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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
9fb1f12e-0b01-4df4-81cb-bad3a26f5d3b | partitioning-guided-k-means-extreme-empty | 2306.14031 | null | https://arxiv.org/abs/2306.14031v1 | https://arxiv.org/pdf/2306.14031v1.pdf | Partitioning-Guided K-Means: Extreme Empty Cluster Resolution for Extreme Model Compression | Compactness in deep learning can be critical to a model's viability in low-resource applications, and a common approach to extreme model compression is quantization. We consider Iterative Product Quantization (iPQ) with Quant-Noise to be state-of-the-art in this area, but this quantization framework suffers from preventable inference quality degradation due to prevalent empty clusters. In this paper, we propose several novel enhancements aiming to improve the accuracy of iPQ with Quant-Noise by focusing on resolving empty clusters. Our contribution, which we call Partitioning-Guided k-means (PG k-means), is a heavily augmented k-means implementation composed of three main components. First, we propose a partitioning-based pre-assignment strategy that ensures no initial empty clusters and encourages an even weight-to-cluster distribution. Second, we propose an empirically superior empty cluster resolution heuristic executed via cautious partitioning of large clusters. Finally, we construct an optional optimization step that consolidates intuitively dense clusters of weights to ensure shared representation. The proposed approach consistently reduces the number of empty clusters in iPQ with Quant-Noise by 100x on average, uses 8x fewer iterations during empty cluster resolution, and improves overall model accuracy by up to 12%, when applied to RoBERTa on a variety of tasks in the GLUE benchmark. | ['Lizhong Chen', 'Victor Agostinelli', 'Tianhong Huang'] | 2023-06-24 | null | null | null | null | ['quantization', 'model-compression'] | ['methodology', 'methodology'] | [ 5.66699505e-02 3.10177393e-02 -2.42929116e-01 -3.20869476e-01
-9.88277018e-01 -1.89704075e-01 5.55163443e-01 4.82977808e-01
-6.77514136e-01 6.03193700e-01 -8.21440574e-03 -2.60806143e-01
-5.22898495e-01 -7.80551076e-01 -8.00571680e-01 -8.97027433e-01
-6.52832612e-02 6.22199953e-01 3.15880895e-01 8.50527138e-02
4.45068717e-01 4.31263000e-01 -1.96477246e+00 3.92364144e-01
9.53623772e-01 1.03581202e+00 5.20773470e-01 5.79267442e-01
-1.26398578e-01 7.12867856e-01 -5.82589805e-01 -4.09361362e-01
2.52563208e-01 -1.73994035e-01 -9.36715007e-01 -1.69320524e-01
4.01258767e-01 -3.43671143e-01 9.78485048e-02 1.18794584e+00
6.13722086e-01 2.25502327e-01 7.37664878e-01 -1.16106057e+00
-1.29382774e-01 1.15852070e+00 -6.50709033e-01 3.03822994e-01
-4.76626396e-01 5.88003732e-02 1.11462522e+00 -7.38697886e-01
3.58331203e-01 1.28438818e+00 7.20420241e-01 2.73238301e-01
-1.49376774e+00 -8.50658059e-01 -1.82180136e-01 4.58882779e-01
-1.98336053e+00 -5.36384523e-01 4.30283666e-01 -3.01449329e-01
1.17282295e+00 3.63878697e-01 6.02120817e-01 5.85331500e-01
2.53713373e-02 5.89432597e-01 7.38041401e-01 -5.84570944e-01
8.78557265e-01 -2.25381143e-02 3.04550409e-01 5.13947666e-01
3.23329151e-01 -1.19854599e-01 -4.48880374e-01 -2.32541949e-01
4.77358937e-01 -1.89722940e-01 -2.44173389e-02 -2.99980760e-01
-8.95490289e-01 1.05451071e+00 4.85051751e-01 3.41965526e-01
-4.82820094e-01 6.44024491e-01 5.48399627e-01 -2.03490764e-01
4.70084608e-01 4.24946815e-01 -2.42319167e-01 -5.38723469e-01
-1.56980753e+00 5.19890010e-01 4.45272505e-01 9.28760052e-01
7.05571294e-01 -1.08525060e-01 -3.32717597e-01 9.76555049e-01
2.33488888e-01 2.02488974e-01 6.84909463e-01 -1.48689508e+00
4.43950057e-01 2.87504017e-01 -1.30863592e-01 -1.01783764e+00
-3.79641086e-01 -7.17677057e-01 -1.24455214e+00 4.37387414e-02
1.04559287e-01 1.57450989e-01 -8.01180482e-01 1.70389247e+00
2.46416584e-01 1.88152865e-01 3.96012440e-02 5.96574306e-01
4.51463670e-01 6.43970490e-01 5.21712080e-02 -2.02248961e-01
1.34026885e+00 -7.33749092e-01 -8.11620891e-01 1.38121784e-01
8.14963102e-01 -4.71437365e-01 1.13853455e+00 8.66966128e-01
-1.31610334e+00 -6.82882428e-01 -1.16557574e+00 -8.57382268e-02
-3.43805939e-01 1.27631530e-01 4.32439685e-01 8.25375438e-01
-1.21386075e+00 9.56088364e-01 -1.08220315e+00 -1.46362722e-01
6.97370112e-01 5.42384267e-01 8.48569125e-02 2.35019461e-03
-1.05647278e+00 6.68491423e-01 9.90946710e-01 -2.86713123e-01
-5.76356471e-01 -1.00115168e+00 -6.24295235e-01 5.53577542e-01
3.42505634e-01 -4.76354182e-01 1.18876886e+00 -4.19218034e-01
-1.38092661e+00 3.96036357e-01 -2.85024852e-01 -1.00737727e+00
4.12194341e-01 -1.61614239e-01 4.97826152e-02 3.70099872e-01
-1.52994692e-01 1.16960967e+00 7.60136843e-01 -1.30756640e+00
-7.02325404e-01 -2.83131331e-01 -3.38853836e-01 1.49812505e-01
-6.92226529e-01 -2.06891015e-01 -7.46927679e-01 -7.06445992e-01
1.59976199e-01 -7.63578534e-01 -3.76952976e-01 -2.69500226e-01
-3.19122553e-01 -4.09607321e-01 5.04423261e-01 -2.04796910e-01
1.76936257e+00 -2.20746827e+00 2.35325028e-03 4.62567627e-01
5.82751513e-01 3.91806901e-01 1.30030766e-01 -1.06545277e-02
7.47948810e-02 3.01891953e-01 -4.04955208e-01 -9.94089186e-01
3.57562423e-01 4.42028761e-01 -1.22035466e-01 2.80750930e-01
-1.34232249e-02 6.55275047e-01 -7.96530187e-01 -8.67545426e-01
3.20127338e-01 6.11913085e-01 -9.18196321e-01 -1.66669503e-01
-7.82164261e-02 -2.70668566e-01 7.14263469e-02 2.73251355e-01
9.55115259e-01 -3.67073506e-01 4.86260168e-02 -3.14847529e-01
-9.19485465e-02 3.14265132e-01 -1.55100369e+00 1.56263828e+00
-2.01451927e-01 3.69046569e-01 -5.16413292e-03 -9.96274352e-01
7.91315436e-01 -2.76269130e-02 4.45432723e-01 -4.86418307e-01
3.23589891e-01 1.18176386e-01 -4.27515991e-02 1.50734959e-02
1.01257181e+00 -4.60443459e-02 7.11844116e-02 2.86516577e-01
2.07180381e-01 -1.08433411e-01 4.77861613e-01 3.17340434e-01
1.03464484e+00 -4.12867010e-01 2.28242263e-01 -5.25205970e-01
1.95602193e-01 -1.12839855e-01 5.42995512e-01 8.28059673e-01
-2.75620759e-01 7.89005578e-01 4.85882252e-01 -1.24371849e-01
-1.27173173e+00 -9.20898855e-01 -4.37475413e-01 1.15902591e+00
-1.51200265e-01 -8.44705641e-01 -1.16744280e+00 -2.69681662e-01
-8.36814791e-02 9.58813846e-01 -5.52656114e-01 -2.47974291e-01
-4.35687184e-01 -1.23864627e+00 9.08591986e-01 4.95391548e-01
5.46215594e-01 -8.28798413e-01 -8.45514238e-01 3.20429415e-01
-2.40318358e-01 -9.18579876e-01 -2.05728799e-01 7.99954176e-01
-1.01098490e+00 -7.69763589e-01 -5.38093805e-01 -4.12065953e-01
3.43274623e-01 5.09826876e-02 1.20461071e+00 -6.37848349e-03
-5.64446710e-02 -1.80070862e-01 -5.07705510e-01 -2.56168932e-01
-5.07696509e-01 5.17359018e-01 1.59039214e-01 -2.31952593e-01
4.40858364e-01 -5.39720953e-01 -5.39698482e-01 6.63056523e-02
-1.15494299e+00 -6.50308877e-02 6.08818054e-01 6.03888571e-01
1.01186574e+00 5.49101114e-01 3.89908880e-01 -7.22580492e-01
6.77200317e-01 -3.69487703e-01 -4.87453341e-01 -2.70602759e-02
-8.51973474e-01 2.60945618e-01 8.93324852e-01 -2.64896631e-01
-6.04080558e-01 2.94276010e-02 -2.75899470e-01 -8.66894424e-01
-5.47978617e-02 2.99501687e-01 -1.06099159e-01 1.63977608e-01
8.05384338e-01 2.55973965e-01 -7.01461211e-02 -3.91047269e-01
6.73190355e-01 6.23996735e-01 5.89542210e-01 -5.66004038e-01
3.26297492e-01 3.48875314e-01 4.57887016e-02 -8.58325720e-01
-3.60304594e-01 -5.53546131e-01 -7.95506537e-01 1.17268555e-01
8.80740404e-01 -8.65257144e-01 -6.84922099e-01 2.46090859e-01
-8.76300931e-01 -5.71389616e-01 -4.42228764e-01 3.30556452e-01
-5.30658841e-01 5.91033101e-01 -6.48186862e-01 -6.54786229e-01
-5.56451797e-01 -1.33050382e+00 9.60338950e-01 -8.96322131e-02
-3.42759162e-01 -6.43412650e-01 -2.45959729e-01 3.51712927e-02
2.77220279e-01 -1.07464001e-01 8.29858303e-01 -5.67630529e-01
-1.80777296e-01 1.78512380e-01 -3.11196446e-01 4.79026020e-01
-2.92125136e-01 9.73611921e-02 -9.36742187e-01 -4.09108162e-01
-1.93983346e-01 -2.59222984e-01 1.07366598e+00 5.39747417e-01
1.65754271e+00 -3.64335179e-01 -3.34856987e-01 7.05906928e-01
1.43977487e+00 -1.25471794e-03 6.60101295e-01 2.54309326e-01
6.62187636e-01 3.16830486e-01 6.01581991e-01 8.80268514e-01
2.98928738e-01 6.42208934e-01 6.07199311e-01 5.73938899e-02
-2.04658851e-01 9.30572897e-02 -6.32933155e-02 1.06725919e+00
1.56952724e-01 -2.62103081e-01 -8.84645998e-01 6.77323461e-01
-1.65336335e+00 -8.12057674e-01 -2.15263620e-01 2.22015643e+00
1.05851877e+00 4.80837017e-01 8.06302875e-02 7.04141915e-01
5.76807261e-01 -8.89901072e-02 -3.30482155e-01 -4.13352191e-01
6.66217580e-02 5.08482277e-01 7.77693748e-01 5.89429796e-01
-1.08701384e+00 9.24355268e-01 5.86931515e+00 1.59330952e+00
-8.66220176e-01 2.68720210e-01 8.22818816e-01 -4.06975150e-01
3.09180585e-03 -4.38351244e-01 -1.07198644e+00 7.54508972e-01
1.25331008e+00 3.86445299e-02 3.84059012e-01 9.76579666e-01
2.70046771e-01 -1.17001422e-01 -1.04444051e+00 1.11964965e+00
-9.65869334e-03 -1.46220601e+00 1.88095093e-01 2.48962283e-01
6.33346498e-01 1.18532613e-01 2.30101183e-01 2.62486160e-01
4.25106555e-01 -9.96511161e-01 9.95000005e-01 2.34994844e-01
7.05197215e-01 -1.39351857e+00 7.75295317e-01 4.35130417e-01
-1.13667250e+00 -1.29071921e-01 -8.01979840e-01 1.38462678e-01
-1.07702449e-01 9.45247471e-01 -8.14167142e-01 3.89884323e-01
1.01425040e+00 3.59743357e-01 -6.60718143e-01 9.64878023e-01
3.21646571e-01 8.24401677e-01 -6.74849033e-01 3.96063104e-02
4.95814860e-01 6.00269623e-02 6.51008934e-02 1.54662693e+00
3.07091117e-01 1.63699359e-01 1.13622844e-02 9.91481602e-01
-6.96043521e-02 -2.41344031e-02 -5.07359728e-02 5.33761322e-01
1.08944392e+00 9.45170343e-01 -9.96032953e-01 -8.51775348e-01
2.56098807e-01 8.15892696e-01 3.91759902e-01 -1.35283992e-01
-9.26903069e-01 -4.95379835e-01 5.74819744e-01 1.30747110e-01
7.73508251e-01 -3.44956107e-02 -5.06783485e-01 -5.21209359e-01
-2.78106183e-01 -7.86368072e-01 3.12566578e-01 -5.65315664e-01
-8.88979137e-01 6.08498871e-01 3.13978285e-01 -1.10390651e+00
-2.58042336e-01 -1.19637296e-01 -1.91114053e-01 6.74583077e-01
-1.42439723e+00 -6.03516281e-01 -3.21015835e-01 4.72476065e-01
4.93427128e-01 6.08588457e-02 7.83569455e-01 6.07584178e-01
-4.87625182e-01 1.14335167e+00 6.03814542e-01 -2.59056628e-01
4.71939445e-01 -1.45662534e+00 3.44310760e-01 7.57415891e-01
-1.04746195e-02 6.19177878e-01 7.77220428e-01 -3.23054105e-01
-9.03031826e-01 -1.50853729e+00 7.90773451e-01 -1.20563492e-01
3.19477260e-01 -3.52744281e-01 -1.06084776e+00 3.44401836e-01
-9.58378986e-02 4.77306880e-02 7.29078352e-01 -1.07765764e-01
-1.95687339e-01 -2.55806565e-01 -1.31661296e+00 4.18246955e-01
6.73818648e-01 -1.75992161e-01 -4.41841036e-01 1.95505887e-01
8.10903847e-01 -2.97063410e-01 -1.14242482e+00 2.85254836e-01
2.17279419e-01 -1.13664401e+00 1.03583312e+00 4.94985748e-03
4.31725711e-01 -3.78641486e-01 -5.00817597e-01 -1.19349289e+00
-6.65795147e-01 -4.05668616e-01 -2.81948090e-01 1.29531276e+00
1.55517608e-01 -1.65419653e-01 1.03609371e+00 2.31952116e-01
-4.21348035e-01 -8.28739643e-01 -1.15937603e+00 -7.89600492e-01
2.33061314e-01 -7.50358105e-01 6.46718025e-01 9.37841058e-01
-3.05561610e-02 1.25373006e-01 -6.42989650e-02 9.19417813e-02
9.01293218e-01 -4.26938802e-01 6.59966588e-01 -1.16484952e+00
-2.92594761e-01 -9.33980525e-01 -2.69404709e-01 -1.01535749e+00
-1.77348450e-01 -9.16594267e-01 1.39184013e-01 -1.28156340e+00
1.47909209e-01 -6.05555654e-01 -2.80612051e-01 4.65177000e-01
-2.00368583e-01 5.14783323e-01 4.49093550e-01 3.89219075e-01
-1.06151628e+00 6.11462235e-01 7.45306492e-01 -3.46491821e-02
-2.89104909e-01 -3.49710435e-01 -5.50175667e-01 6.39140248e-01
8.56104374e-01 -5.75638711e-01 -4.36936796e-01 -2.86439598e-01
4.25719470e-02 -4.98089671e-01 1.54679060e-01 -1.45198154e+00
3.13787997e-01 2.55033463e-01 2.16281563e-01 -9.84101832e-01
3.40190083e-01 -6.98561311e-01 1.69803560e-01 6.49936736e-01
-4.74561036e-01 1.79815087e-02 2.89360523e-01 3.53002846e-01
-1.36257544e-01 -4.46971983e-01 1.10665667e+00 -9.94371548e-02
-5.87316632e-01 2.85527613e-02 -3.00840408e-01 7.41127087e-03
8.46974254e-01 -2.25072965e-01 -7.08009005e-02 1.43672958e-01
-5.80509603e-01 1.94002792e-01 3.30605090e-01 -1.06186597e-02
4.74668503e-01 -1.32199490e+00 -6.73126400e-01 5.83283566e-02
-1.76259521e-02 4.11720008e-01 2.57401079e-01 6.48013711e-01
-8.03939283e-01 6.10287488e-01 6.95760325e-02 -9.17863488e-01
-1.42528450e+00 6.28874779e-01 1.14248402e-01 -5.11046648e-01
-6.23267710e-01 1.06030715e+00 -1.79764792e-01 -1.92481816e-01
6.74692690e-01 -8.53352070e-01 -2.40781263e-01 2.19537690e-01
7.32064724e-01 7.87059426e-01 5.56772053e-01 -5.07759571e-01
-1.43097460e-01 2.67883837e-01 -2.15829313e-01 -1.83939300e-02
1.20516360e+00 -1.34256154e-01 -5.39196916e-02 3.62861425e-01
1.29173732e+00 -5.09857774e-01 -1.27933145e+00 -1.84256569e-01
-1.27060652e-01 -2.83810347e-01 4.54854131e-01 -4.70324904e-01
-1.10226464e+00 9.48975444e-01 8.45118284e-01 3.45681489e-01
1.09588528e+00 -9.55442712e-02 8.13948750e-01 4.66140896e-01
2.51846462e-01 -1.26391613e+00 -8.00062865e-02 3.95335466e-01
4.44698662e-01 -8.28260958e-01 2.01050669e-01 -1.21710733e-01
-4.77250636e-01 6.68444455e-01 3.95211130e-01 -7.87610412e-02
6.84556782e-01 3.16450953e-01 -3.92358691e-01 -1.03399374e-01
-8.16495001e-01 -3.76387835e-02 5.17998300e-02 4.93772417e-01
1.41322568e-01 2.27962658e-01 -4.13687557e-01 7.57893741e-01
-7.48654366e-01 -4.15583737e-02 3.58524650e-01 7.35514283e-01
-8.65842402e-01 -7.63096988e-01 -6.53954208e-01 7.54837334e-01
-4.99187917e-01 -3.19750339e-01 9.74477604e-02 6.64786935e-01
5.95606744e-01 8.64638805e-01 4.99394447e-01 -5.13557911e-01
7.52381161e-02 -5.95091730e-02 3.54460359e-01 -3.83904427e-01
-6.74113572e-01 2.32298777e-01 -3.93387258e-01 -6.03473246e-01
-2.74230719e-01 -5.89551747e-01 -1.43783903e+00 -6.80009663e-01
-3.78656805e-01 2.83948123e-01 6.78809345e-01 8.38533819e-01
5.06727040e-01 8.57681394e-01 2.62525588e-01 -9.61443245e-01
-6.20126545e-01 -1.21399677e+00 -7.07381308e-01 1.89002961e-01
1.60068482e-01 -7.41028190e-01 -4.28605795e-01 8.82176012e-02] | [8.595242500305176, 3.1163887977600098] |
e4a7732c-44d5-4d8d-ae4d-d9552c689b1e | magnification-invariant-medical-image | 2302.11488 | null | https://arxiv.org/abs/2302.11488v1 | https://arxiv.org/pdf/2302.11488v1.pdf | Magnification Invariant Medical Image Analysis: A Comparison of Convolutional Networks, Vision Transformers, and Token Mixers | Convolution Neural Networks (CNNs) are widely used in medical image analysis, but their performance degrade when the magnification of testing images differ from the training images. The inability of CNNs to generalize across magnification scales can result in sub-optimal performance on external datasets. This study aims to evaluate the robustness of various deep learning architectures in the analysis of breast cancer histopathological images with varying magnification scales at training and testing stages. Here we explore and compare the performance of multiple deep learning architectures, including CNN-based ResNet and MobileNet, self-attention-based Vision Transformers and Swin Transformers, and token-mixing models, such as FNet, ConvMixer, MLP-Mixer, and WaveMix. The experiments are conducted using the BreakHis dataset, which contains breast cancer histopathological images at varying magnification levels. We show that performance of WaveMix is invariant to the magnification of training and testing data and can provide stable and good classification accuracy. These evaluations are critical in identifying deep learning architectures that can robustly handle changes in magnification scale, ensuring that scale changes across anatomical structures do not disturb the inference results. | ['Amit Sethi', 'Nikhil Cherian Kurian', 'Pranav Jeevan'] | 2023-02-22 | null | null | null | null | ['breast-cancer-histology-image-classification'] | ['medical'] | [ 1.67939380e-01 -1.26707211e-01 -1.51003376e-02 -3.37066829e-01
-4.80066985e-01 -3.79693657e-01 3.59056741e-01 8.02406594e-02
-8.14433396e-01 2.96858937e-01 -3.82967964e-02 -5.61998069e-01
-1.81878671e-01 -7.05098093e-01 -5.07744014e-01 -7.09307611e-01
-1.38175383e-01 1.81015596e-01 3.73501867e-01 -1.81422189e-01
5.75157739e-02 7.86667407e-01 -9.88099694e-01 7.30425477e-01
5.30362904e-01 1.05694234e+00 2.06796914e-01 1.22203088e+00
5.23467697e-02 1.16970420e+00 -8.48125875e-01 -3.95088136e-01
5.65903708e-02 -7.87524506e-02 -1.03816462e+00 -4.57806319e-01
5.56644857e-01 -4.08912361e-01 -5.61470926e-01 1.09901321e+00
9.32826340e-01 -3.99200052e-01 5.72664082e-01 -1.00618410e+00
-8.03392410e-01 6.07452691e-01 -3.53102267e-01 1.16684556e+00
-4.95403081e-01 3.77005279e-01 3.53423685e-01 -5.09279191e-01
4.17889774e-01 1.04502714e+00 1.25552654e+00 5.90737700e-01
-1.09392703e+00 -6.45645499e-01 -3.33010823e-01 4.33146745e-01
-1.21940005e+00 -3.69589269e-01 4.59319025e-01 -2.76109517e-01
9.32292521e-01 3.31762731e-01 5.04369199e-01 1.21713364e+00
1.00314903e+00 3.59626919e-01 9.87155914e-01 -3.06213051e-01
5.39300106e-02 -5.62392268e-03 -5.11818156e-02 7.39728272e-01
3.01955968e-01 -1.81540594e-01 -2.77820289e-01 -6.38319999e-02
1.07113779e+00 9.03768465e-02 -2.20815629e-01 4.18290287e-01
-1.41234863e+00 7.05588341e-01 9.64598477e-01 7.64079988e-01
-2.62517154e-01 4.97344285e-01 6.53481781e-01 5.23414254e-01
4.33485091e-01 7.44511724e-01 -4.79542315e-01 1.47038296e-01
-9.29995596e-01 -1.66118547e-01 3.28866690e-01 5.01906693e-01
9.64091420e-02 -3.82924499e-03 -3.14713240e-01 9.09985185e-01
-1.15008555e-01 2.64924824e-01 1.33097088e+00 -7.29070187e-01
1.33067697e-01 6.13895416e-01 -6.18625939e-01 -1.04436350e+00
-1.09486389e+00 -7.99328804e-01 -1.25173008e+00 1.81104407e-01
4.59214389e-01 -1.45206554e-02 -1.20080817e+00 1.57088494e+00
-5.37222996e-02 -9.30229798e-02 7.11788759e-02 4.62151259e-01
1.35107040e+00 1.15698136e-01 3.40561807e-01 2.65959710e-01
1.55668998e+00 -8.23972285e-01 -5.26306331e-01 -2.76308268e-01
7.09131539e-01 -8.34470749e-01 1.09877932e+00 2.30208617e-02
-1.37451005e+00 -8.00975382e-01 -1.18274033e+00 -3.30814093e-01
-6.12097621e-01 2.64507979e-01 5.13734639e-01 6.01964772e-01
-1.36596680e+00 8.13277423e-01 -1.17787039e+00 -5.10763586e-01
7.43403018e-01 6.08182490e-01 -3.39772314e-01 1.59801394e-01
-1.04232049e+00 1.23648834e+00 2.41692707e-01 3.60681206e-01
-7.50828564e-01 -1.07222092e+00 -4.75835115e-01 9.79591757e-02
-5.25971770e-01 -8.35657239e-01 1.38637364e+00 -1.25714624e+00
-1.14038634e+00 1.05828238e+00 3.57452154e-01 -5.40382504e-01
6.55931473e-01 3.76801401e-01 -4.11269218e-01 3.86798471e-01
-5.27990749e-03 9.57092822e-01 5.96330345e-01 -6.99170411e-01
-4.78839248e-01 -2.77348727e-01 -1.50829628e-02 -9.72912982e-02
-5.32827437e-01 -7.47712627e-02 -1.52242184e-01 -6.91067994e-01
-9.26081762e-02 -8.26892316e-01 -2.31885567e-01 1.99513510e-01
-5.09650111e-01 1.51749581e-01 6.52413964e-01 -6.59172058e-01
8.27275157e-01 -2.16195130e+00 -2.04471245e-01 7.63059482e-02
4.85663652e-01 1.72350273e-01 -3.80119592e-01 -7.97895566e-02
-3.58549386e-01 2.53481656e-01 2.21025065e-01 5.49537875e-02
-6.03161931e-01 1.15571007e-01 3.34258854e-01 7.65875876e-01
2.38022566e-01 1.32626820e+00 -6.55890346e-01 -7.09855914e-01
1.53839186e-01 6.17314458e-01 -4.17217225e-01 1.08070627e-01
3.37970793e-01 2.38577932e-01 -2.62727737e-02 7.67101765e-01
5.97746968e-01 -7.83332348e-01 5.76350614e-02 -7.89738357e-01
2.44614333e-01 -1.91769287e-01 -3.37395608e-01 1.32916534e+00
-6.11595571e-01 1.04182267e+00 -3.95382680e-02 -9.14676130e-01
4.44728583e-01 2.43538544e-01 4.25797522e-01 -8.67700875e-01
4.51553285e-01 1.84637308e-01 4.73712236e-01 -7.30261385e-01
3.39751929e-01 -2.61777014e-01 3.21957827e-01 1.76469348e-02
4.07495171e-01 8.15120414e-02 1.24747328e-01 -2.19243526e-01
1.39480853e+00 -8.91768456e-01 9.87587497e-02 -5.38163722e-01
2.18575805e-01 -6.52863756e-02 9.80048776e-02 7.41301537e-01
-2.80631691e-01 7.12356627e-01 6.76987648e-01 -8.01644206e-01
-1.27465284e+00 -9.79959607e-01 -6.01839602e-01 1.03383982e+00
-5.17363846e-03 1.38194174e-01 -7.87861645e-01 -5.70800662e-01
-3.33829671e-01 7.16005936e-02 -1.27874112e+00 -4.47490543e-01
-4.62632596e-01 -1.40418077e+00 1.10389483e+00 9.11368251e-01
7.57888198e-01 -1.05846727e+00 -9.19494212e-01 1.10824957e-01
-1.07150115e-01 -1.01334178e+00 -2.77259678e-01 4.47886646e-01
-1.16046369e+00 -1.38222921e+00 -9.18632805e-01 -1.06999373e+00
1.16391015e+00 1.64544526e-02 1.05490208e+00 3.78715128e-01
-6.49760783e-01 2.23430037e-01 -5.15143089e-02 -4.99360532e-01
-7.04941332e-01 6.83206439e-01 -3.44979078e-01 -3.22333872e-01
-2.27898303e-02 -4.48465198e-01 -1.04784882e+00 4.11215752e-01
-1.23527145e+00 1.56210631e-01 9.40261662e-01 1.03618646e+00
3.57186824e-01 1.56825155e-01 4.00753021e-01 -8.45809102e-01
7.89997458e-01 -4.21277940e-01 3.26737203e-02 3.17580223e-01
-4.65906382e-01 -2.55682208e-02 6.96277738e-01 -7.88773954e-01
-5.50355971e-01 -5.17091691e-01 -3.07769984e-01 -4.36141074e-01
8.56482610e-02 4.48544502e-01 3.78700197e-01 -4.88047272e-01
1.21967757e+00 -1.30257681e-01 4.15204853e-01 -6.80636019e-02
-1.67702630e-01 3.93677771e-01 7.48024881e-01 7.09712058e-02
3.79121542e-01 7.41564751e-01 1.46701425e-01 -7.08922505e-01
-5.16130328e-01 -4.42829691e-02 -4.36332643e-01 -2.55259246e-01
8.27269137e-01 -6.01231456e-01 -6.32216454e-01 7.82870591e-01
-9.13032651e-01 -6.86615646e-01 -2.93519646e-01 3.29979807e-01
-1.28869817e-01 -8.17165151e-02 -1.32264817e+00 2.15068370e-01
-8.92269194e-01 -1.38564515e+00 7.37745702e-01 4.48266834e-01
-2.73722589e-01 -1.36171758e+00 -1.35141850e-01 1.90490577e-02
1.07318330e+00 4.17716652e-01 1.30416763e+00 -6.76296890e-01
4.85926867e-02 -3.16188961e-01 -4.21483427e-01 3.96573097e-01
3.12310547e-01 2.66904116e-01 -1.07635248e+00 -4.93113250e-01
-1.37050629e-01 -2.71886379e-01 8.49638343e-01 8.05863082e-01
1.53101575e+00 -3.36034924e-01 -3.38456959e-01 9.85235572e-01
1.26373613e+00 3.86441685e-02 8.67431462e-01 6.45825267e-01
5.08978009e-01 2.67383993e-01 -3.98540348e-01 -2.58932382e-01
1.17060132e-01 5.03702350e-02 5.47087550e-01 -8.70762408e-01
-4.21636641e-01 1.79633588e-01 -2.77276158e-01 6.73537135e-01
-2.16868371e-01 -3.24142762e-02 -9.22396183e-01 3.81514490e-01
-1.27952790e+00 -8.80378067e-01 1.21082351e-01 1.64137399e+00
8.48153889e-01 2.35549614e-01 -9.83829051e-02 1.99619979e-01
6.25220180e-01 -6.43986836e-02 -6.21487737e-01 -2.62009948e-01
1.73740964e-02 3.20002288e-01 9.52459395e-01 4.20511365e-02
-1.05033326e+00 4.14261550e-01 7.51071405e+00 7.50753105e-01
-1.81800210e+00 3.86548191e-01 1.21921706e+00 -1.72924206e-01
7.92767480e-02 -7.52590477e-01 -2.25576028e-01 5.08990407e-01
1.01768041e+00 1.68849364e-01 5.86407967e-02 7.63317943e-01
7.36448262e-03 1.20576672e-01 -1.02160084e+00 8.20434630e-01
-3.94127965e-02 -1.70298851e+00 -5.23355491e-02 -7.25010857e-02
5.78550816e-01 4.18466777e-01 5.46333730e-01 2.19654694e-01
1.03657544e-01 -1.33278906e+00 5.00900388e-01 2.93552399e-01
1.18196619e+00 -5.57605088e-01 1.29961371e+00 -2.32959598e-01
-7.03110993e-01 -1.47359714e-01 -5.27090430e-01 4.50299114e-01
-5.08907318e-01 4.07728404e-01 -1.10086954e+00 8.12168494e-02
9.49284732e-01 3.24156195e-01 -1.10280991e+00 1.03874946e+00
1.87203437e-01 5.04430115e-01 -1.27673359e-03 -1.16750650e-01
1.78059369e-01 7.95419276e-01 5.77332191e-02 1.51344955e+00
2.50982374e-01 -3.34916323e-01 -3.01620066e-01 6.02428854e-01
-2.83348411e-01 -4.35072817e-02 -4.21364084e-02 2.09786922e-01
1.08609095e-01 1.54500425e+00 -1.27681327e+00 -2.03913286e-01
-1.57536745e-01 6.81014597e-01 2.11695999e-01 3.32255304e-01
-9.20159400e-01 -3.55335742e-01 4.19339955e-01 3.14740658e-01
8.74639116e-03 1.85407668e-01 -3.67863804e-01 -7.43434489e-01
-3.54920179e-01 -8.37099612e-01 5.86957812e-01 -7.66377151e-01
-1.29229105e+00 9.29312289e-01 -6.53204620e-02 -1.11009145e+00
2.50690907e-01 -9.80383694e-01 -9.59551275e-01 4.66630250e-01
-1.49560034e+00 -1.33641839e+00 -6.55928016e-01 6.63741410e-01
5.02761960e-01 -1.34056911e-01 7.34748185e-01 3.58178943e-01
-4.10503536e-01 8.97516370e-01 3.68358521e-03 4.59471077e-01
5.76987922e-01 -1.19333375e+00 2.84330159e-01 4.94933039e-01
-4.24815387e-01 6.15574181e-01 5.28262734e-01 -1.60489425e-01
-1.01126063e+00 -1.34415221e+00 4.11939949e-01 -1.82638571e-01
5.47859967e-01 1.13005908e-02 -8.11248660e-01 6.24360621e-01
2.38358632e-01 4.78389174e-01 8.03834617e-01 -9.92018655e-02
-3.10586691e-01 -3.43309879e-01 -1.34746766e+00 5.85347056e-01
6.26049161e-01 -4.83725816e-01 -2.35605478e-01 4.65434223e-01
4.26795155e-01 -8.26497614e-01 -1.01980329e+00 6.50512576e-01
5.91144264e-01 -8.04575920e-01 1.11958921e+00 -6.36001229e-01
8.18322480e-01 1.92690894e-01 1.92871779e-01 -1.40049732e+00
-6.87823832e-01 1.74175634e-03 2.81653583e-01 6.47458673e-01
5.81875265e-01 -7.40854323e-01 6.87019467e-01 3.18841904e-01
-1.17495432e-01 -9.70265985e-01 -1.18256819e+00 -4.99199539e-01
3.59812260e-01 -4.58747037e-02 4.43651468e-01 8.78565609e-01
-1.43184856e-01 -9.75289498e-04 2.23228976e-01 5.33857048e-02
2.79757023e-01 -4.38944340e-01 2.85167396e-01 -8.27161789e-01
-1.30063385e-01 -8.64117324e-01 -9.15127099e-01 -2.50521392e-01
-1.83050022e-01 -1.07757831e+00 -2.70318747e-01 -1.46884906e+00
3.59533042e-01 -3.35172653e-01 -5.45798898e-01 5.92791855e-01
-6.74998090e-02 7.01896548e-01 -1.10240743e-01 3.17505181e-01
-2.97871590e-01 -4.05136086e-02 1.49320281e+00 -5.91071010e-01
4.54649888e-02 -7.55777434e-02 -8.87657523e-01 7.43821979e-01
9.65373456e-01 -3.13275188e-01 -2.91574448e-01 -6.24201775e-01
1.14397272e-01 -3.13491374e-01 5.88847578e-01 -1.25707018e+00
4.74281490e-01 1.53414503e-01 1.13511395e+00 -1.42601892e-01
-1.61099240e-01 -4.91079003e-01 1.94912642e-01 9.87968087e-01
-7.48930812e-01 5.06655991e-01 5.60375690e-01 -4.46183747e-03
2.66957395e-02 -1.92171305e-01 1.24726677e+00 -3.36992532e-01
-2.30431885e-01 3.13656598e-01 -4.83757824e-01 -2.95624416e-02
7.37480462e-01 -2.41369158e-01 -8.06503177e-01 -4.31322381e-02
-6.28622174e-01 -8.27997848e-02 1.51211813e-01 3.31053019e-01
4.86165941e-01 -1.29568553e+00 -7.11932719e-01 7.98000768e-02
2.03098413e-02 -1.08613223e-02 6.41568184e-01 1.18013608e+00
-1.19112170e+00 1.34931862e-01 -5.77922523e-01 -7.97153473e-01
-1.35770738e+00 4.05832231e-01 1.04899704e+00 -5.57934940e-01
-5.09471357e-01 9.86257136e-01 1.50726601e-01 -3.23774338e-01
1.91801414e-01 -9.49132085e-01 -2.00398162e-01 -2.64979750e-01
6.73034012e-01 3.62565339e-01 5.56215882e-01 -1.21975660e-01
-2.55287677e-01 3.09064865e-01 -3.95744950e-01 2.71551669e-01
1.14809012e+00 3.07835519e-01 -3.72935608e-02 2.29978472e-01
1.48913109e+00 -5.14055669e-01 -8.62101257e-01 1.41918082e-02
-3.93760949e-01 -5.45234159e-02 2.96773672e-01 -8.22866261e-01
-1.53891349e+00 7.46752203e-01 1.16343617e+00 4.15733486e-01
1.21738231e+00 -4.34157737e-02 5.49012184e-01 1.78726301e-01
-1.09685794e-01 -9.19432640e-01 3.47602874e-01 9.19208527e-02
8.09083164e-01 -1.10464406e+00 -1.24983393e-01 1.31075472e-01
-2.50616550e-01 1.39606559e+00 7.65549541e-01 -9.62566584e-02
6.44473553e-01 7.12349117e-01 2.25262061e-01 -4.81743127e-01
-6.46608055e-01 4.16802824e-01 1.52060628e-01 4.75489855e-01
5.51123321e-01 -9.07670259e-02 1.43473595e-01 3.08221787e-01
-4.32526171e-01 6.72235712e-02 4.51042950e-01 9.02360499e-01
-1.13910757e-01 -4.37074602e-01 -3.24469835e-01 8.69594276e-01
-1.01394820e+00 -1.36522517e-01 -9.31785852e-02 1.02405655e+00
1.86279267e-01 5.49159646e-01 3.96807402e-01 -3.27869564e-01
4.73052233e-01 -2.91217387e-01 6.19256198e-01 -1.58168435e-01
-9.84999716e-01 -2.58792967e-01 -2.80952603e-01 -3.25140178e-01
-2.83666581e-01 -8.27037767e-02 -1.10345435e+00 -7.19256878e-01
-3.18479568e-01 -1.52256250e-01 6.86520875e-01 7.32262254e-01
1.61769599e-01 1.28498209e+00 3.49476606e-01 -7.36868978e-01
-4.78275716e-01 -1.24612057e+00 -3.55288506e-01 2.37922281e-01
5.29068172e-01 -2.82145888e-01 -4.01767254e-01 1.28804058e-01] | [15.02774715423584, -2.6805691719055176] |
0cc56657-fe4c-4506-ba31-ef750c4bd9b1 | efficient-re-parameterization-residual | 2109.05479 | null | https://arxiv.org/abs/2109.05479v2 | https://arxiv.org/pdf/2109.05479v2.pdf | Efficient Re-parameterization Residual Attention Network For Nonhomogeneous Image Dehazing | This paper proposes an end-to-end Efficient Re-parameterizationResidual Attention Network(ERRA-Net) to directly restore the nonhomogeneous hazy image. The contribution of this paper mainly has the following three aspects: 1) A novel Multi-branch Attention (MA) block. The spatial attention mechanism better reconstructs high-frequency features, and the channel attention mechanism treats the features of different channels differently. Multi-branch structure dramatically improves the representation ability of the model and can be changed into a single path structure after re-parameterization to speed up the process of inference. Local Residual Connection allows the low-frequency information in the nonhomogeneous area to pass through the block without processing so that the block can focus on detailed features. 2) A lightweight network structure. We use cascaded MA blocks to extract high-frequency features step by step, and the Multi-layer attention fusion tail combines the shallow and deep features of the model to get the residual of the clean image finally. 3)We propose two novel loss functions to help reconstruct the hazy image ColorAttenuation loss and Laplace Pyramid loss. ERRA-Net has an impressive speed, processing 1200x1600 HD quality images with an average runtime of 166.11 fps. Extensive evaluations demonstrate that ERSANet performs favorably against the SOTA approaches on the real-world hazy images. | ['Peng Chen', 'XinRui Huang', 'ErKang Chen', 'Tian Ye'] | 2021-09-12 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [-1.13123149e-01 -3.41936052e-01 3.08752984e-01 -2.75053620e-01
-6.85722530e-01 3.00148875e-01 -1.33825503e-02 -5.07698298e-01
-4.63165820e-01 6.11278296e-01 2.34986782e-01 -2.20367193e-01
-7.26183727e-02 -9.29446518e-01 -8.73583853e-01 -1.28398705e+00
8.76660496e-02 -3.16198438e-01 3.83475572e-01 -2.94899702e-01
1.11772880e-01 4.77192730e-01 -1.60275972e+00 3.77789825e-01
1.29170394e+00 1.19682658e+00 4.46828097e-01 7.00188279e-01
-6.02731295e-03 8.68836880e-01 -5.40649056e-01 -3.88584360e-02
4.86013949e-01 -2.96189517e-01 -3.22465807e-01 -6.76231161e-02
8.30908895e-01 -8.07398200e-01 -6.24277890e-01 1.24739838e+00
7.43056059e-01 2.78346390e-01 3.04604053e-01 -9.36628163e-01
-7.98761904e-01 3.19526881e-01 -8.58815968e-01 6.25271261e-01
-2.83971757e-01 3.74344260e-01 7.16443360e-01 -1.09216118e+00
2.37766504e-02 1.41891909e+00 4.44836736e-01 2.03280672e-01
-8.25907528e-01 -1.01746786e+00 3.20447922e-01 6.86377168e-01
-1.42468250e+00 -4.14930314e-01 6.35233819e-01 1.70198455e-02
8.38447034e-01 2.43767828e-01 6.42634630e-01 3.77319098e-01
4.94147152e-01 7.18960047e-01 1.06992280e+00 -2.36020967e-01
-1.06623478e-01 -2.95540720e-01 2.47437596e-01 8.27296078e-01
2.40807191e-01 7.31716976e-02 -3.75597507e-01 3.52586895e-01
9.94330406e-01 3.49237680e-01 -7.04350054e-01 1.69618994e-01
-7.61253655e-01 7.17046499e-01 9.86554146e-01 2.03942999e-01
-5.59180260e-01 4.44054514e-01 -1.23077698e-01 4.28456962e-01
5.91061592e-01 1.22794092e-01 -2.80334413e-01 4.40973103e-01
-8.48733008e-01 9.61609110e-02 8.65085497e-02 8.08802307e-01
1.32465875e+00 5.10845840e-01 -3.29082906e-01 8.64350200e-01
3.29288065e-01 8.02877963e-01 3.41176271e-01 -1.04950285e+00
5.02353668e-01 2.27109686e-01 -8.74349028e-02 -9.32773530e-01
-1.24359623e-01 -5.05746424e-01 -1.26665497e+00 6.09000027e-01
-3.43947448e-02 -1.44580543e-01 -1.43187559e+00 1.33079696e+00
1.87454835e-01 5.13807118e-01 -7.77994916e-02 1.28510916e+00
1.00726914e+00 1.35382497e+00 -2.03222886e-01 -1.56163469e-01
1.23991311e+00 -1.34985030e+00 -9.56937611e-01 -1.63553823e-02
-1.59803689e-01 -7.15724587e-01 9.20223713e-01 4.41103488e-01
-1.50545871e+00 -8.46850336e-01 -1.21309030e+00 -6.16908669e-01
-3.22497338e-01 5.43871969e-02 4.69928652e-01 1.83626264e-01
-1.35977304e+00 3.72209549e-01 -6.65950894e-01 1.80502385e-01
5.30190766e-01 5.27523637e-01 -1.14973947e-01 -5.47098637e-01
-1.06792140e+00 6.63212061e-01 1.13027751e-01 5.54945588e-01
-1.02183986e+00 -8.42971623e-01 -9.07250941e-01 4.48800951e-01
2.19996095e-01 -9.06744778e-01 8.24438989e-01 -1.18842173e+00
-1.54372776e+00 1.70950055e-01 -3.68336409e-01 -2.79557198e-01
6.41503111e-02 -5.25616407e-01 -4.51261610e-01 4.28101242e-01
-2.01295186e-02 6.95241153e-01 1.25906038e+00 -1.20753562e+00
-8.44261408e-01 -2.25957274e-01 4.31873053e-02 3.98032099e-01
-1.91726282e-01 -1.95673868e-01 -8.59269738e-01 -9.55457926e-01
2.12222531e-01 -4.54546601e-01 -1.64554402e-01 1.44321248e-01
-2.58948773e-01 2.30666533e-01 8.81593108e-01 -9.38918710e-01
1.26476896e+00 -2.43986773e+00 9.25887078e-02 3.66314277e-02
4.77423906e-01 5.81826806e-01 -2.32589230e-01 -2.51887180e-03
-1.44052133e-01 -1.03063308e-01 -3.47143084e-01 -2.04346657e-01
-3.73620629e-01 -1.53995678e-02 -4.63494897e-01 5.27922690e-01
1.53208405e-01 8.68551254e-01 -4.70122427e-01 -3.86426508e-01
4.41462874e-01 8.90441537e-01 -6.74724758e-01 5.36142886e-01
2.90794969e-01 1.74352556e-01 -1.38907477e-01 6.59415901e-01
1.13785779e+00 -2.09591731e-01 -6.60792589e-01 -5.31815648e-01
-2.19684750e-01 -6.86130952e-04 -9.15928900e-01 1.41675305e+00
-7.11964250e-01 6.53247833e-01 4.53079015e-01 -5.98272681e-01
6.77159786e-01 -4.25013192e-02 4.26186472e-02 -9.36561227e-01
1.39088422e-01 -5.88738024e-02 -1.58682629e-01 -6.54181182e-01
4.06321347e-01 -1.27015188e-01 3.56526941e-01 -2.06878297e-02
4.90628183e-02 8.33161548e-02 -3.01027268e-01 2.87096184e-02
6.58188581e-01 -2.31743515e-01 5.25930058e-03 -2.38049820e-01
7.96117306e-01 -3.69708329e-01 6.30995333e-01 7.42200136e-01
-1.21571079e-01 8.56229186e-01 -1.44343516e-02 -5.68726182e-01
-7.98631489e-01 -1.12450969e+00 -3.85167487e-02 9.88089561e-01
6.25947058e-01 -9.61124673e-02 -5.23873210e-01 -3.15059036e-01
-1.47962227e-01 5.08318484e-01 -6.94482148e-01 -2.84456879e-01
-7.50482738e-01 -9.23184395e-01 3.37266475e-02 3.87565285e-01
1.15485382e+00 -9.72370148e-01 -2.57715642e-01 1.65912136e-01
-3.86625648e-01 -8.38861287e-01 -5.87140262e-01 2.05606967e-02
-7.70171225e-01 -7.77611613e-01 -7.78452277e-01 -8.68198276e-01
7.34267473e-01 9.11092222e-01 8.11090350e-01 3.95034969e-01
-2.90103048e-01 -2.27478743e-02 -1.35550350e-01 -4.73869711e-01
2.66206563e-01 -4.06576514e-01 -3.61636490e-01 4.16693747e-01
9.16324332e-02 -7.54011989e-01 -1.01164460e+00 7.08036721e-02
-9.30643857e-01 1.18557222e-01 9.77133334e-01 1.02808344e+00
6.72768116e-01 4.30791676e-01 1.46093741e-01 -5.50842106e-01
1.79095238e-01 -4.25779551e-01 -6.81703866e-01 -2.77368631e-02
-5.34648895e-01 -1.43479884e-01 7.37097740e-01 -2.00787738e-01
-1.43150759e+00 -2.76739717e-01 -2.61682361e-01 -7.54843056e-01
-3.44220996e-02 1.56298280e-01 -3.38222265e-01 -3.80773246e-01
1.18321314e-01 4.29490268e-01 1.51004791e-02 -6.10261500e-01
2.45254949e-01 4.39406663e-01 8.73286247e-01 -2.22744703e-01
1.01599848e+00 7.15040267e-01 -1.52048260e-01 -9.71333444e-01
-6.75293863e-01 -4.86126572e-01 -2.51148731e-01 4.96665854e-03
9.90025163e-01 -1.47483885e+00 -7.30334044e-01 7.86924303e-01
-1.08667374e+00 -4.28459108e-01 -4.82824333e-02 5.35925508e-01
-2.20177881e-02 3.82811904e-01 -1.03522098e+00 -6.43031716e-01
-8.15056145e-01 -1.11557901e+00 1.02662539e+00 4.97364610e-01
8.42574418e-01 -7.05551386e-01 -4.75618333e-01 2.86041200e-01
7.48449266e-01 -2.44867638e-01 9.07095671e-01 2.54079014e-01
-1.22983539e+00 1.89517885e-01 -7.03964114e-01 6.84012115e-01
-1.65168434e-01 -8.62731636e-02 -1.18828857e+00 -4.51019526e-01
1.78298339e-01 4.89165373e-02 1.44639516e+00 7.70985961e-01
1.46865928e+00 -3.56817871e-01 1.42012194e-01 1.26034355e+00
1.71343732e+00 -4.15517054e-02 1.11311960e+00 3.51181269e-01
9.22339916e-01 1.93415806e-01 5.40707469e-01 6.71520904e-02
5.10292351e-01 4.39217776e-01 6.65026546e-01 -7.50181377e-01
-5.88521183e-01 6.43143654e-02 5.49148560e-01 9.16229904e-01
-5.10815531e-02 -3.43804061e-01 -3.24682832e-01 6.17424071e-01
-1.62324524e+00 -8.23773146e-01 -1.53139189e-01 1.98885953e+00
5.28645694e-01 -1.02651753e-01 -4.41662312e-01 -1.52921572e-01
5.59521139e-01 3.19111377e-01 -3.69609416e-01 -1.02187753e-01
-2.67987370e-01 2.98311740e-01 5.79127789e-01 8.54664624e-01
-9.14449275e-01 8.41725469e-01 5.88726568e+00 1.01772630e+00
-1.31509197e+00 9.35508385e-02 7.56712198e-01 -2.67601460e-01
-1.61270350e-01 -1.71127617e-01 -8.00669551e-01 7.77435720e-01
7.61381865e-01 1.72109425e-01 7.06435502e-01 3.19038302e-01
2.43052781e-01 -1.98940992e-01 -3.47791761e-01 1.21230090e+00
2.21008211e-01 -1.27158582e+00 3.70873094e-01 3.98433730e-02
5.88397741e-01 3.12785864e-01 2.77855217e-01 3.73152375e-01
2.21036822e-01 -9.41414177e-01 6.98064148e-01 6.86510146e-01
9.09009874e-01 -8.88947546e-01 1.01529932e+00 1.15664490e-01
-1.15003300e+00 -3.23696822e-01 -7.62518644e-01 1.07855529e-01
9.38190520e-02 7.60704517e-01 -3.32444787e-01 7.65277863e-01
1.31391120e+00 6.37118638e-01 -5.29245853e-01 1.22921908e+00
-3.59959394e-01 7.16733217e-01 -2.83448100e-01 5.43610513e-01
4.02935207e-01 -2.99095690e-01 3.01211894e-01 1.18293536e+00
4.15522933e-01 5.36215603e-01 -1.34220824e-01 6.68265700e-01
-3.32581475e-02 -3.14528644e-01 -6.66227788e-02 7.27502048e-01
1.49957746e-01 1.15207040e+00 -2.32424647e-01 -4.40430850e-01
-5.25819540e-01 1.11159360e+00 1.01890601e-01 8.17281783e-01
-9.82401013e-01 -7.64721453e-01 8.30919325e-01 7.76496455e-02
8.78172994e-01 1.33284209e-02 -2.88992643e-01 -1.23434436e+00
8.93927738e-03 -7.93296278e-01 3.04155797e-01 -1.26050937e+00
-1.19153702e+00 6.87655032e-01 -2.14181170e-01 -9.48413670e-01
4.59859461e-01 -4.53260154e-01 -8.05847406e-01 1.31554186e+00
-2.47018290e+00 -1.14665663e+00 -9.37612653e-01 9.42832649e-01
6.02236748e-01 1.31098986e-01 3.31047893e-01 5.80190539e-01
-8.84024322e-01 7.04313338e-01 5.37160560e-02 1.57109704e-02
7.76661038e-01 -1.08627117e+00 2.35161543e-01 1.29385614e+00
-3.57930362e-01 5.99570870e-01 4.29591656e-01 -5.04215658e-01
-1.23549819e+00 -1.27309752e+00 5.51650822e-01 9.71126894e-04
2.13955775e-01 -1.80336803e-01 -1.32352436e+00 4.54505891e-01
3.62484545e-01 3.73471856e-01 1.64129108e-01 -3.54827374e-01
-4.61562335e-01 -6.42994165e-01 -9.41157520e-01 3.80083442e-01
5.43436944e-01 -4.18455601e-01 -3.58594894e-01 -3.14983949e-02
9.60279703e-01 -4.12641674e-01 -5.00994861e-01 3.86095166e-01
2.37509310e-01 -1.17969501e+00 1.12784636e+00 -1.05332330e-01
3.74456435e-01 -7.47969270e-01 -2.06350416e-01 -1.15499091e+00
-7.59519637e-01 -7.78430939e-01 -3.23872745e-01 8.79331887e-01
2.00095996e-01 -7.07696497e-01 4.40444440e-01 1.98251814e-01
-5.34793973e-01 -8.98012161e-01 -7.27189839e-01 -3.60211611e-01
-2.52878405e-02 2.69695763e-02 8.14153552e-01 6.37653172e-01
-6.33735895e-01 3.11846077e-01 -6.74693286e-01 7.77738452e-01
7.36573696e-01 2.50338107e-01 6.36179507e-01 -8.29461873e-01
-2.37583578e-01 -1.80353999e-01 -2.57705063e-01 -1.51584542e+00
-1.40369281e-01 -4.76532847e-01 1.68119162e-01 -1.54490662e+00
1.81794211e-01 -2.89681047e-01 -5.43066978e-01 6.00822091e-01
-5.27485251e-01 4.80015039e-01 1.64498001e-01 1.64397284e-01
-5.24118662e-01 9.06856656e-01 1.51484561e+00 -2.57304519e-01
-7.38460347e-02 -1.93289235e-01 -7.85423636e-01 5.98829091e-01
4.87084270e-01 -3.24760199e-01 -4.02285963e-01 -8.92454863e-01
-3.16274583e-01 3.09316553e-02 5.25076926e-01 -1.04261136e+00
4.83674586e-01 -1.74028426e-02 8.09486687e-01 -7.21892476e-01
5.90715945e-01 -8.98439288e-01 -1.97442234e-01 2.89971888e-01
1.15032412e-01 -2.70176642e-02 2.83881634e-01 6.74586773e-01
-4.97961581e-01 2.61959314e-01 1.23138046e+00 -4.88751903e-02
-8.24680567e-01 6.04956627e-01 -2.18157381e-01 -2.04521120e-01
6.73799813e-01 -2.82129407e-01 -5.40588915e-01 -6.10875070e-01
-4.33891475e-01 2.73215383e-01 4.33398515e-01 2.62963533e-01
9.94474769e-01 -1.02474260e+00 -9.70418453e-01 7.12049305e-01
-3.61087322e-01 2.71914005e-01 8.44752491e-01 9.12711084e-01
-8.94672036e-01 7.78266713e-02 -2.81452447e-01 -5.46852946e-01
-1.09264338e+00 5.66046715e-01 5.15940130e-01 9.81118008e-02
-1.06292713e+00 9.97598648e-01 9.10658181e-01 2.15600759e-01
2.10662782e-01 -3.10943395e-01 -1.48414478e-01 -3.39795589e-01
9.94973958e-01 4.77863401e-01 8.95257592e-02 -6.01920545e-01
-1.33244246e-02 6.25833333e-01 -3.18155229e-01 9.22209397e-02
1.56050897e+00 -5.30885458e-01 -5.10496855e-01 2.14338843e-02
1.30230963e+00 4.24241833e-02 -1.63080692e+00 -2.66439795e-01
-9.43630099e-01 -8.17897737e-01 6.63927615e-01 -8.94321680e-01
-1.53509223e+00 1.13773990e+00 8.43702674e-01 -1.53906733e-01
1.75094593e+00 -4.53534156e-01 1.13177752e+00 2.46927351e-01
6.08915016e-02 -5.90393424e-01 4.92920317e-02 5.06498575e-01
9.11455214e-01 -9.73617136e-01 2.19582543e-01 -4.76834953e-01
-3.37751418e-01 9.73515987e-01 8.06894362e-01 -3.16040158e-01
8.50213289e-01 7.96277449e-02 2.09348604e-01 -2.94345379e-01
-5.59003890e-01 -1.16768293e-01 3.52153063e-01 3.88103187e-01
-1.14929035e-01 -2.91638494e-01 2.28030920e-01 5.09452105e-01
1.21945664e-01 -3.48851383e-01 6.32741690e-01 5.10385334e-01
-7.51106143e-01 -3.45385104e-01 -5.74924588e-01 2.98226267e-01
-5.47485709e-01 -5.43023586e-01 2.74413139e-01 5.71815312e-01
2.78418630e-01 1.09075224e+00 1.10387817e-01 -3.21219325e-01
2.53577799e-01 -5.35370946e-01 2.23221362e-01 -2.29541913e-01
-4.94586974e-01 4.31525260e-01 -3.65693480e-01 -8.44036162e-01
-2.03263223e-01 -9.51235965e-02 -1.24345815e+00 -5.34428537e-01
-3.94805253e-01 1.94905341e-01 4.87152040e-01 5.22021294e-01
6.45166397e-01 8.29454601e-01 8.11096430e-01 -1.15121758e+00
-1.23424239e-01 -8.58146191e-01 -6.37431443e-01 1.68035135e-01
1.11640501e+00 -4.44619268e-01 -6.10633492e-01 -2.65999734e-02] | [10.983869552612305, -3.081284761428833] |
ca23d85a-b457-4185-9a49-139cf591ddea | fact-or-factitious-contextualized-opinion-1 | 2010.15296 | null | https://arxiv.org/abs/2010.15296v1 | https://arxiv.org/pdf/2010.15296v1.pdf | Fact or Factitious? Contextualized Opinion Spam Detection | In this paper we perform an analytic comparison of a number of techniques used to detect fake and deceptive online reviews. We apply a number machine learning approaches found to be effective, and introduce our own approach by fine-tuning state of the art contextualised embeddings. The results we obtain show the potential of contextualised embeddings for fake review detection, and lay the groundwork for future research in this area. | ['Andrew McCarren', 'Jennifer Foster', 'Kirils Sloka', 'Niall Walsh', 'Stefan Kennedy'] | 2020-10-29 | fact-or-factitious-contextualized-opinion | https://aclanthology.org/P19-2048 | https://aclanthology.org/P19-2048.pdf | acl-2019-7 | ['spam-detection'] | ['natural-language-processing'] | [-7.26902336e-02 3.09105635e-01 -5.91785312e-01 -4.30684179e-01
-4.85332221e-01 -4.54278409e-01 1.14539695e+00 4.73653257e-01
-3.61107111e-01 5.10595024e-01 2.99987555e-01 -5.63113511e-01
8.90766159e-02 -4.91236091e-01 -2.85262704e-01 -2.27502897e-01
-5.43905199e-02 2.29865804e-01 2.26794332e-01 -7.23393977e-01
1.17072010e+00 4.82433587e-01 -1.31046391e+00 5.07573426e-01
5.39664149e-01 6.95578992e-01 -7.55820692e-01 6.51489913e-01
7.71233998e-03 8.34926248e-01 -1.13888204e+00 -1.35630822e+00
-4.29402888e-02 -3.37358922e-01 -6.25968933e-01 1.65717915e-01
6.62466705e-01 -5.17408073e-01 -4.30216283e-01 9.62498248e-01
2.38857269e-01 -8.19874182e-02 7.72628427e-01 -1.32323694e+00
-1.65208530e+00 2.08334014e-01 -1.50958806e-01 7.36700892e-01
3.66732746e-01 7.59845972e-03 1.16262186e+00 -1.09735656e+00
6.84855163e-01 1.29754663e+00 8.00775409e-01 8.06522012e-01
-7.03035653e-01 -4.66093928e-01 1.19253933e-01 2.44350746e-01
-7.99216688e-01 -4.92673695e-01 7.77304232e-01 -2.54217505e-01
9.95377839e-01 -1.25912264e-01 6.07029140e-01 1.39242828e+00
4.72674042e-01 7.11363375e-01 1.49170578e+00 -6.78810537e-01
2.03049332e-02 8.93353045e-01 4.55824435e-01 7.20434070e-01
8.96758318e-01 2.23672315e-01 -6.21355534e-01 -6.38219178e-01
4.11840618e-01 -2.16568604e-01 6.42264783e-02 -2.69355327e-01
-6.65766418e-01 1.62830150e+00 4.05355364e-01 4.94395763e-01
-7.93138444e-02 3.31115246e-01 7.68505931e-01 5.61646700e-01
1.14618540e+00 9.11186099e-01 -3.43233436e-01 -2.32318372e-01
-9.76105094e-01 2.26593316e-01 1.01011336e+00 6.97732866e-01
4.25524294e-01 6.15695529e-02 2.91549623e-01 8.56172085e-01
4.44922209e-01 2.21204869e-02 6.89731121e-01 -6.24649048e-01
3.45299959e-01 3.24441284e-01 6.59368575e-01 -1.61137772e+00
2.55073488e-01 1.98152363e-01 1.13717623e-01 1.66396365e-01
1.31962165e-01 1.66042507e-01 -6.15353346e-01 5.87618530e-01
-1.54128987e-02 -8.18230733e-02 -6.36343062e-02 8.18961561e-01
6.76942706e-01 3.82212013e-01 -1.21523337e-02 -2.36145277e-02
1.14534795e+00 -1.38461268e+00 -9.86730695e-01 -4.16812897e-01
8.64029646e-01 -9.53895867e-01 1.00338769e+00 4.79243815e-01
-8.77427220e-01 -4.05453295e-02 -1.38106346e+00 -2.63095647e-01
-1.01342714e+00 -2.23998547e-01 9.71709311e-01 1.22475111e+00
-8.24300468e-01 8.51697445e-01 -3.01015049e-01 -2.25780264e-01
7.30646968e-01 7.37963393e-02 -3.87652993e-01 -1.95678234e-01
-1.24049366e+00 1.51493621e+00 -2.16097787e-01 1.26062259e-01
-6.88684344e-01 -1.05585396e-01 -7.50101209e-01 -3.38644922e-01
1.19700715e-01 -4.13720822e-03 1.40483403e+00 -1.24786139e+00
-1.26799238e+00 1.21432567e+00 4.51926962e-02 -5.27154028e-01
2.81576931e-01 -3.68130952e-01 -7.50129759e-01 4.19298977e-01
8.79458040e-02 2.09121704e-01 1.24809921e+00 -1.41419125e+00
-2.48826325e-01 -2.83720404e-01 3.08468621e-02 -2.67134517e-01
-7.43162632e-01 5.49349427e-01 3.00410777e-01 -7.48555660e-01
-4.27864313e-01 -7.59888351e-01 7.30695575e-02 -8.30394402e-02
9.51772183e-03 -3.17930341e-01 1.00293875e+00 -6.42019451e-01
1.33291340e+00 -1.79997468e+00 -2.95188099e-01 3.65121290e-02
3.83290529e-01 9.82735574e-01 -3.32951955e-02 1.02341771e+00
1.79097489e-01 8.58939946e-01 -1.41222343e-01 -7.26373553e-01
2.31191456e-01 2.02648610e-01 -4.79406983e-01 9.03697133e-01
5.12842834e-01 1.17406702e+00 -1.21715522e+00 -3.92140567e-01
2.21411437e-01 5.58097184e-01 -6.28003897e-03 -1.64395198e-01
1.90866262e-01 -4.42197382e-01 -3.20994735e-01 9.08481777e-01
5.02322793e-01 1.43106446e-01 5.54111749e-02 2.41851851e-01
1.36996105e-01 9.11294043e-01 -5.00946462e-01 5.85861623e-01
-3.63367230e-01 1.19019747e+00 2.21470103e-01 -8.82617414e-01
1.05270469e+00 2.06334770e-01 -3.16235363e-01 -4.98336881e-01
4.65651423e-01 4.04593974e-01 -2.19425425e-01 -5.16561031e-01
1.02463555e+00 -4.77166593e-01 -4.29389663e-02 7.55870581e-01
-6.34415373e-02 -4.71165895e-01 -3.46273005e-01 4.64339525e-01
1.04927242e+00 -3.63582641e-01 4.10471767e-01 -1.51495755e-01
5.71861267e-01 2.85822242e-01 -2.76838928e-01 8.55134666e-01
-1.15998900e+00 2.75842935e-01 8.22943926e-01 -8.32136452e-01
-1.24206948e+00 -6.52840376e-01 -9.06408504e-02 1.04253423e+00
1.86301053e-01 -5.98996520e-01 -4.99572396e-01 -1.22638845e+00
3.91721308e-01 6.98059678e-01 -1.13902235e+00 -3.38158667e-01
-4.47389096e-01 -6.98802471e-01 5.16638279e-01 2.00647578e-01
-3.94722857e-02 -1.24175823e+00 -3.88300568e-01 4.89496044e-04
3.67878854e-01 -9.01014805e-01 -2.99751639e-01 -5.57730868e-02
-1.00663340e+00 -1.20695543e+00 -1.49720684e-01 -7.99227417e-01
4.67672020e-01 7.44125903e-01 1.24840319e+00 8.03369880e-01
-2.18065664e-01 4.17534977e-01 -9.34872746e-01 -5.39750576e-01
-7.68624961e-01 -9.20588374e-02 4.66746800e-02 -2.74176419e-01
9.43962395e-01 -9.89406276e-03 -3.73475283e-01 2.54223287e-01
-1.08897710e+00 -1.01899707e+00 2.57241249e-01 1.00158179e+00
-7.92868733e-02 -1.16963260e-01 9.54810739e-01 -1.25869548e+00
1.60089362e+00 -7.15353847e-01 -2.96641260e-01 -1.36683643e-01
-1.02834606e+00 -3.47025901e-01 4.13745195e-01 -3.90187532e-01
-4.74057198e-01 -7.44299054e-01 3.68337482e-02 -2.56433159e-01
2.45270640e-01 2.26601735e-01 6.93352759e-01 -5.95458627e-01
8.70467424e-01 -1.69282123e-01 3.11352581e-01 -1.91770598e-01
6.14273190e-01 1.20152032e+00 -1.08392328e-01 1.66769512e-02
8.80187988e-01 5.52079082e-01 -4.48047191e-01 -8.71961892e-01
-9.16953504e-01 -5.16817927e-01 -5.93135118e-01 -2.05608010e-02
3.14569026e-01 -4.59194779e-01 -1.24979563e-01 2.25381732e-01
-1.16876292e+00 3.50099467e-02 -1.27000976e-02 3.83141413e-02
-1.79618642e-01 7.05729902e-01 -8.79694521e-01 -1.08844161e+00
-2.84010738e-01 -7.34946251e-01 8.88793468e-01 -2.55785406e-01
-4.32200253e-01 -1.42602813e+00 3.82106453e-01 6.68115914e-01
4.86788243e-01 1.16532102e-01 4.40262705e-01 -1.04333210e+00
-1.97548777e-01 -8.02098751e-01 -3.39042425e-01 8.76417279e-01
1.58759505e-02 2.92582005e-01 -9.59827781e-01 -1.81239039e-01
1.03006035e-01 -5.63852012e-01 9.73904312e-01 -1.60198703e-01
5.61753869e-01 -4.80607182e-01 -2.90745080e-01 -1.34252325e-01
1.34143555e+00 -3.10307324e-01 8.46895874e-01 6.74530685e-01
5.82022786e-01 8.81089509e-01 1.03973007e+00 1.28662795e-01
2.04717413e-01 4.27687287e-01 6.30284488e-01 4.67941821e-01
2.24565029e-01 -1.61426604e-01 3.82536381e-01 7.99517334e-01
4.37828064e-01 -2.18569934e-01 -6.78619385e-01 1.14113581e+00
-1.60380101e+00 -1.15205133e+00 -3.35942537e-01 1.74591351e+00
3.71974707e-01 1.90710321e-01 2.51667112e-01 2.63509780e-01
7.95720935e-01 5.04332542e-01 1.08267985e-01 -1.81058049e+00
1.74983323e-01 5.42098343e-01 6.45142794e-01 8.15750659e-01
-1.11276722e+00 1.45920813e+00 8.09835529e+00 4.74069864e-01
-6.63341403e-01 4.41766292e-01 2.67928779e-01 -2.96440441e-03
-3.66883278e-01 -1.35454103e-01 -5.06653488e-01 4.54176754e-01
1.35924292e+00 9.04970020e-02 3.97195637e-01 9.35953975e-01
2.01489091e-01 -1.91900566e-01 -7.91310191e-01 6.43559098e-01
6.59128010e-01 -1.32354212e+00 -9.96610746e-02 1.87274978e-01
7.81563640e-01 5.12055680e-02 1.39381409e-01 3.79276723e-01
1.56090811e-01 -1.14336288e+00 4.47410941e-01 7.67817572e-02
2.39479288e-01 -7.60758102e-01 1.13448107e+00 -9.43681970e-02
-1.49076968e-01 -1.12440042e-01 -6.08433783e-01 -5.51453710e-01
1.31975383e-01 5.34523487e-01 -9.09471512e-01 1.04734385e-02
5.58284402e-01 1.02047670e+00 -8.68409097e-01 6.43653989e-01
-7.33325899e-01 7.24188864e-01 2.54974723e-01 -7.49963462e-01
5.20917416e-01 -7.33408555e-02 4.12516505e-01 1.39950073e+00
-3.71635884e-01 -3.56659532e-01 -4.42179620e-01 4.19334322e-01
-2.35339329e-01 1.39070094e-01 -1.13581634e+00 -5.53987086e-01
4.41615641e-01 1.23117816e+00 -6.11538112e-01 -5.18765867e-01
-4.45921808e-01 1.23199499e+00 4.56033885e-01 -1.33237705e-01
-7.72957742e-01 -6.37127399e-01 8.15904260e-01 1.68539375e-01
6.08954966e-01 -3.06232631e-01 -4.56945866e-01 -1.32496560e+00
1.57578140e-01 -1.03089237e+00 4.87788282e-02 -6.09809756e-01
-1.87937891e+00 3.80796194e-01 -5.31024933e-02 -7.72525609e-01
-3.08191491e-04 -1.08359396e+00 -6.72252655e-01 6.18574798e-01
-1.79801047e+00 -8.12862039e-01 1.74950622e-02 -4.20365930e-02
3.71180862e-01 -1.05124496e-01 8.21400046e-01 7.23518729e-02
-2.17426613e-01 5.06237328e-01 2.19853312e-01 -1.48307845e-01
9.54071403e-01 -1.07611191e+00 7.71263659e-01 3.81941020e-01
2.32516646e-01 7.94809341e-01 7.30612278e-01 -7.51469493e-01
-1.01907158e+00 -7.37341762e-01 1.47972631e+00 -1.06317008e+00
1.37976980e+00 -4.95779306e-01 -8.61128569e-01 6.41305387e-01
4.41730499e-01 1.37542993e-01 8.74619067e-01 8.86266157e-02
-8.30902815e-01 3.50819826e-01 -1.57305253e+00 3.35543275e-01
6.02723777e-01 -9.20109332e-01 -1.05155873e+00 4.52191204e-01
4.46358174e-01 1.84852049e-01 -6.69251502e-01 -2.21707419e-01
8.02234173e-01 -1.24996126e+00 7.75480568e-01 -8.29099536e-01
9.38531876e-01 3.19825977e-01 1.57754213e-01 -1.48096156e+00
-1.69251755e-01 -4.90404934e-01 -2.68882751e-01 9.96733010e-01
2.37259462e-01 -1.11726201e+00 7.86154449e-01 2.40110129e-01
-4.69822213e-02 -1.12575281e+00 -8.37344408e-01 -7.83395112e-01
6.10586166e-01 -2.95318067e-01 1.31212771e-01 1.21272528e+00
3.35640371e-01 2.84360379e-01 -4.66812372e-01 -9.69471335e-02
2.63149172e-01 -4.60475117e-01 6.01241589e-01 -9.53869402e-01
1.93912536e-01 -5.65789700e-01 -7.24040151e-01 -7.25033879e-01
3.83610219e-01 -4.30869639e-01 -3.36494714e-01 -1.38264465e+00
-1.10443339e-01 -8.21539015e-02 -2.28111610e-01 5.86067513e-02
-1.42968506e-01 8.18648577e-01 1.07965380e-01 2.37359822e-01
-7.02381492e-01 5.45438588e-01 1.00091136e+00 -1.18616864e-01
8.42633992e-02 -3.88512373e-01 -1.11991680e+00 4.67185020e-01
1.18377650e+00 -7.52700686e-01 -2.08940461e-01 -2.65261114e-01
3.19040000e-01 -8.12873006e-01 5.85940838e-01 -3.09718460e-01
-3.61384243e-01 6.00089610e-04 9.47225913e-02 -1.96182653e-01
5.13089299e-01 -5.96990466e-01 -1.04916453e+00 2.43274689e-01
-2.92136580e-01 4.23637927e-01 2.16752082e-01 9.32108045e-01
-2.03231588e-01 -7.40909994e-01 7.35702217e-01 -2.85297036e-01
-1.71952367e-01 -3.19778472e-01 -8.31189632e-01 1.14522606e-01
9.55683589e-01 -2.96341211e-01 -7.80756831e-01 -7.16415882e-01
-2.97079444e-01 -1.59535497e-01 8.35992515e-01 7.61070728e-01
8.64175320e-01 -1.14676213e+00 -4.64303523e-01 -2.23103344e-01
4.48583424e-01 -1.29955876e+00 -5.58774769e-01 4.83094990e-01
-7.11381018e-01 5.47199726e-01 -1.09157786e-01 1.33129522e-01
-1.18551052e+00 9.62081552e-01 1.13033816e-01 -2.09554937e-02
-2.20111206e-01 6.48677528e-01 -7.92384207e-01 -4.40539509e-01
-2.04458460e-01 7.28036538e-02 -2.68212914e-01 2.29616567e-01
6.09733224e-01 5.89170575e-01 1.47142455e-01 -9.70355272e-01
-4.92413014e-01 1.02344766e-01 -3.12104642e-01 -1.03406228e-01
1.21704924e+00 -8.37088749e-02 -3.40503633e-01 5.56583524e-01
1.46580458e+00 3.39436769e-01 -4.42334235e-01 1.71630278e-01
1.13850631e-01 -1.06728649e+00 2.01970369e-01 -9.28443372e-01
-5.05210340e-01 8.05571437e-01 1.26327395e-01 9.69645560e-01
4.44483697e-01 6.31553214e-03 6.96788549e-01 2.99489111e-01
2.15208575e-01 -1.33380461e+00 4.25850868e-01 2.76993096e-01
6.61714613e-01 -1.65452015e+00 5.32331586e-01 -5.10149121e-01
-7.25748718e-01 1.26485121e+00 2.86220223e-01 -9.52671945e-01
7.12964594e-01 -3.74840140e-01 3.30507040e-01 -8.96452069e-01
-5.84752917e-01 1.39214024e-01 1.10691160e-01 7.54131198e-01
5.43203890e-01 5.94602292e-03 -1.03531837e+00 1.60580277e-01
-9.30498242e-02 -3.99666071e-01 1.18004978e+00 1.16307724e+00
-4.83380467e-01 -1.50928617e+00 -3.89834315e-01 6.28146529e-01
-7.98520923e-01 -3.38918358e-01 -1.42782080e+00 1.17741251e+00
-2.19595686e-01 1.45236874e+00 -3.33071530e-01 -4.16640341e-01
2.02592596e-01 2.54436195e-01 4.83737499e-01 -7.85028994e-01
-1.05127859e+00 -5.66183567e-01 6.16563380e-01 -3.74337405e-01
-3.66628498e-01 -6.25804245e-01 -3.68926764e-01 -6.24591470e-01
-8.91111851e-01 2.08035335e-01 8.87174785e-01 8.74318242e-01
4.11449343e-01 -7.26631954e-02 7.28384495e-01 -9.05159473e-01
-7.82099068e-01 -1.06484842e+00 -5.58793068e-01 5.11263549e-01
6.51270986e-01 -7.44257271e-01 -1.11623466e+00 -3.65156204e-01] | [8.15803050994873, 10.246224403381348] |
971e938a-d386-4389-9c73-845ceac8dc46 | the-politics-of-deceptive-borders-biomarkers | 1911.09156 | null | https://arxiv.org/abs/1911.09156v4 | https://arxiv.org/pdf/1911.09156v4.pdf | The politics of deceptive borders: 'biomarkers of deceit' and the case of iBorderCtrl | This paper critically examines a recently developed proposal for a border control system called iBorderCtrl, designed to detect deception based on facial recognition technology and the measurement of micro-expressions, termed 'biomarkers of deceit'. Funded under the European Commission's Horizon 2020 programme, we situate our analysis in the wider political economy of 'emotional AI' and the history of deception detection technologies. We then move on to interrogate the design of iBorderCtrl using publicly available documents and assess the assumptions and scientific validation underpinning the project design. Finally, drawing on a Bayesian analysis we outline statistical fallacies in the foundational premise of mass screening and argue that it is very unlikely that the model that iBorderCtrl provides for deception detection would work in practice. By interrogating actual systems in this way, we argue that we can begin to question the very premise of the development of data-driven systems, and emotional AI and deception detection in particular, pushing back on the assumption that these systems are fulfilling the tasks they claim to be attending to and instead ask what function such projects carry out in the creation of subjects and management of populations. This function is not merely technical but, rather, we argue, distinctly political and forms part of a mode of governance increasingly shaping life opportunities and fundamental rights. | ['Javier Sánchez-Monedero', 'Lina Dencik'] | 2019-11-20 | null | null | null | null | ['deception-detection'] | ['miscellaneous'] | [ 2.65877813e-01 5.42785883e-01 -8.74228328e-02 -5.19892693e-01
-1.98384330e-01 -6.87757432e-01 8.96320105e-01 -1.82418019e-01
-4.79398042e-01 4.86601859e-01 5.98118901e-01 -5.70648253e-01
-2.93618500e-01 -5.27793169e-01 -2.22709194e-01 -4.63382155e-01
1.22488342e-01 -9.40874368e-02 -3.85600448e-01 -3.34426343e-01
5.26029766e-01 6.17577493e-01 -1.37365520e+00 5.40395752e-02
3.95740300e-01 1.21430612e+00 -9.62630391e-01 7.34967291e-01
3.13805401e-01 1.13813698e+00 -7.70789385e-01 -8.82219791e-01
6.24671541e-02 -6.14157915e-01 -7.96557009e-01 -2.04085469e-01
1.70119718e-01 -7.02146292e-01 -1.41502529e-01 8.91170263e-01
3.09717447e-01 -3.64205301e-01 5.30739427e-01 -1.30030155e+00
-6.78970397e-01 -5.43271452e-02 -1.06020004e-01 1.69049844e-01
4.45280313e-01 4.30038095e-01 6.07856393e-01 -4.33979660e-01
5.56388259e-01 1.24772823e+00 8.99550140e-01 7.46258438e-01
-1.00062287e+00 -7.41219223e-01 -1.05668589e-01 -1.26497418e-01
-1.49040651e+00 -1.12801659e+00 4.49781209e-01 -7.69495189e-01
7.10376263e-01 6.80171788e-01 9.39955890e-01 1.36709726e+00
1.64369315e-01 3.82330388e-01 1.16657889e+00 -4.26834911e-01
2.64169335e-01 1.67686224e-01 8.98519624e-03 5.11463046e-01
6.43679619e-01 5.10218441e-01 -5.28604209e-01 -6.24394119e-01
5.34145772e-01 -5.32418489e-01 -9.72913131e-02 1.51751116e-01
-7.20983446e-01 9.39284980e-01 -1.72491550e-01 5.05539358e-01
-3.72461885e-01 3.30687612e-01 5.66407859e-01 2.91767538e-01
4.99311417e-01 2.42861494e-01 -5.56355193e-02 -6.67260110e-01
-8.79222512e-01 1.08907633e-01 8.18504632e-01 1.40514329e-01
3.12851109e-02 -5.10749035e-03 2.10090041e-01 3.51497859e-01
7.79272437e-01 4.61506039e-01 -1.46334823e-02 -1.23911250e+00
-2.88428843e-01 4.20410395e-01 3.64219248e-01 -1.28581107e+00
-3.45356107e-01 -2.97242571e-02 -3.25291157e-01 2.77382672e-01
3.27161551e-01 -6.57929957e-01 -4.85937119e-01 1.50585163e+00
1.85339227e-01 -3.51042837e-01 -4.37354259e-02 1.01924622e+00
3.92736435e-01 1.68733284e-01 3.54860157e-01 -1.66328892e-01
1.30010021e+00 3.19369435e-01 -7.17065036e-01 -1.00489326e-01
5.95567584e-01 -3.86435926e-01 5.50482333e-01 7.20330834e-01
-8.13026369e-01 1.60625190e-01 -1.08181798e+00 1.86708286e-01
-2.18505338e-01 -2.89761871e-01 1.09198368e+00 1.69123685e+00
-8.26317370e-01 3.05427343e-01 -6.41693830e-01 -6.60800457e-01
6.03935421e-01 -4.90328111e-02 -3.07224929e-01 2.83968896e-01
-1.01594687e+00 1.22843134e+00 -2.26911586e-02 7.19397724e-01
-5.93133569e-01 -1.18235357e-01 -4.73653555e-01 -3.48788232e-01
8.88527408e-02 -4.55980748e-01 7.65481412e-01 -1.53519583e+00
-1.44058979e+00 1.30269206e+00 3.18280369e-01 -2.47959316e-01
5.61487377e-01 -1.37005448e-01 -9.21834111e-01 1.66146785e-01
-2.86959350e-01 3.58375192e-01 5.73398650e-01 -1.16470397e+00
-3.65774959e-01 -5.74562311e-01 -4.45100898e-03 -2.65447438e-01
-2.06104517e-01 8.05423141e-01 6.01666331e-01 -1.39430165e-01
-3.24675530e-01 -7.94435203e-01 2.89348960e-01 8.06459561e-02
-3.37870792e-02 -9.94513184e-02 7.04965293e-01 -6.80264235e-01
1.14083195e+00 -2.26426220e+00 -5.05986273e-01 4.30216402e-01
3.05841058e-01 4.15874630e-01 2.86838979e-01 7.91112900e-01
1.20754845e-01 6.63319886e-01 -5.96531667e-02 2.14519545e-01
5.78538954e-01 4.61547077e-02 -3.95899326e-01 8.94548237e-01
1.69502169e-01 9.06253874e-01 -8.10162604e-01 -2.30659574e-01
-9.56796929e-02 5.56426644e-01 -2.96830893e-01 -2.15288967e-01
4.34906632e-01 1.00995637e-01 -4.14276153e-01 8.86504114e-01
4.88600701e-01 4.29978102e-01 4.18287307e-01 2.21073508e-01
-4.57171947e-01 -2.75655813e-03 -5.20417392e-01 1.02646685e+00
2.52580225e-01 7.89734364e-01 6.69924915e-01 -8.34690571e-01
1.06488359e+00 3.89060438e-01 2.56574601e-01 -6.72205627e-01
3.64128113e-01 3.86939645e-01 2.94553787e-01 -8.81629825e-01
3.16574693e-01 -8.03740740e-01 -3.24717492e-01 5.49353480e-01
-2.74433494e-01 -4.69962992e-02 -7.53465950e-01 -9.74868983e-02
1.16128135e+00 4.11487311e-01 1.46134943e-01 -2.87299961e-01
2.00234950e-01 -3.89971323e-02 6.08583033e-01 3.77947658e-01
-9.06351566e-01 1.81885436e-01 7.92591810e-01 -7.14993894e-01
-7.26232231e-01 -9.09112811e-01 -3.25325370e-01 7.44787455e-01
-2.72841960e-01 -2.31187135e-01 -6.79879069e-01 -7.18030035e-01
1.78257883e-01 8.85973394e-01 -7.51718879e-01 -3.48270148e-01
2.29048431e-01 -9.75249231e-01 1.39653575e+00 -8.43275860e-02
4.53703374e-01 -7.40934253e-01 -1.50916672e+00 -2.19131753e-01
2.98301931e-02 -7.37570286e-01 2.78973371e-01 -2.24963799e-01
-4.11079049e-01 -1.05136526e+00 -2.87207842e-01 7.14057386e-02
2.20619455e-01 -2.45584205e-01 8.12560558e-01 3.87518317e-01
-2.40148202e-01 8.04414690e-01 -4.41781104e-01 -6.97264135e-01
-5.14885366e-01 -4.49115992e-01 -5.53260045e-03 1.22199096e-01
9.20932472e-01 -7.62769222e-01 -7.94254363e-01 2.44337190e-02
-1.01125181e+00 -4.26155210e-01 6.15154028e-01 2.95734763e-01
-6.49984539e-01 -5.50288141e-01 1.01963508e+00 -6.77868366e-01
7.88802087e-01 -6.71627998e-01 -2.43226409e-01 2.06512257e-01
-4.55295086e-01 -6.19672239e-01 -1.71662346e-01 8.67256448e-02
-1.04246962e+00 -5.85501671e-01 -2.80147195e-01 -8.56965780e-02
-3.99741828e-01 4.53913957e-01 -1.10439539e-01 -2.86513031e-01
8.25677335e-01 -3.25932801e-02 4.71281141e-01 1.37681635e-02
1.67133249e-02 1.09355474e+00 3.86105090e-01 -6.02175653e-01
5.04760981e-01 7.49003768e-01 -1.12806372e-01 -1.08357060e+00
-3.78939688e-01 7.49659631e-03 -2.77356714e-01 -7.61783004e-01
7.06953824e-01 -7.22958326e-01 -1.30458736e+00 2.98922598e-01
-9.68523324e-01 -2.68349051e-01 -1.09520189e-01 3.58918786e-01
-3.26441258e-01 3.11136335e-01 -3.75963062e-01 -1.68670332e+00
-3.06087703e-01 -2.38183483e-01 7.92723894e-01 -2.07459584e-01
-7.68853486e-01 -8.74659717e-01 1.47239625e-01 7.79726565e-01
7.32922852e-01 9.57291007e-01 6.40764892e-01 -5.14290512e-01
2.22857650e-02 -5.25434196e-01 -1.48468450e-01 4.40581024e-01
-5.99615760e-02 2.82427847e-01 -1.25030172e+00 -3.06987949e-02
4.80472982e-01 -5.59552252e-01 2.01252431e-01 -7.76484162e-02
2.75788218e-01 -5.54771662e-01 2.97951829e-02 1.87216118e-01
1.36511338e+00 2.22367898e-01 1.10604715e+00 2.09081039e-01
1.54178604e-01 1.26876020e+00 4.59660411e-01 5.71524620e-01
3.79845381e-01 3.84700418e-01 2.76569158e-01 3.29331875e-01
3.95624638e-01 -1.73591778e-01 7.68512666e-01 3.30447882e-01
-1.71904027e-01 1.61522195e-01 -1.13020861e+00 5.08813381e-01
-1.65405166e+00 -1.21143305e+00 -3.42692435e-01 2.17225909e+00
4.41358030e-01 -3.88860069e-02 4.19431001e-01 -1.59639522e-01
5.00107527e-01 -5.25615141e-02 -1.70898080e-01 -1.16324782e+00
1.31654084e-01 -2.59626587e-03 2.28776947e-01 3.85461003e-01
-5.45906305e-01 6.11414194e-01 7.32436514e+00 2.71251738e-01
-9.92845416e-01 1.08735956e-01 9.37590957e-01 -2.20614642e-01
-4.44167584e-01 7.75785446e-02 -1.75709292e-01 4.40556556e-01
1.37775743e+00 -3.71526368e-02 4.23747241e-01 2.42231160e-01
5.35738051e-01 -5.72718143e-01 -9.16042268e-01 6.08786702e-01
2.98332304e-01 -8.40089619e-01 -4.40972030e-01 5.26661038e-01
7.53865242e-02 -1.12176828e-01 -7.30003864e-02 -3.90711427e-03
3.01024348e-01 -1.40247679e+00 1.09684193e+00 7.19259679e-01
6.15717888e-01 -7.35108674e-01 7.14055479e-01 2.81196952e-01
1.11072503e-01 -1.06306717e-01 -3.41606140e-02 -6.93676591e-01
-3.48724499e-02 4.43703115e-01 -3.20272863e-01 1.92094684e-01
5.07816136e-01 -7.52569064e-02 -2.25336254e-01 5.11666298e-01
-1.47166267e-01 8.34711909e-01 -3.31625283e-01 4.11213711e-02
1.91625267e-01 -1.51207909e-01 5.68966269e-01 1.26183760e+00
1.15825795e-01 2.78439403e-01 -6.86245263e-01 8.74146223e-01
4.31956917e-01 -3.07107687e-01 -7.57613182e-01 -5.23337066e-01
4.61704880e-01 1.36128557e+00 -5.45418680e-01 1.06841072e-01
-3.27559114e-01 7.02502549e-01 -3.04502398e-01 1.09122567e-01
-7.45319486e-01 -2.21225545e-01 8.03945899e-01 1.40476838e-01
3.93910408e-02 -1.00589693e-01 -3.54986042e-01 -1.04910004e+00
1.06363203e-02 -1.08390355e+00 2.00062156e-01 -6.00176334e-01
-9.78208601e-01 1.57745987e-01 -1.04211025e-01 -4.35508311e-01
-1.77797064e-01 -4.57410365e-01 -4.10422266e-01 7.65290797e-01
-1.12990367e+00 -1.44925797e+00 -3.54344212e-02 1.47299513e-01
-6.40903533e-01 2.92922139e-01 1.08859003e+00 -3.90097313e-02
-4.74721789e-01 4.10671830e-01 -1.24496721e-01 1.00964494e-01
3.64541441e-01 -6.06146336e-01 1.54258639e-01 6.88578069e-01
-1.88017696e-01 8.09039354e-01 7.18031466e-01 -6.55558050e-01
-1.55708039e+00 -3.59132469e-01 9.41587985e-01 -7.49290466e-01
7.43219972e-01 -6.69453621e-01 -3.97695750e-01 7.00320244e-01
3.11935872e-01 -2.37789407e-01 1.21885169e+00 3.54086816e-01
-2.21074596e-01 3.67422514e-02 -1.73423874e+00 3.29346180e-01
9.63793576e-01 -7.42064118e-01 -6.67060554e-01 -3.95142846e-02
-6.06719591e-02 3.65059674e-01 -9.55263376e-01 5.15133068e-02
1.35138023e+00 -1.31345761e+00 4.87454027e-01 -6.62582815e-01
2.68879205e-01 -1.72409099e-02 -1.43776536e-01 -7.98375368e-01
-1.03480652e-01 -1.00122797e+00 3.42656493e-01 1.60981774e+00
-1.92318782e-02 -1.10827005e+00 8.64632905e-01 1.24971318e+00
1.92987740e-01 -4.74579990e-01 -1.30420446e+00 -5.11992633e-01
3.36645812e-01 -6.26220286e-01 4.21827048e-01 1.25337517e+00
6.78341687e-01 1.04064852e-01 -1.96161449e-01 -3.42147634e-03
5.47770619e-01 -4.77680594e-01 7.46517360e-01 -1.25159943e+00
9.14918110e-02 -3.01371425e-01 -8.37360561e-01 -6.20280299e-03
-2.04372138e-01 -2.16866568e-01 -1.62160501e-01 -9.90743458e-01
8.10648426e-02 -2.04224288e-01 -3.98694165e-02 3.20185483e-01
4.01053339e-01 3.20303470e-01 2.13837549e-01 2.40377218e-01
-3.76530796e-01 2.70951629e-01 5.30695379e-01 3.15274984e-01
1.23129427e-01 -5.51985264e-01 -1.41536915e+00 7.05871880e-01
7.49011576e-01 -2.49624699e-01 -1.23313330e-01 -9.11747068e-02
9.15710807e-01 -1.48708418e-01 1.04700768e+00 -7.39069581e-01
1.22714024e-02 -4.67906594e-01 3.84459257e-01 1.23409487e-01
2.65461326e-01 -9.06804740e-01 5.20174384e-01 4.89502341e-01
-3.25647518e-02 -3.19400996e-01 3.68537396e-01 8.84629264e-02
1.97123557e-01 -2.90284455e-01 5.81269205e-01 -1.31411344e-01
-2.33730495e-01 -4.15493697e-01 -8.88852477e-01 -6.39787167e-02
9.08982217e-01 -6.40175104e-01 -5.35373807e-01 -6.14820063e-01
-4.88664597e-01 9.11562070e-02 7.99332857e-01 9.50751454e-02
4.39218909e-01 -1.13363564e+00 -7.27447271e-01 1.73391812e-02
2.34236456e-02 -7.73040116e-01 1.00203559e-01 1.14555848e+00
-5.61535239e-01 4.01428103e-01 -3.08508575e-01 1.48974191e-02
-9.59538698e-01 1.22859456e-01 6.23272240e-01 4.39185947e-01
-4.29566830e-01 5.88932812e-01 4.99656573e-02 -1.51009262e-01
-1.08505853e-01 1.58279523e-01 3.65417711e-02 2.20678240e-01
6.35921776e-01 2.62153625e-01 -2.22973853e-01 -8.44839036e-01
-5.99310458e-01 -8.58635902e-02 2.51180708e-01 -4.45975512e-01
1.54334009e+00 -2.61373490e-01 -3.73917162e-01 3.56360674e-01
8.93149495e-01 1.52718872e-01 -1.03813231e+00 7.28463173e-01
2.50953645e-01 -6.10432684e-01 2.96756595e-01 -1.30695820e+00
-5.94561994e-01 6.07814372e-01 5.10556698e-01 4.30515081e-01
8.62728000e-01 -3.09329599e-01 4.51917142e-01 7.90879223e-03
2.64645129e-01 -1.23079038e+00 -4.97143805e-01 -1.94728553e-01
1.07622266e+00 -6.55770183e-01 1.98663771e-01 -2.93991446e-01
-4.77462441e-01 1.09766924e+00 9.36185643e-02 -6.06044941e-02
2.56890148e-01 1.11483172e-01 1.78881332e-01 -7.04932272e-01
-7.10235715e-01 -4.59464407e-03 -3.00281584e-01 9.03714299e-01
4.48066801e-01 1.98882833e-01 -7.87576675e-01 8.03002834e-01
-1.50382504e-01 6.19351149e-01 3.74289155e-01 9.90963697e-01
-4.52511132e-01 -9.61276174e-01 -7.38448918e-01 3.58986378e-01
-8.81898224e-01 2.16323882e-01 -1.32185423e+00 1.04409981e+00
5.52853346e-01 1.14930284e+00 2.27456596e-02 -3.69954854e-01
1.43070027e-01 4.32344228e-01 5.75082958e-01 -1.02166049e-01
-9.33319449e-01 -1.95169240e-01 8.47974122e-01 -4.00375783e-01
-4.33053136e-01 -1.07698452e+00 -4.70982641e-01 -7.78616905e-01
-1.05217390e-01 1.54751688e-01 7.15864480e-01 9.13610458e-01
4.82179284e-01 -1.44152597e-01 4.76992786e-01 -3.56046587e-01
-4.22688454e-01 -8.40820670e-01 -5.96440136e-01 2.05338392e-02
2.73861349e-01 -3.13739568e-01 -4.47296053e-01 -2.38993272e-01] | [9.025540351867676, 6.324026107788086] |
de30de09-ac32-4568-9079-a4f8a132efa6 | cope-conditional-image-generation-using | 2104.05077 | null | https://arxiv.org/abs/2104.05077v3 | https://arxiv.org/pdf/2104.05077v3.pdf | CoPE: Conditional image generation using Polynomial Expansions | Generative modeling has evolved to a notable field of machine learning. Deep polynomial neural networks (PNNs) have demonstrated impressive results in unsupervised image generation, where the task is to map an input vector (i.e., noise) to a synthesized image. However, the success of PNNs has not been replicated in conditional generation tasks, such as super-resolution. Existing PNNs focus on single-variable polynomial expansions which do not fare well to two-variable inputs, i.e., the noise variable and the conditional variable. In this work, we introduce a general framework, called CoPE, that enables a polynomial expansion of two input variables and captures their auto- and cross-correlations. We exhibit how CoPE can be trivially augmented to accept an arbitrary number of input variables. CoPE is evaluated in five tasks (class-conditional generation, inverse problems, edges-to-image translation, image-to-image translation, attribute-guided generation) involving eight datasets. The thorough evaluation suggests that CoPE can be useful for tackling diverse conditional generation tasks. The source code of CoPE is available at \url{https://github.com/grigorisg9gr/polynomial_nets_for_conditional_generation}. | ['Yannis Panagakis', 'Markos Georgopoulos', 'Grigorios G Chrysos'] | 2021-04-11 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 6.68916464e-01 2.36052260e-01 1.13475986e-01 -2.07159951e-01
-9.63380694e-01 -3.39257598e-01 6.90006733e-01 -4.63494033e-01
1.36523306e-01 1.00386584e+00 -1.52605534e-01 -1.40388638e-01
4.86568548e-02 -9.99521673e-01 -8.42534006e-01 -9.76866543e-01
1.70482561e-01 5.46381533e-01 -2.52167612e-01 -1.81467533e-01
-1.00036167e-01 3.50259900e-01 -1.45704150e+00 2.63638377e-01
7.32225835e-01 8.76047850e-01 3.35964322e-01 9.31402802e-01
5.68416864e-02 4.03128028e-01 -6.47238731e-01 -4.40123498e-01
2.03742981e-01 -7.68692076e-01 -6.86003745e-01 -1.85466409e-02
2.66521156e-01 -4.66402061e-02 -1.30522296e-01 1.07832623e+00
5.97887099e-01 1.81961685e-01 8.32040787e-01 -1.58086753e+00
-1.23515999e+00 6.00462675e-01 -4.60127145e-01 -2.38416612e-01
2.69042522e-01 9.60665122e-02 1.01364875e+00 -1.11161542e+00
8.56087625e-01 1.33653128e+00 6.09215200e-01 6.50353849e-01
-1.79485607e+00 -5.31051278e-01 -3.75148654e-01 1.00221351e-01
-1.40254474e+00 -2.50083089e-01 8.52892995e-01 -5.63631117e-01
7.17923522e-01 2.57768273e-01 4.26776022e-01 1.56416023e+00
2.18448624e-01 5.71866095e-01 1.28773296e+00 -5.54419160e-01
3.04375216e-02 2.13806987e-01 -5.91260493e-01 5.50496042e-01
-7.20089674e-02 3.86327714e-01 -6.14656448e-01 5.17705232e-02
1.16570032e+00 -4.00241882e-01 -2.46682167e-01 -1.15099117e-01
-1.28651190e+00 1.07661712e+00 4.27562684e-01 7.66005740e-02
-3.42222005e-01 3.66506785e-01 -1.72755532e-02 3.18436801e-01
5.35597742e-01 6.49377644e-01 -3.84013593e-01 9.32648107e-02
-8.18258643e-01 5.11690438e-01 7.08634913e-01 1.24393189e+00
1.00356328e+00 4.96550083e-01 -4.34124261e-01 7.23728418e-01
-1.80883110e-01 6.21691287e-01 3.02708089e-01 -9.40024257e-01
1.32899776e-01 1.45802006e-01 -1.01452686e-01 -1.05294824e+00
-3.18991482e-01 -4.50428545e-01 -1.32418656e+00 4.51226503e-01
2.46148512e-01 -5.06097376e-01 -1.12088227e+00 2.00280952e+00
1.39916256e-01 1.99142829e-01 6.82015121e-02 7.38302350e-01
1.06334031e+00 6.76135838e-01 -1.04372025e-01 -7.86813870e-02
1.03181958e+00 -8.25974703e-01 -5.21345913e-01 -6.81093112e-02
1.10590108e-01 -9.04027283e-01 9.18848395e-01 3.68604243e-01
-1.23806965e+00 -7.32708335e-01 -7.03229666e-01 -1.38018370e-01
-3.96243900e-01 2.41514891e-01 7.05128431e-01 2.60881573e-01
-1.41911495e+00 5.74171066e-01 -6.24806523e-01 -8.81891176e-02
4.86463040e-01 2.51931697e-01 -4.77285087e-01 -5.85086606e-02
-1.29637313e+00 6.41846895e-01 3.69324028e-01 2.55352905e-04
-9.99878109e-01 -5.53696394e-01 -8.13344002e-01 9.21049062e-03
3.07921708e-01 -1.24226403e+00 1.06345332e+00 -1.21278298e+00
-1.42377055e+00 7.11067557e-01 -1.79460049e-01 -4.75836873e-01
4.88394469e-01 2.71425140e-03 -1.47598833e-01 6.65563494e-02
2.24837318e-01 1.16087210e+00 1.17856729e+00 -1.50159609e+00
-2.69344836e-01 -1.27957106e-01 2.28090631e-03 9.50326473e-02
3.31877731e-04 -2.17996947e-02 -3.34279060e-01 -8.59590888e-01
1.05846887e-02 -1.01489902e+00 -4.18010741e-01 -2.09466871e-02
-6.60660565e-01 -4.92762364e-02 5.95036685e-01 -5.53916395e-01
8.76341462e-01 -1.91114986e+00 6.01404965e-01 9.40196067e-02
2.14294076e-01 -7.89541379e-02 -1.86223000e-01 4.37311590e-01
-4.33420777e-01 2.65125126e-01 -5.20632029e-01 -4.52568620e-01
1.24739923e-01 3.04046154e-01 -2.68523604e-01 6.98862672e-02
6.94077730e-01 1.19431889e+00 -7.83083558e-01 -4.42764193e-01
6.02969676e-02 8.38907182e-01 -5.02837241e-01 1.83587477e-01
-5.27268112e-01 7.68007457e-01 -2.51988977e-01 6.06927097e-01
6.48813844e-01 -3.69703621e-01 -2.15695038e-01 -1.33037224e-01
1.60690770e-01 -2.40603283e-01 -1.18979359e+00 1.64644289e+00
-3.69230360e-01 8.67250204e-01 -9.05854926e-02 -8.73949468e-01
1.06186402e+00 4.79403526e-01 4.05310541e-01 -5.41209280e-01
8.45119078e-03 -1.48955761e-02 -2.37807289e-01 -4.34713960e-01
4.85322177e-01 -3.28999430e-01 -8.22690949e-02 1.77369043e-01
3.03820699e-01 -4.52687889e-01 2.67658591e-01 1.80710897e-01
8.18457365e-01 4.95873690e-01 2.67395914e-01 -3.14925537e-02
2.87806660e-01 -3.83768417e-02 4.03090477e-01 7.17890561e-01
3.24721217e-01 1.05071270e+00 7.71552980e-01 -1.31114796e-01
-1.34959793e+00 -1.35589671e+00 -4.52574193e-02 9.82671738e-01
-2.50938714e-01 -3.80479336e-01 -8.16468954e-01 -1.96095839e-01
-2.55582005e-01 7.86185980e-01 -7.05915034e-01 2.76379399e-02
-3.84458810e-01 -6.33952737e-01 5.38579822e-01 5.62270582e-01
3.22210222e-01 -1.26184893e+00 -3.38086575e-01 3.50701958e-02
-3.40477973e-01 -1.08753371e+00 -1.40037745e-01 1.98461577e-01
-5.63370645e-01 -5.99836290e-01 -7.92003453e-01 -7.77621269e-01
9.41822052e-01 -2.41580158e-01 1.39459145e+00 -2.34222233e-01
-4.21468943e-01 4.52528208e-01 -1.68307304e-01 -4.10287529e-01
-5.17951071e-01 -1.16368614e-01 -1.68668061e-01 -7.87286460e-02
-5.73085509e-02 -9.28750873e-01 -4.41452742e-01 1.64066166e-01
-1.15324724e+00 5.44681311e-01 7.99448133e-01 1.33790243e+00
9.91516113e-01 4.50115278e-02 3.93052608e-01 -9.40138876e-01
7.52328157e-01 -4.31312114e-01 -4.51142013e-01 1.48711652e-01
-3.88277620e-01 1.41695261e-01 7.43065894e-01 -5.91815710e-01
-1.05858696e+00 3.13840955e-01 -1.65943608e-01 -6.83073282e-01
-3.20716888e-01 7.05619693e-01 -1.48615226e-01 1.50907531e-01
8.79373372e-01 4.05779749e-01 -1.80960119e-01 -2.32485592e-01
5.62992811e-01 9.81182531e-02 8.52072060e-01 -7.15795100e-01
1.05329359e+00 3.01158845e-01 5.46406388e-01 -7.52366006e-01
-4.69313920e-01 1.08060710e-01 -6.74012065e-01 2.87246741e-02
9.44848120e-01 -7.93299854e-01 -2.41502538e-01 3.85561585e-01
-1.21119404e+00 -4.75485533e-01 -5.42890847e-01 1.61824107e-01
-8.86425197e-01 -2.23275125e-02 -4.28284764e-01 -5.98170578e-01
-2.97293395e-01 -1.15859497e+00 1.05191517e+00 3.25419992e-01
-1.23106837e-01 -1.10825038e+00 -1.50812328e-01 5.23554981e-02
5.80428898e-01 7.88472354e-01 9.11973596e-01 -2.07762226e-01
-7.65935481e-01 -8.05222392e-02 -1.49815589e-01 4.62229580e-01
-5.98608665e-02 2.12879434e-01 -9.57289755e-01 -7.73330033e-02
-2.12513819e-01 -4.19730633e-01 8.98275435e-01 5.72945595e-01
1.28381371e+00 -3.66124928e-01 -9.59923938e-02 9.33859110e-01
1.41797531e+00 -5.56633547e-02 8.87000620e-01 -4.25685979e-02
7.18653083e-01 3.90388638e-01 1.77844971e-01 4.83776122e-01
9.67801213e-02 6.52126014e-01 4.57578957e-01 -3.89789701e-01
-3.65211874e-01 -2.58132160e-01 1.07633218e-01 3.78012717e-01
-3.93345177e-01 -4.03252393e-01 -8.35819960e-01 4.24424142e-01
-1.62807345e+00 -9.96698201e-01 -2.58060068e-01 1.87153518e+00
8.57873142e-01 -1.61081523e-01 -1.71069935e-01 -1.58706874e-01
6.12957478e-01 -3.23691815e-02 -5.39362013e-01 -2.53879488e-01
-4.31301266e-01 6.84220612e-01 1.63048819e-01 5.44337213e-01
-1.07155120e+00 1.03262472e+00 5.83594370e+00 8.93832505e-01
-1.11736369e+00 7.27610663e-02 8.63101304e-01 1.40922502e-01
-6.45818412e-01 8.48782957e-02 -6.43577158e-01 2.05615103e-01
8.35439324e-01 -2.70131350e-01 5.79952657e-01 8.63344908e-01
5.91164827e-02 9.09987018e-02 -1.01339900e+00 8.91720653e-01
9.62408707e-02 -1.33465457e+00 2.44079858e-01 -1.25245735e-01
1.14154220e+00 -2.74398595e-01 5.37225068e-01 2.32996777e-01
3.64111185e-01 -1.53349960e+00 4.72787440e-01 7.14565217e-01
1.34421480e+00 -6.07944548e-01 4.34003830e-01 1.09942898e-01
-1.02375960e+00 2.75330365e-01 -3.17341983e-01 6.63499460e-02
2.74687737e-01 5.11552870e-01 -9.98114407e-01 6.39015794e-01
5.68498433e-01 5.10946870e-01 -4.69160348e-01 6.19840086e-01
-5.11073053e-01 4.71282363e-01 -3.36036049e-02 2.70325303e-01
8.28407332e-02 -3.21598411e-01 6.86241865e-01 1.20160437e+00
6.42050683e-01 -4.05453108e-02 2.53540725e-02 1.49635494e+00
-5.92129864e-02 -7.30608031e-02 -6.16731763e-01 -1.60393156e-02
1.81109339e-01 1.36426938e+00 -6.66738629e-01 -1.90337881e-01
-1.41898513e-01 1.18942881e+00 2.26978108e-01 7.65271068e-01
-9.62863445e-01 -3.78089756e-01 4.77452874e-01 2.55586095e-02
5.05474269e-01 -2.64252961e-01 -4.87481743e-01 -9.70776200e-01
-5.24042845e-02 -7.68003821e-01 3.20969895e-02 -1.30332756e+00
-1.22846317e+00 8.49671006e-01 1.41340181e-01 -1.01000965e+00
-6.95165396e-01 -5.61760128e-01 -6.02501869e-01 1.25729740e+00
-1.18555367e+00 -1.46000731e+00 -4.45779502e-01 8.17345619e-01
3.66604537e-01 -1.02275997e-01 1.14363527e+00 8.43719468e-02
-3.29229176e-01 6.50791109e-01 -5.52613698e-02 7.59180263e-02
7.35247195e-01 -1.41626000e+00 4.45184350e-01 9.40673828e-01
2.43010044e-01 5.48538268e-01 9.66696978e-01 -5.28610051e-01
-1.28140748e+00 -1.06682277e+00 7.73250997e-01 -3.43604654e-01
5.33314049e-01 -4.39914018e-01 -5.92236280e-01 8.33937585e-01
2.20266625e-01 1.87996551e-02 3.48612666e-01 -2.03583628e-01
-2.42257372e-01 6.32166043e-02 -9.72884059e-01 7.69688904e-01
9.66289878e-01 -3.28936458e-01 8.74752402e-02 4.02118027e-01
8.99536610e-01 -7.81271994e-01 -8.89215052e-01 3.84459734e-01
3.93817037e-01 -9.38781798e-01 1.11565685e+00 -5.94186544e-01
1.16814828e+00 -2.45214239e-01 -2.09274203e-01 -1.49714303e+00
-4.88711566e-01 -7.55343974e-01 -2.08527409e-02 1.17583442e+00
7.30711997e-01 -5.52081823e-01 6.10936403e-01 4.87597495e-01
-2.65620440e-01 -9.34262753e-01 -8.50170493e-01 -4.65797722e-01
1.56530291e-01 -4.33507264e-01 4.48899537e-01 8.28254223e-01
-6.12103581e-01 4.77768600e-01 -6.83574557e-01 1.05771497e-01
4.43276942e-01 1.07366294e-01 9.40810740e-01 -1.04158390e+00
-7.34325290e-01 -4.97969151e-01 -3.74405116e-01 -8.68837714e-01
3.81604477e-04 -9.19081748e-01 9.27116498e-02 -1.79867899e+00
1.02984034e-01 -2.64018089e-01 4.19409685e-02 5.00732958e-01
-1.86850011e-01 4.75928098e-01 1.68224931e-01 -1.78195983e-01
-1.69084743e-01 5.17717838e-01 1.68138635e+00 -5.60434721e-02
-1.01224951e-01 1.53190464e-01 -7.95057237e-01 4.44967359e-01
9.91976023e-01 -3.09475005e-01 -4.81898427e-01 -3.53402346e-01
3.74979496e-01 3.06388170e-01 7.29397237e-01 -9.52306688e-01
1.67804107e-01 -2.97733217e-01 6.81956410e-01 -4.19539243e-01
6.37819290e-01 -5.57521999e-01 7.79914975e-01 5.55644147e-02
-2.85941094e-01 2.27271035e-01 1.44107297e-01 3.64411026e-01
-2.84988075e-01 -1.53391689e-01 6.22383714e-01 -3.18701953e-01
-5.21828890e-01 4.60987806e-01 -9.22217369e-02 8.60552713e-02
8.19295049e-01 -2.08014503e-01 -3.47567588e-01 -6.42211676e-01
-1.06776822e+00 -1.35465994e-01 1.98094517e-01 2.33912349e-01
8.99140835e-01 -1.44594872e+00 -1.15981364e+00 3.13237786e-01
-4.17464189e-02 2.67067492e-01 3.90025616e-01 8.31007123e-01
-2.84510851e-01 1.88830212e-01 -3.25467944e-01 -8.13519120e-01
-1.20434189e+00 4.69388127e-01 1.90578163e-01 -3.72086883e-01
-5.03541470e-01 1.19616199e+00 3.82808447e-01 -2.96785921e-01
-5.82402125e-02 -1.42874092e-01 1.75434381e-01 -1.30886391e-01
2.82525569e-01 6.71107993e-02 -3.47251803e-01 -6.53970957e-01
1.70298979e-01 4.57818449e-01 2.21798554e-01 -2.25665659e-01
1.43525922e+00 7.32937753e-02 -2.12728932e-01 3.95740718e-01
1.12750602e+00 -2.31103495e-01 -1.33768952e+00 -2.27464870e-01
-5.35835743e-01 -2.84633577e-01 -1.88573718e-01 -8.42755914e-01
-1.22088790e+00 8.60850275e-01 3.57262224e-01 1.50692165e-01
1.44637454e+00 5.26937917e-02 3.66919547e-01 8.23534653e-02
2.62391001e-01 -7.94668078e-01 2.96382606e-01 5.26076734e-01
1.23914182e+00 -1.20185888e+00 -6.19173460e-02 -3.59592855e-01
-7.75358319e-01 1.13746309e+00 5.42688310e-01 -1.61267713e-01
5.66587090e-01 3.86472911e-01 -2.07270995e-01 -6.74618483e-02
-7.86466718e-01 -1.94470987e-01 4.04256314e-01 9.19323087e-01
4.56680954e-01 3.47663641e-01 -4.76434603e-02 3.80753934e-01
-6.40285671e-01 -3.69060226e-02 5.32572925e-01 4.50393915e-01
6.53254539e-02 -1.15611994e+00 -4.54916060e-01 4.70929861e-01
-3.32886189e-01 -3.78914028e-01 -3.86643976e-01 7.98877716e-01
4.06531304e-01 6.39847934e-01 -1.29520997e-01 -3.79384518e-01
3.19825411e-02 4.57751006e-02 5.35360575e-01 -5.81556618e-01
-2.26376534e-01 2.39616945e-01 -3.82903032e-02 -3.96350592e-01
-4.46569324e-01 -5.77375293e-01 -9.45101559e-01 -2.54771233e-01
-9.08920318e-02 -2.02619061e-01 5.80010414e-01 4.93366838e-01
2.52480626e-01 7.96297550e-01 3.43341678e-01 -1.15975535e+00
-3.27335775e-01 -9.84907508e-01 -3.76323491e-01 4.29783165e-01
1.97695315e-01 -5.08684695e-01 -2.89060503e-01 4.76760358e-01] | [11.537797927856445, -0.3645278513431549] |
ebd66b05-df6c-4849-9968-f766b80b663e | mirrorwic-on-eliciting-word-in-context | 2109.09237 | null | https://arxiv.org/abs/2109.09237v1 | https://arxiv.org/pdf/2109.09237v1.pdf | MirrorWiC: On Eliciting Word-in-Context Representations from Pretrained Language Models | Recent work indicated that pretrained language models (PLMs) such as BERT and RoBERTa can be transformed into effective sentence and word encoders even via simple self-supervised techniques. Inspired by this line of work, in this paper we propose a fully unsupervised approach to improving word-in-context (WiC) representations in PLMs, achieved via a simple and efficient WiC-targeted fine-tuning procedure: MirrorWiC. The proposed method leverages only raw texts sampled from Wikipedia, assuming no sense-annotated data, and learns context-aware word representations within a standard contrastive learning setup. We experiment with a series of standard and comprehensive WiC benchmarks across multiple languages. Our proposed fully unsupervised MirrorWiC models obtain substantial gains over off-the-shelf PLMs across all monolingual, multilingual and cross-lingual setups. Moreover, on some standard WiC benchmarks, MirrorWiC is even on-par with supervised models fine-tuned with in-task data and sense labels. | ['Ivan Vulić', 'Anna Korhonen', 'Nigel Collier', 'Fangyu Liu', 'Qianchu Liu'] | 2021-09-19 | null | https://aclanthology.org/2021.conll-1.44 | https://aclanthology.org/2021.conll-1.44.pdf | conll-emnlp-2021-11 | ['contextualised-word-representations'] | ['natural-language-processing'] | [ 4.30042952e-01 2.12134242e-01 -4.37940806e-01 -5.96489668e-01
-1.23556638e+00 -8.07562709e-01 8.67661953e-01 3.71091366e-01
-9.20314312e-01 7.60041535e-01 7.38187015e-01 -3.81945491e-01
1.76183194e-01 -6.34006262e-01 -8.62745345e-01 -1.35890931e-01
-1.75871309e-02 4.51462805e-01 -1.07289050e-02 -7.37249374e-01
2.58506257e-02 -1.44802034e-01 -1.10681951e+00 4.88110751e-01
1.05781686e+00 3.63057792e-01 5.97923517e-01 6.09893680e-01
-6.33681715e-01 8.13606679e-01 -2.17183486e-01 -6.69629931e-01
-5.24924472e-02 -1.49415031e-01 -9.87512112e-01 -2.66194791e-01
5.45393348e-01 3.08799148e-01 4.14103903e-02 7.80045748e-01
4.48771119e-01 1.20291658e-01 5.42616248e-01 -6.42970145e-01
-1.06951177e+00 1.44676113e+00 -2.87254542e-01 2.45465208e-02
3.23527902e-01 1.59583345e-01 1.42259109e+00 -1.11993825e+00
8.52212727e-01 1.37114561e+00 7.56488025e-01 7.65101194e-01
-1.47700071e+00 -6.76600039e-01 4.98110324e-01 1.35115966e-01
-1.22989511e+00 -6.54269040e-01 7.76139021e-01 -1.64860234e-01
1.68346286e+00 -7.32052252e-02 2.27519825e-01 1.53371716e+00
-2.08958045e-01 1.16208863e+00 1.11789668e+00 -1.02901435e+00
6.26797676e-02 2.07568988e-01 3.48816723e-01 6.45833850e-01
2.17150688e-01 -1.33434907e-01 -7.36419022e-01 1.33579299e-01
1.21059254e-01 -3.50355953e-01 -1.38540104e-01 -1.98481679e-01
-1.51098156e+00 8.91389549e-01 1.05368398e-01 6.02946877e-01
-3.63520421e-02 2.65609384e-01 6.83297455e-01 4.54580992e-01
8.43113601e-01 8.14883947e-01 -9.15898681e-01 -2.64801592e-01
-9.32829976e-01 6.80524185e-02 7.07177639e-01 1.13864398e+00
8.96623492e-01 1.32050112e-01 -3.22024912e-01 1.07763278e+00
1.74003422e-01 8.64382684e-01 9.23451841e-01 -5.92086077e-01
7.72548795e-01 4.67567712e-01 -2.28179380e-01 -5.25350869e-01
-3.72844338e-01 -6.86486125e-01 -6.91451073e-01 -3.77857119e-01
1.40015380e-02 -2.86039114e-01 -6.27214134e-01 2.04552650e+00
-2.00384229e-01 1.25833035e-01 3.59582007e-01 3.53643656e-01
7.41317213e-01 6.01531267e-01 3.36886406e-01 4.39835154e-02
1.24381924e+00 -1.15951717e+00 -7.06018150e-01 -8.59840035e-01
9.79813039e-01 -8.20742130e-01 1.56327486e+00 1.43207341e-01
-1.07381535e+00 -4.80020911e-01 -1.28415740e+00 -4.65936005e-01
-7.58376956e-01 1.61143407e-01 6.86766565e-01 3.62528145e-01
-1.11750317e+00 5.07101119e-01 -6.48162305e-01 -5.59083462e-01
1.29345387e-01 -1.48451358e-01 -4.46811706e-01 -2.90218145e-01
-1.48439956e+00 1.17938840e+00 6.41333580e-01 -2.51018941e-01
-9.01872158e-01 -1.01980829e+00 -1.26564765e+00 -1.36690378e-01
4.77865487e-01 -4.50990856e-01 1.35876167e+00 -8.28436255e-01
-1.45985818e+00 1.07764161e+00 -3.14817995e-01 -7.73307264e-01
4.19601724e-02 -7.21982658e-01 -5.33462107e-01 -2.60292560e-01
2.45787159e-01 7.82833517e-01 5.34473836e-01 -1.21056330e+00
-4.58692461e-01 7.42122158e-02 2.24373490e-01 1.29113376e-01
-3.37854296e-01 6.91393167e-02 -3.79117489e-01 -8.00462365e-01
-5.46825588e-01 -6.12296879e-01 -2.65434891e-01 -4.40692425e-01
-4.38763201e-01 -5.58096290e-01 2.55111575e-01 -3.23338807e-01
1.24108565e+00 -1.85672641e+00 7.53083080e-02 -2.06312746e-01
-9.88235548e-02 4.86311704e-01 -8.16178799e-01 8.64265680e-01
-7.37790093e-02 2.30909228e-01 -4.04851615e-01 -9.45089996e-01
2.79646784e-01 5.53041697e-01 -3.46941829e-01 1.51365370e-01
5.79358935e-01 1.12945735e+00 -1.31555653e+00 -5.21622539e-01
1.80815384e-01 4.41547602e-01 -7.38896072e-01 1.42205283e-01
-3.98729026e-01 -5.92222437e-04 -3.24922614e-02 2.15066373e-01
3.91141117e-01 -9.93690789e-02 5.33779323e-01 8.65239054e-02
-1.37798712e-01 8.65548253e-01 -8.00067902e-01 2.37187052e+00
-1.25157714e+00 5.07890105e-01 -2.28326604e-01 -1.21916032e+00
9.65876341e-01 3.31563950e-01 1.25562176e-02 -7.44150460e-01
-3.11297059e-01 4.59282637e-01 -2.17890322e-01 -1.91659763e-01
7.61794209e-01 -1.45150304e-01 -3.87783974e-01 5.76157570e-01
8.47741842e-01 -1.33797765e-01 3.55064720e-01 3.20553750e-01
9.89422917e-01 3.94910872e-01 6.41374588e-01 -7.71526754e-01
7.40991116e-01 -5.45816030e-03 4.08245772e-01 7.73671389e-01
-1.15713425e-01 3.90665889e-01 1.58918485e-01 -2.07964882e-01
-7.99575210e-01 -1.04772258e+00 -5.98995686e-02 1.67971873e+00
-3.06503177e-01 -8.07488143e-01 -7.39912629e-01 -8.17426682e-01
-2.46240869e-02 1.11807501e+00 -6.65388286e-01 -2.87472289e-02
-7.94027090e-01 -3.71433467e-01 8.64040911e-01 5.48812568e-01
2.07393110e-01 -1.30354059e+00 -5.44981882e-02 5.23762405e-01
-1.30407333e-01 -1.37751698e+00 -7.09689677e-01 4.93271470e-01
-5.01428604e-01 -8.55328023e-01 -4.37304318e-01 -1.06330872e+00
2.65792429e-01 2.54025757e-01 1.74768019e+00 -1.57876626e-01
-6.60005212e-02 3.03185463e-01 -5.66272199e-01 -4.86101389e-01
-4.79719847e-01 6.96951747e-01 1.26005128e-01 -2.07548648e-01
6.35822296e-01 -5.09582937e-01 -3.66654545e-01 -4.02445585e-01
-7.71516085e-01 9.82873812e-02 5.21398127e-01 8.32197249e-01
5.23865163e-01 -6.51425838e-01 8.94842923e-01 -1.43277955e+00
7.05687642e-01 -4.58994269e-01 -3.43048215e-01 4.69981343e-01
-7.57273316e-01 6.47345781e-01 8.44473779e-01 -2.40509406e-01
-1.23087347e+00 -3.55461575e-02 -4.02017802e-01 2.85793990e-01
-7.65909180e-02 6.66875184e-01 -9.35445428e-02 3.71964514e-01
7.50524700e-01 3.38249505e-01 -2.60791898e-01 -7.17151642e-01
1.04963148e+00 6.93427265e-01 7.24894881e-01 -9.82003868e-01
8.28034639e-01 1.62864864e-01 -6.70912683e-01 -6.28471017e-01
-1.39888287e+00 -6.25655234e-01 -8.27042758e-01 2.04507291e-01
7.94717193e-01 -1.26816356e+00 -2.00456962e-01 2.06627637e-01
-1.24941552e+00 -7.79425263e-01 -3.76337439e-01 3.05114120e-01
-3.31513673e-01 2.87180603e-01 -6.02863610e-01 -5.79966128e-01
-7.50260890e-01 -9.17519450e-01 1.29711938e+00 -1.92335218e-01
-4.01122451e-01 -1.61324406e+00 4.65418518e-01 8.10529590e-02
7.65293479e-01 -1.03466757e-01 1.02004361e+00 -6.93116486e-01
6.08489336e-03 -5.67834079e-02 -6.14789538e-02 6.00391388e-01
1.39999464e-01 -3.41256648e-01 -1.10351157e+00 -3.67177576e-01
-6.34154916e-01 -7.73604214e-01 1.11605763e+00 -4.99246120e-02
9.02811944e-01 -3.30333441e-01 -1.76221468e-02 7.52742112e-01
1.70220840e+00 -4.64320749e-01 1.81298301e-01 3.81415099e-01
5.23970008e-01 3.45174342e-01 3.29871982e-01 1.82395250e-01
7.75815666e-01 4.91308004e-01 1.33123383e-01 -3.04572191e-02
-3.89640480e-01 -6.79096222e-01 6.97661698e-01 1.54940128e+00
1.17800869e-01 -1.39511168e-01 -9.45989668e-01 1.04494226e+00
-1.63839281e+00 -6.92312837e-01 5.57267815e-02 1.85963762e+00
1.63218844e+00 1.94221750e-01 -3.34962726e-01 -1.56459793e-01
2.81470448e-01 6.13942504e-01 -2.16632858e-01 -6.48333490e-01
-5.04481137e-01 7.92801440e-01 7.06493676e-01 8.69413257e-01
-1.19086814e+00 1.54138362e+00 6.09330559e+00 1.08907700e+00
-8.89364123e-01 4.85691547e-01 1.66361555e-01 -7.87901655e-02
-8.63021493e-01 -5.06498367e-02 -9.82707858e-01 1.13661900e-01
1.16025519e+00 -3.13883215e-01 4.22856927e-01 8.98243546e-01
1.05748223e-02 3.06820273e-01 -1.17369437e+00 8.56492937e-01
2.94866771e-01 -1.50610554e+00 -8.64885822e-02 -3.89514297e-01
9.52664196e-01 6.37578249e-01 -1.15156241e-01 1.02133310e+00
9.48144555e-01 -8.95288885e-01 6.88085556e-01 2.45292187e-01
1.16917825e+00 -7.27163494e-01 8.15548480e-01 3.02411109e-01
-1.14033175e+00 4.55386341e-01 -4.86325741e-01 -1.88324079e-01
9.64418948e-02 5.55189192e-01 -6.28758729e-01 4.74571735e-01
3.17996651e-01 1.00057399e+00 -9.28180099e-01 2.93854266e-01
-5.30946791e-01 1.09244871e+00 -3.69008146e-02 -2.02841744e-01
5.93959689e-01 2.37524271e-01 2.66438603e-01 2.05755925e+00
-1.81000419e-02 -4.15779233e-01 2.73814112e-01 6.27621531e-01
-3.35672587e-01 4.77282494e-01 -6.05030358e-01 -7.19875693e-02
5.83581924e-01 1.25957751e+00 -6.17394522e-02 -5.81000865e-01
-6.19518697e-01 1.01896751e+00 1.01036477e+00 2.95984954e-01
-3.89546514e-01 -4.09438580e-01 9.02025461e-01 -1.75811619e-01
2.79787838e-01 -3.94193888e-01 -3.16340774e-01 -1.41303563e+00
-1.22254096e-01 -8.59342039e-01 1.68460265e-01 -4.04700905e-01
-1.64558208e+00 7.29779661e-01 -1.31947607e-01 -9.38095629e-01
-4.94753093e-01 -8.60058963e-01 -5.76168239e-01 8.86822939e-01
-2.27534890e+00 -1.43010855e+00 2.30341151e-01 5.94736457e-01
8.75262916e-01 -4.12245899e-01 1.28128815e+00 7.32373968e-02
-4.37614769e-01 8.98418844e-01 2.44843215e-01 3.45751852e-01
1.04172182e+00 -1.72552299e+00 8.89586687e-01 9.74805593e-01
5.47472775e-01 9.65406239e-01 3.99209082e-01 -4.61987108e-01
-1.35111678e+00 -1.46002805e+00 1.62022734e+00 -4.31862354e-01
1.11920404e+00 -9.23615217e-01 -7.61880696e-01 7.42023647e-01
8.66670609e-01 8.47273022e-02 8.69330347e-01 7.83404231e-01
-8.50051343e-01 -8.38150010e-02 -6.21826410e-01 6.66340828e-01
1.24902093e+00 -1.02111900e+00 -9.93615031e-01 5.22949159e-01
1.24070966e+00 -4.79841195e-02 -7.40996003e-01 8.39438066e-02
3.18817139e-01 -3.73709410e-01 9.48857546e-01 -9.46230829e-01
5.96151292e-01 8.46921578e-02 -4.17014509e-01 -1.81569946e+00
-3.27625751e-01 -7.09688365e-01 8.80207494e-02 1.50551128e+00
8.33284378e-01 -4.89383429e-01 2.41207302e-01 -1.66115284e-01
-4.30920124e-01 -5.97842813e-01 -7.16892421e-01 -8.70099127e-01
7.63685405e-01 -1.06648934e+00 5.29934466e-01 9.70114946e-01
2.45393246e-01 7.07327068e-01 -2.22346067e-01 -1.22905284e-01
6.40054226e-01 -1.27699375e-01 3.54370385e-01 -9.07802820e-01
-3.60323638e-01 -3.46822977e-01 -2.48603057e-03 -9.37667191e-01
7.75336742e-01 -1.62534094e+00 3.97387177e-01 -1.51113248e+00
1.55094400e-01 -3.49099129e-01 -5.13733566e-01 6.88887000e-01
-5.39451063e-01 1.61330670e-01 3.23809862e-01 -1.13196172e-01
-8.92259836e-01 5.05952954e-01 9.06728923e-01 -3.56252134e-01
2.37398773e-01 -5.62614501e-01 -8.77304196e-01 6.97204053e-01
7.56870687e-01 -3.39625448e-01 -4.56158847e-01 -1.08308733e+00
5.38608372e-01 -4.77159679e-01 -8.50043539e-03 -7.53621399e-01
7.22168908e-02 7.19823083e-03 -1.31097198e-01 -3.37402284e-01
-3.25156264e-02 -3.64781529e-01 -8.77869964e-01 1.37128934e-01
-8.59670997e-01 6.24236055e-02 1.97918519e-01 3.40058863e-01
-3.11880887e-01 -3.95771682e-01 5.77051818e-01 -3.63777459e-01
-9.60851967e-01 -1.35976285e-01 -3.02676171e-01 8.50868344e-01
3.62228870e-01 6.17372453e-01 -4.49140251e-01 -3.02166969e-01
-2.95278162e-01 2.63078690e-01 6.23450950e-02 6.95975602e-01
3.66259813e-01 -1.41951156e+00 -1.08583403e+00 2.10666671e-01
7.35720098e-01 -1.77488446e-01 -4.39182892e-02 3.75523120e-01
-2.53821224e-01 8.80355477e-01 3.51685196e-01 -2.66036987e-01
-6.57702565e-01 5.76550364e-01 -2.81118378e-02 -7.26409972e-01
-4.62489575e-01 1.05793858e+00 5.36773354e-02 -1.15368783e+00
1.15564242e-01 -7.00420201e-01 -1.60349309e-01 -1.26248961e-02
6.21891797e-01 -1.20600894e-01 2.37135664e-01 -5.47505498e-01
-4.60456669e-01 3.60929132e-01 -2.56179273e-01 -1.34015471e-01
1.38028622e+00 -1.77281305e-01 1.58144176e-01 6.80439591e-01
1.39221299e+00 2.80606002e-01 -1.01187265e+00 -8.24896097e-01
5.40493488e-01 2.31248215e-01 6.61699548e-02 -1.02298319e+00
-6.61238790e-01 1.01805937e+00 1.94765970e-01 -3.24845970e-01
8.72225404e-01 3.82480472e-02 9.12760377e-01 7.57220924e-01
5.03251314e-01 -1.32580829e+00 -8.74898881e-02 1.03670049e+00
9.09473538e-01 -1.39866257e+00 -3.94847661e-01 -5.44924624e-02
-8.87356281e-01 8.54410231e-01 3.41703206e-01 -4.74071294e-01
5.56560636e-01 4.99316782e-01 1.75332546e-01 1.37107700e-01
-1.28676987e+00 -6.52467728e-01 3.38953316e-01 7.44990647e-01
1.11810303e+00 3.69556844e-01 -4.67374951e-01 8.88031244e-01
-4.19834077e-01 -1.03197053e-01 4.43059921e-01 6.46254241e-01
-4.29830641e-01 -1.43055654e+00 2.77573496e-01 5.79053648e-02
-3.91502261e-01 -8.13304484e-01 -1.22059032e-01 8.55709255e-01
1.39679879e-01 1.05920529e+00 -7.16019869e-02 -1.89607799e-01
2.91048139e-01 4.67071712e-01 3.01487744e-01 -1.05678129e+00
-6.54953063e-01 -1.53270945e-01 4.98062998e-01 -6.55660927e-01
-5.33563733e-01 -5.68189502e-01 -1.14493799e+00 2.20780373e-01
6.97551444e-02 8.74768123e-02 5.60760081e-01 1.21248794e+00
2.31676906e-01 5.67640483e-01 3.74808252e-01 -5.18992186e-01
-6.28718376e-01 -1.50851786e+00 -2.42859855e-01 6.81420982e-01
2.98959374e-01 -4.52854842e-01 -1.65435046e-01 1.12599961e-01] | [10.875801086425781, 9.536581993103027] |
7ba6769a-3117-4da0-a38c-b6bf6c154607 | the-effect-of-counterfactuals-on-reading | 2304.00487 | null | https://arxiv.org/abs/2304.00487v1 | https://arxiv.org/pdf/2304.00487v1.pdf | The Effect of Counterfactuals on Reading Chest X-rays | This study evaluates the effect of counterfactual explanations on the interpretation of chest X-rays. We conduct a reader study with two radiologists assessing 240 chest X-ray predictions to rate their confidence that the model's prediction is correct using a 5 point scale. Half of the predictions are false positives. Each prediction is explained twice, once using traditional attribution methods and once with a counterfactual explanation. The overall results indicate that counterfactual explanations allow a radiologist to have more confidence in true positive predictions compared to traditional approaches (0.15$\pm$0.95 with p=0.01) with only a small increase in false positive predictions (0.04$\pm$1.06 with p=0.57). We observe the specific prediction tasks of Mass and Atelectasis appear to benefit the most compared to other tasks. | ['Akshay Chaudhari', 'Matthew Lungren', 'Anuj Pareek', 'Evan Zucker', 'Sovann En', 'Rupert Brooks', 'Joseph Paul Cohen'] | 2023-04-02 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 2.34008178e-01 1.18548381e+00 -7.33608067e-01 -5.09336174e-01
-7.88789392e-01 -3.77650172e-01 5.25238931e-01 5.73322415e-01
-4.84888911e-01 1.19766736e+00 5.31498849e-01 -1.07287407e+00
-2.67513663e-01 -4.33333755e-01 -7.51774728e-01 -2.92105436e-01
2.88865399e-02 5.60144007e-01 4.47238907e-02 2.25942895e-01
6.18335485e-01 1.69944577e-02 -8.89297485e-01 5.26989043e-01
6.53796911e-01 6.51736736e-01 -1.52554540e-02 7.24937499e-01
2.30217516e-01 1.13122869e+00 -5.14506102e-01 -7.56092787e-01
5.21057099e-02 -4.76919085e-01 -7.35555649e-01 -1.06840633e-01
2.15996206e-01 -4.88526255e-01 -2.14660555e-01 6.94302142e-01
4.53618228e-01 3.90864126e-02 9.12503660e-01 -1.23739707e+00
-6.15014315e-01 8.57189715e-01 -6.89753532e-01 5.66695452e-01
6.07976317e-01 2.70028532e-01 7.31969118e-01 -2.92543888e-01
5.79930305e-01 9.47095156e-01 7.95627952e-01 5.07642984e-01
-1.36889791e+00 -9.93190229e-01 1.26031071e-01 -1.57664910e-01
-8.82686257e-01 -1.70320198e-02 2.04236686e-01 -4.94700998e-01
6.74767256e-01 6.14092469e-01 7.10798323e-01 9.30128992e-01
1.08143497e+00 -5.96088776e-03 1.50609016e+00 -2.14111909e-01
2.46665344e-01 2.40456358e-01 3.19227904e-01 4.81217206e-01
9.19643223e-01 5.87147176e-01 -2.39658490e-01 -7.97311068e-01
7.57031500e-01 1.93655401e-01 -3.55019391e-01 -5.29993931e-03
-1.50514078e+00 9.03896868e-01 5.92295825e-01 -5.36079407e-02
-6.94604993e-01 2.40906045e-01 1.83541253e-01 -5.48222065e-02
1.50529683e-01 7.51263857e-01 -2.39216819e-01 5.08407503e-02
-7.08100677e-01 4.68332976e-01 8.16295087e-01 3.00628185e-01
-1.08187407e-01 -1.17001578e-01 -2.84213543e-01 4.24809068e-01
1.88795671e-01 6.40082955e-01 8.54696095e-01 -1.08556092e+00
2.41259396e-01 1.09505346e-02 5.77459872e-01 -9.02010679e-01
-6.73781097e-01 -6.67564631e-01 -3.97180200e-01 3.49545836e-01
5.12570322e-01 -1.85233504e-01 -7.09361792e-01 1.64522171e+00
-9.54336450e-02 -2.51823157e-01 -2.96899304e-02 7.05808222e-01
3.03722173e-01 3.02725490e-02 6.99459851e-01 -6.75710320e-01
1.57205510e+00 -4.71519113e-01 -8.03556800e-01 -3.39833468e-01
7.74917066e-01 -1.01854801e+00 1.17662477e+00 3.51866812e-01
-1.12675202e+00 -2.71966249e-01 -7.77391076e-01 5.47075629e-01
4.33450103e-01 -1.97945029e-01 5.10809362e-01 8.21202815e-01
-6.21521354e-01 7.40844727e-01 -6.99494302e-01 4.15948406e-02
5.07636726e-01 1.79773644e-01 2.59183254e-03 1.61500305e-01
-9.13455725e-01 1.17585576e+00 3.25954884e-01 -6.55953467e-01
-4.91435558e-01 -9.70013618e-01 -5.71099937e-01 1.72004253e-01
3.75260949e-01 -9.59403813e-01 1.84616077e+00 -8.13553452e-01
-6.74698949e-01 9.12850022e-01 -1.53026029e-01 -9.34996367e-01
9.14614737e-01 -1.88204467e-01 -4.47495818e-01 2.16086850e-01
7.26669073e-01 6.22328222e-01 3.50003541e-01 -1.35300183e+00
-7.57371724e-01 -1.88809246e-01 -4.30982038e-02 3.74845415e-01
4.42004383e-01 -1.56062663e-01 7.12381005e-01 -8.34832788e-01
3.70275795e-01 -1.17643249e+00 -4.74296838e-01 -3.97193851e-03
-5.17354965e-01 -1.47092164e-01 9.91441086e-02 -4.90070164e-01
1.02241564e+00 -1.71232498e+00 -1.26967573e+00 -8.58654734e-03
5.78517377e-01 -3.84690106e-01 8.56280148e-01 -2.72951946e-02
-7.14275718e-01 4.55890387e-01 -2.78814912e-01 3.47290158e-01
-2.03930482e-01 1.20968871e-01 -5.63201547e-01 3.03040087e-01
-1.49673283e-01 8.19879293e-01 -9.10823226e-01 -6.34583235e-01
2.76133418e-01 1.38570229e-03 -3.13306242e-01 -9.31075290e-02
2.26951420e-01 1.64166968e-02 -1.84981361e-01 1.92096233e-01
5.64661324e-01 -9.12536800e-01 3.34940523e-01 2.11356938e-01
-1.18636414e-01 6.84872031e-01 -5.58378398e-01 8.86174798e-01
-2.60006666e-01 4.73667502e-01 -5.92562854e-01 -3.42098922e-01
5.46583593e-01 5.98264754e-01 7.10917637e-02 -2.93225765e-01
6.11722888e-03 2.45310023e-01 5.47048151e-01 -2.31126010e-01
2.95293927e-01 -1.41053081e+00 5.60702421e-02 1.04970360e+00
-7.66371489e-01 -2.24842414e-01 -6.32567465e-01 4.15997118e-01
1.02161133e+00 -5.66902339e-01 8.94977391e-01 -4.97181326e-01
-1.70044437e-01 4.88186806e-01 4.12566364e-01 1.42982864e+00
-3.19267929e-01 7.55644560e-01 5.16089797e-01 -4.27897185e-01
-7.28974581e-01 -1.44244206e+00 -3.90476972e-01 5.96060276e-01
7.44502023e-02 3.04389864e-01 -5.02486944e-01 -1.07635033e+00
7.81497434e-02 2.04632998e+00 -8.55425298e-01 -3.45943093e-01
-1.12253144e-01 -8.31459582e-01 3.49525481e-01 8.16986501e-01
1.62686214e-01 -8.76946867e-01 -1.29328573e+00 -1.42067686e-01
-4.51604009e-01 -5.05243123e-01 -3.93320650e-01 8.23840350e-02
-1.20672190e+00 -1.13291919e+00 -6.83759212e-01 2.08933204e-01
6.32278383e-01 3.96852463e-01 1.27313697e+00 3.38055134e-01
-4.25447971e-02 4.77346331e-01 -1.33624122e-01 -8.49798918e-01
-6.02320790e-01 -5.80064535e-01 7.20365420e-02 -1.00594449e+00
3.93982112e-01 -6.71756342e-02 -9.90540445e-01 3.10403794e-01
-5.40419817e-01 2.34773591e-01 8.09797883e-01 9.26102400e-01
6.18866324e-01 -2.27604195e-01 5.06916046e-01 -1.46285868e+00
7.01423109e-01 -7.94104159e-01 1.66406706e-02 2.23088339e-02
-1.29891622e+00 -2.18797728e-01 3.99778813e-01 -4.97418672e-01
-1.46434927e+00 -3.99215966e-01 2.98491955e-01 -1.71975791e-02
-3.05495799e-01 2.93933630e-01 7.63409615e-01 3.63624781e-01
1.14221871e+00 -2.13127837e-01 -2.71259435e-02 3.36057208e-02
-1.25824317e-01 1.46185175e-01 5.57683051e-01 -1.09966278e-01
4.87641990e-01 4.92874175e-01 -1.59669533e-01 -1.37888342e-01
-1.22438717e+00 -5.37351929e-02 1.20616145e-02 -2.31513549e-02
9.22194898e-01 -6.92558289e-01 -6.26880407e-01 -7.14367449e-01
-8.05004120e-01 -1.69598758e-01 -5.97643375e-01 1.41921854e+00
-7.29635954e-01 2.24371180e-01 -3.00111562e-01 -7.00936854e-01
-1.50552213e-01 -9.82956648e-01 4.72749829e-01 1.88203827e-02
-1.14100790e+00 -8.60735953e-01 -1.92474481e-02 5.85522175e-01
1.65252507e-01 2.95779675e-01 9.15044725e-01 -1.03998566e+00
-1.53220892e-01 -3.53073716e-01 -3.19598407e-01 -3.16202968e-01
2.48987004e-01 -4.32931304e-01 -8.00594151e-01 5.04116900e-02
6.64773822e-01 -2.12490454e-01 5.62444150e-01 9.10899222e-01
1.46538091e+00 -6.01417482e-01 -7.87249386e-01 -7.44600296e-02
1.18256760e+00 6.49860322e-01 3.94264579e-01 2.53011912e-01
-5.05132042e-03 4.66240108e-01 1.00120783e+00 4.53054339e-01
3.23893607e-01 2.56916434e-01 3.47098947e-01 -9.14538801e-02
7.35134780e-02 -4.85258192e-01 -3.56429875e-01 -1.38697758e-01
-8.18762258e-02 -5.00462614e-02 -1.23666000e+00 3.23736548e-01
-1.45588195e+00 -9.39537585e-01 -4.39562529e-01 2.34703040e+00
3.64712894e-01 6.32536829e-01 -2.44278252e-01 1.45829201e-01
8.50399375e-01 -9.67479646e-02 -5.63172698e-01 -4.43829507e-01
2.37641677e-01 -4.63534473e-03 6.40157938e-01 4.07762349e-01
-5.61954439e-01 -5.01379706e-02 7.70448875e+00 1.89419553e-01
-7.09696054e-01 1.23180874e-01 1.29040933e+00 -3.12211722e-01
-5.69330513e-01 2.23202094e-01 -1.13900788e-01 6.49238944e-01
1.14214921e+00 -7.40224659e-01 -5.51969528e-01 1.00369108e+00
4.04768944e-01 -7.41578341e-01 -1.11622667e+00 5.95271111e-01
-1.75459445e-01 -1.52259338e+00 1.08435296e-01 2.77098179e-01
6.67585611e-01 -2.34463871e-01 1.75270498e-01 2.33034659e-02
5.49154520e-01 -1.13409030e+00 9.70325530e-01 5.35213470e-01
8.74626160e-01 -5.18417358e-01 1.03505707e+00 3.36248040e-01
-3.52080733e-01 7.47559890e-02 -3.12378049e-01 -6.76141441e-01
2.38899276e-01 5.75719416e-01 -1.69114125e+00 1.52267084e-01
6.42969906e-01 -1.48023084e-01 -8.43026023e-03 7.17718422e-01
-4.03284758e-01 1.01314390e+00 -4.86049429e-02 1.06560707e-01
3.20824310e-02 5.79348981e-01 3.80113095e-01 1.03034735e+00
2.58931845e-01 8.95185292e-01 -3.68166089e-01 7.75547624e-01
-3.16832145e-03 -1.05878571e-03 -6.37952745e-01 4.73973542e-01
5.40604949e-01 5.38284242e-01 -1.07741284e+00 -8.80739927e-01
-2.32816085e-01 5.44807196e-01 -2.39498466e-01 -1.22656018e-01
-9.42817688e-01 7.68039525e-02 -4.23788391e-02 6.90019429e-01
-2.58085996e-01 5.91208816e-01 -1.06895101e+00 -8.59155297e-01
-3.10656428e-01 -5.28376997e-01 8.27571332e-01 -1.37569737e+00
-1.13965571e+00 3.40828538e-01 2.85773963e-01 -1.08272481e+00
-4.36437339e-01 -2.66271412e-01 -7.96496809e-01 9.37898278e-01
-9.85009372e-01 -4.36011195e-01 -1.74419373e-01 -1.95682533e-02
3.71465236e-01 2.98330188e-01 8.37744951e-01 -6.36290491e-01
9.97358784e-02 4.84866589e-01 -3.50479782e-01 -2.43150130e-01
9.68371034e-01 -1.43153560e+00 2.08916187e-01 4.00803834e-01
-3.59879285e-01 7.99764872e-01 1.05318940e+00 -1.00251377e+00
-2.98372716e-01 -6.66352034e-01 1.15735483e+00 -6.60730898e-01
3.59117568e-01 6.55667245e-01 -1.00531077e+00 9.37675834e-01
5.65397181e-02 -2.11045131e-01 1.26124334e+00 1.81904063e-01
-1.60365433e-01 4.45226073e-01 -1.53381276e+00 5.34662664e-01
6.99370980e-01 -1.31275877e-01 -1.41812313e+00 4.44762439e-01
6.12127244e-01 -2.99902141e-01 -9.44247603e-01 3.13362390e-01
9.08875406e-01 -1.33167768e+00 9.54111099e-01 -5.98141789e-01
9.03512955e-01 4.25675958e-02 3.63813601e-02 -1.11937559e+00
-5.12716830e-01 -9.86152589e-02 4.50620145e-01 2.31717736e-01
8.02431047e-01 -1.02157986e+00 8.62314224e-01 1.24003255e+00
-1.03529692e-01 -7.11754918e-01 -9.17581201e-01 -2.91811496e-01
1.65604517e-01 -6.38366759e-01 5.15066385e-01 1.40596759e+00
3.58546525e-01 8.53382275e-02 1.42496442e-02 1.11485116e-01
7.37992108e-01 3.65359306e-01 1.91437006e-01 -9.74884629e-01
-5.69213450e-01 -2.27435604e-01 -7.38557056e-02 -3.90940309e-01
-3.54020834e-01 -6.62596941e-01 -4.13973719e-01 -1.56638873e+00
7.83496618e-01 -3.29837590e-01 -3.20796698e-01 4.82842833e-01
-5.93559742e-01 -5.27794585e-02 2.93036491e-01 5.41858196e-01
-1.10681981e-01 1.08774751e-02 1.09281445e+00 2.85121083e-01
1.87172160e-01 3.65477085e-01 -1.32424343e+00 9.39906120e-01
1.04232061e+00 -9.37991679e-01 -5.44268966e-01 -8.56053904e-02
2.14936845e-02 9.03298199e-01 5.99625111e-01 -7.53269911e-01
-1.81977496e-01 -3.06193620e-01 9.40307677e-01 -5.97363234e-01
2.44564731e-02 -4.78939056e-01 5.16686976e-01 1.49804449e+00
-7.06974089e-01 2.95020103e-01 3.52038354e-01 7.25758314e-01
1.64550796e-01 -3.66000265e-01 6.98265493e-01 -5.15269816e-01
3.99444140e-02 -5.30904233e-01 -5.16845107e-01 -8.14886298e-03
1.05541885e+00 -3.41833770e-01 -3.10500622e-01 -6.11095011e-01
-1.10155785e+00 -1.49052620e-01 4.84402657e-01 -6.60593137e-02
5.83588839e-01 -9.86153960e-01 -5.97097158e-01 -5.16730249e-01
1.06017001e-01 -5.59919417e-01 3.89941245e-01 9.59933102e-01
-4.14512187e-01 5.94108164e-01 -1.00993522e-01 -3.60125273e-01
-1.19467509e+00 5.61096549e-01 3.90473634e-01 -5.21122575e-01
-5.95966578e-01 6.17777407e-01 9.06608820e-01 3.80564220e-02
-5.00723600e-01 -2.27198318e-01 1.57167569e-01 -4.91842151e-01
6.13606036e-01 8.11679780e-01 -3.36642742e-01 -1.69660091e-01
-3.02312464e-01 -4.67151403e-05 -4.17915314e-01 -4.59816068e-01
9.36677754e-01 -1.82700351e-01 3.09564024e-01 6.72573268e-01
4.23643380e-01 -9.19344947e-02 -8.97723913e-01 2.75323510e-01
-2.00700060e-01 -8.28394353e-01 -1.50308907e-01 -1.52335310e+00
-2.00916663e-01 6.46609545e-01 6.23528421e-01 2.26292551e-01
7.16033518e-01 2.58024186e-01 5.79208098e-02 1.57947332e-01
2.37474814e-01 -6.90149724e-01 -6.55584708e-02 -5.00980377e-01
9.54035163e-01 -1.28038490e+00 4.36536223e-01 -4.57756609e-01
-1.16485238e+00 7.44573534e-01 3.93779248e-01 -4.84541766e-02
4.94510055e-01 -1.36884317e-01 6.52475655e-02 -4.10097122e-01
-1.06071377e+00 6.90033138e-01 4.20389175e-02 3.31341207e-01
8.82960141e-01 8.29843342e-01 -9.53302801e-01 8.95406365e-01
-7.13880122e-01 8.84244889e-02 9.86990690e-01 7.54400432e-01
-3.75689685e-01 -5.40293694e-01 -9.18696404e-01 1.37784374e+00
-8.42026353e-01 9.35612470e-02 -2.67698526e-01 1.26199806e+00
-1.66630074e-01 8.64333093e-01 3.00314784e-01 -1.61161616e-01
3.05316448e-01 2.06192151e-01 1.32481664e-01 -7.36610949e-01
-5.29686928e-01 8.33910406e-02 2.53433257e-01 -4.63566273e-01
-2.66141236e-01 -9.78706002e-01 -1.52718318e+00 -2.74515897e-01
-4.18881416e-01 4.61716294e-01 1.73553124e-01 6.72237635e-01
-4.26649349e-03 6.49519145e-01 2.64980972e-01 6.64200410e-02
-9.69144821e-01 -8.92441988e-01 -4.76974130e-01 3.04300576e-01
2.00638145e-01 -8.56587470e-01 -6.22333348e-01 -1.38010010e-01] | [8.522860527038574, 5.588902473449707] |
5c097825-a2cb-4fd3-ac51-f390f76d7b7d | predicting-solar-flares-with-remote-sensing | 2110.07658 | null | https://arxiv.org/abs/2110.07658v1 | https://arxiv.org/pdf/2110.07658v1.pdf | Predicting Solar Flares with Remote Sensing and Machine Learning | High energy solar flares and coronal mass ejections have the potential to destroy Earth's ground and satellite infrastructures, causing trillions of dollars in damage and mass human suffering. Destruction of these critical systems would disable power grids and satellites, crippling communications and transportation. This would lead to food shortages and an inability to respond to emergencies. A solution to this impending problem is proposed herein using satellites in solar orbit that continuously monitor the Sun, use artificial intelligence and machine learning to calculate the probability of massive solar explosions from this sensed data, and then signal defense mechanisms that will mitigate the threat. With modern technology there may be only safeguards that can be implemented with enough warning, which is why the best algorithm must be identified and continuously trained with existing and new data to maximize true positive rates while minimizing false negatives. This paper conducts a survey of current machine learning models using open source solar flare prediction data. The rise of edge computing allows machine learning hardware to be placed on the same satellites as the sensor arrays, saving critical time by not having to transmit remote sensing data across the vast distances of space. A system of systems approach will allow enough warning for safety measures to be put into place mitigating the risk of disaster. | ['Erik Larsen'] | 2021-10-14 | null | null | null | null | ['solar-flare-prediction'] | ['time-series'] | [ 3.59095871e-01 1.83645785e-01 -1.92004681e-01 -6.41505048e-02
-3.15464623e-02 -6.58352852e-01 4.54908937e-01 1.27873495e-01
-1.51152909e-01 9.93760169e-01 -2.54566878e-01 -6.28654599e-01
-2.43278638e-01 -1.04763103e+00 -4.29777741e-01 -7.12128043e-01
-2.00078115e-01 2.76050240e-01 2.16423869e-01 -5.57311058e-01
2.32203454e-01 6.84717417e-01 -1.83217776e+00 -1.42297089e-01
9.24103737e-01 1.35522115e+00 9.32558104e-02 7.11257577e-01
3.65825236e-01 6.79363787e-01 -6.98284805e-01 4.60060447e-01
4.63969380e-01 -6.22784913e-01 -5.42067960e-02 -5.07536411e-01
1.10523507e-01 -1.13613131e-02 5.88551424e-02 8.55409503e-01
4.33142304e-01 -2.01093718e-01 3.24822158e-01 -1.26250970e+00
1.51379868e-01 -5.24781533e-02 -5.48056722e-01 7.26302564e-01
2.46188834e-01 2.50178546e-01 5.23277700e-01 -3.77618521e-01
1.69189602e-01 3.93353701e-01 9.14502561e-01 2.36064140e-02
-5.40866613e-01 -4.98081893e-01 -6.68757558e-01 3.01175207e-01
-1.05322444e+00 -4.92194802e-01 4.18423563e-01 -4.46863651e-01
1.62835479e+00 8.48195016e-01 1.17791533e+00 1.86187014e-01
7.99009204e-01 -6.66357800e-02 6.65819168e-01 -6.21118784e-01
4.50720459e-01 -1.34707287e-01 -2.81995744e-01 5.23813725e-01
9.37106192e-01 6.49847567e-01 -4.92133021e-01 -6.01539388e-02
9.05340016e-02 -2.80623943e-01 -6.38175547e-01 8.82945284e-02
-9.66703713e-01 7.85601318e-01 5.99893570e-01 6.10766530e-01
-6.03879750e-01 -1.25726178e-01 1.19431898e-01 3.90870571e-01
1.48480430e-01 6.76188171e-01 -7.06032634e-01 -1.15869986e-02
-1.04864717e+00 -5.35059199e-02 7.61999011e-01 5.15826046e-02
6.19544208e-01 5.29881716e-01 6.80803061e-01 -6.91472813e-02
2.63262331e-01 1.13827980e+00 4.80167568e-01 -5.83731592e-01
6.11084793e-03 6.73946619e-01 7.98872188e-02 -1.11184180e+00
-8.42211366e-01 -6.10308051e-01 -8.00782979e-01 7.68110275e-01
-3.66152644e-01 -5.95741451e-01 -7.95723379e-01 1.06299436e+00
4.83394980e-01 1.32923141e-01 1.99525505e-01 8.81432652e-01
1.09420791e-01 1.08248413e+00 -3.25561702e-01 -4.37385589e-01
1.13328862e+00 -2.88460463e-01 -5.56141317e-01 -9.12055254e-01
6.94400668e-01 -3.72968137e-01 1.74599230e-01 4.14874017e-01
-6.07521951e-01 -8.87388885e-02 -1.83108699e+00 5.17663300e-01
-3.92947793e-01 -2.75770724e-01 7.90580451e-01 7.35106587e-01
-8.13197553e-01 5.19945264e-01 -1.07895160e+00 -4.48767453e-01
1.32115886e-01 2.97657877e-01 1.10933684e-01 4.12652403e-01
-1.24846113e+00 1.39800406e+00 4.57261175e-01 1.52891815e-01
-3.83856803e-01 -7.63864577e-01 -6.56571209e-01 1.02945410e-01
-1.74144447e-01 -8.43718886e-01 9.30516005e-01 -8.92283976e-01
-7.88241684e-01 4.67215002e-01 2.70706713e-01 -1.03618777e+00
5.44596799e-02 2.07467616e-01 -8.28970790e-01 3.19592911e-03
-5.05813770e-03 3.12124670e-01 7.37629712e-01 -7.15621054e-01
-1.14351118e+00 -6.07852340e-01 -7.81971633e-01 1.54284045e-01
-4.27133054e-01 -1.70316517e-01 7.22017765e-01 -1.33684859e-01
2.09841236e-01 -1.00082994e+00 -1.82635233e-01 -2.89034117e-02
2.35293314e-01 3.07159759e-02 1.19702220e+00 -8.22621524e-01
9.10782516e-01 -1.84345031e+00 -1.35993510e-01 9.58062187e-02
-1.83997720e-01 3.86886597e-01 4.80394125e-01 3.68491679e-01
4.67555523e-02 -2.13383183e-01 -3.87219697e-01 4.84378606e-01
-4.05856490e-01 2.88389027e-01 -3.83913279e-01 4.81454551e-01
-1.87192425e-01 4.82465416e-01 -5.23303807e-01 2.05163330e-01
1.71878442e-01 2.83514291e-01 4.09287624e-02 -1.26169324e-01
-4.06843841e-01 2.28044331e-01 -3.44442248e-01 7.35987484e-01
3.48378241e-01 -1.81584954e-01 1.36194214e-01 1.68657154e-01
-4.00524259e-01 2.92442888e-01 -8.37247133e-01 1.11297464e+00
-3.02771866e-01 7.67716587e-01 2.58029848e-01 -1.17214715e+00
1.08285129e+00 1.64249837e-01 6.33839607e-01 -1.28020883e+00
3.31079543e-01 4.57304269e-01 -1.68456435e-01 -5.77190697e-01
2.90325522e-01 -5.44832826e-01 -1.21170312e-01 3.51959586e-01
-5.64039290e-01 -2.46085018e-01 -9.03996900e-02 -2.93443613e-02
1.35903275e+00 -3.67718577e-01 3.39757591e-01 -4.04991001e-01
1.24187469e-02 5.11463940e-01 8.39075923e-01 7.24688992e-02
1.02826335e-01 -1.78232074e-01 -2.06335172e-01 -7.68665969e-01
-8.91962528e-01 -7.48137355e-01 -2.62716770e-01 4.41645652e-01
3.76561321e-02 -7.58519992e-02 -3.44697356e-01 -8.64241421e-02
3.03955078e-01 1.06410074e+00 -1.52609468e-01 -5.36643684e-01
-8.83319527e-02 -1.23749506e+00 2.77956545e-01 3.08142811e-01
5.07035434e-01 -8.41533840e-01 -1.76595557e+00 5.99874705e-02
1.57215968e-02 -7.04499781e-01 4.77046758e-01 4.70129669e-01
-9.86943960e-01 -1.13329804e+00 1.33545190e-01 -2.91436464e-01
4.27797049e-01 4.48545784e-01 1.01033676e+00 4.40235287e-01
-7.96826899e-01 1.59982815e-01 -2.60376811e-01 -9.01558697e-01
-2.18437493e-01 -3.52945775e-01 5.49375594e-01 -4.75860894e-01
2.88406610e-01 -8.39624643e-01 -5.89905500e-01 1.44488481e-03
-7.06474423e-01 -1.18571967e-01 5.14465392e-01 3.34226727e-01
1.97634399e-01 5.76523840e-01 9.31823969e-01 -1.84592232e-01
5.29495142e-02 -9.28084135e-01 -1.10823965e+00 9.88324061e-02
-1.18051445e+00 -1.26061499e-01 2.91197836e-01 1.25708416e-01
-8.27240229e-01 1.98559403e-01 2.18101889e-01 2.08421536e-02
2.36550674e-01 5.23986340e-01 5.09390980e-02 -5.06851315e-01
9.78282571e-01 -1.66447029e-01 -1.48687929e-01 -6.54901862e-02
-9.97353643e-02 7.62922704e-01 7.89208889e-01 2.71677822e-01
7.91551769e-01 6.32431567e-01 2.50092596e-01 -1.26245117e+00
-6.44982100e-01 -2.50822276e-01 -2.66446054e-01 -4.99560088e-01
7.61994541e-01 -9.93006349e-01 -3.83536577e-01 1.92227557e-01
-7.95979440e-01 7.12933466e-02 -1.80331990e-01 4.59489644e-01
-1.92759097e-01 1.06133953e-01 1.86635569e-01 -1.04243016e+00
-6.88706875e-01 7.38704912e-05 5.65070152e-01 8.57666612e-01
-3.10212225e-01 -8.49416137e-01 6.22425020e-01 4.08119649e-01
7.06962526e-01 6.82754338e-01 4.81977046e-01 -1.62047952e-01
-8.04093301e-01 -4.28333074e-01 2.25808367e-01 2.18985319e-01
1.74215227e-01 -3.03668380e-01 -7.94468522e-01 -6.16212070e-01
5.86012959e-01 -1.78136960e-01 8.85259032e-01 2.74105698e-01
2.72686303e-01 -2.84005553e-01 -9.26249266e-01 8.53023231e-01
1.58569622e+00 6.87340558e-01 7.47230589e-01 4.71796006e-01
1.59988463e-01 1.69809505e-01 5.26973844e-01 4.15090710e-01
6.34216294e-02 1.43786475e-01 7.16124237e-01 -8.01330730e-02
2.40434036e-01 9.60984826e-02 1.99949995e-01 4.72903490e-01
1.18383318e-01 -2.60679424e-01 -9.74355638e-01 6.59609377e-01
-1.50274611e+00 -1.15741444e+00 -2.50496954e-01 2.57331824e+00
2.45851964e-01 3.85496676e-01 -3.75209033e-01 3.35296661e-01
3.43427896e-01 -1.78566314e-02 -8.04672360e-01 -3.77390832e-01
-1.68634325e-01 -6.26615286e-02 1.00101697e+00 5.06925344e-01
-6.83039129e-01 1.79698080e-01 6.20505810e+00 -8.54888335e-02
-1.51309597e+00 9.49272048e-03 1.94106072e-01 -3.64863247e-01
-5.57456136e-01 3.94897044e-01 -6.38678551e-01 6.80510044e-01
1.44133401e+00 -5.09597540e-01 4.16217327e-01 4.63417739e-01
2.46083945e-01 -7.36073732e-01 -4.47762698e-01 8.64686966e-01
3.96543860e-01 -1.47902811e+00 -6.97167158e-01 -4.10664966e-03
5.70899308e-01 6.49456918e-01 -6.71061993e-01 -1.12925515e-01
9.80510861e-02 -6.77972794e-01 2.94564456e-01 8.81800175e-01
4.88007665e-01 -7.71529734e-01 4.23016191e-01 7.68597186e-01
-8.15365195e-01 -4.87661093e-01 -2.47646466e-01 -7.31624484e-01
5.20116746e-01 1.14594460e+00 -9.97284234e-01 4.86325473e-01
1.02269006e+00 2.61155248e-01 -2.55510092e-01 1.20021808e+00
-7.10539892e-02 4.92925227e-01 -1.19542885e+00 -1.35763317e-01
-3.28361005e-01 -1.33181915e-01 7.52162099e-01 3.55748922e-01
8.83142233e-01 4.02984679e-01 -5.76920286e-02 2.38067418e-01
3.52125168e-01 -5.06512165e-01 -9.92065907e-01 -1.09211355e-01
5.34067690e-01 1.26121187e+00 -7.62193322e-01 -1.54515281e-01
-2.50592619e-01 7.86600113e-01 -3.53165865e-01 -3.86647999e-01
-6.03483677e-01 -4.23645765e-01 4.74753290e-01 5.22104502e-01
1.74225457e-02 -2.99058378e-01 -6.05250597e-01 -8.55676055e-01
1.26724616e-02 -4.96941388e-01 5.93881130e-01 -1.08635211e+00
-8.21332872e-01 4.35786217e-01 -4.25693840e-01 -1.00285614e+00
-3.93936157e-01 -2.27387503e-01 -9.31637287e-01 5.80603480e-01
-1.36549628e+00 -6.27132297e-01 -2.42859423e-01 1.21293880e-01
2.21363436e-02 -2.68140256e-01 7.61500478e-01 2.23045871e-01
-5.83459586e-02 -1.50498539e-01 1.87373832e-01 -6.46224797e-01
3.52230519e-01 -7.22154379e-01 3.38099986e-01 9.59179103e-01
1.96529955e-01 -3.79479796e-01 1.00270939e+00 -1.21705079e+00
-1.71336401e+00 -8.24796617e-01 1.08474743e+00 -4.18359369e-01
6.96904063e-01 -2.26964196e-03 -7.97143579e-01 7.41855383e-01
3.88539076e-01 -1.61162376e-01 4.62657660e-01 -4.12394144e-02
2.37522081e-01 -3.09924185e-01 -1.24555898e+00 7.47269066e-03
5.11589527e-01 -1.01474285e-01 -6.75388336e-01 7.55240381e-01
3.32139730e-01 1.93559974e-02 -5.07945359e-01 9.89149511e-01
2.83015847e-01 -1.07830501e+00 5.72990656e-01 -3.43843818e-01
-2.39204884e-01 -6.18941009e-01 5.74321300e-02 -1.37347782e+00
-5.52312210e-02 -6.43401980e-01 -7.11212978e-02 7.41560340e-01
6.63526416e-01 -7.93042839e-01 7.35787451e-01 2.78769523e-01
-4.16023940e-01 -3.42690319e-01 -1.26849043e+00 -7.27445483e-01
-1.77627236e-01 -1.76631674e-01 2.66453356e-01 1.09189034e+00
1.75532594e-01 5.51863790e-01 -1.42635867e-01 7.33651459e-01
8.68018866e-01 2.83853859e-01 2.96431988e-01 -1.72667682e+00
-9.80535895e-02 1.15704238e-01 -6.84238434e-01 5.43899760e-02
-4.27208453e-01 -7.23298073e-01 4.98641953e-02 -1.64558518e+00
-2.78687954e-01 -5.32568634e-01 -1.62003692e-02 7.34466672e-01
2.66864091e-01 3.51120472e-01 -3.11899818e-02 2.35613272e-01
-1.51079550e-01 4.88597631e-01 3.89800191e-01 -8.74017179e-02
-3.35788764e-02 -4.52548377e-02 -3.84077698e-01 7.30320275e-01
1.03899920e+00 -5.63826561e-01 -3.64751667e-01 -3.55398923e-01
1.24555182e+00 4.19770151e-01 3.95724714e-01 -1.85050166e+00
6.95604146e-01 -1.18757442e-01 7.44896054e-01 -4.79334682e-01
1.31960526e-01 -1.02388322e+00 6.00919068e-01 1.01571894e+00
7.81325281e-01 7.20702996e-03 4.14889216e-01 4.62308407e-01
8.86936784e-02 -1.12237133e-01 1.00357425e+00 2.01807395e-01
-5.99023342e-01 -2.05397353e-01 -6.16123438e-01 -5.46494663e-01
1.62030506e+00 1.47293806e-01 -5.24748325e-01 -2.14857936e-01
-5.06173193e-01 6.82861626e-01 6.61252797e-01 6.72742426e-01
2.08972052e-01 -9.65447068e-01 -5.64353406e-01 5.41845262e-01
-2.23480448e-01 -4.22364116e-01 3.00451308e-01 6.27103329e-01
-6.58667803e-01 4.54605699e-01 -5.08593619e-01 -4.40798461e-01
-1.01740062e+00 6.92922890e-01 7.35821426e-01 2.52475768e-01
-6.93220317e-01 6.13243401e-01 -6.84293866e-01 -1.18086696e-01
-2.48449922e-01 -2.45081168e-02 8.42336714e-02 1.71282813e-01
8.34667861e-01 1.93227425e-01 5.99150181e-01 -1.58554569e-01
-5.58727801e-01 3.04358602e-01 4.54430223e-01 1.46646574e-01
1.42441666e+00 -7.39907026e-02 -1.97844684e-01 2.90252566e-01
3.68052036e-01 4.69630882e-02 -1.04946887e+00 4.88289803e-01
1.36155128e-01 -1.86885595e-01 7.45178163e-01 -1.23899972e+00
-1.00620830e+00 5.20666420e-01 9.75653529e-01 7.21738935e-01
1.52377951e+00 -5.70705347e-02 1.04387856e+00 4.94123042e-01
6.63846016e-01 -1.16764617e+00 -5.43811262e-01 5.54354340e-02
5.97718060e-01 -9.67178941e-01 2.77977884e-01 1.87166676e-01
-1.56889588e-01 8.89162898e-01 4.20676857e-01 1.50565919e-03
6.95647717e-01 8.40273857e-01 2.02071806e-03 -4.31369215e-01
-9.99005735e-01 6.06238656e-02 -3.59433889e-01 3.65677774e-01
-2.65472889e-01 1.81437179e-01 -4.54942167e-01 2.52341449e-01
-1.92720577e-01 1.39790624e-01 4.88892764e-01 1.30163169e+00
-1.54974926e+00 -7.74629712e-01 -9.16409850e-01 7.82123625e-01
-4.96221595e-02 1.49105296e-01 -3.38194251e-01 2.56140918e-01
2.80086935e-01 9.25520837e-01 4.92958128e-01 -2.25704134e-01
1.65978462e-01 1.27691403e-01 2.49007016e-01 4.31122119e-03
-1.27411157e-01 -5.15958786e-01 1.98134452e-01 -2.59701937e-01
2.94984262e-02 -7.41299868e-01 -1.69816196e+00 -2.01620013e-01
-3.36453050e-01 3.89636695e-01 1.34305954e+00 1.04349947e+00
6.98602140e-01 4.45719481e-01 8.01889539e-01 -7.98046827e-01
-4.92433399e-01 -8.09084117e-01 -5.89239359e-01 -4.09640431e-01
4.20739770e-01 -6.09225988e-01 -8.12615633e-01 -7.74668679e-02] | [6.64409875869751, 2.8139593601226807] |
ec8298d1-4e4a-43ce-a8b1-ded3befc0818 | scalable-coupling-of-deep-learning-with | 2305.07617 | null | https://arxiv.org/abs/2305.07617v1 | https://arxiv.org/pdf/2305.07617v1.pdf | Scalable Coupling of Deep Learning with Logical Reasoning | In the ongoing quest for hybridizing discrete reasoning with neural nets, there is an increasing interest in neural architectures that can learn how to solve discrete reasoning or optimization problems from natural inputs. In this paper, we introduce a scalable neural architecture and loss function dedicated to learning the constraints and criteria of NP-hard reasoning problems expressed as discrete Graphical Models. Our loss function solves one of the main limitations of Besag's pseudo-loglikelihood, enabling learning of high energies. We empirically show it is able to efficiently learn how to solve NP-hard reasoning problems from natural inputs as the symbolic, visual or many-solutions Sudoku problems as well as the energy optimization formulation of the protein design problem, providing data efficiency, interpretability, and \textit{a posteriori} control over predictions. | ['Thomas Schiex', 'Sophie Barbe', 'Marianne Defresne'] | 2023-05-12 | null | null | null | null | ['protein-design', 'logical-reasoning'] | ['medical', 'reasoning'] | [ 4.50032860e-01 6.69104218e-01 -8.81962031e-02 -7.61563182e-01
-8.39698613e-01 -5.93241930e-01 2.63513118e-01 2.38315076e-01
-4.66744602e-01 1.20384765e+00 -2.39016458e-01 -8.25339913e-01
-6.99833274e-01 -9.68442261e-01 -1.07908869e+00 -5.21252513e-01
-2.40031481e-01 9.93730426e-01 -2.83337057e-01 -8.88070166e-02
-6.51783589e-03 7.58011639e-01 -1.54393411e+00 5.53196907e-01
1.00120234e+00 1.34732246e+00 -3.90033424e-02 8.31966758e-01
-9.95527282e-02 1.09711242e+00 -2.75850147e-01 -3.91474038e-01
1.57069713e-01 -4.22873646e-01 -8.84781063e-01 -6.94559276e-01
8.05872738e-01 -3.15035671e-01 -1.54674038e-01 1.20592892e+00
2.98836559e-01 8.86100754e-02 9.81306851e-01 -1.28086734e+00
-9.15156960e-01 6.38353050e-01 4.09634635e-02 -1.11100696e-01
3.42508227e-01 2.73826689e-01 1.39948344e+00 -4.95076746e-01
6.46496773e-01 1.44599521e+00 5.85926890e-01 5.77869117e-01
-1.55296004e+00 -3.25275332e-01 8.09956565e-02 4.15089518e-01
-1.16321301e+00 -1.77395836e-01 4.50813532e-01 -1.25670120e-01
1.58883178e+00 5.31095743e-01 5.26846290e-01 8.58594894e-01
1.89921707e-01 7.49504745e-01 9.42635059e-01 -6.06959403e-01
7.35327601e-01 -1.43687755e-01 3.03038836e-01 1.16179252e+00
2.27479979e-01 2.88616776e-01 -1.91431776e-01 -4.90623951e-01
5.09882450e-01 -1.66118592e-01 -1.20593451e-01 -3.78996938e-01
-5.20649314e-01 9.82459664e-01 5.76013446e-01 -3.76087844e-01
-9.89173576e-02 5.69825113e-01 2.25945860e-01 3.90002817e-01
-1.84002426e-02 7.98083901e-01 -8.32519233e-01 1.62899699e-02
-5.99901974e-01 5.26254535e-01 1.21255136e+00 6.82285190e-01
6.35229528e-01 3.84890735e-02 1.04501501e-01 5.62196374e-01
3.61912966e-01 2.12969214e-01 -5.90919480e-02 -1.07761931e+00
5.94011664e-01 6.81380928e-01 1.43577531e-01 -7.12921143e-01
-7.56595731e-01 -7.68831819e-02 -7.10161388e-01 7.65947163e-01
6.46558583e-01 -3.11032951e-01 -1.01185012e+00 1.99457955e+00
9.00288075e-02 -4.42037672e-01 1.43052861e-01 6.90392613e-01
7.15358734e-01 7.72237897e-01 1.49662286e-01 -3.17748129e-01
1.08490443e+00 -5.54851055e-01 -5.53289115e-01 -2.95755446e-01
5.78107536e-01 3.09481677e-02 8.45694184e-01 9.04252350e-01
-1.32129264e+00 -3.10440630e-01 -1.20449197e+00 -5.37332177e-01
-7.54525959e-01 7.61963651e-02 1.08854592e+00 4.77932274e-01
-1.02970827e+00 9.08317745e-01 -8.13143373e-01 2.30210930e-01
4.33625877e-01 9.88544643e-01 -1.15706883e-01 -1.73368361e-02
-1.39007080e+00 1.01371193e+00 8.36624324e-01 2.23830894e-01
-5.45016885e-01 -4.96979177e-01 -8.05826008e-01 5.23706436e-01
6.89816892e-01 -5.07044971e-01 1.12099874e+00 -7.91058779e-01
-1.37535954e+00 4.77103323e-01 3.03590268e-01 -8.47835839e-01
5.49655855e-01 -1.51781263e-02 -1.61143094e-01 -1.46901459e-01
-5.43304145e-01 1.10952222e+00 2.53624558e-01 -7.14953423e-01
-1.79611638e-01 -2.54881740e-01 3.62202913e-01 6.23669513e-02
-4.90584075e-02 -4.91992474e-01 1.60605267e-01 -1.79844543e-01
-4.77056131e-02 -8.31596375e-01 -3.33527654e-01 3.73297215e-01
-4.45549160e-01 -4.54564571e-01 4.61370468e-01 -7.28949904e-01
8.94498348e-01 -1.71277881e+00 6.90111637e-01 3.70827377e-01
1.81897938e-01 1.43736184e-01 -7.40203857e-02 3.23979050e-01
-2.56938279e-01 3.38676989e-01 -4.53954160e-01 2.67910123e-01
6.72252297e-01 6.06436193e-01 -2.08003372e-01 7.15583563e-02
5.75743020e-01 1.23277187e+00 -5.04310131e-01 -2.73077428e-01
-6.14105836e-02 2.63840675e-01 -8.80848706e-01 9.67827812e-02
-1.19583774e+00 -2.05428749e-01 -2.35293120e-01 4.93864208e-01
4.93606627e-01 -4.22363043e-01 6.83422446e-01 5.79306670e-02
8.11953396e-02 2.87655741e-01 -1.23761165e+00 1.39281642e+00
-2.13234574e-01 5.08763790e-01 2.38171425e-02 -1.01075196e+00
6.75078750e-01 4.38708179e-02 1.13305315e-01 -9.51119304e-01
7.44052529e-02 1.98993400e-01 6.84775710e-02 -4.56734031e-01
6.80664927e-02 -2.09481522e-01 -8.50818604e-02 1.85430795e-01
1.71149194e-01 -3.34232658e-01 2.65707970e-01 -2.86431730e-01
1.14745808e+00 4.20627683e-01 3.40939432e-01 -7.87933096e-02
3.33838880e-01 1.04195789e-01 5.21859527e-01 9.17360663e-01
3.38853300e-01 1.61171719e-01 1.15266871e+00 -9.88781750e-01
-1.07566035e+00 -1.05890489e+00 -8.85286033e-02 1.07457948e+00
-3.86792183e-01 -1.65629596e-01 -6.63966596e-01 -5.37223995e-01
1.99119374e-01 9.99023259e-01 -6.25610709e-01 -2.75896132e-01
-6.96546078e-01 -8.50562394e-01 4.76268381e-01 6.69088960e-01
4.05752435e-02 -1.19793069e+00 -5.20882130e-01 1.12662695e-01
2.40899593e-01 -6.50325119e-01 1.51454657e-01 1.33963239e+00
-8.21702361e-01 -1.04568362e+00 -1.50639072e-01 -5.32920063e-01
5.24215817e-01 -7.07555473e-01 1.16781425e+00 -1.02219358e-01
-6.79136455e-01 -3.19672287e-01 2.89845496e-01 -2.81982511e-01
-5.51327467e-01 -4.98493947e-02 -3.97878475e-02 -7.40285277e-01
4.42323864e-01 -6.23377502e-01 -1.00361168e-01 4.71842885e-02
-1.05544102e+00 1.34864777e-01 5.74500680e-01 1.18851793e+00
8.86822104e-01 7.05821365e-02 4.06759918e-01 -8.33008409e-01
7.52913117e-01 -3.73795688e-01 -1.16031623e+00 6.58658564e-01
-7.45775104e-01 7.83483565e-01 1.18356323e+00 -3.19856793e-01
-6.51403248e-01 4.54134762e-01 -1.83103636e-01 -1.98256820e-01
-3.07444751e-01 7.25150883e-01 -2.82886326e-01 -1.09484196e-01
8.36484432e-01 2.00332608e-02 -2.88579643e-01 -1.87125295e-01
5.71800113e-01 1.21902548e-01 3.68765771e-01 -1.04131722e+00
2.66845047e-01 -1.41796783e-01 5.33924818e-01 -5.07860303e-01
-9.29563522e-01 2.45638177e-01 -5.57067752e-01 3.33324283e-01
1.11353183e+00 -4.77308333e-01 -1.35518289e+00 -1.22198820e-01
-1.23419535e+00 -4.76906627e-01 -3.96955132e-01 1.32321820e-01
-1.03394020e+00 1.82970598e-01 -5.85657179e-01 -9.44657683e-01
-1.89521506e-01 -1.25518882e+00 7.00442433e-01 1.80216432e-01
-3.31364095e-01 -9.50010657e-01 2.05343477e-02 2.75979012e-01
7.35708773e-02 6.24986947e-01 1.80158007e+00 -7.04807580e-01
-6.56028152e-01 -9.11570936e-02 -3.53250831e-01 2.22836301e-01
-4.08325493e-01 -1.61469489e-01 -8.48958790e-01 -1.78474426e-01
-1.67531535e-01 -1.07486796e+00 8.11614990e-01 3.93214852e-01
1.54383028e+00 -7.90557206e-01 6.75861984e-02 7.89581478e-01
1.56182063e+00 2.95001686e-01 5.68968654e-01 -7.75142685e-02
3.00842196e-01 4.16538954e-01 2.44631153e-02 1.07795812e-01
-3.75527143e-02 4.22523320e-01 7.00253785e-01 6.96565881e-02
3.26269388e-01 -2.51760811e-01 1.52239993e-01 -2.00919863e-02
-1.15283348e-01 -4.33183759e-01 -1.28183484e+00 -1.78014666e-01
-2.08725047e+00 -1.09272385e+00 9.95738283e-02 1.78562379e+00
1.04023075e+00 3.96421820e-01 -1.70260102e-01 -1.34939384e-02
2.19215870e-01 -2.68724501e-01 -1.06964290e+00 -8.63727212e-01
-6.61653802e-02 4.58193779e-01 4.13174570e-01 7.32326448e-01
-9.52577233e-01 5.99092960e-01 7.37964916e+00 7.56618798e-01
-6.11637473e-01 -4.23262417e-01 8.99117827e-01 -2.83091933e-01
-3.40041548e-01 -2.60473073e-01 -7.87578285e-01 -2.80018393e-02
1.39982688e+00 4.38470840e-01 9.14892375e-01 9.15675819e-01
-1.72389552e-01 -6.40438963e-03 -1.68787956e+00 7.76044667e-01
-2.70329863e-01 -1.69373059e+00 9.52874646e-02 1.98120549e-01
5.24399519e-01 -1.16156854e-01 1.06731176e-01 4.84502137e-01
8.64881337e-01 -1.69342232e+00 6.72441602e-01 6.13817155e-01
8.06963921e-01 -9.78981316e-01 4.52097327e-01 5.52351117e-01
-6.98789120e-01 -4.97659922e-01 -4.49986130e-01 -3.36266518e-01
-2.04831079e-01 2.07840875e-01 -7.47042894e-01 2.63318539e-01
2.86465824e-01 1.26106173e-01 -2.84572482e-01 6.10835373e-01
-2.73278952e-01 2.22558305e-01 -5.90353131e-01 -5.17011046e-01
4.85891759e-01 -3.11450452e-01 -3.95015702e-02 9.68542695e-01
-9.13909823e-02 4.02849525e-01 2.61647284e-01 1.39765644e+00
-1.50235459e-01 -4.04845655e-01 -4.43747133e-01 -4.97135073e-01
1.96576901e-02 7.88906813e-01 -6.80708230e-01 -4.87692133e-02
-7.80061111e-02 5.69231868e-01 1.00155711e+00 4.63320076e-01
-8.83798242e-01 -4.49218750e-01 4.82780755e-01 -2.15929791e-01
2.61112392e-01 -3.44873279e-01 -5.34725964e-01 -8.61283004e-01
-2.39051692e-02 -1.06367838e+00 5.16625404e-01 -8.97238910e-01
-1.13696277e+00 2.71085173e-01 1.69012278e-01 -2.87539780e-01
-4.39119041e-01 -1.48519647e+00 -4.06862587e-01 8.81177306e-01
-1.22059488e+00 -8.26664090e-01 1.76917404e-01 6.19209697e-03
1.04338765e-01 -1.69889808e-01 1.10385334e+00 2.90499274e-02
-3.26423943e-01 4.55001354e-01 3.45544398e-01 1.27729196e-02
-9.46426019e-02 -1.64585900e+00 2.83546478e-01 2.07902148e-01
3.02752759e-02 6.53622329e-01 9.09559846e-01 -2.95744538e-01
-1.76994061e+00 -5.48865438e-01 5.34738064e-01 -2.54164875e-01
6.84848905e-01 -6.86160684e-01 -7.44352043e-01 5.51173151e-01
-1.54603302e-01 -7.82822147e-02 6.13916993e-01 3.72165143e-01
-4.21770334e-01 2.09174111e-01 -1.29144764e+00 5.55789351e-01
9.83012319e-01 -6.63669050e-01 -6.29904091e-01 6.61896706e-01
7.73581684e-01 -5.61796606e-01 -7.47438967e-01 2.49326855e-01
5.72272897e-01 -8.39330852e-01 1.20244527e+00 -1.36552310e+00
7.76031196e-01 9.80980173e-02 -4.09730852e-01 -9.05562758e-01
-2.77965456e-01 -5.12151122e-01 -6.80249035e-01 5.08587480e-01
6.89144909e-01 -4.62157905e-01 9.39729393e-01 1.05300713e+00
1.56796172e-01 -1.18930304e+00 -1.10419822e+00 -4.32704538e-01
4.56735611e-01 -2.33938739e-01 3.69484395e-01 6.76609695e-01
-6.01962022e-02 2.06039235e-01 -1.31480336e-01 1.57012224e-01
5.46311080e-01 2.22141907e-01 1.45173430e-01 -1.35230672e+00
-8.85893226e-01 -5.08300483e-01 -3.16799670e-01 -8.54531467e-01
4.33006823e-01 -1.02832866e+00 -3.04685924e-02 -1.60802972e+00
2.31690910e-02 -1.24068096e-01 -3.49499166e-01 1.09155715e+00
5.16453385e-01 -1.96425065e-01 -7.83151761e-02 -3.97251993e-01
-5.71500361e-01 4.17300314e-01 9.49692547e-01 -5.32443166e-01
-2.80537196e-02 -2.91449070e-01 -4.73429292e-01 6.77574813e-01
5.17911911e-01 -5.63590527e-01 -4.98869032e-01 -3.61188114e-01
1.21047390e+00 6.29930854e-01 5.33787966e-01 -9.83094037e-01
1.59597307e-01 -6.48248851e-01 7.06249356e-01 -6.73621476e-01
4.77263927e-01 -8.35754037e-01 2.87131727e-01 6.12081707e-01
-8.09335113e-01 1.96885467e-02 3.88987362e-01 4.80179846e-01
6.12052679e-02 -4.22114283e-01 7.17874408e-01 -3.35547924e-01
-3.91248256e-01 7.28312284e-02 -1.97861224e-01 1.52797922e-01
7.56911874e-01 1.14091180e-01 -4.34895515e-01 -1.25132665e-01
-1.16107440e+00 5.52310586e-01 2.27601856e-01 8.17840546e-02
6.90420926e-01 -1.01369822e+00 -3.16537321e-01 6.52412027e-02
-2.19210178e-01 1.14227019e-01 -1.01973116e-01 3.40575546e-01
-9.04957175e-01 6.89403355e-01 -4.32188690e-01 -2.14654058e-01
-1.02621722e+00 5.61954796e-01 7.00470328e-01 -4.07346725e-01
-2.88081765e-01 8.90224814e-01 -2.26872057e-01 -8.12488198e-01
3.67687523e-01 -6.39050126e-01 2.50729233e-01 -9.13442075e-02
3.04814488e-01 3.05090189e-01 4.05055471e-02 2.64852494e-01
-3.06509674e-01 -4.37967256e-02 -1.23949073e-01 9.34912711e-02
1.71492445e+00 6.91883922e-01 -2.63291121e-01 3.75758767e-01
1.09904885e+00 -7.51105964e-01 -1.31983018e+00 1.17648467e-01
1.76585466e-01 2.71570772e-01 -2.49295402e-02 -1.47164273e+00
-2.62203634e-01 1.03624630e+00 6.66945338e-01 4.27030802e-01
7.79900432e-01 -1.56407192e-01 3.68128717e-01 1.52397585e+00
2.91884273e-01 -1.20528710e+00 -2.54142046e-01 8.59190762e-01
8.09410393e-01 -1.17540097e+00 2.55470008e-01 6.67123049e-02
-1.08282551e-01 1.56403327e+00 5.09393394e-01 6.46866187e-02
3.01562250e-01 6.08060002e-01 -3.79363477e-01 -2.93454885e-01
-1.10282290e+00 2.17338130e-01 4.95465517e-01 3.56851280e-01
3.39637339e-01 4.85020354e-02 -6.75036013e-02 3.89801651e-01
-2.11777538e-01 7.82736316e-02 2.13685989e-01 8.38107705e-01
-7.35000730e-01 -1.13333249e+00 -1.05642363e-01 5.94945550e-01
-2.03418404e-01 -1.22993670e-01 -7.59786129e-01 6.69944406e-01
1.88732073e-01 3.14377666e-01 -3.04191321e-01 -5.71236983e-02
-3.19578350e-02 5.94027340e-01 7.41029799e-01 -5.09742856e-01
-4.02332932e-01 -4.81833547e-01 3.53363603e-01 -5.61141133e-01
1.26112536e-01 -2.73551643e-01 -1.63728082e+00 -1.36655837e-01
-2.24702686e-01 2.25026250e-01 5.60512662e-01 1.04640436e+00
9.57900137e-02 6.23932421e-01 5.23423180e-02 -7.06360340e-01
-1.11916447e+00 -5.00615895e-01 -2.92874187e-01 -4.64961678e-03
4.26420122e-01 -3.86263251e-01 -2.28124335e-01 -3.45393032e-01] | [8.69253921508789, 6.9519362449646] |
0e3d31fa-1cea-49fa-83c4-da783b25ad89 | systematic-assessment-of-the-quality-of-fit | 2201.01658 | null | https://arxiv.org/abs/2201.01658v1 | https://arxiv.org/pdf/2201.01658v1.pdf | Systematic assessment of the quality of fit of the stochastic block model for empirical networks | We perform a systematic analysis of the quality of fit of the stochastic block model (SBM) for 275 empirical networks spanning a wide range of domains and orders of size magnitude. We employ posterior predictive model checking as a criterion to assess the quality of fit, which involves comparing networks generated by the inferred model with the empirical network, according to a set of network descriptors. We observe that the SBM is capable of providing an accurate description for the majority of networks considered, but falls short of saturating all modeling requirements. In particular, networks possessing a large diameter and slow-mixing random walks tend to be badly described by the SBM. However, contrary to what is often assumed, networks with a high abundance of triangles can be well described by the SBM in many cases. We demonstrate that simple network descriptors can be used to evaluate whether or not the SBM can provide a sufficiently accurate representation, potentially pointing to possible model extensions that can systematically improve the expressiveness of this class of models. | ['Tiago P. Peixoto', 'Felipe Vaca-Ramírez'] | 2022-01-05 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 3.72118413e-01 4.10125613e-01 -3.78807783e-01 -5.67994602e-02
-1.63743615e-01 -6.76526725e-01 1.04759026e+00 1.96246088e-01
6.35656863e-02 8.10819685e-01 -2.19184995e-01 -8.51385176e-01
-6.56094551e-01 -1.10220265e+00 -6.36648297e-01 -7.06952453e-01
-3.46221685e-01 1.06753218e+00 7.25833774e-01 -1.27812117e-01
8.66365954e-02 5.99982083e-01 -1.25596130e+00 -3.97772752e-02
7.64609635e-01 8.83222640e-01 -2.75725901e-01 4.43994790e-01
1.52898267e-01 4.85259861e-01 -2.50624239e-01 -6.10219598e-01
8.89531672e-02 -4.25012529e-01 -9.36423361e-01 -1.72501162e-01
8.22289512e-02 4.22294810e-02 -4.94138777e-01 1.05068958e+00
-1.06761701e-01 -1.67706177e-01 9.98053432e-01 -1.51394665e+00
1.23248354e-01 9.10940886e-01 -1.66590765e-01 2.03818738e-01
4.35038447e-01 6.97057769e-02 1.27194881e+00 -3.34384292e-01
7.33951092e-01 1.20620358e+00 8.61276150e-01 1.10069752e-01
-1.98298061e+00 -5.09997368e-01 -4.72158827e-02 -1.85898319e-01
-1.67464530e+00 -3.81203562e-01 6.47124410e-01 -4.24085081e-01
7.08502948e-01 3.53716642e-01 8.09401989e-01 9.91387308e-01
1.13991812e-01 -1.23885132e-01 1.14523780e+00 -3.98801357e-01
3.39481682e-01 1.26822263e-01 2.42233336e-01 6.65848672e-01
9.84848320e-01 7.28966668e-02 -2.94684649e-01 -6.31427646e-01
7.44907975e-01 -4.68635052e-01 3.22619709e-03 -6.81986094e-01
-1.13031209e+00 6.77885890e-01 2.03972772e-01 4.24868077e-01
-2.10799605e-01 2.54769564e-01 4.29620475e-01 3.17036033e-01
3.80594313e-01 4.42399889e-01 -3.11485380e-01 -2.12210074e-01
-1.03588295e+00 4.40681875e-01 1.21002722e+00 9.65193272e-01
6.05356157e-01 -1.19145490e-01 2.55573690e-01 3.59316200e-01
3.59905779e-01 2.18843073e-01 -2.31989279e-01 -1.04148400e+00
2.76905298e-01 7.23908842e-01 6.26243353e-02 -1.07983792e+00
-4.13980544e-01 -5.30500591e-01 -1.02532423e+00 1.78210326e-02
9.40742612e-01 3.62204850e-01 -3.10255170e-01 1.61739278e+00
8.87997970e-02 -1.89091638e-01 -2.81355351e-01 2.08905086e-01
2.77028054e-01 3.01134557e-01 -1.38326421e-01 -5.06900847e-01
9.48469639e-01 -3.53691518e-01 -1.98885202e-01 8.31682757e-02
7.90290058e-01 -3.62622738e-01 7.31168628e-01 5.26750684e-01
-1.18968320e+00 -2.64007390e-01 -9.34206605e-01 4.93504643e-01
-5.52020501e-03 -4.95178610e-01 7.05940187e-01 8.63572478e-01
-1.06595886e+00 1.07284081e+00 -9.46301997e-01 -5.80188513e-01
2.70077676e-01 4.06760305e-01 -4.60284412e-01 -2.95611799e-01
-8.99320185e-01 9.20944035e-01 4.56505984e-01 1.53991774e-01
-8.11771810e-01 -2.32799590e-01 -6.23261809e-01 2.41672218e-01
4.06864613e-01 -6.42886937e-01 7.58496642e-01 -7.96790719e-01
-9.39596593e-01 6.84697747e-01 -3.37463379e-01 -7.22914875e-01
8.40527892e-01 6.03758454e-01 -5.51969886e-01 2.84408599e-01
-2.49824062e-01 1.49536222e-01 5.11517286e-01 -1.39915049e+00
1.91054389e-01 -1.86064586e-01 3.91514570e-01 -3.82486373e-01
-4.20420505e-02 -5.01561090e-02 2.13889983e-02 -3.53120297e-01
3.78590524e-01 -1.11318445e+00 -3.71516824e-01 2.60645375e-02
-7.75121987e-01 -1.42370969e-01 2.63810337e-01 -6.68798685e-02
1.53547144e+00 -1.62566268e+00 1.64016839e-02 1.05177009e+00
4.53126520e-01 -2.39734259e-03 -8.14830363e-02 9.82884049e-01
-2.18040794e-01 6.87999129e-01 -3.52482647e-01 -1.77362427e-01
2.13277251e-01 5.61754644e-01 -4.04330976e-02 8.23461711e-01
6.32374808e-02 6.60071671e-01 -6.32884026e-01 -6.95744455e-01
1.30295372e-02 1.80287912e-01 -5.94780922e-01 -2.54712641e-01
-1.39180079e-01 2.51460165e-01 -3.59207153e-01 4.33203280e-01
4.69777882e-01 -7.84973443e-01 7.41189361e-01 5.90271968e-03
2.57061422e-01 5.15704632e-01 -1.17220688e+00 8.56384099e-01
4.74409834e-02 5.76832652e-01 1.20694665e-02 -1.05526519e+00
9.35473263e-01 1.77372754e-01 2.93653011e-01 -2.35924721e-01
5.13388887e-02 2.72444516e-01 6.54284418e-01 -2.97124777e-02
1.79130092e-01 -3.29151750e-01 1.66081443e-01 6.25032961e-01
-1.19617157e-01 -2.41534226e-02 4.50164229e-01 3.73013407e-01
1.37527514e+00 -2.04734236e-01 3.99471939e-01 -6.82194054e-01
4.12264407e-01 7.48141389e-03 3.44567567e-01 1.01519167e+00
-2.14503519e-02 3.10038507e-01 1.16564000e+00 -3.41798455e-01
-1.30849838e+00 -1.04311228e+00 -2.88251311e-01 5.07492304e-01
3.20206247e-02 -6.95056796e-01 -5.73050439e-01 -4.04425323e-01
8.19584951e-02 3.96586090e-01 -6.09646678e-01 -2.83354491e-01
-2.05779761e-01 -7.73966193e-01 7.14076638e-01 3.01105052e-01
-1.02875128e-01 -6.98830068e-01 -4.61835898e-02 4.16790508e-03
9.43096429e-02 -1.14659655e+00 1.81814864e-01 2.25043118e-01
-9.83014286e-01 -1.51189137e+00 -2.06732363e-01 -1.98502809e-01
7.76389122e-01 -5.71029037e-02 1.30343962e+00 5.32043099e-01
1.69254631e-01 -3.68477963e-02 -2.01387718e-01 3.37396771e-01
-1.16630042e+00 3.29923868e-01 2.68147677e-01 -2.50650853e-01
5.24713397e-02 -8.69522989e-01 -3.21018726e-01 5.95338583e-01
-9.46797550e-01 -1.16438568e-01 3.77896547e-01 5.57621241e-01
2.42477983e-01 4.96103019e-01 3.63384485e-01 -9.97022629e-01
5.34854114e-01 -5.83357513e-01 -7.68092036e-01 1.67317554e-01
-1.06433821e+00 1.77302122e-01 7.15651453e-01 -3.18085372e-01
-3.86426181e-01 -3.24326277e-01 -3.33385877e-02 -3.75828939e-03
-5.10806181e-02 7.92211711e-01 -1.37825176e-01 -3.67636621e-01
4.74848211e-01 4.86638732e-02 2.57320732e-01 -3.75681072e-01
1.04338257e-02 3.28268707e-01 8.54443982e-02 -7.24952817e-01
9.09951508e-01 5.05974531e-01 7.26037502e-01 -7.38746762e-01
-4.14131880e-01 -1.99258327e-01 -7.07664549e-01 -1.84434727e-01
1.51129544e-01 -3.55971456e-01 -8.81962538e-01 2.05496967e-01
-9.51907337e-01 -4.12052661e-01 -4.89101298e-02 3.14055532e-01
-6.29133821e-01 6.80377662e-01 -6.30305409e-01 -1.03585243e+00
2.92251527e-01 -1.11726916e+00 5.53198993e-01 -4.68007326e-01
-7.15085030e-01 -1.28671372e+00 2.78977454e-02 1.59897476e-01
2.57376552e-01 4.36099470e-01 1.49674416e+00 -1.05578494e+00
-3.85267198e-01 -5.07957220e-01 -2.88919717e-01 2.15492502e-01
-1.65252492e-01 7.04981804e-01 -6.25768900e-01 -4.27582204e-01
-4.04440552e-01 1.09981246e-01 7.60756910e-01 2.55733192e-01
9.92396712e-01 -3.83189410e-01 -5.77591300e-01 2.00958505e-01
1.50111628e+00 -1.28848895e-01 8.59348297e-01 1.99706838e-01
2.41888076e-01 7.89037764e-01 1.20344169e-01 2.57879317e-01
1.27927870e-01 7.19339490e-01 6.97441339e-01 4.57332015e-01
2.74982929e-01 -4.75527555e-01 3.23962085e-02 9.31963086e-01
-2.07271725e-01 -3.19754004e-01 -1.30666602e+00 4.05264169e-01
-1.65364766e+00 -1.14294481e+00 -5.54987907e-01 2.42106986e+00
5.40997028e-01 9.98725951e-01 5.09069502e-01 6.49720907e-01
7.71988451e-01 3.19040298e-01 -2.24768683e-01 -2.69147724e-01
-3.21168043e-02 5.63135557e-02 6.65029407e-01 5.24897575e-01
-5.05158901e-01 3.54433477e-01 8.10505676e+00 1.03669417e+00
-5.80993056e-01 -2.37674654e-01 5.58584034e-01 3.14402044e-01
-6.03733242e-01 6.48105621e-01 -5.12809575e-01 5.87731719e-01
1.25779736e+00 -1.89951882e-01 4.74497706e-01 5.41342199e-01
2.55980104e-01 -1.49410427e-01 -1.07038307e+00 3.65940571e-01
-5.18856525e-01 -1.46283972e+00 8.90820026e-02 6.21995389e-01
7.63818383e-01 -1.00609511e-01 -3.36799949e-01 -6.21383376e-02
4.54664022e-01 -1.39027214e+00 6.44334853e-01 7.12023675e-01
8.01807642e-01 -6.45538568e-01 6.13388419e-01 5.74034750e-01
-1.04048812e+00 1.22233205e-01 -1.95866510e-01 -3.37235659e-01
-6.61491528e-02 7.30899513e-01 -6.59205198e-01 6.34887874e-01
3.77359837e-01 5.73772907e-01 -5.28498650e-01 1.09663045e+00
1.75862461e-01 9.45484221e-01 -7.70264804e-01 -1.59244612e-02
1.68314904e-01 -2.91355878e-01 5.99528491e-01 9.40379739e-01
4.07748893e-02 -3.99581760e-01 -7.88533986e-02 8.56844068e-01
1.41943201e-01 -2.55261064e-01 -7.87230909e-01 -3.06337029e-01
7.49489903e-01 9.50060010e-01 -1.15246928e+00 -2.11827293e-01
-1.49585307e-01 1.71826687e-02 1.87489659e-01 2.67965198e-01
-6.81195438e-01 -4.07446250e-02 1.76172018e-01 7.41157949e-01
3.30657363e-01 -3.93876940e-01 -2.03917965e-01 -9.71651077e-01
-1.17851444e-01 -8.43433619e-01 1.59366474e-01 -6.58806741e-01
-1.23309100e+00 6.44352853e-01 2.17625484e-01 -1.14760137e+00
-3.13786209e-01 -6.19469404e-01 -6.34494066e-01 4.83074963e-01
-1.02835429e+00 -8.93277168e-01 -1.81513995e-01 1.87503651e-01
-3.30337375e-01 2.42373079e-01 7.08112597e-01 6.60388619e-02
-4.71424162e-01 2.72220105e-01 1.32522419e-01 -2.89269924e-01
2.74831802e-01 -1.12766671e+00 3.52098912e-01 6.24712348e-01
1.50344282e-01 6.74450338e-01 1.31818509e+00 -5.87270617e-01
-1.12591064e+00 -8.71652305e-01 8.16607416e-01 -2.83220530e-01
1.02567065e+00 -4.19172227e-01 -8.74430060e-01 7.03949869e-01
-3.52798462e-01 1.82424501e-01 2.53738582e-01 2.86586225e-01
-5.08246660e-01 2.52046273e-03 -1.04307854e+00 5.57420492e-01
1.26432419e+00 -5.25527775e-01 -2.21851781e-01 2.33754709e-01
3.58125061e-01 2.60634154e-01 -1.26081192e+00 3.95162970e-01
5.39755285e-01 -1.31264210e+00 6.85760319e-01 -6.41483068e-01
4.92873877e-01 -1.04186177e-01 -2.68486679e-01 -8.20683479e-01
-3.26694399e-01 -7.80750871e-01 -2.13984013e-01 1.19018328e+00
4.75583673e-01 -8.71190786e-01 9.09870684e-01 4.65986371e-01
3.69651675e-01 -8.88933778e-01 -1.23389602e+00 -1.12553632e+00
1.83824345e-01 -6.11581683e-01 7.02943623e-01 7.80593932e-01
2.06116572e-01 1.09885059e-01 -7.18919933e-02 -1.30914316e-01
6.38842762e-01 -5.28730527e-02 7.16900289e-01 -1.74720860e+00
-2.66659647e-01 -8.19391668e-01 -6.37672842e-01 -8.38714004e-01
2.58086860e-01 -8.50322604e-01 -3.94801468e-01 -1.44104850e+00
2.87856221e-01 -7.79346883e-01 1.23104057e-03 8.29521660e-03
4.73227084e-01 4.27383572e-01 -1.61830485e-01 4.67483431e-01
-7.00402260e-01 2.59013951e-01 9.77494121e-01 1.93801120e-01
3.58617455e-01 2.76274085e-01 -6.03165329e-01 9.25781071e-01
5.49572349e-01 -6.27755523e-01 -2.44350493e-01 3.24329078e-01
9.76704776e-01 3.19105566e-01 7.35152185e-01 -8.41141105e-01
2.50047743e-01 -1.08758405e-01 4.70942110e-02 -4.22029644e-01
2.91774362e-01 -9.98626053e-01 7.54978597e-01 5.44015586e-01
-4.74839211e-01 1.67906940e-01 -6.20734245e-02 8.31433058e-01
-7.33631104e-02 -4.85022783e-01 6.04627013e-01 -2.86322180e-02
-4.69884649e-02 1.74177155e-01 -4.90459561e-01 -1.82475690e-02
7.97188163e-01 -5.44648111e-01 -3.57571512e-01 -5.31765640e-01
-6.91343248e-01 -8.10863972e-02 1.00748527e+00 -2.10604385e-01
1.44036129e-01 -1.14090943e+00 -3.41766894e-01 1.04709752e-01
1.68636680e-01 -2.36556634e-01 -1.70239747e-01 1.15448380e+00
-6.97997272e-01 4.80153412e-01 1.50955141e-01 -5.91117144e-01
-1.11582267e+00 5.72445393e-01 5.37043393e-01 -6.42733872e-01
-3.05672318e-01 1.25813007e-01 -1.99996114e-01 -3.37691188e-01
-1.07705094e-01 -4.64676887e-01 2.72519737e-01 -2.12307066e-01
-4.64427806e-02 5.72294593e-01 -2.75439639e-02 -7.37632155e-01
-4.20172483e-01 2.57664621e-01 1.55656055e-01 1.22452922e-01
1.30203581e+00 -1.70363471e-01 -5.40729105e-01 4.61002529e-01
1.01810348e+00 1.32437527e-01 -1.02063286e+00 -4.33813855e-02
2.30822548e-01 -2.21121818e-01 -3.77896667e-01 -3.82812887e-01
-7.58529007e-01 6.42434418e-01 -2.87384182e-01 1.15918911e+00
6.55691564e-01 1.64331734e-01 1.31921992e-01 1.50439456e-01
5.93557358e-01 -6.36766374e-01 -3.97560090e-01 3.70975167e-01
5.39177179e-01 -8.97012889e-01 3.68560106e-01 -6.89157426e-01
-2.26252198e-01 1.11861360e+00 1.10285684e-01 -4.78194505e-01
8.31184983e-01 1.70208260e-01 -7.38466203e-01 -3.18276793e-01
-1.12896860e+00 1.24597894e-02 1.37351364e-01 6.21187329e-01
-6.27897382e-02 -5.40479533e-02 -2.30221614e-01 3.94827008e-01
-2.76131332e-01 -3.23220551e-01 7.89126098e-01 3.29801559e-01
-2.69890130e-01 -1.09540939e+00 -4.73332666e-02 7.49266088e-01
-3.74815196e-01 6.06886894e-02 -5.49569905e-01 1.30028725e+00
-3.05549026e-01 9.53469336e-01 1.30837306e-01 -2.99462289e-01
-3.70847434e-02 -1.41562680e-02 5.39668620e-01 -2.92230487e-01
-3.29328746e-01 -8.23292732e-02 6.24867320e-01 -3.38897020e-01
-5.99834323e-01 -3.88633996e-01 -5.21596074e-01 -9.70578611e-01
-6.19854212e-01 2.09388793e-01 2.19333857e-01 1.02733338e+00
5.37195392e-02 1.29919097e-01 2.88643539e-01 -4.90734756e-01
-7.29088902e-01 -9.45201635e-01 -8.88365209e-01 1.54146403e-01
2.03552961e-01 -7.99098849e-01 -6.95542812e-01 -4.08842742e-01] | [6.910609245300293, 5.3136420249938965] |
844f362e-5cfd-4728-97ff-670605661c83 | topological-guided-actor-critic-modular | 2304.10041 | null | https://arxiv.org/abs/2304.10041v1 | https://arxiv.org/pdf/2304.10041v1.pdf | Topological Guided Actor-Critic Modular Learning of Continuous Systems with Temporal Objectives | This work investigates the formal policy synthesis of continuous-state stochastic dynamic systems given high-level specifications in linear temporal logic. To learn an optimal policy that maximizes the satisfaction probability, we take a product between a dynamic system and the translated automaton to construct a product system on which we solve an optimal planning problem. Since this product system has a hybrid product state space that results in reward sparsity, we introduce a generalized optimal backup order, in reverse to the topological order, to guide the value backups and accelerate the learning process. We provide the optimality proof for using the generalized optimal backup order in this optimal planning problem. Further, this paper presents an actor-critic reinforcement learning algorithm when topological order applies. This algorithm leverages advanced mathematical techniques and enjoys the property of hyperparameter self-tuning. We provide proof of the optimality and convergence of our proposed reinforcement learning algorithm. We use neural networks to approximate the value function and policy function for hybrid product state space. Furthermore, we observe that assigning integer numbers to automaton states can rank the value or policy function approximated by neural networks. To break the ordinal relationship, we use an individual neural network for each automaton state's value (policy) function, termed modular learning. We conduct two experiments. First, to show the efficacy of our reinforcement learning algorithm, we compare it with baselines on a classic control task, CartPole. Second, we demonstrate the empirical performance of our formal policy synthesis framework on motion planning of a Dubins car with a temporal specification. | ['Zhentian Qian', 'Lening Li'] | 2023-04-20 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 1.06921516e-01 5.19552886e-01 -8.77864003e-01 5.30283339e-02
-8.15700293e-01 -6.20417118e-01 4.19786960e-01 -1.38724893e-01
-2.48991609e-01 8.36565912e-01 9.02341008e-02 -7.46297061e-01
-4.05975401e-01 -6.99191451e-01 -1.02643192e+00 -5.70911109e-01
-4.36779946e-01 4.44963276e-01 5.51369153e-02 -3.84341687e-01
8.09704512e-02 4.07231957e-01 -1.16807735e+00 -4.60776575e-02
7.47196138e-01 1.05909860e+00 2.27899179e-01 5.98014057e-01
3.43123645e-01 8.87458980e-01 -3.12690973e-01 1.37000099e-01
6.40905619e-01 -2.78330922e-01 -8.22447121e-01 1.77434713e-01
5.63099273e-02 -6.43371761e-01 -3.42175007e-01 9.70257163e-01
-5.72198676e-03 3.62475365e-01 6.65342927e-01 -1.67859411e+00
-3.83216649e-01 9.63863730e-01 -1.55492499e-01 -2.80461907e-01
1.12522647e-01 6.75424933e-01 1.40109253e+00 -2.10355952e-01
5.20845115e-01 1.34532571e+00 3.10251206e-01 5.50578237e-01
-1.37085426e+00 -2.58769214e-01 3.54444295e-01 -8.81165266e-02
-1.08549929e+00 -1.12239964e-01 6.38822079e-01 -3.65040600e-01
1.17622423e+00 -2.38884091e-01 8.86690259e-01 8.59836817e-01
6.36487961e-01 9.48748648e-01 1.02597141e+00 -2.21430212e-01
5.90927422e-01 -2.86144674e-01 -2.36140162e-01 1.00531530e+00
-1.18115067e-01 8.22602212e-01 1.43853366e-01 -2.63126016e-01
8.73720288e-01 -4.86222021e-02 1.41443044e-01 -6.06308877e-01
-9.36605692e-01 9.52942073e-01 3.00345510e-01 -1.13222517e-01
-4.14774269e-01 9.01467621e-01 2.27652118e-01 5.19482195e-01
-1.46622092e-01 9.54340219e-01 -6.31691992e-01 -4.68541905e-02
-5.69688082e-01 6.07722878e-01 9.50186610e-01 9.40675259e-01
6.58282578e-01 4.10980076e-01 -3.84705156e-01 4.27378356e-01
3.48885566e-01 6.01894736e-01 8.15448388e-02 -1.79996753e+00
2.94701576e-01 3.46060663e-01 5.37411630e-01 -5.22115648e-01
-4.28772390e-01 -4.06465977e-01 -4.23353851e-01 6.01745784e-01
4.73564208e-01 -3.56597245e-01 -7.23794818e-01 2.11179733e+00
1.58260748e-01 1.55731127e-01 2.95825183e-01 7.38683105e-01
-5.59070230e-01 8.58115494e-01 -7.14620352e-02 -4.75506008e-01
1.12828338e+00 -9.13493097e-01 -5.73757827e-01 -4.65205647e-02
6.92258298e-01 -1.38876408e-01 1.25790429e+00 4.80542779e-01
-1.25921762e+00 -2.54288077e-01 -1.38611424e+00 4.64313984e-01
4.63644014e-04 -5.57020456e-02 6.35586441e-01 2.61795551e-01
-1.04714370e+00 9.91987407e-01 -1.19437253e+00 2.75467020e-02
1.35461941e-01 5.80826640e-01 2.18685731e-01 3.21694165e-01
-1.31141567e+00 9.86424148e-01 5.50911963e-01 -2.44537726e-01
-1.52045035e+00 -6.22801542e-01 -8.54730546e-01 1.91798463e-01
1.06203365e+00 -6.40300095e-01 1.86440313e+00 -7.22986281e-01
-2.05759001e+00 -6.24873303e-03 3.25380385e-01 -8.54511678e-01
2.16548190e-01 3.79974358e-02 -1.45721793e-01 3.57286893e-02
-3.97861488e-02 8.02618980e-01 1.02303958e+00 -1.08412313e+00
-9.55469489e-01 1.76877856e-01 6.51509702e-01 2.77773216e-02
-6.76194876e-02 -5.98260283e-01 -3.56096700e-02 -3.43372434e-01
-2.16784477e-01 -1.36469531e+00 -6.58624411e-01 -3.82802576e-01
-1.70167431e-01 -4.17765796e-01 3.77066106e-01 -2.80270159e-01
1.40256834e+00 -1.92655575e+00 3.58291000e-01 4.16890889e-01
-1.64627403e-01 -2.58835942e-01 -2.79392719e-01 4.20164198e-01
1.64139256e-01 1.00549452e-01 -3.09469968e-01 2.21531908e-03
4.49750185e-01 5.71958184e-01 -5.42149603e-01 4.42803591e-01
3.16396236e-01 1.00506353e+00 -9.48792338e-01 -2.89416403e-01
-1.08739004e-01 -3.04302484e-01 -9.53243911e-01 5.01612276e-02
-1.04367912e+00 1.88398734e-01 -6.32321119e-01 5.30449390e-01
1.00734748e-01 -1.16700254e-01 5.25007069e-01 2.20731068e-02
-1.40301540e-01 2.33301073e-01 -1.17780209e+00 1.44249272e+00
-6.69331729e-01 9.70985740e-03 6.75996095e-02 -1.01439881e+00
7.25750625e-01 1.16887569e-01 7.69553006e-01 -7.58040667e-01
1.40432775e-01 2.76072264e-01 -5.41792847e-02 -2.01896131e-01
6.97916150e-01 -3.40232074e-01 -6.35242164e-01 6.86936319e-01
-1.13784701e-01 -5.07635593e-01 2.44263574e-01 1.20243737e-02
1.22011459e+00 4.64775831e-01 2.49317825e-01 -2.72134006e-01
3.94960582e-01 2.17196658e-01 6.46399200e-01 8.85334373e-01
-1.85941339e-01 -2.45866179e-01 1.22908235e+00 -3.53950709e-01
-1.19589162e+00 -9.50805843e-01 3.03531557e-01 1.07781065e+00
2.09232680e-02 -3.11913371e-01 -5.89743853e-01 -7.83906221e-01
2.43197203e-01 9.57675755e-01 -5.18777192e-01 -3.81830335e-01
-7.46284306e-01 -7.07846880e-02 4.08542275e-01 4.62922096e-01
2.18427852e-01 -1.01421821e+00 -9.46721673e-01 3.46349776e-01
2.58163899e-01 -7.86137879e-01 -8.51441264e-01 3.86111349e-01
-8.58564794e-01 -7.92651474e-01 -1.68942258e-01 -6.04519308e-01
4.41896558e-01 -3.23790699e-01 8.69904876e-01 -1.32559389e-01
4.73951221e-01 6.59747362e-01 1.94880664e-01 -3.43920328e-02
-8.00032854e-01 3.85928035e-01 3.09562802e-01 -3.31371695e-01
-4.93640929e-01 -4.88104761e-01 -3.01139534e-01 1.89246982e-01
-9.60889816e-01 4.56107259e-02 5.54492056e-01 1.17832613e+00
7.19771385e-01 3.20310444e-01 5.35517752e-01 -3.44475210e-01
8.23557496e-01 -3.21024716e-01 -1.46880841e+00 2.65505344e-01
-8.47576618e-01 9.01813745e-01 8.55400026e-01 -5.51222146e-01
-7.45131493e-01 4.46183652e-01 2.40189865e-01 -4.93297666e-01
2.78444797e-01 7.33225465e-01 -3.55173051e-02 2.95636564e-01
4.72265273e-01 -4.71650027e-02 3.69389832e-01 1.93377063e-01
5.96661508e-01 1.85927376e-01 5.21459103e-01 -1.22703922e+00
6.70420766e-01 2.10654140e-02 3.20736140e-01 -1.57593712e-01
-6.16227925e-01 1.84152544e-01 -4.40557227e-02 -1.14683278e-01
5.85180402e-01 -5.36382437e-01 -1.45578170e+00 -4.58077807e-03
-9.31295633e-01 -1.01645911e+00 -6.38729990e-01 2.76312202e-01
-1.41590440e+00 -6.50293827e-02 -5.38381398e-01 -1.17323327e+00
1.40958399e-01 -1.40020263e+00 8.89111876e-01 2.71492768e-02
-1.24605194e-01 -7.85949588e-01 2.18854636e-01 -3.09973449e-01
1.85881704e-01 3.03260714e-01 1.19730330e+00 -2.84694791e-01
-8.59842658e-01 6.87692985e-02 3.66280764e-01 2.78579295e-01
-1.50347605e-01 -1.52784839e-01 -3.27135652e-01 -4.22776937e-01
-1.04854271e-01 -4.09996480e-01 5.12218237e-01 4.83858109e-01
8.13378930e-01 -8.50555837e-01 -2.21547380e-01 3.43870103e-01
1.47984409e+00 6.10954463e-01 3.45435321e-01 4.62500066e-01
3.07034433e-01 5.07402360e-01 9.66711819e-01 6.55366123e-01
3.78905296e-01 6.65675402e-01 6.81170106e-01 6.28925323e-01
4.39122975e-01 -6.50271595e-01 8.74556303e-01 1.99730784e-01
1.89111620e-01 4.73258980e-02 -8.27227235e-01 4.42517966e-01
-2.12167525e+00 -9.06841874e-01 8.28491390e-01 2.17656040e+00
9.66639817e-01 4.34635043e-01 5.38365066e-01 -1.98815092e-01
2.74556339e-01 -1.02158442e-01 -9.53549922e-01 -7.93872833e-01
3.01896691e-01 1.17096327e-01 9.27897274e-01 8.77827048e-01
-9.09644127e-01 1.11441481e+00 6.44796753e+00 8.84720027e-01
-1.07594001e+00 -1.09096423e-01 5.44285357e-01 -5.12455329e-02
-3.65270674e-01 1.51336238e-01 -7.74938941e-01 2.02604204e-01
1.27708471e+00 -2.98104703e-01 1.23289418e+00 1.09125888e+00
5.25712490e-01 -3.90055403e-02 -1.33969486e+00 2.36513555e-01
-7.80811965e-01 -1.46499932e+00 -3.22600782e-01 3.07913542e-01
8.69250238e-01 -1.72929779e-01 3.64749163e-01 7.95544147e-01
1.01452804e+00 -9.97694075e-01 1.02850413e+00 3.79768074e-01
6.23237431e-01 -1.11626101e+00 3.39353740e-01 4.72631186e-01
-1.10139096e+00 -7.72717416e-01 -1.19825890e-02 -1.94184795e-01
3.74732971e-01 -1.54036120e-01 -9.48364854e-01 3.06024939e-01
3.00143789e-02 4.77075547e-01 -1.65343322e-02 4.89041150e-01
-2.22092941e-01 3.99775922e-01 -4.72468913e-01 -2.30042696e-01
6.29023969e-01 -3.03915024e-01 4.48485672e-01 6.79797351e-01
2.95239508e-01 1.97753850e-02 6.73529506e-01 1.14140427e+00
1.92298457e-01 -4.05242175e-01 -7.66648829e-01 -5.38779736e-01
3.68698537e-01 9.57450509e-01 -6.29179418e-01 -1.94645330e-01
-1.97010398e-01 3.59937161e-01 3.33772838e-01 4.89537686e-01
-1.03459525e+00 1.65097583e-02 7.08791852e-01 -1.95924759e-01
4.42593038e-01 -6.36417627e-01 -2.45243251e-01 -6.57652378e-01
-2.54650027e-01 -1.23788106e+00 3.67887676e-01 -3.85041505e-01
-9.67770219e-01 1.61303461e-01 3.75925899e-01 -1.31498051e+00
-8.73301506e-01 -5.66020072e-01 -2.56199718e-01 5.18983662e-01
-1.14427745e+00 -8.11250091e-01 6.29220366e-01 3.42441469e-01
4.07867223e-01 -3.09173286e-01 5.44761121e-01 -2.51925498e-01
-4.86650437e-01 4.70363587e-01 -1.57519445e-01 -4.32961971e-01
2.38276526e-01 -1.34597850e+00 1.85691729e-01 5.85462630e-01
-3.89053464e-01 5.61610818e-01 7.92410553e-01 -7.66512692e-01
-2.00606513e+00 -1.02255142e+00 1.06368333e-01 -2.68088877e-01
1.11145437e+00 1.26393452e-01 -5.10805130e-01 8.07651162e-01
1.98722288e-01 -1.28599301e-01 -1.21825822e-01 -1.32353634e-01
-1.16886489e-01 -2.16364041e-01 -1.06512070e+00 1.02848041e+00
8.80539238e-01 -3.19747061e-01 -4.33761388e-01 2.10965142e-01
1.20216548e+00 -5.89372098e-01 -9.94792402e-01 4.30046141e-01
7.05244362e-01 -3.38841796e-01 7.91003168e-01 -8.85820627e-01
5.23873031e-01 -5.10169804e-01 -4.24169987e-01 -1.45516622e+00
-4.88766611e-01 -1.11858809e+00 -6.86168015e-01 6.04567528e-01
6.22691989e-01 -6.07998073e-01 7.14699328e-01 6.25699878e-01
-4.14690822e-01 -1.19930863e+00 -9.54856098e-01 -1.16813922e+00
4.62618500e-01 -5.70457220e-01 7.53166258e-01 3.82171065e-01
2.85103291e-01 2.64158398e-01 -4.83970642e-01 2.70985276e-01
3.49878848e-01 1.36585698e-01 4.90913332e-01 -5.47409296e-01
-9.08399463e-01 -6.58183992e-01 3.38213980e-01 -1.18368363e+00
6.59262896e-01 -8.38254929e-01 3.87208909e-01 -1.10260701e+00
-2.95991123e-01 -4.41469908e-01 -4.13758397e-01 6.30815446e-01
3.65672648e-01 -6.18587136e-01 4.81520146e-01 7.42114615e-03
-6.70686483e-01 6.53669953e-01 1.50459158e+00 -4.55372989e-01
-7.05595434e-01 7.85875618e-02 -6.24947488e-01 3.51509959e-01
8.93681407e-01 -2.97391683e-01 -6.58054709e-01 -2.63140853e-02
5.15653133e-01 7.27961659e-01 1.46528810e-01 -7.53399014e-01
1.32748485e-01 -9.30357695e-01 -3.06692094e-01 -4.28998113e-01
1.50581941e-01 -9.39084172e-01 7.90029392e-02 9.60277140e-01
-7.22600758e-01 2.77593434e-01 1.49735272e-01 5.85228264e-01
1.88766256e-01 -1.86794490e-01 7.28986323e-01 4.18511033e-02
-6.38956964e-01 3.76511633e-01 -6.63678050e-01 -1.33940816e-01
9.31670725e-01 2.17195198e-01 -7.93072674e-03 -4.95152563e-01
-7.44725823e-01 6.98653340e-01 4.04698849e-01 2.43803114e-01
4.26674813e-01 -1.50134397e+00 -9.84231308e-02 -3.96440504e-03
-1.05034597e-01 -2.91329831e-01 -1.13385715e-01 8.38908911e-01
-8.01930428e-02 4.68484432e-01 -2.68909276e-01 -4.68524665e-01
-3.62984836e-01 8.35584342e-01 5.30358136e-01 -7.55983889e-01
-3.73796850e-01 9.45813023e-03 -1.47681072e-01 -3.76778483e-01
7.79651478e-02 -8.19510877e-01 8.32433403e-02 -1.45572588e-01
1.22658938e-01 3.37289602e-01 -4.03149754e-01 4.25909944e-02
-5.55882603e-02 1.59576371e-01 1.63048938e-01 -7.94064045e-01
1.07760680e+00 9.87183005e-02 1.77972704e-01 3.64966482e-01
8.15520525e-01 -4.78567868e-01 -1.75169730e+00 -1.10264257e-01
2.45886594e-01 1.07214302e-01 -5.15305959e-02 -6.28058255e-01
-8.28312039e-01 2.01985791e-01 4.78724152e-01 4.55073833e-01
9.68170583e-01 -1.03804320e-01 7.00592995e-01 8.39184761e-01
5.56758404e-01 -1.53556132e+00 3.28401059e-01 1.12027395e+00
8.16930115e-01 -8.33494067e-01 -1.79701149e-01 2.13895053e-01
-7.94600666e-01 1.04543686e+00 6.37616813e-01 -4.49123472e-01
3.47527266e-01 5.66703618e-01 -4.41902399e-01 8.81929323e-02
-1.13322294e+00 -2.79392362e-01 -7.74260238e-03 3.89509022e-01
-7.60450140e-02 2.78718263e-01 -2.90911853e-01 6.10806167e-01
-2.16090366e-01 1.03055105e-01 5.51391721e-01 1.03091550e+00
-5.12196243e-01 -1.22384048e+00 -2.94720381e-01 1.66678652e-01
-9.77548659e-02 2.94711798e-01 1.66591763e-01 8.55548561e-01
-1.52809247e-01 7.07398295e-01 5.95902242e-02 -3.92909765e-01
4.06138927e-01 8.99642035e-02 7.61220515e-01 -5.67907393e-01
-2.67921299e-01 1.53309211e-01 5.67296445e-02 -8.79753888e-01
-5.53651201e-03 -6.46563709e-01 -1.65456283e+00 -1.76844358e-01
8.46975446e-02 7.07685053e-02 5.80498159e-01 1.11807585e+00
3.03520799e-01 7.25087762e-01 8.33392620e-01 -7.18262732e-01
-1.35312331e+00 -4.57085222e-01 -4.99601305e-01 -7.51953050e-02
5.19399822e-01 -7.65083253e-01 -9.57272127e-02 -1.66746646e-01] | [4.449658393859863, 2.07556414604187] |
a432dc28-6673-4405-b137-7926ee312333 | variational-autoencoder-with-disentanglement | 2202.13363 | null | https://arxiv.org/abs/2202.13363v3 | https://arxiv.org/pdf/2202.13363v3.pdf | Variational Autoencoder with Disentanglement Priors for Low-Resource Task-Specific Natural Language Generation | In this paper, we propose a variational autoencoder with disentanglement priors, VAE-DPRIOR, for task-specific natural language generation with none or a handful of task-specific labeled examples. In order to tackle compositional generalization across tasks, our model performs disentangled representation learning by introducing a conditional prior for the latent content space and another conditional prior for the latent label space. Both types of priors satisfy a novel property called $\epsilon$-disentangled. We show both empirically and theoretically that the novel priors can disentangle representations even without specific regularizations as in the prior work. The content prior enables directly sampling diverse content representations from the content space learned from the seen tasks, and fuse them with the representations of novel tasks for generating semantically diverse texts in the low-resource settings. Our extensive experiments demonstrate the superior performance of our model over competitive baselines in terms of i) data augmentation in continuous zero/few-shot learning, and ii) text style transfer in the few-shot setting. | ['Gholamreza Haffari', 'Tianyang Zhan', 'Tongtong Wu', 'Qiongkai Xu', 'Lizhen Qu', 'Zhuang Li'] | 2022-02-27 | null | null | null | null | ['text-style-transfoer'] | ['natural-language-processing'] | [ 4.65411842e-01 2.79013097e-01 -2.04988822e-01 -2.55721390e-01
-1.01185143e+00 -4.39174980e-01 1.00989389e+00 -5.11051655e-01
-3.80675077e-01 9.54918563e-01 7.56718040e-01 3.33895385e-02
1.97029501e-01 -7.60146618e-01 -5.42060912e-01 -7.74429917e-01
7.04171062e-01 8.34912956e-01 -3.82564694e-01 -3.70403320e-01
-2.05641121e-01 -5.58180101e-02 -1.63337982e+00 3.58794034e-01
8.93197596e-01 6.13899291e-01 3.86112958e-01 6.45433664e-01
-2.26706639e-01 7.80088067e-01 -6.58860028e-01 -5.84193349e-01
2.90095091e-01 -5.89987397e-01 -6.03780329e-01 1.83105528e-01
2.76734352e-01 -4.64299589e-01 -5.47847927e-01 8.53091657e-01
6.45667851e-01 5.44530511e-01 1.23470736e+00 -1.33977497e+00
-1.34403586e+00 9.30365205e-01 -5.35437226e-01 -2.02164948e-02
2.45641433e-02 3.84214014e-01 1.41397667e+00 -1.22128057e+00
6.92524970e-01 1.30978131e+00 1.40995886e-02 1.10733628e+00
-1.58419812e+00 -8.99290979e-01 2.04864264e-01 -9.00602117e-02
-1.10408056e+00 -5.96633017e-01 8.55765402e-01 -5.57433963e-01
8.82859349e-01 -5.38276173e-02 2.29168653e-01 2.08512878e+00
-3.06387603e-01 1.03432310e+00 8.38471115e-01 -3.67451370e-01
2.17766464e-01 1.65717021e-01 3.67198959e-02 5.21122992e-01
4.17376965e-01 1.86617136e-01 -7.20182240e-01 -2.00289518e-01
7.87490785e-01 1.68130845e-01 -4.36616212e-01 -4.49265301e-01
-1.60927045e+00 1.31544161e+00 1.40040129e-01 3.38626541e-02
-2.36777619e-01 3.11716884e-01 4.39332783e-01 2.27729410e-01
8.66665125e-01 6.37447298e-01 -4.01145190e-01 9.49607491e-02
-7.95221865e-01 2.61049479e-01 6.94815040e-01 1.50696874e+00
6.63868666e-01 7.20531642e-01 -8.73838067e-01 8.43148410e-01
2.09165826e-01 5.20356238e-01 8.11787307e-01 -7.64183879e-01
6.78719699e-01 1.82227269e-01 2.67150998e-01 -3.75621825e-01
2.02080026e-01 -2.72659510e-01 -1.03868699e+00 6.43329397e-02
2.32200012e-01 -4.72592264e-01 -1.05660355e+00 2.23209810e+00
-7.85768032e-02 3.69409561e-01 4.41717744e-01 7.46577859e-01
1.00532556e+00 8.26034129e-01 1.67934000e-01 -1.10999485e-02
1.52178657e+00 -1.13234794e+00 -9.92778957e-01 -3.41490060e-01
3.21788341e-01 -5.09688437e-01 1.48746777e+00 2.27170866e-02
-1.13393748e+00 -6.58867121e-01 -1.10787225e+00 -5.58000565e-01
-2.11653665e-01 3.29593688e-01 4.27262992e-01 3.39423776e-01
-8.12588334e-01 3.14412385e-01 -5.71685851e-01 2.62990557e-02
3.60230893e-01 -1.50981382e-01 -1.98650062e-01 -2.00622007e-01
-1.46850467e+00 7.67294168e-01 5.17901361e-01 -3.95068496e-01
-1.35161209e+00 -7.70985484e-01 -1.18040764e+00 3.61070573e-01
4.46791649e-01 -1.27786195e+00 1.30433130e+00 -6.37026191e-01
-1.72002339e+00 8.53247285e-01 -1.27344146e-01 -2.74941087e-01
5.35778880e-01 -3.17723870e-01 -1.60114914e-01 -4.16470282e-02
2.16738239e-01 7.88392246e-01 1.16670394e+00 -1.22404778e+00
-1.00385301e-01 -1.84700880e-02 1.24337068e-02 4.70799178e-01
-3.89494658e-01 -4.22438025e-01 1.75438393e-02 -1.09805596e+00
-3.53408247e-01 -7.31972992e-01 -4.39809039e-02 -1.05582140e-01
-3.71576756e-01 -5.28593481e-01 4.77718592e-01 -6.47848070e-01
5.72612464e-01 -2.11857414e+00 7.01237023e-01 -5.04076004e-01
3.41029823e-01 -4.54952568e-02 -5.27800858e-01 4.65159029e-01
-6.81193545e-02 2.19513476e-01 -2.53995866e-01 -8.47184181e-01
3.81233543e-01 3.01671773e-01 -8.33696842e-01 1.23328008e-01
4.44733530e-01 9.92775202e-01 -1.22692037e+00 -2.40186498e-01
-8.62598345e-02 4.07305390e-01 -5.83386540e-01 5.80757439e-01
-5.86836278e-01 3.99509698e-01 -3.78509194e-01 -3.44259571e-03
3.70755553e-01 -5.05090773e-01 9.14753508e-03 -2.12887954e-02
3.63403708e-01 4.25164372e-01 -7.78290510e-01 2.20240426e+00
-7.14275181e-01 5.48099279e-01 -1.56824708e-01 -8.24016750e-01
9.88883317e-01 8.77573907e-01 -1.70178339e-01 -1.61128461e-01
2.95361072e-01 -5.81610985e-02 -1.75819471e-01 -4.78923470e-01
7.30387568e-01 -6.15105987e-01 -3.00535470e-01 1.00839818e+00
7.51241326e-01 -3.09317350e-01 2.45058089e-01 5.29322743e-01
6.15659177e-01 4.65893745e-01 3.03171366e-01 -3.80635142e-01
5.26944138e-02 -4.16434824e-01 4.36283022e-01 8.74680638e-01
1.24471844e-03 8.76397014e-01 4.59154367e-01 -4.02783304e-02
-1.29787695e+00 -1.46379006e+00 1.44066691e-01 1.46337914e+00
-1.38146743e-01 -1.53219059e-01 -5.00693440e-01 -5.90656936e-01
-1.86524555e-01 1.37059176e+00 -8.27367663e-01 -3.45364660e-01
-2.94863701e-01 -4.91474599e-01 5.21717370e-01 7.05727875e-01
2.30700746e-01 -1.14465976e+00 -3.50830913e-01 -4.67921346e-02
-4.56939280e-01 -1.12432384e+00 -7.27567017e-01 6.38610199e-02
-6.23360634e-01 -6.61337554e-01 -1.10424864e+00 -7.45563686e-01
5.13970792e-01 4.07528937e-01 1.34660673e+00 -4.97043580e-01
-8.82980973e-02 2.32200921e-01 -4.17261630e-01 -3.66988271e-01
-4.73695040e-01 5.52039705e-02 4.73914444e-02 -6.12997450e-03
4.09493953e-01 -7.53432274e-01 -5.21140933e-01 -8.88265893e-02
-9.71208870e-01 5.18357575e-01 5.68575025e-01 1.32567978e+00
3.49528611e-01 -6.05240166e-01 8.09449255e-01 -1.02826035e+00
1.01060188e+00 -8.08031738e-01 -2.79309273e-01 2.61770517e-01
-2.17619434e-01 5.49058616e-01 6.84323430e-01 -6.20883822e-01
-1.50570273e+00 -3.91870141e-01 2.65035242e-01 -7.72121131e-01
-2.42051125e-01 2.62811512e-01 -3.26785088e-01 1.02521133e+00
1.01038980e+00 3.55044156e-01 -4.67397785e-03 -3.72931302e-01
1.18092835e+00 4.58502769e-01 4.09703761e-01 -1.00965309e+00
8.76424730e-01 4.39098239e-01 -4.31614697e-01 -6.23701632e-01
-1.28479862e+00 -2.13401780e-01 -5.07351100e-01 4.20067161e-01
1.22701240e+00 -1.43708050e+00 -1.16346791e-01 5.45680299e-02
-1.44188023e+00 -2.66617596e-01 -8.07165802e-01 4.89579976e-01
-7.91306615e-01 1.24234661e-01 -6.26626611e-01 -6.35939896e-01
-5.33380568e-01 -1.15440929e+00 1.21970439e+00 1.02694556e-02
-3.01446050e-01 -1.11918795e+00 2.43454516e-01 4.40157473e-01
2.91533321e-01 1.74989015e-01 1.02302027e+00 -8.89327228e-01
-3.98095071e-01 4.45241779e-02 -3.13924164e-01 4.15477335e-01
1.14138409e-01 -5.24401426e-01 -1.24541056e+00 -3.48761678e-01
1.93094894e-01 -8.02219689e-01 1.17264247e+00 8.02760050e-02
1.06145275e+00 -4.29787427e-01 5.26876040e-02 6.20675504e-01
1.10711873e+00 -2.80260950e-01 4.35639322e-01 -4.26937550e-01
8.34290266e-01 6.13851011e-01 5.68878315e-02 5.86625159e-01
1.76239565e-01 3.43227088e-01 2.37842590e-01 2.10563764e-01
-2.82596856e-01 -5.18966377e-01 3.92890930e-01 9.83843863e-01
-1.78082943e-01 -5.12633801e-01 -4.65050995e-01 6.40881538e-01
-1.72574520e+00 -1.09071875e+00 3.91912282e-01 1.87650013e+00
1.19206071e+00 -2.62624860e-01 -7.38900527e-02 -4.48186606e-01
8.77333403e-01 5.64710677e-01 -6.82486176e-01 -1.87692627e-01
-7.95163065e-02 5.67524493e-01 -6.15699440e-02 4.31981295e-01
-8.03902686e-01 9.98444676e-01 5.78068256e+00 8.14509869e-01
-4.85027760e-01 4.18128729e-01 3.01855147e-01 -5.07507682e-01
-9.17593539e-01 -5.22167571e-02 -7.11033344e-01 3.31483662e-01
7.92306542e-01 -4.91470933e-01 3.74471039e-01 9.37976897e-01
-2.29390219e-01 5.77340007e-01 -1.45131195e+00 9.68689263e-01
4.12805736e-01 -1.26952493e+00 5.48015118e-01 -8.30399096e-02
1.09911799e+00 -8.44210610e-02 4.31277931e-01 7.12662697e-01
1.03026724e+00 -1.16372824e+00 6.56389177e-01 3.25263083e-01
1.18038344e+00 -5.48879027e-01 3.45338702e-01 5.51036716e-01
-6.63389802e-01 2.30220169e-01 -4.70690310e-01 -7.47793168e-02
2.67828047e-01 2.98734486e-01 -7.54638255e-01 5.20583391e-01
9.71342027e-02 6.27247274e-01 -5.68559952e-02 2.19545707e-01
-8.20320904e-01 4.90086228e-01 2.67491519e-01 -1.69202507e-01
2.18685567e-01 -2.42914900e-01 6.41243458e-01 1.06967962e+00
4.02052224e-01 2.19256669e-01 -9.01059713e-03 1.66947055e+00
-5.34982264e-01 -1.67058960e-01 -7.47545004e-01 -2.86931545e-01
4.39680427e-01 1.12734056e+00 -1.36848152e-01 -6.70493662e-01
-4.13427711e-01 1.23966110e+00 5.18109620e-01 8.70083332e-01
-6.53707802e-01 -4.72872525e-01 7.84643352e-01 -3.26313615e-01
3.48147035e-01 -1.86177716e-01 -2.33553454e-01 -1.91192055e+00
-2.53670305e-01 -7.14888752e-01 2.58421004e-01 -9.40801620e-01
-1.97117960e+00 6.32282972e-01 3.04532111e-01 -1.05589128e+00
-6.61114633e-01 -6.60516739e-01 -7.30611145e-01 1.40358829e+00
-1.37004542e+00 -1.24186742e+00 -6.48887679e-02 6.02423310e-01
1.11365569e+00 -5.99757373e-01 1.17018855e+00 -3.41153145e-01
-4.72704113e-01 6.54858649e-01 2.22831801e-01 1.65775999e-01
7.62657940e-01 -1.40132189e+00 6.91149414e-01 8.41369629e-01
3.03276747e-01 6.52069807e-01 8.04583192e-01 -5.15527487e-01
-9.10818338e-01 -1.11452091e+00 6.12404406e-01 -5.97552419e-01
6.36639357e-01 -8.73956621e-01 -8.04909170e-01 1.05055118e+00
8.02816212e-01 -1.45880952e-01 1.13856626e+00 2.28330567e-01
-8.95698071e-01 6.33069754e-01 -9.03057635e-01 9.96717334e-01
1.09184074e+00 -6.92812324e-01 -1.07124519e+00 4.41471487e-01
1.45735848e+00 -1.81168720e-01 -5.63460708e-01 1.72457956e-02
3.39258820e-01 -4.87687171e-01 8.57777178e-01 -1.14190805e+00
9.93004620e-01 1.12215392e-01 -3.28506470e-01 -1.73357856e+00
-5.10391474e-01 -7.01090515e-01 -4.48818415e-01 1.08121979e+00
4.91542578e-01 -4.00343984e-01 6.17283642e-01 4.81496245e-01
-2.79492885e-01 -4.35222298e-01 -7.14081168e-01 -7.05262423e-01
5.22572756e-01 -1.50718123e-01 6.03846371e-01 1.08103347e+00
-1.06039971e-01 1.04585350e+00 -7.66065419e-01 -2.54213393e-01
7.22824752e-01 1.22127227e-01 7.42878437e-01 -1.13917196e+00
-7.82628596e-01 -2.98161924e-01 2.23880962e-01 -1.20674527e+00
7.20613599e-01 -1.20428264e+00 3.15188989e-02 -1.57335413e+00
5.40036678e-01 6.33160174e-02 -2.07288817e-01 2.66651154e-01
-4.91093099e-01 -6.01664335e-02 2.11875305e-01 1.67257503e-01
-3.77537161e-01 1.28513205e+00 1.53647649e+00 -2.58965909e-01
8.77647772e-02 -2.94235647e-01 -1.03237760e+00 5.45708895e-01
6.73223674e-01 -4.04932648e-01 -9.10308123e-01 -8.28400075e-01
1.93375811e-01 1.04045972e-01 2.94499457e-01 -4.10553843e-01
-2.14253396e-01 -2.48904347e-01 3.19157213e-01 -4.43939865e-01
7.14580655e-01 -3.86620164e-01 -2.64971733e-01 7.00374600e-03
-8.55965078e-01 -2.66990304e-01 -8.07222500e-02 9.43883777e-01
9.41641454e-04 -3.98313642e-01 6.64958298e-01 -3.41248333e-01
-4.58153874e-01 4.81573820e-01 -2.41690904e-01 6.27654731e-01
7.58055091e-01 1.01312667e-01 -5.82641184e-01 -5.51986337e-01
-9.50236738e-01 1.60151899e-01 3.05286884e-01 6.11864984e-01
7.74950027e-01 -1.65050673e+00 -1.15003598e+00 1.83607474e-01
3.93620938e-01 2.21601367e-01 4.54289168e-01 1.91784605e-01
-1.67287420e-02 1.27650306e-01 -2.95405835e-01 -2.10660622e-01
-5.91525555e-01 7.12548614e-01 -3.95842008e-02 -4.05160934e-01
-6.43037975e-01 1.03844368e+00 7.68694818e-01 -4.81108248e-01
1.69463158e-02 3.89875546e-02 -1.25985414e-01 1.60379618e-01
5.80078542e-01 2.66183943e-01 -4.71543282e-01 -4.66611266e-01
3.34627718e-01 6.64735306e-03 -4.10639226e-01 -7.30082989e-01
1.24399590e+00 -1.59511808e-02 2.29874790e-01 9.00370061e-01
1.21358597e+00 -3.39907944e-01 -1.59688294e+00 -6.59432828e-01
-5.38924158e-01 -4.06817645e-01 -1.75233409e-01 -6.15180790e-01
-7.49008179e-01 1.36267114e+00 8.47913511e-03 -2.66377795e-02
5.92607677e-01 1.42198503e-01 7.08797097e-01 3.13073158e-01
1.51972724e-02 -8.88444662e-01 6.62377715e-01 6.71107948e-01
1.18883586e+00 -1.12532020e+00 -2.58776605e-01 -1.34195089e-01
-1.15318108e+00 7.18818963e-01 6.99438274e-01 -4.08535391e-01
3.91127050e-01 -6.73659518e-02 -2.54502237e-01 -5.78114204e-02
-1.20062351e+00 -1.96857661e-01 4.09328580e-01 7.95608759e-01
6.16293609e-01 2.09397599e-01 -2.95847915e-02 9.03769076e-01
-3.78439486e-01 -3.52538347e-01 5.42032182e-01 4.94131088e-01
-3.05787712e-01 -8.39019895e-01 1.19940192e-01 3.58228773e-01
-1.60136059e-01 -4.31806028e-01 -2.85535216e-01 6.76910102e-01
-1.38823986e-01 7.57978857e-01 2.98997164e-01 3.11187515e-03
9.85964984e-02 6.08114600e-01 5.47213256e-01 -1.12732136e+00
-2.17087612e-01 -6.37719184e-02 -1.48162665e-02 -3.26817073e-02
-5.63227795e-02 -4.97652650e-01 -8.53903472e-01 6.52907044e-02
-2.47246966e-01 1.98262647e-01 3.89967650e-01 8.36183786e-01
4.61994946e-01 7.87348330e-01 3.61933529e-01 -9.21126664e-01
-1.15893960e+00 -1.44124496e+00 -6.54379308e-01 8.20491314e-01
2.21534252e-01 -7.37054408e-01 -4.84326363e-01 2.72271425e-01] | [11.808056831359863, 9.090681076049805] |
3cf4b375-e285-4e51-b84b-aec61a6fa69c | naist-at-the-nli-2013-shared-task | null | null | https://aclanthology.org/W13-1717 | https://aclanthology.org/W13-1717.pdf | NAIST at the NLI 2013 Shared Task | null | ['Yuji Matsumoto', 'Keisuke Sakaguchi', 'Tomoya Mizumoto', 'Yuta Hayashibe', 'Mamoru Komachi'] | 2013-06-01 | null | null | null | ws-2013-6 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.198925495147705, 3.846264600753784] |
