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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
8f1efdd5-ce49-434a-9ea6-0c334ada9958 | query-based-keyphrase-extraction-from-long | 2205.05391 | null | https://arxiv.org/abs/2205.05391v1 | https://arxiv.org/pdf/2205.05391v1.pdf | Query-Based Keyphrase Extraction from Long Documents | Transformer-based architectures in natural language processing force input size limits that can be problematic when long documents need to be processed. This paper overcomes this issue for keyphrase extraction by chunking the long documents while keeping a global context as a query defining the topic for which relevant keyphrases should be extracted. The developed system employs a pre-trained BERT model and adapts it to estimate the probability that a given text span forms a keyphrase. We experimented using various context sizes on two popular datasets, Inspec and SemEval, and a large novel dataset. The presented results show that a shorter context with a query overcomes a longer one without the query on long documents. | ['Pavel Smrz', 'Martin Docekal'] | 2022-05-11 | null | null | null | null | ['keyphrase-extraction'] | ['natural-language-processing'] | [ 1.72689572e-01 7.89359678e-03 -2.86095262e-01 -2.23785132e-01
-7.98911989e-01 -7.05531657e-01 1.12824488e+00 9.51282084e-01
-1.22462440e+00 9.18480754e-01 6.35082603e-01 -3.72544885e-01
-2.47137681e-01 -9.42851722e-01 -6.41305268e-01 -3.62165004e-01
-5.65873226e-04 5.85840285e-01 6.16485775e-01 -1.76444620e-01
7.88740516e-01 3.15551043e-01 -1.41860390e+00 8.26667130e-01
5.17108679e-01 7.54012764e-01 6.05717719e-01 8.64469528e-01
-8.14491689e-01 8.82863104e-01 -1.08067560e+00 -1.39061168e-01
2.63313562e-01 -6.44231364e-02 -1.20130777e+00 -2.15121970e-01
6.64459765e-01 -2.82162666e-01 -6.13008738e-02 6.86304569e-01
1.48561954e-01 1.24166220e-01 5.33933401e-01 -9.74140584e-01
-1.26482949e-01 1.35727859e+00 -2.43923530e-01 7.71595716e-01
3.50378186e-01 -4.56020504e-01 1.24198294e+00 -1.01466894e+00
7.32802629e-01 1.00741088e+00 3.68722647e-01 1.76124908e-02
-9.54659581e-01 -5.21009147e-01 3.95463370e-02 3.44165206e-01
-1.29452431e+00 -3.57610583e-01 6.76521063e-01 -2.47616187e-01
1.57409966e+00 2.99056470e-01 5.94720244e-01 6.73435807e-01
6.35969698e-01 8.14253151e-01 9.27210569e-01 -7.59485781e-01
1.24541126e-01 3.13664407e-01 7.46818304e-01 2.49405488e-01
1.80922776e-01 -1.00054689e-01 -6.59813166e-01 -1.79508165e-01
8.44400227e-02 -2.17232749e-01 1.64809436e-01 1.95885196e-01
-1.37326407e+00 9.04507220e-01 -3.12847830e-02 9.23012197e-01
-5.85980713e-01 -1.49509847e-01 8.25209856e-01 5.45225859e-01
3.34948480e-01 6.39218867e-01 -6.81490481e-01 -8.17173719e-02
-1.60494590e+00 7.23515213e-01 1.15300226e+00 9.85927820e-01
7.27135003e-01 -3.84601116e-01 -3.49052489e-01 5.21282256e-01
-1.34607613e-01 2.77288318e-01 5.84394753e-01 -2.57283002e-01
7.39179730e-01 5.99418402e-01 2.81329036e-01 -1.03977466e+00
-4.39304888e-01 -1.96475878e-01 -5.16880095e-01 -2.25623101e-01
2.97194362e-01 -1.06627174e-01 -9.92952764e-01 1.28406477e+00
2.39256069e-01 -3.83559585e-01 6.47458509e-02 2.38241807e-01
8.40331733e-01 1.05697429e+00 2.94118702e-01 -4.32585150e-01
1.74607909e+00 -6.94399536e-01 -8.82076383e-01 -2.70179212e-01
4.04367864e-01 -1.25427806e+00 9.08818543e-01 5.45800090e-01
-8.68053675e-01 -5.36740601e-01 -1.11130047e+00 -3.83595854e-01
-8.94271612e-01 3.54060531e-01 2.53423542e-01 3.28852504e-01
-9.34309840e-01 3.10873240e-01 -3.49566102e-01 -5.46424210e-01
-1.23738848e-01 1.28272191e-01 -2.71194160e-01 2.65416503e-01
-1.49544549e+00 1.07568932e+00 1.12889886e+00 -2.25460559e-01
-6.66854084e-01 -8.09851408e-01 -5.65885961e-01 3.06393206e-01
7.70350218e-01 -4.60616112e-01 1.30937910e+00 -4.61238742e-01
-9.89282727e-01 9.17368591e-01 -3.36867839e-01 -1.00322974e+00
1.17878094e-01 -7.24066556e-01 -4.77531761e-01 3.07570875e-01
3.02218139e-01 3.92406493e-01 1.21019793e+00 -6.62395239e-01
-1.17679608e+00 -2.39290878e-01 1.28622949e-01 1.23197235e-01
-4.00254279e-01 4.38065410e-01 -2.37431049e-01 -8.20586920e-01
-1.49877369e-01 -3.90366018e-01 1.79278273e-02 -8.62762749e-01
-5.60039818e-01 -4.84081388e-01 1.00227189e+00 -6.12212837e-01
1.79760361e+00 -1.68635905e+00 -1.14307880e-01 9.24006924e-02
9.77178812e-02 2.06096202e-01 -4.55468111e-02 1.11227334e+00
1.26763284e-01 6.64442480e-02 5.73348477e-02 8.64768550e-02
-8.70858058e-02 5.82637787e-02 -1.02726030e+00 -3.69012207e-02
-1.04353912e-01 6.34545207e-01 -6.29908264e-01 -1.08893812e+00
7.85936415e-02 4.60650176e-02 -1.38634667e-01 3.86583656e-01
-5.64984739e-01 -2.93077916e-01 -4.35661644e-01 1.59325331e-01
4.58347857e-01 1.20050907e-01 1.14681311e-02 -4.27559644e-01
-3.64190608e-01 8.01479459e-01 -1.16785240e+00 1.30490637e+00
-7.10527956e-01 6.81939125e-01 -1.27762556e-02 -9.36141014e-01
8.26785088e-01 5.03709495e-01 4.94134545e-01 -4.65522498e-01
8.37280080e-02 2.21124031e-02 -3.97823244e-01 -4.69640195e-01
1.34830713e+00 -1.65242150e-01 -2.19501942e-01 7.69684672e-01
2.43474215e-01 -2.82638073e-01 8.26183438e-01 6.26588941e-01
1.01333797e+00 -1.64248124e-01 6.13691986e-01 -6.29901648e-01
8.37753594e-01 1.00785010e-01 8.57978314e-02 8.60146821e-01
2.00924903e-01 7.75910318e-02 4.72699791e-01 -7.03582346e-01
-1.18576097e+00 -7.76532412e-01 1.57272983e-02 1.49869394e+00
-2.28578866e-01 -1.05656064e+00 -6.02031410e-01 -8.96891296e-01
-1.53174192e-01 8.82272005e-01 -5.87068975e-01 1.89036146e-01
-8.11075926e-01 -4.43899572e-01 6.04421198e-01 1.73509985e-01
2.22048670e-01 -1.28272009e+00 -9.63197052e-01 5.20948052e-01
-2.75751829e-01 -1.25773811e+00 -5.64813197e-01 4.31180269e-01
-7.06585646e-01 -9.70176935e-01 -6.14161968e-01 -8.10242653e-01
1.73363745e-01 1.12823524e-01 1.37571204e+00 -1.09802194e-01
-6.11197427e-02 1.78889662e-01 -6.03545547e-01 -6.60859406e-01
-5.90987325e-01 7.24684775e-01 -1.48108065e-01 -3.92541200e-01
9.23974335e-01 -4.73231256e-01 -2.08367407e-01 -1.13071546e-01
-1.13670969e+00 2.88874246e-02 7.18567014e-01 6.29925489e-01
3.14409971e-01 3.48320693e-01 5.14010012e-01 -9.05687094e-01
1.04604399e+00 -4.66394097e-01 -6.12724960e-01 4.19826478e-01
-9.13156211e-01 4.47206646e-01 8.59048903e-01 -3.19265902e-01
-9.46978867e-01 -1.76792383e-01 1.46428682e-02 1.93086207e-01
-3.08305442e-01 6.74761772e-01 9.58691388e-02 6.34280324e-01
6.70177698e-01 3.90303552e-01 -6.69336021e-01 -6.07123375e-01
4.12663430e-01 6.67163968e-01 3.25916737e-01 -6.14133596e-01
7.91377962e-01 1.31819800e-01 -2.17289522e-01 -1.19387627e+00
-9.08477664e-01 -9.21427190e-01 -7.65724063e-01 9.69672650e-02
5.38913488e-01 -7.18930602e-01 -1.48548201e-01 2.28728801e-01
-1.25954521e+00 1.38848096e-01 -4.90236163e-01 3.42631638e-01
-1.89374685e-01 4.97996867e-01 -5.06060719e-01 -5.86220443e-01
-1.05098784e+00 -5.15268445e-01 9.33714747e-01 8.76030475e-02
-5.73247790e-01 -8.33894968e-01 2.46346891e-01 -1.91857502e-01
3.38048100e-01 -1.54200971e-01 9.43284035e-01 -1.17523658e+00
-3.95481110e-01 -3.38739157e-01 -9.14154574e-02 3.99620924e-03
2.81397074e-01 -1.18210949e-02 -7.93608725e-01 -1.85569450e-01
1.72748059e-01 -2.68317163e-01 9.95957792e-01 1.35058448e-01
6.44788861e-01 -7.02713490e-01 -3.53009492e-01 1.96731225e-01
1.41273880e+00 1.53479367e-01 3.44366103e-01 5.94504714e-01
4.58258569e-01 5.98573864e-01 9.39640284e-01 3.95025760e-01
2.58916527e-01 3.15032363e-01 -1.22597113e-01 4.28556919e-01
1.60101622e-01 -4.65057373e-01 2.82608151e-01 8.71806860e-01
3.75901520e-01 -3.35486472e-01 -8.11931193e-01 9.88510549e-01
-1.43149221e+00 -1.06231177e+00 1.56381100e-01 1.88732100e+00
1.25284982e+00 5.64142942e-01 6.88000768e-02 3.92880410e-01
3.30712765e-01 6.31565809e-01 1.21905603e-01 -5.48839986e-01
-7.45950965e-03 4.53057081e-01 4.84760165e-01 8.57319534e-01
-1.25919282e+00 1.34251642e+00 6.32895422e+00 8.22693527e-01
-1.07344162e+00 -1.38278991e-01 7.52961338e-02 6.80218413e-02
-2.82057643e-01 2.21887499e-01 -1.42938018e+00 1.63824961e-01
1.35787606e+00 -4.95309711e-01 -8.81951209e-03 7.85457373e-01
1.40403509e-01 -6.97225749e-01 -1.35123062e+00 6.56837881e-01
1.53047159e-01 -1.39138925e+00 4.00073946e-01 -1.52754858e-01
1.14155225e-01 -1.31055778e-02 -4.03835177e-01 2.87360638e-01
7.03113303e-02 -7.73933113e-01 7.62944818e-01 3.98573220e-01
4.08954263e-01 -6.60162508e-01 5.83835065e-01 5.29902339e-01
-1.35766745e+00 8.05357918e-02 -5.21417618e-01 3.97665314e-02
1.04134440e-01 7.99654305e-01 -1.39593029e+00 3.83081436e-01
5.15210450e-01 3.10304910e-01 -6.13285184e-01 8.08009684e-01
-2.25982025e-01 6.24244571e-01 -5.68169594e-01 -3.60121131e-01
5.17940760e-01 1.12246089e-01 8.53244781e-01 1.93687820e+00
1.17786013e-01 -1.10393792e-01 1.05724387e-01 6.84193492e-01
6.88746274e-02 6.03690684e-01 -5.61522722e-01 -2.99528241e-01
5.93746185e-01 1.37208128e+00 -1.06711006e+00 -7.85831571e-01
-2.02343196e-01 6.59250617e-01 2.17890721e-02 2.11553916e-01
-2.77417600e-01 -8.29701722e-01 7.58653507e-02 1.41319320e-01
5.04381061e-01 -2.59592593e-01 -8.19827020e-02 -1.11448538e+00
9.75730047e-02 -8.86020124e-01 7.23960638e-01 -5.04229128e-01
-8.13721955e-01 7.65676558e-01 5.23954391e-01 -8.71580541e-01
-7.25187719e-01 -3.53828162e-01 -4.92163330e-01 9.69795763e-01
-1.29404283e+00 -1.23730719e+00 1.31610334e-01 5.26294827e-01
9.88296211e-01 -2.25471288e-01 8.31172764e-01 2.14686021e-02
9.04053152e-02 3.30171257e-01 -2.39933908e-01 3.20071965e-01
7.06429720e-01 -1.43726325e+00 4.72564757e-01 1.07219839e+00
6.05173409e-01 8.95398021e-01 1.10264981e+00 -7.00102925e-01
-1.14305782e+00 -6.75915420e-01 1.66503417e+00 -3.88981640e-01
7.64070392e-01 -4.07328486e-01 -9.37351942e-01 6.55635715e-01
7.36815989e-01 -5.09540558e-01 5.22239506e-01 1.36555538e-01
-5.11806905e-01 -2.79813975e-01 -8.96860123e-01 2.97109872e-01
1.96227893e-01 -9.03930902e-01 -1.57992589e+00 1.28340751e-01
1.02362251e+00 -4.01745513e-02 -8.06932211e-01 -1.72140375e-02
5.57174742e-01 -7.53340125e-01 7.62847364e-01 -5.71372867e-01
2.30610073e-01 -1.67865783e-01 -1.07402764e-02 -1.09214139e+00
4.51619104e-02 -6.70434713e-01 -4.16560054e-01 1.25377870e+00
4.54637796e-01 -2.59712577e-01 7.01861620e-01 3.72145534e-01
2.69107550e-01 -5.58828950e-01 -7.86561430e-01 -5.07998228e-01
2.48980045e-01 -4.24581856e-01 6.36614084e-01 7.22822189e-01
2.55100250e-01 9.41337347e-01 -1.21533073e-01 -1.91932783e-01
4.26741809e-01 3.04177910e-01 7.26868272e-01 -1.21562374e+00
8.39521959e-02 -2.39906177e-01 8.99659470e-02 -8.83564115e-01
-2.10596453e-02 -5.32084465e-01 -1.10029891e-01 -1.33670747e+00
1.31844968e-01 -1.59555510e-01 -2.63855785e-01 3.46549839e-01
-6.43707216e-02 -4.87495303e-01 1.47086516e-01 8.01450610e-02
-3.70445848e-01 1.63297787e-01 6.72014773e-01 -2.03928694e-01
-3.83915514e-01 1.63664475e-01 -2.97465771e-01 5.59894204e-01
8.11687231e-01 -7.09609270e-01 -5.62303483e-01 -1.03648692e-01
5.73916495e-01 -8.22413936e-02 -1.18524313e-01 -9.30279493e-01
5.18266380e-01 -2.20734522e-01 1.57387078e-01 -1.30605459e+00
-7.87123293e-02 -9.62518573e-01 -3.42320770e-01 3.27824801e-01
-6.92108274e-01 4.08153057e-01 4.80442256e-01 2.30214491e-01
-3.57084662e-01 -5.89391828e-01 5.34353673e-01 -3.34663808e-01
-7.04918742e-01 1.19110413e-01 -7.45916486e-01 2.58999527e-01
6.93836451e-01 -2.19072402e-02 6.12294525e-02 -2.62219816e-01
-5.27990103e-01 -7.27098843e-04 1.59255806e-02 5.16348660e-01
7.71639824e-01 -7.95539916e-01 -8.11820567e-01 1.16778664e-01
1.70085758e-01 -1.98180482e-01 -1.66682631e-01 4.01983738e-01
-4.86663491e-01 9.69032705e-01 -1.36254683e-01 -1.19499296e-01
-1.53397238e+00 8.73066783e-01 -3.71969566e-02 -8.72214973e-01
-7.86332250e-01 5.88081002e-01 -1.57866925e-01 -1.38076335e-01
3.35802138e-01 -9.95909691e-01 -6.68673277e-01 7.19137549e-01
1.00281739e+00 8.32061172e-02 3.16383570e-01 -5.57692409e-01
3.07428613e-02 3.37710172e-01 -6.86497688e-01 -4.49719787e-01
1.14158058e+00 -3.63968343e-01 -4.22856003e-01 6.94981694e-01
1.01583326e+00 3.57326597e-01 -5.14178932e-01 -7.65823066e-01
7.03863680e-01 -1.43365309e-01 1.45001084e-01 -6.72443569e-01
-4.00946200e-01 7.10162759e-01 2.29199171e-01 5.81539631e-01
1.04589880e+00 2.08299272e-02 1.12387824e+00 8.96953344e-01
2.88713783e-01 -1.47227323e+00 -2.11743057e-01 9.55000579e-01
1.02393937e+00 -8.49733412e-01 4.59934920e-01 2.56725680e-02
-2.35162795e-01 1.36204207e+00 3.36154044e-01 -7.36057386e-02
9.14246321e-01 3.55694890e-01 -3.13532911e-02 -3.56438786e-01
-8.83930206e-01 -9.93560031e-02 4.92835045e-01 2.30223596e-01
4.91262794e-01 -8.04160610e-02 -7.40602136e-01 5.17382503e-01
-7.14172423e-01 -3.83967161e-03 5.98015964e-01 1.11310387e+00
-8.10277104e-01 -1.14324296e+00 -5.35027504e-01 5.18921256e-01
-9.60277319e-01 -4.90831584e-01 -5.82230449e-01 8.36364865e-01
6.23259135e-02 7.75916636e-01 8.99951607e-02 -2.93774545e-01
-4.32616146e-03 3.88668060e-01 2.60403723e-01 -8.09869111e-01
-1.01944971e+00 9.85765606e-02 2.92592257e-01 -3.59147191e-01
-4.50146049e-01 -3.85944515e-01 -1.05113387e+00 -1.46368938e-02
-2.62754440e-01 6.81916475e-01 6.07422173e-01 1.11366475e+00
6.11336827e-02 2.06652671e-01 2.85862446e-01 -3.78363907e-01
-4.91463542e-01 -1.50315428e+00 -3.49007428e-01 1.29228368e-01
5.71038365e-01 -1.21275239e-01 -2.45191991e-01 3.78169179e-01] | [12.25340461730957, 8.866875648498535] |
45eb6c21-7b38-434a-9e28-11e149a122a7 | clipface-text-guided-editing-of-textured-3d | 2212.01406 | null | https://arxiv.org/abs/2212.01406v2 | https://arxiv.org/pdf/2212.01406v2.pdf | ClipFace: Text-guided Editing of Textured 3D Morphable Models | We propose ClipFace, a novel self-supervised approach for text-guided editing of textured 3D morphable model of faces. Specifically, we employ user-friendly language prompts to enable control of the expressions as well as appearance of 3D faces. We leverage the geometric expressiveness of 3D morphable models, which inherently possess limited controllability and texture expressivity, and develop a self-supervised generative model to jointly synthesize expressive, textured, and articulated faces in 3D. We enable high-quality texture generation for 3D faces by adversarial self-supervised training, guided by differentiable rendering against collections of real RGB images. Controllable editing and manipulation are given by language prompts to adapt texture and expression of the 3D morphable model. To this end, we propose a neural network that predicts both texture and expression latent codes of the morphable model. Our model is trained in a self-supervised fashion by exploiting differentiable rendering and losses based on a pre-trained CLIP model. Once trained, our model jointly predicts face textures in UV-space, along with expression parameters to capture both geometry and texture changes in facial expressions in a single forward pass. We further show the applicability of our method to generate temporally changing textures for a given animation sequence. | ['Matthias Nießner', 'Angela Dai', 'Justus Thies', 'Shivangi Aneja'] | 2022-12-02 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 4.72884864e-01 5.16699970e-01 2.71808326e-01 -5.68453789e-01
-5.96645474e-01 -8.05934429e-01 6.53554380e-01 -8.23806107e-01
3.10723156e-01 3.25041592e-01 3.97933573e-02 1.32799178e-01
3.30683947e-01 -9.34033930e-01 -1.08041990e+00 -5.36191821e-01
-2.05902588e-02 4.80025053e-01 -5.01897037e-01 -3.89899731e-01
-2.66406417e-01 1.03904295e+00 -1.70700336e+00 3.38549912e-01
6.69758737e-01 1.13042045e+00 -3.02632838e-01 9.71782386e-01
2.80748438e-02 6.54782176e-01 -2.08768576e-01 -4.35948163e-01
4.76401776e-01 -4.94052052e-01 -3.86675477e-01 6.11894190e-01
7.18601942e-01 -5.24025738e-01 -3.42075437e-01 6.67343974e-01
3.64259064e-01 -1.93549264e-02 9.50418651e-01 -1.17092311e+00
-9.22877669e-01 -1.97868973e-01 -3.94497335e-01 -7.35478401e-01
5.63905954e-01 4.89753753e-01 6.03019774e-01 -9.03853416e-01
1.01843441e+00 1.56526601e+00 4.94881511e-01 1.06789148e+00
-1.61608601e+00 -7.28040874e-01 5.01872860e-02 -6.52988374e-01
-1.20787430e+00 -8.10029387e-01 1.22044945e+00 -4.82695013e-01
4.86844271e-01 5.32115400e-01 1.02996314e+00 1.34408426e+00
1.09468251e-01 3.74566048e-01 1.30332148e+00 -3.72330725e-01
2.70364851e-01 -1.75049696e-02 -8.72834444e-01 1.29796696e+00
-7.51719236e-01 2.49486163e-01 -6.39782846e-01 -1.92107528e-01
1.43484890e+00 -2.36242115e-01 -1.01548042e-02 -5.88177800e-01
-8.45115244e-01 6.13297701e-01 1.10951096e-01 -4.90925699e-01
-2.02833951e-01 3.70065957e-01 2.99940631e-02 5.67637801e-01
1.03555429e+00 3.50922406e-01 -4.67710733e-01 -1.21495895e-01
-5.73359311e-01 4.77569699e-01 6.98211670e-01 1.18862855e+00
9.49433923e-01 4.29101437e-01 -1.86856940e-01 7.53863990e-01
2.27257520e-01 1.08608711e+00 -1.09922744e-01 -1.40666556e+00
-4.33517881e-02 4.90067363e-01 -1.06814746e-02 -1.04328418e+00
9.55907777e-02 2.05523804e-01 -8.72865081e-01 8.79145741e-01
6.54033944e-02 -1.22330077e-01 -1.07827044e+00 2.10790491e+00
7.57000148e-01 -1.33543342e-01 -2.68195361e-01 8.06222677e-01
5.12666404e-01 5.76633811e-01 -1.63106143e-01 -6.24739081e-02
8.09590280e-01 -5.17107666e-01 -5.58655679e-01 1.85487866e-01
1.29609793e-01 -7.12138295e-01 1.48479509e+00 1.40374689e-03
-1.46730769e+00 -3.76841486e-01 -6.79177761e-01 -2.28151575e-01
-4.25842730e-03 1.76118333e-02 5.08413076e-01 4.71466005e-01
-1.21850204e+00 6.97336197e-01 -1.03182042e+00 -8.06967840e-02
6.35358751e-01 4.93856132e-01 -5.82703590e-01 3.08268398e-01
-7.67524064e-01 5.11588454e-01 -6.40399516e-01 1.24548033e-01
-1.14289618e+00 -8.41709971e-01 -1.05003881e+00 -3.12290519e-01
1.86660178e-02 -9.47246134e-01 1.07418072e+00 -1.42651772e+00
-2.45179653e+00 1.39143503e+00 -1.58466488e-01 2.79918492e-01
7.93962300e-01 4.11150455e-02 -8.97752345e-02 2.01923788e-01
-1.40564263e-01 8.10991943e-01 1.48105073e+00 -1.46385550e+00
1.16328299e-01 -2.98222572e-01 2.07939789e-01 2.17436478e-01
-2.81625420e-01 -5.10782041e-02 -5.00872970e-01 -8.81970942e-01
-1.13825701e-01 -1.06309140e+00 -9.90314707e-02 1.00948906e+00
-5.49045682e-01 4.42871988e-01 1.01703489e+00 -4.98236746e-01
4.78921086e-01 -2.04480195e+00 5.34989774e-01 4.37899768e-01
1.68159217e-01 -2.42193371e-01 -4.57784265e-01 8.35606605e-02
-2.42735781e-02 1.70774400e-01 -2.97203124e-01 -8.10616910e-01
2.19839826e-01 3.32335144e-01 -5.40998280e-01 3.40109885e-01
6.89661145e-01 1.02051580e+00 -6.43889666e-01 -4.44629312e-01
2.63025314e-01 9.13549900e-01 -9.65556145e-01 7.41663396e-01
-8.02516818e-01 1.18214345e+00 -7.06442595e-01 8.87264609e-01
6.99874461e-01 1.45874202e-01 5.00123724e-02 -2.11594492e-01
-4.53451276e-02 -3.25442441e-02 -6.36578083e-01 1.81595409e+00
-9.19357181e-01 3.94209474e-01 4.82312858e-01 -4.13394362e-01
1.11749601e+00 2.28391543e-01 4.49098587e-01 -5.62650263e-01
2.12145999e-01 3.13293338e-02 -6.65777802e-01 -5.30171454e-01
2.25184843e-01 -3.21914405e-01 -7.87176266e-02 6.75571918e-01
-8.21798071e-02 -1.02068090e+00 -6.82738841e-01 -8.73886272e-02
9.39304829e-01 7.12680161e-01 -4.58996356e-01 -1.00433111e-01
1.88978493e-01 -4.91109103e-01 1.85348570e-01 1.63217336e-01
3.65373909e-01 7.53609061e-01 6.22495472e-01 -5.33922255e-01
-1.45435333e+00 -1.26574218e+00 5.58941439e-02 1.00166416e+00
-2.03600243e-01 -4.85822260e-02 -9.91322219e-01 -4.89114165e-01
1.30657956e-01 2.12500080e-01 -1.13269615e+00 -1.98159039e-01
-7.12127090e-01 2.85485201e-03 4.66509879e-01 1.47096708e-01
3.34375679e-01 -9.67206299e-01 -3.86384666e-01 -1.17934257e-01
3.02583635e-01 -1.00394666e+00 -8.05261314e-01 -2.13056430e-01
-6.59270287e-01 -8.77421558e-01 -6.84191406e-01 -7.47504294e-01
1.15061140e+00 -4.25198734e-01 9.86520231e-01 2.41526160e-02
-4.19805884e-01 8.08366477e-01 6.50741979e-02 -7.56043047e-02
-8.43313098e-01 -1.86305836e-01 1.19671375e-01 6.06807292e-01
-7.20459580e-01 -1.09543240e+00 -5.42150736e-01 3.51528913e-01
-9.76702750e-01 4.97162938e-01 -5.27347326e-02 6.62449121e-01
9.27905142e-01 -5.52989662e-01 9.75790694e-02 -9.21252608e-01
3.37238252e-01 5.70899472e-02 -7.00152457e-01 1.05444856e-01
-3.93237025e-02 1.34314239e-01 7.44493067e-01 -7.78341234e-01
-1.19097412e+00 3.76770556e-01 -2.47518048e-01 -1.02344978e+00
-2.07279213e-02 -2.17514992e-01 -5.84273636e-01 -5.82975388e-01
5.95057845e-01 1.21404408e-02 3.07151169e-01 -2.39537030e-01
7.80940056e-01 3.77221376e-01 5.85285485e-01 -1.13582146e+00
1.30322552e+00 7.81188071e-01 3.17528009e-01 -6.89426959e-01
-5.57289302e-01 6.56206429e-01 -6.38425708e-01 -4.87378061e-01
7.41684973e-01 -8.24558377e-01 -9.14576232e-01 8.37859750e-01
-1.00583088e+00 -1.22441959e+00 -5.85963786e-01 -2.34555304e-01
-1.20812714e+00 -2.00788736e-01 -6.45530403e-01 -7.49575496e-01
-3.82329136e-01 -9.57059145e-01 1.68834913e+00 -1.91299468e-01
-2.64790744e-01 -9.63288307e-01 -7.56300762e-02 1.07501820e-01
6.03710711e-01 1.13407624e+00 9.23005581e-01 4.48521346e-01
-7.61138260e-01 7.54011655e-03 1.68706939e-01 2.48545036e-01
5.38033962e-01 3.25682074e-01 -1.09929383e+00 -1.33820921e-01
-2.01385498e-01 -7.80778170e-01 2.28742912e-01 3.87008227e-02
1.44115174e+00 -7.52789795e-01 -1.79580897e-02 1.32004535e+00
9.43791151e-01 -1.73399031e-01 3.84074777e-01 -2.60425448e-01
1.02085328e+00 6.61611676e-01 2.74717093e-01 6.05317235e-01
3.06929946e-01 7.92136848e-01 4.49818730e-01 -3.28012854e-01
-1.89289197e-01 -7.71462023e-01 3.65763694e-01 5.83179295e-01
-2.60122865e-01 7.35124052e-02 -4.65513408e-01 -8.44662264e-03
-1.20432746e+00 -6.15474463e-01 5.25614560e-01 1.97459900e+00
1.22310936e+00 -2.09449977e-01 -7.03618079e-02 -3.64038527e-01
3.89063716e-01 1.80457830e-01 -8.61222565e-01 -5.72296917e-01
-1.72726527e-01 7.80997574e-01 3.59803252e-02 6.97037935e-01
-8.93918335e-01 1.25128376e+00 6.16256094e+00 6.53899670e-01
-1.53496099e+00 -8.19433182e-02 9.26265538e-01 -4.25922960e-01
-1.01311111e+00 -1.60823569e-01 -1.85521260e-01 1.92812726e-01
5.20638645e-01 -6.69804960e-02 8.87947142e-01 5.90340436e-01
4.59562749e-01 5.67585886e-01 -1.18130839e+00 8.61593902e-01
2.92791259e-02 -1.39722800e+00 3.82829368e-01 7.69814774e-02
9.85114872e-01 -6.31165385e-01 6.03816867e-01 -8.82718414e-02
4.10107315e-01 -1.20153546e+00 1.00841999e+00 7.92271137e-01
1.80700183e+00 -6.71087027e-01 -4.51240927e-01 6.66177347e-02
-9.13173139e-01 4.03921813e-01 2.98804865e-04 7.05208257e-02
1.87778935e-01 2.85368562e-01 -2.70695239e-01 -3.82386195e-03
4.20556277e-01 5.91198623e-01 -2.59610936e-02 -8.47772658e-02
-3.27963322e-01 4.57292974e-01 -3.86574656e-01 3.16673219e-01
-2.32196882e-01 -3.53129536e-01 6.58835471e-01 8.28263640e-01
2.53927320e-01 4.15556341e-01 2.07468662e-02 1.43101680e+00
-4.15006965e-01 2.09152978e-02 -8.28731000e-01 -1.41268834e-01
2.45304972e-01 1.13749361e+00 -1.90689936e-01 -2.81543285e-02
1.38309255e-01 1.42283583e+00 5.18721640e-01 4.67308670e-01
-8.04474950e-01 -5.58621809e-02 1.03862596e+00 3.64235550e-01
7.22892135e-02 -4.26466405e-01 -2.54935741e-01 -1.22739995e+00
8.27284753e-02 -8.55000436e-01 -4.96306717e-01 -1.10314941e+00
-1.19808114e+00 6.87660813e-01 -3.51541191e-01 -1.05365658e+00
-2.55479515e-01 -4.44757700e-01 -5.69818795e-01 8.92449141e-01
-1.34080398e+00 -1.83184195e+00 -3.93916547e-01 8.58391345e-01
2.22060278e-01 -1.10581577e-01 1.15888405e+00 -1.02392353e-01
-3.13264340e-01 8.78800511e-01 -3.41954976e-01 6.14126399e-02
6.84046566e-01 -9.75361168e-01 7.20217824e-01 2.18445554e-01
-1.82098642e-01 3.94487083e-01 4.96567726e-01 -5.19047737e-01
-2.00749445e+00 -1.40079367e+00 2.84342021e-01 -5.89631021e-01
3.51915896e-01 -7.88871288e-01 -6.21517420e-01 7.32754886e-01
-2.30856746e-01 3.46159488e-01 4.18573588e-01 -4.17685717e-01
-5.44671595e-01 -6.43883795e-02 -1.40591955e+00 8.95100176e-01
1.42810345e+00 -1.00499952e+00 1.25353992e-01 3.51224035e-01
7.48858690e-01 -1.03286588e+00 -9.61009681e-01 3.32768440e-01
8.88897717e-01 -7.47645497e-01 8.16950142e-01 -6.63318753e-01
7.36477375e-01 -8.15279409e-02 -1.72821432e-01 -1.25839555e+00
5.20832613e-02 -1.35086477e+00 -5.03802523e-02 1.14978707e+00
2.31428653e-01 -4.02196586e-01 9.78929102e-01 9.62045610e-01
-1.20834604e-01 -8.70349586e-01 -8.11341047e-01 -4.78497446e-01
2.33794481e-01 -3.57164353e-01 9.09273148e-01 1.00865340e+00
-3.23158950e-01 -1.91346914e-01 -5.42280495e-01 1.07867643e-01
5.71381092e-01 2.72276670e-01 1.22114086e+00 -6.79129601e-01
-3.92868757e-01 -1.43593743e-01 -2.19041869e-01 -1.10639405e+00
7.23849535e-01 -9.00084138e-01 6.27561845e-03 -7.56928563e-01
-1.06935717e-01 -8.38010192e-01 5.41910112e-01 7.07600772e-01
2.02629641e-01 4.62074727e-01 1.52497277e-01 1.56881511e-02
8.03252775e-03 1.03165460e+00 1.81536126e+00 -1.08269677e-01
-3.71896595e-01 -2.12326288e-01 -4.19231743e-01 7.28102982e-01
4.52928543e-01 -7.39419237e-02 -4.06702101e-01 -7.21438766e-01
1.54233620e-01 3.82380903e-01 6.54544711e-01 -4.83351856e-01
-2.38268033e-01 -4.46903735e-01 5.09856582e-01 2.39580035e-01
7.97966480e-01 -6.56206250e-01 4.13358688e-01 -1.22401476e-01
-6.81481719e-01 -1.29630283e-01 2.48439208e-01 3.34826261e-01
1.61377788e-01 7.44880736e-01 8.74451876e-01 -3.43089178e-02
4.15656939e-02 1.02001190e+00 -5.46786971e-02 6.32921383e-02
7.57363141e-01 -1.35667682e-01 3.11142743e-01 -7.47041523e-01
-9.86528337e-01 -1.12249255e-01 1.15273917e+00 3.61156374e-01
7.15605021e-01 -1.61311901e+00 -5.89595914e-01 9.04553056e-01
9.36043933e-02 2.86722749e-01 2.12139770e-01 1.22177906e-01
-6.25222921e-01 -4.48308855e-01 -3.26534569e-01 -5.15527248e-01
-1.06954551e+00 1.46495715e-01 7.17574060e-01 2.19587505e-01
-5.58791578e-01 8.24414194e-01 4.06267285e-01 -8.85113776e-01
-1.69061590e-02 -2.04617485e-01 4.97375160e-01 -5.80041170e-01
1.79364428e-01 -1.14168689e-01 -2.54892677e-01 -6.47293210e-01
1.11690320e-01 1.01176715e+00 4.26697493e-01 -4.35315132e-01
1.32675850e+00 3.41855958e-02 -2.17637360e-01 3.30477268e-01
1.34909785e+00 3.49248797e-01 -2.03071833e+00 1.07956015e-01
-9.35490072e-01 -5.98238468e-01 -2.43668482e-01 -6.52486682e-01
-1.31975818e+00 8.07340682e-01 2.48152971e-01 -4.57311809e-01
1.19253671e+00 -8.72586817e-02 7.95728803e-01 6.95363283e-02
3.89976472e-01 -7.72538662e-01 3.96701336e-01 3.94882530e-01
1.35378468e+00 -8.19696009e-01 -4.76071358e-01 -6.35916173e-01
-4.50377882e-01 1.17689085e+00 5.55223644e-01 -2.86423445e-01
7.41599143e-01 6.53344154e-01 2.00271323e-01 -1.54398903e-01
-7.64421582e-01 4.18219537e-01 3.00428987e-01 7.84560323e-01
1.78192765e-01 1.58804819e-01 5.63563049e-01 2.23628491e-01
-5.69149375e-01 1.14378445e-01 9.68451276e-02 6.84567988e-01
9.57866907e-02 -1.12007201e+00 -2.63138562e-01 9.16099697e-02
-8.75710025e-02 1.11722119e-01 -5.62604368e-01 4.81194437e-01
1.03908420e-01 4.60994333e-01 2.05416396e-01 -4.64504600e-01
4.82475579e-01 -5.41109927e-02 9.61600721e-01 -4.73512530e-01
-2.01158240e-01 -5.52716739e-02 -4.73974869e-02 -8.19079161e-01
-2.12578565e-01 -5.31555057e-01 -9.59956050e-01 -4.89206821e-01
2.83865839e-01 -4.09835160e-01 5.64447403e-01 5.77029705e-01
6.16492987e-01 1.84863016e-01 1.38768566e+00 -1.38098001e+00
-3.73758405e-01 -4.82524097e-01 -5.83391607e-01 6.88481152e-01
5.47454596e-01 -5.78103423e-01 -3.21993411e-01 5.33476949e-01] | [12.686161041259766, -0.3821890354156494] |
7963b40f-4112-49fe-bd21-12552bbed7b0 | trusted-multi-view-classification-with | 2204.11423 | null | https://arxiv.org/abs/2204.11423v3 | https://arxiv.org/pdf/2204.11423v3.pdf | Trusted Multi-View Classification with Dynamic Evidential Fusion | Existing multi-view classification algorithms focus on promoting accuracy by exploiting different views, typically integrating them into common representations for follow-up tasks. Although effective, it is also crucial to ensure the reliability of both the multi-view integration and the final decision, especially for noisy, corrupted and out-of-distribution data. Dynamically assessing the trustworthiness of each view for different samples could provide reliable integration. This can be achieved through uncertainty estimation. With this in mind, we propose a novel multi-view classification algorithm, termed trusted multi-view classification (TMC), providing a new paradigm for multi-view learning by dynamically integrating different views at an evidence level. The proposed TMC can promote classification reliability by considering evidence from each view. Specifically, we introduce the variational Dirichlet to characterize the distribution of the class probabilities, parameterized with evidence from different views and integrated with the Dempster-Shafer theory. The unified learning framework induces accurate uncertainty and accordingly endows the model with both reliability and robustness against possible noise or corruption. Both theoretical and experimental results validate the effectiveness of the proposed model in accuracy, robustness and trustworthiness. | ['Joey Tianyi Zhou', 'Huazhu Fu', 'Changqing Zhang', 'Zongbo Han'] | 2022-04-25 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [-3.45772505e-01 3.18143293e-02 -2.33696461e-01 -5.53686321e-01
-1.28367424e+00 -6.82625234e-01 6.53253555e-01 3.75122607e-01
5.38242273e-02 8.30699027e-01 4.37919572e-02 2.90054381e-01
-5.35634160e-01 -7.68685222e-01 -5.70099413e-01 -1.15389705e+00
2.05217466e-01 4.65837926e-01 -7.30764046e-02 2.33178467e-01
-1.33137312e-02 1.77573100e-01 -1.63237643e+00 2.90183812e-01
9.63409901e-01 1.43826628e+00 -4.15842682e-01 3.74040812e-01
5.49076378e-01 6.49463117e-01 -5.09902775e-01 -1.00762212e+00
1.53231680e-01 -3.73592712e-02 -4.53146100e-01 2.44834736e-01
3.45692486e-02 -4.49735999e-01 5.33437610e-01 1.36505342e+00
2.43452445e-01 -1.62599772e-01 9.15029407e-01 -1.61372817e+00
-4.97588426e-01 7.30054438e-01 -5.19076765e-01 -2.27703258e-01
3.64128977e-01 -3.52606475e-01 1.18073142e+00 -1.03707266e+00
4.81771767e-01 1.13144457e+00 4.91778970e-01 1.69398129e-01
-9.80190098e-01 -4.40662056e-01 3.52909297e-01 5.72629094e-01
-1.33420944e+00 -3.68641376e-01 9.60242450e-01 -4.16677207e-01
-1.86435282e-01 1.47965714e-01 6.00529194e-01 1.38806009e+00
5.59862256e-01 9.01699424e-01 1.61908197e+00 -2.40600824e-01
5.54436445e-01 5.62302232e-01 2.90361285e-01 4.55227971e-01
7.80393481e-01 1.00468978e-01 -6.31952763e-01 -5.78880668e-01
-1.05532045e-02 3.40078212e-02 -3.59501839e-01 -7.67630816e-01
-8.78282845e-01 8.42979252e-01 1.90746531e-01 6.09223135e-02
-4.56750512e-01 -1.81453541e-01 6.20402873e-01 1.46773085e-01
6.88421249e-01 -3.91366094e-01 -1.82714254e-01 3.08617204e-01
-8.16548109e-01 -5.63044585e-02 6.85679972e-01 7.63278723e-01
3.72405052e-01 -8.91592875e-02 1.17637314e-01 4.36957687e-01
7.61644125e-01 6.50571585e-01 3.94600071e-02 -1.04537368e+00
3.30011189e-01 3.98122638e-01 1.47128448e-01 -1.00366509e+00
-1.18131742e-01 -7.00867116e-01 -1.19263434e+00 5.86497486e-01
2.37395063e-01 3.68821733e-02 -1.75099000e-01 1.63654745e+00
7.64994860e-01 -2.85262950e-02 4.17533010e-01 7.36468732e-01
4.65773791e-01 2.25178212e-01 -2.01421246e-01 -5.92438638e-01
1.47257268e+00 -5.00795066e-01 -9.22486722e-01 4.23741817e-01
2.50205576e-01 -4.10770833e-01 3.51616681e-01 9.93245244e-01
-8.41968775e-01 -4.45183426e-01 -1.38167799e+00 4.71474797e-01
-1.55322641e-01 1.13903686e-01 6.47288188e-02 9.13529575e-01
-6.18423879e-01 5.10153472e-01 -7.55616724e-01 1.83685869e-01
4.74890649e-01 -2.28912337e-03 -7.61456490e-01 -3.11078548e-01
-1.24596179e+00 1.04841590e+00 2.94128478e-01 2.83494025e-01
-1.01078498e+00 -4.10040170e-01 -7.43889749e-01 -4.70174998e-02
4.53770638e-01 -8.43008041e-01 7.81894386e-01 -7.34639227e-01
-1.17681360e+00 4.97428060e-01 1.34893373e-01 -4.55801815e-01
9.53932703e-01 -2.66639024e-01 -4.40084994e-01 3.77739787e-01
1.44749448e-01 -1.71580210e-01 1.22565782e+00 -2.00854945e+00
-4.86403972e-01 -9.49928343e-01 -3.35826315e-02 2.15434223e-01
-9.85104218e-02 -2.99450368e-01 -7.40370825e-02 -4.07269686e-01
2.92119265e-01 -7.20628262e-01 1.23488054e-01 9.25935656e-02
-3.26776594e-01 -1.92400470e-01 8.19984734e-01 -6.12970769e-01
8.71759892e-01 -1.89211941e+00 3.67499739e-01 4.23156202e-01
3.50065529e-01 1.62697006e-02 4.48445857e-01 3.44689816e-01
2.47859061e-01 -1.43415872e-02 -1.02296539e-01 -6.32357955e-01
-1.02143310e-01 2.71578580e-01 -1.07154511e-01 1.00139880e+00
-1.91290438e-01 4.23079044e-01 -8.78834188e-01 -6.16754770e-01
2.63642162e-01 6.60571098e-01 -1.43992692e-01 4.50007804e-02
2.33845949e-01 5.25542676e-01 -5.44166505e-01 8.09766710e-01
8.85772109e-01 -2.83194423e-01 4.01727825e-01 -3.99971455e-01
4.81816232e-01 -4.11562175e-01 -1.48829842e+00 1.18215096e+00
-5.03519297e-01 -9.00005549e-02 3.89257491e-01 -1.07977700e+00
7.73222268e-01 6.91479027e-01 4.21716779e-01 -3.40869129e-02
3.46204519e-01 2.92742223e-01 -3.66353363e-01 -2.02347815e-01
2.03615770e-01 -4.39703554e-01 -1.52558550e-01 6.77899897e-01
2.93650180e-01 1.93346548e-03 -3.52816582e-01 3.03172231e-01
3.40254843e-01 4.50212471e-02 7.73160279e-01 -5.87295927e-02
8.15069199e-01 -6.41066909e-01 5.84463239e-01 5.28562844e-01
-4.22523826e-01 4.32018250e-01 5.38606703e-01 -1.53110862e-01
-8.19323778e-01 -1.04600489e+00 -4.71150041e-01 4.89656031e-01
1.21399336e-01 -1.99090153e-01 -6.30965054e-01 -1.15950632e+00
1.85601767e-02 6.73204720e-01 -7.99639881e-01 -1.73171520e-01
2.87510663e-01 -5.36466837e-01 2.39519462e-01 4.45896268e-01
5.02849042e-01 -1.21139377e-01 -5.05018950e-01 2.89900973e-02
-3.65203232e-01 -1.27673554e+00 6.35879412e-02 1.50527269e-01
-8.11113358e-01 -1.46839046e+00 -5.09627640e-01 1.92242384e-01
5.09308696e-01 2.57708102e-01 9.36753154e-01 -7.31469393e-02
4.36160654e-01 9.33001161e-01 -5.80120862e-01 -2.74553597e-01
-6.84637666e-01 -5.12175798e-01 5.00031233e-01 6.26247466e-01
-1.81026906e-01 -6.91516638e-01 -3.90072316e-01 4.60718989e-01
-8.85961950e-01 -2.00888246e-01 2.89912581e-01 1.00941074e+00
6.58430755e-01 3.34379017e-01 7.93288469e-01 -9.29508984e-01
3.24132651e-01 -6.21312737e-01 -7.75893331e-01 8.05790186e-01
-1.01439643e+00 -1.10488288e-01 6.69697762e-01 4.53086458e-02
-1.28253090e+00 -3.28323632e-01 1.29630983e-01 -6.95805013e-01
6.29774183e-02 5.37746310e-01 -6.21811509e-01 -2.64595188e-02
4.33503419e-01 2.01825380e-01 7.82107115e-02 -1.66282773e-01
5.84375799e-01 7.72566259e-01 2.00610220e-01 -7.93643117e-01
6.44031346e-01 8.61586511e-01 3.03721815e-01 -2.88910329e-01
-1.25734973e+00 -2.58622885e-01 -8.82319272e-01 -6.98658466e-01
5.32585084e-01 -1.14783752e+00 -9.05091524e-01 5.57939351e-01
-9.83361959e-01 7.36695290e-01 1.08643789e-02 5.99977851e-01
-5.99979639e-01 6.72258139e-01 -2.52663076e-01 -1.34556460e+00
-4.59858567e-01 -1.33672822e+00 1.12270987e+00 -5.58456667e-02
3.01682591e-01 -1.23712194e+00 -2.51707882e-01 7.93898821e-01
-4.63327840e-02 4.42228913e-01 6.17606878e-01 -9.22060370e-01
-4.58181679e-01 -5.39406538e-01 -3.28833796e-02 8.76810610e-01
-4.96082881e-04 2.38657258e-02 -1.28434265e+00 -3.93708527e-01
6.45262182e-01 -5.10427237e-01 7.21246123e-01 3.08786482e-01
9.08142090e-01 -3.01216125e-01 -5.25583401e-02 1.45908013e-01
1.47635829e+00 -1.78739339e-01 1.34004235e-01 2.35940777e-02
3.78385246e-01 8.06202650e-01 9.03398156e-01 9.69427168e-01
5.95810831e-01 5.79650283e-01 7.91662395e-01 6.98496282e-01
4.49759334e-01 -1.67847369e-02 3.48877102e-01 9.46785629e-01
-2.55869657e-01 -2.75067598e-01 -5.87785482e-01 2.14330494e-01
-1.93412793e+00 -1.04558611e+00 -2.68775206e-02 2.50467157e+00
4.11543012e-01 1.14605688e-01 -1.19572841e-01 4.36172009e-01
7.17638135e-01 1.57843202e-01 -4.08668339e-01 1.19350843e-01
-1.81998774e-01 -4.52861071e-01 8.77262428e-02 4.55583155e-01
-1.03153121e+00 1.36770476e-02 5.51845264e+00 8.77227366e-01
-6.01458192e-01 4.87652421e-01 8.81575108e-01 9.04392824e-02
-6.88180149e-01 4.57395725e-02 -8.05888653e-01 3.84219408e-01
6.92121327e-01 -2.34085351e-01 2.37078816e-02 9.16641057e-01
4.08916734e-02 -4.95302200e-01 -1.10001421e+00 9.46547031e-01
5.41658580e-01 -1.07559466e+00 7.28516281e-03 2.18345687e-01
7.42869735e-01 -4.42391604e-01 8.49317536e-02 9.13745072e-03
2.38342151e-01 -4.57068056e-01 9.63438451e-01 8.44993949e-01
3.85269523e-01 -9.93136048e-01 1.06098187e+00 5.39733887e-01
-1.04273736e+00 -2.38880605e-01 -1.39351025e-01 4.90127534e-01
1.49448365e-01 1.20931041e+00 -1.71450526e-01 1.44506490e+00
4.89842594e-01 8.31784368e-01 -4.17493373e-01 5.89374542e-01
-2.57709712e-01 2.71041870e-01 -1.70496672e-01 3.39087635e-01
-2.54439831e-01 -3.65073651e-01 6.64841413e-01 2.73554653e-01
4.50028330e-01 -1.91696193e-02 1.49049163e-01 5.77858567e-01
9.24970135e-02 -1.79701924e-01 -6.81177378e-01 5.47950506e-01
6.57647789e-01 1.29977655e+00 -5.03079236e-01 -5.83127379e-01
-2.14597195e-01 5.45949697e-01 3.07263553e-01 1.22384690e-01
-8.89401674e-01 3.74464065e-01 2.68169433e-01 -4.09173489e-01
2.07895592e-01 2.71670893e-02 -3.60061228e-01 -1.56404436e+00
3.26046407e-01 -9.34510469e-01 6.29565656e-01 -6.88536406e-01
-1.75927341e+00 5.17810225e-01 3.53207707e-01 -1.84580922e+00
-1.46926567e-01 -5.76310337e-01 -2.15924636e-01 5.01164556e-01
-1.40452325e+00 -1.57863855e+00 -9.28601846e-02 6.92436337e-01
1.08469576e-01 -2.00067535e-01 7.41113067e-01 3.51610631e-02
-3.72224718e-01 5.20483792e-01 2.78400034e-01 -3.07059675e-01
8.30785394e-01 -1.20223558e+00 -6.68104589e-01 7.54144669e-01
1.14372462e-01 3.93324524e-01 6.41155005e-01 -5.04980683e-01
-1.30929744e+00 -8.36208820e-01 2.77799308e-01 -6.48268461e-01
6.69072270e-01 -8.28695893e-02 -8.82513523e-01 4.02113527e-01
6.88878074e-02 4.21222329e-01 1.14283729e+00 2.19206721e-01
-7.86999524e-01 -4.18421477e-01 -1.44221210e+00 2.47120522e-02
4.36366349e-01 -6.68177009e-01 -4.32512194e-01 2.06271484e-01
3.55108917e-01 -2.64743328e-01 -1.30326378e+00 4.75429684e-01
7.68093288e-01 -1.38912225e+00 8.71746898e-01 -3.26313913e-01
3.84835452e-01 -2.06196770e-01 -7.34500289e-01 -1.38211191e+00
-8.45463574e-03 1.43398732e-01 -3.71782780e-01 1.49711454e+00
8.99980366e-02 -7.62477040e-01 3.69066149e-01 5.66454172e-01
2.57255733e-01 -7.02187181e-01 -1.29419088e+00 -7.73516774e-01
4.33759913e-02 -7.71382093e-01 4.32339221e-01 7.35489547e-01
-3.72780114e-02 -5.75967319e-02 -8.04137647e-01 7.41447806e-01
1.43588281e+00 3.11461717e-01 3.91957998e-01 -1.52962136e+00
-3.22278678e-01 2.00354420e-02 -4.52407807e-01 -1.84480995e-01
2.66043246e-01 -6.22600794e-01 -1.89919010e-01 -1.24652123e+00
4.72655654e-01 -2.74027228e-01 -3.66915584e-01 -8.47734418e-03
-2.70555705e-01 -5.65600656e-02 3.61375511e-01 4.13246334e-01
-8.08314323e-01 9.08032417e-01 1.18343019e+00 -3.78669277e-02
5.26305020e-01 4.57827598e-01 -6.88564777e-01 8.57666433e-01
3.12772512e-01 -4.47138309e-01 -4.97438312e-01 1.17604591e-01
4.08639371e-01 6.11821651e-01 6.28605485e-01 -1.01782525e+00
2.12102190e-01 1.58201754e-02 3.76863599e-01 -7.29542494e-01
5.24069846e-01 -1.28257489e+00 1.95326462e-01 3.42852622e-01
-2.16780066e-01 -3.05553287e-01 -3.37055683e-01 1.37178123e+00
-4.27842021e-01 -3.76266539e-01 9.27790046e-01 -4.14926372e-02
-1.24975860e-01 1.05561242e-01 -1.50173098e-01 -2.55875587e-01
1.33102810e+00 5.07599115e-02 -3.44861478e-01 -6.59283519e-01
-1.07201684e+00 3.53073478e-01 3.08896035e-01 2.68263184e-02
6.25252903e-01 -1.66258502e+00 -6.71334147e-01 4.99547310e-02
3.96086872e-01 -2.29213759e-01 6.65524006e-01 9.59760547e-01
2.70731509e-01 4.99189943e-02 -2.77706310e-02 -8.13502789e-01
-1.38111055e+00 6.60059869e-01 2.89738327e-01 -4.90372837e-01
-7.80017972e-02 4.41276789e-01 -1.52028695e-01 -3.42381954e-01
8.99638385e-02 1.00251362e-01 -4.70074624e-01 6.59338534e-01
4.75524992e-01 5.73561072e-01 1.80931598e-01 -7.39115715e-01
-3.29376549e-01 5.88436723e-01 7.37837628e-02 -2.80619621e-01
1.11346805e+00 -5.07649481e-01 -2.27712214e-01 8.53366494e-01
9.75545764e-01 6.79925531e-02 -1.10399687e+00 -3.84908915e-01
-2.30817616e-01 -5.60238063e-01 1.11660592e-01 -7.50533700e-01
-1.07421601e+00 9.08748269e-01 5.81962705e-01 3.74200255e-01
9.97643173e-01 -1.22085385e-01 2.25642264e-01 9.92337316e-02
1.05526698e+00 -9.10613954e-01 9.09360945e-02 5.04148975e-02
9.71583962e-01 -1.65242553e+00 3.76983017e-01 -3.46734345e-01
-9.24868703e-01 1.12614238e+00 3.07970762e-01 1.85596243e-01
1.08276606e+00 -2.30212207e-03 -1.33480743e-01 -1.25406295e-01
-9.76408243e-01 2.38241971e-01 4.49482173e-01 6.51554048e-01
-1.92873254e-02 3.24423224e-01 -1.49332523e-01 9.29438651e-01
3.39057833e-01 -1.60634890e-01 6.11941040e-01 7.08344400e-01
-1.76332310e-01 -1.07387686e+00 -7.57368565e-01 2.96353281e-01
-4.60926682e-01 3.95747781e-01 -2.96190791e-02 4.79213804e-01
9.10259709e-02 1.18770194e+00 -5.54175198e-01 -4.09521461e-01
5.62619530e-02 2.10157827e-01 3.44023854e-01 -3.31729442e-01
-3.22937548e-01 1.59764126e-01 6.82032108e-02 -4.35049564e-01
-9.12262857e-01 -9.73614514e-01 -6.20168328e-01 -2.00504854e-01
-7.92082787e-01 3.14048916e-01 6.67128325e-01 1.12762904e+00
3.15457314e-01 2.35843152e-01 1.10661304e+00 -6.12318933e-01
-1.26374209e+00 -6.93399429e-01 -9.54829097e-01 1.98130071e-01
2.90975988e-01 -9.78123605e-01 -7.04890430e-01 -1.80826440e-01] | [8.511534690856934, 4.498816967010498] |
1b7cd0d8-56ff-4533-882b-4899c237c8e0 | end-to-end-optimized-arrhythmia-detection | 2111.11789 | null | https://arxiv.org/abs/2111.11789v1 | https://arxiv.org/pdf/2111.11789v1.pdf | End-to-End Optimized Arrhythmia Detection Pipeline using Machine Learning for Ultra-Edge Devices | Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia worldwide, with 2% of the population affected. It is associated with an increased risk of strokes, heart failure and other heart-related complications. Monitoring at-risk individuals and detecting asymptomatic AF could result in considerable public health benefits, as individuals with asymptomatic AF could take preventive measures with lifestyle changes. With increasing affordability to wearables, personalized health care is becoming more accessible. These personalized healthcare solutions require accurate classification of bio-signals while being computationally inexpensive. By making inferences on-device, we avoid issues inherent to cloud-based systems such as latency and network connection dependency. We propose an efficient pipeline for real-time Atrial Fibrillation Detection with high accuracy that can be deployed in ultra-edge devices. The feature engineering employed in this research catered to optimizing the resource-efficient classifier used in the proposed pipeline, which was able to outperform the best performing standard ML model by $10^5\times$ in terms of memory footprint with a mere trade-off of 2% classification accuracy. We also obtain higher accuracy of approximately 6% while consuming 403$\times$ lesser memory and being 5.2$\times$ faster compared to the previous state-of-the-art (SoA) embedded implementation. | ['Vineeth Vijayaraghavan', 'Shanthakumar S', 'Vishal Nagarajan', 'Sachin Krishan T', 'Sideshwar J B'] | 2021-11-23 | null | null | null | null | ['arrhythmia-detection', 'atrial-fibrillation-detection'] | ['medical', 'medical'] | [ 2.35152066e-01 -9.22986045e-02 -1.66498810e-01 -1.21558547e-01
-6.50689423e-01 -4.52730566e-01 -1.85796440e-01 6.12684667e-01
-3.03956330e-01 8.40241373e-01 -2.49985754e-01 -8.28634441e-01
-1.85894087e-01 -9.75666285e-01 -5.50359450e-02 -3.50069195e-01
-4.29027230e-01 1.85542211e-01 -1.67059585e-01 4.09793377e-01
-3.60766165e-02 5.34596384e-01 -1.49905884e+00 3.06935340e-01
9.49384034e-01 1.58172178e+00 -4.23856676e-01 7.86363482e-01
3.49650949e-01 2.06210762e-01 -7.97038853e-01 -1.32272333e-01
1.21326536e-01 -5.23174226e-01 -1.17072299e-01 -3.50760937e-01
1.38976127e-01 -9.82184783e-02 2.12633550e-01 5.02881348e-01
1.07161319e+00 -7.75206327e-01 1.97505340e-01 -6.85678899e-01
1.40318945e-02 3.82740706e-01 -3.01697373e-01 5.86571932e-01
2.71898717e-01 1.27120972e-01 4.21635270e-01 -7.00263798e-01
1.45196915e-01 3.88159662e-01 1.01319444e+00 3.15921754e-01
-1.11486316e+00 -6.05574548e-01 -5.16556621e-01 9.36899185e-02
-1.62647295e+00 -6.26541972e-01 4.88742292e-01 -2.40251422e-01
1.21418130e+00 7.90329874e-01 1.21537971e+00 6.48690999e-01
4.01012629e-01 6.46147178e-03 1.10655808e+00 -4.41192806e-01
3.60295922e-01 1.06017493e-01 1.96904480e-01 7.15237498e-01
9.70707417e-01 -3.04499753e-02 -5.12996137e-01 -5.92438936e-01
4.85112101e-01 1.17469616e-01 -2.84709454e-01 1.77803293e-01
-1.36641228e+00 3.36129278e-01 9.85334739e-02 3.61996979e-01
-5.64345896e-01 2.56511748e-01 4.17536646e-01 6.28604814e-02
2.70985276e-01 3.62461358e-01 -6.58357441e-01 -6.59698009e-01
-9.15870905e-01 -2.01826960e-01 7.52635360e-01 5.84030211e-01
3.12567264e-01 3.47579569e-01 -2.92658418e-01 2.16060936e-01
2.67623901e-01 8.28740835e-01 5.09901226e-01 -4.63014841e-01
2.10199818e-01 9.59229052e-01 1.35123730e-01 -7.39857614e-01
-7.85466373e-01 -1.15567017e+00 -1.32191682e+00 -7.52340853e-02
2.51104742e-01 -4.93018538e-01 -5.43554306e-01 1.20539570e+00
1.67881742e-01 4.26765859e-01 3.31060961e-02 4.56621647e-01
5.18157423e-01 1.49085313e-01 1.81268945e-01 -7.08141088e-01
1.86476088e+00 -1.01275332e-01 -5.34764588e-01 -2.08632499e-01
5.21208286e-01 -6.00607455e-01 8.18040550e-01 4.03792173e-01
-9.82252061e-01 -4.62878883e-01 -1.37713623e+00 4.29037929e-01
2.54746922e-03 4.97965783e-01 8.64617229e-01 1.80557191e+00
-1.00027955e+00 6.09470367e-01 -9.66694832e-01 -4.67393398e-01
9.50472593e-01 6.44165158e-01 8.88476521e-02 4.63575631e-01
-6.27671421e-01 4.99546379e-01 -2.87480414e-01 1.40840113e-01
-3.68447810e-01 -1.11161721e+00 -3.67469102e-01 6.62261695e-02
-1.62157685e-01 -1.00007057e+00 5.45537055e-01 -5.22893071e-01
-1.47285378e+00 6.10220730e-01 -2.87102550e-01 -7.57745504e-01
3.07349652e-01 -4.19349104e-01 -8.83660018e-01 -2.14085691e-02
-1.77947521e-01 -1.89494118e-01 7.23827660e-01 -2.24665210e-01
-5.44776976e-01 -7.33122826e-01 -4.28099871e-01 -1.53947771e-01
-5.63382626e-01 -4.48132269e-02 1.24389410e-01 -6.17067039e-01
1.72791500e-02 -9.80522752e-01 -2.32836649e-01 6.65488318e-02
1.71864722e-02 4.54430550e-01 8.96855772e-01 -6.42907977e-01
1.61722863e+00 -1.95549822e+00 -2.43948683e-01 2.75232285e-01
3.89170021e-01 6.33581281e-01 6.44864619e-01 1.57769248e-01
2.97579199e-01 3.67124736e-01 -2.31508970e-01 1.85764104e-01
-5.49916327e-01 -2.10968941e-01 1.18562184e-01 6.40922725e-01
-1.13749318e-02 8.84513080e-01 -4.91942823e-01 -7.01438859e-02
1.95053175e-01 8.13157737e-01 -3.27492565e-01 -8.74807611e-02
3.40008557e-01 4.29495305e-01 -4.56823915e-01 1.01011944e+00
3.28702897e-01 -4.23574597e-01 6.33772671e-01 -4.10951853e-01
-1.46474138e-01 2.63846219e-01 -1.10018504e+00 1.62406874e+00
-6.03741765e-01 3.73927921e-01 4.81332764e-02 -9.91821647e-01
9.54621494e-01 6.78002596e-01 5.40133238e-01 -5.58757722e-01
2.77986199e-01 4.60683435e-01 3.13793689e-01 -3.73746812e-01
-3.71787935e-01 1.07462429e-01 2.46647224e-02 4.35187399e-01
-5.66968620e-01 5.09312510e-01 -1.77121922e-01 -1.79706156e-01
1.44899595e+00 -1.03006922e-02 5.46613276e-01 -8.82822394e-01
6.47758961e-01 -2.18001649e-01 6.30041540e-01 6.94068551e-01
-4.23743337e-01 1.03405401e-01 -4.39329259e-02 -8.98622036e-01
-2.87722409e-01 -1.04945409e+00 -3.21365058e-01 9.40854326e-02
-2.45156884e-01 -7.47085869e-01 -7.13448465e-01 -4.31927204e-01
-1.22106798e-01 1.69493675e-01 -4.56177332e-02 -9.55226794e-02
-6.41207457e-01 -1.25965166e+00 7.99749076e-01 4.59658086e-01
9.45697606e-01 -6.37419701e-01 -1.79944253e+00 4.45949078e-01
9.15451348e-02 -7.18449175e-01 2.24008374e-02 8.08787942e-02
-1.31930745e+00 -9.31569278e-01 -4.85234290e-01 -3.84634286e-01
5.15114427e-01 -2.62673914e-01 1.23076844e+00 4.33713615e-01
-1.02029133e+00 1.49613678e-01 -6.50981218e-02 -6.57569885e-01
-3.49113382e-02 6.74634948e-02 2.66712695e-01 1.81352839e-01
3.69691670e-01 -8.16315413e-01 -1.43747914e+00 -1.74714793e-02
-1.01685338e-01 -2.23283827e-01 8.12868476e-01 4.90830958e-01
6.06473804e-01 -9.27979723e-02 1.03515637e+00 -1.02308607e+00
4.11129296e-01 -2.53373742e-01 -5.38334846e-01 1.02597013e-01
-1.26088774e+00 -1.95215866e-01 6.69208169e-01 -1.41745269e-01
-5.08323908e-01 3.73630285e-01 -4.63265218e-02 2.46757358e-01
-1.55834574e-02 2.79677689e-01 -2.14030191e-01 -1.88227519e-01
6.64547026e-01 -4.71836105e-02 -9.70990881e-02 -1.73754469e-01
-1.13150083e-01 9.75748062e-01 3.28397751e-01 -2.43218571e-01
3.12555850e-01 4.77898002e-01 4.37921077e-01 -1.04173017e+00
-3.10572058e-01 -2.79944688e-01 -2.61251450e-01 -1.97781071e-01
6.48212850e-01 -1.15618992e+00 -1.05332446e+00 2.15545237e-01
-7.92954385e-01 1.64075062e-01 -1.64473638e-01 4.30871218e-01
-2.86595896e-02 2.06895873e-01 -2.73902327e-01 -1.01153326e+00
-1.29789412e+00 -8.94570529e-01 7.55866408e-01 1.30478933e-01
-6.72524750e-01 -5.04591644e-01 -2.32797503e-01 2.86858737e-01
9.81944144e-01 6.81248307e-01 8.40513468e-01 -1.93136215e-01
-4.26839381e-01 -4.12944227e-01 -3.28573212e-02 4.28886712e-02
4.86592889e-01 -3.84392947e-01 -1.12695265e+00 -4.60837126e-01
1.32854611e-01 5.97236633e-01 2.25999326e-01 5.91178894e-01
1.07891464e+00 -1.44142956e-01 -7.31715322e-01 5.69717467e-01
1.40875244e+00 3.30896258e-01 7.08728731e-01 -9.97583792e-02
4.25394028e-01 -2.72253186e-01 3.74227971e-01 6.36321127e-01
-1.92667879e-02 6.77888870e-01 2.08659902e-01 -1.19697399e-01
-2.10448876e-01 4.67442483e-01 1.28839016e-01 5.07106304e-01
-3.96513969e-01 8.94333199e-02 -1.00959527e+00 4.49478507e-01
-1.54450583e+00 -7.78511584e-01 -4.19663906e-01 2.87602758e+00
5.94353497e-01 2.76675582e-01 3.03370982e-01 7.13838041e-01
4.27735150e-01 -5.03948510e-01 -5.65005779e-01 -4.82528925e-01
1.30455736e-02 9.10723567e-01 4.76477772e-01 1.06439151e-01
-1.01314890e+00 1.82901919e-01 5.56412745e+00 1.15054660e-02
-1.50871694e+00 3.79839361e-01 1.05681634e+00 -2.77988434e-01
2.85925537e-01 -1.27757313e-02 -6.17427766e-01 6.31094813e-01
1.67061055e+00 -1.19086243e-01 5.86384796e-02 6.11714303e-01
4.35126036e-01 5.08246832e-02 -9.16559219e-01 1.40960169e+00
-1.25457510e-01 -1.60942435e+00 -4.01542634e-01 1.85995594e-01
2.07804218e-01 -2.30404004e-01 -4.15176928e-01 -2.46577144e-01
-8.05028796e-01 -8.43742609e-01 3.37284952e-01 4.52219069e-01
1.33265686e+00 -8.58007669e-01 8.78565907e-01 7.15516359e-02
-1.42676783e+00 -2.41630182e-01 2.11398259e-01 -5.53458393e-01
2.15363145e-01 1.35768163e+00 -6.94561303e-01 4.23520952e-01
1.04204297e+00 5.10680914e-01 -4.67142940e-01 9.94083166e-01
2.78127909e-01 8.91401291e-01 -6.49416685e-01 -2.25289881e-01
-7.45530665e-01 -7.69171938e-02 3.50060970e-01 1.11060441e+00
8.76385987e-01 1.34269908e-01 4.55254093e-02 2.17947423e-01
1.18170813e-01 1.19724095e-01 -4.11569923e-01 1.85750246e-01
6.46100700e-01 1.30086553e+00 -1.02660286e+00 -4.94954377e-01
-3.18500906e-01 9.59837914e-01 -2.42094070e-01 -3.34536672e-01
-8.34928453e-01 -5.52898824e-01 5.69074750e-01 7.81993032e-01
-8.52705073e-03 -9.54533443e-02 -7.93139398e-01 -8.78613293e-01
3.89098436e-01 -7.94898272e-01 3.39080900e-01 1.38929874e-01
-6.54785931e-01 9.48371649e-01 -5.71328104e-01 -1.41186452e+00
-1.77337199e-01 -3.78535360e-01 -4.35145974e-01 9.92113948e-01
-1.10716939e+00 -9.48030114e-01 -5.08617342e-01 3.36463422e-01
8.61290991e-02 -1.21183954e-01 1.65741646e+00 7.59817779e-01
-6.04803562e-01 6.03417039e-01 -5.25554478e-01 -3.16920727e-01
3.62782866e-01 -9.23977613e-01 2.88608491e-01 1.04038179e+00
1.24524273e-01 7.37257063e-01 4.15699035e-01 -6.91846251e-01
-1.80666220e+00 -9.96262729e-01 1.03462970e+00 -4.16121244e-01
1.44904196e-01 -3.41408759e-01 -2.96078175e-01 5.24568604e-03
-3.57923023e-02 4.69792515e-01 1.04221141e+00 -7.91830570e-03
1.09133698e-01 -6.59212708e-01 -1.32967401e+00 3.67420048e-01
1.15209115e+00 -3.08456630e-01 1.83034196e-01 1.93768367e-01
-3.70045453e-02 -1.55746669e-01 -1.13523090e+00 7.19158173e-01
8.86654496e-01 -1.17980278e+00 9.97273624e-01 -1.64138764e-01
-4.68256205e-01 -4.19747621e-01 2.04902560e-01 -7.50046730e-01
-2.45414317e-01 -1.14784563e+00 -5.78074157e-01 1.08396065e+00
6.07112050e-01 -8.54298174e-01 1.01129830e+00 5.48258960e-01
-1.55346585e-03 -9.29799855e-01 -1.18484294e+00 -6.12789452e-01
-6.09373569e-01 -4.67227995e-01 3.43404442e-01 4.26284611e-01
1.84014559e-01 3.15524697e-01 -1.98545143e-01 2.87256360e-01
6.19909525e-01 1.65577337e-01 2.88462877e-01 -1.52797616e+00
-4.27558154e-01 -1.11120082e-01 -7.63189435e-01 -2.50028700e-01
-7.51431227e-01 -7.36174703e-01 -6.54145718e-01 -1.32173169e+00
-1.60852864e-01 -6.82723463e-01 -5.00022411e-01 4.84158814e-01
-1.23201326e-01 7.97399879e-01 -3.07042092e-01 -8.03138688e-02
-5.88121526e-02 -2.68005788e-01 3.43067497e-01 -1.09817930e-01
-4.55576628e-01 3.16348493e-01 -7.18528688e-01 5.82557261e-01
1.04288197e+00 -3.80335957e-01 -4.63007033e-01 -2.26996839e-01
2.79941022e-01 3.42653506e-02 2.00475618e-01 -1.67156482e+00
1.05169266e-01 5.63854516e-01 6.33469760e-01 -7.87275806e-02
3.07436764e-01 -1.00329852e+00 6.54290795e-01 1.25287044e+00
5.11390328e-01 3.82389545e-01 4.30770308e-01 3.33728880e-01
2.98895299e-01 4.62041706e-01 5.45581102e-01 1.83774516e-01
-4.28431667e-02 2.74535418e-02 -5.72580755e-01 -3.45117331e-01
1.34732497e+00 -3.14675719e-01 -2.69365400e-01 6.11143149e-02
-7.33280420e-01 -4.05177981e-01 2.19409075e-02 1.29987881e-01
4.60528851e-01 -8.14242125e-01 -6.22253895e-01 5.40672004e-01
-5.84202372e-02 -3.34733427e-01 2.55897909e-01 1.18009222e+00
-9.25893128e-01 4.06234860e-01 -1.67414784e-01 -6.97425008e-01
-1.49267673e+00 2.20500842e-01 3.68158847e-01 -6.73465729e-02
-1.00500250e+00 5.27342081e-01 -8.51987720e-01 4.89739120e-01
7.14295208e-02 -3.96123737e-01 1.27872914e-01 -1.13502145e-01
7.67368317e-01 7.90979207e-01 7.05179274e-01 -3.82453725e-02
-9.22754407e-01 7.37013817e-01 3.36194187e-01 2.58912057e-01
1.05402935e+00 6.95850253e-02 -1.37691259e-01 1.11666679e-01
4.95158643e-01 4.44562733e-01 -5.46386957e-01 5.12854576e-01
9.30777490e-02 -2.25215390e-01 3.64772946e-01 -1.16395199e+00
-1.27333558e+00 7.32553422e-01 1.49578214e+00 2.63242304e-01
1.37443340e+00 -2.63396204e-01 7.57091463e-01 8.47309604e-02
6.92701995e-01 -5.58607399e-01 -6.89060211e-01 -4.72571522e-01
2.28086531e-01 -8.73274505e-01 4.12910014e-01 -6.41525447e-01
-2.15546880e-02 1.08569598e+00 5.59197664e-02 -2.01528668e-02
1.19462502e+00 6.25219464e-01 2.49278620e-01 -1.60590515e-01
-4.52725232e-01 2.41124317e-01 3.07603497e-02 7.40341008e-01
7.84711838e-01 3.88237745e-01 -8.41497064e-01 8.39051485e-01
1.63914077e-02 4.73811120e-01 1.40669957e-01 1.15516806e+00
-7.64977932e-02 -1.49454701e+00 -1.40917107e-01 1.13698375e+00
-9.34660554e-01 -2.09763259e-01 -3.35279442e-02 1.80469200e-01
4.38087553e-01 1.21245778e+00 2.97497213e-02 -1.31173044e-01
2.89880544e-01 2.20558435e-01 3.04849327e-01 -5.06309986e-01
-1.01295686e+00 2.29052510e-02 3.70071650e-01 -6.05015039e-01
-3.81716758e-01 -7.09124744e-01 -1.13018024e+00 -6.24054670e-03
-3.98465991e-02 -4.90718633e-02 8.53766561e-01 7.33209550e-01
1.32371235e+00 8.55676234e-01 1.77495509e-01 -4.17608052e-01
-5.52934296e-02 -7.76094913e-01 -3.82286102e-01 -1.57408074e-01
6.79319128e-02 -2.97382236e-01 -1.81827411e-01 2.57887393e-01] | [14.14015007019043, 3.2403175830841064] |
f21f8619-b0ef-4d74-9b70-7669dcae7116 | indoor-smartphone-slam-with-learned-echoic | 2210.08493 | null | https://arxiv.org/abs/2210.08493v1 | https://arxiv.org/pdf/2210.08493v1.pdf | Indoor Smartphone SLAM with Learned Echoic Location Features | Indoor self-localization is a highly demanded system function for smartphones. The current solutions based on inertial, radio frequency, and geomagnetic sensing may have degraded performance when their limiting factors take effect. In this paper, we present a new indoor simultaneous localization and mapping (SLAM) system that utilizes the smartphone's built-in audio hardware and inertial measurement unit (IMU). Our system uses a smartphone's loudspeaker to emit near-inaudible chirps and then the microphone to record the acoustic echoes from the indoor environment. Our profiling measurements show that the echoes carry location information with sub-meter granularity. To enable SLAM, we apply contrastive learning to construct an echoic location feature (ELF) extractor, such that the loop closures on the smartphone's trajectory can be accurately detected from the associated ELF trace. The detection results effectively regulate the IMU-based trajectory reconstruction. Extensive experiments show that our ELF-based SLAM achieves median localization errors of $0.1\,\text{m}$, $0.53\,\text{m}$, and $0.4\,\text{m}$ on the reconstructed trajectories in a living room, an office, and a shopping mall, and outperforms the Wi-Fi and geomagnetic SLAM systems. | ['Guosheng Lin', 'Rui Tan', 'Zhenyu Yan', 'Qun Song', 'Wenjie Luo'] | 2022-10-16 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [-3.14682163e-02 -4.37632143e-01 1.04845785e-01 -7.74956718e-02
-1.13062251e+00 -2.63654619e-01 4.26027961e-02 2.58236498e-01
-4.75248754e-01 8.76880229e-01 4.28749248e-02 -4.15843219e-01
-1.90364406e-01 -8.45232427e-01 -8.88381660e-01 -6.27469897e-01
-5.77664196e-01 -1.72583926e-02 2.11039767e-01 -9.64263082e-02
2.60973543e-01 1.44875228e-01 -1.41966295e+00 -4.08243984e-01
9.51854825e-01 1.40552795e+00 3.57447952e-01 8.31223845e-01
2.45092839e-01 6.49575949e-01 -1.00503445e+00 4.32962507e-01
-1.72212124e-01 -2.41763145e-03 -7.82772973e-02 -6.80270433e-01
8.81676525e-02 -3.48247170e-01 -2.36343607e-01 7.89657593e-01
7.67695129e-01 1.37033299e-01 1.87601656e-01 -1.32371724e+00
2.07025021e-01 2.35502437e-01 -4.30883348e-01 1.16114192e-01
1.13457131e+00 -5.41842461e-01 2.27626368e-01 -6.51792288e-01
-2.86098272e-02 4.67883378e-01 1.34567451e+00 -1.05978876e-01
-7.89885938e-01 -1.05000758e+00 -4.36253399e-01 -2.09609374e-01
-1.96446764e+00 -6.10373437e-01 5.95253706e-01 -2.06822366e-01
9.27115440e-01 4.79301482e-01 4.60004359e-01 1.00311363e+00
7.34723508e-01 1.03383310e-01 7.59623110e-01 -4.30962861e-01
5.05851984e-01 -1.75478514e-02 -3.05457503e-01 7.37949610e-01
2.60950685e-01 9.09374878e-02 -1.01917195e+00 -6.94196224e-01
3.46898288e-01 1.70509234e-01 -4.45504367e-01 -1.61274597e-02
-1.35302317e+00 2.77243704e-01 3.73177111e-01 1.54338598e-01
-3.53991777e-01 5.23160338e-01 -1.27425730e-01 -1.07098274e-01
2.23216742e-01 3.14884901e-01 -9.88040641e-02 -8.54762435e-01
-1.10519028e+00 -2.38611042e-01 7.81081855e-01 1.25759959e+00
8.49189401e-01 9.26497951e-02 5.87136805e-01 4.48057771e-01
4.56632167e-01 1.33620942e+00 4.72247839e-01 -7.54906535e-01
6.13723755e-01 6.08035550e-02 6.42362833e-01 -1.51992285e+00
-8.77404928e-01 -3.92473191e-01 -6.89369678e-01 -6.02458835e-01
7.73591027e-02 -6.29302740e-01 -3.72872464e-02 1.43469727e+00
5.53810298e-01 8.81455839e-01 -1.20279446e-01 5.33486724e-01
3.33905756e-01 4.14157867e-01 -4.29282784e-01 -2.23919868e-01
9.84443069e-01 -4.71893191e-01 -7.50696182e-01 -4.30269331e-01
8.96773040e-01 -8.63712549e-01 9.01770115e-01 2.08436921e-01
-5.50262332e-01 -6.01958811e-01 -1.43867517e+00 6.53423786e-01
-5.44895232e-02 6.03278130e-02 4.18104827e-01 1.09802830e+00
-7.31868327e-01 2.00008124e-01 -1.27983427e+00 -9.06331614e-02
-3.03580493e-01 4.94033307e-01 7.10523352e-02 1.67651117e-01
-1.24444520e+00 3.95759225e-01 -3.13359410e-01 2.47442007e-01
-4.58318263e-01 -4.63677406e-01 -9.01205003e-01 -3.28064620e-01
-5.02446778e-02 -3.71566385e-01 1.00940621e+00 1.86533377e-01
-1.37967086e+00 1.66796461e-01 -5.56920648e-01 -5.46900868e-01
2.07974599e-03 -2.53362447e-01 -1.09924257e+00 5.25048450e-02
6.08826041e-01 -2.23466888e-01 5.09780526e-01 -1.12348080e+00
-1.07838190e+00 -2.98337430e-01 -2.81465918e-01 -7.48305395e-02
-1.14454232e-01 -4.95037973e-01 1.15796581e-01 -1.38183564e-01
8.82336795e-01 -1.23024905e+00 4.56225872e-02 -8.61019909e-01
-2.58029103e-01 5.91695070e-01 6.23573899e-01 -4.07139361e-01
1.52362335e+00 -2.32591128e+00 -8.03523302e-01 4.34552878e-01
-7.14949444e-02 -5.20366311e-01 7.12445617e-01 4.20976430e-01
7.44417906e-01 -2.16279387e-01 1.51470304e-01 -6.78094506e-01
2.01278068e-02 1.29972294e-01 -4.54512089e-01 9.46711123e-01
-1.02947092e+00 3.85120183e-01 -1.11489165e+00 -1.76365852e-01
3.49877298e-01 4.07100320e-01 -4.44881827e-01 -9.58752111e-02
7.26092279e-01 6.36432409e-01 -5.38350999e-01 7.81877458e-01
7.22212255e-01 6.41564056e-02 -5.20949513e-02 3.32942903e-02
-5.17590404e-01 5.99852741e-01 -1.58594835e+00 1.73095572e+00
-9.26301479e-01 4.66719985e-01 3.69836897e-01 -4.32913452e-01
1.08777499e+00 1.94818303e-01 6.65243626e-01 -9.35037971e-01
7.05388039e-02 6.51795030e-01 -7.30127394e-01 -3.22651654e-01
8.19982886e-01 2.33666286e-01 -6.50839448e-01 6.09582007e-01
-3.20449352e-01 -3.36253084e-02 -5.27683318e-01 -1.85802132e-02
1.47522187e+00 -4.59558032e-02 2.05660760e-01 -2.91127950e-01
6.41299784e-01 -2.37628922e-01 6.53216779e-01 1.05998492e+00
-2.22465813e-01 1.53641537e-01 -4.99491781e-01 2.70937057e-03
-3.07441026e-01 -1.02575731e+00 -1.54150844e-01 8.86697352e-01
8.75726342e-01 -5.59747696e-01 -5.11333048e-01 -1.93404153e-01
2.00359002e-01 5.73940814e-01 -8.50968584e-02 -2.30631724e-01
-6.20611966e-01 -5.46977580e-01 9.19428289e-01 2.56856531e-01
7.46336639e-01 -2.43718445e-01 -7.66155720e-01 4.07551050e-01
-7.76662707e-01 -1.10915899e+00 -4.37734336e-01 3.66230011e-02
-4.39002723e-01 -5.34911036e-01 -7.35098496e-02 -5.09533405e-01
4.71823007e-01 8.28367949e-01 5.05300939e-01 -2.42229506e-01
1.82092279e-01 7.60895193e-01 -1.66603610e-01 -4.77085322e-01
1.71025231e-01 -6.57845661e-02 9.51150060e-01 1.78252041e-01
2.09052056e-01 -8.69924188e-01 -7.30669737e-01 8.08319271e-01
9.68963280e-02 -4.35836017e-01 -7.53044486e-02 4.31254655e-01
6.59092903e-01 3.05733860e-01 4.07862127e-01 -1.15474924e-01
3.74033898e-01 -5.76450527e-01 -6.66874588e-01 -3.34486187e-01
-5.72969496e-01 -5.55260181e-01 4.56124753e-01 -2.33059719e-01
-6.92868352e-01 -2.27874424e-02 -2.27199525e-01 2.74103373e-01
-7.34782293e-02 3.69409114e-01 -2.15418458e-01 -6.12128019e-01
7.21173108e-01 5.68148077e-01 -5.71428657e-01 -3.09721559e-01
-1.43961385e-01 1.11732173e+00 8.47771466e-01 -3.25449377e-01
7.91939080e-01 9.75864232e-01 -7.33744726e-02 -1.18343878e+00
-3.63296330e-01 -7.05141127e-01 -1.68063924e-01 -2.13133112e-01
2.90246576e-01 -1.26693320e+00 -1.21978605e+00 3.03989291e-01
-7.48250782e-01 -1.23317428e-01 2.08708391e-01 9.60127413e-01
-5.56808531e-01 4.34967548e-01 -2.60722905e-01 -1.18380392e+00
-1.03709497e-01 -9.84706163e-01 1.26904500e+00 3.43242049e-01
-5.29327810e-01 -7.01449096e-01 1.38759166e-01 4.59083438e-01
6.24581039e-01 2.44978532e-01 -8.83501321e-02 -8.31865519e-02
-5.98388314e-01 -7.33535409e-01 4.04232144e-01 -3.76201540e-01
2.75481611e-01 -8.45726371e-01 -9.78495598e-01 -4.51453865e-01
5.35489857e-01 1.34165078e-01 7.72451013e-02 5.08571863e-01
6.16886079e-01 -3.95796090e-01 -9.61966276e-01 1.23847735e+00
1.14685559e+00 4.66227770e-01 5.37796319e-01 5.61136782e-01
6.30845368e-01 -2.52424628e-01 8.92859280e-01 7.18015552e-01
6.72660112e-01 6.73437297e-01 4.84124690e-01 2.65665442e-01
3.40183437e-01 -5.79267263e-01 4.17446613e-01 8.78289640e-01
1.80352122e-01 -2.22193226e-01 -8.70224953e-01 3.02406192e-01
-1.59292114e+00 -6.61984384e-01 -2.36483529e-01 2.62450671e+00
3.63454401e-01 2.00561166e-01 -3.72655183e-01 2.30795652e-01
6.51569426e-01 2.13958174e-01 -7.25027621e-02 6.34930888e-03
2.02015013e-01 -7.49607459e-02 1.08860385e+00 9.84262407e-01
-1.05441511e+00 4.26462948e-01 4.96189833e+00 6.00532413e-01
-1.17755651e+00 3.18917304e-01 -3.60407144e-01 8.06020200e-02
-2.72107512e-01 2.33751466e-03 -1.03191936e+00 9.67304528e-01
1.26878178e+00 7.51875341e-02 4.28190082e-01 1.03081131e+00
4.04819757e-01 -9.59247887e-01 -7.31373429e-01 1.37085319e+00
4.69082296e-02 -1.24652100e+00 -1.07640362e+00 3.90123665e-01
4.62253273e-01 2.65266359e-01 1.40484005e-01 1.99117526e-01
-1.81385964e-01 -6.15305305e-01 8.23798716e-01 3.83751273e-01
7.50551522e-01 -9.30756807e-01 6.16305113e-01 6.76692486e-01
-1.75549829e+00 -4.51108366e-02 -7.27674887e-02 -5.57231128e-01
5.82830429e-01 9.41203117e-01 -1.21025372e+00 4.81260985e-01
7.91743875e-01 2.38986373e-01 -2.32696280e-01 8.78742576e-01
-1.91142395e-01 7.48995185e-01 -8.54968905e-01 -3.38117629e-01
3.56831844e-03 -3.96717153e-02 6.58683360e-01 1.01909542e+00
1.08142877e+00 -5.92571646e-02 2.90160298e-01 3.26316267e-01
1.90306947e-01 -1.95594728e-01 -7.15849221e-01 6.51180089e-01
1.06781876e+00 8.97313833e-01 -5.63843548e-01 -2.90296357e-02
-1.24734249e-02 9.76843536e-01 -4.68083680e-01 4.04608816e-01
-1.16714859e+00 -9.58526254e-01 6.04991555e-01 3.20757210e-01
-1.54533386e-01 -9.13697600e-01 -2.32999951e-01 -9.73123610e-01
1.92982778e-01 -2.37077013e-01 -2.15866536e-01 -5.32876372e-01
-3.73867929e-01 3.01990509e-01 -5.56276262e-01 -1.50766301e+00
-3.58728945e-01 1.51358455e-01 -2.73837417e-01 7.94530094e-01
-9.75562334e-01 -6.62480295e-01 -6.25164270e-01 6.32789135e-01
-2.91647583e-01 5.34814484e-02 1.03189778e+00 7.46398211e-01
-1.75666943e-01 6.26882672e-01 6.69596732e-01 2.00927351e-02
6.18883789e-01 -7.35573530e-01 3.95042896e-01 7.17760563e-01
7.15351775e-02 9.30215538e-01 7.70212710e-01 -9.69501019e-01
-1.72156465e+00 -8.81857812e-01 1.22997105e+00 -5.85742056e-01
6.45769715e-01 -8.25312018e-01 -2.53346533e-01 6.30114734e-01
-4.22386587e-01 6.94912970e-02 7.46746480e-01 1.15232378e-01
1.34843096e-01 -4.43283528e-01 -1.36482501e+00 4.57242876e-01
1.06499743e+00 -7.49976397e-01 -2.50375807e-01 3.97983491e-01
5.99069953e-01 -8.34488034e-01 -7.37848341e-01 3.88887852e-01
7.98182845e-01 -9.43044305e-01 1.18496192e+00 5.60954571e-01
-7.96883821e-01 -5.36546707e-01 -6.72457278e-01 -1.00203955e+00
-4.86967452e-02 -9.94275033e-01 -3.00055712e-01 1.21102679e+00
7.37079456e-02 -1.21711934e+00 6.55121267e-01 1.25415206e-01
-3.65039259e-02 -2.97388852e-01 -1.56199193e+00 -7.73648739e-01
-8.96687925e-01 -9.78664339e-01 9.51964259e-01 6.06589854e-01
2.09076703e-01 1.78578094e-01 -5.18208206e-01 9.25362289e-01
8.39144111e-01 -1.48068443e-01 9.05650198e-01 -8.13022375e-01
-1.40296936e-01 2.74770856e-01 -4.35415179e-01 -1.48880529e+00
-3.18618238e-01 -3.40282261e-01 3.12148124e-01 -1.28321922e+00
-6.96663976e-01 -1.05494583e+00 -1.31162435e-01 -9.32310373e-02
3.52699101e-01 2.10651845e-01 -4.00771171e-01 2.32055202e-01
-8.91403139e-01 3.71918559e-01 2.03533888e-01 -2.73483973e-02
-4.45121229e-01 6.48945391e-01 -1.37966007e-01 9.90821123e-01
6.92078173e-01 -5.64416111e-01 -2.92163014e-01 -5.92341065e-01
6.59263611e-01 3.18260103e-01 2.60383189e-01 -1.70836782e+00
7.72986412e-01 1.06912494e-01 3.41233402e-01 -6.96612000e-01
5.60456693e-01 -7.94470727e-01 3.17394108e-01 7.22585499e-01
4.14947420e-01 6.20009564e-02 2.04793110e-01 7.32874990e-01
-1.77273005e-02 1.70593843e-01 2.75320381e-01 1.42718032e-01
-4.66331840e-01 -4.28974489e-03 -6.14437282e-01 -2.21347868e-01
7.82461584e-01 -3.75776917e-01 -2.21001416e-01 -8.40317070e-01
-2.28611887e-01 1.70923427e-01 5.09811878e-01 1.28471568e-01
6.31554544e-01 -1.44937515e+00 8.25043172e-02 3.77692819e-01
3.81912175e-03 -1.52494013e-01 5.12348950e-01 1.11343622e+00
-5.27285337e-01 6.98264956e-01 4.37558174e-01 -9.72632647e-01
-8.48792136e-01 4.43149209e-02 4.36355889e-01 3.41218680e-01
-1.96227416e-01 9.33267295e-01 -4.00136262e-01 -5.91212332e-01
4.62402463e-01 -4.17613983e-01 2.76495516e-01 -1.91128999e-01
7.08954990e-01 8.68378043e-01 1.80787489e-01 -8.36978197e-01
-1.00442362e+00 1.01641178e+00 7.79456139e-01 -4.65109587e-01
7.22442865e-01 -1.04950011e+00 -4.56850452e-04 5.17292798e-01
1.24511480e+00 1.05593073e+00 -7.57401645e-01 -9.52170230e-03
-3.18015926e-02 -5.25655031e-01 -1.92656666e-01 -3.23813170e-01
-1.28691033e-01 3.85968983e-01 8.99862707e-01 2.61940330e-01
9.58048582e-01 -1.82971060e-01 1.26554143e+00 3.78933042e-01
1.54450583e+00 -1.20918238e+00 -2.40309492e-01 6.81620717e-01
1.49785861e-01 -8.46446872e-01 -3.95383611e-02 7.78077403e-03
4.76771519e-02 5.21427751e-01 1.68259665e-01 -3.32288146e-02
8.30570459e-01 6.22587204e-01 1.76380083e-01 1.28207520e-01
2.18095228e-01 1.59239337e-01 -3.03211153e-01 8.10630977e-01
2.58987155e-02 2.91741848e-01 4.69049029e-02 7.38841593e-01
-6.76610470e-01 -3.08684379e-01 4.02900726e-01 1.26872718e+00
-8.08153689e-01 -5.63635409e-01 -9.37902987e-01 1.57280341e-01
-3.45207065e-01 1.31247550e-01 3.56232584e-01 4.45877701e-01
4.63426501e-01 1.41515028e+00 -3.66681106e-02 -1.01926231e+00
4.94700521e-01 -8.18711221e-02 2.42187232e-01 -8.88035372e-02
-9.53503251e-02 1.29216075e-01 1.10116616e-01 -1.09803712e+00
-4.52241488e-02 -6.62735164e-01 -1.73639870e+00 -3.71467233e-01
-2.10336477e-01 4.70594764e-01 1.05283010e+00 7.27831423e-01
6.03073359e-01 4.46997911e-01 1.03313017e+00 -9.15009856e-01
-1.83533117e-01 -6.45090938e-01 -8.17002416e-01 -3.10229361e-01
7.97003210e-01 -6.96504474e-01 -5.94530106e-01 -4.89442557e-01] | [6.3529887199401855, 0.9583912491798401] |
c9ca2bd9-c0ea-459c-ba20-65aeebede5a5 | unraveling-cold-start-enigmas-in-predictive | 2305.08120 | null | https://arxiv.org/abs/2305.08120v1 | https://arxiv.org/pdf/2305.08120v1.pdf | Unraveling Cold Start Enigmas in Predictive Analytics for OTT Media: Synergistic Meta-Insights and Multimodal Ensemble Mastery | The cold start problem is a common challenge in various domains, including media use cases such as predicting viewership for newly launched shows on Over-The-Top (OTT) platforms. In this study, we propose a generic approach to tackle cold start problems by leveraging metadata and employing multi-model ensemble techniques. Our methodology includes feature engineering, model selection, and an ensemble approach based on a weighted average of predictions. The performance of our proposed method is evaluated using various performance metrics. Our results indicate that the multi-model ensemble approach significantly improves prediction accuracy compared to individual models. | ['A. Patra', 'K. Ganguly'] | 2023-05-14 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [-4.32223260e-01 -5.23518562e-01 -2.39259794e-01 -4.83781576e-01
-8.22334766e-01 -7.01536775e-01 6.81210756e-01 2.57287204e-01
-2.25620151e-01 5.30898988e-01 2.04123661e-01 -9.65768099e-02
-8.88553634e-02 -6.54369891e-01 -4.29100506e-02 -5.37946522e-01
9.09128189e-02 2.29833618e-01 4.75864977e-01 -5.52742481e-01
7.94691145e-01 1.06192790e-01 -1.83405650e+00 4.10819381e-01
1.20664203e+00 1.25662339e+00 -2.07335174e-01 8.03651214e-01
-1.81882799e-01 8.84477198e-01 -5.04285336e-01 -8.53083253e-01
4.18118000e-01 1.03894427e-01 -3.63301814e-01 -5.02665818e-01
4.41680223e-01 6.61101639e-02 -3.77320498e-02 5.78357160e-01
5.28215408e-01 5.10019064e-01 4.06559527e-01 -1.39387488e+00
4.23176261e-03 8.42320323e-02 -3.35683584e-01 5.68141282e-01
3.48612666e-01 1.21978268e-01 1.03349316e+00 -7.91945517e-01
6.85235858e-01 5.14633954e-01 6.93792582e-01 -4.04074090e-03
-9.82260525e-01 -8.37065995e-01 5.17035007e-01 5.47074199e-01
-1.37693965e+00 -4.92085785e-01 5.10485530e-01 -8.34266961e-01
9.47763741e-01 6.28342092e-01 5.50030887e-01 5.27869701e-01
5.53239763e-01 5.12313783e-01 1.32567334e+00 -2.51433879e-01
3.82820785e-01 6.32968247e-01 6.53338850e-01 1.69036791e-01
6.10700436e-03 -1.08208489e-02 -8.04398060e-01 -4.85931754e-01
-1.71843484e-01 -8.16055946e-03 1.45926014e-01 1.78393960e-01
-6.65723741e-01 7.15129733e-01 5.41523434e-02 1.92836344e-01
-3.18369478e-01 -3.57581437e-01 3.33744794e-01 5.19553006e-01
1.10882258e+00 7.15681076e-01 -5.65942943e-01 -3.20916802e-01
-1.54581559e+00 6.54375255e-01 1.12643301e+00 9.65309978e-01
5.03461659e-01 -2.69631296e-01 -2.81352669e-01 5.62289357e-01
2.12325558e-01 1.29020587e-01 3.97978634e-01 -4.99736935e-01
4.31332737e-01 6.08614087e-01 3.55806261e-01 -1.10280561e+00
-3.47668827e-01 -6.38780475e-01 -4.55637157e-01 6.59615248e-02
-4.64682616e-02 -3.49336147e-01 -6.70953751e-01 1.30917609e+00
5.60079992e-01 3.18278849e-01 -3.86019051e-02 7.97490954e-01
1.12639391e+00 7.52442479e-01 3.23784143e-01 -5.60363889e-01
9.04228866e-01 -1.23627651e+00 -7.68598914e-01 1.29972652e-01
7.18884706e-01 -1.08606243e+00 5.08921146e-01 8.55229437e-01
-7.93221891e-01 -7.03243732e-01 -8.26604486e-01 3.87929022e-01
-6.23780191e-01 -2.49954239e-01 4.24849540e-01 6.53626382e-01
-8.69226158e-01 5.09729803e-01 -1.49764881e-01 -5.57284415e-01
4.95903008e-02 3.89468670e-01 -1.45259753e-01 3.75329703e-02
-9.57815766e-01 1.02519667e+00 2.51355171e-01 -2.20758155e-01
-5.17524064e-01 -7.18932569e-01 -2.24971056e-01 -1.22207843e-01
3.37168992e-01 -4.96528745e-01 1.18740451e+00 -1.05068600e+00
-1.29387903e+00 2.95653850e-01 -2.32996613e-01 -2.31867477e-01
4.18097466e-01 -7.50976443e-01 -1.16071677e+00 -4.89612103e-01
-3.33814144e-01 -2.68285424e-01 7.85190344e-01 -1.03197014e+00
-9.27889228e-01 -2.90113419e-01 6.06214739e-02 2.37280980e-01
-3.67275804e-01 5.00445724e-01 -3.66615027e-01 -3.78965557e-01
-1.28780335e-01 -1.06307268e+00 -5.78890920e-01 -6.65493190e-01
-3.26252013e-01 -1.24748699e-01 9.57168698e-01 -8.50741267e-01
2.21492863e+00 -1.92074692e+00 -8.11453015e-02 6.10613167e-01
5.91443717e-01 2.58944988e-01 2.07418367e-01 7.39515364e-01
3.62115055e-01 2.91827857e-01 6.85224354e-01 -1.82467833e-01
-1.97761014e-01 -2.10875571e-01 -9.44908429e-03 1.44651486e-02
-3.84213567e-01 2.30063736e-01 -4.18086976e-01 -6.10079169e-01
-1.14537664e-02 2.26186484e-01 -4.10686344e-01 5.39181471e-01
-3.45697224e-01 3.96614403e-01 -4.31536108e-01 7.60770082e-01
7.13961482e-01 -1.19198352e-01 2.78410286e-01 -1.11455083e-01
-6.21132374e-01 -8.57814699e-02 -1.17462456e+00 1.20450008e+00
-3.16414177e-01 4.07476693e-01 -3.59596848e-01 -3.71506125e-01
1.17316449e+00 1.38011366e-01 2.87766993e-01 -5.15155554e-01
1.29314706e-01 1.61174402e-01 1.51540875e-01 -7.30999470e-01
9.48824883e-01 9.02642608e-02 -7.15029761e-02 2.36518592e-01
1.87296093e-01 3.64102870e-01 3.23236525e-01 4.56540912e-01
1.05523872e+00 1.58127248e-01 3.77148569e-01 -2.29881912e-01
4.35768008e-01 3.63334715e-02 7.50047445e-01 7.33772337e-01
-2.40957662e-01 3.63239169e-01 3.13962042e-01 -6.79024696e-01
-9.69746113e-01 -4.50637758e-01 8.92372206e-02 1.39926171e+00
1.50659904e-02 -1.01262832e+00 -4.71913010e-01 -8.64889145e-01
2.29514483e-02 9.58517015e-01 -4.72165257e-01 1.52803436e-01
-3.26077312e-01 -7.71448612e-01 -1.81335673e-01 2.21841574e-01
2.30970234e-01 -4.59609658e-01 -1.42856538e-01 4.05334413e-01
-4.78312895e-02 -8.43320549e-01 -5.58720529e-02 3.48653551e-03
-8.74951065e-01 -7.93749571e-01 -2.44597599e-01 -7.14929625e-02
1.64900884e-01 1.66398928e-01 1.32554591e+00 1.31596848e-01
1.42875865e-01 1.54059708e-01 -7.34404325e-01 -6.01178288e-01
-7.52861649e-02 5.48435330e-01 1.05415620e-01 3.28143001e-01
5.49863100e-01 -5.07876754e-01 -6.30153000e-01 3.93706352e-01
-3.62938672e-01 -6.05330840e-02 3.41266870e-01 3.57245475e-01
5.12806952e-01 -9.83194187e-02 5.93539238e-01 -1.15629697e+00
8.19814444e-01 -1.19656980e+00 -6.19475663e-01 6.13291919e-01
-1.02039707e+00 -3.09249818e-01 4.64086443e-01 -1.78055331e-01
-1.16369677e+00 -9.45942029e-02 -7.24318177e-02 -2.25056503e-02
-4.03877907e-02 6.65839434e-01 1.08802609e-01 -4.58483905e-01
5.37976921e-01 -2.21791137e-02 -2.98610598e-01 -8.56722176e-01
1.31274208e-01 6.83619678e-01 -1.10740058e-01 -2.52263397e-01
7.06839621e-01 -9.79992300e-02 -1.36787534e-01 -5.66915810e-01
-1.02507210e+00 -8.61279428e-01 -7.29377985e-01 -9.39178646e-01
6.03544712e-01 -9.33519959e-01 -5.71130097e-01 2.14138165e-01
-6.61969006e-01 2.63647497e-01 2.68257141e-01 4.58849400e-01
-1.54587075e-01 -1.85530528e-01 5.05765062e-03 -1.12284589e+00
-7.27175713e-01 -7.90923953e-01 2.61060506e-01 6.44759893e-01
-3.13932747e-01 -7.81322420e-01 7.40729570e-01 5.39223552e-01
1.00572360e+00 4.98468786e-01 6.53674841e-01 -1.40809464e+00
-6.01734281e-01 -7.86487877e-01 2.69761831e-01 5.95405363e-02
-6.57177866e-01 6.08112276e-01 -9.74408031e-01 -1.53005376e-01
-3.30213755e-01 2.27119830e-02 6.46527886e-01 1.98629305e-01
1.07771420e+00 -2.43529916e-01 -5.38002670e-01 2.94791639e-01
1.86656570e+00 -6.25344459e-03 5.14776826e-01 5.47982752e-01
4.46276426e-01 3.48971218e-01 1.06620514e+00 9.57505226e-01
6.05240762e-01 9.19002473e-01 3.45069677e-01 1.73084915e-01
4.18217003e-01 1.03820130e-01 5.15287220e-02 1.03504515e+00
-5.73080480e-01 -3.83235514e-01 -1.14169836e+00 1.36625513e-01
-2.10896134e+00 -1.06091452e+00 -4.31210637e-01 2.35862851e+00
2.28450447e-01 3.01351607e-01 5.19396126e-01 -9.87770632e-02
4.55728203e-01 3.63798998e-02 -3.98058832e-01 -8.18200111e-01
6.85021132e-02 1.49373844e-01 4.13612098e-01 1.58419028e-01
-1.13762379e+00 6.57872438e-01 7.11309671e+00 6.42506421e-01
-8.33704770e-01 5.09573042e-01 6.11121833e-01 -5.96672237e-01
-9.57471430e-02 3.83811414e-01 -1.23276305e+00 7.07392216e-01
1.39485466e+00 -5.70995510e-01 -4.05109040e-02 1.01622593e+00
2.10876361e-01 -3.64498943e-01 -6.97872341e-01 6.89728737e-01
3.45914781e-01 -1.35269749e+00 -4.42756861e-01 3.23108360e-02
1.09287310e+00 3.13218445e-01 -3.08530815e-02 5.53633213e-01
2.93805689e-01 -5.75749457e-01 5.89339554e-01 1.38793445e+00
6.50014877e-02 -8.48806918e-01 9.07564759e-01 4.87796575e-01
-1.03161716e+00 -3.86068761e-01 6.32212013e-02 -1.43145263e-01
1.65169999e-01 8.81458879e-01 -3.77508640e-01 6.98974371e-01
8.71861041e-01 4.26960945e-01 -8.58679950e-01 1.82805312e+00
4.52318341e-01 8.43775928e-01 -3.09958369e-01 -2.81843334e-01
-6.78912690e-03 2.63091419e-02 5.02003014e-01 1.27520847e+00
4.56300825e-01 1.71108603e-01 3.80948514e-01 1.32837817e-02
2.11930826e-01 7.84427226e-01 -4.20602709e-01 3.66227090e-01
4.41394895e-01 1.53809404e+00 -2.74504006e-01 -5.41761518e-01
-5.95814109e-01 3.64500135e-01 2.48262912e-01 2.33440980e-01
-8.92082512e-01 -2.88578331e-01 5.52303851e-01 3.13066453e-01
2.84031719e-01 -1.30694970e-01 -2.96930283e-01 -1.21285594e+00
-4.36078664e-03 -8.26373398e-01 4.68037337e-01 -5.44319451e-01
-1.32444930e+00 1.15103567e+00 1.00738090e-02 -1.65404832e+00
-4.67908494e-02 1.02845363e-01 -1.06502986e+00 7.74731338e-01
-1.20343590e+00 -1.04920447e+00 -6.31900430e-01 2.99811631e-01
3.76145661e-01 -5.81318974e-01 7.70504415e-01 3.09035629e-01
-8.19616735e-01 6.47246003e-01 6.27668262e-01 -7.58175373e-01
8.29181135e-01 -9.28412020e-01 6.23848177e-02 8.01274657e-01
-1.48401484e-01 4.40419197e-01 9.16140735e-01 -6.67604625e-01
-9.92435813e-01 -9.16166544e-01 1.13504970e+00 -5.57398498e-01
5.23736596e-01 -3.71783078e-01 -7.22396314e-01 2.99310654e-01
4.82954174e-01 2.69833617e-02 1.52654660e+00 7.81969249e-01
-3.32677782e-01 -6.42902792e-01 -1.20453250e+00 2.08945841e-01
3.21597248e-01 7.84918144e-02 -2.54498591e-04 4.71829385e-01
2.40116164e-01 -1.30357757e-01 -1.37841785e+00 2.18152687e-01
1.01049387e+00 -1.18392992e+00 3.15932661e-01 -8.46973240e-01
3.76743019e-01 -2.10515469e-01 -3.78480643e-01 -1.26416957e+00
-5.57464838e-01 -6.39896095e-01 -6.52116895e-01 1.39976442e+00
7.25409091e-01 -2.20701009e-01 7.27852046e-01 1.32626784e+00
1.26574367e-01 -9.05971646e-01 -7.26026773e-01 -6.05151594e-01
-2.88719028e-01 -3.07175756e-01 5.95606327e-01 6.51925921e-01
3.62567641e-02 1.71435744e-01 -8.87111723e-01 -2.28971824e-01
5.66957951e-01 5.38226552e-02 1.25827408e+00 -1.55160546e+00
-2.61767000e-01 -2.42819753e-03 -5.79968810e-01 -3.42255086e-01
-3.13288510e-01 -5.03882408e-01 -3.72048259e-01 -1.11523879e+00
3.44521970e-01 -5.73598623e-01 -8.77936661e-01 -3.34874243e-02
-2.02223063e-01 1.37721881e-01 5.52379489e-01 3.74285519e-01
-1.23172462e+00 9.58978683e-02 6.49680853e-01 3.64438772e-01
-3.19974065e-01 3.99903536e-01 -6.30295038e-01 6.66458368e-01
1.02295041e+00 -6.90665603e-01 -2.64311254e-01 -2.76514348e-02
4.63285774e-01 -1.21861055e-01 -3.42040770e-02 -1.42492867e+00
2.90421188e-01 -3.53714705e-01 3.39471787e-01 -7.41405666e-01
2.31412128e-01 -7.15971828e-01 6.00720644e-01 -3.58632905e-03
-4.57548887e-01 3.45458686e-01 7.04053566e-02 7.02295780e-01
-3.51580769e-01 -6.34518042e-02 5.60996950e-01 -3.67360301e-02
-9.40022528e-01 1.98003292e-01 -2.65110254e-01 -2.14788064e-01
1.47062075e+00 -2.76652109e-02 -4.65974003e-01 -4.49942857e-01
-1.07627690e+00 2.23969713e-01 1.77433431e-01 6.36716545e-01
3.11201423e-01 -1.21063936e+00 -7.91053236e-01 -7.84119740e-02
3.87983620e-01 -9.83288765e-01 5.16712308e-01 1.09959745e+00
-2.08323330e-01 1.83005691e-01 -2.03993723e-01 -1.95736483e-01
-1.78189087e+00 3.19062948e-01 4.21367139e-02 -6.63854420e-01
-4.01311107e-02 9.05046105e-01 -4.14515167e-01 -9.07093734e-02
1.17559031e-01 3.13709676e-01 -6.72482193e-01 2.03839689e-01
5.77552259e-01 6.54328525e-01 2.38519654e-01 -7.76349604e-01
-3.76118183e-01 3.34056646e-01 -6.47038937e-01 8.31091776e-02
1.46723616e+00 -1.29495412e-01 1.04984336e-01 5.54171085e-01
9.29500401e-01 -1.16987847e-01 -6.89945340e-01 -3.02779138e-01
6.93553910e-02 -8.36905539e-01 3.94868344e-01 -1.18831933e+00
-9.81773794e-01 4.80118424e-01 9.37428653e-01 3.72431129e-01
1.28881550e+00 -2.80601472e-01 6.25663817e-01 2.27596611e-01
3.07699174e-01 -1.66237676e+00 -3.79724681e-01 2.85047680e-01
7.07094252e-01 -1.57949495e+00 4.09066886e-01 -1.32886946e-01
-9.64041948e-01 9.72413778e-01 1.05020082e+00 -1.39149562e-01
1.31475914e+00 9.16039869e-02 -5.91118000e-02 -3.65484357e-02
-1.62463093e+00 -1.63724139e-01 5.83117127e-01 3.14396441e-01
7.44097531e-01 1.74167529e-01 -7.87783980e-01 8.54941070e-01
-9.85596888e-03 2.97309160e-01 3.48763317e-01 1.08085060e+00
-5.54016173e-01 -1.14069724e+00 -2.21780464e-01 1.16189623e+00
-6.20141029e-01 -2.20384359e-01 -2.78986782e-01 4.79106635e-01
4.06762332e-01 1.13472068e+00 -1.97474241e-01 -1.25073087e+00
3.62953901e-01 2.91124195e-01 -1.66311190e-01 -3.43785077e-01
-1.38313437e+00 6.19604476e-02 5.59943855e-01 -5.44031084e-01
-3.97946000e-01 -7.59143949e-01 -1.37649849e-01 -6.55755520e-01
-7.45197237e-01 2.77442873e-01 1.02951050e+00 9.12080050e-01
6.54207647e-01 1.51683807e-01 1.04695833e+00 -4.93928611e-01
-2.51868576e-01 -1.03195620e+00 -4.79378879e-01 1.83629066e-01
-4.95108664e-02 -6.60111904e-01 -2.52552718e-01 -1.67967603e-01] | [10.098807334899902, 5.762589931488037] |
ac516375-654b-4aa3-a741-1f07c1907e82 | unsupervised-paraphrasability-prediction-for | null | null | https://aclanthology.org/2022.naacl-main.237 | https://aclanthology.org/2022.naacl-main.237.pdf | Unsupervised Paraphrasability Prediction for Compound Nominalizations | Commonly found in academic and formal texts, a nominalization uses a deverbal noun to describe an event associated with its corresponding verb. Nominalizations can be difficult to interpret because of ambiguous semantic relations between the deverbal noun and its arguments. Automatic generation of clausal paraphrases for nominalizations can help disambiguate their meaning. However, previous work has not identified cases where it is awkward or impossible to paraphrase a compound nominalization. This paper investigates unsupervised prediction of paraphrasability, which determines whether the prenominal modifier of a nominalization can be re-written as a noun or adverb in a clausal paraphrase. We adopt the approach of overgenerating candidate paraphrases followed by candidate ranking with a neural language model. In experiments on an English dataset, we show that features from an Abstract Meaning Representation graph lead to statistically significant improvement in both paraphrasability prediction and paraphrase generation. | ['Carol Carol Webster', 'Ho Hung Lim', 'John Sie Yuen Lee'] | null | null | null | null | naacl-2022-7 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 4.78331745e-01 2.79061049e-01 -4.54320431e-01 -6.75525069e-01
-7.48677552e-01 -1.09566379e+00 8.38036001e-01 8.83889794e-01
-3.95423770e-01 1.05481768e+00 8.31215799e-01 -7.05419779e-01
-2.01366797e-01 -8.38163733e-01 -7.69081652e-01 -3.63350749e-01
6.05708063e-01 7.69635379e-01 1.96260419e-02 -3.69525880e-01
6.31359577e-01 6.60547197e-01 -1.47945404e+00 7.47815251e-01
8.83916676e-01 5.07696450e-01 2.27427945e-01 3.37670073e-02
-5.01850307e-01 4.54273373e-01 -1.01291108e+00 -6.38815165e-01
1.17684700e-01 -6.75387681e-01 -1.31956518e+00 -3.91243584e-02
5.74353635e-01 3.14423770e-01 2.77465254e-01 9.47310150e-01
8.68538395e-02 7.28666633e-02 9.29449201e-01 -9.30934489e-01
-6.81978405e-01 1.21646225e+00 -1.75369248e-01 3.79964232e-01
9.79777932e-01 -3.73115689e-01 1.55354345e+00 -9.32840407e-01
7.98057497e-01 1.44361115e+00 3.87958378e-01 4.40278202e-01
-1.62849355e+00 -5.77744544e-01 -6.14924394e-02 2.59487271e-01
-1.15467012e+00 -2.54714996e-01 7.78578401e-01 -2.35798553e-01
1.31277382e+00 2.49896243e-01 9.58716512e-01 1.23587108e+00
5.07785201e-01 4.63451296e-01 1.12503314e+00 -6.79917216e-01
2.45884150e-01 -2.71485391e-04 3.42173427e-01 1.09556913e-01
5.13957024e-01 -3.91187966e-01 -4.82636511e-01 -3.08485627e-01
2.99345642e-01 -3.66542011e-01 -5.31387925e-02 -1.30593240e-01
-1.20731831e+00 1.06510031e+00 3.82589579e-01 4.81575876e-01
-5.26995122e-01 -4.15246338e-02 4.36250627e-01 5.64184308e-01
-3.40987593e-02 1.04309380e+00 -3.65216762e-01 -3.98331024e-02
-9.19213712e-01 6.47559941e-01 8.97762656e-01 7.51865804e-01
4.82380956e-01 -2.65455663e-01 1.01979062e-01 9.51782882e-01
-2.53047850e-02 2.17423901e-01 9.88949180e-01 -9.72775638e-01
5.09214044e-01 8.74234974e-01 1.79042831e-01 -7.94147015e-01
-2.33145148e-01 9.92644280e-02 -2.79360592e-01 -2.59565383e-01
4.04848963e-01 8.66502300e-02 -3.47342849e-01 1.73851526e+00
8.25674981e-02 -3.69978219e-01 5.72508693e-01 6.28163815e-01
9.45452154e-01 8.36214423e-01 2.84193665e-01 -4.64553922e-01
1.41647673e+00 -3.58306170e-01 -4.35453206e-01 -3.63608181e-01
5.96752703e-01 -1.04169977e+00 1.40522373e+00 2.01890156e-01
-1.39865029e+00 -2.99530476e-01 -1.09312165e+00 -2.10316032e-01
-2.97618061e-01 -3.15342039e-01 4.17611390e-01 3.79740238e-01
-4.18676823e-01 7.02973843e-01 -4.88901854e-01 -4.07260448e-01
-1.18052185e-01 3.48108470e-01 -4.53299433e-01 3.89293820e-01
-1.37757230e+00 1.09315515e+00 8.37051868e-01 -3.69108886e-01
-4.94384430e-02 -7.13501632e-01 -1.01418078e+00 5.74245900e-02
7.61960521e-02 -6.99017823e-01 1.28311169e+00 -1.28649545e+00
-1.09501565e+00 1.38644862e+00 -4.09999400e-01 -7.73845673e-01
-4.77006473e-02 1.05470724e-01 -3.73854965e-01 1.06727025e-02
6.21263385e-01 6.41384006e-01 9.01355922e-01 -7.69119322e-01
-7.09982038e-01 -2.77840436e-01 4.20519799e-01 3.38723749e-01
-1.61888793e-01 3.64746004e-01 3.18949431e-01 -9.22825277e-01
6.23232007e-01 -8.65154028e-01 2.17224002e-01 -6.51065290e-01
-4.82035875e-01 -8.85210693e-01 4.46171373e-01 -2.13400692e-01
1.45040858e+00 -1.78654909e+00 4.20928091e-01 7.68698826e-02
-1.23195566e-01 -2.64031682e-02 2.31076911e-01 5.63014627e-01
-7.05014110e-01 3.55715543e-01 -7.53369182e-02 4.28898335e-01
1.80588767e-01 3.87013882e-01 -7.70982146e-01 6.26969114e-02
8.23307931e-02 9.47131515e-01 -8.51674378e-01 -2.60373831e-01
-2.25695316e-02 -1.18828632e-01 -3.40293854e-01 -4.36554551e-02
-1.52611494e-01 -3.40215743e-01 -2.24367186e-01 4.23379123e-01
3.81477594e-01 -6.55968487e-02 4.14738774e-01 -4.03936058e-01
8.80516022e-02 1.17258942e+00 -1.13623905e+00 1.14030814e+00
-8.18622947e-01 3.43468994e-01 -3.87545705e-01 -1.08207488e+00
7.18905866e-01 1.99207172e-01 -1.77214444e-01 -5.33272982e-01
-1.01998381e-01 7.52008617e-01 2.10507244e-01 -2.17393428e-01
5.35622358e-01 -8.63273084e-01 -6.96326375e-01 5.64272106e-01
-1.98939696e-01 -6.94701731e-01 5.48249006e-01 2.94558316e-01
7.64482796e-01 4.10151593e-02 1.03614485e+00 -4.92559075e-01
7.54756749e-01 2.32858703e-01 5.79976976e-01 6.52949154e-01
3.60640079e-01 3.97403359e-01 6.09210312e-01 -4.84335601e-01
-1.04455030e+00 -1.63246191e+00 -4.57592547e-01 9.40368116e-01
5.27767390e-02 -5.39676070e-01 -5.81221700e-01 -7.17842937e-01
1.13570720e-01 1.45024085e+00 -4.00242418e-01 -2.65572459e-01
-8.86535406e-01 -6.27923846e-01 5.21091878e-01 5.86040497e-01
-6.96814731e-02 -1.35569596e+00 -6.46406472e-01 4.62072134e-01
-5.06631255e-01 -9.16800022e-01 -3.50222766e-01 4.98822570e-01
-7.31885731e-01 -8.92081857e-01 -1.69339046e-01 -1.13931584e+00
6.15464211e-01 -1.30959690e-01 1.29453766e+00 -5.91918733e-03
1.51304388e-02 -2.57782459e-01 -3.64300936e-01 -4.59984332e-01
-9.53127265e-01 6.90998212e-02 1.51034385e-01 -5.27305543e-01
8.94864440e-01 -6.01573229e-01 -3.17269415e-02 -1.35988459e-01
-9.81724679e-01 -2.27405429e-02 1.10839590e-01 8.22417259e-01
5.21818757e-01 -2.56514817e-01 7.81451941e-01 -1.06301641e+00
1.08518839e+00 -4.93045628e-01 -3.06330174e-01 1.10188305e-01
-4.54454243e-01 4.82982993e-01 8.98758590e-01 -4.85955328e-01
-9.64909494e-01 -1.53344572e-02 -2.94230998e-01 1.96201339e-01
-1.89949080e-01 6.79630518e-01 -2.50586905e-02 5.60667455e-01
1.01318157e+00 1.95243403e-01 -2.89067894e-01 -7.51070678e-02
2.59130120e-01 6.34551108e-01 4.96939570e-01 -9.08598602e-01
8.16800773e-01 1.49498716e-01 1.23786487e-01 -6.93572283e-01
-1.26967883e+00 -4.04582173e-01 -5.96459985e-01 3.46210569e-01
6.76924706e-01 -7.90826440e-01 -3.50438952e-01 -2.15650514e-01
-1.40503561e+00 3.16440284e-01 -6.39109612e-01 2.62142003e-01
-7.19803393e-01 5.96525192e-01 -5.51499486e-01 1.19347908e-01
-2.17064574e-01 -1.08202958e+00 9.35045898e-01 1.76537320e-01
-1.41403055e+00 -9.12313163e-01 -2.01342180e-01 4.39527541e-01
8.45883712e-02 -5.34805022e-02 1.79336143e+00 -1.18096220e+00
1.23342209e-01 -9.67204943e-02 -6.19590282e-02 9.62251127e-02
3.74020934e-01 -2.25324497e-01 -3.64699215e-01 4.09050807e-02
1.02319412e-01 -6.32557750e-01 5.69962800e-01 1.11060850e-01
6.34954929e-01 -6.95667267e-01 -1.01339251e-01 3.17966223e-01
1.24385941e+00 1.07950866e-01 4.90531445e-01 5.86720169e-01
1.65683702e-01 7.19478190e-01 6.44192219e-01 5.63976094e-02
2.35983595e-01 7.20958054e-01 -1.32525921e-01 5.21147609e-01
-1.14511870e-01 -3.60326827e-01 4.79215950e-01 4.67928439e-01
3.12001288e-01 -1.79612324e-01 -7.73233056e-01 7.21207976e-01
-1.46992517e+00 -1.12986517e+00 -3.43667835e-01 2.14380288e+00
1.34986103e+00 4.47257221e-01 -1.47187397e-01 2.48435110e-01
5.29337168e-01 7.25163966e-02 -1.32973343e-01 -9.76664245e-01
-3.98132175e-01 5.82091808e-01 -7.19706565e-02 7.62142718e-01
-7.36079693e-01 1.40443540e+00 5.77592564e+00 8.56802464e-01
-8.34698200e-01 -2.29799107e-01 1.87847272e-01 1.57653943e-01
-6.77331686e-01 2.52705842e-01 -1.01967227e+00 3.18195641e-01
8.96190941e-01 -7.67180562e-01 1.79938599e-01 5.45600951e-01
3.38665336e-01 -1.15769781e-01 -1.56740499e+00 5.70041060e-01
2.94464618e-01 -1.61745262e+00 7.99105585e-01 -4.36556250e-01
6.37232423e-01 -5.24797738e-01 -2.88395345e-01 2.85654813e-01
9.91594642e-02 -9.73347366e-01 7.01683223e-01 9.00514349e-02
2.38645777e-01 -8.09596360e-01 7.02154696e-01 1.19164199e-01
-1.04320312e+00 -4.48690429e-02 -5.94703794e-01 -3.30425352e-01
1.16143607e-01 1.96989290e-02 -9.45418894e-01 2.28393912e-01
1.93700239e-01 4.22660083e-01 -5.55656374e-01 6.15956128e-01
-8.24460983e-01 6.05852127e-01 -2.43111297e-01 -2.99901664e-01
2.53162354e-01 -2.36971304e-01 8.31657708e-01 1.19964314e+00
2.56624848e-01 1.76293418e-01 1.95813119e-01 1.00436676e+00
-3.60530764e-02 6.85908258e-01 -5.62969983e-01 -1.02635369e-01
6.95567429e-01 8.70751739e-01 -7.64382601e-01 -5.61351418e-01
-2.85768092e-01 1.09982145e+00 2.37855315e-01 2.54392475e-01
-6.87292039e-01 -5.92501581e-01 4.53720689e-01 3.19532305e-01
9.41289291e-02 2.34331474e-01 -3.28859299e-01 -1.14025700e+00
1.86455071e-01 -8.38861346e-01 4.61885333e-01 -1.07400894e+00
-1.53284252e+00 3.78036499e-01 3.44393700e-01 -9.54029202e-01
-8.44682276e-01 -5.02876163e-01 -7.99357951e-01 1.04888141e+00
-8.99990976e-01 -1.11873388e+00 3.89273465e-01 1.69776097e-01
1.00307214e+00 -1.42243102e-01 1.09462786e+00 -2.99006611e-01
-1.32074177e-01 4.24582690e-01 -2.84459502e-01 -1.38326287e-01
6.42068028e-01 -1.40554976e+00 1.71941280e-01 8.06189537e-01
2.09350035e-01 1.15744829e+00 1.18116200e+00 -7.62081504e-01
-8.27793360e-01 -9.05812562e-01 1.90481329e+00 -2.89815962e-01
1.16540432e+00 -9.58944559e-02 -9.89000380e-01 7.59301126e-01
2.93159813e-01 -6.65252030e-01 7.13732362e-01 1.58725068e-01
-5.45082629e-01 2.17777088e-01 -9.12814856e-01 9.98190999e-01
8.15027118e-01 -6.14391983e-01 -1.78171515e+00 6.34234071e-01
7.86076248e-01 -1.45977587e-01 -6.48686707e-01 2.41390124e-01
3.64397973e-01 -6.18544400e-01 1.11398554e+00 -1.00720406e+00
7.98928142e-01 -1.09380044e-01 -2.79902279e-01 -1.29724014e+00
-2.21162036e-01 -4.21698153e-01 1.57680124e-01 1.05171061e+00
7.99165010e-01 -2.76602030e-01 7.56697595e-01 5.10074854e-01
-1.22796163e-01 -5.24332583e-01 -9.67896938e-01 -7.72717774e-01
5.33134818e-01 -2.17134744e-01 4.63435233e-01 9.30557489e-01
7.12614417e-01 1.09560418e+00 2.87339836e-01 -3.14764142e-01
3.29006374e-01 7.31992960e-01 1.72172904e-01 -1.37697196e+00
-2.13800460e-01 -6.76790655e-01 -3.62843841e-01 -8.94212604e-01
1.02882779e+00 -1.56069505e+00 -2.59419024e-01 -1.60233343e+00
1.90513536e-01 7.09877312e-02 1.76563874e-01 5.28172433e-01
6.55968720e-03 2.86461979e-01 1.05638027e-01 4.40465473e-02
-1.25855848e-01 2.15589061e-01 6.37023628e-01 -2.42193997e-01
-4.92107362e-01 5.80865182e-02 -8.51735532e-01 1.26792049e+00
9.81483638e-01 -7.28440106e-01 -5.05032063e-01 -3.37008722e-02
7.61352837e-01 2.07367942e-01 1.98683962e-01 -4.56293583e-01
4.12410460e-02 -4.68060076e-01 2.30039269e-01 -3.34202379e-01
2.08416000e-01 -6.01362824e-01 1.01416267e-01 6.09329283e-01
-7.40354538e-01 7.19455659e-01 2.89457440e-01 1.62550837e-01
-2.75313050e-01 -1.05238080e+00 7.83190846e-01 -2.59528697e-01
-5.94905853e-01 -6.54339492e-01 -8.85345936e-01 4.57996845e-01
8.88421655e-01 -4.54100937e-01 2.19327528e-02 -1.78883418e-01
-8.38238776e-01 -2.22101390e-01 5.60823798e-01 5.33647597e-01
8.60129476e-01 -1.25287271e+00 -7.04916716e-01 1.22067377e-01
2.39138663e-01 -5.16134083e-01 -3.05372864e-01 2.83968151e-01
-5.84215581e-01 3.71915102e-01 -1.61816895e-01 -8.28416869e-02
-1.53363383e+00 4.32823956e-01 2.00565591e-01 -1.89892337e-01
-5.99664927e-01 8.21351409e-01 -2.46357433e-02 -5.55529237e-01
4.54325639e-02 -5.87164402e-01 -3.58916283e-01 4.60475147e-01
2.73784190e-01 -5.62951006e-02 2.00987086e-01 -9.93724585e-01
-3.67200643e-01 3.49418759e-01 -3.17980587e-01 -2.28274614e-01
1.00725627e+00 -3.30769718e-02 -4.12710488e-01 6.64241672e-01
1.12216783e+00 4.09229487e-01 -7.25736395e-02 2.89422590e-02
2.75050431e-01 -1.63988128e-01 -4.28521574e-01 -7.90398061e-01
-8.86240527e-02 2.68542200e-01 -2.19405264e-01 2.32439995e-01
8.20034385e-01 2.83573329e-01 8.51077676e-01 9.01796877e-01
2.08017871e-01 -1.00666785e+00 -5.40905520e-02 7.80581117e-01
1.19739270e+00 -8.35474074e-01 5.24704568e-02 -5.48824549e-01
-7.54266202e-01 1.31702614e+00 4.85384703e-01 -2.55886912e-01
3.44433755e-01 -9.41415597e-03 -6.49217069e-02 -2.58097231e-01
-7.73869753e-01 3.79773378e-01 2.09973261e-01 9.93752927e-02
8.32453787e-01 1.45519152e-01 -1.19753611e+00 5.12673974e-01
-1.05235004e+00 -5.67434728e-01 1.24094713e+00 7.38647223e-01
-4.34706599e-01 -1.38623607e+00 -3.56429517e-01 5.84725082e-01
-6.87896252e-01 -5.52397251e-01 -7.15267539e-01 6.45998299e-01
1.54107413e-03 7.88823843e-01 1.82505861e-01 6.46514520e-02
3.56214643e-01 3.92951369e-01 6.29389584e-01 -1.09793007e+00
-9.16565239e-01 -7.03190267e-02 3.63794416e-01 -1.39823914e-01
-4.57448334e-01 -8.41749609e-01 -1.90892243e+00 -1.05214357e-01
-3.00224796e-02 6.24485314e-01 2.71476954e-01 1.14985394e+00
-5.83992414e-02 6.37413710e-02 1.14084058e-01 -3.26868355e-01
-8.02318394e-01 -7.47326910e-01 -3.92942220e-01 9.43029463e-01
-1.15844920e-01 -5.10755002e-01 -4.82056946e-01 1.14368103e-01] | [10.664488792419434, 9.162837982177734] |
0ffa05e1-2c11-4de8-ba81-2e9c09db5420 | edge-adaptive-l2-regularization-image | 1811.08487 | null | http://arxiv.org/abs/1811.08487v1 | http://arxiv.org/pdf/1811.08487v1.pdf | Edge-adaptive l2 regularization image reconstruction from non-uniform Fourier data | Total variation regularization based on the l1 norm is ubiquitous in image
reconstruction. However, the resulting reconstructions are not always as sparse
in the edge domain as desired. Iteratively reweighted methods provide some
improvement in accuracy, but at the cost of extended runtime. In this paper we
examine these methods for the case of data acquired as non-uniform Fourier
samples. We then develop a non-iterative weighted regularization method that
uses a pre-processing edge detection to find exactly where the sparsity should
be in the edge domain. We show that its performance in terms of both accuracy
and speed has the potential to outperform reweighted TV regularization methods. | [] | 2018-11-20 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [ 3.85241807e-01 -1.11881875e-01 8.42126533e-02 -1.93951920e-01
-9.18105423e-01 -1.16800003e-01 1.40779719e-01 -2.08970621e-01
-4.86164600e-01 6.54025137e-01 2.06951901e-01 -8.26216936e-02
-3.73725414e-01 -4.61008191e-01 -5.70239127e-01 -7.07127631e-01
-1.34649903e-01 6.51678741e-02 1.84217691e-01 1.91281457e-02
2.68183500e-01 3.91027242e-01 -1.01209211e+00 -1.67112879e-03
7.57732034e-01 8.43108058e-01 1.27838343e-01 5.38957834e-01
1.79848932e-02 6.68650389e-01 9.08585191e-02 2.73071349e-01
3.05052668e-01 -5.43091416e-01 -6.63217604e-01 7.44354606e-01
4.31767404e-01 -1.80273965e-01 -5.69873899e-02 1.10205138e+00
4.37661290e-01 4.17875409e-01 4.54320520e-01 -5.19384563e-01
-2.62910873e-01 -4.37729470e-02 -1.07007742e+00 4.04817402e-01
3.06971341e-01 -3.51050496e-01 7.16166317e-01 -1.35407484e+00
8.11783075e-01 7.35641956e-01 1.24669838e+00 2.16608107e-01
-1.56605172e+00 -2.36226823e-02 -1.14529625e-01 -3.93056497e-02
-1.27402198e+00 -6.65069818e-01 9.79937136e-01 -4.67778414e-01
7.22206354e-01 2.31748536e-01 3.85398000e-01 4.21435207e-01
2.06881702e-01 5.91029525e-01 1.07327688e+00 -7.94730783e-01
1.93134114e-01 -1.33641303e-01 2.60168403e-01 6.89082086e-01
1.32011354e-01 1.06809057e-01 -5.80408514e-01 -3.01850408e-01
9.62438285e-01 -2.81901538e-01 -6.59580052e-01 -5.79269469e-01
-1.01752687e+00 8.83582413e-01 1.25112414e-01 6.28469586e-01
-6.00204706e-01 3.55652064e-01 4.50856268e-01 3.04717928e-01
1.02927637e+00 1.15589172e-01 -2.62716681e-01 1.05211064e-01
-1.52337992e+00 5.62479757e-02 5.23943365e-01 4.63658512e-01
7.87615478e-01 2.98971742e-01 1.33311421e-01 9.68817055e-01
4.07093287e-01 1.82408080e-01 2.51045972e-01 -1.19294977e+00
-7.97270685e-02 -2.54875943e-02 1.88097209e-01 -1.23032248e+00
-4.26556259e-01 -3.97448659e-01 -7.78366566e-01 5.17355323e-01
4.07361060e-01 -1.12701751e-01 -8.18446875e-01 1.50880373e+00
5.03842831e-01 5.99216342e-01 -1.69317916e-01 1.11184669e+00
5.12213588e-01 5.10584176e-01 -3.00107211e-01 -6.79699779e-01
9.46157098e-01 -8.47138166e-01 -1.02875865e+00 -2.20665976e-01
5.74194372e-01 -1.11092806e+00 6.93219244e-01 3.86357188e-01
-1.36072171e+00 -3.59794736e-01 -8.97451639e-01 -5.93222193e-02
3.84869218e-01 -5.46679199e-02 2.83094198e-01 4.07949299e-01
-1.10288060e+00 6.98190212e-01 -9.43497181e-01 -1.29623204e-01
5.66631481e-02 1.01454906e-01 -4.00116861e-01 -2.56011695e-01
-6.13344967e-01 9.23676193e-01 -1.78854644e-01 4.84582447e-02
-2.69398302e-01 -1.01639235e+00 -9.84874547e-01 -1.84097141e-01
2.68577635e-01 -3.12594593e-01 9.17843878e-01 -1.28813875e+00
-1.21988237e+00 7.76205301e-01 -5.04505336e-01 -1.87590092e-01
6.75469458e-01 -5.36480211e-02 -1.33375466e-01 2.99510717e-01
2.31798682e-02 -7.21226111e-02 1.20385408e+00 -1.29668677e+00
-2.86608547e-01 -3.34781826e-01 -5.39800525e-01 1.32177740e-01
-1.35653047e-02 -5.73212020e-02 -4.43644911e-01 -9.42575097e-01
6.18899882e-01 -9.74819720e-01 -6.01980627e-01 3.62750709e-01
6.64525758e-03 2.37420425e-01 8.28033984e-01 -9.24094379e-01
1.03871548e+00 -2.21240282e+00 -3.54094468e-02 6.62628174e-01
2.95449644e-01 6.11331612e-02 3.30983661e-02 1.65089652e-01
-2.93281764e-01 -2.66452223e-01 -6.87485337e-01 -2.95895159e-01
-6.91137791e-01 1.47263452e-01 1.26567155e-01 1.09502780e+00
-1.22743547e-01 4.92167592e-01 -9.32677388e-01 -5.12856841e-01
3.83834571e-01 8.75611126e-01 -6.28679991e-01 -1.18382916e-01
3.67321849e-01 6.18425667e-01 -2.79981852e-01 1.32647172e-01
8.16467702e-01 -1.57342121e-01 2.24883109e-01 -3.41344208e-01
-3.97095889e-01 -2.36061022e-01 -1.40361679e+00 1.69282925e+00
-4.34627265e-01 1.10007799e+00 8.96316946e-01 -1.44944906e+00
6.44873798e-01 5.10495722e-01 1.12955248e+00 -4.14904326e-01
-2.26903930e-02 3.68151337e-01 -2.52234131e-01 -6.28193080e-01
3.16071093e-01 -6.28137887e-01 5.52056134e-01 4.22457963e-01
-2.14749649e-01 -2.02388734e-01 1.79880142e-01 8.44555646e-02
1.01779902e+00 4.60812710e-02 3.40296149e-01 -9.50571299e-01
4.94878441e-01 1.51335850e-01 5.42593181e-01 5.14878988e-01
-9.53684002e-02 8.53224635e-01 5.20528331e-02 -3.98111969e-01
-1.11950099e+00 -6.69986904e-01 -6.34792089e-01 6.32513940e-01
5.51827736e-02 -1.55934513e-01 -5.60319543e-01 -2.70399600e-01
-2.00485557e-01 6.11614943e-01 -4.98801470e-01 1.77225292e-01
-6.92169785e-01 -7.47608244e-01 -1.25944898e-01 2.90608436e-01
3.27336699e-01 -6.32000744e-01 -6.24421597e-01 4.16648090e-01
-2.25269631e-01 -1.07704747e+00 -8.34304810e-01 9.95505527e-02
-1.20394802e+00 -1.14735341e+00 -1.09777474e+00 -8.29315901e-01
1.00217259e+00 5.04581690e-01 1.04784477e+00 2.89339095e-01
-3.72091681e-01 8.92705381e-01 -1.60126224e-01 1.18576922e-01
1.16857402e-02 -6.52938902e-01 -1.25297248e-01 2.13250488e-01
-7.08513707e-02 -5.85979521e-01 -6.07762814e-01 2.92370677e-01
-9.09536302e-01 -2.12172285e-01 -4.83272560e-02 1.05527341e+00
9.02246654e-01 7.02332482e-02 3.41862142e-01 -1.03387010e+00
6.58060372e-01 -3.22847217e-01 -4.47206110e-01 9.14455131e-02
-8.15250337e-01 -1.17996288e-02 2.69536763e-01 -4.01492089e-01
-1.15706503e+00 2.98622727e-01 -1.58190042e-01 -3.81584823e-01
4.01953757e-01 8.11625540e-01 6.83338702e-01 -6.82515979e-01
8.01937997e-01 1.22099876e-01 2.39532739e-01 -5.05135596e-01
1.27988234e-01 6.79772422e-02 1.67983487e-01 -4.79935616e-01
3.92896324e-01 6.75804496e-01 3.13302159e-01 -1.33416390e+00
-6.60249650e-01 -7.34123945e-01 -4.43165660e-01 -5.26481152e-01
5.68439841e-01 -5.26713014e-01 -1.63424820e-01 1.98977783e-01
-8.51013422e-01 -3.48745823e-01 -6.30319357e-01 7.42206872e-01
-7.56859064e-01 8.84472907e-01 -6.67112291e-01 -6.93571806e-01
-1.58082962e-01 -9.80016410e-01 5.73851526e-01 -2.21898377e-01
-1.21202946e-01 -1.31102467e+00 1.95230886e-01 3.71879227e-02
6.50931001e-01 2.68809676e-01 4.74430889e-01 -2.66083125e-02
-9.24377590e-02 -3.88652176e-01 -2.29044393e-01 4.20928299e-01
9.66508016e-02 -5.72227538e-02 -4.98700172e-01 -3.86834711e-01
6.30972862e-01 -1.11268247e-02 8.77832174e-01 1.10992348e+00
8.97138655e-01 4.83289594e-03 -1.64224476e-01 7.16520786e-01
1.82835448e+00 -1.17350608e-01 6.85788631e-01 3.25206444e-02
4.64952976e-01 5.28710246e-01 3.58470172e-01 3.26983660e-01
-1.10186443e-01 6.41471505e-01 5.15936948e-02 -3.22229147e-01
-3.63736063e-01 4.15367663e-01 -3.76053937e-02 8.12809050e-01
-1.89374804e-01 1.00804858e-01 -8.31848621e-01 8.60153794e-01
-1.92896485e+00 -9.03863668e-01 -6.61579311e-01 2.18264151e+00
6.54436529e-01 -1.64106339e-01 -2.76302267e-02 3.57752711e-01
7.77667701e-01 7.37957284e-02 -3.48904431e-01 -3.64681602e-01
6.00624382e-02 2.20688164e-01 7.50974476e-01 1.06127036e+00
-8.66439760e-01 3.57181191e-01 7.59401512e+00 7.27364779e-01
-9.34287548e-01 2.54949123e-01 4.20671523e-01 -2.32350603e-02
-2.11456358e-01 3.38763669e-02 -1.16464414e-01 2.90330976e-01
4.58389729e-01 -9.43421125e-02 4.84825194e-01 5.56675315e-01
4.30166751e-01 -3.93803507e-01 -5.91190159e-01 9.79090571e-01
1.32473722e-01 -1.17469811e+00 -5.78317821e-01 7.06957430e-02
9.24404502e-01 -6.87409043e-02 -1.87647834e-01 -3.35338116e-01
-1.85319409e-01 -8.24058175e-01 3.26494843e-01 6.03736460e-01
7.10803092e-01 -5.84701717e-01 4.39086378e-01 3.39441478e-01
-1.05535889e+00 3.70837271e-01 -3.45966101e-01 -4.09440324e-02
5.91502309e-01 1.23338842e+00 -3.04093122e-01 3.16500336e-01
5.15905499e-01 7.66367018e-01 2.62788743e-01 1.14961994e+00
1.21982552e-01 5.98603845e-01 -5.98579466e-01 5.59697568e-01
2.37256452e-01 -5.86821258e-01 7.76113331e-01 1.22024226e+00
5.25293410e-01 5.02179265e-01 3.74612778e-01 4.39991236e-01
2.43362933e-01 2.51641154e-01 -6.49830520e-01 4.97173429e-01
-8.92008245e-02 1.11903918e+00 -8.87220860e-01 -1.74633786e-01
-8.25686753e-01 9.38141823e-01 1.67509064e-01 4.96206611e-01
-3.65802288e-01 -9.60838050e-02 3.67770165e-01 5.18407226e-01
2.96248794e-01 -4.74529028e-01 -3.41427952e-01 -1.08257020e+00
-1.16465896e-01 -5.74320972e-01 3.29517663e-01 -5.62180340e-01
-1.33285248e+00 3.91585648e-01 -1.51147857e-01 -1.04629159e+00
-9.61379856e-02 -5.81250370e-01 -5.19873381e-01 8.27994108e-01
-1.35747230e+00 -7.01870203e-01 -1.81107149e-01 6.04300559e-01
6.93633080e-01 4.58349615e-01 5.36323786e-01 4.88663405e-01
-2.65190899e-01 1.22827269e-01 3.74074697e-01 -2.38800123e-02
5.52350819e-01 -9.74125564e-01 -1.92727968e-01 8.97161305e-01
-3.73122022e-02 3.05640757e-01 9.82317984e-01 -6.55662656e-01
-1.23843050e+00 -5.69824815e-01 7.51253545e-01 1.79695338e-01
4.82846349e-01 3.65925223e-01 -1.11926079e+00 8.61535311e-01
1.35560125e-01 6.09968126e-01 4.80706245e-01 3.12346052e-02
2.53231347e-01 1.85547873e-01 -1.42569649e+00 2.74470121e-01
7.49570429e-01 -2.45473593e-01 -3.52369577e-01 3.87064576e-01
-8.86442587e-02 -3.54612619e-01 -8.34918678e-01 5.11382401e-01
3.56348097e-01 -9.82406676e-01 1.16882420e+00 -1.09581009e-01
2.32274160e-01 -1.49265245e-01 -1.92162469e-01 -1.25072169e+00
-5.74008763e-01 -5.42494655e-01 1.29884899e-01 7.48041511e-01
2.48612866e-01 -5.21867037e-01 8.55121136e-01 7.08591938e-01
-1.16524138e-01 -6.94455802e-01 -1.19090223e+00 -7.44395316e-01
-2.96559602e-01 -6.58209562e-01 -4.91385728e-01 1.07662821e+00
1.27217025e-01 1.36321560e-01 -6.49798572e-01 -1.44654781e-01
1.19126546e+00 -5.11384159e-02 1.90583542e-01 -9.66955543e-01
-8.38874094e-03 -2.79428840e-01 -4.17893797e-01 -1.08432496e+00
-3.72571349e-02 -7.82613337e-01 4.19726342e-01 -1.47585440e+00
-2.34649479e-02 -4.54668611e-01 -1.69279307e-01 2.14173831e-02
-9.40006971e-02 5.40804267e-01 -2.14455351e-01 2.61986554e-01
-1.07929394e-01 1.99216112e-01 1.15736234e+00 1.56769261e-01
-2.25221992e-01 -8.84099975e-02 -1.21085107e-01 9.97734308e-01
6.55001104e-01 -4.55371827e-01 -3.63182306e-01 -6.02537990e-01
6.34304285e-02 2.79784381e-01 1.14924356e-01 -9.11872625e-01
2.66503841e-01 1.60341207e-02 2.46194839e-01 -3.40693980e-01
4.26533967e-01 -1.09622526e+00 2.32223913e-01 3.12589556e-01
-3.17011923e-01 -1.28006682e-01 6.71308562e-02 6.34291768e-01
-1.62694439e-01 -7.85991549e-01 1.19975615e+00 -2.31116146e-01
-5.18916845e-01 -1.74341686e-02 -5.92389286e-01 6.43684193e-02
8.99283051e-01 -4.31482702e-01 5.40436149e-01 -5.54594994e-01
-1.02995932e+00 -5.09968400e-02 4.64508176e-01 -3.97523642e-01
7.55970597e-01 -1.15519071e+00 -8.31242800e-01 3.34143609e-01
-3.86299372e-01 -2.75630474e-01 4.20162469e-01 1.27037168e+00
-7.56743431e-01 -1.45888934e-02 9.27551240e-02 -7.10459530e-01
-1.36472619e+00 6.30320907e-01 5.16989589e-01 -2.76129633e-01
-1.06727695e+00 8.41010809e-01 -1.03569530e-01 -8.78300220e-02
8.55531357e-03 1.12969324e-01 1.83754668e-01 2.29347367e-02
5.74648798e-01 7.08853006e-01 1.81727841e-01 -7.51828134e-01
-3.83974493e-01 1.02553213e+00 3.30680668e-01 -3.41803610e-01
1.50582528e+00 -3.45691800e-01 -3.31635356e-01 3.21601927e-01
1.32749331e+00 3.27266932e-01 -1.38313627e+00 -4.14625436e-01
1.76000923e-01 -7.55631983e-01 8.25992286e-01 -2.59140342e-01
-1.49684298e+00 5.03640234e-01 5.67006886e-01 2.55252182e-01
1.25214708e+00 -2.92198598e-01 8.34496617e-01 -2.20289566e-02
1.18924618e-01 -1.28490233e+00 -1.54363245e-01 2.29411662e-01
8.26075375e-01 -1.23424804e+00 5.20091951e-01 -7.13925779e-01
-3.01936984e-01 1.17393279e+00 1.42484559e-02 -8.55936348e-01
1.25757420e+00 3.80731195e-01 7.05248639e-02 -3.76546949e-01
-2.38516271e-01 -1.00137092e-01 4.67139602e-01 4.40516204e-01
8.26694787e-01 -2.32753247e-01 -9.06482637e-01 -1.28278032e-01
6.12341166e-01 1.81830272e-01 3.21810365e-01 1.12835252e+00
-4.97484952e-01 -8.24497879e-01 -6.08084440e-01 5.10104716e-01
-6.62166536e-01 -1.84887368e-02 2.03653604e-01 5.43359518e-01
-2.36889869e-01 1.05098641e+00 -1.62344620e-01 2.41982475e-01
3.46680105e-01 -5.34123406e-02 7.82676876e-01 -2.62990922e-01
-3.75510573e-01 6.33152962e-01 1.49581358e-01 -5.18705249e-01
-7.25311279e-01 -6.90986574e-01 -1.21248543e+00 -3.72428417e-01
-4.75666165e-01 2.20452964e-01 6.25119269e-01 8.35460961e-01
4.71865945e-02 4.46735501e-01 6.21351421e-01 -1.12894845e+00
-2.28973135e-01 -6.23293757e-01 -7.79083490e-01 3.82462144e-01
5.50962269e-01 -4.84549850e-01 -5.60267389e-01 1.41146287e-01] | [11.704784393310547, -2.432959794998169] |
6bf1d825-b6f1-46ca-b2c5-8f9233a469d0 | detect-distill-and-update-learned-db-systems | 2210.05508 | null | https://arxiv.org/abs/2210.05508v2 | https://arxiv.org/pdf/2210.05508v2.pdf | Detect, Distill and Update: Learned DB Systems Facing Out of Distribution Data | Machine Learning (ML) is changing DBs as many DB components are being replaced by ML models. One open problem in this setting is how to update such ML models in the presence of data updates. We start this investigation focusing on data insertions (dominating updates in analytical DBs). We study how to update neural network (NN) models when new data follows a different distribution (a.k.a. it is "out-of-distribution" -- OOD), rendering previously-trained NNs inaccurate. A requirement in our problem setting is that learned DB components should ensure high accuracy for tasks on old and new data (e.g., for approximate query processing (AQP), cardinality estimation (CE), synthetic data generation (DG), etc.). This paper proposes a novel updatability framework (DDUp). DDUp can provide updatability for different learned DB system components, even based on different NNs, without the high costs to retrain the NNs from scratch. DDUp entails two components: First, a novel, efficient, and principled statistical-testing approach to detect OOD data. Second, a novel model updating approach, grounded on the principles of transfer learning with knowledge distillation, to update learned models efficiently, while still ensuring high accuracy. We develop and showcase DDUp's applicability for three different learned DB components, AQP, CE, and DG, each employing a different type of NN. Detailed experimental evaluation using real and benchmark datasets for AQP, CE, and DG detail DDUp's performance advantages. | ['Peter Triantafillou', 'Meghdad Kurmanji'] | 2022-10-11 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [-6.96788877e-02 1.59071520e-01 -2.11553842e-01 -3.35997641e-01
-7.67546654e-01 -4.53589827e-01 3.30573231e-01 3.59479338e-01
-4.62131798e-01 1.15984845e+00 -2.81702757e-01 -3.38962078e-01
-2.38168478e-01 -1.16755366e+00 -1.53633666e+00 -4.70715463e-01
-1.71103835e-01 1.22685742e+00 5.16447425e-01 5.29220887e-02
1.44793674e-01 8.92268479e-01 -1.80560529e+00 4.04866874e-01
8.41738105e-01 1.36204362e+00 9.55754742e-02 6.98550105e-01
-4.17800903e-01 5.67380071e-01 -8.83754551e-01 -6.16238892e-01
3.97628367e-01 1.32511467e-01 -5.82145870e-01 -3.45686287e-01
4.07882780e-01 -6.91477776e-01 -2.13218123e-01 8.02104950e-01
5.90178490e-01 -7.24265799e-02 6.26658440e-01 -1.26627779e+00
-5.62592626e-01 5.53788066e-01 -1.85479850e-01 1.10917076e-01
4.76076119e-02 8.64717588e-02 5.98042965e-01 -6.76450610e-01
6.01103961e-01 1.08308566e+00 8.52024853e-01 5.24176121e-01
-1.53439605e+00 -6.26286685e-01 3.28534320e-02 2.27748543e-01
-1.31025600e+00 -3.45564216e-01 5.40523171e-01 -1.90204188e-01
1.01643550e+00 4.42431629e-01 5.94050288e-01 1.02007961e+00
1.96296081e-01 9.51493442e-01 7.00416446e-01 -5.08026719e-01
9.24126446e-01 5.67190528e-01 1.05628751e-01 2.13885784e-01
6.31113827e-01 8.60218778e-02 -7.21585095e-01 -5.02849877e-01
4.93175834e-01 -2.49807745e-01 -9.65578780e-02 -7.51795769e-01
-5.33533156e-01 6.79692328e-01 2.37588100e-02 1.55168340e-01
-4.18872505e-01 2.14779407e-01 4.87913698e-01 5.42929888e-01
4.87411261e-01 4.10242260e-01 -9.04158294e-01 -3.49316001e-01
-8.28195393e-01 5.54556787e-01 1.09363520e+00 1.26561987e+00
9.72994149e-01 3.47152120e-03 -2.95777351e-01 9.09516335e-01
-1.39643341e-01 5.29070258e-01 5.69879830e-01 -7.10886657e-01
5.25351405e-01 4.61193085e-01 2.75765002e-01 -7.49282122e-01
-2.67892540e-01 -4.02092040e-01 -9.26339030e-01 -9.02010426e-02
2.25728955e-02 1.21746793e-01 -7.07326114e-01 1.77206302e+00
5.75970292e-01 5.43135628e-02 -1.35341555e-01 6.37961850e-02
2.43793920e-01 7.75635540e-01 -2.92683035e-01 -4.55783278e-01
5.93151569e-01 -3.66728187e-01 -5.10459781e-01 2.05932725e-02
6.55518889e-01 -2.00981334e-01 1.28332317e+00 1.04429853e+00
-1.35787201e+00 -6.31329477e-01 -1.01611733e+00 1.59326777e-01
-5.77132463e-01 -9.87979546e-02 5.02873063e-01 6.21671855e-01
-1.31032538e+00 5.61025918e-01 -8.12060654e-01 -3.14071596e-01
4.48472261e-01 6.16745114e-01 -2.35166643e-02 -1.32055253e-01
-1.12538505e+00 8.50376904e-01 7.67282605e-01 2.48707328e-02
-6.55727208e-01 -1.05621111e+00 -5.31595588e-01 -5.33151142e-02
3.76626641e-01 -7.34100044e-01 1.50005627e+00 -5.79807580e-01
-1.55829084e+00 4.83849436e-01 -1.22062780e-01 -8.83443654e-01
8.34603429e-01 -3.18209201e-01 -4.70520586e-01 -2.45461151e-01
-1.58533305e-01 3.24692249e-01 7.22530723e-01 -1.47491682e+00
-7.59464741e-01 -5.48488200e-01 -2.67713279e-01 -3.58605802e-01
-6.02141798e-01 -5.23857951e-01 -6.31556630e-01 -6.59112096e-01
1.05279326e-01 -6.80281699e-01 1.83863923e-01 2.59652324e-02
-3.92116487e-01 -2.97430336e-01 6.27534032e-01 -6.46249652e-01
1.58983254e+00 -1.91341305e+00 -6.70829182e-03 6.89083695e-01
4.98455986e-02 2.89012253e-01 7.60773718e-02 4.77906317e-01
2.56843656e-01 -5.88039272e-02 -4.61919785e-01 -3.88405561e-01
2.63745248e-01 6.32030666e-01 -5.24343610e-01 6.27582520e-02
9.20465309e-03 7.26960361e-01 -6.49655223e-01 -2.78250217e-01
-9.09270495e-02 2.49642897e-02 -7.65036702e-01 4.08855259e-01
-7.07775295e-01 -7.26928264e-02 -8.16793218e-02 7.81316280e-01
8.86768818e-01 4.30289432e-02 1.21651411e-01 -3.94166857e-01
8.14141780e-02 1.22632016e-03 -1.57214463e+00 1.36033154e+00
-5.21529198e-01 3.07659805e-01 -1.10185765e-01 -1.19008946e+00
1.03439415e+00 -3.32701385e-01 4.08985138e-01 -9.61772561e-01
-2.76155800e-01 4.71070826e-01 -3.50987762e-01 -4.68104273e-01
6.33635759e-01 9.83238965e-02 4.22773771e-02 4.27685440e-01
9.49217305e-02 -5.14522307e-02 3.50231588e-01 1.85122900e-03
1.22941852e+00 -6.66421577e-02 1.21343657e-01 -9.37255397e-02
1.99622080e-01 -2.33552620e-01 5.70341766e-01 1.09038734e+00
2.04863742e-01 2.06429645e-01 5.08889675e-01 -4.77266014e-01
-1.05999863e+00 -1.37647760e+00 -4.68315989e-01 1.01295137e+00
-1.23779342e-01 -1.76823348e-01 -6.47365868e-01 -4.27034467e-01
6.60205722e-01 1.17986226e+00 -5.34073293e-01 -4.36000824e-01
-6.44209504e-01 -8.50601971e-01 6.94027841e-01 5.19604623e-01
3.80631685e-01 -8.74657273e-01 -5.36124229e-01 5.19812942e-01
2.00264126e-01 -7.73760796e-01 -9.39899217e-03 5.24351776e-01
-1.17993927e+00 -8.91513467e-01 -3.54763091e-01 -2.52899408e-01
3.82144272e-01 -4.07209694e-01 1.18706930e+00 -3.50356013e-01
-1.71970323e-01 6.05858624e-01 -6.61061853e-02 -6.50667429e-01
-5.97233474e-01 2.18306452e-01 3.99554819e-01 -5.49542978e-02
4.34966475e-01 -7.99540997e-01 -1.03263900e-01 1.60829991e-01
-1.40049791e+00 -2.60085911e-01 8.14542949e-01 7.62174308e-01
8.44742656e-01 1.19217776e-01 6.73797190e-01 -1.10841584e+00
8.34139228e-01 -5.55437684e-01 -8.02460790e-01 6.16965115e-01
-1.21130908e+00 4.56302077e-01 5.52310169e-01 -6.21098697e-01
-9.09791112e-01 -4.16341782e-01 -1.01556197e-01 -4.98735726e-01
-1.25837415e-01 4.78942305e-01 -4.14037853e-01 2.21549273e-01
8.96466613e-01 4.63242859e-01 5.39307855e-03 -6.18102670e-01
2.30571181e-01 8.90491128e-01 6.29781425e-01 -1.07060981e+00
5.85190773e-01 3.60556245e-01 -1.89990401e-01 -6.18148685e-01
-3.25029254e-01 -1.82756037e-01 -5.09211898e-01 6.28186297e-03
8.02910775e-02 -4.78731930e-01 -6.00517869e-01 7.11174071e-01
-1.19160068e+00 -5.84584057e-01 -7.06492722e-01 1.50740877e-01
-6.75188780e-01 2.48125285e-01 -5.72792888e-01 -8.30397964e-01
-3.49206924e-01 -7.33442843e-01 8.45079064e-01 -1.94182977e-01
-3.58402021e-02 -7.59681106e-01 3.27670991e-01 -1.72903568e-01
5.50336719e-01 1.82627723e-01 1.34724975e+00 -9.08379376e-01
-5.91004968e-01 -3.18299383e-01 -8.64111334e-02 5.79061389e-01
-2.88904049e-02 -2.72252709e-02 -8.91798198e-01 -4.64684635e-01
-7.34724570e-03 -2.91597545e-01 5.94899297e-01 4.68599617e-01
1.55366540e+00 -6.38368368e-01 -3.50010335e-01 5.38705111e-01
1.36887002e+00 2.73433089e-01 7.52540469e-01 4.69866753e-01
1.81580409e-01 2.48333216e-01 5.00864148e-01 8.10815632e-01
3.26475471e-01 7.49038458e-01 3.76841664e-01 3.26802164e-01
1.67123437e-01 -2.67132670e-01 4.68827158e-01 6.73378825e-01
3.71931463e-01 -3.24420214e-01 -9.43062961e-01 3.65987331e-01
-1.86920285e+00 -5.06223619e-01 2.03390270e-01 2.64768481e+00
1.33873904e+00 5.75384140e-01 1.63157821e-01 1.52840927e-01
5.72306991e-01 -4.68375117e-01 -1.16005611e+00 -6.18482411e-01
-1.11445427e-01 4.92886722e-01 5.21474540e-01 2.29123637e-01
-6.07811809e-01 5.27441621e-01 5.46652079e+00 8.97428155e-01
-9.97544467e-01 2.80649010e-02 6.84490204e-01 -2.91017979e-01
-3.90639246e-01 -3.94957870e-01 -1.18505859e+00 7.68672824e-01
1.19811618e+00 -3.78272543e-03 5.47147155e-01 1.13626480e+00
-1.04917385e-01 -3.56692463e-01 -1.61985743e+00 1.07354522e+00
9.37285796e-02 -1.58994973e+00 5.23788571e-01 -5.74482605e-03
5.90376019e-01 2.49270890e-02 -1.39335439e-01 8.04153621e-01
2.94254333e-01 -5.60162902e-01 7.74296165e-01 9.55411732e-01
8.23244274e-01 -8.32433999e-01 8.48707020e-01 4.96646017e-01
-6.68064713e-01 -1.44789815e-01 -4.89676565e-01 3.42434466e-01
-1.35485560e-01 1.15509689e+00 -9.03172135e-01 5.47159612e-01
1.01162779e+00 2.15064496e-01 -6.55735612e-01 1.01272368e+00
4.51201260e-01 4.17300224e-01 -7.30620503e-01 -4.91812751e-02
-3.02817017e-01 4.06259904e-03 4.10581112e-01 1.12209404e+00
4.88394678e-01 -3.78230244e-01 -2.46283382e-01 1.10513747e+00
-1.69283226e-01 -1.16608873e-01 -4.38389689e-01 3.20970267e-01
8.88718426e-01 5.77517211e-01 -1.34449646e-01 -4.37679380e-01
9.82313305e-02 8.19001496e-01 5.05437672e-01 3.66685152e-01
-8.18852246e-01 -2.61475384e-01 3.65743339e-01 4.75432783e-01
3.39907259e-01 1.42867535e-01 -2.21311003e-01 -8.16420913e-01
4.89371926e-01 -8.54111671e-01 5.55295348e-01 -5.87174237e-01
-1.65111136e+00 3.07424635e-01 3.65948856e-01 -9.92250800e-01
-5.70927501e-01 -5.55970848e-01 -1.36316329e-01 5.33734500e-01
-1.34928954e+00 -8.43390107e-01 -3.40281725e-01 5.56233048e-01
2.67362863e-01 -1.71102002e-01 6.44342959e-01 5.29561460e-01
-3.82633358e-01 9.61329818e-01 6.28747344e-01 -3.56128305e-01
7.96442151e-01 -1.27359188e+00 5.46542630e-02 4.51194465e-01
2.84523070e-02 3.62373441e-01 5.97976208e-01 -7.27774441e-01
-1.42976701e+00 -1.31850684e+00 4.67923731e-01 -4.21349943e-01
4.65155542e-01 -4.97968912e-01 -1.15907311e+00 6.61470771e-01
-5.43291390e-01 -8.58263373e-02 5.05673051e-01 2.03455284e-01
-2.73393214e-01 -8.50501716e-01 -1.41757178e+00 2.59353191e-01
9.76654053e-01 -3.71378154e-01 -4.08980757e-01 2.68918842e-01
7.74849534e-01 -4.89727020e-01 -8.92343998e-01 7.09961712e-01
4.58694309e-01 -1.12097669e+00 9.48597789e-01 -7.26451755e-01
-7.51260817e-02 -1.34187922e-01 -4.36106682e-01 -1.05785787e+00
1.78994507e-01 -4.34666246e-01 -1.20366704e+00 1.35047758e+00
2.95191944e-01 -8.40101898e-01 7.64300346e-01 8.45877528e-01
-2.34030530e-01 -9.50764835e-01 -1.24669909e+00 -1.24422359e+00
8.37721452e-02 -8.13212037e-01 9.39382553e-01 6.49299443e-01
-4.96531099e-01 -1.23372838e-01 -2.02714592e-01 1.76861614e-01
4.32085425e-01 -1.41748115e-01 1.12951493e+00 -1.36495841e+00
-6.77078605e-01 -3.23160797e-01 -2.63276994e-01 -1.04208100e+00
-1.64501503e-01 -6.22262299e-01 -1.64266601e-01 -1.24906516e+00
-6.76980242e-02 -7.69733369e-01 -2.32274011e-01 2.14787096e-01
3.38330954e-01 -3.48985046e-01 -2.46409029e-01 3.45290959e-01
-3.89658481e-01 6.52944922e-01 4.89975452e-01 -1.70054197e-01
-6.32870555e-01 3.20618063e-01 -2.82697827e-01 2.96437234e-01
7.73554027e-01 -4.32749033e-01 -3.73220623e-01 -3.57200086e-01
5.01503766e-01 -2.85583049e-01 4.24699962e-01 -1.21627820e+00
4.64321762e-01 4.98556942e-02 4.99936730e-01 -9.32188511e-01
1.54011294e-01 -8.23242784e-01 4.37260330e-01 5.72703004e-01
-2.03266114e-01 2.05113322e-01 5.69275200e-01 7.76532054e-01
-1.03353776e-01 -3.02686036e-01 6.38803422e-01 7.58429244e-02
-5.24692595e-01 2.72920907e-01 -5.22407666e-02 -1.19867707e-02
1.05622649e+00 -3.78852665e-01 -1.78499311e-01 -2.01783344e-01
-7.14867711e-01 1.02378331e-01 1.87818214e-01 1.26635462e-01
5.08558512e-01 -1.17557740e+00 -4.13903326e-01 3.79289329e-01
2.45207742e-01 4.83521461e-01 9.30276364e-02 4.78274584e-01
-5.14897704e-01 2.94607222e-01 1.32934794e-01 -6.97678149e-01
-6.30436540e-01 5.88747025e-01 3.97467315e-01 -5.25996804e-01
-1.79599971e-01 6.97056055e-01 -4.01018709e-01 -8.73474300e-01
7.86710739e-01 -8.55729878e-01 5.32650352e-01 2.32425444e-02
3.01436245e-01 5.95625281e-01 6.61637366e-01 4.27532852e-01
9.38757658e-02 -5.81018031e-02 -3.18048805e-01 1.38262004e-01
1.48374581e+00 7.88404569e-02 -2.52681702e-01 8.69075954e-01
8.87336373e-01 -1.75917536e-01 -9.89528358e-01 -4.51842308e-01
2.84841597e-01 -3.16878617e-01 -2.59909600e-01 -9.40741360e-01
-6.37461066e-01 6.42426372e-01 9.53494251e-01 3.09273064e-01
1.13084888e+00 1.10631242e-01 7.23502457e-01 8.78469825e-01
7.42934465e-01 -1.62050760e+00 1.18485438e-02 2.23586082e-01
1.01013124e+00 -8.42719078e-01 -5.23070917e-02 1.95361793e-01
-3.00242342e-02 1.14942110e+00 9.49057221e-01 3.23469728e-01
7.48695552e-01 3.57046723e-01 -5.73304474e-01 8.00968558e-02
-7.34623611e-01 3.69209349e-01 -1.40339285e-02 7.02272117e-01
-3.50040376e-01 -1.06003083e-01 1.23875633e-01 7.42336810e-01
-1.91664428e-01 3.67467105e-01 3.43275636e-01 9.88816082e-01
-4.20470655e-01 -1.17585552e+00 -3.47407013e-01 1.03637958e+00
1.75276250e-01 -3.17595266e-02 -1.70250177e-01 1.06625628e+00
3.20891500e-01 3.00684303e-01 4.53017324e-01 -3.42355043e-01
5.64707160e-01 7.89213926e-02 4.40402359e-01 -4.14985657e-01
-1.76544756e-01 -2.85707772e-01 -4.24939282e-02 -5.65530002e-01
1.37901809e-02 -6.48890138e-01 -1.03357410e+00 -5.05317390e-01
-2.85284370e-01 6.04727753e-02 7.58412004e-01 7.75714397e-01
6.64588869e-01 2.76231617e-01 4.51368958e-01 -2.70425469e-01
-1.14396513e+00 -9.81560349e-01 -7.33880162e-01 9.56042036e-02
1.72079757e-01 -7.09547162e-01 -3.39932889e-01 -1.07456788e-01] | [9.42155933380127, 3.441514492034912] |
4118d9b0-90d6-4897-bda3-226be2498071 | hico-det-sg-and-v-coco-sg-new-data-splits-to | 2305.09948 | null | https://arxiv.org/abs/2305.09948v4 | https://arxiv.org/pdf/2305.09948v4.pdf | HICO-DET-SG and V-COCO-SG: New Data Splits for Evaluating the Systematic Generalization Performance of Human-Object Interaction Detection Models | Human-Object Interaction (HOI) detection is a task to localize humans and objects in an image and predict the interactions in human-object pairs. In real-world scenarios, HOI detection models are required systematic generalization, i.e., generalization to novel combinations of objects and interactions, because the train data are expected to cover a limited portion of all possible combinations. However, to our knowledge, no open benchmarks or previous work exist for evaluating the systematic generalization performance of HOI detection models. To address this issue, we created two new sets of HOI detection data splits named HICO-DET-SG and V-COCO-SG based on the HICO-DET and V-COCO datasets, respectively. When evaluated on the new data splits, the representative HOI detection models performed much more poorly than when evaluated on the original splits. This reveals that systematic generalization is a challenging goal in HOI detection. By analyzing the evaluation results, we also gain insights for improving the systematic generalization performance and identify four possible future research directions. We hope that our new data splits and presented analysis will encourage further research on systematic generalization in HOI detection. | ['Hisanao Akima', 'Tomotake Sasaki', 'Moyuru Yamada', 'Kentaro Takemoto'] | 2023-05-17 | null | null | null | null | ['human-object-interaction-detection', 'systematic-generalization'] | ['computer-vision', 'reasoning'] | [-1.53832883e-02 -1.01800948e-01 -4.77649644e-02 -3.24488401e-01
-1.16343997e-01 -4.09057885e-01 4.65854853e-01 -1.21529557e-01
-9.73267481e-02 3.54803115e-01 -1.14272386e-01 7.31904656e-02
-1.23221569e-01 -4.39019412e-01 -5.65863371e-01 -5.00815749e-01
-2.94441015e-01 6.06782377e-01 6.50815070e-01 3.96596175e-03
-1.05286755e-01 5.88153481e-01 -2.16119623e+00 5.35109401e-01
7.71505833e-01 1.00918615e+00 4.70164865e-02 6.18037581e-01
7.76609480e-01 4.90352213e-01 -6.87974334e-01 -2.33037576e-01
4.99174774e-01 -4.81290191e-01 -6.48458123e-01 2.74518251e-01
6.79257631e-01 -2.00796232e-01 -3.43518496e-01 7.60244071e-01
5.95227599e-01 1.65809795e-01 6.24227941e-01 -1.88951957e+00
-6.31220698e-01 3.88293922e-01 -6.04350269e-01 3.68713379e-01
6.53971195e-01 4.91403043e-01 8.24828446e-01 -1.20985353e+00
7.49713480e-01 1.44149053e+00 6.38472080e-01 5.87283909e-01
-9.90932524e-01 -8.57411623e-01 1.39671817e-01 5.40055573e-01
-1.66839457e+00 -9.12429914e-02 5.00031114e-01 -6.25481129e-01
1.02318835e+00 4.92310643e-01 6.66555941e-01 1.17911184e+00
-6.83166981e-02 1.34853697e+00 9.59592402e-01 -4.19861585e-01
-6.15688488e-02 5.59320688e-01 6.10372365e-01 5.06249070e-01
5.80553770e-01 3.26257885e-01 -6.58050835e-01 9.07465965e-02
2.48696193e-01 1.78431403e-02 -2.97487706e-01 -3.78310770e-01
-1.47028065e+00 6.69852614e-01 6.07035697e-01 2.21413329e-01
-7.48828724e-02 -4.37556565e-01 4.00908530e-01 9.76732653e-03
1.65932044e-01 5.94874144e-01 -1.46189734e-01 3.23058873e-01
-5.26057601e-01 8.01988006e-01 6.85677171e-01 1.55305731e+00
7.55823255e-01 -5.11143625e-01 -5.31628251e-01 6.38950050e-01
-3.36827710e-02 4.21422094e-01 2.18918547e-01 -6.90025330e-01
5.23270488e-01 9.18048322e-01 1.85319498e-01 -1.01248157e+00
-8.38716030e-01 -6.08643234e-01 -8.17963123e-01 8.30462724e-02
3.86258960e-01 -1.21333130e-01 -7.33783066e-01 1.51690102e+00
6.58727288e-01 -1.89533886e-02 -1.92647591e-01 1.14542305e+00
1.17467248e+00 4.64882880e-01 1.04443333e-03 4.24445085e-02
1.46511042e+00 -1.29606593e+00 -5.04690170e-01 -3.53385597e-01
9.85613883e-01 -5.02185225e-01 1.13528490e+00 2.46961057e-01
-7.74462759e-01 -9.27760065e-01 -1.12294745e+00 6.89242929e-02
-6.37350559e-01 4.26298231e-01 7.77856529e-01 4.85217899e-01
-6.17561638e-01 5.06247245e-02 -5.20050228e-01 -7.75440454e-01
3.05231035e-01 2.32599124e-01 -3.98144931e-01 2.53717955e-02
-1.17488873e+00 8.12867880e-01 6.50517523e-01 2.15305850e-01
-1.09575200e+00 -6.38092160e-01 -7.78595507e-01 3.87915596e-03
7.54247665e-01 -7.22744405e-01 1.15329063e+00 -4.51812416e-01
-7.02373564e-01 1.04866552e+00 -4.85674106e-02 -5.12713850e-01
5.99871397e-01 -3.03769499e-01 -3.81136179e-01 -8.59944746e-02
1.35855913e-01 9.57682252e-01 4.00381505e-01 -1.40568852e+00
-1.13149786e+00 -4.67250586e-01 7.23020732e-02 2.71867275e-01
-2.01178566e-01 2.67959714e-01 -4.99112189e-01 -4.12914634e-01
-1.24035878e-02 -1.45904648e+00 1.29432440e-01 -1.25175759e-01
-7.72644699e-01 -6.30220711e-01 1.15804350e+00 -3.08068335e-01
1.46786058e+00 -2.06539845e+00 2.16410905e-02 -9.02835801e-02
3.06712210e-01 4.69030499e-01 -7.80212879e-02 3.41035008e-01
-1.84351504e-01 8.05052668e-02 1.65401429e-01 -3.12747061e-01
-9.39626154e-03 -4.84095365e-02 -1.74843028e-01 4.27571446e-01
4.69591469e-02 1.01613033e+00 -8.11288178e-01 -5.40758908e-01
2.80708522e-01 1.05922952e-01 -5.84412873e-01 4.76244420e-01
3.73877510e-02 4.88991708e-01 -1.94646299e-01 8.03541541e-01
6.30545855e-01 -3.27303052e-01 -1.80255666e-01 -3.44702333e-01
3.41304136e-03 -9.01253298e-02 -1.10108352e+00 9.47173893e-01
-1.44754658e-02 6.82217777e-01 -3.83561134e-01 -9.64744985e-01
7.05902994e-01 1.95200026e-01 4.64696854e-01 -4.62496817e-01
7.31378654e-03 4.98297485e-03 4.15558457e-01 -5.16779721e-01
3.77602309e-01 2.78885365e-01 -1.79794341e-01 5.48406132e-02
5.07099852e-02 1.84473723e-01 5.46043038e-01 1.32424712e-01
8.78203452e-01 -2.54020751e-01 4.72590804e-01 -8.61595497e-02
4.02479500e-01 7.94442967e-02 6.46851897e-01 1.36012137e+00
-6.71217203e-01 7.91437268e-01 2.11197317e-01 -6.36181831e-01
-8.53595078e-01 -8.98705423e-01 -3.98612350e-01 1.25264788e+00
4.02412117e-01 -5.36626518e-01 -8.36975515e-01 -1.12495649e+00
5.69090396e-02 7.03786194e-01 -7.88791120e-01 -1.47834286e-01
-4.35783088e-01 -7.74364293e-01 2.48268172e-01 7.71687806e-01
6.64645314e-01 -1.40328312e+00 -8.30726027e-01 -2.16512337e-01
-1.93389475e-01 -1.33611286e+00 -5.09383142e-01 -1.56836119e-04
-3.48525882e-01 -1.26293468e+00 -5.38491607e-01 -8.51563752e-01
4.49269861e-01 8.41184139e-01 1.06353378e+00 1.42112881e-01
-8.18884671e-01 4.60162431e-01 -5.75841308e-01 -8.44869316e-01
-2.41560683e-01 1.24696739e-01 2.52414882e-01 -5.24213985e-02
6.97768271e-01 -7.24784657e-02 -6.82382703e-01 1.09394097e+00
-6.09623432e-01 3.23765159e-01 3.90793234e-01 7.62063920e-01
2.93526381e-01 2.88186729e-01 3.38136226e-01 -8.16504717e-01
3.33200127e-01 -4.90505964e-01 -4.95277673e-01 5.48070967e-01
-4.06466037e-01 -3.40093225e-01 2.30207294e-01 -7.29317129e-01
-1.18817699e+00 1.33776560e-01 4.01478857e-01 -5.41202784e-01
-5.35759151e-01 3.27646762e-01 -2.86660731e-01 -6.12921752e-02
9.84036565e-01 -9.58632827e-02 -5.93897879e-01 -3.72043252e-01
1.39701992e-01 7.42015600e-01 5.60729682e-01 -3.25552315e-01
6.55103028e-01 3.16696852e-01 -9.66347456e-02 -9.23830867e-01
-1.12677133e+00 -9.62108254e-01 -7.98552811e-01 -3.18300426e-01
9.91638005e-01 -8.16974580e-01 -8.03767562e-01 6.12975776e-01
-1.27557898e+00 -2.32206300e-01 -1.60328187e-02 4.77741092e-01
-4.16460693e-01 3.80539268e-01 -3.14557552e-01 -1.02521980e+00
-1.39486164e-01 -1.23285282e+00 1.31472373e+00 2.79577196e-01
-3.73837411e-01 -5.48895478e-01 -8.49677622e-02 6.39050424e-01
-1.13654710e-01 2.26089209e-01 5.91118991e-01 -7.78501153e-01
-6.02998674e-01 -5.33279300e-01 -3.78044784e-01 2.94161737e-01
-3.48103829e-02 -7.38570392e-02 -8.41081500e-01 -4.69033033e-01
-6.85345381e-02 -4.34676975e-01 7.07648516e-01 9.88160744e-02
1.27423310e+00 -1.07725337e-01 -8.20614219e-01 5.00699341e-01
7.93389499e-01 4.72817391e-01 5.65034926e-01 2.49179021e-01
7.91342854e-01 8.56872618e-01 1.07474303e+00 5.78513801e-01
3.73119920e-01 1.07783043e+00 3.06895912e-01 -1.51584119e-01
-4.37735051e-01 -2.29039162e-01 1.10153221e-01 1.25636220e-01
-1.85466379e-01 -6.03964269e-01 -1.25342321e+00 5.05557537e-01
-2.13971448e+00 -9.22303617e-01 -3.83820653e-01 2.16620874e+00
1.94868639e-01 2.17989013e-02 5.01666963e-01 1.16267227e-01
8.78762662e-01 -2.52368778e-01 -5.53224444e-01 3.74012560e-01
-1.82718411e-01 -3.94833595e-01 2.76452959e-01 5.87958731e-02
-1.48104525e+00 9.01551843e-01 6.92321444e+00 5.40557206e-01
-7.53886998e-01 1.24918714e-01 3.24376673e-01 -1.80488974e-01
4.97079939e-01 -1.02999337e-01 -1.49555445e+00 4.59913999e-01
4.32637155e-01 -6.77502379e-02 2.52187997e-01 1.07296276e+00
-1.74930900e-01 -2.12340087e-01 -1.59632015e+00 1.24989593e+00
2.53842950e-01 -7.89039552e-01 1.30139023e-01 3.49987224e-02
7.79631495e-01 -1.84164166e-01 1.41683057e-01 7.00411141e-01
5.63872904e-02 -8.92740130e-01 6.73303127e-01 3.17039222e-01
3.15125614e-01 -3.98975134e-01 8.76264513e-01 6.34222448e-01
-1.24023581e+00 -4.05845672e-01 -2.39866048e-01 -4.11883593e-02
4.47827429e-02 2.09606662e-01 -1.03576720e+00 4.61356699e-01
1.08719814e+00 7.03499019e-01 -1.11076260e+00 1.33442318e+00
-1.60918504e-01 4.96533185e-01 -4.53945011e-01 -2.14745738e-02
-7.51010701e-02 3.02104324e-01 7.43552566e-01 1.37799799e+00
3.29536974e-01 3.72677743e-02 5.84370017e-01 8.44976068e-01
2.46215135e-01 -4.87431102e-02 -7.12723970e-01 1.19174026e-01
3.32959712e-01 1.12921989e+00 -7.19482362e-01 -4.09694105e-01
-6.69611037e-01 9.50518250e-01 4.75169450e-01 4.86070812e-01
-1.16134977e+00 -2.63695598e-01 3.81999701e-01 3.88191819e-01
1.14340484e-01 -9.72547457e-02 -8.79674703e-02 -1.28212190e+00
1.49631962e-01 -9.07415450e-01 8.97745013e-01 -8.98408532e-01
-1.45308924e+00 4.66244876e-01 7.01368690e-01 -1.32075870e+00
-1.35073885e-01 -8.17494869e-01 -4.74359363e-01 4.56991255e-01
-8.77570152e-01 -1.37981606e+00 -9.04624641e-01 3.94346505e-01
5.93115270e-01 -1.24359995e-01 4.17218447e-01 2.13260129e-01
-8.74827087e-01 9.20894265e-01 -4.05131727e-01 4.08682488e-02
6.69333577e-01 -9.15125549e-01 2.82309771e-01 7.14393675e-01
1.91211313e-01 6.38407648e-01 8.25261593e-01 -6.42471313e-01
-1.08173406e+00 -1.03546536e+00 6.05256557e-01 -9.40680742e-01
2.75148660e-01 -7.33083725e-01 -1.08909476e+00 1.00132859e+00
1.27146328e-02 1.92524925e-01 3.90437156e-01 2.54987836e-01
-1.21972106e-01 -2.49642860e-02 -9.81993556e-01 5.82328916e-01
1.64000583e+00 -2.27159455e-01 -5.06564081e-01 6.14613473e-01
6.04300976e-01 -4.63723898e-01 -6.44349217e-01 9.24031258e-01
5.64205706e-01 -1.25971925e+00 1.07951224e+00 -5.51794469e-01
3.70102050e-03 -3.61590862e-01 -6.88921586e-02 -1.01611197e+00
-4.92991775e-01 -2.64392406e-01 -2.87770987e-01 1.10072291e+00
2.08536267e-01 -4.73889798e-01 8.08057070e-01 5.79085350e-01
-1.07227750e-01 -7.33043969e-01 -5.80417812e-01 -1.25893211e+00
-3.29108417e-01 -4.04211581e-01 5.49663723e-01 7.15318441e-01
1.24787927e-01 4.27026689e-01 -5.57690918e-01 3.79593104e-01
6.97882473e-01 1.11353241e-01 1.39108288e+00 -1.14052796e+00
-3.20886910e-01 -2.69348860e-01 -6.30928695e-01 -1.09603584e+00
9.78026092e-02 -8.33958030e-01 6.51647449e-02 -1.16942585e+00
8.32622290e-01 -2.59990036e-01 -1.80014178e-01 5.53383052e-01
-4.79563653e-01 2.81315327e-01 3.35082203e-01 4.65319574e-01
-9.83316779e-01 5.40081322e-01 1.27113950e+00 -2.35334009e-01
-3.70736599e-01 2.06420675e-01 -3.44202280e-01 7.24997342e-01
5.57868481e-01 -2.17320412e-01 -4.78254080e-01 -1.38737280e-02
-1.80808648e-01 -2.89136261e-01 7.22507834e-01 -1.50322700e+00
2.97299832e-01 -2.37957358e-01 3.42825860e-01 -1.05980623e+00
2.23537594e-01 -5.92999399e-01 2.04404131e-01 3.06973964e-01
-3.00812185e-01 -2.11683780e-01 1.08214028e-01 3.06847155e-01
-1.74659997e-01 -1.18248671e-01 6.69742405e-01 -7.55042434e-02
-1.08633494e+00 2.84166545e-01 -2.55679667e-01 3.36239561e-02
1.57194245e+00 -3.81322622e-01 -3.32969368e-01 -3.85460883e-01
-7.78382063e-01 5.08872390e-01 2.67246872e-01 7.73069799e-01
5.53012431e-01 -1.31750739e+00 -6.26679659e-01 2.80840307e-01
8.33750784e-01 7.32340217e-02 2.94761658e-01 8.44211400e-01
-9.37463641e-02 8.24712515e-01 -6.16181567e-02 -9.27591503e-01
-1.66594362e+00 1.14283657e+00 3.86013716e-01 -1.10657394e-01
-5.83859324e-01 6.62993610e-01 1.06236529e+00 -4.65557754e-01
6.30520761e-01 -2.96758980e-01 -4.05403487e-02 -2.77302921e-01
6.82135642e-01 7.13691890e-01 -1.83811545e-01 -7.69815862e-01
-5.11280477e-01 2.92709738e-01 -1.27351925e-01 3.42029274e-01
7.95442104e-01 -7.78775960e-02 2.12564841e-01 3.61177683e-01
1.04056609e+00 -5.35227239e-01 -9.88064170e-01 -1.65976554e-01
5.09925075e-02 -5.22884488e-01 -4.49296534e-01 -8.73706162e-01
-7.50645578e-01 8.70393693e-01 8.70587826e-01 2.69770980e-01
9.60671306e-01 5.98245502e-01 3.84529173e-01 6.42824292e-01
6.18906140e-01 -8.99881005e-01 3.17574263e-01 3.71231854e-01
1.33911002e+00 -1.53720379e+00 -1.27860755e-01 -8.76464069e-01
-8.80492866e-01 6.24107838e-01 1.39025795e+00 3.24413776e-01
4.41431671e-01 -5.63481934e-02 -3.11325759e-01 -4.84042972e-01
-5.53614318e-01 -3.96005929e-01 6.54823959e-01 8.25115263e-01
3.48987669e-01 1.75325260e-01 -2.14438483e-01 5.44823885e-01
-1.48792028e-01 -9.88970101e-02 9.37717482e-02 8.25888515e-01
-3.36828709e-01 -6.09405518e-01 -6.03773475e-01 4.32432592e-01
1.93369627e-01 2.82213241e-01 -6.44065678e-01 1.06795144e+00
5.00143647e-01 1.02874923e+00 -1.52814820e-01 -6.31560624e-01
8.30890119e-01 -9.72329155e-02 4.53092903e-01 -7.86811113e-01
-3.65199983e-01 -3.25617403e-01 -4.11712043e-02 -6.13654852e-01
-2.08439246e-01 -7.32157528e-01 -9.75542188e-01 -5.85151948e-02
-6.27905369e-01 -1.54476643e-01 1.89134970e-01 7.73420691e-01
4.22924072e-01 4.03649092e-01 2.95525759e-01 -1.11025012e+00
-4.13423210e-01 -1.13849533e+00 -6.26468599e-01 1.02423871e+00
-1.29118366e-02 -1.24457777e+00 -4.91363466e-01 -2.32373655e-01] | [9.595017433166504, 1.3763809204101562] |
6ba797ec-7654-424d-bb01-e5d9ee4b5335 | transformer-with-peak-suppression-and | 2107.06538 | null | https://arxiv.org/abs/2107.06538v2 | https://arxiv.org/pdf/2107.06538v2.pdf | Transformer with Peak Suppression and Knowledge Guidance for Fine-grained Image Recognition | Fine-grained image recognition is challenging because discriminative clues are usually fragmented, whether from a single image or multiple images. Despite their significant improvements, most existing methods still focus on the most discriminative parts from a single image, ignoring informative details in other regions and lacking consideration of clues from other associated images. In this paper, we analyze the difficulties of fine-grained image recognition from a new perspective and propose a transformer architecture with the peak suppression module and knowledge guidance module, which respects the diversification of discriminative features in a single image and the aggregation of discriminative clues among multiple images. Specifically, the peak suppression module first utilizes a linear projection to convert the input image into sequential tokens. It then blocks the token based on the attention response generated by the transformer encoder. This module penalizes the attention to the most discriminative parts in the feature learning process, therefore, enhancing the information exploitation of the neglected regions. The knowledge guidance module compares the image-based representation generated from the peak suppression module with the learnable knowledge embedding set to obtain the knowledge response coefficients. Afterwards, it formalizes the knowledge learning as a classification problem using response coefficients as the classification scores. Knowledge embeddings and image-based representations are updated during training so that the knowledge embedding includes discriminative clues for different images. Finally, we incorporate the acquired knowledge embeddings into the image-based representations as comprehensive representations, leading to significantly higher performance. Extensive evaluations on the six popular datasets demonstrate the advantage of the proposed method. | ['Xiaoguang Han', 'Lili Wang', 'Xinda Liu'] | 2021-07-14 | null | null | null | null | ['fine-grained-image-recognition'] | ['computer-vision'] | [ 3.39065850e-01 -2.44547635e-01 -2.08199352e-01 -2.60463208e-01
-6.77494586e-01 -4.05836672e-01 5.05444288e-01 3.27022746e-02
-4.80938792e-01 5.67304254e-01 2.98275828e-01 2.28964269e-01
-5.16153276e-01 -7.71353781e-01 -4.97758359e-01 -1.17588925e+00
4.31104124e-01 7.77705461e-02 4.12921369e-01 2.10393056e-01
4.64066386e-01 6.27454102e-01 -1.72030759e+00 3.19230020e-01
8.14879358e-01 1.35378993e+00 6.04690671e-01 1.50866643e-01
-2.95508593e-01 6.28738880e-01 -5.36119878e-01 -3.27284820e-02
8.82788226e-02 -1.93015695e-01 -4.87575948e-01 3.22646409e-01
4.34651673e-01 -2.58332789e-01 -2.28222981e-01 1.04467654e+00
4.56495672e-01 2.28157803e-01 5.63315868e-01 -8.79306197e-01
-9.02107954e-01 3.06656092e-01 -3.91123086e-01 4.45158601e-01
1.91212788e-01 2.29850970e-02 9.39364195e-01 -1.25093329e+00
3.75347733e-01 1.00451124e+00 1.14315934e-01 2.70394981e-01
-9.84494567e-01 -5.47253609e-01 4.84885454e-01 6.77556455e-01
-1.58787525e+00 -2.98028290e-01 1.00967658e+00 -5.49415827e-01
8.64523828e-01 1.12354450e-01 5.31518400e-01 5.62966406e-01
-2.48390678e-02 7.77350545e-01 1.00474274e+00 -3.51417422e-01
-3.26927900e-02 5.04711151e-01 3.47498357e-01 7.27536261e-01
9.83173996e-02 5.88817932e-02 -6.80887401e-01 2.07128480e-01
6.93504930e-01 3.45508307e-01 -4.66594189e-01 -4.98581350e-01
-1.18879211e+00 6.88866138e-01 6.63522661e-01 7.28782952e-01
-7.53414631e-01 -2.66407430e-01 8.91418606e-02 1.50033325e-01
1.25961661e-01 1.72840789e-01 -3.81179959e-01 1.29645437e-01
-9.07559454e-01 -2.35923529e-01 3.13639015e-01 5.73202372e-01
1.28657949e+00 1.66849971e-01 -4.38017458e-01 9.12480175e-01
2.52934277e-01 2.13746309e-01 7.69628704e-01 -2.99611598e-01
3.95094067e-01 1.10353005e+00 -7.37037286e-02 -1.30961919e+00
-1.82225481e-01 -7.79270649e-01 -7.70779669e-01 7.34106228e-02
1.98166952e-01 2.66491532e-01 -7.95648992e-01 1.62501597e+00
2.88840353e-01 9.64651108e-02 -1.68340132e-02 1.03003168e+00
8.34412038e-01 6.36725903e-01 5.83427167e-03 -1.52240798e-01
1.37225711e+00 -1.14964294e+00 -3.89561981e-01 -1.91867426e-01
2.78039753e-01 -7.09516048e-01 9.19603348e-01 3.53399277e-01
-7.08863795e-01 -9.68676031e-01 -1.02948594e+00 1.01951603e-02
-5.85577548e-01 4.86346036e-01 2.94640958e-01 3.27837169e-01
-1.09768796e+00 3.15944225e-01 -2.62339711e-01 -1.58064365e-01
4.05783594e-01 4.04269338e-01 -5.26496232e-01 -2.43598491e-01
-1.01096761e+00 7.61374354e-01 5.97508430e-01 3.54344964e-01
-8.30717027e-01 -6.06482506e-01 -8.53925169e-01 3.83500695e-01
4.02067661e-01 -4.91500705e-01 4.46209580e-01 -1.13555789e+00
-1.10085750e+00 5.35134375e-01 -2.39908397e-01 3.30664366e-02
6.90115336e-03 -1.04923889e-01 -3.06341618e-01 2.83672780e-01
1.65632948e-01 4.90352809e-01 1.28781891e+00 -1.31381989e+00
-7.42755711e-01 -3.98824155e-01 1.38318717e-01 3.60638916e-01
-6.65280521e-01 -2.67187685e-01 -7.30220139e-01 -7.31105149e-01
9.66969281e-02 -4.17857200e-01 5.36750443e-02 -2.90548891e-01
-2.62358993e-01 -2.27197438e-01 8.09891224e-01 -5.93675196e-01
1.25464487e+00 -2.38773584e+00 4.16184098e-01 3.81355792e-01
1.34568408e-01 2.83661395e-01 -4.18279827e-01 2.27978468e-01
-1.85197249e-01 -2.57143915e-01 7.28246644e-02 4.79522869e-02
-1.31941840e-01 1.97786108e-01 -2.27434695e-01 1.57060847e-01
5.34434021e-01 9.42836940e-01 -7.28214324e-01 -5.23372114e-01
5.57567716e-01 6.07794762e-01 -4.58764911e-01 1.55156687e-01
1.82653621e-01 3.47436428e-01 -8.24637949e-01 7.91228652e-01
6.26850963e-01 -2.09988147e-01 -2.30521485e-02 -7.27702141e-01
-2.37683654e-01 -2.53094584e-01 -1.13990986e+00 1.52876651e+00
-4.69412416e-01 1.14772536e-01 -2.58751273e-01 -1.29245901e+00
1.16359651e+00 -7.33034685e-02 2.65787572e-01 -9.87601101e-01
5.12230769e-02 1.10443652e-01 -9.45119411e-02 -6.27864838e-01
4.36034232e-01 -1.35700479e-01 -1.03564315e-01 3.39548647e-01
3.37856352e-01 3.04638326e-01 1.01704165e-01 3.58393379e-02
6.96340740e-01 9.26601216e-02 4.73025888e-01 1.30751163e-01
9.45327103e-01 -1.29322097e-01 3.63527477e-01 5.74637473e-01
-1.46710977e-01 3.46379817e-01 1.52819538e-02 -4.55321759e-01
-5.31065524e-01 -9.95525777e-01 4.56974842e-02 1.29104650e+00
3.92133832e-01 -2.74827451e-01 -4.68415022e-01 -9.76347089e-01
4.68595996e-02 3.65837187e-01 -9.87664461e-01 -4.85669822e-01
-3.11353624e-01 -4.80331361e-01 1.21462904e-01 5.07329762e-01
5.47218323e-01 -1.34747791e+00 -5.35479784e-01 8.43482688e-02
-6.16254322e-02 -6.55245721e-01 -5.13252139e-01 3.35765958e-01
-7.19655395e-01 -1.12084508e+00 -7.49551117e-01 -1.02871180e+00
9.24780130e-01 6.30324125e-01 6.89262748e-01 8.72137696e-02
-3.23879778e-01 5.36195636e-01 -5.11217535e-01 1.55488580e-01
2.06745997e-01 -2.24174500e-01 -1.12963676e-01 6.74565554e-01
4.10295248e-01 -4.07808959e-01 -6.22934759e-01 3.27247918e-01
-9.25234199e-01 -2.77302653e-01 1.18597472e+00 1.24095559e+00
8.92383993e-01 2.43824750e-01 5.66788077e-01 -3.70025069e-01
4.52225715e-01 -4.40029114e-01 -3.19971591e-01 5.18842578e-01
-2.85650432e-01 2.88225383e-01 8.40445280e-01 -4.78615880e-01
-1.17847824e+00 1.06830940e-01 6.10881858e-02 -6.13118172e-01
-2.13478789e-01 5.13742387e-01 -4.52518612e-01 -3.46958309e-01
2.86283731e-01 8.86074245e-01 -3.75073165e-01 -5.03987491e-01
3.97941500e-01 4.40556616e-01 4.29424524e-01 -5.21728694e-01
7.42476106e-01 3.37482125e-01 -4.36174989e-01 -8.03679764e-01
-8.09419394e-01 -6.69761181e-01 -8.44234824e-01 -2.65122533e-01
6.38777852e-01 -7.04698682e-01 -5.44919968e-01 2.68303156e-01
-8.74558151e-01 2.94349492e-01 -5.25657475e-01 4.94404674e-01
-2.61905730e-01 3.59717011e-01 -5.84937334e-02 -5.72362483e-01
-2.27060467e-01 -1.26310098e+00 1.14464557e+00 6.03345752e-01
1.54155463e-01 -8.75082493e-01 -4.43882532e-02 3.20891052e-01
3.02944452e-01 -1.60906300e-01 1.03889728e+00 -6.46279573e-01
-8.55561078e-01 -2.23172724e-01 -5.95723569e-01 4.50017840e-01
3.56901258e-01 -3.68505657e-01 -1.07702661e+00 -1.85264170e-01
5.64526878e-02 -3.82371306e-01 1.30950570e+00 2.57069618e-01
1.09965253e+00 -3.11690360e-01 -3.30252498e-01 5.45880497e-01
1.63756597e+00 3.82206619e-01 4.99087274e-01 4.44917649e-01
7.12489486e-01 6.82848513e-01 6.94356143e-01 3.83682072e-01
2.08605900e-01 5.35242915e-01 4.14135933e-01 -5.53724077e-03
-2.77432591e-01 -1.71169192e-01 3.26502264e-01 9.91075516e-01
-2.84703001e-02 1.97033525e-01 -3.76932055e-01 7.10136235e-01
-1.71906233e+00 -9.87806201e-01 5.55473447e-01 2.01296902e+00
6.07368946e-01 -9.00570974e-02 -1.98744759e-01 2.26671681e-01
7.25891232e-01 3.21695983e-01 -7.12132156e-01 -3.80717754e-01
-3.79827648e-01 1.81671530e-01 -4.02359739e-02 2.75890201e-01
-9.37629282e-01 9.40060794e-01 5.24275398e+00 1.21103644e+00
-1.18392241e+00 9.19862688e-02 3.22920144e-01 -3.50921601e-02
-4.08758491e-01 -7.89339170e-02 -8.18495691e-01 3.62350553e-01
2.36246288e-01 -7.12809637e-02 3.24426502e-01 8.40390861e-01
4.47116084e-02 -3.01871002e-01 -9.70751047e-01 1.07015193e+00
3.97144139e-01 -1.21158254e+00 7.24356890e-01 -1.04046881e-01
5.52505732e-01 -3.76241237e-01 2.95026720e-01 2.72758812e-01
-2.38835722e-01 -9.49678600e-01 6.58773482e-01 9.55420375e-01
6.29930854e-01 -8.04972470e-01 7.50049353e-01 3.56772333e-01
-1.33171463e+00 -2.51484096e-01 -5.57454228e-01 1.01330325e-01
-2.13043988e-01 6.41218007e-01 -7.06294954e-01 8.73857856e-01
5.36428928e-01 8.24509740e-01 -7.66782105e-01 9.39711452e-01
-2.51808256e-01 2.72364020e-01 2.28835754e-02 9.14892107e-02
3.73424590e-01 -2.35293522e-01 3.37049454e-01 1.14492559e+00
2.92152166e-01 -1.26340205e-03 1.43137649e-01 7.14922130e-01
1.89787924e-01 1.95241272e-01 -5.22471607e-01 9.19194520e-02
2.66138226e-01 1.43022072e+00 -6.42316282e-01 -3.01960200e-01
-5.77033222e-01 1.06588447e+00 5.78692436e-01 5.18521428e-01
-4.80458945e-01 -4.65142787e-01 3.94158959e-01 -2.91891187e-01
8.82633448e-01 6.67538717e-02 -4.74003330e-02 -1.10343754e+00
1.78421572e-01 -6.51224852e-01 5.35842299e-01 -6.56215549e-01
-1.15340889e+00 7.79367268e-01 -2.49152511e-01 -1.22269630e+00
-9.23033878e-02 -5.54873705e-01 -4.26898003e-01 1.00887156e+00
-1.94189346e+00 -1.22879624e+00 -4.47449446e-01 1.00226200e+00
6.64989173e-01 -1.71879098e-01 7.64885724e-01 3.73719066e-01
-5.95818937e-01 7.59506762e-01 1.38670444e-01 -1.04846507e-01
5.70498109e-01 -1.05707729e+00 -5.11959136e-01 8.34061265e-01
2.02420682e-01 6.72409475e-01 8.17581564e-02 -5.83549857e-01
-1.26967573e+00 -1.04471016e+00 7.64047265e-01 -1.20267734e-01
5.37944198e-01 -3.82340550e-02 -1.08366191e+00 2.60729313e-01
-1.07014567e-01 7.80657530e-02 7.51684189e-01 -5.41791022e-02
-5.88172972e-01 -3.82465661e-01 -8.95454466e-01 2.79743224e-01
6.71865702e-01 -6.46469235e-01 -8.93781126e-01 -2.26208270e-01
4.34128344e-01 2.16036752e-01 -7.54613042e-01 2.76874363e-01
7.28543460e-01 -9.18811560e-01 1.10615206e+00 -3.24818581e-01
2.25711733e-01 -5.84243000e-01 -3.08326423e-01 -1.31651604e+00
-7.60627925e-01 2.14419037e-01 -2.96193641e-02 1.31241345e+00
6.59736395e-02 -5.49493611e-01 6.04882240e-01 -6.50106966e-02
-3.36834863e-02 -7.76397169e-01 -9.83609259e-01 -5.99722981e-01
-3.25606763e-01 -2.32517570e-01 4.56775248e-01 7.67524302e-01
-1.09277591e-01 2.76541024e-01 -4.15438175e-01 2.60784268e-01
6.40052795e-01 6.67558968e-01 3.67224276e-01 -1.02785885e+00
-3.26806337e-01 -3.97426397e-01 -5.63254893e-01 -1.04762101e+00
-1.10016540e-02 -9.18355584e-01 6.80091083e-02 -1.44002771e+00
6.55173421e-01 -3.14781636e-01 -9.01053488e-01 7.22119629e-01
-3.92689615e-01 4.70825106e-01 3.80723566e-01 3.84076387e-01
-7.50129998e-01 6.98420286e-01 1.41527486e+00 -4.66139555e-01
-5.29309176e-02 -3.08885664e-01 -9.98903871e-01 6.13396943e-01
5.90966344e-01 -1.68886781e-01 -4.88262802e-01 -3.21966290e-01
-1.40214264e-01 -2.55973369e-01 5.45107961e-01 -9.31204915e-01
4.04556394e-01 -4.97694500e-02 8.63052428e-01 -6.26261711e-01
4.51699108e-01 -8.76602530e-01 3.55799496e-02 3.28666359e-01
-1.06240086e-01 -4.28735256e-01 3.15720856e-01 5.49320340e-01
-6.43365622e-01 -3.26749921e-01 6.80129230e-01 -8.92829895e-02
-1.29806924e+00 3.00097436e-01 -1.16825782e-01 -1.42482087e-01
1.03062189e+00 -7.10671127e-01 -2.27762550e-01 4.48268056e-02
-8.80298734e-01 3.60164642e-02 1.48001269e-01 4.77781326e-01
1.18786824e+00 -1.50456953e+00 -5.01810551e-01 6.03868067e-01
4.04768109e-01 -2.95730650e-01 8.39276075e-01 8.26363444e-01
1.72940865e-01 5.96722484e-01 -4.99362886e-01 -5.57697833e-01
-1.19684637e+00 8.79503489e-01 8.62255916e-02 -4.02893275e-01
-4.66333121e-01 7.30595469e-01 9.13849592e-01 -4.58684750e-02
1.29160136e-01 -2.27973387e-01 -8.24778795e-01 5.25806963e-01
6.75571978e-01 1.86302945e-01 1.00603320e-01 -9.22226191e-01
-5.96322656e-01 1.18383253e+00 -3.12499374e-01 3.03235769e-01
1.29424703e+00 -3.13813567e-01 5.21830395e-02 2.98437268e-01
1.32264030e+00 1.46371424e-01 -1.28371942e+00 -6.01531386e-01
-1.49508476e-01 -6.85974300e-01 1.73708797e-02 -9.29060459e-01
-1.16108990e+00 9.08692956e-01 5.20433366e-01 -5.13660647e-02
1.63041639e+00 1.63765058e-01 3.82417142e-01 2.00880051e-01
3.20766866e-01 -8.70933235e-01 3.76411676e-01 4.67069775e-01
1.00440979e+00 -1.01562488e+00 -1.08392969e-01 -2.45508552e-01
-6.01782978e-01 1.21065092e+00 4.54092383e-01 -1.22536540e-01
5.76737642e-01 -7.66470954e-02 -6.09040894e-02 -2.36848846e-01
-5.58442771e-01 -3.77052814e-01 6.50362372e-01 6.69433892e-01
-7.04442933e-02 -9.52633321e-02 -1.56818092e-01 9.59673285e-01
2.02942818e-01 -2.04246659e-02 3.71073000e-02 6.26017451e-01
-7.68391907e-01 -8.86312723e-01 -4.94580746e-01 4.47065085e-01
-5.85187189e-02 -2.48642117e-01 -3.65733445e-01 4.75312382e-01
6.23778462e-01 8.65082860e-01 2.98090111e-02 -6.16931260e-01
4.74416256e-01 6.30623251e-02 4.60167021e-01 -5.91531396e-01
-5.37753880e-01 1.63341090e-01 -2.54623115e-01 -6.58590794e-01
-5.16627848e-01 -5.29247165e-01 -9.79832172e-01 2.16818839e-01
-4.34354246e-01 2.91411161e-01 4.13384587e-01 1.02979493e+00
4.27273721e-01 7.94633090e-01 7.39293456e-01 -9.96564329e-01
-2.91905522e-01 -6.71447814e-01 -6.86391890e-01 3.07375014e-01
4.17263180e-01 -9.69107866e-01 -3.55503529e-01 -5.64349368e-02] | [9.749601364135742, 1.9437679052352905] |
081b26f1-6233-494a-81bf-d24f72af03a8 | benchmarking-transformers-based-models-on | 2207.09152 | null | https://arxiv.org/abs/2207.09152v1 | https://arxiv.org/pdf/2207.09152v1.pdf | Benchmarking Transformers-based models on French Spoken Language Understanding tasks | In the last five years, the rise of the self-attentional Transformer-based architectures led to state-of-the-art performances over many natural language tasks. Although these approaches are increasingly popular, they require large amounts of data and computational resources. There is still a substantial need for benchmarking methodologies ever upwards on under-resourced languages in data-scarce application conditions. Most pre-trained language models were massively studied using the English language and only a few of them were evaluated on French. In this paper, we propose a unified benchmark, focused on evaluating models quality and their ecological impact on two well-known French spoken language understanding tasks. Especially we benchmark thirteen well-established Transformer-based models on the two available spoken language understanding tasks for French: MEDIA and ATIS-FR. Within this framework, we show that compact models can reach comparable results to bigger ones while their ecological impact is considerably lower. However, this assumption is nuanced and depends on the considered compression method. | ['Sophie Rosset', 'Christophe Servan', 'Sahar Ghannay', 'Oralie Cattan'] | 2022-07-19 | null | null | null | null | ['spoken-language-understanding', 'spoken-language-understanding'] | ['natural-language-processing', 'speech'] | [ 2.41999760e-01 -1.03981115e-01 -5.96879199e-02 -3.87806773e-01
-9.70519125e-01 -5.35089731e-01 9.02080238e-01 3.63790482e-01
-9.39036429e-01 7.96031952e-01 8.78960341e-02 -2.30526701e-01
-1.74944133e-01 -4.53509629e-01 -6.98165715e-01 -4.60120618e-01
1.42009556e-01 9.93163764e-01 2.26483226e-01 -4.41167653e-01
2.48373643e-01 3.47134233e-01 -2.04418421e+00 4.03368145e-01
1.20468879e+00 8.75888228e-01 5.66221178e-01 7.43089676e-01
-3.27996433e-01 5.38785577e-01 -3.60259891e-01 -4.63906139e-01
-3.32175910e-01 -2.64283627e-01 -9.62597787e-01 -2.81114399e-01
4.67471242e-01 1.57047942e-01 -5.84035106e-02 8.11940432e-01
5.11596441e-01 -2.39119958e-02 6.04612350e-01 -8.78963470e-01
-4.47654426e-01 9.08779144e-01 1.36992320e-01 4.45708156e-01
3.68009657e-01 -8.28351974e-02 1.15827954e+00 -1.21374929e+00
5.40042102e-01 1.32994723e+00 5.39307356e-01 5.00508904e-01
-1.11940312e+00 -2.17519343e-01 3.17852944e-01 6.05060577e-01
-1.34171879e+00 -8.14441919e-01 3.87394816e-01 -7.17430860e-02
1.41042316e+00 3.50893319e-01 4.62054074e-01 1.23317838e+00
-5.03733531e-02 8.49451005e-01 1.19379973e+00 -7.71328568e-01
1.89877838e-01 4.77204412e-01 1.98902458e-01 5.01314342e-01
-1.53356880e-01 -5.56994043e-02 -8.40965748e-01 1.95158765e-01
1.46331608e-01 -3.72745365e-01 -5.63120127e-01 -2.36736387e-01
-1.25348353e+00 9.15998101e-01 3.72657180e-02 9.52772975e-01
-3.06987613e-01 -3.28908384e-01 6.18507981e-01 4.91563559e-01
7.75813162e-01 2.57710814e-01 -7.87515402e-01 -7.17654884e-01
-1.19050217e+00 2.30167314e-01 9.43766952e-01 7.50058830e-01
4.52171862e-01 1.03021823e-01 3.50192897e-02 1.10912359e+00
1.91143692e-01 4.71676171e-01 6.46076679e-01 -4.13992286e-01
5.90231717e-01 4.01512712e-01 -1.91506132e-01 -6.67824924e-01
-4.74405289e-01 -6.83122873e-01 -9.41879809e-01 -2.39458337e-01
7.23149121e-01 2.67040640e-01 -6.53471887e-01 1.87468565e+00
-1.08686998e-01 6.84218779e-02 2.18015388e-01 6.20148242e-01
5.31743467e-01 5.82916319e-01 2.39732321e-02 -3.35800469e-01
1.34341550e+00 -1.07396245e+00 -9.12369013e-01 -2.42584586e-01
6.90954924e-01 -8.65297079e-01 1.68481386e+00 7.53598690e-01
-1.20196855e+00 -6.82494879e-01 -7.23863423e-01 -1.33937612e-01
-6.45056784e-01 3.03004175e-01 3.31299096e-01 7.29546964e-01
-1.26911283e+00 5.80743074e-01 -6.30900323e-01 -7.38973975e-01
1.16524637e-01 2.04858050e-01 -3.10159236e-01 -1.05106413e-01
-1.28196955e+00 1.20900810e+00 2.75773019e-01 1.72886163e-01
-1.00820017e+00 -6.70240402e-01 -5.77119172e-01 3.04791570e-01
4.06313479e-01 -4.55020815e-01 1.29781687e+00 -9.17420506e-01
-1.64429021e+00 1.18431926e+00 -2.22136274e-01 -8.49224985e-01
7.25646675e-01 -6.70417964e-01 -5.31544209e-01 -1.29245102e-01
-3.96835029e-01 5.82920790e-01 5.77266395e-01 -1.00615370e+00
-5.62103093e-01 -4.61710125e-01 1.67589933e-01 1.45372584e-01
-5.65847933e-01 1.26735121e-01 -2.78286695e-01 -5.48663318e-01
-1.42119259e-01 -7.50119030e-01 1.71359822e-01 -3.00765067e-01
-1.04391128e-01 -5.08635998e-01 5.88370502e-01 -4.48389977e-01
1.49334824e+00 -1.93196392e+00 5.57265759e-01 -4.61264879e-01
-2.11107865e-01 5.32062769e-01 -3.57887179e-01 6.00110888e-01
-5.52170463e-02 1.42256960e-01 -3.52102995e-01 -6.48173392e-01
1.33108363e-01 2.33608171e-01 -5.67640007e-01 1.93241164e-01
1.62566200e-01 6.65550828e-01 -5.60201764e-01 -3.18221897e-01
3.35939050e-01 7.35428631e-01 -6.59266949e-01 2.65858054e-01
-4.18176800e-01 2.79320836e-01 -1.36000514e-01 4.11836058e-01
5.05280316e-01 -6.50537983e-02 -1.66611690e-02 -2.21534029e-01
-3.90348852e-01 6.76967442e-01 -8.37993145e-01 2.01147985e+00
-9.49677408e-01 5.12683213e-01 2.94653233e-02 -1.20211375e+00
7.67349064e-01 4.68869895e-01 1.58813193e-01 -6.57357335e-01
1.65821910e-01 5.01104414e-01 3.40673476e-01 -3.05617869e-01
4.69640166e-01 -1.19209155e-01 1.89989582e-01 4.22883898e-01
4.30221856e-01 -3.84028465e-01 4.72416997e-01 -1.13720624e-02
8.08466673e-01 4.84421365e-02 3.61334682e-01 -6.72241688e-01
9.93615925e-01 -2.87134141e-01 -9.39221680e-02 7.22569585e-01
-8.98678750e-02 5.49928606e-01 1.25254095e-01 -4.33236092e-01
-8.15294504e-01 -9.82501686e-01 -3.76755983e-01 1.51869738e+00
-3.02933455e-01 -6.23882115e-01 -1.03677583e+00 -5.23683906e-01
-4.33165193e-01 1.04515386e+00 -4.39725369e-01 -5.42677194e-02
-4.73941028e-01 -6.60452902e-01 7.21013665e-01 1.52724117e-01
3.55774015e-01 -1.29627228e+00 -5.53434908e-01 4.11696076e-01
-3.47803563e-01 -1.27794588e+00 -7.81479552e-02 1.61263287e-01
-8.46277833e-01 -8.65941644e-01 -8.59715283e-01 -4.66674358e-01
-2.11898498e-02 -3.47285043e-03 1.74219477e+00 -2.61347704e-02
1.79237183e-02 4.62253869e-01 -5.57508647e-01 -6.81079865e-01
-5.42299390e-01 7.08293378e-01 2.11168379e-01 1.32059798e-01
5.05843759e-01 -5.12566090e-01 -5.16916141e-02 1.43303722e-01
-8.40991974e-01 -9.16111190e-03 4.88336831e-01 7.15096593e-01
4.21037674e-01 -3.04817885e-01 6.75984800e-01 -7.36386955e-01
5.51848769e-01 -8.27233791e-02 -5.68098664e-01 5.08950174e-01
-5.40052176e-01 2.94699818e-01 8.52724671e-01 -2.75854617e-01
-1.01571453e+00 -2.70717353e-01 -5.26752174e-01 -9.10377875e-02
-4.16257471e-01 6.15323246e-01 -1.56433791e-01 1.96620479e-01
5.42364419e-01 5.18912137e-01 -3.89041901e-01 -7.60844409e-01
2.77859330e-01 7.28261113e-01 2.58246034e-01 -6.00476503e-01
2.27474511e-01 1.89956620e-01 -5.77957213e-01 -1.21537852e+00
-8.85167837e-01 -2.15018615e-01 -6.38820708e-01 -4.24016379e-02
6.92131937e-01 -7.63838530e-01 -5.36930442e-01 5.19548655e-01
-1.37149656e+00 -2.56904125e-01 -2.40728587e-01 4.97400165e-01
-6.49270833e-01 2.31448129e-01 -4.42139119e-01 -9.24968541e-01
-3.46719056e-01 -1.30087781e+00 1.22891057e+00 -2.55307555e-01
-1.31538197e-01 -1.03779888e+00 2.20371291e-01 2.76094496e-01
7.66930103e-01 -5.39721847e-01 1.04022229e+00 -7.73234189e-01
-3.73925775e-01 4.62551676e-02 -1.37611166e-01 4.34190452e-01
-1.11246660e-01 9.99849197e-03 -1.47085690e+00 -3.26584399e-01
1.80534095e-01 -5.48375845e-01 9.85066533e-01 2.70068407e-01
1.20096529e+00 5.93146821e-03 -1.44655749e-01 1.90102965e-01
1.13679063e+00 1.41161144e-01 4.98011023e-01 4.85293493e-02
4.11088973e-01 7.99994111e-01 5.56501865e-01 1.74070820e-01
3.38510811e-01 9.20816958e-01 3.72881085e-01 1.79617077e-01
3.76385227e-02 4.20202240e-02 5.47547519e-01 1.50289202e+00
-3.76967400e-01 -8.27677488e-01 -1.01482737e+00 5.62042058e-01
-1.59437239e+00 -7.35684872e-01 -2.29578763e-02 2.37799859e+00
8.47815752e-01 3.41013908e-01 5.12999669e-02 4.76385891e-01
3.83741766e-01 2.53777921e-01 -2.32887924e-01 -5.79905033e-01
-4.55379367e-01 3.61593604e-01 -1.71647459e-01 7.18703508e-01
-9.51448798e-01 9.85420048e-01 6.25803089e+00 1.11758471e+00
-1.33848524e+00 4.53623593e-01 7.11283088e-01 -1.07819125e-01
-3.47547799e-01 -3.86471897e-01 -9.02351081e-01 1.94634229e-01
1.62781525e+00 -2.78558344e-01 3.49113852e-01 6.43479943e-01
2.63431668e-01 -7.54152238e-02 -1.19231212e+00 9.88189518e-01
2.04643503e-01 -1.05859149e+00 2.36536488e-01 -1.27438277e-01
2.94717401e-01 5.17022908e-01 8.99037346e-02 4.12747979e-01
-1.66192889e-01 -1.24153519e+00 7.15740263e-01 4.38110650e-01
7.95361161e-01 -6.50610030e-01 7.59004653e-01 4.91415113e-01
-1.16676950e+00 7.29922578e-02 -4.20434594e-01 -5.31279575e-03
3.06025922e-01 6.58177435e-01 -3.86886895e-01 5.95027924e-01
7.02661335e-01 6.33271098e-01 -5.11457741e-01 6.81628406e-01
1.02836713e-01 8.77895057e-01 -5.09804130e-01 -1.58581764e-01
4.95785356e-01 -5.99719323e-02 4.53268468e-01 1.42083251e+00
4.76762414e-01 -2.97698110e-01 -1.82777748e-01 6.75128698e-01
8.48633063e-04 6.26297534e-01 -7.11391628e-01 -3.52560490e-01
1.88013032e-01 1.13547075e+00 -5.17215371e-01 -4.30103302e-01
-6.23005509e-01 7.78487861e-01 4.56554025e-01 1.04245722e-01
-9.78746653e-01 8.65732357e-02 4.74972606e-01 2.07822189e-01
1.78609595e-01 -2.70294636e-01 9.42461863e-02 -1.38047290e+00
-6.91403076e-02 -1.21042025e+00 1.79529876e-01 -5.54872453e-01
-1.09431005e+00 1.29849482e+00 -1.84090212e-02 -8.29867244e-01
-5.07858753e-01 -8.42355490e-01 -4.20389920e-02 7.08147287e-01
-1.69465518e+00 -9.86731291e-01 -1.90916955e-01 5.07020772e-01
1.11007464e+00 -4.35449839e-01 1.32579160e+00 6.81713879e-01
-3.61827284e-01 5.19913495e-01 1.25119597e-01 -4.80423659e-01
7.40002871e-01 -1.11237800e+00 3.12159538e-01 6.45993829e-01
7.59405673e-01 3.59064341e-01 8.22088242e-01 -5.47318440e-03
-1.14113772e+00 -5.85858703e-01 1.41749859e+00 -3.94540340e-01
7.10973680e-01 -7.17029750e-01 -1.23040056e+00 4.61779684e-01
5.31995952e-01 -2.14939505e-01 2.54404485e-01 2.69312441e-01
-3.68341595e-01 -1.79364428e-01 -7.64583409e-01 3.50225061e-01
1.21531045e+00 -7.60828197e-01 -6.40216291e-01 4.77728754e-01
5.02311468e-01 -1.59567237e-01 -6.64161205e-01 5.06050706e-01
4.60868388e-01 -1.53221619e+00 7.41354227e-01 -5.38357675e-01
3.46986115e-01 1.51764601e-01 -2.78539628e-01 -1.40326357e+00
1.12543203e-01 -4.56231713e-01 1.87540710e-01 1.25439608e+00
5.53823948e-01 -6.86192453e-01 5.00248671e-01 3.07813101e-02
-2.37817690e-01 -7.49169409e-01 -1.24885166e+00 -6.56756461e-01
2.87202060e-01 -6.33178115e-01 2.48978868e-01 7.43251026e-01
-1.80398956e-01 6.29568815e-01 -2.27799639e-01 -3.31587225e-01
1.09916039e-01 -9.34535190e-02 5.43596447e-01 -1.40557110e+00
-1.89808041e-01 -6.29850745e-01 -9.14066508e-02 -9.61364746e-01
3.95293415e-01 -8.14637423e-01 4.22603562e-02 -1.01496565e+00
1.09657003e-02 -2.93263614e-01 -3.69452059e-01 2.69090116e-01
1.43265584e-02 1.39384747e-01 2.24569499e-01 -9.05679911e-02
-5.74566245e-01 6.94625556e-01 8.79902065e-01 -2.04336911e-01
1.33712515e-01 2.50305813e-02 -2.76770234e-01 8.94243777e-01
6.56037450e-01 -2.79360861e-01 -5.85397422e-01 -8.25395226e-01
1.93745151e-01 -2.28837535e-01 1.26664668e-01 -1.27741456e+00
1.19100139e-01 3.79720688e-01 -3.74020994e-01 -5.42033017e-01
5.61370254e-01 -8.76153052e-01 -1.78777620e-01 4.35242653e-01
-4.18002367e-01 4.67887580e-01 4.75356966e-01 1.57364905e-01
-7.18053758e-01 -3.25770557e-01 8.91369343e-01 -9.25509483e-02
-6.11696601e-01 1.37672424e-01 -3.83685917e-01 3.63370061e-01
5.98783731e-01 -6.71840413e-03 1.02532459e-02 -4.91694450e-01
-7.66806364e-01 -3.72983485e-01 7.31678382e-02 7.29394853e-01
4.19916928e-01 -8.79915059e-01 -1.08177555e+00 2.97789752e-01
1.92878380e-01 -3.13444346e-01 5.14224946e-01 9.99737501e-01
-3.66016775e-01 1.11136782e+00 -7.25270063e-02 -8.47837389e-01
-1.45646977e+00 3.84547204e-01 3.94110262e-01 -6.19388402e-01
-2.71483332e-01 8.88063788e-01 3.91508788e-01 -7.13417649e-01
2.72979081e-01 -6.82554185e-01 -2.42913097e-01 2.96816170e-01
6.02211893e-01 3.88666391e-01 4.39410985e-01 -7.57889092e-01
-5.28834879e-01 6.90371692e-01 -6.15820847e-02 -2.31397226e-01
1.38207829e+00 -1.86086595e-01 -8.41401797e-03 8.90313566e-01
9.83146191e-01 4.97934502e-03 -8.42604220e-01 -1.88073963e-01
1.86037973e-01 -3.09406463e-02 1.00565568e-01 -8.03851604e-01
-9.28900898e-01 1.55061126e+00 7.09429920e-01 4.69380200e-01
1.22676539e+00 -7.27468506e-02 3.61101925e-01 5.45819342e-01
5.27114689e-01 -1.00378847e+00 -1.85013145e-01 9.64925826e-01
1.23112297e+00 -1.23738348e+00 -3.18106949e-01 -4.20404732e-01
-4.90321010e-01 1.06392491e+00 3.06803852e-01 9.77973565e-02
5.49955547e-01 4.10997421e-01 7.73987100e-02 6.78410530e-02
-1.01974499e+00 -2.60082632e-01 4.04930890e-01 4.34165418e-01
9.66855764e-01 -2.08925113e-01 -3.98416370e-01 4.73261118e-01
-4.44987297e-01 1.10902116e-01 2.85670847e-01 2.91936070e-01
-3.10852081e-01 -1.20693803e+00 -1.23515122e-01 1.52071759e-01
-6.42623723e-01 -5.99047601e-01 -2.30394542e-01 1.09543502e+00
-4.70316894e-02 1.00313830e+00 7.45052174e-02 -1.01874741e-02
4.29706007e-01 3.64798844e-01 6.68718100e-01 -5.59987962e-01
-7.06762850e-01 -3.45937125e-02 2.80473888e-01 -5.87989151e-01
-7.78199017e-01 -6.66269243e-01 -8.29991281e-01 -2.12042451e-01
-4.51630712e-01 3.00417662e-01 7.45096266e-01 1.19754589e+00
3.24698418e-01 6.15101039e-01 6.81505352e-02 -8.62100542e-01
-5.74422538e-01 -1.37041628e+00 -4.41306829e-01 1.31635994e-01
8.58017802e-02 -7.18887866e-01 -4.49938238e-01 -5.17633930e-02] | [13.836960792541504, 6.869466304779053] |
7f435983-4da4-4784-aa14-65d5669ee8ae | scaling-instruction-finetuned-language-models | 2210.11416 | null | https://arxiv.org/abs/2210.11416v5 | https://arxiv.org/pdf/2210.11416v5.pdf | Scaling Instruction-Finetuned Language Models | Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance and generalization to unseen tasks. In this paper we explore instruction finetuning with a particular focus on (1) scaling the number of tasks, (2) scaling the model size, and (3) finetuning on chain-of-thought data. We find that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups (zero-shot, few-shot, CoT), and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation). For instance, Flan-PaLM 540B instruction-finetuned on 1.8K tasks outperforms PALM 540B by a large margin (+9.4% on average). Flan-PaLM 540B achieves state-of-the-art performance on several benchmarks, such as 75.2% on five-shot MMLU. We also publicly release Flan-T5 checkpoints, which achieve strong few-shot performance even compared to much larger models, such as PaLM 62B. Overall, instruction finetuning is a general method for improving the performance and usability of pretrained language models. | ['Adams Yu', 'Dasha Valter', 'Kevin Robinson', 'Marie Pellat', 'Alex Castro-Ros', 'Yunxuan Li', 'Jason Wei', 'Quoc V. Le', 'Denny Zhou', 'Adam Roberts', 'Jacob Devlin', 'Jeff Dean', 'Ed H. Chi', 'Slav Petrov', 'Hongkun Yu', 'Andrew Dai', 'Yanping Huang', 'Vincent Zhao', 'Gaurav Mishra', 'Sharan Narang', 'Aakanksha Chowdhery', 'Xinyun Chen', 'Mirac Suzgun', 'Zhuyun Dai', 'Shixiang Shane Gu', 'Albert Webson', 'Siddhartha Brahma', 'Mostafa Dehghani', 'Xuezhi Wang', 'William Fedus', 'Yi Tay', 'Barret Zoph', 'Shayne Longpre', 'Le Hou', 'Hyung Won Chung'] | 2022-10-20 | null | null | null | null | ['multi-task-language-understanding', 'cross-lingual-question-answering', 'paraphrase-identification'] | ['methodology', 'natural-language-processing', 'natural-language-processing'] | [-1.50050357e-01 -5.18448293e-01 -5.20569444e-01 -3.18545759e-01
-1.25858450e+00 -6.87224805e-01 5.41180670e-01 -1.09870378e-02
-6.82425559e-01 5.98262787e-01 2.30638832e-01 -9.10902321e-01
1.07210360e-01 -4.13477033e-01 -8.54307532e-01 -3.83950651e-01
1.16810717e-01 5.41366041e-01 3.64795923e-01 -6.32377803e-01
2.30138510e-01 1.85555797e-02 -1.40924752e+00 4.93402094e-01
7.77675152e-01 5.55074275e-01 1.80686921e-01 9.71141100e-01
-8.73378217e-02 5.68289101e-01 -5.69915473e-01 -2.85813481e-01
-7.33317211e-02 7.37000182e-02 -9.75012302e-01 -4.19268757e-01
6.38219178e-01 -4.06484604e-01 -3.00766230e-01 5.98163068e-01
4.48460251e-01 5.03836870e-01 4.72007930e-01 -1.06216407e+00
-9.21296775e-01 1.06876373e+00 -6.83163762e-01 4.45724547e-01
8.91130120e-02 5.60463905e-01 7.88538277e-01 -8.99737656e-01
1.51895046e-01 1.59583914e+00 5.66531956e-01 8.08435023e-01
-1.28362286e+00 -9.09405768e-01 2.71124959e-01 -8.12210515e-02
-1.29039335e+00 -5.65753281e-01 -1.04489684e-01 -3.20680112e-01
1.57977545e+00 4.16338861e-01 -2.45460063e-01 1.66657722e+00
7.20005035e-01 8.06237817e-01 1.47949243e+00 -6.56816483e-01
1.41979665e-01 -2.38004446e-01 1.14038563e+00 6.33210242e-01
6.42939880e-02 2.85767168e-01 -7.82109141e-01 -3.41518432e-01
1.17789716e-01 -1.55768022e-01 -1.55907214e-01 3.63120645e-01
-1.41111267e+00 8.22243452e-01 1.26659330e-02 2.22239584e-01
4.81590070e-02 4.55224037e-01 6.29906118e-01 5.33870459e-01
4.40040678e-01 7.39535511e-01 -1.00122297e+00 -7.29993105e-01
-7.08712101e-01 2.28336677e-01 8.81382108e-01 1.05662644e+00
7.54682481e-01 2.75234640e-01 -7.15928197e-01 6.84992373e-01
-5.05727194e-02 6.82020009e-01 1.14126301e+00 -7.93916404e-01
6.37950778e-01 3.79149392e-02 -8.31115097e-02 -9.21845064e-02
-5.73598385e-01 -2.73298800e-01 -4.54542667e-01 -1.57945499e-01
2.98092961e-01 -3.05302024e-01 -1.29186964e+00 1.90781927e+00
-1.99030563e-01 9.85691175e-02 9.93986279e-02 5.02022326e-01
8.51620793e-01 9.22795713e-01 4.19081628e-01 -4.17419374e-02
1.37392044e+00 -1.52983606e+00 -5.92075229e-01 -6.91961169e-01
1.24199891e+00 -1.02071989e+00 1.84893727e+00 5.47600091e-01
-7.83274531e-01 -9.04089808e-01 -9.06610310e-01 -2.56727874e-01
-5.35329759e-01 -3.03909723e-02 7.30631411e-01 7.65509009e-01
-1.19910264e+00 3.70539606e-01 -1.05364335e+00 -4.49108452e-01
-1.73127517e-01 3.98396462e-01 -1.58715934e-01 -2.45163918e-01
-1.15265644e+00 7.81582594e-01 7.43778467e-01 -5.83756804e-01
-1.26654029e+00 -9.57529366e-01 -8.85085225e-01 1.59321979e-01
6.33725643e-01 -4.65635121e-01 1.72010028e+00 -1.44153327e-01
-1.52852702e+00 7.49475777e-01 -3.19649875e-01 -5.50873041e-01
1.23158261e-01 -7.07022011e-01 -5.83791137e-01 -6.90910935e-01
-7.93618113e-02 8.78326118e-01 6.18091524e-01 -9.86795127e-01
-4.93027985e-01 -2.06351265e-01 1.88670512e-02 -5.86647773e-03
-5.07906258e-01 6.80796877e-02 -4.57518160e-01 -6.93461776e-01
-5.46322703e-01 -1.18520153e+00 4.27913889e-02 -1.10488939e+00
-5.62633753e-01 -6.26981199e-01 9.49776113e-01 -4.48351264e-01
1.60237014e+00 -2.17560196e+00 -5.89097515e-02 -3.33858073e-01
1.71428993e-01 5.73219717e-01 -7.39178061e-01 4.04191881e-01
-5.67637198e-02 2.52202034e-01 9.32396650e-02 -5.29751599e-01
2.81901747e-01 3.85848373e-01 -3.92349839e-01 -9.10291001e-02
-1.19373888e-01 1.32547486e+00 -7.69875288e-01 -1.94783330e-01
4.22312208e-02 9.96799171e-02 -7.24877715e-01 2.79377311e-01
-5.35689652e-01 -5.59270792e-02 -2.27282122e-01 6.30136311e-01
2.75110096e-01 -3.25166017e-01 -2.17828766e-01 9.78008285e-02
-1.66755896e-02 4.31598932e-01 -6.73381448e-01 1.93041813e+00
-8.68313909e-01 5.84289193e-01 -2.87724346e-01 -2.63918608e-01
7.59026110e-01 1.07107192e-01 -1.70907333e-01 -7.74544299e-01
-5.55696674e-02 1.00664809e-01 7.30742216e-02 -3.93488109e-01
8.93400013e-01 8.67658705e-02 -4.97533262e-01 6.78638577e-01
1.93661809e-01 -3.91028896e-02 3.99684310e-01 3.46422493e-01
1.21764350e+00 -1.67484768e-02 2.57426202e-01 -5.68436980e-01
2.24724412e-01 -7.10680038e-02 3.16347065e-03 1.12802076e+00
-1.69444814e-01 5.40304966e-02 1.89912736e-01 -3.79380018e-01
-4.70326364e-01 -1.00058317e+00 4.23571393e-02 2.31621337e+00
-3.08852911e-01 -9.10506725e-01 -9.51011598e-01 -8.61386120e-01
1.11445718e-01 1.45304203e+00 -6.69917941e-01 -7.09691703e-01
-6.91696048e-01 -9.47301686e-01 7.95053244e-01 7.25199401e-01
3.66973817e-01 -9.88129795e-01 -3.25693011e-01 5.88478446e-02
-2.17523083e-01 -1.06859851e+00 -1.09140944e+00 8.72684181e-01
-7.11933136e-01 -5.89414895e-01 -2.72370219e-01 -6.33074939e-01
7.35039189e-02 4.37792122e-01 1.70661032e+00 -2.58834280e-05
9.35110170e-03 3.79808217e-01 -3.55105489e-01 -3.44384760e-01
-7.08802760e-01 6.76065147e-01 4.21830207e-01 -5.69320083e-01
7.10710585e-01 -7.83187896e-02 2.16799360e-02 4.56891835e-01
-1.00523591e+00 -1.40071526e-01 7.64179111e-01 1.03522468e+00
4.49042290e-01 -1.52589649e-01 5.37163734e-01 -1.06660581e+00
9.57104266e-01 -4.68984663e-01 -3.83175790e-01 4.23194617e-01
-9.12010372e-01 4.48888838e-01 6.59190178e-01 -7.86636472e-01
-1.03941202e+00 -4.73502368e-01 -2.14554384e-01 -4.92297977e-01
-2.69236207e-01 4.99837518e-01 1.23760916e-01 1.05378762e-01
1.04582548e+00 2.59754419e-01 -3.65618497e-01 -6.85254753e-01
6.33782804e-01 8.42676342e-01 6.34533703e-01 -1.15977180e+00
4.94254172e-01 -1.49851665e-01 -4.21014637e-01 -7.31794000e-01
-1.03571570e+00 -4.62651640e-01 -4.72082078e-01 4.95393097e-01
8.18770826e-01 -8.83682013e-01 -6.75405443e-01 4.15877998e-01
-8.70232403e-01 -1.19287026e+00 -6.92781806e-02 2.80504733e-01
-3.87316287e-01 1.05153084e-01 -1.11930323e+00 -2.53882617e-01
-5.86845517e-01 -1.71894515e+00 1.36146379e+00 2.25453693e-02
-4.95033205e-01 -1.01085436e+00 1.40280828e-01 4.87175882e-01
5.06169319e-01 -4.45541501e-01 1.30039084e+00 -9.30993199e-01
-1.43492132e-01 1.39177650e-01 -4.95781861e-02 4.66867477e-01
1.27727643e-01 -1.16246775e-01 -1.21213806e+00 -6.73629284e-01
-1.91902369e-01 -8.74444127e-01 1.15932441e+00 2.50112057e-01
1.32895756e+00 -1.51216462e-01 -4.13451821e-01 6.85115397e-01
9.45366323e-01 2.19640076e-01 3.15001249e-01 2.84565091e-01
7.38439798e-01 5.17323473e-03 6.83148205e-01 1.37316212e-01
3.51599842e-01 7.08543360e-01 1.87690064e-01 3.87008399e-01
-3.17635685e-01 -2.21293375e-01 9.92671967e-01 1.19527924e+00
3.69599372e-01 -4.59764719e-01 -1.17517006e+00 2.75959939e-01
-1.64775550e+00 -2.23420262e-01 -4.13481779e-02 2.12537837e+00
1.08699381e+00 4.38317835e-01 4.76690307e-02 -2.67366439e-01
1.23579055e-01 1.93927065e-01 -5.08406639e-01 -8.60807955e-01
8.60468391e-03 4.84197319e-01 5.43770969e-01 7.46374726e-01
-1.03041720e+00 1.52998435e+00 6.76795530e+00 1.40458894e+00
-1.17036784e+00 3.92451942e-01 8.30328584e-01 -2.02415422e-01
-1.43083751e-01 -2.19824523e-01 -1.58270705e+00 5.20021141e-01
1.69181442e+00 -2.21379042e-01 7.10381389e-01 8.87863815e-01
-1.30965903e-01 -2.62795445e-02 -1.45189047e+00 8.40827346e-01
2.12005585e-01 -1.16871536e+00 2.43123040e-01 2.55514719e-02
1.04111445e+00 4.75907773e-01 2.64080256e-01 1.33836246e+00
7.75031924e-01 -1.29813862e+00 5.36717355e-01 1.35422200e-01
9.72755134e-01 -8.49051237e-01 6.07956588e-01 5.80377340e-01
-7.48971999e-01 -3.43899392e-02 -3.20370317e-01 -3.47567461e-02
-2.38819346e-01 2.30995879e-01 -7.26898909e-01 2.45299533e-01
7.67783105e-01 1.97819859e-01 -1.06282377e+00 2.72371650e-01
-1.01849213e-01 1.17572749e+00 -1.84848443e-01 -6.95599765e-02
6.03858173e-01 4.50622618e-01 1.25818402e-01 1.57742631e+00
1.11835673e-01 9.80709940e-02 6.26854777e-01 6.23046398e-01
-2.74927109e-01 -1.62221640e-01 -3.09679210e-01 -1.77612215e-01
5.05658209e-01 1.24179351e+00 -2.72097796e-01 -4.88297075e-01
-4.36345845e-01 8.50523412e-01 4.08648849e-01 3.69557768e-01
-9.05149341e-01 -2.31962457e-01 7.98441291e-01 2.64858995e-02
-3.74142975e-02 -3.87296557e-01 9.07140225e-02 -9.87235606e-01
-5.16861081e-01 -1.54801893e+00 5.40340662e-01 -8.30298841e-01
-1.14694881e+00 8.33737552e-01 2.32877359e-01 -4.11655575e-01
-5.25676489e-01 -1.01655054e+00 -4.74054515e-01 8.57479572e-01
-1.28633440e+00 -9.54996765e-01 -3.47095579e-01 5.52102983e-01
1.11361539e+00 -2.72685319e-01 1.04187214e+00 1.78792447e-01
-8.28336835e-01 1.07306886e+00 1.93705380e-01 -1.69829965e-01
1.24840450e+00 -1.31113458e+00 8.98683667e-01 6.56485677e-01
2.33786687e-01 9.21771228e-01 6.40060484e-01 -5.41139960e-01
-1.69832826e+00 -1.21015060e+00 6.62131727e-01 -9.58180308e-01
1.03003156e+00 -5.24381816e-01 -9.66737449e-01 1.14842796e+00
3.89370620e-01 -6.33615032e-02 6.19712353e-01 4.49138671e-01
-6.00268960e-01 1.00836255e-01 -5.51439106e-01 7.10126638e-01
8.85406494e-01 -5.80331683e-01 -6.50175214e-01 5.30234218e-01
1.46768415e+00 -6.45550430e-01 -9.09623742e-01 2.25352272e-01
3.15351039e-01 -8.81532729e-01 7.35060036e-01 -1.27664280e+00
3.26574147e-01 3.52464795e-01 -4.44708586e-01 -1.70285714e+00
-5.80128431e-01 -7.78327584e-01 -3.92792881e-01 1.09060073e+00
5.53947628e-01 -6.12784028e-01 4.78573710e-01 2.83343494e-01
-5.35197139e-01 -9.27549660e-01 -4.91645187e-01 -1.13783443e+00
6.40435338e-01 -5.86565733e-01 5.73763132e-01 6.27771020e-01
-2.34635159e-01 6.45790815e-01 -2.76643664e-01 -3.25714499e-01
2.77173758e-01 -9.29899961e-02 1.00003064e+00 -7.14167476e-01
-6.27659678e-01 -3.48336697e-01 3.34511787e-01 -1.32526493e+00
5.99451363e-01 -9.70280170e-01 6.08918397e-03 -8.42573941e-01
3.56703013e-01 -4.02321607e-01 -5.60125530e-01 8.63233447e-01
-6.24761879e-01 -2.08573788e-02 2.07488984e-01 6.42505810e-02
-7.67802656e-01 3.34532976e-01 1.10445833e+00 -1.18774354e-01
-1.48319945e-01 -1.85349897e-01 -7.95674980e-01 5.72195947e-01
8.30423713e-01 -2.78797418e-01 -4.47762042e-01 -7.61910319e-01
-6.28217086e-02 -1.72317907e-01 -1.56791229e-02 -9.12174344e-01
5.44853695e-03 -2.46228933e-01 -7.30313212e-02 -3.73190314e-01
2.73764014e-01 -1.61054999e-01 -4.02889013e-01 5.30821741e-01
-6.25853360e-01 6.31019473e-01 8.48658800e-01 4.37653333e-01
1.16471395e-01 -2.34737098e-01 6.08852625e-01 -6.70832321e-02
-1.17486012e+00 1.06658876e-01 -2.36915275e-01 5.34426451e-01
7.52155125e-01 2.60652840e-01 -9.49548900e-01 -6.92749172e-02
-4.72336918e-01 3.70606869e-01 1.73963055e-01 1.04437757e+00
2.67981529e-01 -1.18261600e+00 -7.41832256e-01 3.87964249e-01
5.47358096e-01 -3.53949815e-01 2.11614296e-01 7.38019347e-01
-2.60841787e-01 1.11767471e+00 6.87801614e-02 -7.13482082e-01
-1.29309738e+00 7.38303304e-01 9.05735195e-02 -8.08901429e-01
-1.69939727e-01 1.09921241e+00 4.44470197e-01 -6.75221980e-01
2.45247141e-01 -8.15350056e-01 3.69873434e-01 -2.44870111e-01
5.87934673e-01 5.52329421e-01 4.08853203e-01 -3.03968430e-01
-2.86504388e-01 1.65770456e-01 -6.04917824e-01 1.75700009e-01
8.30790341e-01 2.09779069e-01 -1.45634720e-02 6.76733851e-01
1.19964373e+00 -5.96750341e-02 -1.08072841e+00 -3.38061005e-01
1.24308154e-01 -1.84488297e-02 2.33882681e-01 -1.10535955e+00
-6.79322958e-01 7.98822939e-01 5.03910422e-01 -1.26275644e-01
8.34394336e-01 -4.08999138e-02 1.05188894e+00 8.96822095e-01
6.42858267e-01 -7.26342976e-01 3.68188143e-01 9.98135686e-01
6.68611526e-01 -1.43043232e+00 -3.41924191e-01 7.94187039e-02
-4.83391255e-01 8.70743096e-01 1.08090413e+00 2.85779774e-01
5.51611304e-01 6.88119054e-01 -2.31152191e-03 5.09177819e-02
-1.41318083e+00 1.24828450e-01 4.48116332e-01 2.67146200e-01
7.88728178e-01 5.31004608e-01 -3.03212758e-02 8.17641675e-01
-3.92971635e-01 -1.15669511e-01 3.31101835e-01 8.69525969e-01
-5.45063615e-01 -1.07581532e+00 -5.68055093e-01 6.23708844e-01
-4.58333731e-01 -6.82757795e-01 -8.82504657e-02 1.04155576e+00
-1.14972614e-01 1.04883265e+00 -6.15850799e-02 -6.30201340e-01
2.88294613e-01 4.33313221e-01 4.10085142e-01 -9.47116077e-01
-9.76533353e-01 -2.26605907e-02 1.48618832e-01 -7.36507356e-01
5.92496514e-01 -1.42829895e-01 -1.04015326e+00 -6.47458553e-01
-3.03843737e-01 7.47519508e-02 3.34449589e-01 1.01050329e+00
4.57830518e-01 8.12699974e-01 3.43080796e-02 -5.90189636e-01
-1.30939233e+00 -1.36007690e+00 -2.86391139e-01 3.27973723e-01
1.49279848e-01 -5.23484290e-01 -3.10825557e-01 -7.39281327e-02] | [10.641461372375488, 8.341826438903809] |
1b8143c6-df17-47de-ab10-4c507a038ca9 | learning-deep-sketch-abstraction | 1804.04804 | null | http://arxiv.org/abs/1804.04804v1 | http://arxiv.org/pdf/1804.04804v1.pdf | Learning Deep Sketch Abstraction | Human free-hand sketches have been studied in various contexts including
sketch recognition, synthesis and fine-grained sketch-based image retrieval
(FG-SBIR). A fundamental challenge for sketch analysis is to deal with
drastically different human drawing styles, particularly in terms of
abstraction level. In this work, we propose the first stroke-level sketch
abstraction model based on the insight of sketch abstraction as a process of
trading off between the recognizability of a sketch and the number of strokes
used to draw it. Concretely, we train a model for abstract sketch generation
through reinforcement learning of a stroke removal policy that learns to
predict which strokes can be safely removed without affecting recognizability.
