.97145665e-02 -1.47395110e+00 -2.08948135e-01 1.06807625e+00
-5.13333261e-01 -4.77913171e-01 -4.04879272e-01 8.10419679e-01
-7.78952334e-03 9.64731038e-01 3.30113739e-01 -6.78908587e-01
3.04908454e-01 1.76890865e-01 2.46327549e-01 -9.45511222e-01
-4.10265356e-01 -8.98935646e-02 6.38211846e-01 -5.28416693e-01
9.43271667e-02 -6.83754742e-01 -1.04575562e+00 -6.30189836e-01
1.76890805e-01 3.79707605e-01 4.60796148e-01 8.41201663e-01
5.30838072e-01 9.24646914e-01 4.40887660e-01 1.51454017e-01
-3.99195582e-01 -1.14100432e+00 -4.19360161e-01 7.60023519e-02
1.98143959e-01 -6.40020311e-01 5.20683751e-02 5.30308560e-02] | [9.616878509521484, 7.894745826721191] |
da59c980-ac95-49f3-915e-fafc9e60a35c | assessing-the-unitary-rnn-as-an-end-to-end | 2208.05719 | null | https://arxiv.org/abs/2208.05719v1 | https://arxiv.org/pdf/2208.05719v1.pdf | Assessing the Unitary RNN as an End-to-End Compositional Model of Syntax | We show that both an LSTM and a unitary-evolution recurrent neural network (URN) can achieve encouraging accuracy on two types of syntactic patterns: context-free long distance agreement, and mildly context-sensitive cross serial dependencies. This work extends recent experiments on deeply nested context-free long distance dependencies, with similar results. URNs differ from LSTMs in that they avoid non-linear activation functions, and they apply matrix multiplication to word embeddings encoded as unitary matrices. This permits them to retain all information in the processing of an input string over arbitrary distances. It also causes them to satisfy strict compositionality. URNs constitute a significant advance in the search for explainable models in deep learning applied to NLP. | ['Shalom Lappin', 'Jean-Philippe Bernardy'] | 2022-08-11 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [ 2.75826812e-01 2.27737904e-01 -3.32868338e-01 -6.52591646e-01
-7.47285485e-01 -6.96162701e-01 4.32137489e-01 1.55913681e-01
-6.83002234e-01 7.51285195e-01 5.81774473e-01 -1.08878028e+00
7.47340731e-03 -7.11578071e-01 -7.88918734e-01 -6.14167750e-01
-2.17432618e-01 4.71287042e-01 -9.40946192e-02 -5.16361296e-01
2.08572164e-01 4.16408390e-01 -8.74345839e-01 6.23099566e-01
7.09991574e-01 4.14798677e-01 -2.16479562e-02 6.95484400e-01
-3.48987997e-01 9.52064395e-01 -3.67020816e-01 -4.75981414e-01
-3.87237445e-02 -4.39021289e-01 -1.02050447e+00 -8.17473054e-01
3.65639210e-01 2.06076764e-02 -4.03536588e-01 8.72106254e-01
3.71352673e-01 3.11134845e-01 5.79946935e-01 -6.16066277e-01
-1.54196060e+00 1.42845321e+00 -2.76386857e-01 5.11674225e-01
7.98763260e-02 -2.70224959e-01 1.85491109e+00 -7.88863063e-01
7.41862476e-01 1.42749333e+00 8.75294209e-01 6.78460121e-01
-1.41662216e+00 -3.97607505e-01 1.60865784e-01 1.27331227e-01
-8.80194247e-01 -2.92012423e-01 5.31784892e-01 -1.24836802e-01
1.85423172e+00 2.94694960e-01 1.12856172e-01 1.25379884e+00
5.81232965e-01 5.89856088e-01 7.93502569e-01 -6.63654804e-01
-2.41900794e-02 -4.43968982e-01 7.68001854e-01 5.01107514e-01
9.60489661e-02 2.47521862e-01 -3.98837715e-01 -2.00625136e-01
6.60296261e-01 -2.04907581e-01 -1.18602768e-01 2.00657435e-02
-1.10717583e+00 1.19767547e+00 8.73509169e-01 1.06999052e+00
6.08352311e-02 4.51340884e-01 4.68603522e-01 7.51666248e-01
5.14685869e-01 6.60670400e-01 -9.31706250e-01 -7.11776540e-02
-2.89956868e-01 -8.41430873e-02 9.45425808e-01 5.84391057e-01
5.95855057e-01 2.68953860e-01 2.58734971e-02 7.77596176e-01
-2.19333358e-02 2.79836804e-01 7.61717618e-01 -7.38537073e-01
5.25977969e-01 2.63429463e-01 -3.74421954e-01 -6.51548445e-01
-6.67749286e-01 -3.52695793e-01 -6.52980030e-01 -6.90056756e-02
2.59656578e-01 -2.98645586e-01 -9.76699412e-01 2.30689096e+00
-2.39721164e-01 -1.84897542e-01 3.28375697e-01 7.03253508e-01
6.23248696e-01 8.80575538e-01 2.02424139e-01 -2.50595093e-01
1.18338215e+00 -6.88268065e-01 -8.03163946e-01 -4.08572048e-01
1.16796792e+00 -5.68549454e-01 1.01180565e+00 -1.48118183e-01
-1.21425498e+00 -4.86936271e-01 -9.00550544e-01 -7.57777512e-01
-5.26782215e-01 -4.54237014e-01 7.57252097e-01 4.86534238e-01
-1.35587442e+00 9.08835411e-01 -7.25961447e-01 -3.66555214e-01
-1.67747930e-01 5.86147726e-01 -3.56744230e-01 2.98665822e-01
-1.61576915e+00 1.49910831e+00 4.19924498e-01 3.63183349e-01
6.35524141e-03 -7.13277817e-01 -1.20068526e+00 1.49739638e-01
-1.68950453e-01 -6.27445936e-01 1.30614674e+00 -9.38211858e-01
-1.17888868e+00 8.29678416e-01 -4.97816533e-01 -6.48870409e-01
-5.35099469e-02 -2.21259162e-01 -3.54774028e-01 -6.14234567e-01
-1.75213277e-01 6.53056204e-01 3.04527014e-01 -7.59407282e-01
-2.60937095e-01 -4.03034687e-01 -8.48116428e-02 1.62071645e-01
-5.46117872e-02 3.75415534e-01 2.75510907e-01 -6.54434443e-01
2.23443612e-01 -1.02439833e+00 -2.65299648e-01 -6.07127249e-01
-2.80027688e-01 -6.05553448e-01 6.90974236e-01 -8.05949867e-01
1.12880969e+00 -1.98701167e+00 7.48379648e-01 5.52082993e-02
-7.17428401e-02 2.67598867e-01 -4.69891578e-01 4.64557230e-01
-4.25285488e-01 5.30112684e-01 -5.15568852e-01 -4.68537122e-01
2.16783449e-01 9.40683424e-01 -5.65215230e-01 3.09633642e-01
4.99142259e-01 1.53040278e+00 -6.23088598e-01 -7.66668469e-02
-8.71522427e-02 6.43974185e-01 -4.64938760e-01 -1.77713260e-02
-3.11103076e-01 8.25867876e-02 1.17996246e-01 1.96470097e-01
2.95640856e-01 4.30602357e-02 6.34317875e-01 -2.91775148e-02
-3.77447575e-01 8.57994199e-01 -4.07179564e-01 1.76813126e+00
-7.19245613e-01 7.33358741e-01 -1.50478393e-01 -1.10193419e+00
8.36024463e-01 3.28945577e-01 -5.91397099e-02 -8.88320506e-01
2.37409770e-01 5.36496401e-01 6.50733888e-01 -2.31964946e-01
4.93383586e-01 -4.34383184e-01 -3.11583191e-01 6.98132396e-01
1.77005216e-01 1.48247033e-01 1.48625523e-01 1.48130402e-01
1.05198812e+00 1.42432690e-01 2.18318760e-01 -5.47541440e-01
1.91677257e-01 -5.74200332e-01 6.67581618e-01 7.10274816e-01
1.77425891e-01 4.98151124e-01 6.82524681e-01 -8.40159416e-01
-1.27761030e+00 -1.11221421e+00 -3.43893290e-01 1.57228959e+00
-3.47897500e-01 -2.60614276e-01 -3.15722466e-01 -6.06221199e-01
-2.36777514e-02 1.05809355e+00 -1.11598527e+00 -1.11712247e-01
-1.39280844e+00 -7.60931432e-01 7.18175054e-01 8.58080208e-01
-2.47219920e-01 -1.38105977e+00 -1.70649230e-01 3.95652294e-01
-1.12727463e-01 -1.01108909e+00 -6.57544494e-01 9.30338383e-01
-9.38650250e-01 -6.93596900e-01 -3.10948044e-01 -1.22983444e+00
4.32156861e-01 -2.69255102e-01 1.13860536e+00 1.65212154e-01
4.63952161e-02 -2.48499826e-01 -1.83181077e-01 -1.08627513e-01
-5.97966254e-01 5.48429489e-01 1.61739439e-01 -4.43014532e-01
7.14820027e-01 -8.38917851e-01 3.80763449e-02 -5.13681844e-02
-8.01954508e-01 -4.73131359e-01 4.50306207e-01 1.04985285e+00
3.17765445e-01 -8.38683784e-01 4.70977724e-01 -9.32695150e-01
8.79846990e-01 -4.43767607e-01 -2.30577201e-01 3.46008569e-01
-3.86532485e-01 7.91546583e-01 6.53728247e-01 -4.43117112e-01
-1.00430739e+00 -2.58570880e-01 -4.87789899e-01 6.53454363e-02
7.62051716e-02 5.11582375e-01 2.34437004e-01 1.13190077e-01