5932606e-21ba-4a5e-8726-49723a9f7cc9 | hignn-hierarchical-informative-graph-neural | 2208.13994 | null | https://arxiv.org/abs/2208.13994v1 | https://arxiv.org/pdf/2208.13994v1.pdf | HiGNN: Hierarchical Informative Graph Neural Networks for Molecular Property Prediction Equipped with Feature-Wise Attention | Elucidating and accurately predicting the druggability and bioactivities of molecules plays a pivotal role in drug design and discovery and remains an open challenge. Recently, graph neural networks (GNN) have made remarkable advancements in graph-based molecular property prediction. However, current graph-based deep learning methods neglect the hierarchical information of molecules and the relationships between feature channels. In this study, we propose a well-designed hierarchical informative graph neural networks framework (termed HiGNN) for predicting molecular property by utilizing a co-representation learning of molecular graphs and chemically synthesizable BRICS fragments. Furthermore, a plug-and-play feature-wise attention block is first designed in HiGNN architecture to adaptively recalibrate atomic features after the message passing phase. Extensive experiments demonstrate that HiGNN achieves state-of-the-art predictive performance on many challenging drug discovery-associated benchmark datasets. In addition, we devise a molecule-fragment similarity mechanism to comprehensively investigate the interpretability of HiGNN model at the subgraph level, indicating that HiGNN as a powerful deep learning tool can help chemists and pharmacists identify the key components of molecules for designing better molecules with desired properties or functions. The source code is publicly available at https://github.com/idruglab/hignn. | ['Ling Wang', 'Jianrong Xu', 'Duancheng Zhao', 'Yi Zhang', 'Weimin Zhu'] | 2022-08-30 | null | null | null | null | ['molecular-property-prediction'] | ['miscellaneous'] | [ 2.87702858e-01 -1.18487708e-01 -6.54132545e-01 -2.82791138e-01
-4.40132231e-01 -4.04723644e-01 2.30316013e-01 6.86418355e-01
2.25598976e-01 9.85686898e-01 7.52946138e-02 -7.65309215e-01
-4.37016696e-01 -9.28794026e-01 -9.15562689e-01 -9.39336717e-01
-2.90138096e-01 2.77233899e-01 -1.64275542e-01 -1.35806903e-01
1.82632223e-01 6.43950582e-01 -7.75529504e-01 3.67495149e-01
1.07743096e+00 7.77691305e-01 1.50061443e-01 3.53867054e-01
5.90260215e-02 9.72807109e-01 -1.63205385e-01 -4.11814928e-01
-1.37151450e-01 -5.71622908e-01 -7.11313844e-01 -2.34724104e-01
1.83404997e-01 -2.71369033e-02 -5.72277963e-01 9.92954791e-01
7.33647406e-01 1.71108559e-01 6.77303433e-01 -7.76725769e-01
-9.50457096e-01 5.16404748e-01 -3.08989882e-01 2.44268849e-01
4.05477583e-01 3.89588207e-01 1.35244191e+00 -7.53458500e-01
5.16759872e-01 9.80497658e-01 4.07212675e-01 3.75308573e-01
-1.20657408e+00 -8.80451798e-01 1.58572719e-01 6.75270379e-01
-1.29852557e+00 -2.00122133e-01 8.48711610e-01 -3.62003863e-01
1.13012254e+00 2.14211777e-01 7.67879844e-01 9.84698117e-01
5.41290641e-01 6.07263863e-01 5.41145742e-01 1.66118830e-01
2.34557196e-01 -5.06602347e-01 2.29258537e-01 1.10664999e+00
4.76199239e-01 -8.99498537e-02 -4.59796906e-01 -4.22066003e-01
6.42883897e-01 5.63205838e-01 -5.55106163e-01 -1.45220399e-01
-8.47652793e-01 1.06257558e+00 1.09380805e+00 2.36376837e-01
-5.75832844e-01 3.57242674e-01 3.15615833e-01 9.04426165e-03
3.31232309e-01 7.38197803e-01 -4.95427608e-01 3.07949245e-01
-2.66842186e-01 -1.72988400e-01 6.14698529e-01 5.21140933e-01
6.84247971e-01 1.57512024e-01 -3.37201767e-02 3.99356902e-01
3.98512363e-01 -7.78809935e-02 2.27555603e-01 -1.52537286e-01
1.16991468e-01 9.76192594e-01 -4.07221049e-01 -1.08242285e+00
-7.54811466e-01 -7.35550880e-01 -1.06837094e+00 -2.68099725e-01
-4.80396189e-02 9.00175571e-02 -8.09376955e-01 1.49867260e+00
5.56021154e-01 2.83495992e-01 -1.68466702e-01 7.24507093e-01
1.29007542e+00 7.58713365e-01 5.42727590e-01 -2.38949105e-01
1.37231040e+00 -8.01654279e-01 -5.55342019e-01 2.90912360e-01
8.16655934e-01 -3.40004474e-01 7.67246783e-01 3.11220944e-01
-6.31737232e-01 -2.55956411e-01 -1.23463356e+00 -1.43887848e-01
-5.54760575e-01 -1.50748655e-01 1.26551688e+00 3.69379222e-01
-5.56810081e-01 1.03387618e+00 -8.18488002e-01 1.50727898e-01
9.70996678e-01 9.32417154e-01 -5.25020838e-01 -2.91092932e-01
-1.42164123e+00 5.54598987e-01 5.32923698e-01 3.37372690e-01
-1.00309873e+00 -7.46085465e-01 -8.77712965e-01 1.41693577e-01
3.20528597e-01 -8.95705104e-01 7.22240806e-01 -7.19728410e-01
-1.46002829e+00 1.86472356e-01 -6.56299070e-02 -2.94315487e-01
-1.69052377e-01 2.36274689e-01 -3.86140823e-01 1.29225791e-01
-2.24909902e-01 2.41400123e-01 3.99284720e-01 -8.16223383e-01
-1.18303098e-01 -6.70676351e-01 1.20483581e-02 7.44556636e-02
-1.92915544e-01 -4.01994467e-01 -1.03922963e-01 -4.49452013e-01
-1.41787723e-01 -5.84175050e-01 -4.54626113e-01 -2.59471953e-01
-8.31032217e-01 -3.93259346e-01 2.46780649e-01 -3.97107035e-01
1.32699776e+00 -1.68366480e+00 3.72516483e-01 4.03130233e-01
8.96230042e-01 4.53548223e-01 -3.80727559e-01 8.81464779e-01
-3.99999082e-01 2.87822604e-01 -3.39388219e-03 4.35781568e-01
-3.57972234e-01 -2.61208981e-01 8.82432908e-02 7.91595101e-01
1.70404971e-01 1.28439248e+00 -1.06358552e+00 5.17980382e-02
1.17694594e-01 8.04229558e-01 -5.68177819e-01 1.66007772e-01
-5.42375922e-01 5.74118793e-01 -1.09660947e+00 9.07405078e-01
5.18615842e-01 -8.40344906e-01 6.54887080e-01 -4.51339722e-01
2.65760422e-01 5.05241632e-01 -3.57005775e-01 1.44781911e+00
5.59096672e-02 1.92836747e-01 -5.71215510e-01 -1.01676965e+00
6.35965109e-01 3.60718548e-01 5.04488826e-01 -7.46581674e-01
3.19743782e-01 1.11461714e-01 4.46418107e-01 -3.63778919e-01
-1.52132899e-01 -9.88192409e-02 3.87880325e-01 7.99714178e-02
7.69411325e-02 4.81873721e-01 3.26765589e-02 1.31270185e-01
1.16577590e+00 -1.17806710e-01 6.51609004e-01 -5.43840863e-02
5.33168375e-01 -9.80928540e-02 4.68016148e-01 3.29455972e-01
1.64363161e-01 8.35959539e-02 7.11647391e-01 -7.11313486e-01
-6.95204318e-01 -4.43851858e-01 -8.86931792e-02 1.16261840e+00
7.10650161e-02 -7.16900706e-01 -4.74253178e-01 -7.17705667e-01
2.63752621e-02 2.56843597e-01 -9.42968965e-01 -5.11971831e-01
-1.84957221e-01 -1.06710207e+00 3.66656274e-01 2.13631392e-01
1.25819609e-01 -8.93124819e-01 3.44162405e-01 5.18244565e-01
2.91695207e-01 -6.08971715e-01 -3.96364331e-01 5.07970929e-01
-7.22696066e-01 -1.54947472e+00 -3.59154791e-01 -8.49549592e-01
6.98302567e-01 3.05238366e-01 7.93145359e-01 3.26328367e-01
-3.01582605e-01 -4.05407459e-01 -3.05399448e-01 -3.42809707e-01
-1.87074915e-01 3.20870131e-01 -2.50839859e-01 -5.14847226e-02
2.92051733e-01 -8.15828860e-01 -1.10075986e+00 9.01279822e-02
-7.17783213e-01 7.50602186e-02 7.08493114e-01 8.40939045e-01
9.04506445e-01 -4.42836870e-04 8.04264784e-01 -1.37915349e+00
7.71993876e-01 -7.99137712e-01 -5.55692196e-01 2.47597456e-01
-5.59983015e-01 3.08778465e-01 1.03837359e+00 -2.79626191e-01
-3.16209465e-01 1.66789234e-01 -5.04424393e-01 -1.96751460e-01
9.23599154e-02 9.74681258e-01 -5.73556244e-01 -4.21646953e-01
3.85259837e-01 3.38423282e-01 -1.99326307e-01 -4.22498643e-01
2.89754719e-01 2.73760796e-01 -1.20084673e-01 -3.80592436e-01
3.94205153e-01 8.58902857e-02 5.23256660e-01 -6.14531636e-01
-6.75058961e-01 -3.89038444e-01 -2.33238533e-01 1.64293215e-01
7.84363806e-01 -8.64928484e-01 -1.38172615e+00 1.75335869e-01
-1.12386560e+00 -2.74475515e-01 3.59256238e-01 3.37750137e-01
-1.72952503e-01 4.20922279e-01 -6.89205408e-01 -2.57749259e-01
-7.56745040e-01 -1.32991159e+00 7.77470827e-01 1.06762655e-01
9.07180533e-02 -1.13873267e+00 1.70175597e-01 3.93893331e-01
8.62809494e-02 5.29780984e-01 1.59148180e+00 -1.06318223e+00
-9.71543908e-01 -3.03825498e-01 -2.65632510e-01 -1.45120770e-01
4.26049650e-01 -2.11631432e-01 -7.39798963e-01 -2.41091341e-01
-5.32807767e-01 -2.39447743e-01 9.25035715e-01 5.34114718e-01
1.51649833e+00 -7.33068705e-01 -4.89972264e-01 9.64831173e-01
1.34861135e+00 5.57438791e-01 6.03740513e-01 5.21318130e-02
1.33764720e+00 1.21424034e-01 5.62465861e-02 4.59521204e-01
3.19712520e-01 4.79274929e-01 8.13176095e-01 -2.82091945e-01
6.48875758e-02 -5.33866882e-01 5.08765243e-02 6.17129028e-01
-3.07647884e-01 -5.82216978e-01 -6.76182449e-01 -9.34784412e-02
-1.81639290e+00 -7.85786271e-01 -3.16028446e-01 1.99163032e+00
9.97313559e-01 -3.45394276e-02 7.73983374e-02 -3.00750524e-01
5.78099191e-01 1.51035681e-01 -9.77836847e-01 -2.27600411e-01
-2.10129982e-03 4.15603936e-01 4.04291600e-01 4.86219287e-01
-9.71519172e-01 1.04981089e+00 5.12568378e+00 1.11473072e+00
-1.27738714e+00 -2.09025115e-01 1.04498732e+00 2.83542633e-01
-5.19231677e-01 -1.28089398e-01 -6.18430793e-01 4.16657865e-01
9.02267694e-01 -7.40264580e-02 4.05787855e-01 6.98314488e-01
4.02713478e-01 4.95331913e-01 -1.15894246e+00 7.50285447e-01
-2.65372455e-01 -2.07126474e+00 6.29232347e-01 2.72645026e-01
5.64674556e-01 9.85552967e-02 7.21076652e-02 -1.10736392e-01
1.56316042e-01 -1.45686078e+00 -8.01741108e-02 3.97037715e-01
7.82072902e-01 -9.30759490e-01 5.88873148e-01 -1.09497011e-01
-1.31264234e+00 2.17329502e-01 -5.47554195e-01 1.67008899e-02
-2.98727453e-01 4.38161105e-01 -1.06523609e+00 8.09645653e-01
-4.82308529e-02 1.14438534e+00 -5.19369960e-01 1.10550439e+00
-2.96034217e-01 5.95839918e-01 1.97037950e-01 -3.64390671e-01
4.29044038e-01 -3.08580428e-01 1.06486715e-01 9.19380128e-01
-1.03015073e-01 3.32114518e-01 1.86282113e-01 8.25878918e-01
-5.73172629e-01 3.27956051e-01 -5.58894038e-01 -6.15594923e-01
3.06313902e-01 1.20100725e+00 -6.70240760e-01 -4.59311269e-02
-2.64009535e-01 7.50655293e-01 4.45975453e-01 5.29267311e-01
-8.98110926e-01 -5.63056469e-01 6.37719393e-01 2.40739986e-01
2.70415574e-01 4.51542996e-02 6.77340105e-02 -9.13937211e-01
-4.87804383e-01 -9.40552354e-01 4.43954170e-01 -4.53527302e-01
-1.22601950e+00 6.56695485e-01 -5.74959338e-01 -1.11720526e+00
3.93227398e-01 -7.33100295e-01 -6.71246171e-01 6.85762823e-01
-1.58593428e+00 -1.20727265e+00 -1.34760246e-01 6.14165306e-01
3.25890422e-01 -1.14443518e-01 9.03524995e-01 2.63144881e-01
-1.02574050e+00 5.17681301e-01 3.62387002e-01 -4.07646969e-02
2.57073820e-01 -9.20344830e-01 3.38724881e-01 2.57915229e-01
8.00728798e-02 9.11521971e-01 4.76597220e-01 -7.00379491e-01
-1.87984645e+00 -1.41770661e+00 4.10342395e-01 -2.27291375e-01
8.30660403e-01 -3.73047113e-01 -1.01370049e+00 4.92829293e-01
-7.81236216e-02 7.58584514e-02 1.28312945e+00 5.37370667e-02
-4.14007872e-01 7.83744827e-02 -6.53110802e-01 4.10543740e-01
9.44806039e-01 -5.01203179e-01 1.84377939e-01 9.17868197e-01
1.00678802e+00 -2.23423794e-01 -1.11892271e+00 4.14894253e-01
4.70092416e-01 -6.75743520e-01 1.04442978e+00 -1.14957464e+00
5.22514820e-01 -2.83008963e-01 1.22548558e-01 -1.11275935e+00
-7.09363222e-01 -6.72013402e-01 -3.44855189e-01 4.66292471e-01
6.80990875e-01 -8.26219261e-01 8.29439700e-01 1.84832156e-01
-2.05620244e-01 -1.48989141e+00 -7.32280552e-01 -3.98585081e-01
-1.26452833e-01 6.12818748e-02 6.54837668e-01 1.04769170e+00
1.93385541e-01 9.75442648e-01 -3.34276229e-01 1.30985752e-02
3.17192227e-01 2.18888506e-01 4.26734209e-01 -1.15213537e+00
-4.73556429e-01 -6.09196603e-01 -6.30981743e-01 -9.22135949e-01
1.92120537e-01 -1.26608896e+00 -5.95077693e-01 -1.73102200e+00
5.45142651e-01 -1.71060696e-01 -5.78730464e-01 6.84178174e-01
-4.07027036e-01 -2.00295951e-02 -2.56959438e-01 1.69519305e-01
-8.92728388e-01 7.75468409e-01 1.54220486e+00 -5.39589822e-01
-2.68572360e-01 1.46674022e-01 -1.15050364e+00 2.60111332e-01
8.54815781e-01 -4.26591933e-01 -5.43435931e-01 -1.21977635e-01
4.46328044e-01 5.97301461e-02 1.67339683e-01 -4.66348290e-01
1.94214866e-01 -3.92999858e-01 6.03390455e-01 -1.90506488e-01
2.65712976e-01 -6.57973349e-01 4.98825580e-01 7.32944250e-01
-3.33637327e-01 -1.96280345e-01 1.76544592e-01 9.69390750e-01
-1.08690478e-01 2.12784931e-01 5.34923077e-01 -5.28642125e-02
-4.01266992e-01 1.16953790e+00 -2.45150641e-01 -4.14818734e-01
7.26567686e-01 -2.14528918e-01 -4.38470721e-01 -2.07199648e-01
-5.18161774e-01 6.15181923e-02 1.94833636e-01 8.37232452e-03
8.69791806e-01 -1.13745236e+00 -4.37814891e-01 2.60265395e-02
3.11760366e-01 -4.40521464e-02 4.37706888e-01 8.75413060e-01
-6.41649783e-01 6.24444067e-01 4.98639978e-02 -1.82223439e-01
-1.20379198e+00 7.12884963e-01 5.09909689e-01 -3.19144934e-01
-2.98158377e-01 9.85012054e-01 6.57771289e-01 -1.10230848e-01
1.56502411e-01 -1.52297318e-01 -2.73246288e-01 -1.24015346e-01
4.14057344e-01 1.01257600e-01 1.97309822e-01 -3.49078566e-01
-5.41977882e-01 2.43737757e-01 -5.34903824e-01 1.07457495e+00
1.63099217e+00 5.75100124e-01 -3.04294229e-01 -7.41431788e-02
1.42406309e+00 -3.09746444e-01 -9.57990468e-01 1.93142220e-02
-2.26247042e-01 -5.66085540e-02 1.59035102e-01 -7.93615937e-01
-1.01424670e+00 8.40961635e-01 3.13435674e-01 -1.28893584e-01
8.87942016e-01 5.52436598e-02 7.44412124e-01 6.79353416e-01
-1.68798200e-03 -3.74647737e-01 1.18097201e-01 3.00714999e-01
7.92923093e-01 -1.18650973e+00 4.14454579e-01 -4.84958410e-01
-3.31166059e-01 1.21260893e+00 4.62866426e-01 1.09179787e-01
6.18696630e-01 -4.72435623e-01 -5.96032917e-01 -9.26042974e-01
-7.89253473e-01 -7.34128430e-02 5.76616108e-01 4.32178259e-01
8.65792871e-01 2.84307182e-01 -2.62535572e-01 7.93950260e-01
2.85484225e-01 -1.51513278e-01 6.72434717e-02 6.58367813e-01
-5.42668402e-01 -1.28740144e+00 2.83015817e-01 5.60943782e-01
-5.12615025e-01 -7.83051550e-01 -8.51388812e-01 5.91865361e-01
-1.85307443e-01 7.19444513e-01 -6.37484789e-01 -6.54559076e-01
1.25571981e-01 -3.29664022e-01 5.02695680e-01 -6.94849789e-01
-5.94299555e-01 1.42090291e-01 3.51136588e-02 -4.73264426e-01
-2.08411083e-01 8.79621878e-02 -1.44497621e+00 -4.47775602e-01
-5.97251594e-01 4.32594746e-01 3.87091756e-01 8.32410932e-01
8.96766961e-01 8.12729895e-01 8.00789177e-01 -6.79643035e-01
-5.03009856e-02 -6.79229975e-01 -5.57641923e-01 1.31583791e-02
3.27333212e-01 -3.62113893e-01 6.90229889e-03 -2.27437630e-01] | [5.15496301651001, 5.8891496658325195] |
e1290e57-20f3-4255-a3a9-c6e34ae82095 | a-self-supervised-automatic-post-editing-data | 2111.12284 | null | https://arxiv.org/abs/2111.12284v2 | https://arxiv.org/pdf/2111.12284v2.pdf | A Self-Supervised Automatic Post-Editing Data Generation Tool | Data building for automatic post-editing (APE) requires extensive and expert-level human effort, as it contains an elaborate process that involves identifying errors in sentences and providing suitable revisions. Hence, we develop a self-supervised data generation tool, deployable as a web application, that minimizes human supervision and constructs personalized APE data from a parallel corpus for several language pairs with English as the target language. Data-centric APE research can be conducted using this tool, involving many language pairs that have not been studied thus far owing to the lack of suitable data. | ['Heuiseok Lim', 'Seungjun Lee', 'Jaehyung Seo', 'Sugyeong Eo', 'Chanjun Park', 'Hyeonseok Moon'] | 2021-11-24 | null | null | null | null | ['automatic-post-editing', 'automatic-post-editing'] | ['computer-vision', 'natural-language-processing'] | [ 2.54641414e-01 3.81219923e-01 -6.57731481e-03 -4.50182050e-01
-7.76415884e-01 -5.05111098e-01 7.12332904e-01 6.43228590e-01
-6.41430736e-01 9.16405320e-01 3.80465090e-01 -4.76509482e-01
-2.70237643e-02 -4.80194956e-01 -5.08429289e-01 2.64518499e-01
3.45968783e-01 7.28173077e-01 2.98214942e-01 -7.37068415e-01
5.57682276e-01 9.20789037e-03 -1.32058585e+00 2.79278845e-01
1.44908535e+00 1.33712202e-01 4.44312632e-01 6.97865069e-01
-2.43425652e-01 4.13539648e-01 -8.97991717e-01 -9.03269351e-01
9.88829136e-02 -4.99288738e-01 -9.99289930e-01 7.46152131e-03
3.25470477e-01 1.36836484e-01 2.11822763e-01 9.21813726e-01
5.95405757e-01 -1.42120585e-01 3.70069474e-01 -1.06266057e+00
-7.99598336e-01 1.16889405e+00 1.55010402e-01 1.68823242e-01
7.01267898e-01 -7.59298354e-02 9.79894280e-01 -1.17393339e+00
9.57322299e-01 8.46512973e-01 6.80530548e-01 5.87602019e-01
-1.23395264e+00 -3.18877578e-01 -4.14331436e-01 1.29415747e-02
-1.17492533e+00 -6.13015890e-01 8.19900692e-01 -4.71667588e-01
1.25716782e+00 2.82856405e-01 6.74196899e-01 1.36964595e+00
1.86466321e-01 5.35105944e-01 9.42130029e-01 -1.09695613e+00
7.82158598e-02 7.52599955e-01 1.66897491e-01 4.58088607e-01
3.59350652e-01 -4.13461953e-01 -8.57206762e-01 5.60483597e-02
2.71190375e-01 -6.27724469e-01 -2.08717480e-01 -1.59537084e-02
-1.21612406e+00 5.99870563e-01 -4.15228665e-01 6.57920837e-01
-3.61789495e-01 -5.85162520e-01 7.43479788e-01 9.01110828e-01
6.88297808e-01 1.05390835e+00 -8.75957370e-01 -7.39676356e-01
-1.10921764e+00 2.59821892e-01 1.11463714e+00 1.33879423e+00
5.67913592e-01 -1.90596595e-01 -1.40653074e-01 9.86263394e-01
3.23300391e-01 3.45018715e-01 8.40187728e-01 -5.55654407e-01
8.94909441e-01 9.72000957e-01 1.71083495e-01 -9.26229537e-01
-1.55847281e-01 -1.06874183e-02 -6.99804127e-01 -1.85367361e-01
1.20562799e-01 -3.46446991e-01 -1.69127271e-01 1.51912606e+00
7.06329793e-02 -6.72371149e-01 3.81263763e-01 1.40025079e-01
8.08568597e-01 4.19102788e-01 -1.13007791e-01 -4.47460622e-01
1.06394100e+00 -1.05706632e+00 -9.35854137e-01 -3.27977836e-01
9.46070969e-01 -1.16628051e+00 1.60291779e+00 4.91927415e-01
-1.31386018e+00 -5.39773703e-01 -1.13938344e+00 -2.65737861e-01
-5.43784142e-01 3.83506328e-01 4.86784816e-01 6.40684247e-01
-1.09513557e+00 7.90840805e-01 -6.18107915e-01 -7.93816924e-01
-2.18328051e-02 -9.27161202e-02 -6.36989355e-01 2.09766060e-01
-1.40779746e+00 1.27518439e+00 5.53882599e-01 -3.72434221e-02
-2.17243850e-01 -7.85215318e-01 -9.47594702e-01 -4.17347163e-01
3.02931130e-01 -6.29285336e-01 1.25579071e+00 -9.06921148e-01
-1.77294469e+00 8.89899075e-01 -1.05140671e-01 -3.23349446e-01
6.39922678e-01 -4.08970863e-01 -6.21252537e-01 -3.82604092e-01
2.26444378e-01 3.17102611e-01 5.12129664e-01 -8.25316310e-01
-5.26007771e-01 -1.93996891e-01 -1.80391848e-01 1.99715346e-01
-6.05914414e-01 6.78334355e-01 -4.91296649e-01 -1.17848420e+00
-3.53470236e-01 -8.20080400e-01 -1.62957057e-01 -4.92071897e-01
-4.64460433e-01 -4.67582792e-01 2.85525084e-01 -1.25992334e+00
1.85592544e+00 -1.88472974e+00 2.40561381e-01 2.47135848e-01
-2.02294514e-01 5.17127514e-01 -2.30230749e-01 8.92405570e-01
1.16681822e-01 5.09898126e-01 -3.76677275e-01 -6.82088375e-01
-3.44548002e-02 2.31253114e-02 -9.06240940e-03 -1.75031558e-01
2.04015538e-01 5.79354525e-01 -1.04914546e+00 -5.44136822e-01
-1.30122647e-01 -1.16777852e-01 -6.63959920e-01 4.82307017e-01
-3.37578356e-01 2.08057389e-01 3.16711958e-03 4.61844832e-01
3.19166452e-01 2.81881273e-01 1.88170820e-01 1.36957020e-01
-5.21930277e-01 6.58045650e-01 -1.10414767e+00 1.98492885e+00
-6.26317143e-01 8.50041986e-01 -2.73367971e-01 -5.14966369e-01
1.28506553e+00 4.21054393e-01 6.47927076e-02 -4.90700960e-01
-8.45942944e-02 4.44379032e-01 -6.18235506e-02 -7.72564530e-01
1.10160041e+00 2.64387339e-01 -2.15304792e-01 7.45563626e-01
2.10723490e-01 -4.40818012e-01 9.38669026e-01 3.52167070e-01
1.14063787e+00 5.33939153e-02 4.54012781e-01 -1.89572841e-01
6.76271617e-01 2.93666393e-01 6.06782258e-01 6.52719855e-01
2.10740387e-01 2.73830324e-01 2.61449873e-01 3.90338153e-02
-1.41875279e+00 -6.16060853e-01 -7.05197826e-02 8.84626329e-01
-5.35963893e-01 -1.03204966e+00 -8.68900478e-01 -5.78099430e-01
-4.35787529e-01 1.04426205e+00 -2.53387809e-01 -2.88186804e-03
-3.86509717e-01 -3.47437918e-01 5.33255398e-01 1.81812853e-01
1.50062412e-01 -1.38532567e+00 -1.16447590e-01 5.27434051e-01
-3.87757212e-01 -9.11831677e-01 -4.79486287e-01 -1.94919318e-01
-6.53980136e-01 -9.54376996e-01 -2.63592392e-01 -8.26487839e-01
7.50478804e-01 -2.63411880e-01 1.28540170e+00 2.03762919e-01
-3.08655668e-02 9.64855328e-02 -8.17215502e-01 -8.13848615e-01
-1.18350625e+00 3.94094110e-01 1.81118757e-01 -3.90753865e-01
5.31097174e-01 -4.99647141e-01 2.09723860e-01 3.39416154e-02
-7.12087810e-01 2.82402486e-01 6.00626409e-01 7.86631644e-01
2.68666118e-01 -5.73330745e-02 6.72072530e-01 -1.28951252e+00
1.18082535e+00 -4.11542267e-01 -4.82871056e-01 5.82057416e-01
-8.48404944e-01 -1.00108460e-01 7.25368381e-01 -1.72723204e-01
-1.23455536e+00 -3.49577129e-01 -2.11200058e-01 4.24869746e-01
-1.90709934e-01 9.29251969e-01 -1.88019291e-01 3.49953860e-01
8.36654484e-01 1.45274997e-01 1.10708764e-02 -8.87042224e-01
3.30021620e-01 1.29397309e+00 5.16421378e-01 -5.72509646e-01
7.66164422e-01 -4.15794641e-01 -6.57710135e-01 -8.33504140e-01
-6.07373059e-01 -9.86099169e-02 -9.23745632e-01 -1.02908149e-01
4.74540472e-01 -7.92720854e-01 1.28763333e-01 6.06217504e-01
-1.39134002e+00 -2.91705459e-01 -3.50518197e-01 4.30739820e-01
-3.15060854e-01 4.09466505e-01 -4.07621354e-01 -2.92416871e-01
-7.35540330e-01 -8.30108166e-01 5.54781139e-01 5.05502671e-02
-1.02845502e+00 -1.00285065e+00 5.42127073e-01 5.25082052e-01
4.39639628e-01 -1.43202975e-01 9.33097959e-01 -8.95065725e-01
-1.44419342e-01 -2.47872233e-01 2.54289538e-01 6.26131773e-01
3.14877421e-01 8.12803447e-01 -2.28564233e-01 -2.03206092e-01
-3.32731277e-01 -3.93616974e-01 -1.19329877e-01 -2.32111335e-01
8.86208117e-01 -5.37811399e-01 6.53389394e-02 1.89897403e-01
9.42934632e-01 -8.53778273e-02 3.69822145e-01 5.42999625e-01
4.87741232e-01 7.74408758e-01 8.39850962e-01 6.82971776e-01
7.18126893e-01 7.30375111e-01 -3.80384058e-01 4.56085615e-02
-8.69949907e-02 -4.36260015e-01 5.15725195e-01 1.46977711e+00
7.75353685e-02 -3.05917531e-01 -1.03718305e+00 7.69980192e-01
-1.74762416e+00 -7.92530298e-01 -5.09688020e-01 2.02914619e+00
1.46771348e+00 1.36572391e-01 -1.06069781e-01 2.60045588e-01
5.21992028e-01 -3.01720321e-01 -1.03278011e-01 -8.29436839e-01
-2.45391905e-01 3.03508103e-01 -3.19191702e-02 4.77141410e-01
-7.13615954e-01 9.97534811e-01 6.59703207e+00 5.31683207e-01
-7.99185514e-01 6.75024763e-02 3.57476100e-02 2.19799370e-01
-7.07192183e-01 1.96954697e-01 -8.21347058e-01 7.10328221e-01
1.18494332e+00 -7.11777210e-01 3.99005294e-01 8.31736684e-01
5.84581435e-01 -6.02413155e-02 -1.32270277e+00 7.34756589e-01
2.37857178e-01 -1.22436154e+00 9.34268087e-02 -2.62843043e-01
6.08925521e-01 8.04665983e-02 -4.91838545e-01 2.69912541e-01
3.72614920e-01 -4.30349350e-01 9.11881566e-01 7.11570859e-01
6.05014384e-01 -5.43725967e-01 7.14040160e-01 5.85228145e-01
-4.93501872e-01 2.14400724e-01 -3.61335635e-01 -1.27075449e-01
2.79013902e-01 8.22666824e-01 -9.22749460e-01 5.21341085e-01
7.18509436e-01 8.70174527e-01 -1.16941023e+00 6.89588368e-01
-6.55165315e-01 6.72217250e-01 -1.15095526e-01 -1.66862220e-01
-4.02411193e-01 -4.14248049e-01 7.25344181e-01 1.42456424e+00
5.73571265e-01 -3.53758633e-01 1.65866464e-02 6.68702424e-01
-1.20882437e-01 7.61252165e-01 -6.95307493e-01 -2.73314893e-01
9.60063994e-01 1.13955641e+00 -1.90760940e-01 -3.28537285e-01
-3.87489349e-01 1.09435248e+00 6.30366147e-01 -1.47442088e-01
-2.45854318e-01 -7.59079099e-01 2.90716171e-01 1.32499084e-01
-1.13991916e-01 -2.19668567e-01 -4.09491479e-01 -1.22386241e+00
3.06996763e-01 -1.38915503e+00 2.40760520e-01 -7.08118558e-01
-1.35930729e+00 7.41421819e-01 -1.02106638e-01 -1.05313468e+00
-4.57183659e-01 -2.76627034e-01 -6.92763090e-01 8.32324088e-01
-8.42188716e-01 -9.70573187e-01 -1.21759191e-01 3.95583242e-01
7.06786990e-01 -5.83332837e-01 9.57566142e-01 4.35866773e-01
-1.03505528e+00 8.94177556e-01 -1.54961227e-02 5.40589467e-02
1.07422674e+00 -1.31775188e+00 5.26332080e-01 1.32300639e+00
2.69183125e-02 9.93280351e-01 8.11450422e-01 -9.83126342e-01
-1.09277451e+00 -1.06087947e+00 2.11760020e+00 -6.56146467e-01
1.00820243e+00 -4.76701617e-01 -8.67431045e-01 6.43081844e-01
6.65937841e-01 -5.84587693e-01 9.67919469e-01 2.94379175e-01
2.00921059e-01 -5.85119352e-02 -8.57946217e-01 7.72132277e-01
1.00039923e+00 -5.70261717e-01 -1.16404748e+00 4.79616553e-01
6.72019064e-01 -3.59245300e-01 -1.16397142e+00 1.41742840e-01
-7.79465139e-02 -4.43511277e-01 3.10291618e-01 -5.56593776e-01
6.42780244e-01 -2.46498048e-01 2.76989579e-01 -1.81205595e+00
-1.50874197e-01 -1.12827063e+00 2.02223450e-01 2.06218576e+00
1.07295859e+00 -3.97489488e-01 1.40064970e-01 8.36922944e-01
-3.62256736e-01 -3.03206861e-01 -6.38632357e-01 -7.66520858e-01
-1.68992892e-01 -5.82041025e-01 6.05249763e-01 1.03108990e+00
6.17111087e-01 6.32783711e-01 -4.30313170e-01 -2.11482748e-01
6.70585409e-02 -3.38604540e-01 1.08675492e+00 -1.13580716e+00
-1.84282973e-01 -4.27947611e-01 3.34463678e-02 -3.83811414e-01
1.92538708e-01 -1.04607856e+00 1.59523711e-01 -1.41521084e+00
2.70091221e-02 -5.34650207e-01 3.59560460e-01 5.67556858e-01
-3.43292117e-01 -1.21246569e-01 1.83472000e-02 1.72511861e-01
-3.77422988e-01 5.78868091e-01 7.01341510e-01 8.84737670e-02
-4.81116474e-01 -1.48173526e-01 -8.07321548e-01 6.20158732e-01
7.57539630e-01 -7.17817247e-01 -3.70027006e-01 -5.25095344e-01
7.67451882e-01 -4.13688272e-01 -3.61065000e-01 -9.15999830e-01
2.21240506e-01 -9.42222327e-02 -1.16067067e-01 -4.81268138e-01
-3.85532975e-01 -5.83392203e-01 3.26684892e-01 1.46778405e-01
-5.25988519e-01 7.40604401e-01 1.18448012e-01 1.07926538e-03
-4.23377216e-01 -6.89499378e-01 4.23456758e-01 -3.69534492e-02
-6.16496921e-01 1.77455377e-02 -5.83621085e-01 3.01414222e-01
1.02552438e+00 -9.10615847e-02 -2.10769191e-01 -3.79896984e-02
-3.35579515e-01 9.70957205e-02 6.58855677e-01 6.55589044e-01
2.34510452e-01 -1.14450407e+00 -1.08269286e+00 2.90944070e-01
4.76408213e-01 -9.87313688e-02 -8.07775334e-02 5.48000038e-01
-6.66912198e-01 2.20925376e-01 -3.05657476e-01 -1.14328496e-01
-1.45545805e+00 2.88871318e-01 -1.61615595e-01 -4.21645731e-01
-5.93414962e-01 6.63574338e-01 -1.12648356e+00 -9.35184896e-01
-2.85100527e-02 2.77515985e-02 -5.13856769e-01 2.26864755e-01
5.57540894e-01 5.50517738e-01 4.57132667e-01 -4.81711626e-01
-1.04939677e-01 2.86533963e-02 -2.80655324e-01 -2.06599832e-01
1.46776938e+00 -4.80586290e-01 -5.52366734e-01 5.46624660e-01
6.90776706e-01 4.28754061e-01 -6.03355289e-01 -2.93480128e-01
4.83030856e-01 -4.70252991e-01 -2.35952362e-01 -9.12881136e-01
-3.52126151e-01 4.78959590e-01 3.19835283e-02 9.01686922e-02
7.91404963e-01 -8.70472565e-02 5.66989243e-01 6.50761545e-01
2.27267399e-01 -1.88831985e+00 -1.19789891e-01 9.53073025e-01
1.32781947e+00 -1.29489744e+00 -1.36413917e-01 -3.01785558e-01
-6.31140232e-01 1.01488960e+00 8.36252570e-01 4.71146345e-01
5.45210123e-01 1.83215111e-01 1.97278053e-01 -2.83304527e-02
-9.63746369e-01 2.09760875e-01 2.95198768e-01 6.18213236e-01
8.73710513e-01 -8.93576629e-03 -1.12376535e+00 8.65285516e-01
-7.94527948e-01 2.07519338e-01 8.71563554e-01 1.08651662e+00
-2.49146558e-02 -1.86020315e+00 -3.84062156e-02 4.60253835e-01
-2.70464480e-01 -4.30409431e-01 -7.14857280e-01 6.52057648e-01
1.16736002e-01 1.06271052e+00 9.94417220e-02 -3.07052255e-01
6.83372021e-01 1.42539844e-01 2.49562293e-01 -1.07282054e+00
-8.41023564e-01 -4.46479082e-01 7.05473959e-01 -1.97656423e-01
-3.52135420e-01 -9.45035875e-01 -8.60566199e-01 -4.59636062e-01
-1.77167252e-01 2.91029930e-01 8.97053361e-01 1.04249465e+00
5.80830812e-01 2.31229186e-01 7.10027933e-01 -5.67709744e-01
-4.20041740e-01 -1.50125611e+00 -1.94140241e-01 7.08109260e-01
-3.60011458e-01 -7.69788474e-02 -6.91118836e-02 2.28303134e-01] | [11.227923393249512, 10.35571575164795] |
9f066268-2d7f-41ee-be35-718fe6b257b5 | machine-learning-with-guarantees-using | 1609.02664 | null | http://arxiv.org/abs/1609.02664v1 | http://arxiv.org/pdf/1609.02664v1.pdf | Machine Learning with Guarantees using Descriptive Complexity and SMT Solvers | Machine learning is a thriving part of computer science. There are many
efficient approaches to machine learning that do not provide strong theoretical
guarantees, and a beautiful general learning theory. Unfortunately, machine
learning approaches that give strong theoretical guarantees have not been
efficient enough to be applicable. In this paper we introduce a logical
approach to machine learning. Models are represented by tuples of logical
formulas and inputs and outputs are logical structures. We present our
framework together with several applications where we evaluate it using SAT and
SMT solvers. We argue that this approach to machine learning is particularly
suited to bridge the gap between efficiency and theoretical soundness. We
exploit results from descriptive complexity theory to prove strong theoretical
guarantees for our approach. To show its applicability, we present experimental
results including learning complexity-theoretic reductions rules for board
games. We also explain how neural networks fit into our framework, although the
current implementation does not scale to provide guarantees for real-world
neural networks. | ['Łukasz Kaiser', 'Charles Jordan'] | 2016-09-09 | null | null | null | null | ['board-games'] | ['playing-games'] | [ 1.88294113e-01 8.18030119e-01 -4.66429085e-01 -5.04050374e-01
-6.55420959e-01 -4.68458474e-01 2.93717414e-01 -1.90762721e-03
-2.03055173e-01 9.15582776e-01 -4.14532304e-01 -9.26771283e-01
-5.44729948e-01 -1.19476652e+00 -1.00270391e+00 -4.22052592e-01
-2.80635923e-01 7.29446113e-01 4.80866313e-01 -1.78552002e-01
-1.09608099e-01 3.59799474e-01 -1.66022944e+00 5.97498417e-01
3.30629259e-01 9.97773826e-01 -4.25123513e-01 7.14644074e-01
-2.42117122e-01 1.02212799e+00 -1.19106546e-01 -6.39469802e-01
4.17530417e-01 -4.12409425e-01 -1.37014735e+00 -2.51267374e-01
6.10719383e-01 4.09272611e-02 -2.26495206e-01 1.15499794e+00
-1.60394028e-01 -2.72893459e-01 3.22788835e-01 -1.62566817e+00
-1.35154441e-01 1.17634702e+00 -5.39752916e-02 -5.09827249e-02
1.85194761e-01 -1.68468997e-01 1.63467395e+00 -5.01955673e-02
5.52072644e-01 1.33910358e+00 6.76514447e-01 1.02432096e+00
-1.21916473e+00 -4.26749468e-01 1.70364603e-01 6.24674678e-01
-9.20574546e-01 -4.78639215e-01 5.70685148e-01 3.21030361e-03
1.11381018e+00 7.47208297e-01 8.37360740e-01 3.35576743e-01
-1.34387165e-01 1.14548814e+00 1.16213727e+00 -9.37362134e-01
2.83786714e-01 4.56591308e-01 5.54074764e-01 1.16607237e+00
6.23849332e-01 1.33438215e-01 -3.42975408e-01 8.04665592e-03
7.41721570e-01 -3.30245316e-01 2.00590119e-01 -6.23886049e-01
-7.05291748e-01 1.09926486e+00 8.95452797e-02 3.23070973e-01
4.05591011e-01 5.12666702e-01 4.75482672e-01 7.53525317e-01
-1.27007455e-01 5.05568087e-01 -6.62074029e-01 9.18690264e-02
-7.03934669e-01 3.05627733e-01 1.38112247e+00 9.63667452e-01
6.34611189e-01 -5.36691174e-02 5.35635114e-01 4.22319025e-01
1.21930778e-01 2.43279353e-01 4.46592644e-02 -1.31185389e+00
1.60306588e-01 2.86096424e-01 -1.93603218e-01 -8.50332975e-01
-3.82669628e-01 -2.02433899e-01 -4.29532111e-01 2.62558639e-01
7.22930253e-01 -7.21065551e-02 -2.61326969e-01 1.89293504e+00
5.65355942e-02 1.43771932e-01 2.90055513e-01 3.49212974e-01
4.09459800e-01 5.55874288e-01 -2.98574656e-01 -5.95191658e-01
1.03104854e+00 -8.64597678e-01 -4.76902902e-01 -2.17779100e-01
1.18627763e+00 -1.43311522e-03 8.74227881e-01 9.85410333e-01
-1.49689054e+00 1.23241976e-01 -1.07641339e+00 -2.25688387e-02
-2.61257976e-01 -2.68404841e-01 1.43910074e+00 9.32853043e-01
-9.69318748e-01 5.12257040e-01 -9.15749311e-01 -3.03425372e-01
4.64193553e-01 7.60330975e-01 -4.68506098e-01 -1.12621829e-01
-9.83651817e-01 9.82235730e-01 9.25610006e-01 -1.09565333e-01
-5.22049189e-01 -2.76700586e-01 -9.77920353e-01 1.06858894e-01
8.01517904e-01 -6.34005606e-01 1.69348967e+00 -1.04151666e+00
-1.40411925e+00 9.63200450e-01 -1.70385882e-01 -1.06383562e+00
3.26410830e-01 1.97939143e-01 -3.56757045e-02 8.60638246e-02
-5.78523278e-01 3.42955977e-01 5.57416603e-02 -9.71117020e-01
-9.58229363e-01 -4.29564774e-01 8.47408712e-01 -3.52967322e-01
-4.13009048e-01 8.31887573e-02 -7.78300129e-03 1.90710723e-01
3.22990507e-01 -7.52582192e-01 -2.66676158e-01 1.72082037e-01
-2.83015549e-01 -3.90691161e-01 3.08323145e-01 2.10203856e-01
1.22310770e+00 -1.75521708e+00 -1.24439761e-01 4.47346866e-01
3.35217081e-02 7.87904859e-02 -5.08439578e-02 3.44636172e-01
-2.10647676e-02 3.44155729e-01 1.19934887e-01 4.20151353e-02
5.52441835e-01 8.23219478e-01 -4.80494440e-01 2.89518893e-01
1.29887193e-01 1.01250815e+00 -6.36621118e-01 -6.57508135e-01
1.33010536e-01 -2.38489538e-01 -9.68255937e-01 -9.34566781e-02
-8.02634299e-01 -6.72022283e-01 -3.90171647e-01 4.33432490e-01
4.78048474e-01 -2.98215091e-01 7.38640547e-01 2.59722203e-01
1.71307757e-01 6.95307374e-01 -1.44409597e+00 1.27307808e+00
-3.63939613e-01 5.57251215e-01 2.93260247e-01 -1.59208369e+00
5.35618544e-01 1.70882449e-01 1.72396731e-02 -3.83613318e-01
2.03336552e-01 2.20985129e-01 2.30709881e-01 -7.06165969e-01
-1.66173637e-01 -7.14455545e-01 -7.01991171e-02 4.62792367e-01
3.35162715e-03 -3.38154525e-01 3.69855791e-01 4.25050122e-04
1.15756965e+00 -6.74946681e-02 3.52231920e-01 -2.72551537e-01
6.06001377e-01 4.06119943e-01 7.11011469e-01 9.87391174e-01
1.31518528e-01 6.54550642e-02 1.04737246e+00 -8.56887579e-01
-9.04341221e-01 -8.13815415e-01 -2.09151097e-02 1.11554062e+00
-1.59584895e-01 -7.31641412e-01 -8.14792931e-01 -5.16354382e-01
-3.29626024e-01 6.84657514e-01 -6.04450107e-01 -1.55103669e-01
-3.99676353e-01 -4.79658186e-01 7.38205016e-01 5.48415065e-01
1.98861808e-01 -1.00573587e+00 -7.03343928e-01 -6.80099577e-02
2.11190730e-01 -1.11602938e+00 5.67833722e-01 6.57673538e-01
-1.29035020e+00 -1.37310183e+00 4.21593249e-01 -7.77679145e-01
4.71367836e-01 2.95699313e-02 1.01995742e+00 4.43655163e-01
-1.57231599e-01 2.57309765e-01 -6.65931962e-04 -6.98094726e-01
-5.85842192e-01 1.73802361e-01 3.88637185e-02 -6.74041271e-01
6.57360137e-01 -4.99376088e-01 2.30679259e-01 -1.19365208e-01
-9.84360456e-01 2.77013570e-01 5.07749259e-01 9.59555507e-01
4.36341465e-01 5.31234026e-01 4.56533469e-02 -1.47337770e+00
1.51412398e-01 -1.93435475e-01 -9.30654585e-01 4.77973223e-01
-7.92296052e-01 4.11768675e-01 8.65972579e-01 -1.83648989e-01
-6.50074005e-01 2.19408974e-01 -1.29019007e-01 6.58062547e-02
-3.99325877e-01 8.46835196e-01 -3.61222833e-01 2.20362414e-02
7.21548378e-01 9.97559801e-02 1.44536749e-01 -1.65252507e-01
2.56310254e-01 3.32481772e-01 4.02035207e-01 -1.05000031e+00
9.66119468e-01 4.71855730e-01 4.99234885e-01 -5.68195879e-01
-1.19526625e+00 1.16094472e-02 -3.81014466e-01 2.52503812e-01
2.40593120e-01 -4.39689875e-01 -1.18966627e+00 -4.99664024e-02
-9.34848666e-01 -5.76883674e-01 -4.05599892e-01 3.77193660e-01
-1.17958796e+00 1.40531644e-01 -6.89702988e-01 -1.31592894e+00
4.87859733e-02 -7.02881634e-01 4.01863545e-01 3.67507488e-02
-5.49247749e-02 -9.34139848e-01 3.96151915e-02 4.53226298e-01
4.71184887e-02 2.40637325e-02 1.40374756e+00 -1.03374088e+00
-6.52561247e-01 -2.20459983e-01 -6.03050925e-03 3.74615759e-01
-3.68625909e-01 1.41086280e-01 -1.06792510e+00 1.17877161e-03
-2.09164664e-01 -6.53989851e-01 6.65282130e-01 2.68210918e-01
1.33360291e+00 -6.43941402e-01 -1.42862335e-01 3.96392614e-01
1.59140694e+00 -1.48354039e-01 8.43198121e-01 6.31035089e-01
1.40726805e-01 7.72890687e-01 4.45634454e-01 1.79093897e-01
3.34709585e-01 3.36969644e-01 4.75291520e-01 3.01181942e-01
4.38863009e-01 -1.82136819e-01 2.68896550e-01 5.76486766e-01
-2.29151189e-01 2.46751592e-01 -9.37221169e-01 3.31934929e-01
-1.99743140e+00 -1.40410030e+00 -3.07160199e-01 2.17163348e+00
1.19724846e+00 7.54271686e-01 3.28067243e-01 7.45693982e-01
3.08432668e-01 -4.23455745e-01 -1.61493257e-01 -9.80944991e-01
-8.85167867e-02 5.80605447e-01 2.66580343e-01 7.13963091e-01
-9.77171063e-01 1.11557114e+00 7.43730068e+00 7.88329363e-01
-8.45027328e-01 -3.10834665e-02 2.56996959e-01 -1.53887302e-01
-4.05461669e-01 2.00741321e-01 -6.23003483e-01 -3.21856320e-01
1.14307237e+00 -1.46532461e-01 6.06538475e-01 1.19763052e+00
-2.27539390e-01 -3.36709172e-02 -1.55497050e+00 7.80542791e-01
-1.89626679e-01 -1.52661920e+00 -2.32503191e-01 -3.87034658e-03
4.09798622e-01 -3.33624333e-01 -6.11999817e-02 4.60848302e-01
5.15330493e-01 -1.38703680e+00 6.66008294e-01 1.18759640e-01
4.26112980e-01 -9.74901140e-01 6.94758356e-01 7.45951712e-01
-6.23574793e-01 -4.18787897e-01 -7.24586427e-01 -8.00032914e-01
-5.34955978e-01 2.93905377e-01 -6.56640053e-01 4.11542863e-01
3.93657595e-01 5.36825895e-01 -3.12241256e-01 1.06790411e+00
-3.36920291e-01 7.29771852e-01 -6.98714733e-01 -3.26209545e-01
9.56676900e-02 1.26099825e-01 -2.26429179e-02 1.18700480e+00
-1.11419819e-01 2.21792921e-01 -1.46355659e-01 7.50834763e-01
8.94468799e-02 -1.60942927e-01 -7.13010907e-01 -7.02234581e-02
3.06098282e-01 1.05998385e+00 -6.44048572e-01 -3.70293915e-01
-5.89166403e-01 2.62662411e-01 4.74674135e-01 -1.14669517e-01
-7.06069469e-01 -2.32798442e-01 6.35668933e-01 -6.77899793e-02
1.52415723e-01 -8.23908225e-02 -5.41033506e-01 -1.21453440e+00
1.44276872e-01 -1.13841701e+00 7.87361860e-01 -4.79560047e-01
-1.17421329e+00 2.07992360e-01 2.38960028e-01 -6.49126232e-01
-4.27739501e-01 -1.18442643e+00 -3.47672403e-01 3.19939733e-01
-1.42260587e+00 -1.06432939e+00 1.62122697e-01 4.21187162e-01
1.54766679e-01 6.01573028e-02 1.13258886e+00 -1.51826397e-01
-4.21205252e-01 7.53109276e-01 -2.00642914e-01 1.18443489e-01
9.09241140e-02 -1.42393386e+00 -2.01543607e-02 7.70134032e-01
5.69841325e-01 6.51409566e-01 9.19688642e-01 6.27624393e-02
-1.99428403e+00 -6.89895689e-01 9.87172067e-01 -3.22916120e-01
9.00155008e-01 -2.20248088e-01 -6.59019053e-01 1.07819283e+00
-1.34063274e-01 6.87411427e-02 8.24163675e-01 5.68864167e-01
-8.31102729e-01 -3.37931663e-01 -9.74081755e-01 4.39602345e-01
1.14897573e+00 -6.15074217e-01 -8.93462956e-01 3.94488961e-01
3.98996443e-01 -3.26168388e-01 -5.27715206e-01 3.66174161e-01
8.39625776e-01 -1.16207647e+00 6.72690094e-01 -1.33344257e+00
4.35120016e-01 -1.29413366e-01 -2.78584659e-01 -6.43802941e-01
-1.37133434e-01 -5.61110556e-01 -4.84633923e-01 6.90274060e-01
6.91929519e-01 -7.14967966e-01 1.20158422e+00 1.04220855e+00
1.31278127e-01 -1.04424107e+00 -9.51409519e-01 -1.13341415e+00
7.21337557e-01 -1.15299857e+00 4.45961833e-01 7.40944982e-01
8.33367050e-01 1.92900524e-01 -1.20890059e-01 -1.20390831e-02
5.55222094e-01 4.07387555e-01 7.34045744e-01 -1.45181048e+00
-7.65578687e-01 -8.17993224e-01 -6.13908350e-01 -7.98283935e-01
4.50260192e-01 -1.11431146e+00 -1.08471792e-03 -1.46693325e+00
5.21515608e-01 -5.23294628e-01 -3.30483913e-01 9.61404741e-01
6.08487129e-01 9.47828367e-02 1.95976600e-01 -2.96445340e-01
-8.21187854e-01 -2.24953622e-01 8.62048507e-01 -2.99402267e-01
2.16098145e-01 4.79267016e-02 -8.07899475e-01 9.58072543e-01
9.37168062e-01 -4.24538434e-01 -5.07682323e-01 -3.87414873e-01
8.55416298e-01 -1.50072858e-01 3.61054868e-01 -1.08116543e+00
3.94004911e-01 -7.11336613e-01 -1.21122137e-01 1.28509421e-02
5.52352108e-02 -1.01202428e+00 1.46858588e-01 7.57349730e-01
-8.31532121e-01 -4.16456819e-01 1.17272526e-01 3.61267105e-02
-5.26218377e-02 -6.92822814e-01 7.89943814e-01 -1.37068376e-01
-3.75159055e-01 1.86746329e-01 -3.26386541e-01 2.08738506e-01
1.04218960e+00 -1.27615914e-01 -2.51944363e-01 -3.72325927e-01
-7.15454876e-01 1.27311006e-01 4.19863254e-01 -1.86859295e-01
5.13562083e-01 -9.04251039e-01 -3.24297726e-01 1.63050205e-03
-3.96241955e-02 -9.49916616e-02 -2.50444025e-01 7.36282885e-01
-5.25552213e-01 7.73455918e-01 -2.06494376e-01 -3.53098474e-02
-1.56000078e+00 8.23188841e-01 5.44944108e-01 -2.37318948e-01
-2.57744372e-01 7.44255960e-01 -9.69292074e-02 -3.88700902e-01
6.99568629e-01 -7.05829680e-01 1.01083875e-01 -4.18603361e-01
8.10590625e-01 1.00081973e-01 4.07501981e-02 1.44196242e-01
-3.51848722e-01 1.96656644e-01 6.94461390e-02 -1.81743786e-01
1.55249119e+00 2.70659953e-01 -4.76438016e-01 5.11463344e-01
8.12173486e-01 -2.89488137e-01 -4.59010720e-01 -3.17661673e-01
3.79315466e-01 -6.90641031e-02 -1.35848358e-01 -6.23389006e-01
-9.40658092e-01 1.12860775e+00 5.73439784e-02 6.46438539e-01
1.27082515e+00 2.95440555e-01 6.60573244e-01 1.38351393e+00
8.99494708e-01 -1.23780179e+00 -4.86056030e-01 7.55645275e-01
2.19691306e-01 -9.45891380e-01 -5.09138927e-02 -7.26624489e-01
-9.41107795e-03 1.55575752e+00 5.07226884e-01 -1.99337602e-01
2.33862266e-01 7.27894723e-01 -2.51535207e-01 -3.33370306e-02
-1.50687003e+00 -4.53791559e-01 -1.98249251e-01 5.00217259e-01
2.35845894e-01 3.44812237e-02 -3.20980072e-01 9.00831997e-01
-5.79074860e-01 2.80398995e-01 6.09987974e-01 1.03004372e+00
-7.84742117e-01 -1.43193328e+00 -1.98327437e-01 3.82504970e-01
-7.00640559e-01 -4.04993072e-02 -4.29817587e-01 1.03679502e+00
9.50553268e-02 8.44046414e-01 -1.75790682e-01 -4.92706865e-01
-5.57616241e-02 2.28860095e-01 1.32683682e+00 -7.38227069e-01
-1.61566421e-01 -2.26836652e-01 4.47189897e-01 -4.73026276e-01
-4.33003396e-01 -5.63817143e-01 -1.52578080e+00 -7.00220823e-01
-3.42187852e-01 3.87927741e-01 4.30350870e-01 1.12870681e+00
-3.32561940e-01 -9.58364904e-02 2.97630996e-01 -2.54984200e-02
-8.79267991e-01 -3.51164401e-01 -6.42847538e-01 -2.15439811e-01
1.19844876e-01 -1.86601937e-01 -3.73382479e-01 1.13656737e-01] | [8.757682800292969, 7.010929584503174] |
b8678a73-deb5-4f7e-91f1-44c9adedded7 | superquadric-object-representation-for | 2109.09627 | null | https://arxiv.org/abs/2109.09627v1 | https://arxiv.org/pdf/2109.09627v1.pdf | Superquadric Object Representation for Optimization-based Semantic SLAM | Introducing semantically meaningful objects to visual Simultaneous Localization And Mapping (SLAM) has the potential to improve both the accuracy and reliability of pose estimates, especially in challenging scenarios with significant view-point and appearance changes. However, how semantic objects should be represented for an efficient inclusion in optimization-based SLAM frameworks is still an open question. Superquadrics(SQs) are an efficient and compact object representation, able to represent most common object types to a high degree, and typically retrieved from 3D point-cloud data. However, accurate 3D point-cloud data might not be available in all applications. Recent advancements in machine learning enabled robust object recognition and semantic mask measurements from camera images under many different appearance conditions. We propose a pipeline to leverage such semantic mask measurements to fit SQ parameters to multi-view camera observations using a multi-stage initialization and optimization procedure. We demonstrate the system's ability to retrieve randomly generated SQ parameters from multi-view mask observations in preliminary simulation experiments and evaluate different initialization stages and cost functions. | ['Cesar Cadena', 'Roland Siegwart', 'Juan Nieto', 'Florian Tschopp'] | 2021-09-20 | null | null | null | null | ['semantic-slam'] | ['computer-vision'] | [ 8.33154321e-02 -5.06509542e-01 -2.92705577e-02 -5.48146665e-01
-8.16463292e-01 -8.71401906e-01 6.43153787e-01 6.41928613e-02
-3.39713424e-01 4.08839285e-01 -1.89416677e-01 9.66990367e-02
-8.24762583e-02 -3.30124289e-01 -8.63990843e-01 -3.62191975e-01
2.39356160e-01 1.17422664e+00 3.83637130e-01 -2.47815132e-01
4.05960441e-01 1.13249290e+00 -1.67493963e+00 -4.08794545e-02
5.70748389e-01 9.15445983e-01 6.79237187e-01 5.58448553e-01
-3.89981419e-02 1.55044094e-01 -3.69456977e-01 5.25456928e-02
5.32094121e-01 2.85598040e-01 -4.60887492e-01 4.09859538e-01
9.69698727e-01 -2.20621973e-01 1.77296877e-01 9.72762704e-01
3.62684637e-01 1.26816988e-01 4.52306062e-01 -1.25308228e+00
-5.34387603e-02 -1.93585411e-01 -5.23894072e-01 -7.33610615e-02
8.32621217e-01 1.79590210e-01 7.94609547e-01 -1.10197699e+00
7.22326279e-01 1.16911304e+00 7.55532026e-01 1.31736323e-02
-1.34188604e+00 -4.53740537e-01 5.65437898e-02 1.45724580e-01
-1.67617798e+00 -5.41936040e-01 7.79770553e-01 -4.32763010e-01
1.03359354e+00 4.26746875e-01 8.62147868e-01 8.46058846e-01
1.44101620e-01 3.75610054e-01 1.17737734e+00 -3.67043167e-01
2.58739203e-01 3.12274843e-01 -3.98722529e-01 7.31872261e-01
3.84405255e-01 2.66794041e-02 -8.02494168e-01 -2.36374974e-01
8.64469469e-01 8.11558515e-02 -1.75840676e-01 -1.21449256e+00
-1.50638795e+00 6.88798428e-01 6.34184778e-01 -2.77668506e-01
-3.89553517e-01 3.60043913e-01 -2.12961912e-01 1.91687152e-01
3.94576073e-01 6.23345256e-01 -5.85028112e-01 -2.11027358e-03
-9.06050861e-01 2.12255925e-01 4.04897809e-01 1.31103754e+00
1.11825883e+00 7.46652260e-02 5.06099761e-01 6.40067518e-01
4.16143149e-01 9.11819518e-01 -2.78913509e-03 -1.08803260e+00
3.14087361e-01 7.20199943e-01 4.17453825e-01 -9.66402769e-01
-6.64180934e-01 -4.70466226e-01 -1.66020840e-01 4.05455232e-01
1.14202216e-01 4.98923421e-01 -9.72194135e-01 1.18192852e+00
6.67686343e-01 -6.68615550e-02 -1.62156582e-01 1.15194213e+00
6.10139668e-01 2.02491984e-01 -4.24123853e-01 1.40055239e-01
1.14832389e+00 -5.63756406e-01 -1.91602930e-01 -7.88028538e-01
3.25115383e-01 -1.13980746e+00 8.04574072e-01 2.31393471e-01
-7.87075341e-01 -4.12471801e-01 -1.10690570e+00 -9.49184075e-02
-2.68601686e-01 1.98027238e-01 6.55595243e-01 5.60876608e-01
-9.92853045e-01 1.74941778e-01 -7.81699300e-01 -6.48294389e-01
1.58427954e-01 5.76710224e-01 -7.63003528e-01 -3.21875185e-01
-4.75247800e-01 1.28634918e+00 4.30154324e-01 1.32002503e-01
-9.44630802e-01 -6.10174000e-01 -1.12271202e+00 -3.94098014e-01
5.86240828e-01 -9.43240225e-01 9.10327375e-01 -5.94107449e-01
-1.22645545e+00 1.06695414e+00 -3.13055843e-01 -1.70262769e-01
5.05795658e-01 -2.63303250e-01 -7.88108632e-02 1.82784081e-01
1.46404281e-01 8.14045429e-01 1.03133392e+00 -1.80184662e+00
-5.21088064e-01 -6.47608638e-01 2.79984195e-02 7.42215991e-01
3.87241483e-01 -5.01135886e-02 -5.74899852e-01 -1.13393813e-01
1.02172196e+00 -1.24473226e+00 -3.43249083e-01 3.07099104e-01
-6.67872932e-03 4.89045471e-01 7.86320865e-01 -4.40674543e-01
2.21139282e-01 -1.88667786e+00 2.90495276e-01 2.37346634e-01
-1.39885107e-02 -2.05416739e-01 4.71872762e-02 3.94344062e-01
3.41467977e-01 -4.14338112e-01 1.06107265e-01 -6.03292048e-01
-6.43364862e-02 4.52932268e-01 -5.60410246e-02 8.98434937e-01
-6.26315400e-02 8.68041277e-01 -8.20933044e-01 -3.79307956e-01
8.78253341e-01 3.89847964e-01 -4.55057859e-01 2.02325761e-01
-4.32087183e-01 7.34207392e-01 -1.90935820e-01 1.00996149e+00
8.13263476e-01 -1.24855459e-01 -9.89716873e-02 -4.00516003e-01
-1.45022467e-01 5.57057410e-02 -1.37849069e+00 2.17170119e+00
-5.80321014e-01 6.09936237e-01 3.06808174e-01 -4.60395545e-01
9.58114743e-01 -1.87368333e-01 6.28809214e-01 -5.16500652e-01
7.61562958e-02 3.37653488e-01 -2.75933415e-01 -2.37427965e-01
1.01745892e+00 -1.06626213e-01 -4.95658815e-02 3.69092636e-02
6.44052261e-03 -8.16989005e-01 -3.67050290e-01 -3.64148170e-02
6.79614484e-01 4.54883933e-01 4.91045058e-01 -1.13626800e-01
4.34123367e-01 4.55450058e-01 2.81540632e-01 6.23590469e-01
3.93111035e-02 1.01890492e+00 -3.36717099e-01 -4.99789804e-01
-1.12245595e+00 -1.15994704e+00 -1.47727579e-01 8.03721011e-01
6.52009308e-01 -1.12180129e-01 -2.28652820e-01 -3.25305402e-01
3.72911721e-01 5.09460747e-01 -1.18211858e-01 8.92680883e-02
-3.92511040e-01 -4.47564751e-01 -6.41198829e-02 1.71586767e-01
9.14605632e-02 -6.35541022e-01 -1.02095056e+00 2.29527056e-02
-9.38431248e-02 -1.43376338e+00 -2.24592179e-01 2.75260806e-01
-8.44708085e-01 -1.07520199e+00 -3.35997134e-01 -3.67236257e-01
8.87559354e-01 8.54052365e-01 1.08968508e+00 -1.50862962e-01
-3.44146430e-01 7.91834354e-01 -3.45142126e-01 -4.17514771e-01
-3.90166581e-01 -1.22053608e-01 4.62996244e-01 -2.04303320e-02
1.21382333e-01 -5.39058566e-01 -5.02764821e-01 6.52538657e-01
-4.77838248e-01 2.44773880e-01 6.29796326e-01 3.60400259e-01
8.24957073e-01 -4.37432021e-01 -1.61874175e-01 -3.39856088e-01
-1.37584075e-01 -1.68568820e-01 -1.11078715e+00 1.00620195e-01
-4.03445095e-01 9.62725803e-02 1.58724681e-01 -2.38492027e-01
-7.07613707e-01 6.02587700e-01 1.32418528e-01 -9.15767550e-01
-2.47555256e-01 2.69035548e-01 -2.22943172e-01 -7.79802978e-01
6.26535833e-01 2.29459479e-01 1.71682104e-01 -5.67489207e-01
5.77050924e-01 5.51240921e-01 5.16787708e-01 -5.78065038e-01
9.20976400e-01 1.04799688e+00 2.26545140e-01 -8.73907506e-01
-7.69235194e-01 -1.02640021e+00 -1.04522729e+00 -3.93637091e-01
5.87548077e-01 -1.19423711e+00 -3.90006065e-01 1.67670071e-01
-1.09779084e+00 6.37629628e-03 -3.82590219e-02 6.48207963e-01
-7.02838540e-01 3.85358304e-01 3.12158465e-01 -6.98802769e-01
-4.48020920e-02 -1.41857851e+00 1.79111278e+00 7.29190558e-03
-8.58391747e-02 -7.71598935e-01 -9.67593789e-02 6.59763455e-01
1.15358815e-01 2.49197647e-01 2.99953640e-01 -2.56748915e-01
-1.08810496e+00 -3.52756202e-01 -6.16873726e-02 -9.62961018e-02
1.11988142e-01 -2.80205071e-01 -9.39008772e-01 -6.55242205e-01
-1.08576259e-02 -1.04926370e-01 3.86595219e-01 2.63933986e-01
5.50571740e-01 2.58637965e-01 -5.28026938e-01 1.08909237e+00
1.70393920e+00 -9.87243131e-02 1.38159364e-01 5.81968367e-01
9.13035870e-01 4.43047494e-01 8.97741556e-01 3.43794912e-01
4.02888238e-01 1.35320199e+00 9.77205992e-01 1.25283003e-01
-1.05611734e-01 -2.13565469e-01 2.32938424e-01 6.39541268e-01
2.01748535e-01 1.54025137e-01 -1.10113633e+00 4.32169497e-01
-1.60548067e+00 -5.27150273e-01 -2.59932935e-01 2.45211983e+00
1.64052457e-01 -3.49107645e-02 -3.07378709e-01 -3.69072139e-01
5.25177360e-01 1.59215182e-01 -5.47988832e-01 -7.56220706e-03
-8.08176994e-02 -2.12971121e-01 9.76406872e-01 6.45056307e-01
-9.97956693e-01 1.10722363e+00 6.09082031e+00 3.49015385e-01
-1.10178983e+00 1.81829348e-01 -3.29449296e-01 -1.83335513e-01
-2.40086868e-01 4.86679405e-01 -1.01655757e+00 9.27768126e-02
5.77589035e-01 2.39405051e-01 4.88038868e-01 9.84343290e-01
1.49072468e-01 -5.33323646e-01 -1.07965422e+00 1.53844166e+00
4.61917996e-01 -1.31390822e+00 -7.24169537e-02 1.50153071e-01
7.90828049e-01 6.42442763e-01 -4.07443754e-02 -2.50626951e-01
1.60366416e-01 -8.00354421e-01 1.06549168e+00 2.73842573e-01
7.65537798e-01 -4.13004935e-01 5.76211214e-01 6.12619698e-01
-1.27103364e+00 -1.85020808e-02 -5.68531930e-01 5.84950708e-02
3.72066855e-01 2.99240679e-01 -1.45343029e+00 8.40502441e-01
6.52553082e-01 5.73473275e-01 -9.90255833e-01 1.31283092e+00
1.01683952e-01 -1.73306957e-01 -8.25098515e-01 1.63678259e-01
2.07697347e-01 -4.02564079e-01 8.71980429e-01 8.01439881e-01
3.80659491e-01 -3.25553060e-01 4.21732634e-01 7.42725909e-01
3.59309733e-01 -1.23563841e-01 -6.41842067e-01 6.71318650e-01
6.19303405e-01 1.34300828e+00 -9.53091264e-01 1.20237479e-02
-3.31573218e-01 9.64325666e-01 3.22844833e-01 9.38483402e-02
-6.86038077e-01 3.91438305e-01 7.59164572e-01 2.65828311e-01
2.28895053e-01 -8.41098249e-01 -1.72121465e-01 -1.25496900e+00
2.44203299e-01 -7.52813220e-01 9.07334164e-02 -1.33551228e+00
-7.35405028e-01 3.93146127e-01 3.12462300e-01 -1.56730223e+00
-2.05332786e-01 -5.98815799e-01 1.48394883e-01 8.46263826e-01
-1.37573004e+00 -1.71021914e+00 -7.32494295e-01 3.84025961e-01
6.15394115e-01 -5.88554963e-02 6.69583559e-01 -9.49976817e-02
1.60575405e-01 -1.59760848e-01 3.16456407e-01 -4.42314982e-01
6.54130638e-01 -1.20373142e+00 3.53962541e-01 8.07518184e-01
8.23537409e-01 5.47715127e-01 8.99896622e-01 -6.96223915e-01
-1.96674573e+00 -9.41998482e-01 1.97410107e-01 -1.15209723e+00
2.22506329e-01 -6.56152904e-01 -5.30396044e-01 7.94108093e-01
-2.67119586e-01 2.48751327e-01 1.40821159e-01 -2.03414261e-02
-2.35988930e-01 -8.42756778e-02 -9.27987576e-01 3.78679186e-01
1.00827599e+00 -6.10952735e-01 -6.10499680e-01 4.66996551e-01
3.91819447e-01 -1.03214383e+00 -7.35503852e-01 5.45179725e-01
6.13141000e-01 -9.99154866e-01 1.39602029e+00 -1.43162653e-01
-4.92887229e-01 -9.13616419e-01 -5.66684127e-01 -1.30069435e+00
-9.07571763e-02 -4.59047735e-01 -6.20664880e-02 7.38082647e-01
-1.59479417e-02 -4.96498734e-01 9.24866915e-01 3.52285355e-01
-2.16920152e-01 -1.79027498e-01 -1.17548883e+00 -9.52063382e-01
-6.26936793e-01 -5.97889245e-01 5.34916222e-01 5.81881583e-01
-7.38113761e-01 1.49255499e-01 -3.32319200e-01 7.69976735e-01
7.76595831e-01 4.40806240e-01 1.38846338e+00 -1.33271217e+00
-6.57280385e-02 -8.20329487e-02 -8.34813893e-01 -1.02787638e+00
1.70972571e-01 -9.30598080e-01 1.60347730e-01 -1.63213313e+00
-6.53943280e-03 -7.67074645e-01 1.10069774e-01 4.59793583e-02
7.79005513e-02 4.10118610e-01 2.43767112e-01 5.63794196e-01
-7.44027853e-01 4.79911953e-01 7.94861913e-01 9.22331586e-02
-1.01622038e-01 -1.31762981e-01 -7.81952217e-02 7.76909292e-01
4.14925545e-01 -4.86466318e-01 -3.25329572e-01 -7.33085573e-01
3.21518302e-01 -5.93091510e-02 6.81801438e-01 -1.18202198e+00
3.03828508e-01 -4.55492854e-01 4.50105071e-01 -1.00950372e+00
1.00783992e+00 -1.32144690e+00 8.13054562e-01 2.32529044e-01
2.06068337e-01 2.24295080e-01 1.24361888e-01 6.61920667e-01
6.53261617e-02 -2.50199974e-01 6.87174380e-01 -4.24284577e-01
-1.18491244e+00 3.00423861e-01 6.73257262e-02 -3.60643268e-01
1.08611917e+00 -6.67856514e-01 1.07883871e-01 -3.62629384e-01
-5.44924200e-01 1.25347584e-01 1.30161238e+00 7.40508080e-01
8.66756141e-01 -1.23867989e+00 -4.83896941e-01 5.46301901e-01
6.07739925e-01 3.89782786e-01 -1.96273066e-02 8.39919686e-01
-1.05937696e+00 4.09401923e-01 -1.43821076e-01 -1.31434166e+00
-1.42351186e+00 5.36374331e-01 3.34186971e-01 4.18493450e-01
-6.44470155e-01 7.43827999e-01 -6.35664761e-02 -7.46940970e-01
-2.54143123e-02 -3.40662897e-01 2.06209645e-01 -1.64772779e-01
1.62187785e-01 1.82862490e-01 3.52487773e-01 -1.24891710e+00
-6.60872459e-01 1.07997859e+00 2.41806656e-01 -2.19326437e-01
1.23154640e+00 -5.54073632e-01 -1.06302626e-01 4.42910314e-01
1.08123314e+00 1.20594114e-01 -1.40339327e+00 -1.84801415e-01
5.89174293e-02 -9.13017869e-01 9.96585339e-02 -4.61844206e-01
-5.68419993e-01 6.80129349e-01 7.09934890e-01 -2.41163045e-01
5.00948966e-01 2.46082723e-01 3.09252858e-01 4.83251840e-01
1.18936133e+00 -8.95140707e-01 1.92834774e-03 2.71979928e-01
1.01075280e+00 -1.45994163e+00 5.17895162e-01 -5.74788809e-01
-4.97796714e-01 1.00542104e+00 4.79305297e-01 7.93703347e-02
2.06031770e-01 1.77695692e-01 2.30937898e-01 -3.52168590e-01
-3.29058647e-01 -2.83735365e-01 4.78935629e-01 7.91585386e-01
-2.96112776e-01 -4.55621304e-03 4.66620713e-01 -4.12603050e-01
-2.89716959e-01 -5.26864350e-01 3.06954533e-01 1.03786731e+00
-5.98856270e-01 -9.05870497e-01 -9.63157237e-01 2.36318439e-01
9.23573449e-02 1.41792566e-01 -2.51799673e-01 6.95331335e-01
7.00984597e-02 6.90757692e-01 -1.71378236e-02 -2.46999279e-01
1.79709315e-01 -2.46192776e-02 9.31268573e-01 -8.47791076e-01
-2.58980751e-01 3.78041230e-02 6.88138185e-04 -8.78656864e-01
-4.83629584e-01 -9.25038993e-01 -8.95952463e-01 6.31396845e-02
-5.34242392e-01 -2.85080165e-01 1.34108067e+00 9.23906684e-01
4.15387243e-01 -2.09959783e-02 4.54156309e-01 -1.45003378e+00
-5.16004860e-01 -5.85452914e-01 -3.16921979e-01 3.33992332e-01
4.45948869e-01 -1.05500805e+00 -2.83594042e-01 -1.36360023e-02] | [7.378938674926758, -2.4020793437957764] |
763f258d-6706-4e68-be89-6064ff204f4b | a-hybrid-learning-rule-for-efficient-and | 1907.01167 | null | https://arxiv.org/abs/1907.01167v3 | https://arxiv.org/pdf/1907.01167v3.pdf | A Tandem Learning Rule for Effective Training and Rapid Inference of Deep Spiking Neural Networks | Spiking neural networks (SNNs) represent the most prominent biologically inspired computing model for neuromorphic computing (NC) architectures. However, due to the non-differentiable nature of spiking neuronal functions, the standard error back-propagation algorithm is not directly applicable to SNNs. In this work, we propose a tandem learning framework, that consists of an SNN and an Artificial Neural Network (ANN) coupled through weight sharing. The ANN is an auxiliary structure that facilitates the error back-propagation for the training of the SNN at the spike-train level. To this end, we consider the spike count as the discrete neural representation in the SNN, and design ANN neuronal activation function that can effectively approximate the spike count of the coupled SNN. The proposed tandem learning rule demonstrates competitive pattern recognition and regression capabilities on both the conventional frame-based and event-based vision datasets, with at least an order of magnitude reduced inference time and total synaptic operations over other state-of-the-art SNN implementations. Therefore, the proposed tandem learning rule offers a novel solution to training efficient, low latency, and high accuracy deep SNNs with low computing resources. | ['Haizhou Li', 'Guoqi Li', 'Malu Zhang', 'Jibin Wu', 'Yansong Chua', 'Kay Chen Tan'] | 2019-07-02 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 5.11088848e-01 -6.21423602e-01 6.06077671e-01 -2.13927448e-01
7.29286075e-02 -1.40955806e-01 4.09621596e-01 1.05866984e-01
-9.56901431e-01 1.00197887e+00 -7.58962810e-01 -2.96950992e-02
9.67909545e-02 -9.80341792e-01 -1.00489783e+00 -1.09254646e+00
3.95692050e-01 1.41767174e-01 7.38173664e-01 6.94935173e-02
5.19876122e-01 9.02459919e-01 -1.91707957e+00 3.17920089e-01
6.20374322e-01 1.62639618e+00 3.81280184e-01 5.83210647e-01
-3.66161674e-01 9.01556790e-01 -4.94219810e-01 -1.39203638e-01
-5.09586837e-03 -3.74019384e-01 1.86426193e-01 -5.91103971e-01
1.36078103e-02 1.63347498e-01 -6.85602248e-01 9.08758342e-01
6.52417660e-01 -6.94774762e-02 5.20899355e-01 -1.29362702e+00
-3.90095621e-01 4.66269910e-01 -5.56247309e-02 4.33672279e-01
-2.13737190e-01 3.10973227e-01 3.52719396e-01 -9.96014416e-01
3.38216722e-01 7.37298250e-01 8.46257329e-01 6.20909631e-01
-1.38089800e+00 -1.03614080e+00 -1.38517216e-01 2.38889396e-01
-1.44643664e+00 -5.01928389e-01 5.29503703e-01 -1.75897032e-01
1.33808303e+00 -1.48425087e-01 1.20274282e+00 9.54746306e-01
5.74689269e-01 4.91169482e-01 1.16909969e+00 -2.73639619e-01
7.03799546e-01 -4.48506325e-01 1.25214219e-01 4.55962151e-01
5.63313544e-01 2.73719374e-02 -9.28685427e-01 1.45939011e-02
1.31306100e+00 4.11037236e-01 1.16346832e-02 1.10772751e-01
-8.46423209e-01 1.88227952e-01 7.06445515e-01 2.49203786e-01
-6.16628289e-01 8.62748265e-01 2.56264299e-01 -5.72317019e-02
-1.90623894e-01 1.85711995e-01 -8.19057748e-02 -1.60098970e-01
-9.73457158e-01 2.66676217e-01 7.13508844e-01 5.52531958e-01
6.51807010e-01 5.63472748e-01 -6.49739057e-02 6.11911356e-01
2.23088160e-01 5.98602176e-01 5.59230924e-01 -8.18356454e-01
-2.07301676e-01 8.75217438e-01 -2.36325204e-01 -7.86260247e-01
-3.59058380e-01 -4.18894589e-01 -1.12520039e+00 5.04531443e-01
4.36699778e-01 7.78841823e-02 -9.27419901e-01 1.61717355e+00
-2.12360710e-01 7.63364315e-01 -1.17407486e-01 7.46556997e-01
6.51687801e-01 7.26454139e-01 9.13286582e-03 -2.73824275e-01
1.07940292e+00 -4.95546311e-01 -7.13320374e-01 -4.05772984e-01
-1.36774387e-02 -3.13013464e-01 5.03045678e-01 2.56138504e-01
-1.31681049e+00 -5.74019313e-01 -1.18649936e+00 -1.16488822e-01
-5.99983871e-01 1.50330935e-03 6.28298640e-01 7.02650249e-02
-1.01123905e+00 6.50893807e-01 -1.12465954e+00 -2.03546882e-01
6.97922945e-01 6.93087220e-01 5.68438061e-02 6.45638928e-02
-7.40398645e-01 8.36573839e-01 4.09047306e-01 2.95905054e-01
-5.68710566e-01 -5.28698623e-01 -3.02415848e-01 2.42186740e-01
-3.06322724e-01 -6.22143269e-01 1.04574049e+00 -7.61748731e-01
-1.51945019e+00 8.42192411e-01 -4.96693164e-01 -8.97738934e-01
-5.44030182e-02 2.96709746e-01 -3.46654266e-01 -1.09614506e-01
-3.68533999e-01 7.09989488e-01 6.67545855e-01 -9.45877969e-01
-5.84185421e-01 -4.50013995e-01 -4.40283746e-01 -2.33880907e-01
-4.66582268e-01 -2.33951464e-01 -1.76089317e-01 -6.95328474e-01
3.98794085e-01 -5.73325694e-01 -5.04952371e-02 6.74322665e-01
2.44446918e-01 -2.26075426e-01 6.91776395e-01 -9.77403671e-02
1.19928789e+00 -2.24228525e+00 3.08021884e-02 2.03327239e-01
2.45121151e-01 5.34601569e-01 8.38760659e-02 1.00628860e-01
1.94763348e-01 -5.33229768e-01 -4.67954904e-01 -5.48053607e-02
-3.48434389e-01 4.93811280e-01 -2.75693625e-01 3.08408111e-01
4.60798442e-01 1.13582587e+00 -7.04351485e-01 -2.96960294e-01
5.25621437e-02 5.88209987e-01 -2.26395950e-01 2.86772490e-01
-1.63226292e-01 2.82148242e-01 -9.38117728e-02 7.57039428e-01
5.41026890e-01 -2.70772874e-01 -4.23932076e-02 -2.61604488e-01
-6.44342601e-01 1.95739090e-01 -8.85731161e-01 1.58887720e+00
-1.22976035e-01 7.43617475e-01 -1.71634406e-01 -1.19134128e+00
1.52320385e+00 9.67941713e-03 2.83326805e-01 -1.10453618e+00
3.23190749e-01 7.70812452e-01 3.23619395e-01 -6.54436871e-02
-3.37873191e-01 -4.80933934e-02 4.12753999e-01 2.71566927e-01
3.48141342e-01 1.27419427e-01 1.12962112e-01 -5.17406821e-01
1.36134803e+00 -8.92869756e-03 4.18606699e-02 -1.60894409e-01
5.53382099e-01 -2.51814127e-01 7.02517569e-01 7.43116736e-01
-1.19615585e-01 3.55846077e-01 4.85624522e-02 -7.38798499e-01
-1.10273588e+00 -1.32195354e+00 -4.50593561e-01 7.28741586e-01
1.46461651e-01 1.76844701e-01 -6.87992215e-01 4.37533319e-01
1.32503152e-01 3.28755707e-01 -3.63082230e-01 -2.52786785e-01
-8.35487783e-01 -6.70767725e-01 1.14572477e+00 7.47157156e-01
7.23251104e-01 -1.47377741e+00 -1.06606948e+00 7.12818027e-01
2.73251027e-01 -1.19939983e+00 1.68808281e-01 8.52549434e-01
-1.06891704e+00 -7.78193891e-01 -5.79534769e-01 -1.16385722e+00
7.23089993e-01 -2.68060505e-01 6.06919110e-01 -3.65390517e-02
-5.59686482e-01 -2.08336309e-01 2.41297364e-01 -7.44204998e-01
2.00133324e-01 -2.82213807e-01 1.80507138e-01 2.04418093e-01
8.53745043e-01 -1.04431272e+00 -7.34069467e-01 -9.60792005e-02
-7.84538031e-01 1.98615432e-01 6.14589512e-01 9.12897706e-01
1.24999774e+00 -2.67453462e-01 6.86436355e-01 -4.77556199e-01
2.77699769e-01 -1.48752376e-01 -7.85502970e-01 1.96093172e-01
-5.39078772e-01 1.08146600e-01 1.00723028e+00 -6.03036940e-01
-6.01341486e-01 1.70724705e-01 -2.62405992e-01 -4.14067388e-01
-2.59235296e-02 2.78592646e-01 3.51313084e-01 -6.56805694e-01
5.63905954e-01 7.79708087e-01 1.07694969e-01 -1.98961452e-01
-4.23183650e-01 4.73304629e-01 9.24307048e-01 -3.79473001e-01
3.87798578e-01 5.63589692e-01 2.75072426e-01 -6.65166974e-01
-2.46217564e-01 -1.18957765e-01 -4.19686139e-01 -3.93355012e-01
5.43576300e-01 -6.75742328e-01 -1.02053905e+00 1.09067047e+00
-1.56196046e+00 -2.53077507e-01 -3.39389145e-01 4.17777091e-01
-7.25011706e-01 -2.07521170e-01 -9.15305018e-01 -1.07657707e+00
-6.42076969e-01 -7.51692832e-01 5.31600118e-01 6.58038676e-01
1.36003613e-01 -6.00245178e-01 -2.38838810e-02 -3.77145648e-01
7.57860124e-01 2.74996459e-01 9.44078863e-01 -5.51644266e-01
-6.02856517e-01 -2.97216356e-01 -5.49303293e-01 1.09779455e-01
-1.63853198e-01 7.71254897e-02 -1.20158339e+00 -1.85090154e-01
2.22127140e-01 -3.99002701e-01 1.02959621e+00 7.14544237e-01
1.50143671e+00 7.99256265e-02 -4.28598315e-01 8.37011933e-01
1.83006072e+00 5.08918166e-01 9.29117143e-01 2.84635007e-01
2.55092651e-01 7.70303458e-02 -1.84919521e-01 5.68529963e-01
1.54091250e-02 3.09509337e-01 5.42472124e-01 1.22479290e-01
-1.05504431e-01 -1.70970947e-01 1.54587090e-01 1.05139112e+00
-3.34547967e-01 -4.69839461e-02 -8.60766709e-01 2.37678483e-01
-2.01967287e+00 -1.01744628e+00 -1.25320151e-01 2.29922986e+00
8.23489487e-01 4.60885346e-01 -2.72705555e-01 3.22889507e-01
7.87394047e-01 -3.35338712e-01 -1.22134626e+00 -3.89637262e-01
-5.72801113e-01 8.25133801e-01 4.20851797e-01 -2.86959410e-01
-5.19091070e-01 5.89194000e-01 6.03325367e+00 6.00886524e-01
-1.35671997e+00 -1.71220183e-01 2.68243998e-01 -2.29615271e-01
5.04711419e-02 -3.24804515e-01 -9.02228832e-01 8.56384933e-01
1.16928077e+00 -1.25774324e-01 8.54111254e-01 4.12794054e-01
3.45400088e-02 8.83687809e-02 -1.10239863e+00 1.55389822e+00
-2.21806645e-01 -1.86063671e+00 1.63723812e-01 -2.18639567e-01
4.45554316e-01 1.91097572e-01 -8.21230486e-02 1.73981071e-01
-3.80738080e-01 -1.10968995e+00 6.61830485e-01 1.12640786e+00
7.69997537e-01 -5.95190763e-01 7.71039605e-01 4.90670621e-01
-1.31908023e+00 -4.06511366e-01 -6.82920396e-01 -4.34927851e-01
-6.68743178e-02 5.99153996e-01 7.25628436e-02 -1.81725964e-01
9.33653235e-01 7.09640384e-01 -2.85149008e-01 1.60108685e+00
2.86760986e-01 4.00276721e-01 -6.68287456e-01 -4.78500426e-01
1.23395890e-01 -8.94677714e-02 1.30327046e-01 1.25445950e+00
6.64316952e-01 3.68211389e-01 -4.26877260e-01 1.34862411e+00
-4.78806525e-01 -3.42510492e-01 -6.57679319e-01 -3.14710557e-01
1.05055892e+00 1.11160123e+00 -8.82125378e-01 -1.37015313e-01
-3.53894114e-01 8.26073766e-01 5.54387331e-01 3.46307218e-01
-7.09086359e-01 -6.65068448e-01 6.22981250e-01 1.30880758e-01
4.31131244e-01 -2.53356099e-01 -9.62286830e-01 -8.02752912e-01
3.22620422e-02 -2.45410889e-01 -1.77379802e-01 -7.28841603e-01
-1.27848685e+00 5.92381060e-01 -8.58033001e-01 -1.17899060e+00
-7.01308530e-03 -1.14613128e+00 -8.49522948e-01 9.57947612e-01
-1.51220357e+00 -6.91562533e-01 -4.64272499e-01 6.82496965e-01
2.91543692e-01 -1.76633701e-01 1.06289303e+00 3.31110239e-01
-6.26105368e-01 4.19617683e-01 2.62187093e-01 1.99390918e-01
2.82405466e-01 -7.72079706e-01 2.68856615e-01 8.24652135e-01
-5.36989013e-04 5.33212662e-01 2.31240854e-01 -3.30383360e-01
-1.59135199e+00 -1.00213337e+00 7.91670859e-01 2.53237635e-01
6.41853929e-01 -4.86843377e-01 -1.17907059e+00 2.17021897e-01
-2.00625643e-01 4.78738606e-01 6.36996448e-01 -7.83541858e-01
-2.66953588e-01 -6.38970971e-01 -1.14836109e+00 5.39705157e-01
1.07878387e+00 -4.88317698e-01 -5.03010273e-01 -1.49553940e-01
3.10503989e-01 -1.80930898e-01 -6.58946931e-01 4.71636385e-01
8.78587186e-01 -9.16842580e-01 7.87801921e-01 -2.04914421e-01
3.42820585e-01 -4.59876359e-01 -2.62006015e-01 -7.63755620e-01
-4.23267096e-01 -7.22904503e-02 -6.98446214e-01 9.09926951e-01
2.64127851e-01 -8.87268722e-01 8.43541622e-01 4.54798907e-01
-1.79891944e-01 -1.11171687e+00 -1.36003315e+00 -6.89370096e-01
-2.07243085e-01 -2.84661114e-01 2.74342269e-01 2.33609527e-01
-2.03607127e-01 -5.48616871e-02 2.40236878e-01 3.07996292e-02
9.03342068e-01 -5.73566109e-02 1.74415726e-02 -1.47758627e+00
-1.89628780e-01 -7.14808285e-01 -8.85898769e-01 -8.68779302e-01
-6.76757023e-02 -7.54952133e-01 2.76827544e-01 -1.50100338e+00
-8.23129788e-02 -3.57716471e-01 -9.38136697e-01 5.10957301e-01
3.22063833e-01 5.36260247e-01 -7.21982196e-02 4.47540164e-01
-2.93536872e-01 3.57200682e-01 8.14105749e-01 4.08013202e-02
1.17898121e-01 -8.70216936e-02 -7.07302839e-02 7.55430222e-01
7.34349370e-01 -5.29884279e-01 -4.75126207e-02 -6.05106771e-01
1.03075899e-01 -1.98432297e-01 6.28739893e-01 -1.63561773e+00
1.24393880e+00 6.75074011e-02 9.62053716e-01 -5.10330439e-01
4.32658494e-01 -8.66034448e-01 1.72141194e-01 8.70712399e-01
-2.63042539e-01 4.62772436e-02 4.00578260e-01 5.50910056e-01
-2.39809304e-01 -2.08953992e-01 9.86608863e-01 -8.39735195e-02
-6.51494980e-01 4.31005359e-01 -6.76093817e-01 -2.49226108e-01
8.90940249e-01 -8.60705197e-01 -3.92789572e-01 3.87786210e-01
-3.49348605e-01 -1.04427159e-01 1.75139159e-01 -8.25506672e-02
1.19299996e+00 -1.41271591e+00 -2.91274577e-01 7.21395850e-01
1.88058577e-02 3.00093126e-02 3.55336294e-02 7.77889132e-01
-5.87572157e-01 4.07509089e-01 -1.00025511e+00 -6.37879252e-01
-6.35665774e-01 1.47877052e-01 5.96051931e-01 1.34085670e-01
-2.91850984e-01 9.95682001e-01 -3.82356375e-01 -1.33550689e-01
5.01851320e-01 -1.71871409e-01 -4.73228022e-02 -3.62391025e-01
6.09863162e-01 4.55994159e-01 1.85468733e-01 -1.62662700e-01
-4.67509001e-01 5.67725301e-01 1.18700646e-01 1.76446810e-02
1.45521176e+00 2.58569360e-01 -5.21677375e-01 9.81585324e-01
7.96425581e-01 -9.86490309e-01 -1.39889038e+00 -2.87634760e-01
-3.09372619e-02 8.54964703e-02 2.29378119e-02 -7.23188996e-01
-9.41629052e-01 1.08416009e+00 8.27884912e-01 -8.25756639e-02
1.30102122e+00 -5.25380433e-01 1.04109573e+00 7.36104250e-01
6.92107081e-01 -1.19550216e+00 -9.14213434e-02 6.62798285e-01
6.27725124e-01 -7.13167369e-01 -5.11571169e-01 1.70202311e-02
2.09771693e-01 1.59277368e+00 1.11249673e+00 -7.59514749e-01
7.20406651e-01 8.91199291e-01 -3.74979198e-01 3.47739924e-03
-9.59412813e-01 4.37258072e-02 -2.07664862e-01 5.32382071e-01
3.97161812e-01 -1.37529463e-01 -3.51912796e-01 7.89647877e-01
2.84009725e-01 7.42397487e-01 8.26392695e-02 9.55064833e-01
-7.97480702e-01 -7.24630415e-01 -8.49261805e-02 9.59302843e-01
-1.70088097e-01 -2.14700893e-01 -6.84467033e-02 2.33208567e-01
2.88918823e-01 4.18682396e-01 5.82343280e-01 -3.81912678e-01
3.84202421e-01 2.72413790e-01 7.49327719e-01 -3.57251078e-01
-6.82006419e-01 -1.82280347e-01 -7.33128428e-01 -4.90632415e-01
-2.36131489e-01 -1.69130221e-01 -1.98440385e+00 -4.38965619e-01
-1.69866502e-01 -3.01763892e-01 8.82144392e-01 1.04957962e+00
4.03246045e-01 6.46125615e-01 3.76903147e-01 -1.08044827e+00
-3.01879883e-01 -6.65218830e-01 -6.14405513e-01 1.30400777e-01
9.37744305e-02 -5.91928840e-01 -2.52362937e-01 2.17071157e-02] | [8.218690872192383, 2.487032413482666] |
e6fb963b-f52d-466e-b0e9-a4de03517e97 | a-time-series-analysis-based-stock-price | 2004.11697 | null | https://arxiv.org/abs/2004.11697v2 | https://arxiv.org/pdf/2004.11697v2.pdf | A Time Series Analysis-Based Stock Price Prediction Using Machine Learning and Deep Learning Models | Prediction of future movement of stock prices has always been a challenging task for the researchers. While the advocates of the efficient market hypothesis (EMH) believe that it is impossible to design any predictive framework that can accurately predict the movement of stock prices, there are seminal work in the literature that have clearly demonstrated that the seemingly random movement patterns in the time series of a stock price can be predicted with a high level of accuracy. Design of such predictive models requires choice of appropriate variables, right transformation methods of the variables, and tuning of the parameters of the models. In this work, we present a very robust and accurate framework of stock price prediction that consists of an agglomeration of statistical, machine learning and deep learning models. We use the daily stock price data, collected at five minutes interval of time, of a very well known company that is listed in the National Stock Exchange (NSE) of India. The granular data is aggregated into three slots in a day, and the aggregated data is used for building and training the forecasting models. We contend that the agglomerative approach of model building that uses a combination of statistical, machine learning, and deep learning approaches, can very effectively learn from the volatile and random movement patterns in a stock price data. We build eight classification and eight regression models based on statistical and machine learning approaches. In addition to these models, a deep learning regression model using a long-and-short-term memory (LSTM) network is also built. Extensive results have been presented on the performance of these models, and the results are critically analyzed. | ['Sidra Mehtab', 'Jaydip Sen'] | 2020-04-17 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-8.29804599e-01 -4.82070595e-01 -8.16525295e-02 -2.92576641e-01
-4.54644829e-01 -5.82134962e-01 6.47288859e-01 -2.07810655e-01
-1.39989272e-01 8.14838052e-01 1.98847279e-01 -9.00751352e-01
-3.33742768e-01 -1.10681617e+00 -5.25464833e-01 -7.24679530e-01
-3.84892792e-01 6.23799562e-01 -1.31405313e-02 -4.10320163e-01
6.66305840e-01 6.85626924e-01 -1.26013434e+00 1.52400613e-01
3.29246134e-01 1.69660497e+00 5.34353778e-02 3.46040368e-01
-3.95731926e-01 1.46000433e+00 -6.46001458e-01 -2.14707017e-01
7.85220623e-01 -2.04310775e-01 -4.36601758e-01 -1.64087772e-01
-3.18549305e-01 -5.05562901e-01 -2.20666975e-01 7.03095853e-01
1.24835812e-01 -1.11984268e-01 4.78841394e-01 -1.13122582e+00
-9.26798284e-01 9.26740706e-01 -4.81633604e-01 7.70054460e-01
-5.11518359e-01 -3.53543796e-02 1.25850093e+00 -8.16135406e-01
2.00179815e-01 7.48791635e-01 8.26822937e-01 -1.90013498e-01
-9.31229889e-01 -9.66021657e-01 2.51324521e-03 1.48489654e-01
-1.07381225e+00 1.25040039e-01 8.88128519e-01 -7.34708130e-01
1.18941319e+00 8.94877985e-02 1.03347373e+00 5.88553786e-01
7.55569100e-01 5.13501644e-01 1.37955821e+00 -1.99458435e-01
2.36076564e-01 2.35938311e-01 8.40537846e-02 -6.02835901e-02
2.60630608e-01 4.60915476e-01 -2.61798233e-01 -1.90092370e-01
8.42114747e-01 4.65301991e-01 2.90734112e-01 2.31084064e-01
-8.42151761e-01 1.17591774e+00 5.28364599e-01 8.72489095e-01
-9.78358984e-01 2.40763314e-02 3.13066900e-01 7.48149395e-01
8.95921826e-01 4.52952027e-01 -1.05861640e+00 -1.49210617e-01
-1.44898081e+00 4.58165497e-01 9.30710614e-01 2.89497524e-01
2.95288414e-01 7.31059551e-01 2.45298490e-01 1.75648063e-01
3.36712778e-01 4.26899940e-01 1.18941259e+00 -4.30843413e-01
2.53383607e-01 6.03312969e-01 3.27768683e-01 -1.20878768e+00
-5.56492031e-01 -7.53279030e-01 -6.78151667e-01 4.61050153e-01
1.33373603e-01 -4.97067183e-01 -6.78622663e-01 1.15894604e+00
-2.56329805e-01 3.25928420e-01 2.15534851e-01 3.17457557e-01
1.99203819e-01 1.11956632e+00 -2.23072797e-01 -5.14357567e-01
8.39603364e-01 -6.98052347e-01 -7.93839991e-01 1.69076771e-01
4.71808285e-01 -6.45063281e-01 4.24325526e-01 3.99426103e-01
-1.11351180e+00 -5.76851726e-01 -9.87639248e-01 2.92836308e-01
-9.34862733e-01 -2.40211338e-01 6.26870811e-01 1.73847079e-01
-1.06757772e+00 1.05478644e+00 -8.12493682e-01 4.05339748e-01
6.30459711e-02 6.28466666e-01 1.78058892e-01 8.97430539e-01
-1.30804789e+00 1.27824819e+00 6.33340538e-01 4.59195405e-01
-1.72140524e-01 -5.97563744e-01 -2.24098593e-01 7.73136839e-02
-2.03308254e-01 -2.11903989e-01 1.30148363e+00 -1.34246004e+00
-1.28291082e+00 3.62522721e-01 3.58997077e-01 -1.08080578e+00
3.44586819e-01 -9.23993811e-02 -7.70616114e-01 -4.93933141e-01
-1.56641737e-01 5.46586066e-02 6.35565102e-01 -7.25422382e-01
-1.06297660e+00 -4.28199619e-01 -3.30219239e-01 -1.71479240e-01
-6.43228590e-02 1.33931518e-01 5.14084399e-01 -8.75995576e-01
1.81115255e-01 -7.69020021e-01 -1.67765781e-01 -1.03704929e+00
2.90799048e-02 -3.95799935e-01 8.42936397e-01 -1.14412093e+00
1.45116377e+00 -1.96248603e+00 -4.46933389e-01 5.61211228e-01
-1.46376982e-01 6.14307029e-03 3.42343956e-01 5.83175838e-01
-5.04242182e-01 9.01600346e-02 5.40208891e-02 1.43421460e-02
3.29834163e-01 1.85323462e-01 -1.17629087e+00 3.20358902e-01
2.15679817e-02 1.15765846e+00 -3.18193674e-01 1.46930903e-01
2.35670075e-01 2.14932621e-01 -8.81429005e-04 -5.20579852e-02
-2.72454977e-01 5.85921295e-02 -3.41035247e-01 4.89799261e-01
5.58571517e-01 -3.68374914e-01 -3.23751755e-02 1.16354145e-01
-5.85461080e-01 5.97422302e-01 -1.13879597e+00 5.66436946e-01
-1.41733766e-01 5.89041352e-01 -5.82020402e-01 -1.04064536e+00
1.30649412e+00 5.17926395e-01 8.14158916e-01 -9.93186533e-01
1.42224595e-01 8.90500247e-01 2.51071990e-01 -1.51956350e-01
5.19956470e-01 -6.41528547e-01 -7.32704531e-03 7.03081310e-01
-2.79017568e-01 1.65059179e-01 7.98093081e-02 -5.34293652e-01
5.85015833e-01 8.16815626e-03 1.72141567e-01 -4.04936939e-01
2.80984610e-01 1.08655378e-01 5.94750583e-01 3.58522564e-01
6.45619109e-02 1.07741922e-01 5.29486537e-01 -1.29495430e+00
-1.20586336e+00 -6.97600007e-01 -3.18513274e-01 8.57182562e-01
-4.80864435e-01 2.64695793e-01 -3.21944088e-01 -2.37896871e-02
3.75070006e-01 1.05475235e+00 -7.05327094e-01 2.99927235e-01
-6.13744497e-01 -1.23834503e+00 8.78199935e-02 6.24246180e-01
4.57975000e-01 -1.50897193e+00 -6.67097509e-01 5.61013877e-01
3.87526423e-01 -6.89534187e-01 1.00641169e-01 5.37876666e-01
-1.12867320e+00 -8.46675098e-01 -3.20674777e-01 -7.20837116e-01
8.38850215e-02 -1.55110508e-01 1.23998165e+00 -9.49308649e-02
4.34549302e-01 -2.60486394e-01 -1.57529190e-01 -1.00272274e+00
-1.58183023e-01 1.33095399e-01 4.06393744e-02 2.51573492e-02
8.59290242e-01 -7.13064909e-01 -4.61076558e-01 -5.19440547e-02
-8.94686699e-01 -2.42210224e-01 6.98782682e-01 6.23279810e-01
4.39510226e-01 6.73805177e-01 8.68690550e-01 -5.11640131e-01
8.08102012e-01 -8.73202145e-01 -1.07409012e+00 -1.98374055e-02
-1.07446849e+00 -3.85309570e-02 5.51022232e-01 -6.09231219e-02
-6.86235785e-01 -3.81887376e-01 -8.34753439e-02 -3.67860407e-01
2.58810163e-01 1.08697510e+00 5.97716630e-01 3.00890177e-01
1.30888686e-01 6.83604717e-01 -2.22465191e-02 -5.34979880e-01
-1.15051977e-01 6.65429473e-01 3.40915650e-01 -4.11326326e-02
9.67455208e-01 3.08525980e-01 -1.33146718e-01 -5.08957088e-01
-6.62842453e-01 -1.59724668e-01 -8.80702198e-01 5.45606669e-03
5.95013380e-01 -9.36285257e-01 -3.07972461e-01 8.14179599e-01
-6.89060509e-01 -3.35319966e-01 -4.20256019e-01 4.44519192e-01
-4.70022768e-01 -4.07785058e-01 -8.34665656e-01 -1.11701596e+00
-5.54158509e-01 -8.57776761e-01 5.80893040e-01 9.26672891e-02
-3.09603482e-01 -1.52834392e+00 2.48273998e-01 1.73809212e-02
7.88868845e-01 6.09623969e-01 8.71155739e-01 -1.35802770e+00
-6.36971235e-01 -6.31636322e-01 9.45096314e-02 5.81428528e-01
2.77145505e-01 1.33892521e-01 -6.57103658e-01 -1.56710625e-01
6.76034808e-01 -8.90134424e-02 5.85810781e-01 7.41554320e-01
6.73900664e-01 -5.29837668e-01 1.59848273e-01 5.07306159e-01
1.60232043e+00 9.24017370e-01 5.56073129e-01 1.01012254e+00
2.21778825e-01 3.88249129e-01 3.19602132e-01 5.05933940e-01
3.55766773e-01 2.09143519e-01 9.64371562e-02 2.23572385e-02
8.83400083e-01 -9.17455330e-02 4.42027599e-01 1.20136750e+00
-2.75329202e-01 3.11640412e-01 -8.56148005e-01 3.42418283e-01
-1.74220896e+00 -1.39549172e+00 -3.52768302e-02 1.82690811e+00
6.38328135e-01 4.98823047e-01 3.89297277e-01 4.50332731e-01
3.04895163e-01 3.61700267e-01 -5.26799142e-01 -5.32014310e-01
-3.11977565e-01 2.10363209e-01 8.34621131e-01 2.77307183e-01
-1.23706520e+00 7.21778989e-01 6.87407017e+00 4.53594387e-01
-1.62137163e+00 -2.73588419e-01 1.04937816e+00 4.58293827e-03
-2.68448651e-01 -2.80973464e-01 -8.66872728e-01 9.39042389e-01
1.55699289e+00 -4.98485535e-01 2.74782807e-01 9.66689765e-01
5.64243555e-01 2.38467216e-01 -6.61936522e-01 6.86780035e-01
-2.69852549e-01 -1.69709980e+00 -5.39220050e-02 4.20575589e-01
8.82353663e-01 3.74704331e-01 4.63567853e-01 4.14385766e-01
3.54023397e-01 -1.06964815e+00 9.52560902e-01 9.31479394e-01
-2.60210246e-01 -1.03629923e+00 1.24914241e+00 5.53117275e-01
-1.03430307e+00 -6.19635224e-01 -2.74687558e-01 -6.30626857e-01
2.39074603e-03 4.64699894e-01 -2.88350075e-01 3.38233232e-01
8.75853360e-01 6.43438756e-01 -3.01506460e-01 6.60054266e-01
2.17301518e-01 5.02680957e-01 -3.05491745e-01 -8.00689086e-02
8.11852038e-01 -6.53023005e-01 -8.97010192e-02 7.21684337e-01
6.86137438e-01 9.16687846e-02 -7.98401758e-02 8.53395760e-01
3.00218701e-01 1.41890839e-01 -5.84079266e-01 -1.67419523e-01
2.43584514e-01 8.07749212e-01 -6.79850519e-01 -3.64325315e-01
-7.05404341e-01 2.94122696e-01 -9.00167674e-02 1.60507962e-01
-7.14861572e-01 3.99742648e-03 4.94909644e-01 2.85592318e-01
5.75693548e-01 -3.38688523e-01 -7.48236358e-01 -9.13648784e-01
2.06686303e-01 -8.07457209e-01 3.24470162e-01 -6.12043738e-01
-1.60033762e+00 7.21252620e-01 -1.31038204e-01 -1.10364318e+00
-7.79096484e-01 -8.87659967e-01 -9.77678895e-01 1.11800361e+00
-1.65543830e+00 -7.20829904e-01 3.40132684e-01 5.60131788e-01
4.91074294e-01 -7.94294000e-01 5.65627456e-01 -1.35548552e-02
-5.41512370e-01 -2.08862454e-01 7.61331797e-01 4.97861266e-01
-1.55612856e-01 -1.53937328e+00 7.78907418e-01 6.85843289e-01
3.08864266e-01 3.72343063e-01 5.96494555e-01 -7.84436226e-01
-8.45406175e-01 -1.08383083e+00 1.37160897e+00 -3.16351265e-01
1.47945487e+00 1.32795259e-01 -1.05709767e+00 1.09777761e+00
3.39204967e-01 -1.93812892e-01 7.38742173e-01 -3.24265778e-01
7.65091777e-02 -3.85116428e-01 -7.92900205e-01 1.94307908e-01
-4.21526618e-02 -3.50754946e-01 -1.11218309e+00 2.82700747e-01
3.53415519e-01 -3.27682868e-02 -9.52369332e-01 1.66720018e-01
5.26013494e-01 -1.33807516e+00 8.18485677e-01 -6.05282545e-01
1.52785376e-01 -2.58784331e-02 -2.60582030e-01 -1.28810024e+00
-6.21168911e-01 -5.00000954e-01 -2.08679482e-01 8.35995436e-01
6.74372077e-01 -1.06396115e+00 6.75149620e-01 1.04899216e+00
4.97983880e-02 -8.69848073e-01 -8.47413063e-01 -8.65983665e-01
6.43051147e-01 -3.79188031e-01 1.04833484e+00 1.03726590e+00
-7.71914795e-02 3.89564186e-02 -3.30247909e-01 7.42034614e-02
2.51917481e-01 5.86468399e-01 4.40786391e-01 -1.49289608e+00
-1.26773998e-01 -8.21308613e-01 -3.03550303e-01 -5.17109275e-01
1.74306139e-01 -4.23084676e-01 -5.11484802e-01 -1.35880566e+00
-2.88966924e-01 -4.12112534e-01 -8.67082298e-01 2.11961299e-01
3.24791133e-01 -1.06915146e-01 1.64485067e-01 8.58760595e-01
2.19129071e-01 3.22777748e-01 7.43228912e-01 9.91995037e-02
-3.91047120e-01 3.23148221e-01 -5.70412099e-01 8.67895067e-01
1.21197331e+00 -1.96047261e-01 -2.28579968e-01 -1.48542196e-01
4.84323889e-01 -2.88623180e-02 1.25184178e-01 -8.02604795e-01
1.69809684e-01 -2.67475545e-01 8.05940926e-01 -1.20783877e+00
3.00614256e-02 -9.08605695e-01 5.41365981e-01 7.22865701e-01
-8.88718516e-02 9.96818483e-01 1.57703415e-01 8.09507910e-03
-7.39491701e-01 -8.71324092e-02 5.75732350e-01 -4.82451707e-01
-7.40809381e-01 4.24869120e-01 -2.32157946e-01 -3.22594374e-01
1.22686148e+00 -3.10256809e-01 5.53925298e-02 -4.03216869e-01
-6.38842702e-01 4.11729924e-02 -1.03805795e-01 4.62893754e-01
3.29050839e-01 -1.53432417e+00 -7.33354032e-01 3.47797692e-01
-6.25034690e-01 -3.64751279e-01 -1.94163084e-01 7.83225596e-01
-9.35754538e-01 8.71515930e-01 -2.75299281e-01 -4.89117354e-02
-3.59145671e-01 8.29159439e-01 6.11687183e-01 -5.67656517e-01
-7.11279452e-01 4.47584242e-01 -2.22384021e-01 5.13674282e-02
-7.82442987e-02 -7.77574003e-01 -3.88365805e-01 6.81469679e-01
6.95516884e-01 2.91685402e-01 1.59707487e-01 -9.07602191e-01
6.57319278e-02 5.29961348e-01 1.34829566e-01 -1.38540164e-01
2.10581636e+00 -8.18497911e-02 -3.24723959e-01 1.14924359e+00
1.04956877e+00 -4.55448121e-01 -1.02518415e+00 -1.50074854e-01
8.19997728e-01 -8.78843367e-02 2.51511574e-01 -5.74193299e-01
-1.29755986e+00 5.49790800e-01 4.32719290e-01 1.06195629e+00
1.04851234e+00 -4.80174959e-01 1.02263200e+00 2.04973117e-01
1.83970019e-01 -1.42180622e+00 -4.06766683e-01 6.28507018e-01
8.26442719e-01 -1.16901612e+00 -2.56272405e-01 5.85836709e-01
-5.20502508e-01 1.29554760e+00 1.39019443e-02 -7.31225431e-01
1.30606341e+00 4.98369217e-01 3.66767645e-01 -2.51828313e-01
-9.78810847e-01 1.48244590e-01 1.19615577e-01 2.55580824e-02
3.61517191e-01 1.48706719e-01 -1.79588906e-02 9.12782013e-01
-8.83269787e-01 9.62996557e-02 4.00722921e-01 8.13354135e-01
-6.46814883e-01 -7.26673722e-01 -4.78893936e-01 6.56171978e-01
-9.66626942e-01 -3.09096307e-01 -2.17410177e-01 1.16070044e+00
6.45207912e-02 6.23251498e-01 7.88283169e-01 -2.61073261e-01
2.40617573e-01 4.83716190e-01 -2.67371893e-01 -1.50730819e-01
-8.17144811e-01 2.64932990e-01 -3.21958661e-01 -1.58561543e-01
-2.56480068e-01 -8.22324276e-01 -1.12039661e+00 -7.45289445e-01
3.50355282e-02 3.48633260e-01 6.09769464e-01 1.22321880e+00
4.61047888e-02 3.30743849e-01 1.21608639e+00 -8.30127835e-01
-1.09521246e+00 -1.09875095e+00 -1.34461665e+00 2.39225090e-01
4.75588709e-01 -4.25628752e-01 -5.70390165e-01 6.63462430e-02] | [4.4752936363220215, 4.225456237792969] |
91c90b3a-7b2d-4144-ae0a-e89be5c0c5d2 | a-comparative-study-on-e-branchformer-vs | 2305.11073 | null | https://arxiv.org/abs/2305.11073v1 | https://arxiv.org/pdf/2305.11073v1.pdf | A Comparative Study on E-Branchformer vs Conformer in Speech Recognition, Translation, and Understanding Tasks | Conformer, a convolution-augmented Transformer variant, has become the de facto encoder architecture for speech processing due to its superior performance in various tasks, including automatic speech recognition (ASR), speech translation (ST) and spoken language understanding (SLU). Recently, a new encoder called E-Branchformer has outperformed Conformer in the LibriSpeech ASR benchmark, making it promising for more general speech applications. This work compares E-Branchformer and Conformer through extensive experiments using different types of end-to-end sequence-to-sequence models. Results demonstrate that E-Branchformer achieves comparable or better performance than Conformer in almost all evaluation sets across 15 ASR, 2 ST, and 3 SLU benchmarks, while being more stable during training. We will release our training configurations and pre-trained models for reproducibility, which can benefit the speech community. | ['Shinji Watanabe', 'Prashant Sridhar', 'Suwon Shon', 'Jiyang Tang', 'William Chen', 'Siddhant Arora', 'Brian Yan', 'Felix Wu', 'Kwangyoun Kim', 'Yifan Peng'] | 2023-05-18 | null | null | null | null | ['spoken-language-understanding', 'spoken-language-understanding'] | ['natural-language-processing', 'speech'] | [ 4.04210538e-01 -1.19742438e-01 -1.33864582e-01 -6.44858778e-01
-1.02177203e+00 -5.74308276e-01 7.14937329e-01 -2.34242886e-01
-6.06956661e-01 4.44813281e-01 7.01283097e-01 -8.77462506e-01
3.73714149e-01 -1.73055261e-01 -5.94081521e-01 -2.31899723e-01
-7.62889087e-02 5.02408743e-01 1.97891165e-02 -3.62025112e-01
-2.95741946e-01 8.13866928e-02 -1.08023143e+00 6.54418111e-01
5.40238321e-01 9.62695837e-01 3.87377262e-01 8.43899548e-01
-2.54818518e-02 6.70750618e-01 -6.29521966e-01 -5.99565804e-01
-1.07803725e-01 -3.62801731e-01 -8.58178735e-01 -1.78649008e-01
2.02306449e-01 -2.50430465e-01 -6.23046219e-01 7.42034137e-01
8.64949226e-01 2.84242451e-01 3.65421116e-01 -8.29898059e-01
-9.12743688e-01 1.15941739e+00 -1.27861470e-01 5.48785627e-01
6.30812168e-01 -3.48031521e-02 1.12165380e+00 -1.28088272e+00
4.89643991e-01 1.48421884e+00 5.77065408e-01 7.33028829e-01
-1.01986098e+00 -7.02577055e-01 -1.32861650e-02 3.77372146e-01
-1.40722501e+00 -1.32405806e+00 2.92190909e-01 1.31494235e-02
1.76786149e+00 4.22032118e-01 1.72273383e-01 1.56366086e+00
-1.15741350e-01 1.24769223e+00 9.01652455e-01 -2.93653518e-01
2.60301322e-01 -2.29133517e-01 -3.68209630e-02 4.96970147e-01
-4.21860009e-01 4.40634668e-01 -8.80115569e-01 2.10892763e-02
4.18888003e-01 -6.82496846e-01 -6.98197901e-01 4.04707134e-01
-1.43007576e+00 6.16445601e-01 1.99058235e-01 4.83183980e-01
-2.85884768e-01 -7.38795102e-02 7.11412072e-01 6.48708582e-01
5.61221004e-01 1.87784851e-01 -5.47492921e-01 -8.26645255e-01
-8.42467487e-01 -6.69557899e-02 8.08306873e-01 1.15044093e+00
7.63169825e-02 7.13352263e-01 -2.79902101e-01 1.42090595e+00
2.78552413e-01 5.60691476e-01 8.26120496e-01 -5.42679727e-01
7.32296228e-01 -1.16866820e-01 -3.06072712e-01 -1.00287035e-01
-8.01424682e-02 -6.99623168e-01 -9.51075494e-01 -4.20195520e-01
-4.50435542e-02 -1.23176679e-01 -1.26951802e+00 1.71238637e+00
-2.54795551e-01 2.85790026e-01 3.99475276e-01 8.91115487e-01
1.14718711e+00 1.24155593e+00 -1.91656113e-01 -3.21779311e-01
1.27897477e+00 -1.11119211e+00 -1.00362647e+00 -5.03297865e-01
6.81696832e-01 -1.01558328e+00 1.03625941e+00 2.76633888e-01
-1.19683266e+00 -4.47515875e-01 -1.01930380e+00 -1.18466310e-01
-1.95745260e-01 2.39200294e-01 1.29143715e-01 7.18709290e-01
-1.46536589e+00 2.97931463e-01 -9.66658950e-01 -4.21049953e-01
1.16027951e-01 1.36829183e-01 -4.65728074e-01 4.57137972e-02
-1.49145699e+00 9.54590917e-01 4.24454063e-01 -1.53691182e-02
-7.25111067e-01 -6.10488117e-01 -9.81992066e-01 2.11361349e-01
7.24898726e-02 -3.23991656e-01 1.98967469e+00 -5.38840830e-01
-2.03048158e+00 6.55012548e-01 -4.87419516e-01 -9.60616827e-01
1.48426250e-01 -1.94708884e-01 -8.44901860e-01 -2.67459065e-01
-3.01125795e-01 5.95175028e-01 6.49101496e-01 -6.52186394e-01
-4.00863975e-01 -7.49890134e-02 -4.45958376e-01 2.37023115e-01
-1.56383306e-01 5.69012344e-01 -2.08161652e-01 -9.61325228e-01
1.32455304e-01 -6.90802574e-01 1.56324521e-01 -5.19035399e-01
-3.86428148e-01 -5.59052110e-01 9.37244177e-01 -9.26774919e-01
1.31133556e+00 -2.18058109e+00 1.52778059e-01 -3.19788486e-01
-7.50150308e-02 9.40551221e-01 -4.57766205e-01 7.52724230e-01
-3.11277360e-01 8.94161016e-02 -3.60859543e-01 -7.27061450e-01
9.45970938e-02 2.93297023e-01 -5.87501168e-01 1.86709672e-01
1.64979845e-01 9.74996507e-01 -8.11213493e-01 7.38665462e-02
5.02013624e-01 5.86178958e-01 -2.96320945e-01 4.29841876e-01
-1.24622747e-01 3.52250814e-01 2.74482407e-02 4.08915192e-01
3.06887507e-01 5.37098683e-02 -7.29146227e-02 1.88004434e-01
-3.55423614e-02 1.20468903e+00 -5.65285265e-01 1.66247523e+00
-8.59841406e-01 1.03615570e+00 2.16018066e-01 -7.08908975e-01
1.12698245e+00 9.43663180e-01 2.68773846e-02 -1.00467205e+00
1.22074131e-02 5.27190387e-01 3.85860860e-01 -1.25953212e-01
5.37148535e-01 -9.38616022e-02 3.88159215e-01 4.89953518e-01
2.76027292e-01 -1.90578923e-01 5.21460548e-02 2.55510122e-01
1.30869424e+00 -3.47198457e-01 4.27234828e-01 3.07225008e-02
5.21157920e-01 -5.78990579e-01 2.18989015e-01 4.84056503e-01
-2.61158764e-01 6.55066490e-01 -9.42530707e-02 -2.81503111e-01
-1.11397219e+00 -1.20132756e+00 -2.70131886e-01 1.02041125e+00
-4.74367857e-01 -6.94708824e-01 -7.12518036e-01 -2.81257451e-01
-4.55294728e-01 1.09603286e+00 -3.01887617e-02 -4.43409495e-02
-7.13287532e-01 -3.32637876e-01 9.88042891e-01 6.01401925e-01
4.44620669e-01 -1.16603696e+00 1.03060693e-01 5.47882855e-01
-4.16053861e-01 -1.53649735e+00 -9.17640865e-01 2.35713243e-01
-4.42918956e-01 -3.55038226e-01 -8.84160757e-01 -8.37942958e-01
-4.48899679e-02 2.80314922e-01 1.05237246e+00 -2.64385402e-01
3.01806957e-01 7.19672590e-02 -8.43209088e-01 -2.08385050e-01
-1.10525835e+00 3.49826902e-01 3.25033784e-01 9.18996707e-03
2.19540358e-01 -5.12628794e-01 -1.03789560e-01 3.44009876e-01
-6.31854415e-01 1.09252959e-01 4.68564451e-01 1.01274359e+00
2.79998451e-01 -3.89813006e-01 7.96834230e-01 -4.31987762e-01
8.83997321e-01 -1.27803743e-01 -3.84397894e-01 2.07021147e-01
-4.37439770e-01 1.56412974e-01 7.86960959e-01 -2.14077711e-01
-9.96952057e-01 -3.37812841e-01 -9.74959791e-01 -5.28942883e-01
-4.80882488e-02 6.93050742e-01 -6.31557852e-02 3.81859809e-01
5.01281619e-01 7.13097036e-01 -1.48001686e-01 -6.95071101e-01
3.16405982e-01 1.46921861e+00 5.83734810e-01 -2.59679884e-01
4.96000111e-01 -2.92713672e-01 -8.22264373e-01 -1.32270122e+00
-5.71488440e-01 -5.52698612e-01 -2.79683266e-02 1.75387219e-01
6.24780595e-01 -1.01004505e+00 -4.53180313e-01 5.67688525e-01
-1.59645689e+00 -4.71754700e-01 3.71942222e-02 6.61084056e-01
-5.45114636e-01 3.40067089e-01 -8.27921212e-01 -8.40213358e-01
-7.09456742e-01 -1.47027552e+00 1.22769332e+00 -2.82273024e-01
-4.19502258e-01 -8.40904295e-01 1.06948949e-01 3.83922696e-01
7.53216326e-01 -8.24885547e-01 7.01364398e-01 -1.01520622e+00
-4.28817779e-01 1.16020837e-03 -5.55778928e-02 7.61241853e-01
2.02502400e-01 -2.04933465e-01 -1.16098809e+00 -7.21653759e-01
-2.70735800e-01 -3.94030720e-01 8.68076384e-01 1.58664495e-01
1.03150737e+00 -4.24306393e-01 3.51700839e-03 7.31614530e-01
6.34193361e-01 5.82587421e-01 7.18440950e-01 -1.46292103e-02
2.37930402e-01 1.18252352e-01 1.22084014e-01 9.42069814e-02
2.68515646e-01 1.14339960e+00 -1.64155290e-02 1.53013468e-01
-5.01733422e-01 -4.06970382e-01 9.70156670e-01 1.80917835e+00
2.86058217e-01 -8.18138719e-01 -8.74151587e-01 5.63758016e-01
-1.49409354e+00 -9.65679586e-01 2.20921427e-01 2.13477969e+00
8.25159371e-01 2.77294256e-02 -3.68094258e-02 1.64404556e-01
6.03119016e-01 5.05548716e-01 -2.36362249e-01 -8.63507688e-01
-2.30403259e-01 5.58649600e-01 1.92770958e-01 6.83212101e-01
-7.34965324e-01 1.19835472e+00 6.42645454e+00 1.04331982e+00
-1.37175763e+00 4.01974559e-01 5.22161245e-01 1.46531954e-01
-3.21654290e-01 -2.68532604e-01 -7.43750691e-01 4.13288623e-01
1.52518630e+00 -2.98621565e-01 7.86766052e-01 5.80201209e-01
3.71194452e-01 5.02049029e-01 -1.20417500e+00 1.23002994e+00
-8.06839764e-02 -1.35829556e+00 6.56413659e-02 -2.90868580e-01
4.27820325e-01 7.81532109e-01 4.95565310e-02 4.98683423e-01
4.09430623e-01 -1.17938089e+00 8.59376371e-01 -3.32957298e-01
1.22239614e+00 -5.64517438e-01 6.50764406e-01 2.69341230e-01
-1.29093730e+00 2.48076975e-01 -9.51716006e-02 2.17392057e-01
5.31446576e-01 2.17619792e-01 -1.39884949e+00 5.83032072e-01
4.47684467e-01 7.86899745e-01 -2.76921224e-02 8.19834411e-01
-2.56936491e-01 1.24704587e+00 -3.47684860e-01 -8.69973749e-02
5.45264840e-01 1.79886371e-01 9.52836931e-01 1.48484492e+00
3.99810255e-01 -6.14852353e-04 -1.67471468e-02 5.40091991e-01
-3.42817307e-01 1.22581288e-01 -5.22739887e-01 -4.68896836e-01
8.96146715e-01 7.17634082e-01 -1.06569082e-01 -3.14900666e-01
-5.31917632e-01 1.01858091e+00 3.41492176e-01 5.76450765e-01
-4.52261508e-01 -3.92454833e-01 9.61438298e-01 -1.01224948e-02
3.58362824e-01 -5.13460636e-01 5.02927974e-02 -1.19961309e+00
6.91496655e-02 -1.24310517e+00 1.50849568e-02 -8.90619576e-01
-1.01237118e+00 1.49515498e+00 -2.32998952e-01 -1.11345637e+00
-8.37046385e-01 -6.17103815e-01 -4.41902846e-01 1.00782800e+00
-1.42498505e+00 -8.95466685e-01 2.68875241e-01 3.63232166e-01
1.19437039e+00 -6.74532294e-01 9.94244039e-01 4.47608769e-01
-6.73916221e-01 9.66315985e-01 1.93421349e-01 2.13436261e-01
4.95421261e-01 -1.06252027e+00 1.20561552e+00 1.09543800e+00
7.35742927e-01 5.00531018e-01 6.25512719e-01 -2.83256203e-01
-1.33498979e+00 -1.11237979e+00 1.20078039e+00 -1.68480769e-01
7.65219271e-01 -5.57998478e-01 -9.94356871e-01 9.24491048e-01
7.49325335e-01 -2.44630173e-01 4.00532693e-01 1.16460443e-01
-4.96900022e-01 -2.60800659e-03 -5.67498386e-01 6.61463797e-01
1.23572838e+00 -8.95482659e-01 -4.89027411e-01 2.04859585e-01
1.23960459e+00 -7.14832067e-01 -8.92106771e-01 3.76681179e-01
3.15038294e-01 -6.90441608e-01 6.16921425e-01 -3.54167163e-01
2.72300512e-01 -1.05585076e-01 -4.92736608e-01 -1.91347957e+00
-1.12289123e-01 -1.14782333e+00 -2.22771406e-01 1.12613082e+00
7.18946695e-01 -8.74373019e-01 3.64437282e-01 -9.43027735e-02
-9.66566801e-01 -8.41667712e-01 -1.45877433e+00 -1.07842028e+00
3.11401561e-02 -7.55612850e-01 8.40783000e-01 6.36722744e-01
-6.43090680e-02 6.83290184e-01 -3.27678412e-01 1.17899746e-01
1.74429715e-01 -3.32452804e-01 4.84796554e-01 -4.66935217e-01
-3.59681457e-01 -5.10240912e-01 -3.98019910e-01 -1.69979906e+00
3.64604622e-01 -1.25331616e+00 2.13957518e-01 -1.43376565e+00
-3.27902734e-01 -2.04158887e-01 -6.15908317e-02 6.18158996e-01
2.40949225e-02 -4.94272187e-02 1.28626496e-01 -4.76800799e-02
-2.28266150e-01 1.01867425e+00 1.16045630e+00 -2.17734188e-01
-1.09571412e-01 1.55538470e-01 -2.53694564e-01 8.89249295e-02
7.72782207e-01 -8.55891928e-02 -4.88865316e-01 -8.44732642e-01
-4.64615554e-01 3.35072547e-01 -9.75139216e-02 -9.53079581e-01
1.55572444e-01 3.37980092e-01 -2.66562402e-01 -6.81769967e-01
6.09279633e-01 -3.14617068e-01 -8.05177018e-02 2.20405370e-01
-5.67111611e-01 3.27789485e-01 2.73036867e-01 6.54793978e-02
-6.65280342e-01 7.29607344e-02 8.70175958e-01 2.16288209e-01
-8.58993709e-01 3.97834808e-01 -7.26265132e-01 2.12555796e-01
2.69355297e-01 -1.32299006e-01 -4.15322453e-01 -8.97213399e-01
-5.29490888e-01 -2.55404096e-02 -1.08659133e-01 9.82573867e-01
9.54689085e-01 -1.24839413e+00 -1.18370271e+00 5.46387196e-01
2.95747876e-01 -1.99565843e-01 2.92185936e-02 6.77937746e-01
-2.41375223e-01 9.04884696e-01 3.59431893e-01 -6.13473833e-01
-1.26870644e+00 7.77895749e-02 4.86306340e-01 1.29466839e-02
-6.11797571e-01 9.77123499e-01 1.94784328e-02 -7.50617921e-01
5.01330435e-01 -4.23632562e-01 3.50752324e-01 -5.30047536e-01
7.75156498e-01 1.99004918e-01 6.33864164e-01 -8.77969623e-01
-4.38122064e-01 -2.02488258e-01 -5.30222893e-01 -4.68274742e-01
1.13094497e+00 -1.88178718e-01 1.71318874e-01 4.05713767e-01
1.30377209e+00 -2.78595001e-01 -7.83788025e-01 -6.06008470e-01
-5.04562855e-02 -1.23537384e-01 3.52210820e-01 -9.75149632e-01
-8.73657048e-01 1.14494455e+00 3.55344504e-01 3.85760106e-02
9.69066441e-01 1.43063965e-03 1.32776368e+00 5.43006718e-01
4.92414564e-01 -7.42695630e-01 -2.10093230e-01 1.17931259e+00
1.22498524e+00 -9.49729264e-01 -7.50049651e-01 -2.96450257e-01
-5.72031379e-01 9.46226060e-01 3.88446033e-01 5.32322168e-01
5.76843441e-01 5.67339063e-01 3.23963732e-01 3.07940334e-01
-1.22168851e+00 -1.96880355e-01 1.85923591e-01 5.40513456e-01
9.19780791e-01 3.01185250e-01 6.22677291e-03 3.16888899e-01
-6.95515692e-01 -1.04754806e-01 3.10812086e-01 5.04875183e-01
-3.15588117e-01 -1.42350221e+00 -8.84312019e-02 4.00711238e-01
-2.90405065e-01 -6.09202266e-01 -4.94758606e-01 3.88781309e-01
-5.50654411e-01 1.19677174e+00 1.86202601e-01 -6.60642624e-01
3.22735935e-01 1.07044347e-01 3.26134026e-01 -7.28392303e-01
-5.68481386e-01 2.01929793e-01 6.33894861e-01 -4.86246645e-01
-1.43505470e-03 -4.91836816e-01 -1.07305038e+00 -4.69896823e-01
-4.26090777e-01 2.98585266e-01 7.55822062e-01 9.87536192e-01
4.18565959e-01 5.84511876e-01 7.43953347e-01 -4.72450048e-01
-8.68962169e-01 -1.47428119e+00 -3.33639145e-01 5.71547709e-02
5.96432388e-01 -4.12261754e-01 -1.97211415e-01 -1.65313631e-01] | [14.47826099395752, 6.956158638000488] |
9ce6ee86-d9b9-4c4f-9e68-efa365a40dd5 | an-image-is-worth-16x16-words-transformers-1 | 2010.11929 | null | https://arxiv.org/abs/2010.11929v2 | https://arxiv.org/pdf/2010.11929v2.pdf | An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale | While the Transformer architecture has become the de-facto standard for natural language processing tasks, its applications to computer vision remain limited. In vision, attention is either applied in conjunction with convolutional networks, or used to replace certain components of convolutional networks while keeping their overall structure in place. We show that this reliance on CNNs is not necessary and a pure transformer applied directly to sequences of image patches can perform very well on image classification tasks. When pre-trained on large amounts of data and transferred to multiple mid-sized or small image recognition benchmarks (ImageNet, CIFAR-100, VTAB, etc.), Vision Transformer (ViT) attains excellent results compared to state-of-the-art convolutional networks while requiring substantially fewer computational resources to train. | ['Neil Houlsby', 'Jakob Uszkoreit', 'Sylvain Gelly', 'Georg Heigold', 'Matthias Minderer', 'Mostafa Dehghani', 'Thomas Unterthiner', 'Xiaohua Zhai', 'Dirk Weissenborn', 'Alexander Kolesnikov', 'Lucas Beyer', 'Alexey Dosovitskiy'] | 2020-10-22 | an-image-is-worth-16x16-words-transformers | https://openreview.net/forum?id=YicbFdNTTy | https://openreview.net/pdf?id=YicbFdNTTy | iclr-2021-1 | ['document-image-classification'] | ['computer-vision'] | [ 3.18294197e-01 -1.40647992e-01 5.97956143e-02 -2.50711054e-01
-3.29235792e-01 -4.66240525e-01 7.23752737e-01 -2.70823568e-01
-7.77436078e-01 4.40601468e-01 -2.65250951e-01 -6.72875702e-01
3.14485699e-01 -7.76778162e-01 -7.09852099e-01 -5.83785295e-01
3.24158132e-01 2.87286639e-01 5.70572019e-01 -2.67358452e-01
1.14024365e-02 5.73645115e-01 -1.43023849e+00 4.29435968e-01
2.82643378e-01 1.22696161e+00 2.25829825e-01 8.24831247e-01
-1.92174643e-01 1.17489827e+00 -4.44227695e-01 -5.13505995e-01
3.31287175e-01 -1.02696575e-01 -9.93438542e-01 1.75076380e-01
7.86102414e-01 -2.73772150e-01 -6.61753774e-01 1.03846407e+00
1.89182177e-01 -1.89160183e-01 3.66324872e-01 -8.67468238e-01
-8.80240858e-01 3.17657173e-01 -4.03347790e-01 5.31112492e-01
-1.06824911e-03 1.93705261e-01 8.83955717e-01 -1.01807976e+00
3.39160711e-01 1.13698208e+00 8.62108767e-01 3.89436275e-01
-1.39339602e+00 -4.04384822e-01 -4.55275364e-02 2.51125515e-01
-1.12367964e+00 -5.19350827e-01 4.59141344e-01 -2.59344518e-01
1.50115907e+00 -1.24577530e-01 7.31738687e-01 1.08182859e+00
3.96350443e-01 8.61169517e-01 8.18315744e-01 -4.71842557e-01
-3.70412357e-02 -7.11754858e-02 2.08393440e-01 9.80676293e-01
8.59846324e-02 6.96557909e-02 -2.70995021e-01 6.57467544e-02
9.26935971e-01 9.13867876e-02 -2.40712643e-01 -1.36213243e-01
-1.31407583e+00 8.05352509e-01 7.36003041e-01 6.55423343e-01
-5.36253214e-01 3.27669740e-01 5.93674004e-01 7.55342960e-01
3.53956163e-01 5.34453571e-01 -5.43790162e-01 -6.68464170e-04
-8.91378284e-01 1.88353702e-01 4.91105020e-01 8.20156872e-01
8.96578610e-01 4.71578985e-01 -1.74567308e-02 8.58022749e-01
-1.69116527e-01 1.69914246e-01 7.07480073e-01 -7.64169276e-01
2.33948201e-01 8.51192772e-01 -2.95664579e-01 -7.05977976e-01
-2.58673698e-01 -5.73734283e-01 -1.12958229e+00 3.11990023e-01
5.56876242e-01 7.07228184e-02 -1.32901657e+00 1.44648039e+00
-2.31610432e-01 1.73483521e-01 2.21961513e-01 5.12584448e-01
8.24077308e-01 5.64432204e-01 2.04631910e-02 2.01275855e-01
1.27971721e+00 -1.06699336e+00 -2.08530650e-01 -7.64474452e-01
4.48511839e-01 -8.30579817e-01 1.06637859e+00 5.78815699e-01
-1.16851151e+00 -8.61711860e-01 -1.03686035e+00 -2.83097833e-01
-5.61818182e-01 6.31245822e-02 5.68307459e-01 4.49222445e-01
-1.50668025e+00 4.57788140e-01 -7.47061849e-01 -4.49381322e-01
7.42900789e-01 5.48910797e-01 -5.27994037e-01 -2.75453150e-01
-8.25987220e-01 1.10767436e+00 3.44045371e-01 1.82295382e-01
-1.15045071e+00 -6.24567270e-01 -1.05726302e+00 1.87923893e-01
1.01083137e-01 -5.96953988e-01 1.65439546e+00 -1.50350022e+00
-1.21456063e+00 1.00855315e+00 -2.98369247e-02 -9.76227880e-01
4.25311565e-01 -8.19209069e-02 -2.46279314e-01 3.51880282e-01
-8.44017714e-02 1.02767920e+00 1.07865143e+00 -7.66023159e-01
-6.96792543e-01 -7.64648765e-02 3.28070104e-01 -2.27818251e-01
-4.29964811e-01 2.49529362e-01 -4.06683445e-01 -6.17450535e-01
-6.34405464e-02 -8.57981741e-01 -2.11765632e-01 1.34610370e-01
-1.81850553e-01 -3.77818644e-01 1.24452817e+00 -4.05348510e-01
5.45346916e-01 -2.10438991e+00 2.89194603e-02 -1.20925635e-01
3.28204334e-01 8.98450971e-01 -4.37216967e-01 1.21395305e-01
-2.66124099e-01 -3.71711748e-03 -3.04562956e-01 -3.05849135e-01
-3.57795328e-01 3.89844775e-01 -2.64288098e-01 4.66676325e-01
4.24217433e-01 1.29059660e+00 -7.63804972e-01 -2.02495530e-01
5.43241084e-01 6.04716182e-01 -3.33384186e-01 -1.24848813e-01
-8.57362822e-02 4.44343314e-03 -2.44838744e-01 4.68783677e-01
3.40533733e-01 -5.66019714e-01 9.66778547e-02 -3.48782599e-01
-3.61667387e-02 4.42378879e-01 -5.20618558e-01 1.53859961e+00
-4.89030898e-01 1.16452146e+00 -1.20980423e-02 -1.56874967e+00
7.38392770e-01 5.62876701e-01 3.81672651e-01 -9.31305408e-01
3.26934129e-01 1.94385590e-03 3.85058045e-01 -3.45247567e-01
3.74081731e-01 -1.09239809e-01 1.28013358e-01 3.06591928e-01
5.26275516e-01 -1.16835348e-02 3.21286261e-01 -1.71574317e-02
1.35827720e+00 -2.03962460e-01 3.47998142e-01 -3.66234213e-01
5.84756911e-01 3.33373845e-01 2.44601682e-01 6.24562919e-01
-2.50769228e-01 6.16362929e-01 3.48067194e-01 -9.53510106e-01
-1.40302038e+00 -9.84545112e-01 -8.85873958e-02 9.23681319e-01
-2.67592847e-01 -2.47660235e-01 -7.56109118e-01 -6.23741269e-01
-1.41496390e-01 2.70041585e-01 -7.99506664e-01 -2.95142502e-01
-5.90978324e-01 -6.75583065e-01 9.68059897e-01 7.76593029e-01
1.07295847e+00 -1.37118113e+00 -7.85570741e-01 2.98628479e-01
1.31819084e-01 -1.27494824e+00 -2.58119166e-01 4.76546943e-01
-6.50308788e-01 -1.28161168e+00 -7.94096291e-01 -1.19816732e+00
5.50190389e-01 5.43488324e-01 1.38657486e+00 2.96979070e-01
-3.53586286e-01 4.22409564e-01 -1.89643264e-01 -3.56574386e-01
-2.49809563e-01 1.82523653e-01 -2.45936677e-01 -8.04378018e-02
4.40590203e-01 -4.85011250e-01 -4.12568718e-01 1.20027311e-01
-9.04599190e-01 -4.92757075e-02 6.26070321e-01 1.23334920e+00
2.76797146e-01 -2.77092587e-02 4.30091947e-01 -9.41958308e-01
4.62798685e-01 3.37305330e-02 -5.90453029e-01 2.14396745e-01
-4.58386809e-01 -1.86833426e-01 8.16972017e-01 -5.58063507e-01
-7.76140451e-01 1.71885952e-01 -2.25949511e-01 -6.96357608e-01
-1.65313184e-01 6.06840253e-01 1.74813017e-01 -4.43548083e-01
7.80308187e-01 4.97096300e-01 4.86183427e-02 -1.27009436e-01
5.00206314e-02 4.03201103e-01 8.06646705e-01 -1.42050982e-01
7.58340657e-01 3.45136017e-01 -2.60862321e-01 -1.01731074e+00
-8.30949426e-01 -2.75150090e-01 -7.95576751e-01 2.22571582e-01
9.46345866e-01 -9.22725379e-01 -4.14607197e-01 6.75140083e-01
-1.19748747e+00 -5.20118833e-01 -1.93644345e-01 1.48764685e-01
-3.11881721e-01 1.26676545e-01 -7.29673147e-01 -1.49263427e-01
-3.58203411e-01 -1.21510720e+00 6.98721051e-01 7.31310174e-02
6.82274103e-02 -1.06103730e+00 -2.50351727e-01 1.28588542e-01
6.27198279e-01 -1.19647928e-01 8.05638850e-01 -5.01287937e-01
-3.90656263e-01 -4.17010397e-01 -5.12988031e-01 9.22269464e-01
2.71317542e-01 2.71396358e-02 -1.18336630e+00 -2.91795254e-01
-1.93921223e-01 -6.31373346e-01 1.16996622e+00 4.30941463e-01
1.34845674e+00 -1.04520537e-01 -9.42838863e-02 4.81468052e-01
1.43510818e+00 2.38268211e-01 9.26600397e-01 3.68433058e-01
5.93533099e-01 1.91649109e-01 -1.12123281e-01 -1.17247425e-01
2.40342930e-01 3.92628580e-01 4.50143486e-01 -4.45446581e-01
-1.01434760e-01 6.37497380e-02 2.29717821e-01 4.69971776e-01
2.27675159e-02 -2.31343985e-01 -1.09656346e+00 8.29047740e-01
-1.72640562e+00 -1.08412743e+00 4.64413948e-02 1.85246813e+00
6.07926965e-01 5.04264712e-01 -1.63504153e-01 4.03369963e-01
4.16635931e-01 1.56275526e-01 -4.37816679e-01 -4.49861199e-01
-2.18278646e-01 6.95925832e-01 5.70810199e-01 2.28494614e-01
-1.21048665e+00 1.13807154e+00 7.60160494e+00 5.45935273e-01
-1.64142597e+00 5.05110472e-02 6.92581892e-01 1.71103999e-01
8.33503082e-02 -2.46546641e-01 -3.56006116e-01 -4.03483175e-02
9.21064138e-01 -4.37709950e-02 5.45613170e-01 8.39830041e-01
-2.02057511e-01 9.44094434e-02 -1.26006055e+00 1.13365233e+00
-5.86650297e-02 -1.68201864e+00 6.15743995e-02 -4.16118875e-02
5.89759707e-01 6.03154242e-01 1.49015173e-01 4.38124537e-01
5.77318728e-01 -1.39103198e+00 6.30680501e-01 9.42988973e-03
7.73790717e-01 -5.06678402e-01 8.93149316e-01 2.40246728e-01
-1.18154156e+00 -2.47766520e-03 -7.36513734e-01 -1.86361492e-01
-3.68750423e-01 3.72962117e-01 -7.65312076e-01 1.93791226e-01
1.00866306e+00 9.13553953e-01 -7.22020209e-01 8.10872257e-01
-1.49280891e-01 7.96427846e-01 -1.67714760e-01 1.14970855e-01
8.32289577e-01 3.12334150e-01 -3.80305853e-03 1.18265593e+00
-2.28231978e-02 -1.30531341e-01 1.24690071e-01 6.55598342e-01
-3.58604789e-01 -3.04907113e-01 -8.06359589e-01 -2.54319191e-01
1.39504254e-01 1.36329758e+00 -6.45955026e-01 -7.65439272e-01
-9.03717935e-01 9.60341632e-01 5.04006982e-01 5.07454991e-01
-6.09573483e-01 -3.34677577e-01 7.75724769e-01 -7.86621496e-02
7.69899249e-01 -2.82454401e-01 -1.89471245e-01 -1.22003233e+00
-1.96697161e-01 -1.00018108e+00 1.85333028e-01 -8.55483413e-01
-1.15873921e+00 1.04171002e+00 -4.00013685e-01 -9.95658815e-01
-2.88955837e-01 -1.16780424e+00 -7.19070017e-01 7.96524882e-01
-1.69709432e+00 -1.11123502e+00 -4.19052631e-01 1.05397415e+00
6.64546907e-01 -4.38394666e-01 9.72607791e-01 4.18300807e-01
-4.94095445e-01 5.95347881e-01 -1.93389848e-01 6.32138610e-01
3.10252398e-01 -1.08126855e+00 6.73709810e-01 9.31502044e-01
4.44081008e-01 5.52786350e-01 4.28267360e-01 1.43235791e-02
-1.38637638e+00 -1.26079905e+00 9.54623878e-01 -4.07397002e-02
6.41329467e-01 -2.85442293e-01 -1.00288022e+00 1.08373702e+00
7.58648217e-01 5.47059000e-01 1.69001192e-01 -1.15779899e-01
-7.01888621e-01 -2.03608587e-01 -9.70833838e-01 5.23289561e-01
8.02642941e-01 -8.73283446e-01 -7.20516860e-01 1.94078073e-01
4.74619716e-01 -3.15006822e-01 -6.55123234e-01 2.66683519e-01
4.15389866e-01 -8.75585556e-01 1.03465950e+00 -7.23560870e-01
5.64696550e-01 -1.90685347e-01 -1.91347867e-01 -1.24524570e+00
-5.41099489e-01 -3.55709642e-01 4.62787807e-01 8.59280884e-01
5.16463518e-01 -6.88221514e-01 7.79296756e-01 2.48344004e-01
-3.39289904e-01 -5.53807974e-01 -8.57994437e-01 -6.19153321e-01
1.46713570e-01 -4.75593269e-01 2.83368319e-01 9.05908883e-01
-3.61105680e-01 6.07035160e-01 -2.46166408e-01 -7.18410015e-02
3.47795784e-01 -1.28610447e-01 7.26388931e-01 -1.19350982e+00
2.33219732e-02 -6.77979767e-01 -6.26714289e-01 -9.42337334e-01
3.25447023e-01 -7.96694040e-01 -6.48606988e-03 -1.48205388e+00
-9.63613093e-02 -3.28755349e-01 -5.11764944e-01 1.00704598e+00
2.03307092e-01 8.09112251e-01 1.21618040e-01 4.06940803e-02
-3.22190583e-01 4.44052577e-01 1.15839159e+00 -6.43956900e-01
3.12318236e-01 -7.13645220e-02 -6.05129540e-01 8.37671518e-01
7.92570829e-01 -2.72389412e-01 -5.70644617e-01 -8.08510482e-01
5.98413609e-02 -3.08068842e-01 6.97380185e-01 -1.20563626e+00
3.52308154e-01 8.88856724e-02 5.26029885e-01 -4.06989098e-01
3.10606986e-01 -9.43212688e-01 -1.11701177e-03 6.48939312e-01
-2.76979864e-01 3.91687572e-01 5.08774221e-01 5.71372248e-02
-6.97873890e-01 -1.66925266e-01 1.07257104e+00 -5.84802628e-01
-9.26846802e-01 4.08411056e-01 -7.33483434e-01 -1.01459168e-01
9.02130127e-01 -3.53405029e-01 -3.28072995e-01 -6.98933452e-02
-4.25017118e-01 -7.71346539e-02 3.52980793e-01 3.77326459e-01
8.48055720e-01 -1.10705352e+00 -5.91040075e-01 2.91109711e-01
4.07791249e-02 1.70301460e-02 -1.04681373e-01 5.99201202e-01
-7.70375490e-01 6.86062515e-01 -5.55078626e-01 -8.09614718e-01
-1.22124147e+00 5.82798541e-01 7.10962117e-01 -3.19342047e-01
-8.94204617e-01 9.43489611e-01 3.56681257e-01 -2.98767298e-01
1.94668561e-01 -5.62860012e-01 -1.60077885e-02 -2.29615629e-01
6.30020261e-01 -2.66805232e-01 3.37380379e-01 -4.04818684e-01
-4.07517105e-01 3.84859562e-01 -4.46995050e-01 5.19454591e-02
1.41489637e+00 2.67414570e-01 -8.31926689e-02 1.83304012e-01
1.13079202e+00 -5.54481804e-01 -1.21399570e+00 -5.85054994e-01
-5.18014431e-02 -6.91651776e-02 3.95019889e-01 -4.93413329e-01
-1.57977581e+00 1.13467443e+00 3.94092917e-01 2.79355586e-01
1.28631747e+00 -1.52253196e-01 7.22827613e-01 9.09121215e-01
3.47291619e-01 -6.65879488e-01 2.32812628e-01 8.75787854e-01
8.16185057e-01 -1.25197601e+00 -2.76359856e-01 -7.89306238e-02
-4.87586975e-01 1.22210479e+00 6.74426973e-01 -6.46049559e-01
8.86410534e-01 5.00734687e-01 1.59388855e-01 -1.59082308e-01
-1.05945158e+00 -2.55958349e-01 2.30926260e-01 6.45404458e-01
5.23927212e-01 -3.29539418e-01 1.81933239e-01 -7.75544420e-02
-1.35748819e-01 1.96494132e-01 4.51406062e-01 9.36904728e-01
-4.86818463e-01 -9.54901636e-01 -5.95414005e-02 6.53397501e-01
-6.71601593e-01 -4.77691144e-01 -3.71267468e-01 9.17801797e-01
1.13119461e-01 7.48592377e-01 3.60663295e-01 -4.69993293e-01
1.93896368e-01 -2.55177869e-03 6.02307975e-01 -6.22455716e-01
-1.04390311e+00 -9.48663801e-02 1.22521697e-02 -4.70433205e-01
-6.67883694e-01 -4.40639228e-01 -9.74625349e-01 -6.08008981e-01
-3.86759900e-02 -1.41110256e-01 3.19218367e-01 1.20851898e+00
6.20849840e-02 7.92837739e-01 1.40871927e-01 -7.26153791e-01
-3.13257277e-01 -1.14370787e+00 -2.75995970e-01 1.10092275e-01
6.68954492e-01 -5.52172065e-01 5.41047677e-02 3.41482222e-01] | [9.390957832336426, 1.7169214487075806] |
fa2890df-11ce-4dd2-a145-dc2fafe8f359 | a-multimodal-framework-for-video-ads | 2108.12868 | null | https://arxiv.org/abs/2108.12868v1 | https://arxiv.org/pdf/2108.12868v1.pdf | A Multimodal Framework for Video Ads Understanding | There is a growing trend in placing video advertisements on social platforms for online marketing, which demands automatic approaches to understand the contents of advertisements effectively. Taking the 2021 TAAC competition as an opportunity, we developed a multimodal system to improve the ability of structured analysis of advertising video content. In our framework, we break down the video structuring analysis problem into two tasks, i.e., scene segmentation and multi-modal tagging. In scene segmentation, we build upon a temporal convolution module for temporal modeling to predict whether adjacent frames belong to the same scene. In multi-modal tagging, we first compute clip-level visual features by aggregating frame-level features with NeXt-SoftDBoF. The visual features are further complemented with textual features that are derived using a global-local attention mechanism to extract useful information from OCR (Optical Character Recognition) and ASR (Audio Speech Recognition) outputs. Our solution achieved a score of 0.2470 measured in consideration of localization and prediction accuracy, ranking fourth in the 2021 TAAC final leaderboard. | ['Yu-Gang Jiang', 'Zuxuan Wu', 'Rui Wang', 'Lingchen Meng', 'Zejia Weng'] | 2021-08-29 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 5.32959282e-01 -9.42679718e-02 -1.97089717e-01 -6.09862328e-01
-1.16978419e+00 -6.88059092e-01 6.37198269e-01 1.90036878e-01
-5.05092919e-01 9.33095813e-02 3.08708489e-01 -2.12490946e-01
1.16002165e-01 -3.68710756e-01 -7.05135584e-01 -4.76673365e-01
-1.28631249e-01 -1.44080166e-03 4.22918290e-01 1.54327052e-02
2.77613789e-01 3.98405105e-01 -1.95112514e+00 1.20208573e+00
4.64542091e-01 1.64856017e+00 2.92727053e-01 9.32900131e-01
-3.53000641e-01 1.10313094e+00 -3.62819105e-01 -4.61489260e-01
-9.76541825e-03 -1.88931093e-01 -6.53052032e-01 5.76145172e-01
4.99820411e-01 -1.58769846e-01 -2.63009936e-01 9.48878765e-01
1.27960145e-01 1.18067972e-02 5.89478165e-02 -1.12269235e+00
-1.10524967e-01 6.58266842e-01 -4.69657898e-01 3.61966759e-01
5.28486550e-01 -7.47289211e-02 1.22221363e+00 -9.75186884e-01
8.33920300e-01 1.24328732e+00 3.09077203e-01 1.63245201e-01
-9.67513621e-01 -1.65648997e-01 4.87862170e-01 5.08490086e-01
-1.15793777e+00 -6.97148860e-01 7.55635917e-01 -6.11923516e-01
5.84575355e-01 5.30107737e-01 6.38018548e-01 8.67824197e-01
2.36224338e-01 1.12898397e+00 8.43066216e-01 -2.52836227e-01
1.27862215e-01 1.97851375e-01 7.12361038e-02 5.20720422e-01
-5.32929957e-01 -3.53458375e-01 -7.60476530e-01 1.79854393e-01
4.17042315e-01 -1.67015254e-01 3.27377677e-01 -1.34387508e-01
-1.08652866e+00 7.88736284e-01 1.26538500e-01 3.29151303e-01
-5.16598046e-01 2.15730920e-01 4.81080353e-01 1.48742571e-01
5.44102132e-01 1.61791816e-01 -3.05092037e-01 -2.24623084e-01
-1.09610915e+00 1.28109962e-01 3.17197144e-01 7.30665386e-01
4.97783512e-01 -1.48250148e-01 -2.71532327e-01 9.06621695e-01
4.06510115e-01 3.42787027e-01 3.72001529e-01 -1.11019886e+00
3.30605119e-01 5.36916256e-01 2.81648319e-02 -1.17040741e+00
-3.83085966e-01 -3.18196684e-01 -1.10151254e-01 -1.49411425e-01
2.60791063e-01 -2.87372265e-02 -1.01075137e+00 1.34261382e+00
2.53578305e-01 7.55003095e-02 -2.73432225e-01 9.75758553e-01
7.57542014e-01 9.94178176e-01 2.42035970e-01 -3.04601341e-01
1.69494307e+00 -9.54816580e-01 -7.87463903e-01 -1.54265940e-01
5.57375491e-01 -1.14735186e+00 7.09000230e-01 7.80138612e-01
-1.29734862e+00 -8.33504140e-01 -8.82754028e-01 -1.23796493e-01
-5.32129169e-01 3.84809971e-01 4.71134245e-01 5.99520564e-01
-9.85366225e-01 7.62418807e-02 -5.66060781e-01 -2.57800490e-01
1.94813848e-01 1.77282080e-01 -1.77808002e-01 -4.58063819e-02
-1.14181960e+00 4.76986229e-01 2.64361292e-01 2.52750486e-01
-9.96545196e-01 -2.40842760e-01 -8.98424089e-01 -1.27331734e-01
6.12042069e-01 1.50084898e-01 1.39174581e+00 -1.37697923e+00
-1.44304097e+00 8.09251606e-01 -1.93205133e-01 -6.82125151e-01
2.71359921e-01 -4.22581792e-01 -8.98236752e-01 5.63565135e-01
7.46405125e-02 9.48277593e-01 1.19963372e+00 -1.21617532e+00
-1.32573330e+00 -2.27928430e-01 7.47572482e-02 2.40154266e-01
-3.86697531e-01 5.08490384e-01 -1.18385363e+00 -4.89231050e-01
2.17149362e-01 -8.80057156e-01 -2.09964722e-01 -2.23997474e-01
-2.59397894e-01 -5.50758541e-02 1.09583402e+00 -1.00953007e+00
1.35478044e+00 -2.31981254e+00 1.32254496e-01 3.11671853e-01
3.40437293e-02 1.08000055e-01 -2.89100796e-01 5.79147153e-02
-6.94174170e-02 -1.42455295e-01 3.23402256e-01 -2.75703400e-01
-1.69187300e-02 -1.68113455e-01 -3.31413925e-01 2.06809297e-01
2.71900564e-01 8.36884677e-01 -6.30678713e-01 -7.43490696e-01
3.27712655e-01 4.10783589e-01 -5.77060580e-01 -2.16725156e-01
-5.82907856e-01 1.90854535e-01 -4.97784406e-01 8.03591251e-01
5.34039915e-01 -1.17364831e-01 2.27331340e-01 -2.61776805e-01
-4.23325658e-01 1.95016250e-01 -1.11250055e+00 1.64620280e+00
-3.70422006e-01 9.22516286e-01 3.45560461e-01 -9.07597065e-01
6.47692263e-01 2.57836342e-01 1.01421654e+00 -1.02625120e+00
2.20732704e-01 -5.63319959e-02 -2.25861788e-01 -5.18599331e-01
9.43738997e-01 5.09245157e-01 -4.98729557e-01 3.49195860e-02
-1.21488022e-02 2.68082023e-01 3.20780277e-01 2.37011462e-01
9.67514336e-01 2.24946767e-01 -3.65016252e-01 9.17632952e-02
8.47249031e-01 1.32140264e-01 3.72656882e-01 5.07343352e-01
-3.68816733e-01 5.58912694e-01 5.81725478e-01 -5.16505837e-01
-1.04897082e+00 -7.98799217e-01 -5.78152109e-03 1.48583817e+00
1.74735963e-01 -6.89648628e-01 -9.86235559e-01 -5.78046739e-01
-2.22514763e-01 5.16777813e-01 -3.99091214e-01 1.92764431e-01
-5.72600126e-01 -3.87557864e-01 1.65892288e-01 2.23720446e-01
3.48320752e-01 -9.66555476e-01 -4.22644198e-01 3.69965017e-01
-3.62630218e-01 -1.73578346e+00 -5.08384109e-01 6.85720593e-02
-3.80016297e-01 -6.84955537e-01 -5.73911369e-01 -8.92389238e-01
1.88034952e-01 3.91197354e-01 7.36949265e-01 -3.70226324e-01
-4.33923781e-01 5.17819822e-01 -4.73823249e-01 -1.57334413e-02
-1.53948188e-01 9.58586950e-03 -2.01214910e-01 7.00145781e-01
4.62742567e-01 9.64589044e-02 -4.94938046e-01 5.02198577e-01
-9.17425632e-01 1.58407465e-01 6.69175923e-01 2.21348599e-01
7.14767694e-01 1.67046860e-01 1.94222540e-01 -5.05585670e-01
3.77993435e-01 -2.61485517e-01 -8.37023318e-01 1.45243183e-01
2.16895685e-04 -4.02271271e-01 3.02543432e-01 -2.96651214e-01
-1.01226437e+00 6.06189668e-01 -9.66869369e-02 -3.75082076e-01
-1.69080615e-01 6.72335804e-01 -2.00138092e-01 9.12680626e-02
1.63453683e-01 3.35499823e-01 -1.49637118e-01 -4.16882932e-01
5.30947864e-01 8.97048652e-01 5.79616189e-01 -2.16986641e-01
4.66759950e-01 4.43312913e-01 -3.58475149e-01 -1.23109734e+00
-1.06763256e+00 -7.11189985e-01 -5.31929970e-01 -9.20415103e-01
1.53512037e+00 -1.06280959e+00 -6.96115911e-01 4.21041012e-01
-9.86752033e-01 -3.17379124e-02 3.88532467e-02 4.94630486e-01
-5.49660563e-01 3.65872175e-01 -5.70621908e-01 -8.73802602e-01
9.61720794e-02 -1.38255084e+00 1.30891848e+00 9.69395712e-02
5.30783844e-04 -3.99279833e-01 -3.98171037e-01 9.23764408e-01
6.43772781e-02 -1.82159878e-02 5.16324878e-01 -6.01808012e-01
-8.18140447e-01 -4.00914222e-01 -2.91415930e-01 4.25731897e-01
-4.42193866e-01 1.64899752e-01 -1.08965087e+00 4.76151370e-02
-2.46276408e-01 -1.02581978e-01 8.47844958e-01 7.80088842e-01
1.52496088e+00 2.89200787e-02 -2.49843404e-01 2.90787905e-01
1.03669012e+00 5.99576175e-01 9.05623674e-01 3.35239470e-01
5.78361988e-01 7.98274040e-01 1.06742251e+00 4.44473922e-01
2.83524424e-01 1.09022844e+00 4.90811586e-01 -7.08810464e-02
-5.48740886e-02 5.85725065e-03 7.98632622e-01 6.24057293e-01
1.54855058e-01 -2.73374259e-01 -7.29195654e-01 4.24530834e-01
-1.89718378e+00 -1.08803940e+00 -3.10897529e-01 1.97626591e+00
3.15469593e-01 4.98999119e-01 3.93403262e-01 -1.06937118e-01
1.10003340e+00 2.49958873e-01 -1.66478217e-01 -6.01398230e-01
-2.00538710e-01 -1.94191054e-01 5.33137262e-01 4.79201704e-01
-1.66469073e+00 1.02006900e+00 6.19403315e+00 1.07546711e+00
-1.05179954e+00 2.94183847e-02 9.64123726e-01 -2.37901181e-01
-1.51730433e-01 -1.82846844e-01 -8.82551074e-01 6.38205349e-01
1.29696250e+00 4.63002622e-01 3.21804851e-01 8.59898210e-01
3.98393869e-01 -1.78255334e-01 -6.83228195e-01 9.42045093e-01
3.22934598e-01 -1.46426594e+00 -8.48995075e-02 2.96140194e-01
6.03805900e-01 -1.90756485e-01 3.06700408e-01 3.04931760e-01
-4.30896021e-02 -6.91406310e-01 1.10319936e+00 4.66358542e-01
3.68280947e-01 -7.47733057e-01 5.17911732e-01 -1.80309474e-01
-1.42751384e+00 -3.46082449e-01 -2.23758817e-03 2.70605445e-01
5.14969528e-01 4.56792414e-01 -5.80670714e-01 3.73294681e-01
6.83200955e-01 6.70867622e-01 -6.41427219e-01 9.32165205e-01
1.20603338e-01 3.27889591e-01 -6.92975447e-02 1.02107050e-02
5.16378820e-01 -1.16700172e-01 4.55702841e-01 1.17135394e+00
2.68115193e-01 -2.11760074e-01 3.88811648e-01 1.98023349e-01
7.35146254e-02 2.21872956e-01 -1.49763033e-01 -3.57939780e-01
5.07740155e-02 1.57841349e+00 -1.00863886e+00 -3.71189177e-01
-6.52249157e-01 9.86031830e-01 -3.96064132e-01 1.89599976e-01
-1.34721971e+00 -1.91508844e-01 4.95947003e-01 2.52148330e-01
6.73567295e-01 -3.65324855e-01 -5.34601845e-02 -9.23450530e-01
1.67528577e-02 -8.95318389e-01 3.49568069e-01 -1.01593554e+00
-6.36929393e-01 7.28657246e-01 -2.35246554e-01 -1.23160398e+00
-9.17939767e-02 -8.40310276e-01 -2.72131860e-02 2.76062757e-01
-1.23418355e+00 -1.27529848e+00 -1.76116526e-01 5.27325392e-01
9.79723036e-01 -1.75783917e-01 2.81384617e-01 7.33488500e-01
-5.18330395e-01 3.16006422e-01 1.45084634e-02 -1.35588413e-02
5.29136300e-01 -8.59441102e-01 2.05339670e-01 9.63223279e-01
2.29912832e-01 8.23951066e-02 5.61844409e-01 -5.89347243e-01
-1.53272867e+00 -1.15053761e+00 1.08060777e+00 -2.08636552e-01
1.01636028e+00 -6.82770014e-01 -3.51278454e-01 4.89019245e-01
1.21510282e-01 -1.30451053e-01 5.43584585e-01 -1.46396384e-01
-3.75416763e-02 -2.04639718e-01 -6.77457988e-01 4.03361410e-01
6.17656648e-01 -7.73283303e-01 -7.77304322e-02 3.22514951e-01
7.17176318e-01 -1.98521167e-01 -8.72671425e-01 2.56967116e-02
7.15650678e-01 -6.67947829e-01 1.02587605e+00 -4.06653643e-01
5.92657506e-01 -3.57415468e-01 -5.43062866e-01 -4.83753949e-01
-1.30935580e-01 -6.38941348e-01 1.65391922e-01 1.25556171e+00
5.49596667e-01 3.88985500e-02 7.98498213e-01 3.77053678e-01
-2.93374866e-01 -3.43853623e-01 -9.04237211e-01 -3.74625653e-01
-7.39513457e-01 -9.26429152e-01 1.32879540e-01 5.06797552e-01
-1.32103011e-01 3.35950643e-01 -5.43895066e-01 2.04076275e-01
3.89132470e-01 1.75975725e-01 5.26446223e-01 -1.22206104e+00
-1.13675423e-01 -2.99774915e-01 -6.06595814e-01 -1.24091339e+00
-9.66446847e-03 -5.99425972e-01 8.26124251e-02 -1.23239410e+00
1.49160832e-01 3.56299840e-02 -2.93188334e-01 1.88455150e-01
3.91499639e-01 6.30574644e-01 3.71083677e-01 5.50135113e-02
-1.17713571e+00 1.21148378e-01 1.11432278e+00 -1.94303572e-01
-1.39930397e-01 2.96654161e-02 -4.28359479e-01 6.31646216e-01
4.96344715e-01 -2.05759285e-03 -1.51340991e-01 -2.65055448e-01
1.65049449e-01 2.10587755e-01 3.20964158e-01 -1.10466492e+00
1.13644414e-01 -1.25104219e-01 4.57683772e-01 -1.04459023e+00
7.93405235e-01 -8.99027109e-01 7.27431965e-04 3.03631365e-01
-7.30830312e-01 -2.28939466e-02 1.96049720e-01 4.66517568e-01
-4.93672937e-01 -8.62065256e-02 5.62909245e-01 1.67791732e-02
-1.28369009e+00 7.70950392e-02 -9.70085144e-01 -4.94989038e-01
1.12153876e+00 -2.23900527e-01 -2.02658206e-01 -5.84387302e-01
-1.27733350e+00 1.58687994e-01 1.01356760e-01 8.07163894e-01
5.20727456e-01 -1.31537223e+00 -4.60985214e-01 2.84173667e-01
9.60968509e-02 -6.58470750e-01 5.00912070e-01 8.38801086e-01
-4.90431637e-01 7.51401067e-01 -1.15295835e-01 -8.43073487e-01
-1.45258737e+00 5.11546254e-01 -1.98495425e-02 -9.03432965e-02
-2.81426489e-01 7.37047315e-01 8.59706476e-02 1.63908362e-01
4.11748320e-01 -2.91433722e-01 -4.58308756e-01 6.02348924e-01
7.34232366e-01 6.37104064e-02 2.48909190e-01 -1.12359667e+00
-3.66723776e-01 3.76103163e-01 -2.99004436e-01 -4.11893606e-01
1.30609536e+00 -4.81293082e-01 1.48658618e-01 2.55443364e-01
1.20067453e+00 -3.19565386e-02 -1.28513670e+00 -1.11145943e-01
2.55816042e-01 -3.45179796e-01 5.02840161e-01 -7.05602407e-01
-1.04842806e+00 8.49440694e-01 8.12567651e-01 6.90270901e-01
1.23646069e+00 5.75058982e-02 8.67516160e-01 -2.30375826e-02
-7.49655515e-02 -1.52050269e+00 2.62737811e-01 4.94347990e-01
6.80879295e-01 -9.86062884e-01 -1.56896666e-01 -4.87210065e-01
-8.31436038e-01 1.13466978e+00 3.82323861e-01 1.30825251e-01
4.12308276e-01 2.18762577e-01 6.52321577e-02 -2.30577052e-01
-7.26685166e-01 -4.54577178e-01 6.30174458e-01 2.76946574e-01
3.61507207e-01 -8.54330584e-02 -8.45695063e-02 5.20498693e-01
1.11250199e-01 -2.65736997e-01 2.87355304e-01 7.37740934e-01
-9.47023511e-01 -9.97350991e-01 -3.72185111e-01 1.40732184e-01
-5.89795411e-01 -1.91382151e-02 -3.83286387e-01 5.20740092e-01
2.85880994e-02 1.16971731e+00 3.72162908e-01 -6.71897471e-01
1.80750579e-01 9.41495076e-02 2.22793490e-01 -3.17331642e-01
-5.38929224e-01 9.24574971e-01 4.14495438e-01 -8.92224491e-01
-4.98112023e-01 -1.06687248e+00 -8.65200520e-01 5.14995232e-02
-5.29218987e-02 2.09460724e-02 1.11856306e+00 7.75391400e-01
3.59345019e-01 6.86857998e-01 7.25827754e-01 -9.78931904e-01
5.63518517e-02 -5.75336218e-01 -4.21257526e-01 2.78167844e-01
2.03211069e-01 -4.09136474e-01 -2.64530699e-03 4.63639975e-01] | [9.998601913452148, 0.6448584198951721] |
f8f1acd0-420e-4a76-981a-7e92d8f81857 | relational-message-passing-for-fully | 2210.03994 | null | https://arxiv.org/abs/2210.03994v2 | https://arxiv.org/pdf/2210.03994v2.pdf | Relational Message Passing for Fully Inductive Knowledge Graph Completion | In knowledge graph completion (KGC), predicting triples involving emerging entities and/or relations, which are unseen when the KG embeddings are learned, has become a critical challenge. Subgraph reasoning with message passing is a promising and popular solution. Some recent methods have achieved good performance, but they (i) usually can only predict triples involving unseen entities alone, failing to address more realistic fully inductive situations with both unseen entities and unseen relations, and (ii) often conduct message passing over the entities with the relation patterns not fully utilized. In this study, we propose a new method named RMPI which uses a novel Relational Message Passing network for fully Inductive KGC. It passes messages directly between relations to make full use of the relation patterns for subgraph reasoning with new techniques on graph transformation, graph pruning, relation-aware neighborhood attention, addressing empty subgraphs, etc., and can utilize the relation semantics defined in the ontological schema of KG. Extensive evaluation on multiple benchmarks has shown the effectiveness of techniques involved in RMPI and its better performance compared with the existing methods that support fully inductive KGC. RMPI is also comparable to the state-of-the-art partially inductive KGC methods with very promising results achieved. Our codes and data are available at https://github.com/zjukg/RMPI. | ['Jeff Z. Pan', 'Song Jiang', 'Huajun Chen', 'Mingyang Chen', 'Wen Zhang', 'Jiaoyan Chen', 'Yuxia Geng'] | 2022-10-08 | null | null | null | null | ['inductive-knowledge-graph-completion'] | ['knowledge-base'] | [ 1.74425095e-02 6.35560930e-01 -4.18839008e-01 -2.44172692e-01
-3.60701978e-01 -2.73340970e-01 5.52441597e-01 6.35676324e-01
-1.62660152e-01 9.16857481e-01 2.67806888e-01 -3.62435013e-01
-4.04730976e-01 -1.49478817e+00 -1.03106034e+00 -2.83875197e-01
-5.51904798e-01 8.49087596e-01 6.03894532e-01 -4.27029371e-01
-3.23936790e-01 2.22310841e-01 -1.76406240e+00 5.07222354e-01
8.37744594e-01 6.25484824e-01 -2.56570041e-01 4.26238447e-01
-4.02987480e-01 1.27690625e+00 8.90256744e-03 -7.42511928e-01
8.65296349e-02 7.43065998e-02 -1.28943169e+00 -4.92943913e-01
1.59671262e-01 -1.96261033e-02 -7.21896708e-01 9.34696555e-01
1.94119051e-01 1.57275289e-01 2.87399441e-01 -1.53378534e+00
-6.94033682e-01 1.34292889e+00 -3.88218284e-01 -3.39351818e-02
7.32031226e-01 -3.08157206e-01 1.21661532e+00 -9.83425438e-01
8.88006210e-01 1.23811793e+00 7.40949392e-01 4.08237338e-01
-9.54379916e-01 -6.56983137e-01 3.33494931e-01 7.29593873e-01
-1.75371611e+00 -1.11897677e-01 5.63284039e-01 -5.72098698e-03
1.38385308e+00 6.06710732e-01 6.33036137e-01 6.44155383e-01
-4.79865879e-01 7.31972396e-01 6.35229886e-01 -4.18384045e-01
5.37149720e-02 2.73550749e-01 5.17230392e-01 8.69705379e-01
5.80902696e-01 -1.04514144e-01 -4.84148711e-01 -2.17442080e-01
1.81308798e-02 2.24083569e-02 -5.91315806e-01 -3.73958021e-01
-1.07561874e+00 5.97622454e-01 7.50439882e-01 3.03256035e-01
-1.99125260e-01 2.91821480e-01 3.67374003e-01 2.72398710e-01
3.80953848e-01 2.44809315e-01 -7.37954676e-01 1.76988214e-01
-3.72797817e-01 3.02735537e-01 1.24011517e+00 1.49450231e+00
1.12487853e+00 -4.05172020e-01 -5.84331006e-02 5.89584649e-01
3.00552726e-01 3.56229335e-01 1.18661210e-01 -4.09732163e-01
8.98895383e-01 1.53496325e+00 -1.84636682e-01 -1.27834368e+00
-4.73492265e-01 -2.91267812e-01 -7.89856851e-01 -4.90129709e-01
1.56268862e-03 6.40144497e-02 -9.52289641e-01 1.66952837e+00
8.23400617e-01 5.21719277e-01 4.32667643e-01 5.90631485e-01
1.42103481e+00 6.12308621e-01 2.37034082e-01 1.84557736e-02
1.37949228e+00 -6.90544665e-01 -6.12019956e-01 -1.39074743e-01
1.22142017e+00 -2.48078242e-01 7.21088350e-01 -9.99537110e-02
-6.75898373e-01 -1.28275126e-01 -9.37432408e-01 -3.52191061e-01
-1.09262574e+00 -1.81576520e-01 1.15054035e+00 4.47368652e-01
-1.05676603e+00 6.73610687e-01 -6.59973800e-01 -6.10478103e-01
2.98799783e-01 5.20737827e-01 -6.23704255e-01 -5.48808396e-01
-1.80867910e+00 8.14785540e-01 1.13886333e+00 1.40510798e-01
-5.47785282e-01 -9.12863672e-01 -1.23563600e+00 2.30932623e-01
9.60480392e-01 -6.41120553e-01 5.89961350e-01 -2.69750834e-01
-7.77982950e-01 5.83803296e-01 -9.99549329e-02 -5.63549995e-01
1.53943095e-02 -2.76191741e-01 -8.23421478e-01 8.14711899e-02
4.54188846e-02 6.34645820e-01 1.24113031e-01 -1.29588842e+00
-7.20252275e-01 -3.65327448e-01 5.29578626e-01 3.14558238e-01
-4.39465523e-01 -3.86066675e-01 -7.52391338e-01 -1.86171710e-01
2.36180499e-01 -7.62688696e-01 1.73518546e-02 -7.10185528e-01
-8.16780090e-01 -6.65824950e-01 8.69961083e-01 -3.50639552e-01
1.46788239e+00 -1.76896060e+00 1.06065907e-01 5.84411979e-01
4.55521017e-01 3.83109748e-01 6.72925040e-02 8.56498778e-01
-2.61589199e-01 3.37450087e-01 -1.06560811e-01 -3.85329053e-02
-2.68963389e-02 5.47224641e-01 -2.63859093e-01 1.43091455e-01
1.41028628e-01 1.14022505e+00 -9.79937315e-01 -5.69482148e-01
8.92343596e-02 1.88538775e-01 -5.17741263e-01 -1.57261491e-01
-4.15731370e-01 -2.05541804e-01 -4.20840770e-01 8.43827486e-01
7.74318516e-01 -6.74060762e-01 6.90074325e-01 -4.09958571e-01
2.90661007e-01 1.39540836e-01 -1.38914049e+00 1.28543603e+00
-2.61054840e-02 -4.48412746e-02 -4.55245227e-01 -1.01251066e+00
6.82997227e-01 3.31527829e-01 2.16679260e-01 -2.87761539e-01
-1.50833398e-01 2.62732983e-01 -3.39470387e-01 -5.53004622e-01
5.55702150e-01 3.52397636e-02 6.56461269e-02 3.07287090e-02
1.89130247e-01 2.82956868e-01 5.55903018e-01 9.54352438e-01
1.46402693e+00 4.84902561e-02 3.27908844e-01 -2.14231059e-01
6.78582251e-01 3.51662159e-01 8.21821153e-01 6.36700451e-01
3.36172491e-01 2.30632678e-01 6.48540795e-01 -4.90068167e-01
-4.52143699e-01 -7.92522490e-01 2.33197346e-01 8.23011160e-01
4.56804723e-01 -1.09506118e+00 -2.61540979e-01 -1.03730762e+00
2.93274850e-01 7.34488606e-01 -6.52989745e-01 -3.87621492e-01
-5.33511102e-01 -7.72058606e-01 6.24532044e-01 5.44852972e-01
4.49546367e-01 -9.31753337e-01 4.19051826e-01 1.47864476e-01
-3.85990381e-01 -1.51317501e+00 1.10932156e-01 1.38025675e-02
-7.32074142e-01 -1.71008611e+00 3.80198471e-02 -8.93048882e-01
9.08092201e-01 1.83029503e-01 1.27634418e+00 6.05091631e-01
-1.43701330e-01 4.13845718e-01 -6.47822022e-01 -1.81693181e-01
-1.26792803e-01 3.71284336e-01 -2.48947423e-02 -6.68696016e-02
6.80514038e-01 -6.62034929e-01 -3.47219884e-01 1.58464983e-01
-9.06986654e-01 3.33700240e-01 4.05658245e-01 4.29295897e-01
6.03916943e-01 4.57594365e-01 5.52752316e-01 -1.83662474e+00
1.09632820e-01 -8.20849121e-01 -4.20841008e-01 6.91357136e-01
-8.81430447e-01 2.09273592e-01 5.43112218e-01 -1.70268416e-01
-1.00932264e+00 -8.13172162e-02 2.23978851e-02 -2.63519913e-01
1.33327842e-01 9.43239570e-01 -2.91005850e-01 -4.45846468e-02
5.57147086e-01 -9.56639498e-02 -4.61291254e-01 -3.97643358e-01
4.98154640e-01 1.54585302e-01 2.62858808e-01 -8.04289162e-01
1.18455875e+00 5.23431897e-01 1.09679431e-01 -5.10432959e-01
-8.28243911e-01 -6.21176541e-01 -3.21891159e-01 -3.45018394e-02
5.58412552e-01 -9.90110159e-01 -8.62598240e-01 2.07507551e-01
-9.30505812e-01 -2.34641299e-01 -3.55263174e-01 5.42662442e-01
1.08498238e-01 3.91705990e-01 -7.00113356e-01 -5.22098303e-01
-3.47444147e-01 -6.14072025e-01 6.67302012e-01 1.97722122e-01
-5.79858606e-04 -1.06876183e+00 -7.89419487e-02 3.54992956e-01
2.50507116e-01 4.05656815e-01 1.35296953e+00 -9.90528643e-01
-1.04861414e+00 -2.62603313e-01 -3.54273498e-01 -5.31652048e-02
4.92532030e-02 -2.03865349e-01 -6.69151962e-01 -2.17647165e-01
-8.11242640e-01 -3.01897466e-01 8.73248577e-01 -3.94437402e-01
8.98583174e-01 -4.41835195e-01 -8.72997463e-01 5.76160133e-01
1.65100110e+00 -2.37944290e-01 8.43101859e-01 9.08425823e-02
1.09661281e+00 5.79832613e-01 5.49096942e-01 8.35513398e-02
1.09649515e+00 2.92833298e-01 5.52835166e-01 1.80611573e-03
-1.10627785e-01 -6.78831995e-01 3.35382633e-02 9.41698611e-01
-2.93653131e-01 -3.90917480e-01 -1.02000427e+00 7.05673039e-01
-2.18202448e+00 -6.38306201e-01 -6.41049385e-01 1.97769892e+00
8.85396004e-01 2.47512478e-02 -4.21251416e-01 2.08472937e-01
8.23680580e-01 -1.51952639e-01 -3.41483474e-01 2.66755410e-02
-2.60846406e-01 3.06447089e-01 6.50890470e-01 5.26966989e-01
-9.23999846e-01 1.22878432e+00 5.26337528e+00 7.65756190e-01
-3.47229123e-01 4.27509956e-02 1.98277548e-01 9.18124393e-02
-7.17748761e-01 5.37364304e-01 -1.08075857e+00 -1.69345811e-02
7.82363474e-01 -2.12992340e-01 4.02974039e-01 8.30968738e-01
-5.08520246e-01 -1.39847189e-01 -1.24085391e+00 7.10318804e-01
-9.19742212e-02 -1.45487082e+00 2.96978951e-01 -2.77351923e-02
6.65251017e-01 -6.83767423e-02 -4.80500549e-01 9.93304968e-01
7.09165633e-01 -8.76769006e-01 1.72292590e-01 5.12869239e-01
6.40626371e-01 -6.90741837e-01 1.00282121e+00 2.13238254e-01
-1.63491023e+00 3.98351513e-02 -4.08764213e-01 -4.00688425e-02
-1.15900740e-01 8.06190610e-01 -9.11527693e-01 1.26452637e+00
8.73763084e-01 7.78243184e-01 -7.69243956e-01 5.56624055e-01
-5.13561487e-01 6.05035901e-01 -6.22094870e-01 -1.24741420e-01
-1.45170037e-02 8.46917406e-02 4.03086662e-01 1.07798409e+00
2.23598659e-01 4.70297128e-01 1.32305801e-01 6.32339716e-01
-4.40509915e-01 3.09814632e-01 -7.79452860e-01 9.79425851e-03
5.80531776e-01 1.34179282e+00 -5.84209442e-01 -5.58271289e-01
-6.75986111e-01 4.96492118e-01 7.87728131e-01 5.41485429e-01
-8.65656435e-01 -4.46151465e-01 4.26706821e-01 2.55964845e-01
3.06246966e-01 2.83429861e-01 2.77161568e-01 -1.26264942e+00
2.67246306e-01 -6.01936460e-01 1.04714584e+00 -6.28036439e-01
-1.21721447e+00 3.98192823e-01 2.47252718e-01 -8.15647840e-01
2.13166565e-01 -4.03160959e-01 -4.43936348e-01 5.02084315e-01
-1.66249287e+00 -1.34737194e+00 -3.70685607e-01 9.23563540e-01
-2.42291182e-01 7.62641206e-02 9.01262403e-01 6.08947217e-01
-5.30801237e-01 4.54921037e-01 -4.37506348e-01 1.04265280e-01
3.77234697e-01 -1.41637301e+00 2.33436480e-01 8.84344876e-01
9.20745581e-02 8.76990378e-01 3.52939099e-01 -1.05029809e+00
-1.79121804e+00 -1.50069880e+00 1.18088996e+00 -2.84683973e-01
5.74438334e-01 -5.00154376e-01 -1.19193399e+00 1.13536918e+00
-1.22688591e-01 3.79332215e-01 6.15056455e-01 7.05222845e-01
-3.51319134e-01 -4.48759139e-01 -1.10284042e+00 5.92541456e-01
1.42425358e+00 -3.61578912e-01 -4.31218505e-01 4.89648461e-01
1.15471768e+00 -4.98328865e-01 -1.13020670e+00 9.25512254e-01
5.59417345e-02 -5.31584799e-01 9.77816761e-01 -8.35493803e-01
6.55228719e-02 -6.74014211e-01 -9.03610364e-02 -1.01233113e+00
-3.76474202e-01 -3.36127073e-01 -7.13616431e-01 1.52910864e+00
8.31728220e-01 -8.72448921e-01 9.28746402e-01 7.96638012e-01
-8.80443770e-03 -8.55143130e-01 -5.80600619e-01 -5.47855854e-01
-4.40670669e-01 -4.62077081e-01 1.09528303e+00 1.21197438e+00
4.04699028e-01 4.61125374e-01 -2.36983180e-01 7.21717775e-01
5.31845927e-01 3.25411022e-01 6.91593051e-01 -1.26180303e+00
-6.18336461e-02 2.92161554e-01 -6.57422364e-01 -5.23003519e-01
2.06105426e-01 -1.46511877e+00 -3.78471583e-01 -2.00420380e+00
2.51749367e-01 -7.12455153e-01 -1.98675960e-01 9.69820142e-01
-3.01217377e-01 4.08249907e-02 -6.19897023e-02 -3.64883952e-02
-8.72747719e-01 5.03810883e-01 9.45415676e-01 -2.61086941e-01
-7.52901733e-02 -2.04378664e-01 -9.06157315e-01 6.27607822e-01
8.11933458e-01 -5.09143114e-01 -6.79647505e-01 -1.88111156e-01
8.18069041e-01 -9.42514762e-02 4.05734152e-01 -9.11622763e-01
6.84898555e-01 1.23458179e-02 1.83675960e-01 -7.09145069e-01
1.48089677e-01 -8.64862204e-01 6.13428414e-01 3.06272119e-01
-1.59440443e-01 -2.76747406e-01 9.69388634e-02 8.09487224e-01
-2.78153539e-01 -1.32873163e-01 1.88492358e-01 -1.78715408e-01
-1.07221806e+00 4.55535442e-01 2.70278364e-01 4.51537758e-01
1.08580959e+00 5.11950403e-02 -6.32637382e-01 -1.06508546e-01
-9.16408658e-01 8.63815904e-01 1.41432837e-01 4.03429031e-01
7.66008914e-01 -1.48475766e+00 -6.79445684e-01 -3.04939933e-02
4.70892876e-01 3.18520933e-01 3.07980120e-01 7.81466246e-01
-4.36639905e-01 2.63430566e-01 6.35745347e-01 -1.91842869e-01
-1.24033999e+00 8.32765639e-01 1.07584037e-01 -5.95541656e-01
-8.20993006e-01 7.96660304e-01 6.31097481e-02 -8.61924112e-01
1.71660945e-01 -4.39225286e-01 -2.97654063e-01 -1.66866779e-01
2.99259633e-01 4.09613043e-01 2.77354836e-01 -3.39743882e-01
-5.94775677e-01 2.63215125e-01 -2.67739266e-01 6.19317174e-01
1.31108367e+00 3.00212074e-02 -5.12560844e-01 1.44222394e-01
1.10381317e+00 -4.64712158e-02 -3.74300510e-01 -6.63787663e-01
3.66703957e-01 -2.55136132e-01 -3.44049811e-01 -7.81946182e-01
-1.18838096e+00 2.43326753e-01 3.84249166e-02 3.00237715e-01
9.19660985e-01 3.64819914e-01 8.61613810e-01 6.22859716e-01
5.65683126e-01 -9.86772716e-01 -5.28968513e-01 5.16991079e-01
6.51108325e-01 -1.19513857e+00 2.30998591e-01 -1.14173460e+00
-4.91787344e-01 8.10332596e-01 8.65978062e-01 1.28355324e-01
8.83645236e-01 2.18931124e-01 -4.45312977e-01 -6.55296326e-01
-1.02285695e+00 -5.49526215e-01 2.57480115e-01 5.60025513e-01
1.66712821e-01 2.57335395e-01 -2.83743531e-01 5.66253603e-01
-6.65499941e-02 -1.53537411e-02 2.42847040e-01 7.87298620e-01
-2.36014932e-01 -1.19113696e+00 3.26339304e-02 6.86087966e-01
-3.04280758e-01 -3.60482574e-01 -4.36310828e-01 1.18641734e+00
2.61018038e-01 8.83390546e-01 -4.65390533e-01 -5.31218827e-01
5.79950869e-01 1.76352471e-01 2.31638640e-01 -7.87916362e-01
-4.58935380e-01 -7.24188983e-01 6.80736661e-01 -6.61997080e-01
-3.67565006e-01 -3.40494126e-01 -1.67634583e+00 -5.13291359e-01
-5.31069219e-01 4.45092469e-01 7.80384541e-02 7.95555294e-01
6.41576886e-01 5.58030367e-01 1.65864363e-01 -6.15398362e-02
-8.28513410e-03 -8.77689123e-01 -6.71911120e-01 6.31308377e-01
-1.39688507e-01 -6.92984998e-01 -3.21642160e-01 -2.42245525e-01] | [8.854522705078125, 7.949342727661133] |
4be333e8-0427-40ed-a7e3-e4d70080bb4d | an-edge-map-based-ensemble-solution-to-detect | 2201.06098 | null | https://arxiv.org/abs/2201.06098v1 | https://arxiv.org/pdf/2201.06098v1.pdf | An Edge Map based Ensemble Solution to Detect Water Level in Stream | Flooding is one of the most dangerous weather events today. Between $2015-2019$, on average, flooding has caused more than $130$ deaths every year in the USA alone. The devastating nature of flood necessitates the continuous monitoring of water level in the rivers and streams to detect the incoming flood. In this work, we have designed and implemented an efficient vision-based ensemble solution to continuously detect the water level in the creek. Our solution adapts template matching algorithm to find the region of interest by leveraging edge maps, and combines two parallel approach to identify the water level. While first approach fits a linear regression model in edge map to identify the water line, second approach uses a split sliding window to compute the sum of squared difference in pixel intensities to find the water surface. We evaluated the proposed system on $4306$ images collected between $3$rd October and $18$th December in 2019 with the frequency of $1$ image in every $10$ minutes. The system exhibited low error rate as it achieved $4.8$, $3.1\%$ and $0.92$ scores for MAE, MAPE and $R^2$ evaluation metrics, respectively. We believe the proposed solution is very practical as it is pervasive, accurate, doesn't require installation of any additional infrastructure in the water body and can be easily adapted to other locations. | ['David Koop', 'Michael. E. Papka', 'Priyanjani Chandra', 'Pratool Bharti'] | 2022-01-16 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [-2.08640665e-01 -8.62248316e-02 6.47813618e-01 -3.45600307e-01
-4.03086990e-01 -4.70391929e-01 1.09720312e-01 4.11860436e-01
-8.26734662e-01 6.58613086e-01 -2.17158169e-01 -2.89206028e-01
4.17817049e-02 -1.42419708e+00 -2.42405042e-01 -7.03377724e-01
-4.98918861e-01 -1.75440148e-01 1.77360207e-01 -5.04964173e-01
3.81553590e-01 6.43413603e-01 -1.28678942e+00 -3.28060865e-01
9.87130344e-01 9.83864307e-01 1.71829358e-01 7.00222552e-01
-2.01113477e-01 5.74187823e-02 -2.56529242e-01 -1.95627660e-01
7.48257220e-01 -3.75259101e-01 -7.45909736e-02 -1.20928809e-01
3.98679018e-01 -7.11765409e-01 -1.95061952e-01 9.55789864e-01
7.09988475e-01 -4.05253582e-02 6.35253191e-01 -9.35654879e-01
1.67398490e-02 2.09419593e-01 -1.11082089e+00 5.30173123e-01
2.27966651e-01 5.15877545e-01 6.30276740e-01 -1.01544213e+00
2.00047389e-01 7.67593920e-01 1.09450710e+00 -1.03596170e-02
-8.68150592e-01 -9.66806054e-01 -1.23298883e-01 4.21745554e-02
-1.58720517e+00 -4.67105031e-01 4.78295773e-01 -3.75226110e-01
1.03644991e+00 2.73388892e-01 9.01495993e-01 -1.24711215e-01
4.11649436e-01 5.35437316e-02 1.36492479e+00 -4.97633725e-01
1.98128894e-01 3.81984236e-03 2.06289012e-02 9.16037381e-01
4.72676575e-01 1.41002029e-01 -2.02755809e-01 -2.53554821e-01
7.34766901e-01 9.02847722e-02 -2.18184069e-01 3.42968881e-01
-7.21877396e-01 8.39531541e-01 6.43269420e-01 6.36820644e-02
-4.81249392e-01 1.32547572e-01 8.41866508e-02 4.73155901e-02
3.02945733e-01 -1.11983530e-01 -3.00501227e-01 -4.40779664e-02
-1.15509760e+00 8.91666040e-02 5.26867330e-01 4.13780212e-01
9.86920536e-01 5.33781797e-02 2.66734749e-01 5.26849091e-01
7.25722671e-01 1.31244969e+00 1.02455638e-01 -8.38375330e-01
5.12836695e-01 7.14274347e-01 3.88746470e-01 -1.39874971e+00
-7.45148718e-01 7.15159625e-02 -1.05087328e+00 4.19654995e-01
4.33952719e-01 -6.82625234e-01 -9.80253398e-01 1.27530015e+00
3.24130446e-01 3.45961660e-01 -4.50849310e-02 6.82815254e-01
8.61794949e-01 7.47818470e-01 2.61732519e-01 -1.94799483e-01
1.44356596e+00 -8.72322693e-02 -4.11204368e-01 -3.02908957e-01
4.22031209e-02 -7.03151703e-01 5.97197831e-01 -1.71011195e-01
-9.07660842e-01 -2.97385752e-01 -1.02062118e+00 6.95804954e-01
-4.28130180e-01 1.29318535e-01 4.06955451e-01 8.39260519e-01
-1.00784624e+00 4.60146546e-01 -7.90661871e-01 -7.78022468e-01
3.49195093e-01 2.57792205e-01 -1.79369345e-01 2.19724849e-01
-1.13220155e+00 9.10240948e-01 -2.32037410e-01 5.99250257e-01
-5.22510171e-01 -4.85489845e-01 -7.50978947e-01 3.97023149e-02
-5.12012243e-01 -2.13715300e-01 5.59039056e-01 -4.10019070e-01
-1.02855289e+00 8.51385713e-01 -1.74246058e-01 -3.22494745e-01
5.27112842e-01 -3.18611488e-02 -3.84192079e-01 1.11584872e-01
2.97300547e-01 5.32015860e-01 3.94533873e-01 -9.26887572e-01
-1.03206325e+00 -6.47848725e-01 -3.81517202e-01 2.07707927e-01
-1.20951459e-01 -5.12895919e-02 4.70978059e-02 -4.41145182e-01
5.54742157e-01 -6.98285043e-01 -6.36945844e-01 3.80822837e-01
1.17227755e-01 1.47243023e-01 7.09329069e-01 -7.62972116e-01
1.26123369e+00 -1.95865381e+00 -8.32510889e-01 4.21492130e-01
1.68015044e-02 1.77143693e-01 3.12586695e-01 5.48027635e-01
4.11872298e-01 -3.56899351e-02 -9.06432331e-01 1.72679126e-01
-2.98732132e-01 2.03792676e-02 -1.16867542e-01 7.61323929e-01
4.65600798e-03 5.55879116e-01 -7.77444899e-01 -4.37145054e-01
3.76502246e-01 3.58782083e-01 5.06722629e-02 3.23188156e-01
4.35323596e-01 7.87535459e-02 -3.99788707e-01 7.58913040e-01
1.27016354e+00 2.27700368e-01 -1.59784526e-01 -1.07771628e-01
-7.79449284e-01 -5.22463679e-01 -1.57569957e+00 1.27429032e+00
-2.69154996e-01 6.32024586e-01 3.66072565e-01 -1.01025629e+00
1.34763002e+00 4.08610195e-01 5.15888035e-01 -7.70939589e-01
-3.25886048e-02 3.21087688e-01 -2.88474739e-01 -9.19156492e-01
2.51400352e-01 -2.02936023e-01 -1.75807342e-01 3.94125342e-01
-6.31656110e-01 -1.81770027e-01 5.02916649e-02 -1.21621512e-01
1.13471866e+00 -4.73082401e-02 3.87960255e-01 -5.74498713e-01
3.19880724e-01 2.23510370e-01 6.59224749e-01 6.80634797e-01
-8.13934863e-01 3.29731643e-01 -2.12157890e-01 -7.32429326e-01
-6.44454718e-01 -1.29440749e+00 -4.23108190e-01 6.80513501e-01
5.36925256e-01 5.15822649e-01 -6.13152564e-01 -1.71455607e-01
1.49312273e-01 2.70900846e-01 -4.51754272e-01 2.68554658e-01
-7.01732039e-01 -1.46593797e+00 6.59135461e-01 4.11311209e-01
1.41404855e+00 -1.07572365e+00 -1.44905758e+00 4.95964199e-01
-1.15209952e-01 -5.99438667e-01 -1.49878159e-01 5.59717529e-02
-1.09064960e+00 -8.92961621e-01 -8.29939425e-01 -7.36358941e-01
7.43089914e-01 2.48012736e-01 1.06194282e+00 4.18685786e-02
-7.61588573e-01 1.65961877e-01 -1.56099498e-01 -5.19539416e-01
4.20543551e-01 -4.58210766e-01 -2.88616210e-01 -6.84622079e-02
7.76192725e-01 -5.50271630e-01 -1.32747388e+00 1.34344786e-01
-4.81385231e-01 -4.91044104e-01 3.71158630e-01 4.36840355e-01
2.35499457e-01 -7.05613929e-04 7.47252166e-01 -4.09201950e-01
3.31545621e-01 -6.57971621e-01 -8.90733838e-01 1.48116395e-01
-6.26305401e-01 -2.08448350e-01 1.50407106e-01 3.29018712e-01
-8.81008208e-01 6.71118557e-01 3.76568921e-02 4.91418898e-01
-1.58654585e-01 6.02646708e-01 2.59775877e-01 -6.21168762e-02
6.81492388e-01 1.67624310e-01 -2.62186497e-01 -4.20126230e-01
7.97962472e-02 9.01824117e-01 6.52620256e-01 -8.20684358e-02
9.80795920e-01 9.15664673e-01 1.27063841e-01 -1.15706909e+00
8.29516202e-02 -6.04514837e-01 -3.77451569e-01 -5.74980855e-01
8.03106129e-01 -1.15141928e+00 -8.38265479e-01 8.24529350e-01
-9.39274073e-01 -3.13607678e-02 2.49902904e-01 3.45298707e-01
2.04278544e-01 4.85288888e-01 -2.00699657e-01 -1.25534618e+00
-1.00049841e+00 -7.40203202e-01 5.49737453e-01 6.95014715e-01
2.05189791e-02 -6.13711357e-01 2.46025085e-01 -8.37106034e-02
8.85018408e-01 6.21532440e-01 3.11701059e-01 1.85204506e-01
-3.42303038e-01 -3.62380862e-01 -5.81228018e-01 -2.08568841e-01
3.82396549e-01 1.80546239e-01 -1.08524215e+00 -1.65394410e-01
-1.20128982e-01 3.02038461e-01 8.33882451e-01 8.00290465e-01
2.65446633e-01 4.73261587e-02 -4.45139408e-01 3.83320570e-01
2.02061677e+00 5.64787805e-01 8.87723684e-01 4.26094264e-01
9.99191403e-02 4.52567786e-01 3.86332333e-01 9.27411675e-01
6.78356528e-01 2.77835220e-01 7.04288602e-01 -5.10695577e-01
2.68225729e-01 1.83347449e-01 3.46903622e-01 2.11779684e-01
-2.24312678e-01 1.17186539e-01 -1.11237442e+00 9.68781173e-01
-1.58817554e+00 -1.05268049e+00 -4.04563099e-01 2.35891366e+00
3.98779929e-01 -2.20178336e-01 -3.64024229e-02 4.65152524e-02
8.15783501e-01 -1.06082901e-01 -5.49626112e-01 -3.07305396e-01
-6.54353946e-02 6.12201512e-01 1.14491260e+00 5.83737016e-01
-1.39115512e+00 5.66879690e-01 5.89715624e+00 1.11499988e-02
-1.16990995e+00 -1.10075854e-01 7.56311715e-01 1.44500017e-01
5.98628558e-02 9.75285545e-02 -6.30464911e-01 5.83346188e-01
7.82797873e-01 -8.49296302e-02 2.10473724e-02 3.94876808e-01
9.32956278e-01 -8.15133154e-01 -2.88332850e-01 6.85123324e-01
-3.21090847e-01 -1.07954967e+00 -6.03959858e-01 -1.11764148e-01
5.01732767e-01 3.01231205e-01 -3.36260080e-01 -1.26567841e-01
4.91833985e-01 -1.09866428e+00 3.25073421e-01 6.67689085e-01
6.93876386e-01 -6.11890256e-01 8.67196262e-01 1.97505921e-01
-1.80034935e+00 3.10706217e-02 -1.98863104e-01 -4.27597374e-01
1.78931624e-01 6.36461437e-01 -4.98815477e-01 1.47847325e-01
1.38214910e+00 3.28642964e-01 -2.36369640e-01 1.32774043e+00
1.49316872e-02 4.93099600e-01 -1.02882910e+00 -1.45738453e-01
4.06913161e-01 -4.71188635e-01 1.77474841e-01 1.30895889e+00
7.60536790e-01 5.87350845e-01 7.96328112e-02 4.11949307e-01
2.33967468e-01 2.88137078e-01 -6.75843894e-01 8.49579632e-01
6.91387475e-01 1.16245234e+00 -8.91075730e-01 -1.59146041e-01
-4.26607907e-01 6.80937946e-01 -2.45885789e-01 1.59452081e-01
-8.39698434e-01 -9.48404610e-01 4.46241021e-01 8.39707851e-02
2.47139007e-01 -4.77123260e-03 -5.58971107e-01 -6.86158597e-01
2.34597906e-01 -1.00732394e-01 4.11328256e-01 -4.70803946e-01
-1.04307079e+00 6.01491570e-01 -1.24700934e-01 -1.25893664e+00
2.21580908e-01 -2.44257808e-01 -1.09345901e+00 1.16106796e+00
-1.71293712e+00 -7.69212067e-01 -9.38186944e-01 3.53882998e-01
2.75463194e-01 1.94630161e-01 9.24097240e-01 3.61114889e-01
-5.79595208e-01 5.14446676e-01 3.66911292e-01 3.69853139e-01
4.75527346e-01 -9.78841603e-01 3.08635265e-01 1.07401645e+00
-3.73031884e-01 1.47115380e-01 8.51583123e-01 -5.77220261e-01
-1.15950668e+00 -8.97110105e-01 9.89529192e-01 1.79204103e-02
3.83749008e-01 1.83245152e-01 -5.88679910e-01 1.77043512e-01
3.01546872e-01 2.60096103e-01 3.24791014e-01 -6.75782204e-01
4.40928638e-02 -4.79182839e-01 -1.85124791e+00 2.57302254e-01
5.20711601e-01 -4.15540896e-02 -3.43157560e-01 -3.10467333e-02
-1.38697237e-01 -1.05834447e-01 -9.66775119e-01 4.96881545e-01
7.72808373e-01 -1.23885143e+00 7.94092298e-01 3.14057082e-01
1.88653961e-01 -5.35013974e-01 -4.00429666e-01 -9.57738578e-01
-9.13830101e-02 -2.49214649e-01 7.05236554e-01 1.01402426e+00
7.93392599e-01 -8.19004059e-01 9.62436497e-01 8.20123732e-01
7.28306100e-02 -5.22144973e-01 -1.28536642e+00 -2.36951888e-01
1.76200762e-01 -2.31523737e-01 5.38749039e-01 6.30698264e-01
7.35481232e-02 -3.10723782e-01 -1.16039544e-01 7.57985890e-01
9.78480279e-01 1.77840546e-01 5.31564653e-01 -1.29356945e+00
2.09808573e-01 -2.50349343e-01 -4.25644875e-01 -6.46862745e-01
-6.89646482e-01 -3.85304123e-01 1.35169283e-01 -1.78507102e+00
1.10689662e-01 -7.56579220e-01 -3.02524447e-01 6.81398392e-01
-2.82021910e-01 4.92208153e-01 -1.94498077e-01 1.00921273e-01
3.54185551e-01 2.84815818e-01 5.98316312e-01 -2.87116230e-01
-3.09103280e-01 -1.09149784e-01 -2.89904416e-01 6.67044342e-01
9.28976953e-01 -4.67219681e-01 1.50133952e-01 -5.93514621e-01
4.51043397e-02 1.67095214e-01 3.06723446e-01 -1.16428375e+00
3.09927851e-01 -4.62819278e-01 3.78030509e-01 -5.34268677e-01
2.07142364e-02 -1.18064666e+00 3.71053368e-01 9.13412988e-01
4.50370580e-01 2.99399465e-01 5.03727078e-01 4.70939159e-01
-3.29185352e-02 -3.30686480e-01 9.99109864e-01 -4.16213930e-01
-9.46768999e-01 1.93902776e-01 -5.05263209e-01 -3.22293907e-01
1.19763267e+00 -5.10721385e-01 -2.66408473e-01 -1.87985823e-01
-3.88013393e-01 4.19165820e-01 2.42547065e-01 -1.33758083e-01
6.70489728e-01 -7.72382200e-01 -1.11345983e+00 2.06639633e-01
1.25627428e-01 -1.74614966e-01 3.48062962e-01 5.48550367e-01
-1.22816420e+00 -1.99715450e-01 -2.47180060e-01 -5.45715213e-01
-1.15505767e+00 -3.22509736e-01 6.18600368e-01 -5.03399372e-02
-7.70406663e-01 8.25393617e-01 -3.70353162e-01 -8.35294202e-02
2.08256356e-02 -1.89732537e-01 -3.68468255e-01 1.74704343e-01
5.59348524e-01 5.98490119e-01 -9.71980691e-02 -6.96463287e-01
-6.71838760e-01 1.22611606e+00 3.81230444e-01 -3.90737981e-01
1.52229738e+00 -3.77148747e-01 -1.23417914e-01 -1.11933602e-02
8.33428621e-01 -1.84506699e-01 -1.36463523e+00 5.11277169e-02
-8.15174133e-02 -4.04538542e-01 3.32426488e-01 -8.22742939e-01
-1.42651117e+00 6.71046197e-01 1.57072830e+00 1.53288782e-01
1.36325264e+00 -5.17236233e-01 8.24438810e-01 1.63111806e-01
3.83245081e-01 -1.04120040e+00 -6.66752756e-01 2.39733234e-01
5.07753730e-01 -1.55580020e+00 1.94341853e-01 -4.16180911e-03
-2.09828287e-01 1.05788064e+00 4.38267857e-01 -2.58659184e-01
1.14007723e+00 4.98863429e-01 5.63327551e-01 -4.73702818e-01
-1.12319335e-01 -1.83882818e-01 -5.68290353e-01 4.98735487e-01
4.02584851e-01 2.99243063e-01 -4.15649414e-01 -6.49086088e-02
1.66189641e-01 3.61015685e-02 4.96643335e-01 1.26867104e+00
-1.08363163e+00 -5.54519773e-01 -7.22471356e-01 7.06386626e-01
-3.09534311e-01 -2.74339736e-01 1.56249776e-01 5.23400068e-01
2.66270995e-01 1.34999824e+00 3.34980965e-01 -5.65390959e-02
5.63970447e-01 -3.44594242e-03 -6.49678707e-02 -4.79207523e-02
-7.51588106e-01 -1.13081723e-01 -2.49128014e-01 -2.46624500e-01
-6.42725229e-01 -8.07482660e-01 -1.44650555e+00 -4.27399993e-01
-2.00524017e-01 2.47938886e-01 7.83715665e-01 6.97843134e-01
2.41852179e-01 -2.90987134e-01 1.03745508e+00 -7.99833596e-01
-1.38609022e-01 -1.10608029e+00 -6.91700339e-01 6.57227114e-02
2.52164841e-01 -4.64198679e-01 -7.24734843e-01 1.07803814e-01] | [9.410316467285156, -1.4031308889389038] |
66eebde2-7613-432f-ab0a-647e6e231407 | exclusive-independent-probability-estimation | 1809.08168 | null | http://arxiv.org/abs/1809.08168v3 | http://arxiv.org/pdf/1809.08168v3.pdf | Exclusive Independent Probability Estimation using Deep 3D Fully Convolutional DenseNets: Application to IsoIntense Infant Brain MRI Segmentation | The most recent fast and accurate image segmentation methods are built upon
fully convolutional deep neural networks. In this paper, we propose new deep
learning strategies for DenseNets to improve segmenting images with subtle
differences in intensity values and features. We aim to segment brain tissue on
infant brain MRI at about 6 months of age where white matter and gray matter of
the developing brain show similar T1 and T2 relaxation times, thus appear to
have similar intensity values on both T1- and T2-weighted MRI scans. Brain
tissue segmentation at this age is, therefore, very challenging. To this end,
we propose an exclusive multi-label training strategy to segment the mutually
exclusive brain tissues with similarity loss functions that automatically
balance the training based on class prevalence. Using our proposed training
strategy based on similarity loss functions and patch prediction fusion we
decrease the number of parameters in the network, reduce the complexity of the
training process focusing the attention on less number of tasks, while
mitigating the effects of data imbalance between labels and inaccuracies near
patch borders. By taking advantage of these strategies we were able to perform
fast image segmentation (90 seconds per 3D volume), using a network with less
parameters than many state-of-the-art networks, overcoming issues such as
3Dvs2D training and large vs small patch size selection, while achieving the
top performance in segmenting brain tissue among all methods tested in first
and second round submissions of the isointense infant brain MRI segmentation
(iSeg) challenge according to the official challenge test results. Our proposed
strategy improves the training process through balanced training and by
reducing its complexity while providing a trained model that works for any size
input image and is fast and more accurate than many state-of-the-art methods. | ['Seyed Raein Hashemi', 'Ali Gholipour', 'Simon K. Warfield', 'Sanjay P. Prabhu'] | 2018-09-21 | null | null | null | null | ['infant-brain-mri-segmentation'] | ['medical'] | [ 3.95304769e-01 3.49817336e-01 -5.07746786e-02 -6.09219551e-01
-5.63309133e-01 -4.15932626e-01 4.73606773e-02 3.84801239e-01
-8.58200371e-01 5.34125149e-01 -4.54370558e-01 -2.62044936e-01
-1.94917023e-01 -5.36031544e-01 -8.13940823e-01 -5.72709143e-01
-2.05992535e-01 8.38626564e-01 6.66699469e-01 1.16928816e-01
1.52553141e-01 6.54338658e-01 -1.42783201e+00 2.83798903e-01
9.95798588e-01 1.15934968e+00 1.61976650e-01 4.82293487e-01
-1.94621012e-01 4.43080306e-01 -3.76378953e-01 -2.76733667e-01
4.30723011e-01 -1.71448976e-01 -1.09705341e+00 -2.18604356e-01
8.97919059e-01 -2.95014858e-01 1.89101055e-01 8.89045060e-01
8.17717373e-01 -7.33786449e-02 7.49885201e-01 -7.66128480e-01
-1.18350342e-01 7.89571524e-01 -8.72355461e-01 6.47699714e-01
-3.03703099e-01 -7.94909298e-02 4.87175435e-01 -5.35414159e-01
5.27264416e-01 6.61594748e-01 1.05766511e+00 7.99976289e-01
-1.40108538e+00 -7.02463627e-01 1.29752785e-01 2.21372560e-01
-1.07690275e+00 -8.02356601e-02 4.65491802e-01 -7.24782467e-01
1.03899181e+00 1.57342538e-01 8.16694915e-01 6.36872232e-01
4.62738276e-02 8.24121058e-01 1.19542110e+00 -1.63186505e-01
8.44314918e-02 -1.21170908e-01 1.60135821e-01 7.39205599e-01
2.48082370e-01 -1.96849495e-01 1.13160953e-01 1.33571595e-01
6.69321120e-01 -9.59280282e-02 -1.27149656e-01 -5.47847688e-01
-1.09410501e+00 7.67744184e-01 4.20353174e-01 7.97620058e-01
-4.27623034e-01 1.51426002e-01 5.83926857e-01 7.75236730e-03
8.07124138e-01 2.90555507e-01 -5.84651530e-01 1.32495299e-01
-1.47408271e+00 1.21927641e-01 4.35586780e-01 3.64620388e-01
6.53501809e-01 -1.86825201e-01 -1.74154654e-01 1.12659395e+00
-7.27797747e-02 2.50461578e-01 6.63194001e-01 -7.97368467e-01
3.99716139e-01 5.99261045e-01 -5.44280291e-01 -6.84132218e-01
-1.11209905e+00 -6.10563219e-01 -8.60885501e-01 3.99013519e-01
5.98973930e-01 -3.69826674e-01 -1.54182589e+00 1.69514775e+00
6.03943944e-01 3.90789025e-02 -4.46291625e-01 8.96269619e-01
9.02480185e-01 1.88561499e-01 1.08398825e-01 -1.58961579e-01
1.21343994e+00 -8.55925739e-01 -2.90784240e-01 -1.57633379e-01
8.46092701e-01 -5.24673462e-01 5.70306063e-01 3.52977633e-01
-1.35201418e+00 -4.50756997e-01 -1.07891417e+00 1.25387549e-01
-3.18428069e-01 -1.99660838e-01 6.83398724e-01 9.21101272e-01
-1.41902173e+00 9.31423128e-01 -1.04987240e+00 -2.99023557e-02
9.77037966e-01 1.03505445e+00 -3.90682071e-01 1.41180709e-01
-1.00498414e+00 1.10726762e+00 5.28011203e-01 -5.80486171e-02
-6.47684097e-01 -1.30653393e+00 -5.69833875e-01 -1.27494991e-01
1.12443186e-01 -3.53029191e-01 9.42212760e-01 -1.21605325e+00
-1.15165782e+00 1.31036985e+00 3.90015781e-01 -7.65513659e-01
7.68536985e-01 1.37209967e-01 -7.43598565e-02 6.10253632e-01
7.66191855e-02 1.30723214e+00 7.61382699e-01 -9.42439675e-01
-5.51489532e-01 -5.53109944e-01 -2.53720969e-01 -3.83461304e-02
6.58744648e-02 8.54145959e-02 9.04757623e-03 -4.36202884e-01
2.60549933e-01 -7.24090636e-01 -2.95774847e-01 -1.60935745e-01
-6.33596480e-02 -1.62604854e-01 6.68540597e-01 -9.38200116e-01
6.32020712e-01 -1.76852655e+00 -1.00455739e-01 3.44900340e-01
5.53617239e-01 5.27242184e-01 -1.41542777e-01 -4.71157163e-01
-4.32649791e-01 8.78445879e-02 -6.81233585e-01 -2.65351683e-01
-2.70077318e-01 -8.79717097e-02 3.09286028e-01 7.27934301e-01
3.24589014e-02 8.20212543e-01 -6.48335993e-01 -7.06926823e-01
1.93762541e-01 5.79605401e-01 -5.77714324e-01 1.57004260e-02
2.35079019e-03 6.69824064e-01 -3.37321423e-02 3.08348536e-01
9.53525126e-01 6.86324686e-02 -1.04314670e-01 -1.44594431e-01
-1.17652960e-01 -3.16878222e-02 -9.03554797e-01 1.74969792e+00
-4.71168399e-01 4.62612927e-01 2.63192505e-01 -1.62923419e+00
8.24232399e-01 4.42167968e-01 1.07923317e+00 -1.03156388e+00
3.57739717e-01 5.59871137e-01 3.37025881e-01 -4.19333577e-01
-3.14122051e-01 -2.50921965e-01 5.26594400e-01 5.30958116e-01
3.51741552e-01 -2.71669537e-01 4.26711917e-01 -2.04728216e-01
8.39552820e-01 1.20773353e-02 -2.53647119e-01 -5.23845434e-01
5.44176817e-01 -2.13258475e-01 3.81151497e-01 6.08088732e-01
-2.82448530e-01 8.69903982e-01 5.62691510e-01 -7.41071224e-01
-1.14971697e+00 -6.98762655e-01 -4.62691426e-01 9.63566184e-01
-1.70198560e-01 3.98818672e-01 -1.41756940e+00 -8.42908621e-01
-3.26896667e-01 3.96782130e-01 -9.38144743e-01 1.09603971e-01
-1.03075564e+00 -1.18484628e+00 7.21380889e-01 3.96855503e-01
4.03797537e-01 -1.13734174e+00 -9.70856845e-01 3.41120511e-01
1.16088670e-02 -1.09787071e+00 -3.80062103e-01 4.30276155e-01
-1.08870614e+00 -9.85120833e-01 -1.40266025e+00 -1.00739205e+00
7.90544987e-01 -2.45320454e-01 1.07238972e+00 1.82384595e-01
-6.26147032e-01 4.76802289e-02 -1.96724698e-01 -4.46509778e-01
-2.84841329e-01 4.94522452e-01 -3.51461619e-01 -1.30844206e-01
-1.62885189e-02 -7.67358899e-01 -9.94098604e-01 1.86062798e-01
-1.03863037e+00 2.94338286e-01 5.55433393e-01 7.65831947e-01
6.84187055e-01 -2.63548106e-01 5.43677747e-01 -9.13128793e-01
1.39710203e-01 -4.30245638e-01 -6.48239374e-01 2.48636737e-01
-7.90663958e-01 -8.97016935e-03 5.11523306e-01 -6.24055982e-01
-7.62981057e-01 1.22656845e-01 -5.73936045e-01 1.61548350e-02
-3.90466094e-01 7.93139413e-02 4.00540262e-01 -5.62720895e-01
5.70648015e-01 5.64045347e-02 1.67571977e-01 -3.92437726e-01
1.13671146e-01 2.74386615e-01 3.14310253e-01 -4.16001976e-01
1.29792541e-01 5.87522686e-01 2.06391037e-01 -3.29307437e-01
-7.63544679e-01 -4.15079266e-01 -9.93690014e-01 -1.89199433e-01
1.18710852e+00 -3.50226730e-01 -4.20721173e-01 6.92071617e-01
-1.10196936e+00 -5.86859763e-01 -2.23420575e-01 5.13348579e-01
-4.72628772e-01 2.66127378e-01 -5.21792471e-01 -2.10944414e-01
-8.72834265e-01 -1.58085823e+00 7.32989967e-01 2.85149753e-01
-2.61078831e-02 -1.12583232e+00 1.34588867e-01 5.97888112e-01
6.83546960e-01 4.85732079e-01 1.11215436e+00 -1.03798234e+00
-7.62277320e-02 1.98743399e-02 -4.77675736e-01 5.83586574e-01
-9.53724682e-02 -2.49532685e-01 -8.25957060e-01 -2.73717552e-01
6.89729825e-02 -1.28589645e-01 1.02903080e+00 8.81902337e-01
1.29391110e+00 1.89908549e-01 -1.06306911e-01 8.63703668e-01
1.36504042e+00 2.99777359e-01 2.32412457e-01 2.97457665e-01
8.27651441e-01 9.86634672e-01 8.18948001e-02 3.29563431e-02
2.61147320e-01 6.00349426e-01 5.17843544e-01 -7.35208750e-01
-5.34011602e-01 4.54821020e-01 -1.69412613e-01 5.09263456e-01
-7.46779963e-02 3.25692832e-01 -1.26447558e+00 1.03914320e+00
-1.43392837e+00 -5.77847362e-01 -2.57464051e-01 2.09442854e+00
1.08660972e+00 1.51575670e-01 3.80343914e-01 -4.87532374e-03
7.70488441e-01 -2.15713587e-02 -6.91553473e-01 -6.13156676e-01
-2.51081046e-02 8.69665027e-01 8.05374444e-01 3.16503972e-01
-1.17758095e+00 6.68831825e-01 5.78075123e+00 9.30890679e-01
-1.61591268e+00 5.40243149e-01 1.14257264e+00 -3.06889832e-01
8.55152402e-03 -4.22255903e-01 -5.15195310e-01 4.74110186e-01
9.53170121e-01 5.27067721e-01 3.12130332e-01 5.91044843e-01
-5.93450181e-02 -1.78300053e-01 -9.85444844e-01 7.49681592e-01
1.14389015e-02 -1.25931478e+00 -3.77953202e-01 -2.05188304e-01
8.31354082e-01 6.04638100e-01 3.97106707e-02 1.78864971e-02
-3.11303318e-01 -1.26753044e+00 8.38345706e-01 2.15454638e-01
8.20047379e-01 -7.23125756e-01 8.37580264e-01 5.40938228e-02
-7.22377121e-01 1.20711997e-01 -1.78671136e-01 3.54158163e-01
2.75100507e-02 7.42037594e-01 -8.46831143e-01 1.46057695e-01
8.91365528e-01 7.01928288e-02 -5.44514656e-01 1.27900136e+00
2.01674685e-01 5.20497620e-01 -4.32468235e-01 1.33395821e-01
5.34438074e-01 -5.84673509e-02 1.95024312e-01 1.41472518e+00
1.16585836e-01 4.84736301e-02 -7.41602033e-02 8.95306289e-01
4.17646766e-02 4.02717352e-01 -8.35318342e-02 5.63659847e-01
-1.83549136e-01 1.28489184e+00 -1.38939595e+00 -4.06559080e-01
-2.65658706e-01 5.30724227e-01 3.11052948e-01 2.61372756e-02
-7.88735807e-01 -3.01802814e-01 1.18197717e-01 3.62723172e-01
4.76561368e-01 1.21539764e-01 -7.08292425e-01 -6.38096988e-01
-6.15736619e-02 -5.99410474e-01 2.85055906e-01 -1.96074188e-01
-8.38332295e-01 8.21601748e-01 1.95652992e-01 -5.00133991e-01
-2.64867041e-02 -4.74529415e-01 -4.90904391e-01 7.09495485e-01
-1.64190209e+00 -9.81367528e-01 -8.44865143e-02 4.15295213e-01
2.75967956e-01 7.23017156e-02 3.57862145e-01 7.21599996e-01
-3.85323554e-01 7.06382871e-01 -1.75890222e-01 2.71116775e-02
3.79003584e-01 -1.32971454e+00 4.25796539e-01 6.84037447e-01
-9.06289145e-02 2.10906655e-01 3.90027791e-01 -4.76930439e-01
-4.59508657e-01 -8.68616641e-01 4.81908083e-01 1.83922872e-01
4.14456367e-01 -2.98034310e-01 -1.03281164e+00 1.86241776e-01
1.08028047e-01 2.21526489e-01 5.36047339e-01 -2.77437389e-01
-9.38377455e-02 -1.86997741e-01 -1.69121158e+00 1.06086656e-01
8.17021251e-01 -6.06558798e-03 -3.88031125e-01 5.47526121e-01
5.74460804e-01 -6.87375605e-01 -9.91639853e-01 8.04540694e-01
4.73227531e-01 -1.12572682e+00 9.85535622e-01 -9.30867493e-02
2.97998607e-01 1.09121017e-01 4.82656628e-01 -1.09864557e+00
-5.44277057e-02 -2.16614857e-01 2.54207790e-01 8.86642337e-01
5.89073122e-01 -7.47493923e-01 9.75352287e-01 6.52510703e-01
-2.74148703e-01 -1.19117606e+00 -1.22933984e+00 -5.19384563e-01
4.42435563e-01 -2.15591863e-01 6.22659683e-01 7.22571313e-01
-5.64479291e-01 -1.30519494e-01 1.21292084e-01 -1.71509221e-01
7.72420049e-01 -3.23549733e-02 -7.45344535e-02 -1.24824190e+00
-1.84407793e-02 -8.91881943e-01 -3.66841346e-01 -5.85698307e-01
1.11778185e-01 -1.05040884e+00 -4.53068363e-03 -1.55491018e+00
1.95004642e-01 -7.43822455e-01 -5.43745637e-01 8.05196702e-01
-1.02279931e-02 6.81977987e-01 1.37550496e-02 -1.26461595e-01
-2.90469646e-01 -1.76929250e-01 1.41584027e+00 -2.71433592e-01
-3.45135070e-02 1.87410433e-02 -3.36130291e-01 6.66961133e-01
8.02573502e-01 -6.41191244e-01 -3.92166734e-01 -6.68681145e-01
-1.98643580e-02 -8.49312767e-02 2.33636349e-01 -1.10087299e+00
8.12836811e-02 2.18256265e-01 4.96654302e-01 -6.30261183e-01
-3.37845273e-02 -8.09060752e-01 -2.71124616e-02 6.85156047e-01
-2.67570436e-01 -1.00251734e-01 4.36211705e-01 -4.20821548e-01
-2.30310261e-02 -5.44842064e-01 1.40339255e+00 -2.39755556e-01
-3.71042311e-01 6.80992424e-01 -1.61865994e-01 1.77783087e-01
1.07445168e+00 -5.82309306e-01 9.35816858e-03 1.81918129e-01
-9.27165329e-01 8.43607783e-02 3.13486546e-01 2.06721619e-01
3.82322997e-01 -6.92047119e-01 -8.29832852e-01 2.10139066e-01
-4.68984812e-01 2.94368207e-01 6.65255427e-01 1.48848009e+00
-1.04745567e+00 3.23811471e-01 -6.42857194e-01 -8.35201621e-01
-1.19684720e+00 2.48551935e-01 8.29910219e-01 -5.41739047e-01
-6.66635394e-01 1.24636424e+00 2.88696319e-01 -5.88025391e-01
2.82229662e-01 -5.46413422e-01 -4.83321756e-01 2.63282508e-01
4.70100284e-01 4.00435060e-01 7.12125778e-01 -6.80078149e-01
-5.49073040e-01 8.77117753e-01 -3.25124651e-01 1.89442411e-01
1.61802542e+00 1.29841298e-01 -3.95780236e-01 -8.24087635e-02
1.46035123e+00 -4.34170395e-01 -1.27217829e+00 9.20928791e-02
4.13695090e-02 -4.61006165e-02 4.09831762e-01 -1.06457150e+00
-1.78823519e+00 1.07706928e+00 1.30009723e+00 1.97980806e-01
1.16648281e+00 -5.75493425e-02 1.21049142e+00 -3.90603483e-01
1.83258384e-01 -1.10579157e+00 -2.95849293e-01 2.35384092e-01
5.98320544e-01 -1.22404099e+00 -3.20403511e-03 -1.43400997e-01
-3.60593408e-01 1.07186830e+00 6.39231741e-01 -1.84720054e-01
4.46756423e-01 3.94551009e-01 2.30335593e-01 -3.99955690e-01
-2.34854445e-01 -6.59996495e-02 5.57679415e-01 7.40927458e-01
4.49233890e-01 1.01342417e-01 -6.45996869e-01 4.01010066e-01
-8.86534154e-02 -2.48941392e-01 1.78056255e-01 4.83915985e-01
-3.55547041e-01 -1.13649082e+00 -1.47026584e-01 6.92608058e-01
-1.04003215e+00 -1.52925462e-01 1.54596688e-02 5.89226365e-01
7.46107399e-01 4.10709590e-01 1.46420211e-01 1.33689806e-01
2.42088988e-01 -1.56385258e-01 8.25407147e-01 -4.71519411e-01
-1.03551590e+00 -1.69358581e-01 -1.76839665e-01 -5.83083212e-01
-5.02519131e-01 -6.69483244e-01 -1.50493157e+00 -4.44617867e-02
-2.61252552e-01 -1.71338450e-02 1.07884514e+00 1.11890638e+00
1.89094052e-01 5.13831496e-01 4.36349481e-01 -1.07318377e+00
-2.36150205e-01 -7.62122214e-01 -4.79890645e-01 3.31508338e-01
3.04036558e-01 -7.25372076e-01 -6.30231202e-02 -2.11882114e-01] | [14.21423053741455, -2.357029676437378] |
3282f1c0-72bb-4d36-a34c-15a8a649a4d5 | continual-learning-in-task-oriented-dialogue | 2012.15504 | null | https://arxiv.org/abs/2012.15504v1 | https://arxiv.org/pdf/2012.15504v1.pdf | Continual Learning in Task-Oriented Dialogue Systems | Continual learning in task-oriented dialogue systems can allow us to add new domains and functionalities through time without incurring the high cost of a whole system retraining. In this paper, we propose a continual learning benchmark for task-oriented dialogue systems with 37 domains to be learned continuously in four settings, such as intent recognition, state tracking, natural language generation, and end-to-end. Moreover, we implement and compare multiple existing continual learning baselines, and we propose a simple yet effective architectural method based on residual adapters. Our experiments demonstrate that the proposed architectural method and a simple replay-based strategy perform comparably well but they both achieve inferior performance to the multi-task learning baseline, in where all the data are shown at once, showing that continual learning in task-oriented dialogue systems is a challenging task. Furthermore, we reveal several trade-offs between different continual learning methods in term of parameter usage and memory size, which are important in the design of a task-oriented dialogue system. The proposed benchmark is released together with several baselines to promote more research in this direction. | ['Zhiguang Wang', 'Eunjoon Cho', 'Zhou Yu', 'Bing Liu', 'Paul Crook', 'Seungwhan Moon', 'Zhenpeng Zhou', 'Zhaojiang Lin', 'Andrea Madotto'] | 2020-12-31 | null | https://aclanthology.org/2021.emnlp-main.590 | https://aclanthology.org/2021.emnlp-main.590.pdf | emnlp-2021-11 | ['intent-recognition'] | ['natural-language-processing'] | [ 1.24131851e-01 2.29453370e-01 -1.04788905e-02 -4.60623860e-01
-8.66838098e-01 -9.02381122e-01 1.08651614e+00 -2.07177568e-02
-5.37090421e-01 9.28496361e-01 2.64691293e-01 -5.82879484e-01
3.99749801e-02 -1.57354981e-01 -3.31284374e-01 -4.44171101e-01
-1.65038958e-01 8.00236166e-01 4.92708653e-01 -8.50123882e-01
2.55443394e-01 1.64640963e-01 -1.12332284e+00 4.48513120e-01
8.90918374e-01 6.14064753e-01 2.46161371e-01 8.64043236e-01
-1.32246733e-01 1.04603589e+00 -1.02096522e+00 -3.19721580e-01
1.13172777e-01 -3.48636329e-01 -1.48837280e+00 2.02170238e-01
3.72970790e-01 -5.74012041e-01 -1.91098392e-01 5.00664890e-01
7.84509718e-01 4.38121587e-01 5.19191384e-01 -1.22476912e+00
-1.25296414e-01 5.11897862e-01 -6.89912066e-02 6.40431186e-04
6.61046624e-01 6.21328771e-01 9.04332757e-01 -6.21486187e-01
3.78795952e-01 1.39744890e+00 5.21446943e-01 8.82488549e-01
-1.26019228e+00 -2.59128332e-01 3.28462839e-01 -4.39569168e-02
-6.98247612e-01 -9.34489131e-01 6.74514711e-01 -2.90156215e-01
1.41804373e+00 3.00817817e-01 2.34863639e-01 1.40049577e+00
2.48804837e-01 1.04953539e+00 1.19575441e+00 -4.93086100e-01
4.00214642e-01 3.01935613e-01 3.64907354e-01 6.51135266e-01
-3.66477668e-01 -1.51252136e-01 -4.68291610e-01 -3.96135628e-01
3.26083213e-01 -2.66592592e-01 1.90221500e-02 -3.07590038e-01
-1.21823633e+00 7.00348735e-01 -1.56482264e-01 3.27976525e-01
-2.18403310e-01 -9.60891619e-02 9.47237074e-01 9.18740749e-01
3.98503453e-01 7.05806792e-01 -7.90354848e-01 -5.45612991e-01
-3.70813370e-01 4.08082932e-01 1.32388735e+00 1.09445083e+00
6.34538352e-01 1.51533335e-01 -4.37618613e-01 1.19045734e+00
-1.28134713e-01 1.13723740e-01 7.34181941e-01 -1.18612111e+00
6.55839562e-01 6.44128859e-01 2.38342285e-01 -1.70655400e-01
-6.37813687e-01 1.33518443e-01 -7.85309494e-01 -5.37157021e-02
5.58720767e-01 -7.81800210e-01 -3.19523275e-01 1.61191237e+00
2.85522878e-01 -1.16883330e-01 5.38397133e-01 4.88144785e-01
6.06343389e-01 7.64716268e-01 1.94060206e-02 -4.34651852e-01
1.38630140e+00 -1.30227017e+00 -7.65244544e-01 -4.36232716e-01
9.27974463e-01 -7.47124493e-01 1.59591985e+00 5.65194428e-01
-1.20491064e+00 -5.45947552e-01 -9.19671476e-01 -1.31641291e-02
-1.71648100e-01 -1.57456860e-01 7.23126411e-01 7.15363383e-01
-1.17689252e+00 4.50252682e-01 -8.08292389e-01 -4.71498072e-01
-4.79379535e-01 4.32529241e-01 -5.54085374e-02 4.03766721e-01
-1.29714251e+00 1.20074952e+00 4.58155096e-01 -2.12846905e-01
-9.67157722e-01 -3.44292849e-01 -6.33489192e-01 -5.37200980e-02
6.86321259e-01 -5.50071776e-01 2.16729903e+00 -6.15837514e-01
-2.24540091e+00 4.33033615e-01 7.33031482e-02 -6.77457809e-01
6.66279554e-01 -3.76768351e-01 -9.89373475e-02 -2.16485009e-01
-1.50368810e-01 4.22312319e-01 5.94004571e-01 -9.32306945e-01
-7.93983161e-01 -1.42786801e-01 6.10489905e-01 5.11417687e-01
-6.14707828e-01 -4.87035438e-02 -1.93192065e-01 -2.23997846e-01
-4.82708156e-01 -1.06528997e+00 -2.02740029e-01 -5.27098656e-01
-2.80090064e-01 -8.37642312e-01 9.26594615e-01 -3.86275113e-01
1.04691470e+00 -1.93306899e+00 2.08618701e-01 -5.80790401e-01
-1.76934525e-02 4.08892512e-01 -2.48103380e-01 7.74896860e-01
1.95770934e-01 -1.77352279e-01 -1.68546841e-01 -4.96741444e-01
1.08947814e-01 1.84646741e-01 -2.24413007e-01 1.21346235e-01
2.42885277e-01 7.25885749e-01 -1.01054764e+00 -3.99314165e-01
2.66082734e-01 -1.17020331e-01 -4.24552023e-01 6.73922777e-01
-6.86083138e-01 4.72034872e-01 -2.25258216e-01 1.93284899e-01
3.86054963e-01 -9.89498347e-02 5.01217425e-01 1.63543865e-01
-6.24906532e-02 7.63763070e-01 -8.12291265e-01 2.09573483e+00
-9.31720674e-01 5.34134269e-01 1.10212779e-02 -9.93290246e-01
9.08460617e-01 5.94034493e-01 1.49470389e-01 -8.97168517e-01
-2.72144616e-01 1.25518637e-02 1.91790715e-01 -4.92131501e-01
8.71440709e-01 -1.42813817e-01 -4.58635986e-01 8.33231986e-01
3.95229846e-01 -2.83874154e-01 2.68964887e-01 1.74751207e-01
1.13650239e+00 -6.07628971e-02 4.29246277e-01 -1.74846262e-01
7.51085162e-01 1.13081187e-01 2.86214471e-01 8.51885200e-01
-2.64398813e-01 -1.09003317e-02 8.56807649e-01 -5.92357278e-01
-9.18290913e-01 -5.11444092e-01 1.58572450e-01 1.97116292e+00
-1.58076555e-01 -2.78135628e-01 -7.58624732e-01 -1.06703722e+00
-2.61287868e-01 1.09340894e+00 -3.25486183e-01 -1.87101215e-01
-9.26760375e-01 -8.02009404e-01 7.29816973e-01 1.92707717e-01
7.80509353e-01 -1.09795475e+00 -6.99948430e-01 3.91380519e-01
-3.61412674e-01 -1.05324435e+00 -6.11725211e-01 4.25553590e-01
-1.01481187e+00 -7.89026380e-01 -7.72737503e-01 -8.76489043e-01
2.87384372e-02 3.97723913e-01 1.17747593e+00 -8.36502165e-02
2.37261891e-01 5.72502494e-01 -3.77667576e-01 -9.63377357e-02
-1.13360703e+00 7.24872649e-01 1.85478047e-01 -2.38373071e-01
3.05300467e-02 -2.92934388e-01 -4.19249475e-01 5.42885840e-01
-6.35840535e-01 -6.49619196e-03 5.22247791e-01 1.25262070e+00
-3.05888504e-01 -3.19756657e-01 1.06162059e+00 -1.18534708e+00
1.34717643e+00 -3.69800717e-01 -6.22729599e-01 5.51163554e-01
-7.65981734e-01 2.88513690e-01 9.62310195e-01 -4.54387009e-01
-1.52733934e+00 7.58208409e-02 -2.31628314e-01 2.87503660e-01
-2.55940109e-01 1.36602849e-01 2.98464261e-02 2.32269496e-01
1.00384021e+00 5.30872285e-01 3.15849453e-01 -4.38728839e-01
5.19324124e-01 9.69867051e-01 2.56200075e-01 -7.66420186e-01
2.70471722e-01 -1.69834748e-01 -5.09616494e-01 -1.02123260e+00
-6.15369558e-01 -6.40481412e-01 -7.09351778e-01 -3.85837294e-02
1.81094319e-01 -8.25753629e-01 -9.29844201e-01 6.47168398e-01
-1.45538247e+00 -9.65814412e-01 -1.46633893e-01 -4.09561880e-02
-8.97613645e-01 6.73316300e-01 -7.39259422e-01 -9.69630837e-01
-6.24023080e-01 -1.22082317e+00 9.53387856e-01 1.30957458e-02
-3.66473645e-01 -1.28442514e+00 2.72413135e-01 3.52097958e-01
5.55660367e-01 -3.65943342e-01 9.02850151e-01 -1.21791530e+00
-4.83529530e-02 -1.64205059e-02 4.15109359e-02 3.99421334e-01
2.83731610e-01 -4.83042568e-01 -1.07971621e+00 -7.48422623e-01
3.13114911e-01 -9.89474297e-01 3.95858139e-01 -1.96875021e-01
5.94776988e-01 -7.75864124e-01 6.08871840e-02 8.74333009e-02
7.57931113e-01 4.94459569e-01 4.10803080e-01 3.75213385e-01
3.22950274e-01 8.53767037e-01 8.47470701e-01 5.33879101e-01
4.56382036e-01 9.45441306e-01 8.86153057e-02 7.07881004e-02
-3.27403694e-02 1.52953789e-01 7.81678617e-01 8.40861857e-01
1.25438243e-01 -2.19308555e-01 -9.86839950e-01 3.54505181e-01
-2.01052713e+00 -7.79246271e-01 2.52128422e-01 2.29980445e+00
1.30006158e+00 2.06891730e-01 6.40553355e-01 -1.73334762e-01
3.72686923e-01 1.96503088e-01 -6.65674567e-01 -7.26263762e-01
1.89197421e-01 -1.39160544e-01 -9.44603048e-03 7.44232714e-01
-1.12324452e+00 1.22941852e+00 6.62643051e+00 7.13354051e-01
-1.16308868e+00 2.45243892e-01 4.94545311e-01 1.17031164e-01
1.21859446e-01 -6.38946891e-02 -9.69374955e-01 2.84317106e-01
1.37676811e+00 -3.02574992e-01 3.53073388e-01 9.60131347e-01
1.53244659e-02 -8.40628818e-02 -1.33434188e+00 5.05911291e-01
-1.17849141e-01 -1.01537836e+00 -1.48844168e-01 -2.07772851e-01
4.40527797e-01 -4.16744500e-02 -6.27150685e-02 9.26516354e-01
6.79396570e-01 -5.24825215e-01 1.56371832e-01 -2.04510037e-02
5.90536535e-01 -4.69550520e-01 7.17756629e-01 8.38386059e-01
-6.47524655e-01 -1.61144704e-01 -2.80773968e-01 -8.87634829e-02
4.50942926e-02 -1.36030883e-01 -1.66510117e+00 3.82262975e-01
1.57229081e-01 2.33434036e-01 -4.44038838e-01 7.22337902e-01
6.45294040e-02 5.16674876e-01 -7.82978237e-02 -4.15208161e-01
4.06895041e-01 1.67049050e-01 4.31572229e-01 1.51947641e+00
-1.75690174e-01 -2.15047225e-01 3.48072678e-01 3.78927499e-01
7.31681138e-02 1.15378812e-01 -7.77172089e-01 2.81448394e-01
3.36838096e-01 1.18853676e+00 -2.86000192e-01 -4.15158510e-01
-4.01997864e-01 1.30491710e+00 5.22383392e-01 1.41371623e-01
-5.92951655e-01 -4.10644352e-01 2.90070087e-01 -1.22373253e-01
-1.92911699e-01 -3.50760311e-01 -4.65944521e-02 -1.15642643e+00
-7.75621384e-02 -1.29905760e+00 3.33749026e-01 -1.02839783e-01
-1.27433717e+00 8.09385240e-01 1.48157462e-01 -1.14016128e+00
-1.00763261e+00 -4.15077239e-01 -5.43296039e-01 5.99071562e-01
-1.40463471e+00 -9.22388136e-01 -1.44779995e-01 7.30534196e-01
1.41891515e+00 -5.55005789e-01 1.19931269e+00 1.05063505e-02
-6.27938330e-01 8.31470132e-01 2.23922655e-01 -1.04873151e-01
1.15879250e+00 -1.53426516e+00 7.86307454e-01 2.93481439e-01
-2.09032238e-01 6.48241282e-01 6.39420807e-01 -4.40714538e-01
-1.68703103e+00 -8.75252247e-01 6.55603945e-01 -4.89126682e-01
7.36209512e-01 -6.65662348e-01 -9.92356777e-01 7.13953257e-01
7.18360543e-01 -5.33288479e-01 5.51890492e-01 6.38277173e-01
-1.39308885e-01 -1.02719650e-01 -9.04888451e-01 5.42579234e-01
7.72718787e-01 -3.27191561e-01 -7.84046352e-01 7.67057717e-01
1.00441921e+00 -5.41332185e-01 -7.47775018e-01 2.01694176e-01
4.72399175e-01 -1.02629745e+00 8.19065511e-01 -7.74244428e-01
1.19794413e-01 3.81346524e-01 1.68660343e-01 -1.55419731e+00
3.31355981e-03 -1.22893631e+00 -1.60997927e-01 1.20579576e+00
4.85451311e-01 -8.49044204e-01 5.74764431e-01 5.87302566e-01
-2.64984548e-01 -5.02593696e-01 -8.68051946e-01 -1.04503000e+00
4.85500306e-01 -4.73358929e-02 3.59036416e-01 7.92975724e-01
4.94621575e-01 1.20765364e+00 -6.84324980e-01 -3.55549365e-01
2.41039649e-01 -2.12238077e-02 1.31883597e+00 -1.02543318e+00
-4.13160354e-01 -2.82863796e-01 4.21092540e-01 -1.66646039e+00
2.52269328e-01 -4.88282263e-01 2.18831211e-01 -1.06925583e+00
-3.85102094e-03 -6.43711865e-01 5.32156304e-02 6.04270637e-01
-1.85023203e-01 -5.26457906e-01 2.40605891e-01 4.61025596e-01
-1.13108850e+00 7.22691059e-01 1.17029643e+00 -3.26170206e-01
-7.98163414e-01 4.96209592e-01 -6.28684044e-01 6.42149687e-01
7.87921548e-01 -1.40517622e-01 -6.63160026e-01 -3.75017613e-01
-2.58398831e-01 6.58967137e-01 -1.10602021e-01 -8.43105853e-01
5.75452745e-01 -2.23958522e-01 -3.51072133e-01 -3.68208170e-01
4.85656321e-01 -5.80465376e-01 -5.47543466e-01 5.83577454e-01
-6.43028498e-01 -2.43581273e-02 3.87172878e-01 4.16780472e-01
-5.83112873e-02 -4.14471924e-01 6.26543283e-01 -2.47383252e-01
-8.57300103e-01 -2.24393532e-01 -7.16830492e-01 3.34388256e-01
1.00563443e+00 1.04457676e-01 -6.25551462e-01 -6.69071138e-01
-6.19447947e-01 5.21860242e-01 4.44916897e-02 5.41947126e-01
3.34245414e-01 -7.11720645e-01 -7.77439296e-01 -7.54948407e-02
1.78305149e-01 8.43790639e-03 2.06485435e-01 6.37159884e-01
-3.02648097e-01 7.48713613e-01 -2.60079056e-01 -7.80163586e-01
-1.26792228e+00 3.38181227e-01 3.66674721e-01 -7.69654751e-01
-4.81903821e-01 5.39246798e-01 1.65605724e-01 -1.00286365e+00
6.05223775e-01 9.21234768e-03 -2.87352115e-01 1.85609698e-01
3.71482223e-01 5.03578067e-01 3.10044855e-01 3.01178452e-02
-1.01837464e-01 -3.55907492e-02 -7.16968179e-01 -2.97142267e-01
1.14796376e+00 -2.77533531e-01 8.53943005e-02 7.58221507e-01
7.82794714e-01 -4.47308421e-01 -1.27164006e+00 -5.33887267e-01
3.07214558e-01 -8.12678933e-02 -3.64665776e-01 -1.14534092e+00
-3.51964027e-01 9.62897956e-01 5.74458301e-01 6.96193874e-01
8.83689165e-01 -2.22183496e-01 7.77074873e-01 1.16655755e+00
5.45304954e-01 -1.23588526e+00 6.40758097e-01 1.18392909e+00
1.02999783e+00 -1.45469689e+00 -1.84500203e-01 -2.77592521e-02
-1.04034793e+00 1.09077716e+00 1.02919269e+00 2.02173024e-01
6.53382912e-02 2.32075766e-01 1.00736901e-01 2.27214582e-02
-1.33994484e+00 5.61846010e-02 -1.09533787e-01 5.98753273e-01
6.42145813e-01 -1.91948771e-01 -2.35487923e-01 3.73676121e-01
-9.73653048e-02 -1.53604165e-01 6.06834888e-01 9.63075638e-01
-3.97944391e-01 -1.48477244e+00 -1.35965899e-01 9.47873294e-02
-2.35829055e-01 5.42226955e-02 -6.90132320e-01 7.80070305e-01
-6.23945892e-01 1.07210386e+00 -1.94013193e-01 -3.27999026e-01
6.99477255e-01 5.61661899e-01 3.89605790e-01 -8.80572259e-01
-1.10260963e+00 1.10012427e-01 6.61503613e-01 -2.05552429e-01
-9.86907706e-02 -5.32683134e-01 -9.51042533e-01 -2.65897334e-01
-4.66463178e-01 2.95071840e-01 5.00777185e-01 8.20227742e-01
4.30829555e-01 4.25375104e-01 1.02322102e+00 -6.08625352e-01
-1.25343120e+00 -1.49423075e+00 -1.63902849e-01 1.61526561e-01
4.31759477e-01 -4.61688399e-01 -2.59216517e-01 -5.99570684e-02] | [12.856510162353516, 8.003944396972656] |
cab2fae5-9cc1-4553-9531-e89ab4e8f6b1 | training-deep-spiking-neural-networks-using | 1608.08782 | null | http://arxiv.org/abs/1608.08782v1 | http://arxiv.org/pdf/1608.08782v1.pdf | Training Deep Spiking Neural Networks using Backpropagation | Deep spiking neural networks (SNNs) hold great potential for improving the
latency and energy efficiency of deep neural networks through event-based
computation. However, training such networks is difficult due to the
non-differentiable nature of asynchronous spike events. In this paper, we
introduce a novel technique, which treats the membrane potentials of spiking
neurons as differentiable signals, where discontinuities at spike times are
only considered as noise. This enables an error backpropagation mechanism for
deep SNNs, which works directly on spike signals and membrane potentials. Thus,
compared with previous methods relying on indirect training and conversion, our
technique has the potential to capture the statics of spikes more precisely.