We show that our abstraction model can be used for various sketch analysis
tasks including: (1) modeling stroke saliency and understanding the decision of
sketch recognition models, (2) synthesizing sketches of variable abstraction
for a given category, or reference object instance in a photo, and (3) training
a FG-SBIR model with photos only, bypassing the expensive photo-sketch pair
collection step. | ['Timothy M. Hospedales', 'Yi-Zhe Song', 'Yongxin Yang', 'Umar Riaz Muhammad', 'Tao Xiang'] | 2018-04-13 | learning-deep-sketch-abstraction-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Muhammad_Learning_Deep_Sketch_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Muhammad_Learning_Deep_Sketch_CVPR_2018_paper.pdf | cvpr-2018-6 | ['sketch-based-image-retrieval', 'sketch-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.39989448e-01 1.73497032e-02 -1.57535553e-01 -1.41168073e-01
-2.86688745e-01 -7.51291037e-01 9.57875073e-01 -8.01025778e-02
1.61165565e-01 2.58785814e-01 2.79993534e-01 -2.30094492e-01
1.18532656e-02 -8.77451837e-01 -6.82647169e-01 -3.81667823e-01
3.80316556e-01 5.30690491e-01 1.93853885e-01 -1.04509324e-01
6.78466380e-01 1.24748945e+00 -1.61262441e+00 5.85572481e-01
4.76734787e-01 8.82992744e-01 1.73163950e-01 1.06075013e+00
-4.14942861e-01 5.58694959e-01 -6.27445221e-01 -5.64133346e-01
3.47359240e-01 -4.63000745e-01 -5.29853880e-01 3.06827843e-01
9.18528199e-01 -6.76538289e-01 -4.35181469e-01 7.16126382e-01
2.35689610e-01 8.42408538e-02 9.49668765e-01 -1.53390431e+00
-9.38160717e-01 4.60748851e-01 -3.47077340e-01 -3.14127356e-01
2.57686704e-01 2.76479363e-01 1.04869056e+00 -1.00801611e+00
8.83564591e-01 1.63863063e+00 1.26817405e-01 9.05161619e-01
-1.23482192e+00 -7.23721981e-01 5.87112308e-01 -8.50871578e-02
-1.09447992e+00 -3.48937005e-01 1.04841113e+00 -5.30904770e-01
5.23290813e-01 4.27673906e-01 7.57853150e-01 1.11861897e+00
-4.27722901e-01 1.33520985e+00 8.85411978e-01 -4.55262840e-01
3.46574426e-01 -3.16736475e-02 -1.08817950e-01 8.03772986e-01
-1.24507904e-01 -3.14787775e-02 -8.05866659e-01 -2.45121568e-01
1.52989686e+00 3.64269972e-01 9.40537527e-02 -4.14749652e-01
-1.19109356e+00 6.21209443e-01 5.06888628e-01 -4.91108224e-02
-3.38361949e-01 4.07515079e-01 2.08644778e-01 2.37532988e-01
9.25543979e-02 7.10246682e-01 -8.68290439e-02 2.95952380e-01
-1.06939816e+00 6.25056088e-01 5.41306973e-01 1.26292145e+00
5.35242498e-01 7.96389282e-02 -7.26480186e-01 9.02457356e-01
-2.62440424e-02 5.65530300e-01 5.10673597e-02 -1.03850746e+00
1.95473388e-01 6.54397130e-01 2.55006433e-01 -7.42129982e-01
3.23695958e-01 3.09029728e-01 -7.93181121e-01 6.44363463e-01
3.99819940e-01 3.76415223e-01 -9.85904157e-01 1.58229709e+00
-6.57863319e-02 1.07355021e-01 -3.72312427e-01 9.62046981e-01
6.41488910e-01 6.07162058e-01 2.62343436e-01 4.35603380e-01
1.41102076e+00 -1.02970719e+00 -2.79547840e-01 -1.14109762e-01
-1.32033527e-01 -6.17331564e-01 1.53546405e+00 5.08639574e-01
-1.40417588e+00 -7.28982389e-01 -1.04870701e+00 -4.72311556e-01
-4.62777883e-01 5.09656131e-01 5.96494913e-01 3.40102881e-01
-8.49087000e-01 8.48288834e-01 -4.11198348e-01 -3.97670120e-01
8.30722809e-01 4.53564264e-02 -1.61075801e-01 -1.05431348e-01
-6.16230428e-01 6.23950899e-01 -1.37743130e-01 -1.42365053e-01
-9.51637149e-01 -8.50638986e-01 -3.67695749e-01 4.57385212e-01
3.14298421e-01 -9.35570598e-01 9.61113095e-01 -1.25653780e+00
-1.69485211e+00 8.13682795e-01 -1.37964323e-01 -8.72569308e-02
7.07107425e-01 -2.66499430e-01 1.81124676e-02 1.16608001e-01
-1.90052554e-01 1.11854446e+00 1.53633320e+00 -1.37031829e+00
-5.33106089e-01 -3.62751633e-01 2.11708635e-01 1.88063830e-01
4.18079272e-03 -2.24155694e-01 -4.96960193e-01 -1.07887506e+00
-1.21567987e-01 -8.82600963e-01 -4.82521839e-02 9.36521053e-01
-4.38566297e-01 -4.21119243e-01 8.14560056e-01 -5.45434177e-01
8.31425130e-01 -1.94552720e+00 5.29979289e-01 2.33958110e-01
2.07564503e-01 3.91372889e-01 -5.24400532e-01 7.60997176e-01
2.77805001e-01 2.39653766e-01 -3.32899876e-02 -2.99373657e-01
2.16028944e-01 1.04965702e-01 -1.09433734e+00 -1.51936010e-01
6.19222462e-01 1.21753526e+00 -1.04381466e+00 -2.56233752e-01
3.83734345e-01 2.95407355e-01 -5.13835728e-01 5.73661983e-01
-5.45358837e-01 1.20254025e-01 -5.53559184e-01 7.57594168e-01
6.14219546e-01 3.98261435e-02 9.46827009e-02 -2.59872437e-01
1.35922626e-01 -1.39508262e-01 -1.10908580e+00 1.66577959e+00
-6.27710879e-01 6.98125303e-01 -1.06590711e-01 -3.90901387e-01
9.43696797e-01 -7.86124691e-02 -2.32824400e-01 -6.10969365e-01
-3.54395509e-01 -5.03031313e-02 -4.33775097e-01 -2.31655929e-02
6.40029550e-01 -1.02098770e-01 -1.51284388e-03 7.97053099e-01
-2.01190978e-01 -5.29853821e-01 2.77769864e-02 3.49234670e-01
8.68983686e-01 1.97024822e-01 6.70332685e-02 -2.50906259e-01
4.51294661e-01 -4.07312274e-01 -8.64178389e-02 1.03554714e+00
1.24754235e-01 7.54966915e-01 6.00443602e-01 -8.02290738e-01
-1.32825017e+00 -1.48451412e+00 4.10441786e-01 1.27148640e+00
2.46474266e-01 -2.14294598e-01 -7.57827520e-01 -6.22641206e-01
3.17220658e-01 7.54590452e-01 -5.97141385e-01 -1.61513418e-01
-6.05639458e-01 3.52310836e-01 6.21934652e-01 6.95980072e-01
1.46597922e-01 -1.66396177e+00 -8.04193795e-01 -1.58473670e-01
3.89391571e-01 -7.50904918e-01 -8.57135952e-01 -3.97740602e-01
-8.15430820e-01 -7.97842860e-01 -1.00979125e+00 -7.02170432e-01
1.04770052e+00 4.04595226e-01 1.24830759e+00 6.68537199e-01
-4.91429090e-01 6.05681360e-01 -1.41272336e-01 -4.21439141e-01
-4.05522287e-01 8.05455819e-02 -2.14565337e-01 1.13901049e-01
-1.46371275e-01 -4.71033901e-01 -9.19852912e-01 2.67603308e-01
-1.20532930e+00 4.76929963e-01 7.45900631e-01 9.02925134e-01
4.67716664e-01 -6.18869066e-01 6.68084621e-01 -8.44740748e-01
9.87000525e-01 1.51121259e-01 -4.90005493e-01 7.73971677e-01
-3.30256522e-01 4.03633922e-01 6.71187937e-01 -6.46904111e-01
-9.68125284e-01 9.96127129e-02 1.17243171e-01 -8.14645231e-01
-1.63688913e-01 -4.02355015e-01 -2.14404047e-01 3.91946547e-02
3.75400245e-01 2.63952553e-01 -1.45161241e-01 -3.44491631e-01
1.02385759e+00 5.20264626e-01 4.94414300e-01 -1.06544232e+00
7.92722762e-01 5.39842963e-01 1.86157308e-03 -1.08001995e+00
-5.84122181e-01 -1.54214725e-01 -8.41785192e-01 -2.84340441e-01
6.88453496e-01 -6.02795303e-01 -8.54024708e-01 3.50254178e-01
-1.18135405e+00 -6.18266284e-01 -4.25533742e-01 -4.33352768e-01
-7.33916461e-01 2.44225740e-01 -4.16257024e-01 -9.99416351e-01
-4.30031687e-01 -1.10755777e+00 1.74227953e+00 1.85475931e-01
-3.83026838e-01 -4.84255403e-01 -2.48113886e-01 -5.70864640e-02
3.40293646e-01 4.76896055e-02 1.37929523e+00 -3.71879688e-03
-1.15460992e+00 -8.48634169e-02 -6.75195813e-01 9.18490142e-02
8.38083327e-02 3.31467867e-01 -9.61254895e-01 -1.67732924e-01
-8.40844035e-01 -5.01557589e-01 7.64282346e-01 1.88167840e-01
1.72054756e+00 -3.87316406e-01 -1.51851907e-01 4.52376276e-01
1.16153383e+00 -2.49682739e-02 8.93985033e-01 -2.27091849e-01
7.96473742e-01 5.61203897e-01 4.03436184e-01 3.43892872e-01
-2.76168138e-02 6.97701812e-01 5.04476093e-02 -7.55374804e-02
-5.16847908e-01 -1.03010976e+00 -3.85663621e-02 1.95775956e-01
-8.13628808e-02 -2.12977707e-01 -4.07789648e-01 4.77237672e-01
-1.82570553e+00 -1.13263917e+00 5.29929042e-01 2.38310242e+00
4.12722528e-01 3.08917593e-02 2.80183226e-01 -1.19780123e-01
5.84406853e-01 2.68911809e-01 -7.67690837e-01 -7.48071969e-01
5.44965789e-02 3.40191990e-01 1.98465809e-01 4.50460911e-01
-6.58921063e-01 1.33845150e+00 6.22943354e+00 8.62975836e-01
-1.14211380e+00 -5.50816536e-01 6.31346643e-01 3.03370319e-02
-5.59725761e-01 6.52107522e-02 -4.27753150e-01 3.41421425e-01
1.46638781e-01 2.02659816e-02 1.03802669e+00 7.34869182e-01
-2.71094650e-01 2.46632934e-01 -1.62486017e+00 1.10130811e+00
2.92699225e-02 -1.65690839e+00 1.14725876e+00 -2.70317346e-01
6.13011658e-01 -7.03588784e-01 1.25736162e-01 2.29678646e-01
1.73170909e-01 -1.19879413e+00 9.91692305e-01 8.98288071e-01
9.64384317e-01 -6.11221671e-01 -2.60435045e-01 2.96759963e-01
-1.10016131e+00 7.53968768e-03 -3.64555389e-01 -4.57753576e-02
-5.86114861e-02 -4.76630479e-02 -6.69583440e-01 1.98089778e-01
1.69366792e-01 3.72409642e-01 -5.05276322e-01 5.43277800e-01
-3.60183090e-01 1.65870169e-03 -3.49807786e-03 -2.91079909e-01
8.41468275e-02 -2.76045173e-01 3.60826939e-01 1.19821501e+00
3.06018174e-01 3.73062760e-01 2.49849148e-02 1.36572659e+00
-3.48277628e-01 -2.98923850e-01 -5.98999560e-01 -3.77735972e-01
7.50642180e-01 1.18373835e+00 -9.38452542e-01 -6.36335075e-01
-7.29966462e-02 1.47070694e+00 5.24783373e-01 4.20834601e-01
-4.13764685e-01 -4.35768872e-01 8.08652818e-01 2.44286850e-01
4.72663343e-01 -1.47538885e-01 -5.01473844e-01 -1.09188855e+00
-1.06612913e-01 -6.83603048e-01 1.73772648e-02 -1.29463923e+00
-1.47205615e+00 4.80176836e-01 -1.38071463e-01 -9.89396334e-01
-1.12413809e-01 -8.01271617e-01 -9.56141472e-01 9.60874438e-01
-1.30712426e+00 -1.50374484e+00 -3.88622701e-01 3.56718272e-01
1.03660440e+00 -1.62508145e-01 7.22350240e-01 -1.05335884e-01
-2.12707832e-01 7.10249960e-01 -3.39941502e-01 1.05383173e-01
6.26679659e-01 -1.35895181e+00 1.03509998e+00 5.08785963e-01
4.63546634e-01 5.87373257e-01 4.45911318e-01 -6.34441257e-01
-1.68232012e+00 -9.59804535e-01 5.74905336e-01 -6.67667627e-01
3.52958232e-01 -6.45685554e-01 -8.64441633e-01 3.80238980e-01
-3.58166844e-02 -1.03399180e-01 2.47208737e-02 -5.06133698e-02
-7.11335659e-01 3.13176662e-02 -1.05154490e+00 1.25589395e+00
1.32102275e+00 -9.12538528e-01 -4.71769094e-01 7.12631866e-02
3.60505283e-01 -1.65739208e-01 -6.31809354e-01 -1.67311028e-01
1.30599988e+00 -7.45716989e-01 1.48234904e+00 -1.00835431e+00
7.73441017e-01 -2.42741495e-01 2.44376785e-03 -1.24561954e+00
-4.81658876e-01 -4.61418480e-01 -1.76868454e-01 9.79879618e-01
-6.13962263e-02 -2.44969921e-03 9.26628351e-01 7.93935776e-01
3.11427593e-01 -7.99262702e-01 -4.19421285e-01 -5.35933912e-01
3.97940874e-02 9.80193987e-02 9.51867878e-01 5.77101290e-01
-2.33517945e-01 2.41789639e-01 -4.39541936e-01 4.29035351e-03
6.03565991e-01 6.70849621e-01 1.14104939e+00 -1.33352745e+00
-1.64293140e-01 -9.34790552e-01 -1.80321485e-01 -1.32506788e+00
1.60413012e-01 -6.88503981e-01 -3.10760252e-02 -1.50585735e+00
1.70469955e-01 -6.05564654e-01 -4.45922539e-02 2.73927897e-01
-2.99157441e-01 6.82425275e-02 7.92379081e-01 3.11893016e-01
-3.59145075e-01 4.01000589e-01 1.44818509e+00 -3.19461972e-01
-1.09568216e-01 -3.92025672e-02 -6.21957123e-01 6.27275825e-01
3.55881870e-01 -5.25361858e-02 -5.43216705e-01 -4.39058006e-01
7.53169432e-02 3.49633962e-01 6.80128098e-01 -4.78805929e-01
2.12257639e-01 -5.25392771e-01 6.75719321e-01 -6.50230587e-01
3.73669565e-01 -6.81196392e-01 8.02932158e-02 3.94010544e-01
-8.32110941e-01 1.58173010e-01 4.77432571e-02 5.58815241e-01
2.23287895e-01 -1.71147749e-01 6.55749619e-01 -3.11028361e-01
-7.30047405e-01 5.24419606e-01 -1.36852160e-01 -1.70346349e-01
6.62301600e-01 -2.22491428e-01 -3.26776356e-01 -2.53062040e-01
-4.90485817e-01 -6.84590340e-02 7.53369868e-01 8.00067961e-01
9.50000703e-01 -1.64100742e+00 -6.10068321e-01 4.82388943e-01
4.12292629e-01 -1.45085171e-01 2.77788609e-01 -2.23927200e-01
-4.76192296e-01 1.96537629e-01 -6.30412936e-01 -3.57478559e-01
-1.11778438e+00 8.30465019e-01 -1.90949235e-02 -4.43733744e-02
-7.77725637e-01 6.46689415e-01 7.75968254e-01 -2.47800693e-01
3.90631199e-01 -5.39907098e-01 3.77318501e-01 -2.43515372e-01
7.04077363e-01 4.49110627e-01 -2.51663446e-01 -6.87975883e-02
-2.48021074e-02 6.81049287e-01 -1.57818913e-01 -1.69647470e-01
1.03643775e+00 2.44844034e-01 1.41392142e-01 3.58463466e-01
8.25923443e-01 -1.07311785e-01 -1.69480145e+00 -8.82640760e-03
-5.48533686e-02 -7.93908596e-01 -2.50580668e-01 -1.01100171e+00
-7.92513669e-01 1.17500484e+00 3.10072303e-01 -2.19638273e-02
8.47045779e-01 -3.89738977e-02 7.24168003e-01 5.67337573e-01
4.39917922e-01 -9.20968056e-01 4.83538300e-01 4.25379984e-02
1.50525212e+00 -8.77225101e-01 7.79721290e-02 -1.03976160e-01
-8.71065497e-01 1.28230250e+00 4.56027269e-01 -6.26458228e-01
3.02542478e-01 1.27173036e-01 -2.14777619e-01 -2.34683886e-01
-7.31010735e-01 2.32708976e-02 6.38115525e-01 5.23792326e-01
4.61130328e-02 3.33032727e-01 2.37613022e-01 3.99246812e-01
1.04113504e-01 9.83797610e-02 2.81792223e-01 6.38523459e-01
-3.23823154e-01 -1.16907239e+00 -8.38235617e-02 6.35230422e-01
2.10211575e-01 -2.76323844e-04 -9.54899251e-01 4.71933991e-01
-1.94197103e-01 1.63273364e-01 7.65460208e-02 5.54678738e-02
5.15326560e-01 1.78514123e-01 8.88169229e-01 -4.55588967e-01
-4.46945071e-01 -4.15069580e-01 -3.05324972e-01 -5.27615130e-01
2.24477295e-02 -5.09640694e-01 -8.10487270e-01 -2.65172005e-01
6.25597164e-02 -2.22240016e-01 5.99380255e-01 6.43928468e-01
4.87491250e-01 3.56231898e-01 4.18795228e-01 -1.37076485e+00
-4.80326563e-01 -4.95344967e-01 -6.79349542e-01 6.95978880e-01
3.82080853e-01 -6.23590589e-01 6.30761823e-03 1.44653320e-01] | [11.71695613861084, 0.40357282757759094] |
549171ec-bb1a-497a-bb94-dba5ad5d9686 | degpr-deep-guided-posterior-regularization | 2304.00741 | null | https://arxiv.org/abs/2304.00741v1 | https://arxiv.org/pdf/2304.00741v1.pdf | DeGPR: Deep Guided Posterior Regularization for Multi-Class Cell Detection and Counting | Multi-class cell detection and counting is an essential task for many pathological diagnoses. Manual counting is tedious and often leads to inter-observer variations among pathologists. While there exist multiple, general-purpose, deep learning-based object detection and counting methods, they may not readily transfer to detecting and counting cells in medical images, due to the limited data, presence of tiny overlapping objects, multiple cell types, severe class-imbalance, minute differences in size/shape of cells, etc. In response, we propose guided posterior regularization (DeGPR), which assists an object detector by guiding it to exploit discriminative features among cells. The features may be pathologist-provided or inferred directly from visual data. We validate our model on two publicly available datasets (CoNSeP and MoNuSAC), and on MuCeD, a novel dataset that we contribute. MuCeD consists of 55 biopsy images of the human duodenum for predicting celiac disease. We perform extensive experimentation with three object detection baselines on three datasets to show that DeGPR is model-agnostic, and consistently improves baselines obtaining up to 9% (absolute) mAP gains. | ['Mausam', 'Prathosh AP', 'Lalita Mehra', 'Govind Makharia', 'Prasenjit Das', 'Chirag Mohapatra', 'Aayush Kumar Tyagi'] | 2023-04-03 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tyagi_DeGPR_Deep_Guided_Posterior_Regularization_for_Multi-Class_Cell_Detection_and_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tyagi_DeGPR_Deep_Guided_Posterior_Regularization_for_Multi-Class_Cell_Detection_and_CVPR_2023_paper.pdf | cvpr-2023-1 | ['medical-object-detection', 'cell-detection'] | ['computer-vision', 'computer-vision'] | [ 2.40880758e-01 -3.63465190e-01 -1.97151765e-01 -7.28739798e-02
-9.71348941e-01 -5.77331543e-01 4.12878841e-01 6.95337534e-01
-7.70068288e-01 6.74566031e-01 -2.17892770e-02 -1.43518403e-01
3.55839789e-01 -4.51080054e-01 -4.30116564e-01 -9.21014369e-01
-1.54599935e-01 8.31355035e-01 3.46162856e-01 2.58898914e-01
1.30469754e-01 5.55774927e-01 -9.82807994e-01 5.09611011e-01
6.49703324e-01 9.03914869e-01 2.42925689e-01 1.02756882e+00
2.20003035e-02 9.12056923e-01 -3.55538458e-01 -9.24160480e-02
8.38400498e-02 -2.38513529e-01 -6.78996563e-01 2.15924248e-01
3.58294487e-01 -3.18855077e-01 4.68519293e-02 1.16851509e+00
5.82758367e-01 -3.93705755e-01 1.25796533e+00 -7.78282940e-01
-5.47801197e-01 1.55915767e-01 -1.03351271e+00 5.91945946e-01
-7.74738938e-02 3.50657612e-01 7.65633523e-01 -6.75503075e-01
5.69762468e-01 1.19996917e+00 9.10857618e-01 6.49463534e-01
-1.52848434e+00 -4.43733901e-01 -2.37637356e-01 -2.08296195e-01
-1.30092955e+00 -2.53187835e-01 9.48517397e-02 -6.70536816e-01
7.86119521e-01 3.61777246e-01 7.87455380e-01 9.31615829e-01
4.09061015e-01 1.10784686e+00 1.19166970e+00 -4.84683305e-01
2.20782936e-01 3.35820198e-01 1.33291230e-01 1.00958788e+00
6.63841784e-01 -3.17044137e-03 -1.78410187e-02 -3.78145188e-01
1.03442514e+00 2.87506372e-01 -2.73800701e-01 -4.06756610e-01
-1.42563283e+00 9.68013465e-01 3.41879755e-01 1.41370326e-01
-1.99716166e-01 1.27176672e-01 6.58112288e-01 -2.39450540e-02
4.08562183e-01 3.71244252e-01 -3.00897062e-01 3.34648401e-01
-6.84434295e-01 5.40205538e-02 8.57143044e-01 7.95553744e-01
1.59254387e-01 -3.72924477e-01 -5.19391358e-01 9.57065105e-01
2.30304942e-01 2.25674942e-01 6.57385528e-01 -6.88900411e-01
-2.11079657e-01 8.72633994e-01 5.41118793e-02 -8.90049398e-01
-7.42878973e-01 -4.96839285e-01 -1.11509275e+00 2.25707367e-01
7.33620644e-01 3.83273624e-02 -1.01740575e+00 1.31633997e+00
3.59094918e-01 -2.79747434e-02 -3.45471770e-01 1.15357685e+00
8.68494511e-01 3.17247957e-01 2.75349468e-01 -2.04147503e-01
1.86395311e+00 -9.01356399e-01 -3.47556770e-01 -3.77652407e-01
1.11735392e+00 -6.10833108e-01 1.02931535e+00 4.61101532e-01
-8.13610375e-01 1.11296877e-01 -9.08726633e-01 -2.52753526e-01
-2.27214977e-01 5.35176873e-01 8.50314796e-01 4.09913927e-01
-6.60963416e-01 8.53963420e-02 -1.24855149e+00 -5.09734511e-01
8.90536666e-01 2.80976653e-01 -2.68417835e-01 -1.78355590e-01
-3.81487221e-01 7.67327845e-01 2.37311348e-01 -7.37443268e-02
-8.44735801e-01 -9.11779702e-01 -5.41063488e-01 -4.79060039e-02
1.58364713e-01 -9.56889868e-01 1.10327852e+00 -7.16072679e-01
-9.18248355e-01 1.49702287e+00 -4.25682329e-02 -4.87339109e-01
1.00243950e+00 3.16933662e-01 -1.69363082e-03 2.11346030e-01
1.24352805e-01 8.47588062e-01 2.73748994e-01 -1.12518847e+00
-8.47313583e-01 -4.47618037e-01 -2.25513622e-01 -2.03034226e-02
-9.92165953e-02 3.54822353e-02 -4.53010768e-01 -6.79311752e-01
-3.76970731e-02 -9.10020888e-01 -5.03494561e-01 6.16882205e-01
-5.94191968e-01 2.31055878e-02 4.39211845e-01 -4.93697017e-01
6.85510099e-01 -2.17189407e+00 -2.24892139e-01 -1.60970509e-01
5.10498524e-01 1.12951413e-01 5.71463741e-02 -2.15278342e-01
3.73195231e-01 -3.39108855e-02 -1.58870071e-01 -5.06866395e-01
-1.61363930e-01 2.23395973e-01 2.08046108e-01 1.01004696e+00
1.21694855e-01 7.71168947e-01 -1.06862593e+00 -1.03320289e+00
1.78973064e-01 6.58089101e-01 -5.74356198e-01 -1.05200961e-01
-1.11712337e-01 5.37168026e-01 -2.67378598e-01 1.05310643e+00
6.06439233e-01 -9.81326759e-01 2.95810938e-01 -2.71319568e-01
3.14981878e-01 -3.14371616e-01 -1.03878284e+00 1.30060804e+00
-2.65327364e-01 5.12691200e-01 1.87615126e-01 -8.07296753e-01
4.59222645e-01 1.12155825e-01 4.32655007e-01 -3.89411986e-01
4.39769357e-01 4.68279988e-01 8.06115009e-03 -4.28615481e-01
1.40518218e-01 -3.84177476e-01 8.60268697e-02 1.36146531e-01
1.61687329e-01 6.09049909e-02 5.65318704e-01 6.78202659e-02
1.20196152e+00 -6.51481569e-01 8.08027208e-01 -5.56897998e-01
4.38707918e-01 2.37674415e-01 7.77717531e-01 7.52314568e-01
-6.65397644e-01 7.65657306e-01 7.65080273e-01 -7.12925971e-01
-9.32147443e-01 -8.64601731e-01 -5.97565472e-01 9.21750426e-01
1.83772057e-01 1.02123506e-01 -2.60913312e-01 -7.82619119e-01
2.68258899e-01 2.38907620e-01 -8.95512760e-01 1.27800003e-01
-4.78183895e-01 -1.48279095e+00 4.50327396e-01 7.35037863e-01
-1.05998656e-02 -8.45965803e-01 -7.24797428e-01 3.77349466e-01
-9.32288244e-02 -1.14897585e+00 -6.01705313e-01 3.55208755e-01
-9.71878052e-01 -1.47904682e+00 -9.92017090e-01 -1.18232596e+00
1.04905605e+00 1.90657854e-01 1.33581543e+00 3.34423214e-01
-1.08878922e+00 -3.10938153e-02 8.95682424e-02 -7.27354705e-01
-6.76783323e-01 -8.64805058e-02 -9.91538465e-02 -2.84112275e-01
6.41894758e-01 -8.95298272e-02 -9.76238012e-01 2.61353582e-01
-5.49220264e-01 9.21813548e-02 7.78475404e-01 1.23486435e+00
1.07425332e+00 -4.31126326e-01 3.92234862e-01 -1.28239942e+00
3.31576049e-01 -2.76508063e-01 -8.01946044e-01 1.56219274e-01
-3.75522494e-01 -3.13792557e-01 4.45447385e-01 -6.10212803e-01
-4.48695987e-01 1.11804858e-01 8.67776871e-02 -3.38995188e-01
-1.39690891e-01 1.79243460e-01 6.27180874e-01 -1.73937246e-01
8.33392799e-01 2.04717368e-01 1.68604836e-01 -2.62246996e-01
-1.75165161e-01 4.03426200e-01 4.85661596e-01 -1.45902395e-01
-4.33005542e-02 9.07792032e-01 1.21268511e-01 -5.19697309e-01
-9.05093610e-01 -8.82332444e-01 -3.28964591e-01 1.90292358e-01
6.15085781e-01 -1.03784394e+00 -7.78956532e-01 4.94860142e-01
-1.01412559e+00 -4.18684214e-01 -1.80846453e-01 5.88277340e-01
-5.13246536e-01 3.40933830e-01 -1.45076489e+00 -6.78729475e-01
-7.03841746e-01 -1.22798896e+00 1.24060500e+00 8.51974189e-02
-2.56901920e-01 -1.22212911e+00 2.44005919e-01 1.81953132e-01
3.31870437e-01 5.46703100e-01 1.12564385e+00 -8.07715595e-01
-4.46730822e-01 -6.00134254e-01 -5.16733229e-01 1.41264558e-01
1.50028110e-01 9.78862271e-02 -7.06755161e-01 -5.86206794e-01
-3.19971591e-01 -6.29866123e-01 1.23906529e+00 8.04733276e-01
1.34193957e+00 -3.35538566e-01 -9.11386728e-01 7.43119240e-01
1.67865491e+00 -1.78267449e-01 2.83562690e-01 5.87749556e-02
5.51860034e-01 2.87930220e-01 3.02538484e-01 4.25712466e-01
2.17388451e-01 3.53740394e-01 4.82246727e-01 -4.96987432e-01
-2.32716009e-01 1.57997489e-01 -2.01735586e-01 7.91423142e-01
3.42308951e-04 -2.31621712e-01 -9.12244022e-01 8.97836924e-01
-1.53009844e+00 -5.45418441e-01 -2.25051492e-01 1.93708396e+00
1.14426315e+00 -6.32930920e-02 2.63049036e-01 -1.01001188e-01
6.76992297e-01 -3.65101486e-01 -5.08915365e-01 1.07193172e-01
-1.20085023e-01 -1.94995865e-01 7.27349818e-01 3.33453208e-01
-1.30895007e+00 4.78739500e-01 6.55774641e+00 1.04396379e+00
-1.14207232e+00 2.05526412e-01 1.08053398e+00 -3.51259321e-01
3.22681874e-01 -5.75335026e-01 -8.76262546e-01 4.68452036e-01
1.75494760e-01 6.02104627e-02 7.76808634e-02 7.07445323e-01
-8.65463838e-02 -3.07755321e-01 -1.42386663e+00 1.14011884e+00
1.51286265e-02 -1.51354384e+00 -1.43920975e-02 7.16477484e-02
7.98534214e-01 2.49173939e-01 -8.92258659e-02 4.58896667e-01
4.08830613e-01 -1.04439735e+00 3.51164877e-01 2.20509723e-01
9.34573472e-01 -4.53433573e-01 1.19539869e+00 4.37200099e-01
-7.96005964e-01 3.65157239e-02 -6.07624948e-01 4.12005484e-01
-6.13376983e-02 7.73374081e-01 -1.23642099e+00 -2.20235705e-01
5.00779271e-01 4.39961761e-01 -5.31610847e-01 1.48529589e+00
2.66861677e-01 5.54348171e-01 -3.71028334e-01 -2.57310808e-01
8.41828883e-02 1.01396658e-01 4.16468948e-01 1.67829728e+00
1.41774222e-01 -7.84513652e-02 2.86938548e-01 8.98779988e-01
-2.51277357e-01 2.91423708e-01 -1.75549105e-01 1.02893502e-01
2.90614516e-01 1.76679516e+00 -1.31964338e+00 -3.52239132e-01
-4.01779056e-01 7.26705968e-01 2.91282713e-01 1.29844904e-01
-8.24405015e-01 -1.67215131e-02 2.66781926e-01 2.46226802e-01
1.28691778e-01 2.65760839e-01 -3.24070334e-01 -9.94973004e-01
-3.30045491e-01 -7.46127665e-01 6.61475539e-01 -1.68902606e-01
-1.77488458e+00 1.39695555e-01 -4.79674220e-01 -1.27108634e+00
1.03728138e-01 -1.05677748e+00 -4.61915880e-01 5.37465334e-01
-1.50113094e+00 -1.05035079e+00 -5.31374991e-01 1.89099461e-01
4.09054488e-01 -3.46069783e-02 9.59032893e-01 4.98947620e-01
-4.53001171e-01 9.07731175e-01 2.14483380e-01 4.61869299e-01
7.76788175e-01 -1.55209827e+00 -1.24034621e-01 3.28392386e-01
-4.04996425e-03 2.78905660e-01 6.04455948e-01 -2.56920397e-01
-1.22465909e+00 -1.22476482e+00 4.75670695e-01 -5.04143000e-01
5.56862354e-01 -4.73198593e-01 -8.90679538e-01 7.18546569e-01
-2.85531193e-01 8.29223454e-01 8.95487607e-01 -1.32919371e-01
-1.02353074e-01 1.25968605e-01 -1.50439835e+00 5.43569446e-01
6.38589203e-01 -1.05586976e-01 -7.97283873e-02 7.80954957e-01
2.52241373e-01 -6.90019369e-01 -8.41711462e-01 3.97733301e-01
5.41796982e-01 -8.82075846e-01 8.60564113e-01 -4.84425336e-01
2.96799034e-01 -1.09680161e-01 -2.57611033e-02 -1.09693205e+00
-4.61393923e-01 1.56634208e-02 -1.66355491e-01 7.06779897e-01
3.85526031e-01 -6.62295163e-01 1.05194330e+00 -4.69244868e-02
-8.56322423e-02 -1.08555150e+00 -9.38855171e-01 -6.78284585e-01
3.76062691e-01 8.73875022e-02 1.30818397e-01 1.11791754e+00
1.62796482e-01 1.94089502e-01 2.27616690e-02 1.04497338e-03
6.43867493e-01 1.16213635e-01 6.33355975e-01 -1.20336533e+00
-5.83904922e-01 -5.92793405e-01 -6.18456006e-01 -8.85595262e-01
-4.15618896e-01 -1.06893623e+00 2.56192740e-02 -1.50232089e+00
9.31831539e-01 -5.28629601e-01 -3.81381750e-01 4.11314934e-01
-3.87325704e-01 6.02382720e-01 -1.11726336e-01 3.80111545e-01
-8.36183190e-01 -9.43264365e-03 1.55182719e+00 -4.42983806e-01
2.59196479e-02 -1.20120868e-01 -6.79445446e-01 1.03710687e+00
5.40339828e-01 -7.28346109e-01 1.84266463e-01 -2.73891449e-01
1.55290917e-01 9.10589322e-02 7.18676925e-01 -9.97128308e-01
3.08888972e-01 8.94171894e-02 7.71092594e-01 -5.84254742e-01
1.21875495e-01 -4.21976924e-01 -1.46662876e-01 9.89600003e-01
-2.02844039e-01 -4.80356604e-01 2.79996693e-01 7.25404680e-01
-9.48743224e-02 -1.11755311e-01 1.32394648e+00 -4.65091228e-01
-2.23896414e-01 3.07402283e-01 -4.04488057e-01 2.61436522e-01
9.00687635e-01 -3.74608070e-01 -4.23884988e-01 2.12460011e-01
-6.72759354e-01 4.71013248e-01 4.68700558e-01 -1.45162910e-01
1.76597565e-01 -1.18116951e+00 -1.07025540e+00 4.91502555e-03
5.10032654e-01 3.75981033e-01 2.56275594e-01 1.45025849e+00
-9.86479104e-01 4.73438829e-01 -6.15640283e-02 -1.13869798e+00
-1.24170280e+00 5.37202954e-01 5.91446459e-01 -7.42573977e-01
-6.97866857e-01 1.26155412e+00 6.34102523e-01 -4.45573300e-01
1.17181435e-01 -7.88819790e-01 -7.61158541e-02 -1.65626016e-02
5.73055625e-01 3.46471548e-01 1.64237037e-01 -2.16545984e-01
-4.95770782e-01 2.90919542e-01 -5.70676625e-01 5.89893460e-01
9.78963196e-01 1.86708376e-01 1.20332979e-01 4.90987569e-01
1.10190618e+00 -4.25162502e-02 -1.23395717e+00 -2.32920006e-01
-1.55561730e-01 -4.08943564e-01 1.00425005e-01 -8.91478777e-01
-1.12565851e+00 7.61608779e-01 8.70210588e-01 1.30051911e-01
8.12209785e-01 1.22833267e-01 6.54064596e-01 1.66058227e-01
5.62180042e-01 -8.64890933e-01 2.80912928e-02 1.88892096e-01
6.14298761e-01 -1.60356224e+00 1.83161899e-01 -4.75246131e-01
-5.00348806e-01 9.35360014e-01 5.72959304e-01 -1.32869259e-01
3.29459935e-01 6.52502477e-01 1.90634191e-01 -1.91430673e-01
-9.19786930e-01 1.43882468e-01 -3.90594848e-03 4.44363654e-01
6.68796599e-01 1.63320512e-01 -2.90577859e-01 5.89060843e-01
2.96172053e-01 2.79177248e-01 4.65112299e-01 8.57795119e-01
-4.67921644e-01 -4.13790852e-01 -3.59899312e-01 1.03359759e+00
-8.72481227e-01 7.06641376e-02 -9.45508555e-02 9.77859020e-01
4.14713584e-02 3.82045835e-01 4.14856106e-01 4.17543143e-01
6.37748837e-02 -5.14412761e-01 6.82691157e-01 -5.92505455e-01
-5.15966654e-01 3.22755545e-01 -1.23332992e-01 -2.53043115e-01
-1.42658338e-01 -7.19893098e-01 -1.41835558e+00 -3.50343555e-01
-5.57687104e-01 -4.75347131e-01 6.56301016e-03 6.54316723e-01
4.96783201e-03 4.74388093e-01 1.84422970e-01 -5.53123534e-01
-6.81659639e-01 -1.11539662e+00 -9.85135853e-01 4.46511358e-01
3.91746640e-01 -6.55924320e-01 -5.53068817e-01 2.16919541e-01] | [14.991652488708496, -3.0605175495147705] |
f49c76d7-160e-428b-9542-6d4372600314 | weakly-supervised-3d-medical-image | 2302.02125 | null | https://arxiv.org/abs/2302.02125v1 | https://arxiv.org/pdf/2302.02125v1.pdf | Weakly-Supervised 3D Medical Image Segmentation using Geometric Prior and Contrastive Similarity | Medical image segmentation is almost the most important pre-processing procedure in computer-aided diagnosis but is also a very challenging task due to the complex shapes of segments and various artifacts caused by medical imaging, (i.e., low-contrast tissues, and non-homogenous textures). In this paper, we propose a simple yet effective segmentation framework that incorporates the geometric prior and contrastive similarity into the weakly-supervised segmentation framework in a loss-based fashion. The proposed geometric prior built on point cloud provides meticulous geometry to the weakly-supervised segmentation proposal, which serves as better supervision than the inherent property of the bounding-box annotation (i.e., height and width). Furthermore, we propose contrastive similarity to encourage organ pixels to gather around in the contrastive embedding space, which helps better distinguish low-contrast tissues. The proposed contrastive embedding space can make up for the poor representation of the conventionally-used gray space. Extensive experiments are conducted to verify the effectiveness and the robustness of the proposed weakly-supervised segmentation framework. The proposed framework is superior to state-of-the-art weakly-supervised methods on the following publicly accessible datasets: LiTS 2017 Challenge, KiTS 2021 Challenge, and LPBA40. We also dissect our method and evaluate the performance of each component. | ['Jing Liao', 'Yan Xu', 'Qihua Dong', 'Hao Du'] | 2023-02-04 | null | null | null | null | ['weakly-supervised-segmentation'] | ['computer-vision'] | [ 2.44709641e-01 1.57019690e-01 -1.58706367e-01 -3.20707262e-01
-6.18813694e-01 -9.42393169e-02 2.69663811e-01 2.30913237e-01
-4.20590043e-01 4.42391574e-01 6.07964136e-02 -1.39688000e-01
4.91734548e-03 -6.54612124e-01 -4.19964582e-01 -1.16311836e+00
8.60395581e-02 2.71962970e-01 5.76006174e-01 -1.47591352e-01
-2.06036642e-02 4.26455975e-01 -1.19494879e+00 -1.08207896e-01
1.26652253e+00 8.88195813e-01 4.79885966e-01 1.28287286e-01
-2.65509903e-01 4.14498955e-01 -1.73898637e-01 -1.37971848e-01
2.68457025e-01 -3.95252794e-01 -7.62304127e-01 4.85917360e-01
6.54021353e-02 -1.42392457e-01 2.35545244e-02 1.39997911e+00
3.61265004e-01 -2.94753704e-02 7.68124461e-01 -8.11027646e-01
-3.01175267e-01 2.70020634e-01 -1.02831829e+00 2.24683255e-01
-4.58497480e-02 2.28697062e-01 5.51270068e-01 -7.08951652e-01
3.95551980e-01 1.02769351e+00 6.68957651e-01 3.33213210e-01
-1.04248667e+00 -4.25803661e-01 1.57825053e-01 -2.40212604e-02
-1.41153908e+00 -1.61356390e-01 1.09103489e+00 -3.03543210e-01
1.50235265e-01 3.57647866e-01 5.53814590e-01 4.67379600e-01
1.38855249e-01 8.91438603e-01 1.13620996e+00 -2.82039404e-01
8.14947635e-02 2.80110985e-01 2.10524589e-01 1.03513002e+00
2.61074275e-01 -1.24755718e-01 3.36873502e-01 -1.99673250e-01
9.15777981e-01 1.49844795e-01 -6.16019607e-01 -5.52334666e-01
-1.30313361e+00 4.29491818e-01 7.31853664e-01 4.75167960e-01
-2.97039896e-01 -2.01729760e-01 3.28758150e-01 -3.59165877e-01
5.96865892e-01 8.34086165e-02 -2.66320527e-01 2.73649901e-01
-9.32654977e-01 -2.12636694e-01 3.50697726e-01 7.69188404e-01
6.45673454e-01 -4.38255668e-02 -2.16406628e-01 9.15963948e-01
6.28583670e-01 1.45126000e-01 5.48322439e-01 -4.46155488e-01
2.73621589e-01 8.23885620e-01 -7.29891360e-02 -1.14556825e+00
-4.93661433e-01 -7.12541461e-01 -1.21468687e+00 2.34188624e-02
6.51000142e-01 1.07247263e-01 -1.08251679e+00 1.35581470e+00
7.44650245e-01 4.53056365e-01 -8.40584114e-02 1.11915588e+00
9.08361495e-01 6.12682998e-01 1.70919910e-01 -2.60282964e-01
1.43091881e+00 -1.08590043e+00 -7.95403421e-01 -1.69735819e-01
6.41203463e-01 -7.73317516e-01 1.22622943e+00 1.11807823e-01
-1.11201930e+00 -5.15939593e-01 -1.04555428e+00 2.01708917e-02
6.08738884e-02 3.71236563e-01 5.87911606e-01 5.47809303e-01
-6.64437950e-01 5.91743112e-01 -1.07726336e+00 -9.24890935e-02
7.42620945e-01 2.11613372e-01 -1.53524294e-01 -2.39617035e-01
-8.39052379e-01 5.06655574e-01 4.02067959e-01 3.46599549e-01
-3.71862918e-01 -8.12180877e-01 -8.47472012e-01 -1.05790064e-01
4.48536605e-01 -4.53198999e-01 6.65468752e-01 -8.06437731e-01
-1.27005100e+00 9.95112836e-01 -8.63452721e-03 -1.19371332e-01
8.13578248e-01 2.06922531e-01 -5.88959455e-02 6.08065665e-01
4.85431440e-02 4.61926579e-01 6.93041503e-01 -1.51830852e+00
-3.91733527e-01 -5.67934871e-01 -2.34420955e-01 3.15508306e-01
-3.92092198e-01 -2.97306001e-01 -7.56982565e-01 -9.49778557e-01
5.08065879e-01 -9.54026580e-01 -4.67252642e-01 5.97606838e-01
-5.98408222e-01 1.84122510e-02 9.08058226e-01 -8.10989916e-01
1.14753604e+00 -2.42158747e+00 -1.62501261e-01 3.09592873e-01
1.54133573e-01 1.19208008e-01 6.48787841e-02 -4.77567792e-01
-1.53730894e-02 1.47967190e-01 -8.45802605e-01 -3.09703797e-01
-2.84369498e-01 2.40621030e-01 3.14861476e-01 8.69079411e-01
2.06315801e-01 7.55839705e-01 -1.02490902e+00 -1.14525402e+00
4.48245525e-01 4.94129449e-01 -3.31276536e-01 1.85884327e-01
-3.24592888e-02 7.25793600e-01 -7.05257773e-01 9.48330939e-01
8.95771444e-01 -3.24278295e-01 -1.62436679e-01 -5.68526328e-01
1.39610900e-03 -2.89996296e-01 -1.06619668e+00 1.73926497e+00
-3.02949369e-01 8.12688321e-02 3.47402036e-01 -1.15429521e+00
7.91916966e-01 2.12880403e-01 7.13526666e-01 -4.91087586e-01
4.13519353e-01 2.34607548e-01 -3.05980220e-02 -5.20408869e-01
6.31083399e-02 -3.15358579e-01 3.01891774e-01 1.53423056e-01
-2.35877171e-01 -2.79785782e-01 -8.90655909e-03 6.29012883e-02
6.27280295e-01 4.91983928e-02 2.79997975e-01 -6.61630630e-01
8.91997695e-01 -5.67814596e-02 8.91156554e-01 1.49016157e-01
-4.94722635e-01 8.28043342e-01 2.62054205e-01 -1.48554847e-01
-7.00247049e-01 -9.00093257e-01 -5.52823186e-01 4.13460672e-01
7.35040486e-01 4.19250391e-02 -7.55605817e-01 -9.75152731e-01
-2.27140725e-01 2.51250654e-01 -6.54090047e-01 -2.76479453e-01
-5.77282190e-01 -1.02389634e+00 2.72013903e-01 5.45530140e-01
8.54854584e-01 -8.07634950e-01 -4.55954760e-01 8.80597681e-02
-3.58491212e-01 -1.11241019e+00 -6.21003926e-01 -1.75916567e-03
-1.16928875e+00 -1.15465868e+00 -1.18662584e+00 -1.26132977e+00
1.20175946e+00 2.73390740e-01 9.58740532e-01 5.41616917e-01
-3.36858600e-01 -4.86387648e-02 -2.93621272e-01 -1.72808379e-01
-2.72500932e-01 -1.41655028e-01 -4.13504869e-01 5.17815910e-02
-4.63090874e-02 -1.81839585e-01 -9.34327006e-01 5.31659067e-01
-1.01424563e+00 2.52103209e-01 8.40143919e-01 1.14347422e+00
1.00024617e+00 2.90310502e-01 1.55275255e-01 -1.06196284e+00
1.35919005e-01 -3.34788203e-01 -3.50459367e-01 3.17225248e-01
-4.99471992e-01 -8.46451372e-02 5.30886650e-01 -3.74116153e-01
-1.23928714e+00 2.35659257e-01 -3.57569516e-01 -2.08310679e-01
-2.55303711e-01 3.82507682e-01 -2.17727244e-01 -3.72447848e-01
2.76141107e-01 3.31876367e-01 2.47230783e-01 -4.11117613e-01
1.37512237e-02 4.74697024e-01 4.76311266e-01 -5.35605848e-01
8.64321411e-01 9.24003839e-01 7.34606534e-02 -9.10270989e-01
-6.34522498e-01 -6.24019265e-01 -4.84894603e-01 -1.36710197e-01
9.11491752e-01 -6.20219588e-01 -3.87536287e-01 6.66033268e-01
-9.97167051e-01 -2.38639846e-01 -2.39606827e-01 5.11103988e-01
-3.41014773e-01 8.46467674e-01 -8.63810301e-01 -5.29606819e-01
-3.96389961e-01 -1.51850450e+00 1.03855920e+00 3.57001513e-01
1.69971392e-01 -1.12887907e+00 -2.20621362e-01 7.28002310e-01
2.80461907e-01 6.19009256e-01 1.02562118e+00 -2.64285743e-01
-3.74789864e-01 2.62153056e-02 -5.46902299e-01 5.17082095e-01
3.42081606e-01 3.46621461e-02 -8.02171111e-01 -2.37051368e-01
1.17336065e-01 -2.10199937e-01 9.18302417e-01 6.55286193e-01
1.53121817e+00 -4.93718088e-02 -4.36786264e-01 8.86701465e-01
1.41935945e+00 5.48485965e-02 6.77835941e-01 3.56501967e-01
9.12440240e-01 8.27278614e-01 9.60872591e-01 2.52686262e-01
2.76114702e-01 4.64788347e-01 4.73360538e-01 -9.45515513e-01
-1.09068215e-01 5.27179837e-02 -1.88298330e-01 9.47782278e-01
-1.56222470e-02 1.58884183e-01 -8.83669555e-01 6.21448994e-01
-1.69793475e+00 -4.46073979e-01 -2.88907200e-01 1.92717505e+00
1.12587309e+00 2.24713564e-01 -1.68882892e-01 2.29615659e-01
7.17449248e-01 9.17800441e-02 -5.13178229e-01 1.87237695e-01
-2.79059168e-02 1.45864695e-01 5.03699780e-01 2.79963315e-01
-1.26963854e+00 6.51661336e-01 4.87800694e+00 1.15870333e+00
-1.16355562e+00 -6.59608934e-03 1.10170627e+00 3.83853644e-01
-2.65031546e-01 -3.25868487e-01 -3.02580535e-01 7.51439929e-01
1.28659546e-01 2.57504642e-01 -1.86353639e-01 7.04663754e-01
2.79575050e-01 -2.04691544e-01 -8.74477029e-01 9.21741784e-01
3.78427543e-02 -1.06349242e+00 -5.29836267e-02 -1.12451455e-02
7.55789518e-01 -3.48691583e-01 1.59952670e-01 4.47146744e-02
-2.30521888e-01 -8.49465072e-01 5.41785479e-01 2.66786546e-01
7.01405644e-01 -6.66601121e-01 1.06629539e+00 2.85171121e-01
-1.28468502e+00 1.51196718e-01 -3.36172134e-01 5.18863618e-01
8.87098089e-02 9.23316240e-01 -5.50407350e-01 5.30484259e-01
6.97659731e-01 7.60317564e-01 -5.68967104e-01 1.08332038e+00
-2.80261859e-02 6.26338542e-01 -2.80865937e-01 2.57824868e-01
3.76214206e-01 -4.91665483e-01 4.93582159e-01 1.23749995e+00
1.01048857e-01 2.88706839e-01 4.39190269e-01 7.46253908e-01
-1.77664887e-02 4.32701766e-01 -6.50090128e-02 3.35923672e-01
1.29869372e-01 1.33722758e+00 -1.18856502e+00 -3.68591130e-01
-4.40695018e-01 8.26912224e-01 -2.85518408e-01 3.07640046e-01
-7.39654183e-01 -1.88786179e-01 1.53607145e-01 2.80796289e-01
-3.04984692e-02 5.84747568e-02 -4.96024996e-01 -1.16786671e+00
1.26994729e-01 -6.94442749e-01 3.60038251e-01 -5.20838261e-01
-1.21969640e+00 5.78823328e-01 -1.69243172e-01 -1.49634123e+00
3.87016892e-01 -2.46436357e-01 -7.67113030e-01 7.84993052e-01
-2.02863193e+00 -1.19739890e+00 -6.80590332e-01 5.93548000e-01
5.17132342e-01 2.02942938e-01 4.51958656e-01 4.94856745e-01
-7.13012934e-01 5.31618178e-01 -1.75092053e-02 2.41192743e-01
7.22098768e-01 -1.33208001e+00 -2.29738504e-01 7.36616492e-01
-3.74212593e-01 5.85598946e-01 5.16981781e-01 -5.26993573e-01
-9.80347395e-01 -1.25348651e+00 3.68046165e-02 1.86612636e-01
3.90678883e-01 -7.81709105e-02 -1.24987078e+00 1.99613824e-01
5.98082179e-03 4.84093904e-01 6.08475447e-01 -4.65070039e-01
2.65800536e-01 -1.34783670e-01 -1.57610941e+00 5.68055034e-01
4.83463287e-01 -3.68582942e-02 -5.41056395e-01 2.92267889e-01
5.78588426e-01 -5.70358336e-01 -7.24284232e-01 7.45602787e-01
3.52749765e-01 -8.27547491e-01 1.08583295e+00 1.21042982e-01
4.44195062e-01 -4.56948668e-01 2.92239904e-01 -1.08843160e+00
-2.55352139e-01 -1.74502805e-01 2.50225931e-01 1.21644473e+00
2.18545899e-01 -6.22435808e-01 1.15276718e+00 3.34278107e-01
-3.78167480e-01 -1.21455979e+00 -8.15459549e-01 -5.57344496e-01
5.08415885e-02 7.51253441e-02 3.41832668e-01 1.20523727e+00
-3.56974185e-01 -1.52695745e-01 7.26280957e-02 1.97039634e-01
9.58621979e-01 -2.35152170e-02 3.64810884e-01 -1.21016204e+00
-2.81180710e-01 -4.03582066e-01 -4.28902864e-01 -1.09097302e+00
-1.02868661e-01 -8.26996922e-01 2.76923656e-01 -1.50339687e+00
3.52952212e-01 -9.30772603e-01 -5.14119029e-01 4.15258735e-01
-5.26961505e-01 4.46199566e-01 -2.02109188e-01 3.51792753e-01
-3.49452883e-01 7.41610467e-01 1.75334775e+00 -4.65691656e-01
-2.54886389e-01 -3.07101116e-04 -6.90974474e-01 1.02946436e+00
8.12929094e-01 -2.30756477e-01 -4.41063583e-01 -3.72100323e-01
-3.35455686e-01 -2.61650775e-02 4.39996153e-01 -8.17388237e-01
1.65900782e-01 -2.69076019e-03 4.25723374e-01 -6.06889307e-01
1.18782692e-01 -1.11319256e+00 -4.09902722e-01 6.37292147e-01
-7.18086511e-02 -4.80043173e-01 2.52911091e-01 6.21604025e-01
-3.85941178e-01 -1.19161241e-01 1.26395941e+00 -2.37863839e-01
-4.63108897e-01 4.01309371e-01 2.14639053e-01 4.61980961e-02
1.21668816e+00 -5.08656263e-01 -2.68469118e-02 1.77845016e-01
-5.72660744e-01 4.52632964e-01 5.45455873e-01 -3.10011022e-02
7.79815853e-01 -1.01815832e+00 -7.27719724e-01 1.25292361e-01
2.03300282e-01 6.29160345e-01 4.59733963e-01 1.21755719e+00
-8.93039107e-01 -8.47860575e-02 -1.15776688e-01 -9.38179195e-01
-1.42671561e+00 2.84545481e-01 2.78119594e-01 -4.42267537e-01
-7.94599891e-01 7.62661874e-01 5.17925560e-01 -3.30669105e-01
2.64410436e-01 -7.59251893e-01 -2.77589381e-01 -3.00446749e-01
3.35090578e-01 2.72018313e-01 -1.43527088e-03 -7.26315618e-01
-5.33961833e-01 8.30414116e-01 -1.60191461e-01 3.64535451e-01
1.19449925e+00 -1.49106026e-01 -1.26592368e-01 1.71134174e-01
1.23645890e+00 -9.86421108e-02 -1.24866736e+00 -3.03328365e-01
-2.70751268e-01 -4.08846974e-01 4.97962296e-01 -5.78239679e-01
-1.41873789e+00 9.76532578e-01 8.75731885e-01 -7.67872483e-02
1.12392688e+00 -3.59612629e-02 1.17403328e+00 -1.59793347e-01
1.88758895e-01 -1.03070378e+00 1.11346163e-01 -2.17355996e-01
6.34318471e-01 -1.50874734e+00 2.92261124e-01 -9.72186148e-01
-6.25969470e-01 9.43770528e-01 5.93496323e-01 -4.30803597e-02
6.90926850e-01 1.54060781e-01 1.31077677e-01 -1.20918766e-01
-2.74343640e-02 -4.91839349e-02 4.36360776e-01 4.48954761e-01
3.78244281e-01 -1.19657062e-01 -4.93350148e-01 5.74423730e-01
3.46712530e-01 -1.77885607e-01 2.29121149e-01 8.39179099e-01
-3.10130745e-01 -7.22206771e-01 -5.04215300e-01 5.15210330e-01
-5.95193088e-01 1.19602285e-01 2.08817534e-02 6.82968080e-01
3.56337756e-01 7.96720624e-01 -5.43006472e-02 3.22050117e-02
2.53265172e-01 -5.84360838e-01 3.11684728e-01 -5.69237947e-01
-3.88092756e-01 4.47249621e-01 -5.18437445e-01 -4.18200076e-01
-5.26701152e-01 -5.14399350e-01 -1.64919138e+00 1.63500175e-01
-4.81463820e-01 9.24317986e-02 5.61308444e-01 8.97831619e-01
-1.76526204e-01 6.96680903e-01 7.72941649e-01 -6.30686522e-01
-4.72710252e-01 -7.25223243e-01 -7.03427315e-01 6.35382354e-01
3.21332067e-01 -6.35503113e-01 -4.95880127e-01 2.02707857e-01] | [14.626012802124023, -2.235063076019287] |
1084b85b-df19-4f93-ab59-1afaef19e505 | anabranch-network-for-camouflaged-object-1 | 2105.09451 | null | https://arxiv.org/abs/2105.09451v1 | https://arxiv.org/pdf/2105.09451v1.pdf | Anabranch Network for Camouflaged Object Segmentation | Camouflaged objects attempt to conceal their texture into the background and discriminating them from the background is hard even for human beings. The main objective of this paper is to explore the camouflaged object segmentation problem, namely, segmenting the camouflaged object(s) for a given image. This problem has not been well studied in spite of a wide range of potential applications including the preservation of wild animals and the discovery of new species, surveillance systems, search-and-rescue missions in the event of natural disasters such as earthquakes, floods or hurricanes. This paper addresses a new challenging problem of camouflaged object segmentation. To address this problem, we provide a new image dataset of camouflaged objects for benchmarking purposes. In addition, we propose a general end-to-end network, called the Anabranch Network, that leverages both classification and segmentation tasks. Different from existing networks for segmentation, our proposed network possesses the second branch for classification to predict the probability of containing camouflaged object(s) in an image, which is then fused into the main branch for segmentation to boost up the segmentation accuracy. Extensive experiments conducted on the newly built dataset demonstrate the effectiveness of our network using various fully convolutional networks. \url{https://sites.google.com/view/ltnghia/research/camo} | ['Akihiro Sugimoto', 'Minh-Triet Tran', 'Zhongliang Nie', 'Tam V. Nguyen', 'Trung-Nghia Le'] | 2021-05-20 | anabranch-network-for-camouflaged-object | https://www.researchgate.net/publication/332806868_Anabranch_network_for_camouflaged_object_segmentation | https://www.researchgate.net/publication/332806868_Anabranch_network_for_camouflaged_object_segmentation | computer-vision-and-image-understanding-2019 | ['camouflaged-object-segmentation'] | ['computer-vision'] | [ 6.47309899e-01 9.96991470e-02 -2.10650757e-01 -8.76538232e-02
-3.29948127e-01 -6.99525356e-01 4.71577704e-01 -7.69068450e-02
-4.21550274e-01 7.36587822e-01 -2.28088319e-01 -4.23596025e-01
2.75217086e-01 -8.18501890e-01 -6.96442068e-01 -7.72064209e-01
1.12715632e-01 1.65747344e-01 6.00423634e-01 -2.19663400e-02
3.01003277e-01 5.78611434e-01 -1.29632318e+00 3.68186459e-02
7.88746893e-01 1.24344277e+00 2.08408132e-01 4.83497918e-01
5.93629759e-03 5.16633689e-01 -7.10940897e-01 -3.54683787e-01
6.43574834e-01 -4.12446439e-01 -7.46665597e-01 1.95365220e-01
5.22824764e-01 -2.83192486e-01 -2.69823611e-01 1.32907414e+00
1.60337478e-01 3.60976607e-02 3.48442763e-01 -1.36351347e+00
-6.64327621e-01 2.51689434e-01 -7.50679612e-01 4.44531083e-01
-2.64573872e-01 3.73595208e-01 6.67657912e-01 -4.77555722e-01
4.75212246e-01 1.08890748e+00 5.31813204e-01 6.53349936e-01
-1.02668762e+00 -8.27304721e-01 7.72887543e-02 8.77922401e-02
-1.06597638e+00 -1.61487088e-01 6.98755026e-01 -4.16025668e-01
1.92742229e-01 4.23489600e-01 6.77345037e-01 7.76142359e-01
1.36411116e-01 9.51108754e-01 1.10049319e+00 -2.43247431e-02
3.02367538e-01 -3.31395924e-01 1.94120854e-01 5.32105565e-01
5.14832556e-01 1.86035886e-01 -6.54218495e-02 -7.26502016e-02
6.28341198e-01 3.51774603e-01 -4.78507817e-01 -2.16563612e-01
-1.08473635e+00 7.84459651e-01 6.76588297e-01 6.71939179e-02
-1.26959756e-01 2.06628650e-01 3.60575169e-01 6.96734414e-02
6.41268253e-01 4.98369426e-01 -5.39671242e-01 1.55075518e-02
-8.83507133e-01 3.29840869e-01 6.64980114e-01 5.68693399e-01
8.19808245e-01 -6.26577586e-02 -2.11269170e-01 6.46556795e-01
3.44266146e-02 4.57976848e-01 2.75375128e-01 -7.53767192e-01
4.55399416e-02 7.61088431e-01 1.42151415e-01 -1.01831365e+00
-2.01131806e-01 -4.68315810e-01 -9.36896682e-01 2.22867072e-01
3.27302217e-01 -2.22604603e-01 -1.31805897e+00 1.53386188e+00
6.28309667e-01 8.19630265e-01 -2.04406574e-01 1.20698547e+00
6.94242120e-01 7.40343034e-01 1.73330531e-02 1.19534031e-01
1.39201152e+00 -1.23188853e+00 -4.38705802e-01 -4.88775283e-01
3.27770829e-01 -5.88491499e-01 8.73886585e-01 3.28487247e-01
-6.52751684e-01 -3.57272327e-01 -9.36316550e-01 3.24164554e-02
-5.05582213e-01 -1.54222205e-01 7.83091009e-01 5.85859060e-01
-7.36925364e-01 3.87245536e-01 -8.28459263e-01 -1.34945467e-01
1.07039034e+00 2.97332317e-01 -1.41768411e-01 -8.87813568e-02
-8.52134109e-01 5.49677193e-01 5.15081763e-01 2.22345233e-01
-9.13361669e-01 -7.01344371e-01 -8.16419482e-01 1.03405960e-01
8.13327432e-01 -3.38027596e-01 1.04484153e+00 -1.30068958e+00
-9.43551362e-01 9.20361280e-01 1.49756119e-01 -6.96408510e-01
4.71790671e-01 -2.22604632e-01 -2.27984771e-01 1.93763867e-01
2.39585847e-01 9.86866593e-01 9.56393719e-01 -1.35876811e+00
-9.32648003e-01 -2.80251563e-01 1.03995927e-01 -1.85231298e-01
-1.35192305e-01 -1.46784469e-01 -4.65220541e-01 -9.99979138e-01
-1.70955211e-01 -9.73227859e-01 -3.53952557e-01 2.17198685e-01
-6.09696448e-01 6.65904954e-02 1.46574688e+00 -6.19115591e-01
8.46431434e-01 -2.16750646e+00 -1.08913310e-01 -5.29492833e-03
4.52950776e-01 7.49669194e-01 -3.23038161e-01 -2.63603474e-03
6.67274967e-02 2.01687902e-01 -6.32591724e-01 -1.52158827e-01
-2.88678080e-01 2.24838123e-01 -5.42392612e-01 6.55744135e-01
4.57471639e-01 1.02656794e+00 -8.57888401e-01 -3.71556073e-01
2.46435747e-01 3.64874452e-01 -5.04412711e-01 2.18563199e-01
-4.59177136e-01 8.04089844e-01 -4.52303410e-01 8.68861616e-01
8.98160636e-01 -2.81641096e-01 -2.27330506e-01 1.15455382e-01
2.10248232e-02 -1.10886067e-01 -8.36448789e-01 1.13530254e+00
3.81463356e-02 6.18368208e-01 2.89280832e-01 -1.03448904e+00
6.39656842e-01 -7.46429488e-02 5.07843971e-01 -7.04891324e-01
5.25911808e-01 2.43929416e-01 4.72051948e-02 -2.83727735e-01
5.54003596e-01 -5.57151139e-02 -1.60642818e-01 2.76185036e-01
-2.31091216e-01 6.30569160e-02 2.81856239e-01 7.58667663e-02
1.13035238e+00 -5.76089881e-02 1.46026075e-01 -2.08687052e-01
2.97719270e-01 2.19402507e-01 7.14234948e-01 6.77870810e-01
-3.95214140e-01 6.22927308e-01 4.70115960e-01 -5.46981215e-01
-7.69181788e-01 -7.46934235e-01 -6.82353824e-02 8.34038317e-01
6.87603951e-01 1.75567195e-01 -8.93326104e-01 -8.73380303e-01
8.31583217e-02 4.28990930e-01 -7.46594191e-01 -2.17093810e-01
-6.17349565e-01 -7.19301224e-01 5.63605070e-01 2.54939109e-01
8.58628988e-01 -1.23049128e+00 -1.05106628e+00 -1.22862168e-01
-1.53709933e-01 -1.10993397e+00 -4.92308915e-01 1.12857390e-02
-5.75646877e-01 -1.22240758e+00 -8.53517175e-01 -8.24920952e-01
6.79715633e-01 6.69646680e-01 9.11010325e-01 4.31762487e-01
-3.41329873e-01 1.45929605e-01 -4.09651250e-01 -4.99582082e-01
-2.77772367e-01 2.72969212e-02 -3.27181250e-01 1.86917961e-01
1.66932553e-01 -3.61787528e-01 -9.37783659e-01 4.30090576e-01
-1.34427869e+00 4.26257700e-02 5.81381023e-01 8.10589910e-01
7.03227997e-01 1.45845786e-01 5.59415579e-01 -9.61018622e-01
3.10831964e-01 -6.14103258e-01 -9.57282364e-01 -1.34741087e-02
-1.77245468e-01 -4.75809216e-01 4.44289833e-01 -6.14500761e-01
-5.26646078e-01 9.32252258e-02 -8.96067247e-02 -5.88058174e-01
-4.56906259e-01 3.14505458e-01 -1.62378117e-01 -2.06888542e-01
1.93246618e-01 3.51450816e-02 -7.21618831e-02 -4.88083601e-01
4.16213930e-01 4.76226687e-01 8.88847470e-01 -1.63258970e-01
1.02599537e+00 8.19923043e-01 6.36770902e-03 -8.15651000e-01
-1.03099895e+00 -5.37653565e-01 -2.75528401e-01 -1.53570905e-01
1.01625812e+00 -6.07661426e-01 -6.97858155e-01 6.46361649e-01
-1.10443151e+00 -4.48839724e-01 -3.51095408e-01 -6.10308275e-02
-2.29300931e-02 5.29808104e-01 -2.73447484e-01 -6.33623540e-01
-3.62397462e-01 -1.09141111e+00 1.13361466e+00 6.00692034e-01
1.45078719e-01 -8.86221349e-01 -8.37353319e-02 6.56240106e-01
3.74823421e-01 6.55112028e-01 6.70290768e-01 -7.70088196e-01
-9.51951563e-01 -2.94531137e-01 -4.76442814e-01 5.04026353e-01
2.43125021e-01 -3.09711933e-01 -7.87818015e-01 -4.31024671e-01
1.29974606e-02 -2.66635448e-01 1.17259824e+00 3.87227565e-01
1.46933341e+00 -5.13405979e-01 -4.42377836e-01 6.06560647e-01
1.19321024e+00 3.47686559e-01 7.20364451e-01 2.70678878e-01
8.39974165e-01 3.07839900e-01 7.11070955e-01 2.37954006e-01
9.21922922e-02 4.32921380e-01 9.48276699e-01 -4.37769115e-01
-2.30557367e-01 4.04944122e-02 -1.12121981e-02 1.88185811e-01
4.05364245e-01 -6.42995358e-01 -7.84186602e-01 7.29481637e-01
-1.56749058e+00 -6.57252491e-01 -1.14064552e-01 1.97281003e+00
5.06212473e-01 1.35791302e-01 5.98995723e-02 -4.43637297e-02
8.11995327e-01 2.70530641e-01 -8.61775398e-01 -7.54228532e-02
-1.69898853e-01 3.77957493e-01 7.53203452e-01 2.18701541e-01
-1.38885999e+00 1.13114011e+00 4.93726444e+00 9.54076350e-01
-1.45791197e+00 2.16101766e-01 9.91406083e-01 2.15899572e-01
1.25202715e-01 9.83347669e-02 -6.65595055e-01 7.57744312e-01
3.48991126e-01 1.62544325e-01 3.43644261e-01 7.14545608e-01
1.71617210e-01 -4.21203643e-01 -6.36039734e-01 6.77250326e-01
7.17386231e-02 -1.46558642e+00 -3.56930005e-03 1.63658217e-01
7.97272205e-01 8.22596550e-02 1.77109703e-01 1.47315606e-01
3.13785940e-01 -8.93302143e-01 7.29737878e-01 5.40160462e-02
5.90313017e-01 -5.73795795e-01 6.59694016e-01 2.94403285e-01
-9.31109905e-01 -1.21963009e-01 -1.30384535e-01 -1.07365847e-01
1.88806906e-01 6.75208747e-01 -5.92607737e-01 1.48111477e-01
7.45188713e-01 5.04634857e-01 -6.47467256e-01 1.35251284e+00
-1.33189738e-01 9.04773891e-01 -2.82183528e-01 2.92393029e-01
4.05888617e-01 -2.22007066e-01 7.23487735e-01 9.26786244e-01
6.66429326e-02 8.82074162e-02 4.21318680e-01 1.02355433e+00
-3.53261441e-01 -2.45681331e-01 -4.25408721e-01 -1.82630882e-01
2.39528865e-01 1.34876215e+00 -1.31792188e+00 -2.76471674e-01
-2.00616539e-01 9.78427231e-01 1.03385381e-01 3.19182754e-01
-9.81750071e-01 -4.41920877e-01 6.39164150e-01 1.58064932e-01
5.84114075e-01 -1.80350602e-01 -4.11866754e-01 -1.12883806e+00
-8.12182948e-02 -7.64259338e-01 2.56015062e-01 -5.73538244e-01
-1.20896184e+00 3.38832825e-01 -1.45183057e-01 -1.10277534e+00
4.11038846e-01 -6.14174962e-01 -7.68593550e-01 5.20492494e-01
-1.66233945e+00 -1.23229086e+00 -4.53507930e-01 2.65323848e-01
6.34320617e-01 6.06040694e-02 2.43965968e-01 3.72252375e-01
-5.58796525e-01 2.50162899e-01 -1.51466936e-01 4.04862851e-01
3.58595669e-01 -1.00581026e+00 5.74771881e-01 1.17169642e+00
-6.33677244e-02 2.19626814e-01 5.43386579e-01 -9.73804176e-01
-1.16022146e+00 -1.57889998e+00 3.00268322e-01 -2.14576662e-01
5.89386761e-01 -4.02249396e-01 -9.09309149e-01 5.04727602e-01
5.29022403e-02 3.13337654e-01 2.88325310e-01 -5.94324410e-01
-2.37321705e-01 1.65274039e-01 -1.26328099e+00 6.91062868e-01
8.65126252e-01 1.31307036e-01 -2.77836502e-01 3.07637751e-01
9.74301398e-01 -4.52906460e-01 -2.99599648e-01 5.19194782e-01
3.59983057e-01 -7.89364994e-01 8.37910831e-01 -5.39879858e-01
5.55429399e-01 -4.45627302e-01 -4.04201038e-02 -1.08674216e+00
-4.84856777e-02 -5.33459485e-01 -1.93634685e-02 1.08530664e+00
1.32834362e-02 -7.73593903e-01 1.03991604e+00 2.81037092e-01
-3.00831199e-01 -8.35806906e-01 -9.98025835e-01 -7.90473998e-01
-1.61353804e-04 -1.49805889e-01 5.70630193e-01 9.54308450e-01
-7.32921243e-01 1.99091464e-01 -4.88689721e-01 2.56476343e-01
5.63402891e-01 3.42273474e-01 8.36120725e-01 -1.11785436e+00
-1.34720847e-01 -4.38630521e-01 -5.33927798e-01 -1.09481668e+00
1.67256549e-01 -8.14653635e-01 1.88197210e-01 -1.36027384e+00
8.88041705e-02 -5.22062480e-01 -1.12086371e-01 7.15948403e-01
-3.30209047e-01 8.92030060e-01 3.47804993e-01 1.18872456e-01
-6.59518719e-01 4.78232056e-01 1.32696617e+00 -3.65450054e-01
-1.89622164e-01 1.04995497e-01 -8.35636854e-01 8.14276516e-01
9.53501403e-01 -6.45869255e-01 -1.90124944e-01 -4.44702297e-01
-1.28702447e-01 -8.90576690e-02 8.57989192e-01 -8.80455136e-01
-1.61560893e-01 -3.53126109e-01 9.83402207e-02 -6.36557162e-01
2.49731556e-01 -9.08339202e-01 -1.02343503e-03 6.85796976e-01
1.93677619e-02 -1.84215829e-01 2.52117395e-01 7.88496554e-01
-8.26670974e-02 4.21633273e-02 1.20190954e+00 9.94857308e-03
-8.68096828e-01 4.27427560e-01 -4.47607607e-01 2.15089843e-01
1.22585547e+00 -3.53051633e-01 -7.03951716e-01 -1.29918188e-01
-2.71116793e-01 4.20319527e-01 5.77424288e-01 4.78540450e-01
5.13672769e-01 -9.30060565e-01 -5.46070993e-01 1.19715601e-01
3.97369936e-02 3.33328694e-01 1.68014556e-01 7.82854378e-01
-7.04146326e-01 8.61380994e-02 -2.91958362e-01 -5.35685360e-01
-1.22841859e+00 6.33365095e-01 2.86767006e-01 -1.29632741e-01
-5.94510913e-01 7.51859128e-01 4.60720450e-01 -2.25839913e-01
8.83261263e-02 -5.11378825e-01 -1.55799445e-02 -2.79913574e-01
3.79599243e-01 4.19892520e-01 -6.18787184e-02 -7.92672455e-01
-2.23791406e-01 1.82014182e-01 -1.64819568e-01 3.32552254e-01
1.27311528e+00 3.20986696e-02 -2.97791660e-01 -1.51180811e-02
1.07979405e+00 -1.80210412e-01 -1.29358339e+00 -1.27410412e-01
-6.35835975e-02 -7.12885976e-01 -5.06961048e-02 -8.44312549e-01
-1.60586095e+00 6.62044823e-01 7.01660037e-01 2.80818343e-01
1.20474553e+00 4.41583581e-02 1.30808592e+00 2.88960248e-01
2.68788546e-01 -6.67853296e-01 1.75310522e-01 2.97749043e-01
5.19145906e-01 -1.18116319e+00 -1.24969728e-01 -5.74644923e-01
-4.55914050e-01 5.01597047e-01 6.12711310e-01 -4.01915580e-01
6.30790174e-01 3.32151353e-01 1.32475093e-01 -2.12142751e-01
-3.51420850e-01 -2.68938333e-01 1.15396440e-01 5.76361477e-01
-2.32349187e-01 1.50537491e-01 -2.70450950e-01 5.09188175e-01
1.41775468e-02 -4.01097722e-02 4.84516412e-01 1.17090905e+00
-5.80648005e-01 -7.81332970e-01 -4.79743659e-01 6.71167672e-01
-6.81830823e-01 1.18663505e-01 -6.24874234e-01 5.80678940e-01
5.85065842e-01 9.38021183e-01 1.61044493e-01 -1.13197222e-01
6.95756152e-02 -3.68768483e-01 1.27401156e-02 -5.84893346e-01
-6.07086062e-01 6.52741641e-02 -2.53577769e-01 -4.08636391e-01
-4.35914218e-01 -4.87114251e-01 -1.11706018e+00 -1.37932599e-01
-3.43444675e-01 3.03713698e-02 5.34475803e-01 8.66949797e-01
3.13531548e-01 4.16131169e-01 3.98400009e-01 -1.06850469e+00
-1.88012689e-01 -6.12237692e-01 -5.06993830e-01 4.84958857e-01
5.15485168e-01 -7.31091678e-01 -2.29210421e-01 2.32244030e-01] | [9.625077247619629, -0.12727481126785278] |
d32ac7f1-8932-4359-8c20-3cdcd5db0c23 | predicting-retrosynthetic-reaction-using-self | 1907.01356 | null | https://arxiv.org/abs/1907.01356v2 | https://arxiv.org/pdf/1907.01356v2.pdf | Predicting Retrosynthetic Reaction using Self-Corrected Transformer Neural Networks | Synthesis planning is the process of recursively decomposing target molecules into available precursors. Computer-aided retrosynthesis can potentially assist chemists in designing synthetic routes, but at present it is cumbersome and provides results of dissatisfactory quality. In this study, we develop a template-free self-corrected retrosynthesis predictor (SCROP) to perform a retrosynthesis prediction task trained by using the Transformer neural network architecture. In the method, the retrosynthesis planning is converted as a machine translation problem between molecular linear notations of reactants and the products. Coupled with a neural network-based syntax corrector, our method achieves an accuracy of 59.0% on a standard benchmark dataset, which increases >21% over other deep learning methods, and >6% over template-based methods. More importantly, our method shows an accuracy 1.7 times higher than other state-of-the-art methods for compounds not appearing in the training set. | ['Yuedong Yang', 'Jiahua Rao', 'Jun Xu', 'Zhongyue Zhang', 'Shuangjia Zheng'] | 2019-07-02 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 7.82962918e-01 2.00672850e-01 -3.80972773e-01 -5.76580837e-02
-7.37662852e-01 -9.95078504e-01 8.14231157e-01 3.22057694e-01
-3.07239443e-01 1.09087837e+00 7.57941231e-02 -7.92111039e-01
5.26845336e-01 -7.88606882e-01 -9.30985272e-01 -7.08045244e-01
2.71498501e-01 4.24959779e-01 -7.10119978e-02 -3.49860907e-01
4.61646885e-01 7.78805852e-01 -1.07161891e+00 4.47482944e-01
9.23788428e-01 1.16556239e+00 1.37160838e-01 4.27818686e-01
-1.90337718e-01 8.79519999e-01 -4.72764224e-01 -7.48950541e-01
2.18979999e-01 -4.19656664e-01 -7.99735785e-01 -6.36952221e-01
5.08613050e-01 -1.87797710e-01 -1.02554440e-01 7.32254744e-01
3.33992332e-01 7.50018135e-02 8.12063277e-01 -5.35646677e-01
-5.29521406e-01 1.02194822e+00 5.53914607e-02 -1.23444878e-01
2.48952642e-01 1.18655495e-01 1.14471650e+00 -1.38483739e+00
5.70630074e-01 8.62529874e-01 5.90923369e-01 5.74387550e-01
-1.38636768e+00 -1.03504932e+00 -1.28144501e-02 -4.01277058e-02
-1.27875769e+00 -7.23189652e-01 5.30109704e-01 -4.94087785e-01
1.66124463e+00 9.62544903e-02 6.46786571e-01 1.16149461e+00
8.15611362e-01 2.60721177e-01 6.56159520e-01 -1.24820538e-01
1.86725214e-01 -3.09782624e-01 -4.98471588e-01 8.38006735e-01
-7.59332860e-03 3.21233541e-01 -4.48915303e-01 2.36501500e-01
5.82344651e-01 3.75906974e-02 -1.27553511e-02 6.19507656e-02
-1.37905967e+00 8.52249324e-01 7.94999123e-01 2.85114735e-01
-3.91933799e-01 2.50215501e-01 4.30417925e-01 -1.98352244e-02
3.74559999e-01 1.33204019e+00 -7.61378467e-01 1.27448186e-01
-1.27854431e+00 1.14324436e-01 7.61216164e-01 9.24164534e-01
4.53435957e-01 5.33555746e-01 7.97485486e-02 3.55750591e-01
-9.14559886e-02 -3.29484455e-02 3.27230424e-01 -5.67642510e-01
4.08680320e-01 5.84385514e-01 3.70120890e-02 -4.50622201e-01
-6.55651867e-01 -4.62677836e-01 -8.37472856e-01 -1.82719499e-01
4.37559575e-01 -2.28934735e-02 -1.06692505e+00 1.59047306e+00
1.39798164e-01 -2.94240028e-01 1.65549055e-01 5.28259575e-01
1.11233282e+00 1.33836210e+00 3.85242373e-01 -5.43108165e-01
8.90145898e-01 -1.37015986e+00 -5.52565992e-01 1.09782360e-01
6.83392644e-01 -9.61761773e-01 5.40701091e-01 1.02842534e+00
-1.29371119e+00 -5.19848764e-01 -1.45891607e+00 -3.95112544e-01
-9.13421333e-01 3.96265447e-01 7.84796119e-01 7.35468924e-01
-7.94378996e-01 1.38940394e+00 -4.76365417e-01 7.23525211e-02
3.42708677e-01 7.47715056e-01 -4.77458984e-01 3.28912526e-01
-9.33522046e-01 1.09956324e+00 9.65249896e-01 5.53284399e-02
-1.33606565e+00 -1.28052735e+00 -6.40936136e-01 2.01693401e-01
5.74539959e-01 -6.36876404e-01 1.51363206e+00 -8.22205544e-01
-2.04148841e+00 3.69644970e-01 -5.50949797e-02 -4.27113265e-01
3.17134947e-01 7.78006241e-02 -4.95318085e-01 -1.64172903e-01
-2.99009718e-02 9.49527860e-01 7.30333745e-01 -8.60603750e-01
-4.41599488e-01 5.57000153e-02 -6.34051710e-02 -1.44488677e-01
2.70627379e-01 -3.93542647e-02 1.70532037e-02 -7.88935602e-01
-8.74990374e-02 -7.54079640e-01 -4.39642668e-01 -8.14490989e-02
-5.85493684e-01 -2.32931152e-01 1.62785634e-01 -7.04615474e-01
1.03499544e+00 -1.48754179e+00 3.36788386e-01 3.68713178e-02
2.84284681e-01 5.26750267e-01 -1.71323195e-01 7.63402224e-01
-6.89531505e-01 3.86769503e-01 -2.09122121e-01 1.01740484e-03
-1.11100256e-01 -3.45058978e-01 -6.30015254e-01 2.05028296e-01
4.88513917e-01 8.91344965e-01 -1.07697606e+00 3.71819213e-02
1.33009225e-01 3.27466160e-01 -6.70669138e-01 7.65292645e-02
-8.11390936e-01 4.93750989e-01 -2.54284106e-02 9.65013027e-01
4.55121160e-01 -2.82622337e-01 7.54103661e-01 -4.68593508e-01
-5.79361856e-01 1.02356362e+00 -4.27815139e-01 2.06279588e+00
-4.39568281e-01 2.16615558e-01 -6.44766212e-01 -8.25869083e-01
1.08527100e+00 5.39223373e-01 3.34233344e-01 -9.06364202e-01
2.28784502e-01 5.31693876e-01 1.46181226e-01 1.41463816e-01
8.75010133e-01 -5.21503389e-01 1.13826469e-01 2.38923788e-01
3.39939952e-01 -1.80483803e-01 5.19148707e-01 2.77756192e-02
7.57358909e-01 5.97117305e-01 7.61227489e-01 -1.69388413e-01
7.79639304e-01 5.22093438e-02 4.97957259e-01 3.88920426e-01
2.69856900e-01 1.82233304e-01 3.50956559e-01 -9.45088625e-01
-1.43824303e+00 -9.27175760e-01 1.79517016e-01 1.08598197e+00
-4.28526461e-01 -1.00080836e+00 -6.10264421e-01 -7.09517300e-01
-4.35998619e-01 7.39798188e-01 -3.96321177e-01 -2.88283348e-01
-6.61567509e-01 -5.69647312e-01 6.80349648e-01 6.29237056e-01
2.21631527e-01 -8.10847878e-01 1.22867502e-01 7.61392295e-01
1.28634483e-01 -8.49404037e-01 -5.66668570e-01 6.58762991e-01
-8.62085581e-01 -9.65093493e-01 -5.06242871e-01 -6.80052638e-01
4.50465679e-01 -1.99831445e-02 9.46662903e-01 -1.14922173e-01
-1.54332593e-01 -7.72374988e-01 -4.89439368e-02 -3.96394640e-01
-8.71749878e-01 4.44253743e-01 1.72907393e-02 -5.04130244e-01
-1.52861327e-01 -6.13207459e-01 -6.89716756e-01 4.70756292e-02
-7.44342029e-01 3.46411705e-01 5.92961431e-01 8.07892025e-01
6.63631856e-01 -2.01829970e-01 7.19625056e-01 -6.76084638e-01
1.31323054e-01 -2.86537319e-01 -9.89558160e-01 3.79654706e-01
-1.03135753e+00 4.98135626e-01 1.46592557e+00 -3.52076560e-01
-9.80759144e-01 5.40787160e-01 -3.23395967e-01 5.40964827e-02
2.55315393e-01 5.46242177e-01 -3.35551679e-01 -9.49936137e-02
7.41659999e-01 2.13169053e-01 -3.49919379e-01 -3.76880497e-01
6.46610558e-01 4.83996384e-02 3.59481126e-01 -6.28071427e-01
4.74004716e-01 -9.00467709e-02 6.18385136e-01 -4.31440264e-01
-7.97630250e-01 6.95742443e-02 -6.81548655e-01 2.28383899e-01
6.24758899e-01 -8.67722213e-01 -9.68756735e-01 -9.15414914e-02
-1.41306508e+00 -3.50646615e-01 2.19024494e-01 1.59032747e-01
-5.41450679e-01 3.03090870e-01 -5.90486944e-01 -2.13259533e-01
-9.00260389e-01 -1.54650009e+00 8.27049732e-01 -6.00004643e-02
-2.90078819e-01 -7.13270485e-01 1.83432754e-02 2.40731969e-01
9.68841985e-02 3.26485217e-01 1.23800194e+00 -8.51838768e-01
-7.92705417e-01 8.52330178e-02 -2.32523888e-01 7.67367706e-02
3.63701433e-01 2.83351034e-01 -8.94011140e-01 -5.70880845e-02
-6.00464165e-01 -3.30860585e-01 1.09791481e+00 7.91479945e-02
1.31133819e+00 -6.18898988e-01 -4.33420509e-01 7.28051424e-01
1.27591944e+00 8.12486649e-01 6.01635277e-01 7.53411427e-02
7.88364530e-01 4.60488558e-01 5.25730908e-01 2.60914534e-01
-3.24442349e-02 4.24288452e-01 5.70145011e-01 1.75622880e-01
4.18672971e-02 -6.87652111e-01 4.14900959e-01 3.72018605e-01
-3.34199965e-01 -4.44051474e-01 -8.07434559e-01 -1.14268102e-01
-1.37343979e+00 -9.38502967e-01 -1.75403878e-02 2.07740283e+00
1.22048950e+00 3.12800594e-02 2.24502206e-01 2.05131531e-01
2.70238370e-01 6.13219775e-02 -5.72374642e-01 -7.77744174e-01
2.20754966e-01 9.35276568e-01 7.55265951e-01 3.18355620e-01
-1.15903473e+00 1.53802443e+00 6.98559523e+00 9.53270018e-01
-1.63856828e+00 -2.48741955e-01 4.92208153e-01 -9.57094878e-02
-1.18991427e-01 1.94725782e-01 -8.39490235e-01 3.11677635e-01
1.36640453e+00 -1.06017150e-01 6.98011220e-01 7.46117711e-01
2.28269681e-01 3.71445745e-01 -1.72623754e+00 7.50888526e-01
-1.34065092e-01 -2.33188820e+00 4.44797963e-01 -1.11884810e-01
6.62976325e-01 -2.86945224e-01 1.09359331e-01 2.61080295e-01
1.29260466e-01 -1.50344694e+00 1.06656539e+00 1.60292745e-01
1.10059547e+00 -1.07727885e+00 1.02937460e-01 1.60421189e-02
-1.29361856e+00 -5.08875400e-02 -1.73020378e-01 6.13792203e-02
1.05266450e-02 1.95506886e-01 -1.48947191e+00 7.87052989e-01
2.02106368e-02 9.59822297e-01 -2.74501234e-01 6.54708922e-01
-3.88126493e-01 3.51906061e-01 -3.02552599e-02 -2.58472532e-01
4.67982709e-01 -3.31574142e-01 -2.25243159e-02 1.30150449e+00
4.46871698e-01 1.18928947e-01 -4.68655750e-02 1.23450434e+00
-6.34643316e-01 1.66468456e-01 -4.98250455e-01 -7.97298729e-01
4.20011789e-01 1.23879993e+00 -8.52616489e-01 -5.39608657e-01
-3.11245210e-02 7.63067544e-01 1.55089080e-01 1.54054791e-01
-1.03083432e+00 -4.00557309e-01 3.33200485e-01 -1.69822156e-01
5.16148150e-01 -2.56292731e-01 -2.05711693e-01 -9.39882755e-01
-3.43442082e-01 -1.10812402e+00 2.92077921e-02 -6.98166549e-01
-7.42563367e-01 5.66252470e-01 -4.82546061e-01 -1.01383448e+00
-1.38255507e-02 -1.09487605e+00 -6.12346292e-01 6.24452710e-01
-1.31143713e+00 -1.11433661e+00 3.29089999e-01 -3.78004368e-03
7.57692695e-01 -3.64523262e-01 1.11427915e+00 2.92018563e-01
-8.00090373e-01 5.72474122e-01 1.33717105e-01 -3.94240499e-01
8.30457807e-01 -1.21834779e+00 7.18312323e-01 7.59241819e-01
-1.02609918e-01 1.04536724e+00 6.04719460e-01 -6.70294464e-01
-1.52078843e+00 -1.43758798e+00 1.11423302e+00 -3.13339055e-01
8.23036492e-01 -4.46663231e-01 -4.26147699e-01 5.03274679e-01
1.71339482e-01 -4.54332530e-01 9.72996294e-01 -2.68845230e-01
-7.99697399e-01 -8.28854889e-02 -8.66290152e-01 8.36688101e-01
1.00334251e+00 -3.38721842e-01 -3.32961082e-01 5.67118406e-01
1.08519578e+00 -4.82195079e-01 -1.17523432e+00 2.16687411e-01
8.19143176e-01 -6.04570329e-01 1.16046298e+00 -7.81534672e-01
1.03354812e+00 -1.72335848e-01 -1.63407415e-01 -1.15364707e+00
-2.87379891e-01 -9.86663043e-01 3.63070220e-02 7.46361852e-01
9.76320744e-01 -3.43512326e-01 7.21672058e-01 3.74865443e-01
-6.41748607e-01 -6.48421824e-01 -6.86322033e-01 -8.02300394e-01
4.62344080e-01 -8.98344144e-02 8.42036843e-01 7.82362461e-01
2.39207819e-01 8.69701684e-01 -3.37804854e-01 -1.99273244e-01
5.42151816e-02 4.71007638e-02 5.10322154e-01 -1.03260374e+00
-1.59021802e-02 -7.87160635e-01 2.00858504e-01 -1.00327659e+00
3.90592635e-01 -1.36864555e+00 -8.24003592e-02 -1.22850811e+00
-5.78854308e-02 -1.94362372e-01 -1.90748259e-01 6.29567683e-01
2.69473672e-01 -3.54304276e-02 4.43088673e-02 -1.54643193e-01
-3.78595859e-01 7.32308805e-01 1.24077427e+00 -6.09144986e-01
-2.01529056e-01 -3.69303167e-01 -7.15651155e-01 4.10522491e-01
9.19596434e-01 -5.47834873e-01 -2.17026487e-01 4.92272265e-02
6.28524303e-01 1.33412793e-01 -4.33071069e-02 -6.40762985e-01
1.43124968e-01 -3.41542602e-01 5.47123253e-01 -1.04666328e+00
5.34830809e-01 -5.76599240e-01 4.40806329e-01 7.33416736e-01
-4.84584033e-01 1.05391098e-02 3.22729647e-01 2.63768852e-01
2.46652037e-01 -8.65908861e-02 5.48318326e-01 -3.93800527e-01
-2.84341037e-01 5.22672117e-01 -7.52199352e-01 -6.92244828e-01
5.55881143e-01 -2.28805393e-02 -4.65731859e-01 2.54638344e-01
-4.86589283e-01 -2.70853788e-01 3.43551189e-01 8.79936516e-02
6.69456184e-01 -1.10284781e+00 -4.41505648e-02 -4.85777371e-02
7.56818950e-02 3.05588469e-02 -2.37553447e-01 7.89149046e-01
-8.96454036e-01 1.02725816e+00 -2.02536941e-01 -1.55769372e-02
-9.61756647e-01 9.82844770e-01 4.32737559e-01 -3.28892082e-01
-4.60366607e-02 7.65712202e-01 1.28732428e-01 -5.71056247e-01
1.95787042e-01 -8.63752961e-01 5.30296937e-02 7.12763658e-03
5.03956616e-01 2.55138010e-01 4.34897214e-01 -2.75460750e-01
-3.30063939e-01 2.37159789e-01 -4.69186038e-01 2.80364245e-01
1.30397666e+00 7.75166750e-01 -2.62219667e-01 -1.54281169e-01
1.09841013e+00 -2.89419621e-01 -1.00246692e+00 -1.99995320e-02
-1.16392635e-02 1.12483747e-01 9.88800153e-02 -1.19014585e+00
-6.78666234e-01 9.17261839e-01 1.27566755e-01 -2.40383625e-01
8.25955987e-01 -3.98552418e-01 8.02587509e-01 1.06398118e+00
4.83798496e-02 -9.17041659e-01 1.86955899e-01 6.23874247e-01
9.66209829e-01 -1.14576411e+00 2.96256483e-01 -4.01815861e-01
-1.98879018e-01 1.66036022e+00 2.69714981e-01 4.10378054e-02
1.03715345e-01 -1.34080155e-02 -5.88272691e-01 1.89418010e-02
-9.61497903e-01 2.37671703e-01 3.45492721e-01 3.32851857e-01
9.19724405e-01 1.51020251e-02 -4.10655528e-01 7.21980751e-01
-4.86148000e-01 2.14976519e-02 4.29209709e-01 8.25026333e-01
-4.07603413e-01 -1.57847536e+00 -4.64230999e-02 -9.15047079e-02
-5.64134061e-01 -7.77427375e-01 -9.21774447e-01 6.56673074e-01
2.07908794e-01 8.93948674e-01 -2.93748677e-01 -4.22992498e-01
6.93779960e-02 2.16132119e-01 9.43147838e-01 -6.29461884e-01
-9.74474967e-01 8.66026953e-02 5.07521391e-01 -6.67996347e-01
-3.17838103e-01 -1.50551319e-01 -1.35666537e+00 -5.20068228e-01
-1.74800932e-01 1.14998586e-01 8.32626820e-01 9.29722846e-01
2.78976381e-01 5.93716204e-01 6.08206570e-01 -1.10480320e+00
-3.66141945e-01 -5.20714581e-01 1.03733443e-01 -5.11599720e-01
2.42357761e-01 -4.66856211e-01 1.74072906e-01 1.39645904e-01] | [4.501057147979736, 6.100210666656494] |
fc378ae8-148d-4cdd-95db-846e69545a3c | chunk-aware-alignment-and-lexical-constraint | 2207.11401 | null | https://arxiv.org/abs/2207.11401v2 | https://arxiv.org/pdf/2207.11401v2.pdf | Chunk-aware Alignment and Lexical Constraint for Visual Entailment with Natural Language Explanations | Visual Entailment with natural language explanations aims to infer the relationship between a text-image pair and generate a sentence to explain the decision-making process. Previous methods rely mainly on a pre-trained vision-language model to perform the relation inference and a language model to generate the corresponding explanation. However, the pre-trained vision-language models mainly build token-level alignment between text and image yet ignore the high-level semantic alignment between the phrases (chunks) and visual contents, which is critical for vision-language reasoning. Moreover, the explanation generator based only on the encoded joint representation does not explicitly consider the critical decision-making points of relation inference. Thus the generated explanations are less faithful to visual-language reasoning. To mitigate these problems, we propose a unified Chunk-aware Alignment and Lexical Constraint based method, dubbed as CALeC. It contains a Chunk-aware Semantic Interactor (arr. CSI), a relation inferrer, and a Lexical Constraint-aware Generator (arr. LeCG). Specifically, CSI exploits the sentence structure inherent in language and various image regions to build chunk-aware semantic alignment. Relation inferrer uses an attention-based reasoning network to incorporate the token-level and chunk-level vision-language representations. LeCG utilizes lexical constraints to expressly incorporate the words or chunks focused by the relation inferrer into explanation generation, improving the faithfulness and informativeness of the explanations. We conduct extensive experiments on three datasets, and experimental results indicate that CALeC significantly outperforms other competitor models on inference accuracy and quality of generated explanations. | ['Min Zhang', 'Yuxing Ding', 'Lin Ma', 'Baotian Hu', 'Yunxin Li', 'Qian Yang'] | 2022-07-23 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [ 1.40073538e-01 5.55850208e-01 -1.88166529e-01 -5.42129993e-01
-2.97564209e-01 -4.05127853e-02 7.72545159e-01 -8.50579739e-02
4.18675914e-02 4.81518716e-01 6.30016744e-01 -3.26469332e-01
3.27560008e-01 -8.06475341e-01 -9.28487659e-01 -3.33387375e-01
7.24182665e-01 3.01600099e-01 2.71072686e-01 -2.57696770e-02
3.96998137e-01 -1.10715456e-01 -1.52223516e+00 8.27757716e-01
1.17100477e+00 8.64064276e-01 7.08046734e-01 4.76488948e-01
-7.78311551e-01 1.27277017e+00 -2.33688310e-01 -6.29532993e-01
-5.62117733e-02 -9.01605725e-01 -8.07280719e-01 4.32920903e-01
3.86801839e-01 -4.73131686e-01 -1.22464530e-01 1.14593136e+00
3.72032858e-02 -1.23208709e-01 7.06006110e-01 -1.41544878e+00
-1.32064426e+00 7.84624159e-01 -5.76236069e-01 1.82952266e-02
4.12437260e-01 4.36966717e-01 9.61117446e-01 -1.16632819e+00
4.73288089e-01 1.39830995e+00 3.35381866e-01 7.11179912e-01
-8.49672735e-01 -4.14018750e-01 5.25014102e-01 6.40185416e-01
-1.02149391e+00 -2.62488693e-01 7.31702209e-01 -3.04612607e-01
1.13882113e+00 7.88032413e-02 9.21724737e-01 9.14926648e-01
4.02414382e-01 8.34438682e-01 8.68348718e-01 -5.38580716e-01
1.14334323e-01 1.51422709e-01 9.28996131e-02 1.07451677e+00
2.03934103e-01 1.16669193e-01 -6.66239679e-01 3.23660105e-01
1.02485120e+00 8.48397091e-02 -2.85040349e-01 -7.39167109e-02
-1.28790593e+00 9.14171994e-01 7.63559520e-01 5.77265210e-02
-6.29991174e-01 2.07088947e-01 2.08723918e-01 -2.14500070e-01
1.59975439e-01 2.95922011e-01 -8.37093890e-02 4.74250019e-01
-7.11830437e-01 1.02042861e-01 5.01874208e-01 1.12635708e+00
9.53788400e-01 2.34173685e-01 -6.38491929e-01 4.98469770e-01
7.57738531e-01 4.02722746e-01 5.16002774e-01 -1.09433091e+00
4.18483436e-01 8.89822960e-01 -2.62609631e-01 -1.16600835e+00
3.45046557e-02 -2.01558277e-01 -7.40168750e-01 1.66828617e-01
2.11877003e-02 1.89583242e-01 -9.86278772e-01 1.64302528e+00
2.77727574e-01 2.57017791e-01 3.76856059e-01 1.32201123e+00
1.38163531e+00 6.68132126e-01 2.49233052e-01 -2.62016505e-01
1.71958733e+00 -1.58631921e+00 -8.06692481e-01 -6.43487334e-01
1.02199264e-01 -9.00478125e-01 1.30325544e+00 -1.70604855e-01
-1.18356860e+00 -7.96663463e-01 -9.40745175e-01 -4.57764685e-01
-3.45862024e-02 1.95596918e-01 6.06880724e-01 -1.53563082e-01
-8.59788537e-01 -1.25030622e-01 -3.52385253e-01 -2.46491298e-01
4.98471022e-01 -6.76737055e-02 -1.89691886e-01 -2.59756327e-01
-1.22315443e+00 9.71019804e-01 4.49215084e-01 1.44065827e-01
-8.37481141e-01 -5.01057565e-01 -1.09853530e+00 1.59246579e-01
3.93982947e-01 -1.49222338e+00 1.16217744e+00 -1.37528300e+00
-1.33995795e+00 8.23995471e-01 -5.75455904e-01 -4.90161330e-01
1.67926252e-01 -3.15506086e-02 -5.79605997e-03 4.43924129e-01
4.75060463e-01 1.14928412e+00 9.61131692e-01 -1.43563557e+00
-5.76681912e-01 -1.48272097e-01 4.36007500e-01 5.92608213e-01
1.72081023e-01 -2.68465459e-01 -7.77311265e-01 -5.80172360e-01
2.72243440e-01 -5.90088129e-01 -2.29600161e-01 1.41794682e-01
-5.76235056e-01 -3.02514374e-01 5.24834514e-01 -5.98914623e-01
8.67107332e-01 -1.88353992e+00 1.02053270e-01 -2.84843624e-01
3.88572872e-01 6.23155348e-02 -3.73337746e-01 2.35075310e-01
-1.01264112e-01 -2.79693715e-02 -5.93602248e-02 -3.62660557e-01
-1.08071543e-01 6.59173191e-01 -7.01410413e-01 -7.85879567e-02
3.13189805e-01 1.21254468e+00 -9.61486518e-01 -9.28869724e-01
4.20844018e-01 3.79150480e-01 -7.22814023e-01 4.86726254e-01
-7.73177028e-01 2.73037910e-01 -6.76558554e-01 4.24158931e-01
5.36454976e-01 -3.67292374e-01 -1.04778893e-01 -7.03018665e-01
-1.01941109e-01 2.37606481e-01 -5.38807452e-01 1.84315586e+00
-5.19383490e-01 5.37959158e-01 -4.51229244e-01 -8.70124698e-01
8.14561188e-01 2.04370335e-01 -1.11800291e-01 -6.68637693e-01
3.09179351e-02 -8.26394781e-02 -5.89327514e-02 -9.62825894e-01
2.99472183e-01 -3.65292817e-01 2.41685256e-01 3.97072881e-01
-9.17265713e-02 -2.37277910e-01 -9.53870490e-02 7.40297914e-01
6.00453079e-01 3.00941944e-01 4.85303462e-01 8.23402852e-02
6.92022026e-01 3.33077908e-01 5.58545411e-01 6.02750003e-01
-3.16669755e-02 8.04557025e-01 5.55066288e-01 -4.53179568e-01
-8.75370860e-01 -1.06290948e+00 3.64063621e-01 6.82140410e-01
7.43471861e-01 -5.14630675e-01 -8.64031374e-01 -7.65307486e-01
-2.25048065e-01 1.20655239e+00 -6.94445252e-01 -2.76267380e-01
-1.18140280e-01 -1.40489027e-01 1.21824786e-01 7.15530217e-01
8.86466980e-01 -1.53055155e+00 -6.35836899e-01 -1.12909555e-01
-6.22923493e-01 -1.43761241e+00 -6.88616216e-01 -3.26410860e-01
-5.75572252e-01 -1.16975129e+00 -2.10043505e-01 -7.33371198e-01
1.10555446e+00 5.83746433e-01 1.24348879e+00 5.51524580e-01
-1.06751136e-01 4.51712996e-01 -4.83933657e-01 -4.88301903e-01
-2.49184072e-01 -5.59384584e-01 -2.63079643e-01 -6.97532147e-02
5.09318709e-01 -2.94781983e-01 -7.36354649e-01 1.51047692e-01
-7.71598339e-01 9.25582290e-01 9.21616375e-01 9.73006427e-01
8.37373734e-01 -2.45959610e-01 3.32913518e-01 -6.55704439e-01
4.12159592e-01 -3.58517855e-01 -1.65166900e-01 4.22800004e-01
-5.38327634e-01 3.35381716e-01 6.24031723e-01 -3.67835432e-01
-1.43163681e+00 6.83060139e-02 -2.89684031e-02 -6.11267447e-01
-1.90598935e-01 5.14035463e-01 -3.49595815e-01 1.94875628e-01
4.34362918e-01 5.87466359e-01 1.09545782e-01 2.50743628e-01
7.92631924e-01 3.41858000e-01 7.72305846e-01 -5.51897824e-01
6.72366738e-01 4.49767232e-01 -2.26133391e-01 -5.07199764e-01
-1.44891334e+00 -7.26360530e-02 -4.36785072e-01 -2.55064189e-01
1.42459083e+00 -9.09416378e-01 -4.90512699e-01 -5.58826476e-02
-1.88451457e+00 -8.55221376e-02 -4.36254054e-01 5.24791300e-01
-7.75411785e-01 5.00437081e-01 -3.99661869e-01 -5.05158603e-01
-3.93510967e-01 -1.22579670e+00 9.41017270e-01 5.41725755e-01
-1.76538631e-01 -8.82302403e-01 -3.22774947e-01 7.78406382e-01
1.50289506e-01 5.74210994e-02 1.02413905e+00 -3.42255980e-01
-9.43223953e-01 2.20759928e-01 -8.65089893e-01 1.10500142e-01
-5.49623258e-02 -2.92184744e-02 -8.35443318e-01 4.78452653e-01
2.14745104e-01 -3.79925221e-01 9.32049930e-01 5.71822405e-01
1.20589399e+00 -6.01864815e-01 -1.30270720e-01 3.97876829e-01
1.38178110e+00 -4.13572863e-02 8.92917395e-01 1.22697875e-01
9.39795971e-01 7.14799702e-01 7.19074368e-01 1.54305741e-01
9.70313728e-01 5.47590077e-01 7.55643725e-01 -2.70802557e-01
-6.92013860e-01 -6.02755785e-01 4.33447689e-01 7.63417184e-01
-2.03540564e-01 -6.43114932e-03 -5.64324498e-01 3.71477127e-01
-2.35969782e+00 -1.23619580e+00 -4.67705488e-01 1.68025589e+00
9.06444192e-01 -6.89633414e-02 -4.48911846e-01 -5.22584438e-01
5.95134377e-01 5.39598167e-02 -7.43416905e-01 -5.18194258e-01
-6.38808608e-02 -2.83544958e-01 -1.29333973e-01 6.58537447e-01
-5.22508919e-01 1.36070466e+00 5.39948130e+00 7.56859362e-01
-8.95654619e-01 1.32206395e-01 3.77455562e-01 1.69809133e-01
-7.18435407e-01 3.53354961e-01 -6.52537405e-01 3.10042828e-01
3.37666541e-01 -2.15681627e-01 3.42769951e-01 7.65495002e-01
4.29780990e-01 -3.43167275e-01 -1.06318319e+00 1.00980365e+00
5.12334406e-01 -1.70508957e+00 6.97583914e-01 -1.31672382e-01
6.54933214e-01 -3.66362900e-01 -5.74097820e-02 2.34916419e-01
1.64017320e-01 -1.12532866e+00 1.01703274e+00 8.28663290e-01
5.69786549e-01 -4.21828896e-01 8.57043862e-01 3.60946596e-01
-1.33815277e+00 1.29757464e-01 -7.23758936e-01 -2.40417957e-01
2.30354071e-01 5.94738722e-01 -7.60631442e-01 6.66565359e-01
4.30887043e-01 7.90794075e-01 -4.42175239e-01 4.90685463e-01
-1.02964187e+00 2.96840221e-01 2.17626452e-01 9.54313055e-02
3.97619069e-01 -5.03605418e-02 3.86848569e-01 9.94593859e-01
8.56296346e-02 5.79601407e-01 2.10836023e-01 1.48690379e+00
1.59838647e-01 -1.39912590e-02 -5.41497707e-01 1.27763242e-01
4.07809198e-01 1.23303723e+00 -6.06863022e-01 -5.85148394e-01
-6.61972761e-01 1.27568007e+00 4.10172373e-01 4.62244719e-01
-1.11580467e+00 4.84317876e-02 4.49667424e-01 -8.17204884e-04
2.76454985e-01 -2.72563193e-02 -6.19268954e-01 -1.26541138e+00
-6.69031367e-02 -7.58137465e-01 1.43364564e-01 -1.58161819e+00
-1.29794896e+00 6.55063748e-01 -9.70981363e-03 -1.14764380e+00
-2.32677370e-01 -3.87142330e-01 -1.06323802e+00 1.00081253e+00
-1.70768917e+00 -1.54333806e+00 -6.52378976e-01 7.27472425e-01
1.25888193e+00 -1.33523405e-01 6.68527186e-01 -4.29757863e-01
-5.34813643e-01 1.69720367e-01 -1.04227698e+00 8.69657006e-03
4.58551288e-01 -1.15609050e+00 -2.54051834e-02 1.00195432e+00
3.59324425e-01 7.72663414e-01 7.94666529e-01 -7.15664744e-01
-1.08823860e+00 -1.14021623e+00 1.06394935e+00 -3.67282778e-01
3.86826098e-01 9.41534266e-02 -9.72495914e-01 6.47799790e-01
7.29577959e-01 -1.07507080e-01 6.34175003e-01 -2.76000142e-01
-4.94066685e-01 1.35078013e-01 -8.16168189e-01 8.87928069e-01
1.08712101e+00 -6.05717719e-01 -1.27170849e+00 4.01790977e-01
1.11702144e+00 -4.11739737e-01 -1.46230608e-01 1.03490010e-01
4.96990532e-01 -1.33601809e+00 9.97564375e-01 -4.14951861e-01
1.29953420e+00 -6.82319283e-01 -1.10560104e-01 -9.40468729e-01
-3.22728127e-01 1.38985544e-01 -1.04679421e-01 1.24401212e+00
3.46598506e-01 -3.53077680e-01 4.69985664e-01 7.78032839e-01
-3.36350739e-01 -9.05786633e-01 -4.48714346e-01 -3.35895926e-01
-3.53461295e-01 -4.52477813e-01 5.28211296e-01 7.98499107e-01
2.71447226e-02 8.13413024e-01 -2.40761578e-01 4.04857129e-01
7.31746376e-01 5.62921524e-01 7.70768106e-01 -7.06975520e-01
-3.34896177e-01 -3.62890810e-01 -1.67135671e-01 -1.13865459e+00
4.75532204e-01 -9.54432428e-01 3.12682122e-01 -2.19603109e+00
5.82843304e-01 -4.18512113e-02 1.40763894e-01 8.20147634e-01
-4.99149412e-01 1.53668568e-01 3.68642330e-01 4.24835384e-01
-6.95874333e-01 7.64409840e-01 1.70143902e+00 -1.87017351e-01
8.94003734e-02 -4.39288646e-01 -1.08489299e+00 1.10062730e+00
5.65130770e-01 -2.61883169e-01 -7.86592722e-01 -6.96114540e-01
3.63333859e-02 1.05224438e-01 8.26352537e-01 -6.74320459e-01
5.21909058e-01 -5.08159757e-01 4.78009194e-01 -5.72571456e-01
1.26669958e-01 -6.17097676e-01 -1.23234640e-03 5.06609738e-01
-4.57440108e-01 -1.05166085e-01 -6.44924641e-02 6.40733719e-01
-3.36390138e-01 -2.50645548e-01 5.20774126e-01 -3.77195030e-01
-1.00673270e+00 3.14289302e-01 -1.87898174e-01 -7.63645992e-02
1.06611443e+00 -4.97609764e-01 -5.22119164e-01 -4.61025000e-01
-5.95444322e-01 5.24376690e-01 3.91360283e-01 2.89314091e-01
1.28171599e+00 -1.42196381e+00 -7.59574771e-01 1.63593635e-01
2.19839126e-01 2.42504120e-01 4.52334911e-01 8.28655958e-01
-4.39258248e-01 3.53934616e-01 -2.65507370e-01 -5.89551747e-01
-1.18493593e+00 5.64156950e-01 2.48048589e-01 -1.07455797e-01
-7.68093944e-01 9.51209545e-01 8.01011205e-01 -1.06091209e-01
-6.41557947e-02 -3.50565135e-01 -3.00101668e-01 -3.14785331e-01
6.23347819e-01 -3.11889172e-01 -6.98984742e-01 -6.25187635e-01
-2.88763583e-01 7.40722954e-01 6.98119551e-02 -8.15931037e-02
8.05694640e-01 -3.81054610e-01 -3.40506285e-01 2.66976893e-01
5.32075942e-01 -1.17976129e-01 -1.41716015e+00 -1.63846880e-01
-4.57258582e-01 -3.40742022e-01 -4.79267463e-02 -8.66433263e-01
-1.13136828e+00 1.12950182e+00 -1.55258998e-01 -1.32284120e-01
1.14925492e+00 3.35909396e-01 8.06717873e-01 4.30169562e-03
1.75474897e-01 -6.15634859e-01 5.98631382e-01 4.05574173e-01
1.33121312e+00 -1.23763132e+00 3.51113528e-02 -7.89293468e-01
-1.16736031e+00 1.22508121e+00 1.20842409e+00 -4.35203798e-02
1.79222539e-01 -5.10354936e-02 1.57173634e-01 -2.54147738e-01
-9.10253108e-01 -4.23352063e-01 4.60656673e-01 7.67502367e-01
3.00616890e-01 -9.61409733e-02 -3.98687452e-01 9.38271224e-01
-1.52524099e-01 1.87734477e-02 4.06863332e-01 2.25959554e-01
-6.67678356e-01 -8.20724130e-01 -3.37499946e-01 1.06699824e-01
2.20838323e-01 -5.01612008e-01 -5.03054619e-01 4.70577776e-01
5.96421301e-01 1.00090957e+00 1.81634709e-01 -2.90803552e-01
-1.02932118e-02 -7.90055767e-02 4.58399802e-01 -8.19037497e-01
-2.49484897e-01 -7.84161836e-02 -6.73817620e-02 -7.83601403e-01
-7.25768447e-01 -2.42621720e-01 -2.02664804e+00 -1.85498208e-01
-2.13098884e-01 8.38946477e-02 3.81907970e-01 1.54460204e+00
3.97980183e-01 8.70603919e-01 2.97323108e-01 -6.63007855e-01
-1.07006773e-01 -6.17271304e-01 -2.20715106e-01 3.87200564e-01
8.14312771e-02 -5.51123917e-01 -2.60416090e-01 5.35676420e-01] | [10.819284439086914, 1.703214406967163] |
2cdd6204-4a93-468b-977c-78cdc4d0f5e2 | generative-view-correlation-adaptation-for | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2130_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123590307.pdf | Generative View-Correlation Adaptation for Semi-Supervised Multi-View Learning | Multi-view learning (MVL) explores the data extracted from multiple resources. It assumes that the complementary information between different views could be revealed to further improve the learning performance. There are two challenges. First, it is difficult to effectively combine the different view data together while still fully preserve the view-specific information. Second, multi-view datasets are usually small. This situation is easily cause overfitting for general model. To address the challenges, we propose a novel View-Correlation Adaptation ( extit{VCA}) framework in semi-supervised fashion. A semi-supervised data augmentation method is designed to generate extra features and labels based on labeled and even unlabeled samples. In addition, a cross-view adversarial training strategy is proposed to explore the structural information from one view and help the representation learning of the other view. Moreover, a simple yet effective fusion network is proposed for the late fusion stage. In our model, all networks are jointly trained in an end-to-end fashion. Extensive experiments demonstrate that our approach is effective and stable compared with other state-of-the-art methods. | ['Yunyu Liu', 'Lichen Wang', 'Yun Fu', 'Yue Bai', 'Can Qin', 'Zhengming Ding'] | null | null | null | null | eccv-2020-8 | ['multi-view-learning'] | ['computer-vision'] | [ 2.92737484e-01 -2.53481511e-02 -2.22725064e-01 -3.59507143e-01
-9.39238846e-01 -6.09231949e-01 5.51052451e-01 -3.41577381e-01
2.87137590e-02 5.54427028e-01 3.52210104e-01 1.50927082e-01
1.99544683e-01 -5.16632438e-01 -5.10478199e-01 -8.87227178e-01
5.21224678e-01 1.86629500e-02 2.34836675e-02 -1.14599757e-01
-2.84089386e-01 2.37092093e-01 -1.31100953e+00 4.82631475e-01
8.72904658e-01 9.43105817e-01 2.62327164e-01 1.83931574e-01
1.39429480e-01 8.76705468e-01 -3.11240613e-01 -5.72186947e-01
3.62812042e-01 -4.74156618e-01 -4.10069615e-01 6.02398336e-01
6.01557195e-01 -3.38260710e-01 -3.74221742e-01 1.16647375e+00
6.80598378e-01 7.32562542e-02 2.96177447e-01 -1.30328834e+00
-5.34289479e-01 3.01158488e-01 -9.26005542e-01 -5.00622876e-02
1.62661836e-01 -9.79358517e-03 8.00472260e-01 -1.38995945e+00
5.83508193e-01 1.37433708e+00 4.06671941e-01 5.69975674e-01
-1.10446632e+00 -7.64654100e-01 5.68772912e-01 -8.69162567e-03
-1.16682279e+00 -5.38580716e-01 1.30439508e+00 -1.62480384e-01
8.88016745e-02 1.64408043e-01 5.84120631e-01 1.34635663e+00
-2.84074876e-03 1.12165582e+00 1.53615820e+00 -2.91350335e-01
-1.37843400e-01 3.31017226e-01 -3.74551386e-01 8.34978521e-01
9.38394573e-03 2.16713220e-01 -5.79040468e-01 -1.42305866e-01
6.48236692e-01 6.39302373e-01 -4.08458859e-01 -8.11575353e-01
-1.31296027e+00 7.53570139e-01 5.11004448e-01 2.28422612e-01
-2.66355693e-01 -3.03627342e-01 4.37740475e-01 1.74811766e-01
4.39767271e-01 -3.03560458e-02 -5.19569755e-01 4.28547770e-01
-7.37207353e-01 -9.54659134e-02 3.82554203e-01 9.28157806e-01
8.31115723e-01 2.49512047e-01 -2.84319557e-02 8.82959664e-01
4.25847441e-01 5.50404608e-01 4.13252771e-01 -8.75528991e-01
9.27697361e-01 7.08154023e-01 -2.33577102e-01 -9.91817474e-01
-1.75454944e-01 -7.22342074e-01 -1.28361607e+00 2.38802239e-01
8.55940506e-02 -1.60610929e-01 -9.06966209e-01 1.93110204e+00
5.21489918e-01 2.94309944e-01 2.85546869e-01 7.69564211e-01
9.60164368e-01 4.01034802e-01 -1.30654588e-01 -3.80944818e-01
1.30677426e+00 -1.33794916e+00 -8.48349512e-01 -4.46433812e-01
3.50107253e-01 -7.61333644e-01 7.63364911e-01 3.52867752e-01
-1.00912011e+00 -7.57149577e-01 -1.23003197e+00 6.94987923e-02
-3.14042509e-01 3.68874818e-01 4.72974092e-01 4.00798947e-01
-5.48340380e-01 2.45258257e-01 -7.79803574e-01 4.77732830e-02
6.25829399e-01 1.32256404e-01 -7.25442588e-01 -5.87080956e-01
-1.02815628e+00 6.05472445e-01 4.24741834e-01 2.00309113e-01
-8.52367818e-01 -4.51150030e-01 -1.14225101e+00 -9.72340629e-02
7.11991847e-01 -8.09030235e-01 7.72388518e-01 -1.03131163e+00
-1.12853265e+00 6.67470634e-01 -1.57565624e-01 -2.25774683e-02
4.91249144e-01 -2.39864781e-01 -6.55184329e-01 2.60115683e-01
1.98867083e-01 4.71177965e-01 1.10990894e+00 -1.91584480e+00
-5.40910900e-01 -6.70457244e-01 1.48289338e-01 5.53037763e-01
-3.49036306e-01 -2.79238969e-01 -7.45610535e-01 -9.13069367e-01
4.18752760e-01 -1.01017296e+00 -2.77598083e-01 3.12634674e-03
-4.01536912e-01 2.64151305e-01 1.03978693e+00 -6.85150981e-01
1.00109494e+00 -2.04230833e+00 3.85537148e-01 2.70732399e-02
4.34026748e-01 3.75159472e-01 -8.28566551e-02 3.77267301e-01
-2.39595979e-01 -2.21147388e-01 -3.03389609e-01 -5.25302052e-01
-4.31281537e-01 1.57954380e-01 -2.57383347e-01 4.52619642e-01
1.38934225e-01 9.00875628e-01 -9.07526553e-01 -5.29926181e-01
3.52871954e-01 5.85476577e-01 -4.02432144e-01 4.39168513e-01
3.01308781e-01 9.00025725e-01 -4.68264788e-01 8.70793343e-01
9.20206606e-01 -3.88974726e-01 2.43969172e-01 -6.96655393e-01
3.57854784e-01 -1.95525333e-01 -1.19468284e+00 2.11559224e+00
-5.85158885e-01 -5.97575307e-03 1.03627600e-01 -9.75344241e-01
7.29398668e-01 4.23562586e-01 4.61917669e-01 -3.91517907e-01
1.98174231e-02 6.44909963e-02 -9.82998237e-02 -3.45212519e-01
1.48676962e-01 -4.84499872e-01 -8.48652795e-02 4.16406065e-01
3.66398066e-01 2.16674849e-01 -2.10860938e-01 1.75197765e-01
5.56839526e-01 3.63542080e-01 5.66837788e-01 2.53300935e-01
9.68848050e-01 -5.99644542e-01 9.81248856e-01 2.72910923e-01
-2.33467504e-01 8.54236007e-01 1.72341838e-01 -3.04390341e-01
-7.85171568e-01 -9.63893592e-01 2.89345890e-01 9.03516650e-01
2.39025339e-01 -3.91981065e-01 -5.72386146e-01 -1.28034568e+00
-2.05618173e-01 3.80633265e-01 -6.40277028e-01 -3.88551623e-01
-3.98969948e-01 -5.43901145e-01 1.35824278e-01 8.04331839e-01
7.13741601e-01 -6.62933409e-01 -1.98878050e-01 6.71663359e-02
-3.12886715e-01 -1.45066261e+00 -6.94780052e-01 -8.72999895e-03
-1.03731799e+00 -1.10422981e+00 -6.91898882e-01 -5.50037920e-01
9.69747424e-01 8.71237457e-01 9.04799342e-01 -4.60209772e-02
3.12924415e-01 3.87101918e-01 -4.00758713e-01 -1.92634538e-01
-3.22804868e-01 -1.52464941e-01 1.23229846e-01 4.95047510e-01
2.70687580e-01 -7.21833706e-01 -6.06731236e-01 3.75011712e-01
-9.98489380e-01 3.25976908e-01 7.85276294e-01 1.28539908e+00
8.26379895e-01 5.83549626e-02 6.09506786e-01 -1.20797920e+00
6.16710372e-02 -3.35510015e-01 -2.02255160e-01 5.82141876e-01
-5.53498209e-01 -1.25190645e-01 8.75205457e-01 -3.01790148e-01
-1.19001281e+00 2.96037197e-01 9.29594785e-02 -1.16608405e+00
-9.93053913e-02 4.51466352e-01 -8.63321483e-01 -4.59657609e-02
1.38776571e-01 4.37721014e-01 1.26021400e-01 -4.41867620e-01
6.04839206e-01 3.68464142e-01 5.54293334e-01 -3.35943192e-01
1.14118135e+00 7.38393128e-01 2.56214626e-02 -3.52186680e-01
-1.30631578e+00 -4.55195427e-01 -9.30002093e-01 -1.83166847e-01
6.22368395e-01 -1.52250016e+00 -2.46274069e-01 4.83252376e-01
-7.94952393e-01 3.67594391e-01 -1.68586180e-01 5.30216157e-01
-4.91590559e-01 5.62530518e-01 -1.87815458e-01 -7.01712728e-01
-3.71108532e-01 -1.24532402e+00 1.02280653e+00 3.04769456e-01
4.00331706e-01 -1.13776469e+00 -2.07122356e-01 8.47981095e-01
1.78041141e-02 4.14043158e-01 4.27962512e-01 -8.56824875e-01
-4.82143372e-01 -2.69479781e-01 -1.84924692e-01 7.59645045e-01
4.41223890e-01 -3.59512329e-01 -1.11127496e+00 -7.51922786e-01
3.71320635e-01 -6.15587056e-01 9.38046634e-01 -3.07517946e-02
1.15535164e+00 -2.33465061e-01 -2.44551048e-01 6.43601120e-01
1.48262274e+00 -6.19869418e-02 3.56703699e-01 -2.03955900e-02
1.29398263e+00 5.71369469e-01 7.92246342e-01 2.75302231e-01
6.03662074e-01 4.82742578e-01 4.77648944e-01 -3.30279291e-01
-9.76874009e-02 -5.78788221e-01 4.46645677e-01 1.19645619e+00
3.20234112e-02 -1.94338635e-01 -4.96673524e-01 3.03562015e-01
-1.83877456e+00 -9.73180354e-01 2.56932497e-01 2.18554521e+00
3.68800014e-01 6.71793893e-02 -1.11940183e-01 3.77977230e-02
7.23471761e-01 6.17145956e-01 -6.37861133e-01 2.65286267e-01
-2.21990436e-01 -2.18858823e-01 2.73142874e-01 1.29539147e-01
-1.36219954e+00 6.45287037e-01 5.17923355e+00 7.73852766e-01
-1.05545712e+00 3.71055424e-01 6.52505696e-01 -8.02598596e-02
-4.72300112e-01 3.09979003e-02 -6.01565957e-01 4.25709873e-01
4.50214237e-01 1.26532421e-01 3.05605233e-01 7.92956769e-01
-1.38403118e-01 2.25821733e-01 -8.86163056e-01 1.11061299e+00
5.69701850e-01 -9.70299184e-01 1.57981634e-01 1.11208178e-01
9.56714332e-01 -2.14557618e-01 1.59080595e-01 4.22094345e-01
1.26834393e-01 -6.18284345e-01 4.57117707e-01 5.73615968e-01
9.15989876e-01 -1.05018175e+00 7.04141259e-01 5.13640285e-01
-1.47262633e+00 -2.79721528e-01 -2.29981557e-01 3.96147043e-01
2.74257094e-01 5.01084864e-01 -3.09144825e-01 1.28971124e+00
3.44428271e-01 1.05276954e+00 -7.08125770e-01 5.21382749e-01
-3.28841001e-01 2.36539871e-01 -8.66421591e-03 8.62673759e-01
2.63802677e-01 -2.62744665e-01 5.52183032e-01 6.89703107e-01
2.70206839e-01 6.71548769e-02 5.81469953e-01 5.53626060e-01
-2.14264899e-01 -5.59436455e-02 -8.83926392e-01 1.72075987e-01
4.08468246e-01 1.62787092e+00 -3.77835572e-01 -3.35361540e-01
-9.90565658e-01 1.09966123e+00 4.35999781e-01 3.01567078e-01
-6.96329117e-01 1.70688555e-01 2.42563203e-01 -2.00328276e-01
5.03094792e-01 9.24705118e-02 2.49859486e-02 -1.81583214e+00
1.63390785e-01 -1.26182234e+00 4.90105867e-01 -7.14909434e-01
-1.61245430e+00 5.06727576e-01 -1.73004106e-01 -1.99674308e+00
-1.71417817e-01 -3.08597863e-01 -5.96138060e-01 9.16548610e-01
-1.58337855e+00 -1.89116311e+00 -3.71314794e-01 7.59048581e-01
5.59971511e-01 -6.12887323e-01 7.52589345e-01 3.25544685e-01
-6.06130183e-01 7.79489398e-01 6.96992129e-02 1.80371895e-01
8.85300934e-01 -1.20448422e+00 4.38885614e-02 1.12896717e+00
2.33091041e-01 6.31394804e-01 1.66438371e-01 -5.24917066e-01
-1.42765427e+00 -1.24636567e+00 2.97751367e-01 -3.73194218e-01
3.36177498e-01 -2.85876065e-01 -1.04654908e+00 7.35385895e-01
3.36824626e-01 5.37362337e-01 1.11698568e+00 -1.11373942e-02
-6.24804378e-01 -3.08212727e-01 -1.02479041e+00 3.84462506e-01
8.51931095e-01 -7.01737285e-01 -5.33666074e-01 1.28516806e-02
7.25754619e-01 -5.06525338e-01 -9.87519860e-01 7.82320857e-01
5.24092376e-01 -1.11526752e+00 1.05604589e+00 -6.68473601e-01
5.29597223e-01 -4.16587681e-01 -4.85753447e-01 -1.51187050e+00
-2.05938026e-01 -4.91900414e-01 -6.48250878e-01 1.58004034e+00
1.07140817e-01 -5.55822492e-01 6.82189941e-01 3.18256527e-01
-2.90816147e-02 -1.08226395e+00 -8.20906222e-01 -5.84444582e-01
-1.07947245e-01 -5.94938137e-02 5.53343415e-01 1.22247756e+00
-3.49906713e-01 7.52608597e-01 -8.11201751e-01 2.92126179e-01
8.90243351e-01 5.19469380e-01 8.53470445e-01 -1.07096791e+00
-3.97798687e-01 1.27888456e-01 -2.34856918e-01 -8.46998632e-01
1.66263714e-01 -9.46681678e-01 -3.26158106e-01 -1.33790767e+00
5.42764723e-01 -2.08926439e-01 -6.79140508e-01 3.01639140e-01
-6.95193946e-01 9.39461514e-02 5.64789176e-01 3.02399874e-01
-6.22716010e-01 9.46163833e-01 1.68774319e+00 6.16524890e-02
1.41001672e-01 1.87573478e-01 -1.00846374e+00 9.93245661e-01
5.14204085e-01 -3.46955985e-01 -7.98353493e-01 -2.33548567e-01
-1.94682121e-01 4.35473442e-01 3.84466350e-01 -9.27410543e-01
1.74731463e-01 9.03504714e-02 7.29164124e-01 -9.84646916e-01
5.15603364e-01 -1.22562873e+00 -1.74472500e-02 1.39147043e-01
-1.22351184e-01 7.44845048e-02 -1.16904967e-01 1.06314766e+00
-5.37274599e-01 5.73666319e-02 7.77200520e-01 -3.20545405e-01
-4.32359636e-01 5.87591529e-01 3.09633315e-01 -2.11741123e-02
1.03924739e+00 2.53960434e-02 -1.92610130e-01 -4.57707942e-01
-8.03249478e-01 4.66501176e-01 5.18017709e-01 5.22307456e-01
6.40379012e-01 -1.87818635e+00 -5.76008499e-01 4.05404836e-01
4.45593297e-01 1.72978953e-01 7.16766059e-01 8.60178530e-01
1.57249928e-01 3.97532210e-02 -2.09999651e-01 -5.12310863e-01
-1.29757583e+00 1.04456890e+00 1.74180552e-01 -7.08927631e-01
-4.30655539e-01 6.11159623e-01 5.83643794e-01 -7.47988582e-01
-6.57452270e-02 2.23242685e-01 -3.45069230e-01 3.92306745e-01
6.35574758e-01 1.63937852e-01 -2.26221323e-01 -9.69087720e-01
-2.73540795e-01 5.83620131e-01 -3.45687002e-01 -2.35224664e-02
1.24132478e+00 -4.43693429e-01 8.05340186e-02 5.94986498e-01
1.37399364e+00 1.99309424e-01 -1.45490777e+00 -7.55258203e-01
-5.97409546e-01 -5.73919773e-01 -7.52223702e-03 -7.60529101e-01
-1.64254797e+00 1.04856741e+00 6.18136108e-01 -1.24975331e-01
1.42635655e+00 -2.82896280e-01 7.96738267e-01 1.41004056e-01
3.10667336e-01 -9.13099170e-01 2.32732862e-01 7.38625452e-02
8.63990784e-01 -1.62221909e+00 3.96612585e-01 -6.30029261e-01
-1.07877243e+00 9.61571395e-01 8.97014678e-01 1.30136043e-01
6.89002037e-01 -7.12236837e-02 1.16948262e-01 -2.01942667e-01
-7.82690167e-01 -1.79849833e-01 4.24026579e-01 4.84877795e-01
1.28348202e-01 -1.47827357e-01 1.17489494e-01 6.52429640e-01
4.65798050e-01 -1.38293952e-01 3.02265108e-01 9.89705622e-01
-5.11018224e-02 -1.19483840e+00 -4.04557198e-01 3.59370440e-01
-5.82392097e-01 4.22149710e-02 -1.67717427e-01 6.88163638e-01
2.70122468e-01 7.86540926e-01 -3.83424431e-01 -6.57712519e-01
3.49857062e-01 1.12598285e-01 3.68899673e-01 -6.02278411e-01
-4.10201341e-01 5.55393875e-01 -1.09383926e-01 -5.32158017e-01
-9.29209292e-01 -8.38281512e-01 -8.31200302e-01 8.30543190e-02
-4.74502087e-01 -1.69273734e-01 2.25019053e-01 9.70516503e-01
3.92279416e-01 6.98351979e-01 1.19398177e+00 -7.64746130e-01
-6.03558660e-01 -8.79240870e-01 -5.62455356e-01 5.49275696e-01
3.79212856e-01 -8.56430531e-01 -2.28069916e-01 1.06756635e-01] | [8.493098258972168, 4.541285991668701] |
a5f9868b-6708-4a22-951d-d7163d069975 | learning-graph-structure-with-a-finite-state | 2007.04929 | null | https://arxiv.org/abs/2007.04929v2 | https://arxiv.org/pdf/2007.04929v2.pdf | Learning Graph Structure With A Finite-State Automaton Layer | Graph-based neural network models are producing strong results in a number of domains, in part because graphs provide flexibility to encode domain knowledge in the form of relational structure (edges) between nodes in the graph. In practice, edges are used both to represent intrinsic structure (e.g., abstract syntax trees of programs) and more abstract relations that aid reasoning for a downstream task (e.g., results of relevant program analyses). In this work, we study the problem of learning to derive abstract relations from the intrinsic graph structure. Motivated by their power in program analyses, we consider relations defined by paths on the base graph accepted by a finite-state automaton. We show how to learn these relations end-to-end by relaxing the problem into learning finite-state automata policies on a graph-based POMDP and then training these policies using implicit differentiation. The result is a differentiable Graph Finite-State Automaton (GFSA) layer that adds a new edge type (expressed as a weighted adjacency matrix) to a base graph. We demonstrate that this layer can find shortcuts in grid-world graphs and reproduce simple static analyses on Python programs. Additionally, we combine the GFSA layer with a larger graph-based model trained end-to-end on the variable misuse program understanding task, and find that using the GFSA layer leads to better performance than using hand-engineered semantic edges or other baseline methods for adding learned edge types. | ['Hugo Larochelle', 'Daniel Tarlow', 'Daniel D. Johnson'] | 2020-07-09 | null | http://proceedings.neurips.cc/paper/2020/hash/1fdc0ee9d95c71d73df82ac8f0721459-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/1fdc0ee9d95c71d73df82ac8f0721459-Paper.pdf | neurips-2020-12 | ['variable-misuse'] | ['computer-code'] | [ 3.35905522e-01 8.43018234e-01 -5.39603889e-01 -3.75631124e-01
-3.79996359e-01 -8.13112020e-01 5.59966803e-01 5.68054259e-01
3.99771146e-03 3.60556930e-01 2.03363180e-01 -1.07437766e+00
2.17728410e-03 -1.43428338e+00 -1.29573834e+00 -8.06752220e-02
-5.06884754e-01 3.59291613e-01 4.40269262e-01 -4.17584151e-01
4.26717661e-03 4.17320818e-01 -1.29577458e+00 2.68878132e-01
6.70532882e-01 6.14498079e-01 -2.32577860e-01 7.85243094e-01
-2.58553356e-01 1.11954844e+00 -3.02417219e-01 -5.26465416e-01
1.02161579e-01 -2.80327529e-01 -1.13002729e+00 -2.01758295e-01
2.74317443e-01 -1.92280337e-01 -1.05133563e-01 1.10738719e+00
-2.32348084e-01 2.02852488e-01 5.00548184e-01 -1.43856061e+00
-5.55991888e-01 1.18362176e+00 -2.41994679e-01 -2.13509356e-03
3.85677189e-01 4.22080576e-01 1.51161611e+00 2.39284476e-03
9.15623307e-01 1.45872378e+00 7.89767802e-01 5.58672547e-01
-1.72348619e+00 -1.46328092e-01 3.63002151e-01 -1.63325414e-01
-8.52632523e-01 -1.20563023e-01 7.26815760e-01 -5.14865935e-01
1.63721371e+00 1.36609599e-01 7.02863932e-01 8.75868857e-01
4.81695205e-01 5.30077815e-01 6.34265304e-01 -4.32114631e-01
2.83239901e-01 -3.73640269e-01 6.18427455e-01 1.42911208e+00
2.81634033e-01 1.60163701e-01 -1.48176804e-01 -4.02546197e-01
6.40669942e-01 -3.10616285e-01 3.91118750e-02 -8.37216198e-01
-7.30505645e-01 9.79162991e-01 7.78243721e-01 1.17752049e-02
-1.08043857e-01 8.67292941e-01 6.28574312e-01 3.86680186e-01
1.06895976e-01 5.72475791e-01 -4.79011208e-01 1.01879314e-02
-4.12569553e-01 2.34515026e-01 1.31857920e+00 7.87546933e-01
9.79619086e-01 3.38372290e-02 -5.98005839e-02 1.84230968e-01
4.53813434e-01 2.08692610e-01 5.41831041e-03 -1.10049081e+00
4.69360501e-01 9.60908115e-01 -2.41747528e-01 -8.88550103e-01
-4.81613576e-01 -2.95542210e-01 -2.56338716e-01 2.63165414e-01
6.54210150e-01 -1.57476887e-01 -8.34497094e-01 2.11851597e+00
3.74566078e-01 1.33906454e-01 6.86357841e-02 4.96683508e-01
5.22157848e-01 6.35862350e-01 2.45709404e-01 1.89386547e-01
1.25235379e+00 -7.66793549e-01 -1.74937606e-01 -5.37207007e-01
1.37037635e+00 3.19651961e-01 1.17022192e+00 1.53389767e-01
-9.41686630e-01 -1.57573432e-01 -1.16297746e+00 -1.99654773e-01
-5.83112717e-01 -5.89817226e-01 9.66086686e-01 4.81352597e-01
-1.40377903e+00 9.11863565e-01 -1.15336442e+00 -1.89315915e-01
3.65564466e-01 6.61110997e-01 -2.87264109e-01 1.40824020e-01
-1.20679760e+00 7.58775294e-01 7.03561008e-01 -2.13892028e-01
-1.01912987e+00 -5.67214727e-01 -1.51767409e+00 2.58563668e-01
8.44790339e-01 -8.99875998e-01 1.40227747e+00 -1.00519860e+00
-1.32926059e+00 7.84987450e-01 5.57549410e-02 -8.04070294e-01
3.75375338e-02 2.85900533e-01 -1.58681244e-01 -9.08838063e-02
-9.04440954e-02 5.43389738e-01 7.20327914e-01 -1.02447081e+00
-4.69285727e-01 -3.60263944e-01 1.00597179e+00 -2.37806767e-01
2.94926204e-02 -4.28529121e-02 -4.05547351e-01 -1.55336618e-01
-1.22608706e-01 -1.14346874e+00 -2.97816604e-01 -5.08501381e-02
-5.04740775e-01 -3.90510648e-01 5.95022202e-01 -8.10102820e-01
1.20275962e+00 -1.91079879e+00 3.71540815e-01 4.19432223e-01
4.53790426e-01 2.20865496e-02 -1.11504413e-01 3.73085439e-01
-1.42247036e-01 3.82983834e-01 -7.05765963e-01 8.98027644e-02
3.48742485e-01 6.81328356e-01 -2.17284679e-01 2.84728318e-01
6.42046273e-01 1.24095547e+00 -1.31284010e+00 -3.19403023e-01
-1.08836628e-01 5.91126271e-02 -9.15079594e-01 1.39909945e-02
-1.13009715e+00 1.02545612e-01 -4.78190690e-01 3.27114522e-01
2.07439050e-01 -4.02622789e-01 6.89289629e-01 1.65265337e-01
3.78440052e-01 7.23146439e-01 -1.10357726e+00 1.81151962e+00
-8.90424013e-01 3.97092819e-01 1.48368031e-01 -9.98352468e-01
6.27708495e-01 -1.61046863e-01 -4.92353551e-02 -4.17975485e-01
-1.07581377e-01 -1.21186502e-01 3.67449522e-01 -3.88509721e-01
3.65688264e-01 -1.02846280e-01 -5.79537809e-01 6.96659327e-01
1.41203269e-01 -4.10150230e-01 2.45428428e-01 5.32070816e-01
1.46723604e+00 6.21479094e-01 4.24579352e-01 -3.18883717e-01
6.06778204e-01 2.59087116e-01 5.79436183e-01 7.20863044e-01
1.66150451e-01 -8.17130059e-02 1.41403234e+00 -5.29233038e-01
-7.54518509e-01 -9.72345412e-01 3.67177457e-01 1.26032889e+00
-1.51732191e-01 -8.20919633e-01 -8.60543430e-01 -1.31651413e+00
1.55329764e-01 9.54857528e-01 -5.81908941e-01 -6.25822425e-01
-9.31135714e-01 -2.49446556e-01 8.14675987e-01 8.68910015e-01
2.33993635e-01 -1.06695628e+00 -6.25012219e-01 2.19735965e-01
7.80213773e-02 -9.11742330e-01 -3.42015773e-01 4.77446675e-01
-9.85879660e-01 -1.17786419e+00 3.62344503e-01 -7.16529608e-01
7.64944553e-01 -4.33394372e-01 1.66716504e+00 4.43939090e-01
1.14810199e-01 5.31173468e-01 -1.83474855e-03 -2.17705503e-01
-9.20827150e-01 2.81251460e-01 -4.70654935e-01 -4.36610758e-01
1.51523367e-01 -6.63836539e-01 7.51334578e-02 -3.66630293e-02
-1.11973739e+00 1.78676993e-02 6.22776747e-02 7.53146946e-01
4.17682588e-01 9.25103296e-03 1.79154441e-01 -1.32798302e+00
8.79755080e-01 -5.04491389e-01 -1.23486209e+00 2.53890038e-01
-5.33823848e-01 9.17631984e-01 9.66843367e-01 -6.78534210e-02
-8.01445425e-01 1.36800408e-01 8.93284939e-03 -1.19232886e-01
-3.36485393e-02 1.08828115e+00 -2.28462771e-01 -5.45459101e-03
9.36748743e-01 -1.51071027e-01 1.98033214e-01 4.03203517e-02
7.01581061e-01 7.32042491e-02 6.04612172e-01 -1.36445022e+00
7.80135691e-01 1.02533214e-01 5.71461797e-01 -2.72522092e-01
-5.85780442e-01 7.92206228e-02 -5.27438581e-01 2.77070671e-01
8.13487113e-01 -4.57801282e-01 -9.19432640e-01 9.29918960e-02
-1.18500948e+00 -1.07598579e+00 -3.97023559e-01 -1.15540750e-01
-6.14963174e-01 1.69190571e-01 -8.44051003e-01 -7.03229308e-01
-1.14283130e-01 -1.41179788e+00 1.07230926e+00 -6.69800118e-02
-4.52644408e-01 -1.36663294e+00 3.80049795e-02 -1.34643599e-01
9.43042040e-02 3.71452600e-01 1.72255743e+00 -7.10557640e-01
-7.71128654e-01 -4.46135700e-02 -4.89560924e-02 2.60521919e-01
-7.57252946e-02 1.14192583e-01 -7.07823753e-01 -3.61200273e-02
-2.85620898e-01 -2.66681939e-01 6.73164904e-01 2.02434063e-01
1.10485911e+00 -4.49275762e-01 -3.94588262e-01 5.91158092e-01
1.30415332e+00 8.53883382e-03 4.73267436e-01 1.65802583e-01
8.60626519e-01 5.06733119e-01 7.04509541e-02 -6.92612007e-02
5.48301041e-01 4.42614138e-01 7.83591866e-01 2.96055377e-01
-7.99372345e-02 -6.99012756e-01 4.96566713e-01 3.11973631e-01
1.99433602e-02 -6.33172169e-02 -1.39434254e+00 4.43951279e-01
-1.91529906e+00 -5.53437412e-01 -8.63603726e-02 2.14210939e+00
8.98573339e-01 4.76426065e-01 3.12534213e-01 -1.37677222e-01
4.09612685e-01 3.45675439e-01 -7.05108583e-01 -7.80670106e-01
4.53454524e-01 4.51957494e-01 6.25618637e-01 1.15396059e+00
-8.63459945e-01 1.14268804e+00 5.45516300e+00 3.40529084e-01
-1.06114125e+00 -1.30576491e-01 3.17923516e-01 4.35374737e-01
-5.37425339e-01 4.57815051e-01 -5.24902999e-01 -9.10950899e-02
1.22785866e+00 -1.97615877e-01 1.02887559e+00 8.09915662e-01
-3.80654037e-01 5.21414839e-02 -1.72376430e+00 2.17384070e-01
-4.39087451e-01 -1.43060768e+00 -5.02170287e-02 -9.80605837e-03
3.45538139e-01 1.17616616e-02 -2.07491487e-01 7.24731684e-01
1.15119493e+00 -9.99662340e-01 6.65816605e-01 1.17347665e-01
5.75400233e-01 -6.37846053e-01 3.84095758e-01 4.19370115e-01
-1.21739948e+00 -7.98736885e-02 3.37081030e-02 -3.19800884e-01
-1.91006124e-01 9.36547518e-02 -1.05181587e+00 7.08938181e-01
8.40177909e-02 7.58066595e-01 -5.21062016e-01 4.30068314e-01
-8.81728411e-01 5.40505528e-01 -3.24730426e-01 -1.47846311e-01
4.66101915e-01 -3.30079287e-01 5.00148475e-01 1.15051818e+00
-2.07476676e-01 1.88032277e-02 1.43433660e-01 1.39411175e+00
-3.81245822e-01 -5.37243128e-01 -1.02496767e+00 -4.20626014e-01
8.44367743e-02 1.18306160e+00 -7.83707082e-01 -4.92291808e-01
-5.46246350e-01 4.99306887e-01 7.09602654e-01 4.91295695e-01
-7.40834355e-01 -3.04346710e-01 5.91044724e-01 1.67352617e-01
2.55443305e-01 -4.12483335e-01 -1.46695852e-01 -1.07001913e+00
6.22252841e-03 -1.16611874e+00 7.06304789e-01 -7.23884106e-01
-8.79685104e-01 3.03691089e-01 2.84880072e-01 -4.44534779e-01
-6.64323568e-01 -7.03915298e-01 -7.21644759e-01 1.01823747e+00
-1.06835806e+00 -9.04691100e-01 1.02936983e-01 4.04904127e-01
-1.01501625e-02 3.40899169e-01 8.98857653e-01 -2.70080149e-01
-4.41167027e-01 3.95742327e-01 -6.19370461e-01 3.85327071e-01
-1.77844733e-01 -1.45909977e+00 1.11098385e+00 1.10289693e+00
2.78955281e-01 9.77637112e-01 6.44184351e-01 -8.29519808e-01
-2.03617334e+00 -1.10623360e+00 4.92371440e-01 -6.58204257e-01
1.12751687e+00 -5.23687780e-01 -1.01623857e+00 1.44923818e+00
5.30363582e-02 4.14271265e-01 7.00273216e-02 4.09598947e-01
-7.12767065e-01 6.56050444e-02 -1.10545874e+00 7.45712280e-01
1.56877422e+00 -8.83624852e-01 -6.70986414e-01 1.15738072e-01
1.11285007e+00 -7.99145222e-01 -8.57973456e-01 1.65862352e-01
2.37677842e-01 -5.79057276e-01 8.79673600e-01 -1.30994463e+00
6.96060181e-01 -3.84103686e-01 -1.49281725e-01 -1.61263072e+00
-1.49925500e-01 -6.77001834e-01 -5.14361739e-01 8.30207288e-01
7.12082028e-01 -9.60838199e-01 8.33422005e-01 9.35697377e-01
-2.98995584e-01 -6.90656364e-01 -5.15275180e-01 -6.84114337e-01
2.06119508e-01 -7.38624990e-01 7.83072352e-01 9.00772333e-01
4.32278425e-01 6.82407856e-01 3.73537570e-01 4.11979586e-01
4.70340252e-01 3.83935124e-01 7.34313011e-01 -1.26763189e+00
-8.60350668e-01 -5.78135967e-01 -3.09675336e-01 -9.09865916e-01
8.37950408e-01 -1.51298249e+00 5.25990920e-03 -1.74858880e+00
-1.21695347e-01 -4.33201671e-01 -5.06219901e-02 1.00305033e+00
-3.67441103e-02 -5.88100195e-01 1.13557853e-01 -3.59172761e-01
-3.67247403e-01 1.79793522e-01 1.08693862e+00 -5.64460397e-01
-2.90330082e-01 -1.35813907e-01 -7.49640524e-01 8.02430212e-01
6.92412615e-01 -6.60809636e-01 -6.82203650e-01 -4.63833421e-01
6.72534347e-01 5.75231016e-01 5.28350413e-01 -6.02217078e-01
1.28147930e-01 -3.36919665e-01 -3.42789024e-01 1.85725704e-01
1.00103192e-01 -7.75352597e-01 1.81089118e-01 7.76540279e-01
-6.00436985e-01 6.25285804e-02 4.60429519e-01 6.10669255e-01
5.10700829e-02 -4.72578079e-01 3.27783614e-01 -3.21814865e-01
-8.34939301e-01 9.26830992e-02 -1.47763923e-01 3.57091337e-01
7.34436691e-01 1.20071724e-01 -6.14552319e-01 -3.64232302e-01
-8.59566689e-01 2.69688368e-01 6.80026829e-01 3.51776004e-01
5.40988371e-02 -9.25165594e-01 -8.83158371e-02 1.94112524e-01
1.03041038e-01 3.93351674e-01 -4.78070557e-01 6.84670568e-01
-5.22313774e-01 8.79369080e-02 -1.88211769e-01 -5.10998964e-01
-9.83412325e-01 7.80462801e-01 6.22447193e-01 -7.10524023e-01
-6.14692926e-01 6.20823801e-01 1.80936232e-01 -8.74718130e-01
3.66372131e-02 -1.13082528e+00 1.14514872e-01 -3.64183038e-01
4.04550880e-02 -5.01308516e-02 1.62784442e-01 -2.09034190e-01
-4.22377586e-01 1.80769563e-01 2.65935570e-01 -3.46049856e-05
1.30663824e+00 5.94086766e-01 -3.94847929e-01 2.78819680e-01
9.61942375e-01 4.57913205e-02 -1.15774393e+00 -2.74545580e-01
3.19501907e-01 -9.40174516e-03 -7.65458196e-02 -6.58957422e-01
-1.06194043e+00 8.28772902e-01 -2.22635344e-01 5.49572825e-01
7.58450627e-01 2.18402207e-01 4.63760912e-01 7.92727649e-01
6.18124485e-01 -7.25876808e-01 -2.03239456e-01 8.22995663e-01
5.97868800e-01 -9.00127590e-01 -1.41868144e-01 -5.61967552e-01
-3.19356412e-01 1.18093443e+00 5.91689944e-01 -2.47898787e-01
3.27074766e-01 5.64528286e-01 -4.87384647e-01 -3.80948514e-01
-6.96390331e-01 -1.33556262e-01 1.12411365e-01 8.53525639e-01
3.06981981e-01 1.72441602e-01 -3.12378071e-02 6.80981576e-01
-1.02200665e-01 1.30168840e-01 6.74544632e-01 1.15057588e+00
-9.06930417e-02 -1.31292808e+00 -1.40291318e-01 6.52604759e-01
-2.94986278e-01 -2.60917127e-01 -4.52987880e-01 1.00824571e+00
-1.58493459e-01 6.65261149e-01 6.86330907e-03 -3.92474025e-01
3.59449089e-01 2.56479293e-01 7.82760739e-01 -1.00323641e+00
-8.07199717e-01 -6.61466599e-01 7.24077225e-01 -7.85007536e-01
-3.69098857e-02 -5.04504323e-01 -1.78294599e+00 -2.98261255e-01
1.27647966e-02 8.20926130e-02 3.88350278e-01 8.96121740e-01
3.77767444e-01 9.27241385e-01 -7.98511133e-02 -5.98869979e-01
-6.09911263e-01 -4.52227980e-01 -3.18194151e-01 4.33569491e-01
2.71117568e-01 -3.93064857e-01 -1.40571386e-01 1.71490442e-02] | [8.81637191772461, 7.350799560546875] |
d8559872-3eb7-4ed1-9172-9342bddf0187 | learning-to-scale-temperature-in-masked-self | 2302.06130 | null | https://arxiv.org/abs/2302.06130v1 | https://arxiv.org/pdf/2302.06130v1.pdf | Learning to Scale Temperature in Masked Self-Attention for Image Inpainting | Recent advances in deep generative adversarial networks (GAN) and self-attention mechanism have led to significant improvements in the challenging task of inpainting large missing regions in an image. These methods integrate self-attention mechanism in neural networks to utilize surrounding neural elements based on their correlation and help the networks capture long-range dependencies. Temperature is a parameter in the Softmax function used in the self-attention, and it enables biasing the distribution of attention scores towards a handful of similar patches. Most existing self-attention mechanisms in image inpainting are convolution-based and set the temperature as a constant, performing patch matching in a limited feature space. In this work, we analyze the artifacts and training problems in previous self-attention mechanisms, and redesign the temperature learning network as well as the self-attention mechanism to address them. We present an image inpainting framework with a multi-head temperature masked self-attention mechanism, which provides stable and efficient temperature learning and uses multiple distant contextual information for high quality image inpainting. In addition to improving image quality of inpainting results, we generalize the proposed model to user-guided image editing by introducing a new sketch generation method. Extensive experiments on various datasets such as Paris StreetView, CelebA-HQ and Places2 clearly demonstrate that our method not only generates more natural inpainting results than previous works both in terms of perception image quality and quantitative metrics, but also enables to help users to generate more flexible results that are related to their sketch guidance. | ['Yi Gong', 'Yuan Zeng', 'Xiang Zhou'] | 2023-02-13 | null | null | null | null | ['image-inpainting', 'patch-matching'] | ['computer-vision', 'computer-vision'] | [ 1.69156268e-01 -2.17779100e-01 6.28851578e-02 -4.08897787e-01
-6.77854478e-01 -2.06140056e-01 4.01324570e-01 -5.98864615e-01
-1.40778139e-01 7.87576854e-01 2.95814186e-01 1.40606001e-01
2.89913714e-01 -1.13393438e+00 -1.30029023e+00 -7.85744429e-01
2.39911705e-01 4.84505929e-02 1.59331143e-01 -4.57116753e-01
2.68437535e-01 5.08773029e-01 -1.32317281e+00 5.01210570e-01
1.27129126e+00 1.00039208e+00 4.61493373e-01 6.26827002e-01
-1.46090120e-01 7.43787169e-01 -8.66419077e-01 -4.34542686e-01
4.21340644e-01 -5.59355378e-01 -4.96070176e-01 -1.51507626e-03
6.62727237e-01 -5.23274601e-01 -5.16098678e-01 8.62909853e-01
6.65275276e-01 2.47462273e-01 5.88216126e-01 -1.12349498e+00
-1.30447435e+00 3.84287536e-01 -7.73638785e-01 4.49276343e-02
2.00438201e-01 2.82494813e-01 6.53414607e-01 -7.04634547e-01
3.09431791e-01 1.33402514e+00 8.13926697e-01 7.21359015e-01
-1.23045206e+00 -8.56646419e-01 7.34426677e-02 3.30747813e-01
-1.20667541e+00 -2.58674592e-01 1.34471452e+00 -1.09575778e-01
7.79180884e-01 3.10987860e-01 6.83598459e-01 1.35030663e+00
5.03465474e-01 7.64981925e-01 1.22872961e+00 -4.71866578e-01
4.55591157e-02 8.29162367e-04 -6.94176674e-01 6.12954080e-01
-3.47526222e-01 2.52086312e-01 -3.69377106e-01 -2.18012050e-04
1.40060544e+00 3.28708082e-01 -1.80154502e-01 -7.24709332e-02
-1.02366638e+00 8.22283506e-01 9.61225152e-01 2.78662950e-01
-5.33814430e-01 5.26989758e-01 2.40891725e-01 3.70669127e-01
6.74200177e-01 5.50707817e-01 -2.21733987e-01 2.45703131e-01
-1.10427082e+00 2.42663980e-01 1.32107750e-01 9.29705620e-01
9.23227906e-01 3.84594500e-01 -6.12223744e-01 1.14101648e+00
-1.15721948e-01 5.06786704e-01 7.13569045e-01 -8.70550811e-01
3.98307562e-01 4.22770262e-01 3.64352986e-02 -1.06633174e+00
1.86176941e-01 -3.52839857e-01 -1.21425831e+00 5.69603205e-01
-4.17277627e-02 -7.85462707e-02 -1.23762453e+00 1.92915750e+00
9.62546170e-02 3.59221458e-01 -1.86245993e-01 1.02389622e+00
7.66543031e-01 8.90591919e-01 -3.48804221e-02 4.45106700e-02
1.19297874e+00 -1.37320030e+00 -8.30883145e-01 -2.41655484e-01
-1.08130798e-01 -9.55043554e-01 1.38736331e+00 3.17489415e-01
-1.32847738e+00 -1.18868685e+00 -1.13069201e+00 -2.98655123e-01
-3.48823935e-01 -7.02731758e-02 5.71719885e-01 3.27698261e-01
-1.09069145e+00 1.04421306e+00 -4.61090386e-01 -1.67048082e-01
6.62572324e-01 1.44235894e-01 -2.59584010e-01 4.87414375e-02
-1.35340893e+00 7.84600019e-01 1.11988723e-01 9.85535700e-03
-8.43414187e-01 -6.87301755e-01 -7.66541243e-01 1.24685220e-01
6.89100176e-02 -8.44382226e-01 8.51466477e-01 -1.70889330e+00
-1.72628570e+00 6.30289257e-01 2.59555504e-02 -3.92166346e-01
6.00876629e-01 -3.61376315e-01 -4.16467994e-01 5.94120957e-02
1.24144748e-01 1.16101551e+00 1.43936729e+00 -1.25598633e+00
-2.33182579e-01 -1.16367474e-01 -8.82180333e-02 2.32621133e-01
-2.76918471e-01 -1.45313218e-01 -5.84505737e-01 -1.40185940e+00
-2.79283792e-01 -5.97217202e-01 -2.35993773e-01 2.49089077e-01
-1.58406079e-01 1.48392648e-01 1.17056048e+00 -8.37091923e-01
9.87478435e-01 -1.94274724e+00 2.10516736e-01 -2.85625979e-02
-2.23883614e-02 2.95772105e-01 -3.75590086e-01 3.70652050e-01
-2.25743175e-01 1.02123290e-01 -4.84158009e-01 -5.87477565e-01
-2.52453923e-01 4.53276902e-01 -5.45447707e-01 1.29896730e-01
6.53204024e-01 1.20879948e+00 -6.41195953e-01 -5.30888140e-01
4.21232939e-01 7.36025691e-01 -5.56340933e-01 7.17635870e-01
-3.43201220e-01 5.74178636e-01 -1.98905855e-01 6.33449256e-01
1.05388272e+00 1.72566757e-01 -5.93079507e-01 -3.72804314e-01
1.34632410e-02 -1.01695694e-01 -8.82373810e-01 2.16772151e+00
-7.94944942e-01 4.63770479e-01 4.51957546e-02 -8.53028715e-01
1.14809024e+00 8.53309780e-02 3.21216792e-01 -9.51249719e-01
5.88588417e-02 -1.84619933e-01 -3.50335747e-01 -6.34342670e-01
4.82237339e-01 -2.34615013e-01 1.23981461e-01 4.13121104e-01
5.45574278e-02 -3.01563978e-01 -2.32632861e-01 -4.70276177e-02
7.76322067e-01 5.43128788e-01 -1.60480037e-01 -1.73351238e-03
4.16598827e-01 -4.47910577e-01 4.23348337e-01 6.08259559e-01
7.49304965e-02 1.23105180e+00 1.70484066e-01 -5.79537153e-01
-1.43777001e+00 -8.67977798e-01 1.53338224e-01 1.06025875e+00
1.16504684e-01 -1.09123299e-02 -8.76426995e-01 -6.29936695e-01
-4.54926193e-02 5.24373412e-01 -1.17639947e+00 -4.25686926e-01
-6.41229928e-01 -5.24760544e-01 4.67907846e-01 7.04892933e-01
9.67115343e-01 -1.62262905e+00 -4.10874218e-01 2.56574541e-01
-2.08451942e-01 -6.70577288e-01 -9.70378041e-01 7.43189007e-02
-8.70088160e-01 -6.87970638e-01 -1.26545584e+00 -8.59704852e-01
7.05547631e-01 1.16902441e-01 1.08414865e+00 8.38900208e-02
-4.01953220e-01 -2.34178584e-02 -3.08290035e-01 -3.18655282e-01
-4.35322851e-01 2.00807620e-02 -3.95209342e-01 2.81063408e-01
-1.75866768e-01 -9.75464940e-01 -9.77165520e-01 2.68479943e-01
-1.18350470e+00 1.21919066e-01 8.38203728e-01 1.20803428e+00
4.25245851e-01 -1.87900051e-01 6.11127019e-01 -7.95593977e-01
7.33402789e-01 -3.12760532e-01 -2.48284325e-01 1.66020289e-01
-5.37291884e-01 2.95872360e-01 9.08360183e-01 -6.59704566e-01
-1.20683825e+00 -8.76312554e-02 -4.53646928e-01 -9.49112594e-01
-6.34819940e-02 -1.48218378e-01 -2.35572815e-01 -4.05282885e-01
6.34831607e-01 4.66501534e-01 1.30812988e-01 -3.08061033e-01
4.65128690e-01 3.23087484e-01 6.61678135e-01 -7.32274055e-01
8.76432538e-01 4.33496624e-01 -1.46632686e-01 -2.85841316e-01
-5.47348261e-01 2.14833692e-01 -3.90333593e-01 -1.62689850e-01
7.59928226e-01 -7.81276286e-01 -4.73772436e-01 5.42509675e-01
-1.16407287e+00 -4.28534627e-01 -4.96630639e-01 -5.81206419e-02
-7.51311779e-01 3.63902241e-01 -6.70381844e-01 -6.28557920e-01
-6.81611121e-01 -1.16138315e+00 1.22957325e+00 3.67283612e-01
1.16248131e-01 -7.89409280e-01 1.33152485e-01 1.61134079e-01
8.99202526e-01 4.73971575e-01 7.44934320e-01 1.82191253e-01
-5.41423559e-01 8.21941420e-02 -2.95827717e-01 6.97354376e-01
2.09879354e-01 9.35514271e-03 -1.09884441e+00 -2.29121804e-01
-5.58721684e-02 -3.74128908e-01 1.14147055e+00 5.09113252e-01
1.56499445e+00 -4.02458817e-01 -5.74239977e-02 1.04288185e+00
1.58674192e+00 3.00627530e-01 1.28920746e+00 3.03366095e-01
7.17318654e-01 3.27082127e-01 4.51016605e-01 1.54025972e-01
-4.09365594e-02 5.98449707e-01 5.72629035e-01 -8.19602370e-01
-3.70792508e-01 -5.36596775e-01 1.73233330e-01 3.46190810e-01
-2.48964295e-01 -8.82507935e-02 -1.13752708e-01 4.24043536e-01
-1.87304366e+00 -1.15262663e+00 4.54862505e-01 1.96452034e+00
1.08048058e+00 -5.09565435e-02 9.20348391e-02 -1.17691465e-01
7.26556242e-01 3.12625974e-01 -7.28788137e-01 -6.06287777e-01
-2.70032912e-01 6.18639588e-01 4.16670531e-01 4.54336196e-01
-9.44901586e-01 1.09915447e+00 6.12717676e+00 1.09287488e+00
-1.49648368e+00 2.59382129e-01 1.05297005e+00 -3.65704894e-02
-3.89929444e-01 -2.57034838e-01 -7.34716654e-02 7.18230844e-01
3.83112222e-01 3.05568755e-01 6.84494495e-01 8.19578409e-01
1.75735846e-01 3.37755345e-02 -8.50472808e-01 1.07479405e+00
2.08055526e-01 -1.47898519e+00 2.98267961e-01 -2.34930649e-01
1.01137960e+00 -3.17721635e-01 4.02431369e-01 3.18669766e-01
6.45258427e-02 -1.23969567e+00 7.14591086e-01 7.71546483e-01
1.10818338e+00 -1.11237550e+00 6.39179528e-01 2.96304319e-02
-9.33376670e-01 -1.35827273e-01 -5.55432320e-01 2.92950291e-02
-8.16187263e-02 3.74534786e-01 -3.27832729e-01 4.73883420e-01
8.64305556e-01 5.79246521e-01 -4.75570828e-01 8.11789155e-01
-2.21954674e-01 3.47138584e-01 3.61295119e-02 2.04429299e-01
2.71798491e-01 -2.79217005e-01 2.37389579e-01 1.13411427e+00
3.94415736e-01 -9.37391073e-03 -2.38723382e-01 1.46241319e+00
-2.02162921e-01 1.01651452e-01 -5.46056032e-01 4.43744183e-01
2.64107943e-01 1.30189180e+00 -4.06983942e-01 -4.21218634e-01
-1.23401023e-01 1.73479056e+00 4.32174176e-01 4.19880211e-01
-1.17246723e+00 -7.22770035e-01 6.30239010e-01 1.67671233e-01
4.87564623e-01 4.10487950e-02 -3.37041646e-01 -1.03481126e+00
1.99579269e-01 -7.60311723e-01 -1.44267321e-01 -1.08725190e+00
-1.33437419e+00 8.46405208e-01 -3.32317978e-01 -1.20457935e+00
6.55123917e-03 -2.23573759e-01 -1.14472783e+00 1.07169318e+00
-1.56932747e+00 -1.49453568e+00 -5.18788397e-01 8.21447611e-01
8.14860582e-01 -2.13342667e-01 7.16098249e-01 3.64947528e-01
-4.16625738e-01 8.48755658e-01 4.70837615e-02 -5.17168790e-02
1.20845246e+00 -1.09094036e+00 5.59393346e-01 7.77455509e-01
2.32678745e-02 5.50309420e-01 5.54150820e-01 -6.18115962e-01
-1.17672896e+00 -1.32549274e+00 1.72655538e-01 -8.94439220e-02
6.64084032e-02 -4.01208639e-01 -9.37543333e-01 4.09736663e-01
6.70772791e-01 1.09961368e-01 5.60031161e-02 -4.15243447e-01
-4.50746477e-01 -4.48370606e-01 -1.53869188e+00 6.41170204e-01
9.99092400e-01 -4.02057946e-01 -3.95003468e-01 2.12541789e-01
7.78424203e-01 -4.19856966e-01 -6.64820671e-01 1.76945001e-01
5.16598701e-01 -1.17059708e+00 1.02654779e+00 -3.05659801e-01
7.77141333e-01 -2.34535143e-01 1.74652010e-01 -1.51338542e+00
-6.91097379e-01 -8.22729230e-01 5.25991023e-02 1.33876002e+00
-3.45017835e-02 -5.01589298e-01 9.25540864e-01 3.91096920e-01
-2.33571500e-01 -9.09987271e-01 -7.17325509e-01 -4.53284979e-01
1.60068706e-01 -1.00451753e-01 8.42538357e-01 8.39641631e-01
-5.92596114e-01 5.23810498e-02 -8.64251077e-01 -1.82351340e-02
5.47023058e-01 1.66157529e-01 8.37574422e-01 -7.02357113e-01
-4.18658376e-01 -4.35392946e-01 -1.64459974e-01 -1.05053318e+00
1.11669064e-01 -4.38457370e-01 9.27648600e-03 -1.29578710e+00
1.82932869e-01 -4.19375986e-01 -1.94200322e-01 4.49411094e-01
-3.39201927e-01 6.82531953e-01 1.22867733e-01 1.37987942e-01
-1.93115711e-01 9.25092876e-01 1.59424007e+00 -3.32120359e-01
-4.70033586e-02 -1.52101472e-01 -7.69485831e-01 3.86460662e-01
7.33884633e-01 -2.74331331e-01 -3.49243879e-01 -5.95971406e-01
-1.35066628e-01 -3.35776433e-02 5.02174258e-01 -1.11686707e+00
1.62990112e-02 -2.14975059e-01 1.00712407e+00 -4.06279385e-01
4.43787217e-01 -9.47872996e-01 1.92820564e-01 3.12496662e-01
-3.46524477e-01 1.25187039e-01 3.60820681e-01 4.21133041e-01
-2.55297869e-01 -1.76571775e-02 1.07767045e+00 -4.08370525e-01
-6.31198287e-01 4.64603990e-01 5.05515784e-02 -1.84922561e-01
1.04894423e+00 -2.98012733e-01 2.52315179e-02 -6.48988307e-01
-4.98315632e-01 4.22361540e-03 5.53785264e-01 6.15971088e-01
8.50222588e-01 -1.69763184e+00 -7.30659723e-01 6.04862154e-01
-5.64600416e-02 4.38919403e-02 6.02464497e-01 2.91424036e-01
-5.93036592e-01 -9.70313549e-02 -8.47986698e-01 -4.17352378e-01
-9.65851903e-01 7.37655461e-01 3.84429961e-01 -9.77299586e-02
-5.72129667e-01 9.98352051e-01 3.30939114e-01 -2.12859064e-01
2.49264672e-01 -4.33960855e-01 1.60915792e-01 -3.71201932e-01
4.88437861e-01 1.73216294e-02 -1.20439336e-01 -3.32208633e-01
-2.67260168e-02 7.64692426e-01 -1.98399544e-01 1.13002062e-02
1.40624166e+00 -2.90989340e-03 -7.95074105e-02 4.25524712e-02
1.11217511e+00 2.86755431e-03 -1.77878916e+00 -1.12983644e-01
-8.14653039e-01 -5.76475918e-01 -3.03421393e-02 -9.01861429e-01
-1.56521928e+00 8.23698342e-01 8.32116961e-01 -5.33228628e-02
1.46265745e+00 -2.21745789e-01 1.17102909e+00 -1.50915295e-01
4.28699479e-02 -1.15523851e+00 5.65057695e-01 6.06779158e-02
1.47073019e+00 -1.21442509e+00 -2.05411404e-01 -4.68727835e-02
-5.75519204e-01 1.14340341e+00 9.58854556e-01 -7.24659085e-01
2.53136843e-01 4.25522983e-01 1.20627254e-01 1.14636719e-01
-5.81622124e-01 1.56141579e-01 2.21573293e-01 7.19322026e-01
3.86381596e-01 -2.39045382e-01 -1.12163864e-01 3.91641974e-01
-6.73050508e-02 3.55928242e-02 6.57496601e-02 6.06616795e-01
-1.32980973e-01 -1.13013196e+00 -7.02256083e-01 1.81080818e-01
-3.53960723e-01 -2.62702674e-01 -1.94636613e-01 5.45644760e-01
5.67493737e-01 5.48363805e-01 1.84129193e-01 -5.29897332e-01
3.23744476e-01 -1.52068526e-01 5.92414975e-01 -2.13435143e-01
-8.19943309e-01 1.31707728e-01 -4.52353448e-01 -5.76915920e-01
-1.84099078e-01 -6.99990243e-02 -7.93900311e-01 -1.79969341e-01
-3.03746790e-01 -4.94926460e-02 5.91628551e-01 5.51331937e-01
5.85917890e-01 9.13802326e-01 9.23366964e-01 -1.28087568e+00
-3.09127241e-01 -1.28158748e+00 -3.51016015e-01 6.22408271e-01
3.64306360e-01 -5.20598650e-01 -7.39518851e-02 -1.01317335e-02] | [11.44791316986084, -1.021410584449768] |
f0331bbf-8f3f-403f-bd89-ef7ad0e175cf | a-self-supervised-joint-training-framework | null | null | https://aclanthology.org/2022.findings-naacl.79 | https://aclanthology.org/2022.findings-naacl.79.pdf | A Self-supervised Joint Training Framework for Document Reranking | Pretrained language models such as BERT have been successfully applied to a wide range of natural language processing tasks and also achieved impressive performance in document reranking tasks. Recent works indicate that further pretraining the language models on the task-specific datasets before fine-tuning helps improve reranking performance. However, the pre-training tasks like masked language model and next sentence prediction were based on the context of documents instead of encouraging the model to understand the content of queries in document reranking task. In this paper, we propose a new self-supervised joint training framework (SJTF) with a self-supervised method called Masked Query Prediction (MQP) to establish semantic relations between given queries and positive documents. The framework randomly masks a token of query and encodes the masked query paired with positive documents, and uses a linear layer as a decoder to predict the masked token. In addition, the MQP is used to jointly optimize the models with supervised ranking objective during fine-tuning stage without an extra further pre-training stage. Extensive experiments on the MS MARCO passage ranking and TREC Robust datasets show that models trained with our framework obtain significant improvements compared to original models. | ['Hai Liu', 'Fu Lee Wang', 'Sijie Cheng', 'Tianyong Hao', 'Xiaozhi Zhu'] | null | null | null | null | findings-naacl-2022-7 | ['passage-ranking'] | ['natural-language-processing'] | [ 5.61070383e-01 -3.58049273e-02 -4.92508411e-01 -8.28934133e-01
-1.24063432e+00 -4.49788570e-01 8.40785921e-01 4.33250874e-01
-8.66773963e-01 5.05463302e-01 7.03747571e-01 -2.61163086e-01
-1.07991926e-01 -4.67814118e-01 -7.05327392e-01 -1.13999575e-01
9.97959748e-02 8.00637722e-01 6.42326057e-01 -5.68492949e-01
5.72040796e-01 -1.24262728e-01 -1.45578980e+00 1.26348162e+00
9.65202868e-01 6.88338578e-01 5.45147359e-01 1.00393617e+00
-4.80403185e-01 1.01576757e+00 -6.35989308e-01 -2.79849619e-01
-7.69137442e-02 -2.43789032e-01 -1.20991051e+00 -2.41337121e-01
2.77655005e-01 -4.13332641e-01 -2.60152042e-01 7.24146605e-01
5.17989278e-01 1.33174092e-01 6.47538900e-01 -6.44085646e-01
-8.27626646e-01 1.03257704e+00 -4.29962605e-01 2.81862020e-01
3.69226754e-01 -5.76432586e-01 1.62682199e+00 -1.00976253e+00
2.99563974e-01 1.43184781e+00 6.94687590e-02 5.13828278e-01
-1.00518382e+00 -4.69425112e-01 2.70147651e-01 2.39821017e-01
-1.32239556e+00 -3.67001593e-01 5.80527484e-01 -1.88528121e-01
1.20468783e+00 4.43290204e-01 -4.03851531e-02 6.04716003e-01
5.60235679e-02 1.28718162e+00 6.39696181e-01 -6.13933504e-01
-1.22554079e-01 3.16923171e-01 4.83194232e-01 5.89065969e-01
-2.91422427e-01 -2.49399543e-01 -6.07240200e-01 -2.91204154e-01
5.74902669e-02 1.61497444e-02 -1.02502905e-01 -3.77500914e-02
-1.25293136e+00 7.24142373e-01 4.26676065e-01 3.95145208e-01
-1.35383919e-01 1.18898988e-01 4.15420294e-01 5.22168159e-01
6.66748822e-01 6.57687664e-01 -7.43434191e-01 3.57891768e-01
-1.24619436e+00 7.61085898e-02 5.28250694e-01 6.73479140e-01
7.82329619e-01 -6.71906352e-01 -8.70474517e-01 1.19705248e+00
6.34442210e-01 4.27043647e-01 8.90523255e-01 -5.72716534e-01
8.80812287e-01 6.61357820e-01 8.66590664e-02 -8.07254553e-01
-1.36533350e-01 -7.19495952e-01 -5.99711657e-01 -6.05443835e-01
-7.64037147e-02 3.36284310e-01 -9.76371527e-01 1.52298903e+00
-1.24309666e-01 -1.90728560e-01 4.20161515e-01 7.25656390e-01
7.43178785e-01 9.04311359e-01 2.95514259e-02 1.25729432e-03
1.27704966e+00 -1.38035703e+00 -5.54730177e-01 -4.48470265e-01
1.14773643e+00 -1.14925861e+00 1.58374131e+00 7.74538964e-02
-8.50706697e-01 -7.13778853e-01 -9.06062245e-01 -5.57903171e-01
-1.90078378e-01 4.51718360e-01 3.01285535e-01 2.27715328e-01
-1.16078699e+00 2.32097819e-01 -5.02765000e-01 -3.57339650e-01
-3.48631516e-02 4.73537505e-01 1.01928495e-01 -6.03879541e-02
-1.59229290e+00 6.83985293e-01 6.60670698e-01 -3.46621983e-02
-8.62629414e-01 -3.88569176e-01 -6.04955018e-01 3.53050500e-01
2.41520226e-01 -6.39157772e-01 1.54793668e+00 -7.51753449e-01
-1.49696696e+00 9.89763379e-01 -7.28142083e-01 -4.76652890e-01
2.70639628e-01 -4.83831823e-01 -4.53163654e-01 -4.62792106e-02
4.00192887e-01 9.41721678e-01 7.62609363e-01 -1.02363324e+00
-8.34789217e-01 -1.23563491e-01 1.60964221e-01 5.71037829e-01
-5.68391442e-01 1.87981337e-01 -7.70936847e-01 -5.78569949e-01
1.51042223e-01 -5.31376123e-01 -1.65977672e-01 -7.27205575e-01
-5.74820459e-01 -7.62149572e-01 3.57632279e-01 -5.27326047e-01
1.63999629e+00 -2.08773589e+00 -2.41541296e-01 2.57862478e-01
7.61905080e-03 1.77691713e-01 -7.09339082e-01 7.05721080e-01
1.99831843e-01 1.56073168e-01 3.70803922e-02 -4.21643853e-01
6.58702999e-02 8.47342536e-02 -9.10333693e-01 -2.34130979e-01
8.54249522e-02 1.12269497e+00 -9.87537980e-01 -7.36597717e-01
-2.93262899e-01 9.32526663e-02 -7.54085600e-01 3.24051797e-01
-5.19720197e-01 3.01533550e-01 -6.44625127e-01 2.16854736e-01
3.39873374e-01 -5.32782555e-01 1.17481939e-01 -1.05618969e-01
2.93175012e-01 9.02543545e-01 -6.91275775e-01 1.63393593e+00
-4.80387479e-01 2.93659210e-01 -4.35913235e-01 -7.57643163e-01
8.31610441e-01 2.69416451e-01 2.59720832e-01 -1.10315239e+00
-3.04058105e-01 3.01718384e-01 -3.19417655e-01 -5.37402391e-01
1.12251794e+00 1.75790206e-01 -2.57677436e-01 7.01808691e-01
-1.70069247e-01 1.43775875e-02 4.57917541e-01 7.56460011e-01
9.12930310e-01 -1.56148106e-01 -1.00083902e-01 4.82386770e-03
1.06597698e+00 9.64454561e-02 2.22408846e-01 1.07357585e+00
3.55015248e-01 4.61252987e-01 2.54514366e-01 3.02395970e-02
-6.62983537e-01 -6.94278717e-01 1.97187290e-01 2.12736154e+00
1.45085394e-01 -7.10690141e-01 -4.12933588e-01 -9.36935365e-01
-9.03039647e-04 8.05724382e-01 -3.14797997e-01 -3.57597351e-01
-6.34994388e-01 -6.07579648e-01 6.29700005e-01 4.42422569e-01
3.78641784e-01 -8.95156443e-01 2.62718588e-01 2.11294904e-01
-4.79812175e-01 -1.07898343e+00 -8.67810249e-01 1.65980577e-01
-7.82069206e-01 -7.78064668e-01 -6.89037085e-01 -1.24059999e+00
8.92151475e-01 4.57260787e-01 1.15436268e+00 2.67738819e-01
3.70388061e-01 3.06739271e-01 -5.02619743e-01 -2.18925044e-01
-4.54821050e-01 5.46539724e-01 -1.99585229e-01 1.09389819e-01
4.97556329e-01 1.30175769e-01 -5.29840827e-01 3.95948023e-01
-1.23329628e+00 1.85293168e-01 8.33724439e-01 9.20282006e-01
5.69492161e-01 -1.11488126e-01 6.26182437e-01 -1.16620481e+00
1.17083633e+00 -1.99613035e-01 -3.59639823e-01 9.93994355e-01
-8.21338058e-01 6.82542741e-01 3.89989883e-01 -2.48875201e-01
-1.15308082e+00 -1.50653273e-01 -1.77559644e-01 3.31980169e-01
3.26128036e-01 8.30945313e-01 1.74835593e-01 4.20336157e-01
5.23415089e-01 3.62987399e-01 -4.86148000e-01 -8.15850854e-01
4.67135608e-01 1.00914800e+00 3.67804468e-01 -4.34827536e-01
7.44946122e-01 1.93450630e-01 -5.31454921e-01 -3.14547181e-01
-1.50097942e+00 -9.80980396e-01 -5.75427294e-01 1.63448751e-01
5.90395927e-01 -1.10877216e+00 -4.03611124e-01 9.95114073e-02
-1.25357807e+00 -1.88739702e-01 -2.04236358e-02 6.67387486e-01
-1.34720951e-01 4.83197033e-01 -6.59270346e-01 -5.02771795e-01
-4.50613230e-01 -1.07006884e+00 1.31830680e+00 -2.56641675e-02
-1.51017919e-01 -1.04274786e+00 2.99409449e-01 8.29597533e-01
4.74460304e-01 -1.01367128e+00 1.31681275e+00 -1.04097760e+00
-4.27894861e-01 -3.62114757e-01 -2.87719339e-01 4.75911945e-01
1.15355276e-01 -3.86236042e-01 -9.44929838e-01 -3.62872481e-01
-2.70392954e-01 -6.78089142e-01 1.42643428e+00 -1.10524185e-01
1.36528838e+00 -3.21837395e-01 -3.03874373e-01 1.55104190e-01
1.01874065e+00 -1.42529562e-01 2.99804181e-01 3.32352668e-01
5.56455433e-01 6.04129970e-01 7.06178069e-01 1.07537515e-01
7.36745119e-01 5.98834276e-01 5.10220928e-03 -9.63355228e-03
-3.35141681e-02 -6.24881923e-01 5.82761467e-01 1.18666279e+00
4.23619717e-01 -4.83175188e-01 -7.90476024e-01 3.17600101e-01
-1.86267233e+00 -7.28369296e-01 1.50765538e-01 2.24008751e+00
1.27679396e+00 2.01715678e-01 -4.29071128e-01 -7.80553073e-02
6.75557375e-01 3.07389915e-01 -3.60742152e-01 -1.57851040e-01
-1.93997890e-01 2.24628057e-02 4.11197633e-01 1.03883541e+00
-1.07708633e+00 1.32114005e+00 6.18678713e+00 9.93909001e-01
-9.65621948e-01 -1.34004857e-02 5.40573180e-01 -6.55419007e-02
-6.28263831e-01 1.09319478e-01 -1.19865918e+00 1.95617065e-01
7.71447062e-01 -2.94869959e-01 2.97453165e-01 5.05231261e-01
1.58958986e-01 2.23394185e-02 -1.25540566e+00 7.47392952e-01
3.80094439e-01 -1.11234856e+00 5.68794191e-01 -3.75236124e-01
7.12806880e-01 2.03041032e-01 7.85311982e-02 8.44469428e-01
2.26542771e-01 -7.65521765e-01 5.11135161e-01 5.84533155e-01
4.68074918e-01 -4.36074674e-01 9.06089365e-01 6.09421015e-01
-9.27703083e-01 -6.25101924e-02 -6.19834185e-01 1.14529714e-01
-1.71420649e-02 5.54606795e-01 -1.37325037e+00 3.68693769e-01
3.53313178e-01 6.45621359e-01 -1.04015911e+00 8.30936372e-01
-4.22378987e-01 8.07553828e-01 3.20102572e-02 -2.09656388e-01
2.86827475e-01 4.46408927e-01 2.69703716e-01 1.30353892e+00
-3.14505920e-02 -2.72280931e-01 3.81687462e-01 2.76953816e-01
-3.95822555e-01 5.12788594e-01 1.22040443e-01 -3.34617466e-01
2.21291766e-01 1.02656281e+00 -5.03787696e-01 -6.68819129e-01
-3.20222884e-01 1.19712973e+00 5.46804667e-01 6.04365826e-01
-4.55776542e-01 -5.59027731e-01 7.06396028e-02 -1.67866948e-03
1.06396042e-01 -7.41092414e-02 -2.95593999e-02 -1.32539248e+00
1.74209848e-01 -6.77206635e-01 5.78556538e-01 -6.72246695e-01
-1.24189758e+00 6.34527564e-01 -2.17680037e-01 -1.09152400e+00
-4.44886148e-01 -3.83620918e-01 -3.13397884e-01 7.77098298e-01
-2.10539293e+00 -9.11064029e-01 1.34909943e-01 5.70821345e-01
5.95750332e-01 -3.28595042e-01 7.07560539e-01 3.69982183e-01
-2.64376223e-01 7.64204144e-01 4.54446256e-01 2.73327380e-01
1.21634936e+00 -1.25207186e+00 1.72245726e-01 6.31416798e-01
4.62223828e-01 8.78156841e-01 2.77503580e-01 -7.05938220e-01
-9.58556890e-01 -1.19294846e+00 1.90516698e+00 -4.25824285e-01
5.07736921e-01 -4.54291016e-01 -8.34678769e-01 5.16994834e-01
3.92364383e-01 -3.68984163e-01 9.37764585e-01 2.72551656e-01
-3.81404907e-01 -3.37554961e-01 -5.13453186e-01 3.43814194e-01
6.13326013e-01 -1.05281329e+00 -9.41784859e-01 7.89844573e-01
1.17990792e+00 -1.22170806e-01 -3.92542511e-01 4.16238487e-01
3.59777361e-01 -4.08545852e-01 1.06787634e+00 -1.00017035e+00
3.02317739e-01 -4.03970063e-01 -2.96727926e-01 -1.01261771e+00
-3.81628633e-01 -3.40757340e-01 -1.36068955e-01 1.26239979e+00
8.05079579e-01 -4.34542716e-01 7.24441111e-01 3.98987263e-01
-6.95489123e-02 -6.98235214e-01 -4.26604956e-01 -4.94731009e-01
-1.96450546e-01 -4.02679324e-01 5.88657677e-01 5.07576942e-01
5.80750480e-02 8.61091793e-01 -2.39737138e-01 1.44372299e-01
2.71987081e-01 1.77229762e-01 5.56702077e-01 -1.07242393e+00
-3.32625836e-01 -2.95042008e-01 1.99370921e-01 -1.84309590e+00
3.66314054e-01 -1.53349400e+00 1.95262611e-01 -1.78291774e+00
5.37104607e-01 -4.54398215e-01 -8.47945511e-01 4.89245713e-01
-6.34256124e-01 1.22399785e-01 1.56040769e-03 6.22565031e-01
-1.35285664e+00 6.28430128e-01 1.09221125e+00 -5.02729297e-01
-3.33382756e-01 2.83387065e-01 -7.73430407e-01 4.26247776e-01
6.15909696e-01 -7.37423122e-01 -7.73831666e-01 -1.15698683e+00
7.80683339e-01 4.17670384e-02 9.13780257e-02 -6.19020224e-01
6.40541255e-01 3.48194130e-03 3.57000381e-01 -9.49520051e-01
5.33098206e-02 -5.22807956e-01 -9.05402601e-01 3.54083806e-01
-1.48093605e+00 8.41765702e-02 -2.76197046e-01 5.78685701e-01
-6.52673960e-01 -4.57719386e-01 2.45224237e-01 5.30608697e-03
-8.14178705e-01 6.25391826e-02 -3.25347543e-01 2.89702594e-01
3.16939443e-01 2.74248779e-01 -2.47135341e-01 -5.42227805e-01
-5.82536340e-01 8.07840526e-01 -1.25974417e-02 7.46277928e-01
6.70889020e-01 -1.16907334e+00 -9.08617973e-01 2.81208098e-01
4.27582681e-01 -1.48136556e-01 6.91430047e-02 4.84582812e-01
-1.88209847e-01 1.23445415e+00 5.18048763e-01 -5.63900292e-01
-1.31588900e+00 3.41506690e-01 -4.13570739e-02 -1.00006473e+00
8.21792055e-03 1.07195389e+00 2.22528279e-01 -8.91976893e-01
5.02941191e-01 -5.07489145e-01 -5.97514987e-01 2.45847907e-02
7.06464410e-01 -1.44031979e-02 2.12191075e-01 -4.58479792e-01
-3.71282488e-01 3.94166142e-01 -7.34498739e-01 -3.36607486e-01
1.05184853e+00 -4.81962323e-01 -4.79567379e-01 4.81853515e-01
1.52312064e+00 2.32777968e-01 -6.05557740e-01 -9.98177886e-01
7.11099386e-01 -1.64363533e-01 7.53010884e-02 -1.03519762e+00
-5.61317682e-01 8.35283041e-01 3.27563912e-01 -1.05003193e-01
1.10508442e+00 1.48595020e-01 8.82517040e-01 1.07615101e+00
9.81507525e-02 -1.20390105e+00 5.03316283e-01 1.08357978e+00
8.47910106e-01 -1.20576191e+00 -1.96004614e-01 -5.60613237e-02
-8.25565636e-01 8.97300899e-01 5.89688301e-01 2.47066081e-01
4.78990197e-01 -3.08047861e-01 2.29180485e-01 4.48887460e-02
-1.04368150e+00 -3.13853681e-01 9.56204474e-01 -1.25529647e-01
8.87619138e-01 -2.04259709e-01 -3.64059746e-01 5.63563228e-01
-3.26164454e-01 7.03545436e-02 -4.86893766e-02 9.33057070e-01
-8.18038821e-01 -1.46213078e+00 -7.91514739e-02 5.51883101e-01
-5.73987424e-01 -5.51522613e-01 -4.54807013e-01 -7.49999359e-02
-2.01516002e-01 1.14372897e+00 3.44313830e-02 -6.26235962e-01
8.98604542e-02 3.11354667e-01 6.83566779e-02 -1.02099252e+00
-6.68969095e-01 9.51732546e-02 1.01986164e-02 -3.57273400e-01
-4.54285175e-01 -1.63834184e-01 -1.38611794e+00 3.59185934e-01
-3.73472959e-01 8.84842217e-01 4.74749029e-01 1.10432374e+00
5.53427339e-01 4.67100888e-01 8.47417176e-01 -2.29989156e-01
-7.88608849e-01 -1.20874333e+00 -3.26308697e-01 3.52278769e-01
5.39522588e-01 4.83041890e-02 -1.43438980e-01 -4.23398688e-02] | [11.455636024475098, 7.677432537078857] |
3ccf1969-3169-40ed-89d2-8c90c479a10a | recurrent-instance-segmentation-using | 1911.02103 | null | https://arxiv.org/abs/1911.02103v1 | https://arxiv.org/pdf/1911.02103v1.pdf | Recurrent Instance Segmentation using Sequences of Referring Expressions | The goal of this work is to segment the objects in an image that are referred to by a sequence of linguistic descriptions (referring expressions). We propose a deep neural network with recurrent layers that output a sequence of binary masks, one for each referring expression provided by the user. The recurrent layers in the architecture allow the model to condition each predicted mask on the previous ones, from a spatial perspective within the same image. Our multimodal approach uses off-the-shelf architectures to encode both the image and the referring expressions. The visual branch provides a tensor of pixel embeddings that are concatenated with the phrase embeddings produced by a language encoder. Our experiments on the RefCOCO dataset for still images indicate how the proposed architecture successfully exploits the sequences of referring expressions to solve a pixel-wise task of instance segmentation. | ['Xavier Giro-i-Nieto', 'Ionut-Teodor Sorodoc', 'Alba Herrera-Palacio', 'Carina Silberer', 'Gemma Boleda', 'Carles Ventura'] | 2019-11-05 | null | null | null | null | ['referring-expression-segmentation'] | ['computer-vision'] | [ 4.89504695e-01 4.98543531e-01 -2.05374137e-01 -7.19930708e-01
-6.37976885e-01 -5.99295318e-01 5.46920657e-01 -3.06525618e-01
-3.35922599e-01 1.72672153e-01 6.88504130e-02 -1.24585584e-01
3.87335062e-01 -6.02675080e-01 -1.11973548e+00 -6.09905720e-01
3.81853431e-01 3.82274956e-01 -1.07361861e-01 -2.16497004e-01
3.21018487e-01 5.89387596e-01 -1.29887354e+00 7.57085323e-01
2.50847846e-01 1.14119744e+00 4.52420712e-01 7.16316581e-01
-2.91107893e-01 9.46441829e-01 -3.17523658e-01 -6.28393114e-01
9.85699594e-02 -2.43781358e-01 -9.17025924e-01 5.86213648e-01
6.12105370e-01 -6.74939677e-02 -2.90340960e-01 9.73100007e-01
-8.02047551e-02 9.14755911e-02 5.05140066e-01 -9.14901257e-01
-1.29229617e+00 7.38259494e-01 -5.00820100e-01 -7.84185454e-02
3.69450212e-01 1.50448486e-01 1.33866167e+00 -9.44491565e-01
8.86633754e-01 1.24118757e+00 1.34237081e-01 6.42695963e-01
-1.39733255e+00 -1.07731573e-01 5.27804434e-01 1.82603061e-01
-1.36913574e+00 -1.89938664e-01 9.11485553e-01 -6.92414105e-01
1.09363806e+00 7.83760473e-02 5.67741215e-01 1.08282959e+00
8.96332636e-02 8.19636345e-01 7.30118155e-01 -1.28815293e-01
8.32803845e-02 3.04381102e-01 1.28902733e-01 9.11710680e-01
-6.62727714e-01 -1.95102453e-01 -5.75340688e-01 2.18250021e-01
8.21298242e-01 3.20307240e-02 -3.03945065e-01 -3.68645906e-01
-1.11095417e+00 8.26650381e-01 8.95571411e-01 3.60014945e-01
-6.63674951e-01 5.55687904e-01 1.21787377e-01 -1.28007576e-01
3.81287038e-01 4.63277817e-01 -4.60351229e-01 2.11421654e-01
-8.96418393e-01 -1.05610937e-01 4.80174571e-01 1.09514749e+00
9.76418257e-01 -3.13340604e-01 -4.20545697e-01 6.49670601e-01
5.82724273e-01 2.25449413e-01 1.56349033e-01 -1.17860675e+00
6.15296423e-01 7.79981256e-01 1.90538958e-01 -1.11033165e+00
-2.28467956e-01 -3.24017197e-01 -4.33737725e-01 -1.94982454e-01
1.91421583e-01 3.59106734e-02 -1.16158128e+00 1.75220299e+00
-2.27053314e-02 -3.15885916e-02 8.28955248e-02 1.26443577e+00
8.88725519e-01 9.68924642e-01 3.69573772e-01 2.84455776e-01
1.53795910e+00 -1.27065253e+00 -7.20546067e-01 -5.85710287e-01
4.25847322e-01 -7.29174852e-01 1.06735718e+00 6.31067008e-02
-1.29974139e+00 -6.88926518e-01 -7.41734922e-01 -5.43609440e-01
-5.84816277e-01 5.72232366e-01 2.26563215e-01 1.26523063e-01
-1.26360691e+00 2.45158941e-01 -5.25178015e-01 -3.05155635e-01
5.06427526e-01 2.91784108e-01 -4.51060236e-01 2.26441041e-01
-8.78119409e-01 8.82912934e-01 3.66434991e-01 6.00561380e-01
-9.01068449e-01 -2.58136302e-01 -1.13505530e+00 2.85445899e-01
8.35589021e-02 -7.01165080e-01 9.05566275e-01 -1.55100715e+00
-1.28500164e+00 1.49851465e+00 -6.24993920e-01 -6.23971283e-01
3.00486177e-01 -2.36168861e-01 -2.71393918e-02 5.75089514e-01
8.48770291e-02 1.62192535e+00 1.04335809e+00 -1.71753573e+00
-5.99905193e-01 -3.04212421e-01 3.38927180e-01 1.59016728e-01
1.41244635e-01 1.19264036e-01 -1.00489283e+00 -4.51720804e-01
1.97419867e-01 -8.03011775e-01 -2.90056676e-01 3.18259175e-04
-8.76174331e-01 -1.99428305e-01 8.47799540e-01 -8.33912373e-01
8.03422332e-01 -2.26854777e+00 8.09275210e-01 5.24096861e-02
-1.45134062e-01 -1.72761798e-01 -3.97178859e-01 2.53002286e-01
-4.42724109e-01 4.15930331e-01 -5.89004517e-01 -7.40981460e-01
2.38640774e-02 6.03521883e-01 -6.27905965e-01 3.60145479e-01
8.17205608e-01 1.19613576e+00 -6.04068995e-01 -4.85554487e-01
1.45175338e-01 7.37151802e-01 -3.84868741e-01 6.61982179e-01
-8.45784843e-01 6.58933103e-01 -3.35843921e-01 4.72523361e-01
4.72376287e-01 -2.57243365e-01 -9.87185687e-02 -4.36354071e-01
-1.02810793e-01 2.94670593e-02 -5.15770555e-01 2.05647111e+00
-5.65390706e-01 7.33077168e-01 -5.17068878e-02 -1.06310499e+00
9.34320509e-01 2.67080218e-01 1.86375618e-01 -6.32184327e-01
2.56118625e-01 4.04112265e-02 -3.20415676e-01 -1.03433287e+00
5.33987224e-01 1.30389836e-02 -3.10313195e-01 4.24903065e-01
1.43144220e-01 -1.77815720e-01 3.33071679e-01 2.11638927e-01
6.14110529e-01 6.73108578e-01 -2.74322689e-01 1.19840205e-01
7.99854755e-01 4.06112373e-02 1.81651637e-01 4.84242022e-01
4.61155623e-02 8.96717906e-01 9.88504529e-01 -6.20828032e-01
-1.20759666e+00 -8.19642365e-01 4.69994508e-02 1.30809891e+00
1.26774609e-01 -3.19972523e-02 -7.25680709e-01 -7.40813375e-01
-2.81207055e-01 9.51270819e-01 -9.99406993e-01 1.77670658e-01
-7.90824294e-01 -1.59945115e-01 2.47548878e-01 6.71873927e-01
3.66597623e-01 -1.38341928e+00 -1.06986880e+00 -1.11687727e-01
-1.83879301e-01 -1.45295620e+00 -5.60198128e-01 4.03781295e-01
-4.71217602e-01 -8.39827240e-01 -7.23192692e-01 -1.10897899e+00
1.00833249e+00 -2.61509210e-01 1.10107338e+00 -6.24771900e-02
-6.18529879e-02 6.26005352e-01 -2.87442893e-01 2.79738426e-01
-2.67132074e-01 1.05086260e-01 -7.72515655e-01 5.99774301e-01
2.49058783e-01 -2.54255772e-01 -5.22210419e-01 3.44981439e-02
-1.05318379e+00 3.51464212e-01 6.04082584e-01 8.11760366e-01
9.99734342e-01 -7.06340969e-01 -8.12420622e-02 -6.88880265e-01
2.58260161e-01 -4.16374922e-01 -5.37122548e-01 4.18457776e-01
3.27883750e-01 3.33186835e-01 4.78579640e-01 -3.25433582e-01
-9.89323020e-01 6.25896037e-01 -1.11604206e-01 -6.10740960e-01
-4.97460842e-01 3.83803666e-01 -2.05714144e-02 1.75248146e-01
1.88183755e-01 1.61400795e-01 -2.34209090e-01 -3.91787499e-01
1.06976032e+00 4.48491633e-01 7.94007182e-01 -5.38121581e-01
2.87178069e-01 5.31067312e-01 -1.28692135e-01 -5.32794833e-01
-1.24023175e+00 -3.22548300e-01 -1.16581774e+00 -1.40233457e-01
1.63671613e+00 -8.34328532e-01 -6.52271032e-01 -1.56587109e-01
-1.79327416e+00 -2.93495923e-01 -2.60436207e-01 -1.70394093e-01
-8.52697432e-01 -2.80205179e-02 -4.70623523e-01 -8.00179064e-01
-1.40993863e-01 -1.56002903e+00 1.47463048e+00 5.38166523e-01
-1.77333072e-01 -8.03269088e-01 -4.14091021e-01 4.65606302e-01
2.80974694e-02 9.15535465e-02 1.16460419e+00 -4.31291193e-01
-9.53265429e-01 -4.21547964e-02 -5.81152678e-01 2.42078811e-01
-1.79279372e-01 2.87597150e-01 -1.09999287e+00 1.82279438e-01
-2.19455406e-01 -2.96433300e-01 1.12266946e+00 3.77057284e-01
1.45937979e+00 -2.41805702e-01 -2.03705877e-01 8.09672177e-01
1.49442613e+00 3.12281191e-01 6.50267124e-01 1.47286236e-01
9.84658957e-01 8.86036575e-01 3.80008638e-01 -2.22119428e-02
3.93119991e-01 7.43588090e-01 9.27912295e-01 -4.23482865e-01
4.04949449e-02 -2.04179987e-01 1.94016248e-01 2.01056629e-01
3.30282271e-01 -1.57422274e-01 -9.56610441e-01 8.18222702e-01
-1.92430508e+00 -7.33655870e-01 -1.19737379e-01 1.51702940e+00
5.48049092e-01 -2.64788747e-01 -3.21765095e-01 -5.49165785e-01
6.02022111e-01 2.98520356e-01 -5.81279218e-01 -9.84250546e-01
-3.07693481e-01 2.30207831e-01 3.60164344e-01 5.58358312e-01
-1.36025286e+00 1.26726651e+00 6.12325764e+00 2.12506741e-01
-1.32860112e+00 4.94080298e-02 9.40007567e-01 1.79320201e-01
-3.58689964e-01 -1.77966803e-02 -6.75766170e-01 9.46536437e-02
7.75772214e-01 4.21331555e-01 4.58121359e-01 6.17094874e-01
2.80034155e-01 -2.94420183e-01 -1.32201529e+00 9.08979535e-01
4.53657359e-01 -1.30365264e+00 3.43626380e-01 -2.79018462e-01
6.82872713e-01 -9.62662976e-03 5.61877131e-01 -4.46166806e-02
-5.75238988e-02 -1.39453483e+00 1.14895916e+00 8.53041351e-01
6.11275315e-01 -4.66151088e-01 6.61921799e-01 6.23267032e-02
-8.99185479e-01 -3.27525973e-01 -8.42167810e-02 2.02300072e-01
4.63303357e-01 -1.54185770e-02 -8.02938879e-01 3.64880532e-01
5.83244145e-01 9.54987168e-01 -6.74236596e-01 5.64027846e-01
-7.42760718e-01 2.77298272e-01 -4.70228819e-03 2.20686391e-01
7.62303889e-01 -4.47608054e-01 2.74145812e-01 1.19535148e+00
1.23015031e-01 -9.97266322e-02 -1.29604831e-01 1.75929368e+00
-1.94501355e-01 8.19367170e-03 -6.51046276e-01 -2.71520913e-01
1.86890941e-02 1.34547865e+00 -4.14035827e-01 -2.71570861e-01
-3.40376914e-01 1.35666263e+00 4.97546852e-01 7.95108199e-01
-9.17373776e-01 -9.50015709e-02 3.60975504e-01 -1.95111766e-01
8.08749080e-01 -1.91503137e-01 -6.06108010e-01 -8.28498125e-01
9.51340944e-02 -4.84500706e-01 2.67547190e-01 -1.41793370e+00
-1.08911121e+00 7.62068927e-01 -1.86623096e-01 -9.20423627e-01
-1.79295853e-01 -9.32259440e-01 -3.90424073e-01 1.33698916e+00
-1.52707589e+00 -1.38984787e+00 -7.93099627e-02 2.15908334e-01
6.86765194e-01 5.34845963e-02 8.19126844e-01 2.69571748e-02
-8.94356668e-01 -5.41377030e-02 -5.73319733e-01 3.28013241e-01
2.42758870e-01 -1.11895573e+00 2.53995150e-01 7.83043087e-01
5.26025593e-01 7.44880795e-01 7.74134755e-01 -3.30781430e-01
-1.08537245e+00 -9.12755251e-01 1.09042895e+00 -3.27198058e-01
6.97466671e-01 -2.88463414e-01 -8.70560944e-01 1.09903443e+00
7.04807043e-01 1.13232464e-01 4.45351064e-01 -3.93400460e-01
-4.78247851e-01 1.69319987e-01 -8.46781313e-01 6.26056790e-01
6.23509467e-01 -9.61194158e-01 -6.53437436e-01 2.35320330e-01
7.63684571e-01 -5.75100660e-01 -6.13041937e-01 3.70150469e-02
3.31662297e-01 -8.91763508e-01 9.20753837e-01 -8.82111728e-01
1.03461647e+00 -4.62971091e-01 -3.38475138e-01 -8.99360061e-01
1.19670019e-01 -1.62967592e-01 1.94495693e-01 1.09047675e+00
7.42724180e-01 -1.07786216e-01 5.96288085e-01 7.31224775e-01
1.08913956e-02 -9.33541179e-01 -9.63628173e-01 9.51167941e-02
-9.82400030e-02 -4.01052862e-01 2.80159086e-01 6.14706755e-01
-2.76259154e-01 4.83641922e-01 -2.26335496e-01 4.35051709e-01
1.20846301e-01 2.81154662e-01 4.27966386e-01 -6.53101504e-01
-9.98986438e-02 -4.49226707e-01 -3.02224725e-01 -1.52303469e+00
6.55286491e-01 -9.30613339e-01 2.14445934e-01 -1.57498693e+00
2.16249928e-01 -7.82371387e-02 -6.70333952e-03 8.05553198e-01
5.36792167e-02 2.35976174e-01 4.78781760e-01 -1.70008019e-02
-6.16097748e-01 4.92040783e-01 1.13204992e+00 -5.88552177e-01
-8.29947963e-02 -4.43716496e-01 -4.73975807e-01 8.07481349e-01
4.64577436e-01 -3.25697362e-01 -2.57702116e-02 -1.07174432e+00
3.54520231e-01 2.36736059e-01 7.33498693e-01 -4.50759977e-01
1.27769992e-01 8.29757899e-02 3.09206754e-01 -8.63630235e-01
7.22853839e-01 -1.11826408e+00 -3.25856842e-02 1.11326464e-01
-8.16394091e-01 1.49010107e-01 7.57692009e-02 3.21619838e-01
-3.75766635e-01 -3.43392432e-01 6.38840497e-01 -2.83153504e-01
-8.23434114e-01 6.95707947e-02 -8.99148434e-02 -3.54314059e-01
1.03565323e+00 7.14254677e-02 -2.10158572e-01 -3.61631632e-01
-1.32593894e+00 3.39648306e-01 4.20698732e-01 3.75524610e-01
8.17806065e-01 -1.11872315e+00 -2.73045570e-01 6.20723926e-02
1.87983751e-01 1.19568609e-01 1.85485080e-01 9.07112539e-01
-5.20032525e-01 5.35745561e-01 -2.83616602e-01 -9.31164086e-01
-1.03893757e+00 6.41298831e-01 8.56338799e-01 -5.63042015e-02
-5.59083760e-01 1.31493151e+00 3.51433545e-01 -2.02150762e-01
3.00924182e-01 -5.34875274e-01 -5.07982671e-01 1.41255438e-01
1.83370173e-01 -4.39801842e-01 -4.13869828e-01 -1.30836320e+00
-7.97954574e-02 9.72248316e-01 1.26660064e-01 -3.56685132e-01
1.29778790e+00 -2.83729374e-01 -4.55218017e-01 8.25690627e-01
1.38302338e+00 -4.20242965e-01 -1.47106373e+00 -1.10130787e-01
1.46915808e-01 -1.33910626e-01 -1.49133682e-01 -7.70580351e-01
-1.31885433e+00 1.37103450e+00 3.89531910e-01 1.49507508e-01
1.15490520e+00 4.50789064e-01 5.20981967e-01 1.38026267e-01
-4.64430973e-02 -8.61688375e-01 1.09641328e-01 5.49428701e-01
1.16534531e+00 -1.02983797e+00 -5.43650508e-01 -4.83200639e-01
-1.07951784e+00 1.35545981e+00 4.20228273e-01 -2.03726307e-01
1.68918699e-01 -9.96993557e-02 2.75440156e-01 -5.25191724e-01
-7.79880285e-01 -4.92588967e-01 4.14692938e-01 4.45197731e-01
6.10881090e-01 -8.67136940e-03 -1.62636325e-01 5.02420008e-01
5.78444004e-02 -3.11336100e-01 3.60988319e-01 3.87313426e-01
-1.06825933e-01 -1.01371360e+00 -4.16992843e-01 -6.17199168e-02
-2.16637328e-01 -9.64119285e-02 -6.78624749e-01 3.16051275e-01
3.70330423e-01 8.09938252e-01 5.72496235e-01 -1.42101645e-01
5.50578713e-01 3.45576197e-01 2.58647352e-01 -5.91084361e-01
-6.48200572e-01 1.62524104e-01 -6.76080287e-02 -7.17043459e-01
-6.98415339e-01 -4.77387369e-01 -1.55370069e+00 5.54633260e-01
3.87319565e-01 -1.36277571e-01 7.42825329e-01 1.04565644e+00
2.31481344e-01 7.62515008e-01 4.77589160e-01 -1.02891123e+00
5.25355004e-02 -7.43151426e-01 -2.45856315e-01 7.36480951e-01
4.50258821e-01 -3.20423305e-01 -6.40648752e-02 5.76516628e-01] | [10.35303783416748, 1.271130919456482] |
2256d6f2-e352-49bd-9fa8-f6089d2869e7 | ai-techniques-for-cone-beam-computed | 2306.03025 | null | https://arxiv.org/abs/2306.03025v2 | https://arxiv.org/pdf/2306.03025v2.pdf | AI Techniques for Cone Beam Computed Tomography in Dentistry: Trends and Practices | Cone-beam computed tomography (CBCT) is a popular imaging modality in dentistry for diagnosing and planning treatment for a variety of oral diseases with the ability to produce detailed, three-dimensional images of the teeth, jawbones, and surrounding structures. CBCT imaging has emerged as an essential diagnostic tool in dentistry. CBCT imaging has seen significant improvements in terms of its diagnostic value, as well as its accuracy and efficiency, with the most recent development of artificial intelligence (AI) techniques. This paper reviews recent AI trends and practices in dental CBCT imaging. AI has been used for lesion detection, malocclusion classification, measurement of buccal bone thickness, and classification and segmentation of teeth, alveolar bones, mandibles, landmarks, contours, and pharyngeal airways using CBCT images. Mainly machine learning algorithms, deep learning algorithms, and super-resolution techniques are used for these tasks. This review focuses on the potential of AI techniques to transform CBCT imaging in dentistry, which would improve both diagnosis and treatment planning. Finally, we discuss the challenges and limitations of artificial intelligence in dentistry and CBCT imaging. | ['Suraiya Jabin', 'Saba Sarwar'] | 2023-06-05 | null | null | null | null | ['super-resolution'] | ['computer-vision'] | [ 1.79960296e-01 5.23673117e-01 -4.68932092e-01 -3.69022310e-01
-6.65994883e-01 1.96251124e-01 2.11982094e-02 4.01286095e-01
-3.68004173e-01 1.59035504e-01 3.10015917e-01 -2.02069864e-01
-9.02508423e-02 -8.46993923e-01 1.27508849e-01 -1.00100243e+00
8.52294415e-02 1.16938174e+00 2.84982294e-01 2.89389919e-02
3.79587561e-01 8.88141513e-01 -1.67507994e+00 4.51414198e-01
5.22597611e-01 9.54005957e-01 3.46597463e-01 5.24260640e-01
-3.87799740e-01 6.16936922e-01 -2.88533926e-01 -5.08014917e-01
-3.12280029e-01 -2.47475892e-01 -7.39017844e-01 2.31503412e-01
3.33140999e-01 -4.96442944e-01 -3.83241661e-02 7.92301536e-01
7.10965276e-01 -4.17218238e-01 8.29267383e-01 -5.48681676e-01
-6.69782102e-01 1.68501258e-01 -5.29793382e-01 5.83792746e-01
2.49945015e-01 1.87480286e-01 3.81742030e-01 -5.97696364e-01
6.03358567e-01 1.30458844e+00 5.58787584e-01 8.88329566e-01
-6.11055553e-01 -3.13626677e-01 -3.83286059e-01 5.78877389e-01
-5.47856212e-01 -1.98404595e-01 5.14335752e-01 -4.10804510e-01
8.09961975e-01 8.48340765e-02 1.13490236e+00 3.08050364e-01
6.36816621e-01 9.19745803e-01 1.04619312e+00 -1.00198019e+00
7.89914966e-01 -2.64582127e-01 1.98471874e-01 6.78927541e-01
5.26830964e-02 3.25899780e-01 6.93201870e-02 -2.69894958e-01
7.94552147e-01 7.51403198e-02 -5.12539595e-03 2.01596171e-01
-3.87062371e-01 1.26924610e+00 7.58409262e-01 5.28916717e-01
-7.44001687e-01 -1.69383362e-01 4.44477111e-01 -2.82840550e-01
4.69349384e-01 1.64875519e-02 -1.22973397e-01 2.30936036e-01
-2.58762896e-01 -2.64317513e-01 3.27485353e-01 1.70520574e-01
-1.83694318e-01 -5.66796307e-03 3.20799172e-01 1.53761482e+00
7.92772472e-01 5.94416142e-01 1.14628005e+00 -1.17861366e+00
-3.14496547e-01 6.39291883e-01 -6.97952926e-01 -3.48081380e-01
-5.99617839e-01 3.83618593e-01 -6.76973462e-01 5.08838892e-01
2.16087133e-01 4.73636776e-01 -1.83457208e+00 6.24496400e-01
9.66655612e-01 -3.96979779e-01 -4.15114254e-01 6.87206149e-01
9.20651734e-01 2.60879040e-01 3.16314697e-01 -3.27378005e-01
1.82002378e+00 -6.54354274e-01 -8.53570402e-01 -5.81178367e-01
4.03225988e-01 -8.27148438e-01 8.08828235e-01 2.90781766e-01
-1.20199299e+00 -7.88648725e-02 -7.11177468e-01 -4.33353871e-01
-2.58666307e-01 -3.75142395e-01 9.13646042e-01 9.84727323e-01
-5.64385355e-01 2.24550381e-01 -1.55817163e+00 -2.57454634e-01
1.02312279e+00 5.85495055e-01 -4.38937634e-01 -5.90923250e-01
-3.34240973e-01 1.15815854e+00 -9.23469514e-02 1.22588322e-01
1.56376660e-01 -6.50800943e-01 -8.51064086e-01 -3.09833437e-01
-2.35912904e-01 -8.68518114e-01 1.73386502e+00 -3.10036957e-01
-2.00403047e+00 1.40457356e+00 -7.06184655e-02 -1.29764780e-01
6.68064356e-02 -1.50421802e-02 -1.11887783e-01 4.89479154e-01
2.15505674e-01 6.23123288e-01 2.42855370e-01 -9.84502316e-01
-8.62403274e-01 -1.64190400e+00 -5.70454121e-01 -7.74456263e-02
2.19760746e-01 -7.94002507e-03 -4.45095032e-01 -1.51972711e-01
1.11000180e+00 -1.12978232e+00 -6.37635887e-01 5.52439511e-01
-2.90348828e-01 -6.14718139e-01 6.47511959e-01 -7.65273988e-01
6.53712571e-01 -1.96889043e+00 -3.60711724e-01 1.45810485e-01
1.22209333e-01 2.75737584e-01 3.81894141e-01 -2.24805139e-02
1.88361496e-01 4.32804460e-03 -3.32629710e-01 -2.14254051e-01
-5.95511675e-01 7.77597189e-01 6.37159228e-01 5.40884495e-01
-5.15968084e-01 8.45131516e-01 -4.65820551e-01 -9.58451688e-01
7.07393050e-01 7.17918515e-01 -4.62544292e-01 -2.41380394e-01
-1.95635915e-01 6.31988585e-01 -4.99768823e-01 1.13315678e+00
5.86061418e-01 -2.23870695e-01 2.27618933e-01 -2.31114477e-01
1.52789861e-01 3.06124121e-01 -6.72234058e-01 1.26218081e+00
-8.33808005e-01 3.98752660e-01 3.35958034e-01 -5.16075313e-01
5.56089580e-01 4.79399532e-01 9.30521190e-01 -1.45713711e+00
2.98612118e-01 5.05837917e-01 2.50001103e-01 -1.09278464e+00
-5.45661807e-01 -7.16243565e-01 6.26646459e-01 6.18104577e-01
-2.44562164e-01 -8.44013453e-01 -1.77172571e-01 -4.61352855e-01
4.49778348e-01 -2.40512103e-01 6.67094350e-01 2.64950126e-01
3.64220023e-01 1.66100189e-02 2.70791709e-01 5.44836000e-02
-3.30287099e-01 7.07281172e-01 -3.51433754e-01 -8.93360376e-01
-1.07789707e+00 -9.29163039e-01 -8.35183859e-01 5.45824945e-01
-1.70386344e-01 6.30448878e-01 -6.25066280e-01 -2.61786729e-01
-8.90596136e-02 6.53653979e-01 -7.71193266e-01 2.36987565e-02
-1.13411617e+00 -1.03741324e+00 -1.45770445e-01 5.70334494e-01
4.10347015e-01 -1.61621618e+00 -7.15799451e-01 4.32729661e-01
-3.02410442e-02 -9.06079233e-01 -2.90694484e-03 8.88005123e-02
-1.65309238e+00 -1.27541709e+00 -8.58524442e-01 -1.42678809e+00
5.84481180e-01 -6.55465946e-03 7.80090868e-01 2.58066833e-01
-9.89346206e-01 3.48765820e-01 1.45865753e-01 -8.53298068e-01
-8.38119984e-01 -3.02526295e-01 -3.88576090e-01 -7.57600725e-01
5.92476726e-01 -7.17470944e-01 -9.89028394e-01 5.28764725e-02
-8.32229733e-01 -3.52087051e-01 6.71832144e-01 7.43755460e-01
1.03791380e+00 -9.10443738e-02 2.08084762e-01 -1.07125974e+00
6.08305812e-01 -3.61006588e-01 -2.59160429e-01 1.72811985e-01
-7.44827390e-01 -2.14112967e-01 -1.50733590e-01 -1.07296415e-01
-1.11135054e+00 8.51590037e-02 -7.35912263e-01 2.59690434e-01
-4.54583913e-01 3.90826523e-01 3.09269607e-01 -1.38546089e-02
6.65171385e-01 3.31766158e-02 5.58727562e-01 -7.14080930e-01
5.08226529e-02 1.01516950e+00 6.07252300e-01 -8.49625915e-02
1.42589835e-02 1.08503962e+00 -8.31159055e-02 -1.07189131e+00
-9.53972340e-01 -6.08635843e-01 -9.06353176e-01 -1.82130635e-01
1.07770455e+00 -3.15577418e-01 -9.55701411e-01 5.08528650e-01
-9.84341085e-01 9.22766477e-02 -4.24234509e-01 7.16953397e-01
-3.26380849e-01 2.24882036e-01 -1.23132670e+00 -7.66905248e-01
-9.42819715e-01 -1.49985266e+00 1.32801294e+00 5.16415417e-01
-2.91053262e-02 -1.05340171e+00 3.79472375e-01 1.10929263e+00
1.71286181e-01 1.80098414e-01 1.53099573e+00 -1.30201548e-01
-6.41961247e-02 -4.15446937e-01 9.36969146e-02 3.26025307e-01
4.21057165e-01 1.84541289e-02 -9.79962707e-01 5.10449827e-01
5.23000717e-01 -1.74786076e-01 5.87444246e-01 1.33436561e+00
9.95953858e-01 -1.68362468e-01 -7.22134113e-01 5.24394095e-01
1.27263916e+00 1.23621690e+00 8.23726296e-01 6.41378760e-01
3.70828658e-01 8.90913427e-01 1.75126299e-01 -4.32757363e-02
5.24568558e-01 5.97731829e-01 6.12801135e-01 -1.18143834e-01
-5.57357132e-01 2.92281955e-01 -4.01044607e-01 9.43003237e-01