7.06646383e-01 -2.41160989e-02 -1.65514722e-01 -4.77120191e-01
6.76550686e-01 2.78456479e-01 4.65231448e-01 -7.76601732e-01
5.85942090e-01 -1.27245272e-02 -1.70721576e-01 -6.91241443e-01
-9.82753575e-01 4.27743904e-02 -1.02634561e+00 6.13518178e-01
1.17345166e+00 -3.38774115e-01 -5.60006082e-01 2.34439149e-02
-1.64525139e+00 -5.03859401e-01 -3.47518981e-01 4.24963921e-01
-2.26124391e-01 2.78305560e-01 -1.58492470e+00 -5.22979081e-01
-5.49504399e-01 -7.25296736e-01 7.67083347e-01 -3.10421996e-02
-6.72164679e-01 -1.44695210e+00 2.53670037e-01 -1.92908287e-01
6.01805687e-01 -1.19903617e-01 1.51260185e+00 -9.16670620e-01
-2.39318371e-01 2.17255712e-01 -2.10886002e-01 2.91979074e-01
9.40648615e-02 1.45668089e-01 -7.17616856e-01 -1.60853922e-01
-4.01035286e-02 -3.82055908e-01 1.25129402e+00 5.26279330e-01
5.93945265e-01 -3.31582576e-01 -1.33532688e-01 6.75888419e-01
9.92026448e-01 2.83615649e-01 5.24309456e-01 1.88668519e-01
7.79917777e-01 6.12399995e-01 -2.81719923e-01 -3.32423329e-01
2.34726056e-01 4.03333992e-01 1.40150264e-01 -7.42842406e-02
-4.80165184e-02 -2.32995152e-01 4.60795313e-01 1.58465111e+00
-4.66825143e-02 -2.91965827e-02 -9.29345131e-01 5.43403208e-01
-1.91272938e+00 -8.50287437e-01 -3.90108824e-01 1.69999146e+00
9.11357701e-01 1.99917197e-01 -4.27236170e-01 -1.43557206e-01
6.38030469e-01 3.71446878e-01 -4.22363311e-01 -1.39713156e+00
-5.01868546e-01 7.98132300e-01 3.49897802e-01 1.01994002e+00
-8.36612642e-01 1.25244200e+00 7.02097178e+00 3.25978547e-01
-1.01619875e+00 3.89172435e-01 4.66850162e-01 -4.31208499e-03
-8.04635823e-01 -2.74958219e-02 -6.85945630e-01 9.75385308e-02
1.34079063e+00 8.32616985e-02 2.21303761e-01 4.60322440e-01
-2.62682617e-01 3.55315238e-01 -1.38289428e+00 5.10202408e-01
1.27034150e-02 -1.38555956e+00 1.55320853e-01 -4.68486100e-02
7.02567816e-01 6.17051005e-01 1.58852890e-01 4.54436630e-01
8.28488827e-01 -1.35942030e+00 3.88112724e-01 -4.72764932e-02
5.89698374e-01 -8.55545282e-01 7.77815402e-01 9.92821380e-02
-1.02015877e+00 -1.67300060e-01 -7.77709305e-01 -5.04069090e-01
4.09748346e-01 3.02368611e-01 -3.52725565e-01 3.66447300e-01
3.96324158e-01 5.52654922e-01 -2.26725385e-01 9.82266292e-02
-5.63299596e-01 6.40430689e-01 -3.01442146e-01 -2.43680775e-01
6.88202977e-01 -3.14609408e-01 3.03095579e-01 1.50857329e+00
1.92245051e-01 2.07451239e-01 -2.29920685e-01 8.16745579e-01
-1.80824488e-01 1.31969899e-01 -6.33353055e-01 1.45846277e-01
2.22658023e-01 8.64135206e-01 -3.93842191e-01 -8.54131356e-02
-3.33171427e-01 9.64368224e-01 9.97982204e-01 5.05299628e-01
-7.22714484e-01 -3.29449534e-01 9.34196174e-01 -4.02491212e-01
4.00840074e-01 -6.18697405e-01 -3.70245516e-01 -1.33348274e+00
-6.30592853e-02 -5.72329402e-01 6.12336695e-01 -5.57674646e-01
-1.39227378e+00 9.22347069e-01 -4.47744608e-01 -2.43924960e-01
-5.66252410e-01 -9.62627888e-01 -7.25715876e-01 1.05012405e+00
-1.59082878e+00 -1.31405342e+00 7.02384293e-01 3.14575195e-01
5.28831184e-01 1.66472182e-01 1.22565079e+00 9.32970718e-02
-5.35878778e-01 5.06660461e-01 1.03510536e-01 4.43633199e-01
4.84995604e-01 -1.44134307e+00 1.09504199e+00 9.23183262e-01
7.02048957e-01 1.14055598e+00 6.31902814e-01 -4.47487622e-01
-1.31319058e+00 -7.55482137e-01 1.58202040e+00 -6.36274993e-01
1.08547807e+00 -6.81896389e-01 -1.21681273e+00 1.44436145e+00
6.14263535e-01 2.54123658e-02 6.93852782e-01 8.24903488e-01
-9.34380949e-01 2.35117614e-01 -4.92447466e-01 6.37814581e-01
1.32640147e+00 -9.45858240e-01 -1.27565253e+00 5.21761812e-02
1.26481938e+00 -3.10103089e-01 -8.22229981e-01 4.64118630e-01
5.65365255e-01 -8.22760582e-01 5.60396850e-01 -1.23940647e+00
4.51267600e-01 2.04180136e-01 -3.57923955e-01 -1.46901906e+00
-7.69203603e-01 -6.09847546e-01 1.89089421e-02 9.36736941e-01
1.09021592e+00 -8.88592660e-01 3.29790682e-01 7.28566706e-01
-3.69361818e-01 -5.87859213e-01 -1.09771407e+00 -6.79992855e-01
9.61011469e-01 -4.50645238e-01 3.24722379e-01 1.19286060e+00
3.65580678e-01 7.41353571e-01 -2.46287093e-01 -9.24559534e-02
3.80385160e-01 3.18584591e-01 -6.15941640e-03 -1.20976722e+00
-3.09532374e-01 -6.62912011e-01 -2.12873697e-01 -1.01819324e+00
8.06298673e-01 -1.28552687e+00 -8.02009553e-02 -1.40299487e+00
8.29561800e-03 -3.74895066e-01 -5.14676332e-01 6.78317249e-01
-8.49171728e-02 9.12505016e-02 7.75449574e-02 3.45366374e-02
-1.48599893e-01 4.02132750e-01 6.47325277e-01 -4.34478000e-03
-1.07694104e-01 -3.92378718e-01 -6.21561706e-01 5.72866976e-01
1.04854405e+00 -5.91216385e-01 -7.95155317e-02 -1.05892730e+00
6.78426385e-01 -1.28871739e-01 -1.43316820e-01 -3.81889105e-01
1.71164185e-01 1.04546264e-01 2.82218784e-01 -2.87122607e-01
1.42145678e-01 -4.37908977e-01 -1.15279667e-01 6.97453082e-01
-8.13258648e-01 5.81863225e-01 1.82976782e-01 8.61208327e-03
-1.25151962e-01 -1.55894756e-01 6.21888280e-01 -1.48661077e-01
-5.59650898e-01 -3.02199144e-02 -4.46247250e-01 1.15704812e-01
5.94495595e-01 1.68476179e-01 -6.07185781e-01 -1.98505193e-01
-9.96047437e-01 2.98422499e-04 6.47028014e-02 7.02328622e-01
2.41629198e-01 -1.32877135e+00 -8.42814088e-01 3.07567775e-01
-3.52029890e-01 -4.04001117e-01 -1.14999572e-02 5.86309969e-01
-2.02858314e-01 7.11481929e-01 -1.30857527e-01 -2.62903720e-01
-1.13048959e+00 6.06392205e-01 4.59370822e-01 -3.88311535e-01
-5.00670552e-01 1.20159400e+00 1.06368572e-01 -1.10226226e+00
-2.70380750e-02 -8.31286728e-01 1.03701681e-01 1.39155447e-01
3.45194817e-01 -1.81744307e-01 1.15957208e-01 -7.36307144e-01
-3.00817460e-01 6.18181109e-01 -2.77332634e-01 -3.89288664e-02
1.51713598e+00 1.33705735e-01 -5.76799452e-01 9.28739369e-01
1.48947799e+00 -5.35844527e-02 -7.36183286e-01 -4.09317851e-01
4.83055443e-01 2.69338965e-01 -3.52777451e-01 -5.32121003e-01
-8.27340662e-01 1.13116407e+00 -6.13544807e-02 3.18415701e-01
4.09589171e-01 2.69899935e-01 1.04335582e+00 6.33097410e-01
-1.97084118e-02 -8.44916523e-01 -1.78653911e-01 1.23274541e+00
7.41647720e-01 -9.09539402e-01 -4.82520849e-01 3.34563963e-02
-3.07564765e-01 1.32017791e+00 5.02221107e-01 -1.98569074e-01
5.18293798e-01 6.63849592e-01 2.32980594e-01 1.98273920e-02
-1.06989694e+00 -2.12049574e-01 1.84611440e-01 3.16527158e-01
9.98864770e-01 6.24307878e-02 -5.65038621e-01 3.44828278e-01
-3.68724555e-01 -6.14104331e-01 4.10418689e-01 7.03858316e-01
-3.22632492e-01 -1.44244623e+00 -3.95797454e-02 2.45295897e-01
-5.63908458e-01 -6.72865748e-01 -4.65669334e-01 8.04187536e-01
-1.82252809e-01 5.91332555e-01 4.27186638e-01 -1.92741677e-01
1.63664132e-01 5.94078958e-01 4.56395090e-01 -6.08915806e-01
-7.81872332e-01 -2.59438485e-01 2.55318493e-01 -4.52389538e-01
-2.49921173e-01 -6.08555913e-01 -1.58377790e+00 -4.77499038e-01
-3.86593878e-01 3.74120235e-01 6.69815660e-01 9.53769207e-01
2.67162979e-01 5.20449102e-01 2.34251574e-01 -7.36989915e-01
-7.41328180e-01 -9.79500353e-01 -3.07299137e-01 4.94687319e-01
4.98194247e-01 -7.65057281e-02 -4.52442855e-01 -3.74507695e-01] | [10.570244789123535, 9.146241188049316] |