Our novel framework outperforms all previously reported results for SNNs on the
permutation invariant MNIST benchmark, as well as the N-MNIST benchmark
recorded with event-based vision sensors. | ['Tobi Delbruck', 'Michael Pfeiffer', 'Jun Haeng Lee'] | 2016-08-31 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 7.57435501e-01 -5.13982594e-01 3.58770013e-01 -9.19274986e-02
-2.66752273e-01 -6.82494938e-01 6.41578853e-01 3.32078040e-01
-8.68768573e-01 1.07746947e+00 -4.09258306e-01 5.39862551e-03
-5.25321178e-02 -1.09316289e+00 -1.20509207e+00 -1.06429076e+00
-7.13985413e-02 1.03132360e-01 5.64907789e-01 1.87719297e-02
4.64791626e-01 8.26855600e-01 -1.63895738e+00 4.11024004e-01
4.55809176e-01 1.29161429e+00 1.31117120e-01 5.17821252e-01
-2.32975408e-01 7.66944706e-01 -6.60272121e-01 -9.86463651e-02
8.96711498e-02 -5.15460074e-01 -7.22197145e-02 -7.19942212e-01
1.57459363e-01 2.19889075e-01 -6.73781514e-01 1.10521555e+00
6.81732833e-01 -7.77453035e-02 4.80720937e-01 -1.35739768e+00
-3.44908386e-01 6.36055171e-01 -2.05851514e-02 5.18610775e-01
-1.35133833e-01 2.82642543e-01 4.91319299e-01 -6.24122024e-01
5.99314570e-01 7.01629281e-01 1.13690317e+00 6.19979501e-01
-1.60996032e+00 -9.07772958e-01 -2.25491866e-01 3.58266652e-01
-1.23915803e+00 -5.41348577e-01 6.74213588e-01 -2.15538293e-01
1.44510853e+00 -7.27282614e-02 1.02287436e+00 1.52074659e+00
7.38334298e-01 5.92006505e-01 1.21287537e+00 -1.85898263e-02
9.40052390e-01 -6.97417855e-01 2.26332113e-01 1.57162488e-01
4.94861603e-01 2.60385007e-01 -8.55354846e-01 2.00416595e-02
1.03666508e+00 3.34896296e-01 -7.48290643e-02 -4.07315753e-02
-1.35418880e+00 3.04988772e-01 5.46683967e-01 2.03093663e-01
-8.08207750e-01 1.11744022e+00 4.45746660e-01 2.36099124e-01
-3.21602583e-01 4.15723860e-01 -4.33842748e-01 -3.99097621e-01
-9.35817361e-01 3.20633203e-01 8.11862051e-01 5.95618725e-01
6.64077044e-01 3.46116483e-01 -3.76595020e-01 2.69887924e-01
-1.27102122e-01 5.96860170e-01 6.38843417e-01 -9.34128165e-01
-4.79088388e-02 7.13184893e-01 -1.45050123e-01 -5.89492321e-01
-5.44106722e-01 -2.87238568e-01 -1.17225504e+00 4.54603732e-01
7.04056621e-01 -8.82241353e-02 -1.22633839e+00 1.75952983e+00
-3.80839080e-01 6.61056995e-01 2.91059166e-01 3.49155098e-01
7.53531337e-01 5.12711465e-01 -6.10616133e-02 -2.55337238e-01
1.05236328e+00 -2.96271741e-01 -8.16592813e-01 -2.54432142e-01
-1.79755669e-02 -1.32473975e-01 5.35385311e-01 3.75309795e-01
-1.36939073e+00 -3.08991402e-01 -1.03452325e+00 1.84723884e-02
-7.24545002e-01 -2.70218402e-01 8.07801902e-01 3.73051256e-01
-1.37009490e+00 8.62633109e-01 -1.51074922e+00 -1.97941080e-01
8.36245775e-01 7.94172943e-01 4.17487100e-02 4.37678397e-01
-9.22210217e-01 6.44737899e-01 3.96243036e-01 9.82312113e-02
-8.90402138e-01 -6.31335557e-01 -4.14231569e-01 2.89555639e-01
-2.57058978e-01 -3.47290188e-01 1.09403300e+00 -8.19907427e-01
-1.68478572e+00 6.71458662e-01 -3.69171321e-01 -1.16816878e+00
1.41318366e-01 5.50010026e-01 -2.20916435e-01 -9.62156430e-02
-4.74610955e-01 8.16398144e-01 3.31458986e-01 -7.46006250e-01
-3.96091461e-01 -3.57433826e-01 -2.39530608e-01 -4.24591959e-01
-1.72367349e-01 -2.44677112e-01 -1.10555261e-01 -7.48511374e-01
2.51308680e-01 -8.60447764e-01 -1.88923851e-01 3.60501528e-01
-1.05508715e-01 -2.38826931e-01 5.40047526e-01 -4.00969200e-02
7.23757863e-01 -2.22431207e+00 9.29938331e-02 2.04286352e-01
1.80303946e-01 2.08655536e-01 -9.24747810e-02 4.27311957e-01
7.37511143e-02 -3.11587095e-01 -5.23094475e-01 -2.07625958e-03
-1.07214585e-01 5.38867772e-01 -3.26728433e-01 2.41253719e-01
5.55863976e-01 1.38151240e+00 -8.58914316e-01 7.92961568e-02
-6.25902265e-02 7.94787705e-01 -1.96568310e-01 -3.01183522e-01
-1.66919008e-01 3.61801624e-01 9.50252544e-03 5.67483068e-01
6.15051568e-01 -2.10992903e-01 -3.26586328e-02 -2.45578155e-01
-4.17912483e-01 4.82639045e-01 -9.32521284e-01 1.81012762e+00
-1.72968000e-01 8.68620992e-01 -2.86438286e-01 -1.35347342e+00
1.01891363e+00 2.06641749e-01 4.72189724e-01 -1.15263522e+00
3.45540553e-01 4.68563259e-01 2.32942536e-01 1.15803286e-01
-1.66246578e-01 2.03902066e-01 1.13580212e-01 2.30949357e-01
1.64963648e-01 2.66661227e-01 2.61799961e-01 -2.13068873e-01
1.66483521e+00 -6.39829859e-02 3.16082910e-02 -2.29930326e-01
6.31474331e-03 -1.63772419e-01 6.45764649e-01 9.04006243e-01
-1.46321908e-01 3.61302674e-01 4.37017709e-01 -4.08379197e-01
-8.13505292e-01 -1.64298391e+00 -4.93831009e-01 6.09188676e-01
3.36775869e-01 1.13855779e-01 -7.07723260e-01 1.02565028e-01
-2.24281512e-02 5.59382200e-01 -6.67717099e-01 -3.62913668e-01
-7.17028081e-01 -1.00553644e+00 1.14934254e+00 9.25469816e-01
4.65158284e-01 -1.51044488e+00 -1.18216181e+00 8.19380820e-01
4.96714115e-02 -1.21240270e+00 9.74421054e-02 1.31143427e+00
-1.16021144e+00 -6.55391872e-01 -5.26983559e-01 -9.41956460e-01
5.62090635e-01 -3.92014265e-01 8.48467350e-01 -2.21390858e-01
-5.61772823e-01 -1.62149966e-01 7.90917352e-02 -7.97940910e-01
-7.33522549e-02 -2.25817591e-01 -7.88516849e-02 -7.22982213e-02
7.93048680e-01 -1.23500013e+00 -7.44855225e-01 -1.30001888e-01
-9.47253823e-01 -3.10167614e-02 5.69144249e-01 8.45717013e-01
1.30403101e+00 -8.79375041e-02 7.50438333e-01 -5.43160141e-01
2.39718229e-01 -2.56829590e-01 -7.93230534e-01 3.75413820e-02
-5.07446170e-01 3.58708054e-01 8.38297725e-01 -7.94039845e-01
-5.49497247e-01 4.73062932e-01 -1.01397149e-01 -6.30853251e-02
-2.49939606e-01 1.06448129e-01 2.34069258e-01 -4.98144507e-01
7.66389132e-01 6.80183291e-01 -1.77384108e-01 -1.35752469e-01
-4.80269492e-01 9.56760123e-02 9.03718472e-01 -2.97255427e-01
3.68629783e-01 7.71182835e-01 3.89128864e-01 -4.26968992e-01
4.60727364e-02 2.87122037e-02 -3.32428485e-01 -1.51921110e-02
5.33903480e-01 -6.93990409e-01 -1.13502717e+00 1.17385304e+00
-1.44915223e+00 -4.65173960e-01 -5.11256456e-01 4.82770920e-01
-7.06268489e-01 -1.35061353e-01 -1.06879175e+00 -7.80482173e-01
-4.47915345e-01 -8.66114855e-01 7.64814258e-01 5.25579989e-01
1.03751168e-01 -7.70500541e-01 1.35070801e-01 -7.37892807e-01
7.25684464e-01 5.23115277e-01 6.92015588e-01 -7.77994692e-01
-7.22042739e-01 -2.07045332e-01 -3.24302852e-01 4.97104265e-02
-4.77071330e-02 5.86732961e-02 -1.11789775e+00 -1.80085346e-01
4.95894402e-02 -2.56018251e-01 1.06614006e+00 8.07814837e-01
1.37749016e+00 9.42283403e-03 -6.41322732e-01 8.64199579e-01
1.78280854e+00 5.00977635e-01 1.06167746e+00 2.64328450e-01
2.49818757e-01 -1.18625844e-02 -4.17685688e-01 4.37722623e-01
1.36103064e-01 3.76306474e-01 6.86150312e-01 5.57992458e-02
-1.25211403e-01 5.79732955e-02 3.65523547e-01 6.95616424e-01
-2.62184203e-01 -3.71304095e-01 -8.70271862e-01 6.46072209e-01
-1.75692439e+00 -1.06379843e+00 -3.36495340e-01 2.27430892e+00
9.85790372e-01 3.83050114e-01 -2.95201361e-01 5.14007270e-01
5.90249479e-01 -3.16445500e-01 -1.11512196e+00 -2.98877776e-01
-6.17893577e-01 9.74981606e-01 8.37320805e-01 -3.71774137e-01
-5.64361036e-01 4.11145538e-01 6.25385523e+00 4.35286939e-01
-1.42532825e+00 -8.52653161e-02 1.19912520e-01 -3.04679811e-01
-1.27244756e-01 -4.10758972e-01 -8.77056599e-01 8.27128708e-01
1.35849929e+00 -1.33691862e-01 9.49476242e-01 1.07193425e-01
-3.01983468e-02 -3.86748463e-02 -1.45831931e+00 1.32188666e+00
-5.11262059e-01 -1.83977950e+00 -2.07629651e-01 -1.36694506e-01
6.97532117e-01 5.96674085e-01 -7.39756785e-03 1.92439944e-01
1.85809195e-01 -1.10625768e+00 6.05262578e-01 7.40046442e-01
5.53291619e-01 -7.48444438e-01 6.35506451e-01 2.55186617e-01
-1.16484332e+00 1.23125417e-02 -3.85922492e-01 -3.33851963e-01
-1.18018854e-02 7.73321629e-01 -3.73453379e-01 -7.40726292e-02
9.57909703e-01 8.18747044e-01 -3.35382283e-01 1.46400321e+00
3.02045077e-01 6.96198225e-01 -7.48961747e-01 -3.88807565e-01
-6.80578174e-03 1.30423799e-01 2.26661474e-01 1.34107649e+00
6.51672244e-01 4.63805683e-02 -5.50107121e-01 1.18180633e+00
-4.97399837e-01 -4.23255205e-01 -5.79344749e-01 -3.86398137e-01
7.16325283e-01 1.09950221e+00 -1.24906707e+00 -1.42733842e-01
-2.61752188e-01 1.12415528e+00 1.56422570e-01 4.84653026e-01
-8.46803010e-01 -7.31635153e-01 6.16645575e-01 -9.02844295e-02
5.52402079e-01 -2.69713610e-01 -8.16827655e-01 -8.87447834e-01
2.00081602e-01 -4.33101892e-01 -7.62378573e-02 -4.86010790e-01
-1.10635149e+00 6.43783450e-01 -6.73387825e-01 -9.78044808e-01
-1.11836411e-01 -9.28560436e-01 -8.23916912e-01 5.38679659e-01
-1.65696609e+00 -4.72455025e-01 -3.55073333e-01 7.62870610e-01
1.34980232e-01 1.94933981e-01 1.00745034e+00 3.58242273e-01
-4.19742912e-01 5.67626119e-01 4.77121294e-01 2.28237197e-01
3.07884932e-01 -1.07333171e+00 8.96765053e-01 7.13183880e-01
2.37935066e-01 4.31868494e-01 5.33209324e-01 -2.38573030e-01
-1.81881583e+00 -1.13418198e+00 6.34911954e-01 -1.47121534e-01
7.23186016e-01 -7.11069763e-01 -1.14519382e+00 4.57774013e-01
7.67783299e-02 3.63231838e-01 3.89996767e-01 -6.20439768e-01
-2.11731434e-01 -2.79901415e-01 -1.11912096e+00 3.73933822e-01
1.39187968e+00 -4.64470178e-01 -4.11733299e-01 6.40565008e-02
3.06592852e-01 -3.28814268e-01 -6.64197981e-01 5.49066961e-01
5.48040509e-01 -7.31635153e-01 9.87344921e-01 -2.43781209e-01
9.20185745e-02 -3.48800033e-01 -9.98958424e-02 -1.18669093e+00
-1.23471238e-01 -4.11560118e-01 -4.08129275e-01 9.39228535e-01
3.06738675e-01 -9.87305522e-01 8.28986466e-01 4.92930748e-02
-1.02774665e-01 -5.10069549e-01 -1.35596693e+00 -1.05592704e+00
-7.89863467e-02 -3.11384261e-01 3.86553138e-01 4.58656490e-01
-3.31100702e-01 -2.35078171e-01 2.92389870e-01 1.24714985e-01
8.32123876e-01 -6.27248138e-02 -1.19800933e-01 -1.61493683e+00
-1.23338953e-01 -7.96778381e-01 -9.25423384e-01 -7.49531269e-01
7.02307224e-02 -8.90259802e-01 4.90998775e-01 -1.36228311e+00
1.78332999e-02 -1.59380421e-01 -9.55887735e-01 6.12279534e-01
3.96981418e-01 7.03421891e-01 -2.99012065e-01 -1.08260460e-01
-3.41985643e-01 2.86415726e-01 7.24961340e-01 -9.68719870e-02
-6.69417754e-02 -8.77783522e-02 -4.91064601e-02 6.39957726e-01
8.30440164e-01 -7.41704464e-01 -7.66617507e-02 -5.25190473e-01
5.23328125e-01 -1.59396812e-01 8.19894969e-01 -1.54079854e+00
9.44020748e-01 4.76444401e-02 6.39162123e-01 -6.04745269e-01
3.74727964e-01 -7.45008767e-01 5.08801937e-01 8.86386037e-01
-2.81804711e-01 1.78469449e-01 5.53367555e-01 7.21385717e-01
-2.13344693e-01 4.62654755e-02 7.35911846e-01 3.81604359e-02
-7.62160540e-01 2.87444741e-01 -8.26877773e-01 1.72973856e-01
9.70039368e-01 -6.19361162e-01 -5.56123257e-01 2.41219744e-01
-4.95307028e-01 -1.85948581e-01 3.50506246e-01 -6.91875257e-03
7.34070897e-01 -1.35809755e+00 -4.08872604e-01 4.88630235e-01
6.96494197e-03 -2.90414598e-02 -1.38385296e-02 9.21597779e-01
-3.38897943e-01 2.67359883e-01 -9.62063015e-01 -8.09504211e-01
-5.74681520e-01 3.89471233e-01 5.83438993e-01 1.64955378e-01
-4.84936565e-01 8.15837562e-01 -1.58490866e-01 -6.84527308e-02
4.88899767e-01 -7.39137530e-01 1.94957051e-02 -3.21990512e-02
5.39711595e-01 3.15311760e-01 4.25337106e-01 -4.06136326e-02
-6.35933876e-01 4.06274438e-01 1.97392553e-01 5.96951135e-02
1.58383799e+00 3.83778632e-01 -2.66176760e-01 7.62811363e-01
1.07792771e+00 -6.14603817e-01 -1.47861123e+00 -1.54134482e-01
1.91819787e-01 3.70609134e-01 6.87645674e-02 -7.57742286e-01
-1.16544068e+00 1.12014425e+00 9.12199378e-01 9.69727486e-02
1.44402182e+00 -4.27947730e-01 1.14362037e+00 8.56772661e-01
6.26324356e-01 -9.31949079e-01 -8.28041136e-02 6.70400560e-01
3.86827201e-01 -8.26146722e-01 -8.58811975e-01 1.52660459e-01
1.51364639e-01 1.44486880e+00 4.23847020e-01 -5.70509613e-01
5.33773243e-01 1.16695118e+00 -2.91747600e-01 4.18747701e-02
-9.57622707e-01 -1.00771971e-01 -1.56393647e-01 6.53463721e-01
1.85612187e-01 -1.02460273e-01 -2.36395434e-01 6.80730641e-01
2.92372942e-01 7.79483497e-01 6.02388263e-01 1.11858928e+00
-3.54177326e-01 -9.02255654e-01 -1.44519852e-02 7.10949242e-01
-4.63302344e-01 -5.12133539e-01 -2.02571869e-01 2.09480256e-01
2.19329190e-03 4.43025261e-01 5.41654050e-01 -3.10804546e-01
4.49286610e-01 2.13917136e-01 7.04121530e-01 -2.62438506e-01
-7.15758920e-01 -4.23011541e-01 -4.28996682e-01 -7.38536775e-01
-4.38502461e-01 -5.25729597e-01 -2.05455160e+00 -3.80462527e-01
-1.09487787e-01 -3.99719775e-01 1.13129759e+00 8.21190059e-01
5.20593822e-01 1.02442193e+00 2.76231229e-01 -9.80894268e-01
-5.14323115e-01 -4.02449161e-01 -4.34275687e-01 1.77899808e-01
4.30587143e-01 -4.18836266e-01 -4.45824444e-01 1.76554441e-01] | [8.201155662536621, 2.483377456665039] |
bb5f3b2e-c87f-41b8-85b5-e0c003facb9f | triplettrack-3d-object-tracking-using-triplet | 2210.16204 | null | https://arxiv.org/abs/2210.16204v1 | https://arxiv.org/pdf/2210.16204v1.pdf | TripletTrack: 3D Object Tracking using Triplet Embeddings and LSTM | 3D object tracking is a critical task in autonomous driving systems. It plays an essential role for the system's awareness about the surrounding environment. At the same time there is an increasing interest in algorithms for autonomous cars that solely rely on inexpensive sensors, such as cameras. In this paper we investigate the use of triplet embeddings in combination with motion representations for 3D object tracking. We start from an off-the-shelf 3D object detector, and apply a tracking mechanism where objects are matched by an affinity score computed on local object feature embeddings and motion descriptors. The feature embeddings are trained to include information about the visual appearance and monocular 3D object characteristics, while motion descriptors provide a strong representation of object trajectories. We will show that our approach effectively re-identifies objects, and also behaves reliably and accurately in case of occlusions, missed detections and can detect re-appearance across different field of views. Experimental evaluation shows that our approach outperforms state-of-the-art on nuScenes by a large margin. We also obtain competitive results on KITTI. | ['Luc van Gool', 'Marc Proesmans', 'Nicola Marinello'] | 2022-10-28 | null | null | null | null | ['3d-object-tracking'] | ['computer-vision'] | [-4.44709241e-01 -3.55284512e-01 -1.64848208e-01 -3.59530509e-01
-3.68538588e-01 -7.54877985e-01 6.79409444e-01 1.07428104e-01
-7.15044320e-01 8.30438957e-02 -3.06670010e-01 -1.02968693e-01
1.76662147e-01 -5.80400646e-01 -7.45056570e-01 -6.91454530e-01
6.15833998e-02 5.14756143e-01 1.10587919e+00 -1.59668431e-01
8.04999098e-02 1.19089723e+00 -1.94378221e+00 -2.79709399e-01
7.67216161e-02 1.02099526e+00 2.67558932e-01 8.07888865e-01
2.29445305e-02 4.73834366e-01 -2.43972376e-01 -2.00068012e-01
4.08126980e-01 1.76457033e-01 -4.68635000e-02 3.93520027e-01
7.74992645e-01 -3.76045942e-01 -5.40886045e-01 1.10592699e+00
1.08478367e-01 1.33260056e-01 6.61508977e-01 -1.65928400e+00
-3.47297519e-01 -4.50856835e-01 -5.53755820e-01 4.51401949e-01
1.78318113e-01 2.61153281e-01 6.87552094e-01 -1.01568091e+00
7.95554698e-01 1.19604969e+00 7.10897386e-01 5.88326693e-01
-8.88499975e-01 -5.25396407e-01 3.83703768e-01 4.40878361e-01
-1.26028371e+00 -4.55455333e-01 6.35398626e-01 -6.62131011e-01
8.73158753e-01 1.11527056e-01 6.30258381e-01 7.85126448e-01
3.42638731e-01 7.27394581e-01 6.48453414e-01 -2.25319624e-01
8.24690908e-02 5.58573246e-01 1.84486315e-01 7.78376043e-01
5.85623324e-01 5.01372516e-01 -4.12979335e-01 -5.61222434e-02
3.67676854e-01 3.59345734e-01 9.34313387e-02 -1.30950034e+00
-1.17365706e+00 7.48692155e-01 5.35961509e-01 1.62511602e-01
-2.01724738e-01 4.65803623e-01 1.62108079e-01 1.52672738e-01
3.79085511e-01 -9.87570286e-02 -1.39996216e-01 -7.22908750e-02
-4.84922469e-01 4.37492758e-01 3.68021965e-01 1.37367940e+00
8.17044973e-01 -1.12662211e-01 6.27161711e-02 3.66172820e-01
5.66041827e-01 9.23241556e-01 1.74756646e-01 -9.32700276e-01
2.06770703e-01 7.70843267e-01 4.57428426e-01 -9.06751573e-01
-2.41524383e-01 -1.09474286e-01 -1.50637895e-01 8.08815777e-01
2.40267381e-01 3.52776110e-01 -8.74549747e-01 1.49058366e+00
7.75785565e-01 2.32049808e-01 5.10576293e-02 1.00412655e+00
6.83639109e-01 3.21650982e-01 -1.79422811e-01 1.39183849e-01
1.36797512e+00 -8.00116718e-01 -5.39026678e-01 -2.65856713e-01
5.33703089e-01 -6.89989686e-01 1.73593774e-01 -1.86233729e-01
-8.84482145e-01 -7.23453581e-01 -1.11791074e+00 -4.53512110e-02
-7.64846444e-01 2.59489596e-01 2.27311701e-01 8.40979278e-01
-1.00392282e+00 1.84776738e-01 -1.10293734e+00 -7.70413160e-01
2.32609093e-01 5.11904657e-01 -6.63679123e-01 -1.91131487e-01
-6.24532580e-01 1.43002975e+00 3.15318346e-01 -3.36165652e-02
-9.08580422e-01 -3.87446135e-01 -1.00269771e+00 -3.19265038e-01
2.07647249e-01 -5.17414331e-01 1.11879480e+00 -3.60991299e-01
-1.20564878e+00 1.12742448e+00 -3.94627333e-01 -5.22474051e-01
5.42420328e-01 -2.29086027e-01 -4.19733822e-01 5.84409386e-02
7.38110170e-02 7.96380758e-01 9.33374047e-01 -1.24122381e+00
-1.15303755e+00 -6.95220649e-01 -1.12187669e-01 1.71181723e-01
-1.42118961e-01 -1.09417075e-02 -5.82233489e-01 4.18302566e-02
1.86848283e-01 -1.41523731e+00 -1.49876118e-01 5.84121048e-01
1.30680516e-01 -4.45041269e-01 1.46182036e+00 -2.53410563e-02
6.04129553e-01 -2.24334407e+00 -6.50380924e-02 4.15014923e-02
3.66936773e-01 3.90160859e-01 6.57623410e-02 1.77089393e-01
3.20336252e-01 -4.42429006e-01 1.78677231e-01 -4.24210787e-01
1.08073056e-01 1.85769260e-01 -1.01850353e-01 9.49474514e-01
3.72024417e-01 8.63755345e-01 -9.82838452e-01 -3.44960064e-01
5.99423409e-01 4.52570260e-01 -2.76386589e-01 1.04020320e-01
-1.12950586e-01 1.23064898e-01 -6.99855089e-01 6.11803472e-01
6.91626668e-01 7.89983571e-03 -3.32071215e-01 -2.80420166e-02
-4.86631125e-01 -3.24406773e-02 -1.12004018e+00 1.43189299e+00
-5.09904847e-02 9.01387334e-01 -1.10332988e-01 -6.63589895e-01
1.07267892e+00 1.50251925e-01 3.84502918e-01 -4.53540087e-01
2.36296788e-01 2.55361855e-01 -1.55426994e-01 -4.63363439e-01
7.01643467e-01 2.40756929e-01 -7.66294822e-02 2.01952383e-01
-6.28365502e-02 -9.42497998e-02 9.63898748e-02 4.50606532e-02
1.20776451e+00 2.76654184e-01 1.84849218e-01 -2.44685058e-02
5.39824843e-01 3.19184393e-01 2.96199471e-01 5.05600929e-01
-5.13031185e-01 3.64309102e-01 -2.69105621e-02 -5.50073504e-01
-1.16370356e+00 -1.05538738e+00 -2.24302068e-01 7.46115685e-01
7.33392060e-01 1.68895500e-03 -1.79213673e-01 -7.35564470e-01
5.97568810e-01 5.21050513e-01 -8.30455005e-01 -3.20066243e-01
-5.54956377e-01 -1.78916901e-01 6.39057234e-02 8.35135579e-01
2.12353438e-01 -5.25266767e-01 -1.17549646e+00 1.68570325e-01
3.61735255e-01 -1.43736374e+00 -3.48732233e-01 2.46990755e-01
-7.74564087e-01 -9.66001689e-01 -4.94041026e-01 -6.86263680e-01
5.94233930e-01 1.09092391e+00 8.43493104e-01 -1.63180634e-01
-4.40200776e-01 7.59888411e-01 -3.03254068e-01 -6.56412721e-01
-4.31944430e-01 -2.46031076e-01 3.73631179e-01 1.67931780e-01
9.30884063e-01 -1.65287048e-01 -5.25234640e-01 6.64632618e-01
-4.01883781e-01 -3.82533967e-01 5.06870687e-01 4.53042537e-01
5.51354289e-01 -3.29800785e-01 2.58045588e-02 -3.85008514e-01
-2.15619817e-01 -2.80183405e-01 -1.16143048e+00 3.95361744e-02
-2.13444442e-01 4.54889797e-02 -2.21230695e-03 -5.99815190e-01
-6.49623573e-01 5.96780658e-01 3.02500099e-01 -1.06847727e+00
-2.88996220e-01 -3.29142332e-01 -3.66421826e-02 -4.86490428e-01
4.82464671e-01 4.09026779e-02 7.19697475e-02 -4.34105456e-01
5.10990798e-01 6.06523931e-01 3.94110680e-01 -2.60847136e-02
1.22905314e+00 1.00757849e+00 2.27422759e-01 -9.80008304e-01
-5.96684515e-01 -1.14289594e+00 -8.93369615e-01 -5.15900552e-01
1.07445478e+00 -1.10842621e+00 -7.81397104e-01 1.56797752e-01
-1.29659081e+00 1.63070783e-01 -1.55342072e-01 7.97143579e-01
-4.84922647e-01 7.92744756e-03 -7.31889606e-02 -1.00958681e+00
2.55586773e-01 -1.05002284e+00 1.45891941e+00 2.16029301e-01
7.52655268e-02 -7.57642806e-01 1.96651995e-01 1.09360754e-01
1.94752127e-01 3.13610971e-01 3.10558558e-01 -4.49354351e-01
-1.04772592e+00 -6.74996853e-01 -3.07415456e-01 1.10619374e-01
8.54997784e-02 -8.26074854e-02 -1.23836732e+00 -3.07045668e-01
-1.72341228e-01 7.68531635e-02 8.85800898e-01 2.51339346e-01
4.45230752e-01 4.65234697e-01 -8.57653260e-01 3.62035066e-01
1.27987218e+00 2.94291258e-01 2.48489872e-01 3.18548650e-01
6.57410443e-01 5.99918365e-01 8.47011983e-01 1.76100463e-01
4.21414107e-01 9.75134134e-01 6.96768880e-01 1.26844063e-01
-2.49235928e-01 -1.07157253e-01 4.70906675e-01 4.25428480e-01
1.50145069e-01 -4.99251820e-02 -7.56612957e-01 8.12219977e-01
-1.99244130e+00 -9.87652719e-01 -3.16367060e-01 2.29749870e+00
9.79023278e-02 3.36008340e-01 2.00731114e-01 -4.10429835e-02
7.46902525e-01 -7.22398236e-02 -6.87340379e-01 -4.11741473e-02
1.10566363e-01 -3.38047266e-01 1.01263309e+00 3.75166774e-01
-1.26249123e+00 8.98972332e-01 5.78639555e+00 2.32520297e-01
-9.25008893e-01 2.28489444e-01 4.85691242e-03 -1.17102653e-01
1.69548854e-01 -1.17910661e-01 -1.54895079e+00 2.59138077e-01
6.37252688e-01 -8.97265077e-02 -2.81821579e-01 1.03496969e+00
-1.02164261e-01 -1.35281175e-01 -1.26313961e+00 9.39978778e-01
4.03918594e-01 -1.15113080e+00 -3.89186472e-01 3.23961407e-01
6.04907453e-01 3.11239958e-01 1.01430781e-01 8.77081081e-02
2.57741064e-01 -4.84565675e-01 9.64097083e-01 3.72350246e-01
3.51426691e-01 -4.46950644e-01 7.25717962e-01 3.41151536e-01
-1.55015445e+00 -5.93984202e-02 -5.92210472e-01 1.23835817e-01
1.44442394e-01 1.43746614e-01 -9.41893458e-01 2.46033669e-01
8.55368912e-01 9.12537217e-01 -6.37890935e-01 1.45323443e+00
1.02578804e-01 -2.29773913e-02 -5.11058092e-01 -3.57393950e-01
3.80661726e-01 2.59398706e-02 8.08805645e-01 1.00930619e+00
3.71964008e-01 -5.97226666e-03 3.18623960e-01 6.93031490e-01
1.56479910e-01 -3.90810370e-01 -1.12651563e+00 2.43107423e-01
3.30591679e-01 1.17043912e+00 -7.20306754e-01 -2.75828868e-01
-6.80314898e-01 6.73610091e-01 1.72985658e-01 1.40053287e-01
-8.60378683e-01 -6.80398494e-02 1.09411681e+00 2.71881223e-01
9.44874763e-01 -6.18919790e-01 3.98872167e-01 -8.85923386e-01
1.91976145e-01 -8.83551463e-02 2.87164673e-02 -8.10486972e-01
-1.04321623e+00 4.97574747e-01 7.59514049e-02 -1.63537478e+00
-1.22117668e-01 -8.95865202e-01 -2.76160508e-01 5.23110628e-01
-1.80622387e+00 -1.14270234e+00 -5.88451922e-01 4.82603073e-01
4.68094051e-01 -6.55886680e-02 4.72406149e-01 2.45915756e-01
-1.21535003e-01 1.41045585e-01 2.29741365e-01 -4.56971675e-03
6.36594296e-01 -1.00138068e+00 6.07539594e-01 8.13908756e-01
3.12105209e-01 2.58079380e-01 7.60473728e-01 -5.39525151e-01
-1.70683026e+00 -1.26210749e+00 8.12533081e-01 -1.09820390e+00
6.87826693e-01 -6.09423280e-01 -7.61872172e-01 6.67442024e-01
-1.10396095e-01 5.36556840e-01 2.78644323e-01 -2.34802499e-01
-3.33673120e-01 -3.09667766e-01 -9.44437146e-01 3.53270590e-01
1.19785655e+00 -3.10088903e-01 -5.75771093e-01 2.69164711e-01
6.55931294e-01 -6.31924391e-01 -4.57603395e-01 2.87571818e-01
6.72797084e-01 -9.08558190e-01 1.07281709e+00 -5.86311400e-01
-5.45504630e-01 -7.98484147e-01 -3.34685504e-01 -8.14377606e-01
-4.42898512e-01 -2.67697096e-01 -2.07019582e-01 7.62591600e-01
4.22594957e-02 -5.65479994e-01 9.12780821e-01 3.91671032e-01
-7.97879323e-02 -2.97469795e-01 -1.12891757e+00 -1.17060757e+00
-4.47134227e-01 -2.94811606e-01 3.26502591e-01 3.98474276e-01
-5.39378703e-01 1.20281123e-01 -1.26350954e-01 6.84518456e-01
8.82998466e-01 3.14157814e-01 1.08011043e+00 -1.54330063e+00
1.88218728e-01 -3.66575718e-01 -1.30362380e+00 -1.17457807e+00
1.54293090e-01 -6.44024968e-01 2.87328154e-01 -1.14639878e+00
7.85773173e-02 -4.05067116e-01 -2.70976186e-01 1.46509513e-01
9.32059512e-02 4.46545422e-01 1.81568369e-01 1.14348158e-01
-8.96748364e-01 6.69854879e-01 8.58438849e-01 -1.45789430e-01
1.16037149e-02 1.04720555e-01 4.32736613e-03 6.04839385e-01
5.48467040e-01 -6.14909887e-01 -1.71752587e-01 -5.28784752e-01
-1.64097697e-01 -2.76615560e-01 8.06023896e-01 -1.24927747e+00
4.77457047e-01 6.90362975e-02 5.37298322e-01 -9.91846263e-01
8.64492655e-01 -1.28008127e+00 2.66208928e-02 4.61367160e-01
2.41933782e-02 3.78490537e-01 4.64663476e-01 9.61840987e-01
-8.34144801e-02 -1.76133007e-01 8.50788414e-01 5.52592240e-02
-1.10238326e+00 4.84931588e-01 -3.60764295e-01 -2.90824860e-01
1.54324460e+00 -6.39351547e-01 -6.36334270e-02 5.85797569e-03
-3.85071993e-01 2.63155252e-01 7.55105376e-01 9.20075178e-01
6.53155208e-01 -1.60253656e+00 -5.59693575e-01 4.06013846e-01
6.37299478e-01 -3.95186394e-01 -1.00731276e-01 7.66977489e-01
-3.97219568e-01 7.07854390e-01 -1.38399765e-01 -1.18958664e+00
-1.66923869e+00 7.46912181e-01 3.09920728e-01 3.67813557e-01
-6.99348330e-01 6.92795455e-01 2.72204906e-01 -1.62896663e-01
3.99372756e-01 -4.96544838e-01 -3.18164863e-02 -1.25836164e-01
4.73077506e-01 2.45662346e-01 1.55807793e-01 -1.03317177e+00
-7.53567815e-01 1.05184960e+00 -7.15710074e-02 -4.93372045e-02
1.04112589e+00 -1.91981047e-01 4.59245712e-01 6.38097703e-01
1.25746119e+00 -1.91510275e-01 -1.53796303e+00 -4.45795089e-01
1.24823809e-01 -7.99013793e-01 1.92937985e-01 -1.89145178e-01
-7.76142359e-01 9.21571553e-01 1.12742412e+00 1.42110884e-01
4.74235564e-01 3.53195995e-01 5.72265923e-01 5.06181538e-01
5.92543125e-01 -7.19066381e-01 1.60292193e-01 4.47968304e-01
4.17342901e-01 -1.48563993e+00 3.09853572e-02 -2.62793571e-01
-4.88009334e-01 9.12138283e-01 4.78730053e-01 -3.41097027e-01
6.13424182e-01 1.05372749e-01 1.31903380e-01 -2.24296391e-01
-8.27207744e-01 -6.90048158e-01 2.92074710e-01 7.47190952e-01
-1.56644061e-01 -9.01906490e-02 4.20352966e-01 -2.15016529e-01
3.09161097e-01 -3.79726708e-01 2.14890495e-01 1.05535495e+00
-8.36643219e-01 -9.89061832e-01 -4.84938622e-01 8.00524354e-02
6.52859360e-02 4.73479837e-01 -3.01076412e-01 1.07387018e+00
2.20903799e-01 7.87918746e-01 2.71157265e-01 -2.70630330e-01
6.02474630e-01 5.04133813e-02 6.90793633e-01 -5.00306427e-01
-1.52946368e-01 -5.30428588e-02 -1.59750879e-01 -6.86866224e-01
-6.07200742e-01 -1.05824971e+00 -9.21567142e-01 -1.85420457e-02
-6.46457553e-01 -1.11118354e-01 1.07226038e+00 6.78782165e-01
3.81747156e-01 2.76808232e-01 6.03001595e-01 -1.04407263e+00
-5.19215286e-01 -5.47834516e-01 -4.27573264e-01 2.84401894e-01
7.20286191e-01 -1.27043712e+00 -3.38291556e-01 1.27714187e-01] | [6.682973384857178, -2.1865553855895996] |
9f5042de-b6fd-4962-9b63-98022abaea28 | pretrain-kges-learning-knowledge | null | null | https://openreview.net/forum?id=HJlv-Fz-pS | https://openreview.net/pdf?id=HJlv-Fz-pS | Pretrain-KGEs: Learning Knowledge Representation from Pretrained Models for Knowledge Graph Embeddings | Learning knowledge graph embeddings (KGEs) is an efficient approach to knowledge graph completion. Conventional KGEs often suffer from limited knowledge representation, which causes less accuracy especially when training on sparse knowledge graphs. To remedy this, we present Pretrain-KGEs, a training framework for learning better knowledgeable entity and relation embeddings, leveraging the abundant linguistic knowledge from pretrained language models. Specifically, we propose a unified approach in which we first learn entity and relation representations via pretrained language models and use the representations to initialize entity and relation embeddings for training KGE models. Our proposed method is model agnostic in the sense that it can be applied to any variant of KGE models. Experimental results show that our method can consistently improve results and achieve state-of-the-art performance using different KGE models such as TransE and QuatE, across four benchmark KG datasets in link prediction and triplet classification tasks. | ['Bin He', 'Xu sun', 'Qi Su', 'Yi Zhang', 'Xiaoqian Liu', 'Zhiyuan Zhang'] | 2019-12-01 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-2.95004159e-01 2.65888095e-01 -6.91613376e-01 -1.80136517e-01
-2.53851146e-01 -3.53174120e-01 3.88393641e-01 3.68733943e-01
-3.91201049e-01 6.82807446e-01 1.67177990e-01 -3.01373571e-01
-3.94088715e-01 -1.04425728e+00 -9.07327473e-01 -1.74304232e-01
-1.38611585e-01 6.59951150e-01 1.57494739e-01 -2.66316026e-01
-3.67166191e-01 1.95939735e-01 -1.04434526e+00 -9.60770622e-02
1.04467177e+00 8.63287151e-01 -2.17958346e-01 2.16467977e-01
-4.07933414e-01 1.10567141e+00 -7.90144727e-02 -1.02600253e+00
2.36487910e-02 1.42226964e-01 -9.46231663e-01 -3.30476612e-01
3.11250240e-01 -1.28454104e-01 -8.59875143e-01 8.47741187e-01
3.35908741e-01 2.40376487e-01 7.65870154e-01 -1.41398406e+00
-1.57685292e+00 1.09231746e+00 -3.43524009e-01 2.83533689e-02
3.40193987e-01 -2.68419921e-01 1.50685251e+00 -1.17392492e+00
5.77950835e-01 1.19852233e+00 8.03734660e-01 4.12458241e-01
-1.22959352e+00 -5.03603220e-01 3.18462163e-01 7.38938153e-01
-1.80212903e+00 -2.63532162e-01 8.01327109e-01 -4.32453722e-01
1.26403582e+00 -2.86435992e-01 5.84768534e-01 9.29241955e-01
-3.01812530e-01 8.25779438e-01 6.50048792e-01 -5.20552754e-01
-1.67081505e-01 7.35344142e-02 5.43093979e-01 1.11320388e+00
6.55642211e-01 -1.04399554e-01 -7.28572607e-01 -1.74586207e-01
5.82251608e-01 -2.13954911e-01 -3.91636401e-01 -8.13358545e-01
-1.21338725e+00 7.35786438e-01 7.12540090e-01 6.98749945e-02
-3.51122826e-01 3.43304336e-01 4.03362513e-01 4.35843349e-01
3.47767532e-01 2.87235856e-01 -6.15673006e-01 2.14398608e-01
-2.80387014e-01 -5.74935526e-02 9.81695533e-01 1.01754832e+00
9.90201712e-01 3.32507193e-02 -2.02539966e-01 9.47644770e-01
3.84499341e-01 3.05721402e-01 2.49621183e-01 -2.48511195e-01
5.89343429e-01 8.46894205e-01 -7.74194226e-02 -9.92633343e-01
-2.35413760e-01 -4.37746972e-01 -5.52529633e-01 -5.04986405e-01
5.90227195e-04 -8.66746828e-02 -1.00377095e+00 1.68617499e+00
3.93333942e-01 5.41600168e-01 4.56921160e-01 6.25623345e-01
1.24631035e+00 4.21689481e-01 2.86979377e-01 1.32522330e-01
1.10828733e+00 -1.13063383e+00 -7.10825741e-01 -1.27964601e-01
8.53658915e-01 -3.19718659e-01 9.09742415e-01 -4.83032921e-03
-4.51183140e-01 -3.41153055e-01 -1.01890028e+00 -2.91005760e-01
-9.56009805e-01 1.76039904e-01 1.06326389e+00 5.60171068e-01
-9.62862492e-01 3.38895410e-01 -9.12817180e-01 -4.56693232e-01
6.00992382e-01 3.44730824e-01 -6.32030070e-01 -4.59813952e-01
-1.49021709e+00 1.07066810e+00 1.23961020e+00 1.77613303e-01
-5.84335566e-01 -1.00787759e+00 -1.34341657e+00 2.91204751e-01
7.04217672e-01 -9.34424877e-01 5.56287944e-01 -2.90736169e-01
-1.40130985e+00 5.52360773e-01 1.20654374e-01 -5.01805305e-01
-1.81966741e-02 -2.92745382e-01 -7.75175869e-01 -7.13186190e-02
-1.88977152e-01 4.77349937e-01 4.66093630e-01 -1.15430903e+00
-8.37034732e-02 -1.17734298e-01 1.60064921e-01 5.29338606e-03
-8.91909063e-01 -2.34761119e-01 -9.06584680e-01 -5.31440854e-01
-2.86850482e-01 -6.88487530e-01 8.82484466e-02 -1.26493737e-01
-6.18405640e-01 -6.79051578e-01 7.87544727e-01 -6.98180974e-01
1.39878476e+00 -1.69579089e+00 4.01078105e-01 5.28590262e-01
4.14544493e-01 6.73667729e-01 -3.56792122e-01 5.51315546e-01
-1.08248942e-01 1.24221005e-01 5.04495054e-02 -2.36333579e-01
4.21561778e-01 6.48830771e-01 -1.14089735e-01 -1.82582457e-02
3.40520203e-01 1.53404880e+00 -1.18130827e+00 -6.61073923e-01
6.44973293e-02 7.37366199e-01 -4.99797910e-01 1.25852123e-01
-3.97671282e-01 -9.95614603e-02 -6.27295554e-01 8.00382853e-01
4.65006620e-01 -6.01787031e-01 6.53401673e-01 -5.15178323e-01
5.36134481e-01 1.90528169e-01 -1.12956417e+00 1.73857677e+00
-4.92621720e-01 3.08421016e-01 -4.97142076e-01 -1.19761109e+00
7.99381614e-01 1.44688874e-01 2.21347079e-01 -2.70124882e-01
6.65445104e-02 1.67967692e-01 -1.98829934e-01 -4.62247342e-01
4.95130211e-01 9.37996954e-02 4.31897156e-02 2.79288828e-01
5.61040878e-01 3.77880812e-01 4.09386426e-01 6.33362830e-01
1.01641631e+00 1.94280297e-01 3.73079807e-01 1.54990539e-01
4.59839851e-01 -3.06099355e-01 7.17305899e-01 4.40452129e-01
6.11094460e-02 -1.73215628e-01 5.66952586e-01 -3.30070108e-01
-4.45537806e-01 -1.14607096e+00 1.49485096e-01 1.02848995e+00
5.88872805e-02 -9.03371334e-01 -2.04893366e-01 -1.21205056e+00
5.04676640e-01 6.13353133e-01 -7.24688053e-01 -4.36737925e-01
-2.95793533e-01 -7.01999843e-01 4.84931439e-01 8.64125073e-01
3.41151655e-01 -8.03095222e-01 3.83361697e-01 2.19157308e-01
7.88203552e-02 -1.55112803e+00 -3.45706373e-01 -1.53301433e-01
-5.51962137e-01 -1.38886249e+00 -3.19319546e-01 -9.25573111e-01
6.99555337e-01 2.60669459e-02 1.28910291e+00 1.74418822e-01
-1.94408536e-01 7.31158137e-01 -5.81912160e-01 -2.18357414e-01
-3.77638228e-02 2.15385288e-01 2.07654044e-01 1.41181082e-01
6.11231506e-01 -6.95125937e-01 -2.11179942e-01 -7.68519985e-03
-8.01922321e-01 -2.37475201e-01 6.37395918e-01 9.62616980e-01
8.40274155e-01 1.97651070e-02 9.41530287e-01 -1.16944194e+00
7.38608420e-01 -5.84041059e-01 -5.46707988e-01 9.78744447e-01
-9.26171124e-01 3.62440616e-01 6.22746766e-01 -2.35265315e-01
-8.11664939e-01 -1.45727098e-01 1.13952950e-01 -8.06442678e-01
5.08946776e-01 1.21547055e+00 -2.48562291e-01 -5.46322882e-01
3.99773806e-01 1.87243983e-01 -3.84037346e-01 -6.44537091e-01
1.03734350e+00 3.28699708e-01 5.83333373e-01 -8.83213401e-01
1.08419347e+00 5.40024824e-02 -1.50570750e-01 -5.18117011e-01
-9.41365123e-01 -4.07498568e-01 -6.25922203e-01 1.52327105e-01
5.42109847e-01 -1.03595853e+00 -6.11021519e-01 1.56180263e-01
-8.72989297e-01 -1.59998909e-01 -2.50038177e-01 6.98194325e-01
-1.64668262e-01 5.17996132e-01 -4.54238892e-01 -3.69488418e-01
-3.96349132e-01 -6.13429546e-01 5.44408023e-01 2.15975225e-01
2.99019665e-01 -1.59903622e+00 9.56443325e-02 4.80620265e-01
3.10723960e-01 2.46207565e-01 1.30771387e+00 -8.10917914e-01
-6.93323493e-01 -1.33729473e-01 -4.54155535e-01 4.91158694e-01
2.56393135e-01 -8.14715028e-02 -6.04810596e-01 -2.24576905e-01
-1.18550932e+00 -8.00104797e-01 1.18762147e+00 -3.50431889e-01
1.10864937e+00 -3.07615101e-01 -6.12440705e-01 7.26388097e-01
1.64973617e+00 -5.50641119e-01 3.98171812e-01 3.32951471e-02
1.34407997e+00 1.47982612e-01 3.77243668e-01 1.90960467e-01
1.09083068e+00 6.56967103e-01 2.12797329e-01 1.91537708e-01
-4.05189097e-01 -7.68713474e-01 2.78747648e-01 1.28766978e+00
-1.12317823e-01 -2.12660581e-01 -1.03755295e+00 9.80382144e-01
-2.01196575e+00 -5.92472911e-01 1.28567874e-01 1.80955040e+00
1.09373331e+00 -1.78608686e-01 -6.26169890e-02 -1.01312503e-01
5.33115029e-01 2.18417540e-01 -5.12415588e-01 -1.94301963e-01
-1.69981167e-01 4.87911969e-01 5.59249461e-01 3.62413168e-01
-9.65931594e-01 1.47967958e+00 5.77384758e+00 8.41363668e-01
-7.79074848e-01 3.25002581e-01 -2.53539562e-01 2.46037200e-01
-4.17823970e-01 -4.38845865e-02 -7.21795142e-01 1.77076459e-01
7.86322415e-01 -4.57769424e-01 4.84496653e-01 7.70044267e-01
-4.96807843e-01 4.45623964e-01 -1.23458874e+00 1.05579329e+00
1.50125355e-01 -1.50554478e+00 4.31266338e-01 -1.89516500e-01
8.80406320e-01 1.90292373e-01 -1.08799249e-01 9.47994053e-01
8.25944662e-01 -1.01687741e+00 1.07348256e-01 6.11755431e-01
7.45913804e-01 -6.67206049e-01 6.65989101e-01 3.17067169e-02
-1.39253330e+00 2.07611978e-01 -5.21081030e-01 2.94113934e-01
1.15333602e-01 5.32317877e-01 -1.02293861e+00 1.40189755e+00
3.68127584e-01 1.02214634e+00 -7.46183455e-01 9.46101725e-01
-7.08518147e-01 6.42425060e-01 -3.55393410e-01 3.74237373e-02
-3.63497958e-02 -2.79146917e-02 1.45422444e-01 1.10287488e+00
6.34477213e-02 -1.11268394e-01 4.89765406e-01 7.84354091e-01
-7.62814343e-01 2.45083690e-01 -5.67422211e-01 -6.40611291e-01
7.28597879e-01 1.24945033e+00 -2.11402148e-01 -3.56557667e-01
-8.89137089e-01 9.17297184e-01 1.17252529e+00 6.59581125e-01
-6.98488653e-01 -6.42688930e-01 5.72227061e-01 -3.21634680e-01
8.54820728e-01 -2.73930728e-01 3.98125529e-01 -1.59789073e+00
1.11042768e-01 -4.13100630e-01 9.12517488e-01 -6.63655579e-01
-1.74191356e+00 3.36056888e-01 1.39646162e-03 -6.46856666e-01
-1.05647027e-01 -8.17083538e-01 -3.78147840e-01 7.06985474e-01
-2.24969220e+00 -1.77238834e+00 -1.97237611e-01 8.76750767e-01
-3.37087184e-01 -2.74721533e-01 1.01258099e+00 5.70512593e-01
-8.10530484e-01 1.17003274e+00 1.41786069e-01 6.68586731e-01
6.49458945e-01 -1.33932698e+00 2.56639957e-01 6.63850904e-01
6.60031199e-01 8.03073704e-01 5.12813888e-02 -7.13778555e-01
-1.83373225e+00 -1.51539540e+00 1.09701204e+00 -6.27992868e-01
1.13523626e+00 -2.44426787e-01 -1.06997716e+00 1.30879331e+00
-3.13461795e-02 8.08829367e-01 1.05375171e+00 7.95693099e-01
-9.28077519e-01 -1.11478120e-01 -7.65241563e-01 3.56917351e-01
1.32037342e+00 -9.29740548e-01 -8.11971605e-01 4.27964300e-01
8.03034484e-01 -3.99852931e-01 -1.61254549e+00 4.59566206e-01
4.18672472e-01 -2.39529178e-01 1.13527286e+00 -1.09261036e+00
3.93547341e-02 -4.52406585e-01 -8.39686990e-02 -1.53136182e+00
-2.85479575e-01 -4.33434606e-01 -9.91516352e-01 1.43629527e+00
6.21679127e-01 -8.20949912e-01 7.54464269e-01 3.20956975e-01
-7.53837600e-02 -9.95289743e-01 -8.37022007e-01 -1.19012117e+00
-3.67813967e-02 -3.07111561e-01 7.59281158e-01 1.47677732e+00
1.46807373e-01 5.90062916e-01 -2.90746480e-01 4.77240443e-01
6.58853650e-01 3.12996179e-01 8.14884543e-01 -1.32863462e+00
-2.55650192e-01 -8.04312676e-02 -8.32944453e-01 -9.45976079e-01
8.33498061e-01 -1.55207622e+00 -5.29957950e-01 -1.99383342e+00
1.98695347e-01 -5.84879637e-01 -7.93104947e-01 1.04671502e+00
-5.52872181e-01 6.48035407e-02 1.44389244e-02 -8.41698721e-02
-1.00986600e+00 1.03714657e+00 1.08319998e+00 -3.74717712e-01
-1.07048880e-02 -6.17402017e-01 -7.77560055e-01 3.80503029e-01
3.73394102e-01 -3.13946605e-01 -7.43502080e-01 -7.85114884e-01
4.65609729e-01 -4.31138009e-01 4.74889517e-01 -7.25626588e-01
5.10846853e-01 4.41348515e-02 1.72335103e-01 -2.68697143e-01
2.45884225e-01 -7.44666576e-01 2.40790490e-02 -6.16087280e-02
-5.64769879e-02 -3.76890093e-01 2.48087972e-01 1.09297025e+00
-4.08368409e-01 6.63903728e-02 1.10471427e-01 3.04164559e-01
-1.30907774e+00 7.99753129e-01 4.47244436e-01 3.00170034e-01
8.21555257e-01 1.38625264e-01 -6.03913069e-01 -1.24638407e-02
-8.73645127e-01 8.83114278e-01 1.54391006e-01 6.26605332e-01
7.45117188e-01 -1.81305742e+00 -6.23607278e-01 -6.34835362e-02
8.40522945e-01 -8.72755572e-02 1.03977896e-01 7.85403132e-01
-2.32682645e-01 4.98703659e-01 1.28540576e-01 -4.13024463e-02
-8.44729543e-01 8.17668080e-01 2.96290070e-01 -5.38142681e-01
-5.52993834e-01 9.63309467e-01 -1.67987868e-01 -7.67202020e-01
2.12032109e-01 -3.79353940e-01 -3.38691890e-01 -5.50689586e-02
2.49882579e-01 1.70599774e-01 1.15348570e-01 -5.41261375e-01
-6.04373515e-01 6.00484431e-01 -3.75226468e-01 5.63729763e-01
1.28160012e+00 2.41014689e-01 -1.11550830e-01 1.16092600e-01
1.22778165e+00 2.73645166e-02 -7.67307758e-01 -8.74674618e-01
2.08112091e-01 -3.86438936e-01 2.67225593e-01 -9.13895786e-01
-1.26502299e+00 6.21847868e-01 7.79080018e-02 -1.89275697e-01
7.96870351e-01 2.85279572e-01 9.22345102e-01 8.41196835e-01
4.65355247e-01 -9.14332986e-01 -1.67848572e-01 5.56173861e-01
7.04591691e-01 -1.24940038e+00 1.50842741e-01 -7.80600846e-01
-4.34414476e-01 9.10474002e-01 6.76335514e-01 5.28298318e-02
9.13306832e-01 -3.46189469e-01 -2.60374576e-01 -4.80542392e-01
-7.93233395e-01 -6.08702123e-01 6.68255866e-01 9.13989127e-01
2.65038490e-01 2.01661736e-01 -2.37266824e-01 9.76282477e-01
8.78810287e-02 3.03264350e-01 5.34208231e-02 7.98824549e-01
-5.61045948e-03 -1.44090044e+00 3.46476346e-01 3.75979215e-01
-1.31716132e-01 -4.01320964e-01 -5.54578185e-01 1.04200244e+00
-1.47863161e-02 6.83790147e-01 -4.19287294e-01 -6.39278471e-01
5.09634316e-01 3.37094545e-01 6.99368298e-01 -7.96230972e-01
-1.23926163e-01 -7.51569986e-01 4.97373343e-01 -4.26570654e-01
-4.72603381e-01 7.35153444e-03 -1.24102962e+00 -3.50006312e-01
-5.70358872e-01 3.65402967e-01 2.01539844e-01 7.84764647e-01
6.32396996e-01 6.86196506e-01 1.72130942e-01 -1.54901192e-01
-4.01213557e-01 -8.05732071e-01 -6.91668987e-01 3.78278315e-01
-2.91176233e-02 -1.07953870e+00 -1.50160296e-02 1.05911400e-02] | [8.791321754455566, 7.90716552734375] |
10003c79-8a49-42f8-98a5-cd9d0f1dc2ec | qt30-a-corpus-of-argument-and-conflict-in | null | null | https://aclanthology.org/2022.lrec-1.352 | https://aclanthology.org/2022.lrec-1.352.pdf | QT30: A Corpus of Argument and Conflict in Broadcast Debate | Broadcast political debate is a core pillar of democracy: it is the public’s easiest access to opinions that shape policies and enables the general public to make informed choices. With QT30, we present the largest corpus of analysed dialogical argumentation ever created (19,842 utterances, 280,000 words) and also the largest corpus of analysed broadcast political debate to date, using 30 episodes of BBC’s ‘Question Time’ from 2020 and 2021. Question Time is the prime institution in UK broadcast political debate and features questions from the public on current political issues, which are responded to by a weekly panel of five figures of UK politics and society. QT30 is highly argumentative and combines language of well-versed political rhetoric with direct, often combative, justification-seeking of the general public. QT30 is annotated with Inference Anchoring Theory, a framework well-known in argument mining, which encodes the way arguments and conflicts are created and reacted to in dialogical settings. The resource is freely available at http://corpora.aifdb.org/qt30. | ['Chris Reed', 'Ray Becker', 'Kamila Gorska', 'Wassiliki Siskou', 'Zlata Kikteva', 'Annette Hautli-Janisz'] | null | null | null | null | lrec-2022-6 | ['argument-mining'] | ['natural-language-processing'] | [-6.86563328e-02 1.05774009e+00 -4.57070231e-01 -3.99115324e-01
-1.13775909e+00 -1.39956760e+00 1.59510446e+00 7.30763912e-01
-4.31763917e-01 1.15426266e+00 1.70222139e+00 -1.30325091e+00
-2.49101669e-01 -6.26128495e-01 -3.58741581e-01 -4.36321527e-01
5.44329405e-01 8.14166367e-01 5.35526536e-02 -1.05514371e+00
5.06148756e-01 -3.28098208e-01 -9.02795970e-01 7.35418439e-01
7.00753510e-01 6.55080318e-01 -4.04697001e-01 7.65320778e-01
-5.84454596e-01 1.56804800e+00 -8.18074524e-01 -1.00338018e+00
-2.21960455e-01 -5.58666945e-01 -1.60983264e+00 -4.83613491e-01
3.85004133e-01 1.80829644e-01 -1.46719873e-01 8.27349842e-01
8.32082152e-01 -4.36544031e-01 2.98462123e-01 -5.32479346e-01
-2.29060844e-01 1.42603838e+00 -1.67361647e-01 7.17834055e-01
6.38075650e-01 -1.61739469e-01 1.44061458e+00 -5.43361723e-01
1.25461972e+00 1.83788109e+00 3.65611851e-01 2.38781974e-01
-8.99278283e-01 -3.52030426e-01 1.06912449e-01 6.43800423e-02
-1.29890218e-01 -8.69359016e-01 5.65142512e-01 -6.83807850e-01
4.88123655e-01 8.45518053e-01 9.51227069e-01 1.41696358e+00
8.08052495e-02 3.24427009e-01 1.65976727e+00 -5.33793390e-01
2.28815719e-01 3.53406742e-02 2.64414549e-01 5.85577674e-02
3.79332155e-02 -3.23578298e-01 -4.00657475e-01 -9.30973053e-01
-8.08488652e-02 -9.14851367e-01 -2.45443821e-01 3.39330107e-01
-1.32675993e+00 1.47462177e+00 2.08723500e-01 5.13759673e-01
-3.81045192e-01 4.51571010e-02 9.57864285e-01 7.83897042e-01
5.49107552e-01 3.79307330e-01 -3.68105888e-01 -6.54326200e-01
-1.33231863e-01 6.33748055e-01 1.73770869e+00 1.85819805e-01
-8.85805935e-02 -5.30765355e-01 -4.04753014e-02 9.59457397e-01
7.50379086e-01 6.88096046e-01 -6.68354109e-02 -1.39782441e+00
8.38860691e-01 6.65165842e-01 2.76331961e-01 -1.40002537e+00
-1.97561830e-01 -1.61838010e-01 -2.25505292e-01 -1.48932800e-01
7.69492745e-01 -6.13190591e-01 1.32307559e-02 1.58112490e+00
9.02410984e-01 -1.18949068e+00 6.34834588e-01 8.11550856e-01
1.47697997e+00 7.65631855e-01 2.44233627e-02 -5.06724536e-01
2.01133156e+00 -3.28804284e-01 -9.70199466e-01 -1.84803471e-01
3.20070386e-01 -1.35074735e+00 5.83800077e-01 1.55884013e-01
-1.13983643e+00 4.28018838e-01 -6.95677757e-01 -3.67496133e-01
-1.12596095e-01 -5.92741549e-01 5.28415620e-01 3.82051378e-01
-5.78768790e-01 5.24852499e-02 6.82767332e-02 -4.98718500e-01
3.89937341e-01 -3.69816035e-01 2.45322771e-02 4.54791397e-01
-1.72970068e+00 1.30222034e+00 3.47166955e-02 2.27174073e-01
-3.25110674e-01 -2.57443547e-01 -5.35922348e-01 -6.69464111e-01
8.96105289e-01 -4.20129329e-01 1.72263861e+00 -8.20043087e-01
-1.51932311e+00 1.37235272e+00 1.07485078e-01 -6.12860918e-01
8.19225013e-01 -3.01005632e-01 -6.08772337e-01 1.37715369e-01
5.83046198e-01 1.16186373e-01 3.91668767e-01 -1.05173588e+00
-7.16082990e-01 -8.00886471e-03 6.33556604e-01 2.83185303e-01
4.73753244e-01 8.14030111e-01 4.26828414e-01 -3.42807055e-01
1.91802010e-02 -7.14611948e-01 -2.06311449e-01 -3.04865479e-01
-3.58392000e-01 -7.52339661e-01 8.39455605e-01 -1.04064167e+00
1.16129422e+00 -1.55102348e+00 1.14234695e-02 2.69972645e-02
5.15155971e-01 -4.04997058e-02 5.42709291e-01 1.41189539e+00
6.03972971e-02 4.32669938e-01 1.06976822e-01 8.17959130e-01
5.26228905e-01 6.46418929e-01 -1.01168740e+00 7.62486160e-01
-3.76690120e-01 9.42758799e-01 -1.13527369e+00 -7.31021345e-01
2.01691315e-02 -4.06822376e-02 -2.51227409e-01 -4.56080317e-01
-5.27703464e-01 7.59115934e-01 -8.73796046e-01 4.44719225e-01
5.13364226e-02 -3.22237581e-01 7.55395114e-01 -9.27069113e-02
-9.33307350e-01 1.22186553e+00 -5.57824969e-01 1.25103819e+00
-2.38389999e-01 1.25368690e+00 7.18223214e-01 -9.20312345e-01
7.84658730e-01 6.56170368e-01 -1.23216333e-02 -8.87775123e-01
8.53274584e-01 4.30257440e-01 5.95324457e-01 -3.98380905e-01
3.06343734e-01 1.21766031e-02 -8.82651865e-01 7.79148102e-01
-6.25763774e-01 -6.15796804e-01 2.21865714e-01 6.78398907e-01
8.19602907e-01 -2.97447205e-01 8.41791332e-01 -8.96168351e-01
7.56601095e-01 4.22156155e-01 4.68831778e-01 5.22039473e-01
-1.06088378e-01 -2.59668022e-01 1.08446956e+00 -1.01790953e+00
-1.18201876e+00 -3.79621774e-01 -5.94512880e-01 1.08022094e+00
-1.65424466e-01 -6.02930903e-01 -5.58327079e-01 -4.04566944e-01
-1.65127203e-01 9.46688235e-01 -6.33752465e-01 1.08895910e+00
-1.06052387e+00 -3.04585218e-01 3.70078593e-01 -5.78997135e-01
7.26244032e-01 -1.11633861e+00 -1.05393171e+00 5.72342813e-01
-1.05830014e+00 -1.01858401e+00 2.32425272e-01 5.82901239e-02
-1.72423497e-01 -1.70604575e+00 -3.88402194e-01 -3.18028569e-01
3.46520543e-02 -3.76007855e-01 1.27365565e+00 1.18628265e-02
7.58518130e-02 2.06671268e-01 -3.71077180e-01 -8.98016155e-01
-1.34500241e+00 -1.13300554e-01 -5.07038534e-01 -5.08365095e-01
-3.28378081e-01 -3.89927983e-01 -5.21747530e-01 1.26187816e-01
-4.45503086e-01 2.18957618e-01 -7.15884045e-02 7.62561083e-01
-2.00140849e-01 -7.37725139e-01 4.54933733e-01 -1.05157757e+00
1.16041481e+00 -6.55117333e-01 -3.72809172e-01 1.63147449e-01
-1.19243398e-01 -2.28364766e-01 1.34756073e-01 9.76028293e-02
-1.17864609e+00 -1.28956020e+00 -4.69652712e-01 1.14333522e+00
1.66312456e-01 7.01032102e-01 3.85240704e-01 1.81188047e-01
1.11577117e+00 -5.62915862e-01 1.55023843e-01 -3.00152093e-01
8.57108295e-01 7.23802209e-01 4.13970143e-01 -1.02040291e+00
5.72933078e-01 6.17302895e-01 -3.83659363e-01 -9.26976442e-01
-1.32444417e+00 -1.80202872e-01 -2.97376532e-02 -7.33040690e-01
6.56807065e-01 -8.76677513e-01 -1.17004180e+00 -1.61641732e-01
-1.39179730e+00 -3.40491742e-01 -3.65727812e-01 1.05844006e-01
-4.22069252e-01 2.84864783e-01 -5.14323592e-01 -8.55302572e-01
-6.35005295e-01 -6.37431204e-01 5.36656380e-01 1.83709785e-01
-1.01703978e+00 -1.15577662e+00 3.13708156e-01 7.89769351e-01
4.56050456e-01 1.15391624e+00 1.08070862e+00 -6.53417706e-01
3.84899825e-01 4.19692934e-01 -1.30115479e-01 -1.37881011e-01
1.03284568e-01 3.29232737e-02 -5.99169672e-01 1.08638681e-01
3.62980068e-01 -8.14633131e-01 2.57125467e-01 2.31663406e-01
4.99240458e-02 -1.27720892e+00 -1.02070365e-02 -6.85514748e-01
1.07814443e+00 -1.40104480e-02 4.27999824e-01 1.00509858e+00
-2.32013226e-01 1.13005829e+00 4.66446191e-01 3.57751131e-01
8.91088367e-01 4.45905089e-01 2.14589611e-01 6.71012998e-02
-7.05696195e-02 -5.92734702e-02 4.21292894e-02 1.09243548e+00
-3.58597755e-01 -3.86191579e-03 -1.38789821e+00 6.63826108e-01
-2.10837603e+00 -1.32366073e+00 -8.79005373e-01 1.31750071e+00
1.20721674e+00 3.29574108e-01 1.14135658e-02 2.80803800e-01
7.19225347e-01 6.03739798e-01 1.20739914e-01 -1.07852554e+00
-4.94770736e-01 1.78506196e-01 2.71919727e-01 9.80182409e-01
-9.43134546e-01 7.93976486e-01 5.59445190e+00 6.00666940e-01
-5.31533718e-01 3.48935336e-01 8.24727416e-01 2.44217291e-01
-7.23416626e-01 6.21340036e-01 -2.97343880e-01 2.91446984e-01
9.91151392e-01 -7.03683734e-01 -8.06562975e-02 4.09416437e-01
3.43629956e-01 -5.14031351e-01 -4.70750928e-01 4.61926550e-01
-3.80003363e-01 -2.04707837e+00 -4.57896084e-01 1.65426686e-01
6.56222522e-01 1.39594242e-01 -2.77075917e-01 5.22714388e-03
1.28163910e+00 -8.02873135e-01 1.40854788e+00 6.15022294e-02
4.46932495e-01 -2.63139665e-01 6.82239234e-01 5.07689893e-01
-6.42504334e-01 -3.39128554e-01 -5.43936044e-02 -5.25965512e-01
8.59585345e-01 5.10810375e-01 -4.36368048e-01 5.15836179e-01
6.50524914e-01 3.09567839e-01 -4.24464084e-02 1.99387431e-01
-7.99674153e-01 1.10565734e+00 -3.47162873e-01 -5.08237958e-01
9.23138022e-01 -2.54799277e-01 1.09827232e+00 1.08863640e+00
-3.21365952e-01 7.77145565e-01 -1.46393282e-02 -3.79398465e-02
2.31771186e-01 4.19422418e-01 -3.39707851e-01 -7.94140399e-02
6.57074034e-01 1.12456667e+00 -7.30972409e-01 -3.61171663e-01
-5.79760969e-02 -4.61823121e-02 1.13788292e-01 6.39736950e-02
-7.92860031e-01 4.25723195e-01 2.10895136e-01 6.08749203e-02
-1.86987266e-01 6.68821633e-02 1.48284391e-01 -5.40370882e-01
-2.55002499e-01 -1.74100709e+00 7.72559166e-01 -2.31589586e-01
-1.24056995e+00 6.34863853e-01 1.34064972e-01 -4.32931662e-01
-1.77698433e-01 -2.46250108e-01 -6.02182209e-01 5.94057202e-01
-1.14267170e+00 -1.11656594e+00 1.15411207e-01 1.28509328e-01
5.15410185e-01 3.76391202e-01 8.65917206e-01 -1.93593532e-01
-1.54847875e-02 -5.23388565e-01 -2.56837994e-01 3.03753525e-01
6.65895998e-01 -1.00208688e+00 4.58715290e-01 3.90521884e-02
-5.56256883e-02 4.15639579e-01 1.30936813e+00 -4.64869738e-01
-1.22460663e+00 -1.34567037e-01 1.35852027e+00 -6.12424314e-01
1.45543349e+00 -1.52207643e-01 -3.87081563e-01 2.02093139e-01
1.18076909e+00 -7.79314637e-01 9.46557224e-01 1.70024619e-01
-2.88897663e-01 3.94605309e-01 -9.81532335e-01 8.00526679e-01
8.72416556e-01 -4.07794565e-01 -1.53301001e+00 8.18109810e-01
6.43782020e-01 -7.36505985e-01 -7.31954336e-01 9.20519009e-02
7.86112607e-01 -7.27549613e-01 6.44107044e-01 -6.93868876e-01
5.71886539e-01 3.78197469e-02 -4.03457046e-01 -8.81480277e-01
1.67820111e-01 -1.27776778e+00 2.43273348e-01 1.25941432e+00
7.59015918e-01 -1.10501850e+00 1.71931952e-01 5.40981948e-01
-9.28463601e-03 -4.87736434e-01 -1.48267508e+00 2.20351726e-01
4.25565869e-01 -1.80969745e-01 1.33731335e-01 1.28635633e+00
3.07727516e-01 7.09928811e-01 8.93485174e-03 -4.85348612e-01
3.87520343e-01 5.07821560e-01 9.52798128e-01 -1.43809700e+00
6.59272745e-02 -6.42882705e-01 2.84503788e-01 -8.08247566e-01
-2.12302953e-01 -7.19086528e-01 -2.80560613e-01 -1.99599147e+00
-2.54751474e-01 -3.78768951e-01 7.31638074e-01 -1.25921294e-02
5.23452163e-01 -1.84287459e-01 4.46696468e-02 4.21479970e-01
-9.69524458e-02 2.00276017e-01 1.50456178e+00 -3.74378376e-02
-6.18963800e-02 -3.71265471e-01 -1.34282768e+00 1.17308748e+00
7.88572431e-01 -4.70046759e-01 2.82448661e-02 -1.77865416e-01
1.18047774e+00 2.63060510e-01 3.02586228e-01 -3.89834717e-02
2.83965051e-01 -3.90491784e-01 -1.40073910e-01 -7.90192962e-01
4.45962958e-02 -2.99983025e-01 3.91024828e-01 8.64254951e-01
-6.39395535e-01 2.25208178e-01 9.81869623e-02 3.52318734e-01
-3.74423414e-01 -4.14727479e-02 4.21801597e-01 -3.03630620e-01
-2.74795800e-01 -5.04877567e-01 -6.99344814e-01 7.68215716e-01
7.02786565e-01 9.08800364e-02 -1.18334067e+00 -8.56225371e-01
-7.40799963e-01 5.05101323e-01 4.27122675e-02 2.33434036e-01
-7.05142319e-02 -1.10292327e+00 -1.70243275e+00 -1.09999764e+00
-2.85664946e-01 -2.56451547e-01 -3.81782353e-02 1.14196861e+00
-7.90615857e-01 5.36170602e-01 4.46528465e-01 -2.08973721e-01
-1.29387772e+00 -1.87299326e-01 9.03037265e-02 -3.02092940e-01
-1.04460812e+00 7.13101745e-01 -1.62673384e-01 -4.94964331e-01
-2.72869736e-01 -2.19111890e-02 -7.91576862e-01 8.82158697e-01
4.72699523e-01 2.69569039e-01 -5.17409921e-01 -1.07950985e+00
-4.04862136e-01 2.06327781e-01 2.05600694e-01 -8.58713090e-01
1.30329561e+00 -4.10538673e-01 -6.93500161e-01 7.02210009e-01
6.16575778e-01 6.20817006e-01 -3.64435822e-01 -1.97402924e-01
3.52740616e-01 -3.41905236e-01 -2.19018251e-01 -1.11735880e+00
-5.02188280e-02 1.41231745e-01 -3.57027858e-01 1.03924274e+00
4.98250946e-02 5.82923830e-01 4.54719782e-01 2.41725564e-01
6.55705333e-02 -1.35667396e+00 -8.91701728e-02 7.04097748e-01
1.96283519e+00 -9.15092409e-01 2.52825677e-01 -5.22387743e-01
-6.55199111e-01 1.21892357e+00 -1.24775186e-01 1.54224545e-01
5.77060044e-01 1.83999807e-01 7.91632295e-01 -1.04548907e+00
-1.05363870e+00 1.61362141e-01 3.35967578e-02 -9.39514339e-02
6.68187320e-01 2.16825351e-01 -1.55883288e+00 2.19743460e-01
-9.79253411e-01 -8.61713707e-01 5.03405094e-01 9.05816019e-01
-5.89815140e-01 -1.07316864e+00 -6.24795735e-01 6.01165704e-02
-8.52345049e-01 -2.17163309e-01 -1.22611582e+00 1.13119960e+00
-1.16430469e-01 1.50457966e+00 -1.69556007e-01 5.63292921e-01
2.24965110e-01 -1.97873130e-01 2.40107372e-01 -1.90306872e-01
-1.06758380e+00 8.06025043e-02 1.76893640e+00 -2.63516277e-01
-1.05047584e+00 -8.62000048e-01 -9.10766065e-01 -7.75642335e-01
-3.71453732e-01 1.01165783e+00 8.21171761e-01 1.00506127e+00
-9.64584798e-02 1.40663177e-01 5.57440102e-01 -7.39206746e-02
-5.61052084e-01 -1.02599514e+00 1.82071045e-01 -7.37311915e-02
2.00618014e-01 -3.28377336e-01 -4.01377171e-01 -3.12167227e-01] | [9.072022438049316, 9.833664894104004] |
d9c721a2-acbd-43a6-abae-1651b3b5d8fd | clustering-for-graph-datasets-via-gumbel | 2005.02372 | null | https://arxiv.org/abs/2005.02372v2 | https://arxiv.org/pdf/2005.02372v2.pdf | Community Detection Clustering via Gumbel Softmax | Recently, in many systems such as speech recognition and visual processing, deep learning has been widely implemented. In this research, we are exploring the possibility of using deep learning in community detection among the graph datasets. Graphs have gained growing traction in different fields, including social networks, information graphs, the recommender system, and also life sciences. In this paper, we propose a method of community detection clustering the nodes of various graph datasets. We cluster different category datasets that belong to Affiliation networks, Animal networks, Human contact networks, Human social networks, Miscellaneous networks. The deep learning role in modeling the interaction between nodes in a network allows a revolution in the field of science relevant to graph network analysis. In this paper, we extend the gumbel softmax approach to graph network clustering. The experimental findings on specific graph datasets reveal that the new approach outperforms traditional clustering significantly, which strongly shows the efficacy of deep learning in graph community detection clustering. We do a series of experiments on our graph clustering algorithm, using various datasets: Zachary karate club, Highland Tribe, Train bombing, American Revolution, Dolphins, Zebra, Windsurfers, Les Mis\'erables, Political books. | ['Deepak Bhaskar Acharya', 'Huaming Zhang'] | 2020-05-05 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-4.06199276e-01 9.24379304e-02 -1.99176893e-02 -1.54931858e-01
4.72420573e-01 -5.04067838e-01 6.65118098e-01 4.57236588e-01
-3.08211923e-01 4.62464541e-01 2.17024043e-01 -3.77519190e-01
-5.99506319e-01 -9.89902556e-01 -3.43957514e-01 -6.60084069e-01
-9.57526803e-01 7.91978598e-01 1.42170340e-01 -3.23385268e-01
3.55785608e-01 4.78334814e-01 -8.50655019e-01 1.57499626e-01
4.38727558e-01 4.04470414e-01 -1.41110003e-01 6.30604982e-01
-9.11979079e-02 1.27770531e+00 -4.72411603e-01 -4.27109987e-01
3.77738811e-02 -1.61965594e-01 -8.02219927e-01 9.32785794e-02
1.64630674e-02 1.69466987e-01 -8.59004855e-01 1.08413982e+00
3.96995157e-01 6.10321835e-02 7.72206068e-01 -1.62859190e+00
-6.74403012e-01 1.27334654e+00 -9.71473515e-01 2.89279461e-01
2.12364972e-01 -2.58888453e-01 1.03023720e+00 -2.83504456e-01
8.63691032e-01 1.72284746e+00 9.98892486e-01 2.18256563e-01
-8.82543921e-01 -9.03883338e-01 9.30768922e-02 4.70714003e-01
-1.41112411e+00 1.26730621e-01 8.32273424e-01 -8.88898969e-01
6.70529306e-01 -2.17726827e-01 8.26719582e-01 1.01523924e+00
2.41334319e-01 4.70989197e-01 9.32783067e-01 -5.47505505e-02
2.22424507e-01 -3.83611172e-01 3.50046217e-01 9.05853808e-01
4.08837050e-01 5.90540059e-02 -2.58550763e-01 -2.86651999e-01
7.32374012e-01 3.15831959e-01 5.23896143e-02 -4.11525607e-01
-1.02860701e+00 1.39115739e+00 1.01267481e+00 4.17043090e-01
-1.02234073e-01 4.20089662e-01 4.36313510e-01 6.93604410e-01
4.91850019e-01 7.06274211e-02 7.09623545e-02 4.61200058e-01
-6.18286252e-01 -1.53027013e-01 1.17412722e+00 8.01056206e-01
4.37260181e-01 -5.20231482e-03 4.86133218e-01 7.77424812e-01
8.06587875e-01 -2.08851206e-03 3.08425397e-01 -4.39737976e-01
2.65458018e-01 8.02451313e-01 -6.72527969e-01 -1.52775002e+00
-9.84933972e-01 -3.54573965e-01 -1.48671186e+00 8.31716433e-02
1.08412966e-01 -3.70777696e-01 -7.71094620e-01 1.26917946e+00
3.26107144e-01 4.30328310e-01 1.96035355e-02 6.89711154e-01
1.45424342e+00 3.76509309e-01 2.57448740e-02 1.67715982e-01
1.24128234e+00 -7.04649866e-01 -4.08450007e-01 9.42214727e-02
3.05026621e-01 -4.05007392e-01 2.56742358e-01 2.69786924e-01
-2.96509326e-01 -2.79220939e-01 -8.66498470e-01 4.96348143e-01
-5.37663400e-01 -4.46474940e-01 1.17475581e+00 4.02632356e-01
-1.26100171e+00 8.14963818e-01 -6.92567170e-01 -9.52947319e-01
6.66831434e-01 7.22437143e-01 -6.59942925e-01 1.15461424e-02
-1.21076846e+00 5.14440060e-01 5.84276080e-01 -6.52454272e-02
-9.65662301e-01 2.49188870e-01 -7.59608328e-01 2.20599473e-01
3.01343530e-01 -6.10581338e-01 3.12527865e-01 -8.46565187e-01
-9.31766629e-01 1.00818002e+00 7.55466521e-01 -8.27470720e-01
3.62950623e-01 3.81040484e-01 -4.72908229e-01 8.48607975e-04
-3.82205620e-02 6.05619311e-01 5.77171266e-01 -1.02088845e+00
-2.83519655e-01 -3.82861108e-01 7.14106187e-02 -8.30478072e-02
-2.26260290e-01 2.05871150e-01 -1.13080181e-01 -5.24701476e-01
2.04360709e-02 -1.12815559e+00 -3.92072588e-01 -5.16659081e-01
-7.13565707e-01 -6.20080054e-01 8.49669933e-01 -3.86897832e-01
1.11328816e+00 -2.01902795e+00 3.71913701e-01 7.32812405e-01
1.08184457e+00 4.81841067e-04 -1.32959098e-01 9.70963955e-01
-4.15394865e-02 3.38475466e-01 -2.88256168e-01 1.70308072e-02
-1.34307176e-01 2.49819741e-01 3.65364552e-01 8.02982569e-01
-8.23819265e-02 9.07827079e-01 -9.30522323e-01 -5.64826667e-01
3.39412391e-02 3.57186675e-01 -4.26451743e-01 -4.72370461e-02
8.27896371e-02 3.35417658e-01 -2.55730152e-01 6.11737609e-01
4.27835077e-01 -7.19934404e-01 8.22636425e-01 1.86617211e-01
3.11653614e-01 -4.00638789e-01 -1.01836812e+00 1.31670904e+00
2.62536556e-01 9.52616274e-01 3.26682925e-01 -1.49596512e+00
9.43162322e-01 6.47762790e-02 6.31576538e-01 -1.01406179e-01
5.60357451e-01 -3.36038381e-01 6.79615080e-01 -6.17682159e-01
1.23387665e-01 -1.58504527e-02 9.51104239e-02 4.63634789e-01
2.88425595e-01 2.97031671e-01 3.45225364e-01 6.87240481e-01
1.43226802e+00 -6.90988004e-01 2.60429382e-01 -4.26932484e-01
2.04143599e-01 -2.25711852e-01 1.27629220e-01 5.20903409e-01
-3.10751289e-01 2.30211109e-01 8.86932969e-01 -3.10029775e-01
-7.75245130e-01 -9.92897332e-01 2.84723580e-01 1.12655950e+00
1.18069448e-01 -3.35025907e-01 -5.96607387e-01 -5.74093461e-01
2.65808284e-01 -2.60606796e-01 -7.14295149e-01 -1.49594009e-01
-5.87255239e-01 -1.01914132e+00 6.94205582e-01 2.91908801e-01
3.50956649e-01 -1.30927181e+00 1.61862314e-01 1.11264519e-01
4.49815020e-02 -1.13221109e+00 -2.12157622e-01 1.89571396e-01
-1.00801325e+00 -1.47811627e+00 -3.07974100e-01 -1.29772329e+00
6.02942407e-01 1.52683035e-01 1.16910326e+00 5.36543250e-01
-2.05697387e-01 3.74196589e-01 -4.38628703e-01 -1.31346241e-01
-4.72548395e-01 9.99684632e-02 1.43518463e-01 1.15339030e-02
3.93038839e-01 -7.19970942e-01 -3.18513483e-01 7.01198876e-02
-6.36805058e-01 -3.22651953e-01 6.79123700e-01 7.74488568e-01
-1.71912804e-01 3.22291285e-01 7.99883902e-01 -1.22500253e+00
1.02614295e+00 -1.12845933e+00 -5.07514179e-01 -6.93699792e-02
-4.87200230e-01 -8.09767097e-02 3.27696145e-01 -5.05286455e-01
-2.34475911e-01 -7.27112889e-02 5.26595563e-02 -5.54765582e-01
-2.78794598e-02 1.09439743e+00 1.94205090e-01 -1.55129030e-01
5.98749578e-01 -2.60433614e-01 2.86993295e-01 -3.25547755e-01
3.91328335e-01 7.37589836e-01 3.20092618e-01 -5.48765901e-03
8.41830254e-01 5.18131137e-01 1.32095620e-01 -1.17566812e+00
4.15015370e-02 -7.81833112e-01 -7.40913570e-01 -5.86381316e-01
1.02364671e+00 -9.27279353e-01 -1.10268342e+00 3.83869737e-01
-1.04613483e+00 -3.39325279e-01 4.03200150e-01 5.75234711e-01
-3.14041004e-02 8.19787920e-01 -9.69647110e-01 -7.62347996e-01
-3.41817021e-01 -6.31822765e-01 5.99560201e-01 7.14005232e-02
-7.19114468e-02 -1.41861367e+00 2.24544346e-01 4.27763909e-01
1.68281794e-01 8.28167617e-01 8.72035086e-01 -9.55762386e-01
-5.64251363e-01 -1.13675594e-01 -3.03599179e-01 1.61473423e-01
-1.58725097e-03 2.54338503e-01 -4.89909023e-01 -5.49270451e-01
-6.47852957e-01 -2.38782745e-02 1.09147942e+00 4.70075518e-01
8.13474119e-01 -2.97581166e-01 -6.39800251e-01 5.19780934e-01
1.18043232e+00 2.92892247e-01 4.74275172e-01 -6.57293051e-02
1.26707327e+00 7.01238692e-01 -1.88364819e-01 2.29517147e-01
5.50020993e-01 4.71066646e-02 7.43255377e-01 -2.70763457e-01
-1.55893043e-01 1.83688682e-02 2.77679801e-01 1.32659626e+00
-3.81393254e-01 -5.67645311e-01 -1.17996800e+00 4.85495716e-01
-2.07537651e+00 -1.22555327e+00 -5.94431639e-01 1.69516778e+00
3.32482457e-01 8.41751844e-02 5.48610449e-01 -3.69333886e-02
1.26030135e+00 -5.97370304e-02 -2.26808891e-01 -1.96970433e-01
-1.41005665e-01 7.31268376e-02 5.20724475e-01 5.71024418e-01
-1.40615129e+00 1.07755840e+00 6.12856531e+00 5.45866072e-01
-8.86555672e-01 3.26442383e-02 5.31094670e-01 2.54734337e-01
4.36530672e-02 -1.86517611e-02 -2.22600341e-01 4.69616175e-01
7.97451377e-01 -4.15509418e-02 6.81495845e-01 7.32045174e-01
-1.59603253e-01 3.07302594e-01 -9.82341290e-01 9.62454259e-01
6.33513778e-02 -1.37185431e+00 -3.05191904e-01 4.24668968e-01
6.71213508e-01 7.60915697e-01 -2.69495249e-01 3.90232235e-01
9.94968474e-01 -1.42782819e+00 -1.24543300e-03 4.52249527e-01
4.83091861e-01 -8.33324373e-01 8.69220257e-01 4.03231472e-01
-1.11055267e+00 -2.84502894e-01 -4.93812561e-01 -2.73085564e-01
-1.86341286e-01 4.38137770e-01 -1.24826109e+00 4.53218549e-01
7.19775140e-01 9.82791185e-01 -6.50153756e-01 1.11986506e+00
8.24021697e-02 8.42090547e-01 -2.60995060e-01 -5.91511488e-01
3.89790028e-01 -5.33280849e-01 4.74221617e-01 1.23900330e+00
-9.00460109e-02 9.91415903e-02 3.18249822e-01 6.23307824e-01
-4.13211167e-01 7.80450925e-02 -9.50614214e-01 -4.23958838e-01
2.42743984e-01 1.50758183e+00 -1.40030229e+00 -2.17816800e-01
-2.05812991e-01 3.37978333e-01 5.40320694e-01 3.32146823e-01
-8.29020739e-01 -5.17667770e-01 2.09424958e-01 2.74299741e-01
1.18730880e-01 -4.52665687e-01 1.72094405e-01 -9.81962860e-01
-7.53077269e-01 -7.85673082e-01 6.80049777e-01 -3.96673530e-01
-1.89874864e+00 6.95440292e-01 1.35160670e-01 -1.00407934e+00
9.18628499e-02 -5.45286596e-01 -9.63278890e-01 1.58200011e-01
-8.79016101e-01 -1.12381709e+00 -3.73045892e-01 8.40833962e-01
-4.31016833e-03 -8.54270875e-01 2.54633278e-01 6.53323710e-01
-3.91163051e-01 1.92613065e-01 2.88098246e-01 8.34956884e-01
4.00683075e-01 -1.33553576e+00 6.50913894e-01 6.14258409e-01
4.10640299e-01 6.88123822e-01 3.07664394e-01 -9.95602787e-01
-1.24321866e+00 -1.10742009e+00 4.14243758e-01 -2.00582877e-01
1.07187772e+00 -6.33174241e-01 -7.03831732e-01 6.45692170e-01
5.50382197e-01 -2.03886136e-01 7.58966565e-01 3.41892451e-01
-1.37283608e-01 6.23402447e-02 -1.12020755e+00 4.77189332e-01
1.12169838e+00 -4.03559089e-01 -2.76026547e-01 8.00969899e-01
6.29810154e-01 1.91715971e-01 -9.50015187e-01 1.78904623e-01
4.39160436e-01 -8.34788322e-01 9.10579145e-01 -8.73948276e-01
2.91871488e-01 -3.15469414e-01 1.78667352e-01 -1.40872824e+00
-6.80196345e-01 -5.85053027e-01 -9.53629240e-02 9.52300131e-01
3.28633226e-02 -6.95599616e-01 1.00108194e+00 -2.77844191e-01
-6.02421388e-02 -2.93937653e-01 -9.27094638e-01 -4.92863327e-01
1.58202469e-01 8.99121538e-02 3.16548616e-01 1.62128842e+00
1.12086162e-01 9.98813331e-01 -3.84154975e-01 2.37514134e-02
9.05456126e-01 -1.51920736e-01 8.74359012e-01 -2.06846642e+00
-3.87220591e-01 -7.34448552e-01 -9.48417783e-01 -5.10180771e-01
2.82740295e-01 -1.30788243e+00 -3.59916419e-01 -1.92742741e+00
1.93437934e-01 -3.40273857e-01 -2.26180136e-01 3.23376805e-01
1.29949629e-01 2.14776099e-01 1.85890481e-01 8.91008750e-02
-8.51317883e-01 6.07306659e-02 1.05066526e+00 -5.70739985e-01
1.14017129e-01 5.30670993e-02 -6.00467503e-01 9.25946951e-01
8.33002269e-01 -6.34527564e-01 -2.32313424e-01 -2.08558753e-01
4.48578149e-01 6.05090149e-02 3.17986429e-01 -9.18322086e-01
6.03288233e-01 9.00687724e-02 1.74227133e-01 -5.14726579e-01
1.25624081e-02 -7.59772420e-01 3.50242287e-01 8.81894529e-01
-4.39984165e-02 8.17657728e-03 -2.65520722e-01 1.07276225e+00
-2.24376783e-01 5.21238372e-02 8.42710555e-01 -2.67182350e-01
-6.94428205e-01 6.30819678e-01 -7.27578759e-01 8.91415123e-03
1.00069225e+00 -9.01046023e-02 -4.25371081e-01 -6.04725182e-01
-1.05325687e+00 6.22039676e-01 6.09558448e-02 5.34442723e-01
5.89700520e-01 -1.10115087e+00 -9.98359025e-01 -4.89792228e-02
-2.89849155e-02 -2.97497511e-01 -2.09183451e-02 9.14000213e-01
-8.95636797e-01 -5.03944531e-02 -2.38303155e-01 -7.95854807e-01
-1.56754470e+00 7.49409080e-01 1.41557738e-01 -1.28958106e-01
-4.70853627e-01 9.28981960e-01 5.58885597e-02 -7.18474448e-01
4.18899983e-01 -3.21736932e-01 -8.79491031e-01 1.96508124e-01
-2.04978332e-01 4.60494012e-01 -3.05966198e-01 -7.71271646e-01
-5.35007298e-01 3.57997149e-01 1.49877921e-01 5.58537185e-01
1.43860650e+00 1.21778712e-01 -6.49996758e-01 3.43227744e-01
1.14483643e+00 -8.83051008e-02 -4.31220829e-01 -2.81643006e-03
3.31000149e-01 1.79774389e-01 -3.09011221e-01 -3.92931402e-01
-1.31406307e+00 7.58519053e-01 5.79939544e-01 1.03874362e+00
5.32195985e-01 4.00588632e-01 4.09878045e-01 6.77330077e-01
2.60296285e-01 -8.81245136e-01 2.35101834e-01 7.00238168e-01
6.55924678e-01 -1.44887972e+00 2.02840775e-01 -2.28476986e-01
-5.81155658e-01 1.05432749e+00 3.19140404e-01 -6.24762475e-01
1.16992009e+00 1.67674780e-01 -1.32064864e-01 -9.02400613e-01
-5.78242242e-01 -3.80455732e-01 1.45459622e-01 8.47913444e-01
5.03134549e-01 3.62250447e-01 -2.45619327e-01 2.50601798e-01
-2.19248980e-01 -5.17913938e-01 7.29628265e-01 3.41626704e-01
-6.51375234e-01 -7.88172066e-01 -2.08721489e-01 8.21466386e-01
-3.08872163e-01 -2.48831153e-01 -1.17492235e+00 1.02744734e+00
8.71092603e-02 1.22343814e+00 8.99268687e-02 -7.02544928e-01
-1.24433868e-01 -4.23743397e-01 1.64628372e-01 -5.83215356e-01
-7.73333311e-01 2.34416276e-02 2.48672515e-01 -5.82360197e-03
-6.83469951e-01 -2.42666051e-01 -1.24744129e+00 -7.37883925e-01
-6.21917248e-01 4.49043840e-01 4.14086282e-01 6.29718244e-01
9.67772380e-02 5.36713004e-01 5.67063034e-01 -6.62637949e-01
8.20867866e-02 -1.31075096e+00 -8.08816135e-01 5.02491832e-01
4.81958240e-02 -7.80734420e-01 -4.66206253e-01 -1.12796135e-01] | [7.075381755828857, 5.925144195556641] |
a9a1da19-6aa5-4a54-be8e-f2c6280e47f5 | mia-2022-shared-task-submission-leveraging | 2207.01940 | null | https://arxiv.org/abs/2207.01940v3 | https://arxiv.org/pdf/2207.01940v3.pdf | MIA 2022 Shared Task Submission: Leveraging Entity Representations, Dense-Sparse Hybrids, and Fusion-in-Decoder for Cross-Lingual Question Answering | We describe our two-stage system for the Multilingual Information Access (MIA) 2022 Shared Task on Cross-Lingual Open-Retrieval Question Answering. The first stage consists of multilingual passage retrieval with a hybrid dense and sparse retrieval strategy. The second stage consists of a reader which outputs the answer from the top passages returned by the first stage. We show the efficacy of using a multilingual language model with entity representations in pretraining, sparse retrieval signals to help dense retrieval, and Fusion-in-Decoder. On the development set, we obtain 43.46 F1 on XOR-TyDi QA and 21.99 F1 on MKQA, for an average F1 score of 32.73. On the test set, we obtain 40.93 F1 on XOR-TyDi QA and 22.29 F1 on MKQA, for an average F1 score of 31.61. We improve over the official baseline by over 4 F1 points on both the development and test sets. | ['Sarguna Janani Padmanabhan', 'Zhucheng Tu'] | 2022-07-05 | null | https://aclanthology.org/2022.mia-1.10 | https://aclanthology.org/2022.mia-1.10.pdf | naacl-mia-2022-7 | ['cross-lingual-question-answering', 'passage-retrieval'] | ['natural-language-processing', 'natural-language-processing'] | [-4.82946217e-01 -1.53909037e-02 9.87420157e-02 -1.69158444e-01
-2.37660766e+00 -7.67518759e-01 6.35297000e-01 2.89769232e-01
-9.21204150e-01 8.33236039e-01 4.62116688e-01 -3.08052450e-01
-1.18409604e-01 -4.70333934e-01 -1.01709640e+00 -2.38600671e-01
1.55126810e-01 8.76250625e-01 1.56034261e-01 -7.21709430e-01
-9.91700217e-02 -2.88065076e-01 -1.12474990e+00 6.95556581e-01
1.00816917e+00 8.66418660e-01 1.55527711e-01 9.27964389e-01
9.98260155e-02 9.68668461e-01 -7.68717587e-01 -7.48849750e-01
-1.57670215e-01 -3.34234506e-01 -1.17785943e+00 -7.92207897e-01
7.73538351e-01 -4.74036485e-01 -5.47070026e-01 6.33327007e-01
8.72866750e-01 2.49664173e-01 8.35315466e-01 -5.37017763e-01
-1.00181997e+00 8.04107487e-01 -1.57644406e-01 3.71925652e-01
9.84360635e-01 -1.02670677e-01 1.52354395e+00 -1.44094968e+00
8.09480190e-01 1.14383149e+00 3.29277337e-01 4.82822984e-01
-9.99554574e-01 -4.65437770e-01 -1.99028134e-01 2.79299408e-01
-1.51029634e+00 -7.66236126e-01 -7.19865412e-03 -6.40879720e-02
1.35092986e+00 1.72317073e-01 -4.12421823e-02 8.24486077e-01
-2.77406741e-02 1.35344589e+00 7.68103182e-01 -6.02524161e-01
-2.10437566e-01 1.31917879e-01 3.84295523e-01 5.84037244e-01
-8.65663588e-03 -2.00062886e-01 -5.43325603e-01 -1.90743003e-02
2.32357770e-01 -6.29424214e-01 -4.94695932e-01 3.96616668e-01
-1.09618855e+00 8.51614118e-01 5.11104882e-01 2.97093570e-01
-4.30353940e-01 -1.53918803e-01 3.76845479e-01 8.81715834e-01
3.62886578e-01 9.35283661e-01 -7.80288875e-01 -1.17257759e-02
-9.13077652e-01 6.59589291e-01 1.01885331e+00 9.18857813e-01
6.60000443e-01 -4.44265991e-01 -6.29434049e-01 1.35322094e+00
3.85318518e-01 1.24919569e+00 4.80522901e-01 -7.84685314e-01
7.84332752e-01 1.66909873e-01 1.24606974e-01 -6.85448170e-01
-1.43480435e-01 -4.90082085e-01 -3.37008744e-01 -7.01639652e-01
5.64069569e-01 -3.10843706e-01 -7.91447818e-01 1.82263136e+00
-1.27948433e-01 -6.30475283e-01 5.77607155e-01 8.51130843e-01
1.28116298e+00 1.24439061e+00 5.09907492e-02 9.03744623e-02
1.51282382e+00 -1.43891919e+00 -8.13704371e-01 -7.75289983e-02
1.17432928e+00 -1.26350617e+00 1.13875532e+00 1.64947882e-01
-1.47953069e+00 -5.98267972e-01 -8.25358689e-01 -7.16761768e-01
-4.21117902e-01 4.43258107e-01 -3.80128399e-02 1.95818022e-01
-1.38807225e+00 -2.64731180e-02 -2.43555680e-01 -3.34612578e-01
-2.32380420e-01 6.08888231e-02 -3.48727912e-01 -8.58286679e-01
-1.82596374e+00 1.14871323e+00 5.09698689e-02 -2.83186585e-02
-1.23686683e+00 -8.14523637e-01 -1.01401865e+00 1.84677765e-01
1.07897125e-01 -6.93979502e-01 1.56825662e+00 -3.76498729e-01
-1.27295744e+00 8.62732589e-01 -1.38940319e-01 -4.72454101e-01
1.34520173e-01 -7.70695031e-01 -7.15357482e-01 4.06988472e-01
4.76847082e-01 8.41402173e-01 2.47426257e-01 -8.86086464e-01
-5.78596234e-01 -5.39233536e-02 1.60337836e-01 6.87882721e-01
7.73368403e-02 2.17827454e-01 -8.57007325e-01 -3.13512206e-01
-3.13641131e-01 -8.36195409e-01 1.71490610e-01 -6.55295193e-01
-3.77003551e-01 -5.55423260e-01 2.01000556e-01 -1.43533897e+00
1.42740679e+00 -1.92695582e+00 1.77861169e-01 1.34601712e-01
-1.14207849e-01 2.79769421e-01 -7.55279720e-01 6.66102409e-01
1.72577292e-01 5.10824956e-02 8.45886841e-02 -3.89816374e-01
1.89067110e-01 -2.45423049e-01 -2.96823949e-01 -1.72873575e-03
3.76655132e-01 1.10110438e+00 -9.38482165e-01 -3.82953972e-01
-3.87111098e-01 4.38196152e-01 -5.33193111e-01 4.29127932e-01
-2.34956324e-01 2.63850033e-01 -5.14516890e-01 5.63954115e-01
1.46715879e-01 -2.54849017e-01 -2.14568511e-01 4.61921692e-02
1.16872787e-01 9.39305782e-01 -6.95336461e-01 2.27662492e+00
-7.80846417e-01 6.20748699e-01 1.76226854e-01 -3.58719617e-01
6.56877697e-01 7.07316279e-01 8.55739936e-02 -1.48823071e+00
-1.26188278e-01 7.73352504e-01 -9.35255662e-02 -4.69084948e-01
8.99916112e-01 1.25746861e-01 -5.23286283e-01 6.60716474e-01
4.70197111e-01 -7.45703131e-02 4.01145786e-01 8.29982698e-01
1.18744159e+00 -8.02734569e-02 -3.00129652e-01 -3.45667452e-01
8.74516428e-01 8.58712643e-02 -2.48711050e-01 8.84126246e-01
1.88714862e-01 7.12606609e-01 1.09903932e-01 6.00395687e-02
-7.32959986e-01 -1.00718725e+00 -1.27254397e-01 1.33235264e+00
-2.02807069e-01 -4.63304490e-01 -6.87681258e-01 -6.10064507e-01
-8.06507245e-02 7.96676576e-01 -2.29289055e-01 -2.64640629e-01
-6.40713573e-01 -1.46855220e-01 8.83917332e-01 1.05295293e-01
3.09082568e-01 -1.09779084e+00 2.26481602e-01 1.40126169e-01
-8.95058751e-01 -1.06628525e+00 -7.37470269e-01 -1.95828527e-01
-3.55076224e-01 -7.39571691e-01 -1.37827957e+00 -1.06016374e+00
1.66325361e-01 3.40412967e-02 1.68697298e+00 1.11147359e-01
1.67007938e-01 5.78646183e-01 -4.60975647e-01 7.01029897e-02
-3.09858203e-01 7.66998649e-01 -1.32880256e-01 -3.54522169e-01
2.92901814e-01 3.65228534e-01 -6.39007390e-01 1.47597551e-01
-6.56301439e-01 -2.45278820e-01 7.50557184e-01 8.33433330e-01
7.46767282e-01 -5.95516920e-01 9.54023600e-01 -5.95806599e-01
8.67456615e-01 -7.38652229e-01 -3.95486712e-01 7.62199998e-01
-3.57916504e-01 2.30915502e-01 4.47692573e-01 -5.59417307e-02
-1.01985586e+00 -6.16855681e-01 -5.06749451e-01 -1.14933312e-01
2.16241464e-01 1.06108248e+00 -6.77370802e-02 3.43983829e-01
7.47371316e-01 1.78144276e-02 -2.40322560e-01 -8.46059620e-01
6.63680077e-01 7.76076019e-01 5.13685644e-01 -7.44499981e-01
3.75424474e-01 -5.15549541e-01 -9.06735718e-01 -6.51602626e-01
-1.12886333e+00 -7.64940143e-01 -1.20028384e-01 -2.05974020e-02
8.99712801e-01 -1.47395802e+00 -3.61782908e-01 2.64196515e-01
-1.02809048e+00 -4.12241399e-01 -2.23937899e-01 5.39685726e-01
-2.71292865e-01 -1.18495494e-01 -8.82541358e-01 -3.33784103e-01
-7.79897511e-01 -1.17432356e+00 1.28525054e+00 1.16502315e-01
-2.88444132e-01 -9.42701459e-01 4.05185401e-01 8.18864405e-01
5.41611969e-01 -5.22913039e-01 8.88593853e-01 -1.04502535e+00
-5.12418807e-01 -4.53871936e-01 -1.66075617e-01 3.79330575e-01
-2.52022624e-01 -5.43232918e-01 -8.61994028e-01 -6.48012817e-01
-3.21147025e-01 -1.10532999e+00 7.72099674e-01 1.13567896e-01
5.14145374e-01 -5.82815148e-02 -4.10996052e-03 -4.03472707e-02
1.34667051e+00 -1.35058850e-01 4.98769134e-01 1.00328743e-01
5.42767465e-01 6.24319017e-01 6.13812268e-01 -2.45264247e-01
8.37709844e-01 7.64686406e-01 -3.40763688e-01 5.93795404e-02
-5.27368188e-01 -6.20278299e-01 6.60113633e-01 1.55956614e+00
5.32492280e-01 -3.71800601e-01 -1.08155167e+00 1.00759172e+00
-1.49060392e+00 -5.31992435e-01 -5.28978370e-02 2.23728871e+00
1.20458412e+00 -3.30012232e-01 -1.20234132e-01 -3.88413906e-01
2.57332921e-01 -8.84872600e-02 -2.47950628e-01 -4.09366995e-01
-1.96547508e-01 6.91756308e-01 1.43211007e-01 9.95219529e-01
-8.01782727e-01 1.16436720e+00 5.92617941e+00 1.05690372e+00
-6.16729438e-01 1.73616618e-01 5.29408395e-01 -8.02370310e-02
-5.91126382e-01 -1.34873450e-01 -1.11608696e+00 1.95156008e-01
1.70832551e+00 -4.12034065e-01 3.48893702e-01 3.40922266e-01
-3.82891864e-01 -2.61574864e-01 -1.09114242e+00 7.22756267e-01
4.01207656e-01 -1.03107071e+00 3.37136239e-01 -3.07390779e-01
9.63175595e-01 4.53458965e-01 6.62058368e-02 1.03630674e+00
3.26559544e-01 -1.16127145e+00 4.95628655e-01 5.94523907e-01
1.05272150e+00 -9.18069422e-01 9.45645034e-01 3.44665557e-01
-9.28801894e-01 4.70425338e-01 -1.43858343e-01 2.77565300e-01
2.19999537e-01 1.32373422e-01 -6.60830021e-01 8.72362375e-01
6.32080555e-01 5.48878253e-01 -6.20431960e-01 8.85525644e-01
-5.46003401e-01 5.85425138e-01 -4.08995390e-01 -1.10653959e-01
4.53774333e-01 2.28763163e-01 4.04921442e-01 1.22475517e+00
8.62639844e-02 -3.42916250e-02 1.42888278e-02 3.56637836e-01
-6.62168086e-01 5.77907562e-01 -5.41936696e-01 -1.02654524e-01
4.38295245e-01 9.62413311e-01 2.77051598e-01 -4.90219504e-01
-3.99470657e-01 9.75911081e-01 5.44338048e-01 6.44518614e-01
-5.11501491e-01 -7.84834921e-01 1.73622146e-01 -3.12709302e-01
2.73387581e-01 -1.63447000e-02 4.47443932e-01 -1.46602201e+00
1.49380062e-02 -1.52782738e+00 7.47605324e-01 -7.96351254e-01
-1.35006237e+00 7.91236699e-01 -2.25460365e-01 -8.26035380e-01
-6.43776774e-01 -2.86504179e-01 6.82046711e-02 1.46515071e+00
-1.79753625e+00 -1.02159429e+00 3.67146283e-01 7.88547754e-01
7.30400085e-01 -2.47795478e-01 1.20239305e+00 9.73713875e-01
-2.93543488e-01 1.01859415e+00 3.56014699e-01 3.60655218e-01
1.19843745e+00 -1.07357216e+00 1.27850190e-01 7.37256944e-01
3.46633911e-01 9.00285006e-01 1.93098515e-01 -3.88387054e-01
-1.58906078e+00 -8.67331028e-01 1.68389225e+00 -7.40851581e-01
7.49521375e-01 -1.22751981e-01 -1.06148136e+00 6.58943176e-01
7.35422373e-01 -4.74108607e-01 6.86775506e-01 4.82243627e-01
-4.18817967e-01 -6.30371878e-03 -9.51430976e-01 3.89839649e-01
2.94764936e-01 -1.20959353e+00 -9.06511366e-01 5.20308077e-01
9.44139063e-01 -5.61403155e-01 -1.31267202e+00 2.33135521e-01
4.66666579e-01 -6.08737059e-02 1.06678689e+00 -6.13116562e-01
5.23914635e-01 -1.38703063e-01 -4.60631132e-01 -1.34139562e+00
-6.81814598e-03 -3.37775350e-01 8.01605508e-02 1.00165009e+00
1.25857735e+00 -3.12831253e-01 -1.14325233e-01 3.49649847e-01
-2.02503711e-01 -6.37487054e-01 -8.68137300e-01 -4.37962562e-01
6.73316658e-01 -1.82459772e-01 1.66572422e-01 7.99579322e-01
-1.64542109e-01 1.03616810e+00 -6.59162775e-02 1.83296397e-01
2.08532065e-01 -7.54720494e-02 6.31280661e-01 -6.38906062e-01
-2.52875239e-01 -2.18930900e-01 4.16242838e-01 -1.41397965e+00
1.63014978e-01 -1.19855571e+00 2.42403373e-01 -1.74999070e+00
1.62984625e-01 -3.39494556e-01 -4.46149141e-01 2.48699114e-01
-3.83697242e-01 1.70499951e-01 2.55778357e-02 3.66026998e-01
-1.14840281e+00 6.00493371e-01 1.26968539e+00 -1.01485774e-01
7.87377134e-02 -3.75571609e-01 -8.69507015e-01 4.45013046e-02
4.84793335e-01 -4.38214093e-01 -4.26063240e-01 -1.15451849e+00
3.91494989e-01 4.46132421e-01 -1.08472668e-01 -7.42624700e-01
3.03045720e-01 7.40494072e-01 1.98041618e-01 -6.92175448e-01
3.78316849e-01 -4.14506167e-01 -4.24173057e-01 3.08770418e-01
-7.58288860e-01 3.69206458e-01 4.86904681e-01 1.89847633e-01
-7.17868805e-01 -9.52555686e-02 3.06816876e-01 -4.69670594e-02
-5.03710747e-01 1.33523628e-01 -3.24343890e-01 7.21042275e-01
2.39379197e-01 6.28343642e-01 -5.59051156e-01 -8.97944212e-01
-6.99183047e-01 9.20562923e-01 8.87494441e-03 7.63725638e-01
5.79625010e-01 -1.38243294e+00 -1.23425996e+00 1.60828084e-01
2.59212077e-01 -1.36306241e-01 2.44457230e-01 8.37719142e-01
-5.38582921e-01 1.14660192e+00 2.23940164e-01 -3.42769027e-01
-8.84285033e-01 -4.26439941e-02 3.05531263e-01 -9.04246509e-01
-8.52680877e-02 1.12055171e+00 -4.53769602e-02 -7.97040164e-01
1.55274376e-01 2.45198291e-02 -4.06228304e-01 2.04341382e-01
8.02162826e-01 3.13773274e-01 4.76169169e-01 -7.82079339e-01
-3.05871218e-01 5.32167196e-01 -6.46707714e-01 -8.01778138e-01
9.52562094e-01 -2.14870572e-01 -1.02964208e-01 4.37062234e-01
1.73761451e+00 2.19685212e-01 -2.80685872e-01 -5.47751307e-01
9.37602222e-02 2.64083315e-02 1.50299489e-01 -1.54934847e+00
-7.39255130e-01 9.07496154e-01 4.14686948e-01 -9.27268863e-02
8.52935493e-01 2.77517706e-01 1.11127865e+00 8.43441844e-01
3.09134394e-01 -9.68654096e-01 -7.85199404e-02 1.14631808e+00
1.22727776e+00 -1.34291303e+00 -2.84681499e-01 2.54624903e-01
-6.74548745e-01 6.08804762e-01 4.03973043e-01 -5.41307144e-02
3.89214545e-01 -3.51326972e-01 4.21743959e-01 -4.92943108e-01
-8.76169443e-01 -1.36747226e-01 8.05715144e-01 -1.38568670e-01
7.96700716e-01 5.85738160e-02 -2.87871242e-01 5.24598598e-01
-2.84045547e-01 -3.14153641e-01 -7.21094757e-02 7.70584106e-01
-2.38589808e-01 -9.67130184e-01 -5.39713204e-02 2.25519612e-01
-7.38969684e-01 -6.00840449e-01 -2.02118307e-01 7.17702508e-01
-6.00071847e-01 1.21056271e+00 -2.61020008e-02 -1.07866168e-01
6.30888999e-01 4.63553101e-01 2.86316037e-01 -6.91350281e-01
-8.68883610e-01 5.67729808e-02 6.87699437e-01 -5.18462181e-01
-8.46732706e-02 -6.48979664e-01 -9.95499194e-01 -7.71036744e-02
-4.77164835e-01 7.08165765e-01 6.57311857e-01 9.32624638e-01
5.30712903e-01 3.33607942e-01 4.39122438e-01 1.37546346e-01
-5.13048470e-01 -1.31219542e+00 -1.25928059e-01 2.09022716e-01
5.43818474e-01 -9.91294309e-02 -3.50428998e-01 -2.42927283e-01] | [11.451630592346191, 8.320489883422852] |
a469f09b-dd08-4397-959a-aed5da1bca46 | gaining-and-losing-influence-in-online | null | null | https://aclanthology.org/L18-1110 | https://aclanthology.org/L18-1110.pdf | Gaining and Losing Influence in Online Conversation | null | ['Arun Sharma', 'Tomek Strzalkowski'] | 2018-05-01 | gaining-and-losing-influence-in-online-1 | https://aclanthology.org/L18-1110 | https://aclanthology.org/L18-1110.pdf | lrec-2018-5 | ['dialogue-understanding'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.370425224304199, 3.736316680908203] |
6673d2ae-adba-42f0-8701-ac822d49d7fd | improving-few-shot-user-specific-gaze | 1904.10638 | null | http://arxiv.org/abs/1904.10638v1 | http://arxiv.org/pdf/1904.10638v1.pdf | Improving Few-Shot User-Specific Gaze Adaptation via Gaze Redirection Synthesis | As an indicator of human attention gaze is a subtle behavioral cue which can
be exploited in many applications. However, inferring 3D gaze direction is
challenging even for deep neural networks given the lack of large amount of
data (groundtruthing gaze is expensive and existing datasets use different
setups) and the inherent presence of gaze biases due to person-specific
difference. In this work, we address the problem of person-specific gaze model
adaptation from only a few reference training samples. The main and novel idea
is to improve gaze adaptation by generating additional training samples through
the synthesis of gaze-redirected eye images from existing reference samples. In
doing so, our contributions are threefold: (i) we design our gaze redirection
framework from synthetic data, allowing us to benefit from aligned training
sample pairs to predict accurate inverse mapping fields; (ii) we proposed a
self-supervised approach for domain adaptation; (iii) we exploit the gaze
redirection to improve the performance of person-specific gaze estimation.
Extensive experiments on two public datasets demonstrate the validity of our
gaze retargeting and gaze estimation framework. | ['Jean-Marc Odobez', 'Yu Yu', 'Gang Liu'] | 2019-04-24 | improving-few-shot-user-specific-gaze-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Yu_Improving_Few-Shot_User-Specific_Gaze_Adaptation_via_Gaze_Redirection_Synthesis_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Yu_Improving_Few-Shot_User-Specific_Gaze_Adaptation_via_Gaze_Redirection_Synthesis_CVPR_2019_paper.pdf | cvpr-2019-6 | ['gaze-redirection'] | ['computer-vision'] | [ 3.15726191e-01 6.97242990e-02 -7.30171055e-02 -6.67567432e-01
-2.00601950e-01 -4.21626747e-01 5.01728058e-01 -4.60056603e-01
-3.09288681e-01 8.30092192e-01 2.10626647e-01 7.64558837e-02
5.32864965e-03 -2.61120796e-01 -8.18953574e-01 -4.82271105e-01
3.66990179e-01 -4.59770346e-03 7.31591955e-02 -2.31769472e-01
5.56147277e-01 1.87436417e-01 -2.09896564e+00 -2.22708508e-01
1.19437242e+00 1.02063465e+00 1.70456842e-01 4.92731661e-01
1.40330195e-01 6.52124286e-01 -5.68186224e-01 -1.18819878e-01
2.67555445e-01 -7.84836948e-01 -7.62282252e-01 1.12905614e-01
8.26303601e-01 -3.28595728e-01 2.47961774e-01 9.59879994e-01
5.24877787e-01 3.59321445e-01 6.20597124e-01 -1.62196493e+00
-8.94307315e-01 -5.66509226e-03 -9.12909687e-01 2.78500676e-01
4.95635629e-01 5.30109584e-01 7.45777607e-01 -6.93265319e-01
5.33555806e-01 1.15852034e+00 4.21580762e-01 1.08166111e+00
-1.23778725e+00 -1.06927228e+00 2.32631981e-01 4.25605297e-01
-1.25629210e+00 -7.79506683e-01 9.36264694e-01 -4.89489198e-01
4.03494120e-01 2.18894213e-01 5.53285182e-01 1.52194023e+00
-3.76073629e-01 6.91014409e-01 1.57332051e+00 -7.76895106e-01
-1.61522347e-02 4.31668997e-01 2.55423546e-01 4.34466422e-01
1.40666470e-01 1.91832736e-01 -8.39493632e-01 2.28748739e-01
4.81775284e-01 -6.36097118e-02 -6.44917905e-01 -4.46283758e-01
-1.32302225e+00 5.37998557e-01 9.01327312e-01 -2.37473287e-03
-3.19223344e-01 -2.90630817e-01 -1.41887993e-01 1.12385601e-01
5.81903040e-01 4.42321211e-01 -2.12684914e-01 -1.48775086e-01
-8.57301474e-01 3.76241803e-01 5.06331921e-01 9.88226771e-01
1.02994132e+00 -3.44531000e-01 -3.56549829e-01 8.11175644e-01
3.95946771e-01 8.40496480e-01 5.01290798e-01 -7.85578787e-01
4.22130853e-01 5.82540810e-01 4.11568135e-01 -8.98437560e-01
-3.98729682e-01 -1.44272998e-01 -4.44156736e-01 3.90980154e-01
9.04950500e-01 -1.33910373e-01 -7.21844077e-01 2.25153613e+00
4.62419093e-01 -9.05073900e-03 -3.58682364e-01 1.22497118e+00
6.53111279e-01 -1.83006972e-02 6.76142648e-02 -9.41808000e-02
1.36526024e+00 -9.50037837e-01 -6.88086689e-01 -3.93357784e-01
4.28726226e-01 -4.88565743e-01 1.57622898e+00 8.05467740e-02
-1.04076326e+00 -6.93533063e-01 -8.75131965e-01 -2.28651538e-01
-2.03963637e-01 1.04872733e-01 4.35420245e-01 7.63619781e-01
-1.01527786e+00 1.28265545e-01 -5.66543460e-01 -6.81058288e-01
6.56526148e-01 4.58120048e-01 -3.56567234e-01 1.14251576e-01
-8.80508900e-01 8.23841393e-01 1.04442844e-02 1.45820960e-01
-3.86441082e-01 -8.47986221e-01 -8.79572749e-01 5.54866670e-03
5.02123296e-01 -8.19368482e-01 1.23578465e+00 -1.59388351e+00
-1.63034654e+00 1.03094602e+00 -8.40981662e-01 -6.66816859e-03
5.10590971e-01 -2.04305574e-01 -1.99402869e-01 -1.60765663e-01
1.74176097e-01 9.12353396e-01 1.01733875e+00 -1.17715275e+00
-7.58031964e-01 -7.60571778e-01 9.48000401e-02 2.40482524e-01
-4.40936655e-01 -5.68647981e-02 -1.54675782e-01 -2.33551160e-01
-1.91121176e-01 -1.14244461e+00 3.61317575e-01 4.70137298e-02
-5.90261638e-01 -2.89106905e-01 5.92043817e-01 -4.47688520e-01
1.12027729e+00 -2.12510490e+00 1.83122188e-01 -1.29314540e-02
6.03601456e-01 1.73782140e-01 5.55356964e-02 -3.38379033e-02
-2.33538523e-01 -2.37071179e-02 -2.53664088e-02 -5.97905993e-01
1.23261670e-02 -2.83273876e-01 -2.41880655e-01 3.32225502e-01
3.34783614e-01 1.10861540e+00 -9.05907154e-01 -4.33066756e-01
4.85635772e-02 3.91089618e-01 -4.27846998e-01 4.16603267e-01
-2.08512358e-02 9.52195048e-01 -2.24203452e-01 4.58219618e-01
7.43094444e-01 -4.20097172e-01 -3.18409681e-01 -2.13414788e-01
-9.88103822e-02 3.10406834e-01 -8.12390864e-01 1.53893483e+00
-2.97798634e-01 9.76271629e-01 -3.10784280e-01 -3.53963286e-01
7.90840209e-01 -1.76851869e-01 9.85931084e-02 -9.35365558e-01
4.84732151e-01 -1.06074274e-01 9.94453058e-02 -7.29451895e-01
5.26015460e-01 3.89488861e-02 3.21361095e-01 9.01753724e-01
-3.55446860e-02 4.12239939e-01 7.10802004e-02 -1.77818581e-01
5.47173083e-01 4.26365495e-01 3.29060793e-01 -2.28514090e-01
7.15268791e-01 -2.16947377e-01 3.84592772e-01 3.40938866e-01
-6.05317831e-01 7.52313733e-01 4.28513050e-01 -3.48915040e-01
-8.73496652e-01 -6.56779230e-01 -1.56981349e-02 1.43056941e+00
2.78789312e-01 2.72395462e-02 -1.16393900e+00 -8.47702801e-01
-2.05651060e-01 8.02094758e-01 -1.19085884e+00 -1.38370439e-01
-4.58153099e-01 -5.52136481e-01 3.00024897e-01 4.79758263e-01
5.06896138e-01 -9.82068896e-01 -6.87542975e-01 -5.92394054e-01
-3.18625867e-01 -8.65073979e-01 -7.44798541e-01 -4.82457310e-01
-5.12505829e-01 -1.33094668e+00 -1.07974589e+00 -3.45247000e-01
9.64673281e-01 7.74119198e-01 1.02773833e+00 -4.91464511e-03
1.77808329e-01 3.02050412e-01 -1.00041062e-01 -7.87184060e-01
-1.18616641e-01 4.04052973e-01 1.56343102e-01 3.31470877e-01
1.02537346e+00 -4.03789699e-01 -8.21926892e-01 5.19343078e-01
-4.60299283e-01 1.96270362e-01 3.10391009e-01 7.20018089e-01
1.57649025e-01 -8.25915515e-01 5.22617817e-01 -9.99622643e-01
8.03143919e-01 -3.43057811e-01 -6.22990310e-01 1.99562460e-01
-8.06092441e-01 1.07934475e-01 1.15028575e-01 -5.13761401e-01
-1.20701063e+00 -3.02493453e-01 2.92240739e-01 -4.49342459e-01
-5.88033378e-01 -1.50816515e-01 -8.71214345e-02 -1.13899842e-01
1.29152894e+00 -9.20551568e-02 3.99115533e-01 -3.36985946e-01
2.30588868e-01 8.78671646e-01 3.30751359e-01 -3.33162844e-01
6.56717658e-01 5.68172812e-01 7.81838223e-02 -6.92290604e-01
-1.09499371e+00 -2.70538241e-01 -9.51699436e-01 -4.28717136e-01
6.93079829e-01 -8.42827499e-01 -1.24270844e+00 5.16382456e-01
-8.32197785e-01 -4.07311350e-01 -8.81919339e-02 3.66885424e-01
-3.75903785e-01 1.62735716e-01 2.04632849e-01 -7.53579080e-01
-3.46985757e-01 -1.13862205e+00 1.19907546e+00 7.06828654e-01
-3.95055890e-01 -8.95451069e-01 2.43726239e-01 3.48328412e-01
3.63905668e-01 2.63503104e-01 3.44667226e-01 -3.58037382e-01
-5.30533075e-01 3.69675718e-02 -5.17627895e-01 1.67886510e-01
4.63974148e-01 -1.90984365e-02 -1.43964362e+00 -2.61570215e-01
-2.11493209e-01 -4.33923334e-01 5.40704906e-01 3.60927284e-01
9.84891474e-01 1.15683479e-02 -5.09446800e-01 8.26182306e-01
8.63123477e-01 -4.12141204e-01 4.76546645e-01 4.65912610e-01
9.88282144e-01 1.08884251e+00 6.20972455e-01 3.09922528e-02
8.65072131e-01 8.23993087e-01 2.94111341e-01 1.24745443e-02
-3.39805633e-01 -1.35334671e-01 3.21737044e-02 1.31751392e-02
-3.99957865e-01 -1.45243779e-01 -9.36689079e-01 5.44826567e-01
-1.73223245e+00 -1.00817287e+00 -3.02812696e-01 2.41971684e+00
8.18475425e-01 -1.66818902e-01 6.33475363e-01 -2.49924771e-02
8.15190077e-01 -1.65208220e-01 -7.15103209e-01 1.00434378e-01
1.53137803e-01 -2.70118285e-02 3.73060197e-01 3.10229689e-01
-7.77099192e-01 7.03397751e-01 5.53002834e+00 1.05360404e-01
-1.56450820e+00 1.11598827e-01 3.03907275e-01 -5.26823699e-01
-9.97508392e-02 -3.34833771e-01 -9.55008268e-01 8.02704632e-01
7.81207621e-01 -5.25869690e-02 6.03769541e-01 5.89282632e-01
1.53669447e-01 -1.36134624e-01 -1.24838305e+00 1.22897351e+00
4.77619737e-01 -6.81596577e-01 -4.88782138e-01 3.94493967e-01
5.61873674e-01 3.72445062e-02 2.64071554e-01 4.16297317e-02
-1.69836298e-01 -9.37454462e-01 7.96681464e-01 8.09166074e-01
1.02109909e+00 -5.06980956e-01 2.83687204e-01 3.58489782e-01
-5.08791447e-01 -1.80077747e-01 -1.05535723e-01 -2.41921708e-01
7.79010877e-02 4.59869951e-02 -1.08113194e+00 2.35456139e-01
9.40531015e-01 8.65482211e-01 -1.14796114e+00 9.62034225e-01
-4.25785363e-01 4.65431571e-01 -2.62149632e-01 -2.63719913e-02
-2.86719203e-01 1.39116831e-02 5.49413383e-01 6.16589844e-01
8.24984536e-02 -3.17064881e-01 -6.68367267e-01 1.29646671e+00
-6.03510775e-02 -2.11314097e-01 -5.49776375e-01 3.50245923e-01
5.49978614e-01 1.17268276e+00 -2.70231098e-01 1.38068005e-01
-3.81426871e-01 7.97025979e-01 5.97234070e-01 6.29456818e-01
-7.36929119e-01 -4.04936463e-01 8.69752228e-01 3.53358448e-01
2.26741880e-01 1.61127582e-01 -4.67923880e-01 -1.26063931e+00
1.82913482e-01 -7.74078786e-01 4.55583185e-02 -1.05507648e+00
-1.21019053e+00 6.49503112e-01 2.55356841e-02 -1.31728983e+00
-6.05556428e-01 -5.16556978e-01 -4.18168724e-01 1.45351458e+00
-1.81191468e+00 -1.46643591e+00 -1.20829594e+00 7.69795060e-01
2.03512341e-01 -1.69109881e-01 6.21708155e-01 9.51056555e-02
-6.88997865e-01 1.05903792e+00 -2.29166806e-01 -1.17348082e-01
1.27556896e+00 -1.23396897e+00 4.54691231e-01 8.59313488e-01
-1.84442356e-01 9.32535946e-01 6.92015529e-01 -1.83326751e-01
-9.08077478e-01 -7.17626989e-01 7.44214892e-01 -1.10476565e+00
2.26260036e-01 -4.96144205e-01 -1.08802330e+00 8.14131796e-01
2.82848805e-01 -2.37775985e-02 8.11127126e-01 4.44812119e-01
-2.71673054e-01 -2.19234660e-01 -1.08709419e+00 8.06119025e-01
1.20192671e+00 -5.23376524e-01 -5.85216105e-01 -9.54510421e-02
3.46932828e-01 -5.18437028e-01 -3.53811949e-01 1.06584072e-01
7.95969903e-01 -1.20600533e+00 7.67714977e-01 -4.74019498e-01
2.23814309e-01 -3.42923880e-01 3.75226349e-01 -1.20336223e+00
-2.16817468e-01 -5.46077251e-01 -2.09797069e-01 1.38366616e+00
1.59269515e-02 -8.90087962e-01 7.08241999e-01 9.78734732e-01
3.52911204e-01 -4.00281578e-01 -4.03891057e-01 -3.11772466e-01
-1.82071134e-01 -1.33348659e-01 1.08796811e+00 8.15578580e-01
-1.75042450e-01 6.09080970e-01 -4.13435727e-01 1.67384557e-02
7.15031922e-01 2.50909049e-02 1.49952674e+00 -1.52328241e+00
6.19232841e-02 -3.51638377e-01 -2.29265809e-01 -1.20780051e+00
3.32698315e-01 -3.63559335e-01 9.17449147e-02 -7.98294604e-01
1.03218034e-01 -3.94743413e-01 -2.06280768e-01 3.53287309e-01
-6.06100798e-01 5.21896660e-01 2.20428020e-01 5.87130845e-01
-4.53756183e-01 4.32869852e-01 1.39321208e+00 2.07928866e-01
-3.90999258e-01 2.68938690e-01 -1.21654117e+00 4.88425672e-01
4.62518841e-01 -2.78271884e-01 -5.73088288e-01 -5.46375513e-01
3.36835474e-01 -4.93475169e-01 7.73208320e-01 -8.68039906e-01
2.40845159e-01 -1.94836214e-01 4.93897706e-01 -3.95365179e-01
2.41147146e-01 -6.56662345e-01 -3.64566505e-01 -1.79161146e-01
-3.44662607e-01 -2.80467659e-01 1.28797308e-01 6.25815153e-01
1.04518808e-01 -1.98523298e-01 7.18341172e-01 1.50714889e-01
-6.51272237e-01 9.36604664e-02 3.70047212e-01 1.90683439e-01
1.04439080e+00 -6.61737859e-01 -4.76114482e-01 -3.54713261e-01
-2.32505932e-01 1.56875119e-01 1.13944328e+00 7.35908568e-01
4.77620922e-02 -1.09827185e+00 -3.37724626e-01 5.93842685e-01
5.88550329e-01 1.13378517e-01 8.95649791e-02 1.06761491e+00
-1.00458659e-01 4.89193141e-01 -5.59075952e-01 -9.44235206e-01
-1.25852537e+00 6.47983909e-01 4.14505005e-01 5.65121830e-01
-2.09485620e-01 8.45190883e-01 6.28380477e-01 -2.40815878e-01
2.47165300e-02 -1.11083530e-01 -5.55568933e-01 -4.51929718e-02
9.23512936e-01 3.98831815e-01 -6.71190992e-02 -1.06341779e+00
-3.16505283e-01 7.13517487e-01 -1.79882273e-01 -3.77091691e-02
1.15714622e+00 -7.71325767e-01 9.28537697e-02 5.17998159e-01
9.51721489e-01 5.67573383e-02 -1.52367127e+00 -3.41913044e-01
-2.67036378e-01 -9.07005966e-01 -1.63123801e-01 -8.10455561e-01
-7.45075822e-01 1.06920779e+00 8.16431105e-01 1.84047043e-01
1.31993318e+00 -8.52525234e-02 5.67495108e-01 2.59725660e-01
1.97886914e-01 -8.53863776e-01 -1.24620222e-01 1.68704405e-01
6.98301196e-01 -1.79585385e+00 -1.83200344e-01 -1.55450791e-01
-8.07323277e-01 8.54118884e-01 8.41191709e-01 1.88947558e-01
4.86893862e-01 -5.67632020e-01 2.53397703e-01 -2.48184830e-01
-4.75042671e-01 -4.68960494e-01 6.77695155e-01 9.87081110e-01
4.80110854e-01 -3.63909960e-01 5.60947545e-02 4.82420474e-01
-4.54897463e-01 3.91202956e-01 3.07161152e-01 6.08928800e-01
-3.63023840e-02 -9.34621155e-01 -4.29886132e-01 2.59541243e-01
-3.42425346e-01 1.85958028e-01 -3.68715495e-01 7.84759402e-01
2.22130511e-02 8.34710360e-01 1.36752948e-01 -2.11954385e-01
4.23436582e-01 1.08980224e-01 6.65127397e-01 -5.98721981e-01
-2.89482653e-01 -3.90791804e-01 -4.23936665e-01 -6.17182076e-01
-7.21497118e-01 -8.97869527e-01 -5.50234914e-01 -3.49725485e-01
-3.88908565e-01 -3.12306762e-01 6.10727668e-01 1.03535438e+00
7.89914608e-01 5.10361671e-01 6.01484895e-01 -1.19005346e+00
-3.02532643e-01 -1.33293140e+00 -3.60329300e-01 8.92541349e-01
8.05999756e-01 -1.23636961e+00 -3.77977967e-01 4.61949527e-01] | [14.12009048461914, 0.0506814680993557] |
03e96ef4-913b-433b-a664-26bb0ad18a3b | indoor-positioning-via-gradient-boosting | 2211.08752 | null | https://arxiv.org/abs/2211.08752v1 | https://arxiv.org/pdf/2211.08752v1.pdf | Indoor Positioning via Gradient Boosting Enhanced with Feature Augmentation using Deep Learning | With the emerge of the Internet of Things (IoT), localization within indoor environments has become inevitable and has attracted a great deal of attention in recent years. Several efforts have been made to cope with the challenges of accurate positioning systems in the presence of signal interference. In this paper, we propose a novel deep learning approach through Gradient Boosting Enhanced with Step-Wise Feature Augmentation using Artificial Neural Network (AugBoost-ANN) for indoor localization applications as it trains over labeled data. For this purpose, we propose an IoT architecture using a star network topology to collect the Received Signal Strength Indicator (RSSI) of Bluetooth Low Energy (BLE) modules by means of a Raspberry Pi as an Access Point (AP) in an indoor environment. The dataset for the experiments is gathered in the real world in different periods to match the real environments. Next, we address the challenges of the AugBoost-ANN training which augments features in each iteration of making a decision tree using a deep neural network and the transfer learning technique. Experimental results show more than 8\% improvement in terms of accuracy in comparison with the existing gradient boosting and deep learning methods recently proposed in the literature, and our proposed model acquires a mean location accuracy of 0.77 m. | ['Pedro H. J. Nardelli', 'Mehdi Rasti', 'Jaber Babaki', 'Ashkan Goharfar'] | 2022-11-16 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-6.76917732e-02 -3.50959808e-01 1.61057517e-01 -6.98267996e-01
-4.58683580e-01 -7.79180005e-02 3.26519161e-01 6.45237043e-02
-6.40184104e-01 1.26957881e+00 -7.03958645e-02 -5.60300648e-01
-6.07397616e-01 -1.05301881e+00 -8.07232976e-01 -7.85259068e-01
-1.14942141e-01 1.73128828e-01 -4.69179451e-02 -1.99038282e-01
-5.92411011e-02 5.30600369e-01 -1.23872674e+00 -1.88031837e-01
9.38511312e-01 1.64543474e+00 1.51842043e-01 5.55876374e-01
9.36614797e-02 7.71037161e-01 -8.20573330e-01 -1.43727273e-01
3.29867572e-01 9.07258876e-03 -3.45948368e-01 -6.75294757e-01
1.34252518e-01 -3.67608108e-02 -1.35092825e-01 5.10960937e-01
1.19881284e+00 8.72253627e-02 2.04141513e-01 -1.36435175e+00
-2.07434282e-01 5.35166562e-01 -3.80773455e-01 4.72445041e-01
2.48963177e-01 -5.60951293e-01 2.18784466e-01 -7.17246115e-01
-2.87806898e-01 5.40559649e-01 1.56909978e+00 1.27688825e-01
-6.73246205e-01 -9.59526002e-01 -3.70406806e-01 5.68386137e-01
-1.48068523e+00 -2.54923552e-01 1.09487045e+00 5.67921028e-02
6.40196800e-01 1.82706028e-01 8.51550698e-01 1.17440104e+00
5.61120391e-01 3.12988430e-01 1.19889510e+00 -5.27596891e-01
6.20654464e-01 9.04863402e-02 -1.35947928e-01 3.90260041e-01
4.19439524e-01 1.23075373e-01 -4.23819244e-01 -1.48652211e-01
4.01272565e-01 5.89132547e-01 1.76695466e-01 -3.20959389e-01
-9.53625143e-01 4.34132129e-01 1.29422176e+00 8.33314657e-01
-6.82155728e-01 4.72356856e-01 3.27197015e-02 2.22534370e-02
1.98436618e-01 7.17164502e-02 -8.10612381e-01 -1.33661821e-01
-8.73712599e-01 -3.02646190e-01 7.62769997e-01 8.08765173e-01
6.48830771e-01 1.78750400e-02 2.08505914e-01 8.23815823e-01
5.62288046e-01 5.73514223e-01 8.30415070e-01 -4.89684552e-01
4.17173982e-01 4.14814830e-01 3.58437002e-01 -1.10565734e+00
-7.71958470e-01 -1.18653905e+00 -1.26173341e+00 -8.70822221e-02
2.52798527e-01 -6.79843187e-01 -5.67403853e-01 1.54201567e+00
4.57917750e-01 5.39787114e-01 -6.85814675e-03 2.43500561e-01
6.82731569e-01 5.78938663e-01 4.19494286e-02 1.51787251e-01
1.09953558e+00 -8.87606740e-01 -5.70609987e-01 -2.49643579e-01
3.29220951e-01 -5.18513441e-01 6.00549519e-01 3.51701826e-01
-5.14467597e-01 -1.10430193e+00 -1.42478943e+00 4.70133871e-01
-8.76157761e-01 1.41456306e-01 5.58439791e-01 1.20625985e+00
-1.03251803e+00 4.40889001e-01 -7.86544085e-01 -5.24344087e-01
4.14267898e-01 9.85627770e-01 -6.51505291e-02 -4.59728949e-03
-1.15388358e+00 6.60593271e-01 2.43319631e-01 4.92826015e-01
-2.32899621e-01 -2.28878140e-01 -4.78038102e-01 2.33170018e-02
-3.74451518e-01 -8.38162541e-01 1.00890684e+00 -8.22468638e-01
-1.63489246e+00 1.98424235e-02 -7.51532242e-02 -6.40397549e-01
3.69773626e-01 -2.38808438e-01 -8.45567942e-01 -7.57424891e-01
6.65706620e-02 2.19762683e-01 5.41344941e-01 -1.13740611e+00
-1.07060397e+00 -6.49097085e-01 -2.01136451e-02 -2.58879387e-04
-4.50229675e-01 -5.44988215e-01 2.60450631e-01 -4.17884231e-01
4.19712991e-01 -8.93340707e-01 -3.02322030e-01 -5.13424933e-01
-7.40538090e-02 -1.67353228e-01 8.65914524e-01 -5.21988928e-01
1.06873298e+00 -1.87494290e+00 -8.03856850e-01 4.57580149e-01
-2.27335170e-01 2.10381418e-01 4.45810586e-01 2.59388208e-01
1.66393548e-01 -2.89441466e-01 1.54473260e-02 -4.20288384e-01
-8.95193368e-02 2.84435719e-01 1.60685241e-01 3.81994963e-01
-6.10542536e-01 8.38594735e-01 -7.94774354e-01 -2.52744377e-01
5.25383592e-01 7.25728333e-01 -1.91706166e-01 9.95258689e-02
6.51206434e-01 1.05735302e+00 -5.02698481e-01 6.49713933e-01
8.92844617e-01 -2.03421071e-01 -9.73735750e-02 -1.28133282e-01
-2.35758349e-01 1.60143107e-01 -1.28277922e+00 1.73447406e+00
-1.19006395e+00 4.49037850e-01 -8.44343845e-03 -1.21337461e+00
1.14418912e+00 1.90784633e-01 6.76994264e-01 -1.06112182e+00
3.21242154e-01 5.34899294e-01 -5.68607226e-02 -2.07009390e-01
6.16331920e-02 4.87287417e-02 -1.16657354e-01 1.36989728e-02
-3.85780409e-02 3.34193259e-01 -4.47136998e-01 -5.04974365e-01
1.28929257e+00 3.58744040e-02 3.30452263e-01 -1.06829740e-01
9.43120480e-01 -3.58147144e-01 4.14979517e-01 1.11057520e+00
-2.34860361e-01 -4.18016054e-02 -9.78922665e-01 -8.72494698e-01
-5.70567250e-01 -8.74277771e-01 -4.60085005e-01 1.11920786e+00
2.73666322e-01 1.14890844e-01 -5.03496587e-01 -6.33233964e-01
4.22528423e-02 3.64370555e-01 -4.23704565e-01 1.59603328e-01
-8.17511857e-01 -1.01905107e+00 5.82580447e-01 7.22801685e-01
1.47677708e+00 -1.00233018e+00 -7.07021952e-01 5.26538730e-01
-2.43262798e-01 -1.00575578e+00 4.38391149e-01 7.44990170e-01
-8.30748200e-01 -6.19028509e-01 -5.49937904e-01 -1.09687316e+00
4.54073578e-01 9.12630409e-02 9.61019456e-01 -1.07338037e-02
3.99865434e-02 3.15347373e-01 -2.39040762e-01 -8.00323248e-01
2.09329829e-01 5.33023655e-01 2.07618847e-01 4.90551107e-02
4.75267380e-01 -1.11596358e+00 -1.15429068e+00 5.74305773e-01
-2.01535180e-01 -4.27880764e-01 9.07744288e-01 7.75350034e-01
3.36025387e-01 4.40640181e-01 1.06251657e+00 -3.59308898e-01
3.92449051e-01 -8.91792893e-01 -5.97628117e-01 -8.05169269e-02
-7.40095973e-01 -3.01393211e-01 9.35745358e-01 -3.69041972e-02
-9.63035762e-01 1.94539279e-01 -7.63564110e-01 6.31736398e-01
-3.65559459e-01 3.03844154e-01 -2.42023587e-01 -6.46546781e-01
7.61603653e-01 2.67275214e-01 -7.44627118e-01 -4.70364124e-01
-2.62176152e-02 1.29346788e+00 4.84243333e-01 -3.90667230e-01
7.39016414e-01 5.27457416e-01 2.41172448e-01 -5.93798280e-01
-7.81555951e-01 -5.42409539e-01 -4.42868739e-01 -2.56333351e-01
5.05939424e-01 -9.27545786e-01 -9.35715914e-01 4.58410591e-01
-7.82528043e-01 -2.40029186e-01 -1.51636675e-01 5.42137027e-01
-2.58377969e-01 -4.62007057e-03 -2.23455057e-01 -8.36251080e-01
-6.68174744e-01 -7.86268711e-01 8.02205265e-01 5.50636411e-01
7.63015673e-02 -9.24192190e-01 1.40212014e-01 4.59071308e-01
1.18524230e+00 2.70855159e-01 2.54598230e-01 -6.87965095e-01
-3.38960141e-01 -6.48217499e-01 2.87363748e-03 2.40199506e-01
3.35136086e-01 -1.04004252e+00 -1.17449355e+00 -3.62712294e-01
3.86521757e-01 5.70173897e-02 1.74622849e-01 6.05117977e-01
1.40628362e+00 -3.20924968e-01 -5.28214276e-01 7.37772882e-01
1.58722818e+00 6.05577707e-01 6.09313667e-01 8.44262362e-01
4.41320837e-01 -9.60298851e-02 3.31954271e-01 4.62597638e-01
4.73272681e-01 7.43928492e-01 7.09453046e-01 -2.91877329e-01
5.11707179e-02 -2.21864328e-01 -2.58922845e-01 6.65756345e-01
-2.70026606e-02 -1.63361207e-01 -9.13134694e-01 3.70365351e-01
-1.90756392e+00 -6.63697004e-01 -6.39894754e-02 2.17285800e+00
3.63955319e-01 2.10592061e-01 -2.25476921e-02 7.45886981e-01
5.64237237e-01 -5.89362979e-02 -3.77870321e-01 -6.55855536e-02
3.03703606e-01 4.70277011e-01 8.95640075e-01 3.88464868e-01
-1.23495328e+00 3.59272063e-01 5.34670305e+00 5.92562318e-01
-1.43717849e+00 4.15914625e-01 5.78696251e-01 5.39863467e-01
5.25638521e-01 -4.32396770e-01 -6.68682516e-01 9.89609659e-01
1.15860271e+00 6.17404699e-01 3.87198180e-01 1.23723495e+00
2.99750715e-02 -1.49211243e-01 -7.12093174e-01 1.27693760e+00
-1.40898243e-01 -9.97892261e-01 -7.41547406e-01 5.20808883e-02
8.73716712e-01 4.09359843e-01 1.53875604e-01 4.88092184e-01
2.43623346e-01 -9.00686800e-01 2.97114819e-01 5.17706573e-01
2.66717583e-01 -8.70513439e-01 1.49913716e+00 6.29126012e-01
-1.28367102e+00 -7.13466585e-01 -2.13376105e-01 -5.71906626e-01
-8.72674808e-02 9.09506559e-01 -1.09779608e+00 6.36603236e-01
1.26591790e+00 3.40856731e-01 -5.67967832e-01 1.42175496e+00
-1.76634207e-01 7.21063197e-01 -7.68834293e-01 -6.03533268e-01
4.69597355e-02 5.38639650e-02 -7.74618387e-02 9.30410743e-01
9.31790531e-01 -1.96273834e-01 -5.93908429e-02 1.54326007e-01
-1.53763875e-01 3.18943113e-02 -6.29982114e-01 1.10979581e+00
7.12478399e-01 1.31843209e+00 -5.03120482e-01 -1.88381359e-01
-2.01496780e-01 8.60044837e-01 -1.85142919e-01 2.02167958e-01
-8.86506200e-01 -5.95568299e-01 -1.07754387e-01 4.15042713e-02
4.25582230e-01 -2.09107906e-01 -6.48361921e-01 -5.94274700e-01
6.79668486e-02 -1.58278331e-01 9.47278067e-02 -6.96212769e-01
-1.28331399e+00 5.88640153e-01 -4.15238172e-01 -1.15885532e+00
-1.31622270e-01 -5.25394320e-01 -6.32004261e-01 5.96211791e-01
-1.58813500e+00 -1.21926308e+00 -1.06976271e+00 8.00896227e-01
3.87241215e-01 -6.23651482e-02 8.82730126e-01 8.99473488e-01
-2.20396444e-01 5.75016797e-01 6.82266593e-01 8.24879706e-02
4.63805437e-01 -1.28186011e+00 2.45162562e-01 6.03697896e-01
2.08677575e-02 5.76637864e-01 5.70335209e-01 -2.91815042e-01
-1.16014683e+00 -1.22626507e+00 9.80493963e-01 -4.64190423e-01
2.03921646e-01 -2.27003053e-01 4.30259816e-02 6.92834020e-01
1.75208136e-01 4.05201644e-01 9.24889147e-01 7.36863911e-02
4.11209047e-01 -1.18105626e+00 -1.71363103e+00 1.47552311e-01
9.60214555e-01 6.25440180e-02 -2.80406237e-01 1.45541146e-01
-4.41559479e-02 -2.72153080e-01 -7.76630044e-01 7.50925839e-01
8.44801903e-01 -9.54920590e-01 1.18020582e+00 2.27038562e-01
-4.83409554e-01 -5.81158280e-01 -4.04800117e-01 -1.35702777e+00
-3.37266237e-01 -2.75744647e-01 -2.94069856e-01 1.10328400e+00
1.41894192e-01 -9.34668541e-01 1.09840357e+00 1.14203759e-01
-3.54281068e-02 -6.77564383e-01 -1.54624212e+00 -5.76371789e-01
-4.16034341e-01 -4.54406947e-01 1.10975850e+00 6.58248842e-01
-2.99915254e-01 5.82672358e-01 -4.35349852e-01 3.04686069e-01
6.04550004e-01 -4.53555197e-01 9.02095079e-01 -1.63006222e+00
-1.24369130e-01 9.97831374e-02 -7.04728663e-01 -1.34669292e+00
-4.35343295e-01 -5.51288784e-01 1.51973158e-01 -1.82260227e+00
-4.06291127e-01 -1.05225277e+00 -9.39979196e-01 3.31882030e-01
3.25130761e-01 5.75751185e-01 -4.02695775e-01 -1.35258615e-01
-8.47016990e-01 4.58021075e-01 4.78570193e-01 -3.48307401e-01
-1.69427440e-01 8.86048079e-01 -4.99684483e-01 8.83361280e-01
1.10173821e+00 -4.76618737e-01 -2.59544045e-01 -5.03112197e-01
3.08672160e-01 -1.60560995e-01 3.12597454e-01 -2.14032292e+00
6.17593348e-01 5.24124742e-01 1.34077930e+00 -7.50976264e-01
4.04689640e-01 -1.84425390e+00 3.93231928e-01 7.54634321e-01
2.38434732e-01 2.56018817e-01 8.54499117e-02 5.78372061e-01
7.20218867e-02 4.64342386e-02 4.48732764e-01 1.55225307e-01
-6.05644524e-01 2.34704465e-02 -2.48466536e-01 -8.54349434e-01
9.70774591e-01 -4.25474942e-01 -7.17552528e-02 -4.85554188e-01
-5.45511305e-01 -9.89726633e-02 -3.11411005e-02 1.16545193e-01
3.94088984e-01 -1.63892591e+00 -4.11056988e-02 2.51352608e-01
-4.48963344e-02 -5.75389266e-02 9.39976051e-02 7.62826741e-01
-4.57692891e-01 4.58547264e-01 -3.52145582e-01 -6.82316542e-01
-8.63848209e-01 1.81403518e-01 6.23722851e-01 -4.41925675e-01
-6.41695410e-02 6.80209637e-01 -7.81203389e-01 -7.95298398e-01
6.20681643e-01 -3.68277729e-01 -3.48459274e-01 -4.24122393e-01
2.14280143e-01 6.80987477e-01 6.01016045e-01 -5.92309713e-01
-6.79693520e-01 7.37212658e-01 5.73921382e-01 8.46071392e-02
1.44288599e+00 -2.66409487e-01 9.58372131e-02 2.33759314e-01
1.06600595e+00 3.42408478e-01 -7.85644174e-01 -1.28809318e-01
2.04419106e-01 -2.04088077e-01 -1.98866669e-02 -1.25441599e+00
-7.49042809e-01 5.84393859e-01 1.67071676e+00 4.11175847e-01
1.14728379e+00 -3.26779217e-01 1.08073199e+00 8.00424159e-01
1.02302301e+00 -9.60133016e-01 -1.87193498e-01 2.97193497e-01
3.36487979e-01 -1.41426218e+00 -2.68261969e-01 1.80197224e-01
3.70874107e-01 8.87371540e-01 3.78114402e-01 -1.55963331e-01
1.13270128e+00 1.89641789e-01 1.54937059e-01 1.56411812e-01
3.46768647e-01 5.04424088e-02 -1.70540392e-01 8.87056947e-01
3.17087263e-01 1.13249384e-02 -2.28614345e-01 7.19969153e-01
-4.87193525e-01 3.18813652e-01 -3.73998344e-01 1.28739297e+00
-5.08682370e-01 -1.10731995e+00 -6.45092726e-01 4.12493169e-01
-7.35067368e-01 1.80341437e-01 2.04358146e-01 4.26769584e-01
8.60219240e-01 1.40524900e+00 -1.87100708e-01 -6.09205067e-01
3.43176156e-01 -1.61272749e-01 2.30406359e-01 1.28619494e-02
-6.68574452e-01 -4.25379872e-01 -1.77289486e-01 -1.99402094e-01
-5.45199513e-01 -1.16237871e-01 -1.03240120e+00 -1.19306415e-01
-3.64531636e-01 4.53941047e-01 1.42287219e+00 9.81964290e-01
5.18212259e-01 8.53638053e-01 1.00752652e+00 -1.10587478e+00
-2.38536626e-01 -1.15515053e+00 -2.37620756e-01 -5.75128384e-02
4.26882893e-01 -5.37433803e-01 -1.53420851e-01 -4.87015307e-01] | [6.427677631378174, 0.9303526878356934] |
4c13f85b-15e6-461a-b0f4-791bb9f557cc | improving-rgb-d-point-cloud-registration-by | 2208.14893 | null | https://arxiv.org/abs/2208.14893v2 | https://arxiv.org/pdf/2208.14893v2.pdf | Improving RGB-D Point Cloud Registration by Learning Multi-scale Local Linear Transformation | Point cloud registration aims at estimating the geometric transformation between two point cloud scans, in which point-wise correspondence estimation is the key to its success. In addition to previous methods that seek correspondences by hand-crafted or learnt geometric features, recent point cloud registration methods have tried to apply RGB-D data to achieve more accurate correspondence. However, it is not trivial to effectively fuse the geometric and visual information from these two distinctive modalities, especially for the registration problem. In this work, we propose a new Geometry-Aware Visual Feature Extractor (GAVE) that employs multi-scale local linear transformation to progressively fuse these two modalities, where the geometric features from the depth data act as the geometry-dependent convolution kernels to transform the visual features from the RGB data. The resultant visual-geometric features are in canonical feature spaces with alleviated visual dissimilarity caused by geometric changes, by which more reliable correspondence can be achieved. The proposed GAVE module can be readily plugged into recent RGB-D point cloud registration framework. Extensive experiments on 3D Match and ScanNet demonstrate that our method outperforms the state-of-the-art point cloud registration methods even without correspondence or pose supervision. The code is available at: https://github.com/514DNA/LLT. | ['Dong Xu', 'Lu Sheng', 'Jing Zhang', 'Zhenghao Chen', 'Xiaoliang Huo', 'ZiMing Wang'] | 2022-08-31 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-1.68288335e-01 -4.04187828e-01 -1.20455259e-02 -5.13754189e-01
-9.52992022e-01 -4.78497624e-01 6.00722730e-01 2.13706002e-01
-4.14361566e-01 5.59063479e-02 -7.28115514e-02 4.44164798e-02
-1.43625826e-01 -9.15180206e-01 -6.47994936e-01 -6.67591214e-01
2.91611493e-01 5.48707485e-01 1.86901495e-01 -2.32261375e-01
2.96730280e-01 9.76352930e-01 -1.52972293e+00 -1.99818879e-01
8.46544027e-01 1.04859090e+00 2.10621402e-01 1.57224774e-01
-3.34330350e-01 -2.08345521e-02 7.85341486e-02 -2.00696290e-01
5.98667562e-01 -5.13733141e-02 -5.82577109e-01 5.36589585e-02
7.00656474e-01 -3.22468281e-01 -4.35299009e-01 1.01795471e+00
5.44119895e-01 1.43993422e-01 3.51687223e-01 -1.26334774e+00
-4.52311575e-01 -2.25187853e-01 -8.24539959e-01 -1.55109853e-01
5.02373755e-01 2.45144859e-01 7.80078948e-01 -1.22810924e+00
5.13281345e-01 1.18612564e+00 6.59969091e-01 2.39231497e-01
-1.03192842e+00 -8.59885633e-01 -2.50777036e-01 1.44138813e-01
-1.72626626e+00 -3.11413795e-01 1.13342094e+00 -3.27440530e-01
8.50976825e-01 4.07042712e-01 8.99828196e-01 4.46781486e-01
-2.26147356e-04 3.09377730e-01 9.87314284e-01 -2.81765521e-01
-1.16245538e-01 -2.04378456e-01 -3.04947197e-01 6.87875390e-01
7.23565146e-02 2.52460420e-01 -4.25028920e-01 -2.45895341e-01
1.22353852e+00 6.63479090e-01 -3.71555895e-01 -7.46760905e-01
-1.52984571e+00 6.80125713e-01 1.12542558e+00 2.38223121e-01
-4.40030634e-01 2.01657668e-01 -4.48828861e-02 1.59145415e-01
5.58973253e-01 8.82504880e-02 -2.76998907e-01 -4.42673229e-02
-7.47983396e-01 7.89738148e-02 2.63720334e-01 1.05404902e+00
1.37763727e+00 -3.91427815e-01 1.88688874e-01 6.76659226e-01
6.19312644e-01 7.85570443e-01 2.45230258e-01 -8.19614291e-01
5.07977784e-01 9.58763540e-01 -1.03016987e-01 -1.29070926e+00
-3.17945510e-01 -1.07319683e-01 -8.11463833e-01 2.92153805e-01
2.43082017e-01 5.62874317e-01 -7.75369346e-01 1.19516075e+00
7.25067019e-01 4.11410064e-01 -3.14824164e-01 1.14393055e+00
8.83478045e-01 3.38552147e-01 -1.87674329e-01 2.42441908e-01
1.34127748e+00 -4.90478545e-01 -2.86072969e-01 -8.09698552e-02
3.64628553e-01 -1.01797724e+00 9.21145380e-01 -2.21117988e-01
-9.48756754e-01 -4.53835934e-01 -9.74430621e-01 -4.10308063e-01
-2.28881896e-01 -5.69577515e-02 4.70248818e-01 2.63933867e-01
-9.29042280e-01 4.51209038e-01 -1.13022912e+00 -3.42816740e-01
3.91935229e-01 5.39355934e-01 -7.54624426e-01 -3.37844133e-01
-7.94460952e-01 7.96372056e-01 1.18170440e-01 3.79758209e-01
-1.61721393e-01 -7.47612596e-01 -9.70986247e-01 -2.18243673e-01
1.81768328e-01 -7.97700405e-01 8.50088120e-01 -4.66053247e-01
-1.22005594e+00 1.11325264e+00 -2.74480373e-01 3.64268690e-01
4.71998900e-01 -2.98002213e-02 -7.33219907e-02 1.48276627e-01
1.25202715e-01 6.01030409e-01 7.41643310e-01 -1.44772267e+00
-4.51265633e-01 -6.63644075e-01 -2.20996633e-01 3.50149006e-01
-7.75923505e-02 -1.67268813e-01 -7.63284683e-01 -3.50913972e-01
9.07855690e-01 -9.60372686e-01 -1.63020596e-01 5.82653582e-01
-1.69249013e-01 -2.34426975e-01 8.14079762e-01 -5.39770901e-01
5.87673903e-01 -2.27205706e+00 9.96439457e-02 6.16873741e-01
4.42960620e-01 1.90111790e-02 -1.31523803e-01 3.48114133e-01
-1.00978702e-01 -1.85267538e-01 -2.09429935e-01 -4.59729284e-01
-7.69365579e-02 1.44888923e-01 -3.89911532e-02 9.56642270e-01
2.51860440e-01 1.10348153e+00 -9.42685902e-01 -5.59316218e-01
7.06983030e-01 8.17244589e-01 -3.52765173e-01 2.75099784e-01
1.99847788e-01 8.08380723e-01 -7.57846415e-01 8.33276689e-01
1.09442747e+00 -7.87751153e-02 -4.57324982e-01 -5.51818073e-01
-1.74035549e-01 1.35694832e-01 -1.05598545e+00 2.21851754e+00
-4.24214900e-01 3.23837638e-01 -1.86315492e-01 -6.62620425e-01
1.19560945e+00 1.55712605e-01 7.86303580e-01 -7.25116789e-01
2.88327396e-01 3.27528924e-01 -3.16346318e-01 -2.30681509e-01
2.61499166e-01 -1.17345192e-01 4.74260040e-02 1.26856819e-01
-1.20404623e-01 -6.63022935e-01 -5.59026837e-01 -5.97643107e-02
8.19946051e-01 2.26950347e-01 3.82965744e-01 1.49343252e-01
6.17765546e-01 8.75572190e-02 4.56284612e-01 1.73740178e-01
7.62482584e-02 9.39092636e-01 -2.58874893e-01 -5.64204514e-01
-1.02195954e+00 -1.21606278e+00 -2.43600920e-01 2.87918836e-01
6.29319787e-01 -3.08769941e-01 -4.01312023e-01 -3.97791743e-01
2.35093310e-01 8.69485438e-02 -4.07428086e-01 -5.04464731e-02
-6.93011105e-01 -5.31807542e-02 1.78877383e-01 3.94033909e-01
6.07973218e-01 -5.87450266e-01 -4.07561034e-01 6.62168348e-03
-1.24480687e-01 -1.12117243e+00 -6.96807086e-01 -2.80278385e-01
-1.11369455e+00 -1.02445436e+00 -5.39342880e-01 -6.97868705e-01
9.97392535e-01 8.19679201e-01 6.99697495e-01 4.25142348e-01
-3.07137996e-01 5.35823047e-01 -3.80447447e-01 -1.37918025e-01
-8.50246474e-02 1.84668656e-02 -3.76095362e-02 4.15320173e-02
5.65759957e-01 -9.17830706e-01 -7.74152517e-01 5.13157725e-01
-8.93821418e-01 2.48891681e-01 6.33371234e-01 6.27097726e-01
9.65987623e-01 -4.27734017e-01 -2.40428373e-01 -2.46165171e-01
8.88407305e-02 -2.56017536e-01 -5.58753312e-01 1.52858019e-01
-3.24916869e-01 6.83056936e-02 1.50601849e-01 -1.83229104e-01
-6.74241543e-01 4.83421385e-01 -1.97258562e-01 -9.29988682e-01
-1.98529929e-01 4.64978218e-01 -2.45669737e-01 -7.02729821e-01
4.28365648e-01 3.48795801e-01 2.27124602e-01 -5.45833766e-01
4.93995428e-01 6.11150742e-01 6.21410191e-01 -6.23842359e-01
1.54545200e+00 8.10882866e-01 1.35059208e-01 -7.26453602e-01
-3.48201811e-01 -8.27205837e-01 -1.07441354e+00 -1.91922396e-01
8.04997802e-01 -1.10256422e+00 -6.92123175e-01 4.32122558e-01
-1.26385188e+00 1.58686668e-01 -1.59113109e-01 6.53030217e-01
-5.66272795e-01 5.59591115e-01 -2.74770945e-01 -3.22961897e-01
-3.90842795e-01 -1.21855497e+00 1.49561608e+00 3.12917620e-01
1.33875951e-01 -7.80237973e-01 1.77746549e-01 2.80696243e-01
3.18088323e-01 5.01896739e-01 5.52910388e-01 -1.68567538e-01
-9.88247931e-01 -5.05041957e-01 -3.58425081e-01 2.06126720e-02
4.68351126e-01 -6.10930473e-02 -8.38931739e-01 -4.34767067e-01
1.07494742e-01 1.19581580e-01 4.26647604e-01 1.24984626e-02
9.16351557e-01 1.60142764e-01 -3.35206032e-01 1.16183138e+00
1.53135264e+00 -2.63754845e-01 6.76507235e-01 5.09792984e-01
1.18797040e+00 3.44244242e-01 7.80236244e-01 3.85396183e-01
6.11748695e-01 1.02148020e+00 6.19887888e-01 -3.99161875e-01
-1.04519486e-01 -4.14415002e-01 -3.99689414e-02 1.17526829e+00
-5.17905593e-01 4.71756458e-01 -1.16994262e+00 1.94471747e-01
-1.83973312e+00 -6.17357612e-01 -2.55811840e-01 2.34221077e+00
7.69162059e-01 -2.82486200e-01 -2.66707480e-01 -7.60505944e-02
6.74458444e-01 -1.73241831e-02 -4.53638941e-01 1.16209529e-01
4.20565233e-02 3.98025632e-01 6.05562747e-01 3.07421684e-01
-9.70248520e-01 8.27057838e-01 4.32955885e+00 8.04885507e-01
-1.27831876e+00 1.81234628e-01 -6.30838610e-03 2.93674693e-02
-4.18252498e-01 2.52773702e-01 -3.75583887e-01 1.84024975e-01
2.79643744e-01 -1.43713489e-01 3.27914387e-01 7.54230201e-01
1.00085318e-01 9.38294381e-02 -1.04529333e+00 1.51577663e+00
-5.04091717e-02 -1.23419380e+00 -5.30651957e-02 3.22199076e-01
4.55112457e-01 3.27297032e-01 -1.43102795e-01 -2.75473863e-01
-3.58106457e-02 -8.37507546e-01 6.93832755e-01 6.27864897e-01
9.47941601e-01 -6.01942658e-01 5.71222305e-01 2.22959384e-01
-1.56677830e+00 4.24909145e-01 -4.88195151e-01 6.91694841e-02
7.79059306e-02 5.58349550e-01 -7.42170155e-01 9.42371488e-01
9.03020442e-01 9.56833124e-01 -6.18654013e-01 1.27912128e+00
-2.67355651e-01 -3.36206518e-02 -5.64182997e-01 4.35361624e-01
2.82129273e-02 -4.78839934e-01 4.84328449e-01 6.64874434e-01
6.31390750e-01 2.34910369e-01 1.90836564e-01 9.46961641e-01
6.59190789e-02 2.08913773e-01 -5.57940602e-01 4.80220735e-01
5.95054805e-01 1.48102736e+00 -7.51035929e-01 5.45946062e-02
-6.66895688e-01 1.03087234e+00 3.32240462e-01 1.15679629e-01
-6.34335339e-01 -2.59300470e-01 9.92588341e-01 2.01566458e-01
1.32433295e-01 -6.81207955e-01 -2.70741105e-01 -1.28243434e+00
2.65738964e-01 -4.98259038e-01 -7.11627081e-02 -7.81389177e-01
-1.37236631e+00 5.19523501e-01 2.68410426e-02 -1.90039754e+00
-1.54555030e-02 -3.59696627e-01 -5.55605829e-01 1.19548476e+00
-1.57115078e+00 -1.46010244e+00 -9.08336043e-01 9.37127471e-01
7.38379434e-02 2.42616564e-01 6.95781946e-01 3.53227377e-01
-1.39290884e-01 4.18539584e-01 -3.46824862e-02 2.75067776e-01
7.99116075e-01 -9.44760978e-01 6.95700049e-01 6.53523982e-01
1.67661190e-01 8.00180256e-01 2.82416821e-01 -5.68304360e-01
-1.84343779e+00 -9.50358987e-01 5.68207502e-01 -5.28546512e-01
3.33066553e-01 -3.73292029e-01 -1.08794975e+00 4.78129029e-01
-2.85639673e-01 5.79335749e-01 4.53419507e-01 -3.97720516e-01
-5.02870739e-01 -2.27348998e-01 -1.15610921e+00 2.87458569e-01
1.17813396e+00 -7.70015836e-01 -5.40568054e-01 1.59932643e-01
5.88213384e-01 -9.18598473e-01 -1.20471549e+00 4.03166205e-01
5.50988555e-01 -7.43104756e-01 1.34469986e+00 3.63147371e-02
8.48555937e-02 -7.90038943e-01 -1.82711124e-01 -1.20744586e+00
-2.26732641e-01 -2.02675506e-01 3.69428456e-01 1.15964031e+00
-1.66650116e-01 -8.18379223e-01 6.59908712e-01 7.02187538e-01
-2.15509504e-01 -5.25583327e-01 -1.27963102e+00 -7.01617897e-01
-1.43618777e-01 -4.95615572e-01 9.79585290e-01 1.15040374e+00
-2.76274562e-01 -1.42475978e-01 -4.80379583e-03 4.75354344e-01
8.23154271e-01 4.21934575e-01 1.05515301e+00 -1.38551879e+00
1.56366870e-01 -3.97467643e-01 -1.02155530e+00 -1.05384314e+00
-2.74398196e-02 -1.14169502e+00 1.05947942e-01 -1.40954483e+00
1.30797112e-02 -8.87267172e-01 -9.55074579e-02 6.37259364e-01
-1.38202712e-01 5.12510240e-01 2.50136554e-01 5.87445319e-01
-1.29623964e-01 7.93449044e-01 1.40024829e+00 -1.05329029e-01
-2.54284263e-01 -1.38141185e-01 -3.97326440e-01 4.65837836e-01
7.31263041e-01 -4.58926052e-01 -4.68398966e-02 -6.15412176e-01
6.47575110e-02 -1.20838277e-01 7.03512132e-01 -9.52384591e-01
5.62018931e-01 -2.54990608e-01 3.83883834e-01 -6.80148184e-01
4.25363988e-01 -1.28706563e+00 5.92442930e-01 2.01805532e-01
2.09617138e-01 1.82517126e-01 9.53657702e-02 4.50435519e-01
-3.76794070e-01 2.34013274e-01 5.25110126e-01 -1.23322576e-01
-7.02698052e-01 9.29964960e-01 5.76916754e-01 -3.76675069e-01
1.01326096e+00 -6.10144973e-01 -1.12508573e-01 -1.85582057e-01
-3.60914439e-01 9.28646401e-02 1.04103422e+00 5.31625926e-01
1.11994863e+00 -1.76007032e+00 -7.03640342e-01 4.41711038e-01
5.32215059e-01 5.28577805e-01 3.35347623e-01 1.12989342e+00
-8.02595735e-01 6.14491329e-02 -2.89631933e-01 -1.11801183e+00
-1.31450033e+00 2.67266273e-01 3.26721132e-01 2.77031898e-01
-7.04065382e-01 5.76113284e-01 4.99595739e-02 -7.27306843e-01
-1.13712199e-01 -4.20575827e-01 2.60675460e-01 -2.63140798e-01
2.82139570e-01 1.54817283e-01 3.97584498e-01 -1.08867693e+00
-7.05356777e-01 1.39722824e+00 1.87694758e-01 1.09271958e-01
1.57745600e+00 -2.52019376e-01 -4.01988834e-01 2.29695573e-01
1.57200289e+00 -3.07384646e-03 -1.19190466e+00 -6.89036310e-01
-2.52498031e-01 -1.07461691e+00 2.27055341e-01 -7.48997405e-02
-1.38915634e+00 9.36407506e-01 8.16580772e-01 -2.12569565e-01
1.17510080e+00 2.40970403e-01 8.50530744e-01 8.61645192e-02
7.73050845e-01 -4.14117962e-01 -2.58665413e-01 1.92141011e-01
9.69157040e-01 -1.42142987e+00 3.42655778e-01 -7.75178552e-01
-1.86452508e-01 1.31960618e+00 4.30455744e-01 -2.95783192e-01
6.66712701e-01 -8.92079547e-02 1.65809482e-01 -4.25474495e-01
-2.02534553e-02 -3.56404603e-01 6.34980381e-01 6.70923412e-01
2.02648252e-01 -4.95093502e-03 -2.70595849e-02 7.99235255e-02
-2.88328499e-01 -1.25442863e-01 -4.73116189e-02 1.03462076e+00
-1.97863653e-01 -1.42145562e+00 -6.60865247e-01 2.51188248e-01
2.36331150e-01 3.45199220e-02 -1.76565379e-01 7.79949009e-01
1.18677855e-01 5.54136574e-01 2.26282820e-01 -6.31783009e-01
5.44360340e-01 -2.13494658e-01 5.51988900e-01 -4.73228067e-01
-4.60981876e-01 1.63866848e-01 -5.35458744e-01 -1.03702986e+00
-7.76942551e-01 -7.73387849e-01 -1.32575059e+00 -4.79981035e-01
-3.74157548e-01 -1.39204502e-01 8.77576351e-01 7.93329000e-01
3.90918672e-01 -1.19544677e-02 8.35769415e-01 -1.42375743e+00
-1.61788017e-01 -5.89720130e-01 -4.35723543e-01 6.31926000e-01
3.97542506e-01 -8.16526115e-01 -3.17809463e-01 -6.32066354e-02] | [7.678435802459717, -3.014394998550415] |
62e5215d-11b6-4037-8e67-865f710cebb3 | an-upper-bound-for-the-distribution-overlap | 2212.08701 | null | https://arxiv.org/abs/2212.08701v2 | https://arxiv.org/pdf/2212.08701v2.pdf | An Upper Bound for the Distribution Overlap Index and Its Applications | This paper proposes an easy-to-compute upper bound for the overlap index between two probability distributions without requiring any knowledge of the distribution models. The computation of our bound is time-efficient and memory-efficient and only requires finite samples. The proposed bound shows its value in one-class classification and domain shift analysis. Specifically, in one-class classification, we build a novel one-class classifier by converting the bound into a confidence score function. Unlike most one-class classifiers, the training process is not needed for our classifier. Additionally, the experimental results show that our classifier can be accurate with only a small number of in-class samples and outperform many state-of-the-art methods on various datasets in different one-class classification scenarios. In domain shift analysis, we propose a theorem based on our bound. The theorem is useful in detecting the existence of domain shift and inferring data information. The detection and inference processes are both computation-efficient and memory-efficient. Our work shows significant promise toward broadening the applications of overlap-based metrics. | ['Farshad Khorrami', 'Siddharth Garg', 'Prashanth Krishnamurthy', 'Hao Fu'] | 2022-12-16 | null | null | null | null | ['one-class-classifier', 'one-class-classification'] | ['methodology', 'miscellaneous'] | [ 4.20571893e-01 -2.65275210e-01 -5.30236244e-01 -5.43047488e-01
-8.78472686e-01 -7.62001812e-01 3.37788194e-01 4.02617604e-01
-3.21096808e-01 8.69444072e-01 -7.60579646e-01 -5.42411745e-01
-4.42051649e-01 -8.81233633e-01 -5.15147567e-01 -8.12275589e-01
-1.03989281e-01 7.54326582e-01 7.04744399e-01 3.25235903e-01
4.90230739e-01 5.36000192e-01 -1.81811106e+00 1.41254619e-01
8.48842561e-01 1.44376385e+00 -8.02005678e-02 7.02883303e-01
3.74362385e-03 3.01343679e-01 -9.09863293e-01 -2.78269887e-01
3.44903260e-01 -3.83032590e-01 -7.25246072e-01 -1.12486355e-01
3.93587857e-01 -1.72285974e-01 1.20328844e-01 1.25452089e+00
2.52660036e-01 -1.41783487e-02 9.36858475e-01 -1.73155296e+00
-1.59830809e-01 3.80575180e-01 -6.91954255e-01 3.25564057e-01
4.38079447e-01 -5.56160808e-01 7.86114573e-01 -6.94984734e-01
1.23996392e-01 1.10203171e+00 7.31865525e-01 2.32647676e-02
-1.07741618e+00 -9.26612496e-01 -1.17800318e-01 3.74961674e-01
-1.64011049e+00 -8.27740729e-02 3.93855482e-01 -5.10227680e-01
7.48862624e-01 4.40752715e-01 3.14640373e-01 5.55236459e-01
8.16337913e-02 7.31162667e-01 1.20252562e+00 -6.47682607e-01
5.88508964e-01 3.10249627e-01 3.62775654e-01 5.27067125e-01
6.77641809e-01 6.69912472e-02 -2.72391260e-01 -6.90700531e-01
6.43040717e-01 1.84186503e-01 -3.19253094e-02 -3.21195096e-01
-8.89562368e-01 7.88017571e-01 -2.49574378e-01 1.35719284e-01
5.85266314e-02 -1.33678064e-01 3.37780535e-01 2.93243170e-01
6.60172224e-01 1.48916855e-01 -5.33255577e-01 -3.13402653e-01
-1.14133966e+00 3.28505248e-01 1.08183134e+00 1.13253272e+00
8.05477917e-01 -5.36491036e-01 -6.49331063e-02 7.38704205e-01
-1.31035879e-01 6.91966057e-01 3.25009674e-01 -6.94415808e-01
2.66369194e-01 4.94251043e-01 2.13707179e-01 -9.09205556e-01
-7.19064772e-02 -3.10487986e-01 -5.84845185e-01 -5.82949892e-02
7.87711143e-01 2.11729050e-01 -5.59806824e-01 1.58180702e+00
3.92993689e-01 2.81392634e-01 -1.14791237e-01 4.37558204e-01
-2.09387811e-03 3.17447275e-01 -2.46704772e-01 -3.89213055e-01
1.55875099e+00 -4.64162439e-01 -5.33226728e-01 4.66831662e-02
7.19322205e-01 -7.88827002e-01 8.45824599e-01 5.51551580e-01
-8.07209253e-01 -3.37218314e-01 -1.20664120e+00 4.33843851e-01
-5.85889161e-01 1.13767333e-01 7.79091001e-01 1.13630700e+00
-3.38135570e-01 5.52787185e-01 -5.92301011e-01 -3.98670346e-01
5.01220942e-01 2.77236909e-01 -1.48942664e-01 -1.96496710e-01
-1.00471199e+00 7.11696744e-01 3.85580420e-01 -2.95449287e-01
-3.86464536e-01 -6.62725747e-01 -5.01872420e-01 5.54923862e-02
3.33779812e-01 -2.61656344e-01 1.20136559e+00 -5.34453869e-01
-1.20151091e+00 8.94770324e-01 -2.54669815e-01 -4.17406261e-01
5.77235997e-01 5.44251688e-02 -5.76620817e-01 1.58104196e-01
2.38340348e-01 -2.95626596e-02 7.43358374e-01 -7.13445246e-01
-1.06250250e+00 -5.04132807e-01 -1.47059023e-01 -2.71511078e-01
-1.79266855e-01 3.72669734e-02 -1.78300604e-01 -4.30303752e-01
3.42319191e-01 -8.01277459e-01 3.32898982e-02 1.26420379e-01
-2.59120077e-01 -4.60837811e-01 8.25518250e-01 -1.92416042e-01
1.56326294e+00 -2.07809019e+00 -7.41666853e-01 5.19512057e-01
1.52339861e-01 2.90725321e-01 2.74277091e-01 2.26286963e-01
1.34985954e-01 3.81667092e-02 -4.22707409e-01 1.95978999e-01
1.13379322e-01 1.60484388e-01 -5.07139802e-01 7.22620904e-01
1.89183298e-02 2.08980292e-01 -8.69609892e-01 -5.36884189e-01
5.36257513e-02 1.68183371e-02 -4.53734010e-01 1.31579235e-01
5.18050827e-02 -1.49043441e-01 -2.67749935e-01 5.71845591e-01
1.33641469e+00 -3.70182425e-01 3.68932903e-01 -4.55276035e-02
5.54027855e-02 3.21942002e-01 -1.51838934e+00 1.03505242e+00
-3.12401742e-01 5.43063223e-01 -3.02933067e-01 -1.39730406e+00
1.08291888e+00 -1.24222852e-01 2.73878485e-01 -3.89719695e-01
9.60235968e-02 4.85970467e-01 -7.99194723e-02 -1.50631025e-01
2.30127573e-01 -1.25723377e-01 -2.29201004e-01 6.34247243e-01
-2.22641826e-01 1.24635786e-01 3.92517596e-01 -1.03103347e-01
1.08429432e+00 -2.62476712e-01 6.94780111e-01 -4.13434774e-01
7.14804471e-01 -2.65402168e-01 5.78160763e-01 9.10479128e-01
-4.03464198e-01 2.22840250e-01 6.35280013e-01 -2.49814227e-01
-7.00250745e-01 -1.23501730e+00 -6.85804188e-01 1.15894568e+00
4.15747523e-01 -1.69929072e-01 -7.26030290e-01 -6.92852676e-01
3.01331460e-01 5.68656504e-01 -5.77931285e-01 -5.84116057e-02
-2.33017504e-01 -8.20498168e-01 7.80533791e-01 6.63532197e-01
4.40342575e-01 -2.06466094e-01 -6.61965728e-01 8.91289767e-03
-1.55763656e-01 -1.23443377e+00 -3.31902087e-01 2.17954189e-01
-8.96262884e-01 -1.51144624e+00 -4.68079478e-01 -8.45892191e-01
6.05290174e-01 3.11657310e-01 8.50379169e-01 -1.53464854e-01
-2.16924772e-01 2.10481972e-01 -2.91244328e-01 -6.00338936e-01
-2.49989972e-01 5.52650578e-02 1.42253980e-01 -1.05464101e-01
9.30325031e-01 -6.68976426e-01 -4.63297814e-01 8.52860332e-01
-5.68461359e-01 -5.09264112e-01 5.90353131e-01 7.70976424e-01
7.18969882e-01 3.69942874e-01 8.06261003e-01 -8.88679206e-01
6.10089600e-01 -6.40452981e-01 -9.71071124e-01 4.98688132e-01
-7.81432867e-01 2.38155872e-01 6.62352920e-01 -7.34125376e-01
-8.24119925e-01 2.22074762e-02 4.00618583e-01 -2.33776808e-01
-1.41123578e-01 7.03583658e-02 -1.62580833e-01 9.35343280e-02
6.95328772e-01 3.13430935e-01 -5.93379177e-02 -5.34812748e-01
1.10105261e-01 1.09983170e+00 6.22839868e-01 -8.83012712e-01
4.35180843e-01 5.49839079e-01 3.73999894e-01 -7.79760659e-01
-9.52288330e-01 -8.66847515e-01 -7.36572981e-01 1.65384308e-01
2.52994299e-01 -6.00263119e-01 -1.02113986e+00 3.56283426e-01
-9.71971214e-01 7.66512528e-02 -6.74307868e-02 4.98856485e-01
-5.21213949e-01 6.20352566e-01 -1.12225577e-01 -1.18763185e+00
-1.91394776e-01 -8.62525642e-01 1.02823150e+00 -1.68128908e-02
-1.28336638e-01 -9.89795923e-01 3.35669331e-02 3.60354856e-02
3.71052921e-02 9.93581116e-02 9.53433871e-01 -1.19128859e+00
-2.91700989e-01 -6.23943090e-01 -4.70126867e-01 2.84033477e-01
6.48973957e-02 -1.48680657e-01 -1.02520430e+00 -2.81907290e-01
-1.61026686e-01 -8.95848647e-02 6.65004432e-01 2.46052191e-01
1.68221760e+00 -2.30713695e-01 -6.92069411e-01 3.43829185e-01
1.25701261e+00 3.29059809e-01 4.03217733e-01 -4.51998785e-02
4.70098741e-02 4.68564242e-01 1.00861084e+00 8.94538164e-01
9.75895450e-02 6.15755737e-01 4.16136952e-03 3.67029399e-01
1.31438926e-01 -6.29624948e-02 6.66319355e-02 4.32827920e-01
1.82314172e-01 -1.26754001e-01 -9.13006306e-01 5.83505332e-01
-1.78550220e+00 -1.10166776e+00 -2.52985537e-01 2.78179288e+00
9.25983787e-01 1.55908510e-01 3.13265651e-01 6.90584123e-01
1.10551333e+00 -3.76067191e-01 -6.00046992e-01 -3.49539667e-01
1.17328659e-01 4.66902822e-01 6.27998352e-01 4.87774312e-01
-1.11540103e+00 4.51439261e-01 7.28699875e+00 1.18605316e+00
-8.61182034e-01 -3.48458178e-02 3.92231911e-01 9.42314416e-02
2.17705760e-02 1.32190421e-01 -9.89434123e-01 7.50630081e-01
9.77876365e-01 -5.10293007e-01 1.74770072e-01 1.22642612e+00
-4.04326320e-01 -5.82580268e-01 -1.51225293e+00 1.13372850e+00
-2.64412630e-02 -1.02293730e+00 -2.81087846e-01 2.15782493e-01
3.34342450e-01 -5.40173650e-01 -1.18791595e-01 1.94523722e-01
6.23697368e-03 -7.72822022e-01 5.71603417e-01 1.50745332e-01
1.11242509e+00 -1.01582623e+00 8.96019280e-01 4.76913035e-01
-1.33988142e+00 -9.97569039e-02 -6.54179275e-01 -3.11478674e-01
-3.08616281e-01 1.10796344e+00 -1.15200341e+00 3.84061396e-01
4.73429769e-01 3.01416874e-01 -4.61088687e-01 1.16043949e+00
6.79077581e-02 5.97020924e-01 -7.03368485e-01 -4.35908705e-01
-3.09701025e-01 2.59376876e-02 3.42090547e-01 1.32351625e+00
4.10373837e-01 -6.95456564e-02 3.63339543e-01 6.24196112e-01
1.66025445e-01 -1.46541268e-01 -4.25969720e-01 1.05196640e-01
1.17146492e+00 7.90126562e-01 -8.84071410e-01 -4.17557061e-01
-2.84820408e-01 8.27600300e-01 1.77642420e-01 4.97167334e-02
-9.35051262e-01 -1.06493819e+00 7.58020520e-01 1.65649623e-01
4.68323082e-01 2.37182621e-03 -5.45816600e-01 -8.41358125e-01
2.58358955e-01 -3.59826177e-01 6.52037978e-01 -1.16884395e-01
-1.64767110e+00 1.67590022e-01 4.75313216e-01 -1.36959672e+00
-3.52089643e-01 -9.39135909e-01 -2.17433989e-01 7.69053638e-01
-1.30898380e+00 -7.29620576e-01 -2.49643430e-01 5.93875825e-01
-4.07124218e-03 -1.25312865e-01 8.55915308e-01 4.11544502e-01
-4.59244519e-01 1.13548088e+00 3.25401694e-01 1.68481141e-01
9.20761406e-01 -1.13897717e+00 -6.68385401e-02 8.20120335e-01
-2.90369570e-01 5.20025611e-01 7.19552755e-01 -4.35845464e-01
-1.06542706e+00 -8.04213762e-01 8.47245336e-01 -3.59152526e-01
6.68640554e-01 -4.14462298e-01 -8.15827727e-01 3.76441538e-01
-4.69038248e-01 2.78263479e-01 1.15028799e+00 4.27619129e-01
-8.03241074e-01 -2.81947345e-01 -1.48429275e+00 -9.14200917e-02
1.13292551e+00 -4.94045377e-01 -6.72151506e-01 3.40635329e-01
2.32465982e-01 -2.94887066e-01 -1.00640404e+00 3.21260661e-01
9.74946201e-01 -8.09065461e-01 8.18120837e-01 -5.50929666e-01
-6.54767035e-03 -3.56231332e-01 -5.55520833e-01 -7.51821816e-01
-1.50193051e-01 -4.92050886e-01 -3.91194314e-01 1.01337290e+00
3.97366107e-01 -1.08241200e+00 6.79437459e-01 5.21510303e-01
2.40309536e-01 -8.02018344e-01 -1.02745152e+00 -1.43384361e+00
1.34123027e-01 -6.83456719e-01 8.45041513e-01 8.97770584e-01
1.41170129e-01 9.21015441e-02 -8.63630027e-02 2.98917919e-01
8.37972224e-01 5.26577473e-01 7.99268901e-01 -1.70890307e+00
-3.18791807e-01 -4.09119904e-01 -7.70816922e-01 -1.12163675e+00
7.14723915e-02 -7.66872346e-01 -1.50125563e-01 -9.02075768e-01
5.12088418e-01 -7.63226628e-01 -2.99878299e-01 4.59727168e-01
-3.91229503e-02 -1.34563539e-02 -2.53086388e-01 3.12362999e-01
-7.10674226e-01 1.39711857e-01 8.59018207e-01 1.21157415e-01
7.30021670e-02 3.10630798e-01 -5.68332613e-01 6.86509311e-01
7.35289872e-01 -7.72815406e-01 -4.68041271e-01 3.68155003e-01
1.58940852e-01 -7.09836930e-02 2.22417325e-01 -1.12638092e+00
2.80111164e-01 -4.21851039e-01 3.62877101e-01 -9.41288948e-01
9.81784314e-02 -7.29251027e-01 -3.58522356e-01 5.00640392e-01
-1.98403478e-01 -2.02619091e-01 8.60953853e-02 8.64694536e-01
-7.67702237e-02 -3.59962165e-01 1.04839730e+00 3.02666575e-01
-4.91220325e-01 2.22096384e-01 -2.16644898e-01 3.45946908e-01
1.28440249e+00 -4.27120566e-01 -4.14128393e-01 -1.46206856e-01
-3.48872304e-01 -1.06126152e-01 1.99515864e-01 -2.00452264e-02
3.95877272e-01 -1.36010075e+00 -4.22080278e-01 3.88094157e-01
3.31746936e-01 -8.77391249e-02 -1.02317795e-01 6.76043630e-01
-3.34275365e-01 5.03342927e-01 -9.01254267e-02 -9.57505882e-01
-1.30065095e+00 5.59495151e-01 9.68588367e-02 -4.45432186e-01
-2.57027559e-02 6.66840613e-01 1.33094788e-01 -4.37577099e-01
3.22357833e-01 -5.02916455e-01 3.87753546e-01 -9.29081626e-03
9.08747852e-01 9.12612855e-01 1.78165406e-01 -5.60500026e-02
-8.01042855e-01 4.79040951e-01 -6.15714006e-02 -1.71184257e-01
8.99422705e-01 1.06519602e-01 -8.46193954e-02 4.63948876e-01
1.36322296e+00 -9.42654014e-02 -8.67019176e-01 -2.58819968e-01
1.47688583e-01 -7.38285542e-01 -2.87399411e-01 -7.49672294e-01
-4.28531289e-01 9.62635815e-01 7.19132900e-01 4.64467824e-01
1.27420020e+00 -4.96818237e-02 6.00695908e-01 6.31876826e-01
6.49562180e-01 -1.27930808e+00 4.35537286e-02 5.65120459e-01
5.26319563e-01 -1.20671856e+00 3.64499390e-01 -8.40107620e-01
-9.58736315e-02 1.14047086e+00 4.92418587e-01 6.02929071e-02
1.09309638e+00 4.53389019e-01 -5.01452267e-01 2.31023327e-01
-6.46065056e-01 -1.28439799e-01 2.66036570e-01 8.28197956e-01
3.24296117e-01 4.31376517e-01 -4.28266555e-01 8.53586733e-01
-4.22698408e-01 -6.35575457e-03 2.89056063e-01 8.96447480e-01
-7.54140317e-01 -1.08788645e+00 -4.57376242e-01 4.38876390e-01
-3.22009295e-01 1.47350490e-01 -2.74418503e-01 7.86443233e-01
3.29199024e-02 1.05117738e+00 4.11225021e-01 -3.30461353e-01
1.43902153e-01 2.00851515e-01 7.60345876e-01 -3.28633785e-01
1.14631042e-01 -3.03476185e-01 7.79806450e-02 -4.80251968e-01
-3.74434263e-01 -6.49745643e-01 -1.03760862e+00 -6.28646553e-01
-5.86697340e-01 1.09544270e-01 6.25114739e-01 9.33043301e-01
4.90645736e-01 6.86298907e-02 7.89859772e-01 -2.55351633e-01
-9.46869850e-01 -8.80462229e-01 -8.20194602e-01 1.46681935e-01
1.64338887e-01 -1.04178500e+00 -5.80179572e-01 -2.29888409e-01] | [8.22064208984375, 4.164764881134033] |
bba063a4-7a8f-41b3-849d-5268555a70d4 | a-convex-and-feature-rich-discriminative | null | null | https://aclanthology.org/P15-1133 | https://aclanthology.org/P15-1133.pdf | A convex and feature-rich discriminative approach to dependency grammar induction | null | ["No{\\'e}mie Elhadad", "{\\'E}douard Grave"] | 2015-07-01 | a-convex-and-feature-rich-discriminative-1 | https://aclanthology.org/P15-1133 | https://aclanthology.org/P15-1133.pdf | ijcnlp-2015-7 | ['dependency-grammar-induction'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.275077819824219, 3.8479018211364746] |
f7b08aec-efdc-4efe-8caf-41c3e53e746f | sas-video-qa-self-adaptive-sampling-for | 2307.04192 | null | https://arxiv.org/abs/2307.04192v1 | https://arxiv.org/pdf/2307.04192v1.pdf | SAS Video-QA: Self-Adaptive Sampling for Efficient Video Question-Answering | Video question--answering is a fundamental task in the field of video understanding. Although current vision--language models (VLMs) equipped with Video Transformers have enabled temporal modeling and yielded superior results, they are at the cost of huge computational power and thus too expensive to deploy in real-time application scenarios. An economical workaround only samples a small portion of frames to represent the main content of that video and tune an image--text model on these sampled frames. Recent video understanding models usually randomly sample a set of frames or clips, regardless of internal correlations between their visual contents, nor their relevance to the problem. We argue that such kinds of aimless sampling may omit the key frames from which the correct answer can be deduced, and the situation gets worse when the sampling sparsity increases, which always happens as the video lengths increase. To mitigate this issue, we propose two frame sampling strategies, namely the most domain frames (MDF) and most implied frames (MIF), to maximally preserve those frames that are most likely vital to the given questions. MDF passively minimizes the risk of key frame omission in a bootstrap manner, while MIS actively searches key frames customized for each video--question pair with the assistance of auxiliary models. The experimental results on three public datasets from three advanced VLMs (CLIP, GIT and All-in-one) demonstrate that our proposed strategies can boost the performance for image--text pretrained models. The source codes pertaining to the method proposed in this paper are publicly available at https://github.com/declare-lab/sas-vqa. | ['Soujanya Poria', 'Min-Yen Kan', 'Hui Chen', 'Wei Han'] | 2023-07-09 | null | null | null | null | ['visual-question-answering', 'video-question-answering', 'video-understanding', 'question-answering'] | ['computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing'] | [ 2.68434018e-01 -4.26008590e-02 -3.65560472e-01 -3.26964080e-01
-7.73405552e-01 -3.97838712e-01 3.55406791e-01 -1.56033188e-01
-4.30719525e-01 3.92046720e-01 1.97168753e-01 -3.27459365e-01
1.76431984e-01 -6.00500584e-01 -1.01067960e+00 -7.04118609e-01
1.90210477e-01 1.88412234e-01 4.45624799e-01 -6.94115739e-03
3.20206612e-01 -1.35546193e-01 -1.75696135e+00 6.20382488e-01
8.14549088e-01 9.81776893e-01 7.95086265e-01 7.32499957e-01
-4.20694470e-01 1.22982728e+00 -3.83207113e-01 -5.80576897e-01
2.44276524e-01 -4.94821548e-01 -7.39391744e-01 3.41436058e-01
6.81651652e-01 -7.35408485e-01 -7.06821024e-01 1.09833229e+00
2.08888978e-01 2.23312795e-01 1.35616720e-01 -1.42266643e+00
-4.31160957e-01 4.70054060e-01 -4.69871998e-01 5.28714120e-01
5.15855730e-01 3.79441440e-01 9.19953883e-01 -9.81708348e-01
7.79713154e-01 1.18210196e+00 3.08824837e-01 5.63726485e-01
-6.70730889e-01 -6.29514277e-01 4.23689425e-01 7.98552513e-01
-1.25779641e+00 -7.95424759e-01 8.04561317e-01 -1.95006043e-01
6.62748158e-01 3.75209749e-01 7.45279968e-01 1.24830222e+00
-2.88267788e-02 1.08751333e+00 5.25616109e-01 -2.27233976e-01
8.49274695e-02 1.22368023e-01 2.95362175e-02 7.57447064e-01
-1.27849251e-01 -2.68477589e-01 -9.22468901e-01 3.28364335e-02
6.72876835e-01 1.35166012e-02 -6.39727354e-01 -1.56278446e-01
-1.29439783e+00 8.15948486e-01 -5.35526462e-02 2.58080512e-01
-4.53883529e-01 1.34339735e-01 3.06522220e-01 4.35783982e-01
3.60290647e-01 -1.24533892e-01 -5.20279408e-01 -3.02475691e-01
-1.03503227e+00 2.12719500e-01 6.50989115e-01 1.10763490e+00
7.92720139e-01 -9.63160545e-02 -1.45540759e-01 6.27273560e-01
1.08866170e-01 3.27441543e-01 2.19114333e-01 -1.37669265e+00
7.77439117e-01 4.05387551e-01 2.95124054e-02 -1.22428608e+00
8.10288489e-02 -1.10774115e-01 -5.75942874e-01 -4.97385561e-01
5.38304806e-01 3.50520127e-02 -6.32956505e-01 1.59422815e+00
4.26766485e-01 5.72709322e-01 -1.46684498e-01 1.03633308e+00
8.99757624e-01 9.70647693e-01 8.74713180e-04 -5.02976954e-01
1.50492704e+00 -1.05941069e+00 -7.54082918e-01 -4.42676842e-01
5.18710673e-01 -8.39674056e-01 1.16803765e+00 2.63964802e-01
-1.15591586e+00 -6.72775626e-01 -6.17644429e-01 -2.28348061e-01
1.06827766e-01 8.06945860e-02 4.07001227e-01 4.12273765e-01
-1.05022824e+00 1.43290028e-01 -6.23635054e-01 -2.64216691e-01
4.53883916e-01 -3.59134428e-05 -2.68015087e-01 -4.00953472e-01
-1.08587754e+00 4.74582642e-01 3.12828481e-01 1.53420821e-01
-1.07878006e+00 -7.08610475e-01 -6.26960397e-01 1.01679079e-01
8.88209999e-01 -6.35146797e-01 1.19678271e+00 -1.48119557e+00
-1.15991879e+00 7.62949944e-01 -6.61844730e-01 -6.24539614e-01
6.17725372e-01 -3.88092965e-01 -2.37951219e-01 7.55323887e-01
9.06219855e-02 1.00627708e+00 1.34210825e+00 -1.07056081e+00
-8.66082489e-01 -8.63064826e-02 4.92802531e-01 2.43780404e-01
-3.85706425e-01 8.38800147e-02 -1.15762830e+00 -5.71217179e-01
8.40901583e-02 -8.43079567e-01 -2.41496041e-02 1.36406824e-01
-1.91886872e-01 -1.31834626e-01 9.63859975e-01 -8.50128353e-01
1.31558561e+00 -2.06321597e+00 1.56744644e-01 -2.77976543e-01
8.48970264e-02 3.23353171e-01 -2.56244183e-01 3.20579380e-01
6.32509813e-02 -7.99559951e-02 -2.34097857e-02 -1.72050864e-01
-5.06617844e-01 3.00542951e-01 -4.69994903e-01 3.43541622e-01
6.80491999e-02 7.21207201e-01 -9.01596725e-01 -7.70874619e-01
3.71471018e-01 3.68735224e-01 -7.87391543e-01 3.55657518e-01
-6.80184126e-01 4.43440348e-01 -5.82042396e-01 6.06540263e-01
6.14983082e-01 -5.51186025e-01 1.32592693e-01 -5.12928069e-01
7.43436068e-02 1.55036837e-01 -1.13284647e+00 1.72452509e+00
-2.13189766e-01 8.11132491e-01 4.50853705e-02 -1.24958062e+00
4.55507606e-01 4.22627866e-01 5.06708741e-01 -6.70732856e-01
4.68938202e-02 -1.42026842e-01 -1.94127768e-01 -1.06118226e+00
4.96807069e-01 4.21169639e-01 3.50129902e-01 2.11027831e-01
-5.67393936e-03 2.48795912e-01 4.80436206e-01 4.63346481e-01
8.57012033e-01 9.06356275e-02 1.63750067e-01 6.61929548e-02
8.27146590e-01 -1.53941084e-02 6.80516601e-01 6.81638181e-01
-3.59221280e-01 7.08194315e-01 6.57122672e-01 -5.08046985e-01
-9.69404280e-01 -6.52412713e-01 2.71382183e-01 1.15486169e+00
3.68552893e-01 -7.24584639e-01 -8.08932543e-01 -6.38004184e-01
-3.92416507e-01 7.58823454e-01 -3.17527115e-01 4.46010800e-03
-6.59904718e-01 -2.88231045e-01 2.25202829e-01 9.45924968e-02
4.77647752e-01 -1.02698338e+00 -7.94988334e-01 1.24697290e-01
-9.96293783e-01 -1.54022062e+00 -6.97934210e-01 -4.59242076e-01
-7.67226338e-01 -1.41072452e+00 -6.37450874e-01 -7.54113019e-01
6.98194623e-01 8.57807159e-01 1.12776554e+00 3.62430423e-01
6.71824217e-02 5.41663051e-01 -6.32878006e-01 -5.92034124e-02
-3.36204588e-01 -2.13178560e-01 -3.14552248e-01 3.75546128e-01
4.27719295e-01 -3.17501336e-01 -8.18925142e-01 4.60383505e-01
-1.07947469e+00 4.68813270e-01 2.78066248e-01 7.00453937e-01
6.25682831e-01 -4.50767018e-03 3.99753869e-01 -7.63167322e-01
1.22827679e-01 -5.64862370e-01 -4.84905720e-01 3.43618959e-01
-1.35352761e-01 -2.07646102e-01 5.77521741e-01 -6.43569112e-01
-9.44180906e-01 -6.17338829e-02 -1.00841053e-01 -9.39999044e-01
-9.15891752e-02 2.95210660e-01 -8.81572887e-02 1.95671052e-01
1.89411163e-01 6.41886830e-01 3.00805178e-03 -2.07569703e-01
2.93155283e-01 3.85000259e-01 4.23811555e-01 -5.13075590e-01
5.96095324e-01 6.01329505e-01 -4.31002766e-01 -1.12286353e+00
-8.11390877e-01 -5.46705663e-01 -1.94378093e-01 -6.17844045e-01
8.97881269e-01 -1.17647135e+00 -5.70002079e-01 3.30900460e-01
-1.21036255e+00 -1.28537208e-01 7.84475543e-03 3.81892711e-01
-6.19189560e-01 7.71033287e-01 -4.28058922e-01 -7.70360172e-01
-1.93679482e-01 -1.36310017e+00 9.27492857e-01 3.53795439e-01
5.67882583e-02 -5.60560048e-01 -5.68052888e-01 8.60258162e-01
2.08604172e-01 -2.69074410e-01 8.14685404e-01 -3.35692763e-01
-1.14903176e+00 9.76719558e-02 -1.95312798e-01 2.69881040e-01
-1.41851798e-01 4.12714966e-02 -8.03025305e-01 -2.41294652e-01
2.05553025e-01 -2.74351239e-01 7.50186503e-01 5.84830105e-01
1.41732907e+00 -4.10311192e-01 -9.77441445e-02 5.91179967e-01
1.29760790e+00 3.37437719e-01 7.91199088e-01 3.08770478e-01
6.51724279e-01 5.62050402e-01 9.09343064e-01 5.01935542e-01
6.40549660e-01 5.85293233e-01 5.80438733e-01 2.00749174e-01
-9.18908417e-02 -3.56292963e-01 6.08238876e-01 8.83298516e-01
-3.22219506e-02 -4.40856934e-01 -5.36791027e-01 6.96934521e-01
-1.97606301e+00 -1.15502632e+00 -5.70660755e-02 2.01917958e+00
7.78504252e-01 2.66289823e-02 4.67443578e-02 -3.02929208e-02
7.36345828e-01 4.68660504e-01 -5.17670989e-01 1.50692359e-01
-1.76678568e-01 -2.49694586e-01 2.23696426e-01 3.95961374e-01
-7.92957306e-01 9.72066879e-01 5.06133986e+00 9.23332810e-01
-8.80503476e-01 2.43571043e-01 8.21496248e-01 -3.54210943e-01
-3.08497101e-01 3.14457327e-01 -7.96103060e-01 7.41574585e-01
9.77783203e-01 -6.43047765e-02 3.90873075e-01 6.44010842e-01
5.55261075e-01 -4.27434623e-01 -9.82028782e-01 1.26257062e+00
2.02098981e-01 -1.58445883e+00 3.56172085e-01 -2.91560858e-01
4.02992576e-01 -1.71233311e-01 -9.18680876e-02 1.29658267e-01
-4.82887506e-01 -5.50315320e-01 1.01070452e+00 5.24886012e-01
5.93362868e-01 -4.90592629e-01 4.35788959e-01 5.35945177e-01
-1.17826569e+00 -8.76582041e-02 -4.68256772e-01 1.55969381e-01
3.18114609e-01 5.20518184e-01 -5.19695520e-01 4.28185225e-01
1.00602305e+00 7.03254044e-01 -5.65666735e-01 7.37154186e-01
-8.24254602e-02 7.77715385e-01 -2.13048264e-01 2.03412160e-01
1.34060279e-01 -2.92782813e-01 7.43365705e-01 9.00100172e-01
3.68415862e-01 3.04418027e-01 1.22870635e-02 5.80473900e-01
7.63927549e-02 1.09592583e-02 -4.24332649e-01 -8.30710754e-02
5.55102646e-01 9.40147936e-01 -7.29920566e-01 -4.53602880e-01
-9.19662356e-01 8.40596557e-01 -9.04434081e-03 5.42123735e-01
-1.09209585e+00 1.69937983e-01 6.51473165e-01 2.51155168e-01
4.79592532e-01 -2.19050080e-01 2.89493412e-01 -1.49327052e+00
3.23981106e-01 -1.35281968e+00 5.47510982e-01 -9.16475058e-01
-8.86641622e-01 4.29083884e-01 8.98115560e-02 -1.37324965e+00
-2.83392727e-01 -2.54410923e-01 -3.58084410e-01 2.86553681e-01
-1.54306698e+00 -8.33350062e-01 -3.68570119e-01 9.41690922e-01
1.28310311e+00 2.18756497e-02 2.71258146e-01 4.61623907e-01
-5.40479600e-01 4.86342877e-01 -1.83310732e-01 7.92526901e-02
6.84508920e-01 -5.22592068e-01 1.69193760e-01 1.13679254e+00
3.57161045e-01 3.69355500e-01 8.83681476e-01 -4.19832766e-01
-1.74878454e+00 -8.42171013e-01 8.86582255e-01 -2.00702503e-01
4.62476999e-01 -1.71623424e-01 -1.00343263e+00 6.05113328e-01
1.76106229e-01 -3.41109708e-02 3.71954560e-01 -4.40197766e-01
-1.96572587e-01 -1.48645222e-01 -6.95005357e-01 6.65061533e-01
1.07814789e+00 -5.99800825e-01 -4.82104152e-01 4.48143393e-01
9.50773478e-01 -3.72552186e-01 -5.17311275e-01 1.79918811e-01
4.27441329e-01 -1.40440810e+00 1.14986014e+00 -3.75986159e-01
5.51491320e-01 -4.72788244e-01 -3.56669545e-01 -6.24204218e-01
8.58413875e-02 -6.55402899e-01 -4.60781366e-01 1.23545718e+00
1.50980540e-02 -2.50391960e-01 8.43281925e-01 4.72746074e-01
1.46359727e-01 -5.54310203e-01 -9.36669111e-01 -4.07813877e-01
-5.79995811e-01 -7.99431384e-01 2.86321610e-01 8.18482101e-01
-4.37062413e-01 1.50433138e-01 -7.22030103e-01 3.04341882e-01
5.94623744e-01 8.82785395e-02 9.14245605e-01 -7.11312056e-01
-2.92673647e-01 -2.18114287e-01 -1.73658147e-01 -1.43386447e+00
1.55692592e-01 -3.49701524e-01 -1.43561214e-01 -1.24924612e+00
1.76615164e-01 -9.14933532e-03 6.74743429e-02 1.93700820e-01
-3.25153679e-01 -8.93933550e-02 4.24227148e-01 4.03899729e-01
-7.71356225e-01 4.17934269e-01 1.22893631e+00 -1.64433107e-01
-4.99153249e-02 2.95863929e-03 -5.19279420e-01 7.15993643e-01
4.67048496e-01 -3.65954787e-01 -7.90254354e-01 -6.72056019e-01
2.49495387e-01 5.51487207e-01 4.87637371e-01 -8.04942608e-01
3.29116225e-01 -2.92307347e-01 1.78015873e-01 -7.38878608e-01
4.26071107e-01 -9.04959261e-01 3.32353473e-01 3.23860914e-01
-3.48384082e-01 1.53939396e-01 4.02230583e-02 7.35763371e-01
-3.65207434e-01 -3.05506915e-01 6.39470756e-01 -2.24935934e-01
-1.07237649e+00 5.37952542e-01 -2.12686524e-01 1.20159522e-01
1.06674576e+00 -3.32731575e-01 -3.47068042e-01 -8.34590793e-01
-4.44084644e-01 3.75494212e-01 4.29432392e-01 7.14879572e-01
8.90692592e-01 -9.38290298e-01 -7.09338129e-01 1.46905348e-01
7.35757127e-02 2.74351444e-02 6.95644975e-01 9.69607890e-01
-4.88565654e-01 3.79643440e-01 9.45429411e-03 -7.77726352e-01
-1.37234914e+00 6.74863279e-01 1.74529403e-01 -9.02949553e-03
-9.24378574e-01 9.27635968e-01 5.51242411e-01 3.37845236e-01
5.17469823e-01 -3.60309869e-01 -1.99883536e-01 1.61971986e-01
7.94874787e-01 1.69393018e-01 -1.23475596e-01 -7.92859852e-01
-1.92912325e-01 4.75858718e-01 -1.39049694e-01 1.23363331e-01
1.12075996e+00 -6.16450250e-01 2.02241186e-02 2.59570211e-01
1.11582470e+00 -3.45440984e-01 -1.50470078e+00 -4.80359316e-01
-2.08308712e-01 -7.80362368e-01 9.51761827e-02 -1.70258135e-01
-1.33070719e+00 7.71243215e-01 3.84614944e-01 1.53727531e-01
1.47169316e+00 8.03841278e-03 1.12130022e+00 2.12618709e-01
3.15261662e-01 -9.78058815e-01 1.67197406e-01 3.01598817e-01
6.53869390e-01 -1.16260445e+00 -1.83860302e-01 -4.78069425e-01
-8.05980444e-01 1.10483301e+00 6.77334309e-01 1.00188941e-01
3.87700260e-01 -8.05398747e-02 -6.95892423e-03 -3.26279551e-02
-1.18349481e+00 -1.57286286e-01 1.36681736e-01 4.39136893e-01
8.98959264e-02 -3.34144682e-01 -1.51467338e-01 3.62303972e-01
6.78244308e-02 1.42604873e-01 5.99902451e-01 7.99475908e-01
-5.22021592e-01 -7.81052828e-01 -4.07520652e-01 3.72176647e-01
-5.05377948e-01 -1.75982863e-01 1.50450706e-01 6.60499513e-01
6.90912083e-02 1.17695928e+00 9.07973498e-02 -2.33666912e-01
9.52823386e-02 1.26877218e-01 4.16920096e-01 -1.86861575e-01
-2.36032829e-01 2.06076548e-01 3.38551998e-02 -8.58237624e-01
-7.95250237e-01 -6.50478303e-01 -1.07934272e+00 -4.03703183e-01
-2.67620534e-01 8.47286079e-03 2.88167655e-01 1.10839963e+00
3.50827694e-01 2.89392918e-01 5.15913188e-01 -7.36918986e-01
-2.34082282e-01 -6.52253985e-01 -1.29356354e-01 4.19933915e-01
4.76505816e-01 -3.83233905e-01 -4.04753327e-01 6.19935811e-01] | [10.098535537719727, 0.7492615580558777] |
02d0c090-6205-4747-8329-39e346790d6a | deep-autoregressive-models-for-the-efficient | 1902.04057 | null | https://arxiv.org/abs/1902.04057v3 | https://arxiv.org/pdf/1902.04057v3.pdf | Deep autoregressive models for the efficient variational simulation of many-body quantum systems | Artificial Neural Networks were recently shown to be an efficient representation of highly-entangled many-body quantum states. In practical applications, neural-network states inherit numerical schemes used in Variational Monte Carlo, most notably the use of Markov-Chain Monte-Carlo (MCMC) sampling to estimate quantum expectations. The local stochastic sampling in MCMC caps the potential advantages of neural networks in two ways: (i) Its intrinsic computational cost sets stringent practical limits on the width and depth of the networks, and therefore limits their expressive capacity; (ii) Its difficulty in generating precise and uncorrelated samples can result in estimations of observables that are very far from their true value. Inspired by the state-of-the-art generative models used in machine learning, we propose a specialized Neural Network architecture that supports efficient and exact sampling, completely circumventing the need for Markov Chain sampling. We demonstrate our approach for two-dimensional interacting spin models, showcasing the ability to obtain accurate results on larger system sizes than those currently accessible to neural-network quantum states. | ['Yoav Levine', 'Or Sharir', 'Noam Wies', 'Amnon Shashua', 'Giuseppe Carleo'] | 2019-02-11 | null | null | null | null | ['variational-monte-carlo'] | ['miscellaneous'] | [ 1.69743359e-01 1.64092090e-02 -5.86856417e-02 -1.89323246e-01
-7.11735427e-01 -2.02409506e-01 7.70699322e-01 -3.23595285e-01
-7.47620046e-01 1.28501225e+00 4.66376124e-03 -4.28618670e-01
-4.00115103e-02 -1.15113640e+00 -6.14536226e-01 -1.01818621e+00
-7.86706656e-02 1.14968348e+00 9.79849845e-02 -7.85113573e-02
2.33948693e-01 6.32903576e-01 -1.51335633e+00 -4.59780358e-02
7.13689387e-01 7.21188962e-01 -4.25843596e-02 6.36155128e-01
1.84855182e-02 8.49848390e-01 -3.55688453e-01 -1.78239688e-01
-6.95328638e-02 -1.02446198e+00 -6.44798398e-01 -5.49506426e-01
5.85659593e-02 -4.42725480e-01 -7.60877430e-01 1.32003045e+00
3.47336024e-01 2.87815332e-01 9.50512946e-01 -7.12986171e-01
-5.79824388e-01 8.04242015e-01 2.63038844e-01 2.58664154e-02
-4.73309271e-02 4.68617886e-01 9.73862112e-01 -5.04127920e-01
6.74506724e-01 1.03356349e+00 7.34741926e-01 9.18184638e-01
-1.77190959e+00 -4.58117008e-01 -6.90148950e-01 2.04167336e-01
-1.48872817e+00 -5.64254165e-01 6.91373467e-01 -1.87543541e-01
1.34569240e+00 7.42264092e-02 1.07303488e+00 1.37450278e+00
3.07494462e-01 4.78240907e-01 1.48565340e+00 -7.34035134e-01
8.51664007e-01 -9.95425582e-02 2.54362524e-02 6.97016120e-01
3.36045712e-01 6.96624875e-01 -4.35285628e-01 -3.70853812e-01
9.91739929e-01 -1.59545362e-01 -5.48250638e-02 -4.14269745e-01
-1.18064439e+00 1.13056314e+00 2.29431927e-01 3.74143273e-01
-3.78901362e-01 8.02209377e-01 4.07176375e-01 -6.62201596e-03
-9.70374327e-03 4.78968769e-01 -4.34007682e-02 -3.83537292e-01
-1.32385170e+00 4.43262964e-01 1.09033811e+00 4.93745923e-01
8.86552036e-01 2.92522430e-01 2.32361816e-02 2.15521395e-01
2.71647334e-01 9.01790142e-01 3.50348115e-01 -1.07424116e+00
-3.71936858e-02 -1.19207375e-01 2.53969312e-01 -4.47405547e-01
-2.51105696e-01 -2.72923738e-01 -1.26477075e+00 1.96060434e-01
5.15244424e-01 -7.20476434e-02 -8.15952241e-01 1.79413450e+00
7.10061472e-03 1.11044273e-01 -2.77900305e-02 5.93750954e-01
2.94563681e-01 5.95704138e-01 -6.82973713e-02 -3.14805955e-01
1.01510787e+00 -5.40743768e-01 -5.95103860e-01 -6.49348870e-02
5.87363541e-01 -2.70596534e-01 5.82994699e-01 2.95784712e-01
-1.20364881e+00 -3.62071514e-01 -1.09840894e+00 1.05971046e-01
-3.71716827e-01 -3.17109168e-01 1.24924445e+00 1.05679750e+00
-1.08403885e+00 1.27014244e+00 -1.28006947e+00 -4.56482545e-03
3.49393785e-01 4.91176307e-01 -1.78554520e-01 1.59589842e-01
-1.34674299e+00 1.13817632e+00 8.39506269e-01 3.27126712e-01
-1.11430442e+00 5.52196056e-03 -6.95046425e-01 1.66607440e-01
9.57983658e-02 -8.28625441e-01 1.11363161e+00 -4.08402950e-01
-1.93513989e+00 4.45235491e-01 -1.59468219e-01 -6.93252206e-01
1.96943909e-01 3.16876709e-01 -2.68221349e-01 2.13601410e-01
-3.74262214e-01 5.83989084e-01 4.89751965e-01 -6.92819595e-01
4.13094461e-01 -1.77746922e-01 -1.98494956e-01 -2.59605348e-01
7.92097151e-02 -4.44732785e-01 4.79757935e-02 9.65584517e-02
3.34654599e-01 -1.19912982e+00 -5.34407377e-01 -5.49854577e-01
-4.03778136e-01 -1.33214876e-01 2.28570193e-01 -1.04005635e-01
1.05543697e+00 -1.45210552e+00 2.68783629e-01 3.36385399e-01
4.00397703e-02 4.30015743e-01 1.51881585e-02 8.40786755e-01
1.02757469e-01 -1.39868483e-01 -2.13963330e-01 -3.06257784e-01
4.02609229e-01 7.16342747e-01 -2.31968090e-01 5.18684626e-01
1.78083658e-01 1.28009009e+00 -7.36496210e-01 -5.08320570e-01
2.68638730e-01 4.79342341e-01 -8.30939353e-01 -2.17871413e-01
-3.50726992e-01 5.42570770e-01 -4.48991597e-01 4.97335792e-02
4.99924034e-01 -5.63019097e-01 4.29115117e-01 -8.61728340e-02
-4.12572455e-03 7.37004578e-01 -1.25771391e+00 1.69998229e+00
-3.97131175e-01 4.60303605e-01 -3.00141722e-01 -1.06027293e+00
5.94245970e-01 3.49346042e-01 -3.41105945e-02 -6.23856068e-01
2.91784853e-01 4.88752246e-01 1.60197109e-01 -1.73337013e-01
5.95851898e-01 -7.26412356e-01 -2.23573446e-01 7.52573907e-01
5.28345525e-01 -5.58173239e-01 3.86748940e-01 2.41398454e-01
8.62391829e-01 3.75880152e-01 4.11333770e-01 -4.13378924e-01
2.67105222e-01 -3.22103679e-01 2.49584377e-01 1.36837637e+00
-1.48070127e-01 4.44442302e-01 7.03129768e-01 -5.71725428e-01
-1.52819824e+00 -1.05475748e+00 -4.51878905e-01 4.83292431e-01
-2.12895185e-01 -2.91084588e-01 -9.65436339e-01 -1.11198373e-01
-2.59810150e-01 7.85773098e-01 -6.67977989e-01 -2.16616213e-01
-4.58503067e-01 -1.01479793e+00 6.32475555e-01 3.24637830e-01
3.07530433e-01 -1.34740031e+00 -5.66112876e-01 4.16045845e-01
8.21606889e-02 -9.50959146e-01 1.39032081e-01 6.01761281e-01
-1.06297636e+00 -7.00230598e-01 -4.97932553e-01 -6.17360026e-02
5.17979324e-01 -5.38374543e-01 1.23489189e+00 -3.07211846e-01
-2.41300508e-01 -1.22697696e-01 2.61757612e-01 1.79642394e-01
-1.00922120e+00 2.97980815e-01 3.31700921e-01 -5.03434181e-01
5.76476276e-01 -9.88865137e-01 -5.12631595e-01 -1.83137968e-01
-9.03791904e-01 4.16618623e-02 6.83763564e-01 1.16526401e+00
3.73464227e-01 -1.23875231e-01 2.64736027e-01 -8.39956284e-01
5.11603653e-01 -1.17574014e-01 -8.48909914e-01 1.23523381e-02
-6.17774248e-01 8.31075072e-01 6.22590363e-01 -2.47834638e-01
-8.61133158e-01 -2.05190122e-01 -4.27271128e-01 2.13188510e-02
-2.39056990e-01 4.87390012e-01 3.24788034e-01 -2.31192946e-01
8.07130218e-01 4.62106109e-01 -4.57335785e-02 -1.99371114e-01
3.28135222e-01 4.46405023e-01 3.51535589e-01 -9.00823951e-01
5.87868929e-01 4.88943130e-01 6.73357189e-01 -6.63875937e-01
-8.11479330e-01 9.48345512e-02 -7.45050669e-01 6.71351328e-02
9.53442276e-01 -6.89186990e-01 -1.30848348e+00 4.98982757e-01
-1.20585179e+00 -1.83704272e-01 -4.54790771e-01 7.40958333e-01
-8.76854062e-01 4.14534986e-01 -1.07827532e+00 -1.24851143e+00
-1.82183981e-01 -1.26244330e+00 7.61406898e-01 3.33931923e-01
-2.09586084e-01 -1.06248355e+00 4.72371876e-01 -3.65310418e-03
6.61486566e-01 1.16654180e-01 9.36080992e-01 -3.16760153e-01
-8.86340380e-01 -5.96755922e-01 4.10233103e-02 4.29214001e-01
-6.21971905e-01 -1.83109060e-01 -1.02236962e+00 -2.32588336e-01
-3.77170928e-02 -6.87099338e-01 1.04451776e+00 4.07151729e-01
8.42693031e-01 -2.11373996e-03 -2.63117462e-01 3.84287000e-01
1.55854273e+00 -1.82460278e-01 9.06418622e-01 -2.36498117e-01
4.41413432e-01 -4.52133976e-02 -4.02125269e-01 3.64646226e-01
-1.84243359e-02 5.53138018e-01 3.41046780e-01 4.59318995e-01
1.85738221e-01 -5.09623349e-01 1.09381877e-01 1.15131283e+00
-5.46989024e-01 1.36457756e-01 -7.72022963e-01 3.69451284e-01
-1.78282011e+00 -1.50158548e+00 -1.74763471e-01 2.41368842e+00
1.06555831e+00 4.21174914e-01 -1.50022551e-01 -8.90993625e-02
3.01267117e-01 2.40882799e-01 -5.05359650e-01 -5.79048932e-01
7.33387750e-03 8.73426914e-01 4.60401714e-01 4.91390944e-01
-7.74682164e-01 8.20365369e-01 7.35363817e+00 1.05610025e+00
-7.23215163e-01 3.62523735e-01 3.47713023e-01 -2.14976728e-01
-3.69304925e-01 3.25445443e-01 -9.83745873e-01 5.79623938e-01
1.60127282e+00 4.27481711e-01 9.09612060e-01 6.53801203e-01
-1.04962029e-01 -4.02104169e-01 -1.08309412e+00 8.35522413e-01
-2.69371271e-01 -1.70188618e+00 -4.10457104e-02 4.27479148e-01
9.69283283e-01 5.61130345e-01 -1.76627204e-01 4.88616377e-01
5.46659470e-01 -1.28297079e+00 6.69525862e-01 6.82097375e-01
7.57239163e-01 -9.79248047e-01 8.56242418e-01 6.38080478e-01
-3.88381332e-01 3.25932682e-01 -8.17084730e-01 -3.84099811e-01
3.89935255e-01 8.44544768e-01 -4.48030740e-01 9.05061793e-03
4.41556089e-02 -1.86749008e-02 -7.08802268e-02 5.74367285e-01
-2.63131678e-01 6.80775881e-01 -6.74428225e-01 -7.39288449e-01
5.46227455e-01 -6.93669498e-01 2.67023414e-01 9.95831370e-01
2.97980934e-01 -1.91137433e-01 -2.69388288e-01 1.57648182e+00
1.05482906e-01 -5.43704093e-01 -6.16207898e-01 -4.89648402e-01
4.78560507e-01 9.56225514e-01 -6.63920999e-01 -4.45024252e-01
-7.52299801e-02 8.33190441e-01 5.41139483e-01 3.47895116e-01
-7.01479733e-01 -4.32917655e-01 1.10582612e-01 -1.98047981e-01
6.64323151e-01 -7.41525054e-01 2.24198010e-02 -1.42895627e+00
-4.04768676e-01 -6.70174956e-01 -2.86197215e-01 -4.94597495e-01
-1.05226874e+00 5.56262136e-01 -3.73399667e-02 -5.64603031e-01
-7.83799171e-01 -8.68069589e-01 -5.04274786e-01 1.21747696e+00
-9.60765123e-01 -8.58369887e-01 2.87158549e-01 2.71321595e-01
-3.86062086e-01 -5.93862310e-02 1.53292394e+00 -6.78089708e-02
-4.10267115e-01 9.11737308e-02 6.54241502e-01 1.05521217e-01
-9.27839056e-02 -1.18916380e+00 6.37251794e-01 6.78234398e-01
4.55841124e-01 1.11516595e+00 9.29118097e-01 -2.81996369e-01
-1.55026495e+00 -2.48014227e-01 8.92848432e-01 -4.59122926e-01
8.10276687e-01 -6.48477554e-01 -6.77617431e-01 5.24450004e-01
4.48238365e-02 2.39877969e-01 6.24672532e-01 4.73578453e-01
-3.62974554e-01 3.31497133e-01 -1.02274334e+00 5.66019416e-01
7.37397134e-01 -1.06716776e+00 -3.43188524e-01 2.97352821e-01
7.06574693e-02 -2.88031876e-01 -6.96063876e-01 -6.92117214e-02
8.15631866e-01 -1.48669958e+00 8.45782816e-01 -6.90498829e-01
4.55711395e-01 -7.75123611e-02 -1.44992217e-01 -1.05454433e+00
-2.77272761e-01 -7.34706342e-01 -5.32454550e-01 5.15251637e-01
2.77153701e-01 -7.41412699e-01 1.01052451e+00 6.08093739e-01
2.66723812e-01 -6.12794995e-01 -1.29913306e+00 -7.34323561e-01
5.19738913e-01 -5.72058201e-01 5.28860271e-01 5.30126035e-01
5.88823296e-02 3.07959616e-01 -3.97034913e-01 -1.79203972e-01
9.93058562e-01 -4.30380576e-04 3.29862416e-01 -1.10898161e+00
-9.69349742e-01 -3.26672524e-01 -4.35373634e-01 -1.12885749e+00
2.77739346e-01 -9.10951972e-01 7.49623477e-02 -1.03995204e+00
5.26780963e-01 -2.60854870e-01 -2.80012071e-01 3.94840837e-02
2.69125372e-01 6.11999512e-01 -6.84146956e-02 2.48459458e-01
-6.42458379e-01 6.55561566e-01 1.10697401e+00 2.13402137e-01
2.74081886e-01 -2.96902448e-01 -1.23257816e-01 6.50002003e-01
4.55054551e-01 -7.50713706e-01 1.01831686e-02 6.00985475e-02
8.01555037e-01 3.88514489e-01 6.44118905e-01 -1.34188747e+00
2.79173046e-01 2.62004361e-02 4.27236050e-01 -6.17871046e-01
5.20924807e-01 -2.11613834e-01 3.67287040e-01 7.41553307e-01
-2.56439686e-01 -5.38960755e-01 -1.54276416e-01 5.06457746e-01
-8.60965773e-02 -9.04704392e-01 1.00119507e+00 -6.14960372e-01
-1.79596141e-01 1.88793883e-01 -6.28052175e-01 -1.48969695e-01
4.05921817e-01 8.48251656e-02 -1.24761842e-01 -4.17022735e-01
-7.55862534e-01 -5.73340178e-01 5.85988581e-01 -4.50086683e-01
1.81086421e-01 -1.32362747e+00 -3.29065621e-01 3.00021678e-01
-2.31896684e-01 -3.59679520e-01 4.16918844e-01 9.71147895e-01