-4.18840051e-01 4.33909863e-01 -1.04571342e+00 4.39192146e-01
-9.32727098e-01 -6.82934523e-01 -1.12070188e-01 1.59381044e+00
6.02019966e-01 -1.63442850e-01 3.58321518e-02 6.61148071e-01
5.86951196e-01 -7.52249658e-01 -1.09438694e+00 -4.98872131e-01
3.39970618e-01 9.52872217e-01 1.09060757e-01 5.15661657e-01
-6.23028576e-01 5.56339920e-01 6.77028275e+00 4.30039167e-01
-1.40586507e+00 2.41625458e-01 9.42132711e-01 1.97947353e-01
-6.50865734e-02 -7.57525682e-01 -5.66420496e-01 2.95172721e-01
5.37861347e-01 4.20319468e-01 -2.31844783e-01 1.00636661e+00
7.85271451e-02 -6.49180293e-01 -8.19812477e-01 7.77148843e-01
-5.71425706e-02 -1.40942979e+00 -1.95246160e-01 4.44071054e-01
3.90354067e-01 1.20791748e-01 1.79506958e-01 -1.83270182e-02
1.63304016e-01 -1.09155810e+00 5.88457435e-02 -3.64792913e-01
4.97005701e-01 -3.22811693e-01 8.33786845e-01 -1.31989375e-01
-6.20103002e-01 -1.27015710e-01 -1.26460239e-01 3.82477552e-01
2.78762877e-01 4.75315779e-01 -1.37434089e+00 -2.02254444e-01
9.40409303e-01 -3.04844379e-02 1.16135851e-02 1.41851461e+00
-6.09171130e-02 4.26636577e-01 -7.42208183e-01 1.02596045e-01
1.52411059e-01 -3.27265710e-01 -8.86191577e-02 4.29196745e-01
2.15371117e-01 7.50887334e-01 -4.69731353e-02 3.75631154e-01
1.86594293e-01 4.01591748e-01 -2.24740624e-01 7.26703763e-01
5.56051672e-01 7.16563463e-01 -1.23833799e+00 -5.32183684e-02
-3.59213620e-01 5.32645583e-01 -2.28647247e-01 -1.68034837e-01
-2.18973055e-01 6.96419597e-01 3.08530152e-01 4.28186506e-01
-9.80454087e-02 2.60214925e-01 -7.33075798e-01 -3.28908861e-01
-1.92157224e-01 -7.39470720e-01 7.61317015e-01 -5.56233585e-01
-1.19145572e+00 6.11751497e-01 -1.04685917e-01 -6.09280586e-01
-3.80161583e-01 -8.10605466e-01 -4.95499820e-01 1.78530470e-01
-1.25667727e+00 -1.22553813e+00 -2.38349617e-01 4.98187095e-01
8.15306246e-01 2.10587844e-01 1.13563418e+00 2.63848603e-01
-5.37183821e-01 5.98590635e-02 3.14547986e-01 -4.21089903e-02
-3.69024388e-02 -1.21707726e+00 -1.42693788e-01 -2.47752070e-01
-5.87185621e-02 1.76446930e-01 3.66945148e-01 -5.87079883e-01
-1.00422955e+00 -3.58235657e-01 5.02958119e-01 1.30117223e-01
-8.35618079e-02 4.06476647e-01 -5.41154325e-01 7.45646179e-01
-3.47333103e-01 1.71129957e-01 1.04987466e+00 -1.78459510e-01
5.63778691e-02 1.19073018e-01 -1.81632793e+00 4.16027784e-01
3.35645109e-01 -2.20402002e-01 -4.03877497e-01 7.07139969e-01
2.16995105e-01 -6.69460356e-01 -1.20124602e+00 5.22341728e-01
1.02736652e+00 -1.00728619e+00 1.44606960e+00 -1.48411870e-01
3.40927601e-01 4.71412629e-01 5.58180315e-03 -7.71662772e-01
-1.37621284e-01 3.82961303e-01 1.32752091e-01 3.64109814e-01
1.46879449e-01 -6.71771049e-01 1.44480348e+00 7.79687285e-01
-5.52239776e-01 -1.06181848e+00 -1.44937360e+00 -9.75272730e-02
5.85706770e-01 -4.32448745e-01 2.50049651e-01 6.45819902e-01
-1.87576339e-01 1.17801934e-01 7.37510324e-01 -4.13589887e-02
6.44944787e-01 2.14828357e-01 -1.09730802e-01 -1.45354164e+00
8.12084377e-02 -6.79647028e-01 -5.93118846e-01 -5.01517057e-01
-4.62854236e-01 -6.92352951e-01 -2.46539131e-01 -2.23456979e+00
1.90432876e-01 -5.80801845e-01 3.06253791e-01 2.61668622e-01
3.47369388e-02 4.41022277e-01 -1.13588855e-01 5.29969215e-01
7.73034215e-01 2.66927630e-01 2.15818357e+00 -3.54289442e-01
-4.76311475e-01 7.25276589e-01 -2.53839850e-01 1.10898936e+00
7.60355592e-01 -4.87812757e-01 -1.27558246e-01 -4.13090050e-01
-5.02740383e-01 1.96058005e-01 -3.54847498e-02 -7.74301052e-01
2.17174292e-01 -2.28988547e-02 3.13955247e-01 -5.71644068e-01
6.86306834e-01 -9.34980154e-01 -2.14900151e-02 1.21686673e+00
5.73677346e-02 -2.80178428e-01 2.16020048e-01 6.04548976e-02
-4.54420060e-01 -2.63086379e-01 1.40154278e+00 -8.02733898e-01
-4.47693795e-01 2.63989925e-01 -5.59402823e-01 -5.02717972e-01
1.05965853e+00 -6.83248997e-01 -6.84850290e-02 -1.13241427e-01
-1.23770010e+00 -3.16941231e-01 -1.13307625e-01 7.04463199e-03
8.67171288e-01 -8.22006762e-01 -5.82032859e-01 4.33894759e-03
-7.92434141e-02 6.63060665e-01 4.15522248e-01 8.20200026e-01
-1.11896610e+00 5.97214401e-01 -3.51259649e-01 -9.63213265e-01
-1.50731158e+00 2.28659868e-01 7.52460837e-01 -4.12269562e-01
-9.33523774e-01 8.91645253e-01 7.83108473e-02 -4.09347594e-01
1.36821315e-01 -6.20069087e-01 -4.84281182e-01 -1.39328986e-01
5.38830817e-01 4.68239605e-01 5.54283619e-01 -6.23709083e-01
-3.31959754e-01 1.22188294e+00 -3.86791468e-01 8.83913860e-02
1.61180949e+00 -1.74191296e-01 -7.38866702e-02 -6.40241988e-03
1.11197948e+00 -4.27690089e-01 -3.75564992e-01 -1.05318137e-01
1.74587071e-01 -1.61929913e-02 8.12981546e-01 -9.34379756e-01
-1.64165127e+00 1.15190613e+00 1.40007436e+00 6.98525161e-02
1.09652650e+00 5.51962793e-01 1.11397159e+00 -3.24291557e-01
4.45655167e-01 -8.04273605e-01 3.15874010e-01 -3.44533473e-01
7.81549096e-01 -1.66311586e+00 2.85199940e-01 -9.27222610e-01
-2.78572351e-01 1.54019153e+00 5.58033586e-01 2.55639732e-01
1.29128683e+00 4.11839098e-01 5.95474362e-01 -8.53858352e-01
-1.96854938e-02 -1.57181788e-02 4.92870137e-02 9.45514083e-01
8.03778231e-01 5.62028550e-02 -1.70445874e-01 -2.75535852e-01
-4.13731337e-01 1.78198546e-01 1.88765693e-02 1.18905413e+00
-5.80524921e-01 -1.13724053e+00 -7.38954008e-01 4.02120769e-01
-8.63369405e-01 7.90957883e-02 -1.21284500e-02 1.05489063e+00
9.42780972e-02 8.17532182e-01 1.52376026e-01 3.90297145e-01
2.55199164e-01 -2.42348492e-01 7.89374709e-01 -8.47660780e-01
-4.56797451e-01 4.13101882e-01 -1.22266240e-01 -3.69237155e-01
-4.20291632e-01 -7.16710508e-01 -1.59973085e+00 -3.61027010e-02
-5.15978813e-01 -3.96970570e-01 1.41375351e+00 1.03996217e+00
-1.27877086e-01 4.52040821e-01 4.43642974e-01 -6.66907668e-01
-9.22956243e-02 -1.12759566e+00 -5.05045354e-01 1.39648154e-01
9.79999453e-02 -4.23154235e-01 -2.34481141e-01 6.21162765e-02] | [13.786505699157715, -2.2265753746032715] |
4001ac46-fd76-4c0c-acc6-8fc1b713326f | convergence-of-communications-control-and | 2307.02663 | null | https://arxiv.org/abs/2307.02663v1 | https://arxiv.org/pdf/2307.02663v1.pdf | Convergence of Communications, Control, and Machine Learning for Secure and Autonomous Vehicle Navigation | Connected and autonomous vehicles (CAVs) can reduce human errors in traffic accidents, increase road efficiency, and execute various tasks ranging from delivery to smart city surveillance. Reaping these benefits requires CAVs to autonomously navigate to target destinations. To this end, each CAV's navigation controller must leverage the information collected by sensors and wireless systems for decision-making on longitudinal and lateral movements. However, enabling autonomous navigation for CAVs requires a convergent integration of communication, control, and learning systems. The goal of this article is to explicitly expose the challenges related to this convergence and propose solutions to address them in two major use cases: Uncoordinated and coordinated CAVs. In particular, challenges related to the navigation of uncoordinated CAVs include stable path tracking, robust control against cyber-physical attacks, and adaptive navigation controller design. Meanwhile, when multiple CAVs coordinate their movements during navigation, fundamental problems such as stable formation, fast collaborative learning, and distributed intrusion detection are analyzed. For both cases, solutions using the convergence of communication theory, control theory, and machine learning are proposed to enable effective and secure CAV navigation. Preliminary simulation results are provided to show the merits of proposed solutions. | ['Choong Seon Hong', 'Walid Saad', 'Omid Semiari', 'Aidin Ferdowsi', 'Tengchan Zeng'] | 2023-07-05 | null | null | null | null | ['autonomous-vehicles', 'autonomous-navigation', 'intrusion-detection', 'navigate', 'decision-making'] | ['computer-vision', 'computer-vision', 'miscellaneous', 'reasoning', 'reasoning'] | [-7.81086460e-02 1.52840674e-01 -3.92714769e-01 1.57525286e-01
-3.24443758e-01 -5.96137881e-01 4.57092673e-01 -1.06292777e-01
-3.79463285e-01 6.01210177e-01 -4.15073305e-01 -9.68475819e-01
-4.92932409e-01 -9.14209843e-01 -3.61088544e-01 -7.41310358e-01
-2.73492128e-01 -6.56065568e-02 4.87256497e-01 -6.72427833e-01
3.27498704e-01 6.50323629e-01 -1.21046770e+00 -8.66298199e-01
1.01160216e+00 1.29277587e+00 -1.58075482e-01 6.15368724e-01
9.05202627e-02 3.84075373e-01 -3.57576370e-01 4.45310883e-02
4.20237690e-01 1.07743748e-01 -2.93969922e-03 -3.00238460e-01
-2.06921279e-01 -1.61490351e-01 -3.45393836e-01 8.88971329e-01
2.32424498e-01 3.42879802e-01 8.13389659e-01 -2.06404209e+00
-6.64733211e-03 1.47977158e-01 -6.13263488e-01 -1.64448857e-01
3.18801589e-02 1.07489467e-01 4.86473769e-01 -2.14949071e-01
1.88497230e-01 9.96548355e-01 7.12533414e-01 6.20439351e-01
-5.22606075e-01 -1.18226779e+00 2.97173589e-01 1.68355823e-01
-1.25535011e+00 -6.33653462e-01 3.15501094e-01 -3.19670558e-01
3.74448419e-01 6.40721843e-02 4.25135434e-01 7.75360882e-01
6.84891820e-01 2.75540739e-01 3.97381037e-01 1.12179548e-01
5.95476449e-01 1.89526260e-01 -9.17217601e-03 5.57964027e-01
7.83641517e-01 4.22868013e-01 -7.43917376e-02 -2.14652956e-01
1.04102522e-01 -2.47564036e-02 -6.68533817e-02 -4.68560606e-01
-7.66927242e-01 7.84231365e-01 4.86262947e-01 -1.38405353e-01
-5.58309972e-01 1.84188187e-02 2.53115267e-01 3.06542426e-01
-3.05821270e-01 2.08542496e-01 -2.40854397e-01 -1.23734653e-01
-1.16956256e-01 -5.28834499e-02 8.05177748e-01 1.33042479e+00
3.82754803e-01 4.35317159e-01 6.54498637e-01 1.62490737e-02
6.51663005e-01 1.11636615e+00 6.08092509e-02 -1.19094169e+00
6.66548848e-01 3.04169774e-01 2.94539690e-01 -1.51318228e+00
-7.58811474e-01 -1.73913270e-01 -1.21640480e+00 4.11810696e-01
1.15648732e-01 -8.20638061e-01 -2.76891738e-01 1.52354729e+00
6.66626394e-01 3.30376923e-01 6.45794809e-01 3.57313305e-01
1.20291613e-01 5.79378724e-01 -1.75771073e-01 -3.33002955e-01
6.81046367e-01 -5.00996470e-01 -9.02229249e-01 -2.13457093e-01
5.97380459e-01 -3.68117213e-01 3.16436797e-01 4.19867001e-02
-5.24886072e-01 -2.03162044e-01 -1.48620236e+00 7.94279873e-01
-5.23478687e-01 -2.00755268e-01 -7.04426318e-02 1.03275609e+00
-8.47838044e-01 -3.65926236e-01 -8.02203655e-01 -3.73295307e-01
2.00212866e-01 4.97261554e-01 1.39526114e-01 1.30923167e-01
-1.30294824e+00 9.04488981e-01 -1.41690895e-01 8.13047215e-02
-5.75055778e-01 -5.11954904e-01 -7.44979262e-01 -2.92561293e-01
6.53400481e-01 -5.49822390e-01 1.09248292e+00 -1.68825671e-01
-1.57660115e+00 -2.65837699e-01 -1.09158017e-01 -7.62042046e-01
3.67448121e-01 3.55024897e-02 -6.64306045e-01 2.95632128e-02
4.12383467e-01 1.03970036e-01 5.57592750e-01 -1.13809192e+00
-1.19296241e+00 -2.97425240e-01 -2.09376529e-01 5.94361834e-02
-4.90149051e-01 -5.93384922e-01 -8.63768235e-02 7.67644197e-02
1.28695741e-01 -1.35801578e+00 -6.42256320e-01 7.16912895e-02
-4.79190558e-01 3.73036563e-02 1.56983840e+00 1.61691502e-01
9.33018267e-01 -2.13220525e+00 -3.58090162e-01 8.02924693e-01
2.42956832e-01 2.14245379e-01 -4.67726821e-03 6.75474763e-01
6.86259806e-01 5.29079922e-02 3.63839746e-01 -1.38733149e-01
-1.10852204e-01 3.34307790e-01 -4.99944478e-01 7.48653769e-01
-4.29182410e-01 2.52466202e-01 -5.71226120e-01 -8.76789913e-02
3.60446095e-01 3.87254536e-01 -4.24768001e-01 -1.71067014e-01
3.95713508e-01 7.36773729e-01 -1.06088483e+00 5.51162422e-01
5.57070613e-01 1.51664183e-01 1.41198225e-02 1.05177596e-01
-5.59097707e-01 -2.14603379e-01 -1.26276922e+00 5.75287640e-01
-5.77724755e-01 7.50013471e-01 8.28404069e-01 -9.13537025e-01
8.59219611e-01 2.57344395e-01 7.38460541e-01 -9.63222027e-01
4.82032984e-01 4.06510234e-01 -3.12647015e-01 -3.80778432e-01
4.58558679e-01 4.84360814e-01 -5.12294650e-01 2.74010926e-01
-7.96930015e-01 -2.34858487e-02 -3.90143812e-01 4.59127188e-01
1.16916895e+00 -9.05515194e-01 1.95321158e-01 -1.71373226e-02
9.28579748e-01 1.61129281e-01 6.24732614e-01 8.58821690e-01
-6.03029490e-01 -4.95019257e-01 1.25943869e-01 7.99505115e-02
-3.76170129e-01 -9.10984397e-01 3.15415740e-01 4.95608628e-01
1.08055985e+00 -8.35733712e-02 -3.05230141e-01 -4.00371104e-01
2.24048465e-01 5.54849267e-01 -1.97127178e-01 -4.41269636e-01
-4.35998261e-01 -1.06914937e-01 6.43571794e-01 7.22738057e-02
6.61345661e-01 1.16753066e-02 -6.26324296e-01 2.33373255e-01
-3.53624811e-03 -1.26557732e+00 -4.17961419e-01 -3.20265502e-01
-2.21023206e-02 -1.62083721e+00 1.38691053e-01 -6.37099624e-01
5.09587586e-01 1.11991787e+00 -1.97939843e-01 1.04779296e-01
1.69275105e-01 9.29347336e-01 -1.63368657e-01 -6.95610821e-01
-4.61944818e-01 1.82531834e-01 8.12161863e-01 3.52643341e-01
-1.47245497e-01 -4.07636583e-01 -5.22713065e-01 8.58032346e-01
-3.11164141e-01 -4.72850263e-01 3.25771719e-01 3.78914386e-01
2.91223735e-01 5.30622482e-01 8.63203883e-01 -2.54477948e-01
6.74137115e-01 -9.30301726e-01 -1.12681890e+00 -1.27326846e-01
-8.59028399e-01 -4.21551913e-01 6.78160846e-01 -2.55490690e-01
-7.27449298e-01 1.24275215e-01 1.09512046e-01 1.37572810e-01
2.40864791e-02 2.13368431e-01 -1.87007397e-01 -6.93722427e-01
4.85414624e-01 3.81666534e-02 7.93608129e-01 4.83144701e-01
4.14702386e-01 1.13786018e+00 4.83307838e-01 7.60457441e-02
1.33315015e+00 8.40290725e-01 4.38611209e-01 -1.25652802e+00
-6.17681742e-02 -5.89729011e-01 -1.12777144e-01 -6.83750451e-01
5.86905360e-01 -9.43057239e-01 -1.37562168e+00 3.83484483e-01
-9.15104508e-01 -1.16031937e-01 3.24065477e-01 5.81062675e-01
-3.77801925e-01 1.46503195e-01 -1.07376680e-01 -1.20074117e+00
-1.90322861e-01 -1.05206561e+00 2.92735517e-01 5.94857991e-01
-6.72515333e-02 -1.01941180e+00 -4.97367345e-02 2.09547594e-01
1.02743459e+00 4.47277606e-01 3.75211507e-01 -3.76691520e-01
-8.15981150e-01 -6.63085878e-01 1.23918287e-01 -2.17883825e-01
1.65022969e-01 -2.65924577e-02 -1.54059410e-01 -4.68051702e-01
-1.66008919e-01 1.39252171e-01 2.25442186e-01 3.84049952e-01
3.11533540e-01 -5.80137491e-01 -9.19709504e-01 4.03501570e-01
1.05750728e+00 7.66291261e-01 1.70180321e-01 4.89153653e-01
1.57736734e-01 6.98470831e-01 9.28948224e-01 2.89504439e-01
8.91542375e-01 4.25294042e-01 1.07462072e+00 4.26328599e-01
5.05345702e-01 -7.71712288e-02 6.13026619e-01 5.72082400e-01
5.96292794e-01 -4.53638196e-01 -6.40038073e-01 3.41248006e-01
-1.89879334e+00 -7.17719078e-01 -3.57741296e-01 2.00013185e+00
-2.02755406e-01 3.22964102e-01 -5.27919270e-02 -4.00719121e-02
8.89445961e-01 -1.74855530e-01 -7.70304561e-01 -5.46086468e-02
1.53014451e-01 -7.57680476e-01 1.24781966e+00 6.69689357e-01
-8.95616710e-01 6.04376972e-01 5.63468790e+00 7.03651130e-01
-1.15756583e+00 6.20370992e-02 2.38772910e-02 3.35871756e-01
2.13689301e-02 -1.44870162e-01 -9.37480092e-01 2.80362636e-01
1.03536487e+00 -2.83410490e-01 1.61912903e-01 9.92624819e-01
7.41402149e-01 -1.60473585e-01 -1.40966356e-01 5.90193391e-01
-2.66034991e-01 -1.42779243e+00 -4.53309327e-01 3.36406916e-01
6.99406803e-01 2.95554101e-01 2.70449489e-01 1.59499809e-01
4.34354752e-01 -4.37155366e-01 4.61098939e-01 2.49405131e-01
3.41948241e-01 -1.09067166e+00 5.14236033e-01 5.78505158e-01
-1.74956799e+00 -7.48057485e-01 6.00769520e-02 -5.94292358e-02
6.13555849e-01 -5.30488975e-02 -2.71087855e-01 5.52053630e-01
3.47261935e-01 5.33594728e-01 -1.34789601e-01 1.12301958e+00
-9.53143910e-02 2.83848941e-01 -5.93853593e-01 -5.54117143e-01
3.90412599e-01 -1.86113924e-01 1.00261533e+00 2.98498720e-01
4.66701239e-01 2.68912852e-01 4.25082505e-01 1.19402133e-01
6.05772555e-01 -2.42379248e-01 -1.09186673e+00 4.77400392e-01
1.25086820e+00 1.06930780e+00 -4.25706655e-01 2.75119450e-02
-3.83250087e-01 3.63080278e-02 -3.29280555e-01 5.11779606e-01
-8.76855612e-01 -9.98883247e-01 1.24739563e+00 2.30657950e-01
4.84544523e-02 -8.61071169e-01 -5.35758078e-01 -2.93989331e-01
-3.39102387e-01 -4.05977190e-01 -6.74818363e-03 -7.95414150e-02
-8.64442050e-01 4.49537784e-01 -2.96383739e-01 -1.58403373e+00
-2.31592342e-01 -1.79873899e-01 -8.78399312e-01 1.09856494e-01
-1.48314393e+00 -7.19011724e-01 -2.65739828e-01 1.00772107e+00
1.54635400e-01 -6.97064161e-01 3.84281129e-01 3.59237254e-01
-6.81729138e-01 5.24580002e-01 4.07317728e-01 -1.51793480e-01
4.01974618e-01 -3.23181897e-01 9.17966887e-02 1.04636037e+00
-5.84434569e-01 3.86689484e-01 4.92652088e-01 -7.52123713e-01
-1.93741333e+00 -1.39692676e+00 4.04707938e-01 7.11104050e-02
7.81201184e-01 6.60185441e-02 -1.00419700e-01 3.64966393e-01
-6.43344894e-02 -4.21051413e-01 2.63634175e-01 -4.69100326e-01
-1.01050049e-01 -3.42558116e-01 -1.07460749e+00 1.06185174e+00
7.21428037e-01 -3.15647312e-02 3.10794264e-01 1.73950821e-01
6.84610724e-01 -2.17625186e-01 -2.52457231e-01 2.89375961e-01
5.16961992e-01 -4.17566836e-01 8.11325550e-01 -1.36147276e-01
-8.34622383e-01 -4.68334794e-01 -3.24621201e-01 -1.14750862e+00
2.24328697e-01 -1.08937371e+00 1.64262593e-01 8.33662689e-01
3.73168051e-01 -1.47286928e+00 6.65725410e-01 5.32272756e-01
-1.90863609e-01 -3.52135956e-01 -1.36444950e+00 -9.57309783e-01
-2.30424911e-01 -7.99416125e-01 3.77736360e-01 4.97504264e-01
2.21020579e-01 3.97915572e-01 -4.16445643e-01 1.08174849e+00
8.73346329e-01 -5.43910325e-01 1.23764384e+00 -1.15626645e+00
6.99558735e-01 -4.46291238e-01 -4.81087089e-01 -1.20776772e+00
3.22358578e-01 -2.96541512e-01 2.45874301e-01 -1.33971691e+00
-9.77970719e-01 -8.11790287e-01 3.17205518e-01 1.75236061e-01
7.05146134e-01 -1.24576792e-01 -5.87220751e-02 1.89701930e-01
-7.86852598e-01 6.91718817e-01 7.60274172e-01 -2.41890535e-01
-5.20790160e-01 8.55099797e-01 -5.77648222e-01 5.85068703e-01
1.25974917e+00 -2.14431256e-01 -7.65154839e-01 -1.03879683e-01
-3.14430222e-02 5.28184891e-01 2.48761952e-01 -1.26976705e+00
1.24776697e+00 -4.63646233e-01 -6.10318780e-01 -6.26776576e-01
4.29179579e-01 -1.43582880e+00 -4.71639894e-02 1.05473113e+00
1.06450178e-01 3.32351699e-02 8.60426277e-02 1.42030251e+00
-5.17496765e-02 4.80006754e-01 7.83101976e-01 6.64240718e-01
-5.99196315e-01 3.90542060e-01 -1.51348400e+00 -1.51606336e-01
1.74551153e+00 -2.77018249e-01 -5.45363426e-01 -1.03720069e+00
-3.46541643e-01 1.11110473e+00 -8.94869715e-02 7.21647918e-01
7.88771093e-01 -1.05183399e+00 7.73717836e-03 4.35369045e-01
-2.19909415e-01 -2.68334329e-01 3.20791543e-01 1.13600254e+00
-1.87388211e-01 6.87810004e-01 -7.52592683e-02 -5.12028277e-01
-1.13843656e+00 1.75634071e-01 4.23190594e-01 5.57075739e-01
-1.20016582e-01 2.91539311e-01 -3.27106237e-01 -6.07166827e-01
5.95952809e-01 1.15409970e-01 -2.76468933e-01 -8.31262842e-02
5.37295818e-01 6.33496404e-01 -3.15168291e-01 -7.95826256e-01
-3.59996736e-01 7.38821089e-01 2.38280565e-01 -1.29289761e-01
6.13929987e-01 -1.14907110e+00 2.37269536e-01 -1.96810618e-01
8.63904655e-01 1.17888689e-01 -1.05285013e+00 -1.29433334e-01
-2.48437226e-02 -5.04490547e-02 -1.48156174e-02 -5.55444472e-02
-1.19695616e+00 2.25170717e-01 5.06295025e-01 3.83618504e-01
6.22106314e-01 -4.60056156e-01 1.14736450e+00 6.73058808e-01
8.93152058e-01 -1.04461098e+00 2.46955082e-01 7.96054184e-01
1.36495948e-01 -1.06399620e+00 -4.37156200e-01 -6.29079342e-01
-4.07873690e-01 1.03996289e+00 9.56467569e-01 -4.97127250e-02
1.21572232e+00 2.98255235e-01 2.58466423e-01 -8.37347433e-02
-5.89256585e-01 -2.10081741e-01 -3.49061102e-01 9.62010026e-01
-8.75745535e-01 -8.55727717e-02 -1.16895381e-02 3.73041421e-01
-3.32134999e-02 -6.51974857e-01 8.43837380e-01 1.09856772e+00
-9.13007319e-01 -7.96208799e-01 -4.05334353e-01 2.60170609e-01
1.10729679e-01 6.81718588e-01 -4.07022983e-02 8.97746027e-01
-1.36686981e-01 1.74539149e+00 -4.99702878e-02 -8.95592809e-01
3.54431897e-01 -6.24733567e-01 -5.74584126e-01 6.33959994e-02
2.11712182e-01 -3.67370009e-01 -3.11351810e-02 -6.34889841e-01
-1.63124397e-01 -5.52282333e-01 -1.48339725e+00 -5.69356561e-01
-3.90495688e-01 4.16501880e-01 1.00781739e+00 9.45858896e-01
7.74816334e-01 3.48140597e-01 1.12343812e+00 -5.20370066e-01
-6.44593477e-01 -1.57995328e-01 -3.23384196e-01 -7.07050800e-01
6.02426589e-01 -8.63806307e-01 -6.73923075e-01 -8.59803855e-01] | [5.586440563201904, 1.6057730913162231] |
be72dd3e-9b77-4ead-9722-2962240574ed | ecological-sampling-of-gaze-shifts | null | null | https://ieeexplore.ieee.org/abstract/document/6502674?casa_token=1byinAjZ5pcAAAAA:rTMkAo5EL0GZ0i_cn00bG3r4fKPrmSkMII18iE6DPq9UZ9xN9HMWOid8cMjfrh8p1O_bB14 | https://ieeexplore.ieee.org/abstract/document/6502674?casa_token=1byinAjZ5pcAAAAA:rTMkAo5EL0GZ0i_cn00bG3r4fKPrmSkMII18iE6DPq9UZ9xN9HMWOid8cMjfrh8p1O_bB14 | Ecological Sampling of Gaze Shifts | Visual attention guides our gaze to relevant parts of the viewed scene, yet the moment-to-moment relocation of gaze can be different among observers even though the same locations are taken into account. Surprisingly, the variability of eye movements has been so far overlooked by the great majority of computational models of visual attention. In this paper we present the ecological sampling model, a stochastic model of eye guidance explaining such variability. The gaze shift mechanism is conceived as an active random sampling that the foraging eye carries out upon the visual landscape, under the constraints set by the observable features and the global complexity of the landscape. By drawing on results reported in the foraging literature, the actual gaze relocation is eventually driven by a stochastic differential equation whose noise source is sampled from a mixture of α-stable distributions. This way, the sampling strategy proposed here allows to mimic a fundamental property of the eye guidance mechanism: where we choose to look next at any given moment in time, it is not completely deterministic, but neither is it completely random To show that the model yields gaze shift motor behaviors that exhibit statistics similar to those displayed by human observers, we compare simulation outputs with those obtained from eye-tracked subjects while viewing complex dynamic scenes. | ['Mario Ferraro', 'Giuseppe Boccignone'] | 2013-04-16 | null | null | null | ieee-transactions-on-cybernetics-2013-4 | ['gaze-estimation', 'eye-tracking'] | ['computer-vision', 'computer-vision'] | [ 2.25421950e-01 -2.94834301e-02 3.59192759e-01 1.17029458e-01
4.06417251e-01 -6.19924009e-01 4.94751304e-01 -6.90112785e-02
-6.62293494e-01 5.68560064e-01 -5.13865352e-02 -1.71461061e-01
-4.79958504e-01 -2.95454681e-01 -6.71969056e-01 -1.12425053e+00
-9.42109972e-02 -1.36539638e-01 3.50520164e-01 -2.57629484e-01
4.98408258e-01 3.42178851e-01 -2.26974940e+00 -4.01499808e-01
5.48364937e-01 6.49206221e-01 6.81934118e-01 1.01320136e+00
1.81168407e-01 6.40650749e-01 -3.81999522e-01 1.04885072e-01
2.23233044e-01 -8.63493443e-01 -4.92859215e-01 3.18922997e-01
2.09475607e-01 1.65930808e-01 1.12037487e-01 1.08535230e+00
3.78180027e-01 4.25354987e-01 9.77654755e-01 -9.22974169e-01
-8.75282228e-01 1.25485912e-01 -6.83116019e-01 5.80657184e-01
4.73559946e-01 5.42401075e-01 8.29329908e-01 -4.61580694e-01
7.39596009e-01 1.02670765e+00 1.76584125e-01 5.77472150e-01
-1.58381832e+00 4.61134426e-02 4.45176542e-01 1.08337410e-01
-1.34314919e+00 -3.87734711e-01 5.42837024e-01 -6.94865048e-01
4.77110118e-01 3.94987136e-01 1.00606036e+00 9.08088088e-01
8.73005390e-01 2.58198917e-01 1.34292936e+00 -6.67287827e-01
5.39399207e-01 -1.39436163e-02 8.36459640e-03 4.79128599e-01
3.61784518e-01 3.39899987e-01 -8.21640968e-01 -8.35952386e-02
5.32083154e-01 -4.74114791e-02 -6.76121593e-01 -6.29263103e-01
-9.52990890e-01 6.09976411e-01 4.43656176e-01 1.22631967e-01
-5.52932322e-01 5.65453172e-02 -3.95106256e-01 3.13958138e-01
3.17661822e-01 2.34936774e-01 -1.61579996e-01 -4.02444683e-04
-5.73614061e-01 2.15373039e-01 6.18871093e-01 6.55352712e-01
8.59585762e-01 -4.39238936e-01 -1.39990211e-01 2.91095853e-01
6.18792117e-01 1.02948654e+00 2.66985595e-01 -1.00308406e+00
-1.93252265e-01 4.27565932e-01 4.79797602e-01 -9.58973944e-01
-4.99473542e-01 -3.36659908e-01 -4.07744229e-01 7.47099578e-01
9.25278008e-01 -4.38521385e-01 -5.27193606e-01 2.11662579e+00
3.37800831e-01 -3.72558653e-01 -3.17131788e-01 1.13614881e+00
1.04355596e-01 3.80129039e-01 2.26783883e-02 -5.11571527e-01
1.31184697e+00 -1.77586243e-01 -7.76939094e-01 -2.22019300e-01
1.91387102e-01 -5.81540465e-01 1.12546349e+00 3.65110040e-01
-1.10502911e+00 -2.02310011e-01 -7.16916323e-01 1.96277127e-01
-1.52695671e-01 -2.50697583e-01 1.71904132e-01 4.57746983e-01
-1.35703909e+00 2.13480532e-01 -7.95556605e-01 -9.52849030e-01
6.64267540e-02 6.99497908e-02 -8.36812612e-03 4.62522596e-01
-5.82695723e-01 8.80224526e-01 -3.54907006e-01 2.37069800e-01
-8.81693125e-01 -2.73401290e-01 -4.17380720e-01 1.64463088e-01
3.95660073e-01 -8.06730032e-01 1.53195703e+00 -1.38208735e+00
-1.60570979e+00 8.04581583e-01 -8.89122605e-01 -2.06979007e-01
4.52881932e-01 4.61157225e-02 1.47906110e-01 -2.36958507e-02
-5.21454997e-02 4.48449582e-01 9.24601853e-01 -1.27659059e+00
-5.99826753e-01 -6.70748234e-01 4.82707806e-02 4.38724786e-01
4.20615375e-01 1.70181528e-01 8.35719779e-02 -1.56466842e-01
-6.78710267e-02 -1.21604311e+00 -2.72056520e-01 2.41730347e-01
-9.92233306e-02 -7.71152526e-02 2.40155235e-01 1.38515487e-01
1.24982142e+00 -2.12277198e+00 3.89805228e-01 -2.66892035e-02
3.11513633e-01 -3.17161739e-01 2.25981668e-01 7.05887139e-01
1.43950477e-01 -8.73527303e-02 -1.45659715e-01 -3.13358188e-01
-5.55616468e-02 1.93173029e-02 -2.40791723e-01 8.89028430e-01
-2.33868081e-02 6.57577217e-01 -9.41439033e-01 -5.16364165e-02
-1.85314104e-01 4.08597976e-01 -6.02382541e-01 1.35439813e-01
-1.51033625e-01 6.80837929e-01 -4.27592695e-01 1.52923062e-01
4.08938497e-01 -3.87900591e-01 -2.22436637e-02 6.17105544e-01
-6.94831789e-01 -3.60158905e-02 -1.04307008e+00 1.21891046e+00
-1.82201281e-01 1.05848825e+00 -1.16402730e-02 -1.98373556e-01
4.58863318e-01 -2.44774465e-02 -7.28035420e-02 -6.49456739e-01
4.21675682e-01 -1.61641717e-01 5.75742126e-01 -6.50433362e-01
5.11349916e-01 -1.40517935e-01 2.98456758e-01 9.18968439e-01
-2.69959688e-01 2.02240691e-01 1.60863966e-01 -1.01300098e-01
7.84698844e-01 1.36701599e-01 4.47004288e-01 -8.35704148e-01
2.47233570e-01 -1.32894004e-02 2.10501209e-01 8.97828639e-01
-3.05591971e-01 5.23126066e-01 6.43404245e-01 -1.94081411e-01
-7.48966038e-01 -9.40236747e-01 -2.12171659e-01 1.21398854e+00
6.20158136e-01 3.88878845e-02 -1.29304397e+00 2.00591326e-01
-2.42043883e-01 7.34665275e-01 -1.32160366e+00 -2.70767838e-01
-1.46718025e-01 -6.53015256e-01 -2.53228515e-01 -2.35363096e-01
-9.23243463e-02 -1.41328657e+00 -1.85228884e+00 -9.57610160e-02
1.40631005e-01 -4.24009949e-01 -6.39389932e-01 1.12338409e-01
-5.17571092e-01 -1.25892735e+00 -8.24883580e-01 -1.93333060e-01
8.30498457e-01 5.47438025e-01 1.12085235e+00 7.22174942e-02
-3.78755867e-01 6.85412049e-01 -1.98214263e-01 -7.10170925e-01
-2.21444070e-01 -1.14374422e-01 2.28300437e-01 5.17545938e-01
5.67605197e-01 -3.20150256e-01 -8.89285684e-01 3.56072187e-01
-7.65625656e-01 -1.97514042e-01 5.92448041e-02 3.00464064e-01
4.23905551e-01 -1.48213476e-01 -2.07317639e-02 -2.78808206e-01
6.40268564e-01 -5.87686479e-01 -8.94883931e-01 2.11738676e-01
-3.39133441e-01 1.60716817e-01 1.56702116e-01 -7.23558843e-01
-9.38224316e-01 -1.79538712e-01 6.03291571e-01 9.14175436e-02
-3.83862019e-01 1.91941127e-01 2.24393398e-01 -1.38903975e-01
8.49790812e-01 3.71213853e-01 2.93941677e-01 -3.05240661e-01
1.85596853e-01 4.13874865e-01 -3.13500948e-02 -1.89972892e-01
4.56889957e-01 7.61881828e-01 2.05339685e-01 -1.15534818e+00
-6.56339586e-01 7.06892684e-02 -6.65296733e-01 -5.94118714e-01
9.63536024e-01 -4.13095117e-01 -1.44788063e+00 6.19083464e-01
-9.35134351e-01 -7.52933443e-01 -4.94339168e-01 4.91321266e-01
-8.00073087e-01 -1.82082281e-01 2.22261533e-01 -1.56938112e+00
2.66179025e-01 -1.16845906e+00 1.03566122e+00 7.33096421e-01
-1.92780554e-01 -9.56045330e-01 3.48843932e-01 -5.99688113e-01
5.96205771e-01 -2.14384980e-02 6.51796341e-01 6.55340850e-02
-9.06437933e-01 2.87040383e-01 1.69669420e-01 -2.14901209e-01
2.52329528e-01 5.25985599e-01 -8.69185090e-01 -7.56874532e-02
3.07155818e-01 3.56785208e-01 5.50116241e-01 1.03684962e+00
3.42643350e-01 -5.70618436e-02 -2.88256973e-01 5.57439506e-01
1.44210529e+00 2.52297878e-01 2.09801912e-01 3.22621763e-01
9.67438817e-02 1.06833100e+00 3.24148268e-01 4.84250933e-01
4.65053171e-01 5.23004770e-01 7.85093606e-01 3.00152808e-01
2.56120801e-01 -3.35870832e-02 1.94495186e-01 -1.07429817e-01
-2.01025471e-01 -5.71239054e-01 -7.43746161e-01 6.52811944e-01
-1.60350382e+00 -1.10792911e+00 -2.93419868e-01 2.73401761e+00
4.26212788e-01 -5.28404005e-02 3.91982347e-01 -2.90006638e-01
6.83446169e-01 -4.99893166e-02 -6.54967308e-01 -2.80622691e-01
-1.89240187e-01 -3.72093290e-01 5.62536299e-01 8.19825172e-01
-3.97969514e-01 5.39281249e-01 7.10595036e+00 6.21627010e-02
-1.22460437e+00 -4.31808233e-02 2.21542835e-01 -5.07987678e-01
-2.79909670e-01 3.84406932e-02 -9.01464105e-01 7.84896314e-01
8.75761509e-01 -4.39338624e-01 8.15626025e-01 1.37352899e-01
7.24929869e-01 -1.06419194e+00 -9.48024750e-01 6.41237617e-01
2.66565084e-02 -6.24305367e-01 -2.61745602e-01 5.82543671e-01
3.59392881e-01 9.09847170e-02 4.63713109e-01 -4.89610791e-01
4.35239941e-01 -8.75473738e-01 1.07600963e+00 1.20233905e+00
4.05124426e-01 -3.76363993e-01 1.35037169e-01 8.28252494e-01
-6.97959602e-01 -3.47637683e-01 -3.98213804e-01 -5.20986021e-01
3.05541843e-01 2.10339159e-01 -2.41891459e-01 -1.57376036e-01
7.47305930e-01 1.73074648e-01 -6.66600585e-01 1.20373964e+00
-1.50459543e-01 3.68631810e-01 -4.79191273e-01 -4.85814720e-01
5.04201502e-02 -4.52881396e-01 9.32904899e-01 4.69417006e-01
3.87702286e-01 7.07984567e-02 -5.87254763e-01 1.21684146e+00
5.50981045e-01 7.30706425e-03 -8.61754417e-01 3.72202486e-01
9.46226865e-02 9.18448091e-01 -7.12897301e-01 2.55354583e-01
-1.84582978e-01 7.06895232e-01 2.79069841e-01 7.64824271e-01
-5.75176418e-01 -6.28028139e-02 9.74852860e-01 4.25533772e-01
4.17080343e-01 -2.85778582e-01 2.16718204e-02 -9.12813306e-01
-1.56612188e-01 -2.79934347e-01 -1.37128636e-01 -1.12598825e+00
-9.64952767e-01 7.87416160e-01 1.38060138e-01 -9.80826795e-01
-2.14094341e-01 -4.80748028e-01 -4.77735251e-01 1.33853185e+00
-1.30679309e+00 -3.55813295e-01 -3.30687732e-01 4.86313939e-01
2.77418971e-01 1.69849083e-01 4.38718557e-01 -4.88477111e-01
-5.39936841e-01 1.88199073e-01 2.67024100e-01 -5.43193460e-01
4.11617160e-01 -1.27786160e+00 2.15820387e-01 9.39087808e-01
1.67403948e-02 8.90473783e-01 1.15687251e+00 -4.22815710e-01
-1.27338076e+00 -4.72037494e-01 8.40213835e-01 -9.16025043e-01
7.23422289e-01 -4.17117655e-01 -8.08751583e-01 4.97468263e-01
6.14220142e-01 4.33260202e-02 5.52421033e-01 -2.19577998e-01
2.55248964e-01 2.16841400e-01 -7.96202481e-01 8.86310577e-01
1.00217533e+00 -2.96198726e-01 -4.97611195e-01 -1.00218780e-01
3.18862766e-01 -2.13031128e-01 9.89978388e-03 -1.17170744e-01
6.62964821e-01 -1.45709682e+00 3.16665202e-01 -6.39259517e-01
2.52533220e-02 -4.90909606e-01 -8.88102129e-03 -1.38810444e+00
-5.42444229e-01 -1.00229406e+00 1.41002670e-01 8.41220438e-01
2.29377076e-01 -9.25840080e-01 3.06627512e-01 3.12091261e-01
5.00103474e-01 -4.01715249e-01 -9.51933086e-01 -4.96019274e-01
-2.68415473e-02 -5.85267805e-02 4.11433250e-01 6.15714751e-02
-7.97805786e-02 2.89577395e-01 4.37661037e-02 3.94403368e-01
8.31461191e-01 7.70045146e-02 7.54516721e-01 -1.45694029e+00
-3.02045733e-01 -6.13081336e-01 -1.48585483e-01 -1.17917538e+00
-1.11425608e-01 -1.92356899e-01 4.26242620e-01 -1.05864072e+00
2.31166914e-01 9.63201467e-03 -1.13814525e-01 -3.10897142e-01
-3.49761635e-01 5.42673573e-04 1.44018665e-01 3.89186680e-01
-3.97095352e-01 5.22145689e-01 1.30567658e+00 5.62376082e-01
-4.94862735e-01 2.60713041e-01 -9.02942777e-01 6.29671633e-01
3.64277273e-01 -4.90432054e-01 -5.24871469e-01 -3.97343189e-01
7.22321033e-01 7.17530213e-03 5.41872144e-01 -7.55884349e-01
3.94580781e-01 -4.25737560e-01 -1.65929690e-01 -2.58980513e-01
3.76382247e-02 -9.76310849e-01 2.87373900e-01 3.82792026e-01
-3.21669728e-01 1.92857116e-01 -1.71284657e-02 8.52066100e-01
3.96817803e-01 -3.63821328e-01 8.96783292e-01 -1.34867072e-01
-1.91709861e-01 8.79507735e-02 -1.13121307e+00 1.15068682e-01
1.12403715e+00 -3.53502512e-01 -2.73467481e-01 -4.59509879e-01
-8.67916703e-01 1.14168255e-02 1.06781733e+00 1.74170226e-01
-1.43820792e-01 -7.31957555e-01 -3.74870330e-01 3.19156110e-01
1.20515667e-01 -1.35964200e-01 3.12636018e-01 1.10160816e+00
-4.42568302e-01 3.29968989e-01 -2.03874663e-01 -9.44535136e-01
-1.12471664e+00 8.81568193e-01 6.58756971e-01 5.28841853e-01
-3.45323801e-01 6.98041558e-01 7.61632860e-01 3.19497883e-01
-1.16035402e-01 -4.92895752e-01 -4.68652636e-01 2.49444619e-01
7.81636059e-01 2.87162572e-01 -3.21668029e-01 -9.25954700e-01
-2.78311908e-01 1.02096128e+00 4.06188726e-01 -1.54699221e-01
9.54562306e-01 -1.01826644e+00 -8.93119723e-02 1.09350038e+00
3.98651391e-01 2.32472464e-01 -1.73598361e+00 -1.54138938e-01
-2.54760891e-01 -4.42351609e-01 -1.32599816e-01 -5.53058803e-01
-6.38492048e-01 9.36731815e-01 6.12170935e-01 9.07438338e-01
1.17929041e+00 2.39978775e-01 -1.66888610e-01 1.38859630e-01
5.24652064e-01 -6.43027246e-01 -2.86770135e-01 3.33269238e-01
8.19246769e-01 -9.61381137e-01 -3.81070107e-01 -3.94310504e-02
-6.44477308e-01 6.56955421e-01 1.96637586e-01 -3.06070268e-01
8.46328199e-01 9.02546793e-02 2.03056306e-01 -2.36914366e-01
-1.12940204e+00 -6.86681271e-01 2.99899608e-01 7.31709778e-01
2.09107861e-01 -2.41873339e-01 -2.53334612e-01 7.25816982e-03
-2.04710111e-01 2.96949851e-03 9.11798835e-01 6.86424375e-01
-6.91983879e-01 -4.71448392e-01 -4.00928080e-01 -6.00601509e-02
-5.03574073e-01 -3.42780761e-02 -3.66865575e-01 7.19900906e-01
-1.41721731e-02 1.00782764e+00 4.80773807e-01 2.08437040e-01
2.96252102e-01 -6.71065897e-02 6.20285869e-01 -7.07704365e-01
-2.93144017e-01 3.81845236e-02 -6.06575489e-01 -5.22014737e-01
-6.05591595e-01 -1.23928201e+00 -7.95967638e-01 -1.47268534e-01
-2.50536054e-01 9.25221518e-02 6.01142168e-01 8.91171038e-01
4.44232434e-01 4.85930026e-01 5.50223291e-01 -1.12312210e+00
-3.70956987e-01 -8.92824650e-01 -1.00237846e+00 5.77920601e-02
9.97232080e-01 -9.31483805e-01 -9.78439748e-01 2.05295011e-01] | [10.062028884887695, 1.6374067068099976] |
d09dc1db-bc30-4894-9fc6-6a60a008fbfa | sci-a-spectrum-concentrated-implicit-neural | 2209.15180 | null | https://arxiv.org/abs/2209.15180v5 | https://arxiv.org/pdf/2209.15180v5.pdf | SCI: A Spectrum Concentrated Implicit Neural Compression for Biomedical Data | Massive collection and explosive growth of biomedical data, demands effective compression for efficient storage, transmission and sharing. Readily available visual data compression techniques have been studied extensively but tailored for natural images/videos, and thus show limited performance on biomedical data which are of different features and larger diversity. Emerging implicit neural representation (INR) is gaining momentum and demonstrates high promise for fitting diverse visual data in target-data-specific manner, but a general compression scheme covering diverse biomedical data is so far absent. To address this issue, we firstly derive a mathematical explanation for INR's spectrum concentration property and an analytical insight on the design of INR based compressor. Further, we propose a Spectrum Concentrated Implicit neural compression (SCI) which adaptively partitions the complex biomedical data into blocks matching INR's concentrated spectrum envelop, and design a funnel shaped neural network capable of representing each block with a small number of parameters. Based on this design, we conduct compression via optimization under given budget and allocate the available parameters with high representation accuracy. The experiments show SCI's superior performance to state-of-the-art methods including commercial compressors, data-driven ones, and INR based counterparts on diverse biomedical data. The source code can be found at https://github.com/RichealYoung/ImplicitNeuralCompression.git. | ['Qionghai Dai', 'Jinli Suo', 'Jinyuan Qu', 'Qianni Cao', 'Yuxiao Cheng', 'Tingxiong Xiao', 'Runzhao Yang'] | 2022-09-30 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 4.20270115e-01 -1.85538799e-01 -4.64622885e-01 -1.68250248e-01
-4.44539666e-01 5.32955751e-02 -2.90738009e-02 1.88065007e-01
-2.67386049e-01 7.45734155e-01 3.09548110e-01 -1.79222509e-01
-4.90486145e-01 -5.27846634e-01 -5.34260929e-01 -8.84194016e-01
-3.97936970e-01 3.71448189e-01 -1.99592769e-01 -3.24138738e-02
1.91794097e-01 5.97634315e-01 -1.41068184e+00 3.45617741e-01
7.93219984e-01 1.19870412e+00 6.12438381e-01 5.72175205e-01
1.42575100e-01 6.83732748e-01 -5.47326446e-01 -9.74461734e-02
2.76084721e-01 -4.37548816e-01 -3.94973695e-01 -1.60437688e-01
-1.05429135e-01 -2.52990305e-01 -9.05211866e-01 1.09237564e+00
9.38498914e-01 2.11226922e-02 8.78368914e-01 -1.04319775e+00
-1.04813230e+00 3.41055602e-01 -6.12989843e-01 5.19379199e-01
-1.32497117e-01 -2.17706454e-03 4.94732678e-01 -8.32937539e-01
5.40939867e-01 8.85611892e-01 6.97867274e-01 8.19842398e-01
-9.08264458e-01 -6.34616315e-01 -3.81798357e-01 3.17189991e-01
-1.67609084e+00 -6.02225006e-01 8.52267146e-01 -1.17105559e-01
7.35880435e-01 5.35893798e-01 9.27985966e-01 9.97025073e-01
3.33921462e-01 7.99678564e-01 8.80483150e-01 -2.80332863e-01
2.32335806e-01 -5.46547845e-02 6.92001581e-02 6.06273949e-01
2.90419996e-01 -1.22686485e-02 -5.57819486e-01 -1.34488806e-01
9.96838152e-01 3.00108910e-01 -7.55391359e-01 4.24214564e-02
-9.24630046e-01 7.10411966e-01 5.00242531e-01 3.50718647e-01
-4.24833298e-01 1.44093350e-01 6.81713641e-01 2.53384501e-01
1.74796104e-01 1.02534235e-01 -1.30573958e-01 -8.12600628e-02
-1.13478553e+00 1.52347594e-01 5.94551504e-01 1.22331953e+00
1.72687158e-01 4.02330548e-01 -2.25672647e-01 1.23225820e+00
1.17324442e-01 3.74325126e-01 9.53420937e-01 -8.55476677e-01
3.16644490e-01 3.66117895e-01 -5.71913540e-01 -1.33759463e+00
-4.72773314e-01 -8.72544706e-01 -1.68041134e+00 -2.33439445e-01
-1.67705595e-01 1.13743223e-01 -9.81306493e-01 1.53358293e+00
7.78706148e-02 3.74722220e-02 8.67188200e-02 1.00253713e+00
9.11915302e-01 8.51247609e-01 -5.96258360e-05 -6.85927272e-01
1.35133195e+00 -6.17297769e-01 -9.25770760e-01 1.79045692e-01
2.09163591e-01 -4.19163585e-01 7.76299238e-01 5.66227794e-01
-1.38361347e+00 -4.54615742e-01 -1.16598761e+00 -2.41480231e-01
-1.42325938e-01 1.69767871e-01 3.80305767e-01 3.66433144e-01
-1.06172955e+00 6.20159924e-01 -8.00033391e-01 2.97573097e-02
8.52881312e-01 3.24459910e-01 -1.05900794e-01 -2.48339698e-01
-1.22292423e+00 5.76470733e-01 5.86184680e-01 1.53160319e-01
-7.74726927e-01 -8.11603546e-01 -5.58308661e-01 8.34967345e-02
-3.06949466e-02 -9.27103102e-01 8.28972399e-01 -8.17710817e-01
-1.09021270e+00 6.51532888e-01 1.90492809e-01 -7.02391565e-01
3.72583449e-01 2.15244949e-01 -3.99122298e-01 6.72851562e-01
-3.63267690e-01 7.30988145e-01 8.61974299e-01 -1.12262094e+00
-1.38585091e-01 -5.26878953e-01 -5.80024362e-01 3.17620844e-01
-8.18934679e-01 -2.32198223e-01 -5.98810375e-01 -1.11084485e+00
4.97005992e-02 -5.54297388e-01 -2.54893959e-01 4.68512356e-01
-2.97516048e-01 1.41495029e-02 8.21799517e-01 -9.41482306e-01
1.60970414e+00 -2.17804098e+00 2.81829566e-01 1.99449837e-01
5.09103477e-01 4.90021884e-01 -9.19070169e-02 3.67873967e-01
-2.15842817e-02 -3.31292450e-02 -4.87170696e-01 -2.59298712e-01
-2.64209807e-01 3.07214528e-01 -6.66650385e-02 5.50590217e-01
-1.85827106e-01 8.89892638e-01 -4.97473538e-01 -6.58754766e-01
-1.30219892e-01 8.40858757e-01 -4.84496236e-01 2.58522511e-01
7.55305514e-02 7.42939860e-02 -5.04499257e-01 1.08403957e+00
8.65173459e-01 -4.81811643e-01 1.59425154e-01 -6.08688951e-01
1.72642156e-01 -3.77811849e-01 -6.73812330e-01 1.60403526e+00
1.33463830e-01 6.53169513e-01 2.75103748e-01 -1.53719544e+00
1.09285522e+00 3.68629098e-01 7.03168035e-01 -8.53948236e-01
3.52031618e-01 3.30644399e-01 -7.95601159e-02 -8.08287323e-01
3.67109507e-01 -2.11378008e-01 3.54936719e-01 8.11748281e-02
-1.06198564e-01 2.14747116e-01 1.03621930e-01 2.06540674e-01
1.05029428e+00 -2.47097492e-01 3.58799011e-01 -2.01855302e-01
2.35421076e-01 -7.19882175e-02 4.53006595e-01 4.77136284e-01
-3.72132033e-01 8.23828340e-01 2.07653329e-01 -3.66851538e-01
-1.45103407e+00 -7.41675735e-01 -4.87694114e-01 6.39511168e-01
9.47476327e-02 -2.62934476e-01 -8.73341262e-01 1.49841443e-01
-1.54067622e-02 1.62989482e-01 -5.96379697e-01 -2.59045869e-01
-6.46144688e-01 -8.42040837e-01 8.41983378e-01 3.76763135e-01
4.71825600e-01 -1.16208291e+00 -8.63358200e-01 1.34421751e-01
-2.17247695e-01 -8.36352766e-01 -4.91702497e-01 1.75168812e-01
-1.16556478e+00 -8.07950735e-01 -1.16284561e+00 -9.34067011e-01
6.94047511e-01 2.63070941e-01 8.20443809e-01 3.22745591e-01
-8.15408826e-01 2.03051656e-01 -3.28705192e-01 -4.17765200e-01
-3.34641337e-01 -9.76327211e-02 6.07972406e-02 -3.29360008e-01
6.59620389e-02 -9.17035043e-01 -1.17788279e+00 -1.93556154e-03
-1.28181815e+00 3.52542400e-01 8.91677380e-01 9.49462652e-01
8.95545781e-01 -2.65054256e-02 8.39117825e-01 -5.63716888e-01
9.27091837e-01 -8.69166970e-01 -2.33746171e-02 1.02912039e-01
-7.81701148e-01 -1.78575948e-01 7.94703960e-01 -5.85182905e-01
-4.71295416e-01 -4.20701355e-01 -5.92125431e-02 -1.12205696e+00
7.76010603e-02 9.51469600e-01 2.46306390e-01 6.46483600e-02
7.84791708e-01 7.21416116e-01 4.65221405e-01 -4.38817233e-01
5.80503754e-02 9.26861644e-01 6.63163543e-01 -5.19793928e-01
1.94280222e-01 3.72538716e-01 7.27054849e-02 -9.84509170e-01
-2.30650514e-01 -3.83727878e-01 -1.02747954e-01 -2.72259027e-01
5.82470000e-01 -7.96165109e-01 -5.04947782e-01 3.04010361e-01
-1.00027776e+00 -9.57948864e-02 -3.56754005e-01 4.24499571e-01
-6.64263904e-01 4.74561572e-01 -9.41632986e-01 -7.69965708e-01
-1.02675200e+00 -9.49591756e-01 5.70568860e-01 1.14295006e-01
4.02401723e-02 -6.02226794e-01 -6.20821491e-02 3.38373899e-01
6.50931776e-01 4.29652095e-01 1.11008441e+00 -4.51786101e-01
-4.92388129e-01 -4.47241180e-02 -4.17668134e-01 4.28529531e-01
2.86284722e-02 -2.82760113e-01 -5.74485600e-01 -5.46485364e-01
2.21259221e-01 -4.14849222e-01 7.40419865e-01 7.24457562e-01
1.73325217e+00 -7.52952456e-01 -3.98295075e-01 1.00461709e+00
1.62635553e+00 3.94870162e-01 7.55124271e-01 7.08795488e-02
4.57238615e-01 4.53529388e-01 1.68238327e-01 7.22065568e-01
-4.78790067e-02 3.69500577e-01 4.99011338e-01 -2.92033888e-03
-2.48106956e-01 7.68168643e-02 -7.98904002e-02 1.49157941e+00
-4.10273373e-01 -3.17649394e-01 -6.08443975e-01 5.46737015e-01
-1.70273983e+00 -1.02933884e+00 2.92720824e-01 2.03443193e+00
1.14829504e+00 -2.02246353e-01 6.39987662e-02 3.43762040e-01
7.15273261e-01 -6.27136827e-02 -8.92528057e-01 -2.84482419e-01
-1.93007931e-01 2.19989836e-01 5.03033638e-01 -7.75550306e-02
-6.71268284e-01 7.69939572e-02 5.85219479e+00 1.45513237e+00
-1.19870543e+00 1.41487777e-01 9.46084678e-01 -4.45325792e-01
-3.78497720e-01 -6.88140750e-01 -3.69074464e-01 5.14403880e-01
1.14500964e+00 -4.22781080e-01 5.93209147e-01 6.66088998e-01
3.35335016e-01 4.09473687e-01 -6.95683300e-01 1.49931049e+00
1.54631987e-01 -1.76741242e+00 4.85378116e-01 1.63831130e-01
4.40552503e-01 -4.19534631e-02 2.50512719e-01 1.95798948e-02
-5.14615119e-01 -1.29552495e+00 6.40109479e-01 6.35738134e-01
1.19178605e+00 -8.22988927e-01 4.04728740e-01 4.06712055e-01
-1.09903169e+00 -3.25824142e-01 -7.24903345e-01 3.97275567e-01
1.42545938e-01 5.35472631e-01 -4.05955613e-01 5.70846379e-01
6.89033449e-01 8.06773245e-01 -1.69926450e-01 1.18368793e+00
7.37277687e-01 5.09871960e-01 -8.85937139e-02 -3.44649628e-02
1.62849203e-02 -1.82528257e-01 6.77669048e-01 1.52159631e+00
6.61695898e-01 5.40793419e-01 -1.14544414e-01 7.67596543e-01
-2.12955102e-01 3.02108079e-01 -5.75083375e-01 -2.86794305e-01
6.80542946e-01 1.03406286e+00 -5.61878204e-01 -2.93074220e-01
-1.60301119e-01 7.39689648e-01 3.80193710e-01 3.25415939e-01
-7.25275517e-01 -3.53536606e-01 2.02468008e-01 2.64302850e-01
1.55972287e-01 -6.53368235e-02 -2.80412525e-01 -7.64214456e-01
1.95498355e-02 -1.06537998e+00 5.73428631e-01 -8.61649811e-01
-1.16868460e+00 7.12274373e-01 2.15310365e-01 -1.51103044e+00
1.17631666e-01 -5.86391985e-01 -1.18817806e-01 6.44691110e-01
-1.34264386e+00 -9.54542398e-01 -3.98926973e-01 8.44840348e-01
7.14334428e-01 -4.75414634e-01 7.28686988e-01 6.45934522e-01
-4.62721437e-01 7.38568246e-01 3.43857467e-01 -1.39164448e-01
2.42443785e-01 -5.97130060e-01 -2.67593503e-01 3.11393797e-01
-4.04066235e-01 8.38488996e-01 6.93488002e-01 -6.05456233e-01
-1.80900061e+00 -1.07909894e+00 1.79764554e-01 4.73026633e-01
3.21703672e-01 1.23695713e-02 -1.17681324e+00 2.31934756e-01
3.26169729e-01 1.10592164e-01 7.98643649e-01 -6.29945099e-01
-8.56475234e-02 -1.24619760e-01 -1.30553818e+00 5.12917757e-01
1.09394872e+00 -2.76167572e-01 -2.05533028e-01 4.87884521e-01
7.50398159e-01 -5.04694164e-01 -1.10394943e+00 6.09125435e-01
7.10245669e-01 -7.83522248e-01 1.08939004e+00 -3.33761662e-01
7.93764710e-01 -5.48615418e-02 -4.47043896e-01 -7.73605883e-01
-5.50992250e-01 -6.00971222e-01 -6.24476492e-01 6.28033757e-01
1.02910824e-01 -4.10747379e-01 8.35391700e-01 2.93849528e-01
-4.12668467e-01 -1.62182283e+00 -1.10934782e+00 -6.09211981e-01
-9.21891108e-02 -1.76637948e-01 2.82403350e-01 9.76080298e-01
1.21387847e-01 -1.04524247e-01 -7.10356534e-01 -1.25271469e-01
7.10238338e-01 -1.73748881e-01 1.56383708e-01 -8.07601452e-01
-3.54934305e-01 -5.23370683e-01 -6.48729801e-01 -1.00131702e+00
-3.65398914e-01 -1.03960967e+00 -2.37453490e-01 -1.45916450e+00
4.88285780e-01 -6.24333203e-01 -4.61818606e-01 3.36939931e-01
2.36213803e-01 3.95848155e-01 2.26186156e-01 7.32543707e-01
-2.14525089e-01 6.84003890e-01 1.51015055e+00 -4.01425302e-01
-6.97510466e-02 -3.71619254e-01 -7.12049901e-01 2.71465510e-01
8.85102630e-01 -2.71818727e-01 -7.65667140e-01 -3.49554896e-01
-7.42621049e-02 4.92635667e-01 1.91167966e-01 -1.21098614e+00
4.50449646e-01 -1.20589547e-01 5.05084395e-01 -5.79921782e-01
4.90880489e-01 -8.12260687e-01 5.98831713e-01 7.39657879e-01
-4.55033213e-01 1.99882671e-01 1.49174273e-01 7.11149514e-01
-2.52782136e-01 -4.09459203e-01 8.55124950e-01 -2.54608989e-01
-3.53470206e-01 8.55213106e-01 -1.60475999e-01 2.26976862e-03
8.71999443e-01 -6.82004511e-01 -3.43921989e-01 -2.91021943e-01
-7.53582299e-01 -9.60728675e-02 1.32509083e-01 1.18509762e-01
1.28989935e+00 -1.51044273e+00 -9.06535625e-01 2.61330098e-01
-2.03106806e-01 1.79729443e-02 6.77971601e-01 8.63574266e-01
-8.07668507e-01 3.21524382e-01 -4.55931246e-01 -6.21220469e-01
-1.27100396e+00 6.52853489e-01 2.66971201e-01 -9.37801227e-02
-9.61769640e-01 5.59888661e-01 1.48515776e-01 1.02266975e-01
3.80994231e-01 -1.35555267e-01 -3.82815033e-01 -1.82171345e-01
6.43631816e-01 5.09501636e-01 -8.35916623e-02 -5.74791551e-01
-7.81443641e-02 5.10172546e-01 3.37256603e-02 2.73756504e-01
1.54493725e+00 -5.81313819e-02 -1.16535239e-01 1.27913296e-01
1.57971382e+00 -4.56318706e-01 -9.89604831e-01 -9.46967378e-02
-6.97689593e-01 -3.94876510e-01 1.50671750e-02 -4.20458019e-01
-1.44481945e+00 7.33912289e-01 8.39122415e-01 2.72293657e-01
1.61760163e+00 -1.86489329e-01 9.63470817e-01 1.01444185e-01
2.24072203e-01 -1.08655310e+00 1.44564837e-01 -2.22376976e-02
1.26370955e+00 -6.63671076e-01 3.61821264e-01 -2.46097535e-01
-5.32201469e-01 1.22803748e+00 2.69229621e-01 -1.39834732e-01
6.92998350e-01 4.25747514e-01 -2.98615754e-01 -3.87136400e-01
-7.65906274e-01 5.69682837e-01 3.11482400e-01 7.79364228e-01
3.39509249e-01 1.52576044e-01 -4.52870280e-01 5.11080921e-01
-1.03637658e-01 3.50927740e-01 2.24667683e-01 9.89746332e-01
-5.77720046e-01 -6.17161930e-01 -2.51694679e-01 8.28775942e-01
-5.69677711e-01 -2.02740803e-01 4.35238153e-01 5.93903065e-01
2.73565780e-02 3.86282504e-01 -2.14026477e-02 -2.17774361e-01
8.66454542e-02 -3.17240626e-01 4.95765924e-01 -2.53206994e-02
-2.57062078e-01 1.43622398e-01 -2.33572304e-01 -3.23794425e-01
-4.85163718e-01 -4.36052233e-01 -1.18313158e+00 -4.02446061e-01
3.82138528e-02 1.00432724e-01 6.15384877e-01 5.68592072e-01
5.66934824e-01 6.36544645e-01 3.48943919e-01 -9.44725931e-01
-6.53974533e-01 -8.01170707e-01 -5.65752268e-01 3.80792588e-01
4.71039206e-01 -3.94211948e-01 -1.42890841e-01 2.80029297e-01] | [11.316238403320312, -1.598046898841858] |
540b3758-7be2-4f30-b65c-ef183965699c | hindiwsd-a-package-for-word-sense | null | null | https://aclanthology.org/2022.wildre-1.4 | https://aclanthology.org/2022.wildre-1.4.pdf | HindiWSD: A package for word sense disambiguation in Hinglish & Hindi | A lot of commendable work has been done, especially in high resource languages such as English, Spanish, French, etc. However, work done for Indic languages such as Hindi, Tamil, Telugu, etc is relatively less due to difficulty in finding relevant datasets, and the complexity of these languages. With the advent of IndoWordnet, we can explore important tasks such as word sense disambiguation, word similarity, and cross-lingual information retrieval, and carry out effective research regarding the same. In this paper, we worked on improving word sense disambiguation for 20 of the most common ambiguous Hindi words by making use of knowledge-based methods. We also came up with “hindiwsd”, an easy-to-use framework developed in Python that acts as a pipeline for transliteration of Hinglish code-mixed text followed by spell correction, POS tagging, and word sense disambiguation of Hindi text. We also curated a dataset of these 20 most used ambiguous Hindi words. This dataset was then used to enhance a modified Lesk’s algorithm and more accurately carry out word sense disambiguation. We achieved an accuracy of about 71% using our customized Lesk’s algorithm which was an improvement to the accuracy of about 34% using the original Lesk’s algorithm on the test set. | ['Chethan Sharma', 'Praatibh Surana', 'Mirza Yusuf'] | null | null | null | null | wildre-lrec-2022-6 | ['word-sense-disambiguation', 'word-similarity', 'transliteration', 'cross-lingual-information-retrieval'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.72957197e-01 -1.95300817e-01 3.09297830e-01 -2.28804931e-01
-8.15973163e-01 -9.78842497e-01 4.33658123e-01 5.38825512e-01
-8.10465634e-01 1.11403561e+00 2.77894109e-01 -6.42845869e-01
-2.49193415e-01 -6.29240513e-01 -1.63932562e-01 -3.77423823e-01
2.61082053e-01 5.59901178e-01 5.49965620e-01 -8.06851566e-01
6.74222887e-01 1.25746444e-01 -1.29936218e+00 -2.85611842e-02
1.17709851e+00 -7.94225261e-02 7.16507375e-01 3.68933350e-01
-3.76895577e-01 4.46402371e-01 -6.55537128e-01 -4.06835735e-01
-3.59175466e-02 -3.17822397e-01 -1.33054924e+00 -5.47939956e-01
1.44339457e-01 3.94645780e-01 5.52723221e-02 1.06850874e+00
8.53194594e-01 1.64789170e-01 2.35377833e-01 -7.80739486e-01
-4.64547753e-01 1.09094441e+00 -3.18016022e-01 3.59912425e-01
6.80719256e-01 -1.94171265e-01 9.96746898e-01 -8.16284180e-01
1.09360373e+00 1.05764353e+00 6.24181747e-01 3.17465007e-01
-5.89092493e-01 -7.40784824e-01 -3.91687244e-01 5.40471852e-01
-1.57598925e+00 1.61321964e-02 3.31312418e-01 -3.43184888e-01
1.30254662e+00 1.40492305e-01 3.81344914e-01 6.55827999e-01
9.83099565e-02 7.84938514e-01 1.13438261e+00 -9.77490425e-01
-9.14522409e-02 1.37866676e-01 4.49264526e-01 5.46642482e-01
2.19545200e-01 -4.78671670e-01 -3.05550158e-01 1.40391737e-02
1.48526907e-01 -2.12290615e-01 -4.96622035e-03 1.25132099e-01
-1.46415913e+00 9.16476488e-01 6.98411092e-03 1.12252998e+00
-5.87888025e-02 -5.62233150e-01 6.83650017e-01 4.55474198e-01
1.73617885e-01 1.02674878e+00 -7.26764381e-01 -4.92960602e-01
-8.98110986e-01 3.61985385e-01 8.99875879e-01 9.72940624e-01
9.35528040e-01 -3.76051873e-01 8.77304450e-02 1.25708330e+00
3.57860029e-01 5.34748793e-01 9.71837640e-01 -4.38105613e-01
5.91100216e-01 7.75369704e-01 -1.00063644e-01 -8.70143414e-01
-4.61745530e-01 -1.11092426e-01 -3.21778595e-01 -1.33966014e-01
3.50585431e-01 -3.16047877e-01 -8.65115464e-01 1.41239083e+00
2.33964711e-01 -3.48023415e-01 2.07816675e-01 6.80181324e-01
6.54931188e-01 5.84251642e-01 9.66917202e-02 -1.43283959e-02
1.43455517e+00 -7.01543987e-01 -5.72497725e-01 -3.21592182e-01
8.76099110e-01 -1.61829257e+00 1.13027906e+00 1.09466150e-01
-7.45566368e-01 -2.72582382e-01 -1.22450912e+00 1.48158576e-02
-1.00337756e+00 -2.53009111e-01 3.85860413e-01 6.81308687e-01
-1.15807462e+00 5.31221747e-01 -4.91772652e-01 -1.08651888e+00
-1.81970417e-01 2.60931462e-01 -4.73761380e-01 -2.57145852e-01
-1.58531892e+00 1.27597010e+00 8.32596958e-01 -4.03260082e-01
-3.07705790e-01 -5.47483563e-01 -9.93890464e-01 -4.08142030e-01
2.68158078e-01 -2.06124514e-01 9.58019972e-01 -4.49921817e-01
-9.93329048e-01 1.28509331e+00 1.56732947e-02 -2.40683034e-01
-1.47032961e-02 -1.48393199e-01 -7.43077099e-01 -4.34814245e-01
6.24312937e-01 2.46493772e-01 1.42650753e-01 -7.40448475e-01
-9.02098238e-01 -5.13215661e-01 4.26304042e-02 1.80903390e-01
-1.13369197e-01 6.68244302e-01 -3.87895912e-01 -7.60674298e-01
2.44263858e-02 -1.27382946e+00 -5.73148206e-02 -8.79328549e-01
-4.79930453e-02 -3.84144157e-01 7.49946773e-01 -1.08046842e+00
1.75768912e+00 -2.16241503e+00 -1.11347519e-01 3.04010421e-01
-5.03023230e-02 7.23677754e-01 -1.34553298e-01 8.74487936e-01
-1.92527995e-01 1.76682472e-01 -3.78526300e-01 9.91206169e-02
-9.15338397e-02 5.55606365e-01 1.78055227e-01 1.08645946e-01
7.95727447e-02 5.25553167e-01 -1.27809989e+00 -6.32928193e-01
3.35058361e-01 2.50176519e-01 -2.73752570e-01 8.72594044e-02
2.11896449e-01 1.87104806e-01 -2.08002746e-01 5.78595161e-01
4.50249702e-01 4.03256536e-01 4.19930756e-01 3.12450677e-01
-6.60577118e-01 5.76952338e-01 -1.31073785e+00 1.87141907e+00
-9.57801223e-01 6.77706301e-01 -2.26733863e-01 -7.72400379e-01
1.02198660e+00 2.96353251e-01 2.93251336e-01 -6.67082787e-01
3.18488628e-02 7.37924814e-01 2.39914671e-01 -6.30206525e-01
8.38312566e-01 1.86254039e-01 -4.12873089e-01 3.78894925e-01
8.91346782e-02 -2.92149872e-01 7.20885336e-01 2.82493830e-01
9.31476057e-01 1.19752958e-01 8.79397094e-01 -8.40147972e-01
7.81530023e-01 3.47211629e-01 2.23060250e-01 5.03938138e-01
-1.08686373e-01 6.12833977e-01 -6.46292195e-02 -3.59304667e-01
-9.74437654e-01 -8.19446802e-01 -3.08404684e-01 1.30683565e+00
2.36293487e-02 -8.17117095e-01 -7.44734228e-01 -6.18469894e-01
-3.80958259e-01 8.26118112e-01 -2.79100358e-01 1.46885976e-01
-8.78605425e-01 -8.74372005e-01 7.08866239e-01 8.82270560e-02
5.42146564e-01 -1.20464039e+00 -3.68344992e-01 4.87988651e-01
-3.19823623e-01 -7.61975765e-01 -5.27881801e-01 3.09542388e-01
-2.24890813e-01 -9.63790476e-01 -5.60404539e-01 -1.17560995e+00
1.34812251e-01 1.47772402e-01 1.22554839e+00 2.26203710e-01
-4.12426770e-01 -1.88004911e-01 -7.96957552e-01 -5.57237267e-01
-7.03138709e-01 6.46309197e-01 -1.36910781e-01 -7.69647837e-01
6.27967000e-01 -3.34184498e-01 -2.04023257e-01 3.51479322e-01
-1.14325511e+00 -5.09056091e-01 3.32927018e-01 7.37892151e-01
6.61116764e-02 -2.06151918e-01 3.86898011e-01 -1.08311105e+00
6.96228266e-01 -5.56190550e-01 -3.31982583e-01 4.25169766e-01
-7.08486140e-01 5.23148000e-01 5.07357717e-01 -1.39202535e-01
-1.03026462e+00 -5.38664907e-02 -8.08554530e-01 7.34910429e-01
-1.00935623e-01 6.78488255e-01 -2.12789118e-01 -9.45013687e-02
7.55805731e-01 1.49292471e-02 -2.94839382e-01 -5.43008566e-01
2.96285063e-01 1.29383445e+00 2.44693458e-01 -6.57126665e-01
5.18089235e-01 -2.75099754e-01 -4.56845552e-01 -9.39555526e-01
-6.69230282e-01 -9.68542516e-01 -9.05618548e-01 1.70038164e-01
9.57108378e-01 -5.56755066e-01 -1.22825928e-01 5.72304308e-01
-1.06684577e+00 -3.74026000e-02 2.00367123e-01 3.26369613e-01
-4.23201621e-02 8.17196548e-01 -3.71686667e-01 -3.51280451e-01
-4.45109218e-01 -1.15342987e+00 9.83309805e-01 1.91067040e-01
-7.86712885e-01 -1.27300310e+00 4.49403822e-01 2.79807270e-01
6.09345913e-01 -1.61761522e-01 1.25963974e+00 -9.96543825e-01
6.63822442e-02 5.16282581e-03 -8.01797062e-02 2.70071656e-01
1.45964786e-01 2.66783917e-03 -3.81604582e-01 -1.70696855e-01
-4.20646280e-01 -9.40983891e-02 2.88854599e-01 -2.24652603e-01
3.56830627e-01 -9.19132009e-02 -1.33069754e-01 1.83377162e-01
1.59243894e+00 4.00685072e-01 7.44436622e-01 9.85504031e-01
5.73620617e-01 4.05195415e-01 9.45654929e-01 4.00411129e-01
6.52767777e-01 7.75358915e-01 -8.09788890e-03 1.53525472e-01
-1.71711937e-01 2.79510338e-02 2.17018098e-01 1.31617522e+00
2.76131295e-02 6.58907369e-02 -1.56677735e+00 1.08937812e+00
-1.67374110e+00 -5.93256116e-01 -3.51937681e-01 2.19287801e+00
1.31765497e+00 -1.11144640e-01 2.88840011e-02 1.98515385e-01
6.58252656e-01 -5.14278300e-02 3.14563692e-01 -7.78036296e-01
-1.86453298e-01 6.37617290e-01 6.01679981e-01 9.03132260e-01
-1.04240644e+00 1.52331531e+00 5.58575487e+00 9.94516253e-01
-1.21048939e+00 2.00829655e-01 -8.58843625e-02 5.39613485e-01
-1.69205800e-01 1.53546289e-01 -9.91863549e-01 5.94638348e-01
8.58182907e-01 -4.32176203e-01 5.31479836e-01 7.91791975e-01
-1.35712296e-01 -2.51005918e-01 -5.20242512e-01 9.92610037e-01
2.18894348e-01 -9.11365569e-01 -2.89969146e-01 -4.34175164e-01
8.73140752e-01 5.12055993e-01 -7.11163759e-01 5.61493456e-01
6.41488075e-01 -7.99191594e-01 4.56097484e-01 -2.90623517e-04
4.52026129e-01 -8.86228085e-01 1.16306138e+00 1.92360625e-01
-1.10054457e+00 2.20636830e-01 -4.42842543e-01 -1.67955503e-01
-1.80885926e-01 4.19069618e-01 -1.16857886e+00 6.46819651e-01
6.59365416e-01 6.21390581e-01 -8.94895732e-01 1.00446010e+00
-4.27729070e-01 6.73478484e-01 -3.68419558e-01 -4.73699331e-01
5.03818393e-01 5.30908443e-02 5.35524189e-01 1.70836627e+00
4.38467741e-01 -1.11478075e-01 3.38571757e-01 1.16551749e-03
3.51905972e-01 6.70251489e-01 -4.58661437e-01 1.65469304e-01
6.06665909e-01 1.28228486e+00 -8.42157364e-01 -4.63677198e-01
-3.61036986e-01 1.13805234e+00 2.03642175e-01 1.47974724e-02
-6.11814499e-01 -1.06612086e+00 6.86622679e-01 4.20110635e-02
5.24872765e-02 -4.46091384e-01 -2.11135104e-01 -1.11813903e+00
-6.63193837e-02 -1.04679942e+00 7.21295953e-01 -6.72210336e-01
-1.07785916e+00 8.35170507e-01 5.06195650e-02 -8.81619275e-01
-4.76992607e-01 -7.68963516e-01 -3.55096936e-01 9.94748652e-01
-1.29941952e+00 -1.12023056e+00 -6.95937574e-02 6.37669504e-01
6.56898022e-01 -1.90230533e-01 9.16365743e-01 6.16469800e-01
-1.66604847e-01 5.47832429e-01 3.49327445e-01 2.17115626e-01
1.11041725e+00 -1.41208959e+00 3.36568296e-01 9.40192699e-01
1.29114389e-01 1.00252628e+00 8.45831037e-01 -7.59100497e-01
-1.02836084e+00 -8.63087177e-01 1.70822346e+00 -5.36981165e-01
1.21458077e+00 -2.02086866e-01 -8.76061976e-01 4.43977386e-01
5.59044778e-01 -6.79843307e-01 6.88960612e-01 4.61880527e-02
-9.68986228e-02 2.73727663e-02 -9.83011663e-01 6.51184916e-01
1.02880073e+00 -4.46167469e-01 -1.02462327e+00 3.40228766e-01
5.34981549e-01 -2.98845202e-01 -8.90265226e-01 2.34487001e-02
3.70105594e-01 -4.42285001e-01 7.59936452e-01 -4.66893494e-01
2.64113039e-01 -4.82728392e-01 -3.05316120e-01 -1.72594631e+00
-1.92443117e-01 -6.36504650e-01 1.00055897e+00 1.53809047e+00
8.02732110e-01 -5.77725112e-01 7.25717144e-03 3.09105217e-01
-2.45687798e-01 -5.50365224e-02 -7.19253719e-01 -7.11401403e-01
2.96070635e-01 -4.33891594e-01 6.00503683e-01 1.28126609e+00
5.16702592e-01 6.77100122e-01 -1.96960762e-01 -1.87190771e-01
9.50865671e-02 3.62455137e-02 4.31302935e-01 -1.05924809e+00
-4.59987670e-03 -1.87230572e-01 -7.34908462e-01 -3.35458606e-01
2.31569871e-01 -1.24552810e+00 2.65240401e-01 -1.59511971e+00
4.33408655e-02 -3.80999267e-01 -8.88001770e-02 7.81483471e-01
-5.12728453e-01 4.09951389e-01 2.97597349e-01 6.22599125e-02
-6.24163151e-01 -1.25170648e-01 8.34465146e-01 1.12997010e-01
-6.67286664e-02 -3.36176842e-01 -6.99709475e-01 5.53330421e-01
8.06042373e-01 -6.71783090e-01 -4.94659506e-03 -6.11648440e-01
5.88925064e-01 -3.03396910e-01 -3.26544613e-01 -1.02866971e+00
1.01080045e-01 -1.02550946e-01 -1.77150771e-01 -3.99541289e-01
-4.07260090e-01 -7.33039916e-01 -8.09354782e-02 4.56952482e-01
-7.00290650e-02 6.04099154e-01 2.30370417e-01 -4.42747682e-01
-3.69479686e-01 -7.12128043e-01 8.03614140e-01 -4.45120186e-01
-1.39715588e+00 -2.71615952e-01 -7.49056816e-01 4.70958322e-01
1.05836773e+00 1.87638924e-02 -2.04032585e-02 3.47635597e-02
-4.66731697e-01 1.68968216e-01 6.05343282e-01 7.23818481e-01
8.77238438e-02 -9.49286699e-01 -8.91265213e-01 1.49155587e-01
4.21657294e-01 -5.52819371e-01 -1.79025218e-01 5.36025822e-01
-9.34929788e-01 4.62473452e-01 -4.33812618e-01 -2.33322471e-01
-1.51605177e+00 3.87987852e-01 -1.50328681e-01 -4.04060632e-01
-1.95645839e-01 5.23660302e-01 -4.66783643e-01 -9.24334228e-01
-1.95478097e-01 -1.22864693e-01 -4.96920973e-01 1.61000535e-01
5.15470445e-01 3.52290064e-01 4.56199884e-01 -8.57374609e-01
-9.38110113e-01 6.37046456e-01 2.00384893e-02 -1.03467658e-01
1.24284685e+00 -4.14607674e-01 -6.48200274e-01 2.68706053e-01
1.19287193e+00 4.94448811e-01 -7.53742596e-03 -2.61212848e-02
6.14585698e-01 -4.04000878e-01 -4.31534290e-01 -1.07576966e+00
-4.96497005e-01 6.63069427e-01 5.01086354e-01 7.86945522e-02
8.98918152e-01 4.29060273e-02 9.23531234e-01 5.05347848e-01
5.70545316e-01 -1.33578396e+00 -6.39976025e-01 1.40366459e+00
4.64350939e-01 -1.31688714e+00 -3.66437316e-01 -2.23805249e-01
-5.94418108e-01 1.11492932e+00 3.21061134e-01 7.18205050e-02
6.10415995e-01 3.52094203e-01 5.03864825e-01 1.01324227e-02
-1.62495553e-01 -4.95818526e-01 3.42660636e-01 6.61787987e-01
1.04746723e+00 9.94054526e-02 -1.18855107e+00 4.75909770e-01
-6.74924791e-01 -9.62389447e-03 5.63439012e-01 1.02935517e+00
-5.58764994e-01 -1.56013525e+00 -3.91672760e-01 1.04533866e-01
-7.66003728e-01 -6.93113863e-01 -4.08786446e-01 1.01309955e+00
3.36999923e-01 1.03354335e+00 -3.27340037e-01 -3.09978843e-01
2.59632558e-01 2.13125646e-01 1.84915319e-01 -8.93537819e-01
-8.88473749e-01 -3.67398828e-01 3.41635019e-01 -2.16051072e-01
-3.03900629e-01 -6.98137462e-01 -1.36675823e+00 -3.31881553e-01
-1.30883515e-01 4.96970147e-01 8.26230645e-01 1.33512127e+00
-2.97009386e-02 1.22978352e-01 3.21565568e-01 -2.79684484e-01
-2.19884664e-01 -1.15170944e+00 -4.31190699e-01 7.10776508e-01
-2.74560094e-01 -3.03823650e-01 -1.50890350e-02 4.67185257e-03] | [10.449542045593262, 9.630961418151855] |
2842920e-89c6-4f05-befa-f95e7b5d619c | generating-video-description-using-sequence | null | null | https://aclanthology.org/C16-1005 | https://aclanthology.org/C16-1005.pdf | Generating Video Description using Sequence-to-sequence Model with Temporal Attention | Automatic video description generation has recently been getting attention after rapid advancement in image caption generation. Automatically generating description for a video is more challenging than for an image due to its temporal dynamics of frames. Most of the work relied on Recurrent Neural Network (RNN) and recently attentional mechanisms have also been applied to make the model learn to focus on some frames of the video while generating each word in a describing sentence. In this paper, we focus on a sequence-to-sequence approach with temporal attention mechanism. We analyze and compare the results from different attention model configuration. By applying the temporal attention mechanism to the system, we can achieve a METEOR score of 0.310 on Microsoft Video Description dataset, which outperformed the state-of-the-art system so far. | ['Sang Phan', 'Raphael Shu', 'Yusuke Miyao', 'Yo Ehara', 'Noriki Nishida', 'Naoaki Okazaki', 'Natsuda Laokulrat', 'Hideki Nakayama'] | 2016-12-01 | generating-video-description-using-sequence-1 | https://aclanthology.org/C16-1005 | https://aclanthology.org/C16-1005.pdf | coling-2016-12 | ['video-description'] | ['computer-vision'] | [ 5.02198696e-01 -8.80686566e-02 -2.89604384e-02 -2.50681132e-01
-6.74230993e-01 -2.86982715e-01 8.25547636e-01 -3.12087744e-01
-2.69210517e-01 9.22131181e-01 6.98439658e-01 1.02904342e-01
3.77027422e-01 -2.33760729e-01 -6.28655434e-01 -5.02736807e-01
8.54991972e-02 1.61559150e-01 1.89381883e-01 -1.37683272e-01
5.48115432e-01 6.34421483e-02 -1.52280438e+00 5.73015332e-01
1.21896386e-01 5.99177957e-01 7.64897943e-01 1.09908998e+00
-2.39340171e-01 1.30165994e+00 -8.59127522e-01 -2.49028206e-01
-9.07756686e-02 -8.90796542e-01 -8.16248357e-01 2.72315681e-01
3.26091856e-01 -2.02130824e-01 -6.28738046e-01 9.09875751e-01
6.74160302e-01 4.42074269e-01 6.05551124e-01 -1.08578110e+00
-1.12597406e+00 5.84794998e-01 -5.25681436e-01 6.00566983e-01
8.10842872e-01 6.09009676e-02 7.21346438e-01 -8.26980472e-01
9.07503784e-01 9.27347898e-01 1.75062716e-01 1.15079904e+00
-8.20793033e-01 -5.68636954e-01 1.41229481e-01 7.15226173e-01
-1.44101608e+00 -4.21220332e-01 7.16762245e-01 -4.07791793e-01
1.33276677e+00 6.68772161e-02 6.16662264e-01 1.38961613e+00
3.55889112e-01 7.45274007e-01 5.27329206e-01 -4.68887120e-01
-8.92222449e-02 1.09663755e-02 -2.30389863e-01 3.30022722e-01
-2.25777507e-01 -1.72931582e-01 -5.91479778e-01 3.62862766e-01
8.80678236e-01 -1.25945508e-01 -3.92664194e-01 2.78597891e-01
-1.27890885e+00 7.46772349e-01 8.44854787e-02 3.67460728e-01
-8.22946608e-01 5.82250476e-01 7.10907638e-01 -1.06603168e-02
2.79859334e-01 1.95696831e-01 -5.21763228e-02 -6.16127670e-01
-1.08260429e+00 3.81537288e-01 4.72516179e-01 1.27877831e+00
1.24955639e-01 3.88746589e-01 -6.66086435e-01 5.70227027e-01
4.35643084e-02 1.27048567e-01 8.26455951e-01 -8.47366452e-01
4.65647757e-01 8.45946744e-02 1.48860395e-01 -7.55595505e-01
2.78681517e-04 -1.31763145e-01 -7.83273280e-01 -1.49955854e-01
-2.92510949e-02 -2.04014480e-01 -1.20182335e+00 1.59064209e+00
-3.04603249e-01 4.35429096e-01 1.98520616e-01 1.08904660e+00
9.21983302e-01 1.20534587e+00 2.09516227e-01 -4.84979779e-01
1.36115611e+00 -1.26838863e+00 -1.20003974e+00 2.29586642e-02
2.00588942e-01 -8.54083121e-01 7.18514800e-01 1.79064587e-01
-1.33892179e+00 -7.80158579e-01 -9.71645057e-01 -2.45280117e-02
-4.17514555e-02 -9.51400623e-02 3.20456445e-01 9.41849500e-02
-1.29855740e+00 4.27030325e-01 -5.10101795e-01 -5.80153644e-01
1.49583593e-01 2.34363332e-01 -4.06783074e-01 1.40740871e-02
-1.24961114e+00 9.49954689e-01 6.46970093e-01 -1.52461842e-01
-1.19824898e+00 -2.98937678e-01 -7.52427518e-01 7.21102878e-02
2.06809893e-01 -8.05363238e-01 1.45550406e+00 -1.35024023e+00
-1.54338634e+00 6.12873077e-01 -3.86575669e-01 -8.32882881e-01
2.20474005e-01 -2.76003361e-01 -4.61923987e-01 4.21543866e-01
5.56526743e-02 1.29955840e+00 9.60152447e-01 -9.86959398e-01
-6.64150417e-01 3.75864148e-01 2.07401767e-01 2.34479323e-01
2.36552563e-02 4.70456451e-01 -5.22533119e-01 -7.87741661e-01
-4.29382712e-01 -1.06332958e+00 -1.49712205e-01 -4.65289086e-01
-1.21062987e-01 -3.70015472e-01 9.25053835e-01 -7.09467649e-01
1.54857183e+00 -1.93204069e+00 1.97445080e-01 -6.88359022e-01
-3.03514063e-01 4.20526743e-01 -8.68233740e-02 6.02900505e-01
-5.35751820e-01 3.28333795e-01 -1.67810917e-02 -3.74872327e-01
-3.60450417e-01 1.43432081e-01 -4.44904923e-01 1.19570464e-01
4.12098408e-01 7.74208546e-01 -8.38188946e-01 -5.29710531e-01
5.32937124e-02 7.81564832e-01 -5.83083212e-01 4.98510510e-01
-3.78120065e-01 3.13961506e-01 -2.47569665e-01 3.60806644e-01
2.17670470e-01 -2.41440549e-01 -1.78784341e-01 -1.00828074e-01
-2.94104397e-01 1.03075422e-01 -5.25185227e-01 2.02492905e+00
-3.30778807e-01 1.12663138e+00 -5.99167705e-01 -7.79592931e-01
7.49178588e-01 1.08526421e+00 3.62615198e-01 -4.80186015e-01
1.64244071e-01 -1.14540733e-01 2.72496969e-01 -9.38427866e-01
7.95749187e-01 -2.28765145e-01 5.35706729e-02 4.68202055e-01
1.89978525e-01 3.01392764e-01 5.90028524e-01 3.25782865e-01
1.00445390e+00 6.73326433e-01 3.57712030e-01 1.49908006e-01
7.28144228e-01 5.64101301e-02 2.70922035e-01 7.68388391e-01
-3.43907684e-01 1.23460996e+00 3.43802124e-01 -5.09032547e-01
-1.59971416e+00 -5.75582087e-01 4.49206531e-01 9.16183174e-01
-9.29679126e-02 -4.37425554e-01 -9.04844224e-01 -3.94925743e-01
-6.96378767e-01 8.32040668e-01 -6.69669032e-01 -1.00052662e-01
-7.40631998e-01 -3.48483294e-01 4.27713066e-01 5.88622391e-01
6.56992197e-01 -1.69757843e+00 -8.59137118e-01 4.44146663e-01
-4.06854898e-01 -1.31035674e+00 -8.30226898e-01 -2.97877729e-01
-4.78172213e-01 -5.02316594e-01 -1.24470305e+00 -8.33534658e-01
5.09310782e-01 2.29334757e-01 1.14421535e+00 -7.30589777e-02
-2.69815028e-01 3.48395616e-01 -6.95954919e-01 -4.35832888e-01
-4.05863196e-01 1.21299669e-01 -6.42293692e-02 3.98789868e-02
2.73797929e-01 -2.95941889e-01 -5.22774041e-01 -1.03449076e-01
-1.17503607e+00 6.13127649e-01 5.83712876e-01 7.31979549e-01
2.64618933e-01 -2.33391568e-01 7.09622264e-01 -5.13790488e-01
7.97728002e-01 -7.31633246e-01 -1.94531888e-01 1.08376279e-01
-1.92423895e-01 2.41827652e-01 4.63447571e-01 -6.30729437e-01
-1.15959382e+00 2.58473426e-01 -1.63170211e-02 -8.16677988e-01
-1.06480062e-01 4.69234854e-01 3.61819148e-01 5.01837432e-01
1.82230562e-01 6.03918254e-01 -4.67411019e-02 -1.37727810e-02
8.65881518e-02 7.04703987e-01 5.37773073e-01 -1.45720929e-01
4.75616246e-01 2.76573330e-01 -1.16477773e-01 -6.27550840e-01
-8.66427124e-01 -4.87534881e-01 -3.66259784e-01 -6.64754033e-01
1.29874170e+00 -9.48894203e-01 -6.26450956e-01 1.61587864e-01
-1.77849174e+00 -1.29937027e-02 2.59462893e-02 4.94066447e-01
-7.70401597e-01 1.32839546e-01 -4.71955627e-01 -9.21296954e-01
-4.58449841e-01 -1.08538735e+00 8.31274569e-01 3.82671773e-01
-3.53381097e-01 -7.95686305e-01 1.06180839e-01 1.90195069e-01
7.26465344e-01 2.42071882e-01 3.14623654e-01 -4.10869688e-01
-6.27576411e-01 -1.97175249e-01 -3.06821048e-01 2.30268866e-01
1.68428078e-01 7.36313090e-02 -8.06661665e-01 -5.95091172e-02
2.60068364e-02 -1.29939526e-01 8.00351739e-01 4.17750329e-01
1.10286093e+00 -4.80429620e-01 -1.94885172e-02 2.18967438e-01
1.59432530e+00 6.42982483e-01 1.00039196e+00 3.82586509e-01
6.12023354e-01 4.57314849e-01 6.64302945e-01 5.47401011e-01
1.51310965e-01 7.69785047e-01 3.92028272e-01 1.38587043e-01
-3.58940661e-01 -1.78688437e-01 6.14705265e-01 6.65674329e-01
-3.97061288e-01 -8.04527462e-01 -7.31762826e-01 7.56016195e-01
-2.12083054e+00 -1.58676457e+00 2.78095342e-02 1.83813083e+00
5.49935520e-01 8.80391747e-02 6.49976656e-02 -1.70621529e-01
1.11233413e+00 2.64354646e-01 -2.46803507e-01 -7.91172802e-01
-1.32727250e-01 -1.01640619e-01 3.10515583e-01 3.74753058e-01
-8.69124949e-01 9.80402410e-01 6.74959803e+00 6.81315064e-01
-1.24943566e+00 5.16754538e-02 5.33615649e-01 -3.11676621e-01
-5.91798779e-03 -1.00094542e-01 -7.60363281e-01 6.68685436e-01
1.43544948e+00 -5.77188253e-01 2.88862139e-01 6.98043346e-01
6.26846910e-01 -6.20068051e-02 -9.72323716e-01 1.18448770e+00
7.72397637e-01 -1.49993467e+00 6.13595247e-01 -2.00881362e-01
9.54587817e-01 -2.27436334e-01 2.64481120e-02 3.32661092e-01
-2.15581745e-01 -1.34445322e+00 7.50874341e-01 8.24028432e-01
8.88503551e-01 -8.42663407e-01 1.03747296e+00 1.80843502e-01
-1.12099981e+00 -1.60295349e-02 -3.36871266e-01 -3.10314357e-01
7.57824719e-01 -6.61488771e-02 -1.12750518e+00 2.30457500e-01
5.05015492e-01 7.65438259e-01 -2.75674760e-01 9.98793304e-01
-7.91919157e-02 4.60444808e-01 3.35464388e-01 -1.95309222e-01
6.37346268e-01 8.36606324e-03 6.85774982e-01 1.52030921e+00
5.83029032e-01 3.51188779e-01 -8.10799450e-02 7.60638893e-01
-1.01868615e-01 -7.45699927e-02 -8.72524559e-01 -3.29092592e-01
3.50684151e-02 1.09931147e+00 -5.18915951e-01 -7.15954304e-01
-5.03377914e-01 1.32907712e+00 -2.37884484e-02 2.96405017e-01
-1.31806767e+00 -3.63668650e-01 4.03715819e-01 2.96722472e-01
6.34010851e-01 -2.62484938e-01 4.40900743e-01 -9.69514966e-01
-1.23745866e-01 -5.28706789e-01 1.15612358e-01 -1.53308749e+00
-8.15549135e-01 1.08089566e+00 2.42472768e-01 -1.37155449e+00
-7.63468921e-01 -1.93606764e-01 -5.48142016e-01 8.11245203e-01
-1.35574222e+00 -9.48207378e-01 -3.12361956e-01 4.45841730e-01
1.35820067e+00 -3.18295509e-01 6.30256832e-01 4.17857140e-01
-4.56363767e-01 1.74503446e-01 -4.36110914e-01 6.44149557e-02
6.98277771e-01 -9.43746686e-01 3.81660253e-01 9.39199030e-01
1.14303298e-01 5.66397846e-01 1.07670081e+00 -6.50290132e-01
-1.11716092e+00 -1.10780156e+00 1.14863431e+00 -2.04084694e-01
4.91415560e-01 -2.36557163e-02 -7.61433601e-01 7.51673281e-01
1.05317295e+00 -2.20422208e-01 4.86083508e-01 -7.51985908e-01
3.87784168e-02 2.72610575e-01 -8.22110116e-01 6.86805308e-01
9.70186174e-01 -3.77500027e-01 -6.40803635e-01 3.22427511e-01
9.14848149e-01 -4.29615796e-01 -4.65923220e-01 1.18596472e-01
4.94622588e-01 -9.24769163e-01 6.43331289e-01 -6.74607217e-01
9.39469635e-01 -5.44076264e-01 -6.69865236e-02 -9.30022776e-01
-4.95578974e-01 -8.89782071e-01 4.75242734e-03 1.42421293e+00
3.09265733e-01 1.35848418e-01 6.94183826e-01 4.31455165e-01
-2.81830221e-01 -4.32369530e-01 -6.95956230e-01 -5.69139063e-01
-5.02912879e-01 -1.22958563e-01 2.06664726e-01 4.92327482e-01
-1.58506349e-01 6.43884718e-01 -1.00705445e+00 -1.87809527e-01
1.79576710e-01 -2.89101124e-01 4.46224511e-01 -5.06997645e-01
-1.25055432e-01 -2.26391226e-01 -6.43145919e-01 -9.65635955e-01
3.22843969e-01 -7.06222236e-01 2.19091296e-01 -1.68068326e+00
6.59807801e-01 5.51537812e-01 -1.98347509e-01 3.77211958e-01
-1.52408987e-01 5.84866345e-01 5.24912596e-01 2.02731058e-01
-8.33982825e-01 4.46082085e-01 1.15365517e+00 -8.51423591e-02
3.55118550e-02 -4.21927094e-01 -5.05062401e-01 3.85156363e-01
8.66445124e-01 -4.93349314e-01 -4.48115826e-01 -4.93480921e-01
1.00555286e-01 3.69952261e-01 1.26753479e-01 -1.18182135e+00
3.20781112e-01 -2.90573551e-03 3.36459368e-01 -7.41413593e-01
4.43018019e-01 -5.80814779e-01 5.72988689e-01 4.39598143e-01
-6.26689136e-01 7.37252831e-01 3.19812715e-01 4.58378226e-01
-5.20848274e-01 -4.75088656e-01 5.96927047e-01 -5.17915726e-01
-1.03006625e+00 2.67682016e-01 -6.62265301e-01 -2.10844770e-01
1.21489584e+00 -4.19751108e-01 -3.95443812e-02 -8.40406775e-01
-6.48379683e-01 -3.51491831e-02 2.51536727e-01 7.36972451e-01
9.15892661e-01 -1.53561318e+00 -1.09283972e+00 -2.48876303e-01
4.96062962e-03 -4.24677521e-01 3.59043509e-01 4.49088752e-01
-6.81955457e-01 8.51990700e-01 -5.38250029e-01 -4.50155467e-01
-1.45541191e+00 7.42745042e-01 4.63409647e-02 -1.56949162e-01
-5.69460630e-01 6.57301605e-01 1.42303437e-01 7.35293806e-01
1.12623677e-01 8.30570012e-02 -7.07332611e-01 -1.98499691e-02
8.35480690e-01 1.03566721e-01 -4.56182241e-01 -9.20101643e-01
-2.45879933e-01 5.46390414e-01 -3.89928669e-01 -4.49968487e-01
1.22893286e+00 -3.11879426e-01 2.55332887e-01 5.40023923e-01
1.17245901e+00 -4.87857342e-01 -1.31204855e+00 2.85497367e-01
-1.69585332e-01 -3.40437323e-01 -1.20115012e-01 -5.49387574e-01
-9.80782211e-01 7.31587648e-01 4.26236182e-01 2.59445131e-01
1.00948393e+00 -1.58578679e-01 9.20352757e-01 2.26239059e-02
9.33039002e-03 -1.02647734e+00 4.29664403e-01 6.47189140e-01
1.22119498e+00 -1.25353217e+00 -8.08528811e-02 5.32783717e-02
-1.12664950e+00 1.24665260e+00 4.85713750e-01 -4.36119556e-01
1.20834924e-01 9.71285403e-02 -1.59599092e-02 3.60078812e-02
-1.34229052e+00 -7.99547061e-02 1.90199912e-01 5.55145383e-01
8.99978817e-01 -3.92495215e-01 -4.00255084e-01 1.74730226e-01
5.17327301e-02 4.10765707e-01 1.08256507e+00 7.91451037e-01
-3.18672776e-01 -1.04626250e+00 -2.88969547e-01 1.42181128e-01
-9.00411963e-01 -3.73142391e-01 -1.00417696e-01 7.19118357e-01
4.04102542e-03 8.50300729e-01 3.09559554e-01 -6.05339520e-02
8.44705775e-02 4.44462560e-02 5.06604195e-01 -7.44842052e-01
-5.89717031e-01 9.38311368e-02 7.30255097e-02 -4.30742472e-01
-8.60259414e-01 -7.76289344e-01 -1.30428469e+00 -5.48365386e-03
-1.34366423e-01 2.64886409e-01 7.21711278e-01 6.91804171e-01
3.36621493e-01 9.58708763e-01 3.53638768e-01 -1.09846628e+00
-8.09778869e-02 -1.32343078e+00 -2.09778130e-01 5.39417267e-01
4.16969687e-01 -3.78756523e-01 -2.24179730e-01 5.98322570e-01] | [10.580971717834473, 0.611694872379303] |
63adfb87-cacd-4826-a9cf-7fec7fd904a4 | bayesian-optimisation-assisted-neural-network | 2203.04032 | null | https://arxiv.org/abs/2203.04032v1 | https://arxiv.org/pdf/2203.04032v1.pdf | Bayesian Optimisation-Assisted Neural Network Training Technique for Radio Localisation | Radio signal-based (indoor) localisation technique is important for IoT applications such as smart factory and warehouse. Through machine learning, especially neural networks methods, more accurate mapping from signal features to target positions can be achieved. However, different radio protocols, such as WiFi, Bluetooth, etc., have different features in the transmitted signals that can be exploited for localisation purposes. Also, neural networks methods often rely on carefully configured models and extensive training processes to obtain satisfactory performance in individual localisation scenarios. The above poses a major challenge in the process of determining neural network model structure, or hyperparameters, as well as the selection of training features from the available data. This paper proposes a neural network model hyperparameter tuning and training method based on Bayesian optimisation. Adaptive selection of model hyperparameters and training features can be realised with minimal need for manual model training design. With the proposed technique, the training process is optimised in a more automatic and efficient way, enhancing the applicability of neural networks in localisation. | ['Ziming Zhu', 'Peizheng Li', 'Xingchi Liu'] | 2022-03-08 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 4.99368966e-01 -3.51528823e-01 3.15991528e-02 -5.41329443e-01
-3.36897135e-01 -2.79811621e-01 5.65233052e-01 9.58261117e-02
-5.45124590e-01 8.06357861e-01 -2.58032307e-02 -3.49669278e-01
-1.05806208e+00 -7.90369570e-01 -3.44490439e-01 -1.09486914e+00
-1.90186575e-01 5.60831964e-01 -9.79161169e-03 3.16658393e-02
3.80418926e-01 7.37219930e-01 -1.37961185e+00 -3.98157120e-01
6.28782511e-01 1.11755693e+00 5.48726499e-01 8.46182704e-01
-2.47771457e-01 2.86785156e-01 -8.68587673e-01 1.58863530e-01
9.06124711e-02 -1.92133337e-01 -3.01294506e-01 -3.55371892e-01
-5.26715040e-01 4.56022561e-01 1.67051390e-01 6.42597377e-01
9.40924525e-01 3.26510131e-01 7.46175587e-01 -9.10123229e-01
3.62703741e-01 6.54842615e-01 -1.82639003e-01 6.81849122e-02
3.54163677e-01 -1.02124892e-01 2.55623251e-01 -3.42968911e-01
-1.13666423e-01 8.65796685e-01 8.66518021e-01 -2.57229321e-02
-1.24153697e+00 -9.66137409e-01 -3.27800721e-01 1.82635784e-01
-1.76927793e+00 -4.43793029e-01 9.18865502e-01 -1.60505965e-01
6.14066005e-01 3.04173440e-01 7.32252896e-01 1.05968976e+00
3.55616182e-01 2.26064682e-01 9.99592006e-01 -7.48732507e-01
4.19787943e-01 3.23403776e-01 -4.68751609e-01 1.45268604e-01
2.45963782e-01 1.69882208e-01 -2.97667682e-01 -1.78664222e-01
8.16177547e-01 -1.46611467e-01 -2.47366950e-01 -4.85610366e-01
-1.12120593e+00 7.97919631e-01 4.84798580e-01 7.26226628e-01
-7.13433743e-01 2.19475284e-01 1.55339375e-01 2.91192420e-02
-1.59211889e-01 7.44953156e-01 -8.48639846e-01 -3.10603499e-01
-8.82954419e-01 -2.91224599e-01 7.87589729e-01 6.58730268e-01
6.94504678e-01 3.23654786e-02 3.07501018e-01 1.02215457e+00
6.94759369e-01 5.85180819e-01 7.59443641e-01 -5.31252146e-01
3.46327960e-01 2.13944077e-01 7.06481412e-02 -1.23041534e+00
-8.74564826e-01 -9.28079844e-01 -1.05291152e+00 -1.32056624e-01
3.54584038e-01 -4.39886212e-01 -7.24194169e-01 1.23768055e+00
5.01429319e-01 2.47728199e-01 2.03101248e-01 4.49675828e-01
4.25415188e-01 6.02418303e-01 1.11234404e-01 -1.41241893e-01
1.08522213e+00 -1.87036529e-01 -5.50338268e-01 -3.70009780e-01
5.18137336e-01 -9.77623880e-01 4.79767382e-01 6.34216964e-01
-5.10166526e-01 -6.15768611e-01 -1.09806812e+00 8.83859813e-01
-3.62973332e-01 1.55646786e-01 8.56082320e-01 1.01264894e+00
-5.95828116e-01 5.59251308e-01 -7.08196044e-01 -5.47960341e-01
-2.12507136e-03 1.14762616e+00 -1.44634441e-01 1.20089874e-01
-1.31002319e+00 9.37903464e-01 1.04108822e+00 7.36002684e-01
-2.62185782e-01 -3.47357132e-02 -7.14565516e-01 -8.72486979e-02
8.14905390e-02 -5.94690084e-01 8.98961663e-01 -8.08702171e-01
-1.91060185e+00 -1.25201762e-01 1.20293140e-01 -4.91732866e-01
2.25633547e-01 1.28343388e-01 -6.12254500e-01 -2.62192070e-01
-4.55687791e-01 4.25743878e-01 1.03577578e+00 -1.05563235e+00
-8.04904878e-01 -6.42566197e-03 -3.23647618e-01 1.57487020e-01
-1.36331350e-01 -5.17106205e-02 -2.64609128e-01 -3.27356070e-01
3.73472571e-01 -8.07184756e-01 -4.78876203e-01 -7.51808643e-01
-8.59695077e-02 -7.95648061e-03 5.62059581e-01 -5.32268047e-01
1.11925542e+00 -1.92918432e+00 -3.02893251e-01 1.16325688e+00
-4.58469361e-01 1.47808194e-01 2.40604967e-01 2.44664758e-01
8.32495242e-02 -2.73118079e-01 1.12385809e-01 2.64673620e-01
7.63260573e-02 3.21802527e-01 4.63870078e-01 3.94029081e-01
-1.71748117e-01 3.88098598e-01 -6.38327420e-01 -6.12034202e-01
6.72603607e-01 9.08638597e-01 -1.91744238e-01 6.71887919e-02
3.30184758e-01 9.43662882e-01 -7.64302373e-01 4.68395650e-01
2.83288300e-01 6.66040182e-02 2.62475878e-01 -1.74356103e-01
1.73244670e-01 1.26053095e-01 -1.61257935e+00 1.26369298e+00
-1.24962187e+00 4.47435677e-01 2.46879105e-02 -1.21481514e+00
1.59015107e+00 5.46709001e-01 6.02065563e-01 -6.74774170e-01
5.68075418e-01 2.91183323e-01 1.40586019e-01 -3.74489009e-01
3.10531080e-01 9.66465771e-02 -2.42333204e-01 1.91748232e-01
-2.48386916e-02 -9.36340392e-02 -9.18806270e-02 -6.86467290e-01
7.83478081e-01 1.63400471e-01 4.60117191e-01 9.23047662e-02
8.84635031e-01 -3.91643226e-01 4.46036845e-01 9.44640934e-01
2.67515540e-01 8.28826874e-02 -2.39375979e-01 -3.31592649e-01
-6.74631059e-01 -4.18018878e-01 -3.97197187e-01 7.71020830e-01
-1.11977622e-01 1.60804629e-01 -4.56722349e-01 -4.20374602e-01
-2.04003602e-01 6.26906812e-01 -2.02373698e-01 -1.90852717e-01
-6.23238623e-01 -9.40569758e-01 4.40706909e-01 1.60694137e-01
4.03992683e-01 -1.22537541e+00 -7.30713606e-01 6.40855312e-01
9.94448289e-02 -7.83654094e-01 3.01356763e-01 8.36753547e-01
-1.10126865e+00 -6.01487219e-01 -4.92408305e-01 -6.14721060e-01
7.47240245e-01 -1.14669971e-01 8.02256167e-01 7.83969983e-02
-8.63267051e-04 1.17017165e-01 -3.45642865e-01 -6.10632539e-01
-3.11849654e-01 6.05416536e-01 -1.05498426e-01 7.05028996e-02
1.98293298e-01 -6.97277367e-01 -4.36277300e-01 7.64748216e-01
-5.55187047e-01 -3.31397057e-01 1.20614660e+00 8.88961256e-01
3.52770090e-01 9.82988298e-01 7.06059098e-01 -5.10972500e-01
7.77382374e-01 -4.54801649e-01 -8.00453365e-01 1.04682691e-01
-6.60815418e-01 2.48838574e-01 6.46724999e-01 -5.18976808e-01
-9.69806194e-01 4.40206766e-01 -3.79042536e-01 1.70355022e-01
-6.17945850e-01 6.19780838e-01 -4.21819538e-01 -6.50404215e-01
7.16486573e-01 1.52788207e-01 -2.55147368e-01 -4.24948573e-01
4.95846458e-02 9.72499311e-01 3.57872099e-01 -4.98738825e-01
7.30257094e-01 -1.90004364e-01 2.36893639e-01 -8.85482252e-01
-2.17718720e-01 -6.29970849e-01 -7.32434034e-01 -3.30185324e-01
4.24531907e-01 -5.03651619e-01 -7.25287318e-01 8.75119790e-02
-7.54373014e-01 -4.53555137e-02 1.80582732e-01 1.05110264e+00
-3.77233922e-01 -4.90350090e-02 3.46471608e-01 -1.13576126e+00
-2.37889439e-01 -1.14559436e+00 7.43037403e-01 5.33011675e-01
-6.43871069e-01 -1.19871747e+00 -1.43774807e-01 1.75534472e-01
7.59125352e-01 1.69584274e-01 5.67826629e-01 -8.25699925e-01
-3.88668239e-01 -8.02509785e-01 1.68364063e-01 3.55177373e-02
2.54833877e-01 -3.29335570e-01 -7.49815583e-01 -1.89241633e-01
-4.15355824e-02 1.80720896e-01 2.18091130e-01 7.55625308e-01
9.86395478e-01 -2.55909264e-01 -5.07719219e-01 6.82377875e-01
1.32570374e+00 6.58290029e-01 5.54181039e-01 7.81504214e-01
2.68850416e-01 4.76810306e-01 7.88464069e-01 3.96847904e-01
-9.32659879e-02 8.50914299e-01 4.33970720e-01 -5.96377477e-02
5.11187673e-01 6.35950565e-02 -7.03874081e-02 4.67049420e-01
-1.03994869e-01 -1.60112083e-01 -8.76537025e-01 2.68492311e-01
-1.66897476e+00 -6.98357522e-01 1.08027250e-01 2.52569652e+00
5.17503917e-01 3.83633167e-01 2.32072473e-02 8.39145660e-01
7.81563699e-01 -1.56470045e-01 -1.44149005e-01 -5.46494305e-01
3.90493631e-01 3.26026112e-01 9.87989187e-01 4.48579937e-01
-1.14757335e+00 5.18535674e-01 5.76665783e+00 9.07027483e-01
-1.34749985e+00 -1.22231245e-01 9.97104645e-02 6.65998533e-02
1.94905892e-01 -5.92793450e-02 -7.86213696e-01 6.04367912e-01
1.14402342e+00 4.04931754e-01 4.98541802e-01 8.12471330e-01
5.45744002e-01 -4.40232426e-01 -5.06370604e-01 1.17376614e+00
-2.71636307e-01 -8.65176857e-01 -5.09700775e-01 2.64569640e-01
4.87690359e-01 -2.70933539e-01 -1.35089189e-01 1.45546526e-01
1.71762049e-01 -1.04524207e+00 3.94867182e-01 6.75752103e-01
5.76356314e-02 -1.39134049e+00 1.44726419e+00 4.12811548e-01
-1.06727147e+00 -4.50808793e-01 -2.71009326e-01 1.30118877e-01
1.80920675e-01 7.61871874e-01 -1.41150594e+00 7.31810331e-01
5.99193335e-01 6.90655336e-02 -4.39980328e-01 1.67505610e+00
-3.19288641e-01 6.86329722e-01 -8.42220604e-01 -6.88483179e-01
6.33435324e-02 -6.02565892e-02 3.18336576e-01 1.13169336e+00
6.67428136e-01 -3.62398446e-01 -4.73354943e-02 1.41546816e-01
6.74559116e-01 2.38372669e-01 -4.44401234e-01 2.39367068e-01
7.31303990e-01 1.25401330e+00 -1.05545354e+00 3.68705601e-01
1.67818680e-01 2.88299441e-01 -4.91777450e-01 3.25276107e-01
-6.94347858e-01 -7.83370376e-01 9.09784213e-02 -3.00753545e-02
3.97945970e-01 -4.37944561e-01 -1.75725594e-01 -2.67296523e-01
-3.56370002e-01 -6.78164065e-01 2.31549144e-04 -5.00042558e-01
-6.88415468e-01 5.36265135e-01 8.75438303e-02 -1.15842426e+00
-8.90704870e-01 -4.82614815e-01 -5.50169468e-01 7.62794137e-01
-1.12583530e+00 -9.07192826e-01 -2.85279036e-01 4.21143562e-01
1.07729919e-01 -2.18497187e-01 8.09190810e-01 3.29728127e-01
-2.93074429e-01 3.73949409e-01 4.11041349e-01 -1.17325693e-01
4.12765801e-01 -9.79119539e-01 -1.40728876e-01 5.90622127e-01
2.36884385e-01 5.18556356e-01 9.97470260e-01 -2.63378829e-01
-1.09163451e+00 -8.51509035e-01 8.61079097e-01 -1.10979386e-01
4.00263280e-01 -2.65437603e-01 -4.06451106e-01 2.07970276e-01
-1.45524785e-01 -2.96059757e-01 6.75189495e-01 4.12508458e-01
6.84101641e-01 -4.95279878e-01 -1.21172321e+00 2.72932172e-01
3.88418108e-01 -6.63170069e-02 -2.28500545e-01 9.57752988e-02
-1.70720220e-01 -2.87428260e-01 -1.05055773e+00 4.05291438e-01
1.00196874e+00 -6.68649316e-01 1.16753232e+00 2.07478747e-01
-6.77079678e-01 -5.47610879e-01 -1.26942685e-02 -1.35274994e+00
-3.52697670e-01 -6.07572794e-01 3.15666236e-02 1.35769987e+00
7.05989301e-01 -8.23110878e-01 8.47568989e-01 3.32862616e-01
4.23084497e-01 -6.53915405e-01 -1.06668425e+00 -6.82248771e-01
-8.17367554e-01 -7.35183418e-01 1.06200886e+00 5.08290946e-01
-5.49878895e-01 2.96204925e-01 -3.97414088e-01 4.01366532e-01
5.28845727e-01 -4.99620527e-01 1.00223112e+00 -1.63076723e+00
-3.65050703e-01 -2.79665887e-01 -7.85677493e-01 -8.12190711e-01
-2.30709910e-01 -3.16100210e-01 3.20580482e-01 -1.53943789e+00
-6.34370029e-01 -1.06501687e+00 -4.55595702e-01 2.74955869e-01
3.15111637e-01 2.09985614e-01 -2.11649269e-01 -4.68241274e-02
-3.77651989e-01 1.72129199e-01 7.80678689e-01 2.98250858e-02
-6.33538306e-01 8.71172309e-01 -4.01666492e-01 7.56174445e-01
1.11244071e+00 -6.93872094e-01 -5.23386121e-01 -2.15288296e-01
3.90380472e-01 -8.02791491e-03 1.11383669e-01 -1.36639440e+00
4.27408487e-01 -4.80232723e-02 1.06999922e+00 -4.51053470e-01
3.01618725e-01 -1.48966217e+00 7.63398767e-01 3.40163052e-01
-2.42205504e-02 -1.39608264e-01 5.59988022e-02 5.11270702e-01
1.97400227e-02 -6.69583201e-01 4.60993260e-01 4.35057543e-02
-5.51337063e-01 -2.44927824e-01 -4.72714067e-01 -7.91061401e-01
8.10932755e-01 -9.00799990e-01 5.88833332e-01 -5.71698368e-01
-8.36316049e-01 1.38485819e-01 5.66249266e-02 4.53215092e-01
3.64178956e-01 -1.23512161e+00 -1.71564490e-01 3.87583762e-01
-1.24870837e-01 1.99817449e-01 3.56925372e-03 9.88126755e-01
-4.03237373e-01 6.68593049e-01 -6.81943670e-02 -8.38542461e-01
-1.16218996e+00 7.57010430e-02 3.15605223e-01 -3.19779783e-01
-1.18353978e-01 4.81604815e-01 -6.50255144e-01 -6.53272033e-01
4.90408629e-01 -2.88366824e-01 -6.61521614e-01 -2.06329245e-02
2.06494898e-01 3.45554739e-01 2.53814846e-01 -6.77277505e-01
-5.13946295e-01 9.17975307e-01 4.25243497e-01 -1.48477823e-01
1.23192275e+00 -3.39091033e-01 2.94888347e-01 5.69682941e-02
8.52173626e-01 -1.06577871e-04 -9.26595449e-01 -7.51491189e-02
3.91706377e-01 -3.89131665e-01 4.70145196e-01 -8.62917781e-01
-7.44749248e-01 3.99691314e-01 9.15706336e-01 2.60405540e-01
1.26967478e+00 -3.42664957e-01 4.53519970e-01 6.23571336e-01
8.29923868e-01 -1.16576457e+00 -5.67637920e-01 2.91191131e-01
5.29640496e-01 -9.68273163e-01 3.15939859e-02 1.02354564e-01
-2.20738411e-01 1.23212588e+00 2.39677876e-02 1.75897032e-01
8.10116529e-01 1.80111587e-01 1.91108137e-01 1.65636182e-01
-1.89538434e-01 -1.01029605e-01 3.10222566e-01 9.46585834e-01
3.04182649e-01 -6.11421950e-02 -8.59895721e-02 3.60862345e-01
-5.96456289e-01 5.15126213e-02 -8.01551044e-02 1.06046629e+00
-6.20744228e-01 -1.24195290e+00 -9.62781608e-01 4.72362816e-01
-4.60992992e-01 3.30696642e-01 1.20512120e-01 7.60710120e-01
3.34657490e-01 1.18824482e+00 -1.46027982e-01 -4.50812757e-01
3.36897552e-01 -8.77787024e-02 5.15866339e-01 -2.10682407e-01
-6.31390631e-01 2.29008004e-01 2.03024253e-01 -1.93194613e-01
-4.50354755e-01 -5.58752000e-01 -9.70267177e-01 3.83094372e-03
-1.03933799e+00 6.04693592e-01 1.49682379e+00 1.20199704e+00
1.02135241e-01 6.81198955e-01 7.75051594e-01 -1.16589141e+00
-2.08307594e-01 -1.07526743e+00 -4.28080410e-01 -3.70085597e-01
-3.97438556e-02 -8.67352426e-01 -2.73032010e-01 -3.58099580e-01] | [6.425114631652832, 1.0043600797653198] |
973cd699-3474-4bed-9fb0-f3ae36cd4d2a | unsupervised-multi-hop-question-answering-by | 2010.12623 | null | https://arxiv.org/abs/2010.12623v2 | https://arxiv.org/pdf/2010.12623v2.pdf | Unsupervised Multi-hop Question Answering by Question Generation | Obtaining training data for multi-hop question answering (QA) is time-consuming and resource-intensive. We explore the possibility to train a well-performed multi-hop QA model without referencing any human-labeled multi-hop question-answer pairs, i.e., unsupervised multi-hop QA. We propose MQA-QG, an unsupervised framework that can generate human-like multi-hop training data from both homogeneous and heterogeneous data sources. MQA-QG generates questions by first selecting/generating relevant information from each data source and then integrating the multiple information to form a multi-hop question. Using only generated training data, we can train a competent multi-hop QA which achieves 61% and 83% of the supervised learning performance for the HybridQA and the HotpotQA dataset, respectively. We also show that pretraining the QA system with the generated data would greatly reduce the demand for human-annotated training data. Our codes are publicly available at https://github.com/teacherpeterpan/Unsupervised-Multi-hop-QA. | ['William Yang Wang', 'Min-Yen Kan', 'Wenhan Xiong', 'Wenhu Chen', 'Liangming Pan'] | 2020-10-23 | null | https://aclanthology.org/2021.naacl-main.469 | https://aclanthology.org/2021.naacl-main.469.pdf | naacl-2021-4 | ['multi-hop-question-answering'] | ['knowledge-base'] | [-4.62984703e-02 6.54221475e-01 3.96364421e-01 -6.69084072e-01
-2.21789312e+00 -9.96420026e-01 1.19811974e-01 2.71084696e-01
-4.32795852e-01 8.46077442e-01 5.44495694e-02 -5.11560619e-01
-1.14420475e-02 -9.53804970e-01 -7.17239916e-01 -3.19802642e-01
6.61435902e-01 1.22978699e+00 6.38371527e-01 -6.09966159e-01
-1.60717294e-01 -1.65348560e-01 -1.53971958e+00 6.75040603e-01
1.65320098e+00 7.22883582e-01 1.40936449e-01 1.19370949e+00
-2.49269366e-01 1.11590815e+00 -9.12755907e-01 -8.02911103e-01
-1.89198796e-02 -9.09945905e-01 -1.59988725e+00 -2.25649089e-01
7.04570591e-01 -2.97306955e-01 5.32756634e-02 5.45851886e-01
7.78993905e-01 1.27972528e-01 3.18846047e-01 -1.29136467e+00
-7.56982744e-01 5.74232161e-01 -4.28842790e-02 -4.39821929e-02
8.09987307e-01 2.12953523e-01 1.21940291e+00 -9.84152496e-01
7.04944193e-01 1.07110596e+00 2.69105643e-01 8.06095719e-01
-7.87327826e-01 -2.38318741e-01 -6.92973316e-01 2.88133115e-01
-1.11781406e+00 -5.62150359e-01 5.20452321e-01 -1.41775593e-01
5.58632493e-01 4.70254719e-01 3.12241558e-02 8.16640675e-01
-3.50752592e-01 1.01205993e+00 1.08935678e+00 -8.47143650e-01
8.87888446e-02 1.53149396e-01 4.03108954e-01 9.70580280e-01
-2.21975207e-01 -4.34047699e-01 -4.97479588e-01 -4.02438968e-01
2.05430202e-02 -6.56146467e-01 -1.95024729e-01 5.48868962e-02
-1.14089727e+00 1.01685643e+00 3.98878813e-01 1.71812817e-01
-3.06340247e-01 -2.75583982e-01 1.06598228e-01 5.05402863e-01
7.87792578e-02 8.87206197e-01 -4.87098843e-01 -1.05972931e-01
-7.91187823e-01 2.52899468e-01 1.07441783e+00 1.22890317e+00
1.14942420e+00 -4.85229850e-01 -4.45226490e-01 8.26099396e-01
2.12417603e-01 6.69159114e-01 4.06218588e-01 -1.44645774e+00
9.45961118e-01 8.77400279e-01 2.60063112e-01 -4.18645352e-01
-2.39830852e-01 1.70144308e-02 -5.01058638e-01 -5.01720607e-01
7.68538833e-01 -6.01155579e-01 -8.26829135e-01 1.49809408e+00
7.58141935e-01 -2.98576564e-01 5.12127638e-01 7.46982872e-01
1.30237949e+00 8.33658814e-01 2.51338035e-01 1.25979871e-01
1.55682540e+00 -1.50429296e+00 -5.82296312e-01 -7.64500126e-02
1.07489872e+00 -7.76879311e-01 1.56092906e+00 7.20077679e-02
-1.38672829e+00 -7.44559884e-01 -6.12157941e-01 -5.81793547e-01
-3.36174637e-01 3.90459411e-02 5.41862920e-02 4.53532338e-01
-1.18076253e+00 -1.46944195e-01 -3.83009851e-01 -2.00782314e-01
1.98449641e-01 9.05664638e-02 -4.53458399e-01 -5.77829123e-01
-1.50478196e+00 7.75494933e-01 3.62550706e-01 -2.01890916e-01
-1.21288538e+00 -6.26461923e-01 -8.33147645e-01 -3.89250964e-02
4.53448266e-01 -7.42391527e-01 1.69618642e+00 -6.95974171e-01
-1.34884763e+00 8.41987729e-01 -3.52288336e-01 -2.45023817e-01
4.00578588e-01 -2.01641172e-01 -4.17136282e-01 7.70883381e-01
5.88165283e-01 1.06332386e+00 5.01121521e-01 -1.32694411e+00
-6.39436126e-01 -2.07800359e-01 3.95437717e-01 3.37691784e-01
-1.11994147e-01 -1.06390737e-01 -4.46095765e-01 7.43993744e-02
-1.66410893e-01 -8.41152132e-01 -2.04770401e-01 -3.42961282e-01
-5.00789225e-01 -6.55721426e-01 5.99795759e-01 -9.20028508e-01
9.77277994e-01 -1.72432220e+00 -1.86979830e-01 -6.73609972e-02
2.58316755e-01 3.81517321e-01 -6.80070758e-01 7.32525170e-01
4.80065167e-01 4.39660437e-02 -4.48029548e-01 -1.42535597e-01
8.80930759e-03 2.60808170e-01 -1.38076082e-01 -1.89655513e-01
5.48763156e-01 1.19135380e+00 -1.38354969e+00 -9.84275818e-01
-2.41567612e-01 2.06182584e-01 -5.67342579e-01 9.19079006e-01
-6.59683228e-01 6.90003276e-01 -5.62039256e-01 6.70493603e-01
3.84917289e-01 -7.06625521e-01 4.07927223e-02 -1.05187029e-01
4.42230701e-01 4.81382877e-01 -6.86797500e-01 1.96884906e+00
-7.76590407e-01 4.10876781e-01 -9.42564011e-02 -6.20184004e-01
9.28897262e-01 6.80868864e-01 2.24754050e-01 -1.04417074e+00
-8.33304524e-02 5.91847777e-01 -1.40225381e-01 -8.28642666e-01
7.50462592e-01 1.30636841e-01 -4.60989267e-01 7.46464491e-01
5.53732872e-01 -3.13079059e-01 6.14623547e-01 4.76556718e-01
1.29012620e+00 -1.50274694e-01 -2.07528308e-01 1.27313249e-02
7.89748013e-01 6.36059582e-01 1.40650451e-01 8.05132270e-01
-2.54782230e-01 6.48478448e-01 1.74440145e-01 7.00314641e-02
-7.47267723e-01 -1.11536384e+00 3.07509124e-01 1.67024338e+00
9.39256698e-03 -2.33197436e-01 -1.05259597e+00 -1.21739173e+00
-3.23029369e-01 8.80970597e-01 -2.69529849e-01 -3.86183150e-03
-7.02669501e-01 -1.14211135e-01 9.61969197e-01 1.91890955e-01
7.40579903e-01 -1.22948098e+00 -6.08964741e-01 3.31294745e-01
-1.00331545e+00 -1.06743896e+00 -3.06600332e-01 -4.81801741e-02
-4.73190695e-01 -1.30993760e+00 -7.90535748e-01 -1.01197720e+00
8.01116109e-01 8.76144916e-02 1.77973020e+00 3.16863447e-01
-7.52887875e-02 6.04671478e-01 -8.63524973e-01 -2.04687148e-01
-7.45637298e-01 6.01333678e-01 -5.98905325e-01 -1.62396580e-01
4.70734447e-01 1.35319242e-02 -7.08151877e-01 5.83755076e-01
-9.40457761e-01 7.96271637e-02 5.67009866e-01 8.49972785e-01
6.15294456e-01 -3.96077037e-01 1.13398409e+00 -1.17889798e+00
9.59218919e-01 -7.58456826e-01 -3.81807268e-01 9.17651594e-01
-9.19157863e-02 1.06778644e-01 7.26493537e-01 -1.47805199e-01
-1.27891147e+00 5.17191030e-02 -6.27159059e-01 9.86079425e-02
-3.94738376e-01 4.76786226e-01 -2.56545335e-01 8.22617635e-02
1.07094133e+00 5.83633669e-02 -3.44130754e-01 -3.61533463e-01
1.01520956e+00 1.05458081e+00 6.94003582e-01 -8.44787061e-01
8.35544467e-01 9.35242027e-02 -6.58207774e-01 -5.08310556e-01
-1.30972099e+00 -7.66300678e-01 -6.70152843e-01 -2.28859752e-01
1.07678914e+00 -8.99776638e-01 -5.23806334e-01 1.26023307e-01
-1.31115496e+00 -6.67971373e-01 -4.89080817e-01 5.85285723e-02
-4.87324744e-01 4.21949998e-02 -5.71878493e-01 -5.05149603e-01
-6.74054384e-01 -9.87017155e-01 1.15685272e+00 5.08318484e-01
-1.58957034e-01 -9.94113564e-01 3.63603413e-01 1.43430316e+00
2.12394193e-01 -3.55538875e-02 7.87227333e-01 -1.06544209e+00
-5.80642402e-01 -2.26798981e-01 -6.73882887e-02 2.12766215e-01
-1.69847514e-02 -3.61299932e-01 -1.08051038e+00 -1.82316586e-01
-1.67582497e-01 -1.29887652e+00 2.14250565e-01 -3.36897045e-01
8.86995077e-01 -3.98418784e-01 2.30266795e-01 -1.09771974e-01
1.24984014e+00 -1.24078423e-01 4.32080150e-01 4.18036282e-02
6.26321018e-01 1.02242553e+00 1.09368694e+00 -2.42332499e-02
1.10937095e+00 2.87324309e-01 2.77985126e-01 -1.04634754e-01
-1.74555480e-01 -4.05008942e-01 1.31419986e-01 1.29551792e+00
4.82164115e-01 -6.11401618e-01 -1.27191985e+00 1.14888752e+00
-1.67716324e+00 -5.86517215e-01 -4.59237784e-01 1.71427989e+00
1.24627793e+00 -4.33335811e-01 1.11413248e-01 -1.07807465e-01
6.07811928e-01 -2.43894294e-01 -4.98042077e-01 -1.01080045e-01
-2.51362715e-02 4.84386593e-01 4.65433113e-02 8.00383091e-01
-7.16697156e-01 9.81702864e-01 5.63658953e+00 6.87309265e-01
-4.68978971e-01 5.63131511e-01 6.42844260e-01 3.00621390e-01
-7.15635657e-01 5.85658569e-03 -8.41002226e-01 2.21200347e-01
1.57663655e+00 -1.62263125e-01 3.42177898e-02 7.09325492e-01
-2.47774720e-01 -1.45495966e-01 -8.53505492e-01 5.62544167e-01
1.35753855e-01 -1.21841526e+00 3.10613900e-01 -3.72840255e-01
8.55729938e-01 8.50207433e-02 -2.17297792e-01 5.97470760e-01
9.07603025e-01 -8.31479430e-01 3.34162772e-01 3.20787817e-01
7.38268912e-01 -4.96729195e-01 9.16874409e-01 7.15110481e-01
-9.89906788e-01 6.30573649e-03 -3.68393153e-01 5.50186217e-01
4.45868880e-01 4.21249002e-01 -1.04073787e+00 9.49401140e-01
6.59092009e-01 -1.51883110e-01 -1.02498901e+00 8.93855035e-01
-5.42676330e-01 8.75145257e-01 -2.16171890e-01 2.67316066e-02
3.64221811e-01 -9.24600009e-03 -7.08198696e-02 9.62745190e-01
4.18167055e-01 2.25725815e-01 1.76957771e-01 5.40643394e-01
-4.70973492e-01 3.41483414e-01 -2.45889798e-01 -1.03515059e-01
7.39493310e-01 1.41927624e+00 -6.21317141e-02 -6.28711641e-01
-4.83246863e-01 9.39049661e-01 6.06023192e-01 2.79097497e-01
-6.27227843e-01 -7.23808825e-01 -1.62570134e-01 -9.18876827e-02
-1.30503207e-01 1.14070676e-01 2.06200272e-01 -1.05905378e+00
1.58765554e-01 -1.34263921e+00 1.03618896e+00 -1.18112624e+00
-1.45160770e+00 1.02779675e+00 -2.52598196e-01 -1.05674720e+00
-7.63390183e-01 -2.36402720e-01 -4.29347843e-01 9.57992435e-01
-1.72867155e+00 -1.32911325e+00 -7.96497166e-01 9.90658820e-01
6.08959973e-01 5.59172500e-03 1.05095232e+00 4.49943125e-01
-1.73596501e-01 7.81900346e-01 -3.15202534e-01 3.24332565e-01
1.06732869e+00 -1.46236396e+00 3.23504627e-01 6.90001965e-01
3.55191559e-01 2.97483504e-01 3.93550456e-01 -3.45513642e-01
-1.37945914e+00 -1.24249339e+00 1.32368159e+00 -8.62685204e-01
6.14112139e-01 -2.05634415e-01 -1.31468642e+00 5.97356319e-01
6.36845231e-01 -1.67819872e-01 1.15698171e+00 -1.87720805e-01
-3.37195724e-01 -3.33834626e-02 -1.29376280e+00 3.78088683e-01
5.22670805e-01 -8.71061206e-01 -9.73403931e-01 6.56186461e-01
1.04352403e+00 -3.58850628e-01 -1.24156284e+00 2.79780567e-01
-1.79373324e-01 -8.18891168e-01 6.88792408e-01 -4.95146304e-01
3.07602972e-01 -3.25574666e-01 -8.30028355e-02 -1.20160067e+00
1.81382939e-01 -5.83425760e-01 -7.77440518e-02 1.35155833e+00
9.57123518e-01 -5.09327531e-01 7.38375664e-01 6.46928966e-01
-3.31110179e-01 -5.36412776e-01 -8.34661305e-01 -4.53172863e-01
2.86564648e-01 -8.78542138e-04 8.60728323e-01 9.63147938e-01
-1.66508451e-01 7.73053110e-01 -2.87785828e-02 4.41164404e-01
4.81318176e-01 2.28413969e-01 9.78612304e-01 -8.56978655e-01
-2.52331972e-01 3.21294338e-01 4.38259602e-01 -1.28890932e+00
3.52266952e-02 -8.18554282e-01 5.29106677e-01 -1.95877385e+00
-1.97942138e-01 -7.57834911e-01 4.78033461e-02 5.25154114e-01
-6.09119117e-01 3.83701801e-01 -6.76757023e-02 1.98109001e-01
-1.31512487e+00 4.45721269e-01 1.45092678e+00 -2.80804764e-02
2.87844613e-02 -2.16136187e-01 -7.02972293e-01 1.63833126e-01
1.01504612e+00 -6.05010271e-01 -8.71620297e-01 -9.52974975e-01
1.58238530e-01 5.42670250e-01 1.99278906e-01 -9.98609662e-01
5.00883698e-01 5.12222201e-02 -7.53174275e-02 -6.81366980e-01
4.11717743e-01 -5.33852339e-01 -3.86090994e-01 2.16176018e-01
-5.45033634e-01 2.59578258e-01 -3.92492339e-02 7.42007494e-02
-6.35591030e-01 -5.88364363e-01 5.19932508e-01 -4.37252909e-01
-6.19149983e-01 1.58635363e-01 -1.23905070e-01 7.63210475e-01
8.74701142e-01 3.99120092e-01 -1.06018519e+00 -7.46142566e-01
-6.11759961e-01 8.26165617e-01 1.32904872e-01 2.20932037e-01
4.98595953e-01 -1.23741138e+00 -1.10265970e+00 -2.01942682e-01
4.57833022e-01 3.49159718e-01 4.96067047e-01 4.81223077e-01
-6.83286369e-01 5.50527632e-01 -1.26048267e-01 -5.81407905e-01
-1.06022537e+00 2.25631371e-01 2.08230138e-01 -5.75653434e-01
-2.37887551e-04 9.61969078e-01 -3.28579009e-01 -1.21048474e+00
-1.34522349e-01 2.10596636e-01 -1.52373031e-01 1.41763180e-01
4.90102351e-01 2.58746862e-01 1.00347780e-01 -6.89501584e-01
-1.00689627e-01 3.33832979e-01 1.04751602e-01 -4.14099753e-01
8.77406359e-01 -1.01086773e-01 -1.36955917e-01 2.88616329e-01
1.20683110e+00 -3.57947722e-02 -8.07959080e-01 -2.18729436e-01
2.83395559e-01 -7.57090151e-02 -5.74479759e-01 -1.07077861e+00
-6.25885844e-01 1.07024634e+00 3.07802171e-01 4.64877993e-01
1.07741880e+00 4.39115703e-01 1.18679357e+00 7.95751512e-01
4.38189059e-01 -1.05396998e+00 5.21944344e-01 5.44000387e-01
9.27967012e-01 -1.47178376e+00 -6.50661528e-01 -4.04031336e-01
-7.96739876e-01 6.14474893e-01 1.10425913e+00 2.54958540e-01
1.39877602e-01 -4.05863464e-01 8.52044106e-01 -4.16462928e-01
-8.98191273e-01 -4.19781148e-01 3.76890838e-01 7.58304894e-01
4.78398979e-01 7.47803412e-03 3.55840661e-02 4.31213170e-01
-5.72348177e-01 -1.53615803e-01 4.28504437e-01 1.09741998e+00
-5.30741513e-01 -1.30477715e+00 -2.09409267e-01 3.65964979e-01
-2.44641706e-01 -1.74266368e-01 -5.13517678e-01 6.32547438e-01
-1.05201669e-01 1.62704885e+00 -1.05526119e-01 -1.56054169e-01
5.29302001e-01 6.20880961e-01 1.67773604e-01 -8.93557906e-01
-8.83406460e-01 -4.99997526e-01 6.23225570e-01 -3.90640616e-01
-4.43534076e-01 -1.04658954e-01 -1.24694347e+00 1.64227821e-02
-3.50566477e-01 1.01373971e+00 3.72090846e-01 9.42776620e-01
6.36722088e-01 2.63089210e-01 5.74306607e-01 2.20208932e-02
-4.34512794e-01 -1.12723184e+00 1.75390631e-01 5.77307105e-01
1.89789176e-01 1.15142949e-03 -2.52350241e-01 1.37335747e-01] | [11.296396255493164, 8.14024829864502] |
31c693e9-6833-4825-9a95-0bae091ae7a6 | dating-ancient-texts-an-approach-for-noisy | null | null | https://aclanthology.org/2020.lt4hala-1.3 | https://aclanthology.org/2020.lt4hala-1.3.pdf | Dating Ancient texts: an Approach for Noisy French Documents | Automatic dating of ancient documents is a very important area of research for digital humanities applications. Many documents available via digital libraries do not have any dating or dating that is uncertain. Document dating is not only useful by itself but it also helps to choose the appropriate NLP tools (lemmatizer, POS tagger ) for subsequent analysis. This paper provides a dataset with thousands of ancient documents in French and present methods and evaluation metrics for this task. We compare character-level methods with token-level methods on two different datasets of two different time periods and two different text genres. Our results show that character-level models are more robust to noise than classical token-level models. The experiments presented in this article focused on documents written in French but we believe that the ability of character-level models to handle noise properly would help to achieve comparable results on other languages and more ancient languages in particular. | ['Ga{\\"e}l Lejeune', 'Ana{\\"e}lle Baledent', 'Nicolas Hiebel'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['document-dating'] | ['natural-language-processing'] | [-1.35733277e-01 -4.62102383e-01 -2.12855220e-01 -4.01857406e-01
-9.51284170e-01 -9.88342285e-01 1.15203011e+00 4.61429060e-01
-1.00942659e+00 1.07064402e+00 2.97634304e-01 -5.31568229e-01
-1.05019957e-01 -8.94202530e-01 -4.20655638e-01 -3.68138254e-01
-2.12832093e-02 9.70703900e-01 3.57092321e-01 -2.82887667e-01
9.34429169e-01 6.09645426e-01 -1.50197744e+00 3.70461524e-01
6.72514558e-01 1.47365734e-01 1.80440411e-01 9.00552392e-01
-7.96119571e-01 2.90812075e-01 -8.40600014e-01 -6.81034684e-01
8.10579658e-02 -1.85229346e-01 -9.22942460e-01 -7.14175761e-01
5.78685522e-01 -2.49317531e-02 -1.77601762e-02 1.07261264e+00
6.25147641e-01 -2.26867557e-01 9.11734760e-01 -3.43586445e-01
-5.74270785e-01 1.28305387e+00 -4.98273611e-01 6.27717257e-01
7.34793127e-01 -5.91023088e-01 9.69801784e-01 -7.53365874e-01
1.02286291e+00 1.48056185e+00 8.52075815e-01 3.77533942e-01
-1.07102001e+00 -4.02531713e-01 -1.87779084e-01 2.21140832e-01
-1.35367405e+00 -4.65955108e-01 5.26104987e-01 -4.05786335e-01
6.28695369e-01 3.96864802e-01 5.10411561e-01 1.30640817e+00
1.82874948e-01 5.28641522e-01 1.46818554e+00 -9.93289948e-01
1.60353221e-02 1.82293907e-01 5.64760506e-01 3.85316849e-01
6.39847577e-01 -9.48387086e-02 -6.76437616e-01 -4.15857971e-01
4.68544900e-01 -5.57055295e-01 -6.08945675e-02 6.47852540e-01
-1.31661284e+00 8.11496854e-01 -4.77342963e-01 1.07454634e+00
-6.42961860e-02 2.12913454e-01 6.92443073e-01 2.25842699e-01
4.37428415e-01 3.89123559e-01 -4.61788535e-01 -6.57457530e-01
-1.58709013e+00 4.27012920e-01 1.07398593e+00 1.00020945e+00
2.06527919e-01 -1.31130278e-01 -1.16425296e-02 1.06490588e+00
2.92386085e-01 5.48359632e-01 6.21109188e-01 -7.19353795e-01
3.51083815e-01 7.44738728e-02 2.25713670e-01 -7.30159342e-01
-2.51570463e-01 1.74631611e-01 -4.30516958e-01 -1.97152495e-01
8.92418265e-01 3.09033513e-01 -1.13475358e+00 1.41581023e+00
6.79414794e-02 -3.37614775e-01 -2.16534823e-01 3.98605734e-01
8.42640996e-01 9.68901932e-01 5.41097894e-02 -3.43500793e-01
1.97567666e+00 -1.82318509e-01 -8.99303377e-01 -5.29508367e-02
3.61259758e-01 -1.45319819e+00 1.17219985e+00 7.86669314e-01
-1.09567595e+00 -2.11816400e-01 -9.98329937e-01 -3.49916220e-01
-6.74738407e-01 1.55326098e-01 7.45586097e-01 1.19774246e+00
-7.71250010e-01 8.64487171e-01 -6.75087214e-01 -7.38133788e-01
-1.11801393e-01 6.90345317e-02 -1.01645686e-01 4.57915775e-02
-1.31456876e+00 9.97466803e-01 5.39403915e-01 1.44283683e-03
-2.81763911e-01 -2.50615448e-01 -4.73688781e-01 -2.89509714e-01
5.68284392e-02 -1.54850766e-01 1.41641998e+00 -3.28089356e-01
-1.18058646e+00 1.23687410e+00 -3.21917295e-01 -3.34850401e-01
9.31844115e-01 -2.93556333e-01 -5.90298057e-01 4.60286736e-02
3.88236701e-01 1.80424377e-03 5.69754601e-01 -1.01108813e+00
-8.17853987e-01 -3.93274933e-01 -6.47350371e-01 -4.79918599e-01
-3.39294553e-01 6.52549982e-01 -7.93928802e-01 -1.05581939e+00
3.65978628e-01 -7.47008264e-01 1.99495599e-01 -4.61910367e-01
-9.40365419e-02 -4.16716099e-01 2.77430803e-01 -1.16008377e+00
1.61333132e+00 -1.61962819e+00 -2.34948799e-01 4.22462583e-01
-4.11512524e-01 2.32099202e-02 3.04771900e-01 6.48961604e-01
3.91696811e-01 5.41086674e-01 -1.61571264e-01 -3.19593370e-01
-5.81050254e-02 5.89970052e-01 -3.12562197e-01 5.69913208e-01
-3.67760628e-01 3.78412992e-01 -1.01823986e+00 -9.64715600e-01
-2.53873438e-01 5.01651347e-01 9.54763517e-02 -6.56163871e-01
-1.64367497e-01 -7.26081654e-02 -3.64829481e-01 1.02685535e+00
5.15147269e-01 5.46679735e-01 2.87352026e-01 2.50397444e-01
-5.20136535e-01 4.72613245e-01 -1.35432255e+00 1.57178211e+00
-4.23549414e-01 9.70318556e-01 -2.81906843e-01 -2.86737025e-01
1.13649297e+00 1.89586788e-01 6.90000951e-02 -6.20795548e-01
2.55043864e-01 8.52453887e-01 -7.39633143e-02 -4.31045860e-01
1.31190550e+00 -2.03176931e-01 -3.77522618e-01 5.38333535e-01
-1.46162868e-01 3.38765932e-03 9.66848791e-01 3.56661797e-01
8.36617649e-01 3.82431835e-01 1.77947596e-01 -7.37497985e-01
5.61163902e-01 7.09754899e-02 7.07279861e-01 8.52081478e-01
8.70557725e-02 7.43776381e-01 4.68580544e-01 -3.97954315e-01
-1.55996537e+00 -8.93265188e-01 -7.68044591e-01 8.90854836e-01
-4.05738354e-01 -7.40621924e-01 -5.08342266e-01 -2.66168207e-01
1.31244540e-01 8.44891429e-01 -6.53905213e-01 6.20379031e-01
-8.17391217e-01 -9.55363989e-01 1.06130087e+00 3.34476531e-01
3.03985402e-02 -9.81191456e-01 -2.95395136e-01 5.20723045e-01
-3.89855921e-01 -8.40251267e-01 -3.22595499e-02 1.54529020e-01
-1.01093268e+00 -9.39709067e-01 -9.94788766e-01 -6.64095283e-01
4.65937585e-01 -2.36826062e-01 1.15075111e+00 5.65843210e-02
-2.39137217e-01 3.71011943e-02 -6.90764606e-01 -5.65736353e-01
-8.32514882e-01 4.06499535e-01 9.95867327e-02 -5.57839036e-01
9.25985754e-01 -2.56085336e-01 4.54744324e-02 -5.29056676e-02
-8.01701844e-01 -7.19296813e-01 2.59962797e-01 6.10834599e-01
3.24972898e-01 -5.38265370e-02 3.07490021e-01 -1.33416355e+00
8.41050625e-01 -2.66408056e-01 -6.06580079e-01 4.11706418e-01
-8.33913565e-01 2.43076056e-01 4.89912003e-01 -3.73722553e-01
-1.15198374e+00 -3.07570010e-01 -2.39152730e-01 6.24782562e-01
1.20696738e-01 6.48315847e-01 7.47503787e-02 3.23482722e-01
9.34274733e-01 1.03516802e-01 -5.59311986e-01 -1.06003630e+00
1.01272307e-01 9.48068440e-01 9.36369479e-01 -1.19579840e+00
6.74768627e-01 3.96852434e-01 -1.38164401e-01 -1.09222126e+00
-3.14166874e-01 -5.54024696e-01 -8.49115133e-01 -3.12178671e-01
3.21018249e-01 -6.73609257e-01 -4.53709483e-01 3.26457649e-01
-1.26429069e+00 2.61650980e-01 3.17110047e-02 4.68625665e-01
-1.55276015e-01 7.97970772e-01 -9.17831004e-01 -1.06148791e+00
-4.27897066e-01 -4.34215367e-01 9.13658023e-01 2.94599116e-01
-5.85294187e-01 -9.90923405e-01 4.67478126e-01 1.23524860e-01
-9.98650640e-02 4.00668144e-01 1.01137447e+00 -5.57038188e-01
-1.25492752e-01 -4.55812901e-01 1.07121259e-01 -8.84362087e-02
6.39109686e-02 9.11506951e-01 -9.41139758e-01 -4.60381843e-02
-1.99006617e-01 1.99651644e-01 1.09838080e+00 2.28728697e-01
5.06030738e-01 -5.64616546e-02 -3.30826730e-01 2.42618859e-01
1.53310823e+00 1.46504477e-01 9.77125704e-01 1.06638050e+00
2.82604754e-01 6.52565777e-01 7.78659701e-01 5.84629714e-01
1.58805698e-01 3.47650349e-01 -3.92998219e-01 6.86989784e-01
4.70480993e-02 -1.90976813e-01 4.67938662e-01 9.10071552e-01
-6.00655377e-01 -1.63515121e-01 -1.35170972e+00 9.76518512e-01
-1.66865790e+00 -1.22145629e+00 -7.87541151e-01 2.31962752e+00
1.30831790e+00 4.48337018e-01 3.17134351e-01 5.52661896e-01
9.03862119e-01 -9.28482786e-02 2.91008681e-01 -9.19150412e-01
-5.08911490e-01 4.03465092e-01 6.86267018e-01 5.45929432e-01
-8.01695585e-01 9.83187318e-01 6.87793732e+00 8.62172604e-01
-7.71375716e-01 -7.20302463e-02 2.13605940e-01 3.39535400e-02
-5.37475348e-01 3.18937987e-01 -1.26684237e+00 8.32472384e-01
1.23988461e+00 -4.04150784e-01 7.39111449e-04 6.36869073e-01
2.84723699e-01 -5.88932633e-01 -1.07953584e+00 7.36837804e-01
2.02018046e-03 -1.25420761e+00 3.61927263e-02 9.26676691e-02
5.50898075e-01 -2.40480512e-01 -3.36850017e-01 -6.59092963e-02
3.40624511e-01 -8.02532494e-01 1.21258247e+00 6.19531393e-01
6.27908945e-01 -9.75072205e-01 7.61875570e-01 6.98374212e-02
-8.72510850e-01 3.16064060e-01 -7.91279316e-01 1.60253271e-01
4.13942188e-01 8.32494438e-01 -7.49145031e-01 6.30988836e-01
9.38234210e-01 4.29362506e-01 -9.02053416e-01 1.24480057e+00
-3.05324197e-01 8.48566353e-01 -6.60446823e-01 -3.55733126e-01
1.32166356e-01 -3.20575595e-01 5.42047262e-01 1.86349571e+00
6.11887872e-01 -1.26307189e-01 -6.20437562e-01 3.64465714e-01
1.36012539e-01 5.45596242e-01 -3.25393528e-01 -3.16040576e-01
8.25881183e-01 9.66329157e-01 -1.32586873e+00 -3.78639728e-01
-3.06398153e-01 7.37515092e-01 -4.52757031e-02 -1.77463032e-02
-4.50137734e-01 -7.00232446e-01 1.79927066e-01 3.56747717e-01
7.61248404e-03 -6.64772630e-01 -5.74319482e-01 -1.05753744e+00
5.28324097e-02 -7.11381793e-01 6.88478529e-01 -3.80887270e-01
-1.30419612e+00 4.98302400e-01 1.64910465e-01 -9.14555490e-01
-4.37434256e-01 -8.45143497e-01 -3.32585990e-01 8.02942693e-01
-1.24673986e+00 -1.07428563e+00 1.00384749e-01 2.36417577e-01
3.51235777e-01 -2.38851994e-01 8.02635133e-01 5.40254414e-01
-1.74946964e-01 5.33065617e-01 8.01440895e-01 3.24629009e-01
1.16182137e+00 -1.50307739e+00 6.30801201e-01 9.93552029e-01
4.32272553e-01 1.13924348e+00 1.12766242e+00 -1.03639519e+00
-1.20429838e+00 -3.23084086e-01 1.80545294e+00 -7.54371941e-01
7.76114941e-01 -4.84575987e-01 -9.62682426e-01 4.70597506e-01
1.07453108e-01 -8.43127787e-01 5.86767673e-01 5.14553130e-01
-3.87069613e-01 1.55199647e-01 -1.00679517e+00 6.89506769e-01
7.22524047e-01 -5.36111712e-01 -1.06587088e+00 3.40023398e-01
-3.90176252e-02 -2.15988502e-01 -9.47287560e-01 -1.22649767e-01
9.20120239e-01 -6.94724977e-01 7.55294383e-01 -3.47615719e-01
4.19062108e-01 -2.83125341e-01 -3.67621221e-02 -6.83075726e-01
-2.05897614e-01 -7.28628695e-01 1.33034214e-01 1.92926502e+00
5.47384441e-01 -3.70362371e-01 7.15268135e-01 5.32485545e-01
3.75890248e-02 -7.37351738e-03 -9.51785266e-01 -1.19260430e+00
4.69559550e-01 -6.89016104e-01 4.84764814e-01 1.01001787e+00
-2.13938385e-01 1.55249521e-01 -3.14031094e-01 -2.52764344e-01
6.28034830e-01 -3.62837687e-02 4.63912040e-01 -1.66447699e+00
6.20501824e-02 -5.66014826e-01 -2.96356797e-01 -2.64871359e-01
-9.49238613e-03 -7.90027440e-01 1.88910086e-02 -1.69343674e+00
-5.77793419e-02 -5.66032529e-01 2.60473251e-01 3.37085247e-01
-6.78659305e-02 2.61315256e-01 6.09443197e-03 3.63182873e-01
3.10215533e-01 -6.67196810e-02 5.67759395e-01 1.13633454e-04
-1.96370810e-01 1.30916853e-02 -3.61744046e-01 6.91580772e-01
6.06384993e-01 -9.80095446e-01 2.59959042e-01 -3.74691755e-01
7.69183338e-01 -4.91648138e-01 -1.92645505e-01 -7.16142237e-01
3.21301371e-01 -1.91115350e-01 5.48169494e-01 -8.02119553e-01
-2.34108288e-02 -4.93877262e-01 1.91336453e-01 4.36486185e-01
1.05665892e-01 3.97610754e-01 1.63062796e-01 2.88994014e-01
-1.10737100e-01 -1.12711263e+00 6.27876103e-01 -4.02695894e-01
-7.00828314e-01 -1.97508395e-01 -7.05620885e-01 -4.29406390e-02
5.81435740e-01 -5.52670240e-01 -2.93150634e-01 -3.22079696e-02
-5.36209047e-01 -6.53876588e-02 8.38677585e-01 3.66514623e-01
1.90054879e-01 -1.22994387e+00 -1.04334688e+00 -3.25906247e-01
1.67990282e-01 -5.01587152e-01 -4.02706504e-01 2.09597424e-01
-1.28790569e+00 3.71652901e-01 -2.49777436e-01 -2.24944651e-01
-1.51707315e+00 2.54519165e-01 -3.44440401e-01 -9.40036178e-02
-5.62835038e-01 7.36860812e-01 -1.04005933e+00 -1.80142313e-01
2.58532107e-01 -1.74127266e-01 -3.81708354e-01 9.26907182e-01
6.54265404e-01 6.80294991e-01 1.86325788e-01 -6.72495961e-01
-5.24811447e-01 6.31898701e-01 -4.04284567e-01 -8.58126640e-01
1.40692139e+00 -5.36424033e-02 -4.20560479e-01 8.14716995e-01
7.02648282e-01 6.29880190e-01 -2.63019830e-01 -1.76937163e-01
8.51448596e-01 -7.74770498e-01 -1.35231122e-01 -7.23965824e-01
-4.91921961e-01 6.66818559e-01 2.38806963e-01 3.13757509e-01
6.87505484e-01 -2.03991145e-01 6.11900508e-01 4.35695022e-01
4.44954216e-01 -1.96326423e+00 -5.30126512e-01 7.45813191e-01
5.46226382e-01 -8.02856445e-01 6.24058366e-01 -6.98242262e-02
-2.13352188e-01 1.49989748e+00 6.37597544e-03 2.29163602e-01
3.77199531e-01 3.41255009e-01 1.46645114e-01 -2.91251037e-02
-2.10544646e-01 -1.77314416e-01 1.06137186e-01 4.76440519e-01
1.05041599e+00 -2.89029270e-01 -1.50209236e+00 6.02335572e-01
-7.22073913e-01 -7.65693113e-02 9.44897234e-01 1.32522714e+00
-5.34379005e-01 -1.84927571e+00 -9.39150572e-01 4.26499516e-01
-1.09034669e+00 -2.85131395e-01 -7.70206332e-01 1.09480071e+00
-2.02833093e-03 7.20469713e-01 6.14130683e-03 2.55527478e-02
-4.19931412e-02 3.79187822e-01 8.42149436e-01 -5.08060694e-01
-8.04979503e-01 3.06480885e-01 6.63453996e-01 1.75156906e-01
-4.07431215e-01 -1.32498181e+00 -1.18917501e+00 -1.01066637e+00
-3.30767959e-01 4.72092003e-01 1.08206940e+00 5.66889822e-01
-4.04361099e-01 3.61126587e-02 6.42330870e-02 -5.07575989e-01
-3.27527285e-01 -1.00263965e+00 -9.48383033e-01 -9.42848921e-02
-1.65439114e-01 -3.54555935e-01 -2.93642491e-01 3.85680050e-01] | [10.216482162475586, 10.263041496276855] |
e16d1fd4-4517-4065-995c-d436c5bf4f34 | large-scale-cloze-test-dataset-created-by | 1711.03225 | null | http://arxiv.org/abs/1711.03225v3 | http://arxiv.org/pdf/1711.03225v3.pdf | Large-scale Cloze Test Dataset Created by Teachers | Cloze tests are widely adopted in language exams to evaluate students'
language proficiency. In this paper, we propose the first large-scale
human-created cloze test dataset CLOTH, containing questions used in
middle-school and high-school language exams. With missing blanks carefully
created by teachers and candidate choices purposely designed to be nuanced,
CLOTH requires a deeper language understanding and a wider attention span than
previously automatically-generated cloze datasets. We test the performance of
dedicatedly designed baseline models including a language model trained on the
One Billion Word Corpus and show humans outperform them by a significant
margin. We investigate the source of the performance gap, trace model
deficiencies to some distinct properties of CLOTH, and identify the limited
ability of comprehending the long-term context to be the key bottleneck. | ['Qizhe Xie', 'Zihang Dai', 'Guokun Lai', 'Eduard Hovy'] | 2017-11-09 | large-scale-cloze-test-dataset-created-by-1 | https://aclanthology.org/D18-1257 | https://aclanthology.org/D18-1257.pdf | emnlp-2018-10 | ['cloze-test'] | ['natural-language-processing'] | [-3.09868246e-01 6.49667680e-02 -1.57194152e-01 -7.16916099e-02
-1.14493060e+00 -9.60518539e-01 3.44424874e-01 5.62495947e-01
-4.30632412e-01 5.80604911e-01 4.32000041e-01 -1.23173237e+00
-3.27745497e-01 -7.75833070e-01 -6.82735980e-01 2.00489581e-01
4.12977517e-01 7.57203773e-02 3.22160631e-01 -5.23152292e-01
5.14475882e-01 1.37473214e-02 -1.59348583e+00 4.33517516e-01
1.54288912e+00 1.11131601e-01 2.55342603e-01 9.09861088e-01
-1.83157057e-01 1.19301140e+00 -8.70797515e-01 -6.79455936e-01
-2.14684024e-01 -7.64733970e-01 -9.25615489e-01 -3.46434340e-02
1.11198199e+00 -1.46421179e-01 -5.51569015e-02 9.77517307e-01
5.12739897e-01 2.97776222e-01 2.94718176e-01 -4.27108943e-01
-1.32017589e+00 8.86096716e-01 -2.58889049e-01 5.61503828e-01
7.68122494e-01 5.61111927e-01 1.06172144e+00 -7.26327717e-01
6.32865787e-01 9.59506571e-01 6.44545555e-01 9.08821583e-01
-9.44733024e-01 -5.63037932e-01 2.83389866e-01 1.55135021e-01
-1.09718192e+00 -2.92196602e-01 3.78164858e-01 -7.11556256e-01
8.48286688e-01 3.00897062e-01 8.16612899e-01 1.16428399e+00
-3.20838019e-02 7.33436942e-01 1.53352416e+00 -8.39064240e-01
-1.27660083e-02 4.49385077e-01 6.12530649e-01 8.64198387e-01
5.90055704e-01 -2.70113498e-01 -6.06144190e-01 1.11393824e-01
2.74326712e-01 -5.28411388e-01 -4.38849926e-01 -4.55943905e-02
-8.95704091e-01 8.39034677e-01 1.63979307e-01 3.43516052e-01
2.88300782e-01 -2.09758684e-01 2.03814119e-01 6.32568896e-01
2.03994326e-02 1.23960876e+00 -3.79210353e-01 -6.36595309e-01
-9.27848697e-01 4.21620041e-01 7.63354599e-01 1.03263879e+00
2.86733340e-02 -1.02087572e-01 -5.72708011e-01 6.31187320e-01
2.08635658e-01 3.03413570e-01 8.75641167e-01 -5.67082703e-01
7.72561371e-01 7.34950066e-01 -1.33141592e-01 -5.28771758e-01
3.52506228e-02 -7.07216024e-01 -1.98823735e-02 5.76255806e-02
9.12618458e-01 -3.42620304e-03 -8.03313673e-01 1.59392977e+00
-1.78696904e-02 -1.30720418e-02 -1.44877344e-01 6.37334168e-01
1.25593305e+00 9.49821249e-02 5.07563770e-01 2.15018690e-01
1.53578544e+00 -9.91583765e-01 -7.97134936e-01 -5.88923872e-01
1.03254175e+00 -8.66947412e-01 2.05120778e+00 6.39262557e-01
-1.53056037e+00 -7.15011239e-01 -1.36006117e+00 -5.13694406e-01
-4.40444380e-01 -8.72460678e-02 5.57168841e-01 1.14520681e+00
-1.00612009e+00 5.53469896e-01 -4.94943500e-01 -5.77249639e-02
1.92550406e-01 -3.60450037e-02 -7.71583021e-02 -3.61035854e-01
-1.21382272e+00 8.73304009e-01 -3.30864973e-02 -3.78752470e-01
-8.84900153e-01 -1.01192498e+00 -1.00522947e+00 1.95332378e-01
2.84568608e-01 -3.72116566e-01 1.51693070e+00 -2.60485262e-01
-1.46536326e+00 1.27966738e+00 -8.51159990e-02 1.43298864e-01
7.14389741e-01 -5.09795785e-01 -2.89361775e-01 -2.50887841e-01
1.88894019e-01 -7.84791335e-02 1.57111213e-01 -6.20312810e-01
-3.65131110e-01 1.34965524e-01 3.21615338e-02 -1.66431144e-02
-5.65108001e-01 1.24301851e-01 -3.18245381e-01 -4.56274390e-01
8.41565579e-02 -5.45527756e-01 2.13810615e-03 -7.38333523e-01
-2.11983621e-01 -6.64247513e-01 -1.15293544e-03 -8.00206661e-01
2.08862638e+00 -2.11553407e+00 -3.69768649e-01 1.15187906e-01
6.16240382e-01 4.15352762e-01 -3.78643066e-01 3.56794208e-01
-3.25740501e-02 6.66682184e-01 4.37229097e-01 -1.89656600e-01
2.92838246e-01 -4.88481730e-01 -2.45948553e-01 3.93858910e-01
6.07289374e-02 1.05141759e+00 -1.43286383e+00 -3.66697222e-01
-9.26082134e-02 -1.02597840e-01 -7.39807904e-01 4.43404108e-01
-2.63168335e-01 2.47757554e-01 -4.49432939e-01 6.21426165e-01
3.47629547e-01 -2.27773905e-01 -1.44313186e-01 8.52007151e-01
-1.36774495e-01 1.27706897e+00 -7.19464242e-01 1.69722044e+00
-5.98058581e-01 9.07048702e-01 -3.87188792e-01 -1.16492800e-01
7.95834184e-01 2.47279316e-01 -3.81702036e-01 -8.59466851e-01
1.35070518e-01 3.17220628e-01 4.68757808e-01 -8.52060199e-01
5.76952934e-01 -1.88549664e-02 -9.33153927e-02 4.18139994e-01
-1.21492080e-01 -5.93869388e-01 6.73184574e-01 3.63592386e-01
1.37128556e+00 -6.09416850e-02 3.02684009e-01 -4.98554468e-01
4.90530968e-01 9.66734141e-02 1.48941636e-01 1.11123276e+00
-3.75963330e-01 5.71416080e-01 4.57890540e-01 -2.09206473e-02
-8.96175623e-01 -9.55010414e-01 -1.43009499e-01 1.55414557e+00
-3.01621079e-01 -9.25656915e-01 -8.15300822e-01 -8.37095022e-01
-3.17808948e-02 1.17457962e+00 -5.94230950e-01 -3.78992766e-01
-6.39851153e-01 -1.95891693e-01 5.98717153e-01 4.59886849e-01
1.50197908e-01 -1.02700961e+00 -4.27582026e-01 4.37813587e-02
3.86791863e-02 -1.10604727e+00 -5.49352348e-01 5.39555326e-02
-5.87869942e-01 -9.44028497e-01 -4.13082868e-01 -1.11412585e+00
4.71331924e-01 2.19783857e-01 1.78425539e+00 7.67893672e-01
-4.24692482e-02 2.88292974e-01 -3.18655133e-01 -5.76773107e-01
-4.87543792e-01 3.11514288e-01 -2.61729151e-01 -9.58143711e-01
1.18707168e+00 -2.38383159e-01 -3.78416687e-01 -9.26349685e-02
-7.06018448e-01 -2.07499042e-03 6.00863159e-01 6.13522112e-01
-1.40663870e-02 -3.39188695e-01 4.09580290e-01 -1.07663584e+00
1.38272321e+00 -4.44067121e-01 -6.54667318e-01 3.75492215e-01
-6.37346447e-01 -6.49791658e-02 4.88359988e-01 -7.70989776e-01
-5.72614908e-01 -7.49121666e-01 -2.78920352e-01 6.64965808e-02
-1.74396902e-01 7.09544599e-01 -2.32566386e-01 -4.26517725e-02
1.07912087e+00 -4.22068164e-02 -5.62165976e-01 -5.75981975e-01
2.57942766e-01 5.69999814e-01 5.11183023e-01 -1.23545372e+00
9.71614122e-01 -4.54018682e-01 -4.41349655e-01 -5.09630203e-01
-1.43863571e+00 -5.89057922e-01 -4.40116644e-01 5.81088886e-02
6.85505390e-01 -1.29397690e+00 -1.03540611e+00 -3.68310791e-03
-1.04807830e+00 -7.28499234e-01 -1.67638823e-01 6.32997751e-01
1.42709934e-03 3.82529013e-02 -1.02160633e+00 -5.43479025e-01
1.50846615e-01 -1.11265492e+00 5.56111038e-01 4.33044374e-01
-6.98922873e-01 -1.08471584e+00 3.54136229e-01 7.99492836e-01
2.43184879e-01 -6.80214241e-02 1.09952188e+00 -9.12930310e-01
-5.33504546e-01 -1.39348164e-01 3.82369757e-02 2.71745622e-01
-2.09852010e-01 2.99266800e-02 -9.76340234e-01 -2.62847573e-01
-1.07431985e-01 -7.89032757e-01 7.88718164e-01 -3.74246575e-02
1.01635718e+00 -1.19634971e-01 1.21037357e-01 6.56807482e-01
1.23643029e+00 -2.45105088e-01 3.84716004e-01 7.66109705e-01
6.88673973e-01 5.79369187e-01 1.09996513e-01 -9.38063338e-02
4.48278129e-01 4.70855460e-02 -2.20504642e-01 3.48029673e-01
-4.94033158e-01 -8.52685690e-01 5.62345445e-01 1.46230495e+00
2.58903593e-01 -2.59059101e-01 -1.15325069e+00 1.00372040e+00
-1.09733605e+00 -7.43597388e-01 -5.82932591e-01 2.23612761e+00
1.28951609e+00 5.54823637e-01 -6.30249903e-02 6.65923804e-02
3.48547965e-01 -1.24800108e-01 -2.39097625e-02 -7.58806705e-01
2.94659287e-02 6.36994958e-01 3.20153445e-01 9.29353476e-01
-6.94642067e-01 1.26124132e+00 7.00675058e+00 8.23158443e-01
-8.07885528e-01 6.59349561e-02 2.91899025e-01 -1.94498226e-01
-8.10342133e-01 -1.29250437e-01 -1.12391329e+00 2.83598304e-01
1.12342811e+00 -2.40838796e-01 2.75066495e-01 7.64020622e-01
1.45410568e-01 1.25224039e-01 -1.23276162e+00 5.59511065e-01
1.79004550e-01 -8.99747133e-01 -1.35256872e-01 5.73454536e-02
1.20344698e+00 -2.85670012e-01 4.46626127e-01 7.48459101e-01
7.43361890e-01 -1.45722437e+00 1.01652467e+00 3.01035363e-02
8.00874114e-01 -3.69900286e-01 4.36454207e-01 2.52903730e-01
-7.15366125e-01 -3.37372869e-02 -4.10965711e-01 -7.70186305e-01
-5.61342120e-01 4.35542017e-01 -8.89715791e-01 -1.59560397e-01
3.52053076e-01 2.85389185e-01 -1.49823046e+00 1.22565997e+00
-9.36550081e-01 1.29436612e+00 2.58751422e-01 -6.62174582e-01
5.83190799e-01 3.97563912e-02 2.52825975e-01 1.26597154e+00
3.76651764e-01 2.95099150e-02 1.18210778e-01 1.03792000e+00
-2.17013225e-01 2.77530044e-01 -4.19867247e-01 -2.70167828e-01
6.88126683e-01 9.64983523e-01 -3.90287459e-01 -2.28710711e-01
-6.84801042e-01 6.27865434e-01 6.07957423e-01 3.90083045e-01
-5.39035738e-01 -5.34473002e-01 6.20661974e-01 5.60820282e-01
1.81814805e-02 -1.88650325e-01 -7.32086837e-01 -1.21855593e+00
1.06493220e-01 -1.38805747e+00 3.00698549e-01 -6.88901424e-01
-1.39751589e+00 3.99755299e-01 -4.53850269e-01 -1.00669432e+00
-1.30582243e-01 -9.32135761e-01 -9.57607865e-01 1.39744496e+00
-1.61826050e+00 -7.20537424e-01 -5.27788877e-01 3.66029978e-01
5.38237989e-01 1.45585435e-02 5.38597524e-01 2.03228325e-01
-5.43375254e-01 1.19533563e+00 -1.28321961e-01 4.24916506e-01
1.06804836e+00 -1.75719118e+00 5.93486369e-01 1.00371182e+00
4.57037985e-03 1.22117555e+00 8.68603826e-01 -8.11983347e-01
-1.03791916e+00 -6.56979680e-01 1.54587245e+00 -1.32930303e+00
1.02093172e+00 -5.31252146e-01 -1.16179180e+00 5.68066061e-01
4.14304107e-01 -2.29565114e-01 9.47453618e-01 5.21877527e-01
-6.16278768e-01 3.41265678e-01 -6.30551755e-01 7.39574134e-01
1.14405632e+00 -8.06805432e-01 -1.23863029e+00 2.12165877e-01
9.03579712e-01 -7.52287805e-01 -6.29756391e-01 4.76151630e-02
3.01506788e-01 -7.88237333e-01 7.19628036e-01 -1.14512074e+00
1.01212060e+00 9.88705754e-02 4.34852511e-01 -1.32310128e+00
-5.07686734e-01 -8.33779037e-01 1.36530429e-01 1.24075842e+00
5.20613790e-01 -1.13500006e-01 8.88584852e-01 8.63744259e-01
-3.13738227e-01 -6.31640553e-01 -4.42464381e-01 -6.93476200e-01
8.98066700e-01 -4.30878639e-01 6.25675321e-01 1.17145157e+00
4.95805919e-01 4.68650162e-01 2.74485737e-01 -1.36074558e-01
2.75610209e-01 -1.93963930e-01 6.44477963e-01 -1.14110672e+00
-1.82878777e-01 -8.86478364e-01 1.60785075e-02 -1.06636202e+00
9.92568657e-02 -8.56606364e-01 3.51578407e-02 -1.30153811e+00
2.62843043e-01 -3.42409074e-01 -1.93955973e-01 1.98615223e-01
-1.05843019e+00 4.94168587e-02 1.45572126e-01 -4.60247278e-01
-8.40347469e-01 -2.02332307e-02 1.70503616e+00 6.51185513e-02
-3.36634308e-01 -1.31513149e-01 -1.34284568e+00 7.26028442e-01
6.24013007e-01 -7.09372237e-02 -5.08003831e-01 -7.78231323e-01
6.30551696e-01 -3.57288808e-01 3.60476300e-02 -9.64750111e-01
3.82980764e-01 -5.21320641e-01 4.05422032e-01 3.19745652e-02
-4.54146117e-01 -1.35685503e-01 -8.19916070e-01 1.40567645e-01
-8.21262956e-01 4.10394639e-01 5.69765806e-01 -1.84905931e-01
-2.89746374e-01 -6.54141903e-01 5.25566638e-01 -5.43844759e-01
-4.17512625e-01 -1.83796182e-01 -3.29056859e-01 7.60480523e-01
6.40701592e-01 -1.28829041e-02 -6.17950320e-01 -3.68443549e-01
-3.68931890e-01 2.37422809e-01 3.63539606e-01 5.53012311e-01
3.32771361e-01 -1.16734815e+00 -9.66103256e-01 1.07203640e-01
2.99556583e-01 -3.35218981e-02 1.42047226e-01 3.82393867e-01
-7.99947262e-01 8.81215572e-01 1.82180986e-01 -1.52297374e-02
-1.18474019e+00 5.39204001e-01 1.12039246e-01 -7.15543509e-01
-3.97844613e-01 1.25072467e+00 -1.59667239e-01 -3.98940653e-01
3.27660948e-01 -6.58669829e-01 -2.93226004e-01 -1.44339010e-01
1.08503175e+00 3.55342358e-01 5.06028652e-01 -8.46089721e-02
2.03494221e-01 4.34015483e-01 7.29774265e-03 5.22800675e-03
9.04800594e-01 -6.26359433e-02 2.03176692e-01 3.94800633e-01
7.97611594e-01 1.15186810e+00 -7.10478723e-01 -1.54665366e-01
4.89922702e-01 -4.68877017e-01 3.01954281e-02 -1.24098754e+00
-3.90423089e-01 1.00942111e+00 6.99914247e-02 1.28949553e-01
6.63613617e-01 -1.42688826e-01 6.14289284e-01 3.27622920e-01
-7.67622665e-02 -1.01156914e+00 2.30542049e-01 8.08881402e-01
4.78634864e-01 -1.30363798e+00 1.33383041e-02 -4.60597903e-01
-3.70174170e-01 9.11037087e-01 1.35422361e+00 6.23391233e-02
1.60728261e-01 6.66759610e-02 3.39029223e-01 -8.72276649e-02
-9.63578224e-01 -4.06973481e-01 8.69732440e-01 2.97265053e-01
1.29350793e+00 -1.04247786e-01 -6.14099503e-01 1.13420904e+00
-8.61457169e-01 -1.65870816e-01 7.98489630e-01 9.01148498e-01
-5.90449750e-01 -1.05458128e+00 -5.76514065e-01 2.59342670e-01
-7.42082477e-01 -6.69336021e-01 -6.33245826e-01 8.17359090e-01
2.07479849e-01 1.13067782e+00 -3.27037796e-02 -3.07766348e-01
4.61528629e-01 5.19574404e-01 5.62545717e-01 -1.18657351e+00
-1.12353027e+00 -3.98466051e-01 -1.40960515e-01 -1.98052257e-01
4.14780080e-01 -4.12643433e-01 -8.94913077e-01 -5.85041583e-01
-2.34940231e-01 4.95691925e-01 2.91509122e-01 1.06217182e+00
-1.63992450e-01 8.34695876e-01 -3.59057449e-02 2.94168234e-01
-9.63279426e-01 -1.40627444e+00 -3.17860603e-01 6.43775105e-01
5.97068489e-01 -2.40606382e-01 -5.74112117e-01 -1.27307490e-01] | [10.202585220336914, 7.638521671295166] |
b0eb6589-dce2-4482-ba6e-90b3e9595395 | cem-commonsense-aware-empathetic-response | 2109.05739 | null | https://arxiv.org/abs/2109.05739v2 | https://arxiv.org/pdf/2109.05739v2.pdf | CEM: Commonsense-aware Empathetic Response Generation | A key trait of daily conversations between individuals is the ability to express empathy towards others, and exploring ways to implement empathy is a crucial step towards human-like dialogue systems. Previous approaches on this topic mainly focus on detecting and utilizing the user's emotion for generating empathetic responses. However, since empathy includes both aspects of affection and cognition, we argue that in addition to identifying the user's emotion, cognitive understanding of the user's situation should also be considered. To this end, we propose a novel approach for empathetic response generation, which leverages commonsense to draw more information about the user's situation and uses this additional information to further enhance the empathy expression in generated responses. We evaluate our approach on EmpatheticDialogues, which is a widely-used benchmark dataset for empathetic response generation. Empirical results demonstrate that our approach outperforms the baseline models in both automatic and human evaluations and can generate more informative and empathetic responses. | ['Minlie Huang', 'Chujie Zheng', 'Sahand Sabour'] | 2021-09-13 | null | null | null | null | ['empathetic-response-generation'] | ['natural-language-processing'] | [-2.25544170e-01 4.72806811e-01 -3.35432822e-03 -5.79488814e-01
-2.42313921e-01 -4.13700134e-01 6.92990661e-01 1.37024149e-01
-3.59186202e-01 8.63132477e-01 8.66513431e-01 3.13565046e-01
3.58548015e-01 -6.96095109e-01 5.73698401e-01 -2.77899891e-01
6.87552750e-01 4.61359084e-01 -4.87855315e-01 -8.72469246e-01
5.23092747e-01 4.24465865e-01 -1.00611353e+00 5.74734569e-01
1.05836380e+00 5.19070089e-01 -4.05152380e-01 8.13455701e-01
-7.24506900e-02 1.62789273e+00 -9.52026367e-01 -9.37555134e-01
-2.37536177e-01 -1.10309088e+00 -1.40969837e+00 -1.81614935e-01
-3.96782160e-01 -6.54801428e-01 1.12562664e-01 7.73145437e-01
6.94097638e-01 4.79540765e-01 6.36404932e-01 -1.42832470e+00
-4.98593330e-01 7.87014842e-01 -1.47392273e-01 -1.11871891e-01
1.10870588e+00 2.48921096e-01 8.38176489e-01 -5.16054928e-01
3.24585587e-01 1.33514118e+00 6.98686123e-01 1.31545496e+00
-7.52319098e-01 -6.06938362e-01 -3.61794025e-01 4.45648059e-02
-5.33251166e-01 -4.93282169e-01 1.11024129e+00 -4.74159390e-01
5.32208085e-01 3.73156101e-01 7.81526804e-01 1.50361454e+00
-3.43309700e-01 9.45918560e-01 1.35030484e+00 -4.24725085e-01
1.93503380e-01 5.97033679e-01 3.32676858e-01 1.81947619e-01
-5.32979012e-01 -8.78856629e-02 -6.21447384e-01 -5.30958116e-01
5.56101799e-01 -3.73134941e-01 -3.47643495e-01 3.38461846e-01
-1.21136618e+00 1.23569679e+00 3.60823750e-01 3.48353714e-01
-9.96881843e-01 -9.87269282e-02 7.25187302e-01 2.22939283e-01
3.20942938e-01 1.12083328e+00 2.70157725e-01 -1.03194094e+00
-5.17707705e-01 4.25061584e-01 1.30625319e+00 5.63569605e-01
2.74344474e-01 -2.44053498e-01 -6.00588858e-01 1.21328163e+00
-1.27713140e-02 1.45409778e-01 4.70831484e-01 -1.40252185e+00
3.86493579e-02 9.12080824e-01 5.70223868e-01 -9.59455192e-01
-4.16476041e-01 6.64098188e-02 -6.10745549e-01 1.01601616e-01
3.90997231e-01 -6.16809547e-01 2.16165289e-01 1.88181984e+00
5.04379094e-01 -5.00977278e-01 4.48351800e-01 1.27235186e+00
1.14803648e+00 4.55264479e-01 3.32620680e-01 -1.50894597e-01
1.41108453e+00 -1.05960989e+00 -7.02784538e-01 -1.42960876e-01
8.14357936e-01 -8.44811559e-01 1.08215690e+00 3.06923658e-01
-1.39087391e+00 -2.80593663e-01 -5.43513954e-01 -1.50970116e-01
1.68233678e-01 2.56148905e-01 8.78480673e-01 4.79614258e-01
-6.59928620e-01 6.96680784e-01 5.84541298e-02 -7.32986271e-01
8.66758749e-02 8.93968791e-02 -3.53791475e-01 3.19103330e-01
-1.69346786e+00 1.26069701e+00 -7.84841031e-02 -5.26069924e-02
9.08058435e-02 -5.53114295e-01 -6.17765605e-01 3.11762467e-03
-3.56399775e-01 -9.52942371e-01 1.69926810e+00 -1.18432438e+00
-2.08753967e+00 1.16685665e+00 1.24146134e-01 -3.43652040e-01
7.26654947e-01 -1.62772506e-01 -1.77416623e-01 5.19209206e-01
-3.59662026e-02 1.00764310e+00 3.45788032e-01 -1.00017440e+00
-2.92192996e-01 -8.35349336e-02 4.61224407e-01 6.03394568e-01
-4.54455763e-01 7.27105498e-01 4.36264753e-01 -4.21454877e-01
-4.67550486e-01 -8.39189112e-01 -1.24409236e-01 -8.58580507e-03
-3.82856071e-01 -5.44668019e-01 4.07282293e-01 -6.22706592e-01
9.77952659e-01 -1.68514884e+00 -7.09841326e-02 -7.25614326e-03
4.62188005e-01 3.64803284e-01 2.36999355e-02 1.17749965e+00
4.23309244e-02 -1.84744939e-01 4.99559268e-02 -3.75850141e-01
2.40482733e-01 -1.68337524e-01 -5.06084383e-01 9.49821174e-02
-1.62365902e-02 1.02323318e+00 -1.09950221e+00 -6.32607341e-01
3.28275599e-02 3.96543205e-01 -4.91554886e-01 7.92136729e-01
2.27122128e-01 8.38646770e-01 -6.06716990e-01 8.54543149e-02
3.30979168e-01 -6.10052869e-02 -2.25477546e-01 2.11423650e-01
8.72046724e-02 4.61736321e-01 -3.83048207e-01 1.04078293e+00
-6.41182244e-01 4.91943926e-01 7.12341582e-03 -4.57193047e-01
1.19957137e+00 5.95791698e-01 3.61481577e-01 -5.08917868e-01
4.40435141e-01 -9.33026299e-02 1.50611624e-01 -6.43855691e-01
6.65382683e-01 -5.80247045e-01 -4.36498702e-01 1.38972151e+00
-5.96404016e-01 -4.41113949e-01 -1.13141917e-01 6.23767734e-01
9.67882514e-01 -2.83260178e-03 7.16297984e-01 1.20629184e-01
7.71241009e-01 5.39886951e-02 2.94781774e-01 5.23795605e-01
-7.19295740e-01 1.65875688e-01 7.94811130e-01 -3.12773943e-01
-5.63244641e-01 -4.62592334e-01 4.10812765e-01 1.25370407e+00
7.88040981e-02 -2.19025850e-01 -1.13641739e+00 -4.18984443e-01
-3.73040050e-01 1.14466882e+00 -8.03273439e-01 -4.16868150e-01
-4.02102858e-01 -2.82913834e-01 9.47272718e-01 5.89793265e-01
6.93149686e-01 -1.77109694e+00 -1.12307096e+00 2.41966560e-01
-9.13001359e-01 -1.02480459e+00 -3.41259271e-01 -5.77198803e-01
-5.31914234e-01 -7.66610265e-01 -6.14404082e-01 -3.28191519e-01
5.18449306e-01 3.06036204e-01 1.12307703e+00 3.72271448e-01
1.14189172e-02 5.95767021e-01 -7.41943538e-01 -3.88681114e-01
-9.42283690e-01 -3.30276042e-01 -2.13365823e-01 -2.57347941e-01
5.62144578e-01 -5.15476584e-01 -7.21993268e-01 3.79227638e-01
-3.87647837e-01 4.76552188e-01 2.19909713e-01 8.55560780e-01
-5.38100362e-01 -7.55835116e-01 9.91061568e-01 -9.24326837e-01
1.70161402e+00 -6.29347086e-01 5.39370835e-01 1.45120574e-02
-9.16678458e-02 -3.18929613e-01 8.28017652e-01 -5.21163166e-01
-1.46706188e+00 -2.88890868e-01 -3.15444082e-01 1.44756868e-01
-3.45715225e-01 7.06607252e-02 2.43946701e-01 -1.54159650e-01
8.47258866e-01 -2.39657104e-01 3.01133007e-01 7.18777105e-02
4.47213680e-01 1.03287363e+00 7.19200671e-01 -9.94868159e-01
2.72402257e-01 3.53879750e-01 -4.12106603e-01 -6.10708773e-01
-9.68672335e-01 -6.74391508e-01 -3.29068780e-01 -6.90267146e-01
6.32200837e-01 -7.24796414e-01 -1.50976312e+00 4.31983113e-01
-1.70764923e+00 -3.24766010e-01 -1.18143529e-01 5.20665646e-01
-9.42309141e-01 5.99538982e-01 -1.05369008e+00 -1.06077826e+00
-9.01774824e-01 -6.88467205e-01 8.17397833e-01 4.42875475e-01
-1.66107237e+00 -1.08236110e+00 3.68671387e-01 8.97970557e-01
5.38655519e-01 2.29485884e-01 6.09847128e-01 -9.93302584e-01
5.13176858e-01 -4.32292730e-01 -2.34456003e-01 1.23885237e-01
3.12327361e-03 -2.04418540e-01 -9.65170562e-01 3.72649431e-01
4.10828859e-01 -1.05930543e+00 1.46576554e-01 -3.57433856e-01
6.11122668e-01 -6.13326490e-01 1.22866347e-01 5.32775233e-03
4.16939169e-01 -1.10565007e-01 8.47278655e-01 -2.88756173e-02
2.30091467e-01 1.43694246e+00 1.02329171e+00 1.11198306e+00
8.09894562e-01 5.72014511e-01 2.03296408e-01 -1.32349670e-01
2.94863760e-01 -2.71976441e-01 4.29768145e-01 5.64292610e-01
-2.06741959e-01 -4.52915430e-02 -5.51219583e-01 2.93920904e-01
-1.89614522e+00 -1.48737204e+00 -4.59562719e-01 1.76097703e+00
1.10410666e+00 -5.50497353e-01 5.14409125e-01 2.54627857e-02
7.80906498e-01 3.07383984e-02 -2.72147745e-01 -1.17377055e+00
9.84131843e-02 2.39818111e-01 -5.37069142e-01 4.84785378e-01
-4.67631042e-01 1.14247680e+00 5.80298376e+00 1.32887319e-01
-9.36985970e-01 -5.74422255e-02 4.40594763e-01 4.53156084e-02
-2.64664024e-01 -1.32133320e-01 -1.69288859e-01 3.55869919e-01
6.15112901e-01 -4.33354944e-01 1.90140277e-01 8.84075105e-01
5.50160766e-01 -1.44581020e-01 -1.21474612e+00 9.53873396e-01
1.61188871e-01 -6.36662126e-01 -4.03355449e-01 -2.31513306e-01
3.57407421e-01 -9.37799037e-01 -2.26877287e-01 5.10208189e-01
7.04114139e-01 -1.08326697e+00 2.20173717e-01 5.87370455e-01
2.48957932e-01 -7.79637694e-01 9.83297706e-01 5.66287398e-01
-3.45604062e-01 1.74244061e-01 -1.40047491e-01 -6.36515558e-01
4.38630164e-01 4.53970134e-02 -1.14348197e+00 -7.46052042e-02
2.22090840e-01 5.23388863e-01 -7.09055364e-02 5.32574654e-01
-7.84014285e-01 3.61238509e-01 9.35990512e-02 -5.07662416e-01
9.24592987e-02 -3.22490901e-01 4.08788979e-01 1.21731842e+00
1.18889332e-01 6.77717507e-01 -6.09966293e-02 1.15911412e+00
-1.47246361e-01 4.54956919e-01 -4.68752623e-01 -2.45988831e-01
7.09426105e-01 1.72470236e+00 -1.78993508e-01 -3.25941265e-01
1.43829957e-01 1.33808160e+00 4.95098054e-01 -1.43882427e-02
-9.98492301e-01 -3.09919149e-01 6.13302052e-01 -3.54547918e-01
-5.80808818e-01 5.43275654e-01 -3.70852917e-01 -8.24925244e-01
-2.25292623e-01 -1.08188450e+00 3.24643582e-01 -1.34147274e+00
-1.42059040e+00 4.86627847e-01 -3.98519099e-01 -1.04634833e+00
-8.20859313e-01 -2.06257522e-01 -1.35390007e+00 9.68586683e-01
-9.74473476e-01 -1.38607872e+00 -8.14513206e-01 4.05310512e-01
2.25710019e-01 2.95695096e-01 1.09729671e+00 -2.81532854e-01
-3.36311370e-01 6.62086606e-01 -9.53130841e-01 2.36843571e-01
1.30216646e+00 -1.03886497e+00 1.62501141e-01 2.83083856e-01
-6.43440664e-01 8.65023911e-01 9.69335973e-01 -3.25616986e-01
-7.50820279e-01 -4.32628334e-01 1.25868821e+00 -4.59502310e-01
7.26591170e-01 1.59821764e-01 -8.85096431e-01 3.52553308e-01
5.94827235e-01 -6.61839366e-01 1.32504845e+00 2.74186973e-02
-2.74834037e-01 5.92116654e-01 -1.58773625e+00 1.00805116e+00
9.67491686e-01 -5.10863423e-01 -1.03319156e+00 3.46445143e-01
2.35103562e-01 -2.55740613e-01 -1.07743847e+00 -3.45311053e-02
5.81982374e-01 -1.53185081e+00 7.25081861e-01 -9.19018686e-01
1.03621066e+00 3.38664085e-01 4.32131290e-01 -1.67190063e+00
-3.22814509e-02 -1.07973289e+00 1.78172275e-01 1.34402311e+00
-1.19407415e-01 -6.96671426e-01 7.35267222e-01 1.29799962e+00
-2.79836711e-02 -6.72135293e-01 -2.93718040e-01 -1.10613173e-02
1.82111278e-01 -1.38875306e-01 7.50919819e-01 1.20803702e+00
1.22224784e+00 6.88701034e-01 -8.56569111e-01 -4.85873520e-01
1.63524076e-01 2.82674611e-01 1.31229401e+00 -1.33715951e+00
-2.11313948e-01 -7.89810121e-01 -1.66554973e-01 -1.00799096e+00
6.46329820e-01 -4.91421312e-01 8.49851891e-02 -1.50899148e+00
3.48625332e-01 -1.39831424e-01 4.35654908e-01 3.70592833e-01
-4.10816163e-01 9.64672789e-02 2.56756097e-01 1.78820193e-01
-5.91335714e-01 7.22611487e-01 1.33687901e+00 5.00558853e-01
-2.38696873e-01 1.24535598e-01 -1.30971682e+00 8.53578866e-01
9.73735988e-01 -1.80299848e-01 -2.96671808e-01 2.78517246e-01
1.97203398e-01 6.20445430e-01 5.56854665e-01 -5.47614396e-01
4.96410042e-01 -5.08704543e-01 3.26496325e-02 -2.45646149e-01
8.59452188e-01 -2.26432383e-01 -2.86655575e-01 3.92471880e-01
-7.86827743e-01 1.77943334e-02 -2.52894253e-01 -3.43307108e-02
-3.00918907e-01 -3.71185154e-01 9.90958512e-01 -2.68095076e-01
-3.57033163e-01 -2.66191363e-01 -5.79711914e-01 1.91672996e-01
1.20481372e+00 -3.89233440e-01 -6.74361467e-01 -1.39272392e+00
-3.17348301e-01 3.95782739e-01 7.70368099e-01 2.63659567e-01
7.24304020e-01 -1.30771995e+00 -8.89283240e-01 -4.26537305e-01
3.38177294e-01 -7.12549567e-01 5.14280558e-01 1.07990849e+00
-3.76038939e-01 5.36985286e-02 -4.89100635e-01 1.03604905e-01
-1.53881443e+00 8.21715146e-02 4.53766257e-01 -2.62410998e-01
-4.56453323e-01 9.16596472e-01 1.32048726e-01 -6.04559124e-01
-9.59633142e-02 4.22558248e-01 -7.20040619e-01 2.19497845e-01
7.52212286e-01 5.47127366e-01 -5.62700629e-01 -9.38875735e-01
-5.95208779e-02 1.50277779e-01 -4.49392274e-02 -3.21158171e-01
1.01310205e+00 -3.21863033e-02 -4.81095850e-01 2.73543715e-01
8.12138379e-01 3.18034917e-01 -6.71032548e-01 3.53436358e-02
-9.48554799e-02 -5.64675748e-01 -6.89371109e-01 -8.68933499e-01
-4.22259480e-01 1.12024081e+00 -4.87502187e-01 1.46610245e-01
9.20732200e-01 -9.30579379e-02 1.21460426e+00 3.76600593e-01
4.01619703e-01 -1.34084892e+00 6.99278176e-01 5.01245618e-01
1.16764402e+00 -1.28660381e+00 -6.68306574e-02 -3.95022303e-01
-1.62955546e+00 1.30156004e+00 1.06385088e+00 8.85942876e-02
-1.86213642e-01 -9.78350341e-02 6.00865960e-01 -1.05833359e-01
-1.00747716e+00 1.10731497e-01 -3.70689780e-02 5.59911549e-01
8.42226028e-01 1.23554140e-01 -6.24416590e-01 9.70084846e-01
-8.65897000e-01 -3.22605930e-02 9.88404810e-01 5.71917832e-01
-2.71652311e-01 -1.15612614e+00 -2.85168082e-01 9.95187089e-02
-3.14682513e-01 1.23389177e-01 -1.49881089e+00 4.94195133e-01
-5.44094682e-01 1.38152528e+00 -3.36097956e-01 -3.39255720e-01
2.90767789e-01 9.35495421e-02 3.37733686e-01 -6.91418290e-01
-1.36126137e+00 -7.01749980e-01 6.80849373e-01 -5.35319805e-01
-5.03004789e-01 -5.38940787e-01 -1.49122214e+00 -8.01360905e-01
2.20396705e-02 4.65561867e-01 3.68718833e-01 1.02085221e+00
1.96419954e-01 -1.24852367e-01 8.69551420e-01 -5.59469104e-01
-9.74532545e-01 -1.14403069e+00 -8.01239088e-02 1.06868768e+00
-2.57059693e-01 -3.47290784e-01 -5.68301976e-01 -4.31161582e-01] | [13.153701782226562, 7.626633644104004] |
8e1ebf8e-ee6b-4daf-95da-8cf480f0e368 | msanii-high-fidelity-music-synthesis-on-a | 2301.06468 | null | https://arxiv.org/abs/2301.06468v1 | https://arxiv.org/pdf/2301.06468v1.pdf | Msanii: High Fidelity Music Synthesis on a Shoestring Budget | In this paper, we present Msanii, a novel diffusion-based model for synthesizing long-context, high-fidelity music efficiently. Our model combines the expressiveness of mel spectrograms, the generative capabilities of diffusion models, and the vocoding capabilities of neural vocoders. We demonstrate the effectiveness of Msanii by synthesizing tens of seconds (190 seconds) of stereo music at high sample rates (44.1 kHz) without the use of concatenative synthesis, cascading architectures, or compression techniques. To the best of our knowledge, this is the first work to successfully employ a diffusion-based model for synthesizing such long music samples at high sample rates. Our demo can be found https://kinyugo.github.io/msanii-demo and our code https://github.com/Kinyugo/msanii . | ['Kinyugo Maina'] | 2023-01-16 | null | null | null | null | ['audio-inpainting', 'music-generation', 'music-generation'] | ['audio', 'audio', 'music'] | [-1.42150834e-01 -3.06065649e-01 3.88652682e-02 4.10898894e-01
-8.16823125e-01 -6.80744708e-01 4.64497000e-01 -3.67872119e-01
8.36634915e-03 5.90541482e-01 6.40740454e-01 -2.78225243e-01
-1.47368833e-01 -6.35953724e-01 -5.33584118e-01 -2.59991944e-01
-3.21787477e-01 4.96739000e-02 4.66027856e-02 -2.90510118e-01
6.32418245e-02 1.87730074e-01 -1.52289581e+00 4.14960533e-01
6.39667094e-01 7.89543867e-01 4.60603416e-01 1.27339780e+00
3.84518921e-01 8.59288573e-01 -6.63681865e-01 -1.00541994e-01
3.81696641e-01 -8.92870426e-01 -3.98393363e-01 -4.00145173e-01
2.86120117e-01 -6.20956242e-01 -8.77647758e-01 8.41257632e-01
7.86740363e-01 1.94237441e-01 5.25613129e-01 -8.56026053e-01
-7.95730531e-01 1.12311554e+00 5.03055379e-02 3.37146699e-01
3.16720486e-01 4.94554639e-01 1.05186749e+00 -9.22550321e-01
6.92137361e-01 9.01705265e-01 6.31899357e-01 5.03309548e-01
-9.98386681e-01 -1.03155506e+00 -4.19914275e-01 7.57175311e-02
-1.60742402e+00 -8.34107578e-01 8.02232623e-01 -3.70477140e-01
1.04863620e+00 3.51991653e-01 9.51003790e-01 1.46025658e+00
1.34028941e-02 8.70537996e-01 7.13201582e-01 -2.53037244e-01
2.49983415e-01 -4.35594946e-01 -3.40926200e-01 3.93721163e-01
-2.03713208e-01 4.17931646e-01 -9.09088373e-01 -1.21585988e-01
1.14658618e+00 -1.40547320e-01 -3.65538478e-01 3.41954947e-01
-1.29593813e+00 5.87613225e-01 2.37380758e-01 5.24442792e-01
-4.51028228e-01 8.06179404e-01 3.81402314e-01 5.61100304e-01
2.52696753e-01 5.73709905e-01 1.13476263e-02 -8.90824437e-01
-1.34710658e+00 2.17159227e-01 7.72779167e-01 1.09821582e+00
3.63389850e-02 8.01434994e-01 -3.95263452e-03 8.50670993e-01
-6.70392811e-02 6.30929291e-01 7.66579568e-01 -1.57137299e+00
3.15195084e-01 -3.19510669e-01 -1.30474061e-01 -6.37775660e-01
-3.40494923e-02 -6.01296008e-01 -9.14396703e-01 -2.65263226e-02
3.35055292e-01 -4.39433753e-01 -4.96374935e-01 1.51077044e+00
-1.61765814e-01 4.81933415e-01 8.00991803e-02 8.84193540e-01
6.13447785e-01 9.62242186e-01 -5.72870433e-01 -5.80451041e-02
8.93300176e-01 -1.02015030e+00 -6.39972091e-01 2.42360786e-01
1.88200131e-01 -1.16985273e+00 1.13093555e+00 7.13553667e-01
-1.38234234e+00 -5.78319669e-01 -1.19569325e+00 -1.49758309e-01
1.34219810e-01 1.58225209e-01 6.27430201e-01 4.34666663e-01
-1.05553031e+00 1.02277458e+00 -1.00250542e+00 -1.15339458e-01
2.88856596e-01 7.51433223e-02 2.55640686e-01 2.68957525e-01
-1.13376629e+00 1.90004095e-01 1.49945512e-01 -2.16358453e-01
-1.24691486e+00 -8.44325542e-01 -3.43585968e-01 1.27889276e-01
1.66190818e-01 -6.52086914e-01 1.56432974e+00 -8.61250579e-01
-1.86851406e+00 -3.90724316e-02 4.39085402e-02 -7.54990339e-01
4.70122457e-01 -4.02981907e-01 -6.26282573e-01 1.83078498e-01
-3.96411508e-01 5.38633943e-01 9.04210269e-01 -8.31725180e-01
-2.48550668e-01 1.94382682e-01 -1.12060279e-01 -6.33822903e-02
-5.14157832e-01 -3.53514969e-01 -2.38207802e-01 -1.21427381e+00
-2.67922074e-01 -1.07070923e+00 4.91890609e-02 -2.87705958e-01
-3.67382556e-01 2.06110314e-01 7.53986478e-01 -5.97240627e-01
1.74610627e+00 -2.64970708e+00 9.93077382e-02 -6.41468018e-02
5.16130291e-02 2.32533783e-01 -3.97628218e-01 9.19378519e-01
1.39885843e-01 1.62473619e-01 -1.10288836e-01 -2.96519786e-01
9.24247131e-02 -8.81928280e-02 -7.36018062e-01 1.56099617e-01
-1.12536289e-01 9.52202976e-01 -8.94268870e-01 3.28011923e-02
1.06979378e-01 6.89822555e-01 -8.17127764e-01 -5.35256192e-02
-1.62776053e-01 4.32255507e-01 -4.61901911e-02 5.36489725e-01
6.49873912e-02 -1.57824650e-01 1.67216763e-01 -2.63461292e-01
-3.47575992e-01 7.20507801e-01 -1.13153207e+00 2.13654280e+00
-6.00494564e-01 9.22197938e-01 -4.59485799e-02 -1.29220828e-01
7.81767368e-01 5.33337176e-01 4.93172586e-01 -5.15143573e-01
1.05657071e-01 6.27655089e-01 1.67089865e-01 1.16192633e-02
6.38692141e-01 -1.20551303e-01 2.05184259e-02 6.70784235e-01
2.23459274e-01 -5.34191370e-01 3.39473099e-01 2.61214107e-01
1.05455339e+00 8.52013901e-02 1.21291317e-01 -1.21762335e-01
-1.63556099e-01 -1.70268759e-01 1.92598090e-01 7.66282856e-01
8.23858008e-02 8.05003703e-01 -1.84030402e-02 1.77133922e-03
-1.59194005e+00 -1.19365740e+00 4.79365587e-02 8.26859117e-01
-2.74958640e-01 -9.79723454e-01 -6.50331020e-01 2.09714323e-01
-5.47232516e-02 8.16883504e-01 -1.12499073e-01 2.83471495e-02
-5.02142847e-01 -1.27333194e-01 1.18571341e+00 4.04601157e-01
2.81762689e-01 -9.86946225e-01 -4.52691019e-01 4.36194837e-01
-1.67797133e-01 -5.60644090e-01 -9.12914395e-01 -1.52723677e-02
-9.81537282e-01 -5.71288645e-01 -9.64484453e-01 -3.46967161e-01
-2.51705766e-01 4.37061302e-02 9.34530616e-01 -5.84185161e-02
-2.85578936e-01 1.48070544e-01 -3.78487140e-01 -3.42995882e-01
-5.83986461e-01 2.85660356e-01 3.00246984e-01 -3.46068263e-01
-1.82577148e-01 -1.29720318e+00 -7.76737273e-01 9.61426273e-02
-9.97614861e-01 2.75934100e-01 3.80661100e-01 5.37302732e-01
4.59700972e-01 7.01898895e-03 5.67855239e-01 -5.04259288e-01
8.66055489e-01 -4.34049696e-01 -5.25215864e-01 -4.46889281e-01
-3.63165200e-01 -1.32612631e-01 7.16079712e-01 -7.37103462e-01
-5.31607568e-01 -3.06343853e-01 -4.44571257e-01 -7.20525622e-01
9.69924703e-02 4.07187700e-01 4.65212435e-01 3.63019288e-01
9.00489748e-01 2.90384024e-01 -6.63396269e-02 -6.33006096e-01
6.49557292e-01 1.03243041e+00 5.94217062e-01 -4.93953645e-01
5.54565251e-01 2.96289712e-01 -3.91809404e-01 -8.89114201e-01
-2.33912781e-01 -1.26008853e-01 -2.35004276e-01 -1.57479391e-01
3.91681403e-01 -1.07334065e+00 -5.85199893e-01 5.57149291e-01
-7.94199407e-01 -9.53090012e-01 -7.12831795e-01 6.98495090e-01
-8.44910145e-01 3.17923993e-01 -1.28872883e+00 -8.07537258e-01
-6.17347300e-01 -7.69954264e-01 6.68335378e-01 4.97078113e-02
-6.23795986e-01 -6.53055131e-01 3.86130549e-02 6.75885156e-02
7.16692626e-01 -1.72159951e-02 4.53230917e-01 -1.47721142e-01
-8.21150661e-01 1.37209609e-01 3.14587295e-01 3.17263395e-01
1.13601089e-01 2.81541616e-01 -1.01613832e+00 -3.04135770e-01
-5.64773977e-02 -3.50185633e-01 7.73253143e-01 3.66490722e-01
9.51661825e-01 -5.06824911e-01 2.28842407e-01 8.10842454e-01
1.26207411e+00 4.39991713e-01 7.57090449e-01 5.95329553e-02
4.81983542e-01 -1.52898207e-01 2.60950208e-01 9.46127415e-01
1.03846267e-01 6.10181510e-01 2.03995347e-01 2.22130090e-01
-7.98672318e-01 -5.56834936e-01 5.61745048e-01 1.83492446e+00
-3.72922838e-01 -4.32990730e-01 -7.56811678e-01 6.96003497e-01
-1.52831137e+00 -1.33423960e+00 -1.00459168e-02 2.01410794e+00
1.00319421e+00 -3.50893065e-02 3.15112233e-01 3.26024354e-01
3.55498821e-01 3.97807986e-01 -6.08663976e-01 -4.67027128e-01
-5.92883751e-02 2.99472988e-01 4.07016695e-01 6.13112509e-01
-7.76528537e-01 7.58518457e-01 6.03722095e+00 1.26044440e+00
-1.31041694e+00 2.82656550e-01 2.32946858e-01 -9.80423987e-01
-4.98629808e-01 -1.25845313e-01 -4.61011350e-01 6.41110241e-01
1.47256386e+00 -3.66107762e-01 1.11884797e+00 4.63020116e-01
3.18529844e-01 4.05976176e-01 -8.17971826e-01 1.07557285e+00
-1.21230632e-01 -1.58812523e+00 -3.21556590e-02 2.19354481e-01
8.81073177e-01 3.75121415e-01 5.10627806e-01 3.34648974e-02
3.57590228e-01 -9.35072601e-01 1.14918125e+00 5.13839066e-01
1.16957116e+00 -8.02705646e-01 8.52735713e-02 2.38511994e-01
-1.20212185e+00 -1.49013996e-01 -1.09997414e-01 -2.39335626e-01
3.60459566e-01 6.36755705e-01 -6.26419961e-01 4.07175422e-01
4.85181749e-01 1.03908181e+00 -7.48398528e-02 9.88694727e-01
-2.09626462e-02 1.08899128e+00 -4.39050138e-01 -4.15871367e-02
2.04760060e-01 -9.16265771e-02 7.28241861e-01 1.32973170e+00
1.02686548e+00 1.61478147e-01 -2.20325366e-01 8.40672314e-01
-1.40259102e-01 -1.19318120e-01 -6.43474936e-01 -5.51836431e-01
7.09785938e-01 7.29349971e-01 -1.62022904e-01 -4.03967530e-01
-2.21963152e-02 1.09589136e+00 -1.89204946e-01 4.48774844e-01
-8.53829384e-01 -5.01338184e-01 7.05524504e-01 1.08009316e-01
4.77506399e-01 -7.83675253e-01 -3.01466018e-01 -1.27907586e+00
-2.01092452e-01 -1.22528672e+00 1.13041058e-01 -1.01629817e+00
-1.04611731e+00 7.24300981e-01 -3.30371767e-01 -1.52997613e+00
-6.04407549e-01 -7.12886499e-03 -2.94493735e-01 7.18588352e-01
-8.90994191e-01 -7.09549189e-01 8.81382227e-02 5.84034085e-01
7.77950287e-01 -3.47147912e-01 8.75540257e-01 4.71682161e-01
-5.77729084e-02 4.10455048e-01 5.16635895e-01 -5.44583537e-02
5.55740118e-01 -9.74476755e-01 7.96107769e-01 6.78315043e-01
4.93880242e-01 6.13405406e-01 6.94153249e-01 -5.86501539e-01
-1.51783693e+00 -9.28046823e-01 6.66072547e-01 -2.41659135e-01
9.77188110e-01 -3.82744551e-01 -7.46008396e-01 5.55243015e-01
6.04552269e-01 -4.74749923e-01 8.47446203e-01 -7.57642016e-02
-3.70991588e-01 -1.11771338e-01 -4.92230713e-01 9.38188374e-01
1.28671813e+00 -9.26345229e-01 -2.60521203e-01 1.62365615e-01
8.40066671e-01 -5.30417681e-01 -1.29393518e+00 -2.62807049e-02
8.41196239e-01 -9.78987515e-01 1.01596916e+00 9.83156487e-02
7.28300035e-01 -3.05814207e-01 -3.79792839e-01 -1.26360440e+00
-3.57701808e-01 -1.08344352e+00 -4.52785581e-01 1.17844760e+00
5.75779438e-01 -5.26577711e-01 3.55867118e-01 -9.47896689e-02
-2.75392264e-01 -4.80947137e-01 -8.49015594e-01 -1.05254662e+00
3.32642496e-02 -7.20756233e-01 7.52051950e-01 7.97534525e-01
2.32644618e-01 1.12503342e-01 -8.11936438e-01 9.66878200e-04
3.45276356e-01 2.36025631e-01 7.13395953e-01 -5.58824122e-01
-9.72603440e-01 -6.49621189e-01 -3.02469917e-03 -1.22103727e+00
-5.03611624e-01 -9.97629464e-01 -1.51547909e-01 -1.26549721e+00
-3.07468057e-01 -4.81224686e-01 -2.49533758e-01 2.70921916e-01
4.77425814e-01 4.69527870e-01 8.25843871e-01 6.18476510e-01
-1.89681470e-01 6.30232215e-01 1.27904832e+00 2.49594823e-02
-3.85642022e-01 -2.18298644e-01 -5.73035121e-01 3.69234353e-01
1.16583896e+00 -5.23019493e-01 -6.02495432e-01 -6.43450618e-01
8.04401562e-02 3.87489855e-01 2.12276727e-01 -1.46856868e+00
3.65114212e-01 4.52131927e-02 7.84948468e-02 -3.21711451e-01
7.55505562e-01 -2.85608202e-01 8.59772086e-01 5.92209816e-01
-3.81727606e-01 1.21841967e-01 3.89554381e-01 3.21269572e-01
-3.40069711e-01 3.00292410e-02 4.46506649e-01 -1.01337641e-01
-4.45701182e-01 1.55737191e-01 -5.80391407e-01 4.70279753e-02
4.15226221e-01 1.59061760e-01 -3.68168503e-01 -7.24236727e-01
-7.13810623e-01 -4.17110622e-01 5.66337705e-01 4.66047764e-01
6.84305370e-01 -1.51221812e+00 -6.94073319e-01 1.91164792e-01
-2.47025982e-01 -4.73145723e-01 3.02137375e-01 7.28314698e-01
-7.64741123e-01 4.32909250e-01 -1.99268669e-01 -2.12616384e-01
-1.07323992e+00 3.80548835e-01 9.25932378e-02 1.55347511e-01
-8.51403236e-01 7.14767873e-01 -8.60817283e-02 2.04837158e-01
1.87693292e-03 -4.13312256e-01 3.56859893e-01 -1.64674103e-01
5.96085310e-01 4.87710834e-01 -2.02835336e-01 -2.49935910e-01
-5.56033254e-02 3.25864792e-01 3.91016692e-01 -6.96671009e-01
1.17113006e+00 -3.29263508e-02 2.36598000e-01 8.15467536e-01
9.85474110e-01 4.28777367e-01 -1.23397863e+00 6.77905679e-02
-5.35725057e-01 -4.36593294e-01 4.31134216e-02 -7.30101228e-01
-8.84595394e-01 9.93389368e-01 3.96645546e-01 1.02236733e-01
1.25071704e+00 -1.65714100e-01 1.25536966e+00 1.72199622e-01
4.24658537e-01 -8.62147331e-01 2.69409537e-01 6.49440110e-01
1.04810643e+00 -2.78066635e-01 -3.09834778e-01 7.98600614e-02
-6.74048781e-01 1.08861578e+00 -7.10183829e-02 -4.57676083e-01
5.99385023e-01 6.30760014e-01 1.69865578e-01 1.28910914e-01
-1.13044345e+00 -9.73594859e-02 1.46700740e-01 3.74074787e-01
5.05881190e-01 3.14775527e-01 -3.01951259e-01 4.73467946e-01
-8.05170834e-01 2.21498907e-01 6.84487283e-01 6.41377449e-01
-2.90807843e-01 -1.04388571e+00 -2.87301600e-01 2.22210944e-01
-3.93273681e-01 -6.32296860e-01 -1.47936359e-01 3.38139117e-01
-1.37841567e-01 9.91619051e-01 3.39025743e-02 -7.07836270e-01
9.75267068e-02 1.15586482e-01 3.67269903e-01 -3.07626903e-01
-7.28499472e-01 5.94034553e-01 2.36243993e-01 -4.94207025e-01
-1.46849334e-01 -5.65469444e-01 -1.06449723e+00 -7.16367126e-01
1.19110599e-01 2.87135001e-02 7.00417459e-01 2.30819046e-01
7.53159285e-01 8.07293177e-01 4.46342826e-01 -8.73258591e-01
-5.02136588e-01 -9.75064337e-01 -7.55959332e-01 1.87407866e-01
4.35658723e-01 -1.81608051e-02 -2.50314385e-01 1.17512077e-01] | [15.572113990783691, 5.7970356941223145] |
44e33e07-deb9-43bd-a345-35cf73869346 | characterizing-the-influence-of-features-on | 1808.09718 | null | http://arxiv.org/abs/1808.09718v1 | http://arxiv.org/pdf/1808.09718v1.pdf | Characterizing the Influence of Features on Reading Difficulty Estimation for Non-native Readers | In recent years, the number of people studying English as a second language
(ESL) has surpassed the number of native speakers. Recent work have
demonstrated the success of providing personalized content based on reading
difficulty, such as information retrieval and summarization. However, almost
all prior studies of reading difficulty are designed for native speakers,
rather than non-native readers. In this study, we investigate various features
for ESL readers, by conducting a linear regression to estimate the reading
level of English language sources. This estimation is based not only on the
complexity of lexical and syntactic features, but also several novel concepts,
including the age of word and grammar acquisition from several sources, word
sense from WordNet, and the implicit relation between sentences. By employing
Bayesian Information Criterion (BIC) to select the optimal model, we find that
the combination of the number of words, the age of word acquisition and the
height of the parsing tree generate better results than alternative competing
models. Thus, our results show that proposed second language reading difficulty
estimation outperforms other first language reading difficulty estimations. | ['Yeali S. Sun', 'Yi-Ting Huang', 'Meng Chang Chen'] | 2018-08-29 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [-3.84678207e-02 2.21872821e-01 -2.84846663e-01 -2.10282534e-01
-8.57265413e-01 -5.05572855e-01 4.84301984e-01 9.85808611e-01
-9.04901385e-01 3.21165413e-01 7.99350202e-01 -6.30813956e-01
-2.00582117e-01 -5.62258601e-01 -2.13303864e-01 2.22765580e-02
4.28795367e-01 9.71281305e-02 5.13683379e-01 -4.01113123e-01
8.70813191e-01 -1.45220801e-01 -1.27736545e+00 -1.17815144e-01
1.51900852e+00 5.97230017e-01 8.80785763e-01 4.50301975e-01
-5.82583189e-01 4.72460985e-01 -6.60398960e-01 -4.99213099e-01
-3.52598757e-01 -4.55555528e-01 -9.56750154e-01 -2.75470227e-01
3.54166240e-01 -3.83951396e-01 -2.88252309e-02 9.13726687e-01
6.02504671e-01 3.10672875e-02 7.61479795e-01 -4.29522514e-01
-7.81818569e-01 1.04976094e+00 -1.64656430e-01 7.04765081e-01
7.21113920e-01 -2.89888680e-01 1.10744524e+00 -6.60604417e-01
4.56928104e-01 1.40401876e+00 3.80235046e-01 3.06601584e-01
-7.01797843e-01 -2.90795356e-01 5.92620075e-01 3.79486561e-01
-9.20114458e-01 -3.22681785e-01 6.47874594e-01 -4.83419180e-01
1.06015337e+00 -3.12538780e-02 5.59985459e-01 9.07565653e-01
3.16225797e-01 6.50935531e-01 1.40452909e+00 -1.06740236e+00
6.06960282e-02 2.74698406e-01 7.74224520e-01 5.35288036e-01
4.83556956e-01 -3.48766148e-01 -9.49007094e-01 6.86292499e-02
2.03360900e-01 -5.19481897e-01 -2.22605586e-01 3.46133232e-01
-1.00337660e+00 8.82321656e-01 -1.71181217e-01 5.17735004e-01
-1.11766636e-01 -5.35037398e-01 2.34161660e-01 3.19210112e-01
5.25933146e-01 7.48125374e-01 -8.24544191e-01 -5.13845503e-01
-7.45176077e-01 7.87814632e-02 9.45078194e-01 7.67783344e-01
2.93274343e-01 -2.52123922e-01 -1.74970686e-01 1.06776845e+00
6.12874210e-01 5.93506098e-01 1.03203857e+00 -4.13229406e-01
7.52242088e-01 6.09397233e-01 -2.00227648e-01 -1.02396691e+00
-5.08362353e-01 -4.13170516e-01 -3.61310363e-01 -4.76657629e-01
6.06577814e-01 -1.74613789e-01 -6.17070556e-01 1.95473599e+00
-4.20396738e-02 -7.40530968e-01 -4.48735170e-02 4.44809645e-01
7.89389193e-01 6.41895711e-01 4.59979475e-01 -5.76068580e-01