5d690c90-37f2-4392-b3cf-2c912100b21e | learning-structured-embeddings-of-knowledge | 2004.07265 | null | https://arxiv.org/abs/2004.07265v1 | https://arxiv.org/pdf/2004.07265v1.pdf | Learning Structured Embeddings of Knowledge Graphs with Adversarial Learning Framework | Many large-scale knowledge graphs are now available and ready to provide semantically structured information that is regarded as an important resource for question answering and decision support tasks. However, they are built on rigid symbolic frameworks which makes them hard to be used in other intelligent systems. We present a learning method using generative adversarial architecture designed to embed the entities and relations of the knowledge graphs into a continuous vector space. A generative network (GN) takes two elements of a (subject, predicate, object) triple as input and generates the vector representation of the missing element. A discriminative network (DN) scores a triple to distinguish a positive triple from those generated by GN. The training goal for GN is to deceive DN to make wrong classification. When arriving at a convergence, GN recovers the training data and can be used for knowledge graph completion, while DN is trained to be a good triple classifier. Unlike few previous studies based on generative adversarial architectures, our GN is able to generate unseen instances while they just use GN to better choose negative samples (already existed) for DN. Experiments demonstrate our method can improve classical relational learning models (e.g.TransE) with a significant margin on both the link prediction and triple classification tasks. | ['Xiaoqing Zheng', 'Jiehang Zeng', 'Lu Liu'] | 2020-04-15 | null | null | null | null | ['triple-classification'] | ['graphs'] | [ 2.46641055e-01 8.04743350e-01 -2.79614061e-01 -1.67314276e-01
-5.12859106e-01 -7.03333437e-01 6.65263414e-01 1.30879745e-01
-1.81799904e-02 8.94836485e-01 -2.24359244e-01 -3.72543842e-01
-6.37956560e-02 -1.59777486e+00 -1.06679630e+00 -6.84921145e-01
1.13631912e-01 9.86956537e-01 3.85244548e-01 -6.21423364e-01
-2.79215604e-01 5.05345941e-01 -1.49927127e+00 4.08944130e-01
9.82833982e-01 9.37684417e-01 -3.76721412e-01 3.64759982e-01
-1.70505136e-01 1.15599525e+00 -8.24995458e-01 -1.01790726e+00
2.51779169e-01 -5.08322418e-01 -9.68395054e-01 -4.21750784e-01
3.15912068e-01 -7.49193057e-02 -5.75361669e-01 1.11832750e+00
2.79964507e-01 1.36405483e-01 8.25150788e-01 -1.66854882e+00
-1.13710487e+00 9.68926072e-01 4.23224606e-02 -1.78048164e-01
5.69789052e-01 -2.18642578e-01 1.19311559e+00 -9.00105894e-01
8.51896703e-01 1.34884429e+00 4.68065679e-01 6.90694213e-01
-1.13267982e+00 -6.82880163e-01 -1.99771672e-01 5.17306805e-01
-1.42501414e+00 -2.38013119e-01 9.21977162e-01 -1.81521580e-01
6.99230313e-01 4.87085640e-01 6.70537829e-01 1.42229760e+00
2.14223638e-02 6.31352484e-01 6.18184209e-01 -4.43524539e-01
3.14036399e-01 2.97431797e-01 8.80732536e-02 8.49760592e-01
2.82966524e-01 2.76358556e-02 -3.00352097e-01 -2.19179899e-01
6.16204858e-01 -6.00008555e-02 -2.79136509e-01 -6.67536914e-01
-9.87081468e-01 9.43315089e-01 1.04919803e+00 2.62503922e-01
-4.77585375e-01 -2.65312102e-03 1.13538727e-01 6.63858235e-01
4.49743681e-02 3.34530622e-01 -1.70557171e-01 3.60402465e-01
-3.68923753e-01 3.89078438e-01 1.06195128e+00 8.66363049e-01
8.03209126e-01 2.48654798e-01 -2.90248245e-01 5.58745861e-01
1.83115497e-01 6.32257283e-01 4.78980511e-01 -5.50743103e-01
5.36547482e-01 1.19141209e+00 -2.41590232e-01 -1.34354699e+00
-1.72306105e-01 -3.84757698e-01 -8.98712993e-01 1.93348393e-01
4.82869804e-01 9.90255550e-02 -1.06835413e+00 1.68825424e+00
4.34748262e-01 2.15467572e-01 3.99523646e-01 7.03166842e-01
1.28813863e+00 4.45050806e-01 6.73451424e-02 2.77680069e-01
1.21509004e+00 -6.81800425e-01 -4.09700900e-01 -2.70603120e-01
7.22264647e-01 -3.26161444e-01 6.64229870e-01 1.96698532e-01
-8.31574082e-01 -4.18721884e-01 -1.08702111e+00 -8.16315040e-02
-8.74976277e-01 -8.13904777e-02 6.09349787e-01 6.32341862e-01
-9.85506117e-01 7.01934099e-01 -5.69937289e-01 -8.85254741e-02
5.34897268e-01 4.45480138e-01 -6.80067837e-01 -3.61166000e-01
-1.69457757e+00 9.88249600e-01 9.03762043e-01 2.02734604e-01
-8.88382018e-01 -4.80940312e-01 -1.02309358e+00 3.67840938e-02
6.85087621e-01 -8.54734421e-01 5.34526229e-01 -8.87927532e-01
-1.10414708e+00 8.22434068e-01 2.66732186e-01 -4.67382938e-01
6.32620752e-01 1.97673395e-01 -6.09634042e-01 2.78009698e-02
4.73924428e-02 5.51398575e-01 1.02039182e+00 -1.51998925e+00
-2.49349698e-01 -3.39075327e-01 3.27949882e-01 5.96619658e-02
-7.25740492e-02 -3.61831158e-01 -1.94374517e-01 -6.00933850e-01
3.83633167e-01 -9.46897030e-01 8.13919976e-02 -2.17828467e-01
-8.95876586e-01 -3.05791289e-01 8.95441771e-01 -7.32383847e-01
8.60077262e-01 -1.75614560e+00 5.01506090e-01 6.59995854e-01
5.56446195e-01 7.04082668e-01 -6.47190884e-02 2.89599538e-01
-4.77043211e-01 2.85115242e-01 -1.36565328e-01 -8.11573025e-03
-8.20161477e-02 6.46748960e-01 -4.20179665e-01 2.10021570e-01
3.40362877e-01 1.23172855e+00 -1.07884610e+00 -3.15749317e-01
4.27637994e-02 4.80848730e-01 -4.10006195e-01 2.54225343e-01
-3.52332383e-01 1.76770166e-01 -5.89950800e-01 7.27239549e-01
3.24380904e-01 -1.74544841e-01 2.30634466e-01 -4.15641248e-01
9.12600696e-01 4.48097996e-02 -1.18250334e+00 1.03145885e+00
-2.12722927e-01 1.82839245e-01 -5.18289387e-01 -1.21985626e+00
1.21217859e+00 3.12500924e-01 -8.53525251e-02 -5.05469859e-01
1.69626147e-01 1.65875047e-01 9.54412520e-02 -3.17843229e-01
4.61233914e-01 -2.48775154e-01 -1.42951563e-01 9.50823277e-02
2.98514336e-01 2.54631676e-02 1.36160078e-02 5.03507972e-01
1.21877050e+00 2.12977484e-01 1.28222331e-01 3.40126961e-01
7.52479911e-01 -3.07187960e-02 6.40683353e-01 6.10177934e-01
1.07230842e-01 4.20276582e-01 7.89586723e-01 -4.21768308e-01
-9.14656222e-01 -1.37214541e+00 3.02188754e-01 8.37038696e-01
8.51320103e-03 -3.27817649e-01 -5.48190653e-01 -1.26592731e+00
1.38986036e-01 1.10843039e+00 -8.37346137e-01 -7.62300909e-01
-4.28617537e-01 -3.22574437e-01 7.52564728e-01 5.55927157e-01
3.38464111e-01 -1.29105234e+00 1.44559830e-01 7.16248006e-02
-3.09419572e-01 -9.36694026e-01 1.34341136e-01 1.49105713e-01
-3.73191595e-01 -1.49653482e+00 -2.38546476e-01 -7.99989223e-01
9.41257119e-01 -1.87256604e-01 1.17532027e+00 1.93933517e-01
-1.23533241e-01 2.43681028e-01 -4.59629595e-01 -2.00934738e-01
-9.43057597e-01 -2.02869978e-02 1.53789809e-02 2.66019076e-01
4.29671198e-01 -5.99676132e-01 -4.24806029e-02 3.11853915e-01
-1.25119841e+00 -2.38464817e-01 3.37917000e-01 9.72116768e-01
6.43902898e-01 1.16067685e-01 7.76991665e-01 -1.42219150e+00
6.01181448e-01 -7.34017313e-01 -2.83570319e-01 6.45180643e-01
-4.57298458e-01 2.58029550e-01 9.19955492e-01 -4.11004752e-01
-7.52340615e-01 -2.76688844e-01 -7.51593858e-02 -7.22819984e-01
6.40003532e-02 5.35208106e-01 -4.19134349e-01 -1.07935168e-01
1.04908359e+00 6.60781488e-02 -1.34601921e-01 -2.25409508e-01