-7.73216784e-01 8.30533206e-01 -1.19030237e-01 -4.49732929e-01
-4.57048386e-01 3.99969935e-01 4.29444313e-01 -6.16473317e-01
-4.28957164e-01 5.95741391e-01 -4.25826669e-01 -6.11934960e-01
-4.51567441e-01 -2.21884593e-01 5.82782626e-01 -3.22803766e-01
3.82580429e-01 2.23840043e-01 -6.24403805e-02 -4.11303014e-01
-1.25311479e-01 1.10805884e-01 -4.40617800e-02 -4.00326699e-01
1.25741017e+00 2.82553494e-01 -3.48208606e-01 4.91390586e-01
1.12927580e+00 -4.02358145e-01 -1.12365627e+00 -2.04467833e-01
-2.62732118e-01 1.15692042e-01 3.39193881e-01 -5.75980842e-01
-3.84883463e-01 1.28755605e+00 4.34610873e-01 5.09649217e-01
2.95570135e-01 -4.97405082e-02 9.35093880e-01 9.95681286e-01
8.89684379e-01 -1.19497585e+00 -3.27751398e-01 7.57930636e-01
1.74686268e-01 -1.09620285e+00 8.95448104e-02 4.39458847e-01
-9.49854702e-02 1.23982930e+00 -1.99254528e-01 -3.93402755e-01
3.70487809e-01 2.19441637e-01 -4.85458374e-01 -3.27443570e-01
-6.18449152e-01 -5.87111637e-02 8.14273283e-02 4.30725157e-01
4.01201665e-01 3.33070993e-01 -1.55214489e-01 6.33639842e-02
-2.01080829e-01 2.42915004e-01 6.67301893e-01 6.61407113e-01
-4.77463663e-01 -1.17545366e+00 -1.37796387e-01 2.93774068e-01
-2.79325068e-01 -4.49115872e-01 2.53534168e-01 7.03852654e-01
-4.27927300e-02 5.13121843e-01 6.34807199e-02 1.19600810e-01
-3.46339375e-01 4.84947056e-01 1.24316657e+00 -2.80885190e-01
-5.57183735e-02 -2.18699127e-01 2.30471611e-01 -5.75854480e-01
-4.08039480e-01 -7.58845925e-01 -1.08814561e+00 -6.24138713e-01
-5.62766492e-01 3.52902949e-01 9.05031323e-01 1.22645938e+00
2.31219888e-01 3.39093566e-01 1.89679280e-01 -1.25907922e+00
-1.25107634e+00 -1.15511012e+00 -9.42934155e-01 2.55437285e-01
2.05126211e-01 -4.98115569e-01 -4.76176262e-01 -4.33389962e-01] | [5.614297866821289, 4.909027099609375] |
a05462ae-955a-42be-8e60-835ab0b85678 | hi-fi-hierarchical-feature-integration-for | 1801.01849 | null | http://arxiv.org/abs/1801.01849v4 | http://arxiv.org/pdf/1801.01849v4.pdf | Hi-Fi: Hierarchical Feature Integration for Skeleton Detection | In natural images, the scales (thickness) of object skeletons may
dramatically vary among objects and object parts, making object skeleton
detection a challenging problem. We present a new convolutional neural network
(CNN) architecture by introducing a novel hierarchical feature integration
mechanism, named Hi-Fi, to address the skeleton detection problem. The proposed
CNN-based approach has a powerful multi-scale feature integration ability that
intrinsically captures high-level semantics from deeper layers as well as
low-level details from shallower layers. % By hierarchically integrating
different CNN feature levels with bidirectional guidance, our approach (1)
enables mutual refinement across features of different levels, and (2)
possesses the strong ability to capture both rich object context and
high-resolution details. Experimental results show that our method
significantly outperforms the state-of-the-art methods in terms of effectively
fusing features from very different scales, as evidenced by a considerable
performance improvement on several benchmarks. | ['Shang-Hua Gao', 'Ming-Ming Cheng', 'Wei Shen', 'Dandan Li', 'Kai Zhao'] | 2018-01-05 | null | null | null | null | ['object-skeleton-detection'] | ['computer-vision'] | [ 1.22782141e-01 -1.88469052e-01 -1.92041442e-01 -2.51712829e-01
-5.45843542e-01 -8.60934630e-02 4.27608818e-01 6.95679933e-02
-2.67381549e-01 2.93663442e-01 2.85062939e-01 3.86592418e-01
-1.60018995e-01 -1.02567172e+00 -5.97266078e-01 -5.67183495e-01
-1.88178360e-01 9.44269598e-02 1.14502823e+00 -1.60485595e-01
5.27236499e-02 8.07933152e-01 -1.74322653e+00 4.86764014e-01
5.05714118e-01 1.33757341e+00 1.24055780e-01 3.63774449e-01
-2.18058854e-01 5.57227671e-01 -1.51484966e-01 -3.48441660e-01
3.51795524e-01 -1.29745319e-01 -8.35267305e-01 2.85447538e-01
7.60479867e-01 -2.85037458e-01 -2.07724109e-01 1.03286827e+00
2.82401145e-01 -2.10258260e-01 4.54618901e-01 -7.77850389e-01
-5.05687118e-01 4.43607330e-01 -1.01724172e+00 1.14491582e-01
5.60749024e-02 1.95880115e-01 1.13962448e+00 -1.05363631e+00
4.66109633e-01 1.43176174e+00 7.18261302e-01 4.05236989e-01
-1.14137626e+00 -6.66045368e-01 3.70162070e-01 9.14235413e-02
-1.41787088e+00 -7.22781718e-02 9.33903635e-01 -4.22479749e-01
5.26289105e-01 -4.18016985e-02 8.72535050e-01 6.87614262e-01
8.51823390e-02 8.09305489e-01 9.22742307e-01 -8.27297643e-02
-1.10254653e-01 -4.50595945e-01 4.23110425e-02 1.11575890e+00
3.58224899e-01 -1.50737688e-01 -7.50613570e-01 -1.08161449e-01
1.32289684e+00 2.48105630e-01 -1.57204032e-01 -5.00450730e-01
-1.54343224e+00 7.41521776e-01 9.18735087e-01 3.45454276e-01
-4.68575090e-01 4.07265186e-01 4.11032528e-01 -8.95039514e-02
5.30562215e-02 1.28136873e-01 -6.79828465e-01 3.78380716e-01
-7.45516360e-01 2.08377466e-01 3.18576247e-01 8.33157182e-01
9.49372411e-01 6.24630600e-02 -1.67138651e-01 9.15007889e-01
2.46996984e-01 5.57758436e-02 5.25328398e-01 -9.95083749e-01
3.09991628e-01 1.06590259e+00 -1.84525102e-01 -1.01837778e+00
-5.51531374e-01 -5.78804076e-01 -1.09951472e+00 4.38591748e-01
3.74920875e-01 2.25686088e-01 -1.06611192e+00 1.67597914e+00
7.03385115e-01 -1.74042359e-02 -1.52337462e-01 9.44894731e-01
1.27669668e+00 3.13105822e-01 8.35871324e-02 2.18313783e-01
1.78366685e+00 -1.05204022e+00 -2.67336547e-01 -1.70185775e-01
9.77814123e-02 -7.43630528e-01 1.02105260e+00 9.05166566e-02
-1.34321272e+00 -9.22762454e-01 -1.12390494e+00 -3.33408386e-01
-2.26416692e-01 2.24375680e-01 8.47543597e-01 1.42151281e-01
-9.32128072e-01 6.75022125e-01 -9.79429603e-01 -3.30589682e-01
8.27562690e-01 5.30791521e-01 -6.28894746e-01 -6.10509031e-02
-7.91943491e-01 3.17094713e-01 4.95897084e-01 1.31176546e-01
-7.36924827e-01 -5.82537353e-01 -9.93102729e-01 1.74332097e-01
4.16867852e-01 -1.16398776e+00 1.03163397e+00 -7.33709693e-01
-1.34624290e+00 6.65824056e-01 1.57829542e-02 2.86457520e-02
6.15652978e-01 -3.91298413e-01 9.96883661e-02 4.38346714e-01
2.29171380e-01 9.19178665e-01 8.00676286e-01 -1.22689700e+00
-9.78690803e-01 -6.25593603e-01 1.91730216e-01 1.00960329e-01
-6.39481306e-01 1.74516469e-01 -7.74428368e-01 -1.00545335e+00
5.36724687e-01 -4.69164938e-01 -3.91470999e-01 7.02419221e-01
-3.70298684e-01 -3.19480032e-01 9.97019410e-01 -1.95087835e-01
9.78896439e-01 -2.03438163e+00 1.92023560e-01 -1.18407393e-02
4.44152027e-01 1.33953333e-01 -1.09333381e-01 5.97354844e-02
1.41932920e-01 3.85563113e-02 -3.97292882e-01 -2.26744503e-01
-3.16947252e-01 1.56812757e-01 3.50681454e-01 3.56730044e-01
6.79120600e-01 1.04751170e+00 -8.86048079e-01 -8.65425408e-01
2.63819039e-01 6.62204146e-01 -5.29027402e-01 1.58593774e-01
6.92396332e-03 3.16613674e-01 -6.87937379e-01 1.05312204e+00
4.75796551e-01 -6.40769482e-01 -1.02792636e-01 -5.71091294e-01
-8.08214769e-02 -1.34195089e-01 -1.34390092e+00 1.74140024e+00
-1.75736979e-01 3.49466652e-01 2.15074494e-01 -6.42984927e-01
7.59883285e-01 1.29138008e-01 5.95258236e-01 -4.23054069e-01
4.87115197e-02 1.12847239e-01 1.35397306e-02 -3.31158429e-01
2.65399247e-01 4.19547558e-02 8.87937918e-02 8.83374736e-02
2.22046524e-01 1.42927736e-01 1.22415483e-01 -3.01408824e-02
1.07698858e+00 1.15631700e-01 4.66114044e-01 -3.17075938e-01
6.92425847e-01 -4.46938455e-01 8.23823094e-01 5.41940629e-01
-2.22769007e-01 8.59361589e-01 3.20126444e-01 -6.28637671e-01
-8.00399601e-01 -1.09932339e+00 -2.67353088e-01 1.08313215e+00
4.10398781e-01 -4.99097347e-01 -4.74233419e-01 -7.74119139e-01
1.37866765e-01 -4.31982726e-01 -1.02757764e+00 7.13890567e-02
-8.96647274e-01 -5.60491741e-01 2.73429483e-01 1.11911917e+00
9.11261857e-01 -1.16649663e+00 -7.75423229e-01 2.55067825e-01
-1.06142417e-01 -1.34486198e+00 -4.08092827e-01 2.23318681e-01
-1.23082614e+00 -1.08625245e+00 -9.60361779e-01 -9.98963237e-01
5.41843176e-01 3.80866438e-01 1.19068825e+00 4.99270290e-01
-6.80358768e-01 -2.59649120e-02 -3.80757838e-01 -9.96349528e-02
1.25561222e-01 1.70649260e-01 -2.54563421e-01 1.90178886e-01
-1.60556003e-01 -7.47678578e-01 -9.04581130e-01 4.76138830e-01
-1.13586855e+00 1.42396078e-01 9.21040952e-01 7.50457168e-01
7.63021111e-01 1.78975776e-01 5.33747494e-01 -6.73603475e-01
2.65294723e-02 -1.04086757e-01 -3.54842842e-01 1.10833608e-01
-1.54752329e-01 -3.48895825e-02 4.72856879e-01 -3.70036870e-01
-9.65036035e-01 2.81827003e-01 -1.95452735e-01 -2.64113456e-01
-4.41867083e-01 1.92337036e-01 -3.85752499e-01 -3.65875155e-01
3.75237495e-01 9.66524705e-02 -1.66355863e-01 -7.89311767e-01
3.16378534e-01 1.90767393e-01 8.34658921e-01 -6.98509872e-01
9.77184296e-01 7.96894252e-01 2.47352645e-01 -8.64145160e-01
-9.82542336e-01 -6.02564037e-01 -1.19603848e+00 -1.82126954e-01
9.30252731e-01 -9.31010306e-01 -4.12246138e-01 7.68329322e-01
-1.11054754e+00 -9.19599831e-02 -3.11863750e-01 2.37067893e-01
-5.01871407e-01 4.85956669e-01 -1.02664781e+00 -2.36319095e-01
-3.57664704e-01 -1.20406210e+00 1.39427900e+00 5.36383986e-01
9.35174152e-02 -7.56934643e-01 -2.35492587e-01 3.50335509e-01
2.50549674e-01 7.27746010e-01 9.92490590e-01 -2.51137704e-01
-8.35956156e-01 -9.45478827e-02 -7.24040747e-01 1.35457247e-01
2.73049206e-01 2.27910310e-01 -7.68572807e-01 -2.16092661e-01
-3.31103951e-01 -4.65953708e-01 1.15688920e+00 3.38673949e-01
1.39113486e+00 4.36887145e-02 -2.98344344e-01 7.01009870e-01
1.41571081e+00 -5.00416875e-01 4.89063114e-01 4.21606928e-01
1.03670847e+00 6.54422224e-01 4.74368602e-01 3.70014042e-01
3.14652979e-01 7.52664864e-01 7.64023662e-01 -6.10614479e-01
-4.84049946e-01 1.40986398e-01 1.26945212e-01 6.41499281e-01
-4.31407839e-01 5.25047958e-01 -5.44593334e-01 5.42452991e-01
-1.94932103e+00 -6.94691122e-01 -3.56367677e-01 1.85253322e+00
9.02123451e-01 3.43388706e-01 3.99325550e-01 5.79910576e-02
6.58319116e-01 2.23838806e-01 -5.61565399e-01 1.66477025e-01
-1.78641140e-01 2.75627226e-01 2.01109871e-01 -7.96976462e-02
-1.31859243e+00 8.81590307e-01 6.48802996e+00 8.58716786e-01
-8.48876476e-01 -1.56665184e-02 2.69743115e-01 1.78886995e-01
2.09585965e-01 -3.16801220e-01 -7.86254406e-01 1.20630398e-01
5.24724089e-02 1.81544676e-01 -3.06302398e-01 9.50549662e-01
-2.04695150e-01 1.49607778e-01 -1.07935023e+00 7.12630093e-01
-7.87672848e-02 -1.33126295e+00 3.09559613e-01 9.91085768e-02
8.13943088e-01 5.53165413e-02 -1.93148047e-01 -1.97215658e-02
2.48805270e-01 -8.73603821e-01 9.23252583e-01 2.78907806e-01
6.53858662e-01 -8.65909815e-01 7.54554749e-01 2.29441464e-01
-1.95799112e+00 -2.45191976e-01 -4.22739387e-01 1.68264374e-01
5.01297880e-03 5.80734968e-01 -1.37468159e-01 7.34624803e-01
1.05255401e+00 9.99665141e-01 -8.43519628e-01 1.28177655e+00
-1.37454063e-01 3.35587263e-01 -3.07334542e-01 3.68874729e-01
2.56599188e-01 1.67627797e-01 2.53691494e-01 1.45799470e+00
9.30270404e-02 6.99710771e-02 3.37719619e-01 7.59365559e-01
-2.21342146e-01 1.51990622e-01 -2.45426476e-01 3.19644570e-01
9.55253243e-02 1.54852355e+00 -1.17116892e+00 -4.07758385e-01
-7.04206109e-01 8.21014643e-01 5.14104128e-01 1.26761064e-01
-6.99321926e-01 -3.40186894e-01 8.31056297e-01 1.66643560e-01
6.97770536e-01 -3.03538799e-01 -3.08179706e-01 -1.11019361e+00
2.20380783e-01 -4.85205591e-01 6.36230648e-01 -5.50140560e-01
-1.38589895e+00 5.93831241e-01 -2.22643167e-01 -1.30332482e+00
3.20916146e-01 -6.55283570e-01 -4.85742033e-01 4.82151061e-01
-1.78264153e+00 -1.60193777e+00 -5.76222956e-01 8.29334319e-01
8.12065959e-01 1.52427271e-01 5.37645280e-01 3.21513116e-01
-4.54443276e-01 4.79889691e-01 -3.88097227e-01 5.13690293e-01
3.84751976e-01 -1.24387681e+00 1.94099471e-01 6.45884395e-01
1.84187680e-01 6.55323625e-01 1.03837892e-01 -4.65417892e-01
-9.93412077e-01 -1.31639135e+00 3.98744017e-01 -2.10612148e-01
5.21401107e-01 -1.33336440e-01 -1.08823252e+00 3.36776942e-01
-1.87560573e-01 4.81427103e-01 5.79055190e-01 1.42120402e-02
-6.23814642e-01 -5.38355857e-02 -8.10706615e-01 3.95021886e-01
1.21862805e+00 -3.29830885e-01 -6.42559886e-01 1.22018695e-01
7.09566057e-01 -3.35078090e-01 -1.14499092e+00 1.02939606e+00
9.36509311e-01 -1.35513210e+00 1.28008270e+00 -5.62534988e-01
7.25204766e-01 -4.77115721e-01 -1.05913408e-01 -5.29151738e-01
-7.81185985e-01 -1.35446206e-01 -4.30845708e-01 1.14679635e+00
1.12159304e-01 -1.95504427e-01 8.33065629e-01 8.16887990e-02
-1.26755923e-01 -1.26811969e+00 -8.13455284e-01 -6.56478941e-01
-3.77055742e-02 -2.48780355e-01 5.57583451e-01 6.61447823e-01
-6.15753710e-01 2.36587673e-01 -8.85858312e-02 1.79064691e-01
6.94545031e-01 5.91470659e-01 8.08644712e-01 -1.58757281e+00
-1.54679939e-01 -8.21768105e-01 -8.94981980e-01 -1.18161333e+00
-1.83431342e-01 -6.75894618e-01 1.77879594e-02 -1.60870326e+00
6.07841790e-01 -9.54991207e-02 -5.20739734e-01 6.74497306e-01
-3.93443197e-01 9.01993752e-01 1.23684049e-01 3.05856466e-01
-7.05387592e-01 6.49335980e-01 1.60252118e+00 2.60059107e-02
6.56309202e-02 -2.96190977e-02 -6.93149328e-01 1.33693564e+00
3.90130460e-01 -3.35617155e-01 7.26294741e-02 -4.11048472e-01
-1.92939967e-01 -2.97058523e-01 6.76853001e-01 -1.31069362e+00
1.36960909e-01 -1.70352250e-01 9.08703744e-01 -6.58210814e-01
3.37080032e-01 -8.24400306e-01 -7.14219585e-02 3.78141373e-01
-2.54150599e-01 -2.03294381e-01 -2.52143554e-02 7.05331206e-01
-3.47980708e-01 1.74005345e-01 1.13709080e+00 -4.30137187e-01
-6.62158251e-01 6.82168543e-01 -2.56191846e-02 -1.13716438e-01
8.10002744e-01 -4.68069732e-01 -1.83177982e-02 1.56869441e-01
-8.13898623e-01 2.02832729e-01 7.14512587e-01 5.17346621e-01
7.85225749e-01 -1.47911215e+00 -5.64388573e-01 1.59873396e-01
3.73625576e-01 5.39874315e-01 7.43628889e-02 7.98860371e-01
-4.15105492e-01 2.12923288e-02 -3.80751818e-01 -9.29374397e-01
-1.26663649e+00 9.59775373e-02 2.66235948e-01 -3.15186024e-01
-1.03752053e+00 1.05966485e+00 6.36737883e-01 -2.17637211e-01
3.09644103e-01 -4.68559116e-01 -3.28051716e-01 -4.94404770e-02
5.69843411e-01 2.46458903e-01 5.62017113e-02 -9.53538120e-01
-3.95339400e-01 1.23665702e+00 -1.45347133e-01 3.43065739e-01
1.40166819e+00 -1.69454254e-02 -1.78151250e-01 2.22775072e-01
1.17825949e+00 -1.69815913e-01 -1.47600961e+00 -6.32338822e-01
-1.09063974e-02 -6.32695317e-01 5.01183346e-02 -4.28301960e-01
-1.53260612e+00 8.57000172e-01 4.44860131e-01 1.58288442e-02
1.13965392e+00 3.90728116e-01 1.08065963e+00 2.73041248e-01
4.21172172e-01 -9.21757579e-01 7.00821400e-01 3.64750236e-01
8.89451563e-01 -1.33198583e+00 2.72666126e-01 -6.68138564e-01
-2.57299334e-01 1.43411577e+00 1.01394391e+00 -2.18583822e-01
7.32153535e-01 3.07324827e-01 -6.16690191e-03 -6.43512547e-01
-4.51602399e-01 -6.57423556e-01 5.88431656e-01 4.97633100e-01
4.74225104e-01 -1.23712070e-01 -1.14288516e-01 7.02804148e-01
9.02803168e-02 -9.08535570e-02 -8.37078039e-03 1.04798675e+00
-6.65109038e-01 -9.61817801e-01 -3.91498804e-01 4.51358050e-01
-6.35359585e-01 1.83317557e-01 -4.60567623e-01 9.26537156e-01
4.30279911e-01 5.00638187e-01 -8.69517848e-02 -2.69921660e-01
3.70597750e-01 -3.67677033e-01 4.30297107e-01 -8.19076717e-01
-6.88186646e-01 3.07394564e-01 -3.35828632e-01 -1.01109445e+00
-7.57222176e-01 -5.87074876e-01 -1.43325543e+00 2.33058870e-01
-4.55324203e-01 -3.23312134e-01 3.64808470e-01 9.02706206e-01
2.17195436e-01 8.29998970e-01 4.62728351e-01 -1.31035864e+00
-2.44451284e-01 -8.15059245e-01 -5.60738325e-01 3.87410820e-01
4.70946610e-01 -1.12237430e+00 -1.66357949e-01 1.67082831e-01] | [9.558755874633789, -0.5878788828849792] |
643002ca-711e-43a1-b8f4-aca90a83d4c2 | mixture-encoder-for-joint-speech-separation | 2306.12173 | null | https://arxiv.org/abs/2306.12173v1 | https://arxiv.org/pdf/2306.12173v1.pdf | Mixture Encoder for Joint Speech Separation and Recognition | Multi-speaker automatic speech recognition (ASR) is crucial for many real-world applications, but it requires dedicated modeling techniques. Existing approaches can be divided into modular and end-to-end methods. Modular approaches separate speakers and recognize each of them with a single-speaker ASR system. End-to-end models process overlapped speech directly in a single, powerful neural network. This work proposes a middle-ground approach that leverages explicit speech separation similarly to the modular approach but also incorporates mixture speech information directly into the ASR module in order to mitigate the propagation of errors made by the speech separator. We also explore a way to exchange cross-speaker context information through a layer that combines information of the individual speakers. Our system is optimized through separate and joint training stages and achieves a relative improvement of 7% in word error rate over a purely modular setup on the SMS-WSJ task. | ['Reinhold Haeb-Umbach', 'Ralf Schlüter', 'Christoph Boeddeker', 'Peter Vieting', 'Simon Berger'] | 2023-06-21 | null | null | null | null | ['speech-separation', 'automatic-speech-recognition'] | ['speech', 'speech'] | [ 3.51740360e-01 1.77357361e-01 8.48883316e-02 -6.68003738e-01
-1.39003694e+00 -4.25812483e-01 5.71321487e-01 -1.94522902e-01
-5.08947849e-01 6.85526803e-02 3.64238977e-01 -7.63406038e-01
4.60416794e-01 1.34330034e-01 -4.90294576e-01 -4.80186164e-01
1.98825523e-01 4.48778838e-01 2.35137120e-01 -4.63438034e-01
-2.87913799e-01 2.81662166e-01 -1.23110664e+00 6.20128512e-01
6.79453433e-01 7.44694054e-01 3.45674425e-01 1.21306944e+00
-3.35527956e-01 7.66733408e-01 -9.88990664e-01 -2.17080504e-01
1.49412647e-01 -2.55846888e-01 -6.69897199e-01 2.06150204e-01
4.82255101e-01 -1.52514800e-01 -4.20923293e-01 6.72736168e-01
9.17029619e-01 2.62977362e-01 1.51381180e-01 -7.10781157e-01
-2.17453256e-01 1.10232675e+00 -3.96522194e-01 3.40428680e-01
2.94008642e-01 2.49798726e-02 8.52565229e-01 -9.56760943e-01
-2.03352690e-01 1.51431024e+00 5.31275511e-01 8.64879906e-01
-1.36796367e+00 -6.62055075e-01 5.42560577e-01 -1.28852576e-01
-1.31505179e+00 -1.53790784e+00 5.97805500e-01 -4.93228771e-02
1.55150568e+00 6.00261509e-01 1.17906742e-02 1.18920040e+00
-4.28217143e-01 1.09187496e+00 7.79367983e-01 -5.89844644e-01
1.08668700e-01 1.84004009e-01 3.39432150e-01 2.08354473e-01
-4.20533478e-01 -8.66763443e-02 -8.26718211e-01 1.77360885e-03
1.59261391e-01 -1.53114960e-01 -1.48656994e-01 3.87254894e-01
-9.65550303e-01 5.60884178e-01 -2.42975950e-02 4.85789567e-01
-2.30048567e-01 1.20675499e-02 3.75593603e-01 3.26164961e-01
7.25350201e-01 -8.13444629e-02 -5.24337888e-01 -1.52295217e-01
-1.71761239e+00 -2.34583348e-01 8.47723305e-01 1.00068104e+00
2.92262018e-01 5.67672074e-01 -2.46443614e-01 1.44596183e+00
7.60593414e-01 4.25464571e-01 8.28366458e-01 -4.38292295e-01
8.11029077e-01 8.19485858e-02 -3.70173484e-01 -3.09631322e-02
-1.29223064e-01 -6.00522637e-01 -4.72464651e-01 6.87385798e-02
4.87006046e-02 -2.24863648e-01 -1.33472979e+00 1.60708129e+00
2.54220188e-01 3.59212101e-01 2.11245909e-01 7.02022970e-01
8.18175852e-01 7.83691049e-01 -6.50605857e-02 -1.50611736e-02
1.50573599e+00 -1.59056747e+00 -7.42129624e-01 -8.05915534e-01
4.12820876e-01 -1.03533745e+00 9.10034657e-01 3.16613376e-01
-1.45577550e+00 -4.87431675e-01 -1.09412944e+00 -1.72437906e-01
-2.83381134e-01 5.03162205e-01 -5.69213927e-02 1.07090497e+00
-1.52075362e+00 7.26308227e-02 -1.04257965e+00 -1.72092944e-01
1.48397565e-01 4.92297620e-01 -2.18764052e-01 2.32870296e-01
-9.82992291e-01 9.03442740e-01 1.10852905e-01 2.24100456e-01
-8.61360967e-01 -5.60105503e-01 -1.05497181e+00 1.79650068e-01
2.62770444e-01 -4.59004134e-01 1.85209548e+00 -1.02994454e+00
-2.26797986e+00 7.64206171e-01 -8.99410903e-01 -5.86469591e-01
4.07354414e-01 -3.97819936e-01 -7.69865274e-01 -1.51077062e-01
-4.23121542e-01 2.42144182e-01 1.04342163e+00 -1.13429511e+00
-6.36354029e-01 -3.41958046e-01 -3.80365849e-01 1.78602546e-01
-2.00638503e-01 7.56094337e-01 -5.00214040e-01 -7.71227062e-01
4.88405786e-02 -9.29509044e-01 -1.14814088e-01 -7.64306068e-01
-6.45978808e-01 -1.17269486e-01 7.75624514e-01 -1.28165495e+00
1.38910866e+00 -2.23981023e+00 9.29460488e-03 6.80552498e-02
-9.73369628e-02 7.36084640e-01 -3.56007129e-01 3.69285971e-01
-3.13996255e-01 -4.50872071e-03 -3.32437336e-01 -1.38360977e+00
9.65428799e-02 -1.21666342e-02 -4.31431592e-01 2.43491665e-01
3.22291464e-01 6.16512775e-01 -4.07994181e-01 4.32337709e-02
3.00555319e-01 9.06248152e-01 -2.76311487e-01 3.70496243e-01
3.61677930e-02 2.34232709e-01 2.91464001e-01 4.07848060e-01
6.96543396e-01 2.17021152e-01 1.71943575e-01 3.26325268e-01
-1.30731881e-01 1.33088362e+00 -1.22304642e+00 1.67879748e+00
-8.79998982e-01 6.79692507e-01 9.92264211e-01 -8.88174236e-01
7.12422550e-01 7.93284655e-01 -6.59690052e-02 -3.32360238e-01
8.66598785e-02 4.39832956e-01 1.59840733e-02 -3.55378091e-02
6.21206999e-01 -2.19808236e-01 5.02053648e-02 2.34548181e-01
2.71838337e-01 1.67338997e-01 -1.63348332e-01 2.62877285e-01
1.11360490e+00 -3.50894570e-01 9.09801051e-02 8.11295435e-02
7.56022811e-01 -6.11298323e-01 2.64270604e-01 6.79797649e-01
-1.54900759e-01 9.06195641e-01 -2.02028617e-01 2.14613914e-01
-6.76079154e-01 -1.07700241e+00 2.34943435e-01 1.52425826e+00
-4.47162718e-01 -4.18355793e-01 -7.83827305e-01 -5.36228716e-01
-1.33553788e-01 1.06039310e+00 -9.00159255e-02 2.20030900e-02
-7.86316752e-01 -5.05901456e-01 1.03586829e+00 5.10809183e-01
-1.96303166e-02 -6.41729414e-01 1.45259231e-01 3.80676270e-01
-2.40707442e-01 -1.38796735e+00 -8.91085029e-01 5.38000584e-01
-5.57306349e-01 -2.79536635e-01 -9.92201686e-01 -7.84817576e-01
2.59510607e-01 6.61557376e-01 9.66788650e-01 -1.53604493e-01
5.88177107e-02 1.78032756e-01 -3.13362390e-01 -2.72088289e-01
-1.04918218e+00 3.89041215e-01 1.67215064e-01 3.01386476e-01
4.66752321e-01 -5.51848054e-01 -3.06627035e-01 3.86394978e-01
-6.24802709e-01 -1.83163360e-01 6.46944880e-01 5.30284345e-01
1.25078738e-01 -3.02117318e-01 6.92151368e-01 -4.47846413e-01
5.53943634e-01 -3.11708391e-01 -3.79845947e-01 2.88843244e-01
-2.70295322e-01 3.40873525e-02 5.17262638e-01 -3.67189795e-01
-1.25588536e+00 1.98027432e-01 -8.20442080e-01 -3.28885883e-01
-4.35096532e-01 1.84341162e-01 -4.61931705e-01 2.53307253e-01
2.50319093e-01 6.09886229e-01 7.38692507e-02 -8.78561080e-01
5.73567271e-01 1.50157046e+00 4.52956855e-01 -1.89758465e-01
5.91388822e-01 3.41030173e-02 -9.41024184e-01 -1.36311519e+00
-4.63892192e-01 -9.87611175e-01 -2.89042920e-01 1.99265584e-01
5.37926435e-01 -1.43210053e+00 -5.69474220e-01 6.42831683e-01
-1.36691856e+00 -4.23011392e-01 9.74954758e-03 5.29915392e-01
-1.08433604e-01 4.15000200e-01 -7.53580332e-01 -1.27136612e+00
-5.27240217e-01 -1.29429030e+00 1.43586218e+00 -2.18366068e-02
-3.30345958e-01 -7.19754994e-01 -6.15744106e-02 8.00907016e-01
9.00503695e-01 -1.06340349e+00 2.52806425e-01 -1.31736863e+00
-2.99308062e-01 -1.73449010e-01 1.19672455e-01 8.45444322e-01
1.30939841e-01 -7.28881955e-02 -1.59861517e+00 -4.15846527e-01
1.50969118e-01 -4.92238589e-02 1.16398799e+00 3.95084828e-01
6.42672360e-01 -3.68334323e-01 -2.61187315e-01 2.94730753e-01
8.30715477e-01 1.76287413e-01 5.14351666e-01 -1.57378092e-01
7.03213394e-01 5.95556617e-01 -3.32839936e-01 -2.73507349e-02
5.54629922e-01 7.90462494e-01 -9.87755209e-02 -1.81911558e-01
-5.84407926e-01 1.36443973e-01 1.15465391e+00 1.42236161e+00
4.30944145e-01 -4.61640865e-01 -8.03465128e-01 6.03811264e-01
-1.49957442e+00 -1.00860429e+00 4.19557318e-02 2.38171554e+00
9.45423484e-01 1.85723066e-01 4.15041566e-01 2.57002831e-01
8.70730877e-01 3.95352811e-01 -1.58204600e-01 -6.52319968e-01
-2.38414869e-01 2.55925268e-01 3.53983432e-01 1.07005167e+00
-1.00163126e+00 1.06879413e+00 6.76293421e+00 1.05692792e+00
-1.33740449e+00 3.13410431e-01 5.05310237e-01 -5.15126109e-01
-7.05282092e-02 -2.32895479e-01 -1.33027828e+00 4.22363281e-01
1.69264579e+00 2.53734291e-01 5.60564995e-01 6.49722099e-01
3.87587428e-01 8.36798325e-02 -1.19784796e+00 8.53974402e-01
2.66552210e-01 -8.03675473e-01 -8.60217810e-02 -5.86563051e-02
2.06038758e-01 5.61550438e-01 5.51097170e-02 4.26392347e-01
4.15766060e-01 -8.66898894e-01 9.64594364e-01 1.88747086e-02
5.36849439e-01 -5.66297114e-01 4.14928734e-01 3.97883177e-01
-1.35165739e+00 -3.20794024e-02 1.91339195e-01 5.91211431e-02
5.03201962e-01 6.02651119e-01 -1.13699472e+00 5.36962152e-01
2.34672591e-01 8.52982700e-02 -4.49662268e-01 8.67762804e-01
-1.74169317e-01 1.24143505e+00 -6.19951248e-01 3.36899877e-01
4.00666781e-02 1.69284359e-01 8.96518409e-01 1.84301293e+00
1.78734019e-01 -4.50969189e-01 5.68017848e-02 4.92926419e-01
-7.00872988e-02 -3.85256074e-02 -7.87519738e-02 -1.02573186e-01
5.68383217e-01 1.06590438e+00 -3.71786952e-01 -5.23276150e-01
-4.73172963e-01 1.30880868e+00 3.77261341e-01 4.93918300e-01
-5.98087728e-01 -3.78885329e-01 9.00379121e-01 -5.49623370e-03
5.97004175e-01 -4.77267325e-01 -3.02281648e-01 -1.21684146e+00
2.70734996e-01 -1.14803398e+00 1.00705989e-01 -3.29664767e-01
-1.11936474e+00 1.01179743e+00 -5.74124515e-01 -7.44353354e-01
-3.37447882e-01 -5.37552536e-01 -7.71322966e-01 1.38446057e+00
-1.81377804e+00 -1.13925719e+00 2.15495601e-01 3.74057949e-01
1.07676578e+00 -2.68148899e-01 9.21955943e-01 6.12539649e-01
-1.01874149e+00 1.09794605e+00 1.49075195e-01 2.60831058e-01
7.55256176e-01 -1.19609547e+00 1.10638928e+00 1.15256786e+00
3.28330338e-01 7.80904472e-01 4.97523338e-01 -3.56485665e-01
-1.06134534e+00 -1.02056110e+00 1.30851316e+00 -4.95502561e-01
5.04845262e-01 -9.06424701e-01 -1.05272710e+00 8.05341542e-01
5.40555060e-01 -2.32129380e-01 1.00985003e+00 3.41224462e-01
-5.58298171e-01 -1.70864239e-01 -7.09061742e-01 5.83979309e-01
7.74513721e-01 -9.12454545e-01 -7.58440018e-01 -4.96102311e-02
1.16100955e+00 -3.24318290e-01 -2.82491475e-01 -7.56100118e-02
3.16041976e-01 -7.48760283e-01 9.46503341e-01 -4.85455424e-01
-9.73673984e-02 -3.13773572e-01 -3.01013499e-01 -1.63680768e+00
-8.53820369e-02 -1.18329191e+00 -3.48065495e-01 1.50726783e+00
9.00151312e-01 -7.85267472e-01 4.79751885e-01 6.44074023e-01
-5.94780922e-01 -2.84707963e-01 -1.15177107e+00 -1.06852877e+00
-1.02653041e-01 -7.69804716e-01 4.52416122e-01 5.84614336e-01
-3.56105864e-02 7.03051388e-01 -2.59021521e-01 5.59973478e-01
3.36772531e-01 -4.78570461e-01 5.73322713e-01 -7.08310127e-01
-6.12242222e-01 -7.51410127e-01 -9.86221805e-02 -1.55496979e+00
2.35193551e-01 -1.03172362e+00 5.03668547e-01 -1.43373716e+00
-3.14158469e-01 -2.00243235e-01 -3.12376559e-01 3.97611946e-01
-3.56868505e-01 -1.21341869e-01 2.91177154e-01 -9.12815779e-02
-5.78714013e-01 6.31902277e-01 2.95772493e-01 -1.50420681e-01
-5.03964901e-01 2.75360227e-01 -7.06041276e-01 4.91986394e-01
6.85673594e-01 -4.57860947e-01 -2.00045392e-01 -7.69509315e-01
-4.26980436e-01 4.01568674e-02 1.03405789e-01 -1.04680967e+00
4.85626727e-01 3.80686015e-01 -5.90608381e-02 -4.60918695e-01
7.04122603e-01 -6.74693227e-01 -1.64023772e-01 1.81544535e-02
-5.25325775e-01 -3.10607821e-01 4.01389271e-01 3.93975675e-01
-3.79713625e-01 3.30778286e-02 8.75948191e-01 1.57607540e-01
-1.03380159e-01 -1.88536912e-01 -7.85876930e-01 -4.62354533e-02
5.22632837e-01 -7.31044561e-02 -1.10322192e-01 -4.15024966e-01
-9.03119802e-01 1.23731956e-01 -1.05451167e-01 6.64468884e-01
3.95311505e-01 -8.73536229e-01 -9.99745548e-01 4.98027623e-01
-9.72062722e-02 -4.21749204e-02 2.49050915e-01 8.78844321e-01
2.82932818e-02 5.14545858e-01 7.55051136e-01 -5.34282207e-01
-1.67764270e+00 1.88582316e-01 6.51598215e-01 -1.45623252e-01
-2.91379124e-01 1.21904862e+00 -1.80763230e-02 -6.84905112e-01
6.43108606e-01 -3.23142707e-01 1.44050315e-01 8.41190293e-03
1.04066861e+00 3.91833872e-01 6.45809054e-01 -1.07177055e+00
-6.71358049e-01 9.58373919e-02 -4.54350293e-01 -6.10833824e-01
1.23661292e+00 -5.17735720e-01 2.26801097e-01 6.43718421e-01
1.18832386e+00 4.41184193e-01 -9.99520957e-01 -5.24542749e-01
-4.94879764e-03 -5.35128033e-03 2.16259271e-01 -1.03877485e+00
-8.46234858e-01 1.09267998e+00 3.24796259e-01 3.20493609e-01
8.27740252e-01 5.87996095e-02 1.01870072e+00 3.03432792e-01
-2.70029545e-01 -1.08638954e+00 -2.98845440e-01 7.63294220e-01
8.32441807e-01 -1.30367887e+00 -6.66940212e-01 -4.08662349e-01
-5.91125727e-01 8.45038593e-01 1.52934209e-01 2.04111531e-01
6.92690849e-01 6.97173059e-01 5.12986422e-01 2.75251627e-01
-8.44316483e-01 -2.69289821e-01 4.22556758e-01 2.36258164e-01
7.95733869e-01 2.12613240e-01 2.03880116e-01 9.00125802e-01
-2.23491296e-01 -4.14290458e-01 1.06062777e-01 9.04763162e-01
-5.07366717e-01 -1.17968321e+00 -5.04846931e-01 -1.48120606e-02
-6.43664658e-01 -5.99321723e-01 -5.42200983e-01 1.63621902e-02
-4.99127954e-01 1.63736117e+00 5.28979860e-02 -4.42818701e-01
4.75437671e-01 6.95009112e-01 1.07739627e-01 -8.38772237e-01
-9.63443100e-01 5.22096157e-01 3.78356487e-01 -3.63458306e-01
-1.47895247e-01 -5.41183054e-01 -1.16978776e+00 -2.12539528e-02
-5.54114699e-01 3.60427238e-02 1.17902148e+00 1.15936363e+00
7.51915455e-01 8.99868667e-01 6.94789886e-01 -9.98137951e-01
-8.07249308e-01 -1.26260459e+00 -3.51423085e-01 -1.77498788e-01
9.61893559e-01 8.55654031e-02 -5.85925639e-01 5.05382456e-02] | [14.599628448486328, 6.335084915161133] |
7e739eb6-4649-47df-98dc-47960639b459 | shifted-diffusion-for-text-to-image | 2211.15388 | null | https://arxiv.org/abs/2211.15388v2 | https://arxiv.org/pdf/2211.15388v2.pdf | Shifted Diffusion for Text-to-image Generation | We present Corgi, a novel method for text-to-image generation. Corgi is based on our proposed shifted diffusion model, which achieves better image embedding generation from input text. Unlike the baseline diffusion model used in DALL-E 2, our method seamlessly encodes prior knowledge of the pre-trained CLIP model in its diffusion process by designing a new initialization distribution and a new transition step of the diffusion. Compared to the strong DALL-E 2 baseline, our method performs better in generating image embedding from the text in terms of both efficiency and effectiveness, resulting in better text-to-image generation. Extensive large-scale experiments are conducted and evaluated in terms of both quantitative measures and human evaluation, indicating a stronger generation ability of our method compared to existing ones. Furthermore, our model enables semi-supervised and language-free training for text-to-image generation, where only part or none of the images in the training dataset have an associated caption. Trained with only 1.7% of the images being captioned, our semi-supervised model obtains FID results comparable to DALL-E 2 on zero-shot text-to-image generation evaluated on MS-COCO. Corgi also achieves new state-of-the-art results across different datasets on downstream language-free text-to-image generation tasks, outperforming the previous method, Lafite, by a large margin. | ['Jinhui Xu', 'Changyou Chen', 'Xiao Yang', 'Yizhe Zhu', 'Bingchen Liu', 'Yufan Zhou'] | 2022-11-24 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_Shifted_Diffusion_for_Text-to-Image_Generation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_Shifted_Diffusion_for_Text-to-Image_Generation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['zero-shot-text-to-image-generation'] | ['natural-language-processing'] | [ 3.56932908e-01 3.68184358e-01 -1.35657728e-01 1.32890731e-01
-9.58450079e-01 -6.03850424e-01 1.17908406e+00 -4.13303941e-01
-3.66239876e-01 6.98052824e-01 4.49411482e-01 -1.88721761e-01
4.80244637e-01 -8.93036008e-01 -8.12742710e-01 -6.20864511e-01
3.46564978e-01 5.36176085e-01 9.41240042e-02 -3.30045491e-01
-1.27467453e-01 6.10022321e-02 -1.15220976e+00 3.97880077e-01
1.00962389e+00 7.72820473e-01 5.44695139e-01 8.75646710e-01
-3.99400555e-02 6.41788661e-01 -7.37555206e-01 -6.44032717e-01
3.32804948e-01 -8.04384470e-01 -6.23326719e-01 3.19021821e-01
5.80709100e-01 -5.72685063e-01 -5.23664236e-01 6.39562309e-01
8.74158084e-01 -7.79477805e-02 9.48280156e-01 -1.27855527e+00
-1.30016804e+00 6.75310731e-01 -6.87596917e-01 -1.23731814e-01
2.83450395e-01 4.33265835e-01 8.21674883e-01 -1.28760397e+00
1.15779555e+00 1.18743658e+00 3.00070643e-01 8.08427274e-01
-1.29051757e+00 -5.45575976e-01 -1.42944843e-01 -2.75752604e-01
-1.34329545e+00 -4.93201017e-01 4.88505483e-01 -3.69488120e-01
8.18026006e-01 -2.66924351e-02 4.39189732e-01 1.43150532e+00
1.05358794e-01 1.13282371e+00 9.11025703e-01 -7.39791751e-01
-2.96466630e-02 2.68036008e-01 -6.61974013e-01 6.50076687e-01
-8.97965878e-02 9.01888534e-02 -5.74683964e-01 1.05284236e-01
8.29113483e-01 -4.12363142e-01 -3.72774214e-01 -1.31263375e-01
-1.55377150e+00 9.27064955e-01 4.22680080e-01 3.70108843e-01
-4.82425451e-01 4.08094287e-01 2.57924438e-01 1.14454366e-01
8.54955614e-01 3.15810919e-01 1.91715673e-01 -7.93258846e-02
-1.30114853e+00 2.42536560e-01 4.92610306e-01 1.32017326e+00
4.17238116e-01 2.74062574e-01 -1.02025175e+00 8.37939560e-01
5.02488017e-03 8.87634099e-01 6.32883549e-01 -8.68376791e-01
7.52833128e-01 1.81469440e-01 2.37277031e-01 -4.77450877e-01
2.53702193e-01 -3.49734247e-01 -1.03007543e+00 1.02820776e-01
1.39804959e-01 -4.55534488e-01 -1.29093945e+00 1.65853834e+00
1.55613814e-02 1.19458586e-01 1.59482881e-01 6.53671026e-01
7.20895290e-01 1.22882402e+00 -6.94785267e-02 -1.86076760e-02
1.14886868e+00 -1.45765305e+00 -7.99244523e-01 -3.26317251e-01
5.68526328e-01 -1.08596516e+00 1.09267867e+00 1.52006373e-01
-1.30824459e+00 -6.58700287e-01 -9.27665114e-01 -1.48115829e-01
-2.79351354e-01 5.68294108e-01 3.50270092e-01 5.67441702e-01
-1.45495820e+00 2.29124725e-01 -3.24018836e-01 -3.32900822e-01
5.33017874e-01 -8.67261142e-02 -3.25985730e-01 -3.12298298e-01
-1.20519412e+00 6.50734961e-01 3.05733085e-01 -1.78174064e-01
-9.80105340e-01 -6.91772699e-01 -9.09042895e-01 -1.17566429e-01
1.11845233e-01 -1.02180183e+00 1.25532937e+00 -7.95459092e-01
-1.57182670e+00 8.98249686e-01 -2.87380129e-01 -6.67958200e-01
9.80822206e-01 -1.95489705e-01 -5.07769100e-02 4.37775582e-01
4.43270206e-01 1.73040402e+00 1.09662271e+00 -1.47011447e+00
-4.67340887e-01 3.73159140e-01 6.23010471e-02 1.79423258e-01
-4.96361583e-01 -1.46471620e-01 -1.01476276e+00 -1.12933254e+00
-6.00405335e-01 -1.12974095e+00 -1.36176288e-01 2.66950995e-01
-5.66666543e-01 -1.20164707e-01 9.36227560e-01 -6.25237226e-01
1.18343723e+00 -1.94315398e+00 1.01050235e-01 -2.14794710e-01
2.63626605e-01 4.65604186e-01 -6.41287804e-01 8.02017748e-01
-2.58189663e-02 3.58820528e-01 -2.79847920e-01 -9.48133945e-01
8.22821707e-02 1.43832415e-02 -7.36311674e-01 -1.78248044e-02
3.49864006e-01 1.35066426e+00 -1.02853894e+00 -6.18756890e-01
3.44492555e-01 5.63945889e-01 -4.08097923e-01 2.96038836e-01
-3.85072351e-01 2.51513422e-01 -1.19018123e-01 3.12608212e-01
6.53425395e-01 -4.23614651e-01 -1.88510820e-01 -1.55791529e-02
1.17382511e-01 -2.46930495e-01 -5.85433364e-01 1.90495658e+00
-8.71465564e-01 8.90178919e-01 -4.42253947e-01 -4.17554617e-01
6.50270402e-01 6.22408271e-01 2.71221995e-01 -7.86780298e-01
-3.82653214e-02 2.84497682e-02 -3.99267793e-01 -2.10255727e-01
7.83855796e-01 9.59083065e-02 3.21322307e-02 7.48350441e-01
2.21375167e-01 -4.68182027e-01 7.65900552e-01 9.43757594e-01
8.30181539e-01 3.20484757e-01 -2.44596630e-01 2.25805510e-02
2.18054026e-01 -1.18196011e-01 -1.12380892e-01 8.70366514e-01
1.64762869e-01 1.11642265e+00 3.77821714e-01 1.02743171e-01
-1.38296092e+00 -1.10976529e+00 7.32775331e-02 7.28228331e-01
6.26314282e-02 -4.76247698e-01 -1.04272175e+00 -8.38910103e-01
-3.72866362e-01 1.01107574e+00 -8.16741824e-01 -8.05360898e-02
-3.36977929e-01 -6.22921586e-01 7.78677225e-01 3.68124843e-01
6.58945739e-01 -1.17428505e+00 -1.26174316e-01 1.66349366e-01
-5.78425705e-01 -1.42346275e+00 -1.22466373e+00 -4.73507166e-01
-6.58783734e-01 -4.40775186e-01 -1.59950554e+00 -9.26165223e-01
9.94631946e-01 4.80337262e-01 1.08249855e+00 -1.71059743e-02
-4.45817351e-01 3.63196135e-01 -5.88718116e-01 -3.35260689e-01
-8.55892360e-01 1.16735861e-01 -4.72842604e-01 2.55593210e-02
-2.99050182e-01 -1.33592978e-01 -8.65381002e-01 2.59392589e-01
-1.34938383e+00 5.64932406e-01 8.29925776e-01 9.79339182e-01
4.09703910e-01 -2.11901605e-01 5.73211908e-01 -7.49470055e-01
9.30528164e-01 -3.70901316e-01 -3.43955457e-01 2.43979603e-01
-7.52362192e-01 1.00794077e-01 6.17995799e-01 -5.32159448e-01
-1.23766911e+00 -1.61492154e-01 -4.25566956e-02 -5.91430068e-01
1.87709510e-01 1.45149663e-01 2.48198956e-01 2.97623277e-01
7.96211839e-01 5.16878366e-01 -4.65591252e-02 -8.58280212e-02
1.00384629e+00 6.42144501e-01 5.03748238e-01 -4.34516877e-01
1.11790597e+00 5.33805668e-01 -3.68557125e-01 -6.72612488e-01
-6.13655865e-01 -1.79954857e-01 -5.11571467e-01 -2.49136448e-01
1.01786017e+00 -1.13200903e+00 3.36522646e-02 8.06753933e-01
-1.39249170e+00 -6.79625213e-01 -4.87457812e-01 1.40637249e-01
-4.79007870e-01 4.18355942e-01 -7.76879609e-01 -5.32609224e-01
-7.01087892e-01 -1.22229910e+00 1.62414932e+00 -3.31540294e-02
-2.40651295e-02 -1.10006285e+00 1.70342594e-01 4.35987800e-01
5.50008595e-01 1.46587491e-01 6.67658031e-01 5.20949066e-02
-6.20088458e-01 -2.87844986e-01 -4.89770442e-01 5.38807631e-01
1.93446688e-02 5.16415723e-02 -7.74698913e-01 -4.44139957e-01
-5.80073476e-01 -5.65598488e-01 9.94440198e-01 3.43294799e-01
7.61617780e-01 -3.02386284e-01 -3.99965584e-01 3.61009657e-01
1.38660944e+00 -3.56093757e-02 1.01190019e+00 4.47685122e-02
5.33607781e-01 3.78475487e-01 5.63095212e-01 4.03508455e-01
3.00219268e-01 7.76149154e-01 1.17244042e-01 -6.09847367e-01
-8.56006682e-01 -6.54072881e-01 6.48684263e-01 6.86658800e-01
1.57979742e-01 -9.76756632e-01 -5.38330078e-01 8.22917104e-01
-1.78893745e+00 -9.91613746e-01 -1.52307883e-01 2.07637405e+00
1.12787497e+00 8.23390763e-03 -1.55694894e-02 -2.40834609e-01
7.75248647e-01 3.09723049e-01 -2.90505379e-01 -1.30400479e-01
-2.73388356e-01 3.39258224e-01 4.02984619e-01 5.41267872e-01
-9.00865257e-01 1.33201075e+00 6.68768787e+00 1.37312150e+00
-1.13035417e+00 2.76691109e-01 8.97420704e-01 -3.31703722e-01
-4.86347944e-01 -1.67228162e-01 -6.58882558e-01 5.27482688e-01
7.94962764e-01 -3.71438533e-01 2.69871533e-01 5.95933855e-01
3.86633098e-01 -1.37493595e-01 -8.64614964e-01 8.31963480e-01
3.98799390e-01 -1.72023952e+00 5.00566721e-01 2.46960580e-01
1.33274901e+00 -9.86397266e-02 2.45311171e-01 2.93857396e-01
4.23503160e-01 -9.19161558e-01 1.02931750e+00 3.56779158e-01
1.36188972e+00 -5.50541103e-01 5.82117796e-01 2.23692462e-01
-1.07895148e+00 4.33018684e-01 -2.25562379e-01 4.89590377e-01
5.70717335e-01 8.44103098e-01 -9.02904630e-01 6.43805504e-01
1.89566210e-01 5.89281321e-01 -6.51556671e-01 6.79050326e-01
-5.55528224e-01 6.67098820e-01 -5.01855649e-02 1.64630190e-01
5.06795883e-01 -1.26596332e-01 3.96407932e-01 1.31393647e+00
6.30843222e-01 -2.45521277e-01 8.65709409e-03 1.22337091e+00
-5.35693765e-01 1.02261133e-01 -8.69576752e-01 -2.50674665e-01
3.89893115e-01 1.30475700e+00 -6.92197859e-01 -7.29701579e-01
-2.11548880e-01 1.63811028e+00 1.15745924e-01 5.78370750e-01
-1.09431255e+00 -7.32271254e-01 4.22906391e-02 1.62985608e-01
5.32384932e-01 -1.59192994e-01 1.05803318e-01 -1.08974123e+00
8.78640562e-02 -7.02204108e-01 -1.55434191e-01 -1.28350449e+00
-1.14204919e+00 8.61843526e-01 1.23176277e-01 -1.28010380e+00
-6.40174568e-01 -2.35026315e-01 -7.40853250e-01 1.04674184e+00
-1.58604860e+00 -1.64734387e+00 -2.95609593e-01 4.85333651e-01
7.73103833e-01 -1.83934823e-01 6.36350751e-01 2.20530108e-01
-3.25879455e-01 8.14796984e-01 1.77074343e-01 2.09187508e-01
1.06416714e+00 -1.11708736e+00 9.92461503e-01 1.14905751e+00
3.02857459e-01 2.83065677e-01 5.01798272e-01 -7.64951468e-01
-1.14470887e+00 -1.37392890e+00 9.76749420e-01 -3.82683724e-01
6.04533613e-01 -6.22240961e-01 -3.70868236e-01 4.81418371e-01
9.43254292e-01 -2.97996044e-01 3.26351672e-01 -6.43302441e-01
-2.58686095e-01 1.69701755e-01 -9.22628403e-01 8.81573617e-01
1.14442623e+00 -3.84037584e-01 -2.77591437e-01 6.02020502e-01
1.06537688e+00 -4.00902659e-01 -6.29957378e-01 -9.80466455e-02
2.38201216e-01 -7.69382358e-01 9.25764501e-01 7.49830380e-02
1.15979874e+00 -1.35021046e-01 2.45671377e-01 -1.47818148e+00
-1.74872100e-01 -1.10678756e+00 7.53813162e-02 1.38271344e+00
5.78415990e-01 -3.89868528e-01 6.77793980e-01 1.64526805e-01
-1.00883134e-01 -6.44483209e-01 -6.23349905e-01 -9.42716122e-01
2.21082181e-01 -2.76828498e-01 2.98527926e-01 5.98797321e-01
-3.28008443e-01 5.72683871e-01 -7.58667946e-01 -3.65504652e-01
6.11209273e-01 2.24450771e-02 1.01333380e+00 -4.91166234e-01
-2.36344412e-01 -4.95047390e-01 8.10981393e-02 -1.46174800e+00
6.37374744e-02 -1.12881756e+00 2.08554715e-01 -1.97676075e+00
2.20182508e-01 -2.49143392e-01 3.71772319e-01 5.46124816e-01
-3.97235364e-01 7.46614516e-01 4.96784985e-01 3.21748614e-01
-4.25633550e-01 9.09180582e-01 1.77474439e+00 -3.79967332e-01
-3.23985405e-02 -4.16556329e-01 -6.59297585e-01 1.16515830e-01
5.06491065e-01 -4.05132741e-01 -6.71047807e-01 -6.51520908e-01
2.95501873e-02 2.38320101e-02 2.59925634e-01 -8.39852750e-01
6.94765449e-02 1.10454127e-01 1.90169692e-01 -4.46536601e-01
2.81418264e-01 -2.71615803e-01 -2.43608430e-02 3.62321138e-01
-3.96266192e-01 -2.54432503e-02 1.74632147e-01 5.53818166e-01
-2.32594922e-01 -3.30465734e-01 6.72606707e-01 1.06524406e-02
-4.70600307e-01 4.65212733e-01 -4.69641715e-01 3.39472353e-01
1.04215324e+00 -1.02136746e-01 -5.20255029e-01 -8.43350649e-01
-4.13134336e-01 2.15639502e-01 4.79409069e-01 7.06046700e-01
6.63508773e-01 -1.58307755e+00 -1.14872336e+00 -3.37264210e-04
2.80698776e-01 -1.71048090e-01 2.55376458e-01 6.20278001e-01
-4.90941703e-01 6.04824007e-01 1.31083637e-01 -4.95922983e-01
-1.00134969e+00 3.85776877e-01 -2.95520872e-02 -6.79910958e-01
-5.48703134e-01 8.00812125e-01 5.74962616e-01 -1.44969136e-01
-2.55580395e-02 -1.04094207e-01 2.88191140e-01 -2.88488436e-02
5.67299664e-01 1.09794714e-01 -1.82589129e-01 -5.54934859e-01
1.98891342e-01 4.40422386e-01 -1.95182383e-01 -7.49144971e-01
1.00985384e+00 -1.00199856e-01 1.08635873e-01 -8.51607248e-02
1.12008190e+00 8.07188526e-02 -1.29683530e+00 -1.93196699e-01
-7.10067809e-01 -5.00928879e-01 1.70470446e-01 -1.05126917e+00
-1.18787074e+00 9.97423112e-01 4.89275455e-01 -1.49859026e-01
9.18393195e-01 -2.60944478e-02 1.19970822e+00 8.06766897e-02
2.31821805e-01 -1.07887816e+00 6.91221654e-01 3.39476496e-01
1.27539933e+00 -1.25374591e+00 -1.42098382e-01 -2.01131180e-01
-1.04044664e+00 7.82004416e-01 4.14390624e-01 9.96683165e-02
2.35651821e-01 2.39249095e-01 1.83294535e-01 1.85849622e-01
-9.30119932e-01 -1.99616149e-01 3.85215908e-01 8.31986666e-01
4.91209686e-01 -1.98812112e-01 -1.96016058e-01 1.25441328e-02
-2.09076896e-01 2.31343418e-01 5.76175094e-01 7.04528213e-01
-5.23008443e-02 -1.35150898e+00 -2.01653093e-01 2.34709665e-01
-1.79285362e-01 -5.95364988e-01 -3.71852189e-01 7.81695187e-01
-6.64676055e-02 1.06286991e+00 6.10251464e-02 -1.37118623e-01
1.09525062e-01 1.41968913e-02 4.45251316e-01 -6.50164127e-01
-4.94351208e-01 3.50482047e-01 1.38628958e-02 -3.44364017e-01
-2.47439772e-01 -3.04091960e-01 -1.06081009e+00 -2.74820417e-01
-3.84953797e-01 -1.76316760e-02 6.40057445e-01 6.06700540e-01
7.37218976e-01 5.22060513e-01 7.53986478e-01 -1.14858830e+00
-3.07254076e-01 -1.05246413e+00 -3.30872834e-01 6.15368962e-01
-2.36070785e-03 -2.01445431e-01 -2.15591595e-01 5.17727971e-01] | [11.294049263000488, 0.06761688739061356] |
c2b45b6a-4893-4d63-9f87-a17d651f3abf | decomposing-cryptocurrency-dynamics-into | 2306.17095 | null | https://arxiv.org/abs/2306.17095v1 | https://arxiv.org/pdf/2306.17095v1.pdf | Decomposing cryptocurrency dynamics into recurring and noisy components | This paper investigates the temporal patterns of activity in the cryptocurrency market with a focus on bitcoin, ether, dogecoin, and winklink from January 2020 to December 2022. Market activity measures - logarithmic returns, volume, and transaction number, sampled every 10 seconds, were divided into intraday and intraweek periods and then further decomposed into recurring and noise components via correlation matrix formalism. The key findings include the distinctive market behavior from traditional stock markets due to the nonexistence of trade opening and closing. This was manifest in three enhanced-activity phases aligning with Asian, European, and US trading sessions. An intriguing pattern of activity surge in 15-minute intervals, particularly at full hours, was also noticed, implying the potential role of algorithmic trading. Most notably, recurring bursts of activity in bitcoin and ether were identified to coincide with the release times of significant US macroeconomic reports such as Nonfarm payrolls, Consumer Price Index data, and Federal Reserve statements. The most correlated daily patterns of activity occurred in 2022, possibly reflecting the documented correlations with US stock indices in the same period. Factors that are external to the inner market dynamics are found to be responsible for the repeatable components of the market dynamics, while the internal factors appear to be substantially random, which manifests itself in a good agreement between the empirical eigenvalue distributions in their bulk and the random matrix theory predictions expressed by the Marchenko-Pastur distribution. The findings reported support the growing integration of cryptocurrencies into the global financial markets. | ['Stanisław Drożdż', 'Jarosław Kwapień', 'Maria Skupień', 'Marcin Wątorek'] | 2023-06-29 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-6.84230268e-01 -1.88585758e-01 -1.05075292e-01 3.20706964e-01
-1.95627615e-01 -1.09746468e+00 1.14411676e+00 8.68854448e-02
-8.84509161e-02 8.55654180e-01 4.00942922e-01 -5.32544255e-01
-4.56009209e-01 -6.41467810e-01 -1.78694814e-01 -7.30460286e-01
-7.08951712e-01 1.92049369e-01 3.63136083e-02 -4.45274621e-01
4.19699490e-01 2.15213597e-01 -8.80811632e-01 -1.11611113e-01
2.58819431e-01 1.42414677e+00 -4.89477992e-01 3.21604729e-01
-3.27051319e-02 1.19714832e+00 -5.11318922e-01 -3.61290187e-01
8.78781855e-01 -4.65200037e-01 -4.74400595e-02 9.24453586e-02
-6.26932263e-01 -1.63961127e-01 -2.16909438e-01 9.02136505e-01
-1.39855102e-01 -2.86811143e-01 3.23100358e-01 -1.03735745e+00
-1.75803125e-01 8.38867188e-01 -7.96598792e-01 8.09061050e-01
1.44488573e-01 2.88190782e-01 1.56945527e+00 -9.08671081e-01
5.19082367e-01 3.94922823e-01 6.42737627e-01 -4.74983960e-01
-1.30059659e+00 -6.90730751e-01 -4.11433786e-01 -1.14459554e-02
-9.95377302e-01 -5.39920032e-02 8.51296604e-01 -7.09198117e-01
8.77580047e-01 4.12118822e-01 1.15171862e+00 8.08867276e-01
4.92356211e-01 1.33806542e-01 1.48550773e+00 -2.53128499e-01
1.92995310e-01 6.24625757e-02 -1.43773742e-02 -2.12413654e-01
8.13933849e-01 5.79505503e-01 -7.37365127e-01 -6.50808454e-01
6.95380390e-01 3.62021616e-03 -2.99295038e-01 -1.23034157e-01
-1.56782627e+00 8.36637020e-01 -2.00410277e-01 8.51199090e-01
-9.07890260e-01 -1.58201590e-01 6.17863715e-01 8.82933974e-01
4.87739354e-01 2.40521938e-01 -9.73094106e-01 -7.08442152e-01
-9.55940306e-01 3.38082701e-01 1.41878831e+00 4.71650362e-01
4.20336038e-01 1.83850586e-01 3.84462416e-01 5.71073964e-02
3.39187175e-01 5.29916286e-01 7.75350869e-01 -8.50262880e-01
5.91363251e-01 2.80650705e-01 1.78500280e-01 -9.04049873e-01
-3.18930775e-01 -6.61168456e-01 -6.30734444e-01 1.19954832e-01
8.76460612e-01 -3.89332533e-01 3.12197506e-01 1.25888896e+00
3.41572892e-03 -1.66676685e-01 5.90763092e-02 5.56559980e-01
-4.23339665e-01 5.00132978e-01 -5.85928142e-01 -9.08635557e-01
1.42358303e+00 -4.50569451e-01 -6.22436047e-01 3.02029818e-01
2.15620309e-01 -8.92577946e-01 2.56406635e-01 5.01388431e-01
-9.42473590e-01 3.62480851e-03 -8.49463642e-01 9.90723550e-01
-3.93086746e-02 -6.66834950e-01 6.22292399e-01 6.80985093e-01
-6.46606386e-01 8.03218663e-01 -7.36045659e-01 2.28672743e-01
-3.53201240e-01 -6.78127632e-02 2.08045021e-02 1.00364196e+00
-1.08782589e+00 6.14374280e-01 4.19603795e-01 3.25137854e-01
-3.24631095e-01 -6.74981415e-01 -2.82964617e-01 1.86464682e-01
2.84163445e-01 4.08920273e-02 1.22962523e+00 -1.14972615e+00
-1.40624499e+00 3.07415992e-01 5.33633232e-01 -9.65553641e-01
7.06669986e-01 1.89442337e-01 -7.54049420e-01 1.49978369e-01
-3.92642021e-02 -5.96927941e-01 4.96981114e-01 -6.01310372e-01
-4.89791065e-01 -2.97276199e-01 -4.96388048e-01 -1.65737018e-01
7.60008916e-02 2.42386401e-01 5.51701307e-01 -1.26712739e+00
3.61109316e-01 -7.39480495e-01 3.84191088e-02 -9.62187827e-01
2.53421143e-02 -6.51285052e-02 4.07212228e-01 -7.16520190e-01
1.42203128e+00 -2.06450248e+00 -6.27313018e-01 8.09118211e-01
1.68896690e-01 -6.30829334e-01 4.58786726e-01 1.16084254e+00
-5.83127677e-01 6.76279515e-02 1.49927080e-01 4.75656927e-01
3.45910490e-01 5.42796887e-02 -7.35149324e-01 8.43231320e-01
-7.55971223e-02 9.81603563e-01 -5.57734072e-01 2.98798233e-01
-6.57736510e-02 -3.20750326e-01 -1.85695559e-01 -9.29730460e-02
-1.27512023e-01 2.80665666e-01 -1.01811476e-01 7.43572712e-01
4.56096172e-01 -4.27044511e-01 5.00323474e-01 1.75886586e-01
-8.24413776e-01 6.85547113e-01 -1.09695852e+00 7.50449121e-01
3.07324767e-01 6.95783257e-01 2.65974402e-01 -9.03587759e-01
7.87874401e-01 6.41630948e-01 8.86312127e-01 -1.15336132e+00
6.12588152e-02 7.30610073e-01 7.11897731e-01 -1.73243843e-02
5.15391469e-01 -4.22746330e-01 1.28646269e-02 1.03385484e+00
-1.07113086e-01 3.63926850e-02 5.83233356e-01 -1.07208021e-01
9.29911017e-01 -1.45368114e-01 6.57603800e-01 -6.83279097e-01
1.72465649e-02 -2.20376194e-01 7.01680779e-01 2.65960217e-01
-2.23620832e-01 1.75302356e-01 1.24298775e+00 -3.89869630e-01
-1.01667547e+00 -8.39664102e-01 -2.21455082e-01 4.83654708e-01
-3.08615953e-01 -5.17538071e-01 -3.33812237e-01 -1.44092873e-01
2.59669632e-01 3.50255460e-01 -7.09837437e-01 4.04359639e-01
-3.40720534e-01 -1.03703988e+00 6.17267787e-02 -2.82979086e-02
6.66371644e-01 -9.70952570e-01 -1.13980210e+00 5.34474969e-01
6.96342140e-02 -8.38857532e-01 -5.00051558e-01 4.20718074e-01
-9.58868444e-01 -1.27281237e+00 -6.31501198e-01 1.23437993e-01
3.35840471e-02 -1.38735011e-01 1.09967613e+00 -3.64482015e-01
1.04502402e-01 1.82900950e-01 -2.32860535e-01 -4.15574521e-01
-2.96729863e-01 -3.84036273e-01 1.73071653e-01 2.62205452e-01
3.61208886e-01 -7.17044771e-01 -8.42786908e-01 3.74182999e-01
-7.79409349e-01 -5.44288397e-01 3.70744735e-01 7.22572148e-01
-4.16431390e-02 8.92492980e-02 6.41680419e-01 -3.37019891e-01
8.07947218e-01 -8.32321584e-01 -1.10241175e+00 -9.12115946e-02
-1.30363166e+00 -1.74788758e-01 6.90162107e-02 -2.99876601e-01
-7.58734465e-01 -7.39867091e-01 5.96209466e-01 -3.44266705e-02
4.10579741e-01 1.05685699e+00 6.32115245e-01 1.33953035e-01
3.01181287e-01 2.98008084e-01 2.48547137e-01 -4.45394337e-01
-1.50617942e-01 4.03677851e-01 1.46104738e-01 -4.03716326e-01
9.34117377e-01 2.79895782e-01 -3.36764991e-01 -6.01917922e-01
-8.43785554e-02 -5.22226453e-01 8.39945860e-03 -1.26919746e-01
4.11147684e-01 -9.41626310e-01 -1.07866144e+00 7.19621062e-01
-8.06516588e-01 -2.42825389e-01 -6.40443265e-01 9.48624372e-01
-3.00920486e-01 3.44102681e-01 -1.18537724e+00 -1.12392068e+00
-1.29317716e-01 -5.35452366e-01 -1.78718362e-02 1.86958089e-01
-6.26233578e-01 -1.02479839e+00 8.12393308e-01 1.81611627e-01
7.29774833e-01 5.69824338e-01 6.75655305e-01 -1.21085835e+00
-6.61330521e-01 -1.76778331e-01 -6.75860718e-02 4.22853827e-01
3.55052263e-01 1.58121988e-01 -4.72719699e-01 -3.26523513e-01
6.47685468e-01 2.05463931e-01 7.33421370e-02 1.28576741e-01
-2.29269058e-01 -6.13502502e-01 4.58584845e-01 1.48944408e-01
1.32276046e+00 6.28454208e-01 2.49654830e-01 7.37932742e-01
-4.77443457e-01 4.85285074e-01 1.80753708e-01 9.85151529e-01
6.60818070e-02 3.63762707e-01 3.49160045e-01 2.27445483e-01
6.33865714e-01 -1.20401643e-01 6.21735990e-01 1.32838130e+00
-4.79358673e-01 5.79819441e-01 -8.10003698e-01 4.18500453e-01
-1.59713769e+00 -1.24605155e+00 -8.66946056e-02 2.26720548e+00
5.65463006e-01 4.56355482e-01 5.54540277e-01 1.27779827e-01
2.74874389e-01 4.03483242e-01 -2.08646223e-01 -1.51456699e-01
-5.18885195e-01 2.66544074e-01 9.47895408e-01 -2.29165293e-02
-4.80777293e-01 8.22309405e-02 6.89134264e+00 1.50776833e-01
-1.12665200e+00 4.56580892e-02 7.34140754e-01 -7.37811550e-02
-4.50202376e-01 3.61461222e-01 -9.91701707e-02 9.48745787e-01
1.35137510e+00 -9.43177044e-01 5.58625877e-01 3.95722985e-01
4.68477070e-01 -3.55188519e-01 -6.04982197e-01 8.55926335e-01
-5.21253109e-01 -1.29721558e+00 -5.63303351e-01 8.43114793e-01
8.54546607e-01 4.74329382e-01 -5.56096472e-02 1.27197921e-01
4.98040803e-02 -4.04155225e-01 1.10797060e+00 4.66363937e-01
1.59982949e-01 -7.05105126e-01 7.85667717e-01 2.69963682e-01
-1.14135623e+00 -2.87259340e-01 1.48256913e-01 -5.00419497e-01
2.19172195e-01 8.46428216e-01 -3.58209938e-01 7.32793927e-01
7.41112649e-01 5.11132181e-01 3.76769118e-02 6.38686299e-01
1.45249277e-01 9.30671215e-01 -4.23936129e-01 9.82170030e-02
3.09942067e-01 -1.14341950e+00 6.34564340e-01 6.56907678e-01
5.40671825e-01 5.42716905e-02 -6.02920949e-01 9.46332455e-01
2.94550769e-02 4.57556061e-02 -5.25467753e-01 -6.14260912e-01
2.94789374e-01 1.13622606e+00 -9.02441382e-01 -6.16733246e-02
-9.56334174e-01 4.97307628e-01 -6.03944898e-01 4.73452598e-01
-7.42985249e-01 -3.48529756e-01 5.92624068e-01 3.65137458e-01
5.11003017e-01 -5.96348763e-01 -4.56407666e-01 -1.42996669e+00
5.78488708e-01 -1.10312462e+00 2.05776036e-01 -1.03908166e-01
-1.12365544e+00 2.84364700e-01 -1.59884810e-01 -1.39553547e+00
-7.10232019e-01 -2.98425466e-01 -7.36252546e-01 8.52192104e-01
-1.23701668e+00 -3.42363827e-02 4.72512335e-01 6.85944736e-01
-9.99664888e-02 -5.00965357e-01 4.97534513e-01 1.01716384e-01
-1.67057917e-01 -7.67878443e-02 6.55800223e-01 3.79405022e-01
1.44759357e-01 -1.39904714e+00 5.79136431e-01 7.78937340e-01
6.08426094e-01 8.07206094e-01 6.87744021e-01 -7.65903711e-01
-1.29246366e+00 -2.60844063e-02 1.03379655e+00 -2.81965226e-01
1.90715230e+00 -1.87722772e-01 -5.24913192e-01 5.59708059e-01
8.51156414e-01 -1.00188784e-01 6.12475336e-01 -2.32865557e-01
-3.70833904e-01 -1.30963296e-01 -6.33037210e-01 1.78528994e-01
1.59372196e-01 -6.77303016e-01 -9.76857483e-01 -1.86840873e-02
3.16981554e-01 1.02895647e-01 -1.02114105e+00 2.05213040e-01
7.91149616e-01 -1.30230737e+00 2.90653259e-01 -1.91126630e-01
9.57986806e-03 -5.05455434e-02 -1.49327487e-01 -1.12705326e+00
-1.11161195e-01 -1.55851197e+00 -1.25290409e-01 9.80498970e-01
5.54389954e-01 -1.37809408e+00 6.59836054e-01 6.28026128e-01
6.21225715e-01 -3.44681352e-01 -1.47560728e+00 -8.22324455e-01
2.40898486e-02 -1.12063065e-01 5.04048049e-01 1.23433363e+00
6.83124602e-01 -8.40696767e-02 -4.19997461e-02 -4.82157707e-01
7.18022883e-01 5.21372199e-01 3.92920315e-01 -1.18725204e+00
-7.07588851e-01 -6.82762682e-01 -2.18999505e-01 -7.85410523e-01
-5.46749353e-01 -6.06760681e-01 -5.90325296e-01 -2.88655668e-01
1.43826772e-02 -7.44989663e-02 -5.49953640e-01 -2.07883120e-01
5.27013361e-01 -1.09076872e-01 4.70449835e-01 6.88775420e-01
1.30469501e-01 2.82836229e-01 6.18202567e-01 1.79598808e-01
-1.23489805e-01 1.57249361e-01 -3.65633905e-01 3.70062470e-01
7.54329741e-01 -8.47707316e-02 8.22772533e-02 3.90299976e-01
8.78110290e-01 6.25170112e-01 2.58132488e-01 -7.54518867e-01
8.40447247e-02 -1.47418633e-01 -1.26330525e-01 -5.61580598e-01
-3.01564932e-01 -9.42934275e-01 8.15499961e-01 7.35034347e-01
8.49496946e-02 8.17605793e-01 -8.20433721e-02 4.83335227e-01
-8.25120628e-01 1.48022339e-01 2.30716065e-01 -1.70154661e-01
7.69483894e-02 -1.03989571e-01 -7.21252501e-01 9.94428173e-02
1.18030930e+00 -2.30585620e-01 -2.18188465e-01 -4.68780994e-01
-7.94676900e-01 -4.18703072e-02 3.24685931e-01 3.40529591e-01
-4.68267202e-02 -1.21540749e+00 -7.52388775e-01 1.17541268e-01
-3.57392371e-01 -7.22543061e-01 -1.00727975e-01 1.45020795e+00
-6.63840771e-01 7.11546481e-01 -1.62211165e-01 -1.70219868e-01
-4.60487247e-01 2.37873226e-01 5.23451507e-01 -6.87108874e-01
-5.27403235e-01 1.85033120e-02 -1.25048742e-01 2.00829878e-01
-8.45800787e-02 -7.26815283e-01 2.51600116e-01 8.99460971e-01
4.79000926e-01 5.34538448e-01 1.12760298e-01 -3.44769537e-01
-1.21693671e-01 1.56065926e-01 2.77111292e-01 -3.74364942e-01
1.42055547e+00 -7.93336183e-02 -6.19196117e-01 9.31504130e-01
9.74749267e-01 4.54001486e-01 -1.12208438e+00 -1.33603424e-01
1.06470573e+00 -6.40860274e-02 -5.89581072e-01 -5.75988710e-01
-1.23752415e+00 7.69259706e-02 1.48117110e-01 1.10785651e+00
6.45520210e-01 -9.49442685e-02 6.35432124e-01 -1.56230517e-02
6.79218829e-01 -1.19251776e+00 -3.28514755e-01 4.23362166e-01
8.34805727e-01 -6.53665781e-01 -2.34885588e-01 2.66808361e-01
-3.79522145e-01 1.23739362e+00 -3.79424691e-01 -2.98787922e-01
1.18281806e+00 5.61110377e-01 2.46703979e-02 -2.81546742e-01
-9.67457354e-01 3.73748958e-01 -2.42224157e-01 -4.23955113e-01
4.70338315e-01 4.10127133e-01 -9.13086534e-01 7.34736025e-01
-3.74256521e-01 -1.29817933e-01 7.51069427e-01 7.93267965e-01
-6.21642508e-02 -1.07828820e+00 -4.81239378e-01 3.55910212e-01
-1.01466286e+00 -2.83873826e-01 -3.22870731e-01 1.09668386e+00
-3.25875401e-01 5.79878092e-01 4.56847668e-01 -1.50809065e-01
2.16427967e-01 2.41994962e-01 -1.61341444e-01 -1.10020274e-02
-1.24821317e+00 6.11089706e-01 5.15664332e-02 -6.33955836e-01
-3.57444048e-01 -1.24997711e+00 -9.29774702e-01 -4.55150813e-01
-3.48776102e-01 7.81327367e-01 5.03275454e-01 7.76405811e-01
-1.68346521e-02 -1.02927178e-01 9.68471706e-01 -4.37707543e-01
-1.11463177e+00 -1.24297476e+00 -1.39756525e+00 2.06205279e-01
4.87841576e-01 -1.83799267e-01 -1.00132251e+00 -1.12357840e-01] | [4.704764366149902, 4.124517440795898] |
ea1f46ee-d771-40a6-9c37-688f6d536708 | gallery-sampling-for-robust-and-fast-face | 2305.07495 | null | https://arxiv.org/abs/2305.07495v1 | https://arxiv.org/pdf/2305.07495v1.pdf | Gallery Sampling for Robust and Fast Face Identification | Deep learning methods have been achieved brilliant results in face recognition. One of the important tasks to improve the performance is to collect and label images as many as possible. However, labeling identities and checking qualities of large image data are difficult task and mistakes cannot be avoided in processing large data. Previous works have been trying to deal with the problem only in training domain, however it can cause much serious problem if the mistakes are in gallery data of face identification. We proposed gallery data sampling methods which are robust to outliers including wrong labeled, low quality, and less-informative images and reduce searching time. The proposed sampling-by-pruning and sampling-by-generating methods significantly improved face identification performance on our 5.4M web image dataset of celebrities. The proposed method achieved 0.0975 in terms of FNIR at FPIR=0.01, while conventional method showed 0.3891. The average number of feature vectors for each individual gallery was reduced to 17.1 from 115.9 and it can provide much faster search. We also made experiments on public datasets and our method achieved 0.1314 and 0.0668 FNIRs at FPIR=0.01 on the CASIA-WebFace and MS1MV2, while the convectional method did 0.5446, and 0.1327, respectively. | ['Jongju Shin', 'Pyoung-gang Lim', 'Myung-Cheol Roh'] | 2023-05-12 | null | null | null | null | ['face-recognition', 'face-identification'] | ['computer-vision', 'computer-vision'] | [-3.02579015e-01 -4.39243972e-01 2.62574136e-01 -5.56311727e-01
-7.25506604e-01 -4.31750536e-01 3.55355233e-01 -4.34278011e-01
-4.62403744e-01 8.26780796e-01 -2.83905685e-01 2.64045507e-01
-1.27437323e-01 -8.49647999e-01 -5.74405372e-01 -7.81892180e-01
-2.18898393e-02 3.34396750e-01 -2.32604340e-01 3.53689156e-02
4.79655266e-01 6.46495461e-01 -1.73254776e+00 2.96151221e-01
7.68479824e-01 1.39080429e+00 -3.50733064e-02 3.14044088e-01
-8.89081955e-02 5.57178676e-01 -9.07863140e-01 -8.49652112e-01
6.73242927e-01 -1.19211026e-01 -5.38762569e-01 3.97614568e-01
8.38022232e-01 -3.73291671e-01 -9.14277881e-02 1.45255744e+00
8.10127676e-01 -8.89136419e-02 6.63283765e-01 -1.59330654e+00
-9.60753977e-01 2.37363830e-01 -1.17905581e+00 4.45409343e-02
-5.78020327e-02 -3.71267349e-02 3.14962476e-01 -1.23920083e+00
2.68150061e-01 1.48273218e+00 8.43193889e-01 7.63434529e-01
-8.77337754e-01 -1.32128537e+00 -4.78955179e-01 3.55566591e-01
-1.72975254e+00 -7.86318123e-01 3.77256453e-01 -2.19555393e-01
4.52126980e-01 3.49266201e-01 1.98039904e-01 6.29044294e-01
-1.46368429e-01 5.56185603e-01 1.13327241e+00 -4.08531189e-01
-1.80802703e-01 4.48608905e-01 1.35648802e-01 8.76597583e-01
5.13623953e-01 -1.09968767e-01 -4.09719527e-01 -4.96121645e-02
7.50152469e-01 2.04623476e-01 -1.01641603e-01 2.50273228e-01
-7.11292624e-01 9.37371612e-01 3.09812903e-01 2.38209337e-01
-1.85935274e-01 -2.75760323e-01 3.32559556e-01 4.90841836e-01
3.01975608e-01 1.43289432e-01 -3.80972952e-01 2.49786764e-01
-1.00299537e+00 -1.02061359e-02 4.25239712e-01 8.21059763e-01
7.59109914e-01 2.74370074e-01 -1.30728647e-01 1.41170180e+00
1.28709599e-01 6.61296725e-01 6.66427016e-01 -8.95064652e-01
1.52409419e-01 7.27520943e-01 1.32421598e-01 -1.29344034e+00
-2.33435616e-01 -3.23354363e-01 -1.15115237e+00 3.99084210e-01
5.48681855e-01 -1.19840488e-01 -1.13073802e+00 1.54427445e+00
3.00129324e-01 -6.47741705e-02 -4.76548709e-02 7.85573661e-01
1.09132218e+00 5.18208265e-01 -1.00536086e-01 -2.20966235e-01
1.36888623e+00 -7.62170851e-01 -6.87338650e-01 1.50527000e-01
3.60953867e-01 -1.24282074e+00 9.71539736e-01 7.89329410e-01
-8.40497971e-01 -8.32218587e-01 -7.86622286e-01 2.86306530e-01
-3.34036052e-01 8.79122674e-01 5.08907914e-01 1.01000786e+00
-1.10115755e+00 5.01196742e-01 -2.50337180e-02 -3.63197267e-01
1.03203523e+00 5.35039306e-01 -7.50567377e-01 -1.55180156e-01
-7.49608815e-01 4.19067174e-01 2.38801375e-01 1.10226214e-01
-9.39558566e-01 -4.32923526e-01 -4.07819331e-01 -1.67812839e-01
2.80722052e-01 1.88795790e-01 8.43084812e-01 -1.16427565e+00
-9.31127071e-01 1.08410358e+00 -6.23988956e-02 -2.46683851e-01
4.74722117e-01 -4.07365635e-02 -9.31718886e-01 4.56494689e-02
3.07380140e-01 9.20803785e-01 7.65687585e-01 -1.17537081e+00
-8.05861235e-01 -7.20874727e-01 -4.01104391e-01 -1.73779681e-01
-5.63954115e-01 2.77875930e-01 -4.46726292e-01 -5.00041187e-01
5.76812029e-02 -8.16580474e-01 2.10623518e-01 1.77358732e-01
-2.73131728e-01 -3.43885243e-01 9.78472233e-01 -8.87555003e-01
9.40474391e-01 -2.12386441e+00 -7.15755641e-01 2.53195345e-01
-2.29536444e-02 8.06400836e-01 -2.16861516e-01 8.20729211e-02
-1.80944636e-01 3.43964815e-01 5.53360805e-02 -2.33345926e-01
-2.76017278e-01 -1.47697255e-01 5.20941913e-02 5.61483204e-01
7.61710480e-02 4.00882453e-01 -3.28504711e-01 -7.12770581e-01
-1.43859405e-02 5.29367924e-01 -4.52871382e-01 1.81301847e-01
4.00413930e-01 5.70257492e-02 -1.68699455e-02 1.05601692e+00
1.25666928e+00 -8.71298239e-02 -1.83730438e-01 -4.64404106e-01
1.40117943e-01 -5.38285136e-01 -1.65127575e+00 1.14387167e+00
-2.72702157e-01 6.16210580e-01 7.89147541e-02 -8.73792410e-01
1.21897101e+00 1.92981288e-01 4.56084937e-01 -7.57269263e-01
2.31982276e-01 1.33562326e-01 -1.29893854e-01 -6.59395635e-01
2.96473563e-01 1.06900416e-01 2.78208822e-01 4.25686955e-01
6.30499125e-02 5.32235444e-01 3.11076283e-01 2.93377955e-02
5.30723095e-01 -1.86639458e-01 6.69900551e-02 -3.65663320e-01
7.69673586e-01 -1.71465993e-01 8.43259692e-01 5.62474191e-01
-3.35642368e-01 6.27247989e-01 1.98799595e-01 -7.89669096e-01
-1.04731679e+00 -6.23167813e-01 -4.42583650e-01 8.21636498e-01
-1.20046042e-01 -1.89049914e-01 -8.75099421e-01 -8.02713335e-01
6.28608614e-02 1.49517938e-01 -5.50365329e-01 -3.82847935e-02
-3.74154419e-01 -1.08431423e+00 7.37830341e-01 1.48042291e-01
1.19997489e+00 -1.16579425e+00 -5.96005246e-02 -1.66020676e-01
-4.96847890e-02 -8.76508892e-01 -5.10745168e-01 -5.75494409e-01
-4.66561377e-01 -1.36075366e+00 -6.18764579e-01 -1.01880634e+00
9.75322902e-01 3.44517350e-01 9.42817807e-01 4.41912830e-01
-7.26894796e-01 -2.55410433e-01 -2.84837037e-01 -4.21953738e-01
-1.46945417e-01 -4.06250387e-01 3.60757768e-01 4.64350909e-01
8.69157493e-01 -8.68297517e-02 -5.91861129e-01 6.54828131e-01
-7.97685862e-01 -4.66030061e-01 5.51078796e-01 1.14732134e+00
3.57660085e-01 2.86460519e-01 8.55969548e-01 -6.83334291e-01
5.70211589e-01 -2.98938751e-01 -7.95015931e-01 4.60304022e-01
-8.36426497e-01 -7.82292560e-02 6.72939062e-01 -5.26647806e-01
-9.88724470e-01 1.21105678e-01 -1.56090483e-01 -4.17273849e-01
-2.02748641e-01 -2.96856403e-01 -3.59002352e-01 -4.26193506e-01
7.46464729e-01 2.09590957e-01 1.80330589e-01 -6.65338457e-01
-1.64113835e-01 1.10465384e+00 4.33239192e-01 -2.90096372e-01
7.04697311e-01 3.45550716e-01 -1.98994391e-02 -9.59328175e-01
-4.22838509e-01 -2.89257079e-01 -2.86126792e-01 -2.16471970e-01
4.32401896e-01 -8.40917170e-01 -1.17575479e+00 8.35060120e-01
-8.29495609e-01 3.70676160e-01 9.30644870e-02 2.53823847e-01
1.67186201e-01 6.37851536e-01 -5.40911674e-01 -1.04349291e+00
-6.96098089e-01 -1.19822788e+00 8.22623074e-01 5.46047926e-01
1.81336001e-01 -4.17528361e-01 -5.07055521e-01 4.90266562e-01
3.73220056e-01 1.75202802e-01 2.72046387e-01 -7.15571344e-01
-2.99939901e-01 -4.36750650e-01 -6.25142276e-01 6.85911179e-01
4.26699698e-01 3.64772305e-02 -1.20111287e+00 -4.15756881e-01
1.75036564e-01 -5.80131829e-01 8.06291401e-01 2.13613525e-01
1.58495140e+00 -5.48418760e-01 -1.27943829e-01 7.01689482e-01
1.60400391e+00 4.16274965e-01 8.35110247e-01 9.91087258e-02
6.08927429e-01 5.66283524e-01 6.76898956e-01 5.31578600e-01
-1.55165821e-01 5.63032210e-01 2.56673306e-01 -1.07263699e-02
-1.24319986e-01 -1.14396095e-01 8.78352746e-02 1.88862279e-01
-6.08119275e-03 5.99970445e-02 -6.75115108e-01 5.45771599e-01
-1.49789727e+00 -1.16776073e+00 -3.26630384e-01 2.26517129e+00
1.02561224e+00 -1.76133305e-01 1.76981911e-01 3.59843075e-01
9.23363924e-01 -2.70576328e-01 -4.01056439e-01 -1.14662185e-01
-2.78479636e-01 7.57887810e-02 7.34138727e-01 3.17849845e-01
-1.14002717e+00 9.04296041e-01 5.63553381e+00 1.21571827e+00
-1.11479747e+00 2.08760351e-01 1.18789542e+00 -1.22075073e-01
3.84368569e-01 -4.33274597e-01 -1.31320000e+00 7.82053232e-01
6.41365647e-01 2.71797657e-01 4.35972005e-01 1.05681825e+00
-8.61636624e-02 -1.85859710e-01 -6.99932754e-01 1.74666035e+00
4.05996412e-01 -1.16738451e+00 2.33608902e-01 1.21984139e-01
7.92382360e-01 -5.32253981e-01 1.63669467e-01 9.11229774e-02
2.49720573e-01 -1.44012356e+00 2.08335802e-01 4.68958348e-01
9.92608309e-01 -9.50946808e-01 9.51667011e-01 2.25076705e-01
-9.78109658e-01 -1.36365548e-01 -9.04095352e-01 1.09911598e-01
-4.33753490e-01 6.87722206e-01 -8.85882318e-01 2.29210243e-01
1.10737169e+00 2.91944534e-01 -7.92965472e-01 1.10689807e+00
3.82313073e-01 4.29554373e-01 -4.64469999e-01 4.67669684e-03
-3.16100307e-02 -1.21263564e-01 1.48066413e-02 8.78918111e-01
4.35562253e-01 -8.20636973e-02 1.90222345e-03 4.28778678e-01
-4.93823588e-01 3.32288146e-01 -4.34525758e-01 2.13611394e-01
6.73549294e-01 1.57495117e+00 -5.38982093e-01 -3.90225887e-01
-1.70303568e-01 8.56285393e-01 1.39082491e-01 4.09603715e-02
-8.63564193e-01 -5.12239456e-01 5.61119258e-01 2.19110511e-02
1.15121491e-01 2.32426450e-01 -1.55730069e-01 -1.06348646e+00
5.90480119e-02 -1.23014402e+00 5.80836177e-01 -4.77219075e-01
-1.43528223e+00 7.72625387e-01 -3.64300281e-01 -9.53926563e-01
9.18012783e-02 -8.24673057e-01 -3.46240669e-01 9.91898298e-01
-1.09535134e+00 -1.06121111e+00 -5.96392751e-01 8.56519341e-01
7.25723624e-01 -9.03222680e-01 7.53016829e-01 7.83188939e-01
-7.71453857e-01 1.14217067e+00 2.79516041e-01 5.55799603e-01
1.01218462e+00 -7.72352517e-01 -1.33187538e-02 8.88090849e-01
2.39036426e-01 6.00566268e-01 3.10664624e-01 -5.20075202e-01
-1.16718411e+00 -1.13045037e+00 6.92679882e-01 -2.99436748e-01
-1.58914149e-01 -8.37283432e-02 -8.10502291e-01 3.24941963e-01
8.62737074e-02 1.40567571e-01 5.72047591e-01 -1.95900843e-01
-2.69555777e-01 -7.41258323e-01 -1.83818769e+00 3.07830215e-01
8.48080516e-01 -1.51281357e-01 -3.98971699e-02 6.18295431e-01
-5.53592630e-02 3.91098522e-02 -7.17001081e-01 2.51390934e-01
6.06264114e-01 -1.10182619e+00 1.02324426e+00 -3.80821556e-01
9.38407890e-03 -3.50684106e-01 -2.23674446e-01 -9.14032757e-01
-3.45618993e-01 -3.51192385e-01 3.57126236e-01 1.68795729e+00
2.02624172e-01 -5.05920231e-01 1.02399945e+00 4.83519793e-01
2.54434764e-01 -5.60998619e-01 -7.08283782e-01 -7.75860190e-01
-2.19058812e-01 7.49361366e-02 9.53117430e-01 9.99683380e-01
-7.09798217e-01 9.57930833e-02 -7.66841173e-01 1.48781063e-02
9.93652344e-01 -1.56626955e-01 8.64584744e-01 -1.32709920e+00
1.46882907e-01 -2.19248384e-01 -3.20630819e-01 -4.53082502e-01
-1.32609397e-01 -5.65498352e-01 -3.27123225e-01 -1.11494136e+00
3.22491348e-01 -6.07740641e-01 -1.34433731e-01 7.39279091e-01
-1.15258135e-01 9.83849406e-01 -1.04884341e-01 3.73893499e-01
-2.05789581e-01 2.25596070e-01 1.06972635e+00 -1.26093328e-01
1.97468668e-01 -7.20720142e-02 -7.44128168e-01 6.80275321e-01
1.18291080e+00 -3.57330889e-01 -2.61950523e-01 -4.41119313e-01
-2.08299428e-01 -3.84482771e-01 7.55616352e-02 -1.08963692e+00
2.26094961e-01 2.90542748e-02 1.14404356e+00 -6.91761136e-01
4.42767709e-01 -7.58722126e-01 3.43868136e-01 4.14776266e-01
7.28544686e-03 2.35613540e-01 1.70704469e-01 1.68371111e-01
-1.62318349e-01 -4.04212475e-01 1.24744296e+00 -3.65183204e-01
-1.01123834e+00 4.91264969e-01 2.84379244e-01 3.41673242e-03
1.20484579e+00 -4.10455078e-01 -2.17934519e-01 -2.36290529e-01
-4.18283850e-01 3.74579020e-02 3.00857991e-01 3.59232247e-01
7.63813496e-01 -1.66394508e+00 -9.90151346e-01 7.04614341e-01
-8.51702318e-02 -4.28547949e-01 2.04026312e-01 3.28313649e-01
-6.19399667e-01 2.18613610e-01 -7.14736998e-01 -5.11008501e-01
-1.90370476e+00 4.41197038e-01 3.73200655e-01 1.65873945e-01
-1.48969889e-01 1.15636289e+00 -6.99289739e-02 -3.44291806e-01
4.13173556e-01 3.25039417e-01 -3.92769486e-01 2.43792146e-01
9.21290159e-01 5.08248270e-01 1.89707801e-01 -9.10524487e-01
-4.49789107e-01 7.76157558e-01 -2.91455686e-01 2.20710814e-01
1.27321577e+00 9.28556267e-03 -3.46974820e-01 -2.46320322e-01
1.40926051e+00 -6.98480010e-03 -9.39626217e-01 -1.14737637e-01
-2.18363404e-01 -9.81764019e-01 -1.45004079e-01 -8.84009242e-01
-1.55577481e+00 7.54960179e-01 1.27147305e+00 2.44378641e-01
1.28125107e+00 -4.32586402e-01 7.56035030e-01 3.19426656e-01
4.47747469e-01 -1.26600707e+00 7.41267875e-02 8.86449963e-02
8.72792304e-01 -1.84449577e+00 3.56149860e-02 -1.91762090e-01
-6.26533210e-01 1.08035815e+00 8.73650372e-01 -3.07635013e-02
5.80655873e-01 2.97668148e-02 1.12063743e-01 -1.08600385e-01
-3.02575380e-01 1.46270782e-01 2.15992033e-01 6.07783556e-01
3.36452305e-01 -6.94263875e-02 -3.96816015e-01 4.56199616e-01
-2.08134249e-01 6.45781979e-02 3.57759655e-01 6.65345430e-01
-4.35092241e-01 -1.13772500e+00 -9.18380558e-01 6.91326320e-01
-9.28972483e-01 1.28310984e-02 -2.86005527e-01 6.83401108e-01
5.18248439e-01 1.20900869e+00 -1.96470171e-02 -5.79995453e-01
2.18884693e-03 8.20394829e-02 3.48483562e-01 -1.86852202e-01
-3.76977026e-01 -1.03220701e-01 5.36239799e-03 -4.22039419e-01
-2.57227689e-01 -4.08408582e-01 -8.19306076e-01 -8.56697321e-01
-4.38144863e-01 3.65724444e-01 7.30825424e-01 4.21934694e-01
3.95683050e-01 -5.98117076e-02 8.20273340e-01 -3.04474115e-01
-7.29613602e-01 -1.20227873e+00 -7.52058268e-01 8.16181421e-01
3.18396911e-02 -6.09817743e-01 -2.95731544e-01 1.65538102e-01] | [13.140542984008789, 0.8133897185325623] |
31d2ef92-eec4-4a2a-8eaa-3559cb4f0670 | narrationbot-and-infobot-a-hybrid-system-for | 2111.03994 | null | https://arxiv.org/abs/2111.03994v2 | https://arxiv.org/pdf/2111.03994v2.pdf | NarrationBot and InfoBot: A Hybrid System for Automated Video Description | Video accessibility is crucial for blind and low vision users for equitable engagements in education, employment, and entertainment. Despite the availability of professional and amateur services and tools, most human-generated descriptions are expensive and time consuming. Moreover, the rate of human-generated descriptions cannot match the speed of video production. To overcome the increasing gaps in video accessibility, we developed a hybrid system of two tools to 1) automatically generate descriptions for videos and 2) provide answers or additional descriptions in response to user queries on a video. Results from a mixed-methods study with 26 blind and low vision individuals show that our system significantly improved user comprehension and enjoyment of selected videos when both tools were used in tandem. In addition, participants reported no significant difference in their ability to understand videos when presented with autogenerated descriptions versus human-revised autogenerated descriptions. Our results demonstrate user enthusiasm about the developed system and its promise for providing customized access to videos. We discuss the limitations of the current work and provide recommendations for the future development of automated video description tools. | ['Pooyan Fazli', 'Ilmi Yoon', 'Abhishek Das', 'Yash Kant', 'Jose M. Castanon', 'Lothar Narins', 'Aditya Bodi', 'Yue-Ting Siu', 'Shasta Ihorn'] | 2021-11-07 | null | null | null | null | ['video-description'] | ['computer-vision'] | [-7.35892951e-02 -9.34531353e-03 -6.07381836e-02 -3.90350848e-01
-8.51603985e-01 -6.43205523e-01 2.09589958e-01 -1.19405471e-01
-6.24926925e-01 7.55517304e-01 8.52767587e-01 -2.89184868e-01
-4.75400686e-02 -2.59542406e-01 -3.80134463e-01 -4.59738187e-02
3.81176263e-01 -1.65436476e-01 3.86527419e-01 -1.48298576e-01
4.85550702e-01 2.20811084e-01 -2.41976833e+00 3.91563356e-01
1.06626451e+00 4.47190106e-01 7.22246349e-01 9.41059947e-01
-3.26960124e-02 1.01791406e+00 -5.40084004e-01 -4.91735965e-01
1.09278649e-01 -6.83908641e-01 -5.32291949e-01 5.53284109e-01
1.08420587e+00 -1.27355134e+00 -4.94757414e-01 9.38654244e-01
8.48263860e-01 2.22656712e-01 1.45917639e-01 -9.74160135e-01
-7.67067790e-01 -1.20339863e-01 1.22218601e-01 1.24657966e-01
1.43351042e+00 5.36338508e-01 4.65418071e-01 -8.11999679e-01
6.50100052e-01 8.48195910e-01 5.01659274e-01 6.65273309e-01
-9.54502761e-01 -4.61695135e-01 -9.24491137e-02 5.05260289e-01
-1.13211977e+00 -9.80914474e-01 -1.28363237e-01 -9.46163654e-01
1.21297228e+00 2.38664150e-01 1.44380295e+00 8.67670476e-01
-1.93087116e-01 2.40804344e-01 8.00390959e-01 -6.16319954e-01
1.05069496e-01 7.22042680e-01 -2.75099874e-01 5.99325180e-01
4.48739767e-01 -2.49517217e-01 -5.43557167e-01 1.12816878e-01
8.42734158e-01 -2.22839072e-01 -9.53746617e-01 -3.64167360e-03
-1.09444964e+00 6.02104247e-01 -2.78776232e-03 3.85914259e-02
-3.95352632e-01 -3.01935703e-01 2.34181285e-01 4.61510092e-01
1.38831183e-01 3.72013658e-01 2.02022400e-02 -7.51445115e-01
-7.26665020e-01 2.95731425e-01 6.66608632e-01 1.22008836e+00
3.45940351e-01 -2.79517740e-01 -4.42090601e-01 7.92059243e-01
3.18140656e-01 3.85904342e-01 5.29154837e-01 -1.10893023e+00
2.01821730e-01 3.96130621e-01 6.64022744e-01 -7.49194205e-01
-6.67338595e-02 -1.05634145e-01 1.72279745e-01 4.61094111e-01
6.90516651e-01 -2.63431042e-01 -8.30323458e-01 1.26777530e+00
-1.91266418e-01 -5.27015090e-01 -3.74604344e-01 1.33537781e+00
9.84123826e-01 2.94527054e-01 2.77293772e-01 -5.18230915e-01
1.15554547e+00 -9.88086820e-01 -1.09173036e+00 -7.00433329e-02
5.90319335e-01 -1.21572661e+00 1.45675933e+00 2.43802503e-01
-1.52237082e+00 -6.85468197e-01 -7.07238019e-01 -1.46334261e-01
1.19139031e-01 3.19482446e-01 3.19966406e-01 9.80068326e-01
-1.59024692e+00 2.27727309e-01 -3.51036936e-01 -9.95038509e-01
1.96637630e-01 3.69082958e-01 -6.41765893e-01 -3.94518435e-01
-6.09639227e-01 9.36066806e-01 -1.01652304e-02 -7.13042244e-02
-3.85943353e-01 -5.10229111e-01 -9.71312344e-01 -1.44503936e-01
-1.00140676e-01 -9.70485926e-01 1.81867635e+00 -1.24764049e+00
-1.31902492e+00 9.13613498e-01 -5.37590146e-01 -1.63252838e-02
7.26977706e-01 -4.62450713e-01 -3.00100178e-01 6.18409514e-01
4.75898623e-01 6.96042180e-01 8.39338720e-01 -7.90461421e-01
-1.20100200e+00 -3.78088094e-02 3.94432396e-01 5.16335130e-01
-6.09827340e-01 3.37955445e-01 -5.17961800e-01 -3.45611989e-01
-1.45578146e-01 -5.80819309e-01 1.53984085e-01 4.80796307e-01
2.18695015e-01 -5.71899116e-03 7.77175784e-01 -1.37968671e+00
1.18736792e+00 -2.14327312e+00 -9.05253291e-02 -2.31594563e-01
3.72214347e-01 4.84594434e-01 2.52961129e-01 7.82953680e-01
1.81876317e-01 2.43791237e-01 4.34568197e-01 -1.60458311e-01
-7.03285709e-02 -1.83591276e-01 4.08803165e-01 3.77572864e-01
-2.00879991e-01 2.24760279e-01 -1.06788635e+00 -3.40966016e-01
5.26781499e-01 8.74805331e-01 -4.86870915e-01 5.31599820e-01
8.27481896e-02 3.19297314e-01 4.41901051e-02 7.01709211e-01
3.38965833e-01 -1.05331525e-01 -7.46771917e-02 -9.43225920e-02
-5.64084172e-01 2.78077841e-01 -7.91095376e-01 1.46483934e+00
-4.56979603e-01 1.28964436e+00 -9.64620560e-02 -6.77668676e-03
3.58355612e-01 6.90382540e-01 8.41426849e-02 -5.98260105e-01
-2.73994431e-02 2.02978998e-01 -1.14187904e-01 -1.62365103e+00
4.44295377e-01 1.21553630e-01 6.53962553e-01 3.16349596e-01
-1.75358802e-01 -3.11090574e-02 4.90828872e-01 1.87039688e-01
9.75959837e-01 2.75635242e-01 5.15439630e-01 3.15565109e-01
4.74187851e-01 7.61071593e-02 -2.51180023e-01 7.43815482e-01
-4.43237841e-01 7.99104095e-01 2.87220061e-01 -2.58276612e-01
-1.28203905e+00 -8.14687550e-01 2.29465485e-01 8.41773689e-01
-2.58279741e-01 -6.79527044e-01 -7.80440986e-01 -1.91409364e-01
-4.08579797e-01 6.20857894e-01 7.00503066e-02 3.27240288e-01
9.37689748e-03 2.98327446e-01 -3.06052029e-01 2.97927082e-01
7.05782771e-01 -9.80030179e-01 -9.51387346e-01 6.85865879e-02
-6.42971218e-01 -1.10675752e+00 -7.49049485e-01 -1.08646488e+00
-5.56256831e-01 -9.00017202e-01 -1.14789784e+00 -1.32188034e+00
9.81151581e-01 9.49920774e-01 5.81515670e-01 -2.52972115e-02
-4.21972513e-01 1.12727010e+00 -6.51178479e-01 6.41098320e-02
-2.36350447e-01 -3.37929547e-01 1.01174772e-01 -3.08250904e-01
3.58366519e-01 -2.95895815e-01 -8.10167611e-01 2.55567819e-01
-5.66692889e-01 2.40132436e-01 4.94821966e-01 4.79441971e-01
-5.95395900e-02 9.67640728e-02 3.07284445e-01 -1.84251055e-01
9.00517285e-01 -2.23878071e-01 -4.74041820e-01 -5.55819832e-02
-3.38488728e-01 -5.55818558e-01 2.12575272e-01 -5.78174472e-01
-9.99991834e-01 2.21504018e-01 -3.37449908e-02 -1.32296607e-01
-8.43098238e-02 3.51907551e-01 -3.33801061e-02 -5.46419442e-01
7.01412201e-01 -5.29913493e-02 3.24775755e-01 -1.22858532e-01
-5.69313243e-02 1.23192441e+00 5.39625227e-01 1.37301274e-02
5.18680274e-01 7.47143403e-02 -7.44292378e-01 -1.24742138e+00
-3.09896380e-01 -6.65143013e-01 -1.61077335e-01 -1.03754663e+00
7.43809402e-01 -1.10455811e+00 -8.99652600e-01 5.24579823e-01
-1.05951524e+00 -2.06358030e-01 2.23244987e-02 1.08795011e+00
-6.16602421e-01 4.65923041e-01 3.41507681e-02 -1.12893462e+00
8.95210207e-02 -1.26324177e+00 6.72198176e-01 8.47320557e-01
-4.05514240e-01 -5.19629955e-01 -4.14238214e-01 9.26988065e-01
6.23948097e-01 3.98245780e-03 5.40156126e-01 8.13062787e-02
-6.80086017e-01 -6.45960927e-01 -4.93049860e-01 3.41853529e-01
3.75936240e-01 4.34282646e-02 -7.24336147e-01 -4.74580705e-01
-3.25348228e-01 -4.32095319e-01 1.82917997e-01 5.49803138e-01
6.62690520e-01 -7.07885742e-01 5.60661182e-02 1.64956674e-01
1.40550709e+00 2.60177583e-01 9.07467723e-01 5.92509806e-01
2.59641320e-01 8.46604586e-01 5.77995658e-01 5.38483322e-01
3.85001540e-01 6.91849887e-01 2.68416047e-01 1.16101295e-01
-4.05716062e-01 -3.81741017e-01 5.35588562e-01 2.42381215e-01
-6.03553593e-01 -2.09399149e-01 -7.49239922e-01 8.12026322e-01
-1.59430134e+00 -1.19961894e+00 -3.48761410e-01 2.72186661e+00
3.39985311e-01 -1.28381312e-01 5.20264745e-01 6.87913299e-02
9.86860156e-01 -6.40475631e-01 -3.18848670e-01 -1.32104708e-02
3.28775309e-02 -3.80465649e-02 5.49579740e-01 6.37375891e-01
-6.06398582e-01 6.25122070e-01 7.08340454e+00 -1.55952394e-01
-7.83385456e-01 4.53007631e-02 -2.54424810e-02 -5.19831419e-01
-1.31072998e-01 -2.25921143e-02 -3.55664670e-01 3.94822091e-01
8.06812942e-01 -4.06788796e-01 5.18060982e-01 6.49325788e-01
1.00255263e+00 -4.03880805e-01 -7.94619799e-01 1.14453828e+00
3.53129774e-01 -1.18470311e+00 -1.64168235e-02 3.85515168e-02
3.68172228e-01 -2.77800083e-01 9.09725577e-03 -6.69766068e-02
-5.63115299e-01 -9.02186692e-01 7.57568359e-01 6.46912932e-01
1.17717886e+00 -4.40066159e-01 7.40427852e-01 -1.06996417e-01
-8.77216339e-01 -1.88217223e-01 -1.69342935e-01 -7.12647319e-01
3.69299442e-01 -3.89287949e-01 -1.19847941e+00 -1.28063813e-01
8.89938831e-01 5.17899156e-01 -7.04096377e-01 1.82725680e+00
-2.28420854e-01 2.39934638e-01 1.39857784e-01 -1.78015828e-01
-2.40067795e-01 -4.93169613e-02 6.46831214e-01 1.02354634e+00
6.60347104e-01 6.03105947e-02 -3.19007039e-01 1.62888393e-01
4.77255732e-01 3.68517488e-01 -7.18993366e-01 -3.96647573e-01
5.08735716e-01 7.41357505e-01 -3.49893570e-01 -8.14955533e-02
-1.05941868e+00 1.00290275e+00 -2.38501355e-01 5.56301236e-01
-5.04424334e-01 -7.77586102e-01 5.41329443e-01 9.67444897e-01
4.09583785e-02 -4.46103990e-01 2.74468571e-01 -9.06668603e-01
5.38325667e-01 -9.97298121e-01 1.31827176e-01 -1.49598169e+00
-6.83718383e-01 5.60063243e-01 1.47205191e-02 -1.59522402e+00
-3.72746646e-01 -5.54754794e-01 -1.20154656e-01 1.12263906e+00
-1.00308251e+00 -9.23165321e-01 -1.08698857e+00 6.86587036e-01
8.28795552e-01 -2.05355763e-01 7.84291863e-01 4.98945177e-01
-1.50709555e-01 4.76313710e-01 -2.25718498e-01 -3.75150442e-01
1.09587717e+00 -7.63724029e-01 -6.52194619e-02 9.63204741e-01
-6.22232854e-01 6.19644880e-01 9.76089537e-01 -6.94462836e-01
-1.23571301e+00 -5.04447460e-01 1.18203545e+00 -1.16422929e-01
4.31878448e-01 1.30349964e-01 -4.27487731e-01 5.79561353e-01
3.92454118e-01 -5.87677479e-01 9.29510474e-01 -3.92238706e-01
-6.63644299e-02 9.42547768e-02 -1.31059027e+00 9.00622070e-01
1.31527293e+00 -7.51897812e-01 -5.09306312e-01 6.34263992e-01
7.30954051e-01 -2.37983882e-01 -6.51528955e-01 -1.28230453e-01
1.23106682e+00 -1.36363828e+00 5.23161709e-01 -2.19986573e-01
4.41530406e-01 -3.24106842e-01 -1.79336444e-01 -8.31117094e-01
-2.70646363e-01 -1.00700009e+00 3.61420959e-01 1.08491743e+00
3.01702291e-01 -7.07740664e-01 4.10702914e-01 1.20978296e+00
-2.15763658e-01 1.00563191e-01 -5.45749485e-01 -3.80010337e-01
-9.90541399e-01 -2.36484334e-01 -7.14175031e-02 2.51901388e-01
4.14715558e-01 -1.38751447e-01 -5.29297590e-01 2.11298689e-01
3.63561541e-01 -8.38477612e-01 7.98696876e-01 -9.81618404e-01
-1.83250815e-01 -2.77163446e-01 -1.00775099e+00 -5.98329544e-01
-2.90794522e-01 -2.58184910e-01 -3.32594454e-01 -2.25675249e+00
2.37668321e-01 5.86432099e-01 7.08083749e-01 1.91043615e-01
1.21777855e-01 2.27007583e-01 7.60242641e-02 5.10821044e-01
-2.00506449e-01 1.98880048e-03 1.24020457e+00 3.12870473e-01
-6.71756744e-01 2.08767563e-01 -9.55830872e-01 3.34763169e-01
6.52201116e-01 1.07397966e-01 -8.45708907e-01 -4.12506402e-01
4.02871579e-01 1.39761299e-01 5.07330835e-01 -1.18692529e+00
2.64659315e-01 -2.70863567e-02 3.57291341e-01 -1.83545366e-01
3.23007643e-01 -6.61080539e-01 4.28066850e-01 4.61139202e-01
4.77078669e-02 -1.09307468e-02 3.61520976e-01 2.56956458e-01
-1.82825595e-01 -3.21812809e-01 5.85045040e-01 -3.14419925e-01
-6.88454509e-01 -9.85227451e-02 -1.07378459e+00 -2.68422753e-01
1.06104410e+00 -7.72602260e-01 -6.63555980e-01 -1.27225673e+00
-8.73656511e-01 -5.35113830e-03 8.81863892e-01 4.20299768e-01
9.73891735e-01 -1.02973652e+00 -5.40427923e-01 2.97743112e-01
3.31252486e-01 -2.84672350e-01 4.88761067e-01 6.50703132e-01
-1.14435101e+00 5.06463170e-01 -6.70769215e-01 -2.24758863e-01
-1.62738621e+00 1.07789904e-01 1.20828822e-01 9.72248375e-01
-8.68889391e-01 5.85305572e-01 -2.25381300e-01 5.56616902e-01
7.26477802e-01 1.41248122e-01 -2.51012117e-01 8.50576535e-02
1.04895294e+00 5.93088627e-01 -2.04012305e-01 -5.67514241e-01
-3.58280577e-02 4.62360620e-01 -2.42826894e-01 -5.27461112e-01
9.86101866e-01 -7.04306543e-01 -1.13351336e-02 1.69258326e-01
6.93469703e-01 1.00368895e-01 -1.06974828e+00 3.46992970e-01
-6.90834463e-01 -1.24424899e+00 -2.68580206e-03 -7.74096847e-01
-4.31860745e-01 4.01781112e-01 7.12799489e-01 1.00038826e-01
1.41478801e+00 -9.47406963e-02 4.59394962e-01 2.54854083e-01
2.53897130e-01 -9.11825359e-01 1.38923302e-01 -3.43334079e-01
9.92380321e-01 -1.12736988e+00 -2.23058000e-01 -3.79072428e-01
-6.67639196e-01 1.20174813e+00 8.05295110e-01 5.28271616e-01
1.77665770e-01 -4.64841813e-01 1.37029588e-01 7.27988109e-02
-3.73545915e-01 -5.40028691e-01 2.00281367e-01 1.28895593e+00
6.94264829e-01 -2.06648663e-01 -7.03570008e-01 1.91960588e-01
-3.71178776e-01 4.94867325e-01 1.20881832e+00 9.22159970e-01
-7.50999629e-01 -7.87948370e-01 -4.82460320e-01 5.34156263e-01
-2.13147238e-01 6.48322850e-02 -2.67983854e-01 8.28543305e-01
8.95301718e-03 1.46525300e+00 -1.10341594e-01 -2.94675022e-01
4.27313179e-01 -5.31993546e-02 6.02463245e-01 -6.27448499e-01
-1.35387480e-01 3.11968243e-03 5.58487356e-01 -4.47292984e-01
-3.98579270e-01 -8.67485642e-01 -4.75149989e-01 -3.38808566e-01
5.11463098e-02 -1.27982795e-01 8.56994271e-01 6.49949253e-01
3.35082114e-01 1.58526227e-01 1.59672379e-01 -7.93813527e-01
4.41356376e-03 -1.13253760e+00 -3.22525918e-01 6.28742129e-02
7.19939351e-01 -4.77505028e-01 -6.21633708e-01 4.30864096e-01] | [10.964506149291992, 0.761188805103302] |
8fb767e6-463a-402d-9d5a-989a9d0abf49 | pr-net-preference-reasoning-for-personalized | 2109.01799 | null | https://arxiv.org/abs/2109.01799v1 | https://arxiv.org/pdf/2109.01799v1.pdf | PR-Net: Preference Reasoning for Personalized Video Highlight Detection | Personalized video highlight detection aims to shorten a long video to interesting moments according to a user's preference, which has recently raised the community's attention. Current methods regard the user's history as holistic information to predict the user's preference but negating the inherent diversity of the user's interests, resulting in vague preference representation. In this paper, we propose a simple yet efficient preference reasoning framework (PR-Net) to explicitly take the diverse interests into account for frame-level highlight prediction. Specifically, distinct user-specific preferences for each input query frame are produced, presented as the similarity weighted sum of history highlights to the corresponding query frame. Next, distinct comprehensive preferences are formed by the user-specific preferences and a learnable generic preference for more overall highlight measurement. Lastly, the degree of highlight and non-highlight for each query frame is calculated as semantic similarity to its comprehensive and non-highlight preferences, respectively. Besides, to alleviate the ambiguity due to the incomplete annotation, a new bi-directional contrastive loss is proposed to ensure a compact and differentiable metric space. In this way, our method significantly outperforms state-of-the-art methods with a relative improvement of 12% in mean accuracy precision. | ['Wenping Wang', 'Xing Sun', 'Pai Peng', 'Nenglun Chen', 'Wenzhe Wang', 'Penghao Zhou', 'Runnan Chen'] | 2021-09-04 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Chen_PR-Net_Preference_Reasoning_for_Personalized_Video_Highlight_Detection_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Chen_PR-Net_Preference_Reasoning_for_Personalized_Video_Highlight_Detection_ICCV_2021_paper.pdf | iccv-2021-1 | ['highlight-detection'] | ['computer-vision'] | [ 8.63325670e-02 -2.16735020e-01 -4.84986573e-01 -5.43327034e-01
-8.80498350e-01 -3.63962919e-01 4.02094364e-01 2.74335712e-01
-3.39760154e-01 4.50818896e-01 6.59828365e-01 4.31422234e-01
-2.64635831e-01 -5.35480797e-01 -2.88432747e-01 -7.20086157e-01
3.76027897e-02 -1.14724934e-01 5.16191781e-01 -1.23412035e-01
4.64834213e-01 2.95265228e-01 -1.79641747e+00 4.88587499e-01
9.02757823e-01 1.31998158e+00 3.75381917e-01 1.28303871e-01
1.75903037e-01 8.73058558e-01 -5.40333152e-01 -4.41058755e-01
7.69872814e-02 -1.65740028e-01 -4.92998630e-01 1.66868731e-01
5.98053396e-01 -3.94595742e-01 -3.35080981e-01 1.02299678e+00
5.56084514e-01 5.02938092e-01 3.51404667e-01 -1.15709925e+00
-4.73838806e-01 4.09131348e-01 -6.09592795e-01 3.02305967e-01
4.24033910e-01 -1.37671521e-02 1.40078247e+00 -9.40146565e-01
7.82347262e-01 1.06764448e+00 1.67008489e-01 2.60100573e-01
-9.66507375e-01 -4.40189958e-01 6.67194545e-01 6.44615531e-01
-1.40964437e+00 -2.67139882e-01 1.36280322e+00 -4.35938627e-01
3.64058763e-01 7.94362903e-01 7.14748204e-01 6.78905010e-01
-2.82006562e-01 1.04059267e+00 5.08284628e-01 3.24605517e-02
1.69924185e-01 1.13111466e-01 -2.33413726e-01 5.24508119e-01
-2.64870405e-01 -3.24258775e-01 -8.34911525e-01 -3.10338866e-02
4.86823201e-01 1.90228000e-01 -5.30431747e-01 -5.47553420e-01
-1.38645160e+00 5.72898567e-01 2.55878299e-01 1.49443239e-01
-4.17900294e-01 -2.35479340e-01 5.97556174e-01 -2.97640506e-02
6.16125226e-01 3.90825033e-01 -4.55577165e-01 -2.21667618e-01
-1.16079426e+00 2.70403385e-01 3.23551625e-01 1.06718886e+00
8.65819395e-01 -2.76681483e-01 -8.67057502e-01 9.06215727e-01
2.82710582e-01 5.34812398e-02 2.22192779e-01 -1.26975977e+00
4.39958751e-01 7.84579635e-01 3.79321337e-01 -1.57387006e+00
-2.63458043e-02 -6.76873565e-01 -5.73284090e-01 -1.74624190e-01
7.00810030e-02 1.34822667e-01 -1.89428076e-01 1.90444672e+00
2.76272595e-01 2.97564179e-01 -2.65348613e-01 1.40163267e+00
7.72686779e-01 7.29226410e-01 1.19452953e-01 -6.77602053e-01
1.37888730e+00 -1.06185710e+00 -4.63120401e-01 3.00618224e-02
2.84487426e-01 -8.04980695e-01 1.29153585e+00 4.03776199e-01
-1.10755265e+00 -5.22161663e-01 -8.40860069e-01 -1.80517524e-01
-7.09027499e-02 3.96266878e-01 4.37781185e-01 3.51709843e-01
-7.66637921e-01 4.14901704e-01 -1.39946386e-01 -2.29939044e-01
3.47963959e-01 1.18187010e-01 2.55624913e-02 -1.51266698e-02
-1.32629585e+00 5.64729512e-01 3.50614607e-01 -1.71213195e-01
-5.03065050e-01 -1.07109189e+00 -5.33665895e-01 2.01199576e-01
7.67725945e-01 -5.91842830e-01 1.05297065e+00 -9.47581351e-01
-1.30419850e+00 8.40518236e-01 -4.38546807e-01 -1.20728694e-01
7.57132292e-01 -2.75416374e-01 -6.05978549e-01 4.19734538e-01
1.27632573e-01 6.92662179e-01 6.76540375e-01 -1.30430794e+00
-1.20735598e+00 -9.20986384e-02 5.93091071e-01 5.43672144e-01
-7.08381414e-01 6.23757206e-02 -1.10977674e+00 -1.01429832e+00
1.53120995e-01 -4.90843058e-01 6.59611076e-02 2.41588742e-01
-5.36703803e-02 -3.66762668e-01 1.00056958e+00 -4.73194391e-01
1.91645026e+00 -2.32453942e+00 2.24953011e-01 9.81286392e-02
2.42097989e-01 -5.59609830e-02 -2.37493068e-02 2.65264601e-01
3.11960071e-01 -1.49249300e-01 -1.51431441e-01 -2.35195085e-01
1.41408533e-01 1.16475560e-01 -2.58422852e-01 2.55680025e-01
9.28234085e-02 4.12930101e-01 -1.37452924e+00 -8.15271318e-01
2.53266305e-01 5.09038329e-01 -7.70574212e-01 2.49062598e-01
-2.91172057e-01 3.68254870e-01 -6.78027391e-01 7.64813185e-01
6.89663053e-01 -1.50133163e-01 1.29285410e-01 -8.48787069e-01
-3.28335166e-01 -5.37732057e-02 -1.01819026e+00 1.81832457e+00
-3.38496864e-01 5.87600827e-01 -2.29478315e-01 -7.61192501e-01
1.00665927e+00 1.57682881e-01 8.37588370e-01 -6.39392078e-01
-1.35811806e-01 8.99385214e-02 -4.80340689e-01 -6.30675673e-01
9.11330819e-01 1.78974345e-01 -4.28068228e-02 2.07658634e-01
-4.57183599e-01 2.66570836e-01 4.04337972e-01 2.04305366e-01
6.65980935e-01 2.80236810e-01 1.72130406e-01 -3.09066892e-01
9.34156418e-01 -3.75776529e-01 9.64105844e-01 3.96026582e-01
-4.19500321e-01 8.92189324e-01 4.39669490e-01 -4.15942013e-01
-6.36075974e-01 -9.11156654e-01 2.97781378e-01 1.54514468e+00
7.99783409e-01 -5.96965432e-01 -5.66206574e-01 -7.61275172e-01
-1.85596004e-01 7.02653944e-01 -4.44235176e-01 -1.88992307e-01
-6.30611718e-01 -3.51199806e-01 3.65281664e-02 2.93141097e-01
5.92873037e-01 -1.07859707e+00 -8.91373694e-01 1.19697355e-01
-7.10955799e-01 -8.68580520e-01 -1.16613722e+00 -4.96942073e-01
-5.88371575e-01 -1.02493072e+00 -1.01574993e+00 -6.88562334e-01
6.47553980e-01 5.73933482e-01 1.12042844e+00 5.98825850e-02
4.82527167e-03 4.12268639e-01 -5.26445329e-01 2.12222517e-01
3.68770719e-01 -1.10192753e-01 -6.72984272e-02 4.21872437e-01
4.35360283e-01 -3.85505468e-01 -1.19377112e+00 4.60808456e-01
-9.62162733e-01 2.65558481e-01 3.16033185e-01 6.65118039e-01
8.72473240e-01 1.97478086e-01 2.62037218e-01 -6.20278418e-01
5.29720545e-01 -4.81431663e-01 -1.23548456e-01 5.03206074e-01
-5.12134254e-01 -1.91029996e-01 5.03759742e-01 -5.31501651e-01
-1.36095548e+00 -1.62495732e-01 2.04450518e-01 -5.94930351e-01
1.62062585e-01 5.34386814e-01 -4.89682108e-01 1.92111835e-01
2.31835499e-01 1.86617464e-01 -5.22532821e-01 -2.45042279e-01
4.22809809e-01 3.37250978e-01 6.39980435e-01 -8.46088648e-01
3.13703150e-01 4.13466424e-01 -1.41418397e-01 -4.19034868e-01
-1.13210058e+00 -7.00162709e-01 -3.87092024e-01 -6.15168452e-01
6.28362477e-01 -8.98289740e-01 -6.67741299e-01 1.82563230e-01
-1.13392079e+00 1.80280060e-01 -2.03316689e-01 4.48719472e-01
-6.34221137e-01 7.32862294e-01 -3.55822414e-01 -7.46739089e-01
-1.81118757e-01 -1.22619057e+00 9.63732541e-01 4.87937212e-01
-3.76783699e-01 -6.82513595e-01 -3.97590429e-01 3.12102020e-01
2.52020121e-01 2.30382994e-01 6.84937775e-01 -2.30665296e-01
-7.91670322e-01 -3.75179835e-02 -6.54605329e-01 1.51535988e-01
2.07743555e-01 2.01948658e-01 -7.65428543e-01 -1.74020857e-01
-1.69508625e-02 1.87379748e-01 7.35714078e-01 3.83344561e-01
1.43614578e+00 -5.05611539e-01 -1.44363359e-01 3.80131721e-01
1.27752995e+00 3.22114468e-01 6.00846767e-01 3.93761635e-01
6.77731156e-01 7.85187066e-01 1.27279663e+00 9.76905107e-01
5.71431994e-01 8.21735740e-01 5.00682354e-01 1.89900145e-01
9.77817401e-02 -2.16669455e-01 1.74908131e-01 5.07372439e-01
-2.21394882e-01 -3.80389482e-01 -2.20892563e-01 4.89967555e-01
-2.20534110e+00 -1.33344114e+00 3.05129178e-02 2.39474320e+00
6.97191834e-01 1.61617339e-01 1.41076148e-01 2.93023530e-02
9.84003723e-01 6.50488615e-01 -6.03590488e-01 -1.83728442e-03
-2.22800449e-01 -4.30825621e-01 1.80714831e-01 3.34496260e-01
-1.07565963e+00 7.22207069e-01 5.08254719e+00 1.26698160e+00
-1.19215548e+00 5.47484048e-02 8.53710592e-01 -3.48307073e-01
-7.28268504e-01 -1.71742849e-02 -4.02776808e-01 8.73677492e-01
8.19253847e-02 -4.22263265e-01 3.16679299e-01 6.93349183e-01
4.99889523e-01 -1.66646704e-01 -9.73032713e-01 1.23289192e+00
2.25026891e-01 -1.30230570e+00 2.43277311e-01 -2.07042992e-01
5.90496063e-01 -6.50612831e-01 3.54809195e-01 1.42740101e-01
-5.13003945e-01 -2.85976648e-01 1.03429413e+00 9.06146288e-01
7.58348763e-01 -1.00317478e+00 3.23616683e-01 1.40296042e-01
-1.51814795e+00 -1.95682660e-01 -2.49204129e-01 5.06299222e-03
3.49894911e-01 7.86072135e-01 -1.98515728e-01 5.63790977e-01
7.03277826e-01 9.92004871e-01 -2.53008813e-01 1.05312204e+00
-5.12167774e-02 2.83037156e-01 -1.30604357e-01 1.22579865e-01
1.80203781e-01 -3.15203220e-01 7.03682244e-01 1.41937327e+00
6.67448401e-01 3.43270242e-01 2.17283770e-01 5.84688902e-01
-8.97890031e-02 3.31863880e-01 1.73442543e-01 2.08281606e-01
6.37593687e-01 1.34259927e+00 -5.58828294e-01 -5.31869888e-01
-4.24948603e-01 1.10822499e+00 9.74424835e-03 4.17057008e-01
-1.00781393e+00 -2.34925702e-01 7.46352971e-01 2.67560780e-02
2.66677082e-01 2.36653537e-01 -1.40735984e-01 -1.17617846e+00
2.81880438e-01 -3.51199865e-01 6.47518039e-01 -8.30227733e-01
-1.24224257e+00 5.67547798e-01 -1.23166971e-01 -1.86169660e+00
1.07229941e-01 -1.93695351e-01 -7.36978531e-01 8.04142535e-01
-1.55129206e+00 -1.06885529e+00 -5.63259840e-01 4.52926010e-01
8.33051562e-01 -1.81230187e-01 4.52493191e-01 5.70621192e-01
-4.41457719e-01 6.74433053e-01 -2.90713787e-01 -3.87134075e-01
9.83669043e-01 -9.70066249e-01 -2.73934871e-01 8.16796184e-01
-2.97459424e-01 5.37619770e-01 1.06435859e+00 -4.25412506e-01
-1.04462636e+00 -1.07268250e+00 1.24498916e+00 -5.63632930e-03
4.40760493e-01 3.35297525e-01 -8.12152386e-01 1.81602091e-02
-1.23167969e-01 9.82282385e-02 6.89540327e-01 -1.11362040e-01
-4.73565161e-01 -5.03499031e-01 -1.25085127e+00 8.86078894e-01
1.19666433e+00 -4.99491721e-01 -3.88962835e-01 -2.52697859e-02
7.33271182e-01 -3.12060505e-01 -8.64314020e-01 5.98384678e-01
8.95091534e-01 -1.43218815e+00 1.07271111e+00 -3.75986360e-02
5.52030563e-01 -6.93455637e-01 -4.24676239e-01 -8.83304358e-01
-5.48133194e-01 -5.93449652e-01 -4.03537810e-01 1.30043256e+00
2.53213234e-02 -8.77566040e-02 8.38617027e-01 7.41245806e-01
-2.62452602e-01 -1.04393589e+00 -6.37269676e-01 -3.14901680e-01
-7.59171963e-01 -2.91915327e-01 7.97422707e-01 9.70084488e-01
1.44501686e-01 -6.21970184e-02 -7.31738269e-01 1.12939015e-01
5.04309654e-01 4.25427407e-01 3.82535666e-01 -1.25226057e+00
-9.95377749e-02 -7.98413873e-01 -2.28990406e-01 -1.41360378e+00
-8.32009092e-02 -4.05213714e-01 -9.84440371e-02 -1.34824288e+00
5.44476569e-01 -3.24935853e-01 -8.77442658e-01 -3.03875841e-02
-2.46763214e-01 3.71009082e-01 2.63430029e-01 3.00449610e-01
-1.17401052e+00 5.23307145e-01 1.44233346e+00 -2.67166018e-01
-2.76258469e-01 3.92950960e-02 -1.02222812e+00 6.14082873e-01
5.83400369e-01 -9.67204720e-02 -7.51197219e-01 -3.66481692e-01
3.42593968e-01 1.71931982e-01 1.29397944e-01 -8.69928300e-01
3.18475693e-01 -5.66985667e-01 2.16457501e-01 -7.98858523e-01
5.49805403e-01 -7.48806417e-01 8.64772573e-02 -2.27846578e-02
-5.27747273e-01 -1.46882668e-01 -2.52280116e-01 5.86345911e-01
-5.23060024e-01 1.79766007e-02 5.36146879e-01 -1.95892733e-02
-1.20689654e+00 6.59362018e-01 7.49724060e-02 -5.08677363e-02
1.03804266e+00 -5.01056314e-01 -1.48693532e-01 -4.26939249e-01
-6.43923104e-01 4.14526761e-01 6.36765480e-01 4.95069504e-01
8.01508904e-01 -1.53162730e+00 -5.80457330e-01 -1.43105224e-01
4.91185784e-01 -2.31764764e-01 8.86864305e-01 8.97188723e-01
-9.23623741e-02 1.35494366e-01 -1.47212997e-01 -3.70932043e-01
-1.41415405e+00 6.17774606e-01 -1.03378899e-01 -1.26932800e-01
-4.24112678e-01 7.91425526e-01 6.03551447e-01 1.83829784e-01
3.32226485e-01 -2.46606961e-01 -5.30625343e-01 5.23050904e-01
5.84677756e-01 6.55028164e-01 -3.38151932e-01 -9.54571187e-01
-4.01956856e-01 6.29006207e-01 -2.59844046e-02 2.28688538e-01
1.02288103e+00 -6.48030341e-01 5.60569055e-02 3.28350455e-01
1.22646260e+00 3.44526321e-01 -1.67822564e+00 -3.96742046e-01
-5.93250319e-02 -8.90420616e-01 -6.39337748e-02 -6.74096107e-01
-1.24435723e+00 6.30354464e-01 4.71091270e-01 -3.10832472e-03
1.59572113e+00 3.04563041e-03 8.78801286e-01 -2.49711927e-02
2.34691918e-01 -1.29451990e+00 1.86284482e-01 2.28800416e-01
9.05452371e-01 -1.15562081e+00 4.10123095e-02 -5.46432614e-01
-1.05306995e+00 1.02508283e+00 6.97831571e-01 8.20573866e-02
4.17759925e-01 -3.85415941e-01 -4.43102643e-02 -3.41547118e-03
-5.31092465e-01 -1.05146661e-01 7.96973646e-01 3.32709700e-01
5.79854250e-01 9.45998430e-02 -5.56756556e-01 8.25814664e-01
1.13013715e-01 -1.88387576e-02 2.03072548e-01 5.04271746e-01
-5.89595735e-01 -7.98985839e-01 -6.81482106e-02 3.78049940e-01
-3.17868799e-01 6.16790215e-03 6.08172826e-02 2.80014217e-01
4.19341922e-01 9.38540459e-01 1.01133280e-01 -4.48843956e-01
3.17773432e-01 -3.55416447e-01 2.16531232e-01 -3.37881535e-01
-2.20011294e-01 3.25595886e-01 -7.33492151e-02 -6.45812571e-01
-7.27015913e-01 -5.52771747e-01 -1.22862315e+00 -2.12755144e-01
7.12366402e-02 2.06605703e-01 -2.67515033e-02 7.16839254e-01
3.93340707e-01 4.01751161e-01 7.86740720e-01 -9.27736700e-01
-5.55817299e-02 -5.09862185e-01 -6.37441337e-01 6.61190510e-01
2.73688555e-01 -7.37069070e-01 -2.54229337e-01 1.78851634e-01] | [10.013283729553223, 0.4398844838142395] |
cb60b793-5f45-4c87-9fa0-e16b0cb8dd56 | degree-of-linear-polarization-based-color | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Ono_Degree-of-Linear-Polarization-Based_Color_Constancy_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Ono_Degree-of-Linear-Polarization-Based_Color_Constancy_CVPR_2022_paper.pdf | Degree-of-Linear-Polarization-Based Color Constancy | Color constancy is an essential function in digital photography and a fundamental process for many computer vision applications. Accordingly, many methods have been proposed, and some recent ones have used deep neural networks to handle more complex scenarios. However, both the traditional and latest methods still impose strict assumptions on their target scenes in explicit or implicit ways. This paper shows that the degree of linear polarization dramatically solves the color constancy problem because it allows us to find achromatic pixels stably. Because we only rely on the physics-based polarization model, we significantly reduce the assumptions compared to existing methods. Furthermore, we captured a wide variety of scenes with ground-truth illuminations for evaluation, and the proposed approach achieved state-of-the-art performance with a low computational cost. Additionally, the proposed method can estimate illumination colors from chromatic pixels and manage multi-illumination scenes. Lastly, the evaluation scenes and codes are publicly available to encourage more development in this field. | ['Yusuke Moriuchi', 'Teppei Kurita', 'Legong Sun', 'Yuhi Kondo', 'Taishi Ono'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['color-constancy'] | ['computer-vision'] | [ 1.24920912e-01 -7.36961722e-01 8.37618113e-02 -3.18087488e-01
-1.07311860e-01 -4.19396758e-01 1.90995619e-01 -5.82875609e-01
-2.95118511e-01 8.02196801e-01 -3.87647092e-01 -1.83480352e-01
1.29708856e-01 -9.44135964e-01 -5.72335184e-01 -1.15628505e+00
3.28135580e-01 -2.39757359e-01 3.14678073e-01 -1.10285930e-01
4.34949338e-01 5.02228379e-01 -1.86244130e+00 -8.71045068e-02
1.20734107e+00 1.07927752e+00 2.62259632e-01 4.07576561e-01
-1.88205376e-01 5.11686862e-01 -3.81944895e-01 -3.41571361e-01
5.58229625e-01 -4.53456938e-01 -2.54816204e-01 2.12454304e-01
6.80518091e-01 -5.42860925e-01 -3.30452651e-01 1.52132678e+00
4.78115380e-01 9.56517160e-02 3.63741636e-01 -1.22768641e+00
-9.74351466e-01 -2.47289687e-01 -9.54515278e-01 -2.73833632e-01
-7.17943013e-02 3.21097940e-01 6.64850295e-01 -6.40640438e-01
9.55508277e-02 8.66321981e-01 5.72093844e-01 2.76885957e-01
-9.36570346e-01 -9.46531057e-01 4.15625721e-02 4.68896210e-01
-1.26115215e+00 -3.72022688e-01 1.00334275e+00 -1.84975430e-01
4.63190764e-01 2.07486570e-01 7.36567140e-01 6.29045963e-01
-2.10928288e-03 4.12701339e-01 1.88849640e+00 -4.76543009e-01
-4.01576012e-02 1.79220662e-01 8.01741630e-02 7.97687292e-01
4.98303950e-01 2.51196384e-01 -1.83115155e-01 1.70940712e-01
9.92262185e-01 3.53502445e-02 -5.10439754e-01 -2.95308530e-01
-9.06972408e-01 4.35213804e-01 5.72554469e-01 -1.42833609e-02
-1.61089599e-01 4.18199934e-02 1.06824003e-03 -1.68909088e-01
2.59113818e-01 3.44989121e-01 -1.85321867e-01 7.17359185e-02
-5.53692698e-01 2.53286809e-02 4.04958397e-01 7.99471855e-01
1.00289714e+00 1.90506935e-01 1.23540170e-01 9.47882771e-01
2.85655409e-01 9.28410232e-01 5.84322251e-02 -1.34265685e+00
4.24043760e-02 4.05032486e-01 3.50237727e-01 -1.27227914e+00
-3.01033199e-01 -4.04741883e-01 -1.14209890e+00 6.08671486e-01
3.43045712e-01 -2.89115906e-02 -9.71610188e-01 1.63145220e+00
2.90566981e-01 6.16751164e-02 -7.00531006e-02 1.20844626e+00
5.91111124e-01 8.48800421e-01 -2.22161412e-01 -4.91363525e-01
1.13387227e+00 -9.73283231e-01 -9.16758418e-01 1.07407048e-02
-1.39029756e-01 -1.04573774e+00 1.21764898e+00 8.16120982e-01
-9.60288942e-01 -6.31815910e-01 -1.08115852e+00 -1.51713163e-01
-1.47180706e-01 3.45754147e-01 1.01590919e+00 9.69659388e-01
-1.13515329e+00 3.96481007e-01 -4.96957392e-01 -1.40316114e-01
3.07906955e-01 8.80036056e-02 5.38808517e-02 -3.15385818e-01
-1.07141829e+00 6.75794661e-01 1.32925227e-01 5.29123247e-01
-4.92328972e-01 -2.28879839e-01 -3.95515054e-01 -8.38768333e-02
4.02184933e-01 -4.81345534e-01 9.58833694e-01 -1.18562448e+00
-1.89375412e+00 8.07576299e-01 -2.54508823e-01 1.82906389e-02
4.37769622e-01 -1.46948293e-01 -6.61670804e-01 1.63701236e-01
-2.90641397e-01 2.46718407e-01 7.01929808e-01 -1.51625371e+00
-5.03043294e-01 -1.26795903e-01 2.49991298e-01 1.62191540e-01
-4.17255849e-01 1.51154082e-02 -8.13104033e-01 -1.39693901e-01
3.24243337e-01 -8.30898285e-01 8.21352378e-02 3.98806781e-01
-4.44259763e-01 1.54805750e-01 7.77815998e-01 -4.31875616e-01
8.96385372e-01 -2.12546873e+00 -2.63258964e-01 -6.33991137e-02
4.25059088e-02 4.94198054e-01 -5.63336611e-02 2.15849712e-01
6.69358820e-02 -1.27087042e-01 -2.96470672e-01 -9.00029466e-02
-1.64441973e-01 9.84946638e-02 -1.75120488e-01 5.98626912e-01
-1.69733211e-01 3.80051523e-01 -6.26278281e-01 -5.16674340e-01
5.08600950e-01 5.86690426e-01 -4.79760349e-01 1.31032482e-01
-2.46634200e-01 4.01820809e-01 -1.58727795e-01 7.45925903e-01
1.42581832e+00 -1.06131606e-01 1.37744606e-01 -6.44591987e-01
-6.52729928e-01 -1.81820288e-01 -1.30934978e+00 1.45510530e+00
-4.00684804e-01 7.69270480e-01 1.52694911e-01 -7.69561768e-01
9.35194254e-01 -1.47194594e-01 4.31693405e-01 -9.77176189e-01
1.91403374e-01 2.57351458e-01 -8.69917125e-02 -6.27007067e-01
6.22404814e-01 -1.48278162e-01 5.14021575e-01 1.04868688e-01
-5.37885666e-01 -1.13669321e-01 8.62414688e-02 -2.21586421e-01
2.64481544e-01 3.65013570e-01 1.06954485e-01 -1.01260923e-01
4.87369329e-01 -6.73832465e-03 8.94894242e-01 5.48449874e-01
-3.08382720e-01 6.45764589e-01 2.19404295e-01 -4.74995792e-01
-8.55754197e-01 -1.02059853e+00 -3.32180470e-01 5.60302436e-01
8.49707961e-01 8.59610066e-02 -4.92678344e-01 1.51772052e-02
-2.92308152e-01 6.69828713e-01 -2.62981951e-01 1.66519105e-01
-4.06115115e-01 -9.71842587e-01 2.20971152e-01 1.74953714e-01
1.39521360e+00 -7.39374995e-01 -5.46600103e-01 -2.09134459e-01
-2.88898379e-01 -1.17462611e+00 -8.14328566e-02 -2.89065808e-01
-7.43541300e-01 -1.36778033e+00 -6.24617875e-01 -7.16471791e-01
7.35459089e-01 8.55376065e-01 8.48747790e-01 1.91737667e-01
-4.08340901e-01 1.28579140e-01 -2.29424581e-01 -3.41761291e-01
-7.48377200e-03 -4.80588347e-01 -8.82720575e-02 2.92790681e-01
2.74937689e-01 -3.76038164e-01 -8.59210432e-01 4.57679987e-01
-8.66654456e-01 5.29807150e-01 7.28098452e-01 7.37273037e-01
6.68273628e-01 4.93938506e-01 6.74353838e-02 -8.45504344e-01
2.23125592e-01 6.76789284e-02 -1.16439939e+00 2.70967185e-01
-8.51338267e-01 -1.92587167e-01 7.26713181e-01 -1.16269156e-01
-1.64901578e+00 -9.49585587e-02 2.05074772e-02 -3.27659220e-01
-2.99919933e-01 6.47689700e-02 -3.72768849e-01 -4.79064137e-01
3.36767614e-01 3.77588868e-01 -1.22994706e-01 -4.35010493e-01
2.89617568e-01 6.51168585e-01 6.84454501e-01 -5.97144663e-01
7.84560382e-01 7.51737356e-01 3.37629706e-01 -9.47549224e-01
-7.69400716e-01 -1.62958294e-01 -3.50386024e-01 -4.69356209e-01
7.86632299e-01 -8.77253115e-01 -1.00852680e+00 1.09164274e+00
-1.12084520e+00 -1.84572324e-01 3.33586276e-01 4.57827419e-01
-9.20920968e-02 6.14177346e-01 -5.51092684e-01 -9.29561973e-01
-2.57486943e-02 -1.08958757e+00 6.71079934e-01 8.27890933e-01
7.41904438e-01 -8.10470283e-01 -4.43622023e-02 2.59101063e-01
6.62714541e-01 2.79709160e-01 9.19668257e-01 6.95482075e-01
-1.12401581e+00 1.20993868e-01 -8.26008677e-01 4.35914010e-01
2.89344281e-01 4.08068806e-01 -1.35944533e+00 -1.29831001e-01
-1.63032692e-02 -1.05750978e-01 8.84092510e-01 5.23149610e-01
1.63638031e+00 -9.30133276e-03 -3.45808752e-02 1.11197305e+00
1.89181435e+00 3.58673185e-01 1.06802213e+00 4.92653459e-01
7.37884045e-01 5.66345453e-01 5.62387049e-01 2.91198999e-01
3.52724522e-01 6.02960885e-01 7.35315621e-01 -6.63740039e-01
-1.84683189e-01 1.25454605e-01 4.31808084e-02 7.36263573e-01
-4.41505462e-01 -2.63626873e-01 -6.17883205e-01 2.19848752e-01
-1.64557862e+00 -1.07123709e+00 -7.28602409e-01 2.19578886e+00
8.89768898e-01 -1.67475179e-01 -2.09836453e-01 -4.25476804e-02
8.73866677e-01 1.73424765e-01 -6.95587158e-01 -8.61695558e-02
-5.66321969e-01 -1.90237649e-02 7.26584315e-01 2.89124340e-01
-9.84664261e-01 7.70962536e-01 6.83406734e+00 4.98030812e-01
-1.50918067e+00 -7.27581903e-02 6.33936644e-01 6.11250056e-03
-2.40578577e-01 3.89511809e-02 -4.22255099e-01 5.57893693e-01
2.93591768e-01 1.20893352e-01 6.78136885e-01 6.01264238e-01
3.48853797e-01 -5.76281846e-01 -4.86731231e-01 1.35758293e+00
1.10077016e-01 -1.07243359e+00 -1.95225641e-01 -2.39130780e-02
9.65859950e-01 -1.28321201e-01 4.15494800e-01 -1.12734564e-01
8.49633962e-02 -7.49369979e-01 2.49641299e-01 7.98076391e-01
9.75699365e-01 -5.52275240e-01 5.74162245e-01 7.95867145e-02
-9.00507927e-01 5.70696779e-02 -6.72943592e-01 -2.38172874e-01
6.96086213e-02 9.05146420e-01 -1.94889069e-01 7.25943387e-01
8.62816870e-01 6.94960177e-01 -6.07591748e-01 1.31486106e+00
-4.55202997e-01 4.95836616e-01 -2.02634707e-01 -8.75047594e-02
6.73974119e-03 -8.52939844e-01 2.05156356e-01 8.52218866e-01
3.10620159e-01 1.77407712e-01 -2.76192240e-02 1.14076722e+00
5.89450672e-02 -9.87119153e-02 -2.64386415e-01 2.14030117e-01
2.00980067e-01 1.44049919e+00 -5.99412978e-01 -8.66856202e-02
-5.72574377e-01 1.00949061e+00 -1.01954348e-01 7.85034955e-01
-9.99663055e-01 -5.43846250e-01 7.15574920e-01 -7.77117759e-02
-8.14655274e-02 -4.12632048e-01 -3.89731109e-01 -1.33088279e+00
-7.17252942e-06 -6.16795421e-01 -1.76733583e-01 -1.31961203e+00
-1.14676130e+00 3.44818592e-01 -1.38544559e-01 -1.32585549e+00
4.75507319e-01 -9.95086968e-01 -6.21235430e-01 1.02263844e+00
-2.14475942e+00 -9.82766032e-01 -8.84729505e-01 6.97566986e-01
2.15226501e-01 1.74755320e-01 8.05511177e-01 4.74552721e-01
-7.91097581e-01 1.06285654e-01 5.68435609e-01 3.01069040e-02
1.03734338e+00 -1.01846397e+00 -1.89430714e-01 1.22003293e+00
-2.23868728e-01 6.92796946e-01 7.39461124e-01 -2.74394959e-01
-1.62081707e+00 -8.85277212e-01 2.27145866e-01 9.69309881e-02
3.07861716e-01 -1.30781159e-01 -8.59972417e-01 1.55192196e-01
5.60497642e-01 -3.91904730e-03 4.19534296e-01 2.01373547e-02
-2.70690203e-01 -5.68252861e-01 -9.71678436e-01 6.03541195e-01
9.09184337e-01 -3.29359233e-01 -1.38009921e-01 2.02621222e-01
5.03684700e-01 -4.53485131e-01 -1.87596500e-01 2.58828610e-01
7.44553030e-01 -1.51469696e+00 7.93024421e-01 1.21714763e-01
4.21145380e-01 -8.50412011e-01 -1.24018744e-01 -1.12691736e+00
-4.21743542e-01 -3.96700710e-01 2.46525213e-01 1.27770770e+00
-4.46049422e-02 -1.00024271e+00 5.75825334e-01 6.41904354e-01
-1.48748428e-01 -4.94223297e-01 -4.37990665e-01 -6.66798830e-01
-2.49050915e-01 -2.25394279e-01 5.43781757e-01 9.52012777e-01
-8.40452850e-01 -9.57161039e-02 -7.39226878e-01 4.73332018e-01
1.10874963e+00 6.47816300e-01 7.21036196e-01 -1.18892169e+00
-2.81126201e-01 -4.28187013e-01 -1.18923984e-01 -9.70429420e-01
-1.38170728e-02 -3.92146438e-01 9.87351686e-02 -1.62217629e+00
3.75560462e-01 -6.85999155e-01 -3.92244905e-01 3.18653554e-01
-1.95842206e-01 5.27000606e-01 -4.07188237e-02 3.31527561e-01
-4.89253819e-01 5.94060183e-01 1.50314856e+00 -1.69916525e-01
-1.39874548e-01 -1.74707338e-01 -8.97089601e-01 7.51953542e-01
9.01620090e-01 -5.36064021e-02 -3.75123352e-01 -7.03832507e-01
3.67155701e-01 -4.62623924e-01 5.16685128e-01 -9.57881987e-01
2.65747756e-01 -5.73138833e-01 6.23271704e-01 -5.94841719e-01
4.59940553e-01 -8.50745797e-01 1.69845298e-01 2.35141158e-01
1.72922343e-01 -2.96592772e-01 2.02168137e-01 4.06212300e-01
-2.14438632e-01 3.39341350e-02 1.14989960e+00 -5.61590642e-02
-1.03075969e+00 3.64397734e-01 6.86547011e-02 -1.70224249e-01
9.51240718e-01 -3.55588794e-01 -7.61419952e-01 -3.25464964e-01
1.04629755e-01 -5.26836552e-02 8.59331071e-01 -3.94901633e-02
5.00880897e-01 -1.25476587e+00 -3.27309251e-01 3.68174642e-01
5.52189425e-02 -8.30744132e-02 4.60305989e-01 9.33842421e-01
-9.92531180e-01 2.46522516e-01 -4.95303690e-01 -6.98882103e-01
-1.11648810e+00 4.35644746e-01 4.85899627e-01 4.56382543e-01
-4.40594345e-01 6.42345428e-01 4.11160439e-01 -1.31194115e-01
3.28662172e-02 -2.58674115e-01 2.06854828e-02 -3.88723403e-01
4.07807559e-01 3.78010482e-01 -2.09968984e-01 -5.22938132e-01
-1.70582160e-01 1.01823306e+00 3.22495967e-01 9.40863341e-02
1.25353277e+00 -3.78845096e-01 -4.30898845e-01 3.91244531e-01
1.00529182e+00 2.56968081e-01 -1.45117319e+00 -1.70025051e-01
-8.22607577e-01 -9.73262608e-01 3.52559596e-01 -8.92613530e-01
-1.32641649e+00 1.06219864e+00 7.91559577e-01 1.69565216e-01
1.72060323e+00 -6.11362875e-01 5.78434348e-01 4.47067261e-01
3.82462025e-01 -1.21949267e+00 -1.79449707e-01 3.59550983e-01
4.72757906e-01 -1.38870335e+00 2.57880241e-01 -5.85962117e-01
-4.73338187e-01 1.40428519e+00 9.20520544e-01 1.46623299e-01
3.90911400e-01 4.70510946e-04 3.72000933e-01 3.56637612e-02
-1.63912788e-01 -1.84992030e-01 1.16302773e-01 6.80253029e-01
5.16376197e-01 3.09090484e-02 -2.73176789e-01 1.51718020e-01
1.03845343e-01 -2.38972553e-03 5.77595949e-01 3.66816908e-01
-5.38733065e-01 -1.09582114e+00 -5.55821776e-01 1.70837805e-01
-2.04717025e-01 -1.00151651e-01 -1.88678339e-01 8.59154046e-01
3.19349080e-01 1.00466621e+00 -4.81877401e-02 -2.12883592e-01
1.29689395e-01 -3.74170661e-01 6.19878411e-01 -1.36672795e-01
8.53030756e-02 1.87821507e-01 -2.33947814e-01 -6.17337346e-01
-9.13605869e-01 -3.95587534e-01 -1.09664702e+00 -4.48705941e-01
-5.10445178e-01 -1.91527247e-01 8.27256918e-01 5.31293273e-01
2.90376782e-01 4.48486924e-01 7.89772391e-01 -6.85383558e-01
-5.94887249e-02 -6.84552550e-01 -9.32633221e-01 3.72187912e-01
2.87284553e-01 -8.62075388e-01 -4.06188399e-01 1.42912865e-02] | [10.573589324951172, -2.6102418899536133] |
8d3dcac7-5e61-434c-a2d8-50571f7754df | graph-free-multi-hop-reading-comprehension-a | 2107.11823 | null | https://arxiv.org/abs/2107.11823v1 | https://arxiv.org/pdf/2107.11823v1.pdf | Graph-free Multi-hop Reading Comprehension: A Select-to-Guide Strategy | Multi-hop reading comprehension (MHRC) requires not only to predict the correct answer span in the given passage, but also to provide a chain of supporting evidences for reasoning interpretability. It is natural to model such a process into graph structure by understanding multi-hop reasoning as jumping over entity nodes, which has made graph modelling dominant on this task. Recently, there have been dissenting voices about whether graph modelling is indispensable due to the inconvenience of the graph building, however existing state-of-the-art graph-free attempts suffer from huge performance gap compared to graph-based ones. This work presents a novel graph-free alternative which firstly outperform all graph models on MHRC. In detail, we exploit a select-to-guide (S2G) strategy to accurately retrieve evidence paragraphs in a coarse-to-fine manner, incorporated with two novel attention mechanisms, which surprisingly shows conforming to the nature of multi-hop reasoning. Our graph-free model achieves significant and consistent performance gain over strong baselines and the current new state-of-the-art on the MHRC benchmark, HotpotQA, among all the published works. | ['Hai Zhao', 'Zhuosheng Zhang', 'Bohong Wu'] | 2021-07-25 | null | null | null | null | ['multi-hop-reading-comprehension'] | ['natural-language-processing'] | [ 1.36366218e-01 8.37745130e-01 -1.75276734e-02 -2.74655223e-01
-7.84682572e-01 -5.92435956e-01 5.24878323e-01 8.58053744e-01
-1.95085660e-01 7.86568046e-01 6.47313237e-01 -9.10715461e-01
-3.88339847e-01 -1.14808393e+00 -1.17953992e+00 6.08290685e-03
1.56301022e-01 7.19720721e-01 5.41144788e-01 -6.80781484e-01
4.73713815e-01 -8.82686973e-02 -1.17054665e+00 6.35339677e-01
1.31314719e+00 4.40542072e-01 1.84750095e-01 1.03381109e+00
-3.91733289e-01 1.50095439e+00 -5.44022143e-01 -1.14011610e+00
-2.67133713e-01 -6.10879779e-01 -1.52547669e+00 -3.05943102e-01
6.68284535e-01 -2.94577062e-01 -4.25630778e-01 9.35327053e-01
2.93806642e-01 4.53508273e-02 6.42428041e-01 -8.94274235e-01
-1.08003914e+00 1.09744656e+00 -5.47954917e-01 5.17302454e-01
8.93578112e-01 1.26986448e-02 1.67138720e+00 -3.53940487e-01
8.09798479e-01 1.24123311e+00 5.54572642e-01 2.69496024e-01
-7.21350729e-01 8.98649171e-02 5.94043732e-01 8.72688115e-01
-8.63529146e-01 9.19012818e-03 6.65198743e-01 -1.80475891e-01
1.29331625e+00 5.08029938e-01 7.79127359e-01 1.04988635e+00
2.86140829e-01 8.00620556e-01 9.49492574e-01 -5.60031891e-01
-4.38215472e-02 -3.59000862e-01 6.85256720e-01 1.16168761e+00
3.32200199e-01 -4.99913603e-01 -5.62631845e-01 1.62091225e-01
4.17043447e-01 -4.91346329e-01 -7.14909017e-01 1.34064913e-01
-1.05935705e+00 6.94244981e-01 8.57855022e-01 2.17075303e-01
-4.63678032e-01 1.10550374e-01 2.27162451e-01 5.46891987e-01
2.09701821e-01 6.38174474e-01 -3.06289315e-01 -1.52944192e-01
-5.58284044e-01 3.73773634e-01 1.07597399e+00 9.94050622e-01
4.19760495e-01 -4.48577017e-01 -6.48256063e-01 4.45649236e-01
3.31330180e-01 3.62329781e-02 -2.31120493e-02 -6.45391881e-01
1.00959337e+00 9.01703656e-01 -2.98724145e-01 -1.33743560e+00
-5.49296796e-01 -1.01884663e+00 -8.31154525e-01 -4.13993210e-01
4.50719416e-01 2.41632625e-01 -5.78514636e-01 1.60965466e+00
2.53098220e-01 -1.42374849e-02 5.87251782e-02 9.56993759e-01
1.26816535e+00 5.90927482e-01 -3.02308556e-02 7.85582811e-02
1.46617961e+00 -1.43183148e+00 -7.09900141e-01 -3.15156758e-01
7.38557339e-01 -3.62701118e-01 1.43382025e+00 4.46653724e-01
-1.23910952e+00 -4.38886940e-01 -1.16868448e+00 -6.38839662e-01
-4.47050899e-01 -4.68039215e-01 5.81519723e-01 3.43528837e-01
-1.29946077e+00 5.39719582e-01 -3.08445096e-01 -2.65597254e-01
2.09405944e-01 -1.60533115e-01 -2.68735528e-01 -5.90956628e-01
-1.62418544e+00 1.15421045e+00 4.88781750e-01 3.03038985e-01
-4.85198051e-01 -5.89134276e-01 -7.33814299e-01 3.27681839e-01
9.13693964e-01 -1.40584958e+00 1.11390996e+00 -2.56681889e-01
-1.03081453e+00 6.97778761e-01 -1.90536350e-01 -6.53568208e-01
7.95681775e-01 -3.61695439e-01 -3.58152419e-01 3.15470964e-01
1.80776507e-01 2.35454991e-01 5.08322299e-01 -1.19070554e+00
-2.49475092e-01 -2.47272134e-01 7.88633704e-01 4.00448829e-01
9.03890133e-02 -3.94621909e-01 -7.05200315e-01 -2.92127460e-01
1.02406390e-01 -4.03800607e-01 7.67231882e-02 -5.39315045e-01
-8.13361824e-01 -6.05268180e-01 1.23571053e-01 -1.02212262e+00
1.81075215e+00 -1.50800741e+00 5.09176135e-01 -2.23746384e-03
7.44324088e-01 1.19133592e-01 -2.56876469e-01 9.49328303e-01
1.72426671e-01 3.73524845e-01 -4.12348770e-02 -2.19125420e-01
3.62275064e-01 1.91117257e-01 -2.97070324e-01 1.10488102e-01
8.46808553e-02 1.39992940e+00 -1.18893409e+00 -6.46631062e-01
-1.41860068e-01 4.38551158e-02 -5.15313625e-01 1.74409598e-01
-6.23899043e-01 2.58711934e-01 -5.44100106e-01 4.95196253e-01
4.35624719e-01 -8.70866239e-01 1.30748495e-01 -1.89728528e-01
3.02471578e-01 6.54821575e-01 -6.78615928e-01 1.70882010e+00
-2.44762465e-01 3.64169091e-01 -2.66703129e-01 -7.32748806e-01
5.51286638e-01 2.04138290e-02 -3.63773912e-01 -9.02875304e-01
-9.15728137e-02 1.73270106e-01 3.17214012e-01 -6.29940867e-01
7.06602335e-01 3.41417521e-01 7.17564821e-02 3.46667260e-01
-6.73734397e-02 -6.97625056e-02 5.69435000e-01 1.02596962e+00
1.45449531e+00 6.21038899e-02 2.72841126e-01 -1.30372912e-01
6.22335136e-01 7.32694119e-02 3.19470651e-02 1.20117569e+00
2.61528641e-01 6.57423317e-01 9.05073583e-01 -2.80047059e-01
-6.75916672e-01 -7.19188571e-01 5.31173944e-01 9.55474317e-01
1.94406077e-01 -9.04318631e-01 -9.16724503e-01 -9.38557029e-01
-1.61872953e-01 1.24225628e+00 -7.62732387e-01 -2.57417113e-01
-7.58554459e-01 -4.39860076e-01 5.52862763e-01 5.19639313e-01
7.28889048e-01 -9.35164571e-01 -3.37390333e-01 4.70227808e-01
-5.50171554e-01 -1.33365631e+00 -1.36682153e-01 -1.79472044e-01
-6.01913273e-01 -1.30473769e+00 -3.30619514e-01 -5.28678715e-01
4.67731804e-01 3.46192509e-01 1.89071262e+00 1.01163828e+00
1.60837740e-01 4.41326678e-01 -7.21904516e-01 -4.24783260e-01
-4.55434114e-01 4.06691164e-01 -7.56946921e-01 -3.32152426e-01
4.90378886e-02 -4.21957225e-01 -6.55390382e-01 -3.22063938e-02
-6.66248083e-01 4.79420215e-01 7.34390736e-01 4.66363370e-01
4.21775550e-01 3.88017334e-02 7.45067894e-01 -1.24590611e+00
1.14623988e+00 -5.20284235e-01 -1.90990075e-01 9.09319103e-01
-7.77448058e-01 1.73073754e-01 8.99954379e-01 3.57758224e-01
-8.03861082e-01 -9.41014647e-01 -3.70393455e-01 9.85345617e-02
2.09404841e-01 9.87603605e-01 -6.57255650e-02 1.96178630e-01
4.56271470e-01 2.39272326e-01 -3.55909467e-01 -2.13215694e-01
6.73622489e-01 1.49855897e-01 4.80318606e-01 -8.01588237e-01
6.75325572e-01 -9.61484946e-03 1.98532820e-01 -5.95873237e-01
-1.33763301e+00 -4.18014526e-01 -5.97634375e-01 -3.66169035e-01
1.05864429e+00 -6.58003390e-01 -7.01280653e-01 2.33556002e-01
-1.54367959e+00 -3.23242307e-01 -2.47797966e-02 -3.99212986e-02
-3.16541135e-01 5.81239760e-01 -7.50371993e-01 -4.69578028e-01
-4.56683576e-01 -9.41230714e-01 9.63560283e-01 1.05632670e-01
-2.00536028e-01 -1.32680702e+00 -6.91701323e-02 1.02919924e+00
4.11909342e-01 2.97420084e-01 1.51713538e+00 -7.64949024e-01
-1.06824923e+00 1.54033890e-02 -5.28437257e-01 -1.06446825e-01
-3.56419683e-01 -2.29798183e-01 -6.30775928e-01 -1.92976780e-02
-2.95475602e-01 -4.03887630e-01 9.28111792e-01 4.52045612e-02
1.22436345e+00 -5.22511244e-01 -6.09270334e-02 2.23472625e-01
1.29489434e+00 -4.58674878e-01 8.11779618e-01 5.33076286e-01
1.00113297e+00 6.31318271e-01 3.49123657e-01 -1.31303608e-01
1.37338197e+00 4.44097102e-01 8.14109445e-01 -1.09312534e-01
-4.50304389e-01 -7.06976414e-01 -1.86181009e-01 1.24909997e+00
-3.15356284e-01 -9.26234305e-01 -9.57451940e-01 4.60819215e-01
-2.06107378e+00 -8.52932692e-01 -8.92533123e-01 1.66680121e+00
5.44256389e-01 3.93745124e-01 -8.38651359e-02 2.36732200e-01
4.12387013e-01 4.81461465e-01 -2.85548121e-01 -4.66689438e-01
-2.85764247e-01 3.28670852e-02 6.67182729e-02 7.11200416e-01
-5.13971925e-01 8.27644944e-01 5.32871675e+00 6.96414232e-01
-4.47987229e-01 -1.74763370e-02 5.16567707e-01 6.13185644e-01
-7.70304620e-01 7.76909366e-02 -5.98989725e-01 1.31853875e-02
7.94007063e-01 -1.39628425e-01 4.56513286e-01 2.80376583e-01
-8.90726969e-02 -9.02562290e-02 -1.07371414e+00 6.43748522e-01
2.80457526e-01 -1.46803892e+00 4.57231194e-01 -1.93546221e-01
4.16885406e-01 -1.93562075e-01 -3.30264777e-01 5.24074554e-01
2.40624607e-01 -1.17268515e+00 7.56668210e-01 6.02829933e-01
2.74008393e-01 -4.68517870e-01 9.37689543e-01 7.54569769e-01
-1.20965469e+00 2.95619126e-02 -3.78703564e-01 -1.75362468e-01
3.62574071e-01 4.28505868e-01 -7.97911286e-01 1.31810081e+00
4.49851513e-01 6.11589193e-01 -1.05140102e+00 8.74218822e-01
-9.16313589e-01 7.14434624e-01 1.98434398e-01 -4.52497959e-01
4.75630850e-01 -7.53990784e-02 7.31105864e-01 1.12214315e+00
5.78685068e-02 3.65558028e-01 -7.71514326e-02 7.82616854e-01
-3.96568626e-01 2.15900049e-01 -3.81246626e-01 -1.45973399e-01
2.47769251e-01 1.01327491e+00 -7.10924089e-01 -4.37341064e-01
-4.81056899e-01 9.92837369e-01 1.09465921e+00 2.46537924e-01
-9.05732751e-01 -2.99410284e-01 -3.09960604e-01 2.09571928e-01
2.07219765e-01 -1.61553025e-01 -1.64415941e-01 -1.18414688e+00
3.93052191e-01 -1.08081675e+00 8.27376485e-01 -1.03248894e+00
-1.34006834e+00 8.70328128e-01 -2.00918261e-02 -5.78114867e-01
-8.80253017e-02 -4.72450346e-01 -7.75580406e-01 7.66348660e-01
-1.86290109e+00 -1.22869205e+00 -6.77363634e-01 5.36996782e-01
5.95768213e-01 3.42543602e-01 6.17822587e-01 4.03876007e-02
-2.78701425e-01 5.13518453e-01 -4.73759264e-01 6.46248534e-02
2.38590613e-01 -1.78716576e+00 8.30098748e-01 9.87510324e-01
5.05857825e-01 7.20266044e-01 7.82557249e-01 -6.44997895e-01
-1.61354601e+00 -8.12845528e-01 1.26791990e+00 -6.89733982e-01
8.56967390e-01 -2.40535200e-01 -1.45672703e+00 7.94260561e-01
6.01262629e-01 -3.86055350e-01 1.70942351e-01 4.04976040e-01
-4.43589479e-01 1.11893937e-01 -6.73619092e-01 7.31508851e-01
1.45151079e+00 -4.13239300e-01 -1.12404656e+00 5.69416404e-01
1.13199103e+00 -5.66472769e-01 -8.31874013e-01 2.96239406e-01
2.46606380e-01 -1.26876438e+00 7.98957527e-01 -8.33391726e-01
7.41955221e-01 -1.95219293e-01 6.21064492e-02 -1.45093846e+00
-5.41739047e-01 -6.77026927e-01 -4.21805501e-01 1.16542101e+00
8.75457406e-01 -7.40315616e-01 5.03622711e-01 2.85311788e-01
-5.27631700e-01 -1.06353879e+00 -6.95874691e-01 -5.12139738e-01
9.39801559e-02 -2.37280875e-01 7.78830469e-01 6.91217601e-01
1.48866117e-01 8.39866281e-01 -8.02532136e-02 2.68443733e-01
5.60683608e-01 2.51869500e-01 7.09621787e-01 -1.23434150e+00
-5.74432492e-01 -5.20558834e-01 -9.06069800e-02 -1.43913472e+00
1.97572440e-01 -1.08860326e+00 -2.89268941e-01 -2.74342465e+00
2.00191766e-01 -4.65808176e-02 -1.43750206e-01 1.05928175e-01
-7.12222934e-01 -2.76823759e-01 2.36454830e-01 5.82844811e-03
-1.20077550e+00 3.34744364e-01 1.73974478e+00 -2.29445651e-01
2.60927439e-01 -9.55611914e-02 -1.01056981e+00 5.42740524e-01
5.87810278e-01 -1.47997051e-01 -8.99139524e-01 -8.90594304e-01
1.11635900e+00 4.13978994e-01 4.21559960e-01 -7.31773257e-01
5.15087724e-01 1.86972991e-01 -2.49198135e-02 -5.41288078e-01
4.46802378e-02 -4.21303153e-01 -9.33976844e-02 2.91438729e-01
-5.01412332e-01 5.92048883e-01 -4.52059396e-02 9.32397008e-01
-2.78525233e-01 -2.48474434e-01 4.50426564e-02 -4.07483578e-01
-7.88581789e-01 1.22229233e-01 -2.27335319e-02 5.24331629e-01
5.71576059e-01 -6.57187179e-02 -1.15782630e+00 -7.40729332e-01
-4.67704415e-01 4.45675999e-01 1.27167597e-01 3.91689181e-01
5.86743355e-01 -7.80994475e-01 -1.00415862e+00 -3.98055911e-01
1.99996814e-01 2.35713914e-01 3.70857596e-01 9.62350667e-01
-6.85541511e-01 5.33330560e-01 1.43296167e-01 -3.21022034e-01
-1.12816298e+00 6.20536447e-01 1.47484571e-01 -9.36031699e-01
-9.09671128e-01 9.35444951e-01 -1.59017950e-01 -2.78284937e-01
1.18285837e-02 -5.54572880e-01 -4.84346390e-01 -1.13042526e-01
3.72206718e-01 5.06562173e-01 4.89408135e-01 -2.87846774e-01
-1.60819605e-01 5.98000944e-01 -2.24090129e-01 2.21494198e-01
1.13876915e+00 -3.34229171e-01 -3.03888470e-01 3.46278310e-01
7.55826294e-01 8.96481723e-02 -7.17301130e-01 -7.99078215e-03
1.19043730e-01 -2.83685148e-01 -1.81492776e-01 -1.19083166e+00
-5.01257539e-01 8.82015049e-01 -5.69977164e-01 7.62766004e-01
8.51038933e-01 1.27756938e-01 8.43986452e-01 4.59192097e-01
5.79446137e-01 -6.46956801e-01 1.72039315e-01 6.29387319e-01
1.31387687e+00 -1.19113123e+00 1.43844619e-01 -7.82956779e-01
-7.09411979e-01 1.12736487e+00 6.56549096e-01 1.17107064e-01
6.89611137e-02 -3.08453262e-01 -2.59199142e-01 -6.77655458e-01
-1.08025372e+00 -1.29192591e-01 7.34703898e-01 3.67663413e-01
4.14551735e-01 -2.93091722e-02 -3.80371720e-01 4.87164766e-01
-5.52996755e-01 -2.65022933e-01 5.55165887e-01 6.83226764e-01
-4.59733754e-01 -7.17352509e-01 9.13874358e-02 4.27847922e-01
-3.31997365e-01 -3.79117429e-01 -7.05442548e-01 1.12987328e+00
-3.98166597e-01 1.35567224e+00 -3.65244269e-01 -1.12796523e-01
6.05360329e-01 -4.51571494e-02 7.15558469e-01 -3.05220574e-01
-8.27161372e-01 -5.92331231e-01 6.88510656e-01 -6.25259995e-01
-1.91232711e-01 -9.92332995e-02 -1.28758037e+00 -4.97629285e-01
-3.74056399e-01 3.15397620e-01 2.71970689e-01 1.10482132e+00
3.33057553e-01 1.09065819e+00 -5.91389611e-02 6.03615083e-02
-5.52397549e-01 -9.92549062e-01 -3.09164286e-01 4.47470099e-01
2.43152887e-01 -3.05434078e-01 -3.45638990e-01 -4.24903899e-01] | [10.807378768920898, 7.950563430786133] |
3dae4b6c-0e1c-4c97-bcb3-0b8d7d02e825 | sparse-random-hypergraphs-non-backtracking | 2203.07346 | null | https://arxiv.org/abs/2203.07346v3 | https://arxiv.org/pdf/2203.07346v3.pdf | Sparse random hypergraphs: Non-backtracking spectra and community detection | We consider the community detection problem in a sparse $q$-uniform hypergraph $G$, assuming that $G$ is generated according to the Hypergraph Stochastic Block Model (HSBM). We prove that a spectral method based on the non-backtracking operator for hypergraphs works with high probability down to the generalized Kesten-Stigum detection threshold conjectured by Angelini et al. (2015). We characterize the spectrum of the non-backtracking operator for the sparse HSBM and provide an efficient dimension reduction procedure using the Ihara-Bass formula for hypergraphs. As a result, community detection for the sparse HSBM on $n$ vertices can be reduced to an eigenvector problem of a $2n\times 2n$ non-normal matrix constructed from the adjacency matrix and the degree matrix of the hypergraph. To the best of our knowledge, this is the first provable and efficient spectral algorithm that achieves the conjectured threshold for HSBMs with $r$ blocks generated according to a general symmetric probability tensor. | ['Yizhe Zhu', 'Ludovic Stephan'] | 2022-03-14 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 2.09424466e-01 5.63433886e-01 -2.30651572e-01 4.39763606e-01
-4.60083187e-01 -6.69447839e-01 1.65171232e-02 -5.06625623e-02
1.07403249e-01 4.98126119e-01 -1.76712364e-01 -6.09416485e-01
-5.42901695e-01 -1.20146513e+00 -8.35387349e-01 -1.04531229e+00
-6.85022652e-01 9.28913593e-01 1.81893229e-01 -1.48618951e-01
1.95078850e-01 2.26578087e-01 -1.18206024e+00 -9.66512784e-02
4.28349733e-01 4.17858928e-01 -1.55488029e-01 9.58805859e-01
-5.04511520e-02 5.10038674e-01 -8.95056799e-02 -5.27527809e-01
8.40703189e-01 -5.31865776e-01 -1.13656044e+00 2.88807213e-01
4.08060193e-01 3.87330025e-01 -8.22639585e-01 1.52908075e+00
1.40828252e-01 -3.13434303e-01 5.02802908e-01 -1.33363092e+00
-3.13324481e-01 1.58022380e+00 -1.24459553e+00 4.04169649e-01
4.55210209e-01 -9.10791829e-02 1.62120223e+00 -3.91013771e-01
8.98365974e-01 1.07165360e+00 4.56351995e-01 1.65855005e-01
-1.78074253e+00 -9.35186982e-01 -2.34974280e-01 2.25240603e-01
-1.82283342e+00 1.59611106e-01 7.38329232e-01 -6.59168720e-01
6.50441468e-01 6.14050090e-01 9.89994824e-01 4.99609292e-01
-3.40832174e-01 2.37494797e-01 1.08492339e+00 -6.26977026e-01
8.28166083e-02 -3.62952113e-01 2.78446943e-01 1.10727370e+00
1.30586302e+00 -6.14170320e-02 -4.89588469e-01 -4.85376388e-01
4.87141460e-01 -5.07075012e-01 -2.97003686e-01 -8.05574596e-01
-7.83620000e-01 9.09684956e-01 3.07117522e-01 3.33096921e-01
-1.31609321e-01 5.12585223e-01 1.32686287e-01 2.17803463e-01
-4.02413122e-03 1.68712795e-01 -5.07845357e-02 2.25445583e-01
-9.23620820e-01 -4.99396399e-02 1.09144843e+00 1.16325963e+00
9.52498674e-01 -1.18064635e-01 3.55015665e-01 -6.72097271e-03
1.83728158e-01 1.16341686e+00 -5.86940408e-01 -9.12214220e-01
5.08826911e-01 5.85758269e-01 -1.35505781e-01 -1.26887858e+00
-3.77114624e-01 -7.32031465e-01 -1.20161903e+00 -3.87112737e-01
4.82730359e-01 1.37971222e-01 -4.75324869e-01 2.05107427e+00
3.70185554e-01 1.33589864e-01 -3.12083453e-01 5.91106772e-01
2.91919172e-01 5.30137420e-01 -4.47542578e-01 -5.08662105e-01
1.21193707e+00 -4.82707918e-01 -1.34068638e-01 -8.52370635e-02
7.80353189e-01 -4.36021119e-01 4.28063124e-01 3.40517044e-01
-1.17610502e+00 2.69463778e-01 -8.72639000e-01 5.97786367e-01
4.94627386e-01 -4.15876240e-01 6.52989924e-01 1.15973389e+00
-1.18103456e+00 4.61971313e-02 -5.31368554e-01 -3.04270476e-01
5.32529391e-02 2.91584373e-01 -3.31519097e-01 -5.13729990e-01
-8.51913571e-01 2.21142620e-01 2.44655773e-01 -4.82679531e-02
-8.39553237e-01 -3.30404609e-01 -5.90639055e-01 1.72841415e-01
9.14182007e-01 -7.26569533e-01 6.43739820e-01 -5.54756284e-01
-6.18459642e-01 9.84912276e-01 2.45682746e-02 -6.07870579e-01
8.73907190e-03 8.23015690e-01 1.89323783e-01 7.14649439e-01
2.63540655e-01 -4.12793577e-01 4.38753456e-01 -1.08678770e+00
-2.52508730e-01 -6.43609405e-01 6.03210330e-02 -2.32748657e-01
4.22508679e-02 -9.43091959e-02 -9.44013521e-02 -1.01802737e-01
7.37901151e-01 -1.48123658e+00 -4.52970624e-01 -8.42784822e-01
-8.23725402e-01 -3.39545831e-02 1.09915547e-01 -2.68155396e-01
1.42920005e+00 -1.93416786e+00 5.68122983e-01 1.20932031e+00
7.52671123e-01 -2.19032899e-01 -2.42837116e-01 1.07810986e+00
-2.53404111e-01 4.54278529e-01 -2.30762556e-01 2.58672357e-01
-2.98950030e-03 -1.54074937e-01 -7.71550983e-02 1.11598647e+00
-7.39783406e-01 2.93063581e-01 -7.80325413e-01 -2.96579123e-01
-5.26771605e-01 -1.20147049e-01 -7.39315808e-01 -3.48056585e-01
-1.36134159e-02 -2.17359111e-01 -1.21471025e-01 4.09835488e-01
1.12274778e+00 -7.83506155e-01 1.04875016e+00 5.89896813e-02
1.57133639e-01 -1.15633504e-02 -1.77737451e+00 1.18177438e+00
1.59353465e-01 3.83726358e-01 8.06354165e-01 -1.06120384e+00
4.46808219e-01 1.31241962e-01 5.98531485e-01 -8.33024457e-02
-8.69317353e-02 3.39130610e-01 3.89853716e-01 -6.19036444e-02
1.95057780e-01 -1.70487136e-01 -1.26127452e-01 8.21969867e-01
-1.34013250e-01 2.09041998e-01 7.30579734e-01 1.36003590e+00
1.91876030e+00 -8.26203704e-01 -7.76623487e-02 -8.31157446e-01
4.01285410e-01 -1.39575735e-01 5.05125761e-01 1.09941888e+00
2.28343438e-02 1.63508922e-01 1.13920522e+00 1.11884817e-01
-1.17543149e+00 -8.92898202e-01 2.02640384e-01 6.15467489e-01
1.88733876e-01 -9.27952766e-01 -9.02532458e-01 -1.63030744e-01
5.04615605e-02 2.33195469e-01 -7.51856804e-01 -1.35004684e-01
-1.77575514e-01 -9.05573368e-01 2.35753119e-01 -4.70187478e-02
2.48000175e-01 -3.26163381e-01 -1.17382787e-01 6.86610565e-02
-3.95483881e-01 -1.15166605e+00 -6.31519794e-01 5.48662506e-02
-7.10757732e-01 -1.67266452e+00 -1.83855459e-01 -7.02496052e-01
8.42483401e-01 6.53832853e-01 9.34897244e-01 2.54989564e-01
-4.85480100e-01 4.56199676e-01 -1.88884616e-01 3.64612579e-01
-2.76571572e-01 1.92652538e-01 3.09906695e-02 -4.32769954e-02
5.13914414e-02 -5.97638667e-01 -7.83103704e-01 1.35586888e-01
-8.24946880e-01 2.80927997e-02 3.62044454e-01 5.82788050e-01
4.31616932e-01 5.85486412e-01 -1.79109201e-01 -1.04872417e+00
1.27773359e-01 -6.32705688e-01 -1.15154982e+00 -2.11108457e-02
-7.17322290e-01 3.26768875e-01 2.91403711e-01 1.05129816e-01
-3.16383749e-01 1.43214345e-01 3.71128231e-01 3.81061733e-02
7.21257091e-01 7.94430077e-01 -6.24411888e-02 -3.98662060e-01
5.78683794e-01 2.17927396e-01 -1.70714110e-01 -1.74729154e-01
4.04235125e-01 9.30536538e-02 -6.85260668e-02 -5.96368253e-01
1.19052720e+00 8.73984277e-01 1.07119656e+00 -7.62946725e-01
-5.67781687e-01 -7.40223289e-01 -1.88677036e-03 -1.62917376e-01
2.05904111e-01 -7.05276489e-01 -1.18777895e+00 2.42427766e-01
-7.55941093e-01 -2.89208829e-01 -1.59097333e-02 3.52023095e-01
-5.14033198e-01 8.49255621e-01 -8.32949340e-01 -1.21504271e+00
-8.39228630e-02 -9.20244336e-01 5.21534801e-01 -5.13327658e-01
2.85323918e-01 -3.82303625e-01 4.70521003e-01 4.18400437e-01
1.30082056e-01 1.25072941e-01 1.10494328e+00 -3.18971097e-01
-1.18937421e+00 -2.27939591e-01 -3.62426549e-01 -3.68076742e-01
-5.65755904e-01 1.49437159e-01 -9.45599377e-02 -6.20558739e-01
-2.00752184e-01 1.45715013e-01 9.96023476e-01 3.64702463e-01
7.46865630e-01 -7.00085461e-01 -4.53704357e-01 4.58979815e-01
1.84497905e+00 -3.81676883e-01 4.98134553e-01 8.00167844e-02
5.99850655e-01 3.81551415e-01 -3.58438939e-02 5.88656247e-01
2.46109113e-01 2.56146371e-01 4.52436626e-01 5.38867176e-01
1.70019090e-01 -1.80366114e-01 1.12283491e-01 1.10514474e+00
-8.75499845e-02 -2.54116833e-01 -1.10157824e+00 6.81483805e-01
-1.64958096e+00 -1.24905252e+00 -8.92976880e-01 2.36327744e+00
8.46157014e-01 1.87140658e-01 4.75300401e-01 2.29978934e-01
1.05639684e+00 -4.62578125e-02 8.50108862e-02 -1.55371651e-01
-3.91277015e-01 4.97529417e-01 1.20551169e+00 8.42874229e-01
-4.27929997e-01 6.19066477e-01 6.00721264e+00 8.96547854e-01
-2.83251166e-01 1.93437174e-01 2.89722472e-01 -2.07195476e-01
-6.89377785e-01 4.81087893e-01 -7.32669175e-01 4.19309825e-01
1.03549290e+00 -5.22358596e-01 7.07239032e-01 9.42158163e-01
-3.81801099e-01 -3.83091390e-01 -5.35720050e-01 1.15455747e+00
-3.46837938e-02 -1.38836682e+00 -1.60644725e-01 8.11673462e-01
1.12413228e+00 -1.93177029e-01 -4.15814901e-03 -2.27141634e-01
7.73832560e-01 -7.40916729e-01 3.09988409e-01 -1.06734624e-02
7.50613689e-01 -1.03707945e+00 2.91948617e-01 3.34430814e-01
-1.41966879e+00 -1.19028397e-01 -3.17532450e-01 1.49605438e-01
-3.29108313e-02 9.19182539e-01 -5.58172822e-01 4.18114424e-01
4.96384054e-01 -9.46610495e-02 -1.62512511e-01 1.04815888e+00
9.92056280e-02 9.94076490e-01 -7.07551062e-01 1.03288434e-01
6.84985220e-02 -6.11237407e-01 9.60593164e-01 7.80981481e-01
5.24482965e-01 4.24186289e-01 7.67892599e-02 6.32578492e-01
-2.96438754e-01 3.28166455e-01 -6.36519969e-01 -4.58627164e-01
4.56276506e-01 1.11886191e+00 -1.10564101e+00 -6.90843537e-02
-1.32823423e-01 5.72368979e-01 2.82007515e-01 1.18533581e-01
-4.32309359e-01 -4.07911003e-01 3.60538542e-01 6.53695822e-01
7.80747771e-01 -3.22260201e-01 -1.84655815e-01 -1.25697792e+00
-1.94371417e-01 -9.66503084e-01 5.79335570e-01 -3.42425287e-01
-1.11749339e+00 4.65883672e-01 -2.17665762e-01 -4.12837565e-01
1.24557391e-01 -3.16636533e-01 -4.43774648e-02 6.21760905e-01
-6.63350940e-01 -5.67107320e-01 1.53129548e-01 5.26425183e-01
-6.23099566e-01 2.51820952e-01 6.71750069e-01 5.78878522e-02
-6.99766278e-01 3.02217990e-01 3.31170231e-01 -6.66212663e-02
-2.24432796e-02 -1.28775167e+00 1.01857111e-01 1.37609112e+00
1.09960936e-01 6.70989215e-01 1.00200582e+00 -8.66225481e-01
-2.05103135e+00 -6.79913044e-01 6.61044657e-01 1.26528829e-01
1.03166616e+00 -5.82840621e-01 -4.16887522e-01 6.78160846e-01
1.53116658e-01 -2.35126227e-01 8.24771225e-01 2.82727003e-01
-7.53327847e-01 -2.34327674e-01 -1.04167449e+00 4.07246113e-01
1.39182782e+00 -4.97798175e-01 2.31097966e-01 6.68060601e-01
4.16062683e-01 -9.18894336e-02 -6.73026919e-01 2.27680996e-01
1.18736990e-01 -7.59772003e-01 6.95066273e-01 -3.09380084e-01
2.25465953e-01 -3.42714131e-01 -3.03025097e-01 -8.62670422e-01
-7.58138120e-01 -1.27159953e+00 -1.54026315e-01 5.97316861e-01
4.22335863e-01 -4.58661050e-01 1.12823033e+00 9.90577117e-02
5.39215863e-01 -2.33888745e-01 -1.25915718e+00 -6.97383225e-01
-8.47436562e-02 -2.16477767e-01 4.00027424e-01 7.94474304e-01
3.77713650e-01 4.92875129e-01 -3.86386395e-01 4.07646984e-01
1.39895749e+00 4.13918823e-01 6.54024005e-01 -1.37966943e+00
-8.18186641e-01 -4.90546525e-01 -3.97062451e-01 -6.19695008e-01
1.18322298e-01 -1.24157214e+00 -2.15197191e-01 -1.33676243e+00
1.13949561e+00 -5.55147946e-01 1.97163686e-01 -9.07727182e-02
1.31988928e-01 3.93676341e-01 1.37248427e-01 1.34251580e-01
-9.60868061e-01 -1.40128136e-02 1.00559080e+00 -9.15839896e-02
2.10846111e-01 5.57395928e-02 -9.17634666e-01 2.79591769e-01
6.70880020e-01 -6.88759804e-01 -1.71631500e-01 1.27336919e-01
1.21596372e+00 4.68601048e-01 8.79949331e-02 -6.45456553e-01
3.21592450e-01 -2.90053040e-01 -6.03430212e-01 -5.16162217e-01
-1.00827612e-01 -5.37630022e-01 7.50630379e-01 9.51136053e-01
-1.67164102e-01 3.57201807e-02 -3.12577963e-01 1.03377664e+00
3.75065714e-01 -3.10298681e-01 7.33256578e-01 -1.27865031e-01
1.20109454e-01 4.93882090e-01 -4.92037982e-01 2.96968162e-01
1.05655169e+00 -5.89198992e-02 -6.25991523e-01 -6.17745757e-01
-6.68925643e-01 3.50389397e-03 4.79809135e-01 -6.40032232e-01
2.50446141e-01 -1.07084739e+00 -7.85403550e-01 1.14288032e-01
1.03754692e-01 -2.48676881e-01 6.07054412e-01 1.12491059e+00
-6.96890950e-01 4.27493244e-01 3.02737176e-01 -5.22238612e-01
-1.42344999e+00 9.24904466e-01 1.80077180e-01 -4.63033855e-01
-3.00603658e-01 1.02739775e+00 7.94430226e-02 2.41928935e-01
-2.94882264e-02 3.08037668e-01 6.18427396e-01 -2.67108798e-01
3.46545011e-01 6.59511983e-01 -5.06930426e-02 -7.94276595e-01
-3.71232808e-01 3.57289791e-01 -8.69565159e-02 -3.91845524e-01
1.11690998e+00 -2.06921816e-01 -9.38714325e-01 -9.36404839e-02
9.96305168e-01 7.20139205e-01 -4.32543486e-01 -3.24038446e-01
-2.00220615e-01 -5.35833955e-01 -1.43443882e-01 -1.28453849e-02
-1.40972090e+00 2.52970219e-01 1.74755938e-02 9.52288330e-01
7.18972266e-01 5.31366229e-01 5.68002462e-01 3.79139602e-01
7.65387475e-01 -8.36658001e-01 -4.47793193e-02 3.91978800e-01
4.53816980e-01 -3.49601209e-01 -1.47820590e-02 -8.76139522e-01
-9.44532156e-02 7.63930798e-01 3.22831124e-01 -2.96901435e-01
9.38780904e-01 1.89788312e-01 -8.45409632e-01 -4.48352575e-01
-7.55466819e-01 -4.87489074e-01 -2.73095161e-01 2.82406658e-01
-6.54531345e-02 4.80315804e-01 -7.10516095e-01 2.10999995e-01
-3.34962010e-01 -5.31751573e-01 9.63819385e-01 4.81780440e-01
-5.37403584e-01 -1.27991581e+00 -4.59190875e-01 5.31125605e-01
-4.14690346e-01 -3.82652283e-01 -5.59634745e-01 3.97704273e-01
-2.53846914e-01 9.16359425e-01 -2.72225410e-01 -4.46383595e-01
-1.93805873e-01 -2.73224741e-01 7.87018239e-01 -8.61054957e-01
-1.80896491e-01 1.11477889e-01 -5.69194593e-02 -2.42284462e-01
-2.82516688e-01 -6.26131654e-01 -8.06441545e-01 -1.10387671e+00
-6.46301270e-01 7.33120620e-01 2.65205562e-01 4.16292727e-01
3.90334517e-01 -5.26694544e-02 9.10786152e-01 -4.45984975e-02
-6.08674824e-01 -7.43792415e-01 -1.37979031e+00 2.28983507e-01
-1.25109330e-01 -3.45706910e-01 -1.03372324e+00 -5.31528354e-01] | [6.854725360870361, 5.132497787475586] |
8ddfceb3-aa94-4c54-adaf-756ce8d352d4 | instance-weighted-central-similarity-for | 2108.05274 | null | https://arxiv.org/abs/2108.05274v5 | https://arxiv.org/pdf/2108.05274v5.pdf | Instance-weighted Central Similarity for Multi-label Image Retrieval | Deep hashing has been widely applied to large-scale image retrieval by encoding high-dimensional data points into binary codes for efficient retrieval. Compared with pairwise/triplet similarity based hash learning, central similarity based hashing can more efficiently capture the global data distribution. For multi-label image retrieval, however, previous methods only use multiple hash centers with equal weights to generate one centroid as the learning target, which ignores the relationship between the weights of hash centers and the proportion of instance regions in the image. To address the above issue, we propose a two-step alternative optimization approach, Instance-weighted Central Similarity (ICS), to automatically learn the center weight corresponding to a hash code. Firstly, we apply the maximum entropy regularizer to prevent one hash center from dominating the loss function, and compute the center weights via projection gradient descent. Secondly, we update neural network parameters by standard back-propagation with fixed center weights. More importantly, the learned center weights can well reflect the proportion of foreground instances in the image. Our method achieves the state-of-the-art performance on the image retrieval benchmarks, and especially improves the mAP by 1.6%-6.4% on the MS COCO dataset. | ['Hanyu Peng', 'Zhiwei Zhang'] | 2021-08-11 | null | null | null | null | ['multi-label-image-retrieval'] | ['computer-vision'] | [-8.43074620e-02 -4.68930334e-01 -4.51755822e-01 -4.13748711e-01
-1.25260055e+00 -3.64076704e-01 3.96487266e-01 6.20809257e-01
-6.47393107e-01 2.85208791e-01 3.58304530e-02 1.91188768e-01
-1.58937529e-01 -8.63989055e-01 -7.06959724e-01 -1.14618862e+00
-1.28367422e-02 6.67537153e-01 2.76282132e-01 1.58859998e-01
4.66392398e-01 4.06728804e-01 -1.48860884e+00 1.42228633e-01
4.23561931e-01 1.39562345e+00 9.02583599e-02 2.92548388e-01
-7.70992716e-04 5.27879119e-01 -6.08333826e-01 -2.78366625e-01
3.47901374e-01 -3.63831148e-02 -5.96323550e-01 -3.44880015e-01
5.62394261e-01 -6.04773700e-01 -7.01147854e-01 1.16099799e+00
8.14535975e-01 2.28126481e-01 7.96010852e-01 -1.23739159e+00
-9.34156835e-01 3.55169564e-01 -8.91494632e-01 1.15187079e-01
2.73824278e-02 -8.89714956e-02 1.32772911e+00 -8.72242093e-01
4.25303966e-01 1.18482959e+00 4.30491149e-01 2.13018730e-01
-1.00681376e+00 -9.56405222e-01 -1.96102738e-01 4.65145022e-01
-1.97643435e+00 1.28117474e-02 6.32210314e-01 -1.43734500e-01
6.55319750e-01 9.12478417e-02 5.39760113e-01 4.85301226e-01
1.03236020e-01 9.62863982e-01 8.32717955e-01 -1.68312386e-01
-1.01734940e-02 7.18167564e-03 5.63801527e-02 7.64051497e-01
1.35066926e-01 -1.27330065e-01 -3.75774294e-01 -3.69515836e-01
4.89206880e-01 5.43295085e-01 -1.33664101e-01 -3.71057272e-01
-1.35789835e+00 1.19230640e+00 1.08018792e+00 1.75827146e-01
-3.37345093e-01 5.06171346e-01 5.32434762e-01 -1.45262564e-02
2.50155836e-01 1.47979379e-01 -8.99201855e-02 3.34633857e-01
-1.09588313e+00 4.44122881e-01 3.77618849e-01 7.56774604e-01
1.18406975e+00 -7.04403877e-01 -6.23352885e-01 1.02946603e+00
5.59631467e-01 6.85943305e-01 6.61014259e-01 -7.40302384e-01
4.02847797e-01 6.13109827e-01 -1.95669845e-01 -1.30509210e+00
-6.56322241e-02 -2.77488619e-01 -8.44102800e-01 -3.59152645e-01
1.21360123e-01 4.27892596e-01 -9.08267260e-01 1.62907481e+00
4.78389472e-01 2.03439698e-01 -3.45662653e-01 1.07918799e+00
7.18552887e-01 9.58053589e-01 -9.41800699e-02 3.74322757e-02
1.40642381e+00 -1.08730602e+00 -5.52707493e-01 1.48989912e-02
5.62103510e-01 -8.22426856e-01 7.57989645e-01 2.03389581e-03
-9.17886734e-01 -2.28307977e-01 -1.11458409e+00 -3.39932770e-01
-6.35555446e-01 -1.03688397e-01 3.98217171e-01 2.29922175e-01
-8.24086130e-01 4.85959381e-01 -4.81803417e-01 1.59301236e-01
4.57886964e-01 3.84032607e-01 -1.33420408e-01 -4.24288154e-01
-1.47555149e+00 6.81483388e-01 5.11974335e-01 -1.77732080e-01
-8.84856284e-01 -6.49731874e-01 -8.91030192e-01 3.14966410e-01
-8.65257531e-03 -3.94914299e-01 9.46363389e-01 -2.73418158e-01
-9.74165440e-01 9.55835104e-01 -5.55882528e-02 -2.84160972e-01
1.64757028e-01 -2.98603289e-02 -1.24347508e-01 3.66012543e-01
2.79199809e-01 1.16986060e+00 8.57335210e-01 -1.09949481e+00
-6.45201981e-01 -3.44383508e-01 -1.83087647e-01 2.33067617e-01
-6.72130048e-01 -5.54930046e-03 -8.48353744e-01 -6.15514457e-01
2.12142810e-01 -1.01646459e+00 1.95769165e-02 3.44866455e-01
-1.98313683e-01 -4.74040657e-01 6.73621893e-01 -3.68029505e-01
1.51506126e+00 -2.27033591e+00 9.20475721e-02 4.61234093e-01
1.86157107e-01 9.24056396e-02 -2.67032713e-01 1.68024734e-01
2.26629436e-01 -2.70144999e-01 -2.10111946e-01 -4.20461625e-01
2.91042715e-01 1.92086056e-01 -3.36564422e-01 6.35779262e-01
5.70160896e-02 8.84730756e-01 -8.57097030e-01 -7.84462154e-01
5.73329115e-03 7.06232965e-01 -5.63993096e-01 1.90716967e-01
7.95632228e-02 -3.56072158e-01 -3.39738756e-01 6.54805362e-01
8.37638974e-01 -5.52419782e-01 -2.02479795e-01 -3.52095544e-01
1.68118477e-01 2.12946132e-01 -1.26146436e+00 1.70343518e+00
-2.67648995e-01 3.46273333e-01 -2.72846878e-01 -8.47493112e-01
9.12092686e-01 -5.51946983e-02 6.78258836e-01 -8.42553139e-01
6.27024993e-02 3.11169863e-01 -3.06664497e-01 -1.00608796e-01
6.52718246e-01 4.17261086e-02 -2.14757279e-01 6.19410038e-01
-7.64349625e-02 8.27389359e-02 4.01648469e-02 1.93131000e-01
7.34689534e-01 -4.27809715e-01 3.50049920e-02 -7.94662833e-02
5.66385210e-01 -3.17976862e-01 5.41353047e-01 7.31886268e-01
-2.02447936e-01 9.97089803e-01 2.29449004e-01 -5.52509069e-01
-1.05363941e+00 -9.19197023e-01 -3.70694250e-01 1.20712721e+00
6.13815486e-01 -3.97571772e-01 -5.45931339e-01 -5.74115157e-01
3.69628400e-01 -9.04780626e-03 -5.91919959e-01 -5.82131088e-01
-5.57920098e-01 -9.34447467e-01 5.79846680e-01 2.46690065e-01
4.54863131e-01 -1.11122465e+00 -3.67102861e-01 1.33235857e-01
-3.55886400e-01 -8.03614974e-01 -1.12903380e+00 1.77226931e-01
-4.15646076e-01 -8.42864275e-01 -1.21168351e+00 -9.80242848e-01
6.81543529e-01 5.54019868e-01 9.22291160e-01 4.15412873e-01
-6.40858114e-01 3.69048715e-02 -2.21764714e-01 4.53023054e-02
1.14802249e-01 3.28100383e-01 -2.24881902e-01 4.67247143e-02
4.76543844e-01 -1.57689825e-01 -1.16281319e+00 4.17345971e-01
-1.25028634e+00 -4.87764955e-01 5.79899669e-01 9.54704702e-01
1.00947702e+00 -2.27133278e-02 1.97049618e-01 -4.37629014e-01
4.94022578e-01 -5.34630775e-01 -5.17927051e-01 4.47599500e-01
-8.21014583e-01 3.13468575e-01 2.51989007e-01 -6.34781599e-01
-1.65125221e-01 -1.02370173e-01 4.38583679e-02 -6.68521106e-01
3.38492185e-01 4.24926311e-01 1.76086217e-01 -3.08486015e-01
2.97561318e-01 3.86539131e-01 -5.03672697e-02 -1.18332162e-01
4.21533376e-01 7.55902171e-01 2.52374083e-01 -5.47222912e-01
7.31891870e-01 4.21205997e-01 -9.31732915e-03 -1.67081892e-01
-9.05258358e-01 -9.19276595e-01 -2.41484344e-01 1.17353790e-01
7.11662233e-01 -1.05665028e+00 -7.72655666e-01 5.68605661e-01
-1.02880621e+00 -2.47027073e-02 5.63375242e-02 3.63925368e-01
-3.14995348e-01 4.86300528e-01 -8.77777576e-01 -2.96944827e-01
-7.47282863e-01 -1.34431243e+00 1.43930840e+00 2.72300631e-01
2.42086127e-01 -6.21054649e-01 2.33081952e-01 1.71383485e-01
5.19281864e-01 -1.90002874e-01 9.62844551e-01 -4.60943371e-01
-7.43033409e-01 -4.93225336e-01 -6.29420877e-01 8.34643543e-02
-1.23007655e-01 -8.44745114e-02 -7.07464874e-01 -5.91854692e-01
-3.82480413e-01 -5.81168473e-01 1.02750826e+00 3.16739351e-01
1.43497348e+00 -3.48580182e-01 -3.50992501e-01 6.45091355e-01
1.61215639e+00 -7.03783110e-02 6.93547726e-01 5.09024441e-01
6.57896459e-01 3.05516809e-01 7.13907778e-01 5.91243684e-01
6.81895196e-01 8.89824450e-01 4.57801282e-01 2.42131576e-02
5.52404625e-03 -2.67224282e-01 3.70552810e-03 8.46557438e-01
5.47628880e-01 -6.95701241e-02 -7.69604385e-01 6.30575955e-01
-1.83154225e+00 -7.71962106e-01 3.89049172e-01 2.43807268e+00
1.06007612e+00 -1.72375031e-02 3.30499895e-02 -7.32372478e-02
1.09112251e+00 4.65400606e-01 -5.82247138e-01 1.00990966e-01
5.85007155e-03 -2.10608579e-02 7.56526113e-01 3.66393209e-01
-1.23126066e+00 7.42975414e-01 5.71123695e+00 1.31036127e+00
-1.20491469e+00 4.66396734e-02 6.12698257e-01 -3.84430319e-01
-2.29182333e-01 -1.56445563e-01 -1.11002028e+00 7.82645643e-01
6.46042168e-01 1.58028513e-01 5.98837912e-01 7.69632101e-01
-5.25068462e-01 1.60322711e-01 -8.18276286e-01 1.27272642e+00
2.22819060e-01 -1.19790852e+00 3.90464693e-01 1.83583528e-01
7.58613467e-01 2.48543084e-01 3.80513370e-01 3.66450578e-01
4.26633470e-02 -6.97411895e-01 7.00757921e-01 2.62969315e-01
7.50575423e-01 -9.50126052e-01 7.31424093e-01 1.11441769e-01
-1.37057161e+00 -1.53555945e-01 -6.95271432e-01 5.95176518e-01
-3.94702852e-02 6.08067274e-01 -4.74967748e-01 1.00018188e-01
8.39530110e-01 4.75028068e-01 -6.53023720e-01 1.39839005e+00
5.15276305e-02 -4.96460684e-03 -5.09626985e-01 -4.93726060e-02
4.84965444e-01 -1.11890566e-02 6.01212829e-02 1.18212223e+00
3.84390086e-01 -2.28079289e-01 2.43960336e-01 5.43049097e-01
-4.38264579e-01 1.92353547e-01 -2.70743787e-01 6.90115541e-02
8.51186752e-01 1.39744520e+00 -6.07851028e-01 -3.18593532e-01
-3.22532922e-01 8.52481186e-01 4.64083701e-01 7.37609044e-02
-9.20224071e-01 -8.33872974e-01 5.63786209e-01 -8.12168196e-02
6.12660408e-01 3.38807814e-02 2.48722121e-01 -8.76167595e-01
3.99776660e-02 -8.29371691e-01 7.10046947e-01 -5.56575835e-01
-1.29106748e+00 3.39986265e-01 -1.17677920e-01 -1.29757452e+00
-1.72423422e-02 -2.99226671e-01 -3.07554632e-01 8.04077089e-01
-1.91881788e+00 -9.09979403e-01 -3.69604737e-01 5.57926953e-01
-2.00562999e-02 1.08720968e-02 7.41493106e-01 8.16200137e-01
-5.14962316e-01 1.04462004e+00 4.62652624e-01 2.05343917e-01
1.17739499e+00 -1.15130699e+00 1.67356133e-01 2.02120215e-01
1.61449358e-01 7.93111503e-01 2.26664215e-01 -1.92650482e-01
-1.31667829e+00 -9.58837688e-01 9.03433383e-01 -1.75921712e-02
6.41748905e-01 -2.64298230e-01 -1.16908801e+00 2.38099396e-01
-3.01091354e-02 6.16355479e-01 6.72666430e-01 -2.18468025e-01
-9.15019393e-01 -5.18877447e-01 -1.20777166e+00 3.03896993e-01
4.67734724e-01 -8.73264432e-01 -2.20786616e-01 4.76480544e-01
8.43730688e-01 -3.56830090e-01 -1.14683449e+00 4.31105465e-01
7.20406115e-01 -5.28761089e-01 1.29822826e+00 -2.54863888e-01
5.01746476e-01 -4.90839660e-01 -3.94627690e-01 -8.89167011e-01
-5.43074846e-01 -1.28547475e-01 -2.84874231e-01 1.06207931e+00
2.18105361e-01 -4.72572327e-01 6.81955993e-01 3.68455172e-01
3.66723776e-01 -1.12737858e+00 -9.10804510e-01 -4.51562136e-01
1.02928542e-01 2.36872241e-01 8.39765429e-01 8.38398218e-01
-2.21953630e-01 6.82150424e-02 -3.17540050e-01 1.31258979e-01
7.97570765e-01 4.97921675e-01 4.00961220e-01 -1.04559994e+00
-1.44724563e-01 -6.22335792e-01 -5.00141740e-01 -1.18390405e+00
2.07389325e-01 -1.00671744e+00 2.38157004e-01 -1.25575733e+00
7.27177560e-01 -7.08068192e-01 -9.30913925e-01 5.07451296e-01
-4.60757852e-01 6.73851907e-01 1.91319183e-01 5.78944743e-01
-1.15461159e+00 6.72438025e-01 1.16561186e+00 -4.63409603e-01
2.72112072e-01 -3.26974183e-01 -5.55444419e-01 1.09689422e-01
4.66941059e-01 -8.90341580e-01 -7.81814307e-02 -4.80945170e-01
3.73065025e-01 -1.77743793e-01 2.85141498e-01 -8.34083080e-01
6.74253166e-01 1.03792526e-01 3.58152479e-01 -7.10689008e-01
2.72497535e-01 -8.75210226e-01 -1.70817494e-01 5.70603371e-01
-6.04731560e-01 1.07241109e-01 -1.63575262e-01 5.54424345e-01
-4.80364084e-01 -4.99449998e-01 8.24717939e-01 -3.95500809e-02
-3.53505820e-01 8.01566005e-01 5.87647296e-02 4.46666107e-02
7.79925108e-01 8.78754780e-02 -4.81984407e-01 -1.23442017e-01
-3.00857034e-02 5.32409668e-01 4.31526601e-01 4.34298396e-01
7.20867991e-01 -1.80929959e+00 -5.79718173e-01 1.55945674e-01
3.82407367e-01 1.47446722e-01 1.59203976e-01 5.35530388e-01
-5.01961291e-01 4.55285341e-01 -9.21193231e-03 -6.47572339e-01
-1.12178898e+00 7.22023487e-01 1.90598220e-01 -2.97663629e-01
-4.20845091e-01 8.28835666e-01 1.79751158e-01 -4.22274768e-01
5.41521013e-01 2.23430023e-01 2.47029588e-02 2.56215602e-01
7.38677204e-01 2.77675867e-01 -7.18436623e-03 -5.79944849e-01
-3.33389044e-01 8.81978810e-01 -6.44440830e-01 7.29956329e-02
1.20690668e+00 -5.28576784e-02 -4.28709060e-01 2.46091738e-01
2.02578306e+00 -5.12991309e-01 -1.05755305e+00 -6.11291826e-01
-1.71345830e-01 -5.08105099e-01 2.87334114e-01 -2.94277549e-01
-1.25270367e+00 8.42708349e-01 7.82979429e-01 3.76725160e-02
9.90436435e-01 1.04793847e-01 1.29687572e+00 5.60340762e-01
1.92052454e-01 -1.11467528e+00 2.79216737e-01 3.82446319e-01
5.51116049e-01 -1.32846081e+00 1.91843480e-01 6.31285459e-02
-3.90252948e-01 8.14281821e-01 3.67602199e-01 -1.76016659e-01
6.03455961e-01 2.34942846e-02 -6.17299639e-02 -3.08565468e-01
-5.30255973e-01 6.00707680e-02 4.61073995e-01 -6.38716621e-03
1.24154642e-01 1.08349593e-02 -2.49497101e-01 -6.37695715e-02
2.23065406e-01 -3.42898399e-01 -2.06189454e-01 8.17668021e-01
-6.39042974e-01 -9.34564769e-01 -4.29379880e-01 5.13420403e-01
-5.73072255e-01 -2.96711266e-01 1.55699123e-02 3.21006000e-01
-5.88853471e-02 3.44893336e-01 2.20487550e-01 -2.73323536e-01
-2.74745803e-02 -4.54100855e-02 3.74626040e-01 -2.68563181e-01
-5.86671889e-01 7.91134462e-02 -8.79475176e-01 -3.97658706e-01
-2.19574630e-01 -4.56073880e-01 -1.32330036e+00 -3.07668626e-01
-6.82421327e-01 1.86657220e-01 6.85641885e-01 6.13352835e-01
2.81228215e-01 1.18756138e-01 9.89307404e-01 -7.17960835e-01
-8.05594862e-01 -7.30195105e-01 -5.09201288e-01 5.99287868e-01
3.97102296e-01 -5.80036163e-01 -5.26250124e-01 -4.40216064e-01] | [11.310572624206543, 0.9475767612457275] |
bba14efc-880c-44e4-9691-1bc104cbc524 | bertgen-multi-task-generation-through-bert | 2106.03484 | null | https://arxiv.org/abs/2106.03484v1 | https://arxiv.org/pdf/2106.03484v1.pdf | BERTGEN: Multi-task Generation through BERT | We present BERTGEN, a novel generative, decoder-only model which extends BERT by fusing multimodal and multilingual pretrained models VL-BERT and M-BERT, respectively. BERTGEN is auto-regressively trained for language generation tasks, namely image captioning, machine translation and multimodal machine translation, under a multitask setting. With a comprehensive set of evaluations, we show that BERTGEN outperforms many strong baselines across the tasks explored. We also show BERTGEN's ability for zero-shot language generation, where it exhibits competitive performance to supervised counterparts. Finally, we conduct ablation studies which demonstrate that BERTGEN substantially benefits from multi-tasking and effectively transfers relevant inductive biases from the pre-trained models. | ['Lucia Specia', 'Pranava Madhyastha', 'Ozan Caglayan', 'Faidon Mitzalis'] | 2021-06-07 | null | https://aclanthology.org/2021.acl-long.503 | https://aclanthology.org/2021.acl-long.503.pdf | acl-2021-5 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 3.07391346e-01 4.50663537e-01 -2.98495412e-01 -1.93071038e-01
-1.81289244e+00 -4.79783535e-01 1.23491549e+00 -5.68097711e-01
-3.45900267e-01 1.14613032e+00 5.05871654e-01 -4.52707559e-01
8.17514777e-01 -3.87358904e-01 -1.23636293e+00 -3.86644304e-01
2.75048286e-01 1.02779305e+00 -4.81753170e-01 -3.01208436e-01
-4.91722196e-01 -2.06554204e-01 -7.75078118e-01 6.09573543e-01
8.68149638e-01 5.54420888e-01 1.79535151e-01 9.46436524e-01
6.09427132e-02 8.59176278e-01 -5.60445070e-01 -1.18477130e+00
-1.03141472e-01 -6.43515766e-01 -8.74523044e-01 5.93412742e-02
3.77804279e-01 -5.78882337e-01 -3.17381024e-01 6.59503162e-01
8.06158900e-01 -6.48822263e-02 1.05847049e+00 -1.40385103e+00
-1.18381751e+00 8.21605504e-01 -4.60129619e-01 -1.69220120e-01
4.02182758e-01 4.93583769e-01 1.00806534e+00 -1.28809130e+00
8.33101869e-01 1.77562630e+00 3.51487130e-01 1.09021175e+00
-1.50622189e+00 -6.53259814e-01 6.66425750e-02 -1.68882325e-01
-1.13014257e+00 -7.33044088e-01 3.23351622e-01 -3.65609348e-01
1.14930964e+00 -6.81907833e-02 1.03763089e-01 2.12867689e+00
3.27508062e-01 1.21003079e+00 1.10624552e+00 -4.39038754e-01
-2.19088063e-01 1.04648070e-02 -5.12357950e-01 7.54418731e-01
-1.30001634e-01 8.64768252e-02 -6.49359941e-01 -1.50932267e-01
6.52126133e-01 -6.81959569e-01 -1.13294214e-01 -7.10648745e-02
-1.64979553e+00 8.84259582e-01 2.03196481e-01 -1.04727641e-01
-2.35806897e-01 7.62250245e-01 4.71413374e-01 3.66243511e-01
7.63331056e-01 4.07176018e-01 -1.82553634e-01 -6.82383329e-02
-6.13980651e-01 8.53419676e-02 6.39309049e-01 1.66314435e+00
7.58744895e-01 4.32143390e-01 -8.89958739e-01 8.48384559e-01
1.97689965e-01 1.12230027e+00 4.99876291e-01 -5.48038125e-01
8.16201270e-01 1.38323812e-03 5.52780516e-02 -1.05594620e-01
1.38767883e-01 -3.17204952e-01 -8.86993945e-01 -4.92771238e-01
-1.62683025e-01 -7.28343427e-01 -1.33096373e+00 2.09087682e+00
-2.08987072e-01 4.81268391e-02 4.95399684e-01 5.43386459e-01
1.17764699e+00 9.23168898e-01 5.41738629e-01 9.60767716e-02
1.13878739e+00 -1.51316833e+00 -6.03313267e-01 -6.43389702e-01
4.77458239e-01 -8.88402045e-01 1.20185101e+00 -4.84579690e-02
-1.30371439e+00 -4.70132530e-01 -6.90841079e-01 -2.69721746e-01
-1.80274084e-01 4.08672035e-01 7.72030532e-01 1.63664818e-01
-1.55239844e+00 -8.82232487e-02 -6.80358708e-01 -2.77840942e-01
3.04178476e-01 2.01145187e-01 -3.27040583e-01 -1.82211116e-01
-1.28558469e+00 1.11708498e+00 3.08367342e-01 9.12349299e-03
-1.98256576e+00 -1.80440858e-01 -1.29427934e+00 -1.97449610e-01
7.65309557e-02 -1.28259051e+00 1.77410996e+00 -1.02999318e+00
-1.54199362e+00 1.00361621e+00 -4.54634070e-01 -6.07663035e-01
6.18737996e-01 -2.71193177e-01 -4.16081518e-01 1.04549229e-01
2.74082750e-01 1.64343965e+00 8.97488356e-01 -1.36401963e+00
-1.39079556e-01 3.50886434e-01 -1.90866403e-02 4.79894280e-01
-1.35933369e-01 4.71281223e-02 -6.88415766e-01 -5.71213186e-01
-7.81942546e-01 -1.02589917e+00 -3.31703097e-01 -7.44516075e-01
-8.60601783e-01 -1.72193214e-01 3.22812766e-01 -8.54380608e-01
5.99148393e-01 -1.83061755e+00 5.38497210e-01 -3.79982769e-01
2.23200507e-02 2.54640356e-02 -9.03831601e-01 7.15846062e-01
8.85843337e-02 1.89231515e-01 -2.23494634e-01 -1.03717458e+00
2.06114873e-01 1.38358429e-01 -6.60764933e-01 -1.22232825e-01
5.40093422e-01 1.95338094e+00 -8.58922184e-01 -5.11121511e-01
-2.98321191e-02 4.87104118e-01 -2.22070903e-01 5.57387173e-01
-7.45678186e-01 5.24609745e-01 -1.96056962e-01 6.72870874e-01
3.28584015e-01 -4.86390769e-01 1.86189283e-02 2.37923791e-03
4.12935883e-01 2.25916088e-01 6.69567287e-02 2.06060863e+00
-8.57904792e-01 6.52165592e-01 -6.05501421e-02 -4.70397949e-01
6.57563627e-01 8.18564773e-01 -1.36913940e-01 -7.90110111e-01
-1.86957438e-02 3.60147774e-01 -2.93758482e-01 -2.49376655e-01
5.51981091e-01 -2.76537508e-01 -5.15829265e-01 6.22034609e-01
6.86604440e-01 -2.74042130e-01 5.93434088e-02 7.05507457e-01
7.28389442e-01 4.26856130e-01 6.58046380e-02 2.82889307e-02
1.24032564e-01 3.91195975e-02 -3.29522826e-02 8.91361356e-01
1.44133136e-01 5.92365444e-01 2.89274901e-01 1.52257308e-01
-1.10647321e+00 -1.42796838e+00 4.45888788e-01 1.35612154e+00
-4.70153764e-02 -2.74071604e-01 -8.29116762e-01 -8.05774808e-01
-2.03047737e-01 1.07878482e+00 -5.73293626e-01 -2.91911006e-01
-3.80108416e-01 -8.19653571e-01 1.05207801e+00 4.14969742e-01
3.60509396e-01 -1.19051921e+00 3.59725952e-03 1.62030071e-01
-6.58333361e-01 -1.40317547e+00 -6.26296759e-01 -1.20764747e-01
-6.50624216e-01 -4.00719464e-01 -1.27430749e+00 -9.80274320e-01
6.64721489e-01 1.44728810e-01 1.71592855e+00 -3.35397303e-01
-5.18013202e-02 6.73195779e-01 -2.21586660e-01 -4.94620234e-01
-9.00799394e-01 4.98213232e-01 -4.31470156e-01 -1.91355705e-01
-1.02501111e-02 -2.54040688e-01 -3.01376969e-01 8.53080079e-02
-6.93694890e-01 8.54039669e-01 1.03812313e+00 1.10905337e+00
3.51026893e-01 -1.06897438e+00 9.79813039e-01 -8.36580455e-01
9.84549642e-01 -4.94768500e-01 -2.18670145e-01 5.76238394e-01
-2.58737713e-01 1.17243662e-01 3.90100539e-01 -5.34485459e-01
-1.24563289e+00 -2.21147388e-01 -1.85824007e-01 -3.95740211e-01
-6.32687137e-02 3.69975984e-01 -2.11735263e-01 2.73525357e-01
5.45243382e-01 4.76009458e-01 -1.71868071e-01 -1.90118387e-01
1.03185868e+00 6.07281327e-01 7.81665921e-01 -7.51152456e-01
8.27747345e-01 2.61956334e-01 -4.62774843e-01 -3.88834924e-01
-8.02659452e-01 -1.25043455e-03 -2.90832609e-01 -1.18745111e-01
1.07993770e+00 -1.68878007e+00 -1.90776110e-01 3.85100812e-01
-1.69576120e+00 -8.52099121e-01 -1.16095766e-02 2.44267642e-01
-8.07352185e-01 -2.20746957e-02 -9.96650815e-01 -7.19250739e-01
-8.06726933e-01 -1.22522342e+00 1.73107493e+00 -9.91538540e-02
-8.10525119e-02 -1.34518468e+00 1.68123379e-01 4.44184601e-01
3.73192102e-01 1.73676327e-01 1.04106486e+00 -4.95739847e-01
-6.92701519e-01 1.54969737e-01 -2.56940335e-01 1.79163903e-01
-1.43514767e-01 -3.72017175e-01 -1.06415057e+00 -4.78257269e-01
-6.00528300e-01 -1.08754766e+00 1.12724543e+00 2.17059180e-01
5.87902784e-01 -4.39688593e-01 -4.73881125e-01 5.70969224e-01
1.08488429e+00 -1.22405365e-02 6.73163950e-01 2.27495469e-02
8.54066789e-01 3.56505126e-01 2.12027237e-01 -9.01518390e-02
7.34470904e-01 5.44242740e-01 2.76414692e-01 -6.76938117e-01
-4.78090197e-01 -7.54096389e-01 8.42688203e-01 9.36834157e-01
2.33758390e-01 -7.26836681e-01 -7.48337984e-01 7.20061183e-01
-1.91324079e+00 -7.50974596e-01 1.25249729e-01 1.73995554e+00
1.14349198e+00 -2.64548004e-01 1.19460595e-03 -1.05398619e+00
7.45373428e-01 1.18368961e-01 -4.81957674e-01 -6.18392825e-01
-5.18549204e-01 2.33788878e-01 3.37598175e-01 6.53906524e-01
-1.00297236e+00 1.68712020e+00 7.65586805e+00 9.51542199e-01
-8.35793316e-01 6.50466263e-01 1.02222633e+00 -1.62134767e-01
-8.56211901e-01 -3.45929451e-02 -7.26510227e-01 1.21324994e-01
1.11750257e+00 -3.14009547e-01 6.41995966e-01 5.53467095e-01
-1.31050840e-01 2.16666088e-01 -1.51074862e+00 9.30561602e-01
5.86541653e-01 -1.18803680e+00 6.18141294e-01 1.03948042e-01
1.27680206e+00 4.87269104e-01 4.42659318e-01 8.33828330e-01
1.07184184e+00 -1.55343330e+00 8.19369555e-01 2.44543463e-01
1.31966174e+00 -5.66355169e-01 5.79943657e-01 3.14358741e-01
-6.90121770e-01 3.33218098e-01 -1.54909760e-01 2.64048100e-01
7.20139384e-01 2.41982132e-01 -1.26548803e+00 6.78144217e-01
-1.06510587e-01 5.28965175e-01 -6.62409186e-01 5.25310814e-01
-6.29183054e-01 6.63080037e-01 2.46251538e-01 1.48498580e-01
5.47411501e-01 3.91770713e-02 4.25298393e-01 1.52130497e+00
4.77377206e-01 -5.64403355e-01 1.85150117e-01 1.11707723e+00
-6.22147560e-01 -4.11452949e-02 -9.74244893e-01 -2.96136558e-01
1.00893982e-01 1.24717748e+00 -1.21437954e-02 -7.15863049e-01
-3.69216591e-01 1.56432378e+00 4.31005210e-01 9.44210351e-01
-1.02641904e+00 -1.04965037e-02 2.27081969e-01 -3.96667600e-01
6.74563721e-02 -2.64482886e-01 -1.24210957e-02 -1.44198012e+00
-4.21408892e-01 -1.10293067e+00 2.07402334e-01 -1.21238518e+00
-1.29901075e+00 9.98560190e-01 1.45864978e-01 -7.85596251e-01
-1.11777103e+00 -3.16598594e-01 -4.81736273e-01 1.08935666e+00
-1.50438511e+00 -1.90272629e+00 6.46439046e-02 5.62288105e-01
8.57179046e-01 -4.98227715e-01 1.01107216e+00 1.12533554e-01
-4.94815201e-01 8.23486209e-01 -1.31219611e-01 1.63621292e-01
9.80015755e-01 -1.17294347e+00 1.09280384e+00 9.50205505e-01
3.40659261e-01 4.49420214e-01 4.62034941e-01 -8.42140198e-01