1.54244792e+00 -6.36203527e-01 -8.29090416e-01 -6.39090359e-01
6.32638693e-01 -8.95463765e-01 1.26357222e+00 4.24397916e-01
-1.42609978e+00 -7.21251011e-01 -1.01856911e+00 -2.78552711e-01
-4.16390955e-01 -7.48296902e-02 3.53468806e-01 9.08554494e-01
-8.61177802e-01 2.96069384e-01 -6.90964162e-01 -5.52286386e-01
-1.41618833e-01 -7.58386180e-02 -1.52390420e-01 -3.75116952e-02
-1.42694259e+00 1.27823746e+00 4.57620770e-01 -2.57775933e-01
-1.21027350e-01 -5.02513051e-01 -8.67989242e-01 1.26862466e-01
1.74934417e-01 -4.30076599e-01 9.34324265e-01 -7.52061367e-01
-1.52649522e+00 7.03969955e-01 -2.90618867e-01 7.06353039e-02
3.62643488e-02 -5.89905083e-01 -3.18959743e-01 1.22899860e-01
2.88624138e-01 2.28525266e-01 6.39285386e-01 -6.43741429e-01
-7.70441234e-01 -5.21073878e-01 1.15374826e-01 6.63017273e-01
-4.20821428e-01 2.51365721e-01 -2.88391739e-01 -5.90181053e-01
2.87335485e-01 -3.35409254e-01 4.05722782e-02 -3.50000262e-01
-1.69189736e-01 -7.96519458e-01 -9.84160379e-02 -1.33413720e+00
1.79098797e+00 -2.38187099e+00 3.06061864e-01 2.45203212e-01
3.08046997e-01 8.23387429e-02 -9.88786817e-02 5.50976515e-01
4.20839071e-01 6.08144462e-01 2.63979077e-01 4.02924279e-03
6.65780753e-02 -2.65962094e-01 1.53846249e-01 9.28460062e-03
-3.62204798e-02 6.86449826e-01 -8.50341678e-01 -5.75425565e-01
-1.25700474e-01 -7.86394696e-04 -4.10569072e-01 1.72307849e-01
1.12234876e-02 -4.66732867e-03 -4.65600759e-01 2.77179241e-01
3.47463280e-01 -1.53228119e-01 2.77495742e-01 3.54802936e-01
-2.02254966e-01 1.00355482e+00 -9.44686055e-01 1.34405422e+00
-6.93796337e-01 4.94467020e-01 -2.07533970e-01 -5.66134810e-01
7.44593799e-01 -1.72579624e-02 -2.57104427e-01 -9.11374748e-01
1.20451152e-01 1.26752675e-01 5.16525328e-01 -6.43528640e-01
4.17022705e-01 2.33309448e-01 -1.74483001e-01 3.23899835e-01
5.71747357e-03 -1.19028591e-01 6.07461512e-01 3.68803948e-01
8.03886116e-01 -1.19766571e-01 7.60685980e-01 -5.60893416e-01
4.23048168e-01 -3.76828313e-01 2.29998723e-01 9.19511855e-01
-4.00027595e-02 -7.41013288e-02 5.52885413e-01 3.20327461e-01
-8.05031121e-01 -1.27967525e+00 -2.16050833e-01 1.43667400e+00
2.76459634e-01 -7.42418110e-01 -9.98917699e-01 -2.17375442e-01
-3.04498971e-01 1.15359616e+00 -4.88158502e-02 -3.32379550e-01
-5.31482637e-01 -3.26645821e-01 1.77534044e-01 4.63609219e-01
5.70736349e-01 -9.15299892e-01 -4.44610238e-01 3.68689150e-01
-5.20635068e-01 -1.09798801e+00 -4.55121011e-01 -1.69738308e-01
-7.91164339e-01 -8.63390148e-01 -4.29527998e-01 -9.10328865e-01
3.93104762e-01 5.71541712e-02 1.12544703e+00 5.06276190e-01
8.66185278e-02 7.64741421e-01 -8.02005112e-01 -6.52125895e-01
-4.23411041e-01 6.82506740e-01 -1.07299119e-01 -7.02848613e-01
6.72607064e-01 -1.60876304e-01 -2.93722332e-01 -1.41156912e-01
-5.93374193e-01 2.48641353e-02 6.83058500e-01 4.92749333e-01
1.51021490e-02 1.37920126e-01 6.36536658e-01 -5.56294858e-01
1.11720872e+00 -5.94377875e-01 -8.93858522e-02 5.83403170e-01
-8.32555592e-01 1.09341741e-01 3.00380349e-01 -4.80704308e-01
-1.25870860e+00 -7.76544333e-01 -4.41762060e-01 9.28653061e-01
-2.01310933e-01 9.58706439e-01 -2.75039911e-01 4.07774001e-01
4.77278620e-01 1.79024085e-01 6.93607926e-02 -5.31835556e-01
8.70118663e-02 8.33909392e-01 -2.32305750e-02 -6.08128428e-01
5.33306777e-01 -5.89685261e-01 -4.54303533e-01 -1.42631912e+00
-1.16618419e+00 -4.04473335e-01 -7.10115612e-01 -2.51364380e-01
8.60016584e-01 -7.91613400e-01 -4.03330505e-01 7.97723651e-01
-9.19803202e-01 -1.54450253e-01 1.73172072e-01 7.86390543e-01
-1.05199262e-01 6.99746668e-01 -5.25888383e-01 -7.01591611e-01
-1.01030365e-01 -9.65619087e-01 7.38261402e-01 3.47627014e-01
-5.75733960e-01 -1.22211301e+00 -2.59293348e-01 4.95135665e-01
4.50583309e-01 -2.85986096e-01 1.59403169e+00 -8.46441627e-01
-2.85973728e-01 1.55303597e-01 3.69124487e-02 3.67958009e-01
1.24585005e-02 -2.92707384e-01 -1.32291451e-01 -8.75626951e-02
1.37286618e-01 -3.25312525e-01 7.08227992e-01 5.96800685e-01
7.33326852e-01 -2.06645876e-01 -1.81891322e-01 -5.74272824e-03
1.42135835e+00 -8.30909014e-02 4.77479160e-01 3.29696983e-01
5.16906261e-01 8.10882688e-01 5.27104259e-01 3.20442170e-01
9.70854700e-01 1.90133318e-01 -2.30835080e-01 3.57342333e-01
-6.43891916e-02 -3.75046015e-01 5.92276514e-01 1.56579423e+00
1.11785894e-02 -3.94694775e-01 -1.10363233e+00 5.25755584e-01
-1.07010722e+00 -5.33822358e-01 -1.40897989e-01 2.11097813e+00
1.02670586e+00 4.15565193e-01 8.15306529e-02 9.09724012e-02
4.60198700e-01 3.71553153e-02 -2.43988395e-01 -5.44632256e-01
-1.10382982e-01 2.84874022e-01 3.41395646e-01 8.74523342e-01
-5.05170584e-01 1.11204886e+00 6.64676619e+00 7.89741218e-01
-7.53362894e-01 -9.77070555e-02 4.90689605e-01 4.51182067e-01
-5.59225619e-01 -6.76109269e-02 -1.20382535e+00 4.60017622e-01
9.54535246e-01 -5.36777079e-01 3.79331797e-01 4.22989249e-01
7.08233863e-02 -7.22516477e-01 -7.83171058e-01 4.87835616e-01
3.78240854e-01 -4.09618199e-01 2.36094724e-02 -2.28330627e-01
1.41341045e-01 -3.65811646e-01 -9.21782404e-02 4.54582632e-01
5.81486989e-03 -6.68764472e-01 8.65179956e-01 5.78734756e-01
4.35910106e-01 -3.98881435e-01 6.26982927e-01 8.03502142e-01
-8.39008570e-01 -1.34517476e-01 -2.70336390e-01 -5.39116204e-01
6.78897724e-02 3.25086653e-01 -4.12234098e-01 1.45940766e-01
5.56689918e-01 3.01689208e-01 -1.18671679e+00 6.80211246e-01
-5.04654408e-01 9.45383489e-01 -3.38137597e-01 -6.20196879e-01
-1.60683870e-01 9.46951881e-02 3.92227173e-01 9.34992015e-01
4.96971488e-01 3.39944750e-01 2.16628104e-01 2.28142768e-01
2.91176736e-01 8.63993466e-01 -2.08260342e-01 -2.58423984e-01
7.30358422e-01 7.55335212e-01 -7.86176145e-01 -9.14436430e-02
-7.61332095e-01 6.01222754e-01 6.37672067e-01 9.97279137e-02
-1.85724527e-01 -4.41453725e-01 1.44421279e-01 2.10488439e-01
2.50348523e-02 -7.11067975e-01 -3.72693568e-01 -9.92999911e-01
1.72401056e-01 -8.89340341e-01 2.25821212e-01 -4.75139380e-01
-1.39737511e+00 2.96825737e-01 3.41623813e-01 -3.66810381e-01
-2.58544654e-01 -8.76309276e-01 -2.78397650e-01 1.00751340e+00
-1.24255669e+00 -5.99848866e-01 -1.12560146e-01 8.45355093e-02
9.38317299e-01 -2.69435555e-01 6.93885028e-01 -1.15842104e-01
-4.75945830e-01 7.10342050e-01 9.90522653e-02 -7.32324868e-02
6.41244650e-01 -1.44027472e+00 2.65959650e-01 8.76680434e-01
-1.02384731e-01 7.38750696e-01 4.81468260e-01 -8.36571336e-01
-8.97740006e-01 -2.12406412e-01 1.57345665e+00 -4.80352134e-01
5.93505383e-01 -3.12832773e-01 -8.07118654e-01 3.50708306e-01
6.49058163e-01 -1.25073159e+00 9.47439730e-01 5.90907454e-01
-1.37681216e-02 9.54395458e-02 -7.55229473e-01 6.73756957e-01
1.01868391e+00 -3.40785414e-01 -1.15951288e+00 6.42475709e-02
8.30863595e-01 -9.63867605e-02 -9.73877072e-01 1.26931086e-01
4.31528032e-01 -7.82151580e-01 6.53995395e-01 -2.95254916e-01
6.09917462e-01 3.09633464e-01 -1.14026982e-02 -1.57066870e+00
-5.90438902e-01 -2.24692509e-01 1.39306292e-01 1.28559291e+00
6.65396869e-01 -5.31738520e-01 6.50031343e-02 8.66112232e-01
-1.70973968e-02 -7.46380985e-01 -6.19416773e-01 -4.81806993e-01
4.80507463e-01 -2.02567026e-01 2.08560914e-01 5.47391534e-01
2.22702771e-01 8.03274572e-01 1.43607527e-01 1.01765871e-01
2.18200326e-01 -5.53135216e-01 1.71004087e-01 -1.47416246e+00
-2.61719674e-01 -5.06764412e-01 -1.54915854e-01 -1.25911415e+00
3.88194650e-01 -6.58597291e-01 -4.12149355e-02 -1.77273822e+00
1.53088376e-01 -7.84388036e-02 -3.14002708e-02 2.44812835e-02
-9.23867643e-01 -5.81583083e-01 2.83087939e-01 -8.59827623e-02
-4.56375241e-01 1.62051052e-01 1.11620092e+00 1.92224443e-01
-1.86213389e-01 7.19802380e-02 -9.07107770e-01 9.18538094e-01
1.03402424e+00 -1.95183545e-01 -6.02180600e-01 -6.69790983e-01
8.45069528e-01 1.30743951e-01 -2.25528181e-01 -7.77992725e-01
2.00171858e-01 -2.49080658e-01 4.71237391e-01 -4.62265223e-01
-1.87371925e-01 -4.10444736e-01 -6.17514014e-01 1.43553779e-01
-5.74129879e-01 4.87341762e-01 1.09352894e-01 2.36646354e-01
-1.37669325e-01 -1.04162395e+00 5.30234993e-01 -2.18733579e-01
-6.07102394e-01 -2.47833043e-01 -7.56437063e-01 6.09113038e-01
6.01104856e-01 -1.80737033e-01 -1.38929978e-01 -5.43971419e-01
-5.41260898e-01 1.06242754e-01 8.40595663e-02 5.63004971e-01
5.21304309e-01 -6.97367728e-01 -7.51990795e-01 1.44886494e-01
-1.12505317e-01 -3.01637739e-01 5.41273914e-02 5.47042370e-01
-5.42899489e-01 4.02725160e-01 -9.47269648e-02 -9.89297964e-03
-1.26484752e+00 3.62599730e-01 -1.03002146e-01 -3.58398616e-01
-2.50015080e-01 8.16870868e-01 -1.96107253e-02 -1.62725464e-01
3.45660865e-01 -2.63735116e-01 -7.10112810e-01 3.25819790e-01
5.49617946e-01 6.94390655e-01 -7.91861638e-02 -3.53909254e-01
-4.69208993e-02 5.71995199e-01 -4.87614721e-01 -2.58875579e-01
8.67546439e-01 -7.25072324e-01 -2.31596664e-01 8.31860542e-01
7.54009187e-01 5.42441070e-01 -4.96174902e-01 -3.40563506e-01
3.86816621e-01 -2.95561701e-01 9.56809819e-02 -9.56808865e-01
-1.75665811e-01 7.08027065e-01 3.43305677e-01 3.88089359e-01
9.93271828e-01 1.93248019e-01 5.30557215e-01 4.88596708e-01
2.69748986e-01 -1.62883890e+00 6.64190575e-02 9.40048456e-01
6.59211040e-01 -1.19686866e+00 5.44626378e-02 -5.58288574e-01
-4.90335315e-01 1.02262866e+00 8.47328961e-01 3.32586735e-01
8.36419582e-01 -1.96465716e-01 7.98653066e-02 8.54977518e-02
-5.08571327e-01 -2.11463749e-01 4.82348442e-01 3.68022591e-01
9.33783710e-01 1.22000594e-02 -1.20886064e+00 8.76354277e-01
-8.45104635e-01 -6.20404482e-01 4.93456036e-01 7.81109631e-01
-8.62780988e-01 -1.33542013e+00 -1.73999131e-01 8.02387297e-01
-6.61921918e-01 -6.06653392e-01 -3.55563790e-01 7.90688157e-01
-2.09495146e-02 1.20438623e+00 1.27670169e-01 1.41811714e-01
3.45568448e-01 2.82995075e-01 6.88674986e-01 -8.29314888e-01
-4.34519678e-01 -1.32495001e-01 1.12011485e-01 -2.11855583e-02
-2.98338056e-01 -8.68773103e-01 -8.52990866e-01 -2.16886133e-01
-6.46834493e-01 2.02729329e-01 6.76792443e-01 1.35937738e+00
-4.30845581e-02 2.87083983e-01 3.06233138e-01 -1.77239850e-01
-7.20428050e-01 -1.48253477e+00 -6.03964448e-01 9.86065194e-02
-6.05165698e-02 -4.57309842e-01 -5.04127026e-01 -1.09739713e-01] | [10.816519737243652, 10.30258560180664] |
0e6aab73-92c3-455e-8a45-a7af7ff366b3 | a-refined-deep-learning-architecture-for | 2007.07922 | null | https://arxiv.org/abs/2007.07922v1 | https://arxiv.org/pdf/2007.07922v1.pdf | A Refined Deep Learning Architecture for Diabetic Foot Ulcers Detection | Diabetic Foot Ulcers (DFU) that affect the lower extremities are a major complication of diabetes. Each year, more than 1 million diabetic patients undergo amputation due to failure to recognize DFU and get the proper treatment from clinicians. There is an urgent need to use a CAD system for the detection of DFU. In this paper, we propose using deep learning methods (EfficientDet Architectures) for the detection of DFU in the DFUC2020 challenge dataset, which consists of 4,500 DFU images. We further refined the EfficientDet architecture to avoid false negative and false positive predictions. The code for this method is available at https://github.com/Manugoyal12345/Yet-Another-EfficientDet-Pytorch. | ['Saeed Hassanpour', 'Manu Goyal'] | 2020-07-15 | null | null | null | null | ['diabetic-foot-ulcer-detection'] | ['medical'] | [-6.18636198e-02 -1.95330307e-01 -3.77215594e-01 -2.22997352e-01
-8.89155090e-01 -1.77596346e-01 1.20873503e-01 -1.14035290e-02
-3.62712353e-01 1.17463982e+00 1.74818367e-01 -4.66146469e-01
1.44903079e-01 -1.09844613e+00 -6.93769038e-01 -3.06264848e-01
-1.50244489e-01 3.49400878e-01 2.16655090e-01 -8.87832493e-02
7.76794702e-02 1.50543898e-01 -1.21994698e+00 7.15459943e-01
1.01959693e+00 1.06439459e+00 3.39420624e-02 8.32969844e-01
3.64076346e-01 5.85049748e-01 -1.62312552e-01 -4.15909886e-01
4.52955186e-01 -5.91254890e-01 -5.03720701e-01 -1.75425813e-01
6.32893503e-01 -1.02564478e+00 -4.65927035e-01 8.50202978e-01
9.95064676e-01 -5.50659060e-01 7.49909699e-01 -8.61641228e-01
-2.53133088e-01 1.12806320e-01 -6.75777555e-01 3.75280708e-01
1.58142269e-01 4.02074605e-01 6.45114303e-01 -6.67738616e-01
6.99557424e-01 1.23334324e+00 7.81770468e-01 5.81534684e-01
-1.18466544e+00 -5.27805388e-01 -1.68389201e-01 5.55094719e-01
-1.17711711e+00 -3.99535418e-01 9.13533643e-02 -5.22564828e-01
7.34055758e-01 1.37039945e-01 9.26595330e-01 1.59976017e+00
7.71648526e-01 1.01249564e+00 8.81750047e-01 -2.95006812e-01
3.00545573e-01 -4.08306450e-01 3.18660766e-01 7.36774206e-01
5.14334679e-01 4.01601851e-01 -8.51253122e-02 -1.78492412e-01
1.36021256e+00 -3.04392669e-02 -2.21356168e-01 -3.68128121e-01
-1.03747308e+00 9.39155757e-01 4.28012550e-01 -5.40168524e-01
-5.09894848e-01 1.47978857e-01 7.77745068e-01 1.53542995e-01
4.58627865e-02 1.20374945e-03 -5.85142076e-01 -2.64760464e-01
-1.29465193e-01 4.51255023e-01 5.25884748e-01 7.88611531e-01
2.13585541e-01 -4.26173478e-01 -1.15776479e-01 8.55460167e-01
1.28800318e-01 2.17763871e-01 2.02414125e-01 -1.07276177e+00
1.94698974e-01 7.37519562e-01 7.15137050e-02 -4.89054322e-01
-4.47704107e-01 -4.76852864e-01 -1.05982065e+00 6.95502579e-01
5.14942765e-01 -4.61925298e-01 -1.38614154e+00 1.18154180e+00
5.07642291e-02 7.99291655e-02 -3.44349295e-01 1.24134910e+00
7.17888057e-01 1.31821975e-01 1.44935563e-01 4.50781196e-01
1.38749385e+00 -7.88593590e-01 -3.31733912e-01 -4.58264679e-01
7.98550069e-01 -6.24688268e-01 9.03832376e-01 6.70423687e-01
-6.67879462e-01 -1.84427872e-01 -1.01507068e+00 -3.73872608e-01
-1.24859147e-01 4.31580126e-01 6.44002855e-01 4.24633235e-01
-6.93174183e-01 5.21334231e-01 -1.06204987e+00 -9.07423973e-01
8.22434843e-01 1.36889964e-01 -3.58474940e-01 -4.82790768e-01
-1.34621465e+00 1.04517758e+00 2.22039640e-01 3.77607420e-02
-6.33094907e-01 -5.83087802e-01 -5.28432608e-01 -3.96081448e-01
1.41358569e-01 -1.19198287e+00 1.20963156e+00 -3.98890257e-01
-1.08898187e+00 8.78801584e-01 -6.87326565e-02 -6.04470670e-01
1.26902652e+00 -7.72761881e-01 -2.71580040e-01 -8.57326165e-02
1.68424904e-01 5.13146639e-01 2.91943580e-01 -6.04889512e-01
-1.12551975e+00 -4.30473655e-01 1.61551812e-03 -1.74783736e-01
3.28538507e-01 -2.29959905e-01 -3.51937830e-01 -3.74131262e-01
-6.54724911e-02 -7.58923173e-01 -5.00523210e-01 6.04687989e-01
-7.95604706e-01 -4.65211160e-02 5.06875217e-01 -9.01812553e-01
1.06786835e+00 -1.93459892e+00 -1.07829057e-01 -1.04703475e-02
3.93828362e-01 4.01672304e-01 -1.37864769e-01 2.49537691e-01
-5.55079244e-02 -5.54572977e-02 -1.43761218e-01 2.51750499e-01
-2.66320527e-01 3.30870599e-01 1.73142135e-01 4.00747776e-01
3.94229859e-01 7.11870968e-01 -7.36205637e-01 -2.42217913e-01
6.53158128e-01 6.52537048e-01 -6.63160980e-01 -1.62705868e-01
1.18366722e-02 3.14788342e-01 -5.21856785e-01 1.00645900e+00
6.30270004e-01 -2.47984424e-01 9.11888108e-02 -3.57188880e-01
-4.65134978e-02 7.05055669e-02 -9.84115422e-01 1.43004692e+00
-3.66228782e-02 4.46448684e-01 -8.25427622e-02 -8.43111098e-01
4.80728209e-01 5.59473522e-02 5.21278262e-01 -9.01205003e-01
3.49813372e-01 5.03099263e-01 8.41824338e-02 -6.39708817e-01
-5.00764310e-01 9.54369158e-02 2.29538262e-01 -2.23386317e-01
-1.07424222e-01 7.91507840e-01 6.26422763e-01 -7.67461732e-02
1.57208252e+00 2.77680427e-01 6.57337666e-01 -2.11875767e-01
1.06392570e-01 1.77624688e-01 8.70963454e-01 8.19393754e-01
-7.19679832e-01 8.56743753e-01 7.58369565e-01 -9.69997108e-01
-9.88318980e-01 -1.48711920e+00 -7.28646517e-01 3.04420710e-01
-7.60431364e-02 -3.35614562e-01 -3.87974799e-01 -5.10040700e-01
8.03894818e-01 5.45008242e-01 -7.23031223e-01 -3.02827448e-01
-4.76483852e-01 -1.08273804e+00 3.61555994e-01 6.92336380e-01
7.22375751e-01 -5.65615475e-01 -6.16690814e-01 5.28927803e-01
-3.65720809e-01 -7.87548184e-01 2.27092523e-02 2.64259815e-01
-7.87615657e-01 -1.42283928e+00 -1.00312626e+00 -7.13974476e-01
2.83292651e-01 -1.27534971e-01 9.21437204e-01 -2.62839258e-01
-1.12666845e+00 -2.38594398e-01 -1.54546037e-01 -5.24879158e-01
-2.73886949e-01 1.91395789e-01 5.85268475e-02 -3.91713470e-01
8.26987624e-01 -5.81897855e-01 -1.15960395e+00 3.54790479e-01
-1.75632253e-01 2.56083041e-01 9.72397029e-01 6.52181983e-01
6.61453366e-01 -5.02253413e-01 8.03508937e-01 -7.46235251e-01
5.10619819e-01 -4.99443531e-01 -5.93600035e-01 -6.51588142e-02
-4.45381850e-01 -3.45528185e-01 1.59376785e-01 -1.03990898e-01
-4.33499545e-01 3.56555879e-01 -4.30530161e-01 -1.01574883e-01
-3.73165905e-01 5.70216060e-01 1.37584552e-01 1.39993116e-01
1.01387477e+00 -2.05949411e-01 2.81368613e-01 -7.31041551e-01
-3.26458476e-02 7.34005511e-01 4.73586708e-01 -1.95532501e-01
-1.47564514e-02 3.19362253e-01 -7.85039663e-02 -8.53771687e-01
-6.81836247e-01 -3.65635663e-01 -4.90723819e-01 -2.59339839e-01
6.61707938e-01 -1.19604552e+00 -4.83742535e-01 7.06502497e-01
-9.93036330e-01 -3.05611968e-01 1.04527354e-01 6.06701076e-01
-5.89784980e-01 3.74167114e-01 -9.58619714e-01 -4.08934921e-01
-4.46258456e-01 -8.98849010e-01 8.02812994e-01 9.27868187e-02
-4.99453127e-01 -5.47948003e-01 1.61143541e-01 3.25211585e-01
4.03851449e-01 6.31431699e-01 1.07537603e+00 -1.20698951e-01
-4.55515236e-01 -5.97548723e-01 -5.95090210e-01 6.43705964e-01
5.80638230e-01 1.21533714e-01 -5.79436958e-01 -1.47995025e-01
-6.67425513e-01 -1.77618608e-01 1.29670656e+00 1.05115426e+00
8.21649969e-01 1.60767853e-01 -4.65739965e-01 5.53561509e-01
1.40122569e+00 3.17165285e-01 1.14404404e+00 6.33462667e-01
4.66766894e-01 1.27253653e-02 4.35251474e-01 6.80553436e-01
1.81720078e-01 4.75843221e-01 5.05199134e-01 -4.35666561e-01
-4.07482237e-01 2.17016667e-01 9.66883078e-02 -7.80198798e-02
-5.56336045e-01 -1.09968662e-01 -1.05214143e+00 5.90540290e-01
-2.03003430e+00 -7.37846553e-01 -5.25397837e-01 2.11238098e+00
7.06343353e-01 4.31025386e-01 2.50853539e-01 -9.57849324e-02
7.77605653e-01 -5.26326180e-01 -8.60369742e-01 -1.21372983e-01
-8.19277912e-02 1.44487932e-01 6.83929443e-01 2.18096629e-01
-1.45763791e+00 5.57262182e-01 5.88815212e+00 2.53856808e-01
-1.06195688e+00 -2.00518563e-01 6.90000296e-01 -1.37217224e-01
5.33496261e-01 -4.00485694e-01 -7.37406433e-01 5.29562116e-01
5.71732938e-01 -1.79459602e-01 -3.41639258e-02 9.14683640e-01
5.03713548e-01 -3.41930509e-01 -9.53869164e-01 7.23091125e-01
-5.73993266e-01 -1.38072407e+00 2.13263426e-02 3.15104693e-01
2.14156836e-01 4.83980507e-01 -2.45002896e-01 3.15876305e-01
2.73055017e-01 -6.92961752e-01 1.99953765e-01 5.19599974e-01
8.29595923e-01 -7.09779322e-01 1.07164001e+00 -1.62272453e-01
-6.47628367e-01 -3.34838293e-02 -3.96990448e-01 -2.54376888e-01
2.78527379e-01 9.86535370e-01 -7.80432403e-01 2.07237437e-01
8.51446092e-01 9.09170091e-01 -4.97035742e-01 1.75373542e+00
-2.59676248e-01 5.58486521e-01 -4.06400859e-01 2.13147387e-01
-6.29922226e-02 -1.80419028e-01 5.19863009e-01 9.21159208e-01
4.67676252e-01 -1.97439447e-01 1.70074537e-01 5.27110577e-01
2.28134796e-01 8.32376182e-02 -4.37846839e-01 1.00337639e-01
-4.34242152e-02 7.82707334e-01 -1.89869091e-01 -1.20486356e-01
-6.82169795e-01 1.14052522e+00 -3.43568698e-02 3.06313783e-01
-8.72619629e-01 -3.61244678e-01 1.15584099e+00 7.43809342e-01
-4.98840474e-02 1.09777547e-01 -4.59837109e-01 -1.31367314e+00
1.81308255e-01 -6.59095585e-01 5.84974825e-01 -5.63630939e-01
-1.65760612e+00 1.39494523e-01 -6.60469770e-01 -1.41334796e+00
-7.26072788e-02 -1.02285862e+00 -3.94733816e-01 9.97651815e-01
-1.48533988e+00 -9.66546595e-01 -4.74346101e-01 4.00822937e-01
4.64493036e-01 7.68430978e-02 1.16690409e+00 7.30325997e-01
-8.58760953e-01 2.40586787e-01 1.68380126e-01 3.46676767e-01
1.12551332e+00 -1.11448705e+00 6.23485863e-01 6.15839124e-01
-9.72142339e-01 3.84363383e-01 6.72498941e-01 -9.94148731e-01
-1.07447088e+00 -1.32484710e+00 6.50893271e-01 -2.34215900e-01
5.47233403e-01 -4.99762297e-02 -5.64874291e-01 8.09660614e-01
-5.88259771e-02 5.86420223e-02 6.43456578e-01 1.96719587e-01
-1.37777952e-02 -9.36410353e-02 -1.15848291e+00 6.34784341e-01
1.10141993e+00 1.28756166e-01 -2.00040653e-01 5.78648269e-01
-1.29491284e-01 -4.22546387e-01 -8.44703376e-01 7.20175326e-01
1.02522218e+00 -9.24988747e-01 8.96931171e-01 -7.89774060e-01
1.03613055e+00 -4.50842865e-02 -2.15727299e-01 -1.08406162e+00
-6.91256225e-01 4.58017830e-03 -1.46960974e-01 5.40308774e-01
3.13355237e-01 -8.44198644e-01 9.96663630e-01 2.80639499e-01
-1.34814426e-01 -8.56244504e-01 -1.01541424e+00 -7.35882342e-01
1.32995188e-01 1.43878227e-02 -2.30370939e-01 6.93025410e-01
-7.24932104e-02 3.13895345e-01 -4.85682607e-01 1.72179252e-01
9.78348970e-01 -1.80278718e-01 6.74413025e-01 -1.47180271e+00
-1.37562782e-01 -3.21510255e-01 -9.70028043e-01 -6.17448449e-01
-8.10076475e-01 -6.17650509e-01 -3.63717526e-01 -2.24370074e+00
1.54773742e-01 9.98861119e-02 -6.08284414e-01 7.20602274e-01
-1.58246070e-01 1.68539926e-01 -3.39884698e-01 -1.51809648e-01
2.93677468e-02 2.67375171e-01 1.23785949e+00 -5.39314635e-02
-3.47056501e-02 4.09667313e-01 -6.36835217e-01 6.35291100e-01
1.14427924e+00 -1.49684876e-01 -7.14960918e-02 -3.83787215e-01
-7.46291503e-02 -6.85964078e-02 7.91860223e-01 -1.19736528e+00
-2.86584556e-01 -1.54347226e-01 9.38000977e-01 -6.24955893e-01
1.53368399e-01 -4.54901814e-01 7.02103525e-02 1.13226068e+00
1.02318242e-01 -4.20483321e-01 2.39598766e-01 3.37161392e-01
4.95699644e-02 2.77937859e-01 1.17048419e+00 -3.14678788e-01
-1.02846801e+00 1.75028086e-01 -8.90715599e-01 -3.78083169e-01
1.09723449e+00 -1.15268782e-01 -6.72215819e-01 -2.63611704e-01
-9.83552516e-01 5.79990089e-01 3.06423873e-01 4.95078981e-01
7.46700823e-01 -1.25625372e+00 -1.20615351e+00 1.53847605e-01
4.32957292e-01 -1.58420667e-01 4.90831047e-01 1.06525183e+00
-1.10170412e+00 3.40681672e-01 -8.60274971e-01 -5.01744866e-01
-1.16957211e+00 1.70530230e-01 7.40691483e-01 7.51098897e-03
-1.22060013e+00 6.92328215e-01 1.60643429e-01 -3.03514868e-01
3.27122897e-01 -3.28733265e-01 1.50704339e-01 -1.12589888e-01
6.37552321e-01 6.14450574e-01 2.67248899e-01 2.74714947e-01
-4.87422347e-01 2.83442676e-01 -6.53349340e-01 4.79355752e-01
1.34750509e+00 5.06228693e-02 9.79534015e-02 -5.39292805e-02
1.01976752e+00 -6.06991231e-01 -1.16554344e+00 1.30824164e-01
-2.42691040e-01 -4.61654574e-01 2.94686258e-01 -1.37516105e+00
-1.12937355e+00 7.31404245e-01 1.46692932e+00 -3.16817284e-01
8.98925126e-01 -2.15892345e-01 1.02976096e+00 2.15867415e-01
6.56956494e-01 -8.28321576e-01 -4.27825838e-01 2.92503864e-01
8.53293777e-01 -1.31071222e+00 6.12909682e-02 -7.64187098e-01
-2.52498776e-01 1.18093276e+00 6.94740176e-01 -4.30792093e-01
6.29470408e-01 2.31379062e-01 4.20527130e-01 5.07916808e-02
-6.56535804e-01 -1.69140652e-01 -2.24933252e-01 5.75630307e-01
4.95830834e-01 3.62408489e-01 -9.07535672e-01 6.76313758e-01
3.11781615e-01 8.85899484e-01 5.52715957e-01 1.39998591e+00
-5.13725221e-01 -1.24942970e+00 6.41569272e-02 1.12696421e+00
-2.59167105e-01 7.99949002e-03 -5.15447676e-01 7.62197733e-01
1.50385901e-01 4.33080226e-01 -1.51563227e-01 -3.24351341e-01
7.51146257e-01 -1.22143626e-02 6.05886519e-01 -4.03261125e-01
1.97066084e-01 1.73488662e-01 6.14149511e-01 -8.19719970e-01
1.43479452e-01 -5.73701501e-01 -1.11002839e+00 -3.31872404e-01
2.46669620e-01 -7.10371673e-01 3.45449239e-01 4.73613948e-01
6.29808128e-01 6.37231946e-01 1.33895740e-01 -4.58550334e-01
-4.38913107e-01 -9.25980210e-01 -7.63242960e-01 -8.44338387e-02
3.29355478e-01 -8.97961915e-01 -2.11011201e-01 7.64894113e-02] | [15.754172325134277, -3.8429386615753174] |
02df2b28-d714-4317-9828-dc04462d3245 | cluster-and-aggregate-face-recognition-with | 2210.10864 | null | https://arxiv.org/abs/2210.10864v3 | https://arxiv.org/pdf/2210.10864v3.pdf | Cluster and Aggregate: Face Recognition with Large Probe Set | Feature fusion plays a crucial role in unconstrained face recognition where inputs (probes) comprise of a set of $N$ low quality images whose individual qualities vary. Advances in attention and recurrent modules have led to feature fusion that can model the relationship among the images in the input set. However, attention mechanisms cannot scale to large $N$ due to their quadratic complexity and recurrent modules suffer from input order sensitivity. We propose a two-stage feature fusion paradigm, Cluster and Aggregate, that can both scale to large $N$ and maintain the ability to perform sequential inference with order invariance. Specifically, Cluster stage is a linear assignment of $N$ inputs to $M$ global cluster centers, and Aggregation stage is a fusion over $M$ clustered features. The clustered features play an integral role when the inputs are sequential as they can serve as a summarization of past features. By leveraging the order-invariance of incremental averaging operation, we design an update rule that achieves batch-order invariance, which guarantees that the contributions of early image in the sequence do not diminish as time steps increase. Experiments on IJB-B and IJB-S benchmark datasets show the superiority of the proposed two-stage paradigm in unconstrained face recognition. Code and pretrained models are available in https://github.com/mk-minchul/caface | ['Xiaoming Liu', 'Anil Jain', 'Feng Liu', 'Minchul Kim'] | 2022-10-19 | null | null | null | null | ['video-recognition'] | ['computer-vision'] | [ 1.53415009e-01 -6.10942304e-01 -8.09735805e-02 -7.46479511e-01
-6.48010790e-01 -2.16431454e-01 4.34129626e-01 -2.05059022e-01
-3.50703508e-01 3.79592121e-01 -7.98052773e-02 4.93032672e-02
-4.80849952e-01 -6.02206707e-01 -6.71136200e-01 -8.64630580e-01
-3.53097916e-01 -9.93112624e-02 -4.19113338e-02 -6.77058697e-02
3.05227399e-01 6.24143839e-01 -1.78996944e+00 4.68933076e-01
5.95516324e-01 1.49255300e+00 1.85203984e-01 6.68160439e-01
-1.08391158e-01 5.48369884e-01 -5.04595280e-01 -3.28601509e-01
3.66742373e-01 -4.11397368e-01 -6.79808497e-01 1.62801400e-01
4.38067853e-01 -1.90555453e-01 -5.01515687e-01 1.10649109e+00
6.58060491e-01 3.34911644e-01 4.08736557e-01 -1.26935852e+00
-7.39203751e-01 4.01416838e-01 -6.92760825e-01 6.52382791e-01
2.17244148e-01 1.56647906e-01 9.40331638e-01 -1.34361124e+00
2.93930888e-01 1.41500771e+00 3.54120970e-01 6.53894603e-01
-1.05476987e+00 -8.29276383e-01 5.89357078e-01 5.84389269e-01
-1.48297858e+00 -7.48918056e-01 7.75160074e-01 -2.63689518e-01
1.12488031e+00 4.40475643e-01 4.20226038e-01 6.27424300e-01
1.88656911e-01 8.43965888e-01 7.08558440e-01 -1.54380783e-01
7.03808889e-02 -9.11267623e-02 3.79917860e-01 7.19470680e-01
2.30317730e-02 -9.79999527e-02 -9.11258459e-01 3.24062333e-02
6.27871096e-01 5.77020764e-01 -4.01612043e-01 9.60147530e-02
-8.03937435e-01 6.72716141e-01 7.31436551e-01 3.89506459e-01
-6.57729268e-01 9.86294001e-02 1.44513771e-01 6.38525903e-01
2.32916981e-01 -5.96406695e-04 -4.88843292e-01 1.37176543e-01
-9.33289111e-01 -6.84461221e-02 3.46130222e-01 8.08161139e-01
1.00650358e+00 4.22375612e-02 -2.90227175e-01 9.33794022e-01
2.81281084e-01 5.74236691e-01 6.45422459e-01 -8.73150587e-01
3.95793825e-01 7.65442669e-01 -1.54124126e-01 -1.23669076e+00
-1.25585914e-01 -3.24453533e-01 -1.06074309e+00 2.56732814e-02
7.79835209e-02 4.64560986e-02 -1.02274394e+00 1.99994528e+00
2.25711375e-01 1.94497555e-01 -1.55692890e-01 6.42720401e-01
7.36718237e-01 7.36331165e-01 -9.37935859e-02 -6.06248856e-01
1.36938798e+00 -7.01005757e-01 -6.45387888e-01 -1.01760082e-01
1.23302611e-02 -7.89524555e-01 6.40735269e-01 2.11146653e-01
-1.26554787e+00 -9.18975174e-01 -9.48655307e-01 1.41993880e-01
-3.14751118e-01 8.60117152e-02 6.24404013e-01 4.20307934e-01
-1.28578413e+00 6.58055246e-01 -7.11613178e-01 -2.02823162e-01
8.10778975e-01 7.17408657e-01 -3.84068161e-01 -3.02544624e-01
-8.92999649e-01 3.51346403e-01 7.04485625e-02 6.34917736e-01
-7.83704281e-01 -6.32400513e-01 -5.75555146e-01 1.60589069e-01
4.05151814e-01 -2.80192316e-01 1.04038978e+00 -1.26345205e+00
-1.16924310e+00 5.16107678e-01 -7.44079113e-01 -8.20910782e-02
1.63122430e-01 -1.94211259e-01 -4.15358752e-01 1.94332227e-01
-6.43360754e-03 6.14776850e-01 1.15559530e+00 -8.58326316e-01
-8.67237806e-01 -6.92274630e-01 -7.49650821e-02 2.87703156e-01
-6.48851633e-01 8.79142359e-02 -7.61178553e-01 -4.40227747e-01
3.79924595e-01 -7.24977016e-01 -2.64502078e-01 2.23496519e-02
-1.18516851e-02 -4.43783373e-01 8.14557135e-01 -5.34767091e-01
1.45774460e+00 -2.34293818e+00 2.00851560e-01 3.30692828e-01
7.16152489e-02 2.36445546e-01 -3.21758598e-01 2.35876217e-01
-2.03331441e-01 1.41167700e-01 -3.90997110e-03 -3.65765750e-01
-2.28507414e-01 3.50754941e-04 -2.75713682e-01 2.85691887e-01
6.05964482e-01 8.25730383e-01 -6.37247443e-01 -3.66090506e-01
-2.00895160e-01 3.72854829e-01 -6.64973736e-01 3.05462837e-01
1.60813093e-01 1.95875555e-01 -4.86279547e-01 9.03125644e-01
8.08529258e-01 -4.80728388e-01 8.47599357e-02 -3.60316068e-01
-2.48950291e-02 -6.02387898e-02 -1.18071675e+00 1.51136613e+00
-9.32634175e-02 3.16191912e-01 1.22320481e-01 -1.15161467e+00
7.28383005e-01 2.71805316e-01 5.43555081e-01 -6.99243188e-01
1.01389140e-01 9.12341475e-03 2.46093720e-01 -2.85352558e-01
2.14799404e-01 5.01073450e-02 1.06342569e-01 4.10879284e-01
3.08852822e-01 3.24010611e-01 3.81340593e-01 2.64825493e-01
1.01839042e+00 -2.70877898e-01 3.19180608e-01 -1.27351314e-01
7.87614822e-01 -5.16512275e-01 9.73602533e-01 7.34332979e-01
-3.64776164e-01 3.23843151e-01 4.24141377e-01 -4.67235595e-01
-5.93980074e-01 -8.81570160e-01 -7.62517899e-02 1.23428440e+00
1.17762424e-01 -1.56078562e-01 -3.97458494e-01 -7.07119524e-01
6.34330735e-02 -5.78438193e-02 -8.81752789e-01 -3.43231976e-01
-6.67898834e-01 -8.81634772e-01 -3.67645919e-02 5.71736813e-01
5.77891707e-01 -1.28417170e+00 -5.27110517e-01 2.15810582e-01
-2.24054847e-02 -6.53673828e-01 -8.63188386e-01 1.30945027e-01
-8.88678610e-01 -1.04526043e+00 -5.67451477e-01 -8.82265866e-01
1.09823883e+00 4.65483248e-01 8.64936709e-01 4.06022340e-01
-4.59805846e-01 4.64015543e-01 -3.71969372e-01 -2.38225967e-01
2.30803132e-01 -1.29511148e-01 1.89938828e-01 5.54759502e-01
3.76050264e-01 -6.34767711e-01 -9.55648422e-01 3.30809355e-01
-9.49995399e-01 -3.22837442e-01 8.06385398e-01 1.07150650e+00
5.61428905e-01 1.69670787e-02 7.50054002e-01 -6.84589267e-01
3.13828707e-01 -5.57048798e-01 -2.99063385e-01 3.75109911e-01
-4.82628644e-01 -9.50751733e-03 6.18404746e-01 -6.37147963e-01
-9.27619755e-01 -2.09463071e-02 8.58164653e-02 -6.98455751e-01
2.68808961e-01 4.29468364e-01 -3.90324593e-01 -4.24556024e-02
3.00379038e-01 4.39885587e-01 1.18327074e-01 -3.62533987e-01
1.69751048e-01 5.46075284e-01 2.62706399e-01 -4.82613295e-01
6.89861774e-01 4.49352652e-01 -1.73478484e-01 -7.25882232e-01
-4.33538765e-01 -2.59540528e-01 -5.44735909e-01 -3.37141037e-01
3.21553946e-01 -7.79592395e-01 -1.03830409e+00 6.96940124e-01
-8.88655066e-01 -1.92727670e-02 -2.31848627e-01 1.80504933e-01
-1.28655478e-01 2.60619998e-01 -5.31298339e-01 -8.93751144e-01
-5.32946527e-01 -1.18715918e+00 6.82844639e-01 7.58339465e-01
1.92065760e-01 -3.82202655e-01 -4.42328990e-01 -3.02813165e-02
3.77555698e-01 -1.55846655e-01 6.86799526e-01 -5.20865023e-01
-7.38816679e-01 -2.61465281e-01 -1.98359832e-01 4.69162703e-01
4.55368340e-01 8.48951116e-02 -9.09412920e-01 -7.20262825e-01
1.42218471e-01 -3.28350246e-01 1.03330159e+00 3.93651247e-01
1.19923866e+00 -5.41629195e-01 -2.59908408e-01 5.03137469e-01
1.35130310e+00 6.68433845e-01 4.94296461e-01 -3.22553188e-01
4.48877484e-01 4.74095464e-01 4.44245845e-01 6.39830709e-01
1.17684141e-01 4.80111629e-01 1.85785264e-01 9.39553753e-02
-8.35144743e-02 2.20885277e-01 4.87967759e-01 1.07099020e+00
-6.28800169e-02 -9.18073803e-02 -5.48993409e-01 5.94866037e-01
-1.74915648e+00 -1.09951246e+00 5.42750061e-01 2.18676567e+00
8.88455868e-01 1.10777132e-02 -3.09329238e-02 5.61645329e-02
7.33954370e-01 8.45790803e-02 -9.73072946e-01 -2.25210488e-01
-1.05377592e-01 3.84504437e-01 -3.12284343e-02 2.72197902e-01
-9.43270087e-01 7.66110420e-01 5.50355530e+00 1.00719726e+00
-1.26183903e+00 8.32957104e-02 9.71497715e-01 -4.52851921e-01
6.33642497e-03 -1.96613029e-01 -9.64173734e-01 7.02524185e-01
6.93157732e-01 -1.42015070e-01 5.56951523e-01 5.65326929e-01
2.21746415e-02 -1.21503048e-01 -1.18557644e+00 1.15336800e+00
1.18288949e-01 -1.14804935e+00 2.54722923e-01 -3.06162070e-02
7.52207458e-01 -2.03517407e-01 4.82856989e-01 2.37961426e-01
8.42832029e-02 -9.44359303e-01 6.17000818e-01 8.75314713e-01
7.29877591e-01 -8.24658871e-01 5.45711935e-01 1.15862288e-01
-1.62012219e+00 -5.83476424e-01 -4.63226199e-01 1.59187764e-01
-1.33381739e-01 3.66563112e-01 -3.05461973e-01 6.80889606e-01
1.02732515e+00 7.88162053e-01 -6.27654910e-01 6.59054041e-01
1.81403548e-01 3.03371161e-01 -3.58714819e-01 1.71280261e-02
4.87422086e-02 -7.30606914e-02 2.42183626e-01 9.31426764e-01
4.25735742e-01 5.02613723e-01 1.08297653e-01 4.01750594e-01
-2.34234214e-01 1.47057816e-01 -3.70307863e-01 -5.54656349e-02
6.59857154e-01 1.20722985e+00 -6.78251326e-01 -3.47655714e-01
-5.95813274e-01 1.05317509e+00 5.57534039e-01 4.23514426e-01
-6.44676268e-01 -4.15441960e-01 9.75993216e-01 -3.11547399e-01
7.79782593e-01 -5.52556589e-02 -1.33142788e-02 -9.99135196e-01
2.08636120e-01 -9.99420822e-01 6.33593202e-01 -2.24208355e-01
-1.24144292e+00 8.85242760e-01 -3.66059810e-01 -1.13144267e+00
-7.79313371e-02 -4.84275818e-01 -6.09286129e-01 8.30238342e-01
-1.33591008e+00 -8.71583521e-01 -2.07705110e-01 8.82618070e-01
7.19703496e-01 -2.09205851e-01 5.77665687e-01 4.08088535e-01
-9.39298093e-01 9.91339087e-01 -2.85775959e-02 1.70854941e-01
4.41926628e-01 -7.33366728e-01 6.00632653e-02 1.07921875e+00
2.17992365e-01 1.00087070e+00 2.70831943e-01 -5.07193625e-01
-1.60504663e+00 -1.04252529e+00 5.84581852e-01 -8.96603391e-02
2.54066229e-01 -1.91008851e-01 -1.04432809e+00 4.46842134e-01
-3.89002562e-02 3.41211557e-01 6.57901525e-01 -2.13738177e-02
-4.02671129e-01 -8.16552937e-01 -1.03730726e+00 3.59831750e-01
1.14677727e+00 -6.17126822e-01 -1.84294477e-01 -3.52359936e-02
4.74121988e-01 6.18314417e-03 -8.37116718e-01 6.36853576e-01
6.44200742e-01 -8.55112731e-01 7.81741083e-01 -6.80958331e-01
1.31686866e-01 -2.56556123e-01 -2.90366501e-01 -8.57325852e-01
-5.41866243e-01 -6.98228657e-01 -8.74429494e-02 1.30781162e+00
3.67826760e-01 -6.68915808e-01 4.60244983e-01 5.36588490e-01
2.46245220e-01 -1.01279747e+00 -1.19636989e+00 -5.32861531e-01
-3.38800699e-01 -2.00671256e-01 6.82114899e-01 5.65509081e-01
-1.75829753e-01 7.75937438e-02 -3.35643858e-01 1.65869102e-01
5.55896401e-01 3.30745786e-01 4.20639277e-01 -1.00139380e+00
-7.65795857e-02 -5.76726913e-01 -4.78178710e-01 -1.13716412e+00
-9.05630440e-02 -8.03814709e-01 7.10993931e-02 -9.05782342e-01
4.72035408e-01 -4.89749134e-01 -9.11361873e-01 8.26598108e-01
-4.71340626e-01 3.37536126e-01 2.52372175e-01 4.30719584e-01
-6.56302631e-01 6.05539918e-01 1.08082759e+00 -7.21490532e-02
-1.70300677e-01 -1.91912856e-02 -7.09815741e-01 4.01018858e-01
6.30356371e-01 -3.34202826e-01 -3.84567738e-01 -4.12141860e-01
-1.07704796e-01 -8.60454421e-03 -1.15320124e-02 -9.85127151e-01
5.47869444e-01 -9.21849757e-02 8.11084747e-01 -5.10903537e-01
4.26429093e-01 -7.24923551e-01 1.80655420e-01 4.91263449e-01
-1.93885356e-01 2.58179605e-01 2.31257454e-01 5.24866164e-01
-2.42992595e-01 1.12013698e-01 9.58072007e-01 -4.31822538e-02
-7.35410094e-01 8.97726655e-01 -1.03095576e-01 -2.65744776e-01
9.12174463e-01 -2.18865529e-01 3.37996781e-02 -2.87586808e-01
-6.99449778e-01 3.43408078e-01 3.00002806e-02 5.37544370e-01
9.17338073e-01 -1.49425077e+00 -6.34463727e-01 6.77753687e-01
-2.37005875e-01 -1.38997778e-01 6.75732553e-01 7.80608535e-01
2.71286126e-02 2.84751177e-01 -3.05870235e-01 -6.65552258e-01
-1.43644106e+00 5.02131045e-01 2.99988538e-01 -1.91475555e-01
-8.33798647e-02 1.29019821e+00 2.85507083e-01 3.11678648e-02
2.52141923e-01 -1.00102760e-01 -2.24711284e-01 3.23903799e-01
9.17937696e-01 7.48181567e-02 5.98534830e-02 -7.66902626e-01
-6.40237808e-01 6.93650186e-01 -5.18547058e-01 5.87674528e-02
1.44635057e+00 -2.43091747e-01 -3.11562419e-01 3.43132973e-01
1.29263186e+00 -4.93182003e-01 -1.43321955e+00 -5.81663609e-01
-5.85434996e-02 -5.25712729e-01 -1.32510856e-01 -4.68358725e-01
-1.57756042e+00 7.12578654e-01 9.41479981e-01 -2.27959547e-02
1.74085665e+00 -2.00985428e-02 4.88447964e-01 3.40413004e-01
2.26684317e-01 -9.40848768e-01 3.85440201e-01 3.58500034e-01
1.04174554e+00 -1.10319996e+00 -4.39551882e-02 -1.11231178e-01
-3.26783687e-01 9.38153684e-01 7.64031827e-01 4.99759987e-03
9.57964838e-01 1.34437501e-01 -6.13530725e-02 -1.88348487e-01
-1.09952974e+00 -2.45822430e-01 4.03209984e-01 1.41725644e-01
2.48529434e-01 -8.68623033e-02 -2.42530152e-01 8.42332065e-01
1.83527142e-01 -5.22388406e-02 -1.71953559e-01 1.08235121e+00
-4.70841646e-01 -1.06178153e+00 -3.25526536e-01 6.83162928e-01
-4.86339450e-01 -1.01399727e-01 5.64096756e-02 4.04973716e-01
2.34240845e-01 9.61456537e-01 1.75530255e-01 -6.05619431e-01
2.50633359e-01 2.33701974e-01 4.52066928e-01 -3.58084381e-01
-5.52576363e-01 1.67511716e-01 -6.34150088e-01 -8.78386438e-01
-2.86936224e-01 -8.48984540e-01 -1.02965486e+00 -2.51452506e-01
-3.28024417e-01 2.60444820e-01 3.92022818e-01 7.88086057e-01
5.93729913e-01 4.16654587e-01 1.28452778e+00 -8.71730626e-01
-5.48192263e-01 -1.10608387e+00 -5.52857339e-01 5.41861475e-01
4.35267419e-01 -4.97217178e-01 -3.00856829e-01 2.47278824e-01] | [13.171126365661621, 0.6967378258705139] |
00a2a77e-944f-46b9-9288-9a638f2f1aa7 | textoir-an-integrated-and-visualized-platform-1 | 2110.15063 | null | https://arxiv.org/abs/2110.15063v1 | https://arxiv.org/pdf/2110.15063v1.pdf | TEXTOIR: An Integrated and Visualized Platform for Text Open Intent Recognition | TEXTOIR is the first integrated and visualized platform for text open intent recognition. It is composed of two main modules: open intent detection and open intent discovery. Each module integrates most of the state-of-the-art algorithms and benchmark intent datasets. It also contains an overall framework connecting the two modules in a pipeline scheme. In addition, this platform has visualized tools for data and model management, training, evaluation and analysis of the performance from different aspects. TEXTOIR provides useful toolkits and convenient visualized interfaces for each sub-module (Toolkit code: https://github.com/thuiar/TEXTOIR), and designs a framework to implement a complete process to both identify known intents and discover open intents (Demo code: https://github.com/thuiar/TEXTOIR-DEMO). | ['Kai Gao', 'Kang Zhao', 'Panpan Zhang', 'Hua Xu', 'Xiaoteng Li', 'Hanlei Zhang'] | 2021-09-13 | textoir-an-integrated-and-visualized-platform | https://aclanthology.org/2021.acl-demo.20 | https://aclanthology.org/2021.acl-demo.20.pdf | acl-2021-5 | ['intent-recognition', 'open-intent-detection', 'open-intent-discovery', 'intent-discovery'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-5.26747942e-01 -4.48806643e-01 -2.66651183e-01 -2.34779224e-01
-4.31543946e-01 -6.72577083e-01 5.22633672e-01 -1.04095429e-01
-9.31049064e-02 1.47332042e-01 6.67360544e-01 -2.41147906e-01
-4.95041125e-02 -5.77195227e-01 3.15423310e-02 -3.72183442e-01
-2.32155189e-01 4.63345706e-01 -1.25398412e-01 -3.31355184e-01
5.30203402e-01 -2.65919790e-02 -1.79505646e+00 3.77202809e-01
8.20799649e-01 7.41316378e-01 1.17356330e-01 1.10918689e+00
-1.52459338e-01 1.09038556e+00 -4.16132838e-01 1.36564404e-01
1.39046133e-01 1.25008849e-02 -9.38156307e-01 -4.13143933e-01
3.31402421e-01 -5.15300512e-01 -4.51928318e-01 6.80133998e-01
3.58455747e-01 8.39355343e-05 7.82903612e-01 -1.69133937e+00
-8.06197941e-01 2.14511976e-01 -9.31898132e-02 3.69287133e-01
1.16366112e+00 5.07322550e-01 1.12426579e+00 -9.81400490e-01
8.69423866e-01 6.93047941e-01 8.55459571e-01 4.24212933e-01
-8.33230019e-01 -5.40306985e-01 -2.06078142e-01 3.11225086e-01
-1.42952514e+00 -4.63543028e-01 3.70781809e-01 -8.81402969e-01
1.17262340e+00 9.21377420e-01 5.75160444e-01 1.44678032e+00
4.53279585e-01 1.06645799e+00 8.20876896e-01 -7.10980669e-02
2.61181712e-01 1.58837110e-01 1.00131333e+00 9.59797442e-01
-5.51039316e-02 -1.94194093e-01 -7.49365449e-01 -5.83736002e-01
1.53627452e-02 5.95177710e-01 -3.08961749e-01 -7.16537759e-02
-1.35711682e+00 7.44108200e-01 6.00774527e-01 4.06578809e-01
-2.02727616e-01 -1.08050898e-01 7.98314810e-01 2.14997247e-01
4.94029611e-01 4.31075096e-01 -3.21446955e-01 -5.16840279e-01
-6.77047193e-01 1.79955140e-01 1.08182728e+00 9.85639036e-01
7.79264212e-01 -2.29913667e-01 -2.02734560e-01 8.91076028e-01
6.03350878e-01 5.19546047e-02 5.95159113e-01 -4.61730182e-01
1.77388579e-01 1.04909277e+00 -8.03779662e-02 -6.78752720e-01
-5.28136194e-01 1.67241827e-01 -5.16898155e-01 -1.76826715e-02
-2.05187559e-01 -2.10864455e-01 -8.18743467e-01 9.79044139e-01
3.51258010e-01 -2.68150359e-01 1.57449454e-01 7.70554602e-01
1.57447362e+00 7.29076564e-01 3.73128094e-02 1.09676123e-01
1.77490270e+00 -1.20648682e+00 -7.45404840e-01 -5.07269561e-01
1.02004075e+00 -9.01655197e-01 1.26824403e+00 1.69554949e-02
-2.95310974e-01 -2.72017092e-01 -1.01633155e+00 -3.63001615e-01
-1.03520477e+00 2.41426006e-01 8.67042601e-01 2.86152482e-01
-1.06061542e+00 1.21563002e-01 -6.41312659e-01 -1.15610611e+00
2.18095869e-01 1.40703157e-01 -3.76813024e-01 5.80526590e-02
-1.03334773e+00 4.51138049e-01 4.04524446e-01 -3.59514326e-01
-9.56085265e-01 -7.79434025e-01 -1.01073408e+00 -1.58695027e-01
2.00627625e-01 -1.05979311e+00 1.31834495e+00 -3.06313902e-01
-9.36381757e-01 8.86213899e-01 -7.86507800e-02 7.31821731e-02
3.72126967e-01 -7.20420837e-01 -4.68172938e-01 -1.77453145e-01
5.88868082e-01 7.40710318e-01 5.73119879e-01 -7.98258722e-01
-5.65695345e-01 -5.33924997e-01 -1.94379315e-01 3.03105742e-01
-5.72069168e-01 5.39816439e-01 -5.26283205e-01 -5.11036098e-01
-2.73539454e-01 -9.09239531e-01 -4.65551466e-02 9.87261832e-02
-5.63136399e-01 -4.77890402e-01 1.45158792e+00 -6.39131665e-01
1.59023392e+00 -2.18876290e+00 -1.65897027e-01 -4.11768258e-01
4.47481424e-01 -1.37033006e-02 3.11816812e-01 1.33628464e+00
-4.00930792e-02 2.11570650e-01 -1.12235330e-01 -6.35053456e-01
2.78938025e-01 1.32796869e-01 -3.60768437e-01 5.18662274e-01
-5.73279977e-01 8.77030611e-01 -7.72899091e-01 -4.74721998e-01
7.00896621e-01 3.95116031e-01 -2.46038571e-01 4.25559402e-01
-4.50536460e-01 2.08753675e-01 -4.84731227e-01 1.01311517e+00
7.19446421e-01 -4.13432032e-01 7.70003721e-02 2.08091214e-01
-7.09762394e-01 3.34069222e-01 -1.05961692e+00 1.75248301e+00
-3.76165658e-01 8.99119556e-01 -8.45151895e-04 9.61978734e-02
7.55909324e-01 2.39615068e-01 4.01099235e-01 5.17124534e-02
2.98498809e-01 -1.34456173e-01 -8.28418732e-01 -8.33052695e-01
9.11512792e-01 6.51980698e-01 -4.61975455e-01 7.98804522e-01
1.10814087e-01 2.10135996e-01 2.20052376e-01 3.42943400e-01
1.42492080e+00 5.16159199e-02 8.32079887e-01 -2.95419335e-01
4.34133232e-01 3.87062669e-01 1.96458116e-01 5.04073203e-01
-2.26452768e-01 3.38142008e-01 3.43454063e-01 -9.20602798e-01
-6.03687823e-01 -8.25227141e-01 -3.69152069e-01 1.68386257e+00
-2.66951043e-02 -1.43158674e+00 -3.69717568e-01 -7.68494666e-01
-1.56883225e-01 7.22301424e-01 -8.53274167e-01 4.46984470e-01
1.57720312e-01 -6.22682035e-01 7.90222585e-01 2.24505946e-01
7.19190001e-01 -1.29778850e+00 -9.17995036e-01 -4.16267902e-01
-5.89631200e-01 -7.50299096e-01 -6.28713310e-01 3.41990501e-01
-5.15207052e-01 -1.18074524e+00 -1.52874529e-01 -6.80579484e-01
5.41078389e-01 3.91966134e-01 1.11092460e+00 3.34365398e-01
-7.31195092e-01 6.24009490e-01 -6.10842943e-01 -3.19799364e-01
-3.19564670e-01 1.92479655e-01 1.56073689e-01 -1.71248972e-01
5.10033548e-01 -4.41941351e-01 -6.93835855e-01 4.54232216e-01
-7.76536524e-01 4.27529931e-01 7.62202591e-02 5.52997231e-01
1.09811090e-01 -4.26105946e-01 -2.25864321e-01 -3.94066632e-01
9.27681208e-01 -8.78855944e-01 -6.47900224e-01 5.87023348e-02
-6.30958617e-01 -1.35912582e-01 5.20474970e-01 7.04890043e-02
-7.55449295e-01 5.28262705e-02 -5.47430158e-01 -5.00591755e-01
-8.57867301e-01 3.79620731e-01 1.64190516e-01 1.79636717e-01
7.84826994e-01 5.50433397e-01 -1.15070365e-01 -7.61109293e-01
2.06766605e-01 1.54834795e+00 2.15263695e-01 1.66232735e-01
4.55308944e-01 5.35053670e-01 -6.83688223e-01 -1.07602155e+00
-7.72605896e-01 -1.22444546e+00 -8.45257699e-01 -4.23867583e-01
9.52192307e-01 -8.00498068e-01 -8.46141875e-01 6.64551318e-01
-9.26000357e-01 -4.75211054e-01 4.28361259e-02 4.43897098e-02
-3.50076944e-01 2.59745866e-01 -5.50253093e-01 -6.21661246e-01
-1.05897605e+00 -1.05069852e+00 1.13243175e+00 2.56298631e-01
-7.16877341e-01 -1.24709213e+00 8.42949629e-01 6.39380991e-01
2.21330404e-01 9.33520123e-02 2.37049401e-01 -9.43575323e-01
-3.53291035e-01 -2.32164085e-01 -1.62483398e-02 -1.41601413e-01
1.88011169e-01 3.24963689e-01 -1.14825070e+00 -3.48917693e-02
-1.97371304e-01 -5.08784711e-01 6.74248874e-01 -1.10401824e-01
8.41466367e-01 -6.46152437e-01 -8.63731623e-01 1.01428425e+00
1.36774111e+00 -3.16094793e-02 7.31161296e-01 6.62404478e-01
7.37601876e-01 3.48794580e-01 8.21861088e-01 7.92414904e-01
6.76365793e-01 7.61270285e-01 5.22412539e-01 1.26989305e-01
1.34641781e-01 -2.76757389e-01 5.47578037e-01 3.98034394e-01
1.90893129e-01 -2.18812883e-01 -1.48640740e+00 4.39533830e-01
-2.05202365e+00 -1.17286932e+00 -7.24002659e-01 1.87826014e+00
2.80621141e-01 -4.43759263e-01 1.81258917e-01 -1.89495996e-01
3.01781863e-01 6.02186680e-01 -2.89012879e-01 -5.73421538e-01
3.84010315e-01 -6.59117281e-01 6.15847223e-02 3.78402889e-01
-1.37917924e+00 8.39526951e-01 6.91675615e+00 7.72122920e-01
-1.04489505e+00 3.57921809e-01 1.71149388e-01 9.20559689e-02
1.19606853e-01 1.04674369e-01 -8.81955743e-01 3.91121000e-01
7.96096385e-01 -1.17829345e-01 3.76165301e-01 1.36247301e+00
1.08242154e-01 -3.82276177e-01 -9.66205597e-01 9.30806935e-01
3.16664696e-01 -1.58058441e+00 -2.31509864e-01 2.74190873e-01
2.70654559e-01 8.85113180e-01 -2.84440815e-01 4.70388055e-01
3.78363222e-01 -6.89595640e-01 6.04490876e-01 2.52407074e-01
7.99625933e-01 -2.54831523e-01 6.19419336e-01 5.42464972e-01
-1.53199375e+00 -3.35191995e-01 -5.84281683e-02 -2.97470540e-01
-1.12612732e-01 7.35691041e-02 -8.62883508e-01 5.58346212e-01
1.06139481e+00 1.19796133e+00 -9.40883636e-01 1.05576384e+00
-2.94567853e-01 3.10014218e-01 -2.12486088e-01 -1.95092812e-01
7.49224424e-02 -2.61840045e-01 8.64420116e-01 1.44711518e+00
-1.84081927e-01 -3.30074757e-01 3.67041707e-01 5.25800467e-01
2.11021647e-01 3.07370275e-01 -1.11576509e+00 7.78196752e-02
1.55773237e-01 1.64594460e+00 -7.12972879e-01 -2.26151273e-01
-4.85376805e-01 1.16260707e+00 4.98937786e-01 -1.78000122e-01
-9.74748969e-01 -3.44299346e-01 1.07921040e+00 -1.64134670e-02
-1.34330308e-02 -8.52513611e-02 -2.12053865e-01 -1.24763536e+00
-1.96591794e-01 -7.61895955e-01 1.27199137e+00 -1.06726062e+00
-1.10967171e+00 6.69824183e-01 9.61640105e-02 -1.66308820e+00
-5.33996932e-02 -5.46195030e-01 -1.12361538e+00 2.68536091e-01
-8.32558393e-01 -1.32993388e+00 -8.69519830e-01 7.66543150e-01
1.02358973e+00 2.02622097e-02 1.33240449e+00 1.31785154e-01
-1.05839956e+00 2.89321423e-01 1.97644755e-01 2.81221122e-01
6.67732596e-01 -1.05774486e+00 6.21764064e-01 7.43078470e-01
-2.02178000e-03 8.78640234e-01 7.30998635e-01 -7.61527777e-01
-1.76528156e+00 -1.10590065e+00 6.26965582e-01 -9.60799813e-01
9.73892927e-01 -8.11731339e-01 -5.38205087e-01 1.16509235e+00
5.91404676e-01 -1.30422294e-01 1.23115623e+00 4.51816380e-01
-3.65804315e-01 2.46688664e-01 -8.25283766e-01 8.45140040e-01
9.85366404e-01 -3.10141146e-01 -7.45773315e-01 8.74103487e-01
3.24401319e-01 -4.18124169e-01 -9.32648420e-01 1.66290566e-01
7.79740572e-01 -1.44203675e+00 6.65106356e-01 -5.96279763e-02
2.41856754e-01 -3.64039898e-01 -4.31354418e-02 -7.86284268e-01
-2.01602280e-01 -7.69434214e-01 -3.47765893e-01 1.03665280e+00
2.71791756e-01 -8.79243314e-01 3.51043642e-01 1.66892573e-01
-4.75787520e-01 -7.61923075e-01 -7.99109876e-01 -4.49507713e-01
-7.51205027e-01 -8.21098566e-01 3.35155249e-01 8.62029076e-01
6.35689676e-01 6.81277692e-01 -4.22020048e-01 1.53720081e-01
3.64169091e-01 3.07759047e-01 1.28992498e+00 -1.01798880e+00
4.71315272e-02 -2.25790009e-01 -4.06583279e-01 -1.00223053e+00
-1.86592579e-01 -9.72782493e-01 -1.92424431e-01 -2.10391140e+00
4.73814100e-01 -4.45818931e-01 1.41945481e-01 9.88958716e-01
-4.68429103e-02 3.46885949e-01 1.99947700e-01 9.22368109e-01
-1.15555084e+00 5.65057993e-01 5.01634657e-01 -7.05453977e-02
-5.20545542e-01 -1.39698619e-02 -7.24488556e-01 8.51960957e-01
1.16878104e+00 -6.26355112e-01 -1.48283288e-01 -5.77499978e-02
2.28320789e-02 -2.94294626e-01 4.95723218e-01 -1.40915442e+00
3.79528046e-01 -1.24877103e-01 1.65837899e-01 -1.23025799e+00
3.09441119e-01 -6.61870897e-01 4.62321192e-01 6.74158275e-01
1.44149780e-01 1.19066074e-01 3.18173438e-01 1.20845601e-01
1.83656216e-01 -2.06550404e-01 3.87833804e-01 -1.01776838e-01
-1.28332913e+00 1.50762334e-01 -8.10588479e-01 -7.90773183e-02
1.50578499e+00 -3.52563828e-01 -8.63773584e-01 -2.68969595e-01
-3.72569442e-01 5.60724318e-01 7.79539466e-01 9.15246606e-01
5.70118368e-01 -9.02702332e-01 -2.92962343e-01 4.31198567e-01
9.40509796e-01 -5.37794054e-01 5.38497925e-01 9.43529665e-01
-8.08053434e-01 7.91308582e-01 -1.98288664e-01 -7.08387792e-01
-1.79439557e+00 7.37755358e-01 3.35437655e-01 -5.72615564e-01
-8.08219373e-01 5.07414997e-01 1.13150187e-01 -9.20162201e-01
1.48932233e-01 8.34472701e-02 -4.10037220e-01 1.66082960e-02
8.69441450e-01 5.24850965e-01 9.35420990e-02 -8.34585369e-01
-8.62349153e-01 4.06693637e-01 -7.05543021e-03 3.77029061e-01
1.37835741e+00 -1.44451186e-01 -2.67167926e-01 6.55850112e-01
1.07964504e+00 4.84545007e-02 -6.20458186e-01 5.17308176e-01
-2.95410186e-01 -4.30216312e-01 5.81183061e-02 -8.12838018e-01
-2.60086030e-01 3.13543677e-01 6.75358534e-01 3.44175547e-01
9.76632059e-01 4.43705112e-01 3.55039954e-01 2.92259127e-01
3.23803842e-01 -6.93824232e-01 -2.22911257e-02 9.03087795e-01
1.19386828e+00 -1.20234787e+00 3.66518468e-01 -3.26677352e-01
-7.48741508e-01 8.98293853e-01 7.05744624e-01 3.65739226e-01
7.48075426e-01 4.26622093e-01 5.38335025e-01 -7.53024340e-01
-1.05579472e+00 -1.87982738e-01 2.52908736e-01 4.87298280e-01
4.06079739e-01 7.21906424e-02 -2.13207901e-01 4.71785247e-01
-3.02535266e-01 1.01540357e-01 3.60566586e-01 1.24062645e+00
-1.87723488e-01 -9.64311242e-01 -8.30703735e-01 5.95972478e-01
-7.11857751e-02 -2.64877617e-01 -5.44338644e-01 7.24832773e-01
8.36165771e-02 9.88100767e-01 -5.88037772e-03 -9.96556222e-01
1.02406600e-02 3.18465084e-01 -4.71951753e-01 -7.16446400e-01
-1.02679050e+00 -2.59080499e-01 2.05140054e-01 -7.70708621e-01
2.33864635e-01 -5.66974163e-01 -1.07523143e+00 -6.27161622e-01
-2.38150939e-01 2.21822917e-01 6.45193696e-01 3.38107377e-01
9.51509833e-01 2.35692456e-01 4.81020004e-01 -1.08287764e+00
1.95775688e-01 -1.23537493e+00 -3.83778095e-01 2.34447643e-02
2.29254156e-01 -4.25043553e-01 -6.90716028e-01 -1.65206537e-01] | [12.425951957702637, 7.572813034057617] |
d96a8ff5-21e9-4246-affb-6d69e6668c80 | should-we-hard-code-the-recurrence-concept-or | 2005.09297 | null | https://arxiv.org/abs/2005.09297v1 | https://arxiv.org/pdf/2005.09297v1.pdf | Should we hard-code the recurrence concept or learn it instead ? Exploring the Transformer architecture for Audio-Visual Speech Recognition | The audio-visual speech fusion strategy AV Align has shown significant performance improvements in audio-visual speech recognition (AVSR) on the challenging LRS2 dataset. Performance improvements range between 7% and 30% depending on the noise level when leveraging the visual modality of speech in addition to the auditory one. This work presents a variant of AV Align where the recurrent Long Short-term Memory (LSTM) computation block is replaced by the more recently proposed Transformer block. We compare the two methods, discussing in greater detail their strengths and weaknesses. We find that Transformers also learn cross-modal monotonic alignments, but suffer from the same visual convergence problems as the LSTM model, calling for a deeper investigation into the dominant modality problem in machine learning. | ['George Sterpu', 'Naomi Harte', 'Christian Saam'] | 2020-05-19 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 3.86411756e-01 1.37651160e-01 -3.47235128e-02 -1.21061541e-01
-1.33321202e+00 -4.40069288e-01 9.75052297e-01 -7.02981502e-02
-3.84182245e-01 2.99402267e-01 4.76171672e-01 -5.08925855e-01
2.25008294e-01 -1.34255097e-04 -6.16543293e-01 -9.91635740e-01
2.14714050e-01 2.80537993e-01 3.43266651e-02 -4.80522998e-02
-2.71170549e-02 3.85932565e-01 -1.69353294e+00 8.41532290e-01
1.44832537e-01 1.14342058e+00 3.43129933e-01 9.39631939e-01
-4.27752584e-01 9.20724094e-01 -6.41388953e-01 -2.24846393e-01
-9.73126739e-02 -2.21240461e-01 -8.17365110e-01 -1.38218105e-01
8.08268487e-01 2.44752064e-01 -3.43407571e-01 6.53481305e-01
8.79161775e-01 1.17641471e-01 4.01938319e-01 -1.13458586e+00
-4.72281128e-01 7.37827241e-01 -6.80544257e-01 3.63024294e-01
4.58308727e-01 9.12562683e-02 1.19227707e+00 -1.19350564e+00
3.97821456e-01 1.42137229e+00 5.49163163e-01 5.91511488e-01
-1.34456742e+00 -4.62941796e-01 2.38869503e-01 6.87135577e-01
-1.24961460e+00 -9.03179109e-01 5.83325326e-01 -3.49545866e-01
1.62798893e+00 3.83740813e-01 5.38155138e-01 1.66870487e+00
6.69736266e-02 9.25713480e-01 1.29801357e+00 -5.61797976e-01
2.00508186e-03 2.00264454e-01 4.14404124e-02 2.41333723e-01
-5.65547764e-01 2.32658431e-01 -9.89910066e-01 -5.29061817e-02
1.26390502e-01 -4.87492889e-01 -2.94307888e-01 -2.22898334e-01
-1.24315941e+00 6.09390080e-01 3.79183382e-01 6.90011382e-01
-2.91513532e-01 3.08933616e-01 7.55410254e-01 5.76026380e-01
4.92818922e-01 -3.48615423e-02 -4.06929612e-01 -2.49504551e-01
-1.23546672e+00 -1.46605119e-01 2.31256261e-01 4.92417753e-01
3.48536551e-01 6.66898608e-01 -4.70069379e-01 1.04689121e+00
6.16669476e-01 5.32813311e-01 6.05302453e-01 -5.74918032e-01
6.60447717e-01 -1.37050301e-01 -3.73779863e-01 -5.51780343e-01
-3.68386596e-01 -5.78859448e-01 -8.06929588e-01 3.91545892e-01
2.65201360e-01 2.63851434e-01 -1.20515287e+00 1.78230059e+00
9.50845554e-02 1.91073984e-01 2.38074139e-01 7.07562685e-01
1.39853680e+00 8.46103251e-01 2.69370228e-01 -3.30266058e-01
1.45165586e+00 -8.69708836e-01 -8.53083313e-01 -3.47986817e-01
3.33484799e-01 -9.17364180e-01 1.19945085e+00 4.12649095e-01
-1.33581090e+00 -5.13398945e-01 -1.25552976e+00 -2.35410213e-01
-2.80694664e-01 2.53019005e-01 6.15251847e-02 5.83727896e-01
-1.59162378e+00 2.72762030e-01 -6.80676997e-01 -3.99636149e-01
7.84492493e-02 2.51839250e-01 -4.95093226e-01 4.09236014e-01
-1.13046563e+00 1.08486319e+00 1.33798316e-01 2.43837893e-01
-9.87906694e-01 -4.52046305e-01 -9.12459016e-01 1.02892157e-03
2.29100615e-01 -7.20272422e-01 1.33122170e+00 -1.01352739e+00
-1.70733380e+00 9.61340964e-01 -4.54194903e-01 -7.38288581e-01
3.76774967e-01 2.30989512e-02 -3.88553351e-01 1.60264924e-01
-4.81926262e-01 7.45246172e-01 1.20488226e+00 -1.27694595e+00
-3.74091387e-01 -3.67147356e-01 -3.33572656e-01 1.92453742e-01
-1.54553473e-01 2.68801987e-01 -1.34627610e-01 -8.88869226e-01
4.86574322e-03 -8.55926216e-01 2.62057066e-01 -4.08654481e-01
-3.15186530e-01 -2.84441888e-01 8.93816352e-01 -7.94916153e-01
9.68552232e-01 -2.28461266e+00 6.14558518e-01 3.23530193e-03
1.39429197e-01 4.20534611e-01 -4.94292945e-01 5.59916019e-01
-5.35017192e-01 -7.92184919e-02 7.69574335e-03 -9.88193989e-01
-4.24661525e-02 1.33963898e-01 -8.30785930e-01 3.77689064e-01
-1.86222252e-02 9.70672786e-01 -3.83553833e-01 -3.18660647e-01
3.59063208e-01 9.59726572e-01 2.01067075e-01 -2.97949724e-02
7.33734742e-02 4.30488020e-01 3.29737395e-01 5.92504501e-01
4.03762877e-01 -1.39065072e-01 -7.05837011e-02 -3.42775792e-01
-9.77954343e-02 7.46619761e-01 -7.76861906e-01 1.80022466e+00
-6.35201216e-01 1.17258012e+00 1.83722600e-01 -9.07885134e-01
7.44552910e-01 9.01454031e-01 1.20992430e-01 -1.06502235e+00
-5.11536142e-03 2.61893094e-01 -4.12570052e-02 -2.75766641e-01
3.35970968e-01 -5.55647910e-01 3.07929635e-01 4.00853045e-02
4.92636353e-01 2.12140068e-01 -2.76630551e-01 2.54640967e-01
8.60084116e-01 2.02944707e-02 2.20240638e-01 1.19531564e-01
6.66464329e-01 -3.84902865e-01 -2.13209353e-02 7.46941149e-01
-1.43068507e-01 7.41699934e-01 3.41096491e-01 -2.26674937e-02
-9.42096770e-01 -1.34327126e+00 -3.31831835e-02 1.13960671e+00
-4.19279754e-01 -6.07248008e-01 -4.12445158e-01 -4.49784249e-01
-3.32131743e-01 7.29216516e-01 -6.11731291e-01 -1.06670849e-01
-4.64585006e-01 -3.96237493e-01 9.84345675e-01 6.87781751e-01
5.32331355e-02 -1.18484223e+00 -4.06566471e-01 -1.10595293e-01
-3.34675431e-01 -1.31748354e+00 -9.42139700e-02 3.05760324e-01
-5.62679291e-01 -4.67445970e-01 -1.06496823e+00 -6.44405782e-01
-1.13990337e-01 2.62180775e-01 9.05796528e-01 -3.26124519e-01
-1.73386887e-01 7.02173531e-01 -2.23921821e-01 -1.12835705e-01
-5.30027568e-01 -1.47529859e-02 -1.46037051e-02 6.33039922e-02
6.30217884e-03 -6.80690706e-01 -1.23097435e-01 -1.55036384e-02
-5.16883135e-01 1.80927381e-01 5.62955320e-01 9.74177837e-01
5.64772785e-01 -5.95871866e-01 5.05755484e-01 -9.21462476e-02
3.93858343e-01 -1.06711835e-01 -3.47526282e-01 2.63884038e-01
-5.46153843e-01 8.71673822e-02 2.51192540e-01 -4.68643397e-01
-1.05778909e+00 1.04162164e-01 -3.35317522e-01 -8.85762811e-01
-8.25810507e-02 5.10715008e-01 -1.16537474e-02 3.44372801e-02
2.97461271e-01 4.52482969e-01 1.82085042e-03 -6.62069976e-01
4.00192589e-01 6.20900214e-01 6.39719844e-01 2.80743316e-02
5.67484140e-01 2.84594536e-01 -2.86838729e-02 -1.14320230e+00
-3.73341382e-01 -3.27987999e-01 -3.65069807e-01 -3.68551642e-01
9.86240149e-01 -1.01589978e+00 -7.31028736e-01 5.01259387e-01
-1.39676046e+00 -3.22823972e-01 -3.24371994e-01 6.06090486e-01
-6.12682879e-01 3.84292543e-01 -8.68851244e-01 -9.91451204e-01
-3.97609144e-01 -1.28484380e+00 1.24298692e+00 -2.43633658e-01
-1.25313535e-01 -7.76387215e-01 1.38907999e-01 5.14294863e-01
4.42352712e-01 -1.98086739e-01 8.07159364e-01 -7.33311594e-01
-3.27859908e-01 1.00383401e-01 -2.34943897e-01 2.73820311e-01
-2.54688859e-01 6.22356273e-02 -1.59989071e+00 -4.61445570e-01
1.44799024e-01 -3.00633073e-01 1.42965806e+00 5.61141491e-01
6.02782071e-01 -1.69636011e-01 -5.78082539e-02 4.02165353e-01
9.56123888e-01 1.53752372e-01 6.64366484e-01 1.41659439e-01
8.23655069e-01 7.99528122e-01 2.98965186e-01 -5.33916205e-02
1.93712279e-01 1.07448125e+00 4.57499236e-01 -2.12570950e-01
-6.82530642e-01 -1.20832212e-01 9.31192100e-01 1.10116613e+00
1.09454945e-01 -2.23726913e-01 -9.75372136e-01 4.85429496e-01
-1.65533376e+00 -1.05641699e+00 -1.85069636e-01 2.30522656e+00
6.15500033e-01 4.40810516e-04 3.24005336e-01 5.22830427e-01
6.07006431e-01 5.72674453e-01 3.48985603e-04 -6.14850640e-01
-5.25424004e-01 2.82778919e-01 2.83242702e-01 7.67613947e-01
-9.58964050e-01 7.93248236e-01 7.24800396e+00 1.21518552e+00
-1.63437498e+00 4.36915070e-01 3.57217699e-01 -5.22409439e-01
-3.49800110e-01 -4.79417183e-02 -5.19855976e-01 1.13538004e-01
1.33050382e+00 4.45396066e-01 3.73459816e-01 2.47966155e-01
1.64236531e-01 4.53258716e-02 -9.25694406e-01 1.30682862e+00
3.01758587e-01 -1.13323653e+00 1.14865929e-01 7.55680948e-02
2.20955089e-02 5.16460240e-01 5.45799613e-01 1.59031317e-01
-2.58684129e-01 -1.36080265e+00 1.07236671e+00 3.42856437e-01
8.60849798e-01 -5.99889755e-01 3.72518182e-01 3.03797089e-02
-1.45762527e+00 -5.41958995e-02 1.47668198e-01 2.96133518e-01
4.36782539e-01 2.15872645e-01 -8.12769473e-01 5.44359624e-01
8.79714847e-01 6.98474050e-01 -6.91243112e-01 7.17937291e-01
-3.32293585e-02 6.27870262e-01 -1.74330011e-01 4.17785376e-01
2.48453423e-01 2.41808549e-01 1.06478846e+00 1.28607559e+00
3.22825640e-01 -7.14388013e-01 -2.84835637e-01 6.45975471e-01
3.38980734e-01 1.36433482e-01 -6.73972428e-01 -3.18666577e-01
1.36940226e-01 1.08451462e+00 -4.05586869e-01 -3.32434028e-01
-4.81633306e-01 8.63216579e-01 1.10103123e-01 4.91317034e-01
-7.67964184e-01 7.00083449e-02 3.59719902e-01 -5.81524484e-02
5.43749750e-01 -2.13550419e-01 -3.47954601e-01 -8.36762786e-01
1.87830627e-01 -9.23377395e-01 4.32892621e-01 -1.10244763e+00
-9.64213252e-01 9.24599230e-01 2.04365198e-02 -1.14684677e+00
-5.26788533e-01 -5.65419614e-01 -3.28166157e-01 8.87881696e-01
-1.45741129e+00 -1.52385259e+00 1.01735942e-01 6.29682899e-01
6.79857552e-01 -4.52998549e-01 7.56375253e-01 4.48162585e-01
-3.80064338e-01 6.84120893e-01 -2.77649045e-01 -4.38512981e-01
7.10348606e-01 -1.11988592e+00 5.51377535e-01 6.81719422e-01
7.06020832e-01 2.27348834e-01 9.61740613e-01 -3.00286144e-01
-1.21962726e+00 -6.10770166e-01 1.14901900e+00 -2.63118088e-01
6.12241507e-01 -5.36139429e-01 -9.96178091e-01 7.96659648e-01
8.73161852e-01 -6.51945993e-02 5.55018306e-01 6.75101876e-02
-8.51352811e-01 -1.59127004e-02 -7.31597900e-01 4.78898019e-01
7.53665864e-01 -1.21334016e+00 -6.09768152e-01 -1.15405276e-01
8.71024549e-01 -2.70390928e-01 -6.14594817e-01 5.14800847e-01
6.04539692e-01 -9.99939561e-01 1.19654071e+00 -3.86711955e-01
5.44807501e-02 -3.31879318e-01 -3.65351319e-01 -1.18406796e+00
6.47007972e-02 -5.99965990e-01 -1.42772198e-01 1.28480709e+00
5.42810261e-01 -7.51500845e-01 1.68277606e-01 -2.43960112e-01
-2.84566522e-01 -5.51907539e-01 -1.62090147e+00 -6.54806137e-01
-1.92146555e-01 -8.07466149e-01 -5.17818891e-02 6.90374374e-01
1.18551143e-02 7.35664785e-01 -6.74813271e-01 -6.39555752e-02
1.60315529e-01 -1.86083168e-01 5.48293948e-01 -9.51015234e-01
-4.97389942e-01 -6.62646055e-01 -4.35469449e-01 -8.91963303e-01
2.45651498e-01 -9.69272435e-01 -1.16029501e-01 -1.41507125e+00
4.17389497e-02 2.85714775e-01 -4.76225019e-01 7.56428301e-01
3.58058661e-01 4.46606696e-01 4.91368115e-01 1.23374224e-01
-4.28184718e-01 7.62447119e-01 7.18855560e-01 -4.91385490e-01
-1.24592274e-01 -7.77569339e-02 -1.60707831e-01 5.29984057e-01
5.06585002e-01 -2.66630769e-01 -4.22684729e-01 -4.10315722e-01
2.73589462e-01 3.36081505e-01 6.17473781e-01 -8.71353269e-01
3.14994991e-01 3.55674326e-01 9.83715206e-02 -9.29840267e-01
1.06162262e+00 -6.23727620e-01 1.60838082e-01 2.96568871e-01
-5.76732218e-01 3.18958014e-01 5.55089295e-01 5.01860142e-01
-5.91758847e-01 2.09219217e-01 7.78405964e-01 1.46539703e-01
-4.17515397e-01 -2.40092516e-01 -8.79075527e-01 -3.37163180e-01
6.24935865e-01 -2.61552423e-01 -3.88213575e-01 -6.39360785e-01
-1.24657369e+00 -2.14433640e-01 3.17870453e-02 6.98090792e-01
8.13970447e-01 -1.38588953e+00 -7.82669663e-01 2.45278239e-01
2.44100109e-01 -7.09731877e-01 5.01023173e-01 1.48146260e+00
1.40197873e-01 6.56185627e-01 -5.09753339e-02 -1.01363707e+00
-1.92767966e+00 6.10271990e-01 5.49988210e-01 -4.28047730e-03
-8.12651098e-01 9.52983379e-01 2.80024707e-01 -2.83171590e-02
6.45669162e-01 -1.70820624e-01 -4.13443595e-01 4.74531084e-01
5.50848067e-01 3.44530553e-01 3.78870249e-01 -1.15859294e+00
-6.03376448e-01 5.25775552e-01 4.46161851e-02 -7.06080437e-01
1.05921781e+00 -3.11890602e-01 2.02382561e-02 1.18198287e+00
1.18593442e+00 7.47152939e-02 -7.09890664e-01 -3.91705155e-01
-1.26803098e-02 -1.06212974e-01 2.65063226e-01 -8.27512383e-01
-1.28890431e+00 1.37856948e+00 9.96761680e-01 1.93210796e-01
1.21370792e+00 2.56467491e-01 6.07938111e-01 -5.17711043e-02
-8.93733799e-02 -8.31039190e-01 3.51163559e-02 4.53363746e-01
1.30419648e+00 -8.72679949e-01 -3.24927270e-01 4.51033413e-02
-8.29868555e-01 1.06160557e+00 1.41358733e-01 1.36588559e-01
4.02001798e-01 3.80946904e-01 2.21111432e-01 -1.61318332e-01
-1.17704022e+00 -3.76408696e-01 6.56142056e-01 5.90128481e-01
4.28112447e-01 -8.42324793e-02 3.82189713e-02 2.21135095e-01
-1.83393970e-01 -5.79828262e-01 5.60285561e-02 4.75438535e-01
-8.01035315e-02 -1.08514929e+00 -5.24824440e-01 -6.39576539e-02
-4.09667701e-01 -4.10171449e-01 -6.69118464e-01 5.68316877e-01
-2.81104475e-01 1.07453084e+00 1.57459751e-01 -6.40320361e-01
1.52368337e-01 4.32794899e-01 6.66948140e-01 -2.27950752e-01
-8.14154327e-01 5.66461086e-01 1.77574247e-01 -6.96401536e-01
-5.36274910e-01 -7.66076982e-01 -1.11068428e+00 5.39031054e-04
-1.85686946e-01 -1.58737361e-01 8.77214670e-01 1.10630584e+00
3.48702610e-01 6.58849418e-01 2.67908752e-01 -1.04232585e+00
-4.13304061e-01 -1.14985776e+00 -2.85338223e-01 -4.35179062e-02
9.84214485e-01 -7.16818333e-01 -5.96598208e-01 -3.15790713e-01] | [14.362855911254883, 5.174940586090088] |
3bb907a4-6034-4790-971b-43dc38c04f1a | self-inspection-method-of-unmanned-aerial | 2303.09013 | null | https://arxiv.org/abs/2303.09013v1 | https://arxiv.org/pdf/2303.09013v1.pdf | Self-Inspection Method of Unmanned Aerial Vehicles in Power Plants Using Deep Q-Network Reinforcement Learning | For the purpose of inspecting power plants, autonomous robots can be built using reinforcement learning techniques. The method replicates the environment and employs a simple reinforcement learning (RL) algorithm. This strategy might be applied in several sectors, including the electricity generation sector. A pre-trained model with perception, planning, and action is suggested by the research. To address optimization problems, such as the Unmanned Aerial Vehicle (UAV) navigation problem, Deep Q-network (DQN), a reinforcement learning-based framework that Deepmind launched in 2015, incorporates both deep learning and Q-learning. To overcome problems with current procedures, the research proposes a power plant inspection system incorporating UAV autonomous navigation and DQN reinforcement learning. These training processes set reward functions with reference to states and consider both internal and external effect factors, which distinguishes them from other reinforcement learning training techniques now in use. The key components of the reinforcement learning segment of the technique, for instance, introduce states such as the simulation of a wind field, the battery charge level of an unmanned aerial vehicle, the height the UAV reached, etc. The trained model makes it more likely that the inspection strategy will be applied in practice by enabling the UAV to move around on its own in difficult environments. The average score of the model converges to 9,000. The trained model allowed the UAV to make the fewest number of rotations necessary to go to the target point. | ['Haoran Guan'] | 2023-03-16 | null | null | null | null | ['q-learning'] | ['methodology'] | [-1.87472165e-01 3.50293726e-01 -1.32624224e-01 1.72180772e-01
2.18014747e-01 -7.79392898e-01 3.79382074e-01 9.79162902e-02
-2.73010761e-01 9.68536377e-01 -6.04439497e-01 -6.84579492e-01
-5.13030589e-01 -1.13942170e+00 -5.70538461e-01 -6.90562010e-01
-3.07497114e-01 3.11914563e-01 4.25426550e-02 -6.70132220e-01
2.83577085e-01 7.58074164e-01 -1.35139489e+00 -4.61855769e-01
8.05048227e-01 7.80559003e-01 8.58016610e-01 9.67799723e-01
5.23892164e-01 8.58418405e-01 -7.25270391e-01 5.07616162e-01
4.40255553e-01 -2.19867885e-01 -9.17597473e-01 3.32693636e-01
-5.45505881e-01 -7.33425677e-01 -4.83844131e-02 8.21857810e-01
3.73829961e-01 3.63593638e-01 5.32013953e-01 -1.39969718e+00
-1.97952345e-01 3.11984628e-01 -1.05666175e-01 9.53239575e-02
3.54728967e-01 6.60500884e-01 7.60495245e-01 -2.91013956e-01
2.29890794e-01 9.63659823e-01 5.39141417e-01 1.06108680e-01
-7.02435374e-01 -3.05229783e-01 -9.20007080e-02 1.37514442e-01
-6.76362991e-01 3.33564043e-01 3.76602530e-01 -4.09579873e-01
1.22560120e+00 -2.57401675e-01 1.21219873e+00 5.47975957e-01
8.02158654e-01 3.76557231e-01 8.39347839e-01 -4.98181403e-01
6.28388107e-01 -8.59376490e-02 -7.81310380e-01 6.00555182e-01
3.93234670e-01 8.83887827e-01 3.05361152e-01 2.84435272e-01
9.25049245e-01 -1.96247786e-01 -1.13854688e-02 -7.23989487e-01
-8.76388371e-01 1.12565196e+00 7.12792695e-01 3.51828933e-01
-7.13618338e-01 1.47812575e-01 4.05957162e-01 6.19266093e-01
-4.12402242e-01 1.15392244e+00 -8.64230037e-01 -2.90735234e-02
-5.42633593e-01 3.92392516e-01 6.76844180e-01 6.23399854e-01
8.25186729e-01 7.19446242e-01 1.08750194e-01 3.08078855e-01
4.34887737e-01 5.84106684e-01 3.36127847e-01 -1.38934577e+00
-4.19544503e-02 5.75630426e-01 5.45506775e-01 -6.99137568e-01
-6.52881801e-01 -2.99979448e-01 -4.96856153e-01 8.46334934e-01
1.51193783e-01 -1.03943038e+00 -7.66585588e-01 1.25019073e+00
4.77771163e-01 -1.61655676e-02 4.43367362e-01 6.68121040e-01
8.48490521e-02 8.83088350e-01 -8.22372213e-02 -3.76129508e-01
8.64075243e-01 -8.67385209e-01 -6.58456504e-01 -2.85979301e-01
8.11600685e-01 -4.12918717e-01 5.95718920e-01 5.35315454e-01
-9.08520937e-01 -9.30011511e-01 -1.33161390e+00 9.52479243e-01
-7.24216700e-01 9.36945006e-02 6.65567636e-01 4.49179441e-01
-1.25123382e+00 1.14591956e+00 -8.41608286e-01 -5.52243531e-01
-1.05161823e-01 6.26688600e-01 -6.85133636e-02 8.42958316e-02
-1.45957136e+00 1.48609459e+00 8.94842863e-01 1.57124624e-01
-1.63557363e+00 -2.53116876e-01 -1.06375754e+00 5.92894964e-02
4.60314333e-01 -5.06064951e-01 1.60095322e+00 -9.36363459e-01
-1.90689170e+00 7.17838705e-02 7.69593060e-01 -5.97964108e-01
-1.80918183e-02 -6.04664162e-02 -1.40842140e-01 1.67056128e-01
1.37694299e-01 7.54533410e-01 9.62865531e-01 -1.05249298e+00
-9.99131501e-01 -6.76789358e-02 5.60445607e-01 4.58875060e-01
-1.84343550e-02 -4.17978555e-01 5.44990003e-01 -1.57926530e-01
-3.94713163e-01 -9.61999655e-01 -7.15166688e-01 -4.08826143e-01
7.21832216e-02 -3.00065547e-01 1.14210188e+00 -3.38437408e-01
5.64170122e-01 -1.75990736e+00 1.17987737e-01 2.08111703e-01
-4.96396601e-01 4.06929016e-01 -2.50395149e-01 8.48680019e-01
-2.51675248e-01 -2.67214507e-01 -2.87500229e-02 7.16196954e-01
-7.90896937e-02 6.76985025e-01 1.14450403e-01 1.07447714e-01
5.63760042e-01 5.40284574e-01 -1.39728653e+00 2.54562069e-02
5.23391008e-01 -1.67292356e-01 -3.47978115e-01 5.53900599e-01
-3.22612554e-01 4.42973942e-01 -5.50089598e-01 6.44191444e-01
5.20024002e-01 1.27085865e-01 3.02788436e-01 9.45789963e-02
-4.50725406e-01 -1.81135461e-01 -1.23626184e+00 1.36491561e+00
-5.07578194e-01 2.85506278e-01 4.58287776e-01 -1.40723515e+00
1.06924272e+00 5.01406014e-01 7.42447197e-01 -5.06229103e-01
1.73123628e-01 5.40980361e-02 2.37486303e-01 -6.92693830e-01
5.18323600e-01 -4.54808734e-02 1.46865815e-01 1.54699922e-01
3.82507920e-01 -1.16906393e+00 4.49592799e-01 -1.62193641e-01
1.32054293e+00 4.80334878e-01 4.63864446e-01 -1.97028905e-01
5.35072029e-01 3.36336374e-01 5.20440161e-01 4.56652492e-01
-4.19962198e-01 -4.66892868e-01 5.04876673e-01 -5.10414839e-01
-1.00360394e+00 -6.50810361e-01 1.69301406e-01 8.88068855e-01
5.62127829e-02 1.76560488e-02 -6.84516668e-01 -6.62037015e-01
1.29503563e-01 9.47001696e-01 -4.03997838e-01 -4.74505156e-01
-7.75260553e-02 -5.49558938e-01 -2.51924936e-02 4.18985993e-01
5.16659617e-01 -1.48358667e+00 -1.31978917e+00 5.63664198e-01
4.09727663e-01 -8.53010833e-01 4.17442471e-01 8.49491537e-01
-1.09041262e+00 -1.27313483e+00 -4.46249545e-01 -7.80786633e-01
6.38777673e-01 5.57427555e-02 8.54790151e-01 9.09336284e-02
-5.30484170e-02 6.32822871e-01 -6.42719686e-01 -6.76457226e-01
-6.45257473e-01 -7.15780556e-02 2.42763817e-01 -9.82422292e-01
-1.96279705e-01 -2.79212117e-01 -3.59383970e-01 3.48054767e-01
-6.97998881e-01 -5.48099816e-01 8.37766886e-01 1.07894754e+00
1.97466269e-01 8.91979456e-01 7.66164899e-01 -2.21624240e-01
8.73885870e-01 -3.62483710e-01 -1.05781078e+00 8.71520713e-02
-8.10828745e-01 -1.32721499e-01 7.42429316e-01 -1.65771022e-01
-5.30683279e-01 3.23040634e-01 -5.54436073e-02 -2.66490608e-01
-3.89293998e-01 6.28528833e-01 4.31903750e-02 -2.13420808e-01
3.98151457e-01 -3.96269783e-02 4.28928554e-01 4.21935022e-01
1.23861104e-01 3.85969877e-01 1.03696428e-01 -2.55272031e-01
1.10046434e+00 -2.21012846e-01 2.21610963e-01 -6.58862352e-01
-4.39881921e-01 2.55817398e-02 -6.44017935e-01 -5.91539621e-01
6.64599061e-01 -6.54500008e-01 -9.58372414e-01 3.58940005e-01
-9.20110703e-01 -9.41735923e-01 -7.20396399e-01 5.68229914e-01
-7.78737664e-01 1.73246324e-01 -3.44165057e-01 -1.00688481e+00
-2.33293474e-01 -1.30990458e+00 4.48868215e-01 7.17522621e-01
1.27270535e-01 -9.80323672e-01 2.12051317e-01 -1.52634010e-01
4.60878968e-01 3.48603457e-01 8.08124959e-01 -3.33408237e-01
-3.08738172e-01 -1.05873622e-01 2.65594423e-01 7.53876269e-01
2.28470162e-01 3.13818991e-01 -4.09575939e-01 -6.98638260e-01
-1.11386977e-01 -7.25312769e-01 -5.23919016e-02 6.02355957e-01
6.97336495e-01 -3.49621028e-02 -1.15426317e-01 4.08622921e-02
1.45036006e+00 9.73485708e-01 4.64742243e-01 8.72606575e-01
7.51536265e-02 5.48070908e-01 1.51683676e+00 6.72974169e-01
1.96186945e-01 2.05253318e-01 1.47264123e+00 -1.16659559e-01
8.27257693e-01 -8.01772550e-02 5.23067772e-01 4.19010371e-01
-1.96730793e-01 -3.40830944e-02 -7.20039904e-01 3.24398875e-01
-1.59445703e+00 -9.69811976e-01 3.91526818e-01 1.97559571e+00
9.18487534e-02 4.55455989e-01 1.00345217e-01 4.13862705e-01
3.14276963e-01 -8.34925994e-02 -8.56667459e-01 -9.76423264e-01
3.16659659e-01 1.38079882e-01 7.71553338e-01 3.68277490e-01
-1.07820976e+00 7.98188925e-01 6.66523027e+00 1.79249704e-01
-1.01213288e+00 -4.92546499e-01 3.60403925e-01 6.30143166e-01
2.13352859e-01 2.21749842e-01 -3.77208203e-01 6.86194450e-02
1.09302020e+00 1.25028446e-01 7.23227859e-01 9.76788819e-01
7.46108115e-01 -5.78190744e-01 -7.22478211e-01 2.90685683e-01
-4.70906585e-01 -8.24083030e-01 -3.22209626e-01 5.41693568e-02
6.76484823e-01 3.24772000e-02 -2.55899280e-01 8.21348310e-01
7.62056589e-01 -8.62302125e-01 4.39167976e-01 4.25523639e-01
3.70095134e-01 -1.10800421e+00 1.16702938e+00 6.46449089e-01
-1.04521263e+00 -1.00973380e+00 -6.19187295e-01 -4.37239677e-01
7.79832006e-02 1.69256300e-01 -1.30882752e+00 9.49835896e-01
7.54179120e-01 6.20446861e-01 -2.18962342e-01 9.15920377e-01
-4.85429555e-01 4.71593529e-01 -1.40018195e-01 -5.48369944e-01
8.53571653e-01 -5.37182927e-01 4.29822654e-01 4.36791688e-01
6.39322698e-01 -5.20370156e-02 5.28786421e-01 5.66941679e-01
6.14937127e-01 -2.14630738e-01 -1.09177732e+00 -1.84043139e-01
4.33870971e-01 1.39306021e+00 -6.48997605e-01 -1.04108863e-01
5.71879931e-02 6.39461756e-01 3.18409652e-02 1.33700222e-01
-7.32384861e-01 -6.13495588e-01 4.15155888e-01 -2.08387673e-01
5.44987023e-01 -3.19295049e-01 1.47750676e-01 -1.66756153e-01
-5.98158658e-01 -9.15118098e-01 5.74315861e-02 -1.11899483e+00
-6.66734576e-01 2.99486935e-01 -8.79674852e-02 -1.34103346e+00
-8.27262044e-01 -1.06824732e+00 -9.55615282e-01 5.90899408e-01
-1.64738214e+00 -6.67491257e-01 -2.47903615e-01 3.10473979e-01
6.25544786e-01 -5.01286983e-01 9.87898529e-01 -3.70443016e-01
-6.26055896e-01 -3.17065597e-01 1.10949054e-01 -1.12822771e-01
2.78950900e-01 -1.58361435e+00 -1.19849235e-01 6.17829978e-01
-4.51911718e-01 -1.63896419e-02 8.96885335e-01 -6.07663095e-01
-1.51297987e+00 -9.63204503e-01 8.75036195e-02 5.62541485e-02
7.55500853e-01 4.11287040e-01 -3.51737559e-01 3.83123755e-01
7.75607526e-01 -1.84937090e-01 2.54931629e-01 -3.36643487e-01
8.91641378e-01 -3.52323771e-01 -1.35986900e+00 4.15563494e-01
1.89363807e-01 2.31523644e-02 -4.39704746e-01 4.58721787e-01
5.89221954e-01 -5.12352943e-01 -1.01767325e+00 7.17844844e-01
2.55679488e-01 -8.00090492e-01 6.03773952e-01 -6.41395092e-01
1.89255774e-01 -3.54749590e-01 1.68935478e-01 -2.20142961e+00
-5.24291158e-01 -7.30655491e-01 -8.83329585e-02 6.33971930e-01
4.33149219e-01 -5.40123105e-01 6.82678401e-01 -7.12106645e-04
-3.40660006e-01 -1.09106088e+00 -6.18477166e-01 -7.06711769e-01
2.27285661e-02 4.94958274e-02 5.63989937e-01 7.44040191e-01
1.50833786e-01 2.84278631e-01 -8.71390775e-02 6.83088899e-01
2.04830706e-01 -2.37278625e-01 6.83355570e-01 -1.14982784e+00
-2.81104058e-01 -1.92379504e-02 -3.83744717e-01 -6.48996472e-01
3.65990400e-02 -2.97219366e-01 1.99018583e-01 -1.99893653e+00
-7.46386528e-01 -3.12786639e-01 -2.63109177e-01 5.93214691e-01
1.45748213e-01 -7.36041248e-01 1.76962078e-01 -3.92102808e-01
-1.92132801e-01 5.50472856e-01 1.66353631e+00 -2.16777280e-01
-3.69516194e-01 4.94192481e-01 -3.01109195e-01 7.05869794e-01
1.33072698e+00 -1.73269853e-01 -7.38001645e-01 -1.14004269e-01
2.16514915e-01 4.85567510e-01 1.44408524e-01 -1.21325982e+00
1.34417504e-01 -3.84556115e-01 6.44582629e-01 -6.01136863e-01
8.74277577e-03 -1.27322400e+00 7.52301712e-05 1.32721353e+00
1.86035618e-01 6.22366250e-01 4.35624957e-01 4.22603518e-01
-2.03810245e-01 -7.93779910e-01 8.87238741e-01 -4.17792171e-01
-8.77633452e-01 -7.66601264e-02 -1.06044829e+00 -4.98716474e-01
1.61864507e+00 -3.74143004e-01 9.62216854e-02 -4.77024496e-01
-6.86444283e-01 7.93093443e-01 3.59225757e-02 1.87417939e-01
6.08705938e-01 -8.49398673e-01 -4.07765955e-01 2.45045215e-01
-4.28249419e-01 1.29708871e-01 -7.70347491e-02 5.95679522e-01
-7.90901542e-01 4.39951360e-01 -8.49241972e-01 -5.55146754e-01
-5.82826614e-01 6.90908909e-01 8.59083474e-01 -6.57903552e-01
-7.38079399e-02 3.42911720e-01 -5.46272933e-01 -7.35187173e-01
2.44753603e-02 -4.58461821e-01 -6.58199430e-01 1.46187305e-01
4.81400229e-02 4.80456144e-01 1.26662314e-01 -6.58537596e-02
8.55778530e-02 2.88150102e-01 2.30126277e-01 7.65751079e-02
1.30372572e+00 3.83386202e-02 1.28278762e-01 1.67304769e-01
3.31908107e-01 -7.38745451e-01 -1.52968347e+00 6.21770620e-01
-8.90052039e-03 4.68409210e-02 2.25513995e-01 -9.67193723e-01
-9.54270244e-01 6.25547647e-01 9.62776065e-01 7.15519369e-01
1.23484933e+00 -7.94787288e-01 2.94892937e-01 8.13976228e-01
5.26636839e-01 -1.58868384e+00 3.68334681e-01 8.48007917e-01
7.60963738e-01 -1.16222048e+00 6.10001478e-03 1.94710225e-01
-7.11906731e-01 1.36356354e+00 9.49136972e-01 -3.74914497e-01
6.22760355e-01 3.77035052e-01 8.35938305e-02 -1.23258680e-01
-7.10972726e-01 -1.76360816e-01 -5.00909328e-01 1.06191421e+00
-6.16820306e-02 1.28667757e-01 -2.98752692e-02 -6.22384697e-02
3.33906822e-02 4.04001437e-02 7.57452011e-01 1.25007296e+00
-9.41926956e-01 -1.01115096e+00 -5.29783785e-01 3.92962545e-01
3.21028419e-02 4.64526296e-01 7.83949420e-02 1.02992451e+00
4.20440078e-01 1.08997869e+00 -9.24835354e-02 -2.27706432e-01
7.32176661e-01 -4.79886323e-01 2.54546493e-01 -6.29258275e-01
-6.96095169e-01 -1.96456268e-01 -6.79610670e-02 -3.90024573e-01
-3.30923110e-01 -5.74403048e-01 -1.51201928e+00 1.60212994e-01
-5.29264748e-01 3.63702029e-01 7.18454897e-01 9.00944173e-01
-3.79567742e-02 8.99185061e-01 1.19750023e+00 -1.17524338e+00
-1.10210526e+00 -1.05073416e+00 -8.43386412e-01 -4.50251251e-01
2.15576038e-01 -9.53194916e-01 -3.05825353e-01 -3.91231447e-01] | [4.628355026245117, 1.8902912139892578] |
1f6832ed-50c6-4ed5-92ce-c9a33e044dd2 | ternary-twitter-sentiment-classification-with | null | null | https://aclanthology.org/w18-6215 | https://aclanthology.org/w18-6215.pdf | Ternary Twitter Sentiment Classification with Distant Supervision and Sentiment-Specific Word Embeddings | null | ['Björn Gambäck', 'Frederik Gørvell de Lichtenberg', 'Mats Byrkjeland'] | 2018-10-01 | null | https://aclanthology.org/W18-6215 | https://aclanthology.org/W18-6215.pdf | emnlp-2018-10 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-2.44508207e-01 3.89024585e-01 -2.65282035e-01 -2.15905145e-01
-8.60921741e-02 -7.76765764e-01 4.48510379e-01 -7.23253429e-01
-5.48377395e-01 1.31954515e+00 3.66348401e-02 -9.49533224e-01
-2.40340635e-01 -1.05564880e+00 -8.44053447e-01 -8.75781775e-01
-7.42435038e-01 6.86515033e-01 1.44298598e-01 -6.52004302e-01
8.47113907e-01 6.29996777e-01 -1.62287033e+00 5.77558100e-01
6.81926727e-01 5.49597681e-01 5.66466339e-02 1.04480565e+00
6.13576770e-02 1.59847176e+00 -6.05450153e-01 -4.56729174e-01
1.43710867e-01 -1.88022718e-01 -5.35770595e-01 -3.36825520e-01
7.91536197e-02 -5.88986814e-01 -3.10633808e-01 6.07356608e-01
1.14270830e+00 -5.53557873e-02 1.07025254e+00 -1.51067197e+00
-7.97060013e-01 5.44234335e-01 2.30399013e-01 1.20632313e-01
6.10446513e-01 -1.79472014e-01 3.51212233e-01 -1.36621380e+00
6.32680655e-01 8.64870071e-01 9.47046995e-01 5.99273622e-01
-1.20647264e+00 -5.60552299e-01 -7.17137158e-01 -2.57835120e-01
-1.59567785e+00 -6.30042374e-01 6.23617955e-02 -3.34735900e-01
1.69142139e+00 6.72587395e-01 1.35647905e+00 1.47341585e+00
1.36649036e+00 4.72517729e-01 1.13904178e+00 -3.16340059e-01
3.65351558e-01 6.43754780e-01 -2.29148138e-02 5.51889002e-01
1.19686353e+00 6.76312149e-01 -5.60917914e-01 -7.92779744e-01
1.11365557e+00 -3.12949389e-01 2.60771513e-01 -7.06989706e-01
-1.06566191e+00 7.88466871e-01 -1.33808568e-01 2.20882341e-01
-3.46203089e-01 -9.92989019e-02 2.24548295e-01 4.12961632e-01
-2.94082850e-01 4.77862865e-01 -8.89795244e-01 -1.87686309e-01
-1.01472771e+00 1.91863477e-01 1.30615628e+00 1.45811689e+00
4.72722203e-03 3.70100170e-01 1.66493103e-01 2.89523482e-01
5.29429734e-01 1.05927479e+00 4.36951250e-01 -1.33365822e+00
-9.55119133e-02 4.20878716e-02 6.12505674e-01 -1.13473701e+00
-7.12930143e-01 -2.41300568e-01 -9.54724610e-01 5.18771529e-01
-1.59270123e-01 -4.59180593e-01 -7.59686172e-01 4.79171664e-01
-1.81984559e-01 -3.19489211e-01 5.45619786e-01 2.69257158e-01
7.28577793e-01 1.45816103e-01 9.25893057e-03 -4.73371416e-01
8.38296890e-01 -1.26899445e+00 -1.11426139e+00 2.33304992e-01
6.82289064e-01 -1.11704338e+00 4.06794518e-01 4.29345578e-01
-1.87200487e+00 -9.69799384e-02 -1.08317137e+00 1.19022481e-01
-7.97166348e-01 -2.57194549e-01 6.66624963e-01 1.49922979e+00
-1.63972652e+00 6.47176981e-01 -6.05021298e-01 5.40335439e-02
-4.73218322e-01 9.07614470e-01 -3.47142309e-01 2.90655226e-01
-1.28751957e+00 9.52449858e-01 -1.04219764e-01 3.03568155e-01
-3.30324143e-01 -1.84501112e-01 -1.03385210e+00 -6.24127567e-01
-2.59417385e-01 -1.07312346e+00 1.36302292e+00 1.09815992e-01
-1.30379498e+00 1.00962746e+00 -4.14315462e-01 -1.14109404e-01
6.61359608e-01 -1.52096570e-01 -8.18174660e-01 1.58321753e-01
2.87593510e-02 3.47163439e-01 5.30214727e-01 -9.69232559e-01
-6.31655395e-01 -1.00153096e-01 -2.54188716e-01 2.48816490e-01
1.43904146e-03 2.05205917e-01 6.38460100e-01 -9.53604504e-02
1.91062525e-01 -8.61961663e-01 -3.21419090e-01 -7.61404037e-01
-1.77353188e-01 -3.84736449e-01 5.59762537e-01 -2.92163283e-01
1.33848476e+00 -1.84365213e+00 -4.54683900e-01 2.94456817e-03
2.53527164e-01 3.89016122e-02 2.33101144e-01 1.23831403e+00
-2.43849114e-01 1.06667292e+00 3.98064762e-01 9.52541176e-03
3.13382268e-01 5.07273376e-01 2.86674555e-02 2.51206756e-01
9.43330750e-02 1.15280068e+00 -1.09214199e+00 -5.20940363e-01
5.57267785e-01 4.60968643e-01 -3.84588838e-01 6.24593735e-01
9.30012882e-01 2.99172718e-02 -8.94021094e-02 1.33241999e+00
1.05377018e+00 -2.90812284e-01 -2.46628653e-02 3.44544262e-01
-6.44541264e-01 6.00222833e-02 -7.41804183e-01 9.49429095e-01
1.31349964e-02 1.99757457e-01 1.74998537e-01 -6.80728197e-01
3.94654661e-01 1.10545933e+00 4.45659161e-01 -9.24297392e-01
-1.69569388e-01 9.00310338e-01 1.03920102e-01 -9.62983370e-01
6.40755475e-01 -1.72503099e-01 7.69825354e-02 1.40852645e-01
-3.61913115e-01 -3.84120375e-01 5.52458428e-02 4.19574350e-01
1.00224769e+00 -1.33707494e-01 5.48685491e-01 -8.72328460e-01
6.88946426e-01 -2.01589778e-01 3.00597936e-01 1.23978651e+00
-2.90862411e-01 9.21478033e-01 3.19022417e-01 -5.37574828e-01
-5.72780192e-01 -1.20098352e+00 -4.80556756e-01 9.86326516e-01
-1.58388391e-01 -3.87671381e-01 -7.10155964e-01 -2.63155133e-01
-1.85071185e-01 4.93483543e-01 -5.93591571e-01 5.23159266e-01
-5.14705896e-01 -6.85349941e-01 7.65277922e-01 3.17972392e-01
6.49624616e-02 -1.49228525e+00 -5.70813596e-01 3.34754556e-01
-1.07132711e-01 -5.39811492e-01 -1.57590687e-01 6.12205923e-01
-1.19366693e+00 -1.01526415e+00 -8.75530541e-01 -1.03603637e+00
7.18889177e-01 2.44572356e-01 1.18463767e+00 2.84042627e-01
-2.87709564e-01 6.49904788e-01 -1.31030381e-02 -2.67939389e-01
-3.16364132e-02 -5.78750819e-02 5.01899779e-01 -1.07851994e+00
6.46393359e-01 -3.92214775e-01 -1.00622165e+00 5.35368681e-01
-3.85821253e-01 -2.04611585e-01 7.85597622e-01 1.04281425e+00
1.42742008e-01 -1.33182362e-01 2.41971418e-01 -1.16735983e+00
1.11498392e+00 -4.85333532e-01 1.30341025e-02 2.62960047e-01
-8.53362501e-01 -4.53584343e-01 2.27881029e-01 -1.33037493e-02
-6.33760452e-01 -2.51423597e-01 -4.14545417e-01 3.22794706e-01
-4.06866640e-01 3.34393084e-02 2.24873364e-01 -5.26000679e-01
5.21005154e-01 1.50258854e-01 -4.07556817e-02 1.21735156e-01
6.30975887e-02 3.75316888e-01 3.67786407e-01 -7.01224983e-01
5.19809723e-01 1.93325520e-01 -3.00995037e-02 -1.28791058e+00
2.66786069e-01 -1.64763972e-01 -6.60750508e-01 -2.69126564e-01
7.75121748e-01 -9.50198710e-01 -6.35705650e-01 5.21708906e-01
-9.54387486e-01 -1.36340424e-01 -5.25335252e-01 6.30965471e-01
-7.02474535e-01 2.94757709e-02 -4.17852998e-01 -1.17976546e+00
-4.78847146e-01 -1.38869667e+00 9.59160328e-01 3.85633826e-01
-2.05811903e-01 -1.19790471e+00 3.38975668e-01 8.99848714e-02
4.29175526e-01 7.39965662e-02 5.03531873e-01 -2.80004859e-01
-2.98304528e-01 -2.85470933e-01 1.59573574e-02 -1.65876746e-01
-3.09365243e-02 4.80770379e-01 -5.48391879e-01 -4.37962860e-01
2.44738594e-01 -4.86500829e-01 -7.51581416e-02 7.11018145e-01
4.09077644e-01 -6.02920949e-01 -7.12544680e-01 5.56552589e-01
1.27709007e+00 4.72497940e-01 5.88340998e-01 6.77552521e-01
9.28869769e-02 5.17871320e-01 4.48484391e-01 2.08113909e-01
1.07506551e-01 1.08483166e-01 1.14217520e-01 -2.17137560e-01
9.45900455e-02 -1.75644696e-01 4.44304377e-01 1.09980536e+00
-8.23837519e-01 -2.15223029e-01 -4.88283783e-01 6.68056428e-01
-1.69156313e+00 -1.31570101e+00 -6.70152605e-01 9.76623476e-01
3.78074080e-01 3.38884830e-01 1.58041969e-01 1.21232189e-01
5.47049284e-01 -4.76693302e-01 -1.27305150e-01 -9.16352808e-01
-3.82233024e-01 4.21136379e-01 7.14838982e-01 6.51479006e-01
-6.10602558e-01 5.08361042e-01 1.25505590e+01 5.49717486e-01
-5.10860123e-02 1.27314463e-01 3.14815581e-01 1.99914098e-01
-2.24020526e-01 1.62632689e-01 -1.04165065e+00 9.26906094e-02
1.64754760e+00 -4.36735392e-01 3.93212438e-01 6.44444764e-01
3.94769877e-01 -1.96429595e-01 -9.78620768e-01 5.21999180e-01
1.12709433e-01 -1.50397491e+00 -3.76861334e-01 9.64888573e-01
5.80093563e-01 -1.81895152e-01 6.81941509e-01 4.33209062e-01
7.39425182e-01 -1.14519978e+00 7.88427055e-01 3.18626672e-01
8.87266219e-01 -8.18826377e-01 1.07315242e+00 2.76328743e-01
-1.04355454e+00 1.15487054e-02 -5.90443075e-01 -9.22581375e-01
1.62244976e-01 2.02188566e-01 -5.61583102e-01 4.06869203e-01
6.19966745e-01 2.15250537e-01 -8.53338003e-01 1.45581341e+00
-4.29231405e-01 4.22687940e-02 -3.00976187e-01 -4.39922899e-01
4.83290881e-01 5.28384298e-02 3.26606274e-01 1.02295005e+00
2.89785862e-01 5.45581102e-01 -2.62208551e-01 2.92286038e-01
6.63503349e-01 5.84383309e-03 -1.38496268e+00 -8.68683681e-02
3.54620427e-01 9.69217539e-01 -3.99494410e-01 -3.56748998e-02
-6.11456513e-01 6.62115753e-01 -1.66942239e-01 5.75146377e-01
-2.92761147e-01 -8.89406264e-01 4.33969021e-01 -2.30117049e-02
-2.88469017e-01 -1.35338321e-01 -3.23372036e-01 -1.02459288e+00
-6.06416404e-01 -3.90122980e-01 7.32179061e-02 -8.63867998e-01
-1.79313409e+00 6.85844719e-01 -3.50327820e-01 -1.03178740e+00
-9.29034173e-01 -9.72984850e-01 -2.90242791e-01 6.67755604e-01
-6.59234166e-01 -1.30476952e+00 1.77773401e-01 4.65253919e-01
1.82401046e-01 -5.88519633e-01 1.09945750e+00 -1.42670095e-01
-3.18711132e-01 1.08882773e+00 3.51253241e-01 -3.05974871e-01
8.32965672e-01 -1.16271269e+00 4.23785180e-01 1.09189451e-02
-6.59259140e-01 1.28460062e+00 6.81860268e-01 -6.81785107e-01
-1.16639841e+00 -1.52893141e-01 1.38110602e+00 -9.15031374e-01
5.27874053e-01 -1.02810413e-01 1.46300867e-01 6.57190859e-01
9.07351732e-01 -5.09461224e-01 1.08935249e+00 -1.13616258e-01
4.93123859e-01 8.23448122e-01 -1.38805819e+00 5.99272728e-01
1.21222687e+00 -4.81854081e-01 -8.27020168e-01 3.28937650e-01
9.94227231e-01 -5.60786903e-01 -1.34066856e+00 6.99696779e-01
1.14132881e+00 -1.04883564e+00 1.54237401e+00 -1.16912270e+00
4.08816934e-01 1.70770600e-01 -2.18455836e-01 -6.52935803e-01
-5.24122715e-01 -8.07741940e-01 -3.97186369e-01 3.40213925e-01
8.64144087e-01 -1.16800272e+00 5.53630650e-01 1.39166510e+00
-3.58625412e-01 -4.51715857e-01 -8.63314152e-01 -1.06149745e+00
-4.66870656e-03 1.81816339e-01 3.85523647e-01 8.18638563e-01
7.09820271e-01 2.02331603e-01 -2.05201268e-01 -4.37674336e-02
6.59316301e-01 -4.87512946e-01 5.82342803e-01 -8.74011815e-01
2.49149442e-01 -3.85976404e-01 -2.94489831e-01 -2.90531307e-01
-3.26795012e-01 -4.87107635e-01 -8.59293997e-01 -1.53084815e+00
8.49664882e-02 1.97367743e-01 -2.33256146e-01 3.44385266e-01
6.77823246e-01 2.45789096e-01 -2.27382466e-01 1.17054820e-01
-1.86927155e-01 -9.06788632e-02 9.18564737e-01 1.10557920e-03
-2.17724726e-01 3.92412305e-01 -1.01716614e+00 9.60534811e-01
7.85404518e-02 -6.09580100e-01 -5.51487744e-01 1.64939865e-01
5.71885824e-01 3.57775450e-01 4.49405432e-01 -9.70509052e-01
9.87445056e-01 -4.32846308e-01 6.12257063e-01 -1.43385470e+00
-2.65485227e-01 -9.00021493e-01 3.61826897e-01 6.29400074e-01
4.08042461e-01 3.53796333e-01 2.26811334e-01 -9.13986787e-02
-2.01378226e-01 -6.44658387e-01 4.43001747e-01 -7.45609999e-01
-4.55086857e-01 -3.90635371e-01 -1.02313221e+00 -9.67486948e-03
7.88939476e-01 -4.61360693e-01 -6.47767961e-01 -1.95199717e-02
-1.44608676e+00 7.30553791e-02 8.33678424e-01 1.34732248e-02
6.98022604e-01 -1.44444084e+00 -1.11748077e-01 6.18802965e-01
-5.42432606e-01 -4.98194307e-01 2.20674574e-01 1.06156552e+00
-1.31528914e+00 1.26915324e+00 -5.21896303e-01 -2.05424115e-01
-8.96585643e-01 7.19095051e-01 3.15837234e-01 1.75098896e-01
-3.93049836e-01 7.00151443e-01 1.12239346e-02 -6.58594191e-01
1.24011397e-01 2.97559589e-01 -4.87763733e-01 -2.52128989e-01
9.22277749e-01 1.17918742e+00 1.32526740e-01 -3.91715407e-01
-5.58191121e-01 4.39446121e-01 2.08604142e-01 -6.12734079e-01
1.38662541e+00 -2.20989168e-01 -7.44191706e-01 5.00318348e-01
7.50003338e-01 -7.33495504e-02 -3.28966863e-02 1.27011871e+00
4.80256975e-03 -5.47239244e-01 -3.11336160e-01 -7.82959402e-01
-3.42039734e-01 5.93214929e-01 6.17733955e-01 4.59080935e-01
9.69997942e-01 -6.47576571e-01 7.70057857e-01 7.64313757e-01
7.83133745e-01 -1.22997761e+00 -6.36951804e-01 3.69483382e-01
9.39836800e-01 -9.30032432e-01 1.30400375e-01 -3.14221263e-01
-5.55534840e-01 1.08689797e+00 5.38173854e-01 -1.62142932e-01
1.17091513e+00 6.96656525e-01 -1.75611794e-01 -4.53250527e-01
-1.04549897e+00 9.83817875e-02 2.10727617e-01 1.32608259e+00
8.26061428e-01 1.41275764e-01 -1.40426433e+00 8.46834600e-01
-6.34404242e-01 4.76996809e-01 9.47268784e-01 1.55018985e+00
-5.12773991e-01 -1.41692734e+00 -4.41162825e-01 2.31142953e-01
-2.82854587e-01 -3.49930972e-01 -8.62966061e-01 1.09929097e+00
1.97083559e-02 1.37330377e+00 -4.48758155e-01 -4.63731140e-01
8.27405751e-01 3.19600284e-01 4.41484600e-01 -1.25142381e-01
-8.94688129e-01 6.48750722e-01 3.84649158e-01 -1.05672014e+00
-7.94483483e-01 -1.27049911e+00 -1.08188879e+00 -1.16740942e+00
-3.07217389e-01 4.65777248e-01 2.31674016e-01 3.19533348e-01
-2.09303293e-02 1.56313553e-03 6.67447627e-01 -1.00111675e+00
-1.97397545e-01 -6.07395887e-01 -1.20775104e+00 -5.56946039e-01
3.94425273e-01 -5.85869551e-01 -1.15223050e+00 -6.24734610e-02] | [-1.5392014980316162, 15.869206428527832] |
63e37d62-53ef-46f1-9211-0c6a660a4233 | representation-justification-and-explanation | 1812.05362 | null | https://arxiv.org/abs/1812.05362v2 | https://arxiv.org/pdf/1812.05362v2.pdf | Representation, Justification and Explanation in a Value Driven Agent: An Argumentation-Based Approach | Ethical and explainable artificial intelligence is an interdisciplinary research area involving computer science, philosophy, logic, the social sciences, etc. For an ethical autonomous system, the ability to justify and explain its decision making is a crucial aspect of transparency and trustworthiness. This paper takes a Value Driven Agent (VDA) as an example, explicitly representing implicit knowledge of a machine learning-based autonomous agent and using this formalism to justify and explain the decisions of the agent. For this purpose, we introduce a novel formalism to describe the intrinsic knowledge and solutions of a VDA in each situation. Based on this formalism, we formulate an approach to justify and explain the decision-making process of a VDA, in terms of a typical argumentation formalism, Assumption-based Argumentation (ABA). As a result, a VDA in a given situation is mapped onto an argumentation framework in which arguments are defined by the notion of deduction. Justified actions with respect to semantics from argumentation correspond to solutions of the VDA. The acceptance (rejection) of arguments and their premises in the framework provides an explanation for why an action was selected (or not). Furthermore, we go beyond the existing version of VDA, considering not only practical reasoning, but also epistemic reasoning, such that the inconsistency of knowledge of the VDA can be identified, handled and explained. | ['Michael Anderson', 'Beishui Liao', 'Susan Leigh Anderson'] | 2018-12-13 | null | null | null | null | ['epistemic-reasoning'] | ['miscellaneous'] | [ 1.38074696e-01 1.35558951e+00 -2.19021648e-01 -4.61246014e-01
3.30197453e-01 -5.04687607e-01 1.02785647e+00 5.52817285e-01
-2.66607642e-01 9.31813657e-01 9.40781385e-02 -6.66960180e-01
-5.97649217e-01 -1.01271093e+00 -5.18972695e-01 -5.24879754e-01
6.00436389e-01 4.84498769e-01 8.65264460e-02 -3.71244818e-01
2.27369711e-01 3.31191599e-01 -1.55603039e+00 2.52449602e-01
8.86578023e-01 7.93174207e-01 -3.76813591e-01 1.37080416e-01
-3.06742519e-01 1.30444908e+00 -3.76321793e-01 -9.83608127e-01
-1.20456330e-01 -8.51196885e-01 -1.48609376e+00 2.16726065e-02
-4.35399234e-01 -2.41964608e-01 6.17571950e-01 1.30047810e+00
-3.13238442e-01 -2.43183643e-01 8.61508131e-01 -1.95209718e+00
-6.02871120e-01 1.18656135e+00 -5.15522324e-02 -6.08431995e-01
5.33207655e-01 -7.37895891e-02 8.94719660e-01 -2.18584642e-01
7.69776762e-01 1.33785677e+00 2.53150582e-01 7.96326160e-01
-9.00723934e-01 3.07643637e-02 3.93085748e-01 5.78619540e-01
-6.98222876e-01 -2.15069801e-01 7.96885252e-01 -5.75498104e-01
4.71794009e-01 4.59191203e-01 1.24097097e+00 8.09820473e-01
3.46452951e-01 4.93340731e-01 1.18701327e+00 -9.52784657e-01
9.48816836e-01 7.19931781e-01 6.31777525e-01 2.07148641e-01
1.12670290e+00 -9.67762526e-03 -2.59991437e-01 -3.75169039e-01
1.01283930e-01 -5.40634513e-01 -3.80126201e-02 -7.43707895e-01
-9.60065365e-01 1.04364717e+00 4.62120362e-02 4.62220848e-01
-6.35215580e-01 1.41153947e-01 3.90022576e-01 3.06067526e-01
-1.09440416e-01 3.08335483e-01 -3.29313874e-01 9.29323286e-02
-3.16119611e-01 3.42499048e-01 1.26735842e+00 2.11079165e-01
5.27159035e-01 -7.84647837e-02 3.31532747e-01 -3.40209156e-01
1.23401678e+00 6.19106054e-01 2.20411986e-01 -1.39443016e+00
-3.37162852e-01 1.11082947e+00 5.07286847e-01 -1.02999783e+00
-4.10448872e-02 -9.15689766e-02 -3.09384555e-01 1.00527620e+00
4.69089031e-01 4.25338857e-02 -1.12716369e-01 1.72119403e+00
8.20300877e-01 -1.96680143e-01 1.12112200e+00 8.58198047e-01
9.34314549e-01 3.21576148e-01 2.34620154e-01 -5.00867009e-01
1.49237072e+00 -4.32042539e-01 -1.30852866e+00 2.80464906e-02
8.12378883e-01 -2.80639350e-01 6.29469931e-01 5.52137017e-01
-1.03544247e+00 3.32528710e-01 -1.26087987e+00 6.06899802e-03
-2.71274686e-01 -1.22309618e-01 5.55515289e-01 5.04102170e-01
-7.97910213e-01 2.81008899e-01 -2.00724646e-01 -5.22311568e-01
1.89232439e-01 6.77592084e-02 -2.88696826e-01 4.92408186e-01
-1.48361313e+00 1.34390259e+00 7.01459467e-01 2.73741782e-01
-4.62808281e-01 1.44023374e-02 -8.12607884e-01 3.84273864e-02
6.56670868e-01 -7.91015387e-01 1.19144928e+00 -1.51294804e+00
-1.60500181e+00 1.22258079e+00 8.33187923e-02 -1.23259115e+00
1.06784594e+00 -4.67842110e-02 -6.00567520e-01 1.85895786e-01
1.56580493e-01 1.83524132e-01 6.22912943e-01 -1.62622070e+00
-6.02995038e-01 -2.88215220e-01 6.81918442e-01 8.42859317e-03
5.74253201e-02 7.88370427e-03 6.34530067e-01 8.22474062e-02
1.05933100e-01 -7.47165322e-01 1.56469885e-02 1.25241205e-01
-3.50870550e-01 -3.79349142e-01 8.13220561e-01 -1.53145269e-01
1.01430392e+00 -1.81073165e+00 -2.44558770e-02 5.06161153e-01
4.14618254e-01 2.93698251e-01 5.20187914e-01 4.70929354e-01
1.44077182e-01 4.29332525e-01 -3.56833607e-01 6.05618060e-01
5.37296474e-01 4.90534693e-01 -4.45149392e-01 3.50667804e-01
-1.07203551e-01 5.00863791e-01 -7.65349448e-01 -5.79397380e-01
4.71988171e-01 3.81957054e-01 -3.24074119e-01 -1.15050331e-01
-5.08413911e-01 3.46990436e-01 -9.45449591e-01 1.89436629e-01
4.69135076e-01 -9.71646085e-02 7.71932542e-01 -1.18987866e-01
-2.29731277e-01 1.55608915e-02 -1.34916270e+00 1.10683942e+00
-3.20227653e-01 3.49827081e-01 7.51811340e-02 -9.79090273e-01
1.01464081e+00 6.58352971e-01 6.61996379e-02 -3.62360746e-01
6.78077042e-01 7.92058110e-01 2.97830671e-01 -8.09529185e-01
3.39706577e-02 -6.11805171e-02 1.58855796e-01 9.93450940e-01
-6.14943683e-01 -2.81789660e-01 1.08350046e-01 3.55381995e-01
4.84043479e-01 4.62351948e-01 1.10315180e+00 -4.85537887e-01
1.25051498e+00 3.62731665e-01 4.88344193e-01 4.36666071e-01
-3.15253884e-01 -3.42410624e-01 9.34965611e-01 -1.05783451e+00
-8.53558660e-01 -4.83845800e-01 -2.35190108e-01 2.05099180e-01
4.44399059e-01 -6.75613284e-02 -1.25728977e+00 -8.62872005e-01
-8.60631745e-03 1.49924529e+00 -7.55617857e-01 -1.93954274e-01
-1.69880062e-01 -1.34461686e-01 2.51725823e-01 -2.18975261e-01
7.30108917e-01 -1.25563645e+00 -1.71225405e+00 4.48970385e-02
-1.42881528e-01 -1.01264083e+00 8.33362460e-01 6.02923706e-02
-6.07536852e-01 -1.60978961e+00 1.25643492e-01 -1.14061512e-01
7.81587541e-01 -2.72659570e-01 9.71999645e-01 6.10627830e-01
4.86214191e-01 3.37618560e-01 -4.94570076e-01 -9.36046004e-01
-1.25653803e+00 -5.47330618e-01 8.66575837e-02 1.89779118e-01
6.02268219e-01 -3.01960200e-01 -2.93822438e-01 8.71031433e-02
-9.89542305e-01 3.06921363e-01 1.72723278e-01 3.59019756e-01
2.46125564e-01 2.94897566e-03 7.14597702e-01 -1.10795021e+00
8.00415695e-01 -3.47976655e-01 -6.47461951e-01 7.36086309e-01
-1.08175421e+00 2.86802292e-01 2.99572349e-01 6.51298836e-02
-1.41359234e+00 -3.78775954e-01 2.84969807e-01 3.12859535e-01
-3.90412867e-01 5.69125354e-01 -7.34637797e-01 1.42574668e-01
7.08965063e-01 -3.94335628e-01 2.84263551e-01 1.51225463e-01
6.60712600e-01 6.94948554e-01 1.92679569e-01 -8.97218466e-01
5.09563982e-01 7.00895071e-01 4.57042843e-01 -4.81028706e-02
-7.02723801e-01 5.02156436e-01 -5.76278687e-01 -6.11295462e-01
7.65971005e-01 -3.73637438e-01 -1.19161355e+00 -1.83919165e-02
-1.47291076e+00 -4.54362063e-03 -7.82600462e-01 8.07883680e-01
-9.23431814e-01 2.10404143e-01 1.42226502e-01 -1.41713345e+00
-3.05139571e-01 -1.22481525e+00 2.31860802e-01 8.82458314e-02
-8.09614062e-01 -1.04409862e+00 -7.86065981e-02 5.26035428e-01
4.71359082e-02 5.43458045e-01 1.21323824e+00 -1.17698812e+00
-1.10819995e-01 -1.55461490e-01 1.86691850e-01 2.30271682e-01
1.33569196e-01 4.64662313e-01 -8.17627490e-01 3.67469907e-01
4.75822389e-01 -7.72452429e-02 2.96346247e-02 3.52388948e-01
1.82947487e-01 -8.25731993e-01 -5.13319224e-02 -2.97977209e-01
1.34680939e+00 5.39218128e-01 5.67653954e-01 9.72517252e-01
-1.65508404e-01 1.26307154e+00 9.98445392e-01 4.18517530e-01
3.96571010e-01 5.34679949e-01 9.15447354e-01 2.66656458e-01
1.60445765e-01 1.34008914e-01 1.73211351e-01 3.54313105e-02
-3.86244476e-01 -1.68268587e-02 -9.88778889e-01 2.10788891e-01
-2.47157788e+00 -1.05978465e+00 -8.56165707e-01 2.14512205e+00
6.19078577e-01 8.76488239e-02 -4.16187271e-02 5.34640074e-01
9.02085006e-01 -2.43164405e-01 -3.29786390e-01 -1.17498505e+00
-3.99176478e-01 -5.44975877e-01 -1.31030500e-01 9.05404568e-01
-5.48877299e-01 6.52839839e-01 5.41435385e+00 4.64249164e-01
-6.72301352e-01 9.62494910e-02 3.64045709e-01 3.89102459e-01
-1.06171370e+00 5.26932120e-01 -2.41365671e-01 1.12254374e-01
8.98075581e-01 -7.61349261e-01 2.52888024e-01 8.75677764e-01
5.13201296e-01 -5.15218973e-01 -1.18098438e+00 2.25556359e-01
-1.00374348e-01 -1.42655087e+00 3.20822656e-01 1.81551456e-01
5.54358900e-01 -1.05503559e+00 -2.61295557e-01 -3.80563080e-01
5.00859380e-01 -8.60172331e-01 1.49088657e+00 4.29328710e-01
3.09951743e-03 -7.71491289e-01 1.35998833e+00 4.51940566e-01
-5.05427182e-01 -5.47300018e-02 3.74004580e-02 -4.01398391e-01
1.17703393e-01 5.21249950e-01 -3.77034277e-01 8.61069858e-01
4.11158174e-01 3.72098714e-01 1.32215733e-03 4.22908992e-01
-9.32604194e-01 3.11548769e-01 -6.46478147e-04 -2.78744400e-01
3.12265992e-01 -7.04283237e-01 6.93915904e-01 5.02507985e-01
2.54612684e-01 3.28287542e-01 -5.74455798e-01 1.13761187e+00
3.16248983e-01 3.03702861e-01 -7.46630549e-01 2.36043140e-01
3.94654155e-01 8.73734057e-01 -7.99392760e-01 -4.91052359e-01
-6.84656277e-02 2.96673119e-01 -2.58283824e-01 3.44118357e-01
-8.23355496e-01 1.24469727e-01 2.62030721e-01 -2.18819380e-02
-5.07344663e-01 6.27932906e-01 -5.06167829e-01 -9.38014865e-01
1.61452428e-01 -1.03558695e+00 2.33897552e-01 -9.20704901e-01
-7.85238087e-01 7.41967857e-01 2.77890861e-01 -1.17621505e+00
-4.83554572e-01 -5.12916744e-01 -5.48529088e-01 5.36235690e-01
-1.52036369e+00 -1.32903755e+00 -1.29929572e-01 4.47376549e-01
-2.05949694e-01 -1.90466847e-02 1.09718800e+00 -4.93093312e-01
-2.03490481e-01 -3.26589793e-01 -4.45477903e-01 -4.07833874e-01
2.90980458e-01 -1.12015462e+00 -3.43934178e-01 9.12763238e-01
-6.39056042e-02 5.68988740e-01 1.21282089e+00 -7.11768389e-01
-9.44288611e-01 -3.93196493e-01 1.43060303e+00 -2.62015104e-01
8.31061959e-01 3.93859267e-01 -9.72007811e-01 8.32404554e-01
6.66729867e-01 -2.17608377e-01 7.21496284e-01 -1.30271867e-01
-3.19966704e-01 -3.84217203e-02 -1.72776151e+00 7.93732941e-01
6.35838032e-01 -1.55615091e-01 -1.26431489e+00 1.49103060e-01
5.93745708e-01 -2.77895923e-03 -6.08897626e-01 1.98874205e-01
6.86017334e-01 -1.29825079e+00 4.79776680e-01 -8.57955277e-01
3.36928010e-01 -8.32269251e-01 -1.20186642e-01 -8.53910446e-01
2.22766504e-01 -4.96731788e-01 -8.96637589e-02 1.24821579e+00
4.50154901e-01 -1.12971103e+00 4.68366414e-01 1.17046630e+00
2.52233386e-01 -4.25616801e-01 -1.13485670e+00 -4.08041775e-01
1.14232153e-01 -4.07470226e-01 1.11872101e+00 1.06282890e+00
5.37682474e-01 -9.34329331e-02 -1.92677081e-01 5.75798750e-02
8.59714627e-01 2.88510054e-01 5.27558982e-01 -1.86085951e+00
2.44852841e-01 -4.47803199e-01 -3.59753460e-01 1.68715596e-01
4.86488074e-01 -6.67739809e-01 -3.13136578e-01 -2.09243751e+00
-1.87043190e-01 -2.27855280e-01 -1.88543310e-03 4.75457311e-01
4.06460732e-01 -2.90130526e-01 4.97768492e-01 5.90933442e-01
-2.89342761e-01 3.76833320e-01 9.58042979e-01 8.66385084e-03
-9.97893233e-03 -1.59492046e-01 -8.30941260e-01 1.53081238e+00
7.92159021e-01 -6.51073754e-01 -5.40908277e-01 -1.72374062e-02
1.10797012e+00 -2.70800442e-02 6.61220968e-01 -5.62496960e-01
1.55411020e-01 -6.91415191e-01 -3.45233232e-01 7.22818151e-02
-1.44502088e-01 -1.51980805e+00 7.61148095e-01 1.04525638e+00
-8.03476214e-01 -5.19307435e-01 -2.30030209e-01 2.12445185e-01
-2.44817093e-01 -1.00262821e+00 5.04617214e-01 1.08001575e-01
-5.91929555e-01 -5.18076241e-01 -7.04030812e-01 -1.71007991e-01
1.61416590e+00 -5.12075365e-01 -5.39944112e-01 -2.91331440e-01
-9.53974307e-01 2.04387635e-01 7.56005764e-01 -2.04064161e-01
4.31304216e-01 -1.12894106e+00 -7.79333234e-01 -2.77105093e-01
1.30527586e-01 -2.37835661e-01 2.40941390e-01 7.43179679e-01
-8.18924725e-01 3.42729360e-01 -4.84028131e-01 -1.46012485e-01
-1.21225822e+00 5.48041761e-01 5.21458149e-01 5.14295697e-02
-4.74189818e-01 4.09073476e-03 3.89295034e-02 -2.35750169e-01
-3.98872793e-02 -1.45078808e-01 -9.40239131e-01 1.20561145e-01
3.28985751e-01 3.28477979e-01 -3.23875487e-01 -1.06084144e+00
-6.41923368e-01 6.66107178e-01 5.37888050e-01 -4.18938756e-01
1.00648463e+00 -3.37049812e-01 -6.48005724e-01 5.12323618e-01
7.33148530e-02 2.67953038e-01 -6.24310911e-01 1.49579972e-01
1.27264857e-01 -1.17426440e-01 -1.19409181e-01 -1.14580917e+00
-6.74955249e-01 5.50177455e-01 1.48407072e-01 8.15147221e-01
7.81493902e-01 -3.40782292e-02 -8.07978809e-02 5.29302299e-01
5.08497357e-01 -1.08428323e+00 -5.52750885e-01 -9.46632549e-02
1.21246493e+00 -1.03966379e+00 1.99156165e-01 -6.68067098e-01
-1.09547865e+00 1.64360130e+00 2.55711794e-01 2.82165974e-01
4.50614601e-01 2.19727308e-01 4.62725192e-01 -4.25881267e-01
-6.77745819e-01 -1.14975227e-02 -8.32774565e-02 6.27402306e-01
8.39351341e-02 3.06669950e-01 -1.24556053e+00 9.29985225e-01
-1.40803099e-01 4.56896007e-01 9.51814651e-01 8.34709704e-01
-4.79381353e-01 -1.10086250e+00 -6.35832906e-01 -3.19714606e-01
-4.40131903e-01 2.80703068e-01 -8.68320286e-01 1.01101816e+00
3.98736626e-01 1.47378564e+00 -3.31510127e-01 1.99419916e-01
2.59559572e-01 -1.00296289e-01 3.29354368e-02 -1.64264694e-01
-7.48634875e-01 -3.46439630e-01 6.04268730e-01 -3.30542326e-01
-1.28302932e+00 -4.43051964e-01 -1.97599626e+00 -4.16329890e-01
-2.35855296e-01 9.50984359e-01 8.12051356e-01 1.39317966e+00
-3.81782949e-02 3.79647255e-01 2.39740580e-01 2.04948917e-01
-4.83482987e-01 -2.35049829e-01 -5.40165246e-01 3.04608852e-01
1.15345581e-03 -5.69288671e-01 -8.44702661e-01 5.00092357e-02] | [8.849723815917969, 6.631740093231201] |
d723d2da-dd57-45a1-92df-4f057460330c | unsupervised-homography-estimation-with | 2205.03821 | null | https://arxiv.org/abs/2205.03821v1 | https://arxiv.org/pdf/2205.03821v1.pdf | Unsupervised Homography Estimation with Coplanarity-Aware GAN | Estimating homography from an image pair is a fundamental problem in image alignment. Unsupervised learning methods have received increasing attention in this field due to their promising performance and label-free training. However, existing methods do not explicitly consider the problem of plane-induced parallax, which will make the predicted homography compromised on multiple planes. In this work, we propose a novel method HomoGAN to guide unsupervised homography estimation to focus on the dominant plane. First, a multi-scale transformer network is designed to predict homography from the feature pyramids of input images in a coarse-to-fine fashion. Moreover, we propose an unsupervised GAN to impose coplanarity constraint on the predicted homography, which is realized by using a generator to predict a mask of aligned regions, and then a discriminator to check if two masked feature maps are induced by a single homography. To validate the effectiveness of HomoGAN and its components, we conduct extensive experiments on a large-scale dataset, and the results show that our matching error is 22% lower than the previous SOTA method. Code is available at https://github.com/megvii-research/HomoGAN. | ['Shuaicheng Liu', 'Qijun Zhao', 'Chunyu Lin', 'Nianjin Ye', 'Yuhang Lu', 'Mingbo Hong'] | 2022-05-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Hong_Unsupervised_Homography_Estimation_With_Coplanarity-Aware_GAN_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Hong_Unsupervised_Homography_Estimation_With_Coplanarity-Aware_GAN_CVPR_2022_paper.pdf | cvpr-2022-1 | ['homography-estimation'] | ['computer-vision'] | [ 5.75996041e-01 1.34713784e-01 -4.71808985e-02 -2.97511846e-01
-6.91880643e-01 -4.18506056e-01 5.00155449e-01 -5.66342294e-01
1.94673032e-01 4.16540474e-01 2.14671835e-01 5.16159683e-02
2.02032402e-01 -1.01409471e+00 -8.85074675e-01 -9.81839120e-01
5.16900301e-01 3.09269905e-01 1.94030300e-01 -3.06721032e-03
2.83670694e-01 2.77813524e-01 -1.16281438e+00 1.41773671e-01
9.65079546e-01 7.60156393e-01 2.25601450e-01 2.24500164e-01
3.67517024e-01 5.18608809e-01 -3.37856025e-01 -4.99937147e-01
7.54314780e-01 -9.31566477e-01 -4.06186879e-01 4.63914931e-01
7.45352566e-01 -4.42181915e-01 -4.75721180e-01 1.27689385e+00
3.24966162e-01 -2.56104439e-01 5.78003526e-01 -1.26912355e+00
-4.19893324e-01 4.11931157e-01 -7.39105642e-01 -3.26980770e-01
1.82101250e-01 2.10983455e-01 1.05446005e+00 -5.80696762e-01
5.99598348e-01 9.00552392e-01 2.90880650e-01 1.73268929e-01
-1.14845395e+00 -9.78696406e-01 -3.72174710e-01 1.24097504e-01
-1.45024192e+00 -4.15315092e-01 1.20996249e+00 -4.60207283e-01
4.13346678e-01 1.12647623e-01 5.61992884e-01 7.82459378e-01
2.35259622e-01 4.82130170e-01 1.32664275e+00 -5.95898688e-01
-1.30286172e-01 -9.38924327e-02 -5.98243535e-01 8.97868216e-01
6.94271401e-02 5.15597999e-01 -4.27309662e-01 1.99174747e-01
1.24057901e+00 1.23676471e-01 -4.74500656e-01 -6.63046539e-01
-1.38387311e+00 7.60715485e-01 6.94465101e-01 2.36728042e-01
-1.83989912e-01 -1.92408487e-02 -2.22601190e-01 8.99747759e-02
2.99574316e-01 4.41225678e-01 1.32061601e-01 2.55888551e-01
-8.53593767e-01 -1.14873253e-01 4.69724327e-01 1.13470006e+00
1.23172545e+00 -5.18968366e-02 1.27132922e-01 7.42404819e-01
1.54947117e-01 6.53606951e-01 3.51686418e-01 -9.88755941e-01
5.21899402e-01 5.73569894e-01 -2.89884508e-01 -1.49853408e+00
-4.09919880e-02 -4.46343690e-01 -1.17302883e+00 1.32161170e-01
2.14492291e-01 7.02917576e-02 -9.34918284e-01 1.42138982e+00
4.06352431e-01 2.98549861e-01 -1.55828193e-01 1.06384051e+00
4.77463603e-01 5.12154222e-01 -7.54477620e-01 -2.98697855e-02
1.19067109e+00 -1.17246079e+00 -4.56677616e-01 -2.75012404e-01
5.54215789e-01 -1.17627716e+00 7.52236187e-01 3.06187540e-01
-8.10768366e-01 -5.51792860e-01 -1.17855513e+00 -5.59793487e-02
8.29753429e-02 3.53145480e-01 4.07407939e-01 5.15048206e-01
-7.20816791e-01 3.92580777e-01 -6.04603112e-01 -1.02547437e-01
2.65247762e-01 2.95991778e-01 -4.27323192e-01 -2.35812157e-01
-8.94674420e-01 5.32752931e-01 4.76382822e-01 3.02633680e-02
-5.49741745e-01 -3.38818461e-01 -8.52395177e-01 8.42411891e-02
3.64364803e-01 -6.59823716e-01 7.47226179e-01 -9.44850266e-01
-1.58452451e+00 9.86624300e-01 -2.81845089e-02 -2.20964774e-01
5.75704157e-01 9.87672061e-02 -1.64279297e-01 2.66193241e-01
1.36798725e-01 6.74725950e-01 9.52006161e-01 -1.38185310e+00
-6.27226889e-01 -1.86414525e-01 -1.19166998e-02 2.63197601e-01
-1.36584476e-01 -2.68735439e-01 -5.72279096e-01 -7.06268966e-01
6.06413186e-01 -1.02098739e+00 -6.42551333e-02 -2.15959594e-01
-7.09386289e-01 3.42272580e-01 5.86799979e-01 -8.28367829e-01
9.55736160e-01 -2.12598276e+00 7.25802258e-02 5.15787244e-01
2.73660094e-01 -2.80159041e-02 -9.67464410e-03 3.82551640e-01
-1.03218943e-01 -1.18866555e-01 -4.06228185e-01 -1.64241135e-01
-2.16276959e-01 3.49923186e-02 -4.21268344e-01 6.83370173e-01
-7.72318169e-02 7.51157224e-01 -6.13335431e-01 -5.44552386e-01
4.19371992e-01 2.06098497e-01 -5.70134342e-01 3.75194222e-01
4.76061665e-02 7.59802639e-01 -2.99666256e-01 5.25125563e-01
9.84035075e-01 -3.58508617e-01 3.18151206e-01 -6.23153746e-01
-2.46583402e-01 1.76749781e-01 -9.43271577e-01 1.56749237e+00
-2.60410964e-01 4.64571416e-01 -3.10566127e-01 -1.01452422e+00
1.17746091e+00 1.28495917e-01 5.89354098e-01 -7.92084515e-01
3.24060053e-01 3.58363718e-01 1.77510202e-01 -1.15909465e-01
1.49528354e-01 5.31419041e-03 8.26066285e-02 3.07774544e-01
-5.75008355e-02 -3.78284842e-01 -3.23381349e-02 -8.64727306e-04
7.72396088e-01 7.76456147e-02 3.42664987e-01 -9.16237086e-02
6.43214762e-01 -1.16873555e-01 8.28206003e-01 4.79845941e-01
2.06430450e-01 1.15073800e+00 3.63994420e-01 -3.60517383e-01
-1.39846396e+00 -7.30205655e-01 -1.39149483e-02 1.95517227e-01
6.09940350e-01 -4.11637634e-01 -8.01066518e-01 -6.08944476e-01
-2.07593024e-01 1.93949699e-01 -2.71933258e-01 -1.18456461e-01
-7.37150073e-01 -3.89477104e-01 2.25079760e-01 2.87528545e-01
1.07874656e+00 -8.27597201e-01 -3.53698820e-01 -1.35284811e-01
-3.73532742e-01 -1.38822281e+00 -6.67146385e-01 -1.86365813e-01
-6.33126676e-01 -1.22773588e+00 -7.26514280e-01 -1.04640937e+00
1.10543692e+00 4.84950513e-01 7.27351725e-01 2.23280236e-01
-3.53066288e-02 -5.97422980e-02 -3.48929137e-01 1.30948231e-01
-3.06955487e-01 1.88193887e-01 -2.19802648e-01 2.30747253e-01
1.29706711e-01 -7.62029886e-01 -7.05033600e-01 9.77182031e-01
-7.58528948e-01 7.37197459e-01 8.73207271e-01 9.96147156e-01
7.51228392e-01 1.77023888e-01 -1.01227283e-01 -9.16626573e-01
-4.86201160e-02 -3.88413779e-02 -1.04185867e+00 2.60759652e-01
-5.20055294e-01 -5.58278039e-02 6.83372140e-01 -2.21423149e-01
-8.65432560e-01 4.12147343e-01 -2.25445703e-02 -6.34642482e-01
-1.48229316e-01 2.48477831e-01 -6.27767622e-01 -5.03082156e-01
3.19050938e-01 5.62049627e-01 2.85840370e-02 -2.61187732e-01
1.68645009e-01 5.18395245e-01 6.59980774e-01 -3.78916681e-01
1.44602513e+00 6.77264750e-01 6.91815689e-02 -6.62761927e-01
-6.46945536e-01 -3.97699356e-01 -7.88652539e-01 -6.59317747e-02
8.64659548e-01 -9.45770621e-01 -3.47704351e-01 7.46464133e-01
-1.09064662e+00 -3.96925479e-01 1.21452451e-01 5.56587577e-01
-6.12035215e-01 5.70528328e-01 -3.30451131e-01 -1.53156459e-01
-2.58594453e-01 -1.22789514e+00 9.69619513e-01 3.33052754e-01
2.01835915e-01 -7.14124143e-01 5.73048182e-03 5.73358297e-01
2.13866740e-01 2.02755496e-01 6.99592471e-01 -2.73067623e-01
-1.21932173e+00 -2.15684418e-02 -3.08018804e-01 3.75183284e-01
3.58024091e-01 -8.92973095e-02 -7.53095627e-01 -2.73068756e-01
1.82549208e-01 -1.73420668e-01 5.77314615e-01 1.75211743e-01
1.04942966e+00 -3.27268034e-01 -2.99240947e-01 1.19081569e+00
1.34603250e+00 3.43636692e-01 9.33757246e-01 4.53188717e-01
1.10584569e+00 4.88401800e-01 6.29335582e-01 1.87324613e-01
2.19507724e-01 9.61869180e-01 4.35127139e-01 -2.99754471e-01
-2.42431104e-01 -6.67711198e-01 1.30946219e-01 9.83612001e-01
6.99394047e-02 -2.53117293e-01 -7.14215398e-01 3.03788751e-01
-1.66558480e+00 -8.07253242e-01 -5.59358671e-02 2.32511258e+00
6.38573468e-01 4.83012684e-02 -1.39717430e-01 4.24135439e-02
9.52797353e-01 3.61176908e-01 -3.03406149e-01 2.06710532e-01
-2.92006940e-01 -2.87659001e-02 6.48167908e-01 5.87744415e-01
-1.04964459e+00 1.19108915e+00 5.24630070e+00 8.84648025e-01
-1.37007225e+00 -1.26386687e-01 5.46945930e-01 3.90809476e-01
-4.53535557e-01 5.29815316e-01 -6.69679582e-01 5.94824374e-01
-2.97913328e-02 -6.58481792e-02 4.80272770e-01 6.37793362e-01
-1.53256133e-01 1.37473447e-02 -7.88584173e-01 1.13325727e+00
3.60183656e-01 -1.26201260e+00 4.47592698e-02 4.73610580e-01
1.08474076e+00 -2.21538410e-01 1.88432857e-02 -2.64676362e-01
1.07249007e-01 -7.86111772e-01 3.72947633e-01 2.96389014e-01
9.10879135e-01 -6.05936587e-01 6.43595517e-01 3.62694740e-01
-1.11061025e+00 3.48406523e-01 -5.28196573e-01 1.66718870e-01
2.96659589e-01 7.84525871e-01 -8.11144531e-01 8.05920899e-01
3.86239409e-01 7.48516262e-01 -4.37983632e-01 1.11264646e+00
-8.33130419e-01 5.12412965e-01 -3.56538147e-01 4.18001443e-01
-8.88243988e-02 -6.57970965e-01 3.74134958e-01 6.31889224e-01
6.15016282e-01 6.05335161e-02 2.18957022e-01 9.12509561e-01
-1.59679949e-01 1.14836775e-01 -7.41437554e-01 9.38709900e-02
4.63062137e-01 1.35271621e+00 -7.32923865e-01 -1.53934523e-01
-3.34933728e-01 1.24355924e+00 9.83757004e-02 2.26170510e-01
-8.03442180e-01 -4.36418742e-01 2.15997413e-01 1.57260314e-01
3.92393231e-01 -2.79263824e-01 -2.67124623e-01 -1.46457672e+00
1.84188843e-01 -1.02309287e+00 8.73448998e-02 -8.05759668e-01
-1.02404165e+00 5.21028399e-01 -1.76149890e-01 -1.60945821e+00
-2.56814778e-01 -4.18830574e-01 -7.90165246e-01 8.46667111e-01
-1.36936164e+00 -1.44527888e+00 -7.48365939e-01 4.50587869e-01
2.12726206e-01 -2.83396155e-01 5.49248278e-01 3.44426423e-01
-3.39511752e-01 6.39583826e-01 -1.12131238e-02 4.09461021e-01
1.02549922e+00 -8.99366736e-01 4.16338116e-01 1.14424551e+00
2.92094707e-01 4.91685182e-01 5.12761354e-01 -7.43359745e-01
-1.23115110e+00 -9.05497193e-01 7.83419967e-01 -1.53537035e-01
3.31213981e-01 -5.41226566e-01 -7.71052539e-01 8.31979036e-01
2.67038226e-01 -8.80814809e-03 3.39084148e-01 -4.18616146e-01
-3.31007272e-01 -2.99366176e-01 -9.14588690e-01 5.00115335e-01
1.15892720e+00 -5.58486223e-01 -1.82816908e-01 2.64596492e-01
4.73648071e-01 -6.97748244e-01 -7.48153448e-01 6.39071047e-01
6.36494875e-01 -1.37053108e+00 7.04116821e-01 2.08222777e-01
9.03351426e-01 -5.79392612e-01 -1.05707996e-01 -1.26280129e+00
-4.08549249e-01 -5.91005623e-01 1.71701148e-01 1.36660922e+00
6.92363977e-02 -9.48641121e-01 1.00051284e+00 6.22177906e-02
-1.27041250e-01 -6.31959498e-01 -6.65582240e-01 -7.50365198e-01
-1.97549507e-01 2.74128467e-01 7.72365153e-01 1.12558031e+00
-3.01151067e-01 3.84506643e-01 -7.84113467e-01 4.82021898e-01
8.08029354e-01 5.16523778e-01 1.25737536e+00 -7.35459268e-01
-5.49204528e-01 -3.47979516e-01 -6.81654632e-01 -1.52147067e+00
8.02547485e-02 -8.08376968e-01 8.67786556e-02 -1.09465837e+00
1.81045875e-01 -4.78003889e-01 1.19672194e-01 3.41141880e-01
7.25698769e-02 5.96117914e-01 4.21017334e-02 5.75600326e-01
-1.42906964e-01 7.97539651e-01 1.43286788e+00 -1.09894417e-01
6.49068579e-02 -1.50091574e-01 -5.36193311e-01 7.12320745e-01
9.33640122e-01 -4.14356053e-01 -3.90878052e-01 -5.60822606e-01
2.41577618e-05 1.84708443e-02 3.12953204e-01 -1.20583129e+00
3.71094823e-01 -1.69731155e-01 3.88590187e-01 -6.20752573e-01
1.74065486e-01 -8.69848788e-01 4.52181309e-01 2.79476553e-01
-3.12578343e-02 8.32633581e-03 -3.61623496e-01 1.49772421e-01
-4.75129902e-01 -9.82362553e-02 7.40113974e-01 5.22603244e-02
-4.69428509e-01 5.95377088e-01 2.75546551e-01 -2.19070300e-01
9.34286892e-01 -3.06042671e-01 -4.84147966e-01 -4.88434732e-01
1.08718192e-02 7.72087425e-02 1.09783351e+00 2.16180548e-01
7.37836182e-01 -1.45624542e+00 -5.49712062e-01 6.86114609e-01
2.25492612e-01 2.01573089e-01 1.92472175e-01 8.06568801e-01
-8.09193552e-01 2.60578930e-01 -3.72864038e-01 -6.98225021e-01
-1.29352999e+00 3.53542954e-01 3.66080850e-01 -1.76510662e-01
-7.79812098e-01 4.82069224e-01 7.12907135e-01 -5.41441083e-01
-1.11464627e-01 7.96874985e-02 8.86985362e-02 -4.61357504e-01
1.86871529e-01 -5.08452170e-02 -1.31079882e-01 -8.24959815e-01
-6.36422560e-02 1.04041004e+00 -1.02337770e-01 -1.08635426e-01
1.01587570e+00 -8.88628960e-02 -1.75888374e-01 -1.08137958e-01
1.33743179e+00 4.49638337e-01 -1.32019365e+00 -2.74244308e-01
-4.89184380e-01 -8.93728435e-01 -1.44886106e-01 -3.88360351e-01
-1.41322529e+00 9.02899683e-01 4.06566113e-01 -1.00019174e-02
1.17820227e+00 -1.45559818e-01 9.53243256e-01 4.08610702e-03
4.33876157e-01 -7.21838534e-01 1.70892715e-01 3.94690692e-01
7.98239172e-01 -1.21458089e+00 1.40515953e-01 -9.03250396e-01
-4.94611621e-01 1.13133359e+00 8.34223151e-01 -3.71342480e-01
3.66973370e-01 -9.18432474e-02 2.09704027e-01 -1.81120500e-01
-1.63571358e-01 -2.23630294e-02 3.05394560e-01 4.53768641e-01
2.36205712e-01 -3.36998478e-02 -2.29436219e-01 5.11561669e-02
-6.72711670e-01 -2.14960814e-01 3.97680312e-01 6.85997844e-01
-1.96071178e-01 -1.41443610e+00 -3.52019519e-01 1.48796216e-01
-1.27513245e-01 -4.65550423e-02 -4.13805932e-01 6.26812637e-01
1.72023550e-01 5.73006451e-01 8.27181339e-02 -8.81919324e-01
8.23739693e-02 -5.49931943e-01 5.97936809e-01 -5.33503950e-01
-1.19625211e-01 3.88495862e-01 -2.13012382e-01 -4.40192610e-01
-4.64297295e-01 -3.22296917e-01 -9.30296600e-01 -1.94390923e-01
-4.16417956e-01 2.97727156e-02 3.32012355e-01 8.71536136e-01
3.51298779e-01 1.14712097e-01 1.19437408e+00 -8.47586095e-01
-3.03217500e-01 -7.06556082e-01 -4.88245040e-01 5.22780359e-01
1.26468807e-01 -6.49410009e-01 -5.64893603e-01 1.33805022e-01] | [8.847541809082031, -2.307605504989624] |
d79c6fb7-5592-4654-b496-5581a89d8ecb | audioslots-a-slot-centric-generative-model | 2305.05591 | null | https://arxiv.org/abs/2305.05591v1 | https://arxiv.org/pdf/2305.05591v1.pdf | AudioSlots: A slot-centric generative model for audio separation | In a range of recent works, object-centric architectures have been shown to be suitable for unsupervised scene decomposition in the vision domain. Inspired by these methods we present AudioSlots, a slot-centric generative model for blind source separation in the audio domain. AudioSlots is built using permutation-equivariant encoder and decoder networks. The encoder network based on the Transformer architecture learns to map a mixed audio spectrogram to an unordered set of independent source embeddings. The spatial broadcast decoder network learns to generate the source spectrograms from the source embeddings. We train the model in an end-to-end manner using a permutation invariant loss function. Our results on Libri2Mix speech separation constitute a proof of concept that this approach shows promise. We discuss the results and limitations of our approach in detail, and further outline potential ways to overcome the limitations and directions for future work. | ['Thomas Kipf', 'John R. Hershey', 'Klaus Greff', 'Scott Wisdom', 'Pradyumna Reddy'] | 2023-05-09 | null | null | null | null | ['speech-separation'] | ['speech'] | [ 4.74078298e-01 1.39083520e-01 1.36100635e-01 -4.28755343e-01
-1.42484832e+00 -6.92031324e-01 8.62441659e-01 -4.45494413e-01
-1.91446021e-01 3.20335507e-01 9.60125983e-01 -2.81061623e-02
-2.40439773e-01 -2.07231611e-01 -7.09986508e-01 -9.53197300e-01
-1.72917977e-01 6.08448207e-01 -5.92641858e-03 1.58107936e-01
3.36346589e-02 1.42686337e-01 -1.61930668e+00 6.19346678e-01
3.47409874e-01 8.36053193e-01 3.19861442e-01 1.13128126e+00
4.44577448e-02 9.72554386e-01 -6.05800867e-01 -7.78834894e-02
3.72884303e-01 -8.12541425e-01 -8.21072698e-01 3.16846281e-01
4.56352204e-01 -2.53872752e-01 -5.20702183e-01 8.93996835e-01
7.76295066e-01 -7.75571121e-03 7.63701200e-01 -1.18325293e+00
-6.60601079e-01 7.29812562e-01 -1.42260805e-01 2.39607230e-01
3.24768782e-01 -1.69660658e-01 1.21035314e+00 -1.13301897e+00
4.07062054e-01 1.31957340e+00 5.54870605e-01 3.73235255e-01
-1.48317885e+00 -4.13559437e-01 -9.56590101e-02 2.10566044e-01
-1.24205053e+00 -1.10246730e+00 8.06131184e-01 -4.22406793e-01
1.07556200e+00 3.09088111e-01 4.03235584e-01 1.17082298e+00
-7.54272342e-02 1.00504148e+00 9.79561687e-01 -6.13026619e-01
2.82163173e-01 1.44740298e-01 -1.73435971e-01 5.38506627e-01
-1.01179965e-01 6.61164075e-02 -1.00948441e+00 -1.24454878e-01
6.96834266e-01 -4.08209264e-01 -2.14608625e-01 -8.51202428e-01
-1.34859407e+00 9.33586597e-01 3.68615091e-01 2.16388017e-01
-1.32301345e-01 4.36170965e-01 1.89698592e-01 3.36919695e-01
4.34341550e-01 4.29249525e-01 -2.33810097e-02 -1.98587887e-02
-1.22086048e+00 1.50309816e-01 7.50460923e-01 8.88156056e-01
4.11175698e-01 4.95289713e-01 -1.36916265e-01 1.00853634e+00
6.97354019e-01 4.02579397e-01 6.52529180e-01 -1.03544843e+00
3.39905947e-01 -1.53687164e-01 -1.06943533e-01 -5.34104824e-01
-1.49573699e-01 -4.23560828e-01 -4.77377772e-01 1.74519688e-01
2.14459866e-01 -1.77313536e-01 -1.23724568e+00 1.62189913e+00
-2.31512431e-02 2.80367762e-01 1.95781991e-01 1.06845105e+00
6.99656129e-01 6.56959534e-01 -2.69936293e-01 1.10452130e-01
1.17482996e+00 -1.04101217e+00 -5.09906292e-01 -3.14186245e-01
9.10596400e-02 -1.02434194e+00 7.75428474e-01 3.43990892e-01
-1.08620059e+00 -4.29726720e-01 -9.91816700e-01 -1.78377852e-01
-8.90508816e-02 5.15253305e-01 5.11089981e-01 7.22703874e-01
-1.35845768e+00 1.25428677e-01 -1.08586073e+00 -5.69039702e-01
4.11238939e-01 1.44293889e-01 -2.40278021e-01 2.14200452e-01
-7.71517694e-01 3.76980573e-01 4.08809423e-01 -8.19300190e-02
-1.65949035e+00 -4.55854177e-01 -1.02662039e+00 1.08676195e-01
-3.94294038e-02 -8.68513882e-01 1.46369708e+00 -9.63082254e-01
-1.66483653e+00 6.98136926e-01 -3.05767179e-01 -8.46515477e-01
1.93987727e-01 -9.87868682e-02 -3.26886833e-01 4.22745377e-01
4.43055719e-01 7.75165558e-01 1.17832243e+00 -1.19746864e+00
-5.85468352e-01 -1.88271597e-01 -1.61787584e-01 4.51330781e-01
-3.95250887e-01 2.74210572e-01 -4.95208323e-01 -9.42316234e-01
2.22967058e-01 -9.69163775e-01 -4.84928675e-02 -2.75269568e-01
-5.81523716e-01 2.09338859e-01 6.39913678e-01 -5.42081416e-01
7.74712920e-01 -2.41600394e+00 4.14124817e-01 -5.33044860e-02
5.74071780e-02 2.95208506e-02 -4.31343555e-01 7.34996259e-01
-3.99652392e-01 -3.00301790e-01 -5.01904368e-01 -7.58281529e-01
6.16348200e-02 -8.01099017e-02 -8.70680213e-01 4.71283644e-01
1.55008927e-01 5.33893585e-01 -8.99972558e-01 -7.96993971e-02
1.60547838e-01 6.21243596e-01 -7.18047917e-01 3.73506933e-01
-1.24790661e-01 2.25526482e-01 -3.70561033e-02 4.21202838e-01
2.93666035e-01 1.87303871e-01 6.55452684e-02 -3.07130013e-02
6.68453053e-02 9.26109970e-01 -1.24039865e+00 2.29812288e+00
-4.17036057e-01 9.89896953e-01 3.72874677e-01 -9.81429994e-01
5.98492503e-01 6.40712738e-01 3.19458812e-01 -2.33068854e-01
-3.18292975e-02 2.18077391e-01 -6.20140843e-02 -3.71408045e-01
4.22908127e-01 -3.28952700e-01 -1.44356415e-01 4.89206165e-01
7.90961325e-01 -3.06167722e-01 -1.88776897e-03 4.89003450e-01
1.06670344e+00 1.65799201e-01 -5.00357337e-02 -8.50378424e-02
2.53299057e-01 -2.88876537e-02 -8.33514030e-04 7.04458594e-01
8.83290097e-02 1.18816078e+00 2.38806143e-01 1.68025754e-02
-1.00595522e+00 -1.59431362e+00 5.34612089e-02 1.04856062e+00
-2.01179937e-01 -6.96780086e-01 -7.28975415e-01 -4.24873143e-01
-2.20838234e-01 7.64564216e-01 -2.38652572e-01 -1.68505326e-01
-2.32971236e-01 -7.32989967e-01 8.73997331e-01 7.11055279e-01
2.88189262e-01 -8.20429087e-01 -4.55089808e-01 2.23790005e-01
-4.22722429e-01 -1.16008615e+00 -3.44483018e-01 5.06243944e-01
-6.06500089e-01 -7.89173424e-01 -1.00712109e+00 -9.45545018e-01
4.86582488e-01 5.98175585e-01 7.06943452e-01 -8.53787839e-01
-3.84807616e-01 7.34350443e-01 -1.61203519e-01 -4.66110319e-01
-3.68647486e-01 -3.65210511e-02 2.61505812e-01 3.52111280e-01
2.45658964e-01 -7.71970451e-01 -4.11897153e-01 6.23734593e-02
-9.42270994e-01 7.61125889e-03 3.32944095e-01 7.28092492e-01
5.12536526e-01 8.71227309e-02 4.48780835e-01 -4.85054731e-01
7.31310725e-01 -3.99004728e-01 -4.55227643e-01 -1.84627339e-01
-1.84099928e-01 2.15760678e-01 5.47931075e-01 -1.39132023e-01
-1.02928805e+00 3.93156290e-01 -8.35669860e-02 -7.25688636e-01
-3.81692350e-01 3.52774888e-01 -2.07839876e-01 1.67470127e-01
8.06353688e-01 5.05971193e-01 3.05288192e-02 -6.84960246e-01
7.43966699e-01 9.45951939e-01 8.61047566e-01 -2.90466815e-01
7.67491102e-01 5.97710788e-01 -4.77442682e-01 -1.09337974e+00
-5.78386724e-01 -7.09326804e-01 -3.70677829e-01 4.51852120e-02
9.42462444e-01 -1.34705865e+00 -1.13391899e-01 4.11363393e-01
-1.15424931e+00 -3.63977909e-01 -5.58972776e-01 6.74252808e-01
-1.00976205e+00 2.05994740e-01 -4.29619223e-01 -9.61218476e-01
4.71858978e-02 -1.09347427e+00 1.38220692e+00 -8.33771154e-02
-3.29980761e-01 -7.62202919e-01 4.30290431e-01 4.50224370e-01
1.73094526e-01 -2.18022570e-01 6.30687952e-01 -7.68045604e-01
-8.02705705e-01 1.22333905e-02 1.81783158e-02 4.79575306e-01
1.77474067e-01 -4.39516157e-01 -1.49722302e+00 -4.30292279e-01
2.74564713e-01 -4.22879189e-01 1.33179712e+00 4.56889689e-01
9.02571499e-01 -2.65324116e-01 -1.49049222e-01 7.55271375e-01
1.29487741e+00 1.41196668e-01 5.71104467e-01 2.00287610e-01
7.55756378e-01 3.94086093e-01 -2.45831497e-02 1.37727112e-01
4.47443068e-01 7.59034753e-01 3.42407674e-01 3.74234729e-02
-8.06574583e-01 -5.07408082e-01 8.29859972e-01 7.28489041e-01
2.89214760e-01 -3.11082155e-01 -7.10730553e-01 1.15439582e+00
-1.63483417e+00 -1.04613197e+00 7.10835084e-02 2.16508913e+00
6.96012914e-01 -1.25987619e-01 4.15231377e-01 3.28157276e-01
4.57852066e-01 4.16409284e-01 -4.47694100e-02 -2.95704156e-01
-1.18293315e-01 2.47616306e-01 3.53096932e-01 7.00746179e-01
-1.30407476e+00 1.06086457e+00 7.12985945e+00 7.33935535e-01
-1.01389372e+00 2.34737113e-01 4.20490392e-02 -4.16306049e-01
-4.02673036e-01 1.59375653e-01 -4.60733086e-01 3.30657721e-01
1.13333595e+00 9.17212665e-02 6.92604780e-01 6.17214978e-01
-8.75322297e-02 1.51258716e-02 -1.26305997e+00 1.07167852e+00
5.15479505e-01 -1.21610951e+00 -8.12320039e-02 -4.00861800e-02
5.37769675e-01 4.31578517e-01 2.98681885e-01 -1.20947793e-01
5.64860702e-01 -9.30768728e-01 1.09638548e+00 2.34249398e-01
8.34786177e-01 -5.32306850e-01 1.54916584e-01 1.75745592e-01
-1.03351617e+00 -4.50904891e-02 -2.89721221e-01 1.02313720e-01
2.85611033e-01 2.69619524e-01 -1.43656003e+00 5.84110439e-01
7.23322451e-01 6.71542645e-01 -4.17916238e-01 1.24600565e+00
-4.17301536e-01 9.68970418e-01 -3.51840496e-01 4.02731955e-01
1.69581667e-01 7.61683136e-02 1.08901966e+00 1.40750706e+00
3.30186993e-01 -3.90043467e-01 -3.94253759e-03 7.98876524e-01
7.08013698e-02 -3.76306534e-01 -8.69288087e-01 -3.38717699e-01
1.80233106e-01 1.03395593e+00 -6.70672715e-01 -8.33729282e-02
-1.42764166e-01 1.21562922e+00 6.93531483e-02 6.09556556e-01
-5.66928566e-01 -5.89514256e-01 7.02015460e-01 2.82754879e-02
6.66576803e-01 -3.29653770e-01 -2.34758168e-01 -1.42384732e+00
-1.07495099e-01 -8.79947484e-01 2.19189435e-01 -9.34586346e-01
-9.97661412e-01 6.84453964e-01 1.35278612e-01 -1.33318150e+00
-6.76425040e-01 -4.22509402e-01 -4.28033680e-01 8.17449987e-01
-1.29095650e+00 -1.12687099e+00 1.36085793e-01 6.53435886e-01
8.52973104e-01 -6.74909413e-01 1.08758891e+00 1.27826840e-01
-1.56750530e-01 4.92232054e-01 2.80530155e-01 2.45277599e-01
7.44313240e-01 -1.30127072e+00 6.27084911e-01 1.22075498e+00
1.19719338e+00 5.05665302e-01 8.80925000e-01 -2.83333898e-01
-1.31451643e+00 -1.10660911e+00 9.03620183e-01 -6.54783130e-01
6.64720297e-01 -8.41342926e-01 -4.62735146e-01 9.37946677e-01
5.45586467e-01 -2.95725584e-01 8.31180334e-01 -6.28699549e-03
-7.27478027e-01 -2.07206771e-01 -7.27335572e-01 5.67167163e-01
9.48854089e-01 -9.80524778e-01 -8.74426484e-01 3.57164174e-01
5.50829768e-01 -4.46792305e-01 -1.86831325e-01 -1.44751668e-01
4.49387461e-01 -1.04345298e+00 1.24224055e+00 -3.08637083e-01
3.56905431e-01 -5.49830198e-01 -3.66582096e-01 -1.56252050e+00
-4.13195550e-01 -8.58516574e-01 1.86854023e-02 1.24899697e+00
3.27632785e-01 -4.48301703e-01 5.71801364e-01 -6.13685586e-02
-4.28706408e-01 -1.19661711e-01 -1.05850935e+00 -8.72118115e-01
-1.91646278e-01 -7.95044780e-01 2.75841385e-01 4.21373427e-01
-1.02594942e-01 7.08891630e-01 -4.84591275e-01 3.35660249e-01
6.61336899e-01 1.12290561e-01 7.46944666e-01 -8.94620478e-01
-5.82856178e-01 -2.98164994e-01 -5.22735119e-01 -1.10797882e+00
1.56206354e-01 -1.13100910e+00 4.32743609e-01 -1.66583359e+00
-3.18652429e-02 -9.61992294e-02 -1.92837194e-01 4.32592392e-01
3.21802646e-01 4.46618110e-01 4.21006352e-01 1.95405334e-01
-5.15540123e-01 5.75136423e-01 4.40802932e-01 -4.66444463e-01
-1.31198898e-01 8.44636001e-03 -8.52423966e-01 6.45050824e-01
6.23114586e-01 -5.82626283e-01 -8.45808506e-01 -9.03813779e-01
-1.49395630e-01 -1.41205937e-01 6.43406034e-01 -1.19729853e+00
4.14886475e-01 2.76883602e-01 2.57066041e-01 -2.94748485e-01
7.74375260e-01 -8.37638199e-01 -8.63063149e-03 8.35309252e-02
-5.89880288e-01 -3.91665697e-01 1.90407142e-01 7.98066318e-01
-4.57314342e-01 -5.85331470e-02 5.60605943e-01 1.60580486e-01
-5.44120133e-01 -4.30982448e-02 -7.58908212e-01 4.02416475e-02
5.84890246e-01 -1.79074928e-01 1.51028752e-01 -6.94825470e-01
-8.09424818e-01 -2.17438549e-01 1.45053864e-01 5.51921546e-01
7.65373290e-01 -1.42809463e+00 -8.62287402e-01 5.76268792e-01
1.21270612e-01 -2.96032101e-01 -7.00778589e-02 5.93322515e-01
-2.28632152e-01 7.39143729e-01 6.75803982e-03 -6.82965279e-01
-1.31399441e+00 4.01036918e-01 3.54374588e-01 2.65379608e-01
-4.97668892e-01 1.23706150e+00 4.79770601e-01 -2.05143124e-01
4.13259149e-01 -2.58950293e-01 -8.04490503e-03 -1.96134187e-02
5.73813558e-01 2.62546122e-01 1.61104679e-01 -8.27187836e-01
-5.72417676e-01 1.97246909e-01 5.95331453e-02 -9.58377242e-01
1.44491112e+00 -1.23442918e-01 1.15143964e-02 4.00363594e-01
1.39838827e+00 2.21295923e-01 -1.23506117e+00 -2.06587002e-01
-2.16967314e-01 -3.13630939e-01 1.39180303e-01 -8.59715462e-01
-7.84761727e-01 1.12169385e+00 5.16665041e-01 3.07228893e-01
1.33715212e+00 3.07798773e-01 4.75848079e-01 5.28749451e-02
2.96604544e-01 -7.84497678e-01 1.13639534e-01 5.07475078e-01
1.02631164e+00 -7.95109153e-01 -5.04684091e-01 -2.76367694e-01
-6.37456715e-01 1.07172143e+00 1.84375662e-02 -3.54977012e-01
4.92005199e-01 4.91437346e-01 9.17215645e-02 -1.17079839e-01
-6.47262514e-01 -4.32806343e-01 4.37147439e-01 9.90722477e-01
3.58774453e-01 -4.40676622e-02 3.52844626e-01 5.09778261e-01
-5.30983448e-01 -1.63068533e-01 4.09305602e-01 6.62473261e-01
-3.04793149e-01 -1.24014151e+00 -6.86626732e-01 -1.00948755e-03
-3.07655692e-01 -3.60323638e-01 -7.28522956e-01 2.43132144e-01
-5.54708838e-02 1.27331066e+00 9.95497480e-02 -4.09938633e-01
1.19917437e-01 3.45383078e-01 5.61848819e-01 -9.19435918e-01
-2.80443043e-01 7.46444464e-01 1.10484354e-01 -3.89292806e-01
-4.57154602e-01 -6.48233414e-01 -1.09321845e+00 3.66630077e-01
4.88942154e-02 1.96218014e-01 5.83787143e-01 7.34180272e-01
4.82635319e-01 6.78717434e-01 5.22979796e-01 -9.40285861e-01
-4.24767852e-01 -8.14424574e-01 -6.14442050e-01 8.32995586e-03
8.21217060e-01 -3.21562350e-01 -3.51213187e-01 5.47713399e-01] | [15.23212718963623, 5.285333633422852] |
238268cf-cbfd-4752-87d0-e946aeb31633 | visual-relationship-detection-with-deep | null | null | https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=Visual+relationship+detection+with+deep+structural+ranking&btnG= | https://pdfs.semanticscholar.org/0709/f6328229a79c44be715db90df028786a3129.pdf | Visual relationship detection with deep structural ranking | Visual relationship detection aims to describe the interactions between pairs of objects. Different from individual object learning tasks, the number of possible relationships is much larger, which makes it hard to explore only based on the visual appearance of objects. In addition, due to the limited human effort, the annotations for visual relationships are usually incomplete which increases the difficulty of model training and evaluation. In this paper, we propose a novel framework, called Deep Structural Ranking, for visual relationship detection. To complement the representation ability of visual appearance, we integrate multiple cues for predicting the relationships contained in an input image. Moreover, we design a new ranking objective function by enforcing the annotated relationships to have higher relevance scores. Unlike previous works, our proposed method can both facilitate the co-occurrence of relationships and mitigate the incompleteness problem. Experimental results show that our proposed method outperforms the state-of-the-art on the two widely used datasets. We also demonstrate its superiority in detecting zero-shot relationships. | ['Xilin Chen', 'Hong Chang', 'Yuhong Guo', 'Kongming Liang'] | 2018-04-27 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 1.76266924e-01 -1.45674154e-01 -2.63852924e-01 -3.54188949e-01
-3.85958068e-02 -2.49697626e-01 4.87395614e-01 4.74984527e-01
-1.95809066e-01 4.70010012e-01 9.93919596e-02 7.72265643e-02
-4.14602548e-01 -7.31046855e-01 -5.36055446e-01 -4.56210315e-01
1.31898122e-02 1.88691318e-01 5.75766504e-01 -2.25469526e-02
1.46340579e-01 3.47065061e-01 -1.81805182e+00 2.71518350e-01
7.70714998e-01 1.09717441e+00 5.47209263e-01 -3.03230970e-03
1.94904264e-02 9.49866533e-01 -2.81257987e-01 -3.26331884e-01
1.60780605e-02 -3.93987447e-01 -4.81259644e-01 2.62341172e-01
4.80722904e-01 -4.52394456e-01 -5.32589138e-01 1.18739128e+00
2.18584254e-01 3.97639602e-01 5.06478071e-01 -1.38357389e+00
-8.83945763e-01 4.55941767e-01 -8.78454685e-01 2.35440999e-01
4.04767215e-01 -1.63298741e-01 1.47412395e+00 -1.01742804e+00
6.23141289e-01 1.44146419e+00 2.62605876e-01 2.57514387e-01
-1.04214466e+00 -7.79128969e-01 4.94130313e-01 5.36599159e-01
-1.48420060e+00 -3.39694053e-01 9.47092474e-01 -4.68899429e-01
6.45972311e-01 2.70615935e-01 6.39209032e-01 5.76302528e-01
-4.25144285e-01 7.87663519e-01 7.51606762e-01 -3.69100660e-01
-7.13358223e-02 1.85351640e-01 2.13997915e-01 6.60105348e-01
3.47371072e-01 -1.95902720e-01 -4.94259298e-01 -1.24423824e-01
8.20915639e-01 2.82061547e-01 -3.87790531e-01 -7.40042865e-01
-1.22107017e+00 4.87756431e-01 7.15562940e-01 1.96234599e-01
-1.50402978e-01 5.78806698e-02 2.33289525e-01 -1.08126700e-01
2.39275992e-01 2.42414162e-01 3.68729569e-02 3.00357759e-01
-4.67509598e-01 -7.90286344e-03 2.71257102e-01 9.97977436e-01
8.04153621e-01 -4.47906077e-01 -3.87029290e-01 1.09810698e+00
5.65445364e-01 -1.16503537e-01 1.16584331e-01 -7.04448640e-01
5.75635970e-01 9.75868523e-01 8.00732970e-02 -1.57478178e+00
-1.92139879e-01 -3.51210356e-01 -8.15589845e-01 -7.40818381e-02
1.24988422e-01 4.94587243e-01 -5.32519758e-01 1.62519586e+00
3.87274057e-01 2.94847250e-01 -1.18661016e-01 1.15814495e+00
1.21260357e+00 4.99535978e-01 4.35447097e-02 -3.59639108e-01
1.33515251e+00 -1.22472882e+00 -9.22052503e-01 -2.50803113e-01
2.50972569e-01 -8.15429211e-01 9.95781898e-01 8.97413194e-02
-9.34753358e-01 -7.13925421e-01 -1.02626240e+00 -1.27852872e-01
-1.99836776e-01 2.20829695e-01 1.01177430e+00 -4.51632328e-02
-6.74813747e-01 2.98610866e-01 -5.53287685e-01 -3.50223005e-01
5.35187960e-01 2.97341764e-01 -4.13070440e-01 -1.47025540e-01
-1.09584904e+00 7.45532691e-01 7.44252563e-01 3.42455328e-01
-7.94469774e-01 -3.43775094e-01 -8.53752732e-01 1.83797330e-01
6.61239028e-01 -4.00026053e-01 8.21265340e-01 -6.98357940e-01
-7.63508797e-01 7.38513529e-01 -2.29853496e-01 -6.63838610e-02
4.47692335e-01 -1.65285215e-01 -2.73774117e-01 1.61293119e-01
-1.42930234e-02 6.89219415e-01 4.53357548e-01 -1.77881229e+00
-6.72963679e-01 -1.32752985e-01 5.05934894e-01 4.24824059e-01
-8.29603255e-01 2.07506448e-01 -1.10264683e+00 -5.81402004e-01
2.99860775e-01 -6.67068422e-01 -1.65228516e-01 4.59467918e-01
-3.53409320e-01 -4.40403908e-01 1.07027245e+00 -2.74537474e-01
1.41201878e+00 -2.15633225e+00 1.31255835e-01 1.73711374e-01
5.16023576e-01 3.02501053e-01 -1.80033475e-01 3.73188406e-01
-3.45790386e-02 1.57170799e-02 -1.29240587e-01 -3.31987977e-01
-1.74643621e-01 3.62180114e-01 -1.45302311e-01 2.20438510e-01
1.77747875e-01 7.59451270e-01 -1.22336078e+00 -9.01106834e-01
2.97560394e-01 5.37349224e-01 -2.60142416e-01 4.21314716e-01
-8.41116533e-02 3.54678392e-01 -5.12740314e-01 5.78864634e-01
7.62572646e-01 -5.39806902e-01 2.11411372e-01 -4.25785661e-01
8.73808339e-02 3.51491831e-02 -1.12158298e+00 1.53939259e+00
-2.66599357e-01 5.44119835e-01 -3.87433887e-01 -9.45663750e-01
1.19105029e+00 9.10506546e-02 4.19232726e-01 -6.86456501e-01
-4.58875783e-02 -1.30053535e-01 1.85794592e-01 -7.06447065e-01
3.35598022e-01 2.67950416e-01 3.06580126e-01 1.58498511e-01
-2.54027307e-01 2.89604008e-01 3.97792906e-01 3.21757168e-01
7.45682716e-01 2.17605144e-01 5.60812116e-01 1.50679812e-01
6.67179942e-01 -4.08187062e-01 7.76977003e-01 5.43921947e-01
-1.88848421e-01 5.28930902e-01 5.70298970e-01 -3.09230298e-01
-6.94624901e-01 -9.21686888e-01 -8.96345526e-02 9.58748639e-01
9.63162303e-01 -7.11130738e-01 -1.02568701e-01 -8.34338546e-01
1.36444364e-02 3.32546383e-01 -6.63060725e-01 -2.01416895e-01
-5.38682759e-01 -5.69344521e-01 -1.06890082e-01 6.52186751e-01
4.21598971e-01 -1.08453071e+00 -4.88692731e-01 -6.50173426e-02
-3.47308904e-01 -1.44381618e+00 -4.26964581e-01 -2.58481741e-01
-7.66564965e-01 -1.18699360e+00 -3.77383381e-01 -1.12167299e+00
1.00243413e+00 7.86514461e-01 1.03246903e+00 5.02554893e-01
-2.62382060e-01 1.50559559e-01 -5.03770471e-01 -1.45416781e-01
-1.18539613e-02 -4.13360298e-01 -2.28882536e-01 2.06770897e-01
2.98878759e-01 -4.97159719e-01 -8.16559017e-01 4.66809183e-01
-7.77170300e-01 2.99696743e-01 6.93893731e-01 8.62425029e-01
7.54238427e-01 4.39294398e-01 2.67146975e-01 -7.03286469e-01
3.62878561e-01 -3.79119337e-01 -4.12061960e-01 6.03647530e-01
-5.38960457e-01 -2.02388987e-02 4.68993038e-01 -5.67740679e-01
-1.10836339e+00 1.66329250e-01 5.56696057e-01 -6.50515497e-01
6.88612685e-02 5.41912556e-01 -2.40307942e-01 -6.54768720e-02
1.31345108e-01 -5.99424802e-02 -3.12740564e-01 -4.48318452e-01
3.12307596e-01 3.81755441e-01 4.49904203e-01 -2.90269583e-01
7.84154475e-01 4.93805856e-01 2.19142020e-01 -5.03418207e-01
-1.02922142e+00 -7.29543269e-01 -7.90041685e-01 -4.51579064e-01
7.58962810e-01 -8.85367870e-01 -6.16334975e-01 -8.75977278e-02
-1.33504748e+00 3.14303458e-01 1.61779955e-01 4.41497684e-01
-1.73056081e-01 6.99225843e-01 -3.08272272e-01 -1.02047002e+00
-1.31911114e-01 -9.34389472e-01 9.97412026e-01 3.72542620e-01
-5.80505356e-02 -7.39413798e-01 -1.87119037e-01 4.68344748e-01
-8.51063505e-02 2.86925375e-01 9.58409786e-01 -3.88869464e-01
-7.61405170e-01 -8.65107477e-02 -8.19461763e-01 6.77423701e-02
3.04291308e-01 1.98298663e-01 -7.39667237e-01 -1.47814766e-01
-3.46730530e-01 -3.59389722e-01 1.05381989e+00 -4.30640206e-02
1.43458867e+00 -1.11326218e-01 -6.41346872e-01 1.68655083e-01
1.27368307e+00 3.40910465e-01 5.55049419e-01 3.15376252e-01
1.08703423e+00 1.02770042e+00 1.19726920e+00 5.90248644e-01
4.14860785e-01 9.34889197e-01 8.43705475e-01 -3.48373711e-01
-2.03486308e-01 -2.74850041e-01 -8.71584192e-02 7.36056030e-01
-2.58671463e-01 -2.69580305e-01 -7.85901725e-01 5.39795160e-01
-2.42848301e+00 -9.07998860e-01 -3.45121175e-01 2.33104634e+00
7.23188221e-01 1.62614986e-01 -5.45272082e-02 -4.16606590e-02
9.95169342e-01 1.80728331e-01 -3.31919074e-01 1.77119464e-01
-1.99566353e-02 -2.96449631e-01 6.64101774e-03 2.40071312e-01
-1.14964640e+00 7.93078125e-01 5.43435717e+00 7.10052013e-01
-6.78065062e-01 -8.32417160e-02 5.38428605e-01 7.73095945e-03
-2.70993590e-01 1.20386913e-01 -7.18052864e-01 3.38640094e-01
-7.17239380e-02 -5.62154129e-02 1.72842979e-01 7.20387816e-01
1.42194599e-01 -1.12054184e-01 -1.18461382e+00 1.10329020e+00
2.65627384e-01 -8.47161770e-01 2.44763434e-01 -4.58469577e-02
6.75916076e-01 -6.07877016e-01 4.42915224e-02 8.68928209e-02
2.26311442e-02 -8.83590877e-01 6.25756145e-01 5.65003753e-01
4.75507379e-01 -7.94579566e-01 6.88110352e-01 2.01879323e-01
-1.70056415e+00 -1.95558697e-01 -5.92205107e-01 -1.34423152e-01
9.30794775e-02 4.52689886e-01 -7.11037934e-01 6.73059642e-01
7.58469582e-01 1.07765925e+00 -8.20233881e-01 1.28378308e+00
-3.73362780e-01 1.23868853e-01 -8.81552994e-02 -7.36817271e-02
4.01642621e-02 -6.58029914e-02 3.97834003e-01 9.03601289e-01
9.01927203e-02 1.53909594e-01 4.92194384e-01 7.83794343e-01
-7.23393038e-02 2.57861614e-01 -6.17262185e-01 2.37073675e-02
6.00217462e-01 1.46689248e+00 -8.38926792e-01 -1.88638538e-01
-6.22401655e-01 8.28657329e-01 6.63402259e-01 2.62049317e-01
-9.18975055e-01 -2.06539005e-01 3.90680790e-01 -1.33648952e-02
3.35254908e-01 -1.24478653e-01 4.80483174e-02 -1.07195318e+00
3.10277700e-01 -5.44486582e-01 5.46576977e-01 -8.46112072e-01
-1.37801409e+00 5.42206645e-01 2.52061933e-01 -1.63260567e+00
1.28806025e-01 -3.42355639e-01 -5.32805562e-01 4.59319443e-01
-1.68183374e+00 -1.33872521e+00 -7.56254494e-01 2.77164966e-01
4.58941489e-01 1.56670675e-01 5.71950436e-01 3.88018161e-01
-6.97614729e-01 4.74510550e-01 -3.25363606e-01 9.88853350e-02
7.26040721e-01 -9.71841991e-01 -2.11309902e-02 8.49787831e-01
2.47604728e-01 7.77183712e-01 5.60668409e-01 -6.98768020e-01
-9.84360695e-01 -9.51288104e-01 7.85411000e-01 -3.05835735e-02
7.78209627e-01 -3.65988374e-01 -1.15427279e+00 4.24552441e-01
8.76159891e-02 3.00537109e-01 6.73506856e-01 9.55116898e-02
-6.90688610e-01 -2.98839152e-01 -6.65823042e-01 7.50980616e-01
1.48739004e+00 -3.93227637e-01 -5.34808278e-01 3.05917472e-01
7.52747536e-01 -6.25708327e-02 -7.04591393e-01 8.06498408e-01
7.10460424e-01 -9.47869778e-01 1.23235166e+00 -4.70644593e-01
5.67024469e-01 -6.57630324e-01 -1.25187859e-01 -8.23496282e-01
-5.45426726e-01 -1.23963162e-01 -5.11364996e-01 1.64331770e+00
1.81722701e-01 -3.24037015e-01 4.99702215e-01 4.57405597e-01
1.89138085e-01 -8.64731014e-01 -5.62024474e-01 -7.42804706e-01
-7.70586073e-01 6.83816224e-02 3.67997885e-01 1.11472881e+00
1.32737793e-02 5.94871700e-01 -6.22567832e-01 2.98387617e-01
6.71362281e-01 5.00242651e-01 6.11330330e-01 -1.42102849e+00
-6.41610861e-01 -5.28532982e-01 -7.02551544e-01 -9.95703578e-01
1.41252220e-01 -6.26492023e-01 1.00418970e-01 -1.75170910e+00
8.71459007e-01 -6.11181200e-01 -7.96078265e-01 6.63423181e-01
-6.37368858e-01 3.82665902e-01 2.83487409e-01 4.75289822e-01
-1.05986774e+00 7.08363175e-01 1.50769758e+00 -2.46048868e-01
-9.40730423e-02 -2.84492075e-01 -6.10088170e-01 7.77921736e-01
5.38787127e-01 -2.55832404e-01 -7.48101115e-01 -4.23718214e-01
2.86877930e-01 -9.92412195e-02 4.63617295e-01 -8.69362712e-01
2.79410213e-01 -2.76028365e-01 4.25115705e-01 -8.29747677e-01
5.02327800e-01 -1.06564128e+00 4.12632488e-02 2.29606748e-01
-4.27643299e-01 -9.54276621e-02 -3.79334278e-02 8.66014600e-01
-3.63667548e-01 -2.47521698e-01 5.01542747e-01 -2.02118158e-02
-9.39862072e-01 3.48106980e-01 2.34734416e-01 -2.92376161e-01
1.28452694e+00 -8.50225538e-02 -4.87944305e-01 -3.38137448e-01
-4.59940702e-01 5.95591903e-01 3.62488151e-01 6.20331585e-01
9.12508368e-01 -1.56174517e+00 -4.57550704e-01 -1.69520006e-01
7.21124530e-01 1.27210692e-01 3.16898286e-01 7.41584957e-01
-1.28945976e-01 8.96014422e-02 -1.12814941e-01 -5.81321955e-01
-1.81876671e+00 9.33848023e-01 -3.36135961e-02 -1.61957175e-01
-6.00257397e-01 8.13678443e-01 8.11612666e-01 8.05417970e-02
4.83115762e-01 2.96977889e-02 -8.78941774e-01 1.22350693e-01
5.71174979e-01 2.05991119e-01 -3.37968707e-01 -7.33699083e-01
-5.16937017e-01 7.59034276e-01 -3.46315026e-01 3.17202896e-01
1.26186740e+00 -2.17216298e-01 -4.41661417e-01 5.79243481e-01
1.07850802e+00 -9.74246934e-02 -1.23354936e+00 -4.36790466e-01
-6.43903390e-02 -8.28790843e-01 -5.50110079e-02 -4.47086722e-01
-1.05847597e+00 8.53624940e-01 4.09467816e-01 2.05358818e-01
1.28052843e+00 2.26708308e-01 5.80666304e-01 3.47335786e-01
1.44862235e-01 -6.91296041e-01 4.98367697e-01 1.09431148e-01
8.45913351e-01 -1.51056302e+00 3.78323406e-01 -1.09787774e+00
-7.19791591e-01 1.00466573e+00 1.16449904e+00 1.44525453e-01
4.34028536e-01 -8.90937597e-02 -1.36895731e-01 -2.94086844e-01
-7.74264336e-01 -4.85446543e-01 7.16331303e-01 4.74571079e-01
4.97171670e-01 -2.02583387e-01 -5.15452206e-01 2.05621019e-01
4.11145121e-01 -3.59308779e-01 1.34355739e-01 7.69270778e-01
-3.47251028e-01 -1.14702690e+00 -2.14247972e-01 3.43541443e-01
-1.50513440e-01 -7.95993730e-02 -6.54751599e-01 6.34423852e-01
2.58701354e-01 1.10495293e+00 2.14715064e-01 -3.77455413e-01
2.24611685e-01 -5.35373688e-01 3.76537710e-01 -7.30110109e-01
-1.39953986e-01 1.63059801e-01 2.81861518e-02 -3.69877905e-01
-7.90801227e-01 -4.53357935e-01 -1.32731509e+00 2.03101009e-01
-7.26581633e-01 -4.70981933e-02 7.47702643e-02 8.46583009e-01
2.15049312e-01 6.70015633e-01 6.82750762e-01 -5.34570634e-01
-7.24308565e-02 -7.69645929e-01 -3.04998368e-01 7.45860100e-01
1.40420839e-01 -1.16941571e+00 -9.08239484e-02 -4.97187786e-02] | [10.197175979614258, 1.6504255533218384] |
ab1b714b-bbc9-4ae4-8e3f-e97cec48cfe4 | a-benchmark-for-edge-preserving-image | 1904.01579 | null | http://arxiv.org/abs/1904.01579v1 | http://arxiv.org/pdf/1904.01579v1.pdf | A Benchmark for Edge-Preserving Image Smoothing | Edge-preserving image smoothing is an important step for many low-level
vision problems. Though many algorithms have been proposed, there are several
difficulties hindering its further development. First, most existing algorithms
cannot perform well on a wide range of image contents using a single parameter
setting. Second, the performance evaluation of edge-preserving image smoothing
remains subjective, and there lacks a widely accepted datasets to objectively
compare the different algorithms. To address these issues and further advance
the state of the art, in this work we propose a benchmark for edge-preserving
image smoothing. This benchmark includes an image dataset with groundtruth
image smoothing results as well as baseline algorithms that can generate
competitive edge-preserving smoothing results for a wide range of image
contents. The established dataset contains 500 training and testing images with
a number of representative visual object categories, while the baseline methods
in our benchmark are built upon representative deep convolutional network
architectures, on top of which we design novel loss functions well suited for
edge-preserving image smoothing. The trained deep networks run faster than most
state-of-the-art smoothing algorithms with leading smoothing results both
qualitatively and quantitatively. The benchmark is publicly accessible via
https://github.com/zhufeida/Benchmark_EPS. | ['Yizhou Yu', 'Xixi Jia', 'Feida Zhu', 'Zhetong Liang', 'Lei Zhang'] | 2019-04-02 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 1.73719138e-01 -1.04223236e-01 -4.29990217e-02 -3.67414087e-01
-7.99036205e-01 -1.37864262e-01 7.00170696e-01 -1.17436342e-01
-5.96959710e-01 5.81846058e-01 3.16121101e-01 -7.29074609e-03
1.70891911e-01 -6.55441463e-01 -6.51616633e-01 -7.63882518e-01
-7.50113279e-02 -2.89888054e-01 7.40634620e-01 -2.23121196e-01
2.86975592e-01 5.10922074e-01 -1.65018010e+00 1.91539332e-01
1.15969002e+00 1.06953406e+00 2.91508555e-01 6.97917759e-01
-1.03320695e-01 3.61748993e-01 -3.34316850e-01 -5.06628275e-01
4.68189090e-01 -1.78244025e-01 -4.47589964e-01 1.97808683e-01
1.15291536e+00 -3.28888923e-01 -5.26001036e-01 1.58111322e+00
7.44910777e-01 2.96180040e-01 5.50215542e-01 -1.19005513e+00
-1.17844248e+00 2.86455780e-01 -6.88808322e-01 8.19112808e-02
-3.23467642e-01 2.73441166e-01 6.87748671e-01 -1.03759992e+00
5.58743536e-01 1.20846975e+00 9.92524743e-01 6.66863322e-01
-1.20854223e+00 -3.37240696e-01 1.03588484e-01 2.22341701e-01
-1.05092525e+00 -3.52686495e-01 8.89476418e-01 -2.32003331e-01
5.01907408e-01 1.48097217e-01 4.07098979e-01 7.57945001e-01
4.77694064e-01 8.33796084e-01 1.14923394e+00 -2.77421087e-01
1.09951772e-01 1.51831135e-02 3.26639533e-01 6.88094079e-01
3.83614600e-01 2.95951068e-01 -3.64000618e-01 1.00395739e-01
9.05173898e-01 1.08219102e-01 -3.68532032e-01 -5.40065527e-01
-1.12568343e+00 6.87187254e-01 8.05042624e-01 1.18775442e-01
-3.16240638e-01 2.16050118e-01 6.15429342e-01 1.30632594e-01
7.43623972e-01 -1.05684269e-02 -2.48965308e-01 3.55694801e-01
-9.77446318e-01 4.25472856e-01 5.02242625e-01 7.38564849e-01
8.15793812e-01 2.25103974e-01 -5.39879143e-01 8.99105191e-01
7.20438659e-02 3.63869160e-01 2.42723614e-01 -1.07329881e+00
1.26532093e-01 4.01179194e-01 1.22614883e-01 -1.10799170e+00
-3.92546326e-01 -2.43893817e-01 -1.28467119e+00 9.15418983e-01
7.69178867e-01 -1.74721144e-02 -1.13894641e+00 1.52008665e+00
2.48817131e-01 4.03694928e-01 -4.22444977e-02 9.57768440e-01
1.30480218e+00 4.87708598e-01 1.88482478e-01 7.22341985e-02
1.48549354e+00 -1.35751188e+00 -6.96391046e-01 -3.57754886e-01
2.50844061e-01 -1.00447273e+00 1.41710865e+00 4.81044114e-01
-1.33054399e+00 -8.24122369e-01 -9.12810087e-01 -6.43970668e-01
-5.03009021e-01 2.60886490e-01 8.52181077e-01 5.47037661e-01
-1.54843986e+00 7.23979175e-01 -7.30976522e-01 -2.59679496e-01
6.02970600e-01 -1.09837212e-01 -4.15621787e-01 -4.20362875e-02
-1.05396497e+00 8.36538970e-01 2.53195196e-01 2.24193007e-01
-5.41393995e-01 -1.14254522e+00 -9.08622622e-01 6.03645518e-02
3.10862064e-01 -6.42266452e-01 1.20770931e+00 -7.36874878e-01
-1.36804461e+00 1.14081275e+00 -9.68673006e-02 -5.77274263e-01
9.27823603e-01 -3.01937193e-01 -4.07349348e-01 -1.37780428e-01
-1.61587670e-02 8.91929686e-01 1.01594627e+00 -1.48150146e+00
-6.57867253e-01 -1.07592806e-01 -2.55764216e-01 -1.53170764e-01
-2.98876673e-01 5.67863956e-02 -8.01155746e-01 -9.72231567e-01
-3.39732587e-01 -6.66190386e-01 -4.85351145e-01 4.64892089e-01
-4.07776862e-01 -2.40738690e-01 8.45231414e-01 -6.85913026e-01
1.13320172e+00 -2.13503337e+00 -1.93761751e-01 -2.55210042e-01
3.41319561e-01 7.13740051e-01 -4.17925090e-01 2.53555298e-01
-1.20825335e-01 -1.14548542e-02 -5.37461698e-01 -6.99814618e-01
8.21572766e-02 -6.82514459e-02 -4.22985286e-01 5.18463492e-01
1.10717863e-01 1.05412066e+00 -7.43052185e-01 -4.42802370e-01
5.28027177e-01 9.16132331e-01 -4.62625891e-01 7.77379051e-02
-1.07557617e-01 2.09164545e-01 -1.65775955e-01 6.44596219e-01
1.13583219e+00 -2.46401399e-01 -4.13474947e-01 -5.66099346e-01
-2.56865144e-01 -3.34907562e-01 -1.35221839e+00 1.66532040e+00
-1.40875757e-01 8.64908814e-01 2.05699071e-01 -8.84498060e-01
8.93992960e-01 -3.99731770e-02 2.21732721e-01 -5.24381399e-01
1.71728820e-01 9.59524587e-02 -3.95405263e-01 -3.21907848e-01
7.83143640e-01 2.87454695e-01 3.30049962e-01 -7.76931597e-03
-1.23158276e-01 -3.37127686e-01 4.15632248e-01 2.59836704e-01
8.13328028e-01 -1.37345523e-01 3.20034772e-01 -5.59391618e-01
6.21441364e-01 -1.64423048e-01 6.01073563e-01 8.09556961e-01
-7.14105487e-01 1.01766491e+00 2.43271902e-01 -7.20723212e-01
-1.11478257e+00 -1.27965796e+00 -3.39503884e-01 1.19603467e+00
4.47229445e-01 -2.13446736e-01 -9.99856353e-01 -4.06461507e-01
1.20483167e-01 4.74263221e-01 -8.42561781e-01 3.45256962e-02
-4.45490003e-01 -6.41494334e-01 2.96851426e-01 6.08242691e-01
1.16287613e+00 -1.25078404e+00 -2.22765699e-01 8.01302418e-02
2.48459466e-02 -1.08231747e+00 -1.03313959e+00 -2.54245460e-01
-8.11209440e-01 -1.06352997e+00 -8.66907179e-01 -1.02202690e+00
6.95770979e-01 7.15666533e-01 1.31111085e+00 3.95765007e-01
-4.26622570e-01 2.77053237e-01 3.59277539e-02 -5.79267025e-01
-4.18711126e-01 -1.58979103e-01 -3.12682986e-01 -9.00953636e-02
3.24114352e-01 -2.96670437e-01 -9.59227026e-01 2.95272321e-01
-1.15260994e+00 6.62947819e-03 6.56390131e-01 8.11331809e-01
7.75329113e-01 6.53831586e-02 4.78301138e-01 -8.44483674e-01
7.12986469e-01 -5.38642816e-02 -7.87290871e-01 2.57289171e-01
-4.67698008e-01 -8.54291469e-02 5.24563909e-01 -3.93535376e-01
-1.27474964e+00 9.65604708e-02 -3.21528107e-01 -1.17319532e-01
-3.52662653e-01 1.32029802e-01 -5.19295707e-02 -4.35660571e-01
8.49495649e-01 2.53209740e-01 2.32573062e-01 -5.48411310e-01
6.15796149e-01 2.15077385e-01 1.11571026e+00 -2.65880972e-01
8.41130376e-01 8.08133483e-01 2.93692630e-02 -1.10794008e+00
-9.39857960e-01 -4.24911201e-01 -4.99719650e-01 -1.93679288e-01
9.45767701e-01 -8.41952562e-01 -5.34273863e-01 9.46802795e-01
-1.18868911e+00 -6.75411940e-01 -1.02082498e-01 1.71700254e-01
-5.69080830e-01 7.08438694e-01 -6.85193419e-01 -3.67238671e-01
-5.33857107e-01 -1.31283474e+00 1.08332896e+00 6.01939857e-01
1.86638087e-01 -1.28338957e+00 -1.63011730e-01 6.42062947e-02
8.35094750e-01 3.63231599e-01 3.52811396e-01 -9.35590938e-02
-6.87756419e-01 -2.12671086e-01 -6.40510917e-01 6.36835754e-01
1.42459720e-01 1.50050491e-01 -1.00817728e+00 -5.20385861e-01
-1.19947672e-01 -3.16864669e-01 1.41220069e+00 9.92911220e-01
1.53950715e+00 4.90274020e-02 -1.90987870e-01 9.82958913e-01
1.41434574e+00 -3.26683432e-01 1.16679609e+00 3.77752393e-01
6.42525673e-01 3.00578952e-01 3.44398767e-01 2.49891672e-02
3.44765991e-01 5.11943877e-01 4.33065772e-01 -6.38022065e-01
-6.23701096e-01 -5.34387790e-02 8.23235437e-02 3.32802206e-01
-2.78764486e-01 -3.97306830e-02 -5.69664776e-01 5.65057337e-01
-1.69618821e+00 -1.00982440e+00 -3.39940697e-01 2.17905450e+00
9.25589800e-01 5.38400374e-02 1.84010088e-01 -1.79501064e-02
8.63138735e-01 4.38695461e-01 -5.07718623e-01 -1.18077330e-01
-2.14212298e-01 -8.82844329e-02 6.81127667e-01 6.40800655e-01
-1.59044790e+00 9.93744731e-01 6.09305620e+00 9.43544745e-01
-1.16210222e+00 -1.41849011e-01 7.54855871e-01 3.03782761e-01
-1.00711152e-01 -3.59560549e-01 -9.73766387e-01 4.79527265e-01
4.02297527e-01 -4.00356710e-01 2.17291743e-01 1.03327560e+00
3.50484759e-01 -2.50262953e-02 -7.23734260e-01 1.02223694e+00
6.51023239e-02 -1.69982636e+00 -6.70169815e-02 -1.99982241e-01
1.11724603e+00 2.41224423e-01 2.57664531e-01 1.26894489e-01
4.23138410e-01 -9.66593623e-01 4.37659562e-01 7.21274734e-01
6.38504207e-01 -3.74873936e-01 6.42502367e-01 3.52921933e-02
-1.33036983e+00 3.03110689e-01 -7.45512605e-01 1.86636016e-01
2.02586427e-01 8.82761657e-01 -3.10098417e-02 3.39867979e-01
9.56731617e-01 7.38104522e-01 -7.83774137e-01 1.48860872e+00
-2.80083597e-01 4.28479910e-01 -7.27425367e-02 2.75240064e-01
3.02566469e-01 -3.96607965e-01 4.56736952e-01 1.48089790e+00
2.11063549e-01 -5.39382994e-02 1.58530205e-01 8.66438627e-01
-2.66578764e-01 -2.55404525e-02 -5.39545238e-01 3.62378210e-01
3.97503167e-01 1.52609086e+00 -7.84956574e-01 -4.38920289e-01
-7.40772545e-01 9.07788634e-01 3.53251696e-01 5.86040616e-01
-7.41068959e-01 -4.76911187e-01 9.31575119e-01 -3.55489887e-02
1.56730205e-01 -1.92876518e-01 -5.99211395e-01 -1.15187848e+00
-1.59064569e-02 -7.54397690e-01 4.41849738e-01 -6.77246928e-01
-1.63780069e+00 5.85535228e-01 -2.23119810e-01 -1.11334991e+00
4.47230756e-01 -6.97903514e-01 -1.13993704e+00 9.22076225e-01
-1.98412299e+00 -1.11433053e+00 -9.05394673e-01 4.68343288e-01
6.66445136e-01 -2.57075783e-02 4.11574841e-01 2.86120743e-01
-3.17845613e-01 5.84081292e-01 1.68961555e-01 1.75054073e-01
1.11916125e+00 -1.41022325e+00 9.56848145e-01 1.23188436e+00
-1.51869267e-01 5.27441680e-01 7.76563406e-01 -5.90909958e-01
-1.09566820e+00 -1.30724490e+00 7.19779015e-01 -3.06516767e-01
6.83375835e-01 -1.70208693e-01 -1.58367765e+00 5.34206748e-01
4.42884326e-01 5.65046430e-01 1.82265742e-03 -3.74797583e-01
-2.24635541e-01 -1.21979155e-01 -1.12290180e+00 9.09821808e-01
1.10393214e+00 -1.08572677e-01 -4.85100448e-01 2.37841785e-01
5.92163682e-01 -6.16984785e-01 -7.04372287e-01 3.90670776e-01
2.79921502e-01 -1.40792120e+00 1.40883398e+00 -3.66804302e-01
5.50028920e-01 -3.63974482e-01 2.11851895e-01 -1.52651238e+00
-4.63317931e-01 -5.66090167e-01 1.55656323e-01 1.14078379e+00
9.06690806e-02 -7.61797905e-01 9.32650626e-01 6.93758190e-01
-5.13616383e-01 -6.23013258e-01 -5.38325191e-01 -8.90915573e-01
3.32838446e-01 -2.36620739e-01 4.25092340e-01 6.73426390e-01
-5.36881030e-01 -1.91058770e-01 -4.78530586e-01 1.95791498e-01
1.34542966e+00 -5.39231673e-02 1.00883007e+00 -1.14225948e+00
2.99182385e-01 -9.32109952e-01 -3.88329625e-01 -1.08013105e+00
1.45146579e-01 -8.04463208e-01 1.18399493e-01 -1.80168819e+00
1.02815464e-01 -1.37312680e-01 -1.21913545e-01 3.87821734e-01
-7.27623761e-01 5.54728568e-01 1.02678366e-01 -1.47927962e-02
-4.05936331e-01 5.60424328e-01 1.62418211e+00 -2.37046301e-01
-1.04239747e-01 -4.83501591e-02 -6.89194262e-01 1.00279009e+00
9.77901459e-01 3.48897725e-02 -3.19146156e-01 -5.54691315e-01
-3.10491860e-01 -4.18637156e-01 7.50440001e-01 -9.48215485e-01
3.80678743e-01 -2.66085982e-01 3.39232594e-01 -6.56210303e-01
-7.44718453e-03 -5.66094100e-01 -1.26910791e-01 4.63719696e-01
-4.70370442e-01 -1.94411680e-01 4.31432307e-01 6.08157039e-01
-2.15307608e-01 7.79615529e-03 1.58975136e+00 1.30923077e-01
-1.47598398e+00 6.46205246e-01 1.36094406e-01 3.71265262e-01
8.61793756e-01 -2.69204199e-01 -5.31085908e-01 -2.23271847e-01
-4.45370466e-01 6.61830530e-02 6.57823563e-01 4.62781698e-01
5.57034791e-01 -1.45029068e+00 -8.78866434e-01 3.02578419e-01
6.89778402e-02 -7.14531392e-02 5.42656600e-01 7.37156451e-01
-5.98123968e-01 4.72346619e-02 -4.62904960e-01 -5.19312620e-01
-1.26376879e+00 6.21246099e-01 3.76290560e-01 4.71600369e-02
-1.05005646e+00 8.61387312e-01 4.20455813e-01 -2.35395536e-01
3.89698386e-01 -6.49500251e-01 -1.20711094e-02 -3.52163851e-01
1.01473153e+00 5.29674530e-01 1.00717656e-01 -3.90695602e-01
-5.70905134e-02 7.05817342e-01 -7.87821859e-02 5.25129855e-01
1.29043686e+00 -1.77770242e-01 -4.43829894e-02 -1.45547405e-01
1.16601932e+00 -9.06410068e-02 -1.82645798e+00 -5.00593781e-01
-1.54000714e-01 -9.10220623e-01 3.54160756e-01 -5.13098419e-01
-1.38946760e+00 8.11593950e-01 6.53957129e-01 3.41193438e-01
1.12236440e+00 -2.26905361e-01 8.04742396e-01 1.63297564e-01
4.58449125e-02 -1.18976641e+00 1.69519380e-01 4.04533476e-01
1.21392548e+00 -1.86110163e+00 5.84047064e-02 -7.02062070e-01
-7.36506402e-01 9.64711070e-01 5.13141751e-01 -4.67895418e-01
7.54124403e-01 4.08223271e-01 3.76714289e-01 -8.23121667e-02
-3.32812279e-01 -3.80697608e-01 6.40269578e-01 8.12282443e-01
6.21405780e-01 -6.96624964e-02 -5.60684919e-01 3.20362002e-01
4.14610840e-02 2.75805235e-01 5.35534322e-01 4.06181008e-01
-7.07198322e-01 -9.33012426e-01 -2.84905791e-01 4.81627852e-01
-3.84807736e-01 -6.00898229e-02 1.21488841e-02 9.18882310e-01
-2.29563132e-01 6.65060103e-01 -5.83228841e-02 1.25303522e-01
4.99317765e-01 -3.53525788e-01 1.76844493e-01 -1.70825452e-01
-3.45464557e-01 -1.47908285e-01 -1.76508591e-01 -7.56042838e-01
-3.49563360e-01 -4.17025238e-01 -1.07501781e+00 -6.97243631e-01
2.45782174e-02 -3.22154522e-01 4.51920062e-01 4.44547296e-01
2.88282722e-01 4.48659360e-01 2.39441529e-01 -1.44438589e+00
-7.28068173e-01 -7.31411755e-01 -7.73785591e-01 9.69740152e-01
2.94349462e-01 -7.41437078e-01 -5.72049320e-01 4.65471625e-01] | [10.868101119995117, -1.4345905780792236] |