7.12244868e-01 2.52832711e-01 5.96235454e-01 -5.81322849e-01
1.25459266e+00 1.56764403e-01 2.06035793e-01 -4.18899953e-01
-8.20161700e-01 5.95495738e-02 -7.17374921e-01 1.98103618e-02
6.98960721e-01 -7.46345401e-01 -4.59957927e-01 8.66861790e-02
-9.87829864e-01 4.94472533e-02 -5.73163688e-01 1.24751016e-01
-2.30906650e-01 -3.08246515e-03 -2.70935237e-01 -5.10593653e-01
-1.80736408e-01 -7.89483666e-01 5.80584347e-01 1.47153363e-01
-1.52379245e-01 -1.32628512e+00 -5.71156926e-02 5.37669718e-01
1.52119517e-01 6.11422241e-01 1.32760322e+00 -1.26899362e+00
-5.20279169e-01 -5.68306804e-01 1.38458937e-01 5.32735229e-01
-6.58865720e-02 8.81395712e-02 -8.26936543e-01 -7.18893111e-02
-2.47951016e-01 -6.47473454e-01 7.95017838e-01 -4.58110094e-01
8.22086453e-01 -5.56220829e-01 -5.09302199e-01 4.90047216e-01
1.23111951e+00 1.03011958e-01 9.72256839e-01 1.56777576e-01
1.16433287e+00 4.34389859e-01 3.71974111e-01 -1.93315104e-01
4.24147815e-01 4.77030218e-01 6.49091721e-01 1.16921142e-01
-2.08461910e-01 -6.33628607e-01 3.37885767e-01 5.33673763e-01
-2.38515705e-01 -4.39395219e-01 -1.02855921e+00 4.71268445e-01
-1.83573890e+00 -1.05484557e+00 -2.01453850e-01 2.20395112e+00
9.13413882e-01 2.10227132e-01 -2.29530901e-01 3.33323181e-01
7.55640090e-01 -8.26816708e-02 -5.35871685e-01 -2.27353215e-01
-3.07791859e-01 4.59398210e-01 2.01979697e-01 3.27477306e-01
-9.14970040e-01 1.05235267e+00 5.40808201e+00 8.10832143e-01
-8.10715497e-01 -5.35922907e-02 4.02849019e-01 3.36733848e-01
-7.68783629e-01 7.26928860e-02 -5.65769792e-01 8.02301541e-02
8.93286645e-01 -1.95445105e-01 5.41102767e-01 8.72780085e-01
-5.93076706e-01 2.99001992e-01 -1.37103522e+00 8.78520250e-01
2.39823252e-01 -1.20402622e+00 5.69337666e-01 -1.55443773e-01
6.69714868e-01 -2.72974998e-01 -8.35557282e-02 8.11001182e-01
7.26042211e-01 -1.28674281e+00 4.73107725e-01 7.82895625e-01
5.50385237e-01 -8.56831968e-01 9.80617404e-01 3.66245896e-01
-8.12747717e-01 5.77115081e-02 -4.41481590e-01 3.02242935e-01
-2.39649028e-01 5.96417308e-01 -1.24543381e+00 1.09588075e+00
2.73919940e-01 3.86503518e-01 -8.38563502e-01 5.50012231e-01
-8.72805536e-01 4.79632616e-01 -2.28345543e-01 1.31400570e-01
-1.35770980e-02 -2.01837078e-01 6.82681799e-01 5.53321838e-01
1.68994784e-01 -1.11208502e-02 2.43600756e-01 1.12935233e+00
-5.81415594e-01 -5.69562754e-03 -9.84198034e-01 -2.09153816e-01
4.82805371e-01 1.24083686e+00 -4.83818561e-01 -3.63389820e-01
-2.03462914e-01 9.43693101e-01 7.34393775e-01 4.46066827e-01
-8.13304782e-01 -4.39123333e-01 2.38059312e-01 5.68160564e-02
2.37653762e-01 1.92589670e-01 4.06421088e-02 -1.24628425e+00
1.01842940e-01 -8.66819680e-01 5.16244590e-01 -8.71760309e-01
-1.36341512e+00 7.40166306e-01 -2.52780974e-01 -1.06240916e+00
-7.01303780e-01 -5.21151245e-01 -5.44807792e-01 9.96499538e-01
-1.18012583e+00 -1.50186980e+00 -3.53700638e-01 9.59412038e-01
-4.14350867e-01 -5.28501213e-01 1.16715515e+00 2.31886551e-01
-4.91416931e-01 7.35161543e-01 -4.41850089e-02 7.17810869e-01
5.42319059e-01 -1.34236312e+00 1.36188388e-01 7.66565621e-01
4.36311007e-01 6.95617378e-01 4.65681136e-01 -8.80198181e-01
-1.34223080e+00 -1.35968161e+00 1.03371584e+00 -8.02917600e-01
7.01655328e-01 -2.69743711e-01 -1.23156786e+00 1.09755504e+00
-3.45406175e-01 3.88855845e-01 7.22984433e-01 2.89099310e-02
-7.29916811e-01 -9.44580734e-02 -1.44506216e+00 5.98111331e-01
8.74381423e-01 -5.72953939e-01 -9.35437560e-01 3.20811093e-01
6.15063429e-01 -4.28073317e-01 -9.37083483e-01 2.97085792e-01
1.86312169e-01 -8.20312023e-01 1.08687985e+00 -1.11461222e+00
5.34521461e-01 -2.81375617e-01 -4.23160903e-02 -1.53759980e+00
-1.50113076e-01 -3.88650358e-01 -5.38854063e-01 1.27383173e+00
5.43770432e-01 -8.55679989e-01 7.67647922e-01 7.25563049e-01
8.71007144e-02 -6.53794944e-01 -9.81880784e-01 -7.99018919e-01
-1.92815496e-03 -1.56547368e-01 5.56717753e-01 1.22637510e+00
-1.41517058e-01 7.35587060e-01 -2.04278499e-01 2.05707386e-01
6.55558109e-01 1.66379750e-01 6.99498713e-01 -1.33197534e+00
-1.20166376e-01 -1.17230996e-01 -8.10284734e-01 -3.87584835e-01
4.12468106e-01 -1.54614687e+00 -4.02484983e-01 -1.61497664e+00
-2.23637834e-01 -6.98578954e-01 -2.30566323e-01 1.02075708e+00
-2.59299427e-01 3.20145816e-01 1.02866389e-01 -1.75148230e-02
-2.97136694e-01 6.61189318e-01 1.21665359e+00 -3.38516802e-01
-6.88017085e-02 6.52018636e-02 -8.80598783e-01 5.93393505e-01
6.69090509e-01 -5.24408996e-01 -6.06782615e-01 1.61112826e-02
5.00140548e-01 1.76367640e-01 7.56872952e-01 -6.77523375e-01
2.64882594e-01 1.21514380e-01 6.01427794e-01 -2.32447937e-01
3.02038312e-01 -9.22286153e-01 3.78114492e-01 3.47343236e-01
-1.50741369e-01 -3.53970438e-01 -4.24641632e-02 6.89710855e-01
-3.84797782e-01 -4.47273254e-01 5.14970303e-01 -4.97140922e-02
-6.04072452e-01 2.69095540e-01 3.61763060e-01 2.45524243e-01
1.20006084e+00 -8.57573971e-02 -5.26801407e-01 -3.93110961e-01
-1.13857412e+00 2.92554438e-01 2.36066446e-01 5.74622273e-01
7.52203643e-01 -1.66652787e+00 -7.12455034e-01 3.18358600e-01
8.90045762e-02 2.03538775e-01 2.97780428e-02 3.11167389e-01
-4.65683281e-01 3.15832555e-01 -1.84089452e-01 -2.39156544e-01
-9.84876692e-01 8.13265264e-01 3.30525488e-01 -4.45881993e-01
-4.30521905e-01 9.09583271e-01 -1.25494108e-01 -7.22092271e-01
7.17089549e-02 -7.28721172e-02 -3.72408658e-01 1.76752046e-01
3.15026045e-01 3.72762918e-01 3.05674404e-01 -5.91509461e-01
-3.57190460e-01 2.17603594e-02 -5.26065901e-02 2.84385681e-01
1.40202236e+00 5.06134033e-01 -2.46636048e-01 2.37334386e-01
1.15249860e+00 4.87020388e-02 -6.09196186e-01 -4.20123756e-01
-2.90764093e-01 -2.83373028e-01 -2.05396518e-01 -8.72517109e-01
-1.16573334e+00 6.07393086e-01 1.68981612e-01 4.64443445e-01
6.41225636e-01 2.14044824e-01 4.66865897e-01 6.24823928e-01
4.66377199e-01 -6.14826858e-01 3.42616700e-02 4.09153163e-01
1.09499609e+00 -1.17778718e+00 -1.67265028e-01 -6.89657629e-01
-6.89961135e-01 1.11696696e+00 5.20124078e-01 -1.44306615e-01
4.97509629e-01 -1.59715235e-01 -1.66250885e-01 -3.41851950e-01
-5.19512415e-01 -1.95465758e-01 5.82921445e-01 1.02955532e+00
-1.15584880e-01 1.62041664e-01 1.08345881e-01 6.28856182e-01
-3.92104179e-01 -2.10643426e-01 3.78884465e-01 6.63675249e-01
-1.84191301e-01 -1.38749862e+00 -3.53927433e-01 5.87526500e-01
-1.72666728e-01 -7.94987082e-02 -6.38090074e-01 8.71030271e-01
2.50106663e-01 7.66050518e-01 -2.71058232e-01 -7.00851262e-01
5.52280307e-01 3.77068728e-01 4.43187714e-01 -7.01638460e-01
-6.12351179e-01 -8.17596734e-01 3.22423100e-01 -4.51568812e-01
-1.83983132e-01 -3.25486273e-01 -1.36969626e+00 -3.77511472e-01
-2.15704486e-01 2.01342240e-01 2.77884841e-01 9.50874329e-01
9.58508924e-02 7.71871626e-01 4.91099447e-01 -2.86973894e-01
-6.06096506e-01 -6.34966493e-01 -5.44577837e-01 8.64013672e-01
1.11414120e-02 -8.54786813e-01 -3.56754303e-01 6.02865033e-02] | [8.698064804077148, 7.884810924530029] |