-1.32164884e+00 -1.48726475e+00 9.94225144e-01 -4.88633126e-01
6.23692691e-01 -9.13950861e-01 -2.64085144e-01 1.29715121e+00
1.03629220e+00 -6.21018529e-01 5.77187479e-01 6.32796902e-03
-4.72236991e-01 3.77567768e-01 -6.69221997e-01 1.07366836e+00
1.06265640e+00 -7.75647461e-01 -3.97874922e-01 6.73764288e-01
1.06239951e+00 -3.95375937e-01 -4.59120572e-01 3.33610117e-01
2.38262206e-01 -3.37759137e-01 9.55851555e-01 -7.22512603e-01
8.83460045e-01 2.77519554e-01 -3.14859115e-02 -1.67722464e+00
-3.87517884e-02 -1.04612792e+00 -1.26038730e-01 1.23921132e+00
1.00933814e+00 -5.09256959e-01 2.19425157e-01 1.37934670e-01
-2.85250396e-01 -5.10494828e-01 -8.18338633e-01 -7.22101748e-01
3.75692099e-01 -2.77324259e-01 5.28992653e-01 5.23241580e-01
-2.34854057e-01 1.06927741e+00 -9.84307885e-01 -3.33649546e-01
5.19259214e-01 -7.11086020e-02 1.14099216e+00 -4.42160696e-01
-5.41481018e-01 -1.99789032e-01 3.01443309e-01 -1.39434683e+00
6.55693054e-01 -1.23001456e+00 3.40692788e-01 -1.92708194e+00
5.94605803e-01 -4.65316605e-03 7.24104717e-02 5.36027133e-01
-3.53059530e-01 6.75621331e-01 1.53779596e-01 3.45363468e-01
-9.03174102e-01 1.00107288e+00 1.54699564e+00 -4.03811216e-01
1.55126721e-01 -3.56212974e-01 -8.67749989e-01 3.26137364e-01
5.16933441e-01 -1.32367387e-01 -6.27107978e-01 -1.13538837e+00
1.37401879e-01 1.53995395e-01 4.88718808e-01 -4.92903322e-01
-1.22130394e-01 -8.26744456e-03 3.10763806e-01 -2.84710169e-01
6.11841977e-01 8.85111280e-03 1.21092781e-01 1.14074074e-01
-7.20579326e-01 3.88596565e-01 3.00396174e-01 4.10437793e-01
-2.36649275e-01 7.85428286e-02 4.05227929e-01 -3.04386228e-01
-4.95793730e-01 3.58690530e-01 -4.06679600e-01 2.76389778e-01
7.36610055e-01 3.96600187e-01 -6.10961497e-01 -9.93612945e-01
-5.79764009e-01 4.20156389e-01 3.10458630e-01 5.98364234e-01
7.00217783e-01 -1.64275110e+00 -1.13192177e+00 -1.89524934e-01
3.85687858e-01 -1.47856042e-01 3.96657251e-02 8.13506067e-01
-1.98845476e-01 6.83827579e-01 1.35675877e-01 -6.11055434e-01
-9.24623847e-01 3.80773574e-01 2.05433577e-01 -4.73123282e-01
-4.06260759e-01 9.18444455e-01 7.13062346e-01 -5.38229346e-01
2.80459318e-02 1.36355292e-02 2.26925880e-01 -3.87604743e-01
2.58755088e-01 -1.08251497e-01 -1.90019980e-01 -7.46227980e-01
-1.49731832e-02 6.84428867e-03 -2.55200267e-01 -1.00871336e+00
8.57079089e-01 -2.56895125e-01 5.67982011e-02 4.78169352e-01
1.03850389e+00 -3.13428611e-01 -1.30199599e+00 -7.74434432e-02
-2.46030003e-01 1.03082776e-01 -2.81031102e-01 -1.35840750e+00
-6.30769491e-01 1.07558525e+00 5.27216047e-02 -6.47158623e-01
9.25129354e-01 3.77749741e-01 1.02260137e+00 4.32056487e-01
4.60688919e-01 -8.30417335e-01 3.71177971e-01 6.21904969e-01
1.12656033e+00 -1.25902295e+00 -5.11286259e-01 -2.05234602e-01
-1.26568913e+00 4.01443332e-01 7.44669139e-01 2.54751474e-01
-2.04694718e-01 3.37732673e-01 2.10924864e-01 1.44683719e-01
-1.38666785e+00 -3.43378633e-01 5.62066436e-01 7.83415556e-01
6.68060124e-01 2.95768321e-01 1.63256414e-02 4.04281139e-01
-3.47080022e-01 -1.02705687e-01 2.38950774e-01 4.80750203e-01
2.67641600e-02 -1.14510608e+00 -1.70863390e-01 2.66598798e-02
-2.44643286e-01 -7.68030286e-01 -7.37612963e-01 7.09608316e-01
-1.78915456e-01 1.01650226e+00 1.37671102e-02 -3.42123955e-01
-1.67645827e-01 2.71772444e-01 7.41593421e-01 -7.31796265e-01
-5.01812935e-01 3.90871733e-01 5.02333283e-01 -1.95720360e-01
-1.30197927e-01 -2.59496480e-01 -8.20470810e-01 -3.52250114e-02
-3.84351611e-02 2.01440573e-01 6.09010816e-01 8.80178928e-01
5.58048427e-01 4.52315867e-01 3.49506497e-01 -7.42687404e-01
-5.94642580e-01 -1.30985320e+00 1.87257528e-01 4.04549241e-01
2.29976907e-01 -3.43267500e-01 3.34771816e-03 3.41737270e-01] | [11.30931282043457, 1.4399302005767822] |
f0aec122-d3b3-44a7-9931-2aad4d56a9b6 | video-mask-transfiner-for-high-quality-video | 2207.14012 | null | https://arxiv.org/abs/2207.14012v1 | https://arxiv.org/pdf/2207.14012v1.pdf | Video Mask Transfiner for High-Quality Video Instance Segmentation | While Video Instance Segmentation (VIS) has seen rapid progress, current approaches struggle to predict high-quality masks with accurate boundary details. Moreover, the predicted segmentations often fluctuate over time, suggesting that temporal consistency cues are neglected or not fully utilized. In this paper, we set out to tackle these issues, with the aim of achieving highly detailed and more temporally stable mask predictions for VIS. We first propose the Video Mask Transfiner (VMT) method, capable of leveraging fine-grained high-resolution features thanks to a highly efficient video transformer structure. Our VMT detects and groups sparse error-prone spatio-temporal regions of each tracklet in the video segment, which are then refined using both local and instance-level cues. Second, we identify that the coarse boundary annotations of the popular YouTube-VIS dataset constitute a major limiting factor. Based on our VMT architecture, we therefore design an automated annotation refinement approach by iterative training and self-correction. To benchmark high-quality mask predictions for VIS, we introduce the HQ-YTVIS dataset, consisting of a manually re-annotated test set and our automatically refined training data. We compare VMT with the most recent state-of-the-art methods on the HQ-YTVIS, as well as the Youtube-VIS, OVIS and BDD100K MOTS benchmarks. Experimental results clearly demonstrate the efficacy and effectiveness of our method on segmenting complex and dynamic objects, by capturing precise details. | ['Fisher Yu', 'Chi-Keung Tang', 'Yu-Wing Tai', 'Martin Danelljan', 'Henghui Ding', 'Lei Ke'] | 2022-07-28 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 2.56021857e-01 -3.52403164e-01 -2.82471627e-01 -1.08597897e-01
-7.85285950e-01 -5.22491634e-01 4.21812624e-01 -1.18524238e-01
-3.86136651e-01 7.42869377e-01 6.96476176e-02 1.88844487e-01
6.70180172e-02 -2.67354041e-01 -6.50439620e-01 -4.04661059e-01
-2.07335949e-02 5.64841926e-01 1.07366478e+00 -7.11042210e-02
1.19487951e-02 2.22271919e-01 -1.72207797e+00 5.66651344e-01
8.93972039e-01 1.17867720e+00 3.02704871e-01 6.10835254e-01
-1.22565717e-01 6.38315856e-01 -3.91970873e-01 -4.24372256e-01
3.99123400e-01 -3.75265718e-01 -9.29603398e-01 4.15097356e-01
8.64937186e-01 -1.99010059e-01 -3.24469298e-01 7.32506752e-01
9.51233804e-02 1.33182600e-01 5.83067060e-01 -1.19300056e+00
1.78179786e-01 4.37768936e-01 -9.05496538e-01 5.84940255e-01
3.32962841e-01 2.73592532e-01 8.77008259e-01 -9.83654797e-01
1.13696754e+00 8.72907639e-01 1.00098622e+00 4.55632955e-01
-1.40677154e+00 -4.65311021e-01 5.97555399e-01 3.50819647e-01
-1.59815454e+00 -5.51844358e-01 5.25884509e-01 -7.79313982e-01
9.23015058e-01 3.70408803e-01 9.29294169e-01 1.16697443e+00
7.80938864e-02 9.80777085e-01 8.79233062e-01 -9.99665707e-02
1.70581024e-02 -1.31783202e-01 -8.86812657e-02 7.03410029e-01
-8.30724686e-02 -1.13998223e-02 -7.15639532e-01 1.26989588e-01
7.37266183e-01 -1.28520370e-01 -5.86000383e-01 -4.88636732e-01
-1.49459445e+00 1.78563297e-01 8.96396562e-02 4.77098405e-01
-5.33498228e-01 1.20344400e-01 5.78006446e-01 3.72212529e-02
7.55979836e-01 1.78013563e-01 -6.04669809e-01 -5.06039202e-01
-1.72484696e+00 2.11408123e-01 6.08893156e-01 1.01148951e+00
7.31819808e-01 6.67375606e-03 -6.57495618e-01 6.82665884e-01
5.71647957e-02 6.94643706e-02 3.41598153e-01 -1.03611445e+00
5.15550196e-01 5.06380737e-01 1.15929812e-01 -9.13175762e-01
-2.61606187e-01 -3.88749212e-01 -6.75615728e-01 4.15683016e-02
4.98579770e-01 2.47876838e-01 -1.29563093e+00 1.37035525e+00
6.03026569e-01 8.55374396e-01 -4.02294159e-01 9.59624529e-01
7.52551079e-01 5.17516494e-01 1.41245335e-01 -4.31806028e-01
1.06862104e+00 -1.25527203e+00 -7.42520511e-01 5.73302293e-03
4.81908500e-01 -6.02866888e-01 9.90851879e-01 4.66521591e-01
-1.16158783e+00 -7.48000026e-01 -6.92345560e-01 1.93846181e-01
-6.82866294e-03 -3.64448912e-02 2.66167909e-01 3.10579956e-01
-9.42557633e-01 8.34129095e-01 -1.10865712e+00 -2.33636558e-01
7.54128337e-01 2.63452679e-01 -4.07525629e-01 -1.00966632e-01
-7.25354791e-01 3.98709506e-01 4.65587437e-01 1.63656250e-01
-1.03009534e+00 -9.28118527e-01 -8.51245999e-01 -8.80835205e-02
7.62117445e-01 -5.20743906e-01 1.05924428e+00 -1.03019714e+00
-1.18589532e+00 8.61095011e-01 -4.69654053e-01 -4.88824815e-01
1.00365508e+00 -1.73408508e-01 -4.22532171e-01 3.95281672e-01
2.56270617e-01 8.05095077e-01 1.02908874e+00 -1.40783465e+00
-9.97567356e-01 -2.66794534e-03 -3.00605744e-01 -3.25807999e-03
-1.89990237e-01 -6.98924586e-02 -1.19693840e+00 -9.44576085e-01
1.69972833e-02 -7.92051673e-01 -1.82566658e-01 -1.33936718e-01
-5.73584080e-01 -3.91477756e-02 1.11765552e+00 -6.90504014e-01
1.68776846e+00 -2.09231567e+00 3.50086629e-01 2.02143162e-01
4.79474992e-01 4.77446377e-01 -1.16018087e-01 1.72935985e-02
1.24084152e-01 1.11731552e-01 -4.21124488e-01 -6.97384596e-01
-1.47480175e-01 2.66447127e-01 -1.07010446e-01 4.41328734e-01
2.97498167e-01 9.89595771e-01 -9.66115057e-01 -9.04999495e-01
3.96704972e-01 3.82761359e-01 -5.90864360e-01 1.65125698e-01
-5.05530834e-01 7.23795414e-01 -3.23509157e-01 8.73428166e-01
5.09431362e-01 -3.79182726e-01 -2.42227893e-02 -4.63184357e-01
-1.91188037e-01 -1.58274285e-02 -1.29652762e+00 1.86758602e+00
1.31618798e-01 6.85943723e-01 1.13229834e-01 -8.03284407e-01
4.63079035e-01 2.85069615e-01 1.13317549e+00 -6.99946165e-01
5.91257662e-02 1.71147689e-01 -4.14852381e-01 -6.02926910e-01
6.56904459e-01 2.96414971e-01 2.78656870e-01 1.34944141e-01
1.16893306e-01 8.73241723e-02 6.48800552e-01 2.35734582e-01
1.25684834e+00 6.39312208e-01 5.91051504e-02 -1.81449920e-01
4.60217953e-01 2.94992983e-01 9.15618896e-01 5.22391737e-01
-5.11934936e-01 1.08072293e+00 3.34144235e-01 -4.51474339e-01
-9.10293639e-01 -7.20647573e-01 -1.74989209e-01 9.17093217e-01
4.85920995e-01 -8.74692976e-01 -9.57381129e-01 -1.05107522e+00
-5.88103347e-02 2.85999209e-01 -7.54610837e-01 2.57796317e-01
-7.81979024e-01 -3.73636752e-01 3.48471224e-01 4.77997243e-01
4.45324600e-01 -1.02712679e+00 -6.17414176e-01 4.54661965e-01
-4.17330533e-01 -1.54411530e+00 -7.18819559e-01 -6.17470555e-02
-7.14528084e-01 -1.16947854e+00 -7.13089168e-01 -6.26585186e-01
4.20825005e-01 3.67416292e-01 1.43445122e+00 3.22653115e-01
-3.47926825e-01 3.87433708e-01 -4.31231499e-01 1.28523871e-01
-2.89787650e-01 1.59462497e-01 -4.28654999e-02 2.65715659e-01
8.14598352e-02 -4.44963843e-01 -7.53311336e-01 8.34474564e-01
-9.18156147e-01 4.43648428e-01 1.94679305e-01 6.99740946e-01
1.21697330e+00 -3.67408693e-02 2.62809873e-01 -9.51260448e-01
-5.94225675e-02 -4.48237389e-01 -5.73420703e-01 2.68455803e-01
-4.73216742e-01 -3.34915429e-01 2.86689609e-01 -4.91212785e-01
-9.09197867e-01 3.21608096e-01 -1.34650469e-01 -9.06805873e-01
-2.10213467e-01 3.05734485e-01 2.66472772e-02 -1.24041833e-01
5.29723704e-01 1.19053207e-01 -3.58576387e-01 -5.08581758e-01
1.18711889e-01 2.86104679e-01 7.98331857e-01 -5.29160023e-01
7.75633454e-01 5.63738585e-01 -3.21031749e-01 -5.70519209e-01
-9.25728202e-01 -8.05401027e-01 -9.61428225e-01 -5.45620382e-01
9.65972006e-01 -9.85764086e-01 -2.46210620e-01 3.44078600e-01
-8.35775793e-01 -7.99903631e-01 -5.59021533e-01 2.22279832e-01
-7.83304393e-01 4.20603096e-01 -7.21369803e-01 -3.95616293e-01
-1.31811753e-01 -1.29677927e+00 1.43874252e+00 -6.74800947e-02
-4.33690459e-01 -7.68280327e-01 1.29335269e-01 4.21753764e-01
1.22075669e-01 5.98119020e-01 2.29697168e-01 -2.67123580e-01
-7.97933519e-01 1.34464785e-01 -2.37452522e-01 1.15601733e-01
7.56015480e-02 3.35033953e-01 -8.83184314e-01 -2.87288845e-01
-3.89100909e-01 -1.40294239e-01 1.11561835e+00 4.73557889e-01
1.44116783e+00 -5.18670306e-02 -6.06511533e-01 9.22930181e-01
1.30047357e+00 -1.45785615e-01 6.57663941e-01 3.20139796e-01
1.15458882e+00 3.23691219e-01 1.08932483e+00 4.25128490e-01
4.11936939e-01 1.09885490e+00 4.06170905e-01 -9.18690115e-02
-4.79154348e-01 -1.71171874e-01 2.08196521e-01 6.57786787e-01
-4.38452423e-01 -3.36138040e-01 -9.53865707e-01 7.94331491e-01
-2.25674629e+00 -9.96329546e-01 -3.73508781e-01 1.97813451e+00
8.13520908e-01 3.46915364e-01 3.45185965e-01 2.18956739e-01
6.56791687e-01 2.30299801e-01 -4.61966217e-01 2.77174205e-01
-1.42573372e-01 5.14412597e-02 3.05314183e-01 2.77883023e-01
-1.33908844e+00 1.24091756e+00 5.99318695e+00 1.15624666e+00
-1.06515205e+00 2.58167088e-01 7.54308164e-01 -2.36904338e-01
-1.30614504e-01 -2.39127979e-01 -7.51552463e-01 6.41483665e-01
6.93687916e-01 1.98215574e-01 2.56772518e-01 5.12416780e-01
4.37310725e-01 -1.76580116e-01 -1.09540880e+00 1.09546328e+00
-5.30726090e-02 -1.73011482e+00 -1.84673339e-01 -8.28019306e-02
1.19372547e+00 2.39675164e-01 -2.26018816e-01 2.18676388e-01
-1.47511467e-01 -9.57321942e-01 1.16280878e+00 4.65630770e-01
1.04574239e+00 -4.32303578e-01 6.11256599e-01 1.03891484e-01
-1.66878319e+00 2.04863220e-01 6.11080695e-03 2.81390429e-01
4.02238369e-01 5.77368915e-01 -6.19472682e-01 5.91192961e-01
1.04303229e+00 1.30046797e+00 -5.63451171e-01 1.38683414e+00
1.37553409e-01 7.48517752e-01 -3.63666445e-01 6.44053698e-01
2.65515953e-01 4.04355079e-02 5.82959652e-01 1.47989666e+00
2.42896855e-01 7.14076012e-02 5.46025515e-01 7.70151079e-01
4.73065069e-03 -1.24144796e-02 -1.31978825e-01 7.65659511e-02
1.89340413e-01 1.31852126e+00 -1.21445203e+00 -4.90884572e-01
-3.56566638e-01 1.12497079e+00 2.21198753e-01 4.53139603e-01
-1.15339661e+00 5.16556144e-01 8.40264499e-01 3.86050314e-01
8.50435197e-01 -1.87780499e-01 -2.05553323e-01 -1.37184072e+00
1.11516975e-01 -8.65969062e-01 4.81815517e-01 -4.59833860e-01
-1.02570260e+00 8.62356722e-01 -5.62149361e-02 -1.49117935e+00
-8.75034109e-02 -1.40026867e-01 -4.17583615e-01 3.34405512e-01
-1.62883580e+00 -1.16840911e+00 -6.52769268e-01 7.82325447e-01
1.04961956e+00 1.34624347e-01 3.82868618e-01 5.75379252e-01
-8.95683885e-01 5.14332116e-01 -1.16719045e-01 -5.75500429e-02
6.23408616e-01 -1.03500175e+00 4.60506886e-01 1.03581727e+00
4.59688663e-01 -9.33347493e-02 6.67143941e-01 -8.66988897e-01
-9.92025614e-01 -1.57845652e+00 5.08178294e-01 -6.88929915e-01
5.61527729e-01 -3.41453999e-01 -1.14679444e+00 6.04337633e-01
-9.03759673e-02 5.49180508e-01 1.99239999e-01 -3.71804595e-01
-1.79911911e-01 9.78644267e-02 -9.41238761e-01 4.50630516e-01
1.56765842e+00 -2.31892586e-01 -2.56969839e-01 1.91823438e-01
6.92853153e-01 -7.87607849e-01 -9.56323743e-01 6.60680950e-01
4.29561853e-01 -1.26330936e+00 9.25019443e-01 -3.63255441e-01
3.95906597e-01 -5.73745191e-01 4.31525111e-02 -9.67732489e-01
-3.25396985e-01 -7.65024424e-01 -4.92667198e-01 1.39691269e+00
1.62987337e-01 5.58472499e-02 1.17380691e+00 4.18479115e-01
-4.10779774e-01 -9.61935997e-01 -1.06698883e+00 -6.32573247e-01
-6.70401037e-01 -7.58442163e-01 4.64429468e-01 9.20531631e-01
-3.20512891e-01 -1.35789514e-01 -4.66377348e-01 -6.63409308e-02
4.33773279e-01 1.58208042e-01 8.97815704e-01 -1.14506280e+00
-1.61385596e-01 -5.05672872e-01 -5.21413207e-01 -1.16620171e+00
1.30511388e-01 -5.48353493e-01 2.93142051e-01 -1.42843688e+00
1.32253826e-01 -6.00394368e-01 -3.14475805e-01 5.33840001e-01
-3.75700355e-01 8.60863805e-01 1.80484831e-01 4.20077562e-01
-1.19002521e+00 5.12632608e-01 1.32278192e+00 -1.64444536e-01
-3.63578886e-01 -5.67798726e-02 -1.48688331e-01 7.81693816e-01
3.93251508e-01 -4.45763469e-01 -3.11258525e-01 -3.43445152e-01
-1.08535960e-01 -1.17927298e-01 4.22743946e-01 -1.15309954e+00
1.76460762e-02 -2.82801628e-01 3.62412632e-01 -7.80663610e-01
3.86685640e-01 -9.04015362e-01 3.99018466e-01 3.28144163e-01
4.53824773e-02 -7.15701357e-02 3.31468791e-01 7.56335080e-01
-4.01263624e-01 1.90391481e-01 6.83491290e-01 5.72453365e-02
-1.25109124e+00 7.79611826e-01 -1.68716967e-01 2.78528392e-01
1.27141631e+00 -6.60540402e-01 -7.35723898e-02 -4.00250852e-02
-7.08724797e-01 4.62020367e-01 7.58828640e-01 4.99715090e-01
6.74507141e-01 -1.14427757e+00 -6.57601118e-01 1.93010032e-01
2.34531209e-01 2.51073986e-01 4.83480453e-01 1.32304358e+00
-5.57431042e-01 8.23073015e-02 -7.81115294e-02 -1.11413825e+00
-1.38858116e+00 4.11375523e-01 2.69025147e-01 -3.57641071e-01
-1.16025400e+00 9.54711795e-01 3.06468427e-01 2.32503936e-01
3.46963286e-01 -6.03965342e-01 -2.97042191e-01 1.52993441e-01
5.39203644e-01 2.37250999e-01 1.50527090e-01 -9.88042831e-01
-2.75893569e-01 7.09221303e-01 4.82123122e-02 1.89850360e-01
1.26051950e+00 -4.07075346e-01 2.21091285e-01 2.73048997e-01
8.36442232e-01 -8.94596614e-03 -1.82686222e+00 -2.91559607e-01
7.99822733e-02 -8.19300413e-01 -1.31380796e-01 -4.49216008e-01
-1.34436214e+00 4.34607267e-01 4.08987612e-01 2.43234128e-01
1.11160004e+00 7.10657537e-02 1.19304371e+00 -4.08896059e-01
3.85575861e-01 -1.13315618e+00 1.57601655e-01 3.48498762e-01
6.66807473e-01 -1.34292626e+00 -9.38216373e-02 -8.34434152e-01
-7.26348341e-01 7.86610901e-01 8.48412395e-01 8.66917893e-02
4.11700487e-01 3.53660196e-01 -2.08604839e-02 -2.49716863e-01
-7.10051894e-01 -3.33648473e-01 6.89460158e-01 5.04383624e-01
1.35834530e-01 -1.98613226e-01 -9.65957865e-02 4.23858225e-01
5.46981022e-02 8.93430486e-02 2.98864841e-01 7.15074718e-01
-2.77855396e-01 -8.61791909e-01 -2.90067345e-01 5.33347487e-01
-3.84882003e-01 1.31986991e-01 -1.62631854e-01 7.26423144e-01
3.57062817e-01 6.33662343e-01 7.08625466e-02 -5.39596081e-01
2.62201786e-01 -1.61568552e-01 3.44260305e-01 -5.51466227e-01
-6.07747972e-01 2.93542117e-01 8.32981318e-02 -1.23008358e+00
-7.58063078e-01 -7.80036986e-01 -1.20884192e+00 -1.15766339e-01
-1.97401777e-01 -4.53702779e-03 2.14193255e-01 1.03690743e+00
4.45132405e-01 8.46771598e-01 1.93587720e-01 -1.38186264e+00
1.36968434e-01 -6.70405149e-01 -3.53247613e-01 7.14261353e-01
3.34245563e-01 -8.42251718e-01 -1.59858063e-01 4.49416608e-01] | [9.138483047485352, -0.10466591268777847] |
ab968d58-4cda-43c3-9f87-1bc386b6eb2d | stance-in-depth-deep-neural-approach-to | null | null | https://www.sciencedirect.com/science/article/pii/S1877050918308640 | https://ac.els-cdn.com/S1877050918308640/1-s2.0-S1877050918308640-main.pdf?_tid=556a1792-f871-4058-97de-e3cc8bc4ed63&acdnat=1551979346_642c1f818cbfdeb437938a5298dc364f | Stance-In-Depth Deep Neural Approach to Stance Classification | Understanding the user intention from text is a problem of growing interest. The social media like Twitter, Facebook etc. extract
user intention to analyze the behaviour of a user which in turn is employed for bot recognition, satire detection, fake news
detection etc.. The process of identifying stance of a user from the text is called stance detection. This article compares the
headline and body pair of a news article and classifies the pair as related or unrelated. The related pair is further classified into
agree, disagree, discuss. We call related as detailed classification and unrelated as broad classification. We employ deep neural
nets for feature extraction and stance classification. RNN models and its extensions showed significant variations in the
classification of detailed class. Bidirectional LSTM model achieved the best accuracy for broad as well as detailed classification | ['Bhadrachalam Chitturi', 'Gayathri Rajendran', 'Prabaharan Poornachandran'] | 2018-06-08 | null | null | null | international-conference-on-computational | ['satire-detection'] | ['natural-language-processing'] | [ 1.69701397e-01 7.01629519e-02 -5.67802668e-01 -2.48919353e-01
-2.79432237e-01 -5.60625553e-01 1.08976138e+00 3.48028123e-01
-4.32662129e-01 7.94367313e-01 6.86377466e-01 -5.43905020e-01
5.84297299e-01 -8.48323226e-01 -2.74647355e-01 -5.13276339e-01
6.83621824e-01 5.42261004e-01 2.99002409e-01 -6.04764223e-01
9.33355749e-01 5.95017744e-04 -1.20967352e+00 7.91393697e-01
3.42273206e-01 1.13495088e+00 -3.10295194e-01 6.39808297e-01
-3.62573087e-01 1.38311481e+00 -7.87738681e-01 -2.60104507e-01
-1.53775916e-01 -6.16546094e-01 -1.29460347e+00 -7.33362650e-03
-1.44269302e-01 -3.43893081e-01 -8.80908407e-03 1.09418535e+00
3.18545580e-01 -2.84840137e-01 6.34665966e-01 -1.34050882e+00
-6.14755571e-01 1.06392872e+00 -8.39362979e-01 5.79156935e-01
5.37952542e-01 -4.82336730e-01 8.46967041e-01 -5.16011357e-01
6.14637911e-01 1.00906885e+00 6.93433702e-01 2.70977020e-01
-7.37530172e-01 -6.26745999e-01 -9.30719450e-03 2.92134404e-01
-6.71105385e-01 -3.64341706e-01 1.03562069e+00 -7.34092832e-01
6.98002458e-01 2.52201706e-01 6.83421731e-01 1.63125348e+00
4.92657125e-01 9.57054615e-01 1.55224884e+00 -2.25335211e-01
-1.25108123e-01 4.75785851e-01 1.12569082e+00 5.08737028e-01
1.56266361e-01 -3.32025528e-01 -2.37821624e-01 -3.41633350e-01
-3.85300815e-02 5.29330969e-02 -3.71936262e-02 6.24694467e-01
-1.02606273e+00 1.30122221e+00 4.96900529e-01 6.45529449e-01
-5.88712931e-01 -2.54795223e-01 9.50603545e-01 6.27365947e-01
8.55895221e-01 1.17437407e-01 -1.50400922e-01 -3.21739502e-02
-9.29954410e-01 1.84071153e-01 1.16701353e+00 4.89389390e-01
4.16726947e-01 2.78121121e-02 -1.31376788e-01 6.91594064e-01
2.67613173e-01 3.11461598e-01 9.76537228e-01 -3.32408190e-01
4.60228056e-01 8.79225254e-01 1.78600609e-01 -1.55954099e+00
-6.79021835e-01 -5.18392980e-01 -7.41871536e-01 -1.96593478e-01
2.85068333e-01 -3.25492650e-01 -6.75163209e-01 1.06753945e+00
1.85606495e-01 -3.19234043e-01 1.01618767e-01 4.81625497e-01
1.46588612e+00 8.68102729e-01 -5.74299246e-02 -4.55481827e-01
1.90648472e+00 -7.23563194e-01 -7.98298657e-01 -2.85959393e-01
5.30354440e-01 -1.03215897e+00 5.40001452e-01 2.25125775e-01
-7.98113763e-01 -3.43835354e-01 -9.94499087e-01 2.66937055e-02
-8.29628944e-01 3.80022861e-02 1.05505705e-01 5.16745508e-01
-5.48006952e-01 6.11772180e-01 -4.56863970e-01 -3.98759335e-01
5.10676861e-01 2.70561010e-01 -1.23342112e-01 7.03769326e-01
-1.47226858e+00 9.58747387e-01 4.50429559e-01 -1.78596266e-02
-2.72773504e-01 1.51331812e-01 -7.10373580e-01 -3.54143977e-01
1.89501390e-01 -1.90153524e-01 1.28186011e+00 -1.20884168e+00
-1.30601299e+00 1.33360350e+00 -1.04315549e-01 -8.67374539e-01
3.04436177e-01 1.43182859e-01 -5.50392032e-01 -3.06559443e-01
4.00649756e-01 1.33618370e-01 1.35893798e+00 -7.44233847e-01
-8.35809588e-01 -5.05297244e-01 -1.29073858e-01 1.75412931e-02
-1.88196674e-01 6.17265463e-01 4.12899345e-01 -4.09439921e-01
4.20356005e-01 -1.02330124e+00 5.15278935e-01 -1.05409276e+00
-8.07242692e-01 -7.37152934e-01 1.57107711e+00 -1.18756425e+00
1.30128002e+00 -1.63556111e+00 -2.00879976e-01 -6.47413507e-02
5.00953794e-01 1.73084512e-01 5.05059302e-01 5.71472049e-01
-2.04335645e-01 3.33508372e-01 3.47442240e-01 1.80597469e-01
3.09197027e-02 -2.08555102e-01 -4.84446019e-01 7.91581511e-01
-1.78804398e-01 7.92271554e-01 -6.60500586e-01 -3.84700358e-01
-3.62761408e-01 1.69679731e-01 -2.41318077e-01 3.76621783e-02
2.15208735e-02 4.14037079e-01 -5.68023086e-01 5.40425777e-01
4.27618980e-01 -4.58801091e-01 7.26621225e-02 -6.33584321e-01
-4.00947392e-01 9.40489411e-01 -2.54977167e-01 3.08087528e-01
-2.83076435e-01 1.36557078e+00 -8.33020657e-02 -1.19139361e+00
1.11560667e+00 5.06326914e-01 1.41284630e-01 -6.90287828e-01
8.54172111e-01 2.81571269e-01 3.29064637e-01 -6.25875652e-01
6.98487759e-01 -8.47320911e-03 -2.87509024e-01 8.68656278e-01
-6.20798588e-01 6.66961670e-01 8.22879467e-03 -5.84790669e-02
6.87506318e-01 3.09662353e-02 8.36240053e-01 -2.12824792e-01
8.18504632e-01 3.52165732e-03 2.04405606e-01 4.22288686e-01
-6.50612473e-01 8.14669430e-02 8.42804492e-01 -5.64926922e-01
-8.93127322e-01 -2.17074782e-01 1.34655684e-01 1.30335343e+00
-4.39875424e-02 -1.76057473e-01 -7.45564938e-01 -1.19740236e+00
-2.26243049e-01 7.90848315e-01 -7.94853330e-01 1.91889748e-01
-7.92088091e-01 -9.17027950e-01 4.02889580e-01 2.79514790e-01
1.13390720e+00 -1.49488080e+00 -6.24717832e-01 1.81211919e-01
-7.86728084e-01 -9.70380902e-01 -1.49978593e-01 2.31887341e-01
-3.26225698e-01 -1.08114588e+00 -5.02858460e-01 -8.86313558e-01
3.30207884e-01 1.42697558e-01 8.18337917e-01 -1.84873375e-03
5.11481881e-01 -3.71530801e-01 -4.83453125e-01 -4.04914021e-01
-8.06693256e-01 1.30701408e-01 6.34015724e-02 8.21213275e-02
5.61277390e-01 -4.00348008e-01 -3.16796064e-01 1.44955575e-01
-4.32227939e-01 -1.01359129e-01 1.05282985e-01 7.37797976e-01
-2.91196555e-01 -8.88350084e-02 5.88088930e-01 -1.33583009e+00
1.19076693e+00 -8.63034844e-01 -1.35604367e-01 -2.94257700e-01
-4.78640616e-01 -1.21206120e-01 4.88826603e-01 -3.23472887e-01
-7.32444167e-01 -5.06929398e-01 -4.96608824e-01 4.63709742e-01
-3.22353423e-01 7.54178524e-01 4.08417314e-01 3.63472313e-01
7.92641461e-01 2.73664087e-01 -9.22566503e-02 -3.08879584e-01
-1.86151385e-01 1.48690152e+00 4.61960724e-03 1.61075905e-01
3.23188871e-01 4.86754715e-01 -4.82766420e-01 -1.20956850e+00
-1.18421638e+00 -7.40817547e-01 -4.63281691e-01 -3.22087318e-01
1.03040469e+00 -4.13923472e-01 -8.82224917e-01 7.20562696e-01
-1.54885209e+00 2.14130148e-01 3.88661832e-01 1.38599262e-01
-2.42595300e-01 3.99739385e-01 -1.08035886e+00 -7.73499072e-01
-8.16214025e-01 -1.04034233e+00 6.36283278e-01 5.51273562e-02
-9.44941223e-01 -9.83799517e-01 2.44744793e-02 8.06169689e-01
3.16332638e-01 3.86943340e-01 7.45427668e-01 -1.54908144e+00
3.68110389e-01 -5.59574366e-01 -4.04702306e-01 1.92412019e-01
1.72774002e-01 -5.12904078e-02 -9.04187620e-01 1.77720115e-02
4.75535780e-01 -2.45936781e-01 8.58654797e-01 3.68439108e-01
4.46112275e-01 -1.09785843e+00 -3.79967541e-01 -5.95857427e-02
8.70049655e-01 4.47970837e-01 6.47709787e-01 8.23636115e-01
6.20460749e-01 6.70604825e-01 4.71280694e-01 1.37761936e-01
1.50907546e-01 5.32463729e-01 3.26758355e-01 2.61863142e-01
1.96080968e-01 1.55652957e-02 7.01391757e-01 8.35166276e-01
7.04714581e-02 -3.57882738e-01 -8.99700761e-01 1.54741094e-01
-1.90195310e+00 -1.55653858e+00 -7.54419088e-01 1.68315053e+00
7.27992892e-01 5.65238118e-01 6.53077424e-01 7.58471310e-01
1.10113323e+00 4.19510305e-01 5.60948774e-02 -9.68765438e-01
1.85030758e-01 -3.27110767e-01 5.84496558e-01 6.29529357e-01
-1.36975384e+00 9.34124351e-01 5.60733843e+00 6.54484570e-01
-1.49747992e+00 4.82037395e-01 8.00352573e-01 7.33275831e-01
2.89961249e-01 -7.66842961e-02 -1.10004056e+00 6.46057129e-01
1.19142199e+00 1.37601634e-02 -5.79984859e-02 8.05323124e-01
3.88739020e-01 -1.52896523e-01 -6.56596661e-01 6.57121718e-01
3.24679583e-01 -1.53830922e+00 -1.52254134e-01 1.62982613e-01
5.61037302e-01 2.05322027e-01 -1.63544327e-01 4.37931716e-01
-1.30820526e-02 -6.48595572e-01 6.94377601e-01 3.07433277e-01
1.83797598e-01 -5.70090592e-01 1.15145576e+00 7.39727616e-01
-6.16789520e-01 -1.77123442e-01 1.54536068e-02 -5.35678387e-01
5.25807850e-02 2.55178273e-01 -1.13312304e+00 -2.84318663e-02
3.83364409e-01 1.07038713e+00 -4.83005583e-01 3.93956751e-01
-2.67680548e-02 9.69605923e-01 -6.97401837e-02 -5.98833442e-01
4.84118521e-01 -2.85560876e-01 7.88013995e-01 1.22192299e+00
-2.07617179e-01 -1.79232121e-01 2.78517216e-01 6.11390829e-01
-1.58304334e-01 1.17141791e-01 -7.06354022e-01 -4.40419704e-01
4.48991917e-02 1.21596742e+00 -1.18609083e+00 -6.77067161e-01
-3.73309888e-02 9.61667299e-01 3.23088281e-02 -3.90979461e-02
-8.87529135e-01 -2.75004715e-01 -3.93760167e-02 3.93539935e-01
1.08399782e-02 1.56965971e-01 -3.65071625e-01 -8.63838851e-01
-3.98546576e-01 -1.06374085e+00 3.86991352e-01 -5.44646561e-01
-1.12171423e+00 8.62462163e-01 -1.98720038e-01 -1.13159037e+00
-2.77715892e-01 -5.32693863e-01 -9.51147020e-01 5.57630062e-01
-9.61705685e-01 -1.23659396e+00 -2.17520967e-01 3.73332590e-01
8.24525535e-01 -2.82487601e-01 4.54640836e-01 1.07006781e-01
-7.11809814e-01 8.73798355e-02 -2.94988096e-01 6.60285413e-01
4.19519395e-01 -9.92353082e-01 3.04211289e-01 3.34401578e-01
-2.61860430e-01 5.30331254e-01 1.07742083e+00 -1.00384855e+00
-4.44497764e-01 -8.46944749e-01 1.64038932e+00 -4.42719460e-01
9.81127799e-01 -2.50998535e-03 -5.26495218e-01 7.35232532e-01
6.36176288e-01 -7.27321267e-01 6.62672341e-01 -4.15608473e-03
-3.24253470e-01 2.00102940e-01 -1.12317383e+00 6.79945230e-01
4.24648225e-01 -4.79153872e-01 -8.40416908e-01 8.51862192e-01
4.56883878e-01 -2.59476930e-01 -3.64542246e-01 -3.79593670e-02
5.68949759e-01 -1.24762869e+00 6.28406167e-01 -8.91836405e-01
8.63309503e-01 7.37586766e-02 7.46827498e-02 -8.84609997e-01
-3.97578120e-01 -3.77142519e-01 -1.27079859e-01 1.04702067e+00
5.28471768e-01 -9.07664180e-01 8.92814159e-01 -1.46450981e-01
1.70940429e-01 -5.47468841e-01 -3.42204779e-01 -1.41097128e-01
2.59883553e-02 -1.32951379e-01 1.98897738e-02 1.30144095e+00
4.44026798e-01 1.09135997e+00 -6.58052921e-01 -2.59301007e-01
4.27723497e-01 5.25697470e-01 4.63326156e-01 -1.43340194e+00
2.47207005e-02 -7.47494638e-01 -3.79812688e-01 -9.01215911e-01
2.27418214e-01 -8.17974329e-01 -2.70176023e-01 -1.68051255e+00
1.97306156e-01 1.23591729e-01 2.30533913e-01 3.07394385e-01
2.95784861e-01 7.11129665e-01 -1.47352412e-01 6.47712469e-01
-3.59665424e-01 2.04044790e-03 7.98575819e-01 -3.71323854e-01
-3.00554991e-01 6.76806748e-01 -6.89778805e-01 1.16957974e+00
1.27551305e+00 -6.42192602e-01 1.10582210e-01 4.11058456e-01
6.76579535e-01 1.02668740e-02 2.29387641e-01 -6.70356870e-01
-1.70245357e-02 -1.32806137e-01 2.02074945e-01 -1.12891901e+00
1.36322707e-01 -6.56217813e-01 -4.34214979e-01 9.88085270e-01
-6.20676935e-01 1.65685058e-01 -2.67240644e-01 5.34238756e-01
-9.35514197e-02 -4.98107523e-01 8.96377385e-01 -2.83546627e-01
-3.96851093e-01 -2.28475317e-01 -8.39845419e-01 4.89043854e-02
8.15487683e-01 -2.59459674e-01 -6.90035582e-01 -9.13831651e-01
-9.25003707e-01 -2.78648436e-01 -1.29682988e-01 4.92225379e-01
3.61004502e-01 -1.10483301e+00 -7.72384822e-01 -6.74945340e-02
-1.82422429e-01 -9.52115655e-01 -3.38667840e-01 1.47190344e+00
-5.43695867e-01 4.62365925e-01 -2.15984717e-01 -4.53490824e-01
-1.84116924e+00 3.04804176e-01 1.05572097e-01 -3.44828367e-01
-4.21911836e-01 4.57499176e-01 -3.76950741e-01 -2.50930190e-01
-1.14327513e-01 -7.10747316e-02 -1.12550819e+00 6.65342927e-01
5.43419003e-01 6.01813674e-01 -1.13431737e-02 -1.43651593e+00
-2.53334284e-01 1.01792820e-01 -3.99850637e-01 1.38183668e-01
1.04404724e+00 -1.33902878e-01 -5.70769787e-01 7.83896625e-01
1.45903242e+00 1.28273815e-01 -1.87327206e-01 -1.24533080e-01
2.94470906e-01 4.54267673e-02 7.18237087e-02 -5.56102455e-01
-6.07979655e-01 6.76139951e-01 7.63341710e-02 1.16996312e+00
3.83316159e-01 -2.63184547e-01 1.17425203e+00 5.27686775e-01
-8.56994018e-02 -1.31395030e+00 -3.52105051e-02 1.10494661e+00
9.32468355e-01 -1.48341656e+00 9.84571129e-02 -1.10851616e-01
-8.45651865e-01 1.21733248e+00 1.73446402e-01 -3.80565554e-01
8.17277431e-01 8.49975422e-02 1.30326629e-01 -4.79073077e-01
-3.85520577e-01 -3.42157967e-02 3.61615300e-01 3.43614757e-01
8.71274889e-01 -1.24924593e-01 -9.53103900e-01 3.88493240e-01
-5.48865318e-01 -3.44441712e-01 6.88695967e-01 9.52708960e-01
-8.76851857e-01 -6.28274024e-01 -4.21072394e-01 6.56367779e-01
-8.93389106e-01 -9.55782160e-02 -1.06173098e+00 6.61893845e-01
9.86470561e-03 1.29013395e+00 -6.12708330e-02 -6.32700443e-01
-9.54242796e-02 3.73224825e-01 -1.60133585e-01 -4.72174644e-01
-1.21670365e+00 5.35197556e-02 7.91459143e-01 -7.18908682e-02
-6.24052584e-01 -4.36056972e-01 -1.21694160e+00 -6.90662920e-01
-4.61500466e-01 2.93522507e-01 7.34079599e-01 1.51555192e+00
-2.21827850e-01 3.06702375e-01 7.51827717e-01 -7.26305246e-01
-5.81938207e-01 -1.41926599e+00 -1.54713944e-01 4.20037955e-01
6.47336066e-01 -3.33131015e-01 -5.25033712e-01 5.28562441e-02] | [8.316808700561523, 10.119523048400879] |
b735bf1e-22a0-4cac-9229-e875b735342b | towards-interpretable-sleep-stage | 2208.06991 | null | https://arxiv.org/abs/2208.06991v2 | https://arxiv.org/pdf/2208.06991v2.pdf | Towards Interpretable Sleep Stage Classification Using Cross-Modal Transformers | Accurate sleep stage classification is significant for sleep health assessment. In recent years, several machine-learning based sleep staging algorithms have been developed, and in particular, deep-learning based algorithms have achieved performance on par with human annotation. Despite the improved performance, a limitation of most deep-learning based algorithms is their black-box behavior, which has limited their use in clinical settings. Here, we propose a cross-modal transformer, which is a transformer-based method for sleep stage classification. The proposed cross-modal transformer consists of a novel cross-modal transformer encoder architecture along with a multi-scale one-dimensional convolutional neural network for automatic representation learning. Our method outperforms the state-of-the-art methods and eliminates the black-box behavior of deep-learning models by utilizing the interpretability aspect of the attention modules. Furthermore, our method provides considerable reductions in the number of parameters and training time compared to the state-of-the-art methods. Our code is available at https://github.com/Jathurshan0330/Cross-Modal-Transformer. | ['Chamira U. S. Edussooriya', 'Anjula C. De Silva', 'Simon L. Kappel', 'Dhinesh Suntharalingham', 'Vinith Kugathasan', 'Mithunjha Anandakumar', 'Jathurshan Pradeepkumar'] | 2022-08-15 | null | null | null | null | ['sleep-stage-detection', 'multimodal-sleep-stage-detection', 'sleep-staging', 'automatic-sleep-stage-classification', 'classification'] | ['medical', 'medical', 'medical', 'medical', 'methodology'] | [-4.27213572e-02 2.00503320e-02 -3.52702290e-01 -4.58751917e-01
-8.22834730e-01 -1.10089101e-01 6.59931302e-02 1.30703062e-01
-5.21635711e-01 5.27649701e-01 1.89617500e-01 -4.09308910e-01
5.12925610e-02 -4.98099208e-01 -1.40862435e-01 -5.97806752e-01
2.47205481e-01 4.50743377e-01 2.49973059e-01 -2.58257408e-02
-1.12768717e-01 -1.73346382e-02 -1.32922781e+00 3.51227850e-01
8.97439659e-01 1.28045762e+00 2.22190738e-01 5.64851999e-01
1.25147188e-02 9.07072902e-01 -5.18510818e-01 -3.63894790e-01
-2.34851599e-01 -5.88320017e-01 -8.83622944e-01 -3.18885416e-01
3.72601412e-02 -1.60660550e-01 -4.61805284e-01 8.15708101e-01
7.72551119e-01 -9.97179747e-02 3.39948475e-01 -1.04829824e+00
-5.91235161e-01 9.87005085e-02 -3.08348864e-01 6.78490698e-01
5.41913994e-02 -6.51118904e-03 9.45734739e-01 -6.90115571e-01
-3.03595643e-02 6.73759103e-01 9.01533782e-01 8.90512824e-01
-1.14046919e+00 -7.68990815e-01 -2.45589226e-01 5.56181550e-01
-1.47868633e+00 -6.57161295e-01 6.16392970e-01 -2.70221949e-01
1.13792610e+00 7.37614855e-02 9.87590551e-01 9.23529983e-01
6.07166111e-01 8.76557708e-01 1.10801303e+00 -1.11876696e-01
2.67712981e-01 6.68694824e-02 3.76530476e-02 1.15963173e+00
1.07507810e-01 -1.91667229e-01 -5.49370408e-01 5.57453185e-02
4.94101644e-01 4.86855358e-01 -1.65411219e-01 -9.90788564e-02
-9.19989169e-01 8.10618162e-01 6.69492245e-01 6.54400945e-01
-1.15492068e-01 9.07682627e-02 6.41589999e-01 -6.99152127e-02
7.84513175e-01 3.12207460e-01 -4.07847732e-01 -4.21864003e-01
-1.52072930e+00 -2.88560957e-01 5.49929798e-01 5.94337881e-01
4.80606556e-01 -3.66791449e-02 -3.48306000e-01 8.19124877e-01
3.35365772e-01 2.38214061e-01 1.05356693e+00 -7.66676962e-01
6.40479773e-02 9.72563207e-01 -2.65316457e-01 -4.21433449e-01
-9.31874156e-01 -6.62532210e-01 -1.10732436e+00 -1.46481618e-01
1.00244842e-01 1.81797147e-01 -8.65061700e-01 1.58775091e+00
1.59564018e-02 6.39122203e-02 -3.36912274e-02 6.77998722e-01
1.03012252e+00 2.94733554e-01 -5.30691668e-02 -1.67324785e-02
1.76349902e+00 -1.34920883e+00 -8.88943493e-01 -4.95802253e-01
5.35064399e-01 -4.95919526e-01 1.22648251e+00 4.68937218e-01
-1.20469677e+00 -4.30247664e-01 -1.09838033e+00 -6.13615572e-01
-3.21179867e-01 3.89330655e-01 5.59540391e-01 6.45296931e-01
-1.31407487e+00 5.86205602e-01 -1.40781629e+00 -5.38910627e-01
8.42763245e-01 7.26102948e-01 -2.19614118e-01 2.05773264e-01
-9.41914320e-01 1.06684709e+00 -8.26423243e-02 2.11409315e-01
-1.05633712e+00 -7.17431962e-01 -8.55886221e-01 3.68716210e-01
2.43830085e-02 -1.05633664e+00 1.57512581e+00 -6.60888970e-01
-1.42432296e+00 9.97681379e-01 -6.21233881e-01 -5.05919933e-01
1.47484794e-01 -2.88094223e-01 -3.63747150e-01 2.24311143e-01
1.55589640e-01 5.06728411e-01 6.13433301e-01 -6.53952658e-01
-4.79387701e-01 -3.40718389e-01 1.62570372e-01 5.71719408e-02
-5.64850569e-01 -1.75380893e-02 -6.28198326e-01 -2.92027444e-01
-2.50908136e-01 -9.35433447e-01 -2.92262524e-01 3.42608035e-01
-2.35376582e-01 -3.16188097e-01 4.83050436e-01 -3.86305273e-01
1.61164224e+00 -2.21139550e+00 2.67379405e-03 -5.86293519e-01
7.08076835e-01 3.49511027e-01 3.05828601e-01 2.64570445e-01
8.22118670e-02 7.17528239e-02 -3.78327459e-01 -9.59111094e-01
-1.13788329e-01 2.94456869e-01 2.88211823e-01 6.80544853e-01
-2.79121567e-02 9.42693651e-01 -8.76222968e-01 -7.07871556e-01
3.60862464e-01 6.43749952e-01 -5.90358615e-01 4.30522978e-01
2.34214857e-01 3.86148751e-01 -2.77554512e-01 5.81121087e-01
1.98542997e-01 -7.74730504e-01 3.52549478e-02 -2.56907403e-01
-1.91066518e-01 7.25988567e-01 -1.62309676e-01 2.13549232e+00
-7.24403560e-01 6.57967925e-01 -1.66494071e-01 -1.04574645e+00
4.56312358e-01 4.65757072e-01 5.80536664e-01 -7.51592815e-01
4.07417268e-01 1.14175104e-01 1.30237088e-01 -4.43594575e-01
3.10534984e-01 -4.56428260e-01 -7.15186000e-02 3.81697774e-01
1.72781125e-01 2.49783814e-01 2.33002573e-01 3.89937572e-02
1.39938557e+00 -2.06070840e-01 5.98415315e-01 -4.16910261e-01
4.72093284e-01 -2.64852673e-01 9.12135243e-01 3.00905854e-01
-5.68159044e-01 6.26190662e-01 4.52250093e-01 -5.06845891e-01
-6.60838127e-01 -8.23601544e-01 -2.33184114e-01 9.08853054e-01
6.81996569e-02 -9.08860564e-01 -7.74280131e-01 -5.31361163e-01
-4.29410905e-01 4.38555479e-01 -1.01554620e+00 -5.96462488e-01
-2.68403798e-01 -9.06860650e-01 8.60378861e-01 8.05648088e-01
5.71294844e-01 -8.64472449e-01 -8.74458790e-01 1.32954389e-01
-4.22644049e-01 -1.07103205e+00 -5.84242404e-01 4.90209997e-01
-1.03205085e+00 -1.15703249e+00 -6.51983917e-01 -7.01747179e-01
6.78366244e-01 1.65380821e-01 1.21584153e+00 1.65974692e-01
-4.88002807e-01 6.97217509e-02 -2.94813305e-01 -4.67839092e-01
-9.21943933e-02 4.55431640e-01 1.29160792e-01 -1.57978520e-01
5.86980462e-01 -5.96207500e-01 -1.36297226e+00 2.69815385e-01
-6.56121731e-01 2.54528463e-01 8.25808048e-01 9.70420659e-01
6.85665250e-01 -9.82944071e-02 3.89387399e-01 -8.47043812e-01
4.03006613e-01 -4.41899270e-01 -2.28864715e-01 -4.58480716e-02
-1.00236440e+00 5.40333129e-02 6.41013086e-01 -7.76594207e-02
-7.16112733e-01 -1.59061506e-01 -4.77668911e-01 -4.62022275e-01
6.84093013e-02 3.77298981e-01 8.77552032e-02 2.09802777e-01
4.63033885e-01 4.22505289e-01 1.55838549e-01 -6.08779490e-01
-1.23905860e-01 8.23353589e-01 3.32572460e-01 -2.99175326e-02
4.66664284e-01 5.41904211e-01 -1.27731517e-01 -7.02201247e-01
-1.41874731e+00 -6.81232870e-01 -5.44765711e-01 7.80611262e-02
1.30251777e+00 -1.13817155e+00 -6.27754033e-01 5.06351769e-01
-7.76015043e-01 -7.17332900e-01 -2.88240880e-01 2.76823163e-01
-5.31209111e-01 1.76992372e-01 -8.35299492e-01 -5.60059428e-01
-1.03055799e+00 -1.24527895e+00 1.17433953e+00 3.14769238e-01
-4.30884421e-01 -1.09225833e+00 2.27215812e-01 7.13889718e-01
6.59598351e-01 -1.67027071e-01 8.97959471e-01 -5.09495020e-01
-1.90100089e-01 3.51464599e-02 -1.07603684e-01 3.09606373e-01
3.37063134e-01 -5.12580633e-01 -1.19947851e+00 -4.88814265e-01
1.11971080e-01 -2.82659769e-01 7.86090493e-01 5.02590001e-01
1.46885395e+00 -1.21319316e-01 -4.74127680e-01 8.29985559e-01
1.15616035e+00 4.76417989e-02 5.34814537e-01 3.23936194e-01
4.90696818e-01 -1.50843009e-01 9.01670828e-02 3.58693361e-01
9.01493371e-01 6.48600578e-01 2.88686186e-01 -6.76126838e-01
-1.54625192e-01 2.37276312e-02 2.91801602e-01 1.00545239e+00
1.35444462e-01 -1.81395128e-01 -7.68254578e-01 5.56688666e-01
-1.71040583e+00 -7.96652973e-01 1.56707242e-01 1.92378974e+00
8.55821490e-01 2.57691860e-01 2.82446682e-01 3.39191675e-01
2.31582955e-01 1.69361115e-01 -7.53493905e-01 -5.02233803e-01
4.70955104e-01 5.30221999e-01 2.07718685e-01 -2.23057973e-03
-1.00138009e+00 6.51117504e-01 6.19115257e+00 5.70092142e-01
-1.24703431e+00 7.61905730e-01 5.16892493e-01 -4.48930860e-01
8.47354382e-02 -4.20634747e-01 -6.03712022e-01 6.27381742e-01
1.32084405e+00 -9.26853716e-02 3.41005176e-01 8.43672872e-01
4.68821764e-01 -4.80224155e-02 -1.17293751e+00 1.11793339e+00
2.34753236e-01 -1.19542658e+00 -5.55350363e-01 1.07985988e-01
2.20269263e-01 5.19845665e-01 4.69769686e-02 5.13953924e-01
-1.70846194e-01 -9.64987516e-01 5.51428318e-01 3.00094336e-01
1.12251067e+00 -4.67162937e-01 8.58322144e-01 1.23328820e-01
-1.38491535e+00 -2.58583397e-01 -3.13770801e-01 1.95909943e-02
-3.06777228e-02 5.96440315e-01 -4.90067154e-01 3.34082514e-01
1.07902086e+00 1.01898050e+00 -9.08618212e-01 1.12881470e+00
-2.20407084e-01 8.57788563e-01 -1.13903701e-01 4.01056968e-02
2.09325496e-02 3.70864980e-02 4.53289831e-03 1.19067681e+00
3.24753553e-01 -4.91027795e-02 -5.30821867e-02 8.26984227e-01
-2.21260533e-01 -2.63692886e-01 -2.68435627e-01 -7.87843764e-02
2.05936432e-01 1.50801647e+00 -8.43040109e-01 -3.24317902e-01
-6.58709943e-01 1.12589133e+00 4.83531028e-01 -1.47213966e-01
-8.45608950e-01 -2.00776279e-01 7.97954857e-01 2.92108744e-01
3.71828049e-01 -2.62570158e-02 -3.46418291e-01 -1.28522420e+00
-9.28107798e-02 -7.32608140e-01 5.26402771e-01 -6.29828632e-01
-1.13902116e+00 9.45059299e-01 -3.20873320e-01 -1.32171738e+00
7.48873651e-02 -4.03234273e-01 -7.86890507e-01 3.91650170e-01
-1.64546239e+00 -1.13757825e+00 -4.15540755e-01 5.82929254e-01
7.94815063e-01 2.00081337e-02 1.06043923e+00 8.02070439e-01
-1.03340399e+00 8.97898078e-01 1.30806088e-01 1.66161850e-01
6.46492600e-01 -1.45674884e+00 2.55597264e-01 6.66946113e-01
-1.65426373e-01 7.67267168e-01 3.18253726e-01 -3.96287516e-02
-1.24559140e+00 -1.19850075e+00 9.17555928e-01 -4.88777071e-01
4.94698256e-01 -4.94916320e-01 -6.91027284e-01 5.98904848e-01
2.90426731e-01 1.89034998e-01 1.28418124e+00 3.32371116e-01
-3.13958637e-02 -5.49494445e-01 -9.88297999e-01 3.16185355e-01
8.08294952e-01 -6.79858088e-01 -5.39737284e-01 1.97510898e-01
2.96186924e-01 -4.53979701e-01 -8.72640371e-01 1.89626202e-01
6.45650387e-01 -1.02218688e+00 6.39118075e-01 -1.35308847e-01
5.09396791e-01 -3.45859617e-01 3.33231241e-01 -1.05877781e+00
-5.94175518e-01 -4.09669578e-01 -3.88556451e-01 9.42117631e-01
4.42564398e-01 -5.33688366e-01 9.18194056e-01 4.56221431e-01
-6.48805320e-01 -1.27704465e+00 -1.20791900e+00 -4.59669232e-01
2.73463223e-02 -1.39487520e-01 2.89206266e-01 3.85711819e-01
1.85971886e-01 7.57522047e-01 -3.48935366e-01 -1.21992968e-01
2.92855024e-01 1.85554504e-01 3.37185502e-01 -1.11943305e+00
-2.25545198e-01 -4.97377753e-01 -4.79000717e-01 -8.70960414e-01
1.32580549e-02 -9.81409669e-01 5.16724512e-02 -1.95002401e+00
6.33020222e-01 -2.21320316e-01 -8.35942328e-01 9.56924081e-01
-1.27956465e-01 5.86837053e-01 -1.57488044e-02 2.67353147e-01
-9.68576431e-01 7.20787168e-01 1.02722442e+00 -1.14138536e-01
-1.06725097e-01 8.88616741e-02 -1.08481681e+00 7.75752366e-01
1.00281966e+00 -7.15981901e-01 -5.66474855e-01 -4.39863473e-01
7.71056488e-02 -1.54074267e-01 1.60321206e-01 -1.33679855e+00
3.60201359e-01 4.27383393e-01 5.21878958e-01 -4.37943131e-01
6.77343607e-01 -6.45766139e-01 6.63773790e-02 8.74237180e-01
-1.90877944e-01 2.07743570e-01 2.56914675e-01 3.63588363e-01
-7.50209093e-02 4.09304276e-02 9.91874754e-01 -1.64624199e-01
-2.76232034e-01 4.42080945e-01 -7.85503089e-01 1.68781132e-01
6.58694208e-01 -2.06413686e-01 -3.72498870e-01 -1.22265883e-01
-4.81125116e-01 2.10716650e-01 4.78314161e-01 2.40068346e-01
4.71271813e-01 -9.47791040e-01 -2.83970684e-01 2.19720885e-01
2.39524156e-01 1.12490602e-01 2.46185273e-01 1.44362962e+00
-4.09967721e-01 5.30989408e-01 -2.44398445e-01 -5.23531795e-01
-1.25642240e+00 4.59696978e-01 7.38294005e-01 -6.92677557e-01
-8.17947388e-01 7.38902986e-01 4.47117686e-01 -1.51832402e-01
3.58662531e-02 -6.11800194e-01 -1.85334623e-01 -8.03666562e-02
5.07393718e-01 2.50463754e-01 4.13878709e-01 -4.22709823e-01
-7.43263841e-01 3.51815075e-01 -1.79638878e-01 3.44502091e-01
1.29012465e+00 -1.42087728e-01 -7.08814859e-02 6.68378532e-01
1.16535449e+00 -3.99051815e-01 -7.78532684e-01 -1.03771742e-02
-2.27356672e-01 -4.34160344e-02 4.58292574e-01 -8.47334325e-01
-1.33093011e+00 1.10591829e+00 9.72214639e-01 -3.79059575e-02
1.50520647e+00 1.13624148e-01 1.22260869e+00 2.47755960e-01
1.71971828e-01 -8.00074935e-01 9.50124413e-02 2.37106591e-01
3.93395454e-01 -1.32111335e+00 1.94730639e-01 -4.51337110e-04
-5.60303569e-01 8.33093464e-01 6.14638209e-01 1.45664841e-01
8.46040189e-01 2.27190375e-01 2.07190424e-01 -4.63863224e-01
-8.96580160e-01 -2.67951995e-01 2.79139102e-01 2.67114013e-01
7.74455011e-01 -5.29001467e-02 -2.65326589e-01 1.00776684e+00
-3.55231762e-01 3.88152897e-01 3.51690561e-01 8.57854605e-01
-2.17840105e-01 -1.09453750e+00 2.79855609e-01 8.03136766e-01
-9.71548796e-01 -3.52026045e-01 -1.09716035e-01 5.07316530e-01
1.92269847e-01 1.14409316e+00 1.66094840e-01 -4.88451570e-01
2.44143859e-01 9.24593210e-02 3.62436861e-01 -8.26438785e-01
-7.46639073e-01 -1.15036942e-01 -9.64162685e-03 -7.35212743e-01
-4.47866261e-01 -4.59574521e-01 -1.18906343e+00 -2.76950240e-01
-2.90446103e-01 2.87289202e-01 3.35886776e-01 1.03459227e+00
6.92867815e-01 8.73989463e-01 3.80496889e-01 -8.17396820e-01
-8.88929144e-02 -1.04615760e+00 -4.95678931e-01 9.15388167e-02
5.77090681e-01 -8.21247697e-01 -2.42012948e-01 -4.34758253e-02] | [13.524077415466309, 3.5212693214416504] |
807d086e-6946-4818-864e-f872338f472d | the-magnitude-vector-of-images-1 | 2110.15188 | null | https://arxiv.org/abs/2110.15188v2 | https://arxiv.org/pdf/2110.15188v2.pdf | The magnitude vector of images | The magnitude of a finite metric space has recently emerged as a novel invariant quantity, allowing to measure the effective size of a metric space. Despite encouraging first results demonstrating the descriptive abilities of the magnitude, such as being able to detect the boundary of a metric space, the potential use cases of magnitude remain under-explored. In this work, we investigate the properties of the magnitude on images, an important data modality in many machine learning applications. By endowing each individual images with its own metric space, we are able to define the concept of magnitude on images and analyse the individual contribution of each pixel with the magnitude vector. In particular, we theoretically show that the previously known properties of boundary detection translate to edge detection abilities in images. Furthermore, we demonstrate practical use cases of magnitude for machine learning applications and propose a novel magnitude model that consists of a computationally efficient magnitude computation and a learnable metric. By doing so, we address the computational hurdle that used to make magnitude impractical for many applications and open the way for the adoption of magnitude in machine learning research. | ["Leslie O'Bray", 'Edward De Brouwer', 'Bastian Rieck', 'Michael F. Adamer'] | 2021-10-28 | the-magnitude-vector-of-images | https://openreview.net/forum?id=-3Qj7Jl6UP5 | https://openreview.net/pdf?id=-3Qj7Jl6UP5 | null | ['boundary-detection'] | ['computer-vision'] | [ 4.90160584e-01 1.93183005e-01 -2.05240794e-03 -2.24769026e-01
-3.05771023e-01 -4.84307766e-01 6.15355849e-01 3.25238854e-01
-5.33590019e-01 3.30899954e-01 -3.25134955e-02 -1.43383771e-01
-4.87045854e-01 -8.70174527e-01 -5.09933352e-01 -7.54610300e-01
-5.80220163e-01 -1.86822519e-01 2.80245375e-02 -1.22732744e-01
5.12439907e-01 6.52565897e-01 -1.64294004e+00 -2.44284973e-01
6.77096844e-01 1.36981404e+00 -7.77969211e-02 7.50631988e-01
1.65636256e-01 4.73843485e-01 -4.67133284e-01 -2.84425579e-02
5.05259871e-01 -5.29162824e-01 -7.05193460e-01 1.79662123e-01
6.34475648e-01 -6.08802624e-02 -1.46292463e-01 1.14574957e+00
2.19883829e-01 6.08995371e-02 9.41704094e-01 -1.36608970e+00
-5.49528539e-01 1.98521435e-01 -4.55651224e-01 4.89331961e-01
4.81923431e-01 -3.96305136e-02 1.26829827e+00 -7.54018426e-01
6.28896058e-01 7.98744977e-01 7.38371253e-01 2.30321214e-01
-1.05514884e+00 1.55684382e-01 -3.32694978e-01 3.55290234e-01
-1.46368563e+00 -1.22435719e-01 1.12706900e+00 -5.91788590e-01
4.29087371e-01 5.87617993e-01 7.76435196e-01 2.07840398e-01
1.69218183e-01 6.23650610e-01 1.25954115e+00 -7.72708237e-01
5.06086469e-01 9.05231852e-03 1.19656235e-01 8.32468808e-01
2.67290354e-01 -1.00021176e-01 -4.05154794e-01 1.64051697e-01
8.42656791e-01 -2.66766787e-01 -4.77020591e-01 -6.83543444e-01
-1.28590584e+00 7.68590629e-01 6.63177192e-01 6.18043900e-01
-2.49960959e-01 1.71003297e-01 4.35628146e-02 2.26264492e-01
5.06633818e-01 5.79757035e-01 -6.06612116e-02 -2.05322966e-01
-7.52178490e-01 7.62998760e-02 7.87999392e-01 4.64898646e-01
7.70937145e-01 -3.89848948e-01 7.71760568e-02 4.41504687e-01
-6.23879470e-02 3.40544999e-01 2.29745761e-01 -9.61886406e-01
3.60105373e-02 9.66748536e-01 -7.49334469e-02 -1.41552901e+00
-7.55485117e-01 -2.99830079e-01 -8.35527897e-01 4.68773991e-01
7.90017128e-01 3.53076279e-01 -2.35802948e-01 1.80172932e+00
4.84211981e-01 2.11231709e-01 -1.93912506e-01 9.16662216e-01
1.75817102e-01 2.34280795e-01 -2.79740304e-01 -1.27792612e-01
1.16133869e+00 -3.41375202e-01 -3.59970033e-01 9.31301415e-02
9.37536538e-01 -4.90014523e-01 1.18598306e+00 5.26664138e-01
-1.13961279e+00 -2.36612022e-01 -1.46409953e+00 -4.77563478e-02
-6.22267902e-01 1.96794048e-02 6.80731237e-01 8.05857122e-01
-1.16850758e+00 8.88597012e-01 -6.38277888e-01 -3.40969175e-01
3.47144037e-01 2.44058564e-01 -4.23292458e-01 9.25834700e-02
-8.44827831e-01 1.00733411e+00 2.27854267e-01 1.12103879e-01
-6.85203895e-02 -7.61477530e-01 -1.00684416e+00 -9.47323814e-02
-5.47430106e-02 -3.32165927e-01 6.82225466e-01 -8.18799675e-01
-1.06248271e+00 1.16607046e+00 3.38915110e-01 -3.95759553e-01
7.44843543e-01 9.63453855e-03 -1.68076813e-01 4.16424602e-01
-1.66997701e-01 4.11713064e-01 8.24584544e-01 -8.63200963e-01
-4.97544914e-01 -4.66955721e-01 3.51911366e-01 -9.11549479e-02
-7.15095639e-01 -3.39652807e-01 2.70229787e-01 -5.71575224e-01
4.00310189e-01 -7.86561728e-01 -8.04422349e-02 5.33118546e-01
-1.40558660e-01 -1.51747659e-01 6.40716553e-01 -3.43844086e-01
1.35660267e+00 -2.40341425e+00 3.13963532e-01 2.67836958e-01
5.34040272e-01 -3.35477991e-03 -1.39180169e-01 2.61987925e-01
-1.62422985e-01 1.60317704e-01 -5.92109621e-01 1.63200557e-01
5.47819734e-02 -2.15263548e-03 -3.07482178e-03 1.16701341e+00
3.86441857e-01 8.36031854e-01 -1.01117098e+00 -5.03453493e-01
4.90776628e-01 5.61488271e-01 -4.37310517e-01 -2.94840150e-02
2.51023084e-01 9.77354795e-02 -3.91225249e-01 2.71629006e-01
8.13840508e-01 -1.75585449e-01 -1.28740981e-01 -3.01713496e-01
-1.80905133e-01 -5.18378541e-02 -1.48619902e+00 1.33315420e+00
-4.57923293e-01 9.32133973e-01 1.57923043e-01 -1.27702820e+00
8.34752619e-01 -1.06967315e-01 7.43536711e-01 -8.07351828e-01
2.00062975e-01 3.21314245e-01 -1.04491666e-01 -4.86730158e-01
3.57049674e-01 -1.56488702e-01 8.50469545e-02 5.48724949e-01
-2.65950024e-01 -4.10398453e-01 2.50261635e-01 -2.32871342e-02
1.23402703e+00 -1.87457234e-01 6.31861269e-01 -6.13830447e-01
7.17982352e-01 -1.44575059e-01 1.36950314e-01 2.72517383e-01
-4.94640231e-01 7.20433176e-01 5.60477674e-01 -3.90702784e-01
-1.04109788e+00 -1.11317265e+00 -5.05116582e-01 8.37444782e-01
4.66167331e-01 -1.60107926e-01 -1.03382361e+00 -4.90475833e-01
3.36459391e-02 2.16801018e-01 -7.85486341e-01 -3.24773878e-01
-6.91711187e-01 -7.45231748e-01 2.97627091e-01 3.88478518e-01
4.19179589e-01 -6.41528487e-01 -1.10873580e+00 -1.27428919e-01
-6.40257969e-02 -1.14099061e+00 -4.68100309e-01 -9.68931615e-02
-9.89803851e-01 -1.42237520e+00 -6.91710830e-01 -6.44260764e-01
7.60314524e-01 3.93189341e-01 1.05269933e+00 3.17702234e-01
-8.54620397e-01 7.97692537e-01 -3.11628968e-01 -6.24004565e-02
-3.24300051e-01 -2.37310842e-01 1.49411306e-01 2.79843748e-01
9.59375799e-02 -6.19270563e-01 -9.96260405e-01 2.17154235e-01
-1.14107263e+00 -9.28755179e-02 5.47463417e-01 4.69202310e-01
3.43349248e-01 -4.53672046e-03 4.36782509e-01 -3.13747853e-01
4.91225272e-01 -2.83169121e-01 -4.92520690e-01 1.15184590e-01
-5.19411206e-01 3.73822182e-01 3.84939283e-01 -3.12499136e-01
-4.66936618e-01 1.88187259e-04 1.16576649e-01 1.43634737e-01
1.72007561e-01 3.42891216e-01 6.98567107e-02 -5.95103741e-01
8.13631713e-01 -8.19308101e-04 3.88298295e-02 -3.10187936e-01
6.24052823e-01 4.54783410e-01 5.41537941e-01 -3.45801950e-01
7.84530044e-01 8.51365566e-01 6.69064403e-01 -1.22329855e+00
-5.86236179e-01 -7.36222506e-01 -8.82245541e-01 -3.80057395e-01
7.97003806e-01 -3.12214315e-01 -7.57322729e-01 3.58267605e-01
-1.00763512e+00 -7.11307675e-02 -5.21884382e-01 5.04813492e-01
-9.15590763e-01 5.55074036e-01 -3.44982654e-01 -8.16008925e-01
-1.86817963e-02 -6.91861510e-01 9.44218040e-01 1.03551202e-01
-2.27140114e-01 -1.31560361e+00 2.06918344e-01 7.46910796e-02
3.61900747e-01 5.67754447e-01 8.65363419e-01 -1.53892785e-01
-4.21120942e-01 -2.13138282e-01 -3.39063972e-01 5.45759022e-01
9.64066386e-02 -3.52688134e-02 -9.65339184e-01 -8.39069709e-02
4.00106698e-01 -2.51275953e-03 7.97696590e-01 4.06994969e-01
9.66667712e-01 -2.30256125e-01 -1.07355770e-02 6.35820568e-01
1.58986580e+00 -3.51844758e-01 6.00648761e-01 4.74522561e-01
4.29069638e-01 8.74795675e-01 5.01391351e-01 4.89243567e-01
8.41306522e-02 6.93181992e-01 6.07630253e-01 -2.16917500e-01
-2.92145550e-01 1.06881827e-01 5.47993965e-02 7.38688409e-01
-1.68068096e-01 1.92673549e-01 -6.71938598e-01 5.01794100e-01
-1.49935412e+00 -7.98462927e-01 -2.99745619e-01 2.55511379e+00
6.71553910e-01 1.68915123e-01 1.06988065e-01 7.61704683e-01
5.65549314e-01 2.67601460e-01 -1.49821192e-01 -3.98492873e-01
-3.09668124e-01 2.08331257e-01 6.59974158e-01 5.67594588e-01
-1.22090948e+00 1.87655479e-01 6.83517885e+00 5.73762715e-01
-1.29094100e+00 -1.22451209e-01 5.13938427e-01 1.56282574e-01
-2.86275685e-01 -1.15115687e-01 -1.27370313e-01 3.64400983e-01
5.04560649e-01 -3.77783805e-01 2.92654157e-01 7.61642396e-01
1.99270129e-01 -3.92687291e-01 -1.40222168e+00 1.22616923e+00
1.90453559e-01 -1.09012127e+00 -2.21172437e-01 2.70486474e-01
5.91266930e-01 -3.37977946e-01 2.21199185e-01 -3.67923439e-01
-6.27672553e-01 -1.14516306e+00 7.41086721e-01 4.21116531e-01
8.43862593e-01 -4.58533138e-01 3.43631357e-01 2.31304765e-01
-1.16548336e+00 -1.17369011e-01 -4.06821370e-01 -6.26033306e-01
-5.79490922e-02 1.05969751e+00 -7.26575851e-01 2.11824223e-01
2.24751234e-01 7.05974579e-01 -7.66380370e-01 1.18303430e+00
1.02571100e-01 1.84474349e-01 -5.34932673e-01 -4.19049375e-02
2.25213096e-01 -5.59183061e-01 5.47879815e-01 1.15457976e+00
4.55633670e-01 -2.43349969e-02 -1.71464771e-01 8.53326797e-01
-6.92093819e-02 3.02015662e-01 -8.11225474e-01 3.14086340e-02
3.55814286e-02 1.37496328e+00 -1.02454960e+00 2.02460811e-02
-4.07230318e-01 9.75964308e-01 2.65147686e-01 7.90646300e-02
-5.29250145e-01 -4.45819646e-01 7.17541277e-01 3.15084189e-01
-3.29568721e-02 -4.92361307e-01 -7.59464085e-01 -9.76926684e-01
6.41064346e-01 -3.29914719e-01 2.13779926e-01 -4.23208475e-01
-9.37475502e-01 1.38418660e-01 -1.08512752e-01 -1.41300654e+00
-9.37993526e-02 -1.06205511e+00 -5.60106993e-01 4.04880285e-01
-1.46822071e+00 -7.83030272e-01 -3.98496598e-01 2.36696303e-01
5.51307425e-02 5.06423295e-01 5.80721200e-01 1.27261236e-01
-1.99817285e-01 2.87238091e-01 1.23643294e-01 1.32776827e-01
2.72602767e-01 -1.57962620e+00 2.99496114e-01 8.62092435e-01
3.55342925e-01 3.53408992e-01 8.93651485e-01 -7.77150616e-02
-1.57183111e+00 -5.60555100e-01 8.53696704e-01 -6.75952077e-01
7.53457785e-01 -3.48147035e-01 -8.16637039e-01 2.03706399e-01
-2.11332440e-01 4.27506506e-01 4.37252343e-01 -2.80548275e-01
-4.74803001e-01 -1.77622005e-01 -1.26485443e+00 5.19623637e-01
1.23567879e+00 -5.54594219e-01 -4.93294299e-01 1.91831559e-01
2.75087088e-01 2.80902795e-02 -9.29118931e-01 4.64481413e-01
5.29895544e-01 -1.34993637e+00 1.13876653e+00 -8.78451318e-02
3.45722675e-01 -2.15537742e-01 -1.70767039e-01 -1.13460410e+00
-1.73145551e-02 -5.47561407e-01 -2.58789569e-01 8.25836420e-01
1.44343555e-01 -6.37100160e-01 6.89504266e-01 5.72046518e-01
9.41797197e-02 -1.00652564e+00 -1.33778906e+00 -9.77854908e-01
1.13970950e-01 -5.52044094e-01 2.83093154e-01 8.21918488e-01
4.76445585e-01 -1.43733442e-01 -1.64113566e-02 -1.10781282e-01
7.89982021e-01 -4.14127819e-02 2.97912061e-01 -1.36564159e+00
-1.08048044e-01 -9.17879999e-01 -1.27971268e+00 -9.76861596e-01
-2.47205272e-01 -8.76814425e-01 -9.71589684e-02 -1.21899974e+00
9.47497189e-02 -5.49859285e-01 -3.77187669e-01 -9.76597890e-02
-3.01753636e-02 5.76772809e-01 1.75076082e-01 -1.20485481e-02
-3.70189011e-01 2.35604838e-01 1.18417692e+00 -1.28288746e-01
-8.64573270e-02 -1.72833234e-01 -4.74559307e-01 8.09862733e-01
6.40434325e-01 -1.36301920e-01 -2.19853505e-01 -7.06482530e-02
6.27137959e-01 -4.29491788e-01 5.20889044e-01 -1.18054259e+00
1.39716372e-01 -1.41825289e-01 1.44282430e-01 -9.96756926e-03
1.88027740e-01 -7.21799493e-01 -2.81499833e-01 5.78454971e-01
-2.12353840e-01 -2.19823092e-01 -2.38502532e-01 3.66675496e-01
-1.75848588e-01 -3.86610925e-01 9.29813445e-01 1.42061442e-01
-9.32926059e-01 2.60237083e-02 1.88527033e-01 3.06504369e-01
1.39812410e+00 -5.72161019e-01 -1.20739520e-01 -2.47090220e-01
-3.19800586e-01 -2.01615348e-01 7.66520739e-01 2.41093233e-01
6.51019156e-01 -1.33777225e+00 -6.20728254e-01 2.49663964e-01
2.20998019e-01 -4.51515675e-01 1.78718001e-01 1.06286943e+00
-8.74341071e-01 2.68464416e-01 -3.12702924e-01 -6.75842583e-01
-1.16335905e+00 7.92336702e-01 3.10519665e-01 5.33885956e-02
-7.36008644e-01 5.77477157e-01 1.77813232e-01 1.27056733e-01
1.17619745e-01 -7.20827043e-01 -1.01587623e-01 1.36605516e-01
7.02192307e-01 8.06509972e-01 2.86907256e-01 -7.14537978e-01
-5.30794680e-01 1.14988899e+00 6.15862370e-01 -1.91620380e-01
1.11367357e+00 -2.23583877e-01 -1.99269578e-01 5.16276121e-01
1.61842287e+00 -1.26316950e-01 -1.27153683e+00 2.91878404e-03
2.65481085e-01 -4.98569548e-01 1.35360003e-01 -2.49801174e-01
-9.09320891e-01 9.21861291e-01 6.88795269e-01 8.03977430e-01
1.37598515e+00 1.13437951e-01 6.61346376e-01 1.17612436e-01
3.16530228e-01 -1.26426232e+00 8.72089192e-02 1.14638761e-01
8.58929694e-01 -1.27583134e+00 5.61393276e-02 -6.36837184e-01
3.80650461e-02 1.23063183e+00 -1.50119394e-01 -2.27664351e-01
8.50985408e-01 2.36208916e-01 3.32405716e-02 -1.43900350e-01
-3.61604877e-02 -4.07760620e-01 5.08168757e-01 6.55439615e-01
4.95220482e-01 2.00154975e-01 -7.60364652e-01 -9.48112980e-02
-3.36646140e-01 -2.42701657e-02 7.39762723e-01 9.81893897e-01
-7.07350671e-01 -8.13939631e-01 -3.43325466e-01 2.83590853e-01
-3.11800987e-01 1.24438442e-01 -4.43365157e-01 7.28762925e-01
8.95566419e-02 8.75306189e-01 1.58109069e-01 -2.35745430e-01
2.87043303e-01 -7.60169253e-02 7.81129479e-01 -3.81195128e-01
-1.50663350e-02 -6.82279110e-01 -2.82099128e-01 -6.85040116e-01
-4.91034538e-01 -5.44299245e-01 -1.09107924e+00 -1.95062339e-01
-2.11982697e-01 -1.14119448e-01 9.42684889e-01 9.67025399e-01
1.41916782e-01 9.15280208e-02 8.39078665e-01 -7.68920958e-01
-6.62521839e-01 -7.17986643e-01 -8.71889293e-01 8.01133573e-01
6.69425368e-01 -7.75226712e-01 -7.65238106e-01 1.52792096e-01] | [7.544579029083252, 4.066821098327637] |
27733ef2-1b0f-4df0-9610-0cbf8043b203 | dockstring-easy-molecular-docking-yields | 2110.15486 | null | https://arxiv.org/abs/2110.15486v1 | https://arxiv.org/pdf/2110.15486v1.pdf | DOCKSTRING: easy molecular docking yields better benchmarks for ligand design | The field of machine learning for drug discovery is witnessing an explosion of novel methods. These methods are often benchmarked on simple physicochemical properties such as solubility or general druglikeness, which can be readily computed. However, these properties are poor representatives of objective functions in drug design, mainly because they do not depend on the candidate's interaction with the target. By contrast, molecular docking is a widely successful method in drug discovery to estimate binding affinities. However, docking simulations require a significant amount of domain knowledge to set up correctly which hampers adoption. To this end, we present DOCKSTRING, a bundle for meaningful and robust comparison of ML models consisting of three components: (1) an open-source Python package for straightforward computation of docking scores; (2) an extensive dataset of docking scores and poses of more than 260K ligands for 58 medically-relevant targets; and (3) a set of pharmaceutically-relevant benchmark tasks including regression, virtual screening, and de novo design. The Python package implements a robust ligand and target preparation protocol that allows non-experts to obtain meaningful docking scores. Our dataset is the first to include docking poses, as well as the first of its size that is a full matrix, thus facilitating experiments in multiobjective optimization and transfer learning. Overall, our results indicate that docking scores are a more appropriate evaluation objective than simple physicochemical properties, yielding more realistic benchmark tasks and molecular candidates. | ['Sergio Bacallado', 'Andreas Bender', 'José Miguel Hernández-Lobato', 'Austin J. Tripp', 'Gregor N. C. Simm', 'Miguel García-Ortegón'] | 2021-10-29 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 4.90053333e-02 -4.66371387e-01 -5.03348351e-01 -1.97086796e-01
-1.12383258e+00 -8.77659261e-01 4.14400935e-01 6.72596097e-01
-5.41414797e-01 1.35039008e+00 -2.04739481e-01 -6.30549431e-01
-2.20379546e-01 -3.97907853e-01 -8.13644350e-01 -9.22575712e-01
-3.38164002e-01 6.80334985e-01 -8.59400164e-03 -2.47600198e-01
2.65991628e-01 8.74209285e-01 -1.18804848e+00 -8.61724187e-03
1.06765354e+00 8.13100219e-01 1.74273476e-01 1.84401959e-01
3.26556027e-01 2.85813034e-01 -4.13110882e-01 -3.37123454e-01
-5.46327680e-02 -3.91233385e-01 -6.90810442e-01 -4.53484237e-01
2.11790010e-01 3.69824246e-02 3.51424485e-01 7.96622097e-01
1.02974236e+00 2.07576424e-01 8.97625506e-01 -6.25672936e-01
-3.11465144e-01 -1.04885995e-01 4.24410813e-02 -6.39336333e-02
7.07376599e-01 5.26256323e-01 1.10864878e+00 -1.14335763e+00
5.86977124e-01 8.78152490e-01 5.49133956e-01 2.69796312e-01
-1.43397212e+00 -7.65420914e-01 -2.29916602e-01 1.31750226e-01
-1.48299885e+00 -4.16848719e-01 3.26072186e-01 -7.13720798e-01
1.08336747e+00 4.81720448e-01 4.88103271e-01 1.10797596e+00
3.36323202e-01 2.06783205e-01 1.12613428e+00 -1.16305798e-01
7.70545661e-01 2.72636384e-01 -2.43509129e-01 4.70254689e-01
1.30547076e-01 2.71882743e-01 -4.11645889e-01 -7.19716311e-01
4.68272954e-01 -1.02839954e-01 -3.99420261e-01 -6.07493162e-01
-1.18915415e+00 1.12794006e+00 4.46611315e-01 -4.51744832e-02
-5.17713070e-01 -2.87293285e-01 1.84308127e-01 1.49996817e-01
1.25652805e-01 9.77910399e-01 -9.45999086e-01 -2.33273511e-03
-5.26186466e-01 4.72363174e-01 1.00822175e+00 3.13365191e-01
6.91662908e-01 -2.70034641e-01 1.53966501e-01 6.79324687e-01
3.39383155e-01 3.74045819e-01 3.12534064e-01 -4.46352392e-01
1.08660825e-01 3.77450883e-01 3.97977591e-01 -8.08510780e-01
-5.99594831e-01 -3.71581793e-01 -5.16709983e-01 4.94344503e-01
5.74598372e-01 -1.01245213e-02 -6.55468881e-01 1.77456999e+00
5.73432267e-01 -2.34678730e-01 1.95191771e-01 9.05323148e-01
9.72281456e-01 4.82863575e-01 4.88019079e-01 -5.93246222e-01
1.12460899e+00 -5.99317670e-01 -2.98066258e-01 1.72570318e-01
7.06206620e-01 -9.41087663e-01 8.46534193e-01 6.12247586e-01
-9.88342404e-01 -2.63774931e-01 -1.02193499e+00 2.48553365e-01
-6.38048589e-01 -8.09070021e-02 1.03449214e+00 6.59832358e-01
-7.24279702e-01 1.12297666e+00 -6.45432651e-01 -8.97749811e-02
3.83633852e-01 9.66448009e-01 -7.65357673e-01 1.20354556e-01
-1.00768197e+00 1.31506968e+00 5.30456245e-01 -1.35116741e-01
-7.96493053e-01 -9.08619881e-01 -7.41950750e-01 -1.86178401e-01
2.42445856e-01 -6.40775979e-01 8.85407388e-01 -6.28444970e-01
-1.68703318e+00 6.86091483e-01 1.47381695e-02 5.30895889e-02
3.51312429e-01 1.59127861e-01 -1.38384044e-01 -1.35540634e-01
5.98584395e-03 4.00049239e-01 2.62080133e-01 -7.97642350e-01
-1.28534660e-01 -4.19553459e-01 -5.86195216e-02 3.51034224e-01
2.76080836e-02 1.44109517e-01 -2.31178448e-01 -3.68234187e-01
-1.62871376e-01 -8.22503150e-01 -6.46394968e-01 -6.25423044e-02
-5.47339857e-01 -9.14922208e-02 -4.10082154e-02 -4.80607778e-01
9.74985778e-01 -1.74114287e+00 4.67853099e-01 4.92382854e-01
2.72079557e-01 2.55338967e-01 -1.81562915e-01 6.88611805e-01
-5.45578361e-01 1.10228330e-01 -3.14508602e-02 2.45587662e-01
-2.16873690e-01 -1.68789163e-01 -3.90261188e-02 7.01555967e-01
2.09382489e-01 8.96284223e-01 -8.46731365e-01 -2.48990864e-01
1.63797215e-01 4.82362598e-01 -7.33650506e-01 1.11374490e-01
-4.90456283e-01 9.01468933e-01 -8.46423209e-01 8.21821034e-01
4.37909752e-01 -4.71883297e-01 4.42238241e-01 -1.17782243e-01
-2.21212000e-01 4.93668735e-01 -9.78505909e-01 1.45291328e+00
8.26676413e-02 -1.93210915e-01 -2.62483835e-01 -6.94705665e-01
8.09725583e-01 3.19409609e-01 8.62570584e-01 -5.64022362e-01
2.71587372e-01 4.45406824e-01 2.95240551e-01 -1.55507594e-01
-2.11093277e-01 -2.71138996e-01 2.17068151e-01 1.12380423e-01
-3.68334167e-02 -1.37867220e-02 3.40760976e-01 -3.05402040e-01
1.18800855e+00 2.12073609e-01 6.86961591e-01 -3.37082863e-01
5.45486510e-01 2.63176352e-01 6.03298903e-01 1.37973651e-01
2.15659931e-01 1.98163033e-01 3.38453382e-01 -6.08878911e-01
-9.52459157e-01 -1.02679658e+00 -7.05467403e-01 9.59024012e-01
-1.98670663e-02 -4.84037429e-01 -4.60368633e-01 -2.87507355e-01
1.75265953e-01 2.09387124e-01 -5.35667777e-01 5.89289423e-03
-1.85108766e-01 -1.36181855e+00 2.68881291e-01 1.15272120e-01
-4.25847232e-01 -9.14843559e-01 5.14062382e-02 5.19997835e-01
3.01749438e-01 -6.54756486e-01 -2.96640903e-01 6.56594574e-01
-7.22510040e-01 -1.62725103e+00 -7.69208968e-01 -6.45979404e-01
4.84914988e-01 -1.44711494e-01 9.01611984e-01 -6.03492856e-02
-3.60100448e-01 -1.78708583e-01 -1.01026408e-01 -5.82589269e-01
-4.93592232e-01 -1.11261189e-01 3.93873394e-01 -2.66086340e-01
3.19305837e-01 -7.88530886e-01 -6.64715707e-01 6.00367367e-01
-5.66684842e-01 -2.21156344e-01 7.88826406e-01 8.39085221e-01
1.08868384e+00 -3.36846322e-01 5.32157958e-01 -7.07925737e-01
7.27805078e-01 -4.39725429e-01 -7.51312375e-01 1.73277333e-01
-5.38957119e-01 6.72887266e-02 7.35315561e-01 -5.17248869e-01
-3.00328642e-01 4.29709554e-01 -4.67812240e-01 -2.14618724e-02
-7.95138702e-02 9.09550190e-01 -3.19574684e-01 -6.26145959e-01
1.01430345e+00 9.76123065e-02 1.94191024e-01 -7.30569720e-01
-5.69462813e-02 3.14730167e-01 1.15892619e-01 -7.77395725e-01
7.72924066e-01 -2.88090985e-02 3.73358428e-01 -8.79730225e-01
-4.56438124e-01 -3.98246229e-01 -5.93348920e-01 2.54278153e-01
6.65373743e-01 -8.70878279e-01 -1.36439443e+00 -2.29288712e-02
-8.83418679e-01 -3.81610453e-01 2.67799765e-01 9.68135595e-01
-4.79320645e-01 4.71996963e-01 -3.69438857e-01 -3.67611259e-01
-3.79297972e-01 -1.72346258e+00 8.26048255e-01 -6.48956746e-02
-2.41649687e-01 -9.52355981e-01 4.45731640e-01 2.95827210e-01
1.61418155e-01 7.94129133e-01 1.17620420e+00 -9.25673723e-01
-5.02872288e-01 -3.30012888e-01 9.87098962e-02 1.18045472e-01
2.45637596e-01 8.17100927e-02 -8.41750979e-01 -4.30202216e-01
-3.67873937e-01 -6.14918709e-01 5.41421354e-01 4.83627796e-01
9.02937174e-01 -1.94248319e-01 -5.36893010e-01 6.88329339e-01
1.27270663e+00 4.66224760e-01 6.33753717e-01 4.12685841e-01
5.21539867e-01 3.58188152e-01 6.67684376e-01 5.31624377e-01
8.51222724e-02 1.01600003e+00 6.01567328e-01 -5.26996315e-01
5.96882522e-01 1.04231849e-01 2.27759108e-01 3.92503113e-01
-5.38382053e-01 8.39962717e-03 -9.43088472e-01 -2.79437482e-01
-1.55631232e+00 -8.80870700e-01 -1.20182663e-01 2.63210440e+00
1.46833360e+00 -1.46736369e-01 3.80597144e-01 -1.35079414e-01
2.23205388e-01 -4.18335259e-01 -8.73322725e-01 -1.93070099e-01
-1.85041353e-01 6.43722653e-01 3.57219070e-01 4.87732589e-01
-1.12486637e+00 6.98783755e-01 6.82869959e+00 9.98588800e-01
-1.24870825e+00 -3.30879509e-01 6.44935548e-01 3.30858082e-02
-8.00048709e-02 -7.32069016e-02 -8.29454303e-01 4.76564437e-01
9.81067240e-01 -2.32345521e-01 3.53107721e-01 9.36254978e-01
5.02489209e-01 -3.54962200e-02 -1.44134736e+00 8.64302397e-01
-4.03432578e-01 -1.63705266e+00 9.06722769e-02 4.03364748e-01
4.79643583e-01 9.86657590e-02 1.47873610e-01 4.33031321e-02
1.49375722e-01 -1.56419420e+00 3.17853957e-01 2.61671036e-01
9.78556633e-01 -7.48581231e-01 5.01617432e-01 2.29528219e-01
-7.83022642e-01 1.78955272e-01 -4.80592906e-01 1.78509355e-01
-2.18848020e-01 4.46874082e-01 -8.59633863e-01 5.44833243e-01
3.29318255e-01 6.42514825e-01 -5.13507545e-01 1.39604008e+00
2.80588642e-02 1.82913348e-01 -3.43336552e-01 -1.54732406e-01
2.74598956e-01 -4.69635546e-01 1.54925048e-01 9.72680688e-01
-3.44508663e-02 1.44827932e-01 5.52321076e-01 6.78933024e-01
-3.22758034e-02 7.42128074e-01 -3.93142402e-01 -1.34302959e-01
2.67578006e-01 1.14383996e+00 -4.52520549e-01 1.19508483e-01
-3.91071767e-01 5.40850282e-01 2.01123536e-01 3.98577094e-01
-9.60885882e-01 -1.96133465e-01 1.24725902e+00 -6.58298060e-02
-8.51870049e-03 -7.77377486e-02 1.15788750e-01 -8.74184012e-01
-1.62048042e-01 -1.31017125e+00 1.84619367e-01 -3.81176829e-01
-1.22772384e+00 5.10897577e-01 -2.49885485e-01 -1.12950850e+00
9.24266316e-03 -1.14580095e+00 -4.11033750e-01 1.13547015e+00
-1.61703050e+00 -7.19537139e-01 3.20858508e-02 5.11765778e-01
7.11138025e-02 -8.60158429e-02 1.20207822e+00 6.03832722e-01
-7.21955180e-01 4.25893307e-01 5.75880468e-01 -3.88967752e-01
9.68310356e-01 -1.11537337e+00 -3.11371591e-02 8.58608051e-04
-2.24213898e-01 9.62775946e-01 8.11580181e-01 -6.18629098e-01
-1.63596082e+00 -8.90106499e-01 4.83220041e-01 -7.94170201e-01
6.61352336e-01 -2.02749416e-01 -8.37469459e-01 2.48164982e-01
-5.06305695e-01 -1.89085335e-01 1.24940741e+00 5.01284488e-02
-1.94810763e-01 2.66313314e-01 -9.01691556e-01 5.22062898e-01
6.18317604e-01 -2.25104690e-01 -1.59558188e-02 1.05236578e+00
2.17615440e-01 -6.42439663e-01 -1.35616088e+00 5.41221619e-01
5.55485964e-01 -7.47045577e-01 1.32538509e+00 -1.01565981e+00
1.45866632e-01 -4.33992416e-01 -9.39142704e-02 -1.02722621e+00
-4.74146783e-01 -6.79396808e-01 2.74988174e-01 4.86800879e-01
9.90191638e-01 -5.76662958e-01 8.39914799e-01 6.70992076e-01
-9.22099650e-02 -1.33992827e+00 -9.27181184e-01 -8.43676209e-01
2.51284748e-01 -1.31131858e-01 3.70542884e-01 1.03150892e+00
1.13374077e-01 6.00520909e-01 -2.52332628e-01 4.08992022e-02
3.67887080e-01 4.51201312e-02 7.73602426e-01 -1.60206532e+00
-5.59501886e-01 -5.06622910e-01 -3.56277615e-01 -7.13406980e-01
-2.86548138e-02 -1.03883266e+00 -3.49542379e-01 -1.13806081e+00
3.93990815e-01 -4.47916001e-01 -1.17565542e-01 5.93368649e-01
-1.28338113e-01 1.35090604e-01 -4.57633674e-01 3.46051812e-01
-4.13501114e-01 5.26417851e-01 1.29942036e+00 -9.55585912e-02
-5.57791114e-01 1.47848725e-01 -6.32599950e-01 5.24499655e-01
6.76200151e-01 -4.28401053e-01 -1.91059381e-01 3.88807535e-01
3.22460979e-01 -2.49206815e-02 1.89768553e-01 -6.21172488e-01
-1.37473211e-01 -6.20160758e-01 6.78025663e-01 -2.01353565e-01
5.40156662e-01 -5.98443806e-01 5.22407532e-01 5.83094954e-01
1.18273152e-02 -2.18493029e-01 3.17623705e-01 4.64779079e-01
-3.75028700e-02 4.76636775e-02 8.89556408e-01 -6.74254149e-02
-3.65067601e-01 6.31261826e-01 -2.32817292e-01 -1.86007693e-01
9.58305359e-01 -5.02317488e-01 -6.81503862e-02 -7.57216439e-02
-8.23821604e-01 7.83892348e-02 6.30456567e-01 1.53913079e-02
3.91131103e-01 -9.98815477e-01 -6.38486087e-01 -5.97924404e-02
2.40947232e-01 -3.63230050e-01 -2.30347589e-01 1.03410172e+00
-6.83274269e-01 7.44467437e-01 -1.42611727e-01 -4.50029701e-01
-1.40304685e+00 7.80691981e-01 4.50115740e-01 -1.95460439e-01
-1.22295208e-01 5.73523521e-01 3.61027867e-01 -3.77890885e-01
2.53319800e-01 -1.34449154e-01 -2.52502322e-01 2.24297866e-02
5.99362850e-01 7.69193023e-02 3.44269454e-01 -5.78251660e-01
-7.33575225e-01 6.01714313e-01 -8.49789456e-02 4.86204445e-01
1.63357151e+00 5.11874259e-01 -1.01624660e-01 -1.00215778e-01
1.11422527e+00 -1.50370419e-01 -9.57013905e-01 5.35976179e-02
1.52324826e-01 -2.08258748e-01 -8.48261267e-02 -1.05585420e+00
-3.09427202e-01 4.56566244e-01 4.85240191e-01 -3.64479065e-01
7.82070458e-01 -2.88107954e-02 3.77538562e-01 8.39505136e-01
4.70466703e-01 -7.18419254e-01 1.33231103e-01 3.39153320e-01
9.80801523e-01 -1.33576572e+00 3.20240557e-01 -4.79015470e-01
-3.48860174e-01 1.06992507e+00 2.44011387e-01 4.29619282e-01
4.39255893e-01 -3.53034064e-02 -9.91591513e-02 -3.78514171e-01
-5.59939384e-01 -1.78135380e-01 5.55261970e-01 6.07997477e-01
7.72490144e-01 -2.83894271e-01 -4.87664253e-01 5.74072659e-01
1.23778339e-02 -1.46195844e-01 -2.17561517e-02 8.27748716e-01
-4.65945065e-01 -1.65541303e+00 -4.37220007e-01 2.42136240e-01
-6.29533708e-01 -3.08892846e-01 -7.10941017e-01 8.10083091e-01
-1.50577500e-01 7.58603513e-01 -7.19348848e-01 -1.46447152e-01
4.12782967e-01 -1.10369876e-01 6.08717740e-01 -7.47696757e-01
-5.36209583e-01 1.48164481e-01 4.85705808e-02 -6.46756232e-01
-2.63791054e-01 -2.77777404e-01 -1.10369682e+00 -3.87456536e-01
-6.49447858e-01 5.86260438e-01 8.49709809e-01 8.88322234e-01
3.63612175e-01 9.47871804e-02 6.44549906e-01 -1.12204826e+00
-5.94545364e-01 -6.58243060e-01 -7.09477663e-01 2.06140548e-01
1.15143932e-01 -8.20834756e-01 -1.43769562e-01 -5.67058697e-02] | [4.973506927490234, 5.558060646057129] |
7e6ed829-2fe6-48e2-8632-0a5e11432d18 | global-wheat-challenge-2020-analysis-of-the | 2105.06182 | null | https://arxiv.org/abs/2105.06182v1 | https://arxiv.org/pdf/2105.06182v1.pdf | Global Wheat Challenge 2020: Analysis of the competition design and winning models | Data competitions have become a popular approach to crowdsource new data analysis methods for general and specialized data science problems. In plant phenotyping, data competitions have a rich history, and new outdoor field datasets have potential for new data competitions. We developed the Global Wheat Challenge as a generalization competition to see if solutions for wheat head detection from field images would work in different regions around the world. In this paper, we analyze the winning challenge solutions in terms of their robustness and the relative importance of model and data augmentation design decisions. We found that the design of the competition influence the selection of winning solutions and provide recommendations for future competitions in an attempt to garner more robust winning solutions. | ['Ian Stavness', 'Frederic Baret', 'Wei Guo', 'Franklin Ogidi', 'Etienne David'] | 2021-05-13 | null | null | null | null | ['head-detection', 'plant-phenotyping'] | ['computer-vision', 'computer-vision'] | [-1.05040431e-01 -1.22294940e-01 -9.77268219e-02 -4.50231701e-01
-5.00322998e-01 -1.17435658e+00 2.51141489e-01 7.73636162e-01
-2.75123924e-01 4.38809216e-01 1.75404653e-01 -5.07864296e-01
-1.76142737e-01 -8.56112778e-01 -6.86120450e-01 -3.34719747e-01
1.24993213e-01 5.83412409e-01 5.33603072e-01 -6.24300599e-01
1.20242611e-01 7.25678027e-01 -1.65966427e+00 3.12895954e-01
7.24873543e-01 5.09927988e-01 5.66973686e-01 7.54009247e-01
-2.46135481e-02 7.19282866e-01 -1.04951119e+00 -6.25269488e-02
3.61652970e-01 -3.37693607e-03 -9.38773692e-01 2.76581675e-01
3.00894260e-01 -1.82309553e-01 1.46589205e-01 6.21821165e-01
9.76892769e-01 -5.55403829e-02 -4.67511527e-02 -1.53758562e+00
-8.87742937e-01 4.76539284e-01 -7.60059893e-01 3.99348348e-01
5.32973051e-01 3.46495718e-01 8.73447120e-01 -5.93810141e-01
9.77803886e-01 1.00994658e+00 8.58867109e-01 1.92434892e-01
-1.39793444e+00 -3.72921586e-01 3.53347927e-01 4.77359854e-02
-1.44597995e+00 -3.70053768e-01 4.68754619e-01 -4.41737175e-01
6.31347120e-01 5.82971871e-01 8.25542152e-01 7.49864757e-01
-1.10216945e-01 9.49942470e-01 1.08860636e+00 -4.58108366e-01
5.47115326e-01 -4.51754928e-02 6.09174445e-02 2.72927552e-01
4.80136871e-01 -1.47203298e-03 -6.45334959e-01 -4.53673631e-01
5.40753424e-01 -4.84921545e-01 -9.74669829e-02 -6.61072731e-01
-9.73635733e-01 1.05603290e+00 7.85638154e-01 1.75372049e-01
-2.15861008e-01 -4.06667233e-01 3.61760467e-01 -2.43990328e-02
4.81340557e-01 1.00800645e+00 -7.65174448e-01 2.65798539e-01
-7.45945632e-01 9.30113196e-01 7.44462430e-01 8.94640565e-01
5.68441689e-01 -1.01760499e-01 -1.49539709e-01 7.36225426e-01
1.54643089e-01 3.26813072e-01 1.87674239e-01 -1.11953545e+00
1.57395691e-01 8.99600446e-01 6.01012647e-01 -9.66630578e-01
-7.51080096e-01 -3.53458617e-03 2.66699046e-02 3.42262149e-01
8.31131637e-01 -5.00342667e-01 -8.66397500e-01 1.42075157e+00
5.53645849e-01 -3.28436166e-01 -1.90014854e-01 1.17618370e+00
1.02243960e+00 2.58950263e-01 2.51281887e-01 3.37730408e-01
1.21952951e+00 -4.83670264e-01 -6.05744898e-01 -6.73156023e-01
1.09133589e+00 -9.71221030e-01 1.23292804e+00 2.12060601e-01
-9.01982605e-01 -4.53104228e-01 -8.95468712e-01 -1.78059176e-01
-7.22982705e-01 7.87413344e-02 7.37035930e-01 7.71962583e-01
-9.15396035e-01 2.90456325e-01 -6.32066667e-01 -1.07425702e+00
5.39421380e-01 2.13510439e-01 -3.36319625e-01 -1.00277439e-01
-7.26494133e-01 1.11161673e+00 1.64524078e-01 2.36761421e-01
-9.70666409e-01 -7.05056846e-01 -6.02149546e-01 -3.31263840e-01
3.95441443e-01 1.94058928e-03 1.26005995e+00 -5.44662356e-01
-1.10403728e+00 1.38297474e+00 -1.35181937e-02 -2.92027295e-01
1.26012355e-01 4.29448821e-02 -4.37931120e-01 -4.56449836e-01
5.10266960e-01 1.03490686e+00 9.97049734e-02 -1.11454046e+00
-1.02411163e+00 -7.55251229e-01 2.55395710e-01 1.72011539e-01
3.78912948e-02 2.30508447e-01 1.57544494e-01 -5.44059634e-01
-1.16600022e-01 -1.08624458e+00 -5.34221232e-01 3.59617770e-01
-9.16293706e-04 -1.28204813e-02 9.15669024e-01 -4.75326747e-01
5.31671822e-01 -2.18080568e+00 -7.37546906e-02 -1.65536299e-01
1.27624109e-01 4.86441374e-01 -4.14426684e-01 5.51276684e-01
3.60433251e-01 2.73259014e-01 1.38453841e-01 3.31519961e-01
-2.38964409e-01 2.96650708e-01 -1.16119698e-01 4.49836731e-01
4.28649127e-01 7.40501225e-01 -9.11480486e-01 9.56765097e-03
3.50514323e-01 1.36770681e-01 -1.96198702e-01 7.48047158e-02
-4.48435456e-01 3.93771410e-01 -3.17772359e-01 9.70338643e-01
9.59015310e-01 -1.03138126e-01 9.05852169e-02 1.17427893e-01
-4.87933010e-01 -2.98332244e-01 -1.26687050e+00 1.71968079e+00
8.96964595e-02 6.93043411e-01 6.23643517e-01 -5.89906871e-01
1.30148840e+00 8.34745765e-02 2.98861265e-01 -4.13528860e-01
3.95637043e-02 2.98305780e-01 2.86334068e-01 -4.21371251e-01
6.08944237e-01 3.18871498e-01 7.92223364e-02 8.45456198e-02
-2.37772584e-01 -4.88522112e-01 3.36914062e-01 -1.13721311e-01
1.21131396e+00 3.22444111e-01 2.06481442e-01 -8.33455741e-01
-4.52092588e-02 1.00519490e+00 5.71087241e-01 7.57051647e-01
-9.45044577e-01 7.66541481e-01 1.06032178e-01 -9.51931536e-01
-8.52342844e-01 -6.76514447e-01 -8.77352729e-02 1.48663139e+00
3.59173924e-01 -3.38912904e-01 -6.01543903e-01 -4.63972896e-01
2.49410033e-01 4.16431069e-01 -7.16922224e-01 4.16573100e-02
-1.86281670e-02 -9.81702626e-01 6.80490017e-01 8.02950144e-01
1.90360993e-01 -1.27196133e+00 -1.16696513e+00 2.18905374e-01
-3.25927347e-01 -1.13434267e+00 4.31479551e-02 8.45163226e-01
-4.91534054e-01 -1.09700882e+00 -7.65200019e-01 -8.40352118e-01
3.91026795e-01 9.32387412e-01 1.11957574e+00 3.00513711e-02
-7.73216665e-01 7.56290331e-02 -7.84837425e-01 -1.28431225e+00
-3.49638537e-02 4.38022494e-01 -2.78546363e-01 -6.25667393e-01
9.20415163e-01 1.05812535e-01 -4.07478720e-01 6.33342326e-01
-9.09245312e-01 -4.33987468e-01 1.97086576e-02 4.19365734e-01
4.39517766e-01 -2.39903912e-01 4.86546278e-01 -7.68665969e-01
8.16533685e-01 -5.20730376e-01 -9.62961555e-01 4.64715570e-01
-7.13374391e-02 -3.31432581e-01 2.02863172e-01 -1.46174610e-01
-8.56296778e-01 5.00737071e-01 1.75585508e-01 1.79874808e-01
-6.19674563e-01 7.03239977e-01 -1.39748737e-01 -3.98659647e-01
1.48511660e+00 -8.56151342e-01 -9.72241685e-02 -4.20853853e-01
3.35721999e-01 4.84620571e-01 2.70614594e-01 -5.73197842e-01
4.24661189e-01 6.86930716e-01 -2.28006944e-01 -9.19708014e-01
-7.87666023e-01 -4.51345652e-01 -6.85694277e-01 -6.32351637e-02
9.33378398e-01 -1.13390625e+00 -5.94073772e-01 4.59679902e-01
-9.94160652e-01 -6.69212103e-01 -7.50459731e-01 1.69734463e-01
-6.18853373e-03 -9.99617726e-02 -1.83554649e-01 -6.30654931e-01
1.83629796e-01 -1.25545740e+00 1.26453078e+00 5.94328344e-01
-4.65693891e-01 -7.48529196e-01 2.99160659e-01 5.05255640e-01
4.83587980e-01 5.44379771e-01 1.84603959e-01 -5.04126072e-01
-6.31439090e-02 -4.23712045e-01 -3.59509587e-01 -2.91516483e-01
2.77373642e-01 3.92951518e-01 -1.30988467e+00 -3.43152285e-01
-4.61889595e-01 -6.14941537e-01 2.52099305e-01 6.35767937e-01
9.69081342e-01 5.91631651e-01 -3.45633090e-01 2.32214078e-01
1.23686135e+00 1.11703306e-01 4.68943447e-01 6.89005494e-01
5.71131051e-01 1.02827454e+00 9.03453350e-01 5.72731912e-01
3.28278214e-01 5.45856893e-01 6.34381294e-01 -4.97899860e-01
4.41310592e-02 3.15986685e-02 -5.21653183e-02 -1.02930821e-01
2.84656156e-02 -5.16042113e-01 -1.27262115e+00 8.31242144e-01
-2.02751732e+00 -5.50955951e-01 -5.46669662e-01 2.06310606e+00
2.76512474e-01 -1.29432544e-01 5.50161004e-01 8.93974230e-02
7.25539625e-01 8.50090608e-02 -5.11456132e-01 -3.73550534e-01
-8.29655230e-01 2.42779970e-01 9.78636920e-01 4.52144235e-01
-1.06842613e+00 1.06984460e+00 7.80008793e+00 3.78255963e-01
-9.01237190e-01 -9.22891349e-02 3.99130195e-01 -4.08355659e-03
1.25570856e-02 2.61778265e-01 -8.08694243e-01 -1.16216065e-02
6.20618880e-01 -2.55481843e-02 3.84524137e-01 6.36837363e-01
2.00235784e-01 -4.65113819e-01 -6.69569314e-01 2.69135207e-01
-2.92102635e-01 -1.36617100e+00 -5.42590141e-01 1.07984900e-01
6.91519439e-01 4.28532630e-01 -3.87826264e-01 -4.22160700e-02
1.20777798e+00 -8.40631485e-01 6.79961622e-01 -1.17835671e-01
-1.00385593e-02 -3.56183529e-01 7.36694574e-01 2.68660307e-01
-1.21175039e+00 -2.44254962e-01 -7.14675725e-01 -6.17146611e-01
4.01196294e-02 4.89053011e-01 -8.24879110e-01 4.95208412e-01
1.20129359e+00 3.27858925e-01 -1.26215649e+00 1.28146553e+00
-4.78761755e-02 4.87826496e-01 -6.22362435e-01 -9.43460986e-02
7.77004808e-02 2.16648668e-01 2.27694109e-01 7.84742057e-01
7.19114691e-02 -1.99746490e-01 6.25684142e-01 4.74711388e-01
2.76526630e-01 1.48466393e-01 -9.24417555e-01 -9.43933874e-02
6.60976291e-01 1.30357504e+00 -1.21962106e+00 4.81423199e-01
-3.84314209e-01 7.04354882e-01 3.65574002e-01 1.80883646e-01
-4.51806813e-01 -1.43133506e-01 5.20424962e-01 3.47686023e-01
2.37707600e-01 -1.29698023e-01 -4.71107483e-01 -6.64326370e-01
-6.97017834e-02 -1.07740057e+00 5.08510828e-01 -1.09457052e+00
-1.39967906e+00 1.02859199e-01 -1.59860760e-01 -8.68721724e-01
2.87800103e-01 -8.57965112e-01 -2.98085004e-01 1.01964629e+00
-1.05857050e+00 -1.54980052e+00 -9.52356815e-01 3.15311730e-01
2.75191277e-01 -1.85181677e-01 1.17383158e+00 -2.03795522e-03
-5.44530988e-01 2.47828618e-01 8.85623768e-02 -1.11321963e-01
8.98371398e-01 -1.02819645e+00 8.35692406e-01 7.37610877e-01
3.47493321e-01 2.10949890e-02 8.89482677e-01 -7.68224537e-01
-1.27467740e+00 -8.47008824e-01 5.04937887e-01 -8.89458716e-01
5.65148592e-01 -4.06465769e-01 -6.45073593e-01 7.13590026e-01
1.95061028e-01 3.40543747e-01 7.74921656e-01 3.75217021e-01
3.75190675e-02 -2.63067055e-02 -1.39737153e+00 5.05510390e-01
8.96870315e-01 -3.51082265e-01 3.22871357e-01 5.49048185e-01
4.32658881e-01 -5.97601771e-01 -6.91439688e-01 4.17593777e-01
4.51052457e-01 -6.12102091e-01 5.86777210e-01 -6.55447960e-01
1.73028871e-01 -5.62199414e-01 -5.49567342e-01 -1.40394056e+00
-7.29603350e-01 -4.80854690e-01 5.93706965e-01 1.30971444e+00
4.86421794e-01 -2.66708046e-01 1.05730712e+00 7.47341454e-01
6.43428415e-02 -2.31584653e-01 -4.76147890e-01 -6.64520562e-01
3.01431388e-01 -1.21854268e-01 8.81997764e-01 9.21094298e-01
-2.06530556e-01 2.39022017e-01 -1.12733148e-01 4.36061740e-01
1.08557492e-01 -8.19690153e-02 1.20566833e+00 -1.63522172e+00
1.35727704e-01 -1.77845508e-01 -6.77963614e-01 -3.38225871e-01
-3.03595304e-01 -5.52171946e-01 1.56638905e-01 -1.68781734e+00
-2.41145566e-01 -5.15489042e-01 2.87288010e-01 5.97271919e-01
-3.71929049e-01 1.55840114e-01 3.83445084e-01 1.15053598e-02
-2.02982098e-01 -7.57457763e-02 9.81251657e-01 -7.91864023e-02
-4.11914051e-01 -1.55656785e-01 -1.09808373e+00 6.36821866e-01
1.03031743e+00 -3.22785228e-01 -4.34998989e-01 -1.09380126e+00
2.22276926e-01 -4.96627003e-01 5.14747322e-01 -1.00165236e+00
4.92616631e-02 -4.74634051e-01 5.89446604e-01 -4.03765291e-01
1.60429686e-01 -9.02675927e-01 1.12409428e-01 2.82564431e-01
-3.53573859e-01 2.96226174e-01 7.86779046e-01 2.21199363e-01
1.19527273e-01 -3.91846210e-01 5.18298090e-01 -3.04777890e-01
-9.24915195e-01 -1.20612584e-01 -5.68748772e-01 1.96692824e-01
1.40863478e+00 -2.91203171e-01 -5.63211918e-01 -8.71079788e-02
-8.94246757e-01 3.86144191e-01 7.69637465e-01 7.01372743e-01
-1.11452334e-01 -9.69989777e-01 -9.31465805e-01 7.77093843e-02
5.61117768e-01 1.08590014e-01 -1.17920384e-01 3.56746405e-01
-9.16903019e-01 5.97039238e-02 -7.03252971e-01 -6.54959977e-01
-1.14533353e+00 5.24632454e-01 1.93014070e-01 -3.71179394e-02
-2.97246397e-01 9.20610070e-01 -8.42209160e-03 -9.60572898e-01
2.46278688e-01 -2.93999285e-01 -1.13466620e-01 2.32874811e-01
4.37954336e-01 4.72014189e-01 4.86539692e-01 -4.90009427e-01
-5.34777403e-01 -1.20664081e-02 8.12205002e-02 -9.80255604e-02
1.35026145e+00 1.04368553e-01 2.81840622e-01 5.41966915e-01
2.61924893e-01 -1.35783225e-01 -1.10595787e+00 1.02569237e-01
3.25146675e-01 -6.92532480e-01 1.40306741e-01 -1.04262114e+00
-9.96591389e-01 5.03591716e-01 1.20718873e+00 6.14184380e-01
8.92806470e-01 -2.25026067e-02 2.25286990e-01 1.70414895e-01
4.84517783e-01 -1.27494097e+00 -4.81277138e-01 1.64910078e-01
9.77250338e-01 -1.61194301e+00 1.39973342e-01 -4.81585681e-01
-7.74844468e-01 8.11567664e-01 9.52198148e-01 -1.41134724e-01
6.69096291e-01 7.20400155e-01 3.81873041e-01 -5.80185533e-01
-3.68364096e-01 -4.83246773e-01 -3.08491856e-01 1.42324615e+00
7.97814310e-01 4.70228970e-01 -2.27547050e-01 4.90803957e-01
4.47350629e-02 4.12828326e-01 5.97671211e-01 1.66493511e+00
-4.32532877e-01 -1.16436267e+00 -8.08206856e-01 4.63230997e-01
-9.30342532e-04 4.59300220e-01 -1.35791326e+00 7.50714004e-01
4.81970608e-01 1.40012753e+00 -1.63337961e-01 -2.47223631e-01
8.61092567e-01 -2.10063055e-01 3.64095420e-01 -8.61350954e-01
-1.07870007e+00 -1.03171736e-01 1.46200836e-01 -3.68982941e-01
-5.34776926e-01 -8.26781094e-01 -8.88594806e-01 -5.62343478e-01
-8.73506248e-01 -1.71796829e-01 6.73057973e-01 7.24487185e-01
6.95876181e-01 3.24800819e-01 2.23868191e-01 -8.91222298e-01
-2.52339005e-01 -8.62533450e-01 -6.13744020e-01 1.40308559e-01
-7.41746724e-02 -6.50165856e-01 1.83577538e-01 8.80443975e-02] | [9.219963073730469, -1.5273334980010986] |
0845cbec-6009-4aa1-a49f-22163d35d220 | submodboxes-near-optimal-search-for-a-set-of | null | null | http://papers.nips.cc/paper/5779-submodboxes-near-optimal-search-for-a-set-of-diverse-object-proposals | http://papers.nips.cc/paper/5779-submodboxes-near-optimal-search-for-a-set-of-diverse-object-proposals.pdf | SubmodBoxes: Near-Optimal Search for a Set of Diverse Object Proposals | This paper formulates the search for a set of bounding boxes (as needed in object proposal generation) as a monotone submodular maximization problem over the space of all possible bounding boxes in an image. Since the number of possible bounding boxes in an image is very large $O(#pixels^2)$, even a single linear scan to perform the greedy augmentation for submodular maximization is intractable. Thus, we formulate the greedy augmentation step as a Branch-and-Bound scheme. In order to speed up repeated application of B\&B, we propose a novel generalization of Minoux’s ‘lazy greedy’ algorithm to the B\&B tree. Theoretically, our proposed formulation provides a new understanding to the problem, and contains classic heuristic approaches such as Sliding Window+Non-Maximal Suppression (NMS) and and Efficient Subwindow Search (ESS) as special cases. Empirically, we show that our approach leads to a state-of-art performance on object proposal generation via a novel diversity measure. | ['Dhruv Batra', 'Qing Sun'] | 2015-12-01 | null | null | null | neurips-2015-12 | ['object-proposal-generation'] | ['computer-vision'] | [ 3.79063815e-01 4.07862782e-01 -4.61551636e-01 -2.73407072e-01
-9.16976988e-01 -6.02789938e-01 1.98793769e-01 1.62720263e-01
-4.12029713e-01 8.63960385e-01 -3.02613556e-01 -1.86268568e-01
-1.17512442e-01 -8.84809554e-01 -7.17141092e-01 -7.97117531e-01
-1.41818970e-01 7.23251283e-01 6.71141207e-01 -1.46196455e-01
3.88014287e-01 8.34109664e-01 -1.32145524e+00 -6.25508204e-02
8.88591468e-01 1.20324576e+00 4.40639585e-01 5.30343592e-01
-2.89863318e-01 1.71303496e-01 -5.24204910e-01 -3.84541571e-01
8.16431642e-01 -4.63940769e-01 -8.73673499e-01 6.47247910e-01
7.51772046e-01 -3.97131443e-01 1.13235842e-02 1.09982920e+00
4.39084828e-01 1.47030383e-01 5.22893369e-01 -1.29792130e+00
-1.16025195e-01 5.18138111e-01 -1.05861032e+00 2.18378767e-01
1.14808656e-01 -6.77171722e-02 1.12768555e+00 -9.44867849e-01
1.00467944e+00 1.03856242e+00 3.03772658e-01 4.45504367e-01
-1.38733602e+00 -3.85268509e-01 3.37604016e-01 6.11223392e-02
-1.62113261e+00 -2.25223780e-01 5.46124339e-01 -1.49234638e-01
5.69004178e-01 6.93129778e-01 8.64661157e-01 4.69131805e-02
-1.16403773e-01 1.11105430e+00 9.16288078e-01 -5.25770307e-01
4.61722165e-01 2.91010946e-01 4.85601500e-02 7.83320844e-01
6.48072302e-01 -3.19369435e-01 -4.99850541e-01 -2.79715240e-01
7.89004803e-01 -8.43047276e-02 -1.32931501e-01 -8.02626610e-01
-1.11130285e+00 1.01760399e+00 5.24244666e-01 -9.96136889e-02
-3.33928436e-01 4.34448034e-01 7.45589333e-03 1.93853979e-03
2.62779772e-01 6.33181632e-01 -2.90953010e-01 3.61663759e-01
-1.46720254e+00 8.57579112e-01 7.13916540e-01 1.22737563e+00
8.61150265e-01 -2.54815787e-01 -2.45433241e-01 6.38871670e-01
2.29312971e-01 3.86843771e-01 -2.16528460e-01 -9.84995723e-01
5.52907705e-01 5.23706257e-01 3.18267316e-01 -8.70640755e-01
-4.16333109e-01 -3.54311466e-01 -5.31701565e-01 1.82370141e-01
4.22734380e-01 -6.43761232e-02 -9.49352801e-01 1.51342273e+00
8.91562939e-01 -4.06053543e-01 -4.29563671e-01 8.92258346e-01
5.63602209e-01 7.31397867e-01 -3.23795587e-01 -7.30131447e-01
1.23161721e+00 -1.09938514e+00 -5.45586228e-01 -2.74688601e-01
5.69366395e-01 -7.45095015e-01 5.48020780e-01 5.06137073e-01
-1.53978205e+00 -2.05040678e-01 -1.06743991e+00 -4.28887047e-02
-8.07447955e-02 -7.29101598e-02 9.20872509e-01 8.18718374e-01
-1.00149775e+00 1.92167372e-01 -3.93979967e-01 -2.84548640e-01
7.17080712e-01 5.97950459e-01 -1.92060784e-01 -2.75123477e-01
-4.41858530e-01 7.58382976e-01 6.84870720e-01 4.80583310e-02
-8.53866160e-01 -4.13634539e-01 -5.81465483e-01 -9.34671834e-02
9.97286797e-01 -5.49119890e-01 1.00631833e+00 -4.18434530e-01
-9.08122063e-01 9.82579708e-01 -4.39520448e-01 -6.74542248e-01
5.92969000e-01 6.80941716e-02 4.30228710e-01 2.15109289e-01
-3.63202766e-03 1.26740730e+00 9.77471828e-01 -1.35150611e+00
-8.56507778e-01 -4.70100760e-01 2.37724662e-01 2.42370129e-01
-1.76919892e-01 2.05009757e-03 -5.30932128e-01 -6.28630817e-01
9.06432211e-01 -9.89914000e-01 -8.14544320e-01 2.50012189e-01
-6.30812585e-01 -2.58929878e-01 5.60099483e-01 -4.50353086e-01
1.57617831e+00 -1.78485048e+00 1.98605776e-01 5.51410675e-01
2.51768768e-01 -1.53869554e-01 -6.04075752e-02 2.79066145e-01
2.67070651e-01 1.56401098e-01 -4.24656361e-01 -2.89226025e-01
-1.10055678e-01 2.08885849e-01 -3.24117988e-01 6.26727521e-01
-5.66394702e-02 7.36396611e-01 -6.58693612e-01 -8.79802942e-01
-1.49041265e-01 -2.48518780e-01 -8.33288074e-01 -2.28809774e-01
-4.93638158e-01 -7.27491230e-02 -4.76438284e-01 1.01279449e+00
1.28513181e+00 -2.46423274e-01 2.20791861e-01 4.19708788e-02
-2.58302778e-01 5.11628948e-02 -1.70381737e+00 1.57266569e+00
-1.64874736e-02 3.60176086e-01 2.26910681e-01 -1.07292914e+00
9.76361096e-01 -2.14710578e-01 6.75299525e-01 -2.05181360e-01
9.27369203e-03 4.65318769e-01 -1.82296917e-01 1.10706259e-02
7.44075477e-01 -1.83469713e-01 3.86168696e-02 4.49209124e-01
-1.65957808e-01 -4.90020901e-01 5.73492169e-01 2.58589208e-01
9.77298796e-01 -1.71570778e-01 7.71948695e-01 -3.51386517e-01
4.86443341e-01 4.82650310e-01 5.01527786e-01 1.12377489e+00
-1.41512007e-01 6.10399902e-01 6.63669586e-01 -3.85432571e-01
-1.05280685e+00 -1.03981507e+00 -3.81318390e-01 9.12150979e-01
5.14723837e-01 -3.22689950e-01 -8.14706624e-01 -7.67942369e-01
6.22764379e-02 3.89305443e-01 -5.43974936e-01 4.50997680e-01
-8.25459898e-01 -9.46528971e-01 -8.58518630e-02 3.51282626e-01
5.05881071e-01 -9.16326761e-01 -9.17850792e-01 2.30542719e-01
-4.59310710e-02 -9.33676541e-01 -6.65429950e-01 3.16525757e-01
-1.34924316e+00 -8.42439592e-01 -9.05611932e-01 -7.69647181e-01
1.08992898e+00 4.93449807e-01 1.04698169e+00 1.34193406e-01
-5.41278839e-01 4.15213108e-02 -2.87703633e-01 -3.73497039e-01
5.22991754e-02 8.24191272e-02 -2.33539581e-01 -3.27893794e-01
-1.43244201e-02 -2.39876330e-01 -7.81502604e-01 4.96401787e-01
-9.12749469e-01 -2.71266159e-02 6.51207089e-01 7.21501410e-01
1.18649483e+00 1.04028985e-01 1.95360467e-01 -5.66512346e-01
2.98342407e-01 -3.07318300e-01 -1.22304249e+00 2.22988367e-01
-4.95284319e-01 1.15915984e-02 2.68830098e-02 -3.00868213e-01
-7.91328192e-01 3.80439162e-01 2.19976559e-01 -1.98506698e-01
3.18177879e-01 2.89406449e-01 -8.06211159e-02 -4.40739840e-01
4.77095544e-01 3.90336275e-01 -3.09838742e-01 -3.16607058e-01
5.49279809e-01 3.44411433e-01 1.52639359e-01 -3.99606079e-01
9.03286278e-01 1.10011709e+00 4.24349099e-01 -7.86576331e-01
-8.33695531e-01 -6.27707958e-01 -6.32513762e-01 -1.26072794e-01
6.22618914e-01 -5.22701383e-01 -7.54140615e-01 -5.83391786e-02
-1.25160313e+00 -7.22125694e-02 -6.38322771e-01 1.38303518e-01
-6.65822148e-01 3.55137378e-01 1.92206219e-01 -1.15221322e+00
-2.13984773e-01 -1.14695001e+00 1.02420199e+00 1.75694317e-01
-7.23357797e-02 -3.31516236e-01 -3.06900907e-02 3.99675608e-01
-8.25181231e-02 2.20265031e-01 6.84232891e-01 -4.82751727e-01
-1.40435743e+00 -2.29409263e-01 -3.50189209e-01 1.18512079e-01
-4.06103969e-01 -3.48929584e-01 -3.53769213e-01 -4.10898089e-01
-4.99536581e-02 -1.85762629e-01 1.03456128e+00 7.74716735e-01
1.18585646e+00 -4.31725085e-01 -5.62032402e-01 6.94647074e-01
1.56154966e+00 3.77167732e-01 6.78349555e-01 5.88778198e-01
2.11387858e-01 5.10788023e-01 1.06714106e+00 8.43036294e-01
9.64476839e-02 8.23580205e-01 7.57041812e-01 1.25245333e-01
2.03885496e-01 1.70667898e-02 -3.09559014e-02 5.64687550e-02
-6.51617255e-03 -4.18393821e-01 -6.85311496e-01 8.90231431e-01
-1.66420257e+00 -8.65557790e-01 -1.20290279e-01 2.45407963e+00
5.93695045e-01 1.02693021e-01 3.89423966e-01 5.85525967e-02
7.19676614e-01 3.60691428e-01 -4.37747926e-01 -5.02559960e-01
-2.11889312e-01 2.08364785e-01 8.88372123e-01 4.85311031e-01
-1.06023061e+00 9.37340319e-01 7.26530886e+00 1.16050160e+00
-7.70091891e-01 8.92450213e-02 7.73343086e-01 -5.76583266e-01
-4.42722589e-01 1.00165024e-01 -1.45327389e+00 1.13996223e-01
8.04670677e-02 -9.26488116e-02 4.29372877e-01 1.16434503e+00
-3.88456732e-02 -7.03916609e-01 -8.65535736e-01 9.52075839e-01
1.34353384e-01 -1.57366824e+00 4.11527790e-03 3.80428880e-01
1.17365956e+00 -3.60197395e-01 1.00314185e-01 -1.94476053e-01
1.35430032e-02 -7.75708735e-01 9.20301497e-01 -2.92982697e-01
5.66893518e-01 -7.40495503e-01 3.14067751e-01 4.57547784e-01
-1.19462681e+00 -2.07964376e-01 -6.06730223e-01 3.50378007e-01
5.34554660e-01 7.24133670e-01 -1.05268002e+00 1.15836039e-01
6.09791160e-01 -1.06886178e-01 -2.33816668e-01 1.34499669e+00
-1.23639822e-01 1.60921529e-01 -8.30808997e-01 -2.35169813e-01
4.05371517e-01 -2.01296613e-01 9.11546350e-01 1.02037334e+00
3.34946781e-01 2.50854343e-01 2.16002777e-01 9.34108078e-01
-2.38942116e-01 2.48867124e-01 -2.75192052e-01 1.02917843e-01
5.20423889e-01 1.11979115e+00 -1.28443015e+00 -3.74539167e-01
-2.09010944e-01 8.38764608e-01 2.66461641e-01 1.16518103e-01
-9.62769151e-01 -2.08368108e-01 2.93788731e-01 4.48784590e-01
8.53143632e-01 -3.60068768e-01 -7.37782598e-01 -7.02004731e-01
3.06108862e-01 -5.92646182e-01 4.84072894e-01 -3.90674949e-01
-6.98189795e-01 2.78363794e-01 3.53165627e-01 -1.13140988e+00
1.33561075e-01 -3.62687677e-01 -3.95800948e-01 6.91546798e-01
-1.41407406e+00 -8.59280288e-01 -2.32466191e-01 3.72840762e-01
7.40369856e-01 2.81507820e-01 1.51468456e-01 1.58922821e-02
-1.90835997e-01 4.66761321e-01 3.73950675e-02 -4.02977914e-01
9.78711993e-02 -1.15937543e+00 1.63716525e-01 1.01778567e+00
8.26017261e-02 5.58222830e-01 8.11235249e-01 -7.12930501e-01
-1.29030049e+00 -7.88088858e-01 8.53360355e-01 9.85912532e-02
3.42173815e-01 -2.17125863e-01 -5.50580204e-01 5.83469272e-01
-1.02746271e-01 1.08744696e-01 4.00610596e-01 -1.33816123e-01
9.40403715e-02 -1.20326281e-01 -1.40863633e+00 6.77000999e-01
1.37534332e+00 3.05871725e-01 -3.19752455e-01 6.16178751e-01
4.55752045e-01 -5.96438050e-01 -4.62994963e-01 4.48896259e-01
5.23322940e-01 -9.67939913e-01 1.11646378e+00 -4.26563621e-01
3.15557718e-01 -3.07677776e-01 -3.03194314e-01 -5.96054018e-01
-9.02187750e-02 -9.93625283e-01 -3.58519047e-01 5.83107710e-01
4.06854331e-01 -4.35166866e-01 1.31418538e+00 3.32843691e-01
2.74425894e-02 -1.34736323e+00 -1.29928470e+00 -9.53411281e-01
-3.09656173e-01 -4.21207041e-01 4.86627728e-01 5.31397521e-01
-1.43807858e-01 -4.14814502e-01 -2.98818856e-01 1.73553690e-01
8.33584845e-01 5.44513464e-01 8.86460602e-01 -9.29142475e-01
-3.41697007e-01 -3.55866194e-01 -3.14971209e-01 -1.70573783e+00
-4.22739208e-01 -7.04487503e-01 1.05192922e-01 -1.52039981e+00
5.33514559e-01 -7.51344204e-01 4.49664295e-02 1.96156368e-01
-8.94443411e-03 3.08100790e-01 4.59159970e-01 1.36113986e-01
-6.55279040e-01 2.43363380e-01 1.53247678e+00 -6.06755354e-02
-4.81144249e-01 1.66276455e-01 -6.99216425e-01 7.71526873e-01
3.51076990e-01 -7.16190755e-01 -1.44758970e-01 -2.84378856e-01
3.54387343e-01 1.75017253e-01 1.21994793e-01 -6.36018574e-01
9.00090784e-02 -5.53869605e-01 4.65028249e-02 -1.30073202e+00
4.88519490e-01 -6.40585721e-01 -4.77676652e-02 6.55245960e-01
-1.50606334e-01 -2.72705346e-01 7.16920272e-02 4.25476670e-01
-1.76613897e-01 -7.54466236e-01 1.09800684e+00 -3.58999550e-01
-7.26828456e-01 4.53735828e-01 -9.59895030e-02 -8.88060257e-02
1.47752500e+00 -8.32782924e-01 -1.41509280e-01 -2.23147959e-01
-8.65376174e-01 4.16393250e-01 4.19776767e-01 -1.58715323e-01
7.06156194e-01 -1.13964272e+00 -5.62274754e-01 -2.96623781e-02
-1.08668916e-01 3.45066041e-01 3.54713410e-01 1.05834031e+00
-7.58090138e-01 5.80984831e-01 1.05230689e-01 -6.92764997e-01
-1.52591896e+00 5.54875791e-01 1.08092934e-01 -5.46194077e-01
-3.44180226e-01 1.49357724e+00 3.24755132e-01 8.66519883e-02
1.71762854e-01 -1.03171863e-01 1.41476274e-01 2.73026198e-01
3.16104293e-01 8.10083866e-01 -1.97823063e-01 -3.52854043e-01
-4.59442079e-01 5.10192275e-01 -2.43976608e-01 -3.76767218e-01
1.16352189e+00 -7.82682076e-02 -3.30251873e-01 -2.28955835e-01
8.86198103e-01 1.33123606e-01 -9.75822031e-01 -4.99007665e-02
-1.25264332e-01 -8.90411615e-01 9.83764604e-02 -3.33515197e-01
-8.93886805e-01 4.07458901e-01 4.41514730e-01 3.52232128e-01
1.25678205e+00 4.46019292e-01 8.05348456e-01 6.22294307e-01
5.20358443e-01 -1.19537354e+00 -1.09087871e-02 2.01414049e-01
1.11552942e+00 -1.15899909e+00 6.23425484e-01 -9.27811623e-01
-3.06515247e-01 9.93846238e-01 6.27393246e-01 -1.35979965e-01
5.42243361e-01 1.36914715e-01 -4.88758802e-01 -1.67226449e-01
-4.80244309e-01 -3.59801143e-01 2.72228003e-01 2.00985834e-01
-7.65919778e-03 -1.35171548e-01 -8.63630474e-01 1.75315708e-01
-1.73910901e-01 -1.03358619e-01 4.35671568e-01 1.18314159e+00
-1.07319057e+00 -9.45141375e-01 -9.09018457e-01 5.59681594e-01
-1.68633744e-01 -8.34288448e-02 -2.65234470e-01 9.33290839e-01
3.86283994e-01 5.50096571e-01 1.21305719e-01 1.39272884e-01
1.34441733e-01 -2.82975018e-01 8.54897738e-01 -7.30399609e-01
-1.68343619e-01 3.01764339e-01 -1.89299911e-01 -6.33431911e-01
-2.82712549e-01 -8.05343688e-01 -1.10434031e+00 -1.18461065e-01
-7.99243391e-01 -4.89209928e-02 6.94895446e-01 7.63576746e-01
-6.90552816e-02 8.82556662e-02 5.43298900e-01 -8.08394611e-01
-6.74135804e-01 -6.97571397e-01 -8.58879805e-01 -3.15372236e-02
7.41263255e-02 -6.01813078e-01 -2.47095779e-01 -1.10483982e-01] | [9.471566200256348, 0.22436439990997314] |
3297a909-c700-44e1-970b-d1a0ae100821 | rand-robustness-aware-norm-decay-for | 2305.15536 | null | https://arxiv.org/abs/2305.15536v1 | https://arxiv.org/pdf/2305.15536v1.pdf | RAND: Robustness Aware Norm Decay For Quantized Seq2seq Models | With the rapid increase in the size of neural networks, model compression has become an important area of research. Quantization is an effective technique at decreasing the model size, memory access, and compute load of large models. Despite recent advances in quantization aware training (QAT) technique, most papers present evaluations that are focused on computer vision tasks, which have different training dynamics compared to sequence tasks. In this paper, we first benchmark the impact of popular techniques such as straight through estimator, pseudo-quantization noise, learnable scale parameter, clipping, etc. on 4-bit seq2seq models across a suite of speech recognition datasets ranging from 1,000 hours to 1 million hours, as well as one machine translation dataset to illustrate its applicability outside of speech. Through the experiments, we report that noise based QAT suffers when there is insufficient regularization signal flowing back to the quantization scale. We propose low complexity changes to the QAT process to improve model accuracy (outperforming popular learnable scale and clipping methods). With the improved accuracy, it opens up the possibility to exploit some of the other benefits of noise based QAT: 1) training a single model that performs well in mixed precision mode and 2) improved generalization on long form speech recognition. | ['Yanzhang He', 'Oleg Rybakov', 'Shaojin Ding', 'David Rim', 'David Qiu'] | 2023-05-24 | null | null | null | null | ['model-compression'] | ['methodology'] | [ 5.02883673e-01 -1.72203153e-01 -2.27618337e-01 -4.07036513e-01
-9.50530171e-01 -3.03678662e-01 5.34480155e-01 -6.23191334e-03
-7.52660155e-01 6.86343133e-01 1.94571558e-02 -5.23630381e-01
-1.65886372e-01 -4.34280336e-01 -8.22615504e-01 -6.60932660e-01
-1.50819078e-01 4.53720123e-01 3.03201586e-01 -8.42428133e-02
3.15582991e-01 3.69169384e-01 -1.48303032e+00 3.35324943e-01
6.79874420e-01 1.17251194e+00 2.88498342e-01 1.00678718e+00
-9.81741250e-02 7.47872353e-01 -8.10706079e-01 -4.18316543e-01
4.57223296e-01 -2.27659419e-01 -5.90381920e-01 -2.79546350e-01
6.33095264e-01 -3.19544405e-01 -3.61732841e-01 1.20647943e+00
8.70363414e-01 1.88239992e-01 3.55580986e-01 -1.02541053e+00
-3.38029116e-01 8.25880647e-01 -3.03664058e-01 5.92399120e-01
-2.45787397e-01 2.29306936e-01 6.23600006e-01 -7.64845729e-01
3.98472399e-01 1.53195381e+00 9.39316630e-01 5.21080971e-01
-9.57802057e-01 -7.22240567e-01 -2.45159477e-01 4.57582384e-01
-1.40237427e+00 -7.92355180e-01 3.66997659e-01 2.83914357e-02
1.57588613e+00 3.21797848e-01 4.43368584e-01 1.01178300e+00
2.59366065e-01 5.24820805e-01 1.02055800e+00 -6.60314083e-01
6.33022845e-01 -6.82012225e-03 1.22478217e-01 5.14500797e-01
-6.62517920e-03 1.61336869e-01 -8.74448061e-01 -2.31814086e-01
6.66270852e-01 -4.07207012e-01 -1.15590408e-01 1.10405035e-01
-9.57451344e-01 7.94897735e-01 1.47759423e-01 1.24525778e-01
-2.34617397e-01 4.95897382e-01 8.10267508e-01 4.66796041e-01
4.84151632e-01 3.35721940e-01 -7.69982696e-01 -8.44651341e-01
-1.25122058e+00 1.27890453e-01 7.19317853e-01 9.19758856e-01
4.33250844e-01 4.78664845e-01 -2.26789236e-01 9.89880979e-01
1.05788939e-01 5.49243927e-01 1.11081100e+00 -9.05368745e-01
8.22836041e-01 1.28123611e-01 -2.98180521e-01 -6.95102274e-01
-2.72539794e-01 -6.59278750e-01 -1.16873360e+00 -1.50943398e-01
5.54823801e-02 -2.65754312e-01 -1.18296444e+00 1.60197031e+00
1.49584934e-01 4.29313421e-01 1.06720045e-01 6.64414227e-01
4.47404861e-01 9.46889937e-01 7.81852100e-03 -3.93040806e-01
1.15170264e+00 -1.18620801e+00 -9.77795482e-01 -2.60848999e-01
9.11913097e-01 -8.12635899e-01 1.16393244e+00 7.00730681e-01
-1.06691301e+00 -6.22158468e-01 -1.15667164e+00 -1.49346784e-01
-5.09765148e-01 -5.21187559e-02 4.42127883e-01 9.76338506e-01
-1.18984020e+00 8.53224277e-01 -8.97099674e-01 -1.98018536e-01
4.40094233e-01 5.27103305e-01 -6.16140924e-02 2.64181686e-03
-1.40332150e+00 9.86384809e-01 5.88307798e-01 -1.14952251e-01
-5.57827830e-01 -8.41519713e-01 -4.67404902e-01 3.71038258e-01
1.32528216e-01 -4.18788821e-01 1.42065036e+00 -1.11035371e+00
-1.72745812e+00 1.78229645e-01 -1.46812171e-01 -1.21134531e+00
3.15442771e-01 -2.85972357e-01 -2.89671421e-01 6.71258150e-03
-4.80836183e-01 8.52600098e-01 1.09106600e+00 -4.40075457e-01
-4.77197856e-01 -2.08161533e-01 -4.42273647e-01 1.42014921e-01
-7.42890120e-01 -5.67733906e-02 -3.64133269e-01 -9.07627761e-01
-1.00427888e-01 -8.58898044e-01 -2.05012575e-01 -1.92715138e-01
-5.41702332e-03 -1.41715571e-01 9.38282311e-01 -7.21181810e-01
1.56495810e+00 -2.20222282e+00 1.59134511e-02 8.49914178e-02
-3.31681222e-01 8.16723287e-01 -1.75685108e-01 3.43318820e-01
-2.43818667e-02 3.93460363e-01 -4.67852831e-01 -5.59630036e-01
-4.78579253e-02 5.08525789e-01 -2.77880579e-01 1.10887676e-01
1.85920253e-01 7.42299497e-01 -4.43390697e-01 -4.53910828e-01
8.60768706e-02 5.89123249e-01 -6.55959606e-01 -1.34888664e-01
-2.30421260e-01 -3.91860902e-02 2.13504490e-02 3.30734730e-01
7.11595833e-01 -1.00099146e-01 -1.01280712e-01 -1.67955533e-02
7.71318451e-02 4.68246430e-01 -1.19379318e+00 1.66761851e+00
-4.62352633e-01 7.36795068e-01 -2.90473122e-02 -1.02302599e+00
7.67172337e-01 4.39475894e-01 6.43986557e-03 -8.39976251e-01
-4.46431600e-02 2.30514765e-01 1.91928912e-02 -4.39531565e-01
4.08102721e-01 -1.57116607e-01 3.87865156e-01 5.30269817e-02
1.74338013e-01 -3.34075123e-01 1.44939721e-01 -7.51450211e-02
1.06256735e+00 -2.75469720e-01 2.32825801e-01 -2.03207999e-01
3.02538723e-01 -1.90505192e-01 3.87298465e-01 6.85589373e-01
-3.61527532e-01 4.51783776e-01 3.94553959e-01 -3.74196202e-01
-1.33546424e+00 -6.63347960e-01 -2.20612481e-01 1.19774032e+00
-4.89995748e-01 -6.38857007e-01 -1.02032399e+00 -1.26011059e-01
-1.87490478e-01 6.06688142e-01 -2.21180409e-01 -2.43496701e-01
-8.19385648e-01 -9.37381446e-01 9.21970785e-01 4.21054453e-01
6.56379223e-01 -7.57721424e-01 -6.26270115e-01 2.08172217e-01
3.96274701e-02 -1.19890320e+00 -4.81350482e-01 7.07077205e-01
-1.33606040e+00 -3.14917356e-01 -6.52358890e-01 -6.69533789e-01
1.05664395e-01 -7.36322105e-02 9.77234602e-01 -4.61230054e-02
-8.44641402e-02 -1.05498724e-01 -4.07809615e-01 -6.46370113e-01
-6.22970819e-01 4.43779290e-01 3.30470532e-01 -3.11672330e-01
1.96894959e-01 -5.80096066e-01 -3.82173777e-01 -1.51338115e-01
-9.72931683e-01 1.11538172e-03 7.14165449e-01 9.84552085e-01
6.41355217e-01 1.23896785e-01 7.17863381e-01 -7.12315857e-01
7.52531350e-01 -2.53001481e-01 -6.53522551e-01 1.71377361e-01
-8.91202748e-01 3.49342287e-01 7.73217082e-01 -6.35914028e-01
-7.23763704e-01 4.17881906e-02 -3.57941657e-01 -6.82491779e-01
1.73924312e-01 4.39503610e-01 2.60771602e-01 -5.44065759e-02
7.74716675e-01 2.76833296e-01 7.77043626e-02 -5.00251889e-01
3.74739617e-01 9.24091160e-01 2.28865162e-01 -9.47437361e-02
5.44180632e-01 -9.86323282e-02 7.72496909e-02 -1.03559601e+00
-5.57983279e-01 -3.80761743e-01 -5.16700327e-01 2.27891713e-01
6.56923711e-01 -8.52859437e-01 -3.26319069e-01 3.83738488e-01
-1.13428187e+00 -4.99884009e-01 -2.28520110e-01 5.73172092e-01
-4.33110744e-01 4.76991981e-01 -7.84181535e-01 -8.49828899e-01
-7.95693636e-01 -1.22815621e+00 1.05267227e+00 2.85175852e-02
2.40483275e-03 -7.47632563e-01 -1.64620087e-01 2.98182726e-01
8.58854771e-01 -3.91266316e-01 1.00254118e+00 -6.04640305e-01
-4.67374921e-01 1.80707406e-02 -4.98938821e-02 7.11430848e-01
-9.07633230e-02 -6.37932271e-02 -1.18659365e+00 -4.63634610e-01
2.39369854e-01 -3.73580396e-01 8.77397239e-01 4.14198041e-01
1.30993724e+00 -5.80549479e-01 4.87329774e-02 7.15608776e-01
1.19850945e+00 3.53063554e-01 5.19890070e-01 2.11976588e-01
3.95185411e-01 1.53483495e-01 4.18020666e-01 4.11452949e-01
-2.41349176e-01 7.75128901e-01 2.81760365e-01 1.67165130e-01
-3.05017710e-01 -6.87112659e-02 5.40692747e-01 1.41814172e+00
2.38999575e-01 -2.11951301e-01 -9.75697339e-01 1.54419795e-01
-1.54395521e+00 -8.56250644e-01 1.84269771e-01 2.27074313e+00
1.06317246e+00 4.65650260e-01 -1.13925904e-01 4.35301334e-01
5.30018866e-01 4.27470170e-02 -5.31837761e-01 -9.31119859e-01
-5.30906245e-02 5.25651574e-01 8.03318501e-01 5.66378117e-01
-1.09896421e+00 9.45861578e-01 6.78634930e+00 1.44351649e+00
-1.52527320e+00 3.51538450e-01 8.68313849e-01 -2.66489118e-01
1.88844532e-01 -4.11357820e-01 -8.80295694e-01 4.68521416e-01
1.84126592e+00 2.73052622e-02 7.08808661e-01 1.03015232e+00
3.71855617e-01 1.82431176e-01 -9.91258681e-01 1.23175943e+00
2.64565903e-03 -1.28225923e+00 2.66475350e-01 -1.36953965e-01
6.75659537e-01 4.26456094e-01 2.72542834e-01 5.17028511e-01
-1.74861863e-01 -1.25141597e+00 6.67738676e-01 2.25713149e-01
9.45443511e-01 -8.37106407e-01 9.06086743e-01 6.10097051e-01
-9.06617880e-01 -2.24945396e-01 -7.48543859e-01 -2.12049276e-01
-2.19053924e-02 6.01067305e-01 -1.29381764e+00 2.21702367e-01
6.33659720e-01 3.00444722e-01 -6.50150597e-01 9.28171992e-01
2.64428645e-01 1.07241070e+00 -5.58673918e-01 -2.17450187e-01
4.65198070e-01 -2.71561854e-02 2.71214753e-01 1.45692647e+00
4.46978629e-01 1.96432248e-02 -3.10877532e-01 2.71988332e-01
-1.14236265e-01 1.40239418e-01 -4.78167176e-01 2.19691321e-02
7.12450802e-01 7.80415833e-01 -5.08411586e-01 -5.94424665e-01
-2.64492065e-01 8.30337644e-01 2.84988523e-01 1.67699113e-01
-8.18413675e-01 -2.45444134e-01 4.40485328e-01 -7.47095719e-02
5.59564769e-01 -3.43679488e-01 -5.27540863e-01 -7.82093704e-01
1.42791972e-01 -1.23917019e+00 -7.70442421e-03 -6.37196004e-01
-7.62705088e-01 6.17546916e-01 -9.02630612e-02 -1.06323481e+00
-2.68595248e-01 -5.09816527e-01 -1.46460012e-01 8.29249561e-01
-1.46927941e+00 -5.12118638e-01 1.15423545e-01 1.50236756e-01
1.07080984e+00 -2.22910166e-01 8.97444010e-01 5.38909554e-01
-5.19686282e-01 9.15705383e-01 4.51744080e-01 -2.16619104e-01
6.30324960e-01 -9.53091562e-01 7.94497788e-01 7.91583121e-01
3.34924489e-01 5.95691442e-01 6.18187785e-01 -4.15809333e-01
-1.46915984e+00 -1.20838320e+00 8.56896698e-01 -2.07073063e-01
3.96778286e-01 -4.44342554e-01 -9.92369294e-01 3.96743745e-01
1.55667469e-01 1.25384415e-02 3.61194849e-01 -2.42088422e-01
-1.92844346e-01 -3.80959749e-01 -1.05913591e+00 4.31845188e-01
7.82275856e-01 -4.78234053e-01 -3.99454057e-01 3.94353420e-01
1.08886099e+00 -5.41928172e-01 -9.33991432e-01 3.94543469e-01
4.18988615e-01 -8.17004383e-01 9.24822986e-01 -4.79187071e-01
1.92170128e-01 2.02507421e-01 -3.67299438e-01 -1.24600196e+00
-3.24468017e-02 -8.02854657e-01 -2.46502146e-01 9.90875065e-01
5.15442550e-01 -4.96080369e-01 8.75823617e-01 3.66945177e-01
-3.11848849e-01 -1.04058266e+00 -1.53839266e+00 -8.35615695e-01
2.23007426e-01 -4.81432825e-01 4.88742232e-01 6.46031797e-01
-2.39694014e-01 4.16140795e-01 -4.97622669e-01 -1.13463767e-01
3.05370957e-01 -6.19460762e-01 4.52451944e-01 -8.69664490e-01
-4.81250048e-01 -3.61854672e-01 -4.26646590e-01 -1.25791645e+00
-2.09516689e-01 -7.31955588e-01 -2.12062467e-02 -9.67605412e-01
-1.30378306e-01 -3.95867974e-01 -1.71770707e-01 3.15670848e-01
-2.44152956e-02 4.10909578e-02 3.63939375e-01 1.79154828e-01
-4.03635561e-01 6.22744262e-01 9.93162692e-01 -1.02548271e-01
-1.36234080e-02 -6.27102256e-02 2.70709172e-02 5.95049500e-01
9.73299980e-01 -6.38280451e-01 -6.44882083e-01 -7.13969529e-01
3.64145711e-02 1.72489285e-01 -1.99300945e-02 -1.31726468e+00
3.92635733e-01 1.58664182e-01 1.60707399e-01 -5.15576899e-01
6.40878439e-01 -6.67922735e-01 -6.66298866e-02 8.05407941e-01
-4.56444472e-01 4.16456014e-01 3.57533664e-01 4.89053339e-01
-3.12889814e-01 -5.16449928e-01 9.59764659e-01 -5.85700385e-02
-5.92130125e-01 1.25009939e-01 -3.52942169e-01 -3.73489857e-02
6.07856095e-01 -1.87750697e-01 -1.77462086e-01 -2.14355871e-01
-5.26044011e-01 -1.68700173e-01 -8.52433033e-03 3.61236751e-01
4.73928660e-01 -1.09215200e+00 -5.66452444e-01 4.53185976e-01
-3.77531379e-01 5.98441288e-02 1.56846605e-02 6.52345121e-01
-6.60786450e-01 8.61970186e-01 -1.23408949e-02 -7.14411080e-01
-1.41978383e+00 5.17080188e-01 3.71472239e-01 -2.68469363e-01
-4.25061256e-01 9.71748471e-01 -3.97503793e-01 -2.85095125e-01
7.90735185e-01 -9.01000738e-01 6.66936859e-02 -2.18735114e-01
5.58542848e-01 5.51113904e-01 6.49268687e-01 -3.25039715e-01
-2.06310377e-01 4.69724894e-01 5.36190718e-03 -1.88677683e-01
1.16146803e+00 7.77347432e-03 9.45170298e-02 6.52457237e-01
1.23050320e+00 -5.14480531e-01 -1.12204003e+00 -1.31450683e-01
1.66714042e-01 -1.95432782e-01 2.57879078e-01 -7.30317116e-01
-8.39886904e-01 1.20231605e+00 1.13884580e+00 2.89947033e-01
1.20824516e+00 -5.35764277e-01 8.98021042e-01 6.89252555e-01
3.40096235e-01 -1.25967610e+00 3.52944806e-02 8.82326543e-01
7.39284456e-01 -1.17273211e+00 -1.27536148e-01 -2.75371313e-01
-4.05593932e-01 1.13489449e+00 3.02296847e-01 1.71989858e-01
6.70443416e-01 6.17935896e-01 3.19721960e-02 2.83979386e-01
-1.24848402e+00 3.35694194e-01 -5.44891804e-02 4.76724744e-01
4.14196432e-01 -5.35155386e-02 -2.55008876e-01 4.33558255e-01
-4.70768690e-01 1.47020578e-01 3.33852828e-01 7.78544247e-01
-6.64915025e-01 -1.02140057e+00 -4.30564970e-01 6.90875709e-01
-5.69369376e-01 -5.48085153e-01 -6.05949685e-02 5.02261043e-01
7.08922669e-02 8.69963408e-01 1.00978449e-01 -3.50841582e-01
1.57254457e-01 4.53699976e-01 2.25063562e-01 -3.87251675e-01
-7.19843030e-01 -1.11870002e-02 5.21502085e-02 -4.78902787e-01
-2.21954077e-01 -4.43578035e-01 -1.13028586e+00 -2.54798025e-01
-5.06199241e-01 1.05289347e-01 1.02317238e+00 9.02485669e-01
6.16022766e-01 5.72921395e-01 2.45943129e-01 -6.78511143e-01
-1.30505335e+00 -1.40563524e+00 -2.86352068e-01 1.62538975e-01
4.24460232e-01 -4.03349698e-01 -4.30731356e-01 1.86125755e-01] | [14.040766716003418, 6.425772666931152] |
d7647134-512e-417e-981b-9d609a68f6c9 | are-3d-face-shapes-expressive-enough-for | 2207.01113 | null | https://arxiv.org/abs/2207.01113v2 | https://arxiv.org/pdf/2207.01113v2.pdf | Are 3D Face Shapes Expressive Enough for Recognising Continuous Emotions and Action Unit Intensities? | Recognising continuous emotions and action unit (AU) intensities from face videos requires a spatial and temporal understanding of expression dynamics. Existing works primarily rely on 2D face appearances to extract such dynamics. This work focuses on a promising alternative based on parametric 3D face shape alignment models, which disentangle different factors of variation, including expression-induced shape variations. We aim to understand how expressive 3D face shapes are in estimating valence-arousal and AU intensities compared to the state-of-the-art 2D appearance-based models. We benchmark four recent 3D face alignment models: ExpNet, 3DDFA-V2, DECA, and EMOCA. In valence-arousal estimation, expression features of 3D face models consistently surpassed previous works and yielded an average concordance correlation of .739 and .574 on SEWA and AVEC 2019 CES corpora, respectively. We also study how 3D face shapes performed on AU intensity estimation on BP4D and DISFA datasets, and report that 3D face features were on par with 2D appearance features in AUs 4, 6, 10, 12, and 25, but not the entire set of AUs. To understand this discrepancy, we conduct a correspondence analysis between valence-arousal and AUs, which points out that accurate prediction of valence-arousal may require the knowledge of only a few AUs. | ['Michel Valstar', 'Timo Giesbrecht', 'Elisabeth André', 'Björn W. Schuller', 'Ömer Sümer', 'Mani Kumar Tellamekala'] | 2022-07-03 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [-3.06586266e-01 -7.09972084e-02 7.74323493e-02 -7.63849676e-01
-1.41985908e-01 -6.56681001e-01 6.85510874e-01 -4.59577054e-01
-1.52016506e-01 3.45551163e-01 -1.05063496e-02 4.72599357e-01
8.99743568e-03 -2.87837386e-01 -2.34752968e-01 -7.56574392e-01
-3.23768914e-01 2.50557035e-01 -7.40778625e-01 -4.97208685e-01
-2.64312744e-01 8.93105328e-01 -1.79139650e+00 3.11842203e-01
1.95481122e-01 1.55120707e+00 -8.68906736e-01 5.85916579e-01
-1.60325333e-01 2.87714332e-01 -5.15802145e-01 -9.24440026e-01
3.78638923e-01 -6.41107321e-01 -3.54871511e-01 1.53654352e-01
8.07077885e-01 -2.39517659e-01 -2.00320985e-02 7.99462795e-01
5.79682410e-01 -1.26370326e-01 1.00997949e+00 -1.60639024e+00
-6.95672452e-01 -1.64761111e-01 -9.88050759e-01 2.09596574e-01
4.79667246e-01 9.50072333e-02 7.21982121e-01 -1.04050851e+00
6.72853529e-01 1.55354595e+00 7.59383917e-01 9.53491867e-01
-1.28275096e+00 -8.08485091e-01 -1.41394019e-01 7.63978064e-02
-1.48696315e+00 -8.15258384e-01 1.10781801e+00 -5.10750592e-01
9.68177795e-01 4.43165541e-01 8.91076386e-01 1.49154365e+00
-4.40332517e-02 4.60283220e-01 1.70649862e+00 -2.93708116e-01
1.31803297e-03 2.24841490e-01 5.36804758e-02 6.23126090e-01
-3.13663721e-01 1.19167455e-01 -7.24668086e-01 -3.83691549e-01
6.67538047e-01 -5.72729409e-01 -2.05654874e-02 4.55127144e-03
-3.56712580e-01 6.99713230e-01 -3.15208882e-02 2.97211975e-01
-5.37781775e-01 -1.60945728e-01 6.48844481e-01 5.08738160e-01
1.01990104e+00 1.51426122e-01 -6.40599132e-01 -5.67191899e-01
-7.62611389e-01 6.73015863e-02 6.01964295e-01 5.71616113e-01
8.78496826e-01 2.97003388e-01 -1.87334821e-01 1.14225292e+00
2.85414934e-01 7.02362835e-01 -2.25818548e-02 -9.74695444e-01
-3.29108000e-01 5.49872994e-01 -6.25244454e-02 -1.10386193e+00
-5.73229074e-01 5.90378121e-02 -7.27876961e-01 4.26552027e-01
4.21080619e-01 -2.40807936e-01 -6.99244201e-01 2.03166556e+00
4.41822350e-01 1.74906448e-01 -1.81964964e-01 1.09470809e+00
1.05589628e+00 2.50493616e-01 1.97208434e-01 -6.33370638e-01
1.37752569e+00 -3.35445046e-01 -9.77032721e-01 1.16101176e-01
2.17248306e-01 -7.43385077e-01 8.33614230e-01 4.36678916e-01
-1.36383343e+00 -7.79980063e-01 -7.23422527e-01 1.72669396e-01
-2.75534958e-01 3.07791412e-01 7.00113118e-01 1.01324642e+00
-1.31346643e+00 5.02850115e-01 -4.74900663e-01 -3.51981848e-01
4.92402494e-01 5.32916546e-01 -7.44641721e-01 5.45492351e-01
-1.03196621e+00 1.09575415e+00 -2.98600078e-01 4.16727781e-01
-5.28752148e-01 -7.71790147e-01 -8.39345038e-01 -2.66165346e-01
-3.58940810e-01 -2.67715901e-01 9.54895139e-01 -1.77904367e+00
-1.89339733e+00 1.55320203e+00 -3.00897986e-01 9.00010094e-02
2.87208736e-01 -4.39976603e-02 -7.08277762e-01 1.29640058e-01
-6.59815133e-01 8.05141628e-01 9.75703716e-01 -1.22870171e+00
1.74266338e-01 -8.70708764e-01 -1.35114074e-01 8.39792043e-02
-4.72623199e-01 4.38625604e-01 -2.42221102e-01 -1.66252047e-01
-3.96875829e-01 -8.89010549e-01 4.02658343e-01 2.71913856e-01
1.97255373e-01 -3.68091851e-01 8.58849049e-01 -6.16035640e-01
9.95600402e-01 -2.25118423e+00 2.30829686e-01 3.21160495e-01
2.26538673e-01 2.53991157e-01 -4.60893363e-01 -9.73419938e-03
-3.88844967e-01 6.67086691e-02 1.19852476e-01 -6.29783988e-01
2.28993997e-01 1.63547993e-01 -4.26212884e-03 8.06621552e-01
4.94160444e-01 9.13329303e-01 -5.08516729e-01 -4.45113301e-01
1.43167898e-01 8.80799413e-01 -5.32824993e-01 4.03349906e-01
1.34856626e-01 4.31630939e-01 -6.04156405e-02 8.65440309e-01
1.04517841e+00 3.91597986e-01 2.38026425e-01 -4.47256356e-01
-4.15539779e-02 -4.66261894e-01 -7.57404745e-01 1.36537898e+00
-5.14663637e-01 8.87984097e-01 4.63668108e-01 -7.95405328e-01
1.35849214e+00 4.69539046e-01 7.42858529e-01 -9.44235265e-01
6.68076158e-01 2.34015603e-02 1.56214118e-01 -4.86042827e-01
3.03408876e-02 -3.18228573e-01 1.02814548e-01 1.22060038e-01
4.65123028e-01 -1.89142630e-01 -6.31816760e-02 -3.70136976e-01
5.03297925e-01 2.28415027e-01 1.57335773e-01 -3.97517741e-01
5.66516936e-01 -7.13714242e-01 4.70062584e-01 5.11579476e-02
-6.78953111e-01 4.62887645e-01 9.06514168e-01 -5.56995034e-01
-8.59855175e-01 -1.05416799e+00 -5.24642467e-01 1.07851028e+00
-1.92178145e-01 -4.29426491e-01 -1.01119053e+00 -6.19316399e-01
1.97437536e-02 4.47263032e-01 -1.36144960e+00 -7.99805447e-02
-1.33643568e-01 -8.50416958e-01 7.51568854e-01 4.12538707e-01
1.99995175e-01 -9.82559502e-01 -3.30166519e-01 -3.93541425e-01
9.03775841e-02 -1.16630113e+00 -1.56546578e-01 -1.74023136e-01
-3.88580501e-01 -1.02527916e+00 -3.95422548e-01 -1.75376400e-01
4.94949788e-01 -4.08696473e-01 1.37518227e+00 -5.66520840e-02
-3.87312204e-01 9.09635246e-01 -3.29444379e-01 -7.47349679e-01
-3.78066480e-01 -5.96438289e-01 5.91393054e-01 4.53935355e-01
9.51760173e-01 -8.10704708e-01 -4.43939179e-01 4.06353712e-01
-4.29469317e-01 -4.04898733e-01 3.08945656e-01 5.61679780e-01
4.69366014e-01 -7.44576812e-01 6.44333601e-01 -5.10578513e-01
6.55100048e-01 -4.02104169e-01 -2.37978607e-01 1.78875953e-01
-5.00673473e-01 -3.72494370e-01 3.93791944e-01 -5.18067062e-01
-1.21236777e+00 8.94254372e-02 -3.62185836e-01 -1.05524552e+00
-4.98132259e-01 -1.02671243e-01 -4.99297902e-02 -2.99266875e-01
7.61539280e-01 -1.76559001e-01 4.03570503e-01 -1.56254813e-01
1.95836976e-01 7.44645774e-01 2.23650858e-01 -6.17574513e-01
4.36986417e-01 4.69222903e-01 2.10823432e-01 -1.20218766e+00
-5.86534917e-01 -2.92167127e-01 -9.39194202e-01 -6.47118986e-01
8.53778124e-01 -9.37736452e-01 -1.11999774e+00 5.92708588e-01
-1.25609422e+00 -2.44522318e-01 -1.32088408e-01 1.67630032e-01
-6.07030749e-01 2.86535382e-01 -6.26822472e-01 -1.14805734e+00
-5.43343544e-01 -8.27899158e-01 1.35222876e+00 1.63377091e-01
-6.71282589e-01 -9.66637790e-01 3.32471758e-01 1.36036977e-01
3.58076036e-01 6.88261390e-01 4.62435603e-01 -3.25071394e-01
5.62433898e-01 -1.81355346e-02 -1.17187195e-01 5.60971916e-01
1.42016992e-01 6.36347234e-01 -1.53416586e+00 -5.47006018e-02
9.77307931e-02 -7.01904655e-01 3.46467167e-01 3.73416424e-01
1.14141190e+00 -7.48884082e-02 2.27965310e-01 6.41576588e-01
8.91715407e-01 8.34831148e-02 7.90264368e-01 -4.33683276e-01
2.58483887e-01 8.80239546e-01 6.48940444e-01 7.80123830e-01
-3.18669602e-02 9.48331416e-01 6.52525008e-01 -3.01684499e-01
8.02973285e-02 2.53278226e-01 4.31825876e-01 6.57409430e-01
-5.95093310e-01 1.93584457e-01 -4.61891711e-01 1.53895736e-01
-1.32620859e+00 -9.75423336e-01 -2.61798441e-01 1.89804578e+00
8.54349613e-01 -4.53345150e-01 4.87461299e-01 -9.37352180e-02
3.04004461e-01 2.10105568e-01 -3.26056182e-01 -1.16403890e+00
-2.92401820e-01 5.37245631e-01 -2.45739937e-01 3.10365587e-01
-7.89117038e-01 7.61796236e-01 6.66056919e+00 6.69034302e-01
-1.36122084e+00 1.12981014e-01 1.11656380e+00 -4.34611678e-01
-1.38663501e-01 -5.97170830e-01 -5.54757833e-01 2.84636140e-01
1.15190339e+00 2.03786030e-01 5.28342962e-01 9.33743775e-01
2.94653207e-01 1.79209486e-01 -1.24901772e+00 1.52495623e+00
4.39734191e-01 -7.07202554e-01 -2.99108535e-01 9.13500935e-02
6.77900553e-01 -2.77327597e-01 2.66234368e-01 3.95135164e-01
-4.20650512e-01 -1.34132385e+00 4.79158044e-01 6.93284154e-01
1.25653195e+00 -7.36359537e-01 5.62865615e-01 -2.08403349e-01
-1.02451479e+00 1.88694313e-01 -2.44949996e-01 -1.80990800e-01
-8.44983011e-02 2.29886606e-01 -5.72309196e-01 3.26357841e-01
9.06173706e-01 5.64247906e-01 -5.57183325e-01 1.43168867e-01
-3.04330103e-02 5.73535204e-01 -2.70871550e-01 -5.71283624e-02
-3.18812719e-03 -5.17065227e-01 2.85498321e-01 1.47074771e+00
2.28350699e-01 4.71792877e-01 -4.28462625e-01 9.92066503e-01
-4.36254628e-02 3.20656389e-01 -5.72312534e-01 -2.04539850e-01
-2.26324163e-02 1.93267357e+00 -4.01577175e-01 2.33311411e-02
-6.39821649e-01 1.12574172e+00 3.82148564e-01 1.39380887e-01
-1.16398287e+00 2.56152928e-01 1.32678533e+00 8.00180715e-03
-1.22775547e-01 -1.61750272e-01 -1.25766277e-01 -8.94890487e-01
-2.97353595e-01 -8.80321383e-01 1.75058439e-01 -9.52673197e-01
-1.51167381e+00 9.77493107e-01 7.92689845e-02 -8.50085199e-01
-3.33408356e-01 -1.02589858e+00 -6.67790592e-01 8.59842181e-01
-1.15430605e+00 -1.19254851e+00 -5.85323513e-01 6.97535574e-01
2.37147301e-01 9.71872360e-02 1.13371742e+00 2.77560472e-01
-5.63088238e-01 1.04059589e+00 -4.36131716e-01 -1.17759844e-02
8.68979514e-01 -9.69307482e-01 -6.05217405e-02 1.35567069e-01
2.96135753e-01 3.34552586e-01 6.43545866e-01 -2.63479110e-02
-1.50450027e+00 -6.66512132e-01 5.24955690e-01 -8.96672726e-01
3.25105369e-01 -6.42026246e-01 -7.66365051e-01 4.07081515e-01
4.45775926e-01 4.22999382e-01 9.72672760e-01 3.75746399e-01
-4.19542670e-01 -1.87518626e-01 -1.33593571e+00 4.75767821e-01
1.28974509e+00 -5.55902898e-01 -2.29717001e-01 -1.09503262e-01
-1.28255740e-01 -2.97747791e-01 -1.19950533e+00 6.90709174e-01
1.06106722e+00 -1.33440387e+00 8.20664406e-01 -6.76420927e-01
2.38646418e-01 2.31380582e-01 -1.63427770e-01 -1.29951012e+00
-2.42008358e-01 -5.22051930e-01 -1.70323119e-01 1.29518473e+00
8.86264816e-02 -3.85568619e-01 6.40816629e-01 7.05942631e-01
1.53013259e-01 -9.58767533e-01 -9.49993491e-01 -6.19423747e-01
1.10491570e-02 -4.29386139e-01 4.45919812e-01 1.13358986e+00
-3.18059586e-02 2.33656511e-01 -3.39180171e-01 -2.36026332e-01
2.10101321e-01 1.01935938e-02 8.50474298e-01 -1.43767953e+00
7.78241307e-02 -5.59912503e-01 -5.81319928e-01 -3.10198635e-01
9.10333216e-01 -6.96788371e-01 -3.99230212e-01 -6.90728664e-01
2.99693823e-01 -1.05343829e-03 -2.49495447e-01 6.73051596e-01
1.37755483e-01 6.94270194e-01 -1.92100778e-02 -3.02271098e-01
-2.21465290e-01 8.10485065e-01 1.15604901e+00 1.29733846e-01
-1.23200782e-01 -3.24756563e-01 -4.02223468e-01 8.28114569e-01
6.95661247e-01 1.66778073e-01 -4.56036389e-01 3.73323075e-02
-2.72463541e-02 -8.18712562e-02 3.61346006e-01 -5.07788181e-01
-4.32754844e-01 -1.72622398e-01 9.56793964e-01 -1.56691700e-01
9.65454340e-01 -8.71036232e-01 1.50387794e-01 -1.27683759e-01
5.41756079e-02 1.26245543e-01 7.07942247e-01 4.07590196e-02
-1.45637959e-01 1.25110716e-01 9.38044250e-01 7.47225657e-02
-6.19433880e-01 4.92747605e-01 -1.98210761e-01 -5.57570159e-02
1.01606727e+00 -4.17283475e-01 -2.16141157e-02 -5.89139342e-01
-8.22733343e-01 -2.80450702e-01 4.18324262e-01 6.69226468e-01
4.41118777e-01 -1.48719656e+00 -8.45672011e-01 4.70543146e-01
-4.46798233e-03 -7.78691411e-01 5.80327272e-01 1.24088573e+00
2.28619874e-02 4.09155600e-02 -6.58846617e-01 -7.34243810e-01
-1.85578978e+00 3.13724726e-01 6.31253123e-01 2.62796938e-01
3.21364731e-01 9.91288364e-01 3.17036420e-01 -3.76561195e-01
-1.53719157e-01 2.38426849e-01 -3.17910552e-01 5.42577147e-01
4.09870207e-01 2.91516453e-01 5.89525029e-02 -1.31288028e+00
-5.60071170e-01 9.24778461e-01 2.78368622e-01 8.49338621e-03
1.20789933e+00 2.05939426e-03 -3.66441965e-01 4.35129821e-01
1.50569296e+00 8.61242861e-02 -1.21170294e+00 3.22253108e-01
-4.37410146e-01 -5.12290239e-01 -1.18300952e-01 -7.71930516e-01
-1.22989345e+00 1.02292252e+00 9.14579332e-01 -2.21976568e-03
1.55356419e+00 2.13328861e-02 2.06025958e-01 -1.30959302e-01
6.18216880e-02 -1.08337295e+00 2.39836052e-01 2.77937412e-01
1.32229292e+00 -1.25481963e+00 -1.90311953e-01 -4.29018825e-01
-7.98042595e-01 1.32511902e+00 1.06190908e+00 1.78876311e-01
7.16949880e-01 3.61225963e-01 2.26293102e-01 -5.01392245e-01
-7.31259525e-01 -1.52386710e-01 7.30857432e-01 5.28388858e-01
7.88996935e-01 1.84746571e-02 -3.52360517e-01 8.71068180e-01
-3.43164355e-01 -1.73905969e-01 3.06593869e-02 3.83355826e-01
1.93985730e-01 -9.90501523e-01 -1.59223065e-01 1.66853011e-01
-6.73411250e-01 2.50734419e-01 -9.80133057e-01 1.05296361e+00
3.15647602e-01 7.68801510e-01 4.86024708e-01 -5.22692621e-01
4.67916340e-01 5.49129665e-01 1.04728770e+00 -1.83086082e-01
-6.48346543e-01 1.00729980e-01 2.64688373e-01 -6.24197662e-01
-8.18737984e-01 -6.96576416e-01 -9.07854080e-01 -3.39685231e-01
-1.96944937e-01 -5.75657636e-02 7.83202350e-01 6.32145464e-01
4.72341329e-01 -5.52450493e-02 9.70042288e-01 -1.02985322e+00
-2.18294829e-01 -1.25794768e+00 -7.85187423e-01 7.30691254e-01
7.13623464e-02 -9.03248727e-01 -6.28119767e-01 -4.10465896e-03] | [13.534074783325195, 1.8079111576080322] |
5c6da70f-41fe-4763-bdcb-234359b9e519 | reconstruction-and-membership-inference | 1906.03006 | null | https://arxiv.org/abs/1906.03006v1 | https://arxiv.org/pdf/1906.03006v1.pdf | Reconstruction and Membership Inference Attacks against Generative Models | We present two information leakage attacks that outperform previous work on membership inference against generative models. The first attack allows membership inference without assumptions on the type of the generative model. Contrary to previous evaluation metrics for generative models, like Kernel Density Estimation, it only considers samples of the model which are close to training data records. The second attack specifically targets Variational Autoencoders, achieving high membership inference accuracy. Furthermore, previous work mostly considers membership inference adversaries who perform single record membership inference. We argue for considering regulatory actors who perform set membership inference to identify the use of specific datasets for training. The attacks are evaluated on two generative model architectures, Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), trained on standard image datasets. Our results show that the two attacks yield success rates superior to previous work on most data sets while at the same time having only very mild assumptions. We envision the two attacks in combination with the membership inference attack type formalization as especially useful. For example, to enforce data privacy standards and automatically assessing model quality in machine learning as a service setups. In practice, our work motivates the use of GANs since they prove less vulnerable against information leakage attacks while producing detailed samples. | ['Martin Härterich', 'Daniel Bernau', 'Benjamin Hilprecht'] | 2019-06-07 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 3.30926210e-01 6.52217925e-01 -1.11734375e-01 -3.59586298e-01
-1.05569577e+00 -1.18630719e+00 9.43439364e-01 -7.09092841e-02
-2.87245095e-01 6.84385300e-01 -1.15521932e-02 -8.34826946e-01
-1.78579781e-02 -1.14802527e+00 -1.11144137e+00 -6.72841489e-01
2.32517019e-01 5.81513286e-01 -2.05854222e-01 2.71877885e-01
-6.41835257e-02 5.35702646e-01 -1.05580854e+00 2.05636606e-01
5.01226664e-01 8.49539340e-01 -8.43610525e-01 7.60508358e-01
3.19872856e-01 9.08135295e-01 -1.12504041e+00 -1.30575931e+00
4.99267817e-01 -2.97704339e-01 -7.05253065e-01 -1.92030564e-01
4.97432649e-01 -7.69462705e-01 -1.52344570e-01 1.23997247e+00
3.33754212e-01 -3.10475737e-01 5.92553794e-01 -1.75388288e+00
-9.32025015e-01 9.75217402e-01 1.25289857e-01 -2.03067008e-02
1.68077901e-01 2.34365433e-01 9.07878160e-01 -4.85252254e-02
4.98879343e-01 9.90913689e-01 7.34087527e-01 9.67722595e-01
-1.55883253e+00 -8.22063804e-01 -1.17196299e-01 -1.98869124e-01
-1.21091664e+00 -6.07122183e-01 7.10421264e-01 -5.46806812e-01
7.42988825e-01 8.16836953e-01 2.14608371e-01 1.74127305e+00
-1.32961765e-01 6.24552667e-01 1.17278612e+00 -1.67477757e-01
5.86369991e-01 6.60424709e-01 1.16789080e-01 3.00138116e-01
6.42889380e-01 2.72762269e-01 -7.89763033e-02 -9.91057217e-01
6.53349459e-01 -9.97070447e-02 -3.88738692e-01 -3.57187778e-01
-5.36492229e-01 1.17455149e+00 -7.34656230e-02 1.08076237e-01
-2.61245042e-01 4.96817172e-01 4.66847092e-01 2.61545330e-01
2.04161674e-01 4.14939970e-01 -2.71691859e-01 -1.20653830e-01
-8.67010117e-01 3.72764409e-01 1.12155294e+00 1.18813467e+00
5.81696451e-01 2.42480487e-01 -1.69863015e-01 2.15576679e-01
4.07529235e-01 5.03880501e-01 6.33022860e-02 -9.52711582e-01
3.94445956e-01 2.19763234e-01 4.77953292e-02 -5.93810558e-01
3.22157085e-01 -3.11105728e-01 -6.23216331e-01 2.21033096e-01
5.42394519e-01 -3.98263603e-01 -6.92745149e-01 1.95017242e+00
1.07925721e-01 1.88173696e-01 3.21496278e-01 3.38389903e-01
5.83072066e-01 3.79890621e-01 2.05201536e-01 -6.37249798e-02
1.37995279e+00 -3.36074382e-02 -6.93980038e-01 2.05338389e-01
5.92549980e-01 -2.58578002e-01 9.68881309e-01 4.76380527e-01
-1.14075160e+00 -4.17398475e-02 -1.09328961e+00 7.87543207e-02
-5.55916667e-01 -3.38454098e-01 8.13604653e-01 1.74130964e+00
-9.85406697e-01 4.29547101e-01 -8.47070277e-01 1.35484397e-01
9.28240538e-01 4.43522364e-01 -1.56515792e-01 3.86801600e-01
-1.47880673e+00 5.49502850e-01 1.49924621e-01 -2.41935149e-01
-1.34642327e+00 -9.73862946e-01 -9.02354002e-01 2.64262915e-01
2.27584749e-01 -6.93115056e-01 1.20678651e+00 -8.80881011e-01
-1.26056993e+00 8.22475672e-01 4.61385339e-01 -1.14217067e+00
8.97644281e-01 -6.27327571e-03 -4.35317725e-01 -1.93697095e-01
-4.38139021e-01 2.38638565e-01 8.75831485e-01 -1.54791236e+00
-2.38417551e-01 -2.19574928e-01 3.97950172e-01 -4.44750696e-01
-4.74551737e-01 6.98814392e-02 -7.83992708e-02 -4.08374727e-01
-7.29135931e-01 -7.40427077e-01 -7.41003081e-02 -3.57725441e-01
-9.45903182e-01 1.21378839e-01 1.10901701e+00 -5.38872600e-01
1.23318779e+00 -1.94957507e+00 -3.74116689e-01 6.49231076e-01
5.24577312e-02 5.22540212e-01 3.29794884e-01 3.22367758e-01
7.26475641e-02 7.89338350e-01 -2.73733318e-01 -5.91791570e-01
5.14674842e-01 3.40736032e-01 -9.29006517e-01 5.55021584e-01
7.32176825e-02 9.95064259e-01 -4.31818813e-01 -1.31796867e-01
1.73950076e-01 7.54863381e-01 -7.73814738e-01 2.93600589e-01
-4.10966158e-01 2.14633375e-01 -2.77122855e-01 5.04784346e-01
7.00451493e-01 5.50778629e-03 2.18149245e-01 -1.35873733e-02
4.18801427e-01 4.03740972e-01 -9.32682157e-01 1.10823119e+00
-3.09757978e-01 4.03710753e-01 1.45865172e-01 -6.16111815e-01
6.72898710e-01 5.78495622e-01 1.98598560e-02 -1.06423408e-01
2.84303665e-01 -2.04781160e-01 -9.31789950e-02 -1.62307128e-01
1.96394309e-01 -3.58658880e-02 -3.67591441e-01 5.96980810e-01
5.12100123e-02 -9.51459333e-02 -3.02216738e-01 3.12237352e-01
1.10158956e+00 -6.04771003e-02 1.82824194e-01 -3.73023748e-02
2.82051325e-01 -4.20998752e-01 5.31205237e-01 1.09475935e+00
-1.36122674e-01 4.79898572e-01 9.16582048e-01 -1.73279166e-01
-1.06976140e+00 -1.31062973e+00 -2.03695968e-01 5.75374842e-01
-3.89712423e-01 -5.94428301e-01 -1.35980892e+00 -1.31867766e+00
-1.26432739e-02 1.31730926e+00 -9.77533937e-01 -3.32564443e-01
-1.84566990e-01 -7.20413208e-01 1.27729452e+00 4.79994893e-01
4.80848491e-01 -8.33715141e-01 -5.45130193e-01 -1.62781656e-01
1.96295843e-01 -9.20279980e-01 -3.09429735e-01 7.08174035e-02
-5.11422694e-01 -1.12952936e+00 2.14982424e-02 1.12661399e-01
6.66583836e-01 -8.66908908e-01 1.18926752e+00 1.29952803e-02
-1.27504859e-02 8.08756649e-01 -1.51141316e-01 -8.48873496e-01
-1.19772446e+00 1.92433167e-02 -2.12742969e-01 1.05654158e-01
7.86680698e-01 -6.89296663e-01 -2.51538575e-01 6.03328198e-02
-1.31988537e+00 -5.07995903e-01 3.09361130e-01 5.75436532e-01
3.98023456e-01 8.88182297e-02 4.68213022e-01 -1.81595480e+00
8.12224984e-01 -5.94845176e-01 -9.94014978e-01 3.82949173e-01
-9.73642766e-01 -1.11324891e-01 9.18581426e-01 -4.93965715e-01
-8.75890493e-01 -1.87074170e-01 -4.34417784e-01 -6.80519164e-01
-5.62410891e-01 4.00626808e-02 -7.23217785e-01 1.05192825e-01
7.80433476e-01 1.88171476e-01 -7.36903101e-02 -2.25334376e-01
3.16403717e-01 5.46738446e-01 5.66524744e-01 -6.95062101e-01
1.03041267e+00 4.87769514e-01 -8.07555690e-02 -3.31022471e-01
-3.35264444e-01 2.73205787e-01 -1.26690552e-01 8.86074454e-02
7.88500726e-01 -5.31141937e-01 -1.16045702e+00 2.95168310e-01
-8.98162961e-01 -1.79391131e-01 -6.18381023e-01 1.01407900e-01
-6.92643046e-01 3.02953750e-01 -5.62664866e-01 -1.10655844e+00
-4.57105398e-01 -1.16105211e+00 8.68780136e-01 -1.87643886e-01
-3.36203784e-01 -1.26620185e+00 1.30051672e-02 4.39948648e-01
4.39120412e-01 6.74525499e-01 8.61564994e-01 -1.33207977e+00
-6.90247655e-01 -5.37581682e-01 4.28361833e-01 6.57365978e-01
-3.66217867e-02 1.72770411e-01 -1.48703146e+00 -3.30224484e-01
2.54121065e-01 -3.38962138e-01 5.65473974e-01 1.94881260e-01
1.47008228e+00 -1.04952598e+00 1.50590837e-01 7.20424294e-01
1.43781877e+00 2.90035814e-01 1.08181667e+00 9.13171768e-02
4.47894096e-01 3.02525252e-01 -4.13109958e-02 4.70391870e-01
4.07595895e-02 5.04710674e-01 7.84187675e-01 -9.31330174e-02
4.58992571e-01 -5.78562975e-01 5.06484747e-01 -1.86613187e-01
1.68948546e-01 -3.76617610e-01 -5.38991988e-01 3.79136115e-01
-1.42691803e+00 -1.30074799e+00 1.35186628e-01 2.45439982e+00
8.98314953e-01 2.18456715e-01 2.34791011e-01 8.56521428e-02
3.35303187e-01 3.18582058e-02 -4.83958483e-01 -8.55020344e-01
8.30168650e-02 4.32550311e-01 8.58824015e-01 4.39488381e-01
-1.14464092e+00 5.00540495e-01 6.51964712e+00 7.06553280e-01
-7.10069954e-01 4.25673842e-01 8.68779302e-01 -2.67655045e-01
-9.27324295e-01 1.69377893e-01 -8.45979273e-01 7.88144946e-01
1.29873025e+00 -5.69672063e-02 4.80291963e-01 1.05019259e+00
-3.68417174e-01 3.04497123e-01 -1.39606428e+00 5.95412612e-01
-1.41712338e-01 -1.47164452e+00 1.51725128e-01 6.07169986e-01
5.37826777e-01 -2.82474011e-01 5.93139410e-01 2.90297180e-01
1.12333882e+00 -1.33459175e+00 7.29750216e-01 5.43268859e-01
7.34612286e-01 -1.22326517e+00 7.28605807e-01 2.95508295e-01
-3.18333954e-01 2.14018170e-02 -2.57829338e-01 4.00070429e-01
-6.03835247e-02 5.27156711e-01 -9.89258647e-01 4.99454051e-01
5.57657659e-01 -7.61190876e-02 -5.77602744e-01 3.54203284e-01
-3.52509171e-01 1.24114978e+00 -4.77517933e-01 1.18008107e-01
1.12104766e-01 -1.31531104e-01 5.02329051e-01 1.11184478e+00
8.67813230e-02 -2.26942360e-01 -2.33028963e-01 1.46660209e+00
-5.13545215e-01 -3.57487530e-01 -9.15103197e-01 -9.32433680e-02
6.81700826e-01 1.00057459e+00 -9.15865004e-02 -1.91402689e-01
-2.66008556e-01 7.95015514e-01 8.76359716e-02 3.02180499e-01
-9.63424265e-01 -8.12326372e-02 8.98007333e-01 1.52085811e-01
2.04714924e-01 2.45315582e-01 -2.44795859e-01 -9.26765144e-01
-3.21454629e-02 -1.07470417e+00 6.97032928e-01 -2.74122566e-01
-1.29086661e+00 4.39127386e-01 1.86132297e-01 -8.23988438e-01
-6.94020927e-01 -3.67664635e-01 -6.02545559e-01 9.25569832e-01
-1.04914439e+00 -1.36312962e+00 3.16059470e-01 6.29273474e-01
-2.03259662e-01 -1.68793678e-01 1.27842486e+00 -6.76682405e-03
-6.26243293e-01 1.30112505e+00 -1.06547438e-02 6.00131094e-01
2.56640881e-01 -1.50325918e+00 5.42860150e-01 1.21925306e+00
4.98537183e-01 1.12866724e+00 7.84184873e-01 -5.55671096e-01
-1.26861906e+00 -1.30445302e+00 4.09755886e-01 -9.70477462e-01
3.26626688e-01 -9.07422185e-01 -9.42496240e-01 1.25490272e+00
4.04252321e-01 -9.44761857e-02 1.19936359e+00 6.87567368e-02
-6.87708795e-01 1.56108350e-01 -1.79044569e+00 3.00560802e-01
5.68187356e-01 -7.79207528e-01 -2.79033154e-01 1.52753413e-01
6.21165395e-01 -3.61577660e-01 -1.23895931e+00 2.57799268e-01
3.78404438e-01 -1.03441668e+00 8.86123061e-01 -9.27748084e-01
1.01017408e-01 -1.95747197e-01 -2.84839511e-01 -8.75607133e-01
-4.00622077e-02 -9.52433825e-01 -5.40513813e-01 1.85082805e+00
6.22690737e-01 -9.14100528e-01 9.59903419e-01 1.28857648e+00
3.92192155e-01 -3.44825983e-01 -9.91680980e-01 -7.81818271e-01
4.15955871e-01 -5.30990839e-01 1.11392117e+00 1.20822561e+00
-5.67018211e-01 -1.74234450e-01 -5.44847727e-01 5.40271997e-01
7.76458621e-01 -2.77511358e-01 1.01489246e+00 -9.44689274e-01
-6.75165236e-01 -3.31452042e-01 -6.05359018e-01 -3.64414752e-01
2.65476346e-01 -7.52321541e-01 -4.66987699e-01 -9.84381199e-01
-3.40905599e-02 -4.07416254e-01 -1.27725214e-01 7.20778286e-01
6.94977492e-02 2.29195431e-01 -2.68451013e-02 -1.54477954e-01
-1.23247087e-01 1.30007565e-01 3.31891894e-01 -1.42550811e-01
1.82009235e-01 4.50090438e-01 -1.12992096e+00 5.66518128e-01
7.75082409e-01 -5.45644343e-01 -7.72374451e-01 4.20431271e-02
3.20483655e-01 -3.81902337e-01 8.50738227e-01 -7.58390427e-01
1.36111259e-01 -5.48538379e-02 2.84135550e-01 -3.55559707e-01
1.10356241e-01 -1.15472865e+00 8.41700256e-01 2.98721015e-01
-5.02442181e-01 -3.99111569e-01 2.25136891e-01 5.48066914e-01
1.65401369e-01 -3.79277200e-01 7.42392421e-01 -6.25511259e-02
-2.90197227e-02 4.42328125e-01 -3.01505089e-01 8.11341628e-02
1.22573435e+00 -1.79830343e-01 -5.29564917e-01 -6.16577268e-01
-8.04623067e-01 -7.88314193e-02 8.32464397e-01 2.00462133e-01
4.58222300e-01 -1.13613248e+00 -6.66582644e-01 5.09027243e-01
-2.09820848e-02 6.52189925e-02 2.77663376e-02 4.07394975e-01
-1.18619010e-01 1.83967099e-01 1.10630669e-01 -2.34546885e-01
-1.38638532e+00 8.92667353e-01 5.35383880e-01 -3.70415777e-01
-3.07448953e-01 7.60137975e-01 2.49897361e-01 -4.00419742e-01
3.73494416e-01 -2.31286943e-01 1.59436733e-01 -2.38291472e-01
5.58325410e-01 2.08139852e-01 1.55023113e-01 -3.74789566e-01
-2.23494664e-01 -3.33613157e-01 -2.35217243e-01 -1.71011284e-01
1.08667457e+00 2.79906660e-01 4.23066467e-02 8.77460763e-02
1.30024326e+00 4.03765798e-01 -1.19715667e+00 2.01652013e-02
-2.86889374e-01 -5.14587164e-01 -1.42316967e-01 -1.06080019e+00
-1.11665297e+00 7.65399098e-01 3.74226123e-01 6.60730064e-01
1.08618653e+00 -2.37831846e-02 3.26753378e-01 1.93677768e-01
4.14088577e-01 -7.59907782e-01 -7.10482419e-01 -2.89746672e-01
6.18711889e-01 -6.59351468e-01 -8.37837718e-03 -2.78587133e-01
-5.41906357e-01 6.30061924e-01 3.31155539e-01 7.60601908e-02
5.97436726e-01 8.28335047e-01 4.82480302e-02 -1.75620914e-01
-7.13158906e-01 3.77203733e-01 8.13248977e-02 9.37920213e-01
-1.30943909e-01 1.94772512e-01 2.04511136e-01 8.32591295e-01
-4.79130954e-01 -1.71822086e-01 7.50590086e-01 7.23957837e-01
2.84748644e-01 -1.40432799e+00 -4.21016991e-01 7.00523257e-01
-1.20497894e+00 -1.77692354e-01 -8.20985138e-01 9.09194052e-01
4.55903485e-02 9.37484562e-01 4.47472185e-02 -3.44186097e-01
3.33486855e-01 2.42740050e-01 2.27523312e-01 -3.71720642e-01
-1.13302064e+00 -3.29985291e-01 2.98577845e-01 -5.22198856e-01
-1.83951065e-01 -9.90543187e-01 -7.24041820e-01 -6.36693716e-01
-4.63648081e-01 3.38468552e-01 5.57769299e-01 5.56602478e-01
3.96497965e-01 3.26735526e-01 6.38602078e-01 1.46832064e-01
-1.03982627e+00 -7.97340930e-01 -5.83896756e-01 6.25591516e-01
1.56645894e-01 -2.12867707e-01 -5.67139030e-01 2.01898828e-01] | [5.939434051513672, 7.18695592880249] |
44368405-5ce7-4975-9a9b-e91fc225b696 | semantic-frame-induction-with-deep-metric | 2304.14286 | null | https://arxiv.org/abs/2304.14286v1 | https://arxiv.org/pdf/2304.14286v1.pdf | Semantic Frame Induction with Deep Metric Learning | Recent studies have demonstrated the usefulness of contextualized word embeddings in unsupervised semantic frame induction. However, they have also revealed that generic contextualized embeddings are not always consistent with human intuitions about semantic frames, which causes unsatisfactory performance for frame induction based on contextualized embeddings. In this paper, we address supervised semantic frame induction, which assumes the existence of frame-annotated data for a subset of predicates in a corpus and aims to build a frame induction model that leverages the annotated data. We propose a model that uses deep metric learning to fine-tune a contextualized embedding model, and we apply the fine-tuned contextualized embeddings to perform semantic frame induction. Our experiments on FrameNet show that fine-tuning with deep metric learning considerably improves the clustering evaluation scores, namely, the B-cubed F-score and Purity F-score, by about 8 points or more. We also demonstrate that our approach is effective even when the number of training instances is small. | ['Koichi Takeda', 'Ryohei Sasano', 'Kosuke Yamada'] | 2023-04-27 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 1.28125012e-01 4.23877567e-01 -3.99966925e-01 -6.49784267e-01
-9.23300266e-01 -5.52557051e-01 7.74913192e-01 2.83734173e-01
-3.93476993e-01 7.08880424e-01 7.26166368e-01 -1.22392125e-01
-1.03152201e-01 -8.54610503e-01 -6.98386908e-01 -5.26183128e-01
6.29030913e-02 5.68129420e-01 3.53971839e-01 -1.83861777e-01
1.81242883e-01 1.30811147e-02 -1.76013434e+00 3.41002405e-01
5.81037462e-01 9.27083671e-01 3.46426331e-02 4.67491299e-01
-2.33776778e-01 1.10722435e+00 -5.65599501e-01 -4.03101772e-01
-2.23631218e-01 -4.65551645e-01 -1.32041693e+00 3.12150586e-02
8.19185600e-02 -1.54161707e-01 -2.24309906e-01 8.25998724e-01
2.51412004e-01 3.88099164e-01 3.75890255e-01 -1.13853681e+00
-8.51311922e-01 7.75645494e-01 3.53085250e-02 3.27514499e-01
3.86071414e-01 -2.22031072e-01 1.75990391e+00 -9.86912608e-01
7.50746429e-01 1.47203302e+00 6.02927864e-01 6.16580009e-01
-1.24733675e+00 -3.13041210e-01 1.87370941e-01 6.15321636e-01
-1.16615200e+00 -3.26498747e-01 9.53402460e-01 -4.60888445e-01
9.67684448e-01 -1.07877553e-01 5.37559807e-01 1.04624593e+00
-4.20610547e-01 7.20219493e-01 6.20882511e-01 -8.37578058e-01
4.82067198e-01 -1.80261150e-01 4.45515186e-01 6.47011757e-01
2.99461842e-01 -1.45107463e-01 -5.45514047e-01 1.89426297e-03
4.57665354e-01 -2.12258294e-01 -2.45480780e-02 -3.32380086e-01
-1.27696192e+00 1.31899631e+00 2.89692372e-01 4.61319834e-01
-3.19445394e-02 2.80913651e-01 6.99893534e-01 -1.09002814e-01
6.79747283e-01 7.93885052e-01 -5.54710209e-01 -3.36676002e-01
-5.73819697e-01 3.43130320e-01 4.00099695e-01 8.03267002e-01
1.00554097e+00 -2.69159466e-01 -1.22100040e-01 9.05845344e-01
1.56585827e-01 2.91763902e-01 5.88716328e-01 -1.14969718e+00
2.61058390e-01 8.48887682e-01 1.68011457e-01 -9.31098163e-01
-4.07947451e-01 3.79475027e-01 8.90210569e-02 -3.10771257e-01
3.90937895e-01 -3.15625489e-01 -7.32911706e-01 2.11407399e+00
5.62944889e-01 4.65616494e-01 2.18202785e-01 8.95187080e-01
8.99020672e-01 3.81918103e-01 3.40060651e-01 3.95429283e-02
1.39357185e+00 -9.35286939e-01 -8.44630361e-01 -1.89786479e-01
1.05038512e+00 -5.83121657e-01 1.50234544e+00 -2.12638974e-01
-7.63809025e-01 -6.11392021e-01 -1.14753461e+00 -4.46942180e-01
-4.16944891e-01 -2.36829579e-01 7.36352563e-01 6.29131913e-01
-5.22037685e-01 3.93536866e-01 -1.07709599e+00 -3.83709937e-01
5.88636339e-01 2.66704768e-01 -3.97949427e-01 -1.97832614e-01
-1.39342201e+00 8.31107914e-01 7.01269031e-01 -4.94636565e-01
-7.93608010e-01 -6.79666042e-01 -1.17078161e+00 1.24187812e-01
4.83441591e-01 -5.02969623e-01 1.08587122e+00 -7.17208683e-01
-1.27786040e+00 9.14048970e-01 -4.52376306e-01 -6.66044056e-01
-1.79764256e-01 -2.95558393e-01 -4.39644784e-01 4.40684348e-01
3.81514132e-01 8.33404839e-01 4.79667753e-01 -1.05549097e+00
-7.75071204e-01 -2.78638184e-01 7.72717595e-01 1.56493306e-01
-5.15725493e-01 8.31598714e-02 -1.41046718e-01 -4.79341596e-01
-6.02951348e-02 -7.36819744e-01 -9.75493863e-02 -5.11084974e-01
-1.76947609e-01 -7.33875871e-01 9.39470172e-01 -1.35307357e-01
1.23601770e+00 -2.32979918e+00 -4.22984973e-04 -1.37561336e-02
1.27575889e-01 1.52133510e-01 -1.23048492e-01 1.01587474e-01
-1.23867042e-01 3.51648569e-01 -2.76892763e-02 -2.47236997e-01
2.45685980e-01 7.58606434e-01 -3.85773629e-01 1.37387633e-01
5.49337327e-01 8.42531502e-01 -1.24684811e+00 -6.26862347e-01
2.86342472e-01 3.74250919e-01 -9.54719365e-01 2.08487660e-01
-5.28532326e-01 1.90874755e-01 -4.36520517e-01 3.37238342e-01
2.31773138e-01 -4.09604520e-01 4.96898800e-01 -2.52983689e-01
2.70025551e-01 6.20781898e-01 -1.06817031e+00 1.86553299e+00
-5.27278483e-01 7.29930460e-01 -6.83115661e-01 -1.32485187e+00
7.54159987e-01 4.66193616e-01 5.45456588e-01 -5.36181450e-01
9.04789492e-02 -1.93376932e-02 -1.96403071e-01 -5.51929951e-01
5.46856880e-01 -3.87034833e-01 -2.75403589e-01 5.26104033e-01
4.63942677e-01 -1.05542593e-01 5.27347088e-01 1.22045904e-01
1.22224367e+00 1.33218288e-01 1.21903732e-01 -2.89402753e-01
5.31394005e-01 2.19782814e-01 7.82073200e-01 4.31815505e-01
-1.14534952e-01 7.01548576e-01 7.54698455e-01 -7.57682621e-01
-1.01254416e+00 -9.27499115e-01 -2.36310601e-01 1.52384531e+00
3.43640238e-01 -9.74073112e-01 -9.38311756e-01 -1.05001736e+00
-2.15638101e-01 9.11086738e-01 -8.91537249e-01 -9.38336253e-02
-5.74130833e-01 -7.13925898e-01 3.58689249e-01 8.39870155e-01
2.77523726e-01 -9.30478334e-01 -7.07855821e-01 1.53605521e-01
-4.45386380e-01 -1.51646674e+00 -9.50941741e-02 2.27880269e-01
-4.11023080e-01 -1.51161599e+00 8.79813917e-03 -8.90287280e-01
5.95058918e-01 3.24809849e-02 1.23455322e+00 1.03878371e-01
1.10463716e-01 3.72454673e-01 -8.23340654e-01 -4.55288947e-01
-1.43194303e-01 1.51025383e-02 7.57972002e-02 -9.39739197e-02
9.99056935e-01 -2.25805521e-01 -4.25539225e-01 3.50341797e-01
-1.07611322e+00 -2.15834811e-01 -3.82991821e-01 1.16689312e+00
6.10480845e-01 1.65915996e-01 5.61083913e-01 -1.00040042e+00
4.31729794e-01 -2.65057117e-01 -4.38065588e-01 2.21324906e-01
-2.78157473e-01 2.09742874e-01 4.03544903e-01 -2.97920585e-01
-1.11276937e+00 -2.03452408e-01 -1.37300119e-01 -2.05621451e-01
-6.97154105e-02 4.88480717e-01 -1.98252976e-01 3.98990929e-01
8.59100342e-01 -4.80625510e-01 -2.17155606e-01 -1.35793105e-01
6.94924593e-01 4.56436306e-01 3.37349862e-01 -1.09065080e+00
6.21626735e-01 6.46007359e-01 -2.39133194e-01 -6.70177996e-01
-1.59098268e+00 -3.24070066e-01 -6.17207825e-01 1.53085575e-01
1.31097507e+00 -9.83789623e-01 -2.19844684e-01 -4.08853233e-01
-1.14326560e+00 -3.32209826e-01 -6.34763956e-01 5.24124324e-01
-7.31294990e-01 1.31776318e-01 -3.03656936e-01 -3.77672851e-01
1.73020482e-01 -1.28842378e+00 1.04849994e+00 8.53462145e-03
-6.24313056e-01 -1.33204496e+00 1.55051559e-01 5.47796249e-01
-2.01428477e-02 3.91999036e-01 1.09454596e+00 -9.90665734e-01
-1.92954123e-01 1.30916998e-01 -1.57320082e-01 4.61592466e-01
3.19977015e-01 -1.38476104e-01 -1.25344265e+00 2.37474069e-01
-1.43263459e-01 -5.31416655e-01 5.61240435e-01 3.44940424e-01
1.25498009e+00 1.03862295e-02 -2.02217653e-01 3.50232065e-01
1.23406410e+00 4.42445427e-02 6.28451288e-01 5.21002531e-01
7.18199074e-01 3.57085049e-01 7.94395626e-01 3.91499668e-01
4.99945849e-01 6.14542961e-01 9.10983011e-02 1.96011528e-01
-1.50407225e-01 -3.99743766e-01 1.48454562e-01 6.31326437e-01
-8.65718871e-02 6.63338304e-02 -9.38159108e-01 8.93619001e-01
-1.92264414e+00 -1.18422914e+00 2.55241185e-01 1.72481441e+00
1.00256205e+00 3.09658051e-01 1.37533816e-02 3.88497502e-01
8.48121047e-01 1.87049553e-01 -1.19425654e-01 -3.87832165e-01
-9.97104645e-02 7.34256506e-01 1.32040128e-01 6.26408458e-01
-1.37640440e+00 1.42104256e+00 6.57600927e+00 6.49704993e-01
-7.15437055e-01 5.16549706e-01 4.84126806e-01 6.22933656e-02
-4.91382897e-01 2.68429667e-01 -7.31560349e-01 4.43638563e-01
1.12721801e+00 -4.21218723e-01 1.58049911e-01 1.01481342e+00
2.85815019e-02 1.88421160e-02 -1.46645820e+00 8.03576887e-01
4.52251583e-02 -1.78470361e+00 3.61086577e-02 -1.34776920e-01
1.01814115e+00 -4.32584770e-02 -2.45767400e-01 4.82207000e-01
5.38339734e-01 -9.21667457e-01 5.76284707e-01 1.94093976e-02
6.15680635e-01 -8.31727624e-01 8.72147679e-01 -1.82466641e-01
-9.48156774e-01 -9.51332003e-02 -6.07381105e-01 -3.11326921e-01
1.98774233e-01 6.64860845e-01 -8.65843475e-01 2.47309446e-01
4.19623226e-01 6.93767607e-01 -4.49443161e-01 4.15034354e-01
-6.89014256e-01 7.98328698e-01 -2.44569600e-01 5.98500222e-02
3.59848320e-01 1.57158837e-01 1.07042193e-01 1.00794399e+00
3.27587314e-02 1.00586422e-01 2.14810520e-01 7.14659691e-01
-2.54204929e-01 6.45866916e-02 -5.96500576e-01 -1.50650647e-02
4.94667143e-01 9.66412842e-01 -7.48187065e-01 -5.70987225e-01
-6.44738019e-01 4.09851283e-01 3.66962820e-01 2.13446409e-01
-1.12495708e+00 -3.65216404e-01 9.77786601e-01 -8.02983940e-02
7.02544093e-01 -1.86381936e-01 -3.18760484e-01 -1.26160014e+00
-1.16328903e-01 -6.30783916e-01 4.83103454e-01 -6.24910951e-01
-1.09757566e+00 3.32961112e-01 2.20452875e-01 -7.89377034e-01
-3.22739363e-01 -7.19031453e-01 -4.76138085e-01 3.25525105e-01
-1.36043000e+00 -1.10278416e+00 -1.50722682e-01 5.21975696e-01
6.83000207e-01 -1.52370304e-01 1.16557264e+00 2.52822220e-01
-4.48201239e-01 4.86713827e-01 -2.51293272e-01 4.81619686e-01
5.47318280e-01 -1.31506717e+00 2.15342134e-01 8.40072632e-01
5.80533087e-01 7.10146248e-01 7.32406020e-01 -1.92497313e-01
-9.86152470e-01 -1.19875860e+00 1.07856357e+00 -8.93871903e-01
8.92356277e-01 -2.70079911e-01 -7.45481193e-01 9.67240810e-01
2.87485905e-02 4.05947119e-01 1.05884266e+00 6.27674639e-01
-6.43431544e-01 1.23119149e-02 -1.12119401e+00 4.35852140e-01
1.16900718e+00 -7.67658651e-01 -1.27854145e+00 5.84425807e-01
1.17703104e+00 -2.31952474e-01 -1.04670978e+00 3.99950385e-01
1.67712703e-01 -7.87613332e-01 9.40131843e-01 -8.99619579e-01
6.14168286e-01 -4.27439034e-01 -7.19035387e-01 -1.05457580e+00
-2.50071943e-01 -1.87032461e-01 -7.46490732e-02 1.20518124e+00
4.03878182e-01 -3.41052622e-01 7.18278587e-01 6.32941723e-01
-1.92847520e-01 -4.46242899e-01 -1.04722154e+00 -6.61840618e-01
1.66091427e-01 -7.42416203e-01 8.18834424e-01 1.12205219e+00
3.20134521e-01 6.30037725e-01 2.38116592e-01 1.59837857e-01
3.20666224e-01 8.45397934e-02 6.94062769e-01 -1.14971566e+00
-4.64857742e-02 -6.17670119e-02 -7.47244358e-01 -6.82293653e-01
8.21184099e-01 -8.04134369e-01 -8.52602124e-02 -1.48195100e+00
1.44246206e-01 -4.78328943e-01 -5.56034863e-01 5.27400255e-01
-3.28901172e-01 3.70677203e-01 -1.87367145e-02 -2.77879179e-01
-1.02754951e+00 4.83409911e-01 9.54969227e-01 -1.85833424e-01
6.52695000e-02 -6.17410004e-01 -8.99009466e-01 9.31326210e-01
6.71256542e-01 -3.85053366e-01 -7.35512376e-01 -7.03880489e-01
1.56073511e-01 -5.04457712e-01 2.94364065e-01 -1.14357698e+00
-2.40320086e-01 -2.51596749e-01 1.38112172e-01 -2.09853217e-01
4.03929561e-01 -6.59118474e-01 -3.52590352e-01 -1.76510945e-01
-5.94195604e-01 -7.29823560e-02 3.01189441e-03 5.12178183e-01
-3.83348167e-01 -4.17893767e-01 5.09277403e-01 -5.44446893e-02
-1.02983272e+00 -4.31885198e-02 2.44374368e-02 7.38673627e-01
1.01607299e+00 4.47927602e-02 -3.19363892e-01 -7.66363665e-02
-9.79835033e-01 -1.57785378e-02 4.81908500e-01 3.35183054e-01
4.61680055e-01 -1.66047144e+00 -3.11515450e-01 8.03619921e-02
4.05873895e-01 1.56594411e-01 -1.21761866e-01 3.30440313e-01
-3.58186185e-01 2.87205577e-01 1.38593584e-01 -6.44038439e-01
-1.13562405e+00 6.62208378e-01 -7.11211786e-02 -1.18475601e-01
-5.03050268e-01 8.51251900e-01 1.07714660e-01 -3.67667109e-01
1.14497781e-01 -6.74505293e-01 -6.45466521e-02 1.83411807e-01
5.60492933e-01 1.34956941e-01 1.88917350e-02 -7.18513429e-01
-4.85581309e-01 5.40422261e-01 1.54155344e-01 -1.58809841e-01
1.60837710e+00 1.38180897e-01 -7.60203786e-03 4.60563660e-01
1.43279159e+00 -1.13017425e-01 -1.17468560e+00 -3.52997214e-01
3.22155148e-01 -6.00467086e-01 5.06811254e-02 -2.33743668e-01
-7.56981075e-01 7.87513554e-01 3.05080235e-01 2.98502684e-01
8.24426234e-01 3.54048669e-01 7.51555085e-01 5.38549721e-01
3.62515390e-01 -1.45262110e+00 3.43532562e-01 7.99516797e-01
2.86072403e-01 -1.26555395e+00 -3.07014566e-02 -2.94671923e-01
-6.70352876e-01 1.00650847e+00 4.84619290e-01 -1.66626409e-01
6.65600359e-01 -2.43274700e-02 1.48983663e-02 -2.91197717e-01
-7.03271806e-01 -6.09676540e-01 2.29692757e-02 7.06353486e-01
4.73909020e-01 1.95362717e-01 -2.05226198e-01 5.26082516e-01
-4.14070994e-01 -7.42116943e-02 4.13555801e-01 8.37826192e-01
-6.46105170e-01 -1.27151227e+00 -2.54158527e-01 1.16983488e-01
-5.47122657e-01 8.51954054e-03 -3.47079337e-02 7.81508088e-01
4.02371854e-01 1.20394611e+00 3.83026510e-01 -4.07646924e-01
1.16398714e-01 4.62557346e-01 5.55469692e-01 -9.77262974e-01
8.28125924e-02 -2.67979711e-01 4.03630465e-01 -4.79139805e-01
-9.96798217e-01 -5.27558804e-01 -1.52999914e+00 -9.28963423e-02
-3.04817677e-01 4.07146186e-01 3.56308222e-01 1.38045502e+00
2.33980164e-01 4.61339593e-01 5.14220238e-01 -5.67100704e-01
-5.06974161e-02 -8.23836982e-01 -1.43212333e-01 9.59279120e-01
2.40955949e-01 -1.23196781e+00 -3.00586045e-01 3.55151206e-01] | [10.143475532531738, 9.182047843933105] |
fb75d6b0-ab36-44ea-b9be-b41069d61f72 | adversarial-attacks-neutralization-via-data | 2306.12161 | null | https://arxiv.org/abs/2306.12161v1 | https://arxiv.org/pdf/2306.12161v1.pdf | Adversarial Attacks Neutralization via Data Set Randomization | Adversarial attacks on deep-learning models pose a serious threat to their reliability and security. Existing defense mechanisms are narrow addressing a specific type of attack or being vulnerable to sophisticated attacks. We propose a new defense mechanism that, while being focused on image-based classifiers, is general with respect to the cited category. It is rooted on hyperspace projection. In particular, our solution provides a pseudo-random projection of the original dataset into a new dataset. The proposed defense mechanism creates a set of diverse projected datasets, where each projected dataset is used to train a specific classifier, resulting in different trained classifiers with different decision boundaries. During testing, it randomly selects a classifier to test the input. Our approach does not sacrifice accuracy over legitimate input. Other than detailing and providing a thorough characterization of our defense mechanism, we also provide a proof of concept of using four optimization-based adversarial attacks (PGD, FGSM, IGSM, and C\&W) and a generative adversarial attack testing them on the MNIST dataset. Our experimental results show that our solution increases the robustness of deep learning models against adversarial attacks and significantly reduces the attack success rate by at least 89% for optimization attacks and 78% for generative attacks. We also analyze the relationship between the number of used hyperspaces and the efficacy of the defense mechanism. As expected, the two are positively correlated, offering an easy-to-tune parameter to enforce the desired level of security. The generality and scalability of our solution and adaptability to different attack scenarios, combined with the excellent achieved results, other than providing a robust defense against adversarial attacks on deep learning networks, also lay the groundwork for future research in the field. | ['Roberto Di Pietro', 'Mouna Rabhi'] | 2023-06-21 | null | null | null | null | ['adversarial-attack'] | ['adversarial'] | [ 3.16396177e-01 6.20336607e-02 1.26432240e-01 -2.14236364e-01
-4.73905087e-01 -1.09941137e+00 7.73392022e-01 -2.73581475e-01
-4.15531814e-01 4.21701312e-01 -2.56143868e-01 -5.64061940e-01
-1.16397671e-01 -1.16827440e+00 -7.05342650e-01 -9.74432349e-01
-1.54675961e-01 9.71211195e-02 3.24563086e-01 -3.20884377e-01
1.35720342e-01 1.06420088e+00 -1.14311361e+00 1.23880163e-01
6.43042743e-01 1.08756959e+00 -4.90011334e-01 5.47480822e-01
1.95115969e-01 4.18773353e-01 -1.07412767e+00 -7.50520587e-01
7.81459272e-01 -2.18333408e-01 -6.08106375e-01 -1.41547486e-01
5.03630102e-01 -2.83661544e-01 -5.17031133e-01 1.16979575e+00
4.90889698e-01 -2.03546360e-01 5.15005946e-01 -1.48862123e+00
-4.72573370e-01 6.10013127e-01 -3.65374416e-01 1.04518689e-01
-5.34110740e-02 2.97249168e-01 7.34374583e-01 -5.35915732e-01
2.47052491e-01 1.09298146e+00 4.30991054e-01 9.29910481e-01
-1.05595756e+00 -1.12380207e+00 2.01011747e-01 -1.56272110e-03
-1.25236392e+00 -2.23066285e-01 9.00699437e-01 -2.53620416e-01
5.18143475e-01 6.40957355e-01 3.90379012e-01 1.48833179e+00
4.32919055e-01 3.85865331e-01 1.12343895e+00 -2.51694620e-01
4.59038347e-01 3.48497599e-01 1.95278525e-01 4.65230793e-01
4.55152959e-01 4.18965131e-01 -5.78678735e-02 -5.71220577e-01
5.24871767e-01 3.47120985e-02 -4.01119381e-01 -5.03733635e-01
-5.86799622e-01 1.04300416e+00 4.47054446e-01 1.33355290e-01
-3.19624506e-02 3.96916121e-02 4.20914054e-01 3.70092332e-01
2.11487059e-02 5.82165420e-01 -3.84578347e-01 3.87564301e-01
-5.13482153e-01 2.87713587e-01 9.23725486e-01 5.33104479e-01
3.93112153e-01 5.70481539e-01 1.16557740e-02 5.74482858e-01
1.93532258e-01 6.85569346e-01 3.21630627e-01 -5.51039219e-01
4.50349063e-01 3.80394191e-01 -3.42296124e-01 -1.20225322e+00
-2.47591078e-01 -6.91019952e-01 -6.81736887e-01 6.16402507e-01
3.38913172e-01 -4.72630829e-01 -7.86710858e-01 2.05481911e+00
2.14478508e-01 2.66203731e-01 3.48623335e-01 5.67926586e-01
4.24696475e-01 4.37167943e-01 -3.81723382e-02 -1.87005929e-03
1.09877229e+00 -6.69111013e-01 -3.01266849e-01 -2.27341697e-01
2.36793503e-01 -5.19233346e-01 1.01043046e+00 4.68708396e-01
-7.21104622e-01 -3.21409106e-01 -1.40491939e+00 7.67369568e-01
-4.88810480e-01 -3.33436817e-01 5.15429974e-01 1.38122690e+00
-8.56912553e-01 4.23828751e-01 -7.05622852e-01 -2.20934972e-02
4.09797251e-01 6.31902874e-01 -4.25332785e-01 1.58824131e-01
-1.36347532e+00 7.60623097e-01 4.60179597e-01 -5.10658696e-02
-1.20862854e+00 -5.41801393e-01 -6.46416247e-01 1.90922171e-01
1.69090047e-01 -3.94777477e-01 6.12168372e-01 -7.21308768e-01
-1.49431264e+00 5.56408584e-01 5.71392775e-01 -8.94477725e-01
6.33848429e-01 -1.89295709e-01 -5.87296009e-01 2.13641077e-01
-3.19878280e-01 4.08375889e-01 1.19165850e+00 -1.34604585e+00
-3.21194679e-01 -3.68555188e-01 4.01642293e-01 -6.14117756e-02
-1.05466473e+00 6.81689829e-02 -1.16188392e-01 -7.63135374e-01
-6.28210902e-02 -1.29320729e+00 -3.23541790e-01 -2.59039491e-01
-6.15305126e-01 4.45358992e-01 1.28193772e+00 -4.60424274e-02
9.89715993e-01 -2.29118967e+00 -1.45232044e-02 5.04392505e-01
2.17560217e-01 7.82793462e-01 -3.43541384e-01 4.83185709e-01
-3.58418345e-01 3.99714679e-01 -3.65202755e-01 1.12598345e-01
-1.57157525e-01 1.47175357e-01 -9.20969903e-01 5.70645630e-01
2.59525865e-01 6.28608882e-01 -5.24882376e-01 3.67063731e-02
1.74669132e-01 5.02884150e-01 -6.29061937e-01 2.37838611e-01
1.49301425e-01 2.22974733e-01 -4.61479068e-01 4.00431395e-01
9.58154976e-01 2.42458373e-01 2.40893379e-01 -6.43807203e-02
4.52205837e-01 -7.57872015e-02 -1.26776779e+00 8.15401316e-01
-3.44316483e-01 3.90687287e-01 2.37948839e-02 -1.01088345e+00
1.12375045e+00 1.48107395e-01 2.99588978e-01 -2.26427898e-01
2.82310605e-01 1.02518663e-01 2.29754865e-01 -4.21334319e-02
8.14030506e-03 5.95480427e-02 -3.99688154e-01 4.71998096e-01
6.02158383e-02 -6.88015623e-03 -1.98813796e-01 1.50268495e-01
1.27170777e+00 -2.64765412e-01 4.88015488e-02 -2.80254453e-01
9.16319549e-01 -2.64215678e-01 4.84029412e-01 8.69897485e-01
-2.45286420e-01 2.41243050e-01 5.87663651e-01 -4.82609212e-01
-7.41865754e-01 -1.33626831e+00 -3.80980909e-01 8.17647994e-01
1.90445617e-01 -2.70039976e-01 -9.83567715e-01 -1.23373902e+00
1.62042044e-02 6.95939302e-01 -6.07886553e-01 -7.16492295e-01
-6.63385987e-01 -1.06987619e+00 1.18682563e+00 4.02806818e-01
6.51062548e-01 -9.65301096e-01 -5.31263709e-01 -1.16010725e-01
2.31961280e-01 -1.27014196e+00 -1.16853468e-01 3.99112433e-01
-7.42666245e-01 -1.28620100e+00 -2.05786139e-01 -5.66502690e-01
7.24522531e-01 2.40773425e-01 7.06792951e-01 1.19527556e-01
-1.51467651e-01 3.23444635e-01 -2.33172387e-01 -5.12302637e-01
-7.85275340e-01 -3.97981666e-02 4.28234816e-01 2.21598089e-01
1.08216286e-01 -7.30756640e-01 -3.21217954e-01 4.79186267e-01
-1.22600281e+00 -7.06401110e-01 4.19283748e-01 8.98918509e-01
1.21015079e-01 3.65535051e-01 6.22344375e-01 -9.72600460e-01
7.41875708e-01 -3.83276373e-01 -7.56930828e-01 4.21899706e-02
-5.55145860e-01 -3.95952240e-02 1.15325511e+00 -7.65543997e-01
-6.51501179e-01 7.60859400e-02 -4.08948034e-01 -5.56807339e-01
-2.33939126e-01 7.26643801e-02 -4.10196573e-01 -7.04568267e-01
1.00031638e+00 1.94900066e-01 -7.63969049e-02 -1.73483491e-01
3.36918712e-01 3.95853758e-01 4.62748289e-01 -5.67216337e-01
1.54940081e+00 5.40697575e-01 1.49167463e-01 -6.50233090e-01
-6.54711843e-01 2.49371052e-01 -4.53988880e-01 -2.85308957e-02
5.63435972e-01 -5.84918320e-01 -7.77211249e-01 6.06533706e-01
-7.90094018e-01 -4.04488891e-02 -5.55737577e-02 2.38287419e-01
-2.49600083e-01 4.95674014e-01 -6.25152469e-01 -7.39566684e-01
-5.51029384e-01 -1.42320538e+00 4.61556613e-01 5.25015034e-02
1.36238739e-01 -8.91576469e-01 -8.06315914e-02 3.04518759e-01