5a5da9d1-6bed-49d4-bace-29e0fbe79be5 | data-domain-adaptation-aided-deep-table | 2211.06648 | null | https://arxiv.org/abs/2211.06648v1 | https://arxiv.org/pdf/2211.06648v1.pdf | DATa: Domain Adaptation-Aided Deep Table Detection Using Visual-Lexical Representations | Considerable research attention has been paid to table detection by developing not only rule-based approaches reliant on hand-crafted heuristics but also deep learning approaches. Although recent studies successfully perform table detection with enhanced results, they often experience performance degradation when they are used for transferred domains whose table layout features might differ from the source domain in which the underlying model has been trained. To overcome this problem, we present DATa, a novel Domain Adaptation-aided deep Table detection method that guarantees satisfactory performance in a specific target domain where few trusted labels are available. To this end, we newly design lexical features and an augmented model used for re-training. More specifically, after pre-training one of state-of-the-art vision-based models as our backbone network, we re-train our augmented model, consisting of the vision-based model and the multilayer perceptron (MLP) architecture. Using new confidence scores acquired based on the trained MLP architecture as well as an initial prediction of bounding boxes and their confidence scores, we calculate each confidence score more accurately. To validate the superiority of DATa, we perform experimental evaluations by adopting a real-world benchmark dataset in a source domain and another dataset in our target domain consisting of materials science articles. Experimental results demonstrate that the proposed DATa method substantially outperforms competing methods that only utilize visual representations in the target domain. Such gains are possible owing to the capability of eliminating high false positives or false negatives according to the setting of a confidence score threshold. | ['Won-Yong Shin', 'Dongwoo Lee', 'Joungbin An', 'Hyebin Kwon'] | 2022-11-12 | null | null | null | null | ['table-detection'] | ['miscellaneous'] | [ 3.37182909e-01 6.98473901e-02 -2.46622205e-01 -2.50667840e-01
-7.12347984e-01 -5.84692836e-01 5.67173183e-01 5.99574924e-01
-3.97435874e-01 6.82056963e-01 -1.77959755e-01 -3.27699900e-01
1.68151811e-01 -1.03029764e+00 -1.11686957e+00 -2.17959344e-01
1.50181606e-01 8.06395948e-01 4.32939321e-01 -1.35136351e-01
6.12075627e-01 4.10757124e-01 -1.67421353e+00 6.90740645e-01
1.03690600e+00 1.35573137e+00 8.42973068e-02 3.76729757e-01
-3.26524168e-01 7.05402493e-01 -7.24153280e-01 -5.68604231e-01
3.60585123e-01 2.59023346e-02 -5.62654912e-01 5.99760637e-02
3.92339021e-01 -2.64129072e-01 -1.16017431e-01 1.09710729e+00
9.16996896e-02 1.55688105e-02 6.20172083e-01 -1.03961051e+00
-7.73373902e-01 6.01750731e-01 -4.53167737e-01 6.80026263e-02
3.91671628e-01 1.31867111e-01 1.00449264e+00 -1.06682527e+00
6.05220675e-01 1.04245031e+00 7.09370494e-01 2.13580072e-01
-1.30205739e+00 -5.30073702e-01 5.38257539e-01 5.10490954e-01
-1.36358130e+00 -3.16269398e-01 9.97882247e-01 -4.14998204e-01
1.08436894e+00 -4.19391468e-02 4.42665577e-01 1.10931528e+00
2.20761880e-01 7.01671481e-01 9.90627944e-01 -7.19470322e-01
5.30547500e-01 7.87028134e-01 -6.98380098e-02 7.34029830e-01
7.16675639e-01 -1.91642448e-01 -4.52355117e-01 5.92814758e-02
6.26131058e-01 -1.88365906e-01 -6.31287247e-02 -9.11109805e-01
-1.15553772e+00 8.26042354e-01 5.34188211e-01 3.64519566e-01
-4.95207191e-01 -3.77158850e-01 5.16610682e-01 1.49960294e-01
9.60164294e-02 5.46429932e-01 -5.10212004e-01 2.52192348e-01
-9.32963490e-01 5.70297353e-02 1.01839781e+00 9.70552564e-01
6.59321010e-01 -7.10220858e-02 -3.86922508e-01 6.52909040e-01
3.11212808e-01 1.97743744e-01 3.62170130e-01 -3.21031868e-01
6.89766884e-01 1.19071066e+00 1.82818249e-01 -1.16690862e+00
-3.19036782e-01 -5.46904325e-01 -7.28208065e-01 2.46733800e-01
5.79538465e-01 2.47455999e-01 -1.08806837e+00 1.36491609e+00
3.55765700e-01 -1.92666024e-01 1.59310848e-01 6.99483514e-01
5.10819614e-01 4.37549680e-01 -1.22646466e-02 9.97473821e-02
1.39795375e+00 -9.54912901e-01 -5.42660773e-01 -4.57873166e-01
3.94831777e-01 -5.56958616e-01 1.25474691e+00 8.43502343e-01
-6.81984782e-01 -8.11168432e-01 -1.55057681e+00 4.58463421e-03
-9.63566303e-01 4.43010479e-01 3.56093675e-01 7.87880063e-01
-8.09819281e-01 5.01309872e-01 -5.57394445e-01 -4.39528853e-01
6.78951979e-01 3.28554153e-01 -2.28310451e-01 1.36049762e-02
-1.09608793e+00 1.08410585e+00 6.22912943e-01 1.83497414e-01
-8.39733541e-01 -4.52929139e-01 -7.55888462e-01 3.17256600e-01
8.10417175e-01 -4.70760256e-01 1.02200150e+00 -8.86545479e-01
-1.28036082e+00 8.60577703e-01 2.04482779e-01 -6.58748448e-01
5.66782653e-01 -1.74857765e-01 -5.06276429e-01 8.80877152e-02
1.46771923e-01 3.81131917e-01 9.12342310e-01 -1.48068440e+00
-7.52923429e-01 -3.76860797e-01 9.83460322e-02 -7.33872503e-02
-4.83146995e-01 -4.08240527e-01 -7.38035560e-01 -4.35062289e-01
-1.17255278e-01 -5.71037173e-01 -8.88475701e-02 1.22032151e-01
-4.84748721e-01 -6.10468462e-02 4.40384328e-01 -6.02427065e-01
1.25758529e+00 -1.84513927e+00 -2.69720018e-01 5.03388464e-01
1.08052649e-01 4.92419213e-01 2.06116121e-02 2.22410932e-01
1.55346900e-01 -6.71967445e-03 -2.40947589e-01 -8.49382952e-03
-4.61968146e-02 9.01461300e-03 -2.34640002e-01 1.52125642e-01
5.70319712e-01 4.92159873e-01 -6.45646393e-01 -6.19900167e-01
4.57548827e-01 2.55858034e-01 -6.30336642e-01 1.61987841e-01
-4.90208894e-01 7.89462700e-02 -4.41449523e-01 7.81499982e-01
7.17455804e-01 -5.61159492e-01 6.65713966e-01 -3.42566431e-01
3.33275720e-02 1.99315190e-01 -1.36371088e+00 1.54534996e+00
-4.94123459e-01 2.88435966e-01 -3.28002006e-01 -1.27624547e+00
1.29391074e+00 -1.36466652e-01 1.57508343e-01 -9.28102851e-01
1.63253874e-01 2.26322368e-01 -1.29648028e-02 -4.24739093e-01
3.38878632e-01 2.33651131e-01 2.19931100e-02 -7.19857290e-02
7.71159530e-02 1.39330059e-01 4.04997557e-01 2.77296454e-02
1.11518621e+00 3.03441882e-01 4.94567573e-01 -8.59750211e-02
1.02568090e+00 5.29123902e-01 3.23092639e-01 1.00607610e+00
-2.29841322e-01 4.76886779e-01 5.52518368e-01 -6.67206526e-01
-1.15123534e+00 -9.38256681e-01 -2.06897765e-01 1.17153835e+00
1.94296330e-01 -3.78290415e-01 -6.93040967e-01 -1.06721699e+00
2.05242395e-01 8.62090349e-01 -9.02580082e-01 -2.75696129e-01
-4.71567690e-01 -6.53006792e-01 3.16632628e-01 8.02961946e-01
6.68960512e-01 -1.18449175e+00 -7.43152976e-01 4.62844551e-01
1.23223878e-01 -1.18745649e+00 9.25392881e-02 4.81691778e-01
-6.80894375e-01 -1.14813745e+00 -3.52159947e-01 -7.39037931e-01
6.15774155e-01 -1.21493384e-01 1.31003249e+00 -5.09575792e-02
-1.22559957e-01 1.35683209e-01 -2.45877400e-01 -6.54136240e-01
-5.34078538e-01 1.91599041e-01 -1.70972183e-01 4.13745493e-02
7.21481085e-01 -2.65855432e-01 -4.26901609e-01 2.13889867e-01
-6.67291403e-01 -7.03534707e-02 9.73207414e-01 9.81711745e-01
5.45578659e-01 3.18549015e-02 5.50364792e-01 -1.03498793e+00
6.21191978e-01 -3.30380321e-01 -9.53821480e-01 6.80831790e-01
-1.09684598e+00 4.37364668e-01 8.10233772e-01 -3.35598826e-01
-1.07331455e+00 1.77561909e-01 5.66478893e-02 -3.88330132e-01
-2.19700232e-01 6.08067095e-01 -3.36073548e-01 1.28807068e-01
8.76106620e-01 2.12368667e-01 -1.34815931e-01 -5.66687286e-01
1.94327235e-01 6.86602056e-01 5.21234691e-01 -5.07795751e-01
7.45520890e-01 4.27152842e-01 -2.32184723e-01 -1.76060706e-01
-7.18094349e-01 -1.92295685e-01 -8.56599569e-01 -9.66735482e-02
6.58236444e-01 -8.13600779e-01 -7.40901947e-01 1.92443460e-01
-1.07863522e+00 1.96769275e-02 3.63178598e-03 2.27693453e-01
-3.86245787e-01 1.36445627e-01 -1.34501025e-01 -8.13180327e-01
-3.66684467e-01 -9.63253558e-01 8.75631094e-01 2.25362301e-01
-9.37083140e-02 -8.40527713e-01 3.18200924e-02 2.93154597e-01
3.85483712e-01 2.14362279e-01 1.29833186e+00 -1.22307920e+00
-6.22897387e-01 -3.33015531e-01 -4.13327307e-01 2.79597223e-01
-4.52473341e-03 6.54907450e-02 -1.11675167e+00 -5.04612066e-02
-2.71840632e-01 -3.03390801e-01 9.32348490e-01 7.93174058e-02
1.13005340e+00 -1.13254331e-01 -5.15349746e-01 1.90453693e-01
1.54670310e+00 4.32069868e-01 3.66040528e-01 1.03266513e+00
6.12562716e-01 6.04542017e-01 7.20378995e-01 6.08513296e-01
4.75422978e-01 5.54739714e-01 5.26290715e-01 1.46918315e-02
1.70288548e-01 -3.34524840e-01 1.98817670e-01 1.77920952e-01
2.28236035e-01 -2.50450045e-01 -1.03734362e+00 5.91748416e-01
-2.00748634e+00 -5.86838305e-01 2.74296910e-01 2.29642415e+00
8.78148735e-01 9.60261703e-01 1.11709431e-01 4.66684252e-01
7.17687547e-01 -2.22131163e-01 -7.64678478e-01 -7.09486842e-01
1.45347521e-01 1.74855039e-01 3.65536779e-01 -3.07944678e-02
-1.34831011e+00 9.01272655e-01 5.53102970e+00 6.54027641e-01
-1.19285965e+00 -4.06155467e-01 6.61967039e-01 2.28148624e-01
-6.96357489e-02 -2.79557228e-01 -8.42001200e-01 2.44835779e-01
9.04497683e-01 -2.72094142e-02 3.51574779e-01 1.38850784e+00
-2.75538385e-01 -2.04230696e-01 -1.36391747e+00 9.17864978e-01
2.00901791e-01 -1.49980807e+00 2.11999193e-01 -2.22113669e-01
3.18450451e-01 -4.51412737e-01 7.39667341e-02 7.41077721e-01
1.92751676e-01 -8.76300216e-01 9.17334974e-01 3.69856298e-01
6.30146384e-01 -7.30427504e-01 8.87230277e-01 1.94567651e-01
-1.05788970e+00 -1.21840812e-01 -3.90349865e-01 -6.12431811e-03
-5.95627844e-01 4.55284566e-01 -1.32948232e+00 5.16497254e-01
8.62305701e-01 4.94110465e-01 -9.46794391e-01 9.95551407e-01
-1.86183512e-01 3.45189810e-01 -2.59394795e-01 -2.49788225e-01
1.11817777e-01 1.76728666e-01 1.75057158e-01 1.21789098e+00
-3.45718814e-03 -4.23110843e-01 1.01528257e-01 1.05900836e+00
-1.29742220e-01 1.26517460e-01 -7.01241016e-01 1.93676446e-02
6.09991372e-01 1.33418202e+00 -1.04199409e+00 -5.81914544e-01
-7.03122556e-01 7.39358008e-01 4.79375154e-01 1.22996889e-01
-9.15209234e-01 -4.09392893e-01 2.01439679e-01 7.93208405e-02
8.25115502e-01 1.27584442e-01 -8.71752501e-01 -1.03351068e+00
2.75961816e-01 -1.03100455e+00 4.71976846e-01 -5.36130130e-01
-1.26756394e+00 8.73660505e-01 -1.31661698e-01 -1.27673864e+00
-3.61695349e-01 -1.00768566e+00 -2.79650092e-01 6.58404469e-01
-1.56624019e+00 -1.00040543e+00 -3.19878638e-01 5.32634079e-01
4.77540612e-01 -3.31550777e-01 7.67768979e-01 1.48672849e-01
-7.39247322e-01 7.78080106e-01 1.27890930e-01 3.18837345e-01
8.24533403e-01 -1.34574318e+00 4.15370941e-01 9.20435667e-01
1.69691533e-01 6.13322377e-01 7.06612229e-01 -6.90713882e-01
-1.44358110e+00 -1.03160799e+00 6.12841070e-01 -6.08166814e-01
5.69394529e-01 -6.60698891e-01 -1.19388902e+00 5.46747684e-01
2.54721264e-03 1.07607916e-01 4.67283309e-01 1.15879364e-01
-4.79169041e-01 -4.39136147e-01 -1.41999328e+00 4.04815733e-01
8.74769449e-01 -3.18369091e-01 -7.46709406e-01 7.80854672e-02
2.70526975e-01 -4.72312450e-01 -6.76087976e-01 5.26315033e-01
7.24464774e-01 -1.11665261e+00 9.39573646e-01 -4.54742372e-01
4.95453715e-01 -3.21231097e-01 -2.29503155e-01 -9.23881531e-01
-2.50816941e-01 1.25910699e-01 -2.64230669e-01 1.22353494e+00
6.26633584e-01 -2.98435777e-01 9.09510016e-01 5.84144950e-01
1.97849106e-02 -8.15126956e-01 -6.80150926e-01 -6.85385168e-01
-2.00503871e-01 -2.95269728e-01 5.60742319e-01 7.36061990e-01
-1.47732764e-01 4.30757463e-01 -1.08497851e-01 2.60152221e-01
4.26494509e-01 3.35987300e-01 8.14224780e-01 -1.51033664e+00
-1.39948115e-01 -3.48005056e-01 -2.28100091e-01 -6.31739140e-01
-1.43741265e-01 -5.80079854e-01 2.23804310e-01 -1.54507029e+00
7.86840692e-02 -3.11928958e-01 -6.44794524e-01 6.03334188e-01
-9.45483670e-02 1.77642778e-01 1.18097991e-01 4.07412499e-02
-7.64450431e-01 3.32436293e-01 9.18483436e-01 -3.15141737e-01
-3.42241168e-01 -2.60417759e-01 -9.61446524e-01 7.55809426e-01
7.45824933e-01 -4.70281631e-01 -2.72670031e-01 -1.67346001e-01
2.44993523e-01 -1.75714538e-01 2.76518494e-01 -1.39661014e+00
2.58717865e-01 -4.17580130e-03 1.03759015e+00 -7.42709756e-01
4.00981912e-03 -1.00184441e+00 -2.91368008e-01 5.10059893e-01
-3.48653913e-01 -3.45602855e-02 6.26974285e-01 7.73174942e-01
-1.54205024e-01 -1.49660990e-01 6.09708071e-01 4.09482716e-04
-1.05898869e+00 -2.69054472e-01 -4.25300419e-01 -2.38761976e-01
1.00493038e+00 -5.53391457e-01 -2.84824371e-01 9.17422250e-02
-4.89397973e-01 1.21323839e-01 3.87423903e-01 5.34421861e-01
6.65767193e-01 -1.08242249e+00 -3.85587215e-01 4.59218651e-01
5.65520227e-01 -9.89987776e-02 -1.30529746e-01 5.25265753e-01
-4.12859738e-01 6.20001674e-01 -6.67608321e-01 -6.16949499e-01
-7.76169240e-01 1.17161179e+00 6.05588965e-02 -6.07766688e-01
-4.51992959e-01 6.78051293e-01 4.13499661e-02 -3.59325111e-01
5.28218746e-01 -6.20108783e-01 -4.69366729e-01 -3.05019431e-02
3.33892345e-01 4.20913436e-02 6.19931936e-01 -4.75591719e-02
-7.98444033e-01 3.17789167e-01 -4.99456912e-01 1.59436822e-01
1.09466946e+00 9.17897299e-02 5.78484237e-01 2.57901162e-01
6.36979401e-01 1.50711173e-02 -1.18540013e+00 -2.69673824e-01
3.50009352e-01 -3.35773647e-01 -6.28524274e-02 -1.26361120e+00
-8.00521672e-01 7.67673969e-01 7.41785347e-01 1.44502640e-01
1.14972770e+00 -3.47956121e-01 3.33589762e-01 6.19143307e-01
3.99803579e-01 -1.28763151e+00 2.87482411e-01 3.13221067e-01
8.12439680e-01 -1.62933266e+00 -1.98264252e-02 -2.87543207e-01
-8.08695734e-01 1.29319072e+00 9.44260538e-01 -2.19398454e-01
4.17353511e-01 2.29402021e-01 -2.20519751e-02 -1.66381840e-02
-7.76049256e-01 -2.00130597e-01 4.80137259e-01 8.59065950e-01
2.55155683e-01 -2.24462256e-01 -1.40048936e-01 8.97095561e-01
4.28023823e-02 2.49260873e-01 3.15554023e-01 1.05347550e+00
-5.17864645e-01 -1.02120864e+00 -6.40165150e-01 4.79594886e-01
-1.57390147e-01 -9.16736722e-02 -4.96858656e-01 1.04376149e+00
2.15801433e-01 9.11104262e-01 8.75266716e-02 -5.75296760e-01
5.10027111e-01 2.31670752e-01 4.52711821e-01 -5.70235908e-01
-6.61740899e-01 -8.70217457e-02 1.82427943e-01 -4.55480516e-01
-1.38046294e-01 -3.52000892e-01 -1.20282662e+00 -1.61699757e-01
-2.07252830e-01 -4.07042019e-02 6.04844451e-01 8.50873411e-01
5.02461493e-01 6.51912570e-01 2.90976226e-01 -5.26370466e-01
-6.78389251e-01 -8.52824330e-01 -2.30626687e-01 3.19435775e-01
1.21079288e-01 -1.02853882e+00 6.95342347e-02 4.65388671e-02] | [11.632781982421875, 2.974062442779541] |
a0579bf3-5de5-4791-ba27-a1d58cd4b425 | text-generation-with-speech-synthesis-for-asr | 2305.16333 | null | https://arxiv.org/abs/2305.16333v1 | https://arxiv.org/pdf/2305.16333v1.pdf | Text Generation with Speech Synthesis for ASR Data Augmentation | Aiming at reducing the reliance on expensive human annotations, data synthesis for Automatic Speech Recognition (ASR) has remained an active area of research. While prior work mainly focuses on synthetic speech generation for ASR data augmentation, its combination with text generation methods is considerably less explored. In this work, we explore text augmentation for ASR using large-scale pre-trained neural networks, and systematically compare those to traditional text augmentation methods. The generated synthetic texts are then converted to synthetic speech using a text-to-speech (TTS) system and added to the ASR training data. In experiments conducted on three datasets, we find that neural models achieve 9%-15% relative WER improvement and outperform traditional methods. We conclude that text augmentation, particularly through modern neural approaches, is a viable tool for improving the accuracy of ASR systems. | ['Xi Chen', 'Irina-Elena Veliche', 'Jessie Salas', 'Ethan Campbell-Taylor', "Antony D'Avirro", 'David Zhang', 'Duc Le', 'Farnaz Abtahi', 'Nelson Cheng', 'David Goss-Grubbs', 'Shashank Jain', 'Ziran Jiang', 'Gil Keren', 'Zhuangqun Huang'] | 2023-05-22 | null | null | null | null | ['text-augmentation', 'automatic-speech-recognition', 'speech-synthesis'] | ['natural-language-processing', 'speech', 'speech'] | [ 1.03895032e+00 7.15342581e-01 -1.72863826e-01 -3.08754653e-01
-1.16189337e+00 -4.14611816e-01 9.58395541e-01 -4.05388400e-02
-4.01858538e-01 7.48985887e-01 7.06324577e-01 -8.22068274e-01
8.19518387e-01 -3.34857732e-01 -6.23261988e-01 -2.91958004e-01
6.40877366e-01 6.30831659e-01 -2.44319573e-01 -5.42638302e-01
-2.77103484e-02 3.68611455e-01 -1.16343045e+00 3.42888892e-01
8.10783803e-01 7.12992311e-01 6.39575273e-02 9.12672758e-01
-3.40691864e-01 9.22864735e-01 -1.36911607e+00 -3.30004960e-01
-9.33340471e-03 -6.24792516e-01 -8.03599596e-01 2.03750432e-01
2.49023974e-01 -2.81546086e-01 -7.28867114e-01 7.11172104e-01
5.89480698e-01 3.79589319e-01 4.26960319e-01 -7.71352530e-01
-8.19702506e-01 1.34691751e+00 -9.20682326e-02 1.58630550e-01
3.61452878e-01 1.19819351e-01 7.77951419e-01 -1.07030237e+00
3.52622688e-01 1.23610246e+00 2.59804696e-01 1.18615723e+00
-1.36507869e+00 -5.29155910e-01 -4.21930589e-02 -4.23627228e-01
-1.10085702e+00 -1.20584011e+00 6.88043475e-01 8.43737647e-03
1.37046576e+00 4.46388394e-01 3.00328910e-01 1.50826859e+00
-5.53751290e-01 1.10671639e+00 9.12710607e-01 -9.07174170e-01
7.86859468e-02 1.44091278e-01 -1.09245189e-01 1.84026137e-01
4.26903330e-02 1.29680544e-01 -7.77382553e-01 2.89566278e-01
5.85139155e-01 -5.89270175e-01 -2.25879222e-01 6.97230875e-01
-1.36382997e+00 8.06719005e-01 -2.43428745e-03 3.68445218e-01
-3.72029245e-01 2.73628771e-01 5.09807587e-01 3.77347618e-01
1.06062210e+00 7.85564899e-01 -4.59829211e-01 -4.57742184e-01
-1.19055772e+00 2.09714353e-01 5.79551220e-01 1.12525618e+00
2.53065854e-01 1.16239369e+00 -5.68962932e-01 1.26802421e+00
3.17158997e-01 6.83628261e-01 9.44504261e-01 -6.73317730e-01
8.48686635e-01 4.38256711e-01 -1.64583325e-01 -1.24167018e-01
7.26410225e-02 -5.11233568e-01 -7.43578374e-01 -1.94261998e-01
1.09453529e-01 -4.11284328e-01 -1.50705218e+00 1.49986112e+00
-8.20764378e-02 -3.47275212e-02 4.95762199e-01 3.64837646e-01
9.23021734e-01 1.03318655e+00 8.83962288e-02 -3.31408143e-01
9.59577560e-01 -1.15166283e+00 -1.14592445e+00 -5.14831066e-01
9.45920050e-01 -9.59771693e-01 1.18157327e+00 1.23891614e-01
-1.45889616e+00 -4.92787749e-01 -9.78647113e-01 -1.64641157e-01
-4.51891780e-01 5.70115030e-01 1.96140587e-01 1.08002985e+00
-1.19962215e+00 3.10891449e-01 -8.13507974e-01 -2.20850110e-01
2.61218280e-01 3.30326796e-01 -1.16603523e-01 1.31856322e-01
-1.27415562e+00 8.83978128e-01 5.07850766e-01 8.90260637e-02
-8.10790479e-01 -6.90207779e-01 -1.08687985e+00 -1.12793215e-01
3.19864810e-01 -3.41564208e-01 1.98277652e+00 -7.91449845e-01
-2.06978202e+00 5.98926723e-01 -4.23054039e-01 -8.36978793e-01
1.90117329e-01 -2.98111379e-01 -5.94983399e-01 -2.50530034e-01
-3.02191913e-01 7.59259999e-01 8.19981337e-01 -1.12955594e+00
-4.24468935e-01 -1.38520598e-01 -3.69460970e-01 3.57136101e-01
-4.69103843e-01 5.42443335e-01 -1.14725620e-01 -1.28119171e+00
-6.02821819e-02 -9.84852612e-01 -3.32592398e-01 -7.97830462e-01
-6.04687929e-01 -3.76149863e-01 9.55059171e-01 -9.92680848e-01
1.45572555e+00 -1.80506766e+00 -9.71267596e-02 -1.28891081e-01
-9.85619575e-02 9.73632693e-01 -1.53970748e-01 5.13963938e-01
-2.00478226e-01 5.25567412e-01 -3.67504060e-01 -8.48364651e-01
-5.63371591e-02 1.31688952e-01 -7.15998471e-01 -1.52421683e-01
5.04692912e-01 1.16722190e+00 -6.14626527e-01 -1.03876051e-02
5.29425561e-01 4.99103397e-01 -1.90438986e-01 3.81782472e-01
-5.04077017e-01 3.69421005e-01 -2.25138273e-02 4.40654546e-01
6.82230592e-02 2.99753658e-02 -6.56350106e-02 2.10789204e-01
-4.39411886e-02 1.16402137e+00 -6.95049405e-01 1.55387199e+00
-7.31027842e-01 9.64007497e-01 -2.58371145e-01 -9.08651292e-01
1.17005038e+00 9.04030502e-01 -5.16918674e-02 -7.59232104e-01
2.35974640e-01 4.07500356e-01 2.10179150e-01 -1.14643415e-02
1.10671258e+00 -1.43394351e-01 1.46521300e-01 4.28216338e-01
2.02338919e-01 -4.97130275e-01 2.33846366e-01 9.38869640e-02
1.06002462e+00 -3.02165337e-02 1.00938179e-01 2.93102413e-01
4.16104913e-01 1.12954549e-01 -8.40466321e-02 6.92966700e-01
2.78621286e-01 8.35087597e-01 -1.27536461e-01 2.06744671e-02
-1.50820076e+00 -6.24451756e-01 1.28332257e-01 1.04877377e+00
-7.59821951e-01 -4.71866369e-01 -1.10781336e+00 -5.40999055e-01
-4.23151582e-01 1.34221578e+00 -2.83897489e-01 -1.25268564e-01
-9.77280140e-01 -6.14117861e-01 1.23832476e+00 7.81411827e-01
1.82997048e-01 -1.34218824e+00 2.85785437e-01 4.35867220e-01
-3.32776546e-01 -1.47815573e+00 -5.83012700e-01 1.65485770e-01
-9.51400638e-01 -5.87530471e-02 -1.07725298e+00 -6.36614561e-01
4.63359416e-01 1.91237614e-01 8.99114192e-01 -3.34049687e-02
1.36133134e-01 6.21386096e-02 -7.65510619e-01 -7.54751742e-01
-1.46099687e+00 4.53591168e-01 8.86045396e-02 -1.34518504e-01
2.05289155e-01 -2.94089496e-01 -9.27941352e-02 8.36855769e-02
-8.98431897e-01 2.90829957e-01 6.84142053e-01 7.02525854e-01
3.26957226e-01 -4.66585070e-01 9.82208550e-01 -9.85972285e-01
8.86620164e-01 -6.85710534e-02 -4.12258267e-01 -1.41368844e-02
-5.94113767e-01 1.12283602e-01 6.73897028e-01 -5.52172482e-01
-1.39516211e+00 -6.46502674e-02 -7.24965870e-01 -3.93098176e-01
-3.21184754e-01 5.77000380e-01 -1.24839783e-01 2.52684593e-01
9.00735617e-01 5.25869370e-01 5.16100228e-02 -4.13663208e-01
6.34466290e-01 1.23997867e+00 4.50856149e-01 -3.31125587e-01
7.96123087e-01 -1.46235943e-01 -4.99452025e-01 -1.38247430e+00
-7.61992693e-01 -2.17755288e-01 -3.28072280e-01 2.48407759e-02
7.19917595e-01 -1.00099456e+00 7.41086081e-02 3.26461464e-01
-1.28508389e+00 -7.17997789e-01 -6.38015568e-01 4.49750096e-01
-5.02178192e-01 9.05251205e-02 -7.24596739e-01 -1.15814269e+00
-6.85725808e-01 -1.07720625e+00 1.27022624e+00 -1.06817298e-02
-5.25854707e-01 -8.23796213e-01 3.26352976e-02 8.29194725e-01
6.66360319e-01 -4.36872751e-01 5.64234972e-01 -1.24856400e+00
-3.75749111e-01 -3.27982038e-01 4.28763358e-03 7.17092097e-01
2.35009730e-01 -8.04528892e-02 -1.27269709e+00 -1.49119228e-01
-2.71074295e-01 -3.38936597e-01 6.79092050e-01 2.91990846e-01
1.16620874e+00 -5.90785980e-01 -7.63970241e-02 2.91028529e-01
5.87209463e-01 6.24374270e-01 8.41038465e-01 -2.01516934e-02
8.13528001e-01 5.47754109e-01 4.00918186e-01 8.85379016e-02
-2.73975506e-02 6.87827766e-01 -3.99821736e-02 -2.37978607e-01
-7.84603477e-01 -4.14264381e-01 6.05383217e-01 1.24328887e+00
1.93175357e-02 -8.93198252e-01 -9.17374313e-01 6.36466861e-01
-1.39269912e+00 -7.23351479e-01 -1.99009776e-01 2.13577223e+00
1.07805336e+00 1.11381441e-01 7.16602951e-02 4.98780787e-01
7.78688848e-01 2.67112464e-01 -1.90708086e-01 -5.21747231e-01
-2.04590663e-01 7.24737227e-01 4.20415133e-01 5.26837170e-01
-7.70829499e-01 1.39582849e+00 6.70677471e+00 8.95668864e-01
-1.13995552e+00 6.80396706e-02 7.35795856e-01 -2.52794083e-02
-3.39864790e-01 -2.26241633e-01 -9.94217992e-01 2.91020334e-01
1.69597399e+00 -2.44088143e-01 4.44002777e-01 7.40781009e-01
4.98635381e-01 3.77843350e-01 -1.12129378e+00 7.65067458e-01
2.21468925e-01 -1.39499044e+00 5.16068101e-01 4.94688526e-02
8.63362849e-01 1.43661112e-01 1.79840460e-01 5.01994371e-01
5.05432248e-01 -1.32610917e+00 6.42348230e-01 -1.34612516e-01
1.23571038e+00 -7.01872170e-01 6.70962095e-01 1.71978563e-01
-8.92787516e-01 2.54768521e-01 1.84185933e-02 6.33634329e-02
3.64274204e-01 3.60927016e-01 -1.55286002e+00 3.43615055e-01
-1.67937484e-02 2.10888863e-01 -4.85738903e-01 4.86371428e-01
-1.87298402e-01 1.33352137e+00 -1.99683696e-01 -2.37133116e-01
1.55649602e-01 2.54501164e-01 6.71829462e-01 1.49791145e+00
3.12402576e-01 -2.32089069e-02 -1.25048041e-01 7.68293679e-01
-5.11072040e-01 2.72299170e-01 -7.34978855e-01 -9.37975645e-01
5.24626315e-01 8.54618788e-01 -5.00465631e-01 -6.94536448e-01
-2.72663504e-01 9.57486570e-01 8.14708769e-02 4.41505283e-01
-3.84562016e-01 -5.06950974e-01 4.53291178e-01 9.28750038e-02
-4.60421257e-02 -4.57876176e-01 -5.23481905e-01 -1.14242387e+00
1.03538990e-01 -1.34203482e+00 -1.43532813e-01 -8.40546727e-01
-7.89037049e-01 1.05163741e+00 -2.20920682e-01 -8.79657030e-01
-7.31273293e-01 -4.36417431e-01 -3.96330744e-01 1.12733018e+00
-1.31163323e+00 -1.25162470e+00 -3.75966635e-03 1.18831284e-01
1.45021880e+00 -6.39935195e-01 8.77028763e-01 3.37444723e-01
-6.44089043e-01 9.31723356e-01 -3.87873454e-03 3.00378948e-01
2.95028538e-01 -1.28224576e+00 1.47956789e+00 1.17415738e+00
4.18622792e-01 5.35692275e-01 7.63130426e-01 -8.45249057e-01
-1.14400470e+00 -1.32967865e+00 1.00044334e+00 -5.23743749e-01
6.79826975e-01 -5.94982088e-01 -1.03092217e+00 9.01868463e-01
4.49525177e-01 -2.84054130e-01 5.27697265e-01 -1.79151833e-01
-7.56602213e-02 2.19477043e-01 -6.62248135e-01 9.45856750e-01
9.04483259e-01 -6.72148108e-01 -6.17437959e-01 1.54822603e-01
1.53796101e+00 -5.86981773e-01 -6.82819247e-01 5.08900702e-01
6.71991706e-02 -1.69013202e-01 6.96071804e-01 -7.20211208e-01
4.36397761e-01 -1.54216111e-01 -3.28205973e-01 -1.73963070e+00
2.42608026e-01 -1.11028922e+00 -2.49666288e-01 1.63017440e+00
9.86658514e-01 -4.23007160e-01 9.09443319e-01 5.12329996e-01
-7.77905524e-01 -3.42854470e-01 -7.17660904e-01 -8.32763851e-01
2.88320370e-02 -6.02219164e-01 4.93008852e-01 7.62463808e-01
-1.22161314e-01 6.61345959e-01 -3.94590288e-01 -7.29026943e-02
1.75295681e-01 -7.68903375e-01 7.81267226e-01 -7.98504889e-01
-9.30425972e-02 -4.86140132e-01 7.67936632e-02 -1.24246883e+00
3.10969293e-01 -1.11928201e+00 5.05543113e-01 -1.53023458e+00
-4.59203869e-01 -3.66110027e-01 2.66746879e-01 5.28436601e-01
-1.41990796e-01 3.49542975e-01 9.36358422e-02 -1.95495993e-01
3.71020772e-02 9.21003461e-01 1.01536775e+00 -7.25366101e-02
-4.46537405e-01 1.98402137e-01 -6.27765477e-01 4.26108003e-01
9.62035298e-01 -3.78025055e-01 -4.49324608e-01 -5.28847039e-01
-2.28738323e-01 3.56179327e-01 -1.94141343e-01 -9.26464558e-01
7.27415904e-02 -7.29473010e-02 6.56428784e-02 -5.58649540e-01
5.10419607e-01 -3.28108132e-01 -3.42606097e-01 7.08727166e-02
-8.60058665e-01 5.15763313e-02 4.45534170e-01 2.05939502e-01
-3.72704536e-01 -3.03788960e-01 7.05917358e-01 2.98723523e-02
-1.16045088e-01 9.35789347e-02 -8.19373131e-01 1.71357095e-01
4.04365629e-01 -1.91349894e-01 -3.39078993e-01 -5.97212911e-01
-7.31684327e-01 -1.95423603e-01 -3.39765362e-02 6.83175921e-01
7.59440422e-01 -1.15014851e+00 -9.49152589e-01 2.67445147e-01
1.80154130e-01 1.16414197e-01 -1.38170451e-01 1.90738901e-01
-1.41128883e-01 8.07869971e-01 4.65224892e-01 -2.36439854e-01
-1.34283459e+00 2.28772089e-01 2.50308156e-01 -1.85938641e-01
-4.51762646e-01 6.98275208e-01 -8.48680735e-02 -5.33702314e-01
4.10750896e-01 -3.13825548e-01 -1.25435740e-01 -3.16739202e-01
6.42547488e-01 4.25209910e-01 5.45317352e-01 -6.74709439e-01
2.23183915e-01 -2.46597320e-01 -3.24626625e-01 -8.34664404e-01
1.19903946e+00 -4.50768098e-02 2.84738928e-01 4.74224478e-01
8.64543378e-01 5.42162545e-02 -5.59129953e-01 -3.29792261e-01
1.66282311e-01 -1.09035015e-01 1.91688180e-01 -1.02695453e+00
-7.57476509e-01 1.01064599e+00 1.01932004e-01 3.43258560e-01
8.59626353e-01 -4.57213484e-02 1.04809642e+00 7.34615445e-01
-1.52966291e-01 -1.10273898e+00 1.93538919e-01 8.81263494e-01
1.08322084e+00 -1.21179867e+00 -5.67802787e-01 -3.68578136e-01
-7.00688303e-01 9.34849918e-01 7.20119894e-01 1.19180351e-01
1.20281331e-01 5.20148396e-01 8.71029720e-02 2.46029511e-01
-8.46236885e-01 -2.48286054e-01 3.17774802e-01 6.66589200e-01
9.00530219e-01 1.14655914e-02 -5.59557416e-02 1.54950634e-01
-7.97949910e-01 -1.82814762e-01 8.15566182e-01 6.22248113e-01
-3.98879021e-01 -1.39140022e+00 -2.03847215e-01 5.78220725e-01
-6.97596431e-01 -6.50378644e-01 -6.76761746e-01 5.44355512e-01
-5.86695015e-01 1.26076221e+00 1.22680441e-01 -3.22128594e-01
3.28628093e-01 4.72479641e-01 2.19971061e-01 -1.21125555e+00
-6.85444832e-01 2.11810365e-01 7.30851769e-01 2.39258888e-03
-6.91029876e-02 -5.29156566e-01 -1.18446803e+00 3.39478701e-02
-6.00740314e-01 7.15775862e-02 1.09373045e+00 1.07351184e+00
1.95858702e-01 8.72246385e-01 5.42935252e-01 -6.42077684e-01
-5.09411573e-01 -1.50604570e+00 -8.65533277e-02 7.38349780e-02
4.93801355e-01 -6.60984814e-02 -2.78012872e-01 2.14281067e-01] | [14.422961235046387, 6.906684875488281] |
4bc9e854-bebf-4ac0-a3f4-c8a580db0be0 | can-machine-generate-traditional-chinese | 1606.05829 | null | http://arxiv.org/abs/1606.05829v1 | http://arxiv.org/pdf/1606.05829v1.pdf | Can Machine Generate Traditional Chinese Poetry? A Feigenbaum Test | Recent progress in neural learning demonstrated that machines can do well in
regularized tasks, e.g., the game of Go. However, artistic activities such as
poem generation are still widely regarded as human's special capability. In
this paper, we demonstrate that a simple neural model can imitate human in some
tasks of art generation. We particularly focus on traditional Chinese poetry,
and show that machines can do as well as many contemporary poets and weakly
pass the Feigenbaum Test, a variant of Turing test in professional domains. Our
method is based on an attention-based recurrent neural network, which accepts a
set of keywords as the theme and generates poems by looking at each keyword
during the generation. A number of techniques are proposed to improve the
model, including character vector initialization, attention to input and
hybrid-style training. Compared to existing poetry generation methods, our
model can generate much more theme-consistent and semantic-rich poems. | ['Qixin Wang', 'Dong Wang', 'Tianyi Luo'] | 2016-06-19 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [ 4.49137807e-01 3.56726855e-01 1.47848457e-01 3.73515338e-02
-6.24245346e-01 -3.15463424e-01 8.69710147e-01 -4.87349123e-01
-2.68154144e-01 8.86440933e-01 2.58583546e-01 -3.98759097e-01
2.73922861e-01 -1.27482557e+00 -5.48287928e-01 -4.66651678e-01
4.90027726e-01 7.25520492e-01 5.21152187e-03 -8.38064134e-01
5.61932445e-01 2.28606611e-01 -1.34238219e+00 3.72124374e-01
1.05579078e+00 4.67427850e-01 4.40710217e-01 7.87012339e-01
-3.81325185e-01 1.20423722e+00 -1.24352586e+00 -7.63881922e-01
-1.22080639e-01 -1.12953722e+00 -1.15479505e+00 -8.37611854e-02
-2.42612008e-02 5.60502969e-02 -2.98412532e-01 7.38107085e-01
6.37052119e-01 5.27738571e-01 7.64825106e-01 -8.40894938e-01
-1.23934102e+00 1.36181235e+00 -8.77458602e-02 9.52165276e-02
2.79651135e-01 1.19223058e-01 1.01223505e+00 -8.32139552e-01
4.24453139e-01 8.69921088e-01 6.53916419e-01 1.01463974e+00
-9.60196137e-01 -5.18351078e-01 -3.56487632e-01 3.19367230e-01
-1.25550318e+00 -2.24475726e-01 1.13196552e+00 -1.92141026e-01
1.04813135e+00 2.87937641e-01 8.54284644e-01 1.24095643e+00
1.21934570e-01 8.98722172e-01 9.21355665e-01 -7.65612543e-01
1.71234801e-01 7.13832751e-02 -5.08478224e-01 4.06996399e-01
-1.48708448e-01 -1.49262086e-01 -6.28891945e-01 1.13246366e-01
1.13594544e+00 -4.41824734e-01 -1.99447989e-01 4.65430617e-01
-1.30591393e+00 9.69461322e-01 2.63769984e-01 7.20611393e-01
-5.95479786e-01 3.57718021e-01 1.82174921e-01 6.20136876e-03
1.82055414e-01 1.21027672e+00 8.06805640e-02 -3.82951796e-01
-1.17459559e+00 3.29735845e-01 8.52663159e-01 8.46686423e-01
3.04745942e-01 7.55273223e-01 -5.58366299e-01 1.04095721e+00
-1.48618191e-01 2.01382354e-01 1.11109114e+00 -8.09741080e-01
4.07978833e-01 3.24240416e-01 -3.81997041e-02 -8.20050836e-01
1.20476104e-01 -4.39611822e-01 -1.07327688e+00 -1.07138336e-01
1.12358399e-01 -2.37716556e-01 -9.18738961e-01 1.70832348e+00
-2.04253450e-01 2.27587894e-01 1.42834466e-02 7.39477992e-01
9.61851060e-01 1.11398947e+00 7.59744421e-02 -5.06761000e-02
1.13242292e+00 -1.24990726e+00 -6.48811638e-01 -3.82656068e-01
2.69758165e-01 -5.71144044e-01 1.34277189e+00 4.52169836e-01
-1.60607624e+00 -8.94546509e-01 -1.15848470e+00 -8.39639157e-02
-2.93670326e-01 9.32292417e-02 7.29349136e-01 7.17489362e-01
-9.88720596e-01 9.89160597e-01 -6.47890985e-01 -1.42128244e-01
3.57873917e-01 2.29406893e-01 -1.05106235e-02 4.78044182e-01
-1.43866384e+00 1.10662591e+00 8.19628596e-01 2.26019561e-01
-6.46588027e-01 -3.15605313e-01 -6.94208682e-01 2.10044533e-01
3.85805480e-02 -9.11353230e-01 1.52154338e+00 -9.65859294e-01
-2.05302668e+00 9.92928863e-01 7.31735602e-02 -6.57092452e-01
3.64775985e-01 1.45485613e-03 -4.85280275e-01 1.26281381e-01
-2.53797956e-02 7.52366424e-01 6.01590455e-01 -1.05570817e+00
-5.29849529e-01 2.24683121e-01 5.59748616e-03 2.54450917e-01
-2.58368343e-01 5.61660342e-02 -2.80268520e-01 -8.37724566e-01
-3.01579479e-02 -6.16722822e-01 -3.52492243e-01 -7.64868319e-01
-5.79363465e-01 -5.92786133e-01 6.92193136e-02 -6.08027518e-01
1.34466457e+00 -1.56616795e+00 7.92820677e-02 7.37386569e-02
1.28862008e-01 2.43469417e-01 -1.16147965e-01 4.87526327e-01
1.51442692e-01 4.28604841e-01 -2.67895073e-01 -2.24668965e-01
1.99808836e-01 1.65915206e-01 -3.94417465e-01 -2.36830682e-01
4.27616924e-01 1.38458192e+00 -1.00844932e+00 -3.97170365e-01
-1.44688159e-01 3.45686227e-01 -3.97516489e-01 2.11137593e-01
-3.55069309e-01 3.15328807e-01 -2.65372187e-01 2.36644730e-01
-1.48147745e-02 -5.02760112e-01 3.89546454e-02 4.53799129e-01
8.74727368e-02 6.86371624e-01 -6.66920662e-01 1.64644682e+00
-6.57433271e-01 7.37784982e-01 -6.17252290e-01 -1.03921914e+00
1.23178935e+00 4.87389386e-01 -1.03859939e-01 -5.99938393e-01
2.38790780e-01 2.31681824e-01 2.89327174e-01 -2.98075050e-01
9.37318563e-01 -5.38523436e-01 -1.47002161e-01 7.26565063e-01
-3.99290472e-02 -6.14827394e-01 3.60116929e-01 5.71573945e-03
1.05012703e+00 3.74463826e-01 3.32154036e-01 4.63836566e-02
5.27078271e-01 1.46844447e-01 2.62808532e-01 1.11147356e+00
4.23594981e-01 8.35403204e-01 2.80245811e-01 -3.72427046e-01
-1.32150257e+00 -1.05604565e+00 3.18711430e-01 1.13869238e+00
-1.40396118e-01 -3.29743057e-01 -9.72187221e-01 -2.33982295e-01
-6.57159269e-01 1.13099563e+00 -4.23191339e-01 -3.00207943e-01
-9.16487932e-01 -7.15612769e-01 8.31619203e-01 6.28905296e-01
7.92086720e-01 -2.08783126e+00 -6.82469606e-01 6.14365339e-01
-2.61619508e-01 -8.51202726e-01 -3.09440583e-01 -5.50820678e-02
-7.32775092e-01 -4.28356469e-01 -9.49959099e-01 -1.26078737e+00
4.50753987e-01 -2.48063177e-01 1.34256816e+00 1.48996994e-01
-3.15141864e-02 -7.86078796e-02 -4.11523938e-01 -4.54956681e-01
-8.69529366e-01 4.12970781e-01 -2.13101640e-01 -2.42766157e-01
3.74902964e-01 -7.48206198e-01 -1.39276743e-01 -1.61257153e-03
-9.44179773e-01 5.85199118e-01 7.17306972e-01 1.24251735e+00
3.02062571e-01 5.02029210e-02 8.05928946e-01 -1.03268349e+00
1.20721102e+00 -3.30819160e-01 2.93330848e-02 1.58810303e-01
-4.49419826e-01 1.42083913e-01 9.28354442e-01 -8.01414549e-01
-1.11416554e+00 -2.61958003e-01 -4.34349328e-01 -4.03051525e-02
-4.60995175e-03 6.19599938e-01 -3.11123878e-02 2.40335926e-01
9.90265846e-01 6.72546864e-01 -2.19180420e-01 -1.72502443e-01
2.51845896e-01 7.39647627e-01 9.57517207e-01 -6.38344646e-01
9.43082750e-01 -8.11856613e-02 -3.33371609e-01 -7.04354763e-01
-7.58423150e-01 7.60833845e-02 -3.03764850e-01 -1.28093615e-01
6.15118980e-01 -5.26338577e-01 -7.22534597e-01 4.38045174e-01
-1.36249280e+00 -6.12025738e-01 -5.00341952e-01 3.83127302e-01
-7.90734887e-01 1.06047936e-01 -9.12929118e-01 -8.50170374e-01
-7.28083313e-01 -5.43785512e-01 7.32560992e-01 5.69862068e-01
-6.94527209e-01 -9.03421581e-01 1.04448840e-01 2.63664454e-01
5.26738107e-01 1.35436118e-01 9.96598065e-01 -7.56404936e-01
-3.65389168e-01 -8.19173381e-02 2.56765306e-01 3.62373859e-01
-4.13911231e-02 -3.22521001e-01 -8.01817298e-01 3.71841460e-01
6.96860701e-02 -4.18899149e-01 7.73029506e-01 8.50386992e-02
1.16432810e+00 -5.19685149e-01 9.82691906e-03 6.16699338e-01
1.05616641e+00 3.97378922e-01 1.22450113e+00 5.18949330e-01
6.16845429e-01 3.60741884e-01 3.02515589e-02 3.39719325e-01
3.41019928e-01 4.39479113e-01 -1.51942164e-01 -3.74234729e-02
-2.59647425e-02 -7.70929754e-01 3.45960021e-01 1.04832780e+00
-6.42370999e-01 -3.65806818e-01 -8.48061562e-01 7.76367366e-01
-1.62110484e+00 -1.41891694e+00 -1.67757254e-02 1.87235963e+00
1.24928188e+00 3.21835339e-01 3.05558890e-02 3.81608337e-01
9.21305239e-01 4.92240191e-02 -2.35191390e-01 -7.33730972e-01
-3.63050908e-01 1.01332867e+00 9.29891411e-03 2.63584673e-01
-6.61885440e-01 1.43873608e+00 6.55419922e+00 9.80581045e-01
-1.03567672e+00 4.47570607e-02 6.22889459e-01 1.48329124e-01
-4.98271525e-01 -2.35360622e-01 -4.59181756e-01 5.12778580e-01
9.93840277e-01 -5.14318824e-01 5.67401826e-01 8.13188195e-01
7.08239675e-02 3.26563805e-01 -9.50653493e-01 1.06160092e+00
5.33131897e-01 -1.72332513e+00 3.02715003e-01 -9.89664942e-02
1.05083060e+00 -3.68453503e-01 4.55520041e-02 6.48133576e-01
4.31671768e-01 -1.52542162e+00 8.08468699e-01 4.12846774e-01
7.42157221e-01 -6.96505368e-01 5.88378966e-01 4.59498256e-01
-8.20745528e-01 2.74440348e-01 -4.02671069e-01 -3.16381067e-01
2.99218267e-01 -3.25190090e-02 -1.01056468e+00 2.93772697e-01
2.17154354e-01 3.02026123e-01 -4.77680922e-01 1.06816924e+00
-8.15807402e-01 8.39795887e-01 -1.07087411e-01 -6.27282143e-01
3.39164585e-01 -1.10468350e-01 2.37339750e-01 1.15858531e+00
6.61032081e-01 2.88405418e-01 -2.13565066e-01 1.35330558e+00
-2.33102649e-01 2.32893333e-01 -5.95489919e-01 -4.50952828e-01
3.85460109e-01 1.10016537e+00 -7.93512225e-01 -5.34598231e-01
1.72002122e-01 1.44298077e+00 1.92096815e-01 2.37502351e-01
-7.96611428e-01 -8.33445311e-01 -1.71276838e-01 1.22982472e-01
2.13884771e-01 -1.30719125e-01 -5.81908524e-01 -1.00911272e+00
-2.00132914e-02 -8.78310442e-01 -1.49723259e-03 -1.14918935e+00
-1.14560449e+00 1.01602292e+00 -4.09148961e-01 -9.98066425e-01
-7.73664176e-01 -4.57152456e-01 -1.31312275e+00 9.26660836e-01
-9.89216745e-01 -1.16639376e+00 -5.91750331e-02 3.49704266e-01
6.99923158e-01 -4.05863017e-01 9.74452257e-01 -1.18596025e-01
-1.93581194e-01 6.44247770e-01 -2.62869895e-01 4.50874358e-01
2.02356815e-01 -1.13305354e+00 8.56619179e-01 7.26784885e-01
6.00966930e-01 6.14897966e-01 6.38252735e-01 -5.62381387e-01
-9.34603810e-01 -7.88880646e-01 1.46401179e+00 -2.75457621e-01
5.73525429e-01 -2.82613873e-01 -7.27643788e-01 5.35485744e-01
3.41961056e-01 -7.71637261e-01 5.70299149e-01 -5.46801873e-02
1.06102906e-01 3.54549736e-01 -7.14726210e-01 1.02880955e+00
1.18320799e+00 -3.89038742e-01 -1.09293306e+00 4.85290527e-01
5.74298203e-01 -3.97833616e-01 -5.61255336e-01 1.51451882e-02
4.46865052e-01 -7.04003274e-01 7.61652529e-01 -7.56622136e-01
1.11669755e+00 -1.21383317e-01 3.14591706e-01 -1.46446788e+00
-5.49120188e-01 -1.18364561e+00 -1.16174690e-01 1.23279941e+00
6.45229220e-01 -2.49293596e-01 1.03002477e+00 3.10217351e-01
-3.60522538e-01 -6.05778396e-01 -6.90309823e-01 -7.81183422e-01
2.34982729e-01 -3.43755662e-01 7.07219839e-01 8.45103562e-01
2.20260695e-01 9.43505228e-01 -6.72076941e-01 -3.77256542e-01
1.96631595e-01 3.72154526e-02 6.56351268e-01 -1.25112331e+00
-6.34452164e-01 -9.44495261e-01 -2.55021036e-01 -9.22198772e-01
3.17117453e-01 -1.04381549e+00 2.42808476e-01 -1.74519265e+00
-7.48177171e-02 -1.29454583e-01 3.66393547e-03 4.55539733e-01
-2.51646549e-01 5.26148617e-01 3.11317563e-01 1.14149436e-01
-2.64777094e-01 5.94822526e-01 1.41739106e+00 -3.10637444e-01
-5.02927005e-01 5.18600158e-02 -1.11270106e+00 5.05670965e-01
1.12068677e+00 -2.63333142e-01 -1.95764780e-01 -4.61240292e-01
5.34086823e-01 1.32289976e-01 1.39284551e-01 -1.10826552e+00
3.94712210e-01 -2.72148907e-01 6.55604362e-01 -3.51819813e-01
4.29355562e-01 7.41074309e-02 2.50714362e-01 4.57562715e-01
-5.67676544e-01 4.99781445e-02 -7.50131160e-02 -3.47120054e-02
-2.77575970e-01 -5.93932867e-01 6.02692485e-01 -6.96976960e-01
-5.33474147e-01 -3.40543129e-02 -6.39453650e-01 1.50921598e-01
5.98184645e-01 -5.38029253e-01 -5.33374734e-02 -6.75558209e-01
-6.05228901e-01 1.16389571e-02 1.24956705e-01 3.20771784e-01
7.23036766e-01 -1.54227650e+00 -1.11348855e+00 -1.93504598e-02
-1.52933538e-01 -8.63613039e-02 1.08190604e-01 2.57071286e-01
-7.70422339e-01 3.25313449e-01 -3.74564230e-01 -5.50261475e-02
-6.15947306e-01 2.32107863e-01 2.05495119e-01 -4.52475667e-01
-7.01692700e-01 9.40099478e-01 -8.11157152e-02 -2.65912265e-01
-6.12478741e-02 9.62871909e-02 -4.57651198e-01 -2.66725391e-01
7.24570572e-01 1.09734289e-01 -2.38522619e-01 -3.97987992e-01
1.41460687e-01 3.55075628e-01 1.15813524e-01 -6.03874981e-01
1.32556653e+00 4.15458322e-01 -1.28284723e-01 5.40757716e-01
5.27121842e-01 2.44621970e-02 -6.60287082e-01 5.04891910e-02
5.91283552e-02 4.45277011e-03 -3.11983496e-01 -9.10719514e-01
-6.47338212e-01 1.01061451e+00 -2.13456020e-01 2.55634785e-01
9.66679811e-01 -9.78099555e-02 1.13299799e+00 7.16021121e-01
3.06944579e-01 -1.42551541e+00 1.45123824e-01 7.57803559e-01
8.45838010e-01 -7.15201020e-01 -3.88535649e-01 1.16454601e-01
-1.02260125e+00 1.08577275e+00 7.59303391e-01 -3.54103088e-01
-5.87120540e-02 3.97629617e-03 -1.82300925e-01 4.84321006e-02
-8.80222738e-01 -1.94399685e-01 2.03036278e-01 5.56579769e-01
6.83536291e-01 4.63271998e-02 -5.57998002e-01 9.28955853e-01
-1.20235908e+00 1.43919796e-01 8.55477035e-01 6.67267621e-01
-6.39852881e-01 -1.13957107e+00 -3.61326277e-01 3.69188517e-01
-6.59229457e-01 -5.12157321e-01 -5.74875593e-01 6.04434192e-01
5.37945479e-02 7.90580034e-01 2.42128462e-01 -4.91708308e-01
2.78774500e-02 3.27754498e-01 6.56836569e-01 -8.76567662e-01
-1.03294039e+00 -2.08877683e-01 2.29190812e-01 8.51079971e-02
-2.32944965e-01 -2.27351561e-01 -1.26122963e+00 -4.34555054e-01
-1.30846873e-01 5.59388876e-01 5.16578972e-01 9.81322169e-01
-3.92361358e-02 6.38473094e-01 5.04053593e-01 -9.06422913e-01
-5.15192330e-01 -1.37881672e+00 -5.44641018e-01 4.59000975e-01
-3.57802004e-01 -9.38382000e-02 -2.68092841e-01 2.81587124e-01] | [11.585979461669922, 9.369278907775879] |
7e32c481-ee10-445a-9c3f-703c473778e5 | only-a-matter-of-style-age-transformation | 2102.02754 | null | https://arxiv.org/abs/2102.02754v2 | https://arxiv.org/pdf/2102.02754v2.pdf | Only a Matter of Style: Age Transformation Using a Style-Based Regression Model | The task of age transformation illustrates the change of an individual's appearance over time. Accurately modeling this complex transformation over an input facial image is extremely challenging as it requires making convincing, possibly large changes to facial features and head shape, while still preserving the input identity. In this work, we present an image-to-image translation method that learns to directly encode real facial images into the latent space of a pre-trained unconditional GAN (e.g., StyleGAN) subject to a given aging shift. We employ a pre-trained age regression network to explicitly guide the encoder in generating the latent codes corresponding to the desired age. In this formulation, our method approaches the continuous aging process as a regression task between the input age and desired target age, providing fine-grained control over the generated image. Moreover, unlike approaches that operate solely in the latent space using a prior on the path controlling age, our method learns a more disentangled, non-linear path. Finally, we demonstrate that the end-to-end nature of our approach, coupled with the rich semantic latent space of StyleGAN, allows for further editing of the generated images. Qualitative and quantitative evaluations show the advantages of our method compared to state-of-the-art approaches. | ['Daniel Cohen-Or', 'Or Patashnik', 'Yuval Alaluf'] | 2021-02-04 | null | null | null | null | ['face-age-editing'] | ['computer-vision'] | [ 5.37590861e-01 6.52647436e-01 3.03346794e-02 -6.67001069e-01
-3.21695387e-01 -4.32091802e-01 8.12241971e-01 -4.79638577e-01
-1.94898844e-01 5.63066304e-01 3.12539369e-01 2.86542803e-01
4.42792684e-01 -8.00726771e-01 -8.92359495e-01 -7.33771980e-01
3.44667077e-01 3.36398661e-01 -5.81830919e-01 -2.21066456e-02
-1.59708887e-01 4.54518020e-01 -1.51188409e+00 -6.13004081e-02
8.73310149e-01 9.97807622e-01 -3.75175595e-01 7.53745139e-01
3.44017804e-01 6.16827130e-01 -3.19610476e-01 -8.81270945e-01
3.38810235e-01 -7.84119248e-01 -5.34848928e-01 4.16803300e-01
9.36875403e-01 -5.63340247e-01 -4.20559078e-01 9.81507540e-01
3.00313860e-01 -1.77841067e-01 8.22451830e-01 -1.58216882e+00
-1.05058396e+00 2.90923774e-01 -6.61574304e-01 -5.79860687e-01
3.81375492e-01 1.52793124e-01 7.69022644e-01 -7.95963049e-01
6.52318835e-01 1.56156754e+00 4.63785142e-01 1.08726525e+00
-1.55830801e+00 -9.74879563e-01 7.14473799e-02 -1.91823363e-01
-1.21317244e+00 -7.28770435e-01 8.77398670e-01 -5.85917890e-01
8.04894045e-02 9.58898142e-02 8.50639582e-01 1.38312221e+00
2.08163545e-01 5.21759391e-01 1.23667824e+00 -5.83187342e-01
9.46484506e-02 -4.68225218e-02 -5.42955816e-01 1.05323851e+00
4.71334066e-03 3.16616029e-01 -7.57727265e-01 2.61874553e-02
9.37610269e-01 -8.67663175e-02 1.52867898e-01 -6.83983564e-01
-9.97128844e-01 6.76006854e-01 3.21547538e-01 -9.38545987e-02
-3.34847778e-01 6.67593122e-01 1.32561307e-02 2.99721092e-01
7.73661673e-01 2.90133357e-01 -3.00373107e-01 5.37904948e-02
-1.18510091e+00 3.40462834e-01 4.33690399e-01 8.05671155e-01
8.57887030e-01 2.23382294e-01 -2.54434019e-01 4.08105314e-01
2.55697310e-01 6.46569610e-01 5.05290739e-02 -1.36819625e+00
-2.86534447e-02 3.82286221e-01 1.39853349e-02 -6.58700705e-01
7.86272692e-04 -2.57099867e-01 -7.98999131e-01 7.56269038e-01
5.75695455e-01 -2.19945058e-01 -1.33105946e+00 2.62815380e+00
3.74968052e-01 4.72826660e-02 -6.60089105e-02 4.60559160e-01
9.70218107e-02 4.57786828e-01 2.50583827e-01 -2.91167974e-01
1.29694915e+00 -8.08513045e-01 -6.67188644e-01 -2.92303622e-01
6.98248819e-02 -5.59370816e-01 1.09516430e+00 1.26122475e-01
-1.50514543e+00 -4.81588513e-01 -8.45299184e-01 -1.99383304e-01
-3.53917554e-02 3.25373024e-01 4.83026236e-01 5.99575937e-01
-1.37579060e+00 5.43367803e-01 -9.44590032e-01 -5.96803546e-01
5.60655653e-01 5.45155227e-01 -5.91759264e-01 -5.09468354e-02
-1.01109123e+00 8.10741842e-01 -5.12627326e-02 -1.09647460e-01
-9.66970623e-01 -1.05873764e+00 -1.00116372e+00 -6.30795658e-02
1.91039100e-01 -1.24789703e+00 1.28203309e+00 -1.57773519e+00
-1.85747695e+00 1.31508315e+00 -1.61912575e-01 -1.62199080e-01
8.48377883e-01 -2.05955565e-01 -1.01428121e-01 9.83208418e-02
2.01774985e-02 1.35265112e+00 1.51170290e+00 -1.25529277e+00
-2.73985624e-01 -5.78910887e-01 1.63795367e-01 1.80157930e-01
-4.88010049e-01 8.56597275e-02 -4.63055968e-01 -8.71098876e-01
-2.33704045e-01 -1.21987736e+00 -5.94555959e-02 8.00538063e-01
-7.31247663e-02 1.26830518e-01 7.82395899e-01 -9.07080650e-01
9.10500944e-01 -2.11522961e+00 6.85312450e-01 5.50451167e-02
3.97700578e-01 -3.33887249e-01 -2.18025357e-01 2.79387414e-01
-2.29540735e-01 -7.30532408e-02 -2.40333617e-01 -8.00391257e-01
7.88322166e-02 1.38237163e-01 -1.37017578e-01 6.10310972e-01
3.24050009e-01 9.71953690e-01 -8.42600524e-01 -5.08839011e-01
-2.26854071e-01 7.65740335e-01 -8.33424330e-01 4.97395903e-01
-1.37034491e-01 7.77605236e-01 -5.15828244e-02 4.80143160e-01
4.55466241e-01 1.08500108e-01 4.99837026e-02 -3.95825535e-01
1.42967731e-01 -3.56938332e-01 -4.19479251e-01 1.97834694e+00
-6.44890666e-01 4.36390191e-01 1.14772953e-01 -4.66284603e-01
8.39052558e-01 3.12797964e-01 5.34492373e-01 -5.52860975e-01
1.40730739e-01 5.21528050e-02 -1.87160060e-01 -7.32909441e-02
2.53337413e-01 -4.59327817e-01 -1.82024762e-01 5.58251977e-01
2.84682363e-01 -2.63896137e-01 -2.00139781e-04 2.15243950e-01
7.66766131e-01 5.41968822e-01 5.27710095e-02 -1.46161094e-01
2.65250742e-01 -6.84867501e-01 5.48054278e-01 1.64104730e-01
-7.62879923e-02 7.12901533e-01 7.01357067e-01 -2.06836656e-01
-1.54155397e+00 -1.27131104e+00 2.48550296e-01 9.75885868e-01
-4.17042375e-01 -1.67404115e-01 -1.24690425e+00 -7.05848098e-01
-3.73172015e-02 6.03940248e-01 -1.23676944e+00 -5.91097713e-01
-5.53291798e-01 -2.77478755e-01 5.53566277e-01 4.49825823e-01
3.91830027e-01 -8.72472763e-01 -5.09006977e-01 -1.47072881e-01
-8.57300218e-03 -1.09843659e+00 -9.81337965e-01 -5.18134534e-01
-7.82051742e-01 -6.52710319e-01 -9.66670334e-01 -6.76452458e-01
1.42349958e+00 -4.44200873e-01 1.07168782e+00 -9.79880840e-02
-3.53697538e-01 6.51022196e-01 -2.28406172e-02 -3.53239596e-01
-6.26565039e-01 3.85210589e-02 1.86616346e-01 5.01211286e-01
-1.46411523e-01 -9.27674174e-01 -7.98954368e-01 -4.66267429e-02
-8.68418753e-01 6.11267507e-01 4.80678201e-01 8.54345918e-01
1.98964581e-01 -2.35287026e-01 3.48792881e-01 -8.60337377e-01
2.64936149e-01 -5.80624938e-02 -4.61420625e-01 2.19342247e-01
-8.05556595e-01 4.11819756e-01 3.25089395e-01 -6.16929233e-01
-1.22585642e+00 3.82850111e-01 1.35354459e-01 -5.82661450e-01
1.03727713e-01 -8.19644406e-02 -4.97863680e-01 1.45867094e-02
5.13670027e-01 4.61159311e-02 5.33117115e-01 -2.15267330e-01
8.59242380e-01 2.58575559e-01 8.93290281e-01 -7.50357389e-01
1.24027812e+00 6.40629768e-01 1.53072879e-01 -2.70683557e-01
-8.37046087e-01 4.85927373e-01 -9.14127648e-01 -3.97426784e-01
1.07605886e+00 -9.82514799e-01 -5.51707745e-01 9.19433832e-01
-1.02536988e+00 -6.31521106e-01 -6.25357926e-01 4.60542515e-02
-9.68282163e-01 -2.80723348e-02 -5.64202070e-01 -6.88792050e-01
-3.95224094e-01 -8.89939845e-01 1.35524845e+00 1.80377245e-01
-5.09714067e-01 -1.00802338e+00 7.84318801e-03 3.34613770e-01
3.28812122e-01 7.65398979e-01 1.00390995e+00 3.11627507e-01
-3.98838133e-01 -1.02723002e-01 -3.69579010e-02 3.74859154e-01
4.68701869e-01 3.12818319e-01 -9.70106781e-01 -5.21979153e-01
-1.66519791e-01 -4.45291221e-01 5.58177173e-01 2.29760706e-01
1.05754411e+00 -5.98647714e-01 -9.87350848e-03 8.29602778e-01
1.10527277e+00 -1.29592180e-01 7.57074773e-01 -1.43221095e-01
9.01067376e-01 7.92882681e-01 4.52508748e-01 3.78774732e-01
4.73674506e-01 7.03160882e-01 3.16577077e-01 -5.28267920e-01
-3.81013572e-01 -6.36299908e-01 6.88448906e-01 4.85394180e-01
-1.95929021e-01 1.30220518e-01 -4.99154538e-01 2.95328051e-01
-1.55067718e+00 -8.34185421e-01 5.39139569e-01 2.02634192e+00
1.26992762e+00 -9.18277502e-02 1.98339969e-01 -1.25652567e-01
5.80784321e-01 1.94120780e-01 -7.40993798e-01 -4.99381751e-01
5.33820651e-02 3.26417863e-01 4.09201741e-01 5.86348891e-01
-7.21467376e-01 1.05714059e+00 6.65514469e+00 4.84515399e-01
-1.30740225e+00 4.08596173e-02 9.65813994e-01 -2.70467728e-01
-4.92288142e-01 -8.50147381e-02 -3.38816375e-01 3.27611357e-01
8.27056468e-01 -4.20567423e-01 8.60918880e-01 5.33596873e-01
2.53713280e-01 2.00623408e-01 -1.50342345e+00 9.25845861e-01
2.85261691e-01 -9.90389526e-01 1.84179485e-01 3.58907610e-01
8.60503852e-01 -8.67195666e-01 6.35845840e-01 2.26643169e-03
4.37918812e-01 -1.14254630e+00 1.31627655e+00 7.33309209e-01
1.68387687e+00 -7.66928494e-01 2.07809955e-02 -5.31244464e-02
-8.48333001e-01 3.89947481e-02 1.16217561e-01 7.31308311e-02
3.26987579e-02 4.60484982e-01 -4.24539477e-01 1.35152340e-01
3.44368875e-01 5.86390018e-01 -6.68660939e-01 7.99156949e-02
-4.99256909e-01 3.66443455e-01 -1.12881158e-02 7.84514189e-01
-2.10822031e-01 -2.56910920e-01 2.00562686e-01 7.43203044e-01
6.74442410e-01 1.06651865e-01 -3.22466314e-01 1.02323425e+00
-4.92578536e-01 -1.83639020e-01 -5.90283453e-01 -2.65674621e-01
1.96589813e-01 1.26176548e+00 -4.13335174e-01 -2.56908983e-01
-1.44647285e-01 1.53372824e+00 4.17623281e-01 3.82785976e-01
-9.37002242e-01 -1.49100199e-02 1.02659357e+00 3.61473799e-01
8.60114247e-02 -2.83341229e-01 -2.78784186e-01 -1.12989497e+00
-1.14099821e-02 -1.08368921e+00 1.95055380e-02 -1.03108978e+00
-1.11766279e+00 5.13041794e-01 -1.86907221e-02 -8.31454396e-01
-4.68670428e-01 -2.75102139e-01 -6.79064095e-01 8.76616061e-01
-1.17406487e+00 -1.89298737e+00 -3.51917297e-01 6.07009709e-01
3.60505939e-01 -5.66746071e-02 8.66228521e-01 1.68369338e-01
-4.65083569e-01 1.13490200e+00 -3.70384902e-01 2.44698807e-04
1.07087064e+00 -1.20238197e+00 4.31903869e-01 1.02966893e+00
-1.97726607e-01 5.02729475e-01 9.45745707e-01 -5.88315129e-01
-1.27519345e+00 -1.15747893e+00 8.26041043e-01 -5.18728137e-01
4.96638417e-01 -6.69881284e-01 -5.24029434e-01 9.16193008e-01
4.36257988e-01 -1.87848117e-02 5.24878383e-01 -1.24403924e-01
-6.51088238e-01 -3.85258764e-01 -1.25499439e+00 8.73144507e-01
1.33881271e+00 -6.78912044e-01 -9.48609486e-02 -1.42052084e-01
8.32249761e-01 -4.31654572e-01 -8.57911825e-01 2.19452024e-01
9.47106838e-01 -7.14845777e-01 8.84303927e-01 -5.86473405e-01
8.53516221e-01 -7.17667863e-02 8.69576111e-02 -1.51857507e+00
-4.23595309e-01 -9.02076364e-01 -2.57561803e-01 1.49238086e+00
1.39348999e-01 -3.16969037e-01 1.04524815e+00 9.62433577e-01
2.03913704e-01 -6.77345932e-01 -6.44757807e-01 -4.17624027e-01
2.84373194e-01 4.22342159e-02 8.44211578e-01 7.09792733e-01
-3.57639730e-01 1.86596364e-01 -7.82429874e-01 1.14911869e-01
9.70293641e-01 -1.84571028e-01 8.63965392e-01 -1.07880378e+00
-2.64231950e-01 -2.46433392e-01 -4.93834406e-01 -5.89770615e-01
4.86844659e-01 -6.85066760e-01 -9.16276127e-02 -1.10982394e+00
2.65916437e-01 -3.38719487e-01 -5.99404015e-02 6.31757200e-01
-2.65104651e-01 5.42122483e-01 2.88507670e-01 1.80571955e-02
-7.77594820e-02 8.24359119e-01 1.51109552e+00 -9.55142677e-02
8.52142647e-02 -2.27922857e-01 -9.36076641e-01 5.47956467e-01
6.36078179e-01 -4.64198023e-01 -5.73220432e-01 -5.07044077e-01
2.94036835e-01 -7.38954768e-02 3.16640824e-01 -7.47132540e-01
-2.21469048e-02 -3.83202642e-01 5.66026032e-01 1.40758157e-01
5.31963289e-01 -6.80578589e-01 4.36795235e-01 5.98375797e-01
-4.94877666e-01 1.43276557e-01 -3.34794074e-02 2.63881087e-01
-6.03310056e-02 2.98459500e-01 1.10646284e+00 1.35789663e-01
-1.36110067e-01 7.57542610e-01 -7.18435794e-02 -9.33433697e-02
1.08989286e+00 -2.45807633e-01 7.17332587e-02 -7.29455113e-01
-8.63124728e-01 -1.83272697e-02 1.07380199e+00 4.50298786e-01
3.46376181e-01 -1.75831783e+00 -8.62222493e-01 4.22894984e-01
1.01205520e-01 -2.02412635e-01 1.89289138e-01 5.71988583e-01
-2.77245730e-01 -1.72450483e-01 -6.72438204e-01 -3.67567062e-01
-1.53209496e+00 4.27050829e-01 3.61721337e-01 -2.40137264e-01
-2.99091965e-01 7.47327149e-01 6.15475297e-01 -3.22030544e-01
7.27448389e-02 -4.64440286e-02 1.91423610e-01 7.59805664e-02
2.96152502e-01 8.42669755e-02 -3.90481681e-01 -7.16212273e-01
-1.77074261e-02 7.82525241e-01 5.69874346e-02 -5.80539525e-01
1.37804198e+00 -2.95261949e-01 -2.91041613e-01 2.18998864e-01
1.19514489e+00 8.11491609e-02 -2.03335786e+00 -1.58784628e-01
-5.67419350e-01 -5.53975761e-01 -1.68731093e-01 -7.25582123e-01
-1.44888425e+00 8.40343058e-01 7.05299437e-01 -3.81132841e-01
1.32291281e+00 -2.61601191e-02 6.65923357e-01 -3.29113394e-01
2.76064664e-01 -9.45471048e-01 4.93434995e-01 3.77899059e-03
1.05859470e+00 -9.84998167e-01 -4.98671047e-02 -3.62843663e-01
-5.67778885e-01 1.01461053e+00 7.13237107e-01 -1.06243473e-02
4.32313800e-01 1.70995891e-01 1.30472332e-01 3.63924392e-02
-7.43700683e-01 2.30059534e-01 2.80388027e-01 6.42836154e-01
3.67139906e-01 3.49012986e-02 -3.92517000e-02 9.78100076e-02
-6.88238621e-01 2.62197286e-01 5.31421423e-01 6.89517498e-01
1.25919029e-01 -1.47571266e+00 -3.18042219e-01 1.62196442e-01
-4.42031324e-01 1.06430151e-01 -2.98499107e-01 4.68935847e-01
4.55322742e-01 4.18879956e-01 2.27471232e-01 -2.40861371e-01
1.72363847e-01 2.06465140e-01 1.08106637e+00 -6.25445187e-01
-1.06205828e-01 -1.84436664e-01 -1.74742460e-01 -6.25200391e-01
-3.61228466e-01 -8.96556914e-01 -9.03220356e-01 -4.78589058e-01
2.10045323e-01 -2.37298369e-01 6.66899800e-01 7.22626269e-01
2.29412526e-01 4.77198154e-01 8.78307045e-01 -9.83508050e-01
-3.67209703e-01 -6.19425833e-01 -4.54180926e-01 7.02347815e-01
3.71790946e-01 -5.41753292e-01 -1.33057564e-01 6.79952264e-01] | [12.507688522338867, -0.24584843218326569] |
3c08d016-016e-42ed-ae7b-38a98b622d05 | face-morphing-attack-detection-with-denoising-1 | 2306.15733 | null | https://arxiv.org/abs/2306.15733v1 | https://arxiv.org/pdf/2306.15733v1.pdf | Face Morphing Attack Detection with Denoising Diffusion Probabilistic Models | Morphed face images have recently become a growing concern for existing face verification systems, as they are relatively easy to generate and can be used to impersonate someone's identity for various malicious purposes. Efficient Morphing Attack Detection (MAD) that generalizes well across different morphing techniques is, therefore, of paramount importance. Existing MAD techniques predominantly rely on discriminative models that learn from examples of bona fide and morphed images and, as a result, often exhibit sub-optimal generalization performance when confronted with unknown types of morphing attacks. To address this problem, we propose a novel, diffusion-based MAD method in this paper that learns only from the characteristics of bona fide images. Various forms of morphing attacks are then detected by our model as out-of-distribution samples. We perform rigorous experiments over four different datasets (CASIA-WebFace, FRLL-Morphs, FERET-Morphs and FRGC-Morphs) and compare the proposed solution to both discriminatively-trained and once-class MAD models. The experimental results show that our MAD model achieves highly competitive results on all considered datasets. | ['Vitomir Štruc', 'Marija Ivanovska'] | 2023-06-27 | face-morphing-attack-detection-with-denoising | https://ieeexplore.ieee.org/document/10156877 | https://lmi.fe.uni-lj.si/wp-content/uploads/2023/06/IWBF2023___Face_Morphing_Attack_Detection_with_Denoising_Diffusion_Probabilistic_Models.pdf | international-workshop-on-biometrics-and | ['face-verification'] | ['computer-vision'] | [ 5.14258407e-02 -4.62910712e-01 -8.83542672e-02 -4.87767130e-01
-7.68653572e-01 -6.99058235e-01 8.97536278e-01 6.40860498e-02
-2.33707547e-01 6.14747047e-01 -4.83021408e-01 -2.22828329e-01
6.11697547e-02 -7.06210554e-01 -5.35961747e-01 -7.42667079e-01
-3.33430767e-01 5.86937070e-01 1.97614357e-01 -2.27749512e-01
4.37952906e-01 9.47307527e-01 -1.35151947e+00 1.49153084e-01
7.34993219e-01 1.08517909e+00 -4.65377510e-01 7.59443402e-01
2.11384758e-01 4.29476917e-01 -8.93719554e-01 -9.47910964e-01
6.43371105e-01 -6.19486928e-01 -6.22730196e-01 1.48510948e-01
9.26121593e-01 -1.90234289e-01 -3.22589368e-01 1.51504719e+00
5.20304441e-01 -2.09630951e-01 8.51909697e-01 -1.46991706e+00
-6.84981942e-01 4.17576842e-02 -8.46228719e-01 3.17679733e-01
2.38432676e-01 3.75345230e-01 6.23640120e-01 -7.53754318e-01
5.43407440e-01 1.41569698e+00 4.54095215e-01 9.66689348e-01
-1.26039600e+00 -1.04090500e+00 -2.54349917e-01 3.83680820e-01
-1.43368268e+00 -6.21717691e-01 9.43075657e-01 -4.16720361e-01
9.68073532e-02 4.58948076e-01 2.24066213e-01 1.16301775e+00
2.37904608e-01 6.91408098e-01 1.52570748e+00 -1.29499003e-01
1.73196569e-01 5.58852434e-01 -2.49227151e-01 7.68216431e-01
3.92488629e-01 1.83561757e-01 -4.42465901e-01 -3.99584830e-01
4.02758062e-01 -2.02767298e-01 -4.03148681e-01 -3.52226913e-01
-3.20887506e-01 8.79858851e-01 2.72129029e-01 3.81334722e-01
-9.06279087e-02 -3.10024232e-01 2.57338524e-01 4.47126180e-01
5.61609924e-01 2.30267480e-01 -2.65815765e-01 8.42718184e-02
-1.12185657e+00 2.72982180e-01 6.69511139e-01 4.59057420e-01
9.04235065e-01 1.96965933e-01 1.73197031e-01 1.01842666e+00
2.69492418e-01 6.29918098e-01 5.62974036e-01 -2.28009596e-01
3.17559779e-01 6.01774037e-01 -1.50269881e-01 -1.34633040e+00
1.02541767e-01 -2.36891836e-01 -8.71159852e-01 4.48436707e-01
4.60853279e-01 -6.90989057e-03 -1.01733959e+00 1.65151119e+00
4.73943174e-01 3.75910163e-01 -7.36542568e-02 5.61839402e-01
3.01733583e-01 4.91504639e-01 9.61178094e-02 -1.96946412e-01
1.21982300e+00 -5.50906122e-01 -3.94930899e-01 -2.08997995e-01
4.40059453e-01 -1.11122072e+00 7.32270062e-01 5.06012499e-01
-5.96460342e-01 -4.57552880e-01 -1.05211389e+00 5.93915522e-01
-6.29000127e-01 1.84673052e-02 3.56692135e-01 1.44277442e+00
-1.05316794e+00 6.88883424e-01 -4.07558769e-01 -2.63040572e-01
8.79299521e-01 6.83366954e-01 -7.33652472e-01 -3.46766680e-01
-8.62803936e-01 7.52325535e-01 1.68908715e-01 9.39943641e-02
-1.27665818e+00 -4.60379332e-01 -7.18488932e-01 -3.61000240e-01
-3.97285782e-02 7.26790428e-02 8.44046831e-01 -1.19014060e+00
-1.23467016e+00 1.29464734e+00 2.04119701e-02 -5.03800869e-01
6.40487731e-01 1.33430824e-01 -9.07566071e-01 3.51934642e-01
-1.92386910e-01 4.62812543e-01 1.60285425e+00 -1.47278023e+00
-2.86688209e-01 -7.16663301e-01 -1.45280749e-01 -4.28997427e-01
-5.78273296e-01 1.64486945e-01 -2.25962489e-03 -7.97115564e-01
-1.72006056e-01 -1.05017745e+00 6.95079267e-02 3.09478883e-02
-6.63279355e-01 -5.11305854e-02 1.47288823e+00 -8.43236148e-01
1.00064468e+00 -1.99156225e+00 8.56834650e-02 3.68804365e-01
-1.29948974e-01 9.26652789e-01 -1.55214429e-01 2.13368982e-01
-2.23484412e-01 1.31024137e-01 -5.18253505e-01 -4.62580562e-01
-2.10319027e-01 2.10353807e-01 -8.93766582e-02 8.83159697e-01
3.91479105e-01 5.29550612e-01 -5.38405359e-01 -5.99497676e-01
8.39175954e-02 4.43521053e-01 -4.70537007e-01 4.05282646e-01
-2.60373503e-02 6.40343845e-01 -1.71329319e-01 1.08214486e+00
1.29908204e+00 2.89191186e-01 5.27540445e-02 -1.52057841e-01
4.48932201e-01 -4.95441705e-01 -1.17194533e+00 8.96444321e-01
-2.10167781e-01 5.40755153e-01 2.17368156e-01 -9.75659132e-01
9.39604700e-01 1.23908781e-01 1.00941598e-01 -3.37954044e-01
4.80565250e-01 4.29839462e-01 1.62593648e-01 -4.06099677e-01
2.61628151e-01 -3.78455848e-01 1.03343733e-01 4.18419451e-01
4.39041436e-01 9.43165366e-03 1.57166317e-01 2.09748070e-03
8.53598177e-01 -4.63272452e-01 1.82686120e-01 -2.43747622e-01
1.05004168e+00 -5.46103537e-01 4.70087498e-01 3.43228281e-01
-5.98498940e-01 3.91640335e-01 5.53561151e-01 -4.52226520e-01
-6.76693976e-01 -8.50407720e-01 -1.59323826e-01 6.13634288e-01
9.76631716e-02 -5.26989363e-02 -1.11204731e+00 -1.26947308e+00
1.14238791e-01 3.00641984e-01 -7.90730417e-01 -2.72850066e-01
-5.89890242e-01 -9.92234111e-01 8.47215235e-01 -1.38094559e-01
6.47528052e-01 -8.47994566e-01 -2.72037461e-02 -1.27075538e-01
3.30819279e-01 -1.10197186e+00 -4.59253639e-01 -3.26587558e-01
-5.31026065e-01 -1.48950934e+00 -8.48571062e-01 -7.62483120e-01
8.89901936e-01 1.16942979e-01 6.29744768e-01 2.99867064e-01
-7.58675694e-01 3.51755708e-01 -1.18615583e-01 -1.43149376e-01
-8.19566488e-01 -1.77524060e-01 2.38939688e-01 1.05652392e+00
3.77842337e-01 -5.40484428e-01 -4.72102642e-01 5.51225901e-01
-1.16137099e+00 -7.96917915e-01 6.24780297e-01 8.23366523e-01
4.69412580e-02 1.93532377e-01 4.36723709e-01 -1.25711703e+00
4.63977844e-01 -5.57078362e-01 -7.50026822e-01 2.69880414e-01
-5.15151620e-01 -5.64926155e-02 7.18976438e-01 -7.91137934e-01
-8.80444169e-01 7.08395615e-02 -4.13472831e-01 -6.35077655e-01
-3.33328754e-01 -1.88831046e-01 -5.16965747e-01 -9.50949073e-01
6.42501950e-01 4.44567055e-01 1.15552038e-01 -5.54090440e-01
4.99929190e-02 7.98840165e-01 4.23913628e-01 -5.77739000e-01
1.59889197e+00 5.09248197e-01 2.60297626e-01 -1.01655293e+00
-2.22520411e-01 -3.16253126e-01 -3.32535982e-01 -4.31919128e-01
4.94085342e-01 -4.58904117e-01 -6.54582560e-01 1.19360769e+00
-9.62724566e-01 2.15562712e-02 2.09546909e-01 1.29211217e-01
-1.10001013e-01 8.09086144e-01 -6.97910309e-01 -9.95392919e-01
-1.16841279e-01 -1.37878978e+00 9.38470721e-01 4.74677503e-01
4.94363382e-02 -1.08375037e+00 7.48578012e-02 3.75048429e-01
6.11014605e-01 5.42376101e-01 8.84271383e-01 -8.90819252e-01
-5.17272592e-01 -7.00182259e-01 -9.35750082e-02 7.80052960e-01
4.67160851e-01 1.67476609e-01 -1.17657995e+00 -5.01139462e-01
1.57268971e-01 -4.29767758e-01 6.48391545e-01 -2.34769687e-01
1.12204945e+00 -6.42921507e-01 -1.21652164e-01 5.46031356e-01
1.48784006e+00 2.28937820e-01 6.81803942e-01 -1.22072645e-01
5.48322201e-01 7.10111916e-01 5.00317454e-01 4.15205866e-01
-9.03963298e-02 7.27031291e-01 6.96737051e-01 1.80509686e-01
-1.05695672e-01 -2.69453764e-01 4.90176439e-01 4.56116587e-01
1.55890957e-01 -2.01801494e-01 -6.55090570e-01 4.91475552e-01
-1.20645499e+00 -1.03618491e+00 7.24638253e-02 2.27709508e+00
7.76986003e-01 4.31570075e-02 2.60854393e-01 4.22082543e-01
9.75118935e-01 3.40203315e-01 -4.13134784e-01 -4.49022561e-01
-2.70527840e-01 6.54495955e-01 5.06119251e-01 5.86579323e-01
-1.34568167e+00 1.03961861e+00 5.27949142e+00 1.29046726e+00
-1.40914059e+00 2.06984386e-01 9.59357262e-01 4.10531908e-01
1.04773171e-01 -1.61959291e-01 -8.76138806e-01 6.70698583e-01
8.55723321e-01 6.24548867e-02 3.80103022e-01 9.13604975e-01
-4.22629684e-01 -2.52265227e-03 -9.71382678e-01 1.17759538e+00
5.02240658e-01 -1.05220950e+00 1.86573133e-01 2.95455337e-01
7.75090575e-01 -6.42353296e-01 6.54632151e-01 8.65913704e-02
1.18440814e-01 -1.04933703e+00 3.31315875e-01 -1.53883666e-01
7.43509769e-01 -8.48308980e-01 6.73936784e-01 1.57784134e-01
-9.87569153e-01 -5.97058199e-02 -4.60368961e-01 5.55620670e-01
-1.73534349e-01 3.13326389e-01 -9.14300680e-01 5.55622041e-01
2.81684786e-01 4.73700136e-01 -1.03915250e+00 9.30064201e-01
-1.30237505e-01 6.86962962e-01 -3.00655901e-01 6.24859147e-02
1.39178291e-01 4.90979627e-02 5.67874610e-01 1.10211349e+00
3.24074060e-01 -3.81811798e-01 -1.71066090e-01 4.98080879e-01
-3.80783200e-01 2.75610328e-01 -7.14196026e-01 -2.16352686e-01
2.18754888e-01 1.46190679e+00 -6.60042346e-01 -8.59761760e-02
-8.55431259e-02 1.13847101e+00 6.80259466e-02 6.86702430e-02
-6.91696525e-01 -2.66446203e-01 9.62672234e-01 1.75466433e-01
2.43068248e-01 -6.14859387e-02 2.46209174e-01 -1.15184987e+00
-1.02112435e-01 -1.33468509e+00 5.64377189e-01 -7.15630800e-02
-1.40872681e+00 8.54781568e-01 -2.59967923e-01 -1.22611403e+00
-2.40971565e-01 -8.97235572e-01 -5.37153184e-01 6.93071842e-01
-1.71321619e+00 -1.18977070e+00 -1.29322410e-01 9.10791695e-01
4.65722293e-01 -6.84030652e-01 7.52734303e-01 7.54480839e-01
-6.66532576e-01 1.10899055e+00 -2.81400800e-01 5.20656407e-01
8.28991532e-01 -1.01810873e+00 1.72259018e-01 1.07715416e+00
4.00000066e-01 4.76471126e-01 6.79663002e-01 -5.04316986e-01
-1.45314407e+00 -1.12793112e+00 5.69647610e-01 -2.21661553e-01
4.53646421e-01 -3.65593761e-01 -9.06719744e-01 3.79166901e-01
1.01367943e-01 4.34127867e-01 6.80225074e-01 -4.12181526e-01
-8.23026478e-01 -3.71153861e-01 -1.84689736e+00 4.40359503e-01
6.83489501e-01 -6.16639137e-01 -5.39719015e-02 3.94493937e-01
-1.03646517e-01 4.37050872e-03 -4.91796255e-01 3.62049907e-01
3.94183308e-01 -1.28945220e+00 8.58527422e-01 -6.91755950e-01
-4.68204878e-02 -3.43379706e-01 -2.27519810e-01 -1.17648995e+00
3.06333542e-01 -9.12961900e-01 -1.39484823e-01 1.46025109e+00
1.05001807e-01 -7.35512078e-01 1.07559621e+00 2.45889589e-01
5.95941782e-01 -5.03181756e-01 -1.11201370e+00 -1.03873193e+00
1.08895088e-02 -9.81268566e-03 6.51564062e-01 1.19072533e+00
-4.40629005e-01 -2.66754776e-02 -7.41113305e-01 3.93901408e-01
1.11982679e+00 -1.50355622e-01 9.47678626e-01 -1.22458172e+00
-4.14262027e-01 -3.43954176e-01 -1.17592990e+00 -6.60320878e-01
4.37055260e-01 -6.94238126e-01 -2.17814967e-01 -3.24047148e-01
4.89978381e-02 -5.33247054e-01 -1.56762347e-01 3.14653337e-01
-1.39292866e-01 8.98663878e-01 3.39615755e-02 6.29629046e-02
6.56469986e-02 2.93192357e-01 9.97442722e-01 -3.96658331e-01
3.28348994e-01 2.81945437e-01 -4.43555772e-01 5.39227068e-01
8.56819034e-01 -7.05298007e-01 -2.89265126e-01 1.67097718e-01
-3.32674026e-01 -9.67531279e-02 3.13847482e-01 -1.17494380e+00
3.96848619e-02 -1.35312051e-01 3.06344777e-01 -1.55827418e-01
5.07449090e-01 -8.04975748e-01 7.31740817e-02 6.06521964e-01
1.64749864e-02 5.04139364e-02 1.63258478e-01 4.82349724e-01
-3.56224746e-01 -3.01217407e-01 1.65277994e+00 2.52452008e-02
-4.63869423e-01 8.06782782e-01 -3.41373160e-02 1.06290728e-01
1.35658038e+00 -1.90011710e-01 -1.99193791e-01 -4.06510651e-01
-5.27682900e-01 -4.26364988e-01 5.72923243e-01 4.85630065e-01
7.21812725e-01 -1.30468714e+00 -7.75873423e-01 5.59188247e-01
2.76099867e-03 -6.69548512e-01 3.18059713e-01 4.51463252e-01
-6.86291337e-01 4.05891649e-02 -4.87925708e-01 -5.29786587e-01
-1.80448830e+00 7.49839485e-01 3.82459015e-01 -1.83278531e-01
6.70794696e-02 1.08533907e+00 -1.82834193e-02 -2.90806621e-01
-3.34708728e-02 3.45299125e-01 4.38725285e-04 5.56489527e-02
6.75214350e-01 1.82990685e-01 1.39630467e-01 -1.33082855e+00
-3.57648104e-01 6.96746171e-01 -3.10754806e-01 3.21476534e-02
9.92878616e-01 3.01045358e-01 -3.13006461e-01 -8.70780274e-02
1.57348716e+00 3.79957139e-01 -8.78092289e-01 -5.52443191e-02
1.90020457e-01 -1.05640340e+00 -5.06781638e-01 -3.89071912e-01
-1.39820230e+00 1.09403932e+00 9.69057977e-01 3.18339497e-01
1.06146479e+00 -2.35918671e-01 7.76363611e-01 -1.81958094e-01
4.60959882e-01 -6.51087880e-01 3.34067732e-01 -2.26252273e-01
5.45635402e-01 -1.23806083e+00 -1.58838019e-01 -4.83113229e-01
-3.19798261e-01 1.02938771e+00 7.02005148e-01 -1.94955409e-01
8.64950061e-01 6.85495287e-02 2.02073306e-01 7.06006587e-02
-2.03581274e-01 7.89053589e-02 2.17709139e-01 7.85001993e-01
-1.10268697e-01 -6.45287931e-02 -1.07842349e-01 -5.55425622e-02
-2.64907386e-02 -5.37469745e-01 3.47960114e-01 9.93888974e-01
-1.95229664e-01 -1.71282089e+00 -6.17264509e-01 2.04815939e-01
-7.33768165e-01 3.31474602e-01 -7.43114114e-01 8.13740134e-01
1.96406528e-01 9.08210933e-01 -5.02347648e-01 -6.17394805e-01
-1.52446469e-02 1.11249104e-01 7.35599577e-01 -4.05879289e-01
-6.08317614e-01 -3.33793968e-01 -2.32278079e-01 -3.55390340e-01
-3.19811583e-01 -6.24554098e-01 -3.81659895e-01 -5.85156322e-01
-2.76093572e-01 2.78146923e-01 6.62210822e-01 6.98126614e-01
1.77440628e-01 -2.15561286e-01 1.21823025e+00 -7.48834789e-01
-7.78064430e-01 -8.97288561e-01 -7.42408097e-01 6.07625186e-01
4.65218633e-01 -7.39779234e-01 -5.10403335e-01 -7.01820850e-02] | [13.014703750610352, 1.127640962600708] |
c307c6ab-dce0-467b-9811-b85000869022 | precursor-of-anomaly-detection-for-irregular | 2306.15489 | null | https://arxiv.org/abs/2306.15489v2 | https://arxiv.org/pdf/2306.15489v2.pdf | Precursor-of-Anomaly Detection for Irregular Time Series | Anomaly detection is an important field that aims to identify unexpected patterns or data points, and it is closely related to many real-world problems, particularly to applications in finance, manufacturing, cyber security, and so on. While anomaly detection has been studied extensively in various fields, detecting future anomalies before they occur remains an unexplored territory. In this paper, we present a novel type of anomaly detection, called \emph{\textbf{P}recursor-of-\textbf{A}nomaly} (PoA) detection. Unlike conventional anomaly detection, which focuses on determining whether a given time series observation is an anomaly or not, PoA detection aims to detect future anomalies before they happen. To solve both problems at the same time, we present a neural controlled differential equation-based neural network and its multi-task learning algorithm. We conduct experiments using 17 baselines and 3 datasets, including regular and irregular time series, and demonstrate that our presented method outperforms the baselines in almost all cases. Our ablation studies also indicate that the multitasking training method significantly enhances the overall performance for both anomaly and PoA detection. | ['Noseong Park', 'Jaehoon Lee', 'Sheo Yon Jhin'] | 2023-06-27 | null | null | null | null | ['anomaly-detection', 'multi-task-learning', 'irregular-time-series'] | ['methodology', 'methodology', 'time-series'] | [ 2.64726430e-01 -5.25928199e-01 2.47151643e-01 -2.12059692e-01
-3.58908266e-01 -3.13349396e-01 6.81036055e-01 7.32325256e-01
-1.37658969e-01 4.81321394e-01 -3.10616016e-01 -7.78754711e-01
-3.54816347e-01 -5.79980493e-01 -5.12806714e-01 -8.00447285e-01
-6.07897580e-01 3.43559235e-01 2.63558507e-01 -3.28486919e-01
4.80151445e-01 5.32725155e-01 -1.34332395e+00 4.95470278e-02
9.40355122e-01 1.44468343e+00 -4.51766133e-01 2.78786689e-01
-3.37948799e-01 6.56589150e-01 -8.45666230e-01 -3.83289270e-02
2.70617723e-01 -2.30712280e-01 -6.30804598e-01 -1.07010761e-02
4.76363152e-02 -7.36103356e-02 1.67918410e-02 9.16307807e-01
1.48923159e-01 3.50850791e-01 7.05127478e-01 -1.69676781e+00
-4.32180732e-01 1.75870910e-01 -9.38527226e-01 9.52173293e-01
2.69070119e-01 1.78536832e-01 7.30000913e-01 -8.73066902e-01
-1.69056952e-01 9.20261443e-01 6.82637095e-01 1.98069304e-01
-1.06356192e+00 -7.48086810e-01 6.49484992e-01 3.13562423e-01
-1.01024210e+00 -1.14323840e-01 9.94345069e-01 -5.49651980e-01
1.03973687e+00 1.87687114e-01 2.87382096e-01 1.30547023e+00
8.16704154e-01 7.06811905e-01 8.04859042e-01 -1.79312035e-01
3.07569742e-01 -5.46403110e-01 3.04747880e-01 1.89016297e-01
2.04133436e-01 2.43939832e-03 -2.80289739e-01 -4.06215400e-01
3.59879434e-01 4.97544289e-01 9.65828076e-02 1.06974900e-01
-1.32065880e+00 6.99756503e-01 -2.22375318e-01 5.24291039e-01
-6.66845500e-01 -7.75653943e-02 7.37514317e-01 9.00966108e-01
8.04793537e-01 5.79292297e-01 -6.01719022e-01 -2.68606693e-01
-4.92484272e-01 3.75247985e-01 7.28411615e-01 4.91020381e-01
4.95224386e-01 7.34024465e-01 -2.76250750e-01 6.00960612e-01
-1.42584920e-01 2.14237019e-01 7.32422829e-01 -3.83658946e-01
3.49957556e-01 5.95146000e-01 9.84932631e-02 -1.13633537e+00
-7.35184908e-01 -4.15968090e-01 -1.13842750e+00 2.17497081e-01
6.76949799e-01 -3.72654140e-01 -8.26297402e-01 1.34817719e+00
1.90807238e-01 6.43920481e-01 -3.82618494e-02 4.71906751e-01
-8.17796588e-03 8.35356712e-01 -8.02823678e-02 -4.64305907e-01
9.62238550e-01 -6.57007217e-01 -1.02190530e+00 -3.04369479e-01
6.53145075e-01 -9.14374292e-01 7.08263874e-01 7.43958771e-01
-6.00168526e-01 -4.12992299e-01 -7.49476433e-01 6.24406159e-01
-6.05393231e-01 -4.47785527e-01 5.93324721e-01 1.75675899e-01
-6.18990600e-01 6.52219832e-01 -9.76261199e-01 -3.12891424e-01
2.55422503e-01 1.16103582e-01 -1.17280595e-01 4.16009992e-01
-1.15164459e+00 7.31872320e-01 6.03346586e-01 1.92896366e-01
-7.36428440e-01 -6.29256427e-01 -7.83463538e-01 -8.47901553e-02
7.78248966e-01 -1.17176935e-01 1.29680264e+00 -7.12272286e-01
-1.00989485e+00 4.91688013e-01 -2.90184885e-01 -8.11110020e-01
3.42703044e-01 -5.57028890e-01 -1.20660603e+00 -5.74865341e-01
1.05203710e-01 -4.77013618e-01 1.26339138e+00 -6.68972492e-01
-9.15526032e-01 -3.78510803e-01 -3.72920692e-01 -3.56130928e-01
-3.76482964e-01 1.43048421e-01 3.53039145e-01 -1.09136939e+00
2.81196177e-01 -6.32351220e-01 -3.13456029e-01 -3.18020493e-01
-5.53811193e-01 -7.65868187e-01 1.50952220e+00 -4.63494748e-01
1.49409449e+00 -2.22300148e+00 -3.39226305e-01 4.95686650e-01
1.15982391e-01 3.40590388e-01 7.49343187e-02 5.93172252e-01
-5.75413585e-01 -5.10326102e-02 -4.52611446e-01 -1.22010507e-01
-2.77430594e-01 3.08986098e-01 -8.52973104e-01 4.82187063e-01
5.98908424e-01 5.15038311e-01 -7.64515817e-01 1.16761848e-01
3.78731161e-01 -3.07572275e-01 -1.44250020e-01 1.21630505e-01
-3.19366127e-01 7.01097727e-01 -6.49381936e-01 8.62283349e-01
4.86740977e-01 -1.26666620e-01 -5.29593408e-01 4.40230697e-01
-2.04740122e-01 -1.98760152e-01 -1.14545131e+00 1.35899448e+00
-3.47735822e-01 5.09576201e-01 -3.45938236e-01 -1.78053355e+00
1.27083361e+00 4.55892026e-01 9.41020131e-01 -9.15456235e-01
-1.21031307e-01 5.01582563e-01 3.39735687e-01 -6.53352678e-01
3.09651226e-01 1.55899927e-01 -2.37423465e-01 5.68775117e-01
-3.37145120e-01 3.56158376e-01 2.67219454e-01 -2.11173609e-01
1.44285893e+00 -8.56272355e-02 4.12476987e-01 -2.70586818e-01
8.44723642e-01 -1.22964360e-01 7.76166558e-01 7.77899146e-01
-3.03238988e-01 2.73689806e-01 7.07040250e-01 -9.55125511e-01
-8.77362967e-01 -1.06303298e+00 -3.05833310e-01 1.00722885e+00
3.76204099e-03 -1.35702908e-01 1.65632286e-03 -6.99099243e-01
3.30272377e-01 9.97197807e-01 -5.63709378e-01 -3.54584157e-01
-7.97793508e-01 -9.66872096e-01 3.53239477e-01 3.93939197e-01
4.22786534e-01 -1.46243131e+00 -3.97888094e-01 5.20159662e-01
-4.82739881e-02 -1.08924937e+00 -3.84851843e-01 2.82152325e-01
-8.60372841e-01 -1.12177622e+00 -4.17520195e-01 -4.38292533e-01
3.47172022e-01 3.39368209e-02 9.53825116e-01 -2.24485144e-01
-3.31498891e-01 3.29899102e-01 -4.63207901e-01 -1.16245639e+00
-4.04799491e-01 -1.07525051e-01 4.46703643e-01 4.48209018e-01
7.73851812e-01 -8.51041973e-01 -2.77706325e-01 4.31198657e-01
-1.00703287e+00 -6.14490688e-01 4.34759855e-01 6.53102577e-01
5.56142032e-01 3.73007268e-01 9.70734239e-01 -7.42004335e-01
9.53316331e-01 -8.70184720e-01 -6.42980576e-01 7.17433915e-02
-8.42218220e-01 -4.52368520e-02 9.55994129e-01 -4.94599879e-01
-6.92567885e-01 -2.28574440e-01 -1.23204604e-01 -6.05580330e-01
-6.58240914e-01 6.73682034e-01 2.99782425e-01 3.16281468e-01
5.74807048e-01 5.03120542e-01 6.91992566e-02 -5.50058305e-01
-4.11206961e-01 4.12325829e-01 5.61172485e-01 -5.92267513e-01
9.53293025e-01 5.10426939e-01 2.15096354e-01 -1.02531016e+00
-8.65482152e-01 -7.37669051e-01 -4.24971074e-01 -1.88942224e-01
5.75832248e-01 -5.65878153e-01 -6.75211906e-01 8.38237643e-01
-1.15412784e+00 -2.21377220e-02 -4.71734881e-01 3.43218803e-01
-3.65818709e-01 3.94269735e-01 -1.25365674e-01 -1.09851897e+00
-3.15009266e-01 -7.91589200e-01 9.95034277e-01 -3.27750742e-02
-3.78413111e-01 -1.18035090e+00 1.33425727e-01 -2.89761096e-01
6.19030297e-01 7.77339518e-01 9.08474743e-01 -1.55424488e+00
-1.80421442e-01 -5.81672609e-01 1.18345857e-01 3.82432908e-01
6.11164391e-01 -8.96072388e-03 -6.48767114e-01 -3.27803642e-01
3.52717102e-01 1.55500367e-01 7.00642705e-01 2.51263887e-01
1.75263691e+00 -1.94465011e-01 -2.13378057e-01 3.38408917e-01
8.65957916e-01 8.25227559e-01 4.56757754e-01 5.58931172e-01
5.59900105e-01 4.03513134e-01 6.49594605e-01 8.22924674e-01
-2.04216503e-02 5.35814285e-01 7.10549116e-01 -2.34018080e-02
6.65115118e-01 3.43084544e-01 5.10156035e-01 7.16231763e-01
-1.72345906e-01 -2.10554168e-01 -1.18545401e+00 6.45250142e-01
-2.03590012e+00 -1.07818687e+00 -4.09382761e-01 2.17310476e+00
3.09166342e-01 4.14454997e-01 2.65331447e-01 5.59444368e-01
6.75300539e-01 2.57983088e-01 -9.11002159e-01 -6.75480008e-01
-1.42092437e-01 1.81699038e-01 5.61700901e-03 -3.73209827e-02
-1.52048051e+00 5.37374079e-01 5.49737740e+00 5.86706221e-01
-1.12495089e+00 -1.57024235e-01 5.58670282e-01 9.57724079e-02
1.80072427e-01 -2.69662231e-01 -4.51852053e-01 6.38827801e-01
8.14617455e-01 -5.00038803e-01 1.41997933e-01 8.29826176e-01
4.17237818e-01 2.75277756e-02 -1.21125591e+00 9.18603301e-01
-9.45027396e-02 -7.23103523e-01 9.45224315e-02 1.38468698e-01
7.11589754e-01 -9.82514769e-02 1.99760690e-01 5.83661079e-01
-1.09276451e-01 -8.55064154e-01 3.34463298e-01 6.50577128e-01
1.08330354e-01 -7.28698015e-01 9.29709435e-01 3.59775126e-01
-1.18346083e+00 -3.48352641e-01 2.87066251e-02 -4.14493650e-01
2.13672742e-01 1.06777632e+00 -3.50704074e-01 7.86687434e-01
8.51138949e-01 9.50260460e-01 -3.79780084e-01 1.07857835e+00
1.44403368e-01 6.78739071e-01 -3.19346488e-01 2.08503619e-01
4.15015101e-01 -3.37373018e-01 1.06632137e+00 8.45321000e-01
6.43614888e-01 -2.66192853e-01 5.57796240e-01 6.33814156e-01
3.71671826e-01 1.18927017e-01 -9.75202858e-01 -4.69715036e-02
9.02545750e-02 8.90252292e-01 -6.67407274e-01 -4.06124536e-03
-6.27562046e-01 8.92767847e-01 -2.08900720e-01 5.27705669e-01
-7.60189712e-01 -6.12697959e-01 9.56134856e-01 4.21236530e-02
1.36153087e-01 -3.41985881e-01 -2.57389218e-01 -1.24250329e+00
4.83417898e-01 -8.58741760e-01 6.99547172e-01 -4.70905378e-02
-1.68807006e+00 3.65262717e-01 -4.21704985e-02 -1.52428401e+00
-6.00701034e-01 -6.51457667e-01 -1.47444773e+00 7.23273396e-01
-1.38285816e+00 -5.72604477e-01 -3.21650684e-01 8.39629769e-01
8.86734247e-01 -5.18051386e-01 7.90510178e-01 2.25965545e-01
-9.74426806e-01 3.34594935e-01 2.94475943e-01 2.18334928e-01
6.42022431e-01 -1.36895120e+00 8.59769881e-01 1.28283584e+00
3.13637853e-02 2.34890178e-01 8.35080862e-01 -6.51308358e-01
-1.08500063e+00 -1.23526955e+00 6.19149268e-01 -3.58135253e-01
1.13149416e+00 -6.26801476e-02 -1.53485429e+00 7.57440090e-01
1.21695781e-02 2.27314979e-01 3.31168711e-01 3.10716420e-01
-1.80908702e-02 -2.88728058e-01 -8.43270719e-01 4.22567546e-01
8.54163587e-01 -1.96544632e-01 -6.12534881e-01 4.43087012e-01
5.05420148e-01 -3.12330574e-01 -7.65548468e-01 7.66690433e-01
1.53626397e-01 -1.07197130e+00 7.19843745e-01 -8.99789035e-01
2.10827842e-01 -2.48791069e-01 6.41843453e-02 -1.46140373e+00
-8.48025382e-02 -7.69658744e-01 -7.58682668e-01 1.16108263e+00
3.12902182e-01 -1.18766904e+00 3.63023907e-01 1.53880239e-01
-4.97876197e-01 -7.63826907e-01 -1.24004328e+00 -1.31737852e+00
4.89116162e-02 -8.06883633e-01 6.74044132e-01 1.24044287e+00
-2.40051568e-01 -5.98654896e-03 -4.64694113e-01 2.64472306e-01
4.64188844e-01 1.67194620e-01 6.30373359e-01 -1.60344434e+00
1.27485767e-01 -7.01184690e-01 -5.29742718e-01 -5.32974601e-01
1.18571155e-01 -6.29860461e-01 -2.99346782e-02 -1.06997204e+00
-6.45466149e-01 -2.57898688e-01 -8.49506557e-01 5.61177373e-01
-1.53934017e-01 -2.25596160e-01 -4.16835845e-01 1.64567754e-01
-5.42348683e-01 6.92320824e-01 7.85256803e-01 -7.56021664e-02
-3.66164833e-01 6.99275255e-01 -3.14288795e-01 9.56889033e-01
1.02001679e+00 -3.72375816e-01 -1.76174372e-01 -1.13676339e-01
4.89459373e-02 -2.48125363e-02 4.05209184e-01 -1.20969331e+00
1.61155522e-01 -3.07534844e-01 8.87294114e-02 -6.84870303e-01
-3.12534273e-02 -9.30567205e-01 -3.26766044e-01 6.23183787e-01
9.75880176e-02 8.64764392e-01 4.60553318e-01 8.71413529e-01
-4.25963819e-01 1.40910000e-02 4.03357059e-01 2.43266359e-01
-1.04604268e+00 5.56929469e-01 -6.98243797e-01 1.49076298e-01
1.41136289e+00 -1.30895957e-01 -1.60711035e-01 -4.21913564e-01
-7.74248481e-01 6.08694196e-01 -2.75367528e-01 9.46092546e-01
6.91318274e-01 -1.29315639e+00 -7.51653373e-01 5.19442677e-01
3.73574555e-01 2.15503454e-01 1.80208221e-01 1.03488958e+00
-1.03307411e-01 3.81803334e-01 -5.00432253e-02 -8.98767292e-01
-8.41787636e-01 6.70376062e-01 2.19905615e-01 -3.77542675e-01
-8.72697532e-01 2.19342425e-01 -4.85453606e-02 -1.65321186e-01
4.99191761e-01 -3.81825268e-01 -2.80814558e-01 5.32626361e-02
5.56996763e-01 6.73496544e-01 2.61704445e-01 -2.47501329e-01
-2.90671915e-01 3.57771754e-01 -3.84746045e-01 4.48286742e-01
1.19354808e+00 1.77477404e-01 -1.58698693e-01 1.18874037e+00
7.26770222e-01 -3.95141393e-01 -8.05736661e-01 -4.83335853e-01
6.87737286e-01 -2.62551576e-01 -3.39449286e-01 -4.12386268e-01
-8.46290350e-01 7.40768194e-01 6.35983884e-01 9.59477425e-01
1.32049847e+00 -2.14096114e-01 9.91045356e-01 7.27692962e-01
2.55271703e-01 -1.03238416e+00 4.18393582e-01 8.65114808e-01
1.00936472e+00 -1.47974563e+00 -2.99211442e-01 -5.62054776e-02
-5.96186459e-01 1.16767335e+00 8.44186664e-01 -2.49228865e-01
9.34586048e-01 4.40143831e-02 3.48905474e-03 -2.03224242e-01
-6.74878478e-01 -1.37072608e-01 1.99418738e-01 3.65883350e-01
1.97019860e-01 -3.50025952e-01 -1.12902887e-01 2.27029175e-01
-9.56025999e-03 -3.36068720e-01 3.71309757e-01 9.95780349e-01
-1.96835086e-01 -8.82136047e-01 -4.54402238e-01 1.06680763e+00
-7.44868815e-01 2.03042313e-01 4.81837941e-03 7.57053673e-01
-2.28874255e-02 9.36193287e-01 5.90771377e-01 -3.07558596e-01
6.31816745e-01 4.69683826e-01 -2.41521195e-01 -3.73767167e-01
-4.27246481e-01 -5.54018915e-02 -2.13545755e-01 -6.11481249e-01
-2.87673950e-01 -1.01350296e+00 -1.08126390e+00 -2.11567357e-01
-1.01839460e-01 7.30418190e-02 3.67232770e-01 1.46879148e+00
1.50625721e-01 9.84843910e-01 9.65421736e-01 -4.20366794e-01
-6.49819970e-01 -1.00680530e+00 -7.28421330e-01 6.27803445e-01
7.66610742e-01 -7.14996874e-01 -4.98085946e-01 -3.29278320e-01] | [7.350043773651123, 2.666863441467285] |
617f91c9-e445-4f89-894b-ac5aaa2cf422 | specialist-diffusion-plug-and-play-sample | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lu_Specialist_Diffusion_Plug-and-Play_Sample-Efficient_Fine-Tuning_of_Text-to-Image_Diffusion_Models_To_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lu_Specialist_Diffusion_Plug-and-Play_Sample-Efficient_Fine-Tuning_of_Text-to-Image_Diffusion_Models_To_CVPR_2023_paper.pdf | Specialist Diffusion: Plug-and-Play Sample-Efficient Fine-Tuning of Text-to-Image Diffusion Models To Learn Any Unseen Style | Diffusion models have demonstrated impressive capability of text-conditioned image synthesis, and broader application horizons are emerging by personalizing those pretrained diffusion models toward generating some specialized target object or style. In this paper, we aim to learn an unseen style by simply fine-tuning a pre-trained diffusion model with a handful of images (e.g., less than 10), so that the fine-tuned model can generate high-quality images of arbitrary objects in this style. Such extremely lowshot fine-tuning is accomplished by a novel toolkit of finetuning techniques, including text-to-image customized data augmentations, a content loss to facilitate content-style disentanglement, and sparse updating that focuses on only a few time steps. Our framework, dubbed Specialist Diffusion, is plug-and-play to existing diffusion model backbones and other personalization techniques. We demonstrate it to outperform the latest few-shot personalization alternatives of diffusion models such as Textual Inversion and DreamBooth, in terms of learning highly sophisticated styles with ultra-sample-efficient tuning. We further show that Specialist Diffusion can be integrated on top of textual inversion to boost performance further, even on highly unusual styles. Our codes are available at: https://github.com/Picsart-AI-Research/Specialist-Diffusion | ['Humphrey Shi', 'Zhangyang Wang', 'Shant Navasardyan', 'Kai Wang', 'Hazarapet Tunanyan', 'Haoming Lu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['disentanglement'] | ['methodology'] | [ 2.89188921e-01 9.76695716e-02 -1.18613549e-01 -1.73031464e-01
-7.77515411e-01 -6.28894150e-01 8.79946053e-01 -5.21913886e-01
-2.53447205e-01 7.52598703e-01 4.09690946e-01 -5.04837669e-02
-6.60547928e-04 -8.01026285e-01 -6.64039254e-01 -6.97703540e-01
3.24660957e-01 8.34388733e-01 9.89046693e-02 -4.74140674e-01
1.62936762e-01 3.68304372e-01 -1.22378659e+00 2.05746412e-01
8.93159151e-01 6.10504091e-01 4.51635480e-01 9.33378875e-01
-2.57636309e-02 5.44735968e-01 -5.21433353e-01 -5.37665427e-01
3.54265571e-01 -6.49842978e-01 -5.76837778e-01 3.83498877e-01
6.17702723e-01 -5.17800093e-01 -4.07786429e-01 8.46485555e-01
6.25792921e-01 1.40443966e-01 6.43183470e-01 -9.55159664e-01
-1.31331503e+00 6.77160978e-01 -7.34737575e-01 1.84128493e-01
3.61210406e-02 6.46091223e-01 8.53681982e-01 -9.89922225e-01
7.54306078e-01 1.34863782e+00 5.76776505e-01 9.99349117e-01
-1.64995360e+00 -7.64741659e-01 2.21575543e-01 1.14464462e-02
-1.23346603e+00 -6.54270947e-01 7.97295213e-01 -4.48835164e-01
6.85836971e-01 1.15054958e-01 7.01163590e-01 1.72470593e+00
5.26019037e-02 8.35937262e-01 1.16740048e+00 -1.46890774e-01
2.35914603e-01 2.24975929e-01 -3.59155357e-01 7.73803651e-01
5.16504906e-02 6.37416840e-02 -6.09377980e-01 3.24574113e-02
1.10715270e+00 -1.33880183e-01 -2.28552625e-01 -2.20701531e-01
-1.50274813e+00 9.91117120e-01 3.24921101e-01 2.52923686e-02
-2.78142452e-01 3.91657770e-01 2.32697710e-01 4.15661067e-01
7.36628354e-01 6.24411821e-01 -3.92397612e-01 -1.39021292e-01
-1.16479743e+00 3.61838847e-01 7.51675725e-01 1.21698499e+00
9.64962780e-01 3.84566307e-01 -5.45282125e-01 9.33239460e-01
-1.52107164e-01 6.38685167e-01 6.26296401e-01 -1.23650730e+00
1.07912198e-01 1.46081403e-01 8.51130765e-03 -6.41736746e-01
-1.41163506e-02 -5.88198185e-01 -1.15669644e+00 2.81550258e-01
1.70035094e-01 -5.07537484e-01 -1.24745190e+00 1.81753850e+00
3.58984828e-01 2.83774912e-01 -1.87783241e-01 7.00150192e-01
5.66376448e-01 8.89003158e-01 -1.31319270e-01 -2.14269124e-02
1.24998474e+00 -1.30726850e+00 -3.74227762e-01 -3.50455731e-01
3.71299446e-01 -9.61616993e-01 1.67137372e+00 6.04044735e-01
-1.22787094e+00 -5.40878356e-01 -9.13361192e-01 -2.07836077e-01
-3.71048786e-02 -2.76605729e-02 8.49442720e-01 6.45352840e-01
-1.17195785e+00 7.95452297e-01 -6.50447309e-01 -3.94567341e-01
7.76099324e-01 1.60233617e-01 -6.72364533e-02 -2.18948662e-01
-8.29055488e-01 5.91534555e-01 8.53040367e-02 -3.36113751e-01
-1.33420372e+00 -1.17202306e+00 -4.61274713e-01 -2.12716162e-01
4.27748293e-01 -1.41666567e+00 1.11377561e+00 -1.09822083e+00
-2.03000474e+00 7.86206722e-01 -1.90792624e-02 -2.84214914e-01
7.82917500e-01 -2.97046512e-01 -2.75283247e-01 -2.71350015e-02
1.80882156e-01 1.01042855e+00 1.59278286e+00 -1.38658941e+00
-3.76880825e-01 -2.53813211e-02 1.02872446e-01 2.45780662e-01
-6.67331278e-01 -2.75944859e-01 -9.34610903e-01 -1.22996950e+00
-3.88338894e-01 -1.15235317e+00 -4.62010980e-01 3.50076765e-01
-4.37662780e-01 5.49775772e-02 9.74551737e-01 -2.92189330e-01
9.98263538e-01 -2.06148958e+00 6.11564159e-01 -1.62710659e-02
5.10501504e-01 2.07716852e-01 -5.51736355e-01 3.73462081e-01
1.87854633e-01 -1.68549363e-02 -4.62496549e-01 -7.65880108e-01
1.68212522e-02 5.25678575e-01 -3.45102996e-01 2.75827080e-01
1.42187193e-01 9.56465840e-01 -9.13775027e-01 -3.57285172e-01
1.16963841e-01 4.87478673e-01 -1.05448306e+00 1.69117063e-01
-6.01820290e-01 7.87791789e-01 -3.19054902e-01 3.94589990e-01
5.76355934e-01 -4.72279370e-01 -2.22563013e-01 -3.83062094e-01
8.54775235e-02 -2.90407002e-01 -1.18741298e+00 2.18732738e+00
-6.98742449e-01 6.03700519e-01 1.09262414e-01 -6.06760383e-01
6.84446812e-01 -2.17199568e-02 1.78684369e-01 -4.55875903e-01
3.74694280e-02 8.95169899e-02 -3.08982521e-01 -2.76434302e-01
5.88150203e-01 -2.29739308e-01 5.35447635e-02 7.46582031e-01
2.32940570e-01 -7.00715661e-01 5.14240146e-01 5.56391954e-01
9.07338083e-01 2.53948212e-01 -1.31315947e-01 -3.35919380e-01
6.16712682e-02 -8.98001865e-02 3.89747769e-01 9.61658657e-01
2.03490779e-01 8.80743265e-01 2.54599422e-01 -9.58072469e-02
-1.43159997e+00 -1.14886546e+00 -2.61339545e-03 1.28279400e+00
7.73067698e-02 -5.51399469e-01 -8.24742138e-01 -6.24029398e-01
-1.01649068e-01 7.67291903e-01 -7.58432746e-01 -1.30758524e-01
-4.76905614e-01 -9.73270535e-01 6.09596312e-01 3.00165743e-01
5.78939259e-01 -8.97664785e-01 -5.70528246e-02 8.26255828e-02
1.98424459e-01 -7.70519555e-01 -9.57631707e-01 3.33838314e-02
-8.20127606e-01 -4.90500093e-01 -1.23060191e+00 -6.57471955e-01
6.33428633e-01 1.81023642e-01 1.27078068e+00 -1.73632950e-01
-8.11338499e-02 4.67137963e-01 -2.66954809e-01 -1.93412900e-01
-6.30560696e-01 2.00028807e-01 5.78486174e-02 1.52446955e-01
-2.07265615e-01 -9.99763966e-01 -9.49875474e-01 2.57381380e-01
-1.17358494e+00 3.83873254e-01 7.69940674e-01 1.05054963e+00
4.84517217e-01 -6.67878687e-02 4.34704781e-01 -1.25945222e+00
8.41077685e-01 -4.59492832e-01 -3.93783450e-01 -2.94334851e-02
-7.39532530e-01 2.98119813e-01 6.64734662e-01 -9.04299200e-01
-1.31631255e+00 -2.06750855e-01 -2.21586391e-01 -6.98960006e-01
1.00016572e-01 2.28294820e-01 2.72438657e-02 -1.77704170e-01
1.12418413e+00 3.40957701e-01 -1.10518083e-01 -5.87712109e-01
1.00352407e+00 2.70354450e-01 5.50472975e-01 -9.15556490e-01
1.10623682e+00 5.74622750e-01 -1.62743211e-01 -8.22967172e-01
-9.29689527e-01 4.94053848e-02 -3.68414938e-01 -2.53115967e-02
6.52546763e-01 -9.98025954e-01 -2.49079153e-01 7.32221544e-01
-7.84303665e-01 -1.03502500e+00 -7.72225380e-01 6.65606335e-02
-6.28444433e-01 3.73455256e-01 -8.45370829e-01 -9.62179005e-02
-3.09871912e-01 -1.13180113e+00 1.10923398e+00 1.24137878e-01
-4.02497411e-01 -1.15375841e+00 1.29679158e-01 3.04908693e-01
8.23142707e-01 -7.06055388e-02 7.50463724e-01 -1.23036779e-01
-6.54770970e-01 2.31239378e-01 -1.90431580e-01 4.75748599e-01
2.25744285e-02 7.79355923e-03 -8.95271301e-01 -5.42470157e-01
-9.90000293e-02 -4.95348364e-01 1.14027083e+00 2.79782206e-01
1.28940260e+00 -5.05466104e-01 -1.55720830e-01 1.17432284e+00
1.19981468e+00 -1.63702071e-01 5.40257156e-01 1.32833883e-01
1.00725412e+00 1.72826260e-01 7.19327778e-02 7.27076709e-01
2.61136323e-01 5.73624253e-01 7.33669698e-02 -1.68737963e-01
-7.69024134e-01 -2.93373793e-01 2.47791365e-01 7.78615594e-01
-1.19873494e-01 -3.86492133e-01 -6.34978354e-01 5.36932886e-01
-1.61282361e+00 -9.04342532e-01 2.27046520e-01 1.74260402e+00
1.28639364e+00 2.01816693e-01 1.10141456e-01 -2.48302564e-01
4.61289078e-01 4.37490702e-01 -9.69200313e-01 -1.53831899e-01
-2.43010074e-01 3.75563115e-01 4.69287157e-01 6.38427615e-01
-7.93147743e-01 1.33443081e+00 6.10968781e+00 1.31764543e+00
-1.19629061e+00 3.89454901e-01 6.99436426e-01 -4.55789328e-01
-8.01495314e-01 -1.02767337e-03 -7.63157189e-01 4.25637454e-01
6.53870285e-01 -2.69331783e-01 8.31976593e-01 8.70524049e-01
1.30207822e-01 2.04438552e-01 -1.07975292e+00 1.05508316e+00
1.57538071e-01 -1.81456006e+00 4.92449552e-01 -6.93782326e-03
1.23721874e+00 5.10996766e-02 5.63183784e-01 3.47148150e-01
7.42439508e-01 -7.43653715e-01 7.47171342e-01 5.12670219e-01
1.11210227e+00 -5.16800880e-01 -3.02702989e-02 1.69458076e-01
-6.78557873e-01 -2.35259105e-02 -4.18492883e-01 2.35644013e-01
2.80071914e-01 8.75169694e-01 -3.81548733e-01 1.45310581e-01
7.08717167e-01 8.94640803e-01 -4.93048072e-01 6.18608177e-01
-1.34204179e-01 6.97749496e-01 -2.58819938e-01 2.22833693e-01
2.97547758e-01 -2.96808988e-01 8.19115341e-01 1.19343746e+00
5.80776334e-01 1.24902971e-01 -3.74089889e-02 1.08791625e+00
-3.11913192e-01 -1.85598031e-01 -4.76160944e-01 -1.12209572e-02
2.76879489e-01 1.22523820e+00 -4.81254309e-01 -7.21068203e-01
-2.72791743e-01 1.66523373e+00 2.79955119e-01 4.78463948e-01
-8.61867547e-01 -1.59529313e-01 7.86676168e-01 1.09769125e-02
5.60515344e-01 -2.90580720e-01 -4.54181880e-01 -1.46838653e+00
-3.16737533e-01 -1.12009430e+00 8.77365842e-02 -9.27918792e-01
-1.56497276e+00 7.07662642e-01 -1.21284403e-01 -9.15862978e-01
-4.31381464e-02 -4.76540893e-01 -6.29090905e-01 8.80923212e-01
-1.15596056e+00 -1.38655818e+00 -4.14512545e-01 9.44057524e-01
9.06017482e-01 -2.65005887e-01 6.11008942e-01 3.57978046e-01
-6.59380078e-01 7.51429915e-01 2.60310948e-01 -3.43946427e-01
9.86164153e-01 -1.31000054e+00 6.52856588e-01 7.93966830e-01
1.43736646e-01 4.93652582e-01 9.83157218e-01 -6.78855419e-01
-1.47140586e+00 -1.31155550e+00 3.23743731e-01 -6.31651700e-01
8.82454634e-01 -3.67046744e-01 -6.82559252e-01 7.21191466e-01
4.84440774e-01 -1.84145108e-01 3.60073894e-01 1.59744456e-01
-3.80216658e-01 -1.15790881e-01 -8.84100914e-01 1.18041968e+00
1.44716084e+00 -4.50048298e-01 -2.03693032e-01 5.30010402e-01
7.88167834e-01 -5.22025943e-01 -8.91213298e-01 8.87442231e-02
3.77083629e-01 -1.01499414e+00 1.06826663e+00 -3.78273576e-01
6.79252684e-01 -8.04865807e-02 -8.28975663e-02 -1.66614544e+00
-7.28968978e-01 -1.16068721e+00 -2.06995219e-01 1.24191475e+00
5.11736989e-01 -3.99962187e-01 1.04463792e+00 3.16160619e-01
-2.65463561e-01 -6.33400977e-01 -4.83480752e-01 -6.78407133e-01
2.33175814e-01 -3.26221943e-01 4.72561657e-01 8.90452027e-01
-5.87116241e-01 7.22055554e-01 -8.94299924e-01 -8.36397707e-02
8.59443009e-01 -9.40056704e-03 1.09433651e+00 -7.90294349e-01
-8.50886345e-01 -6.67023242e-01 7.78650641e-02 -1.54086161e+00
-1.00042589e-01 -7.54197598e-01 -2.71501660e-01 -1.34884965e+00
1.03667870e-01 -6.01343811e-01 4.59869839e-02 2.71454424e-01
-3.01302344e-01 7.31922030e-01 2.68567473e-01 4.17748779e-01
-3.05046052e-01 8.84345233e-01 1.92830551e+00 -4.25954401e-01
-2.87395000e-01 -7.84414932e-02 -1.04442489e+00 6.13140583e-01
6.69934034e-01 -4.02541161e-01 -7.72286713e-01 -6.99529231e-01
2.96419322e-01 -1.54015258e-01 2.24563777e-01 -9.31745112e-01
1.06000438e-01 -2.17269644e-01 3.64742368e-01 -1.16161518e-01
5.97491264e-01 -4.09935415e-01 2.94888765e-01 1.62277371e-01
-4.56048846e-01 -1.19640410e-01 2.13071242e-01 6.75157607e-01
1.91296324e-01 -1.84102114e-02 9.63244796e-01 -3.03810775e-01
-8.04409146e-01 7.04686821e-01 -2.65426308e-01 3.18165749e-01
8.97436678e-01 -2.13566050e-01 -2.54752517e-01 -5.70871294e-01
-8.83665919e-01 -3.27933989e-02 6.42707407e-01 3.74658853e-01
4.43884015e-01 -1.29704487e+00 -9.05691683e-01 2.67832547e-01
5.67263700e-02 1.06154576e-01 5.07807076e-01 7.02551246e-01
-3.71150315e-01 -1.17753252e-01 -2.70953536e-01 -5.35253823e-01
-9.21168327e-01 6.62978411e-01 -1.90198161e-02 -1.74712628e-01
-8.58310878e-01 1.25532329e+00 7.05567420e-01 -1.92263573e-01
-3.27260047e-03 -7.74337798e-02 3.66144955e-01 -3.37782204e-02
4.88227248e-01 2.99225807e-01 -3.04493785e-01 -2.08180562e-01
1.22861236e-01 6.77284956e-01 -3.10863554e-01 -3.51186037e-01
1.36183691e+00 -3.12399477e-01 4.97872420e-02 2.75324345e-01
1.01467168e+00 2.78843731e-01 -1.87935340e+00 -4.84160125e-01
-6.59608364e-01 -5.04872561e-01 1.91219762e-01 -8.21814716e-01
-1.24149334e+00 6.29647911e-01 3.49216431e-01 -3.76934111e-02
1.20771313e+00 1.18054293e-01 9.70522285e-01 3.27922106e-01
1.21738940e-01 -9.27925706e-01 5.10905147e-01 4.29206252e-01
1.14798892e+00 -1.07252824e+00 -2.65757684e-02 -4.02621217e-02
-9.60992992e-01 7.84680128e-01 5.28344750e-01 -3.94892067e-01
6.85702503e-01 3.65164489e-01 1.00452632e-01 -6.52152300e-02
-9.48089242e-01 -2.15819981e-02 2.72953033e-01 6.98474467e-01
1.51297048e-01 -3.50689366e-02 1.72834903e-01 1.88971788e-01
-3.77890617e-01 7.17220828e-02 5.08371949e-01 5.12802243e-01
-2.79503852e-01 -1.12799728e+00 -1.59066990e-01 5.54458022e-01
-1.17711440e-01 -5.06719828e-01 -5.82541674e-02 5.54414630e-01
2.54878178e-02 6.25693262e-01 -1.60594042e-02 -2.67721325e-01
2.03962073e-01 -3.27574760e-01 6.78205729e-01 -6.50183618e-01
-3.86764526e-01 1.44946188e-01 1.89569995e-01 -6.29422784e-01
-2.36271426e-01 -4.76484507e-01 -5.06681800e-01 -8.71910036e-01
3.89961265e-02 -1.40019834e-01 2.59051889e-01 7.55416811e-01
5.67252278e-01 5.54144025e-01 4.82431680e-01 -1.45325434e+00
-4.82542098e-01 -8.96441877e-01 -6.01997077e-01 4.56287861e-01
3.57808799e-01 -5.40587246e-01 -2.97592610e-01 3.65909189e-01] | [11.347957611083984, -0.3368754982948303] |
832ce4ae-c256-40e1-b886-bb3f5d32a4bf | enforcing-interpretability-and-its | 2010.13764 | null | https://arxiv.org/abs/2010.13764v2 | https://arxiv.org/pdf/2010.13764v2.pdf | Enforcing Interpretability and its Statistical Impacts: Trade-offs between Accuracy and Interpretability | To date, there has been no formal study of the statistical cost of interpretability in machine learning. As such, the discourse around potential trade-offs is often informal and misconceptions abound. In this work, we aim to initiate a formal study of these trade-offs. A seemingly insurmountable roadblock is the lack of any agreed upon definition of interpretability. Instead, we propose a shift in perspective. Rather than attempt to define interpretability, we propose to model the \emph{act} of \emph{enforcing} interpretability. As a starting point, we focus on the setting of empirical risk minimization for binary classification, and view interpretability as a constraint placed on learning. That is, we assume we are given a subset of hypothesis that are deemed to be interpretable, possibly depending on the data distribution and other aspects of the context. We then model the act of enforcing interpretability as that of performing empirical risk minimization over the set of interpretable hypotheses. This model allows us to reason about the statistical implications of enforcing interpretability, using known results in statistical learning theory. Focusing on accuracy, we perform a case analysis, explaining why one may or may not observe a trade-off between accuracy and interpretability when the restriction to interpretable classifiers does or does not come at the cost of some excess statistical risk. We close with some worked examples and some open problems, which we hope will spur further theoretical development around the tradeoffs involved in interpretability. | ['Daniel M. Roy', 'Shai Ben-David', 'Gintare Karolina Dziugaite'] | 2020-10-26 | null | null | null | null | ['misconceptions'] | ['miscellaneous'] | [ 5.06744504e-01 8.78202975e-01 -1.47372156e-01 -6.96901739e-01
-6.40633345e-01 -7.09064901e-01 6.10222042e-01 4.18697596e-01
-5.31014204e-01 6.08716249e-01 2.66453505e-01 -7.60292947e-01
-5.32931626e-01 -4.43766385e-01 -6.08083546e-01 -6.18769825e-01
1.49230585e-01 4.88520682e-01 -3.39434236e-01 1.79029956e-01
3.43936294e-01 2.86838681e-01 -1.40914762e+00 2.85339057e-01
8.85827065e-01 7.39128590e-01 -2.05910295e-01 4.05096501e-01
2.12339073e-01 6.37509346e-01 -5.27106464e-01 -6.05689228e-01
1.57436028e-01 -6.59476936e-01 -1.31680465e+00 3.35068941e-01
2.80946404e-01 -4.85778525e-02 2.74146795e-01 1.08728182e+00
-4.35334258e-02 -1.29950851e-01 9.68721569e-01 -1.41236782e+00
-6.52276158e-01 8.48205388e-01 -1.63497701e-01 2.58752257e-01
1.03258096e-01 1.80799201e-01 1.56740212e+00 -4.38308805e-01
2.48684734e-01 1.14966559e+00 3.46797138e-01 4.98929858e-01
-1.40737569e+00 -4.04679999e-02 2.96256989e-01 1.63193699e-02
-1.12368023e+00 -3.90807569e-01 4.55158144e-01 -5.06896257e-01
3.77583593e-01 8.45479548e-01 4.14169520e-01 9.89301741e-01
1.49174318e-01 6.86615586e-01 1.47819102e+00 -8.75369012e-01
4.58808333e-01 3.62868547e-01 6.09871924e-01 6.01942420e-01
6.76685452e-01 2.23108351e-01 -3.76808703e-01 -2.23085910e-01
3.61663759e-01 -2.89104909e-01 -3.26888949e-01 -2.20409453e-01
-8.38606954e-01 9.61445451e-01 1.51390925e-01 2.48366758e-01
-5.75155653e-02 6.85052723e-02 3.52912128e-01 5.85288823e-01
4.81389135e-01 8.01691532e-01 -5.41658282e-01 7.79961199e-02
-4.73030925e-01 1.80163950e-01 7.48001575e-01 3.56050283e-01
4.80352402e-01 -5.03704667e-01 2.06787199e-01 4.77139443e-01
5.06559968e-01 1.66814148e-01 1.49418265e-01 -9.04387116e-01
2.59962261e-01 4.59391236e-01 2.25140527e-01 -8.38641167e-01
-3.33981454e-01 -7.40742832e-02 -4.72663790e-01 3.60730201e-01
7.73322701e-01 -1.17914207e-01 -5.36008179e-01 1.86097026e+00
-4.82398495e-02 -2.14095592e-01 1.36756320e-02 9.46857870e-01
3.11096874e-03 2.83893406e-01 1.76848963e-01 -3.62829566e-01
1.44010544e+00 -4.33937490e-01 -5.27278721e-01 -2.80362070e-01
8.74011815e-01 -5.53449571e-01 1.45096231e+00 3.89529288e-01
-9.78173494e-01 -8.92657414e-02 -1.03694284e+00 -1.31445318e-01
-5.65543585e-02 -1.44922540e-01 8.33987355e-01 9.20406759e-01
-6.67842031e-01 5.52407086e-01 -9.82138157e-01 -2.57259816e-01
1.43864989e-01 3.37367564e-01 -2.70862430e-01 2.14369357e-01
-9.71882701e-01 1.06666744e+00 4.22881782e-01 2.80594409e-01
-2.52858818e-01 -3.31429005e-01 -7.54168212e-01 1.69210568e-01
6.38060451e-01 -7.16053247e-01 1.24368823e+00 -1.82260680e+00
-1.02743220e+00 9.44845915e-01 -1.26487315e-01 -4.58347470e-01
6.41238630e-01 -1.66422188e-01 -1.66058093e-01 -2.80135870e-01
5.39671183e-02 2.92386919e-01 4.58230644e-01 -1.23033071e+00
-6.70400381e-01 -4.72319275e-01 5.23863316e-01 2.19788492e-01
-2.06713840e-01 1.01621971e-02 -8.65447521e-03 -4.96402591e-01
1.55466020e-01 -1.17175889e+00 -4.24569286e-02 2.02779043e-02
-5.08473456e-01 -3.38221937e-01 2.89808154e-01 -1.77087173e-01
1.26583505e+00 -2.28582692e+00 4.32849452e-02 3.87184978e-01
3.25285345e-01 -8.47363845e-02 3.60329151e-02 9.85930040e-02
-2.79351950e-01 7.69303024e-01 -4.97297913e-01 -3.54502171e-01
2.53182888e-01 4.02595609e-01 -4.69252408e-01 6.88736260e-01
4.94178593e-01 5.34011006e-01 -7.67849803e-01 -2.39365712e-01
6.72242939e-02 1.60118774e-01 -6.49913251e-01 -6.37444183e-02
-6.73236176e-02 1.98536932e-01 -6.32565439e-01 2.89695323e-01
3.81515115e-01 -2.93751776e-01 5.60331047e-01 -4.74022590e-02
-5.76549135e-02 7.47575819e-01 -1.34272647e+00 8.65437090e-01
-3.63573849e-01 7.12392986e-01 -1.03696868e-01 -1.20475066e+00
2.78310031e-01 9.63560119e-02 -7.12492764e-02 -2.40828305e-01
1.85971245e-01 4.46005017e-01 4.85704452e-01 -5.66983700e-01
2.61991650e-01 -7.10543096e-01 -1.30824566e-01 6.71054721e-01
-3.22751462e-01 -2.85689179e-02 -2.39323393e-01 -2.07004905e-01
9.85473216e-01 -2.45413616e-01 5.49584687e-01 -6.88346982e-01
4.46234286e-01 -1.71533391e-01 4.50904578e-01 8.96952748e-01
6.11456223e-02 6.58135355e-01 9.87562418e-01 -5.14890134e-01
-1.00759697e+00 -1.14496386e+00 -4.70729977e-01 7.83542812e-01
2.63484214e-02 -3.43974799e-01 -7.87332058e-01 -9.99089658e-01
-1.20690562e-01 1.05896854e+00 -8.41742814e-01 -1.37295127e-01
-4.22757536e-01 -8.89187992e-01 3.22620451e-01 2.87052244e-01
4.16074134e-02 -7.81690359e-01 -8.83741438e-01 -3.82832736e-01
-1.77314095e-02 -6.40420556e-01 -3.57264519e-01 4.28206205e-01
-9.24773097e-01 -1.16992247e+00 9.60636213e-02 -2.07430854e-01
9.79352295e-01 -1.13095641e-02 1.16799605e+00 5.43983996e-01
1.81062803e-01 4.52179939e-01 -3.21055681e-01 -6.26721680e-01
-7.30318487e-01 1.27252609e-01 6.96240067e-02 -8.52404982e-02
5.05610943e-01 -2.48374164e-01 -2.96397060e-01 2.74311066e-01
-1.28888881e+00 8.83187875e-02 4.88286227e-01 1.02668297e+00
3.04298133e-01 6.01707809e-02 5.01953542e-01 -1.29770148e+00
6.53656483e-01 -5.59048057e-01 -6.18120909e-01 4.67221290e-01
-1.04982269e+00 4.92645264e-01 4.92426634e-01 -3.03555518e-01
-9.15573061e-01 -2.14776352e-01 1.36585444e-01 1.81037515e-01
-2.74721324e-01 7.22592831e-01 -3.63097399e-01 2.82679349e-01
6.99930012e-01 -3.21000338e-01 6.43600151e-02 -2.81258762e-01
2.40069211e-01 8.62723410e-01 -1.64489791e-01 -8.74163628e-01
7.80307174e-01 3.69314730e-01 -5.46544343e-02 -7.17503846e-01
-1.19737935e+00 3.34494933e-02 -5.05109966e-01 6.15631789e-03
7.15067387e-01 -3.10850918e-01 -7.80044973e-01 -2.26604700e-01
-1.01382589e+00 -2.16534376e-01 -4.73562837e-01 4.36513692e-01
-8.39514434e-01 3.91140997e-01 -1.05320066e-01 -1.08498454e+00
1.64776996e-01 -1.29921746e+00 9.09374237e-01 -1.78534418e-01
-9.13472712e-01 -1.26932776e+00 -4.01479274e-01 5.93312442e-01
7.74183646e-02 1.18297614e-01 1.25012493e+00 -1.08666098e+00
-5.19209981e-01 -1.13999844e-01 5.08703142e-02 4.97459531e-01
3.14892501e-01 -1.33669302e-01 -1.05628431e+00 6.15131743e-02
4.61004496e-01 -3.71988207e-01 6.07425511e-01 3.28289688e-01
1.23550296e+00 -8.83906841e-01 -4.04451415e-03 2.29180276e-01
1.29269171e+00 1.04523376e-01 4.58036542e-01 3.39812487e-01
4.25568014e-01 1.06995559e+00 5.72140098e-01 2.13940904e-01
2.42793232e-01 7.26639569e-01 2.76617885e-01 -4.87406664e-02
5.01278579e-01 -1.53546780e-01 1.44024551e-01 2.72151530e-01
-1.53477073e-01 -3.76435757e-01 -9.43608105e-01 2.81454235e-01
-1.73170817e+00 -9.12731230e-01 -2.10416213e-01 2.54399991e+00
9.12264824e-01 5.13553381e-01 7.63191357e-02 3.56307745e-01
5.39602101e-01 -1.28327673e-02 -1.70892194e-01 -6.63498998e-01
1.02423571e-01 -1.17016211e-02 3.20670009e-01 1.07603836e+00
-9.08827782e-01 4.85845864e-01 6.38714743e+00 6.83045268e-01
-1.03770924e+00 -1.76214933e-01 1.15642416e+00 8.11733902e-02
-8.14207673e-01 4.37027723e-01 -4.13812011e-01 3.76506954e-01
8.53033483e-01 -1.83424160e-01 3.99747163e-01 7.13989377e-01
2.89298177e-01 -2.77661115e-01 -1.74727511e+00 2.90544927e-01
-4.47478704e-02 -9.75108385e-01 -6.03316613e-02 3.60742092e-01
2.35065877e-01 -4.07485127e-01 1.37620404e-01 -2.70622484e-02
2.80166984e-01 -1.36328185e+00 1.03299546e+00 2.60906965e-01
4.14903820e-01 -6.26490295e-01 8.97130907e-01 5.29894531e-01
-4.19275343e-01 -1.67911127e-01 -4.58079800e-02 -5.63433945e-01
-2.24788010e-01 5.81284702e-01 -8.30428720e-01 2.77195424e-01
2.17804357e-01 1.52106240e-01 -3.46295238e-01 4.87704784e-01
-2.53276080e-01 8.19476604e-01 -2.99542814e-01 2.02327576e-02
1.60950705e-01 -2.91944325e-01 3.99180114e-01 1.01034069e+00
-4.37216610e-02 3.68590385e-01 -8.98679197e-02 1.03195786e+00
1.42466486e-01 -8.97239521e-02 -6.65015638e-01 -1.41444147e-01
3.55038375e-01 9.62328196e-01 -9.41840351e-01 -1.20898739e-01
-4.08106267e-01 7.15522468e-01 3.27973098e-01 1.83215782e-01
-7.72502959e-01 2.16381311e-01 6.07540846e-01 1.33231118e-01
-3.22213858e-01 4.52285931e-02 -8.81759405e-01 -1.13810539e+00
2.60222584e-01 -1.09386158e+00 4.66703981e-01 -3.79772991e-01
-1.43806815e+00 1.39627084e-01 3.48748028e-01 -8.10669720e-01
-2.90181655e-02 -7.13558912e-01 -4.88504618e-01 8.65284503e-01
-1.15848362e+00 -7.01553106e-01 3.07043701e-01 -2.31656149e-01
6.12206221e-01 4.41683590e-01 7.23471463e-01 -2.57418871e-01
-4.54012603e-01 5.72746456e-01 -1.50791898e-01 -9.78366584e-02
3.17850947e-01 -1.46032035e+00 1.72464877e-01 6.93124652e-01
2.77316779e-01 9.32718396e-01 1.11398220e+00 -3.35085332e-01
-1.14313257e+00 -7.94164419e-01 1.23762858e+00 -9.13240612e-01
6.73786998e-01 -2.32121527e-01 -9.43382442e-01 9.81542349e-01
-4.50564697e-02 -2.07582399e-01 8.53837132e-01 6.01072848e-01
-4.06130016e-01 4.11259346e-02 -1.18554807e+00 7.97409296e-01
9.29809272e-01 -5.17319381e-01 -9.50274646e-01 2.11311489e-01
5.32928288e-01 -1.45500734e-01 -6.16329432e-01 4.02839839e-01
6.20721757e-01 -1.06232250e+00 6.40502512e-01 -8.51822257e-01
4.92574781e-01 -1.56266198e-01 -2.59692013e-01 -1.06556165e+00
-7.94944242e-02 -3.29131693e-01 5.16104162e-01 1.24545276e+00
7.28824854e-01 -7.69273162e-01 6.28749073e-01 1.29113626e+00
-8.70820582e-02 -7.71483362e-01 -1.04631686e+00 -7.07701266e-01
2.45253250e-01 -6.37775898e-01 2.96276689e-01 1.18380368e+00
3.19050401e-01 4.62916762e-01 3.59482714e-03 3.06189716e-01
3.39466661e-01 -6.35529831e-02 3.64583850e-01 -1.17575431e+00
-6.17211580e-01 -4.80170220e-01 -3.59199047e-01 -6.43670917e-01
3.12633276e-01 -8.11534345e-01 5.41302636e-02 -9.84676778e-01
4.31324065e-01 -6.83584929e-01 -2.02362999e-01 5.55059195e-01
-3.91574323e-01 -1.49210051e-01 3.84650767e-01 2.17014924e-01
-3.59790564e-01 1.57241285e-01 7.92637050e-01 -7.24832481e-03
-6.15496077e-02 1.61313802e-01 -1.14091146e+00 1.09630954e+00
8.06314588e-01 -5.37586808e-01 -7.15189278e-01 -4.02329236e-01
4.35246259e-01 4.18483466e-03 5.16412914e-01 -3.60218048e-01
-2.58340359e-01 -6.94662511e-01 5.95861189e-02 3.00817490e-01
1.04881721e-02 -1.05832636e+00 2.12953940e-01 4.68618870e-01
-9.75304246e-01 1.53505772e-01 -4.67013940e-02 4.69958872e-01
-2.28792783e-02 -6.12029076e-01 6.47253096e-01 7.75509700e-02
1.11486549e-02 -2.39403486e-01 -4.99099791e-01 2.83410639e-01
8.48636985e-01 -9.72255915e-02 -1.39002174e-01 -1.37764975e-01
-6.81768477e-01 1.22883497e-02 7.21885264e-01 3.19170117e-01
2.95380384e-01 -9.05670881e-01 -4.80105221e-01 6.04039915e-02
1.25755027e-01 -1.60597116e-01 -1.50802732e-01 7.22895443e-01
-2.43840739e-01 4.10548568e-01 3.01925480e-01 -2.85954535e-01
-1.31643665e+00 3.66836399e-01 5.26033163e-01 -1.34869605e-01
-3.68030012e-01 5.07946074e-01 5.41594982e-01 -1.58973739e-01
1.85949326e-01 -5.45407593e-01 7.85080716e-02 -1.41045853e-01
2.20934734e-01 1.91542327e-01 -2.84685604e-02 -2.91936576e-01
-3.98169160e-01 1.95599735e-01 1.99226048e-02 -3.48739117e-01
1.13526940e+00 -1.23624697e-01 -1.52415484e-01 7.52999008e-01
9.29048538e-01 1.73020720e-01 -1.04318571e+00 2.11920977e-01
4.52192962e-01 -4.73731905e-01 -2.71101356e-01 -9.23149168e-01
-5.68865538e-01 7.35857248e-01 4.11640286e-01 7.54539073e-01
1.09827006e+00 2.60848314e-01 9.39052738e-03 5.01490474e-01
1.32139072e-01 -1.00172806e+00 -2.51581758e-01 2.86610633e-01
9.71204400e-01 -1.35265958e+00 3.46965492e-02 -5.15277803e-01
-5.94780207e-01 1.04248405e+00 4.08420533e-01 6.39669076e-02
3.72588515e-01 6.87269494e-02 -5.21226646e-03 -2.92892367e-01
-7.62732923e-01 6.99771289e-03 3.98534000e-01 3.19004059e-01
7.53189564e-01 3.62612605e-01 -7.88341820e-01 6.25128031e-01
-4.49633777e-01 -1.28759369e-01 5.68463326e-01 7.39599347e-01
-4.03338969e-01 -1.26160073e+00 -4.50201869e-01 5.29353678e-01
-7.79947460e-01 -1.68050513e-01 -6.89224184e-01 9.83745933e-01
2.49465600e-01 9.11709964e-01 1.62422583e-01 -1.40823185e-01
1.30821750e-01 1.89968795e-01 3.81557167e-01 -7.01808393e-01
-3.70526433e-01 -2.16905132e-01 4.57955092e-01 -1.28422379e-01
-3.40718031e-01 -8.10876369e-01 -1.12062013e+00 -3.81633848e-01
-4.96399134e-01 2.25517601e-01 3.33671808e-01 1.43403053e+00
3.66329146e-03 1.67275816e-01 4.99264389e-01 -1.90705135e-01
-9.66177404e-01 -4.92082983e-01 -5.10880053e-01 4.85539734e-01
5.75286627e-01 -4.39455807e-01 -9.55210328e-01 7.37948269e-02] | [8.741117477416992, 5.708701133728027] |
793ec5be-105f-4fde-a8ef-e867425304d4 | salsi-a-new-seismic-attribute-for-salt-dome | 1901.02937 | null | http://arxiv.org/abs/1901.02937v1 | http://arxiv.org/pdf/1901.02937v1.pdf | SalSi: A new seismic attribute for salt dome detection | In this paper, we propose a saliency-based attribute, SalSi, to detect salt
dome bodies within seismic volumes. SalSi is based on the saliency theory and
modeling of the human vision system (HVS). In this work, we aim to highlight
the parts of the seismic volume that receive highest attention from the human
interpreter, and based on the salient features of a seismic image, we detect
the salt domes. Experimental results show the effectiveness of SalSi on the
real seismic dataset acquired from the North Sea, F3 block. Subjectively, we
have used the ground truth and the output of different salt dome delineation
algorithms to validate the results of SalSi. For the objective evaluation of
results, we have used the receiver operating characteristics (ROC) curves and
area under the curves (AUC) to demonstrate SalSi is a promising and an
effective attribute for seismic interpretation. | ['Tariq Alshawi', 'Ghassan AlRegib', 'Zhiling Long', 'Muhammad Amir Shafiq'] | 2019-01-09 | null | null | null | null | ['seismic-interpretation'] | ['miscellaneous'] | [-1.03235140e-01 9.79048386e-02 6.40732527e-01 -4.43670340e-02
-5.06092310e-01 -2.46809945e-01 4.72841799e-01 3.60879004e-01
-5.84067106e-01 2.09500507e-01 4.58989799e-01 5.27012683e-02
-1.80722833e-01 -6.72883630e-01 -4.06668246e-01 -8.18763971e-01
-3.75310987e-01 1.68989614e-01 9.21442509e-01 -3.88425291e-01
9.18862760e-01 5.73059797e-01 -1.44275630e+00 1.55741736e-01
8.46854925e-01 8.06233883e-01 6.59146309e-01 3.63884896e-01
1.16998546e-01 6.45509243e-01 -5.73262036e-01 -7.27739781e-02
1.57323763e-01 -2.49118596e-01 -4.20160174e-01 -1.00125141e-01
4.86133061e-02 -7.67313540e-02 -1.48544908e-01 1.27261484e+00
6.49186075e-01 1.14014052e-01 7.47056067e-01 -9.18494225e-01
-7.92520866e-02 6.36781454e-01 -4.43594515e-01 6.10934258e-01
1.63624674e-01 3.94780971e-02 8.69421303e-01 -1.40634477e+00
4.48548675e-01 1.12449038e+00 6.06993616e-01 4.56345417e-02
-4.35523480e-01 -2.86733240e-01 -2.10562795e-01 4.23257202e-01
-1.48706973e+00 -1.71446264e-01 9.21004593e-01 -6.56831384e-01
3.11719805e-01 4.36518848e-01 8.20020020e-01 1.46238670e-01
1.52891219e-01 1.07411635e+00 1.09214199e+00 -3.64499152e-01
5.79733729e-01 1.88084766e-02 1.43125370e-01 6.52143419e-01
4.51462567e-01 -2.47736145e-02 -6.77464366e-01 -1.67666242e-01
5.89966774e-01 -1.05950780e-01 -5.83357990e-01 -2.93452710e-01
-1.19050491e+00 8.26879025e-01 6.59959733e-01 2.43294984e-01
-6.13979399e-01 -8.79679620e-02 2.69993782e-01 -3.73746753e-01
3.93609196e-01 5.07107556e-01 2.28542432e-01 1.43421859e-01
-1.10144579e+00 2.50238746e-01 4.98674065e-01 2.39544526e-01
4.62486684e-01 2.47962207e-01 -7.92125091e-02 4.56450343e-01
7.49135673e-01 1.04312789e+00 4.89043385e-01 -5.65539598e-01
2.06044808e-01 5.73722899e-01 2.24052384e-01 -1.14685142e+00
-4.54485834e-01 -4.20854181e-01 -3.58274847e-01 4.44444865e-01
8.83414149e-02 -1.09320313e-01 -8.15181077e-01 9.41401422e-01
-1.22421738e-02 -8.77889898e-03 4.24101263e-01 1.42925584e+00
9.75664854e-01 6.33464098e-01 1.18519105e-01 1.13723055e-01
1.33725917e+00 -5.07474065e-01 -6.03991926e-01 -3.84725183e-01
2.87013590e-01 -4.46805179e-01 9.87033784e-01 2.41595712e-02
-8.12265158e-01 -2.79034257e-01 -1.21371305e+00 5.70014596e-01
-1.32629260e-01 2.48572439e-01 2.72918105e-01 3.21830094e-01
-8.13649833e-01 3.33106458e-01 -8.15396309e-01 -2.22198531e-01
2.32726052e-01 -1.88138455e-01 -1.90765522e-02 3.29751045e-01
-1.38075244e+00 9.49399650e-01 3.63371700e-01 4.26678181e-01
-1.24571872e+00 -3.24389100e-01 -9.36280489e-01 3.43306780e-01
1.45717889e-01 -4.13758904e-02 9.03167188e-01 -7.28424907e-01
-6.27008200e-01 7.59870052e-01 1.09423570e-01 -6.68913126e-01
5.30489862e-01 -2.24745899e-01 -1.40839577e-01 6.41533971e-01
2.27718353e-01 3.81459683e-01 5.42430639e-01 -1.65464127e+00
-7.50256360e-01 -2.16199741e-01 -3.03830177e-01 4.01530027e-01
-1.77276321e-02 1.79228276e-01 -2.01819509e-01 -7.50381887e-01
6.12936318e-01 -7.14095414e-01 -8.06402192e-02 -1.69400513e-01
-1.82855904e-01 1.48942629e-02 8.11291516e-01 -1.04936314e+00
1.01216936e+00 -2.57616377e+00 -1.56094313e-01 4.94417101e-01
9.07742903e-02 3.95731479e-01 4.30432767e-01 1.06992826e-01
4.11494106e-01 -5.03281876e-02 -9.49569702e-01 2.53199995e-01
-1.46482140e-01 3.94191369e-02 -2.12286636e-01 3.80166799e-01
-1.90061331e-02 5.42193592e-01 -7.83075452e-01 -6.76352620e-01
2.31261864e-01 -3.87839996e-03 2.87966859e-02 2.54637688e-01
1.14126012e-01 1.44048929e-01 -6.34858370e-01 7.30866551e-01
7.94454098e-01 2.79621959e-01 -4.02761370e-01 -2.07837716e-01
-4.24436271e-01 -2.44040802e-01 -1.05247986e+00 1.16758573e+00
2.43225589e-01 8.82799447e-01 -7.10424408e-02 -4.82209474e-01
1.43857503e+00 1.86821431e-01 9.78379324e-02 -9.54260826e-01
-4.75079268e-02 5.79183221e-01 -2.88444329e-02 -1.05643141e+00
5.87333977e-01 -1.38928831e-01 3.43571641e-02 8.34804103e-02
-5.03019392e-01 -2.77335823e-01 1.24510437e-01 2.77356327e-01
5.32796979e-01 -1.20941646e-01 1.22074522e-01 -6.71868026e-01
3.75105441e-01 1.66911513e-01 3.75136495e-01 7.37027586e-01
-2.41733089e-01 9.07096922e-01 4.15968120e-01 -3.97358716e-01
-1.02278113e+00 -1.15439153e+00 -6.70399442e-02 6.67443573e-01
6.51545644e-01 3.27560902e-01 -8.94906461e-01 -4.33624595e-01
-1.08203813e-01 8.11789393e-01 -5.85213959e-01 -4.67897989e-02
-5.09081781e-01 -8.88319731e-01 5.99147439e-01 5.88398218e-01
7.10436583e-01 -1.39011872e+00 -1.52492797e+00 7.01377839e-02
-2.67111033e-01 -7.96616137e-01 -1.27569512e-01 1.14007272e-01
-7.64552534e-01 -9.77581084e-01 -1.13406765e+00 -8.02942932e-01
5.93256891e-01 2.68401146e-01 7.99150407e-01 4.42006439e-02
-3.82699929e-02 1.51431009e-01 -4.95925426e-01 -6.63045585e-01
-2.92282104e-01 -2.39377379e-01 -4.38822567e-01 1.11999869e-01
3.09925936e-02 -4.04946581e-02 -8.05738091e-01 4.54522550e-01
-8.78466904e-01 1.30722240e-01 6.16358280e-01 3.32203716e-01
1.86635271e-01 -1.54730186e-01 4.31819111e-01 -4.53102440e-01
4.83857214e-01 -4.13045287e-01 -6.16165936e-01 2.60241240e-01
-4.18204933e-01 1.27226025e-01 -1.43337488e-01 2.01544374e-01
-1.02803326e+00 -1.17977269e-01 -1.37935430e-01 -7.82725215e-02
1.10027485e-01 7.85191655e-01 -2.99008377e-02 -4.28465195e-02
7.68659711e-01 4.70754713e-01 -1.15536712e-01 -3.83048981e-01
-2.81072855e-01 9.70382869e-01 8.03614140e-01 -3.24066095e-02
5.09628177e-01 7.27698088e-01 -1.29554514e-02 -8.99394035e-01
-6.93946600e-01 -6.41402066e-01 -2.52628922e-01 -7.56658077e-01
8.21956694e-01 -7.57788777e-01 -5.43275297e-01 4.60184723e-01
-8.79192889e-01 2.91956551e-02 -4.03116532e-02 7.16503561e-01
-2.72144198e-01 5.46506047e-01 -2.42096707e-01 -1.24952328e+00
-7.54745126e-01 -1.04397535e+00 9.67162073e-01 6.63246512e-01
9.19080749e-02 -7.43169665e-01 -5.77614196e-02 1.28115460e-01
4.36379194e-01 4.81795847e-01 5.60960054e-01 -6.20173395e-01
-4.21085089e-01 -1.58097431e-01 -2.14872360e-01 8.71665180e-02
-1.74802572e-01 -3.34955752e-02 -1.01398396e+00 3.71730775e-02
1.51058614e-01 3.36382419e-01 1.18188596e+00 5.60329139e-01
4.84975308e-01 1.03509203e-01 -1.75388381e-01 3.83020520e-01
1.52199125e+00 6.24811351e-01 7.85646737e-01 8.79143953e-01
4.44928229e-01 6.56026244e-01 1.05669975e+00 6.74070895e-01
2.08762810e-01 2.71704912e-01 8.86982203e-01 -4.11968052e-01
-5.33829369e-02 1.45783862e-02 4.23966169e-01 4.55602348e-01
-1.75257638e-01 -1.16662070e-01 -1.29672897e+00 1.01125503e+00
-1.77679002e+00 -8.12036157e-01 -4.49266315e-01 2.08716226e+00
2.39847988e-01 3.24734539e-01 -1.48556024e-01 3.12781096e-01
9.40907598e-01 2.27202162e-01 -1.09291747e-01 -4.08728085e-02
-5.71033657e-01 -2.71717399e-01 7.40138173e-01 3.59203339e-01
-1.19595611e+00 6.66480064e-01 5.81510830e+00 5.07095397e-01
-1.31769288e+00 -2.75123063e-02 3.01247746e-01 5.55292964e-01
-3.86206150e-01 9.82578844e-02 -3.81181687e-01 6.17195666e-01
4.18307573e-01 -2.46851474e-01 -1.59322649e-01 7.00336099e-01
6.22640312e-01 -6.55236602e-01 -3.20232630e-01 5.88100791e-01
3.72218966e-01 -1.17619002e+00 5.19764796e-02 -4.30065453e-01
4.97215807e-01 2.89231986e-02 -9.11918953e-02 -2.78619289e-01
-2.01346755e-01 -6.67770326e-01 1.37783241e+00 8.16658676e-01
1.73844576e-01 -6.05274498e-01 1.34299994e+00 1.05734497e-01
-1.08076000e+00 2.77582798e-02 -1.62260070e-01 1.70620397e-01
2.19704613e-01 3.83019686e-01 -1.18354428e+00 3.43323588e-01
7.53401637e-01 2.84093529e-01 -8.37962568e-01 1.80286574e+00
-4.91019785e-01 6.82359159e-01 -1.15735501e-01 -3.28391939e-01
6.26265109e-01 -1.69169900e-04 8.15720201e-01 1.34124124e+00
3.98736924e-01 1.30044147e-01 -6.50518760e-02 7.97721326e-01
5.08092225e-01 2.43068531e-01 -3.90865058e-01 2.90559471e-01
3.97451937e-01 8.86396289e-01 -9.82512593e-01 -4.48820919e-01
-8.81102011e-02 4.02786374e-01 -5.13009369e-01 2.17509508e-01
-7.11803198e-01 -5.87854743e-01 -1.01814620e-01 2.12154031e-01
3.76240402e-01 -2.74436288e-02 -7.96065509e-01 -5.76918304e-01
4.06215191e-02 -5.01744270e-01 3.80787075e-01 -1.38077295e+00
-5.08315802e-01 4.90516365e-01 7.31963664e-02 -1.43015802e+00
2.41080478e-01 -2.05629036e-01 -6.99821532e-01 8.53618562e-01
-1.34355104e+00 -8.76089394e-01 -6.92920446e-01 9.05992910e-02
5.95723450e-01 -1.45801812e-01 3.02436024e-01 -4.46284562e-02
-4.93257381e-02 7.41045922e-02 2.45369554e-01 4.07207668e-01
1.23967797e-01 -1.26658964e+00 4.02955174e-01 1.17825961e+00
-4.13451381e-02 -4.25045975e-02 1.22201014e+00 -9.05690491e-01
-8.50290418e-01 -7.47802913e-01 6.18032157e-01 8.45178887e-02
3.91766399e-01 2.32685193e-01 -1.26717424e+00 2.31781200e-01
2.45990120e-02 -4.04541016e-01 8.46493617e-02 -6.44248426e-01
1.59384862e-01 5.92187867e-02 -1.13993526e+00 5.68147957e-01
1.78177729e-01 -9.51467231e-02 -1.12803924e+00 -4.83887084e-02
2.79046983e-01 -3.62809539e-01 -5.46065271e-01 6.72206700e-01
3.26153904e-01 -9.75284517e-01 7.99377263e-01 3.06665391e-01
3.29667330e-01 -7.82094002e-01 -2.54378617e-01 -1.22081065e+00
-4.69113886e-02 4.54748794e-02 4.03204739e-01 6.76774442e-01
5.34208179e-01 -3.67416263e-01 5.77785969e-01 1.76813737e-01
-5.09989023e-01 -2.45591223e-01 -9.03001726e-01 -4.60279256e-01
-5.45863509e-01 -5.06523885e-02 4.89159435e-01 7.02662766e-01
2.45026145e-02 -1.37408048e-01 -2.62644440e-01 6.30620301e-01
7.82180607e-01 5.08528613e-02 3.19459409e-01 -1.10400128e+00
1.11025512e-01 -3.43526214e-01 -6.36719525e-01 -3.74322414e-01
-6.17461085e-01 -4.65754688e-01 4.70481157e-01 -1.70973897e+00
3.55291873e-01 -2.16244787e-01 -4.01223958e-01 4.17198867e-01
-4.72127378e-01 1.67539462e-01 3.19273144e-01 4.26564693e-01
-3.92806411e-01 6.78195596e-01 1.10998547e+00 -2.94091433e-01
-8.43444616e-02 -1.84111428e-02 -2.45087773e-01 8.98435473e-01
8.20780516e-01 -6.14340484e-01 -1.76351995e-03 -5.05795896e-01
1.00911774e-01 2.41197124e-01 5.26260018e-01 -1.26209736e+00
2.00723350e-01 -6.76132590e-02 2.00245664e-01 -8.89854074e-01
3.63313816e-02 -4.80057091e-01 -1.03250280e-01 1.11108804e+00
-1.93055600e-01 -6.15142621e-02 -2.44438592e-02 3.99075150e-01
-5.09254217e-01 -7.12716699e-01 8.66494894e-01 -1.32125944e-01
-1.32142687e+00 -2.90114939e-01 -3.98787349e-01 -1.58044212e-02
8.88654768e-01 -3.74620736e-01 -2.12273404e-01 -2.19739750e-01
-6.36655271e-01 1.38964415e-01 3.25360119e-01 1.23892382e-01
1.11319399e+00 -8.26583207e-01 -1.05153167e+00 1.56874225e-01
3.93627673e-01 -1.07126206e-01 3.18734407e-01 8.33755136e-01
-1.32915306e+00 -1.64470911e-01 -2.13403806e-01 -7.27633595e-01
-1.20778096e+00 2.12831646e-01 5.00973582e-01 3.67504030e-01
-8.64923537e-01 6.26151562e-01 3.07016313e-01 2.73962796e-01
1.75365638e-02 -3.09279114e-01 -9.00017321e-01 7.68742859e-02
4.37621295e-01 4.59701627e-01 3.18609993e-04 -7.39088297e-01
-5.63187718e-01 6.14802241e-01 4.03047681e-01 -5.39825499e-01
1.19432640e+00 4.95962501e-02 5.49733676e-02 4.82430339e-01
4.23185319e-01 1.32539719e-01 -1.26435864e+00 -2.29163915e-01
3.65212172e-01 -5.54761052e-01 1.02560796e-01 -7.14721143e-01
-9.50987756e-01 9.19352293e-01 1.04488480e+00 1.75539032e-01
9.63377535e-01 1.50165901e-01 4.81231719e-01 6.28779978e-02
2.01922819e-01 -1.13538253e+00 -1.53001234e-01 3.84960353e-01
1.03544235e+00 -1.14105153e+00 7.60636572e-03 -1.94575205e-01
-1.15734529e+00 1.00943792e+00 2.61369765e-01 -1.46904692e-01
4.59714711e-01 1.99129000e-01 3.89452964e-01 -5.60189724e-01
-6.07693084e-02 -2.41227165e-01 1.71380207e-01 3.01740468e-01
5.30818403e-02 2.92452097e-01 -6.72584713e-01 4.85713691e-01
-1.47712424e-01 -3.34232062e-01 7.51597524e-01 1.09786677e+00
-1.39176369e+00 2.46965885e-03 -1.01115870e+00 1.43514618e-01
-2.29499266e-01 -1.23862118e-01 -1.95734873e-01 5.29833376e-01
3.25537175e-02 7.70586669e-01 9.62862223e-02 -1.38283491e-01
4.35348779e-01 -3.19482177e-01 -2.03731552e-01 -2.85340667e-01
-5.23990393e-01 1.88911781e-01 -8.91870931e-02 1.98375687e-01
-3.11356157e-01 -8.00281703e-01 -1.77842319e+00 2.48159468e-01
-1.42499581e-01 6.07758224e-01 8.77057195e-01 1.00628495e+00
-1.34531572e-01 4.49262232e-01 7.40512908e-01 -7.86867797e-01
-2.42950946e-01 -1.22793448e+00 -8.13657105e-01 5.69294333e-01
1.74917758e-01 -7.23440051e-01 -5.54124177e-01 -3.79585568e-03] | [9.46243953704834, -0.626992404460907] |
94207fd0-be29-47c4-a5e2-e6d9b6fea489 | recovering-aes-keys-with-a-deep-cold-boot | 2106.04876 | null | https://arxiv.org/abs/2106.04876v1 | https://arxiv.org/pdf/2106.04876v1.pdf | Recovering AES Keys with a Deep Cold Boot Attack | Cold boot attacks inspect the corrupted random access memory soon after the power has been shut down. While most of the bits have been corrupted, many bits, at random locations, have not. Since the keys in many encryption schemes are being expanded in memory into longer keys with fixed redundancies, the keys can often be restored. In this work, we combine a novel cryptographic variant of a deep error correcting code technique with a modified SAT solver scheme to apply the attack on AES keys. Even though AES consists of Rijndael S-box elements, that are specifically designed to be resistant to linear and differential cryptanalysis, our method provides a novel formalization of the AES key scheduling as a computational graph, which is implemented by a neural message passing network. Our results show that our methods outperform the state of the art attack methods by a very large margin. | ['Lior Wolf', 'Eliya Nachmani', 'Itamar Zimerman'] | 2021-06-09 | null | null | null | null | ['cryptanalysis'] | ['miscellaneous'] | [ 1.97791785e-01 -2.51634289e-02 -1.82609558e-01 1.31929040e-01
-5.17051935e-01 -9.76273298e-01 5.29948957e-02 3.49038869e-01
-5.13863564e-01 5.69845259e-01 -2.89617926e-01 -1.17437351e+00
1.84474722e-01 -1.13898325e+00 -8.90952349e-01 -7.98285723e-01
-3.91275495e-01 8.90681967e-02 3.04750532e-01 -4.96307909e-01
4.61231768e-01 4.44740206e-01 -1.15572584e+00 1.72695756e-01
3.23729128e-01 6.37719870e-01 -6.20402634e-01 9.28068757e-01
2.73722857e-01 5.19681513e-01 -7.73464739e-01 -4.49821174e-01
6.23523533e-01 -3.24760526e-01 -9.58680511e-01 -7.73689806e-01
1.13869287e-01 -4.91127193e-01 -6.56919122e-01 1.16005898e+00
2.41531909e-01 -3.39244694e-01 1.96222980e-02 -1.34344840e+00
6.12619594e-02 1.07103288e+00 -3.92749250e-01 -9.41942036e-02
4.15795207e-01 2.23065093e-01 9.01249349e-01 -1.30167425e-01
3.89233023e-01 7.69602478e-01 7.59808898e-01 4.41358864e-01
-1.23868608e+00 -1.14138150e+00 -2.59069085e-01 2.59526581e-01
-1.83838463e+00 -2.64248461e-01 4.14539129e-01 2.89626271e-01
1.45265532e+00 5.89393556e-01 4.81663376e-01 5.32887459e-01
1.16080451e+00 1.06314257e-01 1.05165803e+00 -5.57832599e-01
4.49576348e-01 -1.48768663e-01 5.31186819e-01 5.41942835e-01
1.10178530e+00 4.58903164e-01 -2.65001267e-01 -7.73748457e-01
-1.05849080e-01 5.89815248e-03 -4.62011218e-01 -2.59147525e-01
-1.05402386e+00 7.81610906e-01 3.75098556e-01 2.93813825e-01
-9.68257990e-03 8.90496612e-01 5.12877643e-01 5.85282087e-01
-3.27756703e-01 4.75565791e-01 -6.44484878e-01 -2.31369142e-03
-1.00609529e+00 2.11363778e-01 1.40445411e+00 4.36518282e-01
9.12200212e-01 -8.95553455e-02 2.81220913e-01 -6.75773859e-01
3.46023589e-01 6.66389287e-01 3.17158908e-01 -1.43286705e-01
1.42693192e-01 5.20137370e-01 -2.87178755e-01 -9.20276880e-01
-5.20137429e-01 -3.08618754e-01 -1.10891557e+00 5.33255756e-01
4.10281956e-01 -2.54514217e-01 -7.21391916e-01 1.74705350e+00
4.74081375e-02 4.03083354e-01 3.38300526e-01 4.63765353e-01
4.71814051e-02 9.25501168e-01 -2.78633475e-01 -1.05153218e-01
1.58164394e+00 -4.88557547e-01 -6.79766536e-01 -3.24042022e-01
1.00766075e+00 -6.48992717e-01 3.13198894e-01 5.46500087e-01
-1.17624664e+00 -2.60924011e-01 -2.01994109e+00 1.57116428e-01
-4.22176629e-01 -5.57591140e-01 7.72155046e-01 1.48694646e+00
-1.25224149e+00 4.20475721e-01 -8.82238150e-01 4.01623309e-01
2.14226283e-02 9.99087930e-01 -4.97312069e-01 1.24887265e-01
-1.35398543e+00 7.29340434e-01 5.45908213e-01 8.76799598e-02
-7.31317759e-01 -6.99252903e-01 -9.26826298e-01 3.54579836e-01
4.19506967e-01 -4.74159151e-01 1.03034902e+00 -5.27672052e-01
-1.35819936e+00 5.77482879e-01 6.48006275e-02 -9.18648720e-01
-1.02756254e-01 1.65175065e-01 -4.69282717e-01 -1.37830347e-01
-6.39379561e-01 -4.76838499e-02 7.23429203e-01 -8.75628471e-01
-3.39141786e-01 -2.98869073e-01 2.65107512e-01 -7.28708565e-01
-1.52638882e-01 -8.28100741e-03 3.14260535e-02 -4.78161752e-01
-1.13428049e-01 -1.36719632e+00 -6.03352010e-01 -4.54898655e-01
-4.76510912e-01 4.98603165e-01 7.32653320e-01 -2.55158633e-01
1.64621305e+00 -2.20300198e+00 6.48320690e-02 1.09367085e+00
3.69478613e-01 4.77393121e-01 1.66914895e-01 7.46103525e-01
-2.69223332e-01 1.67087331e-01 -3.28604281e-01 7.23240227e-02
8.15995932e-02 1.67757019e-01 -8.33817959e-01 9.22797620e-01
-2.47335374e-01 8.71140122e-01 -6.09301925e-01 2.86250859e-01
-2.80267626e-01 4.04932320e-01 -5.35632133e-01 -1.77517027e-01
-1.52767375e-01 -2.55463511e-01 -2.97493249e-01 3.00923318e-01
1.43947613e+00 -1.67177424e-01 8.09042752e-01 -9.43802223e-02
-1.50826918e-02 5.77818692e-01 -1.56246936e+00 1.45166337e+00
-4.19130445e-01 4.89741653e-01 4.06929739e-02 -5.36673903e-01
4.19967741e-01 2.53634095e-01 -4.34151925e-02 -5.60131848e-01
4.60525960e-01 3.34409088e-01 3.07127476e-01 5.13535976e-01
9.53592002e-01 5.66635951e-02 -7.43290901e-01 1.19826841e+00
-6.59419119e-01 -2.79958156e-04 -1.31642282e-01 4.61229950e-01
1.65160918e+00 -3.55549395e-01 3.32421333e-01 -3.21980178e-01
8.01919699e-01 5.10281622e-02 5.56793511e-01 8.27055216e-01
2.14827210e-01 1.37319058e-01 8.45052123e-01 -4.04223144e-01
-8.83940756e-01 -7.18601525e-01 -1.02126546e-01 3.90427351e-01
3.09163302e-01 -1.21539247e+00 -1.06533802e+00 -7.28331745e-01
-7.72351697e-02 7.30446637e-01 -4.21537697e-01 -7.95608580e-01
-7.57210553e-01 -8.00132930e-01 1.20656705e+00 5.41294336e-01
3.52135450e-01 -5.40879309e-01 -9.03946161e-01 2.20046327e-01
5.41733742e-01 -8.71897936e-01 -4.40891087e-01 9.16016042e-01
-5.23177564e-01 -1.32762754e+00 2.51934946e-01 -5.47332764e-01
7.26472437e-01 2.64064133e-01 8.59218180e-01 8.42830241e-01
-4.89608228e-01 6.15084209e-02 -2.36973912e-01 -9.03110802e-02
-1.00099301e+00 9.33797061e-02 -6.72180504e-02 -2.95041203e-01
5.37173271e-01 -3.31259608e-01 -5.31692863e-01 2.72911847e-01
-1.32392514e+00 -2.53167957e-01 4.58513916e-01 9.31515872e-01
3.74832183e-01 8.52304161e-01 -2.25082770e-01 -1.04520833e+00
2.55689651e-01 -2.35863626e-01 -1.11243486e+00 -1.45415664e-02
-8.17589045e-01 5.33320487e-01 8.25851500e-01 -2.55953103e-01
-5.12869120e-01 2.63090253e-01 -1.32800475e-01 2.47451603e-01
3.17994624e-01 2.70817369e-01 -4.27988499e-01 -7.19757855e-01
8.25895548e-01 1.36924729e-01 -2.05740123e-03 2.36689508e-01
2.68179595e-01 4.76071298e-01 5.03450096e-01 -5.14694631e-01
1.36098969e+00 4.03571904e-01 5.44252634e-01 -5.56141771e-02
-1.12061305e-02 1.76457569e-01 -2.64780611e-01 5.09706676e-01
3.86263728e-01 -6.95828080e-01 -1.37667048e+00 7.19186127e-01
-9.35364485e-01 -5.44751167e-01 -9.59013402e-03 -7.01939641e-03
5.10608777e-02 7.73966134e-01 -7.45372832e-01 -5.15177250e-01
-5.38650393e-01 -1.44692731e+00 8.34561527e-01 7.07318932e-02
-3.45337182e-01 -6.30661964e-01 1.10560030e-01 -3.27574700e-01
4.60228801e-01 1.20951362e-01 1.12312198e+00 -8.04191649e-01
-9.48365688e-01 -7.52072871e-01 9.65067968e-02 2.95133621e-01
-9.46328267e-02 -1.60093278e-01 -7.86306798e-01 -9.43167150e-01
7.95777589e-02 4.00007814e-02 6.91000044e-01 -2.35118419e-01
1.18717456e+00 -2.02735230e-01 -5.35244584e-01 9.19120789e-01
1.46948934e+00 1.35967880e-01 1.15045846e+00 5.59522569e-01
1.47720128e-01 -2.00418606e-01 1.92424864e-01 5.75261116e-01
2.28332534e-01 4.98976141e-01 6.71471238e-01 6.69440851e-02
3.35019946e-01 1.61568150e-01 7.50284255e-01 3.66090596e-01
4.10084754e-01 -2.70928472e-01 -8.06989014e-01 7.42970780e-03
-1.34481382e+00 -6.17336094e-01 -4.65331495e-01 2.65451121e+00
1.10581172e+00 4.67589587e-01 -4.21355277e-01 7.51291513e-01
2.54731685e-01 1.01518340e-01 -6.08201809e-02 -8.43252480e-01
-9.61965546e-02 1.37009490e+00 1.57076156e+00 5.85244358e-01
-7.98601389e-01 7.79387712e-01 6.36305857e+00 7.59595096e-01
-1.09799206e+00 -1.35958403e-01 2.54139692e-01 2.48034373e-01
-4.14496541e-01 7.88860142e-01 -9.92305756e-01 4.25232649e-01
1.49713528e+00 -2.96848685e-01 5.34605622e-01 4.72673118e-01
-4.85219568e-01 -3.01308870e-01 -1.27499902e+00 5.87553144e-01
-1.22340560e-01 -1.17489231e+00 -4.78955299e-01 3.77276301e-01
4.63383645e-01 -6.34709835e-01 2.34431073e-01 1.19780339e-01
2.61143833e-01 -1.09336591e+00 4.24050003e-01 1.63407266e-01
7.11402535e-01 -1.26516986e+00 1.02788746e+00 -6.28932491e-02
-9.04827893e-01 -8.22983980e-02 -2.59380311e-01 -2.18446553e-01
-1.73840344e-01 4.02092695e-01 -5.60707748e-01 5.77475905e-01
5.08915365e-01 -1.37291282e-01 -4.97985363e-01 7.26500154e-01
-5.78530133e-01 7.28270888e-01 -3.76692176e-01 -1.10413186e-01
1.70650497e-01 2.89408743e-01 2.79147685e-01 1.07354879e+00
3.76636863e-01 8.17791522e-02 -1.27757505e-01 4.84778821e-01
-8.55500549e-02 -2.21686006e-01 -2.50385523e-01 1.01992518e-01
4.63732570e-01 1.20383286e+00 -7.70673275e-01 -1.17719680e-01
-3.92556131e-01 9.01272595e-01 -3.19069892e-01 1.41585674e-02
-7.79284835e-01 -7.90107012e-01 9.01092708e-01 2.16317885e-02
3.48075211e-01 -6.15282536e-01 -4.21547920e-01 -9.11654711e-01
-1.49688780e-01 -1.39779806e+00 3.60283345e-01 -2.94447690e-01
-6.60157263e-01 5.10769308e-01 -2.29328513e-01 -8.30614686e-01
-2.78149843e-01 -6.97326839e-01 -5.45273721e-01 9.00749803e-01
-1.51023901e+00 -6.70310915e-01 4.49616695e-03 6.40883625e-01
-5.33262014e-01 -9.54412892e-02 1.31431735e+00 9.00886357e-02
-4.85149920e-01 1.20065987e+00 2.02152789e-01 -7.34727010e-02
7.87035644e-01 -1.00747347e+00 9.97434914e-01 1.27124798e+00
-7.92821869e-02 1.17798126e+00 6.57991588e-01 -7.62518346e-01
-2.56432343e+00 -5.66262007e-01 6.05798781e-01 1.11905873e-01
6.98784828e-01 -6.99868202e-01 -9.86871779e-01 8.48574281e-01
3.66837561e-01 -7.97941089e-02 8.07929039e-01 -3.75554681e-01
-6.93875253e-01 -1.48003757e-01 -9.82933819e-01 5.19458115e-01
2.55438179e-01 -4.47187781e-01 -2.46997207e-01 1.00616768e-01
4.92068440e-01 -8.07844758e-01 -4.58767295e-01 1.46117359e-01
4.24965709e-01 -7.79103458e-01 9.15877759e-01 -3.18680227e-01
-1.48718610e-01 -7.21380353e-01 -2.37058833e-01 -9.20132577e-01
-1.21972442e-01 -1.26684940e+00 -5.75564027e-01 6.53925896e-01
2.48093247e-01 -8.81465137e-01 9.85879064e-01 4.68109965e-01
1.23985149e-01 -9.71218571e-02 -1.06802475e+00 -5.41652501e-01
1.13763906e-01 -5.85781634e-01 1.15563238e+00 6.38122499e-01
1.96983695e-01 6.80056587e-02 -4.66275781e-01 6.13257527e-01
4.97285277e-01 1.04704043e-02 9.25413072e-01 -1.06691206e+00
-6.40866280e-01 -2.96167195e-01 -6.89371407e-01 -7.14442134e-01
4.03899521e-01 -1.04844809e+00 -2.39704445e-01 -6.37382746e-01
1.05171673e-01 -5.66581309e-01 -4.37404633e-01 8.71369302e-01
-3.95012610e-02 4.18417275e-01 -1.95604697e-01 -4.47601855e-01
-6.66503683e-02 -2.06494048e-01 4.15809840e-01 -2.91256309e-01
-1.03462011e-01 1.40837163e-01 -8.06268632e-01 2.51691729e-01
7.14048088e-01 -6.70992315e-01 -2.07063958e-01 1.22068025e-01
1.10005236e+00 6.59437776e-02 2.78743595e-01 -1.05872750e+00
7.00924516e-01 1.79771557e-01 4.33304347e-02 -3.82555574e-01
7.39519745e-02 -1.19598567e+00 5.58204234e-01 1.19461215e+00
6.61906153e-02 4.09385055e-01 6.43408239e-01 3.66304040e-01
5.90304211e-02 -2.87052900e-01 5.44949234e-01 4.09284204e-01
-5.21891594e-01 1.54456615e-01 -7.28702188e-01 -2.58580267e-01
1.15294433e+00 -2.27948233e-01 -6.00171268e-01 -2.37062559e-01
-2.48491153e-01 -1.10319946e-02 9.27128553e-01 4.95382138e-02
5.30281305e-01 -9.50325906e-01 -4.57904428e-01 7.54074991e-01
5.40400706e-02 -2.51091212e-01 2.29308978e-01 6.16329551e-01
-9.78897631e-01 3.62079740e-01 -1.26298428e-01 -3.07555087e-02
-1.39292896e+00 8.91028106e-01 2.82480210e-01 -4.57518995e-01
-6.14276946e-01 5.97152293e-01 -1.32183582e-01 1.40329346e-01
9.70951915e-02 -3.16614896e-01 5.83516300e-01 -5.12743294e-01
9.23784494e-01 8.05286616e-02 5.04623950e-01 -2.96411991e-01
-5.91418028e-01 5.07547557e-01 -2.43906572e-01 1.47907257e-01
1.02289176e+00 1.78554773e-01 -8.07493269e-01 -4.61368114e-01
1.08210599e+00 6.55216157e-01 -3.06708813e-01 -1.70698598e-01
-1.35094687e-01 -3.77207190e-01 4.11207736e-01 -5.21195173e-01
-1.09576714e+00 5.76166868e-01 5.92612267e-01 2.18251377e-01
1.41944516e+00 -7.19797552e-01 1.24084413e+00 7.39115775e-01
7.38162220e-01 -7.21114397e-01 -4.28663731e-01 5.76583922e-01
-3.68373781e-01 -4.77897376e-01 3.93133432e-01 -3.52126002e-01
1.26188546e-01 1.38767278e+00 1.47196919e-01 -1.87018245e-01
6.50611460e-01 1.20595300e+00 -3.71240497e-01 2.01030999e-01
-5.69904268e-01 4.17499542e-01 -1.05775148e-01 3.75217438e-01
-1.63522944e-01 -9.79029983e-02 -4.62512791e-01 7.84207880e-01
-4.03486639e-01 -1.83021381e-01 1.11046243e+00 1.25397921e+00
-2.64190763e-01 -2.09314919e+00 -6.76776648e-01 -9.94573087e-02
-8.25766087e-01 -3.73863786e-01 -1.50429741e-01 8.13396513e-01
-1.79152906e-01 8.51078629e-01 -5.67477793e-02 -6.78958833e-01
-6.90187439e-02 8.66177958e-03 3.83787602e-01 -2.41305500e-01
-1.05174446e+00 -3.85734916e-01 -8.74865577e-02 -7.64399171e-01
2.77482122e-01 -2.74464816e-01 -1.62959504e+00 -9.05176997e-01
-5.67682445e-01 1.90450668e-01 8.40865016e-01 6.23028040e-01
3.96631777e-01 7.46041238e-01 8.76340628e-01 -4.62108523e-01
-6.20511889e-01 -6.90779313e-02 -7.07385063e-01 -4.22016621e-01
3.37946832e-01 -7.15632215e-02 -5.18249333e-01 -4.50833648e-01] | [5.776139259338379, 7.337114334106445] |
d439cd38-a074-47a0-b550-100b1da4166a | on-the-complementarity-between-pre-training-1 | 2209.03316 | null | https://arxiv.org/abs/2209.03316v3 | https://arxiv.org/pdf/2209.03316v3.pdf | On the Complementarity between Pre-Training and Random-Initialization for Resource-Rich Machine Translation | Pre-Training (PT) of text representations has been successfully applied to low-resource Neural Machine Translation (NMT). However, it usually fails to achieve notable gains (sometimes, even worse) on resource-rich NMT on par with its Random-Initialization (RI) counterpart. We take the first step to investigate the complementarity between PT and RI in resource-rich scenarios via two probing analyses, and find that: 1) PT improves NOT the accuracy, but the generalization by achieving flatter loss landscapes than that of RI; 2) PT improves NOT the confidence of lexical choice, but the negative diversity by assigning smoother lexical probability distributions than that of RI. Based on these insights, we propose to combine their complementarities with a model fusion algorithm that utilizes optimal transport to align neurons between PT and RI. Experiments on two resource-rich translation benchmarks, WMT'17 English-Chinese (20M) and WMT'19 English-German (36M), show that PT and RI could be nicely complementary to each other, achieving substantial improvements considering both translation accuracy, generalization, and negative diversity. Probing tools and code are released at: https://github.com/zanchangtong/PTvsRI. | ['DaCheng Tao', 'Weifeng Liu', 'Yu Cao', 'Li Shen', 'Liang Ding', 'Changtong Zan'] | 2022-09-07 | null | https://aclanthology.org/2022.coling-1.445 | https://aclanthology.org/2022.coling-1.445.pdf | coling-2022-10 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 3.80104125e-01 -2.22287308e-02 -5.55471838e-01 -8.21470395e-02
-1.14257979e+00 -5.60701489e-01 8.57142746e-01 -6.77906796e-02
-4.22143191e-01 1.02572572e+00 4.67003793e-01 -6.44201040e-01
2.11956445e-02 -5.33859313e-01 -8.49915802e-01 -6.70024872e-01
3.97927940e-01 8.97688329e-01 -3.44641030e-01 -5.35500050e-01
-1.26300678e-01 1.65946230e-01 -1.16525519e+00 4.02744383e-01
1.03540254e+00 6.78079844e-01 1.03022568e-01 1.91378057e-01
-7.78433010e-02 4.09335762e-01 -3.12040567e-01 -5.80583811e-01
3.59541655e-01 -4.58155334e-01 -9.04105783e-01 -4.73051399e-01
2.43206322e-01 2.31410768e-02 -1.31523088e-01 9.92284596e-01
7.93062806e-01 1.54572800e-01 8.14094305e-01 -8.88419509e-01
-1.12547636e+00 1.23825181e+00 -5.94462574e-01 2.08494037e-01
-2.15977579e-02 1.00099050e-01 1.10682297e+00 -1.13994980e+00
8.16865325e-01 1.23642635e+00 7.68995166e-01 5.91907322e-01
-1.43588221e+00 -6.70044899e-01 9.45342630e-02 9.76427421e-02
-1.33670712e+00 -7.76390970e-01 3.46849978e-01 -9.19241607e-02
1.41721559e+00 3.33063573e-01 2.52819806e-01 1.65938234e+00
1.37429163e-01 8.96168232e-01 1.17960310e+00 -5.05813301e-01
2.79312432e-02 1.78224102e-01 2.59682816e-02 3.47802043e-01
1.68693319e-01 2.50017703e-01 -6.24972165e-01 -8.31433833e-02
5.40241718e-01 -3.36992085e-01 -3.88901234e-01 -1.30461887e-01
-1.51464534e+00 7.43885398e-01 4.76215333e-01 6.54611826e-01
-4.25613940e-01 2.80021504e-02 4.20847893e-01 7.02870965e-01
6.78648591e-01 6.87255979e-01 -5.91490567e-01 -3.05097252e-01
-9.24321115e-01 2.54842099e-02 5.43950915e-01 1.00168395e+00
6.88762546e-01 1.90900624e-01 -4.59044546e-01 1.06998622e+00
-1.63873151e-01 6.92321837e-01 9.53049183e-01 -4.08491969e-01
9.03512418e-01 3.75982404e-01 4.88410480e-02 -5.19336581e-01
-5.03434718e-01 -8.37145686e-01 -1.08827376e+00 -3.45121950e-01
3.36786866e-01 -2.34347224e-01 -8.54420602e-01 2.10946822e+00
-1.73788995e-01 -1.62171587e-01 1.69969127e-01 8.49936485e-01
7.25028634e-01 8.32741380e-01 -7.75306597e-02 -1.79920673e-01
1.08867788e+00 -1.02650654e+00 -5.88672161e-01 -4.78952825e-01
8.19739938e-01 -8.23172569e-01 1.37594616e+00 -8.53034779e-02
-1.25712132e+00 -4.78907526e-01 -9.00124550e-01 -4.65197116e-02
-3.78556460e-01 4.30532634e-01 4.24321026e-01 5.95639944e-01
-1.25515950e+00 8.48833919e-01 -7.41167605e-01 -5.35973191e-01
4.84472752e-01 3.31870079e-01 -2.60542750e-01 -1.25572741e-01
-1.42897487e+00 1.27007306e+00 2.28496194e-01 2.34095082e-02
-6.16684556e-01 -7.87189841e-01 -6.02461874e-01 1.35655656e-01
1.68727607e-01 -8.77259612e-01 1.11606717e+00 -1.27205372e+00
-1.79445314e+00 8.05158675e-01 -3.05159509e-01 -5.04840374e-01
7.49992907e-01 -4.36529070e-01 -1.79610744e-01 -4.11770493e-01
5.63385934e-02 8.27174127e-01 5.12798071e-01 -1.10629284e+00
-2.15247422e-01 -2.30927721e-01 -1.42675102e-01 5.02038538e-01
-4.02378857e-01 1.03267007e-01 -2.68842965e-01 -7.62031913e-01
-1.01701409e-01 -1.00289309e+00 -2.16494426e-01 -6.33050740e-01
-5.92520416e-01 -1.14436440e-01 9.31552574e-02 -6.86919570e-01
1.06963170e+00 -1.88979828e+00 5.90711057e-01 -4.54450212e-02
-4.34336029e-02 3.89243692e-01 -6.43123627e-01 5.35472274e-01
-1.10076979e-01 2.79105872e-01 -2.04733118e-01 -5.10389447e-01
1.23108029e-01 9.48747247e-02 -3.60678285e-01 4.37542409e-01
6.20288789e-01 1.35865867e+00 -6.92651033e-01 -1.08834662e-01
-5.90157174e-02 4.60731000e-01 -4.35327321e-01 -2.70965934e-01
-3.16037893e-01 5.13789594e-01 -2.20597520e-01 6.15963399e-01
4.44591671e-01 -4.57431465e-01 4.07734454e-01 3.64126600e-02
7.96331912e-02 7.15999901e-01 -6.55937076e-01 1.69202042e+00
-7.01371372e-01 7.36568332e-01 -2.33390480e-01 -9.27824080e-01
9.59077716e-01 3.10610235e-01 2.49570578e-01 -1.13249087e+00
2.18412504e-01 5.03496110e-01 1.79088876e-01 6.91210711e-03
7.28925765e-01 -1.30557671e-01 -4.29854868e-03 7.76951849e-01
9.27242562e-02 3.37199956e-01 -7.11347312e-02 -1.60111651e-01
8.22050571e-01 3.79780293e-01 1.15033023e-01 -4.61410165e-01
2.05253899e-01 -2.00221557e-02 5.25401473e-01 6.30681872e-01
1.16922684e-01 6.26584232e-01 3.05667073e-01 -1.56796128e-01
-1.17052841e+00 -1.03225982e+00 -5.28982878e-02 1.31171405e+00
3.58642898e-02 -1.89217001e-01 -7.21824527e-01 -6.00207210e-01
-1.48325771e-01 7.98171580e-01 -5.76638699e-01 -4.30279821e-01
-9.32090163e-01 -1.14200842e+00 7.00878859e-01 5.29460669e-01
3.70899439e-01 -9.66380477e-01 -3.21648628e-01 1.15699194e-01
-6.12099707e-01 -9.52389181e-01 -4.87935722e-01 5.54780900e-01
-8.36245060e-01 -3.58850151e-01 -9.48642135e-01 -6.65351689e-01
4.71250743e-01 2.73505718e-01 1.49559999e+00 1.74987763e-02
3.37738216e-01 -8.87490958e-02 -5.30369222e-01 -2.06846699e-01
-6.49023890e-01 8.43635380e-01 3.72760415e-01 -2.41912186e-01
2.41096795e-01 -6.30925119e-01 -2.62675941e-01 5.71607292e-01
-5.33393204e-01 2.70333081e-01 8.44163358e-01 1.13964105e+00
5.32998323e-01 -4.31114316e-01 8.58181059e-01 -7.35006988e-01
6.73939049e-01 -6.36592507e-01 -1.98242679e-01 4.82879996e-01
-7.97661483e-01 1.07406452e-01 8.13298404e-01 -6.55864656e-01
-8.52234185e-01 -4.59487826e-01 -1.22044183e-01 -5.05877495e-01
2.20382839e-01 5.62871635e-01 -1.37129009e-01 3.30057621e-01
8.75582874e-01 4.01098102e-01 -7.17511401e-02 -6.55935407e-01
5.55781543e-01 5.78604996e-01 2.56233841e-01 -7.85612822e-01
7.33179212e-01 1.27964355e-02 -3.52053761e-01 -4.94917989e-01
-4.00079697e-01 1.20611982e-02 -6.27809763e-01 2.76535392e-01
5.05931497e-01 -9.42507267e-01 -1.01091743e-01 1.94204941e-01
-9.89918947e-01 -6.85859740e-01 -3.45908195e-01 5.14782786e-01
-6.04423404e-01 8.95983875e-02 -6.29627824e-01 -3.95393312e-01
-6.49217784e-01 -1.24832547e+00 1.00159895e+00 -6.96887597e-02
-3.87790531e-01 -9.90474403e-01 -3.45813259e-02 3.57259274e-01
8.56091022e-01 -1.83252782e-01 1.30137038e+00 -9.55581784e-01
-2.71993816e-01 4.93104197e-03 -3.58856022e-01 2.32209072e-01
2.18637154e-01 -2.20561683e-01 -9.38732564e-01 -5.73554099e-01
-2.50287831e-01 -3.45384598e-01 9.41756487e-01 4.73354995e-01
5.74090123e-01 -4.46147323e-01 -2.31252655e-01 7.10058510e-01
1.27031434e+00 -1.15154833e-01 5.45238078e-01 4.14646327e-01
5.91252089e-01 4.22058463e-01 4.58054721e-01 6.64512143e-02
2.81739771e-01 1.14073515e+00 1.00291215e-01 -2.47622997e-01
-4.58108574e-01 -2.58972228e-01 7.70807564e-01 1.15742707e+00
-3.23364258e-01 -3.65871489e-01 -9.89047289e-01 3.84015083e-01
-1.86425793e+00 -6.38004363e-01 1.59395620e-01 2.26990318e+00
1.17033911e+00 8.79541039e-02 1.22650541e-01 -1.99101076e-01
6.88452780e-01 6.10966468e-03 -6.09581828e-01 -6.14184380e-01
-4.97847289e-01 3.27386171e-01 5.93048811e-01 3.86592478e-01
-7.70394921e-01 1.18706417e+00 6.02860546e+00 1.14089072e+00
-1.26227963e+00 4.91788596e-01 9.19479668e-01 -3.77221555e-01
-6.28721833e-01 -1.63471058e-01 -8.63208473e-01 3.91705066e-01
1.22740853e+00 -2.93652743e-01 7.79183030e-01 4.62427020e-01
6.90190792e-02 5.37830651e-01 -1.22004414e+00 7.10698664e-01
-3.43347751e-02 -1.59738660e+00 3.52440804e-01 1.15846716e-01
8.77523065e-01 6.12730861e-01 3.55583757e-01 6.02960825e-01
5.21008432e-01 -1.34901929e+00 9.05909777e-01 2.58077145e-01
1.03778625e+00 -7.24343121e-01 8.39387774e-01 3.66728276e-01
-7.53187954e-01 1.43116906e-01 -5.30238926e-01 2.60844524e-03
-7.48946443e-02 3.96295339e-01 -7.59598911e-01 9.60722983e-01
4.81698751e-01 6.78610682e-01 -5.31033754e-01 3.84286135e-01
-3.77660766e-02 5.81013262e-01 -2.45311454e-01 -9.26165432e-02
3.77896935e-01 -2.08742201e-01 5.96268833e-01 1.30129969e+00
4.13451552e-01 -5.13846993e-01 -2.12963205e-02 9.70229328e-01
-4.86355931e-01 4.63666886e-01 -5.83870113e-01 -1.33821338e-01
7.31042325e-01 8.46998394e-01 -4.77882594e-01 -3.72039974e-01
-2.24149778e-01 8.57166767e-01 5.30631781e-01 4.45673883e-01
-9.64536667e-01 8.12816992e-02 8.97246301e-01 -1.14985518e-01
2.63909191e-01 2.57732277e-03 -5.28721273e-01 -1.27806330e+00
1.30173430e-01 -1.28141487e+00 4.44891900e-02 -4.82609838e-01
-1.29139972e+00 8.84257197e-01 -2.66867161e-01 -1.24568582e+00
-3.39708149e-01 -5.69387972e-01 -4.05656546e-01 1.24251091e+00
-1.49389029e+00 -1.31717002e+00 3.47564995e-01 3.13529640e-01
7.36624002e-01 -3.51450622e-01 9.02605772e-01 3.36909026e-01
-8.94555748e-01 1.18631899e+00 4.80765551e-01 -8.04849248e-03
6.43483341e-01 -8.87983561e-01 8.52050424e-01 7.44822085e-01
4.09267128e-01 6.57208681e-01 5.61255991e-01 -4.65646207e-01
-1.53135252e+00 -1.20991778e+00 1.03007686e+00 -6.90215170e-01
5.15950263e-01 -3.20680350e-01 -8.32056820e-01 6.57937288e-01
4.00137454e-01 -4.94589239e-01 5.13766110e-01 3.34111750e-01
-4.80493248e-01 2.33185798e-01 -8.79768848e-01 8.91414464e-01
1.18817115e+00 -4.61484909e-01 -3.38276625e-01 4.91775513e-01
9.45587754e-01 -3.77627909e-01 -9.57838058e-01 4.79620159e-01
4.79104668e-01 -7.39912927e-01 9.95222628e-01 -8.35305512e-01
6.64292634e-01 2.23974958e-01 -5.29988348e-01 -1.67824960e+00
-4.01764750e-01 -5.68676114e-01 2.58313686e-01 1.30195856e+00
9.74921882e-01 -9.34836984e-01 5.08602679e-01 -1.24627361e-02
-2.54398316e-01 -1.11966872e+00 -1.17799425e+00 -1.12303174e+00
9.03062880e-01 -2.52138257e-01 8.60696673e-01 1.19170249e+00
-7.71865100e-02 6.06350780e-01 -5.04074395e-01 -2.25306466e-01
4.11313735e-02 1.27675429e-01 5.92552602e-01 -8.95024717e-01
-4.28189784e-01 -9.29792643e-01 6.88627511e-02 -9.45693791e-01
1.93643004e-01 -1.49422026e+00 -1.04486933e-02 -1.49894881e+00
2.68298000e-01 -5.23029566e-01 -4.27709788e-01 7.30376601e-01
-2.74728179e-01 3.09437513e-01 1.65746927e-01 5.48613131e-01
-2.35649914e-01 6.55962348e-01 1.03819644e+00 -1.74967974e-01
-2.15085685e-01 -6.79553300e-02 -8.19742024e-01 3.01160783e-01
1.03792214e+00 -3.75096679e-01 -2.57210463e-01 -9.61885214e-01
4.60222036e-01 -5.12599684e-02 2.80280635e-02 -7.70732045e-01
-1.98997945e-01 -1.82161592e-02 2.48969972e-01 -1.50931895e-01
2.64802843e-01 -3.71024311e-01 1.74196124e-01 4.65056300e-01
-5.64479470e-01 4.23668891e-01 2.92232603e-01 2.13312969e-01
4.98879701e-02 -5.64815383e-03 7.03864932e-01 -7.78001174e-02
-2.62157917e-01 1.24213494e-01 -3.19977582e-01 8.87900963e-02
2.72373885e-01 -5.21927932e-03 -6.35632515e-01 -2.67189056e-01
-2.72521436e-01 8.07849914e-02 6.57057703e-01 8.17183077e-01
2.70167112e-01 -1.48294342e+00 -1.16618252e+00 2.12715074e-01
7.33183771e-02 -4.15377438e-01 -5.47383651e-02 1.24897456e+00
-5.21782972e-02 5.42960167e-01 -2.12784410e-01 -5.45090973e-01
-8.78212929e-01 4.02810574e-01 3.88986290e-01 -6.12341464e-01
-6.20850682e-01 8.03460777e-01 8.31784606e-02 -8.33939135e-01
5.39146811e-02 -3.60016376e-01 -1.52327804e-04 5.13161086e-02
1.63611993e-01 3.16009581e-01 3.19996923e-01 -7.30373025e-01
-3.68351936e-01 4.83283669e-01 -2.14554280e-01 -1.57003284e-01
1.30294037e+00 -1.73847958e-01 1.64299365e-02 4.40861315e-01
1.06883037e+00 -6.06230982e-02 -8.68913352e-01 -6.40618622e-01
1.28530815e-01 -2.88964689e-01 -1.63102835e-01 -1.09123695e+00
-1.10120380e+00 9.74844694e-01 4.70529497e-01 -1.47542059e-01
9.52640891e-01 -1.15871549e-01 8.37342024e-01 4.01005924e-01
4.22121495e-01 -9.04500425e-01 -1.83235869e-01 6.33321047e-01
9.35777843e-01 -1.13578749e+00 -2.17372239e-01 7.46055506e-03
-8.09188485e-01 7.61423051e-01 2.15865597e-01 1.67316571e-01
1.97840333e-01 9.19400454e-02 1.61875397e-01 2.45300636e-01
-1.11891556e+00 -1.75717652e-01 2.65243053e-01 5.41402578e-01
6.91402733e-01 2.29938552e-01 -1.70407459e-01 6.95367873e-01
-4.60507751e-01 -2.30707526e-01 1.75316364e-01 4.52589005e-01
-2.13053912e-01 -1.16729379e+00 -4.65206653e-02 4.30766165e-01
-4.49610174e-01 -6.33455515e-01 -5.36023021e-01 7.27245212e-01
-1.15934663e-01 8.18372667e-01 -1.58135518e-02 -6.74590945e-01
2.14853778e-01 2.67051607e-01 3.77829075e-01 -4.70282495e-01
-7.82544255e-01 5.31948991e-02 4.01295394e-01 -4.47650611e-01
-3.11410744e-02 -6.49913013e-01 -8.04026365e-01 -5.50944746e-01
-4.11963075e-01 4.59898002e-02 6.49051666e-01 8.87931883e-01
5.73386908e-01 5.00152171e-01 6.05755210e-01 -8.95492494e-01
-9.59406674e-01 -1.27456868e+00 -5.73129952e-02 2.30712384e-01
1.80504605e-01 -5.45271039e-01 -2.34628573e-01 -4.41486984e-01] | [11.639525413513184, 10.114669799804688] |
213c43a4-62e5-4a1e-b5db-bdd6a864b9e0 | character-aware-neural-morphological | null | null | https://aclanthology.org/P17-2105 | https://aclanthology.org/P17-2105.pdf | Character-Aware Neural Morphological Disambiguation | We develop a language-independent, deep learning-based approach to the task of morphological disambiguation. Guided by the intuition that the correct analysis should be {``}most similar{''} to the context, we propose dense representations for morphological analyses and surface context and a simple yet effective way of combining the two to perform disambiguation. Our approach improves on the language-dependent state of the art for two agglutinative languages (Turkish and Kazakh) and can be potentially applied to other morphologically complex languages. | ['Gulmira Tolegen', 'Alymzhan Toleu', 'Aibek Makazhanov'] | 2017-07-01 | null | null | null | acl-2017-7 | ['morphological-disambiguation'] | ['natural-language-processing'] | [-1.90918043e-01 -5.63513786e-02 1.80197552e-01 -4.11348373e-01
-6.97936714e-01 -1.03932273e+00 5.00910759e-01 7.30821729e-01
-7.28854358e-01 6.04970992e-01 3.06720674e-01 -8.93382430e-01
-7.77689219e-02 -8.81274104e-01 -2.77692258e-01 -4.90634561e-01
-1.31346270e-01 7.44524896e-01 2.66935825e-01 -7.17141449e-01
1.72959477e-01 6.82248473e-01 -1.09402740e+00 2.52378464e-01
1.01254463e+00 4.83335257e-01 2.40572050e-01 2.51419544e-01
-4.91390079e-01 4.06499863e-01 -6.76055074e-01 -7.61886299e-01
2.95075923e-02 -3.10508236e-02 -1.34122884e+00 -2.94175029e-01
6.68387353e-01 5.97749203e-02 8.83704200e-02 1.34401238e+00
4.21426058e-01 4.08852965e-01 6.50784671e-01 -1.41922593e-01
-1.25313437e+00 9.86415625e-01 -3.74301970e-01 7.41354108e-01
3.23047280e-01 -2.57707894e-01 1.35370588e+00 -1.05880070e+00
7.91545630e-01 1.57427061e+00 5.25603533e-01 5.44227898e-01
-1.07003534e+00 -1.96911305e-01 5.79547167e-01 4.15742993e-01
-1.16065204e+00 -3.88240427e-01 7.30132937e-01 -2.68915474e-01
1.42725658e+00 1.24548167e-01 3.55533451e-01 5.41057110e-01
-1.83640301e-01 7.51881599e-01 1.17911398e+00 -9.09865558e-01
1.41680941e-01 -5.27442336e-01 4.19518799e-01 6.92161143e-01
5.21726429e-01 -1.21803895e-01 -2.40872473e-01 -5.77270165e-02
5.84319830e-01 -7.85725594e-01 9.97658633e-03 1.14562772e-01
-1.00850546e+00 1.00901914e+00 4.48161066e-01 8.19812536e-01
-1.54789075e-01 -1.99690893e-01 5.30486643e-01 3.01647753e-01
2.65032291e-01 8.00485134e-01 -8.85527551e-01 3.03812355e-01
-5.65357387e-01 2.26445213e-01 7.21560657e-01 4.42033201e-01
8.69405150e-01 5.29525816e-01 2.65078008e-01 9.32792664e-01
3.62718463e-01 3.86903703e-01 4.68309700e-01 -6.80824399e-01
3.01506460e-01 6.83774769e-01 -6.81589246e-02 -4.78893280e-01
-4.79781568e-01 -7.81756192e-02 -8.10034797e-02 1.05985433e-01
8.50133717e-01 -1.98687807e-01 -9.12571013e-01 1.83927500e+00
4.56290364e-01 -2.67618328e-01 3.11764926e-01 8.04554105e-01
1.00570762e+00 4.74102229e-01 5.98439038e-01 -1.12583548e-01
1.81321096e+00 -3.89616758e-01 -7.26278603e-01 -5.31867266e-01
8.01986396e-01 -9.25141156e-01 1.23594093e+00 2.53139794e-01
-1.33913696e+00 -4.27142739e-01 -1.06613433e+00 -4.67250437e-01
-8.29192281e-01 7.53469467e-02 1.14132202e+00 7.94387698e-01
-1.40116239e+00 5.96052766e-01 -7.70675778e-01 -5.90099394e-01
6.23382889e-02 6.42988384e-01 -5.85999668e-01 3.23033720e-01
-1.39974225e+00 1.33552003e+00 9.26198542e-01 3.28956455e-01
-2.46882424e-01 -2.34771550e-01 -1.32319295e+00 -2.39308462e-01
2.33209968e-01 -3.86122644e-01 1.24448407e+00 -1.03058839e+00
-1.32628906e+00 1.63165057e+00 -1.94204688e-01 -5.00656903e-01
-3.28968972e-01 -4.81685966e-01 -5.47435522e-01 5.65370880e-02
1.41950503e-01 4.95756954e-01 3.66231859e-01 -9.78962183e-01
-4.86340582e-01 -6.32453442e-01 2.37609953e-01 3.20705622e-01
-4.66169454e-02 8.79024088e-01 -1.28419697e-01 -9.63730276e-01
3.24192196e-01 -7.49714255e-01 -2.44458169e-01 -6.47135973e-01
-1.35194704e-01 -7.60847390e-01 4.31341410e-01 -1.30586731e+00
1.10109234e+00 -2.01934433e+00 2.94013530e-01 -3.65132391e-02
-7.61200413e-02 7.38486409e-01 -8.14819261e-02 3.52111965e-01
-1.74442112e-01 3.79582524e-01 -3.72172087e-01 -6.26675412e-02
1.90929070e-01 5.60822427e-01 -1.24023102e-01 4.72942591e-01
8.28383684e-01 1.28569293e+00 -1.19549167e+00 -4.35339361e-01
3.79579723e-01 1.40588060e-01 -2.60970950e-01 -2.13870201e-02
-1.80631906e-01 2.44790822e-01 -1.46145225e-01 8.11223745e-01
6.73265934e-01 2.95158237e-01 1.02459002e+00 3.20408307e-02
-1.95552230e-01 1.14646411e+00 -1.05109704e+00 1.55556095e+00
-3.17619532e-01 4.88017023e-01 3.30110013e-01 -1.00017953e+00
1.08638060e+00 5.03074266e-02 -2.60600358e-01 -5.84229112e-01
2.81931758e-01 5.29066503e-01 3.93948436e-01 1.44331413e-03
6.13541126e-01 -2.96413809e-01 -4.69448388e-01 3.27298790e-01
3.32084417e-01 -2.61710823e-01 2.96962231e-01 -5.12464717e-02
6.92138374e-01 5.78285217e-01 1.04246259e+00 -9.35702026e-01
8.15537095e-01 -1.71332091e-01 7.94154167e-01 2.55889058e-01
-2.16356173e-01 5.50730526e-02 1.55577764e-01 -7.96885669e-01
-6.36485994e-01 -1.43433046e+00 -2.78145075e-01 1.57492077e+00
-8.06269050e-02 -4.26484495e-01 -9.01042283e-01 -8.92268777e-01
-1.21426746e-01 8.44523013e-01 -6.70701087e-01 1.95242792e-01
-1.44952822e+00 -6.95611000e-01 5.63685000e-01 8.47732425e-01
1.43268675e-01 -1.52225018e+00 -2.81488031e-01 4.49640304e-01
9.23503637e-02 -9.32527661e-01 -7.48735219e-02 5.32749593e-01
-7.67646372e-01 -9.56277788e-01 -6.93818778e-02 -1.39136517e+00
2.45165750e-01 -2.55254924e-01 1.46700764e+00 3.82437944e-01
7.88850933e-02 -1.57948472e-02 -4.16586548e-01 -5.48576951e-01
-5.02717078e-01 6.25342876e-02 1.52188137e-01 -4.41650689e-01
8.33815396e-01 -4.04125899e-01 -1.04167670e-01 -3.06212276e-01
-9.47452486e-01 -7.91801870e-01 1.26665577e-01 4.98877555e-01
5.44926226e-01 -3.26295882e-01 5.36877513e-01 -1.16590035e+00
4.55997169e-01 -1.90591171e-01 -6.04870617e-01 3.08937877e-01
-1.18962184e-01 3.77806336e-01 7.83145845e-01 -2.82957792e-01
-1.11197364e+00 2.82559488e-02 -5.36590636e-01 2.51344621e-01
-3.64456534e-01 6.61352932e-01 -7.00368404e-01 -6.34045005e-02
6.19010925e-01 -2.89145172e-01 -5.67176163e-01 -7.93145359e-01
6.95040643e-01 1.65005520e-01 8.60494792e-01 -9.54030871e-01
5.93459368e-01 3.21995020e-01 -9.59956497e-02 -7.80310154e-01
-7.29230464e-01 -2.33766645e-01 -1.22640908e+00 3.68622839e-01
1.00248659e+00 -7.05509841e-01 -4.46539044e-01 3.26565415e-01
-1.39510322e+00 -2.42517501e-01 -3.31771195e-01 2.05615878e-01
-3.50820422e-01 7.99368143e-01 -1.00124407e+00 -5.20148396e-01
-5.14108539e-01 -8.31858218e-01 1.15732718e+00 2.88548768e-01
-5.70639789e-01 -1.59835386e+00 1.54093713e-01 -1.56719446e-01
5.17976359e-02 1.75457075e-01 1.62650084e+00 -1.20483053e+00
2.12687720e-02 2.55342007e-01 -1.89335555e-01 3.06267925e-02
3.75357002e-01 9.79580954e-02 -8.88214052e-01 -4.36205938e-02
2.71889046e-02 -2.06714004e-01 8.43568921e-01 2.23839343e-01
2.75703907e-01 -2.41998240e-01 -8.56022239e-02 4.57794875e-01
1.32677317e+00 1.22847021e-01 5.63173175e-01 6.82807565e-01
6.94158137e-01 6.46735013e-01 5.23658037e-01 -1.38996676e-01
7.13996053e-01 6.59871399e-01 1.17073402e-01 5.05340248e-02
-3.61291528e-01 1.77923683e-02 4.96623129e-01 1.17284048e+00
7.50367939e-02 6.66728541e-02 -1.25942099e+00 1.02137721e+00
-1.60442603e+00 -6.01180673e-01 -2.43042141e-01 2.13193893e+00
1.06531727e+00 -7.82833397e-02 9.31979865e-02 2.63858169e-01
8.16621602e-01 2.58214414e-01 6.61574602e-02 -1.32811368e+00
-5.36902785e-01 1.04404807e+00 -1.16832890e-01 7.24607944e-01
-1.48487651e+00 1.88483846e+00 7.41684151e+00 6.49821758e-01
-1.05941582e+00 1.34518713e-01 3.08735460e-01 6.00804985e-01
-6.28646970e-01 1.85537800e-01 -9.34936404e-01 -7.31858239e-02
9.98043120e-01 1.60953164e-01 2.98768133e-01 4.75217760e-01
-3.05303246e-01 -7.15315640e-02 -1.14807665e+00 6.08544409e-01
-4.44266712e-03 -1.14068472e+00 3.74353766e-01 -2.97896415e-01
4.23562825e-01 -1.41585469e-01 -1.72176972e-01 3.66573572e-01
7.51797795e-01 -9.50345755e-01 7.21662998e-01 -7.52466768e-02
7.23749757e-01 -7.43746459e-01 6.33055568e-01 -1.78997442e-01
-1.55659890e+00 1.67280197e-01 -5.83980858e-01 -2.67702848e-01
1.57878339e-01 2.02288002e-01 -7.09263325e-01 7.29700327e-01
5.29264271e-01 4.25215751e-01 -9.48899031e-01 5.01850784e-01
-1.03054845e+00 6.80137932e-01 -2.17729226e-01 -1.69189814e-02
4.39482719e-01 -3.12932938e-01 6.16340756e-01 1.61296844e+00
-2.08264261e-01 1.73180521e-01 4.10386980e-01 5.61699092e-01
1.62745461e-01 5.82341552e-01 -5.53334594e-01 1.67093173e-01
6.59698606e-01 1.23528230e+00 -1.13450289e+00 -3.79976064e-01
-3.67757142e-01 7.72620857e-01 1.20703745e+00 1.23442881e-01
-3.36246431e-01 -3.49246681e-01 6.74518824e-01 -2.72484004e-01
5.83546698e-01 -6.64605975e-01 -5.49284935e-01 -9.72207606e-01
-8.52633864e-02 -8.12128663e-01 9.87381995e-01 -4.86374199e-01
-1.43565977e+00 8.25675130e-01 -7.13618472e-02 -4.46674764e-01
-3.09391171e-01 -1.22737586e+00 -6.35977209e-01 1.03555012e+00
-1.59324479e+00 -1.32090318e+00 5.87050557e-01 4.50768560e-01
4.55641478e-01 -8.14461634e-02 1.24589670e+00 -9.17439014e-02
-3.40335637e-01 4.83558029e-01 -3.52289021e-01 5.87454021e-01
5.33447623e-01 -1.81346619e+00 1.03693271e+00 1.33463764e+00