2510d729-f9d5-4446-9d47-bd6a054473f3 | federated-semi-supervised-medical-image | 2106.08600 | null | https://arxiv.org/abs/2106.08600v1 | https://arxiv.org/pdf/2106.08600v1.pdf | Federated Semi-supervised Medical Image Classification via Inter-client Relation Matching | Federated learning (FL) has emerged with increasing popularity to collaborate distributed medical institutions for training deep networks. However, despite existing FL algorithms only allow the supervised training setting, most hospitals in realistic usually cannot afford the intricate data labeling due to absence of budget or expertise. This paper studies a practical yet challenging FL problem, named \textit{Federated Semi-supervised Learning} (FSSL), which aims to learn a federated model by jointly utilizing the data from both labeled and unlabeled clients (i.e., hospitals). We present a novel approach for this problem, which improves over traditional consistency regularization mechanism with a new inter-client relation matching scheme. The proposed learning scheme explicitly connects the learning across labeled and unlabeled clients by aligning their extracted disease relationships, thereby mitigating the deficiency of task knowledge at unlabeled clients and promoting discriminative information from unlabeled samples. We validate our method on two large-scale medical image classification datasets. The effectiveness of our method has been demonstrated with the clear improvements over state-of-the-arts as well as the thorough ablation analysis on both tasks\footnote{Code will be made available at \url{https://github.com/liuquande/FedIRM}}. | ['Pheng-Ann Heng', 'Qi Dou', 'Hongzheng Yang', 'Quande Liu'] | 2021-06-16 | null | null | null | null | ['semi-supervised-medical-image-classification'] | ['medical'] | [ 3.17375474e-02 3.83512437e-01 -6.54141486e-01 -7.75188446e-01
-1.22689879e+00 -2.49824584e-01 9.72726643e-02 -2.23173514e-01
-2.82580584e-01 9.33743358e-01 1.56668350e-01 -2.77810931e-01
-4.01642382e-01 -3.14947337e-01 -7.21832752e-01 -9.56325233e-01
4.67539206e-03 5.38755655e-01 -4.98037636e-01 4.26982492e-01
-5.57749152e-01 3.45098436e-01 -9.17559922e-01 4.85221714e-01
8.77579153e-01 1.10968697e+00 1.39978155e-01 1.07491888e-01
-6.67511746e-02 1.33864260e+00 4.43334207e-02 -5.68155289e-01
4.69589949e-01 -4.16613907e-01 -8.83032858e-01 2.00880483e-01
6.05278850e-01 -4.37805057e-01 -6.95169091e-01 1.07812572e+00
5.28654277e-01 -1.22626200e-01 2.68449783e-01 -1.25439715e+00
-7.19910860e-01 6.18999958e-01 -5.44847667e-01 3.92841212e-02
-3.51713002e-01 9.55514386e-02 9.67316628e-01 -9.29712892e-01
8.00948322e-01 6.89453721e-01 6.36082828e-01 8.50629985e-01
-9.34650004e-01 -1.14755917e+00 1.34167641e-01 7.26700425e-02
-1.28329360e+00 -3.93615752e-01 8.54489088e-01 -3.30871761e-01
4.19678867e-01 2.80116737e-01 1.56976879e-01 1.22038138e+00
-1.59810215e-01 1.31275523e+00 1.07467520e+00 -2.82893479e-01
-1.01098120e-01 2.73647130e-01 3.82310480e-01 1.11732471e+00
4.03469205e-01 1.91223398e-02 -5.88280022e-01 -4.62038547e-01
5.53915441e-01 5.70324898e-01 -3.47592622e-01 -5.30259192e-01
-1.08196008e+00 6.80024743e-01 5.78965962e-01 3.11784595e-01
-2.80930370e-01 -3.59295644e-02 4.26068634e-01 9.29185748e-02
5.42474806e-01 -1.49165392e-01 -7.27502584e-01 4.10253704e-01
-9.19483662e-01 -1.06072389e-01 6.51930034e-01 1.08367217e+00
7.76145697e-01 -2.01542974e-01 -2.88098335e-01 8.75945807e-01
3.69273454e-01 4.49157983e-01 4.78201687e-01 -1.01656032e+00
6.04585409e-01 8.65196288e-01 -2.32071847e-01 -6.30681634e-01
-3.27789217e-01 -7.10856616e-01 -1.35497522e+00 -2.64904946e-01
3.71013194e-01 -4.67648894e-01 -7.01977730e-01 1.89511168e+00
5.58811069e-01 5.80451608e-01 -1.20903134e-01 7.46588588e-01
9.00023580e-01 -2.01930944e-02 2.17389036e-02 -2.89809167e-01
1.27536428e+00 -1.23638189e+00 -8.92323375e-01 -1.16073117e-01
9.06520903e-01 -3.97162676e-01 7.07958817e-01 3.85371387e-01
-8.02824378e-01 -7.59257153e-02 -7.12769210e-01 5.38733415e-02
-7.74237467e-03 3.68831724e-01 8.50876033e-01 4.53320205e-01
-7.98520505e-01 4.34175014e-01 -1.07904589e+00 -1.19251214e-01
1.14264214e+00 5.32487094e-01 -6.22149944e-01 -6.29266918e-01
-9.74779606e-01 4.16945457e-01 -1.05458923e-01 1.77804917e-01
-8.44864666e-01 -9.02428746e-01 -7.90169239e-01 -9.90068689e-02
5.21390915e-01 -7.07547188e-01 1.26036251e+00 -1.09019041e+00
-7.29561090e-01 1.16229224e+00 -9.88532975e-02 -3.58866572e-01
7.00804830e-01 -2.21995816e-01 -2.67166704e-01 5.14998324e-02
1.57491118e-01 4.91043866e-01 5.24752557e-01 -1.23913431e+00
-5.77690542e-01 -6.66675687e-01 -3.56933415e-01 -9.06235650e-02
-7.48111725e-01 -1.80703267e-01 -6.38033807e-01 -8.03179622e-01
2.89251264e-02 -9.72681463e-01 -3.99530113e-01 3.47992063e-01
-5.22091568e-01 -1.75778896e-01 9.72962677e-01 -6.13756001e-01
1.16159821e+00 -2.13953781e+00 -2.77675778e-01 1.95276722e-01
4.96459842e-01 2.05781832e-01 -4.35647853e-02 1.03926510e-01
-2.59429276e-01 -1.50862440e-01 -4.15592313e-01 -7.71641076e-01
-1.44841880e-01 4.30574030e-01 -4.81062867e-02 8.09648812e-01
-3.77477109e-02 9.85876083e-01 -8.83699298e-01 -9.20356870e-01
-1.57984465e-01 5.91163874e-01 -5.28995335e-01 4.73034948e-01
4.71827872e-02 8.49700332e-01 -8.46705377e-01 1.05822647e+00
6.30071998e-01 -9.37441647e-01 5.92272580e-01 -2.66646504e-01
4.45726126e-01 -7.29056969e-02 -1.01645422e+00 2.05693698e+00
-3.77881795e-01 8.06552917e-02 4.03229237e-01 -1.18569195e+00
6.83521628e-01 7.48100162e-01 1.09695852e+00 -5.95028937e-01
1.57284856e-01 3.56232107e-01 -3.63038480e-01 -6.48879349e-01
-2.64359921e-01 -1.70228496e-01 1.41991481e-01 5.39304495e-01
2.41630137e-01 6.62501991e-01 -2.79858559e-01 3.78562957e-01
1.16688728e+00 9.88095105e-02 1.91160396e-01 -1.85466811e-01
4.18334007e-01 -2.91328449e-02 8.57520640e-01 6.85156345e-01
-5.17498851e-01 4.55915362e-01 1.74346894e-01 -4.53306288e-01
-5.85131884e-01 -8.69163573e-01 -3.11241418e-01 9.43760872e-01
-7.44236335e-02 -3.67758498e-02 -5.39147258e-01 -1.41474330e+00
2.54132718e-01 1.85832739e-01 -8.08038831e-01 -8.31430778e-02
-6.74187183e-01 -7.83073127e-01 5.62125146e-01 5.39253891e-01
3.93749863e-01 -8.48745525e-01 -2.60231763e-01 -1.56538919e-01
-3.15022826e-01 -1.06608450e+00 -6.38121247e-01 2.71341234e-01
-1.03972447e+00 -1.34450591e+00 -8.13951910e-01 -9.63445842e-01
1.07581365e+00 2.09731445e-01 1.05884671e+00 2.27966160e-01
-5.69230735e-01 3.72228235e-01 -9.69910845e-02 -2.04640776e-01
-8.24520513e-02 -7.06967935e-02 -2.55203601e-02 3.49419206e-01
4.45862710e-01 -3.66673052e-01 -9.96389925e-01 3.27311337e-01
-9.27600145e-01 4.34487984e-02 7.59179056e-01 1.31232417e+00
7.49580681e-01 -4.43252802e-01 7.69125938e-01 -1.66472197e+00
1.63657561e-01 -8.35357070e-01 -2.80205160e-01 5.35343111e-01
-8.33367467e-01 2.96777971e-02 5.15997350e-01 -1.99351102e-01
-1.20370078e+00 4.41422284e-01 3.26144621e-02 -9.25147533e-01
-1.42365903e-01 3.42761934e-01 -1.77174866e-01 -4.93463986e-02
3.51775914e-01 2.62335222e-02 1.82302803e-01 -7.14258850e-01