4.34541792e-01 7.39308417e-01 9.22578573e-01 -1.30245626e+00
-4.61218029e-01 -6.30405128e-01 7.73953497e-02 7.38799810e-01
5.23838215e-02 4.17797267e-02 -1.03506589e+00 -6.60238326e-01
5.71677268e-01 -4.47765529e-01 7.09481001e-01 -7.59054199e-02
1.26950002e+00 -5.82670510e-01 -1.25961930e-01 9.05366778e-01
1.51347864e+00 4.96741772e-01 6.36471331e-01 5.89236200e-01
6.06129944e-01 3.93461019e-01 4.10191357e-01 3.43322664e-01
-3.52514088e-01 6.27210021e-01 1.03278458e+00 -1.49374008e-01
3.16399634e-01 9.79368240e-02 5.36536872e-01 1.65266559e-01
3.68612170e-01 -3.32257032e-01 -8.73310030e-01 1.20638415e-01
-1.36396396e+00 -1.05357003e+00 1.31662235e-01 2.27067471e+00
6.03068590e-01 5.10682225e-01 8.52591638e-03 4.90546703e-01
5.98982811e-01 3.42049688e-01 -5.74230194e-01 -7.00967431e-01
-8.30633938e-02 5.52423656e-01 6.10424161e-01 4.31893736e-01
-1.32170761e+00 1.08195233e+00 6.70710993e+00 7.45691597e-01
-1.39841676e+00 -1.37535959e-01 4.85232323e-01 -9.64223668e-02
-1.37615144e-01 -1.18203357e-01 -9.39795017e-01 3.03629965e-01
9.33739960e-01 -1.54078677e-01 4.00620610e-01 1.04422903e+00
-2.39166766e-01 5.50639451e-01 -8.48922074e-01 5.13307989e-01
1.97513461e-01 -1.25349975e+00 3.35129261e-01 2.79416412e-01
5.02551913e-01 -8.79144818e-02 4.32562709e-01 2.93678522e-01
5.37427425e-01 -1.00784314e+00 5.15873015e-01 -2.44168490e-01
4.49806809e-01 -1.23511243e+00 6.60125613e-01 3.84108841e-01
-8.04383397e-01 -4.78565127e-01 -3.40544939e-01 1.94257542e-01
-2.77288824e-01 1.85561731e-01 -7.61227310e-01 5.31231403e-01
5.05789757e-01 -2.62526143e-02 -6.34660363e-01 4.88651544e-01
-4.21606004e-01 7.32528627e-01 -1.74182445e-01 1.89378783e-01
3.39423448e-01 6.57532290e-02 7.12075710e-01 1.10991848e+00
-5.33142984e-02 -1.36341676e-01 3.32652926e-01 7.32519686e-01
-2.04917267e-02 -5.86559139e-02 -8.14744353e-01 8.25149566e-02
6.95392013e-01 1.30836630e+00 -5.67661345e-01 1.86129920e-02
-7.42008463e-02 7.05185413e-01 2.32774839e-01 2.50858873e-01
-1.27657866e+00 -4.90226865e-01 1.08313632e+00 -8.18017945e-02
1.87100217e-01 -2.46116400e-01 -4.43560779e-01 -1.10002077e+00
-5.80378547e-02 -1.40251160e+00 3.68379146e-01 1.22643910e-01
-1.21545732e+00 1.07898176e+00 6.40551932e-03 -1.24392819e+00
-2.71239758e-01 -6.67162478e-01 -7.82653987e-01 8.17319572e-01
-1.13632035e+00 -9.66419518e-01 -2.23894373e-01 9.55851972e-01
2.12955505e-01 -7.31815577e-01 1.05319297e+00 1.49094403e-01
-8.01540315e-01 1.12223780e+00 -1.14361517e-01 4.31628764e-01
4.55761552e-01 -9.24373209e-01 7.17242062e-01 1.34052360e+00
3.31894070e-01 8.37253571e-01 6.82119787e-01 -3.78379941e-01
-1.33411491e+00 -1.22260213e+00 8.87406394e-02 -5.38488388e-01
7.63705194e-01 -6.68568909e-01 -9.17880476e-01 4.93600905e-01
-8.42037238e-03 6.95657171e-03 8.78418922e-01 -1.12642944e-01
-7.26981580e-01 -2.16193929e-01 -1.43156981e+00 8.52256000e-01
7.55183995e-01 -4.35617983e-01 -3.40308070e-01 2.85014689e-01
8.00447404e-01 -4.61711049e-01 -7.70029604e-01 6.55519962e-01
3.86893451e-01 -1.12477183e+00 1.19922829e+00 -7.30603099e-01
1.87849075e-01 -1.66993529e-01 -3.05612355e-01 -1.05698764e+00
-2.14524597e-01 -5.43733180e-01 -1.58099338e-01 1.12521827e+00
3.99383724e-01 -9.51895297e-01 9.82574046e-01 4.64690000e-01
-1.43293058e-02 -9.69785392e-01 -7.76359022e-01 -1.04699969e+00
3.39056730e-01 -5.42247713e-01 7.40566790e-01 1.04842103e+00
-4.12555814e-01 -5.02930395e-02 -4.73914683e-01 7.63431907e-01
7.21310735e-01 -2.28110582e-01 1.02136862e+00 -9.95232701e-01
-4.58528191e-01 -3.52286965e-01 -5.79694688e-01 -2.90376931e-01
2.85099864e-01 -8.61630917e-01 -3.98894876e-01 -7.71186113e-01
-1.66928649e-01 -4.89574403e-01 -4.21718329e-01 7.61477590e-01
-6.73217028e-02 5.64207673e-01 4.74916667e-01 9.04053450e-02
1.76189274e-01 3.08640450e-01 7.16934860e-01 -1.51167318e-01
-1.95349514e-01 2.42033631e-01 -8.46094966e-01 8.06610048e-01
9.60219383e-01 -6.77782118e-01 -6.82588816e-01 -1.90540656e-01
-3.26411963e-01 -2.30606437e-01 3.83660942e-01 -1.25776565e+00
-1.22199252e-01 -2.08463892e-01 4.77830738e-01 -3.70154455e-02
3.53155881e-01 -9.50528026e-01 1.02152802e-01 8.90112817e-01
-3.01315784e-01 1.76599959e-03 3.69853586e-01 5.17396629e-01
7.69151328e-03 -2.36452490e-01 1.25870717e+00 1.65489092e-01
-3.43025744e-01 3.98467094e-01 -1.27642199e-01 -2.71554552e-02
1.37951064e+00 -2.05900773e-01 -4.26999241e-01 -2.32310534e-01
-7.16226518e-01 -1.29761919e-01 4.41954672e-01 6.22666180e-01
5.45673132e-01 -1.15802801e+00 -6.27715349e-01 5.59822440e-01
-1.59974381e-01 -5.19796550e-01 -7.48611316e-02 9.06723216e-02
-4.05585736e-01 2.65602142e-01 -5.80579162e-01 -4.41697180e-01
-1.40845275e+00 8.68094683e-01 5.18484235e-01 -3.55007142e-01
-5.51499367e-01 7.41780221e-01 3.82890821e-01 -3.00853312e-01
4.45008606e-01 3.23661953e-01 -2.36812651e-01 -2.32048675e-01
5.06221592e-01 7.44408965e-02 9.34244990e-02 -4.40056235e-01
-5.14018118e-01 2.30931506e-01 -2.81204671e-01 -1.58520583e-02
1.15195405e+00 4.22703654e-01 9.62884426e-02 -2.77167290e-01
1.17097783e+00 3.85951519e-01 -1.11149621e+00 -1.15099093e-02
-2.61032879e-01 -4.65955526e-01 -1.81606427e-01 -7.58678317e-01
-1.44510174e+00 7.18742371e-01 6.68329895e-01 4.54448581e-01
1.32768142e+00 -4.42800164e-01 6.67777002e-01 4.41818267e-01
4.42189723e-01 -5.07663369e-01 2.80334085e-01 4.77806687e-01
8.33952188e-01 -7.99311280e-01 -7.59119987e-02 -4.96350884e-01
-6.20922029e-01 1.02718616e+00 9.10048604e-01 -5.10209560e-01
5.84986567e-01 5.94304383e-01 2.78508544e-01 6.76809549e-02
-5.70569277e-01 2.54071772e-01 2.18385741e-01 7.61216223e-01
-2.50348412e-02 -2.22317681e-01 -3.33569467e-01 3.94425362e-01
-4.68732446e-01 -6.79002821e-01 6.16761506e-01 7.62537599e-01
-2.77934819e-01 -1.41996336e+00 -5.19088507e-01 6.39647767e-02
-7.31737018e-01 2.69764941e-02 -6.63212657e-01 8.66087556e-01
8.49828031e-03 9.19724703e-01 -3.14570278e-01 -8.22350025e-01
3.34826231e-01 -5.21863960e-02 1.32759228e-01 -4.99166608e-01
-9.55248117e-01 -3.04956555e-01 -7.00828359e-02 -4.87063289e-01
1.59581318e-01 -3.20863843e-01 -9.57108378e-01 -3.57648492e-01
-3.00623715e-01 9.29169580e-02 7.13097453e-01 7.73738503e-01
2.10160851e-01 4.84842241e-01 1.18784833e+00 -6.59027278e-01
-1.15254211e+00 -4.52840567e-01 -4.21312541e-01 3.19288909e-01
1.13778941e-01 -6.20607376e-01 -6.26916230e-01 -4.26467508e-01] | [5.609267234802246, 7.8031511306762695] |
3e14e80a-ffd3-4a47-b2e8-a274081569e3 | on-the-cross-dataset-generalization-for | 2201.00267 | null | https://arxiv.org/abs/2201.00267v4 | https://arxiv.org/pdf/2201.00267v4.pdf | On the Cross-dataset Generalization in License Plate Recognition | Automatic License Plate Recognition (ALPR) systems have shown remarkable performance on license plates (LPs) from multiple regions due to advances in deep learning and the increasing availability of datasets. The evaluation of deep ALPR systems is usually done within each dataset; therefore, it is questionable if such results are a reliable indicator of generalization ability. In this paper, we propose a traditional-split versus leave-one-dataset-out experimental setup to empirically assess the cross-dataset generalization of 12 Optical Character Recognition (OCR) models applied to LP recognition on nine publicly available datasets with a great variety in several aspects (e.g., acquisition settings, image resolution, and LP layouts). We also introduce a public dataset for end-to-end ALPR that is the first to contain images of vehicles with Mercosur LPs and the one with the highest number of motorcycle images. The experimental results shed light on the limitations of the traditional-split protocol for evaluating approaches in the ALPR context, as there are significant drops in performance for most datasets when training and testing the models in a leave-one-dataset-out fashion. | ['David Menotti', 'Valter Estevam', 'Diego R. Lucio', 'Everton V. Cardoso', 'Rayson Laroca'] | 2022-01-02 | null | null | null | null | ['scene-text-recognition', 'license-plate-recognition', 'license-plate-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 8.09843093e-02 -7.50988007e-01 -9.35752988e-02 -5.30031979e-01
-1.14671862e+00 -9.36367273e-01 6.63981676e-01 -2.40284309e-01
-4.25182551e-01 3.39397907e-01 -3.69488180e-01 -4.15875703e-01
1.18597083e-01 -3.44748884e-01 -9.65328991e-01 -5.16472161e-01
1.83790416e-01 7.05321252e-01 5.74334443e-01 -1.91228576e-02
5.95281541e-01 1.01114357e+00 -1.47169173e+00 4.83009934e-01
5.69423437e-01 8.47163677e-01 -2.96435319e-02 6.32475615e-01
1.74210787e-01 6.36727273e-01 -5.63044369e-01 -7.83122897e-01
5.49100220e-01 4.11363877e-03 -3.55262518e-01 4.92344499e-01
1.12390554e+00 -3.02976519e-01 -5.87008834e-01 7.21213937e-01
3.97738189e-01 7.73470327e-02 8.27163100e-01 -1.23795700e+00
-8.40888143e-01 -7.87916481e-02 -6.70826435e-01 2.69933462e-01
3.92609015e-02 4.81922656e-01 7.43618786e-01 -1.09432769e+00
7.22951293e-01 1.06046534e+00 9.91052926e-01 3.21626633e-01
-1.12505209e+00 -5.43138981e-01 -1.50296435e-01 1.29250720e-01
-1.55337965e+00 -6.49886012e-01 4.41473156e-01 -6.75195575e-01
9.39144075e-01 1.08787253e-01 8.71494189e-02 1.07453001e+00
1.04552776e-01 9.67201352e-01 1.29824722e+00 -2.43185639e-01
1.06201470e-01 4.41397727e-01 2.35268444e-01 3.37188661e-01
4.48991388e-01 -1.36782661e-01 -3.60697687e-01 2.70430177e-01
6.05685174e-01 -3.26647490e-01 1.14275254e-01 -3.33014011e-01
-7.74409115e-01 6.82109475e-01 -9.71107930e-02 -1.34899877e-02
1.82148159e-01 -8.19811597e-02 6.09777331e-01 2.41025612e-01
3.76373798e-01 4.76983547e-01 -4.06879306e-01 -1.58107802e-01
-1.15467930e+00 3.01122367e-01 5.67000985e-01 1.13529074e+00
6.98518872e-01 9.27127749e-02 9.44202126e-04 1.21368098e+00
2.34459743e-01 5.62158525e-01 4.68878806e-01 -6.28011525e-01
8.87054920e-01 3.19599867e-01 -4.66959812e-02 -8.96166921e-01
-3.18244636e-01 -1.65889934e-01 -2.46091470e-01 4.24657494e-01
5.52929461e-01 -1.65782310e-02 -9.97978091e-01 9.17240143e-01
-2.87333965e-01 4.68324646e-02 2.08993360e-01 7.67204702e-01
7.83906102e-01 5.90923488e-01 -1.05650239e-01 3.47452223e-01
1.14900053e+00 -1.16427755e+00 -1.55949920e-01 -5.60383558e-01
7.55031049e-01 -9.00356531e-01 1.13071251e+00 5.34556389e-01
-6.38132155e-01 -5.93237936e-01 -1.27231646e+00 -2.49440134e-01
-6.50556087e-01 6.91836417e-01 4.23787348e-02 8.93748462e-01
-9.11782086e-01 3.00805002e-01 -4.37617868e-01 -5.90101779e-01
6.13836050e-01 2.10790187e-01 -7.97150433e-01 -3.90389860e-01
-6.24431074e-01 1.01540017e+00 2.61267513e-01 2.78350532e-01
-8.02713096e-01 -5.86117089e-01 -6.09293699e-01 -2.82378316e-01
2.83437848e-01 3.67652208e-01 1.05679595e+00 -8.15892577e-01
-1.39305234e+00 1.15085542e+00 3.16369534e-01 -4.03383404e-01
8.25636685e-01 -1.17321633e-01 -7.14341223e-01 1.13483541e-01
-1.05845854e-01 6.10724568e-01 6.18304789e-01 -1.39572191e+00
-6.25286102e-01 -4.17228401e-01 -1.95151553e-01 6.72779903e-02
1.68070078e-01 2.64568210e-01 -7.61357725e-01 -3.01739931e-01
-4.44964208e-02 -1.26445818e+00 2.56006181e-01 -4.32176828e-01
-3.65749151e-01 -2.79850066e-02 8.52987766e-01 -7.90690720e-01
6.21387243e-01 -2.35764384e+00 -7.21701741e-01 7.05119669e-02
-1.65575653e-01 6.04955554e-01 -2.62477547e-01 4.04597312e-01
-1.82124004e-01 2.40548685e-01 7.50214979e-03 -5.89144528e-01
1.57473683e-01 4.35427129e-02 -3.85136962e-01 8.93856406e-01
3.80877107e-01 7.15618968e-01 -2.47438803e-01 -4.88146573e-01
3.13462108e-01 2.61465371e-01 -1.99323431e-01 -8.08555335e-02
5.36446869e-02 1.58284202e-01 -1.03842653e-01 7.65477717e-01
8.59218001e-01 2.06689164e-01 -1.78472862e-01 -1.18226921e-02
-2.98118621e-01 -8.45823660e-02 -9.55014229e-01 1.12732708e+00
-1.74317449e-01 1.52479160e+00 -3.02836955e-01 -6.08337522e-01
1.20096874e+00 9.45383981e-02 -2.45810095e-02 -8.79109621e-01
-6.84783459e-02 5.33526897e-01 -9.78712365e-02 -5.93196332e-01
8.43751252e-01 1.77633345e-01 -5.94634265e-02 -5.42027242e-02
-1.29548237e-01 5.18152229e-02 5.54048061e-01 -1.91001266e-01
7.65077591e-01 1.43257454e-02 -3.97888184e-01 -2.84449160e-01
5.38511872e-01 1.92390680e-01 2.36083001e-01 8.13671768e-01
-3.54915202e-01 1.27256751e+00 6.42228723e-01 -4.92168188e-01
-1.39888346e+00 -9.86101091e-01 -4.40042496e-01 1.00994861e+00
1.74825676e-02 2.18325898e-01 -5.97297788e-01 -6.65600240e-01
1.79488391e-01 7.74275482e-01 -5.43432832e-01 1.68688133e-01
-5.47741413e-01 -7.38005936e-01 1.18215525e+00 7.27185667e-01
5.54590046e-01 -7.13230371e-01 -2.48320714e-01 -2.80891508e-01
1.44331455e-01 -1.61957586e+00 -3.24715018e-01 -1.73107728e-01
-6.87746167e-01 -1.00853109e+00 -7.10683703e-01 -8.24068069e-01
5.26120901e-01 3.10240090e-01 9.89659131e-01 -2.82422990e-01
-1.60182416e-02 3.16593558e-01 -3.55984151e-01 -3.15706015e-01
-8.05383146e-01 4.27403301e-02 -1.79090127e-02 2.23491952e-01
6.52670801e-01 8.76503214e-02 -3.86240065e-01 7.76752710e-01
-1.02714539e+00 -2.71935970e-01 1.08462512e+00 2.43824124e-01
3.61131728e-01 1.49595747e-02 3.50233495e-01 -7.17515826e-01
5.81294954e-01 -2.19602734e-01 -8.97484660e-01 4.18564856e-01
-3.21032703e-01 -2.33055919e-01 5.06150067e-01 -2.53476053e-01
-8.42889607e-01 1.26034450e-02 -1.69064105e-01 -6.04874074e-01
-6.93004787e-01 3.38919222e-01 -1.61244065e-01 -1.38971850e-01
5.72115123e-01 4.48186874e-01 -2.86137685e-02 -5.55576682e-01
7.09910616e-02 1.03718281e+00 6.89405501e-01 -4.43281770e-01
7.06459820e-01 2.60727048e-01 -1.69745848e-01 -1.39539850e+00
-2.70352155e-01 -7.39295244e-01 -6.63775027e-01 -4.49478358e-01
7.96133339e-01 -8.55302274e-01 -5.84712386e-01 9.91874993e-01
-8.59166920e-01 -3.58931601e-01 3.61801200e-02 5.09187043e-01
-4.48731244e-01 4.45357174e-01 -4.61148262e-01 -6.56202316e-01
2.08127335e-01 -1.52698231e+00 1.21451938e+00 1.51234418e-01
2.20305264e-01 -7.41276383e-01 8.93595964e-02 8.94234896e-01
2.49648079e-01 3.81596237e-02 7.17263460e-01 -1.09467769e+00
-5.39317667e-01 -8.08592379e-01 -4.27308291e-01 6.73366368e-01
-3.76767159e-01 5.02927184e-01 -1.22044241e+00 -3.31689358e-01
-5.42286277e-01 -4.64497268e-01 8.96386087e-01 1.36902049e-01
9.13805485e-01 9.43879187e-02 -5.13175800e-02 3.42648596e-01
1.60041535e+00 1.25363573e-01 1.26632130e+00 6.97844923e-01
6.86071754e-01 5.84983468e-01 3.62882704e-01 -8.88030007e-02
1.07484609e-01 8.83192599e-01 1.00823745e-01 -8.28605816e-02
-1.82125673e-01 -1.47439793e-01 7.12839305e-01 4.57357675e-01
1.29710063e-01 -5.22858977e-01 -1.18891335e+00 4.71253037e-01
-1.30876994e+00 -8.04724932e-01 -2.82079041e-01 2.48274112e+00
3.47688496e-01 2.81760722e-01 3.16103399e-01 -1.33834824e-01
8.59840333e-01 1.26196951e-01 -2.80009747e-01 -6.12329423e-01
-3.65445733e-01 -6.18596494e-01 8.94084394e-01 2.55605698e-01
-1.31136620e+00 8.68795872e-01 6.49005461e+00 9.00491536e-01
-1.37564075e+00 -2.83041269e-01 9.26634431e-01 2.47096881e-01
3.27653170e-01 -3.83224875e-01 -1.20076811e+00 3.87371421e-01
1.03265011e+00 4.81875092e-01 8.51183236e-02 8.54949594e-01
1.00457601e-01 -2.28554189e-01 -1.31271458e+00 1.03471792e+00
4.42654014e-01 -1.27903783e+00 -5.18650785e-02 1.61364630e-01
8.35242450e-01 6.51811600e-01 4.86248076e-01 3.16470325e-01
-1.86497390e-01 -1.07844222e+00 8.74451101e-01 3.73024166e-01
9.09398735e-01 -6.28128290e-01 9.43064988e-01 1.45067871e-01
-6.91067278e-01 -3.46471705e-02 -5.69832683e-01 5.03644168e-01
-2.08196998e-01 7.66227916e-02 -1.02671731e+00 2.74516881e-01
5.30898631e-01 5.35128295e-01 -1.22449589e+00 1.34862876e+00
1.78368881e-01 6.97174311e-01 -1.43418252e-01 -2.50127795e-03
4.42202806e-01 -3.31314325e-01 5.23050487e-01 1.79507363e+00
1.62850127e-01 -5.10003686e-01 -1.53579861e-01 6.80007994e-01
-1.20076105e-01 1.07358657e-01 -6.84970260e-01 -1.49359852e-01
1.13281950e-01 1.14360404e+00 -9.22063053e-01 8.29496607e-02
-8.00931036e-01 6.61427081e-01 1.54916849e-03 4.95582014e-01
-9.70562637e-01 -2.56532580e-01 4.16938543e-01 4.66898888e-01
5.99246919e-01 -3.87512267e-01 -1.59870893e-01 -9.23367441e-01
2.63460606e-01 -9.59124863e-01 1.06047720e-01 -8.82180750e-01
-1.49497163e+00 4.45673525e-01 1.14708766e-01 -1.44460917e+00
1.12660222e-01 -1.36072731e+00 -4.15189087e-01 5.75334847e-01
-1.45265853e+00 -1.20113480e+00 -8.86654854e-02 2.38525420e-01
7.85448432e-01 -5.88887453e-01 2.80741483e-01 4.20501590e-01
-8.69305313e-01 9.02462959e-01 8.94946694e-01 6.93148971e-01
7.77409077e-01 -9.90774989e-01 4.09464508e-01 1.12993693e+00
2.07576096e-01 3.81144524e-01 6.32231414e-01 -4.08660412e-01
-1.35043967e+00 -1.24035990e+00 4.72618610e-01 -8.36169660e-01
5.14311850e-01 -6.08626306e-01 -8.70179832e-01 8.55796337e-01
3.24182995e-02 7.71915987e-02 6.92386389e-01 -5.51999845e-02
-5.65943241e-01 -3.93770546e-01 -1.00277293e+00 5.85825086e-01
2.89288193e-01 -4.89599824e-01 -4.45780843e-01 3.21586877e-01
1.04859784e-01 -3.59291643e-01 -7.29496717e-01 1.93461552e-01
7.77012408e-01 -9.81461108e-01 7.11895525e-01 -3.92951280e-01
7.49487042e-01 -4.56036299e-01 -3.46853346e-01 -8.32783103e-01
-1.11324392e-01 1.32233202e-02 5.46477675e-01 1.46664441e+00
7.70439446e-01 -7.45186031e-01 8.09921145e-01 8.28185618e-01
-3.53802145e-01 -6.14022195e-01 -8.57845366e-01 -1.12015975e+00
4.07109678e-01 -6.80216014e-01 2.30569482e-01 5.74630857e-01
-5.20024180e-01 8.84123445e-02 -2.87173778e-01 4.51074302e-01
3.39474797e-01 -1.13893352e-01 1.03194356e+00 -9.91983116e-01
-1.85384646e-01 -6.17545307e-01 -1.05320072e+00 -9.65415835e-01
1.11593485e-01 -7.12609947e-01 3.23143303e-01 -1.21094477e+00
2.95361578e-01 -2.79381365e-01 -5.21440022e-02 1.57408595e-01
2.57753521e-01 6.60173595e-01 3.47842664e-01 5.27690291e-01
-9.53903377e-01 4.51918654e-02 7.86660552e-01 -2.98122078e-01
-2.50340421e-02 4.54402491e-02 -2.40888864e-01 6.59393966e-01
6.74780548e-01 -1.92716405e-01 -1.82519808e-01 -6.27838731e-01
4.57401685e-02 -4.15539593e-01 4.07735765e-01 -1.23695529e+00
1.48416713e-01 2.34013557e-01 6.03253245e-01 -7.81886458e-01
2.24502787e-01 -7.16327250e-01 -8.59766155e-02 -1.37905449e-01
-4.28785235e-01 1.45966545e-01 6.43972874e-01 4.80371445e-01
-1.23539388e-01 -4.54701930e-01 9.43561792e-01 1.25610620e-01
-1.21857154e+00 4.63778749e-02 -6.45935774e-01 1.63280100e-01
9.84192431e-01 -8.17605078e-01 -6.24901474e-01 -9.04612318e-02
-2.62854934e-01 -5.46054803e-02 6.97065711e-01 6.68337882e-01
3.74623537e-01 -1.00634837e+00 -1.03767359e+00 2.67643034e-01
5.60427010e-01 -3.03475052e-01 2.69248843e-01 7.71433711e-01
-1.26175332e+00 8.29037189e-01 -3.30077022e-01 -8.55648279e-01
-1.26381743e+00 4.20643836e-01 4.75409299e-01 -1.47104666e-01
-4.81504738e-01 6.31514907e-01 1.37086079e-01 -6.21031106e-01
-1.48246968e-02 -1.58118770e-01 -1.35061875e-01 -1.27425985e-02
4.50518489e-01 7.07863629e-01 5.89055777e-01 -1.13175833e+00
-3.18181455e-01 6.03868663e-01 -4.11814928e-01 -5.22156097e-02
1.15649629e+00 -7.86267817e-02 1.84571609e-01 5.32361805e-01
1.46490443e+00 7.69962817e-02 -1.52954543e+00 1.83874041e-01
-4.77828234e-02 -7.27997541e-01 -1.73422724e-01 -7.13468432e-01
-1.02006519e+00 9.36486661e-01 7.25532174e-01 4.88970149e-03
6.81502640e-01 -1.00956857e-01 4.11870688e-01 5.86794674e-01
2.52704412e-01 -1.39672756e+00 -1.44269347e-01 5.57510555e-01
1.01315224e+00 -1.31786609e+00 -2.39105895e-02 -8.95301178e-02
-9.34006989e-01 1.30942416e+00 4.94189590e-01 -1.25295609e-01
1.94196552e-01 1.55878022e-01 1.61696449e-01 6.61566108e-02
-3.56143922e-01 1.76147714e-01 2.89787352e-01 5.91983140e-01
3.30142885e-01 -4.68172245e-02 1.65420726e-01 2.68634379e-01
2.47454364e-02 -1.12956971e-01 7.57185757e-01 6.14561796e-01
-2.71333277e-01 -7.42794931e-01 -5.19226849e-01 3.97508502e-01
-3.75537068e-01 2.07578391e-01 -4.60588455e-01 1.32660401e+00
-1.83585298e-03 8.87060285e-01 1.63270637e-01 -3.39694083e-01
5.05691648e-01 2.11505070e-01 3.06230247e-01 -3.09193820e-01
-1.07028529e-01 -2.21760452e-01 2.39881352e-01 -1.71724826e-01
-4.87578195e-03 -1.04549253e+00 -7.95754135e-01 -3.42239380e-01
-3.89156997e-01 -3.78228009e-01 7.38340378e-01 9.95208204e-01
2.29200691e-01 1.32239267e-01 5.70228159e-01 -8.27676833e-01
-6.60199404e-01 -9.01610672e-01 -8.35345805e-01 6.14531696e-01
1.89422205e-01 -4.66512710e-01 -2.54483014e-01 1.93238974e-01] | [9.835419654846191, -4.924130439758301] |
ea4acaba-0674-46bf-9b7e-1f078fd96dc8 | detection-d-objets-dans-les-documents | 2301.11753 | null | https://arxiv.org/abs/2301.11753v1 | https://arxiv.org/pdf/2301.11753v1.pdf | Détection d'Objets dans les documents numérisés par réseaux de neurones profonds | In this thesis, we study multiple tasks related to document layout analysis such as the detection of text lines, the splitting into acts or the detection of the writing support. Thus, we propose two deep neural models following two different approaches. We aim at proposing a model for object detection that considers the difficulties associated with document processing, including the limited amount of training data available. In this respect, we propose a pixel-level detection model and a second object-level detection model. We first propose a detection model with few parameters, fast in prediction, and which can obtain accurate prediction masks from a reduced number of training data. We implemented a strategy of collection and uniformization of many datasets, which are used to train a single line detection model that demonstrates high generalization capabilities to out-of-sample documents. We also propose a Transformer-based detection model. The design of such a model required redefining the task of object detection in document images and to study different approaches. Following this study, we propose an object detection strategy consisting in sequentially predicting the coordinates of the objects enclosing rectangles through a pixel classification. This strategy allows obtaining a fast model with only few parameters. Finally, in an industrial setting, new non-annotated data are often available. Thus, in the case of a model adaptation to this new data, it is expected to provide the system as few new annotated samples as possible. The selection of relevant samples for manual annotation is therefore crucial to enable successful adaptation. For this purpose, we propose confidence estimators from different approaches for object detection. We show that these estimators greatly reduce the amount of annotated data while optimizing the performances. | ['Mélodie Boillet'] | 2023-01-27 | null | null | null | null | ['document-layout-analysis', 'line-detection'] | ['computer-vision', 'computer-vision'] | [ 4.33631837e-01 5.68016805e-02 1.40924439e-01 -2.60234892e-01
-3.11756551e-01 -3.30029160e-01 5.23458898e-01 4.66360897e-01
-4.73987550e-01 4.99802053e-01 -4.90261108e-01 -2.49543354e-01
-1.57840297e-01 -8.35405707e-01 -7.66137123e-01 -4.85709459e-01
2.52586633e-01 7.11977422e-01 5.25261998e-01 8.07895884e-02
5.79735518e-01 1.17193282e+00 -1.87562990e+00 4.69604075e-01
8.72682452e-01 1.01077342e+00 7.66847730e-01 8.45439732e-01
-2.16586560e-01 4.81316060e-01 -9.44529295e-01 -1.00289606e-01
6.49320111e-02 -2.74986953e-01 -4.04899478e-01 4.71795350e-01
6.51440084e-01 -4.79093075e-01 3.16144854e-01 7.26468146e-01
5.24592519e-01 1.98248297e-01 9.65554774e-01 -8.75370622e-01
-2.12264970e-01 4.56731766e-01 -5.57791352e-01 -1.42788976e-01
1.90988347e-01 -1.39409006e-01 7.20530152e-01 -8.46966922e-01
7.00793087e-01 8.46719980e-01 6.73707306e-01 5.79995036e-01
-1.29587376e+00 -9.08176005e-02 1.37470096e-01 1.78915337e-01
-1.42402899e+00 -3.47609580e-01 7.66016781e-01 -6.02643728e-01
7.12322295e-01 4.93962616e-01 4.82323498e-01 1.10066020e+00
-2.09079757e-01 1.03094876e+00 8.30184519e-01 -1.11242533e+00
4.14380699e-01 7.84864783e-01 3.05447042e-01 6.65854394e-01
3.91799122e-01 -3.82886857e-01 -3.18360925e-01 1.87204957e-01
7.55801141e-01 -2.04803705e-01 -3.05124789e-01 -4.80216771e-01
-7.81444013e-01 5.60094893e-01 1.47436410e-01 7.26080775e-01
-2.67055035e-01 -3.07743013e-01 2.90929884e-01 -1.46803811e-01
3.13223839e-01 4.82335776e-01 -3.32141578e-01 1.46124825e-01
-1.30275428e+00 1.24932528e-01 8.74361455e-01 9.74851489e-01
4.93325800e-01 -1.04610339e-01 -3.07981491e-01 9.72936153e-01
4.66003716e-02 2.90746599e-01 3.66684467e-01 -4.25490946e-01
5.30448794e-01 7.67766714e-01 2.52529889e-01 -1.12994576e+00
-6.27912581e-01 -5.26130319e-01 -6.05336189e-01 4.41056192e-01
7.64402211e-01 2.79090535e-02 -7.42042363e-01 1.31352103e+00
2.67587870e-01 -5.71455002e-01 -1.70797318e-01 6.75584912e-01
1.03944227e-01 6.84981823e-01 -2.50188023e-01 -2.72781640e-01
1.24579668e+00 -8.87940764e-01 -7.34713972e-01 -6.87491298e-02
8.79959404e-01 -7.99428284e-01 1.28287709e+00 1.04423547e+00
-9.10147667e-01 -9.23109412e-01 -1.19535279e+00 -6.20903037e-02
-7.28103757e-01 1.18691981e+00 2.16381267e-01 7.84842134e-01
-6.84574842e-01 7.68659592e-01 -5.92444062e-01 -5.53935826e-01
2.11374268e-01 4.64163035e-01 -7.33818710e-02 2.37774506e-01
-5.07641673e-01 9.88265038e-01 5.73496044e-01 3.89577746e-01
-3.41623962e-01 -2.53448874e-01 -3.74485880e-01 4.63100433e-01
5.26478887e-01 -3.30486923e-01 1.10428631e+00 -1.15723109e+00
-1.35272717e+00 7.23695457e-01 -1.32511824e-01 -3.91703188e-01
9.64609623e-01 -3.38709056e-01 -2.91433901e-01 2.28829101e-01
-1.03490919e-01 5.05734622e-01 1.03035748e+00 -1.45791316e+00
-8.29041660e-01 -3.30302298e-01 -2.30363384e-01 -1.69216320e-01
-7.16450930e-01 -1.90533996e-01 -4.95983481e-01 -5.07764161e-01
-1.81399986e-01 -5.71504653e-01 9.24545676e-02 1.38186991e-01
-5.59836149e-01 -2.33524188e-01 9.12327290e-01 -6.40318155e-01
1.20499694e+00 -2.07158542e+00 -6.91459626e-02 3.42685580e-01
-2.81315930e-02 3.73151064e-01 -5.92609011e-02 2.44557098e-01
7.66058043e-02 -4.09542881e-02 -1.47071583e-02 -6.17375433e-01
-2.38519143e-02 -2.01560527e-01 -2.14030743e-01 2.44115919e-01
2.86640704e-01 3.39968055e-01 -3.56949121e-01 -6.30625665e-01
4.35655504e-01 4.49124634e-01 -3.60493630e-01 1.93325877e-01
-3.59249145e-01 -4.93754931e-02 -3.69948000e-01 5.55993438e-01
7.14012802e-01 2.63912920e-02 2.62414068e-01 -3.92511487e-01
-3.44794869e-01 -1.64097592e-01 -1.62675714e+00 1.24945962e+00
-5.11634231e-01 9.48820055e-01 7.75371492e-02 -1.02810013e+00
1.42332840e+00 8.39338154e-02 3.38628411e-01 -7.18912125e-01
4.18003500e-01 4.09289449e-01 -1.10052429e-01 -5.07857323e-01
8.73600304e-01 4.68587786e-01 3.33931893e-01 3.20859075e-01
-5.94093949e-02 -4.67217602e-02 6.84455693e-01 -2.05126017e-01
6.48821592e-01 3.02944183e-01 1.31985933e-01 -1.77699491e-01
6.81023955e-01 -1.81407705e-02 1.67334557e-01 9.22927916e-01
2.17205778e-01 6.88240528e-01 6.85444713e-01 -5.56110442e-01
-1.15596592e+00 -4.85252440e-01 -2.04467028e-01 8.79065812e-01
-2.11624756e-01 -2.40244821e-01 -1.14591396e+00 -6.14976883e-01
-1.24668039e-01 8.59252274e-01 -5.94846368e-01 1.62874863e-01
-6.58906519e-01 -7.71210253e-01 1.84006944e-01 4.20063645e-01
1.35557309e-01 -1.10851455e+00 -1.03124630e+00 1.79723948e-01
2.64548987e-01 -9.81599569e-01 -6.70767501e-02 7.04216719e-01
-9.77092981e-01 -1.05153608e+00 -7.80058682e-01 -7.28505015e-01
9.86197114e-01 -8.13548937e-02 7.81313062e-01 1.53313592e-01
-6.10480189e-01 3.44646811e-01 -4.01041567e-01 -5.21563709e-01
-5.39640069e-01 3.68786186e-01 -1.21921062e-01 2.39023522e-01
3.01729441e-01 2.30969600e-02 -2.87445754e-01 2.70112067e-01
-8.60857487e-01 1.29929080e-03 7.61981487e-01 7.04752445e-01
4.18572932e-01 6.19029999e-03 2.42185026e-01 -8.99582624e-01
7.83390045e-01 2.40206942e-01 -9.25495982e-01 5.19073725e-01
-6.37907565e-01 4.52519432e-02 9.38591719e-01 -5.85571587e-01
-1.05491817e+00 3.33665401e-01 -8.10776949e-02 -1.53612331e-01
-5.82477093e-01 -3.77968028e-02 -2.08729953e-01 5.43285459e-02
8.26650620e-01 1.68099776e-01 -1.51468799e-01 -7.58563340e-01
1.23859510e-01 9.42335188e-01 1.66329861e-01 -4.83758807e-01
4.62181121e-01 1.78699896e-01 1.24641769e-01 -1.32431471e+00
-4.29584473e-01 -3.99446666e-01 -1.13905311e+00 -3.59904796e-01
6.52235150e-01 -2.35451028e-01 -7.43979633e-01 3.54602933e-01
-1.50102961e+00 -3.40995520e-01 -4.77071315e-01 3.33431572e-01
-5.21244526e-01 4.46061313e-01 -3.47646713e-01 -1.17747414e+00
-1.81320757e-01 -1.15854371e+00 1.25232637e+00 3.10971469e-01
-2.46478304e-01 -7.24317610e-01 -1.96950018e-01 2.72516031e-02
2.78872158e-02 -4.59377170e-02 8.24880242e-01 -8.11456442e-01
-7.08459437e-01 -5.72155237e-01 -2.19179213e-01 4.71853375e-01
-1.04408544e-02 4.92450118e-01 -1.21277130e+00 -7.03093633e-02
-1.04724467e-01 -5.00177965e-02 7.85159409e-01 4.96100575e-01
1.51391149e+00 8.61383379e-02 -5.60176194e-01 2.60715127e-01
1.33586884e+00 4.67758745e-01 4.37538266e-01 6.17162287e-01
5.00115514e-01 9.27142978e-01 8.38459194e-01 5.45867085e-01
-2.58955210e-01 8.98510695e-01 2.32701942e-01 -2.24156737e-01
-1.05379090e-01 -1.55172162e-02 -9.04295594e-03 3.10939968e-01
-1.01944298e-01 -6.18088186e-01 -7.85458326e-01 3.73554468e-01
-1.75010121e+00 -7.96768427e-01 -4.55089152e-01 2.35129809e+00
4.45323408e-01 5.36134124e-01 2.49697700e-01 5.29080629e-01
7.50481367e-01 -3.12938720e-01 -1.99368075e-01 -6.28872812e-01
-1.06647622e-03 2.90914886e-02 4.40485626e-01 4.90207493e-01
-1.15663528e+00 5.76908886e-01 5.87286758e+00 8.88360918e-01
-1.15881443e+00 -2.61792958e-01 6.34529352e-01 9.42367613e-02
1.80351347e-01 -3.67877930e-01 -1.35526443e+00 5.03187060e-01
6.54690325e-01 4.43950325e-01 3.35504189e-02 1.10077155e+00
3.87860179e-01 -5.53051114e-01 -1.22344518e+00 7.55339801e-01
3.41030717e-01 -1.16485167e+00 1.07179739e-01 -1.15338666e-03
3.41076404e-01 -7.15051055e-01 -2.28208929e-01 -2.41676830e-02
-5.92397094e-01 -6.71338141e-01 9.29108500e-01 8.59237731e-01
3.37429911e-01 -6.11249030e-01 4.57533538e-01 6.14214122e-01
-8.48947525e-01 -2.93627203e-01 -4.43904400e-01 1.66381136e-01
-1.15936901e-02 7.41568565e-01 -1.07255280e+00 4.28802758e-01
3.47699881e-01 2.74051249e-01 -7.97684133e-01 1.28685260e+00
-8.34421664e-02 2.30697021e-01 -4.13243622e-01 -5.25303245e-01
1.40486844e-02 -1.91236287e-01 3.07127982e-01 1.37623513e+00
5.17306268e-01 -5.82393706e-01 9.06001218e-03 1.13387597e+00
1.99054286e-01 4.18006122e-01 -4.08708870e-01 3.51119876e-01
2.71616012e-01 1.43320704e+00 -1.16474605e+00 -3.04155201e-01
-1.57327771e-01 1.08088410e+00 3.11957270e-01 2.07975179e-01
-6.79312706e-01 -6.88861907e-01 -3.43079358e-01 3.47950637e-01
5.15529156e-01 -1.42759159e-01 -4.51202899e-01 -8.02263618e-01
2.13585436e-01 -6.80259407e-01 3.38020623e-02 -7.79569864e-01
-7.93718338e-01 6.44798040e-01 -7.21524730e-02 -1.12105918e+00
-3.19609523e-01 -1.35876501e+00 -4.37081307e-01 7.91234374e-01
-1.23298573e+00 -1.15424168e+00 -4.14290041e-01 4.21157815e-02
7.22989678e-01 -3.26679409e-01 6.99133515e-01 3.09174806e-01
-8.72715175e-01 4.46986586e-01 3.14870209e-01 -4.32943068e-02
6.97949529e-01 -1.38639307e+00 -9.73429829e-02 9.03659642e-01
2.86180288e-01 4.33071882e-01 6.84620261e-01 -4.61572737e-01
-9.15970564e-01 -7.17376769e-01 8.06696415e-01 -2.83399671e-01
2.33840272e-01 -6.67329848e-01 -9.08520997e-01 3.61151665e-01
-9.58352238e-02 -3.56527478e-01 3.08658928e-01 1.60741910e-01
2.75839478e-01 -4.03893977e-01 -9.07798111e-01 4.50505286e-01
4.07746583e-01 -2.07958356e-01 -4.60753143e-01 4.10416633e-01
-5.33587113e-03 -1.07845612e-01 -5.05157769e-01 6.36430085e-02
7.14916825e-01 -1.24181163e+00 7.13051498e-01 -8.43523145e-02
2.27772802e-01 -2.70661533e-01 1.53038159e-01 -8.88925552e-01
-1.07901476e-01 -1.32795393e-01 -1.68992937e-01 1.35151911e+00
4.97033715e-01 -2.44969085e-01 9.75046694e-01 4.20971274e-01
-4.87165265e-02 -6.49504125e-01 -5.79851091e-01 -7.57582009e-01
-3.02210629e-01 -2.15019524e-01 1.99261054e-01 5.58492720e-01
-3.40805590e-01 1.61365569e-01 -2.49332294e-01 -1.99426198e-04
4.43066329e-01 1.99063882e-01 9.24549699e-01 -1.45082748e+00
-4.25065249e-01 -6.34031475e-01 -2.26890117e-01 -1.22201729e+00
-1.40530720e-01 -2.62182921e-01 5.23112230e-02 -1.32764757e+00
-3.08311582e-02 -2.87556142e-01 9.64829996e-02 2.21708089e-01
1.32657319e-01 2.02781204e-02 3.04051191e-01 -6.07056282e-02
-2.05871582e-01 1.47332802e-01 9.42237437e-01 -2.31672078e-01
-4.47994024e-01 3.46686721e-01 -2.20078491e-02 8.21574628e-01
7.46928334e-01 -2.55984306e-01 -1.66750729e-01 -3.10483187e-01
4.46708873e-02 -3.56461585e-01 2.60169357e-01 -1.29108369e+00
2.16313303e-01 1.79271057e-01 9.43585217e-01 -1.13972306e+00
3.73265892e-01 -1.06238627e+00 -3.58471334e-01 5.85673153e-01
-3.32828820e-01 -2.89811641e-01 2.11628437e-01 2.13429630e-01
1.90526334e-04 -1.20981276e+00 7.41394281e-01 -1.00182667e-01
-8.21965516e-01 -4.13769692e-01 -4.72541094e-01 -6.95614696e-01
1.07995069e+00 -7.33183801e-01 -1.22997858e-01 1.07738890e-01
-9.72533822e-01 -1.33829772e-01 3.12870979e-01 2.90869623e-01
3.39960396e-01 -7.75093317e-01 -2.84507304e-01 2.65649498e-01
3.42031680e-02 -7.77493864e-02 1.10367261e-01 6.86494648e-01
-8.20817471e-01 5.67152858e-01 -3.28569680e-01 -5.79216361e-01
-1.50126135e+00 8.00207675e-01 3.43359232e-01 -2.64178753e-01
-3.57181132e-01 5.16601861e-01 -4.20586839e-02 -9.68593657e-02
5.28873861e-01 -6.10581279e-01 -4.68740016e-01 4.88195658e-01
5.20238459e-01 6.69380665e-01 3.21434587e-01 -1.46945253e-01
-3.48376408e-02 7.15548575e-01 -8.28945935e-02 1.47195563e-01
1.13151991e+00 6.02929927e-02 1.32785887e-01 5.70550740e-01
7.75881708e-01 2.66322434e-01 -1.14962912e+00 3.02738607e-01
3.09053808e-01 -3.55854452e-01 -1.02316746e-02 -7.27049649e-01
-5.81837058e-01 1.14424944e+00 7.81863153e-01 6.17499232e-01
1.22268820e+00 -4.59766209e-01 4.59630182e-03 6.40076816e-01
4.37648855e-02 -1.59895873e+00 4.82090004e-02 1.80031329e-01
8.87622058e-01 -1.06966543e+00 1.50482044e-01 -3.93964797e-01
-2.69184530e-01 1.81896770e+00 7.59236395e-01 6.51782900e-02
6.11270815e-02 4.24687505e-01 -2.66037583e-01 -4.44438346e-02
-3.81000638e-01 -3.30596894e-01 3.80987853e-01 5.60989141e-01
4.20340478e-01 -9.56997797e-02 -5.15366554e-01 5.14555216e-01
6.62143603e-02 8.43993500e-02 6.69330120e-01 8.15934300e-01
-7.48721719e-01 -1.24833953e+00 -7.68251657e-01 5.05538285e-01
-1.14063121e-01 2.23437756e-01 -5.88329375e-01 1.04108000e+00
3.55774075e-01 5.61675012e-01 2.10618109e-01 3.81239541e-02
6.10795498e-01 1.05718397e-01 5.83143950e-01 -4.98240352e-01
-3.36819232e-01 2.57066905e-01 8.91285613e-02 -1.11109652e-01
-1.71594322e-01 -6.72279954e-01 -7.54496098e-01 1.92537889e-01
-7.04434812e-01 -2.79233009e-02 1.05047429e+00 7.82192290e-01
9.43331793e-02 7.71313131e-01 4.75983232e-01 -1.00036168e+00
-6.27130568e-01 -1.07144451e+00 -8.48154306e-01 9.65167955e-02
1.35806441e-01 -5.15564084e-01 -4.55735512e-02 2.28218868e-01] | [11.770841598510742, 2.6366260051727295] |
2a0fe972-e9af-4fa6-afae-5f4d21bd9606 | deep-seam-prediction-for-image-stitching | 2302.05027 | null | https://arxiv.org/abs/2302.05027v2 | https://arxiv.org/pdf/2302.05027v2.pdf | Deep Seam Prediction for Image Stitching Based on Selection Consistency Loss | Image stitching is to construct panoramic images with wider field of vision (FOV) from some images captured from different viewing positions. To solve the problem of fusion ghosting in the stitched image, seam-driven methods avoid the misalignment area to fuse images by predicting the best seam. Currently, as standard tools of the OpenCV library, dynamic programming (DP) and GraphCut (GC) are still the only commonly used seam prediction methods despite the fact that they were both proposed two decades ago. However, GC can get excellent seam quality but poor real-time performance while DP method has good efficiency but poor seam quality. In this paper, we propose a deep learning based seam prediction method (DSeam) for the sake of high seam quality with high efficiency. To overcome the difficulty of the seam description in network and no GroundTruth for training we design a selective consistency loss combining the seam shape constraint and seam quality constraint to supervise the network learning. By the constraint of the selection of consistency loss, we implicitly defined the mask boundaries as seams and transform seam prediction into mask prediction. To our knowledge, the proposed DSeam is the first deep learning based seam prediction method for image stitching. Extensive experimental results well demonstrate the superior performance of our proposed Dseam method which is 15 times faster than the classic GC seam prediction method in OpenCV 2.4.9 with similar seam quality. | ['Wenbing Tao', 'Nanjun Yuan', 'Zhi Chen', 'Fan Yang', 'Senmao Cheng'] | 2023-02-10 | null | null | null | null | ['image-stitching'] | ['computer-vision'] | [ 2.56448984e-01 -1.53397113e-01 -4.27330285e-02 -3.36724311e-01
-3.00918728e-01 -1.07169293e-01 4.36456293e-01 -5.23172140e-01
-3.95763844e-01 3.92726600e-01 -1.14784002e-01 -1.26932755e-01
-2.12941527e-01 -6.29624069e-01 -8.90471876e-01 -7.76880383e-01
4.15389329e-01 -3.31902206e-02 6.94794238e-01 -2.56135494e-01
4.80043709e-01 1.64626166e-01 -1.59465611e+00 3.30599807e-02
1.36946988e+00 1.04596829e+00 5.85898161e-01 1.86757907e-01
-2.04889134e-01 5.37493050e-01 -4.20857519e-01 -5.22468328e-01
7.74924099e-01 -3.56495023e-01 -5.66562116e-01 1.96609095e-01
8.37464452e-01 -5.14788210e-01 -3.32812130e-01 1.26533902e+00
2.45602682e-01 3.00773174e-01 1.51663706e-01 -1.38889706e+00
-5.34782290e-01 3.95998597e-01 -1.19225466e+00 -2.92647798e-02
2.00725012e-02 5.66973984e-02 5.79424918e-01 -7.73150742e-01
4.84736592e-01 1.07689762e+00 7.40732849e-01 4.29492801e-01
-8.84159505e-01 -9.22809362e-01 6.39695153e-02 2.14927331e-01
-1.39237571e+00 -2.18518302e-01 9.71205294e-01 -2.24474102e-01
4.58183438e-01 3.79270256e-01 6.11256242e-01 6.93482101e-01
5.23376048e-01 5.96027374e-01 1.07643485e+00 -2.99541652e-01
-7.10980818e-02 -1.01190031e-01 1.79838836e-02 8.68867755e-01
-6.39303625e-02 4.20536846e-01 -4.84298170e-01 1.18565977e-01
1.12982750e+00 1.79729968e-01 -5.33786714e-01 -3.67286384e-01
-1.32528102e+00 6.90855801e-01 5.84170818e-01 1.47720993e-01
1.04525194e-01 2.99590915e-01 2.61145294e-01 2.94049650e-01
2.65023291e-01 2.28647757e-02 -2.50152260e-01 2.49921679e-01
-1.10982239e+00 1.40513644e-01 3.64611804e-01 1.04428732e+00
9.45574760e-01 1.71662331e-01 2.13796005e-01 9.98686671e-01
2.78273165e-01 6.98441342e-02 5.32349288e-01 -1.13318455e+00
3.09942275e-01 6.37748480e-01 -1.04237691e-01 -1.20430851e+00
-1.71355367e-01 -2.89391339e-01 -9.36947405e-01 6.57120287e-01
1.89435840e-01 -3.04067638e-02 -1.16592681e+00 1.42729318e+00
4.45800036e-01 5.56605577e-01 -1.51731029e-01 1.27285779e+00
8.86612177e-01 8.11874330e-01 -2.93431461e-01 -1.77899241e-01
1.05008650e+00 -1.41568184e+00 -8.07137489e-01 -2.42611349e-01
3.92925799e-01 -1.17722631e+00 8.38473856e-01 6.94523275e-01
-9.57152665e-01 -9.16133702e-01 -1.36639988e+00 -3.22226174e-02
-6.37803599e-02 -3.10518667e-02 8.79689395e-01 5.54782867e-01
-1.00162637e+00 6.68694973e-01 -6.75491095e-01 -1.12693705e-01
2.00917587e-01 4.04889375e-01 -3.09290767e-01 1.27598923e-02
-1.00710702e+00 8.43266368e-01 6.41700566e-01 2.13034883e-01
-6.78097546e-01 -6.74223065e-01 -7.93690622e-01 -2.57976919e-01
5.34377098e-01 -6.62827611e-01 9.47867215e-01 -1.40914690e+00
-1.53619254e+00 8.11359286e-01 3.81747067e-01 -3.91641796e-01
6.19638681e-01 -3.40859085e-01 -4.95132923e-01 -6.06638230e-02
-7.22281709e-02 9.03712630e-01 1.03959012e+00 -1.44355452e+00
-8.20508778e-01 -1.83171406e-01 -1.23866931e-01 3.87612820e-01
-2.42546692e-01 -2.37625629e-01 -7.38802254e-01 -8.26062918e-01
2.91309536e-01 -7.93908060e-01 -3.96876305e-01 3.37625891e-01
-3.91386807e-01 3.08731222e-03 1.17708445e+00 -6.89627469e-01
1.14215004e+00 -2.24334264e+00 7.83314407e-02 1.30497903e-01
1.25699848e-01 3.70389134e-01 5.65843098e-03 1.98851153e-01
-4.72597666e-02 -3.23206872e-01 -5.36576152e-01 -4.25909430e-01
-5.20371556e-01 2.98517793e-01 -2.94787467e-01 4.44420069e-01
-3.48147273e-01 5.21844506e-01 -6.10637724e-01 -8.54636550e-01
6.90428972e-01 3.29233170e-01 -3.76437277e-01 3.17553252e-01
-1.76756918e-01 4.41478819e-01 -1.16274722e-01 4.83651638e-01
1.32951975e+00 1.26667619e-01 -3.15031320e-01 -7.71957099e-01
-3.91498953e-01 -4.83012140e-01 -1.33685315e+00 2.34694695e+00
-3.77452761e-01 4.15823042e-01 2.25271210e-01 -7.49815166e-01
1.19301891e+00 4.72951271e-02 4.12672281e-01 -6.82357311e-01
2.67315149e-01 2.07325712e-01 -2.35496894e-01 -6.15271866e-01
7.02442408e-01 -4.58968654e-02 2.69615710e-01 2.29338437e-01
1.03441097e-01 -2.00268030e-01 -1.63283587e-01 9.06409398e-02
4.41210121e-01 3.72250229e-01 -1.24078512e-01 -3.39168012e-01
5.95908523e-01 -3.24486941e-02 8.41771007e-01 4.16201741e-01
-3.95229608e-02 1.13964939e+00 -1.31043335e-02 -6.81406558e-01
-1.05458939e+00 -9.77336407e-01 -2.11487770e-01 7.03424096e-01
9.46543753e-01 -1.66280791e-01 -7.98240066e-01 -5.54432094e-01
-3.62985700e-01 5.90717912e-01 -4.59024370e-01 -2.05789357e-01
-7.39359319e-01 -4.78916675e-01 3.00069571e-01 1.83620676e-01
1.14409077e+00 -8.28901052e-01 -4.96411920e-01 -1.37504399e-01
6.99483156e-02 -9.45818365e-01 -1.02311969e+00 -1.40748844e-01
-8.97902310e-01 -1.17093563e+00 -7.29280889e-01 -1.20561576e+00
6.50357425e-01 7.30657935e-01 6.44037545e-01 5.11810660e-01
-3.80867839e-01 -7.40641654e-02 -1.84912771e-01 1.54791683e-01
-2.90091783e-01 -3.01108658e-01 -2.79832840e-01 1.89158283e-02
-1.48757979e-01 -6.58725262e-01 -1.04737115e+00 5.02367914e-01
-1.33761775e+00 7.57451892e-01 5.47541976e-01 7.76779473e-01
4.85318840e-01 3.03477108e-01 1.18744105e-01 -5.96158624e-01
2.84069449e-01 -5.98717071e-02 -9.09835160e-01 3.22804064e-01
-9.74722624e-01 -1.04027331e-01 4.87484902e-01 -3.45916241e-01
-1.44753325e+00 1.21537857e-01 -1.70073405e-01 -7.11309075e-01
5.78897744e-02 2.63120204e-01 -8.70073959e-02 -5.73080838e-01
3.10433894e-01 3.56859744e-01 3.67382199e-01 -3.97505820e-01
3.71014595e-01 4.24233079e-01 8.61585855e-01 -2.61464924e-01
7.90023148e-01 5.04620910e-01 -4.66253608e-02 -3.84956211e-01
-5.90877235e-01 -3.05135041e-01 -3.77341926e-01 -4.05235171e-01
1.13074839e+00 -7.94559300e-01 -5.61253250e-01 6.71039104e-01
-1.15646279e+00 -4.38423231e-02 2.02632084e-01 3.41622442e-01
-4.30002600e-01 9.64503825e-01 -5.07273376e-01 -4.86324817e-01
-6.75983608e-01 -1.31198239e+00 8.47485900e-01 5.67513347e-01
2.90676892e-01 -9.15263534e-01 8.24836940e-02 4.40069854e-01
2.72331804e-01 4.34741735e-01 4.91194248e-01 4.89168353e-02
-7.84848988e-01 1.79428279e-01 -3.03653359e-01 5.72146773e-01
3.48914683e-01 9.41998661e-02 -8.28488290e-01 -1.58037871e-01
3.33487064e-01 -4.75678360e-03 9.79856193e-01 4.70513135e-01
1.30586016e+00 -1.88991994e-01 -3.30790251e-01 1.12372649e+00
1.88532424e+00 5.91490865e-01 7.91423023e-01 6.50447309e-01
9.45340991e-01 5.92271030e-01 8.11860085e-01 1.26706719e-01
-3.16382051e-02 7.78663039e-01 7.48802423e-01 -4.07977402e-01
-1.46092623e-01 -2.78884411e-01 1.75574467e-01 5.38161695e-01
7.33897388e-02 -9.08323601e-02 -4.30189759e-01 3.00122261e-01
-2.06174874e+00 -7.40088820e-01 -3.38129133e-01 2.38707829e+00
8.03684890e-01 2.62362450e-01 -1.65814400e-01 -9.45316702e-02
9.99880850e-01 3.70362371e-01 -2.50774831e-01 -3.35792333e-01
-1.20051049e-01 -1.21198125e-01 6.08894706e-01 5.95262408e-01
-1.00888228e+00 9.10628974e-01 5.23569345e+00 1.31845760e+00
-1.01193106e+00 -1.69190634e-02 4.83470619e-01 1.34903640e-01
-2.76863128e-01 3.75776857e-01 -5.17058909e-01 7.28203833e-01
1.08639158e-01 1.99535653e-01 4.11038250e-01 7.31933892e-01
9.11810342e-03 -3.61495674e-01 -7.55431831e-01 1.18213820e+00
1.68046415e-01 -1.52231407e+00 7.79398307e-02 -2.10439652e-01
8.96484792e-01 -2.33921707e-01 8.95063579e-02 -3.21723074e-01
1.74498018e-02 -9.37256694e-01 6.49287403e-01 5.34683645e-01
4.96353149e-01 -1.04216492e+00 6.64486229e-01 2.49832213e-01
-1.09212244e+00 1.12068228e-01 -5.04275799e-01 3.43588471e-01
3.34252924e-01 6.66651428e-01 -2.12004527e-01 8.59642148e-01
8.12555969e-01 7.80021250e-01 -4.62472647e-01 1.29196453e+00
1.76172927e-01 1.85134485e-01 -2.17894122e-01 5.87023795e-01
3.75530273e-01 -6.90792441e-01 4.70845282e-01 9.16284621e-01
4.67444122e-01 -6.24830835e-02 1.05284132e-01 9.15632606e-01
2.32662648e-01 -9.84477773e-02 -4.88649786e-01 5.57197213e-01
4.57655907e-01 1.30195391e+00 -8.47160995e-01 -1.47380158e-01
-4.61855978e-01 1.23655081e+00 -1.46078974e-01 1.11394674e-01
-1.01015186e+00 -3.72531325e-01 4.20649707e-01 3.12055070e-02
1.66430190e-01 -1.52767777e-01 -3.91242206e-01 -9.41596329e-01
-2.47693118e-02 -7.63281822e-01 3.32035571e-01 -9.12803710e-01
-1.14510036e+00 6.17643595e-01 5.68720177e-02 -1.63591254e+00
2.90789157e-01 -3.12191725e-01 -9.96194363e-01 6.00034714e-01
-1.32766116e+00 -1.28752232e+00 -7.05581009e-01 7.42851734e-01
8.34558189e-01 -1.44270793e-01 3.06660891e-01 2.93897659e-01
-6.06165051e-01 4.61413354e-01 9.28458571e-02 -1.29748121e-01
8.26814651e-01 -8.98153245e-01 2.69771457e-01 9.19654071e-01
-9.14911702e-02 5.08601725e-01 7.82532692e-01 -7.03838646e-01
-1.05489588e+00 -9.68373895e-01 1.46739960e-01 1.61067963e-01
3.25713694e-01 -4.15682383e-02 -1.17747450e+00 4.84670073e-01
7.36187458e-01 -1.30586192e-01 1.90275922e-01 -5.75683892e-01
-1.81749851e-01 -4.09105182e-01 -9.62222576e-01 7.31527388e-01
1.05703402e+00 1.81223545e-02 -4.96568978e-01 1.60168663e-01
1.06947768e+00 -7.09156513e-01 -8.33178699e-01 5.97188056e-01
4.51053709e-01 -1.60546529e+00 1.07346344e+00 3.76656592e-01
4.63959903e-01 -6.47400141e-01 2.23681897e-01 -9.89230037e-01
-1.22153759e-01 -7.50536501e-01 2.64393866e-01 1.45556259e+00
-3.98605019e-02 -4.73027140e-01 7.47598469e-01 4.28084582e-01
-5.75155258e-01 -9.00830388e-01 -9.47686553e-01 -7.59193063e-01
-1.96892887e-01 -6.64750040e-02 6.65423453e-01 1.07280362e+00
-4.36989069e-01 -1.12258859e-01 -7.95704722e-01 5.38711436e-02
8.84323776e-01 3.93716812e-01 7.78553486e-01 -9.88195181e-01
-3.58174115e-01 -4.47244197e-01 -2.99677908e-01 -1.02286315e+00
-7.31134638e-02 -5.21860421e-01 1.32721424e-01 -1.42284966e+00
1.00517116e-01 -5.77454686e-01 -3.11973859e-02 1.59845740e-01
-2.33836547e-01 2.28644326e-01 8.56263638e-02 3.52987081e-01
-3.03245962e-01 6.08575106e-01 1.54059684e+00 -2.85941083e-02
-3.18623662e-01 -1.36171654e-01 -3.75600666e-01 7.63694406e-01
7.99526334e-01 -3.51501703e-01 -6.89719260e-01 -7.54486501e-01
1.00129703e-02 2.24096730e-01 4.68188912e-01 -1.17788279e+00
5.29938877e-01 -4.25648719e-01 4.04043436e-01 -8.09539676e-01
3.03597569e-01 -9.96990561e-01 6.34680688e-01 4.32000637e-01
-1.33926228e-01 4.62274486e-03 1.10159338e-01 5.88502467e-01
-3.30929369e-01 -3.05303067e-01 9.03993726e-01 -1.40874177e-01
-1.13322222e+00 4.72764581e-01 1.22522868e-01 -4.00028914e-01
1.15748942e+00 -7.56664693e-01 -3.27147126e-01 -6.69070706e-02
-1.99446455e-01 2.94254184e-01 7.29995489e-01 6.12369835e-01
1.10574925e+00 -1.33732915e+00 -3.20476025e-01 3.33383024e-01
-5.88519312e-02 3.74583602e-01 5.13029635e-01 7.24654973e-01
-8.17950428e-01 -1.36390358e-01 -4.78146166e-01 -6.19723260e-01
-1.27315998e+00 5.73711455e-01 4.39179480e-01 8.95927697e-02
-9.66866314e-01 9.32150483e-01 4.68338221e-01 -2.45568842e-01
3.33838016e-01 3.51079367e-02 -7.60132819e-02 -4.86189783e-01
3.48664492e-01 3.78634214e-01 1.79721005e-02 -4.33629394e-01
-8.54805484e-02 1.03290153e+00 -2.70709872e-01 -8.60513747e-02
1.05710351e+00 -3.03030461e-01 -3.95568997e-01 -4.15751413e-02
1.23229790e+00 -2.08834156e-01 -1.58426237e+00 1.54341257e-03
-3.05117756e-01 -8.15690517e-01 3.02325428e-01 -3.66954088e-01
-1.72304308e+00 6.16489172e-01 9.60684538e-01 -1.83194801e-01
1.31094837e+00 -2.87871569e-01 1.30818188e+00 -1.76404655e-01
3.48168939e-01 -1.22196341e+00 2.74688423e-01 -1.03865296e-01
8.36412728e-01 -1.15101612e+00 1.84456125e-01 -6.34571493e-01
-7.10458159e-01 1.29166305e+00 1.27164054e+00 -3.88465941e-01
5.37392378e-01 2.62603223e-01 1.45175144e-01 -1.90983340e-01
-3.39723766e-01 2.06010118e-01 2.13669971e-01 5.04873335e-01
6.24239370e-02 -4.22252417e-01 -5.89880347e-01 1.07256055e-01
-1.36712730e-01 -3.39514911e-02 4.19784129e-01 8.52778018e-01
-5.81450641e-01 -1.06772959e+00 -5.08137941e-01 2.22839668e-01
-1.52857214e-01 6.49637952e-02 1.77601855e-02 6.85255826e-01
4.92129713e-01 9.24852014e-01 1.46550968e-01 -7.65156806e-01
3.26804370e-02 -4.60781217e-01 5.00185609e-01 -1.04473919e-01
-6.40745997e-01 1.80556238e-01 -2.78973609e-01 -7.27846622e-01
-5.68869889e-01 -5.92703186e-02 -1.33590126e+00 -3.70390117e-01
-6.02919579e-01 5.43826669e-02 5.73284984e-01 8.75344098e-01
1.09477341e-01 1.58842653e-01 7.37083256e-01 -7.04412103e-01
-2.05876634e-01 -5.56872129e-01 -5.87769926e-01 3.25959861e-01
3.08779061e-01 -5.40008426e-01 -4.05261666e-01 1.64378911e-01] | [9.480511665344238, -2.2328014373779297] |
b0ae44a0-0085-4650-9e38-21e0b18254cc | ray3d-ray-based-3d-human-pose-estimation-for | 2203.11471 | null | https://arxiv.org/abs/2203.11471v3 | https://arxiv.org/pdf/2203.11471v3.pdf | Ray3D: ray-based 3D human pose estimation for monocular absolute 3D localization | In this paper, we propose a novel monocular ray-based 3D (Ray3D) absolute human pose estimation with calibrated camera. Accurate and generalizable absolute 3D human pose estimation from monocular 2D pose input is an ill-posed problem. To address this challenge, we convert the input from pixel space to 3D normalized rays. This conversion makes our approach robust to camera intrinsic parameter changes. To deal with the in-the-wild camera extrinsic parameter variations, Ray3D explicitly takes the camera extrinsic parameters as an input and jointly models the distribution between the 3D pose rays and camera extrinsic parameters. This novel network design is the key to the outstanding generalizability of Ray3D approach. To have a comprehensive understanding of how the camera intrinsic and extrinsic parameter variations affect the accuracy of absolute 3D key-point localization, we conduct in-depth systematic experiments on three single person 3D benchmarks as well as one synthetic benchmark. These experiments demonstrate that our method significantly outperforms existing state-of-the-art models. Our code and the synthetic dataset are available at https://github.com/YxZhxn/Ray3D . | ['Wongun Choi', 'Renliang Weng', 'Fenghai Li', 'Yu Zhan'] | 2022-03-22 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhan_Ray3D_Ray-Based_3D_Human_Pose_Estimation_for_Monocular_Absolute_3D_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhan_Ray3D_Ray-Based_3D_Human_Pose_Estimation_for_Monocular_Absolute_3D_CVPR_2022_paper.pdf | cvpr-2022-1 | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [-2.57211089e-01 -3.94003034e-01 -1.60132304e-01 -3.96647364e-01
-6.44259751e-01 -4.64439690e-01 3.84414226e-01 -5.55873036e-01
-5.40032744e-01 5.14546454e-01 3.26709360e-01 2.00447872e-01
2.84767151e-01 -3.09973687e-01 -7.72089183e-01 -3.10244352e-01
2.31350854e-01 4.85655308e-01 2.93697342e-02 -2.04272464e-01
-1.39642265e-02 4.27941889e-01 -1.05420280e+00 -4.38444287e-01
3.18388522e-01 8.12160969e-01 -2.17627764e-01 8.26534331e-01
5.49169004e-01 2.60760605e-01 -6.36745870e-01 -3.49403471e-01
7.67933369e-01 -2.04280913e-01 -3.50926220e-01 1.98199928e-01
9.06259060e-01 -6.75947785e-01 -7.53709972e-01 1.00298572e+00
1.03125679e+00 -5.82538284e-02 5.33564329e-01 -1.39125717e+00
-5.48937261e-01 -1.64620895e-02 -7.80062258e-01 -7.37743005e-02
1.03209519e+00 4.87318009e-01 5.79966247e-01 -1.11854577e+00
6.07736707e-01 1.42172873e+00 9.46491063e-01 6.95267141e-01
-8.32781374e-01 -6.09584033e-01 1.75412640e-01 -1.05878912e-01
-1.67873776e+00 -2.31784105e-01 8.48072529e-01 -3.47296953e-01
5.77398837e-01 1.66361943e-01 9.63262498e-01 1.65496731e+00
1.80659309e-01 7.33909845e-01 9.43462193e-01 -4.35380459e-01
-1.10929780e-01 -1.80907786e-01 2.64907293e-02 7.18713164e-01
5.68788290e-01 3.37043643e-01 -8.22369218e-01 -8.80326182e-02
1.36869454e+00 2.02753484e-01 -4.50140387e-01 -8.41458738e-01
-1.54679763e+00 3.74436498e-01 6.62346303e-01 -3.69324505e-01
-4.82892878e-02 7.04174042e-01 4.79303636e-02 -4.49083820e-02
2.41512850e-01 2.56706208e-01 -4.97320920e-01 -3.14883411e-01
-3.79020065e-01 5.86704314e-01 6.18360877e-01 1.42435920e+00
3.96552324e-01 -2.83912212e-01 7.53844390e-03 5.43753088e-01
3.98383886e-01 1.06060290e+00 -1.25782983e-02 -1.30068171e+00
6.51386976e-01 5.17958164e-01 3.59922886e-01 -1.09965909e+00
-5.46594918e-01 -3.63335937e-01 -6.62660718e-01 2.67591551e-02
8.07159185e-01 -2.34382883e-01 -8.42999697e-01 1.83310890e+00
5.09599447e-01 -3.05671245e-01 -4.02117312e-01 1.44171429e+00
9.08482432e-01 2.66082972e-01 -3.63815635e-01 3.73791933e-01
1.37454998e+00 -9.17335689e-01 -4.32301968e-01 -6.05926752e-01
3.31112891e-01 -8.45318019e-01 1.14011395e+00 2.08564073e-01
-1.02154148e+00 -6.08241379e-01 -9.87854660e-01 -2.99645901e-01
-9.13382247e-02 4.75400239e-01 4.07313168e-01 7.18213379e-01
-7.69700050e-01 -2.90255789e-02 -8.84057760e-01 -7.90946424e-01
2.32561566e-02 3.97563159e-01 -5.96345365e-01 -1.40622675e-01
-1.11547267e+00 9.07564700e-01 3.11125326e-03 2.34028190e-01
-5.13346791e-01 -5.31987429e-01 -8.96410823e-01 -4.46542114e-01
7.28650630e-01 -1.22633374e+00 1.31271100e+00 -1.97620615e-01
-1.56954563e+00 1.01347148e+00 3.98162240e-03 2.99037620e-03
8.84317696e-01 -8.72973204e-01 8.58656764e-02 1.32618561e-01
9.24443528e-02 7.87337482e-01 5.99732161e-01 -1.30745888e+00
-2.00733349e-01 -6.93860054e-01 1.65911749e-01 6.24088168e-01
2.68037524e-02 -2.82369912e-01 -1.28249204e+00 -6.59930110e-01
3.50523800e-01 -1.38568759e+00 -2.13010177e-01 4.60657239e-01
-7.58904159e-01 1.76266819e-01 2.85108924e-01 -4.35566425e-01
6.87907219e-01 -1.92153227e+00 2.58625299e-01 2.58286804e-01
4.16081846e-01 -2.28477672e-01 2.41556272e-01 1.14901088e-01
1.60273015e-01 -4.27672446e-01 1.88077092e-01 -6.33507013e-01
1.77564174e-01 -1.49967507e-01 2.09384352e-01 8.82616699e-01
-3.01447213e-01 1.00772154e+00 -6.30711675e-01 -4.46435899e-01
2.95600414e-01 9.31576550e-01 -5.77419460e-01 1.90037161e-01
2.03485310e-01 6.79737628e-01 -4.22296613e-01 8.74093413e-01
5.74292719e-01 -3.19666207e-01 -2.12008551e-01 -5.63719928e-01
1.54462889e-01 -1.41177744e-01 -1.31262553e+00 2.06434321e+00
1.04064167e-01 6.16117716e-01 -1.99409187e-01 -1.72108203e-01
8.96486223e-01 4.85665761e-02 5.02407134e-01 -4.92213190e-01
5.73107064e-01 9.56405029e-02 -4.64609653e-01 -1.16113357e-01
6.96587801e-01 4.29674536e-01 -3.38328719e-01 2.95072198e-01
-1.33500308e-01 -1.54789329e-01 -3.11861753e-01 1.37408316e-01
8.10713053e-01 5.30051708e-01 3.97367269e-01 -1.19197309e-01
2.15350270e-01 -1.96038455e-01 5.64207673e-01 7.31295109e-01
-5.15421748e-01 1.15881240e+00 3.14876437e-01 -5.76479316e-01
-1.19290864e+00 -1.27681434e+00 1.68306381e-01 4.93650466e-01
6.55408978e-01 -5.85874856e-01 -8.81742001e-01 -4.68810648e-01
2.50286341e-01 -9.64134559e-02 -6.51435912e-01 -1.26164302e-01
-7.16643751e-01 -5.10707319e-01 6.20036483e-01 7.17295289e-01
8.31928074e-01 -2.07084090e-01 -9.35204864e-01 -4.49150920e-01
-4.87017214e-01 -1.41482425e+00 -1.01525366e+00 -1.89250886e-01
-4.41225678e-01 -1.18923068e+00 -1.11627305e+00 -4.03835624e-01
8.55913341e-01 4.82774585e-01 9.56351459e-01 -1.90431952e-01
-4.42253232e-01 9.36638594e-01 -1.91490039e-01 -3.08520377e-01
2.80794680e-01 1.59549683e-01 5.64651489e-01 -2.94564337e-01
5.71565270e-01 -2.91521579e-01 -1.05920625e+00 6.78508937e-01
-7.29138404e-02 7.85086751e-02 4.07728940e-01 5.14231741e-01
6.00645661e-01 -4.44783866e-01 -3.26798767e-01 -5.22112310e-01
4.25523669e-01 5.62717803e-02 -7.25310445e-01 -5.84186018e-02
-2.80788481e-01 -2.15895399e-01 -1.52794831e-02 -4.05322701e-01
-9.27475393e-01 4.24975097e-01 1.01228423e-01 -4.44173604e-01
-1.45842761e-01 8.70204195e-02 -2.92660236e-01 -2.41309792e-01
1.00042427e+00 -8.43477771e-02 8.77031311e-02 -4.00532156e-01
4.05404747e-01 5.68248093e-01 1.04995918e+00 -7.59450078e-01
1.06335270e+00 6.89375162e-01 2.01531336e-01 -6.67055726e-01
-8.50764513e-01 -7.41033077e-01 -1.06741416e+00 -4.23222095e-01
1.07455707e+00 -1.46760821e+00 -9.12348807e-01 7.80137718e-01
-1.27543736e+00 -3.24710935e-01 4.29183505e-02 7.19100952e-01
-7.44196117e-01 3.69132757e-01 -5.20968974e-01 -6.51522160e-01
-1.94462806e-01 -1.39391434e+00 1.45339549e+00 8.13367069e-02
-4.74036336e-01 -6.17255509e-01 1.33037761e-01 3.07871163e-01
-2.56584305e-02 5.55819869e-01 6.23207502e-02 1.28718019e-01
-7.30716527e-01 -4.50656056e-01 -7.60868713e-02 6.47744164e-02
1.41982093e-01 -3.26893270e-01 -7.69157529e-01 -4.54633951e-01
-3.75225656e-02 -2.32568786e-01 4.34801787e-01 5.93956232e-01
6.56356394e-01 1.66498810e-01 -3.66569191e-01 1.19218969e+00
1.09571946e+00 -4.20872122e-01 5.24840534e-01 6.31909668e-01
1.21392214e+00 2.83532619e-01 6.55164659e-01 5.20166814e-01
7.38650680e-01 1.12475193e+00 2.22384214e-01 -7.35869184e-02
-1.55737668e-01 -5.61265469e-01 3.46551359e-01 4.50465322e-01
-3.65024626e-01 -1.60617948e-01 -9.64427769e-01 5.72810993e-02
-1.73002803e+00 -7.21218050e-01 -8.38870853e-02 2.19998527e+00
6.53951824e-01 1.97532073e-01 4.07020092e-01 -4.36339155e-03
7.12238610e-01 1.43934205e-01 -5.92680573e-01 4.88481760e-01
-1.13955982e-01 -3.53412956e-01 9.38542247e-01 4.99685973e-01
-1.10822582e+00 8.71336222e-01 6.06633806e+00 1.32639125e-01
-8.25456917e-01 -1.43212855e-01 2.29599297e-01 -4.59150225e-01
2.44452119e-01 -1.70580044e-01 -1.25548315e+00 3.25505376e-01
2.63095528e-01 2.27447703e-01 2.43611544e-01 8.23021472e-01
-1.32061271e-02 -1.95897460e-01 -1.31423962e+00 1.81268799e+00
3.59616429e-01 -9.09493685e-01 -1.55696318e-01 2.27598488e-01
6.47826970e-01 5.08395303e-03 2.17843592e-01 -2.20798001e-01
-4.06660046e-03 -7.97781885e-01 9.53561664e-01 6.00589514e-01
9.96771097e-01 -6.27687991e-01 5.07641077e-01 2.85201967e-01
-1.24792063e+00 1.56403333e-01 -2.76776969e-01 -5.72743751e-02
2.81113803e-01 3.14876258e-01 -5.25313437e-01 2.86555529e-01
1.13938081e+00 7.56197810e-01 -7.94989347e-01 9.32937205e-01
-3.69435012e-01 3.87854199e-03 -6.03707552e-01 1.46524310e-01
-2.16051430e-01 2.97623239e-02 7.27390587e-01 9.38120723e-01
2.50889570e-01 2.54697800e-01 1.58985302e-01 8.31194043e-01
-1.06329784e-01 -2.81577229e-01 -4.96238559e-01 4.70104873e-01
5.14529705e-01 8.95688713e-01 -4.87780303e-01 6.20412081e-02
-2.95282006e-01 1.25981474e+00 3.56089957e-02 4.44764256e-01
-9.00896668e-01 -1.77645832e-01 7.83422172e-01 2.82769293e-01
4.60226834e-02 -7.77523518e-01 -2.69560367e-01 -1.45595825e+00
3.86738330e-01 -9.66163635e-01 1.66468024e-01 -9.03879344e-01
-1.10495794e+00 3.75446945e-01 3.57523054e-01 -1.53904450e+00
-1.63933292e-01 -8.34399104e-01 -9.02353674e-02 7.73652852e-01
-1.10056055e+00 -1.29539311e+00 -9.45204616e-01 9.21063304e-01
2.22229630e-01 -1.72840148e-01 4.27898675e-01 2.72054285e-01
-3.82354647e-01 9.80321586e-01 -3.31531495e-01 4.24006581e-01
1.22915137e+00 -1.19363570e+00 9.94156897e-01 9.27233577e-01
-1.67312309e-01 9.36531544e-01 7.52373934e-01 -6.89997435e-01
-1.71158397e+00 -7.59625912e-01 4.85979795e-01 -1.37496686e+00
7.46087506e-02 -7.34732151e-01 -1.78980693e-01 8.91078472e-01
-3.12194228e-01 1.58665136e-01 3.44068617e-01 -9.24270451e-02
-3.68252665e-01 -2.30619628e-02 -9.02092755e-01 8.05954397e-01
1.63982046e+00 -4.68349725e-01 -3.88546884e-01 2.93211460e-01
7.42898464e-01 -1.03894997e+00 -7.81317472e-01 1.99760213e-01
7.78484344e-01 -9.83383596e-01 1.46930850e+00 2.91075483e-02
4.82361391e-02 -5.71958661e-01 -4.49415267e-01 -1.06042981e+00
-2.19793200e-01 -8.31524968e-01 -1.98106378e-01 6.64255738e-01
-2.82875565e-03 -4.83487785e-01 1.06013906e+00 7.99403548e-01
2.70776242e-01 -5.54283202e-01 -7.53713667e-01 -8.28446805e-01
-2.47823894e-01 -3.74651432e-01 6.52449906e-01 4.29613978e-01
-2.99763560e-01 2.70274013e-01 -8.22175026e-01 3.26606035e-01
1.11711228e+00 -3.14478576e-01 1.65061271e+00 -1.04315186e+00
-2.66095102e-01 -6.46060109e-02 -6.44397736e-01 -1.74045014e+00
-2.43162051e-01 -2.65348494e-01 6.37533590e-02 -1.10640585e+00
2.73000717e-01 -2.22981870e-01 3.76833886e-01 2.15429068e-02
-3.07754636e-01 4.51237053e-01 3.01558405e-01 3.39033216e-01
-5.99128067e-01 6.11517847e-01 1.29379702e+00 1.74040973e-01
-1.58413187e-01 1.13033131e-01 -6.37139797e-01 9.27977324e-01
6.18880272e-01 -2.92276740e-01 -4.36187327e-01 -7.57175326e-01
3.10231656e-01 -1.41875997e-01 9.11676824e-01 -1.18507409e+00
5.53878665e-01 -9.78830010e-02 8.93395901e-01 -8.62868190e-01
8.35464776e-01 -7.86561728e-01 9.72819328e-02 4.00867969e-01
-8.74264464e-02 3.98287565e-01 -1.85210958e-01 5.73406994e-01
2.19668746e-01 3.81359845e-01 7.06935287e-01 -4.31680262e-01
-6.34036958e-01 6.47747159e-01 3.65683675e-01 2.06681550e-01
8.06808233e-01 -6.48837864e-01 -3.42275262e-01 -7.35847652e-01
-3.29933763e-01 1.45524815e-01 9.90010917e-01 5.80391705e-01
7.36476183e-01 -1.59815133e+00 -4.64260489e-01 2.68074721e-01
3.39342862e-01 3.15714240e-01 2.36976449e-03 7.10823596e-01
-8.63375366e-01 4.59447324e-01 -1.71461046e-01 -1.04681659e+00
-1.20175064e+00 2.35690549e-01 6.49637282e-01 3.29135627e-01
-6.14126265e-01 9.06252742e-01 1.60632178e-01 -8.64351928e-01
4.99905795e-01 -2.23157302e-01 3.92195493e-01 -5.95288098e-01
3.85168821e-01 5.26079118e-01 -3.82621557e-01 -9.95597363e-01
-5.29885232e-01 1.23634923e+00 2.21462935e-01 -2.60022283e-01
1.06744099e+00 -5.69583654e-01 2.91360140e-01 2.63845950e-01
1.28909719e+00 7.83768669e-02 -1.60602069e+00 -3.60262841e-01
-5.65679014e-01 -8.29425991e-01 -2.62981951e-01 -5.95921934e-01
-9.86201525e-01 6.53768957e-01 7.11273909e-01 -6.57457113e-01
7.26178527e-01 1.22715056e-01 8.02920699e-01 5.62822223e-01
6.97409272e-01 -1.18624711e+00 3.03009659e-01 4.57851350e-01
8.08092594e-01 -1.37573266e+00 4.77656633e-01 -6.95453465e-01
-3.61031681e-01 9.18077111e-01 9.72024381e-01 -1.03659242e-01
5.18736184e-01 9.22500193e-02 3.16012174e-01 -3.19742113e-01
-7.42686167e-02 8.72616246e-02 4.18251902e-01 7.62323916e-01
2.53317147e-01 1.81986541e-02 2.06074938e-01 3.51079524e-01
-5.82152426e-01 -8.36948603e-02 2.40407959e-01 1.00246906e+00
-9.89447981e-02 -9.50304270e-01 -8.28188121e-01 -2.09107131e-01
-1.92070112e-01 3.01144630e-01 -6.53143167e-01 1.02722502e+00
-9.32439864e-02 5.53131163e-01 -3.32606971e-01 -4.91127998e-01
7.38793969e-01 -3.84703487e-01 9.18470860e-01 -3.34555358e-01
-1.18813485e-01 1.30841464e-01 -2.04262286e-01 -9.16699708e-01
-4.18593675e-01 -7.10594416e-01 -1.01054299e+00 -5.41891456e-01
-1.46607056e-01 -4.90922987e-01 6.14698231e-01 7.18610942e-01
3.62845868e-01 1.11379981e-01 2.34783038e-01 -1.24444675e+00
-5.98261416e-01 -6.32913470e-01 -3.32736105e-01 4.86056685e-01
4.57422704e-01 -8.72953951e-01 -3.59128416e-01 2.46322472e-02] | [7.031741619110107, -0.9647547006607056] |
10cc1164-5f7b-4cdf-8270-091e30ccac40 | co-generation-of-game-levels-and-game-playing | 2007.08497 | null | https://arxiv.org/abs/2007.08497v2 | https://arxiv.org/pdf/2007.08497v2.pdf | Co-generation of game levels and game-playing agents | Open-endedness, primarily studied in the context of artificial life, is the ability of systems to generate potentially unbounded ontologies of increasing novelty and complexity. Engineering generative systems displaying at least some degree of this ability is a goal with clear applications to procedural content generation in games. The Paired Open-Ended Trailblazer (POET) algorithm, heretofore explored only in a biped walking domain, is a coevolutionary system that simultaneously generates environments and agents that can solve them. This paper introduces a POET-Inspired Neuroevolutionary System for KreativitY (PINSKY) in games, which co-generates levels for multiple video games and agents that play them. This system leverages the General Video Game Artificial Intelligence (GVGAI) framework to enable co-generation of levels and agents for the 2D Atari-style games Zelda and Solar Fox. Results demonstrate the ability of PINSKY to generate curricula of game levels, opening up a promising new avenue for research at the intersection of procedural content generation and artificial life. At the same time, results in these challenging game domains highlight the limitations of the current algorithm and opportunities for improvement. | ['L. B. Soros', 'Julian Togelius', 'Aaron Dharna'] | 2020-07-16 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.38538852e-01 5.63001990e-01 5.14968336e-01 3.53652418e-01
-6.41516298e-02 -7.33145833e-01 7.15975225e-01 -3.57787311e-01
2.53329729e-03 8.63239527e-01 5.11356816e-02 -3.37663829e-01
-5.37673712e-01 -1.27882993e+00 -7.50978351e-01 -5.43271780e-01
-2.06064299e-01 4.91512150e-01 1.89972192e-01 -1.05480731e+00
1.71510577e-01 1.68073520e-01 -2.33521438e+00 -5.14436066e-02
1.14823449e+00 2.09657297e-01 2.01322019e-01 1.07265961e+00
9.06037241e-02 9.82871771e-01 -7.72481441e-01 -3.26949328e-01
2.91537941e-01 -1.03751242e+00 -4.80936170e-01 -1.69835120e-01
6.03286959e-02 1.98044911e-01 -1.25496298e-01 5.46706975e-01
6.92148447e-01 2.66601861e-01 6.30141735e-01 -1.38381636e+00
-6.90900385e-01 6.33023560e-01 -1.10297360e-01 8.24918002e-02
5.05932987e-01 5.96007228e-01 7.54753709e-01 -3.27601045e-01
9.78947699e-01 9.42208588e-01 8.14809740e-01 8.37533951e-01
-1.04456067e+00 -4.36564624e-01 -3.73300284e-01 -8.58217701e-02
-1.21079791e+00 1.04489870e-01 6.12268329e-01 -5.16084313e-01
9.56359625e-01 5.29515803e-01 1.71766853e+00 1.09581769e+00
5.59273899e-01 5.14205098e-01 1.00713027e+00 -5.50044537e-01
7.08778143e-01 -1.30479783e-01 -6.60430014e-01 7.21087456e-01
5.29744446e-01 6.63553119e-01 -7.26890385e-01 7.62445033e-02
1.02364039e+00 -6.68404162e-01 9.63215232e-02 -5.85780323e-01
-7.91830361e-01 9.38116670e-01 2.54271716e-01 2.49515593e-01
-4.03755784e-01 4.26173210e-01 1.92963928e-02 1.94424227e-01
7.06662834e-02 1.43595874e+00 -7.86575153e-02 -7.61802733e-01
-7.49971628e-01 1.03725791e+00 1.10125756e+00 8.00971270e-01
3.38853747e-01 6.75074041e-01 -1.33180380e-01 5.54850876e-01
2.42531493e-01 1.80128977e-01 5.86958766e-01 -1.17760098e+00
-3.77895594e-01 8.39014649e-01 -1.47296607e-01 -9.42951024e-01
-2.86930889e-01 -6.49928033e-01 -9.79959816e-02 1.09327459e+00
1.80235505e-01 -5.25126278e-01 -7.92047441e-01 1.66122437e+00
6.45118535e-01 3.17412019e-01 1.54551625e-01 7.87634373e-01
8.43873739e-01 7.06006587e-01 7.55299479e-02 1.92222267e-01
1.23453736e+00 -8.66339803e-01 -2.68441617e-01 -1.87602490e-01
4.65632945e-01 -3.63652349e-01 1.20055819e+00 4.84738678e-01
-1.56381953e+00 -2.71958560e-01 -1.31634104e+00 5.12076497e-01
-5.41305304e-01 -8.19782078e-01 7.87202835e-01 1.18977451e+00
-1.46846592e+00 7.05313563e-01 -5.78866959e-01 -4.98234689e-01
1.37266845e-01 2.37718239e-01 2.54403502e-01 3.69442731e-01
-1.20069945e+00 1.08246112e+00 3.75411063e-01 -3.39391500e-01
-7.55343437e-01 -1.04689920e+00 -7.55719662e-01 6.06042184e-02
3.59763414e-01 -1.18155611e+00 1.30289400e+00 -1.05057764e+00
-1.91639590e+00 6.91143870e-01 7.66549528e-01 -4.37660933e-01
6.17480814e-01 3.18445951e-01 -3.09723876e-02 -2.29769319e-01
4.94476687e-03 9.09775317e-01 3.79394323e-01 -1.16044521e+00
-4.72233266e-01 4.19912338e-02 1.67394131e-01 7.59382606e-01
-1.17696881e-01 -2.45275378e-01 3.31473388e-02 -8.18909883e-01
-4.05002505e-01 -9.09876287e-01 -4.70495313e-01 -3.30541193e-01
2.16552764e-01 -2.27475259e-02 5.70911825e-01 -2.48672828e-01
1.14325833e+00 -1.52864242e+00 5.99479198e-01 7.89812356e-02
2.80575484e-01 3.01703155e-01 -6.96421042e-02 8.26872349e-01
1.53240308e-01 2.48986274e-01 -5.36975749e-02 3.16276580e-01
1.78017408e-01 3.03820759e-01 1.99257627e-01 -3.85624826e-01
1.18735336e-01 1.26787472e+00 -1.06176233e+00 -3.17679048e-01
6.12493902e-02 3.03734839e-01 -1.00500464e+00 1.08052986e-02
-6.42625809e-01 4.41150337e-01 -2.67581671e-01 3.65693152e-01
2.14176141e-02 1.93472669e-01 6.90585151e-02 6.13498151e-01
-3.27775866e-01 -2.58070230e-01 -1.11966825e+00 1.59810174e+00
-2.25411654e-01 6.04564428e-01 -6.62172139e-02 -4.28725094e-01
1.05409408e+00 7.25370497e-02 5.11064768e-01 -7.90906429e-01
2.66835868e-01 3.24777842e-01 5.02198875e-01 -5.57869554e-01
8.90892506e-01 -2.23221645e-01 -2.51926631e-01 6.03337228e-01
1.74689725e-01 -1.10570061e+00 5.86092055e-01 3.19099240e-02
1.39837492e+00 7.81223118e-01 1.29237249e-01 -5.28008878e-01
-9.16244835e-03 5.02103984e-01 2.99575031e-01 8.90037715e-01
-1.23633504e-01 5.65746844e-01 2.94676363e-01 -3.71853620e-01
-1.46278322e+00 -1.19137311e+00 2.74076402e-01 1.22862756e+00
1.11363530e-01 -6.04542375e-01 -1.15337324e+00 2.27709189e-01
-1.68759927e-01 8.98523033e-01 -7.84070909e-01 -4.00763094e-01
-4.94367719e-01 -6.03869736e-01 8.15734863e-01 1.84694365e-01
2.48865992e-01 -1.60448229e+00 -1.46156192e+00 5.33640027e-01
7.63067827e-02 -3.28891993e-01 8.11233670e-02 1.21338647e-02
-5.75020969e-01 -7.97603786e-01 -7.14836955e-01 -9.15198565e-01
4.74452302e-02 -8.51813853e-02 1.41440868e+00 2.35881642e-01
-8.84560108e-01 6.80019677e-01 -5.10602593e-01 -8.39610457e-01
-1.04663384e+00 -6.57028519e-03 -1.51883006e-01 -7.90970147e-01
-4.56902720e-02 -9.38638568e-01 -6.49604023e-01 1.97170034e-01
-1.03019011e+00 5.03126442e-01 2.65348732e-01 8.97294044e-01
2.54750997e-01 2.41396025e-01 5.86112797e-01 -4.55986172e-01
1.20321393e+00 -5.41239619e-01 -4.55467671e-01 1.88391551e-01
-5.40143907e-01 1.72561258e-02 3.73020709e-01 -6.10113621e-01
-9.65303361e-01 -4.98568416e-01 6.17662966e-02 -9.25197452e-03
-1.26528023e-02 5.98697484e-01 2.43236721e-02 -2.81935841e-01
1.16100144e+00 3.08915615e-01 2.98959404e-01 2.82830179e-01
6.89062238e-01 8.57243910e-02 5.68087041e-01 -1.00958550e+00
7.80045509e-01 -1.12234801e-01 1.53234154e-02 -7.90193558e-01
1.67022526e-01 5.15068948e-01 -5.37617095e-02 -8.63176525e-01
8.70491624e-01 -4.67258722e-01 -7.42286623e-01 6.65415466e-01
-5.03054142e-01 -7.79424429e-01 -1.06237805e+00 1.80995967e-02
-1.02803576e+00 -2.46307179e-01 -4.98523474e-01 -7.62833059e-01
-3.09499681e-01 -9.74446118e-01 6.13498092e-01 8.76357973e-01
-6.69692874e-01 -8.11356366e-01 7.06885993e-01 2.78775781e-01
6.26772285e-01 7.98247635e-01 7.94947505e-01 -1.94370896e-01
-5.39903939e-01 1.64644599e-01 7.06660151e-01 -3.39221328e-01
-9.57863927e-02 3.43476325e-01 -3.36534053e-01 -1.89924501e-02
-2.60059446e-01 -3.60002577e-01 -5.13661206e-02 3.51566851e-01
3.10790390e-01 -2.55246043e-01 3.32330503e-02 5.48495233e-01
1.42593384e+00 8.38461518e-01 9.11066711e-01 8.44705641e-01
3.23659658e-01 7.39017069e-01 4.20079947e-01 5.76124012e-01
3.85206431e-01 6.49913371e-01 3.53416562e-01 7.35923201e-02
-2.28383303e-01 -4.57042813e-01 1.77591726e-01 7.32776761e-01
-3.26766968e-01 -3.37674588e-01 -1.17350292e+00 4.95117247e-01
-1.89025092e+00 -1.38314009e+00 -1.38086885e-01 1.68060029e+00
8.51817131e-01 5.57316914e-02 4.62245107e-01 6.69101700e-02
5.71440697e-01 -2.93095589e-01 -6.41866088e-01 -1.04218876e+00
-1.93121284e-01 8.07321191e-01 -3.70194651e-02 1.87553674e-01
-4.51806247e-01 1.13664150e+00 6.65481234e+00 8.37600470e-01
-6.78251088e-01 -1.71283886e-01 6.07952535e-01 -2.85541385e-01
-6.07512295e-01 -7.96562731e-02 -2.04881519e-01 5.21918356e-01
9.42442596e-01 -9.27050292e-01 9.43118215e-01 7.95015752e-01
2.97471464e-01 -1.41560793e-01 -5.99634469e-01 4.90940541e-01
6.96811751e-02 -1.75516844e+00 -8.38601682e-03 2.10654214e-01
1.06672788e+00 -5.11430442e-01 3.30690235e-01 4.68249917e-01
1.08368528e+00 -1.35293519e+00 1.10811365e+00 3.58096063e-01
5.09577394e-01 -1.10786211e+00 8.68101716e-02 1.83502495e-01
-9.61657584e-01 -1.92151740e-01 4.30924483e-02 -4.35278684e-01
1.83006793e-01 -3.55591089e-01 -7.28013456e-01 3.90131950e-01
7.43486345e-01 1.00239389e-01 -5.45781195e-01 1.35378718e+00
-1.30881399e-01 4.87968713e-01 -1.67440042e-01 -6.00510955e-01
3.02584469e-01 -4.49472725e-01 8.49588871e-01 7.41737962e-01
6.01311445e-01 4.57547128e-01 -3.19310308e-01 1.19051290e+00
5.51189601e-01 1.16223330e-02 -1.00288618e+00 -2.02779830e-01
5.15922189e-01 1.07660198e+00 -9.34301674e-01 -2.33984068e-02
2.80535132e-01 7.50169814e-01 2.98630632e-02 1.59278125e-01
-1.02649522e+00 -4.29658353e-01 7.43491828e-01 2.98131764e-01
2.24875167e-01 -2.79355168e-01 -8.20194125e-01 -6.89246535e-01
-5.82667708e-01 -1.37420523e+00 9.08186510e-02 -1.24007165e+00
-9.00079131e-01 5.02061725e-01 -9.09973308e-02 -9.59528983e-01
-7.04717636e-01 -2.67640263e-01 -8.91076922e-01 5.76564670e-01
-4.24376905e-01 -1.16844034e+00 -5.59411526e-01 3.90300006e-01
4.84950542e-01 -3.93226355e-01 6.96715117e-01 -3.48112315e-01
-4.32918578e-01 4.82361048e-01 1.90508097e-01 -5.61869144e-01
-5.80379255e-02 -1.23061955e+00 7.28626490e-01 7.58450508e-01
2.43686195e-02 2.96950251e-01 1.11761844e+00 -8.83918285e-01
-1.58593643e+00 -6.06003344e-01 7.10000917e-02 -4.20127630e-01
5.38981378e-01 -3.14283937e-01 -4.64333057e-01 1.18355073e-01
4.72930998e-01 -1.04732871e+00 6.26022995e-01 -1.56681791e-01
1.92169875e-01 4.41776663e-01 -1.10568380e+00 1.40741408e+00
1.45844746e+00 5.85061163e-02 -4.85687137e-01 -1.07119374e-01
8.38977993e-01 -7.53755689e-01 -7.63818979e-01 2.42587820e-01
7.78999150e-01 -1.11941421e+00 1.02460134e+00 -5.09913921e-01
9.67352450e-01 -1.75438613e-01 2.26561308e-01 -1.60378039e+00
-4.95452672e-01 -1.14123511e+00 1.11762539e-01 1.12623644e+00
2.61142999e-01 -4.06768322e-01 1.16213238e+00 7.49026895e-01
-4.09472793e-01 -8.15413773e-01 -6.81782603e-01 -8.11070085e-01
5.95118582e-01 -2.23876297e-01 8.51920426e-01 8.73626471e-01
3.07293355e-01 1.62740305e-01 -1.05682947e-01 -5.90480208e-01
3.71687353e-01 -3.13139528e-01 9.98441935e-01 -1.18561184e+00
-8.32640707e-01 -9.84118879e-01 -6.90713227e-01 -6.49682507e-02
-2.91440576e-01 -7.01965332e-01 1.97296992e-01 -1.56440151e+00
-8.26117545e-02 -5.23223579e-01 3.72089863e-01 7.94358552e-02
-1.32425070e-01 1.67163610e-01 4.99973983e-01 -8.14082772e-02
-1.06895268e-01 5.52188635e-01 1.44008017e+00 1.35988116e-01
-7.32222080e-01 -5.05988419e-01 -1.12868845e+00 4.90740031e-01
9.75241184e-01 -1.52033523e-01 -8.60970914e-01 -2.14600578e-01
6.18497074e-01 2.61058081e-02 2.68115968e-01 -1.49698734e+00
1.59418091e-01 -4.27215248e-01 6.93706647e-02 -2.46142875e-02
2.82357633e-01 -2.18464911e-01 9.70257223e-01 7.66382098e-01
-7.41798207e-02 4.41228747e-01 4.72612619e-01 3.85986120e-01
2.67277151e-01 -1.65572003e-01 5.71972847e-01 -4.01062697e-01
-7.34050930e-01 -2.72568882e-01 -1.15921235e+00 3.49046648e-01
1.51340723e+00 -1.11701608e+00 -4.58645403e-01 -5.58064520e-01
-6.31810486e-01 1.73486531e-01 8.47033620e-01 5.40005505e-01
5.91609001e-01 -1.20983231e+00 -8.12041402e-01 8.15396383e-02
-1.00374520e-01 -9.59378108e-02 2.27506593e-01 8.21904764e-02
-1.16670012e+00 -2.22902194e-01 -9.41729724e-01 -2.95708418e-01
-9.88234937e-01 3.30180764e-01 6.56250119e-01 -2.63983846e-01
-5.54091334e-01 1.00156879e+00 5.09214066e-02 -5.06690085e-01
-3.02073359e-01 1.95995867e-01 -2.16441616e-01 -9.20353159e-02
2.32101098e-01 5.40736318e-01 -2.22070113e-01 -2.66629189e-01
7.71400928e-02 1.79753333e-01 4.89987820e-01 -7.42917120e-01
1.47587073e+00 3.17533553e-01 -2.07024515e-02 2.47256964e-01
1.29100725e-01 -2.66883045e-01 -1.15954411e+00 6.42071605e-01
-2.60342598e-01 -1.65943027e-01 -3.88789982e-01 -1.10621941e+00
-6.25591457e-01 3.69516999e-01 3.98898065e-01 2.29020953e-01
1.02327788e+00 -3.19847763e-01 5.23821235e-01 8.65425821e-03
6.42892599e-01 -1.15403032e+00 5.22103488e-01 5.18569291e-01
8.73624206e-01 -2.88098276e-01 -3.71924818e-01 -3.18326689e-02
-7.44178712e-01 9.86339271e-01 1.17237270e+00 -3.09334248e-01
1.95801750e-01 5.72706223e-01 -1.23031981e-01 -4.34639245e-01
-9.89486575e-01 -1.40705690e-01 -5.19393245e-03 1.09373212e+00
3.21871191e-01 6.83099926e-02 -8.46956253e-01 4.14702266e-01
-1.04620910e+00 1.92369580e-01 9.67998564e-01 1.36367702e+00
-5.47833800e-01 -1.19420433e+00 -5.22635281e-01 2.90589035e-01
-3.37325200e-03 -1.03166550e-02 -7.15530097e-01 9.59747732e-01
2.92609185e-01 9.69060063e-01 1.07496334e-02 -6.31739676e-01
1.99172378e-01 -1.99704036e-01 7.65517414e-01 -5.01516402e-01
-1.27840924e+00 -3.77493232e-01 1.11648038e-01 -2.86633939e-01
3.16009410e-02 -6.25561655e-01 -1.05682790e+00 -7.64186502e-01
-4.78259958e-02 3.51418585e-01 6.12441480e-01 2.89454699e-01
5.98327756e-01 8.78872693e-01 1.19001620e-01 -1.09353113e+00
-4.57041413e-02 -3.46297085e-01 -4.86156493e-01 2.06358820e-01
-6.57099783e-01 -7.24321485e-01 2.65098691e-01 -1.43243810e-02] | [3.6364943981170654, 1.5810099840164185] |
8d65851a-4d05-4b27-bff0-afc9f2af7a86 | joint-embedding-in-hierarchical-distance-and | 2303.15655 | null | https://arxiv.org/abs/2303.15655v1 | https://arxiv.org/pdf/2303.15655v1.pdf | Joint embedding in Hierarchical distance and semantic representation learning for link prediction | The link prediction task aims to predict missing entities or relations in the knowledge graph and is essential for the downstream application. Existing well-known models deal with this task by mainly focusing on representing knowledge graph triplets in the distance space or semantic space. However, they can not fully capture the information of head and tail entities, nor even make good use of hierarchical level information. Thus, in this paper, we propose a novel knowledge graph embedding model for the link prediction task, namely, HIE, which models each triplet (\textit{h}, \textit{r}, \textit{t}) into distance measurement space and semantic measurement space, simultaneously. Moreover, HIE is introduced into hierarchical-aware space to leverage rich hierarchical information of entities and relations for better representation learning. Specifically, we apply distance transformation operation on the head entity in distance space to obtain the tail entity instead of translation-based or rotation-based approaches. Experimental results of HIE on four real-world datasets show that HIE outperforms several existing state-of-the-art knowledge graph embedding methods on the link prediction task and deals with complex relations accurately. | ['Fengyu Zhou', 'Chongfeng Fan', 'Jianye Chen', 'Jin Liu'] | 2023-03-28 | null | null | null | null | ['knowledge-graph-embedding'] | ['graphs'] | [-2.45645568e-01 2.84525901e-01 -4.93696809e-01 -1.42227501e-01
-6.31665811e-02 -5.07892549e-01 5.00350058e-01 4.49892223e-01
-6.68951347e-02 5.75368404e-01 3.11316520e-01 -2.69744396e-01
-7.54800677e-01 -1.18832803e+00 -4.99486357e-01 -4.38162327e-01
-2.07853522e-02 6.01303637e-01 3.91364068e-01 -2.01387942e-01
-2.74550289e-01 2.59347588e-01 -9.09493148e-01 -1.91011727e-01
8.32415700e-01 9.52297628e-01 -1.32859439e-01 1.01081119e-03
-3.05152237e-01 9.59205806e-01 -1.62442878e-01 -8.95243645e-01
-2.52265744e-02 -1.56332493e-01 -8.16161335e-01 -3.64892334e-01
2.17788860e-01 -1.26947453e-02 -1.01820874e+00 9.87982869e-01
3.67939264e-01 1.32840186e-01 6.00971818e-01 -1.57320309e+00
-1.10568261e+00 6.85689449e-01 -4.10079420e-01 1.41182765e-01
3.36882263e-01 -4.43709075e-01 1.49377263e+00 -9.19573069e-01
6.88391387e-01 1.20286226e+00 6.75119162e-01 -1.18312784e-01
-1.03893363e+00 -5.76222956e-01 2.36902058e-01 7.91913092e-01
-1.85161364e+00 -1.40005350e-02 1.06096315e+00 -5.77317595e-01
8.27316225e-01 2.34869897e-01 7.11933017e-01 7.24946737e-01
-2.26427466e-01 5.48118591e-01 6.63038194e-01 -1.12400398e-01
-1.87463582e-01 7.52709731e-02 4.80602741e-01 9.28746700e-01
6.53731167e-01 -2.32916605e-03 -4.47911590e-01 -9.54893529e-02
7.44154990e-01 1.47701681e-01 -4.92149085e-01 -8.57116818e-01
-1.52534270e+00 7.84844339e-01 1.01846731e+00 4.19020295e-01
-3.78199846e-01 4.90252189e-02 3.60391498e-01 2.02724129e-01
3.53226393e-01 2.27897480e-01 -4.79697436e-01 1.71445176e-01
-2.98743933e-01 -2.08857641e-01 7.10931897e-01 1.31170547e+00
1.06134403e+00 -2.62652963e-01 -1.89665839e-01 7.05491662e-01
3.98678392e-01 2.19457634e-02 3.74346167e-01 -3.77277285e-01
9.23189461e-01 1.19672048e+00 4.29442013e-03 -1.76313865e+00
-6.76089406e-01 -6.52671695e-01 -8.41865838e-01 -5.89752793e-01
-3.47186513e-02 -1.37161640e-02 -5.82779408e-01 1.54864681e+00
6.03684247e-01 6.65451646e-01 1.40155688e-01 7.43273556e-01
1.20043242e+00 5.21615744e-01 3.43224518e-02 -1.75479427e-01
1.30781770e+00 -1.16225398e+00 -8.18898320e-01 -3.07434034e-02
1.06776762e+00 -3.88291925e-01 7.08274603e-01 -3.13346505e-01
-6.26866639e-01 -5.47559798e-01 -9.85252202e-01 -2.52641380e-01
-8.06755662e-01 2.60628700e-01 6.79705620e-01 2.07972378e-01
-7.45309949e-01 4.36273426e-01 -5.82503378e-01 -3.49290639e-01
1.75530195e-01 3.05261075e-01 -7.34668672e-01 -3.30962092e-01
-1.68404043e+00 9.85367239e-01 1.00113750e+00 4.14195001e-01
-9.11707133e-02 -5.95235169e-01 -1.21545148e+00 2.31642872e-01
8.02108288e-01 -8.32818449e-01 3.00145686e-01 -1.21948943e-02
-1.07811666e+00 4.20620859e-01 1.11837126e-01 -1.54149942e-02
9.51243043e-02 -3.74310054e-02 -7.81142294e-01 -8.78583118e-02
8.75886530e-03 1.23240806e-01 4.43805993e-01 -1.31140661e+00
-5.09159744e-01 -6.69231474e-01 1.95391759e-01 2.95957506e-01
-6.59868181e-01 -6.14834368e-01 -7.39915073e-01 -6.24683321e-01
4.77579743e-01 -8.15289199e-01 2.89389133e-01 -1.62318066e-01
-7.92582572e-01 -4.48004574e-01 9.42320168e-01 -9.33290482e-01
1.68968141e+00 -1.96252656e+00 5.80639243e-01 5.25268853e-01
6.28629386e-01 4.26122397e-01 -8.87223780e-02 7.40105987e-01
-8.33730027e-02 9.17034969e-02 9.79418904e-02 1.34079996e-02
3.01638603e-01 4.38334942e-01 1.02769211e-02 2.96611369e-01
-1.41514838e-01 1.38211036e+00 -1.07680643e+00 -8.11474919e-01
2.26603583e-01 5.97310185e-01 -3.23210239e-01 1.06601957e-02
1.22264914e-01 2.80715108e-01 -9.08331871e-01 4.94234979e-01
4.75487411e-01 -4.33717251e-01 4.20923710e-01 -9.37810004e-01
2.54200548e-01 1.34262919e-01 -1.16156435e+00 1.78807390e+00
-3.08440477e-01 2.41489246e-01 -3.31109405e-01 -1.12291873e+00
1.09019351e+00 1.90151975e-01 5.72297454e-01 -6.39887512e-01
-2.67606527e-02 8.10694396e-02 1.58087220e-02 -4.79443341e-01
3.17729294e-01 2.20883295e-01 9.42808762e-02 -8.66827816e-02
1.52740240e-01 3.67026359e-01 2.88865436e-02 3.27639043e-01
1.17035592e+00 8.93168300e-02 2.58471906e-01 1.54613569e-01
7.87290275e-01 -3.13001961e-01 7.54658520e-01 2.04657272e-01
-1.58421323e-02 -9.95775871e-03 5.28237283e-01 -1.61426306e-01
-6.77094996e-01 -9.82296765e-01 1.12594686e-01 8.89890075e-01
6.50658131e-01 -9.37263966e-01 -2.62357414e-01 -1.10933113e+00
4.04157132e-01 4.91376907e-01 -6.57389462e-01 -6.05985165e-01
-4.67408031e-01 -5.15314877e-01 3.07627380e-01 6.94254935e-01
5.41696906e-01 -5.90387046e-01 3.77317220e-01 2.29059726e-01
-3.21266770e-01 -1.23508525e+00 -4.87959653e-01 -2.13485003e-01
-5.27503908e-01 -1.26866961e+00 -3.54875654e-01 -9.86752033e-01
5.74756742e-01 3.01586062e-01 8.05444300e-01 2.44458914e-01
-6.08221330e-02 4.42289263e-01 -5.96531510e-01 1.86161071e-01
2.31194049e-01 2.43135139e-01 -9.26400721e-02 2.30696082e-01
3.83074522e-01 -6.99462235e-01 -6.43962562e-01 4.60350662e-01
-6.99295521e-01 7.38185197e-02 6.95445597e-01 9.62036371e-01
7.09706187e-01 4.47837532e-01 5.56864619e-01 -8.53380799e-01
4.82694060e-01 -6.16584897e-01 -2.27035597e-01 6.32547975e-01
-7.91455388e-01 2.12411806e-01 4.80906636e-01 -1.72335267e-01
-7.48971105e-01 -3.66201937e-01 1.87693983e-01 -5.36567390e-01
3.60996276e-01 1.11240041e+00 -5.42694211e-01 -1.22473434e-01
1.52524874e-01 2.67949700e-01 -4.78544950e-01 -6.05879188e-01
6.25319600e-01 3.88155788e-01 4.47618455e-01 -4.08067763e-01
1.14429080e+00 2.81265795e-01 4.31364238e-01 -4.15794581e-01
-9.36477780e-01 -5.61028659e-01 -9.61411595e-01 1.91931963e-01
8.14593613e-01 -8.94830704e-01 -6.29682779e-01 -1.84190795e-02
-9.70637500e-01 1.93269879e-01 -6.40756339e-02 6.48914754e-01
-1.54060915e-01 5.52942216e-01 -4.10174519e-01 -3.22397530e-01
-1.81727216e-01 -6.95913732e-01 9.72407222e-01 -3.22933891e-03
2.37146050e-01 -1.41479945e+00 3.68707217e-02 5.51074922e-01
4.19145152e-02 3.40934157e-01 1.35078979e+00 -8.59881520e-01
-5.80750823e-01 -3.05979639e-01 -7.29620636e-01 1.64714128e-01
3.74942482e-01 -2.62571126e-01 -3.10635954e-01 -3.02566528e-01
-7.07649529e-01 5.16298749e-02 8.48664284e-01 -3.43511909e-01
9.03645813e-01 -4.82841343e-01 -7.70836711e-01 5.78846455e-01
1.34387612e+00 -1.41162202e-01 4.45951492e-01 3.07501435e-01
1.57469857e+00 6.14771664e-01 7.82616079e-01 1.55607820e-01
1.13967264e+00 9.97802734e-01 4.74666297e-01 1.54669836e-01
-2.30620354e-01 -7.62219965e-01 -8.22437406e-02 1.24926186e+00
-1.23344600e-01 -2.88530022e-01 -8.55416417e-01 5.76953769e-01
-2.31272840e+00 -7.69887447e-01 -2.71013290e-01 2.00363255e+00
7.18357861e-01 1.53787592e-02 -2.00472906e-01 1.05695188e-01
8.99014592e-01 2.06746519e-01 -4.43561643e-01 3.60283107e-01
3.93568091e-02 -2.61986762e-01 3.95108402e-01 4.17770386e-01
-9.04098749e-01 1.04330921e+00 3.96165013e+00 8.28345358e-01
-7.03236163e-01 1.19919918e-01 -2.80489922e-01 5.26936471e-01
-4.99559283e-01 2.73387939e-01 -7.15322495e-01 5.58302760e-01
3.92621547e-01 -3.85387838e-01 3.75707805e-01 4.86063540e-01
-2.98239261e-01 4.80139643e-01 -1.18523240e+00 1.17570579e+00
6.58697486e-02 -1.13358533e+00 3.08530033e-01 4.00286391e-02
6.62795782e-01 -3.94124001e-01 -1.75669789e-01 7.17203498e-01
1.35579839e-01 -9.07521725e-01 2.28297397e-01 7.99130082e-01
5.12255788e-01 -7.38294303e-01 8.91887128e-01 2.12819874e-01
-1.76028812e+00 -2.07078643e-02 -4.67528760e-01 2.78147191e-01
1.48795664e-01 5.99784434e-01 -6.46460891e-01 1.43926275e+00
4.70947206e-01 1.22342479e+00 -7.61373281e-01 9.59224761e-01
-4.80170459e-01 2.27014035e-01 -1.21650122e-01 1.65885538e-01
1.63970545e-01 -3.84393603e-01 3.92029554e-01 8.70831072e-01
4.65059340e-01 2.29037166e-01 3.59168798e-01 5.64853430e-01
-3.74871016e-01 3.66313964e-01 -5.34458339e-01 -1.55156285e-01
8.53262067e-01 1.26256251e+00 -2.48275459e-01 -1.89118072e-01
-6.52748704e-01 1.01676667e+00 7.76212394e-01 5.36056817e-01
-9.01668608e-01 -6.40795529e-01 6.17274463e-01 -7.72783905e-03
5.31422973e-01 -2.91125268e-01 2.47278228e-01 -1.30428100e+00
3.09862375e-01 -2.34807104e-01 6.58050120e-01 -7.66035080e-01
-1.39470732e+00 2.96560377e-01 -2.37474609e-02 -1.14574122e+00
1.38438493e-01 -5.88262677e-01 -4.95175868e-01 6.69715464e-01
-1.75182533e+00 -1.72644365e+00 -5.56285441e-01 8.13237906e-01
-2.57157981e-01 -1.37386650e-01 6.86233401e-01 5.63327730e-01
-8.37455273e-01 7.46773362e-01 2.73954570e-01 4.59289879e-01
5.57351112e-01 -1.17377460e+00 1.08393006e-01 3.66050273e-01
3.84503931e-01 6.77807808e-01 2.34150380e-01 -7.84358859e-01
-1.61668158e+00 -1.51418602e+00 9.31407332e-01 -5.07643700e-01
1.04941595e+00 -7.46200234e-02 -1.07544303e+00 1.06606507e+00
-2.92232186e-01 4.78256941e-01 7.32506454e-01 3.67775053e-01
-6.25978649e-01 -2.46857405e-01 -8.38264763e-01 5.32738149e-01
1.50228739e+00 -7.70205498e-01 -7.01401055e-01 2.89181739e-01
9.65221882e-01 -3.12299788e-01 -1.64867353e+00 7.35479653e-01
3.85506928e-01 -5.47806203e-01 1.29740167e+00 -6.98812127e-01
8.85977447e-02 -5.39339542e-01 -1.08271740e-01 -1.45560861e+00
-5.75348496e-01 -2.01012179e-01 -8.18283916e-01 1.47968495e+00
2.82284170e-01 -8.02780151e-01 7.08041072e-01 1.94637612e-01
-5.75564541e-02 -8.85646045e-01 -1.07288694e+00 -8.93781066e-01
-1.94342658e-01 8.35996345e-02 8.52814794e-01 1.43576658e+00
1.65805757e-01 7.99488127e-01 -4.23652381e-01 5.36116481e-01
6.77332878e-01 4.08618599e-01 6.91297889e-01 -1.60660493e+00
-1.12242840e-01 -2.80404299e-01 -1.13060975e+00 -9.66324627e-01
3.86506885e-01 -1.18343437e+00 -6.28288507e-01 -2.14431143e+00
4.86775599e-02 -5.61297297e-01 -6.18233204e-01 5.76114357e-01
-3.94261539e-01 4.96944152e-02 6.10040836e-02 2.74860084e-01
-7.61799991e-01 9.85908449e-01 1.34082341e+00 -3.37053150e-01
-1.40832085e-02 -4.31557655e-01 -6.84691548e-01 6.15387917e-01
4.01033133e-01 -1.63701102e-01 -7.53302276e-01 -5.50923586e-01
4.71887469e-01 8.87528062e-02 5.29819071e-01 -7.42086172e-01
4.92034107e-01 -4.43652347e-02 1.96381882e-01 -5.21186769e-01
4.43456709e-01 -1.13638985e+00 3.26225668e-01 -7.79577298e-04
-5.86568788e-02 -1.67746067e-01 -2.10187241e-01 1.02615368e+00
-4.30546045e-01 7.24877939e-02 4.31131385e-02 3.61859500e-01
-9.75427568e-01 7.75009930e-01 7.83846557e-01 -7.08279433e-03
1.18583965e+00 -1.04380650e-02 -7.03591585e-01 -1.53575704e-01
-9.97349918e-01 6.08233750e-01 2.41209745e-01 6.17604136e-01
6.12119377e-01 -1.92396414e+00 -5.53593814e-01 -1.49846926e-01
6.74937963e-01 1.17659494e-01 2.83396423e-01 1.08817708e+00
-1.24570489e-01 5.43976128e-01 1.08998761e-01 -2.08583832e-01
-1.11976576e+00 9.58090544e-01 8.62774998e-02 -5.22730768e-01
-6.72791839e-01 6.46911502e-01 2.42612496e-01 -5.78326106e-01
1.20877497e-01 -7.17998520e-02 -4.73784566e-01 2.93181628e-01
-3.59993577e-02 5.58627903e-01 -4.60904539e-02 -9.85186756e-01
-4.76849616e-01 8.68402839e-01 -8.65148082e-02 4.23091739e-01
1.09300172e+00 -3.74624401e-01 -2.73665994e-01 3.24339449e-01
1.50228882e+00 -7.41170812e-03 -6.91005945e-01 -7.93300390e-01
1.82443023e-01 -4.71528769e-01 1.49812594e-01 -5.42695343e-01
-1.13617051e+00 6.24881387e-01 1.68473527e-01 3.05104196e-01
7.19709396e-01 2.27857396e-01 9.16594386e-01 5.90659499e-01
4.64215994e-01 -8.89728367e-01 3.66422720e-02 3.76385212e-01
8.26485991e-01 -1.13793194e+00 1.64177760e-01 -8.45499158e-01
-4.98917162e-01 8.71428132e-01 7.07387030e-01 2.92861998e-01
9.18756604e-01 -6.06607556e-01 -3.67331386e-01 -5.49868882e-01
-3.63744885e-01 -5.51743269e-01 7.14334011e-01 6.55851126e-01
1.93019986e-01 1.29813552e-01 -4.12860274e-01 6.65916860e-01
-1.82134226e-01 -3.09624314e-01 -1.58478003e-02 8.19426596e-01
-2.95660615e-01 -1.15922713e+00 1.05173640e-01 4.75269556e-01
2.14902490e-01 -1.95878111e-02 -5.64920902e-01 9.15080130e-01
3.08543354e-01 9.15458381e-01 -3.39246660e-01 -8.46063852e-01
5.20536184e-01 6.23520389e-02 3.51294190e-01 -5.57691514e-01
1.65048782e-02 -4.14472133e-01 2.35776693e-01 -3.84431094e-01
-2.50890851e-01 -3.26151520e-01 -1.43259645e+00 -3.09189558e-01
-6.73602939e-01 4.03034925e-01 3.42838854e-01 9.07507181e-01
5.14586747e-01 7.17720330e-01 5.84874451e-01 -2.85061687e-01
-1.90282866e-01 -7.62021303e-01 -8.79873931e-01 6.04616344e-01
1.21939808e-01 -1.14782143e+00 -1.51145115e-01 -2.86186427e-01] | [8.676774024963379, 7.850226402282715] |
563a6101-9822-4ddf-8fd3-7bc0644d0eb3 | 3d-object-tracking-with-transformer | 2110.14921 | null | https://arxiv.org/abs/2110.14921v1 | https://arxiv.org/pdf/2110.14921v1.pdf | 3D Object Tracking with Transformer | Feature fusion and similarity computation are two core problems in 3D object tracking, especially for object tracking using sparse and disordered point clouds. Feature fusion could make similarity computing more efficient by including target object information. However, most existing LiDAR-based approaches directly use the extracted point cloud feature to compute similarity while ignoring the attention changes of object regions during tracking. In this paper, we propose a feature fusion network based on transformer architecture. Benefiting from the self-attention mechanism, the transformer encoder captures the inter- and intra- relations among different regions of the point cloud. By using cross-attention, the transformer decoder fuses features and includes more target cues into the current point cloud feature to compute the region attentions, which makes the similarity computing more efficient. Based on this feature fusion network, we propose an end-to-end point cloud object tracking framework, a simple yet effective method for 3D object tracking using point clouds. Comprehensive experimental results on the KITTI dataset show that our method achieves new state-of-the-art performance. Code is available at: https://github.com/3bobo/lttr. | ['Sifan Zhou', 'Zuoxu Gu', 'Jiayao Shan', 'Zheng Fang', 'Yubo Cui'] | 2021-10-28 | null | null | null | null | ['3d-object-tracking'] | ['computer-vision'] | [-4.68478620e-01 -7.26396203e-01 -8.35379586e-02 -3.04524004e-01
-5.86121798e-01 -3.40848476e-01 4.02972758e-01 -5.05165681e-02
-9.56370533e-02 -2.77348906e-02 -1.15908936e-01 1.35014668e-01
3.92603949e-02 -5.93119383e-01 -7.26921916e-01 -5.68092823e-01
1.49677500e-01 5.37323356e-01 6.22184217e-01 -1.26286764e-02
1.61671504e-01 9.39236224e-01 -1.64686370e+00 1.58480946e-02
6.01818740e-01 1.31088650e+00 5.31683922e-01 2.60290712e-01
-3.74797612e-01 3.84212226e-01 -2.47741595e-01 -1.69045717e-01
5.68068385e-01 1.75576970e-01 -7.47491866e-02 1.58465058e-01
7.97056675e-01 -4.82882053e-01 -5.46776175e-01 1.37690532e+00
6.75165594e-01 4.54121456e-02 1.98758617e-01 -1.48732674e+00
-8.08990538e-01 1.44452989e-01 -6.40125394e-01 5.00074804e-01
1.53762296e-01 4.52728182e-01 7.65016198e-01 -1.13846457e+00
3.56014013e-01 1.59985590e+00 7.83450484e-01 4.55743909e-01
-7.76560187e-01 -1.26012266e+00 4.28301781e-01 3.16720963e-01
-1.54092240e+00 -3.45001668e-01 9.02271271e-01 -3.62210721e-01
8.90582323e-01 1.06047019e-01 1.08661044e+00 4.17834550e-01
8.27719867e-02 1.02670860e+00 6.07523978e-01 1.00097105e-01
-2.99617797e-01 -3.70064117e-02 1.55519187e-01 7.27891862e-01
4.36975211e-01 5.59493303e-01 -2.84699410e-01 -1.50242671e-01
6.65405154e-01 8.20107043e-01 -2.19297782e-01 -6.53557360e-01
-1.24479139e+00 6.17483854e-01 1.04846311e+00 3.31961393e-01
-3.92994702e-01 3.49895865e-01 2.62831897e-01 1.04187064e-01
4.93741930e-01 -3.58774722e-01 -2.11439341e-01 2.48789325e-01
-8.46092105e-01 4.09971595e-01 2.22171128e-01 1.62688243e+00
8.88621747e-01 -1.34949647e-02 -4.21937615e-01 2.49596953e-01
9.05192018e-01 1.11587834e+00 1.76221877e-01 -7.67156422e-01
3.13782066e-01 9.12963688e-01 1.38253160e-02 -9.54770625e-01
-1.55885220e-01 -6.93707943e-01 -5.88879466e-01 4.07249987e-01
-4.22625318e-02 2.73883879e-01 -9.82060134e-01 1.47782815e+00
8.22875559e-01 7.95851886e-01 -3.95641208e-01 1.23980021e+00
9.86301422e-01 4.88464743e-01 5.62510500e-03 -5.02120033e-02
1.50281048e+00 -7.46307611e-01 -6.74749136e-01 4.53746431e-02
3.88441533e-01 -8.89285922e-01 3.71191025e-01 -2.64652848e-01
-1.14486849e+00 -9.69421268e-01 -8.71803820e-01 -2.46958137e-01
-2.74351984e-01 -1.61856953e-02 4.74377275e-01 2.54839182e-01
-7.81338215e-01 4.86881077e-01 -1.06215894e+00 -2.24155620e-01
7.54361510e-01 6.48114324e-01 -6.74744695e-02 -1.03353575e-01
-8.00940931e-01 8.01761031e-01 2.94945985e-01 2.91036040e-01
-6.52834117e-01 -9.34374452e-01 -7.85519183e-01 1.04909025e-01
3.01622123e-01 -1.03268182e+00 1.24650228e+00 -3.26163471e-01
-1.13793361e+00 7.76427031e-01 -3.92738879e-01 -1.72015846e-01
1.59356579e-01 -4.29785699e-01 -2.44786829e-01 -2.68974394e-01
2.98308998e-01 7.95927405e-01 7.78124630e-01 -1.09602511e+00
-1.04673612e+00 -7.57409453e-01 -4.04174387e-01 2.75234759e-01
1.89727426e-01 1.47748321e-01 -8.98300052e-01 -3.02664429e-01
4.06501532e-01 -1.15919256e+00 -1.04741111e-01 5.58543265e-01
-2.10274592e-01 -6.22133017e-01 1.45251834e+00 -1.98629186e-01
7.74350941e-01 -2.30470324e+00 2.83942139e-03 9.49183181e-02
4.54561561e-01 3.28629255e-01 4.94181504e-03 -1.74266726e-01
1.00148760e-01 -4.40096468e-01 1.40221611e-01 -3.38627070e-01
1.33473024e-01 -1.27417400e-01 -3.07258874e-01 6.25884414e-01
4.07581598e-01 1.21937096e+00 -8.65518332e-01 -7.31579661e-01
5.85408628e-01 7.53434300e-01 -5.75584948e-01 1.81579754e-01
-2.40611166e-01 4.27072257e-01 -9.58585680e-01 9.89282668e-01
1.08495557e+00 -3.25889051e-01 -4.89981681e-01 -5.50834596e-01
-4.55616087e-01 1.95316523e-01 -1.03721797e+00 2.06096172e+00
-1.81370359e-02 4.79211599e-01 -1.51764423e-01 -3.10559332e-01
1.04252553e+00 6.30726218e-02 7.93717682e-01 -4.76155072e-01
4.73342955e-01 1.92006901e-01 -6.40792679e-03 -1.95951648e-02
3.59180957e-01 -1.49532221e-02 1.86343938e-01 7.13755265e-02
7.97016099e-02 -1.30505711e-01 -1.26663491e-01 5.59850223e-02
8.11768353e-01 3.65977854e-01 4.63019907e-02 3.93239707e-02
6.36909306e-01 1.89237058e-01 8.12950969e-01 3.51380765e-01
-5.18448114e-01 2.91887164e-01 -4.40594971e-01 -4.24501181e-01
-7.71427989e-01 -1.19836211e+00 -1.67642266e-01 5.55382609e-01
6.49759471e-01 -3.08235765e-01 -1.75100967e-01 -6.02057099e-01
5.26844621e-01 2.43758202e-01 -3.36077958e-01 -3.07806462e-01
-5.74026823e-01 -3.23115401e-02 1.75870433e-01 6.72485769e-01
5.85234165e-01 -7.50852466e-01 -6.28312349e-01 1.74388066e-01
7.18799233e-03 -1.08538294e+00 -7.77498960e-01 -2.29250342e-01
-1.25782621e+00 -8.76209438e-01 -6.64036989e-01 -6.97120428e-01
4.73907471e-01 1.06799400e+00 8.07958007e-01 3.55093360e-01
-2.07137316e-01 3.93619925e-01 -1.33105263e-01 -7.30816424e-01
1.90745547e-01 -2.44220465e-01 9.82359722e-02 -2.61743724e-01
9.62104619e-01 -3.52024585e-01 -5.06242037e-01 3.23098809e-01
-3.74560535e-01 -1.66208878e-01 6.37989283e-01 4.52790588e-01
9.73237514e-01 -3.18008929e-01 -5.22087654e-03 -1.97688714e-01
-1.09258583e-02 -4.26444709e-01 -1.06971025e+00 4.16045114e-02
-1.79606259e-01 -1.04904085e-01 -2.30742106e-03 -5.08262753e-01
-8.31913769e-01 3.33668470e-01 1.94333903e-02 -1.52202690e+00
8.65147710e-02 -7.74178356e-02 -4.01288122e-01 -3.33725750e-01
1.14199229e-01 1.89643368e-01 -2.68867481e-02 -6.51210427e-01
3.76911283e-01 4.40394223e-01 3.90669197e-01 -1.95843041e-01
1.35150170e+00 7.12128401e-01 2.90805586e-02 -3.47036362e-01
-8.26712787e-01 -8.12532663e-01 -6.27372861e-01 -4.24334824e-01
9.38316584e-01 -1.33243930e+00 -1.25856864e+00 2.16557801e-01
-1.37796640e+00 3.51773888e-01 -3.59709769e-01 8.19986284e-01
-4.12904412e-01 2.22451687e-01 -2.80676246e-01 -7.32665539e-01
-5.26545703e-01 -1.15941513e+00 1.51055801e+00 4.68164444e-01
3.21074843e-01 -5.69903433e-01 9.01695192e-02 1.97095852e-02
3.01685005e-01 5.59132434e-02 2.47217089e-01 -4.60286111e-01
-1.32995141e+00 -2.19038367e-01 -5.62305033e-01 -1.20988384e-01
2.61170894e-01 -1.82344869e-01 -9.23296154e-01 -4.75472510e-01
2.32207358e-01 2.54012078e-01 7.89615810e-01 4.31895167e-01
9.67261493e-01 2.09739566e-01 -8.44275355e-01 7.69534409e-01
1.36356580e+00 1.59890071e-01 1.80193633e-01 -3.42614651e-02
1.04555666e+00 4.80677783e-02 1.01657522e+00 3.70130837e-01
5.18334568e-01 9.31339979e-01 5.51137984e-01 1.32816419e-01
-3.43885690e-01 -2.51670122e-01 2.69434482e-01 9.35969591e-01
-6.07033400e-03 1.72736779e-01 -7.54736006e-01 5.03850102e-01
-1.91537905e+00 -1.22345424e+00 -2.77235895e-01 2.05901194e+00
3.93738389e-01 1.74509674e-01 5.91779239e-02 -3.14083695e-01
1.07303929e+00 -4.86407131e-02 -8.05418789e-01 4.02988881e-01
2.22767100e-01 -6.56003281e-02 4.74487185e-01 1.39181510e-01
-1.06311297e+00 9.45784628e-01 4.75304508e+00 6.44987524e-01
-1.17430019e+00 2.93281555e-01 -3.22921693e-01 -3.64920795e-01
-2.86357794e-02 -1.06641818e-02 -1.34185135e+00 7.19165027e-01
5.64796627e-01 -3.32015336e-01 4.40376960e-02 9.83246446e-01
-6.09068237e-02 4.20426339e-01 -1.10248172e+00 1.16794324e+00
-1.71874724e-02 -1.26370966e+00 -2.01971196e-02 1.38288915e-01
3.46871346e-01 6.20459318e-01 4.51599471e-02 3.22285622e-01
2.46477619e-01 -2.56416380e-01 9.00515318e-01 6.25904858e-01
3.65301400e-01 -5.90107024e-01 5.85928440e-01 2.34405547e-01
-1.98156893e+00 1.18712358e-01 -6.56208396e-01 3.00073028e-01
2.72950202e-01 6.30034149e-01 -6.18870199e-01 7.34552324e-01
1.09014654e+00 1.22862005e+00 -5.26751876e-01 1.60101712e+00
2.31715471e-01 -3.08856964e-02 -6.17472947e-01 4.82321270e-02
1.61353961e-01 -2.40054373e-02 9.60822403e-01 9.15391862e-01
7.20978856e-01 2.35573456e-01 5.20407856e-01 1.11588657e+00
9.11503658e-03 -2.16872662e-01 -8.02030623e-01 2.08898485e-01
8.15107524e-01 1.14665020e+00 -4.72521752e-01 -3.97984803e-01
-5.99380851e-01 5.75284362e-01 1.62226453e-01 -6.93220347e-02
-1.08478022e+00 -3.11895311e-01 1.04184580e+00 -4.33398923e-03
8.55852485e-01 -3.66534591e-01 -3.36799696e-02 -1.13248682e+00
3.56480926e-02 -4.17549074e-01 2.68034875e-01 -9.22392249e-01
-1.29020011e+00 4.28917587e-01 -2.67708283e-02 -1.99146044e+00
2.49212101e-01 -5.01150489e-01 -4.94780719e-01 1.01161730e+00
-1.75954807e+00 -1.40851140e+00 -6.28356516e-01 8.09997201e-01
4.20541257e-01 -4.59184907e-02 1.59806907e-01 6.52840674e-01
-2.49833390e-01 2.34530270e-01 -2.38151252e-01 5.89862131e-02
5.07089019e-01 -7.91837096e-01 6.93504870e-01 7.12415457e-01
1.00167133e-01 7.34535873e-01 2.95733720e-01 -1.04217947e+00
-1.75392842e+00 -1.34278917e+00 7.84239531e-01 -6.19034350e-01
4.00762349e-01 -2.41137341e-01 -1.07758236e+00 7.76265264e-01
-4.83995862e-03 4.37753916e-01 4.81095970e-01 -1.13626085e-01
-2.96396792e-01 -1.46149471e-01 -1.01609826e+00 3.66824657e-01
1.49189150e+00 -3.66143256e-01 -7.40043223e-01 3.25819999e-01
1.08554816e+00 -7.84708440e-01 -8.85306239e-01 5.88564634e-01
4.23085719e-01 -6.41565323e-01 1.21145272e+00 -4.03274953e-01
-3.76174808e-01 -1.12112796e+00 -1.51608765e-01 -9.42674458e-01
-9.06931460e-01 -2.08793759e-01 -4.45647299e-01 1.07484770e+00
-2.01274544e-01 -4.60282534e-01 7.46820927e-01 3.26160610e-01
-5.27458012e-01 -4.49093521e-01 -1.07643366e+00 -9.53175664e-01
-3.29916388e-01 -4.38175648e-01 9.46217239e-01 6.82161748e-01
-5.85487366e-01 3.30613226e-01 -1.30707085e-01 5.98149538e-01
1.08059263e+00 6.42992258e-01 8.67021382e-01 -1.62962723e+00
1.52971953e-01 -4.81061101e-01 -8.75476360e-01 -1.40913701e+00
2.00763211e-01 -1.21978104e+00 2.06609264e-01 -1.24805558e+00
1.72279328e-01 -6.49460375e-01 -2.89080322e-01 3.96916002e-01
-2.98567176e-01 2.35322729e-01 8.47000301e-01 4.67584074e-01
-7.96286821e-01 8.58804345e-01 1.32558620e+00 -3.57838213e-01
-9.93483365e-02 1.71144158e-01 -4.94265258e-01 5.22798479e-01
6.42553926e-01 -7.95360446e-01 -2.70624608e-02 -7.69432068e-01
-4.14562762e-01 -1.99523240e-01 6.96473897e-01 -1.34106112e+00
7.25036860e-01 2.95075849e-02 7.21961558e-01 -1.51401937e+00
6.20598614e-01 -1.42224813e+00 3.29859674e-01 7.71256566e-01
2.50323415e-01 2.65512735e-01 4.74429786e-01 6.20206356e-01
-1.31338492e-01 6.48439974e-02 6.77780569e-01 -1.29518136e-02
-8.13297868e-01 9.70635891e-01 2.76624888e-01 -3.40990275e-01
1.13262308e+00 -3.59896958e-01 -1.25671178e-01 2.07746670e-01
-3.84227961e-01 6.29383504e-01 5.48232853e-01 7.26692796e-01
8.43235075e-01 -1.96016669e+00 -6.38892472e-01 2.87183940e-01
1.38354003e-01 2.83483148e-01 3.08119565e-01 9.20681834e-01
1.52951609e-02 6.06106460e-01 -4.19491827e-01 -1.30570257e+00
-1.50020945e+00 8.38839710e-01 2.93978989e-01 1.23605907e-01
-8.37255001e-01 7.92144239e-01 3.79803419e-01 -2.93630898e-01
1.33189484e-01 -5.90933383e-01 7.42809996e-02 -2.75401950e-01
5.72702587e-01 2.37411112e-01 -7.68336281e-02 -8.48603368e-01
-7.29090452e-01 1.21542335e+00 -1.18679620e-01 3.32895547e-01
9.94216859e-01 -1.14531070e-01 1.64599478e-01 2.56255567e-01
1.16723800e+00 -1.90297291e-01 -1.37332761e+00 -6.80723310e-01
-1.81849837e-01 -9.86797214e-01 3.56425405e-01 -2.18437493e-01
-1.41115022e+00 9.19516444e-01 1.03384566e+00 -1.60613105e-01
1.05821955e+00 1.30256057e-01 1.03204870e+00 3.16437304e-01
5.47742724e-01 -3.67166102e-01 -2.31694072e-01 6.02940917e-01
7.12147117e-01 -1.36423075e+00 1.55187279e-01 -5.18695354e-01
-3.56360614e-01 6.34501517e-01 7.39044964e-01 -3.00013512e-01
8.33697915e-01 2.52827287e-01 -1.08004972e-01 -3.90412241e-01
-8.28532994e-01 -7.63595641e-01 5.21153986e-01 6.80853486e-01
1.50601670e-01 -2.47630179e-01 2.14530051e-01 3.86634290e-01
1.03832975e-01 2.02108115e-01 -3.51779014e-01 1.08306110e+00
-7.59705901e-01 -8.70522559e-01 -5.32439291e-01 3.59039843e-01
1.15636382e-02 1.24860696e-01 -1.84978843e-01 7.26195812e-01
3.95401537e-01 5.64305723e-01 4.29412007e-01 -5.48300564e-01
6.33285999e-01 -1.75463572e-01 7.16370761e-01 -5.84766984e-01
-7.23948777e-01 1.87750623e-01 -5.59359193e-01 -6.44321918e-01
-5.26039958e-01 -9.38957334e-01 -1.55031621e+00 -4.77525532e-01
-7.71959722e-01 7.29157329e-02 7.46550500e-01 6.54642701e-01
7.46094644e-01 5.65144360e-01 5.69631398e-01 -1.25254083e+00
-3.57287496e-01 -7.82974660e-01 -1.41861022e-01 2.46825680e-01
4.49926734e-01 -1.23205400e+00 7.99474716e-02 -1.58369437e-01] | [6.5990495681762695, -2.380692958831787] |
9d4b9cc1-7e8b-4b05-ba4f-d4f172b1f789 | simulation-intelligence-towards-a-new | 2112.03235 | null | https://arxiv.org/abs/2112.03235v2 | https://arxiv.org/pdf/2112.03235v2.pdf | Simulation Intelligence: Towards a New Generation of Scientific Methods | The original "Seven Motifs" set forth a roadmap of essential methods for the field of scientific computing, where a motif is an algorithmic method that captures a pattern of computation and data movement. We present the "Nine Motifs of Simulation Intelligence", a roadmap for the development and integration of the essential algorithms necessary for a merger of scientific computing, scientific simulation, and artificial intelligence. We call this merger simulation intelligence (SI), for short. We argue the motifs of simulation intelligence are interconnected and interdependent, much like the components within the layers of an operating system. Using this metaphor, we explore the nature of each layer of the simulation intelligence operating system stack (SI-stack) and the motifs therein: (1) Multi-physics and multi-scale modeling; (2) Surrogate modeling and emulation; (3) Simulation-based inference; (4) Causal modeling and inference; (5) Agent-based modeling; (6) Probabilistic programming; (7) Differentiable programming; (8) Open-ended optimization; (9) Machine programming. We believe coordinated efforts between motifs offers immense opportunity to accelerate scientific discovery, from solving inverse problems in synthetic biology and climate science, to directing nuclear energy experiments and predicting emergent behavior in socioeconomic settings. We elaborate on each layer of the SI-stack, detailing the state-of-art methods, presenting examples to highlight challenges and opportunities, and advocating for specific ways to advance the motifs and the synergies from their combinations. Advancing and integrating these technologies can enable a robust and efficient hypothesis-simulation-analysis type of scientific method, which we introduce with several use-cases for human-machine teaming and automated science. | ['Johann Brehmer', 'David Krakauer', 'Avi Pfeffer', 'Stephan Zheng', 'Samuel Assefa', 'Manuela Veloso', 'Adi Hanuka', 'Haruko Wainwright', 'Jiaxin Zhang', 'Kyle Cranmer', 'Jakob Macke', 'Peter L. McMahon', 'Erik Peterson', 'Olexandr Isayev', 'Carina Prunkl', 'Atılım Güneş Baydin', 'Kamil Rocki', 'Sanjay Choudry', 'Anima Anandkumar', 'Tim Mattson', 'Justin Gottschlich', 'Brooks Paige', 'Hector Zenil', 'Alexander Lavin'] | 2021-12-06 | null | null | null | null | ['probabilistic-programming'] | ['methodology'] | [ 1.47411689e-01 2.45690495e-02 -8.07166621e-02 2.92987287e-01
-1.29989132e-01 -7.02373326e-01 1.18188381e+00 2.35412776e-01
4.50308919e-02 6.86993420e-01 1.83584020e-01 -9.60734725e-01
-7.15195775e-01 -1.02084041e+00 -8.86224568e-01 -1.01738644e+00
-3.92133921e-01 8.32802117e-01 -1.64429203e-01 -1.92976385e-01
4.57462221e-01 6.01477027e-01 -1.83165038e+00 -1.02746651e-01
1.00342643e+00 5.51328301e-01 2.46728897e-01 7.80123353e-01
-1.27463445e-01 7.22737014e-01 -3.56095433e-01 1.84851736e-01
-1.48285717e-01 -5.81013799e-01 -5.39376020e-01 -5.86353958e-01
-6.32586420e-01 3.91111493e-01 -1.88022591e-02 7.58137405e-01
2.70844489e-01 5.07490896e-02 7.12307632e-01 -1.46996176e+00
-4.85634893e-01 3.85803163e-01 -3.86543274e-01 6.94925189e-02
3.04250777e-01 7.94796228e-01 2.61329651e-01 -5.23996174e-01
8.55097234e-01 1.27803314e+00 6.44075811e-01 1.87938705e-01
-1.26546752e+00 -4.45269197e-01 4.34611216e-02 -8.16958621e-02
-1.27885699e+00 -1.89288199e-01 3.26876700e-01 -8.72484267e-01
1.18894207e+00 5.53903103e-01 1.11986995e+00 1.04542542e+00
4.45240200e-01 2.02646032e-01 1.29083467e+00 -4.42307383e-01
7.05476940e-01 -5.71076572e-02 -1.87390149e-02 5.05490363e-01
4.58415121e-01 7.89322197e-01 -8.26074660e-01 -6.07749820e-01
7.42972136e-01 3.26777510e-02 1.96433663e-01 4.59656045e-02
-1.67981923e+00 6.48246884e-01 -2.22270489e-02 1.95139542e-01
-6.17011845e-01 6.26159549e-01 3.73076610e-02 -1.70347139e-01
4.43424076e-01 7.57728696e-01 -5.29354155e-01 -3.76230687e-01
-8.23454738e-01 6.65585399e-01 1.10686255e+00 4.53317255e-01
5.51337123e-01 2.48936057e-01 2.23733947e-01 2.75190324e-01
7.69805789e-01 6.27466917e-01 -8.81060660e-02 -1.32538092e+00
-2.82341361e-01 5.28850198e-01 4.56535429e-01 -8.09184730e-01
-6.30420744e-01 -4.39805865e-01 -7.34815061e-01 3.68980974e-01
2.52089232e-01 -4.83962208e-01 -4.97737467e-01 1.67397642e+00
6.79967582e-01 6.31627262e-01 -1.11906469e-01 6.85429633e-01
4.46487874e-01 9.66182768e-01 3.96505147e-01 -2.62939692e-01
1.41819668e+00 -5.96786499e-01 -3.70372534e-01 1.48067474e-01
2.63544738e-01 -4.56757396e-01 8.69123936e-01 2.80764580e-01
-1.26208532e+00 -1.88372806e-01 -9.39499080e-01 4.55086529e-01
-7.55713582e-01 -5.66889286e-01 9.82109785e-01 6.03804469e-01
-1.10734308e+00 5.58893204e-01 -1.20254612e+00 -4.45479900e-01
-2.59887485e-04 2.13456929e-01 2.44074792e-01 4.84625787e-01
-1.03902781e+00 1.01785612e+00 -1.11704007e-01 -5.26334047e-02
-1.09385812e+00 -1.43395209e+00 -4.37361240e-01 1.18106104e-01
3.33070233e-02 -1.46560001e+00 7.06839144e-01 -5.98162055e-01
-1.61290395e+00 6.51451290e-01 -3.08955461e-01 -1.43783092e-01
4.20036763e-01 2.81651825e-01 -3.05455744e-01 -2.81651884e-01
-9.35376715e-03 3.47620219e-01 7.00866953e-02 -1.65578258e+00
-3.85189772e-01 -4.62809920e-01 -2.46881098e-01 1.14618344e-02
2.60365397e-01 3.44794720e-01 1.01503752e-01 -5.94439268e-01
-6.34385049e-02 -9.73610103e-01 -6.36474788e-01 -2.26442501e-01
-2.38848850e-01 -2.19025150e-01 6.39520824e-01 -4.32734996e-01
1.11818504e+00 -1.43533885e+00 5.69372475e-01 2.53050447e-01
3.08817238e-01 -5.97664900e-02 2.29878142e-01 9.82112586e-01
1.12258554e-01 4.68650192e-01 -3.53227466e-01 2.32370738e-02
3.15634698e-01 -1.43432006e-01 -3.02480578e-01 3.18514347e-01
2.17967834e-02 1.01044476e+00 -1.19455719e+00 -6.68813754e-03
3.53208005e-01 4.92480546e-01 -3.14783692e-01 1.77672341e-01
-8.10612142e-01 8.68881226e-01 -6.47452652e-01 8.74240816e-01
4.00750518e-01 -3.95937085e-01 4.79420930e-01 2.86207974e-01
-7.98245609e-01 2.47141048e-01 -1.08317006e+00 1.38512945e+00
-4.56855506e-01 2.15343967e-01 3.85529995e-01 -8.71144295e-01
7.37394750e-01 2.40044937e-01 6.15257919e-01 -3.46223891e-01
-1.80933729e-01 1.97781861e-01 2.13743865e-01 -5.73737979e-01
1.37713313e-01 -7.43659213e-02 -9.06852037e-02 1.02057242e+00
-2.88979262e-01 -5.59879899e-01 -3.32071967e-02 -4.94400300e-02
1.20429957e+00 7.47915089e-01 1.63856581e-01 -7.45119631e-01
9.60707143e-02 3.91580015e-01 4.70709056e-01 8.86518955e-01
-2.19696447e-01 -4.73627411e-02 4.66066301e-01 -6.87417507e-01
-1.11515284e+00 -1.36204219e+00 -1.80239052e-01 9.90002275e-01
-5.57849146e-02 -4.08384979e-01 -7.80712545e-01 1.38237119e-01
3.58393788e-01 8.42673600e-01 -7.48217702e-01 5.44798672e-02
-3.73483777e-01 -1.60530519e+00 5.95982254e-01 7.29493722e-02
-5.32369949e-02 -9.63582456e-01 -8.15073907e-01 2.23988906e-01
1.41004547e-01 -5.74217856e-01 4.78812367e-01 2.22667694e-01
-7.40678668e-01 -7.43672609e-01 -1.94244519e-01 -2.24305883e-01
3.39864135e-01 8.42432827e-02 1.27733946e+00 2.84617394e-01
-3.69168878e-01 3.54984820e-01 1.01063900e-01 -7.16751635e-01
-8.00990760e-01 -3.94254804e-01 5.44633448e-01 -5.77225208e-01
1.02001671e-02 -1.21185160e+00 -7.30875731e-01 1.43531546e-01
-8.40787649e-01 4.46151644e-01 2.37514287e-01 6.71519995e-01
3.89537752e-01 -2.09637120e-01 6.11843407e-01 -4.73206997e-01
4.53410774e-01 -1.05821168e+00 -9.23179984e-01 4.63652074e-01
-8.10593724e-01 -3.26509820e-03 4.73152071e-01 -2.05648206e-02
-8.79814625e-01 -3.07999790e-01 1.33894682e-01 -5.89392819e-02
-1.46247283e-01 7.79240310e-01 -8.11972916e-02 -3.35941501e-02
4.83605653e-01 2.19283089e-01 3.87507439e-01 -3.30969721e-01
4.92386729e-01 4.23599869e-01 2.86551595e-01 -1.23859322e+00
5.76166391e-01 7.23248601e-01 4.26541597e-01 -8.37119401e-01
-1.15884356e-02 3.21923167e-01 -2.32810885e-01 -5.23512602e-01
7.62672126e-01 -4.87309188e-01 -1.26225603e+00 4.69660014e-01
-1.12999928e+00 -6.42144859e-01 -3.37777615e-01 3.81232411e-01
-5.97081065e-01 -7.79817477e-02 -3.88344884e-01 -1.04630089e+00
-1.08697824e-01 -1.38253510e+00 9.98436868e-01 5.06239831e-01
-3.74230415e-01 -1.29498565e+00 7.30904996e-01 3.44749153e-01
6.77690268e-01 7.34706998e-01 1.16936755e+00 -2.81082213e-01
-6.47413433e-01 3.69086683e-01 -2.01379545e-02 -4.32396561e-01
-2.64695078e-01 8.40667546e-01 -8.93895030e-01 2.62894005e-01
-7.08719045e-02 8.88908580e-02 2.80703068e-01 6.50043607e-01
1.09781659e+00 -1.87241986e-01 -6.97840393e-01 8.40755165e-01
1.35972357e+00 4.06847149e-01 5.12842357e-01 4.81706738e-01
5.26300251e-01 1.01656651e+00 2.35176653e-01 6.59396291e-01
6.53299212e-01 4.74483341e-01 2.43865937e-01 -1.38666242e-01
1.79450229e-01 -2.21692339e-01 4.67154831e-01 1.00182927e+00
-4.35736090e-01 -9.36078206e-02 -1.50023305e+00 2.42650852e-01
-2.08017254e+00 -1.05005026e+00 -4.10914510e-01 2.13766432e+00
7.40607202e-01 -2.70786107e-01 1.22686237e-01 -4.66584414e-01
5.67421734e-01 -4.00922000e-02 -7.33996332e-01 -6.34570539e-01
-7.61775076e-02 2.75172442e-01 3.08263361e-01 4.18610424e-01
-5.55743217e-01 8.14071953e-01 7.79778671e+00 7.44444072e-01
-8.95797551e-01 2.15734661e-01 8.10106277e-01 -1.61757860e-02
-8.79751921e-01 4.87353504e-01 -4.50500131e-01 8.42731774e-01
1.55497706e+00 -3.37952107e-01 9.28419590e-01 3.68738323e-01
8.70755911e-01 -4.23372388e-01 -7.53739476e-01 2.43839845e-01
-5.59617639e-01 -2.14540958e+00 -3.21871713e-02 2.24726200e-01
8.99160922e-01 4.97581571e-01 -2.91153044e-01 -1.86225504e-01
1.08213055e+00 -1.39186263e+00 8.45226526e-01 9.45950210e-01
5.30580699e-01 -4.43788946e-01 6.73058257e-02 3.63179564e-01
-7.77145267e-01 1.69248402e-01 1.53445676e-01 -4.42400903e-01
3.42527986e-01 8.29078019e-01 -1.82781875e-01 7.38295734e-01
8.44287634e-01 3.49240601e-01 9.15659517e-02 6.76626503e-01
9.56162959e-02 7.69800365e-01 -6.29603624e-01 -3.22719395e-01
-3.23575959e-02 -8.64219546e-01 9.04541016e-01 1.28580928e+00
3.07005852e-01 7.77516887e-02 -1.87518090e-01 1.71499133e+00
6.09669566e-01 -7.27346599e-01 -7.69343555e-01 -3.92341524e-01
9.43598151e-01 1.12104785e+00 -1.00439727e+00 -2.16051444e-01
-4.71777171e-02 8.57800469e-02 -2.10770145e-01 4.24689710e-01
-9.64923263e-01 -1.10621631e-01 1.00399637e+00 -3.44160013e-04
-2.22121269e-01 -6.29799306e-01 -1.01932335e+00 -8.07478428e-01
-5.87125957e-01 -1.05765522e+00 8.38046297e-02 -9.37116206e-01
-1.09431875e+00 -4.72985208e-02 1.33614019e-01 -5.63488483e-01
-3.62432361e-01 -5.16253710e-01 -8.78323019e-01 8.88904870e-01
-1.04941738e+00 -1.22996032e+00 -1.55031323e-01 2.35444512e-02
-1.22712374e-01 -5.81796654e-02 9.46924388e-01 -1.27383307e-01
-8.03041160e-01 -2.16242090e-01 7.90453196e-01 -6.77770257e-01
-7.69846439e-02 -1.07561207e+00 6.32470012e-01 6.51872635e-01
-3.63841265e-01 9.07373607e-01 9.64581132e-01 -9.70332444e-01
-2.33965063e+00 -7.47029722e-01 5.40041685e-01 -7.28801966e-01
1.16697979e+00 -3.66329044e-01 -5.03116548e-01 4.14064914e-01
5.32520302e-02 -3.86918455e-01 6.54572546e-01 2.16520086e-01
-2.43250895e-02 1.18462890e-01 -1.02706766e+00 7.91532338e-01
1.05637157e+00 -3.20311189e-01 -2.68275708e-01 6.16792381e-01
7.42853403e-01 -9.65461284e-02 -1.03651512e+00 2.93286860e-01
8.20511699e-01 -8.55068147e-01 1.16679478e+00 -5.99237263e-01
5.00836968e-01 -7.37285674e-01 -9.79002286e-03 -1.06009555e+00
-2.47252807e-01 -1.19838190e+00 -3.33417743e-01 9.69670892e-01
2.76645273e-01 -9.22665954e-01 3.87847096e-01 6.81047559e-01
-2.36894861e-01 -7.78267622e-01 -8.34652424e-01 -6.23519540e-01
4.23606336e-01 -5.08148909e-01 8.31209183e-01 1.12153065e+00
3.63598377e-01 -1.55567884e-01 2.62938827e-01 1.59929663e-01
7.23219216e-01 1.28512427e-01 6.91889226e-01 -1.06909370e+00
-4.87193376e-01 -8.70010495e-01 -1.02015577e-01 -2.78346509e-01
6.68790191e-03 -7.35586286e-01 -1.88530728e-01 -1.45823359e+00
4.78102744e-01 -7.65127182e-01 8.09261799e-02 1.29539603e-02
7.00204913e-03 -2.40046337e-01 -5.24601489e-02 3.40091765e-01
-2.37277374e-01 4.72860396e-01 8.49573851e-01 2.78337419e-01
-8.31795409e-02 -4.14302766e-01 -8.80552769e-01 7.34061301e-01
6.33289814e-01 -3.88548553e-01 -7.74963573e-02 -2.64098883e-01
5.94715834e-01 1.87953725e-01 8.82745385e-01 -7.80161977e-01
4.37707752e-01 -9.34657753e-01 1.45832106e-01 -2.45403871e-01
2.84643546e-02 -3.26653451e-01 1.03193259e+00 8.47377717e-01
7.87340626e-02 3.40832412e-01 3.37009549e-01 3.79168570e-01
3.09663385e-01 1.34426102e-01 5.45593679e-01 -2.28954673e-01
-2.36635029e-01 -1.03765622e-01 -8.86769533e-01 -5.92541136e-02
1.12990367e+00 -6.27192408e-02 -7.57247925e-01 -7.91621134e-02
-5.74014306e-01 4.04875726e-01 9.37495351e-01 7.45771080e-02
1.06845900e-01 -6.72790229e-01 -7.82484591e-01 -1.21448271e-01
-4.07648116e-01 -1.40965268e-01 1.66910201e-01 9.07676518e-01
-7.92584836e-01 3.36508304e-01 -1.02689058e-01 -5.31155825e-01
-3.37532818e-01 3.39390159e-01 6.00751102e-01 -2.31703013e-01
-2.09731534e-01 5.90835035e-01 1.34147003e-01 -7.25906074e-01
-2.02439517e-01 -5.37859090e-02 2.43461251e-01 -2.77288824e-01
3.94774318e-01 7.96142936e-01 -2.18203351e-01 -1.11169301e-01
-4.34067905e-01 4.34990108e-01 7.60400832e-01 -5.06051123e-01
1.69665384e+00 -8.61460045e-02 -8.36157918e-01 6.79573178e-01
3.73467714e-01 -2.14569658e-01 -1.09150600e+00 4.79994029e-01
1.11595243e-01 9.43005756e-02 1.35173842e-01 -1.32251847e+00
-4.58128452e-01 5.97754598e-01 1.90537080e-01 4.10684258e-01
7.24461436e-01 -8.02125931e-02 3.20258111e-01 -3.92432325e-02
5.11553645e-01 -1.02327275e+00 -5.28391898e-01 3.69441420e-01
8.29829872e-01 -6.72734082e-01 2.46513367e-01 -1.84706468e-02
-1.40147597e-01 1.01060951e+00 3.40710044e-01 -1.16525358e-02
1.04661179e+00 9.04656649e-01 -4.03641552e-01 -6.19491935e-01
-1.17367911e+00 5.69307618e-02 -2.58413523e-01 7.21179366e-01
3.78211260e-01 3.55281323e-01 -3.48758459e-01 8.34885895e-01
-7.23121837e-02 2.22641304e-01 2.67250061e-01 1.00574136e+00
-2.46038258e-01 -1.12599409e+00 -8.37865770e-01 3.97190988e-01
-4.16520275e-02 -1.58281863e-01 -3.92829210e-01 6.67001843e-01
2.85923690e-01 9.68601465e-01 1.49320304e-01 -3.61801535e-01
-1.54876232e-01 4.07229736e-03 3.30014229e-01 -2.42690921e-01
-1.02167428e+00 -2.30329469e-01 9.02337730e-02 -5.97941399e-01
-3.76962453e-01 -1.00659788e+00 -1.24080265e+00 -1.03264844e+00
1.22314453e-01 3.46962303e-01 1.20050716e+00 1.00530446e+00
1.05748880e+00 6.29482925e-01 3.83556515e-01 -1.37338758e+00
-7.99116269e-02 -5.24250269e-01 -3.34930867e-01 -4.30282772e-01
-1.44025534e-01 -7.01673567e-01 -5.04410148e-01 -6.03953749e-02] | [6.0532965660095215, 4.12118673324585] |
61390b09-a7c3-4a63-9665-d0d43129c64e | robust-remote-sensing-scene-classification | 2301.05858 | null | https://arxiv.org/abs/2301.05858v1 | https://arxiv.org/pdf/2301.05858v1.pdf | Robust Remote Sensing Scene Classification with Multi-View Voting and Entropy Ranking | Deep convolutional neural networks have been widely used in scene classification of remotely sensed images. In this work, we propose a robust learning method for the task that is secure against partially incorrect categorization of images. Specifically, we remove and correct errors in the labels progressively by iterative multi-view voting and entropy ranking. At each time step, we first divide the training data into disjoint parts for separate training and voting. The unanimity in the voting reveals the correctness of the labels, so that we can train a strong model with only the images with unanimous votes. In addition, we adopt entropy as an effective measure for prediction uncertainty, in order to partially recover labeling errors by ranking and selection. We empirically demonstrate the superiority of the proposed method on the WHU-RS19 dataset and the AID dataset. | ['George Hadjichristofi', 'Min Gan', 'Tao Wang', 'Jinyang Wang'] | 2023-01-14 | null | null | null | null | ['scene-classification'] | ['computer-vision'] | [ 3.70508105e-01 4.17441353e-02 -4.22780812e-02 -7.91579485e-01
-6.19418919e-01 -7.29552746e-01 3.52705389e-01 1.77861765e-01
-4.38369185e-01 6.62390530e-01 1.10858455e-01 -3.31882626e-01
-2.28969529e-01 -8.87038231e-01 -5.89738965e-01 -7.93305933e-01
3.08908731e-01 2.28484407e-01 -1.55697256e-01 1.74703047e-01
2.98037380e-01 4.37920719e-01 -1.40735936e+00 4.75919813e-01
8.80311668e-01 1.41264212e+00 -1.72436222e-01 3.17243695e-01
4.89380211e-01 1.15148497e+00 -2.24143893e-01 -5.71233153e-01
5.75314343e-01 -1.37829795e-01 -9.71507072e-01 1.71413288e-01
6.04425550e-01 -8.82286370e-01 -2.33880341e-01 1.49748719e+00
3.73238832e-01 -1.08931288e-01 8.87557805e-01 -1.07705748e+00
-4.24696624e-01 6.46536469e-01 -6.67917311e-01 -9.89235342e-02
-3.06856215e-01 -1.90884262e-01 1.34553123e+00 -1.05034280e+00
3.87259126e-01 9.13022697e-01 5.97587287e-01 4.36748147e-01
-1.01100016e+00 -8.47293198e-01 6.49205521e-02 3.51089150e-01
-1.42290056e+00 -3.29467386e-01 6.67618990e-01 -6.41041100e-01
1.19193032e-01 4.40340281e-01 2.76410133e-01 6.32933497e-01
-2.40979604e-02 7.83922493e-01 1.37986445e+00 -1.49185419e-01
3.07942122e-01 9.14423987e-02 2.81147957e-01 6.02550507e-01
3.50136220e-01 1.20787501e-01 -2.32338184e-03 -3.62885177e-01
9.32192877e-02 4.29801852e-01 -3.32303107e-01 -3.32580805e-01
-9.40631986e-01 9.34997201e-01 9.05384541e-01 -1.90436602e-01
-3.18397790e-01 -7.95443803e-02 3.93398762e-01 9.08802226e-02
6.39118493e-01 2.61392027e-01 -6.53870046e-01 7.64212549e-01
-1.04811919e+00 -5.71790189e-02 3.29240859e-01 3.52130473e-01
8.56658459e-01 -2.24411801e-01 -1.32593676e-01 8.35590839e-01
3.29091311e-01 6.16792202e-01 9.40439180e-02 -1.16193748e+00
4.04990047e-01 7.32743144e-01 1.71138421e-01 -1.19306457e+00
-3.50499272e-01 -4.05039728e-01 -1.38132966e+00 3.83488953e-01
1.54760763e-01 -9.88835022e-02 -1.03797412e+00 1.56952274e+00
1.79103971e-01 -2.60574464e-02 1.11999393e-01 8.38488877e-01
7.59086490e-01 4.83065724e-01 3.07978820e-02 -4.09318954e-02
1.19914687e+00 -6.13751292e-01 -2.67914832e-01 -1.50417432e-01
3.42827320e-01 -2.93381810e-01 5.21837950e-01 4.41867173e-01
-5.07709205e-01 -4.62949425e-01 -1.04431248e+00 1.94833562e-01
-1.44865364e-01 7.10528016e-01 3.25799942e-01 3.67368877e-01
-7.85549641e-01 6.27218902e-01 -5.29618263e-01 2.41250008e-01
8.36298704e-01 1.64874375e-01 -5.67880750e-01 -1.27078086e-01
-1.22474110e+00 8.30162227e-01 7.01241493e-01 4.04420286e-01
-9.78521585e-01 -3.42433482e-01 -8.17319572e-01 4.01132643e-01
2.96661735e-01 -2.08923429e-01 1.04684114e+00 -1.02953196e+00
-8.70982945e-01 1.05368781e+00 1.67244866e-01 -4.88191038e-01
8.87021482e-01 -1.40052751e-01 -1.27787128e-01 -1.00002009e-02
5.95207959e-02 6.54665172e-01 9.01602745e-01 -1.68286026e+00
-7.73857951e-01 -5.68193018e-01 9.72396582e-02 2.40764201e-01
-2.86029071e-01 -2.82847971e-01 1.80109069e-01 -4.36238557e-01
5.55117548e-01 -1.07994056e+00 -1.68343306e-01 1.14955336e-01
-5.28958559e-01 4.21113195e-03 7.01550901e-01 -9.40172434e-01
9.06553566e-01 -2.21095443e+00 -1.18756004e-01 5.24398506e-01
4.38815504e-01 2.57293344e-01 -8.15648139e-02 -7.25759044e-02
-5.90258464e-02 4.69101280e-01 -5.92289925e-01 -2.06428692e-02
-2.40236044e-01 1.42287195e-01 -7.33921051e-01 5.15251338e-01
4.86166552e-02 5.11295795e-01 -7.37965107e-01 -2.57994920e-01
1.90581992e-01 1.62674353e-01 -4.43482757e-01 8.12109858e-02
5.31215221e-03 3.40385139e-01 -4.71084923e-01 4.19568598e-01
1.23770881e+00 -4.07820493e-01 3.32112908e-01 -3.87229383e-01
2.26021975e-01 4.87910695e-02 -1.18351007e+00 8.77024591e-01
-3.28789771e-01 4.64229405e-01 -6.57897890e-02 -9.35452640e-01
8.82662535e-01 3.80153060e-02 4.41391140e-01 -2.73959845e-01
1.76257342e-01 3.09424102e-01 -3.18112403e-01 -3.73180866e-01
4.74258333e-01 -3.28390524e-02 -1.05186529e-01 4.92983878e-01
-9.16160867e-02 -6.98119625e-02 -2.87977278e-01 4.57102917e-02
5.62049150e-01 -1.75408125e-01 6.08842075e-01 3.01685836e-02
7.13053226e-01 -1.63516611e-01 7.82933295e-01 7.53752828e-01
-4.15994823e-01 7.88240016e-01 2.23977715e-01 -8.45370293e-01
-1.06830323e+00 -7.34928906e-01 -3.16807926e-01 9.43350911e-01
3.60371947e-01 4.80494164e-02 -6.14761770e-01 -1.33363557e+00
1.61609933e-01 7.34796584e-01 -6.38353050e-01 -1.37685671e-01
-2.07639396e-01 -8.75838280e-01 5.44726372e-01 4.18820351e-01
1.02516341e+00 -9.32520032e-01 -7.11561084e-01 -1.81351051e-01
-5.74524045e-01 -8.38567615e-01 -4.48061414e-02 2.66063124e-01
-7.22764790e-01 -1.29935277e+00 -3.36073935e-01 -4.47919339e-01
9.41686749e-01 5.03093064e-01 7.88833439e-01 3.26577395e-01
2.40325511e-01 -2.05766216e-01 -3.82085323e-01 -1.19028546e-01
-4.43608731e-01 -4.26294282e-02 -6.41267449e-02 1.89327165e-01
2.78756529e-01 -2.71637678e-01 -7.56060839e-01 3.19656700e-01
-1.03470874e+00 -2.55897772e-02 4.91336256e-01 1.13868260e+00
6.42201185e-01 2.90258467e-01 4.30442214e-01 -1.02054226e+00
2.02301309e-01 -4.54815567e-01 -6.75396502e-01 5.09187996e-01
-6.41379833e-01 6.74622729e-02 6.03653908e-01 1.16156347e-01
-1.08251655e+00 4.05154258e-01 -1.67297855e-01 -3.31375062e-01
-3.03403527e-01 6.33217812e-01 3.04220710e-02 -1.47380456e-01
6.46185994e-01 1.72860280e-01 -4.30255622e-01 -2.86054432e-01
2.68578589e-01 1.04588807e+00 3.19513470e-01 -1.11925803e-01
6.95104539e-01 6.15163028e-01 -4.77984995e-02 -3.19370687e-01
-1.23626900e+00 -2.06670880e-01 -7.02342093e-01 -1.88824579e-01
6.36055470e-01 -1.13858521e+00 -6.52240336e-01 7.66896307e-01
-1.16034555e+00 1.67002857e-01 -7.47144520e-02 4.28007275e-01
-1.85836986e-01 5.98109901e-01 -3.63903016e-01 -9.40655947e-01
-4.56561565e-01 -9.81088221e-01 1.10304630e+00 2.53368437e-01
1.69527382e-01 -5.03630340e-01 -1.75590172e-01 5.58606446e-01
1.36352569e-01 2.46087372e-01 8.75179589e-01 -9.68550980e-01
-4.20867920e-01 -1.96629182e-01 -5.07698298e-01 8.37154865e-01
2.97256149e-02 8.87251273e-02 -1.35457647e+00 -3.70428324e-01
7.66068250e-02 -7.55987942e-01 1.58511496e+00 3.10928196e-01
1.64110541e+00 -3.32101494e-01 -8.98424685e-02 6.83763444e-01
1.41315019e+00 6.27898052e-02 5.51738858e-01 2.75215924e-01
4.73985076e-01 6.06808007e-01 4.64297771e-01 6.47787631e-01
5.17703533e-01 1.17957927e-02 8.08138311e-01 -5.49107790e-03
4.09207046e-01 -2.85777986e-01 3.09965224e-04 4.88902718e-01
1.16174929e-02 -4.30500954e-01 -8.93299460e-01 3.36507231e-01
-1.86338580e+00 -1.08126473e+00 6.00859784e-02 2.23955584e+00
5.72869182e-01 -1.12372629e-01 -6.21346056e-01 3.19782734e-01
9.21539485e-01 5.58885217e-01 -6.49077535e-01 1.13017917e-01
-9.51475427e-02 -2.04651251e-01 7.76995957e-01 5.00362754e-01
-1.69774926e+00 7.44069159e-01 6.12978363e+00 7.54844904e-01
-1.20068288e+00 -2.98236832e-02 1.22939658e+00 2.87214279e-01
-4.89828467e-01 2.90880837e-02 -4.38604772e-01 3.22235525e-01
5.58406897e-02 2.92958260e-01 3.55223030e-01 1.01391292e+00
-1.70225233e-01 -2.15806127e-01 -7.96014905e-01 7.93788552e-01
3.84416319e-02 -1.15733600e+00 1.21561356e-01 -1.35971144e-01
9.80441570e-01 2.16077566e-01 1.63814947e-01 3.47595960e-02
6.42656624e-01 -9.08401132e-01 8.15007567e-01 3.29857528e-01
8.45654011e-01 -5.56828737e-01 9.89233255e-01 6.27124369e-01
-8.89893115e-01 -4.97552067e-01 -6.59850240e-01 -2.69792043e-02
-2.81175882e-01 8.95506978e-01 -5.39393902e-01 7.03681827e-01
7.41720676e-01 6.39274240e-01 -5.91355562e-01 8.54585052e-01
-4.61600035e-01 5.23866415e-01 -2.46332869e-01 3.62185746e-01
-5.74560575e-02 -2.69236088e-01 1.14800811e-01 6.87750697e-01
4.98564094e-01 3.27080995e-01 1.88470766e-01 5.77855349e-01
-5.87496459e-01 -1.58950061e-01 -6.61231041e-01 4.80153888e-01
6.03931606e-01 1.31091881e+00 -5.65835714e-01 -3.01459372e-01
-8.38229135e-02 8.49239469e-01 4.56632435e-01 2.17094660e-01
-6.70813739e-01 -7.63244033e-02 3.66715223e-01 -5.74663877e-01
2.37185404e-01 1.05426230e-01 -5.11761665e-01 -1.50964856e+00
-2.12630127e-02 -7.89172411e-01 6.21979415e-01 -8.81046951e-01
-1.43584919e+00 6.98459327e-01 -2.20516294e-01 -1.42455232e+00
6.08697943e-02 -4.72342968e-01 -4.16483819e-01 6.67708278e-01
-1.97398341e+00 -1.03556550e+00 -6.54524148e-01 3.44905078e-01
2.13912129e-01 -8.64923671e-02 7.22616732e-01 1.29191950e-01
-2.75932550e-01 3.54260802e-01 4.67945158e-01 3.79005104e-01
6.62474632e-01 -1.05025148e+00 9.17474553e-02 8.30222547e-01
9.03042853e-02 1.63887098e-01 2.81370968e-01 -5.24222732e-01
-4.50416923e-01 -1.33981586e+00 7.84720361e-01 -1.72933325e-01
2.33572081e-01 3.87341343e-02 -8.84146154e-01 6.59494698e-01
-1.23874202e-01 3.19733173e-01 5.77073395e-01 -3.73069733e-01
-6.95625961e-01 -3.30236524e-01 -1.48467100e+00 1.54926211e-01
6.67515576e-01 -7.02147901e-01 -5.04883468e-01 2.42973834e-01
3.25288624e-01 -2.91604638e-01 -4.00068611e-01 9.83668089e-01
6.83019221e-01 -1.00186980e+00 6.62579298e-01 -4.83272612e-01
6.66087389e-01 -5.06116331e-01 -5.51166356e-01 -1.49149990e+00
-4.88644153e-01 3.02632362e-01 4.58939999e-01 1.01267600e+00
3.77791673e-01 -6.30455613e-01 6.00584805e-01 4.79690731e-01
2.82896936e-01 -2.76328236e-01 -9.66487408e-01 -1.08787350e-01
-4.49729059e-03 -1.75550163e-01 6.45518124e-01 1.10162222e+00
-4.21783686e-01 -5.29998057e-02 -8.62974942e-01 5.22581339e-01
6.88961267e-01 4.54393804e-01 3.36314350e-01 -1.60562396e+00
9.75714102e-02 -1.26415327e-01 -2.12351024e-01 -7.25020826e-01
3.67829829e-01 -9.23961163e-01 2.76565611e-01 -1.22744262e+00
7.83640623e-01 -4.45866525e-01 -6.30600095e-01 7.58311033e-01
-3.96818250e-01 2.97101617e-01 2.26408929e-01 4.87760067e-01
-4.63453650e-01 6.58671319e-01 8.82134199e-01 -5.02497792e-01
3.56292576e-01 2.16398701e-01 -6.88549459e-01 9.48464155e-01
1.00039268e+00 -7.20135033e-01 -1.79264143e-01 -5.78414857e-01
4.11628127e-01 -1.42205521e-01 6.26987100e-01 -7.82885134e-01
-6.50864542e-02 -3.54055345e-01 6.50313675e-01 -6.29832506e-01
-1.03392996e-01 -1.29036641e+00 3.11919767e-02 6.96792424e-01
-8.19726527e-01 -3.46556216e-01 -2.74686664e-01 6.30135119e-01
-2.90676862e-01 -3.90462041e-01 9.70380664e-01 -1.35142326e-01
-6.45736456e-01 4.10424948e-01 -3.47251724e-03 -2.39336014e-01
8.98901224e-01 3.01740557e-01 -6.74282014e-01 -4.53290433e-01
-7.17837155e-01 4.61519241e-01 3.61239403e-01 7.03947842e-02
7.31423080e-01 -1.36701143e+00 -9.80696321e-01 2.45553881e-01
2.86341280e-01 -3.98990512e-02 4.64041233e-01 3.50225627e-01
-5.09982288e-01 4.85752635e-02 -2.90840209e-01 -5.18098712e-01
-1.30357826e+00 3.11811388e-01 5.52433729e-01 -3.05104047e-01
-3.53616513e-02 7.39204228e-01 2.56779224e-01 -9.20081377e-01
3.97786796e-02 -1.84269249e-01 -5.10095060e-01 2.12402925e-01
2.75393665e-01 3.12754124e-01 3.99561971e-02 -6.65641010e-01
-3.96881014e-01 4.48962241e-01 1.03991009e-01 1.84173033e-01
1.33841145e+00 -2.66390473e-01 -2.39565805e-01 1.51054248e-01
1.01012444e+00 -8.08172971e-02 -1.32214749e+00 -3.50751817e-01
-6.41883761e-02 -5.89456856e-01 3.19153845e-01 -9.47374642e-01
-1.29706669e+00 1.06585240e+00 8.66055548e-01 2.48673305e-01
1.17023587e+00 -2.60521770e-01 4.79039967e-01 6.88295245e-01
2.21374631e-01 -1.04173911e+00 -1.79825068e-01 6.21887743e-01
7.84892559e-01 -1.70997393e+00 1.40527427e-01 -1.02739938e-01
-8.74152184e-01 1.17387199e+00 6.04234040e-01 6.26342744e-02
6.56376600e-01 -2.19125256e-01 1.40855744e-01 -4.21166234e-02
-4.22150493e-01 3.94024402e-02 3.09810251e-01 2.65588403e-01
-2.30800919e-02 4.57102716e-01 -2.23546445e-01 5.43498516e-01
1.98096991e-01 2.17278972e-02 4.83925015e-01 6.07902706e-01
-7.08677113e-01 -5.05847096e-01 -2.72626638e-01 5.98459482e-01
-4.68071073e-01 -1.19514652e-01 -5.35716176e-01 1.82561040e-01
3.28963578e-01 9.75418091e-01 -1.52210534e-01 -9.25161958e-01
3.61390586e-04 1.71951219e-01 -2.09240705e-01 -2.63484299e-01
-4.35247660e-01 -3.62724036e-01 -2.65616216e-02 -2.13405475e-01
-5.72046816e-01 -4.46508646e-01 -9.60917771e-01 -2.80009419e-01
-7.11487293e-01 -1.72892846e-02 3.91316712e-01 9.07693088e-01
3.66242856e-01 9.22208950e-02 1.41139865e+00 -6.15002871e-01
-1.14932787e+00 -9.12905812e-01 -5.26623011e-01 4.81701076e-01
5.21057010e-01 -4.29681093e-01 -6.59754455e-01 -9.12362561e-02] | [9.529565811157227, 3.819283962249756] |
4101079d-e7cd-478d-a234-8b10251c8a01 | ground-truth-free-meta-learning-for-deep | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qin_Ground-Truth_Free_Meta-Learning_for_Deep_Compressive_Sampling_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qin_Ground-Truth_Free_Meta-Learning_for_Deep_Compressive_Sampling_CVPR_2023_paper.pdf | Ground-Truth Free Meta-Learning for Deep Compressive Sampling | Deep learning has become an important tool for reconstructing images in compressive sampling (CS). This paper proposes a ground-truth (GT) free meta-learning method for CS, which leverages both external and internal learning for unsupervised high-quality image reconstruction. The proposed method first trains a deep model via external meta-learning using only CS measurements, and then efficiently adapts the trained model to a test sample for further improvement by exploiting its internal characteristics. The meta-learning and model adaptation are built on an improved Stein's unbiased risk estimator (iSURE) that provides efficient computation and effective guidance for accurate prediction in the range space of the adjoint of the measurement matrix. To further improve the learning on the null space of the measurement matrix, a modified model-agnostic meta-learning scheme is proposed, along with a null-space-consistent loss and a bias-adaptive deep unrolling network to improve and accelerate model adaption in test time. Experimental results have demonstrated that the proposed GT-free method performs well, and can even compete with supervised learning-based methods. | ['Hui Ji', 'Tongyao Pang', 'Yuhui Quan', 'Xinran Qin'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-reconstruction', 'unrolling'] | ['computer-vision', 'computer-vision'] | [ 4.82380003e-01 9.79369357e-02 -1.61355674e-01 -2.25202695e-01
-1.35843885e+00 -4.08590212e-02 1.61549792e-01 -3.36275935e-01
-1.97648391e-01 4.70786005e-01 2.90193141e-01 -3.36346895e-01
-2.79347777e-01 -4.83128577e-01 -9.25139010e-01 -9.22383845e-01
-2.66508639e-01 1.72370926e-01 -3.03532928e-01 5.31282872e-02
1.82587966e-01 2.12666243e-01 -7.79008746e-01 8.17330405e-02
1.00216877e+00 1.38182342e+00 4.15179908e-01 5.16752660e-01
3.95394683e-01 1.10344934e+00 -7.38426894e-02 -1.84867933e-01
6.67725146e-01 -6.44586444e-01 -3.53977054e-01 1.15783602e-01
4.32686269e-01 -6.09853923e-01 -7.93091536e-01 1.23429406e+00
7.88599432e-01 5.35067245e-02 4.84147161e-01 -6.35985196e-01
-4.98382777e-01 4.41961765e-01 -5.73721111e-01 -3.05199740e-03
-1.90277006e-02 2.70173043e-01 7.40091860e-01 -1.23875499e+00
2.22368315e-01 1.03554499e+00 1.06257653e+00 1.66060597e-01
-9.86734271e-01 -9.38490570e-01 -2.87857682e-01 2.75904417e-01
-1.41936362e+00 -6.27761543e-01 9.09954846e-01 -2.82015234e-01
2.92769969e-01 3.49067934e-02 7.01517403e-01 9.79271233e-01
2.14181334e-01 9.09167707e-01 1.26871860e+00 -5.69058716e-01
2.62318820e-01 -7.56414086e-02 -5.63635767e-01 9.67414618e-01
1.42648116e-01 5.37285030e-01 -4.49338555e-01 -1.66831732e-01
1.25145400e+00 -5.47813959e-02 -5.08465648e-01 -7.84062207e-01
-1.31937277e+00 8.66065443e-01 7.41149366e-01 -3.25159132e-02
-6.03418767e-01 4.63783085e-01 1.87691227e-01 2.55915642e-01
4.06439453e-01 6.15387380e-01 -3.32476199e-01 2.12439984e-01
-1.30869448e+00 -2.63376564e-01 5.80187857e-01 8.32178652e-01
7.83638358e-01 6.13981068e-01 -4.41097200e-01 8.73798311e-01
3.85217220e-01 8.75531316e-01 5.19419491e-01 -1.26615286e+00
4.53946739e-01 -2.84504369e-02 8.09304714e-02 -1.14981234e+00
-1.34964719e-01 -1.53410470e+00 -1.31397402e+00 2.25595571e-02
-1.97133467e-01 -1.93496257e-01 -8.98216486e-01 1.71741188e+00
3.59993637e-01 8.54880393e-01 -2.08897442e-01 9.34623420e-01
2.63996691e-01 4.48656827e-01 -4.93481845e-01 -2.92503774e-01
6.78943217e-01 -9.71724689e-01 -5.89233518e-01 -3.17153096e-01
4.72686172e-01 -5.89915156e-01 7.24884868e-01 4.67583060e-01
-1.19364798e+00 -7.22135901e-01 -1.41000319e+00 1.94635406e-01
5.59481621e-01 2.49438137e-01 2.79165149e-01 4.91573811e-01
-9.43076015e-01 8.12885940e-01 -8.43483567e-01 2.19911903e-01
8.91737938e-01 1.35055020e-01 -1.95713788e-01 -3.79685283e-01
-1.11026800e+00 6.89374626e-01 3.10164511e-01 2.39545181e-01
-1.43163788e+00 -1.10468996e+00 -8.77354145e-01 -1.42464880e-02
5.20568967e-01 -9.85879719e-01 9.05304074e-01 -1.12582743e+00
-1.79719722e+00 6.25800014e-01 2.06280813e-01 -6.18947506e-01
6.88279092e-01 -2.40420386e-01 -3.36831421e-01 5.30219078e-01
3.61197352e-01 1.25612080e-01 1.80921304e+00 -1.43502474e+00
-3.03525239e-01 -3.97985071e-01 -5.01224935e-01 -1.47886127e-02
-1.02241151e-01 -7.02140629e-01 -4.83046532e-01 -1.16583276e+00
6.52128994e-01 -8.78192425e-01 -5.17426491e-01 9.27140415e-02
-2.67533988e-01 8.18110526e-01 2.96512336e-01 -1.09073508e+00
1.12070096e+00 -2.19204283e+00 1.75215378e-01 5.61251521e-01
3.63189965e-01 2.88791239e-01 -4.40280169e-01 2.42391959e-01
-1.87366366e-01 -4.17701066e-01 -5.61587095e-01 -4.30154622e-01
-3.18704903e-01 -3.00105996e-02 -2.56462365e-01 7.80971885e-01
4.62200269e-02 9.16012824e-01 -1.11415040e+00 -2.40364060e-01
3.05828422e-01 4.32261884e-01 -8.46879780e-01 4.72233534e-01
1.75797731e-01 9.98600364e-01 -4.24004763e-01 4.92890626e-01
8.66032243e-01 -6.69636190e-01 2.21825555e-01 -7.08199978e-01
2.12476179e-01 -1.12042852e-01 -1.26535141e+00 2.02512002e+00
-9.20968831e-01 3.07500482e-01 3.02313358e-01 -1.48378682e+00
6.94344044e-01 9.40080509e-02 6.21276379e-01 -7.85619199e-01
1.22213932e-02 4.61417735e-01 -1.85666353e-01 -5.76383889e-01
-7.59356096e-02 -4.12628859e-01 4.08230275e-01 2.64321029e-01
3.12311858e-01 -3.13671559e-01 -7.03373849e-01 2.14921072e-01
1.04399168e+00 2.47308195e-01 3.71608436e-01 -3.23911339e-01
6.65409327e-01 -5.18457353e-01 7.39059865e-01 1.08785188e+00
-6.21693023e-02 7.07571566e-01 -2.41213322e-01 -2.69521505e-01
-1.18266726e+00 -9.98968661e-01 -2.64411807e-01 5.52221358e-01
6.05306663e-02 -2.60311812e-01 -5.15544116e-01 -5.61576188e-01
-5.17297350e-02 6.53039515e-01 -5.79490840e-01 -5.41982949e-01
-4.99730051e-01 -6.93426847e-01 1.63153559e-01 2.95762062e-01
1.04830945e+00 -3.77661794e-01 -2.27799892e-01 3.93593669e-01
-3.51996094e-01 -1.09961748e+00 -3.27389836e-01 1.06662378e-01
-1.03387296e+00 -8.43612432e-01 -9.15137351e-01 -4.82929379e-01
6.61809206e-01 3.86294216e-01 7.17814982e-01 9.06844735e-02
-5.71360998e-02 6.83114111e-01 -2.40162700e-01 -1.26026526e-01
-4.51392770e-01 -3.84745002e-01 5.34489155e-02 6.66691482e-01
-5.40881872e-01 -8.53738725e-01 -1.02699149e+00 1.36719719e-01
-7.88317204e-01 9.33587402e-02 9.57471311e-01 1.44392955e+00
6.88675404e-01 -2.15681419e-01 6.47268474e-01 -8.79448473e-01
1.09418062e-02 -5.31483769e-01 -6.11204505e-01 6.83391169e-02
-8.88047516e-01 2.82859474e-01 5.86468160e-01 -2.04467744e-01
-9.50190365e-01 -3.58870439e-02 -1.78472579e-01 -8.81698310e-01
6.94979727e-01 7.53164649e-01 2.04481184e-02 -6.38986528e-01
5.17055452e-01 6.40850067e-01 2.53741562e-01 -2.77776629e-01
1.70130640e-01 3.71823043e-01 6.88994944e-01 -5.38509727e-01
1.26214790e+00 7.02422857e-01 4.53232288e-01 -7.23787487e-01
-1.47000551e+00 -6.05689228e-01 -4.83783394e-01 -8.24570730e-02
4.17498052e-01 -1.43088341e+00 -2.88220316e-01 5.92767000e-01
-4.70675617e-01 -5.21234572e-01 -5.39514124e-01 8.69952917e-01
-8.84850264e-01 7.18239188e-01 -5.22373796e-01 -6.28706157e-01
-4.34352636e-01 -9.98234510e-01 1.04682219e+00 -3.08415741e-01
4.38884646e-01 -1.11573696e+00 3.31438705e-02 5.33311009e-01
7.92180598e-01 9.71148089e-02 5.84865749e-01 -1.00567579e-01
-8.77174675e-01 -2.71603733e-01 -2.56733149e-01 1.03497481e+00
-1.00177571e-01 -7.92631269e-01 -9.72339809e-01 -1.00532258e+00
6.02055490e-01 -4.56952661e-01 1.04880977e+00 6.79079652e-01
1.57365489e+00 -5.03107429e-01 -7.69044831e-02 1.57732737e+00
1.64938676e+00 -3.08280945e-01 6.82679713e-01 4.36582744e-01
7.18389869e-01 -1.57423113e-02 4.54425037e-01 7.40301311e-01
2.99264312e-01 5.08451283e-01 6.08819246e-01 -1.61861509e-01
-2.59826064e-01 -4.29942578e-01 2.95133978e-01 1.12318349e+00
1.56715453e-01 2.80533761e-01 -4.76884663e-01 2.93277234e-01
-1.67883658e+00 -8.68020654e-01 3.93039614e-01 2.31002975e+00
7.03756452e-01 -1.56986549e-01 -4.39910501e-01 1.09885707e-01
3.91693264e-01 3.29132348e-01 -8.65419030e-01 5.31586647e-01
-1.16202801e-01 3.98236066e-01 6.76179945e-01 5.66981196e-01
-1.10420322e+00 5.89843988e-01 5.70938587e+00 1.04098821e+00
-1.30147243e+00 3.86193961e-01 6.17822289e-01 1.91782072e-01
-2.16567233e-01 1.54218584e-01 -2.25166634e-01 3.34454536e-01
7.11661756e-01 1.35746405e-01 6.56170845e-01 6.26708031e-01
1.14959233e-01 2.70471126e-01 -8.27728271e-01 1.38995683e+00
3.66191059e-01 -1.71549749e+00 1.18502319e-01 7.50982538e-02
9.94952321e-01 1.99731916e-01 3.48650426e-01 4.85596448e-01
-1.13616370e-01 -7.14127779e-01 8.09346497e-01 7.51502216e-01
1.10268021e+00 -5.52324176e-01 6.94934428e-01 3.88788939e-01
-8.58958006e-01 -3.67547154e-01 -3.77448559e-01 2.29361653e-01
-2.49492247e-02 8.27135742e-01 -6.88410103e-01 8.21844816e-01
4.92609441e-01 1.14742184e+00 -3.30742002e-01 1.01722527e+00
-2.54626900e-01 7.09534049e-01 -4.65945248e-03 6.23208940e-01
1.73141196e-01 -4.01147872e-01 9.41691101e-01 9.16844487e-01
7.15001345e-01 2.32528254e-01 4.07305568e-01 7.45944440e-01
-7.70553499e-02 -1.82950795e-01 -3.32687140e-01 2.34850690e-01
3.02478015e-01 1.14137828e+00 -1.49031850e-02 -2.87669539e-01
-1.91170782e-01 1.06887102e+00 1.58978641e-01 5.45213342e-01
-5.44119418e-01 1.62210204e-02 1.15165450e-01 1.96854547e-01
5.55070341e-01 -1.95337519e-01 -5.05126007e-02 -1.67352211e+00
-1.65809914e-01 -1.10961723e+00 2.69512922e-01 -7.75843799e-01
-1.28351867e+00 2.88358718e-01 -4.16784674e-01 -1.64304996e+00
-2.73064196e-01 -3.35582078e-01 -5.10784864e-01 7.36729741e-01
-1.88365602e+00 -1.34194183e+00 -2.88357347e-01 6.56009197e-01
1.72698036e-01 -5.64920425e-01 6.07527792e-01 3.72040451e-01
-4.34666753e-01 9.35065567e-01 4.90004003e-01 2.07494363e-01
6.36312544e-01 -9.22365487e-01 -1.86092094e-01 1.04466891e+00
3.05630732e-02 7.02264667e-01 4.56642687e-01 -5.09249747e-01
-1.80858254e+00 -1.40940821e+00 -1.48213893e-01 6.28464371e-02
6.70122385e-01 5.00797480e-02 -5.91942906e-01 7.85850763e-01
-2.30108500e-01 5.20706117e-01 6.14810586e-01 -2.08546534e-01
-4.82019097e-01 -4.61665779e-01 -1.10754085e+00 2.57125556e-01
8.93129528e-01 -7.54947305e-01 -1.20623618e-01 4.18465078e-01
5.96403241e-01 -5.24601042e-01 -8.65002513e-01 7.86541879e-01
5.22940338e-01 -1.07666147e+00 1.38614357e+00 -4.20667112e-01
4.56944019e-01 -1.33127406e-01 -6.08371556e-01 -1.47317195e+00
-6.53996110e-01 -9.18114305e-01 -5.03723681e-01 5.58238208e-01
-1.58705469e-02 -6.49650812e-01 7.34398127e-01 -2.19295412e-01
-6.38010919e-01 -8.70039761e-01 -1.02027512e+00 -8.25937629e-01
-9.30518284e-02 -5.71512222e-01 4.43502843e-01 1.05287504e+00
-5.31420112e-01 9.44917202e-02 -6.40799582e-01 7.12175608e-01
1.46665442e+00 2.64573544e-01 8.51140976e-01 -6.64491832e-01
-9.06175315e-01 -8.51732120e-02 -4.73957956e-01 -1.39871359e+00
5.95249906e-02 -1.18657112e+00 -5.93870617e-02 -1.24392176e+00
2.78739840e-01 -6.38821959e-01 -6.18755519e-01 -1.69790104e-01
-2.02718958e-01 2.50136584e-01 1.75242379e-01 3.51895124e-01
-5.65113127e-01 1.00906861e+00 1.59656155e+00 -2.80808449e-01
3.29319656e-01 2.61070848e-01 -5.93789458e-01 7.78526008e-01
3.94718528e-01 -4.61340785e-01 -2.41505593e-01 -4.20731157e-01
3.06639940e-01 5.37425697e-01 6.54577374e-01 -1.40961504e+00
2.44662836e-01 7.51852468e-02 6.45962358e-01 -4.33825552e-01
3.01339686e-01 -8.75513494e-01 -6.95558861e-02 8.09475422e-01
-4.57068950e-01 -6.46334231e-01 -1.74192563e-01 9.98487473e-01
-1.90141931e-01 -9.08566564e-02 1.28453112e+00 -1.40242457e-01
-3.89861494e-01 8.39329243e-01 3.69970173e-01 1.90392032e-01
4.93244231e-01 -4.02200855e-02 2.74932504e-01 -8.25185835e-01
-7.00779140e-01 5.64090423e-02 1.63480818e-01 -2.27788731e-01
1.03088295e+00 -1.52437806e+00 -1.04367459e+00 7.23506868e-01
1.14218965e-01 -1.54069915e-01 4.93928254e-01 9.82433796e-01
-4.20775354e-01 -1.33355418e-02 5.72111607e-02 -8.61543536e-01
-2.76701301e-01 3.99587393e-01 6.13608301e-01 -2.82872468e-01
-6.56727672e-01 1.11518133e+00 2.15858012e-01 -7.49209881e-01
1.94082990e-01 -2.50149071e-02 3.86566788e-01 -4.52349812e-01
5.81864357e-01 4.01817143e-01 -5.62591106e-02 -3.59963894e-01
5.96616566e-02 6.87130511e-01 1.95278049e-01 -1.83622763e-01
1.59615445e+00 -3.66479069e-01 -1.49096310e-01 3.27598959e-01
1.57792377e+00 1.36493608e-01 -1.63704264e+00 -8.02376270e-01
-1.47428602e-01 -8.73910010e-01 6.66519105e-01 -7.34277725e-01
-1.41518950e+00 8.35665047e-01 7.68341482e-01 -3.53267103e-01
1.32109928e+00 -4.51507360e-01 1.00543213e+00 3.27340394e-01
4.91258323e-01 -9.48487282e-01 7.26639450e-01 4.41283405e-01
1.14036655e+00 -1.57873356e+00 2.65400410e-01 -4.14347015e-02
-2.69676864e-01 1.05706525e+00 2.01479152e-01 -3.98596436e-01
9.61717069e-01 -3.80336903e-02 -1.49518847e-01 -8.12336653e-02
-2.67014563e-01 3.44736189e-01 4.29342866e-01 5.62663198e-01
-1.93052068e-01 -3.24345380e-02 4.12245601e-01 3.67276371e-01
1.21513307e-01 1.09443478e-01 3.14456552e-01 5.96080720e-01
-4.06968474e-01 -7.70401955e-01 -3.90867025e-01 7.51573861e-01
-2.45918587e-01 -3.92189860e-01 5.40667474e-01 4.20635611e-01
-1.89303666e-01 6.57316625e-01 -4.06768531e-01 -3.90051305e-01
6.69505000e-02 -5.09755313e-01 8.31030726e-01 -6.01757586e-01
-9.28645581e-02 2.83276498e-01 -3.02811235e-01 -8.37161839e-01
-2.47779980e-01 -6.44538641e-01 -8.90769720e-01 -3.75273265e-02
-4.52645719e-01 4.24901135e-02 6.47075236e-01 8.57085824e-01
4.20187533e-01 3.88004690e-01 1.47144687e+00 -8.49782884e-01
-1.40242565e+00 -7.89599299e-01 -7.03024387e-01 2.12675288e-01
8.69142294e-01 -3.82326454e-01 -6.07517898e-01 -2.59709120e-01] | [11.28690242767334, -2.2484071254730225] |
ca8f51cb-e166-4743-85b4-48029a1ea26d | effectively-incorporating-weighted-cost-to-go | 2205.11624 | null | https://arxiv.org/abs/2205.11624v4 | https://arxiv.org/pdf/2205.11624v4.pdf | Effective Integration of Weighted Cost-to-go and Conflict Heuristic within Suboptimal CBS | Conflict-Based Search (CBS) is a popular multi-agent path finding (MAPF) solver that employs a low-level single agent planner and a high-level constraint tree to resolve conflicts. The vast majority of modern MAPF solvers focus on improving CBS by reducing the size of this tree through various strategies with few methods modifying the low level planner. Typically low level planners in existing CBS methods use an unweighted cost-to-go heuristic, with suboptimal CBS methods also using a conflict heuristic to help the high level search. In this paper, we show that, contrary to prevailing CBS beliefs, a weighted cost-to-go heuristic can be used effectively alongside the conflict heuristic in two possible variants. In particular, one of these variants can obtain large speedups, 2-100x, across several scenarios and suboptimal CBS methods. Importantly, we discover that performance is related not to the weighted cost-to-go heuristic but rather to the relative conflict heuristic weight's ability to effectively balance low-level and high-level work, implying that existing suboptimal CBS work misses this subtlety. Additionally, to the best of our knowledge, we show the first theoretical relation of prioritized planning and bounded suboptimal CBS and demonstrate that our methods are their natural generalization. | ['Tushar Kusnur', 'Maxim Likhachev', 'Rishi Veerapaneni'] | 2022-05-23 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 2.62846977e-01 6.49128675e-01 -5.32768011e-01 8.83324742e-02
-7.04297781e-01 -7.14031935e-01 5.61896086e-01 2.04721466e-01
-3.70730817e-01 1.32818675e+00 2.76460469e-01 -5.46092808e-01
-5.86568356e-01 -9.08732533e-01 -3.92978489e-01 -4.53512609e-01
-6.27792358e-01 9.54542279e-01 1.06361818e+00 -7.52392650e-01
5.76282263e-01 4.27762032e-01 -1.07984281e+00 9.64221507e-02
6.67562604e-01 5.29667854e-01 6.62840232e-02 6.59725904e-01
-3.32198590e-02 7.77540207e-01 -6.31645739e-01 1.54206693e-01
6.23159111e-01 -3.59635711e-01 -1.37330282e+00 -2.81842232e-01
-1.60547435e-01 -3.45311522e-01 -4.26198095e-02 8.47446740e-01
2.18903333e-01 2.12208822e-01 3.06593657e-01 -1.98264086e+00
2.50439405e-01 8.38876247e-01 -8.13680172e-01 2.97289968e-01
8.89184415e-01 1.64111570e-01 1.03038430e+00 -4.82684672e-02
9.05804634e-01 1.27793097e+00 5.04998267e-01 3.17358077e-01
-1.16363311e+00 -3.81844372e-01 5.52317500e-01 2.91034520e-01
-1.14977956e+00 -2.44892031e-01 4.41941828e-01 -4.90940809e-02
1.61392677e+00 5.20321012e-01 6.78668499e-01 5.59179068e-01
5.57001889e-01 2.67134100e-01 1.39064491e+00 -5.73399305e-01
4.32783425e-01 -4.43754256e-01 -7.85543099e-02 6.82028294e-01
3.77459288e-01 4.18280363e-01 -4.01215523e-01 -5.59735775e-01
6.27461612e-01 -6.41918421e-01 -1.83383986e-01 -3.93960774e-01
-1.44692075e+00 7.93079019e-01 2.57748783e-01 2.25837350e-01
-3.60658586e-01 4.63300407e-01 3.54130268e-01 3.47890019e-01
-1.69040740e-01 8.30063343e-01 -5.68850160e-01 -2.71548569e-01
-7.26279855e-01 6.93507671e-01 1.15356398e+00 1.07978117e+00
7.24326372e-01 -3.30157846e-01 -3.55644971e-02 6.84474260e-02
9.64433998e-02 5.41999787e-02 -5.68263195e-02 -1.49727821e+00
4.71029699e-01 4.33691353e-01 4.63995874e-01 -9.33629394e-01
-8.72295618e-01 -9.08210576e-02 -1.80648521e-01 7.33464956e-01
4.08152401e-01 -1.40547052e-01 -5.52068353e-01 1.77492106e+00
5.91624737e-01 -4.17790469e-03 1.17189988e-01 8.84029627e-01
6.66297376e-02 6.54564261e-01 -3.24174166e-01 -7.14387298e-01
1.11040545e+00 -1.46163368e+00 -4.02728260e-01 -2.45005116e-01
8.01542163e-01 -5.47771871e-01 5.84879279e-01 4.92293417e-01
-1.60307622e+00 3.17083001e-01 -1.24795783e+00 2.45409176e-01
-2.32937589e-01 -9.76527452e-01 8.90264750e-01 4.73524004e-01
-1.46482122e+00 4.72498029e-01 -7.96751380e-01 -4.78952348e-01
-1.42548472e-01 6.91281736e-01 -2.28469998e-01 -2.65740812e-01
-9.61556315e-01 1.29633498e+00 4.73986626e-01 -3.19449186e-01
-8.00814629e-01 -3.77247870e-01 -5.30406475e-01 4.87429053e-02
1.31521893e+00 -7.53407836e-01 1.61749637e+00 -3.46748590e-01
-1.58672786e+00 2.62938619e-01 -1.24476589e-01 -3.65949869e-01
4.73655939e-01 4.17291105e-01 -2.23109163e-02 9.38677341e-02
5.35616875e-01 5.70403934e-01 1.97252393e-01 -1.36620581e+00
-1.21628237e+00 7.28252977e-02 8.30895603e-01 3.95870984e-01
4.57551211e-01 9.94376689e-02 -2.42555849e-02 2.33496167e-03
2.90322930e-01 -1.09782529e+00 -7.79834390e-01 -3.98502380e-01
-2.55827338e-01 -4.17374969e-01 3.36727023e-01 -4.53411266e-02
1.35834765e+00 -1.63575649e+00 4.36208427e-01 4.07142520e-01
2.94217020e-01 -3.22967023e-01 -4.34593320e-01 9.91321325e-01
3.04760039e-01 2.32653499e-01 -1.36049598e-01 3.11882257e-01
3.08651239e-01 4.17778134e-01 -4.06142510e-02 4.10161197e-01
-6.77810609e-02 5.40897012e-01 -1.39444304e+00 -5.39401770e-01
-7.50395954e-02 -4.05133843e-01 -8.40459466e-01 -2.57775724e-01
-4.51061636e-01 -2.35843286e-02 -4.34879124e-01 6.45988226e-01
4.50211763e-01 1.61193699e-01 7.13246465e-01 3.29074174e-01
-7.84759760e-01 5.98927379e-01 -1.38613594e+00 1.57119286e+00
-2.21752711e-02 2.10351750e-01 6.30229056e-01 -7.39939213e-01
2.58401453e-01 4.82971966e-02 6.70536757e-01 -9.36031461e-01
-1.78520933e-01 5.29158115e-01 2.73573011e-01 -3.99050415e-02
6.73087478e-01 -5.19114248e-02 -4.60884422e-01 6.62699699e-01
-6.07844174e-01 -3.45871717e-01 6.51983798e-01 3.18602741e-01
1.67136121e+00 1.18065208e-01 6.64309740e-01 -5.71749568e-01
3.69399309e-01 9.19525743e-01 9.04610753e-01 9.61838126e-01
-4.79387224e-01 -2.16451868e-01 8.32409441e-01 -5.55109262e-01
-6.95735455e-01 -6.58091426e-01 2.33575284e-01 1.18884063e+00
6.56206548e-01 -7.27068186e-01 -5.34808576e-01 -5.71653366e-01
-2.03460753e-01 8.48297894e-01 -4.05655414e-01 8.49565268e-02
-1.04480290e+00 -5.37028015e-01 5.51262259e-01 2.57039398e-01
3.04827005e-01 -8.12812805e-01 -1.24933958e+00 6.84235811e-01
-1.64416716e-01 -9.04955387e-01 -4.16901827e-01 3.53728682e-01
-4.78135824e-01 -1.36099517e+00 9.22714174e-02 -6.38954043e-01
4.79884446e-01 6.26490295e-01 9.78361726e-01 4.03906196e-01
5.82587793e-02 2.77159929e-01 -5.56233943e-01 -2.90863276e-01
-2.76744783e-01 1.48909405e-01 1.92919057e-02 -1.09414387e+00
-2.42986560e-01 -6.80956721e-01 -3.70247334e-01 5.03416181e-01
-4.71384287e-01 2.84146309e-01 5.01692533e-01 7.37757087e-01
2.81199634e-01 5.02709031e-01 4.17003036e-01 -4.04566109e-01
9.39609110e-01 -4.37842965e-01 -6.68473125e-01 1.73010483e-01
-7.88078368e-01 4.19764295e-02 3.54191244e-01 -1.99117407e-01
-7.61442065e-01 -2.06471607e-01 3.33402872e-01 2.80858517e-01
3.26403342e-02 6.81864560e-01 1.60359502e-01 -3.45456451e-01
4.48926091e-01 -2.32700735e-01 -2.45647550e-01 1.82710066e-01
2.99267918e-01 -3.04232407e-02 3.52974653e-01 -1.13191879e+00
5.65315723e-01 2.89206326e-01 5.29281497e-01 -7.04824254e-02
-2.82000363e-01 -1.30820632e-01 -2.67662317e-01 -2.50488341e-01
5.16579330e-01 -2.86385357e-01 -1.06342793e+00 -1.63232893e-01
-1.15116692e+00 -9.35014367e-01 -1.98181242e-01 1.46368489e-01
-8.87738645e-01 2.49009788e-01 -6.81676865e-01 -8.55029285e-01
1.96635693e-01 -1.46807837e+00 8.14876676e-01 4.43223193e-02
-4.07592714e-01 -5.35236478e-01 4.12888706e-01 -5.47736976e-03
5.25677443e-01 6.82635009e-01 9.55747604e-01 -4.77550030e-01
-6.98995948e-01 3.94998103e-01 -2.48816520e-01 -9.19182777e-01
-1.66748296e-02 -1.83816075e-01 -4.80269790e-02 -5.27063370e-01
-1.62443370e-01 -2.23152321e-02 9.73548889e-02 4.61795092e-01
1.41852796e-01 -6.28484070e-01 -8.24579597e-01 6.87779486e-02
1.57034278e+00 6.77879214e-01 4.97774750e-01 1.16665554e+00
1.11942373e-01 7.02942610e-01 9.58709002e-01 4.73365605e-01
7.68164754e-01 9.99590874e-01 7.21447527e-01 2.75150836e-01
-3.82904112e-02 2.44571298e-01 2.36996815e-01 4.17170227e-01
-4.84589428e-01 -1.68653429e-01 -1.25528717e+00 5.25406063e-01
-2.28310013e+00 -1.02991772e+00 -1.48047313e-01 1.90483499e+00
8.74431252e-01 4.90162194e-01 4.85938311e-01 2.53407180e-01
5.85255265e-01 1.92858092e-03 -3.77213120e-01 -9.01049972e-01
2.80715555e-01 -3.46432924e-02 8.19214344e-01 9.69948709e-01
-4.88595486e-01 1.15027809e+00 7.22334862e+00 5.46477020e-01
-6.43332005e-01 1.85239077e-01 -1.88899904e-01 -3.75500768e-01
-1.53720707e-01 5.02133608e-01 -3.85700881e-01 1.50389656e-01
8.49889457e-01 -6.31906390e-01 9.32263851e-01 7.66127646e-01
2.88581222e-01 -5.31660259e-01 -1.18363118e+00 2.40659833e-01
-3.74407202e-01 -1.42670703e+00 -4.81453657e-01 4.94634598e-01
6.37201011e-01 -2.21495837e-01 -6.09229088e-01 3.86379153e-01
8.61975670e-01 -9.56569254e-01 1.10007179e+00 -1.29085675e-01
3.78637373e-01 -1.05100346e+00 6.20446503e-01 5.83829761e-01
-1.42671716e+00 -2.40578294e-01 4.91259154e-03 -7.79158473e-01
6.73454881e-01 -1.16948765e-02 -8.02009344e-01 8.04627180e-01
5.53709149e-01 -1.12999462e-01 1.73203990e-01 1.04930925e+00
-6.66768402e-02 -1.58201948e-01 -4.95653272e-01 -3.91860902e-02
8.63468289e-01 -4.72108424e-02 8.96526754e-01 1.10789883e+00
5.12352102e-02 5.91100693e-01 8.04892838e-01 4.90827829e-01
6.32733524e-01 -4.04943585e-01 -3.02434564e-01 2.26148874e-01
8.48162293e-01 8.13943028e-01 -9.93245602e-01 -1.86384872e-01
-3.18127394e-01 4.93305176e-01 4.14844006e-01 1.36555046e-01
-9.22361314e-01 -1.60926789e-01 7.15355396e-01 9.14075784e-03
5.69248646e-02 -5.60088992e-01 -3.49716425e-01 -5.29151022e-01
-1.44556969e-01 -1.05690002e+00 4.64331806e-01 -5.80448866e-01
-7.19776511e-01 6.10636830e-01 5.96126795e-01 -7.52281070e-01
-3.81726682e-01 -2.35010087e-01 -4.68818337e-01 5.54931104e-01
-1.60508585e+00 -9.11781073e-01 2.69316547e-02 4.95283961e-01
6.62472963e-01 2.20555291e-01 7.92164147e-01 -8.31085965e-02
-4.88824427e-01 1.65947646e-01 -4.50154305e-01 -5.58735609e-01
3.55911732e-01 -1.19474125e+00 1.73931971e-01 9.77036655e-01
-6.98966384e-01 5.71689308e-01 1.00598061e+00 -7.82679558e-01
-1.71133411e+00 -3.95359993e-01 7.10049987e-01 -1.86609253e-02
7.38502324e-01 2.31501818e-01 -2.60307819e-01 8.52175713e-01
5.68913043e-01 -5.05628288e-01 1.04706153e-01 1.39740348e-01
-9.70391929e-02 2.25477472e-01 -1.37716889e+00 9.21254516e-01
1.36283672e+00 1.51944116e-01 -7.05421448e-01 3.40546608e-01
9.03678298e-01 -7.99474835e-01 -6.44360900e-01 5.48178971e-01
5.46118379e-01 -1.24993324e+00 7.76249826e-01 -6.26680255e-01
1.27339840e-01 -6.97608054e-01 -3.37654799e-01 -1.34656239e+00
-7.62213469e-01 -8.29282999e-01 2.65484042e-02 6.70335948e-01
3.91297251e-01 -9.20832813e-01 7.18663096e-01 7.03852117e-01
-6.65718794e-01 -9.34565544e-01 -1.27774954e+00 -1.04580665e+00
6.22190274e-02 -3.63200307e-01 8.40862036e-01 1.02269006e+00
7.86535919e-01 1.27819821e-01 -8.49651173e-02 4.45489973e-01
6.25097990e-01 1.61184728e-01 5.39567411e-01 -1.08289742e+00
-3.66487563e-01 -9.28065658e-01 4.91195805e-02 -6.34520411e-01
4.90582734e-02 -5.50382257e-01 2.50403821e-01 -1.99427867e+00
5.11053056e-02 -7.57886052e-01 1.12706035e-01 7.59901226e-01
4.06379759e-01 -3.41632009e-01 4.78912890e-01 2.53329873e-01
-7.29600787e-01 -1.64072961e-01 1.26336825e+00 -5.16915508e-02
-5.05153239e-01 -4.86164331e-01 -7.19961166e-01 6.53958976e-01
9.31620300e-01 -6.20431364e-01 -4.64361727e-01 -3.04715365e-01
6.59354866e-01 6.47087812e-01 4.42280397e-02 -6.71154976e-01
7.55458117e-01 -1.22926295e+00 -6.97355151e-01 -2.66440153e-01
2.66227931e-01 -7.50432849e-01 4.15424407e-01 8.75120282e-01
-1.77296042e-01 4.51519340e-01 2.60972619e-01 3.77005935e-01
8.78690183e-02 -3.14268947e-01 3.04768890e-01 -4.45562482e-01
-9.03626084e-01 -2.30385542e-01 -7.74419665e-01 -1.67344719e-01
1.46070659e+00 -4.82153952e-01 -9.21365738e-01 -2.13868409e-01
-4.15672123e-01 6.38584077e-01 6.18647158e-01 -1.03018850e-01
2.52575576e-01 -1.12741315e+00 -4.76364791e-01 -4.89354908e-01
-2.66287386e-01 1.46063626e-01 -1.15405127e-01 1.27313447e+00
-6.52670026e-01 6.33384764e-01 -5.12520194e-01 -7.61127174e-02
-9.46222901e-01 8.14997792e-01 2.66311854e-01 -7.35548317e-01
-5.91783047e-01 5.39651334e-01 -1.75656781e-01 -1.39966697e-01
1.57057554e-01 -3.25175524e-01 2.16885060e-01 -1.56071633e-01
5.53130150e-01 8.46938610e-01 -1.49502009e-01 -3.00004005e-01
-1.01795805e+00 4.78781939e-01 1.35677248e-01 -5.87181091e-01
1.29803181e+00 -3.07380557e-01 -4.31165814e-01 -1.91512689e-01
2.29271799e-01 1.55809462e-01 -9.29042339e-01 1.65145397e-01
3.38250875e-01 -4.53810334e-01 1.07578188e-01 -9.71482158e-01
-5.95239639e-01 -1.37479812e-01 -3.11432898e-01 6.55188143e-01
1.09461713e+00 -3.23680714e-02 7.08279371e-01 4.30757672e-01
1.35058105e+00 -1.22180557e+00 -1.75351024e-01 5.95131159e-01
8.97124529e-01 -6.11006379e-01 3.51321697e-01 -9.20607209e-01
-4.67442870e-01 9.82742429e-01 7.20234156e-01 -1.89203054e-01
-8.07617083e-02 8.53114009e-01 -3.53431612e-01 -2.88202673e-01
-9.82626438e-01 -3.69452238e-01 -5.68300843e-01 7.48684049e-01
-1.93600267e-01 1.50178790e-01 -8.83960247e-01 2.86539972e-01
-3.62087429e-01 -1.50090441e-01 1.02687573e+00 1.57994771e+00
-8.04182172e-01 -1.41589260e+00 -6.45882845e-01 2.40908656e-02
1.11786351e-01 6.74974248e-02 -4.77027059e-01 1.28281736e+00
4.01412919e-02 1.41944551e+00 -1.30486250e-01 -3.07844400e-01
5.02722502e-01 -2.51167625e-01 8.60092402e-01 -4.95020121e-01
-7.79399872e-01 1.03044271e-01 9.36218977e-01 -1.11264908e+00
-6.57757044e-01 -6.56642199e-01 -1.68832207e+00 -8.26003313e-01
-1.03361592e-01 2.68761158e-01 2.96518624e-01 1.04994261e+00
7.38423690e-02 5.78338981e-01 3.05583656e-01 -1.21636128e+00
-4.37749475e-01 -2.89932072e-01 -3.35492045e-01 -3.19141120e-01
2.71709472e-01 -9.60187018e-01 -3.00405741e-01 -6.80233479e-01] | [4.971241474151611, 1.8725779056549072] |
97b536b1-4a68-47a1-af20-618470adea13 | convolutional-recurrent-neural-networks-on | 2001.03538 | null | https://arxiv.org/abs/2001.03538v1 | https://arxiv.org/pdf/2001.03538v1.pdf | Convolutional-Recurrent Neural Networks on Low-Power Wearable Platforms for Cardiac Arrhythmia Detection | Low-power sensing technologies, such as wearables, have emerged in the healthcare domain since they enable continuous and non-invasive monitoring of physiological signals. In order to endow such devices with clinical value, classical signal processing has encountered numerous challenges. However, data-driven methods, such as machine learning, offer attractive accuracies at the expense of being resource and memory demanding. In this paper, we focus on the inference of neural networks running in microcontrollers and low-power processors which wearable sensors and devices are generally equipped with. In particular, we adapted an existing convolutional-recurrent neural network, designed to detect and classify cardiac arrhythmias from a single-lead electrocardiogram, to the low-power embedded System-on-Chip nRF52 from Nordic Semiconductor with an ARM's Cortex-M4 processing core. We show our implementation in fixed-point precision, using the CMSIS-NN libraries, yields a drop of $F_1$ score from 0.8 to 0.784, from the original implementation, with a memory footprint of 195.6KB, and a throughput of 33.98MOps/s. | ['Ricard Delgado-Gonzalo', 'Antonino Faraone'] | 2020-01-08 | null | null | null | null | ['arrhythmia-detection'] | ['medical'] | [ 4.56042230e-01 -6.75130039e-02 -6.44489303e-02 -5.13040781e-01
-4.93638545e-01 -1.37650385e-01 -4.59369361e-01 3.15694332e-01
-6.14597976e-01 7.15975225e-01 -2.44259447e-01 -6.35726810e-01
-1.26594022e-01 -5.08638620e-01 -2.91670144e-01 -6.66914344e-01
-1.19103715e-01 -2.89850801e-01 -2.45120138e-01 1.19632386e-01
-3.70774060e-01 3.97746921e-01 -1.30278957e+00 2.65973151e-01
4.95309263e-01 1.56987870e+00 -2.56324708e-01 8.01011562e-01
5.85331321e-01 5.52236617e-01 -6.87877834e-01 -1.72309145e-01
-7.77149871e-02 -2.73565739e-01 -1.84026152e-01 -8.23863328e-01
-1.71985358e-01 -4.21332717e-01 -2.60234058e-01 7.21598029e-01
1.07910633e+00 -6.33843958e-01 2.91648984e-01 -8.12276483e-01
1.64453745e-01 7.72367775e-01 -1.33591235e-01 3.86129856e-01
1.43264711e-01 1.36879861e-01 3.03120106e-01 -5.20488620e-01
-6.31240159e-02 4.58407551e-01 1.36103094e+00 6.76925421e-01
-1.21524847e+00 -6.91564858e-01 -6.04718864e-01 2.52544675e-02
-1.57685447e+00 -7.12378144e-01 9.04343307e-01 -1.42290711e-01
1.47248077e+00 3.97455126e-01 7.58855343e-01 1.37700236e+00
7.17223167e-01 1.65115654e-01 8.89325678e-01 -2.56591290e-01
5.44139862e-01 6.31640479e-02 1.95234254e-01 2.64824212e-01
7.64266193e-01 -1.66081607e-01 -7.17278063e-01 -5.01684666e-01
7.17006505e-01 4.64337736e-01 -2.71224588e-01 2.84433812e-01
-1.14173412e+00 3.02547127e-01 1.97864071e-01 4.29877073e-01
-6.95817888e-01 6.22603714e-01 7.81950831e-01 4.30820808e-02
2.01747492e-01 3.83287251e-01 -8.28161657e-01 -6.37786448e-01
-7.41691589e-01 -1.48405179e-01 1.01122630e+00 6.02249503e-01
1.03376858e-01 4.29745793e-01 6.51998296e-02 3.35758209e-01
3.23452592e-01 6.52080715e-01 7.50655890e-01 -5.85478604e-01
1.98274866e-01 5.95761120e-01 6.27941489e-02 -8.68011475e-01
-1.13225186e+00 -4.85095471e-01 -1.56849623e+00 -3.92061561e-01
1.05416097e-01 -7.31492639e-01 -3.91724646e-01 1.43174112e+00
8.80524293e-02 -3.79044525e-02 1.90776452e-01 7.88371861e-01
8.21694374e-01 2.76822984e-01 2.38765597e-01 -3.84678394e-01
1.52316725e+00 -9.01570395e-02 -8.69433105e-01 -1.96684077e-01
5.13255179e-01 -1.59999847e-01 8.97814035e-01 5.30686140e-01
-9.64160264e-01 -6.11935735e-01 -1.34913158e+00 1.35315672e-01
-6.95911944e-02 5.73863864e-01 4.39934045e-01 1.11564648e+00
-1.02500153e+00 8.18681836e-01 -1.30671954e+00 -1.93512604e-01
4.48367327e-01 6.78816199e-01 1.34166852e-01 8.50528359e-01
-1.34806895e+00 6.01013601e-01 2.09344894e-01 4.03088927e-01
-3.12225938e-01 -8.53221714e-01 -4.39254314e-01 3.76550913e-01