6.59424424e-01 8.82712185e-01 4.18621063e-01 -7.33880520e-01
-1.31818092e+00 -7.24073946e-01 1.40859652e+00 -8.14643204e-01
1.11454201e+00 -3.58164191e-01 -1.24506235e+00 7.96586215e-01
5.62651455e-01 -1.24524944e-01 8.38482976e-01 6.10152602e-01
-5.96136570e-01 2.05988795e-01 -9.69119191e-01 8.71890664e-01
1.04529893e+00 -6.84823513e-01 -1.63558066e+00 6.98134899e-02
6.84384048e-01 -3.11759442e-01 -8.14106941e-01 4.09087807e-01
4.27497923e-02 -5.44448495e-01 8.31003428e-01 -1.01192832e+00
-1.76063344e-01 -2.96181440e-01 -3.50138128e-01 -1.30614555e+00
-7.10356593e-01 -7.99047053e-01 -3.45698111e-02 1.28776205e+00
5.20568132e-01 -4.50708717e-01 5.20183802e-01 3.37115675e-01
-4.60495919e-01 -2.88274646e-01 -1.16063273e+00 -6.53918147e-01
7.03149617e-01 -2.92109728e-01 7.63734639e-01 1.18574595e+00
2.38972589e-01 5.82233429e-01 4.09668207e-01 2.51238704e-01
1.34428337e-01 3.88423562e-01 -7.42898583e-02 -1.48483503e+00
-2.42521718e-01 -6.33004069e-01 -6.19011343e-01 -4.79031891e-01
8.32431674e-01 -1.19224620e+00 -3.33099067e-02 -1.44235229e+00
-3.94329280e-01 -3.80961210e-01 -4.50778127e-01 7.58797109e-01
-6.91792607e-01 3.29028696e-01 2.57081240e-01 -2.31014282e-01
-5.13229430e-01 2.98805814e-02 6.15419209e-01 7.90810958e-02
-2.86119282e-01 -6.28744602e-01 -8.72819424e-01 9.07528281e-01
8.04322422e-01 -2.35626712e-01 2.22004116e-01 -9.31012332e-01
5.46940863e-01 -5.07742465e-01 -2.64948696e-01 -4.33633476e-01
-1.75876871e-01 -1.94852933e-01 4.13460493e-01 -2.59073257e-01
6.98489025e-02 -2.25659281e-01 -5.11960149e-01 2.42072269e-01
1.65734321e-01 7.39970505e-01 7.50255644e-01 -1.07267104e-01
-2.22471744e-01 -4.43557173e-01 7.61087000e-01 -4.61610913e-01
-1.22442508e+00 4.53490727e-02 -6.24483645e-01 3.63283336e-01
5.99389970e-01 -2.07862973e-01 -4.12756801e-01 2.29225442e-01
-7.82741904e-01 -6.08123466e-02 5.84898829e-01 3.95442754e-01
4.95241642e-01 -1.28672361e+00 -8.66239607e-01 1.75775975e-01
-8.75412673e-02 -2.97730923e-01 -4.90395159e-01 2.78676420e-01
-7.81387806e-01 1.80235520e-01 -3.02789360e-01 -7.65799582e-02
-9.95816648e-01 9.00446594e-01 1.21099465e-01 -2.88892508e-01
-3.44327420e-01 9.68931258e-01 1.93540603e-01 -8.12268376e-01
-2.82221913e-01 -2.89364189e-01 -6.51664317e-01 8.57323706e-02
5.50203979e-01 2.01935023e-01 5.22184789e-01 -1.16225123e+00
-9.27190840e-01 4.31579590e-01 -2.10079789e-01 -2.33505651e-01
1.18498349e+00 -5.81276901e-02 -6.36345923e-01 5.58469951e-01
5.96539319e-01 6.56761110e-01 -5.59377909e-01 -2.06979111e-01
8.35978150e-01 -1.13739766e-01 -3.52761030e-01 -8.12234044e-01
-6.58617973e-01 9.36815679e-01 8.11909884e-02 1.17337808e-01
9.58697259e-01 1.62717640e-01 6.20074987e-01 5.19124985e-01
1.93537265e-01 -1.11951530e+00 -5.37021041e-01 1.22448218e+00
6.72356665e-01 -7.71298945e-01 1.09995697e-02 -6.27904773e-01
-5.36286175e-01 1.13827729e+00 5.90534508e-01 -5.52989841e-01
5.01788437e-01 3.77088726e-01 3.59212846e-01 -2.32284904e-01
-4.31843877e-01 -7.29718089e-01 6.23808086e-01 1.04210997e+00
1.06061471e+00 3.94508898e-01 -7.43304849e-01 6.77906036e-01
-4.65498805e-01 -9.45622981e-01 2.86347508e-01 8.86377275e-01
-5.38020074e-01 -1.58946633e+00 -4.63226527e-01 1.45540401e-01
-8.09991658e-01 -6.39692068e-01 -1.08665538e+00 1.07485688e+00
2.20777661e-01 8.47094893e-01 1.38049215e-01 1.80098161e-01
1.86907217e-01 3.71811420e-01 8.05417895e-01 -1.10506499e+00
-1.11315787e+00 2.78936684e-01 5.30196369e-01 -3.61364514e-01
-4.10315275e-01 -8.99501026e-01 -1.54802370e+00 -1.36307359e-01
-1.98992386e-01 1.39295176e-01 1.85575321e-01 1.23938000e+00
-7.65374303e-02 -1.01240240e-02 8.84168670e-02 -8.07118833e-01
-2.87012726e-01 -7.56815493e-01 -6.04367673e-01 6.61268175e-01
-1.17167514e-02 -6.26118064e-01 -2.09884550e-02 -1.60601258e-01] | [10.407983779907227, 10.132219314575195] |
d23bd0d8-a424-45b9-8bfd-df8a9407e386 | analysis-over-vision-based-models-for | 2305.17451 | null | https://arxiv.org/abs/2305.17451v1 | https://arxiv.org/pdf/2305.17451v1.pdf | Analysis over vision-based models for pedestrian action anticipation | Anticipating human actions in front of autonomous vehicles is a challenging task. Several papers have recently proposed model architectures to address this problem by combining multiple input features to predict pedestrian crossing actions. This paper focuses specifically on using images of the pedestrian's context as an input feature. We present several spatio-temporal model architectures that utilize standard CNN and Transformer modules to serve as a backbone for pedestrian anticipation. However, the objective of this paper is not to surpass state-of-the-art benchmarks but rather to analyze the positive and negative predictions of these models. Therefore, we provide insights on the explainability of vision-based Transformer models in the context of pedestrian action prediction. We will highlight cases where the model can achieve correct quantitative results but falls short in providing human-like explanations qualitatively, emphasizing the importance of investing in explainability for pedestrian action anticipation problems. | ['François Charpillet', 'François Aioun', 'Julien Moreau', 'Lina Achaji'] | 2023-05-27 | null | null | null | null | ['action-anticipation', 'autonomous-vehicles'] | ['computer-vision', 'computer-vision'] | [ 1.28502890e-01 4.40097898e-01 -1.37623772e-01 -7.05585241e-01
-4.48239475e-01 -1.61875024e-01 9.63209033e-01 7.43791535e-02
-2.44503915e-01 4.67494011e-01 3.91134232e-01 -5.47146320e-01
2.14915007e-01 -7.49539852e-01 -8.43792737e-01 -3.07516783e-01
-5.75127378e-02 3.10327321e-01 5.63873410e-01 -4.88926768e-01
1.86291263e-01 4.83231544e-01 -1.82499886e+00 8.49200547e-01
6.19419634e-01 8.24844778e-01 9.65533555e-02 7.73253143e-01
6.18309081e-02 1.04742122e+00 -1.17569745e-01 -6.79625869e-01
7.48581141e-02 -1.47432461e-01 -8.36685717e-01 1.24331966e-01
6.54386163e-01 -3.44845176e-01 -4.81920123e-01 5.92258573e-01
-4.65072654e-02 5.33638112e-02 5.43296039e-01 -1.87597346e+00
-6.48794651e-01 4.24968153e-01 -8.74309763e-02 2.39896908e-01
4.72937435e-01 7.21682250e-01 1.17071128e+00 -7.68790066e-01
4.40580964e-01 1.36567533e+00 7.55366385e-01 8.15986395e-01
-1.05541217e+00 -2.36870080e-01 6.34754777e-01 1.01962233e+00
-8.62715364e-01 -3.93809617e-01 6.23767078e-01 -4.93233293e-01
1.50179756e+00 3.90612811e-01 1.12674892e+00 1.20376945e+00
3.39241415e-01 1.23903441e+00 7.54091978e-01 -1.54130802e-01
8.18370879e-02 -3.32839079e-02 3.53758752e-01 6.35292530e-01
5.95637038e-02 6.67763591e-01 -4.86020654e-01 4.48055923e-01
3.03757787e-01 -4.99676168e-02 -1.46416156e-02 -4.34834272e-01
-1.05953228e+00 7.11199880e-01 8.83292794e-01 -1.09645329e-01
-5.31574190e-01 4.51239616e-01 2.84696043e-01 -9.93935242e-02
2.45704263e-01 9.05964822e-02 -4.08710480e-01 -1.47726029e-01
-7.32087076e-01 6.64104223e-01 3.67702335e-01 9.24045444e-01
7.34849453e-01 3.16194817e-02 -2.92213798e-01 4.12098952e-02
3.86558086e-01 2.70525426e-01 -1.55939370e-01 -1.11549556e+00
3.52362484e-01 6.62979901e-01 3.66442949e-01 -8.36519957e-01
-3.90744388e-01 -7.74391964e-02 -5.03899753e-01 4.52760309e-01
4.23210412e-01 5.69234788e-02 -9.65068936e-01 1.55000091e+00
5.57229891e-02 3.51359487e-01 4.80741300e-02 1.06942403e+00
9.32455838e-01 5.67051709e-01 6.92337632e-01 3.34434539e-01
1.13626420e+00 -1.35086811e+00 -3.48357528e-01 -6.47053838e-01
4.83159125e-01 -5.60656250e-01 8.59357834e-01 -2.02016309e-02
-1.15913463e+00 -8.82559955e-01 -7.15420544e-01 -2.88754791e-01
-5.81373036e-01 1.65131256e-01 7.19754338e-01 3.26148152e-01
-1.28711951e+00 4.76109773e-01 -9.01401401e-01 -7.00578392e-01
4.57435399e-01 2.96743989e-01 -4.38834876e-01 -6.72396719e-02
-1.05320191e+00 1.46116948e+00 5.56923971e-02 1.17631219e-01
-8.55845511e-01 -7.30908096e-01 -1.03658915e+00 2.65932947e-01
-6.44207448e-02 -9.58431721e-01 1.34859788e+00 -9.29820955e-01
-9.24521446e-01 8.09779465e-01 -6.84704065e-01 -1.06517780e+00
6.76052570e-01 -3.69591415e-01 -3.49018097e-01 -5.48655689e-02
1.93917841e-01 1.41710997e+00 2.77274549e-01 -1.22351372e+00
-1.04653192e+00 2.06532255e-02 4.50326234e-01 -1.00708686e-01
2.06535563e-01 -1.10019319e-01 -1.58468142e-01 -1.13672376e-01
-1.80827484e-01 -9.00757551e-01 -6.58708096e-01 2.15026155e-01
-3.59101504e-01 -2.87247241e-01 9.76314962e-01 -3.83945584e-01
6.81263387e-01 -1.83318639e+00 -2.91494519e-01 -2.55055398e-01
2.17251629e-01 4.50073779e-01 -3.67295921e-01 4.95611221e-01
-2.67749161e-01 8.62468034e-02 1.27386287e-01 -5.62286556e-01
1.56114817e-01 4.05166417e-01 -5.82821131e-01 1.59188062e-01
7.79162884e-01 1.26002192e+00 -1.00224674e+00 -4.33852434e-01
7.67563045e-01 7.29562879e-01 -6.94275260e-01 4.97509092e-02
-3.31851572e-01 4.45929557e-01 -3.84595931e-01 4.14401054e-01
5.90949297e-01 -2.77574044e-02 -1.06971189e-01 -3.27381879e-01
-5.22860646e-01 4.01874423e-01 -7.05030918e-01 9.72328603e-01
-2.91082054e-01 1.23684204e+00 -5.02771854e-01 -1.02279127e+00
6.02040768e-01 1.09858006e-01 3.59218687e-01 -7.88174868e-01
-1.09802224e-01 -1.66936189e-01 6.30253255e-02 -6.29320621e-01
7.13551104e-01 -1.80410757e-03 2.61043459e-01 -2.13586707e-02
-3.84740204e-01 2.13448226e-01 2.51779526e-01 1.22153960e-01
8.84401500e-01 3.45704228e-01 1.76930070e-01 -3.60993370e-02
5.80462456e-01 4.75054830e-01 4.85809892e-01 7.40447402e-01
-6.95427895e-01 6.04470432e-01 4.02211696e-01 -1.23678112e+00
-1.17190516e+00 -1.07918799e+00 2.51679271e-01 8.70340466e-01
2.99609244e-01 -4.72117901e-01 -6.00747466e-01 -6.19081676e-01
-1.24059794e-02 1.14317310e+00 -9.23098624e-01 -1.59785822e-01
-7.32002974e-01 -3.64497751e-01 2.96009064e-01 1.11224449e+00
5.22268653e-01 -9.74855959e-01 -1.17795742e+00 6.65351301e-02
-5.70075929e-01 -1.31579995e+00 -4.44851294e-02 -9.88777429e-02
-3.97234678e-01 -1.25129235e+00 -2.70425200e-01 -5.55866897e-01
4.55444604e-01 6.06966972e-01 1.28385377e+00 3.23557913e-01
-1.51114717e-01 6.47660196e-01 -2.15758801e-01 -5.94635546e-01
-1.94359675e-01 -1.92916602e-01 -4.46706489e-02 -9.70740020e-02
7.62684286e-01 -3.15019816e-01 -7.78975844e-01 4.21580225e-01
-3.74912560e-01 5.19123554e-01 3.57073694e-01 6.70868337e-01
2.65484244e-01 -3.64158511e-01 2.03943297e-01 -2.74571776e-01
1.11134581e-01 -2.88769156e-01 -4.05237049e-01 3.05064946e-01
-3.41438234e-01 -2.86459923e-02 5.23870468e-01 -1.19550705e-01
-1.06918383e+00 4.28590924e-01 -3.58583629e-01 -1.98811844e-01
-5.17411828e-01 -1.16087729e-02 -9.88718644e-02 1.85617227e-02
4.22449917e-01 1.87987491e-01 -1.26589403e-01 -3.83556671e-02
5.42820096e-01 6.50211424e-02 4.16525304e-01 -2.74768531e-01
5.63704729e-01 7.03159392e-01 1.92170903e-01 -5.91624737e-01
-9.08350468e-01 -3.10715020e-01 -7.99534142e-01 -6.44755423e-01
1.03307557e+00 -9.07086432e-01 -1.10974908e+00 9.77183580e-02
-1.52715814e+00 -4.70905662e-01 -3.67717773e-01 2.45133594e-01
-1.00099850e+00 3.02135736e-01 -5.25736213e-01 -9.81607616e-01
1.61836207e-01 -1.27339602e+00 1.17295384e+00 2.42166653e-01
-3.43795240e-01 -8.84996057e-01 -2.78006792e-01 4.59974527e-01
3.50296617e-01 1.88940153e-01 6.50112689e-01 -3.52599174e-01
-1.12614226e+00 -2.25983605e-01 -4.43832308e-01 -1.80519089e-01
-3.29423755e-01 2.57067710e-01 -1.05859590e+00 2.09215432e-01
-5.66428304e-01 3.95478122e-02 1.18435955e+00 7.11140454e-01
8.84161055e-01 -1.89717516e-01 -6.65930927e-01 2.39154458e-01
1.02260983e+00 -5.23825474e-02 8.67132246e-01 4.51681286e-01
3.68619591e-01 9.41507816e-01 1.03228474e+00 3.04362386e-01
1.00279045e+00 7.97751307e-01 9.32695150e-01 -2.87979335e-01
-3.69915217e-01 -5.49215257e-01 2.76646227e-01 -1.51429474e-01
-2.12205231e-01 -4.21259254e-01 -1.13453948e+00 8.51510525e-01
-2.42689228e+00 -1.45808005e+00 -8.00148606e-01 1.66570699e+00
-8.67498666e-03 2.88398206e-01 4.01499569e-01 6.81185573e-02
5.48446476e-01 1.58672929e-02 -3.25810313e-01 -4.96303678e-01
-3.26564759e-02 -3.55934978e-01 2.92988271e-01 7.58612096e-01
-1.25681508e+00 1.27619088e+00 7.13057852e+00 1.97650716e-01
-9.28069890e-01 -1.60983056e-01 7.23256946e-01 9.30711776e-02
-2.56536931e-01 3.58813226e-01 -9.06712472e-01 1.90981850e-01
9.40799057e-01 2.94190608e-02 -3.98331992e-02 9.64353740e-01
6.57991529e-01 -1.59012571e-01 -1.42369449e+00 5.48096299e-01
-2.21939728e-01 -1.52186215e+00 2.38681406e-01 -2.43297778e-02
4.26607341e-01 -2.80944351e-02 2.99145192e-01 3.88653457e-01
4.35957462e-01 -1.19049191e+00 1.01513791e+00 7.19400406e-01
1.42309830e-01 -5.08359790e-01 6.32352352e-01 3.79349798e-01
-1.35214210e+00 -2.74936825e-01 -3.16099793e-01 -6.49513781e-01
5.92222631e-01 1.52081043e-01 -1.00505412e+00 1.60020530e-01
8.25269043e-01 1.00830841e+00 -6.55253708e-01 1.23291862e+00
-2.71177143e-01 4.20453906e-01 -6.13390543e-02 -9.44470465e-02
6.04129851e-01 2.31970668e-01 4.33292687e-01 1.37145710e+00
1.78282529e-01 1.16554394e-01 2.49191344e-01 9.45920587e-01
6.96640968e-01 -3.29565436e-01 -9.12374020e-01 3.24920535e-01
3.57225314e-02 9.09054875e-01 -3.51176530e-01 -4.42057103e-01
-5.55081785e-01 6.73047006e-01 3.64033788e-01 4.04615372e-01
-1.17742181e+00 2.84391314e-01 1.23417079e+00 3.39755058e-01
4.04501289e-01 -3.12020808e-01 -6.50987744e-01 -8.58484745e-01
-2.51039304e-02 -3.54574502e-01 1.07339159e-01 -1.24134684e+00
-9.02780294e-01 5.36229253e-01 5.11184148e-03 -1.36522639e+00
-4.17207122e-01 -8.84932399e-01 -9.99928117e-01 7.27612495e-01
-1.86062872e+00 -1.77372587e+00 -4.68831867e-01 3.78127754e-01
8.78126323e-01 1.09097920e-01 5.79229951e-01 2.87235528e-02
-3.39486361e-01 2.53130972e-01 -6.62963688e-01 -6.39368743e-02
2.91796535e-01 -1.10720563e+00 9.69179988e-01 1.01610291e+00
-7.47908512e-03 3.45344305e-01 1.29350221e+00 -4.79375809e-01
-9.06355262e-01 -1.23708093e+00 1.57143879e+00 -9.77786243e-01
5.28727829e-01 9.36883613e-02 -4.91465867e-01 1.09644592e+00
4.07383800e-01 1.03056788e-01 4.48059916e-01 1.43077955e-01
-2.56934255e-01 -1.27692655e-01 -8.93449903e-01 1.05474842e+00
1.17396426e+00 -2.93415338e-01 -4.61733639e-01 3.28894556e-01
4.67173427e-01 -1.73551604e-01 -2.76322186e-01 4.06399786e-01
7.00243115e-01 -1.53429723e+00 1.36338937e+00 -1.19267619e+00
8.65892947e-01 -3.56165588e-01 -1.38638824e-01 -1.03427565e+00
-5.31890154e-01 -5.73943891e-02 1.66636202e-02 8.53145123e-01
3.75909567e-01 -2.91672528e-01 9.05967712e-01 1.06665659e+00
-3.67708325e-01 -7.78863907e-01 -9.27397251e-01 -5.88294983e-01
1.43317394e-02 -7.48687029e-01 6.34355009e-01 2.47351333e-01
1.38803244e-01 1.51489392e-01 -5.17565310e-01 3.26993018e-01
6.59126639e-01 1.49372905e-01 9.38816488e-01 -9.09794927e-01
1.90488860e-01 -5.39349198e-01 -8.61266553e-01 -1.13193536e+00
3.17924798e-01 -4.32581067e-01 1.99953541e-01 -1.80384934e+00
2.32902840e-01 2.65908102e-03 -6.34731203e-02 6.30007863e-01
-1.72590435e-01 1.83275640e-01 6.10145628e-01 -1.97467670e-01
-7.26508200e-01 5.66993952e-01 1.04930127e+00 -3.96744132e-01
3.24840426e-01 2.56021529e-01 -5.03373742e-01 8.28335822e-01
8.92147005e-01 -3.92126441e-02 -3.29391539e-01 -5.72505116e-01
-1.98563579e-02 -1.82473250e-02 1.28399169e+00 -1.28447938e+00
3.79999548e-01 -4.79898691e-01 4.74603415e-01 -8.57560098e-01
6.75328553e-01 -8.71427715e-01 -9.35990736e-03 6.72726572e-01
-5.27068019e-01 3.20259094e-01 2.76721090e-01 6.09827876e-01
-2.71051437e-01 2.63532013e-01 6.58462942e-01 -1.08773328e-01
-1.37599659e+00 2.23696232e-01 -8.07778895e-01 -3.72959316e-01
1.17688704e+00 -4.82361555e-01 -4.79769737e-01 -6.06344700e-01
-8.98588538e-01 4.88536179e-01 4.34066415e-01 7.36022711e-01
8.64818335e-01 -1.29630220e+00 -6.63841248e-01 -8.55414942e-03
2.08282337e-01 -6.63427234e-01 4.35144097e-01 7.90486574e-01
-2.83931822e-01 8.53066444e-01 -5.55271626e-01 -8.01579773e-01
-1.44121981e+00 6.91591740e-01 4.46057975e-01 -1.70192122e-01
-5.65867901e-01 7.77945518e-01 3.48983794e-01 -2.01058045e-01
7.82770291e-02 -4.17120039e-01 -3.73092771e-01 -3.49598378e-01
3.90263885e-01 2.44306043e-01 -2.60903448e-01 -8.86390686e-01
-5.38506806e-01 4.09353107e-01 8.94877687e-02 6.65361434e-02
1.22699821e+00 -2.76083410e-01 4.73079115e-01 1.13083974e-01
7.57932544e-01 -6.31772757e-01 -1.69400811e+00 1.27530634e-01
2.27249246e-02 -4.33419138e-01 -3.07806641e-01 -8.09767962e-01
-8.56259465e-01 1.41409039e+00 7.00650811e-01 3.89529057e-02
8.22963655e-01 -7.89074525e-02 8.44610989e-01 2.20865160e-01
2.83843845e-01 -9.56016719e-01 1.14551475e-02 6.13803625e-01
9.36026394e-01 -1.66460967e+00 -3.09233010e-01 -5.85731208e-01
-7.89373755e-01 1.21572793e+00 1.09848773e+00 -1.16549768e-01
4.46799517e-01 3.36236283e-02 9.27991122e-02 -9.25827473e-02
-1.04124677e+00 -8.13760757e-01 4.19251442e-01 9.61174965e-01
4.30413157e-01 1.70804143e-01 -4.34265174e-02 3.76325667e-01
-1.73764989e-01 9.93862599e-02 3.87702495e-01 6.32215500e-01
-6.23789370e-01 -9.87496555e-01 -4.70181890e-02 1.39541194e-01
1.29118919e-01 -4.18361016e-02 -5.06264150e-01 8.14978838e-01
3.30341816e-01 1.27770948e+00 1.24408320e-01 -6.12234473e-01
4.21441704e-01 -8.06623325e-02 3.11095029e-01 -1.33455262e-01
-5.34058928e-01 -6.46432102e-01 3.64164680e-01 -8.19711983e-01
-4.93404090e-01 -8.59791219e-01 -1.04465258e+00 -5.08480966e-01
2.85073221e-01 -2.72921056e-01 2.99160391e-01 1.21441281e+00
3.90806526e-01 4.91727710e-01 1.88816458e-01 -1.15418863e+00
-5.58270961e-02 -4.48769659e-01 2.20627517e-01 4.55303639e-01
5.90333164e-01 -6.80131555e-01 7.53157139e-02 2.22102731e-01] | [6.236481189727783, 0.6743611097335815] |
6e7ef943-2ca8-4592-a505-82e28764abca | timers-document-level-temporal-relation | null | null | https://aclanthology.org/2021.acl-short.67 | https://aclanthology.org/2021.acl-short.67.pdf | TIMERS: Document-level Temporal Relation Extraction | We present TIMERS - a TIME, Rhetorical and Syntactic-aware model for document-level temporal relation classification in the English language. Our proposed method leverages rhetorical discourse features and temporal arguments from semantic role labels, in addition to traditional local syntactic features, trained through a Gated Relational-GCN. Extensive experiments show that the proposed model outperforms previous methods by 5-18{\%} on the TDDiscourse, TimeBank-Dense, and MATRES datasets due to our discourse-level modeling. | ['Dinesh Manocha', 'Quan Hung Tran', 'Vlad Morariu', 'Franck Dernoncourt', 'Rajiv Jain', 'Puneet Mathur'] | 2021-08-01 | null | null | null | acl-2021-5 | ['temporal-relation-extraction', 'temporal-relation-classification'] | ['natural-language-processing', 'natural-language-processing'] | [-1.31963283e-01 5.13439417e-01 -1.31148648e+00 -6.54248476e-01
-9.00967360e-01 -7.53959298e-01 1.44056046e+00 5.39228439e-01
-3.24887604e-01 9.53392863e-01 1.07603812e+00 -5.79149365e-01
-7.25352466e-02 -5.72996259e-01 -3.60112548e-01 -2.45580971e-01
-5.50461709e-01 5.05865991e-01 4.87179399e-01 -7.30721176e-01
2.76176333e-01 -9.49891508e-02 -8.65131974e-01 7.57416785e-01
3.72074306e-01 1.27329433e+00 -2.06433192e-01 6.04538679e-01
-1.58295795e-01 2.44751668e+00 -8.20107877e-01 -1.56081304e-01
-3.26940686e-01 -3.81218642e-01 -1.52168262e+00 -3.87746543e-01
-1.73641384e-01 -1.40649483e-01 -7.34150231e-01 1.63700581e-01
1.24089792e-01 4.90813851e-01 5.12557387e-01 -8.56918454e-01
-5.13625443e-01 1.49806917e+00 -3.77827048e-01 9.27350044e-01
7.75053859e-01 -3.90961796e-01 1.55364025e+00 -5.00978410e-01
1.21879888e+00 1.79578459e+00 4.99872416e-01 8.12454075e-02
-9.62608755e-01 -2.48532772e-01 5.46482146e-01 6.70392156e-01
-6.22833848e-01 -4.51172709e-01 1.11908388e+00 -3.49060804e-01
1.68550229e+00 2.26376995e-01 5.37884831e-01 1.64604247e+00
2.35332891e-01 9.81597602e-01 1.21009898e+00 -4.69869763e-01
2.25658696e-02 -7.65357852e-01 7.08876073e-01 5.43930829e-01
-8.57546866e-01 -4.46421914e-02 -9.77412641e-01 -2.03209326e-01
2.83571720e-01 -5.05768359e-01 7.60393590e-02 5.26884198e-01
-1.17009616e+00 9.37516868e-01 1.28937721e-01 6.59712553e-01
-2.42959052e-01 3.22686821e-01 1.18987000e+00 5.22239387e-01
1.13651454e+00 3.24316025e-01 -6.53866529e-01 -8.68375480e-01
-5.24224699e-01 2.95845062e-01 9.71324325e-01 8.66928101e-01
-7.91551992e-02 -4.16105241e-02 -5.23088455e-01 8.07430267e-01
4.28934395e-03 1.42496126e-02 5.42863250e-01 -1.32771897e+00
8.28795671e-01 8.36554170e-01 -1.55630052e-01 -8.39217246e-01
-6.94570124e-01 2.90215284e-01 -2.54205436e-01 -8.16017210e-01
2.02628985e-01 1.15305834e-01 -4.27473724e-01 1.69812274e+00
3.60719323e-01 7.60519654e-02 1.87239647e-01 6.09065950e-01
1.31479788e+00 8.32225919e-01 4.27157909e-01 -7.63631403e-01
1.47396040e+00 -9.76924002e-01 -1.31669748e+00 -1.20148860e-01
9.92488205e-01 -5.04467666e-01 1.07092500e+00 -1.85567796e-01
-1.21923161e+00 -6.38105944e-02 -9.94488478e-01 -6.89495027e-01
-1.42855078e-01 -2.59124070e-01 8.77005875e-01 -2.57156432e-01
-6.91754580e-01 5.66688776e-01 -8.90179873e-01 -2.48327032e-01
2.64553130e-01 -3.25424939e-01 5.77713326e-02 4.87730026e-01
-1.79631186e+00 1.06402564e+00 6.02283835e-01 -1.48445845e-01
-9.25115168e-01 -5.20729721e-01 -1.23120475e+00 -3.53038460e-01
7.67241836e-01 -7.23462775e-02 1.78112197e+00 7.72871673e-02
-1.70009732e+00 1.04323912e+00 -4.73481148e-01 -9.92690742e-01
2.80040324e-01 -3.47465366e-01 -4.36500490e-01 3.90493244e-01
4.24121499e-01 -2.17479449e-02 4.23144460e-01 -5.57972789e-01
-5.00619054e-01 -1.74901739e-01 6.22230232e-01 3.45585078e-01
-1.03803046e-01 6.66221023e-01 -2.26928249e-01 -8.62051845e-01
-2.35165674e-02 -6.30453348e-01 3.73632647e-02 -6.77112281e-01
-5.93041062e-01 -1.27337587e+00 1.07026720e+00 -8.21098983e-01
1.62775850e+00 -1.90713549e+00 2.73277968e-01 -3.49128425e-01
1.44264117e-01 -2.98770070e-01 1.00685276e-01 6.00804329e-01
-4.34932373e-02 3.82814147e-02 1.75309479e-01 -1.68473095e-01
-7.07248822e-02 5.61234832e-01 -5.69419026e-01 5.79185367e-01
1.64748915e-02 1.10107756e+00 -1.19573522e+00 -1.00562489e+00
3.26013565e-01 -3.83588634e-02 -8.25387891e-03 1.59106657e-01
-7.53854811e-01 5.12096584e-01 -8.03949952e-01 6.68716133e-01
-3.79606098e-01 -3.07249308e-01 8.54943395e-01 9.21168476e-02
-2.36626118e-01 1.74447119e+00 -2.20009357e-01 1.80270529e+00
-5.34912825e-01 7.18540192e-01 -1.46584496e-01 -1.37963426e+00
7.51427293e-01 8.66621733e-01 6.27385497e-01 -1.34965527e+00
4.27783668e-01 3.10541354e-02 -4.28177118e-02 -4.14779723e-01
4.60889280e-01 2.13686731e-02 -7.63366580e-01 6.26256526e-01
1.29556894e-01 1.24111548e-01 4.96404678e-01 4.87959564e-01
1.49522007e+00 4.29120123e-01 6.94240272e-01 -3.94903690e-01
5.37855446e-01 1.60192668e-01 7.81835020e-01 3.12814265e-01
-2.89457202e-01 -1.47165641e-01 1.06820762e+00 -5.78311920e-01
-6.43830299e-01 -5.27358949e-01 -2.02181041e-01 1.72365761e+00
-9.93165448e-02 -9.80655015e-01 -2.61721820e-01 -9.69068229e-01
-2.04294279e-01 8.45144272e-01 -7.70044684e-01 3.15533221e-01
-1.59462488e+00 -3.67154390e-01 8.58514488e-01 7.35749543e-01
3.75030249e-01 -1.33047163e+00 -6.65950894e-01 5.16861856e-01
-5.47933042e-01 -1.64834511e+00 -2.24967271e-01 5.54929793e-01
-7.92276800e-01 -1.43431294e+00 1.18803814e-01 -6.47844255e-01
-3.68706048e-01 -1.64975002e-01 1.31160152e+00 -2.05573484e-01
2.08142281e-01 5.08086719e-02 -6.13659024e-01 -1.70340553e-01
-4.88580763e-01 2.14808166e-01 -2.20501393e-01 -6.79103911e-01
1.77388728e-01 -7.63499022e-01 -3.19842637e-01 1.46008460e-04
-1.09976403e-01 -1.08423226e-01 -3.59016925e-01 7.81488359e-01
3.23147476e-01 -4.03943747e-01 6.14897013e-01 -1.06634498e+00
8.22163045e-01 -7.07511067e-01 -2.87096441e-01 9.97016951e-02
-4.48947310e-01 9.78247374e-02 3.94964516e-01 -4.80066150e-01
-1.46232808e+00 -8.89885545e-01 1.52530730e-01 1.45987924e-02
3.67171884e-01 8.81976128e-01 1.42670274e-01 8.02242577e-01
7.40672946e-01 -3.03922862e-01 -2.19212919e-01 -2.71113366e-01
9.42879319e-01 1.84927836e-01 7.51646638e-01 -1.12214255e+00
4.47031438e-01 4.66175973e-01 3.86231914e-02 -6.55346751e-01
-1.56191552e+00 -4.53370214e-01 -4.45229381e-01 -8.93764645e-02
8.79601181e-01 -9.34938550e-01 -8.91768813e-01 -1.74529534e-02
-1.51275849e+00 -6.80144370e-01 -4.81217444e-01 2.22965032e-01
-7.44179368e-01 1.80614650e-01 -1.58742845e+00 -9.44493055e-01
-3.88668656e-01 -4.30599630e-01 1.00825095e+00 -1.32636696e-01
-8.70736480e-01 -1.49666953e+00 -1.66408147e-03 4.41636682e-01
-6.62847385e-02 8.15365076e-01 1.19549954e+00 -7.30079889e-01
-9.66211930e-02 3.54901165e-01 1.40987784e-02 -2.25889340e-01
2.38527775e-01 -2.17808619e-01 -7.50124693e-01 3.26889783e-01
1.92027166e-01 -8.86973321e-01 8.87633026e-01 6.65693402e-01
8.17025244e-01 -7.82535911e-01 -4.57487226e-01 1.13391519e-01
5.41149974e-01 4.76824433e-01 1.74695238e-01 6.74823642e-01
4.06347692e-01 5.62327921e-01 1.19551778e+00 6.12348497e-01
8.21336150e-01 6.98125720e-01 1.75015301e-01 1.89958155e-01
-3.44079673e-01 -1.94115311e-01 4.57117975e-01 1.05838752e+00
-3.58672559e-01 -2.23686874e-01 -9.36563909e-01 7.50416756e-01
-2.34086823e+00 -1.24405432e+00 -3.59326214e-01 1.13191843e+00
1.49285209e+00 5.13137043e-01 2.17858717e-01 2.84955204e-01
4.73760456e-01 1.20567441e+00 -2.33699068e-01 -6.42758310e-01
-5.00407815e-01 4.15859878e-01 9.27204341e-02 4.91254747e-01
-1.45167720e+00 1.36771333e+00 6.72381544e+00 8.99019837e-01
-7.69957185e-01 6.44958079e-01 5.22039652e-01 -1.47293091e-01
-1.17240652e-01 2.42069095e-01 -6.01294339e-01 2.05925316e-01
1.42613459e+00 -4.71436203e-01 1.67729750e-01 7.36641228e-01
4.34281290e-01 -1.33353978e-01 -1.29455578e+00 4.40220475e-01
-5.54623976e-02 -2.01010180e+00 -6.28894091e-01 -2.25702360e-01
4.85027760e-01 -3.86120826e-02 -1.81790829e-01 6.45075440e-01
8.09975743e-01 -8.07196021e-01 1.22443652e+00 3.31914574e-02
6.67183459e-01 -2.92235881e-01 4.85860229e-01 2.45247334e-01
-1.53726387e+00 -4.42929193e-02 3.95803243e-01 -5.31127691e-01
5.56066930e-01 2.72136658e-01 -9.20042038e-01 7.73695707e-01
6.17508054e-01 1.21327078e+00 -1.51940376e-01 -3.91928345e-01
-6.84298575e-01 1.29378808e+00 -2.61382014e-01 -1.44509584e-01
4.83576745e-01 2.50123620e-01 6.67685449e-01 1.10989487e+00
-3.59802186e-01 6.32733345e-01 4.85642314e-01 5.32646835e-01
-2.80716032e-01 -6.80702552e-02 -5.61370969e-01 -3.03780824e-01
8.03053439e-01 1.02323794e+00 -5.47606528e-01 -3.29308033e-01
-2.62302667e-01 4.31422174e-01 4.54302609e-01 1.39573902e-01
-1.02931392e+00 2.31614843e-01 1.26382753e-01 -8.32584798e-02
5.54474406e-02 -3.77839446e-01 -1.10424079e-01 -8.37051988e-01
6.80747256e-02 -7.01520622e-01 1.09890735e+00 -3.55417609e-01
-1.21036744e+00 4.89582807e-01 4.70677316e-01 -7.58770585e-01
-9.24382269e-01 -2.96675235e-01 -6.60751700e-01 3.44903469e-01
-1.40251601e+00 -1.58176625e+00 3.81507665e-01 5.34185171e-01
9.50022936e-01 -2.14425638e-01 7.24275589e-01 8.45026523e-02
-2.93283522e-01 1.00697793e-01 -5.69210291e-01 4.67087388e-01
5.35147548e-01 -1.28344095e+00 2.83703178e-01 4.69834983e-01
5.37781715e-02 5.88779032e-01 7.29311585e-01 -6.88698828e-01
-1.22590458e+00 -9.56051290e-01 1.46569943e+00 -7.23257363e-01
1.51329100e+00 -1.80935159e-01 -7.26328313e-01 1.09968245e+00
6.69145823e-01 5.04570408e-03 5.18176615e-01 8.89973402e-01
-7.28950083e-01 2.84733444e-01 -6.51920378e-01 5.25135458e-01
1.53771389e+00 -1.00100243e+00 -1.48526967e+00 6.40322804e-01
1.34102952e+00 -8.26619685e-01 -1.34180343e+00 4.81526703e-01
1.49045199e-01 -2.59514540e-01 9.03899312e-01 -9.23801184e-01
8.38286161e-01 1.00977771e-01 -3.17196876e-01 -7.20658660e-01
4.88008000e-03 -1.04071951e+00 -9.18477118e-01 1.48801017e+00
4.15549219e-01 -2.84607589e-01 2.83384204e-01 3.09833400e-02
-5.18528223e-01 -7.39322901e-01 -1.50373459e+00 -8.92228663e-01
1.64213359e-01 -7.15787053e-01 2.89293796e-01 1.23755455e+00
7.16389418e-01 1.07553494e+00 -2.51089871e-01 -6.02100611e-01
2.91597042e-02 3.77216399e-01 8.64265636e-02 -1.07295752e+00
-1.03170037e-01 -5.08585572e-01 2.50565149e-02 -8.74679983e-01
1.13590670e+00 -9.78234947e-01 -1.61772028e-01 -1.51103234e+00
1.39500890e-02 -3.64014775e-01 -2.53003746e-01 7.55939364e-01
6.02879785e-02 -2.93513149e-01 -1.53582871e-01 2.54190117e-01
-1.07879424e+00 7.96742499e-01 1.11067104e+00 -3.55838895e-01
-2.73906738e-01 -4.01287705e-01 -3.99253815e-01 8.45723748e-01
4.85352367e-01 -4.36432540e-01 -4.68060791e-01 -1.54136285e-01
1.56307921e-01 7.79745996e-01 1.90304756e-01 -2.03927204e-01
1.23261362e-01 -6.23579025e-01 -2.45793477e-01 -8.62008214e-01
6.37313724e-01 1.81538537e-02 -5.55811465e-01 2.01620117e-01
-8.38778794e-01 9.18047428e-02 2.76811626e-02 6.54108703e-01
-5.20685613e-01 2.66428411e-01 2.15797529e-01 6.96212379e-03
-7.75323927e-01 -1.18992619e-01 -4.01957959e-01 6.17331207e-01
8.46134484e-01 2.67631590e-01 -1.11757195e+00 -5.28422773e-01
-5.63264906e-01 3.84033650e-01 -3.54592383e-01 6.22373700e-01
1.58946931e-01 -1.40862036e+00 -6.15057409e-01 -8.89834940e-01
1.36210725e-01 6.08950779e-02 -1.74474698e-02 1.09582543e+00
2.34333873e-02 7.29932368e-01 5.79844773e-01 -5.17426491e-01
-1.11677659e+00 3.88418257e-01 -1.27084762e-01 -8.26895773e-01
-9.93856132e-01 8.10330033e-01 -3.16865653e-01 -1.80421591e-01
1.74882427e-01 -5.95790565e-01 -5.46512246e-01 4.90709245e-01
8.90757293e-02 1.43303946e-01 -2.66400259e-02 -7.03985095e-01
-6.31231725e-01 -6.30161092e-02 -2.82225711e-03 -5.71290135e-01
1.43636024e+00 -6.75909966e-02 -4.96903926e-01 1.07227182e+00
9.68335092e-01 -3.50539852e-03 -7.46597707e-01 -5.91384172e-01
8.80447388e-01 2.30517641e-01 -6.69668987e-02 -6.34843171e-01
-2.94059277e-01 3.96740407e-01 -3.13010663e-01 6.35864556e-01
7.03288019e-01 7.03207731e-01 8.02647293e-01 3.42101753e-01
3.33215833e-01 -1.44955254e+00 4.94562298e-01 1.15424323e+00
1.14060318e+00 -7.37481356e-01 1.75234258e-01 -4.39599723e-01
-6.86048627e-01 8.88069034e-01 4.98535514e-01 -2.17034668e-01
6.11162782e-01 5.23885190e-01 -1.45245688e-02 -5.34568012e-01
-1.44863534e+00 -4.38700989e-02 1.76844150e-01 2.29379833e-01
1.07678151e+00 1.33461848e-01 -9.80990291e-01 8.07040930e-01
-4.31279838e-01 -2.88451612e-01 1.00934096e-01 1.30139554e+00
-8.20193067e-03 -1.26109958e+00 -3.94280404e-02 1.65211126e-01
-6.62244737e-01 -1.75822109e-01 -4.36888963e-01 1.10054123e+00
-3.09734285e-01 1.30212104e+00 2.47524977e-01 -1.53000385e-01
9.92317498e-02 1.93325549e-01 4.97930318e-01 -6.94028139e-01
-7.52607346e-01 2.37781510e-01 1.45736516e+00 -1.06644237e+00
-1.10681272e+00 -8.51035774e-01 -1.89315057e+00 -2.68454194e-01
1.72107518e-01 2.36437768e-01 3.29996757e-02 1.38295889e+00
4.59420457e-02 9.37104583e-01 5.79317689e-01 -5.14664888e-01
-4.47838843e-01 -1.29339135e+00 -1.42158777e-01 3.31314147e-01
1.55392736e-01 -8.74979138e-01 -2.56044924e-01 -2.83648539e-03] | [9.197713851928711, 9.219600677490234] |
783e7608-8452-44eb-a041-a224fd890255 | multi-modal-graph-learning-over-umls | 2307.04461 | null | https://arxiv.org/abs/2307.04461v1 | https://arxiv.org/pdf/2307.04461v1.pdf | Multi-modal Graph Learning over UMLS Knowledge Graphs | Clinicians are increasingly looking towards machine learning to gain insights about patient evolutions. We propose a novel approach named Multi-Modal UMLS Graph Learning (MMUGL) for learning meaningful representations of medical concepts using graph neural networks over knowledge graphs based on the unified medical language system. These representations are aggregated to represent entire patient visits and then fed into a sequence model to perform predictions at the granularity of multiple hospital visits of a patient. We improve performance by incorporating prior medical knowledge and considering multiple modalities. We compare our method to existing architectures proposed to learn representations at different granularities on the MIMIC-III dataset and show that our approach outperforms these methods. The results demonstrate the significance of multi-modal medical concept representations based on prior medical knowledge. | ['Rita Kuznetsova', 'Gunnar Rätsch', 'Manuel Burger'] | 2023-07-10 | null | null | null | null | ['graph-learning', 'knowledge-graphs'] | ['graphs', 'knowledge-base'] | [ 3.15496922e-01 8.43876839e-01 -5.68450809e-01 -4.43874091e-01
-8.65763485e-01 -2.47761086e-01 3.56873155e-01 1.20873165e+00
3.14467810e-02 5.77670932e-01 6.44440413e-01 -6.58759713e-01
-6.12813354e-01 -1.02271450e+00 -4.48976815e-01 -1.45378247e-01
-5.39145887e-01 8.59705210e-01 4.90100384e-02 -9.14817005e-02
-4.25566941e-01 3.69150251e-01 -1.07310700e+00 8.65704179e-01
7.08728373e-01 7.72385120e-01 -2.62815267e-01 9.05561924e-01
-1.37693420e-01 1.37079513e+00 -3.91652226e-01 -4.00365084e-01
-1.11327767e-01 -4.46553320e-01 -1.02481091e+00 3.01498830e-01
1.15324855e-01 -4.61622849e-02 -3.24974418e-01 5.77587843e-01
2.35365242e-01 1.76380381e-01 9.68055189e-01 -9.39123809e-01
-9.64434147e-01 7.22713470e-01 -2.26339445e-01 2.39135697e-01
7.84166813e-01 -1.59797773e-01 1.12446618e+00 -3.83834466e-02
9.60741222e-01 1.34196723e+00 7.63610125e-01 6.01763129e-01
-1.14015055e+00 1.76766545e-01 4.24703330e-01 2.95243710e-01
-1.12182689e+00 1.34243950e-01 3.67086768e-01 -6.32733822e-01
1.16484034e+00 4.20506746e-01 5.83323240e-01 1.06357563e+00
4.78297114e-01 6.18392766e-01 4.34375703e-01 -4.40623879e-01
1.99638039e-01 -1.25076830e-01 4.34291005e-01 1.31815147e+00
5.63706994e-01 -1.51176706e-01 -2.22301841e-01 -8.52041423e-01
6.30710661e-01 5.23995340e-01 -2.03175470e-01 -3.40507329e-01
-1.08443701e+00 9.82700527e-01 8.92985463e-01 4.53994244e-01
-8.44772398e-01 3.03702027e-01 6.76736176e-01 -3.86324637e-02
4.70657855e-01 4.09242034e-01 -3.88729662e-01 3.53040785e-01
-6.57067657e-01 -3.05859391e-02 8.60513628e-01 8.23870063e-01
3.01310450e-01 -2.09611654e-01 -4.94258225e-01 5.66015720e-01
4.05591309e-01 -1.97911933e-01 7.96901643e-01 -5.90735972e-01
3.06237459e-01 1.22077930e+00 -1.03394672e-01 -8.22782815e-01
-7.68506587e-01 -2.17448056e-01 -6.73756063e-01 -5.03349960e-01
-1.19770125e-01 -6.54493049e-02 -1.46888685e+00 1.28532934e+00
2.12631434e-01 2.84736127e-01 3.17669839e-01 4.08790529e-01
1.10571527e+00 3.27307791e-01 5.80374300e-01 -8.99274200e-02
1.48832345e+00 -6.27136230e-01 -6.54864013e-01 1.28953427e-01
1.16149366e+00 4.31965478e-02 4.24600661e-01 1.98731154e-01
-7.30132997e-01 -2.28819206e-01 -7.41425455e-01 2.57328898e-01
-5.61776042e-01 -2.09960282e-01 8.96625400e-01 6.48179948e-01
-1.20013607e+00 7.67174780e-01 -1.30329895e+00 -6.34600282e-01
6.22489333e-01 2.54677951e-01 -1.36499718e-01 -3.48605722e-01
-1.20762384e+00 9.24491763e-01 8.74110222e-01 -3.98565382e-01
-7.49705017e-01 -7.30873108e-01 -1.31189525e+00 1.73403233e-01
3.01826835e-01 -1.32467782e+00 1.08767295e+00 -4.62902486e-01
-7.18150973e-01 7.41803706e-01 -4.85631339e-02 -8.45511436e-01
1.00044601e-01 3.03224653e-01 -7.94871032e-01 4.60621089e-01
-1.98070332e-01 2.46877521e-01 2.53086954e-01 -1.14592969e+00
-5.53586602e-01 -3.77575547e-01 2.37370595e-01 -3.00672930e-02
-1.90156162e-01 -4.85063076e-01 -2.04531670e-01 -3.89096439e-01
1.27454102e-02 -7.68202424e-01 -9.48532283e-01 -2.39231601e-01
-5.55761158e-01 -2.69567162e-01 3.50449860e-01 -8.44462335e-01
1.55485988e+00 -1.66090274e+00 2.70392239e-01 4.69005585e-01
6.31321073e-01 1.14112105e-02 -5.66635020e-02 7.13204980e-01
-3.14118415e-01 2.27607518e-01 -2.40983367e-01 -9.97250974e-02
-3.20026308e-01 4.79231447e-01 1.75272927e-01 1.03997506e-01
1.09528877e-01 1.34043312e+00 -1.01619112e+00 -5.41450918e-01
2.42986992e-01 6.25377238e-01 -5.44304430e-01 1.52826265e-01
-5.45756817e-01 3.00316185e-01 -7.46067286e-01 9.36112881e-01
-5.10824546e-02 -1.12482715e+00 7.52602577e-01 -1.20237559e-01
7.34518647e-01 8.29076543e-02 -6.04517043e-01 1.68495178e+00
-3.20420951e-01 -1.14330187e-01 -6.57847762e-01 -1.13790321e+00
6.38002515e-01 7.13467538e-01 8.48508954e-01 -2.61039257e-01
1.49272963e-01 -2.31612384e-01 -5.58033250e-02 -6.67687297e-01
-3.04263644e-02 -2.51576990e-01 -1.45766094e-01 1.35614276e-01
1.04450539e-01 2.29918808e-01 2.69519299e-01 5.31174004e-01
1.46591842e+00 -1.54162258e-01 1.02942586e+00 6.44170642e-02
3.87527734e-01 3.56297754e-02 2.73766428e-01 6.74616575e-01
1.15730181e-01 2.76964515e-01 6.35709107e-01 -7.92691588e-01
-5.10060608e-01 -1.17570508e+00 -1.17674612e-01 8.08706462e-01
-3.82516772e-01 -9.17033017e-01 -3.00844699e-01 -1.05292368e+00
3.49275887e-01 7.69872606e-01 -1.09664524e+00 -3.54078680e-01
-8.67326781e-02 -7.59842694e-01 2.97445804e-01 9.24274981e-01
-5.30826449e-01 -9.16616619e-01 -7.66714633e-01 4.17680025e-01
6.63939044e-02 -9.75348830e-01 1.71355065e-02 -2.21532416e-02
-1.11328328e+00 -1.52011442e+00 -5.83684504e-01 -5.32872975e-01
7.20539868e-01 -5.13565660e-01 1.19189870e+00 1.06926002e-01
-9.78565931e-01 1.13411224e+00 -5.40754080e-01 -4.21418428e-01
-7.58782744e-01 -5.57761565e-02 -2.04248413e-01 -1.17985286e-01
3.35223287e-01 -3.98440778e-01 -6.43084943e-01 -3.91979039e-01
-1.10874081e+00 -6.61460161e-02 4.66763973e-01 7.32805431e-01
7.23320127e-01 -2.94244558e-01 5.08073211e-01 -1.51953375e+00
8.73153508e-01 -1.04998696e+00 -1.32712513e-01 7.17362404e-01
-7.08305299e-01 4.16209757e-01 3.53830248e-01 -1.71677902e-01
-6.89437330e-01 7.62072206e-02 1.04895942e-01 -5.25778055e-01
-2.69846499e-01 1.12107790e+00 4.97048348e-01 2.15960920e-01
8.98942471e-01 -1.70020796e-02 -2.69471854e-01 -2.95213699e-01
5.79001665e-01 4.93105203e-01 2.44792238e-01 -3.35504472e-01
6.60560057e-02 3.10174078e-01 3.50282013e-01 -5.39313078e-01
-8.66208792e-01 -5.21545291e-01 -5.82396626e-01 6.96938261e-02
1.16682470e+00 -7.37554908e-01 -8.06707561e-01 -4.72518414e-01
-7.45125949e-01 6.71142265e-02 -3.95124614e-01 4.63541508e-01
-6.62233114e-01 4.37809616e-01 -8.24069679e-01 -6.39896095e-01
-4.69110280e-01 -7.60961354e-01 1.02205122e+00 -8.21497217e-02
-5.18869698e-01 -1.67115664e+00 3.27100694e-01 2.41814613e-01
7.33535811e-02 1.01751804e+00 1.55792105e+00 -1.03854370e+00
-3.20991933e-01 -3.37230027e-01 -4.14115638e-02 -1.73574492e-01
7.30729461e-01 -2.76770711e-01 -4.81458604e-01 -3.67225945e-01
-6.16236567e-01 -1.28722385e-01 9.01880622e-01 5.78959823e-01
1.49693811e+00 -5.20640492e-01 -8.32428455e-01 4.15520638e-01
1.60439384e+00 4.28814769e-01 2.57326990e-01 -9.10464525e-02
8.22483540e-01 4.60670054e-01 1.67833075e-01 7.22464859e-01
7.05659032e-01 1.43139139e-01 6.55705512e-01 -6.02113493e-02
5.05305640e-02 -1.46952644e-01 -3.40758651e-01 5.74556530e-01
-2.90672958e-01 -3.87458235e-01 -1.46894181e+00 7.88781881e-01
-2.09415317e+00 -5.90629637e-01 6.86988682e-02 1.96051419e+00
4.97271061e-01 -1.09781407e-01 1.32110000e-01 -2.96668828e-01
4.41457093e-01 -1.13032393e-01 -6.18651032e-01 -4.83346432e-01
4.24556106e-01 4.18253601e-01 3.87021750e-01 3.60435098e-01
-9.13526893e-01 5.61858475e-01 7.24121904e+00 4.80845310e-02
-6.79269612e-01 4.55864742e-02 6.59612000e-01 1.20094419e-02
-4.98514831e-01 -2.85169244e-01 -4.15466398e-01 8.37378949e-02
1.46253633e+00 -3.17134410e-01 9.57087427e-02 6.65927947e-01
-2.45958850e-01 1.92405418e-01 -1.44799697e+00 9.02025998e-01
2.48594746e-01 -1.71883202e+00 7.01679945e-01 1.21167921e-01
6.87367022e-01 -4.41745706e-02 -2.40122095e-01 2.13571280e-01
9.99884307e-01 -1.20600581e+00 -1.82218134e-01 1.04533553e+00
6.70178294e-01 -6.94987833e-01 5.77903330e-01 -3.69721763e-02
-1.34897578e+00 -2.64366865e-01 -5.09856381e-02 4.83054072e-01
1.62320241e-01 2.53881961e-01 -1.74577761e+00 1.37286711e+00
3.13896179e-01 9.90437269e-01 -7.61938334e-01 8.85843039e-01
3.50738391e-02 6.17371023e-01 9.75030437e-02 2.33047441e-01
3.08617681e-01 2.76072919e-01 1.64493084e-01 1.22211123e+00
4.09288466e-01 3.28771681e-01 5.42814910e-01 5.48392296e-01
-5.81532829e-02 2.67805487e-01 -9.79335785e-01 -6.41577125e-01
-6.33730143e-02 9.10618961e-01 -6.68845057e-01 -7.66313612e-01
-5.47026336e-01 8.00076246e-01 3.72868210e-01 3.65283668e-01
-5.22466481e-01 -2.39385925e-02 3.54467154e-01 2.71599382e-01
1.90894902e-01 3.04487586e-01 4.08238918e-02 -1.09185255e+00
-5.49586535e-01 -6.04866207e-01 1.45464575e+00 -6.63070679e-01
-1.39160037e+00 8.74343455e-01 -7.32681528e-02 -1.12541509e+00
-8.61501932e-01 -6.95612609e-01 -1.89974606e-01 7.18836486e-01
-1.36754298e+00 -1.47765005e+00 -2.08055377e-02 8.26160014e-01
1.93482324e-01 -2.54667640e-01 1.43588626e+00 -1.08784609e-01
-1.56989068e-01 3.93063694e-01 -2.40702957e-01 9.77742225e-02
3.01204771e-01 -1.34307635e+00 2.34930500e-01 2.17140615e-01
7.24078789e-02 6.77172065e-01 2.26720914e-01 -1.07868171e+00
-1.05646288e+00 -1.42184031e+00 7.92634130e-01 -5.43655336e-01
6.16573513e-01 2.53344297e-01 -1.05874944e+00 1.31730878e+00
7.43161887e-02 1.06114104e-01 1.43459332e+00 4.00280923e-01
-5.23375630e-01 4.47454095e-01 -1.18709254e+00 2.15195581e-01
8.14646006e-01 -4.85912621e-01 -8.97298574e-01 6.24673605e-01
9.16039109e-01 -3.13616693e-01 -1.50155377e+00 5.23406923e-01
1.21982418e-01 -3.77753556e-01 1.04506111e+00 -1.56131458e+00
4.65792328e-01 4.11214903e-02 -2.35156298e-01 -1.44905281e+00
-4.30233538e-01 -1.50359854e-01 -5.60492039e-01 3.24311316e-01
5.37083387e-01 -6.55920923e-01 4.98039871e-01 6.62070632e-01
2.41092313e-02 -1.07681727e+00 -7.30186522e-01 -3.47258806e-01
-3.05876017e-01 -9.62050632e-02 7.48996556e-01 1.31530559e+00
5.17540514e-01 1.18574567e-01 -1.46868393e-01 4.91874754e-01
4.99021471e-01 2.81743348e-01 -4.40784320e-02 -1.54550135e+00
-5.58181822e-01 -1.42742544e-01 -9.59612846e-01 -1.20654767e-02
-6.24882020e-02 -1.42082918e+00 -5.64783096e-01 -2.44692302e+00
3.39293242e-01 8.66257027e-02 -8.28457236e-01 9.23946202e-01
-2.94031262e-01 -1.87517837e-01 -1.13362521e-01 -2.12273642e-01
-7.30411708e-01 9.43467543e-02 1.13556254e+00 -2.25480437e-01
-2.24181980e-01 -9.53675807e-02 -7.52293468e-01 6.67638183e-01
5.47148526e-01 -3.03486973e-01 -5.76106131e-01 -7.77921900e-02
2.54640758e-01 8.11982214e-01 2.82590687e-01 -8.59160423e-01
7.76526481e-02 -1.23834923e-01 4.65323418e-01 -4.42633450e-01
2.69858479e-01 -7.39142239e-01 2.03907549e-01 9.24664259e-01
-6.63105726e-01 4.42556664e-03 3.87735218e-01 9.67933118e-01
-1.91196501e-01 1.97244048e-01 4.52535629e-01 -4.50629771e-01
-6.88727736e-01 5.43544888e-01 -2.49755234e-01 -1.80265471e-01
1.18681669e+00 1.03846774e-01 -1.41218022e-01 -3.76306891e-01
-1.73973417e+00 3.80295902e-01 7.00024888e-02 4.86017019e-01
7.54748881e-01 -1.24234009e+00 -7.29873538e-01 -1.32013947e-01
6.42043352e-01 -3.00865084e-01 3.90373439e-01 4.71659899e-01
-6.10144913e-01 7.11718082e-01 -6.61381334e-02 -5.65229058e-01
-1.31080854e+00 9.47752059e-01 4.91422027e-01 -7.74796128e-01
-6.92882717e-01 5.54635584e-01 1.18889615e-01 -4.05569255e-01
8.74822214e-02 -6.20600343e-01 -4.49341536e-01 1.87357575e-01
4.52049524e-01 1.09978817e-01 1.63743481e-01 -3.72732669e-01
-4.90679443e-01 1.11415617e-01 -3.71369332e-01 3.35256130e-01
1.56463981e+00 2.30945274e-01 1.10142380e-02 4.99996275e-01
9.68796611e-01 -6.10468507e-01 -5.04347682e-01 -3.91690493e-01
3.29157114e-01 2.29696315e-02 -2.37511352e-01 -1.08183146e+00
-8.25164437e-01 5.37445784e-01 6.01625383e-01 2.30765596e-01
1.03054464e+00 5.38210094e-01 3.27170491e-01 5.68052828e-01
2.80209720e-01 -4.33138877e-01 3.70357707e-02 5.07434197e-02
6.62341118e-01 -1.05692589e+00 1.40537411e-01 -5.33339739e-01
-8.76157403e-01 1.15455210e+00 2.61182427e-01 -8.65656137e-02
6.95093811e-01 -1.23918995e-01 8.09125081e-02 -7.00304985e-01
-9.79327142e-01 -3.56573373e-01 7.49920130e-01 6.40881896e-01
6.23838186e-01 5.70466101e-01 -8.22178647e-02 5.83903015e-01
2.73675770e-01 2.47446448e-01 4.69374567e-01 1.05654764e+00
-2.00197786e-01 -1.23178017e+00 -1.10909887e-01 1.02792227e+00
-5.75159132e-01 -2.36850560e-01 -3.78164351e-01 5.58700740e-01
-3.55214179e-02 5.91202676e-01 -1.38175964e-01 -3.46199036e-01
5.38459480e-01 4.93188918e-01 5.36625445e-01 -1.13862813e+00
-5.40794015e-01 -1.52379453e-01 1.54205054e-01 -7.63289332e-01
-4.58243251e-01 -4.28218365e-01 -1.44480455e+00 3.49163771e-01
2.12457374e-01 1.23422310e-01 2.87305027e-01 7.17647552e-01
6.85966909e-01 1.31315482e+00 9.69408825e-02 -2.85318851e-01
-3.04050863e-01 -7.10822761e-01 -5.30919671e-01 6.09030664e-01
4.92240608e-01 -3.97764266e-01 5.72523475e-02 2.05906719e-01] | [7.901112079620361, 6.934577941894531] |
d458ead6-3415-4288-9f73-2228267262f3 | automatic-fast-and-robust-characterization-of | 1805.12071 | null | http://arxiv.org/abs/1805.12071v2 | http://arxiv.org/pdf/1805.12071v2.pdf | Automatic, fast and robust characterization of noise distributions for diffusion MRI | Knowledge of the noise distribution in magnitude diffusion MRI images is the
centerpiece to quantify uncertainties arising from the acquisition process. The
use of parallel imaging methods, the number of receiver coils and imaging
filters applied by the scanner, amongst other factors, dictate the resulting
signal distribution. Accurate estimation beyond textbook Rician or noncentral
chi distributions often requires information about the acquisition process
(e.g. coils sensitivity maps or reconstruction coefficients), which is not
usually available. We introduce a new method where a change of variable
naturally gives rise to a particular form of the gamma distribution for
background signals. The first moments and maximum likelihood estimators of this
gamma distribution explicitly depend on the number of coils, making it possible
to estimate all unknown parameters using only the magnitude data. A rejection
step is used to make the method automatic and robust to artifacts. Experiments
on synthetic datasets show that the proposed method can reliably estimate both
the degrees of freedom and the standard deviation. The worst case errors range
from below 2% (spatially uniform noise) to approximately 10% (spatially
variable noise). Repeated acquisitions of in vivo datasets show that the
estimated parameters are stable and have lower variances than compared methods. | ['Samuel St-Jean', 'Max A. Viergever', 'Alexander Leemans', 'Alberto De Luca'] | 2018-05-30 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 3.30529213e-01 -4.59822893e-01 3.06436688e-01 -4.46302950e-01
-6.29778028e-01 -6.67750776e-01 4.10887599e-01 1.52026847e-01
-8.17531943e-01 1.11136663e+00 -9.01721269e-02 -2.55793601e-01
-3.33644748e-01 -2.15499848e-01 -5.76974988e-01 -1.16471291e+00
-5.01358688e-01 4.92922157e-01 5.60188234e-01 4.07649547e-01
2.77372062e-01 5.39779246e-01 -1.04262650e+00 -3.11315566e-01
8.89329851e-01 9.50104535e-01 4.64525312e-01 7.07767546e-01
1.28040552e-01 6.40249908e-01 -7.60574162e-01 1.22804470e-01
1.38471082e-01 -5.38063109e-01 -3.82763386e-01 8.91957432e-02
-2.71527261e-01 -3.81772220e-01 1.81993932e-01 1.46400940e+00
8.25587332e-01 2.43867770e-01 1.13461924e+00 -5.85703552e-01
9.11573321e-03 7.69309521e-01 -7.82701313e-01 5.96588552e-01
-8.52011666e-02 1.68263048e-01 -8.79769549e-02 -5.22220194e-01
5.03037691e-01 7.40866005e-01 6.64315164e-01 9.14529432e-03
-1.48304236e+00 -4.37539071e-01 -4.13841516e-01 7.74093568e-02
-1.51410306e+00 -4.47902709e-01 6.80207372e-01 -7.45910108e-01
4.04071689e-01 1.37812763e-01 3.85355860e-01 7.13139951e-01
7.75091648e-01 -3.72330211e-02 1.74073982e+00 -5.40243626e-01
6.16644979e-01 2.14586228e-01 2.09403992e-01 1.92706764e-01
6.44612432e-01 1.80832203e-02 1.10344276e-01 -3.46569151e-01
1.06522381e+00 -4.92704451e-01 -6.17362618e-01 -6.42379105e-01
-1.22294855e+00 7.02890694e-01 3.19918245e-02 7.11280525e-01
-7.83453166e-01 1.43495306e-01 2.89239764e-01 -1.37343511e-01
8.28797668e-02 2.61043340e-01 -2.01785695e-02 -1.96524560e-02
-9.77750838e-01 9.53844562e-02 6.92852914e-01 6.71262026e-01
3.34365487e-01 1.48748055e-01 -1.34083658e-01 5.50175369e-01
1.26889080e-01 9.69939470e-01 4.64625537e-01 -7.88853228e-01
3.62754203e-02 -4.80683178e-01 3.96071196e-01 -7.89545178e-01
-5.90786755e-01 -6.26108110e-01 -8.63224387e-01 1.56150982e-01
9.46574509e-01 -5.50249934e-01 -9.51001525e-01 1.54715312e+00
3.96941483e-01 2.99609895e-03 -3.96292359e-01 1.10959136e+00
2.00683624e-01 1.69086859e-01 3.57067846e-02 -8.12299073e-01
1.30928767e+00 -1.11768126e-01 -1.26010036e+00 -7.95149151e-03
2.32988417e-01 -8.99218380e-01 4.58046198e-01 5.03166497e-01
-1.19288933e+00 1.55167868e-02 -9.90209579e-01 7.06370890e-01
2.32607141e-01 -2.16871485e-01 2.55588084e-01 8.59596193e-01
-9.13994014e-01 5.04851878e-01 -1.03550887e+00 1.64794907e-01
1.43085912e-01 4.26709235e-01 -1.43343493e-01 1.90780200e-02
-1.12887073e+00 1.26282704e+00 2.45870993e-01 2.73127794e-01
-6.64370298e-01 -7.26377845e-01 -4.85187322e-01 -2.04798922e-01
1.48827687e-01 -2.88453072e-01 1.18932033e+00 -8.65354478e-01
-1.56789398e+00 2.93132514e-01 4.08959650e-02 -3.54587734e-01
7.43561447e-01 2.93298930e-01 -3.09861869e-01 5.93178689e-01
4.73756194e-02 -1.00936644e-01 1.06356096e+00 -1.02954483e+00
2.19384417e-01 -3.72299463e-01 -5.82044363e-01 -2.58284602e-02
5.65829039e-01 1.17878085e-02 7.97316581e-02 -4.20577139e-01
3.76335293e-01 -8.84500027e-01 -4.01311517e-01 -2.80702651e-01
-2.50081211e-01 5.61021447e-01 9.76482108e-02 -8.79308343e-01
8.41816366e-01 -1.98442554e+00 -3.05473030e-01 6.46938086e-01
1.51786149e-01 -2.02043846e-01 2.52065867e-01 8.63626078e-02
-3.53657663e-01 -3.97127986e-01 -5.72669029e-01 5.51171422e-01
-3.74168992e-01 -7.50792995e-02 2.29888543e-01 1.25416625e+00
-2.62073070e-01 3.69818926e-01 -7.99981236e-01 -3.37172657e-01
3.42479467e-01 7.87421882e-01 -2.48207614e-01 1.71614252e-02
3.49464834e-01 1.07177877e+00 -5.23943484e-01 -8.43441039e-02
1.15347207e+00 -1.14026181e-01 3.42084289e-01 -4.76075441e-01
-1.02380589e-01 8.70697796e-02 -1.42230833e+00 1.27920759e+00
-3.84650916e-01 3.92601311e-01 4.69040841e-01 -1.01308835e+00
6.65376604e-01 4.70283657e-01 6.32516205e-01 -6.85499310e-01
5.47085464e-01 4.35823947e-01 6.35363460e-01 -5.92148542e-01
-3.32405008e-02 -7.35957682e-01 2.79378593e-01 4.28746164e-01
5.08454107e-02 -2.35969409e-01 1.72806606e-02 -6.76015718e-03
9.26743627e-01 -2.45879650e-01 4.35296386e-01 -1.04845011e+00
3.68732512e-01 -3.11140567e-01 4.26239550e-01 8.12044382e-01
-2.97443300e-01 5.74840128e-01 5.40469587e-01 -6.63180947e-02
-1.05844188e+00 -1.09448433e+00 -8.85509074e-01 -1.25490740e-01
8.82024616e-02 3.06089729e-01 -1.06192362e+00 -1.79648757e-01
-3.13170612e-01 7.52578139e-01 -4.83722329e-01 9.92557704e-02
-6.22347057e-01 -1.18331802e+00 1.33162439e-01 2.07826644e-01
2.24856496e-01 -5.56853712e-01 -1.06768239e+00 4.59287554e-01
-3.82117361e-01 -1.19457877e+00 -3.96069288e-01 4.85259861e-01
-1.09376609e+00 -1.05363989e+00 -1.04000545e+00 1.02553489e-02
8.08497369e-01 -1.69793189e-01 9.38579261e-01 -4.37877804e-01
-3.93226475e-01 3.72269094e-01 -7.99841657e-02 -1.97109699e-01
-4.05955970e-01 -4.96081889e-01 8.19660574e-02 -4.00877930e-02
-8.91076215e-03 -6.46442354e-01 -7.87077308e-01 2.71812379e-01
-9.62326646e-01 -5.73168755e-01 5.39039016e-01 7.85118401e-01
5.43246686e-01 2.99738407e-01 4.39906567e-01 -8.23712349e-01
6.49250448e-01 -3.36059928e-01 -8.59685302e-01 -2.94499937e-02
-4.09778386e-01 4.58810449e-01 3.60336751e-01 -7.60991931e-01
-1.08452880e+00 1.61408186e-02 1.46858394e-01 -2.03704908e-01
-2.33511075e-01 2.98014015e-01 -5.87303564e-02 -2.64661670e-01
8.74574542e-01 2.41479144e-01 2.13175684e-01 -2.35447794e-01
1.70064017e-01 3.86441380e-01 5.34883797e-01 -4.77461785e-01
3.02314818e-01 3.58553320e-01 4.16064411e-01 -1.13642168e+00
-9.79208425e-02 -2.13563770e-01 -7.03594625e-01 -3.64477932e-01
8.82271767e-01 -5.36704183e-01 -6.16668701e-01 6.00740671e-01
-9.37856495e-01 -1.84980422e-01 -7.40893558e-02 1.18158174e+00
-5.65339327e-01 5.94801605e-01 -4.91376698e-01 -1.08464265e+00
-1.89251751e-01 -1.49840999e+00 4.37904894e-01 7.42582008e-02
-2.05839276e-01 -1.07801867e+00 -1.63484335e-01 -1.22785509e-01
8.79060149e-01 3.74768168e-01 8.79236162e-01 -3.55292380e-01
-4.23089027e-01 -1.46915242e-01 2.69402236e-01 2.82335728e-01
9.38398242e-02 -5.87723672e-01 -7.63350487e-01 -2.05703035e-01
9.23894763e-01 2.89180040e-01 3.51034403e-01 1.30290377e+00
9.20813560e-01 -1.39520139e-01 -2.62146413e-01 3.02307516e-01
1.46857846e+00 4.81368303e-01 7.92017996e-01 8.03227276e-02
2.61787683e-01 4.84907001e-01 3.23710144e-01 5.06079435e-01
-1.81858122e-01 6.04295850e-01 7.78149068e-02 2.81974077e-01
3.80906798e-02 3.96382183e-01 -1.40532419e-01 7.10084975e-01
2.42986325e-02 7.15400875e-02 -8.95837724e-01 4.39725816e-01
-1.15589893e+00 -8.63461971e-01 -6.32225096e-01 2.85371232e+00
9.04811203e-01 -7.43673369e-02 -4.91996296e-02 1.70476824e-01
9.28466797e-01 -2.17757285e-01 -4.10988748e-01 -7.04584271e-02
7.57551789e-02 2.67752916e-01 1.11127543e+00 8.74464214e-01
-8.49209964e-01 8.35300013e-02 7.04241228e+00 6.36894166e-01
-1.26248205e+00 2.88683832e-01 3.89035106e-01 1.74588904e-01
-3.60445470e-01 -5.40268384e-02 -3.61283123e-01 8.62061262e-01
1.08854592e+00 -1.23921469e-01 5.16371429e-01 2.70506024e-01
3.62801611e-01 -9.46310699e-01 -6.15974188e-01 1.04222322e+00
4.12019901e-02 -4.89946961e-01 -5.59298038e-01 9.16987732e-02
5.01765847e-01 -9.61215943e-02 -1.15989618e-01 -4.53840494e-01
-2.98895668e-02 -8.94196093e-01 6.26770794e-01 8.88999999e-01
8.06964159e-01 -7.58722961e-01 8.82449090e-01 4.70859915e-01
-6.08488023e-01 3.99220377e-01 -3.76395792e-01 2.43017375e-01
4.99392211e-01 1.40236616e+00 -8.37112606e-01 2.54586399e-01
3.76626015e-01 -3.39621492e-02 -2.32840508e-01 1.41225815e+00
-2.30907813e-01 6.18558764e-01 -7.72400975e-01 8.46677646e-02
-1.30855933e-01 -4.38080639e-01 7.12236583e-01 9.72345114e-01
5.02369046e-01 3.54155183e-01 -3.22285235e-01 8.41423333e-01
4.86203283e-01 -3.05979401e-02 -3.22985023e-01 2.11631387e-01
4.43807989e-01 1.14597142e+00 -1.08282280e+00 -1.24105886e-01
-7.42535219e-02 6.73862636e-01 -2.74495810e-01 5.49502254e-01
-7.02905834e-01 -2.66712546e-01 2.85002496e-03 5.08033335e-01
4.25983876e-01 -3.88859093e-01 -2.41243243e-01 -9.61544693e-01
9.98417884e-02 -7.62066603e-01 -5.65820076e-02 -5.85420430e-01
-1.23403811e+00 7.44961202e-01 5.13882458e-01 -1.00055921e+00
-7.10927963e-01 -3.87966365e-01 -3.80103320e-01 1.34324026e+00
-1.11445904e+00 -1.64628327e-01 4.71007973e-02 6.19577527e-01
-2.40047500e-01 2.65611857e-01 6.12371445e-01 3.76184762e-01
-1.41447624e-02 1.92381084e-01 4.53888655e-01 -9.81493592e-02
6.09136641e-01 -1.05444205e+00 -3.42884868e-01 8.22995961e-01
-4.11638319e-01 7.54684925e-01 1.17005682e+00 -7.33220935e-01
-1.12965095e+00 -3.75458181e-01 3.56865376e-01 -5.02916761e-02
6.58468783e-01 -2.62949944e-01 -1.00779557e+00 3.67813051e-01
7.47433230e-02 3.19860190e-01 4.78265136e-01 -3.43397796e-01
3.40212971e-01 -6.07895814e-02 -1.44185841e+00 -1.16040120e-02
2.94381469e-01 -1.93741888e-01 -3.92570138e-01 1.35995165e-01
-1.17491513e-01 -5.54213345e-01 -8.95655572e-01 2.13941216e-01
4.78023678e-01 -9.74103928e-01 7.11238980e-01 1.70655884e-02
-2.72754312e-01 -4.00105357e-01 4.24043648e-02 -1.48764360e+00
-2.93887615e-01 -3.09658200e-01 3.29219908e-01 7.79212832e-01
1.78137854e-01 -7.84779012e-01 3.30639094e-01 6.56446218e-01
2.39807531e-01 -2.56960601e-01 -1.25189900e+00 -8.00548434e-01
4.23968658e-02 -5.03922284e-01 7.18913227e-02 7.67786503e-01
-2.28699490e-01 7.93010890e-02 -2.09818900e-01 2.95554250e-01
1.30726254e+00 -2.84050703e-01 6.90245628e-02 -1.10375011e+00
-4.29121524e-01 2.26343540e-03 -5.42590201e-01 -7.57111847e-01
-1.58360526e-01 -5.46552420e-01 2.93526262e-01 -1.11127603e+00
1.16292790e-01 -6.04283035e-01 -5.51608913e-02 -3.11977267e-01
-1.12341031e-01 -2.81181615e-02 -2.96100020e-01 8.20107386e-02
1.02734253e-01 2.43374884e-01 1.26880956e+00 2.39796281e-01
-7.82737392e-04 7.13219792e-02 -1.97484225e-01 8.72509778e-01
6.38483644e-01 -7.95928538e-01 -5.11765480e-01 -2.88972348e-01
-1.54339239e-01 5.13520479e-01 2.30682507e-01 -1.14539051e+00
2.31909335e-01 1.45681828e-01 6.43537462e-01 -4.06440377e-01
-2.21698787e-02 -1.09706891e+00 7.25598156e-01 5.12054741e-01
-1.21791400e-01 -7.99184106e-03 6.87968209e-02 3.50842535e-01
-3.86504922e-03 -8.06063831e-01 1.20620143e+00 -1.92803025e-01
6.29889220e-02 -4.69250903e-02 -7.40222573e-01 1.67743117e-01
7.73489356e-01 5.29545508e-02 1.78972796e-01 -5.58905125e-01
-9.81124222e-01 -3.48303080e-01 1.91570938e-01 -3.34894300e-01
3.41594726e-01 -1.10987115e+00 -6.15271389e-01 2.08771542e-01
-4.87819284e-01 -3.29509676e-01 6.10550523e-01 1.59995663e+00
-6.67208672e-01 2.34929502e-01 -2.46969819e-01 -8.51091921e-01
-8.10021996e-01 5.10988235e-01 7.81946778e-01 -1.51336700e-01
-5.23522079e-01 4.77743983e-01 8.73338282e-02 1.22943684e-01
-1.16247214e-01 -2.80349612e-01 -4.76394780e-02 -9.06606689e-02
8.61581683e-01 3.40521693e-01 2.92872965e-01 -5.83828628e-01
-5.34060478e-01 6.02191329e-01 5.93903661e-03 -4.54213887e-01
1.22115135e+00 -3.26536387e-01 -1.66579008e-01 5.44358790e-01
9.62687194e-01 5.48741333e-02 -1.21375084e+00 -5.03588133e-02
-1.05589740e-01 -5.07770002e-01 4.28438902e-01 -9.67076659e-01
-9.38945770e-01 8.99918199e-01 1.05859900e+00 3.34505513e-02
1.09610927e+00 -2.93564916e-01 2.58513510e-01 -2.35948786e-01
7.43209302e-01 -8.57139349e-01 -4.25450474e-01 2.18563557e-01
7.48965323e-01 -8.07293832e-01 2.09382966e-01 -4.75411147e-01
-6.53291523e-01 9.76732194e-01 -1.12879366e-01 -2.09148437e-01
9.08692777e-01 8.60300183e-01 2.25508243e-01 4.09324504e-02
-1.04885183e-01 1.72681049e-01 2.39299923e-01 8.97761345e-01
6.21729374e-01 1.39371291e-01 -8.48551869e-01 5.83265066e-01
1.67457387e-02 1.22486979e-01 7.98102617e-01 7.84246504e-01
-3.47203791e-01 -9.79464591e-01 -8.35274994e-01 4.44682181e-01
-9.06951368e-01 -5.60159869e-02 4.51929390e-01 6.64433599e-01
-1.51849672e-01 7.97578931e-01 1.72899067e-02 4.54936057e-01
2.80248761e-01 -9.14328396e-02 1.01790082e+00 -8.22650865e-02
-3.02447796e-01 5.82665443e-01 -1.00442849e-01 -3.42013031e-01
-5.48629284e-01 -8.70588422e-01 -1.31227970e+00 -5.27155250e-02
-6.54288232e-01 4.44253266e-01 1.06108332e+00 9.62740719e-01
-1.80264547e-01 5.08393586e-01 2.54243255e-01 -6.78754449e-01
-8.91879618e-01 -1.16549170e+00 -1.15062714e+00 1.57223940e-01
2.32258365e-01 -8.20868254e-01 -5.58283448e-01 -6.85262606e-02] | [13.37751579284668, -2.4774341583251953] |
7b65548c-9fe1-48f0-b7aa-7a58c0be9f4e | systematic-review-on-reinforcement-learning | 2305.07466 | null | https://arxiv.org/abs/2305.07466v1 | https://arxiv.org/pdf/2305.07466v1.pdf | Systematic Review on Reinforcement Learning in the Field of Fintech | Applications of Reinforcement Learning in the Finance Technology (Fintech) have acquired a lot of admiration lately. Undoubtedly Reinforcement Learning, through its vast competence and proficiency, has aided remarkable results in the field of Fintech. The objective of this systematic survey is to perform an exploratory study on a correlation between reinforcement learning and Fintech to highlight the prediction accuracy, complexity, scalability, risks, profitability and performance. Major uses of reinforcement learning in finance or Fintech include portfolio optimization, credit risk reduction, investment capital management, profit maximization, effective recommendation systems, and better price setting strategies. Several studies have addressed the actual contribution of reinforcement learning to the performance of financial institutions. The latest studies included in this survey are publications from 2018 onward. The survey is conducted using PRISMA technique which focuses on the reporting of reviews and is based on a checklist and four-phase flow diagram. The conducted survey indicates that the performance of RL-based strategies in Fintech fields proves to perform considerably better than other state-of-the-art algorithms. The present work discusses the use of reinforcement learning algorithms in diverse decision-making challenges in Fintech and concludes that the organizations dealing with finance can benefit greatly from Robo-advising, smart order channelling, market making, hedging and options pricing, portfolio optimization, and optimal execution. | ['Rashid Mehmood', 'Iyad Katib', 'Nadeem Malibari'] | 2023-04-29 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-8.52881074e-01 1.10283501e-01 -4.34665740e-01 -1.67349372e-02
-2.40628764e-01 -3.92467290e-01 2.14407861e-01 2.14902192e-01
-2.47315913e-01 9.71883118e-01 1.18852049e-01 -5.15307963e-01
-8.82214010e-01 -1.05863881e+00 -2.58493513e-01 -5.34218550e-01
-1.17086448e-01 7.73174524e-01 -2.92228371e-01 -5.97830176e-01
1.14424694e+00 8.13013971e-01 -1.22767913e+00 9.48115960e-02
3.62121940e-01 1.19234264e+00 -2.89248000e-03 3.74008417e-01
-2.59820431e-01 1.25834179e+00 -7.40195572e-01 -5.71905673e-01
6.05174363e-01 -8.49389359e-02 -5.79348207e-01 -2.87956774e-01
-4.25718158e-01 -5.83828330e-01 -6.85134456e-02 4.10146326e-01
6.91366613e-01 -1.96811017e-02 5.10042906e-01 -1.14466166e+00
-5.23769379e-01 6.44866943e-01 -3.85364115e-01 4.52829987e-01
2.15507314e-01 2.98482895e-01 1.00149155e+00 -5.56965947e-01
2.33900055e-01 9.02052641e-01 4.25996780e-01 1.08817555e-01
-5.03720224e-01 -7.62446523e-01 -1.97676644e-01 5.77447534e-01
-4.70424026e-01 2.93434352e-01 5.68937659e-01 -5.33505917e-01
1.59179544e+00 -1.23913072e-01 1.22138381e+00 4.91905630e-01
6.65956438e-01 4.57293063e-01 1.46094716e+00 -6.30430520e-01
4.79404569e-01 5.10666847e-01 -2.01179653e-01 1.37568042e-01
4.49476659e-01 8.46679091e-01 -3.98798585e-01 1.89117178e-01
9.99657452e-01 2.92363279e-02 5.27111590e-01 -6.49877638e-02
-5.55251062e-01 1.38511908e+00 4.90635149e-02 3.34482074e-01
-5.95402718e-01 1.04366973e-01 4.54641014e-01 6.82361186e-01
6.65654335e-03 8.82415891e-01 -6.34707808e-01 -6.17276788e-01
-9.04069304e-01 6.19287848e-01 9.78217542e-01 5.87302208e-01
1.77560866e-01 6.72669709e-01 -4.79051620e-02 5.07393181e-01
4.82976913e-01 4.33844864e-01 5.53984106e-01 -1.17523849e+00
3.94507110e-01 4.34483409e-01 3.86663646e-01 -6.37251973e-01
-5.53197384e-01 -4.61113870e-01 -1.56131893e-01 7.57227302e-01
9.83632803e-02 -7.42726862e-01 -8.36428329e-02 6.45087540e-01
1.18601173e-01 -2.09630266e-01 1.86423257e-01 4.61305767e-01
5.65554738e-01 4.93964195e-01 -9.99506488e-02 -3.75489652e-01
1.10456550e+00 -9.19398963e-01 -8.72151434e-01 4.29123729e-01
2.55083919e-01 -9.14854169e-01 5.07932842e-01 1.02280867e+00
-1.24575329e+00 -4.96234328e-01 -1.24367905e+00 8.01077247e-01
-5.08727670e-01 -1.50150955e-01 8.50954115e-01 1.15629256e+00
-6.19410276e-01 1.02014327e+00 -3.08341146e-01 1.79394215e-01
4.50242609e-01 4.86561686e-01 4.02816534e-01 4.44514036e-01
-1.37046540e+00 1.32577956e+00 2.71986991e-01 -3.32811683e-01
-8.20739388e-01 -6.62645221e-01 -2.64763236e-01 9.62915421e-02
4.71935898e-01 -2.30648950e-01 1.33932829e+00 -6.63023829e-01
-1.93472159e+00 2.48823583e-01 8.94179821e-01 -1.11994529e+00
6.96791053e-01 -4.95698512e-01 -4.26558614e-01 1.85685068e-01
-2.44588271e-01 7.65822157e-02 4.46129501e-01 -4.57983404e-01
-1.21165419e+00 -3.18994224e-01 1.38680756e-01 1.35050476e-01
-1.65232465e-01 -2.10622847e-02 4.53299522e-01 -8.99456203e-01
-5.31177759e-01 -3.65844876e-01 -3.20097625e-01 -9.03205752e-01
4.78934139e-01 -4.08577234e-01 6.67207956e-01 -6.95814073e-01
1.42981434e+00 -1.57923710e+00 -5.81381679e-01 3.36031288e-01
-3.52725178e-01 2.69151390e-01 4.82483476e-01 9.98957694e-01
-3.79837416e-02 2.35271156e-01 5.39241731e-01 7.30165064e-01
1.96172342e-01 1.55375347e-01 -2.85885930e-01 1.58898920e-01
1.44433543e-01 8.16409349e-01 -7.28435636e-01 -2.15594262e-01
7.39391446e-01 -2.02424631e-01 -5.30784428e-01 3.94076228e-01
-2.08990455e-01 1.73584759e-01 -6.68850005e-01 9.89473403e-01
5.67059338e-01 9.84809473e-02 2.31908202e-01 1.87636152e-01
-4.41233754e-01 1.59112751e-01 -1.50206792e+00 7.99133301e-01
-6.53673828e-01 3.27951670e-01 -5.30942202e-01 -1.35453475e+00
1.41831565e+00 4.34444815e-01 1.00172460e+00 -1.21061373e+00
6.88288435e-02 4.62972343e-01 7.70783722e-02 -7.85527050e-01
4.20501024e-01 -3.97469133e-01 2.89027572e-01 7.14568853e-01
5.85890003e-02 -2.79242069e-01 3.10131431e-01 -2.50960737e-01
8.46774161e-01 6.34667039e-01 3.81218910e-01 -1.93353206e-01
5.52168190e-01 -5.40300179e-03 3.69514853e-01 7.33359516e-01
-4.41698492e-01 -2.98904955e-01 8.19578528e-01 -5.58285117e-01
-9.86774802e-01 -7.77975500e-01 -1.92369729e-01 5.85892439e-01
-4.41084146e-01 7.68453628e-02 -6.26675963e-01 -7.56377637e-01
4.32647437e-01 9.24540639e-01 -3.94713044e-01 1.53326690e-01
-4.41710413e-01 -6.06584072e-01 1.97724536e-01 5.70192397e-01
7.71423817e-01 -1.65227473e+00 -1.09979558e+00 6.97726369e-01
6.72289550e-01 -7.36732721e-01 4.92117941e-01 3.87437731e-01
-1.32105494e+00 -1.23844540e+00 -7.30410457e-01 -4.45590526e-01
-8.30012560e-02 -2.89142281e-01 1.10865450e+00 -1.68055683e-01
-2.89228737e-01 5.08157670e-01 -5.00930488e-01 -9.16129470e-01
-1.48390293e-01 -2.85011709e-01 -1.51260078e-01 -6.84955478e-01
5.68880618e-01 -2.28819311e-01 -6.86216295e-01 3.08716863e-01
-4.04645711e-01 -8.40096891e-01 7.47613072e-01 1.02632499e+00
4.86823380e-01 5.84596276e-01 1.41246116e+00 -1.06580484e+00
8.11524332e-01 -6.64052546e-01 -1.02174723e+00 2.67078280e-01
-1.55607700e+00 -1.08904086e-01 4.22188640e-01 2.08781838e-01
-1.19328749e+00 -6.32340848e-01 1.95337855e-03 -1.06750719e-01
2.23112389e-01 4.78835195e-01 2.39175811e-01 -1.20617740e-01
4.53561753e-01 -1.39479369e-01 1.80264384e-01 -1.75240681e-01
7.69794360e-02 6.04997933e-01 -1.18980341e-01 -5.89361727e-01
4.46806312e-01 -4.82594967e-02 1.98121712e-01 -3.99197519e-01
-4.94412154e-01 -2.54141808e-01 -3.21770519e-01 -6.51732385e-01
5.08740366e-01 -4.62927729e-01 -1.07607889e+00 3.76277238e-01
-2.63390869e-01 -2.91221350e-01 -7.02850819e-01 8.14194441e-01
-7.56764889e-01 -5.01750000e-02 -5.96283913e-01 -1.14767039e+00
-5.08788466e-01 -1.05914378e+00 -1.07610278e-01 6.81248665e-01
-5.83068244e-02 -1.30182230e+00 1.24313071e-01 6.34311736e-01
5.79708636e-01 4.76468712e-01 8.52334023e-01 -8.29047263e-01
-4.36060250e-01 -2.83340007e-01 1.47987276e-01 7.55046368e-01
9.87473801e-02 4.99346405e-02 -5.57082117e-01 2.88828742e-02
3.19798470e-01 -5.15248716e-01 2.40211740e-01 6.35117054e-01
9.77599859e-01 -2.99104393e-01 3.54187459e-01 1.05627239e-01
1.66180325e+00 9.43689108e-01 6.86957121e-01 1.39947367e+00
-5.75117953e-02 9.90021646e-01 1.73694324e+00 1.26236129e+00
-9.93285179e-02 3.10354590e-01 6.01151049e-01 5.52690387e-01
4.55797583e-01 -1.14938561e-02 4.44330573e-01 3.97636622e-01
-3.63262683e-01 2.15322539e-01 -7.86175191e-01 8.49475041e-02
-1.69947934e+00 -1.17230248e+00 4.91116419e-02 2.24125814e+00
1.83754191e-01 6.43470526e-01 5.57744563e-01 2.92649746e-01
4.19225007e-01 -2.60226488e-01 -6.98761702e-01 -1.16565418e+00
3.41357663e-02 7.61329770e-01 8.99581432e-01 5.92680834e-02
-8.95228267e-01 8.19751978e-01 6.89909077e+00 8.89443934e-01
-1.12177956e+00 -2.02281028e-01 8.86260211e-01 -2.60636419e-01
-2.38094315e-01 -2.22495082e-03 -8.17701638e-01 4.22765255e-01
1.30696666e+00 -4.78408426e-01 4.92087960e-01 1.22470999e+00
6.91204131e-01 -2.00227275e-01 -5.32904148e-01 7.15401649e-01
-4.21219200e-01 -1.80501151e+00 -1.28748879e-01 2.23821089e-01
9.66847420e-01 -3.50387812e-01 3.27166557e-01 6.48261607e-01
4.74144459e-01 -1.29858541e+00 6.79087877e-01 6.92228079e-01
3.73611003e-02 -1.57054842e+00 1.29996860e+00 -1.87502518e-01
-7.64221430e-01 -7.28910089e-01 -6.17724717e-01 -3.99931192e-01
6.84945583e-02 5.53425789e-01 -9.44881320e-01 9.22761858e-01
9.09616530e-01 8.41765761e-01 -1.75842866e-01 1.07575428e+00
-3.05902898e-01 7.95212507e-01 2.65304744e-01 -5.04247665e-01
5.08442283e-01 -6.77959323e-01 2.74971239e-02 9.29107666e-01
2.36673728e-01 1.58151926e-03 -1.80865377e-01 5.12278378e-01
3.48593950e-01 4.98327523e-01 -3.99224907e-01 -4.01385963e-01
4.19553638e-01 1.15580678e+00 -7.31955111e-01 -2.06742749e-01
-5.29610097e-01 9.56302732e-02 -2.31456205e-01 -4.28843834e-02
-7.12386072e-01 -4.11314636e-01 2.63585895e-01 1.97223127e-01
6.17648125e-01 1.14375569e-01 -8.93402338e-01 -2.59960145e-01
-1.97673023e-01 -1.12395370e+00 7.08442509e-01 -3.42323720e-01
-1.06060648e+00 -7.14336932e-02 -1.17726147e-01 -1.29568899e+00
-5.22094846e-01 -1.03051686e+00 -5.93891501e-01 7.07898974e-01
-1.53387177e+00 -5.43353736e-01 2.03822047e-01 3.82019013e-01
5.13136268e-01 -1.20172226e+00 5.53547919e-01 7.44921193e-02
-3.21819991e-01 3.97088259e-01 6.44226551e-01 -3.11499029e-01
4.35190737e-01 -1.50529492e+00 -3.45229983e-01 4.80801389e-02
-3.97924960e-01 1.95782870e-01 3.88211071e-01 -8.00147355e-01
-1.32409048e+00 -5.73170304e-01 4.29300100e-01 -2.01734509e-02
8.78079832e-01 5.67032099e-01 -2.85976619e-01 1.23703577e-01
4.75273997e-01 -7.09714830e-01 7.87623167e-01 -1.09305389e-01
3.71257126e-01 -5.93692422e-01 -1.43869209e+00 2.21403942e-01
2.85860300e-02 7.22207921e-03 -6.01777375e-01 2.56531060e-01
1.62493736e-01 -2.14623302e-01 -1.31756711e+00 2.27720648e-01
7.08538294e-01 -1.39152682e+00 9.24431503e-01 -6.49958253e-01
7.59826779e-01 4.49618101e-01 1.42132118e-01 -8.53765965e-01
-2.40357503e-01 -6.77768230e-01 -2.32108384e-01 1.20426619e+00
2.17421994e-01 -7.73246944e-01 1.12480462e+00 2.49606013e-01
5.00894450e-02 -1.36527312e+00 -9.34998691e-01 -8.69259059e-01
5.40750504e-01 -2.35123008e-01 7.05757976e-01 6.43921912e-01
2.50779334e-02 -1.39102742e-01 -4.09745634e-01 -4.93335456e-01
5.89407682e-01 7.34596848e-02 5.72244227e-01 -9.94098246e-01
-6.41339839e-01 -6.73631072e-01 -2.44271353e-01 -1.41304418e-01
-2.87606806e-01 -4.37694341e-01 -8.83542061e-01 -1.51291037e+00
-3.07545871e-01 -4.09002364e-01 -7.05952644e-01 1.05275325e-01
3.72944802e-01 -4.88733232e-01 2.75873303e-01 1.82564780e-01
-1.18333526e-01 3.18595946e-01 1.51461172e+00 2.59901471e-02
-2.63204038e-01 4.83596414e-01 -7.33974099e-01 6.04326010e-01
1.24540496e+00 -3.36096406e-01 -4.15459156e-01 5.15525460e-01
5.63083172e-01 4.92732078e-01 -1.66231930e-01 -8.09457302e-01
-2.16029689e-01 -5.70974350e-01 7.05526173e-01 -6.85371518e-01
-2.73498476e-01 -8.60857189e-01 -1.11838169e-02 1.10382152e+00
-5.64938523e-02 4.56090957e-01 2.13846594e-01 1.66165709e-01
-2.92989939e-01 -9.17710423e-01 6.10495150e-01 -4.91253346e-01
-7.04067051e-01 -8.09241906e-02 -5.72063446e-01 1.19615272e-01
1.47051322e+00 -4.78003740e-01 1.26384562e-02 -3.77508223e-01
-5.88260591e-01 3.37731987e-01 -1.36280805e-01 3.46370578e-01
6.75183952e-01 -1.01996243e+00 -6.26693189e-01 -3.20202798e-01
-6.10396624e-01 -5.82805395e-01 2.65526801e-01 6.66030467e-01
-9.98130560e-01 7.79829741e-01 -9.13669407e-01 1.27451867e-01
-7.57957160e-01 4.98249739e-01 6.03163719e-01 -1.07315016e+00
-4.01873410e-01 4.57651526e-01 -4.96493906e-01 -1.12300823e-02
1.78894818e-01 1.07658729e-01 -1.02892959e+00 4.83122289e-01
2.67798752e-01 1.17772985e+00 3.25383216e-01 2.11372748e-01
-2.89221834e-02 4.24343556e-01 -3.01305950e-01 -4.09355372e-01
1.83479440e+00 1.97538018e-01 3.01506162e-01 3.89601529e-01
2.77520925e-01 -1.28936335e-01 -1.35457921e+00 5.54566681e-01
6.01158261e-01 -5.97939610e-01 9.12612751e-02 -1.25183940e+00
-1.34310055e+00 7.26906300e-01 8.26331913e-01 2.56871283e-01
1.04397345e+00 -7.64516294e-01 5.29886067e-01 3.51450205e-01
5.57579696e-01 -2.02309752e+00 7.71170855e-01 5.38368940e-01
1.02231205e+00 -1.00168014e+00 2.76925743e-01 2.53151238e-01
-1.04100072e+00 1.77288640e+00 4.63448048e-01 -6.23527169e-01
1.07384753e+00 2.51262814e-01 1.83702245e-01 -1.54582039e-01
-4.97009844e-01 1.95418164e-01 -4.17947084e-01 7.97358155e-01
6.41197860e-01 1.39911860e-01 -7.55832136e-01 9.37696278e-01
-5.73757172e-01 2.65472502e-01 7.80460000e-01 1.16383374e+00
-5.90450644e-01 -1.71972811e+00 -7.05856383e-01 7.04750061e-01
-1.01088619e+00 6.23007417e-02 -9.52536613e-02 1.14432049e+00
3.65749262e-02 9.42581058e-01 -2.21343473e-01 -8.13038126e-02
4.82297629e-01 -1.88124537e-01 4.78740305e-01 -2.36244053e-01
-1.39757693e+00 6.24075830e-02 8.38268921e-02 -3.87127608e-01
-3.44153494e-01 -1.04160035e+00 -1.42010081e+00 -4.36131507e-01
1.11884316e-02 5.05867660e-01 8.05169225e-01 8.73771906e-01
-3.46985017e-03 1.01906037e+00 1.04569173e+00 -3.07961911e-01
-1.13001454e+00 -8.48334849e-01 -1.22499776e+00 -2.28767246e-01
-3.07562381e-01 -9.90874231e-01 -1.47341207e-01 -6.56321168e-01] | [4.456421375274658, 3.9221487045288086] |
65dd750b-aca9-414a-8b4d-88a9f3093cd5 | see-more-know-more-unsupervised-video-object-1 | 2001.06810 | null | https://arxiv.org/abs/2001.06810v1 | https://arxiv.org/pdf/2001.06810v1.pdf | See More, Know More: Unsupervised Video Object Segmentation with Co-Attention Siamese Networks | We introduce a novel network, called CO-attention Siamese Network (COSNet), to address the unsupervised video object segmentation task from a holistic view. We emphasize the importance of inherent correlation among video frames and incorporate a global co-attention mechanism to improve further the state-of-the-art deep learning based solutions that primarily focus on learning discriminative foreground representations over appearance and motion in short-term temporal segments. The co-attention layers in our network provide efficient and competent stages for capturing global correlations and scene context by jointly computing and appending co-attention responses into a joint feature space. We train COSNet with pairs of video frames, which naturally augments training data and allows increased learning capacity. During the segmentation stage, the co-attention model encodes useful information by processing multiple reference frames together, which is leveraged to infer the frequently reappearing and salient foreground objects better. We propose a unified and end-to-end trainable framework where different co-attention variants can be derived for mining the rich context within videos. Our extensive experiments over three large benchmarks manifest that COSNet outperforms the current alternatives by a large margin. | ['Ling Shao', 'Xiankai Lu', 'Jianbing Shen', 'Fatih Porikli', 'Chao Ma', 'Wenguan Wang'] | 2020-01-19 | see-more-know-more-unsupervised-video-object | http://openaccess.thecvf.com/content_CVPR_2019/html/Lu_See_More_Know_More_Unsupervised_Video_Object_Segmentation_With_Co-Attention_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Lu_See_More_Know_More_Unsupervised_Video_Object_Segmentation_With_Co-Attention_CVPR_2019_paper.pdf | cvpr-2019-6 | ['unsupervised-video-object-segmentation', 'video-polyp-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.15412155e-01 -3.17445427e-01 -4.08830523e-01 -3.90375882e-01
-8.03848922e-01 -3.44220102e-01 3.87104660e-01 -2.76847154e-01
-4.64731246e-01 3.91352892e-01 4.47074592e-01 2.69979566e-01
9.68972966e-02 -3.76125365e-01 -9.98898566e-01 -7.24808753e-01
-2.28094414e-01 5.48355058e-02 6.85865760e-01 2.36566305e-01
1.25729799e-01 2.91887462e-01 -1.28449345e+00 5.09127736e-01
7.25887120e-01 1.05245268e+00 5.72474718e-01 7.14757621e-01
7.18534365e-02 1.16532350e+00 -4.14777488e-01 -1.50761634e-01
2.11529464e-01 -4.90434408e-01 -7.89540291e-01 4.94495034e-01
7.90093482e-01 -5.85946143e-01 -5.35206079e-01 9.37425137e-01
9.56268385e-02 4.06874150e-01 3.14278603e-01 -9.81684685e-01
-5.81461966e-01 4.62649763e-01 -8.90848100e-01 8.39674532e-01
2.21797973e-01 3.75396490e-01 1.31268919e+00 -9.45845485e-01
7.48981416e-01 1.32428062e+00 4.38528299e-01 4.21348512e-01
-1.03489602e+00 -4.86508965e-01 9.51178491e-01 5.62549174e-01
-1.14333713e+00 -3.03312778e-01 9.57036436e-01 -3.65009308e-01
9.78954017e-01 2.62949280e-02 8.07718813e-01 1.17218173e+00
-1.07681468e-01 1.59698272e+00 3.58093083e-01 2.40280610e-02
-3.30241136e-02 -5.00107408e-01 1.79991238e-02 8.38931084e-01
1.44720688e-01 -2.40491152e-01 -6.11473560e-01 1.84396058e-01
9.36573565e-01 5.36560774e-01 -3.10235143e-01 -4.04758215e-01
-1.42322195e+00 7.46527374e-01 6.99694574e-01 2.98818707e-01
-6.02207065e-01 5.01655817e-01 4.63843048e-01 -1.84153870e-01
4.84577805e-01 3.05917770e-01 -7.64366627e-01 -4.20100987e-02
-1.13950455e+00 1.64212599e-01 3.08221787e-01 8.66642118e-01
8.16461325e-01 1.02661438e-01 -5.39429784e-01 6.88773513e-01
5.16022682e-01 3.03215295e-01 3.99514973e-01 -1.22872436e+00
5.20640433e-01 5.59238791e-01 -1.58800602e-01 -9.29772258e-01
6.35237759e-03 -5.13887465e-01 -5.39717019e-01 -2.69052267e-01
2.67812848e-01 -1.91875115e-01 -1.02881503e+00 1.78679609e+00
3.94145340e-01 9.40118730e-01 -8.63471106e-02 1.26070154e+00
6.31846368e-01 8.25896621e-01 3.74994159e-01 -1.05263762e-01
1.35399723e+00 -1.63257861e+00 -6.10500872e-01 -4.26771522e-01
2.69294918e-01 -5.93449354e-01 8.82512212e-01 1.95041031e-01
-1.21671271e+00 -8.17695081e-01 -8.79808843e-01 -2.74001271e-01
4.24741069e-03 1.17437214e-01 7.03801095e-01 -8.10483322e-02
-7.80500889e-01 6.42172992e-01 -1.16827321e+00 -1.96068078e-01
1.00411248e+00 2.95326263e-01 -1.67400375e-01 -2.78582841e-01
-8.52716625e-01 2.91229278e-01 2.64798850e-01 2.61465102e-01
-1.32463908e+00 -6.80817604e-01 -1.11033642e+00 1.16872758e-01
8.10067892e-01 -7.40743220e-01 1.02376509e+00 -1.46052194e+00
-1.34169388e+00 6.07568622e-01 -4.93049234e-01 -5.21236181e-01
3.18406403e-01 -7.65855849e-01 -1.38132587e-01 6.30631447e-01
3.40025336e-01 8.72941792e-01 9.82077956e-01 -1.12105286e+00
-8.28831553e-01 -8.19285288e-02 2.29367428e-02 1.75443634e-01
-1.09655112e-01 2.74680376e-01 -1.16537368e+00 -1.03653800e+00
-1.13768771e-01 -7.92865276e-01 -3.32040995e-01 2.10192464e-02
-3.09710741e-01 -3.36571008e-01 1.20171392e+00 -5.74364364e-01
1.21985972e+00 -2.10234666e+00 4.82932776e-01 -1.64305404e-01
4.13375586e-01 4.64422047e-01 -3.63983095e-01 -7.04826787e-02
7.93839917e-02 -1.47942588e-01 -3.34473252e-01 -5.24431229e-01
-2.74898559e-01 3.79496425e-01 -2.19010562e-01 4.81788933e-01
8.46698582e-01 1.23078454e+00 -1.23130417e+00 -6.31072938e-01
2.15635881e-01 5.07791281e-01 -8.70920599e-01 5.61784148e-01
-5.53499162e-01 4.90070581e-01 -6.74769223e-01 8.12298834e-01
4.28099722e-01 -6.17555380e-01 1.40589699e-01 -2.65385598e-01
1.75850481e-01 7.92887211e-02 -8.02390695e-01 2.13109207e+00
-1.40534610e-01 8.78014743e-01 2.04841606e-02 -1.17956996e+00
4.35013682e-01 5.51855601e-02 6.11834526e-01 -3.89529616e-01
9.07662138e-02 -8.75833109e-02 -1.32487699e-01 -8.29265058e-01
2.96111941e-01 3.34609181e-01 1.13951370e-01 2.82884449e-01
5.15953362e-01 5.88868499e-01 3.38909388e-01 2.70635933e-01
8.50707769e-01 6.61813080e-01 1.61641136e-01 -2.23031059e-01
6.37479842e-01 -3.35063040e-01 1.01192355e+00 4.98932809e-01
-5.38160622e-01 7.31905937e-01 4.18371052e-01 -6.18303001e-01
-6.89672709e-01 -9.29912031e-01 3.66650254e-01 1.61870718e+00
4.78138477e-01 -3.64901125e-01 -6.20917678e-01 -1.00265014e+00
-5.26238121e-02 1.93262741e-01 -8.40648353e-01 6.41691387e-02
-9.09843326e-01 -5.52238047e-01 8.26407224e-02 1.05493879e+00
4.55232054e-01 -1.23228645e+00 -8.05361629e-01 2.63961464e-01
-3.55592549e-01 -1.50106609e+00 -9.56339419e-01 9.70021263e-02
-9.26957130e-01 -1.11326110e+00 -8.95743489e-01 -9.21096265e-01
4.19877350e-01 6.98872328e-01 1.21418440e+00 1.45924598e-01
-2.39777818e-01 2.94885665e-01 -3.62512916e-01 4.73464504e-02
2.79044807e-01 -5.31696230e-02 -2.63423860e-01 4.45099860e-01
5.48088551e-01 -5.63042462e-01 -9.58148718e-01 2.78130919e-01
-8.08337510e-01 1.08380005e-01 7.77635276e-01 7.84463525e-01
7.86742032e-01 -5.25037527e-01 4.28764880e-01 -8.69953871e-01
-1.72793254e-01 -7.70932019e-01 -4.57434416e-01 2.62488365e-01
1.72778536e-02 8.75753015e-02 3.43473166e-01 -3.83816868e-01
-1.06729996e+00 6.97541907e-02 6.37044460e-02 -1.05773568e+00
-2.27343917e-01 2.46407688e-01 -3.61708522e-01 1.54683873e-01
-1.06363576e-02 1.88210696e-01 -3.57925117e-01 -4.86480683e-01
4.35678214e-01 7.07874149e-02 7.98343837e-01 -6.06333971e-01
4.72919494e-01 5.80970824e-01 -3.39377224e-01 -5.62677383e-01
-1.29295182e+00 -8.54224026e-01 -7.93631792e-01 -2.62866050e-01
1.41203475e+00 -1.18306255e+00 -4.38292742e-01 2.97782928e-01
-1.17119336e+00 -5.70366085e-01 -2.71945417e-01 4.16628093e-01
-5.90835035e-01 4.86720204e-01 -7.30951071e-01 -5.64035356e-01
-1.46086648e-01 -1.27933455e+00 1.47404492e+00 4.02977645e-01
-1.11105099e-01 -1.01182413e+00 -5.24920635e-02 4.56644118e-01
-1.79636236e-02 4.32750046e-01 3.87605697e-01 -6.17416799e-01
-1.11514819e+00 2.14527473e-01 -5.58892429e-01 1.92209199e-01
8.76999926e-04 1.89275160e-01 -9.26228702e-01 -2.61592060e-01
-2.46955395e-01 -2.67880648e-01 1.39508438e+00 6.30383372e-01
1.48108268e+00 -2.32863918e-01 -3.91612530e-01 9.79888141e-01
1.36421156e+00 -2.67754439e-02 4.85114783e-01 1.08404443e-01
1.14317894e+00 4.42558348e-01 6.90962970e-01 2.98398048e-01
4.59576488e-01 6.09871745e-01 5.13905704e-01 -2.65682846e-01
-2.56751478e-01 -9.87594575e-02 4.53110665e-01 8.52243423e-01
-2.18475640e-01 -8.32973793e-02 -6.05002046e-01 8.59706938e-01
-2.23596382e+00 -1.25491953e+00 1.39557183e-01 1.61907005e+00
6.53593957e-01 2.02442333e-01 3.18214506e-01 -3.53230596e-01
7.54079640e-01 4.96959925e-01 -7.89151788e-01 2.33288780e-01
-9.74534750e-02 5.64245135e-02 2.26461157e-01 3.73761177e-01
-1.45732331e+00 1.23122489e+00 5.98664331e+00 7.14624166e-01
-1.02960229e+00 2.10939124e-01 7.97696114e-01 -3.02881926e-01
6.97689457e-03 -2.06921935e-01 -6.16722107e-01 5.64444661e-01
6.19175494e-01 2.70199955e-01 1.05445296e-01 9.56970751e-01
1.76494867e-01 -8.89487639e-02 -1.16646862e+00 9.51510370e-01
2.65929043e-01 -1.53057492e+00 3.43872420e-02 -1.21403202e-01
1.17856526e+00 4.00291502e-01 -3.15692089e-02 2.14528024e-01
2.82848924e-01 -7.01560974e-01 8.65828395e-01 5.12403488e-01
3.80223304e-01 -6.34579360e-01 6.46003842e-01 -1.95975557e-01
-1.59667623e+00 -2.39340410e-01 -1.78795025e-01 8.03931355e-02
2.67123401e-01 3.26531023e-01 -3.18940789e-01 5.38903594e-01
8.54478836e-01 1.43981719e+00 -5.99940956e-01 1.00237596e+00
-1.82525098e-01 7.74073780e-01 -1.32458642e-01 2.60992438e-01
7.95544624e-01 -9.34614614e-02 4.65631753e-01 1.57186842e+00
-9.49499905e-02 1.33885354e-01 6.21471107e-01 9.47502494e-01
-1.50099248e-01 -2.72574693e-01 -6.41564354e-02 -6.84853345e-02
1.90180600e-01 1.31633961e+00 -1.07235789e+00 -6.85562134e-01
-7.69841373e-01 1.19046926e+00 3.73870581e-01 7.26498067e-01
-1.03501105e+00 -1.89360484e-01 1.04171824e+00 -2.17066795e-01
1.05711710e+00 -2.77957261e-01 4.86319922e-02 -1.42920113e+00
-2.94194147e-02 -7.50228107e-01 5.73486209e-01 -5.46387136e-01
-1.29956102e+00 5.04646778e-01 -1.15185760e-01 -1.09913433e+00
-1.73761711e-01 -5.18199086e-01 -7.81348467e-01 5.00803471e-01
-1.89066994e+00 -1.28011131e+00 -4.29611802e-01 7.36422360e-01
1.21868777e+00 -7.66471550e-02 1.20024838e-01 2.86459059e-01
-9.38258827e-01 4.08648551e-01 -2.71348029e-01 4.98015374e-01
5.98786831e-01 -1.22557139e+00 4.34867471e-01 1.26764953e+00
4.26554769e-01 4.95362133e-01 3.17726821e-01 -5.91736078e-01
-1.26053369e+00 -1.49827707e+00 4.12007421e-01 -5.03905177e-01
7.15677083e-01 -3.09940010e-01 -1.03436363e+00 7.24508286e-01
3.13241035e-01 6.34298682e-01 8.38193238e-01 -3.23269702e-02
-4.30339396e-01 2.53057405e-02 -5.43172657e-01 4.97832209e-01
1.28337300e+00 -7.21721113e-01 -6.38533413e-01 1.39352694e-01
1.10994232e+00 -1.24207452e-01 -5.35089016e-01 3.10957789e-01
4.37132895e-01 -9.18645024e-01 1.01337242e+00 -8.38990867e-01
7.07454443e-01 -4.69247311e-01 -1.98830530e-01 -6.69901431e-01
-4.52144533e-01 -8.70869756e-01 -7.47403204e-01 1.05524373e+00
7.67345726e-02 8.05886313e-02 8.85473549e-01 2.84588367e-01
-2.53074318e-01 -1.12785184e+00 -5.23151636e-01 -4.52039570e-01
-3.56455773e-01 -4.96125102e-01 3.54499996e-01 8.64862442e-01
-4.16487992e-01 3.72810334e-01 -4.71947193e-01 2.18646079e-01
6.25593781e-01 3.95021647e-01 6.34829581e-01 -9.48719859e-01
-5.00948668e-01 -5.56038082e-01 -4.59777772e-01 -1.67583883e+00
2.73303896e-01 -6.93910658e-01 2.11199522e-01 -1.30049920e+00
5.29144287e-01 3.77309062e-02 -7.62108684e-01 3.38826627e-01
-8.11792374e-01 4.66186732e-01 4.87453520e-01 3.21606040e-01
-1.52568221e+00 6.38660073e-01 1.25469017e+00 -5.55593260e-02
-1.52678818e-01 -2.40684107e-01 -6.43277109e-01 9.86667097e-01
2.22283721e-01 -2.94457108e-01 -3.65326971e-01 -7.77500212e-01
-3.40374410e-01 -1.44033045e-01 4.35413927e-01 -9.84509170e-01
2.43870452e-01 -2.88310170e-01 6.10169291e-01 -5.58478057e-01
2.01796696e-01 -7.22788036e-01 -4.02264148e-01 1.79517299e-01
-4.28019434e-01 2.04112381e-02 -2.07948429e-03 1.05796301e+00
-3.85890394e-01 1.98357895e-01 6.69267178e-01 -8.71888548e-02
-1.23001361e+00 7.60316908e-01 -2.95093507e-01 3.24806303e-01
1.11403716e+00 -2.33152226e-01 1.57499313e-01 -1.58384413e-01
-8.08502018e-01 4.72625345e-01 2.49158964e-01 5.38144112e-01
6.73415601e-01 -1.31911647e+00 -4.96953815e-01 2.03675911e-01
-5.82776926e-02 2.32555658e-01 3.70262355e-01 9.47492003e-01
-4.04453665e-01 2.35512719e-01 -2.04207882e-01 -1.01742697e+00
-1.11496663e+00 7.20965564e-01 1.97561562e-01 -2.34794378e-01
-6.07540905e-01 1.25848019e+00 7.34486163e-01 3.51357460e-01
3.36518317e-01 -6.59927905e-01 -1.78661510e-01 1.24780223e-01
6.56325877e-01 1.47314161e-01 -3.86423528e-01 -8.02204192e-01
-3.92995268e-01 6.61088645e-01 -2.63086349e-01 2.48720154e-01
1.50009656e+00 -2.98527718e-01 1.68075919e-01 2.90468186e-01
1.44869506e+00 -2.00489715e-01 -2.25061917e+00 -5.70217669e-01
1.26146689e-01 -6.92517996e-01 3.59634981e-02 -3.52129012e-01
-1.61761832e+00 9.35743928e-01 2.01495036e-01 -1.94016695e-01
1.34713233e+00 2.49993816e-01 1.06531847e+00 2.99301624e-01
-1.19607627e-01 -9.95467186e-01 6.67109489e-01 4.05973941e-01
5.77625930e-01 -1.44899845e+00 -9.13496614e-02 -2.99891740e-01
-9.36776459e-01 9.92323816e-01 8.62516880e-01 -5.66374660e-01
5.38196862e-01 1.23560593e-01 -7.00024664e-02 -1.71659514e-01
-9.44185615e-01 -5.26040077e-01 6.45647824e-01 4.66587245e-01
4.05863434e-01 -3.14185113e-01 1.58448532e-01 7.60863304e-01
6.48008943e-01 -1.70088679e-01 8.12320411e-02 8.62563252e-01
-4.85449612e-01 -7.15113759e-01 -7.62593970e-02 2.09949732e-01
-7.09130526e-01 6.65073171e-02 -2.28808299e-01 5.67896843e-01
3.90040845e-01 6.65456593e-01 3.38719368e-01 -2.17004225e-01
9.35508031e-03 -1.34281576e-01 3.53786051e-01 -5.72240174e-01
-4.40848619e-01 6.07290983e-01 -2.47566625e-01 -1.12990832e+00
-9.54963505e-01 -8.30292284e-01 -1.17512178e+00 2.07763970e-01
-1.95346236e-01 -9.88711119e-02 -1.71768554e-02 1.10304487e+00
4.87817883e-01 9.26114559e-01 4.72744018e-01 -1.41941404e+00
-4.51745689e-02 -6.51850283e-01 -1.45726457e-01 6.55729294e-01
7.04829812e-01 -7.18990803e-01 9.77296382e-02 5.63061535e-01] | [9.246973037719727, -0.16227711737155914] |
edb3a825-67a6-4b2a-8052-715a1321405a | learning-hard-alignments-with-variational | 1705.05524 | null | http://arxiv.org/abs/1705.05524v2 | http://arxiv.org/pdf/1705.05524v2.pdf | Learning Hard Alignments with Variational Inference | There has recently been significant interest in hard attention models for
tasks such as object recognition, visual captioning and speech recognition.
Hard attention can offer benefits over soft attention such as decreased
computational cost, but training hard attention models can be difficult because
of the discrete latent variables they introduce. Previous work used REINFORCE
and Q-learning to approach these issues, but those methods can provide
high-variance gradient estimates and be slow to train. In this paper, we tackle
the problem of learning hard attention for a sequential task using variational
inference methods, specifically the recently introduced VIMCO and NVIL.
Furthermore, we propose a novel baseline that adapts VIMCO to this setting. We
demonstrate our method on a phoneme recognition task in clean and noisy
environments and show that our method outperforms REINFORCE, with the
difference being greater for a more complicated task. | ['Colin Raffel', 'George Tucker', 'Chung-Cheng Chiu', 'Navdeep Jaitly', 'Kevin Swersky', 'Dieterich Lawson'] | 2017-05-16 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 1.73498183e-01 2.38941148e-01 -2.44990945e-01 -2.72432119e-01
-1.22254443e+00 -2.30787724e-01 6.82416916e-01 -2.44837552e-01
-5.90518117e-01 9.70909774e-01 3.64290565e-01 -3.04434747e-01
-1.44540705e-02 -1.35125950e-01 -8.55096161e-01 -7.85546720e-01
2.00640917e-01 6.01505220e-01 -5.59296422e-02 2.09782496e-01
2.25666136e-01 4.68829870e-02 -1.38002157e+00 -2.34738335e-01
8.44604135e-01 7.97040582e-01 5.51861227e-01 8.81238163e-01
-1.33740632e-02 8.77124369e-01 -5.95765352e-01 -2.52079874e-01
-5.14545254e-02 -3.34852368e-01 -8.39395165e-01 -1.27538979e-01
5.56847930e-01 -1.24659277e-01 -6.69882521e-02 9.38117266e-01
6.52187049e-01 6.31325066e-01 7.91636229e-01 -1.27859068e+00
-9.57414627e-01 3.04011852e-01 -5.42411804e-01 3.65033865e-01
4.43785042e-02 1.83386073e-01 1.33742535e+00 -9.47994649e-01
2.93151349e-01 1.57046688e+00 4.66446817e-01 8.57283890e-01
-1.51006949e+00 -3.07358205e-01 4.71517086e-01 4.76814091e-01
-1.11435831e+00 -5.59193254e-01 4.56540316e-01 -4.02435750e-01
1.24258614e+00 2.70230293e-01 3.84391770e-02 1.35322595e+00
6.74265176e-02 1.19169557e+00 1.25562096e+00 -4.50517893e-01
4.34745193e-01 1.53683499e-01 1.29471883e-01 2.23368183e-01
-2.39651218e-01 4.67372574e-02 -3.11462224e-01 -1.60685569e-01
5.53423166e-01 -8.55368972e-02 -2.65023589e-01 -1.57750919e-01
-7.06044972e-01 1.11101830e+00 3.11609596e-01 -5.85917570e-03
-5.50463974e-01 4.64835972e-01 6.56278655e-02 1.48374483e-01
7.88246632e-01 4.53160942e-01 -4.32884067e-01 -5.39011180e-01
-8.33567262e-01 1.23510942e-01 7.92340815e-01 7.06164658e-01
5.44378638e-01 1.82772070e-01 -5.45706689e-01 1.22992086e+00
5.91296315e-01 5.96999049e-01 4.11026448e-01 -1.11141157e+00
4.05599535e-01 -5.14334023e-01 2.42535755e-01 -4.30388778e-01
-8.26267004e-02 -1.90106779e-01 -6.34249985e-01 2.75447458e-01
2.82664388e-01 -2.29368821e-01 -1.34893739e+00 2.00607634e+00
1.51463300e-01 2.17423335e-01 -1.74844071e-01 1.01366186e+00
5.31685650e-01 9.60111439e-01 3.77289534e-01 -2.89114326e-01
1.03556538e+00 -1.25019753e+00 -1.15483725e+00 -6.06221259e-01
7.27584586e-02 -6.29984438e-01 1.22061872e+00 4.45082635e-01
-1.34393108e+00 -5.64615309e-01 -6.39504015e-01 -3.21593553e-01
-8.41203481e-02 -3.08223128e-01 5.58991194e-01 5.02018750e-01
-1.13502598e+00 4.47315454e-01 -9.66369629e-01 -2.85594553e-01
5.61172485e-01 4.38623488e-01 2.96335787e-01 4.75367717e-02
-9.62122738e-01 1.03685164e+00 -5.33388183e-02 6.42034560e-02
-1.20833731e+00 -4.64728177e-01 -9.27960038e-01 4.85514283e-01
6.20565891e-01 -5.17429888e-01 1.55747736e+00 -1.07457352e+00
-1.94515073e+00 3.67266297e-01 -4.89142030e-01 -5.56424737e-01
4.90962476e-01 -7.27028906e-01 5.38697429e-02 6.90251170e-03
-2.08071079e-02 9.06858802e-01 1.30913699e+00 -9.35082555e-01
-3.52633983e-01 3.17829661e-02 6.60094991e-02 2.44425222e-01
-2.35329568e-01 1.24013118e-01 -4.57130283e-01 -7.40147829e-01
-5.61286747e-01 -8.84366393e-01 -3.33537906e-01 -2.21161202e-01
-2.23504856e-01 -6.45166397e-01 7.82694757e-01 -5.85644007e-01
8.67490053e-01 -2.23904371e+00 4.36580718e-01 -1.94354832e-01
1.48425356e-01 4.60063487e-01 -2.65502691e-01 9.54551920e-02
1.35624990e-01 2.70357430e-01 -2.73877889e-01 -9.95528877e-01
2.71426886e-01 6.42090857e-01 -4.08006370e-01 9.71124992e-02
5.28662026e-01 1.13180566e+00 -9.23940063e-01 -5.66546381e-01
1.23850130e-01 6.62181020e-01 -8.33780646e-01 4.60955590e-01
-4.44302648e-01 3.66466135e-01 -1.26363218e-01 3.73867363e-01
4.98689711e-01 -4.94362772e-01 2.13456526e-02 3.51568907e-01
6.74463809e-02 2.72872537e-01 -9.15258765e-01 1.23287952e+00
-6.38037860e-01 9.54701304e-01 4.14213121e-01 -1.16396117e+00
4.44753319e-01 4.06726986e-01 1.21682391e-01 -5.65402508e-01
9.16608423e-03 -1.59022182e-01 -1.29594862e-01 -6.81923568e-01
5.45015156e-01 -3.42635572e-01 2.99534619e-01 3.24993908e-01
1.24427617e-01 -1.72625870e-01 -1.00313313e-01 1.95138812e-01
7.31823623e-01 2.07062528e-01 4.03422676e-02 -2.36605123e-01
5.37981093e-02 -5.47692239e-01 5.62871158e-01 1.29695618e+00
-3.48466992e-01 8.08718145e-01 4.74037617e-01 2.05780059e-01
-9.02684927e-01 -1.11635995e+00 -1.91458046e-01 1.45912552e+00
-1.90192774e-01 -1.69387028e-01 -6.71581864e-01 -6.12764537e-01
5.65503491e-03 7.80611098e-01 -5.85796773e-01 1.24771949e-02
-4.61057425e-01 -6.84824347e-01 1.12908423e-01 8.11084509e-01
-2.54647136e-02 -1.30945599e+00 -4.12670821e-01 2.97111273e-01
-2.75673747e-01 -1.03307319e+00 -6.51597619e-01 3.79484773e-01
-6.12709522e-01 -4.42018479e-01 -1.15309429e+00 -6.70454741e-01
4.19728279e-01 -5.82884289e-02 1.08854151e+00 -1.57146677e-01
-2.03784987e-01 5.34854233e-01 -1.61156461e-01 -6.95708036e-01
-2.23820478e-01 1.37614265e-01 1.02429882e-01 1.41518831e-01
2.40276560e-01 -1.64622009e-01 -4.03399736e-01 7.20425695e-02
-8.21221828e-01 -2.09231749e-01 4.58857447e-01 1.36924601e+00
5.33085525e-01 -7.04807222e-01 8.10367346e-01 -7.91115463e-01
8.74367952e-01 -6.66475475e-01 -6.83246970e-01 1.11373164e-01
-7.31799960e-01 3.40586573e-01 3.27853292e-01 -7.03821421e-01
-1.17116129e+00 -2.21570611e-01 -3.87996733e-01 -6.44763708e-01
-2.09302288e-02 7.16935039e-01 2.72947848e-02 1.57921821e-01
3.17877799e-01 -1.89178333e-01 1.66361146e-02 -6.45434141e-01
3.35533381e-01 7.44261324e-01 2.78844416e-01 -5.00270009e-01
4.46114570e-01 2.33284399e-01 -2.26026565e-01 -1.06131375e+00
-1.02606475e+00 -3.33190918e-01 -1.61341608e-01 1.08711950e-01
1.11032057e+00 -7.29666114e-01 -7.46466219e-01 2.17903346e-01
-1.02115428e+00 -8.31154525e-01 -4.08020735e-01 6.72324419e-01
-7.48887479e-01 5.40403545e-01 -7.61580169e-01 -1.17671013e+00
-1.75145939e-01 -1.25525141e+00 1.10687423e+00 1.67509884e-01
-3.00382912e-01 -1.23852742e+00 1.47600383e-01 2.96272933e-01
7.00548470e-01 -2.03261405e-01 6.70246601e-01 -2.36047953e-01
-7.06187487e-01 1.99274912e-01 -1.90024629e-01 6.27639234e-01
-1.13016061e-01 -6.95855021e-02 -1.36855125e+00 -4.29685891e-01
-1.91816129e-02 -8.51105809e-01 1.29360855e+00 9.72593129e-01
1.46300924e+00 -3.11584592e-01 -2.08937392e-01 5.50210774e-01
1.15499032e+00 7.04755932e-02 6.82161093e-01 4.56248224e-02
6.63212478e-01 3.78865331e-01 4.83795673e-01 2.32659146e-01
2.81583369e-01 8.21898580e-01 4.48848575e-01 -2.78940260e-01
-7.15235844e-02 1.56102097e-02 4.38654989e-01 7.36585736e-01
-1.66013315e-01 -5.50423622e-01 -7.40066171e-01 7.94484675e-01
-2.07656980e+00 -1.07191789e+00 2.02774838e-01 1.96886814e+00
8.09128821e-01 2.08754107e-01 2.21435875e-02 -2.05328196e-01
3.81613523e-01 1.43380240e-01 -4.39810038e-01 -7.61201560e-01
2.93373410e-02 4.95503068e-01 1.90778211e-01 1.02666533e+00
-9.81307447e-01 9.07321334e-01 7.37010622e+00 7.49207199e-01
-1.03070223e+00 4.96913493e-01 6.44658506e-01 -4.98841554e-01
-1.83713704e-01 -4.96353246e-02 -7.11856186e-01 3.88604879e-01
1.04858649e+00 2.29864001e-01 5.08574605e-01 6.99821115e-01
1.00199610e-01 -2.78148592e-01 -1.22229385e+00 1.11529088e+00
2.13955715e-01 -8.87474179e-01 -4.46648866e-01 1.75408661e-01
8.79781663e-01 4.51921225e-01 4.24353391e-01 6.51141822e-01
4.65440005e-01 -1.26248491e+00 5.09357333e-01 5.56886315e-01
5.82144499e-01 -5.55115581e-01 6.79114282e-01 3.27882290e-01
-6.05257213e-01 -1.22581124e-01 -4.17960852e-01 -2.69631207e-01
3.60186189e-01 3.60264599e-01 -7.98804402e-01 -5.16576543e-02
7.49278247e-01 4.65163648e-01 -2.18673527e-01 1.10530722e+00
-4.72187907e-01 1.14105082e+00 -2.45243743e-01 -3.66524905e-01
4.76533592e-01 -1.82051398e-02 6.35300577e-01 1.24446893e+00
2.15061098e-01 7.19197989e-02 2.01976508e-01 1.05438161e+00
-6.80939183e-02 -2.13153720e-01 -3.47578019e-01 -1.09978847e-01
1.27505779e-01 1.01575005e+00 -3.44031483e-01 -6.16978824e-01
-4.45633203e-01 9.58748579e-01 6.45399511e-01 7.41962910e-01
-1.01282990e+00 -4.43876952e-01 7.93016851e-01 -2.77192682e-01
8.15460742e-01 -2.42543653e-01 -4.31688353e-02 -1.25602365e+00
-1.38740420e-01 -7.21363008e-01 2.25107267e-01 -8.76668036e-01
-1.29531252e+00 4.30888087e-01 1.12214193e-01 -4.98130292e-01
-6.05255127e-01 -5.70859075e-01 -5.23513436e-01 1.14426613e+00
-1.75164056e+00 -5.87697029e-01 -7.23999292e-02 3.40549916e-01
9.43528473e-01 2.33398974e-01 6.53983712e-01 2.74536192e-01
-6.17388308e-01 7.62773573e-01 4.88281429e-01 -1.58573061e-01
8.00406754e-01 -1.64019382e+00 4.28264529e-01 6.71271801e-01
2.97116756e-01 4.96043384e-01 8.49578559e-01 -4.16420609e-01
-1.08531499e+00 -8.06076765e-01 8.70931625e-01 -7.52758503e-01
4.56220657e-01 -5.99677205e-01 -1.09781325e+00 8.23287070e-01
6.02740347e-01 -2.34062150e-02 5.07960498e-01 5.08790910e-01
-1.63598180e-01 2.83570141e-01 -8.45016301e-01 4.34338480e-01
6.86477661e-01 -7.25498617e-01 -7.01436937e-01 3.12692732e-01
7.54218936e-01 -3.20304096e-01 -5.60785890e-01 -4.90602324e-05
3.25871706e-01 -3.85841638e-01 8.93692732e-01 -7.27334142e-01
3.42838585e-01 1.44740716e-01 -3.23090330e-02 -1.63566566e+00
-5.65354168e-01 -8.18907976e-01 -4.99700516e-01 1.20615220e+00
4.61670816e-01 -6.87557936e-01 6.53699338e-01 7.17665017e-01
-1.90637067e-01 -5.79422891e-01 -9.68941271e-01 -9.32075977e-01
2.73518950e-01 -5.24004340e-01 1.97634757e-01 7.22545385e-01
-5.27330972e-02 6.52195632e-01 -8.04932892e-01 -1.22350801e-04
6.93217635e-01 -1.84125498e-01 5.98542631e-01 -1.07875431e+00
-7.63848960e-01 -3.61157328e-01 -4.44698185e-02 -1.47096634e+00
2.58118302e-01 -6.63194954e-01 5.75823069e-01 -1.48702669e+00
3.12159806e-01 -2.12333649e-01 -4.73943293e-01 5.06885409e-01
-6.36314809e-01 2.20316142e-01 2.79841781e-01 6.25627115e-02
-7.71804810e-01 8.86521995e-01 1.06586587e+00 -3.39792997e-01
-1.47502676e-01 -9.13953409e-02 -6.20181918e-01 4.97930288e-01
5.94433308e-01 -5.79060078e-01 -3.73471171e-01 -6.98135555e-01
8.29894394e-02 -2.88709644e-02 4.97175455e-01 -6.41232848e-01
-1.06664486e-02 -2.32912317e-01 1.82722703e-01 -3.86494547e-01
7.24300981e-01 -4.35600102e-01 -4.75001514e-01 2.64131129e-01
-5.52368879e-01 -1.65792465e-01 3.63119900e-01 7.37510383e-01
-2.24328026e-01 -4.43956435e-01 8.40419829e-01 1.03074357e-01
-5.08688450e-01 2.79201269e-01 -6.47015631e-01 5.34558654e-01
7.38489270e-01 1.14123031e-01 -1.02822632e-01 -8.18986595e-01
-1.03414571e+00 5.09418190e-01 -6.20435253e-02 5.34788132e-01
6.62814260e-01 -1.27983296e+00 -7.79439211e-01 7.83988535e-02
-1.61058515e-01 -3.90073434e-02 6.39852434e-02 9.86681402e-01
9.02008489e-02 4.65175450e-01 1.13237001e-01 -8.20665777e-01
-1.29154217e+00 7.23802090e-01 1.10441938e-01 -2.01940089e-01
-2.78357506e-01 1.07529950e+00 5.28837144e-01 -1.17971361e-01
6.67727172e-01 -4.82593477e-01 -1.63680986e-01 3.47804502e-02
5.80788672e-01 3.04300368e-01 -3.03556114e-01 -3.46950114e-01
-2.12320849e-01 3.47442359e-01 -3.42602581e-01 -4.13816690e-01
1.39674294e+00 -2.37809464e-01 2.37908304e-01 5.01025736e-01
1.09994411e+00 -2.72189826e-01 -1.82016027e+00 -2.61927694e-01
-3.01212422e-04 -5.47134876e-01 3.09743643e-01 -6.97525859e-01
-7.24735379e-01 1.42575812e+00 7.82474041e-01 4.61670011e-01
7.53212512e-01 2.30378583e-01 5.72801113e-01 3.44163120e-01
-7.69219920e-02 -1.38656747e+00 3.23114425e-01 8.20845127e-01
7.82056749e-01 -1.58609617e+00 -3.17433566e-01 1.47459537e-01
-6.66900873e-01 5.45528710e-01 5.56353390e-01 3.80848311e-02
6.08246744e-01 1.02393970e-01 4.94481921e-02 2.03253590e-02
-9.73295927e-01 -4.01271641e-01 4.30666298e-01 7.75978923e-01
4.36658263e-01 5.28437132e-03 1.70549087e-03 1.46854565e-01
2.70335644e-01 5.18648028e-02 3.50545019e-01 8.58501196e-01
-3.93437237e-01 -9.32117939e-01 -4.51575518e-01 4.14225131e-01
-9.06181216e-01 -4.00103122e-01 -1.25376761e-01 4.01729256e-01
-2.68957645e-01 9.86519516e-01 2.12106869e-01 2.15723217e-01
-4.67001759e-02 2.16959298e-01 4.96231079e-01 -7.01425612e-01
-2.47639284e-01 3.72818768e-01 3.62570919e-02 -5.14327645e-01
-4.74038869e-01 -8.51153851e-01 -8.98003280e-01 8.94526169e-02
-4.96422440e-01 3.06250989e-01 6.14014864e-01 1.07943773e+00
3.37109894e-01 6.36430562e-01 4.04609412e-01 -1.07725012e+00
-1.07108712e+00 -1.21823752e+00 -2.44761541e-01 4.93755400e-01
8.24129343e-01 -6.86518490e-01 -5.01840591e-01 -3.46782245e-02] | [10.09252643585205, 1.9616557359695435] |
b30829df-fe45-4ced-8e0a-c9ec9665d935 | faheem-at-nadi-shared-task-identifying-the | null | null | https://aclanthology.org/2020.wanlp-1.29 | https://aclanthology.org/2020.wanlp-1.29.pdf | Faheem at NADI shared task: Identifying the dialect of Arabic tweet | This paper describes Faheem (adj. of understand), our submission to NADI (Nuanced Arabic Dialect Identification) shared task. With so many Arabic dialects being under-studied due to the scarcity of the resources, the objective is to identify the Arabic dialect used in the tweet, country wise. We propose a machine learning approach where we utilize word-level n-gram (n = 1 to 3) and tf-idf features and feed them to six different classifiers. We train the system using a data set of 21,000 tweets—provided by the organizers—covering twenty-one Arab countries. Our top performing classifiers are: Logistic Regression, Support Vector Machines, and Multinomial Na ̈ıve Bayes. | ['Aqil Azmi', 'Nouf AlShenaifi'] | null | null | null | null | coling-wanlp-2020-12 | ['dialect-identification'] | ['natural-language-processing'] | [-5.62233925e-01 -4.41417068e-01 -2.98720568e-01 -4.71523970e-01
-5.87094426e-01 -9.10893500e-01 9.73619401e-01 3.11635107e-01
-7.36237764e-01 7.40970910e-01 4.39501464e-01 -6.42541051e-01
-9.88106877e-02 -8.98468435e-01 2.98543517e-02 -5.80752969e-01
-3.95726323e-01 9.37910259e-01 -2.55738884e-01 -1.13204634e+00
6.28338754e-01 6.47060990e-01 -1.25362754e+00 1.79777890e-01
1.09608173e+00 6.37774467e-01 -3.38147551e-01 5.91280818e-01
1.68062486e-02 6.25057697e-01 -4.69798207e-01 -6.04993820e-01
6.21422380e-02 -3.75282258e-01 -1.10768425e+00 -2.35779509e-01
1.36515379e-01 -4.15996939e-01 -3.85733768e-02 7.81815052e-01
2.76342392e-01 1.11645430e-01 1.24699068e+00 -8.82749438e-01
-5.98258615e-01 8.21636319e-01 -6.86819017e-01 2.79698104e-01
5.22849560e-01 -3.82447958e-01 1.18374920e+00 -1.54709995e+00
4.71622795e-01 1.56848574e+00 6.70684874e-01 6.19915463e-02
-7.81022429e-01 -6.42146289e-01 -2.50558525e-01 2.61084139e-01
-1.49787629e+00 -3.91548932e-01 2.90965408e-01 -6.17525339e-01
7.32608438e-01 2.38879368e-01 3.93167675e-01 7.41238952e-01
8.18581358e-02 3.92829508e-01 1.45245802e+00 -8.56892347e-01
-1.50568470e-01 3.68789911e-01 3.17425251e-01 7.22017109e-01
2.28141043e-02 -4.42161500e-01 -4.63719606e-01 -5.27966201e-01
1.87997073e-01 -3.41415495e-01 1.31353199e-01 5.92084289e-01
-1.07543170e+00 1.38729274e+00 1.16143182e-01 5.98472774e-01
-4.12163883e-01 -6.75491691e-01 3.94278347e-01 7.87898779e-01
6.65800512e-01 2.22145543e-01 -5.27442515e-01 -1.97196737e-01
-7.22285330e-01 4.12212133e-01 9.60410833e-01 4.10436958e-01
6.88051462e-01 7.36157671e-02 4.94854659e-01 1.40147471e+00
6.34425104e-01 8.34869623e-01 6.56910360e-01 -1.05355240e-01
3.05398554e-01 4.18026894e-01 2.48679638e-01 -1.27934718e+00
-5.41159511e-01 2.22044781e-01 -7.40015566e-01 -2.47838110e-01
8.78851295e-01 -5.98965168e-01 -7.63792336e-01 1.06229508e+00
7.58040696e-02 -6.76887631e-01 -1.29253954e-01 8.93454969e-01
5.24470448e-01 9.59051251e-01 -1.23391151e-02 -1.25264987e-01
1.52774537e+00 -5.69360673e-01 -5.84830821e-01 2.86674649e-02
7.42325544e-01 -1.26027656e+00 7.60763526e-01 7.22253799e-01
-7.49043822e-01 -3.46542746e-02 -9.11360383e-01 3.68331403e-01
-9.74779069e-01 3.75245899e-01 6.61563933e-01 1.03770161e+00
-9.25229192e-01 3.39688025e-02 -2.12861121e-01 -5.83716869e-01
-9.08850059e-02 4.42107648e-01 -4.77184772e-01 -2.62873303e-02
-1.59744048e+00 1.27049708e+00 1.34822235e-01 1.64636612e-01
-4.19196010e-01 -7.15314224e-02 -8.43567550e-01 -6.80316985e-01
-1.73454389e-01 2.64895290e-01 7.14153171e-01 -1.05914032e+00
-1.35240006e+00 1.30054963e+00 -1.95984617e-02 -6.88574985e-02
1.48916945e-01 -1.69632480e-01 -8.93413723e-01 -1.26115382e-01
1.66798174e-01 1.72318950e-01 6.93307936e-01 -9.50451195e-01
-9.71148252e-01 -5.76470315e-01 2.80777067e-02 1.33969679e-01
-4.79382843e-01 9.12373126e-01 1.67151958e-01 -9.47950065e-01
3.77742767e-01 -1.09362829e+00 -8.12681466e-02 -1.05022824e+00
-1.81553647e-01 -4.35479164e-01 6.08290136e-01 -1.26408827e+00
1.54001760e+00 -1.79169226e+00 6.56047612e-02 7.49555051e-01
-5.28605431e-02 7.36096725e-02 6.78990185e-02 7.39415169e-01
-3.58548872e-02 -8.77975821e-02 -1.32317215e-01 2.86417097e-01
-1.00408532e-01 3.17366756e-02 -3.26244444e-01 5.76585352e-01
1.42377079e-01 2.90951222e-01 -8.08087707e-01 -3.81260246e-01
-1.10904671e-01 2.23470166e-01 -8.76774564e-02 -3.78600657e-02
1.73938885e-01 1.42404988e-01 -7.43589103e-02 1.04240227e+00
8.20302725e-01 3.40619385e-01 3.58077645e-01 -1.56628460e-01
-3.89193445e-01 3.81768346e-01 -9.62955713e-01 7.46471703e-01
-3.82897139e-01 5.01471698e-01 1.31439209e-01 -8.67336273e-01
1.19118798e+00 1.52237862e-01 3.21335882e-01 -3.75959396e-01
2.64915347e-01 7.85180092e-01 1.03405021e-01 -3.12813103e-01
6.58397019e-01 -9.02073309e-02 -6.43095315e-01 7.70173728e-01
1.24559589e-01 2.22865626e-01 4.72400427e-01 4.50717211e-02
2.43135691e-01 -2.29476467e-01 6.33253157e-01 -8.01697373e-01
9.10570383e-01 2.67982602e-01 1.08031116e-01 5.08633316e-01
-8.58666003e-02 1.71135962e-01 3.18285823e-01 -7.44249761e-01
-8.13553929e-01 -8.17261279e-01 -4.82312143e-01 1.91850209e+00
-5.20747066e-01 -2.81685114e-01 -8.02704155e-01 -7.47512400e-01
-5.59212267e-02 5.06436408e-01 -4.22794968e-01 4.45128798e-01
-1.07903624e+00 -1.45475996e+00 8.16543698e-01 -8.64568949e-02
3.21924448e-01 -7.97261119e-01 -1.78141316e-04 2.64802247e-01
-4.12519693e-01 -6.05581403e-01 -4.08404410e-01 1.17485195e-01
-3.55769366e-01 -8.73764753e-01 -8.40234637e-01 -1.13309145e+00
4.03224856e-01 6.33273786e-03 1.14754629e+00 -1.02618545e-01
2.71678828e-02 -1.88106731e-01 -6.29282355e-01 -5.59870422e-01
-5.16935766e-01 3.54261279e-01 3.77148658e-01 1.75664306e-01
8.65326762e-01 -1.55871183e-01 -7.56170377e-02 4.90806937e-01
-5.67382812e-01 -5.93677163e-01 2.43578345e-01 7.03418553e-01
-2.06639722e-01 -1.53441682e-01 8.59514475e-01 -1.07436383e+00
7.57529497e-01 -8.89684618e-01 -3.28762591e-01 1.90387025e-01
-5.66504598e-01 -3.96089375e-01 3.88166219e-01 -1.38473079e-01
-9.03923392e-01 -2.79765517e-01 -4.35368508e-01 7.61478424e-01
-8.51451904e-02 1.08299482e+00 -1.94616057e-02 -1.63025945e-01
7.51215458e-01 1.62615687e-01 3.84740718e-02 -3.23041230e-01
4.10295546e-01 1.39052987e+00 9.11911353e-02 -4.88854617e-01
4.53208059e-01 1.92888200e-01 -3.77782851e-01 -1.05540895e+00
-5.72093606e-01 -4.57919091e-01 -9.84178543e-01 -4.33295131e-01
5.67151308e-01 -9.68495965e-01 -4.30464387e-01 9.91128087e-01
-7.89374471e-01 1.24269471e-01 8.60401988e-01 5.49827099e-01
-2.68733412e-01 6.01186678e-02 -9.19174314e-01 -8.71837795e-01
-3.03577840e-01 -9.71110582e-01 4.10583764e-01 -4.79536355e-02
-5.36911845e-01 -1.15300012e+00 2.07876399e-01 1.88979730e-01
5.14226675e-01 1.57575443e-01 1.31427062e+00 -9.79731858e-01
3.31197232e-01 -1.30338132e-01 -2.35174522e-01 -4.36270535e-02
6.02768123e-01 2.04195485e-01 -8.67165864e-01 -1.50359303e-01
-1.55899376e-01 -4.95178640e-01 6.86478674e-01 2.73707360e-01
5.64299822e-01 -6.20725036e-01 9.61971134e-02 -3.92040284e-03
1.21890175e+00 1.18741862e-01 2.36863300e-01 6.68427348e-01
5.13482273e-01 9.57571447e-01 9.43707526e-01 6.58530653e-01
9.25856829e-01 5.21383643e-01 -5.25557995e-02 2.19150364e-01
5.24676681e-01 2.52068818e-01 6.78472340e-01 1.22510660e+00
-4.06972468e-01 -3.19881067e-02 -1.52528489e+00 7.38399982e-01
-1.56927252e+00 -6.35282159e-01 -3.47144842e-01 2.12695169e+00
1.07641041e+00 -2.44862109e-01 8.18140209e-01 4.52848583e-01
4.81802940e-01 1.90899484e-02 3.95958632e-01 -8.50348294e-01
-5.03428638e-01 2.02640072e-01 6.84241295e-01 7.69303620e-01
-1.30052829e+00 1.09250319e+00 6.58264303e+00 8.61652315e-01
-1.22640204e+00 -1.37940720e-01 8.40762794e-01 4.06101793e-01
-9.18129012e-02 -3.59327316e-01 -9.09921825e-01 3.57827544e-01
1.43748927e+00 -1.91922426e-01 5.42522788e-01 3.34113985e-01
7.22694248e-02 -3.28088969e-01 -5.68762600e-01 7.61831760e-01
4.83121753e-01 -9.16049957e-01 2.37090230e-01 -4.82921302e-02
5.31220555e-01 1.92608297e-01 2.13899672e-01 2.45608598e-01
6.42084897e-01 -1.28936267e+00 7.43684888e-01 2.87568480e-01
7.69197702e-01 -1.13286018e+00 9.67448652e-01 3.16423833e-01
-9.08960700e-01 -2.47153491e-02 -1.26937002e-01 -1.74450666e-01
-1.28211984e-02 3.98990333e-01 -9.02398109e-01 4.71470416e-01
7.10048914e-01 6.38846397e-01 -5.26438951e-01 3.65067691e-01
2.08999380e-01 1.08712006e+00 -2.07370535e-01 -3.74968678e-01
7.28439927e-01 -5.88040590e-01 3.64154398e-01 1.66632366e+00
3.03652704e-01 4.13421281e-02 9.11147818e-02 -1.72249332e-01
2.70368397e-01 8.03736687e-01 -4.61243182e-01 2.96397563e-02
4.87036437e-01 1.07235253e+00 -7.82769263e-01 -1.57794766e-02
-5.45072496e-01 7.98184335e-01 1.35318756e-01 1.79110229e-01
-3.88552487e-01 -6.85692370e-01 4.90939021e-01 -8.39859024e-02
-1.49467349e-01 -4.04756874e-01 -3.85362983e-01 -9.36147451e-01
-5.17244577e-01 -1.39261544e+00 8.77604306e-01 -1.02214351e-01
-1.58413541e+00 8.01214278e-01 -9.88249388e-03 -8.81553113e-01
-6.67668641e-01 -1.10416687e+00 -1.32365525e-01 1.24604285e+00
-1.31972218e+00 -1.34957433e+00 2.47678846e-01 5.52306950e-01
3.88151646e-01 -8.07781637e-01 1.25248766e+00 6.32825673e-01
-1.99947745e-01 6.18228793e-01 2.22392932e-01 5.79250216e-01
1.15182984e+00 -1.26747096e+00 2.03232184e-01 1.84799984e-01
-1.19618624e-02 8.40449870e-01 5.82426190e-01 -5.23394227e-01
-1.15447998e+00 -8.11505914e-01 1.91550517e+00 -5.32962441e-01
1.09037793e+00 -2.76411384e-01 -5.09226918e-01 5.64174712e-01
3.02597284e-01 -6.44504070e-01 1.06939292e+00 3.90518427e-01
-4.23868746e-01 1.30352471e-02 -1.31934476e+00 4.15282458e-01
1.42300770e-01 -5.74544013e-01 -5.08563519e-01 4.24043357e-01
-7.36107454e-02 -8.21122900e-03 -9.77574825e-01 -1.11464106e-01
8.08267474e-01 -6.61981761e-01 7.90476441e-01 -8.72338176e-01
2.11823851e-01 8.03645477e-02 -5.61491728e-01 -1.59379530e+00
-9.57280248e-02 -5.27198970e-01 5.11165857e-01 1.29307067e+00
6.98151648e-01 -9.18639541e-01 3.22012872e-01 6.17240109e-02
4.41862106e-01 -4.07643944e-01 -6.19964898e-01 -4.47538376e-01
6.50843441e-01 -1.19137846e-01 6.77803993e-01 1.57354832e+00
3.05803120e-01 4.51048732e-01 -6.03988469e-01 -1.20554276e-01
2.93655336e-01 -1.42289653e-01 4.03729707e-01 -1.22839880e+00
2.68647641e-01 -5.68416357e-01 -1.68868199e-01 -5.88293850e-01
8.34981725e-02 -8.04935157e-01 -2.18977511e-01 -8.76587927e-01
-1.78072020e-01 -7.91189551e-01 1.13712363e-01 4.77324426e-01
-9.60673485e-03 6.31375134e-01 7.99051374e-02 3.83100510e-01
2.04519909e-02 5.99296018e-02 5.52203655e-01 -1.78876340e-01
-1.64395228e-01 2.37051144e-01 -5.24605632e-01 8.00594807e-01
8.67600381e-01 -2.24431217e-01 2.32505754e-01 -3.46992165e-01
5.07604659e-01 -1.29765466e-01 -2.96833605e-01 -3.63340020e-01
-1.54588938e-01 -3.81495863e-01 2.08247945e-01 -7.38748848e-01
3.42871733e-02 -4.84236747e-01 -3.42143685e-01 3.55361521e-01
-1.27753854e-01 6.84281766e-01 6.13445528e-02 -2.96055555e-01
-5.10323644e-01 -3.67463380e-01 7.40850866e-01 -4.08834219e-02
-5.32627940e-01 2.01612711e-01 -1.10453725e+00 -1.06700391e-01
5.62769532e-01 2.43107434e-02 -4.71826047e-02 -4.57158715e-01
-5.65869033e-01 3.13446783e-02 4.49166894e-02 3.94594222e-01
3.30977172e-01 -1.33376932e+00 -1.45845163e+00 3.35201740e-01
3.27288151e-01 -9.27879155e-01 -3.31895590e-01 6.99645996e-01
-1.01745677e+00 6.02836072e-01 -4.96191472e-01 -2.34983400e-01
-1.17432082e+00 -3.09008718e-01 1.64465904e-01 -9.92619097e-02
1.38864130e-01 8.19800556e-01 -4.40258801e-01 -1.05449080e+00
-1.39163658e-01 4.31404203e-01 -8.98061156e-01 9.20098305e-01
7.27014124e-01 6.92389607e-01 3.80149454e-01 -1.46447992e+00
-4.82648134e-01 3.50785613e-01 -3.86964381e-01 -5.79150438e-01
1.31571484e+00 -3.18812639e-01 -9.06715035e-01 7.23933458e-01
9.16865885e-01 4.51679558e-01 -9.04679447e-02 -2.31000409e-01
4.99265850e-01 -4.98404384e-01 -1.87292397e-01 -8.30425262e-01
-3.78322035e-01 5.41263938e-01 4.24284786e-01 4.21678901e-01
8.81021082e-01 -3.35874528e-01 5.54065526e-01 3.57745796e-01
5.07099152e-01 -1.29062593e+00 -5.11103988e-01 1.31126761e+00
7.47544229e-01 -1.27647305e+00 -2.03178257e-01 -2.75385410e-01
-6.56257510e-01 1.37460768e+00 2.02965438e-01 -5.54268174e-02
1.26905811e+00 9.34184529e-03 6.87524438e-01 4.37776037e-02
-2.60454804e-01 -5.99114262e-02 4.00363952e-01 6.51029408e-01
8.70051503e-01 2.73946583e-01 -9.33710694e-01 6.82414174e-01
-6.78976476e-01 -5.09362161e-01 6.24625087e-01 6.07652605e-01
-5.86116970e-01 -1.17958629e+00 -6.90679133e-01 7.49070823e-01
-8.40872586e-01 -3.06311816e-01 -4.25887406e-01 7.60647416e-01
-4.82734703e-02 1.35021019e+00 2.40322761e-02 -4.40539241e-01
-7.40438029e-02 5.98367478e-05 2.98558384e-01 -1.82712272e-01
-9.29863811e-01 3.22685614e-02 5.73259473e-01 2.17050314e-01
-4.77395684e-01 -9.92498338e-01 -1.16642976e+00 -8.37059796e-01
-1.12860702e-01 3.24550420e-01 8.49639595e-01 1.09837508e+00
-4.21627581e-01 -5.30948818e-01 1.16372979e+00 -7.25890517e-01
-5.93336105e-01 -1.28803182e+00 -8.82232070e-01 2.49016136e-01
2.49034703e-01 -4.79437470e-01 -2.92829096e-01 7.79704526e-02] | [10.184222221374512, 10.722079277038574] |
f5e9b226-ab39-4181-b0c9-323597ff3379 | ernie-sparse-learning-hierarchical-efficient-1 | 2203.12276 | null | https://arxiv.org/abs/2203.12276v1 | https://arxiv.org/pdf/2203.12276v1.pdf | ERNIE-SPARSE: Learning Hierarchical Efficient Transformer Through Regularized Self-Attention | Sparse Transformer has recently attracted a lot of attention since the ability for reducing the quadratic dependency on the sequence length. We argue that two factors, information bottleneck sensitivity and inconsistency between different attention topologies, could affect the performance of the Sparse Transformer. This paper proposes a well-designed model named ERNIE-Sparse. It consists of two distinctive parts: (i) Hierarchical Sparse Transformer (HST) to sequentially unify local and global information. (ii) Self-Attention Regularization (SAR) method, a novel regularization designed to minimize the distance for transformers with different attention topologies. To evaluate the effectiveness of ERNIE-Sparse, we perform extensive evaluations. Firstly, we perform experiments on a multi-modal long sequence modeling task benchmark, Long Range Arena (LRA). Experimental results demonstrate that ERNIE-Sparse significantly outperforms a variety of strong baseline methods including the dense attention and other efficient sparse attention methods and achieves improvements by 2.77% (57.78% vs. 55.01%). Secondly, to further show the effectiveness of our method, we pretrain ERNIE-Sparse and verified it on 3 text classification and 2 QA downstream tasks, achieve improvements on classification benchmark by 0.83% (92.46% vs. 91.63%), on QA benchmark by 3.24% (74.67% vs. 71.43%). Experimental results continue to demonstrate its superior performance. | ['Haifeng Wang', 'Hua Wu', 'Hao Tian', 'Yu Sun', 'Zhida Feng', 'Shikun Feng', 'Yuxiang Lu', 'Li Chen', 'Jiaxiang Liu', 'Yang Liu'] | 2022-03-23 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 2.18072027e-01 -6.93119392e-02 -1.99109316e-02 -1.91416696e-01
-1.39469278e+00 -2.90639967e-01 4.27917689e-01 -2.50880271e-01
-2.69575864e-01 5.84148824e-01 6.44369781e-01 -2.59813219e-01
-6.63671494e-02 -3.38000506e-01 -7.24956572e-01 -7.58085907e-01
6.16806000e-03 5.04531741e-01 2.55098313e-01 -4.58010912e-01
3.37536156e-01 7.09341392e-02 -1.19092226e+00 5.48373938e-01
1.19578695e+00 9.50160325e-01 5.95195413e-01 4.86515224e-01
-4.25872989e-02 1.10615683e+00 -5.62229753e-01 -3.44163716e-01
3.43877934e-02 -3.46117407e-01 -1.05208802e+00 -1.89202264e-01
2.09493250e-01 -5.76096699e-02 -3.45771164e-01 9.07683671e-01
6.62426412e-01 1.01015598e-01 5.63341498e-01 -1.08073604e+00
-8.48945856e-01 1.12548697e+00 -9.73250508e-01 6.01636469e-01
2.62692094e-01 1.57225281e-01 1.43596387e+00 -1.02099419e+00
5.08978546e-01 1.37511563e+00 9.46189165e-01 3.17035541e-02
-8.08735549e-01 -6.50171638e-01 3.55611295e-01 5.41158736e-01
-1.52680743e+00 -5.08149326e-01 3.88698459e-01 -2.63805598e-01
1.41364634e+00 3.05687100e-01 5.34105375e-02 1.07567763e+00
2.46717170e-01 1.00916398e+00 1.14802480e+00 -3.04988205e-01
-2.18416721e-01 -2.01076508e-01 4.95054871e-01 7.53715754e-01
-1.35559425e-01 -6.04090728e-02 -6.35395646e-01 -6.62108585e-02
2.91638792e-01 -2.63298571e-01 -5.63907564e-01 1.13773175e-01
-1.27174211e+00 8.54595780e-01 4.68096972e-01 4.26764488e-01
-3.33668828e-01 -1.79256335e-01 5.15769064e-01 3.15828919e-01
4.87968922e-01 3.55595261e-01 -6.98758185e-01 -4.41547602e-01
-6.09101117e-01 -6.47692606e-02 4.40857172e-01 1.09554708e+00
4.55041200e-01 1.88141584e-01 -6.74188673e-01 1.24468422e+00
3.60001773e-01 5.80667555e-01 8.35719764e-01 -6.57718122e-01
8.68675947e-01 2.97605336e-01 -2.80595958e-01 -9.25010204e-01
-3.37440461e-01 -7.66484916e-01 -1.07333374e+00 -7.04383254e-01
1.43488631e-01 5.05456477e-02 -7.45924413e-01 1.69757831e+00
-1.52779326e-01 1.28127232e-01 1.90651305e-02 9.36115801e-01
9.29964304e-01 1.01858592e+00 7.27635100e-02 -1.63302377e-01
1.35303831e+00 -1.35554099e+00 -7.81420112e-01 -2.38128960e-01
8.07113171e-01 -9.62841094e-01 1.48240542e+00 3.24977607e-01
-1.17123544e+00 -5.66122711e-01 -8.52190912e-01 -2.36480013e-01
2.62920037e-02 1.55177787e-01 2.59607583e-01 4.63410825e-01
-1.16129112e+00 3.78318429e-01 -6.24665141e-01 -3.45522940e-01
3.97768050e-01 2.31450275e-01 -7.28097232e-03 -2.29595020e-01
-1.36642349e+00 7.26763010e-01 1.62129194e-01 -1.34398884e-04
-8.89492154e-01 -1.05008984e+00 -8.17283750e-01 5.38619757e-01
2.54531533e-01 -6.47625446e-01 1.27644598e+00 -7.35226691e-01
-1.66054702e+00 5.99087358e-01 -3.42895031e-01 -6.37841344e-01
1.92426443e-01 -6.05629861e-01 -1.42467812e-01 -9.28753521e-03
3.10272872e-01 5.35474420e-01 4.60860461e-01 -9.50018048e-01
-6.06396675e-01 -3.09867710e-01 -1.94315657e-01 3.10559213e-01
-2.97659755e-01 1.31943300e-01 -7.22960651e-01 -8.31369400e-01
-1.56626910e-01 -9.90246773e-01 6.18976839e-02 -8.77967000e-01
-6.34074390e-01 -4.52236086e-01 9.78438258e-01 -8.93513858e-01
1.42936349e+00 -2.10618711e+00 2.67668605e-01 -5.38179018e-02
9.81480107e-02 4.23575044e-01 -5.03552020e-01 5.13871014e-01
-1.46139771e-01 1.61309198e-01 -3.83584797e-01 -5.01920879e-01
-1.06738873e-01 2.57074218e-02 -3.94133836e-01 3.77013564e-01
1.12203330e-01 1.05539000e+00 -6.92043841e-01 -3.43522608e-01
-1.67411134e-01 6.87162995e-01 -7.04294086e-01 2.00793147e-01
3.84339690e-02 3.55342329e-01 -4.81572658e-01 7.59753644e-01
5.47358990e-01 -6.62032962e-01 2.29299963e-01 -4.73777831e-01
-1.02558881e-01 6.54029906e-01 -6.46958411e-01 1.80470932e+00
-4.76647168e-01 6.62979245e-01 -1.50122792e-01 -1.02375567e+00
6.76549017e-01 2.34435230e-01 3.77802223e-01 -1.02237213e+00
2.88598239e-01 -4.59607132e-02 1.37139753e-01 -3.37190688e-01
6.21598959e-01 2.41720621e-02 -2.17370000e-02 3.32721680e-01
-6.56511411e-02 4.41497177e-01 2.66470835e-02 6.82139039e-01
1.38187671e+00 -1.05070271e-01 1.02907255e-01 -5.75668573e-01
5.78583181e-01 -3.51556778e-01 6.51119173e-01 7.76549935e-01
-2.62963951e-01 6.91488862e-01 4.91573691e-01 -4.58192676e-02
-8.34021807e-01 -7.30427802e-01 1.01890847e-01 1.33185887e+00
1.51689246e-01 -7.42308199e-01 -5.89006603e-01 -7.45427489e-01
-2.34342664e-01 5.69728374e-01 -4.84980136e-01 -2.06631109e-01
-6.23000681e-01 -9.13844168e-01 6.66038394e-01 6.87055886e-01
6.64875567e-01 -8.64226043e-01 -1.13462545e-01 4.43258468e-05
-8.41896534e-01 -1.21688199e+00 -9.00209844e-01 -9.35060680e-02
-7.21221149e-01 -6.77407205e-01 -6.07825994e-01 -7.20453799e-01
1.23703606e-01 6.16278946e-01 1.31980050e+00 1.53200969e-01
7.35381991e-02 2.62252390e-01 -6.96158707e-01 6.04292192e-02
-7.49771893e-02 6.81507766e-01 -2.08861455e-01 -3.30769382e-02
4.58644256e-02 -4.48935658e-01 -3.28250527e-01 4.99789804e-01
-4.74729538e-01 -2.82192472e-02 9.27969098e-01 1.06756234e+00
5.39021850e-01 -2.03380167e-01 7.62632012e-01 -9.58071113e-01
6.76815450e-01 -6.14777267e-01 -3.19923282e-01 2.86538541e-01
-6.20286405e-01 2.29430944e-01 5.73555171e-01 -3.94060947e-02
-1.22593760e+00 -3.13668162e-01 -5.00978351e-01 -3.98814678e-01
2.68080950e-01 8.33361387e-01 -3.05073529e-01 1.84273452e-01
2.88000852e-01 3.85380179e-01 -1.12972036e-01 -5.85991263e-01
3.70458066e-01 7.71628618e-01 2.15399250e-01 -4.31084096e-01
5.40989041e-01 9.36112702e-02 -5.74050665e-01 -8.37198138e-01
-1.15228617e+00 -5.98669946e-01 -1.48753673e-01 1.79339916e-01
7.40067780e-01 -1.07225835e+00 -7.16900289e-01 3.87682498e-01
-1.01261222e+00 -4.03663248e-01 5.12050688e-02 3.34896475e-01
-3.17767978e-01 6.70724332e-01 -9.71612871e-01 -4.88651037e-01
-8.30661952e-01 -1.36675763e+00 1.19003725e+00 -1.28215671e-01
-1.37187123e-01 -6.96941257e-01 -1.12826545e-02 8.57006490e-01
6.39912128e-01 -3.54124218e-01 7.46080279e-01 -6.52909696e-01
-7.50940144e-01 3.94554228e-01 -5.57527900e-01 1.36861011e-01
-5.80721833e-02 -1.90164506e-01 -1.03556633e+00 -4.97023672e-01
-1.37375519e-01 -5.25335014e-01 1.00610888e+00 3.56415331e-01
1.27714503e+00 -3.15541208e-01 -3.86262149e-01 6.77428007e-01
1.23903739e+00 2.66308486e-01 8.95193338e-01 3.97439241e-01
9.01210129e-01 5.44820130e-02 6.83385849e-01 3.47353369e-01
5.85516453e-01 8.04531455e-01 1.80047020e-01 4.87443097e-02
-3.41453910e-01 -1.60603553e-01 6.24268234e-01 1.58363581e+00
9.22982097e-02 -5.11187196e-01 -8.85067046e-01 5.64946830e-01
-1.86960542e+00 -9.23967481e-01 -2.14519009e-01 1.78084993e+00
7.97288716e-01 1.19092517e-01 1.18425101e-01 1.88487798e-01
6.28889024e-01 2.00905859e-01 -3.33257437e-01 -3.44098032e-01
-3.20647538e-01 1.75370455e-01 3.83862644e-01 5.85616231e-01
-9.16177630e-01 1.01733756e+00 5.96218204e+00 1.26005173e+00
-9.08731639e-01 3.16683352e-01 7.52440631e-01 1.65166914e-01
-1.91141278e-01 -3.06645095e-01 -9.82191205e-01 5.94082594e-01
1.16597843e+00 -9.63086709e-02 2.42750078e-01 5.36856711e-01
1.20652311e-01 3.15957129e-01 -6.54783666e-01 8.83634925e-01
2.28275776e-01 -1.27947783e+00 1.85916230e-01 -1.15356416e-01
6.55144572e-01 5.20471752e-01 2.10757226e-01 7.38696516e-01
3.99653882e-01 -8.91449213e-01 6.15623295e-01 2.60098964e-01
7.22962201e-01 -7.45969474e-01 1.02295887e+00 3.00383240e-01
-1.39202607e+00 -9.72747654e-02 -3.58606845e-01 3.63011248e-02
3.25901985e-01 6.42823100e-01 -6.51135027e-01 7.42231369e-01
9.39866304e-01 1.09925711e+00 -4.55529392e-01 8.65732551e-01
-1.60812065e-01 1.08627415e+00 -2.26271510e-01 1.50323614e-01
4.71614182e-01 -3.61259319e-02 6.04660332e-01 1.45742142e+00
2.69221872e-01 7.61563778e-02 5.91149703e-02 5.55236578e-01
-2.37013862e-01 2.27473646e-01 -2.48413756e-01 1.13795310e-01
5.41102886e-01 1.03639114e+00 -2.62307853e-01 -3.59453976e-01
-4.78995204e-01 8.89705896e-01 5.65409064e-01 3.11488926e-01
-1.19328356e+00 -4.17281479e-01 5.16185284e-01 -1.41265705e-01
7.50596404e-01 -1.96451489e-02 -1.35646984e-01 -1.18599761e+00
-1.75851882e-01 -1.38597298e+00 6.73807442e-01 -1.03894520e+00
-1.45609593e+00 9.98273313e-01 -3.39717925e-01 -8.69078696e-01
2.33488396e-01 -1.38631999e-01 -2.73936391e-01 1.01897681e+00
-1.80153143e+00 -1.09644544e+00 -2.40408301e-01 5.58122456e-01
9.01921391e-01 -3.24452430e-01 4.61025119e-01 6.10803246e-01
-8.89968038e-01 1.02182460e+00 2.26781517e-01 1.99116096e-02
7.21765161e-01 -9.87361968e-01 4.20899063e-01 7.44927704e-01
1.14613928e-01 5.61811388e-01 2.71994144e-01 -3.57872963e-01
-1.48213232e+00 -1.14940751e+00 1.17894435e+00 -4.67165232e-01
7.27598429e-01 -2.25242242e-01 -1.19428062e+00 8.95596683e-01
6.28266633e-01 -3.19959104e-01 5.58090091e-01 3.51020634e-01
-5.53151190e-01 -1.63959444e-01 -7.62678146e-01 4.34842616e-01
1.15844882e+00 -5.04255831e-01 -5.18561006e-01 2.47187257e-01
1.15921021e+00 -4.09491092e-01 -1.05206192e+00 5.30628204e-01
3.01505059e-01 -1.06547368e+00 1.06483710e+00 -3.54100674e-01
4.10157114e-01 -1.87014729e-01 -2.49452069e-01 -1.36292171e+00
-8.61871064e-01 -7.54568756e-01 8.68056994e-03 1.27561533e+00
4.80450660e-01 -7.50859320e-01 4.83321518e-01 -7.01256320e-02
-6.48064673e-01 -1.01554978e+00 -8.00019801e-01 -8.23683262e-01
1.56471252e-01 -2.86338776e-01 5.84118724e-01 1.06269944e+00
-6.78799897e-02 1.01608741e+00 -5.49316168e-01 2.79799014e-01
3.37952793e-01 1.98576331e-01 4.28031117e-01 -7.12512016e-01
-3.67238283e-01 -5.17449856e-01 -9.03334618e-02 -1.63360798e+00
1.48913294e-01 -9.73570287e-01 -7.38748908e-02 -1.32892406e+00
4.99646902e-01 -5.63261807e-01 -4.43638742e-01 5.52496135e-01
-5.25329828e-01 9.46376100e-02 2.07642511e-01 3.65561664e-01
-8.77849460e-01 1.03696382e+00 1.07774723e+00 -2.45415047e-01
-9.01552325e-04 -3.58496726e-01 -9.95519757e-01 4.18266058e-01
1.02880383e+00 -4.60781932e-01 -4.83435899e-01 -7.62065411e-01
8.54577590e-03 -4.98626046e-02 -1.39543846e-01 -7.12327003e-01
4.56583723e-02 2.83444226e-01 -1.37969837e-01 -7.24015772e-01
8.70759040e-02 -3.55880946e-01 -1.60763487e-01 4.11712170e-01
-4.34596270e-01 3.69592160e-01 2.49646962e-01 5.38321733e-01
-3.56820643e-01 1.42904237e-01 7.98387647e-01 1.82665661e-01
-7.44750202e-01 3.18837941e-01 -3.09945703e-01 5.46115696e-01
5.43334246e-01 1.60453215e-01 -4.85044241e-01 -3.74626517e-01
-4.05425906e-01 3.75467211e-01 -1.30839124e-01 5.43609858e-01
5.59374988e-01 -1.22865486e+00 -1.00497782e+00 8.21766853e-02
2.99106073e-02 -3.81584585e-01 4.11588550e-01 1.16208601e+00
-1.92443058e-01 8.08182001e-01 2.00662315e-01 -7.91305006e-01
-1.33764696e+00 4.86999929e-01 1.82779223e-01 -7.65590668e-01
-7.00912714e-01 1.15907252e+00 3.61347705e-01 -3.90380323e-01
2.26758674e-01 -4.91770327e-01 -1.82125300e-01 -1.56237096e-01
4.80009496e-01 4.13901687e-01 2.87674636e-01 -8.54266763e-01
-5.29116035e-01 7.03325152e-01 -4.55769777e-01 2.56962270e-01
1.23932111e+00 -2.94605643e-01 -2.93338820e-02 1.37558401e-01
1.41727579e+00 6.96949437e-02 -8.69142890e-01 -5.32991409e-01
-7.38168061e-02 -4.76607412e-01 1.64859399e-01 -1.10349131e+00
-1.32730079e+00 8.40461969e-01 4.35227066e-01 2.30265662e-01
1.21138716e+00 1.75847523e-02 1.20733976e+00 1.93694666e-01
2.18920007e-01 -6.43190086e-01 1.32279530e-01 1.01087165e+00
1.10809088e+00 -1.13726795e+00 -1.38577849e-01 -5.39715409e-01
-8.81112278e-01 4.78937685e-01 4.79264468e-01 2.15760157e-01
5.16606390e-01 3.00013274e-01 3.61000653e-03 -2.54082352e-01
-1.09984338e+00 -1.84905365e-01 4.21175092e-01 4.26093876e-01
8.03489208e-01 -2.27037579e-01 -3.06350499e-01 8.03522408e-01
-1.94507107e-01 -2.10922942e-01 1.80894494e-01 6.71660125e-01
-3.76480520e-01 -9.82937276e-01 -2.00602919e-01 3.63624245e-01
-5.38289845e-01 -5.25838971e-01 -1.61687702e-01 6.92606509e-01
-1.79530740e-01 1.03496051e+00 -1.26632899e-01 -4.72255141e-01
4.98128086e-01 -1.72476277e-01 1.35687560e-01 -4.48228091e-01
-8.56555521e-01 2.10175529e-01 3.14040661e-01 -6.48010492e-01
-3.24541420e-01 -7.36849070e-01 -1.08208704e+00 -3.78345996e-01
-4.08607155e-01 3.51226270e-01 1.20442972e-01 9.18908477e-01
8.97197604e-01 9.48756933e-01 5.05015910e-01 -3.48508298e-01
-7.52445340e-01 -1.18633008e+00 -3.09004396e-01 3.35526288e-01
1.46188140e-01 -6.14785910e-01 -4.77793753e-01 -1.32661937e-02] | [10.905529022216797, 6.6412153244018555] |
30f40fce-8853-4faa-99ff-d6e7cad78d25 | transform-contrast-and-tell-coherent-entity | 2302.02124 | null | https://arxiv.org/abs/2302.02124v1 | https://arxiv.org/pdf/2302.02124v1.pdf | Transform, Contrast and Tell: Coherent Entity-Aware Multi-Image Captioning | Coherent entity-aware multi-image captioning aims to generate coherent captions for multiple adjacent images in a news document. There are coherence relationships among adjacent images because they often describe same entities or events. These relationships are important for entity-aware multi-image captioning, but are neglected in entity-aware single-image captioning. Most existing work focuses on single-image captioning, while multi-image captioning has not been explored before. Hence, this paper proposes a coherent entity-aware multi-image captioning model by making use of coherence relationships. The model consists of a Transformer-based caption generation model and two types of contrastive learning-based coherence mechanisms. The generation model generates the caption by paying attention to the image and the accompanying text. The horizontal coherence mechanism aims to the make the caption coherent with captions of adjacent images. The vertical coherence mechanism aims to make the caption coherent with the image and the accompanying text. To evaluate coherence between captions, two coherence evaluation metrics are proposed. The new dataset DM800K is constructed that has more images per document than two existing datasets GoodNews and NYT800K, and are more suitable for multi-image captioning. Experiments on three datasets show the proposed captioning model outperforms 6 baselines according to single-image captioning evaluations, and the generated captions are more coherent than that of baselines according to coherence evaluations and human evaluations. | ['Jingqiang Chen'] | 2023-02-04 | null | null | null | null | ['coherence-evaluation'] | ['natural-language-processing'] | [ 2.98643529e-01 2.78100312e-01 -2.02488959e-01 -3.75230908e-01
-1.21347618e+00 -3.86583000e-01 9.54119146e-01 8.10167752e-03
-2.50211984e-01 9.11188543e-01 8.03026736e-01 1.73104838e-01
3.39849502e-01 -4.14530456e-01 -1.14413857e+00 -5.52745104e-01
3.22854072e-01 6.91938400e-01 3.35608095e-01 -2.14790910e-01
8.72388184e-02 -4.15655583e-01 -1.19950950e+00 9.43905652e-01
7.67349005e-01 6.63382411e-01 6.67651355e-01 5.95112205e-01
-3.15812737e-01 8.17649007e-01 -6.92388177e-01 -6.06296003e-01
-1.81993008e-01 -8.35436165e-01 -7.29518235e-01 3.33761334e-01
6.69336557e-01 -2.20630512e-01 -1.81941688e-01 9.25380051e-01
5.05052328e-01 -2.21993491e-01 6.53331459e-01 -1.65582621e+00
-1.23386943e+00 8.26171577e-01 -8.43584955e-01 1.71195805e-01
4.07771945e-01 -3.28147709e-02 1.06665206e+00 -7.81990469e-01
7.07740426e-01 1.31920815e+00 2.11727992e-01 5.17641723e-01
-7.96142519e-01 -5.23915231e-01 2.35643923e-01 3.42217177e-01
-1.15835392e+00 -2.15620831e-01 6.23152316e-01 -3.46765161e-01
3.68898720e-01 2.30416924e-01 3.43872249e-01 1.29024684e+00
7.86092356e-02 1.07475603e+00 9.79950547e-01 -4.10279721e-01
-2.74796218e-01 3.66134703e-01 9.52282622e-02 2.88174689e-01
2.84969866e-01 -2.76088953e-01 -3.21513146e-01 1.32309347e-01
7.50341594e-01 -2.99405932e-01 -4.59748954e-01 -1.78670853e-01
-1.89692116e+00 8.62832963e-01 5.61626375e-01 4.63617027e-01
-5.49969137e-01 2.73783505e-01 4.16774035e-01 -1.97835475e-01
3.93511355e-01 6.15286887e-01 9.31098983e-02 1.85255453e-01
-9.38475966e-01 2.51361519e-01 2.63080627e-01 1.14367628e+00
6.61604285e-01 -3.30146551e-01 -8.77204120e-01 7.30399549e-01
4.24038470e-01 9.02087986e-01 6.54306650e-01 -7.09951401e-01
9.63799000e-01 4.37331855e-01 4.74368960e-01 -1.26757598e+00
-9.07189026e-03 -2.75802135e-01 -9.26694334e-01 -5.90904295e-01
-5.49090505e-02 -2.99448729e-01 -9.15480673e-01 1.83396864e+00
-1.10374633e-02 4.09731269e-01 6.10638797e-01 1.01774096e+00
1.36616504e+00 1.37933385e+00 3.79235655e-01 -4.39020663e-01
1.79261053e+00 -1.57939923e+00 -1.10966849e+00 -5.73044837e-01
3.31959218e-01 -1.09660339e+00 9.05427456e-01 -3.73251349e-01
-1.13645768e+00 -7.43904352e-01 -8.85961115e-01 1.29793629e-01
-1.94790453e-01 2.39024088e-01 2.22751245e-01 1.33600757e-01
-9.70276892e-01 -3.47883046e-01 -8.13146085e-02 -3.08998197e-01
1.19124904e-01 -1.88866124e-01 -4.99996841e-01 -2.15542540e-01
-1.48769319e+00 1.03117955e+00 8.87849867e-01 -1.29997656e-01
-7.56289601e-01 -3.29729766e-01 -9.60196316e-01 5.20936400e-02
5.19685335e-02 -1.04087853e+00 1.40526319e+00 -1.32345140e+00
-7.89026678e-01 8.91960859e-01 -2.90135920e-01 -4.69165981e-01
3.17440957e-01 -1.52043909e-01 -6.25278711e-01 5.00418961e-01
5.65217018e-01 1.37230217e+00 6.67061567e-01 -1.89038312e+00
-6.54358327e-01 2.87010461e-01 -5.53889833e-02 8.08854342e-01
-3.71525377e-01 1.27530560e-01 -1.08578181e+00 -6.25768960e-01
-3.63955170e-01 -9.31868613e-01 -3.89039470e-03 -4.88823414e-01
-5.33002973e-01 5.01856543e-02 9.82034028e-01 -6.15480244e-01
9.93259370e-01 -2.03751612e+00 -8.67525861e-03 -5.05089700e-01
-1.27909675e-01 2.06862092e-01 -6.22109830e-01 4.59676683e-01
-1.24057181e-01 2.03837186e-01 -1.43531188e-01 -4.39520448e-01
-1.97515845e-01 1.83976933e-01 -5.45258880e-01 -1.18537344e-01
1.83475986e-01 1.40330660e+00 -1.05096614e+00 -1.05911112e+00
2.00039402e-01 4.34552312e-01 7.63005298e-03 5.05874097e-01
-5.35224199e-01 2.92674273e-01 -4.90839779e-01 2.57769853e-01
6.65071189e-01 -8.17530632e-01 -5.51122762e-02 -6.71949506e-01
1.73326477e-01 -2.95900643e-01 -8.44395041e-01 1.46405637e+00
-4.94807005e-01 7.53579617e-01 -3.99523616e-01 -5.77969015e-01
7.66964376e-01 8.33506227e-01 2.66316593e-01 -8.99611175e-01
6.65641157e-03 1.21981539e-01 -3.96122724e-01 -5.96186996e-01
8.93952668e-01 -8.12341943e-02 -4.94547665e-01 4.80203629e-01
-2.67414510e-01 -1.46386340e-01 4.09058422e-01 6.17044806e-01
4.63776648e-01 -6.61519021e-02 3.71572524e-02 2.58821934e-01
5.24263740e-01 3.13507885e-01 1.46435872e-01 6.93736732e-01
-4.88350429e-02 1.30486596e+00 2.83124268e-01 -2.79594451e-01
-1.33214700e+00 -9.13396418e-01 2.27386937e-01 7.53735483e-01
9.61870611e-01 -2.82670081e-01 -6.33024931e-01 -5.97534716e-01
-4.98520106e-01 8.75825703e-01 -7.41581380e-01 5.30502051e-02
-5.63630760e-01 -8.15349579e-01 2.99281955e-01 3.14577490e-01
9.87829328e-01 -1.47253060e+00 -2.61294246e-01 1.97400302e-01
-1.30469358e+00 -1.60263824e+00 -9.35476661e-01 -5.27109504e-01
-3.62965107e-01 -8.02514732e-01 -1.38312209e+00 -1.24690890e+00
7.83011734e-01 6.69182956e-01 1.37172854e+00 -1.70922965e-01
2.45388657e-01 5.62195539e-01 -6.41809464e-01 -4.51368362e-01
-6.58715725e-01 -5.83796054e-02 -3.42998654e-01 2.38242164e-01
3.08467075e-02 -8.36699307e-02 -5.04763782e-01 4.78652388e-01
-1.11103392e+00 1.01473320e+00 1.11054587e+00 7.60163367e-01
6.39195263e-01 -3.78714859e-01 6.66653454e-01 -6.20350122e-01
8.19930673e-01 -8.43697548e-01 -1.05990373e-01 7.88917542e-01
-2.45348871e-01 7.18861353e-03 4.11809146e-01 -5.32304168e-01
-1.20666254e+00 2.81546060e-02 3.00475150e-01 -4.55052614e-01
1.79705434e-02 4.14186835e-01 -9.28274840e-02 5.08131146e-01
3.36800724e-01 6.91011965e-01 -1.94074169e-01 -1.93562475e-03
7.48230457e-01 8.19689870e-01 9.45694923e-01 -4.53684837e-01
6.79554522e-01 1.32639185e-01 -4.95634913e-01 -4.80509937e-01
-1.07177269e+00 -6.09747350e-01 -2.14592472e-01 -5.42616904e-01
1.25088596e+00 -1.33741891e+00 -9.75709483e-02 3.06061685e-01
-1.80138278e+00 2.81664729e-01 -4.93105175e-03 4.80717123e-01
-6.36592269e-01 3.95489454e-01 -3.60543132e-01 -3.48889679e-01
-6.46035373e-01 -1.32756829e+00 1.50195849e+00 5.45275331e-01
3.84034701e-02 -7.93928266e-01 1.53670758e-01 7.67633319e-01
1.85554400e-01 4.42855686e-01 5.59317112e-01 -5.18397331e-01
-5.78634918e-01 -1.19527251e-01 -6.42488241e-01 1.10325843e-01
9.63839889e-02 -1.97499603e-01 -6.88392162e-01 -8.78855214e-02
-4.42311883e-01 -4.60248262e-01 6.67165101e-01 4.02923226e-01
6.80923700e-01 -5.71984768e-01 -5.19551635e-01 -4.14535515e-02
1.56528473e+00 2.36293450e-01 9.44066584e-01 6.73478186e-01
6.98574722e-01 5.62233031e-01 7.31406569e-01 3.76811847e-02
6.45749748e-01 8.40306699e-01 6.20451391e-01 -4.36244696e-01
-4.38879877e-01 -4.98921156e-01 2.93940812e-01 9.09360945e-01
3.42202157e-01 -9.69792902e-01 -7.47557044e-01 1.01408219e+00
-2.25259829e+00 -1.32052767e+00 -4.90101427e-01 1.60019672e+00
7.75759399e-01 -8.58699828e-02 -1.66827679e-01 -5.13763964e-01
1.37596846e+00 3.58893603e-01 -3.56086433e-01 -1.87863499e-01
-5.55507183e-01 -7.58735657e-01 1.52543291e-01 3.24139178e-01
-1.12981784e+00 8.66512060e-01 6.02022362e+00 8.03030431e-01
-8.04624140e-01 2.79279888e-01 9.48935091e-01 1.82422146e-01
-5.01356244e-01 -5.37130311e-02 -8.89151156e-01 7.88477480e-01
9.07436907e-01 -5.22417784e-01 -4.17097807e-01 7.88202882e-01
1.95085242e-01 -9.96658579e-02 -8.14861774e-01 1.20803869e+00
5.78473806e-01 -1.68434954e+00 7.46349037e-01 -2.06930086e-01
1.06615293e+00 -1.14567757e-01 1.07297175e-01 1.53695002e-01
1.31280646e-01 -9.45064902e-01 8.15926492e-01 4.98779744e-01
6.51709020e-01 -4.74526793e-01 1.19822645e+00 1.27403229e-01
-1.24882567e+00 3.11537176e-01 -1.64347470e-01 3.28861713e-01
9.92410839e-01 5.01807213e-01 -8.55187058e-01 8.28297019e-01
4.89754349e-01 7.31836319e-01 -6.93267703e-01 1.19361603e+00
-3.04565698e-01 2.45751277e-01 6.57485574e-02 -6.43101260e-02
6.46271944e-01 1.08687757e-02 4.51750010e-01 1.35982060e+00
3.88916999e-01 2.14874111e-02 1.06973194e-01 6.66347742e-01
-3.77874881e-01 4.35817242e-01 -4.39285129e-01 -7.65620172e-02
4.24198419e-01 1.35249329e+00 -6.79781199e-01 -7.08299577e-01
-5.30167878e-01 1.27589798e+00 -1.01491332e-01 2.59140819e-01
-1.15322340e+00 -1.37128219e-01 4.68045287e-02 -4.57217619e-02
3.26635063e-01 1.45221114e-01 1.27306879e-02 -1.15450811e+00
2.23511666e-01 -8.89183819e-01 3.98106158e-01 -1.70735443e+00
-1.21838355e+00 1.11725509e+00 3.45738024e-01 -1.40955055e+00
-2.19004914e-01 4.42158505e-02 -8.05054963e-01 7.56785214e-01
-1.76161957e+00 -1.52140450e+00 -5.79099715e-01 4.36868221e-01
8.01533937e-01 4.15141396e-02 5.81719041e-01 2.84789681e-01
-5.30668020e-01 2.98223943e-01 3.90540548e-02 1.21831737e-01
1.01846385e+00 -9.39664423e-01 2.94055790e-01 1.00530553e+00
2.62513369e-01 1.40266031e-01 9.66370463e-01 -7.60738730e-01
-6.32205546e-01 -1.35645235e+00 1.15948033e+00 -3.44098508e-01
2.73657143e-01 -3.70138362e-02 -9.42886412e-01 5.92339814e-01
9.29479599e-01 -2.05140159e-01 5.09779692e-01 -4.63559270e-01
-3.05366039e-01 6.17597178e-02 -7.64214456e-01 5.86472094e-01
4.58881855e-01 -1.64834023e-01 -9.87777650e-01 7.54504740e-01
1.17378771e+00 -2.81205893e-01 -5.66313565e-01 3.07315111e-01
2.41707951e-01 -5.81234038e-01 1.02964139e+00 -1.16667286e-01
1.00214148e+00 -6.48762465e-01 3.33766602e-02 -1.18967915e+00
-3.34129363e-01 -3.93400401e-01 1.54024124e-01 1.57720006e+00
6.63510799e-01 -2.60603870e-03 3.59659106e-01 3.69930774e-01
-2.97925562e-01 -1.74395412e-01 -6.19294763e-01 -7.12918639e-01
-1.73446372e-01 1.69145390e-01 5.30382991e-01 9.06562209e-01
-1.80494085e-01 7.93365777e-01 -9.12759423e-01 2.30626494e-01
4.00221020e-01 5.16103029e-01 7.18346536e-01 -6.73018157e-01
-6.99550807e-02 -3.05044383e-01 -8.28188136e-02 -1.06068492e+00
-2.45698076e-02 -6.32804811e-01 2.91793227e-01 -2.24833488e+00
8.72338772e-01 -5.82131594e-02 3.60516235e-02 5.59531391e-01
-5.02192855e-01 6.66905403e-01 3.67306501e-01 7.07944632e-01
-1.23578823e+00 7.45778501e-01 1.59526360e+00 -4.13892955e-01
1.14282228e-01 -5.21099448e-01 -8.91430318e-01 3.16517264e-01
5.68553746e-01 -4.30666715e-01 -5.57149351e-01 -6.34533525e-01
8.57416391e-02 3.15215349e-01 2.36824974e-01 -1.01245725e+00
2.80543238e-01 -2.97409058e-01 2.44402781e-01 -7.07599759e-01
3.64936143e-01 -5.36700785e-01 4.11401510e-01 3.28435421e-01
-4.57151741e-01 3.85544300e-01 3.32885474e-01 5.52485108e-01
-7.71322370e-01 -2.61741400e-01 8.38572919e-01 -2.97576249e-01
-1.02499759e+00 2.67711371e-01 -1.66174605e-01 2.97159910e-01
1.43674803e+00 -1.08011454e-01 -7.06167340e-01 -8.12142611e-01
-3.41964573e-01 5.02458453e-01 3.17173183e-01 8.92964363e-01
6.16846025e-01 -1.69698489e+00 -1.32035208e+00 -5.67495048e-01
6.02592587e-01 -8.78219083e-02 3.87826532e-01 4.03689474e-01
-4.56786603e-01 6.88060164e-01 -3.52579325e-01 -6.36096120e-01
-1.50759685e+00 6.98547363e-01 -6.30990695e-03 -4.27108854e-01
-4.26136494e-01 6.60264015e-01 6.67269170e-01 1.72890291e-01
2.48146299e-02 8.14671963e-02 -6.40627444e-01 1.28085002e-01
8.43565226e-01 -4.19419914e-01 -5.15828907e-01 -1.24772024e+00
-1.25786558e-01 7.06189752e-01 -3.44293416e-01 -1.64636940e-01
1.08058453e+00 -6.05622828e-01 -1.62601992e-02 2.55585164e-01
1.17821836e+00 -4.50991988e-01 -9.13281262e-01 -1.85472757e-01
-3.23953867e-01 -3.04719418e-01 -1.01365745e-01 -9.86105621e-01
-8.30016971e-01 5.88850856e-01 4.68525052e-01 1.20928466e-01
1.08689570e+00 2.57451534e-01 1.10942519e+00 -2.17530336e-02
3.78478281e-02 -7.27907419e-01 8.07860970e-01 2.56089598e-01
1.27797174e+00 -1.61412168e+00 -1.41084373e-01 -4.59351033e-01
-1.25189459e+00 8.58953655e-01 9.79506195e-01 4.11434650e-01
-9.75679904e-02 -2.85234779e-01 1.62183508e-01 -1.69081315e-01
-6.62219822e-01 -1.90332323e-01 5.66779554e-01 5.48611820e-01
2.17423856e-01 -1.32461160e-01 -2.93352395e-01 6.05204403e-01
9.36543867e-02 -1.66017383e-01 7.87027836e-01 5.27244151e-01
-5.16453385e-01 -9.28115904e-01 -6.78817868e-01 4.99019995e-02
-1.83148012e-01 -3.70587707e-01 -3.13036799e-01 7.30838180e-01
1.34252772e-01 1.06553233e+00 3.16313118e-01 -1.31483942e-01
1.71487942e-01 -1.21242955e-01 5.40065877e-02 -6.41919017e-01
-3.17900270e-01 -1.21786088e-01 4.03213389e-02 -7.09687918e-02
-8.98533463e-01 -3.84746939e-01 -1.06910837e+00 1.46843210e-01
-4.46076155e-01 6.86950982e-01 6.56745017e-01 8.70623410e-01
5.26809335e-01 4.62775230e-01 5.95766842e-01 -6.41563177e-01
2.97880560e-01 -1.00340843e+00 -1.45734563e-01 7.89563060e-01
7.74615258e-02 -3.33569616e-01 -2.57224083e-01 5.98274767e-01] | [10.974367141723633, 1.0159038305282593] |
e61156f6-adaf-40cc-84e9-e17f316f340f | viewer-centred-surface-completion-for | 2209.06407 | null | https://arxiv.org/abs/2209.06407v1 | https://arxiv.org/pdf/2209.06407v1.pdf | Viewer-Centred Surface Completion for Unsupervised Domain Adaptation in 3D Object Detection | Every autonomous driving dataset has a different configuration of sensors, originating from distinct geographic regions and covering various scenarios. As a result, 3D detectors tend to overfit the datasets they are trained on. This causes a drastic decrease in accuracy when the detectors are trained on one dataset and tested on another. We observe that lidar scan pattern differences form a large component of this reduction in performance. We address this in our approach, SEE-VCN, by designing a novel viewer-centred surface completion network (VCN) to complete the surfaces of objects of interest within an unsupervised domain adaptation framework, SEE. With SEE-VCN, we obtain a unified representation of objects across datasets, allowing the network to focus on learning geometry, rather than overfitting on scan patterns. By adopting a domain-invariant representation, SEE-VCN can be classed as a multi-target domain adaptation approach where no annotations or re-training is required to obtain 3D detections for new scan patterns. Through extensive experiments, we show that our approach outperforms previous domain adaptation methods in multiple domain adaptation settings. Our code and data are available at https://github.com/darrenjkt/SEE-VCN. | ['Stewart Worrall', 'Eduardo Nebot', 'Mao Shan', 'Julie Stephany Berrio', 'Darren Tsai'] | 2022-09-14 | null | null | null | null | ['multi-target-domain-adaptation'] | ['computer-vision'] | [ 2.53897786e-01 7.43497834e-02 -6.13963231e-02 -6.82785809e-01
-8.52311373e-01 -7.94250786e-01 6.50148809e-01 -5.11713363e-02
-4.38414127e-01 2.93043196e-01 -8.10852125e-02 -1.52016118e-01
1.24674812e-02 -7.21084177e-01 -1.08248293e+00 -3.52488488e-01
3.31717938e-01 8.40857923e-01 5.97190917e-01 -1.27425835e-01
1.60103261e-01 7.48949349e-01 -1.60117757e+00 -2.12102950e-01
8.09670448e-01 7.20900595e-01 3.12069088e-01 4.38106000e-01
-3.47400829e-02 1.74566675e-02 -2.73692280e-01 -1.14035115e-01
7.14025557e-01 2.49729957e-02 -3.90953809e-01 1.77359313e-01
8.34319234e-01 -2.21392855e-01 -2.94873536e-01 9.66453075e-01
3.10264170e-01 2.09370613e-01 9.58742857e-01 -1.34194589e+00
-2.73776114e-01 8.87978598e-02 -6.97385728e-01 -1.55275976e-02
4.94707301e-02 1.75059780e-01 8.37527931e-01 -1.25689518e+00
7.24720180e-01 1.23359394e+00 9.78183925e-01 5.20462990e-01
-1.56312013e+00 -7.00296700e-01 3.67118001e-01 5.82364947e-02
-1.62139761e+00 -4.34447527e-01 9.88407433e-01 -6.49182022e-01
6.58546150e-01 -1.17990606e-01 3.32280874e-01 1.13966358e+00
4.04441766e-02 7.09503770e-01 7.31817007e-01 -1.18437335e-01
3.52707028e-01 2.29762569e-01 -8.47679079e-02 2.26349756e-01
6.48655534e-01 2.70434082e-01 -3.71912718e-01 4.95525524e-02
7.21108556e-01 1.42133042e-01 1.35543510e-01 -1.28030610e+00
-8.50783467e-01 8.73142540e-01 5.24427950e-01 -1.94613993e-01
-2.54613280e-01 -6.24974743e-02 2.46898785e-01 2.98105299e-01
5.17697155e-01 3.70417207e-01 -6.23702705e-01 1.68867022e-01
-6.75223291e-01 5.15616775e-01 5.00767291e-01 1.26221073e+00
1.24630821e+00 -8.20674822e-02 2.97418267e-01 8.54306042e-01
3.25969219e-01 9.17242825e-01 6.15602620e-02 -1.00553977e+00
5.21345198e-01 7.66252518e-01 1.48850664e-01 -6.44410312e-01
-6.28766596e-01 -3.60391915e-01 -5.04328847e-01 4.59739834e-01
4.12571788e-01 -2.06180915e-01 -1.15991390e+00 1.72185373e+00
4.40503448e-01 -3.92581569e-03 1.11587502e-01 9.01265800e-01
6.64369941e-01 2.51580417e-01 5.35494052e-02 5.65355539e-01
9.97650087e-01 -7.15267479e-01 -9.61258933e-02 -8.69509101e-01
6.49188340e-01 -5.79433620e-01 1.04378664e+00 3.02700371e-01
-5.87743521e-01 -7.34993517e-01 -1.22580218e+00 -9.05048922e-02
-5.85983813e-01 1.67951304e-02 2.84683049e-01 4.36592400e-01
-9.22920465e-01 2.85614401e-01 -8.19943964e-01 -7.43396103e-01
5.75392008e-01 2.24654928e-01 -3.65562528e-01 -3.35924864e-01
-7.00451553e-01 1.01762295e+00 4.07507747e-01 -2.75007308e-01
-9.43839669e-01 -8.22723866e-01 -1.00768650e+00 -4.47890669e-01
5.80459654e-01 -5.36620677e-01 1.43916643e+00 -6.82511508e-01
-1.02821636e+00 9.63635206e-01 -2.30839014e-01 -3.35545212e-01
5.38219154e-01 -3.44983906e-01 -5.09888470e-01 -1.89126804e-01
4.50563997e-01 8.18845391e-01 8.76069188e-01 -1.66199517e+00
-6.82102084e-01 -6.06216133e-01 -2.45207280e-01 2.92978764e-01
1.94176689e-01 -4.91297662e-01 -5.84283590e-01 -2.96302676e-01
4.26591218e-01 -1.16445959e+00 -2.77448028e-01 2.41101012e-01
-2.16533363e-01 -1.25440896e-01 1.02943659e+00 -1.91502094e-01
5.37409544e-01 -2.34384012e+00 1.66480429e-02 2.47442856e-01
2.86385298e-01 1.65223077e-01 -3.75132918e-01 4.51644272e-01
7.71939941e-03 -1.55535489e-02 -6.35166883e-01 -3.84128869e-01
-8.99779797e-02 4.49975818e-01 -1.69173911e-01 5.98902583e-01
6.73547447e-01 6.48669779e-01 -7.88766563e-01 -1.65450618e-01
4.13956195e-01 3.20304722e-01 -6.05986178e-01 8.69390965e-02
-3.33697468e-01 6.99079454e-01 -5.73896527e-01 6.08714521e-01
1.22513676e+00 -1.19346961e-01 -6.30771741e-02 -1.61857847e-02
-1.66978046e-01 1.23325236e-01 -1.39551008e+00 1.96061170e+00
-3.29408854e-01 5.50295353e-01 3.23661342e-02 -1.05520868e+00
1.29222929e+00 -1.74796104e-01 4.65285510e-01 -8.09606910e-01
-5.91124687e-03 3.63622248e-01 -1.66645378e-01 -3.44583839e-01
4.79125410e-01 -1.16227269e-01 -2.21423402e-01 1.57206103e-01
1.09614141e-01 -4.75128829e-01 -5.14170974e-02 1.09259233e-01
1.09737349e+00 4.19110239e-01 2.04742655e-01 -1.31242335e-01
2.78578579e-01 5.28898180e-01 6.91358745e-01 8.93455803e-01
-2.88647085e-01 7.40855038e-01 5.31616844e-02 -4.48318601e-01
-1.35575247e+00 -1.47957063e+00 -4.56956357e-01 9.67633426e-01
3.58437300e-01 -2.34495346e-02 -4.13311630e-01 -6.07262313e-01
5.46094000e-01 8.00575495e-01 -5.29413700e-01 -2.88209856e-01
-5.37870586e-01 -2.90040672e-01 4.77409452e-01 6.46537423e-01
3.77802014e-01 -7.36037672e-01 -6.69190228e-01 8.41110945e-02
1.83489680e-01 -1.21065223e+00 -2.82621533e-01 5.17696917e-01
-9.80427623e-01 -1.17543006e+00 -5.10479033e-01 -6.26059771e-01
5.65768898e-01 6.01542711e-01 1.14948833e+00 -3.25872809e-01
-1.61383897e-01 5.96492708e-01 -3.65215898e-01 -9.00561094e-01
-3.76315534e-01 3.98536623e-01 2.47265354e-01 -6.44381121e-02
7.91757762e-01 -7.00443566e-01 -4.65766132e-01 4.60669518e-01
-7.80516684e-01 -1.32877216e-01 7.11433232e-01 4.75133538e-01
7.41916835e-01 -3.63692343e-01 3.07743043e-01 -9.42353189e-01
2.33013287e-01 -6.86047733e-01 -8.37508738e-01 -1.68479696e-01
-6.46604300e-01 2.42996380e-01 4.34996873e-01 -3.63718390e-01
-1.05475318e+00 5.85561275e-01 3.29518951e-02 -7.37306476e-01
-5.83522141e-01 1.81025550e-01 -1.53747201e-01 -1.65561378e-01
1.09321737e+00 5.61918393e-02 1.53989822e-01 -7.36694515e-01
4.74768877e-01 4.79819596e-01 5.98399758e-01 -4.64416206e-01
1.38533115e+00 6.51507318e-01 4.04320732e-02 -8.01801026e-01
-8.62922847e-01 -8.69888246e-01 -1.04581332e+00 -1.34700269e-01
8.33683312e-01 -1.25382245e+00 3.70750353e-02 3.84681553e-01
-1.06212986e+00 -4.29748595e-01 -4.40542668e-01 4.74969417e-01
-4.87389475e-01 6.67703897e-02 1.96090937e-01 -7.30282426e-01
2.09238410e-01 -9.14441049e-01 1.35114229e+00 1.00190245e-01
-7.70620853e-02 -9.74025130e-01 3.89271259e-01 3.22230905e-02
2.08577991e-01 3.10072988e-01 5.26759744e-01 -7.62838721e-01
-5.32600641e-01 -2.65728205e-01 -3.71917635e-01 3.10489148e-01
1.24444954e-01 -2.06765965e-01 -1.14631844e+00 -2.89365590e-01
-2.25201815e-01 -2.47408211e-01 9.74423170e-01 3.88199657e-01
8.40224385e-01 2.99372345e-01 -5.93379319e-01 7.29349315e-01
1.59841204e+00 5.50696403e-02 3.45261037e-01 4.28281188e-01
7.26284504e-01 5.97386301e-01 7.93068588e-01 3.04736108e-01
5.61652601e-01 6.81620300e-01 6.09433770e-01 -2.92201430e-01
-1.39860824e-01 -5.11009753e-01 2.12742865e-01 1.39009699e-01
1.89492539e-01 -1.19503783e-02 -1.27444243e+00 7.71223128e-01
-1.87331164e+00 -5.96840203e-01 -2.01659814e-01 2.22724390e+00
2.53954887e-01 1.92306221e-01 2.51926392e-01 -2.69241452e-01
7.06276298e-01 -3.43613960e-02 -1.23448193e+00 -1.39127910e-01
-1.36949077e-01 1.89380348e-01 1.10848892e+00 3.95481855e-01
-1.15579212e+00 1.03812277e+00 6.11786318e+00 3.99446338e-01
-9.58879530e-01 1.31135970e-01 8.89305100e-02 -2.97747366e-02
-2.07909063e-01 3.34102707e-03 -8.90790045e-01 7.22287223e-03
7.44434714e-01 4.67668548e-02 3.01830113e-01 1.04368496e+00
2.44941056e-01 -8.05986747e-02 -1.17028487e+00 9.50310290e-01
-5.61992712e-02 -1.01193559e+00 -1.43600836e-01 1.17609091e-01
6.06262803e-01 7.74859667e-01 -1.33230776e-01 4.10835952e-01
5.72556973e-01 -6.98967159e-01 8.97488475e-01 3.30455035e-01
8.41270745e-01 -4.37028646e-01 3.83529037e-01 4.34259504e-01
-1.31553173e+00 -5.55653311e-02 -6.39480054e-01 -9.70001668e-02
5.52575141e-02 4.76201713e-01 -1.00707829e+00 6.21839225e-01
8.80795479e-01 8.10552120e-01 -7.11531639e-01 1.00648665e+00
-1.74740449e-01 3.54037702e-01 -5.57192206e-01 2.68708199e-01
7.21485540e-02 -1.45480290e-01 7.06287384e-01 9.91678774e-01
3.94155264e-01 -2.61772990e-01 3.68260294e-01 1.03951061e+00
2.92491019e-02 -1.61663264e-01 -1.08463120e+00 4.06850398e-01
6.84346616e-01 1.15888119e+00 -4.63623494e-01 -1.45219624e-01
-5.07405102e-01 6.78737104e-01 3.97424906e-01 3.93165618e-01
-7.42146492e-01 -1.99530557e-01 9.56214130e-01 3.65725696e-01
5.30704379e-01 -4.69508976e-01 -5.08138537e-01 -9.95918214e-01
1.78838745e-01 -4.81547892e-01 2.18834028e-01 -6.89056695e-01
-1.35566115e+00 2.79983789e-01 2.95848221e-01 -1.56223404e+00
-4.73954789e-02 -7.51901746e-01 -3.50999355e-01 7.42384672e-01
-1.78101718e+00 -1.07883894e+00 -6.18763804e-01 5.83385527e-01
5.90770006e-01 -9.10785645e-02 4.05281156e-01 3.12682986e-01
-3.11745852e-01 3.36529851e-01 1.24870591e-01 -5.11125512e-02
1.02981091e+00 -1.15533531e+00 7.64585376e-01 8.70348990e-01
3.11302431e-02 1.07946776e-01 6.25639200e-01 -6.89856887e-01
-1.33927095e+00 -1.51262116e+00 4.66665059e-01 -9.01254714e-01
4.01349008e-01 -5.64668655e-01 -1.04965520e+00 8.24755251e-01
-2.65870214e-01 -1.42395249e-04 4.00701821e-01 1.94316223e-01
-6.66908622e-01 -2.58151293e-01 -1.13174951e+00 3.94490182e-01
1.49387050e+00 -3.38367552e-01 -5.26703775e-01 1.53549537e-01
7.60542274e-01 -4.53791976e-01 -6.34077489e-01 5.08319438e-01
2.74379373e-01 -8.56555641e-01 8.73038471e-01 -3.94703954e-01
1.76169515e-01 -5.49002290e-01 -3.39922488e-01 -1.42257404e+00
-4.73384619e-01 -9.01590437e-02 2.12010220e-01 8.95054758e-01
5.99971652e-01 -9.08828437e-01 8.34157646e-01 3.74192268e-01
-4.49832350e-01 -1.68901250e-01 -9.75481927e-01 -1.13110316e+00
4.13508415e-01 -8.46633375e-01 7.82593906e-01 7.45550513e-01
-6.10790014e-01 3.98065060e-01 -3.60559253e-03 5.79081953e-01
8.18759322e-01 -9.84615237e-02 1.30577803e+00 -1.77136528e+00
8.77900720e-02 -2.86639184e-01 -4.27728802e-01 -1.24984705e+00
-9.53760371e-02 -9.59613264e-01 3.51574093e-01 -1.36192954e+00
-1.98151562e-02 -6.98031306e-01 -2.49735964e-03 6.05115473e-01
1.82156265e-01 1.15295365e-01 7.91491866e-02 4.15762067e-01
-3.78375053e-01 5.97574115e-01 1.08222330e+00 -1.90332338e-01
-3.27886283e-01 2.82963887e-02 -6.69119835e-01 7.18197405e-01
9.94076073e-01 -6.57489121e-01 -3.63024473e-01 -7.00469851e-01
6.32125884e-02 -3.92527878e-01 5.66120744e-01 -1.33843744e+00
1.49328023e-01 -1.61156878e-01 4.47701722e-01 -8.17890882e-01
4.45523083e-01 -1.04649436e+00 1.08458668e-01 1.77450597e-01
2.08454542e-02 -8.12498778e-02 5.02108753e-01 6.77367151e-01
3.09078787e-02 -1.15879059e-01 8.39838862e-01 -5.11716567e-02
-1.11649072e+00 3.94621581e-01 -2.64369577e-01 1.66376352e-01
1.01026154e+00 -6.17623270e-01 -2.56442521e-02 -1.04596384e-01
-5.30672848e-01 4.02046412e-01 8.43744636e-01 7.59566188e-01
5.28670132e-01 -1.31194913e+00 -6.98697805e-01 4.79146779e-01
7.16971159e-01 6.93211496e-01 1.53929785e-01 5.97193122e-01
-1.77685052e-01 3.51506263e-01 -9.65043232e-02 -1.07876956e+00
-8.03260982e-01 4.14657205e-01 4.20995146e-01 8.74258857e-03
-7.52360880e-01 6.05548978e-01 4.45056975e-01 -1.08855700e+00
4.05961927e-03 -3.41230184e-01 9.69164893e-02 -1.03627153e-01
-7.15927235e-05 1.00883506e-01 1.59003019e-01 -6.25739157e-01
-5.40335298e-01 8.88216257e-01 -2.21690126e-02 4.35133502e-02
1.31542993e+00 -9.51026306e-02 5.14452398e-01 5.14866948e-01
1.02865982e+00 -1.19672380e-01 -1.80764389e+00 -5.02783775e-01
6.03133701e-02 -3.40332031e-01 -1.84741970e-02 -6.50685191e-01
-8.73797894e-01 6.53160989e-01 8.01780879e-01 -9.89543051e-02
9.36905861e-01 3.60639870e-01 3.92921984e-01 4.93824363e-01
4.23810750e-01 -1.22550046e+00 -1.05378993e-01 6.75893486e-01
8.52882087e-01 -1.66790783e+00 -5.56919575e-02 -3.79148185e-01
-7.06305981e-01 8.57448518e-01 7.85875738e-01 -3.66813540e-01
7.75518239e-01 1.06952123e-01 2.37332970e-01 -4.34126288e-01
-2.91553110e-01 -4.86771077e-01 3.00813347e-01 1.04852223e+00
-9.06793550e-02 1.75575659e-01 2.96122402e-01 3.04146528e-01
-2.05632851e-01 -1.67221919e-01 3.60997349e-01 9.87881958e-01
-6.08574927e-01 -1.17049313e+00 -5.09126425e-01 4.35081244e-01
3.98302466e-01 2.65334517e-01 -5.74526787e-01 1.11608720e+00
2.19059035e-01 8.06127429e-01 3.41345757e-01 -5.74555695e-01
8.06274831e-01 2.01379597e-01 3.17870021e-01 -8.57811272e-01
1.14142634e-01 -1.30956411e-01 -7.24127665e-02 -5.69801688e-01
-3.13972622e-01 -1.00940764e+00 -1.22347736e+00 -5.55740595e-02
-4.91326191e-02 -2.33381867e-01 7.74219573e-01 7.44005501e-01
6.57845020e-01 3.48093092e-01 6.16450071e-01 -1.16450346e+00
-5.35207212e-01 -9.36856985e-01 -5.42933226e-01 5.00087559e-01
4.68681574e-01 -1.14163613e+00 -3.18807811e-01 -8.36098939e-02] | [8.146768569946289, -2.6696958541870117] |
5f746465-a0b5-4e29-a8e6-38e9a9265b82 | ffhq-uv-normalized-facial-uv-texture-dataset | 2211.13874 | null | https://arxiv.org/abs/2211.13874v2 | https://arxiv.org/pdf/2211.13874v2.pdf | FFHQ-UV: Normalized Facial UV-Texture Dataset for 3D Face Reconstruction | We present a large-scale facial UV-texture dataset that contains over 50,000 high-quality texture UV-maps with even illuminations, neutral expressions, and cleaned facial regions, which are desired characteristics for rendering realistic 3D face models under different lighting conditions. The dataset is derived from a large-scale face image dataset namely FFHQ, with the help of our fully automatic and robust UV-texture production pipeline. Our pipeline utilizes the recent advances in StyleGAN-based facial image editing approaches to generate multi-view normalized face images from single-image inputs. An elaborated UV-texture extraction, correction, and completion procedure is then applied to produce high-quality UV-maps from the normalized face images. Compared with existing UV-texture datasets, our dataset has more diverse and higher-quality texture maps. We further train a GAN-based texture decoder as the nonlinear texture basis for parametric fitting based 3D face reconstruction. Experiments show that our method improves the reconstruction accuracy over state-of-the-art approaches, and more importantly, produces high-quality texture maps that are ready for realistic renderings. The dataset, code, and pre-trained texture decoder are publicly available at https://github.com/csbhr/FFHQ-UV. | ['Linchao Bao', 'Jinshan Pan', 'Haoxian Zhang', 'Di Kang', 'Haoran Bai'] | 2022-11-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Bai_FFHQ-UV_Normalized_Facial_UV-Texture_Dataset_for_3D_Face_Reconstruction_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Bai_FFHQ-UV_Normalized_Facial_UV-Texture_Dataset_for_3D_Face_Reconstruction_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 3.37370217e-01 1.35324374e-01 2.53742278e-01 -6.82793796e-01
-1.02060437e+00 -2.28850693e-01 6.34003341e-01 -8.92825186e-01
3.40942323e-01 2.50339538e-01 9.37680975e-02 7.70124272e-02
3.22256207e-01 -9.02121127e-01 -9.18879747e-01 -7.51919508e-01
4.61937338e-01 5.93850195e-01 -1.88675880e-01 -4.11304533e-01
-1.41148090e-01 4.63282615e-01 -1.80419779e+00 5.64278483e-01
3.90575171e-01 1.34814227e+00 -6.36523440e-02 4.11144912e-01
-6.38236701e-02 1.89326033e-01 -1.52871504e-01 -8.43014121e-01
3.86438787e-01 -4.67025369e-01 -3.76255780e-01 2.04847232e-01
9.56112146e-01 -7.11350918e-01 -1.20569281e-01 6.53712213e-01
6.74807310e-01 -2.04364389e-01 5.95604360e-01 -9.32119548e-01
-8.65679145e-01 -2.44299889e-01 -6.83663964e-01 -5.43529570e-01
3.42799574e-01 1.39927268e-01 3.14349890e-01 -1.20551670e+00
1.03022110e+00 1.60272324e+00 4.84975010e-01 9.95259702e-01
-1.37525678e+00 -7.06116438e-01 -4.86651540e-01 -2.73992449e-01
-1.23197973e+00 -1.05312538e+00 7.53284097e-01 -2.94555902e-01
5.91223240e-01 3.19532633e-01 8.27710092e-01 1.58714128e+00
2.03393012e-01 2.57520407e-01 1.53159440e+00 -3.20594728e-01
-4.16132249e-02 -1.85010910e-01 -8.80427897e-01 1.16630125e+00
-3.63442630e-01 2.66302913e-01 -7.66951680e-01 -1.91423982e-01
1.27520370e+00 -2.14465037e-01 -8.85152519e-02 -1.72112763e-01
-9.32410181e-01 3.81934315e-01 -1.17272988e-01 -3.10819864e-01
-2.08892882e-01 1.12214766e-01 1.58937082e-01 3.24347556e-01
1.12927222e+00 -3.11257690e-01 -3.26879412e-01 -2.11625323e-01
-6.74629629e-01 2.79302597e-01 4.46579129e-01 1.02781868e+00
9.52163756e-01 2.45198652e-01 -2.25367546e-01 1.29903460e+00
2.35576868e-01 1.11528289e+00 -2.31783524e-01 -1.14503658e+00
7.32678995e-02 2.10546806e-01 -2.01337144e-01 -9.21333373e-01
5.82087152e-02 1.27257362e-01 -6.16293848e-01 4.19291586e-01
2.47530416e-01 2.54742891e-01 -1.16351628e+00 1.77794766e+00
8.57250094e-01 1.18446290e-01 -3.62994552e-01 9.79306936e-01
1.15748823e+00 4.92958516e-01 -3.70897055e-01 4.19995561e-02
1.42113674e+00 -5.95399499e-01 -8.35413933e-01 1.88667178e-01
1.40440419e-01 -1.19642377e+00 1.38343894e+00 4.56806272e-01
-1.14166069e+00 -3.55825424e-01 -6.54341936e-01 -4.50709522e-01
5.62728159e-02 2.78187305e-01 6.80035472e-01 8.60278547e-01
-1.32322741e+00 3.42705071e-01 -6.92421913e-01 -2.66678363e-01
9.26172018e-01 1.99997380e-01 -8.08256865e-01 -3.12122792e-01
-5.57731450e-01 5.24909437e-01 -3.63139391e-01 3.14264774e-01
-1.21062696e+00 -8.48697186e-01 -1.03233492e+00 -4.87569779e-01
1.81871980e-01 -6.92742825e-01 1.17860889e+00 -1.01366413e+00
-2.10426021e+00 1.24818051e+00 -4.85309273e-01 5.35742223e-01
6.72830343e-01 1.08128972e-01 -3.03872466e-01 1.62875205e-01
-2.39044279e-02 8.16558719e-01 1.23624647e+00 -1.32469332e+00
1.05485640e-01 -4.32354957e-01 -4.77722466e-01 -2.52296124e-02
-7.78819695e-02 3.30713928e-01 -9.42572236e-01 -6.79241955e-01
-3.27001996e-02 -7.48707175e-01 3.78016621e-01 5.41132152e-01
-2.30105266e-01 1.84644267e-01 9.30706918e-01 -8.65428627e-01
3.82532924e-01 -2.16961765e+00 2.17874367e-02 3.63258004e-01
1.66804880e-01 -4.47340980e-02 -4.80556309e-01 1.92119807e-01
-4.58224937e-02 -3.38673443e-02 -9.19454023e-02 -8.83922219e-01
1.46733165e-01 9.02396515e-02 -2.36584798e-01 5.22094190e-01
3.70766789e-01 9.85208869e-01 -3.72580975e-01 -4.33554888e-01
3.54043782e-01 1.06173825e+00 -7.22889483e-01 3.59329015e-01
-4.32004660e-01 9.59129691e-01 -2.07824588e-01 1.06840920e+00
1.12107658e+00 1.12286665e-01 9.10872668e-02 -4.27309662e-01
2.85627663e-01 -1.67700604e-01 -5.97708642e-01 1.89657748e+00
-5.31469703e-01 5.22575617e-01 3.37695986e-01 -3.09699535e-01
1.23713219e+00 3.21524262e-01 5.36159039e-01 -1.16054523e+00
4.61144835e-01 3.80862296e-01 -6.14163637e-01 -3.74659091e-01