1.96916357e-01 9.44619656e-01 5.36114454e-01 -7.54815757e-01
4.68036026e-01 8.15036118e-01 -1.63466901e-01 -5.51844016e-03
-1.28916669e+00 -7.43801177e-01 -4.14165199e-01 -5.39175496e-02
5.40566802e-01 -1.17348647e+00 -6.68904901e-01 1.88914508e-01
-8.31555188e-01 -2.56037951e-01 -3.55225384e-01 7.03216612e-01
-3.84783119e-01 2.15616554e-01 -8.67526114e-01 -5.17525196e-01
-6.38214946e-01 -1.07182014e+00 9.21145737e-01 7.82284420e-03
9.93338823e-02 -1.07580447e+00 2.33852401e-01 8.58810365e-01
3.39666814e-01 3.77366364e-01 7.70949185e-01 -9.60007727e-01
-6.80863798e-01 -2.76244313e-01 -3.06496054e-01 4.77596372e-01
5.00235498e-01 -4.10328567e-01 -1.14686525e+00 -5.93410552e-01
1.86434418e-01 -7.22018301e-01 8.00331354e-01 2.84587801e-01
1.51152778e+00 -4.24009919e-01 -6.12491131e-01 8.41976583e-01
1.57632971e+00 -1.84929699e-01 1.24949105e-01 4.58059199e-02
8.36803436e-01 6.33880615e-01 5.05979657e-01 6.55687034e-01
3.74469548e-01 3.09817851e-01 4.60946798e-01 -4.81292099e-01
-2.29291514e-01 -1.04995258e-01 -2.08096951e-01 8.41943264e-01
2.04109609e-01 1.00724194e-02 -1.03610277e+00 6.18382037e-01
-2.20889902e+00 -5.29715359e-01 1.61952749e-01 1.98135221e+00
9.65835750e-01 -3.72092903e-01 -2.10649326e-01 -3.80641371e-01
6.45810366e-01 9.69797745e-03 -9.03584003e-01 3.28308880e-01
-7.06045702e-02 2.68847704e-01 5.85656703e-01 1.13515452e-01
-1.18680966e+00 3.57631981e-01 5.05305672e+00 7.63131738e-01
-1.06710410e+00 5.84615588e-01 1.16290212e+00 -3.47859323e-01
-6.60618246e-02 -3.43617141e-01 -6.83640182e-01 2.53134549e-01
6.47198200e-01 -4.10354137e-02 3.86668108e-02 9.22556043e-01
1.52136862e-01 4.04070050e-01 -1.30228591e+00 1.12061071e+00
3.55158113e-02 -1.56351924e+00 -2.89707519e-02 3.71416658e-03
9.76431608e-01 3.19936335e-01 2.31303141e-01 1.84508458e-01
6.11317396e-01 -9.97456551e-01 2.34537899e-01 5.27158916e-01
1.10427868e+00 -4.01495039e-01 7.07391918e-01 4.41236258e-01
-1.00926137e+00 -2.40061522e-01 8.11699852e-02 4.95613486e-01
-1.92288488e-01 6.03931725e-01 -5.11257708e-01 9.73355353e-01
8.60900760e-01 9.16357219e-01 -2.56284833e-01 8.38667214e-01
1.01972155e-01 7.93592393e-01 -8.34150836e-02 5.61790824e-01
1.93483531e-01 -1.21856235e-01 1.31317275e-03 1.16770434e+00
8.52501318e-02 1.58476755e-01 3.03740144e-01 6.27797306e-01
-5.65988064e-01 1.92285597e-01 -5.66187084e-01 1.76489368e-01
3.60336512e-01 1.47820807e+00 -2.98206240e-01 -2.95039952e-01
-7.48417437e-01 6.20478511e-01 6.96943581e-01 3.86951387e-01
-8.78107727e-01 1.41164467e-01 5.21307826e-01 -4.53984849e-02
1.02845877e-01 3.28574598e-01 -3.24337333e-01 -1.26547694e+00
4.37437296e-02 -1.03403282e+00 9.60328102e-01 -2.17548892e-01
-1.79510057e+00 6.93050027e-01 -2.85009623e-01 -1.37746263e+00
1.18151000e-02 -2.83056855e-01 -1.82250470e-01 7.41618872e-01
-1.69195807e+00 -1.54564929e+00 -3.46827239e-01 1.19525576e+00
2.62742341e-01 -2.94106096e-01 8.81331503e-01 9.12669241e-01
-9.36346650e-01 9.49084580e-01 4.45326865e-01 5.42774916e-01
1.05614066e+00 -8.48935187e-01 -4.58376855e-01 6.18418336e-01
5.25294757e-03 6.45666778e-01 7.63787031e-02 -5.38647056e-01
-1.54627490e+00 -1.51368415e+00 7.90745974e-01 -3.25493425e-01
5.32277346e-01 -3.23456705e-01 -8.40144336e-01 1.06719589e+00
1.38751686e-01 7.54146755e-01 1.15239012e+00 4.57147695e-02
-5.28915942e-01 -3.98216277e-01 -1.42465341e+00 1.82389021e-01
1.07257259e+00 -4.21591133e-01 -2.70462215e-01 8.67627859e-01
6.51249826e-01 -3.66123825e-01 -8.63366663e-01 6.34858131e-01
3.48146617e-01 -6.84964001e-01 7.34621406e-01 -9.50774789e-01
3.27465773e-01 2.70640813e-02 -2.75075614e-01 -6.09586120e-01
-2.33997688e-01 -5.72785199e-01 -4.12168026e-01 1.14299142e+00
3.31376374e-01 -7.05636680e-01 1.33947873e+00 9.01208639e-01
-8.95511061e-02 -1.32323182e+00 -9.75190222e-01 -6.05006874e-01
-4.08857688e-02 -2.69226372e-01 3.98779094e-01 1.50044894e+00
5.53872157e-03 1.89893208e-02 -5.28098822e-01 2.21218541e-01
9.81265247e-01 3.46075922e-01 4.12181258e-01 -8.84079039e-01
-5.58548748e-01 -2.41826829e-02 1.68528378e-01 -8.13192964e-01
2.99473047e-01 -1.27789974e+00 -1.58465415e-01 -1.42416322e+00
7.20410883e-01 -9.97618198e-01 -7.84921885e-01 1.03625321e+00
-1.93494424e-01 2.58629769e-01 -4.14698236e-02 5.60882568e-01
-9.51748312e-01 3.93538713e-01 1.34781837e+00 -2.59505093e-01
2.12904677e-01 6.30043913e-03 -9.00465071e-01 4.70860600e-01
6.02556407e-01 -7.47029662e-01 -4.65982080e-01 -6.34522796e-01
-2.54556388e-01 2.98836142e-01 3.08079839e-01 -6.23400450e-01
3.94844741e-01 -7.97915738e-03 3.28454226e-01 -2.41673872e-01
-1.31104039e-02 -1.16627562e+00 3.22316289e-01 6.22847140e-01
-4.45180207e-01 -1.66598052e-01 3.42587195e-02 7.35537410e-01
-3.64927739e-01 1.13481320e-01 7.70571768e-01 -2.44033098e-01
-2.83955485e-01 9.08608556e-01 4.11813140e-01 1.68024823e-01
1.02839363e+00 2.62750447e-01 -4.96098995e-01 -7.35891610e-02
-8.60652030e-01 6.34622097e-01 1.94745526e-01 9.92272273e-02
5.30691981e-01 -1.26874661e+00 -8.56658995e-01 1.72896430e-01
8.32019821e-02 3.39268208e-01 6.83821619e-01 1.26965714e+00
-2.13522747e-01 5.36982656e-01 1.31709009e-01 -6.11424148e-01
-1.34647977e+00 7.24648535e-01 5.28708279e-01 -7.45016932e-01
-6.63691938e-01 1.01034689e+00 4.43162173e-01 -6.67107880e-01
6.56241596e-01 6.64502233e-02 1.87427774e-01 -2.44741574e-01
3.65175337e-01 1.36434883e-01 3.11852247e-01 -3.67282450e-01
-6.31048024e-01 1.50886714e-01 -4.53629106e-01 6.06447697e-01
1.57773936e+00 8.80948007e-02 -1.18178956e-01 1.28184110e-01
1.48033059e+00 -2.94141799e-01 -1.37485075e+00 -7.32807875e-01
-5.31005077e-02 -4.15552258e-01 -1.41982585e-01 -7.83519387e-01
-1.71112001e+00 5.96213341e-01 7.71857560e-01 -5.23393512e-01
1.20804143e+00 2.29121327e-01 1.03742278e+00 3.29137802e-01
3.90632659e-01 -6.18846834e-01 1.51633546e-01 -3.85412015e-02
4.54373389e-01 -1.60718262e+00 -5.32365590e-02 -3.10311824e-01
-6.18963420e-01 6.43044472e-01 5.58903873e-01 -1.67122204e-02
9.76188302e-01 4.80481535e-01 2.90997207e-01 -3.47531438e-01
-9.45525229e-01 2.59431183e-01 1.08736649e-01 1.99479476e-01
4.99307573e-01 -2.26848032e-02 -1.47160605e-01 8.17302227e-01
5.19433022e-01 4.44539845e-01 -6.73828274e-02 1.11121273e+00
1.79356396e-01 -1.31063378e+00 -1.57705694e-01 8.64328504e-01
-8.65011454e-01 -8.84958506e-02 -1.39770299e-01 6.21041536e-01
3.03222656e-01 6.43281162e-01 -2.58863300e-01 -9.90603566e-02
1.87520504e-01 3.56813660e-03 3.42195451e-01 -6.00269616e-01
-6.61171317e-01 1.08345620e-01 -2.20761284e-01 -7.15962708e-01
-6.04525864e-01 -6.35413826e-01 -1.37318075e+00 3.83643135e-02
-3.11658233e-01 2.38004833e-01 2.90937513e-01 8.05839479e-01
4.74076062e-01 4.19867009e-01 8.67608666e-01 -2.05092922e-01
-7.51215816e-01 -5.92944205e-01 -6.99047208e-01 7.09505558e-01
4.32571679e-01 -4.12209213e-01 -2.31228352e-01 1.17645800e-01] | [6.037525177001953, 6.442866325378418] |