-2.23910734e-01 -7.41820276e-01 1.13471305e+00 -7.20956504e-01
-1.83900595e+00 4.28255796e-01 1.34093445e-02 -9.18311417e-01
5.35617881e-02 -2.57343590e-01 -9.34345722e-01 1.74892530e-01
-5.56778610e-01 3.26697007e-02 8.80901515e-01 2.85546258e-02
-3.43966521e-02 -5.31334043e-01 -4.93541241e-01 -3.60550851e-01
-7.70500660e-01 -3.34365144e-02 3.43963236e-01 -4.14528549e-01
9.67812091e-02 -8.35051954e-01 -3.37734103e-01 6.23618811e-02
-2.23247498e-01 2.61420697e-01 3.96398574e-01 -5.26106596e-01
1.34532785e+00 -2.21958470e+00 -5.18368900e-01 4.14357573e-01
2.50005782e-01 4.95534688e-01 5.96681118e-01 9.18858498e-02
-8.71790498e-02 -2.07880601e-01 -1.65540293e-01 9.60394442e-02
-2.62270093e-01 1.96379069e-02 -8.44288394e-02 6.89359844e-01
1.92921087e-01 1.07610786e+00 -3.28412920e-01 -4.57857794e-04
2.71606773e-01 8.46077621e-01 -1.70064285e-01 4.43398068e-03
3.70379239e-01 3.12936962e-01 -4.44553107e-01 6.29615009e-01
1.53081447e-01 -4.71576333e-01 5.71860492e-01 -5.68581164e-01
-1.13097340e-01 6.23838484e-01 -1.01508701e+00 1.90978205e+00
-4.85382259e-01 4.97196466e-01 7.36086145e-02 -9.78588581e-01
1.16898918e+00 7.25460291e-01 5.80676079e-01 -6.68186665e-01
6.06147885e-01 3.25111181e-01 1.64967641e-01 -7.12046146e-01
-1.08979689e-02 -2.21279636e-02 -1.54800549e-01 3.49048793e-01
-3.37632485e-02 4.20939058e-01 -4.40896362e-01 -2.88527429e-01
1.54508066e+00 1.74786337e-02 5.81844509e-01 -5.97396195e-01
1.99007198e-01 -2.88924396e-01 6.57517374e-01 5.59571743e-01
-2.03021586e-01 1.48245826e-01 1.68317646e-01 -8.13761711e-01
-6.83690429e-01 -8.92823756e-01 -5.18809676e-01 5.53785205e-01
-3.69726568e-01 -4.64806348e-01 -8.49605918e-01 1.72154099e-01
-7.00303912e-02 7.29694068e-02 -6.00522533e-02 -4.01604235e-01
-6.55194461e-01 -1.02548695e+00 1.26491749e+00 8.87190819e-01
3.98759216e-01 -9.01762843e-01 -1.99278438e+00 7.73583531e-01
2.90709436e-01 -1.07632256e+00 2.11635292e-01 7.79387832e-01
-1.22214758e+00 -6.45462275e-01 -4.82839674e-01 -3.34719092e-01
2.46357247e-01 -5.98947465e-01 1.02789426e+00 -2.56407022e-01
-8.61679316e-01 2.05521718e-01 2.92331763e-02 -8.69585812e-01
1.59169212e-01 3.84955555e-01 6.40028894e-01 -4.44611870e-02
7.80678809e-01 -8.52616847e-01 -9.89975214e-01 -1.61179706e-01
-4.73972648e-01 -2.39400759e-01 7.65588105e-01 6.90171778e-01
3.65072608e-01 -3.60598117e-01 1.14087212e+00 -6.34334743e-01
5.87825775e-01 -3.16329896e-01 -3.71455401e-01 -1.57244578e-01
-8.62775087e-01 -1.45808801e-01 1.09241724e+00 -4.91547346e-01
-5.94707549e-01 6.08172357e-01 -5.14185309e-01 -2.71372497e-01
-9.18952599e-02 4.89986122e-01 -1.17064200e-01 1.53394439e-03
1.09390295e+00 1.91489533e-02 1.44927755e-01 -2.69986868e-01
-2.00396016e-01 1.16215885e+00 7.82818735e-01 -2.28060469e-01
1.32500930e-02 1.04354337e-01 8.42659250e-02 -9.62373674e-01
-1.73754707e-01 -3.21143091e-01 -2.35162690e-01 -1.02794662e-01
6.33710504e-01 -1.37456095e+00 -1.37720287e+00 5.09001255e-01
-1.14653313e+00 -2.01035514e-01 -3.57941151e-01 6.94278657e-01
-3.76127690e-01 -1.39173463e-01 -9.32232261e-01 -1.04118264e+00
-1.40036380e+00 -7.73845494e-01 7.94420719e-01 3.73214483e-01
-8.55093062e-01 -5.56671023e-01 -1.49905398e-01 -2.79307485e-01
8.41481030e-01 4.17230964e-01 5.34179032e-01 -4.66590881e-01
1.19660221e-01 -4.26005810e-01 2.63847679e-01 3.44111413e-01
2.01973557e-01 -4.59884644e-01 -1.48053586e+00 -3.56916130e-01
6.47627592e-01 -1.11434415e-01 2.23547623e-01 5.88158369e-01
1.30990255e+00 -1.99510470e-01 -5.53876102e-01 6.11800730e-01
1.39901114e+00 3.56480986e-01 7.34722912e-01 -3.02342057e-01
3.63507450e-01 -1.31422237e-01 -7.40739554e-02 6.43692732e-01
-4.63749021e-02 3.86149198e-01 6.49226233e-02 -2.20060349e-01
5.40798664e-01 2.90588975e-01 3.33374768e-01 1.11423755e+00
-7.96091706e-02 2.27932557e-01 -1.00881624e+00 2.22761512e-01
-1.62950158e+00 -5.06632805e-01 -3.87826145e-01 2.18160820e+00
9.60637391e-01 2.35771582e-01 1.25181109e-01 7.53110647e-01
3.89272988e-01 -4.16010082e-01 -9.27907050e-01 -6.77410841e-01
2.03735083e-01 7.79331803e-01 5.89061379e-01 -1.86051190e-01
-8.39015901e-01 5.85327856e-02 5.99484205e+00 2.48349175e-01
-1.65547419e+00 2.03872785e-01 7.65004218e-01 -4.59328771e-01
4.90028173e-01 -7.42372751e-01 -6.17579818e-01 6.87493145e-01
2.06919169e+00 1.44087346e-02 2.04117242e-02 1.00590050e+00
2.12037846e-01 4.15810868e-02 -1.20822072e+00 1.47330487e+00
-2.04867050e-01 -1.07617831e+00 -6.91601038e-01 -1.59013331e-01
1.29612144e-02 -9.04744640e-02 -1.94241136e-01 9.53878388e-02
-8.02454650e-01 -1.03358400e+00 1.78979069e-01 5.73734760e-01
1.30815804e+00 -8.28025579e-01 1.02413797e+00 3.11756402e-01
-1.12325144e+00 -1.25553504e-01 -2.27480784e-01 -6.92070603e-01
-5.67118563e-02 1.37826669e+00 -7.36002862e-01 1.85253739e-01
1.17023671e+00 5.11030376e-01 -1.32453680e-01 4.86555666e-01
1.92760095e-01 9.35576975e-01 -5.88892877e-01 -5.89388192e-01
-4.52919751e-01 2.26505324e-01 3.95762026e-02 1.26069963e+00
4.00394052e-01 4.16805476e-01 -3.31219763e-01 6.89503253e-01
-1.61398146e-02 -2.52807051e-01 -4.62415993e-01 2.72028983e-01
4.10923779e-01 1.50588584e+00 -5.04172385e-01 -4.25084591e-01
-2.08270937e-01 6.20510221e-01 -1.84059247e-01 -4.46231999e-02
-1.06080604e+00 -1.05860639e+00 6.41387403e-01 2.42233306e-01
-2.15792060e-01 -1.93693563e-01 -7.90882289e-01 -8.33212614e-01
4.63443547e-01 -7.28298545e-01 -2.12790444e-01 -3.26537162e-01
-9.28035736e-01 6.81311250e-01 -4.85841125e-01 -1.14543772e+00
-4.37701970e-01 -6.84470057e-01 -4.36377585e-01 1.01913083e+00
-9.08507943e-01 -2.97485083e-01 -3.86047482e-01 5.83516717e-01
6.03884421e-02 5.09599932e-02 1.44204688e+00 6.01274073e-01
-6.00205898e-01 7.50779808e-01 -1.16280906e-01 1.51088431e-01
3.70498329e-01 -8.03641319e-01 6.54908836e-01 5.95510900e-01
-4.25720185e-01 8.40778887e-01 4.63189244e-01 -3.54836583e-01
-1.90025806e+00 -1.15079904e+00 8.14954400e-01 4.56935950e-02
4.70888764e-01 -6.08683050e-01 -8.71979713e-01 2.68266857e-01
-9.44237933e-02 3.24660867e-01 7.89031208e-01 -1.03162445e-01
7.02417344e-02 -7.49723494e-01 -1.25427878e+00 5.56166887e-01
7.80946910e-01 -6.45828724e-01 -2.61032313e-01 5.58911599e-02
2.84340948e-01 -4.75071222e-01 -1.30930769e+00 6.27495587e-01
9.98993933e-01 -6.67003274e-01 8.73875678e-01 -2.13271767e-01
6.55452954e-03 -1.40884638e-01 1.41927719e-01 -9.72909868e-01
-2.13638678e-01 -9.84287441e-01 -4.69132006e-01 7.83669412e-01
4.16330099e-01 -9.28123832e-01 9.44459498e-01 8.32703233e-01
-1.04832195e-01 -1.13488460e+00 -1.26285899e+00 -5.90862989e-01
-4.06827390e-01 -4.28968757e-01 4.05383795e-01 3.56411695e-01
7.82809079e-01 5.78727186e-01 -3.30141634e-01 -1.19312435e-01
2.99590856e-01 -4.22889322e-01 6.03951737e-02 -1.28852928e+00
-4.09859747e-01 2.48282440e-02 -7.59176075e-01 -6.14656270e-01
-3.75272870e-01 -4.23744917e-01 8.67427140e-02 -8.00050974e-01
-1.18968435e-01 -2.65162498e-01 -6.71504557e-01 6.92774057e-01
1.13972813e-01 4.69971716e-01 -4.71768498e-01 -1.84108153e-01
-2.93270230e-01 -4.88379262e-02 7.07837045e-02 -1.58580720e-01
-3.54298115e-01 -3.96280512e-02 -5.03829598e-01 8.72863591e-01
1.27484822e+00 -5.24961591e-01 -2.97583550e-01 -2.51321048e-01
4.27715421e-01 3.34010869e-01 2.74505734e-01 -1.71826971e+00
5.48830092e-01 7.12121725e-01 8.50240588e-01 -2.35519633e-01
5.40022254e-01 -9.13939655e-01 5.55978656e-01 1.20774090e+00
-2.10341483e-01 3.91967237e-01 3.93298328e-01 1.52999520e-01
1.62977919e-01 9.37687904e-02 5.34653485e-01 1.37048587e-01
3.74977365e-02 -6.99589178e-02 -7.75384963e-01 -2.54711628e-01
8.49068046e-01 -1.60800964e-01 -2.07834885e-01 1.44837528e-01
-6.66706324e-01 -3.69154811e-01 -1.82349190e-01 -1.25497496e-02
7.15882003e-01 -1.14568114e+00 -5.29603899e-01 4.79103655e-01
-3.90413366e-02 -2.11569026e-01 5.36755800e-01 1.04467070e+00
-3.37110966e-01 6.60165727e-01 -4.23316658e-01 -8.82400393e-01
-1.11802447e+00 1.48749307e-01 5.03561735e-01 1.36056662e-01
-7.85973787e-01 2.87818819e-01 -7.09891260e-01 2.19131038e-01
2.39697576e-01 -8.79211843e-01 1.54980561e-02 -8.07000846e-02
7.42949128e-01 6.22740626e-01 8.08503926e-01 2.01857537e-01
-8.65895867e-01 4.67221230e-01 3.80062431e-01 2.66532153e-01
1.11347389e+00 3.54930788e-01 3.44345011e-02 7.19877958e-01
1.12424397e+00 -6.10796869e-01 -6.04010761e-01 1.33775353e-01
8.92443210e-02 4.08434510e-01 7.66012892e-02 -7.69943714e-01
-1.03676641e+00 8.54304373e-01 1.20010221e+00 2.17606530e-01
1.51058722e+00 -6.25261009e-01 9.43496704e-01 7.41295338e-01
5.54613769e-01 -9.70432699e-01 -4.30343300e-01 1.48286387e-01
3.13126355e-01 -8.17556739e-01 8.68186206e-02 2.21872441e-02
-1.13133110e-01 1.24898434e+00 1.59796730e-01 -1.62573189e-01
8.89926493e-01 8.82283449e-01 1.22874983e-01 6.48277700e-02
-7.77371585e-01 4.16668445e-01 -1.22409791e-01 4.91431355e-01
7.85932124e-01 3.52179348e-01 -3.45183730e-01 1.15737164e+00
-8.50324035e-02 8.04076910e-01 4.29556698e-01 1.06603587e+00
-9.86504778e-02 -5.65846741e-01 -1.32597506e-01 1.08782220e+00
-9.71047878e-01 -1.45997733e-01 6.05752729e-02 1.90152824e-01
-2.34417003e-02 1.00691330e+00 3.00296396e-01 -5.71283400e-01
4.38497245e-01 1.99387774e-01 1.76097095e-01 -2.46029139e-01
-1.14710224e+00 2.47106235e-02 -3.25544993e-03 -7.53660500e-01
-3.35176468e-01 -4.61395055e-01 -1.24721658e+00 -2.51873769e-02
1.24247493e-02 -1.13308489e-01 1.11195767e+00 6.17340565e-01
9.13694561e-01 1.15110731e+00 3.03614557e-01 -7.32865691e-01
-7.13661432e-01 -1.15626419e+00 -6.89621091e-01 -3.09974551e-01
3.40684265e-01 -6.60092905e-02 -3.23793888e-02 4.79827635e-02] | [13.98394775390625, 3.1576499938964844] |
a8687153-f5e7-4a1f-956f-73aeece4cbb2 | zeromesh-zero-shot-single-view-3d-mesh | 2208.02676 | null | https://arxiv.org/abs/2208.02676v2 | https://arxiv.org/pdf/2208.02676v2.pdf | Single-view 3D Mesh Reconstruction for Seen and Unseen Categories | Single-view 3D object reconstruction is a fundamental and challenging computer vision task that aims at recovering 3D shapes from single-view RGB images. Most existing deep learning based reconstruction methods are trained and evaluated on the same categories, and they cannot work well when handling objects from novel categories that are not seen during training. Focusing on this issue, this paper tackles Single-view 3D Mesh Reconstruction, to study the model generalization on unseen categories and encourage models to reconstruct objects literally. Specifically, we propose an end-to-end two-stage network, GenMesh, to break the category boundaries in reconstruction. Firstly, we factorize the complicated image-to-mesh mapping into two simpler mappings, i.e., image-to-point mapping and point-to-mesh mapping, while the latter is mainly a geometric problem and less dependent on object categories. Secondly, we devise a local feature sampling strategy in 2D and 3D feature spaces to capture the local geometry shared across objects to enhance model generalization. Thirdly, apart from the traditional point-to-point supervision, we introduce a multi-view silhouette loss to supervise the surface generation process, which provides additional regularization and further relieves the overfitting problem. The experimental results show that our method significantly outperforms the existing works on the ShapeNet and Pix3D under different scenarios and various metrics, especially for novel objects. The project link is https://github.com/Wi-sc/GenMesh. | ['Luping Zhou', 'Guosheng Lin', 'Xianghui Yang'] | 2022-08-04 | null | null | null | null | ['3d-object-reconstruction', 'object-reconstruction'] | ['computer-vision', 'computer-vision'] | [-3.94531712e-02 -2.37095891e-03 2.18422189e-01 -5.28598487e-01
-6.35980546e-01 -4.02480990e-01 5.22911131e-01 -4.14129615e-01
3.76110524e-02 2.00394705e-01 -4.34451140e-02 1.37439907e-01
-2.00482216e-02 -9.20580924e-01 -1.08668387e+00 -5.98105073e-01
5.79910159e-01 7.59626031e-01 3.28294873e-01 -1.89052999e-01
1.00213289e-01 8.70380700e-01 -1.61283422e+00 1.21891968e-01
6.18790090e-01 1.03414881e+00 3.05424124e-01 1.08596265e-01
-5.55390008e-02 2.24834517e-01 3.35803702e-02 -4.14175481e-01
4.95205253e-01 -4.77928445e-02 -5.71351111e-01 5.10462046e-01
6.14750266e-01 -3.51498663e-01 -3.85840625e-01 9.47813869e-01
5.83399475e-01 -1.42234832e-03 6.07463241e-01 -9.65580761e-01
-7.74142325e-01 -2.13308283e-03 -8.17857444e-01 -2.76900083e-01
4.16995585e-01 9.53236520e-02 6.58970594e-01 -1.54182303e+00
7.06283987e-01 1.44327784e+00 7.68935740e-01 5.58953762e-01
-1.06230092e+00 -6.03733778e-01 3.22817296e-01 -1.16130956e-01
-1.39636600e+00 -2.84460247e-01 1.32410288e+00 -3.65078509e-01
5.98814547e-01 1.31122410e-01 5.41735590e-01 1.06626964e+00
-9.91545394e-02 8.16011429e-01 1.05523837e+00 -3.44222635e-02
-5.89325326e-03 5.78912087e-02 -2.81704992e-01 6.88562036e-01
-1.11760395e-02 1.32143021e-01 -2.49150723e-01 1.70984305e-02
1.32658601e+00 4.70199198e-01 -5.64290881e-01 -9.46096182e-01
-1.18273211e+00 5.92834473e-01 5.49849272e-01 -1.80786848e-02
-4.37953174e-01 2.53431704e-02 1.95614338e-01 2.78253853e-03
8.48693669e-01 3.38795915e-04 -6.97871387e-01 2.33503327e-01
-5.07158339e-01 2.35772520e-01 3.14642370e-01 1.21875405e+00
9.63832378e-01 4.65791002e-02 3.38212520e-01 1.03418136e+00
5.49911082e-01 6.56360805e-01 6.93141446e-02 -8.69039714e-01
6.05041504e-01 9.27974999e-01 1.64909605e-02 -9.43025351e-01
-3.64269018e-01 -6.14133954e-01 -8.65729153e-01 2.01124862e-01
4.15120050e-02 3.52015465e-01 -1.15723097e+00 1.32068646e+00
7.57406116e-01 1.32709384e-01 -2.57445604e-01 1.17167199e+00
1.12818003e+00 4.21054721e-01 -3.45307738e-01 1.06152920e-02
1.11710048e+00 -9.88895297e-01 -2.48561442e-01 -1.21344164e-01
1.93413660e-01 -7.65076876e-01 1.28790200e+00 3.90830934e-01
-1.44112861e+00 -6.73204124e-01 -6.83518767e-01 -2.86613464e-01
-1.76336005e-01 2.54950020e-02 5.02181947e-01 5.35958782e-02
-6.21172130e-01 4.83220875e-01 -8.11338305e-01 -3.36260311e-02
7.19611585e-01 1.43555462e-01 -5.88813722e-01 -5.26408076e-01
-7.40940630e-01 5.52717924e-01 1.41391337e-01 2.76951820e-01
-9.51404631e-01 -9.08055604e-01 -8.33883286e-01 -1.86304480e-01
5.12336373e-01 -1.13633299e+00 9.65992451e-01 -7.02166557e-01
-1.39728594e+00 1.23447454e+00 -6.55503646e-02 3.72190028e-01
7.26175904e-01 -1.24327689e-01 -8.95729735e-02 1.36882728e-02
1.40108302e-01 4.02797461e-01 9.68102157e-01 -2.02612233e+00
-3.27684492e-01 -7.85295606e-01 9.34586376e-02 4.32956964e-01
4.05866280e-02 -4.17395562e-01 -7.35894382e-01 -7.41082132e-01
7.88759291e-01 -6.86979234e-01 -3.73368226e-02 3.15721244e-01
-3.65024060e-01 -3.42717975e-01 8.62765431e-01 -5.83792269e-01
5.47893524e-01 -2.10132074e+00 5.54845572e-01 2.76688691e-02
1.87603906e-01 -5.91359250e-02 8.78768228e-03 1.52629435e-01
-8.21722224e-02 -7.79404268e-02 -3.64770234e-01 -7.48825550e-01
-7.49327242e-02 2.86071569e-01 -2.24414781e-01 6.05974793e-01
1.39243990e-01 9.80458200e-01 -7.13128388e-01 -2.20820814e-01
3.95239234e-01 6.46853447e-01 -5.28887868e-01 3.34788382e-01
-2.72384703e-01 8.23988914e-01 -5.73802769e-01 9.01803493e-01
1.21652544e+00 -4.54928637e-01 -4.39050674e-01 -4.79661077e-01
4.00025258e-03 6.02095723e-02 -1.19427991e+00 2.29346323e+00
-7.10215092e-01 -2.13924333e-01 2.90284604e-01 -9.88669395e-01
1.02937865e+00 1.37371525e-01 5.58749139e-01 -4.59491551e-01
5.44639528e-02 3.13225001e-01 -5.21471620e-01 -5.36679804e-01
5.27333096e-02 -5.07294178e-01 2.17647582e-01 2.69437999e-01
3.34705189e-02 -4.35898781e-01 -6.18398786e-01 -1.10241711e-01
4.81082797e-01 6.47811413e-01 -1.58231646e-01 6.24814630e-02
5.12411058e-01 -2.16323256e-01 5.51258922e-01 1.06966078e-01
4.08155710e-01 1.26945269e+00 -9.29497108e-02 -6.84593737e-01
-9.96188521e-01 -1.39443088e+00 -3.55182767e-01 6.53725207e-01
6.68888152e-01 4.03151400e-02 -3.49099636e-01 -8.51689041e-01
1.97463334e-01 5.16921997e-01 -5.35362303e-01 -1.38083011e-01
-6.80889547e-01 -4.81060147e-01 -3.98382097e-02 5.74884772e-01
5.26829541e-01 -1.03204894e+00 -4.56949919e-01 2.90813781e-02
-1.17881104e-01 -1.04755294e+00 -4.38821942e-01 -1.47821262e-01
-1.13012516e+00 -1.07879245e+00 -9.18011606e-01 -8.66618454e-01
8.79987240e-01 5.71786523e-01 1.13347089e+00 1.29701599e-01
-9.97200459e-02 6.93355560e-01 -3.72137129e-01 -2.68089592e-01
-1.15570888e-01 -7.21410215e-02 -2.28418782e-03 2.29554638e-01
7.13192523e-02 -1.00787139e+00 -7.80193269e-01 6.08250439e-01
-9.21843410e-01 3.78564864e-01 6.13922834e-01 7.49489129e-01
1.12528300e+00 -9.34040025e-02 3.90692592e-01 -7.32704639e-01
-9.22017768e-02 -5.05084276e-01 -2.97182083e-01 1.79668948e-01
-3.86729062e-01 -3.65165681e-01 6.18424177e-01 -4.06774372e-01
-1.09184170e+00 2.99264580e-01 -3.42526168e-01 -1.16215968e+00
-3.74192744e-01 2.10303009e-01 -5.60738385e-01 -2.10457146e-01
2.55976915e-01 6.22991264e-01 6.95740506e-02 -9.51189339e-01
2.57681459e-01 4.02730048e-01 2.70526975e-01 -4.76214379e-01
1.01635504e+00 9.05340970e-01 -1.04664922e-01 -6.72062874e-01
-9.45805967e-01 -4.23326522e-01 -7.25291610e-01 -3.67714316e-01
7.16014266e-01 -1.16320801e+00 -4.53100175e-01 5.83148956e-01
-1.19400883e+00 -1.68774456e-01 -3.24341804e-01 4.29353714e-01
-8.58474910e-01 5.03187060e-01 -4.01655704e-01 -4.39444989e-01
-3.58005077e-01 -1.12658906e+00 1.70030701e+00 1.09504752e-01
5.03076792e-01 -8.94955277e-01 -2.29618102e-01 6.83825016e-01
5.58104403e-02 5.10401428e-01 8.58883917e-01 -3.74417193e-02
-8.42759490e-01 -5.66436723e-03 -3.23897809e-01 4.91644114e-01
1.79382220e-01 -3.74122232e-01 -1.04561162e+00 -4.45464492e-01
4.29188341e-01 -3.59995693e-01 7.31656909e-01 2.05470920e-01
1.41880214e+00 -8.53688456e-03 -2.76880205e-01 1.04263699e+00
1.40163946e+00 -1.33567736e-01 6.25275970e-01 1.52470514e-01
1.26669168e+00 5.04420280e-01 5.74045420e-01 3.78121138e-01
6.29334211e-01 8.52220297e-01 8.89411449e-01 -2.56213129e-01
-2.93573290e-01 -5.59220314e-01 4.02765200e-02 9.88824368e-01
-3.07151675e-01 -4.71739136e-02 -7.97163963e-01 4.79015619e-01
-1.64775848e+00 -4.67724562e-01 -3.45612884e-01 2.11541080e+00
5.18192589e-01 5.09768017e-02 -1.70419112e-01 -1.47410020e-01
4.89269584e-01 -4.87587079e-02 -9.22052622e-01 2.16548517e-01
-3.33828889e-02 1.75963253e-01 1.58275947e-01 3.34982187e-01
-8.20191741e-01 8.97092521e-01 4.29862356e+00 9.18534040e-01
-1.22549236e+00 2.26546496e-01 5.21213412e-01 -2.23553888e-02
-6.26112878e-01 -1.56425357e-01 -5.83015442e-01 3.38954896e-01
-5.71879968e-02 4.72468793e-01 4.08630431e-01 8.96010876e-01
3.79809830e-03 2.73285300e-01 -1.14159799e+00 1.26606059e+00
2.55625248e-01 -1.14096701e+00 3.34549546e-01 1.53590083e-01
7.17049301e-01 1.39016822e-01 6.14486448e-02 1.00053765e-01
-2.05506384e-01 -9.27023470e-01 1.05194139e+00 7.88586080e-01
9.37848866e-01 -6.89292133e-01 5.58348060e-01 6.16430223e-01
-1.20945334e+00 2.20680892e-01 -6.09128892e-01 2.02998936e-01
4.20071334e-01 6.02465808e-01 -2.77608424e-01 1.00792491e+00
8.68701994e-01 1.06461489e+00 -4.08634961e-01 8.47658873e-01
-1.50667474e-01 1.25347316e-01 -3.11149001e-01 5.66360056e-01
1.04663163e-01 -3.03351253e-01 6.62745833e-01 5.21323383e-01
3.78723711e-01 2.02212736e-01 2.22989127e-01 1.24478209e+00
-7.79714510e-02 -9.54242721e-02 -6.72897756e-01 6.65377855e-01
1.96596384e-01 1.25151157e+00 -7.02180564e-01 -1.43292949e-01
-4.62090582e-01 1.11499560e+00 3.68158996e-01 3.22404951e-01
-7.94479191e-01 4.29102443e-02 3.45762432e-01 6.54160440e-01
5.29753029e-01 -1.43945962e-01 -3.59508395e-01 -1.36785948e+00
5.01363158e-01 -5.28684855e-01 7.09720713e-04 -1.08874464e+00
-1.67137194e+00 4.76546496e-01 1.17587417e-01 -1.50592995e+00
3.09713513e-01 -5.09988487e-01 -4.73327130e-01 8.16745460e-01
-1.63480949e+00 -1.48605907e+00 -5.84737718e-01 6.85117424e-01
7.52420068e-01 6.40069470e-02 5.10358810e-01 4.37375546e-01
-2.30763599e-01 4.91526723e-01 -1.00218832e-01 -1.66975141e-01
4.62113231e-01 -8.78792167e-01 3.56250793e-01 3.85970324e-01
-4.37919758e-02 5.09949207e-01 1.88043520e-01 -6.06716871e-01
-1.69427812e+00 -1.31345797e+00 3.68211359e-01 -6.71605349e-01
3.68706696e-02 -5.31046867e-01 -1.12330937e+00 6.67969763e-01
-3.84535998e-01 4.08904046e-01 1.59292340e-01 -2.01608807e-01
-3.84832889e-01 7.31631648e-03 -1.26621127e+00 2.94491649e-01
1.65200698e+00 -4.00467664e-01 -5.22514403e-01 3.49074066e-01
7.57729530e-01 -8.27526689e-01 -1.16197670e+00 7.79322684e-01
4.56443489e-01 -1.05608606e+00 1.36134171e+00 -4.80543703e-01
5.87621152e-01 -4.74093318e-01 -3.08221191e-01 -1.24318266e+00
-3.98247600e-01 -2.53090590e-01 -1.29085273e-01 1.12589860e+00
3.89133692e-02 -6.23855829e-01 9.33855653e-01 3.14584136e-01
-6.32254124e-01 -1.28237331e+00 -1.08488512e+00 -7.18280256e-01
2.06387579e-01 -2.87168503e-01 7.23906100e-01 1.09264064e+00
-8.59720588e-01 3.43353957e-01 -2.23640367e-01 3.54328722e-01
7.95480072e-01 6.46984279e-01 8.99221659e-01 -1.37484682e+00
-2.55532116e-01 -3.04441720e-01 -3.37110519e-01 -1.39552534e+00
-5.55282496e-02 -1.12232304e+00 -7.50938207e-02 -1.60970473e+00
1.26689613e-01 -7.77817309e-01 3.41061167e-02 2.41279677e-01
-3.91639136e-02 3.50246459e-01 6.38809726e-02 3.44121933e-01
-4.38471913e-01 1.15733230e+00 1.93176425e+00 4.79579493e-02
-9.09400210e-02 2.85022408e-01 -6.42788827e-01 9.72152233e-01
5.43899894e-01 -4.00811732e-01 -3.54647815e-01 -8.21287870e-01
1.73330769e-01 -6.18798891e-03 7.70037591e-01 -6.27476394e-01
-7.99251273e-02 -2.14832261e-01 4.97398019e-01 -9.56011832e-01
6.46354198e-01 -1.15452695e+00 3.78636390e-01 8.21234286e-02
2.17493549e-01 -1.00581177e-01 -3.07916733e-03 7.72790968e-01
1.50393564e-02 -1.67077761e-02 9.05407250e-01 -4.91208643e-01
-4.52035129e-01 8.55933487e-01 4.50019389e-01 1.07592449e-01
1.00102627e+00 -4.91438419e-01 9.89767537e-02 -5.74456491e-02
-8.56671095e-01 2.53678411e-01 9.08320725e-01 5.77586234e-01
1.15854883e+00 -1.66117644e+00 -7.48870015e-01 4.73207504e-01
3.01740110e-01 1.02546525e+00 5.68767786e-01 8.35424781e-01
-4.74620730e-01 -8.44125152e-02 1.67533830e-02 -1.01654232e+00
-9.18053210e-01 5.99528611e-01 5.03193378e-01 6.37096316e-02
-1.13959384e+00 1.02687716e+00 7.30356276e-01 -1.05339944e+00
3.47334594e-01 -2.51719505e-01 9.19137821e-02 -3.36912364e-01
1.47162095e-01 2.37284347e-01 1.38582110e-01 -7.95771956e-01
-2.74149179e-01 1.30343413e+00 -2.43753120e-02 3.97581279e-01
1.77164710e+00 -1.78240031e-01 -1.87565967e-01 5.60963452e-01
1.30053258e+00 -2.13582262e-01 -1.55277991e+00 -4.04451102e-01
-5.27911425e-01 -6.61514282e-01 3.16622071e-02 -6.28310978e-01
-1.37264931e+00 1.10301220e+00 6.42782450e-01 -1.48590907e-01
9.81055856e-01 3.23333889e-01 1.02248740e+00 1.91942137e-02
6.96696877e-01 -7.38674760e-01 1.96013302e-01 4.82720673e-01
1.36039877e+00 -1.24009693e+00 1.17154911e-01 -6.81370378e-01
-3.94518584e-01 1.02788365e+00 8.06421340e-01 -3.59612882e-01
8.31308722e-01 -2.85264671e-01 -2.43668005e-01 -6.99112415e-01
-2.74109989e-01 1.40315399e-01 4.25697416e-01 5.75467825e-01
-8.04985389e-02 -1.15829883e-02 1.75900951e-01 6.41700685e-01
-1.40863918e-02 -1.40630081e-01 2.39178479e-01 6.97504878e-01
-2.06778079e-01 -8.60445321e-01 -3.47862810e-01 3.90609592e-01
-2.91503947e-02 2.29444891e-01 -2.62261927e-01 7.39869237e-01
4.49698299e-01 4.08385485e-01 5.51933274e-02 -4.84027535e-01
8.63543749e-01 -1.74910009e-01 6.71646714e-01 -8.84567976e-01
-2.88493723e-01 1.41506061e-01 -3.55833828e-01 -5.40099323e-01
-5.36308646e-01 -6.44966364e-01 -1.25840914e+00 -1.62757650e-01
-4.12867069e-01 -3.56800407e-01 5.19999862e-01 7.23026872e-01
4.28086460e-01 4.24510688e-01 1.02381718e+00 -1.40791488e+00
-5.13111949e-01 -7.74641991e-01 -5.83867133e-01 6.52141452e-01
2.74745226e-01 -1.13926327e+00 -4.35513377e-01 -1.82637513e-01] | [8.42149543762207, -3.2146761417388916] |
b8998571-5eab-47ae-8b4c-0f3cc21fa030 | avoiding-inference-heuristics-in-few-shot | 2109.04144 | null | https://arxiv.org/abs/2109.04144v1 | https://arxiv.org/pdf/2109.04144v1.pdf | Avoiding Inference Heuristics in Few-shot Prompt-based Finetuning | Recent prompt-based approaches allow pretrained language models to achieve strong performances on few-shot finetuning by reformulating downstream tasks as a language modeling problem. In this work, we demonstrate that, despite its advantages on low data regimes, finetuned prompt-based models for sentence pair classification tasks still suffer from a common pitfall of adopting inference heuristics based on lexical overlap, e.g., models incorrectly assuming a sentence pair is of the same meaning because they consist of the same set of words. Interestingly, we find that this particular inference heuristic is significantly less present in the zero-shot evaluation of the prompt-based model, indicating how finetuning can be destructive to useful knowledge learned during the pretraining. We then show that adding a regularization that preserves pretraining weights is effective in mitigating this destructive tendency of few-shot finetuning. Our evaluation on three datasets demonstrates promising improvements on the three corresponding challenge datasets used to diagnose the inference heuristics. | ['Iryna Gurevych', 'Victor Sanh', 'Nafise Sadat Moosavi', 'Prasetya Ajie Utama'] | 2021-09-09 | null | https://aclanthology.org/2021.emnlp-main.713 | https://aclanthology.org/2021.emnlp-main.713.pdf | emnlp-2021-11 | ['sentence-pair-classification'] | ['natural-language-processing'] | [ 4.27272797e-01 2.24716693e-01 -1.67004913e-01 -5.52839279e-01
-9.59413469e-01 -5.08605897e-01 8.06299150e-01 5.22851706e-01
-6.37544751e-01 7.33694136e-01 4.22932535e-01 -6.60513878e-01
-1.55977517e-01 -7.21562266e-01 -7.92459130e-01 -6.06839418e-01
3.22314620e-01 5.77386379e-01 1.82319567e-01 -5.06702960e-01
3.71587396e-01 4.02682880e-03 -1.55355930e+00 6.09170437e-01
6.16883039e-01 5.82328200e-01 2.11919919e-01 6.79212451e-01
-2.84134686e-01 8.49007487e-01 -6.34718001e-01 -5.64995527e-01
2.43775770e-01 -4.25734103e-01 -1.01667690e+00 -2.57604212e-01
8.56112897e-01 -2.82098442e-01 -2.59115338e-01 9.38574016e-01
4.25789028e-01 5.69294035e-01 6.73640490e-01 -8.73653948e-01
-5.28704524e-01 1.23779786e+00 -2.19422668e-01 6.48399770e-01
1.63406566e-01 3.52011532e-01 1.39460075e+00 -9.00572538e-01
5.89490056e-01 1.44884193e+00 8.08036566e-01 6.08473778e-01
-1.70036566e+00 -5.39361656e-01 3.60750407e-01 1.78763121e-01
-1.17223549e+00 -7.72940516e-01 3.94278437e-01 -3.08058560e-01
1.74871635e+00 2.57197350e-01 1.73918009e-01 1.48425794e+00
2.81261742e-01 6.65837288e-01 9.37799513e-01 -6.43736243e-01
2.41158813e-01 -4.11278419e-02 7.64645696e-01 6.31464660e-01
1.19810581e-01 2.28877082e-01 -7.26307869e-01 -2.07998633e-01
1.05036445e-01 -1.84072733e-01 -3.95390727e-02 3.19629014e-02
-8.06635797e-01 9.32410657e-01 1.39143348e-01 4.33472306e-01
1.71593595e-02 -1.03343688e-02 6.65012300e-01 6.41678274e-01
5.95183611e-01 8.45870674e-01 -6.80002630e-01 -4.57864404e-01
-1.06345677e+00 2.95582712e-01 7.82671630e-01 7.85686910e-01
5.63124537e-01 -2.52825201e-01 -7.04845965e-01 8.57739568e-01
5.57005443e-02 2.56784745e-02 6.81639075e-01 -7.89242029e-01
6.30603790e-01 4.02317822e-01 -8.57577398e-02 -5.54440320e-01
-3.06622028e-01 -4.40107137e-01 -4.71902937e-01 -3.89304101e-01
5.70906639e-01 -1.30006701e-01 -8.78300011e-01 2.02375817e+00
-2.09975675e-01 1.77617446e-01 4.87738959e-02 5.92719018e-01
6.58456862e-01 5.05010784e-01 5.81567824e-01 -3.76268506e-01
1.20698988e+00 -9.55972493e-01 -5.42342842e-01 -6.96615875e-01
1.16740263e+00 -6.94088042e-01 1.59589064e+00 3.17393392e-01
-1.06256795e+00 -4.70744550e-01 -1.14158392e+00 -3.91623020e-01
-2.95332938e-01 -4.37887073e-01 6.28738999e-01 4.41458493e-01
-9.35954511e-01 8.63146782e-01 -5.24483979e-01 -5.02756596e-01
2.34163418e-01 5.75195625e-02 -1.45436255e-02 -1.32520825e-01
-1.51390886e+00 1.18111181e+00 5.01283824e-01 -3.26607019e-01
-7.56833851e-01 -1.05391335e+00 -6.35404646e-01 6.29387379e-01
5.85681438e-01 -6.55524194e-01 1.56690514e+00 -7.47975707e-01
-1.17138422e+00 1.05443430e+00 -2.41975084e-01 -7.89252222e-01
3.09179425e-01 -3.58374059e-01 -1.91412538e-01 -1.89822018e-01
1.57083273e-01 4.62995380e-01 8.62238348e-01 -8.09281230e-01
-5.90219200e-01 -1.82884648e-01 1.18065126e-01 1.06992371e-01
-4.45339203e-01 -2.40076636e-03 -6.00454025e-02 -5.84064782e-01
-1.94057778e-01 -5.58047712e-01 -2.17037469e-01 -6.29276991e-01
-3.91738683e-01 -4.91235942e-01 4.51085895e-01 -1.85978517e-01
1.56310940e+00 -1.92575479e+00 -4.69660871e-02 -4.34528627e-02
2.23202571e-01 4.74492610e-01 -4.58188593e-01 4.47414011e-01
-1.94344476e-01 4.66092408e-01 -1.45214140e-01 -6.15110993e-01
5.97082153e-02 2.98649639e-01 -7.52651691e-01 4.65268493e-02
1.95944294e-01 8.54187310e-01 -1.05752933e+00 -4.13595557e-01
1.10878274e-01 -7.12066963e-02 -5.26613653e-01 1.40179276e-01
-4.17154312e-01 -1.05263934e-01 1.83130056e-02 1.52536958e-01
2.35063940e-01 -1.28975362e-01 3.07901353e-01 -1.68444559e-01
2.13513866e-01 8.92287612e-01 -6.50522113e-01 1.57190156e+00
-5.00784993e-01 5.33488154e-01 -2.66831368e-01 -1.18360698e+00
5.77613652e-01 2.22077623e-01 -5.29995263e-02 -7.82304883e-01
2.41805181e-01 -9.87820444e-04 3.23703974e-01 -5.79203486e-01
6.52108490e-01 -7.02019513e-01 -2.29408771e-01 5.46930969e-01
3.36884797e-01 -2.43168902e-02 4.16754574e-01 5.72356462e-01
1.32483876e+00 -3.00700307e-01 2.97979325e-01 -4.85715121e-01
1.45603985e-01 1.69216804e-02 4.27732259e-01 1.47693789e+00
-2.06347108e-01 2.74134189e-01 5.72267115e-01 -2.80512720e-01
-1.03714073e+00 -1.16390514e+00 -1.80758744e-01 1.63692057e+00
-1.46048397e-01 -8.94883990e-01 -5.73848009e-01 -8.14976335e-01
3.86338122e-02 1.32353222e+00 -5.24639964e-01 -8.31051409e-01
-3.50909889e-01 -7.92606831e-01 8.66497219e-01 5.60623348e-01
9.50038284e-02 -7.50192940e-01 -4.33829159e-01 3.17467272e-01
-2.98076361e-01 -9.93945897e-01 -4.20723975e-01 7.11502075e-01
-8.59924376e-01 -8.30686450e-01 -6.90417364e-02 -5.49959600e-01
4.45274889e-01 3.61593604e-01 1.69280100e+00 2.70112336e-01
-1.60007238e-01 5.71444966e-02 -3.45861316e-01 -2.19174147e-01
-6.79093599e-01 2.82853305e-01 1.82755649e-01 -4.64683622e-01
8.41770232e-01 -4.62203592e-01 -7.04534575e-02 -1.73719972e-02
-6.36646450e-01 -3.33396643e-02 4.20596063e-01 1.22559798e+00
2.52562135e-01 -1.17801607e-01 6.49364412e-01 -1.13105941e+00
1.04604900e+00 -5.36556065e-01 -2.10464418e-01 4.75277960e-01
-7.64798522e-01 5.48741639e-01 6.95987880e-01 -4.80425477e-01
-1.07810020e+00 -4.68750656e-01 -1.99730769e-01 -3.45316231e-01
-2.42126510e-01 5.63432574e-01 2.10144788e-01 3.27778339e-01
9.14244771e-01 7.49360351e-03 -1.35064915e-01 -5.98854482e-01
3.37543011e-01 4.07359868e-01 2.57241488e-01 -8.26615453e-01
3.96799982e-01 2.81792078e-02 -2.00421825e-01 -6.55993223e-01
-1.57785439e+00 -5.89721859e-01 -2.86295146e-01 1.03037618e-01
6.30555034e-01 -7.68814027e-01 -5.65453172e-01 5.29635549e-02
-1.08994031e+00 -5.64710915e-01 -3.54565531e-01 1.24552526e-01
-4.26447183e-01 2.72993207e-01 -7.83749640e-01 -6.55584514e-01
-3.33265960e-01 -9.32257771e-01 8.30297649e-01 -2.87229382e-02
-6.57059431e-01 -1.13065624e+00 -3.42505574e-02 2.25399867e-01
3.65507334e-01 -4.94597495e-01 1.36221826e+00 -1.08247948e+00
-5.18789738e-02 4.08495218e-02 -2.29330882e-02 3.03904265e-01
2.89356336e-02 -2.62712166e-02 -1.06277609e+00 -1.85548037e-01
9.91484597e-02 -6.76461697e-01 1.38444674e+00 1.34935811e-01
9.94043291e-01 -3.59731644e-01 -3.10285032e-01 3.04458410e-01
1.25975001e+00 -2.09380805e-01 2.73423731e-01 1.13467023e-01
3.77694428e-01 4.86514330e-01 6.59711659e-01 2.13938698e-01
8.64410773e-02 7.29048669e-01 -1.30899483e-02 3.40685427e-01
-8.58447626e-02 -5.29771984e-01 4.58258659e-01 5.95648766e-01
4.04910654e-01 -2.76741356e-01 -1.03671122e+00 5.01344264e-01
-1.95364177e+00 -1.27516210e+00 2.24821642e-02 2.15940928e+00
1.20297539e+00 6.97648883e-01 -2.20070392e-01 -3.13931108e-02
5.29030144e-01 1.98277727e-01 -3.80054623e-01 -8.98840904e-01
-1.85432807e-01 3.71363550e-01 3.06345254e-01 7.94623077e-01
-1.01634026e+00 1.31758928e+00 6.77062225e+00 1.03895557e+00
-9.77048934e-01 2.11668596e-01 7.17073560e-01 -5.81072807e-01
-4.29507494e-01 -4.03253213e-02 -1.17244422e+00 5.36610723e-01
1.35049999e+00 -3.10150087e-01 4.00235653e-01 6.06198192e-01
2.89948642e-01 -2.66339719e-01 -1.48958075e+00 5.62917531e-01
5.85253686e-02 -1.53502238e+00 3.43542963e-01 -2.66236812e-01
5.79022288e-01 2.08804071e-01 -1.00880593e-01 9.74886239e-01
5.17764986e-01 -9.84897435e-01 6.85982585e-01 3.34380746e-01
4.72604036e-01 -7.00042963e-01 6.17981017e-01 6.94358826e-01
-6.07066989e-01 -1.12901568e-01 -4.78790551e-01 -5.58438182e-01
9.29813161e-02 6.78625584e-01 -1.13432741e+00 2.68090591e-02
3.40441525e-01 4.19699520e-01 -7.25846291e-01 5.23149729e-01
-2.30396092e-01 8.58543575e-01 -1.17628165e-01 -6.94184750e-02
3.72008264e-01 1.55470148e-01 6.09782338e-01 1.37101936e+00
-1.28884912e-01 6.86227158e-02 1.90353274e-01 7.31488109e-01
-1.35547489e-01 -1.54671416e-01 -7.40396917e-01 -2.51486778e-01
5.93823969e-01 1.01259983e+00 -4.85593915e-01 -6.78258955e-01
-3.05160105e-01 7.43855238e-01 9.24289227e-01 2.47314200e-01
-6.48665190e-01 -1.20249785e-01 7.84349322e-01 -1.06045632e-02
2.49352753e-01 -8.11756700e-02 -5.66359997e-01 -1.21965814e+00
-2.57950932e-01 -8.31894517e-01 4.49592322e-01 -5.88145018e-01
-1.49857724e+00 2.62107790e-01 1.32430280e-02 -5.51341653e-01
-5.45566976e-01 -5.15099406e-01 -7.20108330e-01 7.71914601e-01
-1.15368032e+00 -6.43915236e-01 2.35996693e-01 3.13514918e-01
9.02631819e-01 1.46699157e-02 9.85627472e-01 1.19138405e-01
-5.89905381e-01 8.86087477e-01 -5.00811078e-02 -5.02828397e-02
9.63394701e-01 -1.28350151e+00 4.64014024e-01 1.06773078e+00
3.33557218e-01 1.02624726e+00 1.09592640e+00 -5.36583722e-01
-1.07823265e+00 -9.50037420e-01 1.41053355e+00 -5.71032226e-01
8.88990819e-01 -4.72999334e-01 -1.05575037e+00 8.16723108e-01
7.51009583e-02 -2.49647826e-01 7.20242560e-01 8.23621929e-01
-5.82243264e-01 1.39168859e-01 -8.41996253e-01 7.57127702e-01
1.25471330e+00 -8.82062256e-01 -1.22115612e+00 4.27843213e-01
8.96359682e-01 -1.59566730e-01 -5.12757301e-01 1.59123063e-01
2.39435703e-01 -8.18093121e-01 7.54344165e-01 -1.40048873e+00
7.14071751e-01 3.38662386e-01 -9.63650793e-02 -1.47801018e+00
-3.78254414e-01 -4.55408186e-01 -3.05511177e-01 1.34748280e+00
5.14294863e-01 -2.33164832e-01 5.40897608e-01 6.79232597e-01
-2.96302289e-01 -7.09444165e-01 -9.56582189e-01 -8.90037715e-01
3.53855401e-01 -5.25448799e-01 2.33441368e-01 9.51443136e-01
3.87856632e-01 9.99032795e-01 -9.92899537e-02 -2.81335980e-01
3.50248426e-01 -8.62805471e-02 3.51093918e-01 -1.06510723e+00
-5.05391121e-01 -6.33531451e-01 -1.66847315e-02 -1.02310193e+00
4.92601514e-01 -1.01498306e+00 4.56698895e-01 -1.14418149e+00
4.08906728e-01 -3.57343554e-01 -4.30274457e-01 6.11170411e-01
-6.33277535e-01 -2.54273601e-02 1.46672711e-01 -1.21702254e-01
-6.22756898e-01 2.61334330e-01 8.42024684e-01 -2.20235705e-01
-2.82655563e-02 -1.85620740e-01 -9.69918072e-01 6.04301155e-01
7.12751448e-01 -5.90579271e-01 -5.97992063e-01 -6.01083994e-01
4.79836524e-01 -1.46051288e-01 3.04668784e-01 -5.93668282e-01
2.83346772e-01 -1.29964933e-01 -1.35199100e-01 -1.61623180e-01
2.62740135e-01 -4.13797081e-01 -4.84003842e-01 4.24252927e-01
-9.97736096e-01 -5.67287430e-02 4.28901672e-01 6.48577988e-01
9.58806649e-02 -4.11750019e-01 7.74779975e-01 -2.45672539e-01
-7.91463733e-01 -1.19564973e-01 -5.41714132e-01 6.15347087e-01
6.17033184e-01 7.86215290e-02 -4.81018037e-01 -7.28447214e-02
-6.92131042e-01 1.64267927e-01 1.82184666e-01 5.93348205e-01
3.49081844e-01 -9.17667806e-01 -6.95105910e-01 2.06578493e-01
2.08812669e-01 -3.44441622e-01 2.14640513e-01 9.43385243e-01
1.87317982e-01 5.44001579e-01 1.27253339e-01 -3.89190435e-01
-1.32785130e+00 5.45868337e-01 3.08296055e-01 -6.23367310e-01
-5.63934743e-01 1.29238248e+00 8.15267861e-02 -2.52621084e-01
3.16514999e-01 -4.96145099e-01 1.94916576e-01 1.58971131e-01
4.51053739e-01 1.27135575e-01 5.24178982e-01 -1.19235173e-01
-4.01158869e-01 -5.80767915e-02 -6.23523057e-01 -7.16214925e-02
1.09186292e+00 -7.68195242e-02 4.22728583e-02 7.47754633e-01
1.19559646e+00 -1.67978108e-01 -1.05039728e+00 -4.37489450e-01
3.45628947e-01 -1.73791289e-01 2.65287161e-01 -1.05170274e+00
-3.57797205e-01 9.24983501e-01 5.69201447e-02 2.00373590e-01
6.92905962e-01 3.77157098e-03 8.14489961e-01 8.19683135e-01
2.54074275e-01 -1.23991382e+00 -1.03959374e-01 1.02802718e+00
5.18800795e-01 -1.23531139e+00 -1.88271970e-01 -1.64680436e-01
-5.12305200e-01 9.99901116e-01 6.78427577e-01 -1.35799110e-01
4.22901154e-01 4.12951499e-01 -1.54766500e-01 -5.06747775e-02
-1.50746393e+00 -8.24523494e-02 1.24727249e-01 1.41817361e-01
5.86818516e-01 8.93355161e-02 -4.75358754e-01 9.00636017e-01
-2.97901660e-01 -1.44516185e-01 3.90506119e-01 8.06555152e-01
-6.05382979e-01 -9.35946405e-01 1.13295332e-01 7.89358318e-01
-3.28324616e-01 -7.42834985e-01 -4.75609601e-01 5.03279805e-01
-2.59062611e-02 9.92913067e-01 3.86945993e-01 -5.03236055e-01
1.01477876e-01 5.34778118e-01 6.19987011e-01 -1.06211269e+00
-9.10996437e-01 -4.11797225e-01 4.39952612e-01 -5.82354009e-01
4.70427386e-02 -5.07037520e-01 -1.13434708e+00 -4.27364379e-01
-1.91039577e-01 2.12773532e-01 9.35835764e-02 1.53579366e+00
4.18057889e-01 6.23162925e-01 2.24676788e-01 -4.59215462e-01
-1.34972394e+00 -1.13862753e+00 -3.49239200e-01 5.88650763e-01
3.79114866e-01 -6.89552784e-01 -6.10247254e-01 -7.28163421e-02] | [10.809849739074707, 8.48558521270752] |
e6b9f268-7951-44a4-a8ea-ad26aa06d1a7 | learning-representations-for-time-series | null | null | http://papers.nips.cc/paper/8634-learning-representations-for-time-series-clustering | http://papers.nips.cc/paper/8634-learning-representations-for-time-series-clustering.pdf | Learning Representations for Time Series Clustering | Time series clustering is an essential unsupervised technique in cases when category information is not available. It has been widely applied to genome data, anomaly detection, and in general, in any domain where pattern detection is important. Although feature-based time series clustering methods are robust to noise and outliers, and can reduce the dimensionality of the data, they typically rely on domain knowledge to manually construct high-quality features. Sequence to sequence (seq2seq) models can learn representations from sequence data in an unsupervised manner by designing appropriate learning objectives, such as reconstruction and context prediction. When applying seq2seq to time series clustering, obtaining a representation that effectively represents the temporal dynamics of the sequence, multi-scale features, and good clustering properties remains a challenge. How to best improve the ability of the encoder is still an open question. Here we propose a novel unsupervised temporal representation learning model, named Deep Temporal Clustering Representation (DTCR), which integrates the temporal reconstruction and K-means objective into the seq2seq model. This approach leads to improved cluster structures and thus obtains cluster-specific temporal representations. Also, to enhance the ability of encoder, we propose a fake-sample generation strategy and auxiliary classification task. Experiments conducted on extensive time series datasets show that DTCR is state-of-the-art compared to existing methods. The visualization analysis not only shows the effectiveness of cluster-specific representation but also shows the learning process is robust, even if K-means makes mistakes. | ['Gary W. Cottrell', 'Sen Li', 'Qianli Ma', 'Jiawei Zheng'] | 2019-12-01 | null | null | null | neurips-2019-12 | ['time-series-clustering'] | ['time-series'] | [ 3.06564808e-01 -6.14693820e-01 -7.36672133e-02 -3.33360881e-01
-5.76717675e-01 -5.11710525e-01 4.29104835e-01 1.90419033e-01
-2.90782452e-01 3.26533705e-01 9.05170590e-02 5.01049508e-04
-5.22835195e-01 -5.62167287e-01 -5.48433542e-01 -1.25857770e+00
-3.84543657e-01 2.83781886e-01 1.71779305e-01 -4.89619374e-02
3.68440419e-01 2.98473448e-01 -1.67323530e+00 2.78477430e-01
1.14215195e+00 8.33676100e-01 4.79552507e-01 5.28967977e-01
-3.56372535e-01 6.68596923e-01 -6.82020783e-01 1.92612514e-01
8.55843276e-02 -7.57686079e-01 -4.41795439e-01 8.78457427e-02
-5.06680787e-01 3.65073234e-01 -1.89132676e-01 8.84744585e-01
4.92803276e-01 4.28039342e-01 5.94170988e-01 -1.16573453e+00
-5.26614547e-01 3.91076118e-01 -5.72015524e-01 1.89870834e-01
9.67227072e-02 3.34130943e-01 5.85351884e-01 -3.51181120e-01
6.05737090e-01 9.20270145e-01 6.91887438e-01 3.82670790e-01
-1.28366137e+00 -6.14540696e-01 1.35433465e-01 5.17708361e-01
-1.52208352e+00 -5.15439473e-02 8.93616557e-01 -6.77871108e-01
7.72879958e-01 3.00332040e-01 6.17737651e-01 1.09607410e+00
7.71501809e-02 6.07682884e-01 8.52707088e-01 -2.08352834e-01
4.40149724e-01 -4.30776745e-01 3.98344398e-02 2.98101395e-01
-3.74640524e-01 1.06143989e-01 -2.82947630e-01 -4.75405380e-02
5.60272813e-01 5.50991654e-01 -3.53781253e-01 -2.69713402e-01
-1.50135183e+00 7.74801910e-01 4.04730380e-01 7.23901570e-01
-5.49307466e-01 5.83829433e-02 7.64458239e-01 3.01278502e-01
4.49743897e-01 2.81318486e-01 -4.25603956e-01 -4.80138004e-01
-9.69761610e-01 -1.25492901e-01 2.89065838e-01 7.76519001e-01
5.36979377e-01 2.18926385e-01 -1.45747185e-01 8.17542732e-01
-2.74864405e-01 3.08910459e-01 1.12903261e+00 -7.71286011e-01
1.14455096e-01 7.69823551e-01 -2.03142151e-01 -1.08275235e+00
-5.60114682e-01 -3.82077247e-01 -1.30466974e+00 -3.17043900e-01
1.44977346e-01 2.52322741e-02 -9.01190042e-01 1.89649713e+00
3.16173017e-01 7.57947445e-01 9.58871171e-02 8.37698936e-01
2.75823951e-01 1.05670679e+00 -1.28133446e-01 -7.67124057e-01
9.04429853e-01 -5.27606189e-01 -9.46735501e-01 3.14534843e-01
7.51914263e-01 -5.65964401e-01 8.86577070e-01 4.01281774e-01
-4.07572895e-01 -7.23584831e-01 -7.12549806e-01 3.51011276e-01
-3.64260316e-01 1.32862762e-01 4.98349845e-01 2.72617042e-01
-6.93191528e-01 8.39256465e-01 -1.05793428e+00 -5.77836335e-01
1.48328811e-01 4.39026922e-01 -3.55465353e-01 1.50883328e-02
-1.03106570e+00 4.30303901e-01 7.98500657e-01 1.29629254e-01
-6.46133423e-01 -6.09322488e-01 -6.14403248e-01 -8.55409168e-03
3.01150918e-01 -1.49887949e-01 6.75532699e-01 -1.12907207e+00
-1.50057054e+00 2.31332153e-01 -3.25329781e-01 -4.48033512e-01
1.90661013e-01 6.57700226e-02 -7.06386864e-01 3.39243203e-01
1.72385406e-02 3.86998296e-01 9.16963100e-01 -8.11181843e-01
-4.03251231e-01 -3.49168837e-01 -6.94518507e-01 -9.25985798e-02
-4.40017283e-01 -1.19955607e-01 -5.07378042e-01 -9.80021000e-01
1.51217103e-01 -9.20566261e-01 -4.71753508e-01 -3.54495704e-01
-2.49160141e-01 -3.68134528e-01 1.09400213e+00 -7.84160793e-01
1.46116376e+00 -2.56473780e+00 3.56287152e-01 3.12296391e-01
-1.67913213e-01 3.32140416e-01 -1.84054807e-01 6.51392400e-01
-4.11588728e-01 3.43897194e-02 -6.43773675e-01 2.19044164e-01
-2.20497906e-01 2.74381220e-01 -1.76933184e-01 5.49812138e-01
4.34658349e-01 7.08960235e-01 -1.05608654e+00 -3.80632371e-01
2.95454592e-01 4.91795093e-01 -5.55169106e-01 2.60040730e-01
-3.42840403e-01 8.59410644e-01 -2.95413107e-01 2.50332236e-01
4.33304757e-01 -3.06740791e-01 4.25769329e-01 -1.71426535e-01
-2.49572933e-01 -1.25259198e-02 -1.05316353e+00 1.91487241e+00
-1.60888910e-01 5.07957637e-01 -4.28036243e-01 -1.60604048e+00
1.18746209e+00 4.38799173e-01 9.82332170e-01 -7.81762600e-01
-1.02653243e-01 1.69143349e-01 2.36960128e-01 -7.83010900e-01
2.03078598e-01 -3.72411162e-02 -3.17642502e-02 3.56105030e-01
-6.86376393e-02 3.87336552e-01 1.99214906e-01 3.75816859e-02
1.22018254e+00 2.72373855e-01 1.81436285e-01 -1.28857926e-01
4.67277378e-01 9.01049823e-02 1.00693226e+00 2.08794206e-01
3.14778415e-03 5.40063381e-01 5.87415516e-01 -3.07338774e-01
-1.15865266e+00 -6.62750125e-01 1.06891878e-01 8.61195028e-01
-8.03950280e-02 -5.36195278e-01 -5.91000557e-01 -5.44309556e-01
-2.41066009e-01 5.46138048e-01 -6.39909983e-01 -5.02345204e-01
-7.98339546e-01 -9.43020701e-01 4.43395138e-01 5.32916307e-01
2.61069924e-01 -9.12459314e-01 -4.76688892e-01 4.97451693e-01
-4.77854043e-01 -7.61782348e-01 -6.36654794e-01 3.96326959e-01
-1.06051993e+00 -9.96133029e-01 -6.35008812e-01 -7.23020732e-01
6.47197545e-01 2.08606228e-01 4.83460724e-01 -9.23706219e-03
-3.78957748e-01 6.03531748e-02 -7.98384309e-01 -6.29625935e-03
-2.30712369e-01 -1.35421008e-02 1.96838319e-01 3.14755529e-01
5.71196496e-01 -8.05764794e-01 -5.63645542e-01 3.98544103e-01
-1.22234631e+00 -1.74713656e-01 4.63382214e-01 1.11695302e+00
6.46036923e-01 6.08765364e-01 9.85916317e-01 -7.14534640e-01
3.74968827e-01 -6.27930045e-01 -6.52631402e-01 9.82460603e-02
-4.35938746e-01 1.19731069e-01 1.24665010e+00 -7.00070202e-01
-7.67090380e-01 3.54903519e-01 3.41096427e-04 -9.47851121e-01
-7.39037246e-02 6.86666965e-01 -7.63825625e-02 4.66346383e-01
5.27866244e-01 8.69128048e-01 4.13557142e-01 -5.67145050e-01
2.67422587e-01 7.09988952e-01 5.20207524e-01 -4.07995522e-01
6.32401526e-01 4.11014378e-01 -1.07985973e-01 -9.07229543e-01
-2.99088955e-01 -8.08212936e-01 -9.68472481e-01 -6.66911453e-02
9.12148476e-01 -7.51796126e-01 -8.87893498e-01 2.56189913e-01
-9.07339334e-01 -1.92241684e-01 -1.89396814e-01 5.21607816e-01
-6.89533412e-01 5.90369642e-01 -3.81478280e-01 -7.66293526e-01
-1.12409092e-01 -1.06021821e+00 8.48921001e-01 -1.13486446e-01
-1.06981538e-01 -8.82108331e-01 2.13806685e-02 -8.89535695e-02
2.43774652e-01 5.19413114e-01 1.10978615e+00 -7.11759984e-01
-4.82711703e-01 -2.24608108e-02 2.63192803e-01 2.51258016e-01
3.85728925e-01 2.46500939e-01 -7.21235871e-01 -3.81001294e-01
7.07099661e-02 2.94480138e-02 8.34060311e-01 4.10730392e-01
1.68766475e+00 -4.10995185e-01 -5.09912312e-01 5.54327190e-01
1.36625981e+00 7.15948343e-01 6.85594916e-01 4.27551344e-02
6.71690822e-01 8.10453653e-01 8.97275031e-01 7.01937914e-01
8.33325535e-02 7.95775056e-01 1.51888251e-01 2.03480512e-01
2.99357593e-01 -1.30942777e-01 5.02018452e-01 1.50127292e+00
-4.32417504e-02 -2.12264899e-02 -8.21044326e-01 6.42852962e-01
-2.21297884e+00 -1.40592074e+00 -9.97337624e-02 2.17893600e+00
8.11668515e-01 -2.44273394e-01 3.54739636e-01 5.61282396e-01
8.42426658e-01 -5.59298657e-02 -7.78222978e-01 -1.14461586e-01
-4.50616218e-02 1.10541880e-02 1.05474561e-01 -1.33586630e-01
-1.11457658e+00 7.91978121e-01 5.37053537e+00 1.07208002e+00
-1.43734992e+00 -4.63266261e-02 3.43247414e-01 -8.41129944e-03
-6.28187880e-02 -1.30876854e-01 -1.60320997e-01 1.01471543e+00
1.19142485e+00 -1.91737607e-01 5.39795101e-01 5.02321839e-01
5.92339754e-01 3.67696106e-01 -1.14808464e+00 1.11942887e+00
-1.94241449e-01 -1.24189413e+00 -1.29790545e-01 -2.47536246e-02
5.71744740e-01 -2.68293470e-01 -8.44354630e-02 2.94493586e-01
8.58211368e-02 -1.01351881e+00 3.50120604e-01 6.62399888e-01
6.69969976e-01 -1.06104815e+00 7.28365064e-01 5.94601929e-01
-1.38373697e+00 -3.00014973e-01 -3.69848907e-01 1.61202744e-01
1.12698965e-01 7.64462054e-01 -8.28370810e-01 9.30769324e-01
7.20038772e-01 1.29931045e+00 -4.58069772e-01 1.09263074e+00
1.45585135e-01 7.78487861e-01 -2.19749376e-01 -5.45932613e-02
2.42703110e-01 -5.02944171e-01 3.26530993e-01 1.24071956e+00
6.15844846e-01 2.17593163e-01 2.98786670e-01 5.96919060e-01
3.14012825e-01 1.07088201e-01 -6.43711269e-01 -4.38009202e-01
6.18242741e-01 9.92806494e-01 -8.60585213e-01 -2.25348607e-01
-1.60915047e-01 1.08130682e+00 2.66720861e-01 3.43171030e-01
-8.34854126e-01 -5.86359739e-01 6.12376213e-01 -1.67506918e-01
6.35102868e-01 -3.50815296e-01 1.41275123e-01 -1.04042733e+00
-1.04741554e-03 -1.09607387e+00 5.43273568e-01 -5.09728193e-01
-1.23360455e+00 4.63155806e-01 -2.64007777e-01 -1.83828163e+00
-5.17423868e-01 -3.38329017e-01 -5.39833665e-01 4.37393844e-01
-1.08998251e+00 -8.47762108e-01 -1.23950213e-01 7.22120225e-01
6.21975482e-01 -1.24615371e-01 7.55225718e-01 4.95374739e-01
-8.25410545e-01 4.23525065e-01 7.12410092e-01 9.75499898e-02
6.30377769e-01 -1.12048614e+00 -5.84408902e-02 9.22228813e-01
-1.05441980e-01 8.03986251e-01 4.94280994e-01 -6.70128703e-01
-1.66706121e+00 -1.57232440e+00 4.49606031e-01 -2.24678051e-02
4.88148004e-01 -4.00239617e-01 -1.27283895e+00 3.76387686e-01
-8.77207965e-02 -1.10952839e-01 9.62125838e-01 -1.17624126e-01
-2.04477146e-01 -3.33388984e-01 -9.11764801e-01 4.62307334e-01
9.91685212e-01 -5.28745294e-01 -3.84083450e-01 3.88230532e-01
9.04963911e-01 5.44486828e-02 -1.08672309e+00 3.47801924e-01
3.30821723e-01 -8.09784889e-01 7.39324212e-01 -5.13385057e-01
3.58336985e-01 -8.04460466e-01 -1.45323560e-01 -1.51093256e+00
-5.70683062e-01 -7.78995812e-01 -7.45966658e-02 1.41993296e+00
-2.73863021e-02 -4.54575926e-01 4.65269566e-01 -2.40876764e-01
-2.99526125e-01 -6.02421403e-01 -9.03197289e-01 -1.20148134e+00
-1.64714664e-01 -3.51535767e-01 8.53660882e-01 1.44116855e+00
2.46601209e-01 1.86776489e-01 -3.96829605e-01 1.64963633e-01
4.53633279e-01 3.74480367e-01 5.89376271e-01 -1.20773005e+00
-2.11067200e-01 -5.33099234e-01 -6.44169509e-01 -8.10759246e-01
2.01134145e-01 -9.84555960e-01 1.87874943e-01 -1.15280211e+00
-5.50681576e-02 -4.05677289e-01 -4.79460239e-01 3.23784292e-01
-2.69208640e-01 -1.52135327e-01 -3.11472900e-02 4.26850975e-01
-5.56560099e-01 7.69676387e-01 1.04459321e+00 -4.26943637e-02
-3.01548660e-01 -1.62218422e-01 -3.90703768e-01 2.92592019e-01
8.36242974e-01 -3.56316745e-01 -5.80961764e-01 -1.20871246e-01
-2.08429307e-01 2.14368716e-01 8.27897713e-02 -9.89906132e-01
4.14621800e-01 -3.10621619e-01 3.95940840e-01 -8.17272186e-01
2.21478894e-01 -9.33998704e-01 5.14235198e-01 6.92393720e-01
-3.08893919e-01 2.70452708e-01 1.05485097e-02 9.46151435e-01
-5.24438620e-01 1.05490759e-01 8.39248121e-01 -7.20706657e-02
-8.32940698e-01 3.48850489e-01 -4.19507116e-01 -2.02177092e-01
1.21850228e+00 -1.94011644e-01 -6.97450191e-02 -2.10914835e-01
-5.70028663e-01 3.59792560e-01 4.22474444e-01 4.64528471e-01
6.60693765e-01 -1.54878104e+00 -5.35304964e-01 2.16570362e-01
2.75702566e-01 -6.44360408e-02 4.18156862e-01 1.13829911e+00
-2.46968120e-01 3.73179108e-01 -4.26914170e-02 -1.02268422e+00
-9.96408403e-01 1.19172859e+00 -6.79044938e-03 2.98256567e-03
-8.27416182e-01 3.29237431e-01 3.74186374e-02 -5.35744011e-01
1.91964537e-01 -2.75696009e-01 -4.25233752e-01 2.12587327e-01
5.47767460e-01 3.83292437e-01 -3.13164666e-02 -5.47821641e-01
-4.38330352e-01 7.91502178e-01 9.66005549e-02 2.38719136e-01
1.43654740e+00 -3.03162374e-02 -2.50210196e-01 7.12995172e-01
1.39966083e+00 -5.16527653e-01 -1.13805425e+00 -1.39522627e-01
4.11184222e-01 -4.00759935e-01 -3.26400161e-01 -3.59598219e-01
-1.08132792e+00 1.09041655e+00 7.26407051e-01 2.54728377e-01
1.43518424e+00 -3.62131745e-01 7.81822801e-01 2.37525061e-01
1.94564328e-01 -1.09149182e+00 4.41875726e-01 4.61404800e-01
5.24373055e-01 -9.84540880e-01 -4.22119886e-01 -2.50110209e-01
-6.94080651e-01 1.08917427e+00 4.78861928e-01 6.14803471e-03
5.17381072e-01 6.76480010e-02 -9.28796306e-02 -4.11306545e-02
-9.50601339e-01 -2.72671819e-01 6.85283542e-02 7.52865016e-01
5.13687432e-01 4.96530533e-02 -3.11188251e-01 5.79978585e-01
-7.33388960e-02 1.31987559e-03 2.37972051e-01 8.88218462e-01
-3.14208210e-01 -1.08346498e+00 -3.19990158e-01 4.94852334e-01
-2.44652256e-01 1.75792545e-01 -1.23054236e-01 3.96844387e-01
1.85994610e-01 8.85770619e-01 1.98892027e-01 -6.15942359e-01
3.66788432e-02 2.46612504e-01 7.45900869e-02 -4.27293569e-01
-5.03773272e-01 4.96271968e-01 -3.24806541e-01 -5.55702090e-01
-5.68812847e-01 -7.76980758e-01 -1.54141498e+00 -1.48501053e-01
-2.87404895e-01 3.96636099e-01 4.58194643e-01 8.49288523e-01
7.51615047e-01 8.16510677e-01 1.17980981e+00 -5.77927589e-01
-2.82220781e-01 -9.61156249e-01 -6.00997746e-01 6.96759701e-01
3.74692827e-01 -5.12248993e-01 -1.65714860e-01 3.78490597e-01] | [7.3092451095581055, 3.3541672229766846] |
9aa080f9-c483-472c-8dc6-a8cbdfda83eb | 3d-object-detection-with-a-self-supervised | 2205.00705 | null | https://arxiv.org/abs/2205.00705v2 | https://arxiv.org/pdf/2205.00705v2.pdf | 3D Object Detection with a Self-supervised Lidar Scene Flow Backbone | State-of-the-art lidar-based 3D object detection methods rely on supervised learning and large labeled datasets. However, annotating lidar data is resource-consuming, and depending only on supervised learning limits the applicability of trained models. Self-supervised training strategies can alleviate these issues by learning a general point cloud backbone model for downstream 3D vision tasks. Against this backdrop, we show the relationship between self-supervised multi-frame flow representations and single-frame 3D detection hypotheses. Our main contribution leverages learned flow and motion representations and combines a self-supervised backbone with a supervised 3D detection head. First, a self-supervised scene flow estimation model is trained with cycle consistency. Then, the point cloud encoder of this model is used as the backbone of a single-frame 3D object detection head model. This second 3D object detection model learns to utilize motion representations to distinguish dynamic objects exhibiting different movement patterns. Experiments on KITTI and nuScenes benchmarks show that the proposed self-supervised pre-training increases 3D detection performance significantly. https://github.com/emecercelik/ssl-3d-detection.git | ['Emeç Erçelik', 'Alois Knoll', 'Yılmaz Kaan Çaylı', 'Maximilian Listl', 'Pınar Topçam', 'Hanzhen Zhang', 'Zhijie Yang', 'MingYu Liu', 'Ekim Yurtsever'] | 2022-05-02 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [-5.77559359e-02 -2.08380729e-01 -6.44855559e-01 -3.28466952e-01
-6.56161845e-01 -7.41148651e-01 5.90975583e-01 -2.45849583e-02
-6.95748851e-02 1.15265183e-01 -2.10717186e-01 -4.50248331e-01
3.60279202e-01 -7.61609674e-01 -7.86184251e-01 -3.87312204e-01
-9.81621221e-02 5.49389839e-01 8.68920147e-01 1.64332062e-01
3.69219065e-01 9.54211652e-01 -1.70826411e+00 6.48066327e-02
6.25024557e-01 8.82240534e-01 3.82069379e-01 9.79190826e-01
-4.39599544e-01 1.02120531e+00 -1.76885903e-01 2.37014994e-01
6.83469832e-01 -1.00269288e-01 -6.16028309e-01 3.31801623e-01
1.02961767e+00 -6.96877241e-01 -5.77881634e-01 6.18189633e-01
3.26248169e-01 1.04539357e-01 8.83964717e-01 -1.52459240e+00
-1.13219142e-01 -8.15385431e-02 -7.54168749e-01 6.70463443e-01
3.89076382e-01 5.84295988e-01 1.02335548e+00 -1.21551383e+00
7.69328594e-01 1.48672807e+00 6.61074698e-01 5.42108417e-01
-1.13853467e+00 -7.83825576e-01 2.37514943e-01 9.06028450e-02
-1.15365767e+00 -4.66363817e-01 9.91482675e-01 -7.97259271e-01
1.10025632e+00 -3.14074934e-01 7.93048918e-01 9.75097537e-01
-8.17513913e-02 1.09589398e+00 7.52913237e-01 -1.75281346e-01
8.35308135e-02 -1.64179638e-01 2.12098867e-01 1.10518205e+00
5.45362294e-01 7.32890546e-01 -8.18530858e-01 1.85510114e-01
8.96593332e-01 1.55211578e-03 6.59263656e-02 -9.17986751e-01
-1.00403988e+00 7.35364318e-01 5.74312210e-01 -2.23844394e-01
-5.56401052e-02 3.78581405e-01 4.73501027e-01 -1.00627042e-01
6.39641106e-01 -2.22350240e-01 -3.76738697e-01 -4.60410118e-03
-1.03233707e+00 1.53201118e-01 5.78907847e-01 1.17560160e+00
1.17026484e+00 1.59471273e-01 1.41716525e-01 2.88629681e-01
6.45022750e-01 8.92907858e-01 2.50151730e-03 -1.20203400e+00
4.29272532e-01 8.14238727e-01 -1.82796288e-02 -7.50434160e-01
-3.28705460e-01 -2.63051301e-01 -3.22713792e-01 6.64869845e-01
3.96618426e-01 2.44504571e-01 -9.95375335e-01 1.20507526e+00
8.44678879e-01 5.70971370e-01 -1.37006521e-01 1.10769081e+00
8.57435644e-01 6.19654477e-01 -2.13862676e-03 9.85972062e-02
7.89666116e-01 -1.03156984e+00 -1.11677870e-02 -4.26326603e-01
7.79222310e-01 -7.54420221e-01 8.17253411e-01 -2.06830263e-01
-1.04485929e+00 -8.65870059e-01 -8.92162621e-01 -3.86800915e-01
-1.73172846e-01 1.59785867e-01 5.80195725e-01 4.74194765e-01
-6.86023891e-01 2.57491171e-01 -1.08595335e+00 -4.90224570e-01
9.57384050e-01 1.43605359e-02 -2.30884403e-01 -1.97710976e-01
-4.20466691e-01 9.01918590e-01 4.07696217e-01 -1.65188417e-01
-1.30326164e+00 -1.01980162e+00 -1.04531002e+00 -2.59463668e-01
3.16006482e-01 -9.40263987e-01 1.26104867e+00 -2.91706800e-01
-1.18193090e+00 1.27977431e+00 -3.87974977e-01 -4.33670640e-01
8.17568064e-01 -3.99719059e-01 1.86199456e-01 4.67517644e-01
5.50817788e-01 1.10438454e+00 9.97523248e-01 -1.35537076e+00
-1.20198941e+00 -2.77530402e-01 -1.77047327e-01 1.86731517e-01
2.38779381e-01 -4.05478537e-01 -4.46799189e-01 -1.33209184e-01
1.51886076e-01 -8.15252900e-01 3.92491976e-03 6.41202331e-01
-4.30592448e-01 -4.44390476e-01 1.47549713e+00 1.71641767e-01
5.25412083e-01 -2.07632899e+00 -1.82930782e-01 -1.31425887e-01
2.33794749e-01 1.89520836e-01 -1.02773070e-01 -7.85899162e-02
2.18935296e-01 -1.23868890e-01 -1.10441349e-01 -6.53776348e-01
-2.37196878e-01 3.57232094e-01 -4.56786156e-01 7.18496740e-01
6.96489692e-01 9.06780183e-01 -1.20968235e+00 -7.94138789e-01
9.94881630e-01 4.02972519e-01 -5.60343206e-01 4.03365701e-01
-3.69885802e-01 5.37230968e-01 -5.05727291e-01 1.25347400e+00
7.15755761e-01 -3.17963511e-01 -4.40630615e-01 -3.06811184e-01
-3.91573191e-01 4.79551971e-01 -1.06617296e+00 1.91058886e+00
-2.26077706e-01 9.18473363e-01 -2.36782789e-01 -9.79812980e-01
1.03389072e+00 -3.66008133e-01 8.18551362e-01 -3.04797351e-01
1.87684208e-01 1.83424890e-01 -4.53450620e-01 -4.13054556e-01
2.62793779e-01 2.40733415e-01 2.23013371e-01 4.55686271e-01
2.91446537e-01 -5.92109382e-01 2.13141263e-01 3.63271385e-01
1.08058918e+00 7.71158457e-01 1.21429615e-01 -6.19071163e-02
5.40788352e-01 4.63606954e-01 4.89869416e-01 7.66914725e-01
-6.94657087e-01 5.52527606e-01 1.24595715e-02 -6.70813024e-01
-1.00033104e+00 -1.26707518e+00 -1.75771907e-01 8.26245725e-01
6.17795587e-01 -3.46938491e-01 -2.79619098e-02 -1.00651491e+00
5.23660541e-01 3.24199051e-01 -4.20391828e-01 -6.69726506e-02
-6.19642913e-01 -6.41629398e-02 4.65992451e-01 6.62706077e-01
4.32412654e-01 -5.42783201e-01 -1.31199944e+00 -3.19636054e-02
6.56125396e-02 -1.59744382e+00 -2.55690366e-01 4.11896259e-01
-1.26634324e+00 -1.36038530e+00 -2.09217161e-01 -6.90522730e-01
5.85842311e-01 1.12594509e+00 1.10602748e+00 -4.29703295e-03
-6.68137848e-01 7.04781651e-01 -3.45093220e-01 -4.91875708e-01
-3.25375378e-01 -1.02366805e-02 2.36129031e-01 -3.50115269e-01
5.78734934e-01 -4.92808759e-01 -7.29627550e-01 3.61870676e-01
-3.52986127e-01 -2.42141783e-02 4.64363098e-01 4.88720357e-01
8.15782964e-01 -5.35601020e-01 -8.69926363e-02 -4.76047277e-01
-4.08484340e-01 -2.57166892e-01 -8.87414098e-01 -2.20154092e-01
-4.31618214e-01 -9.13472380e-03 8.05492140e-03 -4.11169082e-01
-9.90827739e-01 7.33609259e-01 3.14671129e-01 -1.20310271e+00
-4.63602901e-01 -2.94303417e-01 9.62528810e-02 -1.83696806e-01
8.07801247e-01 4.67071533e-02 -3.07025444e-02 -2.43369594e-01
7.84251750e-01 3.31861883e-01 7.47447908e-01 -5.19050479e-01
1.41135633e+00 9.95352030e-01 2.44237304e-01 -9.75355983e-01
-1.11816108e+00 -9.47086334e-01 -1.21036232e+00 -6.45720422e-01
9.22231317e-01 -1.18720412e+00 -4.99319047e-01 2.93984801e-01
-1.31055248e+00 -5.86176097e-01 -6.38411641e-01 5.13183653e-01
-7.61933565e-01 3.57150704e-01 -3.48419040e-01 -1.01498497e+00
-1.48481101e-01 -8.56294215e-01 1.59469593e+00 1.88970536e-01
-2.91940123e-02 -8.05850804e-01 1.65050328e-01 4.11305726e-01
-2.28768513e-01 3.93713236e-01 4.28593367e-01 -1.34300187e-01
-1.19267154e+00 -1.33926198e-01 -3.85697812e-01 2.09664226e-01
1.14191920e-01 2.25888893e-01 -1.17117524e+00 -6.32778406e-02
-3.51408392e-01 -4.32063997e-01 1.08017766e+00 4.55726773e-01
8.75420511e-01 4.11883861e-01 -5.02131701e-01 9.22773838e-01
1.17462361e+00 -1.87786102e-01 1.35939449e-01 2.36959562e-01
9.76882219e-01 4.44927394e-01 8.84962380e-01 3.36110115e-01
5.66787302e-01 3.82813483e-01 7.33141422e-01 -7.63587207e-02
-6.26634002e-01 -5.92705607e-01 4.64354098e-01 2.93496162e-01
-5.00288270e-02 1.72025338e-01 -1.13793647e+00 4.95240569e-01
-1.82212317e+00 -1.06411529e+00 -4.20466691e-01 1.91276789e+00
3.96967947e-01 4.65787143e-01 2.81944484e-01 1.40149549e-01
5.82275271e-01 1.86305419e-01 -9.29139614e-01 3.14735770e-01
-1.15998246e-01 9.07965656e-03 7.15951800e-01 5.90653777e-01
-1.33843458e+00 1.40213633e+00 5.13221169e+00 3.01061124e-01
-1.09157526e+00 1.84105456e-01 1.23689249e-01 -1.99248701e-01
1.16638146e-01 2.99702257e-01 -1.27349114e+00 7.80557021e-02
3.68543774e-01 -9.02780816e-02 -3.25450659e-01 9.66246545e-01
5.66332161e-01 -1.11714184e-01 -1.28084528e+00 1.18601537e+00
-1.08277509e-02 -1.47949648e+00 5.13798110e-02 1.43404365e-01
5.96765280e-01 5.51038802e-01 -3.37875366e-01 1.67032793e-01
2.90022165e-01 -4.92633402e-01 8.81692529e-01 1.60264134e-01
7.14347005e-01 -4.09496367e-01 1.20404445e-01 4.00194913e-01
-1.73159051e+00 -1.14277124e-01 -5.27204275e-01 -1.57939985e-01
4.27306980e-01 5.57289243e-01 -9.21628594e-01 3.61870617e-01
8.55661929e-01 1.43252504e+00 -5.55652261e-01 1.09691525e+00
-4.50258732e-01 4.63248014e-01 -5.79524279e-01 1.76691487e-01
3.22879702e-01 -2.62389556e-02 9.92410064e-01 1.29148877e+00
1.52648389e-01 -3.62098478e-02 6.23513758e-01 9.77655232e-01
9.95285884e-02 -3.05345327e-01 -9.92929995e-01 2.50784516e-01
6.32893741e-01 1.17744720e+00 -8.64810884e-01 -2.38051742e-01
-5.12830496e-01 7.55157828e-01 3.62106383e-01 1.24420293e-01
-7.15132058e-01 1.30758077e-01 8.83183718e-01 3.98856372e-01
5.07335484e-01 -6.83678746e-01 -3.54276001e-01 -1.34188175e+00
-2.62994528e-01 6.78488836e-02 4.38962311e-01 -8.00571024e-01
-1.26274693e+00 4.03042743e-03 9.44404900e-02 -1.64234281e+00
-1.73966274e-01 -6.55589521e-01 -5.99422216e-01 3.53736192e-01
-2.11486292e+00 -1.46550810e+00 -7.47716010e-01 7.36805558e-01
8.68237317e-01 3.17954347e-02 3.57407033e-01 1.27400890e-01
-3.50358099e-01 4.64088395e-02 -6.38346016e-01 2.82164782e-01
5.93009233e-01 -1.17688239e+00 4.26449209e-01 1.10710990e+00
3.68667066e-01 1.89244561e-02 2.10401505e-01 -7.61579096e-01
-1.63806236e+00 -1.46207094e+00 5.85826933e-01 -7.75116324e-01
5.51938415e-01 -4.18959975e-01 -7.62259901e-01 5.94952106e-01
-3.78480524e-01 7.25063026e-01 3.83609325e-01 -3.85456175e-01
-5.46134889e-01 -1.12155020e-01 -8.21779013e-01 2.04573408e-01
1.69982839e+00 -6.44167542e-01 -6.35482490e-01 4.86442089e-01
6.99393272e-01 -6.92312717e-01 -5.42077243e-01 4.42547888e-01
3.88112307e-01 -9.39905345e-01 1.37161183e+00 -5.56659222e-01
4.49882984e-01 -7.14308083e-01 -1.49410263e-01 -6.31255925e-01
-3.75251211e-02 -5.08013427e-01 -7.48292625e-01 7.88817704e-01
-2.82952674e-02 -1.11805379e-01 1.15672946e+00 1.50101975e-01
-3.07552159e-01 -3.72897923e-01 -9.82518137e-01 -9.06542778e-01
-1.07869305e-01 -8.23990285e-01 -1.92547273e-02 7.83930063e-01
-5.15519977e-01 6.20143175e-01 -1.63675528e-02 4.08264130e-01
1.20337808e+00 4.36767906e-01 1.48500443e+00 -1.40966046e+00
1.10530429e-01 -4.98359114e-01 -8.33134592e-01 -1.60366237e+00
3.97711575e-01 -1.11917794e+00 6.33518845e-02 -1.42113066e+00
-2.09220409e-01 -5.61944842e-01 1.74315140e-01 5.21640182e-01
9.65157617e-03 2.54403919e-01 4.52779114e-01 6.39124930e-01
-5.70576549e-01 5.14587045e-01 1.07856834e+00 -1.95083708e-01
-5.35603642e-01 1.08047046e-01 -4.79509309e-03 7.57888854e-01
6.53205991e-01 -4.93429095e-01 -4.66939062e-01 -5.24005532e-01
-2.75788069e-01 -1.63264200e-01 8.86355877e-01 -1.07155550e+00
4.93210226e-01 -1.68748811e-01 5.10597706e-01 -1.29997432e+00
2.91314006e-01 -7.50444472e-01 -5.81345201e-01 7.41297305e-01
9.39075090e-03 -2.53744036e-01 2.56268620e-01 8.53340089e-01
1.64484054e-01 9.76766199e-02 9.77258444e-01 -2.99690038e-01
-1.22078288e+00 6.96786225e-01 -1.39066830e-01 1.08627409e-01
1.19391274e+00 -6.16250157e-01 -2.02615693e-01 -5.55576431e-03
-4.31742907e-01 4.42665130e-01 3.62181127e-01 6.16786659e-01
9.75042343e-01 -1.09631193e+00 -6.63385212e-01 3.35541338e-01
4.08023387e-01 8.06621909e-01 -1.01591386e-01 6.90939248e-01
-6.22303963e-01 2.51195550e-01 -1.24595404e-01 -1.64558232e+00
-9.99555886e-01 2.43992791e-01 4.77812290e-01 3.25988948e-01
-9.52028394e-01 7.50569224e-01 5.14578000e-02 -3.48380119e-01
2.80198336e-01 -5.21555543e-01 1.96454003e-01 3.96607593e-02
2.20695466e-01 5.40436387e-01 -1.25115797e-01 -7.32189357e-01
-5.50982237e-01 1.03295410e+00 1.71275988e-01 8.36551711e-02
1.17105711e+00 -2.52870977e-01 4.72303629e-01 5.57851315e-01
1.22563159e+00 -3.82311672e-01 -1.74691594e+00 -3.77711415e-01
-8.62578154e-02 -8.70401978e-01 2.50319272e-01 -1.05086207e-01
-9.49585557e-01 1.14242697e+00 6.24770403e-01 -2.14983806e-01
6.76158249e-01 3.03020567e-01 5.61770856e-01 4.84076411e-01
4.64721203e-01 -6.72546566e-01 5.05323827e-01 7.69245446e-01
2.74424970e-01 -1.64902759e+00 1.06370933e-01 -5.77328086e-01
-2.32607007e-01 1.32876217e+00 9.49430346e-01 -4.02294636e-01
5.82554996e-01 2.15924472e-01 1.28855869e-01 -2.81238824e-01
-6.13583028e-01 -7.69540668e-01 1.46241724e-01 9.78440344e-01
-1.28792450e-01 -2.70198464e-01 3.69771481e-01 -3.18737179e-01
6.40875846e-02 2.73273736e-02 3.79230052e-01 1.24015236e+00
-6.93198562e-01 -8.76889169e-01 -2.32368901e-01 2.14333907e-01
3.62088650e-01 3.52392614e-01 -3.79893124e-01 9.31502640e-01
1.91810295e-01 8.16668093e-01 2.91088730e-01 -3.02663386e-01
3.19227040e-01 -2.61408109e-02 5.98024368e-01 -7.65190363e-01
1.47508113e-02 3.06828506e-02 -3.27764392e-01 -9.00594234e-01
-8.79258990e-01 -8.18246067e-01 -1.39461458e+00 -1.50707677e-01
-3.74081582e-01 -3.91562313e-01 5.72944760e-01 7.71419406e-01
4.98565227e-01 8.54834691e-02 7.49199212e-01 -1.48184550e+00
-3.31694633e-01 -4.70871866e-01 -7.46546239e-02 3.81181568e-01
6.00133657e-01 -1.02843285e+00 -6.00328624e-01 2.87294328e-01] | [7.803024768829346, -2.5995724201202393] |
8c3f13a1-2712-4ea2-b26c-10a660178dc4 | multichannel-sleep-spindle-detection-using | null | null | https://doi.org/10.1016/j.jneumeth.2017.06.004 | https://www.researchgate.net/publication/317533757_Multichannel_Sleep_Spindle_Detection_using_Sparse_Low-Rank_Optimization | Multichannel sleep spindle detection using sparse low-rank optimization | BACKGROUND:
Automated single-channel spindle detectors, for human sleep EEG, are blind to the presence of spindles in other recorded channels unlike visual annotation by a human expert.
NEW METHOD:
We propose a multichannel spindle detection method that aims to detect global and local spindle activity in human sleep EEG. Using a non-linear signal model, which assumes the input EEG to be the sum of a transient and an oscillatory component, we propose a multichannel transient separation algorithm. Consecutive overlapping blocks of the multichannel oscillatory component are assumed to be low-rank whereas the transient component is assumed to be piecewise constant with a zero baseline. The estimated oscillatory component is used in conjunction with a bandpass filter and the Teager operator for detecting sleep spindles.
RESULTS AND COMPARISON WITH OTHER METHODS:
The proposed method is applied to two publicly available databases and compared with 7 existing single-channel automated detectors. F1 scores for the proposed spindle detection method averaged 0.66 (0.02) and 0.62 (0.06) for the two databases, respectively. For an overnight 6 channel EEG signal, the proposed algorithm takes about 4min to detect sleep spindles simultaneously across all channels with a single setting of corresponding algorithmic parameters.
CONCLUSIONS:
The proposed method attempts to mimic and utilize, for better spindle detection, a particular human expert behavior where the decision to mark a spindle event may be subconsciously influenced by the presence of a spindle in EEG channels other than the central channel visible on a digital screen. | ['Indu Ayappa', 'David M. Rapoport', 'Ricardo S.Osorio', 'Ivan W. Selesnick', 'Ankit Parekha', 'Andrew W. Vargad'] | 2017-08-15 | null | null | null | journal-of-neuroscience-methods-volume-288 | ['spindle-detection'] | ['medical'] | [ 3.06834579e-01 -9.83257964e-02 3.26092720e-01 -9.07562822e-02
-5.26618540e-01 -8.10099006e-01 -9.51487869e-02 1.92294285e-01
-4.45290834e-01 1.07395065e+00 -6.28820360e-02 -1.04848996e-01
-1.75238356e-01 2.27023765e-01 -2.69716382e-01 -7.69280374e-01
-1.68219075e-01 -7.54368082e-02 1.42082423e-01 1.24505349e-01
4.80507642e-01 2.63324887e-01 -1.49259162e+00 1.18375354e-01
1.04909265e+00 1.17779315e+00 3.45977068e-01 7.10263371e-01
5.73789418e-01 2.36525178e-01 -1.09280801e+00 2.69559443e-01
7.84470439e-02 -7.18024194e-01 -2.69154906e-01 1.76609397e-01
7.17429891e-02 1.37433320e-01 -1.57468356e-02 1.27306426e+00
8.42435122e-01 -9.56478268e-02 5.14964402e-01 -1.17246783e+00
1.09856457e-01 2.28302795e-02 -7.69242585e-01 1.02693367e+00
7.31784105e-01 2.50625223e-01 3.67825359e-01 -7.61323512e-01
2.44839758e-01 3.56014848e-01 6.70397997e-01 2.97893509e-02
-1.52890444e+00 -9.63449419e-01 -3.57240468e-01 7.09081352e-01
-1.73304033e+00 -6.20930374e-01 7.24452078e-01 -4.13222462e-01
9.78948593e-01 3.09481740e-01 9.53269243e-01 9.45497632e-01
8.78186643e-01 1.23009838e-01 1.67205262e+00 -1.75987944e-01
4.31257904e-01 2.63715148e-01 4.01576012e-01 3.80032957e-01
3.48580748e-01 -2.75190085e-01 -1.07413578e+00 -1.58561617e-01
5.07637322e-01 -4.86831427e-01 -7.07946658e-01 2.84127295e-01
-1.12083268e+00 2.83450425e-01 -4.41361405e-02 3.94292355e-01
-7.06581593e-01 -4.14964795e-01 3.68366808e-01 3.38703483e-01
5.15556812e-01 6.79026783e-01 -3.34959835e-01 -4.76042688e-01
-1.54319358e+00 -2.43235126e-01 7.63925076e-01 8.48342657e-01
5.45610011e-01 7.37415850e-02 -2.89617807e-01 4.87068415e-01
-6.55296892e-02 5.17324328e-01 9.36549842e-01 -4.68311876e-01
-1.55963202e-03 5.54610550e-01 3.09301496e-01 -8.97297144e-01
-1.01267183e+00 -5.58355987e-01 -6.76052332e-01 9.14410874e-02
3.07597458e-01 -2.43558474e-02 -6.21835768e-01 1.27321017e+00
-1.35936975e-01 4.64258075e-01 -9.49794278e-02 1.22074962e+00
6.32225692e-01 4.88312185e-01 -1.24919705e-01 -8.14248443e-01
1.82575536e+00 -2.95627505e-01 -1.26573026e+00 -2.87160099e-01
-3.54024395e-02 -8.47106516e-01 1.09957957e+00 8.94795239e-01
-1.10845983e+00 -3.87855560e-01 -1.44859540e+00 2.67084658e-01
-1.07410796e-01 7.73965895e-01 1.00517593e-01 6.39882207e-01
-1.20667017e+00 2.28235424e-01 -1.09988713e+00 -5.35731435e-01
1.03341982e-01 6.62443876e-01 -4.22765255e-01 3.97288412e-01
-7.49356389e-01 1.06063771e+00 1.24532342e-01 1.56375602e-01
-6.13841116e-01 -3.28089744e-01 -4.14432883e-01 1.88118145e-01
-3.16480696e-02 -4.59245533e-01 9.28044081e-01 -1.31550479e+00
-1.29762876e+00 8.36686015e-01 -6.60351753e-01 -4.50850785e-01
-3.47623974e-02 2.74794493e-02 -7.80236363e-01 5.34175217e-01
1.43159598e-01 -8.95494148e-02 1.27774191e+00 -7.78788209e-01
-2.72888422e-01 -5.83738208e-01 -4.03197825e-01 2.81416684e-01
-3.60360392e-03 1.43175229e-01 5.14157787e-02 -6.56720757e-01
2.06223056e-01 -5.04575729e-01 4.32990462e-01 -5.09245694e-01
-3.90316486e-01 -9.72063690e-02 5.13979375e-01 -1.02265942e+00
1.50094438e+00 -2.50845718e+00 -1.63532700e-02 2.55535960e-01
4.07878608e-01 -1.90232664e-01 8.56079459e-02 1.85245261e-01
-3.42494607e-01 -3.64806026e-01 -4.59570348e-01 -2.81104714e-01
-2.82478452e-01 -4.15312946e-01 -1.18675735e-02 1.14686322e+00
-1.90099049e-02 2.83808231e-01 -6.86325014e-01 -3.57813425e-02
-1.39060736e-01 2.07662702e-01 2.49949738e-01 3.91050279e-01
6.60132706e-01 3.16380054e-01 5.51062584e-01 6.65960014e-01
7.69717276e-01 4.18949351e-02 -1.07611768e-01 -4.18616176e-01
-4.08662856e-01 4.37941343e-01 -1.27954805e+00 1.61953557e+00
-1.96410175e-02 1.10178232e+00 1.61006257e-01 -7.53876746e-01
8.69872630e-01 6.38966262e-01 -1.46902434e-03 -6.70594215e-01
1.57160863e-01 3.72109741e-01 4.48311180e-01 -6.11999691e-01
1.45004973e-01 -2.62078911e-01 3.46268505e-01 3.67825329e-01
4.80907291e-01 -1.45637214e-01 3.19255531e-01 1.28824040e-01
1.31260204e+00 -2.87477612e-01 7.74871707e-01 -7.86225438e-01
3.12373370e-01 -3.30932677e-01 5.47408760e-01 5.66930354e-01
-3.61005872e-01 5.32520473e-01 7.48824239e-01 6.23793788e-02
-4.67060208e-01 -9.89178836e-01 -2.29457185e-01 5.81274033e-01
3.60569149e-01 -4.04224217e-01 -9.33163106e-01 -8.28283429e-02
-4.48160410e-01 7.22917497e-01 -6.07454062e-01 -3.80243659e-01
1.88964963e-01 -9.46700871e-01 4.58244711e-01 2.85676748e-01
2.48849273e-01 -7.92333663e-01 -1.17805743e+00 1.74511090e-01
-2.31882602e-01 -8.66020262e-01 -6.24591172e-01 7.63848662e-01
-7.61840999e-01 -1.25750124e+00 -3.67097765e-01 -4.66286182e-01
7.16498375e-01 9.82575491e-02 7.29778469e-01 -2.35290483e-01
-6.26266301e-01 5.05244076e-01 -2.08460435e-01 -4.42591012e-01
3.34894419e-01 -2.43118405e-01 3.80646795e-01 3.34607929e-01
6.66525841e-01 -7.90110707e-01 -8.31464231e-01 1.97620496e-01
-5.21591306e-01 -4.58234400e-02 6.95449114e-01 5.73889911e-01
2.36352399e-01 2.90877849e-01 8.00509274e-01 -1.76204950e-01
9.35287714e-01 -5.72464526e-01 -5.43214321e-01 -1.52606517e-01
-5.82715333e-01 -3.66196573e-01 5.18537402e-01 -5.68703353e-01
-8.21342766e-01 2.03513891e-01 5.31768203e-01 -2.45203704e-01
-4.19982314e-01 3.24646622e-01 -4.42771725e-02 -1.48735732e-01
7.92931557e-01 4.85488653e-01 -1.79715112e-01 -1.53240785e-01
-3.65303814e-01 7.51251400e-01 7.62699485e-01 1.91647708e-01
3.24337900e-01 2.10452795e-01 -2.51367241e-01 -1.02710915e+00
-2.73132026e-01 -9.04259920e-01 -5.96276581e-01 -3.41213077e-01
8.21434021e-01 -1.12214160e+00 -7.12014675e-01 6.26334012e-01
-1.16488028e+00 -1.16197646e-01 2.62906641e-01 8.07853639e-01
-4.77041304e-01 2.08938152e-01 -3.60214710e-01 -1.09786546e+00
-5.87410569e-01 -8.79169524e-01 8.29315066e-01 4.86391872e-01
-6.43142462e-01 -5.74306846e-01 9.01166946e-02 5.31594306e-02
2.80468613e-01 -1.69921443e-01 4.55707818e-01 -6.51682377e-01
1.08595848e-01 -2.09089279e-01 7.43833631e-02 1.34324074e-01
3.72629404e-01 -2.71498740e-01 -1.23893476e+00 -2.79809445e-01
7.18720973e-01 -5.77375889e-02 3.57346624e-01 5.33199728e-01
7.02599168e-01 -5.13490438e-02 -8.59776065e-02 4.57901567e-01
1.38235915e+00 5.82062602e-01 7.15340972e-01 5.87676205e-02
-4.93103713e-02 2.90528446e-01 2.40001410e-01 4.66248155e-01
1.10843115e-01 2.66320169e-01 1.74656093e-01 -1.85060263e-01
1.11711107e-01 4.08103377e-01 6.27466559e-01 8.19369197e-01
-1.35270178e-01 6.40730839e-03 -7.81136870e-01 5.67496598e-01
-1.40441525e+00 -7.27095246e-01 -6.05398774e-01 2.52649903e+00
7.23456919e-01 2.08730251e-01 2.51572222e-01 4.74220395e-01
6.89918935e-01 -4.99637753e-01 -6.18689597e-01 -2.71105170e-01
-4.02655065e-01 7.12387204e-01 5.20580649e-01 1.71899587e-01
-6.87722385e-01 3.61550003e-01 6.41888952e+00 6.55829191e-01
-1.28901541e+00 3.03784192e-01 1.87839091e-01 -5.19284010e-01
4.98906702e-01 -2.96054989e-01 -3.25189322e-01 9.30651486e-01
1.25068307e+00 -4.60343778e-01 7.59932697e-01 1.94682389e-01
8.48165810e-01 -1.09896541e+00 -9.71823692e-01 1.59040439e+00
4.29121524e-01 -7.86589563e-01 -6.70833170e-01 -3.20223451e-01
4.21409905e-01 -3.67446095e-01 7.19152810e-03 -1.33190319e-01
-7.79657841e-01 -1.27012813e+00 8.67476940e-01 8.72047126e-01
9.36851203e-01 -7.03956604e-01 1.01958323e+00 2.35580415e-01
-1.21563458e+00 -9.68442038e-02 -5.35598993e-02 -4.38131303e-01
-2.99858660e-01 6.56467438e-01 -7.59835839e-01 3.46234053e-01
7.50045419e-01 6.52294517e-01 -8.96654665e-01 1.61758494e+00
-3.10829550e-01 1.13717246e+00 -3.18753242e-01 1.89997688e-01
-2.87679583e-01 -9.10865813e-02 7.80777752e-01 1.20929885e+00
6.06558204e-01 2.45200127e-01 -4.77353096e-01 1.01809144e+00
3.72215241e-01 -4.55815047e-02 -2.94076383e-01 7.85867274e-02
5.57365417e-01 1.55702722e+00 -1.18276298e+00 -2.09769495e-02
-4.23894763e-01 1.27261984e+00 -1.75356120e-01 5.73517978e-01
-6.14779055e-01 -8.59009922e-01 8.15489516e-02 -5.56343868e-02
-2.56723315e-01 4.61150929e-02 -8.71961057e-01 -9.99467075e-01
2.38604754e-01 -9.21736002e-01 2.18633771e-01 -1.22440779e+00
-9.76645887e-01 6.74862802e-01 -2.81222820e-01 -1.31146419e+00
1.46645769e-01 -4.53893363e-01 -9.12674785e-01 1.12372673e+00
-9.17615235e-01 -4.97229844e-01 -4.09521222e-01 9.04110730e-01
6.49838686e-01 -6.81974590e-02 7.23091722e-01 1.23565361e-01
-4.80841666e-01 1.96014598e-01 -3.09201270e-01 -5.35664916e-01
1.07865846e+00 -1.43986428e+00 -3.88902068e-01 1.03225255e+00
-7.72564486e-02 6.03023410e-01 9.45488155e-01 -5.83333969e-01
-1.12108469e+00 -4.84880298e-01 7.91503727e-01 -1.15403682e-01
6.36585593e-01 -5.56778014e-01 -8.96956861e-01 2.81361908e-01
5.84147394e-01 -3.96519035e-01 7.31264174e-01 -1.93446666e-01
3.77473950e-01 -3.61116379e-01 -1.07102168e+00 1.62134075e-03
3.13319027e-01 -7.41514027e-01 -9.77983952e-01 8.78559202e-02
-2.28135243e-01 -2.63051361e-01 -4.77121294e-01 -1.53984660e-02
4.73848343e-01 -1.06422067e+00 4.75480258e-02 1.98236555e-01
-2.44407952e-01 -6.56829834e-01 3.58971685e-01 -1.64550364e+00
-2.29390740e-01 -1.06210363e+00 7.17572402e-03 6.96018755e-01
1.85944572e-01 -6.55832827e-01 7.94918835e-02 1.86691359e-01
-3.97201180e-01 -4.87097770e-01 -1.19428921e+00 -6.13394558e-01
-7.81955659e-01 -1.84338719e-01 -8.60826001e-02 5.54000795e-01
6.81681931e-01 5.65581203e-01 -8.86446163e-02 5.52095592e-01
2.78978914e-01 -2.81954646e-01 7.23867342e-02 -1.28341949e+00
-8.92144255e-03 -2.03170434e-01 -7.37182140e-01 -5.83363831e-01
-3.96812335e-02 -5.31769633e-01 3.44661325e-01 -1.16839838e+00
3.51719409e-01 3.83689135e-01 -5.42894244e-01 2.76176214e-01
-1.36695087e-01 4.76144612e-01 -3.51335436e-01 -8.97352993e-02
-3.16848755e-01 2.04713821e-01 5.45742512e-01 9.03368965e-02
-4.41126913e-01 6.72418773e-02 -5.70344806e-01 7.70467281e-01
6.44165874e-01 -5.78674793e-01 -5.09652972e-01 1.70231953e-01
2.09986463e-01 2.98658580e-01 5.16538441e-01 -1.41071558e+00
7.07817972e-01 3.88936371e-01 5.64040244e-01 -5.61055541e-01
3.66953015e-01 -6.34938538e-01 2.86269665e-01 2.48334169e-01
2.10242242e-01 3.08371186e-01 5.28465092e-01 5.75668335e-01
-1.05921276e-01 -1.27621114e-01 7.80041099e-01 1.38396189e-01
-3.11721414e-01 -4.26546514e-01 -1.13907027e+00 -2.84275383e-01
1.06268978e+00 -5.45574725e-01 -3.49366337e-01 -4.05191451e-01
-8.89286160e-01 -9.60786082e-03 2.25404501e-01 -1.01218693e-01
6.73493147e-01 -8.12796474e-01 -4.67447519e-01 4.87190872e-01
-5.28900474e-02 -7.60875046e-01 2.53972054e-01 1.72882056e+00
-1.57375827e-01 2.61628896e-01 -6.46379054e-01 -6.09341145e-01
-1.36831605e+00 2.89913118e-01 4.50505078e-01 3.83162260e-01
-3.73366356e-01 8.25037062e-01 7.49214590e-02 8.60989332e-01
2.65259206e-01 -5.67479074e-01 -4.10468161e-01 4.34948295e-01
6.28537416e-01 6.97412193e-01 6.25691354e-01 -5.03695786e-01
-6.93542302e-01 2.85681635e-01 4.24765646e-01 -2.00214863e-01
1.03587496e+00 -2.86939085e-01 -4.41722602e-01 1.18569231e+00
5.60426295e-01 1.89580768e-01 -9.26688731e-01 5.91984808e-01
-1.01342937e-02 -2.02835932e-01 9.41508710e-02 -1.23700058e+00
-6.55986488e-01 6.33569181e-01 1.17897296e+00 3.47494841e-01
1.78474557e+00 -2.75677204e-01 9.59863067e-02 5.83805181e-02
3.32987279e-01 -1.27358055e+00 1.18701179e-02 -1.15437480e-02
8.75409544e-01 -7.32866585e-01 4.62378822e-02 -2.07977325e-01
-5.24811625e-01 1.40751052e+00 4.80545491e-01 -2.84045845e-01
3.22830439e-01 4.41713125e-01 -1.37216076e-01 -3.17656308e-01
-8.11930239e-01 -1.12854421e-01 4.27869737e-01 4.63495433e-01
3.11188728e-01 1.59097388e-02 -8.59295130e-01 1.19339871e+00
-1.36240169e-01 9.73229334e-02 8.71203721e-01 4.37080115e-01
-3.90092343e-01 -1.25476107e-01 -7.55284786e-01 8.21828783e-01
-3.81018877e-01 -2.56263018e-01 -2.74008781e-01 4.96556640e-01
3.86012852e-01 1.41969693e+00 3.61469865e-01 -2.54301727e-01
4.15648103e-01 4.96671975e-01 4.74279583e-01 -6.34592593e-01
-6.92997396e-01 4.70487028e-01 -1.29522592e-01 -6.13323569e-01
-5.22804081e-01 -5.74850082e-01 -1.03096414e+00 3.31357390e-01
-4.40694839e-01 2.06244901e-01 4.91431117e-01 9.61430430e-01
4.29063439e-01 5.82728267e-01 2.27416784e-01 -7.52286911e-01
-5.13845086e-02 -1.48769641e+00 -1.27734756e+00 1.82199199e-02
7.08501756e-01 -8.81100714e-01 -7.49724627e-01 5.46891868e-01] | [13.385936737060547, 3.4607253074645996] |
7fa142fc-58de-4941-a988-e32e0242e632 | sar-image-despeckling-using-quadratic-linear | 1801.04751 | null | http://arxiv.org/abs/1801.04751v1 | http://arxiv.org/pdf/1801.04751v1.pdf | SAR Image Despeckling Using Quadratic-Linear Approximated L1-Norm | Speckle noise, inherent in synthetic aperture radar (SAR) images, degrades
the performance of the various SAR image analysis tasks. Thus, speckle noise
reduction is a critical preprocessing step for smoothing homogeneous regions
while preserving details. This letter proposes a variational despeckling
approach where L1-norm total variation regularization term is approximated in a
quadratic and linear manner to increase accuracy while decreasing the
computation time. Despeckling performance and computational efficiency of the
proposed method are shown using synthetic and real-world SAR images. | ['Fatih Nar'] | 2018-01-15 | null | null | null | null | ['sar-image-despeckling'] | ['computer-vision'] | [ 5.41757286e-01 -4.01516706e-01 4.62508619e-01 -4.45019186e-01
-8.17992091e-01 -4.88244563e-01 4.21277136e-01 -2.96867400e-01
-5.77857196e-01 8.84890318e-01 1.57031149e-01 -1.73251554e-02
-4.10479635e-01 -5.78664660e-01 -4.82900366e-02 -1.09537661e+00
2.20952183e-01 -1.44175887e-01 1.65066898e-01 6.58911839e-02
1.92860872e-01 8.62238944e-01 -8.92296374e-01 -1.54212609e-01
8.68350029e-01 9.55719292e-01 2.90220439e-01 5.52576363e-01
4.15794671e-01 5.20160615e-01 -3.72445583e-01 8.25182498e-02
7.05425739e-01 -5.16909659e-01 6.39000256e-03 5.65397501e-01
4.09834415e-01 -1.38280660e-01 -4.48541015e-01 1.62671161e+00
4.38312918e-01 5.05450130e-01 5.69147408e-01 -1.12164095e-01
-6.77816212e-01 1.43300489e-01 -1.02646935e+00 8.09798658e-01
-5.89634657e-01 2.36987621e-01 4.13569719e-01 -1.07408655e+00
4.85718220e-01 1.01478004e+00 5.42520285e-01 1.13166302e-01
-1.31921875e+00 -2.67371893e-01 -5.44433355e-01 -1.36227885e-04
-1.49091613e+00 -6.50912464e-01 1.12184882e+00 -4.19394910e-01
2.93199509e-01 4.53970343e-01 3.61617297e-01 1.96568102e-01
6.54395878e-01 7.70569220e-02 1.45323074e+00 -1.29152507e-01
1.81758374e-01 -3.15406680e-01 3.84881437e-01 3.87626916e-01
7.08639324e-01 2.06115276e-01 -9.08990763e-03 -1.74392909e-01
8.55530143e-01 -5.28890230e-02 -4.29564744e-01 -2.59658098e-01
-1.17079997e+00 9.43841159e-01 4.34153557e-01 4.62942898e-01
-6.48194551e-01 -9.74196643e-02 1.37686282e-01 2.62739867e-01
7.63243258e-01 3.54417175e-01 1.02436557e-01 4.14004624e-01
-1.59685194e+00 5.47410585e-02 8.90146568e-02 3.76159817e-01
4.90682095e-01 9.36408758e-01 -2.51188397e-01 7.02307343e-01
2.82315314e-01 1.20016932e+00 9.49119702e-02 -8.43285561e-01
7.26063102e-02 2.62048811e-01 3.16948265e-01 -1.35555351e+00
-3.45875740e-01 -8.77116084e-01 -1.32703948e+00 5.57270646e-01
2.20927969e-01 -2.19138458e-01 -1.23281586e+00 1.17694020e+00
2.54967183e-01 1.25899658e-01 4.61164445e-01 1.31700099e+00
5.61718881e-01 8.08826804e-01 -4.36865911e-02 -7.34880328e-01
1.13573301e+00 -5.09708226e-01 -1.16215026e+00 -6.49537325e-01
-5.98433614e-02 -1.33310676e+00 4.83846128e-01 2.28762567e-01
-1.08988285e+00 -3.92574728e-01 -1.17741954e+00 8.94291028e-02
3.57044548e-01 5.71485639e-01 1.70230076e-01 5.37752569e-01
-5.51276386e-01 3.87888134e-01 -1.01308370e+00 8.70221034e-02
6.04699910e-01 -1.66803598e-01 -2.31289178e-01 -2.71330684e-01
-7.49741971e-01 8.79391432e-01 1.18299879e-01 8.35157812e-01
-5.00447392e-01 -7.44677842e-01 -7.97902167e-01 -6.59046322e-02
1.59986079e-01 -3.02068383e-01 5.40685356e-01 -7.33847439e-01
-1.20853949e+00 6.73059285e-01 -2.21414715e-01 -6.92450941e-01
4.55294997e-01 -1.49632290e-01 -6.59231663e-01 2.14044094e-01
-2.15934560e-01 -3.32779884e-01 1.44482946e+00 -1.15991485e+00
1.16219826e-01 -5.21620631e-01 -8.94303560e-01 1.10337988e-01
4.95289028e-01 -1.27147526e-01 4.04362947e-01 -1.09723401e+00
5.74657917e-01 -6.27365530e-01 -5.75813890e-01 -1.31636739e-01
-5.76764010e-02 6.04500413e-01 1.24943674e+00 -1.07515419e+00
1.03134727e+00 -2.48259330e+00 -1.36690186e-02 3.48112434e-01
1.45812005e-01 6.02340102e-01 -2.58096512e-02 -1.02795109e-01
7.76369125e-02 -4.45274740e-01 -9.66276646e-01 2.58019239e-01
-7.27794826e-01 -1.62292108e-01 -2.49703452e-01 1.01233017e+00
1.98375508e-01 7.85056412e-01 -3.85614306e-01 -2.37299517e-01
4.52774346e-01 6.56147718e-01 -1.29758134e-01 -3.96531299e-02
2.07844470e-02 7.86193073e-01 -6.04723930e-01 5.07481217e-01
1.34573305e+00 8.06704909e-02 -8.70158970e-02 -7.43417740e-01
-3.62116218e-01 -7.16238797e-01 -1.21424222e+00 9.81339335e-01
-2.30275601e-01 9.70734835e-01 8.02538216e-01 -1.25422931e+00
1.36872339e+00 -9.48943645e-02 2.57275730e-01 -6.32552862e-01
2.78187960e-01 2.51753211e-01 -1.14718098e-02 -3.76653165e-01
5.33783436e-01 -7.05614626e-01 1.05294444e-01 -1.21859185e-01
-7.58368492e-01 -5.05890429e-01 -3.91388796e-02 -1.10560335e-01
7.33819008e-01 -4.86546040e-01 7.89958775e-01 -7.19603956e-01
7.14264989e-01 5.92979908e-01 6.84644818e-01 4.23616767e-01
-3.20510060e-01 4.92864102e-01 -1.56350985e-01 -5.10022104e-01
-1.25725400e+00 -1.18115473e+00 -4.04393554e-01 1.15013771e-01
1.17475964e-01 6.58682644e-01 -4.72416759e-01 4.37082313e-02
-3.27104807e-01 7.70593822e-01 -5.01468599e-01 1.51759028e-01
-9.56787705e-01 -1.02192605e+00 1.75541043e-01 -2.55587185e-03
1.00507343e+00 -6.03256404e-01 -8.34529459e-01 3.30062717e-01
6.02036826e-02 -1.13295293e+00 -6.05806530e-01 -1.54315799e-01
-1.15038931e+00 -1.11341643e+00 -7.68798649e-01 -7.34245837e-01
8.95930648e-01 6.27104163e-01 6.82655573e-01 -2.59221256e-01
-7.82471478e-01 4.22518216e-02 -9.16944668e-02 1.56340655e-02
-1.87100664e-01 -7.42153227e-01 -2.05270529e-01 3.00433576e-01
1.33228108e-01 -4.11714315e-01 -8.02132487e-01 1.32887051e-01
-1.11056733e+00 -4.31030035e-01 9.19449925e-01 1.19459462e+00
8.69373560e-01 4.90326524e-01 1.68004826e-01 -7.57303059e-01
4.85642940e-01 8.47502425e-02 -1.18462515e+00 7.00400770e-02
-4.06361103e-01 -5.94848208e-02 7.03614235e-01 -2.43958876e-01
-1.55205667e+00 4.57044169e-02 3.16992491e-01 -3.41953248e-01
1.90516368e-01 6.91190600e-01 1.21065348e-01 -5.47463536e-01
7.15196133e-01 5.94414532e-01 2.69606888e-01 -4.75606173e-01
2.68261856e-03 3.89013678e-01 8.17732394e-01 2.96765625e-01
1.31506908e+00 8.41820121e-01 5.13691664e-01 -1.59916556e+00
-9.20287728e-01 -4.97526765e-01 -4.29477125e-01 -1.60025716e-01
9.58496630e-01 -8.52182925e-01 -3.41653645e-01 3.62113804e-01
-7.46233582e-01 -4.53057103e-02 -2.63471991e-01 8.35152805e-01
-1.87100306e-01 5.69596350e-01 -3.24673831e-01 -9.36896443e-01
-5.62195063e-01 -9.15107191e-01 5.02695799e-01 3.67676824e-01
3.60464156e-01 -8.43485594e-01 -4.91987467e-02 4.88593161e-01
8.39514375e-01 6.02377713e-01 6.78137004e-01 -3.83096077e-02
-5.51199079e-01 -1.44581065e-01 -6.12111628e-01 4.57470417e-01
1.60193756e-01 -3.84488851e-01 -4.09426987e-01 -5.15094936e-01
7.60500491e-01 1.63985595e-01 9.89098907e-01 1.14435840e+00
4.01204318e-01 -3.56755763e-01 1.37120426e-01 6.92403436e-01
2.05746531e+00 3.50783050e-01 8.68612170e-01 1.65042356e-01
2.71404564e-01 3.78150940e-01 8.05311978e-01 4.42002922e-01
-6.16983294e-01 2.60969758e-01 1.07994720e-01 -1.32391736e-01
-4.27418560e-01 4.37367022e-01 1.15052141e-01 6.16566420e-01
2.88412836e-03 4.51569492e-03 -6.53114617e-01 7.16995060e-01
-1.47830260e+00 -1.17759669e+00 -6.02434337e-01 1.95424843e+00
4.42969322e-01 -3.01132381e-01 -7.01456606e-01 8.24897885e-02
7.90999413e-01 6.84079826e-01 -6.88288152e-01 -2.75591575e-02
-7.02188849e-01 3.09378892e-01 8.62855673e-01 1.04872990e+00
-1.24397385e+00 7.46665299e-01 6.33518744e+00 9.31117654e-01
-1.24784517e+00 5.18916640e-03 4.95906383e-01 2.11353987e-01
4.90956358e-04 -1.13385662e-01 -3.34185988e-01 2.23837703e-01
5.43314159e-01 -4.05911714e-01 2.91382521e-01 4.82888013e-01
8.15671563e-01 -4.88852650e-01 1.76423788e-01 9.73847210e-01
9.63259041e-02 -1.47806919e+00 4.72761281e-02 -1.96955577e-01
8.00272524e-01 -1.52602538e-01 2.59118497e-01 -3.57779145e-01
6.79920018e-02 -9.89341080e-01 1.73492551e-01 9.36951518e-01
6.51615083e-01 -8.51834714e-01 9.15238917e-01 7.79362470e-02
-1.01376915e+00 1.40939415e-01 -5.71759462e-01 6.33650273e-02
5.91824830e-01 1.16780925e+00 -3.61256599e-01 4.64515924e-01
3.36056888e-01 5.68853796e-01 -3.93161654e-01 9.69733953e-01
2.34299321e-02 6.36484444e-01 -1.25486270e-01 4.17500228e-01
3.97778928e-01 -9.07039404e-01 1.30767727e+00 9.63466465e-01
5.52547574e-01 9.59479153e-01 1.63010508e-01 7.15925932e-01
3.92671704e-01 9.36703291e-03 -4.10853803e-01 -1.82895213e-01
1.33823425e-01 1.13014054e+00 -7.98076510e-01 3.18330079e-02
-1.72704488e-01 6.93381369e-01 -6.37929022e-01 5.20680249e-01
-4.64922547e-01 -5.01289606e-01 4.72143859e-01 2.87873983e-01
5.05666137e-01 -7.25747466e-01 -7.05275834e-01 -9.45600331e-01
-7.44059756e-02 -5.37141085e-01 2.11832479e-01 -6.57319844e-01
-1.05124724e+00 8.47278118e-01 -1.52749315e-01 -1.47011364e+00
3.57885242e-01 -3.02115798e-01 -5.54283977e-01 9.66442764e-01
-1.55832589e+00 -1.05471730e+00 -4.60085869e-01 4.56952006e-01
7.59895921e-01 -4.28449571e-01 2.69532830e-01 3.29618645e-03
-3.22071463e-01 -6.92956075e-02 4.57899719e-01 2.00569034e-02
3.40341598e-01 -5.07794023e-01 2.81866416e-02 1.60203445e+00
-3.89333487e-01 3.92270416e-01 1.42173636e+00 -8.90929580e-01
-1.15476954e+00 -1.17176545e+00 4.87906963e-01 5.68863988e-01
7.12755382e-01 2.29340509e-01 -1.07464898e+00 5.17501175e-01
2.65584767e-01 3.70654941e-01 2.69068360e-01 -9.17830050e-01
1.31728008e-01 -9.87159833e-02 -1.44575715e+00 4.60532963e-01
1.98315427e-01 6.60226494e-02 -6.96962893e-01 1.08755879e-01
1.83378622e-01 -2.79267132e-01 -8.10601592e-01 7.23980308e-01
7.77191818e-02 -7.57282615e-01 1.07731497e+00 -9.08159465e-02
1.53744861e-01 -7.02285469e-01 -2.66628981e-01 -1.32200813e+00
-5.48733175e-01 -7.76886284e-01 2.67966062e-01 6.79959297e-01
2.75495201e-01 -5.76410890e-01 8.29409301e-01 1.50606275e-01
-4.48879525e-02 -1.58173397e-01 -9.50409710e-01 -8.55769694e-01
-1.52847290e-01 1.26870424e-01 -1.68732941e-01 8.19783866e-01
-6.86442316e-01 1.20791234e-01 -7.28346407e-01 4.77517694e-01
1.52740514e+00 1.51340008e-01 3.29229951e-01 -1.21336269e+00
9.81023163e-02 1.73261145e-03 -4.88256574e-01 -7.92445123e-01
-7.40602389e-02 -6.12541497e-01 1.52528897e-01 -1.33985174e+00
-6.28191531e-02 -1.80871889e-01 1.49038732e-01 -2.92518467e-01
-1.02941096e-01 6.86863184e-01 5.35409898e-02 5.51755607e-01
2.31811076e-01 6.75934315e-01 1.40059614e+00 -2.36949921e-01
-2.09860712e-01 1.84111983e-01 -2.43082255e-01 7.93808460e-01
9.33910251e-01 -4.87026364e-01 -2.30296746e-01 -5.38896680e-01
-2.07974061e-01 3.23080927e-01 4.83575791e-01 -1.07011056e+00
3.57608765e-01 -4.54571009e-01 5.60244620e-01 -6.71776950e-01
3.63991797e-01 -9.00608182e-01 3.60071242e-01 6.32706165e-01
-1.05323322e-01 -1.55894130e-01 2.28804067e-01 9.25819635e-01
-5.77745616e-01 -1.78053543e-01 1.67999053e+00 2.81407218e-02
-7.63952792e-01 9.28689688e-02 -4.35613841e-01 -8.13473165e-02
1.17325068e+00 -4.31887716e-01 -2.00275540e-01 -1.12723999e-01
-9.62868392e-01 -2.62551337e-01 1.75752625e-01 -4.30338979e-01
9.60885048e-01 -9.33529317e-01 -1.12382317e+00 5.51287949e-01
-2.43735701e-01 -4.59530532e-01 8.58623624e-01 1.14671063e+00
-1.08676708e+00 1.96156025e-01 -3.51441115e-01 -4.01943535e-01
-1.45923960e+00 3.33508104e-01 3.14107060e-01 -2.86029249e-01
-9.99804258e-01 5.85664570e-01 -3.07370313e-02 2.39547655e-01
-3.79852146e-01 9.59401280e-02 -2.21581116e-01 -1.72614992e-01
7.94179916e-01 6.44621968e-01 -1.29020572e-01 -9.00496066e-01
-3.15247446e-01 9.65371788e-01 4.40088771e-02 -3.08048278e-01
1.44812739e+00 -5.52746534e-01 -2.82312930e-01 7.79593214e-02
9.57564414e-01 4.32222039e-01 -1.26107836e+00 -4.41042900e-01
4.61460799e-02 -8.90940845e-01 8.10080826e-01 -5.16892374e-01
-1.18201852e+00 4.62203234e-01 8.72012138e-01 -5.51910028e-02
1.47912562e+00 -4.14085299e-01 7.92543828e-01 3.59952420e-01
-7.87921697e-02 -9.95461047e-01 -3.09670925e-01 2.63178676e-01
8.83657455e-01 -1.34019971e+00 6.06473804e-01 -4.91285264e-01
-7.07836211e-01 1.10990465e+00 -1.76547654e-03 -7.02503920e-01
7.81333327e-01 4.41420227e-01 4.32231605e-01 -3.90516073e-01
-2.12079346e-01 -1.78147793e-01 3.90425742e-01 3.47969174e-01
2.77530491e-01 7.61634782e-02 -9.26141918e-01 1.07920356e-01
1.61334217e-01 -2.17624322e-01 6.17299736e-01 7.85642564e-01
-9.57686603e-01 -4.29962605e-01 -8.87652397e-01 5.26584029e-01
-6.85830891e-01 -1.77938223e-01 1.17358238e-01 5.83302140e-01
-4.43170458e-01 1.05016136e+00 1.14928866e-02 2.92984903e-01
3.44811529e-01 -3.73458415e-01 1.88046053e-01 -1.32673413e-01
-2.43680477e-01 5.17624378e-01 -1.07979007e-01 -1.76061913e-01
-6.19772971e-01 -5.52268744e-01 -8.77517700e-01 -2.18320072e-01
-1.83752254e-01 3.46499652e-01 5.53932488e-01 8.67034256e-01
8.15192387e-02 3.52267206e-01 6.81751072e-01 -5.10880589e-01
-6.69382513e-01 -9.33700800e-01 -1.00474203e+00 1.30248263e-01
5.41615903e-01 -4.38267887e-01 -8.27750266e-01 3.53437543e-01] | [10.44385814666748, -2.2218053340911865] |
50c46494-87d2-4707-8716-65e4d94450bd | super-nerf-view-consistent-detail-generation | 2304.13518 | null | https://arxiv.org/abs/2304.13518v1 | https://arxiv.org/pdf/2304.13518v1.pdf | Super-NeRF: View-consistent Detail Generation for NeRF super-resolution | The neural radiance field (NeRF) achieved remarkable success in modeling 3D scenes and synthesizing high-fidelity novel views. However, existing NeRF-based methods focus more on the make full use of the image resolution to generate novel views, but less considering the generation of details under the limited input resolution. In analogy to the extensive usage of image super-resolution, NeRF super-resolution is an effective way to generate the high-resolution implicit representation of 3D scenes and holds great potential applications. Up to now, such an important topic is still under-explored. In this paper, we propose a NeRF super-resolution method, named Super-NeRF, to generate high-resolution NeRF from only low-resolution inputs. Given multi-view low-resolution images, Super-NeRF constructs a consistency-controlling super-resolution module to generate view-consistent high-resolution details for NeRF. Specifically, an optimizable latent code is introduced for each low-resolution input image to control the 2D super-resolution images to converge to the view-consistent output. The latent codes of each low-resolution image are optimized synergistically with the target Super-NeRF representation to fully utilize the view consistency constraint inherent in NeRF construction. We verify the effectiveness of Super-NeRF on synthetic, real-world, and AI-generated NeRF datasets. Super-NeRF achieves state-of-the-art NeRF super-resolution performance on high-resolution detail generation and cross-view consistency. | ['Qionghai Dai', 'Yuwang Wang', 'Xiaohang Yu', 'Tao Yu', 'Yuqi Han'] | 2023-04-26 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 3.27391803e-01 2.96826139e-02 4.44462039e-02 -3.62298846e-01
-9.97964501e-01 -2.49769956e-01 4.98564929e-01 -7.35996962e-01
2.63036579e-01 7.12197781e-01 5.45439184e-01 3.21352452e-01
-1.97602719e-01 -1.22897255e+00 -7.94928312e-01 -6.16693497e-01
2.62402594e-01 -5.30664984e-04 1.97253630e-01 -4.64104205e-01
5.16221523e-02 7.74523914e-01 -1.85249519e+00 6.51481748e-01
1.07783425e+00 5.93547642e-01 8.29290986e-01 6.91200554e-01
-2.95154881e-02 4.95578498e-01 -2.21073806e-01 2.72947215e-02
4.23361272e-01 -5.50017178e-01 -4.18379605e-01 1.37294605e-01
7.93548286e-01 -6.35121703e-01 -1.97367892e-01 9.45956230e-01
4.33798671e-01 1.83541954e-01 3.95120978e-01 -4.79160845e-01
-1.07471657e+00 2.88118273e-01 -9.22929287e-01 3.33780974e-01
5.19783020e-01 -1.69999152e-01 6.58050954e-01 -1.11698794e+00
9.79552031e-01 1.34557605e+00 4.93555069e-01 7.06117392e-01
-1.51758182e+00 -6.47282004e-01 5.53546883e-02 -2.64547706e-01
-1.44926536e+00 -2.91650355e-01 8.76733065e-01 -1.80680528e-01
7.82447815e-01 2.43928388e-01 5.05474031e-01 1.13735437e+00
4.47862297e-01 6.33606268e-03 1.44137836e+00 -2.99735367e-01
1.58353299e-01 9.44410488e-02 -4.54504311e-01 4.34778005e-01
1.43572912e-01 5.05194902e-01 -7.22309709e-01 1.12496309e-01
1.66744077e+00 -4.07064110e-02 -4.57296193e-01 -5.45651950e-02
-1.27486813e+00 5.38727224e-01 6.35776877e-01 3.87781084e-01
-6.37274206e-01 -2.03745395e-01 -2.18111113e-01 -1.88866377e-01
7.88410962e-01 6.21114910e-01 -3.01764071e-01 2.81931788e-01
-8.83939981e-01 2.43697628e-01 -7.57328793e-02 9.30346072e-01
8.20137858e-01 4.68207151e-01 -1.66210383e-01 1.07945907e+00
-1.78178355e-01 4.64124203e-01 1.94005728e-01 -1.44502270e+00
3.06013018e-01 5.30780315e-01 1.71985835e-01 -8.52719009e-01
-2.48502523e-01 -5.57631731e-01 -1.34513831e+00 5.87021708e-01
-2.98019558e-01 1.34137064e-01 -9.63650107e-01 1.76611483e+00
4.62516516e-01 5.94879501e-02 3.09618413e-01 1.22156310e+00
9.31847632e-01 1.10743845e+00 -3.69938403e-01 -4.12172616e-01
1.25103283e+00 -7.20236063e-01 -8.50283742e-01 -1.18523359e-01
-3.60324949e-01 -6.36571467e-01 1.11631322e+00 2.32587203e-01
-1.44402945e+00 -9.00960505e-01 -9.79285359e-01 -2.84543097e-01
1.62679795e-02 -1.31586730e-01 4.32486981e-01 1.63485704e-04
-1.00981617e+00 3.94269168e-01 -3.34525853e-01 -6.79838434e-02
3.47659379e-01 -1.88346401e-01 -6.21886194e-01 -4.84324992e-01
-1.27102983e+00 7.46352136e-01 4.64936525e-01 3.13849077e-02
-8.68728220e-01 -1.08960330e+00 -9.48825359e-01 1.05306143e-02
3.86290669e-01 -1.16716921e+00 8.05083930e-01 -7.28424191e-01
-1.48329175e+00 6.90599024e-01 -2.48820186e-01 1.21482089e-02
2.45288596e-01 9.83082652e-02 -5.35502613e-01 2.50347286e-01
2.84603417e-01 7.22333789e-01 8.93687606e-01 -1.90394032e+00
-7.21206069e-01 -2.86419660e-01 1.59725100e-01 5.57444990e-01
2.90638488e-02 -2.79619068e-01 -2.86335468e-01 -8.41172993e-01
3.24157059e-01 -3.82267803e-01 -4.61037040e-01 -8.57388452e-02
-1.57727569e-01 3.91706437e-01 7.06371307e-01 -5.61478555e-01
1.05252576e+00 -2.08601213e+00 3.64672631e-01 -1.87967911e-01
3.81562620e-01 -3.13059427e-02 -1.79786876e-01 4.24555317e-03
-1.32312879e-01 1.42081887e-01 -2.60273397e-01 -1.57754928e-01
-5.93656361e-01 1.99996144e-01 -3.30937952e-01 5.96886836e-02
8.75788629e-02 7.63288498e-01 -8.47826898e-01 -3.36640000e-01
4.40060318e-01 1.23152554e+00 -5.52900851e-01 3.46572757e-01
-8.60923603e-02 8.75255287e-01 -5.39880514e-01 3.54631931e-01
1.08685839e+00 -3.56963485e-01 -6.15520924e-02 -6.96637094e-01
-5.60279071e-01 -2.76604712e-01 -1.25921488e+00 1.86331737e+00
-7.48535872e-01 2.73896188e-01 1.15436308e-01 -1.58939674e-01
1.08424830e+00 1.57076493e-02 4.36365426e-01 -1.13706958e+00
-2.80376285e-01 8.44341367e-02 -6.57888770e-01 -1.83001965e-01
8.80209327e-01 -4.76461917e-01 1.19727869e-02 1.06756292e-01
-7.62927234e-02 -2.80540884e-01 -1.67940110e-01 1.38866290e-01
4.29294318e-01 3.65324497e-01 3.89683187e-01 -6.35696724e-02
5.51139593e-01 -9.91554856e-02 6.63122654e-01 5.02266049e-01
3.90667707e-01 1.34032929e+00 -1.52736664e-01 -3.65767002e-01
-1.65970349e+00 -1.37016726e+00 -4.05147254e-01 6.10882103e-01
3.52678418e-01 -4.02127266e-01 -8.29821944e-01 -1.13846712e-01
-3.56127918e-01 6.98740661e-01 -7.16344655e-01 6.28502443e-02
-6.60290956e-01 -5.36376476e-01 -2.05292907e-02 3.99781466e-01
1.01047611e+00 -9.74102437e-01 -6.89537346e-01 1.46142676e-01
-4.56912130e-01 -1.45151699e+00 -4.90668535e-01 -1.77523986e-01
-9.16811645e-01 -6.38061106e-01 -9.12364900e-01 -5.47079265e-01
6.75155401e-01 7.57494271e-01 1.20763111e+00 -3.89794707e-01
-3.88919830e-01 1.37052327e-01 -1.59219146e-01 2.14109212e-01
-5.55178761e-01 -2.05011681e-01 -4.29557189e-02 -1.89424772e-02
-4.98417288e-01 -8.81140172e-01 -7.02439964e-01 2.94146925e-01
-1.15607643e+00 9.31834519e-01 6.82250559e-01 8.31845105e-01
1.21030819e+00 4.66603786e-01 3.19764823e-01 -7.34602690e-01
2.82997847e-01 -1.56873718e-01 -6.26581967e-01 1.12298414e-01
-4.83299047e-01 2.61526495e-01 9.33126986e-01 -3.83729488e-01
-1.93117678e+00 -1.56069458e-01 -5.81737123e-02 -7.94425786e-01
-1.35464489e-01 1.30460024e-01 -3.29426199e-01 9.48057100e-02
7.67943263e-01 4.03798670e-01 -3.92930031e-01 -6.20973647e-01
5.41241229e-01 3.40015590e-01 8.29529822e-01 -5.21089792e-01
1.01121581e+00 8.13409626e-01 1.89577207e-01 -6.95733666e-01
-1.21780539e+00 3.85548621e-02 -6.45072877e-01 -1.51120678e-01
1.23859870e+00 -1.25718629e+00 -1.41028717e-01 2.09735706e-01
-1.03448379e+00 -2.34090418e-01 -4.61140096e-01 3.65625024e-01
-6.19820476e-01 8.61845165e-02 -4.47576076e-01 -6.70244336e-01
-3.22598606e-01 -1.05006862e+00 1.32351339e+00 4.95068401e-01
2.33647496e-01 -6.34173632e-01 1.04130320e-01 5.45508802e-01
6.09204054e-01 5.80711782e-01 8.75092626e-01 6.77724540e-01
-1.05867064e+00 5.85950673e-01 -5.54723024e-01 3.52278501e-01
2.70238459e-01 -2.13957548e-01 -9.71268833e-01 -2.39750192e-01
5.69401607e-02 -2.50396729e-02 5.84617317e-01 5.69911122e-01
1.05129492e+00 -2.13119179e-01 -5.79229929e-02 1.12044573e+00
1.98686934e+00 1.43607125e-01 8.18484664e-01 3.96685034e-01
9.05457556e-01 5.04634619e-01 6.22737646e-01 4.51358527e-01
4.32891488e-01 1.10598409e+00 3.02496165e-01 -3.36145639e-01
-5.81522405e-01 -5.69498718e-01 7.52911717e-02 4.94789600e-01
-5.10214627e-01 4.83306423e-02 -3.88866305e-01 2.28440642e-01
-1.52720380e+00 -1.42126322e+00 -2.12142542e-02 2.07607007e+00
6.80401862e-01 -1.56328321e-01 -2.71000981e-01 -3.25048596e-01
8.78632605e-01 4.70396876e-01 -7.09460676e-01 -2.45828196e-01
-6.09406888e-01 1.13643043e-01 2.25961179e-01 6.93914652e-01
-6.52980208e-01 8.40837479e-01 5.77663994e+00 7.82141745e-01
-1.04758763e+00 1.95960477e-01 7.50389159e-01 -1.04789719e-01
-8.45829964e-01 -1.41176522e-01 -9.46421027e-01 2.36293212e-01
5.78839958e-01 -2.52710342e-01 7.74321318e-01 7.08928525e-01
4.19687063e-01 -6.01238385e-02 -5.51263392e-01 1.31724024e+00
2.14749277e-01 -1.64956307e+00 4.82782155e-01 1.24901086e-01
1.34626079e+00 -3.57024223e-01 1.55945346e-01 1.40304805e-03
7.36263469e-02 -1.15489912e+00 5.33106327e-01 8.19942951e-01
1.51312888e+00 -1.02708220e+00 1.63128510e-01 2.39605710e-01
-1.51637602e+00 -7.84264356e-02 -5.38257420e-01 4.19565350e-01
5.24340630e-01 7.33803213e-01 -1.86853558e-01 9.74224567e-01
1.04339814e+00 6.79501534e-01 -4.05843973e-01 2.90416956e-01
-1.00823447e-01 -1.94640785e-01 -4.08121676e-04 9.27393734e-01
-4.67514656e-02 -3.26425731e-01 6.64542019e-01 8.87550652e-01
5.57979167e-01 6.02888286e-01 -3.37303169e-02 1.44392562e+00
5.81407398e-02 -2.49420956e-01 -6.93524003e-01 4.59216923e-01
4.65949655e-01 1.37125552e+00 -3.79121661e-01 -2.59482861e-01
-3.25127900e-01 1.15936339e+00 3.30703586e-01 5.67259967e-01
-8.40576768e-01 7.16890693e-02 7.53942013e-01 5.08371592e-01
3.96171093e-01 -1.41601384e-01 -3.78167123e-01 -1.38403881e+00
7.49321878e-02 -8.96051884e-01 1.50429517e-01 -1.49696016e+00
-1.08137393e+00 1.11791158e+00 3.09984237e-01 -1.51404941e+00
-4.17299688e-01 -6.27274886e-02 -1.73309550e-01 1.28902912e+00
-1.79802990e+00 -1.19502079e+00 -9.25203860e-01 6.08178318e-01
8.31194520e-01 1.00028731e-01 7.33737707e-01 1.19752467e-01
-2.67008245e-01 1.59982115e-01 1.37691097e-02 -3.79849344e-01
5.68358541e-01 -8.89066756e-01 3.71395469e-01 1.03673673e+00
-1.72435746e-01 5.10554850e-01 5.83988249e-01 -7.18848050e-01
-1.17944896e+00 -1.24778748e+00 4.00414735e-01 -4.04174507e-01
-1.85386613e-01 -3.12879950e-01 -1.19853961e+00 3.00017089e-01
4.86671925e-02 2.66384482e-01 1.88093767e-01 -3.38156551e-01
-4.81173724e-01 -2.47366890e-01 -1.41304874e+00 7.38730729e-01
1.43754923e+00 -6.05798781e-01 -4.07167792e-01 -2.32619494e-01
9.07836139e-01 -7.89172232e-01 -1.16043305e+00 7.78353572e-01
4.93221283e-01 -1.58504760e+00 1.48132336e+00 -2.57034078e-02
9.57515478e-01 -7.04606116e-01 -3.40967923e-01 -1.41300583e+00
-9.47262466e-01 -3.78519237e-01 -2.99131963e-02 1.19968927e+00
3.66498437e-03 -3.39349210e-01 4.94064480e-01 3.12968940e-01
-1.81146473e-01 -7.91779637e-01 -6.62571311e-01 -4.41929698e-01
-2.24590361e-01 2.52449512e-02 7.27557659e-01 1.06126535e+00
-7.40466774e-01 4.07997161e-01 -4.91009355e-01 5.48365831e-01
1.08744740e+00 4.96510029e-01 6.91870511e-01 -9.65248883e-01
-2.51741290e-01 -1.06805660e-01 5.66675998e-02 -8.72510791e-01
-1.67694986e-02 -5.45650423e-01 -2.16923803e-02 -1.90957892e+00
4.81646746e-01 -2.47402400e-01 -3.85141820e-02 2.03904640e-02
-2.30764613e-01 4.36799079e-01 1.90789670e-01 2.84531683e-01
-1.63744256e-01 6.49733126e-01 1.99091780e+00 2.73214102e-01
-3.60368788e-01 -4.50268239e-01 -9.83065128e-01 5.42816043e-01
5.27718067e-01 -7.00756907e-02 -6.76348329e-01 -5.13318837e-01
2.46473983e-01 5.00814557e-01 3.70687753e-01 -8.91768098e-01
-1.55052856e-01 -4.70990241e-01 8.88679981e-01 -7.15225518e-01
6.15527868e-01 -7.79713750e-01 8.00381541e-01 -2.30874732e-01
-2.58326471e-01 8.42050985e-02 9.24327895e-02 4.28662837e-01
-2.33962402e-01 2.43212804e-01 1.22553384e+00 -5.67470670e-01
-9.35232222e-01 4.70459193e-01 9.26212594e-02 -1.54891713e-02
8.16797137e-01 -5.73529780e-01 -4.54917848e-01 -3.19979191e-01
-4.65773731e-01 -1.62404284e-01 9.02062476e-01 5.65083146e-01
1.13098693e+00 -1.54892862e+00 -9.30221140e-01 5.35685301e-01
1.59197003e-01 6.07699394e-01 1.12844360e+00 7.46798962e-02
-2.10183397e-01 1.38693213e-01 -5.07498562e-01 -6.24488890e-01
-1.14266288e+00 7.30655193e-01 4.78073955e-01 -3.12085211e-01
-1.23350227e+00 3.91562492e-01 8.90468657e-01 -2.96059877e-01
-4.11916345e-01 -1.50891766e-01 -3.34268600e-01 -3.83792222e-01
9.27191377e-01 1.95854947e-01 -2.74312437e-01 -7.98756003e-01
-1.01538543e-02 1.16669202e+00 2.50224918e-02 -2.60482550e-01
1.59255028e+00 -5.04261792e-01 6.24779128e-02 3.16571653e-01
1.06778061e+00 1.61726713e-01 -1.74544942e+00 -1.99518517e-01
-8.17286432e-01 -9.29864109e-01 3.61180663e-01 -7.05164611e-01
-1.18259227e+00 4.88832533e-01 6.19395435e-01 -2.70949721e-01
1.50620091e+00 4.43284884e-02 6.88444436e-01 -1.78434625e-01
7.40789294e-01 -7.34380901e-01 2.09361404e-01 3.89754623e-01
1.20393920e+00 -1.04808164e+00 2.67569721e-01 -6.48765266e-01
-6.75171375e-01 1.02713060e+00 1.04544115e+00 2.51288358e-02
1.69437855e-01 4.08805281e-01 -1.54307932e-01 -1.82985485e-01
-7.92188644e-01 -3.91934849e-02 3.61526698e-01 7.81629443e-01
2.37865433e-01 -1.78446829e-01 1.61771849e-01 4.31075692e-01
-2.24788487e-01 1.14090808e-01 7.39584029e-01 4.56343770e-01
-2.40543500e-01 -6.80882990e-01 -6.44867539e-01 8.69502723e-02
-1.65349573e-01 -1.33336231e-01 2.14054987e-01 5.90180814e-01
2.68713236e-01 5.41962743e-01 1.81139156e-01 -3.81361753e-01
6.10381484e-01 -4.85770106e-01 7.23168433e-01 -5.76947033e-01
-1.35911211e-01 1.19276896e-01 -9.24686715e-02 -8.23607683e-01
-5.68294644e-01 -2.66765743e-01 -1.14873004e+00 -2.70059675e-01
2.67592054e-02 -1.05259299e-01 5.00458241e-01 4.48375821e-01
6.30226612e-01 8.32095623e-01 9.30139065e-01 -1.30736637e+00
2.22282647e-03 -5.95596611e-01 -7.01949000e-01 4.76113826e-01
4.42520201e-01 -6.16944373e-01 -5.15391231e-01 3.17375213e-02] | [10.637517929077148, -2.2724404335021973] |
7ba69009-f29e-4367-bf7c-f7440567b166 | intelligent-3d-network-protocol-for | 2207.11504 | null | https://arxiv.org/abs/2207.11504v1 | https://arxiv.org/pdf/2207.11504v1.pdf | Intelligent 3D Network Protocol for Multimedia Data Classification using Deep Learning | In videos, the human's actions are of three-dimensional (3D) signals. These videos investigate the spatiotemporal knowledge of human behavior. The promising ability is investigated using 3D convolution neural networks (CNNs). The 3D CNNs have not yet achieved high output for their well-established two-dimensional (2D) equivalents in still photographs. Board 3D Convolutional Memory and Spatiotemporal fusion face training difficulty preventing 3D CNN from accomplishing remarkable evaluation. In this paper, we implement Hybrid Deep Learning Architecture that combines STIP and 3D CNN features to enhance the performance of 3D videos effectively. After implementation, the more detailed and deeper charting for training in each circle of space-time fusion. The training model further enhances the results after handling complicated evaluations of models. The video classification model is used in this implemented model. Intelligent 3D Network Protocol for Multimedia Data Classification using Deep Learning is introduced to further understand spacetime association in human endeavors. In the implementation of the result, the well-known dataset, i.e., UCF101 to, evaluates the performance of the proposed hybrid technique. The results beat the proposed hybrid technique that substantially beats the initial 3D CNNs. The results are compared with state-of-the-art frameworks from literature for action recognition on UCF101 with an accuracy of 95%. | ['Faisal Jamil', 'Harun Jamil', 'Ammar Muthanna', 'Abid Ali', 'Muhammad Munawar Iqbal', 'Eman A. Aldhahri', 'Arslan Syed'] | 2022-07-23 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [-3.28421950e-01 -4.78329778e-01 -2.39057988e-01 -1.19989261e-01
1.42904609e-01 -1.06076740e-01 6.63044691e-01 -5.95935166e-01
-6.34332776e-01 4.72245276e-01 3.32150698e-01 -9.87740383e-02
-2.02639446e-01 -5.54634988e-01 -5.20307720e-01 -6.11622095e-01
-5.03686130e-01 -1.21586882e-01 2.13696465e-01 -1.07679009e-01
1.54917479e-01 1.04713964e+00 -1.76954710e+00 5.72059691e-01
5.23992665e-02 1.50659609e+00 -3.06663305e-01 9.33864355e-01
-6.42659664e-02 1.01781642e+00 -7.64848888e-01 4.96185012e-02
4.93945271e-01 -9.61588845e-02 -6.71319067e-01 3.05560499e-01
4.89698470e-01 -8.34937036e-01 -1.00189722e+00 5.90264380e-01
6.73308551e-01 2.29915380e-01 6.58264697e-01 -1.41462255e+00
-7.52609670e-01 -1.59792349e-01 -2.04678208e-01 7.02034771e-01
5.60675442e-01 2.78549016e-01 6.38205260e-02 -6.64636075e-01
6.33355796e-01 1.33183396e+00 9.64722693e-01 4.83710885e-01
-3.45663160e-01 -4.54287589e-01 -1.33795962e-01 6.55697763e-01
-1.15809250e+00 1.83740985e-02 6.74699903e-01 -6.35558188e-01
1.56440532e+00 7.66366273e-02 1.07018781e+00 1.50060582e+00
4.07050341e-01 1.24378657e+00 8.85420918e-01 -1.67588785e-01
-1.06954999e-01 -8.73488486e-02 3.60183626e-01 8.14565361e-01
1.29804209e-01 4.66416687e-01 -6.86307669e-01 4.60496277e-01
1.06452048e+00 4.26432282e-01 -5.43116890e-02 -1.48376584e-01
-1.23972321e+00 4.38460708e-01 4.36910391e-01 6.18038833e-01
-4.72942770e-01 2.77403504e-01 8.08919668e-01 1.82324454e-01
3.67870241e-01 -1.18640922e-02 -4.26312059e-01 -4.68568444e-01
-6.78461552e-01 2.32916653e-01 2.60284156e-01 8.98962021e-01
1.31745279e-01 2.77167171e-01 -3.88068020e-01 4.45307195e-01
1.67182863e-01 4.46827590e-01 6.64194524e-01 -9.79494631e-01
1.68519974e-01 1.01870584e+00 -1.59773827e-02 -1.26237845e+00
-7.60665298e-01 -3.22966814e-01 -1.16473401e+00 4.71841007e-01
3.70832145e-01 3.58164054e-03 -1.01861846e+00 1.21923137e+00
2.63951898e-01 4.13018137e-01 2.12995544e-01 1.11152685e+00
1.27293670e+00 5.41529000e-01 -4.06777561e-02 1.86199844e-01
1.29023218e+00 -9.76542830e-01 -9.60190415e-01 5.88581681e-01
8.25621665e-01 -3.46646160e-01 4.54321295e-01 4.62365836e-01
-9.57762122e-01 -1.30537617e+00 -1.21692050e+00 4.28056419e-02
-9.42169309e-01 3.59724075e-01 7.06398547e-01 7.03104019e-01
-1.10326159e+00 6.30835474e-01 -8.65112185e-01 -6.38680577e-01
8.72753084e-01 4.54928488e-01 -9.25782263e-01 1.38733983e-01
-1.47869945e+00 1.08512259e+00 4.46115494e-01 3.64986837e-01
-1.07110083e+00 -4.59985137e-01 -7.86722362e-01 -2.62967765e-01
9.08175632e-02 -6.59874320e-01 1.00862110e+00 -7.91963935e-01
-1.30344999e+00 1.08828676e+00 2.99153477e-01 -7.86421120e-01
6.24708235e-01 -4.70520586e-01 -6.19330049e-01 5.16978264e-01
-1.42197132e-01 6.53127074e-01 6.10627890e-01 -5.76308370e-01
-6.82081521e-01 -5.17676473e-01 4.51940060e-01 3.12456582e-02
-3.86946023e-01 -7.61276186e-02 -4.55030203e-01 -6.70352817e-01
-1.39786571e-01 -7.31114388e-01 1.87679797e-01 3.69643390e-01
-1.14700682e-01 -3.07649314e-01 1.23984027e+00 -5.12638509e-01
1.21674478e+00 -2.14116693e+00 -2.97332183e-02 -2.67039776e-01
3.10407698e-01 8.39657307e-01 6.75884858e-02 2.70918071e-01
-2.23690033e-01 1.03472762e-01 1.96952641e-01 -1.44932464e-01
9.68281925e-02 2.32568502e-01 3.86109233e-01 7.09256113e-01
4.35929537e-01 1.01484787e+00 -4.99407828e-01 -5.77426314e-01
6.98272347e-01 6.99409246e-01 -5.12651384e-01 2.61886001e-01
3.48004460e-01 1.96811587e-01 -5.14515877e-01 9.23692405e-01
6.11883938e-01 -1.44239768e-01 -3.43991250e-01 -6.59959078e-01
-1.13292515e-01 -4.91650522e-01 -1.17782497e+00 1.83729196e+00
1.27977759e-01 8.91109049e-01 -3.82230252e-01 -1.52933896e+00
9.03802991e-01 5.62252879e-01 7.24167347e-01 -8.59110296e-01
5.33149898e-01 -7.50439912e-02 -2.03083858e-01 -1.37280107e+00
8.81495029e-02 3.17224294e-01 3.75903063e-02 1.57481864e-01
5.81070244e-01 4.33201373e-01 8.87964517e-02 -1.54966384e-01
1.16020083e+00 6.85488641e-01 3.34898442e-01 -1.04861960e-01
8.97465289e-01 3.49156596e-02 2.78217971e-01 6.88837767e-01
-8.93829703e-01 2.26659104e-01 5.15658438e-01 -1.13752544e+00
-9.44032311e-01 -8.07331681e-01 2.97716446e-02 5.05451798e-01
6.54813498e-02 8.80504325e-02 -7.46339440e-01 -9.29558516e-01
1.13616131e-01 8.30145031e-02 -1.10809553e+00 -1.70355052e-01
-6.88788474e-01 -4.78337884e-01 1.10680127e+00 7.93510377e-01
1.36785233e+00 -1.06979573e+00 -1.06214643e+00 -4.68094498e-02
8.60044062e-02 -1.33275187e+00 -6.62729517e-03 6.60061538e-02
-8.45528305e-01 -1.39440811e+00 -8.80617678e-01 -6.56508088e-01
2.31528431e-01 3.78034085e-01 7.06898451e-01 5.00993952e-02
-3.62430722e-01 8.41238260e-01 -5.68980932e-01 -4.25554246e-01
-3.14282184e-03 -3.04052263e-01 3.89378369e-01 -3.06135211e-02
8.78207743e-01 -4.86032516e-01 -5.94112754e-01 4.18703675e-01
-7.90788949e-01 -2.33075634e-01 6.66334331e-01 7.26072073e-01
2.00431690e-01 2.77903765e-01 3.72919649e-01 -1.13499068e-01
3.11279953e-01 -4.35465544e-01 -2.29474097e-01 2.58493219e-02
3.07151247e-02 -1.99302584e-01 3.14995736e-01 -5.33813536e-01
-8.59987676e-01 3.71093526e-02 -1.76949263e-01 -1.06218660e+00
-9.14847255e-01 2.14928091e-01 9.18977931e-02 -2.09240679e-04
6.20738924e-01 3.11198086e-01 3.61916095e-01 -3.77570659e-01
-8.40391889e-02 7.50950515e-01 5.15189171e-01 -1.63603231e-01
3.02002460e-01 5.88553369e-01 2.87273139e-01 -9.93523777e-01
-8.07748735e-01 -3.54657084e-01 -1.19715202e+00 -9.86312389e-01
1.48200881e+00 -9.77263749e-01 -1.14620864e+00 1.08686101e+00
-1.16730607e+00 -1.50210842e-01 -5.38950181e-03 9.50239778e-01
-7.81284928e-01 3.35573316e-01 -5.97444952e-01 -9.76775587e-01
1.02458388e-01 -1.09693766e+00 9.90248919e-01 1.74649745e-01
-9.36808065e-02 -1.16173577e+00 -1.66350141e-01 5.58172047e-01
4.24807757e-01 7.75053263e-01 5.15357018e-01 -7.92778730e-01
-5.04569948e-01 -4.88924801e-01 -2.07053408e-01 6.52983725e-01
-1.19105130e-01 2.25576689e-03 -9.74304736e-01 2.64140993e-01
1.10396817e-01 -2.04330996e-01 7.10601509e-01 6.53715730e-01
1.20269322e+00 9.73521322e-02 -3.06277782e-01 5.27468562e-01
1.09113312e+00 6.54048264e-01 8.58349919e-01 5.87958217e-01
4.65993136e-01 4.77357298e-01 5.85932612e-01 5.01199603e-01
7.40300044e-02 6.08400881e-01 6.03844583e-01 -1.51239648e-01
-3.11373353e-01 1.83724254e-01 5.10271430e-01 6.38720155e-01
-5.64216793e-01 -2.48025969e-01 -8.54439318e-01 3.12826604e-01
-1.84163165e+00 -1.50547302e+00 -2.57054269e-01 1.40436065e+00
2.38885097e-02 3.30962628e-01 2.55403638e-01 5.62221766e-01
6.17465973e-01 2.02803686e-01 -3.80457938e-01 -2.58302003e-01
-3.41342688e-01 8.22864175e-02 3.38326603e-01 -6.12212997e-03
-1.79486942e+00 5.92388332e-01 6.25544357e+00 7.85296142e-01
-1.17155302e+00 -2.97318213e-02 4.47290003e-01 -2.11725712e-01
7.91855454e-01 -8.54041100e-01 -8.77171457e-01 3.40606928e-01
9.36678231e-01 3.59449863e-01 -1.16784014e-01 7.33732760e-01
6.11612439e-01 -1.09915622e-01 -1.07869232e+00 1.52867758e+00
4.34023291e-01 -1.59429729e+00 1.73130766e-01 5.46862707e-02
4.82585520e-01 -2.91251659e-01 1.00023806e-01 4.75184739e-01
-3.09838891e-01 -1.12355375e+00 6.34999335e-01 9.86812592e-01
4.56319660e-01 -6.76795065e-01 1.21479774e+00 2.80915886e-01
-1.04402268e+00 -5.11497676e-01 -1.76548138e-01 -3.86110544e-01
2.61506081e-01 9.88785774e-02 -4.71101880e-01 6.05799496e-01
1.14587808e+00 1.33804595e+00 -4.68743622e-01 9.11804676e-01
1.55646086e-01 2.13043362e-01 3.36868018e-02 -3.62592578e-01
9.45632219e-01 1.33858129e-01 5.19076288e-01 1.35252273e+00
5.38034558e-01 4.77862448e-01 -1.03850618e-01 4.96408820e-01
1.51434973e-01 -2.02781051e-01 -1.07918024e+00 -1.64968029e-01
-2.13317797e-01 9.41679239e-01 -4.15745676e-01 -5.63517153e-01
-5.78732193e-01 9.13993835e-01 -4.03333530e-02 1.95685610e-01
-1.15993047e+00 -1.49464890e-01 8.63878012e-01 2.96758432e-02
4.72132653e-01 -5.32863677e-01 9.91953462e-02 -9.42500770e-01
-9.33052152e-02 -5.63402236e-01 4.92311478e-01 -1.14962828e+00
-1.23933423e+00 5.76551616e-01 2.72896200e-01 -1.65548182e+00
4.96732420e-04 -1.35297620e+00 -3.96734208e-01 3.27551544e-01
-1.26746917e+00 -1.06765759e+00 -6.24899864e-01 1.05069852e+00
6.54315412e-01 -7.51532853e-01 6.45389080e-01 6.46258533e-01
-7.56871402e-01 3.45725626e-01 -2.11723775e-01 5.10088563e-01
4.39623773e-01 -8.89950693e-01 3.54899578e-02 6.38269246e-01
-2.13696450e-01 1.36026382e-01 1.52862862e-01 -4.16761041e-01
-1.59514391e+00 -1.06093144e+00 4.62898701e-01 -5.99883020e-01
3.01050723e-01 -5.92040606e-02 -4.93019640e-01 4.86249030e-01
2.28445500e-01 9.89990607e-02 8.19547534e-01 -4.39038396e-01
-1.15657628e-01 2.20133048e-02 -1.28955007e+00 4.06591296e-01
1.42262292e+00 -3.16820920e-01 -7.27376521e-01 2.32392415e-01
4.54436719e-01 -3.54330271e-01 -1.15369928e+00 6.01636291e-01
8.69479120e-01 -1.43330383e+00 1.20640635e+00 -9.63381052e-01
5.79280734e-01 -3.62678558e-01 -5.25409937e-01 -7.72731721e-01
-2.82633513e-01 -8.85772482e-02 -5.80777407e-01 4.38241452e-01
-1.54553577e-01 -2.61411160e-01 8.52304399e-01 8.10994431e-02
-4.40227181e-01 -9.10787106e-01 -1.00920427e+00 -9.03115630e-01
-2.97002703e-01 -6.46147549e-01 3.61075848e-01 1.00327194e+00
-2.52190500e-01 -9.93812159e-02 -5.64900577e-01 7.31561473e-03
4.42214191e-01 -4.70315218e-01 9.11956668e-01 -1.03039396e+00
1.76939890e-01 -6.61180675e-01 -1.39565158e+00 -1.17629385e+00
9.93542373e-02 -5.98959804e-01 -7.25499332e-01 -1.37680018e+00
2.44583581e-02 1.81488425e-01 -3.82889301e-01 3.55825514e-01
3.96729529e-01 4.13144648e-01 1.39181659e-01 -1.93108812e-01
-7.70349085e-01 5.28306127e-01 1.60741091e+00 -3.43757182e-01
1.10291466e-01 -1.94660857e-01 3.05622779e-02 7.54899085e-01
8.22744906e-01 1.37389854e-01 -3.48642409e-01 -3.79062682e-01
-5.38317740e-01 7.10319877e-02 7.99714804e-01 -1.66223764e+00
2.73609012e-01 2.14784101e-01 1.12253392e+00 -1.24192035e+00
6.02564454e-01 -1.05201733e+00 1.67005569e-01 6.40376031e-01
-2.56343365e-01 5.92858791e-02 3.39940757e-01 3.81851703e-01
-5.84252477e-01 1.33987218e-01 6.80051684e-01 -4.80433762e-01
-1.23420727e+00 5.95116258e-01 -6.08383179e-01 -3.84182811e-01
1.48279977e+00 -9.49825883e-01 -1.49343997e-01 -2.70144105e-01
-1.20632565e+00 -8.89731571e-02 -8.93553123e-02 6.78624988e-01
7.74486184e-01 -1.93015695e+00 -3.97058457e-01 4.48639154e-01
-1.16412379e-01 -4.41525459e-01 8.08105469e-01 1.07108521e+00
-7.69298971e-01 9.25570190e-01 -1.02361548e+00 -7.96321571e-01
-1.27996242e+00 7.18290806e-01 8.78043234e-01 -6.89016804e-02
-5.77441871e-01 5.50700426e-01 -1.38556644e-01 -3.05282980e-01
6.85974360e-01 -3.90698999e-01 -6.83957279e-01 6.73945993e-02
6.04256570e-01 6.47964895e-01 -1.11285768e-01 -8.65336001e-01
-4.47457552e-01 7.05782652e-01 3.33946347e-01 3.25102806e-01
1.48743773e+00 2.76317578e-02 2.11088866e-01 3.68509650e-01
1.72080445e+00 -1.00226772e+00 -1.23554814e+00 8.33329856e-02
-3.28813583e-01 -5.18100679e-01 -2.39200741e-02 -6.79248154e-01
-1.23662686e+00 1.17696464e+00 1.17311144e+00 2.51526415e-01
1.05249059e+00 -3.51454467e-01 8.01619887e-01 6.12323999e-01
1.07482888e-01 -1.14963114e+00 4.62253153e-01 7.24335968e-01
8.53152990e-01 -1.26193309e+00 -9.10750255e-02 1.04835793e-01
-4.37597841e-01 1.61660051e+00 8.96546960e-01 -3.21071476e-01
1.18074489e+00 6.14244305e-02 2.07312647e-02 -5.98887503e-01
-4.36541826e-01 -2.04500422e-01 3.30341876e-01 8.69540274e-01
2.13213876e-01 -2.58292675e-01 2.13288907e-02 6.64055228e-01
3.25569689e-01 3.47029686e-01 2.12303981e-01 1.16267312e+00
-3.41468304e-01 -3.68836850e-01 -3.43190134e-01 2.09146798e-01
-4.73718107e-01 4.56294566e-01 -2.07195073e-01 1.29978991e+00
6.39734864e-01 6.23880744e-01 2.18260452e-01 -6.77963734e-01
6.72322154e-01 1.34478822e-01 4.80249047e-01 2.02195510e-01
-5.76468110e-01 -2.86431462e-01 1.15737721e-01 -9.86469865e-01
-1.29930568e+00 -5.20813823e-01 -8.81047606e-01 -4.78497863e-01
1.63052425e-01 -4.57140833e-01 5.52931249e-01 8.62899601e-01
4.33226258e-01 6.89157903e-01 3.27897280e-01 -1.22724235e+00
-1.49601296e-01 -8.74847054e-01 -6.26349092e-01 4.70649481e-01
3.22116286e-01 -1.33005404e+00 -1.56631187e-01 2.15307847e-01] | [7.9291276931762695, 0.4787179231643677] |
bffc2ad4-4a57-4244-a291-585b42a70994 | refine-re-randomization-before-fine-tuning | 2205.05282 | null | https://arxiv.org/abs/2205.05282v3 | https://arxiv.org/pdf/2205.05282v3.pdf | ReFine: Re-randomization before Fine-tuning for Cross-domain Few-shot Learning | Cross-domain few-shot learning (CD-FSL), where there are few target samples under extreme differences between source and target domains, has recently attracted huge attention. Recent studies on CD-FSL generally focus on transfer learning based approaches, where a neural network is pre-trained on popular labeled source domain datasets and then transferred to target domain data. Although the labeled datasets may provide suitable initial parameters for the target data, the domain difference between the source and target might hinder fine-tuning on the target domain. This paper proposes a simple yet powerful method that re-randomizes the parameters fitted on the source domain before adapting to the target data. The re-randomization resets source-specific parameters of the source pre-trained model and thus facilitates fine-tuning on the target domain, improving few-shot performance. | ['Se-Young Yun', 'Hwanjun Song', 'Jin-Hwa Kim', 'Namgyu Ho', 'Sungnyun Kim', 'Jaehoon Oh'] | 2022-05-11 | null | null | null | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 3.83578271e-01 -2.87904650e-01 -5.32169104e-01 -5.49596071e-01
-9.28189397e-01 -4.81069356e-01 6.01891577e-01 -4.83849794e-02
-5.08602738e-01 8.71585846e-01 1.93118319e-01 3.51553440e-01
-4.84042950e-02 -9.61153030e-01 -4.51434523e-01 -6.29867673e-01
3.06834430e-01 7.29162872e-01 7.40249395e-01 -3.05326939e-01
6.61499947e-02 1.10175952e-01 -1.42697668e+00 2.81999499e-01
8.97842884e-01 8.32291842e-01 5.32043397e-01 3.41834396e-01
-4.19491291e-01 4.37983811e-01 -8.63892853e-01 -1.07326955e-01
2.54073650e-01 -7.25401282e-01 -5.17499089e-01 -5.96871860e-02
-6.53498769e-02 -6.20611198e-02 -2.88607746e-01 1.17744768e+00
6.88042045e-01 6.92414999e-01 9.83894587e-01 -1.20375419e+00
-1.01732326e+00 4.67794001e-01 -4.68913913e-01 2.73440838e-01
-1.11840323e-01 3.63200396e-01 5.92378497e-01 -7.97839046e-01
6.90853357e-01 1.11357725e+00 6.40733659e-01 9.40961897e-01
-1.22703016e+00 -1.00759447e+00 -1.41535118e-01 5.20445406e-01
-1.37802315e+00 -5.78557849e-01 9.54214692e-01 -3.54089350e-01
6.22064590e-01 -5.15678704e-01 3.00304443e-01 1.46363544e+00
9.51035414e-03 4.43267137e-01 7.65537858e-01 -5.61094701e-01
7.78638184e-01 3.39197636e-01 -2.62785908e-02 6.66230991e-02
1.40077323e-01 2.94182092e-01 -4.36372578e-01 -2.80292302e-01
7.29471147e-01 5.15576378e-02 -8.73824805e-02 -7.48532236e-01
-8.89541328e-01 1.07742333e+00 3.07040721e-01 4.94859219e-01
-3.65043104e-01 -4.09662485e-01 8.02688301e-01 2.83114493e-01
5.75967729e-01 4.57756281e-01 -5.42203546e-01 -1.34544075e-01
-8.89288247e-01 -3.19796056e-02 4.56397235e-01 1.07292390e+00
1.06548655e+00 1.05760336e-01 -4.90887940e-01 1.36473477e+00
-7.85881579e-02 3.92736197e-01 9.79997277e-01 -4.17781442e-01
4.53433245e-01 6.05884075e-01 1.46322414e-01 -4.26207572e-01
8.57835785e-02 -2.46564355e-02 -5.87210774e-01 4.42818962e-02
4.31346148e-01 -4.28284138e-01 -1.09361732e+00 1.92385316e+00
4.25002128e-01 4.48629469e-01 3.74435186e-01 7.72350848e-01
7.04657376e-01 8.71270120e-01 3.62758428e-01 -2.17866793e-01
8.77708554e-01 -1.01458526e+00 -4.87532556e-01 -6.24965847e-01
5.65528393e-01 -5.05809128e-01 1.54889405e+00 -9.87757593e-02
-6.28499985e-01 -8.74127030e-01 -1.09857583e+00 2.20533133e-01
-6.01254106e-01 -1.53276278e-02 -7.86145497e-03 3.57864529e-01
-5.08820593e-01 6.12561047e-01 -1.98576927e-01 -6.19609177e-01
5.71477234e-01 -2.11908147e-02 -1.47130564e-01 -5.48344254e-01
-1.77231109e+00 8.53174806e-01 9.12264585e-01 -6.90157712e-01
-1.19016194e+00 -9.31039035e-01 -8.02197278e-01 1.68042257e-01
4.97245342e-01 -2.87076592e-01 1.41735125e+00 -1.13121438e+00
-1.81154978e+00 9.00709689e-01 6.26226813e-02 -2.25590751e-01
3.88958693e-01 9.17712674e-02 -8.21970761e-01 -7.40723312e-02
3.44960839e-01 6.63653433e-01 1.24336135e+00 -8.95797253e-01
-6.47850633e-01 -6.50496483e-02 -2.64525145e-01 2.53625005e-01
-5.30687749e-01 -2.51207769e-01 -4.46054429e-01 -6.27392590e-01
-5.55906475e-01 -5.46049178e-01 -7.51080066e-02 -1.97550744e-01
-2.74645863e-03 -4.45544302e-01 8.26727927e-01 -5.57892919e-02
1.19634366e+00 -2.21397495e+00 -2.07088217e-01 -1.92730933e-01
-3.21153015e-01 8.17467213e-01 -6.37698293e-01 3.48507404e-01
-3.05661947e-01 -3.68319988e-01 -1.75483614e-01 1.69480711e-01
-1.36848181e-01 1.26671642e-01 -3.47633809e-01 3.90158832e-01
2.19271839e-01 8.44141066e-01 -1.16720998e+00 -3.90426904e-01
2.20089480e-01 1.71012312e-01 -3.44065964e-01 5.74181080e-01
-3.58955979e-01 2.62626320e-01 -3.65113586e-01 2.54879892e-01
6.72396541e-01 -2.90625870e-01 -7.17758853e-03 -1.28567949e-01
2.81053841e-01 2.08964702e-02 -1.10515606e+00 1.71572399e+00
-5.11971772e-01 3.92969459e-01 -3.22748244e-01 -1.15997660e+00
1.39111805e+00 1.65531725e-01 3.89374197e-01 -9.07410562e-01
2.51204431e-01 7.37777585e-03 -7.85202011e-02 -4.29697931e-01
2.89761066e-01 -7.48609722e-01 -2.75165021e-01 5.39794385e-01
3.34092617e-01 -1.07695594e-01 1.33429840e-01 -1.64904132e-01
7.54609406e-01 2.37229422e-01 7.81256199e-01 -1.59747526e-01
4.97696251e-01 2.95531988e-01 7.55147219e-01 6.06394529e-01
-5.83838284e-01 6.51393533e-01 1.79624036e-01 -2.00868636e-01
-1.34897804e+00 -1.17282784e+00 8.09382200e-02 1.62312770e+00
2.51856983e-01 -1.05840832e-01 -7.52175152e-01 -9.49065447e-01
-1.01776235e-01 9.83908117e-01 -8.81269276e-01 -8.76090407e-01
-3.25248182e-01 -3.79118353e-01 4.67632741e-01 4.60296124e-01
5.16526341e-01 -1.19362533e+00 -5.14094353e-01 4.20709670e-01
-1.00212507e-01 -7.38407075e-01 -7.01339602e-01 3.98933351e-01
-8.04526150e-01 -9.73032117e-01 -1.17224228e+00 -8.73348951e-01
5.43496311e-01 3.97485822e-01 9.92088556e-01 -7.07482517e-01
-5.96538708e-02 2.38014515e-02 -4.91062999e-01 -4.87993628e-01
-5.58555663e-01 2.46928513e-01 1.95897400e-01 -3.43378596e-02
9.86887097e-01 -4.19880748e-01 -1.60371810e-01 4.61303264e-01
-9.20654178e-01 -2.70415753e-01 4.55045491e-01 9.88243580e-01
5.15645742e-01 1.37888953e-01 1.01866531e+00 -1.21672666e+00
8.16468775e-01 -1.00465763e+00 -4.42741811e-01 5.25841892e-01
-4.71448660e-01 6.55669570e-02 1.03133094e+00 -1.12302685e+00
-1.36677480e+00 -9.62393954e-02 3.18589568e-01 -9.46327269e-01
-3.94878268e-01 2.76865512e-01 -3.59699905e-01 3.70749891e-01
1.38512111e+00 2.88898259e-01 -1.11013278e-01 -4.97973323e-01
4.09104079e-01 8.05654764e-01 3.51796031e-01 -5.86443901e-01
8.31468046e-01 1.66714877e-01 -7.02035248e-01 -6.33674979e-01
-1.05952215e+00 -5.04804015e-01 -8.17185163e-01 7.39925727e-02
6.06651425e-01 -8.17183435e-01 1.83789894e-01 4.88714069e-01
-9.08652961e-01 -7.80249298e-01 -8.37367177e-01 3.15899491e-01
-6.06777906e-01 4.82554361e-02 -1.54976100e-01 -4.50862586e-01
-1.24205880e-01 -7.06119657e-01 6.81250751e-01 5.71381271e-01
-3.24130654e-01 -1.25820005e+00 5.62866032e-01 -2.01501280e-01
5.73141038e-01 -2.61000574e-01 9.38984931e-01 -1.14636779e+00
2.25283414e-01 -1.96089119e-01 -3.40646386e-01 3.34697187e-01
6.22394979e-01 -3.96118611e-01 -9.85482752e-01 -3.20521742e-01
1.10212542e-01 -6.71428680e-01 5.16315877e-01 3.88729304e-01
7.44825840e-01 -1.43588213e-02 -3.41643870e-01 3.72028500e-01
1.45179069e+00 3.70394647e-01 5.18315852e-01 4.61230248e-01
2.12728485e-01 5.04605174e-01 1.11466646e+00 5.88752091e-01
3.25888731e-02 4.78826702e-01 -2.52675444e-01 1.49252832e-01
-2.52237648e-01 -4.89789397e-01 4.85070020e-01 4.60351646e-01
5.31267762e-01 -1.07284777e-01 -8.51212084e-01 8.48956347e-01
-1.73108160e+00 -1.04066837e+00 7.66210914e-01 2.25969100e+00
1.18888271e+00 2.92082310e-01 2.63312250e-01 -3.75120044e-01
1.36989713e+00 1.89847589e-01 -1.21884286e+00 -1.80138439e-01
1.03936695e-01 2.96477765e-01 3.59868586e-01 2.69114614e-01
-1.04637885e+00 1.20042241e+00 6.64117765e+00 1.28696406e+00
-1.42894256e+00 3.52442175e-01 3.27461362e-01 -2.15394527e-01
-7.33202789e-03 -2.05269635e-01 -1.01046932e+00 6.68192983e-01
1.15429652e+00 -9.97977078e-01 3.53500396e-01 1.35038769e+00
1.08122595e-01 3.02567799e-02 -1.06881857e+00 9.49212730e-01
6.79921955e-02 -1.18944740e+00 9.46246758e-02 -2.91960657e-01
9.70095694e-01 6.18408956e-02 -8.10826048e-02 1.03167057e+00
6.44629955e-01 -6.64127111e-01 2.74129391e-01 2.86114752e-01
1.26347923e+00 -8.63174081e-01 4.28936809e-01 5.88805318e-01
-1.08948791e+00 -1.56731367e-01 -1.05695701e+00 1.03173658e-01
3.93460579e-02 4.00802523e-01 -1.10607302e+00 2.47792661e-01
5.63284576e-01 7.99145937e-01 -4.42623943e-01 8.90758216e-01
-5.50977178e-02 6.28775477e-01 1.83140766e-02 -3.93592380e-03
1.65042296e-01 -1.34575948e-01 3.82223696e-01 1.10696185e+00
4.37651068e-01 1.58604812e-02 2.50544846e-01 8.63614500e-01
6.00537322e-02 7.88405836e-02 -7.31052518e-01 -1.99228749e-01
7.52378821e-01 1.02094972e+00 -4.78857577e-01 -7.53182232e-01
-3.06950659e-01 8.53689671e-01 4.70061809e-01 4.92457598e-01
-1.05081201e+00 -7.22461641e-01 6.53272152e-01 2.59030499e-02
5.62913477e-01 3.18887413e-01 -1.96481079e-01 -1.24822927e+00
-4.45971817e-01 -7.71772802e-01 7.00110972e-01 -6.21849358e-01
-1.94343865e+00 5.22084832e-01 1.59607530e-01 -1.70915723e+00
-4.34658021e-01 -2.34189838e-01 -8.73898804e-01 1.03314483e+00
-1.48753190e+00 -8.59451413e-01 -1.81090206e-01 9.83590484e-01
9.96370494e-01 -5.38534880e-01 8.90548348e-01 2.69127995e-01
-4.65724468e-01 7.69104838e-01 5.41729867e-01 7.18118101e-02
1.28798640e+00 -7.81523168e-01 3.21068227e-01 6.79232299e-01
-2.92901576e-01 4.56020683e-01 6.90082252e-01 -8.29131544e-01
-6.62424505e-01 -1.41876161e+00 5.68922937e-01 -8.04674104e-02
6.72727585e-01 -2.33701557e-01 -1.43963039e+00 4.53604370e-01
7.60499910e-02 1.50053173e-01 9.03049648e-01 -2.55331174e-02
-5.15245616e-01 -1.91527426e-01 -1.34157205e+00 5.45807600e-01
7.88982451e-01 -5.90762913e-01 -9.85604823e-01 1.20980740e-01
5.68979800e-01 -2.53629535e-02 -6.18813992e-01 1.04690015e-01
7.20664412e-02 -7.42140830e-01 8.93999279e-01 -9.05820549e-01
3.61236781e-01 -7.26853535e-02 -1.44316867e-01 -1.94140255e+00
-7.08219945e-01 -1.89158112e-01 3.55222374e-02 1.44131732e+00
9.55601931e-02 -3.94008070e-01 9.69177246e-01 5.46483636e-01
1.23131208e-01 -2.57111192e-01 -8.15326631e-01 -1.06866276e+00
4.05792862e-01 -1.26466870e-01 6.41992152e-01 1.07350743e+00
1.74067579e-02 7.59683132e-01 -4.01256859e-01 -4.11135316e-01
5.72327137e-01 1.50546238e-01 7.78254807e-01 -1.37439060e+00
-2.08620355e-01 -2.53813654e-01 1.27968773e-01 -8.00039887e-01
2.99176961e-01 -8.17127645e-01 4.12585676e-01 -1.14293802e+00
1.60411417e-01 -5.29744029e-01 -5.08711874e-01 5.33333361e-01
-4.12581027e-01 5.51994070e-02 2.01688986e-02 2.70313382e-01
-7.08425760e-01 8.50799680e-01 1.21261036e+00 -1.65566310e-01
-5.01277208e-01 5.80553859e-02 -6.48234189e-01 5.56486249e-01
8.33243430e-01 -6.32255673e-01 -8.53774846e-01 -6.27773115e-03
-4.48964089e-01 -1.82247505e-01 -2.24730387e-01 -1.20078468e+00
2.42284313e-01 -6.93304539e-01 5.79513192e-01 -2.33154476e-01
2.60937840e-01 -7.21788585e-01 -7.90918842e-02 3.56226861e-01
-5.01703680e-01 -5.41681349e-01 2.93475300e-01 7.71906972e-01
-2.07980126e-01 -7.01909602e-01 1.52441239e+00 -3.39359105e-01
-1.37490356e+00 3.92784387e-01 -1.90250069e-01 6.48147464e-01
1.40079367e+00 -4.38419998e-01 -2.45654345e-01 -2.14836851e-01
-6.71816170e-01 3.62191275e-02 6.59275830e-01 6.47996604e-01
4.69993532e-01 -1.70213139e+00 -5.79285562e-01 2.55520344e-01
5.89988410e-01 -1.39982462e-01 5.77708125e-01 9.51214507e-02
1.93137169e-01 1.19846091e-01 -6.86722517e-01 -3.87306720e-01
-8.85456860e-01 1.15216362e+00 2.28923082e-01 -2.59273589e-01
-4.43885893e-01 8.32622051e-01 3.77087653e-01 -5.50435305e-01
8.27502012e-02 3.11452627e-01 -2.46636346e-01 4.38645422e-01
7.31330812e-01 4.24905658e-01 -2.89964080e-01 -5.19591510e-01
-2.66585708e-01 4.68139738e-01 -2.60620207e-01 -7.47680143e-02
1.40200686e+00 -2.32790157e-01 5.71003020e-01 8.12116802e-01
1.37778056e+00 -4.59101498e-01 -1.61543882e+00 -9.38936234e-01
5.58440275e-02 -3.87586772e-01 -2.24840477e-01 -7.82884777e-01
-6.34961069e-01 1.03273499e+00 6.05969071e-01 -1.39570534e-01
1.15788400e+00 3.08026951e-02 8.49125087e-01 4.09323603e-01
2.87268251e-01 -1.64935088e+00 5.42235613e-01 8.14673483e-01
5.02624989e-01 -1.31685662e+00 -3.39753270e-01 4.45717573e-02
-1.00770366e+00 1.06452131e+00 9.90614176e-01 -3.81285757e-01
5.41456461e-01 -1.24398030e-01 7.64544681e-02 1.75986767e-01
-5.60585797e-01 -2.41435528e-01 2.46112064e-01 1.16348207e+00
2.72863209e-02 -1.55206561e-01 2.66185671e-01 8.89858842e-01
1.19853035e-01 4.57648277e-01 2.73863137e-01 7.39072859e-01
-8.22952807e-01 -1.16370356e+00 -3.61807793e-01 4.29128826e-01
4.11267094e-02 3.09469141e-02 -2.50930786e-01 5.83464384e-01
2.77171463e-01 6.12337232e-01 1.17982097e-01 -2.90627569e-01
4.35621619e-01 4.62664217e-01 2.11423382e-01 -1.15011013e+00
-2.96222329e-01 2.81903930e-02 -3.03977787e-01 -1.95552528e-01
-1.71132058e-01 -4.48570669e-01 -1.08943403e+00 -1.27213389e-01
-2.77457952e-01 3.79151762e-01 5.59654534e-02 8.27770293e-01
5.22126317e-01 4.28148478e-01 6.80209160e-01 -9.17029917e-01
-8.23361695e-01 -1.16570044e+00 -7.66416013e-01 6.13180995e-01
1.44215196e-01 -9.41588879e-01 -2.34719276e-01 1.72186747e-01] | [10.107619285583496, 3.070810556411743] |
b4250b34-ef4b-4f55-a6a2-15e3a8921daf | enabling-multimodal-generation-on-clip-via-1 | 2203.06386 | null | https://arxiv.org/abs/2203.06386v2 | https://arxiv.org/pdf/2203.06386v2.pdf | Enabling Multimodal Generation on CLIP via Vision-Language Knowledge Distillation | The recent large-scale vision-language pre-training (VLP) of dual-stream architectures (e.g., CLIP) with a tremendous amount of image-text pair data, has shown its superiority on various multimodal alignment tasks. Despite its success, the resulting models are not capable of multimodal generative tasks due to the weak text encoder. To tackle this problem, we propose to augment the dual-stream VLP model with a textual pre-trained language model (PLM) via vision-language knowledge distillation (VLKD), enabling the capability for multimodal generation. VLKD is pretty data- and computation-efficient compared to the pre-training from scratch. Experimental results show that the resulting model has strong zero-shot performance on multimodal generation tasks, such as open-ended visual question answering and image captioning. For example, it achieves 44.5% zero-shot accuracy on the VQAv2 dataset, surpassing the previous state-of-the-art zero-shot model with $7\times$ fewer parameters. Furthermore, the original textual language understanding and generation ability of the PLM is maintained after VLKD, which makes our model versatile for both multimodal and unimodal tasks. | ['Pascale Fung', 'Qun Liu', 'Xin Jiang', 'Lifeng Shang', 'Lu Hou', 'Wenliang Dai'] | 2022-03-12 | null | https://aclanthology.org/2022.findings-acl.187 | https://aclanthology.org/2022.findings-acl.187.pdf | findings-acl-2022-5 | ['multimodal-generation'] | ['natural-language-processing'] | [ 1.58357844e-01 2.70300448e-01 -3.22211236e-02 -1.54201686e-01
-1.12862837e+00 -3.17910433e-01 1.08983207e+00 -1.66890383e-01
-3.47006470e-01 6.74865961e-01 3.43867660e-01 -2.83320397e-01
6.22139156e-01 -7.46342540e-01 -1.06524837e+00 -6.34042442e-01
5.68542540e-01 6.38560772e-01 -2.07525957e-02 -4.29986835e-01
-1.92445979e-01 -3.36352229e-01 -1.65252221e+00 6.97625160e-01
9.96673346e-01 9.08407629e-01 5.53284585e-01 8.24865639e-01
-4.73923653e-01 8.70736122e-01 -3.31893504e-01 -1.03262663e+00
-1.83539882e-01 -7.05153286e-01 -6.12356544e-01 8.19647461e-02
7.22988963e-01 -6.37746990e-01 -6.32820785e-01 6.63092613e-01
7.52737701e-01 -1.01004064e-01 6.74898386e-01 -1.54250741e+00
-1.29250002e+00 5.74009001e-01 -4.81948704e-01 -4.68329430e-01
5.01427710e-01 4.99035329e-01 1.15001357e+00 -1.31250441e+00
6.87101245e-01 1.41885424e+00 2.42720068e-01 8.75937879e-01
-1.13462579e+00 -4.37253326e-01 -2.29993656e-01 1.98477238e-01
-1.32116187e+00 -5.59086382e-01 5.10855258e-01 -3.49951327e-01
9.97446120e-01 3.48197296e-02 4.57712114e-01 1.67048216e+00
2.20343824e-02 1.24333012e+00 8.61579776e-01 -3.66277754e-01
1.24919824e-02 2.06135631e-01 -3.25610846e-01 8.50214660e-01
-1.14265539e-01 -2.59596318e-01 -7.84397662e-01 1.79867800e-02
7.53461301e-01 -1.31661490e-01 -2.78884470e-01 -4.28795367e-01
-1.44157898e+00 9.16554689e-01 3.87333125e-01 5.50793558e-02
-2.30932683e-01 3.75843495e-01 3.95915359e-01 1.49313912e-01
2.36363113e-02 2.47913644e-01 1.36726230e-01 -8.88981298e-02
-8.13822091e-01 2.12858900e-01 5.07489502e-01 1.28289390e+00
7.21667171e-01 3.37104380e-01 -8.45537603e-01 8.61118197e-01
4.45598751e-01 9.92106259e-01 5.15131116e-01 -7.94389009e-01
7.84266114e-01 5.17618358e-01 -2.64092386e-02 -5.91139376e-01
1.27308071e-01 8.55755061e-02 -1.15814376e+00 -2.34106332e-01
1.58190146e-01 -1.03091687e-01 -1.30055809e+00 1.73956692e+00
-5.74933439e-02 5.53564616e-02 6.47804797e-01 8.51680338e-01
1.47429490e+00 1.26085711e+00 2.09970087e-01 2.04310361e-02
1.47855926e+00 -1.28817594e+00 -7.56115735e-01 -2.98176676e-01
3.25340241e-01 -7.47872412e-01 1.44139624e+00 2.21426934e-02
-1.21027207e+00 -9.05372381e-01 -8.10763419e-01 -3.18783134e-01
-2.07881153e-01 1.34309784e-01 6.52709484e-01 3.97876889e-01
-1.19241667e+00 -1.61963508e-01 -4.32356298e-01 -4.12114471e-01
3.71187538e-01 3.04016825e-02 -4.51333404e-01 -4.61322099e-01
-1.29908335e+00 6.36654198e-01 4.96961385e-01 -9.47347134e-02
-1.36119699e+00 -6.72017753e-01 -1.24698949e+00 6.99011832e-02
3.08936328e-01 -1.21502101e+00 1.21969593e+00 -8.64288926e-01
-1.60076237e+00 8.06751847e-01 -3.48702729e-01 -5.00207424e-01
4.68633056e-01 -1.25816911e-01 -1.65860325e-01 4.23924327e-01
6.39541596e-02 1.40765631e+00 1.23122525e+00 -1.52844441e+00
-2.94844538e-01 1.22954035e-02 1.90517362e-02 4.20973152e-01
-5.07205248e-01 -3.08558345e-01 -9.38580155e-01 -4.39270139e-01
-5.16173720e-01 -7.86469579e-01 1.65291242e-02 -1.43422723e-01
-2.33086541e-01 -2.45243728e-01 6.80231690e-01 -7.75533497e-01
9.73318696e-01 -2.05187106e+00 2.44780242e-01 -3.56983006e-01
1.19407222e-01 4.51986164e-01 -7.03118682e-01 8.96075070e-01
1.73088804e-01 -1.29476100e-01 -3.20740849e-01 -7.44368494e-01
2.33960584e-01 3.44251364e-01 -7.08605528e-01 -1.52747214e-01
4.69491780e-01 1.73687458e+00 -8.37808788e-01 -7.70875514e-01
2.85651743e-01 5.43036103e-01 -4.66379285e-01 5.74783742e-01
-6.25176489e-01 2.18212888e-01 -5.83513305e-02 7.33197451e-01
4.90615606e-01 -5.38673699e-01 -2.64309272e-02 -1.30058751e-01
2.05563501e-01 -3.57954293e-01 -4.82647419e-01 2.33672190e+00
-5.05484879e-01 6.30214870e-01 -2.16903880e-01 -6.64429784e-01
9.09326792e-01 6.10571146e-01 1.92033485e-01 -1.12007535e+00
4.63498645e-02 2.81369425e-02 -3.39733034e-01 -6.42497301e-01
7.06570089e-01 -6.74311966e-02 -3.36880684e-01 2.35558033e-01
6.99588478e-01 -3.65875572e-01 3.41716081e-01 7.93226421e-01
7.17945576e-01 3.48466069e-01 -6.54974207e-02 3.51008564e-01
4.95183647e-01 1.37738595e-02 -1.07544295e-01 7.67256200e-01
1.56552896e-01 8.49755049e-01 3.62899244e-01 1.34503737e-01
-1.30445480e+00 -1.38948035e+00 2.46456385e-01 1.08805621e+00
2.15603158e-01 -5.27175486e-01 -7.38949239e-01 -3.66005450e-01
-1.56953678e-01 9.10816073e-01 -3.77765179e-01 -1.78925157e-01
-2.15611845e-01 -5.15513361e-01 8.45910788e-01 4.97099727e-01
6.95963442e-01 -1.02994430e+00 -3.21933895e-01 8.09479952e-02
-6.22367740e-01 -1.49372768e+00 -3.96095246e-01 -4.72114027e-01
-4.76051301e-01 -5.80613434e-01 -1.29731023e+00 -9.90441561e-01
5.22157490e-01 3.41247618e-01 1.27172613e+00 -3.14907581e-01
-2.09250525e-01 8.10366690e-01 -5.12340188e-01 -2.39130914e-01
-5.42669654e-01 -2.69843549e-01 -3.46940130e-01 2.45875567e-01
1.19201034e-01 -2.25004345e-01 -5.36666572e-01 3.68864136e-03
-1.07549393e+00 5.98094344e-01 9.37136531e-01 1.27897906e+00
5.43189347e-01 -5.99894345e-01 7.62571633e-01 -2.44463012e-01
8.31035614e-01 -3.43412727e-01 -2.53711343e-01 7.03236878e-01
-2.59943962e-01 7.00477064e-02 4.94025767e-01 -4.88311231e-01
-1.27083492e+00 -8.79942551e-02 -1.67922199e-01 -7.25590229e-01
-6.09482415e-02 4.17888403e-01 -1.66491136e-01 2.89861917e-01
4.29396480e-01 7.36541748e-01 2.61060953e-01 -1.50409579e-01
1.00456178e+00 6.59591496e-01 9.23732162e-01 -4.90769863e-01
6.61130548e-01 2.57086396e-01 -2.62692511e-01 -8.92308831e-01
-5.02564132e-01 -2.79789925e-01 -2.94902414e-01 -2.22819567e-01
1.36607850e+00 -1.46214449e+00 -6.76365495e-01 5.89741230e-01
-1.46751618e+00 -3.37074757e-01 -9.95310321e-02 1.79489046e-01
-6.84149683e-01 5.93389094e-01 -5.60037017e-01 -8.76923442e-01
-6.89512074e-01 -1.12053668e+00 1.46017849e+00 2.47802734e-01
1.18132442e-01 -8.94935012e-01 -6.73276093e-03 8.46224487e-01
3.17946851e-01 6.17078133e-02 1.00234032e+00 -2.24526644e-01
-7.65816689e-01 4.52565067e-02 -4.98604953e-01 3.63114774e-01
-3.54886204e-01 -2.49446437e-01 -1.05626571e+00 -3.39847028e-01
-4.54025477e-01 -1.08156264e+00 1.08403587e+00 1.40933646e-02
8.79033208e-01 -1.72671616e-01 -7.03003258e-02 4.59771931e-01
1.46375704e+00 -4.22150120e-02 9.23629224e-01 -1.70608371e-01
8.45838547e-01 4.18317795e-01 5.10317445e-01 3.97540808e-01
7.97736645e-01 6.22756660e-01 6.65369630e-01 -3.44764173e-01
-4.94572252e-01 -7.98872232e-01 6.53417706e-01 9.36468303e-01
8.57260674e-02 -5.73324502e-01 -9.29361880e-01 7.17883706e-01
-2.10944891e+00 -9.22985554e-01 -1.96665332e-01 1.84452498e+00
8.99482608e-01 -3.61172110e-01 -1.42152101e-01 -5.95615685e-01
4.53572601e-01 2.19558746e-01 -2.79711604e-01 -4.11619574e-01
-3.94888908e-01 1.85462460e-01 5.22262901e-02 3.29950839e-01
-8.13006103e-01 1.24555266e+00 5.91130066e+00 1.06050503e+00
-8.77628326e-01 2.55816758e-01 4.66962606e-01 3.93012613e-02
-6.58052683e-01 -1.91974029e-01 -7.66059756e-01 3.39806616e-01
8.65490258e-01 -1.40315562e-01 2.86058277e-01 6.44783199e-01
-1.91052288e-01 -1.30467162e-01 -1.05111730e+00 1.49705744e+00
6.90433145e-01 -1.45417798e+00 8.45801651e-01 -1.71468005e-01
8.55257690e-01 -9.32412595e-02 2.81848609e-01 8.11897755e-01
1.69449821e-01 -1.29784286e+00 5.84086239e-01 5.56767464e-01
1.25543749e+00 -6.46319032e-01 7.14353204e-01 4.75091130e-01
-9.76945877e-01 -5.57815060e-02 -4.03791338e-01 2.72533774e-01
6.15222812e-01 7.65580162e-02 -1.02588761e+00 7.77216315e-01
3.27430040e-01 3.18625093e-01 -6.69433594e-01 5.49424350e-01
-1.68275326e-01 3.36724132e-01 1.26860365e-01 -6.49499148e-02
5.80444455e-01 -6.93324879e-02 2.51198411e-01 1.14826190e+00
5.23174465e-01 -1.52057171e-01 -5.27248643e-02 1.07166326e+00
-3.10407043e-01 1.49954513e-01 -8.39361787e-01 -5.20861566e-01
4.53755856e-02 1.25655174e+00 -7.19787776e-02 -6.70103252e-01
-6.77784860e-01 1.46271849e+00 1.96408480e-01 5.27172685e-01
-1.14795792e+00 -2.30884686e-01 3.79649132e-01 -2.66366124e-01
3.60516638e-01 -3.34399670e-01 2.29584165e-02 -1.32090724e+00
-4.67519611e-02 -8.64285529e-01 1.88042715e-01 -1.32081020e+00
-1.29721224e+00 7.56323159e-01 -6.89539732e-03 -1.07171583e+00
-5.90962172e-01 -6.08896017e-01 -4.82286721e-01 8.26878250e-01
-1.52144134e+00 -1.85873163e+00 -5.27409673e-01 9.85829115e-01
7.89414287e-01 -5.11997461e-01 9.60950971e-01 2.16631204e-01
-1.78702727e-01 8.21634650e-01 -1.08582608e-01 1.74073398e-01
9.31489170e-01 -1.02300370e+00 4.29381013e-01 7.87984192e-01
3.46223414e-01 3.58991951e-01 3.96316737e-01 -5.49123824e-01
-1.99515283e+00 -1.09853458e+00 8.65690351e-01 -5.76870322e-01
6.92889154e-01 -5.89410543e-01 -7.69597173e-01 3.89457941e-01
1.00413501e+00 -5.43091774e-01 7.69297600e-01 -4.38151151e-01
-4.87805754e-01 2.71140486e-02 -6.24983609e-01 9.62171674e-01
8.38830650e-01 -6.52821302e-01 -7.98219502e-01 2.48762965e-01
1.14520776e+00 -1.99125767e-01 -7.28716612e-01 4.98195827e-01
4.42408890e-01 -7.53260374e-01 1.06397879e+00 -5.71945965e-01
1.11421037e+00 -8.06385055e-02 -2.28257954e-01 -1.05306625e+00
-2.69343276e-02 -6.81622863e-01 -3.42521280e-01 1.43388498e+00
4.99122232e-01 -1.67433992e-01 4.71252382e-01 4.70374554e-01
-1.61816850e-01 -5.77038050e-01 -8.51819754e-01 -5.49185932e-01
-8.97179320e-02 -3.72101486e-01 2.16406077e-01 6.50286794e-01
-3.33615728e-02 1.05815494e+00 -9.24402714e-01 -2.24841014e-01
5.77748954e-01 1.86010271e-01 1.24796760e+00 -5.15556097e-01
-4.91182208e-01 -2.91723788e-01 -1.22216962e-01 -1.42227888e+00
2.15377480e-01 -1.07376790e+00 1.18265592e-01 -1.90116239e+00
4.34962243e-01 1.18403740e-01 1.50581494e-01 5.10169029e-01
-3.43680263e-01 4.26615745e-01 5.61088502e-01 7.51219839e-02
-7.96679258e-01 1.13242364e+00 1.55277646e+00 -4.56058115e-01
-2.46293172e-02 -5.68650067e-01 -6.44793868e-01 1.89324945e-01
3.78121108e-01 2.39510670e-01 -7.67593622e-01 -8.02926481e-01
1.10636540e-01 3.87699783e-01 6.82009518e-01 -6.63566053e-01
2.47046024e-01 2.68292204e-02 2.07735762e-01 -7.60569155e-01
8.48385036e-01 -4.69226003e-01 -7.99307674e-02 2.17048183e-01
-2.61593759e-01 3.63731049e-02 2.10250303e-01 5.98386049e-01
-5.30778587e-01 -4.18877695e-03 4.24710095e-01 -1.25623316e-01
-1.14835095e+00 3.22472632e-01 -1.73051521e-01 6.63286969e-02
1.03488922e+00 2.90767830e-02 -6.18928134e-01 -7.70864964e-01
-4.02966768e-01 5.22922397e-01 3.80502969e-01 7.66723871e-01
1.12674034e+00 -1.58206940e+00 -9.71157551e-01 1.49470299e-01
5.86633325e-01 -7.44707286e-02 5.51250041e-01 8.21154177e-01
-3.33553553e-01 6.51384950e-01 -3.42010379e-01 -8.54035020e-01
-1.18229747e+00 7.85336673e-01 -1.66358739e-01 -1.69433549e-01
-5.46124876e-01 8.37221324e-01 5.07259607e-01 -1.25651598e-01
1.65388629e-01 1.33449063e-01 -4.06389050e-02 6.43195957e-02
5.58775187e-01 -2.22695228e-02 -3.07950824e-01 -6.24092281e-01
4.28423695e-02 3.73543024e-01 8.55999738e-02 -5.24451196e-01
9.50052977e-01 -1.84520856e-01 -1.09876364e-01 3.50372553e-01
1.08173060e+00 -4.10654992e-01 -1.19854760e+00 -2.43203670e-01
-5.73752284e-01 -2.55251259e-01 -3.74498695e-01 -7.47952044e-01
-6.83037877e-01 1.38109720e+00 3.33923727e-01 -2.36407712e-01
9.53517556e-01 2.45349541e-01 1.11250687e+00 5.16032934e-01
2.54764766e-01 -9.93712306e-01 7.52122641e-01 6.65005505e-01
1.15200210e+00 -1.44422245e+00 -5.01675367e-01 -5.62555008e-02
-1.38059556e+00 8.51702511e-01 8.17672670e-01 2.22754672e-01
-7.36200735e-02 -9.56078172e-02 7.83896744e-02 6.85284957e-02
-1.05460095e+00 -4.95971978e-01 4.91140634e-01 7.92908609e-01
1.55413985e-01 -4.45878915e-02 -1.29526496e-01 5.67642391e-01
-1.43800288e-01 -8.60904828e-02 2.79659122e-01 5.31512856e-01
-2.29281291e-01 -9.32101846e-01 -2.11233914e-01 2.88143549e-02
2.79568266e-02 -4.71319020e-01 -4.69327778e-01 7.25815594e-01
-4.37277593e-02 9.88864362e-01 1.24353670e-01 -3.94859046e-01
4.43466082e-02 3.42807353e-01 6.39355183e-01 -4.22153354e-01
-2.51453698e-01 8.89918432e-02 7.23069385e-02 -4.37859654e-01
-3.71303022e-01 -1.34245858e-01 -1.12090516e+00 -8.78766850e-02
2.52872519e-02 -8.72418657e-02 6.00784898e-01 7.53604710e-01
4.67857569e-01 4.22527760e-01 2.88226932e-01 -8.01429093e-01
-5.07306635e-01 -9.23740804e-01 -1.30493954e-01 6.40416384e-01
1.05269954e-01 -4.04522210e-01 2.15073764e-01 3.99682939e-01] | [10.994552612304688, 1.3175007104873657] |
56b4b5c7-b875-4b43-a07f-0e0b74974dee | fara-future-aware-ranking-algorithm-for | 2305.16637 | null | https://arxiv.org/abs/2305.16637v1 | https://arxiv.org/pdf/2305.16637v1.pdf | FARA: Future-aware Ranking Algorithm for Fairness Optimization | Ranking systems are the key components of modern Information Retrieval (IR) applications, such as search engines and recommender systems. Besides the ranking relevance to users, the exposure fairness to item providers has also been considered an important factor in ranking optimization. Many fair ranking algorithms have been proposed to jointly optimize both ranking relevance and fairness. However, we find that most existing fair ranking methods adopt greedy algorithms that only optimize rankings for the next immediate session or request. As shown in this paper, such a myopic paradigm could limit the upper bound of ranking optimization and lead to suboptimal performance in the long term. To this end, we propose FARA, a novel Future-Aware Ranking Algorithm for ranking relevance and fairness optimization. Instead of greedily optimizing rankings for the next immediate session, FARA plans ahead by jointly optimizing multiple ranklists together and saving them for future sessions. Particularly, FARA first uses the Taylor expansion to investigate how future ranklists will influence the overall fairness of the system. Then, based on the analysis of the Taylor expansion, FARA adopts a two-phase optimization algorithm where we first solve an optimal future exposure planning problem and then construct the optimal ranklists according to the optimal future exposure planning. Theoretically, we show that FARA is optimal for ranking relevance and fairness joint optimization. Empirically, our extensive experiments on three semi-synthesized datasets show that FARA is efficient, effective, and can deliver significantly better ranking performance compared to state-of-the-art fair ranking methods. | ['Qingyao Ai', 'Zhenduo Wang', 'Zhichao Xu', 'Tao Yang'] | 2023-05-26 | null | null | null | null | ['exposure-fairness', 'information-retrieval'] | ['adversarial', 'natural-language-processing'] | [-3.49328309e-01 -2.42883950e-01 -5.64952791e-01 -5.66310763e-01
-8.33824813e-01 -6.30268455e-01 3.62747282e-01 2.13860609e-02
-2.98436046e-01 6.66471779e-01 2.88242787e-01 -3.72489274e-01
-8.43575001e-01 -7.23604858e-01 -3.08689952e-01 -3.40526968e-01
-2.00340390e-01 5.98094344e-01 2.85485722e-02 -3.17487866e-01
5.82984090e-01 2.94645488e-01 -1.60094106e+00 1.19145930e-01
1.27792478e+00 1.05473614e+00 6.25263080e-02 3.22678179e-01
7.24181086e-02 6.62678123e-01 -5.37835479e-01 -6.37124062e-01
6.22216642e-01 -3.38122725e-01 -1.02702987e+00 -6.00113809e-01
1.80380702e-01 -8.42023551e-01 -2.16096029e-01 7.48879790e-01
7.03490317e-01 5.65555215e-01 3.30191553e-01 -1.23008657e+00
-6.08764827e-01 8.35821688e-01 -3.82663965e-01 6.87765107e-02
2.67193705e-01 -3.55580270e-01 1.45497000e+00 -7.00181544e-01
2.04198211e-01 1.25979793e+00 -8.39538947e-02 4.02426749e-01
-8.25614333e-01 -7.37824738e-01 4.46512699e-01 3.18783194e-01
-1.23562634e+00 -3.89229238e-01 5.84833920e-01 -1.17588826e-02
2.41163775e-01 8.52213562e-01 5.56816459e-01 2.15703428e-01
-3.34841162e-02 8.34968090e-01 1.14431977e+00 -3.84447098e-01
3.57423544e-01 1.28104731e-01 4.93727237e-01 2.23404691e-01
2.29480520e-01 3.27908307e-01 -4.86163765e-01 -5.10641158e-01
2.56613046e-01 1.41329601e-01 -8.29866454e-02 -1.52994037e-01
-7.69909918e-01 8.33204269e-01 3.86480808e-01 -2.92041540e-01
-4.60334569e-01 1.48358732e-01 1.11044779e-01 4.79881734e-01
5.36803961e-01 6.24767184e-01 -2.86004335e-01 1.35251597e-01
-1.08428001e+00 4.45107877e-01 7.85739303e-01 5.44905126e-01
4.83156115e-01 -5.76237381e-01 -9.40206766e-01 1.02157152e+00
6.00112379e-01 5.37658811e-01 5.50842024e-02 -1.34362996e+00
3.28586787e-01 3.02776366e-01 3.95368904e-01 -1.07362521e+00
-2.92388257e-03 -7.76617110e-01 -4.47610408e-01 -7.75056854e-02
3.39664251e-01 7.26658329e-02 -2.90049344e-01 1.71903741e+00
1.72322601e-01 -1.97175026e-01 -2.85004616e-01 1.22924721e+00
4.18962508e-01 6.64045215e-01 -6.14405936e-03 -8.38855684e-01
1.12273586e+00 -1.03922415e+00 -6.58443809e-01 1.42616421e-01
2.24766508e-01 -9.78890300e-01 1.19178867e+00 4.92104173e-01
-1.33276749e+00 -2.43380368e-01 -8.66648674e-01 1.56182498e-01
1.07918404e-01 2.35749641e-03 7.72712231e-01 7.96089053e-01
-9.21974838e-01 6.90363526e-01 -1.49858281e-01 8.49236622e-02
2.73502231e-01 5.35158455e-01 4.47548360e-01 -1.02310702e-01
-1.49948514e+00 6.90267086e-01 -2.32816309e-01 1.40753821e-01
-6.49797738e-01 -7.80153275e-01 -2.76460834e-02 5.45101941e-01
7.87942767e-01 -7.41784632e-01 1.43903220e+00 -5.10350108e-01
-1.73033941e+00 4.10647720e-01 -2.65542954e-01 -9.77890790e-02
5.93212247e-01 -4.89644825e-01 -2.67603993e-01 -2.65463740e-01
-5.18372357e-02 1.22835465e-01 3.16707224e-01 -1.27172220e+00
-7.62469411e-01 -2.40291372e-01 7.63974309e-01 6.07112646e-01
-6.70424819e-01 2.83399194e-01 -5.49956620e-01 -2.31509075e-01
-1.43660635e-01 -9.12409186e-01 -3.74848157e-01 -3.70524824e-01
-3.32144976e-01 -4.00369465e-01 1.00102358e-01 -3.15366745e-01
2.04424882e+00 -1.89872968e+00 -2.04876170e-01 7.97249615e-01
1.64955109e-01 1.05177924e-01 -1.40532374e-01 3.37064087e-01
3.03084582e-01 3.99071097e-01 6.19368374e-01 -6.47179857e-02
3.88189435e-01 -8.84665474e-02 -4.41364765e-01 1.65996954e-01
-8.29239607e-01 8.55937302e-01 -1.03470719e+00 -4.14483368e-01
4.36655208e-02 -2.55759787e-02 -7.62999356e-01 4.46851403e-01
-2.41080243e-02 5.51090315e-02 -6.67055368e-01 4.21803772e-01
7.56033897e-01 -3.46181989e-01 3.81145597e-01 -8.76866654e-02
-1.65026411e-01 3.68289500e-01 -1.10832858e+00 1.20309937e+00
-6.81394279e-01 5.68172745e-02 -2.52945870e-01 -5.32374918e-01
6.44175649e-01 2.58545261e-02 6.84965432e-01 -1.16988909e+00
6.10704347e-02 4.53032643e-01 -1.50498569e-01 8.32937881e-02
8.89029801e-01 -2.52633858e-02 -1.89836100e-01 7.97044694e-01
-7.99353898e-01 3.81740391e-01 3.45545948e-01 6.12501442e-01
6.16308033e-01 -3.59820336e-01 1.13656022e-01 -4.20156926e-01
6.34156704e-01 -4.12440479e-01 4.41319138e-01 8.97442222e-01
-1.80538639e-01 4.18615490e-01 4.65824515e-01 -3.32698733e-01
-5.40537775e-01 -9.39867437e-01 -6.43372163e-02 1.47943819e+00
8.00321698e-01 -4.74187344e-01 -5.53283870e-01 -7.64379561e-01
-4.66498770e-02 8.31911027e-01 -2.70397514e-01 -2.25394651e-01
-2.47784674e-01 -7.45384574e-01 -1.65418521e-01 -8.10408965e-02
2.96475917e-01 -7.43792713e-01 -3.77156377e-01 5.25886454e-02
-4.72579002e-01 -4.74995494e-01 -1.00547290e+00 -3.92796546e-01
-7.27490366e-01 -9.91078138e-01 -6.85120463e-01 -1.85724553e-02
6.46520197e-01 7.95332730e-01 1.21752930e+00 5.49583077e-01
2.91119695e-01 1.97121456e-01 -3.31655443e-01 -2.45697886e-01
7.97580034e-02 -2.55319085e-02 1.69847831e-01 -5.85897341e-02
-1.11638019e-02 -3.05230290e-01 -1.22903264e+00 7.25712597e-01
-7.16374815e-01 -2.12783769e-01 4.60796624e-01 7.18505681e-01
7.08662093e-01 2.38672704e-01 7.86408186e-01 -1.30612421e+00
1.18010759e+00 -5.19487321e-01 -7.03823686e-01 7.58756161e-01
-1.43469965e+00 1.60884678e-01 7.46645212e-01 -7.81889409e-02
-1.15511703e+00 -5.69726408e-01 1.90154657e-01 -7.31117651e-02
7.36676574e-01 5.95930874e-01 -2.84162343e-01 1.11865707e-01
4.40342367e-01 -3.81351441e-01 -2.27174133e-01 -3.83971572e-01
4.97578710e-01 7.91463077e-01 4.42458224e-03 -7.63298154e-01
6.64764822e-01 1.75641075e-01 -2.04959158e-02 3.76239754e-02
-1.13892186e+00 -5.60742915e-01 2.02022210e-01 -4.62236404e-01
1.40836522e-01 -6.24101460e-01 -1.21200848e+00 9.36471447e-02
-7.77476430e-01 -5.49410880e-02 -2.00163975e-01 4.16543633e-01
-3.21046919e-01 3.78872752e-01 -3.59912455e-01 -9.99686360e-01
-6.44652724e-01 -1.04122925e+00 6.30981445e-01 5.14429927e-01
-4.76587825e-02 -5.87999940e-01 1.88008115e-01 5.39007425e-01
6.13247991e-01 -3.82897079e-01 8.91575515e-01 -5.05908132e-01
-8.70514512e-01 -2.07981780e-01 -2.66436279e-01 -2.90617216e-02
-1.13286771e-01 -9.19760540e-02 -6.91156983e-01 -4.01403278e-01
-4.59318697e-01 -1.63389705e-02 6.07392669e-01 4.77506369e-01
1.40719724e+00 -6.89560354e-01 -9.28167477e-02 4.01539475e-01
1.33356905e+00 2.97240973e-01 7.11856663e-01 4.70676243e-01
2.50099540e-01 7.16582239e-01 1.32391357e+00 8.47184062e-01
5.19261897e-01 1.05297220e+00 4.57074195e-01 1.26935497e-01
2.52289325e-01 -2.43468508e-01 3.49129468e-01 7.79014409e-01
-2.54034370e-01 -4.66090530e-01 -1.87059045e-01 2.00482756e-01
-2.11665606e+00 -1.02361977e+00 3.90799629e-04 3.03637290e+00
5.86762488e-01 -1.00684188e-01 2.25316226e-01 -9.20097157e-02
6.86609387e-01 3.47238928e-02 -4.95336860e-01 -6.47657633e-01
1.97357297e-01 -2.05038209e-02 5.48761010e-01 6.64209604e-01
-5.57169974e-01 6.07601345e-01 6.60635376e+00 1.11432552e+00
-5.83855510e-01 1.61193721e-02 9.56735551e-01 -7.17010319e-01
-9.00899947e-01 3.61189216e-01 -5.09340525e-01 7.69294262e-01
8.29201102e-01 -9.22983348e-01 9.71300840e-01 6.48661256e-01
5.70977032e-01 -2.40451247e-02 -9.20288920e-01 7.36192524e-01
-3.56261194e-01 -9.77572381e-01 6.25160187e-02 3.59875202e-01
9.20704007e-01 -5.23598373e-01 2.44909480e-01 4.51943696e-01
3.84195507e-01 -8.64661813e-01 6.44332707e-01 7.99486935e-01
6.29264772e-01 -1.18606031e+00 8.03811252e-01 3.12992662e-01
-9.23765063e-01 -2.78558701e-01 -5.25679708e-01 -1.16892569e-01
3.88015687e-01 1.06400180e+00 -8.66016224e-02 8.10227334e-01
5.71309566e-01 1.90373033e-01 -3.69284213e-01 1.19787347e+00
-2.11173892e-01 4.28105533e-01 -1.55046746e-01 -1.17393129e-01
-2.94853866e-01 -3.17012817e-01 3.28573525e-01 4.45458263e-01
4.33350712e-01 4.43112910e-01 2.63245851e-01 3.64592999e-01
-4.50233400e-01 5.18578887e-01 5.95013648e-02 1.70740172e-01
8.90470564e-01 1.30724204e+00 -2.09173143e-01 -2.68724501e-01
-9.80141759e-02 4.89208102e-01 1.52875781e-01 4.48223948e-01
-8.89477670e-01 -2.14053094e-01 6.03564024e-01 2.14445651e-01
-3.58458340e-01 4.20449823e-01 -2.46353194e-01 -9.95143890e-01
1.13587335e-01 -7.94015050e-01 6.14049554e-01 -4.80502903e-01
-1.19129312e+00 4.16690409e-01 1.62677914e-01 -1.31014442e+00
-1.48683161e-01 -6.07863553e-02 -7.22807348e-01 9.57066417e-01
-1.93501246e+00 -7.06175625e-01 -8.11927468e-02 2.99904764e-01
2.51255512e-01 -1.07633412e-01 4.59895700e-01 5.61623752e-01
-5.75856566e-01 1.03750443e+00 4.70134199e-01 -8.15147877e-01
9.80153859e-01 -1.08878493e+00 -2.54685670e-01 7.45091736e-01
-1.69898957e-01 1.01700854e+00 5.38676023e-01 -3.03585291e-01
-1.13867486e+00 -7.28165209e-01 1.17621124e+00 -1.92024931e-01
1.14602193e-01 1.05288334e-01 -6.33827627e-01 -8.88277888e-02
-1.24142915e-01 -1.94949389e-01 8.73030603e-01 6.73119009e-01
-6.21680692e-02 -7.11161375e-01 -9.01546955e-01 7.31647193e-01
1.10850954e+00 -4.47202861e-01 4.17332649e-02 3.81170303e-01
5.62611401e-01 -2.69065440e-01 -7.57162511e-01 4.63054240e-01
1.15272689e+00 -1.16304708e+00 1.16945112e+00 -3.41740429e-01
5.56072235e-01 -2.00286075e-01 -5.45982458e-02 -1.24141252e+00
-6.38261616e-01 -8.56547654e-01 -4.80799556e-01 1.17225039e+00
4.35594559e-01 -5.51660001e-01 7.81117439e-01 1.25167894e+00
1.24236457e-01 -1.01712060e+00 -4.24641490e-01 -8.53467703e-01
-3.97184640e-02 -3.25870439e-02 9.83149827e-01 4.26695675e-01
-1.34454970e-03 2.84048527e-01 -8.50586116e-01 -4.50865813e-02
7.30613232e-01 6.86776698e-01 6.62108481e-01 -1.30896020e+00
-6.07864976e-01 -7.51759529e-01 4.83504146e-01 -1.21329391e+00
-1.43377557e-02 -7.50097036e-01 4.73378738e-03 -1.62198865e+00
7.52730310e-01 -8.87130857e-01 -1.01396620e+00 2.39294723e-01
-5.77233732e-01 -5.20474650e-02 3.51455748e-01 5.99148452e-01
-1.12488627e+00 4.31689829e-01 1.32723999e+00 4.51348647e-02
-2.98725188e-01 5.24071336e-01 -1.42572248e+00 1.57988459e-01
6.83978021e-01 -4.72814381e-01 -8.37918699e-01 -2.13187665e-01
7.80137002e-01 2.08261549e-01 -1.46013170e-01 -1.85744181e-01
2.67301232e-01 -7.80077815e-01 -1.40458703e-01 -5.63811779e-01
-9.74425077e-02 -7.03682840e-01 2.24932238e-01 3.51953685e-01
-6.78384960e-01 -1.45395041e-01 -4.71207321e-01 5.36682606e-01
-1.41120285e-01 -3.30087841e-01 4.97180849e-01 1.95480973e-01
-4.37569112e-01 6.44537210e-01 3.96437384e-02 -6.79180920e-02
8.76172006e-01 6.02503121e-02 -5.15106261e-01 -7.75222003e-01
-2.80450344e-01 8.88368189e-01 4.10433650e-01 2.83309907e-01
4.13002014e-01 -1.47678494e+00 -6.49097502e-01 -3.58953923e-01
7.03723207e-02 -5.00373006e-01 4.99733210e-01 6.37922764e-01
-7.69674778e-02 6.32858694e-01 1.73504889e-01 2.87539307e-02
-1.39969718e+00 4.43406969e-01 3.79553251e-02 -9.29513156e-01
3.27047080e-01 7.00000584e-01 4.40097123e-01 -2.02976659e-01
2.36965328e-01 4.08260494e-01 -4.69261527e-01 -9.82874259e-03
5.64815104e-01 8.49375665e-01 -2.55719088e-02 -1.81702018e-01
-4.24813688e-01 4.79649842e-01 -1.61641285e-01 -1.59141526e-01
9.65165317e-01 -5.91652751e-01 -2.78878659e-01 -9.19461250e-02
1.04078770e+00 4.96294498e-01 -8.22930396e-01 -2.45139495e-01
-1.65173411e-01 -1.18327641e+00 2.74679542e-01 -1.02415943e+00
-1.18624485e+00 5.57057440e-01 3.38644803e-01 2.61589229e-01
1.49007881e+00 -4.03212190e-01 8.66496623e-01 2.30839908e-01
6.81483269e-01 -1.25490606e+00 9.54224020e-02 3.05174261e-01
6.32379293e-01 -1.14283097e+00 4.96592194e-01 -4.39153343e-01
-5.86844385e-01 7.25957274e-01 5.80565274e-01 1.77498817e-01
6.21417344e-01 -6.86603248e-01 -7.21992627e-02 -4.63876203e-02
-9.52354670e-01 -8.47386569e-02 7.97425151e-01 -1.18487515e-01
5.95155001e-01 3.90008092e-01 -1.14596856e+00 9.90093291e-01
-7.77043849e-02 -4.15229164e-02 3.33430260e-01 4.65194941e-01
-4.16486382e-01 -1.77925038e+00 -1.26418382e-01 8.04913223e-01
-7.11366296e-01 -2.24975228e-01 -2.74921417e-01 1.05311908e-01
-1.91286623e-01 1.32195354e+00 -3.73683602e-01 -4.79685485e-01
4.84384596e-01 -4.92111802e-01 1.68733045e-01 -5.07809758e-01
-6.91263258e-01 1.42249882e-01 2.02287421e-01 -7.91925609e-01
-2.66314179e-01 -5.26215494e-01 -1.04783010e+00 -8.23891819e-01
-7.81994462e-01 8.02714348e-01 4.66074914e-01 8.45929265e-01
4.84081089e-01 5.50504088e-01 1.46154702e+00 -2.16351181e-01
-9.36760306e-01 -3.37104380e-01 -7.20957220e-01 4.08859760e-01
-1.53724626e-01 -6.31225824e-01 -3.15296799e-01 -7.87947297e-01] | [9.607768058776855, 5.60270881652832] |
51a4b4ee-bbcc-4fb4-85d7-30279fda4824 | simultaneously-color-depth-super-resolution | 1708.09105 | null | http://arxiv.org/abs/1708.09105v3 | http://arxiv.org/pdf/1708.09105v3.pdf | Simultaneously Color-Depth Super-Resolution with Conditional Generative Adversarial Network | Recently, Generative Adversarial Network (GAN) has been found wide
applications in style transfer, image-to-image translation and image
super-resolution. In this paper, a color-depth conditional GAN is proposed to
concurrently resolve the problems of depth super-resolution and color
super-resolution in 3D videos. Firstly, given the low-resolution depth image
and low-resolution color image, a generative network is proposed to leverage
mutual information of color image and depth image to enhance each other in
consideration of the geometry structural dependency of color-depth image in the
same scene. Secondly, three loss functions, including data loss, total
variation loss, and 8-connected gradient difference loss are introduced to
train this generative network in order to keep generated images close to the
real ones, in addition to the adversarial loss. Experimental results
demonstrate that the proposed approach produces high-quality color image and