2.33358979e-01 -2.16329232e-01 2.66019076e-01 5.73186934e-01
2.80502915e-01 -6.19130790e-01 -6.55403957e-02 -1.42823607e-01
4.44208235e-01 7.74164677e-01 -5.17379045e-01 -2.79770702e-01
2.85741597e-01 -5.87853909e-01 6.06943607e-01 7.80182565e-03
2.18214810e-01 1.10312474e+00 5.89301765e-01 -6.38297856e-01
-1.46297061e+00 -1.10640299e+00 -4.82908100e-01 9.15803909e-01
-2.49955788e-01 -5.83355665e-01 -1.22123218e+00 -2.67677158e-01
-8.51915628e-02 8.85123946e-03 -1.16737044e+00 1.99258700e-01
-3.73430192e-01 -5.60924292e-01 5.38279831e-01 2.24563610e-02
5.98686278e-01 -8.50271583e-01 1.36038601e-01 -2.41117775e-01
-2.70770758e-01 -1.18835592e+00 -6.06139183e-01 -5.35759270e-01
-4.64486867e-01 -1.05024612e+00 -6.83255553e-01 -6.81196153e-01
9.54012990e-01 -6.94913939e-02 1.11878788e+00 2.08285287e-01
-5.00667512e-01 4.21595782e-01 -6.52795061e-02 -4.55213219e-01
-4.54389870e-01 -5.41652799e-01 2.96225026e-02 3.70359361e-01
-1.06191084e-01 -5.18331230e-01 -6.53173745e-01 6.56317949e-01
-9.80378628e-01 4.49726313e-01 1.01426564e-01 8.62964630e-01
1.06574070e+00 -4.43624347e-01 2.46918604e-01 -9.06453669e-01
2.91329503e-01 -2.14792509e-02 -7.95232892e-01 2.10176528e-01
-1.13836527e-01 -1.84175745e-01 3.62854630e-01 -2.42863595e-01
-1.52388370e+00 -1.67669386e-01 -5.49223304e-01 -5.17215133e-01
-1.51317060e-01 -1.81419045e-01 -4.71433491e-01 -5.27416766e-01
5.92286587e-01 2.19466597e-01 3.69128853e-01 -4.71027464e-01
3.40158045e-01 5.70149064e-01 3.73884916e-01 -9.86351728e-01
7.20504284e-01 7.63390899e-01 1.01524293e-01 -8.31383646e-01
-5.75621843e-01 4.30809021e-01 -5.97821474e-01 -5.57364464e-01
6.39461040e-01 -9.45948899e-01 -7.85758674e-01 8.46638620e-01
-8.32944155e-01 -9.82739329e-01 7.89855868e-02 -6.50811270e-02
-7.73289859e-01 1.59735233e-02 -6.44171059e-01 -5.91705084e-01
-4.82802391e-01 -1.44450974e+00 1.85767591e+00 3.94435264e-02
2.02456251e-01 -7.42174983e-01 1.68567337e-02 6.63996756e-01
5.10966659e-01 5.45380712e-01 7.60558128e-01 5.99538326e-01
-5.86812973e-01 2.90472031e-01 -1.85852468e-01 2.29977891e-01
2.32325077e-01 4.53302830e-01 -1.32665074e+00 -2.41206840e-01
-2.03471363e-01 -6.65108979e-01 5.97516418e-01 2.93381870e-01
1.45063496e+00 -2.61575282e-01 9.32447761e-02 1.25339186e+00
1.13415086e+00 -2.57306069e-01 9.10018802e-01 -8.31942037e-02
8.43068242e-01 6.78484619e-01 5.43605685e-01 4.20723796e-01
4.51495796e-01 9.41059470e-01 3.27892333e-01 -5.15591860e-01
-4.37334478e-01 -3.09857666e-01 3.64542991e-01 7.02110529e-01
-3.70972365e-01 9.71432030e-02 -5.62224329e-01 5.63202985e-02
-1.31134200e+00 -7.44427383e-01 7.05220997e-02 2.01229978e+00
9.65700746e-01 -5.00124753e-01 1.65070426e-02 -2.78051794e-01
4.08509493e-01 9.93895829e-02 -3.38456422e-01 -5.12476325e-01
-3.64475459e-01 7.42946625e-01 4.76371720e-02 4.39513326e-01
-8.81235898e-01 1.33080161e+00 6.05257177e+00 1.10980368e+00
-1.49807239e+00 3.36174637e-01 1.13033938e+00 -3.47076654e-01
-7.00408280e-01 -3.44253421e-01 -4.29102212e-01 2.81140655e-01
7.49305189e-01 4.35477555e-01 7.77160525e-01 5.44353247e-01
3.21290791e-01 -5.62622920e-02 -6.70512617e-01 1.23847008e+00
2.04699665e-01 -1.35857058e+00 1.50185034e-01 3.76791149e-01
9.18728530e-01 -1.64706744e-02 5.47220409e-01 -2.62018770e-01
2.40377411e-01 -1.29535711e+00 9.79529917e-01 6.20370924e-01
1.95120645e+00 -8.06735635e-01 2.75652438e-01 -3.47024322e-01
-1.07402039e+00 4.92211968e-01 -5.91978669e-01 5.39712191e-01
6.78672343e-02 7.96781719e-01 -5.03166139e-01 6.60659194e-01
1.00454235e+00 6.74471796e-01 -5.35872698e-01 2.00244829e-01
-3.13826740e-01 4.34062749e-01 -3.96616191e-01 4.41398710e-01
-4.18974310e-01 -5.97330749e-01 -1.52535923e-02 7.88154244e-01
5.32721698e-01 2.10989162e-01 -2.94300586e-01 1.11120808e+00
-1.93181351e-01 2.27489501e-01 -6.61434174e-01 1.03017278e-01
1.81501463e-01 1.67841041e+00 -5.43182373e-01 -1.34021223e-01
-3.15741301e-01 1.02434456e+00 3.38583887e-01 4.03270245e-01
-9.15709078e-01 2.58909553e-01 9.92035329e-01 2.18525410e-01
8.54188353e-02 -1.87846079e-01 -1.12374507e-01 -1.23808396e+00
1.49096876e-01 -1.00634539e+00 -1.81267589e-01 -1.10379410e+00
-1.15409946e+00 8.47365499e-01 -3.17943841e-01 -9.29325819e-01
6.26944762e-04 -7.09973991e-01 -3.01234066e-01 9.58477855e-01
-1.45820034e+00 -1.94711089e+00 -6.47802353e-01 9.00246620e-01
1.57016411e-01 -2.44124845e-01 1.15744197e+00 4.45781648e-01
-6.44098222e-01 8.73674512e-01 -6.97646886e-02 -1.38406418e-02
9.73769188e-01 -6.00037396e-01 6.63780749e-01 4.94548827e-01
-1.71838894e-01 4.49449986e-01 2.43502587e-01 -7.37445295e-01
-1.91783297e+00 -1.25690663e+00 2.12743893e-01 -6.57793283e-01
6.03349283e-02 -1.03528869e+00 -6.46840930e-01 6.39806926e-01
8.49652961e-02 3.21132272e-01 6.49179041e-01 -2.17243835e-01
-5.44522762e-01 -2.96129048e-01 -1.43858898e+00 7.08719134e-01
1.37939775e+00 -7.00313926e-01 2.05682456e-01 3.49268675e-01
4.44584005e-02 -9.01761353e-01 -1.14056253e+00 5.10927141e-01
9.98142898e-01 -1.16490734e+00 8.23927701e-01 -1.51607115e-02
7.06419706e-01 -1.81023240e-01 -4.49200273e-02 -1.21433914e+00
-5.51581420e-02 -8.25329065e-01 3.73752803e-01 1.28648877e+00
1.46219894e-01 -6.78922594e-01 7.03368247e-01 3.69964957e-01
-7.64493421e-02 -7.83965349e-01 -8.74834299e-01 -3.74049097e-01
-1.22109145e-01 -4.28221285e-01 1.11976004e+00 7.80663550e-01
-5.14591634e-01 -2.60716408e-01 -5.45189619e-01 -3.01728338e-01
8.62597406e-01 8.71898308e-02 1.29000962e+00 -8.74419332e-01
7.32804388e-02 -3.11551839e-01 -1.26296520e-01 -5.70872664e-01
2.86896706e-01 -8.82516623e-01 -3.02445274e-02 -1.14175010e+00
1.80696771e-01 -5.82645178e-01 5.31646311e-01 7.31858611e-01
2.32005477e-01 1.04198277e+00 -2.04989940e-01 3.19504403e-02
-1.19621299e-01 9.37559366e-01 1.84320784e+00 2.09665045e-01
2.80584246e-01 -5.10700762e-01 -5.47496378e-01 4.54306275e-01
5.69363832e-01 -1.23580210e-01 -4.08648282e-01 -7.39029765e-01
8.91595855e-02 -3.39812450e-02 4.49294806e-01 -4.78007942e-01
-4.02545214e-01 -2.77602017e-01 8.07506621e-01 -1.94398150e-01
8.81451905e-01 -5.67512929e-01 7.26607442e-01 -2.53228128e-01
1.31768867e-01 -1.11160263e-01 3.61014187e-01 1.08039536e-01
-7.63979694e-03 6.63905025e-01 9.26830530e-01 -8.01531300e-02
-3.00776035e-01 8.71517777e-01 -1.70532404e-03 -2.33671337e-01
6.87892675e-01 -3.20053995e-01 -5.48751831e-01 -4.71526921e-01
-3.91721159e-01 -4.13303405e-01 1.05892575e+00 4.64113861e-01
8.31599355e-01 -1.61713147e+00 -9.76570308e-01 1.04483700e+00
2.87070185e-01 1.02294490e-01 5.30010164e-01 8.26974571e-01
-8.82875085e-01 -1.48500577e-01 -5.35452425e-01 -8.30930948e-01
-1.35028648e+00 -1.56427875e-01 4.74359512e-01 4.41888273e-01
-7.46853650e-01 9.04882371e-01 5.76714873e-01 -8.70527506e-01
-3.53210509e-01 -3.56814079e-02 3.40779692e-01 -4.24899757e-01
6.56932771e-01 -9.38220769e-02 4.82203692e-01 -1.03967786e+00
-3.78957577e-02 1.09061050e+00 3.89929295e-01 -3.87865990e-01
1.65946317e+00 -7.27993324e-02 -5.02631426e-01 -8.52407292e-02
1.23303676e+00 2.09633663e-01 -1.57138586e+00 2.70523392e-02
-7.96781898e-01 -9.93120670e-01 -1.89525057e-02 -6.72104299e-01
-1.54488754e+00 8.60912681e-01 6.15388274e-01 -7.74669111e-01
1.32327271e+00 -4.61669303e-02 7.56821632e-01 -2.17984334e-01
7.16538310e-01 -8.69443119e-01 1.43904164e-01 4.84654278e-01
1.19491005e+00 -9.95344996e-01 -2.19963014e-01 -9.01718736e-01
-4.50377852e-01 1.17877328e+00 6.44760251e-01 1.05781183e-01
6.21679306e-01 6.10000372e-01 4.88610536e-01 -3.80559415e-01
-8.44306588e-01 2.99424708e-01 3.66884679e-01 8.03408742e-01
4.24060434e-01 3.69592533e-02 3.27309459e-01 2.97593385e-01
-6.59248173e-01 2.33808327e-02 1.78765416e-01 5.30397832e-01
2.37144902e-01 -1.24381244e+00 -4.86540705e-01 3.43222618e-01
-4.46783572e-01 -5.66642657e-02 -3.08609456e-01 3.89856040e-01
-1.19621288e-02 6.93822145e-01 2.24877708e-02 -4.46683556e-01
2.05695629e-01 1.03584295e-02 1.16161060e+00 -3.51975113e-01
-2.90200442e-01 5.12604117e-01 1.94251060e-01 -1.08515346e+00
-2.35583633e-01 -5.15133381e-01 -7.30441928e-01 -8.93355429e-01
-2.48816181e-02 -4.23390031e-01 8.79389346e-01 6.66117072e-01
6.36653364e-01 1.74751744e-01 6.92339301e-01 -1.23134339e+00
3.22599828e-01 -9.33274269e-01 -5.93905091e-01 6.93763554e-01
-2.30600368e-02 -9.53824341e-01 -5.19033335e-02 2.45259136e-01] | [12.811223030090332, -0.26517248153686523] |
58867e20-1ba4-47c2-ba75-7d9fa8e6d7cf | retrieval-re-ranking-and-multi-task-learning | null | null | https://aclanthology.org/2021.eacl-main.26 | https://aclanthology.org/2021.eacl-main.26.pdf | Retrieval, Re-ranking and Multi-task Learning for Knowledge-Base Question Answering | Question answering over knowledge bases (KBQA) usually involves three sub-tasks, namely topic entity detection, entity linking and relation detection. Due to the large number of entities and relations inside knowledge bases (KB), previous work usually utilized sophisticated rules to narrow down the search space and managed only a subset of KBs in memory. In this work, we leverage a \textit{retrieve-and-rerank} framework to access KBs via traditional information retrieval (IR) method, and re-rank retrieved candidates with more powerful neural networks such as the pre-trained BERT model. Considering the fact that directly assigning a different BERT model for each sub-task may incur prohibitive costs, we propose to share a BERT encoder across all three sub-tasks and define task-specific layers on top of the shared layer. The unified model is then trained under a multi-task learning framework. Experiments show that: (1) Our IR-based retrieval method is able to collect high-quality candidates efficiently, thus enables our method adapt to large-scale KBs easily; (2) the BERT model improves the accuracy across all three sub-tasks; and (3) benefiting from multi-task learning, the unified model obtains further improvements with only 1/3 of the original parameters. Our final model achieves competitive results on the SimpleQuestions dataset and superior performance on the FreebaseQA dataset. | ['Bing Xiang', 'Ramesh Nallapati', 'Patrick Ng', 'Zhiguo Wang'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-3.61687869e-01 2.05121949e-01 -5.03860652e-01 -1.66618094e-01
-1.42451251e+00 -4.38114643e-01 4.04493570e-01 3.52353990e-01
-7.88083255e-01 9.22632873e-01 4.61330749e-02 -2.95807838e-01
-3.84980321e-01 -9.96092021e-01 -1.03674066e+00 -1.99715853e-01
-1.11195341e-01 9.27235186e-01 9.11934435e-01 -3.26659173e-01
-1.86117247e-01 1.45978794e-01 -1.24659479e+00 5.30918539e-01
1.21873808e+00 1.42678499e+00 2.11145625e-01 4.44208354e-01
-2.85190970e-01 1.02370965e+00 -4.58031654e-01 -7.59342372e-01
-3.77024487e-02 3.25062960e-01 -1.44399405e+00 -6.39087319e-01
4.93360639e-01 -5.67104459e-01 -4.92675513e-01 7.64546514e-01
5.39354324e-01 2.26082459e-01 5.21773815e-01 -8.71371627e-01
-8.90173197e-01 8.44253898e-01 -3.18247676e-01 2.43953452e-01
3.72445256e-01 -4.07056600e-01 1.63898563e+00 -8.86130750e-01
5.59813917e-01 1.20918107e+00 4.12171006e-01 2.12160543e-01
-1.05150199e+00 -7.43293703e-01 4.24092740e-01 6.35397017e-01
-1.73496199e+00 -3.97317946e-01 3.94822627e-01 1.40824332e-03
1.23912954e+00 1.62831694e-01 2.90585428e-01 8.02946031e-01
-2.66809165e-01 1.18554819e+00 4.80854899e-01 -3.48789155e-01
2.87893694e-02 1.10832408e-01 5.66280007e-01 6.23394549e-01
3.83822918e-01 -4.84049231e-01 -4.20359254e-01 -4.59674805e-01
4.37427789e-01 -2.35646814e-01 -3.21802706e-01 -3.76905531e-01
-1.14589715e+00 7.06358910e-01 7.27645934e-01 3.05333614e-01
-3.91956776e-01 1.81346446e-01 4.46137697e-01 3.43189150e-01
4.21947956e-01 7.18533516e-01 -9.32494044e-01 2.53378779e-01
-6.94289327e-01 3.51630479e-01 1.03818691e+00 1.03157413e+00
1.07684982e+00 -6.93500102e-01 -4.11973089e-01 1.07756352e+00
3.19214582e-01 4.81966168e-01 3.81802291e-01 -5.19182265e-01
9.59380329e-01 7.40424156e-01 2.77936161e-01 -9.73373771e-01
-5.09935558e-01 -3.97139043e-01 -5.73728979e-01 -8.49841416e-01
2.04278693e-01 -1.08151399e-02 -9.80000854e-01 1.66381884e+00
4.94208127e-01 8.49394966e-03 1.49305418e-01 8.13177943e-01
1.10389388e+00 7.49086201e-01 3.02932739e-01 -6.86566904e-02
1.74709058e+00 -1.18088984e+00 -5.97157478e-01 -2.55637467e-01
9.32831049e-01 -5.05007327e-01 1.03472769e+00 6.06489368e-02
-8.86537492e-01 -2.81565607e-01 -1.03489125e+00 -5.95034063e-01
-7.00213313e-01 3.16152811e-01 7.41807997e-01 2.12126568e-01
-1.13554358e+00 5.85069843e-02 -6.66803360e-01 -2.00495824e-01
2.87648886e-01 4.74289209e-01 -2.25723416e-01 -2.36692578e-01
-1.83686209e+00 9.19471860e-01 8.78351986e-01 1.75107598e-01
-8.58617365e-01 -6.50973022e-01 -7.40762770e-01 4.27135617e-01
8.75618160e-01 -9.82166946e-01 1.23203838e+00 -3.49134207e-01
-1.37478089e+00 4.75631446e-01 -1.70188695e-01 -3.42306346e-01
8.71906206e-02 -6.96141064e-01 -4.42968369e-01 2.34079957e-01
2.99062133e-01 5.17427504e-01 4.64413911e-01 -1.08112717e+00
-7.44773209e-01 -1.57829896e-01 4.14267808e-01 3.81657809e-01
-5.95541775e-01 -5.46590984e-02 -1.16940558e+00 -4.47260797e-01
-6.28881007e-02 -7.00389683e-01 -2.63045193e-03 -4.91238117e-01
-6.08806074e-01 -8.39928806e-01 4.15493637e-01 -7.10983992e-01
1.62204516e+00 -1.71012318e+00 8.97767693e-02 2.27302387e-01
2.47762680e-01 5.85460663e-01 -2.60241270e-01 3.67441624e-01
2.26352245e-01 1.02828510e-01 1.99186131e-01 -1.00589737e-01
-2.24940721e-02 5.52107394e-02 -4.65204239e-01 -1.42252669e-01
1.77293003e-01 1.31405759e+00 -9.89238322e-01 -7.75137663e-01
-4.72195506e-01 2.00547278e-01 -3.77754420e-01 1.61801904e-01
-5.61138928e-01 -2.04231203e-01 -1.13056457e+00 7.22176731e-01
5.10671675e-01 -7.88392067e-01 2.76000649e-01 -4.05901849e-01
4.05134529e-01 7.53807724e-01 -1.23338318e+00 1.67627037e+00
-5.00595033e-01 9.80682150e-02 -2.99188420e-02 -1.04450834e+00
6.28680289e-01 4.71028090e-01 2.41438359e-01 -9.63041365e-01
-3.70163828e-01 5.75469673e-01 -3.54787618e-01 -5.66627026e-01
6.28846884e-01 3.17837864e-01 -1.73857406e-01 4.97171789e-01
3.43966395e-01 3.48576486e-01 2.97356933e-01 5.66106975e-01
1.27554929e+00 -7.59785473e-02 2.50601828e-01 3.88983190e-02
6.19938374e-01 9.34385136e-03 4.47053552e-01 8.91293049e-01
1.86089173e-01 7.27761164e-02 3.92554134e-01 -4.52613533e-01
-5.62960148e-01 -9.16573346e-01 -1.55044913e-01 1.74335992e+00
2.80132204e-01 -4.34626281e-01 -2.21126065e-01 -1.04814982e+00
3.36409539e-01 4.20953721e-01 -4.46939290e-01 -2.63015628e-01
-7.75075614e-01 -8.61802936e-01 7.23629713e-01 4.38712269e-01
5.15372872e-01 -9.53877151e-01 1.02521740e-02 3.61855447e-01
-5.68438113e-01 -1.19369113e+00 -3.47618461e-01 1.85341522e-01
-6.91762090e-01 -1.10230148e+00 -7.20245242e-01 -7.31689215e-01
3.64631653e-01 2.92958021e-01 1.42857552e+00 5.13869859e-02
4.25699130e-02 3.06209922e-01 -5.26077569e-01 -1.93465188e-01
7.35054836e-02 8.48006368e-01 -2.01176777e-01 -5.47583364e-02
5.27036190e-01 -2.56237566e-01 -5.81919193e-01 2.65224099e-01
-8.27833772e-01 -1.94090590e-01 7.62275040e-01 8.87380958e-01
4.77605283e-01 -7.14907795e-02 1.07653594e+00 -8.60503435e-01
6.74657941e-01 -4.78727847e-01 -6.30663633e-01 9.46440458e-01
-8.13320696e-01 2.72158027e-01 3.80915463e-01 -2.83152491e-01
-1.25798035e+00 -2.79669017e-01 9.80436951e-02 -1.11128271e-01
3.47114891e-01 9.88133192e-01 -2.68624783e-01 1.31603092e-01
6.13144100e-01 2.03505140e-02 -6.15434885e-01 -6.63817167e-01
6.98214173e-01 6.69629276e-01 2.92268991e-01 -8.66963744e-01
7.89926052e-01 1.99571162e-01 -4.94787544e-01 -1.68709010e-01
-1.43937051e+00 -8.79957139e-01 -5.26729465e-01 2.04252526e-01
5.62911153e-01 -1.46079469e+00 -8.14558744e-01 9.26139951e-02
-1.21147871e+00 -1.36050701e-01 9.14407242e-03 4.71193671e-01
-4.93472889e-02 3.02332908e-01 -8.42782915e-01 -6.54753864e-01
-7.63360202e-01 -7.91184008e-01 1.24463367e+00 1.65786073e-01
2.53954232e-01 -8.35494041e-01 1.65149003e-01 7.22314298e-01
4.24205720e-01 -5.34688771e-01 1.24119473e+00 -1.01526284e+00
-9.53763366e-01 -2.58699358e-01 -6.54307485e-01 1.08634844e-01
-2.41371840e-02 -5.17341971e-01 -9.45956469e-01 -2.16896355e-01
-5.38160622e-01 -9.72036421e-01 1.42897999e+00 -3.16906385e-02
1.05985177e+00 -4.08427596e-01 -7.88107693e-01 3.22268337e-01
1.17677081e+00 -6.41562641e-02 5.18610835e-01 6.44979775e-01
8.73005271e-01 4.51769382e-01 7.78788865e-01 3.43138985e-02
1.05841839e+00 8.83176744e-01 1.50037423e-01 -7.26920739e-02
5.96016124e-02 -3.04381490e-01 2.72386849e-01 8.26093912e-01
4.92037646e-02 -2.57846177e-01 -9.97773290e-01 7.62769699e-01
-2.16002178e+00 -6.97335005e-01 3.61561298e-01 2.04752421e+00
1.52408743e+00 -4.57303263e-02 -8.75246990e-03 -3.23110551e-01
4.63398635e-01 1.20648444e-01 -6.04463816e-01 1.31545231e-01
-3.41766924e-02 1.29510507e-01 4.27659065e-01 3.40347290e-01
-1.35246837e+00 1.26222837e+00 5.48714685e+00 1.27797031e+00
-9.09712613e-01 1.15770452e-01 5.71325719e-01 -1.09286167e-01
-2.50612497e-01 -8.80491138e-02 -1.31815100e+00 2.71577597e-01
7.36285508e-01 -2.77656823e-01 2.59157270e-01 8.38163614e-01
-5.52957296e-01 1.13652341e-01 -1.13137400e+00 6.08056784e-01
-2.02938765e-02 -1.19603455e+00 4.51389492e-01 -1.37972981e-01
4.75282282e-01 3.66733283e-01 -2.10030928e-01 1.02708924e+00
4.73658055e-01 -8.19837809e-01 4.52309966e-01 4.50601429e-01
6.84112847e-01 -5.87164521e-01 9.20475483e-01 3.94469351e-01
-1.32466328e+00 -1.32275507e-01 -6.73485100e-01 4.92236823e-01
7.20302984e-02 6.76820695e-01 -1.00022399e+00 9.94314015e-01
8.15062284e-01 3.85524541e-01 -6.31456912e-01 1.01268888e+00
-2.39362016e-01 4.99367744e-01 -5.28960943e-01 -1.28702372e-01
1.64344817e-01 9.89746377e-02 2.91366816e-01 1.22192442e+00
9.81508940e-02 4.00977060e-02 2.68761694e-01 5.40139914e-01
-6.06753111e-01 4.00556743e-01 -1.10325880e-01 7.79921608e-03
7.01152444e-01 1.42594147e+00 -2.21459404e-01 -4.90966469e-01
-4.32476521e-01 6.41558349e-01 1.06601012e+00 5.83518744e-01
-6.85411453e-01 -5.30452311e-01 1.77403986e-01 -1.88404202e-01
7.06423163e-01 2.12450817e-01 2.60016769e-01 -1.35524380e+00
3.35183173e-01 -6.67265117e-01 9.94697928e-01 -5.31650543e-01
-1.34922123e+00 6.17561698e-01 1.32251740e-01 -7.59605944e-01
-3.10941130e-01 -4.18066978e-01 -5.68674989e-02 7.52741098e-01
-2.14668846e+00 -1.39365506e+00 1.47033492e-02 6.20743573e-01
-1.10177165e-02 -1.82523560e-02 8.81046891e-01 7.15827703e-01
-7.63269782e-01 8.00101817e-01 1.35663122e-01 3.76342088e-01
1.04765153e+00 -1.37371802e+00 5.24063557e-02 4.32704508e-01
1.72431305e-01 9.39299107e-01 -1.01204693e-01 -5.74715793e-01
-1.39891994e+00 -1.12645495e+00 1.15591156e+00 -5.43253243e-01
7.50871480e-01 -2.86044091e-01 -1.15641975e+00 7.87961721e-01
-4.16977070e-02 1.12892650e-01 6.99689329e-01 8.38171124e-01
-7.60094225e-01 -5.53462565e-01 -5.78468859e-01 3.72902751e-01
7.28167474e-01 -6.91794217e-01 -7.86415339e-01 4.48242694e-01
1.12121129e+00 -4.31776613e-01 -1.13560987e+00 6.83752358e-01
4.49108630e-01 -2.16209769e-01 1.27388680e+00 -7.69102037e-01
1.13710515e-01 -3.12371433e-01 7.03007951e-02 -1.00061297e+00
-2.83751994e-01 -3.61475766e-01 -8.24692070e-01 1.28959322e+00
1.00656223e+00 -5.85794330e-01 5.56788564e-01 6.76886499e-01
6.62645698e-02 -9.47287500e-01 -1.12444854e+00 -6.03999138e-01
-8.26459154e-02 -2.02125579e-01 7.02460647e-01 9.26086187e-01
1.14469722e-01 9.56686378e-01 -2.10965291e-01 3.94785792e-01
2.04756021e-01 4.19656694e-01 7.14130461e-01 -1.31314397e+00
-3.26406449e-01 -1.75928771e-01 1.94694608e-01 -1.53636789e+00
1.33822978e-01 -1.02307928e+00 5.32900468e-02 -1.74221754e+00
4.61256266e-01 -9.28598166e-01 -8.83565545e-01 8.08531284e-01
-8.02775979e-01 -1.72915384e-01 -8.43657628e-02 4.99004245e-01
-1.54684269e+00 7.11438179e-01 1.09105158e+00 -2.04924732e-01
-2.69015282e-01 -1.41258150e-01 -8.91799390e-01 3.45276833e-01
2.63936996e-01 -5.69644809e-01 -4.21714604e-01 -7.56565750e-01
8.20171654e-01 4.41105999e-02 1.80144966e-01 -5.20936430e-01
6.04858518e-01 1.39674932e-01 2.21115395e-01 -6.02151752e-01
3.63631457e-01 -6.85059428e-01 -4.24775422e-01 -1.09674394e-01
-3.98716837e-01 -3.76923740e-01 1.21262059e-01 8.64651442e-01
-4.77721184e-01 -1.89133450e-01 1.57626569e-01 2.27976665e-02
-6.99444532e-01 4.89095837e-01 1.61151230e-01 2.43434027e-01
6.34729385e-01 4.05742347e-01 -7.03502238e-01 -2.21948773e-01
-4.91660178e-01 8.72595668e-01 -6.51925132e-02 3.87642473e-01
2.88938820e-01 -1.26585054e+00 -5.47337055e-01 -3.16620171e-01
4.79132116e-01 6.69678271e-01 2.13476479e-01 7.85288334e-01
-1.96852475e-01 9.36254621e-01 4.53811526e-01 -3.44819486e-01
-8.75387788e-01 5.26946962e-01 2.26423234e-01 -1.12224698e+00
-3.16090286e-01 1.02922034e+00 2.99576432e-01 -5.20505011e-01
3.97549629e-01 -1.77453429e-01 -6.09728038e-01 4.11249965e-01
5.36059260e-01 2.76775569e-01 3.74682963e-01 -2.54978329e-01
-4.32754427e-01 2.45851874e-01 -8.46891165e-01 1.32929549e-01
1.28716099e+00 -1.00890487e-01 -3.38923633e-01 1.36015132e-01
1.08795083e+00 -1.84186459e-01 -7.38728702e-01 -8.13913703e-01
5.57641625e-01 3.94170657e-02 2.33774945e-01 -9.68392968e-01
-8.54725301e-01 5.23682714e-01 1.25794873e-01 2.19469309e-01
1.13358116e+00 3.62283677e-01 1.06993783e+00 1.21130407e+00
5.08964837e-01 -1.22701418e+00 9.74617302e-02 6.99553013e-01
7.39309907e-01 -1.21336484e+00 1.30209386e-01 -5.09288371e-01
-5.00314593e-01 7.00727940e-01 6.97872281e-01 3.60800147e-01
6.15776896e-01 -2.35722870e-01 -6.89374283e-02 -5.05492330e-01
-1.12120938e+00 -3.45849663e-01 7.86376834e-01 1.73048362e-01
3.64599198e-01 -4.25080210e-02 -3.00400734e-01 9.68341947e-01
1.80926859e-01 1.72858953e-01 -2.25978628e-01 8.50169837e-01
-4.99840409e-01 -1.03261566e+00 1.38334110e-02 6.38577998e-01
-5.63603044e-01 -3.75924915e-01 -2.24297032e-01 6.78765178e-01
-1.39118984e-01 8.42213690e-01 -2.04746798e-01 -1.80336803e-01
4.64164495e-01 3.16463679e-01 1.00112922e-01 -6.49785638e-01
-5.20886779e-01 -1.03940822e-01 4.86878842e-01 -6.19177103e-01
-3.40121806e-01 -1.59057423e-01 -1.07048166e+00 1.08178504e-01
-9.48254049e-01 4.74440932e-01 2.15744659e-01 1.00104821e+00
7.84923553e-01 3.78687471e-01 3.70305687e-01 -8.25432166e-02
-7.80746341e-01 -1.09127879e+00 -4.08868641e-01 1.50057867e-01
1.54651523e-01 -6.94511175e-01 7.29710236e-02 -2.31462613e-01] | [10.336955070495605, 8.060524940490723] |
482e7685-df89-4359-90b3-c22a34e5b9ed | a-hybrid-system-for-systematic-generalization | 2306.17249 | null | https://arxiv.org/abs/2306.17249v1 | https://arxiv.org/pdf/2306.17249v1.pdf | A Hybrid System for Systematic Generalization in Simple Arithmetic Problems | Solving symbolic reasoning problems that require compositionality and systematicity is considered one of the key ingredients of human intelligence. However, symbolic reasoning is still a great challenge for deep learning models, which often cannot generalize the reasoning pattern to out-of-distribution test cases. In this work, we propose a hybrid system capable of solving arithmetic problems that require compositional and systematic reasoning over sequences of symbols. The model acquires such a skill by learning appropriate substitution rules, which are applied iteratively to the input string until the expression is completely resolved. We show that the proposed system can accurately solve nested arithmetical expressions even when trained only on a subset including the simplest cases, significantly outperforming both a sequence-to-sequence model trained end-to-end and a state-of-the-art large language model. | ['Alessandro Sperduti', 'Alberto Testolin', 'Flavio Petruzzellis'] | 2023-06-29 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [ 3.83890122e-01 1.75992459e-01 2.01028228e-01 -4.05384958e-01
-3.93200696e-01 -8.56270373e-01 3.25322926e-01 1.32919833e-01
-3.14992189e-01 6.80934370e-01 -3.88898402e-01 -8.58081281e-01
-2.20926732e-01 -1.23767304e+00 -9.56842482e-01 -1.51810735e-01
3.78646217e-02 9.22724247e-01 3.95726204e-01 -5.84706545e-01
2.74943888e-01 4.65895057e-01 -1.50550878e+00 6.86111748e-01
1.26880372e+00 7.04313517e-01 -8.22621956e-03 8.00725698e-01
-7.77299225e-01 1.60801768e+00 -6.93934858e-01 -7.46474624e-01
1.47471905e-01 -5.13368011e-01 -1.00058961e+00 -6.15667343e-01
5.85170627e-01 -3.06103498e-01 -1.95903629e-02 1.40026116e+00
-1.15263611e-01 -3.65180410e-02 6.07594073e-01 -1.10968268e+00
-7.28119910e-01 1.19016993e+00 -2.88394056e-02 3.62556428e-02
4.96189535e-01 2.97924817e-01 1.01968288e+00 -3.45174670e-01
4.61296797e-01 1.24724078e+00 5.58380306e-01 6.33496225e-01
-1.19538426e+00 -6.38515711e-01 1.42093986e-01 4.51445132e-01
-1.36940241e+00 -7.69250393e-02 4.88155872e-01 -3.53867888e-01
1.45204008e+00 1.04750812e-01 6.20108426e-01 5.47244012e-01
1.12425335e-01 8.91129315e-01 9.14168298e-01 -6.34842277e-01
3.72847229e-01 -2.85652012e-01 4.09850568e-01 1.17274988e+00
8.00752193e-02 -3.04141939e-01 -2.83893198e-01 -9.01880860e-02
7.42345750e-01 -2.78701305e-01 6.70008510e-02 -3.97730172e-01
-1.04266942e+00 7.16529191e-01 3.35928917e-01 3.93434018e-01
-1.39730960e-01 4.21798706e-01 6.73705935e-01 7.13126779e-01
-2.61392027e-01 7.79498935e-01 -6.98719382e-01 -3.67178142e-01
-1.02288985e+00 6.33376539e-01 9.79233623e-01 8.73592615e-01
2.86151171e-01 6.08461127e-02 -5.09200580e-02 5.51629782e-01
-8.50875080e-02 5.78278124e-01 4.63317692e-01 -1.09598160e+00
4.32534307e-01 1.04093695e+00 -7.95122683e-02 -4.98629630e-01
-3.64513367e-01 -5.16271412e-01 -5.01384854e-01 3.06279033e-01
8.36623967e-01 -5.21499217e-02 -4.85562444e-01 2.12986135e+00
7.68854246e-02 2.14895785e-01 3.29987019e-01 6.14819348e-01
4.75253433e-01 6.43455029e-01 5.10342196e-02 -1.14357546e-01
1.24076355e+00 -1.05762208e+00 -4.34328347e-01 -4.15176243e-01
1.05244195e+00 -1.31286874e-01 1.10950899e+00 7.90225565e-01
-1.62559259e+00 -4.44830626e-01 -1.01896727e+00 -3.48221481e-01
-1.96041360e-01 -1.16382502e-01 9.11267102e-01 4.32924151e-01
-1.02459979e+00 4.99043226e-01 -5.74339986e-01 1.82434812e-01
3.53784829e-01 2.75012374e-01 -1.68689355e-01 -3.62492085e-01
-1.29310632e+00 1.08964050e+00 5.94456136e-01 4.53300178e-02
-6.37940288e-01 -7.98604727e-01 -1.09304726e+00 5.19840658e-01
6.89719319e-01 -7.22213924e-01 1.69183803e+00 -1.15127337e+00
-1.66154158e+00 6.91669881e-01 -3.80820483e-01 -6.92196667e-01
6.04599953e-01 -1.63033530e-01 -9.52778831e-02 1.13062672e-01
-5.99997677e-02 3.59382987e-01 5.79161346e-01 -6.50156260e-01
-3.26232255e-01 -3.20718110e-01 5.42900443e-01 -2.89470106e-01
3.13228011e-01 2.56318420e-01 -7.47815147e-02 -3.19041133e-01
2.42629275e-02 -7.36168921e-01 -1.17468081e-01 -2.79949908e-03
2.81541180e-02 -5.40043473e-01 8.86023939e-02 -6.35496140e-01
9.91913617e-01 -2.08300424e+00 6.95262372e-01 2.90752321e-01
-9.21341404e-02 5.58740914e-01 -2.46636197e-01 1.53103098e-01
-1.84911370e-01 -1.35464132e-01 -5.79686999e-01 1.28104553e-01
5.31789124e-01 2.59837776e-01 -6.45062268e-01 4.04267795e-02
5.65869927e-01 1.35706484e+00 -1.17454362e+00 -3.39413404e-01
-5.57672083e-02 -1.32788926e-01 -9.15596664e-01 2.48370841e-01
-1.07357299e+00 5.87188005e-02 -1.31068334e-01 3.66644353e-01
6.17388546e-01 -3.67262542e-01 5.88209271e-01 4.81064320e-01
3.27230841e-01 4.76685852e-01 -1.10740137e+00 1.78059161e+00
-7.12579787e-01 3.29255760e-01 -2.52541810e-01 -1.25621057e+00
7.84974694e-01 3.09175849e-02 -4.63931322e-01 -8.22324216e-01
7.86210671e-02 4.98216569e-01 6.73297167e-01 -4.65817690e-01
1.36987939e-01 -4.45297629e-01 -2.73310393e-01 5.26896954e-01
1.58399413e-03 -3.66320938e-01 6.76207960e-01 2.83570141e-01
1.27563536e+00 3.78640056e-01 4.64262396e-01 -4.21878733e-02
1.05634618e+00 4.29713093e-02 4.04885054e-01 8.69409084e-01
2.26717442e-01 1.70971043e-02 1.14841306e+00 -6.85323179e-01
-1.01592851e+00 -9.12120104e-01 2.08340019e-01 1.15612948e+00
-2.16271490e-01 -3.09548885e-01 -8.89296949e-01 -5.80889523e-01
5.50430082e-02 1.25587416e+00 -4.16348249e-01 -2.99824029e-01
-1.02889764e+00 4.49907556e-02 1.08939147e+00 7.22131968e-01
6.05855644e-01 -1.38347995e+00 -8.87154877e-01 2.28984624e-01
-2.75212284e-02 -1.15115285e+00 2.40012318e-01 3.36605221e-01
-6.68419421e-01 -1.03874862e+00 -2.38346756e-01 -7.70924687e-01
7.47678638e-01 -4.89340991e-01 1.23793221e+00 7.22145855e-01
-1.44763604e-01 -1.21241324e-01 -1.27318695e-01 -2.30014995e-01
-7.22166896e-01 1.93803594e-01 -4.15613353e-01 -4.72611040e-01
4.29971099e-01 -5.85017800e-01 2.42385104e-01 -9.87421349e-02
-9.41475689e-01 3.08810383e-01 4.54276085e-01 7.92091548e-01
1.59486488e-01 2.52270430e-01 4.56789553e-01 -9.15792584e-01
6.06278479e-01 -2.32349917e-01 -1.00254071e+00 6.83224082e-01
-2.86904443e-02 5.89734256e-01 1.28005195e+00 -4.49684858e-01
-9.28886592e-01 -1.00948393e-01 -1.09441265e-01 -1.57460958e-01
-2.36168042e-01 7.43220866e-01 -3.04224323e-02 5.92468157e-02
6.39577508e-01 5.50029874e-01 1.07039334e-02 2.38482803e-02
3.86824131e-01 1.00221246e-01 6.90019965e-01 -1.42584550e+00
7.26696908e-01 -5.77692464e-02 4.33466792e-01 -4.73675758e-01
-9.86696422e-01 8.17254186e-02 -5.80349982e-01 3.20471346e-01
4.69385862e-01 -6.98282599e-01 -9.92156446e-01 5.20502269e-01
-1.42930448e+00 -9.56322253e-01 -2.56949842e-01 -1.59676876e-02
-7.49213159e-01 3.06848586e-01 -6.92949295e-01 -8.33701968e-01
-1.53917223e-01 -1.28558576e+00 8.55754912e-01 -3.04338876e-02
-7.40837872e-01 -8.47264409e-01 -1.20997019e-01 2.99710095e-01
4.06799376e-01 -1.58937812e-01 1.71388268e+00 -7.75995255e-01
-7.50218868e-01 -1.74126983e-01 -3.61942440e-01 4.61794525e-01
-3.97877485e-01 -3.19376141e-02 -5.40168822e-01 1.57219261e-01
-1.91312432e-01 -8.01426649e-01 5.76965272e-01 -8.78921077e-02
1.35519433e+00 -1.17209189e-01 2.05347508e-01 5.90913296e-01
1.05390239e+00 2.00935956e-02 9.14542854e-01 2.29053125e-01
2.17467785e-01 3.58269542e-01 2.06742093e-01 -3.02702785e-02
4.07055140e-01 4.11153823e-01 9.30755809e-02 5.60662091e-01
9.39264894e-02 -2.45199203e-01 4.07764047e-01 3.94976109e-01
4.63151895e-02 2.56163955e-01 -1.22182608e+00 4.74424005e-01
-1.80871737e+00 -1.09738195e+00 3.85908433e-03 1.72654593e+00
1.37884104e+00 2.65951991e-01 -1.32338986e-01 4.24187452e-01
1.33264199e-01 -2.08191112e-01 -3.76605183e-01 -9.24686015e-01
3.39036323e-02 9.90328074e-01 -4.04998213e-02 8.87725055e-01
-6.65685296e-01 1.26248682e+00 6.11487103e+00 8.14395666e-01
-1.09852898e+00 -3.60869706e-01 5.11618704e-02 -3.67123447e-02
-3.18565279e-01 -1.75162420e-01 -4.58616763e-01 1.53216615e-01
9.61744726e-01 4.14656401e-02 9.74805117e-01 8.90351295e-01
-5.80069959e-01 -2.16044441e-01 -1.53277695e+00 5.38930893e-01
4.80775721e-02 -1.23687088e+00 1.60663694e-01 -6.04022324e-01
5.29560983e-01 -4.86392081e-01 -4.14302386e-02 8.52751851e-01
5.11480093e-01 -1.40918136e+00 9.91425872e-01 4.73462522e-01
5.22015035e-01 -7.91751266e-01 7.75455713e-01 8.51186097e-01
-8.04697752e-01 -3.22221488e-01 -2.12457061e-01 -6.94054663e-01
6.00880273e-02 1.72375321e-01 -5.79551399e-01 3.12801957e-01
4.52012271e-02 3.38850409e-01 -6.65815890e-01 7.26797342e-01
-1.15116727e+00 3.53463739e-01 -2.10033938e-01 -3.86981130e-01
2.74016261e-01 -7.17881322e-02 7.38410503e-02 1.14191163e+00
3.09227496e-01 4.97062117e-01 -7.09952340e-02 1.33192015e+00
9.01145861e-02 -2.72587568e-01 -4.39485192e-01 -3.16244096e-01
2.16003329e-01 7.98326492e-01 -6.08819306e-01 -7.56374300e-01
-3.88579041e-01 9.21527088e-01 9.97680128e-01 2.80486733e-01
-8.10884655e-01 -3.68273199e-01 4.18201864e-01 -2.43010134e-01
5.51968813e-01 -5.22913933e-01 -5.91481030e-01 -1.24384737e+00
1.43221796e-01 -1.45577514e+00 2.95661777e-01 -1.05557513e+00
-1.02277184e+00 2.40248039e-01 2.40316391e-01 -3.31972361e-01
-6.62176073e-01 -1.16577661e+00 -4.71412182e-01 1.07739186e+00
-1.47661507e+00 -9.27111328e-01 -2.09449362e-02 4.97902483e-01
3.18945497e-01 -2.76559949e-01 1.14594924e+00 4.32704166e-02
-7.57400468e-02 6.49996638e-01 -4.69454348e-01 3.68753940e-01
1.30825676e-02 -1.22512805e+00 5.41038990e-01 8.06164145e-01
1.08704135e-01 1.34146392e+00 6.53219640e-01 -3.45607579e-01
-1.58901131e+00 -7.16567338e-01 1.14740145e+00 -2.85659254e-01
9.08961535e-01 -3.26738000e-01 -1.24866521e+00 9.99074876e-01
-4.36141752e-02 -3.20437923e-02 4.03865486e-01 1.62235469e-01
-8.94671798e-01 2.12866440e-01 -1.07473421e+00 8.97651315e-01
1.17311609e+00 -7.44681478e-01 -1.21439493e+00 1.28847539e-01
7.73454845e-01 -8.73425126e-01 -3.22754741e-01 2.18107998e-01
4.77159858e-01 -8.10864687e-01 9.65262353e-01 -1.04721749e+00
9.59150732e-01 -5.18253207e-01 -2.41640136e-01 -1.15371454e+00
-2.17246443e-01 -4.29342896e-01 -3.05160582e-01 6.85024798e-01
4.96107966e-01 -5.98291874e-01 5.20767689e-01 6.86456501e-01
9.03567672e-02 -4.90255177e-01 -7.69480467e-01 -8.44055355e-01
4.98401314e-01 -7.17717648e-01 8.12735975e-01 7.94822991e-01
5.87112129e-01 3.12920421e-01 2.33886912e-01 2.18768656e-01
3.86626691e-01 5.67317545e-01 8.92280459e-01 -1.09388542e+00
-6.21902049e-01 -7.25408196e-01 -3.89455587e-01 -9.85344768e-01
1.07092619e+00 -1.31985211e+00 -5.54614365e-02 -1.28619277e+00
3.39919627e-02 -1.72177866e-01 1.18846763e-02 6.99409246e-01
-1.70620196e-02 -1.53990775e-01 5.91694154e-02 -5.55584311e-01
-6.70333028e-01 2.99584895e-01 1.10160148e+00 -3.23763371e-01
3.28816086e-01 -3.53143424e-01 -6.11839890e-01 8.92155230e-01
6.70031071e-01 -2.07697928e-01 -4.52533990e-01 -6.44352198e-01
1.03078568e+00 2.17304602e-01 4.96788025e-01 -1.03511560e+00
3.97538424e-01 -5.60424566e-01 4.83444594e-02 -1.46466896e-01
4.07856293e-02 -8.33741248e-01 -4.76438664e-02 7.53859937e-01
-7.19928503e-01 2.32274652e-01 5.41089594e-01 -1.35193974e-01
-1.74168780e-01 -7.37933040e-01 6.88206732e-01 -3.40240508e-01
-8.43646407e-01 -3.69855255e-01 -2.21162483e-01 3.57096791e-01
9.18246627e-01 3.71686190e-01 -4.38606828e-01 -1.98873803e-01
-4.55162942e-01 2.63225645e-01 4.64614809e-01 -9.65063646e-02
5.73681474e-01 -9.39819813e-01 -5.98914087e-01 3.47035915e-01
3.09729930e-02 3.65322351e-01 3.45153622e-02 7.05116093e-01
-1.14360726e+00 4.46434170e-01 -2.91930646e-01 -2.39583716e-01
-1.22889316e+00 8.70163739e-01 4.86037463e-01 -5.14572263e-01
-5.72031438e-01 9.72031593e-01 -4.50787880e-02 -9.11872864e-01
1.68268725e-01 -1.06981766e+00 2.24508286e-01 -5.01282573e-01
8.09556544e-01 -1.10692002e-01 -4.77836207e-02 -1.19828098e-01
-3.75935286e-01 5.66232681e-01 2.29360443e-02 1.25012724e-02
1.21357167e+00 6.77869260e-01 -7.89568424e-01 3.50880474e-01
5.53410649e-01 -1.29835665e-01 -5.57494164e-01 -5.76357782e-01
3.83505337e-02 -1.52081311e-01 -5.22680759e-01 -1.01763129e+00
-5.47635674e-01 1.17648470e+00 -4.95541453e-01 -2.73122685e-03
7.99844325e-01 -5.47840931e-02 5.44485211e-01 1.21886015e+00
6.37208700e-01 -5.88904202e-01 3.97396088e-02 1.38910401e+00
7.85318077e-01 -7.72788882e-01 -1.54853821e-01 -4.07540768e-01
-4.23693657e-01 1.27053905e+00 5.90907216e-01 -2.94603556e-01
-4.38426696e-02 6.81562245e-01 -2.60457158e-01 -1.18453708e-02
-9.32383060e-01 -1.75858252e-02 -2.00772868e-03 3.77084672e-01
4.94856656e-01 1.71247050e-02 -1.67080790e-01 8.09860766e-01
-6.62532389e-01 4.74780709e-01 2.02602506e-01 1.08960032e+00
-3.42444986e-01 -1.09939480e+00 -3.32533956e-01 -2.14665886e-02
-2.87995756e-01 -3.89604747e-01 -4.04395819e-01 7.13036716e-01
1.53068960e-01 3.35220397e-01 2.44562794e-03 2.49786451e-01
3.06795895e-01 7.82270670e-01 1.23957527e+00 -6.07440233e-01
-6.11343026e-01 -8.70385349e-01 2.07249209e-01 -4.43949908e-01
2.30422691e-02 -6.23438537e-01 -1.86947739e+00 -5.25428474e-01
2.88483411e-01 1.98335517e-02 1.60290092e-01 1.35225558e+00
-1.28420010e-01 8.27743888e-01 -2.06984341e-01 -3.23056668e-01
-1.02511990e+00 -7.12481916e-01 -3.81579399e-01 4.01960403e-01
2.72511244e-01 -2.96832204e-01 -5.75807169e-02 6.31879270e-02] | [9.415355682373047, 7.303221702575684] |
b35f60a3-4c39-44de-886f-d0be2c04a282 | post-hoc-analysis-of-arabic-transformer | 2210.09990 | null | https://arxiv.org/abs/2210.09990v1 | https://arxiv.org/pdf/2210.09990v1.pdf | Post-hoc analysis of Arabic transformer models | Arabic is a Semitic language which is widely spoken with many dialects. Given the success of pre-trained language models, many transformer models trained on Arabic and its dialects have surfaced. While there have been an extrinsic evaluation of these models with respect to downstream NLP tasks, no work has been carried out to analyze and compare their internal representations. We probe how linguistic information is encoded in the transformer models, trained on different Arabic dialects. We perform a layer and neuron analysis on the models using morphological tagging tasks for different dialects of Arabic and a dialectal identification task. Our analysis enlightens interesting findings such as: i) word morphology is learned at the lower and middle layers, ii) while syntactic dependencies are predominantly captured at the higher layers, iii) despite a large overlap in their vocabulary, the MSA-based models fail to capture the nuances of Arabic dialects, iv) we found that neurons in embedding layers are polysemous in nature, while the neurons in middle layers are exclusive to specific properties | ['Hassan Sajjad', 'Fahim Dalvi', 'Nadir Durrani', 'Ahmed Abdelali'] | 2022-10-18 | null | null | null | null | ['morphological-tagging'] | ['natural-language-processing'] | [-1.09645061e-01 3.01474314e-02 1.51465401e-01 -3.87546360e-01
-1.76556230e-01 -1.00128841e+00 7.64650643e-01 3.44567209e-01
-4.71130192e-01 1.64191797e-01 6.79004014e-01 -4.49327528e-01
1.05531521e-01 -8.54088306e-01 -4.32563484e-01 -7.35471606e-01
-1.40179753e-01 6.44374371e-01 8.79848450e-02 -6.94074571e-01
1.71652988e-01 4.88573432e-01 -1.09253788e+00 4.85425264e-01
1.00978208e+00 4.31160629e-01 1.24674529e-01 2.10127756e-01
-1.00632004e-01 6.88510418e-01 -2.60954648e-01 -5.78492939e-01
6.55134395e-02 -6.35077536e-01 -9.32726204e-01 -1.70396060e-01
7.12368786e-01 -1.66455716e-01 -3.44001889e-01 1.19034386e+00
2.02828482e-01 -2.98092216e-01 9.88634527e-01 -6.21459544e-01
-1.28764439e+00 1.23020208e+00 -2.09197506e-01 4.05258685e-01
9.72588435e-02 1.24101847e-01 1.40675330e+00 -9.73190546e-01
6.23571277e-01 1.52365935e+00 9.09435511e-01 3.85110438e-01
-1.42986071e+00 -3.08187097e-01 3.55339020e-01 -1.48098059e-02
-1.33788621e+00 -5.24544477e-01 4.92286414e-01 -4.08742517e-01
1.22366357e+00 -1.28723472e-01 8.04568946e-01 8.14540029e-01
2.70003974e-01 7.95703948e-01 1.53148305e+00 -5.59546292e-01
-2.27888092e-01 2.08430335e-01 5.81093907e-01 1.04004622e+00
1.35489225e-01 -8.37689564e-02 -5.91609597e-01 7.25190267e-02
7.97731519e-01 -4.20559824e-01 -1.82724670e-01 -5.13967536e-02
-1.23580813e+00 9.07386601e-01 5.62082708e-01 7.31683731e-01
-3.05774778e-01 -2.30341971e-01 4.23409104e-01 7.26762772e-01
1.47891089e-01 5.23958683e-01 -7.19503462e-01 2.03861177e-01
-6.49693191e-01 -1.97757557e-01 8.20790470e-01 6.39236093e-01
6.95888400e-01 3.85931402e-01 2.71237433e-01 1.24829352e+00
5.13798833e-01 5.83282053e-01 7.25604773e-01 -4.78449285e-01
6.69557810e-01 6.94284022e-01 -4.97860044e-01 -9.45862830e-01
-4.73415464e-01 -2.78490921e-03 -4.27661717e-01 1.17951572e-01
9.00368750e-01 -1.44835785e-01 -7.24773645e-01 1.99843574e+00
-3.19202185e-01 -7.38213778e-01 2.43248612e-01 5.95945716e-01
5.65095723e-01 6.65618241e-01 3.60302567e-01 3.10505807e-01
1.57072139e+00 -6.23357654e-01 -3.92964691e-01 -6.08596563e-01
8.31918299e-01 -8.37185502e-01 1.40238202e+00 1.72545925e-01
-1.34141767e+00 -4.26688880e-01 -1.20629287e+00 -3.32086712e-01
-8.93654108e-01 1.01977326e-01 6.18840694e-01 1.00068331e+00
-1.44093525e+00 4.43894058e-01 -8.62256110e-01 -8.48297834e-01
2.86631435e-01 4.36927468e-01 -5.82523704e-01 2.06182543e-02
-1.28039801e+00 1.60981154e+00 2.84685344e-01 1.13419704e-01
-7.58755565e-01 -3.03785414e-01 -1.08497727e+00 2.15842966e-02
-5.14124990e-01 -1.65877670e-01 9.05487776e-01 -1.37339246e+00
-1.30970991e+00 1.33207357e+00 -1.71540111e-01 -2.96597369e-02
-3.94844770e-01 1.07556522e-01 -5.25367260e-01 8.83241072e-02
-2.19537720e-01 6.86060727e-01 3.12417150e-01 -1.13716543e+00
-6.34559989e-01 -8.51717591e-01 1.29227102e-01 4.24379289e-01
-4.61246580e-01 4.08057868e-01 1.26793101e-01 -8.19496393e-01
4.43130642e-01 -1.01024687e+00 1.70309246e-01 -3.40191007e-01
2.65703704e-02 -2.45745048e-01 1.46094456e-01 -1.19255829e+00
1.12946808e+00 -1.89741862e+00 1.75889924e-01 2.61691511e-01
-1.05062038e-01 1.70770109e-01 -3.30307662e-01 6.63011491e-01
6.14267122e-03 2.54440904e-01 -3.74578059e-01 1.11042351e-01
3.53371471e-01 2.61361331e-01 -3.97323310e-01 7.27268994e-01
5.32162189e-01 9.53675687e-01 -6.34317517e-01 -1.28055885e-01
-6.34179711e-02 4.86898214e-01 -5.14981985e-01 -2.53871828e-01
-2.99638547e-02 1.81864109e-02 2.25351796e-01 8.04571688e-01
6.49288058e-01 2.33008251e-01 7.44286776e-01 -1.49239972e-01
-1.85428202e-01 1.15218008e+00 -7.88531303e-01 1.41341877e+00
-5.61236084e-01 7.69661427e-01 4.24819291e-01 -8.95268619e-01
6.02471530e-01 4.11502749e-01 -3.50440145e-01 -6.78000808e-01
3.22351038e-01 6.56135798e-01 9.06812727e-01 -1.89701617e-01
3.73946935e-01 -4.76198345e-01 -3.26918483e-01 7.77952731e-01
5.94466209e-01 1.32129192e-01 5.42943291e-02 2.00426206e-02
6.73681676e-01 1.16714567e-01 2.80414879e-01 -8.01223397e-01
5.24199188e-01 2.89768130e-02 1.62320837e-01 4.20187324e-01
-1.89873576e-01 5.01586556e-01 5.17776370e-01 -3.47899914e-01
-7.35457599e-01 -1.27197921e+00 -4.91988748e-01 1.61882353e+00
-1.02560401e-01 -2.37663671e-01 -8.95461261e-01 -6.13304317e-01
-1.80831105e-01 7.81765461e-01 -8.03506374e-01 -1.64513558e-01
-9.14635599e-01 -1.22700405e+00 1.05267596e+00 5.87021232e-01
2.34522611e-01 -1.45957243e+00 -2.39450872e-01 2.16916949e-01
2.98028234e-02 -7.83790410e-01 -4.03179944e-01 5.67513108e-01
-9.41418946e-01 -8.89438748e-01 -3.83385003e-01 -1.47364819e+00
5.89595437e-01 -1.64236218e-01 1.30033338e+00 3.74756455e-02
4.17054534e-01 2.61479437e-01 -1.39686465e-01 -3.12960714e-01
-6.92951977e-01 4.12756145e-01 3.13018598e-02 -1.35770366e-01
1.02286911e+00 -4.87852246e-01 -8.26731324e-02 7.63712227e-02
-8.34579825e-01 -5.72629929e-01 6.09853983e-01 5.57548404e-01
2.38155611e-02 -3.98703247e-01 5.48721373e-01 -1.03660357e+00
5.40020347e-01 -6.39030039e-01 -2.99165934e-01 2.74891078e-01
-4.13377464e-01 2.29788050e-01 5.43694913e-01 -3.64246607e-01
-9.32601810e-01 -2.22376227e-01 -5.69276869e-01 7.85635829e-01
-4.42190349e-01 6.88022733e-01 -4.62995023e-01 -1.87190380e-02
6.66486740e-01 3.34439427e-01 1.87820122e-01 -4.17262763e-01
3.52257878e-01 6.28796399e-01 2.32647404e-01 -6.26617134e-01
6.38081670e-01 3.69296193e-01 -5.44641137e-01 -1.01125360e+00
-5.20971775e-01 5.36670052e-02 -1.13999867e+00 2.64390260e-01
1.00474262e+00 -8.90923440e-01 -4.08376217e-01 8.34607780e-01
-1.03841496e+00 -6.15186691e-01 2.31442694e-02 3.23922783e-01
-3.43580037e-01 2.73053259e-01 -1.20815253e+00 -3.00365299e-01
-6.65038675e-02 -1.09944201e+00 4.14630711e-01 -1.74179867e-01
-6.11298561e-01 -1.59659445e+00 1.08238310e-01 -1.54929519e-01
4.48825777e-01 -3.90994310e-01 1.79935646e+00 -1.00573409e+00
-1.26770943e-01 1.26190126e-01 -1.18229970e-01 2.55713493e-01
2.87675500e-01 -1.06943034e-01 -1.07189739e+00 -1.61850527e-01
5.17780557e-02 -5.09524107e-01 9.40844953e-01 1.56597495e-01
1.85069039e-01 -3.74747038e-01 -3.71130109e-02 5.19680977e-01
1.29771054e+00 1.53122932e-01 5.13250589e-01 3.52956831e-01
6.71153724e-01 1.17099524e+00 -6.22483045e-02 -5.09059072e-01
8.21582139e-01 2.78443247e-01 1.64802000e-01 1.24879062e-01
-3.29572409e-01 -2.09804103e-01 1.21058774e+00 1.34429169e+00
1.95776090e-01 -4.13829237e-02 -1.16475952e+00 1.02268422e+00
-1.32030845e+00 -6.40553772e-01 -9.36881825e-02 2.16034484e+00
1.25312424e+00 1.52885308e-02 9.45010688e-03 -7.06704706e-02
4.06412363e-01 3.24334979e-01 -1.04539514e-01 -8.54312718e-01
-7.09266305e-01 4.57794219e-01 1.12701431e-01 7.65422761e-01
-8.84806097e-01 1.55209088e+00 6.84787369e+00 4.15206075e-01
-1.24236298e+00 1.15965463e-01 3.24991852e-01 3.48632663e-01
-4.76828456e-01 -4.09431495e-02 -9.25360441e-01 1.71936437e-01
1.33320928e+00 5.50552666e-01 5.98465145e-01 2.23911554e-01
-4.12946165e-01 2.17505097e-01 -1.31349432e+00 2.95775592e-01
3.39720070e-01 -9.61556494e-01 4.33970660e-01 4.74678650e-02
3.17084938e-01 4.91676360e-01 1.53241426e-01 4.69521195e-01
7.77377129e-01 -1.43421888e+00 1.00497198e+00 -9.16121714e-03
5.64266562e-01 -7.48687088e-01 4.86570865e-01 1.51644811e-01
-8.34233344e-01 -4.88069691e-02 -5.19633412e-01 -3.06681544e-01
-5.96569367e-02 -2.75361776e-01 -6.22497559e-01 -2.03041479e-01
4.49394256e-01 5.56580544e-01 -9.11103308e-01 3.87256384e-01
-5.66222310e-01 1.00305343e+00 -1.88271433e-01 -1.17188163e-01
5.58481872e-01 -5.12244403e-01 3.03463221e-01 1.60209668e+00
1.58492804e-01 -1.45003781e-01 -1.39760494e-01 6.22618020e-01
1.24743283e-01 4.68169659e-01 -6.43494248e-01 -3.53339016e-01
3.20571244e-01 1.14741981e+00 -8.25993955e-01 -3.19715813e-02
-7.21127152e-01 9.46078718e-01 6.78651512e-01 3.85839880e-01
-3.35013747e-01 -4.41444188e-01 8.11821401e-01 1.10603444e-01
2.58334249e-01 -5.55267274e-01 -5.76163411e-01 -1.10294926e+00
-3.70153189e-01 -1.04077506e+00 4.27446872e-01 -5.95472991e-01
-1.54745328e+00 8.34905446e-01 -6.15849257e-01 -4.62053716e-01
-1.88920841e-01 -1.35891879e+00 -4.02691722e-01 1.28605664e+00
-1.27455711e+00 -1.54931116e+00 3.87550831e-01 6.90208316e-01
1.62351891e-01 -5.32026827e-01 1.41316533e+00 1.19159408e-01
-4.52893853e-01 6.90797448e-01 1.76958099e-01 7.78425753e-01
7.72845268e-01 -1.50803411e+00 3.71303439e-01 8.24879050e-01
6.70403063e-01 1.10570860e+00 1.59552902e-01 -3.34460974e-01
-9.84571218e-01 -7.04091549e-01 1.54998052e+00 -9.64486361e-01
1.01920831e+00 -7.31131375e-01 -1.11418712e+00 1.17482114e+00
9.06958461e-01 -5.08108258e-01 9.69304085e-01 5.05179703e-01
-8.91633332e-01 -6.54894188e-02 -8.22448850e-01 8.20518911e-01
8.46330404e-01 -9.33907449e-01 -1.18359709e+00 -2.23173387e-02
3.49950463e-01 1.86116651e-01 -8.05898666e-01 -4.96399030e-02
5.23943365e-01 -7.69131243e-01 6.51691616e-01 -8.48244250e-01
4.81507748e-01 -5.54900803e-02 -4.94905978e-01 -1.70643044e+00
-5.74102104e-01 -6.10741489e-02 3.84216100e-01 1.43548787e+00
9.08417106e-01 -6.92650735e-01 2.87467331e-01 2.51704842e-01
-2.69238085e-01 -3.60816598e-01 -6.79085553e-01 -4.97774988e-01
9.31428075e-01 -1.54218361e-01 3.96955162e-01 1.40656352e+00
3.74849379e-01 8.32974017e-01 5.52094162e-01 1.99003443e-01
-6.23846985e-03 1.18373655e-01 -1.20206349e-01 -1.34910119e+00
1.19040068e-02 -8.58897984e-01 -1.60913035e-01 -6.71751499e-01
4.59389865e-01 -1.60670817e+00 -3.71496752e-02 -1.46870255e+00
-7.57353157e-02 -6.35004938e-01 -3.05710614e-01 9.03293133e-01
3.13767307e-02 5.66675603e-01 1.11261770e-01 1.42220318e-01
8.93888772e-02 2.20706239e-01 6.74140334e-01 -9.18554291e-02
-2.91838814e-02 -4.07568246e-01 -1.10085273e+00 1.13169909e+00
8.38758707e-01 -3.46530467e-01 -2.09713414e-01 -1.04452717e+00
6.25976443e-01 -5.00987351e-01 6.82478305e-03 -6.88752830e-01
-1.37940064e-01 6.02973215e-02 5.64433753e-01 -3.72851752e-02
1.64575964e-01 -7.97390759e-01 -6.75021350e-01 5.18005550e-01
-1.63167566e-01 4.34279680e-01 3.59189779e-01 -1.51077911e-01
-2.92156100e-01 -4.36772227e-01 1.11608899e+00 -4.53306019e-01
-6.77412093e-01 1.21967625e-02 -1.16946697e+00 2.31850013e-01
2.47286707e-01 -2.63466209e-01 -2.29764253e-01 3.58689353e-02
-7.95170844e-01 -1.44847319e-01 6.48226023e-01 4.49742466e-01
1.61629453e-01 -1.14828372e+00 -8.41676831e-01 4.50895786e-01
7.39886239e-02 -7.11739123e-01 -3.69422287e-01 8.25785220e-01
-6.17614090e-01 3.78354877e-01 -6.49042726e-01 -1.62857488e-01
-8.47526133e-01 2.62487113e-01 4.23139066e-01 -1.06450155e-01
-7.66333342e-02 9.07855213e-01 2.55051017e-01 -1.18700528e+00
-9.49400738e-02 -2.54287392e-01 -7.18998373e-01 8.94582748e-01
2.42968127e-01 2.21585229e-01 1.05325341e-01 -1.28197634e+00
-4.50633228e-01 5.64705491e-01 -2.83442110e-01 -4.17999268e-01
1.43859017e+00 -2.11316526e-01 -7.21400499e-01 7.76169598e-01
9.07858312e-01 6.98872626e-01 -7.78780699e-01 -1.75411418e-01
3.75878155e-01 3.46257627e-01 -2.97662735e-01 -1.07579732e+00
-9.13819671e-01 1.38602483e+00 3.42816591e-01 1.38588250e-01
8.89108598e-01 -8.06406438e-02 5.08923411e-01 2.72178769e-01
1.74602047e-01 -1.07070696e+00 -3.69334459e-01 1.50942457e+00
7.87974060e-01 -9.37059045e-01 -6.51076198e-01 -2.95771044e-02
-7.68071234e-01 1.06140339e+00 5.53415537e-01 -3.81331891e-01
7.33812273e-01 2.44138092e-01 5.35954952e-01 -1.79165110e-01
-4.29214209e-01 -2.45458320e-01 1.79780915e-01 7.32100487e-01
1.02413857e+00 -5.58079518e-02 -4.10281390e-01 8.37443292e-01
-6.68660879e-01 -8.29066396e-01 6.89669967e-01 6.21237159e-01
-3.65126610e-01 -1.10260928e+00 -2.87958771e-01 2.34708682e-01
-6.07009590e-01 -8.00046027e-01 -7.28576541e-01 9.63523149e-01
2.57067680e-01 7.94048846e-01 3.31710130e-01 -1.63934186e-01
9.59166586e-02 7.55393445e-01 7.66778409e-01 -9.73873436e-01
-1.20899343e+00 -1.66731223e-01 2.06175923e-01 -1.23093560e-01
-2.29207560e-01 -9.98967946e-01 -1.50294971e+00 -1.90886751e-01
5.43939769e-02 -6.14929246e-03 3.66421252e-01 9.86547172e-01
1.08309705e-02 1.07981209e-02 9.86946560e-03 -6.58009052e-01
-5.52604556e-01 -1.27617300e+00 -1.06658375e+00 3.28152448e-01
2.93781221e-01 -2.85582334e-01 -2.49877632e-01 1.03453733e-01] | [10.631967544555664, 10.016727447509766] |
22045b0a-386e-41c5-beb6-8d56bd657b86 | slabert-talk-pretty-one-day-modeling-second | 2305.19589 | null | https://arxiv.org/abs/2305.19589v1 | https://arxiv.org/pdf/2305.19589v1.pdf | SLABERT Talk Pretty One Day: Modeling Second Language Acquisition with BERT | Second language acquisition (SLA) research has extensively studied cross-linguistic transfer, the influence of linguistic structure of a speaker's native language [L1] on the successful acquisition of a foreign language [L2]. Effects of such transfer can be positive (facilitating acquisition) or negative (impeding acquisition). We find that NLP literature has not given enough attention to the phenomenon of negative transfer. To understand patterns of both positive and negative transfer between L1 and L2, we model sequential second language acquisition in LMs. Further, we build a Mutlilingual Age Ordered CHILDES (MAO-CHILDES) -- a dataset consisting of 5 typologically diverse languages, i.e., German, French, Polish, Indonesian, and Japanese -- to understand the degree to which native Child-Directed Speech (CDS) [L1] can help or conflict with English language acquisition [L2]. To examine the impact of native CDS, we use the TILT-based cross lingual transfer learning approach established by Papadimitriou and Jurafsky (2020) and find that, as in human SLA, language family distance predicts more negative transfer. Additionally, we find that conversational speech data shows greater facilitation for language acquisition than scripted speech data. Our findings call for further research using our novel Transformer-based SLA models and we would like to encourage it by releasing our code, data, and models. | ['Vera Tobin', 'Alekhya Yadavalli', 'Aditya Yadavalli'] | 2023-05-31 | null | null | null | null | ['language-acquisition', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [-1.02551192e-01 4.41845395e-02 -6.28344834e-01 -4.63811696e-01
-6.71148181e-01 -6.37589693e-01 7.26457179e-01 2.60418266e-01
-5.25502741e-01 5.92464149e-01 6.78313732e-01 -1.05567634e+00
-1.42051384e-01 -6.17119312e-01 -8.39936316e-01 -1.94187179e-01
9.17245671e-02 2.73452044e-01 5.37314676e-02 -5.40629745e-01
3.59948367e-01 2.35944882e-01 -1.28306079e+00 1.97076052e-01
1.13196075e+00 -5.15929833e-02 5.58105826e-01 2.40231872e-01
-3.07815969e-01 8.94695520e-01 -3.23535025e-01 -3.45496178e-01
3.60682979e-02 -7.40603209e-01 -1.10159683e+00 -3.55450422e-01
3.55899721e-01 -2.88232207e-01 7.54377944e-03 7.33303905e-01
3.72599185e-01 -1.49043873e-01 7.42249489e-01 -9.98480201e-01
-8.88350189e-01 1.22069705e+00 -2.52029628e-01 4.54104185e-01
6.71217918e-01 1.67542189e-01 8.78016353e-01 -8.07863832e-01
6.67824090e-01 1.54626358e+00 5.18919826e-01 7.98883438e-01
-1.27952111e+00 -9.84819710e-01 1.16565570e-01 -4.18092310e-02
-9.93212342e-01 -7.57630646e-01 5.38815379e-01 -7.52180338e-01
8.66013527e-01 -2.33001396e-01 9.35953200e-01 1.13238764e+00
1.19546197e-01 7.66869724e-01 1.74841869e+00 -9.56837475e-01
-2.52511680e-01 5.54613173e-01 9.87720788e-02 3.05169702e-01
-4.24405858e-02 3.81997168e-01 -1.12318170e+00 3.23665261e-01
4.99398530e-01 -7.53384769e-01 -2.30655834e-01 1.07241265e-01
-1.08909559e+00 8.23114574e-01 1.01569258e-01 9.49408770e-01
-1.07223779e-01 -3.92891020e-01 3.42731625e-01 1.08099389e+00
3.79609317e-01 4.39270973e-01 -6.68147624e-01 -7.03313828e-01
-5.98972023e-01 -1.01911359e-01 7.24415898e-01 8.10101330e-01
6.19400322e-01 -6.48377016e-02 4.64910626e-01 1.21091247e+00
4.16603655e-01 4.80494469e-01 7.13458836e-01 -3.97849709e-01
4.77001339e-01 2.37462446e-01 -3.37873816e-01 -4.51089621e-01
-9.81272385e-03 -2.91957520e-02 4.54994924e-02 3.45806479e-02
7.02188253e-01 -2.41592035e-01 -4.28480655e-01 2.37003708e+00
1.19236723e-01 -4.07479584e-01 3.00516099e-01 2.87259221e-01
4.37951714e-01 6.81497991e-01 4.71268475e-01 -5.22935331e-01
7.67300725e-01 -6.69371426e-01 -3.88677746e-01 -2.86816627e-01
1.08884275e+00 -8.60847831e-01 1.49667025e+00 2.40347609e-01
-1.38937855e+00 -6.07676923e-01 -8.46500516e-01 2.91558579e-02
-3.91991854e-01 -5.02822876e-01 6.58388495e-01 1.04947197e+00
-1.32605994e+00 4.60545689e-01 -8.37047935e-01 -6.71400070e-01
-6.99501857e-02 2.69388944e-01 -5.23663580e-01 -8.71943533e-02
-1.28777862e+00 1.14008379e+00 2.63190389e-01 -3.92966181e-01
-6.62021279e-01 -8.59442592e-01 -8.26210856e-01 -2.84622967e-01
-2.96648055e-01 -1.71131223e-01 1.25834084e+00 -1.58030593e+00
-1.45105481e+00 1.20145965e+00 -1.20533489e-01 8.28199089e-02
-3.26412395e-02 -2.10257336e-01 -5.40994227e-01 -1.48497775e-01
5.36309838e-01 6.50473893e-01 3.21463794e-01 -9.94894564e-01
-9.03614402e-01 -5.97781479e-01 -3.21354270e-02 5.31855881e-01
-5.07755518e-01 6.73837721e-01 2.82332331e-01 -6.42548680e-01
-1.39313340e-01 -8.23263168e-01 4.82614607e-01 -5.42043447e-01
4.58652169e-01 -7.56551445e-01 2.23860279e-01 -9.67787206e-01
1.27547491e+00 -2.25470543e+00 2.46477887e-01 1.61474705e-01
-1.84658691e-01 1.02703951e-01 -2.34839216e-01 9.08579469e-01
-1.78302705e-01 4.79135424e-01 9.87500325e-02 4.46765386e-02
-2.23414619e-02 1.65333703e-01 1.25441581e-01 2.42097795e-01
1.53500587e-01 7.41558015e-01 -8.23060215e-01 -3.13021094e-01
-7.34094232e-02 4.56144035e-01 -6.58932805e-01 2.39921495e-01
2.29925081e-01 6.93329990e-01 1.50137292e-02 5.10303199e-01
1.92366511e-01 3.52393568e-01 2.20224768e-01 5.14831245e-01
-7.56224692e-01 1.17974913e+00 -4.49522346e-01 1.42756498e+00
-7.71571517e-01 8.44886541e-01 3.35789472e-01 -8.59603882e-01
8.05452645e-01 4.22102660e-01 -3.08384039e-02 -8.12323689e-01
8.22098926e-02 7.38040328e-01 1.07758629e+00 -3.66279721e-01
-2.26850007e-02 -7.50156403e-01 5.35074472e-02 5.95043361e-01
8.41882676e-02 -4.78484482e-01 9.57854837e-02 1.19425587e-01
7.31494665e-01 8.02307799e-02 2.82502800e-01 -8.02680552e-01
5.26283503e-01 -2.18568757e-01 4.56336200e-01 3.74853581e-01
-2.97209591e-01 -4.06798869e-01 3.05437863e-01 3.16715807e-01
-8.28565717e-01 -1.14985740e+00 -2.02410638e-01 1.71834338e+00
-3.10026288e-01 -1.78712197e-02 -8.57420325e-01 -3.46257299e-01
2.56330539e-02 1.44469607e+00 -2.49310374e-01 -2.36610338e-01
-9.90661919e-01 7.38026425e-02 5.32534480e-01 4.47149158e-01
1.34335622e-01 -1.46040893e+00 -1.23660088e-01 2.08470270e-01
-1.23040333e-01 -6.92191005e-01 -6.53453887e-01 -1.32980868e-01
-6.88491762e-01 -5.68585873e-01 -3.19333702e-01 -1.25072658e+00
2.78372139e-01 -3.26779261e-02 9.80410337e-01 3.74088019e-01
2.69997329e-01 1.00541139e+00 -4.97011274e-01 -6.09333813e-01
-1.05389678e+00 3.16578746e-01 5.19241035e-01 -6.42793238e-01
6.85167432e-01 -7.50336766e-01 2.23344535e-01 3.40179354e-02
-5.42851269e-01 -8.65175575e-02 4.01437193e-01 6.28641129e-01
-2.15780735e-01 -2.45546684e-01 9.80282307e-01 -4.50092673e-01
8.97698820e-01 -6.82546616e-01 -2.14480191e-01 2.21856311e-01
-6.67276382e-01 -3.56622159e-01 4.02700096e-01 -5.33810258e-01
-1.23563290e+00 -5.10472417e-01 -2.84039050e-01 3.85174483e-01
-3.07046980e-01 8.16440642e-01 3.77043262e-02 -1.02400491e-02
4.84219342e-01 3.04594219e-01 2.53400207e-01 -4.43129122e-01
-2.49801278e-02 9.83843565e-01 2.41162077e-01 -1.15036035e+00
6.79982126e-01 -3.74087185e-01 -5.64284682e-01 -1.23532629e+00
-4.81411815e-01 -1.83612078e-01 -9.20486033e-01 -2.84968138e-01
8.13114345e-01 -1.19919086e+00 -4.92408186e-01 7.08390653e-01
-8.62966537e-01 -9.89207447e-01 -1.45803522e-02 1.08768570e+00
-6.47285581e-01 -1.57815553e-02 -7.86439896e-01 -7.46619999e-01
4.19604778e-03 -1.11613166e+00 4.55300003e-01 1.21040188e-01
-5.00982046e-01 -1.35122991e+00 1.61032408e-01 3.09533060e-01
5.51868975e-01 -3.19089651e-01 1.44872141e+00 -8.28547359e-01
-2.87070990e-01 5.47816575e-01 2.89406449e-01 5.35067618e-01
6.01120710e-01 -1.70646742e-01 -2.60707557e-01 -5.87782443e-01
1.94835499e-01 -7.46033728e-01 1.54340756e-03 2.14121446e-01
-2.69154971e-03 -2.80416578e-01 -8.61503463e-03 1.22696079e-01
1.11507320e+00 3.82313013e-01 1.39615223e-01 2.99120486e-01
2.90359050e-01 1.07262290e+00 8.14249814e-01 -3.00619930e-01
6.70365214e-01 4.04148668e-01 -4.39134479e-01 3.49612147e-01
-1.53262392e-01 -5.30657053e-01 1.14755476e+00 1.66207290e+00
9.40557420e-02 -8.77884496e-03 -1.38099980e+00 1.03208709e+00
-9.07483637e-01 -6.81973219e-01 5.20440638e-02 2.30259943e+00
1.40423012e+00 1.04350008e-01 2.34797254e-01 -2.51256898e-02
4.70402956e-01 -8.78235400e-02 -4.02449965e-02 -7.52389908e-01
3.99343446e-02 5.38162529e-01 7.63837770e-02 1.13327289e+00
-1.18224941e-01 1.50127912e+00 5.99013090e+00 6.20268583e-01
-1.31962967e+00 7.32956529e-02 3.77563834e-01 1.53677195e-01
-7.02660918e-01 6.79992465e-03 -8.84712219e-01 4.06366259e-01
1.21319866e+00 -5.09469211e-01 6.34675503e-01 3.12539935e-01
2.16153145e-01 -1.45104259e-01 -1.50747287e+00 3.77880305e-01
-4.22960967e-02 -4.77513880e-01 -2.49789774e-01 -3.51341255e-02
4.60382789e-01 1.49374038e-01 -1.76702663e-02 7.16648161e-01
3.40265512e-01 -1.00621760e+00 8.23539555e-01 -4.13147453e-03
9.87933218e-01 -5.40398479e-01 2.89556265e-01 7.22349465e-01
-8.64057660e-01 1.11669619e-02 9.24272910e-02 -5.87659419e-01
3.27043861e-01 -3.58683020e-01 -1.00192559e+00 -1.48682399e-02
5.45114517e-01 4.11385596e-01 -4.34909612e-01 6.47880509e-02
-4.00115490e-01 1.35659242e+00 -2.77960628e-01 -9.31694433e-02
1.72549933e-01 -1.63054630e-01 4.28681821e-01 9.12337422e-01
3.25543314e-01 3.44659835e-01 4.55602258e-03 4.36584681e-01
3.95632923e-01 7.95789361e-01 -1.02549911e+00 -4.65875924e-01
7.42502868e-01 4.47910786e-01 -4.27769631e-01 -6.23461464e-03
-8.48186016e-01 7.20101953e-01 4.80145663e-01 6.54307976e-02
-1.85748890e-01 -7.91515037e-02 7.26361334e-01 4.36688095e-01
-8.61181393e-02 -5.48670232e-01 -7.51714483e-02 -7.41098404e-01
-1.92713097e-01 -1.26245928e+00 -9.61310863e-02 -3.68486375e-01
-1.34759295e+00 1.38446704e-01 1.79739475e-01 -4.61218268e-01
-5.80010951e-01 -6.06539190e-01 -3.85257989e-01 1.31811750e+00
-1.00529790e+00 -1.27606142e+00 4.53296304e-01 6.17834926e-01
7.08982825e-01 -3.97845894e-01 5.72714865e-01 2.93002039e-01
-1.13571949e-01 7.97758460e-01 -3.36241007e-01 4.80671786e-02
8.92524719e-01 -9.73573685e-01 2.47724399e-01 5.45999408e-01
-1.76470086e-01 1.05267990e+00 3.56572092e-01 -9.45979476e-01
-9.07703340e-01 -2.75582463e-01 1.43150628e+00 -5.09081721e-01
8.98347318e-01 -2.92493552e-01 -9.07955110e-01 1.12027550e+00
8.44133258e-01 -1.04610121e+00 8.88109803e-01 3.31246525e-01
-2.19710376e-02 1.47391394e-01 -9.50034142e-01 8.51560175e-01
1.45965683e+00 -7.38165975e-01 -8.28277826e-01 1.17507592e-01
1.13372099e+00 1.08859271e-01 -1.21924245e+00 3.54155928e-01
5.86473703e-01 -1.02232468e+00 5.26878893e-01 -5.66351771e-01
4.06381786e-01 4.47347462e-01 -1.63851842e-01 -1.59339738e+00
-2.84888953e-01 -4.47057962e-01 8.26382756e-01 1.57873523e+00
5.76114237e-01 -9.42342997e-01 1.33148029e-01 3.75447482e-01
-3.00889850e-01 -5.48899472e-01 -1.01087308e+00 -6.62422061e-01
1.03086150e+00 -4.58238155e-01 5.24313211e-01 1.34408927e+00
3.32370639e-01 5.36686897e-01 1.78087220e-01 -2.75898725e-02
5.85216694e-02 -6.15024447e-01 6.64111853e-01 -1.24001610e+00
-2.26371139e-01 -6.42096221e-01 -5.16357087e-02 -7.21016586e-01
6.70972526e-01 -1.06560934e+00 -5.91220446e-02 -1.10956478e+00
-2.25578278e-01 -8.04700851e-01 -7.00309724e-02 4.00405526e-01
-7.18821511e-02 -5.06790042e-01 5.77007651e-01 1.12247586e-01
2.32720286e-01 4.90402371e-01 9.64700162e-01 3.93032104e-01
-6.80349708e-01 -4.42028642e-02 -7.17552722e-01 7.92517126e-01
1.00058460e+00 -4.31507587e-01 -5.92131197e-01 -5.24995804e-01
7.81600103e-02 4.60400939e-01 -3.01662952e-01 -5.57550371e-01
-3.76771241e-02 -4.66163099e-01 -2.41671577e-01 -9.87661853e-02
-1.80681571e-01 -5.48255444e-01 -2.73986220e-01 4.73582327e-01
-3.85747820e-01 4.48890567e-01 4.47863698e-01 -5.06298721e-01
-2.08971247e-01 -4.88291025e-01 6.67542517e-01 4.80344817e-02
-3.73726785e-01 -9.25108343e-02 -7.21739233e-01 4.44663435e-01
8.35260391e-01 -1.07345730e-01 7.01099038e-02 -5.62893331e-01
-5.11776924e-01 1.64667130e-01 6.75859749e-01 8.72780561e-01
2.15524256e-01 -1.26185751e+00 -9.19629216e-01 3.40246588e-01
-1.53606176e-01 -4.72745389e-01 -2.61298448e-01 9.03691947e-01
-2.24542588e-01 5.20971954e-01 -4.21246767e-01 -2.65065610e-01
-1.26932395e+00 2.62745053e-01 8.54804441e-02 4.58123572e-02
1.92410305e-01 1.08266187e+00 3.76810253e-01 -1.00621092e+00
-1.56284779e-01 1.48894846e-01 -1.16331287e-01 1.59374982e-01
1.43761680e-01 4.68340009e-01 -3.95480692e-01 -9.19980943e-01
-2.93903798e-01 5.21618962e-01 -2.97222257e-01 -7.53113270e-01
1.14420974e+00 -2.63040006e-01 -5.32697499e-01 1.10599804e+00
1.08307517e+00 9.00882065e-01 -5.56524992e-01 -1.60227999e-01
8.28305110e-02 -1.40089765e-01 -4.12695050e-01 -8.14447522e-01
-3.31522495e-01 8.81347775e-01 3.33818942e-01 -2.67485455e-02
9.69045162e-01 2.84787238e-01 3.28081369e-01 6.55491557e-03
4.55394536e-01 -1.22537684e+00 -9.19242352e-02 7.01236665e-01
1.09025049e+00 -1.05331862e+00 -5.34835517e-01 -3.94305944e-01
-5.38519681e-01 6.04466975e-01 1.04885888e+00 3.23047817e-01
7.65990555e-01 9.19907093e-02 3.60765547e-01 4.68871035e-02
-9.10407901e-01 -7.00275553e-03 -8.40329751e-03 7.00484395e-01
1.20887375e+00 2.77495176e-01 -1.10617268e+00 3.60783577e-01
-9.88369465e-01 -1.70882866e-01 3.41509908e-01 8.26166391e-01
-2.50330925e-01 -1.63738120e+00 -3.32211286e-01 2.43872315e-01
-4.13812250e-01 -6.62173748e-01 -5.03578961e-01 1.27318585e+00
3.44207108e-01 9.98525202e-01 1.74220651e-01 -3.07182103e-01
1.14386007e-01 5.37832797e-01 6.51362181e-01 -9.73621428e-01
-9.38409984e-01 2.61187293e-02 1.84677809e-01 2.81870924e-02
-5.09113967e-01 -1.21150672e+00 -1.12862825e+00 -4.06301618e-01
-3.33124489e-01 2.21821621e-01 5.18461585e-01 1.09098840e+00
-1.96407914e-01 -6.51030010e-03 4.79872704e-01 -2.99989522e-01
-3.76752138e-01 -1.36703312e+00 -6.38641775e-01 4.58613485e-02
1.46502674e-01 -4.38872814e-01 -6.25456691e-01 -1.60703599e-01] | [10.872736930847168, 9.953985214233398] |
9ad7a769-9915-4d41-92d9-c2fc83450071 | toward-discourse-aware-models-for | null | null | https://aclanthology.org/2021.ranlp-srw.29 | https://aclanthology.org/2021.ranlp-srw.29.pdf | Toward Discourse-Aware Models for Multilingual Fake News Detection | Statements that are intentionally misstated (or manipulated) are of considerable interest to researchers, government, security, and financial systems. According to deception literature, there are reliable cues for detecting deception and the belief that liars give off cues that may indicate their deception is near-universal. Therefore, given that deceiving actions require advanced cognitive development that honesty simply does not require, as well as people’s cognitive mechanisms have promising guidance for deception detection, in this Ph.D. ongoing research, we propose to examine discourse structure patterns in multilingual deceptive news corpora using the Rhetorical Structure Theory framework. Considering that our work is the first to exploit multilingual discourse-aware strategies for fake news detection, the research community currently lacks multilingual deceptive annotated corpora. Accordingly, this paper describes the current progress in this thesis, including (i) the construction of the first multilingual deceptive corpus, which was annotated by specialists according to the Rhetorical Structure Theory framework, and (ii) the introduction of two new proposed rhetorical relations: INTERJECTION and IMPERATIVE, which we assume to be relevant for the fake news detection task. | ['Thiago Pardo', 'Fabrício Benevenuto', 'Francielle Vargas'] | null | null | null | null | ranlp-2021-9 | ['deception-detection'] | ['miscellaneous'] | [-2.52789445e-03 4.44598585e-01 -3.22891563e-01 -2.69559056e-01
-6.85344815e-01 -8.45947683e-01 1.12085199e+00 4.50655878e-01
-1.38728365e-01 1.03237104e+00 8.24996233e-01 -7.53291607e-01
-6.01358153e-02 -2.94031024e-01 -3.99963796e-01 -2.33282983e-01
5.69248796e-01 1.71955153e-01 -1.16537854e-01 -6.64993942e-01
1.26784849e+00 4.31446701e-01 -1.06892264e+00 6.32716894e-01
1.08841538e+00 5.11255562e-01 -3.46889704e-01 4.08490509e-01
1.35589419e-02 1.81788659e+00 -1.28786623e+00 -1.01669252e+00
-3.99049401e-01 -7.96801150e-01 -1.17139220e+00 1.58243701e-01
7.36546338e-01 -5.24554431e-01 -1.46124467e-01 1.26252937e+00
2.78007209e-01 -4.23818797e-01 4.75011230e-01 -1.01841199e+00
-1.38489461e+00 6.35664523e-01 -2.91482005e-02 5.00710905e-01
7.64515221e-01 1.17056752e-02 6.81679845e-01 -8.05183887e-01
9.48321879e-01 1.55859423e+00 5.23704350e-01 6.13871336e-01
-5.68705678e-01 -4.37296122e-01 -2.13605061e-01 6.23055279e-01
-7.89410472e-01 -9.94300485e-01 1.04230785e+00 -7.61476755e-01
7.56803036e-01 4.78212327e-01 7.19446242e-01 1.66930628e+00
4.16479617e-01 5.08852601e-01 1.63926208e+00 -6.95648313e-01
-3.77921350e-02 6.95732594e-01 4.19298798e-01 5.97533882e-01
5.38116932e-01 1.47715062e-01 -7.53898501e-01 -6.01881683e-01
6.06671274e-02 -8.46070111e-01 -3.80792677e-01 2.21096441e-01
-8.68189931e-01 1.34786868e+00 1.91958189e-01 8.69597197e-01
-2.26887450e-01 -1.39276221e-01 9.09132481e-01 6.35048091e-01
8.40761542e-01 6.40052319e-01 2.68068820e-01 -5.96424818e-01
-7.23651588e-01 1.47368550e-01 1.18426025e+00 3.87231499e-01
-2.22625241e-01 3.55777055e-01 1.06651485e-01 7.72270381e-01
4.07680690e-01 3.81999493e-01 6.53918982e-01 -9.00963426e-01
5.40494800e-01 5.24112046e-01 3.79259974e-01 -1.97801256e+00
1.72390174e-02 -3.35684657e-01 -2.35144541e-01 -6.49214536e-02
5.41850448e-01 1.22429095e-01 2.11857073e-03 1.23485899e+00
7.95823410e-02 -6.68150961e-01 3.02754819e-01 1.15654957e+00
6.93734288e-01 5.03858924e-01 -2.63885587e-01 -6.78478718e-01
1.48092854e+00 -6.64326131e-01 -1.13360023e+00 -3.69344205e-01
1.03082156e+00 -1.03867328e+00 9.54658926e-01 5.24790943e-01
-1.12101960e+00 2.44811848e-01 -1.31421137e+00 -3.28258753e-01
-2.92367846e-01 -1.93019770e-02 4.88271654e-01 1.16653323e+00
-5.82721055e-01 2.20866963e-01 -4.97999549e-01 -1.96700081e-01
4.43573385e-01 -6.89213037e-01 -1.03901640e-01 -2.58869212e-02
-1.33219004e+00 1.75317109e+00 2.69429386e-01 4.61140275e-01
-6.19832873e-01 -1.17055299e-02 -8.24303627e-01 -5.05905390e-01
4.10945147e-01 -2.25541294e-01 8.92422140e-01 -1.55920589e+00
-1.28536928e+00 1.41565084e+00 1.29536549e-02 -4.80729014e-01
7.28905320e-01 -2.57442594e-01 -8.13727975e-01 5.49237847e-01
2.66209126e-01 -2.66329765e-01 1.19778740e+00 -1.34464169e+00
-1.81292534e-01 -6.07166409e-01 1.57062054e-01 1.65700257e-01
-3.20523113e-01 9.24839258e-01 1.07437289e+00 -7.45948613e-01
1.31435096e-01 -5.15058100e-01 7.57788718e-01 -3.68459046e-01
-4.64325875e-01 -3.40817988e-01 9.26264286e-01 -1.23282957e+00
1.21101248e+00 -1.84969246e+00 -2.60400742e-01 -2.32528657e-01
4.76377785e-01 3.58043075e-01 5.07935822e-01 7.68550754e-01
-3.25036608e-02 3.96428674e-01 -1.17427409e-01 1.97686806e-01
3.24761122e-01 2.36346453e-01 -6.29898906e-01 1.14172506e+00
-1.93154395e-01 8.70840907e-01 -9.89514291e-01 -4.34307545e-01
-1.04862504e-01 3.01405936e-01 5.95431253e-02 -2.75480598e-01
2.48145089e-01 8.74257833e-02 -3.62023860e-01 7.88766265e-01
4.28135365e-01 2.89847888e-02 1.42063290e-01 1.61199734e-01
-5.19530773e-01 1.02298796e+00 -1.94332600e-01 7.58031964e-01
-7.92157128e-02 1.42933714e+00 3.80641043e-01 -1.07758880e+00
1.05485892e+00 5.60116768e-01 -5.38213670e-01 -6.52924597e-01
2.24339873e-01 7.45707154e-01 2.56514400e-01 -8.42386961e-01
8.00142348e-01 -4.58048254e-01 -2.73571968e-01 7.61317492e-01
-4.75710154e-01 -2.40988716e-01 -1.92687988e-01 4.48611557e-01
7.02666044e-01 -2.01292187e-01 5.00670254e-01 -4.86356467e-01
8.34407985e-01 4.36166704e-01 2.57893950e-01 5.24680614e-01
-4.65967566e-01 -1.03886642e-01 8.92117143e-01 -5.68790793e-01
-9.09515679e-01 -3.81560147e-01 -2.38070190e-01 6.95717990e-01
-6.59863353e-02 3.12364865e-02 -8.28795671e-01 -7.77674258e-01
-3.33204478e-01 1.49926388e+00 -3.87649119e-01 -2.12593883e-01
-7.04047799e-01 -5.09705245e-01 1.09554005e+00 -3.42211425e-01
5.85653543e-01 -9.35010791e-01 -9.36305642e-01 8.56062621e-02
-4.98765260e-01 -9.45101619e-01 -6.00890303e-03 -4.35712636e-01
-4.20993507e-01 -1.31418633e+00 -1.18789673e-01 -5.58226168e-01
3.93758625e-01 3.70849848e-01 7.70950973e-01 6.55036867e-01
2.22675636e-01 2.80572027e-01 -7.31178582e-01 -4.43396062e-01
-1.36926317e+00 -5.22852421e-01 -2.28336871e-01 -3.84769171e-01
2.98391551e-01 -2.28845447e-01 -1.81559585e-02 3.01187346e-03
-9.23412323e-01 -3.46365303e-01 1.81004360e-01 9.24734771e-01
-6.29293680e-01 -3.56561095e-01 8.34349394e-01 -9.48226392e-01
1.39433336e+00 -5.80709398e-01 -1.68839365e-01 2.44202897e-01
-5.60061038e-01 -5.06232679e-01 4.04706627e-01 -1.45555362e-01
-1.30717242e+00 -1.17557323e+00 -6.77494854e-02 3.26630086e-01
-2.34682664e-01 7.11981058e-01 2.80377984e-01 -3.24160457e-01
1.15538585e+00 1.83376625e-01 3.01490486e-01 6.48183422e-03
2.08610952e-01 9.79475260e-01 3.45598370e-01 -5.19086719e-01
5.33156216e-01 3.18465322e-01 -4.53409344e-01 -9.91814613e-01
-1.21484351e+00 -2.88985390e-02 -3.46737832e-01 -5.58785617e-01
4.18016583e-01 -5.76427698e-01 -5.71814597e-01 4.58013266e-01
-1.77011895e+00 1.72065511e-01 2.41202772e-01 4.52697068e-01
-4.08336222e-01 8.79720628e-01 -1.09511173e+00 -1.13042283e+00
-1.35963902e-01 -8.21300626e-01 4.59529936e-01 -5.14216959e-01
-5.67063630e-01 -1.31530499e+00 -1.72759041e-01 1.17737138e+00
3.02012354e-01 6.98514521e-01 1.00672233e+00 -9.84353602e-01
1.11288214e-02 -2.17112079e-01 4.50438298e-02 5.12055278e-01
8.93935636e-02 -1.83422327e-01 -7.01931298e-01 -3.30825299e-02
1.02430189e+00 -9.93866503e-01 2.57566750e-01 -2.04256356e-01
1.77125826e-01 -1.25617075e+00 1.75400198e-01 -3.79672170e-01
9.58163142e-01 9.07956585e-02 7.80772626e-01 6.17310584e-01
3.16531956e-01 1.09960759e+00 7.09173262e-01 3.14930499e-01
5.66024184e-01 6.14104569e-01 4.26280439e-01 5.69456816e-01
2.02019070e-03 -1.66698322e-01 7.98612118e-01 1.10498548e+00
1.29010051e-01 -3.09012622e-01 -9.16408956e-01 4.74998802e-01
-1.56034946e+00 -1.42263031e+00 -9.25588667e-01 1.73443007e+00
7.67311871e-01 1.17068760e-01 -1.27287507e-01 2.59754151e-01
7.30786383e-01 5.25528133e-01 7.52709433e-02 -1.18906975e+00
-3.86095673e-01 -3.64625961e-01 2.66317446e-02 8.39586854e-01
-6.22759998e-01 9.72115755e-01 6.04965115e+00 7.53991842e-01
-8.38489354e-01 4.63525921e-01 5.04821599e-01 3.21100295e-01
-3.47924083e-01 -6.08308502e-02 -2.53805667e-01 6.25736594e-01
9.43276286e-01 -2.21818641e-01 2.29855776e-01 5.91354191e-01
5.92479706e-01 -3.75404358e-01 -7.31474280e-01 5.45668185e-01
9.16963518e-01 -1.24676621e+00 -2.43403479e-01 1.00533620e-01
4.27341253e-01 -6.36359751e-01 3.65839526e-02 -2.23800894e-02
3.63146663e-02 -9.64282990e-01 1.37951577e+00 2.93139100e-01
1.14450315e-02 -5.18677056e-01 1.19315994e+00 7.92462826e-01
3.21968317e-01 9.05848388e-03 -3.70478719e-01 -5.37023306e-01
4.32404786e-01 5.32592893e-01 -8.58863413e-01 3.66573811e-01
4.59491722e-02 5.54982305e-01 -5.33086181e-01 2.46176660e-01
-8.33849490e-01 9.00513828e-01 4.07853544e-01 -3.75172794e-01
4.06565338e-01 -2.98868954e-01 1.09337699e+00 1.05695784e+00
2.06861183e-01 2.14320302e-01 -3.60284686e-01 9.51716244e-01
1.26069069e-01 5.09102196e-02 -7.74045587e-01 -2.73801208e-01
5.96178949e-01 8.52388918e-01 -4.65490997e-01 -2.98432946e-01
-1.46510318e-01 1.05105996e+00 3.27325851e-01 -4.14606929e-02
-5.96474588e-01 6.10926151e-02 2.15011701e-01 1.32246539e-01
-3.33328485e-01 -1.77990749e-01 -4.38457608e-01 -1.01507521e+00
1.98959276e-01 -1.33277810e+00 9.71817821e-02 -9.47560310e-01
-1.36222482e+00 4.51319665e-01 2.18772925e-02 -6.16317809e-01
-4.18390810e-01 -5.31930387e-01 -1.63087890e-01 6.09741092e-01
-1.19242930e+00 -1.37926126e+00 1.00187473e-01 1.55110836e-01
4.21474040e-01 -4.15704995e-02 7.31632292e-01 -9.97137204e-02
-2.06165910e-01 2.59295344e-01 -2.57972807e-01 6.22314438e-02
7.42087424e-01 -7.42603838e-01 -5.75976968e-02 8.09246838e-01
-9.88278687e-02 7.64262676e-01 1.14402997e+00 -8.31441641e-01
-9.21800196e-01 -2.01184988e-01 1.65770352e+00 -7.69346595e-01
1.35415053e+00 -2.94843428e-02 -1.09901273e+00 6.79885268e-01
6.65780306e-01 -4.95961279e-01 7.12656140e-01 -3.13972145e-01
-5.03125966e-01 6.69065654e-01 -1.52269888e+00 3.78591388e-01
8.15436006e-01 -7.29457259e-01 -1.40200329e+00 8.31386328e-01
3.03350538e-01 -4.58268553e-01 -6.84008658e-01 -5.06329417e-01
3.46183151e-01 -1.19940722e+00 5.73534787e-01 -6.18383765e-01
7.72984028e-01 6.63271323e-02 6.27235919e-02 -1.36575806e+00
-5.24232760e-02 -5.64389229e-01 4.86770645e-02 1.21615696e+00
2.38062143e-01 -1.11576641e+00 4.12524551e-01 5.47427773e-01
-6.11480057e-01 -1.54957294e-01 -1.31728256e+00 -5.99749625e-01
3.68646443e-01 -3.45051080e-01 2.87677646e-02 1.82578468e+00
6.80521488e-01 3.37510258e-01 -5.71141541e-01 -1.30783096e-01
4.71415609e-01 -2.53075249e-02 3.67361009e-01 -1.01784647e+00
1.03057899e-01 -4.83607084e-01 -4.71545607e-01 -9.10991013e-01
6.02268040e-01 -6.10988021e-01 -1.88629016e-01 -1.15578997e+00
1.13341630e-01 1.30064428e-01 7.57013857e-01 7.96092227e-02
5.69618195e-02 9.99801829e-02 1.77096426e-01 4.69886005e-01
-2.61326164e-01 5.53160846e-01 1.25473523e+00 -2.11708024e-02
1.41063318e-01 -3.76617789e-01 -1.11967742e+00 1.01357830e+00
9.20002103e-01 -7.34316707e-01 -7.10857213e-02 -3.11766475e-01
6.06618702e-01 4.02086377e-01 8.97137344e-01 -1.66802496e-01
1.54267848e-01 -2.68363625e-01 -1.89768314e-01 -2.26860270e-01
3.83998632e-01 -4.85276759e-01 -1.57246009e-01 6.83600843e-01
-4.41876233e-01 1.84590474e-01 -1.58312276e-01 2.28876188e-01
-3.98027629e-01 -9.79231656e-01 6.83429837e-01 -2.51287460e-01
-2.66086429e-01 -9.40037787e-01 -1.02126503e+00 1.85103357e-01
9.30014610e-01 -4.77381349e-01 -1.25657690e+00 -8.42903078e-01
-3.81186426e-01 -3.62992048e-01 6.67620838e-01 2.31228411e-01
6.89959109e-01 -1.28892612e+00 -1.07904148e+00 -7.23278522e-01
-1.43133206e-02 -1.00468862e+00 -4.08541001e-02 1.15602016e+00
-7.87249207e-01 7.15982974e-01 -1.78746298e-01 1.46629214e-01
-1.29694295e+00 4.68829453e-01 1.43490240e-01 2.15911806e-01
-3.38352293e-01 5.82242429e-01 -4.16296780e-01 -8.33963379e-02
-5.68529844e-01 2.88809121e-01 -3.19200814e-01 1.86665639e-01
5.71472704e-01 8.97801697e-01 7.90029243e-02 -1.46751535e+00
-3.12027633e-01 -3.52487445e-01 -9.12834480e-02 -2.79767424e-01
9.43204105e-01 -4.27983135e-01 -9.26802218e-01 6.09280050e-01
9.34136271e-01 5.55372119e-01 -2.32494786e-01 2.53026277e-01
4.86038774e-01 -1.08553338e+00 -5.75268529e-02 -1.14964199e+00
-2.97091812e-01 5.23820519e-01 -2.07860112e-01 9.17570889e-01
4.30947810e-01 2.75299884e-02 3.92903924e-01 9.46139544e-02
3.16925764e-01 -1.15391886e+00 3.45259517e-01 4.96496320e-01
1.60789549e+00 -9.72045779e-01 2.33332172e-01 -6.89810097e-01
-7.60168433e-01 1.35487294e+00 3.09239775e-01 6.94680912e-03
-1.62420303e-01 -1.39952868e-01 2.70675242e-01 -7.51803160e-01
-7.10115850e-01 5.53742230e-01 5.07648624e-02 5.29464126e-01
7.46851742e-01 4.94831167e-02 -1.50950885e+00 2.93307632e-01
-7.73561001e-01 -5.97957492e-01 1.24504650e+00 8.01478207e-01
-6.45005643e-01 -8.86788368e-01 -1.00081980e+00 1.39444500e-01
-8.22386861e-01 -1.18200600e-01 -1.44776809e+00 1.14969039e+00
-1.93501741e-01 1.48268425e+00 -3.02444458e-01 8.91597569e-02
1.76304132e-02 5.02657294e-02 4.49160606e-01 -2.92145342e-01
-9.83721554e-01 -3.80209386e-01 1.13407695e+00 -2.86469370e-01
-7.54647434e-01 -9.41824675e-01 -6.17353857e-01 -8.30613434e-01
-5.55162847e-01 4.29069549e-01 6.73156142e-01 1.29849935e+00
8.44834819e-02 -7.97485188e-03 2.22502917e-01 -2.09463760e-01
-8.90664458e-01 -1.33251381e+00 -3.43437493e-01 2.22517297e-01
3.00820500e-01 -4.83720958e-01 -7.02055335e-01 -1.28686219e-01] | [8.224410057067871, 10.349774360656738] |
bec47391-6f2f-4c5c-be6c-c94358f1fabf | emotionic-emotional-inertia-and-contagion | 2303.11117 | null | https://arxiv.org/abs/2303.11117v2 | https://arxiv.org/pdf/2303.11117v2.pdf | EmotionIC: Emotional Inertia and Contagion-driven Dependency Modelling for Emotion Recognition in Conversation | Emotion Recognition in Conversation (ERC) has attracted growing attention in recent years as a result of the advancement and implementation of human-computer interface technologies. However, previous approaches to modeling global and local context dependencies lost the diversity of dependency information and do not take the context dependency into account at the classification level. In this paper, we propose a novel approach to dependency modeling driven by Emotional Inertia and Contagion (EmotionIC) for conversational emotion recognition at the feature extraction and classification levels. At the feature extraction level, our designed Identity Masked Multi-head Attention (IM-MHA) captures the identity-based long-distant context in the dialogue to contain the diverse influence of different participants and construct the global emotional atmosphere, while the devised Dialogue-based Gate Recurrent Unit (DialogGRU) that aggregates the emotional tendencies of dyadic dialogue is applied to refine the contextual features with inter- and intra-speaker dependencies. At the classification level, by introducing skip connections in Conditional Random Field (CRF), we elaborate the Skip-chain CRF (SkipCRF) to capture the high-order dependencies within and between speakers, and to emulate the emotional flow of distant participants. Experimental results show that our method can significantly outperform the state-of-the-art models on four benchmark datasets. The ablation studies confirm that our modules can effectively model emotional inertia and contagion. | ['Zhigang Zeng', 'XiaoPing Wang', 'Jiang Li', 'Yingjian Liu'] | 2023-03-20 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [-2.69333005e-01 8.55047554e-02 1.95095748e-01 -8.29303205e-01
-2.17487752e-01 -1.97623104e-01 7.20838785e-01 -2.06776448e-02
-1.53920189e-01 5.04226625e-01 8.67213070e-01 1.69592217e-01
2.83155084e-01 -3.34398031e-01 1.94863379e-01 -7.26811826e-01
-2.25802451e-01 1.63709551e-01 -4.90575671e-01 -5.80986857e-01
1.61849543e-01 2.44773701e-01 -1.47163355e+00 6.69017971e-01
1.00394857e+00 9.23852026e-01 -3.23270410e-02 7.80130625e-01
-4.49613512e-01 1.27648795e+00 -7.78341949e-01 -3.30507100e-01
-5.98597944e-01 -9.32111919e-01 -1.03325558e+00 4.08860408e-02
-5.13896227e-01 1.30220488e-01 2.03020915e-01 5.61557353e-01
7.77645826e-01 3.39386225e-01 4.29132819e-01 -1.18871439e+00
-4.14338678e-01 7.27630794e-01 -2.83562005e-01 8.39020684e-03
7.03553557e-01 -1.72757626e-01 9.93326306e-01 -9.54247236e-01
5.80695152e-01 1.56411088e+00 5.72600842e-01 8.04162621e-01
-8.20554793e-01 -5.21696925e-01 6.75423622e-01 3.86431634e-01
-9.52301919e-01 -3.71043295e-01 1.15749490e+00 -4.07263786e-01
1.34348285e+00 5.38459420e-01 7.32669830e-01 1.56788862e+00
3.20711106e-01 7.30437219e-01 1.28890073e+00 -4.72147375e-01
6.76634759e-02 3.56892109e-01 4.81687158e-01 3.26057792e-01
-9.65999007e-01 -1.94548786e-01 -6.80773795e-01 -2.63708532e-01
1.97906896e-01 -2.38562331e-01 -1.52927577e-01 6.79252982e-01
-6.90464020e-01 1.12641454e+00 1.75866425e-01 7.59586811e-01
-4.83391523e-01 -4.73951876e-01 6.56172693e-01 2.96595484e-01
9.44446802e-01 1.57158643e-01 -4.37904269e-01 -6.00918889e-01
-3.74208421e-01 -1.58785298e-01 1.34268296e+00 8.95001531e-01
6.44195199e-01 -1.66882083e-01 -5.08743286e-01 1.09198797e+00
2.06183314e-01 -1.32306051e-02 6.41695082e-01 -7.21035421e-01
1.48305476e-01 7.77414858e-01 -2.47509614e-01 -9.56242383e-01
-6.52616560e-01 -2.96744764e-01 -1.19643700e+00 -3.12792212e-01
-2.96729743e-01 -6.08206868e-01 -2.09242478e-01 1.87555635e+00
4.86621767e-01 1.16891161e-01 2.40399778e-01 7.65327275e-01
1.09610236e+00 6.98875368e-01 4.76707011e-01 -7.23064721e-01
1.58360362e+00 -1.19341314e+00 -1.40452814e+00 -1.53078124e-01
7.37889111e-01 -8.57546031e-01 8.12739551e-01 2.98991263e-01
-8.25454473e-01 -6.79057777e-01 -6.08685911e-01 -2.75683347e-02
-5.40126503e-01 -4.84141745e-02 8.36840510e-01 5.65146506e-01
-8.27560008e-01 1.87148690e-01 -3.05863827e-01 -2.29559988e-01
-2.27667168e-01 3.50703001e-01 -2.76002020e-01 4.66866583e-01
-1.86530530e+00 1.08492386e+00 -7.50709400e-02 7.63738811e-01
-2.89433092e-01 -1.36261597e-01 -8.92797828e-01 8.56052265e-02
1.87736526e-02 -3.24140489e-01 9.31357622e-01 -1.43039429e+00
-2.13138890e+00 5.38318098e-01 -5.62831342e-01 1.48582160e-01
5.95350116e-02 -3.80937189e-01 -6.55588865e-01 -2.29261652e-01
-4.26239550e-01 3.78956974e-01 4.97602075e-01 -1.19170809e+00
-4.42679316e-01 -1.80078804e-01 -1.45856842e-01 6.33740783e-01
-3.54143649e-01 6.68215036e-01 -3.14885288e-01 -3.73800963e-01
-4.25966114e-01 -8.59234810e-01 -2.27510959e-01 -9.82242286e-01
-3.35038185e-01 -6.62524581e-01 8.30335319e-01 -7.05412745e-01
1.72726846e+00 -2.14726782e+00 4.40052330e-01 1.36918232e-01
-8.59054625e-02 1.53470889e-01 4.41141091e-02 7.51599133e-01
-1.07058614e-01 8.29632133e-02 -1.16366655e-01 -8.98525298e-01
4.90764081e-02 3.80144060e-01 -1.02028564e-01 9.17438790e-02
3.67716879e-01 8.18911850e-01 -7.44162142e-01 -7.22429812e-01
2.93197185e-01 8.81013632e-01 -5.92243731e-01 8.15604210e-01
1.43345013e-01 9.57491338e-01 -4.49043214e-01 3.01427066e-01
4.81637031e-01 5.00278287e-02 4.89590108e-01 3.03727426e-02
-2.69731879e-01 3.09192508e-01 -8.87618661e-01 1.56783140e+00
-1.02668333e+00 3.09976786e-01 4.29018438e-01 -6.45520091e-01
1.17365265e+00 7.46435702e-01 3.79562169e-01 -5.02801657e-01
4.51977670e-01 -1.92959577e-01 4.11784351e-02 -7.91266978e-01
5.38876414e-01 -2.98721105e-01 -6.57691956e-01 3.56589109e-01
2.83898860e-01 2.30636820e-01 -3.55984479e-01 2.24964783e-01
7.10062206e-01 -1.28769472e-01 3.77595693e-01 -1.38598353e-01
1.08570957e+00 -6.27788723e-01 8.65185201e-01 2.54889190e-01
-3.80763203e-01 2.31445372e-01 5.93675494e-01 -2.99675256e-01
-2.77457654e-01 -2.36402765e-01 2.66282223e-02 1.60664797e+00
3.60836536e-02 -6.34848952e-01 -9.23934162e-01 -7.70234704e-01
-5.85292816e-01 6.34552836e-01 -9.01835680e-01 -1.79931447e-01
-6.15454912e-01 -9.89167035e-01 3.81067246e-01 2.95460314e-01
4.52349216e-01 -1.68314350e+00 -2.86721379e-01 5.27486324e-01
-7.94054568e-01 -1.16006362e+00 -4.12211180e-01 3.91808778e-01
-2.84002662e-01 -5.55280268e-01 -1.04913630e-01 -7.63437748e-01
2.37562776e-01 -4.64596808e-01 1.20641601e+00 -6.68987408e-02
-2.82252163e-01 3.16591114e-01 -7.72422433e-01 -1.11641176e-01
-5.05038381e-01 8.38403851e-02 -1.75611272e-01 5.30629039e-01
5.79493761e-01 -6.85247540e-01 -4.23225582e-01 3.70766550e-01
-5.26506424e-01 6.67376891e-02 3.02905023e-01 8.98924232e-01
-7.38986582e-02 -2.15839893e-01 9.72313285e-01 -9.53118622e-01
1.10856140e+00 -6.54191375e-01 3.77571613e-01 3.01860631e-01
-3.02703649e-01 -8.85725394e-02 6.87828124e-01 -6.36114478e-01
-1.84687161e+00 -1.44357517e-01 -4.63206887e-01 4.61445041e-02
-4.77782100e-01 5.94260335e-01 -5.36822140e-01 3.57693344e-01
1.81118339e-01 -1.22606322e-01 -2.74470061e-01 -3.45948756e-01
4.81696695e-01 1.22163641e+00 2.22921684e-01 -6.86980665e-01
-1.90373436e-01 1.49901971e-01 -4.47420478e-01 -9.60249543e-01
-8.71623278e-01 -5.87343514e-01 -7.05751359e-01 -7.67374039e-01
1.22098434e+00 -9.89726782e-01 -1.04291964e+00 5.11448503e-01
-1.47211266e+00 -2.02065334e-01 -1.69532578e-02 4.78830785e-01
-2.92965472e-01 2.71409661e-01 -1.06183517e+00 -1.44405973e+00
-6.77513719e-01 -9.05527472e-01 9.65104997e-01 4.39781100e-01
-6.06905878e-01 -1.27531111e+00 2.84713507e-01 2.73802608e-01
5.43201149e-01 4.41184551e-01 7.66476393e-01 -7.99701631e-01
2.34867841e-01 -3.40736322e-02 2.34105393e-01 4.23178315e-01
7.42356405e-02 -1.26543855e-02 -1.32102835e+00 2.81597763e-01
3.52363199e-01 -4.99209553e-01 4.28758085e-01 -1.73628889e-02
7.36292362e-01 -2.30604544e-01 -3.63942459e-02 1.31019965e-01
8.56926382e-01 1.89102948e-01 6.50536776e-01 -3.21320415e-01
6.67226791e-01 1.27532327e+00 6.18061543e-01 8.73006463e-01
7.23716378e-01 5.70909321e-01 1.15776710e-01 -2.55525649e-01
2.63731152e-01 2.88062338e-02 6.36988878e-01 1.69870341e+00
-3.38758618e-01 -2.15086490e-01 -4.19757038e-01 2.86657631e-01
-2.11284423e+00 -1.04426718e+00 -4.47950631e-01 1.58551061e+00
1.04422677e+00 -1.60695106e-01 -5.92735223e-02 -2.05686435e-01
8.80014420e-01 3.08474302e-01 -4.61492836e-02 -1.28172600e+00
-3.14215630e-01 4.11211178e-02 -4.84540969e-01 9.82322216e-01
-8.44829559e-01 1.08765352e+00 5.19319582e+00 7.33205497e-01
-8.95680785e-01 2.64145017e-01 8.36701214e-01 1.59185916e-01
-1.72442049e-01 -5.28447405e-02 -9.06518996e-01 3.91892999e-01
1.11413109e+00 3.28879088e-01 3.88241708e-01 7.09111452e-01
3.33878785e-01 -7.90075287e-02 -9.33230698e-01 7.94195592e-01
2.94253021e-01 -3.96574080e-01 -4.53678608e-01 -4.55148378e-03
4.59322631e-01 -3.72221112e-01 -2.72124469e-01 8.45522106e-01
2.58135885e-01 -1.01058853e+00 2.19796166e-01 9.01606500e-01
4.43956614e-01 -9.21330273e-01 1.02337003e+00 5.47992706e-01
-1.33606553e+00 -8.72320607e-02 2.40939744e-02 -3.42790663e-01
4.01903510e-01 5.51701605e-01 -5.85565090e-01 7.54824281e-01
5.28000891e-01 5.97897232e-01 -3.18899691e-01 -2.64430530e-02
-3.29776525e-01 7.33029604e-01 -1.65634871e-01 -3.79230589e-01
2.33680010e-01 -3.62728298e-01 4.01150107e-01 1.91778481e+00
-1.22553408e-01 3.26121241e-01 9.98103619e-02 6.05860293e-01
2.01970443e-01 4.34643805e-01 -1.73434675e-01 2.52472758e-01
3.24911445e-01 1.82009256e+00 -3.59336346e-01 -2.80724108e-01
-3.98927331e-01 1.33763373e+00 5.29340804e-01 2.70678550e-01
-9.49889719e-01 -3.51806492e-01 5.02133489e-01 -7.30850339e-01
9.87056792e-02 5.57910874e-02 -2.38666404e-02 -1.21112263e+00
-4.78039756e-02 -8.57696533e-01 3.46778095e-01 -2.66124636e-01
-1.62356865e+00 1.12878358e+00 -2.51794934e-01 -7.36201525e-01
-4.81233776e-01 -1.03246272e-01 -1.10979283e+00 1.20494568e+00
-1.40010297e+00 -1.37603509e+00 -3.87291223e-01 8.10146332e-01
6.47412717e-01 1.91172972e-01 1.42495799e+00 1.79471537e-01
-8.11336517e-01 4.72783148e-01 -5.04124582e-01 1.38075188e-01
8.82167876e-01 -1.15460753e+00 -1.00933202e-01 3.49133283e-01
-3.95642966e-01 6.33305609e-01 5.36898553e-01 -4.65499997e-01
-9.82649744e-01 -7.57396221e-01 1.70546615e+00 -2.58262694e-01
5.15852571e-01 -9.32887554e-01 -9.97867048e-01 4.03866708e-01
7.49543786e-01 -1.86939463e-01 1.23605072e+00 8.50227416e-01
-1.93830788e-01 1.52200609e-01 -8.97220016e-01 4.28878576e-01
8.23985040e-01 -7.78227210e-01 -6.51191235e-01 -7.45109320e-02
7.25987017e-01 -1.98964775e-02 -1.17505121e+00 3.56104523e-01
4.98764485e-01 -1.22916567e+00 4.15022373e-01 -4.12887216e-01
4.21040624e-01 1.68101683e-01 3.91248660e-03 -1.24420631e+00
-3.48509789e-01 -1.03504527e+00 -1.51742727e-01 1.94694507e+00
2.31135249e-01 -5.05814910e-01 1.61798567e-01 8.64336133e-01
-1.35498568e-01 -8.34815979e-01 -8.75377536e-01 1.34432437e-02
-3.08702830e-02 -3.57425332e-01 4.61601317e-01 1.31262290e+00
8.63612056e-01 1.09035850e+00 -8.37976813e-01 -1.40184283e-01
-9.48769599e-02 2.53306657e-01 4.26453590e-01 -1.26764238e+00
-3.36183280e-01 -4.17018235e-01 1.10579893e-01 -8.47617090e-01
6.45435572e-01 -5.40103197e-01 2.30081603e-01 -1.20409298e+00
8.03722441e-02 -3.11704606e-01 -2.33000174e-01 1.32462785e-01
-5.71987867e-01 -2.75515914e-01 9.63516384e-02 -1.25164449e-01
-9.90845799e-01 1.12656295e+00 9.82983708e-01 3.99720490e-01
-5.50194204e-01 -3.61573882e-02 -6.69108987e-01 7.27020860e-01
7.94158518e-01 -2.34750539e-01 -4.22502868e-02 2.52016842e-01
2.11794928e-01 3.22494894e-01 -7.48788565e-02 -6.13047242e-01
2.87645817e-01 4.04417180e-02 1.70953244e-01 -4.23280567e-01
5.34575224e-01 -7.48961031e-01 -9.20332372e-02 -8.18416402e-02
-5.68308353e-01 -5.26642613e-02 -6.69013336e-02 3.96466732e-01
-5.16759098e-01 3.08533162e-02 5.33047438e-01 -3.96562256e-02
-2.63333589e-01 -8.66897926e-02 -8.30846548e-01 -3.92872691e-02
7.03692079e-01 3.64074677e-01 -6.04842603e-02 -6.41209006e-01
-1.00448763e+00 4.20481205e-01 -2.54586101e-01 6.74324393e-01
3.22712570e-01 -1.18644238e+00 -7.51434445e-01 1.86887041e-01
-2.29383241e-02 -2.92403251e-01 7.59898245e-01 8.77718806e-01
3.76689643e-01 1.69337764e-01 -1.57444049e-02 -3.13138604e-01
-1.56977773e+00 2.83784181e-01 3.77562940e-01 -8.90788913e-01
-2.14383602e-01 9.73958194e-01 3.18189263e-01 -6.41588807e-01
4.26574707e-01 -1.43530741e-01 -8.23710382e-01 4.96543497e-01
4.13551867e-01 1.94729283e-01 -2.17961594e-01 -1.03943467e+00
-3.52054417e-01 3.28176469e-01 -6.66166916e-02 -2.08105102e-01
1.18591797e+00 -6.45309567e-01 -4.57785696e-01 8.89166176e-01
1.28455555e+00 1.15345813e-01 -1.03629696e+00 -4.48474959e-02
-2.58223969e-03 2.43384182e-01 -1.52729884e-01 -9.53184724e-01
-6.14515543e-01 1.08928514e+00 2.28254884e-01 5.05620956e-01
1.19728804e+00 -6.90437332e-02 6.69609010e-01 -4.30494472e-02
-3.15711871e-02 -1.47232974e+00 9.48817178e-04 1.04159939e+00
1.06529546e+00 -1.05172336e+00 -3.76626611e-01 -5.49947798e-01
-1.40565467e+00 1.01505005e+00 7.79666126e-01 1.53020084e-01
8.68674457e-01 4.85415936e-01 3.17869246e-01 -2.03872934e-01
-1.26231611e+00 -2.61402518e-01 1.15251556e-01 2.70727098e-01
8.89326811e-01 1.48567423e-01 -4.76600647e-01 1.31918466e+00
-1.32478803e-01 -3.96475881e-01 9.41633433e-02 8.10840130e-01
-1.25780702e-01 -1.25441408e+00 -1.36996970e-01 -2.41019741e-01
-5.18106461e-01 -1.73653200e-01 -7.87017465e-01 4.44976568e-01
5.35324275e-01 1.55416822e+00 5.90752400e-02 -6.24412239e-01
3.07692558e-01 6.23096168e-01 -6.33052215e-02 -4.29721117e-01
-1.65137661e+00 4.38829035e-01 6.28589571e-01 -5.40995300e-01
-9.00639892e-01 -4.67592150e-01 -1.29727125e+00 8.79851878e-02
-5.46328843e-01 6.32265806e-01 6.51509941e-01 1.07352197e+00
6.35661721e-01 6.60395622e-01 1.34632182e+00 -8.46413910e-01
4.70025167e-02 -1.59074330e+00 -4.94321853e-01 6.44165039e-01
1.48618311e-01 -3.28327328e-01 -7.22347140e-01 -7.07269832e-02] | [13.015625, 6.080755710601807] |
935e5008-7bc5-43ed-a03c-a9439b1c5e5b | generating-complement-data-for-aspect-term | null | null | https://aclanthology.org/2022.deeplo-1.21 | https://aclanthology.org/2022.deeplo-1.21.pdf | Generating Complement Data for Aspect Term Extraction with GPT-2 | t | ['Bonan Min and Thien Huu Nguyen', 'Franck Dernoncourt', 'Amir Pouran Ben Veyseh'] | null | null | null | null | deeplo-2022-7 | ['term-extraction'] | ['natural-language-processing'] | [ 1.89939722e-01 2.36788392e-01 -5.91508627e-01 -3.67945790e-01
-5.24475813e-01 -8.09045732e-01 2.93398023e-01 -7.17031658e-01
-1.54879212e-01 7.98508823e-01 -1.99629426e-01 -6.90531909e-01
-3.04593027e-01 -7.58767843e-01 -8.23488533e-01 -8.08756053e-01
-9.00167286e-01 6.62851155e-01 2.46676758e-01 -2.83387691e-01
4.47858244e-01 2.21166894e-01 -1.51938629e+00 8.17170560e-01
1.13089287e+00 1.70284688e+00 2.96642125e-01 1.01283693e+00
-3.35448980e-01 1.18994772e+00 -6.46863997e-01 -5.62546670e-01
4.42817062e-01 -3.37279588e-01 -3.64081897e-02 2.43813559e-01
2.51474857e-01 5.70869893e-02 -3.71525168e-01 9.90638435e-01
4.43276435e-01 1.10163748e-01 1.26252306e+00 -1.44224024e+00
-5.91280758e-01 8.18176508e-01 -4.49393660e-01 -8.83459449e-02
6.28999770e-01 -4.56945032e-01 8.94929349e-01 -1.37190151e+00
8.43617499e-01 1.36689758e+00 1.22254932e+00 2.10383818e-01
-1.14468801e+00 -1.39829740e-01 -7.96726942e-02 1.94943070e-01
-1.05240738e+00 -1.36237845e-01 1.78917408e-01 -2.14169145e-01
1.76460838e+00 1.00378203e+00 1.25295269e+00 1.01856434e+00
2.06551719e+00 1.44298208e+00 1.37016428e+00 -5.20172536e-01
3.43809366e-01 5.18362224e-02 3.77140969e-01 9.10886601e-02
4.43431616e-01 4.23922777e-01 -7.34497309e-01 -5.10606706e-01
8.03467691e-01 -3.21181059e-01 2.86919504e-01 -3.37324262e-01
-1.12428963e+00 6.33509457e-01 4.84345645e-01 4.14401531e-01
-6.08334422e-01 3.44729632e-01 6.13652766e-01 7.06705570e-01
2.46248499e-01 4.65929121e-01 -7.17349231e-01 1.30878568e-01
-5.55795729e-01 4.92499024e-01 8.54222119e-01 1.65030825e+00
-2.28135171e-03 3.22754562e-01 -3.00849706e-01 5.18211365e-01
4.64348763e-01 1.17489815e+00 5.35341680e-01 -1.24818802e+00
9.06072035e-02 -5.67485159e-03 2.47882321e-01 -3.14330548e-01
-2.58001894e-01 -1.64535791e-01 -9.63801563e-01 -1.68291673e-01
3.79071563e-01 -2.28124261e-01 -1.06214964e+00 7.67602026e-01
1.45905241e-01 -7.43074492e-02 3.54180247e-01 8.46829042e-02
4.99283522e-01 4.09673303e-01 -2.05288887e-01 -6.91714466e-01
1.13712680e+00 -1.02304780e+00 -1.74066412e+00 1.15602694e-01
6.87205791e-01 -1.20829093e+00 2.33131707e-01 4.04619932e-01
-1.45673180e+00 -1.05888166e-01 -7.73380756e-01 6.06836192e-02
-5.02120912e-01 -1.83962464e-01 7.62394726e-01 1.54058015e+00
-1.46052659e+00 6.79608524e-01 -3.44809204e-01 -3.08380157e-01
-1.44321144e-01 7.52548873e-01 1.08459078e-01 4.72643852e-01
-1.52220440e+00 1.15118182e+00 6.20099753e-02 4.03061062e-01
-7.46870100e-01 -1.95691302e-01 -7.15343952e-01 -6.12893820e-01
5.06622083e-02 -6.41239464e-01 1.46449745e+00 -7.97403008e-02
-1.79270327e+00 6.65626347e-01 -1.52741894e-01 -4.20509309e-01
7.43422925e-01 -2.87544847e-01 -4.89737481e-01 8.68125036e-02
2.21662924e-01 4.92182225e-01 1.13331234e+00 -1.18401015e+00
-5.70588529e-01 -6.98563397e-01 -3.59988093e-01 3.42236400e-01
-5.86456545e-02 3.35361689e-01 4.22853947e-01 -2.68302470e-01
4.78362650e-01 -9.71662700e-01 -4.87820297e-01 -4.26882505e-01
-4.63944227e-01 -3.00730914e-01 9.90977943e-01 -2.64651239e-01
1.16940653e+00 -1.55490828e+00 -3.26942414e-01 2.36096203e-01
6.82533205e-01 -1.45345451e-02 2.85065323e-02 8.53169143e-01
-3.46390486e-01 6.08425677e-01 3.65286261e-01 1.49751291e-01
3.29615355e-01 4.97534513e-01 -3.74460310e-01 3.38356137e-01
-2.27232222e-02 1.49712217e+00 -7.99672306e-01 -2.79272586e-01
3.97615105e-01 1.03713125e-01 -3.34000736e-01 2.99901992e-01
7.53000155e-02 3.70438099e-01 -6.72875583e-01 8.12499821e-01
8.88974011e-01 1.64589241e-01 4.26477402e-01 4.60821003e-01
-6.40049696e-01 3.99733156e-01 -1.41205534e-01 7.67119408e-01
-6.27802089e-02 5.45784891e-01 2.60998487e-01 -1.06391418e+00
8.65994453e-01 9.31548119e-01 6.42848074e-01 -1.05545890e+00
3.91279310e-01 7.14037895e-01 1.75795510e-01 -8.66502821e-01
4.40493941e-01 -2.36573577e-01 -1.74578980e-01 7.50014305e-01
-4.71572518e-01 -9.29108083e-01 -2.22104475e-01 8.92405137e-02
6.42322004e-01 -1.08119011e-01 2.27961838e-01 -7.36103833e-01
2.39393353e-01 -2.65759289e-01 1.81378976e-01 1.24003875e+00
-2.84355670e-01 -2.02037562e-02 5.22130013e-01 -7.00815678e-01
-9.64008629e-01 -7.65356004e-01 -5.02605259e-01 1.20296323e+00
1.85126424e-01 4.27732915e-02 -1.11524069e+00 -1.79784060e-01
2.68691272e-01 7.90132284e-01 -1.00782096e+00 1.53625146e-01
-6.94981039e-01 -8.79014075e-01 4.97215033e-01 3.19848984e-01
3.55103493e-01 -1.19258249e+00 -5.45660317e-01 1.55578017e-01
-3.89463156e-01 -8.42305481e-01 -6.66711032e-01 7.47166932e-01
-1.63723433e+00 -4.36473489e-01 -5.73449612e-01 -8.55272233e-01
5.81257343e-01 2.04227880e-01 6.50495768e-01 -3.55152130e-01
-1.13953702e-01 4.77806658e-01 -2.22404659e-01 -6.19727969e-01
-1.25880420e-01 -1.67800546e-01 8.95826161e-01 -2.56743848e-01
3.48995835e-01 -1.70074806e-01 -4.69268560e-01 8.06690693e-01
-4.84402537e-01 -2.17795849e-01 2.55734891e-01 9.40218151e-01
3.92726898e-01 1.67035088e-01 6.86087683e-02 -5.37814200e-01
8.21058452e-01 8.68108943e-02 2.91677509e-02 5.25601387e-01
-3.92417401e-01 -4.90596265e-01 3.27895433e-01 -5.36372006e-01
-1.16805279e+00 -7.46271908e-01 1.66443944e-01 -2.42259562e-01
3.42849970e-01 1.81951568e-01 9.71686617e-02 -4.83085275e-01
5.89464605e-01 3.20968062e-01 9.88630354e-02 -1.16144948e-01
3.37989181e-01 3.69506598e-01 9.12515521e-02 -8.24873209e-01
9.76583540e-01 1.89692751e-02 -1.83379762e-02 -1.00814426e+00
-2.53604561e-01 3.94363970e-01 -1.02228272e+00 -9.48803008e-01
7.10936248e-01 -3.10328454e-01 -8.92485619e-01 4.92992878e-01
-7.89266646e-01 -3.16716403e-01 -4.76785332e-01 5.75828910e-01
-1.07606995e+00 5.41701764e-02 -5.11994660e-01 -1.39143908e+00
-4.73697394e-01 -1.26493227e+00 9.89715815e-01 -1.39136344e-01
-1.54875562e-01 -1.21094310e+00 -1.32906422e-01 4.70664620e-01
2.99690098e-01 -1.91290736e-01 8.20965707e-01 -3.06973726e-01
-4.85169500e-01 -2.96215147e-01 4.13204022e-02 -1.51869848e-01
-2.22846996e-02 3.07069451e-01 -1.12824881e+00 -5.63265979e-01
8.34414423e-01 -1.30536035e-01 7.85833970e-02 1.26730096e+00
8.79319489e-01 -5.13120413e-01 -9.46984291e-01 5.44356763e-01
1.07217348e+00 7.16077149e-01 5.12038112e-01 5.47985613e-01
3.31056505e-01 6.74696505e-01 1.27254450e+00 2.77839363e-01
1.52111277e-01 6.04227126e-01 1.46246269e-01 2.76763052e-01
6.26627356e-02 -2.96620935e-01 7.40217209e-01 1.31125784e+00
-5.78064144e-01 -1.47077024e-01 -7.75543690e-01 2.83217520e-01
-1.77835071e+00 -1.21642387e+00 -4.40115660e-01 1.28189552e+00
4.62879270e-01 2.91927338e-01 -1.63142815e-01 1.46978900e-01
1.01554763e+00 -6.46483719e-01 -6.52676344e-01 -8.19531202e-01
-1.52674362e-01 3.66030306e-01 7.62685061e-01 6.91384792e-01
-5.89482367e-01 7.66647577e-01 1.24491730e+01 1.35650063e+00
-4.08224821e-01 1.57630906e-01 1.04111016e+00 1.60436973e-01
-6.93065405e-01 4.70289141e-02 -7.55821705e-01 -1.31277904e-01
1.37683690e+00 -8.06113422e-01 2.59953141e-01 5.14261305e-01
7.27152526e-01 -3.28785270e-01 -1.38832510e+00 8.98782074e-01
8.18762705e-02 -1.14380479e+00 8.83668195e-03 6.76908970e-01
5.90522349e-01 -5.20515859e-01 9.83852625e-01 7.18544185e-01
3.16567928e-01 -1.10500252e+00 1.18558323e+00 -1.67055085e-01
1.03510225e+00 -5.16323447e-01 3.53999943e-01 8.44747052e-02
-1.53256738e+00 -3.83495122e-01 -6.71874464e-01 -8.42298985e-01
1.54422939e-01 1.09240323e-01 -9.35747921e-01 5.74631810e-01
8.30331504e-01 5.26373088e-01 -1.85087740e-01 1.35962188e+00
3.89768928e-02 -1.67607114e-01 -4.34347719e-01 -4.67691749e-01
5.60756147e-01 -3.22952360e-01 7.80376434e-01 3.63160014e-01
3.95563036e-01 2.80316412e-01 3.24063212e-01 3.38242620e-01
5.63331723e-01 1.12846196e-01 -1.23109615e+00 -6.14734478e-02
2.60081440e-01 9.88807976e-01 -7.41849899e-01 -3.81950378e-01
3.41136344e-02 4.15943474e-01 -3.99838723e-02 4.65144902e-01
-7.12056100e-01 -4.13732678e-01 4.97172683e-01 1.19220920e-01
2.13116761e-02 -1.61851928e-01 -6.81930363e-01 -9.26758528e-01
-5.51150322e-01 -5.56569815e-01 -4.99655418e-02 -7.54947960e-01
-1.50612307e+00 5.65921009e-01 -4.40481193e-02 -1.51991355e+00
-3.55551273e-01 -1.44982708e+00 -1.97954893e-01 5.81557564e-02
-1.04040527e+00 -1.01083434e+00 4.30586785e-01 3.97105217e-01
1.18066512e-01 -3.80264312e-01 7.94512272e-01 -5.24258316e-02
-4.35840219e-01 7.79594898e-01 5.88282585e-01 -7.14952052e-01
7.36396372e-01 -1.22806084e+00 5.91358721e-01 -6.73428923e-02
-6.68152094e-01 1.11977923e+00 6.27602398e-01 -8.84531081e-01
-1.92270243e+00 -6.83907509e-01 9.71550167e-01 -8.10284674e-01
9.20165479e-01 -2.55020797e-01 -2.20517814e-02 1.02859819e+00
4.19887543e-01 -7.33902276e-01 8.60386431e-01 -2.09508836e-01
1.13460578e-01 2.91082412e-01 -1.47463858e+00 4.50772256e-01
1.24246180e+00 -7.74584830e-01 -6.38555288e-01 7.79140413e-01
6.31502688e-01 -7.08490431e-01 -1.90750539e+00 3.46303344e-01
1.11826324e+00 -7.66655982e-01 1.19824612e+00 -6.98251009e-01
-1.20459408e-01 3.85222226e-01 -1.81850567e-01 -7.63292313e-01
-1.88167617e-01 -1.32464731e+00 2.45678186e-01 -1.71182334e-01
4.30979103e-01 -1.34319079e+00 2.72918314e-01 1.34494781e+00
-4.65472609e-01 -3.81976753e-01 -1.53015709e+00 -1.27737558e+00
2.33196318e-01 -2.97680706e-01 4.64499563e-01 9.34277773e-01
7.60169208e-01 6.05871193e-02 -1.76320747e-01 -3.72492880e-01
1.04645765e+00 5.77011406e-02 6.22676134e-01 -1.18750143e+00
6.15142763e-01 -6.13965869e-01 1.02366067e-01 -1.26336491e+00
-8.41216445e-02 -8.93945873e-01 -1.95095018e-01 -9.06829536e-01
1.67432263e-01 -3.94941568e-01 -6.44420758e-02 1.92613695e-02
4.16199535e-01 2.03177944e-01 -1.06396683e-01 3.92603129e-01
-3.37048769e-01 9.08262074e-01 2.01500106e+00 -2.59478718e-01
-5.11344187e-02 6.61453679e-02 -2.66384125e-01 3.95984024e-01
-2.30289534e-01 -3.95135969e-01 -7.05061495e-01 -3.04246873e-01
7.87112340e-02 5.15138507e-01 -2.12846085e-01 -5.99075973e-01
4.64595407e-01 -3.36401969e-01 4.13682073e-01 -1.32351279e+00
2.80695796e-01 -9.62273777e-01 1.61175415e-01 7.88183510e-01
8.82592872e-02 1.95026442e-01 1.81165382e-01 -6.48092479e-04
-1.07943185e-01 -4.37374592e-01 6.80991828e-01 -6.03318632e-01
-1.68771401e-01 2.71748930e-01 -1.14019620e+00 -1.94603875e-01
1.04882240e+00 -9.21916187e-01 -4.94318366e-01 -3.26782256e-01
-1.16711247e+00 3.39655548e-01 5.61399519e-01 5.75944036e-02
6.26476705e-01 -1.59493732e+00 -2.11476997e-01 3.13548356e-01
-2.84545928e-01 -5.67733526e-01 1.58858210e-01 1.19876039e+00
-7.24584281e-01 1.01518333e+00 -6.64929450e-01 -6.47873163e-01
-9.36170340e-01 5.97725570e-01 5.99120677e-01 6.84380159e-02
-5.94326317e-01 6.88359976e-01 3.33671093e-01 -7.42121279e-01
-1.36155412e-01 -4.42581415e-01 -6.13011062e-01 -4.30494286e-02
5.85559309e-01 9.41894054e-01 -2.40120575e-01 -6.89351737e-01
-4.56086159e-01 9.60907876e-01 -1.62124112e-01 -3.37578297e-01
1.21202886e+00 -2.27867439e-01 -9.30782139e-01 6.21266186e-01
4.74122018e-01 -7.73310363e-01 -3.10821533e-01 3.54538292e-01
1.20676249e-01 -1.14684784e+00 -2.67380357e-01 -1.44833967e-01
-5.49234569e-01 4.87696886e-01 5.77761650e-01 8.17515671e-01
8.36860001e-01 -4.01973456e-01 9.73162472e-01 9.70731914e-01
7.90355146e-01 -1.47186983e+00 -2.99316552e-02 6.47735655e-01
9.40879166e-01 -4.81309980e-01 1.09276883e-01 -1.00248826e+00
-7.41886795e-01 1.05813515e+00 2.39938542e-01 -1.38619132e-02
6.96265221e-01 6.73565686e-01 1.58846304e-01 -2.46428266e-01
-1.17298937e+00 1.33114040e-01 1.93507865e-01 1.36494064e+00
7.54637897e-01 3.84633869e-01 -4.50943470e-01 8.45573321e-02
-7.78585136e-01 -1.94481492e-01 7.95102298e-01 1.10179341e+00
-4.91532892e-01 -1.44410038e+00 -9.94438827e-01 5.72424948e-01
-2.58122474e-01 1.63886824e-03 -4.46063191e-01 5.63656151e-01
-4.28137332e-01 1.50927961e+00 -2.40980104e-01 -3.95527095e-01
3.20507526e-01 3.32847089e-01 9.46980178e-01 -8.80468413e-02
-1.07144022e+00 9.68575954e-01 2.36820653e-01 -9.15274799e-01
-7.65339196e-01 -1.29927778e+00 -9.04530048e-01 -1.01017177e+00
-7.32975006e-01 1.16703488e-01 3.54904890e-01 6.89339340e-01
-3.98848087e-01 1.16995655e-01 9.99229252e-01 -1.34482634e+00
-5.66089638e-02 -9.86962140e-01 -1.04934478e+00 -3.15998435e-01
4.39148545e-01 -1.11476994e+00 -5.48778951e-01 -2.57884592e-01] | [-7.271073818206787, 3.8188908100128174] |
57da3ff4-36b4-428e-b330-188bded37248 | ltc-se-expanding-the-potential-of-liquid-time | 2304.08691 | null | https://arxiv.org/abs/2304.08691v1 | https://arxiv.org/pdf/2304.08691v1.pdf | LTC-SE: Expanding the Potential of Liquid Time-Constant Neural Networks for Scalable AI and Embedded Systems | We present LTC-SE, an improved version of the Liquid Time-Constant (LTC) neural network algorithm originally proposed by Hasani et al. in 2021. This algorithm unifies the Leaky-Integrate-and-Fire (LIF) spiking neural network model with Continuous-Time Recurrent Neural Networks (CTRNNs), Neural Ordinary Differential Equations (NODEs), and bespoke Gated Recurrent Units (GRUs). The enhancements in LTC-SE focus on augmenting flexibility, compatibility, and code organization, targeting the unique constraints of embedded systems with limited computational resources and strict performance requirements. The updated code serves as a consolidated class library compatible with TensorFlow 2.x, offering comprehensive configuration options for LTCCell, CTRNN, NODE, and CTGRU classes. We evaluate LTC-SE against its predecessors, showcasing the advantages of our optimizations in user experience, Keras function compatibility, and code clarity. These refinements expand the applicability of liquid neural networks in diverse machine learning tasks, such as robotics, causality analysis, and time-series prediction, and build on the foundational work of Hasani et al. | ['Hamdan Abdellatef', 'Ferhat Atasoy', 'Michael Bidollahkhani'] | 2023-04-18 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [-2.96674222e-02 -2.59479195e-01 1.45718932e-01 -3.79531421e-02
1.17367491e-01 -5.99264562e-01 4.17231590e-01 -5.69416463e-01
-5.95196426e-01 7.43513167e-01 -2.00733811e-01 -5.25860846e-01
-1.83549240e-01 -5.03107786e-01 -5.53457379e-01 -9.72013772e-01
-4.41188514e-01 -1.78358018e-01 5.27425230e-01 -2.32558116e-01
2.55849749e-01 8.11151683e-01 -1.46229327e+00 3.04255188e-01
2.46390268e-01 1.21343863e+00 -7.58282514e-03 6.19551718e-01
1.57219216e-01 1.12017179e+00 -3.00286915e-02 1.10411808e-01
-8.22006240e-02 -3.04508328e-01 -3.22353840e-01 -1.11479020e+00
-2.65372306e-01 8.98531228e-02 -7.32259393e-01 5.97411990e-01
3.94402862e-01 5.39824367e-02 4.75949615e-01 -1.62373352e+00
-6.74845338e-01 7.74279535e-01 1.44698750e-02 4.75187361e-01
-4.81549054e-01 3.75445813e-01 6.72611117e-01 -8.56076479e-01
5.36341190e-01 1.04316497e+00 1.30442512e+00 9.26745892e-01
-1.20708668e+00 -8.43325436e-01 2.44628266e-01 -9.63614658e-02
-1.23080266e+00 -6.21120572e-01 2.64603823e-01 -2.95152009e-01
2.06445837e+00 -8.82451534e-02 6.86439157e-01 1.14220881e+00
6.78898811e-01 7.53732562e-01 6.59458101e-01 7.19946027e-02
6.43726468e-01 -5.18031001e-01 6.24273360e-01 4.94027168e-01
2.67772704e-01 1.60445049e-01 -6.28040493e-01 -2.06320718e-01
1.14368403e+00 7.90590644e-02 -2.26121813e-01 -6.63487911e-02
-1.18664026e+00 3.60490501e-01 4.90575463e-01 5.02275229e-01
-4.36628520e-01 9.99458253e-01 6.28206909e-01 2.34732330e-01
1.12868078e-01 3.78344804e-01 -4.72908944e-01 -3.16323340e-01
-7.07753003e-01 4.32976574e-01 9.91038024e-01 9.43695307e-01
3.29881370e-01 8.43175530e-01 1.71746373e-01 4.34650004e-01
2.24060640e-01 4.57861155e-01 7.23449171e-01 -1.47789192e+00
-9.89534184e-02 3.06816339e-01 -2.01809347e-01 -6.17489457e-01
-7.20731914e-01 -6.15657270e-01 -9.20452058e-01 4.41671520e-01
-1.41334336e-03 -1.86757803e-01 -7.71684706e-01 1.77621174e+00
-5.67378759e-01 2.41360769e-01 3.19147855e-01 5.84128141e-01
5.96621573e-01 6.00639641e-01 4.27596420e-02 -1.05854079e-01
6.71317577e-01 -7.91599810e-01 -6.95885181e-01 -4.85047735e-02
7.91419327e-01 -1.47535250e-01 4.74206746e-01 3.05059791e-01
-1.03517437e+00 -1.26543745e-01 -1.10180998e+00 -1.75077468e-01
-7.08216846e-01 -1.61008567e-01 9.33496416e-01 4.32518452e-01
-1.60078645e+00 9.74148750e-01 -1.46081209e+00 -2.95705110e-01
2.96156615e-01 8.61906648e-01 -5.40011041e-02 7.86754429e-01
-8.98388922e-01 8.36471617e-01 2.74128854e-01 2.52976507e-01
-4.14879352e-01 -5.29169977e-01 -4.23713833e-01 1.29865119e-02
-1.75410539e-01 -5.13879001e-01 1.30271256e+00 -5.42577684e-01
-1.63698757e+00 5.33223510e-01 -1.90031618e-01 -1.12849081e+00
-1.76259920e-01 3.40788513e-02 -3.63608450e-01 -3.61534096e-02
-2.96089977e-01 8.00837159e-01 5.11945374e-02 -6.39273763e-01
-2.19437346e-01 -1.70995325e-01 -5.09308159e-01 -1.91411331e-01
4.43843706e-03 1.12939410e-01 -8.83467346e-02 -6.94504797e-01
3.91373247e-01 -9.99784052e-01 -2.36450449e-01 3.40068698e-01
2.05531150e-01 -2.11117178e-01 1.09883332e+00 -2.82913774e-01
1.14138269e+00 -2.37696791e+00 7.07929358e-02 1.53997481e-01
3.52864653e-01 2.90029019e-01 -7.79369324e-02 2.68864691e-01
-2.94755310e-01 7.15501830e-02 -2.41935760e-01 -2.36956522e-01
-1.88018769e-01 3.43668431e-01 -6.69668138e-01 3.88758779e-01
2.42878959e-01 1.23379159e+00 -6.29055262e-01 9.33348015e-02
-2.52054602e-01 5.10328650e-01 -4.58216101e-01 -3.00703198e-01
-1.47692427e-01 1.40264019e-01 -2.78320879e-01 6.33621573e-01
1.87785625e-01 -3.39520425e-01 -2.31119454e-01 -8.87018964e-02
-8.60756755e-01 4.60099727e-01 -8.58256876e-01 1.34270096e+00
-1.56710193e-01 1.02017832e+00 -3.46989669e-02 -7.80002475e-01
1.11397600e+00 4.71664548e-01 5.80165386e-01 -8.49707723e-01
2.23294109e-01 7.45248914e-01 1.19288743e-01 -1.49525926e-01
2.94792593e-01 1.81716442e-01 1.25436157e-01 7.47237623e-01
4.63134825e-01 9.07516032e-02 1.47683591e-01 1.84364289e-01
1.35610557e+00 5.12713611e-01 -7.18166307e-02 -4.20711040e-01
-2.23514363e-01 -4.38875526e-01 5.71630180e-01 7.79774427e-01
-3.28485519e-01 4.95776623e-01 7.51686335e-01 -3.49972039e-01
-1.13188493e+00 -9.88556802e-01 -1.19674765e-01 7.89491415e-01
-2.13596553e-01 -3.35684001e-01 -6.13304198e-01 3.59159708e-01
-1.40656739e-01 6.59460604e-01 -6.07010543e-01 -2.99239784e-01
-5.43531537e-01 -8.72350752e-01 1.37609935e+00 8.14867556e-01
5.12647033e-01 -1.52954280e+00 -1.28060520e+00 5.66170216e-01
1.73377156e-01 -9.02414620e-01 1.46478694e-02 1.05353522e+00
-1.18767977e+00 -6.68402851e-01 -4.37879264e-01 -7.44839430e-01
-1.87843386e-03 3.23228762e-02 7.21192718e-01 -2.07960278e-01
-4.38999325e-01 1.94272593e-01 1.22188710e-01 -9.60628763e-02
1.77339628e-01 -1.24281108e-01 4.02411312e-01 -3.80318403e-01
2.77949989e-01 -1.07570159e+00 -3.05747211e-01 1.00202575e-01
-8.48076344e-01 2.69717246e-01 2.37184055e-02 7.05674112e-01
6.14179492e-01 -2.84934640e-01 6.42932534e-01 -5.01744986e-01
4.82992291e-01 -4.63817060e-01 -6.61061943e-01 1.02859288e-01
-7.89815545e-01 2.92656630e-01 6.10236645e-01 -5.76961219e-01
-6.99714899e-01 3.58012132e-02 -9.41309109e-02 -5.45933425e-01
3.92305911e-01 4.60896343e-01 5.53574502e-01 -5.19799292e-01
7.06621170e-01 2.21485689e-01 7.42515698e-02 -1.69694334e-01
-6.53167516e-02 3.49615604e-01 8.25758040e-01 -4.47756708e-01
-4.59267832e-02 2.99327165e-01 -6.98425546e-02 -3.18314195e-01
-2.20591780e-02 -2.17455924e-01 -4.16933000e-01 -2.30610982e-01
4.69810128e-01 -8.63215327e-01 -1.29945445e+00 1.15344763e+00
-1.38444424e+00 -7.74099410e-01 -8.42283517e-02 4.12887245e-01
-8.35080683e-01 -3.35000545e-01 -1.19422901e+00 -1.05839825e+00
-5.98060906e-01 -8.57055128e-01 5.79932392e-01 4.51028645e-01
-2.55556405e-01 -1.02145910e+00 1.10420465e-01 -8.78633261e-01
9.38321173e-01 2.44306073e-01 8.64233434e-01 -4.21303570e-01
-5.41146338e-01 -3.04341000e-02 -2.00448781e-01 1.63996264e-01
-4.37374026e-01 4.53863263e-01 -1.07481706e+00 -6.08461238e-02
1.57588217e-02 -3.75980675e-01 1.08874643e+00 5.31557143e-01
1.01138580e+00 -8.40725452e-02 -3.80004823e-01 1.00528955e+00
1.51955378e+00 5.53266525e-01 7.54361749e-01 2.27143630e-01
3.54833782e-01 2.65782654e-01 -4.44135785e-01 3.21402162e-01
4.09608893e-02 1.92742512e-01 4.51702535e-01 2.43061483e-01
1.19434491e-01 4.25242111e-02 6.34543538e-01 9.69768882e-01
-4.75806594e-01 -1.04266286e-01 -9.76949036e-01 3.17705959e-01
-2.28283119e+00 -1.03953433e+00 -2.70035684e-01 1.86073279e+00
6.81083977e-01 2.32494578e-01 -5.64881861e-02 7.58646503e-02
4.93732929e-01 -1.84622675e-01 -1.07400572e+00 -4.41865027e-01
-5.35943151e-01 4.47919130e-01 7.09916949e-01 1.96637735e-01
-5.82638562e-01 1.01210535e+00 7.32805872e+00 4.03997332e-01
-1.49359357e+00 3.11090471e-03 4.10407454e-01 -3.78591627e-01
-1.52690366e-01 -1.87929310e-02 -1.03924441e+00 3.13596368e-01
1.55392945e+00 -2.12438568e-01 9.80427086e-01 5.57012200e-01
2.20348328e-01 2.17462927e-02 -9.19454932e-01 1.11468220e+00
-2.16904879e-01 -1.60915565e+00 -4.75806504e-01 -2.20983431e-01
2.55898416e-01 8.23956430e-01 2.98150122e-01 4.92792368e-01
2.86323726e-01 -1.12492609e+00 9.26881433e-01 7.73874879e-01
6.34026766e-01 -4.45115596e-01 5.42533576e-01 -2.62902044e-02
-1.13062692e+00 -2.30424270e-01 -2.79403627e-01 -4.24478889e-01
-5.01395436e-03 3.75866026e-01 -1.33646414e-01 -2.16299638e-01
1.04692972e+00 1.17878699e+00 -3.45244825e-01 9.97711837e-01
2.05424383e-01 6.49719477e-01 -7.36175656e-01 -5.38565576e-01
2.91512758e-01 6.75365282e-03 2.65720516e-01 1.19903255e+00
3.55264932e-01 3.09211969e-01 -6.67871833e-01 1.50257134e+00
2.44961545e-01 -6.30094111e-01 -6.31060660e-01 -3.39840621e-01
6.21440113e-01 1.11994970e+00 -1.11291218e+00 -1.54646887e-02
-8.94313678e-02 6.54106438e-01 2.69462705e-01 8.06479037e-01
-7.91247129e-01 -7.22709775e-01 5.27594388e-01 -2.66935408e-01
2.61877984e-01 -5.75656772e-01 -8.84683430e-01 -1.00858247e+00
-2.50656843e-01 -4.36201274e-01 -1.18678324e-01 -1.27993059e+00
-7.31135130e-01 8.22541952e-01 -4.74687934e-01 -1.04139149e+00
-4.56139207e-01 -1.10877180e+00 -5.10355055e-01 9.86797154e-01
-1.09192479e+00 -6.97317660e-01 1.27782091e-01 6.41896844e-01
-1.64527185e-02 -1.36862844e-01 1.14698529e+00 8.91197100e-02
-7.27019548e-01 3.55376750e-01 3.48645061e-01 1.63070820e-02
1.80916965e-01 -7.52596915e-01 8.66676211e-01 7.52833366e-01
-4.00535822e-01 1.12958968e+00 5.53799331e-01 -4.94332075e-01
-1.40471077e+00 -1.06538427e+00 6.65914536e-01 -2.43760925e-02
1.07231033e+00 -5.01532793e-01 -1.07272387e+00 1.05594528e+00
-8.16848129e-02 1.37843311e-01 2.91191369e-01 -8.47097412e-02
-6.40439391e-01 4.86863852e-02 -7.61103809e-01 9.00253117e-01
1.07070446e+00 -7.45288432e-01 -1.17458738e-01 -2.17524007e-01
6.39508545e-01 -2.33533651e-01 -6.68540478e-01 4.18507069e-01
9.30793226e-01 -9.93021429e-01 5.90047717e-01 -2.35075518e-01
8.52258950e-02 -2.07309738e-01 -1.28665760e-01 -8.53613377e-01
-4.32249248e-01 -9.00849044e-01 -2.70088136e-01 6.95278525e-01
4.11711484e-01 -1.01599026e+00 4.24557120e-01 7.61482418e-01
-5.43169796e-01 -9.33207512e-01 -9.95454311e-01 -8.44708383e-01
1.31491929e-01 -4.79111195e-01 1.67403087e-01 4.13008541e-01
3.55535686e-01 1.40427500e-01 7.47586787e-02 -1.70282245e-01
3.19550127e-01 -2.56917626e-01 -1.21644683e-01 -1.26921070e+00
-3.52184147e-01 -7.52215862e-01 -4.31109786e-01 -7.95669675e-01
3.74575816e-02 -8.75415683e-01 1.78280666e-01 -1.19419849e+00
-1.00808762e-01 -6.68168843e-01 -6.31826043e-01 1.20905566e+00
5.52634537e-01 3.37496519e-01 1.35811463e-01 5.02434313e-01
-5.84201276e-01 3.49096358e-01 5.20774186e-01 3.14747840e-01
-2.82116205e-01 -4.77620065e-02 -3.46713603e-01 5.56068242e-01
8.91379654e-01 -4.77154642e-01 -2.15023100e-01 -4.93337363e-01
5.58164239e-01 -1.69083886e-02 6.72514975e-01 -1.41871119e+00
8.50376189e-01 2.37272024e-01 2.00690418e-01 -5.42804360e-01
3.89376700e-01 -4.45021987e-01 3.50659251e-01 7.87404418e-01
-3.99964422e-01 3.95206034e-01 5.71697950e-01 2.25314423e-01
2.04892959e-02 -1.69116855e-01 8.22062194e-01 -1.87827066e-01
-7.92838097e-01 2.53977746e-01 -8.28740239e-01 -2.52072245e-01
7.41654873e-01 -4.45848733e-01 -8.22760522e-01 -1.39961168e-01
-6.89156175e-01 6.10757358e-02 2.93867081e-01 2.13451475e-01
6.60508096e-01 -1.00319648e+00 -9.37822536e-02 3.99504662e-01
-1.17238425e-01 -2.34275624e-01 -3.56332511e-02 8.88896883e-01
-4.74424034e-01 8.06956589e-01 -6.34954035e-01 -3.78108144e-01
-6.37935996e-01 2.16208816e-01 8.80672216e-01 1.12072818e-01
-6.00275755e-01 9.18420851e-01 -7.91913569e-02 -4.00787920e-01
3.87319475e-01 -7.47712076e-01 -3.11862901e-02 -2.82585144e-01
4.21394110e-01 3.21741760e-01 2.75788963e-01 -8.48822389e-03
-3.62260640e-01 3.54526252e-01 1.39327914e-01 -4.34781522e-01
1.68750918e+00 2.76256025e-01 -5.31334519e-01 1.23718035e+00
1.00050461e+00 -7.42789924e-01 -1.30490553e+00 1.13027897e-02
1.10824838e-01 9.07913268e-01 1.26136139e-01 -8.05508733e-01
-9.83711183e-01 1.02106369e+00 6.31979167e-01 -5.20967841e-02
1.06276619e+00 -4.71222848e-01 7.57485271e-01 7.26981521e-01
6.36982262e-01 -8.19790244e-01 -1.63030118e-01 1.27308083e+00
7.08725750e-01 -8.71142596e-02 -5.54504752e-01 9.80896875e-02
-3.02213520e-01 1.42607081e+00 5.88395298e-01 -5.15817463e-01
7.93269575e-01 1.12585723e+00 -1.67143166e-01 5.69719337e-02
-1.36847806e+00 1.32746935e-01 -5.18194675e-01 4.58031923e-01
5.99950671e-01 -1.38591141e-01 1.04765132e-01 7.85223722e-01
-5.87745309e-02 5.22345483e-01 4.41514909e-01 1.09886754e+00
-4.60687339e-01 -5.23784578e-01 -3.02096680e-02 4.41174001e-01
-2.16977060e-01 -5.69619298e-01 -8.36838335e-02 5.45562446e-01
-2.10991159e-01 4.85416740e-01 4.04676229e-01 -4.27661806e-01
2.27014180e-02 1.75719634e-01 3.74256104e-01 -3.43682021e-01
-8.17654431e-01 9.07312781e-02 -2.25526720e-01 -5.98991334e-01
-3.11753660e-01 -5.07366359e-01 -2.03619719e+00 -3.62616509e-01
-1.94217712e-01 -1.45834237e-01 9.12840784e-01 9.37205732e-01
7.96941161e-01 8.20815504e-01 5.96058667e-02 -9.97856796e-01
-1.82437778e-01 -7.42064118e-01 -7.26356983e-01 -4.68163937e-01
4.12422299e-01 -5.45718133e-01 -4.05472100e-01 1.51628032e-01] | [8.17261791229248, 2.5860421657562256] |
e8952d7d-799f-4673-ab8f-cee23fc2b489 | recogym-a-reinforcement-learning-environment | 1808.00720 | null | http://arxiv.org/abs/1808.00720v2 | http://arxiv.org/pdf/1808.00720v2.pdf | RecoGym: A Reinforcement Learning Environment for the problem of Product Recommendation in Online Advertising | Recommender Systems are becoming ubiquitous in many settings and take many
forms, from product recommendation in e-commerce stores, to query suggestions
in search engines, to friend recommendation in social networks. Current
research directions which are largely based upon supervised learning from
historical data appear to be showing diminishing returns with a lot of
practitioners report a discrepancy between improvements in offline metrics for
supervised learning and the online performance of the newly proposed models.
One possible reason is that we are using the wrong paradigm: when looking at
the long-term cycle of collecting historical performance data, creating a new
version of the recommendation model, A/B testing it and then rolling it out. We
see that there a lot of commonalities with the reinforcement learning (RL)
setup, where the agent observes the environment and acts upon it in order to
change its state towards better states (states with higher rewards). To this
end we introduce RecoGym, an RL environment for recommendation, which is
defined by a model of user traffic patterns on e-commerce and the users
response to recommendations on the publisher websites. We believe that this is
an important step forward for the field of recommendation systems research,
that could open up an avenue of collaboration between the recommender systems
and reinforcement learning communities and lead to better alignment between
offline and online performance metrics. | ['Flavian vasile', 'David Rohde', 'Travis Dunlop', 'Stephen Bonner', 'Alexandros Karatzoglou'] | 2018-08-02 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [-6.05527908e-02 -2.95834262e-02 -5.72956920e-01 -3.66222769e-01
-1.13228306e-01 -6.41034484e-01 6.71418250e-01 1.88421771e-01
-5.37987530e-01 5.50461650e-01 3.24697733e-01 -6.23129904e-01
-7.17767894e-01 -9.62781429e-01 -5.72381735e-01 -5.92814505e-01
-2.76668161e-01 6.39953136e-01 2.98833281e-01 -7.39544749e-01
5.61740875e-01 2.85288274e-01 -1.69582105e+00 1.65801778e-01
3.93471152e-01 6.81792259e-01 1.89803988e-02 7.09963858e-01
5.39036468e-02 8.24499488e-01 -1.27825335e-01 -4.77695376e-01
4.05954123e-01 -3.16570729e-01 -8.50557029e-01 -5.76704070e-02
1.22670822e-01 -1.78464949e-01 -4.83836740e-01 8.46735299e-01
3.03607106e-01 5.91363847e-01 2.64948517e-01 -1.32091951e+00
-8.94353688e-01 9.68467772e-01 -4.53654304e-02 3.38015527e-01
4.84872520e-01 1.48476779e-01 1.31536722e+00 -1.70561224e-01
6.39207840e-01 1.09041202e+00 4.79218423e-01 4.73871231e-01
-1.31330001e+00 -1.81396842e-01 5.17243743e-01 4.58746761e-01
-6.24026537e-01 -2.99985677e-01 3.71450603e-01 -4.25172448e-01
8.78782272e-01 3.55355352e-01 6.08226657e-01 1.34264290e+00
8.63855258e-02 7.78634846e-01 1.26618838e+00 -3.66282314e-01
4.40402716e-01 5.64437270e-01 3.06919366e-01 2.93946832e-01
2.00798020e-01 7.79265285e-01 -4.38359380e-01 -1.31408587e-01
5.81569791e-01 3.95982236e-01 1.03479266e-01 -5.29625833e-01
-8.73409808e-01 1.01344037e+00 1.67692095e-01 4.70787764e-01
-7.26569593e-01 -1.14725575e-01 2.89650559e-01 1.03180885e+00
3.23560297e-01 8.16050708e-01 -4.93869662e-01 -4.00449544e-01
-5.36991417e-01 4.56453353e-01 1.13647866e+00 5.90033591e-01
6.21707678e-01 -2.10838318e-01 -1.41745687e-01 7.85439312e-01
5.17911434e-01 2.42028251e-01 6.82688653e-01 -1.08647501e+00
-2.05648616e-02 3.43809456e-01 4.43405867e-01 -7.64753222e-01
-4.30984139e-01 -4.70404655e-01 -1.96635481e-02 1.01709023e-01
6.43589914e-01 -3.41200709e-01 -2.05359265e-01 1.54713058e+00
3.33567709e-01 2.20856279e-01 -9.98428538e-02 1.08875334e+00
8.64385515e-02 5.58932602e-01 8.83577112e-03 -3.22196633e-01
8.14408779e-01 -1.03779781e+00 -4.82687861e-01 -8.84944201e-02
7.81001985e-01 -4.89535928e-01 1.13750577e+00 9.21772659e-01
-1.04788613e+00 -5.50094485e-01 -8.76410544e-01 5.17525971e-01
-5.27435184e-01 -3.78805488e-01 8.26347649e-01 7.94027448e-01
-1.15512538e+00 1.24771845e+00 -7.43860185e-01 -7.91632593e-01
-1.02190517e-01 5.10809958e-01 1.61561325e-01 -9.93419737e-02
-1.10610867e+00 1.03992033e+00 -1.68120116e-01 -2.82419529e-02
-6.59329891e-01 -3.44946444e-01 -1.26374379e-01 2.83066574e-02
8.08856487e-01 -4.26944047e-01 1.50921619e+00 -1.12958634e+00
-1.99531829e+00 2.27559000e-01 6.09946549e-01 -5.58050036e-01
3.47056776e-01 -3.12092394e-01 -8.07685316e-01 -6.27939641e-01
-3.09201241e-01 -4.93655317e-02 5.35810411e-01 -1.02632236e+00
-1.04445100e+00 -4.93790984e-01 3.82915527e-01 2.07111761e-01
-3.13932151e-01 6.38242736e-02 9.21457727e-03 -1.88358068e-01
-3.28874141e-01 -1.17666316e+00 -4.77124482e-01 -5.70561707e-01
1.60650089e-01 -4.66055870e-01 4.25740898e-01 -2.07086459e-01
1.27290463e+00 -1.72861743e+00 -1.11204265e-02 4.78258222e-01
-1.27342016e-01 2.80487061e-01 -4.82504129e-01 8.90111566e-01
1.60157353e-01 1.51677651e-03 6.60620213e-01 4.82748747e-02
2.25080997e-01 2.93270916e-01 -4.35446471e-01 4.05247599e-01
-3.46907556e-01 7.16195464e-01 -1.22135806e+00 2.50090063e-01
1.86282024e-01 1.35534048e-01 -8.16149235e-01 3.15772682e-01
-4.44868028e-01 3.73315513e-01 -6.92457080e-01 1.25145927e-01
1.01204105e-01 -2.66419858e-01 5.14943361e-01 3.41372937e-01
-3.43976408e-01 7.70580292e-01 -1.34483731e+00 1.31589150e+00
-4.73583281e-01 1.67082936e-01 -1.41063526e-01 -1.20181656e+00
8.40889275e-01 1.81617945e-01 8.17690492e-01 -1.13689864e+00
1.59879044e-01 -6.17660992e-02 4.28932488e-01 -6.87060952e-01
6.91850901e-01 8.21313709e-02 3.21912318e-01 9.92166102e-01
1.00614578e-01 4.63905632e-01 3.61785233e-01 1.78533629e-01
1.31030965e+00 4.32196021e-01 5.02821617e-02 -1.16323203e-01
3.80067974e-01 -3.77834551e-02 2.00432241e-01 1.23256171e+00
-7.32930675e-02 -1.70206666e-01 4.13617313e-01 -5.43874264e-01
-8.32499921e-01 -7.77306736e-01 3.04001719e-01 1.82375109e+00
-2.60303188e-02 -4.75518405e-01 -4.29269671e-01 -7.89772928e-01
4.29065600e-02 9.15963948e-01 -6.08584166e-01 -3.60467374e-01
-4.74849284e-01 -3.33678931e-01 -8.13509896e-02 2.03675732e-01
2.31663343e-02 -1.37626147e+00 -4.94710207e-01 3.77071083e-01
3.18727642e-01 -6.15002513e-01 -4.97244418e-01 2.08436385e-01
-1.11152148e+00 -9.97935832e-01 -2.95938313e-01 -3.37505072e-01
2.93248147e-01 4.20753956e-01 1.17397618e+00 3.35161954e-01
3.62340808e-01 8.82096708e-01 -7.80671358e-01 -2.01508194e-01
-5.89930475e-01 9.00820456e-03 4.63480562e-01 1.09242931e-01
3.83443266e-01 -5.77613592e-01 -8.75059128e-01 6.03435457e-01
-7.93406963e-01 -4.05391246e-01 6.47459686e-01 6.59634054e-01
1.45167977e-01 -4.85421233e-02 7.71855175e-01 -1.56659198e+00
8.88578415e-01 -1.01140797e+00 -6.06248975e-01 2.51109093e-01
-1.46793401e+00 7.61763006e-02 8.18705916e-01 -6.29647374e-01
-1.04091287e+00 -3.60248238e-01 2.83114309e-03 -8.43257084e-02
-2.81738579e-01 5.07458568e-01 3.91775608e-01 2.72492498e-01
9.82335627e-01 -2.88991742e-02 8.51212293e-02 -5.18874288e-01
6.42248154e-01 7.24150777e-01 6.65582493e-02 -5.91081500e-01
6.70254230e-01 4.02874090e-02 -3.21125835e-01 -4.93774861e-01
-8.62868547e-01 -6.50795221e-01 -1.86800495e-01 -4.01887894e-01
2.56770998e-01 -1.98035970e-01 -1.08006692e+00 -2.09989786e-01
-3.99360269e-01 -8.14270914e-01 -5.54812789e-01 6.27882242e-01
-6.02890730e-01 5.31901121e-02 -6.33752763e-01 -9.43525374e-01
1.13706226e-02 -9.21401262e-01 4.05722409e-01 4.50849801e-01
-8.42383355e-02 -1.17158723e+00 6.13107681e-01 3.80497247e-01
9.18337107e-01 -6.32439435e-01 6.93522871e-01 -1.23100865e+00
-3.01945895e-01 -1.65438309e-01 1.90237030e-01 2.77788758e-01
1.47445565e-02 -7.78488815e-02 -7.32908905e-01 -5.01029074e-01
-1.76742792e-01 -1.83603004e-01 3.47415030e-01 1.87170818e-01
9.63864148e-01 -5.44594705e-01 -1.46897778e-01 -7.20747635e-02
1.27997410e+00 4.43127930e-01 4.89753544e-01 5.23114860e-01
1.10619694e-01 7.61019766e-01 8.33587408e-01 4.57283676e-01
4.80420679e-01 8.44777167e-01 3.40931416e-01 1.76162273e-01
2.01267689e-01 -5.91283917e-01 7.61135340e-01 7.58007467e-01
-2.42728159e-01 -3.16135138e-01 -3.78847212e-01 1.11078821e-01
-2.24104857e+00 -1.22129381e+00 1.22747518e-01 2.70254111e+00
4.80669558e-01 1.68539450e-01 8.45783651e-01 -1.55628622e-01
4.68701869e-01 -2.43414536e-01 -6.44278347e-01 -7.66385078e-01
5.58264434e-01 1.29981682e-01 5.73190153e-01 3.69788110e-01
-5.42404115e-01 7.21382558e-01 6.61427498e+00 2.66318709e-01
-1.23410487e+00 6.71440512e-02 4.44679379e-01 -8.28285813e-02
-2.31876731e-01 2.52664566e-01 -6.41385794e-01 4.43976343e-01
1.60217071e+00 -1.80949032e-01 1.21726596e+00 8.85563254e-01
5.31101942e-01 -9.54555497e-02 -1.45059705e+00 6.02883279e-01
-1.04746394e-01 -1.11277688e+00 -3.48293930e-01 3.87322903e-01
7.12790489e-01 4.43890065e-01 1.15176193e-01 8.33171964e-01
9.56922233e-01 -6.23052776e-01 2.96278983e-01 8.67562115e-01
-4.86555919e-02 -4.73408997e-01 6.15464628e-01 6.68961763e-01
-4.39284146e-01 -4.94905114e-01 -4.12327975e-01 -3.11156332e-01
-7.15337172e-02 1.88987017e-01 -8.52615476e-01 2.14561537e-01
5.01978338e-01 7.14840055e-01 -4.71248060e-01 1.33155334e+00
2.11009026e-01 9.57654774e-01 -5.26223890e-02 -4.95925188e-01
3.16910110e-02 -8.12073946e-01 3.62507313e-01 9.07774448e-01
3.47142905e-01 -6.52392805e-02 2.89063245e-01 5.63343763e-01
7.61186983e-03 4.33887601e-01 -6.99048698e-01 -1.35235429e-01
2.34903753e-01 1.42448747e+00 -7.70107090e-01 -1.62700474e-01
-7.27903664e-01 6.18186593e-01 3.05060208e-01 3.35151970e-01
-7.08734870e-01 1.77460283e-01 4.78355020e-01 4.07505959e-01
2.57141650e-01 3.05696372e-02 2.39984408e-01 -9.39286470e-01
-4.62456942e-01 -1.13147509e+00 5.27461827e-01 -5.77262580e-01
-1.42785692e+00 1.96710572e-01 -1.83341011e-01 -1.14504802e+00
-2.75467157e-01 -5.39595306e-01 -5.49899161e-01 2.63182402e-01
-1.48363149e+00 -3.46830338e-01 2.13333905e-01 5.32622814e-01
6.39943719e-01 -2.43085220e-01 7.24896073e-01 4.44764316e-01
-4.87447858e-01 4.29516464e-01 4.82861042e-01 -5.15788019e-01
7.82267153e-01 -1.30561149e+00 1.17216773e-01 4.11173850e-01
7.58705437e-01 6.88102663e-01 8.49036157e-01 -4.01759863e-01
-1.92455935e+00 -7.16957986e-01 5.89548111e-01 -7.06821322e-01
9.44684684e-01 3.78725044e-02 -6.60573304e-01 6.31864548e-01
2.11802900e-01 -2.57530868e-01 7.71278322e-01 6.39486074e-01
-1.47801191e-01 -2.75264174e-01 -9.73871291e-01 6.91484511e-01
1.01156592e+00 -1.69232830e-01 -1.71662077e-01 6.57741666e-01
4.60921764e-01 1.56121105e-02 -1.01141572e+00 -1.96514875e-01
6.36300981e-01 -1.09404755e+00 6.24446094e-01 -1.03272092e+00
2.01218233e-01 -2.99074985e-02 -6.55169934e-02 -1.70395792e+00
-6.13906741e-01 -9.20443296e-01 -1.48410290e-01 9.72216189e-01
5.82793832e-01 -8.23916614e-01 8.29162657e-01 6.53942764e-01
9.11955982e-02 -7.03316867e-01 -3.82001370e-01 -6.23816550e-01
2.02142298e-02 -4.22296733e-01 4.26628888e-01 7.67866850e-01
3.82036746e-01 5.84752262e-01 -4.85882014e-01 -9.57138762e-02
1.88634381e-01 7.42430463e-02 8.64463270e-01 -1.46642756e+00
-8.60581875e-01 -5.46713293e-01 2.00357974e-01 -1.33332765e+00
-2.96258628e-01 -7.63757527e-01 -3.11387572e-02 -1.16243744e+00
-6.23890720e-02 -7.60912299e-01 -8.09017479e-01 1.79833293e-01
3.22028637e-01 -4.89265770e-02 1.25118643e-01 3.23079944e-01
-1.19552588e+00 1.03336319e-01 1.15999794e+00 2.48277768e-01
-5.32867551e-01 6.95761979e-01 -1.11054790e+00 5.07493854e-01
7.30622530e-01 -5.34576237e-01 -4.90357280e-01 1.39972446e-02
8.56536150e-01 2.35384077e-01 -1.10455431e-01 -6.76074505e-01
3.45526457e-01 -5.92816651e-01 -1.67783752e-01 6.63196445e-02
-1.28070250e-01 -9.92348611e-01 2.48537213e-02 3.27996254e-01
-9.76988554e-01 2.02821404e-01 -3.96759331e-01 8.72194171e-01
3.65659028e-01 -3.44844878e-01 3.54525894e-01 -7.88409710e-02
-5.12375593e-01 3.26900393e-01 -5.60212195e-01 -1.73483521e-01
7.27064610e-01 -1.54045433e-01 -4.38285023e-01 -7.22598672e-01
-8.81076097e-01 2.81350255e-01 2.01319218e-01 8.83846581e-01
1.09660111e-01 -9.70225990e-01 -4.03717577e-01 3.91625836e-02
-8.65601748e-02 -8.59790027e-01 2.06808541e-02 9.49704170e-01
1.15232289e-01 3.64466757e-01 -1.18685275e-01 -1.17881916e-01
-8.76816928e-01 7.55926073e-01 2.75886297e-01 -5.39101958e-01
-4.57186550e-01 2.98073202e-01 -4.35510576e-01 -5.95030487e-01
4.43345726e-01 -1.63288146e-01 -6.68052733e-01 -3.78097445e-02
4.00537968e-01 6.64835155e-01 1.26355112e-01 -2.18681112e-01
1.35391995e-01 -4.07144465e-02 -2.74520606e-01 -1.24454394e-01
1.73068678e+00 -2.92055547e-01 2.56880760e-01 5.94743848e-01
6.71833456e-01 1.68392748e-01 -1.07283533e+00 -4.81362909e-01
3.88730049e-01 -4.96667862e-01 1.96509808e-01 -1.14106214e+00
-1.11716402e+00 2.11904004e-01 8.78377557e-01 9.63874817e-01
7.97416329e-01 3.06546297e-02 5.58850825e-01 8.94833565e-01
4.91794944e-01 -1.69899511e+00 2.35179454e-01 3.44405055e-01
5.75987041e-01 -1.19032896e+00 2.43256353e-02 3.15507770e-01
-7.97974110e-01 1.09863997e+00 5.03167808e-01 -5.75040340e-01
8.94558311e-01 -5.40379249e-02 2.23963894e-02 -2.09062025e-01
-1.30031049e+00 -3.79433572e-01 -1.70214493e-02 6.01725459e-01
6.67767227e-01 1.24803811e-01 -4.62585092e-01 6.17519975e-01
-1.47218592e-02 2.10005537e-01 6.70566082e-01 5.40716708e-01
-5.11048317e-01 -1.83446181e+00 3.53978239e-02 8.48690152e-01
-2.50322372e-01 2.20193133e-01 -1.34423494e-01 5.51023722e-01
-1.11562133e-01 1.20782256e+00 -1.49933279e-01 -6.19651616e-01
6.62966847e-01 8.57754871e-02 3.47081035e-01 -6.87281966e-01
-9.19678390e-01 -2.62667518e-02 2.62398422e-01 -9.28041339e-01
-4.26169217e-01 -9.10998285e-01 -8.05040538e-01 -4.02249783e-01
-5.20896733e-01 4.47402716e-01 7.84048498e-01 1.00819600e+00
3.89097512e-01 4.38818306e-01 1.17248082e+00 -6.50547922e-01
-1.09273970e+00 -8.21810603e-01 -6.95742249e-01 5.72398603e-01
3.64468917e-02 -7.58697629e-01 -3.30880016e-01 -3.51906478e-01] | [9.935948371887207, 5.70648193359375] |
e74427ab-4cf9-4e29-96da-3f0ab49f46c3 | learning-dynamic-knowledge-graphs-to | 2002.09127 | null | https://arxiv.org/abs/2002.09127v4 | https://arxiv.org/pdf/2002.09127v4.pdf | Learning Dynamic Belief Graphs to Generalize on Text-Based Games | Playing text-based games requires skills in processing natural language and sequential decision making. Achieving human-level performance on text-based games remains an open challenge, and prior research has largely relied on hand-crafted structured representations and heuristics. In this work, we investigate how an agent can plan and generalize in text-based games using graph-structured representations learned end-to-end from raw text. We propose a novel graph-aided transformer agent (GATA) that infers and updates latent belief graphs during planning to enable effective action selection by capturing the underlying game dynamics. GATA is trained using a combination of reinforcement and self-supervised learning. Our work demonstrates that the learned graph-based representations help agents converge to better policies than their text-only counterparts and facilitate effective generalization across game configurations. Experiments on 500+ unique games from the TextWorld suite show that our best agent outperforms text-based baselines by an average of 24.2%. | ['Marc-Antoine Rondeau', 'Mikuláš Zelinka', 'Marc-Alexandre Côté', 'Ashutosh Adhikari', 'Xingdi Yuan', 'William L. Hamilton', 'Romain Laroche', 'Pascal Poupart', 'Jian Tang', 'Adam Trischler'] | 2020-02-21 | null | http://proceedings.neurips.cc/paper/2020/hash/1fc30b9d4319760b04fab735fbfed9a9-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/1fc30b9d4319760b04fab735fbfed9a9-Paper.pdf | neurips-2020-12 | ['text-based-games'] | ['playing-games'] | [ 1.49494573e-01 4.58333969e-01 -3.20276886e-01 -1.20984115e-01
-7.92299807e-01 -5.54881692e-01 8.74825239e-01 2.64676601e-01
-4.30249184e-01 8.07046413e-01 5.41182101e-01 -4.99315113e-01
-1.75788522e-01 -1.16067398e+00 -6.50878370e-01 -1.49343684e-01
-4.19548839e-01 1.22277641e+00 3.27379555e-01 -7.97267735e-01
3.50018412e-01 -8.72832164e-02 -1.24899733e+00 3.51820081e-01
8.77539933e-01 4.55561876e-01 3.49902153e-01 9.57047164e-01
-8.84193182e-02 1.70833719e+00 -4.95608300e-01 -4.00179744e-01
5.19882619e-01 -5.95809639e-01 -8.75367820e-01 2.42244005e-01
-2.42968109e-02 -5.53761661e-01 -8.14105690e-01 8.21054161e-01
2.63954669e-01 6.76373303e-01 6.23329163e-01 -1.16451442e+00
-4.70873982e-01 1.09783888e+00 -3.62691998e-01 2.33902290e-01
4.84810352e-01 6.91762149e-01 1.19987631e+00 -3.71807851e-02
7.49130666e-01 1.36939847e+00 4.08610195e-01 7.59138584e-01
-1.25531399e+00 -4.41601187e-01 6.50764585e-01 1.42623678e-01
-7.89445698e-01 -1.85131744e-01 5.66579282e-01 -3.21412683e-01
1.69346428e+00 -3.51065487e-01 1.09684813e+00 1.16045332e+00
4.75488096e-01 8.46623242e-01 1.00109553e+00 -1.98239028e-01
4.29772258e-01 -6.92660809e-01 -2.34893084e-01 1.32052732e+00
2.62171149e-01 4.41174656e-01 -7.19465733e-01 -6.79647997e-02
1.03903830e+00 -1.82827413e-02 1.41490802e-01 -5.64207375e-01
-8.00856411e-01 1.18100512e+00 4.60959405e-01 -2.18632653e-01
-7.15224266e-01 5.96272290e-01 5.26117802e-01 4.32077646e-01
3.92944187e-01 1.03202808e+00 -2.85169393e-01 -8.47727239e-01
-5.25538802e-01 6.63049877e-01 8.31441224e-01 1.02416682e+00
6.96411610e-01 4.49236155e-01 -2.62470216e-01 5.46528578e-01
2.02753946e-01 3.05701464e-01 6.21706009e-01 -1.02646601e+00
7.97145188e-01 7.79862583e-01 -8.93490389e-04 -9.16847527e-01
-4.90171075e-01 -2.69733816e-01 -2.36332700e-01 4.86849070e-01
3.07841659e-01 -4.89678860e-01 -1.13837564e+00 1.73326957e+00
-1.67832270e-01 1.82709038e-01 3.27057600e-01 5.92921734e-01
5.50305665e-01 6.10780239e-01 2.58602947e-01 8.60546157e-02
9.00158405e-01 -1.32189381e+00 -3.17455441e-01 -8.91118109e-01
4.99467015e-01 -2.01551779e-03 9.80722606e-01 4.52199787e-01
-1.37939739e+00 -4.39398102e-02 -7.95224190e-01 1.83388472e-01
-1.35035470e-01 -5.05755186e-01 1.02923059e+00 4.36709553e-01
-1.41836727e+00 5.97208858e-01 -1.19688046e+00 -3.38166803e-01
4.99894291e-01 4.53740686e-01 -1.69190884e-01 -8.17813352e-02
-9.65476811e-01 9.74103868e-01 6.51646733e-01 -4.20179158e-01
-1.44369316e+00 -3.11943471e-01 -1.14026797e+00 7.87471235e-02
1.05988872e+00 -9.41281974e-01 1.86491263e+00 -5.16404331e-01
-2.25761318e+00 3.59039873e-01 2.34147310e-01 -9.46779430e-01
4.95647073e-01 -6.22790754e-02 3.84563357e-01 1.73055813e-01
2.76810199e-01 4.41954970e-01 7.63874114e-01 -8.46639156e-01
-7.84300089e-01 -2.09278658e-01 6.51453197e-01 7.65222311e-01
3.34061868e-02 -3.96319777e-01 -5.80253184e-01 -4.43979621e-01
-1.53874755e-01 -9.82436121e-01 -8.08238208e-01 -6.63658798e-01
-2.45812476e-01 -3.91962141e-01 2.01190397e-01 -5.61707735e-01
8.42763484e-01 -1.43975198e+00 4.54128504e-01 7.40560964e-02
6.61844075e-01 -1.29755899e-01 -2.89426774e-01 8.48874390e-01
4.77287084e-01 -3.98987085e-02 9.62727368e-02 -3.96394014e-01
2.72033095e-01 3.54039580e-01 -2.59778261e-01 1.05719902e-02
4.65292931e-02 1.29462433e+00 -1.44246256e+00 -2.52222180e-01
4.22885031e-01 -1.26225784e-01 -1.04260159e+00 2.52294391e-01
-9.71445382e-01 1.91900626e-01 -7.54245460e-01 3.05491656e-01
-9.25403237e-02 -3.38297218e-01 8.12799454e-01 6.61293030e-01
2.50157028e-01 6.65905356e-01 -6.32985830e-01 2.08076978e+00
-6.11351848e-01 4.23735648e-01 -2.33333290e-01 -8.76807034e-01
8.24092567e-01 -1.06580764e-01 3.56226206e-01 -1.02646828e+00
4.46394598e-03 -4.03157800e-01 -2.72766799e-02 -5.71067072e-02
8.34689617e-01 1.76467180e-01 -3.93050581e-01 7.62097836e-01
1.22064821e-01 -6.53365135e-01 5.61213791e-01 6.88215077e-01
1.63336658e+00 3.44430447e-01 5.34822166e-01 6.40148744e-02
-3.01862925e-01 5.98346531e-01 3.86511952e-01 1.37292099e+00
-4.68344055e-02 -7.21641257e-02 7.58750737e-01 -6.30827010e-01
-8.63509476e-01 -8.91685963e-01 1.01857841e+00 1.64875138e+00
6.15028851e-02 -9.61767018e-01 -4.24814463e-01 -8.92932594e-01
-1.16588846e-02 9.22313511e-01 -7.08583355e-01 -4.66304451e-01
-5.40064871e-01 -3.32391679e-01 2.54963666e-01 7.68006802e-01
6.04022026e-01 -1.28232527e+00 -8.22151482e-01 7.31268644e-01
-1.82150712e-03 -1.00651419e+00 -4.07645851e-01 2.20970675e-01
-7.86275089e-01 -1.18716943e+00 -4.13162671e-02 -4.34986204e-01
4.64120477e-01 3.55831027e-01 1.58192003e+00 1.47628248e-01
3.91889028e-02 6.54587150e-01 -5.75839043e-01 -3.37590784e-01
-4.90565926e-01 3.18827689e-01 -1.48187265e-01 -7.64141262e-01
7.51918256e-02 -6.10387743e-01 -3.29449445e-01 -4.18087579e-02
-4.63499695e-01 5.54407179e-01 6.29863799e-01 1.04791796e+00
1.43930778e-01 2.44884536e-01 9.11414847e-02 -9.64694321e-01
1.37301588e+00 -4.38324958e-01 -1.04495156e+00 8.35468546e-02
-7.26567447e-01 4.79248643e-01 8.21400821e-01 -2.60153741e-01
-9.72228706e-01 -8.37672949e-02 2.55925089e-01 -1.68333292e-01
1.24570809e-01 7.13981330e-01 3.25276405e-01 6.96895346e-02
9.24851179e-01 4.82392311e-01 7.18613267e-02 2.19493702e-01
6.34161830e-01 1.24588542e-01 4.32010323e-01 -1.06480742e+00
6.87727511e-01 5.82023002e-02 -2.06855074e-01 -4.12073225e-01
-5.98236024e-01 -1.44110993e-01 6.80916905e-02 -2.44404152e-01
6.85537040e-01 -1.02560759e+00 -1.07459605e+00 2.65014380e-01
-6.59733057e-01 -1.38785744e+00 -4.32713509e-01 2.86865443e-01
-1.18716562e+00 -6.47127703e-02 -7.73134530e-01 -7.71449387e-01
-3.40197593e-01 -1.24432814e+00 9.95855987e-01 3.58032912e-01
-1.09586231e-01 -1.11613905e+00 4.32851762e-01 4.36984450e-01
4.11314160e-01 -1.46875186e-02 7.89423466e-01 -6.57424450e-01
-7.41879523e-01 7.54001662e-02 2.16751304e-02 -5.08647740e-01
7.76355639e-02 -3.95712882e-01 -4.09259409e-01 -4.56859738e-01
-5.82985878e-01 -9.10171986e-01 1.00189018e+00 5.35387278e-01
8.82989585e-01 -4.60604578e-01 -1.73300460e-01 5.39531469e-01
1.18431354e+00 1.75266728e-01 4.78411078e-01 5.96297264e-01
4.93219465e-01 1.06956400e-01 4.50647444e-01 8.01307499e-01
8.64869595e-01 6.11232579e-01 7.43203580e-01 4.86662865e-01
8.91829953e-02 -8.47155631e-01 5.17090499e-01 2.14284539e-01
-4.05261785e-01 -5.00155151e-01 -1.14376926e+00 4.13332134e-01
-2.18614769e+00 -1.19077766e+00 6.85410023e-01 1.77410090e+00
7.66876221e-01 5.00076294e-01 4.94409084e-01 -4.27951932e-01
3.89262527e-01 3.84164095e-01 -9.14314210e-01 -3.36893082e-01
4.63381678e-01 5.45763254e-01 6.75804317e-01 7.90325940e-01
-9.41743553e-01 1.95843637e+00 6.45091105e+00 7.40703523e-01
-7.59368658e-01 -2.10965559e-01 3.26351225e-01 -1.86025456e-01
-1.87780693e-01 -4.12763394e-02 -4.97182459e-01 -1.04980908e-01
9.06029880e-01 -7.03179896e-01 1.15145826e+00 1.10479379e+00
1.72346473e-01 3.83967906e-03 -7.86642134e-01 7.23750591e-01
-1.05600975e-01 -1.78331816e+00 2.03188613e-01 3.38768870e-01
9.06708002e-01 3.01609218e-01 6.77421018e-02 1.18417418e+00
1.99210823e+00 -1.25608194e+00 6.94047987e-01 3.23990315e-01
5.08061826e-01 -8.19847405e-01 3.90449762e-01 4.04028893e-01
-1.35629022e+00 -3.11998606e-01 -3.88384312e-01 -7.39886105e-01
-7.72401467e-02 -3.25987786e-01 -1.64408243e+00 6.29996955e-01
1.58343598e-01 1.04226065e+00 -4.97866273e-01 6.56892300e-01
-7.25475132e-01 7.35734940e-01 -1.33858785e-01 -4.67997879e-01
7.81770885e-01 -1.45876601e-01 4.56731647e-01 6.90790474e-01
-1.04058899e-01 3.06355834e-01 1.01332188e+00 6.94654286e-01
-9.10181776e-02 -1.94344640e-01 -9.26211655e-01 -7.15077162e-01
3.73882145e-01 8.98295164e-01 -8.23180497e-01 -3.72550368e-01
-1.39381841e-01 9.33547735e-01 9.50691640e-01 3.85994107e-01
-5.95565915e-01 5.82164787e-02 8.06371987e-01 -8.98459330e-02
2.78534144e-01 -6.58671856e-01 -5.97662926e-02 -1.16293561e+00
-4.95364517e-01 -1.23048830e+00 5.99357903e-01 -6.50260210e-01
-1.16277659e+00 7.29732752e-01 -3.47937159e-02 -8.04372609e-01
-1.11409557e+00 -5.06760240e-01 -6.91440761e-01 3.83151680e-01
-1.21575367e+00 -9.81494248e-01 -3.46075773e-01 7.54313529e-01
9.60548818e-01 -7.07508504e-01 9.42139328e-01 -7.29849875e-01
-4.20420408e-01 3.02545518e-01 -1.84820116e-01 2.80931801e-01
8.77705887e-02 -1.58856285e+00 1.13897324e+00 7.34079838e-01
4.03287739e-01 3.59625489e-01 7.32683957e-01 -1.07347262e+00
-1.66769612e+00 -1.12782109e+00 -5.47798984e-02 -2.89375782e-01
8.83444965e-01 -2.34547615e-01 -2.47782707e-01 1.12180960e+00
3.36055666e-01 -3.64285499e-01 2.55653381e-01 4.09974903e-01
-2.85956681e-01 3.16319823e-01 -7.43070781e-01 1.09965837e+00
1.47240603e+00 -4.61774915e-01 -8.23046505e-01 5.94540179e-01
6.83112621e-01 -8.20363820e-01 -3.15411776e-01 -2.55999148e-01
9.22886729e-02 -7.66906738e-01 9.21776712e-01 -1.18344593e+00
5.51317513e-01 3.47752608e-02 7.75252432e-02 -1.93949413e+00
-5.96404076e-01 -1.10023534e+00 2.07996536e-02 2.13084117e-01
2.63283610e-01 -6.55053914e-01 1.22853029e+00 5.31856358e-01
-1.90150440e-01 -5.83754003e-01 -4.17830795e-01 -6.56417191e-01
7.43880346e-02 -4.66213465e-01 5.98358154e-01 5.27526021e-01
4.78252262e-01 3.98903370e-01 -4.05225664e-01 -9.02141817e-03
5.95137417e-01 1.60622880e-01 1.15676415e+00 -1.12471569e+00
-5.96543133e-01 -6.99844778e-01 -5.35744369e-01 -1.30799258e+00
5.09291649e-01 -8.58919322e-01 2.85201639e-01 -2.06624246e+00
1.09618090e-01 -1.76116779e-01 7.33582154e-02 8.88803780e-01
-1.54062077e-01 -3.07826430e-01 4.07685131e-01 -2.33699128e-01
-1.17242491e+00 7.14520156e-01 1.30976069e+00 -4.04633135e-01
-4.07568216e-01 -5.05159795e-02 -1.01324189e+00 7.03849316e-01
1.06405842e+00 -3.15120041e-01 -9.50494766e-01 -4.32956547e-01
4.41517502e-01 4.37733412e-01 1.61764547e-01 -1.01391304e+00
6.36428654e-01 -7.73406208e-01 -4.37619388e-02 -1.97109748e-02
5.10491490e-01 -2.30061010e-01 -4.28232163e-01 4.45458978e-01
-5.37708700e-01 3.82731766e-01 3.71467501e-01 1.00670445e+00
1.30678877e-01 1.33693755e-01 1.68630734e-01 -7.52570868e-01
-1.13405001e+00 5.81957757e-01 -8.29720318e-01 6.03716552e-01
1.10072732e+00 -4.77523245e-02 -4.53675538e-01 -1.11197710e+00
-5.84485769e-01 5.71082532e-01 3.94888788e-01 1.99774131e-01
7.67927647e-01 -8.79020989e-01 -8.16311359e-01 -6.56304648e-03
3.12624089e-02 -3.53295580e-02 -1.86097939e-02 5.98549731e-02
-6.21775925e-01 2.16803446e-01 -5.12940824e-01 -1.75787166e-01
-8.84796619e-01 2.28766933e-01 5.09491444e-01 -1.00664139e+00
-9.70683753e-01 9.76660669e-01 -9.45228040e-02 -5.31674206e-01
7.87758902e-02 -3.65882844e-01 -2.15507269e-01 -4.65250254e-01
4.01083320e-01 4.15719002e-02 -1.77159622e-01 -1.87959317e-02
-8.22208300e-02 -2.68913936e-02 -5.13300359e-01 -5.50864100e-01
1.61706114e+00 3.76605839e-01 3.92423958e-01 -1.31038800e-01
2.02096939e-01 -2.65988141e-01 -1.75494909e+00 -2.20024899e-01
-5.73415607e-02 -5.46732485e-01 1.86628819e-01 -8.77228439e-01
-9.57335770e-01 4.10468966e-01 -2.82934070e-01 2.27050737e-01
6.63243115e-01 -1.87397525e-01 4.40506548e-01 1.03439093e+00
9.95547831e-01 -9.94740963e-01 6.54191077e-01 1.10662735e+00
8.02420318e-01 -1.07667530e+00 -3.41762416e-02 4.31707837e-02
-1.11611331e+00 1.03078330e+00 8.63929927e-01 -5.63384593e-01
-8.08523297e-02 1.76188782e-01 -2.53488868e-01 -4.06052977e-01
-1.42811584e+00 -5.32316685e-01 -4.43791188e-02 9.64203537e-01
-3.04033165e-03 2.06103444e-01 3.76451999e-01 5.01296043e-01
-6.77041829e-01 9.91580263e-03 7.06773818e-01 1.18769884e+00
-5.95061004e-01 -1.02842200e+00 1.39447404e-02 7.16594160e-01
1.38673201e-01 -2.02132538e-01 -5.61249495e-01 7.39051640e-01
-6.19177043e-01 1.22396481e+00 -6.27453625e-03 -5.69416046e-01
2.57859647e-01 -4.81944643e-02 8.06997061e-01 -1.07520425e+00
-6.73965216e-01 -4.58200693e-01 4.92973417e-01 -1.07115090e+00
1.87518999e-01 -5.90555251e-01 -1.42382860e+00 -5.87812364e-01
2.14906290e-01 2.66426831e-01 1.93577111e-01 9.69205201e-01
3.30192626e-01 9.83735800e-01 1.48917124e-01 -1.05094194e+00
-7.32680500e-01 -8.51972759e-01 -4.23135132e-01 2.13397577e-01
-1.06588453e-02 -9.69575107e-01 9.53679997e-03 -3.73921633e-01] | [3.7882511615753174, 1.3968206644058228] |
a9e07ab6-3221-4349-890b-5d9839c3a698 | adaptive-plant-propagation-algorithm-for | 1708.07040 | null | http://arxiv.org/abs/1708.07040v1 | http://arxiv.org/pdf/1708.07040v1.pdf | Adaptive Plant Propagation Algorithm for Solving Economic Load Dispatch Problem | Optimization problems in design engineering are complex by nature, often
because of the involvement of critical objective functions accompanied by a
number of rigid constraints associated with the products involved. One such
problem is Economic Load Dispatch (ED) problem which focuses on the
optimization of the fuel cost while satisfying some system constraints.
Classical optimization algorithms are not sufficient and also inefficient for
the ED problem involving highly nonlinear, and non-convex functions both in the
objective and in the constraints. This led to the development of metaheuristic
optimization approaches which can solve the ED problem almost efficiently. This
paper presents a novel robust plant intelligence based Adaptive Plant
Propagation Algorithm (APPA) which is used to solve the classical ED problem.
The application of the proposed method to the 3-generator and 6-generator
systems shows the efficiency and robustness of the proposed algorithm. A
comparative study with another state-of-the-art algorithm (APSO) demonstrates
the quality of the solution achieved by the proposed method along with the
convergence characteristics of the proposed approach. | ['Sayan Nag'] | 2017-08-04 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [ 1.54023275e-01 -1.20561048e-01 2.67696261e-01 1.91919938e-01
2.55968124e-01 -5.07943273e-01 2.39013672e-01 4.45365220e-01
-2.05299497e-01 9.45952296e-01 -3.97495627e-01 -1.36218756e-01
-1.15622354e+00 -7.73888588e-01 -8.48796517e-02 -8.72500300e-01
3.33349183e-02 7.40438521e-01 -1.57323480e-01 -5.35910964e-01
5.74342251e-01 7.68116295e-01 -1.63513398e+00 -4.95273083e-01
1.20777440e+00 9.27200675e-01 7.00448334e-01 5.10304213e-01
-5.28237298e-02 3.15477788e-01 -6.14242136e-01 2.84286112e-01
1.79335624e-01 -4.81999338e-01 -6.47834003e-01 4.01276499e-01
-8.95134926e-01 2.36172155e-01 4.75482464e-01 9.04210925e-01
6.24501407e-01 7.24336922e-01 9.52583492e-01 -1.45586765e+00
-3.95627081e-01 1.17979668e-01 -7.08035767e-01 1.05173446e-01
1.14360154e-01 -1.14434443e-01 6.42183900e-01 -6.78709567e-01
3.91739130e-01 7.74486423e-01 3.23869735e-01 -6.41465038e-02
-1.08789182e+00 -6.71753734e-02 -8.07468146e-02 4.01821107e-01
-1.50421536e+00 1.30737796e-01 6.46415472e-01 -4.89542872e-01
1.04732049e+00 3.57352108e-01 6.67947471e-01 -8.14323872e-02
1.84991509e-01 3.87681015e-02 9.81219590e-01 -5.93022883e-01
4.67459828e-01 1.73648566e-01 1.01285852e-01 5.67916095e-01
5.38565814e-01 6.54127300e-02 1.33201689e-01 1.64192557e-01
2.82915644e-02 -2.80838251e-01 -1.55620918e-01 -2.21003950e-01
-6.02535009e-01 1.03323543e+00 1.84507728e-01 6.69917762e-01
-7.05521762e-01 -2.94501752e-01 2.92313695e-01 -1.02559626e-01
1.46755874e-01 6.95548952e-01 -3.18483829e-01 -5.54333664e-02
-9.71952498e-01 3.64441484e-01 9.46902931e-01 9.01900709e-01
1.58000663e-01 4.94977206e-01 9.21330079e-02 6.76988721e-01
4.32660758e-01 2.40077391e-01 1.32553771e-01 -4.07667994e-01
2.22371534e-01 6.99477911e-01 4.29168612e-01 -1.31826735e+00
-6.36646092e-01 -6.57130122e-01 -6.18153989e-01 6.28328741e-01
2.68509567e-01 -5.27287185e-01 -5.79479396e-01 1.26635969e+00
4.60249335e-01 -4.67063069e-01 5.76935522e-02 7.15269864e-01
2.41376698e-01 1.23026204e+00 -4.46331240e-02 -6.83182299e-01
1.10513437e+00 -9.28093910e-01 -9.76247370e-01 1.15664639e-01
1.80972978e-01 -1.05761743e+00 4.68588084e-01 6.30125880e-01
-1.04945695e+00 -3.76511216e-01 -1.33987761e+00 6.93952799e-01
-6.80966437e-01 4.71333146e-01 2.82470763e-01 7.51461804e-01
-6.24255836e-01 5.57355344e-01 -3.93760860e-01 -4.72710907e-01
-1.31029397e-01 6.92477703e-01 4.85215075e-02 1.68218166e-01
-6.67909980e-01 1.37607324e+00 1.03716362e+00 7.05745280e-01
-3.01155478e-01 -4.71916229e-01 -3.16129059e-01 4.14779365e-01
7.11481452e-01 -4.54703778e-01 7.54626393e-01 -8.03346872e-01
-1.77218425e+00 1.32102072e-01 2.20237464e-01 3.18416068e-03
4.92622882e-01 2.05739468e-01 -3.32147598e-01 -7.80861005e-02
-2.69389778e-01 -1.44789293e-01 2.55985081e-01 -1.03001308e+00
-6.89622402e-01 -1.35966375e-01 -6.59620538e-02 2.74477005e-01
-1.08789906e-01 -8.69535208e-02 2.48644620e-01 -4.63139236e-01
-9.56201777e-02 -9.09341395e-01 -2.92016089e-01 -4.87328678e-01
-4.14215803e-01 -1.29905164e-01 9.91230369e-01 -8.21106553e-01
1.41713428e+00 -1.68154919e+00 6.76133215e-01 7.07183242e-01
-4.24659312e-01 3.34147900e-01 2.29083657e-01 8.30065131e-01
-2.45349556e-01 -1.09756850e-01 -3.87879848e-01 1.89317018e-01
1.98654443e-01 3.20770770e-01 4.56747264e-01 4.76881027e-01
1.70344293e-01 7.92689696e-02 -6.38021350e-01 -3.25236917e-01
4.57830817e-01 2.87716359e-01 -3.31832886e-01 1.59830764e-01
-1.89997226e-01 2.40715444e-01 -6.00632489e-01 6.00830555e-01
6.09669983e-01 3.47991496e-01 1.45420447e-01 -3.14273477e-01
-7.04346895e-01 -6.78338051e-01 -1.74317956e+00 1.04710186e+00
-4.61711794e-01 1.58103585e-01 4.20035511e-01 -1.57091773e+00
1.08658016e+00 5.04183292e-01 9.05451179e-01 -3.49315047e-01
5.65171838e-01 4.77136105e-01 2.16869414e-01 -6.58275545e-01
5.43460667e-01 -5.43292873e-02 4.35687214e-01 1.81282759e-01
3.09376307e-02 -3.76466751e-01 9.16748643e-01 -4.40496713e-01
4.90647644e-01 2.53855228e-01 5.06600082e-01 -7.11418331e-01
1.05462575e+00 2.31999457e-01 7.33272135e-01 1.26943141e-02
9.41748247e-02 -6.25600144e-02 3.55804145e-01 1.92989279e-02
-1.29155171e+00 -2.87959307e-01 -9.86625105e-02 4.45220560e-01
8.97146165e-02 1.80390507e-01 -5.33626318e-01 -1.22583762e-01
-3.85578461e-02 1.05068183e+00 -1.20735168e-01 1.14655681e-01
-4.16513711e-01 -1.17851043e+00 -6.87313378e-02 -1.79116845e-01
5.01906991e-01 -7.29862213e-01 -8.98101509e-01 7.72257328e-01
2.28685945e-01 -7.72762001e-01 1.41682297e-01 3.61618072e-01
-7.22191811e-01 -9.43275928e-01 -7.55459845e-01 -7.94841468e-01
8.94152343e-01 -3.17892849e-01 6.31703377e-01 3.20425451e-01
-6.25572145e-01 -8.77829045e-02 -4.57897007e-01 -6.40192032e-01
-4.87646908e-01 1.54581279e-01 -8.53747502e-02 -5.47058955e-02
-3.82599324e-01 -4.35098678e-01 -2.17929468e-01 4.46167737e-01
-8.95234525e-01 -1.55819133e-01 3.87363106e-01 8.51034582e-01
4.41894710e-01 1.23461258e+00 9.32828069e-01 -6.22329712e-01
9.34241533e-01 -5.54098487e-01 -1.25681949e+00 5.62614739e-01
-9.77407277e-01 1.74560741e-01 9.89920318e-01 -5.51947653e-02
-1.11020541e+00 2.15097532e-01 3.77486572e-02 1.86131030e-01
4.52945903e-02 7.53290892e-01 -2.85572797e-01 -3.96116734e-01
1.86373919e-01 7.60335997e-02 7.75642321e-02 -3.33227515e-01
-8.39498043e-02 3.59934807e-01 2.54702002e-01 -6.67186618e-01
9.74551201e-01 -2.26778254e-01 8.93686473e-01 -7.55700111e-01
-1.80651799e-01 -3.25908214e-01 -2.95116991e-01 -5.13353765e-01
7.59708643e-01 2.87242197e-02 -1.10417151e+00 4.88733292e-01
-9.24318135e-01 2.32130021e-01 -8.42382535e-02 4.04132515e-01
-6.60426497e-01 2.69337445e-01 -1.47605300e-01 -1.32168376e+00
-4.79312122e-01 -1.42913115e+00 1.54008716e-01 4.32758152e-01
-6.87343031e-02 -1.08329844e+00 -2.23410010e-01 9.41112489e-02
6.13550842e-01 9.86511171e-01 1.24766743e+00 -1.58239245e-01
-1.86664596e-01 -4.17586297e-01 1.27807423e-01 1.59875363e-01
2.04720646e-01 5.83632290e-01 -2.52312303e-01 -2.14634761e-01
9.51614603e-02 9.90746692e-02 -4.50097807e-02 5.56243241e-01
4.92869556e-01 -1.65661946e-01 -1.61381349e-01 2.61342645e-01
2.18378925e+00 9.73558366e-01 3.09272110e-01 5.76315105e-01
2.06244707e-01 7.88272381e-01 1.08480990e+00 9.20679748e-01
-2.57570863e-01 7.74386823e-01 5.66085041e-01 -2.06177115e-01
5.69539011e-01 3.55915844e-01 -1.70508504e-01 5.53831935e-01
-4.77459848e-01 -7.60769904e-01 -8.36524785e-01 3.87420088e-01
-1.92114043e+00 -9.29367602e-01 -2.53680319e-01 1.85099030e+00
2.58581012e-01 -5.23915514e-02 1.75528258e-01 8.96366656e-01
8.46518457e-01 -3.57955635e-01 -2.23301053e-01 -1.13132346e+00
-6.97075874e-02 2.61755198e-01 7.97390223e-01 4.82067406e-01
-7.60433376e-01 1.77119359e-01 5.42577171e+00 8.77351344e-01
-9.57391560e-01 -2.47202456e-01 2.28766844e-01 7.90775865e-02
7.57388677e-03 -6.62660897e-02 -3.62120032e-01 5.83980560e-01
6.91368997e-01 -5.90603530e-01 8.07903051e-01 7.08716989e-01
6.51682258e-01 -6.10471368e-01 -5.94392002e-01 6.22453094e-01
-1.66487008e-01 -9.43799078e-01 -3.75246495e-01 1.22280270e-01
9.93664920e-01 -7.38084018e-01 -5.10792173e-02 -7.86752552e-02
-8.05005580e-02 -8.87861133e-01 7.65667498e-01 3.71806145e-01
6.26379624e-02 -1.45910466e+00 9.92988765e-01 4.48786199e-01
-1.28812158e+00 -4.43255246e-01 -3.10813934e-01 1.03433564e-01
5.91963947e-01 3.96424264e-01 -8.88559520e-01 1.15777385e+00
2.66181231e-01 1.11577526e-01 -7.74189979e-02 1.44735229e+00
7.77017996e-02 1.33871824e-01 -6.37998641e-01 -3.87287945e-01
5.86959898e-01 -7.64069080e-01 7.62883902e-01 8.12702417e-01
7.78994024e-01 1.65278092e-01 1.07664771e-01 1.12065852e+00
6.27057076e-01 6.97538018e-01 -4.22921389e-01 -4.08882856e-01
3.09181660e-01 1.27579665e+00 -1.03405452e+00 1.27673388e-01
-1.28638193e-01 5.50878823e-01 -3.30138147e-01 1.57362297e-01
-9.64055181e-01 -7.50950456e-01 1.19809017e-01 -8.06285962e-02
3.52792472e-01 -3.01535726e-01 -2.11663112e-01 -2.18496293e-01
-1.68473408e-01 -6.23063147e-01 4.14142370e-01 -5.31900823e-01
-9.18612659e-01 5.58159232e-01 3.31252038e-01 -1.10381675e+00
-4.10459995e-01 -7.55285442e-01 -6.80516481e-01 1.01807988e+00
-1.44624758e+00 -8.17829788e-01 -2.31774360e-01 3.31592441e-01
6.93121791e-01 -4.34842169e-01 6.13091826e-01 5.28624475e-01
-9.36136246e-01 7.74410591e-02 6.61426425e-01 -7.58821249e-01
-1.18187673e-01 -1.11376798e+00 -6.42409384e-01 1.04512453e+00
-7.53833830e-01 9.08907130e-02 1.17426038e+00 -5.52663743e-01
-1.47215819e+00 -4.56070065e-01 7.21091211e-01 6.07288778e-01
5.08101046e-01 2.60296881e-01 -3.87704492e-01 -1.51715204e-02
4.34882164e-01 -4.96693045e-01 2.44403347e-01 -4.43508923e-01
8.53884399e-01 -3.68714109e-02 -1.64209032e+00 1.99585795e-01
1.66412368e-01 2.90933222e-01 -2.17467085e-01 2.56378770e-01
1.32904544e-01 -2.74115086e-01 -9.42598283e-01 3.35896462e-01
2.69046962e-01 -5.40897191e-01 7.85040319e-01 -2.26498321e-01
5.79370037e-02 -6.67345166e-01 7.29472563e-02 -1.39439583e+00
-4.96179879e-01 -5.37086844e-01 3.16230386e-01 1.47590089e+00
5.23744345e-01 -6.41379118e-01 4.46995080e-01 7.16740727e-01
-2.63888031e-01 -8.16680849e-01 -7.15485513e-01 -8.53466451e-01
-3.64808083e-01 2.54580021e-01 5.87250292e-01 7.13442802e-01
-7.87691921e-02 9.61266011e-02 -1.71374217e-01 2.51679629e-01
5.47595561e-01 3.03095095e-02 1.09548226e-01 -1.16809714e+00
-4.52087998e-01 -4.93666619e-01 -2.51869291e-01 2.62944162e-01
-7.12390020e-02 -6.20093942e-01 8.99700895e-02 -1.60609615e+00
-4.62006629e-01 -5.31492472e-01 -6.50386661e-02 2.32508853e-02
4.52029891e-03 -4.46130693e-01 3.53667468e-01 -2.55538076e-01
2.72288561e-01 5.27464330e-01 1.11812532e+00 -9.57583934e-02
-3.63058299e-01 1.86701611e-01 -9.92575288e-02 4.64884162e-01
9.55432653e-01 -4.22389865e-01 -7.86849141e-01 -1.65092424e-01
2.57598996e-01 1.84217542e-01 -1.91312730e-01 -1.13545966e+00
-1.11477147e-03 -4.65434283e-01 1.06718294e-01 -7.43745565e-01
1.72176972e-01 -1.45199645e+00 8.85352433e-01 7.26340175e-01
1.95849746e-01 6.96295798e-01 1.68732107e-01 2.67955124e-01
-3.46618652e-01 -1.07381213e+00 9.03349161e-01 -2.86804363e-02
-7.09824085e-01 -1.85980037e-01 -4.56827879e-01 -3.52770388e-01
1.64399290e+00 -4.84833896e-01 -1.30618736e-01 1.41966781e-02
-7.79460073e-01 3.94601315e-01 5.96784651e-02 6.96892664e-03
6.73049465e-02 -9.69053328e-01 -7.20865488e-01 -3.44834864e-01
-4.19734865e-01 -2.99633682e-01 1.25573337e-01 8.58663619e-01
-1.31861806e+00 5.43764055e-01 -7.25939393e-01 -1.41688704e-01
-1.18737721e+00 6.61195695e-01 3.67658705e-01 -3.92497718e-01
-9.12623405e-02 4.17292953e-01 -6.17047310e-01 7.78249577e-02
-4.55226526e-02 -1.41689867e-01 -5.02377570e-01 3.01900983e-01
-1.35638028e-01 1.07626224e+00 1.62409022e-01 -5.96746624e-01
-3.85557711e-01 6.57926977e-01 7.32606649e-01 -1.66514993e-01
1.63844097e+00 -1.94885228e-02 -2.62813509e-01 -7.43089011e-03
7.96538293e-01 -1.32241935e-01 -5.68467855e-01 4.63315815e-01
1.72007546e-01 -3.15474123e-01 3.80568713e-01 -9.49131727e-01
-1.12900293e+00 2.88824409e-01 6.48091078e-01 4.09320951e-01
1.53263676e+00 -1.04465759e+00 4.81175125e-01 3.22728246e-01
2.37204790e-01 -1.64699209e+00 -4.33351725e-01 3.23008060e-01
1.00687551e+00 -6.88352525e-01 2.77823687e-01 -5.90656459e-01
-4.11704987e-01 1.49189520e+00 3.98682564e-01 -7.11412728e-02
6.33686483e-01 3.55022311e-01 -5.60395122e-01 8.22683498e-02
-2.99073547e-01 -3.55786383e-01 1.24594748e-01 3.75174075e-01
3.57231855e-01 -5.39228842e-02 -1.22471106e+00 3.92616093e-01
9.17085970e-04 1.52519390e-01 5.52846074e-01 1.30983496e+00
-4.92731363e-01 -1.27625453e+00 -8.03976178e-01 8.99201259e-02
-5.46423435e-01 4.10938561e-01 -4.37309109e-02 1.09723890e+00
5.58283508e-01 1.27026582e+00 -5.30543864e-01 8.50403309e-02
6.83800519e-01 -1.28240392e-01 3.68182600e-01 -2.04284087e-01
-8.62316370e-01 1.24761924e-01 2.68229932e-01 1.40747717e-02
-4.08224046e-01 -6.37416601e-01 -1.39470041e+00 -3.93647224e-01
-8.08465004e-01 6.64215982e-01 1.16726172e+00 8.29035461e-01
-7.80810490e-02 8.99062872e-01 8.45973253e-01 -6.39838755e-01
-5.25454760e-01 -5.95269859e-01 -9.53485012e-01 3.86574939e-02
-1.32384568e-01 -1.02647889e+00 -2.11200774e-01 -1.10526644e-01] | [5.728173732757568, 3.4316728115081787] |
ccd97aaa-675c-426d-bd09-682432b05c70 | recurrent-saliency-transformation-network | 1709.04518 | null | http://arxiv.org/abs/1709.04518v4 | http://arxiv.org/pdf/1709.04518v4.pdf | Recurrent Saliency Transformation Network: Incorporating Multi-Stage Visual Cues for Small Organ Segmentation | We aim at segmenting small organs (e.g., the pancreas) from abdominal CT
scans. As the target often occupies a relatively small region in the input
image, deep neural networks can be easily confused by the complex and variable
background. To alleviate this, researchers proposed a coarse-to-fine approach,
which used prediction from the first (coarse) stage to indicate a smaller input
region for the second (fine) stage. Despite its effectiveness, this algorithm
dealt with two stages individually, which lacked optimizing a global energy
function, and limited its ability to incorporate multi-stage visual cues.
Missing contextual information led to unsatisfying convergence in iterations,
and that the fine stage sometimes produced even lower segmentation accuracy
than the coarse stage.
This paper presents a Recurrent Saliency Transformation Network. The key
innovation is a saliency transformation module, which repeatedly converts the
segmentation probability map from the previous iteration as spatial weights and
applies these weights to the current iteration. This brings us two-fold
benefits. In training, it allows joint optimization over the deep networks
dealing with different input scales. In testing, it propagates multi-stage
visual information throughout iterations to improve segmentation accuracy.