5ff612ca-e6be-49fe-a0a4-e2558c94450f | communicative-subgraph-representation | 2205.05957 | null | https://arxiv.org/abs/2205.05957v1 | https://arxiv.org/pdf/2205.05957v1.pdf | Communicative Subgraph Representation Learning for Multi-Relational Inductive Drug-Gene Interaction Prediction | Illuminating the interconnections between drugs and genes is an important topic in drug development and precision medicine. Currently, computational predictions of drug-gene interactions mainly focus on the binding interactions without considering other relation types like agonist, antagonist, etc. In addition, existing methods either heavily rely on high-quality domain features or are intrinsically transductive, which limits the capacity of models to generalize to drugs/genes that lack external information or are unseen during the training process. To address these problems, we propose a novel Communicative Subgraph representation learning for Multi-relational Inductive drug-Gene interactions prediction (CoSMIG), where the predictions of drug-gene relations are made through subgraph patterns, and thus are naturally inductive for unseen drugs/genes without retraining or utilizing external domain features. Moreover, the model strengthened the relations on the drug-gene graph through a communicative message passing mechanism. To evaluate our method, we compiled two new benchmark datasets from DrugBank and DGIdb. The comprehensive experiments on the two datasets showed that our method outperformed state-of-the-art baselines in the transductive scenarios and achieved superior performance in the inductive ones. Further experimental analysis including LINCS experimental validation and literature verification also demonstrated the value of our model. | ['Yuedong Yang', 'Sijie Mai', 'Shuangjia Zheng', 'Jiahua Rao'] | 2022-05-12 | null | null | null | null | ['gene-interaction-prediction'] | ['graphs'] | [ 2.99008429e-01 2.28285059e-01 -5.90280175e-01 -3.57605994e-01
-4.48175132e-01 -3.02555412e-01 4.75479543e-01 4.49399650e-01
1.30789950e-01 1.21123576e+00 2.34155551e-01 -4.94142145e-01
-5.83765745e-01 -1.06022930e+00 -9.92421865e-01 -8.97344470e-01
-2.54777104e-01 6.44344091e-01 1.19921245e-01 -3.65788609e-01
-3.27770323e-01 1.82048053e-01 -8.86604369e-01 4.03354257e-01
1.01141715e+00 3.01799446e-01 -1.36320248e-01 6.50489926e-02
2.25073230e-02 8.58801782e-01 -9.38460231e-02 -3.93019527e-01
-1.87970877e-01 -6.29777670e-01 -8.42340469e-01 -3.12326789e-01
-5.23205161e-01 8.25370327e-02 -5.32199025e-01 8.51453304e-01
6.42119646e-01 -2.96609938e-01 6.38698936e-01 -1.08314443e+00
-6.63084626e-01 8.29765499e-01 -4.04713988e-01 -5.67489117e-02
7.05350459e-01 1.24585077e-01 1.26849461e+00 -8.42947364e-01
9.26464736e-01 1.35625172e+00 6.31594360e-01 6.11162305e-01
-1.22475553e+00 -8.97714615e-01 1.37295842e-01 1.78896293e-01
-1.54013515e+00 -1.95076540e-01 5.31839967e-01 -4.67564940e-01
1.00788939e+00 2.57965386e-01 5.67566216e-01 1.32290125e+00
2.23821059e-01 5.98372817e-01 7.88461685e-01 -5.44235520e-02
-2.92783836e-03 3.94282304e-02 3.18266861e-02 6.72791898e-01
1.97100565e-01 3.00657779e-01 -3.07784170e-01 -6.93540215e-01
4.62639570e-01 1.77217960e-01 -5.09690046e-01 -1.20366409e-01
-1.19198000e+00 8.33523214e-01 6.48056984e-01 4.07334358e-01
-2.55670667e-01 -2.21825689e-01 2.96706170e-01 1.56865641e-01
6.31395757e-01 2.65038699e-01 -6.90504014e-01 2.91300267e-01
-1.77445918e-01 -1.70131817e-01 7.95022249e-01 9.56628859e-01
6.51759386e-01 -5.30390263e-01 -5.66281788e-02 8.03467810e-01
5.57051539e-01 -4.22165208e-02 3.33031356e-01 2.58236021e-01
3.58761877e-01 9.98290360e-01 -2.53147960e-01 -1.04097271e+00
-6.72996581e-01 -4.84094918e-01 -1.16476679e+00 -7.04179466e-01
1.05249658e-01 -1.97518751e-01 -7.49765694e-01 1.82356000e+00
6.92512333e-01 8.73219013e-01 1.78272501e-01 6.59456193e-01
1.41791070e+00 5.65346777e-01 6.25645697e-01 -4.71685797e-01
1.20214868e+00 -4.40550864e-01 -7.74394512e-01 2.05466822e-01
1.28436077e+00 -5.13187528e-01 6.40797019e-01 1.26495108e-01
-5.28467178e-01 -2.05904767e-01 -8.06553364e-01 1.98547468e-01
-4.78059977e-01 -2.08395869e-01 1.06917870e+00 1.77196965e-01
-5.08598804e-01 8.09356868e-01 -5.49596429e-01 -3.05202127e-01
6.64752781e-01 7.85934567e-01 -5.55304527e-01 -1.84631899e-01
-1.89312720e+00 6.96110368e-01 4.63123441e-01 1.39072984e-01
-9.05161917e-01 -9.54723775e-01 -6.12790048e-01 -1.29378989e-01
2.25193173e-01 -9.07416999e-01 5.03560245e-01 -5.62749565e-01
-1.36956108e+00 7.27430284e-01 8.89519677e-02 -1.88708618e-01
1.15143038e-01 4.71979147e-03 -7.50950813e-01 -1.77638635e-01
-9.87026393e-02 3.46488267e-01 -1.01754973e-02 -8.16942096e-01
-1.64704397e-01 -4.81081367e-01 1.57413363e-01 1.33238554e-01
-3.46980542e-01 -2.76575685e-01 -3.82311702e-01 -4.39542383e-01
-1.45754337e-01 -9.60851431e-01 -4.15504307e-01 -4.93361622e-01
-8.84140491e-01 -5.11240900e-01 5.50469279e-01 -1.40632644e-01
1.41232109e+00 -2.00101900e+00 4.49794203e-01 3.85085016e-01
4.11717534e-01 3.56501520e-01 -2.95452118e-01 8.34547400e-01
-7.11466551e-01 2.62563407e-01 -6.30889311e-02 4.60242033e-01
-4.28624928e-01 3.69597912e-01 -2.18141854e-01 6.16871834e-01
2.05473393e-01 1.07821023e+00 -1.15817726e+00 -6.23679876e-01
-5.41631617e-02 6.22291267e-01 -6.47737622e-01 2.07983002e-01
-5.66706479e-01 9.96064305e-01 -1.07834840e+00 7.51019061e-01
5.43130755e-01 -7.36872911e-01 7.22952604e-01 -3.47635597e-01
4.72813606e-01 3.63005936e-01 -7.09266007e-01 1.63111961e+00
-8.90694037e-02 -3.36969107e-01 -6.25178099e-01 -1.19063187e+00
8.21075320e-01 5.92201531e-01 8.83053720e-01 -4.34706569e-01
1.29173264e-01 1.61391914e-01 3.25316340e-01 -6.22146845e-01
-5.13490677e-01 -3.75348389e-01 1.54493496e-01 -2.97965817e-02
-2.25355104e-02 2.38925308e-01 -1.60496935e-01 3.53147030e-01
1.25055909e+00 1.57092005e-01 5.40425003e-01 -1.81522951e-01
8.17931294e-01 -3.97884957e-02 8.45770359e-01 2.04009041e-01
2.83558458e-01 -7.03433529e-02 7.63913512e-01 -3.37361813e-01
-4.14705813e-01 -6.22442484e-01 -6.62591457e-01 8.51467013e-01
2.75729209e-01 -8.29066277e-01 -4.97948438e-01 -9.48656499e-01
2.07869872e-03 3.02129716e-01 -7.69608498e-01 -4.44956243e-01
-2.33318970e-01 -1.51403797e+00 5.90116382e-01 4.65888232e-02
8.07014778e-02 -6.10552251e-01 8.47073734e-01 4.74047065e-01
-1.06571145e-01 -1.09450650e+00 -2.06049293e-01 7.72189200e-02
-8.40412676e-01 -1.62070191e+00 -2.99497634e-01 -8.30812693e-01
8.38703990e-01 -2.08647162e-01 1.01502395e+00 1.59760296e-01
-3.01574349e-01 -9.01931450e-02 -2.97672987e-01 -3.31277311e-01
-4.35444683e-01 -6.35050684e-02 5.49786910e-02 6.95343912e-02
6.58953786e-01 -8.03020120e-01 -7.04847634e-01 6.48926437e-01
-9.30032492e-01 2.08888829e-01 5.49182355e-01 1.19325519e+00
7.17992425e-01 -9.35185477e-02 1.04397142e+00 -1.58554828e+00
5.82358003e-01 -1.08217776e+00 -3.94364387e-01 3.06429863e-01
-5.06591678e-01 6.89573213e-02 6.64511621e-01 -4.64881629e-01
-8.42193246e-01 3.77369106e-01 -4.70721364e-01 1.36222214e-01
3.45345661e-02 9.10448849e-01 -5.66308677e-01 -6.61362484e-02