depth image from low-quality image pair, and it is superior to several other
leading methods. Besides, we use the same neural network framework to resolve
the problem of image smoothing and edge detection at the same time. | ['Huihui Bai', 'Bing Zeng', 'Yao Zhao', 'Lijun Zhao', 'Jie Liang', 'Anhong Wang'] | 2017-08-30 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 4.94491488e-01 5.53561002e-02 2.22799763e-01 -1.87113866e-01
-7.56861687e-01 -2.61847138e-01 2.57830620e-01 -7.90132701e-01
-1.54698923e-01 9.59475458e-01 1.89415477e-02 2.23966748e-01
1.09706692e-01 -1.03875279e+00 -5.66066384e-01 -9.41485047e-01
5.10485888e-01 -7.53439739e-02 1.97657794e-01 -4.09324877e-02
1.54808834e-01 4.46291745e-01 -1.31760836e+00 1.10616693e-02
1.23566163e+00 1.08593297e+00 2.11524859e-01 4.84956533e-01
-1.82317123e-01 8.57321858e-01 -4.22489852e-01 -4.83726114e-01
4.36045378e-01 -7.30064690e-01 -3.86912584e-01 3.43363374e-01
5.20853400e-01 -7.10161924e-01 -6.46387100e-01 1.38945520e+00
6.69989109e-01 3.78082767e-02 4.82253134e-01 -1.02569222e+00
-9.35174286e-01 -3.47283948e-03 -1.04829478e+00 1.38577744e-01
2.57995754e-01 1.73437580e-01 3.42980236e-01 -7.45663822e-01
5.99026501e-01 1.39701653e+00 3.57455611e-01 7.26488650e-01
-1.22318709e+00 -7.46278346e-01 5.40764220e-02 -9.45084020e-02
-1.18177378e+00 -9.23304558e-02 1.24710166e+00 -2.32756168e-01
2.26607338e-01 -6.14443570e-02 5.98039508e-01 8.58528495e-01
3.25528592e-01 3.98285329e-01 1.34380615e+00 -2.85551727e-01
-5.53490445e-02 1.09110534e-01 -7.94812441e-01 8.47900808e-01
1.42090961e-01 3.68857622e-01 -3.39138359e-01 2.64421433e-01
1.66363382e+00 1.09955505e-01 -6.29009724e-01 -3.85724068e-01
-9.63034630e-01 6.10275149e-01 5.63596785e-01 2.51963973e-01
-3.22326511e-01 9.27013811e-03 -1.72723141e-02 4.84282337e-02
6.28791332e-01 9.42533687e-02 -9.98804569e-02 2.32750773e-01
-6.42872393e-01 5.92474416e-02 3.06110889e-01 1.03388608e+00
8.20070207e-01 5.81950486e-01 -8.40498954e-02 7.69293249e-01
2.75632441e-01 5.38966894e-01 4.66177195e-01 -1.03639495e+00
6.64709985e-01 5.36085010e-01 1.10668868e-01 -1.11763287e+00
9.73185375e-02 -2.25713193e-01 -1.34002459e+00 6.10605240e-01
1.48041755e-01 -1.64254948e-01 -9.50691760e-01 1.70893240e+00
3.92518371e-01 3.25062096e-01 1.54371783e-01 1.16033614e+00
7.76836276e-01 7.28892207e-01 -2.01611400e-01 -3.65256369e-01
9.29077268e-01 -5.97167790e-01 -9.48633432e-01 -3.63710910e-01
-2.85811007e-01 -8.72079730e-01 9.24528241e-01 2.23141596e-01
-1.63188756e+00 -7.90546358e-01 -1.17037642e+00 -3.76953483e-01
5.87223135e-02 -1.58755213e-01 4.35740411e-01 5.17147779e-01
-9.42896485e-01 3.89560759e-01 -4.63976383e-01 1.84937805e-01
5.38393795e-01 4.78059053e-02 -2.99506903e-01 -3.35821927e-01
-1.18658435e+00 5.36402524e-01 4.44109023e-01 1.73445135e-01
-8.07692230e-01 -8.34635556e-01 -7.67278731e-01 -2.58584619e-01
2.08731487e-01 -9.40917492e-01 6.34950995e-01 -1.20455492e+00
-1.84448302e+00 9.42174733e-01 5.90566471e-02 3.44479308e-02
7.51162350e-01 -5.01757525e-02 -4.25490230e-01 1.85595155e-01
6.46898709e-03 3.78720671e-01 1.13713956e+00 -1.48188221e+00
-7.05254078e-01 -6.23071432e-01 -5.59236705e-02 4.70087320e-01
5.76686375e-02 -3.73137683e-01 -6.91938996e-01 -8.09711576e-01
2.45600387e-01 -4.13281888e-01 -7.99209923e-02 1.87984586e-01
-3.49699676e-01 4.73907620e-01 8.44622970e-01 -8.09685469e-01
7.34020650e-01 -2.17174363e+00 5.24839997e-01 5.82492016e-02
2.35485405e-01 9.83347669e-02 -1.63501859e-01 -2.71982729e-01
1.48477256e-02 -4.89174500e-02 -4.12983477e-01 -2.75628418e-01
-5.21428645e-01 1.92187592e-01 -1.01065226e-02 4.64435726e-01
3.01507950e-01 8.73527050e-01 -8.56550694e-01 -6.59198880e-01
4.22981650e-01 1.00731027e+00 -4.94647890e-01 5.46136260e-01
3.52142416e-02 1.02908826e+00 -5.20856321e-01 4.65848476e-01
1.25383139e+00 2.92613497e-03 -1.00055173e-01 -4.57337350e-01
3.48175280e-02 -3.81077200e-01 -1.18141198e+00 1.83226991e+00
-6.41896844e-01 2.94620097e-01 3.16431731e-01 -6.54700935e-01
1.09484208e+00 1.08501919e-01 4.48082209e-01 -1.08179009e+00
1.30894765e-01 7.12683648e-02 -4.85059559e-01 -3.27349156e-01
3.33401442e-01 -3.81624997e-01 2.41718724e-01 7.91559368e-02
-2.29393184e-01 -5.53633809e-01 -2.53968298e-01 -8.60942751e-02
3.57263029e-01 3.14439803e-01 -2.19691575e-01 2.35065341e-01
8.95785213e-01 -6.33678913e-01 8.31372201e-01 2.00843707e-01
1.75766870e-01 1.05791605e+00 3.36112589e-01 -1.10333540e-01
-1.35356593e+00 -1.24539220e+00 7.48635456e-02 3.47600818e-01
6.94762707e-01 3.26864898e-01 -7.90638447e-01 -4.47700769e-01
-2.92033404e-01 4.26779747e-01 -5.89681506e-01 -2.34631360e-01
-6.61298513e-01 -4.99764204e-01 2.53597617e-01 4.59068775e-01
1.06909037e+00 -7.94298172e-01 -2.95346707e-01 2.29887575e-01
-2.11245269e-01 -1.26421452e+00 -7.11907864e-01 -2.32768208e-01
-9.86727059e-01 -1.05351043e+00 -1.19986188e+00 -8.12764287e-01
7.77129710e-01 2.56932378e-01 9.82221484e-01 -1.29843935e-01
-4.70336556e-01 1.18424654e-01 -1.50546744e-01 3.25409956e-02
-5.29504001e-01 -3.76467317e-01 -2.84762502e-01 3.13882291e-01
-3.02037328e-01 -8.15388560e-01 -9.89151776e-01 3.05224031e-01
-1.25192988e+00 3.06644708e-01 8.56043339e-01 1.09328556e+00
8.38292956e-01 3.94844562e-01 2.97639251e-01 -8.27453911e-01
3.29444081e-01 -1.18606769e-01 -7.32377827e-01 9.74732861e-02
-7.11525738e-01 -9.23843384e-02 5.88597238e-01 -3.91964853e-01
-1.69710231e+00 -9.55419801e-03 -1.68979958e-01 -6.62001967e-01
-2.05484009e-03 -8.03965051e-03 -7.11858809e-01 -3.77167791e-01
2.71348983e-01 6.77032590e-01 2.45219663e-01 -3.56208086e-01
5.33759236e-01 2.23612651e-01 7.61524916e-01 -2.65957326e-01
1.09849572e+00 6.64314151e-01 2.30170652e-01 -4.79846895e-01
-7.26101696e-01 1.50743619e-01 -5.53705990e-01 -3.85453105e-02
1.22822452e+00 -1.09291387e+00 -5.09399414e-01 7.65909076e-01
-1.06410694e+00 -2.28631496e-01 -2.13854805e-01 3.88746202e-01
-6.78187728e-01 5.16621828e-01 -7.35790670e-01 -5.67951620e-01
-3.05673420e-01 -1.00266206e+00 1.03382540e+00 7.24388361e-01
6.88798666e-01 -1.07110786e+00 -3.17904204e-01 3.48823488e-01
4.56984401e-01 7.34881580e-01 9.62795734e-01 5.63777685e-01
-9.47926641e-01 3.81399542e-02 -6.62859499e-01 8.17638755e-01
4.53089416e-01 -2.62257189e-01 -8.08342040e-01 -3.21833670e-01
1.86481923e-01 -1.12282299e-01 5.59279621e-01 5.41602075e-01
1.15503216e+00 -2.96140432e-01 6.51739985e-02 1.14779866e+00
1.80078363e+00 3.31580728e-01 1.05938721e+00 1.19528994e-01
1.13737774e+00 3.72868955e-01 5.78948140e-01 2.68995494e-01
1.33625939e-01 7.04009652e-01 5.03814220e-01 -6.61818862e-01
-5.24006546e-01 -3.90510947e-01 1.06391013e-01 3.94370139e-01
-2.53779620e-01 -5.45736104e-02 -1.72693864e-01 1.56012982e-01
-1.40712023e+00 -8.99409592e-01 8.59524310e-02 2.29081440e+00
7.38957047e-01 1.04498088e-01 -1.44968301e-01 -5.17716743e-02
8.23222637e-01 1.65623814e-01 -8.22518229e-01 -1.75834093e-02
-4.97593462e-01 2.00722784e-01 5.35671234e-01 5.59789002e-01
-7.82284558e-01 7.86934435e-01 5.45260143e+00 9.40857768e-01
-1.25228703e+00 1.41623482e-01 1.07386887e+00 8.58306885e-02
-5.94032824e-01 -3.08612764e-01 -3.97403985e-01 6.56657517e-01
2.26664096e-01 -1.16698965e-01 6.59347713e-01 5.74200332e-01
1.74690276e-01 -6.82409629e-02 -6.70850098e-01 1.25440705e+00
2.14305207e-01 -1.01489687e+00 2.81283945e-01 8.59230533e-02
1.15102851e+00 -5.36866188e-01 4.79404747e-01 -2.46946365e-01
1.00965641e-01 -1.04702926e+00 4.36516076e-01 7.08258450e-01
1.26884222e+00 -1.17339337e+00 6.20761812e-01 -1.13753239e-02
-1.17687976e+00 7.20392168e-02 -3.90678138e-01 3.90704274e-01
3.38584930e-01 6.44764125e-01 -6.57968894e-02 1.01064777e+00
6.61798418e-01 7.53137946e-01 -2.94036597e-01 6.83204591e-01
-3.97179782e-01 -1.15639403e-01 1.82667449e-01 5.16803920e-01
1.29716462e-02 -5.51339984e-01 5.82681715e-01 6.87198400e-01
4.06540275e-01 3.08528155e-01 -1.71786830e-01 1.26215482e+00
-1.46177579e-02 -9.78490710e-02 -6.05415702e-01 3.75809878e-01
3.17640036e-01 1.13561404e+00 -3.41397941e-01 -1.36980012e-01
-4.07512188e-01 1.59222710e+00 3.61598395e-02 6.20311618e-01
-9.13151801e-01 -4.12217498e-01 6.17509365e-01 1.63419187e-01
2.23221064e-01 -5.89221455e-02 -3.05266649e-01 -1.13226283e+00
8.61395746e-02 -7.05874383e-01 6.38346225e-02 -1.06143689e+00
-1.34399438e+00 6.87529147e-01 -3.35158199e-01 -1.50921261e+00
-2.08383977e-01 -3.48641783e-01 -5.12360632e-01 1.29584765e+00
-1.92161417e+00 -1.10694551e+00 -6.78962111e-01 8.88758183e-01
4.76402044e-01 -5.04201725e-02 4.28582102e-01 5.35526931e-01
-4.49529529e-01 6.71556056e-01 2.77807683e-01 4.53583375e-02
7.20242262e-01 -1.06227934e+00 1.23940781e-01 9.62147713e-01
-4.87545103e-01 1.41593784e-01 3.86313736e-01 -5.75475454e-01
-1.43693519e+00 -1.29561543e+00 1.06080053e-02 -1.34324029e-01
4.01661918e-02 -3.72441486e-02 -9.23509061e-01 2.56429851e-01
1.00294821e-01 7.46629834e-02 4.92849424e-02 -6.78876221e-01
-1.76507905e-01 -3.53920251e-01 -1.54570031e+00 4.74918455e-01
9.49133337e-01 -4.33332205e-01 -1.54280931e-01 -9.52114835e-02
8.34632099e-01 -9.16700244e-01 -1.04222655e+00 4.01484191e-01
3.92614454e-01 -1.29501772e+00 1.26314080e+00 2.00760532e-02
9.17144716e-01 -5.17499149e-01 7.38656893e-02 -1.28327835e+00
-3.21456343e-01 -5.18346369e-01 1.42767146e-01 1.45929646e+00
1.30797699e-01 -6.71113968e-01 8.08166623e-01 4.43196416e-01
-4.74259630e-03 -5.68676591e-01 -9.20584798e-01 -5.44698179e-01
1.07006989e-01 2.27315634e-01 5.52850187e-01 8.60954344e-01
-7.82055497e-01 3.41989934e-01 -7.01372325e-01 2.74420351e-01
1.13982737e+00 1.46575168e-01 8.03118110e-01 -9.22314703e-01
-3.08672518e-01 -3.82554203e-01 -3.57160121e-01 -1.14743721e+00
-4.36892696e-02 -5.77372372e-01 4.72528487e-02 -1.50367939e+00
1.80432752e-01 -3.33040804e-01 -1.68341205e-01 -1.54773012e-01
-4.50229049e-01 5.38014531e-01 1.58548541e-02 2.97156740e-02
-6.48520887e-02 8.18234801e-01 1.99972951e+00 -1.32180572e-01
-2.80403107e-01 -6.58902749e-02 -7.34654903e-01 5.42659342e-01
5.19076288e-01 -5.71503602e-02 -5.30483186e-01 -5.88850200e-01
-1.22257560e-01 4.66141373e-01 4.48334455e-01 -8.59041631e-01
-6.46491023e-03 -1.81109741e-01 9.66678977e-01 -4.26974684e-01
4.55761671e-01 -9.32208598e-01 1.91911414e-01 2.53233969e-01
-9.16503146e-02 -2.45214641e-01 1.16473556e-01 7.97349155e-01
-4.05135423e-01 1.70617536e-01 1.33587480e+00 -2.55058020e-01
-6.89891100e-01 6.81801379e-01 2.44419798e-01 2.14117587e-01
9.30982232e-01 -4.64122802e-01 -2.00861365e-01 -4.14020449e-01
-3.52984965e-01 -7.17297494e-02 6.26105130e-01 4.42810714e-01
1.03634036e+00 -1.48454404e+00 -8.90598238e-01 4.27520365e-01
-2.35933848e-02 4.29229051e-01 7.43632734e-01 4.37292665e-01
-6.71826780e-01 -7.64879510e-02 -5.94603717e-01 -4.81135368e-01
-9.26810443e-01 6.56463742e-01 4.74850118e-01 -1.17266469e-01
-8.84740174e-01 8.86749804e-01 5.94114959e-01 -1.11059379e-02
1.52446285e-01 -1.20761814e-02 -4.11407324e-03 -5.15682280e-01
5.77947140e-01 2.66965032e-01 -3.63268554e-01 -5.52502036e-01
3.65322828e-02 1.13280106e+00 -5.31505495e-02 -7.94032142e-02
1.09269261e+00 -5.53334296e-01 -9.55699310e-02 1.65559962e-01
1.23944926e+00 9.42202061e-02 -1.73214400e+00 -3.64529908e-01
-7.36401975e-01 -1.03555286e+00 2.97379583e-01 -5.80909550e-01
-1.69840682e+00 7.61611938e-01 1.04398692e+00 6.32176325e-02
1.57247663e+00 -3.59639794e-01 9.64802325e-01 -5.63256562e-01
2.42651165e-01 -9.34117794e-01 4.25242305e-01 9.19856057e-02
6.67150974e-01 -1.30506599e+00 3.16699930e-02 -6.02480948e-01
-6.08542621e-01 9.76846695e-01 9.58663881e-01 -2.49393910e-01
1.92347959e-01 1.56831622e-01 6.61290139e-02 1.26788557e-01
-1.79296553e-01 -4.42117229e-02 3.56203645e-01 8.56824934e-01
3.57102901e-01 -2.62369871e-01 -1.60054997e-01 1.73800588e-01
1.47095010e-01 7.39564672e-02 4.40483063e-01 3.92104208e-01
-1.24760568e-01 -9.38855052e-01 -3.16858977e-01 1.36393225e-02
-4.88507032e-01 -1.46892026e-01 -7.74956271e-02 6.48101270e-01
1.83360547e-01 7.41999209e-01 1.38447061e-01 -3.21053416e-01
3.76587957e-01 -4.83020544e-01 7.46460497e-01 -2.56580025e-01
5.03721759e-02 2.94332415e-01 -4.71872181e-01 -6.13221049e-01
-4.30929184e-01 -1.49474263e-01 -1.07329440e+00 -2.67332494e-01
-1.19908296e-01 -1.30021185e-01 6.41945839e-01 4.93292958e-01
2.31088430e-01 7.87899613e-01 1.11968839e+00 -9.31203961e-01
-1.14843570e-01 -6.98309183e-01 -9.60239887e-01 5.31106651e-01
4.10969734e-01 -3.84283364e-01 -4.68910933e-01 1.16140358e-01] | [11.01171875, -2.033147096633911] |
8e91c718-ec18-44e1-855c-60ee4be66956 | drone-shadow-tracking | 1905.08214 | null | https://arxiv.org/abs/1905.08214v1 | https://arxiv.org/pdf/1905.08214v1.pdf | Drone Shadow Tracking | Aerial videos taken by a drone not too far above the surface may contain the drone's shadow projected on the scene. This deteriorates the aesthetic quality of videos. With the presence of other shadows, shadow removal cannot be directly applied, and the shadow of the drone must be tracked. Tracking a drone's shadow in a video is, however, challenging. The varying size, shape, change of orientation and drone altitude pose difficulties. The shadow can also easily disappear over dark areas. However, a shadow has specific properties that can be leveraged, besides its geometric shape. In this paper, we incorporate knowledge of the shadow's physical properties, in the form of shadow detection masks, into a correlation-based tracking algorithm. We capture a test set of aerial videos taken with different settings and compare our results to those of a state-of-the-art tracking algorithm. | ['Sabine Süsstrunk', 'Xiaoyan Zou', 'Ruofan Zhou', 'Majed El Helou'] | 2019-05-20 | null | null | null | null | ['shadow-removal', 'shadow-detection'] | ['computer-vision', 'computer-vision'] | [ 5.37459254e-01 -4.35109675e-01 3.44413161e-01 1.79659784e-01
2.71914452e-01 -1.19542134e+00 4.12104070e-01 -1.84909865e-01
-7.43680894e-02 6.99056685e-01 -2.00131878e-01 -1.13333747e-01
1.22620232e-01 -4.29236948e-01 -4.22167361e-01 -7.55704522e-01
-3.56153995e-01 -1.51075134e-02 1.00529563e+00 -1.91384777e-01
7.33979885e-03 7.47018874e-01 -1.46889627e+00 -1.75515875e-01
4.57996249e-01 8.30778062e-01 2.26143092e-01 9.83287334e-01
6.82399988e-01 4.32760239e-01 -1.00801563e+00 7.59821152e-03
6.56844914e-01 -3.27809811e-01 1.37707561e-01 5.36941886e-01
1.01436901e+00 -7.07160115e-01 -4.03103650e-01 7.82852948e-01
9.24322903e-02 3.20369750e-01 2.36838669e-01 -1.39192653e+00
1.77205548e-01 -4.40245956e-01 -5.69890738e-01 2.04975501e-01
6.08645737e-01 1.53659970e-01 4.91270155e-01 -4.90280896e-01
8.07942927e-01 7.52125204e-01 9.03591812e-01 1.16506159e-01
-9.87354100e-01 -7.05578089e-01 3.18415761e-01 -3.95418286e-01
-1.36807775e+00 -2.98473299e-01 6.34668767e-01 -4.86447453e-01
2.94505209e-01 4.09240842e-01 1.07023990e+00 7.10003078e-01
4.86673445e-01 4.24693137e-01 9.54532802e-01 -2.27578700e-01
7.84400254e-02 4.75822799e-02 -4.01875466e-01 1.06026828e+00
6.65815055e-01 4.57251161e-01 -6.02643788e-01 -3.99407774e-01
5.17699838e-01 1.05575897e-01 -9.83483672e-01 -9.47342336e-01
-9.41657245e-01 3.26269656e-01 2.97167122e-01 -1.07844248e-01
-1.35422751e-01 2.16515824e-01 -9.47305188e-02 -3.79363894e-02
2.60058016e-01 5.00185490e-01 -3.85586649e-01 -1.61600232e-01
-1.30548954e+00 2.72447914e-01 7.67467082e-01 1.08730030e+00
4.73793536e-01 1.50926262e-01 9.60150734e-02 7.88038149e-02
1.15301713e-01 9.64628696e-01 -4.29798245e-01 -8.19160044e-01
1.40976802e-01 4.92193401e-01 5.98883688e-01 -1.32362711e+00
-2.50139326e-01 -1.89824015e-01 -1.84927300e-01 8.64096045e-01
4.68437195e-01 -6.03508413e-01 -9.27908719e-01 1.20904315e+00
5.61850667e-01 3.59888166e-01 -1.12227514e-01 1.31341755e+00
4.02517498e-01 3.79181147e-01 -5.30275941e-01 -3.03760141e-01
1.18191540e+00 -6.31519198e-01 -7.86222219e-01 -4.89889801e-01
9.83288065e-02 -1.22544861e+00 4.68165100e-01 4.99061584e-01
-7.38479018e-01 -1.91031083e-01 -1.31855619e+00 6.01562262e-01
-1.09718584e-01 2.95032203e-01 4.21133608e-01 6.92728937e-01
-6.94087923e-01 5.55586457e-01 -8.00823629e-01 -4.82963234e-01
5.11970297e-02 5.19509852e-01 -1.22745834e-01 -2.21557245e-01
-7.05965281e-01 8.75052571e-01 -1.46721035e-01 2.83748154e-02
-6.92327738e-01 -6.16949975e-01 -7.02832758e-01 -1.79739788e-01
9.65168893e-01 -4.25114065e-01 1.07307100e+00 -1.15794110e+00
-1.13736165e+00 4.55667049e-01 -3.16186430e-04 -6.74792230e-02
6.78438842e-01 -5.24540186e-01 -5.11924505e-01 3.04252237e-01
-1.32140517e-01 1.07220985e-01 1.42274380e+00 -1.67056286e+00
-7.34658420e-01 -2.19068363e-01 3.75117540e-01 3.41883957e-01
-7.69883022e-02 -2.20290244e-01 -5.58342159e-01 -5.92453837e-01
-1.77112103e-01 -1.43943989e+00 3.22308391e-02 4.47608441e-01
-2.51732439e-01 7.92548120e-01 1.62904322e+00 -5.52599013e-01
1.29926598e+00 -2.16226339e+00 -5.78141771e-02 2.40637153e-01
9.74759459e-02 5.30353129e-01 2.73099452e-01 4.82141674e-01
4.26704317e-01 -4.60749328e-01 -3.77018481e-01 5.10126576e-02
-3.72914135e-01 8.57422948e-02 -3.99799734e-01 7.56056726e-01
-1.71785951e-01 1.45684868e-01 -1.09093082e+00 -2.73355722e-01
2.37400100e-01 6.40379071e-01 -2.00174645e-01 2.33278275e-01
-9.46293920e-02 2.54620519e-02 -3.74485195e-01 9.45649683e-01
8.94232392e-01 1.64267629e-01 3.41197878e-01 -5.75365499e-02
-4.94610339e-01 -2.52751946e-01 -1.18137860e+00 9.58941400e-01
-1.68703571e-01 1.12614417e+00 5.09981513e-01 2.99579233e-01
6.32357478e-01 9.89978239e-02 4.84070271e-01 5.12551852e-02
-4.62076515e-02 -6.59752414e-02 -4.17213934e-03 -3.83018672e-01
9.60693181e-01 -3.62301506e-02 2.39478648e-01 3.20495903e-01
-5.47339618e-01 -2.50869095e-01 7.92652443e-02 1.47100925e-01
1.41661704e+00 4.26972598e-01 1.79498479e-01 -3.28793615e-01
-3.04322299e-02 1.98120788e-01 4.77500141e-01 3.47728133e-01
-1.34019151e-01 7.49003887e-01 9.37163383e-02 -2.32137144e-01
-5.82679272e-01 -8.40774179e-01 -2.25593317e-02 7.13781595e-01
5.61576009e-01 -5.47948718e-01 -6.89030170e-01 -5.09137034e-01
3.20944011e-01 3.77052188e-01 -5.34131646e-01 -1.38442889e-01
-4.84919965e-01 -5.74135721e-01 3.87338847e-01 3.37924451e-01
3.45124692e-01 -4.49521363e-01 -1.82928205e+00 1.12853669e-01
-2.46934630e-02 -1.44734621e+00 -7.73936093e-01 -5.31939007e-02
-7.69599259e-01 -1.39741230e+00 -6.02530003e-01 -3.63717712e-02
8.69299531e-01 1.10046637e+00 1.08993268e+00 5.56621611e-01
-4.79303122e-01 8.52327943e-01 -3.80922735e-01 -6.18516982e-01
-4.34028311e-03 -5.53739130e-01 7.73077235e-02 1.28242245e-03
-3.06152105e-01 -1.49190113e-01 -7.56804407e-01 6.36812508e-01
-7.47084260e-01 -1.79257303e-01 1.76515669e-01 5.32712579e-01
1.11146763e-01 6.44820333e-01 -5.39263427e-01 -6.31213009e-01
1.41765878e-01 -6.96000457e-02 -1.26055348e+00 4.44296859e-02
-1.53031096e-01 -5.88206649e-01 5.53461015e-01 -4.77251977e-01
-8.77728581e-01 5.00292540e-01 9.82816696e-01 -8.29364598e-01
3.21846381e-02 -8.02620649e-02 -1.02095559e-01 -4.76861626e-01
2.03747004e-01 -1.97437540e-01 -2.43991271e-01 -2.27262396e-02
-1.53727651e-01 1.91921070e-01 4.30198431e-01 1.12670377e-01
1.30017042e+00 1.02062786e+00 4.14969265e-01 -1.32725525e+00
-8.98062170e-01 -4.07798588e-01 -8.37877095e-01 -6.97253525e-01
5.61474025e-01 -6.39389515e-01 -5.08644998e-01 3.46338272e-01
-9.39795852e-01 -5.02427399e-01 2.05973089e-01 3.72259796e-01
-7.60605782e-02 3.41718137e-01 3.09742689e-02 -1.16311550e+00
2.43503034e-01 -8.05913687e-01 1.06554782e+00 3.00896645e-01
-2.25075841e-01 -1.10912549e+00 -5.80881722e-03 1.49825200e-01
4.09741960e-02 7.18725622e-01 1.56115860e-01 1.85627773e-01
-1.11441994e+00 -3.31227422e-01 3.22577059e-02 -3.97624215e-03
4.26066697e-01 4.89181340e-01 -1.01458347e+00 -6.41591847e-01
-1.86163470e-01 3.95614207e-01 6.78436279e-01 4.28439558e-01
3.31662953e-01 -7.90682510e-02 -7.09559381e-01 7.23785043e-01
1.37481952e+00 1.90374643e-01 3.26691717e-01 1.25961572e-01
7.47304976e-01 3.48220497e-01 1.37033558e+00 6.32962525e-01
-1.92513511e-01 9.88570809e-01 6.11131489e-01 -3.78571212e-01
-1.93546921e-01 7.95299560e-02 7.40942478e-01 1.53371301e-02
-5.19886911e-01 -7.02592075e-01 -8.39862645e-01 2.15258617e-02
-1.48771536e+00 -8.85371506e-01 -4.91794020e-01 2.51660013e+00
3.18254679e-02 1.30567268e-01 -1.33910030e-03 -2.00767051e-02
5.88655412e-01 2.96312630e-01 -4.46903884e-01 4.73354384e-02
-3.42098773e-02 -3.33240330e-01 1.15488255e+00 6.95982218e-01
-1.23813546e+00 9.10817027e-01 6.37422323e+00 9.66110006e-02
-1.28624284e+00 -3.76527965e-01 -4.24995393e-01 -2.90433556e-01
7.99526423e-02 2.54582971e-01 -8.57933521e-01 4.57245171e-01
2.61967212e-01 1.86343521e-01 5.36129475e-01 5.19065797e-01
9.77335572e-02 -8.01782191e-01 -7.94564605e-01 7.41746962e-01
5.95541000e-01 -7.94025481e-01 -4.71453577e-01 3.75949442e-01
5.53742588e-01 -1.94251060e-01 9.62435901e-02 -3.72868747e-01
1.44889191e-01 -5.78713775e-01 6.33116961e-01 5.15579641e-01
7.20839560e-01 -6.37459695e-01 6.11787319e-01 3.22845206e-02
-1.39309192e+00 -6.19588867e-02 -2.88209289e-01 -1.91257149e-01
3.04732502e-01 4.61900264e-01 -1.13560557e+00 4.17613477e-01
7.04472065e-01 5.45074821e-01 -6.51353478e-01 1.06221068e+00
-3.24940920e-01 3.95445049e-01 -4.57662582e-01 3.87846567e-02
2.76472867e-01 -3.41970116e-01 1.12933445e+00 9.89301085e-01
4.61916924e-01 5.70905089e-01 5.01174510e-01 4.65726346e-01
2.81060874e-01 -4.01714712e-01 -1.08883381e+00 -5.70270140e-03
2.71165550e-01 1.34428704e+00 -1.05473423e+00 -7.98892751e-02
-3.94611120e-01 1.12897420e+00 -4.74594861e-01 4.58250046e-01
-8.93693984e-01 -7.34656751e-02 7.39405751e-01 5.08621812e-01
5.92450619e-01 -4.95982558e-01 2.43974596e-01 -1.02650082e+00
2.92165335e-02 -7.84066081e-01 1.11632429e-01 -9.93282795e-01
-5.06282926e-01 5.83466947e-01 2.00266559e-02 -1.64882755e+00
3.27199966e-01 -5.28160036e-01 -5.68879664e-01 3.74780357e-01
-1.31603396e+00 -1.05514991e+00 -7.55360663e-01 3.67926419e-01
3.97687733e-01 -2.76716836e-02 6.40352011e-01 1.69840619e-01
-1.72572866e-01 -6.58387467e-02 5.74387461e-02 -5.76280057e-02
8.37546945e-01 -1.06643975e+00 1.03644030e-02 1.28557968e+00
-1.03223607e-01 3.20320547e-01 1.21425259e+00 -1.09033132e+00
-1.77794194e+00 -9.70524490e-01 3.21494818e-01 -8.05017412e-01
6.46163821e-01 -4.61550593e-01 -7.06743419e-01 6.05814517e-01
1.40346587e-01 -1.21521004e-01 6.22130871e-01 -8.64905193e-02
1.29879147e-01 -1.82925798e-02 -8.42979074e-01 6.71677172e-01
8.69369447e-01 -2.14012206e-01 -3.41284066e-01 2.61375815e-01
1.80627629e-01 -7.01990545e-01 -4.25602287e-01 3.38625163e-01
9.93656754e-01 -1.14984632e+00 9.55323100e-01 -7.36181065e-02
-2.56469399e-01 -7.54493058e-01 -2.29366675e-01 -1.39578044e+00
-2.02624276e-01 -8.69542658e-01 -7.59333149e-02 8.57575774e-01
-6.32843524e-02 -3.45946252e-01 6.50186062e-01 5.60281336e-01
-1.75010592e-01 -3.90794247e-01 -5.49055576e-01 -9.70566630e-01
-8.76404643e-01 1.66924685e-01 2.79564321e-01 6.35866404e-01
-3.57144058e-01 2.61771858e-01 -7.00469792e-01 7.83907175e-01
6.27013564e-01 4.84121501e-01 1.15446782e+00 -1.43207324e+00
-1.85504377e-01 2.37704962e-01 -3.18570942e-01 -6.88635349e-01
-2.95984566e-01 -1.25876427e-01 3.93943876e-01 -9.93293762e-01
-8.85257497e-02 -3.39918256e-01 2.39357248e-01 1.81307837e-01
-1.42672405e-01 3.30247909e-01 6.31781399e-01 -4.42104563e-02
-4.32569057e-01 2.61798292e-01 1.24767840e+00 1.84528455e-01
-2.28103891e-01 2.08053008e-01 1.04551762e-01 1.25948215e+00
7.08897054e-01 -5.88579118e-01 -4.88439739e-01 -2.81214178e-01
1.59236684e-01 1.43780082e-01 5.15096426e-01 -1.30147552e+00
1.41022459e-01 -2.84570396e-01 6.76857591e-01 -6.73613369e-01
9.87390459e-01 -1.31898880e+00 2.06324965e-01 7.36298680e-01
4.32060331e-01 2.91576833e-01 4.60492790e-01 5.88810325e-01
2.53893226e-01 -1.94442496e-01 7.53853023e-01 -8.28200951e-02
-3.55508387e-01 2.20986456e-01 -7.46151745e-01 -8.44534189e-02
1.20256972e+00 -5.09986341e-01 -6.93676025e-02 -4.45875883e-01
-2.26596877e-01 3.50799635e-02 1.23450601e+00 2.81814694e-01
7.20013738e-01 -8.30691218e-01 -1.59076065e-01 2.67272860e-01
-6.60511926e-02 -4.05312002e-01 -5.66229187e-02 9.32304859e-01
-6.33451819e-01 1.43318057e-01 -1.65363088e-01 -4.98469442e-01
-2.03793025e+00 5.07583201e-01 2.69687742e-01 1.35891795e-01
-7.97160089e-01 4.44142848e-01 2.73465782e-01 2.36595035e-01
4.55150604e-02 -3.59951764e-01 2.17090234e-01 2.16862366e-01
5.00532389e-01 5.20636976e-01 -2.78757095e-01 -6.91346109e-01
-5.28082252e-01 8.41868997e-01 2.93496370e-01 -2.69700028e-02
1.01274061e+00 -2.62419820e-01 2.30947271e-01 4.31086749e-01
4.46445227e-01 8.07872355e-01 -1.72888029e+00 2.26339832e-01
-4.01466697e-01 -1.11726892e+00 2.80339986e-01 -7.13387668e-01
-9.99309063e-01 6.25221074e-01 4.94962364e-01 1.85418546e-01
1.14449298e+00 -5.03694654e-01 8.78673971e-01 2.91977495e-01
6.14978969e-01 -1.13663149e+00 1.07235163e-01 4.10769433e-01
6.68259144e-01 -1.11621404e+00 5.60880423e-01 -6.96406424e-01
-6.89314663e-01 1.19557285e+00 4.52273160e-01 -1.19135082e-01
5.29941916e-01 6.81290686e-01 3.08811754e-01 -1.50464028e-01
-5.45226872e-01 -2.68901110e-01 3.17360669e-01 5.85951090e-01
-3.54419574e-02 7.64972568e-02 3.81405622e-01 -1.72441527e-01
-1.64706960e-01 -1.36489183e-01 9.53754067e-01 1.43401432e+00
-6.12453103e-01 -6.36559248e-01 -8.73828828e-01 1.70669615e-01
-2.21369535e-01 1.25383036e-02 -7.82743871e-01 1.11721194e+00
3.74813586e-01 8.42871785e-01 -6.81203604e-02 -3.61412227e-01
3.86515230e-01 -3.15507859e-01 6.58952296e-01 -5.26836216e-01
-3.30232412e-01 3.13836962e-01 2.52444804e-01 -4.84325528e-01
-5.13285935e-01 -1.02703476e+00 -1.12420154e+00 -2.63094991e-01
-6.98908627e-01 -1.65794149e-01 4.54928070e-01 5.47444046e-01
2.59268619e-02 4.92337227e-01 3.90094876e-01 -8.18532109e-01
8.17556307e-02 -3.51247072e-01 -7.24615455e-01 -3.58946361e-02
8.30917835e-01 -1.26628041e+00 -4.39914167e-01 3.01011682e-01] | [7.389305591583252, -1.5735526084899902] |
be4bb708-05df-4758-9597-98cb50017349 | semi-supervised-multitask-learning-for | 1704.07156 | null | http://arxiv.org/abs/1704.07156v1 | http://arxiv.org/pdf/1704.07156v1.pdf | Semi-supervised Multitask Learning for Sequence Labeling | We propose a sequence labeling framework with a secondary training objective,
learning to predict surrounding words for every word in the dataset. This
language modeling objective incentivises the system to learn general-purpose
patterns of semantic and syntactic composition, which are also useful for
improving accuracy on different sequence labeling tasks. The architecture was
evaluated on a range of datasets, covering the tasks of error detection in
learner texts, named entity recognition, chunking and POS-tagging. The novel
language modeling objective provided consistent performance improvements on
every benchmark, without requiring any additional annotated or unannotated
data. | ['Marek Rei'] | 2017-04-24 | semi-supervised-multitask-learning-for-1 | https://aclanthology.org/P17-1194 | https://aclanthology.org/P17-1194.pdf | acl-2017-7 | ['grammatical-error-detection'] | ['natural-language-processing'] | [ 2.86464959e-01 3.12904179e-01 -6.72837615e-01 -4.75136071e-01
-6.67575181e-01 -6.76183701e-01 3.67032558e-01 6.10507309e-01
-8.77102613e-01 9.46295857e-01 5.15595078e-01 -6.73554897e-01
3.78949672e-01 -3.83338094e-01 -7.03027129e-01 -2.81056076e-01
-8.52322802e-02 3.92967552e-01 1.50561273e-01 -1.22001313e-01
3.34331840e-01 2.20274806e-01 -1.32156777e+00 8.82553875e-01
1.08718741e+00 8.65829349e-01 5.72670519e-01 6.13226414e-01
-8.90560091e-01 1.38530219e+00 -7.55563617e-01 -5.58936357e-01
-2.38774166e-01 -1.94541723e-01 -1.18307650e+00 -2.04988435e-01
1.40183121e-01 3.42235267e-01 3.33997048e-02 9.10118580e-01
4.10521537e-01 2.41711587e-01 3.75781417e-01 -6.34488523e-01
-7.75626659e-01 1.26395142e+00 -3.70871462e-02 2.68309921e-01
5.27416646e-01 -7.63398334e-02 1.32863307e+00 -9.01097894e-01
9.11664367e-01 8.14003885e-01 8.17590892e-01 7.85289049e-01
-8.72557938e-01 -3.21557760e-01 3.44952404e-01 3.83028001e-01
-1.21045542e+00 -4.54187781e-01 2.80327141e-01 -3.52760196e-01
1.61719179e+00 2.52673090e-01 1.30851150e-01 1.17555952e+00
1.88866258e-01 1.30754566e+00 8.30901861e-01 -9.05598879e-01
3.47178668e-01 2.69600600e-01 6.46893740e-01 6.86748087e-01
1.45242110e-01 5.61016379e-04 -8.97478163e-01 -1.60600662e-01
-2.23517552e-01 -2.47896209e-01 -3.59927803e-01 1.94919761e-02
-9.88506258e-01 8.24249685e-01 -1.85213581e-01 4.99804139e-01
-2.69845009e-01 -1.57098979e-01 8.28945816e-01 2.42872521e-01
9.32093024e-01 7.24188685e-01 -1.33479476e+00 -5.26569843e-01
-8.36678386e-01 -1.98474452e-01 9.82848644e-01 1.03066778e+00
6.60126626e-01 -4.54579964e-02 -3.48250210e-01 1.04118001e+00
3.46419424e-01 3.02235305e-01 1.14779925e+00 -2.33931482e-01
7.84450352e-01 7.49977171e-01 -1.96736474e-02 -2.58571237e-01
-7.13890493e-01 -3.95572931e-01 -2.51718849e-01 -2.83020079e-01
3.31402689e-01 -4.06543821e-01 -1.02700591e+00 1.63818228e+00
2.39905849e-01 4.83911157e-01 3.86183560e-01 3.43694061e-01
1.14636230e+00 7.46219993e-01 7.58307517e-01 -1.95198387e-01
1.40035903e+00 -1.21318841e+00 -8.32962155e-01 -5.46522141e-01
1.52406061e+00 -7.18647838e-01 1.20597816e+00 2.00695828e-01
-7.81916916e-01 -3.41347873e-01 -7.34443605e-01 -2.27028858e-02
-5.62318563e-01 3.13420504e-01 3.65613371e-01 6.67428613e-01
-8.55519235e-01 5.26320815e-01 -6.36207819e-01 -8.20269361e-02
3.37628067e-01 -1.36789670e-02 -3.66849840e-01 1.40412018e-01
-1.46181750e+00 1.32584226e+00 9.85166311e-01 -3.81813973e-01
-4.83711898e-01 -1.08442295e+00 -1.23552692e+00 1.99019790e-01
1.32865250e-01 -2.97835469e-01 1.41957211e+00 -1.12230420e+00
-1.38483036e+00 1.07482040e+00 -3.71588200e-01 -7.88319170e-01
1.38396725e-01 -4.66420501e-01 -5.23790956e-01 -2.86269099e-01
-1.04533032e-01 4.61737275e-01 4.50242877e-01 -9.20741141e-01
-6.76633656e-01 -3.21013957e-01 -4.98299509e-01 3.05072337e-01
-5.42880535e-01 3.17940265e-01 -5.54008633e-02 -9.75997806e-01
-5.11339426e-01 -6.84957683e-01 -3.58694851e-01 -8.41562629e-01
-4.53238904e-01 -7.35798717e-01 5.84627151e-01 -1.05763280e+00
1.58719110e+00 -1.88799119e+00 -1.88944992e-02 -4.21978757e-02
-1.43003881e-01 7.40282476e-01 -4.45213586e-01 3.84560823e-01
-1.74615636e-01 3.20677489e-01 -1.47690043e-01 -4.85481501e-01
-5.51260337e-02 2.51850069e-01 -4.57553923e-01 3.14393602e-02
2.16367438e-01 1.25612152e+00 -9.27508652e-01 -2.89046377e-01
8.62153620e-02 1.43682614e-01 -1.46372736e-01 5.46047449e-01
-6.53684676e-01 1.33804351e-01 -7.61134550e-02 3.59622896e-01
2.17637792e-01 -1.60599738e-01 3.73206764e-01 3.64013910e-01
7.73354396e-02 1.10895801e+00 -9.39812362e-01 1.86772156e+00
-8.17847610e-01 3.48723233e-01 -3.40721130e-01 -1.04761708e+00
8.81080866e-01 4.89178240e-01 2.85271972e-01 -6.92421496e-01
-9.39978193e-03 1.57967865e-01 -1.70086071e-01 -7.97104597e-01
7.38596082e-01 -1.33253127e-01 -3.13657761e-01 6.19623423e-01
5.25871277e-01 4.76467490e-01 1.93317950e-01 1.39473349e-01
1.36853874e+00 9.97931436e-02 6.73470795e-01 -2.42283657e-01
6.43075466e-01 2.65634298e-01 5.27581811e-01 5.42191684e-01
-1.24713182e-02 4.87793311e-02 1.77949771e-01 -4.63296294e-01
-1.00157368e+00 -4.19430703e-01 -3.50328051e-02 1.98939300e+00
-3.70165668e-02 -6.03502214e-01 -7.70305872e-01 -1.32430577e+00
-1.04495235e-01 1.23092568e+00 -4.01787013e-01 -1.44737974e-01
-6.58688188e-01 -5.27270854e-01 8.45677614e-01 8.18673074e-01
-1.09608583e-01 -1.53561914e+00 -4.37370092e-01 3.76970649e-01
-1.81133449e-01 -1.19932318e+00 -6.45839155e-01 7.39722192e-01
-7.70531893e-01 -8.17595959e-01 -3.45599592e-01 -1.13235998e+00
4.93865311e-01 -1.95422426e-01 1.50044990e+00 3.23256314e-01
-1.39301524e-01 2.54939556e-01 -7.19701827e-01 -5.65931141e-01
-7.42857456e-01 5.93264103e-01 -1.00273505e-01 -2.64897436e-01
7.95178711e-01 -5.53972311e-02 1.02438703e-01 1.05361324e-02
-7.81567335e-01 -3.60896997e-02 3.95604253e-01 9.80437696e-01
2.57794201e-01 -4.52710390e-01 7.64412522e-01 -1.35589969e+00
5.30012071e-01 -6.14410877e-01 -2.69126296e-01 7.25284100e-01
-6.05357289e-01 5.93740225e-01 8.19072127e-01 -5.61268508e-01
-1.27589464e+00 2.09030494e-01 -6.38858140e-01 5.19425534e-02
-4.91585970e-01 6.92822099e-01 -3.49738091e-01 1.86330214e-01
6.93717539e-01 4.00474757e-01 -2.75651664e-01 -8.06068182e-01
6.17926776e-01 6.62895322e-01 2.13629559e-01 -3.57731044e-01
2.49486193e-01 -1.97760731e-01 -6.40741110e-01 -8.36883426e-01
-1.31321323e+00 -9.74635065e-01 -8.43023241e-01 2.38973320e-01
7.92432308e-01 -9.48159814e-01 -4.92778182e-01 3.35384101e-01
-1.36740375e+00 -6.02763593e-01 -3.86206329e-01 3.56233925e-01
-2.55325258e-01 3.95477742e-01 -6.34463489e-01 -7.16384351e-01
-6.03945255e-01 -7.25709319e-01 8.57564092e-01 8.12024102e-02
-5.75598001e-01 -1.52513540e+00 3.51953238e-01 2.81377256e-01
2.09840283e-01 -4.23289716e-01 1.01657951e+00 -1.67056346e+00
7.34640509e-02 1.32506952e-01 1.92352727e-01 4.10437495e-01
-2.30999082e-01 -3.56354922e-01 -8.93190742e-01 3.07666101e-02
-2.46757448e-01 -6.57387316e-01 9.92395759e-01 -6.66430965e-02
1.29237485e+00 -7.13088989e-01 -3.47194403e-01 4.02469963e-01
1.13930583e+00 1.57604828e-01 3.18815947e-01 3.66201520e-01
6.87970281e-01 7.28716254e-01 5.53808391e-01 3.47544909e-01
3.66509527e-01 5.88938475e-01 8.75600353e-02 1.18142776e-01
-1.80198997e-01 -5.50136626e-01 6.27173066e-01 7.60618627e-01
5.25245070e-01 -5.23087621e-01 -1.31985080e+00 6.52997851e-01
-1.80233788e+00 -8.55119228e-01 -7.80018866e-02 1.80812490e+00
1.27890778e+00 -1.95808038e-01 -1.29541516e-01 -1.90126583e-01
6.89970195e-01 1.53298736e-01 -4.48473245e-01 -5.89737654e-01
-2.18124446e-02 5.74831724e-01 8.13272834e-01 5.22040665e-01
-1.06855536e+00 1.31269717e+00 6.79339409e+00 1.15140247e+00
-1.10293126e+00 3.71291667e-01 5.82457960e-01 2.38218576e-01
-5.36336541e-01 -2.51451731e-01 -1.46594822e+00 6.42372012e-01
1.48771167e+00 -1.36624470e-01 7.52046180e-04 1.04705155e+00
-2.08491273e-02 9.19698086e-03 -8.22491467e-01 5.47019064e-01
1.62770972e-01 -1.59759820e+00 -3.57706696e-02 -4.38191980e-01
6.46782100e-01 3.70297134e-01 -2.98026651e-01 6.49253011e-01
6.05191946e-01 -1.16988134e+00 7.89187908e-01 2.75469691e-01
6.00530267e-01 -7.50420928e-01 8.43844831e-01 8.55739415e-01
-8.50176573e-01 -2.64331736e-02 -1.93269044e-01 -4.44852561e-02
3.27476293e-01 6.26684129e-01 -1.10853553e+00 3.52808595e-01
3.01465094e-01 6.79464817e-01 -6.01066351e-01 8.55870843e-01
-6.98968709e-01 1.08746660e+00 -2.40361150e-02 -4.69614267e-01
2.40482181e-01 1.95765540e-01 3.95454317e-01 1.85647893e+00
-4.59414311e-02 8.04086030e-02 3.14879000e-01 3.12725931e-01
-4.76764560e-01 6.78011358e-01 -3.88518304e-01 -1.18130498e-01
6.63429916e-01 1.26528692e+00 -5.02479851e-01 -4.53285307e-01
-4.77765113e-01 8.53007674e-01 8.28287542e-01 1.10075526e-01
-6.60593271e-01 -2.78704107e-01 7.93308377e-01 -1.27320990e-01
4.93476033e-01 -1.63690612e-01 -7.98173130e-01 -1.46063411e+00
-2.01191008e-01 -8.45324278e-01 6.16452336e-01 -1.98143765e-01
-1.23363996e+00 5.45958042e-01 -4.04173344e-01 -6.81926727e-01
-3.20873260e-01 -9.11991596e-01 -6.07328653e-01 7.74802089e-01
-1.63343596e+00 -8.72432292e-01 3.29790503e-01 1.93899304e-01
7.12715924e-01 -4.17748004e-01 9.47301507e-01 2.50042170e-01
-6.71297967e-01 8.98264349e-01 1.84795246e-01 3.81781071e-01
7.23262966e-01 -1.41630030e+00 5.53781450e-01 8.45140874e-01
7.70555973e-01 3.90482068e-01 5.33497036e-01 -7.26345539e-01
-1.05496240e+00 -1.40569043e+00 1.60179555e+00 -6.42984807e-01
6.95842147e-01 -4.78308588e-01 -1.30879426e+00 7.85602808e-01
1.97220638e-01 9.59894955e-02 1.22098756e+00 4.89027143e-01
-4.23728555e-01 5.44528842e-01 -8.53210688e-01 1.59674793e-01
1.13437986e+00 -4.56819594e-01 -8.89547169e-01 4.73008543e-01
1.10365045e+00 -5.44977903e-01 -6.29818082e-01 3.15966636e-01
1.51241094e-01 -3.46285999e-01 8.03175211e-01 -1.44181204e+00
4.00935858e-01 6.96992129e-02 5.44411577e-02 -1.50880980e+00
7.26546906e-03 -4.47188169e-01 -5.51125169e-01 1.38166606e+00
1.09818840e+00 -4.43173587e-01 9.03081417e-01 6.50617659e-01
-4.32169080e-01 -8.64302218e-01 -7.67736018e-01 -9.27877784e-01
1.44595444e-01 -7.63866305e-01 4.58279371e-01 1.06463969e+00
4.92873430e-01 5.93885481e-01 -2.88121194e-01 -1.31831244e-02
9.27864611e-02 -2.24503726e-01 2.85003453e-01 -8.16907287e-01
-1.47243500e-01 -3.87133420e-01 -3.12209398e-01 -1.16306829e+00
9.40051794e-01 -1.45655191e+00 3.56751114e-01 -1.29719639e+00
-2.32814089e-03 -7.12890685e-01 -2.03581274e-01 9.00497317e-01
-6.43877685e-01 -1.04506105e-01 8.38088393e-02 -6.48615435e-02
-1.04599273e+00 5.91756701e-01 4.02615339e-01 1.82020366e-01
-7.19480664e-02 -5.12116961e-03 -4.75598633e-01 7.41658092e-01
7.77187169e-01 -7.07073867e-01 -1.31164595e-01 -6.19391620e-01
2.49823540e-01 -1.83412611e-01 -3.15421551e-01 -6.66188598e-01
2.53663003e-01 -1.75622210e-01 1.96572214e-01 -4.50182647e-01
-3.26433957e-01 -6.03392005e-01 -2.57209390e-01 4.90279287e-01
-9.74836826e-01 2.55734324e-02 3.07422042e-01 3.13331515e-01
-1.18975312e-01 -9.59348142e-01 6.95391715e-01 9.41949897e-03
-1.26715171e+00 -4.32402566e-02 -4.74919945e-01 6.99483931e-01
1.24140406e+00 2.29845151e-01 -1.96820199e-01 -2.35373210e-02
-8.47527444e-01 2.38041282e-01 1.68198720e-01 6.49453640e-01
3.54775041e-01 -1.10158968e+00 -7.02180803e-01 1.21111579e-01
3.16823065e-01 -2.87179112e-01 5.95533997e-02 2.70604819e-01
-3.33285749e-01 6.75330579e-01 1.16971985e-01 -3.94767642e-01
-1.50053322e+00 7.09579051e-01 1.83474317e-01 -9.46211696e-01
-3.48140895e-01 1.21415555e+00 -2.76679575e-01 -9.52766359e-01
5.25340378e-01 -3.25028718e-01 -4.98558819e-01 1.14038885e-01
8.64945710e-01 3.52043033e-01 4.84210640e-01 -5.54024220e-01
-3.85740489e-01 -1.05659455e-01 -2.83801377e-01 2.15273201e-01
1.20427728e+00 -1.89719677e-01 1.85125962e-01 4.46357787e-01
1.18565845e+00 4.49083112e-02 -8.65294993e-01 -6.49054408e-01
1.07635891e+00 2.08623167e-02 -7.92329758e-02 -1.04243672e+00
-5.88900447e-01 8.29312682e-01 2.05624066e-02 -9.55668837e-03
6.98440194e-01 1.61879823e-01 1.09044707e+00 4.68289614e-01
3.11917722e-01 -1.10276914e+00 -1.02640778e-01 1.37331498e+00
2.93226272e-01 -1.33293259e+00 -3.93075883e-01 -3.12165320e-01
-1.04790783e+00 9.67200100e-01 6.70195103e-01 3.13905060e-01
6.10050440e-01 6.26870215e-01 2.05086052e-01 1.87494550e-02
-1.11419642e+00 -3.05919647e-01 6.05731905e-01 4.35561419e-01
9.36164439e-01 -3.02472655e-02 -6.04597151e-01 8.95568430e-01
-7.99128413e-02 -3.15955579e-01 1.03972763e-01 9.93291020e-01
-7.57133305e-01 -1.46354282e+00 9.36806481e-03 4.34333503e-01
-7.52363324e-01 -5.26610255e-01 -4.08926636e-01 3.26326460e-01
7.90124387e-02 7.44449854e-01 3.65710780e-02 -1.62369564e-01
2.46288985e-01 8.37103426e-01 4.68943547e-03 -1.27998221e+00
-1.12263787e+00 -6.40197277e-01 4.97906774e-01 -5.08968115e-01
-2.44390443e-01 -6.68816447e-01 -1.67454803e+00 1.34817258e-01
-4.83454883e-01 5.17803669e-01 6.52072966e-01 1.26330972e+00
4.87008333e-01 2.32675210e-01 3.42361540e-01 -1.51184395e-01
-5.94393373e-01 -1.17702615e+00 -2.39755780e-01 5.75585306e-01
4.72745225e-02 -2.29744181e-01 -6.90516680e-02 2.42349848e-01] | [10.723347663879395, 8.802068710327148] |
645510c5-40c7-41c7-aa4c-539e976a9046 | fpconv-learning-local-flattening-for-point | 2002.10701 | null | https://arxiv.org/abs/2002.10701v3 | https://arxiv.org/pdf/2002.10701v3.pdf | FPConv: Learning Local Flattening for Point Convolution | We introduce FPConv, a novel surface-style convolution operator designed for 3D point cloud analysis. Unlike previous methods, FPConv doesn't require transforming to intermediate representation like 3D grid or graph and directly works on surface geometry of point cloud. To be more specific, for each point, FPConv performs a local flattening by automatically learning a weight map to softly project surrounding points onto a 2D grid. Regular 2D convolution can thus be applied for efficient feature learning. FPConv can be easily integrated into various network architectures for tasks like 3D object classification and 3D scene segmentation, and achieve comparable performance with existing volumetric-type convolutions. More importantly, our experiments also show that FPConv can be a complementary of volumetric convolutions and jointly training them can further boost overall performance into state-of-the-art results. | ['Zizheng Yan', 'Haibin Huang', 'Yiqun Lin', 'Xiaoguang Han', 'Ligang Liu', 'Shuguang Cui', 'Dong Du'] | 2020-02-25 | fpconv-learning-local-flattening-for-point-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Lin_FPConv_Learning_Local_Flattening_for_Point_Convolution_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Lin_FPConv_Learning_Local_Flattening_for_Point_Convolution_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-object-classification'] | ['computer-vision'] | [-8.59354343e-03 1.01387650e-01 1.13977846e-02 -4.44260985e-01
-3.89338583e-01 -4.60520148e-01 6.16887987e-01 1.83002859e-01
-2.22275257e-01 9.75712985e-02 -3.51767749e-01 -5.91552556e-01
3.63248080e-01 -1.09764338e+00 -9.92428958e-01 -1.11341022e-01
-2.44796321e-01 5.85544705e-01 4.79806811e-01 9.27906558e-02
1.55466393e-01 1.29980671e+00 -1.21209753e+00 1.88130721e-01
6.75192058e-01 1.36848462e+00 2.46945322e-01 4.50318545e-01
-7.34856248e-01 4.66569215e-02 -1.88597202e-01 -7.36112818e-02
4.24081445e-01 3.90407115e-01 -8.88563097e-01 1.00572474e-01
6.54334664e-01 -2.62128264e-01 -2.16674700e-01 8.06990564e-01
2.35710621e-01 6.18402958e-02 8.61890435e-01 -1.00559187e+00
-5.82438231e-01 9.50743258e-02 -6.96581841e-01 4.96785901e-02
4.93396483e-02 2.25806803e-01 7.54855216e-01 -1.37316084e+00
5.49961865e-01 1.40103495e+00 1.16317344e+00 2.75385112e-01
-1.22652292e+00 -4.81130809e-01 2.30935395e-01 -3.34989965e-01
-1.30648565e+00 1.61925837e-01 9.33820963e-01 -4.94797796e-01
1.38472044e+00 3.31994206e-01 9.51829910e-01 4.86293346e-01
9.78667736e-02 8.32502007e-01 8.12377930e-01 8.65854416e-03
1.95355639e-01 -2.60295540e-01 -2.82571800e-02 9.48397219e-01
-2.27845367e-02 -5.00230081e-02 -1.37859553e-01 -2.19060659e-01
1.48343968e+00 1.33349270e-01 -2.37846836e-01 -5.71303010e-01
-1.13666785e+00 8.60248029e-01 1.00302255e+00 -3.99006568e-02
-3.09166074e-01 7.15583265e-01 2.62335092e-01 2.24763736e-01
8.99506748e-01 4.01863605e-01 -6.37808681e-01 1.75875053e-01
-8.43505859e-01 3.30185711e-01 5.07628560e-01 1.09241056e+00
1.17648423e+00 3.36853066e-03 -2.78897017e-01 6.09215319e-01
4.14777845e-01 5.11720121e-01 -2.46982276e-01 -1.08422744e+00
2.68171042e-01 1.07470596e+00 -1.03228122e-01 -6.69154108e-01
-6.33065462e-01 -3.89935523e-01 -6.94930077e-01 8.49092841e-01
1.05340846e-01 7.85293430e-02 -1.32478714e+00 9.74194705e-01
5.35885692e-01 6.80253983e-01 -4.02751058e-01 8.52802098e-01
1.23526955e+00 5.71565032e-01 -2.16217771e-01 3.74476314e-01
1.09038210e+00 -7.52612889e-01 3.36527862e-02 -1.52833998e-01
6.72041714e-01 -6.76606655e-01 1.04689360e+00 6.26970008e-02
-1.17557836e+00 -4.62040901e-01 -8.60211968e-01 -3.77355933e-01
-3.16105276e-01 -2.54395008e-01 1.09113073e+00 2.37091556e-01
-1.30649555e+00 9.25157487e-01 -1.05997849e+00 -5.01399748e-02
1.25683260e+00 5.32513022e-01 -2.31238380e-01 -1.79885440e-02
-4.62293625e-01 6.93475783e-01 1.23836279e-01 -1.19854644e-01
-6.32294714e-01 -1.40786326e+00 -1.06008410e+00 4.84351739e-02
7.93046597e-03 -1.17453158e+00 1.31814480e+00 -2.35536799e-01
-1.39059210e+00 1.21564555e+00 -2.62218952e-01 -3.21394771e-01
3.44610989e-01 -1.33313537e-01 1.94708288e-01 6.43449202e-02
9.61948372e-03 7.11094618e-01 7.49436975e-01 -1.30886412e+00
-2.05497265e-01 -4.50140089e-01 -4.87500429e-02 1.21815033e-01
1.91340223e-01 -1.51960075e-01 -7.58437097e-01 -3.87493163e-01
4.85227466e-01 -4.74890143e-01 -5.61342418e-01 7.13961422e-01
-4.26871568e-01 -5.92058420e-01 1.24250746e+00 -1.00271314e-01
4.12577540e-01 -2.20831966e+00 -4.14998047e-02 4.61020023e-01
6.64225280e-01 2.45807603e-01 -1.44424096e-01 -5.70505634e-02
1.10377580e-01 2.70313829e-01 -4.94340539e-01 -6.09141707e-01
1.94053650e-02 1.80575430e-01 -1.48002997e-01 6.00077927e-01
7.76343942e-01 1.47075462e+00 -7.81681240e-01 -2.58814543e-01
6.58846438e-01 8.96540880e-01 -6.51954293e-01 -7.13144541e-02
-4.91330802e-01 3.49512756e-01 -5.10166705e-01 7.41038084e-01
1.15854728e+00 -6.58754170e-01 -5.46252728e-01 -1.55460224e-01
-2.47044731e-02 3.16232711e-01 -8.24662983e-01 2.10475755e+00
-6.44233763e-01 6.31631792e-01 1.92170992e-01 -9.20077264e-01
1.13185346e+00 -3.39008053e-03 6.19552791e-01 -4.87021089e-01
3.44477743e-01 1.56196445e-01 -5.03490210e-01 9.55154896e-02
3.40837926e-01 -4.38197218e-02 6.63215667e-02 2.19071284e-01
1.50704458e-01 -9.13699389e-01 -5.37261963e-01 4.04329784e-03
1.08768499e+00 3.24546903e-01 -1.74537078e-02 -3.91997993e-01
2.92169690e-01 9.23007429e-02 2.09423289e-01 4.93686885e-01
2.20024809e-01 8.17102015e-01 3.75056863e-01 -6.56645954e-01
-9.61571693e-01 -1.27193308e+00 -5.33524454e-01 5.92139781e-01
3.29209745e-01 -2.21673578e-01 -5.28848886e-01 -7.39063084e-01
6.91673994e-01 5.30236125e-01 -4.97977167e-01 1.53649151e-01
-6.07505381e-01 -3.51033479e-01 3.29471707e-01 8.44455540e-01
4.76227760e-01 -9.82517600e-01 -4.36949372e-01 2.34568074e-01
5.64537108e-01 -1.04778147e+00 -3.98983359e-01 4.52099442e-01
-1.27607775e+00 -1.11131370e+00 -4.86564636e-01 -8.20928395e-01
6.34307206e-01 4.64668542e-01 1.54219210e+00 2.49556109e-01
-1.39078766e-01 3.82339835e-01 -1.25209823e-01 -6.99966252e-01
-7.66023099e-02 1.74634680e-01 -3.20182592e-01 -4.18166459e-01
4.76065129e-01 -1.02792084e+00 -5.78476846e-01 2.83038970e-02
-6.28890634e-01 1.58077046e-01 1.76828519e-01 3.57365489e-01
1.05593348e+00 -3.07324916e-01 9.22106802e-02 -9.11866128e-01
3.96225274e-01 -2.90085912e-01 -6.79351807e-01 -1.78737894e-01
-1.92736447e-01 -9.30689648e-02 3.95328730e-01 -1.29451171e-01
-6.12113178e-01 2.34643564e-01 -4.99357581e-01 -1.00765419e+00
-2.48940334e-01 3.11694831e-01 2.05877408e-01 -6.48307383e-01
5.42617679e-01 1.23787016e-01 1.23392753e-01 -6.65188789e-01
5.31512499e-01 2.75271505e-01 3.78650874e-01 -4.83540803e-01
1.08316278e+00 9.20186639e-01 3.05119038e-01 -8.09653223e-01
-7.86971092e-01 -5.20447016e-01 -8.98656428e-01 -5.94382435e-02
8.67291093e-01 -9.78578568e-01 -8.15806627e-01 4.52099621e-01
-1.55703044e+00 -8.79584551e-01 -5.72531164e-01 1.69091031e-01
-6.94159448e-01 1.36291400e-01 -5.34001350e-01 -4.49786276e-01
-4.75657970e-01 -1.01431513e+00 1.59235537e+00 1.10507436e-01
1.82247423e-02 -1.24795163e+00 -2.24591956e-01 -2.04424471e-01
3.45138103e-01 5.36507487e-01 8.79462838e-01 -2.03156948e-01
-8.93160760e-01 -1.31909430e-01 -6.68620527e-01 2.42150158e-01
-4.04392183e-02 -6.81732148e-02 -1.12084055e+00 -1.56628907e-01
-1.28027394e-01 -1.74959764e-01 1.06029272e+00 5.54399669e-01
1.75119388e+00 6.98657557e-02 -5.66599309e-01 1.41014469e+00
1.56176686e+00 -2.57218003e-01 5.55932105e-01 4.13901918e-02
1.07036924e+00 -7.89321437e-02 1.26849994e-01 2.30134919e-01
4.31724191e-01 4.71139699e-01 8.97073209e-01 -5.58743894e-01
-4.13139850e-01 -1.95074782e-01 -1.70450866e-01 5.69952309e-01
-2.87673116e-01 1.54867426e-01 -9.47869301e-01 2.11197317e-01
-1.63438976e+00 -5.25664508e-01 -5.44056475e-01 1.89187932e+00
4.01470274e-01 1.72673002e-01 -4.72599380e-02 -1.71374857e-01
3.02407235e-01 2.06858590e-01 -6.79845870e-01 -4.84762192e-01
-3.77920531e-02 8.73260677e-01 7.09724069e-01 4.69325572e-01
-1.26688337e+00 1.12393486e+00 6.69245481e+00 8.83792162e-01
-1.35940433e+00 2.36953065e-01 4.95329976e-01 -1.28260711e-02
-7.03391492e-01 -2.93323308e-01 -5.69488108e-01 1.86177298e-01
1.18760735e-01 1.54767722e-01 2.77134269e-01 1.04435658e+00
-3.73028889e-02 2.94705003e-01 -1.13572872e+00 1.17346454e+00
-3.98855984e-01 -1.89507592e+00 -6.90675853e-03 2.62543201e-01
7.50430882e-01 8.40509474e-01 -9.81206447e-02 2.87853748e-01
4.53440547e-01 -1.40124929e+00 5.97833335e-01 3.24215919e-01
1.13332736e+00 -7.85619915e-01 4.88050967e-01 2.37030521e-01
-1.48337483e+00 4.83656377e-01 -7.14534521e-01 -2.21903890e-01
1.98709741e-01 8.15104544e-01 -8.35492194e-01 4.15568173e-01
7.20178068e-01 8.50668013e-01 -3.59858990e-01 1.27022159e+00
-1.48053974e-01 4.60014910e-01 -7.02068210e-01 3.89194004e-02
4.90192771e-01 -1.53283030e-01 5.79211473e-01 1.18905520e+00
2.78644383e-01 3.43086235e-02 3.26472849e-01 1.33342874e+00
-2.98429966e-01 -1.00277027e-03 -9.35520709e-01 3.19174767e-01
3.57492119e-01 1.27384651e+00 -8.99038315e-01 -2.84884721e-01
-6.45627916e-01 9.76419270e-01 6.27383351e-01 3.77454400e-01
-6.96903944e-01 -3.71938616e-01 9.19037580e-01 3.50706965e-01
6.91870749e-01 -4.94134635e-01 -9.15018141e-01 -9.64002967e-01
-2.20075771e-02 1.41210020e-01 -2.09817156e-01 -7.63396859e-01
-1.45283747e+00 5.75774193e-01 -3.21222275e-01 -1.03555119e+00
4.26066667e-01 -1.01146233e+00 -1.05527103e+00 1.12739718e+00
-1.89056361e+00 -1.20507371e+00 -5.41590750e-01 7.10054159e-01
3.02829236e-01 2.13704288e-01 6.99262977e-01 -3.46043222e-02
-3.47284190e-02 3.40925068e-01 -2.88648516e-01 1.46567166e-01
-5.68168145e-03 -1.40193880e+00 9.61114824e-01 4.44715321e-01
3.98080684e-02 3.02848637e-01 -4.76841666e-02 -6.83476567e-01
-1.34350204e+00 -1.61868525e+00 4.28032309e-01 -5.39190888e-01
3.57289433e-01 -5.10436654e-01 -1.16048539e+00 6.30297005e-01
-9.48828161e-02 6.00342512e-01 3.83263230e-01 1.71089679e-01
-4.93617415e-01 3.15161437e-01 -1.29323006e+00 3.28269929e-01
1.61104131e+00 -4.87701863e-01 -3.84102464e-01 4.94684696e-01
1.09431410e+00 -8.96623552e-01 -1.00964725e+00 6.89834416e-01
1.23888049e-02 -8.88020515e-01 1.29149675e+00 -5.50616801e-01
3.97175223e-01 -3.25450748e-01 -8.56281295e-02 -1.26792610e+00
-6.61814928e-01 -3.99068743e-01 -3.41131419e-01 5.70320904e-01
2.68771589e-01 -6.53690457e-01 1.07030725e+00 2.67714292e-01
-8.29425216e-01 -1.22629893e+00 -1.05411398e+00 -7.80687928e-01
3.59331161e-01 -9.83489215e-01 8.84889960e-01 8.69284213e-01
-3.91695231e-01 9.80260074e-02 4.39912140e-01 3.51242393e-01
6.29559278e-01 3.80138099e-01 7.67910421e-01 -1.61389935e+00
1.85736921e-02 -9.45866525e-01 -6.54602408e-01 -1.47033107e+00
1.35833651e-01 -1.62021530e+00 -1.82372138e-01 -1.86317444e+00
-2.13018030e-01 -9.40515578e-01 -5.44182444e-03 5.74876249e-01
-3.76999192e-02 7.05520093e-01 1.54194623e-01 -7.50716729e-03
-3.15967113e-01 6.60169184e-01 1.86813653e+00 -1.91777080e-01
-3.82362157e-01 1.27057023e-02 -6.04596198e-01 8.75717342e-01
7.61212349e-01 -1.06433414e-01 -1.97488353e-01 -7.27334738e-01
8.61274675e-02 -6.77229837e-02 7.13847637e-01 -8.74378920e-01
-6.70824349e-02 -2.25331083e-01 6.32432938e-01 -1.04311633e+00
4.81536597e-01 -7.70174384e-01 -1.05625659e-01 4.71777208e-02
4.01246339e-01 -2.35432446e-01 5.53865433e-01 3.75853628e-01
2.36910041e-02 6.70595914e-02 7.17943788e-01 -3.98302734e-01
-6.58190489e-01 9.63528454e-01 3.36509012e-02 -1.22322924e-01
9.57474887e-01 -5.31622946e-01 5.94469942e-02 7.75155723e-02
-5.38569629e-01 3.24544758e-01 7.56828725e-01 7.39549473e-02
9.84018564e-01 -1.62419367e+00 -6.17606759e-01 4.12550837e-01
-7.29388967e-02 1.15133190e+00 7.81377554e-02 6.00482225e-01
-9.30614412e-01 2.26539746e-01 1.58190623e-01 -1.18058014e+00
-8.04277599e-01 3.63623172e-01 4.23774630e-01 3.54997255e-02
-1.18164241e+00 1.19211197e+00 4.67911750e-01 -8.97075653e-01
1.54368937e-01 -7.44149387e-01 2.63143659e-01 -4.29371357e-01
2.35794976e-01 -2.78689363e-03 2.77039826e-01 -2.96232134e-01
-5.03323078e-01 8.46269965e-01 2.68319547e-01 3.69228184e-01
1.73719752e+00 4.21188533e-01 -1.18622042e-01 3.04585457e-01
1.37287486e+00 -2.25861937e-01 -1.43191302e+00 -3.16064388e-01
-3.52892607e-01 -4.90786165e-01 3.80882174e-01 -3.82910460e-01
-1.44555652e+00 1.09740031e+00 2.89252639e-01 1.31544501e-01
7.61175692e-01 4.78431731e-01 8.18399429e-01 3.62025261e-01
4.30071622e-01 -4.96670157e-01 -1.63584016e-02 8.52125525e-01
1.02037656e+00 -1.10680401e+00 -1.15151837e-01 -9.86963868e-01
-4.98139299e-02 1.25868785e+00 5.27503788e-01 -7.33740866e-01
1.16939628e+00 5.41173518e-01 -1.13808505e-01 -7.89075851e-01
-2.26911977e-01 -3.20554256e-01 6.32755816e-01 7.91245878e-01
3.62266183e-01 1.53187215e-01 3.04107010e-01 1.75449148e-01
-3.04448098e-01 -1.29758753e-02 1.31986812e-01 7.96592891e-01
-4.01737988e-01 -8.80244195e-01 -1.90797105e-01 8.13843966e-01
1.95474848e-02 -1.65840372e-01 -1.04160719e-01 7.46920645e-01
1.85978651e-01 1.15099519e-01 7.14710712e-01 -4.17131424e-01
4.40960050e-01 1.07090669e-02 7.11684048e-01 -9.16500628e-01
-5.03845036e-01 -7.29619118e-04 -3.08897734e-01 -7.96537161e-01
-3.38009566e-01 -4.97456104e-01 -1.54872906e+00 -4.97002244e-01
-2.16102779e-01 -3.52186352e-01 6.42033935e-01 7.82507658e-01
6.11388385e-01 5.85018933e-01 6.11671984e-01 -1.52512777e+00
-1.42698154e-01 -7.40754366e-01 -4.83104974e-01 3.06159109e-01
2.26735160e-01 -7.94547021e-01 -1.58231959e-01 -3.73736471e-01] | [8.031340599060059, -3.692086935043335] |
64355c01-0158-4e27-81f8-d033de32cb65 | prodesign-toward-effective-and-efficient | 2209.12643 | null | https://arxiv.org/abs/2209.12643v4 | https://arxiv.org/pdf/2209.12643v4.pdf | PiFold: Toward effective and efficient protein inverse folding | How can we design protein sequences folding into the desired structures effectively and efficiently? AI methods for structure-based protein design have attracted increasing attention in recent years; however, few methods can simultaneously improve the accuracy and efficiency due to the lack of expressive features and autoregressive sequence decoder. To address these issues, we propose PiFold, which contains a novel residue featurizer and PiGNN layers to generate protein sequences in a one-shot way with improved recovery. Experiments show that PiFold could achieve 51.66\% recovery on CATH 4.2, while the inference speed is 70 times faster than the autoregressive competitors. In addition, PiFold achieves 58.72\% and 60.42\% recovery scores on TS50 and TS500, respectively. We conduct comprehensive ablation studies to reveal the role of different types of protein features and model designs, inspiring further simplification and improvement. The PyTorch code is available at \href{https://github.com/A4Bio/PiFold}{GitHub}. | ['Stan Z. Li', 'Pablo Chacón', 'Cheng Tan', 'Zhangyang Gao'] | 2022-09-22 | null | null | null | null | ['protein-design'] | ['medical'] | [ 1.63450629e-01 -7.50503689e-02 -4.72909510e-02 -3.04379314e-01
-7.88188696e-01 -6.30185604e-01 -6.70499206e-02 -1.86194628e-01
-2.12717935e-01 1.10140121e+00 4.61034141e-02 -4.98049349e-01
2.71044999e-01 -4.21357334e-01 -9.94048834e-01 -8.94326985e-01
1.72570333e-01 4.24498707e-01 -6.80910274e-02 -3.18386853e-01
5.59460446e-02 2.39905208e-01 -1.16577625e+00 3.75467062e-01
1.17113733e+00 5.67999542e-01 5.53974926e-01 6.52730346e-01
8.34105983e-02 5.94462454e-01 -3.87437761e-01 -3.20841223e-01
1.24667689e-01 -7.71429777e-01 -7.21102715e-01 -4.49989080e-01
4.05450948e-02 -2.24925429e-01 -2.98963755e-01 7.83200026e-01
8.31057310e-01 -9.95405540e-02 5.10829628e-01 -6.28993332e-01
-9.71334636e-01 6.39864624e-01 -5.37515044e-01 8.06254894e-02
2.56412745e-01 8.07745636e-01 1.04868913e+00 -1.03801072e+00
4.95474458e-01 1.00503790e+00 4.90670532e-01 7.26536214e-01
-1.44066966e+00 -9.03617382e-01 2.29645781e-02 2.47170672e-01
-1.28360462e+00 -5.62866867e-01 3.44936639e-01 -1.53365597e-01
1.35142946e+00 1.23445176e-01 4.15167302e-01 1.15886104e+00
1.55421808e-01 9.14800882e-01 7.74900198e-01 -2.89104530e-03
8.37310404e-02 -5.19739687e-01 1.71938837e-01 8.00885022e-01
4.67464253e-02 1.00575931e-01 -5.86378276e-01 -3.72283041e-01
6.79503202e-01 1.29298270e-01 -3.99453074e-01 -5.48664480e-02
-1.25031030e+00 7.14998603e-01 5.02367973e-01 1.13300525e-01
-6.62754118e-01 1.53381839e-01 2.35285595e-01 3.02297324e-02
7.43087381e-03 9.20530796e-01 -8.21189225e-01 -3.02786440e-01
-5.24544656e-01 5.16731739e-01 7.17459202e-01 1.15483296e+00
5.96051335e-01 1.34553000e-01 -1.88498318e-01 7.26657808e-01
2.50125639e-02 4.77544159e-01 2.63491541e-01 -8.15344095e-01
2.33118564e-01 4.65101272e-01 2.72650778e-01 -5.38936079e-01
-4.37619060e-01 -1.92317411e-01 -7.95252204e-01 -3.06773126e-01
4.24381554e-01 -3.86412859e-01 -8.75401139e-01 1.82450855e+00
2.16074437e-01 7.50126764e-02 8.20048302e-02 1.03189766e+00
8.08187068e-01 9.44786727e-01 3.16323221e-01 -2.06981480e-01
1.53416836e+00 -9.06173110e-01 -4.98005301e-01 -1.51941910e-01
6.25758410e-01 -7.88315594e-01 1.12435341e+00 4.47501451e-01
-1.21150374e+00 -4.10846323e-01 -1.05643845e+00 -2.94453681e-01
5.98818399e-02 4.04865295e-01 7.10187495e-01 1.71661943e-01
-5.86691260e-01 7.59397686e-01 -9.67758954e-01 -4.10671160e-02
4.43309098e-01 4.79035944e-01 -1.20296657e-01 2.02985294e-02
-1.24961817e+00 7.13761389e-01 5.32711387e-01 1.10213988e-01
-9.75121081e-01 -9.86784995e-01 -5.18191993e-01 1.64575547e-01
4.00858313e-01 -7.67042160e-01 1.20756471e+00 -6.14743948e-01
-1.66207528e+00 5.05299151e-01 -4.77429599e-01 -4.76385742e-01
2.28881687e-01 -4.34947699e-01 2.41890401e-02 -5.84966391e-02
-2.61045218e-01 7.63255715e-01 3.10445964e-01 -6.05511963e-01
-2.58359939e-01 -3.26329559e-01 -1.40526801e-01 6.49918169e-02
1.40944824e-01 1.34890422e-01 -3.57937276e-01 -7.22374618e-01
-1.26919836e-01 -1.14724267e+00 -5.11367261e-01 -2.38854349e-01
-4.40897942e-01 -1.96350008e-01 5.21519780e-02 -8.25013340e-01
1.19752407e+00 -1.71356130e+00 5.63578665e-01 -5.64838247e-03
2.18069136e-01 6.27597451e-01 -3.11249912e-01 6.75771654e-01
-2.60582417e-01 1.62702259e-02 -5.03613949e-01 3.64437610e-01
-2.30340391e-01 -2.23831892e-01 -1.99166909e-01 2.84816593e-01
4.43345189e-01 1.21110928e+00 -7.66188085e-01 4.69676219e-02
-3.90256532e-02 6.51014984e-01 -7.73997009e-01 3.30753386e-01
-6.96288228e-01 6.09372020e-01 -7.39298761e-01 6.31952465e-01
5.42886734e-01 -7.70384848e-01 7.02179730e-01 -2.75745988e-01
-4.66173291e-02 3.91821295e-01 -4.68759716e-01 1.71454751e+00
3.70316096e-02 8.69780779e-02 -2.43932039e-01 -7.34653652e-01
1.05814958e+00 2.08815500e-01 4.95059341e-01 -4.92174149e-01
9.78497714e-02 1.76582441e-01 2.20148906e-01 -2.64661759e-01
2.25653827e-01 -5.71358427e-02 1.42353714e-01 1.89871684e-01
-2.22987130e-01 2.38062799e-01 8.74622986e-02 8.56521726e-03
1.29310298e+00 6.00269258e-01 3.58608931e-01 -3.54566544e-01
3.80359769e-01 2.18520731e-01 1.07140052e+00 3.89199674e-01
-4.86742631e-02 5.77344418e-01 5.54347277e-01 -6.45684183e-01
-1.34209633e+00 -6.11637414e-01 1.06782228e-01 1.25511086e+00
-2.59653814e-02 -6.54039264e-01 -7.45406747e-01 -3.61814767e-01
-2.70109833e-03 6.33211076e-01 -1.93298548e-01 -4.43213314e-01
-9.25215840e-01 -1.21642244e+00 7.20938444e-01 3.94769073e-01
1.05213799e-01 -1.35003376e+00 -3.37624907e-01 4.55441952e-01
-3.10022682e-01 -6.38914466e-01 -8.09116781e-01 3.18899244e-01
-7.38416791e-01 -9.98398602e-01 -9.00726676e-01 -8.11397970e-01
5.76291561e-01 1.57774478e-01 1.12035167e+00 8.24321136e-02
-3.59208018e-01 -7.17476130e-01 -3.38691175e-01 -3.34106028e-01
-3.40000391e-01 3.00906122e-01 1.64481372e-01 -5.28004110e-01
7.45808065e-01 -7.44438827e-01 -9.57779229e-01 4.57876861e-01
-8.36558759e-01 4.31295127e-01 8.07882488e-01 1.09389293e+00
8.40882778e-01 -6.06835186e-01 7.51332998e-01 -9.72728074e-01
4.84844059e-01 -4.18438464e-01 -8.83234918e-01 3.35446805e-01
-7.10184097e-01 5.86965680e-01 1.00204444e+00 -3.89946878e-01
-8.61930668e-01 3.61665457e-01 -5.75211167e-01 -4.46847171e-01
1.17959231e-02 4.06407386e-01 -4.12536442e-01 2.62839407e-01
7.55547047e-01 5.87382257e-01 2.14415535e-01 -7.08757460e-01
3.17725897e-01 4.07782316e-01 2.39189610e-01 -6.95047081e-01
3.11951309e-01 -8.97210911e-02 -3.36044103e-01 -5.08444130e-01
-6.36474133e-01 -2.00291261e-01 -2.44757846e-01 4.29290652e-01
7.40797758e-01 -1.03033900e+00 -1.27303231e+00 5.06222010e-01
-9.99419034e-01 -4.91233945e-01 2.59260178e-01 3.48694205e-01
-7.86455095e-01 5.19243479e-01 -9.01141405e-01 -4.03110385e-01
-9.70202506e-01 -1.47524858e+00 8.65208685e-01 1.35198295e-01
-2.90261120e-01 -1.33050904e-01 1.57206599e-02 5.04996777e-01
2.24748999e-01 2.68219132e-02 1.11549187e+00 -4.72633421e-01
-5.73813975e-01 2.52022773e-01 -2.14825258e-01 2.12883726e-01
5.13658375e-02 -9.44119617e-02 -5.88699460e-01 -4.05626237e-01
-4.26860303e-01 -4.57465172e-01 8.39816928e-01 4.51104313e-01
1.29302168e+00 -4.35869575e-01 -3.61691892e-01 7.90839314e-01
1.09455931e+00 4.90388453e-01 8.25621486e-01 1.24169193e-01
6.82603121e-01 2.58293957e-01 6.19197309e-01 5.30812681e-01
3.24302390e-02 7.74936438e-01 1.50348827e-01 -8.25520232e-03
1.70377627e-01 -4.24815893e-01 4.83802110e-01 8.09059680e-01
-1.45221159e-01 -3.41523916e-01 -8.44167173e-01 1.12879433e-01
-1.95468557e+00 -9.01046515e-01 3.09022386e-02 1.97164476e+00
1.34489572e+00 -2.02326044e-01 2.92881548e-01 -4.72506613e-01
6.41092360e-01 -9.65319797e-02 -1.18592250e+00 -6.56529739e-02
-2.33840510e-01 2.89485812e-01 7.21750915e-01 2.63102949e-01
-7.46408284e-01 1.24281347e+00 6.22977495e+00 9.37506974e-01
-8.73954833e-01 -3.70132089e-01 9.32126939e-01 -2.68324465e-01
-2.63834834e-01 1.48095399e-01 -1.06027687e+00 7.20814109e-01
1.12303066e+00 -1.93693325e-01 5.00726402e-01 8.50618958e-01
4.55099076e-01 5.49717784e-01 -8.46682906e-01 8.25621724e-01
-3.84496212e-01 -1.60565543e+00 5.35618439e-02 6.79975227e-02
3.86664152e-01 1.84801549e-01 -8.62204283e-02 3.03905815e-01
6.22866333e-01 -1.17011905e+00 2.91070461e-01 6.56235814e-01
7.68317938e-01 -9.28461850e-01 4.32048559e-01 3.55264246e-01
-8.43665659e-01 2.68276900e-01 -7.28339612e-01 1.34113938e-01
1.66832328e-01 5.13301551e-01 -7.74381280e-01 3.68608922e-01
5.65313876e-01 7.31702626e-01 -1.11779980e-01 7.11580336e-01
-2.05555186e-01 7.99224794e-01 -3.30883652e-01 -2.71412909e-01
-2.65583843e-02 -5.71998477e-01 3.33443254e-01 1.02398288e+00
1.59874782e-01 6.03341877e-01 1.80199966e-01 1.09670591e+00
-3.09052110e-01 1.23102084e-01 -1.02955498e-01 -4.12428945e-01
5.42483985e-01 9.79827881e-01 -1.86000526e-01 -2.19743147e-01
-1.21868320e-01 9.29916620e-01 5.45878947e-01 3.93629134e-01
-1.11520338e+00 -5.30071437e-01 1.15824282e+00 -8.33706523e-04
3.91827822e-01 -1.46818325e-01 1.90695301e-01 -1.24102926e+00
3.41542587e-02 -1.57515228e+00 1.24850564e-01 -7.17864871e-01
-1.31730819e+00 4.63536173e-01 -4.91050512e-01 -8.08623970e-01
5.98582923e-02 -6.62828326e-01 -2.10297495e-01 1.03468716e+00
-1.12968230e+00 -8.60329509e-01 2.29509082e-02 -9.84884426e-02
6.10575974e-01 -9.83082652e-02 8.30110013e-01 1.82088375e-01
-9.42124307e-01 8.60823512e-01 4.27974194e-01 -1.49385691e-01
6.98967934e-01 -9.18938935e-01 1.02424586e+00 5.74902475e-01
-4.13574576e-01 1.06955075e+00 8.69413316e-01 -7.05049992e-01
-1.81672573e+00 -1.13232076e+00 6.51360393e-01 -4.15375501e-01
3.91122282e-01 -3.44388098e-01 -1.28270674e+00 5.72356820e-01
-6.90398589e-02 -4.80081677e-01 6.87545061e-01 -8.66803303e-02
-2.69678712e-01 1.76872998e-01 -8.60598207e-01 8.50452483e-01
1.31523192e+00 -2.32546419e-01 -1.68984130e-01 4.28494185e-01
1.09640694e+00 -3.74321938e-01 -1.14374232e+00 4.92588162e-01
5.97524524e-01 -6.88420355e-01 1.08169699e+00 -9.25515771e-01
4.74351346e-01 -4.52259779e-01 8.23703706e-02 -8.71129334e-01
-7.74292231e-01 -9.54027057e-01 -2.79983938e-01 8.30743968e-01
7.97860205e-01 -6.63492322e-01 9.82666433e-01 4.77576137e-01
-1.75059870e-01 -1.16698933e+00 -4.67434049e-01 -5.81756175e-01
3.24766576e-01 1.06657602e-01 8.73334527e-01 4.27857071e-01
1.45271182e-01 6.38683558e-01 -8.10564101e-01 -2.04869539e-01
1.58261940e-01 2.34509617e-01 6.71590388e-01 -5.99213123e-01
-6.89910412e-01 -2.54163086e-01 7.26843923e-02 -1.60027802e+00
3.71466577e-02 -8.31882775e-01 2.24800140e-01 -1.14561951e+00
5.01229882e-01 -3.00071269e-01 -3.62403125e-01 6.90067708e-01
-6.73898160e-01 -1.26045078e-01 5.39580472e-02 4.18709934e-01
-5.10979772e-01 8.87239635e-01 1.22556496e+00 7.99607933e-02
-2.14781940e-01 -1.58354118e-01 -1.00560343e+00 2.34146118e-01
1.13898170e+00 -3.18122715e-01 -1.92519248e-01 -2.97424823e-01
2.96638936e-01 1.19861014e-01 -1.57070413e-01 -6.57717466e-01
-2.42483452e-01 -3.24805230e-01 4.28259552e-01 -5.23194015e-01
3.35274547e-01 -2.25045800e-01 6.69251680e-01 6.04048789e-01
-5.19296169e-01 3.36757809e-01 2.10778698e-01 5.67151368e-01
2.19880357e-01 5.25771976e-02 8.25154305e-01 -3.44544590e-01
-2.74565786e-01 6.28607750e-01 -4.18084711e-01 -1.14783138e-01
8.35673034e-01 2.02382252e-01 -3.11834782e-01 -1.15318038e-01
-5.42919040e-01 3.62206429e-01 7.73958504e-01 3.37853372e-01
5.25212765e-01 -8.74298513e-01 -8.67499530e-01 7.00738207e-02
1.53903559e-01 -7.83701763e-02 2.17599675e-01 7.14850008e-01
-8.26442420e-01 7.44792938e-01 -1.43101960e-01 -2.72074968e-01
-1.26585221e+00 7.44813502e-01 4.81883526e-01 -1.63809180e-01
-6.03051186e-01 7.28724539e-01 5.18372178e-01 -4.83154416e-01
-1.05332591e-01 -3.18459272e-02 8.51971805e-02 -5.17541349e-01
6.79026663e-01 2.49362022e-01 -8.10118541e-02 -3.46820951e-01
-2.84258813e-01 2.96940953e-01 -4.92901176e-01 6.63389325e-01
1.49106443e+00 2.84829825e-01 -2.26868875e-02 -1.44181415e-01
9.25306439e-01 -4.62892324e-01 -1.40710568e+00 -1.66505381e-01
6.16070516e-02 -1.57923892e-01 -4.55102265e-01 -9.32481050e-01
-8.64348948e-01 6.00508988e-01 5.07148266e-01 -6.22466922e-01
9.64067042e-01 -1.56668335e-01 1.19607306e+00 5.95691621e-01
4.29938585e-01 -5.89279413e-01 -1.78355768e-01 5.71385920e-01
7.67218173e-01 -1.08590007e+00 1.11081926e-02 -2.85535723e-01
-7.74420619e-01 7.61894464e-01 6.70245349e-01 -1.06929936e-01
3.44915465e-02 1.13512121e-01 -5.95923029e-02 -1.89268261e-01
-1.15599430e+00 4.70897779e-02 -1.78125054e-02 3.16099077e-01
9.19047236e-01 1.24453411e-01 -4.92859185e-01 7.18477726e-01
-1.00829579e-01 2.05110624e-01 2.87837144e-02 8.98720622e-01
-4.56120968e-01 -1.35752583e+00 8.27969313e-02 4.57043886e-01
-7.65135705e-01 -4.38574135e-01 -5.16416192e-01 1.13929600e-01
-3.66926223e-01 7.37119257e-01 -4.65230435e-01 -3.06950420e-01
3.60749424e-01 1.88091055e-01 4.01222229e-01 -3.83544415e-01
-6.77987874e-01 1.24840558e-01 7.93965161e-02 -5.82948983e-01
1.02412291e-01 -4.63972658e-01 -1.59738290e+00 -5.20951331e-01
-4.37070280e-01 3.72494429e-01 2.38088757e-01 4.73896116e-01
1.12621486e+00 4.55269665e-01 3.61408174e-01 -5.41743219e-01
-7.79047668e-01 -9.88580465e-01 -2.41633788e-01 2.15442747e-01
-9.21726152e-02 -3.63706648e-01 5.05843200e-02 1.58779994e-01] | [4.696526050567627, 5.645231246948242] |
6be0600a-d111-4014-a744-6d81b35e219b | using-satellite-image-classification-and | 1903.04347 | null | http://arxiv.org/abs/1903.04347v1 | http://arxiv.org/pdf/1903.04347v1.pdf | Using satellite image classification and digital terrain modelling to assess forest species distribution on mountain slopes.A case study in Varatec Forest District | The relation between ecological conditions and geomorphological factors is
considered the basis for species distribution in Romania. In this context, the
location of each species within parts of the mountain slopes is difficult on a
medium to brad scale level. The paper presents methodology to combine
vegetation data, obtained from IKONOS satellite images, and Digital Elevation
Model obtained from digitized topographic maps. The study area is a northern
slope of the Stanisoarei Mountains with a gradient of species from beech mixed
and coniferous stands. | [] | 2019-03-11 | null | null | null | null | ['satellite-image-classification'] | ['computer-vision'] | [ 4.98839989e-02 -6.11367106e-01 -1.72359839e-01 -1.28748685e-01
2.22574756e-01 -3.58802289e-01 4.08116668e-01 2.71487176e-01
-7.17671096e-01 1.28774405e+00 -3.28185186e-02 -7.19088376e-01
-2.92605996e-01 -1.38170660e+00 -7.41849095e-02 -4.81597841e-01
-4.39318031e-01 2.71859944e-01 2.33092442e-01 -7.24494874e-01
2.40452021e-01 8.46417904e-01 -1.59408092e+00 -2.80159771e-01
1.02611625e+00 3.37332726e-01 7.74385870e-01 6.65819943e-01
-1.51932359e-01 -2.77705967e-01 -3.52629513e-01 1.75764069e-01
2.42314026e-01 -8.83581489e-02 -3.49447906e-01 2.67616287e-03
4.16433364e-01 -5.26594639e-01 9.90171917e-03 1.01249635e+00
9.22587886e-02 -2.01364234e-01 9.36209619e-01 -4.97886717e-01
-2.81658530e-01 5.48138916e-01 -8.65788043e-01 1.83480680e-01
-1.20528512e-01 -1.22166738e-01 7.90374160e-01 -9.96518016e-01
7.79056728e-01 9.98982966e-01 4.94757503e-01 -4.23969895e-01
-1.25777519e+00 -4.62902993e-01 1.54070631e-01 3.66313756e-01
-2.18961883e+00 1.66984927e-02 2.96688050e-01 -6.30315363e-01
5.26763558e-01 3.38092238e-01 9.80765700e-01 1.94907427e-01
5.59794068e-01 -1.19771384e-01 1.46325421e+00 -4.83976245e-01
3.33525449e-01 -6.55563995e-02 -2.32141674e-01 1.95431620e-01
9.47094917e-01 -3.68163995e-02 -3.58203709e-01 -2.71087855e-01
9.88851249e-01 5.78142926e-02 -1.61706567e-01 -5.30096814e-02
-2.89087057e-01 4.80071813e-01 8.73034358e-01 6.76021934e-01
-6.69326186e-01 -2.02016756e-01 2.23098204e-01 1.71433434e-01
2.00893611e-01 1.57862186e-01 -3.67150307e-01 -1.21832266e-01
-1.38023603e+00 2.48238131e-01 4.63835150e-01 4.08020437e-01
6.75933719e-01 2.27229819e-01 8.71797740e-01 9.10128534e-01
4.89983201e-01 8.43569577e-01 -2.76500918e-02 -3.36452305e-01
1.88140586e-01 7.38518715e-01 2.98282266e-01 -1.15458381e+00
-7.00699463e-02 -3.73415411e-01 -7.19042003e-01 6.35259986e-01
3.81913096e-01 -9.96882766e-02 -8.34813952e-01 1.18216407e+00
1.91382736e-01 -3.67383987e-01 -6.69556931e-02 9.24271047e-01
2.55020946e-01 5.79589903e-01 5.10063767e-01 -2.84506202e-01
1.28087234e+00 -6.37918338e-02 -5.51444829e-01 8.04261118e-02
6.32862896e-02 -4.13745910e-01 1.15858388e+00 1.57941058e-01
-7.23278284e-01 -2.77223408e-01 -1.24243689e+00 4.26986992e-01
-7.95112073e-01 2.50109911e-01 2.77099162e-01 6.34165227e-01
-1.07516718e+00 5.42343497e-01 -6.72516823e-01 -6.44850373e-01
7.80813545e-02 2.45069444e-01 -4.74668264e-01 4.57687438e-01
-1.14115405e+00 1.10901237e+00 3.73392344e-01 7.45064199e-01
-1.33983448e-01 -1.63588881e-01 -5.74245811e-01 1.00085653e-01
1.17904472e-03 -3.56142968e-02 5.48385382e-01 -9.63198185e-01
-1.04833341e+00 8.48138213e-01 -3.18406671e-02 -1.78597495e-01
6.47419453e-01 9.29603260e-03 -6.28619492e-01 9.76528078e-02
3.70697975e-01 -2.69164667e-02 1.21239617e-01 -1.32139289e+00
-1.11713791e+00 -8.35052073e-01 -3.12546074e-01 3.33756089e-01
7.31055960e-02 -4.23500827e-03 4.52392876e-01 -5.32376051e-01
3.72089595e-01 -6.36689126e-01 -6.20713711e-01 -3.62679422e-01
1.30845845e-01 3.27254206e-01 7.26921976e-01 -8.36044252e-01
1.30226147e+00 -1.82831776e+00 -2.41730168e-01 6.55321598e-01
-9.75274667e-02 4.96950001e-02 4.85486239e-01 7.39269376e-01
2.05529079e-01 4.38233525e-01 -4.49443072e-01 9.92037058e-01
-3.70678157e-01 5.58767796e-01 -1.30580023e-01 3.19723964e-01
-9.17630270e-02 2.44580239e-01 -8.59490871e-01 -6.06469154e-01
3.39371443e-01 2.53515750e-01 1.77097723e-01 -3.44310433e-01
4.69224036e-01 -2.04164803e-01 -5.29550612e-01 1.01924348e+00
1.15118039e+00 6.63847923e-01 4.26229149e-01 4.53101516e-01
-1.15661752e+00 1.90631315e-01 -1.16558576e+00 7.37944901e-01
-4.22577918e-01 6.30077004e-01 5.59229076e-01 -4.38623607e-01
1.19976556e+00 8.85101259e-02 8.91321599e-02 -2.88197786e-01
-1.44414142e-01 6.13840640e-01 2.72860467e-01 -3.77715319e-01
1.07068980e+00 -2.77564496e-01 1.07930958e-01 -4.16991562e-01
-3.69711250e-01 -2.36372426e-01 2.79264331e-01 -2.95639843e-01
3.11974019e-01 2.47038469e-01 9.27500308e-01 -9.58978772e-01
3.40031832e-01 3.80825222e-01 4.77275729e-01 3.29852611e-01
-1.57548696e-01 3.85812931e-02 4.89868671e-01 -4.98623312e-01
-1.31240988e+00 -1.48324788e+00 -1.19306648e+00 9.55474496e-01
2.70029813e-01 -5.46379685e-02 -4.14249539e-01 1.20761000e-01
2.47732982e-01 4.13392454e-01 -5.63657939e-01 4.93911684e-01
-4.01322454e-01 -8.40039253e-01 2.87448376e-01 3.88832897e-01
8.45938087e-01 -9.00085807e-01 -6.83384895e-01 2.91876107e-01
1.37850404e-01 -3.49139571e-01 2.60795385e-01 2.90590793e-01
-1.15566206e+00 -6.86675787e-01 -5.13260543e-01 -6.48735702e-01
5.02344429e-01 2.14303136e-01 8.73920023e-01 9.74538326e-02
-4.11809295e-01 -2.91995525e-01 -2.65901923e-01 -3.19769084e-01
-2.36612111e-01 3.92020166e-01 -4.73237000e-02 -4.54848289e-01
2.66657084e-01 -1.20123649e+00 -3.04904163e-01 5.72897553e-01
-5.43668389e-01 -1.56247810e-01 5.21641433e-01 2.60604650e-01
4.84523773e-01 2.83123076e-01 6.72958732e-01 -6.74284995e-01
8.83165970e-02 -7.02911973e-01 -7.12853134e-01 1.41185090e-01
-4.67053056e-01 -4.75135803e-01 5.63569844e-01 2.05681235e-01
-8.31044674e-01 2.19044924e-01 4.36111121e-03 5.80770075e-01
-6.54323220e-01 9.08978820e-01 -2.96741158e-01 -8.92922878e-02
8.84554803e-01 2.65535772e-01 -2.02298686e-01 -3.41056526e-01
-4.75515537e-02 1.23751950e+00 9.00880754e-01 -5.07819474e-01
7.29250550e-01 4.09669399e-01 9.57886055e-02 -1.68976068e+00
1.50635496e-01 -4.27434593e-01 -1.24770284e+00 -4.94810998e-01
5.48387527e-01 -8.14960539e-01 -3.69913906e-01 1.27167836e-01
-5.42981446e-01 -1.45132005e-01 -6.81215599e-02 5.89589596e-01
-2.13079318e-01 -6.11832365e-02 -4.02051121e-01 -1.04874253e+00
-8.81345719e-02 -5.30558348e-01 2.87034243e-01 4.30730760e-01
-9.51794013e-02 -9.39382851e-01 2.39192516e-01 -2.73201048e-01
3.75923872e-01 6.53858364e-01 8.17846179e-01 1.77416757e-01
-2.78223306e-01 -3.05007070e-01 -3.07557046e-01 -1.28788590e-01
5.44296026e-01 7.18062937e-01 -4.79944974e-01 1.91463187e-01
-2.09266439e-01 2.65216649e-01 8.21353257e-01 4.19506997e-01
2.91677684e-01 8.40321109e-02 -2.96015352e-01 2.61928618e-01
2.21540880e+00 5.85408390e-01 6.19880676e-01 7.93594480e-01
2.35559031e-01 6.26882851e-01 6.70771599e-01 4.00617361e-01
8.44266042e-02 4.56632435e-01 5.54930627e-01 -1.06751010e-01
3.15139621e-01 1.11769997e-02 3.76655728e-01 3.97394121e-01
-7.56121516e-01 4.17011827e-01 -1.34527826e+00 8.80619407e-01
-1.47764838e+00 -1.04630637e+00 -7.09341228e-01 2.39487219e+00
4.88925010e-01 -2.07553655e-01 7.95477405e-02 1.00768745e-01
9.99396443e-01 4.72935289e-02 -9.32946652e-02 -6.23747647e-01
-4.00871545e-01 1.82774678e-01 1.14444232e+00 9.21356201e-01
-8.27403843e-01 1.08906221e+00 7.25267506e+00 3.73984188e-01
-1.39800310e+00 -4.48569059e-01 3.88162099e-02 2.09751830e-01
-1.27940074e-01 1.27400219e-01 -7.84251690e-01 2.35621661e-01
9.27479327e-01 -6.42101049e-01 1.58387303e-01 6.12998307e-01
9.24516022e-01 -7.23238289e-01 1.26645714e-01 2.13810444e-01
-4.08857763e-01 -8.02284539e-01 -9.10471752e-02 4.43351775e-01
5.50678968e-01 2.17985123e-01 -1.68323532e-01 -7.45751113e-02
3.76152664e-01 -7.72757828e-01 5.13678849e-01 1.81711331e-01
9.45652485e-01 -5.97793281e-01 5.14614642e-01 1.74690500e-01
-1.82288897e+00 4.55131307e-02 -5.61887860e-01 -6.79835021e-01
2.30585724e-01 2.21832201e-01 -8.76561701e-01 7.25369930e-01
8.16071749e-01 4.59543020e-01 -6.64557576e-01 1.26973939e+00
-4.44045216e-01 7.98490942e-01 -5.92072546e-01 -1.67072359e-02
5.57004929e-01 -9.09606636e-01 3.84063363e-01 9.18679237e-01
4.04227316e-01 3.55198532e-02 2.73376346e-01 4.70973194e-01
3.62754166e-01 7.16709077e-01 -9.49090958e-01 3.62303585e-01
6.27826333e-01 1.14005756e+00 -8.35985780e-01 -2.47535527e-01
-3.55240367e-02 3.21427047e-01 -5.46922907e-02 2.05405593e-01
-3.43394071e-01 -4.73694921e-01 3.72510284e-01 8.21326673e-01
-8.43398739e-03 -7.51776159e-01 -6.50289893e-01 -7.90123284e-01
1.51628731e-02 -1.79657966e-01 3.40580106e-01 -4.85546231e-01
-3.68167847e-01 5.14026165e-01 3.43703300e-01 -1.04410768e+00
-1.90445766e-01 -5.13445139e-01 -6.19394124e-01 1.40252483e+00
-1.20001638e+00 -1.08818400e+00 -3.90572786e-01 4.00487669e-02
2.11207643e-01 -9.39157084e-02 1.00644493e+00 -1.86703280e-01
-2.53941923e-01 -4.26188231e-01 9.71850574e-01 -2.18518391e-01
1.69343099e-01 -1.41375196e+00 2.51873940e-01 5.81298888e-01
-4.77024227e-01 2.93301225e-01 6.32241130e-01 -1.18284583e+00
-4.55208004e-01 -9.31913495e-01 1.19519639e+00 4.11851615e-01
9.50200558e-01 -1.02350682e-01 -7.76266992e-01 3.04277539e-01
1.94856003e-01 -4.45288807e-01 6.23394549e-01 5.69236398e-01
1.10522501e-01 -2.04921693e-01 -1.10914779e+00 9.94533956e-01
8.59422565e-01 -2.31042162e-01 -2.40619257e-01 1.71249695e-02
-4.52018797e-01 3.33794147e-01 -9.62692142e-01 2.45377406e-01
1.12591994e+00 -6.52405083e-01 5.21978498e-01 -1.01310506e-01
5.09078145e-01 -4.28330958e-01 -4.74491566e-01 -1.42581654e+00
-4.88519311e-01 8.43237899e-03 1.44919240e+00 9.97725308e-01
5.04494965e-01 -4.85602558e-01 7.10299194e-01 2.74225205e-01
2.45591059e-01 -8.47573578e-02 -8.42712522e-01 -9.07666564e-01
3.42738032e-01 1.16367579e-01 1.21946022e-01 5.04722476e-01
3.57285365e-02 1.73041388e-01 -9.88564342e-02 3.95319819e-01
4.11604166e-01 1.64600223e-01 6.58599913e-01 -1.61650074e+00
4.46803302e-01 -5.01824737e-01 -7.65942454e-01 -1.55442059e-01
-1.29676759e-01 -4.47561055e-01 -9.08214971e-02 -1.59550524e+00
-1.50798798e-01 -6.23795986e-01 7.25302175e-02 1.40283003e-01
-1.04892641e-01 9.22521874e-02 2.78978258e-01 4.00997013e-01
4.53179806e-01 3.26509088e-01 6.70216084e-01 6.07807413e-02
-6.26080215e-01 1.05783427e-02 -3.21990818e-01 8.73236179e-01
1.09549153e+00 -3.14447284e-01 -8.90139118e-02 -5.99159375e-02
4.91437651e-02 1.36140093e-01 1.10150017e-01 -1.03757429e+00
-1.98578417e-01 -7.50398517e-01 9.44370180e-02 -9.05902684e-01
7.25023225e-02 -1.16244686e+00 5.25688827e-01 8.68322551e-01
1.84434608e-01 2.53437553e-02 2.85025656e-01 2.44007140e-01
-7.14913905e-02 -3.02993715e-01 9.51507151e-01 -1.26700640e-01
-9.12189543e-01 8.64711702e-02 -9.42338824e-01 -5.46334028e-01
9.12278831e-01 -7.40657449e-01 3.04201227e-02 -1.72753498e-01
-7.47545540e-01 4.28932980e-02 9.40414071e-01 -3.64220259e-03
1.44573882e-01 -1.08647633e+00 -9.56879675e-01 -1.36002451e-02
3.15259248e-02 -4.54479277e-01 3.09275419e-01 5.01269639e-01
-1.58565998e+00 1.27293408e-01 -1.28063524e+00 -3.46133620e-01
-1.53951919e+00 3.89219844e-03 5.03608942e-01 1.35404468e-01
-4.84347820e-01 -2.04541869e-02 6.94243982e-02 -1.62525386e-01
-3.52520347e-01 -9.68895387e-03 -7.39190042e-01 2.75253326e-01
4.25763816e-01 5.36493421e-01 -1.35827407e-01 -1.02138448e+00
-4.20361966e-01 5.94160795e-01 3.26676369e-01 -4.55637664e-01
1.15412402e+00 -3.37180942e-01 -2.73014635e-01 1.02713883e+00
4.95269984e-01 1.86097309e-01 -7.67763674e-01 2.44233012e-01
9.25545543e-02 -9.15189445e-01 1.06981844e-01 -6.80000067e-01
-5.36292076e-01 6.72675252e-01 6.88740015e-01 3.31005871e-01
1.04445660e+00 -6.37390137e-01 -1.31267339e-01 3.79852653e-01
3.82820398e-01 -1.18209159e+00 -1.33064735e+00 1.97845131e-01
1.05204415e+00 -7.70443201e-01 1.45690739e-01 -4.41104859e-01
-4.56154227e-01 1.23338878e+00 4.15028840e-01 -2.98211515e-01
9.29220498e-01 3.53964597e-01 1.62262559e-01 2.29498848e-01
-5.71292818e-01 -7.31229603e-01 -4.72708285e-01 5.68726242e-01
6.93285644e-01 7.05961347e-01 -1.17691612e+00 5.75165525e-02
-5.91683328e-01 8.58770385e-02 8.97793949e-01 9.32537198e-01
-1.02782035e+00 -9.01088834e-01 -8.77373576e-01 7.67907202e-01
-2.99915344e-01 -2.57767439e-01 -5.28403401e-01 1.24231315e+00
1.72688231e-01 6.03860736e-01 2.33703464e-01 -6.61209822e-02
2.24268481e-01 -3.42962652e-01 1.07129551e-01 -1.91190347e-01
-5.90978980e-01 3.98141444e-01 4.83405329e-02 3.42369974e-01
-2.04568043e-01 -1.00673032e+00 -1.12002552e+00 -8.28734934e-01
-2.62843609e-01 5.31935394e-01 1.18591881e+00 5.85727274e-01
-2.67704755e-01 -1.20780937e-01 8.21999252e-01 -8.35008502e-01
1.19889386e-01 -1.22307241e+00 -1.59879255e+00 -3.96413416e-01
-2.71587878e-01 -6.52158439e-01 -2.26954222e-01 -1.07765928e-01] | [9.408945083618164, -1.599672794342041] |
6683c623-6975-4df2-834d-3e0c46026084 | a-majorization-minimization-algorithm-for-1 | 2204.09741 | null | https://arxiv.org/abs/2204.09741v1 | https://arxiv.org/pdf/2204.09741v1.pdf | A majorization-minimization algorithm for nonnegative binary matrix factorization | This paper tackles the problem of decomposing binary data using matrix factorization. We consider the family of mean-parametrized Bernoulli models, a class of generative models that are well suited for modeling binary data and enables interpretability of the factors. We factorize the Bernoulli parameter and consider an additional Beta prior on one of the factors to further improve the model's expressive power. While similar models have been proposed in the literature, they only exploit the Beta prior as a proxy to ensure a valid Bernoulli parameter in a Bayesian setting; in practice it reduces to a uniform or uninformative prior. Besides, estimation in these models has focused on costly Bayesian inference. In this paper, we propose a simple yet very efficient majorization-minimization algorithm for maximum a posteriori estimation. Our approach leverages the Beta prior whose parameters can be tuned to improve performance in matrix completion tasks. Experiments conducted on three public binary datasets show that our approach offers an excellent trade-off between prediction performance, computational complexity, and interpretability. | ['Cédric Févotte', 'Paul Magron'] | 2022-04-20 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [ 4.35488790e-01 2.95151561e-01 -4.14078355e-01 -5.85759163e-01
-7.87934244e-01 -5.29933393e-01 6.93656921e-01 1.94222294e-02
-4.05502528e-01 5.52692533e-01 1.79823130e-01 -5.84662497e-01
-3.09467942e-01 -5.39437890e-01 -7.41032481e-01 -6.82342112e-01
8.36579874e-03 6.17574811e-01 -1.40928179e-01 4.54760864e-02
1.67572908e-02 1.79730132e-01 -1.12691891e+00 3.41079198e-02
8.90361905e-01 1.00883806e+00 1.21719211e-01 5.81376374e-01
3.44433635e-01 5.96452892e-01 -4.94027466e-01 -8.95686746e-01
5.37728965e-01 -8.18513408e-02 -5.93808353e-01 1.69286042e-01
1.98285609e-01 -3.96027952e-01 -2.26716712e-01 1.03465652e+00
1.60208866e-01 -6.59036934e-02 9.61896718e-01 -1.27682972e+00
-3.46767992e-01 9.10258293e-01 -8.41766953e-01 9.43878815e-02
1.16244666e-01 -1.43931016e-01 1.40064824e+00 -6.72485769e-01
2.22344995e-01 1.34762037e+00 7.81087756e-01 3.86316925e-02
-1.73209167e+00 -5.94649017e-01 1.28649682e-01 2.78962720e-02
-1.55021954e+00 -4.86247092e-01 5.07023096e-01 -5.40825009e-01
3.90173048e-01 3.10393602e-01 3.84388387e-01 1.23837757e+00
2.00259537e-01 7.83750951e-01 1.03196764e+00 -4.66496795e-01
3.79357636e-01 1.91152036e-01 2.79369920e-01 4.03349727e-01
5.27561545e-01 -1.40185624e-01 -5.23364305e-01 -6.20178521e-01
6.84450924e-01 1.80143386e-01 -1.93525478e-01 -6.01553261e-01
-9.66700554e-01 1.05782163e+00 1.00856021e-01 -1.01036765e-01
-3.71231884e-01 3.63152713e-01 1.11802787e-01 4.57768440e-02
4.37149525e-01 1.33707583e-01 -2.91166931e-01 -9.03593376e-02
-1.01455522e+00 2.15652764e-01 9.05509293e-01 9.03841853e-01
8.13309729e-01 -2.45082483e-01 -3.93248469e-01 8.67066622e-01
5.64860284e-01 5.36250591e-01 -1.40111465e-02 -8.67232561e-01
5.51271319e-01 2.06387848e-01 2.20336467e-01 -1.13976061e+00
-1.08216032e-01 -5.08180618e-01 -9.87103760e-01 -1.62855059e-01
5.24284422e-01 2.32703723e-02 -9.81213748e-01 1.86896646e+00
3.23068500e-01 2.12536380e-01 -2.26093978e-01 7.77149856e-01
2.26375252e-01 4.41262394e-01 -2.29185745e-02 -1.45431086e-01
1.53186011e+00 -6.56805098e-01 -6.97090745e-01 -2.95926541e-01
2.45677575e-01 -7.25658059e-01 9.39949930e-01 7.27784216e-01
-8.30951273e-01 -1.42995775e-01 -1.01520634e+00 -1.23789413e-02
1.79456741e-01 3.32609922e-01 1.02127755e+00 1.29552042e+00
-7.07818985e-01 4.32440788e-01 -1.24280739e+00 -2.47518066e-02
3.08649033e-01 4.14691776e-01 -1.53776973e-01 -2.09724650e-01
-9.08501506e-01 5.80300868e-01 3.65970045e-01 2.71668751e-02
-9.34243679e-01 -7.77711511e-01 -6.07713044e-01 2.95973271e-01
6.38197839e-01 -9.20231938e-01 1.19901526e+00 -5.89216352e-01
-1.47542894e+00 3.93006265e-01 -2.93323487e-01 -7.05133259e-01
5.94806075e-01 -2.16034457e-01 3.34450267e-02 7.33699501e-02
-9.28136110e-02 5.11142313e-01 1.28014457e+00 -1.10623515e+00
-3.44093144e-01 -2.80987680e-01 4.65511948e-01 -1.23481251e-01
-5.21247447e-01 -1.43073320e-01 -8.11174393e-01 -8.83051455e-01
1.89881951e-01 -1.20361924e+00 -4.94939417e-01 -1.75196707e-01
-5.00001132e-01 3.90048102e-02 4.07130495e-02 -6.55939519e-01
1.43801522e+00 -2.12445211e+00 2.04324812e-01 5.02007663e-01
3.03182274e-01 -9.39683989e-02 1.74373403e-01 3.58760864e-01
4.87209149e-02 1.53485402e-01 -4.75615859e-01 -6.74618483e-01
2.53204316e-01 3.79240304e-01 -5.01965106e-01 5.57165802e-01
1.70712531e-01 5.32981396e-01 -5.25810778e-01 -2.06308126e-01
-3.21477950e-02 7.46968985e-01 -1.07010090e+00 7.77990185e-03
-1.41436905e-01 3.03463042e-01 -3.70137781e-01 4.18753177e-01
9.97483015e-01 -5.59098840e-01 6.25847518e-01 -2.91702300e-01
4.35454071e-01 2.79368371e-01 -1.61177015e+00 1.41000926e+00
-4.46977854e-01 2.17100158e-01 1.66482791e-01 -9.48827624e-01
6.67658806e-01 8.25135484e-02 4.23043638e-01 1.48047879e-01
1.69903874e-01 -1.42741129e-01 3.28182057e-02 -9.58766565e-02
6.09380424e-01 -2.38644287e-01 -1.39951289e-01 5.21742761e-01
-1.03748485e-01 7.73100406e-02 4.12779510e-01 3.46031129e-01
1.02275443e+00 4.60044518e-02 4.86826479e-01 -3.36779922e-01
2.49195844e-01 -4.63569164e-01 7.04966545e-01 1.11478984e+00
2.65118480e-01 7.29747713e-01 7.03992307e-01 -8.54523331e-02
-9.98700857e-01 -9.41552639e-01 -4.04884607e-01 1.08444715e+00
-1.37583986e-01 -9.86540556e-01 -6.50033474e-01 -5.05367041e-01
4.96673584e-02 6.86130166e-01 -7.28879035e-01 5.98113462e-02
-1.85318291e-01 -1.35509109e+00 5.29729128e-01 4.79463607e-01
-8.49187747e-03 -6.19697906e-02 -4.20515269e-01 7.04773422e-03
-3.10137004e-01 -1.08534718e+00 -3.25805247e-01 4.35943544e-01
-7.33290017e-01 -9.06235039e-01 -6.31570399e-01 9.10780281e-02
7.48053193e-01 2.55091369e-01 9.91894007e-01 -2.71189630e-01
-1.53205708e-01 4.07600403e-01 -3.84189785e-01 -2.88270295e-01
-4.43980098e-01 7.67548680e-02 -5.65251336e-02 3.18468124e-01
3.26903433e-01 -6.99500144e-01 -4.88190144e-01 2.31007397e-01
-1.28671181e+00 3.13385366e-03 7.27262914e-01 9.57990468e-01
4.95943397e-01 3.00460402e-02 -3.44757847e-02 -9.70698953e-01
5.32397985e-01 -6.69710755e-01 -6.89988375e-01 2.30970398e-01
-7.71890938e-01 4.45329994e-01 4.50843006e-01 -5.52485287e-01
-1.01364756e+00 4.92217615e-02 1.19644463e-01 -5.47193706e-01
5.23290001e-02 8.02567840e-01 -2.55961508e-01 1.14846200e-01
4.35219318e-01 -7.46900812e-02 -1.45213842e-01 -7.09024906e-01
4.40049618e-01 5.77426434e-01 3.30257475e-01 -7.97897577e-01
7.18678653e-01 7.06687689e-01 2.82558560e-01 -5.36433220e-01
-9.01231885e-01 -3.71784985e-01 -4.19449270e-01 3.19642693e-01
5.82510769e-01 -1.04804385e+00 -8.78741086e-01 3.70706618e-02
-8.79843354e-01 6.30777702e-02 8.53005722e-02 4.81775016e-01
-4.93956119e-01 7.63930857e-01 -3.96679282e-01 -8.64898264e-01
-2.33894140e-02 -1.23784947e+00 1.12725687e+00 -2.11553484e-01
-3.37790638e-01 -8.87000501e-01 3.71518359e-02 4.36101735e-01
2.49000654e-01 -1.07904620e-01 8.37463081e-01 -6.06357396e-01
-5.36204457e-01 -1.65863290e-01 -2.44323984e-01 3.35306883e-01
-1.08183429e-01 -1.51847854e-01 -8.46205056e-01 -4.25718278e-01
-6.03861623e-02 -5.32044210e-02 1.19925690e+00 3.46117765e-01
1.27565312e+00 -5.02428591e-01 -4.13392961e-01 7.77752459e-01
1.19158244e+00 -2.03474745e-01 3.77042949e-01 5.66480309e-02
5.37808836e-01 2.70741135e-01 5.35518765e-01 1.06872845e+00
5.39824605e-01 8.56573462e-01 3.84991556e-01 2.21972331e-01
1.98109716e-01 -2.28571370e-01 3.77506554e-01 4.12692666e-01
-6.83181882e-02 -3.05057704e-01 -9.83146667e-01 2.54311621e-01
-1.82673633e+00 -8.32818508e-01 3.57248187e-02 2.40924430e+00
9.18011725e-01 2.04027086e-01 1.55495092e-01 2.07985446e-01
5.93430936e-01 1.36475906e-01 -3.45230609e-01 -6.76689148e-02
-5.37960604e-03 1.53249756e-01 7.67978191e-01 6.91754401e-01
-1.22520304e+00 6.18940294e-01 6.75385046e+00 9.89189029e-01
-8.07232022e-01 1.99246690e-01 7.05384791e-01 -1.98666811e-01
-6.42657757e-01 3.55741173e-01 -1.08986068e+00 6.31577134e-01
9.54278886e-01 3.02184168e-02 4.90123779e-01 8.61469746e-01
-3.92410532e-03 -2.19354942e-01 -1.19698346e+00 9.99612629e-01
9.42954700e-03 -8.89786184e-01 -1.09776616e-01 4.27093983e-01
5.36859810e-01 -2.06961393e-01 3.14059615e-01 1.37280568e-01
5.52459776e-01 -1.03772819e+00 8.03209543e-01 5.09759426e-01
4.78312194e-01 -8.44682932e-01 4.24707949e-01 2.78315544e-01
-7.09471166e-01 -2.11194903e-01 -4.89693433e-01 -7.28863850e-02
2.62664825e-01 1.02904832e+00 -1.01143134e+00 5.01627922e-01
6.32033169e-01 4.36557204e-01 -6.85202301e-01 8.37350547e-01
-2.01942235e-01 9.15859640e-01 -7.21662581e-01 3.23003054e-01
2.34837551e-03 -4.91842896e-01 4.81464267e-01 1.24281669e+00
1.65116727e-01 -1.13871694e-01 3.12870383e-01 7.85928547e-01
1.12176597e-01 4.13579866e-03 -1.02213249e-01 -9.45800468e-02
5.07938445e-01 1.25203228e+00 -7.37419963e-01 -2.69463032e-01
-3.22660625e-01 6.53657734e-01 2.99731910e-01 3.44384611e-01
-1.10537302e+00 1.00988500e-01 6.38054490e-01 1.36838689e-01
8.66576195e-01 -3.04982185e-01 -4.95335251e-01 -1.45063174e+00
1.53411418e-01 -1.07931006e+00 4.89125669e-01 -1.70498610e-01
-1.34128904e+00 3.73836368e-01 4.15881932e-01 -9.51798677e-01
-4.52127337e-01 -6.41125262e-01 -1.47966534e-01 8.42719734e-01
-1.35867512e+00 -1.13780069e+00 -1.50229791e-02 3.94924879e-01
1.55454837e-02 1.32604524e-01 6.17878795e-01 2.78039724e-01
-8.29699337e-01 7.90792346e-01 3.40284318e-01 -3.61509442e-01
8.38240385e-01 -1.12949383e+00 1.63743883e-01 1.03373802e+00
1.46089941e-01 1.10608256e+00 1.17014492e+00 -5.67088246e-01
-1.52289426e+00 -7.23798275e-01 5.64410508e-01 -3.11639607e-01
6.85720265e-01 -4.55623657e-01 -5.70933163e-01 8.65169466e-01
-1.67380542e-01 -1.32190540e-01 1.13220227e+00 6.18130803e-01
-7.06331730e-01 -7.25587681e-02 -8.82375777e-01 5.04164815e-01
7.03193009e-01 -2.89265394e-01 -2.68160164e-01 3.58449072e-01
3.64439070e-01 -2.04447433e-01 -1.03562737e+00 5.52757978e-01
6.63376749e-01 -1.06438780e+00 1.06131268e+00 -6.32590055e-01
4.81799364e-01 -4.03182387e-01 -4.88838792e-01 -1.00300598e+00
-4.20764565e-01 -8.18858325e-01 -3.60337794e-01 1.31091225e+00
3.82249415e-01 -4.48401451e-01 7.15682447e-01 8.63045573e-01
2.38990083e-01 -7.46768773e-01 -8.68697822e-01 -6.68987215e-01
-1.26232520e-01 -8.24636281e-01 6.13037169e-01 6.09665632e-01
-1.71947256e-01 3.18559080e-01 -8.08278501e-01 3.02224159e-01
6.66818738e-01 1.15501590e-01 1.02784956e+00 -1.22834229e+00
-1.11974037e+00 -2.72518575e-01 -3.01780045e-01 -1.50801933e+00
1.88298598e-01 -9.18932855e-01 -1.97540656e-01 -1.02106380e+00
6.11231625e-01 -5.88822424e-01 -1.41230464e-01 5.98713398e-01
-4.48872179e-01 2.78829932e-01 1.85102314e-01 1.92203924e-01
-5.03295541e-01 4.85235929e-01 7.53122151e-01 -8.90216678e-02
5.16960546e-02 2.23210990e-01 -1.01248705e+00 6.90122604e-01
5.06305575e-01 -6.11742437e-01 -5.18340468e-01 -4.24151808e-01
5.71768582e-01 -3.89264859e-02 3.98508281e-01 -4.58699375e-01
1.18786827e-01 -1.48503944e-01 2.96659261e-01 -5.10500550e-01
6.23788953e-01 -9.05086696e-01 4.06945527e-01 1.46202564e-01
-2.23727539e-01 -1.06473230e-01 -1.72058389e-01 9.01132226e-01
-6.79871589e-02 -4.17878389e-01 6.80356503e-01 1.62532970e-01
-3.55090983e-02 2.17526481e-01 -3.52045149e-01 -6.46818057e-02
7.38212287e-01 -4.45887968e-02 5.46284355e-02 -6.56273842e-01
-7.15435684e-01 -8.05845335e-02 4.88175780e-01 2.21821085e-01
3.04470330e-01 -1.19637883e+00 -6.21278167e-01 1.75432041e-01
-4.03326079e-02 -1.67389169e-01 5.79192713e-02 1.14950120e+00
-1.87011719e-01 4.44127411e-01 1.33293480e-01 -6.68987095e-01
-1.09515560e+00 6.23441756e-01 -1.43870041e-01 -6.71993077e-01
-2.38556355e-01 6.80197120e-01 2.47687474e-01 -1.80710673e-01
2.73128539e-01 -2.45856196e-01 1.02163889e-01 1.28896192e-01
5.06527305e-01 4.82446641e-01 2.78689936e-02 -4.92074162e-01
-3.56528282e-01 8.15376341e-02 -2.98188537e-01 -2.80180931e-01
1.36265051e+00 -2.91623294e-01 -1.86883092e-01 3.36565584e-01
8.51228118e-01 1.93249792e-01 -1.28958690e+00 -3.62643987e-01
4.45148861e-03 -6.65664971e-01 1.48852214e-01 -4.60929066e-01
-7.12045074e-01 7.33220935e-01 1.96280465e-01 2.22048387e-01
1.06167865e+00 -1.82428151e-01 4.18500155e-01 3.01614583e-01
3.23240548e-01 -6.72520459e-01 -9.92694423e-02 5.18558919e-01
6.28047168e-01 -1.14430404e+00 2.68721879e-01 -7.00517356e-01
-5.64307570e-01 9.15356219e-01 8.37463513e-02 6.17210157e-02
7.69119859e-01 4.12522942e-01 -2.57844150e-01 3.08983952e-01
-7.43539751e-01 1.59004852e-02 4.99603838e-01 4.16055620e-01
3.57754916e-01 1.47483081e-01 -3.90661210e-01 9.27146316e-01
-4.22564089e-01 -2.32315168e-01 4.90847290e-01 7.27035105e-01
-9.98787060e-02 -1.39025652e+00 -6.68130457e-01 3.94905746e-01
-7.96263933e-01 -3.91682118e-01 -4.20784168e-02 3.86237651e-01
-1.56793281e-01 1.11256313e+00 1.32841431e-02 -2.40431607e-01
-1.43747821e-01 -2.30128262e-02 5.74846089e-01 -5.38743615e-01
-1.96507648e-01 3.43973696e-01 2.97646504e-02 -3.48061979e-01
-3.39226216e-01 -7.52059221e-01 -6.28667414e-01 -3.57399732e-01
-3.97295386e-01 1.50039243e-02 6.88671947e-01 9.96768415e-01
4.55727965e-01 2.73246050e-01 5.24585843e-01 -7.04800785e-01
-1.13803816e+00 -1.09038353e+00 -4.98900324e-01 2.85023004e-01
1.33100033e-01 -8.74934375e-01 -3.34448516e-01 1.90385431e-01] | [7.3534111976623535, 4.411067962646484] |
c382e7dd-ddd3-42ea-8bd3-b02a4de7de83 | auc-maximization-for-low-resource-named | 2212.04800 | null | https://arxiv.org/abs/2212.04800v3 | https://arxiv.org/pdf/2212.04800v3.pdf | AUC Maximization for Low-Resource Named Entity Recognition | Current work in named entity recognition (NER) uses either cross entropy (CE) or conditional random fields (CRF) as the objective/loss functions to optimize the underlying NER model. Both of these traditional objective functions for the NER problem generally produce adequate performance when the data distribution is balanced and there are sufficient annotated training examples. But since NER is inherently an imbalanced tagging problem, the model performance under the low-resource settings could suffer using these standard objective functions. Based on recent advances in area under the ROC curve (AUC) maximization, we propose to optimize the NER model by maximizing the AUC score. We give evidence that by simply combining two binary-classifiers that maximize the AUC score, significant performance improvement over traditional loss functions is achieved under low-resource NER settings. We also conduct extensive experiments to demonstrate the advantages of our method under the low-resource and highly-imbalanced data distribution settings. To the best of our knowledge, this is the first work that brings AUC maximization to the NER setting. Furthermore, we show that our method is agnostic to different types of NER embeddings, models and domains. The code to replicate this work will be provided upon request. | ['Lan Du', 'Changyou Chen', 'Richard Beare', 'Wray Buntine', 'Wei Tan', 'Ngoc Dang Nguyen'] | 2022-12-09 | null | null | null | null | ['low-resource-named-entity-recognition'] | ['natural-language-processing'] | [-1.76663101e-01 2.74206605e-02 -3.81187797e-01 -5.72708786e-01
-1.06497931e+00 -5.73991477e-01 3.17991555e-01 4.53641355e-01
-1.04237688e+00 8.60812068e-01 1.85206771e-01 -2.28162855e-01
-1.47437751e-01 -6.79865420e-01 -2.98848271e-01 -2.22903982e-01
1.55997686e-02 4.12616849e-01 -5.22095226e-02 1.30438164e-01
1.50374666e-01 5.43971360e-01 -1.17050278e+00 -2.07589597e-01
9.94470239e-01 7.42283285e-01 -1.15411155e-01 6.80196941e-01
-2.30077609e-01 6.71475291e-01 -7.48981833e-01 -9.23735976e-01
2.17989117e-01 -6.25004619e-02 -1.03995109e+00 -4.91189092e-01
3.43008578e-01 1.35605531e-02 -1.88275591e-01 1.00371408e+00
8.18839729e-01 3.04960012e-01 5.72960436e-01 -1.21067762e+00
-7.36374080e-01 5.21838069e-01 -2.36509547e-01 3.35198849e-01
1.64943188e-01 -2.62981653e-01 1.29271472e+00 -7.52074003e-01
4.36893374e-01 8.38107586e-01 1.16364920e+00 4.92627591e-01
-1.25675488e+00 -5.70087373e-01 -7.11612925e-02 -1.77840799e-01
-1.38465655e+00 -4.40427870e-01 3.32331479e-01 -1.54946268e-01
1.16886795e+00 7.89986253e-02 2.48955563e-02 7.63238609e-01
5.58308698e-02 4.89981085e-01 9.49901521e-01 -5.55632830e-01
2.39480942e-01 3.72958541e-01 3.03512663e-01 3.56262624e-01
5.76131344e-01 -3.32002193e-02 -2.03676164e-01 -5.14080644e-01
2.64461040e-01 -6.15901612e-02 5.49268760e-02 -6.01559170e-02
-7.65295148e-01 8.53582263e-01 1.82440221e-01 4.65467036e-01
-3.18365276e-01 -3.01830694e-02 1.75557613e-01 1.81370094e-01
6.77089214e-01 8.46379817e-01 -8.89084578e-01 -3.38339627e-01
-1.07087874e+00 5.95062338e-02 1.43206239e+00 6.90653741e-01
6.25618935e-01 -2.81803608e-01 -8.71375855e-03 1.10052252e+00
2.31032670e-01 3.10629398e-01 4.82442111e-01 -9.56987202e-01
6.44204378e-01 3.82612824e-01 3.33586663e-01 -7.69742072e-01
-4.60692257e-01 -3.66725296e-01 -4.39685255e-01 -2.00024605e-01
4.97229129e-01 -6.16458178e-01 -6.71506703e-01 1.94644856e+00
3.26680809e-01 -1.58217549e-01 8.62823501e-02 5.88776529e-01
5.79012334e-01 3.34029824e-01 6.56447649e-01 3.05765476e-02
1.43418980e+00 -5.31809866e-01 -8.33827913e-01 -3.49044025e-01
1.04943967e+00 -8.69418800e-01 9.17884707e-01 -1.24254897e-01
-7.87988007e-01 -8.18182826e-02 -9.87888396e-01 -1.31927967e-01
-5.74956477e-01 1.43763930e-01 6.32262170e-01 1.08394027e+00
-8.27115119e-01 6.67358458e-01 -9.35321569e-01 -5.97136140e-01
3.74807656e-01 3.32795620e-01 -5.42325974e-01 -1.92393944e-01
-1.43493676e+00 1.26982796e+00 2.16863140e-01 -9.14442837e-02
-2.19738990e-01 -8.11809540e-01 -8.76763761e-01 1.63438275e-01
-5.35029881e-02 -4.85952497e-01 1.24386322e+00 -6.75560892e-01
-9.55434561e-01 8.13057065e-01 -4.74645644e-02 -3.35839897e-01
3.04832935e-01 -3.41460943e-01 -4.46862340e-01 -2.24867091e-01
1.03713162e-01 5.16801953e-01 -4.07379754e-02 -8.78079236e-01
-3.67072910e-01 -3.31626773e-01 1.13260210e-01 1.84803709e-01
-7.63445675e-01 3.73593360e-01 -9.21982452e-02 -4.73250240e-01
-3.06790948e-01 -6.80532336e-01 -3.95498127e-01 -3.87976795e-01
-2.83648878e-01 -3.13770592e-01 2.64398187e-01 -6.59396052e-01
1.49726999e+00 -2.05535817e+00 -3.97032648e-01 4.88512442e-02
5.02651595e-02 2.86303699e-01 -9.37350243e-02 4.59270060e-01
-1.00035340e-01 7.19109893e-01 -2.85055816e-01 -3.68138015e-01
2.85396308e-01 7.52310604e-02 1.87037840e-01 4.89619732e-01
4.15270716e-01 6.50500357e-01 -7.98409820e-01 -6.34833813e-01
-1.60225824e-01 6.49576247e-01 -5.54040730e-01 2.74845451e-01
3.12628627e-01 -1.19347773e-01 -2.30542719e-01 4.60995018e-01
7.22958505e-01 -2.01575577e-01 4.28414524e-01 -1.35622278e-01
-1.13113604e-01 4.28783476e-01 -1.20542300e+00 1.33946490e+00
-5.61445653e-01 5.32884300e-01 -1.03698522e-01 -9.18011487e-01
8.70152950e-01 2.98779130e-01 5.02493322e-01 -5.38818181e-01
-6.46262839e-02 2.78582633e-01 -1.45075992e-01 -2.89700240e-01
5.96349597e-01 -5.33255160e-01 -4.01741326e-01 3.80617768e-01
2.93920279e-01 1.66247532e-01 1.73626527e-01 -2.05836371e-02
1.30574143e+00 -1.09798558e-01 4.30149615e-01 -2.07266510e-01
-1.52543023e-01 1.89878754e-02 9.59743440e-01 8.83718312e-01
-5.68226397e-01 4.28541929e-01 5.00703216e-01 -1.69344291e-01
-1.30200458e+00 -8.86414349e-01 -7.05068111e-01 1.03795576e+00
-1.84966341e-01 -3.49379152e-01 -7.04739690e-01 -9.43884313e-01
-8.21348950e-02 8.47653210e-01 -3.83421034e-01 1.32665649e-01
-4.10072803e-01 -1.35003233e+00 1.19183040e+00 7.12365031e-01
2.96206415e-01 -8.51983666e-01 -3.15765709e-01 2.07861915e-01
-3.30895215e-01 -1.08141327e+00 -4.61404741e-01 5.14948726e-01
-9.02646124e-01 -9.78935778e-01 -6.41031504e-01 -7.91516960e-01
5.17123342e-01 -2.20913038e-01 1.32721877e+00 -7.04400539e-02
-2.61554271e-01 4.19600844e-01 -3.40350598e-01 -4.58525896e-01
-1.88572537e-02 3.59322667e-01 7.66761452e-02 -5.84759653e-01
8.77071798e-01 -3.13665062e-01 -4.73577082e-01 2.79871047e-01
-9.75866079e-01 -6.74075007e-01 4.67081457e-01 9.35308218e-01
4.48336005e-01 2.53508449e-01 6.42588377e-01 -1.02296412e+00
5.17720819e-01 -9.14863527e-01 -2.45280623e-01 3.32097679e-01
-8.97831440e-01 2.87153900e-01 5.50591707e-01 -2.92391539e-01
-1.16604233e+00 -1.48642227e-01 -3.55897814e-01 -2.60713734e-02
-3.37151974e-01 5.78187048e-01 -2.94773638e-01 2.81989068e-01
7.98565149e-01 -5.15962303e-01 -5.17286897e-01 -7.90582299e-01
1.56076729e-01 9.43745077e-01 1.46854192e-01 -6.16199315e-01
4.46954399e-01 2.07251042e-01 -2.04627007e-01 -6.28119051e-01
-1.20220768e+00 -5.62127531e-01 -4.95728463e-01 3.20983499e-01
8.32368791e-01 -9.30851758e-01 -2.99004734e-01 1.48621574e-01
-1.01097846e+00 -1.60174593e-01 -2.80629903e-01 7.44894862e-01
-2.41371527e-01 4.43817049e-01 -7.12203503e-01 -1.16797316e+00
-3.40723991e-01 -6.67135954e-01 8.15773368e-01 4.48647201e-01
-2.28366449e-01 -1.37265289e+00 3.67268354e-01 1.99997067e-01
4.26964611e-01 1.23445608e-01 6.06403112e-01 -1.35175729e+00
1.16388164e-01 -5.48729062e-01 -1.39838234e-01 4.93336469e-01
5.93112856e-02 -1.49266377e-01 -1.00043643e+00 -1.67361811e-01
-3.15089338e-02 -4.10385787e-01 7.71344721e-01 2.25093856e-01
9.97341871e-01 -2.49783039e-01 -1.70567065e-01 5.06187379e-01
1.70801878e+00 -1.18599661e-01 7.13912427e-01 4.94076014e-01
5.63708901e-01 7.53670692e-01 6.38279021e-01 4.52580303e-01
7.60492861e-01 5.90945363e-01 4.65524606e-02 -2.51055002e-01
9.04438943e-02 -3.45529586e-01 3.88247877e-01 7.96199441e-01
1.67059883e-01 -5.05322456e-01 -1.20438552e+00 7.10862696e-01
-1.52405524e+00 -1.12570727e+00 7.64800087e-02 2.24129891e+00
1.05791891e+00 -1.70598812e-02 2.15466507e-02 -1.11756675e-01
9.72296774e-01 8.48554894e-02 -4.13135380e-01 -5.93975008e-01
-2.09984288e-01 4.17410582e-01 8.33125412e-01 3.34778547e-01
-1.34491515e+00 6.54133618e-01 6.79690933e+00 7.27319002e-01
-6.46453917e-01 2.49191493e-01 5.76356351e-01 5.02418615e-02
-2.62062341e-01 1.87337905e-01 -1.01375341e+00 4.82282400e-01
1.45412683e+00 -1.08531684e-01 2.20603779e-01 8.71956885e-01
-1.21359698e-01 8.23625699e-02 -9.08826828e-01 6.75448000e-01
2.85930000e-02 -7.13989258e-01 -6.23833358e-01 1.92804664e-01
6.60809755e-01 3.65975142e-01 -1.28957689e-01 3.93578529e-01
7.85234749e-01 -1.12068439e+00 2.92639315e-01 5.23710907e-01
9.34508026e-01 -8.09842288e-01 1.23159409e+00 2.26904884e-01
-1.06366730e+00 -9.36012268e-02 -4.29338336e-01 1.73420191e-01
1.22666657e-01 1.10227835e+00 -7.84694254e-01 4.54868108e-01
6.94515646e-01 3.63350332e-01 -3.39015156e-01 1.31868231e+00
4.97124530e-02 7.48542964e-01 -5.45714796e-01 -3.18275467e-02
-9.79665518e-02 3.23910564e-01 3.58741999e-01 1.66974032e+00
2.76659667e-01 -1.10314175e-01 6.98681921e-02 5.77996135e-01
-4.74353611e-01 3.45065087e-01 -5.90728641e-01 -3.08526397e-01
8.16637397e-01 1.29474521e+00 -4.89534736e-01 6.58347830e-02
-4.84236240e-01 7.04500198e-01 7.73302436e-01 1.15580827e-01
-6.91515744e-01 -1.07346332e+00 7.23447382e-01 -1.32889181e-01
4.49007720e-01 -1.48262188e-01 -6.23718798e-01 -1.41121113e+00
-8.48769117e-03 -5.02270818e-01 7.83934057e-01 -2.83801377e-01
-1.94432187e+00 4.42050964e-01 -1.09272882e-01 -1.03287208e+00
7.87190199e-02 -5.94436765e-01 -4.01892871e-01 8.06445658e-01
-1.76260066e+00 -6.21140659e-01 1.08105451e-01 8.17229003e-02
5.48392441e-03 2.48843312e-01 1.10585499e+00 6.78889275e-01
-7.00290322e-01 1.03550589e+00 3.71589094e-01 5.73557377e-01
9.68737364e-01 -1.52504051e+00 1.48684114e-01 7.64782071e-01
1.20446175e-01 7.21431851e-01 5.39643824e-01 -6.40105426e-01
-8.84336293e-01 -9.40369308e-01 1.63086104e+00 -8.37014198e-01
5.20242631e-01 -2.66994059e-01 -7.13557601e-01 6.60527408e-01
9.90215689e-03 5.38996607e-02 1.40664756e+00 5.74089944e-01
-5.32999098e-01 7.08998814e-02 -1.55202615e+00 2.27789447e-01
8.55143189e-01 -4.57585901e-01 -7.26187170e-01 2.98262477e-01
5.79017401e-01 -2.39217743e-01 -1.72443759e+00 4.22942907e-01
5.28714597e-01 -7.09137976e-01 7.57318377e-01 -9.24510002e-01
2.86742389e-01 -1.84684694e-01 -4.66304034e-01 -1.33417654e+00
-2.10323364e-01 -1.33830339e-01 6.77547753e-02 1.63849747e+00
7.77226210e-01 -6.16302073e-01 5.15327215e-01 1.09955847e+00
1.61687002e-01 -6.47016943e-01 -8.81101191e-01 -7.59582102e-01
4.17678118e-01 -5.15076637e-01 7.20812440e-01 1.21430719e+00
1.64256334e-01 2.03477561e-01 -2.43453547e-01 1.35157585e-01
6.49708986e-01 -3.91973257e-01 2.78475642e-01 -1.38650024e+00
-1.64545909e-01 -5.43891564e-02 -4.27587926e-01 -6.78160429e-01
3.58346134e-01 -9.28576708e-01 1.81566641e-01 -1.53800833e+00
3.64703119e-01 -9.06365097e-01 -6.81310892e-01 7.06031859e-01
-4.46839780e-01 1.40977353e-01 2.56764382e-01 2.42689222e-01
-8.31085324e-01 3.43117833e-01 6.10475242e-01 2.53951371e-01
-8.34166352e-03 -1.78110972e-01 -1.09454596e+00 4.56871003e-01
9.85568345e-01 -1.02373350e+00 -1.41365612e-02 -4.83965456e-01
3.77126813e-01 -3.02048236e-01 7.92190731e-02 -7.33389616e-01
3.10244173e-01 -7.47511834e-02 5.65445900e-01 3.13910246e-02
-1.47410249e-02 -7.75765002e-01 -9.97809134e-03 3.51445600e-02
-4.75012243e-01 2.70310372e-01 1.98882669e-01 5.20360470e-01
-2.31798366e-01 -4.25676465e-01 7.97880709e-01 -1.01910241e-01
-3.29762816e-01 2.16547027e-01 -5.15964255e-02 7.03981757e-01
6.72814488e-01 -8.53523239e-02 -4.98479515e-01 -2.29627237e-01
-5.49841523e-01 9.85246375e-02 5.97568750e-01 5.96459508e-02
1.19435653e-01 -1.33512127e+00 -7.49854267e-01 -1.92701325e-01
6.29879236e-02 -2.81137288e-01 1.37369648e-01 8.71861219e-01
-2.30263174e-01 2.80312568e-01 1.15344852e-01 -7.81313181e-02
-8.43620241e-01 1.81446746e-01 5.20499110e-01 -8.02215993e-01
-1.42053485e-01 6.68363094e-01 -2.51718670e-01 -9.87909079e-01
1.66459396e-01 2.22863033e-01 -1.89998195e-01 1.11477956e-01
4.25084412e-01 6.10732555e-01 3.05495888e-01 -6.03662848e-01
-6.47810757e-01 7.28944987e-02 4.07285728e-02 -1.77238360e-01
1.47177947e+00 -4.50864583e-02 5.25331646e-02 1.96776554e-01
1.36367810e+00 2.99666911e-01 -9.54443634e-01 -1.07029162e-01
4.33704853e-01 -3.07261735e-01 1.96419418e-01 -8.91662896e-01
-7.26618350e-01 7.09748566e-01 5.26165068e-01 2.32109591e-01
1.05184567e+00 -1.45569071e-01 7.66163528e-01 3.61202717e-01
1.14959843e-01 -1.32674932e+00 -3.39821607e-01 7.08920240e-01
3.27830389e-02 -1.14527726e+00 -9.46065858e-02 1.86845381e-02
-7.92842805e-01 9.58522975e-01 6.74640179e-01 3.26641090e-02
8.84449542e-01 6.95604146e-01 1.18936189e-01 2.89652310e-02
-7.07859755e-01 -2.88509488e-01 7.65106529e-02 6.56431019e-01
9.09958184e-01 1.13963693e-01 -6.57125473e-01 8.83293509e-01
-2.55822748e-01 -3.42941284e-03 3.53199542e-01 9.16859746e-01
-2.70637959e-01 -1.41171789e+00 -2.79514760e-01 6.82156920e-01
-1.18313229e+00 -3.11232746e-01 -2.47357070e-01 6.35752976e-01
-4.06224700e-03 1.21092534e+00 1.97551832e-01 -2.86074966e-01
5.29620111e-01 4.10251111e-01 2.03179270e-01 -6.15645111e-01
-6.97631538e-01 -5.20634949e-01 4.78791505e-01 -2.36965165e-01
-6.55612648e-01 -6.59132004e-01 -1.29175138e+00 -4.71298873e-01
-7.53269136e-01 4.07261938e-01 8.27279210e-01 8.31894219e-01
5.90072036e-01 4.09944952e-02 5.97090721e-01 -2.81924754e-01
-7.53235757e-01 -9.27388370e-01 -6.11625373e-01 5.00435829e-01
4.64235507e-02 -6.38211489e-01 -6.74615622e-01 -3.34915876e-01] | [9.710752487182617, 9.384809494018555] |
d35e5087-b05e-4eed-a8a3-7adcb6cf8367 | the-future-of-chatgpt-enabled-labor-market-a | 2304.09823 | null | https://arxiv.org/abs/2304.09823v2 | https://arxiv.org/pdf/2304.09823v2.pdf | The Future of ChatGPT-enabled Labor Market: A Preliminary Study | As a phenomenal large language model, ChatGPT has achieved unparalleled success in various real-world tasks and increasingly plays an important role in our daily lives and work. However, extensive concerns are also raised about the potential ethical issues, especially about whether ChatGPT-like artificial general intelligence (AGI) will replace human jobs. To this end, in this paper, we introduce a preliminary data-driven study on the future of ChatGPT-enabled labor market from the view of Human-AI Symbiosis instead of Human-AI Confrontation. To be specific, we first conduct an in-depth analysis of large-scale job posting data in BOSS Zhipin, the largest online recruitment platform in China. The results indicate that about 28% of occupations in the current labor market require ChatGPT-related skills. Furthermore, based on a large-scale occupation-centered knowledge graph, we develop a semantic information enhanced collaborative filtering algorithm to predict the future occupation-skill relations in the labor market. As a result, we find that additional 45% occupations in the future will require ChatGPT-related skills. In particular, industries related to technology, products, and operations are expected to have higher proficiency requirements for ChatGPT-related skills, while the manufacturing, services, education, and health science related industries will have lower requirements for ChatGPT-related skills. | ['HengShu Zhu', 'Meng Chang', 'Yaqi Yang', 'Shiyu Wu', 'Xi Chen', 'Lan Chen'] | 2023-04-14 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-1.11061431e-01 5.50424337e-01 -2.93610036e-01 -1.56755567e-01
4.06792089e-02 -2.53526449e-01 3.14797610e-01 -1.02037288e-01
-4.81483817e-01 5.96086025e-01 3.48564565e-01 -5.00875652e-01
-7.65550673e-01 -9.54756677e-01 -1.21272877e-01 -1.78115204e-01
5.14127016e-01 1.01956618e+00 -3.04523528e-01 -6.77202523e-01
2.66719759e-01 2.67658949e-01 -1.16535676e+00 -7.17822090e-02
1.25045240e+00 5.60406923e-01 4.59014356e-01 4.08838950e-02
-1.26133233e-01 9.61307824e-01 -5.59030712e-01 -1.01906598e+00
2.61810690e-01 -5.34713328e-01 -1.27771056e+00 4.80124056e-02
-1.80388391e-01 2.64193475e-01 -5.38228095e-01 9.64644670e-01
2.95206517e-01 2.98539758e-01 4.04953420e-01 -1.31345284e+00
-1.00849056e+00 8.24795902e-01 -2.69054562e-01 -1.98608965e-01
3.11985523e-01 1.72856748e-01 1.04089868e+00 -5.62491179e-01
9.36275601e-01 1.21357250e+00 5.44335067e-01 2.25050718e-01
-6.82500243e-01 -9.51707721e-01 -1.95194315e-02 1.40132025e-01
-1.35347807e+00 1.35318264e-01 7.41736531e-01 -4.80374396e-01
8.34437370e-01 2.51123428e-01 9.10522223e-01 6.57496691e-01
5.41480184e-01 3.22551936e-01 9.35686171e-01 -4.79528844e-01
-2.03329623e-01 2.99069941e-01 -2.80608684e-02 7.16651380e-01
4.96533126e-01 -3.86679620e-01 -7.83597469e-01 7.30902478e-02
9.12904143e-01 1.91455647e-01 5.30035853e-01 3.33534390e-01
-1.34935033e+00 6.01234794e-01 4.02890265e-01 8.50865126e-01
-4.69493628e-01 -3.26614901e-02 -8.67904350e-03 5.73692977e-01
6.76947594e-01 9.72707987e-01 -3.23892206e-01 -6.44389749e-01
-2.65955091e-01 1.52741417e-01 8.42251062e-01 1.14268935e+00
8.22393656e-01 -2.69094944e-01 -1.25825688e-01 7.42622972e-01
1.09900050e-01 2.07524657e-01 1.95166856e-01 -1.36336625e+00
5.15540659e-01 1.16899490e+00 -1.47452787e-01 -1.36485887e+00
-2.91576922e-01 -8.11226666e-01 -1.01895571e+00 -5.91493189e-01
5.27456880e-01 -5.18693805e-01 -4.29888785e-01 1.24813342e+00
-1.02999825e-02 -6.06045797e-02 -3.70694622e-02 9.54178989e-01
4.58041310e-01 3.99498761e-01 2.20545009e-01 -1.65966749e-01
1.67104983e+00 -8.82420599e-01 -1.03379703e+00 -6.68951333e-01
6.64449334e-01 -6.33167803e-01 9.76209760e-01 1.22306854e-01
-1.17879009e+00 -9.06869829e-01 -4.16691661e-01 -1.49293527e-01
-4.97402608e-01 7.08333626e-02 1.22305584e+00 7.60456741e-01
-7.35780001e-01 4.26465541e-01 -3.54947597e-01 -7.28746295e-01
1.73345879e-01 3.45734656e-01 -3.11513394e-01 -1.49265066e-01
-1.62179279e+00 9.86000299e-01 3.64030778e-01 2.17932463e-01
-2.63500046e-02 -5.67774177e-01 -6.51324570e-01 1.51582152e-01
7.84792542e-01 -8.49598467e-01 1.04805386e+00 -5.72304070e-01
-1.07717299e+00 8.28600109e-01 2.19035856e-02 -1.16056763e-01
1.96750030e-01 -5.83869480e-02 -6.81263387e-01 -2.00986817e-01
4.51299518e-01 2.99087942e-01 8.27313513e-02 -8.85185540e-01
-8.19002807e-01 -3.77111256e-01 1.72382966e-01 5.04152894e-01
-8.88168216e-01 2.17449531e-01 -5.81768870e-01 -4.58410352e-01
1.40900254e-01 -1.06908774e+00 -3.47509056e-01 -6.44161165e-01
-5.11243418e-02 -6.61240220e-01 3.52766603e-01 -5.71186483e-01
1.40283954e+00 -2.04010201e+00 1.52184963e-01 4.41997737e-01
3.32231969e-01 -1.89299449e-01 7.42743090e-02 8.75948548e-01
5.34869790e-01 2.97804803e-01 3.81107628e-01 -2.24813446e-02
1.92767009e-01 4.37473714e-01 1.74662396e-01 -1.92469716e-01
-9.54001322e-02 1.16163492e+00 -1.22174489e+00 -6.84948266e-01
-1.89116374e-01 -4.32569534e-01 -4.00930852e-01 -2.52592236e-01
1.64823562e-01 7.39757955e-01 -7.59809434e-01 8.60593140e-01
1.17943980e-01 -4.47602928e-01 7.01457798e-01 5.50833106e-01
-2.38560736e-01 2.82295384e-02 -3.77451926e-01 1.68013740e+00
-5.18139958e-01 6.40114963e-01 2.78223813e-01 -1.05915356e+00
1.05775797e+00 3.07983518e-01 5.69287539e-01 -6.57628953e-01
2.66507030e-01 2.70193577e-01 5.28670907e-01 -4.22321320e-01
8.64578187e-01 -1.12353407e-01 -5.12868822e-01 4.78755236e-01
-4.74058650e-02 -2.97473639e-01 2.65963972e-01 4.13289577e-01
1.05910861e+00 -1.40226275e-01 -3.70917097e-02 -3.88742507e-01
2.39715219e-01 4.74164665e-01 8.85523438e-01 7.39756107e-01
-2.44903937e-01 -6.39570355e-02 5.93433321e-01 -1.98659211e-01
-6.50295734e-01 -3.99717271e-01 2.61308968e-01 1.30325961e+00
3.48941922e-01 -4.07432050e-01 -5.79696178e-01 -5.29615879e-01
3.89997661e-02 7.75369942e-01 -3.03018652e-02 -2.48820975e-01
-3.66984397e-01 -1.49087429e-01 3.96768481e-01 3.78574640e-01
1.06401169e+00 -1.35940087e+00 -9.85216796e-02 5.70188649e-02
-5.12554049e-01 -1.22787225e+00 -6.02967203e-01 -2.21458882e-01
-5.42153955e-01 -5.66298366e-01 -6.86232090e-01 -1.17530739e+00
6.53018177e-01 1.83641434e-01 1.13767815e+00 2.39516571e-01
-1.29754320e-01 4.29583102e-01 -3.49132627e-01 -8.15605104e-01
-3.87492001e-01 5.64780056e-01 2.47130916e-01 -3.79263729e-01
7.04130590e-01 -5.52289248e-01 -2.86503136e-01 5.85459113e-01
-3.28355491e-01 2.94680119e-01 7.02587545e-01 6.42403781e-01
-6.09436706e-02 8.56048822e-01 8.84988010e-01 -1.00133038e+00
9.85960841e-01 -5.48426032e-01 2.07649339e-02 3.26844513e-01
-9.13469613e-01 -6.29201770e-01 2.21380994e-01 -4.23429549e-01
-1.55848086e+00 -2.39663973e-01 8.81487802e-02 5.19377552e-02
3.51780318e-02 1.19266617e+00 3.19692306e-02 1.59411341e-01
4.93274152e-01 -1.55223608e-02 7.10879341e-02 -3.64670545e-01
-5.52254617e-02 1.15071821e+00 5.60146987e-01 -7.14953125e-01
9.90178525e-01 -3.13971080e-02 5.16092293e-02 -7.77572036e-01
-8.82122278e-01 -5.94683707e-01 -5.11467874e-01 -6.86166942e-01
9.55771744e-01 -7.56282330e-01 -1.38098443e+00 2.71201074e-01
-8.19673061e-01 -3.49840552e-01 -1.53830662e-01 8.05654168e-01
-2.06145570e-01 1.62704423e-01 -8.36549342e-01 -1.06148314e+00
-1.29269734e-01 -6.83679163e-01 6.53605878e-01 4.74861115e-01
-5.54285765e-01 -9.05791700e-01 -4.66439843e-01 1.46672130e+00
2.74018049e-01 -3.31836671e-01 1.21089423e+00 -5.84238291e-01
-4.79333729e-01 -1.76172808e-01 -1.62360653e-01 1.60861567e-01
2.48594791e-01 -8.42655480e-01 -1.51686087e-01 -9.62535590e-02
-7.44698569e-02 -3.32419962e-01 1.89938799e-01 1.13018043e-02
7.98737526e-01 -1.89664196e-02 -4.12048757e-01 -1.57633945e-01
7.66728044e-01 6.81007504e-01 5.92438817e-01 2.25618690e-01
5.78501880e-01 1.16417444e+00 1.24217904e+00 3.53798956e-01
4.58657950e-01 4.24931824e-01 -1.67881638e-01 -1.72429934e-01
1.52325884e-01 -6.14334345e-01 1.15449518e-01 1.23732042e+00
-6.39541924e-01 -3.24369818e-01 -1.35535276e+00 5.33341110e-01
-2.16347194e+00 -8.39795232e-01 -2.80956626e-01 1.70389664e+00
7.45094895e-01 1.99614003e-01 1.06076255e-01 -2.24072844e-01
6.57592654e-01 -4.92199153e-01 -4.15963203e-01 -2.56272048e-01
3.12083483e-01 1.94469824e-01 3.75976622e-01 -5.16616143e-02
-4.80578572e-01 1.37577212e+00 6.19600105e+00 9.31894779e-01
-2.27695510e-01 1.85310692e-01 7.65774488e-01 2.93220788e-01
-1.78745314e-01 2.56298870e-01 -5.67036748e-01 1.52747944e-01
6.28485918e-01 -6.74032629e-01 5.96706092e-01 7.13734984e-01
4.53911424e-02 -6.30917400e-02 -4.77210045e-01 8.20978224e-01
-3.47579151e-01 -1.22837353e+00 -4.00964200e-01 7.13852108e-01
9.01719689e-01 -5.66006482e-01 9.87430513e-02 5.96025348e-01
5.10179162e-01 -8.83885980e-01 4.95044410e-01 6.31729960e-01
7.19780385e-01 -9.63087618e-01 1.02319050e+00 9.06268358e-01
-1.04600823e+00 -3.21320027e-01 -4.88769323e-01 -1.03695595e+00
2.52878129e-01 5.01417518e-01 -7.15336859e-01 9.87141073e-01
7.02836514e-01 4.61394459e-01 -2.33090758e-01 2.29380503e-01
-3.27409953e-02 6.60280883e-01 1.85985580e-01 -2.34131888e-01
7.79519528e-02 -7.36396909e-01 6.15228526e-02 6.47895038e-01
6.64254546e-01 4.23575580e-01 2.21793637e-01 8.17353368e-01
-2.02775538e-01 1.04852021e-03 -8.38062108e-01 -1.00383449e+00
6.45772994e-01 1.21050453e+00 -8.16964924e-01 -1.86913878e-01
-5.87761641e-01 9.43454683e-01 5.78181259e-02 3.53673339e-01
-3.98748934e-01 -6.77061915e-01 8.43642056e-01 1.41430140e-01
-5.72719514e-01 -5.44883907e-01 -4.20608997e-01 -7.25535274e-01
-4.18928802e-01 -8.81032586e-01 1.58605918e-01 -7.28090048e-01
-1.34097767e+00 2.11634085e-01 -1.19837083e-01 -7.33246088e-01
-3.30782682e-01 -4.76002008e-01 -3.04829568e-01 1.00409365e+00
-4.37398553e-01 -1.07582653e+00 -1.41966373e-01 3.28213006e-01
3.40768546e-01 -5.16794562e-01 6.07059777e-01 2.48125106e-01
-4.60260391e-01 3.68938237e-01 -2.18780816e-01 3.19863588e-01
5.55419922e-01 -6.88646317e-01 4.76879418e-01 4.61573452e-01
-2.30347533e-02 9.51549709e-01 3.94075990e-01 -1.37400818e+00
-1.53638399e+00 -8.16654801e-01 1.80031681e+00 -6.24053359e-01
7.48735487e-01 -3.87619108e-01 -2.29643643e-01 8.68797481e-01
4.37142789e-01 -8.80315781e-01 6.86141789e-01 7.29665458e-01
6.41300797e-01 -4.59895320e-02 -9.65254366e-01 7.38122344e-01
1.90352523e+00 -6.85729146e-01 -5.60794592e-01 6.67195976e-01
9.40582156e-01 -2.20491961e-01 -1.34932184e+00 1.09950781e-01
4.61914450e-01 -4.71312076e-01 8.74432743e-01 -4.57141131e-01
3.01692992e-01 1.30183324e-01 4.17289376e-01 -1.09631133e+00
-9.85427022e-01 -9.04395938e-01 6.10210836e-01 1.25289881e+00
2.34630510e-01 -1.06307924e+00 1.11623025e+00 1.03049052e+00
-4.28478599e-01 -4.58692938e-01 -8.63291264e-01 -1.08904397e+00
-1.37281135e-01 -3.52182627e-01 5.80072284e-01 1.19796503e+00
5.86170971e-01 7.52245963e-01 -6.55609548e-01 -2.50428379e-01
1.49100587e-01 1.71022043e-01 7.91978061e-01 -1.71496820e+00
-1.42982304e-01 -1.77909285e-01 -2.23635808e-01 -7.90534139e-01
3.46613795e-01 -9.75583017e-01 -2.52324224e-01 -1.93369210e+00
3.74140233e-01 -3.85864735e-01 -6.02966137e-02 6.55110717e-01
1.31165341e-01 1.70910791e-01 2.22057745e-01 2.29555547e-01
-4.19556171e-01 1.84758440e-01 1.87146640e+00 -5.84333427e-02
-3.63956302e-01 1.69423774e-01 -1.23157370e+00 4.89106417e-01
8.43105614e-01 -1.88539430e-01 -6.51481807e-01 -2.45739564e-01
7.96148956e-01 3.36498648e-01 -8.26297104e-02 -7.27902532e-01
5.25466323e-01 -7.75570035e-01 -3.82812843e-02 -2.30748281e-01
3.10815603e-01 -8.85187805e-01 6.61250293e-01 5.70617795e-01
-2.67978281e-01 1.62451882e-02 -2.41801605e-01 2.86048800e-01
-3.10301125e-01 -2.30882272e-01 -2.51038969e-01 -3.12947124e-01
-5.49744487e-01 9.73288491e-02 -9.14441884e-01 -2.77333170e-01
7.67127395e-01 -3.86238933e-01 -4.30474222e-01 -7.52300441e-01
-6.03577435e-01 4.75467116e-01 2.86224157e-01 5.64555585e-01
3.60900283e-01 -1.26023316e+00 -4.65994239e-01 -5.72754769e-03
5.69159910e-02 -2.59976208e-01 2.63564050e-01 9.52450693e-01
-3.49791497e-01 7.47982502e-01 -2.25246906e-01 3.53619814e-01
-1.05242848e+00 3.74099702e-01 -4.68752265e-01 -3.45210224e-01
-2.19512969e-01 7.28775740e-01 1.02194436e-01 -5.91125369e-01
1.92874167e-02 1.11224152e-01 -8.62506852e-02 -1.47498548e-01
-1.38541237e-01 7.57524490e-01 -3.14706773e-01 -6.01576507e-01
-2.03600690e-01 1.65665373e-01 2.86986321e-01 -2.61459500e-01
9.62034643e-01 -1.44906536e-01 -4.06906277e-01 4.15277064e-01
6.34052873e-01 -3.42893414e-02 -3.68774354e-01 -1.51564822e-01
2.23138273e-01 -6.50232315e-01 -1.96498558e-01 -9.21247482e-01
-8.36121082e-01 4.28963780e-01 -3.48923691e-02 4.49307024e-01
1.09294963e+00 2.84301162e-01 9.18956220e-01 6.71257615e-01
9.26104546e-01 -1.59686470e+00 2.30203420e-01 7.25960255e-01
7.12301314e-01 -7.84728348e-01 -1.63057983e-01 -8.33579004e-01
-1.06093800e+00 6.05616510e-01 9.09848034e-01 6.66668117e-01
4.77852643e-01 -3.28208566e-01 3.84454057e-02 -5.24831831e-01
-6.33041978e-01 -3.89179558e-01 2.99185723e-01 4.90799427e-01
6.93855286e-01 2.24670216e-01 -6.63205445e-01 7.06096768e-01
-5.44860184e-01 1.11471549e-01 1.22072557e-02 8.84929419e-01
-3.30642998e-01 -1.40188432e+00 -2.01469421e-01 7.03046322e-01
-2.91256607e-01 4.34047841e-02 -8.55272532e-01 6.95356786e-01
4.36293006e-01 1.40562105e+00 2.43685637e-02 -8.46008062e-01
3.59539390e-01 2.75526702e-01 2.30618298e-01 -7.78493404e-01
-8.24851513e-01 -8.31588432e-02 5.75324297e-01 -1.24041177e-01
-3.09978962e-01 -4.42216843e-01 -1.34815013e+00 -7.87024498e-01
-2.70076364e-01 4.50034052e-01 2.87761092e-01 1.00265718e+00
6.42768562e-01 5.10775208e-01 1.92432791e-01 -2.53775269e-01
-1.23693988e-01 -1.16608417e+00 -1.06344783e+00 2.17544019e-01
-9.77125943e-01 -4.58897352e-01 1.50555432e-01 -3.08182746e-01] | [9.19057846069336, 6.6236958503723145] |
3cdc8a13-e4bd-47be-850a-6940a93f7cb0 | fastpose-towards-real-time-pose-estimation | 1908.05593 | null | https://arxiv.org/abs/1908.05593v1 | https://arxiv.org/pdf/1908.05593v1.pdf | FastPose: Towards Real-time Pose Estimation and Tracking via Scale-normalized Multi-task Networks | Both accuracy and efficiency are significant for pose estimation and tracking in videos. State-of-the-art performance is dominated by two-stages top-down methods. Despite the leading results, these methods are impractical for real-world applications due to their separated architectures and complicated calculation. This paper addresses the task of articulated multi-person pose estimation and tracking towards real-time speed. An end-to-end multi-task network (MTN) is designed to perform human detection, pose estimation, and person re-identification (Re-ID) tasks simultaneously. To alleviate the performance bottleneck caused by scale variation problem, a paradigm which exploits scale-normalized image and feature pyramids (SIFP) is proposed to boost both performance and speed. Given the results of MTN, we adopt an occlusion-aware Re-ID feature strategy in the pose tracking module, where pose information is utilized to infer the occlusion state to make better use of Re-ID feature. In experiments, we demonstrate that the pose estimation and tracking performance improves steadily utilizing SIFP through different backbones. Using ResNet-18 and ResNet-50 as backbones, the overall pose tracking framework achieves competitive performance with 29.4 FPS and 12.2 FPS, respectively. Additionally, occlusion-aware Re-ID feature decreases the identification switches by 37% in the pose tracking process. | ['Wei Zou', 'Jiabin Zhang', 'Yanwei Li', 'Peng Li', 'Hu Su', 'Guan Huang', 'Zheng Zhu'] | 2019-08-15 | null | null | null | null | ['multi-person-pose-estimation-and-tracking'] | ['computer-vision'] | [-2.00477228e-01 -3.61361772e-01 1.24722056e-01 -2.85014361e-01
-5.48712552e-01 -3.32927316e-01 1.06340967e-01 -3.09622020e-01
-7.51157105e-01 5.57075977e-01 1.89519778e-01 5.15193284e-01
2.39544049e-01 -4.46116060e-01 -5.01851618e-01 -3.75962377e-01
-2.52854079e-02 5.30296266e-01 5.60148358e-01 4.96704876e-02
-2.41537362e-01 6.66983902e-01 -1.64974356e+00 -7.19302967e-02
5.41054964e-01 9.52604949e-01 9.36976522e-02 8.32778394e-01
2.72930622e-01 1.33000672e-01 -6.81112766e-01 -4.70888466e-01
5.17711341e-01 1.45410836e-01 -2.87362039e-01 7.67158940e-02
1.09182250e+00 -7.16305315e-01 -5.95086396e-01 9.26804245e-01
1.10480118e+00 2.43771568e-01 3.97657305e-02 -1.37108064e+00
2.89491136e-02 -5.56839108e-02 -9.92513061e-01 1.93347111e-01
6.16699457e-01 2.24876300e-01 4.35727298e-01 -9.11634624e-01
4.90194112e-01 1.79339683e+00 1.09144378e+00 6.38882399e-01
-9.06623065e-01 -9.73391116e-01 3.10010582e-01 1.72922701e-01
-1.68355381e+00 -4.27278250e-01 5.14430463e-01 -2.48632327e-01
6.56681538e-01 1.61627591e-01 8.24534714e-01 8.93614829e-01
6.64861202e-02 7.61472464e-01 7.65374720e-01 -5.25421314e-02
-2.93484688e-01 -2.05162153e-01 -3.69023569e-02 7.87090063e-01
5.82515657e-01 2.10569456e-01 -7.98441827e-01 -4.49165925e-02
1.11670434e+00 2.44937807e-01 -1.65690184e-01 -3.47361535e-01
-1.08656502e+00 4.22919273e-01 6.11677527e-01 -1.33805841e-01
-3.62332761e-01 4.32207525e-01 5.75793505e-01 -7.24229291e-02
2.89186299e-01 -8.24269652e-02 -4.75721478e-01 -1.31838411e-01
-9.74989295e-01 4.90961611e-01 4.14282411e-01 1.10809147e+00
3.81940305e-01 1.60772931e-02 -5.68590581e-01 7.81428397e-01
4.83276814e-01 1.06390595e+00 1.21251069e-01 -8.75187099e-01
4.89726812e-01 6.52755737e-01 3.88398737e-01 -1.19084525e+00
-7.51075745e-01 -7.60187864e-01 -6.34707451e-01 -1.11227818e-01
6.03661299e-01 -3.89076799e-01 -9.34132159e-01 1.86155999e+00
8.44159782e-01 2.19391450e-01 -4.21838909e-01 1.28524852e+00
7.26258337e-01 2.67667115e-01 2.86581129e-01 -1.83553733e-02
1.80496955e+00 -1.15800881e+00 -5.87551832e-01 -4.53268468e-01
3.48232448e-01 -6.31084144e-01 6.17428124e-01 1.38458535e-01
-9.16637480e-01 -1.03435481e+00 -8.27069342e-01 -3.37603875e-02
2.09742822e-02 5.88937104e-01 4.20920879e-01 7.66874254e-01
-8.63199413e-01 3.21890861e-01 -9.56980348e-01 -6.00469172e-01
4.08795118e-01 7.64561474e-01 -3.22883874e-01 3.17249633e-02
-9.33649778e-01 6.29283607e-01 1.92452893e-01 4.06176686e-01
-5.92083931e-01 -6.87504113e-01 -7.72091150e-01 -7.75484787e-03
4.54961985e-01 -1.00452638e+00 1.09508765e+00 -4.97775346e-01
-1.35696197e+00 4.98963863e-01 -2.55954325e-01 -3.28737408e-01
8.71083796e-01 -7.96232462e-01 -4.29079890e-01 2.94915736e-01
2.12886557e-01 8.33294451e-01 9.08118665e-01 -9.12053704e-01
-8.61787021e-01 -7.63900638e-01 -1.33886799e-01 4.37727511e-01
-5.75069785e-01 2.41549715e-01 -1.07294238e+00 -5.62899709e-01
1.17982417e-01 -1.19997501e+00 -2.38259032e-01 5.43601036e-01
-3.95026594e-01 -1.75668314e-01 9.67584670e-01 -8.44712973e-01
1.15271640e+00 -1.91753888e+00 -3.30878757e-02 5.22729643e-02
3.74943018e-01 4.23189580e-01 1.82372332e-02 -7.34595656e-02
3.06935102e-01 -5.96741378e-01 5.06625235e-01 -7.38271117e-01
-3.54254548e-03 -2.06144750e-01 3.17635804e-01 8.04754794e-01
-1.37081534e-01 9.08647776e-01 -5.79602540e-01 -7.95580447e-01
5.11313498e-01 8.19893301e-01 -5.65567553e-01 1.89137861e-01
2.32668772e-01 5.65945029e-01 -4.26368922e-01 8.11344326e-01
9.62980628e-01 -2.43546978e-01 4.63203713e-02 -7.95802891e-01
-2.46270508e-01 -3.21688980e-01 -1.55043912e+00 1.80901802e+00
-1.53788045e-01 2.96754301e-01 1.94226205e-01 -3.49419743e-01
7.30118394e-01 1.56586856e-01 6.76150858e-01 -4.57655460e-01
3.10035557e-01 -3.79369967e-02 -1.52191639e-01 -8.58975127e-02
6.08035326e-01 3.87348473e-01 -9.80714709e-02 -2.83842161e-02
-9.61907357e-02 6.72659457e-01 2.27028459e-01 -6.92362264e-02
8.72276604e-01 4.99260277e-01 1.46467924e-01 -1.60841599e-01
7.52311230e-01 -2.63162494e-01 9.20044959e-01 5.67831576e-01
-7.20503390e-01 3.85298342e-01 -3.63927841e-01 -7.17254221e-01
-1.02266383e+00 -1.16908252e+00 6.16257377e-02 1.36801708e+00
3.59825283e-01 -5.75463295e-01 -6.78653121e-01 -3.85763884e-01
3.30226392e-01 -2.22272485e-01 -2.90936172e-01 7.40270093e-02
-9.42387164e-01 -6.17454588e-01 6.53442562e-01 9.29757774e-01
1.00095534e+00 -5.40106475e-01 -9.20525849e-01 2.60945737e-01
-2.91995645e-01 -1.52167714e+00 -9.50654864e-01 -6.65321529e-01
-6.47395134e-01 -9.29601133e-01 -1.05353856e+00 -4.84494627e-01
6.33742988e-01 4.70649332e-01 5.20258427e-01 -2.71761399e-02
-4.61810917e-01 5.15235186e-01 1.47823915e-01 -1.40613452e-01
4.33571935e-01 4.70776819e-02 5.18960416e-01 2.89199986e-02
3.07576686e-01 -3.94861609e-01 -1.17052996e+00 5.34606338e-01
-9.24265757e-02 -1.54278157e-02 5.98332763e-01 4.37977701e-01
5.50837696e-01 -3.22872192e-01 2.92746723e-01 -2.66111791e-01
1.73181847e-01 3.48853976e-01 -6.68563426e-01 3.08679402e-01
-3.29938531e-01 -2.24149168e-01 3.85152996e-01 -6.34489059e-01
-1.04639268e+00 5.52786171e-01 6.07057214e-02 -7.02837706e-01
1.38226867e-01 -2.18816802e-01 -1.62373811e-01 -4.91176397e-01
4.07178462e-01 1.51714653e-01 1.07319914e-01 -5.43569326e-01
1.43935770e-01 6.33495152e-01 8.28230917e-01 -4.17938441e-01
9.25285041e-01 5.79632163e-01 1.33659616e-01 -7.67314911e-01
-9.59238768e-01 -8.29849899e-01 -6.99525714e-01 -6.33016229e-01
8.89030039e-01 -1.56884098e+00 -1.34155726e+00 6.14292800e-01
-1.20308411e+00 2.68679619e-01 9.42473859e-02 5.99432170e-01
-2.00353727e-01 4.68756467e-01 -6.89316988e-01 -9.98946607e-01
-8.44932616e-01 -9.84316647e-01 1.50222087e+00 4.79363561e-01
-7.54863098e-02 -4.80643064e-01 -1.30908862e-01 4.28130209e-01
4.47999299e-01 3.24176461e-01 -2.71294296e-01 -2.73337275e-01
-6.28340065e-01 -3.80290359e-01 -5.49276531e-01 -1.86282858e-01
-1.04125597e-01 -5.08824348e-01 -9.00710702e-01 -9.42813933e-01
-3.70293975e-01 -7.19332471e-02 5.58288932e-01 5.48271358e-01
9.58795249e-01 -5.36011066e-03 -6.89056516e-01 8.01195621e-01
1.19399440e+00 -2.02600822e-01 2.31266022e-01 3.44496727e-01
1.20453906e+00 2.80004591e-01 6.89421177e-01 7.10810065e-01
6.92328870e-01 1.18853581e+00 1.63365722e-01 -1.11090839e-01
-6.22782826e-01 -3.05323899e-01 4.88095820e-01 5.29501677e-01
-3.61842275e-01 1.48117691e-01 -6.11334920e-01 1.41103476e-01
-1.95013452e+00 -8.83095503e-01 -6.62815794e-02 2.26889157e+00
4.46150631e-01 1.14934839e-01 6.18525803e-01 -2.59009659e-01
1.30100107e+00 2.95035858e-02 -7.72123814e-01 5.07458746e-01
2.20683828e-01 -1.81211576e-01 8.42205405e-01 2.96960831e-01
-1.44629681e+00 9.55316484e-01 5.78183651e+00 7.28649259e-01
-7.85698771e-01 1.96497500e-01 1.08335011e-01 -3.42356265e-01
5.48967779e-01 -3.93454552e-01 -1.52339089e+00 4.53883499e-01
6.89545333e-01 1.33530334e-01 1.77555397e-01 9.45450366e-01
2.21416727e-01 1.75041780e-02 -8.41251373e-01 1.54417396e+00
1.10049225e-01 -8.20137799e-01 -1.30127177e-01 8.34267884e-02
2.68371463e-01 -2.44687349e-01 -1.27302289e-01 3.47331762e-01
2.97349151e-02 -5.36723316e-01 7.61861205e-01 4.58746850e-01
9.20293450e-01 -9.01488841e-01 6.37447417e-01 7.84561262e-02
-2.14313984e+00 -2.11457551e-01 -4.52342212e-01 -1.67608224e-02
4.60232198e-01 3.27224076e-01 -4.42213804e-01 5.03296793e-01
9.20160055e-01 5.78346968e-01 -7.43237138e-01 1.16755986e+00
5.25320731e-02 -4.67785448e-02 -6.69405580e-01 1.26636729e-01
-2.32959434e-01 2.70452619e-01 6.31609499e-01 1.25563085e+00
1.80088043e-01 -4.39257249e-02 8.56680989e-01 4.23501611e-01
6.67182878e-02 -2.41394773e-01 3.13998684e-02 6.02747202e-01
6.91136181e-01 1.40735292e+00 -6.97675049e-01 -3.77033383e-01
-3.71867597e-01 1.08541620e+00 2.66483337e-01 6.83868602e-02
-1.27122509e+00 -2.17395604e-01 8.19857121e-01 3.14551234e-01
4.26708072e-01 -4.13790792e-01 1.49273247e-01 -1.10086203e+00
1.02497481e-01 -5.57459533e-01 4.49370652e-01 -4.44598287e-01
-1.08607292e+00 5.17266691e-01 8.06904361e-02 -1.32879782e+00
-6.59665540e-02 -4.39872175e-01 -6.79708347e-02 7.29409635e-01
-1.30853534e+00 -1.58307207e+00 -7.53451943e-01 7.29793489e-01
5.84704995e-01 9.17734671e-03 4.65874940e-01 8.11222374e-01
-9.18596923e-01 1.06090033e+00 -2.91968018e-01 4.31531727e-01
9.08960402e-01 -8.78647268e-01 5.03414690e-01 1.05408502e+00
-3.68876904e-01 6.06327891e-01 6.59403324e-01 -9.81550574e-01
-1.64221883e+00 -1.34356296e+00 5.22953093e-01 -3.81074548e-01
2.10498542e-01 -4.53541309e-01 -3.25161189e-01 5.67331493e-01
-4.42567468e-01 5.01412690e-01 4.29782450e-01 1.50001049e-02
-2.32099131e-01 -3.39419007e-01 -1.19996619e+00 5.94141483e-01
1.52855027e+00 -1.55378640e-01 -2.14284673e-01 2.58757502e-01
7.38002896e-01 -8.10825706e-01 -9.63700771e-01 3.60576957e-01
1.00395656e+00 -6.80002511e-01 1.42317104e+00 -1.99885458e-01
-4.54301655e-01 -5.50131261e-01 -1.01247560e-02 -6.97834074e-01
-6.45089209e-01 -6.36565447e-01 -4.62305397e-01 1.01445520e+00
-3.15869898e-01 -5.01327395e-01 1.05367994e+00 7.22528696e-01
6.92449957e-02 -5.08003652e-01 -1.10703111e+00 -9.22545791e-01
-7.35560715e-01 -3.82401906e-02 5.20523965e-01 1.73221484e-01
-5.36320806e-01 4.57893103e-01 -7.70127177e-01 4.06777382e-01
1.14727557e+00 1.22174107e-01 1.18114364e+00 -1.35281861e+00
-1.23914301e-01 7.47715905e-02 -7.05803633e-01 -1.53639483e+00
-2.29855418e-01 -3.59944314e-01 5.88575900e-02 -1.18499100e+00
3.27124596e-01 -1.82106704e-01 -1.19276367e-01 3.13981533e-01
-3.06028247e-01 4.38694865e-01 6.95486188e-01 3.43767077e-01
-1.04411387e+00 5.69016218e-01 1.25445271e+00 1.77520916e-01
-1.86405048e-01 1.06077611e-01 -3.78321797e-01 8.35714638e-01
4.76330608e-01 -3.19257587e-01 -1.68058202e-01 -4.92460161e-01
-3.35120082e-01 3.15131769e-02 7.19044626e-01 -1.59250593e+00
6.90306365e-01 2.77994841e-01 1.02299571e+00 -1.08309734e+00
9.13395226e-01 -8.22020233e-01 2.22378284e-01 8.84894550e-01
1.40772715e-01 4.65953410e-01 2.83554703e-01 6.48501515e-01
2.02296808e-01 4.16475803e-01 8.29763293e-01 -8.04799572e-02
-8.48761439e-01 8.39744687e-01 2.27637082e-01 -4.09567468e-02
1.09219110e+00 -5.06914794e-01 -1.33285224e-01 -1.46811128e-01
-6.89300776e-01 4.43781137e-01 2.14788333e-01 3.80801946e-01
6.57775760e-01 -1.56414866e+00 -6.98523760e-01 8.46831650e-02
-1.28360301e-01 -8.77359137e-02 5.07351816e-01 9.39583421e-01
-3.90246451e-01 4.83040929e-01 -2.36764163e-01 -8.76083732e-01
-1.61170638e+00 2.90045857e-01 3.11172634e-01 -4.41646546e-01
-7.23182499e-01 8.05224419e-01 5.93079254e-02 -2.31875405e-01
4.74271774e-01 1.30709648e-01 -1.47507489e-01 -8.70394185e-02
8.51951122e-01 8.31694424e-01 -2.68175215e-01 -9.46189284e-01
-7.51389921e-01 9.81604338e-01 -2.61437118e-01 3.07897367e-02
1.03794003e+00 -4.62691158e-01 3.13693702e-01 -2.40309998e-01
9.18318570e-01 -1.72585338e-01 -1.73316455e+00 -2.66952008e-01
-2.74868250e-01 -7.45160818e-01 -2.95718089e-02 -5.64759195e-01
-1.24056828e+00 4.53005254e-01 1.10015380e+00 -4.97916013e-01
9.63731766e-01 -2.76513249e-01 1.17277098e+00 2.71590114e-01
6.69974804e-01 -1.23078942e+00 1.59328088e-01 3.35393369e-01
5.96545279e-01 -1.03653669e+00 3.75384659e-01 -7.56570935e-01
-2.76620150e-01 9.54734623e-01 1.10388637e+00 -2.30743095e-01
3.70979041e-01 2.73046851e-01 -2.19185725e-01 -1.11134261e-01
-1.96184129e-01 -3.00719529e-01 4.03200895e-01 7.41413891e-01
1.82319582e-01 3.04215942e-02 -1.60505220e-01 4.37486678e-01
-1.86968371e-01 5.89595102e-02 -1.32070974e-01 8.27829003e-01
-5.70121169e-01 -7.73292422e-01 -7.98226655e-01 2.85925031e-01
-3.80467296e-01 2.66102970e-01 9.12756398e-02 7.44193614e-01
2.00434417e-01 7.85402775e-01 7.98139814e-03 -4.58286643e-01
5.52612662e-01 -3.51604939e-01 7.13375509e-01 -3.74952316e-01
-6.53774619e-01 2.49408469e-01 1.04366884e-01 -9.81164873e-01
-5.39346933e-01 -6.59499884e-01 -1.10905850e+00 -5.89329183e-01
-3.43749940e-01 -2.23749205e-01 4.56243634e-01 7.29332387e-01
6.61690295e-01 5.71745276e-01 2.52227336e-01 -1.19450104e+00
-7.73471236e-01 -9.62063074e-01 -2.54321754e-01 3.90819192e-01
1.38133213e-01 -1.09976733e+00 2.09039226e-01 -1.18710816e-01] | [7.063582420349121, -0.9057454466819763] |
f79c1507-28c8-4ff0-8c56-59089fdccc21 | florence-a-new-foundation-model-for-computer | 2111.11432 | null | https://arxiv.org/abs/2111.11432v1 | https://arxiv.org/pdf/2111.11432v1.pdf | Florence: A New Foundation Model for Computer Vision | Automated visual understanding of our diverse and open world demands computer vision models to generalize well with minimal customization for specific tasks, similar to human vision. Computer vision foundation models, which are trained on diverse, large-scale dataset and can be adapted to a wide range of downstream tasks, are critical for this mission to solve real-world computer vision applications. While existing vision foundation models such as CLIP, ALIGN, and Wu Dao 2.0 focus mainly on mapping images and textual representations to a cross-modal shared representation, we introduce a new computer vision foundation model, Florence, to expand the representations from coarse (scene) to fine (object), from static (images) to dynamic (videos), and from RGB to multiple modalities (caption, depth). By incorporating universal visual-language representations from Web-scale image-text data, our Florence model can be easily adapted for various computer vision tasks, such as classification, retrieval, object detection, VQA, image caption, video retrieval and action recognition. Moreover, Florence demonstrates outstanding performance in many types of transfer learning: fully sampled fine-tuning, linear probing, few-shot transfer and zero-shot transfer for novel images and objects. All of these properties are critical for our vision foundation model to serve general purpose vision tasks. Florence achieves new state-of-the-art results in majority of 44 representative benchmarks, e.g., ImageNet-1K zero-shot classification with top-1 accuracy of 83.74 and the top-5 accuracy of 97.18, 62.4 mAP on COCO fine tuning, 80.36 on VQA, and 87.8 on Kinetics-600. | ['Pengchuan Zhang', 'Luowei Zhou', 'Michael Zeng', 'Jianwei Yang', 'Zhen Xiao', 'Bin Xiao', 'JianFeng Wang', 'Lijuan Wang', 'Yu Shi', 'Yumao Lu', 'Zicheng Liu', 'Mengchen Liu', 'Ce Liu', 'Chunyuan Li', 'Boxin Li', 'Xuedong Huang', 'Houdong Hu', 'Jianfeng Gao', 'Xiyang Dai', 'Noel Codella', 'Yi-Ling Chen', 'Dongdong Chen', 'Lu Yuan'] | 2021-11-22 | null | null | null | null | ['zero-shot-transfer-image-classification', 'zero-shot-cross-modal-retrieval'] | ['computer-vision', 'miscellaneous'] | [ 1.93096429e-01 -4.04622912e-01 -3.86307061e-01 -2.46375993e-01
-9.63744044e-01 -5.84758699e-01 7.89671421e-01 -4.09486413e-01
-2.52539575e-01 6.79580033e-01 1.48665294e-01 -1.38034731e-01
1.53273702e-01 -7.40555882e-01 -9.66329157e-01 -5.47148764e-01
2.17511997e-01 4.26808387e-01 5.30981541e-01 -4.02160853e-01
2.17349425e-01 2.83377141e-01 -1.81300819e+00 6.70235574e-01
2.90794551e-01 1.18756676e+00 4.11654800e-01 9.40122724e-01
-1.49300888e-01 7.42831945e-01 -3.65613371e-01 -1.85429052e-01
3.25883865e-01 -1.26815334e-01 -1.01609063e+00 1.10385530e-01
8.06015074e-01 -6.07710361e-01 -5.18150985e-01 1.01513064e+00
5.18265545e-01 2.84518898e-01 7.94026375e-01 -1.58636975e+00
-1.54217231e+00 7.94748962e-02 -8.18431914e-01 4.09066767e-01
3.01316291e-01 7.27316201e-01 8.67753685e-01 -1.17001164e+00
8.73861015e-01 1.48102880e+00 4.68685061e-01 1.03899550e+00
-1.15712810e+00 -7.47712135e-01 6.99974895e-02 5.76896429e-01
-1.00715709e+00 -4.48225319e-01 1.92004353e-01 -5.69412470e-01
1.35761368e+00 6.46238774e-02 4.74658161e-01 1.56374586e+00
1.70764655e-01 9.04920280e-01 1.05912519e+00 -1.90343723e-01
-5.29433554e-03 9.23976768e-03 1.34211063e-01 6.24515474e-01
-2.45430600e-02 -1.05443420e-02 -6.87529922e-01 2.88230181e-01
7.41797388e-01 1.53493196e-01 -3.33272129e-01 -3.48831683e-01
-1.38390923e+00 8.50981951e-01 7.34395921e-01 -1.12010576e-01
-8.38771388e-02 5.29412627e-01 7.02124476e-01 3.58919650e-01
1.14353523e-01 2.05849379e-01 -6.11957371e-01 -2.34506756e-01
-5.90464950e-01 -5.14335893e-02 3.71065944e-01 1.28984272e+00
9.82272923e-01 2.08981916e-01 -5.93534708e-01 9.60747361e-01
3.23459923e-01 9.12048876e-01 7.46634305e-01 -1.08340430e+00
4.79316443e-01 3.39865744e-01 -1.57873258e-01 -4.76685822e-01
-8.12030584e-02 -1.02646895e-01 -9.37469661e-01 1.83500394e-01
-1.80707760e-02 1.82248011e-01 -1.70812809e+00 1.71507120e+00
1.40600115e-01 3.81456405e-01 2.50137031e-01 9.73846614e-01
1.46881509e+00 1.02530909e+00 3.05202365e-01 2.08729416e-01
1.38837004e+00 -1.52447653e+00 -1.47562176e-01 -5.98423243e-01
2.34870777e-01 -6.70474946e-01 1.42768490e+00 2.24190727e-01
-1.02568400e+00 -9.50647056e-01 -7.77485490e-01 -4.22288150e-01
-7.27188349e-01 -1.32327810e-01 6.98805392e-01 3.61107200e-01
-1.32215512e+00 1.09215237e-01 -5.17822206e-01 -9.78335023e-01
6.39393628e-01 1.22373492e-01 -6.64727330e-01 -5.91106176e-01
-1.01164925e+00 8.86538506e-01 4.37405825e-01 -2.62330383e-01
-1.54828691e+00 -6.64546251e-01 -8.87811780e-01 1.24383308e-01
1.30184054e-01 -1.15121448e+00 1.10091555e+00 -9.35186327e-01
-1.26221788e+00 1.18102145e+00 2.77176827e-01 -4.60164636e-01
2.37092659e-01 -1.66086361e-01 -3.63013297e-01 3.34723800e-01
3.43495876e-01 1.43600965e+00 9.72880900e-01 -1.12344086e+00
-6.61906660e-01 -3.24203879e-01 1.52443036e-01 7.74640217e-02
-3.80001277e-01 -7.95229897e-02 -7.70598292e-01 -3.24941844e-01
-3.95581484e-01 -8.98286879e-01 -6.07414637e-03 4.02090698e-01
5.47518395e-02 -2.75419950e-01 1.13990211e+00 -4.09665316e-01
5.15815735e-01 -2.06217790e+00 1.50067911e-01 -5.11631250e-01
7.16095492e-02 5.65803885e-01 -6.11850023e-01 4.01268929e-01
-1.31973380e-03 -1.18551895e-01 -1.43084489e-03 -1.26179412e-01
-1.61926597e-01 4.03767526e-01 -4.35863942e-01 3.20856631e-01
4.69670862e-01 1.51297033e+00 -9.18576002e-01 -3.76712531e-01
6.10982060e-01 2.62728900e-01 -4.48969096e-01 2.31704623e-01
-2.46449098e-01 1.76584959e-01 -3.86352539e-01 9.00935173e-01
4.81135964e-01 -7.88787484e-01 -5.32276928e-01 -2.64097750e-01
1.75828189e-01 -6.17953897e-01 -6.58959687e-01 2.16742969e+00
-4.46171731e-01 8.04688096e-01 -6.15487806e-02 -1.08441937e+00
9.41865683e-01 -2.33584940e-02 2.44582415e-01 -8.05443764e-01
7.04544187e-02 -1.61530405e-01 -3.11589032e-01 -7.12148428e-01
3.94952774e-01 1.71351805e-01 -1.18446849e-01 2.13239178e-01
6.72717810e-01 -6.05844222e-02 1.68024510e-01 5.48032284e-01
1.12291670e+00 1.52297795e-01 1.84012055e-01 2.95902765e-03
3.18741232e-01 1.66014537e-01 1.79066420e-01 8.37573767e-01
-5.73202431e-01 9.54120994e-01 -9.48818251e-02 -3.91285479e-01
-1.07147312e+00 -1.38212693e+00 -4.30897027e-02 1.60216510e+00
3.10440511e-01 -1.75972313e-01 -5.05183220e-01 -3.19326192e-01
1.31529346e-01 3.38798195e-01 -9.53070879e-01 -6.03748679e-01
8.10820423e-03 -3.90297383e-01 5.75838149e-01 8.57275486e-01
9.30697858e-01 -1.38686109e+00 -7.04393327e-01 5.20185307e-02
-1.92394659e-01 -1.37570596e+00 -4.45083082e-01 9.91284009e-03
-6.30109847e-01 -9.67985928e-01 -9.21240985e-01 -9.58489358e-01
1.39268875e-01 9.03491557e-01 1.21853912e+00 -2.47963712e-01
-8.71505558e-01 1.17080462e+00 -5.34465969e-01 -9.65300277e-02
-6.73457608e-02 -2.74324119e-01 5.87735185e-03 7.03811049e-02
4.11436737e-01 -3.54837090e-01 -7.11984575e-01 3.02257299e-01
-1.00055885e+00 8.68923590e-02 7.53545523e-01 1.15542388e+00
5.55415750e-01 -1.05956340e+00 6.60110950e-01 -4.51755911e-01
3.27546418e-01 -7.85795510e-01 -2.93457776e-01 7.15932012e-01
-4.15579945e-01 -2.18231663e-01 3.25741470e-01 -6.82458282e-01
-1.02488208e+00 1.57554653e-02 1.91839576e-01 -1.04817450e+00
-2.66789526e-01 1.80957720e-01 8.87220427e-02 -1.81440666e-01
1.13000643e+00 6.28407478e-01 -1.66438408e-02 -2.12470040e-01
9.70086038e-01 8.41980875e-01 9.58615661e-01 -5.41515291e-01
6.12834513e-01 5.72584093e-01 -2.30509207e-01 -9.27420616e-01
-8.45809937e-01 -7.77915955e-01 -4.62993085e-01 -2.37143338e-01
1.34658515e+00 -1.15703011e+00 -5.03733814e-01 6.07865989e-01
-1.08071709e+00 -3.66354436e-01 -3.46945822e-01 1.09394275e-01
-8.76816213e-01 3.28909725e-01 -5.69366395e-01 -3.18180054e-01
-7.37367868e-01 -1.23726666e+00 1.38476264e+00 4.61578995e-01
1.59491852e-01 -7.00204670e-01 1.40938476e-01 7.61665881e-01
5.87792933e-01 1.56634584e-01 6.89775407e-01 -2.14994580e-01
-7.69404292e-01 1.59576029e-01 -9.44460690e-01 4.79051411e-01
-9.32198018e-02 5.74475676e-02 -1.10415912e+00 -4.95970070e-01
-6.71580136e-01 -1.33041847e+00 1.33661842e+00 3.21645439e-01
1.28443623e+00 1.86562434e-01 -3.67657721e-01 1.07431400e+00
1.48441064e+00 -6.33880845e-04 8.57666969e-01 4.33199763e-01
6.29313707e-01 1.49983108e-01 5.53013265e-01 4.88801956e-01
4.24427599e-01 4.93385136e-01 7.53648102e-01 4.52183671e-02
-6.13630712e-01 -1.70965821e-01 5.78858554e-01 3.97267044e-01
-1.31528512e-01 -1.94011733e-01 -9.15791869e-01 7.98736572e-01
-1.99581730e+00 -9.40001905e-01 8.70783776e-02 1.73964953e+00
5.62012315e-01 -7.86446333e-02 -9.69761238e-02 -7.14387655e-01
6.75218284e-01 1.79741621e-01 -1.04319179e+00 -3.03211421e-01
-1.94130793e-01 2.20455572e-01 4.64604050e-01 -2.10266069e-01
-1.05983436e+00 1.39519429e+00 6.30605221e+00 8.43953133e-01
-1.10564387e+00 3.54528964e-01 5.92489898e-01 -1.12176389e-01
1.47684757e-02 -2.55964518e-01 -8.32324028e-01 1.27070487e-01
9.04801548e-01 -6.22060634e-02 4.94983882e-01 1.04221702e+00
-4.20218319e-01 5.73591962e-02 -1.17185223e+00 1.55017769e+00
3.38091880e-01 -1.66906202e+00 4.49616581e-01 -1.97056442e-01
1.02495527e+00 8.17414939e-01 3.65828633e-01 8.80310476e-01
4.29588526e-01 -1.22104871e+00 4.84773517e-01 5.46357453e-01
1.32809925e+00 -1.53935283e-01 3.60148549e-01 6.88914806e-02
-1.13332069e+00 -1.50013864e-01 -6.11880302e-01 1.39428154e-01
6.49871528e-02 -3.29273373e-01 -4.82473850e-01 2.10341781e-01
1.27129471e+00 1.12186384e+00 -7.31049180e-01 8.95094156e-01
1.05468392e-01 3.79986703e-01 3.27712782e-02 8.26868489e-02
5.90421438e-01 1.75223798e-01 2.40360305e-01 1.20315278e+00
1.71199650e-01 1.70979142e-01 2.51589119e-01 7.74695277e-01
-1.24980949e-01 -1.79157063e-01 -7.78556466e-01 -6.69352361e-04
1.50621817e-01 1.34522963e+00 -4.78000015e-01 -5.14897287e-01
-7.50285029e-01 1.21727490e+00 4.60129499e-01 5.79161167e-01
-9.58700418e-01 -1.81827620e-01 8.64301741e-01 -2.63974935e-01
5.67850530e-01 -1.34494245e-01 2.78369516e-01 -1.36777818e+00
-2.75738895e-01 -8.00697088e-01 5.43605089e-01 -1.31186819e+00
-1.58297849e+00 5.58617353e-01 8.11627358e-02 -1.16928506e+00
-1.66476831e-01 -1.04393327e+00 -5.68212450e-01 5.03614843e-01
-1.58822668e+00 -1.59740818e+00 -7.81891286e-01 1.35319078e+00
1.09414709e+00 -5.30463040e-01 8.82954061e-01 1.74428672e-01
-3.12997371e-01 5.53554773e-01 2.01497763e-01 2.11328417e-01
1.01185381e+00 -9.92818773e-01 5.42892218e-01 4.87619877e-01
2.82562345e-01 2.36011550e-01 3.12654853e-01 -2.52598643e-01
-1.76895916e+00 -1.37243617e+00 -1.49530515e-01 -5.72641492e-01
8.02107871e-01 -1.79872200e-01 -8.91923964e-01 7.36247778e-01
3.66357267e-01 8.27868521e-01 4.17328060e-01 -1.10860534e-01
-9.10485566e-01 -1.67660922e-01 -1.04326665e+00 3.73471558e-01
1.28072655e+00 -5.96947491e-01 -5.69967031e-01 5.26584804e-01
1.00841093e+00 -2.71386296e-01 -7.87831306e-01 1.83155254e-01
6.80122554e-01 -9.08765376e-01 1.30899549e+00 -1.24781847e+00
7.23914444e-01 -1.28561765e-01 -6.35088384e-01 -1.14553535e+00
-6.73263490e-01 -2.17629343e-01 -2.83227742e-01 1.04729259e+00
2.23552197e-01 -4.58837986e-01 5.65709472e-01 3.28512371e-01
-2.97124445e-01 -5.87251902e-01 -8.12219858e-01 -8.79890621e-01
6.02018572e-02 -2.89412081e-01 1.88679054e-01 8.23309004e-01
-5.12735248e-01 6.55481815e-01 -6.13034129e-01 -1.18955322e-01
5.92349887e-01 1.84762016e-01 1.08631980e+00 -9.63800669e-01
-4.64290887e-01 -3.99665087e-01 -7.79627144e-01 -1.07878089e+00
-3.75922001e-03 -8.58910501e-01 -7.21514598e-02 -1.85670471e+00
6.35643959e-01 -4.75458726e-02 -6.24339581e-01 6.58778906e-01
8.38033706e-02 7.19838262e-01 5.85333347e-01 3.23774070e-01
-1.06676257e+00 8.37311327e-01 1.45521915e+00 -6.99736118e-01
1.65423542e-01 -3.96538675e-01 -6.68386757e-01 4.92467910e-01
5.49170315e-01 1.87774505e-02 -4.98067737e-01 -7.01143563e-01
-1.32022813e-01 -4.79765497e-02 6.71642601e-01 -1.09659123e+00
1.56636626e-01 -4.01374638e-01 5.52863300e-01 -3.89405012e-01
7.18468606e-01 -5.66047549e-01 -1.41248137e-01 4.23039377e-01
-1.12528831e-01 -4.83691655e-02 4.18568939e-01 8.69832397e-01
-2.34911919e-01 -8.67383853e-02 9.73949969e-01 -4.30981845e-01
-1.80508101e+00 6.81044221e-01 -1.75845418e-02 3.28875750e-01
1.47699559e+00 -3.01650137e-01 -1.06204283e+00 -4.27267730e-01
-7.99515486e-01 4.17759478e-01 2.98564345e-01 1.00582671e+00
9.53011811e-01 -1.37771404e+00 -8.81004155e-01 8.55647251e-02
8.66717041e-01 -2.34900668e-01 5.49666643e-01 4.67755258e-01
-4.05434370e-01 4.90510464e-01 -8.74430656e-01 -1.01194680e+00
-1.13492370e+00 8.46797884e-01 6.99871406e-02 1.32020861e-01
-5.89246988e-01 1.10467935e+00 6.31375492e-01 -2.44139209e-01
7.80517831e-02 -2.09070429e-01 -7.27888197e-03 -1.53606400e-01
7.80854344e-01 2.56799132e-01 -2.06110045e-01 -7.22272217e-01
-3.78958672e-01 8.35676312e-01 -2.42314562e-01 2.23824233e-01
1.31532276e+00 -2.21306995e-01 1.05712384e-01 3.78749669e-01
1.36878860e+00 -9.56570268e-01 -1.47533965e+00 -3.75404447e-01
-5.74479043e-01 -4.03109163e-01 -7.45964572e-02 -9.74440575e-01
-1.06015790e+00 1.27453244e+00 9.37138319e-01 -1.60479397e-01
1.15626454e+00 3.57818812e-01 8.28787506e-01 5.97081542e-01
5.94223559e-01 -9.26509500e-01 6.13474548e-01 7.01211810e-01
9.83325899e-01 -1.78663313e+00 -3.31179142e-01 1.29964888e-01
-9.05116260e-01 9.19389963e-01 1.19134414e+00 -1.01054743e-01
4.00041014e-01 -2.54061043e-01 2.96541639e-02 -7.95984864e-02
-1.17701423e+00 -3.89484763e-01 3.93671781e-01 1.02466691e+00
-4.61019315e-02 -6.25244528e-02 4.74114448e-01 1.92774862e-01
2.40251139e-01 -7.64602828e-06 4.31338340e-01 6.86955094e-01
-7.73702323e-01 -5.96367836e-01 -1.70977294e-01 4.77503031e-01
4.28703651e-02 -1.87914386e-01 -2.64427871e-01 7.69985795e-01
-2.47122515e-02 6.71582699e-01 9.54249278e-02 -3.57588738e-01
7.96039030e-02 -4.86546289e-03 6.71559989e-01 -9.14702475e-01
-2.67427504e-01 -3.21378529e-01 -3.14078987e-01 -6.91836536e-01
-4.74024653e-01 -3.71613562e-01 -9.53363180e-01 -5.38923293e-02
-7.57175162e-02 -4.82964277e-01 6.10179961e-01 8.81285906e-01
5.79588532e-01 4.41050678e-01 1.37835309e-01 -9.70872462e-01
-6.10612929e-01 -1.11566627e+00 -3.88855726e-01 5.61853588e-01
3.26224536e-01 -8.02330971e-01 -2.57548653e-02 3.10563475e-01] | [10.326523780822754, 1.7392995357513428] |
dc683019-bb0d-405a-a57a-39876af4dc0c | robust-segmentation-models-using-an | 2109.14879 | null | https://arxiv.org/abs/2109.14879v1 | https://arxiv.org/pdf/2109.14879v1.pdf | Robust Segmentation Models using an Uncertainty Slice Sampling Based Annotation Workflow | Semantic segmentation neural networks require pixel-level annotations in large quantities to achieve a good performance. In the medical domain, such annotations are expensive, because they are time-consuming and require expert knowledge. Active learning optimizes the annotation effort by devising strategies to select cases for labeling that are most informative to the model. In this work, we propose an uncertainty slice sampling (USS) strategy for semantic segmentation of 3D medical volumes that selects 2D image slices for annotation and compare it with various other strategies. We demonstrate the efficiency of USS on a CT liver segmentation task using multi-site data. After five iterations, the training data resulting from USS consisted of 2410 slices (4% of all slices in the data pool) compared to 8121 (13%), 8641 (14%), and 3730 (6%) for uncertainty volume (UVS), random volume (RVS), and random slice (RSS) sampling, respectively. Despite being trained on the smallest amount of data, the model based on the USS strategy evaluated on 234 test volumes significantly outperformed models trained according to other strategies and achieved a mean Dice index of 0.964, a relative volume error of 4.2%, a mean surface distance of 1.35 mm, and a Hausdorff distance of 23.4 mm. This was only slightly inferior to 0.967, 3.8%, 1.18 mm, and 22.9 mm achieved by a model trained on all available data, but the robustness analysis using the 5th percentile of Dice and the 95th percentile of the remaining metrics demonstrated that USS resulted not only in the most robust model compared to other sampling schemes, but also outperformed the model trained on all data according to Dice (0.946 vs. 0.945) and mean surface distance (1.92 mm vs. 2.03 mm). | ['Hans Meine', 'Bram van Ginneken', 'Horst K. Hahn', 'Andrea Schenk', 'Grzegorz Chlebus'] | 2021-09-30 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 1.60123467e-01 5.87354898e-01 -4.83739711e-02 -3.70687366e-01
-1.17013371e+00 -5.75112343e-01 2.73181319e-01 6.67358100e-01
-8.04690301e-01 9.24031019e-01 -9.87539515e-02 -3.55507165e-01
-3.53176534e-01 -7.23630607e-01 -5.94141424e-01 -9.62796450e-01
-4.10728455e-01 6.34588301e-01 4.09367800e-01 5.81128776e-01
9.44840536e-02 3.40588629e-01 -7.61665583e-01 -4.25535850e-02
1.22808957e+00 1.28776133e+00 2.93317348e-01 2.65215814e-01
-5.84346429e-02 4.25936043e-01 -7.70050764e-01 -2.59519845e-01
3.94953519e-01 -5.52131176e-01 -7.03777611e-01 -3.26131321e-02
2.30348632e-01 -3.73717010e-01 1.76030368e-01 9.99931335e-01
6.14311218e-01 2.63479978e-01 8.74092937e-01 -7.01330543e-01
-2.92607039e-01 8.36675704e-01 -5.71555376e-01 1.49100125e-01
-3.35713997e-02 1.07646555e-01 4.87581730e-01 -6.75125420e-01
8.08093965e-01 5.55402100e-01 9.22207415e-01 5.52816451e-01
-1.11031759e+00 -4.64837402e-01 -2.82650501e-01 -2.10569531e-01
-1.50982153e+00 -1.96778089e-01 2.78169930e-01 -6.48930550e-01
5.80430686e-01 4.78468925e-01 8.77247334e-01 5.22037923e-01
3.62667710e-01 5.62491238e-01 1.07765853e+00 -4.63362217e-01
5.87241828e-01 2.33155549e-01 2.35433714e-03 8.91701937e-01
5.80379605e-01 -1.94689155e-01 1.38684347e-01 -2.55890310e-01
9.53362286e-01 -2.22459555e-01 -4.55779821e-01 -3.70379746e-01
-1.28543568e+00 9.48069692e-01 6.28296494e-01 4.94099289e-01
-5.24293005e-01 -2.31222585e-01 4.15619999e-01 -2.80112475e-01
7.21699357e-01 7.09261358e-01 -4.60301816e-01 2.42229670e-01
-1.36722624e+00 -3.11408699e-01 8.26646626e-01 6.02467120e-01
2.77860522e-01 -3.39341052e-02 -2.25040123e-01 8.02167237e-01
3.98596466e-01 2.13979483e-01 5.53270996e-01 -1.07904959e+00
2.08910584e-01 5.66419721e-01 -7.21794665e-02 -7.77066469e-01
-7.28823066e-01 -7.63621569e-01 -1.00715292e+00 2.76407748e-02
8.53948236e-01 -3.11360002e-01 -1.18669939e+00 1.45751274e+00
2.23939970e-01 -2.35102519e-01 -1.99027449e-01 9.59675610e-01
1.01379967e+00 3.36576045e-01 3.15351278e-01 -6.46633208e-01
1.00400114e+00 -6.99247777e-01 -5.36378920e-01 1.69536516e-01
7.71665752e-01 -5.31785846e-01 8.64502370e-01 2.36438125e-01
-1.17203867e+00 -2.95653313e-01 -1.07836628e+00 5.92049122e-01
-3.38676870e-02 -1.66538171e-02 4.31818992e-01 7.73654878e-01
-1.26307380e+00 6.73674047e-01 -1.05254149e+00 -2.76163757e-01
7.03816950e-01 3.14397067e-01 -2.16740787e-01 1.72618672e-01
-9.52285290e-01 9.26009595e-01 5.95986724e-01 -9.41596106e-02
-7.58910120e-01 -9.55642879e-01 -8.11872661e-01 -5.63453957e-02
4.49889094e-01 -4.81247485e-01 9.28867400e-01 -7.88428724e-01
-1.32642758e+00 8.66201103e-01 2.68523484e-01 -7.91233540e-01
9.68067408e-01 1.59385428e-01 1.84702054e-02 5.23286402e-01
1.09015658e-01 8.21351111e-01 2.10221797e-01 -1.27688849e+00
-8.50623026e-02 -3.42867941e-01 -2.32953280e-01 1.88649058e-01
3.08231805e-02 -7.56921768e-02 -3.73909652e-01 -4.90343183e-01
6.02752149e-01 -9.25110340e-01 -5.63231349e-01 5.80873154e-02
-4.70271438e-01 8.50862712e-02 1.86525732e-01 -9.10182238e-01
1.04786026e+00 -1.92346537e+00 -3.19989949e-01 4.63249713e-01
3.89701456e-01 1.58872843e-01 4.91292536e-01 -3.70354295e-01
2.85533001e-03 4.06500667e-01 -9.42400277e-01 -1.30014166e-01
-3.48313093e-01 -4.72290516e-02 5.67117035e-01 5.38245440e-01
-1.50117412e-01 7.72811413e-01 -1.11678290e+00 -9.15155888e-01
3.98379356e-01 3.59921992e-01 -3.57717276e-01 2.07223073e-02
2.30259508e-01 7.26644933e-01 -1.93371713e-01 6.16156757e-01
6.19677961e-01 -4.15821731e-01 1.96843475e-01 -8.68816748e-02
8.66426080e-02 -1.21689223e-01 -9.50856984e-01 1.75801301e+00
-2.90359467e-01 5.10500610e-01 -1.38005584e-01 -8.48132253e-01
1.03980446e+00 6.21971309e-01 1.11281168e+00 -7.51903951e-01
1.65038630e-01 6.07082844e-01 3.29495400e-01 -3.87806714e-01
7.89269712e-03 -2.49587432e-01 -5.98491961e-03 3.81493211e-01
1.13775246e-01 -2.27630496e-01 1.75129145e-01 2.12691441e-01
9.66022253e-01 -1.01017572e-01 5.08237958e-01 -6.67022407e-01
4.23780143e-01 1.68678146e-02 6.35277390e-01 8.24329615e-01
-6.58321321e-01 8.60094190e-01 4.70050514e-01 -2.64616162e-01
-9.01329517e-01 -1.18900263e+00 -6.20014489e-01 2.01162711e-01
6.22799248e-02 -1.93151280e-01 -1.19878674e+00 -1.14416182e+00
-3.22689742e-01 9.77345467e-01 -5.69392562e-01 5.63675426e-02
-4.52721655e-01 -1.14941132e+00 4.65321690e-01 4.47402447e-01
5.98699450e-01 -9.03481960e-01 -9.09753740e-01 1.38282925e-01
-1.93659022e-01 -8.69129479e-01 -2.81884253e-01 1.80188179e-01
-1.27666569e+00 -1.19857657e+00 -1.13473725e+00 -3.55034351e-01
8.88159275e-01 -4.46002662e-01 1.28080630e+00 1.05839945e-01
-1.92999944e-01 1.62098601e-01 -3.24517936e-01 -3.09591264e-01
-4.52228934e-01 -2.52352245e-02 -2.06454799e-01 -3.60054106e-01
-7.86209628e-02 -1.05299287e-01 -8.44835222e-01 4.05331373e-01
-8.94273698e-01 1.98226310e-02 4.33357894e-01 9.56848443e-01
9.41164315e-01 -2.62115877e-02 5.45574486e-01 -9.93775845e-01
1.71497151e-01 -5.02146602e-01 -6.22342348e-01 2.89548784e-01
-1.02592671e+00 -3.56489450e-01 4.20016706e-01 -1.98610350e-01
-1.05797434e+00 2.39826575e-01 -3.33273262e-02 -3.09119403e-01
-1.70691416e-01 4.85560119e-01 3.67515862e-01 -3.96809317e-02
7.98361003e-01 -1.19099207e-01 1.51136518e-01 -1.76748708e-01
-1.59606591e-01 4.46219653e-01 6.44642785e-02 -2.45382130e-01
1.91470504e-01 3.24358135e-01 -6.35505095e-02 -5.59490025e-01
-6.98813438e-01 -2.89408028e-01 -8.69924724e-01 -3.47234219e-01
1.09039068e+00 -4.53113168e-01 -2.03959092e-01 3.07135910e-01
-6.94229722e-01 -5.11027098e-01 -6.28506601e-01 9.13304985e-01
-5.97919285e-01 4.00140166e-01 -5.34001648e-01 -7.51706898e-01
-7.45384216e-01 -1.56558728e+00 5.77701747e-01 1.70420751e-01
-4.50143397e-01 -1.19820666e+00 -2.96981812e-01 4.01962668e-01
5.28409481e-01 6.82435513e-01 7.94304430e-01 -1.00084603e+00
-3.60755801e-01 -5.73250614e-02 -2.15920970e-01 4.23922330e-01
1.56678200e-01 -1.60406217e-01 -7.42378652e-01 -1.52536273e-01
3.21403116e-01 1.24902576e-02 6.88764930e-01 1.00610566e+00
1.48367572e+00 -1.23779461e-01 -1.43739119e-01 5.45625806e-01
1.57935536e+00 8.60742390e-01 4.63827491e-01 2.16696545e-01
4.05687362e-01 4.10175711e-01 5.25979698e-01 2.63625264e-01
-5.10806926e-02 3.66765857e-01 5.90691805e-01 -2.44473755e-01
-1.26093358e-01 2.12318480e-01 -2.08375812e-01 7.19202101e-01
-1.47532389e-01 -8.54305923e-02 -1.41362715e+00 4.78431225e-01
-1.57648671e+00 -4.18141425e-01 -2.20665440e-01 2.51240659e+00
7.56925881e-01 2.97705084e-01 -3.37828286e-02 3.63968164e-02
7.30267406e-01 -1.27892375e-01 -4.70932633e-01 -2.39257812e-01
1.20174386e-01 1.99641049e-01 8.55842829e-01 4.00659561e-01
-1.25872719e+00 4.22658026e-01 5.89391327e+00 7.47103930e-01
-9.53910887e-01 1.58851892e-01 1.16714191e+00 4.48486060e-02
-4.86052483e-02 -2.65168428e-01 -2.46698961e-01 7.99728572e-01
1.04778445e+00 2.52928473e-02 -7.22727999e-02 7.20476031e-01
9.33619812e-02 -5.92649877e-01 -8.64333332e-01 7.81031370e-01
4.76866439e-02 -1.31075394e+00 -2.70901501e-01 -3.98443229e-02
1.04437602e+00 1.20737486e-01 -1.14392906e-01 -4.43948731e-02
-5.82454866e-03 -1.16244364e+00 6.71107411e-01 4.96461898e-01
9.84616160e-01 -3.91911447e-01 1.31692612e+00 4.44665641e-01
-7.20511615e-01 3.13860953e-01 -1.08017296e-01 4.47452545e-01
1.97977111e-01 9.61259425e-01 -1.34261715e+00 6.76033854e-01
7.21532762e-01 1.98609367e-01 -4.34032053e-01 1.40819597e+00
9.17618498e-02 6.95214272e-01 -5.89986742e-01 -3.11725549e-02
3.29021543e-01 -3.04933906e-01 4.95427608e-01 1.16875815e+00
4.29551363e-01 2.13186368e-02 1.30117685e-02 8.42290580e-01
-5.42332642e-02 3.59915167e-01 -1.06820010e-01 2.74252921e-01
3.96888763e-01 1.15198064e+00 -1.49479532e+00 -5.63025832e-01
-1.85100049e-01 6.25215054e-01 -1.97027668e-01 1.36047482e-01
-1.01185238e+00 -3.00758153e-01 -3.39033753e-01 1.35843024e-01
-2.27012455e-01 1.54327258e-01 -6.91800475e-01 -7.54062116e-01
-9.97479483e-02 -3.84843618e-01 7.55168855e-01 -5.11164010e-01
-1.11969733e+00 7.58250952e-01 2.53196180e-01 -1.17946279e+00
-2.39268422e-01 -4.27776039e-01 -4.21274394e-01 9.81182814e-01
-9.58917916e-01 -4.00583893e-01 -4.72389609e-01 2.23543286e-01
4.77393627e-01 7.00328946e-02 9.24981713e-01 3.00419569e-01
-3.56555521e-01 5.25923729e-01 1.22422747e-01 3.46562862e-01
4.32619691e-01 -1.44409430e+00 -1.53658926e-01 5.31625271e-01
-2.12666132e-02 4.72525239e-01 3.44357401e-01 -5.93129277e-01
-6.54848516e-01 -9.83952284e-01 7.32733369e-01 -4.99340892e-02
2.73162931e-01 3.15003872e-01 -9.01534498e-01 5.71682930e-01
-2.43457388e-02 2.68040925e-01 8.05848122e-01 -2.94055462e-01
3.11836004e-01 2.13817954e-01 -1.86017656e+00 2.59324282e-01
6.82848811e-01 5.17041683e-02 -4.47607309e-01 3.72123301e-01
4.64940906e-01 -8.20695817e-01 -1.40519500e+00 7.83619165e-01
5.18238604e-01 -1.18106365e+00 9.31792736e-01 -1.07318215e-01
2.47523353e-01 -6.71722293e-02 -2.44458020e-02 -1.18187153e+00
-2.01282576e-01 -9.71523523e-02 2.12949097e-01 7.75793910e-01
8.85841072e-01 -6.95133746e-01 9.98255849e-01 8.11223865e-01
-5.49017251e-01 -1.32690084e+00 -1.05758798e+00 -5.83591521e-01
2.62597412e-01 -3.07193339e-01 3.73836994e-01 1.16366124e+00
-2.43237436e-01 -2.85712481e-01 4.20883417e-01 -3.14745784e-01
9.06493366e-01 -2.28349328e-01 -5.76387830e-02 -1.27013469e+00
-9.58142877e-02 -5.48889816e-01 -2.41958052e-01 -3.42034787e-01
-2.08054557e-01 -1.06874394e+00 1.53342728e-02 -2.03258753e+00
3.89972836e-01 -8.43393385e-01 -5.57303965e-01 3.85616571e-01
-1.04699560e-01 3.67333263e-01 1.37986749e-01 3.94449264e-01
-2.75514692e-01 6.04526848e-02 1.45594692e+00 -7.84411207e-02
-3.61918092e-01 1.10383406e-01 -4.34924364e-01 1.04400766e+00
1.10239935e+00 -3.30915034e-01 -4.19078559e-01 -2.35989511e-01
-1.96564272e-01 4.25549477e-01 1.53644994e-01 -9.92135048e-01
8.23161453e-02 1.13129318e-01 9.19052601e-01 -6.46465540e-01
1.22691259e-01 -8.77407908e-01 2.91799426e-01 8.96916866e-01
-4.21043009e-01 -3.04912031e-01 8.01661462e-02 1.00827545e-01
-3.20269614e-02 -5.55659533e-01 1.04704189e+00 -6.00627363e-01
-5.26445270e-01 3.32955956e-01 -1.48112953e-01 1.72043085e-01
1.26338470e+00 -4.76625651e-01 3.97023298e-02 -2.65182629e-02
-1.10198462e+00 -4.84028682e-02 3.32425356e-01 -1.85262725e-01
5.84857225e-01 -1.01527500e+00 -6.42124772e-01 1.18121039e-02
-1.43147126e-01 5.63846827e-01 2.74813682e-01 1.32690179e+00
-9.39651310e-01 3.92277956e-01 9.16169025e-03 -9.61746931e-01
-1.06790578e+00 1.03010498e-01 5.53676009e-01 -4.29187924e-01
-6.80962861e-01 9.76573110e-01 7.41140591e-03 -3.55367154e-01
2.18613490e-01 -3.84060770e-01 -1.60525903e-01 1.59718528e-01
9.02699232e-02 5.78438759e-01 4.17752475e-01 -4.47963476e-01
-4.88110989e-01 3.96706790e-01 2.17041776e-01 -7.38333911e-02
1.08835173e+00 4.15815078e-02 -6.76503479e-02 4.06149954e-01
1.19612360e+00 -1.80921808e-01 -1.19147587e+00 -6.31235391e-02
2.11835593e-01 -3.32233429e-01 2.99263269e-01 -1.25974786e+00
-1.34797001e+00 4.95854855e-01 8.64852905e-01 3.73999447e-01
1.05985463e+00 8.00960809e-02 5.78876674e-01 -2.23062947e-01
4.93712336e-01 -1.05318558e+00 -2.97536314e-01 2.49214947e-01
6.83536947e-01 -1.22156608e+00 1.80138379e-01 -4.98947144e-01
-7.85819471e-01 1.01358068e+00 4.04454887e-01 -3.93092744e-02
6.44028902e-01 2.44473353e-01 1.27217591e-01 -1.18165970e-01
-2.19607174e-01 1.59981310e-01 3.99361908e-01 3.64040524e-01
6.13540888e-01 3.52867514e-01 -6.68862700e-01 4.69276845e-01
-2.21442170e-02 -7.38467500e-02 2.99424320e-01 6.38462126e-01
-4.44777310e-01 -4.45951372e-01 -3.90017539e-01 7.60430396e-01
-6.70477271e-01 1.33959139e-02 -1.28508452e-02 9.93355870e-01
2.29168147e-01 6.99326336e-01 2.77335942e-01 -3.58263776e-02
9.13754106e-02 1.61626577e-01 5.00271142e-01 -5.26808977e-01
-6.84853375e-01 2.02149764e-01 1.17686046e-02 -4.87201393e-01
-4.98130292e-01 -6.44029140e-01 -1.59282947e+00 5.94766960e-02
-4.28321749e-01 3.06441635e-01 8.90762866e-01 6.34479463e-01
-3.35631967e-02 5.83942235e-01 4.03819680e-01 -4.82618064e-01
-5.08931279e-01 -9.49725091e-01 -6.34965599e-01 3.43565375e-01
-7.88504630e-02 -6.26375496e-01 -6.02735877e-01 -7.75539503e-02] | [14.426414489746094, -2.4769206047058105] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.