7.37966418e-01 1.21560311e-02 -2.43577629e-01 -3.95086676e-01
2.80556083e-01 6.91252649e-01 -1.85855523e-01 -6.02777243e-01
4.19462800e-01 1.60876036e-01 4.04910266e-01 -4.23202395e-01
-7.41115689e-01 -3.50061923e-01 -3.45824540e-01 2.92593062e-01
6.66496217e-01 -1.10087383e+00 -1.15793073e+00 2.12071076e-01
-1.05999792e+00 6.23571761e-02 1.94362134e-01 8.00627053e-01
-1.63537085e-01 4.53474194e-01 -8.36408317e-01 -2.59082586e-01
-2.92638868e-01 -1.31685591e+00 1.00594068e+00 -1.84574366e-01
-8.67744461e-02 -1.28328156e+00 4.37852472e-01 1.87780842e-01
-6.42629489e-02 4.34165716e-01 1.49825656e+00 -1.11932516e+00
-5.35196245e-01 -1.87903747e-01 -8.82400796e-02 -1.85145870e-01
5.74594438e-01 -2.25773677e-01 -7.77732790e-01 -1.18411526e-01
-4.92280483e-01 -4.33183700e-01 7.00249374e-01 9.11830068e-02
1.26715481e+00 -3.81534874e-01 -1.04748166e+00 4.28406715e-01
1.29453957e+00 2.91012138e-01 8.54713142e-01 -2.67067194e-01
8.65259290e-01 4.15390670e-01 6.07216477e-01 5.32330692e-01
2.83737928e-01 8.17535758e-01 4.95075732e-01 -4.42517102e-01
2.60486215e-01 -5.34139752e-01 1.49568513e-01 9.26218033e-01
-2.55995005e-01 -5.18480122e-01 -8.42418969e-01 3.08016598e-01
-2.07427645e+00 -5.16332269e-01 -4.98640865e-01 1.97952902e+00
1.57463384e+00 -1.46429554e-01 -1.18969753e-01 -2.85457641e-01
7.41803110e-01 -3.64530504e-01 -6.72391951e-01 6.18226379e-02
-2.80959904e-01 4.10028487e-01 1.04184389e-01 5.01529396e-01
-1.00905323e+00 1.14815247e+00 6.03127003e+00 1.11526191e+00
-7.62454689e-01 -1.53793961e-01 9.04119253e-01 2.74439216e-01
-4.81594086e-01 9.26526356e-03 -8.77233505e-01 4.14897084e-01
8.85679185e-01 -1.46904871e-01 1.16952799e-01 5.86629272e-01
1.54664800e-01 3.43497068e-01 -1.50571048e+00 6.87995613e-01
-4.81859297e-01 -1.50455558e+00 2.48235986e-01 2.25120604e-01
6.32873535e-01 1.40600160e-01 -2.03846619e-01 1.77881837e-01
6.84720933e-01 -1.29589128e+00 -6.12890720e-01 6.01494610e-01
6.79516137e-01 -5.28093159e-01 7.96536624e-01 2.61424124e-01
-9.98324096e-01 2.28928670e-01 -3.95141035e-01 1.64021909e-01
-2.04328582e-01 9.10863042e-01 -1.24951482e+00 1.06329906e+00
1.06971204e-01 1.44113433e+00 -2.89422244e-01 8.45999599e-01
-2.92630702e-01 6.51369989e-01 -1.58402950e-01 -1.37091100e-01
1.44166440e-01 -3.08699876e-01 2.18295589e-01 1.15612233e+00
7.75521174e-02 5.63193679e-01 2.29043052e-01 6.42907381e-01
-3.59643728e-01 5.89254379e-01 -9.05797720e-01 -2.28887513e-01
1.56760320e-01 1.08717537e+00 -1.52061477e-01 -2.06601411e-01
-5.07262647e-01 4.77685720e-01 2.60251969e-01 4.62313473e-01
-1.10684502e+00 -1.41279981e-01 6.46754146e-01 2.07619444e-01
-2.36123994e-01 4.61537838e-01 2.04219177e-01 -1.15492237e+00
-4.65663373e-01 -9.17701662e-01 7.10670650e-01 -3.90226573e-01
-1.64650559e+00 6.25838876e-01 -1.60993114e-01 -1.48509312e+00
-8.03501755e-02 -4.79206324e-01 -1.37277678e-01 6.47923708e-01
-1.48172390e+00 -1.24293470e+00 1.86822981e-01 9.16097403e-01
-5.39200753e-02 -1.63321406e-01 1.22604918e+00 7.89749086e-01
-7.92467892e-01 5.25445163e-01 8.73674452e-02 -5.19944448e-03
8.09434533e-01 -7.45610118e-01 -5.88495247e-02 -1.81566421e-02
-5.14390655e-02 8.62030566e-01 4.34972286e-01 -8.75367761e-01
-1.54142833e+00 -1.37539768e+00 8.33243430e-01 -3.27780753e-01
8.91636729e-01 -3.72523457e-01 -1.02337909e+00 7.40070701e-01
1.98538620e-02 2.63576835e-01 1.39018047e+00 4.76700813e-01
-2.60569543e-01 -4.02416289e-02 -9.78751898e-01 5.02171218e-01
1.42729592e+00 -2.98694640e-01 -3.66892546e-01 1.15228951e+00
7.41611838e-01 -4.00675952e-01 -1.32769895e+00 7.63735712e-01
1.46461949e-01 -4.01739001e-01 1.19292140e+00 -1.49481320e+00
5.25582254e-01 -4.55175102e-01 1.42298967e-01 -1.33487594e+00
-5.26128054e-01 -4.45556045e-01 8.34612921e-02 1.01997197e+00
8.10990870e-01 -8.07647586e-01 4.62772191e-01 3.67420495e-01
-5.96507229e-02 -1.14881420e+00 -8.96566272e-01 -3.55180472e-01
-6.13826653e-03 -1.63509578e-01 7.75376499e-01 1.38498926e+00
6.63541377e-01 7.89945662e-01 -6.33774936e-01 2.09089205e-01
1.92759693e-01 1.83982819e-01 5.94395995e-01 -1.29245317e+00
-6.83256388e-01 -6.95973961e-03 -6.38144433e-01 -9.40055132e-01
2.57902503e-01 -1.32713032e+00 -3.17889780e-01 -1.42814589e+00
5.83084762e-01 -6.16575956e-01 -5.36619186e-01 8.82770300e-01
-3.42372388e-01 -1.98037773e-02 -6.40784919e-01 1.03435755e-01
-6.82871819e-01 6.27617002e-01 1.65665376e+00 -3.28930676e-01
-1.96020409e-01 4.79827560e-02 -9.58270252e-01 6.53048933e-01
4.68053311e-01 -6.27194643e-01 -6.85853899e-01 5.32471016e-02
5.30261397e-01 2.63308436e-01 1.30846322e-01 -2.18975112e-01
2.96234965e-01 -3.81206781e-01 2.04954654e-01 -2.33892158e-01
2.27860481e-01 -7.81587601e-01 7.06461787e-01 6.29113793e-01
-3.16949904e-01 -5.65412819e-01 2.80646920e-01 1.00025117e+00
-2.36778185e-01 3.00975412e-01 5.69057167e-01 5.09716533e-02
-1.58556879e-01 7.15833247e-01 2.46124100e-02 2.24816706e-02
1.22256863e+00 2.62527138e-01 -3.21420521e-01 -2.55085975e-02
-9.98325527e-01 2.31912106e-01 -1.36877924e-01 3.66716683e-01
6.28919661e-01 -1.42633474e+00 -8.91090810e-01 2.02348456e-02
3.94723564e-01 -4.71729413e-02 1.94223881e-01 1.09860945e+00
-5.95414974e-02 5.97351909e-01 3.88713121e-01 -6.04926050e-01
-1.26776159e+00 7.59442031e-01 4.35158938e-01 -4.89178956e-01
-3.91434133e-01 8.08484495e-01 8.25940013e-01 -5.12241960e-01
8.17931965e-02 -3.52223702e-02 -3.76080453e-01 -8.56528506e-02
1.86389208e-01 -2.24570796e-01 1.09637216e-01 -5.41509867e-01
-7.43152440e-01 5.39924085e-01 -3.07194412e-01 8.30493748e-01
1.61799288e+00 4.28331286e-01 -4.30393130e-01 8.98678154e-02
1.27746320e+00 -1.70080930e-01 -3.20365161e-01 -4.79019523e-01
6.59810659e-03 -3.12775791e-01 -2.18694761e-01 -9.54088449e-01
-1.00995684e+00 5.95398784e-01 2.51374424e-01 -9.39059630e-02
9.66502964e-01 2.57191211e-01 5.46060920e-01 4.12898690e-01
3.46173584e-01 -5.93295395e-01 -1.87211901e-01 2.22837433e-01
8.89129460e-01 -1.10527587e+00 2.65565246e-01 -1.13151026e+00
-4.76216167e-01 9.52288568e-01 5.28882861e-01 2.79976547e-01
9.44538951e-01 5.46769127e-02 -4.42876160e-01 -5.03724098e-01
-9.25023615e-01 2.00618356e-02 4.33981955e-01 2.85665333e-01
9.02676225e-01 1.04565993e-01 -6.68605089e-01 8.56637836e-01
4.68282670e-01 3.98057550e-01 -5.26432283e-02 4.78139043e-01
6.55966923e-02 -1.73080051e+00 1.60130411e-01 4.10098791e-01
-4.62889761e-01 -5.15631497e-01 -7.10734785e-01 8.06447506e-01
2.35349298e-01 8.89678299e-01 -6.61746919e-01 -6.20006263e-01
3.53115290e-01 -1.62674129e-01 3.38230729e-01 -7.05100179e-01
-5.91054797e-01 3.58281255e-01 3.25769156e-01 -4.51335639e-01
-6.49292886e-01 -2.92904198e-01 -1.56198204e+00 -2.22810701e-01
-7.10513711e-01 4.27128613e-01 5.20719774e-02 1.00355458e+00
8.89632702e-01 7.38780797e-01 8.36919248e-01 7.88635537e-02
-1.87787339e-01 -8.56626272e-01 -6.01522207e-01 4.18927938e-01
-9.83760953e-02 -7.06695199e-01 -5.40077351e-02 -4.26717848e-02] | [5.310478687286377, 5.8765764236450195] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.