Experiments in the NIH pancreas segmentation dataset demonstrate the
state-of-the-art accuracy, which outperforms the previous best by an average of
over 2%. Much higher accuracies are also reported on several small organs in a
larger dataset collected by ourselves. In addition, our approach enjoys better
convergence properties, making it more efficient and reliable in practice. | ['Elliot K. Fishman', 'Yuyin Zhou', 'Qihang Yu', 'Lingxi Xie', 'Alan L. Yuille', 'Yan Wang'] | 2017-09-13 | recurrent-saliency-transformation-network-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Yu_Recurrent_Saliency_Transformation_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Yu_Recurrent_Saliency_Transformation_CVPR_2018_paper.pdf | cvpr-2018-6 | ['pancreas-segmentation'] | ['medical'] | [ 1.89531460e-01 2.20828786e-01 -3.26348573e-01 -2.26414442e-01
-5.92221797e-01 -2.36962184e-01 1.67027116e-01 2.67111361e-01
-4.74082381e-01 6.62995398e-01 6.41851425e-02 -1.02263495e-01
1.56333461e-01 -6.25285923e-01 -7.10444927e-01 -8.61686110e-01
1.16284102e-01 2.87316889e-01 6.64859593e-01 1.29151568e-01
6.15080632e-02 3.22660446e-01 -9.92407322e-01 2.48341218e-01
1.15313721e+00 9.35430288e-01 6.10797465e-01 3.23606342e-01
-3.33330423e-01 5.74345589e-01 -5.03065586e-01 -1.90883353e-01
6.11829571e-02 -6.31625593e-01 -7.10884213e-01 4.44872715e-02
1.30658627e-01 -1.06729761e-01 -7.33810142e-02 1.36510503e+00
3.84501547e-01 -2.46350318e-02 3.70291203e-01 -1.07075667e+00
-6.50326729e-01 8.55171740e-01 -7.90695012e-01 3.53084117e-01
-1.00680299e-01 1.50514454e-01 6.34377420e-01 -6.19021833e-01
5.07553399e-01 8.76946688e-01 8.63711059e-01 5.38402438e-01
-1.31648278e+00 -5.78320742e-01 2.91415215e-01 -8.89637694e-02
-1.27318883e+00 1.28642276e-01 8.10816348e-01 -3.16246301e-01
5.28145254e-01 2.63607383e-01 1.01488221e+00 8.21030140e-01
4.34018850e-01 8.87001932e-01 9.88894045e-01 -1.95055902e-01
7.54894614e-02 2.43643388e-01 6.20573051e-02 1.00586498e+00
3.55809659e-01 4.07462008e-02 3.09107266e-02 8.42962936e-02
1.29986775e+00 2.44664282e-01 -4.18486774e-01 -5.62244654e-01
-1.55126464e+00 8.62204969e-01 1.10922706e+00 6.30281687e-01
-5.37804842e-01 -7.98252150e-02 2.56427497e-01 -1.23385519e-01
1.83587641e-01 4.19747680e-01 -3.01977932e-01 2.17538536e-01
-1.13153052e+00 -3.42550248e-01 6.79841459e-01 5.56536555e-01
5.32925308e-01 -5.35113961e-02 -3.84093285e-01 7.30362654e-01
3.12952697e-01 1.26861975e-01 7.81577647e-01 -5.81388235e-01
2.97701418e-01 6.52038395e-01 -2.28797048e-01 -8.49991024e-01
-8.76776338e-01 -7.64966011e-01 -1.18831730e+00 4.87926573e-01
7.66687691e-01 -1.53580070e-01 -1.48181319e+00 1.77432108e+00
3.77341270e-01 9.54711661e-02 -1.78752542e-01 1.36674178e+00
9.38402414e-01 4.65373397e-01 4.22824264e-01 -1.73131049e-01
1.38005030e+00 -1.30131781e+00 -5.72966635e-01 -3.17333996e-01
2.99797505e-01 -6.92841113e-01 1.03914952e+00 2.19873294e-01
-1.14678347e+00 -5.49628675e-01 -1.07224047e+00 2.22460121e-01
-2.16471538e-01 2.81131446e-01 7.30784059e-01 3.93779695e-01
-1.21460509e+00 6.49545252e-01 -1.15194583e+00 -3.91516209e-01
6.19034469e-01 2.52140641e-01 -5.50330989e-02 3.66224289e-01
-1.05853510e+00 1.04720497e+00 4.05771881e-01 6.66784793e-02
-5.29319465e-01 -8.36774349e-01 -8.30382764e-01 2.01393664e-01
3.12851310e-01 -8.43385220e-01 1.08641374e+00 -1.33798885e+00
-1.50983667e+00 7.66799331e-01 -3.66108976e-02 -4.06940520e-01
8.14828396e-01 2.20478460e-01 -1.16667673e-01 2.77715147e-01
1.55247360e-01 8.95824969e-01 7.31403470e-01 -1.17814112e+00
-5.72472572e-01 -1.74224541e-01 -1.44570246e-01 2.61517555e-01
2.61280853e-02 -3.48376721e-01 -8.18226933e-01 -1.00574005e+00
5.62226355e-01 -8.56690764e-01 -5.89613914e-01 2.88125217e-01
-4.19473529e-01 1.00642659e-01 4.84553874e-01 -6.27627492e-01
1.19878376e+00 -2.07800007e+00 2.01004744e-01 2.41578698e-01
3.32949549e-01 1.38159260e-01 7.36745968e-02 -3.24154258e-01
-2.81697452e-01 1.22861847e-01 -3.76528591e-01 -8.54248032e-02
-3.87270749e-01 1.32414460e-01 9.31733176e-02 5.10636806e-01
1.44570574e-01 1.07782400e+00 -9.20373619e-01 -9.11974490e-01
3.47281516e-01 4.13446665e-01 -4.98317808e-01 -6.35959655e-02
8.68080109e-02 4.99141842e-01 -4.73733574e-01 9.11530137e-01
5.36597311e-01 -7.70237863e-01 6.65999129e-02 -4.49022770e-01
-1.52310506e-01 -1.48461461e-01 -1.11906695e+00 1.84933329e+00
-2.08446234e-01 4.83073324e-01 1.25703946e-01 -1.00285244e+00
5.96569657e-01 2.38826707e-01 6.02700591e-01 -7.61267722e-01
2.33661875e-01 1.78646743e-01 8.71774256e-02 -4.19273645e-01
2.33910143e-01 -2.54016787e-01 1.28699364e-02 8.50475580e-02
1.88360345e-02 -6.58130506e-03 1.51192039e-01 6.67328164e-02
6.82612121e-01 6.95912391e-02 4.81895149e-01 -4.25651103e-01
3.59214783e-01 2.14136511e-01 7.87033081e-01 7.70875752e-01
-5.29291511e-01 9.25278306e-01 4.81742561e-01 -5.34486234e-01
-8.35984051e-01 -1.19936132e+00 -2.09647119e-01 7.23676145e-01
7.24297106e-01 1.55852124e-01 -7.65124083e-01 -9.36597168e-01
-8.18617120e-02 3.54295135e-01 -8.65411401e-01 -1.60549819e-01
-6.70839906e-01 -9.89113748e-01 3.44711035e-01 7.12870538e-01
5.48877060e-01 -1.15480697e+00 -8.83783460e-01 4.15377647e-01
-2.33929947e-01 -7.18843400e-01 -6.44061506e-01 3.46506864e-01
-1.17395782e+00 -1.01760948e+00 -1.32285714e+00 -1.10347390e+00
1.09550071e+00 3.37913856e-02 1.08884144e+00 2.44562432e-01
-4.87953901e-01 -3.13685954e-01 -4.94095050e-02 -2.34642848e-01
-3.56596559e-01 5.62953837e-02 -3.47039312e-01 -3.33308607e-01
-1.04726017e-01 -2.03484952e-01 -8.28070700e-01 3.30459505e-01
-8.15460920e-01 3.70394975e-01 9.59770501e-01 1.10746241e+00
7.42257893e-01 -1.84921116e-01 4.35094804e-01 -7.64176607e-01
3.50752056e-01 -3.22583199e-01 -5.43402672e-01 3.95844549e-01
-5.85190237e-01 1.28348306e-01 6.57779634e-01 -7.00370491e-01
-9.43654656e-01 1.89292550e-01 -2.86966190e-02 -3.72439325e-01
-1.36033103e-01 3.85835022e-01 3.94060493e-01 -1.97399825e-01
5.48144519e-01 1.18709333e-01 1.75414398e-01 -3.19040388e-01
-3.98240704e-03 -2.61541624e-02 5.28450727e-01 -8.75279978e-02
4.66444552e-01 3.13875556e-01 -1.49651468e-01 -2.38442466e-01
-6.13860250e-01 -1.84027076e-01 -5.18209398e-01 -2.26536155e-01
9.38736141e-01 -7.43530810e-01 -4.98066872e-01 4.10684854e-01
-7.98308909e-01 -4.20125216e-01 -3.38521063e-01 6.50963902e-01
-6.15462624e-02 4.91596282e-01 -8.17981839e-01 -2.38736197e-01
-3.86833161e-01 -1.47120845e+00 7.48020113e-01 8.04693937e-01
-2.24817842e-01 -1.16689432e+00 -1.88544139e-01 -2.95347217e-02
6.66205466e-01 4.18213576e-01 7.92800605e-01 -4.29185182e-01
-5.03650248e-01 1.38009161e-01 -4.68977749e-01 8.57120845e-03
2.24923760e-01 7.53264651e-02 -5.79695761e-01 -2.58938223e-01
-6.07652776e-02 8.09411425e-03 1.01773858e+00 8.43715966e-01
1.32621050e+00 -8.81107748e-02 -6.18229091e-01 8.10748994e-01
1.53093505e+00 3.01511139e-01 3.66506487e-01 4.88308728e-01
6.69977248e-01 1.64377615e-01 6.21672153e-01 1.56910807e-01
2.63049930e-01 4.92838353e-01 4.89803672e-01 -9.10120666e-01
-4.11846161e-01 2.70300005e-02 -2.50726379e-03 6.00147069e-01
7.47887045e-02 1.20744221e-01 -8.87004614e-01 6.47272885e-01
-1.72250128e+00 -7.34332800e-01 1.60370674e-02 1.97775590e+00
1.04757524e+00 1.89097628e-01 8.76884311e-02 -2.72857964e-01
7.96759665e-01 9.08588693e-02 -6.50796175e-01 -1.14336468e-01
-9.46851913e-03 1.23119861e-01 6.13657355e-01 3.78622085e-01
-1.30696595e+00 7.03751385e-01 6.14647865e+00 7.89502382e-01
-1.67344856e+00 9.81197506e-02 9.79263723e-01 -4.82648425e-02
-2.33819529e-01 -2.76522666e-01 -4.66826528e-01 6.91705465e-01
2.13063180e-01 5.77989370e-02 2.07377449e-01 7.76829004e-01
-4.95754182e-02 -4.32899535e-01 -8.77243221e-01 8.47456515e-01
-2.44544670e-02 -1.40932763e+00 -8.42832178e-02 -2.13282526e-01
8.07310164e-01 1.81963265e-01 1.61298434e-03 1.61878273e-01
4.63823751e-02 -9.29652929e-01 8.08398604e-01 4.37788606e-01
7.28323996e-01 -6.53076351e-01 8.47875893e-01 3.67237240e-01
-1.16487265e+00 5.81981726e-02 -2.17478558e-01 3.57153565e-01
1.73951671e-01 6.19758725e-01 -9.24095273e-01 3.43922436e-01
8.18103194e-01 3.80342335e-01 -4.82711375e-01 1.47999859e+00
-1.78320065e-01 2.36205146e-01 -3.92159969e-01 -2.27435306e-01
4.72845763e-01 -3.30964953e-01 5.80812454e-01 1.40076780e+00
3.60454470e-01 -5.14130667e-02 2.49218702e-01 1.03455806e+00
3.46043520e-02 5.06637357e-02 -2.16399536e-01 3.73894751e-01
1.14437103e-01 1.49468458e+00 -1.41165423e+00 -6.18017614e-01
-2.93293297e-01 1.01677215e+00 3.15286905e-01 4.15916502e-01
-1.21080601e+00 -3.73231769e-01 2.38148019e-01 -1.17438033e-01
3.67744505e-01 2.01293364e-01 -8.21402371e-01 -1.16822755e+00
-4.59749103e-02 -4.65294331e-01 5.29733777e-01 -4.90220845e-01
-1.11751747e+00 6.98892772e-01 -2.63500005e-01 -1.30100822e+00
-3.30904089e-02 -3.42403471e-01 -7.00897098e-01 8.94191921e-01
-1.48327947e+00 -8.83409202e-01 -4.92056996e-01 4.15897429e-01
6.40309274e-01 3.39780360e-01 6.13970757e-01 2.63311476e-01
-5.71817219e-01 6.34994090e-01 -7.21154362e-02 2.39773884e-01
6.40666306e-01 -1.41009271e+00 -5.32089248e-02 9.03369546e-01
-2.16988400e-02 4.12523419e-01 5.56073189e-01 -6.72246397e-01
-7.12721646e-01 -9.47476983e-01 6.34901047e-01 -2.55089700e-02
4.58777159e-01 6.33106306e-02 -1.01933491e+00 4.02433336e-01
1.57212198e-01 1.25872120e-01 3.36711794e-01 -3.26398671e-01
2.72077233e-01 4.23467020e-03 -1.38546169e+00 7.43175447e-01
6.15180016e-01 1.81012712e-02 -7.54599690e-01 -6.68470711e-02
5.08290589e-01 -9.59312320e-01 -7.74496257e-01 6.11447990e-01
5.22620618e-01 -8.73912990e-01 1.13192534e+00 -8.63863006e-02
3.94487411e-01 -3.84047002e-01 4.72579956e-01 -1.40874720e+00
-6.10967278e-01 -2.15380862e-01 5.29698879e-02 8.66013467e-01
6.39049828e-01 -5.78193247e-01 8.03521812e-01 6.11757278e-01
-2.06779376e-01 -1.03748500e+00 -8.81729007e-01 -3.33164722e-01
-8.13995227e-02 7.41165206e-02 3.06372195e-01 1.09513223e+00
-3.36512364e-02 -8.51404592e-02 -2.19725128e-02 2.23055616e-01
6.98414743e-01 5.09648800e-01 1.41451925e-01 -1.09329128e+00
-1.13320947e-01 -8.35090816e-01 -1.77091822e-01 -1.00536406e+00
-3.43489021e-01 -1.00280535e+00 1.72116831e-01 -1.69771814e+00
4.45987761e-01 -5.78623652e-01 -6.11857176e-01 7.20249116e-01
-6.23356938e-01 5.67049563e-01 2.51714230e-01 2.01337859e-01
-3.86857390e-01 2.04032511e-01 1.77852094e+00 -2.33868688e-01
-4.92225796e-01 1.25500813e-01 -8.31934333e-01 8.63551855e-01
8.32743883e-01 -5.12045622e-01 -7.76866376e-02 -3.23134094e-01
-3.64428908e-01 1.42028213e-01 4.45888966e-01 -9.46088254e-01
4.36575562e-01 1.79904196e-02 9.78544950e-01 -5.83722293e-01
-1.83058064e-02 -7.98666477e-01 2.27499828e-01 8.49471092e-01
-2.07728371e-01 -1.72102794e-01 3.59702975e-01 3.20026159e-01
-3.00843179e-01 -2.20620960e-01 1.15053189e+00 -3.35016787e-01
-7.80647695e-01 2.06863865e-01 -2.19590247e-01 -3.47380899e-02
1.16230595e+00 -4.82273102e-01 5.28987544e-03 2.60290205e-02
-1.06443989e+00 3.06395173e-01 5.96525133e-01 2.47999415e-01
5.26672125e-01 -1.20454514e+00 -4.77434784e-01 3.67022097e-01
-3.13419938e-01 2.39953384e-01 3.21098238e-01 1.40911198e+00
-7.66785979e-01 1.37917474e-01 -3.34725946e-01 -1.06281209e+00
-1.12789464e+00 4.46634322e-01 4.83101159e-01 -5.15726626e-01
-9.30554628e-01 9.86114323e-01 4.57370043e-01 -5.53277098e-02
4.11602914e-01 -8.95430326e-01 -2.97128141e-01 8.61313567e-02
2.76463062e-01 5.81885278e-02 -1.04122072e-01 -4.00732070e-01
-4.69852090e-01 6.84508502e-01 -1.05396919e-01 1.99531376e-01
1.23248672e+00 -1.06610931e-01 5.39792888e-02 2.46586502e-01
1.12370157e+00 -1.77268431e-01 -1.42517734e+00 -2.46067613e-01
-1.95137188e-01 -3.93341184e-01 2.42798150e-01 -1.09535193e+00
-1.50983751e+00 6.09420061e-01 7.96058118e-01 2.42454559e-01
1.27702403e+00 8.95911753e-02 7.97541916e-01 -1.79117173e-01
2.20252454e-01 -1.08226180e+00 -5.46453558e-02 3.23498517e-01
7.24331856e-01 -1.46582067e+00 -5.71932346e-02 -2.71919876e-01
-9.80701387e-01 1.25036311e+00 7.05193102e-01 -5.49363345e-02
4.73503381e-01 4.19037819e-01 3.75072479e-01 -1.13586709e-01
-2.74899840e-01 -6.51333332e-02 4.39232290e-01 2.06657007e-01
4.26317960e-01 6.83285147e-02 -3.80406559e-01 5.80255628e-01
1.02896884e-01 8.92418846e-02 2.83280253e-01 8.18480730e-01
-4.32243645e-01 -6.81451917e-01 -4.04361367e-01 5.08161724e-01
-7.78873086e-01 -2.43484490e-02 9.25028473e-02 1.06194198e+00
1.21573307e-01 4.23396349e-01 4.31491695e-02 8.38575214e-02
2.70093739e-01 -2.50129342e-01 2.83378452e-01 -2.67664582e-01
-8.66490066e-01 4.53224689e-01 -3.67820263e-01 -8.71510744e-01
-3.76088411e-01 -5.77414691e-01 -1.68223882e+00 1.72430381e-01
-2.71636933e-01 8.51119235e-02 4.97637242e-01 6.10567629e-01
1.39236860e-02 1.02654552e+00 3.81057590e-01 -1.08141482e+00
-3.71680915e-01 -7.78810501e-01 -4.43185776e-01 4.32218999e-01
5.47231436e-01 -5.54128408e-01 -3.04202050e-01 -8.91186204e-03] | [14.563542366027832, -2.572664737701416] |
2e248361-037f-4146-a2ba-a3f06de37426 | learning-to-reuse-distractors-to-support | 2210.13964 | null | https://arxiv.org/abs/2210.13964v2 | https://arxiv.org/pdf/2210.13964v2.pdf | Learning to Reuse Distractors to support Multiple Choice Question Generation in Education | Multiple choice questions (MCQs) are widely used in digital learning systems, as they allow for automating the assessment process. However, due to the increased digital literacy of students and the advent of social media platforms, MCQ tests are widely shared online, and teachers are continuously challenged to create new questions, which is an expensive and time-consuming task. A particularly sensitive aspect of MCQ creation is to devise relevant distractors, i.e., wrong answers that are not easily identifiable as being wrong. This paper studies how a large existing set of manually created answers and distractors for questions over a variety of domains, subjects, and languages can be leveraged to help teachers in creating new MCQs, by the smart reuse of existing distractors. We built several data-driven models based on context-aware question and distractor representations, and compared them with static feature-based models. The proposed models are evaluated with automated metrics and in a realistic user test with teachers. Both automatic and human evaluations indicate that context-aware models consistently outperform a static feature-based approach. For our best-performing context-aware model, on average 3 distractors out of the 10 shown to teachers were rated as high-quality distractors. We create a performance benchmark, and make it public, to enable comparison between different approaches and to introduce a more standardized evaluation of the task. The benchmark contains a test of 298 educational questions covering multiple subjects & languages and a 77k multilingual pool of distractor vocabulary for future research. | ['Thomas Demeester', 'Chris Develder', 'Johannes Deleu', 'Lucas Sterckx', 'Amir Hadifar', 'Semere Kiros Bitew'] | 2022-10-25 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [-3.71144116e-01 -1.11535117e-02 1.74521759e-01 -3.48891228e-01
-1.10573947e+00 -1.15143836e+00 5.04365087e-01 5.54890692e-01
-4.06641215e-01 6.54522598e-01 2.10170791e-01 -5.80041111e-01
-4.82498676e-01 -6.73522234e-01 -4.90890145e-01 -8.81153196e-02
4.74896073e-01 4.31141168e-01 9.48160648e-01 -6.25349522e-01
4.61449385e-01 2.93187231e-01 -2.34071684e+00 3.16532433e-01
1.61433744e+00 6.21906459e-01 4.57437485e-01 6.39484704e-01
-6.44777417e-01 8.84556651e-01 -1.17700398e+00 -7.49395549e-01
-2.36508057e-01 -5.23898304e-01 -9.16381121e-01 -1.09595455e-01
1.25520861e+00 -1.95519403e-01 1.14873968e-01 1.02550542e+00
6.67509496e-01 3.79969358e-01 4.55836743e-01 -1.07449210e+00
-1.05123615e+00 4.95585799e-01 2.02634782e-01 3.11838001e-01
1.08064568e+00 1.16198838e-01 8.81978691e-01 -1.10893083e+00
5.54292023e-01 1.03309619e+00 3.32070053e-01 5.96757710e-01
-8.70270967e-01 -7.56041050e-01 2.02537570e-02 7.25732744e-01
-1.16295612e+00 -2.17685789e-01 4.32778329e-01 -7.45560229e-01
7.60658979e-01 5.31482935e-01 6.86446905e-01 9.22128797e-01
9.20749605e-02 5.69601357e-01 1.36397684e+00 -6.35135114e-01
1.83483467e-01 6.45876706e-01 2.31349736e-01 4.38333035e-01
2.68379897e-01 -4.94950444e-01 -4.38283652e-01 7.27332309e-02
1.57376423e-01 -2.03028619e-01 -4.23105836e-01 -1.98954359e-01
-6.83610439e-01 6.39465451e-01 -6.63178936e-02 5.01461387e-01
8.50184262e-02 -6.09278619e-01 -1.90880284e-01 7.69163132e-01
1.75088570e-01 7.85817564e-01 -6.53567314e-01 -5.59650362e-01
-6.07073605e-01 6.77256286e-01 1.04960871e+00 1.10719001e+00
5.29733956e-01 -2.71914929e-01 -5.40676951e-01 1.16186523e+00
3.19477260e-01 6.79849803e-01 9.37903166e-01 -7.58169472e-01
2.99834222e-01 1.14625812e+00 8.57883915e-02 -9.47355568e-01
-1.24239875e-02 -4.36026692e-01 1.86072178e-02 2.28250958e-02
7.69200325e-01 9.52048823e-02 -6.73571885e-01 1.50935888e+00
6.07414186e-01 -8.35234523e-02 -2.69544333e-01 5.76730609e-01
1.41205096e+00 1.37348443e-01 -3.36082317e-02 -1.20553246e-03
1.47727668e+00 -7.51364410e-01 -1.05592334e+00 1.23606570e-01
8.95929098e-01 -1.35991585e+00 1.64908719e+00 7.14516222e-01
-1.21233857e+00 -5.75343907e-01 -7.30680406e-01 -2.78717250e-01
-6.89914465e-01 -2.56992787e-01 -1.05192289e-01 8.64349246e-01
-9.04496908e-01 2.16433018e-01 1.33311346e-01 -1.67394951e-01
3.36405523e-02 -3.26288342e-02 -1.44143164e-01 -2.72442609e-01
-1.13131821e+00 1.02532446e+00 -2.40800694e-01 -7.01625705e-01
-8.73883665e-01 -8.74254465e-01 -6.82868183e-01 1.24245524e-01
5.84232688e-01 -6.61255270e-02 1.76535439e+00 -6.47001266e-01
-1.56575406e+00 8.22324395e-01 1.91469237e-01 3.27278823e-01
4.38639790e-01 -4.61875588e-01 -7.16212213e-01 -4.51709293e-02
3.29173744e-01 1.03731975e-01 6.74373984e-01 -7.85837114e-01
-8.50301325e-01 -9.98818353e-02 3.30370307e-01 1.87693313e-01
-6.23634040e-01 6.96230829e-02 -3.54446173e-01 -4.54080790e-01
9.29021016e-02 -5.73213339e-01 2.14868724e-01 -2.90000677e-01
3.35573591e-02 -8.95482540e-01 5.58720887e-01 -5.11037350e-01
1.81612635e+00 -1.84931576e+00 -2.82143325e-01 3.87295783e-02
3.92012060e-01 5.74341118e-01 -2.45622665e-01 4.54632789e-01
1.36595339e-01 1.86603412e-01 4.12840128e-01 2.07760990e-01
1.94081903e-01 7.88146034e-02 -1.03803410e-03 -1.31993771e-01
-3.21365856e-02 4.78925467e-01 -1.21590817e+00 -6.44977212e-01
2.05071315e-01 6.85247257e-02 -6.33612752e-01 6.27889693e-01
-3.88801008e-01 2.62419552e-01 -3.24063271e-01 5.51772296e-01
5.05342305e-01 -2.12785397e-02 -5.04670329e-02 4.61329222e-01
-2.22807452e-01 8.36663187e-01 -1.54173553e+00 1.40033770e+00
-6.79715335e-01 5.12049794e-01 -2.50592560e-01 -3.43034714e-01
9.36965466e-01 4.12940294e-01 5.94439879e-02 -9.88127649e-01
2.31522769e-02 4.52983707e-01 8.00155327e-02 -1.13373852e+00
6.37295425e-01 2.81063706e-01 6.87704012e-02 4.58986402e-01
4.47869360e-01 -5.96453846e-01 5.86748779e-01 3.90089869e-01
1.22764182e+00 5.96245490e-02 3.24860603e-01 -3.14865828e-01
8.89235198e-01 -1.28697231e-01 2.29966477e-01 7.25542128e-01
-3.82042468e-01 4.42786306e-01 1.73293278e-01 -6.73066676e-02
-2.43221223e-01 -1.04311693e+00 -1.33477494e-01 1.59797001e+00
-2.66499966e-01 -5.67136467e-01 -7.95196354e-01 -9.68892932e-01
1.98516343e-02 8.98551643e-01 -7.22776800e-02 -1.54600767e-02
-3.01451653e-01 1.05297074e-01 3.20366323e-01 3.33602279e-02
1.56579211e-01 -8.43508959e-01 -3.64218235e-01 4.01021898e-01
-3.44929785e-01 -8.80943000e-01 -4.45699275e-01 -2.38511994e-01
-3.97777528e-01 -1.39389324e+00 -3.51124495e-01 -8.32162261e-01
4.28694695e-01 5.88242650e-01 1.73907888e+00 5.14150143e-01
-6.35048747e-02 8.84879827e-01 -6.07133508e-01 -6.15215361e-01
-5.46473324e-01 -4.44336981e-02 -7.64721856e-02 -4.40976173e-01
8.49418044e-01 -4.51230317e-01 -5.70150673e-01 4.49181318e-01
-1.07638359e+00 -3.51998031e-01 3.63385558e-01 5.05575895e-01
4.02295917e-01 -3.08209211e-01 8.56153369e-01 -9.32910025e-01
1.13365388e+00 -6.83147132e-01 -8.70294452e-01 4.65168118e-01
-7.48000026e-01 -1.05636522e-01 6.15332961e-01 -7.43487597e-01
-9.67235923e-01 -4.90915567e-01 -5.51479459e-01 -1.26413152e-01
-2.96321720e-01 3.54653060e-01 -2.10863546e-01 -4.72037137e-01
1.01035273e+00 9.69382450e-02 -2.39201620e-01 -8.29638064e-01
4.38749522e-01 8.87552977e-01 2.44987041e-01 -8.26251984e-01
8.21405649e-01 -5.08870184e-01 -3.83962214e-01 -7.89786994e-01
-1.17117429e+00 -6.43948138e-01 -4.43913162e-01 -7.21037745e-01
3.34057927e-01 -7.93965816e-01 -8.80440295e-01 1.28117174e-01
-8.95983577e-01 -8.41761902e-02 -4.61698800e-01 5.11718392e-01
-5.77687174e-02 1.56603187e-01 -2.92803288e-01 -5.33635497e-01
1.94668725e-01 -1.23584414e+00 4.25927192e-01 6.26932621e-01
-2.12762967e-01 -1.00067234e+00 3.27045500e-01 9.17193413e-01
6.19293034e-01 -3.17103207e-01 1.08355081e+00 -1.32542145e+00
-5.74591994e-01 -1.99354321e-01 2.00199470e-01 5.19926071e-01
1.07811771e-01 1.69128686e-01 -1.02999842e+00 -9.31537375e-02
9.26433131e-02 -5.64253688e-01 5.36275059e-02 -2.97476888e-01
1.07537580e+00 -4.57911700e-01 2.32671112e-01 -1.02013625e-01
1.25001097e+00 -6.16652565e-03 3.74568701e-01 3.13542515e-01
4.82549220e-01 8.25049758e-01 6.81352675e-01 1.63144380e-01
1.00529695e+00 5.51530480e-01 3.68415803e-01 5.99256575e-01
-5.79048872e-01 -4.00302589e-01 3.18957835e-01 1.73188424e+00
3.13880980e-01 -1.45636350e-01 -1.02860916e+00 9.44270134e-01
-1.19814658e+00 -6.92847908e-01 -6.17560327e-01 2.46755242e+00
1.24410570e+00 -5.31936474e-02 8.03196337e-03 2.71687597e-01
3.13398659e-01 -3.97128612e-01 -1.14746364e-02 -4.47248816e-01
2.01776162e-01 8.09463024e-01 -2.39881814e-01 6.57903135e-01
-3.74728024e-01 6.90739512e-01 6.03749514e+00 1.05655539e+00
-8.44494343e-01 3.15612108e-01 8.69946182e-02 -3.11818887e-02
-7.71950305e-01 -2.51618743e-01 -1.06962764e+00 5.14366150e-01
9.82841551e-01 -3.46196026e-01 1.41014725e-01 8.21418822e-01
6.08590357e-02 -2.09499598e-01 -9.68690932e-01 6.81810439e-01
2.66394287e-01 -9.40462708e-01 8.24444517e-02 -1.80082053e-01
1.01337123e+00 -2.23563582e-01 1.13961793e-01 7.59980023e-01
5.20535707e-01 -7.84084260e-01 6.27284765e-01 5.26242673e-01
4.65939403e-01 -5.83019614e-01 5.56755543e-01 4.10177112e-01
-8.19843531e-01 -9.16540325e-02 -2.50827193e-01 -2.77557194e-01
-8.00644279e-01 4.60617602e-01 -1.00981116e+00 3.25799853e-01
8.87707531e-01 1.62663534e-01 -1.29856694e+00 1.37603974e+00
-7.18769372e-01 8.81619036e-01 -1.23707280e-01 -5.48955083e-01
1.06380200e-02 4.41464217e-04 4.58303243e-01 9.21856582e-01
5.84436297e-01 5.27265891e-02 4.34781946e-02 6.47134960e-01
-1.66919589e-01 8.09396267e-01 -5.14665723e-01 2.53604203e-01
9.19810534e-01 1.38090169e+00 -8.37616548e-02 -2.73298919e-01
-8.12513888e-01 3.45773160e-01 4.43823934e-01 1.62542462e-01
-5.05771816e-01 -4.39168870e-01 7.10508823e-01 4.40807730e-01
-5.39419577e-02 5.93603663e-02 5.56360185e-02 -1.13402808e+00
7.71410614e-02 -1.55819619e+00 4.59893137e-01 -6.71481371e-01
-1.41646349e+00 3.52247655e-01 -1.09847896e-01 -1.40795875e+00
-4.19664502e-01 -4.74002212e-01 -5.23524046e-01 8.68609011e-01
-1.58042443e+00 -4.93898273e-01 -5.66724300e-01 7.09883451e-01
5.93243003e-01 -3.67313921e-02 7.88360119e-01 6.87983811e-01
-3.16343784e-01 8.08693767e-01 1.33054599e-01 -2.46911258e-01
1.16885710e+00 -1.57482040e+00 -2.70247102e-01 8.06479275e-01
2.63797641e-01 4.92205203e-01 7.63508737e-01 -5.34069538e-01
-1.34058356e+00 -9.31512773e-01 1.41509819e+00 -8.75952601e-01
6.72448456e-01 -1.66423008e-01 -1.41689968e+00 2.38023743e-01
3.55315000e-01 -1.80771992e-01 1.11449063e+00 5.07535897e-02
-2.85283118e-01 -9.70455259e-02 -1.23722136e+00 5.88343859e-01
7.82366812e-01 -3.70456457e-01 -1.03364372e+00 4.37777102e-01
6.90777481e-01 -4.51055765e-01 -1.06946242e+00 2.33503580e-02
3.04067940e-01 -1.17000496e+00 7.01450050e-01 -5.63597083e-01
3.45095307e-01 -2.18000382e-01 -2.10398678e-02 -1.38043118e+00
-1.17412172e-01 -4.75639164e-01 -8.91337544e-02 1.62723446e+00
2.73233831e-01 -5.66128433e-01 4.23596919e-01 6.99760020e-01
-3.09377700e-01 -6.94084167e-01 -9.48223472e-01 -7.63056159e-01
4.42281276e-01 -4.84447479e-01 9.29596126e-01 1.23069704e+00
-1.24203656e-02 2.37488061e-01 3.72739673e-01 -6.09012581e-02
2.62388110e-01 -1.40710086e-01 8.44766915e-01 -1.38775122e+00
-7.03374902e-03 -6.05354905e-01 -3.11828971e-01 -9.86265600e-01
-2.11218715e-01 -8.98229122e-01 5.84839098e-03 -1.30834067e+00
-1.46054491e-01 -3.95129830e-01 -1.26086086e-01 2.43462130e-01
-6.34176433e-01 1.74031913e-01 5.22785373e-02 -2.41604581e-01
-7.68230975e-01 2.72037029e-01 1.74550807e+00 1.26519129e-01
-7.87502900e-02 9.74446982e-02 -6.90273166e-01 7.43285418e-01
5.06380975e-01 -5.91376185e-01 -6.02847934e-01 -4.68206644e-01
6.81940913e-01 -7.94248283e-02 -6.86229393e-02 -1.13040829e+00
3.55610698e-01 -4.15006369e-01 1.14773557e-01 -4.35899258e-01
-1.55492917e-01 -8.57711554e-01 -4.42241549e-01 1.81246474e-01
-3.07452112e-01 2.37292528e-01 1.73398390e-01 1.04332920e-02
-4.63245898e-01 -7.98484445e-01 4.21631098e-01 -5.84295802e-02
-5.12204349e-01 -1.05171055e-01 -5.50835788e-01 7.19916701e-01
9.06356990e-01 -8.39816481e-02 -5.36370575e-01 -6.20147824e-01
-4.21959907e-01 3.00602317e-01 9.18167531e-02 7.35812485e-01
6.65521741e-01 -1.32108867e+00 -7.47029662e-01 2.92445034e-01
4.88998920e-01 -5.43177351e-02 1.70606375e-01 4.70524579e-01
-4.01250899e-01 4.77908462e-01 -5.53364344e-02 -5.29688895e-01
-1.47580564e+00 1.67000905e-01 2.09487319e-01 -2.53401726e-01
-3.06554232e-02 9.78738666e-01 -2.68876135e-01 -7.58770466e-01
3.42370123e-01 -6.24269605e-01 -7.67075419e-01 3.49224627e-01
7.82348812e-01 5.64569414e-01 5.68224847e-01 -4.05811638e-01
1.32142574e-01 4.00375277e-01 -5.86320162e-02 8.51769745e-02
8.70715141e-01 -2.33491972e-01 1.97284296e-02 4.55995351e-01
6.93779945e-01 6.38154685e-01 -4.47196722e-01 -4.96004671e-01
2.51323402e-01 -7.59785533e-01 -3.15453559e-01 -1.26314247e+00
-6.93970740e-01 8.78386557e-01 6.81353569e-01 5.66715777e-01
1.09512877e+00 -7.24855512e-02 3.53306860e-01 3.96382183e-01
2.67166823e-01 -1.13933587e+00 4.97683555e-01 7.85895348e-01
8.20929646e-01 -1.30327749e+00 -2.31869459e-01 -5.04137158e-01
-1.24487765e-01 1.11679745e+00 1.26257920e+00 4.69836950e-01
5.11369467e-01 -2.33826771e-01 3.81846040e-01 -1.44718885e-01
-1.03023005e+00 -4.17224109e-01 7.31983304e-01 5.14673471e-01
1.00014174e+00 3.40175144e-02 -7.16976166e-01 8.66469979e-01
-5.20559430e-01 -1.48947701e-01 8.80531251e-01 1.10012984e+00
-7.93590069e-01 -1.34290826e+00 -6.33028209e-01 5.60058892e-01
-6.48812652e-01 -7.83382058e-02 -5.13216555e-01 6.07393324e-01
3.56948793e-01 1.50507700e+00 -2.41773307e-01 -2.91774631e-01
7.48673797e-01 5.08044839e-01 6.36058807e-01 -1.04360008e+00
-1.23077404e+00 -5.44774950e-01 5.63730262e-02 -3.18530589e-01
-2.06204727e-02 -4.56072390e-01 -7.65847445e-01 -1.81700796e-01
-6.95366085e-01 3.09534431e-01 7.93134034e-01 1.04558337e+00
2.41193250e-01 6.40938222e-01 6.20151162e-01 2.65251637e-01
-9.08247232e-01 -1.31489074e+00 -1.20285057e-01 7.03455150e-01
1.62161961e-01 -7.54199088e-01 -4.37321275e-01 -2.29627728e-01] | [11.257516860961914, 8.146797180175781] |
f4974305-ca63-4179-9822-85ff0c1c924f | learning-on-gradients-generalized-artifacts | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tan_Learning_on_Gradients_Generalized_Artifacts_Representation_for_GAN-Generated_Images_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tan_Learning_on_Gradients_Generalized_Artifacts_Representation_for_GAN-Generated_Images_Detection_CVPR_2023_paper.pdf | Learning on Gradients: Generalized Artifacts Representation for GAN-Generated Images Detection | Recently, there has been a significant advancement in image generation technology, known as GAN. It can easily generate realistic fake images, leading to an increased risk of abuse. However, most image detectors suffer from sharp performance drops in unseen domains. The key of fake image detection is to develop a generalized representation to describe the artifacts produced by generation models. In this work, we introduce a novel detection framework, named Learning on Gradients (LGrad), designed for identifying GAN-generated images, with the aim of constructing a generalized detector with cross-model and cross-data. Specifically, a pretrained CNN model is employed as a transformation model to convert images into gradients. Subsequently, we leverage these gradients to present the generalized artifacts, which are fed into the classifier to ascertain the authenticity of the images. In our framework, we turn the data-dependent problem into a transformation-model-dependent problem. To the best of our knowledge, this is the first study to utilize gradients as the representation of artifacts in GAN-generated images. Extensive experiments demonstrate the effectiveness and robustness of gradients as generalized artifact representations. Our detector achieves a new state-of-the-art performance with a remarkable gain of 11.4%. The code is released at https://github.com/chuangchuangtan/LGrad. | ['Yunchao Wei', 'Guanghua Gu', 'Shikui Wei', 'Yao Zhao', 'Chuangchuang Tan'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['fake-image-detection'] | ['computer-vision'] | [ 3.08068991e-01 -2.90637732e-01 1.07054478e-02 5.42951785e-02
-9.23577726e-01 -4.26308304e-01 5.83300292e-01 -2.83176512e-01
6.69731721e-02 5.45566082e-01 -4.89085279e-02 -1.92790076e-01
4.63663042e-01 -7.50375748e-01 -7.83482075e-01 -7.47673571e-01
4.19188261e-01 -3.70120674e-01 1.09625593e-01 -9.28537995e-02
3.73791426e-01 3.28982294e-01 -9.81610477e-01 2.95047432e-01
9.40206647e-01 1.07184672e+00 -4.30534109e-02 3.80138904e-01
3.37330818e-01 8.65483463e-01 -8.79168689e-01 -6.15714610e-01
3.67599308e-01 -8.50807548e-01 -3.51594090e-01 3.07256430e-01
3.46137375e-01 -7.46973395e-01 -6.28434539e-01 1.31256104e+00
4.01251197e-01 -3.98526102e-01 5.10090530e-01 -1.54938591e+00
-1.09061766e+00 3.99692476e-01 -8.39725733e-01 1.42388612e-01
1.69093624e-01 4.41710949e-01 5.79432547e-01 -1.17114460e+00
4.56867903e-01 1.02069950e+00 4.51673269e-01 6.83331966e-01
-9.46158886e-01 -9.84684229e-01 -1.79145291e-01 2.33806625e-01
-1.31180847e+00 -2.45935887e-01 1.18982470e+00 -3.98017287e-01
3.62186998e-01 1.55986995e-01 5.56091845e-01 1.50410843e+00
5.29534705e-02 1.04744494e+00 1.22411549e+00 -4.19288337e-01
-7.29958294e-03 2.44373381e-01 -2.56188661e-01 8.82204592e-01
5.21973193e-01 1.80307806e-01 -5.25763571e-01 -1.19864531e-01
9.56764579e-01 1.85904920e-01 -5.79762101e-01 -9.77984592e-02
-7.80292690e-01 9.16602373e-01 6.89380527e-01 3.58710051e-01
-2.50100434e-01 3.73680830e-01 2.22245544e-01 7.30858594e-02
4.54135537e-01 4.59250569e-01 2.36719802e-01 1.07537113e-01
-7.93342650e-01 8.18675533e-02 3.28186929e-01 9.83051956e-01
6.34467959e-01 3.75692934e-01 -2.02224791e-01 6.58737063e-01
5.83900623e-02 4.50495094e-01 6.45335674e-01 -4.37485367e-01
6.04112089e-01 5.93134761e-01 1.84446760e-02 -1.42960131e+00
1.62898555e-01 -6.61074817e-01 -9.73187923e-01 1.11407146e-01
1.74791694e-01 -9.30717797e-05 -9.50703800e-01 1.51738691e+00
2.01942503e-01 4.70069200e-01 -5.85203655e-02 1.00844955e+00
4.64932024e-01 5.91598332e-01 -2.02690125e-01 6.98243007e-02
1.27309453e+00 -1.10913646e+00 -6.36170566e-01 -2.73883015e-01
5.97050369e-01 -8.73083770e-01 1.03633606e+00 5.49475133e-01
-8.09048176e-01 -5.11380494e-01 -1.34468257e+00 4.73950803e-02
-2.33147547e-01 4.22018439e-01 3.82813215e-01 8.38023126e-01
-7.00242341e-01 3.52658987e-01 -6.22896075e-01 -1.44650117e-01
7.61184275e-01 -1.60579503e-01 -1.16786957e-01 -1.68957546e-01
-1.11402583e+00 7.75404930e-01 2.99429506e-01 1.51602283e-01
-1.21180594e+00 -4.99794781e-01 -6.84676647e-01 -2.36412942e-01
3.49044353e-01 -4.09008592e-01 1.02277052e+00 -1.06651056e+00
-1.23779166e+00 8.93642604e-01 1.92469954e-01 -3.88895363e-01
1.09841084e+00 -1.91494465e-01 -7.23626494e-01 1.57216743e-01
2.75180578e-01 1.78494185e-01 1.30900002e+00 -1.59938884e+00
-4.77257550e-01 -3.23156863e-01 -2.34765142e-01 -2.77851045e-01
-6.21963084e-01 -8.50312971e-03 -6.06344938e-01 -1.12936425e+00
1.56790428e-02 -9.16575849e-01 7.97934979e-02 -1.30278200e-01
-7.48898089e-01 1.92794353e-01 1.07869232e+00 -8.92736673e-01
1.30205512e+00 -2.29958010e+00 -2.31865674e-01 7.78915584e-02
3.70474666e-01 4.82106775e-01 -1.22702077e-01 4.35706407e-01
-3.43461893e-02 3.48003775e-01 -4.99542296e-01 -3.62879187e-01
-2.91239768e-01 -2.57520109e-01 -6.62514329e-01 5.93576849e-01
6.63430035e-01 1.01078236e+00 -1.07373452e+00 -3.53150696e-01
1.90200001e-01 3.53962779e-01 -1.97659612e-01 2.88862944e-01
1.34663045e-01 4.80052501e-01 -6.20904446e-01 8.16058993e-01
1.05021846e+00 -2.08987683e-01 -1.03365839e-01 -2.49768823e-01
1.49171382e-01 8.22327938e-03 -7.77084529e-01 1.43371499e+00
-2.35763162e-01 6.25419617e-01 -3.48428458e-01 -7.57628500e-01
1.08973312e+00 8.75011459e-02 1.40440077e-01 -6.15663767e-01
3.02303582e-01 6.12697721e-01 -2.99138337e-01 -5.05800247e-01
4.42490637e-01 1.03930302e-01 -1.20955497e-01 2.69188046e-01
-3.33689377e-02 -1.64934143e-01 -7.93912262e-02 3.16402048e-01
1.12119758e+00 1.37290852e-02 1.50738508e-01 1.98858321e-01
4.13882107e-01 -7.04929754e-02 4.75424200e-01 7.40220606e-01
-1.92948461e-01 9.27530169e-01 3.80014688e-01 -4.32011098e-01
-1.09993827e+00 -9.34426785e-01 5.96661270e-02 2.49629408e-01
4.77777719e-01 -2.98686475e-01 -9.45720434e-01 -9.55104172e-01
2.96104308e-02 6.30321980e-01 -5.66516221e-01 -5.41033328e-01
-6.96708560e-01 -8.10802996e-01 9.41725791e-01 4.31491703e-01
1.04683292e+00 -1.00443399e+00 -3.64834666e-01 6.50450513e-02
-3.13866109e-01 -1.25521648e+00 -7.19020844e-01 -3.10801953e-01
-6.76318884e-01 -1.12955570e+00 -7.74783015e-01 -7.54575610e-01
9.02231216e-01 5.23818493e-01 6.29513681e-01 5.15879810e-01
-4.56469595e-01 -1.32897943e-01 -3.97711694e-01 -2.66766995e-01
-5.50694466e-01 -2.69884944e-01 -8.61999765e-02 4.64634329e-01
6.60336167e-02 -4.52300787e-01 -8.38170469e-01 2.99106836e-01
-1.21802700e+00 1.18489459e-01 9.28713024e-01 1.00973463e+00
4.22560930e-01 4.32000756e-02 4.26348805e-01 -9.28155601e-01
8.12192619e-01 -4.96458501e-01 -5.61291933e-01 1.65666655e-01
-5.87275982e-01 -9.68865752e-02 8.78924072e-01 -5.11027813e-01
-9.76305544e-01 -6.30222959e-03 1.02771148e-01 -7.77835190e-01
-1.69135979e-03 2.85902023e-01 -2.73113012e-01 -3.03690314e-01
7.33461857e-01 6.08525336e-01 -3.56490582e-01 -3.74215364e-01
4.36050832e-01 8.80955219e-01 7.27357328e-01 -4.13756758e-01
1.25064385e+00 5.42768002e-01 -8.71861428e-02 -5.46927750e-01
-7.29371011e-01 -2.89334357e-01 -1.54330820e-01 -2.39934862e-01
5.43658972e-01 -9.13226128e-01 -1.57908723e-01 1.05931568e+00
-1.23142660e+00 -1.16094649e-01 1.01758875e-01 1.60079032e-01
-3.05327475e-01 7.60974348e-01 -7.78057694e-01 -8.03992987e-01
-4.34780747e-01 -1.15336728e+00 9.40457880e-01 2.07089871e-01
2.25821614e-01 -5.17317891e-01 -2.71362990e-01 5.65050662e-01
3.62887561e-01 6.08727574e-01 5.72480440e-01 -3.27259392e-01
-7.14166522e-01 -4.48658228e-01 -4.67135221e-01 7.62710333e-01
3.32674652e-01 -1.61099181e-01 -9.97153997e-01 -4.17169660e-01
4.00050879e-01 -2.90799409e-01 1.01877391e+00 -1.44146577e-01
1.42071331e+00 -4.37353462e-01 -2.72172689e-01 6.96886539e-01
1.55989170e+00 1.93156913e-01 9.99770582e-01 4.22290593e-01
9.54307318e-01 1.54531568e-01 5.43971539e-01 4.83650982e-01
-8.17986578e-02 7.01229513e-01 5.20594656e-01 -2.62846708e-01
-4.51644629e-01 -7.33765543e-01 4.31638151e-01 6.18828833e-01
9.27250236e-02 -5.09886026e-01 -6.87905252e-01 5.36108255e-01
-1.73991871e+00 -8.53495121e-01 -3.59954208e-01 2.10161543e+00
6.04528546e-01 1.89859584e-01 -9.55174416e-02 1.12537608e-01
1.02144909e+00 2.27940589e-01 -5.21900415e-01 -5.60417026e-02
-8.44810009e-02 1.17734417e-01 5.98683834e-01 3.08466516e-02
-1.19421196e+00 1.08875191e+00 4.89913797e+00 1.18686354e+00
-1.32056916e+00 2.00199559e-01 7.61647165e-01 3.12100649e-01
-1.17076539e-01 6.48553222e-02 -4.65148330e-01 9.63218689e-01
3.59189987e-01 -8.75898264e-03 3.34644228e-01 8.81213367e-01
2.32520148e-01 2.06791461e-01 -7.61878312e-01 1.21826220e+00
5.27329922e-01 -1.14918780e+00 2.84823596e-01 9.55953598e-02
8.85699689e-01 -4.37278330e-01 3.97901118e-01 -5.47900610e-02
8.41220766e-02 -7.97577858e-01 8.42038572e-01 1.93170577e-01
9.28749919e-01 -6.70080662e-01 6.76764250e-01 1.41588658e-01
-8.54688287e-01 1.51902819e-02 -3.17974031e-01 2.99513578e-01
1.39087841e-01 7.71851718e-01 -8.38116884e-01 6.25177741e-01
4.36210841e-01 6.84622407e-01 -6.19054556e-01 1.03831911e+00
-8.19191754e-01 7.59521961e-01 6.45477176e-02 1.94802970e-01
2.27485925e-01 -1.37017563e-01 4.49856997e-01 1.09533787e+00
6.34039581e-01 -1.90719306e-01 4.35410589e-02 1.19701910e+00
-5.77291012e-01 -8.51546749e-02 -6.37035191e-01 -9.43202302e-02
4.37361389e-01 1.27561569e+00 -5.72447181e-01 -2.66546369e-01
-2.23251536e-01 1.45713699e+00 2.37344012e-01 2.80052185e-01
-1.16601610e+00 -4.21275258e-01 3.24221283e-01 3.20263319e-02
2.47488022e-01 -2.45413706e-02 -2.53695577e-01 -1.52760422e+00
4.61312056e-01 -1.11662996e+00 2.14075312e-01 -7.06522763e-01
-1.45366335e+00 5.63709915e-01 -3.42806071e-01 -1.55146170e+00
-2.44872533e-02 -4.95602787e-01 -7.29307413e-01 7.72552490e-01
-1.45464683e+00 -1.41098237e+00 -7.97727466e-01 5.11251390e-01
4.78022307e-01 -1.06389694e-01 4.89367306e-01 2.39105314e-01
-7.64253259e-01 8.00317228e-01 1.45571470e-01 5.69955170e-01
6.47249579e-01 -9.12632644e-01 5.45995474e-01 1.43755245e+00
1.60782620e-01 3.70935947e-01 4.51242745e-01 -8.14437270e-01
-1.38885093e+00 -1.34853184e+00 3.37006569e-01 -2.83340693e-01
6.70192838e-01 -4.31723267e-01 -9.20877397e-01 5.54436684e-01
-2.69163190e-03 2.21147597e-01 2.55363792e-01 -7.56283104e-01
-7.17292190e-01 -6.92700744e-02 -1.26270401e+00 5.01131654e-01
1.03528094e+00 -6.61946177e-01 -2.14168236e-01 3.18479538e-01
5.36127388e-01 -4.17344987e-01 -3.87255222e-01 2.21567467e-01
3.69123727e-01 -8.95389974e-01 7.95039535e-01 -1.88304812e-01
7.57907748e-01 -5.17652035e-01 1.57068819e-01 -1.37868190e+00
-2.77262449e-01 -5.90410411e-01 -2.03032672e-01 1.31911349e+00
2.23533317e-01 -7.69579768e-01 7.48333216e-01 1.73878118e-01
-1.47538081e-01 -6.81622922e-01 -6.90809786e-01 -1.04857337e+00
-1.11997150e-01 -3.29379320e-01 6.58600092e-01 1.02992713e+00
-2.00484037e-01 7.41700456e-03 -9.48970914e-01 3.55731606e-01
8.16719174e-01 1.15015894e-01 8.74990404e-01 -4.80361819e-01
-3.84961724e-01 -3.68129730e-01 -6.65722370e-01 -1.10319829e+00
-2.23948911e-01 -7.20739663e-01 6.23464771e-03 -1.11649632e+00
2.91284114e-01 -3.53213042e-01 -3.75152975e-01 2.82740206e-01
-4.59190875e-01 6.73124194e-01 2.92348027e-01 4.50938046e-01
-2.15803206e-01 5.62897086e-01 1.22811615e+00 -3.00459087e-01
8.80210325e-02 -2.75626183e-01 -8.66907954e-01 4.88426656e-01
1.22287369e+00 -6.50840342e-01 -1.78687751e-01 -5.67929327e-01
-2.00625837e-01 -2.96723187e-01 6.62270546e-01 -1.16146266e+00
1.44499671e-02 -3.46573927e-02 5.03546417e-01 -3.23642254e-01
2.05838934e-01 -5.02738178e-01 2.09671631e-01 6.80462062e-01
-9.34054926e-02 -6.62011653e-02 -2.17030793e-02 6.52224123e-01
-3.81939381e-01 -2.29863882e-01 8.40252221e-01 -1.73250109e-01
-6.14272535e-01 3.02950978e-01 1.25251766e-02 -7.11484998e-02
1.11727166e+00 -8.45281631e-02 -6.15565479e-01 -4.44406569e-01
-1.12291314e-01 -9.37713310e-02 6.42723739e-01 5.91115952e-01
9.74813819e-01 -1.44829869e+00 -8.03566337e-01 2.36412361e-01
3.64351660e-01 -2.47653097e-01 2.67314583e-01 6.30772293e-01
-6.36481702e-01 2.54806876e-02 -1.09214224e-01 -4.08720285e-01
-1.10572004e+00 7.15423286e-01 2.62359113e-01 -2.40065858e-01
-5.49229205e-01 6.76932633e-01 3.66019934e-01 2.40108967e-01
-2.68364310e-01 -9.37790647e-02 3.21413934e-01 -3.00963908e-01
5.09321094e-01 2.93576568e-01 1.03836693e-01 -7.02016115e-01
-2.34742120e-01 2.56843626e-01 -1.52267575e-01 1.22252144e-01
1.12721562e+00 1.74587891e-01 -4.04272228e-02 -1.05570748e-01
1.31688714e+00 1.12013280e-01 -1.17056394e+00 -1.12600200e-01
-8.28076676e-02 -1.06910300e+00 5.89917414e-02 -7.49362409e-01
-1.42629945e+00 8.36961806e-01 6.60276890e-01 2.01437891e-01
1.27428651e+00 -7.21815452e-02 1.15915990e+00 -1.81414858e-01
4.74922568e-01 -7.67227650e-01 6.23613894e-01 -9.41700209e-03
1.07637000e+00 -1.27650988e+00 -1.20385565e-01 -6.31679177e-01
-5.15626669e-01 9.93860781e-01 8.20160806e-01 -2.59814113e-01
4.61419225e-02 -2.04960816e-02 6.62786141e-02 -1.23808637e-01
-1.82771794e-02 2.73713805e-02 1.93351656e-02 5.78029156e-01
1.38004019e-03 2.76168942e-01 -3.60769957e-01 3.92242461e-01
-1.72848672e-01 1.05214037e-01 7.20509827e-01 8.73583853e-01
-6.73077255e-02 -1.23379850e+00 -4.92531508e-01 3.37043852e-01
-4.27646607e-01 -1.73992753e-01 -5.85555077e-01 6.88978016e-01
2.47957483e-01 1.11610138e+00 -3.92888963e-01 -7.10567832e-01
2.59546071e-01 -2.61707425e-01 4.11219925e-01 -3.77020329e-01
-4.71421927e-01 -7.95960054e-02 -1.98562503e-01 -4.48327839e-01
-9.39760655e-02 -3.66711855e-01 -8.86276722e-01 -2.22998932e-01
-6.26352429e-01 -6.76500052e-02 5.97577572e-01 5.50100803e-01
4.68166232e-01 2.39250794e-01 1.00646770e+00 -5.89014292e-01
-6.53534472e-01 -9.23406482e-01 -4.79907870e-01 9.73712444e-01
3.33130926e-01 -5.28580844e-01 -5.95858574e-01 -6.15810789e-02] | [12.434298515319824, 1.0358116626739502] |
d42baeea-1c0e-41c5-a115-d7cc9dd70f83 | aspect-is-not-you-need-no-aspect-differential | null | null | https://aclanthology.org/2022.naacl-main.115 | https://aclanthology.org/2022.naacl-main.115.pdf | Aspect Is Not You Need: No-aspect Differential Sentiment Framework for Aspect-based Sentiment Analysis | Aspect-based sentiment analysis (ABSA) is a fine-grained sentiment classification task. Most recent efforts adopt pre-trained model to classify the sentences with aspects. However, the aspect sentiment bias from pre-trained model brings some noise to the ABSA task. Besides, traditional methods using cross-entropy loss are hard to find the potential associations between sentiment polarities. In this work, we analyze the ABSA task from a novel cognition perspective: humans can often judge the sentiment of an aspect even if they do not know what the aspect is. Moreover, it is easier to distinguish positive and negative sentiments than others for human beings because positive and negative are two opposite sentiments. To this end, we propose a no-aspect differential sentiment (NADS) framework for the ABSA task. We first design a no-aspect template by replacing the aspect with a special unbiased character to eliminate the sentiment bias and obtain a stronger representation. To better get the benefits from the template, we adopt contrastive learning between the no-aspect template and the original sentence. Then we propose a differential sentiment loss instead of the cross-entropy loss to better classify the sentiments by distinguishing the different distances between sentiments. Our proposed model is a general framework and can be combined with almost all traditional ABSA methods. Experiments on SemEval 2014 show that our framework is still able to predict the sentiment of the aspect even we don’t konw what the aspect is. Moreover, our NADS framework boosts three typical ABSA methods and achieves state-of-the-art performance. | ['Xu Bai', 'Lei Jiang', 'Huailiang Peng', 'Rui Liu', 'Jiahao Cao'] | null | null | null | null | naacl-2022-7 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 1.46365939e-02 -1.63150147e-01 -3.88273560e-02 -8.59294534e-01
-4.24100488e-01 -6.84580147e-01 6.88655674e-01 2.79806674e-01
-4.75215435e-01 4.37771976e-01 2.18848661e-01 -2.06153825e-01
9.39516276e-02 -1.08927095e+00 -5.32087088e-01 -7.49253750e-01
6.26071274e-01 1.36875153e-01 1.00374028e-01 -7.99676597e-01
3.75825971e-01 -1.22614600e-01 -1.46345806e+00 5.44014335e-01
9.13136780e-01 1.33527684e+00 6.20632386e-03 1.02989815e-01
-4.54627603e-01 5.67838252e-01 -7.34788954e-01 -8.67049038e-01
1.69723362e-01 -5.54470658e-01 -7.85999477e-01 -6.80803508e-02
2.26315647e-01 3.33460748e-01 4.32999194e-01 1.25754392e+00
4.02112812e-01 -1.02015160e-01 7.74861693e-01 -1.15049398e+00
-8.39181721e-01 4.88346964e-01 -7.13853657e-01 -9.86147746e-02
4.80178922e-01 -8.17998126e-02 1.48401618e+00 -8.49306703e-01
2.51651466e-01 1.30671155e+00 8.74327064e-01 4.00817573e-01
-7.82350659e-01 -6.41302764e-01 6.78083122e-01 2.45640129e-01
-8.33295643e-01 1.09200954e-01 1.03642499e+00 -2.77597368e-01
5.25105119e-01 4.71834362e-01 9.42755520e-01 1.03182590e+00
3.79740834e-01 1.02621269e+00 1.70812476e+00 -1.80437043e-01
2.72592068e-01 5.44043005e-01 6.22206569e-01 2.56172419e-01
2.56118298e-01 -3.90480191e-01 -5.31269372e-01 4.67037261e-02
-2.98296034e-01 1.19957946e-01 -2.41191655e-01 -2.66771287e-01
-1.15242934e+00 7.54204750e-01 3.98588717e-01 3.11229676e-01
-3.17077219e-01 -5.02807438e-01 5.00337899e-01 5.08518517e-01
5.40603459e-01 7.05581248e-01 -8.91183257e-01 7.22858533e-02
-6.97734654e-01 3.23439479e-01 8.08375180e-01 5.64692140e-01
1.01801503e+00 -2.20478684e-01 -3.18466961e-01 7.45624840e-01
3.24912876e-01 5.77221394e-01 7.79995739e-01 -2.47918054e-01
2.01490015e-01 1.15026307e+00 -1.57927006e-01 -1.36762738e+00
-4.26977545e-01 -5.41317821e-01 -7.86064863e-01 3.05681646e-01
2.06032798e-01 -9.24733728e-02 -6.69773400e-01 1.67525446e+00
3.41364980e-01 -3.54495853e-01 2.13856190e-01 8.13615501e-01
8.08797956e-01 6.36163592e-01 -2.65536644e-02 6.40639197e-03
1.72790563e+00 -1.00770783e+00 -7.19051838e-01 -5.83684266e-01
5.22623420e-01 -7.80088723e-01 1.49595356e+00 3.58763218e-01
-7.05003381e-01 -4.12659466e-01 -1.34492934e+00 -5.06264158e-02
-8.86452317e-01 5.56011088e-02 8.21527839e-01 7.68216372e-01
-7.04805315e-01 4.68771040e-01 -5.04810870e-01 -7.72218481e-02
2.91983694e-01 2.02698514e-01 -3.28148991e-01 1.38968334e-01
-1.42831254e+00 8.74707580e-01 2.52759635e-01 1.75653875e-01
-1.49423465e-01 -5.82571983e-01 -1.04620481e+00 -2.16414183e-02
3.52113187e-01 -7.17139959e-01 1.05385137e+00 -1.64407122e+00
-1.36728358e+00 8.33278656e-01 -3.11032355e-01 -1.01069666e-01
2.11096644e-01 -2.07427412e-01 -4.83594894e-01 -2.54386842e-01
2.40699381e-01 1.46213472e-01 8.67539585e-01 -1.12507272e+00
-6.12013340e-01 -6.04593039e-01 4.86838251e-01 3.13148052e-01
-7.15836167e-01 -6.95850104e-02 -1.30664393e-01 -6.95729434e-01
7.79861435e-02 -9.06863809e-01 -9.40944105e-02 -1.72600955e-01
-5.19770503e-01 -2.15015739e-01 8.69362056e-01 -1.75792485e-01
1.15838635e+00 -2.00950432e+00 -2.05413342e-01 1.42351598e-01
3.69824059e-02 2.97694892e-01 -1.68295428e-01 1.26380414e-01
-6.81881085e-02 7.46615976e-02 -3.69684130e-01 -1.88360900e-01
7.87765235e-02 -7.53581151e-02 -2.36329556e-01 1.49610966e-01
3.35759968e-01 7.41231084e-01 -8.02435279e-01 -2.48080045e-01
-1.32273272e-01 4.66901124e-01 -5.98530412e-01 9.29814130e-02
-1.08971849e-01 2.59574950e-01 -6.20378971e-01 6.08370721e-01
8.57965171e-01 -1.44259453e-01 4.28543463e-02 -4.89183426e-01
-8.51777643e-02 5.49452007e-01 -9.43302512e-01 1.19413590e+00
-7.13010967e-01 4.46837246e-01 -1.62933096e-01 -1.04297900e+00
1.17002988e+00 2.90926155e-02 5.82077876e-02 -7.26131558e-01
4.27326053e-01 2.23316476e-01 9.10554454e-02 -2.69547403e-01
6.82886302e-01 -7.32980251e-01 -3.72603655e-01 4.25148308e-01
-1.10345960e-01 -3.34338248e-01 2.43573353e-01 -1.17260672e-01
6.23666286e-01 -9.87395737e-03 5.38592815e-01 -4.62360770e-01
9.96839166e-01 -1.08479105e-01 8.29757512e-01 3.27527583e-01
-2.92161047e-01 6.61634445e-01 8.74164701e-01 -5.95988214e-01
-6.60943627e-01 -6.77279592e-01 -1.09936789e-01 1.12052488e+00
4.85542625e-01 -5.58308244e-01 -6.08770251e-01 -1.11837995e+00
-1.97370723e-01 6.90929949e-01 -9.63546574e-01 -2.34671384e-01
-1.24308944e-01 -1.12850285e+00 4.59447578e-02 4.17825401e-01
7.53199100e-01 -1.07766724e+00 -1.35725349e-01 -2.67373472e-01
-2.97172844e-01 -8.57744992e-01 -5.69552481e-01 2.05525070e-01
-4.58774865e-01 -9.14888561e-01 -5.09970546e-01 -8.53777707e-01
6.59110665e-01 2.10846618e-01 1.20045948e+00 -5.16104996e-02
4.85178769e-01 -1.62531018e-01 -7.04094112e-01 -7.48030841e-01
-1.73434705e-01 1.42711163e-01 -5.43687716e-02 4.09896374e-01
8.72218490e-01 -4.04081911e-01 -7.44426370e-01 3.48009020e-01
-9.27335382e-01 -1.44231483e-01 6.72859192e-01 7.51955330e-01
7.14846015e-01 2.99464136e-01 6.66948497e-01 -1.34407747e+00
9.07224655e-01 -4.27767128e-01 -1.12322256e-01 1.14422396e-01
-9.61335540e-01 2.94757746e-02 9.94814217e-01 -3.43358219e-01
-1.06434512e+00 -2.57040173e-01 -3.67921263e-01 1.91225678e-01
9.14169401e-02 6.36658609e-01 -5.46436131e-01 2.20919460e-01
2.92502314e-01 3.11884880e-01 -2.39912361e-01 -1.41697839e-01
1.29974484e-01 8.12063217e-01 2.06511971e-02 -3.63767445e-01
6.30266547e-01 5.74393451e-01 -1.18427597e-01 -4.98338044e-01
-1.56788123e+00 -4.45748985e-01 -3.68176460e-01 -7.82998092e-03
6.99013054e-01 -8.42724502e-01 -7.51681924e-01 6.77058399e-01
-1.01683044e+00 2.36589268e-01 -2.37874389e-01 2.58723557e-01
-2.60787159e-01 3.68521005e-01 -2.84658670e-01 -6.47279084e-01
-6.15856528e-01 -1.27058697e+00 1.00969386e+00 3.82159173e-01
-3.91146958e-01 -1.03806663e+00 1.43999577e-01 4.13649499e-01
5.43516517e-01 4.77677695e-02 9.79133248e-01 -9.12830353e-01
5.28739579e-02 -3.56805623e-01 -1.80020686e-02 8.01741600e-01
3.22456270e-01 -1.00556180e-01 -1.12674463e+00 -1.08185828e-01
4.89252388e-01 -2.56152093e-01 9.47808862e-01 1.54150994e-02
9.52818513e-01 -4.67383295e-01 1.95146471e-01 4.77436215e-01
1.31178153e+00 9.69300047e-02 6.58152401e-01 8.80731404e-01
6.14849210e-01 7.65880287e-01 9.27726924e-01 1.89968407e-01
6.64160132e-01 3.57016891e-01 3.22890460e-01 1.78372301e-02
2.59648561e-01 -2.51962095e-01 7.34913051e-01 1.22909999e+00
1.42640918e-01 -1.50771737e-01 -6.03973687e-01 5.79196274e-01
-1.55614567e+00 -8.18144977e-01 -2.36791655e-01 1.93004584e+00
9.98249531e-01 4.33569580e-01 -7.02390820e-02 4.35511678e-01
5.78801751e-01 5.99986076e-01 -5.29881418e-01 -8.08132768e-01
-5.21419823e-01 5.36953807e-02 1.02647450e-02 2.80897558e-01
-1.41174603e+00 8.14434409e-01 5.20745754e+00 9.34775949e-01
-1.27052855e+00 -2.36098617e-02 8.15144539e-01 1.27310619e-01
-8.14135551e-01 4.83806469e-02 -7.19443440e-01 5.02561927e-01
5.36130250e-01 -2.99422801e-01 -7.18369111e-02 1.11279261e+00
-2.22517774e-02 -3.98323312e-02 -8.78046870e-01 6.81847215e-01
3.36073935e-01 -7.61037171e-01 4.17901933e-01 -2.99082428e-01
7.60259509e-01 -4.15765285e-01 2.21286103e-01 5.57430625e-01
-1.57628894e-01 -7.75377631e-01 7.69946218e-01 4.18210179e-01
2.20759243e-01 -8.45239997e-01 1.25230753e+00 2.39487678e-01
-1.16076589e+00 1.31983444e-01 -3.75712484e-01 -3.72284532e-01
-1.40442848e-01 1.01420736e+00 -1.32596284e-01 6.49257958e-01
7.50741601e-01 1.00914800e+00 -8.16831112e-01 6.16800189e-01
-4.85456854e-01 4.66318697e-01 1.37618452e-01 -4.25310671e-01
2.76865125e-01 -4.47397381e-01 5.37565351e-01 1.04654717e+00
2.09810928e-01 -2.62397647e-01 -1.16809152e-01 5.77709556e-01
-1.87722519e-01 5.25020063e-01 -6.19231343e-01 9.01601538e-02
-5.19017987e-02 1.49334419e+00 -6.29332185e-01 -3.29985917e-01
-5.92658341e-01 9.47410941e-01 2.19323665e-01 8.06291625e-02
-4.19276536e-01 -7.29798555e-01 7.31610119e-01 -1.41322374e-01
3.62262040e-01 4.22956079e-01 -7.52354443e-01 -1.36039007e+00
3.80475610e-01 -1.12208056e+00 1.82269365e-01 -7.48976052e-01
-1.67129862e+00 1.07232594e+00 -4.57581133e-01 -1.48794949e+00
1.00946322e-01 -8.10663223e-01 -8.89819324e-01 7.73336411e-01
-1.88199389e+00 -1.14238095e+00 -1.14507221e-01 5.12071609e-01
3.90317202e-01 -3.67100094e-03 6.70457006e-01 9.96314958e-02
-3.44365597e-01 6.25173450e-01 -1.03526831e-01 7.91058466e-02
8.61517906e-01 -1.50336170e+00 2.49699712e-01 7.32339263e-01
-2.78338343e-01 8.48518014e-01 7.73708344e-01 -4.30897236e-01
-1.09710228e+00 -1.00273454e+00 1.19544303e+00 -5.66134274e-01
7.64398396e-01 -2.53216565e-01 -9.47877288e-01 4.77736294e-01
4.04372782e-01 -2.78361022e-01 8.90921950e-01 4.82382119e-01
-7.07743108e-01 -4.91348088e-01 -1.04826319e+00 6.69692993e-01
4.15487677e-01 -5.41610897e-01 -9.94092345e-01 -4.80081402e-02
6.86242223e-01 2.17670053e-01 -5.32498538e-01 6.85207248e-01
7.21587181e-01 -1.34734821e+00 6.41968250e-01 -4.37905639e-01
7.89099276e-01 -5.42226315e-01 -2.56821692e-01 -1.62085366e+00
4.36718650e-02 -1.38845220e-01 5.07837117e-01 1.45195806e+00
7.03780174e-01 -8.38719726e-01 6.32556558e-01 4.29550797e-01
2.35328712e-02 -1.10161281e+00 -5.66751659e-01 -6.61583841e-01
3.90138566e-01 -5.58534026e-01 9.30621743e-01 1.00197005e+00
2.69051760e-01 8.29732776e-01 -1.94173813e-01 -7.93914199e-02
2.72007376e-01 6.50490761e-01 5.24867415e-01 -1.16590559e+00
-1.90154105e-01 -6.46812022e-01 -4.81685549e-01 -7.60238826e-01
1.86657384e-01 -9.06757057e-01 6.02295622e-02 -1.41772449e+00
4.56718057e-01 -1.54017180e-01 -5.39368272e-01 4.78041351e-01
-6.62960351e-01 3.63539279e-01 1.71142876e-01 -7.82856569e-02
-6.85384393e-01 9.52750206e-01 1.51866257e+00 -4.32210118e-01
-1.64242238e-02 2.72003174e-01 -1.41341889e+00 1.20537817e+00
9.45133805e-01 -4.30569082e-01 -3.66822600e-01 -8.39740038e-02
9.36477065e-01 -6.42467141e-01 -7.96312094e-02 -7.18985498e-01
-2.08329968e-02 -4.24986184e-02 1.53882951e-01 -5.91935933e-01
2.15374798e-01 -7.87926137e-01 -4.52923357e-01 2.36025691e-01
-3.10142934e-01 1.79593563e-01 -6.62605325e-03 4.60540473e-01
-7.12572932e-01 -1.57879263e-01 7.08125591e-01 -1.20102048e-01
-5.43107688e-01 1.36439294e-01 -2.60601848e-01 2.87544012e-01
7.88998187e-01 4.67628520e-03 -4.21959400e-01 -3.50810200e-01
-3.36685508e-01 1.61482722e-01 5.18399596e-01 4.94012833e-01
4.25009578e-01 -1.35984170e+00 -4.79902267e-01 1.59448177e-01
4.32711989e-01 -2.13788316e-01 3.98671329e-01 9.06360269e-01
-1.50061846e-01 1.78745344e-01 -1.52535170e-01 -1.83251083e-01
-1.18840182e+00 5.60654223e-01 1.76037654e-01 -4.74598467e-01
-7.19738454e-02 5.84447503e-01 5.55799723e-01 -9.71349597e-01
-2.89943129e-01 -2.71201283e-01 -6.98455513e-01 5.32535195e-01
7.21622944e-01 4.99322377e-02 1.92786574e-01 -8.57126713e-01
-5.03083706e-01 8.89884651e-01 -2.96228707e-01 1.62157729e-01
1.24853516e+00 -2.36776978e-01 -3.51446480e-01 7.41171956e-01
1.41115630e+00 3.53514999e-01 -7.61711001e-01 -7.09277764e-02
-1.52538449e-01 -3.45235109e-01 -9.74565595e-02 -9.38328266e-01
-1.21548820e+00 1.06536007e+00 2.51312643e-01 5.41175723e-01
1.35117424e+00 -1.06764957e-01 8.52046788e-01 5.04846871e-01
1.06451593e-01 -1.11111569e+00 1.71626210e-02 7.54418612e-01
7.14791417e-01 -1.47537673e+00 2.03822460e-02 -3.00456941e-01
-1.14720881e+00 9.32588398e-01 5.55269539e-01 -2.15638563e-01
7.93717742e-01 3.91393639e-02 5.34569621e-01 -3.00297618e-01
-6.14112616e-01 -2.74177283e-01 4.50553238e-01 3.56342167e-01
5.60788631e-01 1.50186568e-01 -7.64738083e-01 1.20520031e+00
-7.74392664e-01 -3.95478666e-01 6.28839672e-01 6.49665177e-01
-2.88476527e-01 -1.10812032e+00 -1.09069131e-01 3.96470785e-01
-8.01696599e-01 -2.91895002e-01 -4.87042814e-01 5.21506906e-01
1.14692613e-01 1.11233783e+00 -2.53391743e-01 -7.60327160e-01
5.83710909e-01 3.03190798e-02 -1.12906024e-01 -4.83126938e-01
-8.92385304e-01 -2.14564830e-01 -2.87887491e-02 -4.77407843e-01
-6.84471905e-01 -6.10099614e-01 -1.00140274e+00 -2.13047907e-01
-3.44497144e-01 3.10294241e-01 8.39935184e-01 1.14992607e+00
2.38226801e-01 5.13679385e-01 9.91349101e-01 -3.60342205e-01
-3.03080767e-01 -8.41001689e-01 -7.67181516e-01 6.76899612e-01
4.40804660e-01 -4.59525645e-01 -6.56351447e-01 -2.99999595e-01] | [11.424759864807129, 6.682975769042969] |
a702aeca-6a25-4845-bb1d-229fb9a8d3dd | lifting-transformer-for-3d-human-pose | 2103.14304 | null | https://arxiv.org/abs/2103.14304v8 | https://arxiv.org/pdf/2103.14304v8.pdf | Exploiting Temporal Contexts with Strided Transformer for 3D Human Pose Estimation | Despite the great progress in 3D human pose estimation from videos, it is still an open problem to take full advantage of a redundant 2D pose sequence to learn representative representations for generating one 3D pose. To this end, we propose an improved Transformer-based architecture, called Strided Transformer, which simply and effectively lifts a long sequence of 2D joint locations to a single 3D pose. Specifically, a Vanilla Transformer Encoder (VTE) is adopted to model long-range dependencies of 2D pose sequences. To reduce the redundancy of the sequence, fully-connected layers in the feed-forward network of VTE are replaced with strided convolutions to progressively shrink the sequence length and aggregate information from local contexts. The modified VTE is termed as Strided Transformer Encoder (STE), which is built upon the outputs of VTE. STE not only effectively aggregates long-range information to a single-vector representation in a hierarchical global and local fashion, but also significantly reduces the computation cost. Furthermore, a full-to-single supervision scheme is designed at both full sequence and single target frame scales applied to the outputs of VTE and STE, respectively. This scheme imposes extra temporal smoothness constraints in conjunction with the single target frame supervision and hence helps produce smoother and more accurate 3D poses. The proposed Strided Transformer is evaluated on two challenging benchmark datasets, Human3.6M and HumanEva-I, and achieves state-of-the-art results with fewer parameters. Code and models are available at \url{https://github.com/Vegetebird/StridedTransformer-Pose3D}. | ['Wenming Yang', 'Pichao Wang', 'Mengyuan Liu', 'Runwei Ding', 'Hong Liu', 'Wenhao Li'] | 2021-03-26 | null | null | null | null | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [ 1.01975702e-01 -7.74442554e-02 -1.76385026e-02 -4.60635096e-01
-7.14580834e-01 -1.49854943e-01 4.41497535e-01 -3.68136853e-01
-3.89835209e-01 4.24314201e-01 3.80468965e-01 1.57434329e-01
1.46660864e-01 -5.04257619e-01 -9.08209741e-01 -5.52230060e-01
-1.32966908e-02 2.37815127e-01 1.42235741e-01 -2.44660497e-01
-3.45803261e-01 3.29439819e-01 -1.25162089e+00 1.69373721e-01
6.31142557e-01 1.09068036e+00 4.77871269e-01 5.40905714e-01
2.67709166e-01 5.26641011e-01 -4.37730998e-01 -3.57483119e-01
3.39636266e-01 -3.25975567e-01 -5.74130714e-01 3.44911307e-01
4.48544413e-01 -5.67209899e-01 -6.55099332e-01 8.15957129e-01
7.76935637e-01 2.19578087e-01 3.31465393e-01 -9.56375659e-01
-3.79244775e-01 1.67353272e-01 -7.63437033e-01 1.62496381e-02
5.36640048e-01 3.23686063e-01 8.64872932e-01 -9.77841496e-01
5.87722063e-01 1.41021371e+00 7.23100245e-01 3.94450366e-01
-1.08565462e+00 -7.72167146e-01 3.64475995e-01 4.96937446e-02
-1.48026252e+00 -2.12278649e-01 9.04087663e-01 -3.34263861e-01
9.84103441e-01 6.98529556e-02 8.40160787e-01 1.41756058e+00
3.52621853e-01 9.88668621e-01 7.24057555e-01 -8.58183280e-02
-2.07384914e-01 -4.54655021e-01 -2.64874548e-01 9.06355977e-01
1.73087474e-02 5.57705462e-02 -6.65416121e-01 1.61863282e-01
1.09280670e+00 3.17277372e-01 -4.23821688e-01 -6.33425355e-01
-1.24747205e+00 7.30562806e-01 7.54012346e-01 4.15715352e-02
-5.79510152e-01 2.81282216e-01 5.33185005e-01 7.76464641e-02
4.89277035e-01 -5.10498416e-03 -5.47094703e-01 -8.24178234e-02
-6.88116550e-01 5.01830339e-01 3.37519586e-01 1.00181162e+00
6.90972745e-01 1.80178359e-01 -1.87933952e-01 8.38228703e-01
3.38418096e-01 5.62386215e-01 4.12358493e-01 -8.82262647e-01
6.82725310e-01 5.70131183e-01 4.79659736e-02 -8.94300520e-01
-5.09584308e-01 -8.21936011e-01 -9.15461540e-01 -6.97881216e-03
3.58495452e-02 -9.05275345e-02 -1.17438900e+00 1.86230242e+00
5.96741855e-01 2.16031313e-01 -2.11533323e-01 1.27869034e+00
7.61207104e-01 6.58717752e-01 1.73282623e-02 2.27228418e-01
1.23582888e+00 -1.02060294e+00 -4.28360134e-01 -4.14376140e-01
5.59501767e-01 -5.66273332e-01 9.23732281e-01 1.26005799e-01
-1.00770354e+00 -9.35059130e-01 -1.13795888e+00 -3.32168043e-01
2.01657414e-03 3.51889104e-01 3.21773618e-01 7.09254369e-02
-6.87121868e-01 5.43330908e-01 -1.10380483e+00 -1.89276487e-01
4.16550249e-01 3.86450797e-01 -6.56895399e-01 -1.60717607e-01
-1.31051779e+00 8.57301652e-01 3.19956392e-01 3.79991114e-01
-8.09094191e-01 -5.71648955e-01 -1.39095008e+00 -1.30635291e-01
5.41987240e-01 -9.74629581e-01 1.16544557e+00 -5.21737099e-01
-1.42398846e+00 6.40204489e-01 -1.35332048e-01 -4.50849771e-01
5.93144596e-01 -8.26983809e-01 -3.21395472e-02 4.95358668e-02
1.99348629e-01 8.99584830e-01 9.74072218e-01 -1.05850279e+00
-4.23338413e-01 -4.89774168e-01 -2.96725761e-02 5.22804439e-01
-6.90163150e-02 -2.58028895e-01 -8.17065477e-01 -9.71423805e-01
3.28333713e-02 -1.06146228e+00 -3.16951662e-01 3.82936075e-02
-3.59588385e-01 -1.70370325e-01 6.87404573e-01 -8.86847794e-01
1.25699079e+00 -2.09030461e+00 7.57792175e-01 -7.87311345e-02
2.60366172e-01 2.82317460e-01 -2.40571856e-01 2.68598020e-01
-1.78195372e-01 -4.13905501e-01 -2.55314827e-01 -7.38958538e-01
-9.36853886e-02 4.21469986e-01 -9.94134508e-03 5.23211539e-01
4.82159615e-01 1.08625126e+00 -8.16555142e-01 -3.43459845e-01
5.24121046e-01 8.67683291e-01 -7.41793334e-01 3.16517889e-01
-2.47225482e-02 5.91940224e-01 -5.45631528e-01 4.34548438e-01
5.68739355e-01 -3.37972283e-01 8.77570361e-03 -4.88313884e-01
6.48755878e-02 4.27351952e-01 -1.04919302e+00 2.25462842e+00
-4.21412855e-01 2.18799174e-01 -2.22372096e-02 -1.00439835e+00
1.02368653e+00 3.06466639e-01 5.83963215e-01 -6.43615842e-01
3.66049439e-01 4.09808680e-02 -2.72980481e-01 -3.06876838e-01
3.33065599e-01 1.40905166e-02 -3.71553183e-01 5.84764816e-02
3.05203825e-01 5.31154424e-02 4.26354110e-02 -4.42566015e-02
9.22896028e-01 8.47654939e-01 2.75778055e-01 1.63451713e-02
5.77184498e-01 -3.17767918e-01 8.79207194e-01 1.64114088e-01
-1.43845081e-01 7.57808685e-01 2.02400133e-01 -6.57185555e-01
-1.08008206e+00 -1.14810920e+00 3.72191519e-01 8.49414110e-01
2.18148410e-01 -6.40809298e-01 -5.72705925e-01 -7.19058633e-01
1.18064970e-01 3.09101343e-01 -6.83650136e-01 -3.21882367e-01
-1.00106478e+00 -2.69731760e-01 3.61250192e-01 7.35691011e-01
6.78307533e-01 -8.75307500e-01 -8.80072176e-01 2.42001936e-01
-4.13806647e-01 -1.20960283e+00 -9.62424695e-01 2.34950691e-01
-7.46269405e-01 -8.61209095e-01 -8.99276555e-01 -6.24684930e-01
3.70701522e-01 2.16973960e-01 8.22781086e-01 -2.30206132e-01
-1.93923995e-01 -6.13212511e-02 -4.54021811e-01 -1.58147767e-01
1.80768762e-02 1.53185114e-01 1.14079185e-01 2.01759879e-02
2.12091833e-01 -6.01954281e-01 -6.77002847e-01 3.89953285e-01
-6.68284178e-01 3.34021986e-01 7.85903215e-01 1.07001317e+00
6.93193913e-01 -2.74666280e-01 3.25549424e-01 -4.39954281e-01
2.31861621e-01 -1.78691283e-01 -3.11080784e-01 -2.70675812e-02
-5.16051203e-02 2.04865143e-01 5.75723410e-01 -4.51207727e-01
-9.86329794e-01 4.07011002e-01 -3.76742363e-01 -1.03296137e+00
-3.29880528e-02 3.85479569e-01 -2.43756115e-01 2.06298411e-01
4.21235085e-01 3.22120339e-01 2.36847475e-01 -7.57826149e-01
2.35011086e-01 3.95000100e-01 7.41757333e-01 -3.43832463e-01
9.19337094e-01 2.49298096e-01 -7.48911947e-02 -6.34106457e-01
-1.03767431e+00 -4.03503478e-01 -7.47338474e-01 -2.55732238e-01
9.75947201e-01 -1.25782776e+00 -4.49664891e-01 6.75206363e-01
-1.10077167e+00 -4.13361520e-01 -1.60024419e-01 5.74000180e-01
-6.57904267e-01 3.37492734e-01 -6.73339486e-01 -4.20241743e-01
-5.36880314e-01 -1.27025318e+00 1.47934961e+00 8.88622776e-02
-4.31694299e-01 -5.42877913e-01 -1.08278133e-01 3.75981033e-01
5.43312393e-02 4.63040680e-01 4.76977974e-01 -9.84680429e-02
-3.65677476e-01 -2.66336590e-01 -2.71461103e-02 5.35768688e-01
1.68851897e-01 -5.23838341e-01 -6.36120975e-01 -5.26064575e-01
-8.69522318e-02 -3.79969895e-01 7.85328507e-01 3.87998462e-01
1.00250995e+00 -8.92056376e-02 -2.05378711e-01 8.19659412e-01
9.95167017e-01 -3.07181738e-02 4.24337268e-01 2.36295074e-01
9.97566581e-01 4.30662185e-01 7.33769178e-01 5.90325892e-01
4.56880510e-01 1.00775456e+00 3.05131882e-01 -1.33724123e-01
-3.23527694e-01 -5.88856637e-01 4.14394706e-01 8.35325062e-01
-2.50103265e-01 -1.48756588e-02 -5.46390295e-01 2.49012679e-01
-1.80357671e+00 -8.47292244e-01 3.11811984e-01 2.16377234e+00
6.74560189e-01 2.57358342e-01 1.87415779e-01 1.08521566e-01
6.08527005e-01 4.69874084e-01 -5.99585354e-01 -5.97190782e-02
1.02576748e-01 2.72564828e-01 3.83193254e-01 3.89014035e-01
-1.21996593e+00 8.07167411e-01 4.80315065e+00 8.12390625e-01
-1.27901888e+00 -2.22720671e-02 4.11999613e-01 -2.88437665e-01
5.86635470e-02 -2.60283083e-01 -8.40127885e-01 4.28851128e-01
4.67280656e-01 2.34748155e-01 1.36737555e-01 7.96733797e-01
3.65092188e-01 2.24234849e-01 -1.00172353e+00 1.15131819e+00
-4.86688390e-02 -1.03461754e+00 2.06534415e-01 1.07712671e-02
4.00698274e-01 -1.96396932e-02 -5.78179769e-02 2.69657671e-01
-7.76588693e-02 -8.44416440e-01 1.08623755e+00 3.37380886e-01
9.16374683e-01 -8.93751144e-01 6.91806495e-01 3.75380427e-01
-1.65616548e+00 -8.61100927e-02 -1.31695077e-01 -2.77720466e-02
4.63392735e-01 2.98429787e-01 -5.81930578e-01 7.85603881e-01
1.00559914e+00 9.98158455e-01 -3.82548094e-01 7.69854426e-01
-3.45017731e-01 1.72132358e-01 -3.97718132e-01 2.13717431e-01
3.38924468e-01 -1.74594969e-02 6.72115803e-01 1.07834280e+00
2.98031718e-01 5.69570772e-02 4.48642045e-01 4.85026360e-01
2.72531863e-02 -3.18686068e-01 -3.97057682e-01 3.34494501e-01
3.49294990e-01 1.11091030e+00 -2.68309861e-01 -1.98443606e-01
-4.10322726e-01 1.24773276e+00 3.78960758e-01 2.60945410e-01
-1.02670729e+00 -2.54782617e-01 9.08072770e-01 1.01247489e-01
6.70643926e-01 -5.02329469e-01 -1.43312803e-02 -1.16305459e+00
2.07413957e-01 -9.26824152e-01 4.38560665e-01 -7.27292836e-01
-1.04263401e+00 6.84959829e-01 1.96062461e-01 -1.40243852e+00
-4.25681144e-01 -4.79858398e-01 -3.30324829e-01 9.87953663e-01
-1.17059183e+00 -1.28950095e+00 -5.39754927e-01 7.34364569e-01
8.43401790e-01 1.99908018e-01 6.28493965e-01 3.79873812e-01
-6.24763310e-01 6.85792029e-01 -4.31752115e-01 2.62573242e-01
7.36857653e-01 -9.69259083e-01 7.79487789e-01 7.49404609e-01
-9.42594782e-02 5.49154997e-01 5.78601718e-01 -6.97746933e-01
-1.33662856e+00 -1.30112696e+00 8.10485303e-01 -2.64193505e-01
2.81184316e-01 -5.39415658e-01 -7.33792245e-01 6.65278912e-01
-1.22320510e-01 1.05987661e-01 2.43656710e-01 -2.22399637e-01
-2.97708482e-01 -1.16010599e-01 -7.79097378e-01 6.59411490e-01
1.35057485e+00 -4.31586951e-01 -6.83539152e-01 4.34528328e-02
9.39445198e-01 -7.88397014e-01 -9.67483163e-01 6.74203336e-01
6.98439717e-01 -7.62142062e-01 1.25125849e+00 -2.16881901e-01
4.82794166e-01 -4.14563090e-01 -1.32561222e-01 -1.20956481e+00
-4.59319383e-01 -6.63609207e-01 -2.96128094e-01 5.81246912e-01
1.05598837e-01 -3.26983660e-01 8.65309954e-01 1.90310255e-01
-4.46117461e-01 -1.31148005e+00 -1.00981271e+00 -7.35214829e-01
-2.01708645e-01 -2.21431211e-01 4.35307890e-01 4.75387663e-01
-3.48841429e-01 6.71360791e-01 -7.80182242e-01 4.23825234e-02
8.29312444e-01 2.43974477e-02 9.89717424e-01 -8.55112493e-01
-4.45037425e-01 -1.00211821e-01 -5.55774689e-01 -1.75889122e+00
-1.27023488e-01 -6.65214658e-01 1.47398040e-01 -1.26661038e+00
1.64686777e-02 -1.52452588e-01 -1.48397312e-01 5.15410244e-01
-3.50607783e-01 3.76455724e-01 3.49563360e-01 5.95966130e-02
-3.65952551e-01 1.01696718e+00 1.53065968e+00 -3.14131491e-02
-2.31873482e-01 -2.28636600e-02 -3.28174233e-01 6.84046388e-01
5.92218876e-01 -2.05279469e-01 -5.18145919e-01 -6.15764678e-01
-3.20439488e-01 1.48200706e-01 6.33721232e-01 -1.08317745e+00
-2.80136112e-02 1.03356764e-01 7.90729284e-01 -9.23775911e-01
7.63992429e-01 -5.29501975e-01 3.15946430e-01 5.68653762e-01
-1.93977773e-01 1.78319290e-01 8.94282907e-02 5.13667107e-01
-2.56957710e-01 3.14332753e-01 7.03490913e-01 -1.97463468e-01
-7.81078637e-01 7.00051546e-01 1.47117481e-01 -9.62858200e-02
8.83099318e-01 -3.36841226e-01 2.85250634e-01 -2.58529663e-01
-7.64555037e-01 4.05674756e-01 4.04975712e-01 7.57221520e-01
8.34633231e-01 -1.58093321e+00 -7.18239367e-01 3.89098853e-01
5.94694763e-02 4.68598515e-01 6.09978616e-01 8.22306991e-01
-4.24175680e-01 4.90897954e-01 -2.88640589e-01 -8.02807331e-01
-1.22365272e+00 4.14286822e-01 2.84857482e-01 -3.57321620e-01
-1.03462636e+00 1.01383221e+00 4.26286131e-01 -4.29378092e-01
3.04401964e-01 -3.25061977e-01 -3.16958092e-02 -1.65236294e-01
3.60501677e-01 1.19303092e-01 -4.84661870e-02 -1.01657307e+00
-5.02204001e-01 7.23348141e-01 -1.13976091e-01 1.01642966e-01
1.47146189e+00 -1.87679648e-01 3.42309207e-01 1.91530317e-01
1.49475801e+00 -3.16739142e-01 -1.80682647e+00 -3.51659894e-01
-4.05393124e-01 -5.21641970e-01 -1.51702717e-01 -3.48461926e-01
-1.17310762e+00 8.16245377e-01 3.53855997e-01 -5.48039973e-01
1.19606960e+00 -3.24449204e-02 1.17734098e+00 2.14788824e-01
3.44771475e-01 -8.03487182e-01 3.19117725e-01 5.40484726e-01
1.14612067e+00 -9.79365587e-01 3.57620232e-02 -4.23761308e-01
-6.58904791e-01 8.43491495e-01 7.94950366e-01 -3.79512489e-01
4.62984115e-01 2.06538931e-01 -1.33678511e-01 -1.18179046e-01
-5.74512124e-01 -9.77007374e-02 5.30447066e-01 4.12222654e-01
4.20901000e-01 -2.81191673e-02 -1.47954270e-01 5.76800227e-01
-2.43999347e-01 4.99863215e-02 -8.02699178e-02 1.06268752e+00
-2.28837892e-01 -9.93846416e-01 -2.88765609e-01 2.98003137e-01
-2.61011958e-01 1.46495193e-01 -5.22115156e-02 7.59008706e-01
2.91877747e-01 5.09337723e-01 -1.51300833e-01 -7.87532210e-01
5.72412789e-01 -4.14554216e-02 4.69691068e-01 -5.15631139e-01
-4.12064433e-01 3.74862462e-01 7.92925134e-02 -9.77632642e-01
-4.05053228e-01 -6.62652373e-01 -1.21611702e+00 -2.22975701e-01
-2.16882452e-01 -1.30449086e-01 2.43877992e-01 8.27905774e-01
4.74444747e-01 6.58516407e-01 4.61577356e-01 -1.42290962e+00
-7.73579955e-01 -1.06431055e+00 -1.96073189e-01 4.95456189e-01
3.45091015e-01 -1.05687034e+00 1.87352335e-03 4.76972610e-02] | [7.144070625305176, -0.6615970134735107] |
c4905274-6f5f-40b7-8839-ca1b876aab48 | concealed-object-detection | 2102.10274 | null | https://arxiv.org/abs/2102.10274v2 | https://arxiv.org/pdf/2102.10274v2.pdf | Concealed Object Detection | We present the first systematic study on concealed object detection (COD), which aims to identify objects that are "perfectly" embedded in their background. The high intrinsic similarities between the concealed objects and their background make COD far more challenging than traditional object detection/segmentation. To better understand this task, we collect a large-scale dataset, called COD10K, which consists of 10,000 images covering concealed objects in diverse real-world scenarios from 78 object categories. Further, we provide rich annotations including object categories, object boundaries, challenging attributes, object-level labels, and instance-level annotations. Our COD10K is the largest COD dataset to date, with the richest annotations, which enables comprehensive concealed object understanding and can even be used to help progress several other vision tasks, such as detection, segmentation, classification, etc. Motivated by how animals hunt in the wild, we also design a simple but strong baseline for COD, termed the Search Identification Network (SINet). Without any bells and whistles, SINet outperforms 12 cutting-edge baselines on all datasets tested, making them robust, general architectures that could serve as catalysts for future research in COD. Finally, we provide some interesting findings and highlight several potential applications and future directions. To spark research in this new field, our code, dataset, and online demo are available on our project page: http://mmcheng.net/cod. | ['Ling Shao', 'Ming-Ming Cheng', 'Ge-Peng Ji', 'Deng-Ping Fan'] | 2021-02-20 | null | null | null | null | ['camouflaged-object-segmentation', 'dichotomous-image-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.26775244e-01 -1.42274633e-01 -1.46638915e-01 -3.41788918e-01
-5.96935153e-01 -8.74065399e-01 7.35279202e-01 -7.18704611e-02
-4.11828160e-01 2.15879962e-01 2.69923180e-01 -8.42269957e-02
1.85486436e-01 -4.05154079e-01 -8.07118356e-01 -7.96640694e-01
-5.10892451e-01 2.35087305e-01 6.94358230e-01 4.18279599e-03
4.62800533e-01 3.39876413e-01 -1.70401478e+00 5.07103324e-01
4.61499274e-01 1.18486750e+00 3.83387834e-01 5.44669747e-01
3.59739840e-01 4.80882823e-01 -6.83998346e-01 -7.06157982e-01
5.09998083e-01 -1.14725098e-01 -7.31429338e-01 9.26594809e-03
7.89134026e-01 -5.41923285e-01 -1.87635124e-01 1.18909693e+00
4.35016185e-01 -2.95902818e-01 6.70193195e-01 -1.14540684e+00
-8.34277093e-01 5.43976843e-01 -5.95008612e-01 3.99202138e-01
-1.35142818e-01 6.13590896e-01 1.15132153e+00 -9.87854958e-01
5.07802844e-01 1.09625971e+00 9.79429901e-01 6.87285185e-01
-1.04729140e+00 -6.34882390e-01 4.17002827e-01 1.46464974e-01
-1.20291269e+00 -7.86915600e-01 5.97042859e-01 -6.79713428e-01
7.74037659e-01 4.19984043e-01 9.36324358e-01 1.17723072e+00
-2.30514184e-01 1.29024768e+00 8.51078808e-01 -7.31412321e-02
1.79329380e-01 -2.94902120e-02 8.17604288e-02 8.17592859e-01
5.80138147e-01 2.05423206e-01 -5.41146100e-01 8.52808952e-02
4.81374204e-01 -6.31339177e-02 -3.90214831e-01 -5.06435812e-01
-1.28191912e+00 7.14462042e-01 7.33520687e-01 2.10005790e-02
4.82386835e-02 2.95756608e-01 2.05166385e-01 -3.99353495e-03
4.87217903e-01 3.93456399e-01 -4.36926663e-01 1.35263249e-01
-7.51162291e-01 3.27920020e-01 6.82568192e-01 9.77967143e-01
7.04922915e-01 -8.24562013e-02 -4.02190723e-03 8.40392709e-01
4.52519178e-01 3.97950917e-01 2.84422606e-01 -9.69193876e-01
4.24555451e-01 5.93628943e-01 2.04148769e-01 -9.02348340e-01
-4.41884667e-01 -5.42701125e-01 -6.16851926e-01 3.74967009e-01
4.33159113e-01 1.38648758e-02 -1.09895754e+00 1.50292361e+00
2.50763506e-01 1.29944414e-01 -6.15098253e-02 1.21441889e+00
1.17445242e+00 3.77343833e-01 -7.65443668e-02 3.22445929e-01
1.68344986e+00 -1.05427039e+00 -1.16787165e-01 -9.02050078e-01
2.36356437e-01 -5.73666930e-01 9.67423856e-01 2.81120539e-01
-6.67338252e-01 -2.51593262e-01 -9.28238094e-01 8.09239969e-02
-5.80835462e-01 2.22131893e-01 9.37334776e-01 6.01633847e-01
-9.42997932e-01 3.40498596e-01 -7.86313832e-01 -3.52061480e-01
1.07791626e+00 7.92533979e-02 -2.41428614e-01 1.11257769e-02
-4.80419338e-01 6.45554781e-01 1.76235154e-01 5.05811095e-01
-1.74319768e+00 -6.23839557e-01 -8.70325506e-01 -3.60334754e-01
4.70448762e-01 -3.41223955e-01 1.30427170e+00 -8.20017159e-01
-7.83657134e-01 1.57786775e+00 1.53919667e-01 -7.04310954e-01
5.72110951e-01 -2.93486625e-01 5.40216826e-02 3.07981726e-02
2.58288652e-01 9.52638805e-01 9.76239800e-01 -1.64099646e+00
-7.41456866e-01 -3.97295386e-01 2.63159484e-01 -1.40275666e-02
5.35908826e-02 3.14028293e-01 -6.68033540e-01 -8.55157971e-01
1.76595032e-01 -7.98191130e-01 -1.06139965e-01 4.54485327e-01
-5.46318471e-01 4.78807129e-02 6.49995446e-01 -5.55942357e-01
6.62445068e-01 -2.37951016e+00 1.65284961e-01 -4.38395292e-01
5.33568799e-01 4.08695936e-01 -2.86141366e-01 3.21528882e-01
2.32334211e-01 2.75744855e-01 -6.60491765e-01 -5.28243840e-01
2.15951219e-01 9.11158621e-02 -3.36469740e-01 8.01904380e-01
1.92217320e-01 1.15446424e+00 -6.48329496e-01 -3.07393044e-01
-1.01969957e-01 6.45421445e-02 -5.26422381e-01 2.35192895e-01
-5.35561502e-01 2.40609974e-01 -1.69911385e-01 1.26630569e+00
7.90654659e-01 -7.84436986e-02 -3.39499533e-01 -1.47944555e-01
-1.55750021e-01 4.33944277e-02 -8.64446342e-01 1.30284536e+00
-4.09823097e-02 1.14936757e+00 5.97040057e-01 -8.82796288e-01
7.43899465e-01 -3.56756568e-01 1.73911154e-01 -6.07186794e-01
-1.68641172e-02 1.32252648e-01 2.32443716e-02 -6.48289084e-01
6.08702779e-01 2.08971649e-01 -1.99583620e-01 2.32670069e-01
-2.45705321e-01 -5.50822318e-02 -2.66581457e-02 5.60595989e-02
9.75430012e-01 -7.49940947e-02 4.34740968e-02 -3.53592068e-01
-4.62985933e-02 2.00345829e-01 6.32617593e-01 9.55473006e-01
-5.04170954e-01 7.59969354e-01 4.18243349e-01 -7.36188293e-01
-6.39781833e-01 -1.04914725e+00 -3.18451494e-01 1.16159511e+00
7.68519163e-01 -2.65971184e-01 -8.11840475e-01 -6.31856143e-01
4.21381116e-01 -3.16288546e-02 -8.22648644e-01 4.04868871e-02
-4.63001013e-01 -9.98868465e-01 9.31790113e-01 6.58100545e-01
8.90554428e-01 -1.27418864e+00 -9.50291455e-01 -2.12281346e-01
-3.48718733e-01 -1.07850182e+00 -4.74862814e-01 2.75009066e-01
-6.40779018e-01 -1.24918485e+00 -7.73651838e-01 -9.45052683e-01
4.43122506e-01 5.89598954e-01 1.21824682e+00 4.66359556e-01
-6.25546694e-01 3.23158115e-01 -3.56854260e-01 -6.70753002e-01
3.96105051e-02 -2.34850988e-01 -5.48994951e-02 6.66683242e-02
2.05777168e-01 -2.26427242e-01 -8.92993331e-01 6.01479411e-01
-7.13357031e-01 3.58365327e-02 7.47409105e-01 5.07343411e-01
3.46663237e-01 1.88208539e-02 2.12458163e-01 -5.20786285e-01
1.80123746e-01 -5.07057786e-01 -9.84132946e-01 6.54450357e-02
-1.32970095e-01 -2.89696544e-01 -1.86785124e-02 -3.85226160e-01
-7.55552411e-01 -1.47071809e-01 -2.09403619e-01 -8.32790583e-02
-3.72939169e-01 1.11626372e-01 -2.91275114e-01 -3.42695117e-01
5.81637740e-01 3.26289475e-01 -2.95375198e-01 -9.52858210e-01
1.80028379e-01 7.02370703e-01 8.00723016e-01 -4.86328125e-01
8.14329624e-01 7.83698797e-01 -4.45518285e-01 -9.17144656e-01
-9.79861498e-01 -6.46703064e-01 -5.54804564e-01 -5.26397452e-02
9.22328174e-01 -1.10351372e+00 -8.17601085e-01 1.09480274e+00
-1.16694331e+00 -8.61614287e-01 -8.41620043e-02 3.60294320e-02
-3.37558031e-01 2.40231648e-01 -6.82163894e-01 -8.92307222e-01
-1.28669560e-01 -9.86882210e-01 1.46342504e+00 1.57792687e-01
1.35683209e-01 -6.95652843e-01 -3.29243273e-01 5.94751418e-01
3.63659024e-01 3.13808352e-01 5.55767834e-01 -4.54510242e-01
-1.34214878e+00 1.67423174e-01 -5.41807473e-01 2.53827035e-01
-1.49460480e-01 -1.79860041e-01 -1.04599643e+00 -3.21119070e-01
-2.25417569e-01 -2.72852153e-01 1.66855443e+00 3.07312310e-01
1.12156236e+00 -5.26016951e-01 -5.60141325e-01 1.10725856e+00
1.39679682e+00 4.43556644e-02 4.59574014e-01 6.34378672e-01
5.56559563e-01 7.72041082e-01 5.88530123e-01 3.01107347e-01
5.12671828e-01 7.22281575e-01 1.07544482e+00 -9.42334309e-02
-4.28727359e-01 -1.43108159e-01 4.54442650e-01 1.97817639e-01
-2.16186777e-01 -4.16142821e-01 -1.14054966e+00 9.38707650e-01
-1.53852713e+00 -9.30652142e-01 -2.12184355e-01 1.86235785e+00
5.51454663e-01 1.45062581e-01 2.96938062e-01 -2.08768249e-01
5.94791830e-01 3.42441261e-01 -6.78112447e-01 3.03692430e-01
-4.10035849e-01 -3.52101386e-01 5.92842698e-01 3.59517783e-01
-1.60089302e+00 1.03085864e+00 6.89567375e+00 7.09217906e-01
-8.38185191e-01 1.34140998e-01 5.05161405e-01 1.11406624e-01
-3.36124413e-02 -2.06259247e-02 -9.51607406e-01 5.85679650e-01
2.87307054e-01 2.38055229e-01 4.12032366e-01 9.31984127e-01
-1.63013890e-01 -3.55574876e-01 -1.07342482e+00 9.66142595e-01
1.73608035e-01 -1.47708416e+00 -2.71059930e-01 1.27039924e-01
5.73658943e-01 5.04118264e-01 2.05083355e-01 -8.15049745e-03
4.88626689e-01 -9.83304381e-01 1.24667096e+00 2.71918535e-01
6.69948041e-01 -7.91128352e-02 7.12886870e-01 3.37467223e-01
-1.36279249e+00 -2.53500134e-01 -5.09670019e-01 -1.19952179e-01
-1.37283981e-01 1.92536756e-01 -5.89108706e-01 1.25243798e-01
1.31576335e+00 1.01629579e+00 -9.47489023e-01 1.62615991e+00
-1.65086135e-01 6.72631204e-01 -2.36446232e-01 7.32359961e-02
3.20265055e-01 7.14699328e-02 6.83118343e-01 1.47236109e+00
-1.49455875e-01 2.27778070e-02 9.68954936e-02 9.23364639e-01
-1.90258980e-01 -3.77292514e-01 -5.10721505e-01 -4.75519523e-02
3.42384905e-01 1.21943951e+00 -1.09724712e+00 -2.32686937e-01
-2.65821844e-01 8.78438532e-01 1.67423293e-01 2.01303393e-01
-9.23406303e-01 -1.92911595e-01 1.19669473e+00 3.21148597e-02
5.09123683e-01 -2.79100329e-01 -2.76692748e-01 -1.19336390e+00
2.53052443e-01 -8.95397365e-01 3.34354013e-01 -8.09382081e-01
-1.48867333e+00 5.76940179e-01 4.98621613e-02 -1.21617413e+00
4.82145458e-01 -9.49072897e-01 -5.33226490e-01 3.44083190e-01
-1.73811126e+00 -1.27068579e+00 -7.60701835e-01 3.24660391e-01
7.66532958e-01 -2.06468686e-01 3.58058810e-01 2.98965484e-01
-5.18988848e-01 4.42499906e-01 5.66301960e-03 6.94653869e-01
4.57154423e-01 -1.13232756e+00 6.92708671e-01 9.80374515e-01
4.58801955e-01 4.06507343e-01 5.81179142e-01 -6.75998807e-01
-1.29426110e+00 -1.23975360e+00 2.23530263e-01 -9.07829225e-01
7.32816398e-01 -8.22713614e-01 -8.38566601e-01 7.96601593e-01
-2.32470315e-02 1.97843701e-01 2.95967251e-01 -9.45629552e-02
-5.08272231e-01 -2.14904830e-01 -9.28310454e-01 4.76801336e-01
1.70323288e+00 -3.33372265e-01 -6.22218192e-01 4.56799060e-01
7.69334018e-01 -5.07659495e-01 -3.42511743e-01 4.43276554e-01
7.14567244e-01 -1.26596999e+00 1.20464063e+00 -3.45416069e-01
3.05578172e-01 -3.00048590e-01 -5.05029857e-01 -1.00054109e+00
-2.78875649e-01 -4.88665044e-01 1.55798510e-01 1.37726510e+00
3.22482169e-01 -6.45577073e-01 1.01853895e+00 3.38703901e-01
-5.12982845e-01 -6.38085842e-01 -9.51067030e-01 -1.07045555e+00
-1.24013864e-01 -5.19960463e-01 6.80884838e-01 5.45864165e-01
-6.63542330e-01 -8.08478668e-02 -4.88119483e-01 5.08026958e-01
9.95637238e-01 5.49823284e-01 8.28850687e-01 -1.31891215e+00
-1.52099311e-01 -6.98054314e-01 -5.54376900e-01 -1.36964381e+00
-1.34475082e-01 -9.25383270e-01 5.00144839e-01 -1.51823211e+00
3.04083228e-01 -8.75893533e-01 1.66290700e-01 8.71028244e-01
3.22289839e-02 8.43677044e-01 2.62188226e-01 5.31863749e-01
-8.73598158e-01 5.76292992e-01 9.16987956e-01 -4.70822603e-01
1.42208591e-01 1.77857116e-01 -9.05645788e-01 1.08488667e+00
5.81672728e-01 -5.75294435e-01 1.36650845e-01 -8.81448328e-01
-1.96203589e-01 -4.90832478e-01 1.02171290e+00 -8.86871934e-01
2.23068133e-01 -2.67425060e-01 3.57663006e-01 -5.82349300e-01
6.36504591e-01 -8.07298958e-01 7.12138042e-02 4.70408112e-01
3.91248837e-02 -2.76395798e-01 3.30437750e-01 6.86722040e-01
-1.13743603e-01 -2.72414207e-01 8.13391924e-01 -4.21040297e-01
-1.29753327e+00 3.94952297e-01 -3.48319113e-01 2.99096733e-01
1.03967321e+00 -4.25969839e-01 -8.37547243e-01 -2.62724124e-02
-4.83788937e-01 2.56834060e-01 6.12817585e-01 5.15829504e-01
7.31275022e-01 -9.75972056e-01 -8.05892587e-01 3.05727214e-01
3.47158641e-01 1.95586503e-01 5.78257777e-02 6.29824996e-01
-6.09025717e-01 1.41250193e-01 -1.33037776e-01 -9.28154469e-01
-1.39159739e+00 5.56752503e-01 5.55732489e-01 4.37770814e-01
-8.54745448e-01 1.25522637e+00 5.99901140e-01 -2.15623245e-01
4.41721380e-01 -3.45020622e-01 2.99162208e-03 1.82594493e-01
5.75904012e-01 1.53902709e-01 -3.40297148e-02 -6.28253698e-01
-7.39739954e-01 8.44858766e-01 2.60067493e-01 2.86145717e-01
1.60124147e+00 -1.28911540e-01 -3.20473433e-01 1.06387965e-01
8.35939407e-01 8.73070210e-02 -1.68821943e+00 -2.80267894e-01
3.25518280e-01 -8.19521844e-01 -3.02461952e-01 -8.51625562e-01
-1.24806511e+00 8.90011847e-01 5.07692099e-01 2.80981302e-01
1.12022269e+00 6.65973961e-01 4.16791439e-01 4.17757213e-01
5.51167309e-01 -6.10116780e-01 3.33882511e-01 2.67040312e-01
1.17826223e+00 -1.52195704e+00 -1.31229371e-01 -3.29085201e-01
-6.57234728e-01 5.95271885e-01 7.25738406e-01 1.10197207e-02
3.70962352e-01 5.54003596e-01 2.17666432e-01 -3.43422413e-01
-6.63850963e-01 -6.92085981e-01 2.19574600e-01 1.02548921e+00
-2.05063760e-01 -5.18535562e-02 3.53453249e-01 7.76019990e-01
-3.02971065e-01 -6.03511333e-01 3.58712435e-01 6.37679040e-01
-7.90834725e-01 -6.97880149e-01 -4.62883979e-01 2.67694503e-01
-4.15569484e-01 -2.17171729e-01 -7.41986513e-01 7.83065915e-01
4.60428447e-01 8.74214649e-01 -2.06832848e-02 -3.14478308e-01
2.23018676e-01 -4.43875879e-01 3.29370141e-01 -6.17453635e-01
-5.07823408e-01 -1.95023224e-01 2.58038938e-01 -5.57124317e-01
-5.15161097e-01 -6.76907182e-01 -7.97976255e-01 -1.68966696e-01
-2.83741146e-01 3.34031582e-02 5.96416295e-01 6.01541519e-01
2.08791852e-01 1.77749217e-01 1.60373643e-01 -1.26318085e+00
-1.61612451e-01 -7.54083574e-01 -3.82605046e-01 1.74708158e-01
8.70747983e-01 -7.85511732e-01 -4.72391665e-01 2.57766187e-01] | [9.513334274291992, -0.04513311758637428] |
10ce663f-36bb-462e-9464-8563e5cd8db6 | visualgptscore-visio-linguistic-reasoning | 2306.01879 | null | https://arxiv.org/abs/2306.01879v1 | https://arxiv.org/pdf/2306.01879v1.pdf | VisualGPTScore: Visio-Linguistic Reasoning with Multimodal Generative Pre-Training Scores | Vision-language models (VLMs) discriminatively pre-trained with contrastive image-text matching losses such as $P(\text{match}|\text{text}, \text{image})$ have been criticized for lacking compositional understanding. This means they might output similar scores even if the original caption is rearranged into a different semantic statement. To address this, we propose to use the ${\bf V}$isual ${\bf G}$enerative ${\bf P}$re-${\bf T}$raining Score (${\bf VisualGPTScore}$) of $P(\text{text}|\text{image})$, a $\textit{multimodal generative}$ score that captures the likelihood of a text caption conditioned on an image using an image-conditioned language model. Contrary to the belief that VLMs are mere bag-of-words models, our off-the-shelf VisualGPTScore demonstrates top-tier performance on recently proposed image-text retrieval benchmarks like ARO and Crepe that assess compositional reasoning. Furthermore, we factorize VisualGPTScore into a product of the $\textit{marginal}$ P(text) and the $\textit{Pointwise Mutual Information}$ (PMI). This helps to (a) diagnose datasets with strong language bias, and (b) debias results on other benchmarks like Winoground using an information-theoretic framework. VisualGPTScore provides valuable insights and serves as a strong baseline for future evaluation of visio-linguistic compositionality. | ['Deva Ramanan', 'Pengchuan Zhang', 'Deepak Pathak', 'Xinyue Chen', 'Zhiqiu Lin'] | 2023-06-02 | null | null | null | null | ['image-text-matching', 'text-matching'] | ['computer-vision', 'natural-language-processing'] | [ 5.61248958e-01 1.91323325e-01 -9.06133950e-02 -5.90750098e-01
-1.31132400e+00 -6.44323766e-01 1.04137218e+00 6.86192587e-02
-7.08041012e-01 3.54183584e-01 2.33693242e-01 -5.69828033e-01
-3.01546864e-02 -6.74460828e-01 -1.16671503e+00 -7.16055870e-01
2.39866242e-01 3.93874884e-01 -3.66984047e-02 6.56762868e-02
9.78753045e-02 8.26333091e-02 -1.71567059e+00 6.30097389e-01
6.33682907e-01 1.28102970e+00 3.88296396e-01 5.31884491e-01
-3.03628862e-01 9.01595235e-01 -2.01928630e-01 -1.05952919e+00
2.15841845e-01 -6.19295776e-01 -6.48020506e-01 -1.19733652e-02
1.29522836e+00 -3.78923267e-01 -5.95760703e-01 1.48171556e+00
2.36523226e-01 1.67838514e-01 1.01184475e+00 -1.22725427e+00
-9.01923180e-01 6.80591702e-01 -7.92120159e-01 3.57923925e-01
2.07383245e-01 6.54513001e-01 1.72318888e+00 -1.10183096e+00
7.37986267e-01 1.57342219e+00 3.45955104e-01 2.92417407e-01
-1.38112009e+00 -8.41553807e-01 5.08311987e-01 9.69970226e-03
-1.33119738e+00 -3.31382334e-01 4.01336133e-01 -5.18128812e-01
1.10335112e+00 4.35213894e-01 2.84209639e-01 9.38256502e-01
2.66078323e-01 1.22174597e+00 1.25933969e+00 -3.48759502e-01
-2.33708456e-01 2.05253571e-01 7.50239491e-02 1.03094399e+00
-9.53839254e-03 9.52011347e-02 -6.59785509e-01 9.62167606e-02
5.36426604e-01 -2.34793082e-01 -2.01871455e-01 -1.52059942e-01
-1.23747444e+00 7.93743670e-01 5.04159689e-01 1.79698035e-01
-1.24287501e-01 8.08322430e-01 2.55918175e-01 9.22735110e-02
1.53044030e-01 1.94850981e-01 -4.86272722e-02 2.06556320e-01
-1.14868546e+00 3.65915328e-01 2.78603643e-01 9.66516495e-01
9.80483055e-01 -3.75958516e-05 -4.73789841e-01 7.87704885e-01
4.95909840e-01 1.11714041e+00 1.91916272e-01 -1.16215551e+00
7.16343403e-01 3.22266579e-01 -1.78211883e-01 -9.42990065e-01
-9.61422455e-03 -6.83093742e-02 -4.94046539e-01 8.10861215e-02
3.68996084e-01 2.27486923e-01 -1.11523819e+00 1.97710454e+00
-4.67991143e-01 -4.71592247e-01 -2.32109293e-01 9.64794099e-01
9.93355393e-01 7.27725327e-01 5.83550632e-01 4.17698808e-02
1.47308862e+00 -8.46246123e-01 -3.58444273e-01 -7.44400680e-01
3.20025980e-01 -9.71668661e-01 1.52785301e+00 1.97517052e-01
-1.46253765e+00 -5.56930840e-01 -8.30269217e-01 -2.96547443e-01
-2.11259604e-01 -1.05637692e-01 6.08859837e-01 5.64736366e-01
-1.35486460e+00 1.63801104e-01 -6.22983098e-01 -3.48940849e-01
6.17333591e-01 1.70120895e-01 -3.38874251e-01 -5.03194213e-01
-9.54555511e-01 7.82508373e-01 1.98658228e-01 -3.19882751e-01
-1.26655245e+00 -5.35970628e-01 -1.02939773e+00 2.78702918e-02
4.16238576e-01 -8.95636737e-01 9.24926758e-01 -1.10322821e+00
-7.11097658e-01 1.53231764e+00 -3.83081555e-01 -4.21470016e-01
5.89548469e-01 2.68787704e-02 -2.37306088e-01 5.22876084e-01
4.54906523e-01 1.37468791e+00 1.05856788e+00 -1.40579736e+00
-5.26291013e-01 -5.03129661e-01 7.97396377e-02 4.05285120e-01
-4.79314625e-02 2.39191979e-01 -7.37646222e-01 -7.82259166e-01
1.79648519e-01 -8.67825747e-01 2.56333053e-01 2.99408227e-01
-3.92964065e-01 -1.35297626e-01 2.62918174e-01 -6.69561446e-01
8.81714940e-01 -2.09157872e+00 1.25910535e-01 1.61462054e-01
3.57415557e-01 -3.63932326e-02 -2.93197304e-01 2.03248933e-01
-1.37776151e-01 3.52023244e-01 -3.45452696e-01 -4.54357743e-01
2.24522039e-01 1.12915613e-01 -5.26986957e-01 4.11950439e-01
2.10831001e-01 1.14265215e+00 -6.17166102e-01 -7.72821546e-01
3.48427027e-01 3.48791510e-01 -5.23694754e-01 -1.78961813e-01
-5.27866423e-01 -1.03091300e-01 -2.65264571e-01 4.27722335e-01
5.74429989e-01 -5.78980386e-01 -1.52720913e-01 -2.75704354e-01
1.39343247e-01 1.13311768e-01 -6.58711851e-01 1.76992118e+00
-6.68606386e-02 7.32088923e-01 1.05799101e-01 -1.06018317e+00
4.62756425e-01 -1.52533427e-01 3.40576708e-01 -1.14249861e+00
1.46075189e-01 -1.23350419e-01 -1.05476722e-01 -2.13427529e-01
4.39770132e-01 -5.73198438e-01 -1.23338126e-01 5.82338989e-01
3.40443820e-01 -5.44404626e-01 7.12101981e-02 7.02501357e-01
9.96595800e-01 1.45395145e-01 -3.66067767e-01 -2.97229648e-01
4.53582227e-01 4.22261748e-03 -9.49796364e-02 1.29632592e+00
-2.44713783e-01 6.81014895e-01 5.74712038e-01 9.37066153e-02
-1.15926838e+00 -1.70800161e+00 -1.74602300e-01 1.29606962e+00
2.66957402e-01 -2.91198462e-01 -5.91417134e-01 -4.72644895e-01
5.99006005e-02 1.05914867e+00 -5.36835730e-01 -1.25592142e-01
-3.19842994e-01 -6.94512367e-01 9.27337945e-01 5.06563067e-01
4.42829043e-01 -1.08079588e+00 -3.55514020e-01 -2.92202234e-01
-4.61162269e-01 -1.16023839e+00 -6.83173895e-01 1.94384888e-01
-3.51507813e-01 -6.21241927e-01 -8.02684367e-01 -7.66310632e-01
6.66662991e-01 2.01879218e-01 1.42834187e+00 -7.39239305e-02
-4.47569221e-01 1.00311971e+00 -1.28562599e-01 -4.21642214e-01
-3.67003649e-01 -6.55343831e-01 -3.14406544e-01 -1.48007780e-01
5.43898404e-01 -7.72500560e-02 -8.52269411e-01 2.52071440e-01
-1.22487879e+00 2.10449472e-01 6.12211049e-01 8.48144710e-01
1.01070619e+00 -1.51580065e-01 2.53626704e-02 -4.71673697e-01
3.95565689e-01 -1.12032019e-01 -4.96472359e-01 5.06537318e-01
-7.35175550e-01 6.79245517e-02 2.95652568e-01 -3.77283543e-01
-9.31664288e-01 -4.83575523e-01 7.37793520e-02 -7.02851593e-01
6.79168478e-02 4.93960559e-01 1.38618285e-02 3.19938302e-01
6.10916257e-01 6.41842723e-01 9.65542868e-02 -3.39279394e-03
6.48481965e-01 2.06879586e-01 7.44701266e-01 -8.07072759e-01
5.05840719e-01 7.94904292e-01 -1.51752219e-01 -5.51365316e-01
-8.53088856e-01 -3.78501356e-01 -1.15594128e-02 -1.75418317e-01
1.14099967e+00 -1.24899983e+00 -9.04314458e-01 1.35401741e-01
-9.57389951e-01 -2.73453981e-01 -1.88523442e-01 4.15148407e-01
-9.05499220e-01 7.06815302e-01 -4.76738125e-01 -8.82100880e-01
-3.48789155e-01 -1.23076069e+00 1.28348744e+00 -3.01728509e-02
-2.47570544e-01 -7.45016158e-01 -5.60389757e-01 9.56099033e-01
1.31130442e-01 -1.78097486e-01 1.25032771e+00 -2.45929480e-01
-7.52901733e-01 -2.78517395e-01 -9.92543221e-01 6.98468328e-01
-4.97790515e-01 -2.57266134e-01 -1.09835863e+00 -3.35204303e-01
1.19118392e-02 -5.99844277e-01 1.32410371e+00 5.48454225e-01
1.07980371e+00 -2.33301774e-01 -1.17381893e-01 3.80596548e-01
1.60107970e+00 -4.91818879e-03 8.41982245e-01 -9.43325609e-02
5.55893719e-01 4.94250596e-01 2.89956748e-01 1.55846074e-01
4.75344330e-01 4.25532162e-01 4.35132205e-01 3.93353514e-02
-5.45797110e-01 -5.00393748e-01 7.22365201e-01 2.08457902e-01
2.77314872e-01 -6.35139465e-01 -1.05886030e+00 6.22869730e-01
-1.70368278e+00 -1.12029183e+00 1.03240669e-01 1.98098385e+00
7.68934906e-01 1.99095979e-01 -3.28104228e-01 -4.50525224e-01
5.62967837e-01 4.62564170e-01 -4.51412678e-01 -1.74761608e-01
-6.68333113e-01 5.08583844e-01 5.25408030e-01 6.44591451e-01
-9.36984003e-01 1.23419476e+00 5.26597834e+00 1.19710243e+00
-1.06926155e+00 1.22545727e-01 8.79455209e-01 -3.53346705e-01
-7.68478990e-01 6.35323077e-02 -6.97110891e-01 4.28744167e-01
6.42136455e-01 6.55598417e-02 6.03847802e-01 5.48151493e-01
-1.47422493e-01 -5.84891915e-01 -1.25111914e+00 1.30176282e+00
6.12473965e-01 -1.31616151e+00 6.38098717e-01 6.12353720e-02
5.53823352e-01 3.80654901e-01 7.82517076e-01 3.48033965e-01
5.84208727e-01 -1.46313083e+00 1.19379508e+00 4.15768594e-01
1.17798543e+00 -1.85199380e-01 3.21759462e-01 9.53430012e-02
-1.00937510e+00 1.53119221e-01 -2.68336624e-01 2.24856958e-01
4.37640138e-02 2.96162128e-01 -6.32196367e-01 3.51882160e-01
8.99643540e-01 3.12709153e-01 -5.81367731e-01 3.41472715e-01
-1.19833402e-01 4.08763170e-01 -2.66569138e-01 5.40169515e-02
5.07553756e-01 3.03347297e-02 6.15041196e-01 1.21749747e+00
7.91431963e-02 4.91053313e-02 1.95751026e-01 1.44190872e+00
-2.67443925e-01 2.29564607e-01 -5.29824078e-01 -4.15867656e-01
-1.38662020e-02 7.49883711e-01 -7.57168472e-01 -3.32228303e-01
-5.76863647e-01 9.56426144e-01 1.70048967e-01 6.52282178e-01
-9.47349370e-01 -4.43712473e-02 4.86203164e-01 2.74180293e-01
3.49699736e-01 -8.72366801e-02 -4.83570725e-01 -1.18736207e+00
1.05804712e-01 -7.02953458e-01 5.59502065e-01 -1.42355549e+00
-1.50841987e+00 3.07744861e-01 1.91141501e-01 -7.50960469e-01
1.16342045e-02 -8.26370001e-01 1.21278707e-02 1.16466749e+00
-1.39347935e+00 -1.26841867e+00 -5.94625995e-02 8.72033477e-01
4.79868323e-01 -2.38668770e-01 5.72240472e-01 7.54608810e-02
-3.52882817e-02 7.60682166e-01 -1.74646184e-01 2.83239156e-01
7.84500599e-01 -1.05343890e+00 4.39251810e-02 9.11872864e-01
4.15605754e-01 6.97683811e-01 8.42619240e-01 -6.75030470e-01
-1.34736431e+00 -9.01801705e-01 7.54729807e-01 -8.16248417e-01
7.85303295e-01 -1.95376888e-01 -5.14012754e-01 9.22264040e-01
3.57816964e-01 -1.56406149e-01 5.09505093e-01 -2.50282496e-01
-8.46562505e-01 -1.05443308e-02 -1.02820373e+00 8.72320056e-01
1.02684820e+00 -1.02702403e+00 -6.66943312e-01 3.34987015e-01
6.32516444e-01 -3.89168411e-02 -6.58945560e-01 3.91104728e-01
5.55967391e-01 -1.07850504e+00 1.08982074e+00 -5.10953844e-01
8.35641623e-01 -4.55507115e-02 -7.31471241e-01 -5.12286425e-01
-1.84940815e-01 -2.88933665e-01 5.32027423e-01 8.79720986e-01
5.93321562e-01 -2.28887230e-01 6.24832749e-01 9.15562689e-01
-2.93872003e-02 -2.05159038e-01 -1.03636646e+00 -5.98965883e-01
4.45445538e-01 -9.09074008e-01 3.94886583e-02 7.72872448e-01
-2.05593228e-01 2.70471483e-01 -7.17729181e-02 -1.81845888e-01
7.72869349e-01 -1.09931529e-01 4.90308464e-01 -7.73419023e-01
-4.38480854e-01 -8.56768608e-01 -3.81940305e-01 -1.07136655e+00
2.60626763e-01 -1.31555033e+00 1.37281314e-01 -1.51275396e+00
8.89369726e-01 -2.96287209e-01 -4.05329615e-01 5.58127642e-01
-7.54757673e-02 3.67614806e-01 4.87768829e-01 1.81364194e-01
-8.52228105e-01 3.69203597e-01 1.23628843e+00 -6.63791656e-01
5.44993043e-01 -7.85452843e-01 -8.29510689e-01 6.82944477e-01
1.32133603e-01 -3.28850299e-01 -4.44793075e-01 -7.64038980e-01
4.95950848e-01 1.90250948e-01 8.00994337e-01 -6.00201786e-01
1.31661236e-01 -1.32587329e-01 3.24460745e-01 -5.59776068e-01
5.92333674e-01 -6.18899107e-01 3.10950894e-02 3.32463354e-01
-5.78423321e-01 9.25019477e-03 2.98099786e-01 6.21062577e-01
-2.84046263e-01 -1.70439929e-01 7.96845734e-01 -5.23318350e-01
-5.20772576e-01 1.67800605e-01 -1.89658403e-01 3.29140216e-01
6.74748898e-01 -1.58449784e-01 -5.36876440e-01 -6.48472369e-01
-5.96219838e-01 2.94686377e-01 6.46144271e-01 3.60737473e-01
7.07956672e-01 -1.03788233e+00 -7.45878458e-01 2.20009461e-02
5.81182837e-01 -1.14078104e-01 5.19089401e-01 8.53630424e-01
-4.37950790e-01 6.16938949e-01 1.71408858e-02 -9.01203394e-01
-1.06528807e+00 4.90871578e-01 3.67414176e-01 -2.62250811e-01
-4.01045352e-01 1.41108799e+00 9.11052465e-01 3.08792386e-03
2.96221554e-01 -2.45025843e-01 5.19182086e-01 -4.35048118e-02
3.89122754e-01 -9.26481038e-02 -1.80222422e-01 -9.10253167e-01
-3.92424673e-01 4.29008693e-01 -2.50258952e-01 -7.07719862e-01
8.38666618e-01 -2.70123005e-01 -3.85284185e-01 2.35257521e-01
1.34788549e+00 -2.61084855e-01 -1.13967264e+00 -3.95516902e-01
-3.72409731e-01 -5.18585682e-01 3.67940404e-02 -1.25337398e+00
-8.04116786e-01 1.00124156e+00 6.71314657e-01 -2.87726730e-01
8.31085503e-01 5.12838662e-01 5.06866097e-01 1.29930913e-01
3.34826410e-01 -1.10091805e+00 4.43727762e-01 2.82606363e-01
8.94064963e-01 -1.48872399e+00 6.03152327e-02 -6.92075863e-02
-9.16544557e-01 6.30574703e-01 3.07862014e-01 2.50330120e-02
5.05692303e-01 -4.92448546e-02 -1.28496930e-01 -4.02020007e-01
-5.24141252e-01 -3.52721959e-01 6.64904177e-01 3.10569316e-01
4.43175197e-01 1.38887197e-01 -2.04771519e-01 5.81302822e-01
-2.39469379e-01 -2.75897354e-01 3.67709361e-02 6.32445276e-01
-3.50236118e-01 -6.74212754e-01 -2.01251581e-01 6.02125585e-01
-5.91254056e-01 -5.81572771e-01 -3.85601461e-01 6.70422852e-01
1.01145841e-01 8.94414961e-01 2.70818442e-01 -1.29922360e-01
-1.45382866e-01 3.63959074e-01 8.31280529e-01 -4.72872734e-01
-4.48539943e-01 2.72603244e-01 -5.65923713e-02 -4.57980722e-01
-4.44601178e-01 -8.30752254e-01 -1.37425637e+00 -2.36189365e-01
1.39817953e-01 -3.33476812e-01 6.28007650e-01 8.61456454e-01
-4.68784533e-02 1.98760763e-01 -3.26527357e-02 -4.35266912e-01
-1.86913639e-01 -6.47882462e-01 -5.93753874e-01 9.34525847e-01
2.89329905e-02 -4.20178592e-01 -3.56500268e-01 2.96301305e-01] | [10.837247848510742, 1.596927523612976] |
5f1a5f18-29e2-4c4f-b01d-5e8ba68d9e84 | adaptive-unfolding-total-variation-network | 2110.00984 | null | https://arxiv.org/abs/2110.00984v4 | https://arxiv.org/pdf/2110.00984v4.pdf | Adaptive Unfolding Total Variation Network for Low-Light Image Enhancement | Real-world low-light images suffer from two main degradations, namely, inevitable noise and poor visibility. Since the noise exhibits different levels, its estimation has been implemented in recent works when enhancing low-light images from raw Bayer space. When it comes to sRGB color space, the noise estimation becomes more complicated due to the effect of the image processing pipeline. Nevertheless, most existing enhancing algorithms in sRGB space only focus on the low visibility problem or suppress the noise under a hypothetical noise level, leading them impractical due to the lack of robustness. To address this issue,we propose an adaptive unfolding total variation network (UTVNet), which approximates the noise level from the real sRGB low-light image by learning the balancing parameter in the model-based denoising method with total variation regularization. Meanwhile, we learn the noise level map by unrolling the corresponding minimization process for providing the inferences of smoothness and fidelity constraints. Guided by the noise level map, our UTVNet can recover finer details and is more capable to suppress noise in real captured low-light scenes. Extensive experiments on real-world low-light images clearly demonstrate the superior performance of UTVNet over state-of-the-art methods. | ['Wentian Shi', 'Daming Shi', 'Chuanjun Zheng'] | 2021-10-03 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zheng_Adaptive_Unfolding_Total_Variation_Network_for_Low-Light_Image_Enhancement_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zheng_Adaptive_Unfolding_Total_Variation_Network_for_Low-Light_Image_Enhancement_ICCV_2021_paper.pdf | iccv-2021-1 | ['noise-estimation'] | ['medical'] | [ 4.81330514e-01 -5.57831824e-01 5.34538031e-01 -3.05436403e-01
-5.22508025e-01 -5.15284091e-02 1.53128237e-01 -5.63334942e-01
-3.71697575e-01 6.11499608e-01 7.09352195e-02 -4.79123965e-02
-1.82209998e-01 -7.93245316e-01 -4.88546133e-01 -1.36534083e+00
6.43182337e-01 -2.97354490e-01 2.64388859e-01 -3.63787740e-01
4.91366386e-02 2.51118511e-01 -1.48223984e+00 -1.23879556e-02
1.23492360e+00 1.01020217e+00 6.15383267e-01 2.94485152e-01
-2.47561559e-01 6.98643386e-01 -3.52859229e-01 -5.33558875e-02
7.18706548e-01 -4.45444107e-01 1.07762784e-01 2.57917702e-01
6.57040119e-01 -4.95271683e-01 -6.12785459e-01 1.66734743e+00
5.97980380e-01 4.13359553e-01 5.32641187e-02 -9.09690619e-01
-6.21859729e-01 6.82793558e-02 -9.05400991e-01 2.32496485e-01
-5.97235709e-02 7.14667141e-01 5.58462560e-01 -8.12147200e-01
3.60115707e-01 1.13328981e+00 5.82344174e-01 2.74814218e-01
-1.33282614e+00 -8.11489999e-01 1.37079671e-01 2.50230551e-01
-1.19872105e+00 -4.03619140e-01 9.70949948e-01 -1.62570700e-01
3.12766165e-01 1.85892805e-01 6.01684570e-01 6.82850182e-01
1.98449284e-01 2.18339950e-01 1.52079546e+00 -1.96536452e-01
-3.56734805e-02 -2.46840224e-01 4.37021926e-02 4.30384725e-01
4.31700855e-01 4.22356516e-01 -3.56853575e-01 1.37708053e-01
8.69303286e-01 1.81741700e-01 -7.74755776e-01 -2.16497853e-01
-9.44585562e-01 2.61426866e-01 5.52993357e-01 2.50971764e-01
-2.58891284e-01 4.75835167e-02 7.68572912e-02 2.57698447e-01
5.69137931e-01 2.52707809e-01 -3.22078288e-01 1.93471044e-01
-6.60463572e-01 -1.82736013e-02 1.55632064e-01 7.12583482e-01
1.02289689e+00 2.02837512e-01 -3.20537478e-01 1.05219197e+00
3.66794288e-01 6.35143518e-01 -1.88191123e-02 -1.06463003e+00
5.01604617e-01 4.53969002e-01 2.50399798e-01 -1.10800695e+00
-2.93130994e-01 -7.88361371e-01 -1.32734191e+00 5.42050481e-01
4.10955071e-01 7.27775693e-02 -8.97952199e-01 1.63443553e+00
4.07611221e-01 5.12146175e-01 -1.22596428e-01 1.35358751e+00
8.59842300e-01 8.52529645e-01 -2.18951598e-01 -7.71643579e-01
9.90896881e-01 -7.78178394e-01 -1.10552561e+00 -2.67887682e-01
-5.70788383e-02 -9.21244323e-01 1.31590116e+00 7.17083097e-01
-1.02980769e+00 -5.84676087e-01 -8.96295846e-01 -3.00275743e-01
1.83495134e-01 2.09157944e-01 3.51063311e-01 4.13320214e-01
-8.42224300e-01 4.39772397e-01 -7.02500641e-01 1.22212902e-01
2.53264964e-01 -1.13370471e-01 -2.97834538e-02 -5.64058363e-01
-1.02410817e+00 5.41471779e-01 -3.83730046e-02 8.71141076e-01
-7.13129818e-01 -6.87640011e-01 -5.63570201e-01 -6.97614253e-02
8.35182786e-01 -6.91255510e-01 5.92382550e-01 -6.59284890e-01
-1.64168918e+00 5.22098362e-01 -1.27541542e-01 1.72699079e-01
6.22900188e-01 -4.69871648e-02 -4.63658869e-01 -3.58138978e-02
3.90591249e-02 -1.35943756e-01 1.16944945e+00 -1.62517190e+00
-4.76930052e-01 -3.70576799e-01 -8.79857037e-03 1.35254368e-01
-6.34035245e-02 -1.58577949e-01 -8.11565638e-01 -7.18330443e-01
5.69080353e-01 -6.26492202e-01 -3.60539854e-01 2.32733980e-01
-2.62232512e-01 1.95192397e-01 8.57632399e-01 -7.79899001e-01
1.27007270e+00 -2.29916573e+00 -8.52866657e-03 -9.37596262e-02
4.47634101e-01 3.15277278e-01 -2.92209327e-01 -1.45701868e-02
-1.46431234e-02 -2.84697741e-01 -5.13096690e-01 -1.77284613e-01
-3.57002795e-01 2.03124821e-01 -1.07913218e-01 5.55987656e-01
-1.59306169e-01 4.15869951e-01 -9.52491879e-01 -3.98517519e-01
4.18823808e-01 7.33792603e-01 -3.38238388e-01 3.15779150e-01
-1.55914828e-01 8.16390395e-01 -3.58329415e-01 5.09901464e-01
1.30851424e+00 -1.53790079e-02 -1.76378623e-01 -8.36737216e-01
-4.79586959e-01 -2.78962225e-01 -1.49268878e+00 1.50739539e+00
-6.99208021e-01 5.14286637e-01 7.31431425e-01 -5.27020693e-01
9.52688932e-01 -2.09692881e-01 2.64266402e-01 -8.55367362e-01
1.88992128e-01 2.35563904e-01 -1.43893421e-01 -7.45296299e-01
2.52858996e-01 -5.39873481e-01 5.54249167e-01 4.13388424e-02
-5.17591059e-01 -4.26743418e-01 1.00093924e-01 -2.59650294e-02
9.21969116e-01 2.47981444e-01 -1.27383545e-01 -4.65419143e-02
7.77860463e-01 -5.59786260e-01 9.29669440e-01 7.35504687e-01
-3.50729674e-01 7.78622687e-01 1.26928553e-01 -2.95316249e-01
-8.19167256e-01 -9.88469541e-01 -3.07481498e-01 7.93935239e-01
5.55591583e-01 -1.04771227e-01 -4.75616068e-01 -1.84948698e-01
-3.13243866e-01 6.94750130e-01 -2.86027193e-01 -1.49666011e-01
-6.51450813e-01 -1.05407083e+00 -1.64852053e-01 -1.02055706e-01
9.93599653e-01 -7.33196735e-01 -2.14953706e-01 1.33178100e-01
-4.66034889e-01 -1.11422002e+00 -5.20095706e-01 -9.85414386e-02
-6.42312825e-01 -1.08454931e+00 -4.67880696e-01 -4.79926229e-01
5.71419477e-01 8.10665846e-01 9.12312925e-01 2.65826136e-01
-4.06192034e-01 4.13401797e-02 -4.38774452e-02 -1.84832722e-01
-2.13702112e-01 -5.68156362e-01 -1.34347007e-01 4.46116626e-01
1.21609934e-01 -6.77287459e-01 -9.91114736e-01 4.08079833e-01
-1.16025686e+00 2.36827239e-01 5.62773108e-01 9.99673724e-01
8.80251110e-01 7.39834189e-01 1.03025742e-01 -6.37235701e-01
5.45574427e-01 -2.26938073e-02 -1.00323188e+00 1.80533245e-01
-6.79690182e-01 -2.64531791e-01 8.01766634e-01 -3.97015899e-01
-1.59277225e+00 -1.41009226e-01 -2.49746758e-02 -5.40644467e-01
2.68380884e-02 3.13889012e-02 -4.13672477e-01 -3.30266565e-01
4.01499957e-01 4.39111024e-01 -9.96829383e-03 -4.99157131e-01
2.10521147e-01 5.06709337e-01 5.34126222e-01 -3.30597609e-01
1.12602210e+00 8.35845947e-01 3.77608031e-01 -1.01014829e+00
-1.03210449e+00 -3.98862034e-01 -9.63900015e-02 -3.23195666e-01
7.78962612e-01 -9.83203053e-01 -9.06513095e-01 7.60643363e-01
-8.57224941e-01 -4.37617332e-01 -3.01245600e-01 3.38374555e-01
-2.98172385e-01 6.91922009e-01 -7.56383181e-01 -7.31000721e-01
-1.99786142e-01 -1.31864953e+00 8.85597169e-01 3.25805068e-01
8.37480843e-01 -6.20722115e-01 -1.42934099e-01 4.25138831e-01
6.76621675e-01 6.07699417e-02 8.19211900e-01 8.11487973e-01
-1.09376705e+00 1.85270831e-01 -7.13377059e-01 7.08716750e-01
2.67839730e-01 -2.76045911e-02 -9.41947997e-01 -4.26707059e-01
7.16972411e-01 4.23107995e-03 1.04559040e+00 7.78481245e-01
1.36700296e+00 -5.22322506e-02 1.99308276e-01 1.28358388e+00
2.03030229e+00 5.67579754e-02 8.33460689e-01 5.35134792e-01
8.49953592e-01 5.25768697e-01 7.49505699e-01 3.28508019e-01
-1.13597279e-02 7.32968986e-01 7.30255544e-01 -3.96351725e-01
-4.94198859e-01 1.14683360e-01 5.87070137e-02 7.28268921e-01
-1.13790751e-01 -7.40438551e-02 -5.15818715e-01 1.31889746e-01
-1.57981777e+00 -7.86001325e-01 -6.85242772e-01 2.14439464e+00
9.80269790e-01 -7.10521489e-02 -4.52985317e-01 -8.32924247e-02
6.78380787e-01 4.38881606e-01 -6.80841982e-01 3.22323054e-01
-5.85768640e-01 -1.97599471e-01 5.18732131e-01 6.73077881e-01
-6.88954651e-01 6.74576044e-01 5.13278103e+00 1.02510214e+00
-1.12874115e+00 2.03649774e-01 7.07846105e-01 -2.30999589e-01
-3.28067660e-01 7.37548023e-02 -6.19928658e-01 8.06665242e-01
4.54281755e-02 1.76260516e-01 8.45113933e-01 1.96862966e-01
8.55910599e-01 -3.50400925e-01 -3.12156558e-01 1.31540060e+00
-4.92347740e-02 -7.50771284e-01 1.51595837e-02 -1.25354588e-01
7.91002870e-01 1.07132141e-02 1.01112276e-01 -3.52799036e-02
9.15257260e-02 -5.96371472e-01 3.88351262e-01 9.34616983e-01
8.06875646e-01 -5.23035347e-01 6.53036952e-01 3.84331912e-01
-9.95459199e-01 -1.77809268e-01 -8.13685656e-01 5.20247780e-02
4.47766274e-01 1.10797167e+00 7.24288821e-03 5.07752776e-01
8.56565595e-01 8.54985893e-01 -4.59414870e-01 1.05380523e+00
-5.65208018e-01 5.32206833e-01 -2.97896355e-01 4.33102667e-01
-4.91446350e-03 -1.17031956e+00 8.04172754e-01 8.41344118e-01
3.22687954e-01 5.50457120e-01 1.76271260e-01 1.04337633e+00
1.53174981e-01 -8.30056220e-02 -2.55990952e-01 4.99801517e-01
-1.82097815e-02 1.43174076e+00 -4.34693247e-01 -1.35001639e-04
-5.23513079e-01 7.66323686e-01 -1.07785560e-01 9.11726773e-01
-6.18173420e-01 -4.25063938e-01 7.63030648e-01 3.53416562e-01
1.25701008e-02 -1.76147148e-01 -5.68790376e-01 -1.37743139e+00
2.65241414e-01 -8.81120384e-01 4.11482109e-03 -1.24389720e+00
-1.41340542e+00 6.39888763e-01 -3.27694058e-01 -1.39244032e+00
7.71870375e-01 -4.55067068e-01 -5.84489286e-01 9.80452240e-01
-1.98139191e+00 -8.54395568e-01 -8.08963895e-01 6.23027503e-01
4.45668936e-01 2.20679432e-01 -4.33773585e-02 4.97977078e-01
-9.03697610e-01 1.59353390e-02 3.55414689e-01 -3.59917462e-01
9.11024809e-01 -8.06505561e-01 -1.49276614e-01 1.29482543e+00
-5.35565495e-01 5.40464699e-01 9.60220516e-01 -6.50056541e-01
-1.45814943e+00 -1.10520315e+00 2.37971678e-01 -2.95146741e-02
6.45468175e-01 -2.35642001e-01 -1.17187536e+00 4.04695898e-01
3.60853933e-02 3.44838589e-01 1.47419628e-02 -4.47468996e-01
-7.60853812e-02 -7.40621746e-01 -1.08314490e+00 6.78718567e-01
1.26062858e+00 -2.60252923e-01 -2.11664557e-01 2.24358931e-01
8.38411391e-01 -4.70528930e-01 -5.14976621e-01 5.42679965e-01
1.59980953e-01 -1.30040419e+00 1.10173070e+00 8.15484151e-02
4.63650793e-01 -8.91122341e-01 -6.41869828e-02 -1.41338551e+00
-4.89316523e-01 -5.90498269e-01 1.38775021e-01 1.33199584e+00
-1.06177993e-01 -8.60305011e-01 5.33471227e-01 4.90843892e-01
-1.84472278e-01 -5.66221297e-01 -7.83131123e-01 -5.32058835e-01
-5.34225583e-01 -4.20591414e-01 4.84933197e-01 9.31892335e-01
-1.01571441e+00 2.07865864e-01 -5.69883823e-01 4.75762606e-01
1.30662262e+00 3.11087698e-01 8.86913180e-01 -1.12767863e+00
-3.22352171e-01 -2.65551835e-01 -8.16778541e-02 -1.23538411e+00
-1.02424778e-01 -5.13526618e-01 4.47118640e-01 -1.66449416e+00
3.17254484e-01 -5.27749419e-01 -2.58280456e-01 1.79853886e-01
-4.89816248e-01 3.37540478e-01 1.60892367e-01 2.64485091e-01
-4.27363694e-01 8.40921342e-01 1.78633988e+00 -1.53724849e-01
-2.91820586e-01 -5.12896925e-02 -6.46893620e-01 7.92753935e-01
4.04482991e-01 -3.41459215e-01 -3.83271158e-01 -7.92589307e-01
5.25006711e-01 2.00472642e-02 4.13365632e-01 -9.28058028e-01
2.72412509e-01 -4.87032443e-01 2.83180922e-01 -4.90758210e-01
3.23768258e-01 -1.06104875e+00 2.38103062e-01 1.34613708e-01
9.21497568e-02 -5.30479431e-01 -8.31786823e-03 8.77597928e-01
-3.65392447e-01 -2.75857169e-02 1.17450941e+00 -1.75709203e-01
-5.66686153e-01 5.49978793e-01 -2.09301505e-02 -2.39128191e-02
6.55647695e-01 -3.03221881e-01 -5.89339733e-01 -3.81218314e-01
-4.69945639e-01 1.57012597e-01 6.79557502e-01 -5.74101321e-02
7.61737466e-01 -1.07534766e+00 -8.55268061e-01 5.23038566e-01
4.13043797e-02 1.88975483e-01 7.82982111e-01 1.01489186e+00
-5.55044353e-01 -3.00634295e-01 3.17423381e-02 -6.82606816e-01
-1.11174357e+00 4.94557768e-01 5.71255744e-01 -1.66239858e-01
-9.59174752e-01 7.04020858e-01 7.04419911e-01 -2.18226150e-01
-3.14417202e-03 -3.89481753e-01 -1.12389281e-01 -3.64642799e-01
5.93945682e-01 6.33957982e-01 5.35005666e-02 -5.21636486e-01
1.13953844e-01 9.09953356e-01 2.10529640e-01 3.75507772e-01
1.57287180e+00 -7.66497850e-01 -5.06986320e-01 3.25062543e-01
1.02436066e+00 1.73343167e-01 -1.65110016e+00 -3.75428468e-01
-6.78175330e-01 -1.07093430e+00 6.24637425e-01 -5.27101815e-01
-1.50798571e+00 7.47471213e-01 7.59280801e-01 1.05042979e-01
1.80058300e+00 -5.02893507e-01 9.66069758e-01 1.91609159e-01
3.98977041e-01 -1.05521595e+00 8.50650948e-03 4.05610472e-01
7.46693015e-01 -1.33202922e+00 1.80104777e-01 -7.66096711e-01
-1.90342650e-01 1.03826189e+00 6.82779729e-01 -2.27042232e-02
6.63157284e-01 2.27117449e-01 3.92712682e-01 -1.34993762e-01
-4.02224213e-01 -1.50206462e-01 4.79061063e-03 4.65985507e-01
2.24166308e-02 -2.95532227e-01 -3.93725902e-01 2.89136320e-01
2.67199159e-01 4.47967984e-02 7.27820992e-01 4.30555701e-01
-4.75794882e-01 -7.35684037e-01 -6.14139855e-01 3.98513734e-01
-3.84360969e-01 -2.48025760e-01 2.91895539e-01 3.94930989e-01
3.46363187e-01 1.21070445e+00 -2.50444382e-01 1.17970286e-02
6.19735718e-01 -6.31739497e-01 2.24180728e-01 -3.33759546e-01
-2.61343241e-01 4.99700963e-01 -3.32981229e-01 -7.48673320e-01
-5.60409248e-01 -3.48507762e-01 -8.51252377e-01 -9.21757594e-02
-5.67500830e-01 -2.12230593e-01 4.57622647e-01 6.05310202e-01
6.81999847e-02 5.64497232e-01 7.65951753e-01 -8.56272936e-01
-3.67993504e-01 -8.74649465e-01 -1.12187660e+00 6.76384628e-01
5.57198167e-01 -6.38307631e-01 -9.35747206e-01 -2.21516028e-01] | [10.810745239257812, -2.520984411239624] |
91cf5663-83d4-4282-b248-3369c2db4d56 | optimising-the-input-window-alignment-in-cd | 1606.09163 | null | http://arxiv.org/abs/1606.09163v1 | http://arxiv.org/pdf/1606.09163v1.pdf | Optimising The Input Window Alignment in CD-DNN Based Phoneme Recognition for Low Latency Processing | We present a systematic analysis on the performance of a phonetic recogniser
when the window of input features is not symmetric with respect to the current
frame. The recogniser is based on Context Dependent Deep Neural Networks
(CD-DNNs) and Hidden Markov Models (HMMs). The objective is to reduce the
latency of the system by reducing the number of future feature frames required
to estimate the current output. Our tests performed on the TIMIT database show
that the performance does not degrade when the input window is shifted up to 5
frames in the past compared to common practice (no future frame). This
corresponds to improving the latency by 50 ms in our settings. Our tests also
show that the best results are not obtained with the symmetric window commonly
employed, but with an asymmetric window with eight past and two future context
frames, although this observation should be confirmed on other data sets. The
reduction in latency suggested by our results is critical for specific
applications such as real-time lip synchronisation for tele-presence, but may
also be beneficial in general applications to improve the lag in human-machine
spoken interaction. | ['Akash Kumar Dhaka', 'Giampiero Salvi'] | 2016-06-29 | null | null | null | null | ['low-latency-processing'] | ['robots'] | [ 5.00728488e-01 9.65425372e-02 1.38350070e-01 -3.56833696e-01
-5.87784171e-01 -2.67740488e-01 7.03284979e-01 -8.71027634e-02
-8.95039141e-01 4.71295446e-01 3.37580174e-01 -4.15727854e-01
-3.44430259e-03 -2.34149814e-01 -2.63330817e-01 -8.36856544e-01
-1.98836669e-01 2.78562695e-01 5.20270824e-01 -1.08945608e-01
1.57008037e-01 7.38379836e-01 -2.15090346e+00 4.78839934e-01
6.32946417e-02 7.33253419e-01 4.37318444e-01 1.15814447e+00
6.99963570e-02 3.33380938e-01 -8.54440212e-01 -2.72560418e-02
9.68980715e-02 -5.01789212e-01 -9.91901040e-01 -4.16978784e-02
4.80402224e-02 -4.22620744e-01 -1.01633854e-01 6.48830891e-01
8.72573435e-01 3.27126920e-01 2.94682682e-01 -1.15176022e+00
1.93710133e-01 3.69999945e-01 1.23632282e-01 4.26467419e-01
4.38219815e-01 1.94604963e-01 5.97750545e-01 -5.90471208e-01
5.37874401e-01 1.23356116e+00 6.63370609e-01 6.06726885e-01
-1.02440941e+00 -4.43398595e-01 5.71933426e-02 4.41495925e-01
-1.41150320e+00 -1.13861775e+00 3.39229077e-01 -9.83665437e-02
1.54630637e+00 3.91901970e-01 6.05709434e-01 8.71548533e-01
1.21731728e-01 4.58281308e-01 8.16926003e-01 -8.93432200e-01
1.54085770e-01 3.40917818e-02 -2.61401325e-01 2.11193226e-02
-3.57792288e-01 2.87127882e-01 -6.92333102e-01 -2.28099171e-02
5.80832839e-01 -4.38144326e-01 -3.85523170e-01 1.33391917e-01
-1.05589259e+00 6.40319824e-01 -2.77096272e-01 7.56324291e-01
-6.05239451e-01 -7.26917833e-02 8.51581693e-01 5.38558125e-01
2.90228307e-01 2.25056008e-01 -5.13324082e-01 -1.04344797e+00
-1.11794472e+00 2.26484481e-02 1.07121181e+00 5.82088053e-01
2.74475783e-01 1.28306434e-01 1.59594581e-01 9.88676727e-01
3.17159407e-02 2.66249925e-01 6.60290599e-01 -1.28926313e+00
2.39807546e-01 -1.83116302e-01 1.28359258e-01 -4.87893760e-01
-4.26196635e-01 8.43690783e-02 -3.38220239e-01 2.25557089e-01
6.84398830e-01 -3.77719939e-01 -8.52577269e-01 1.57938910e+00
1.08925119e-01 1.56705901e-01 1.89788058e-01 7.04340458e-01
4.10000145e-01 7.75549233e-01 -5.33773890e-03 -5.68419278e-01
1.32761836e+00 -4.67874497e-01 -1.06898737e+00 -9.07096788e-02
9.09963250e-01 -1.28811216e+00 9.73482847e-01 6.31617486e-01
-9.05921459e-01 -7.16588199e-01 -8.98494899e-01 1.78430825e-01
-1.59642279e-01 -2.85978187e-02 1.55133396e-01 9.21285510e-01
-1.31227350e+00 6.03046417e-01 -8.70502114e-01 -5.92728376e-01
-4.83496696e-01 6.52969897e-01 -3.70864898e-01 3.29443485e-01
-1.33928478e+00 9.17313933e-01 4.37413275e-01 3.38544548e-01
-2.66644567e-01 -8.87014866e-02 -6.39753103e-01 1.53651193e-01
1.39749438e-01 -4.54306304e-02 1.72708225e+00 -1.11337924e+00
-2.17175674e+00 6.75990701e-01 -3.99100900e-01 -6.34680629e-01
4.08562869e-01 -2.86723495e-01 -6.01552069e-01 2.14783087e-01
-5.42051077e-01 8.36456001e-01 7.71159708e-01 -9.13454056e-01
-8.74032617e-01 1.45581454e-01 -9.61915329e-02 2.87205249e-01
2.77898014e-02 4.46729004e-01 -3.29139531e-01 -1.93468839e-01
-2.30753317e-01 -1.28290653e+00 7.70846382e-02 -5.33484221e-01
-1.09801389e-01 -1.66523784e-01 1.11305416e+00 -6.99089110e-01
1.26392472e+00 -2.27232480e+00 -2.33783185e-01 6.64148405e-02
-7.07689106e-01 6.32405996e-01 7.70242438e-02 8.26651752e-01
-2.93226629e-01 -2.56641656e-01 6.68337792e-02 -4.51158345e-01
-1.08126439e-01 5.93696415e-01 -2.36263499e-01 4.20408219e-01
-1.83700562e-01 3.56137693e-01 -4.47748721e-01 -3.91657203e-01
4.25933987e-01 7.04886556e-01 -3.12661558e-01 3.39720726e-01
2.03214854e-01 3.55854303e-01 4.05753851e-01 -4.68446724e-02
3.60407531e-01 3.45246941e-01 2.44416699e-01 2.06966363e-02
-4.83002305e-01 8.95921409e-01 -1.41482699e+00 1.42061341e+00
-7.91144729e-01 1.20076787e+00 1.94930375e-01 -8.09987068e-01
8.44187677e-01 8.91573370e-01 2.69329429e-01 -7.33488560e-01
3.34528871e-02 5.45746610e-02 5.63251078e-01 -6.62050247e-01
6.40836418e-01 -1.23913154e-01 2.91418761e-01 5.06494939e-01
-2.02173367e-01 -1.18274473e-01 1.02310456e-01 -2.54185498e-01
7.25189984e-01 -6.63995370e-02 1.14548132e-01 -1.55936435e-01
5.00891507e-01 -5.33351719e-01 3.70168000e-01 4.21104789e-01
-2.69117653e-01 6.03613675e-01 5.14220834e-01 -1.61856979e-01
-1.13942385e+00 -4.83084172e-01 5.60838822e-03 1.22976530e+00
-2.21156463e-01 -5.54903269e-01 -8.80435050e-01 -1.86586991e-01
-6.27790689e-01 8.25994432e-01 -1.61807999e-01 -5.06566791e-03
-8.03814769e-01 -3.27257454e-01 6.96804285e-01 6.06681287e-01
1.17778793e-01 -1.58968818e+00 -1.16122401e+00 5.11332333e-01
-2.33529702e-01 -1.23245132e+00 -3.34844738e-01 4.93119389e-01
-8.23357463e-01 -6.54588640e-01 -7.70589769e-01 -6.98104918e-01
1.17124826e-01 -1.39539108e-01 8.82260263e-01 2.70481072e-02
1.25720426e-01 3.59349072e-01 -4.65927184e-01 -4.33015436e-01
-8.24615777e-01 2.01269910e-02 1.65165499e-01 -1.41099125e-01
4.31632906e-01 -4.39860404e-01 -7.08750427e-01 3.92250091e-01
-9.18912470e-01 6.30434975e-02 3.98585677e-01 8.15515995e-01
6.30832240e-02 2.61972696e-02 5.38985372e-01 -4.55183357e-01
6.74542069e-01 1.40611187e-01 -3.26755941e-01 -1.31549761e-01
-4.79948014e-01 -2.40415130e-02 4.20638889e-01 -5.89333236e-01
-1.09712195e+00 1.30302504e-01 -8.63035202e-01 -8.08817223e-02
-5.99848986e-01 1.90379620e-01 -1.18453853e-01 2.71894038e-01
3.13358456e-01 1.22120239e-01 1.64931402e-01 -5.44252753e-01
3.16039175e-02 1.21763933e+00 5.75659096e-01 -8.82201120e-02
-1.85807534e-02 2.03219861e-01 -2.89005786e-01 -1.23202479e+00
1.06835701e-01 -6.01999164e-01 -7.71915793e-01 -2.36297980e-01
6.72073543e-01 -6.41284347e-01 -1.12155771e+00 6.27106667e-01
-1.30094719e+00 -6.07061207e-01 -1.96884364e-01 7.34356523e-01
-8.71240973e-01 4.70696300e-01 -8.03992450e-01 -1.13621747e+00
-3.13103169e-01 -1.40807211e+00 1.01620924e+00 7.20698759e-02
-7.67777681e-01 -8.47690225e-01 5.63627947e-03 -1.21984437e-01
3.41540098e-01 -2.28133932e-01 7.16011763e-01 -7.37457156e-01
1.28191695e-01 -9.27776843e-02 1.56622976e-01 6.44946337e-01
2.38913745e-01 2.58158743e-01 -1.29950225e+00 -2.48719588e-01
-2.99611855e-02 1.53754354e-01 4.52318221e-01 4.78321195e-01
7.33170033e-01 -2.38366947e-01 -1.98543191e-01 1.47483572e-02
9.58542645e-01 8.28644335e-01 8.11737061e-01 3.00373137e-01
2.55301684e-01 8.52759957e-01 7.56489336e-01 2.98700899e-01
3.03536630e-03 9.97474372e-01 2.30063260e-01 -1.71634451e-01
-6.85743093e-02 1.32927373e-01 4.95897382e-01 9.07176435e-01
-6.24076203e-02 -4.66257632e-01 -9.91361022e-01 7.16916323e-01
-1.56492984e+00 -1.06215537e+00 -4.26656380e-02 2.54180050e+00
7.48029828e-01 3.83012384e-01 4.02201474e-01 7.33872771e-01
9.85725284e-01 4.55628149e-02 3.82616394e-03 -1.41559720e+00
3.01316321e-01 2.72846192e-01 2.83372372e-01 8.35052013e-01
-7.60257900e-01 6.69239461e-01 6.89121342e+00 7.30064452e-01
-1.61972284e+00 1.07006364e-01 2.99959511e-01 -2.96190262e-01
3.87273759e-01 -6.33512512e-02 -7.90601790e-01 5.60938895e-01
1.86058903e+00 1.06899612e-01 1.60201088e-01 5.85465908e-01
5.96282542e-01 -5.44241190e-01 -1.28069448e+00 9.08467174e-01
-2.90913463e-01 -9.09902215e-01 -4.72090513e-01 3.82892847e-01
6.96618855e-02 -1.15627982e-01 -1.09222502e-01 1.75234213e-01
-4.39789504e-01 -9.03295338e-01 6.17108166e-01 2.08812594e-01
9.02098119e-01 -1.19630814e+00 8.30662012e-01 4.78869319e-01
-1.01469767e+00 1.21202148e-01 -1.22931704e-01 -3.12684596e-01
5.28467059e-01 2.39014775e-01 -1.37543535e+00 1.21298932e-01
6.33461952e-01 1.87861882e-02 3.71265523e-02 1.07506824e+00
8.25926661e-02 7.61637151e-01 -6.10182881e-01 5.05364873e-02
3.14186573e-01 3.56680423e-01 4.45474595e-01 1.68575609e+00
3.06627303e-01 -1.92371532e-02 -3.28988642e-01 -1.27467126e-01
4.36873168e-01 5.41617088e-02 -6.06188238e-01 2.28494689e-01
5.06775856e-01 7.58429229e-01 -7.84040213e-01 -3.61516684e-01
-3.10808688e-01 1.09266627e+00 -3.30029696e-01 2.88058966e-01
-6.20954335e-01 -7.64367938e-01 6.01950228e-01 1.66956142e-01
5.62416792e-01 -2.81955451e-01 2.39059448e-01 -3.62373322e-01
-1.44886777e-01 -9.50496197e-01 1.02490835e-01 -6.43575251e-01
-4.23573107e-01 8.82504165e-01 2.63100535e-01 -1.15850723e+00
-1.10207605e+00 -4.87349600e-01 -5.77091694e-01 1.01161087e+00
-1.09990442e+00 -5.35483241e-01 3.53166759e-01 3.32059503e-01
7.20157981e-01 2.80987024e-01 1.17451227e+00 3.12479019e-01
-1.41140774e-01 6.02039099e-01 -9.09923762e-03 -5.91848567e-02
7.84136713e-01 -9.62483764e-01 5.23149014e-01 7.95152009e-01
1.70083106e-01 5.28752685e-01 1.26262069e+00 -1.81784615e-01
-8.13453853e-01 -3.15082937e-01 1.43304741e+00 -4.00035754e-02
1.81296960e-01 -4.19687748e-01 -1.20020044e+00 3.28391880e-01
6.21011317e-01 -3.05279374e-01 7.15851247e-01 8.22533071e-02
1.73901215e-01 -5.63851520e-02 -8.81935775e-01 4.17493761e-01
6.68310881e-01 -7.45380819e-01 -6.13989770e-01 -7.20961690e-02
5.67297518e-01 -5.18534720e-01 -7.18275368e-01 3.36123526e-01
9.52447593e-01 -1.40389717e+00 5.20836651e-01 -3.90360616e-02
-1.27154127e-01 2.63321921e-02 -2.25909892e-02 -1.22172403e+00
1.53659269e-01 -8.72655690e-01 5.33997640e-02 1.24017143e+00
3.13707530e-01 -6.04347706e-01 7.39777744e-01 6.06787086e-01
-1.16246067e-01 -5.70066929e-01 -1.43401992e+00 -8.62268031e-01
-3.58670890e-01 -8.21749985e-01 4.41062272e-01 4.44389611e-01
-5.42690326e-03 2.07683876e-01 -4.06812221e-01 3.56984362e-02
-2.64024526e-01 -1.98514193e-01 5.83901107e-01 -9.73588824e-01
-3.10377061e-01 -3.28896791e-01 -4.20890510e-01 -1.13213527e+00
6.22368753e-02 -9.57805589e-02 4.86677617e-01 -1.06828678e+00
-4.28111225e-01 -1.54927135e-01 -1.25731722e-01 3.35693747e-01
1.98755562e-01 6.33066148e-02 3.40078890e-01 -3.67362462e-02
-1.40624762e-01 2.12771535e-01 7.31682479e-01 3.53918433e-01
-4.88113850e-01 4.66633856e-01 1.89303994e-01 7.45976985e-01
6.56977057e-01 -4.90724027e-01 -4.16084558e-01 -6.18356541e-02
-1.00812785e-01 2.21842885e-01 8.23097378e-02 -1.14401042e+00
3.47044945e-01 1.22034848e-01 5.62019162e-02 -6.87619150e-01
8.31379116e-01 -9.96780574e-01 3.16427171e-01 6.21514976e-01
-3.18583757e-01 4.37259257e-01 6.05084538e-01 1.37712225e-01
-2.56354153e-01 -4.41926360e-01 7.00135529e-01 1.76726639e-01
-8.73672962e-01 -4.45855945e-01 -9.70831037e-01 -4.39632028e-01
6.55390799e-01 -6.03936076e-01 2.47734070e-01 -8.41907918e-01
-1.03864408e+00 -4.40210968e-01 2.84221858e-01 5.02429068e-01
4.47676033e-01 -1.06667995e+00 -2.76142150e-01 4.22460079e-01
-1.18929550e-01 -4.57425982e-01 2.00217530e-01 9.27943289e-01
-5.20666778e-01 8.38185012e-01 -1.51882783e-01 -7.82495141e-01
-2.04749203e+00 2.55867392e-01 3.85670543e-01 -7.54669607e-02
-6.62536323e-01 7.59934843e-01 -1.61096364e-01 1.69354722e-01
6.62739038e-01 -4.48038846e-01 -2.17392430e-01 2.47632816e-01
5.73884904e-01 3.55386764e-01 4.51212227e-01 -8.66089821e-01
-4.06983674e-01 2.28878245e-01 -8.62653553e-02 -6.14162862e-01
1.12563980e+00 -4.80268985e-01 3.10891598e-01 8.97058845e-01
1.29735553e+00 -1.50091341e-03 -1.27948940e+00 -1.40082999e-03
1.52480423e-01 -3.47394913e-01 7.24229356e-03 -5.25437713e-01
-6.43497646e-01 1.02335989e+00 9.22838151e-01 5.14560282e-01
1.29070580e+00 -3.19263071e-01 8.27764034e-01 3.06788445e-01
1.95919782e-01 -1.22915387e+00 -4.16485667e-01 7.40667999e-01
7.81026006e-01 -6.96538329e-01 -4.59345877e-01 -2.39827018e-02
-6.92038953e-01 1.30794799e+00 2.41431743e-01 3.81452590e-01
5.21238506e-01 5.56944549e-01 4.44017023e-01 1.85574248e-01
-1.01188564e+00 -3.82372081e-01 -7.89554231e-03 6.80960536e-01
7.20961928e-01 -9.91990268e-02 -3.27000827e-01 -1.19117118e-01
-3.34032506e-01 -1.13983691e-01 5.87507963e-01 9.38017905e-01
-3.87378961e-01 -1.33235395e+00 -5.25370121e-01 -4.08030115e-02
-7.49498308e-01 -1.27029270e-01 -1.16690435e-01 9.95484769e-01
1.09338522e-01 1.23385060e+00 4.21240211e-01 -2.34817162e-01
3.81819636e-01 6.48776948e-01 3.12399477e-01 -5.86328983e-01
-7.10932314e-01 5.26526988e-01 3.82221341e-01 -4.54878241e-01
-5.34694254e-01 -7.74041414e-01 -1.27841997e+00 -3.84136230e-01
-4.35312331e-01 3.66601557e-01 9.67450321e-01 1.03591633e+00
3.35234642e-01 4.28431511e-01 3.75305027e-01 -1.17127407e+00
-1.08314253e-01 -1.11578572e+00 -3.71433645e-01 1.12601854e-01
5.99352419e-01 -2.88305700e-01 -4.04494196e-01 3.40534091e-01] | [14.684435844421387, 5.927675247192383] |
de54d70a-d5f2-4bd3-8525-34900e4221d8 | universal-deep-network-for-steganalysis-of | 2111.12231 | null | https://arxiv.org/abs/2111.12231v1 | https://arxiv.org/pdf/2111.12231v1.pdf | Universal Deep Network for Steganalysis of Color Image based on Channel Representation | Up to now, most existing steganalytic methods are designed for grayscale images, and they are not suitable for color images that are widely used in current social networks. In this paper, we design a universal color image steganalysis network (called UCNet) in spatial and JPEG domains. The proposed method includes preprocessing, convolutional, and classification modules. To preserve the steganographic artifacts in each color channel, in preprocessing module, we firstly separate the input image into three channels according to the corresponding embedding spaces (i.e. RGB for spatial steganography and YCbCr for JPEG steganography), and then extract the image residuals with 62 fixed high-pass filters, finally concatenate all truncated residuals for subsequent analysis rather than adding them together with normal convolution like existing CNN-based steganalyzers. To accelerate the network convergence and effectively reduce the number of parameters, in convolutional module, we carefully design three types of layers with different shortcut connections and group convolution structures to further learn high-level steganalytic features. In classification module, we employ a global average pooling and fully connected layer for classification. We conduct extensive experiments on ALASKA II to demonstrate that the proposed method can achieve state-of-the-art results compared with the modern CNN-based steganalyzers (e.g., SRNet and J-YeNet) in both spatial and JPEG domains, while keeping relatively few memory requirements and training time. Furthermore, we also provide necessary descriptions and many ablation experiments to verify the rationality of the network design. | ['Jiwu Huang', 'Shunquan Tan', 'Weiqi Luo', 'Kangkang Wei'] | 2021-11-24 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 5.97551167e-01 -2.90324688e-01 9.76449996e-02 7.49469176e-02
-5.67395822e-04 -1.93897456e-01 2.37111241e-01 -6.91826165e-01
-5.00987411e-01 4.20356423e-01 -3.50976169e-01 -6.49017453e-01
3.54765892e-01 -1.24108112e+00 -4.38465238e-01 -9.53709424e-01
-8.21161717e-02 -5.81248999e-01 5.80441594e-01 -5.03611386e-01
3.38570386e-01 2.07656279e-01 -1.26024663e+00 2.69024432e-01
9.13346291e-01 1.10431063e+00 9.73394588e-02 7.28595018e-01
2.12852105e-01 8.52801025e-01 -5.46637177e-01 -4.02050495e-01
4.33111519e-01 -1.03907812e+00 -3.78646284e-01 1.50148138e-01
-2.77770787e-01 -4.86929566e-01 -9.78380680e-01 1.46735477e+00
4.06545520e-01 -1.56082347e-01 2.68596768e-01 -9.63084042e-01
-8.80226016e-01 5.76098442e-01 -5.69682479e-01 1.23919621e-01
-1.80792257e-01 5.16440690e-01 5.25663555e-01 -5.26144981e-01
3.95558059e-01 1.04719305e+00 6.45909786e-01 5.00919640e-01
-6.84843779e-01 -1.29354239e+00 -2.89113849e-01 4.04702812e-01
-1.32838345e+00 -3.01804721e-01 9.06733334e-01 2.10904539e-01
6.53139770e-01 3.84344071e-01 9.58361626e-01 7.55646944e-01
3.66177291e-01 4.93477821e-01 1.26822269e+00 -3.25331956e-01
-2.78503627e-01 -2.12187156e-01 -4.50226754e-01 1.09339261e+00
4.52749908e-01 2.42397815e-01 -8.83786660e-03 3.11345309e-01
1.24101591e+00 3.15441102e-01 -5.25803983e-01 -2.55380385e-02
-1.33824337e+00 9.62089539e-01 8.12711358e-01 5.21648228e-01
-7.32096285e-02 5.26711941e-01 2.49715999e-01 7.15823114e-01
2.20115542e-01 1.06938541e-01 5.45760281e-02 3.17679644e-01
-9.01230812e-01 -3.18096370e-01 7.75993764e-01 9.34030950e-01
8.95808101e-01 3.84757668e-01 2.07312956e-01 5.48116982e-01
3.68562847e-01 5.14002860e-01 7.12247133e-01 -5.10623753e-01
4.68166322e-01 5.24960577e-01 -4.97040570e-01 -1.60467196e+00
-4.31545191e-02 -3.19346666e-01 -1.57976377e+00 2.22539574e-01
8.99329316e-03 -7.99658149e-02 -1.38528955e+00 1.14261830e+00
-1.91790283e-01 4.53093618e-01 3.41923416e-01 8.69068861e-01
8.31162512e-01 7.99454749e-01 -1.25546262e-01 7.77466223e-03
1.39401650e+00 -1.07853878e+00 -5.62676370e-01 -3.70858580e-01
5.56192577e-01 -8.03398550e-01 7.28802443e-01 2.82737672e-01
-9.57525790e-01 -6.27599180e-01 -1.57505345e+00 -9.77916047e-02
-5.54594755e-01 9.21314433e-02 6.43164396e-01 1.06424785e+00
-9.51853395e-01 8.93834054e-01 -7.19830275e-01 -3.68295312e-02
3.59480441e-01 4.92971778e-01 -2.59206176e-01 -2.07753554e-01
-1.70928252e+00 5.07857621e-01 8.31101477e-01 4.06944662e-01
-8.93557787e-01 5.76684922e-02 -8.96901309e-01 3.85151297e-01
1.98811688e-03 -1.07408300e-01 6.05504394e-01 -1.26980853e+00
-1.65870523e+00 6.48138225e-01 4.48619753e-01 -3.69058818e-01
4.53900725e-01 6.23615384e-01 -8.79991829e-01 3.30258995e-01
-3.95067930e-01 5.28348267e-01 1.17665625e+00 -8.87435794e-01
-7.58646250e-01 1.12792470e-01 -9.32434201e-02 -7.21184835e-02
-7.17031062e-01 3.77674624e-02 -7.88136423e-01 -8.29152048e-01
5.54403603e-01 -8.44079733e-01 -3.08011562e-01 -1.58617780e-01
-3.91826838e-01 3.67328286e-01 1.30184746e+00 -1.02157331e+00
1.43765497e+00 -2.34836674e+00 -2.46279091e-01 6.70354247e-01
3.29560101e-01 6.42720759e-01 -1.62147298e-01 6.60174787e-02
-2.68910944e-01 3.64276081e-01 -4.93640512e-01 6.62949011e-02
-3.51969481e-01 -1.32516427e-02 2.91854143e-02 8.59407961e-01
9.61936787e-02 9.64650333e-01 -7.37850904e-01 -6.42024219e-01
2.69143164e-01 4.87169176e-01 -4.28268820e-01 -1.29272237e-01
2.06433669e-01 2.58742452e-01 -4.79837865e-01 5.85526824e-01
9.16693628e-01 -4.61178482e-01 4.80594188e-01 -1.67017683e-01
-7.10711479e-02 -5.68316597e-03 -9.25867260e-01 1.14709878e+00
-3.74972373e-01 8.12978506e-01 1.12824729e-02 -1.10147071e+00
1.15840757e+00 3.89436156e-01 2.27256924e-01 -7.59694755e-01
5.58647394e-01 5.19369960e-01 2.18220130e-01 -4.73689407e-01
4.62663680e-01 -6.64047971e-02 -1.03014708e-01 4.31746304e-01
-9.41149592e-02 -2.04796065e-02 -3.07751670e-02 -7.78683499e-02
1.02659607e+00 -2.79918969e-01 1.22742213e-01 -6.52777851e-02
8.94439399e-01 -1.08826101e-01 4.52829391e-01 4.56928492e-01
-4.07015756e-02 6.49292469e-01 5.09444952e-01 -3.43391538e-01
-1.28427505e+00 -2.56327182e-01 4.02890354e-01 6.16300166e-01
6.47159576e-01 -8.18951463e-04 -6.74574733e-01 -6.22269869e-01
-4.12799567e-01 3.65584791e-02 -4.57313031e-01 -5.62521040e-01
-9.04745400e-01 -7.51187563e-01 9.65845883e-01 1.15543850e-01
1.73171401e+00 -1.39030755e+00 -5.96591890e-01 2.08974943e-01
-1.56414971e-01 -8.65462124e-01 -5.81872225e-01 -8.49239454e-02
-8.21490407e-01 -1.29444218e+00 -8.74643564e-01 -1.42037189e+00
6.72236025e-01 5.98439455e-01 5.16063869e-01 7.95722425e-01
-1.24327421e-01 -5.26258767e-01 -5.56269467e-01 2.09036544e-01
-4.95338112e-01 1.81580633e-01 -5.22349477e-01 7.87253156e-02
1.84525445e-01 -4.70496118e-01 -9.42875087e-01 4.08069491e-01
-1.26638734e+00 2.51635820e-01 1.01899612e+00 1.04754674e+00
4.37552221e-02 7.51522481e-01 -1.93452150e-01 -8.81537616e-01
5.12887418e-01 -2.71480531e-01 -6.09221339e-01 2.33828902e-01
-6.84408784e-01 -1.77329823e-01 7.97139049e-01 -4.12286043e-01
-6.76963985e-01 -2.63971210e-01 -2.78461073e-02 -4.98202384e-01
3.52166802e-01 5.00868320e-01 -3.22396718e-02 -7.95147955e-01
2.32649609e-01 8.01788270e-01 2.63253361e-01 -2.01941594e-01
-3.27704363e-02 8.50526750e-01 5.72169363e-01 2.79734612e-01
1.20490301e+00 5.34995258e-01 -2.10799929e-02 -6.42477214e-01
-4.48679365e-02 2.55933206e-04 -9.56943035e-02 3.42011489e-02
9.85149920e-01 -9.01778042e-01 -9.00166571e-01 1.06044960e+00
-9.97426391e-01 -1.93646789e-01 3.24536681e-01 4.01236653e-01
-1.45655289e-01 7.40184665e-01 -9.50484693e-01 -3.63129050e-01
-4.63635623e-01 -1.22171640e+00 3.01587522e-01 3.97117049e-01
6.49349391e-01 -9.53071356e-01 -4.97455597e-01 -1.50448665e-01
6.72274470e-01 4.06245053e-01 7.30318964e-01 -3.89861315e-01
-7.41380095e-01 -3.49412382e-01 -6.17188513e-01 6.89530373e-01
1.21613503e-01 -1.24636896e-01 -5.80625713e-01 -6.81670606e-01
1.63131520e-01 -3.72703783e-02 1.38036871e+00 4.16826718e-02
1.44281816e+00 -4.99195725e-01 -2.19337240e-01 1.25869226e+00
1.64953649e+00 5.17808497e-01 1.29371059e+00 6.09026670e-01
7.29983270e-01 3.48667145e-01 1.39922231e-01 1.93420425e-01
5.26339263e-02 -2.50234772e-02 4.91472036e-01 -5.22111952e-01
-1.28324404e-01 -1.89121798e-01 3.56746078e-01 9.65987802e-01
-3.14578861e-01 -4.76180285e-01 -4.84477907e-01 2.38745883e-01
-1.38619089e+00 -9.56755936e-01 -2.82265805e-02 1.78522491e+00
6.28560603e-01 2.06986174e-01 -4.29879338e-01 4.05709028e-01
1.13988745e+00 6.62363350e-01 -4.12577271e-01 -1.14691310e-01
-3.23361516e-01 5.25303543e-01 1.19476891e+00 1.17883772e-01
-1.28091943e+00 1.10768628e+00 5.68484020e+00 1.10911918e+00
-1.49201107e+00 1.83686674e-01 7.54165769e-01 5.04076898e-01
-1.85321227e-01 1.83464050e-01 -1.81188762e-01 6.17384851e-01
6.00746155e-01 4.06680316e-01 8.24515760e-01 5.21543324e-01
-7.55706429e-02 3.01806033e-01 -2.14945242e-01 1.01841009e+00
8.24631974e-02 -1.35506320e+00 -1.48879915e-01 1.69875398e-01
8.67820740e-01 -3.09209943e-01 3.12579185e-01 1.18809268e-01
3.03115100e-01 -1.15537739e+00 4.64371264e-01 2.86098756e-02
1.38405204e+00 -7.73020267e-01 8.27465117e-01 -1.43879160e-01
-1.24158990e+00 -2.99915165e-01 -6.51713431e-01 1.87857136e-01
-2.09897935e-01 2.43549481e-01 -1.35582164e-01 6.02619588e-01
7.69303858e-01 9.08973157e-01 -3.76819104e-01 8.33329737e-01
-4.62114245e-01 6.66705251e-01 -1.30980670e-01 -5.30445129e-02
7.00705647e-01 -1.19969234e-01 1.41473308e-01 1.02659810e+00
7.27812350e-01 3.53980422e-01 -2.16711342e-01 6.30847275e-01
-3.19762588e-01 -2.87238032e-01 -3.33733946e-01 -2.46729597e-01
3.97217333e-01 1.21219802e+00 -9.62736189e-01 -5.21077871e-01
-3.72737736e-01 1.25814652e+00 -4.12653178e-01 3.70958179e-01
-8.81838500e-01 -1.27977586e+00 3.00063550e-01 -2.76338756e-01
7.97035992e-01 -2.07131997e-01 -2.20031589e-01 -1.30324090e+00
-4.07785207e-01 -1.15300846e+00 2.23066330e-01 -4.38322604e-01
-5.64405382e-01 5.76191962e-01 -5.49034119e-01 -1.61482131e+00
1.64793968e-01 -6.57243907e-01 -8.73916864e-01 8.61914814e-01
-1.89475513e+00 -1.24487007e+00 -6.48541868e-01 8.67445886e-01
1.09710962e-01 -3.33531052e-01 3.55777830e-01 3.16452086e-01
-7.17949450e-01 6.04087770e-01 4.54990923e-01 6.17559314e-01
4.80099231e-01 -5.19341230e-01 7.98333585e-01 1.22513497e+00
-4.80796933e-01 5.13596654e-01 2.90752679e-01 -7.23864853e-01
-1.22998011e+00 -1.09287310e+00 5.96530437e-01 6.86050355e-01
3.83933067e-01 -2.08161712e-01 -8.09567511e-01 5.65134525e-01
2.38071680e-01 2.10234955e-01 9.07768160e-02 -8.39302599e-01
-3.52221638e-01 1.54738063e-02 -1.31692529e+00 5.45044899e-01
9.46731329e-01 -3.33349198e-01 7.98104331e-02 1.45882238e-02
7.62072384e-01 -4.20379877e-01 -5.57647705e-01 2.75482029e-01
5.98946393e-01 -1.12840116e+00 8.94132078e-01 -2.13355213e-01
7.43907213e-01 -3.31277609e-01 1.22441471e-01 -1.05416656e+00
-6.57272696e-01 -8.13523769e-01 2.26501882e-01 6.51967943e-01
1.15319774e-01 -1.00463521e+00 1.00694060e+00 -2.10537940e-01
-6.34004250e-02 -5.32754540e-01 -6.47387683e-01 -5.75482607e-01
-2.54015595e-01 2.93343943e-02 9.46237803e-01 9.73989487e-01
-1.99794680e-01 -3.65961432e-01 -8.53701293e-01 1.87578455e-01
6.41735971e-01 1.01622045e-01 4.71893221e-01 -7.41879225e-01
-1.16957471e-01 -4.94288743e-01 -5.38360298e-01 -9.68067706e-01
-2.13987544e-01 -6.25615537e-01 5.85222505e-02 -9.96497869e-01
-1.35574490e-01 -6.01021945e-01 -4.54085857e-01 4.98124927e-01
-9.13056582e-02 9.19752419e-01 1.37707680e-01 4.83078152e-01
-1.85731754e-01 3.86845708e-01 1.66412961e+00 -2.28116795e-01
8.94251745e-03 -3.16039354e-01 -6.76789522e-01 5.96205473e-01
8.89861882e-01 -3.87864262e-01 -6.71925917e-02 -4.43681508e-01
-4.56245579e-02 -7.97244813e-03 4.71961468e-01 -1.29487681e+00
3.25172693e-01 -1.34933159e-01 7.11933494e-01 -1.83621958e-01
8.53476003e-02 -8.80952477e-01 2.68817514e-01 1.30342662e+00
-9.53212529e-02 -4.14354317e-02 -3.04255337e-01 4.74316627e-01
-3.81258637e-01 -8.27102736e-02 1.09382248e+00 -5.10525167e-01
-9.68791723e-01 4.52815831e-01 -3.46032947e-01 -4.75876004e-01
1.00171971e+00 -5.83344460e-01 -4.73628432e-01 -4.46793735e-01
-2.70708263e-01 2.60481127e-02 5.64614236e-01 1.91326693e-01
1.07818961e+00 -1.30678260e+00 -5.41271985e-01 7.84159839e-01
-2.26520896e-01 -3.85016054e-01 4.18424815e-01 6.86280012e-01
-1.42163801e+00 4.10783708e-01 -5.82690418e-01 5.53880744e-02
-1.17187262e+00 4.93367702e-01 4.65063989e-01 -4.20862794e-01
-6.09134674e-01 7.41276562e-01 -4.99970876e-02 -6.72083870e-02
-4.31093834e-02 -1.80094734e-01 -3.11705619e-01 -3.89104187e-01
3.62828344e-01 4.56128508e-01 -2.29646921e-01 -5.43658078e-01
-9.59326699e-02 5.17571807e-01 4.71125133e-02 6.19516708e-02
1.25997114e+00 -1.74430087e-01 -4.51223642e-01 -5.34567416e-01
1.51293576e+00 -4.19494241e-01 -1.05861712e+00 -2.02605680e-01
-4.75079119e-01 -5.91745734e-01 1.80597052e-01 -1.96923718e-01
-1.88683593e+00 9.82447684e-01 7.52442241e-01 4.33363199e-01
1.47259188e+00 -5.81977129e-01 1.37470067e+00 2.26286471e-01
1.09966584e-01 -8.88017833e-01 1.34893283e-01 3.70864421e-01
3.03753197e-01 -9.32963133e-01 7.29826391e-02 -4.63013858e-01
-3.66772205e-01 1.35083902e+00 5.85585892e-01 -6.16744280e-01
6.35143816e-01 -1.47512570e-01 -2.64028851e-02 -2.29942694e-01
-2.04215765e-01 -8.04044828e-02 -5.94633073e-02 4.19211745e-01
4.12503630e-02 -1.46481454e-01 -5.33658743e-01 4.01541367e-02
-1.73101440e-01 -7.96386600e-02 6.31219268e-01 9.12234187e-01
-6.71785951e-01 -9.32939529e-01 -4.56182301e-01 4.27902162e-01
-6.84019148e-01 -3.79282981e-01 3.85486856e-02 7.82232702e-01
3.53429645e-01 9.32735622e-01 8.73177722e-02 -9.62578118e-01
-4.84702513e-02 -4.69337523e-01 1.32738784e-01 -2.11201772e-01
-6.71812177e-01 9.42331478e-02 -2.49199137e-01 -4.05245751e-01
-3.94021034e-01 -5.60554415e-02 -9.03191209e-01 -8.80033672e-01
-5.38682759e-01 -1.40306270e-02 5.51957071e-01 6.73017800e-01
-4.88379970e-02 6.22184217e-01 1.05478811e+00 -7.36190736e-01
-4.61998492e-01 -9.63007808e-01 -7.36252189e-01 1.97625518e-01
4.19456899e-01 -9.77035239e-03 -5.10021448e-01 2.32977737e-02] | [4.295793056488037, 8.054793357849121] |
27a39bc8-7cb1-4ecc-bcc3-ec9453183c56 | energy-efficient-downlink-semantic-generative | 2306.05041 | null | https://arxiv.org/abs/2306.05041v1 | https://arxiv.org/pdf/2306.05041v1.pdf | Energy-Efficient Downlink Semantic Generative Communication with Text-to-Image Generators | In this paper, we introduce a novel semantic generative communication (SGC) framework, where generative users leverage text-to-image (T2I) generators to create images locally from downloaded text prompts, while non-generative users directly download images from a base station (BS). Although generative users help reduce downlink transmission energy at the BS, they consume additional energy for image generation and for uploading their generator state information (GSI). We formulate the problem of minimizing the total energy consumption of the BS and the users, and devise a generative user selection algorithm. Simulation results corroborate that our proposed algorithm reduces total energy by up to 54% compared to a baseline with all non-generative users. | ['Jinho Choi', 'Sooyoung Kim', 'Jihong Park', 'Hyein Lee'] | 2023-06-08 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [ 5.37022114e-01 5.37533581e-01 -9.09070745e-02 1.70687847e-02
-9.43060696e-01 -7.38777637e-01 5.45383275e-01 -4.42538559e-01
1.14333797e-02 7.36191809e-01 1.20388687e-01 -3.88387650e-01
4.75467086e-01 -1.06614399e+00 -7.24056780e-01 -1.05406499e+00
8.53660181e-02 2.42944807e-01 -1.04018450e-01 2.88392574e-01
-1.25656441e-01 -8.28879699e-02 -8.79459858e-01 -2.51851141e-01
8.29580843e-01 8.27136576e-01 6.89851403e-01 1.00371909e+00
1.86444461e-01 5.76456249e-01 -1.03247380e+00 -3.18291754e-01
3.62767316e-02 -1.00944078e+00 -5.27538121e-01 3.17555606e-01
-4.06104028e-01 -7.40448952e-01 -7.38571346e-01 8.41776967e-01
9.32023048e-01 -1.51750490e-01 5.40522158e-01 -1.69732499e+00
-8.62520218e-01 8.32762420e-01 -5.32956362e-01 1.24317497e-01
4.08430636e-01 2.23187566e-01 4.34033662e-01 -3.61223787e-01
4.56223786e-01 9.89657581e-01 1.09754108e-01 8.16732347e-01
-9.55024242e-01 -8.90061557e-01 -1.27543330e-01 -3.16299163e-02
-1.41378403e+00 -8.09050739e-01 6.66731000e-01 2.38306746e-01
6.07651472e-01 6.62505925e-01 8.13730121e-01 1.26589823e+00
-8.77365768e-02 9.03941214e-01 8.86728525e-01 -4.93212253e-01
4.28223550e-01 1.14300221e-01 -5.06408036e-01 5.94207287e-01
2.97733814e-01 -2.76050001e-01 -8.12071085e-01 -2.31745377e-01
8.45925152e-01 -3.57368916e-01 -3.25493187e-01 2.95018464e-01
-1.13709140e+00 6.92979872e-01 8.50774199e-02 -3.14179435e-02
-6.66192591e-01 8.86095345e-01 -3.89283597e-01 -4.64597344e-02
5.83834231e-01 -2.95655936e-01 5.24414361e-01 -2.51108557e-01
-8.73064876e-01 -1.61987603e-01 9.25411463e-01 1.83372784e+00
6.56396270e-01 3.01024675e-01 -7.27362454e-01 3.97002161e-01
5.57384849e-01 1.46512651e+00 3.96072596e-01 -8.67281914e-01
3.91934305e-01 1.60374641e-02 -4.27990668e-02 -7.56447971e-01
1.12805165e-01 -5.40207982e-01 -1.01167810e+00 -7.11436212e-01
-3.85168672e-01 -8.94022524e-01 -8.28763366e-01 1.69085002e+00
1.95234194e-02 3.78266782e-01 1.26798555e-01 8.38548958e-01
7.50696421e-01 1.11895096e+00 2.72844713e-02 -4.13548738e-01
1.03332698e+00 -8.09929371e-01 -5.96526861e-01 -2.04602286e-01
-6.19099513e-02 -6.41373396e-01 5.62918007e-01 6.02796813e-03
-1.51472425e+00 -1.93410337e-01 -7.79852927e-01 3.45164210e-01
2.39499480e-01 3.86987865e-01 3.34343851e-01 1.12343097e+00
-1.67006123e+00 -1.91748008e-01 -8.09930325e-01 -6.22925937e-01
4.71888840e-01 4.35854346e-01 6.14679337e-01 2.38459930e-01
-6.79361820e-01 -1.25375807e-01 1.32054716e-01 -2.76909858e-01
-1.26238072e+00 -4.84077185e-01 -3.51216823e-01 3.52972209e-01
3.10375988e-01 -1.26593041e+00 1.22715640e+00 -4.70204413e-01
-1.64954007e+00 3.13135982e-01 -4.81085151e-01 -3.73654574e-01
4.32276666e-01 2.64723420e-01 -1.24534555e-01 3.51416707e-01
2.42438242e-01 9.32922482e-01 8.76893580e-01 -1.68729508e+00
-5.57999134e-01 4.88630794e-02 1.09577104e-01 4.19449717e-01
-5.96379161e-01 -4.17126358e-01 -1.18951738e+00 -7.63731241e-01
1.19233178e-02 -1.48376608e+00 -1.53081357e-01 -5.39360523e-01
-1.30661500e+00 -8.45993534e-02 1.03635108e+00 -2.78811365e-01
1.21895325e+00 -2.08368468e+00 -1.17447600e-01 4.27291989e-01
2.69402981e-01 -1.16829157e-01 -6.53839484e-02 4.29027379e-01
5.00230491e-01 5.47041416e-01 2.54391491e-01 -6.55046165e-01
-1.40443414e-01 2.55897969e-01 -2.61450678e-01 -3.63215581e-02
-4.42470223e-01 1.08776224e+00 -9.10926580e-01 -4.65739191e-01
2.34390497e-01 4.52497393e-01 -4.36190784e-01 3.17202389e-01
-5.15672341e-02 4.65791374e-01 -8.39028478e-01 5.80774903e-01
6.71648324e-01 -8.06331158e-01 4.93718326e-01 -7.01120347e-02
3.65096629e-01 1.19368032e-01 -6.75628722e-01 1.33848393e+00
-7.49712467e-01 6.50490522e-01 1.46227345e-01 -6.19287729e-01
3.22337776e-01 4.59836662e-01 5.82513034e-01 -6.88994825e-01
2.21332163e-01 -1.79948375e-01 -6.55083060e-01 -1.36453822e-01
5.21376729e-01 1.93292528e-01 -2.71292299e-01 9.85879719e-01
-7.07248375e-02 1.94030866e-01 1.81712314e-01 1.03659451e+00
1.06062460e+00 -3.03662479e-01 -6.12340719e-02 -1.11648440e-01
1.86736230e-02 -3.41596305e-01 8.80920291e-02 1.05420744e+00
2.59954214e-01 2.66779989e-01 1.96113855e-01 5.32656848e-01
-8.69559050e-01 -1.31851542e+00 6.05030119e-01 9.92727637e-01
6.16882980e-01 -6.46781027e-01 -1.24689305e+00 -4.64558721e-01
-3.49660844e-01 1.02849674e+00 -8.14834610e-02 -2.08247751e-01
-1.22104280e-01 -8.61965895e-01 3.40872020e-01 2.25508541e-01
8.74946177e-01 -6.19626284e-01 -6.31995261e-01 3.09785098e-01
-9.54704463e-01 -1.27187943e+00 -9.31404591e-01 -2.51743078e-01
-4.50223029e-01 -5.24400234e-01 -7.75513232e-01 -2.99788684e-01
1.09544122e+00 9.11976993e-01 9.30818200e-01 5.41832969e-02
-1.98726952e-01 9.64604020e-01 -4.73944396e-01 -6.13733351e-01
-7.23360896e-01 3.15009296e-01 -3.10756683e-01 -3.22402753e-02
-3.36758584e-01 -5.27759433e-01 -1.01469469e+00 2.23103523e-01
-8.70441318e-01 6.49963915e-01 5.51623523e-01 1.78413447e-02
5.40902317e-01 2.61456817e-01 6.19663656e-01 -5.69795668e-01
8.07726622e-01 -7.99440503e-01 -3.21896821e-01 3.66640717e-01
-6.81214690e-01 -1.38068438e-01 3.93151909e-01 -1.25825424e-02
-1.13767362e+00 1.05464771e-01 1.20370880e-01 -5.45899868e-02
2.26345822e-01 -8.32130834e-02 -3.16853583e-01 -4.03007828e-02
1.74510777e-01 8.17966819e-01 -4.12280828e-01 3.54106128e-02
4.31034505e-01 9.05525029e-01 5.44729054e-01 -2.44597808e-01
1.00307286e+00 4.37582254e-01 7.73027912e-02 -1.01242650e+00
-4.08476830e-01 -3.27882264e-03 1.16776340e-01 -6.16745234e-01
8.48723233e-01 -1.39202821e+00 -7.47727573e-01 4.70936149e-01
-1.10806084e+00 -4.89886940e-01 -2.53886320e-02 -4.63347591e-04
-4.52977985e-01 3.31452370e-01 -2.62468308e-01 -1.10709989e+00
-7.57971585e-01 -9.74603713e-01 1.41984808e+00 6.45358741e-01
1.26386300e-01 -8.93049121e-01 -5.63860238e-01 4.74601477e-01
7.85242856e-01 1.47203326e-01 4.96222436e-01 1.91021792e-03
-1.32108641e+00 1.71270698e-01 -2.25745201e-01 -1.56321615e-01
3.30000520e-01 -4.44328845e-01 -8.80483925e-01 -5.14480472e-01
-4.03802693e-01 4.73655835e-02 5.12521684e-01 5.96694946e-01
1.58269393e+00 -7.63615549e-01 -8.68251801e-01 5.74304760e-01
1.45413649e+00 3.73789072e-01 6.25652313e-01 -3.86644989e-01
5.61651170e-01 -3.59981447e-01 -6.23980090e-02 1.00175476e+00
8.15014541e-01 5.57972193e-01 3.69241267e-01 -3.09157431e-01
-4.60743338e-01 -4.17025030e-01 5.41659236e-01 6.60714567e-01
6.22518733e-02 -1.63268268e+00 -3.58593792e-01 4.11962003e-01
-1.62920320e+00 -6.82320237e-01 -1.84811816e-01 2.12494779e+00
7.84227490e-01 -1.36929765e-01 8.54032263e-02 -1.48269609e-01
7.26816297e-01 -3.97312548e-03 -6.42592728e-01 2.81627744e-01
-5.28656878e-02 2.17798844e-01 1.08698714e+00 1.86892539e-01
-7.27240920e-01 7.92665660e-01 6.59917784e+00 1.12751281e+00
-9.58588898e-01 4.80238825e-01 1.08197069e+00 -6.39710605e-01
-6.55081332e-01 -1.88362479e-01 -7.35206366e-01 1.07934582e+00
1.34439313e+00 -7.38013506e-01 7.39929676e-01 5.49529791e-01
7.40706325e-01 -5.54917276e-01 -6.69125378e-01 1.07060039e+00
2.47719333e-01 -1.40832388e+00 2.01878369e-01 4.53460068e-01
7.93794274e-01 -8.87574181e-02 4.89680171e-02 -1.16128460e-01
7.62736857e-01 -5.35000563e-01 6.37885273e-01 5.03708422e-01
1.02710021e+00 -7.39550471e-01 4.17514667e-02 3.22112799e-01
-1.23794210e+00 7.70269101e-03 -1.33533612e-01 6.22054517e-01
7.00081229e-01 9.42956984e-01 -9.18273926e-01 6.55105174e-01
4.41218466e-01 -1.06701097e-02 -5.11881530e-01 5.23330927e-01
-3.66858333e-01 1.13665497e+00 -6.20379031e-01 -2.90376633e-01
8.08089450e-02 -1.55328169e-01 6.63196683e-01 9.35196996e-01
1.00614989e+00 5.51987231e-01 1.46809116e-01 1.02835655e+00
-5.54062545e-01 -3.14879268e-01 -2.57050395e-01 -2.37634983e-02
9.14925754e-01 1.35881138e+00 -1.05087006e+00 -6.94752634e-01
9.90344435e-02 1.80328155e+00 -6.02372944e-01 5.75917006e-01
-1.14764464e+00 -3.09506804e-01 2.40604579e-01 1.35256007e-01
-3.82225476e-02 -2.89270610e-01 -1.47780001e-01 -7.42570877e-01
-2.91730482e-02 -8.84714201e-02 -1.19708985e-01 -1.15845823e+00
-6.46244466e-01 2.30854198e-01 -2.80993730e-02 -7.22422183e-01
-3.56450617e-01 4.59966779e-01 -9.11941111e-01 7.87455261e-01
-1.39499712e+00 -1.08432567e+00 -9.01578903e-01 6.73266947e-01
4.87517357e-01 4.22663279e-02 4.54913527e-01 2.31706634e-01
-6.92162156e-01 8.73099387e-01 1.23790190e-01 -9.00287777e-02
7.89849088e-02 -1.05352509e+00 5.77611387e-01 9.92371440e-01
2.35391986e-02 3.73217970e-01 4.93140280e-01 -7.34884620e-01
-1.99574101e+00 -1.48109925e+00 6.27155721e-01 1.08624317e-01
-4.32939604e-02 -6.00614905e-01 1.27338439e-01 3.15237850e-01
8.66765857e-01 -3.86028200e-01 8.18502247e-01 -8.52246821e-01
5.90804100e-01 1.41670585e-01 -1.09460497e+00 7.50428259e-01
1.28826356e+00 -2.73331165e-01 7.05166519e-01 6.59204066e-01
6.48703218e-01 -2.48540133e-01 -4.74899650e-01 -3.33925605e-01
1.46252409e-01 -7.25362659e-01 8.97849858e-01 4.70149904e-01
9.66865644e-02 -7.09783733e-02 7.31763095e-02 -1.51606238e+00
-1.30965322e-01 -1.41477406e+00 -2.01157197e-01 1.50778890e+00
3.55597079e-01 -3.15331727e-01 1.00094604e+00 4.98780459e-01
1.02812573e-01 -1.86168656e-01 -7.47965753e-01 -7.92151868e-01
-6.21724725e-01 -2.46068895e-01 4.76941556e-01 1.05971634e-01
-1.63913310e-01 6.08254135e-01 -5.00005960e-01 2.86705852e-01
9.09885883e-01 -3.64957564e-02 8.89940381e-01 -6.21775568e-01
-3.65469664e-01 -1.70475198e-03 2.15143278e-01 -1.64094555e+00
-2.47318164e-01 -1.07344997e+00 1.70759588e-01 -1.91436195e+00
5.75978279e-01 -6.11972213e-01 2.87764937e-01 5.61117649e-01
-3.43712091e-01 4.86352384e-01 3.88917983e-01 2.62039244e-01
-8.04882526e-01 6.51172280e-01 9.88574624e-01 9.74868089e-02
1.75829306e-02 1.62344202e-01 -1.06767845e+00 7.34986290e-02
7.55179703e-01 -3.86866778e-01 -1.03263628e+00 -5.02635479e-01
2.64522582e-01 4.47400212e-01 5.47045469e-01 -1.06035006e+00
2.44688258e-01 7.53036188e-03 3.17173868e-01 -6.35139287e-01
2.15251759e-01 -6.64699554e-01 5.36375403e-01 4.45875555e-01
-1.19034290e-01 -3.95649076e-01 -1.64296061e-01 9.38399315e-01
5.32199144e-01 1.52076006e-01 4.83502090e-01 -6.07567616e-02
-2.13688999e-01 3.79847825e-01 -1.05276179e+00 -4.52924520e-02
1.39346290e+00 -1.05102599e-01 -4.02901888e-01 -1.25386071e+00
-5.61456501e-01 3.27081233e-01 2.69249648e-01 2.14400932e-01
4.40823704e-01 -1.31625414e+00 -6.03628039e-01 1.20548286e-01
-2.28576347e-01 -2.01491922e-01 3.40668499e-01 7.08831131e-01
-3.34047638e-02 5.70390522e-01 4.38050300e-01 -5.57278991e-01
-1.34381926e+00 -1.24824271e-01 -8.77563804e-02 1.24775298e-01
-2.97190130e-01 9.38552141e-01 2.45021641e-01 5.02466679e-01
-5.72971739e-02 5.41792512e-02 5.39643824e-01 -2.85542607e-01
-4.78367740e-03 7.41613328e-01 -2.28064179e-01 -3.34942341e-01
-8.21729228e-02 -3.24623249e-02 4.34423804e-01 -2.68898129e-01
1.01152050e+00 -8.69759381e-01 2.17872903e-01 -4.06536102e-01
1.10479426e+00 1.23549607e-02 -1.00121248e+00 -1.30142659e-01
-8.13796699e-01 -4.10317153e-01 3.92255872e-01 -8.06300700e-01
-1.56843877e+00 3.08034748e-01 6.00291908e-01 5.54062068e-01
1.49446464e+00 3.66249353e-01 1.35853171e+00 -3.29853147e-02
5.93557954e-01 -1.10009575e+00 2.80202240e-01 -1.81226239e-01
4.91854131e-01 -7.23548949e-01 -2.42522135e-01 -7.42540121e-01
-5.80777466e-01 6.99859500e-01 3.16852003e-01 4.14558679e-01
5.65227032e-01 5.61396122e-01 -3.76265615e-01 -1.91099629e-01
-8.03888381e-01 -2.11167246e-01 -2.53606111e-01 6.41927123e-01
2.46267706e-01 3.41396749e-01 -2.17597350e-01 4.67400372e-01
-3.63897413e-01 5.16709760e-02 9.00837362e-01 9.68886256e-01
-4.66731519e-01 -1.13348806e+00 -1.77334934e-01 6.83526635e-01
-3.13157111e-01 -2.07480088e-01 -3.02623838e-01 1.04833394e-01
9.97969974e-03 1.55750155e+00 2.81709909e-01 -3.45202088e-01
-1.89588562e-01 -2.97335923e-01 1.46387175e-01 -4.37480509e-01
-1.22639939e-01 6.27838135e-01 1.64476067e-01 -5.17168581e-01
-4.48026448e-01 -4.77892548e-01 -1.25475514e+00 -4.49082643e-01
-2.68208355e-01 1.68815359e-01 9.97247517e-01 6.87920272e-01
9.43600476e-01 7.88837612e-01 1.33615339e+00 -7.47112274e-01
-4.74529341e-02 -7.06789136e-01 -5.24557054e-01 -2.03106329e-01
-1.41956255e-01 -2.28635166e-02 -1.78105012e-01 4.91580844e-01] | [11.102506637573242, -0.4519113004207611] |
b8c7b0d9-2f83-44f9-930e-c84628a81f27 | are-quantitative-features-of-lung-nodules | 1908.05667 | null | https://arxiv.org/abs/1908.05667v1 | https://arxiv.org/pdf/1908.05667v1.pdf | Are Quantitative Features of Lung Nodules Reproducible at Different CT Acquisition and Reconstruction Parameters? | Consistency and duplicability in Computed Tomography (CT) output is essential to quantitative imaging for lung cancer detection and monitoring. This study of CT-detected lung nodules investigated the reproducibility of volume-, density-, and texture-based features (outcome variables) over routine ranges of radiation-dose, reconstruction kernel, and slice thickness. CT raw data of 23 nodules were reconstructed using 320 acquisition/reconstruction conditions (combinations of 4 doses, 10 kernels, and 8 thicknesses). Scans at 12.5%, 25%, and 50% of protocol dose were simulated; reduced-dose and full-dose data were reconstructed using conventional filtered back-projection and iterative-reconstruction kernels at a range of thicknesses (0.6-5.0 mm). Full-dose/B50f kernel reconstructions underwent expert segmentation for reference Region-Of-Interest (ROI) and nodule volume per thickness; each ROI was applied to 40 corresponding images (combinations of 4 doses and 10 kernels). Typical texture analysis metrics (including 5 histogram features, 13 Gray Level Co-occurrence Matrix, 5 Run Length Matrix, 2 Neighboring Gray-Level Dependence Matrix, and 2 Neighborhood Gray-Tone Difference Matrix) were computed per ROI. Reconstruction conditions resulting in no significant change in volume, density, or texture metrics were identified as "compatible pairs" for a given outcome variable. Our results indicate that as thickness increases, volumetric reproducibility decreases, while reproducibility of histogram- and texture-based features across different acquisition and reconstruction parameters improves. In order to achieve concomitant reproducibility of volumetric and radiomic results across studies, balanced standardization of the imaging acquisition parameters is required. | ['Matthew T. Bigelow', 'Vikash Gupta', "Thomas P. O'Donnell", 'Barbaros S. Erdal', 'Rainer Grimmer', 'Gehan F. M. Ibrahim', 'Andreas Wimmer', 'Richard D. White', 'Mutlu Demirer', 'Chiemezie C. Amadi', 'Luciano M. Prevedello', 'Kevin J. Little'] | 2019-08-14 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 2.18196958e-01 -2.60272443e-01 -3.15396070e-01 1.41718769e-02
-1.02896440e+00 -4.78545964e-01 3.08823556e-01 6.21518910e-01
-5.48842847e-01 2.37764373e-01 2.24266991e-01 -7.80072749e-01
-4.70938832e-01 -8.97975028e-01 -1.98051050e-01 -7.65067816e-01
-2.83678383e-01 6.27347469e-01 9.42202687e-01 3.77593011e-01
-1.07750431e-01 8.79377306e-01 -7.46004045e-01 4.74241197e-01
3.15924048e-01 9.41445768e-01 4.80735451e-01 9.36488628e-01
-1.05696052e-01 6.65099323e-01 -2.03804914e-02 -6.79692253e-02
3.04310739e-01 -6.38898015e-01 -7.29018867e-01 3.84795755e-01
2.27827698e-01 -3.68505985e-01 -2.31845826e-01 6.69085145e-01
3.72774124e-01 -2.46740773e-01 1.09199321e+00 -3.89408737e-01
-2.62272686e-01 3.23643863e-01 -6.18187249e-01 6.13382161e-01
-1.60591543e-01 4.15443897e-01 1.94818720e-01 -7.10551262e-01
4.15951580e-01 4.93604332e-01 8.23994160e-01 2.80484647e-01
-1.33183432e+00 -1.81545287e-01 -7.34509230e-01 -3.37742805e-01
-1.31728399e+00 1.95275515e-01 -1.63762376e-01 -7.81270325e-01
7.92914569e-01 7.30949759e-01 1.01352882e+00 1.64317995e-01
1.05771363e+00 -4.12224621e-01 1.17284691e+00 -4.80656207e-01
1.01713613e-01 2.58419991e-01 -2.55467236e-01 9.07009363e-01
7.90253639e-01 -2.05376539e-02 4.17270213e-01 -4.11729872e-01
1.28112006e+00 -9.94547736e-04 -4.62450266e-01 -2.84639657e-01
-1.45734811e+00 6.11293137e-01 3.76697153e-01 7.59734035e-01
-4.00571913e-01 3.08607221e-01 4.46557164e-01 -2.90186238e-02
1.36344984e-01 2.11899266e-01 1.47283167e-01 1.79362357e-01
-7.35103011e-01 -1.89950272e-01 4.35172558e-01 4.05207306e-01
2.56675810e-01 -1.66148409e-01 -5.53931236e-01 7.88109362e-01
1.42898515e-01 5.25589108e-01 8.18191826e-01 -7.73440003e-01
-4.93076956e-03 5.02949417e-01 -5.04164770e-02 -5.00863314e-01
-6.02775037e-01 -2.39996850e-01 -8.19772899e-01 2.50584275e-01
7.24528372e-01 2.39509657e-01 -1.21938264e+00 8.83789241e-01
3.62445205e-01 -2.14843631e-01 -4.31160241e-01 1.00926626e+00
5.24062097e-01 2.14745030e-01 5.33086300e-01 -5.32430232e-01
1.78458667e+00 -3.38077247e-01 -3.22806418e-01 1.39873043e-01
9.10711765e-01 -8.85288656e-01 1.27781188e+00 -2.87243903e-01
-1.29562664e+00 -4.11915123e-01 -6.79278493e-01 4.91383314e-01
2.52287686e-01 1.46159917e-01 9.02346224e-02 1.00753748e+00
-9.47290242e-01 7.00834692e-01 -1.14016891e+00 -4.05644059e-01
1.75725162e-01 3.76260310e-01 -3.84843618e-01 9.60933324e-03
-6.10838413e-01 1.09769130e+00 -1.19090397e-02 -3.21708292e-01
-4.23728347e-01 -1.13271248e+00 -3.17581922e-01 -1.83908165e-01
-2.22446039e-01 -9.78056431e-01 1.20091879e+00 -4.40318286e-01
-1.21158850e+00 1.01747596e+00 1.51192921e-03 -3.50097328e-01
6.18622899e-01 5.96650600e-01 -3.02162647e-01 5.15118718e-01
2.25325361e-01 5.39541170e-02 5.18809974e-01 -1.11354303e+00
-2.03089222e-01 -4.65000629e-01 -8.21331441e-01 4.02202219e-01
8.52643549e-02 1.73732117e-01 -4.08311248e-01 -2.66765893e-01
5.25227726e-01 -9.33767498e-01 -5.35353482e-01 2.62296140e-01
2.40918864e-02 5.72838545e-01 5.77548623e-01 -8.18682909e-01
1.25094402e+00 -1.88612044e+00 -4.86210644e-01 6.42434239e-01
4.05066967e-01 -4.30752844e-01 3.15302551e-01 -2.70396829e-01
-1.30473211e-01 4.57939386e-01 -5.11740327e-01 4.94734079e-01
-5.47695160e-01 -1.64735600e-01 6.72708988e-01 8.75759661e-01
-2.21748143e-01 9.83752310e-01 -6.76064134e-01 -8.72388422e-01
5.48461556e-01 4.84764606e-01 -3.07213426e-01 -1.95580557e-01
2.13783309e-01 4.52423364e-01 -4.96186137e-01 5.05489409e-01
5.91631770e-01 -5.44818401e-01 3.75218213e-01 -4.46562320e-01
-4.10667360e-01 -1.41858160e-01 -5.96065223e-01 9.82615113e-01
-5.76638639e-01 4.04388845e-01 -3.53254527e-02 4.31332737e-02
7.05743909e-01 5.41416407e-01 1.08164871e+00 -7.45469689e-01
3.16444457e-01 3.66062194e-01 4.94248003e-01 -4.90984768e-01
9.89981368e-02 -7.31619835e-01 2.58799583e-01 5.41091383e-01
-6.33470476e-01 -7.84622014e-01 4.02202979e-02 -6.89456090e-02
1.26817083e+00 -7.72727013e-01 2.56992310e-01 -5.55760562e-01
3.53760362e-01 3.40723813e-01 -1.70258924e-01 5.66646814e-01
-2.82702953e-01 8.47927272e-01 7.48417079e-01 -1.46921143e-01
-1.51522303e+00 -1.36956179e+00 -9.23803329e-01 4.11726981e-01
-3.18794623e-02 4.39047776e-02 -4.86015320e-01 -4.42278355e-01
9.45320800e-02 4.99276191e-01 -7.35471129e-01 1.68495730e-01
-5.91545820e-01 -1.05873346e+00 4.77350384e-01 4.51780260e-01
2.43258938e-01 -6.67695940e-01 -1.04746103e+00 3.47866505e-01
-2.32252702e-02 -7.30969012e-01 -4.67663437e-01 3.20248187e-01
-1.39212739e+00 -1.12652552e+00 -9.93573129e-01 -6.05510056e-01
8.94895852e-01 2.38861322e-01 1.18080723e+00 3.15022916e-01
-6.67188704e-01 7.11089134e-01 -9.39629823e-02 1.58621624e-01
-7.60072887e-01 -4.21469927e-01 -4.06457871e-01 -5.97307801e-01
-2.56251693e-01 -1.22588903e-01 -8.60570610e-01 6.64718986e-01
-9.53542113e-01 -1.52146712e-01 6.56377614e-01 8.70656192e-01
1.22063649e+00 1.21684158e-02 -1.04820147e-01 -8.87188852e-01
6.34411275e-01 -3.26684624e-01 -4.08953607e-01 4.06038582e-01
-4.86663073e-01 -2.48177186e-01 2.90431589e-01 -2.91131377e-01
-1.07937121e+00 -1.97666466e-01 4.13919240e-01 -4.35151368e-01
1.10373050e-01 5.11434197e-01 6.15621567e-01 -4.45815861e-01
1.17940021e+00 2.35936552e-01 3.24402452e-01 2.07436174e-01
-1.11690871e-01 2.40982428e-01 2.87979692e-01 -3.58891129e-01
5.49550235e-01 6.18487239e-01 3.20642292e-01 -7.07729995e-01
-1.08652100e-01 -5.50057888e-01 -7.15073526e-01 -3.02616507e-01
9.90046561e-01 -4.03124571e-01 -2.68514603e-01 1.20540231e-01
-3.63895684e-01 -5.63834310e-01 -7.82761276e-01 1.17413735e+00
-4.95958775e-01 4.88923013e-01 -7.48899519e-01 -2.08359003e-01
-4.02856976e-01 -1.42044318e+00 5.86603940e-01 5.55673726e-02
-3.02619278e-01 -1.16512299e+00 1.68833375e-01 2.16600932e-02
7.35239327e-01 2.93061435e-01 1.25262082e+00 -1.34224102e-01
-2.99773842e-01 -1.48368970e-01 -4.65587497e-01 -5.79631655e-03
5.31971395e-01 1.51338607e-01 -1.54698730e-01 -1.11669272e-01
3.99175733e-01 4.24029343e-02 6.00043774e-01 1.09929466e+00
1.45290673e+00 7.87431076e-02 -3.61537337e-01 5.49135745e-01
1.73995280e+00 3.85712057e-01 6.14664912e-01 4.58481796e-02
4.24370229e-01 3.59003276e-01 1.38013557e-01 3.70975852e-01
-2.64744729e-01 2.57746518e-01 2.02061325e-01 -3.08340728e-01
-4.00803119e-01 1.83530584e-01 -2.66364574e-01 4.27500069e-01
-5.03118575e-01 5.54704256e-02 -1.18299222e+00 4.72887367e-01
-5.24654448e-01 -8.32489014e-01 -6.06376529e-01 2.47632766e+00
5.85327864e-01 1.73002213e-01 2.16813251e-01 -1.73865691e-01
8.92449021e-01 -1.14750192e-01 -3.73240948e-01 -1.63336098e-01
3.36469322e-01 3.86736214e-01 1.06938910e+00 4.82482076e-01
-7.34394491e-01 1.68465544e-02 6.67890692e+00 7.92706609e-01
-1.39005065e+00 2.28261650e-01 9.02822018e-01 -1.02223076e-01
-6.19867921e-01 4.99739498e-02 -3.01735163e-01 2.99710900e-01
1.05947900e+00 -2.36408755e-01 4.50144634e-02 4.73609209e-01
3.16283911e-01 -1.03338468e+00 -7.21659660e-01 5.47363639e-01
-4.01051342e-01 -1.35810566e+00 -1.99699655e-01 3.66006970e-01
5.13985932e-01 1.29667178e-01 3.07953477e-01 -1.72166303e-01
-1.50055289e-01 -1.19267642e+00 3.27127010e-01 3.02140087e-01
1.57820141e+00 -2.58590966e-01 7.80394971e-01 -1.83586463e-01
-1.37258291e+00 5.29519081e-01 -3.93496275e-01 6.66832507e-01
2.59430893e-02 7.57692218e-01 -1.53900731e+00 4.01100248e-01
4.97269154e-01 -9.96333286e-02 -5.67961216e-01 1.04434860e+00
5.91778815e-01 5.86182296e-01 -6.29346013e-01 -9.39189568e-02
3.40523161e-02 -2.30045065e-01 1.45708472e-01 1.24781227e+00
4.41293836e-01 4.91997927e-01 -2.43780166e-01 7.76378214e-01
2.30720162e-01 4.09174085e-01 -2.47644991e-01 2.11283028e-01
6.51778162e-01 1.16025496e+00 -1.60058928e+00 -2.80850261e-01
-6.14305437e-01 4.17087674e-01 -2.79234469e-01 6.63730502e-02
-9.51500475e-01 1.27308011e-01 -2.05030516e-01 1.15092027e+00
-4.56125252e-02 2.93816328e-02 -7.09452689e-01 -4.80796099e-01
-2.67705381e-01 -3.94427806e-01 5.40596962e-01 -4.18294609e-01
-1.20404911e+00 5.36951542e-01 4.10804242e-01 -1.19718122e+00
4.00388911e-02 -4.47079390e-01 -7.94317603e-01 1.18403339e+00
-8.24509323e-01 -7.04120100e-01 -4.79901552e-01 3.68416309e-01
1.85833827e-01 3.69958192e-01 7.69938827e-01 9.06421989e-03
1.06163532e-01 2.95149505e-01 2.29528412e-01 4.10447828e-02
3.21156919e-01 -1.21670640e+00 -1.88646972e-01 -9.27729830e-02
-5.97737312e-01 2.41198257e-01 2.26226598e-01 -9.57696617e-01
-1.27583945e+00 -1.13829780e+00 2.74804145e-01 -3.58056068e-01
6.31810009e-01 4.43636388e-01 -8.49093974e-01 4.70835596e-01
-2.07576379e-01 4.56999362e-01 7.46884227e-01 -4.69450355e-01
3.38365346e-01 3.33090603e-01 -1.66679394e+00 5.13915002e-01
3.11421961e-01 -2.40100071e-01 1.02738246e-01 2.59824306e-01
6.39308766e-02 -7.57750690e-01 -1.63297057e+00 6.84130311e-01
8.44139993e-01 -1.20467806e+00 9.25416052e-01 -1.36061069e-02
-2.47682855e-02 -9.77638513e-02 -7.04957638e-04 -7.36817539e-01
-7.70333767e-01 4.53616560e-01 8.39558959e-01 1.18239723e-01
6.35548711e-01 -5.76633930e-01 9.15197372e-01 7.25233853e-01
-3.54387522e-01 -1.19684494e+00 -1.13792980e+00 -5.02920866e-01
4.15824056e-01 -2.92116970e-01 -1.05614729e-01 5.73000848e-01
-7.82213081e-03 -5.18551767e-01 6.59147859e-01 -2.31212422e-01
4.98423427e-01 -2.42762417e-01 7.32074752e-02 -7.07717657e-01
-3.78708452e-01 -6.02355957e-01 -2.30320022e-01 -1.95223853e-01
-6.22212768e-01 -1.10461485e+00 -4.54922348e-01 -1.83077872e+00
5.44005871e-01 -6.83911920e-01 -2.47181524e-02 7.03513548e-02
9.97785106e-02 1.94481313e-01 -1.36469617e-01 5.98607779e-01
4.72389549e-01 -2.78862100e-02 1.86955547e+00 8.27917382e-02
-2.24835873e-01 1.97327837e-01 -3.65898944e-02 5.03260195e-01
7.90117860e-01 -5.97210467e-01 -4.39144194e-01 1.27764389e-01
-5.11906326e-01 8.41601014e-01 7.54879773e-01 -1.18009782e+00
7.42976218e-02 -4.30726588e-01 1.04164302e+00 -7.00298965e-01
1.47146374e-01 -8.54460895e-01 8.55290711e-01 1.24970984e+00
-3.41571756e-02 1.25769570e-01 3.89175892e-01 2.34271467e-01
1.62756383e-01 -4.66532290e-01 1.35670841e+00 -5.59920490e-01
-7.24521801e-02 1.69862345e-01 -7.14405954e-01 -1.57682449e-01
1.27218306e+00 -8.33967328e-01 -6.10598512e-02 -4.50621359e-04
-1.06371450e+00 -4.38607723e-01 7.06193328e-01 -5.69399893e-01
6.40799761e-01 -1.26929832e+00 -9.10425842e-01 -5.77171072e-02
-2.10376441e-01 -4.80026454e-02 7.81463921e-01 1.58328605e+00
-1.23782349e+00 3.45140457e-01 -9.61643606e-02 -9.79144514e-01
-1.11998487e+00 1.39256805e-01 1.14196277e+00 -8.02433014e-01
-6.29151523e-01 8.07677686e-01 3.94641340e-01 1.58485681e-01
-3.96923691e-01 -5.27975738e-01 4.25235033e-01 -3.99244189e-01
-7.59701282e-02 4.65640962e-01 4.04897809e-01 -2.68800616e-01
-3.04926127e-01 8.14034700e-01 -1.60474461e-02 -2.85061985e-01
4.69191760e-01 -3.45060788e-02 8.66137668e-02 3.23551148e-01
9.69794691e-01 1.46351874e-01 -9.67067719e-01 1.32487386e-01
-3.19697618e-01 -5.29250741e-01 3.03897381e-01 -7.01274335e-01
-7.70618737e-01 3.84160310e-01 9.37656701e-01 2.24696338e-01
1.06039500e+00 3.60398382e-01 2.50575691e-01 -5.64165473e-01
1.04546718e-01 -5.49305856e-01 1.01682730e-01 1.01288944e-01
7.12756097e-01 -9.47218239e-01 5.11946082e-01 -6.86165452e-01
-6.59579158e-01 1.18232846e+00 5.83832324e-01 4.04920094e-02
7.50321090e-01 6.55998468e-01 -3.92417237e-02 -3.91852856e-01
-4.99332964e-01 1.39828593e-01 3.05767745e-01 4.06546086e-01
9.57932353e-01 3.63123924e-01 -6.21070564e-01 1.20543987e-01
8.44652671e-03 -1.53458714e-01 4.68683720e-01 1.00021291e+00
-8.35270703e-01 -7.48698175e-01 -9.10879493e-01 1.13308585e+00
-4.87629533e-01 1.04733603e-02 8.28680098e-02 1.23726726e+00
-1.90674618e-01 1.85017541e-01 3.13492328e-01 -4.30400595e-02
3.98891419e-01 -3.14012319e-01 9.16202128e-01 -6.50197506e-01
-9.13820148e-01 5.07549942e-01 -1.04516856e-01 -1.29582360e-01
-2.40526140e-01 -6.72714889e-01 -1.58770752e+00 -4.45956826e-01
-4.20788944e-01 4.35660668e-02 7.02865481e-01 3.29614431e-01
-3.87705952e-01 6.47616804e-01 5.14002860e-01 -3.07877570e-01
-4.40373152e-01 -9.05867338e-01 -8.58483613e-01 -3.49247195e-02
-1.75340176e-01 -3.62720430e-01 -5.82761467e-01 1.32468939e-01] | [15.069999694824219, -2.162994861602783] |
dbda3837-a1b2-4cf9-bef4-35cfd0b618c4 | spectral-cross-domain-neural-network-with | 2301.10171 | null | https://arxiv.org/abs/2301.10171v1 | https://arxiv.org/pdf/2301.10171v1.pdf | Spectral Cross-Domain Neural Network with Soft-adaptive Threshold Spectral Enhancement | Electrocardiography (ECG) signals can be considered as multi-variable time-series. The state-of-the-art ECG data classification approaches, based on either feature engineering or deep learning techniques, treat separately spectral and time domains in machine learning systems. No spectral-time domain communication mechanism inside the classifier model can be found in current approaches, leading to difficulties in identifying complex ECG forms. In this paper, we proposed a novel deep learning model named Spectral Cross-domain neural network (SCDNN) with a new block called Soft-adaptive threshold spectral enhancement (SATSE), to simultaneously reveal the key information embedded in spectral and time domains inside the neural network. More precisely, the domain-cross information is captured by a general Convolutional neural network (CNN) backbone, and different information sources are merged by a self-adaptive mechanism to mine the connection between time and spectral domains. In SATSE, the knowledge from time and spectral domains is extracted via the Fast Fourier Transformation (FFT) with soft trainable thresholds in modified Sigmoid functions. The proposed SCDNN is tested with several classification tasks implemented on the public ECG databases \textit{PTB-XL} and \textit{MIT-BIH}. SCDNN outperforms the state-of-the-art approaches with a low computational cost regarding a variety of metrics in all classification tasks on both databases, by finding appropriate domains from the infinite spectral mapping. The convergence of the trainable thresholds in the spectral domain is also numerically investigated in this paper. The robust performance of SCDNN provides a new perspective to exploit knowledge across deep learning models from time and spectral domains. The repository can be found: https://github.com/DL-WG/SCDNN-TS | ['Rossella Arcucci', 'Weiping Ding', 'Sibo Cheng', 'Che Liu'] | 2023-01-10 | null | null | null | null | ['electrocardiography-ecg', 'feature-engineering'] | ['methodology', 'methodology'] | [ 3.59037042e-01 -2.48850897e-01 7.77375698e-02 -3.55544090e-01
-7.55272508e-01 -4.48599279e-01 4.61827479e-02 1.84270203e-01
-4.68135297e-01 8.53971243e-01 -4.22678709e-01 -2.30103105e-01
-7.25585699e-01 -5.26370943e-01 -3.24647427e-01 -8.47850323e-01
-5.50254643e-01 1.46073233e-02 -3.34885418e-02 -2.13002607e-01
-1.05545737e-01 5.23376822e-01 -1.18071401e+00 3.72116923e-01
9.61156666e-01 1.40867579e+00 -1.62603080e-01 6.78317249e-01
3.32560480e-01 2.42071345e-01 -7.81414151e-01 -1.01405429e-02
2.89359361e-01 -8.39046359e-01 -4.39322114e-01 -5.17313540e-01
-2.11641058e-01 -3.59976734e-03 -2.52569228e-01 9.03043568e-01
1.11758411e+00 -2.89830774e-01 4.66535389e-01 -1.02544034e+00
-1.76128402e-01 5.69346070e-01 -4.18928057e-01 5.56031942e-01
-1.50703965e-02 4.01936173e-02 5.76565385e-01 -7.59293497e-01
3.66937280e-01 5.34387052e-01 1.24984503e+00 1.18158169e-01
-1.12579608e+00 -6.93298519e-01 -4.94519711e-01 5.41289151e-01
-1.54970312e+00 5.20247854e-02 1.30078042e+00 -4.54102606e-01
6.66213155e-01 3.67636085e-01 8.09185147e-01 1.02764773e+00
3.74097526e-01 3.19469601e-01 1.32438469e+00 -4.06790435e-01
-2.84849089e-02 -4.95021082e-02 1.60406411e-01 5.80653250e-01
-1.37021661e-01 4.35743034e-01 -6.09633148e-01 -3.02164316e-01
8.69471192e-01 -5.25627881e-02 -4.54855978e-01 1.56442914e-02
-1.22716570e+00 6.17741942e-01 2.69857824e-01 6.86276674e-01
-4.73363906e-01 -1.06819563e-01 7.80028760e-01 7.65718043e-01
6.01913989e-01 4.91147220e-01 -8.24177384e-01 -1.50893182e-01
-9.65097308e-01 5.30037731e-02 6.89030051e-01 3.86782914e-01
5.43752730e-01 3.06599706e-01 -2.48436347e-01 8.10028493e-01
5.03186621e-02 2.92361349e-01 6.60376668e-01 -6.75426126e-01
1.98559597e-01 3.93042833e-01 -2.85312533e-01 -9.33211207e-01
-9.70301092e-01 -1.02427840e+00 -1.28953791e+00 8.15052819e-03
5.02718866e-01 -5.21528542e-01 -5.73182523e-01 1.64377797e+00
3.17418247e-01 3.78827751e-01 5.22472039e-02 9.22315955e-01
9.62373674e-01 3.70214999e-01 -2.01977238e-01 -5.14980078e-01
1.42155957e+00 -1.64971113e-01 -9.27813053e-01 2.06189647e-01
6.31828547e-01 -4.99939054e-01 6.68803871e-01 6.09042943e-01
-1.01401186e+00 -9.55713749e-01 -1.40497327e+00 1.91536054e-01
-3.33492398e-01 4.07550156e-01 2.70947069e-01 6.27901793e-01
-8.81410718e-01 1.16751623e+00 -7.46867716e-01 -2.25637220e-02
4.65285778e-01 5.07841408e-01 -2.02779129e-01 3.47428232e-01
-1.86029458e+00 8.27067733e-01 4.56235439e-01 5.04941165e-01
-4.89301383e-01 -8.06654215e-01 -6.28431022e-01 -3.15143913e-02
-2.92553473e-02 -5.01045585e-01 7.22701967e-01 -1.22392929e+00
-1.27766919e+00 8.44140708e-01 3.63625526e-01 -6.67074680e-01
7.34249115e-01 -1.11623466e-01 -7.89598584e-01 4.14052427e-01
-1.40882924e-01 -8.66539218e-03 1.05327713e+00 -7.57489920e-01
-3.00722569e-01 -5.32965481e-01 -4.15188789e-01 -7.08702281e-02
-4.62153852e-01 2.33616866e-02 7.50628710e-02 -9.57146347e-01
2.41821900e-01 -5.72557092e-01 1.07313611e-01 -1.70664161e-01
-2.74020672e-01 -1.09778112e-02 6.80948436e-01 -8.57273459e-01
1.41493165e+00 -2.43101501e+00 1.07689314e-01 4.23409671e-01
4.00824130e-01 4.51163083e-01 8.26233625e-02 4.46422786e-01
-4.47526842e-01 -1.67000398e-01 -4.50652957e-01 1.54703930e-01
-1.72438174e-01 -2.36327443e-02 -3.69007252e-02 8.40197623e-01
1.89351678e-01 7.93261945e-01 -5.63372374e-01 -3.85623127e-01
1.57765269e-01 5.92749357e-01 2.47239228e-02 2.55864467e-02
2.40252152e-01 6.80401504e-01 -2.11824298e-01 4.33998376e-01
7.51574755e-01 -1.98520079e-01 2.45752141e-01 -5.59496820e-01
-1.18334606e-01 4.37829532e-02 -1.13729954e+00 1.80676270e+00
-3.31434384e-02 5.21173716e-01 -1.20503612e-01 -1.49755633e+00
1.24538732e+00 7.26501226e-01 9.72046256e-01 -7.84103274e-01
2.58184791e-01 5.72079539e-01 3.47259313e-01 -6.81551099e-01
-3.09182614e-01 -2.86186337e-01 2.48398911e-02 3.07928175e-01
3.83350253e-01 2.90589899e-01 -1.54930621e-01 -5.87522268e-01
1.07336938e+00 2.89361682e-02 3.37240875e-01 -4.99362946e-01
5.50655007e-01 -3.28553885e-01 7.14240849e-01 7.66023219e-01
-2.17393056e-01 5.60110748e-01 5.30917823e-01 -7.28233635e-01
-7.69555569e-01 -9.61406112e-01 -6.05569839e-01 8.01893532e-01
-4.35086228e-02 -1.52888909e-01 -6.28249466e-01 -3.26087892e-01
-5.45123518e-02 1.02330066e-01 -6.05982840e-01 -4.51886475e-01
-8.23418856e-01 -9.80935633e-01 1.16516423e+00 4.56147969e-01
5.66733539e-01 -1.11416006e+00 -1.07340324e+00 4.81866390e-01
-1.21219404e-01 -9.47989166e-01 -1.44675538e-01 6.77376509e-01
-1.10780978e+00 -9.49273884e-01 -8.77575994e-01 -6.08419418e-01
-5.58695570e-03 -6.18420243e-01 8.15168858e-01 -1.99128762e-02
-6.10819399e-01 5.64971790e-02 -2.75348425e-01 -6.61693931e-01
-3.44651967e-01 1.69641718e-01 -1.44411791e-02 4.21341062e-01
2.74247438e-01 -1.07171357e+00 -7.22754121e-01 2.38287821e-01
-6.87964320e-01 -1.19307756e-01 5.93959987e-01 1.07297421e+00
5.29154360e-01 -1.70784667e-02 1.12339532e+00 -5.64547122e-01
8.23420584e-01 -4.99613702e-01 -4.67785776e-01 2.21739663e-03
-7.29981184e-01 -2.29604259e-01 8.05686295e-01 -5.14354765e-01
-5.27272701e-01 -4.94074263e-02 -3.48757535e-01 -5.46713829e-01
-1.99712828e-01 7.19556749e-01 2.86834072e-02 -7.23567978e-02
8.58617902e-01 4.99359012e-01 8.65220428e-02 -3.82829547e-01
-1.74785316e-01 4.66924906e-01 6.40232205e-01 -5.05409062e-01
5.57156205e-01 3.86963427e-01 1.60800397e-01 -6.83594167e-01
-5.01336575e-01 -3.32132936e-01 -9.26345944e-01 -2.93521971e-01
8.78097296e-01 -6.76380992e-01 -5.98196328e-01 7.60804534e-01
-1.11276376e+00 -1.60131067e-01 -4.27953571e-01 6.98404968e-01
-5.37836909e-01 4.33554888e-01 -7.28332460e-01 -6.91972673e-01
-8.04988444e-01 -7.93813825e-01 7.47000694e-01 1.13209762e-01
-2.65727729e-01 -1.01361728e+00 -5.44135906e-02 -1.56860083e-01
4.66800034e-01 8.49602222e-01 9.86311018e-01 -9.58746076e-01
-6.24330528e-02 -1.98118284e-01 1.73717213e-03 6.44058347e-01
9.22715068e-02 -3.29654336e-01 -1.15995085e+00 -3.30312878e-01
5.10834932e-01 1.46376360e-02 6.72666132e-01 6.73995733e-01
1.28168190e+00 1.77234728e-02 -1.02107130e-01 1.09334481e+00
1.31663990e+00 5.40215015e-01 6.07210755e-01 2.22204730e-01
4.05905038e-01 2.00578570e-01 2.82364249e-01 7.42870390e-01
4.16908227e-02 4.12450731e-01 1.47242695e-01 -6.78357363e-01
2.01149970e-01 4.15564388e-01 -1.94295787e-03 8.27012479e-01
-4.28507745e-01 2.39393428e-01 -1.03025866e+00 3.98576945e-01
-1.76546741e+00 -8.48226547e-01 -2.15743497e-01 2.04338217e+00
1.00956881e+00 2.33225465e-01 1.32024169e-01 7.88108468e-01
6.98496521e-01 -5.78264184e-02 -8.07710052e-01 -3.22056741e-01
-3.76644701e-01 7.57182479e-01 2.43134052e-01 -2.34549269e-02
-1.28737140e+00 8.26558545e-02 5.17475796e+00 6.70787215e-01
-1.50507128e+00 3.21742117e-01 5.44622540e-01 2.45096628e-02
3.56112719e-01 -4.54355359e-01 -2.40465537e-01 3.80938917e-01
1.17957902e+00 -2.74054170e-01 2.65953213e-01 4.16788697e-01
3.46982777e-01 3.76969546e-01 -1.12811422e+00 1.34722269e+00
-2.20088899e-01 -1.31635892e+00 -6.06125176e-01 -1.04575999e-01
2.77977109e-01 6.07078411e-02 1.28592461e-01 1.39244616e-01
-7.49728382e-01 -1.12058175e+00 6.19944930e-01 7.62568533e-01
1.25887084e+00 -7.33517230e-01 9.53670561e-01 2.48191550e-01
-1.34387732e+00 -4.36771780e-01 6.95892870e-02 8.58629346e-02
-1.59305707e-01 8.94564629e-01 -5.51846564e-01 1.02124667e+00
8.42795253e-01 9.26303804e-01 -3.27022314e-01 1.00934398e+00
2.45740280e-01 8.36722076e-01 -2.57545441e-01 2.58432269e-01
-7.63183236e-02 -1.20600902e-01 5.99725187e-01 1.30928934e+00
4.27781373e-01 1.13864064e-01 6.31006062e-02 1.02557945e+00
1.87644154e-01 7.35499710e-03 -3.52180868e-01 1.78060964e-01
3.21138829e-01 1.21261990e+00 -7.06165791e-01 -2.07306430e-01
-2.21970782e-01 5.87316573e-01 -2.55662620e-01 3.33954602e-01
-1.04858804e+00 -8.57012391e-01 2.24505156e-01 1.20444380e-01
4.65342738e-02 -4.71211746e-02 -6.28836453e-01 -7.67010212e-01
2.48094037e-01 -9.21073437e-01 6.15798295e-01 -3.29358220e-01
-1.37211347e+00 9.27845299e-01 6.40849862e-03 -1.63534117e+00
-2.64162362e-01 -4.63578999e-01 -4.71049696e-01 1.19925165e+00
-1.42375469e+00 -7.21037149e-01 -2.78951317e-01 1.02947700e+00
1.69955090e-01 -2.38462836e-01 1.09337807e+00 7.61002004e-01
-1.93164945e-01 7.17955351e-01 -9.63908620e-03 2.98658192e-01
6.42415345e-01 -1.20123470e+00 1.37733191e-03 6.19974613e-01
-2.29862660e-01 3.79285514e-01 5.09579241e-01 -3.19533378e-01
-1.15699029e+00 -9.69226778e-01 5.00411451e-01 1.16931491e-01
4.36428070e-01 -2.88692385e-01 -1.19462562e+00 8.44163150e-02
2.17535406e-01 3.27072263e-01 7.02735901e-01 -7.78990164e-02
-7.10868016e-02 -4.83072519e-01 -9.45071638e-01 -9.11501702e-03
7.54141986e-01 -7.10009873e-01 -4.66452628e-01 2.07187250e-01
3.18232268e-01 -5.21081924e-01 -1.40773141e+00 8.88037801e-01
7.73388565e-01 -9.34461296e-01 8.97989392e-01 -3.66893619e-01
2.04517961e-01 -1.76034257e-01 1.23422250e-01 -1.36009669e+00
-1.53237104e-01 -8.71920586e-01 -3.05596292e-01 8.10890973e-01
3.19258541e-01 -9.90738451e-01 3.04018110e-01 -2.17114225e-01
-3.75181377e-01 -1.03706026e+00 -1.31395376e+00 -7.61695087e-01
3.56096439e-02 -4.99559402e-01 4.74208981e-01 1.16925466e+00
-4.43613790e-02 8.11744407e-02 -2.82240659e-01 1.26843840e-01
5.13667703e-01 6.37779161e-02 -4.32207994e-03 -1.57649004e+00
-4.00292456e-01 -4.75273550e-01 -4.22450751e-01 -2.01163486e-01
-2.05229402e-01 -1.07464147e+00 -3.08914661e-01 -1.11748767e+00
-2.29618713e-01 -4.08166528e-01 -9.93876815e-01 5.56549013e-01
9.85559076e-02 1.89487919e-01 1.60397589e-02 6.69353902e-02
-8.84581283e-02 2.50318319e-01 1.17008102e+00 7.21628498e-03
-3.89169276e-01 1.96154401e-01 -3.00413758e-01 7.26720929e-01
1.00507963e+00 -4.84954387e-01 -3.99021119e-01 -4.19745632e-02
8.56070966e-02 6.07878149e-01 5.25801003e-01 -1.44917977e+00
1.49377182e-01 4.32708621e-01 6.90676332e-01 -6.39402807e-01
2.90684879e-01 -6.94211245e-01 4.14617687e-01 6.71725452e-01
-3.65184128e-01 2.38227561e-01 4.27602142e-01 2.01627806e-01
-3.73259276e-01 1.93725914e-01 8.90744627e-01 -6.86713168e-03
-1.94723248e-01 2.72857875e-01 -2.84963191e-01 8.78835693e-02
8.15318942e-01 -3.48042250e-01 -9.58891660e-02 8.37355256e-02
-1.14421058e+00 -1.40855551e-01 -4.26940143e-01 1.87934637e-01
7.55846143e-01 -1.32039928e+00 -8.71389270e-01 5.26827753e-01
-2.76609398e-02 -2.69481868e-01 6.76754355e-01 1.44443500e+00
-2.95592427e-01 7.49367997e-02 -5.48958123e-01 -9.71167088e-01
-1.14926505e+00 2.25689143e-01 9.70728755e-01 -3.03519130e-01
-8.20880175e-01 5.77898502e-01 -1.59526765e-01 -1.67166889e-01
2.29137823e-01 -4.39485371e-01 -2.82589555e-01 3.33037943e-01
3.15829635e-01 3.62866491e-01 4.55904782e-01 -2.80158073e-01
-4.52815920e-01 6.60599828e-01 3.16934139e-01 9.17664915e-02
1.51844013e+00 2.11852893e-01 -1.63615286e-01 6.30962133e-01
1.36905265e+00 -5.62927008e-01 -8.02360177e-01 -2.72400022e-01
-2.28146985e-02 8.82310197e-02 -1.07447021e-01 -1.09351134e+00
-1.12174535e+00 1.05882108e+00 1.34094763e+00 4.88713145e-01
1.72718430e+00 -3.58809352e-01 7.78363764e-01 8.11537355e-02
8.08187351e-02 -1.12493694e+00 -7.44075701e-02 3.05131316e-01
9.13610756e-01 -7.62692869e-01 -2.15536445e-01 4.16090712e-02
-3.25674683e-01 1.55252016e+00 2.01937556e-01 -1.59628570e-01
1.09632719e+00 4.19083714e-01 1.79180145e-01 -3.29178333e-01
-4.78097945e-01 5.58705553e-02 3.82946908e-01 5.62783778e-01
5.00400126e-01 3.97033542e-02 -6.22758985e-01 1.14299905e+00
-8.84867534e-02 3.42691541e-01 2.65319288e-01 7.46999264e-01
-7.12966993e-02 -9.73025620e-01 -4.32818234e-01 5.33948660e-01
-7.91466713e-01 -8.10674578e-02 3.75927463e-02 7.61575818e-01
6.10171914e-01 7.16930211e-01 -2.80984223e-01 -5.01165509e-01
3.08163583e-01 3.67596239e-01 3.94736558e-01 -1.72689378e-01
-9.70113993e-01 3.55330557e-01 -1.10153131e-01 -4.52743053e-01
-4.23275411e-01 -5.82074583e-01 -1.33215106e+00 1.99115679e-01
-1.01959020e-01 3.05674728e-02 5.61071217e-01 8.74187231e-01
3.23766977e-01 9.92432177e-01 6.70567036e-01 -6.58165276e-01
-6.68717265e-01 -1.01226938e+00 -9.05793965e-01 2.81292081e-01
6.62914693e-01 -4.41006213e-01 -1.75264686e-01 3.54348160e-02] | [14.28073501586914, 3.284977436065674] |
c324cfc2-7518-4ee7-aa83-dd19d5b49048 | a-style-aware-content-loss-for-real-time-hd | 1807.10201 | null | http://arxiv.org/abs/1807.10201v2 | http://arxiv.org/pdf/1807.10201v2.pdf | A Style-Aware Content Loss for Real-time HD Style Transfer | Recently, style transfer has received a lot of attention. While much of this
research has aimed at speeding up processing, the approaches are still lacking
from a principled, art historical standpoint: a style is more than just a
single image or an artist, but previous work is limited to only a single
instance of a style or shows no benefit from more images. Moreover, previous
work has relied on a direct comparison of art in the domain of RGB images or on
CNNs pre-trained on ImageNet, which requires millions of labeled object
bounding boxes and can introduce an extra bias, since it has been assembled
without artistic consideration. To circumvent these issues, we propose a
style-aware content loss, which is trained jointly with a deep encoder-decoder
network for real-time, high-resolution stylization of images and videos. We
propose a quantitative measure for evaluating the quality of a stylized image
and also have art historians rank patches from our approach against those from
previous work. These and our qualitative results ranging from small image
patches to megapixel stylistic images and videos show that our approach better
captures the subtle nature in which a style affects content. | ['Björn Ommer', 'Dmytro Kotovenko', 'Sabine Lang', 'Artsiom Sanakoyeu'] | 2018-07-26 | a-style-aware-content-loss-for-real-time-hd-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Artsiom_Sanakoyeu_A_Style-aware_Content_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Artsiom_Sanakoyeu_A_Style-aware_Content_ECCV_2018_paper.pdf | eccv-2018-9 | ['video-style-transfer', 'image-stylization'] | ['computer-vision', 'computer-vision'] | [ 3.64283234e-01 -8.96458700e-02 1.18923739e-01 -3.52040648e-01
-5.15425265e-01 -7.03607202e-01 7.76983917e-01 -2.48306200e-01
-3.95297438e-01 6.64179146e-01 2.82990426e-01 7.79541284e-02
2.96151400e-01 -9.29919362e-01 -1.14962447e+00 -3.74256134e-01
4.34895873e-01 4.56393778e-01 2.33934909e-01 -2.08899170e-01
3.47180247e-01 4.89575118e-01 -1.33468986e+00 2.37873048e-01
6.48288071e-01 1.15688407e+00 1.52577758e-01 6.19530678e-01
-2.56355286e-01 6.81719244e-01 -8.32432568e-01 -9.12305534e-01
4.08459753e-01 -7.22121775e-01 -8.71966720e-01 4.14979279e-01
1.04632473e+00 -5.45288920e-01 -2.31711850e-01 1.22522640e+00
4.66309905e-01 -2.20099300e-01 6.52303040e-01 -1.02988088e+00
-1.18484509e+00 5.67542315e-01 -5.66679120e-01 -4.09900956e-02
2.66749799e-01 3.59524399e-01 1.00619245e+00 -6.06849670e-01
8.64975810e-01 1.59940910e+00 6.39900744e-01 4.90245849e-01
-1.52907228e+00 -4.52540815e-01 3.92476693e-02 -5.17968796e-02
-9.39205408e-01 -2.70511091e-01 1.03970695e+00 -4.56138074e-01
3.31223816e-01 -3.96007719e-03 8.92509341e-01 1.49298060e+00
-2.88523524e-03 8.58855724e-01 1.37940443e+00 -1.56937122e-01
2.53452688e-01 1.66200116e-01 -5.74140668e-01 5.27648985e-01
1.80048749e-01 -5.98317869e-02 -5.22778213e-01 3.10568124e-01
1.34546816e+00 -2.46614171e-03 -1.90004915e-01 -3.87631476e-01
-1.24840784e+00 7.18353927e-01 6.37394428e-01 3.88629168e-01
-2.20157996e-01 3.26350451e-01 4.78007108e-01 5.19502163e-01
5.60564160e-01 8.65347028e-01 -1.57460928e-01 -3.01213115e-01
-1.28807211e+00 2.55392015e-01 6.24002814e-01 9.17143226e-01
8.43519449e-01 2.26527721e-01 -1.14638381e-01 6.24323010e-01
-2.49776825e-01 5.79111755e-01 1.14861660e-01 -1.08278966e+00
1.47675604e-01 7.43682206e-01 1.35466218e-01 -1.12041831e+00
1.05306953e-01 -2.49665335e-01 -7.87014961e-01 5.20516872e-01
6.33806527e-01 4.53404859e-02 -6.97169602e-01 1.61309373e+00
-1.47228688e-01 -2.09281892e-02 -3.78263980e-01 1.13125384e+00
4.82290089e-01 4.13404912e-01 -1.12693369e-01 3.09503049e-01
1.28048491e+00 -1.06658399e+00 -5.58755577e-01 -2.74444968e-01
7.90774822e-02 -8.80928576e-01 1.66214406e+00 6.28899872e-01
-1.39818656e+00 -5.91431379e-01 -9.70598221e-01 -4.48866814e-01
-4.39458996e-01 3.01324204e-02 3.30654114e-01 6.52434230e-01
-1.24248326e+00 9.00487483e-01 -4.93169874e-01 -5.86678803e-01
7.70285785e-01 5.42844385e-02 -3.53599548e-01 -1.71490852e-02
-8.30184281e-01 9.45095301e-01 3.85217555e-02 -2.45268732e-01
-7.88738787e-01 -8.57379675e-01 -7.59732902e-01 -8.47627670e-02
4.02037770e-01 -9.63337302e-01 9.15151358e-01 -1.82577205e+00
-1.64431918e+00 9.68846202e-01 6.69912100e-02 -3.61896724e-01
8.54341030e-01 -5.02519429e-01 -2.43421763e-01 2.46164501e-01
9.60848480e-02 8.57925832e-01 1.11421072e+00 -1.51913059e+00
-4.22790736e-01 -2.70754099e-01 4.11680937e-01 5.78192575e-03
-4.58596945e-01 -3.42637487e-02 -5.09948552e-01 -1.09126067e+00
-4.13528174e-01 -7.61823237e-01 3.99123691e-02 3.21520239e-01
-3.60720962e-01 2.47830957e-01 8.97494912e-01 -6.29228890e-01
9.53122020e-01 -2.07643819e+00 4.88157302e-01 -2.34768659e-01
2.43604090e-02 1.59218922e-01 -2.09604576e-01 2.19894990e-01
2.21762434e-01 3.94248307e-01 -4.00819153e-01 -4.72081304e-01
9.00381431e-02 1.15895323e-01 -4.67465937e-01 2.39506692e-01
4.81606960e-01 1.05036807e+00 -1.04769468e+00 -4.91747439e-01
3.45395595e-01 5.66926122e-01 -6.89533353e-01 1.60015851e-01
-4.07304704e-01 5.60049474e-01 -9.29882824e-02 7.83904552e-01
5.77776253e-01 -2.78831691e-01 -1.70223922e-01 -3.47319931e-01
9.31530818e-02 6.01994060e-02 -9.92956936e-01 2.18418360e+00
-6.03432119e-01 7.94076383e-01 3.53607386e-02 -6.04368806e-01
6.26930416e-01 2.73892377e-02 2.27134675e-01 -6.59744799e-01
2.85826832e-01 1.54790998e-01 -4.60602850e-01 -2.30417758e-01
7.61365950e-01 -2.04117537e-01 -1.07345410e-01 5.95948100e-01
2.20119655e-02 -5.22891164e-01 1.20078877e-01 1.36980757e-01
1.09219170e+00 5.61433554e-01 3.32779922e-02 -1.42883092e-01
1.46119654e-01 -8.70288312e-02 2.65417844e-01 6.76724494e-01
-1.10062562e-01 1.02717686e+00 5.52472949e-01 -5.09502709e-01
-1.56710994e+00 -1.11684883e+00 -2.02371851e-02 1.03257585e+00
2.88760990e-01 -2.37230822e-01 -1.02275693e+00 -7.43382633e-01
-2.12313179e-02 5.36529481e-01 -8.09232295e-01 5.31329028e-03
-6.35111511e-01 -1.83638185e-01 4.97053951e-01 4.50824738e-01
6.89713895e-01 -1.37737572e+00 -6.20117068e-01 7.41020367e-02
-8.64790231e-02 -1.20641363e+00 -7.66209245e-01 -1.88676566e-01
-9.99719083e-01 -9.64214921e-01 -9.92553592e-01 -5.59555471e-01
5.90102196e-01 2.99275224e-03 1.61215186e+00 -1.17927648e-01
-2.22890705e-01 3.51590276e-01 -3.05036157e-01 -3.18575233e-01
-4.18492883e-01 5.77365980e-02 -7.26652890e-02 3.00537944e-01
1.47594869e-01 -7.81000137e-01 -8.21165025e-01 1.29357144e-01
-1.22383440e+00 4.71580364e-02 9.88935232e-01 9.00422454e-01
3.59262913e-01 -2.45631710e-01 3.74002665e-01 -1.03521216e+00
4.09826368e-01 -8.97724777e-02 -3.64879042e-01 8.88453200e-02
-4.22945946e-01 2.40254551e-01 6.97765291e-01 -3.86694700e-01
-1.00226068e+00 7.00242817e-02 6.67698830e-02 -7.49532819e-01
-2.95852780e-01 -1.92227773e-02 -2.47708157e-01 1.01314113e-02
5.93224823e-01 1.71835542e-01 -1.68827325e-01 -5.49958825e-01
7.11027443e-01 3.35220397e-01 5.95829308e-01 -7.07913637e-01
9.39755082e-01 8.40956748e-01 -1.78514555e-01 -7.39440262e-01
-8.30664217e-01 -1.40115339e-02 -8.23934197e-01 -1.98313400e-01
9.37315524e-01 -8.33893359e-01 -5.25674701e-01 3.38619709e-01
-1.25777256e+00 -4.11052644e-01 -6.13455057e-01 1.07710697e-01
-6.56944990e-01 4.84411031e-01 -6.64345920e-01 -4.79781091e-01
-3.06849360e-01 -1.14988124e+00 1.43621576e+00 -6.73796982e-02
-4.05600607e-01 -8.95182669e-01 -1.83572039e-01 2.48466715e-01
5.61583042e-01 5.53753078e-01 8.77387166e-01 1.52634695e-01
-7.72623897e-01 -3.12472731e-02 -5.90828538e-01 4.78306979e-01
2.27228865e-01 2.41605882e-02 -1.06816733e+00 -1.44615695e-01
-4.82852124e-02 -5.31922817e-01 9.69942391e-01 2.10305244e-01
1.25297379e+00 -4.55120921e-01 8.85190144e-02 6.92839682e-01
1.66877139e+00 -8.11375603e-02 8.59198272e-01 3.64751309e-01
1.08845878e+00 5.58412075e-01 2.05843151e-01 2.84204423e-01
9.58018005e-02 6.43164098e-01 6.40805364e-01 -4.37518328e-01
-4.31899190e-01 -2.69170672e-01 4.19734150e-01 4.75892127e-01
-3.99212122e-01 -1.99892744e-01 -4.19953972e-01 6.90148652e-01
-1.51356697e+00 -1.00992465e+00 2.10998073e-01 2.06387067e+00
1.04126883e+00 2.53730506e-01 3.93409520e-01 1.05703942e-01
6.10242009e-01 3.18383306e-01 -7.05513597e-01 -4.81434345e-01
-4.43830252e-01 3.33584040e-01 4.14417535e-01 2.53689408e-01
-9.49470103e-01 1.18553185e+00 6.58249950e+00 9.42790151e-01
-1.33960140e+00 2.83830538e-02 7.67609179e-01 -2.56923109e-01
-3.35629582e-01 -9.12636966e-02 -2.08086178e-01 6.18896067e-01
4.20994699e-01 2.43805289e-01 6.76868558e-01 9.65843916e-01
7.87065625e-02 -1.48806378e-01 -1.44584906e+00 1.13197649e+00
1.95713863e-01 -1.40362763e+00 5.96595585e-01 1.03294402e-01
9.46060240e-01 -2.32339934e-01 3.59610230e-01 7.64555484e-03
3.60137284e-01 -1.17899084e+00 1.16891563e+00 7.07560599e-01
9.21101868e-01 -6.50804102e-01 4.35261607e-01 -2.37033948e-01
-8.71653318e-01 2.24282160e-01 -3.41363132e-01 -9.04674232e-02
8.71707350e-02 6.85043931e-01 -3.81839484e-01 4.23440903e-01
8.19899261e-01 8.51813614e-01 -8.64105761e-01 5.67603469e-01
-1.73442125e-01 4.97148603e-01 2.18092557e-02 6.07177615e-02
2.88178474e-01 -1.74340367e-01 4.92881775e-01 1.29200137e+00
2.77539819e-01 -2.78765827e-01 -9.80286077e-02 1.37294126e+00
-3.85962725e-01 -2.63817813e-02 -7.21622884e-01 -2.00782850e-01
1.41549945e-01 1.28738976e+00 -8.20013821e-01 -3.97273093e-01
-4.65984315e-01 1.45211935e+00 3.04197162e-01 2.67411411e-01
-6.76297784e-01 -2.59610534e-01 6.16647422e-01 4.04562145e-01
4.65579420e-01 -1.37221217e-01 -7.99180269e-01 -1.28579557e+00
-1.09056961e-02 -9.33015466e-01 -1.47327587e-01 -1.00765872e+00
-1.43552995e+00 6.68621242e-01 -5.44060349e-01 -1.18784380e+00
-2.13913750e-02 -7.45382667e-01 -6.82524979e-01 7.54006028e-01
-1.25260794e+00 -1.29877579e+00 -3.14894199e-01 4.08670932e-01
6.13951445e-01 -2.40072962e-02 6.70981526e-01 3.38989049e-01
-3.16323102e-01 5.30258238e-01 -1.57769930e-04 3.09436768e-01
1.09219718e+00 -1.59173059e+00 4.33775544e-01 8.61605942e-01
6.44373223e-02 4.21984553e-01 8.48377347e-01 -5.22745490e-01
-1.39352739e+00 -1.06299663e+00 6.20036960e-01 -8.65761936e-01
5.93400598e-01 -4.62285757e-01 -7.01694667e-01 5.52547276e-01
6.23851359e-01 -2.54146159e-01 2.12922245e-01 4.06070687e-02
-4.73927289e-01 -1.51952237e-01 -9.85636592e-01 8.74766350e-01
1.21763802e+00 -6.21191859e-01 -4.29298073e-01 -5.08695245e-02
5.92699468e-01 -8.62419233e-02 -9.10148799e-01 1.77533343e-01
6.39820874e-01 -1.23499930e+00 9.55206037e-01 -3.75028998e-01
1.18846297e+00 -3.37290257e-01 -1.17560383e-02 -1.39360011e+00
-1.59094885e-01 -4.11423534e-01 1.96216851e-01 1.27975106e+00
1.98248088e-01 -1.73935555e-02 7.91229129e-01 3.42078775e-01
-5.62530719e-02 -5.36905587e-01 -5.28302312e-01 -8.76788795e-01
1.72484860e-01 -3.09685856e-01 7.00879455e-01 9.94078934e-01
-4.83534843e-01 4.69480723e-01 -6.92093790e-01 -2.84611821e-01
5.65096736e-01 2.87002772e-01 1.02190530e+00 -1.06924832e+00
-2.73983032e-01 -8.75491798e-01 -4.72412556e-01 -7.65497983e-01
-1.33002698e-01 -6.20409191e-01 -1.23119123e-01 -1.54080057e+00
1.67223707e-01 -1.68803900e-01 2.41898769e-03 1.76779330e-01
-1.41692042e-01 8.43775272e-01 2.51823068e-01 2.34190792e-01
-6.43855274e-01 4.57121819e-01 1.64538884e+00 -3.40685189e-01
1.57201126e-01 -5.30989826e-01 -7.56894708e-01 8.27083886e-01
5.27882457e-01 -3.04063201e-01 -7.55614266e-02 -6.92363143e-01
5.08999467e-01 -2.71750778e-01 7.59611845e-01 -1.01490867e+00
-1.51775226e-01 4.29481603e-02 7.11472631e-01 -3.22442472e-01
3.57124448e-01 -8.49046469e-01 1.08342156e-01 3.84101212e-01
-2.59492666e-01 -3.37890498e-02 3.32561582e-02 7.02473164e-01
-2.44504884e-01 1.31156249e-03 9.13954794e-01 -3.89409244e-01
-7.20455229e-01 3.03280771e-01 1.10543132e-01 1.75830796e-01
6.08711421e-01 -3.59694064e-01 -1.17827095e-01 -3.81353289e-01
-5.91030419e-01 -3.15861851e-01 1.14447796e+00 3.68811160e-01
2.61149615e-01 -1.47809434e+00 -6.47313833e-01 -1.13802113e-01
8.32949281e-02 2.29999656e-03 1.47669725e-02 4.51503605e-01
-7.92156160e-01 3.29222740e-03 -6.42956734e-01 -5.81841230e-01
-8.32421064e-01 7.79710472e-01 1.22306988e-01 -9.29428563e-02
-6.01641119e-01 7.12555110e-01 5.04246235e-01 -9.25998688e-02
1.91157967e-01 -5.33891380e-01 5.22893630e-02 1.68758690e-01
2.86741525e-01 3.28702092e-01 -7.54054487e-02 -6.43373430e-01
-3.36425081e-02 9.02667046e-01 1.07200786e-01 -2.26390257e-01
1.22029209e+00 -2.08878785e-01 -1.19214952e-01 6.56201720e-01
9.65432167e-01 1.29293784e-01 -1.70734191e+00 -3.48986387e-01
-1.09219819e-01 -6.32998109e-01 -1.20005637e-01 -8.71553302e-01
-1.26638353e+00 1.10417867e+00 4.86361682e-01 1.51799768e-01
1.26752257e+00 1.52932620e-02 9.63812530e-01 2.26229087e-01
3.85777324e-01 -1.24756026e+00 6.70360982e-01 3.45191568e-01
1.22607422e+00 -1.46977222e+00 -9.21122581e-02 -4.53942120e-02
-7.59336054e-01 1.04532719e+00 4.08525169e-01 -5.50867081e-01
4.13498372e-01 1.32625267e-01 9.48548242e-02 -7.21357912e-02
-1.61331758e-01 -2.74334043e-01 3.84536594e-01 4.00009930e-01
3.64090174e-01 -3.28443898e-03 -1.72722898e-02 5.10388672e-01
-4.86602485e-01 1.23626590e-01 4.51293945e-01 7.17158973e-01
-2.30374008e-01 -9.89438832e-01 -3.03945631e-01 2.64472246e-01
-5.87152719e-01 -1.86253250e-01 -8.56411517e-01 7.26513922e-01
2.54791975e-01 6.23751521e-01 1.91907927e-01 -4.44512784e-01
3.63712311e-01 -1.56035990e-01 8.28187168e-01 -5.00048935e-01
-6.87723577e-01 7.98668638e-02 -2.53635854e-01 -6.73982024e-01
-6.57130063e-01 -5.41101873e-01 -5.99923313e-01 -7.65831053e-01
1.43490404e-01 -3.89938563e-01 6.14716649e-01 7.61226356e-01
1.99899077e-01 6.95611656e-01 4.85934138e-01 -1.24908841e+00
2.71559563e-02 -9.79230165e-01 -7.12526739e-01 7.57744133e-01
1.76680848e-01 -5.50136507e-01 -3.57825518e-01 5.39407611e-01] | [11.445302963256836, -0.3635285496711731] |
1b3ad9d5-57c7-4551-9009-ed0ad1785c26 | deep-learning-human-mind-for-automated-visual | 1609.00344 | null | https://arxiv.org/abs/1609.00344v2 | https://arxiv.org/pdf/1609.00344v2.pdf | Deep Learning Human Mind for Automated Visual Classification | What if we could effectively read the mind and transfer human visual capabilities to computer vision methods? In this paper, we aim at addressing this question by developing the first visual object classifier driven by human brain signals. In particular, we employ EEG data evoked by visual object stimuli combined with Recurrent Neural Networks (RNN) to learn a discriminative brain activity manifold of visual categories. Afterwards, we train a Convolutional Neural Network (CNN)-based regressor to project images onto the learned manifold, thus effectively allowing machines to employ human brain-based features for automated visual classification. We use a 32-channel EEG to record brain activity of seven subjects while looking at images of 40 ImageNet object classes. The proposed RNN based approach for discriminating object classes using brain signals reaches an average accuracy of about 40%, which outperforms existing methods attempting to learn EEG visual object representations. As for automated object categorization, our human brain-driven approach obtains competitive performance, comparable to those achieved by powerful CNN models, both on ImageNet and CalTech 101, thus demonstrating its classification and generalization capabilities. This gives us a real hope that, indeed, human mind can be read and transferred to machines. | ['Concetto Spampinato', 'Simone Palazzo', 'Nasim Souly', 'Isaak Kavasidis', 'Mubarak Shah', 'Daniela Giordano'] | 2016-09-01 | deep-learning-human-mind-for-automated-visual-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Spampinato_Deep_Learning_Human_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Spampinato_Deep_Learning_Human_CVPR_2017_paper.pdf | cvpr-2017-7 | ['object-categorization'] | ['computer-vision'] | [ 3.86128217e-01 8.52444395e-02 3.19893301e-01 -3.09221059e-01
-1.67439833e-01 -3.72114182e-01 8.85167181e-01 -4.12481725e-02
-6.66355193e-01 6.61956549e-01 -1.22092694e-01 -1.20287836e-01
-9.73537490e-02 -5.23350418e-01 -7.27042139e-01 -7.08362579e-01
-5.73240668e-02 2.38558888e-01 -8.84379447e-02 3.93840969e-02
5.29097795e-01 6.25666440e-01 -1.76339519e+00 3.77637088e-01
3.64698350e-01 1.29540837e+00 4.51105177e-01 6.70173049e-01
1.56795651e-01 1.19924033e+00 -6.46354139e-01 1.27228022e-01
-3.27725871e-03 -6.28327310e-01 -9.90768552e-01 -3.30043845e-02
3.01464140e-01 5.55509515e-02 -3.34189206e-01 1.14347363e+00
3.13493431e-01 6.47787228e-02 1.00106585e+00 -1.13563907e+00
-1.11885047e+00 3.10416877e-01 -3.55387896e-01 7.51659870e-01
4.37751919e-01 2.05923051e-01 8.16830993e-01 -9.88482535e-01
3.43217701e-01 8.47460806e-01 1.65051043e-01 7.23226607e-01
-1.46749783e+00 -5.57923913e-01 -1.42427579e-01 6.57355249e-01
-1.33391058e+00 -3.07407141e-01 7.75147140e-01 -7.69210160e-01
1.27464020e+00 7.16501549e-02 1.09963620e+00 1.54739749e+00
5.02962708e-01 5.67445040e-01 1.59090912e+00 -4.63553756e-01
4.35854256e-01 5.02770960e-01 5.35655320e-01 3.27328682e-01
8.15545917e-02 2.14742064e-01 -5.28159261e-01 1.83742806e-01
7.86656797e-01 1.97975844e-01 -5.62628746e-01 -5.13400733e-02
-1.58475554e+00 6.37759149e-01 1.00084090e+00 7.96020508e-01
-8.01610291e-01 2.43901685e-01 3.05486917e-01 4.90021199e-01
9.76772383e-02 6.65726542e-01 -2.70331204e-02 4.05095853e-02
-9.13457513e-01 -2.77663589e-01 5.92778087e-01 5.29715359e-01
5.83559990e-01 3.76456738e-01 -2.05678731e-01 7.49414027e-01
3.12532932e-01 3.79165262e-01 1.03060377e+00 -5.86065829e-01
-3.96956243e-02 5.71799994e-01 -3.27995181e-01 -9.64441836e-01
-6.02832913e-01 -4.28928465e-01 -1.06946254e+00 5.05811870e-01
2.93202758e-01 2.80119836e-01 -1.00433540e+00 1.42239702e+00
-5.55750310e-01 2.21165400e-02 2.12279886e-01 1.12723672e+00
8.27743769e-01 6.28438592e-01 2.52299964e-01 -8.87752324e-02
1.64329922e+00 -4.38474029e-01 -4.52506602e-01 -2.79987663e-01
1.63386270e-01 -9.99463350e-02 9.19121146e-01 5.96276700e-01
-9.03797925e-01 -8.14197183e-01 -1.20536029e+00 1.31238967e-01
-5.99051893e-01 1.26945451e-01 4.32153583e-01 2.94911832e-01
-1.25819433e+00 5.00900805e-01 -6.92071378e-01 -5.89760661e-01
7.58201182e-01 5.81013978e-01 -4.47847903e-01 4.23865139e-01
-9.78138924e-01 1.28040051e+00 5.54722965e-01 2.56817847e-01
-1.25976110e+00 -3.91412288e-01 -7.16392636e-01 3.21670830e-01
-3.20880353e-01 -5.45055091e-01 7.94357479e-01 -1.59188902e+00
-1.53116846e+00 1.23801816e+00 -1.45502703e-03 -7.61103094e-01
-6.27521425e-02 2.31168151e-01 -4.34985101e-01 3.44351590e-01
-2.13995472e-01 1.00999916e+00 1.14600945e+00 -1.01239502e+00
-2.34734371e-01 -5.82540929e-01 -3.89633179e-01 -3.08569670e-01
-4.40861940e-01 -1.36927078e-02 3.31840485e-01 -4.30312991e-01
-1.52615262e-02 -8.30547214e-01 2.35552847e-01 -3.45293462e-01
-2.10267350e-01 -5.89811265e-01 4.77720469e-01 -5.34653425e-01
5.50968051e-01 -2.13437510e+00 4.45118934e-01 3.51622701e-01
4.98570442e-01 2.18689427e-01 -3.76619309e-01 1.29764890e-02
-5.08392513e-01 -2.92474300e-01 -9.04164240e-02 1.75743684e-01
-1.58696726e-01 -1.26743410e-02 -4.89372641e-01 5.68256497e-01
4.94149208e-01 1.36002350e+00 -5.87010384e-01 5.04715219e-02
2.94310421e-01 6.92065477e-01 -3.19189966e-01 3.67463857e-01
2.05292001e-01 5.16158104e-01 -7.29843006e-02 4.19228792e-01
3.47271919e-01 -3.47213417e-01 -2.37539485e-02 -1.39152691e-01
-3.48147303e-02 2.76024222e-01 -4.07157034e-01 1.59598649e+00
-5.24937272e-01 1.43423164e+00 -4.94838297e-01 -1.69634926e+00
1.21127057e+00 4.72441226e-01 4.22742479e-02 -1.33265674e+00
5.26562750e-01 5.07566184e-02 3.44218284e-01 -6.23150826e-01
-1.95260778e-01 -1.68996185e-01 3.92662257e-01 2.81695634e-01
7.41984129e-01 6.31542653e-02 -3.31973583e-01 -1.72147185e-01
9.66689885e-01 -1.54385015e-01 2.98349708e-01 -6.30134106e-01
7.26458907e-01 -2.24578574e-01 -2.22009644e-01 6.68183029e-01
-6.41646534e-02 5.36727011e-01 5.17600000e-01 -8.35384548e-01
-9.27235365e-01 -1.12477779e+00 -4.30178195e-01 9.65294123e-01
-8.12624767e-02 1.88137978e-01 -7.61091948e-01 -1.24585308e-01
-2.35267624e-01 6.88608527e-01 -1.03939426e+00 -4.20953363e-01
-2.71936387e-01 -7.09913194e-01 3.30849349e-01 7.11956322e-01
4.38389450e-01 -1.80141914e+00 -1.10612738e+00 3.07156801e-01
1.76607326e-01 -9.55849528e-01 3.50258946e-01 6.97259367e-01
-7.12212265e-01 -1.10355687e+00 -9.79011714e-01 -9.39365387e-01
6.11807823e-01 -1.29324675e-01 9.86711800e-01 -1.27346829e-01
-7.46284425e-01 6.60371959e-01 -6.86026737e-02 -6.47414744e-01
-2.17066959e-01 1.28527237e-02 1.66230083e-01 2.81593114e-01
7.59101272e-01 -8.66296053e-01 -6.84988737e-01 2.19278852e-03
-7.27398813e-01 -2.94789970e-01 7.51505494e-01 8.33594978e-01
1.66665778e-01 -4.78620708e-01 8.36752355e-01 -2.63131082e-01
6.17123961e-01 -4.65562940e-01 -6.02631152e-01 2.03346670e-01
-4.26872402e-01 7.76319951e-02 7.07743049e-01 -5.71203470e-01
-6.92383826e-01 6.61420375e-02 -5.56800589e-02 -4.48968947e-01
-5.80714643e-01 2.79783458e-01 3.03697407e-01 -2.99932450e-01
1.03799272e+00 6.76338971e-01 -1.33790508e-01 -1.88926011e-01
1.51760906e-01 8.16709876e-01 7.82150447e-01 -2.30172332e-02
3.49723935e-01 2.67054856e-01 -9.50484723e-02 -9.57511365e-01
-5.22913039e-01 -3.59759420e-01 -8.19574773e-01 -3.42067838e-01
1.29656124e+00 -8.71022105e-01 -1.21769631e+00 3.51430744e-01
-1.30281281e+00 -2.88473129e-01 -1.67929381e-01 7.74034202e-01
-8.16095114e-01 -1.55423507e-01 -5.42495549e-01 -6.92776799e-01
-4.73913401e-01 -8.01728666e-01 6.73225582e-01 1.80858999e-01
-2.65702039e-01 -8.67903948e-01 6.62036464e-02 3.50405090e-02
5.65257311e-01 -3.22369812e-03 9.27978039e-01 -8.86846483e-01
-3.08195502e-01 -1.72131732e-01 -4.57117140e-01 5.40537000e-01
-9.04850662e-03 -4.73407477e-01 -1.49556947e+00 -2.87742317e-01
3.90892237e-01 -6.28681958e-01 9.93555069e-01 3.16912562e-01
1.42111218e+00 -1.44063741e-01 -2.33527452e-01 4.62531298e-01
1.47139168e+00 3.76945913e-01 9.85002756e-01 3.87971848e-01
4.72306550e-01 5.71934104e-01 -3.41096669e-01 1.76101550e-01
6.72448426e-02 5.16352415e-01 4.21723217e-01 -1.16883636e-01
2.22567413e-02 8.15959051e-02 2.15623349e-01 6.83036625e-01
-5.35416901e-01 1.19498141e-01 -1.03785467e+00 5.19111276e-01
-1.41853499e+00 -9.76822734e-01 9.24005266e-03 2.10281658e+00
4.67917621e-01 1.23634212e-01 1.98129565e-02 3.00938696e-01
4.27682310e-01 -2.39593998e-01 -4.09372658e-01 -5.50019085e-01
7.30391368e-02 5.39779902e-01 4.93209362e-02 -8.12976584e-02
-8.96767080e-01 5.15109539e-01 6.35883665e+00 2.73684233e-01
-1.58345437e+00 1.30407646e-01 4.47661489e-01 1.35199185e-02
1.97233692e-01 -4.35231090e-01 -2.90663153e-01 3.25118899e-01
1.41683090e+00 -8.21470246e-02 8.10877740e-01 6.97235703e-01
-6.39597401e-02 4.15976346e-02 -1.41080153e+00 1.55410862e+00
3.99608225e-01 -1.16967511e+00 1.13181286e-01 1.69691294e-01
3.52597117e-01 2.09240735e-01 2.51448721e-01 4.78772610e-01
-1.59731030e-01 -1.62346458e+00 6.92935526e-01 8.48942280e-01
5.46666980e-01 -5.24335206e-01 5.47243059e-01 4.00635481e-01
-7.47639954e-01 -4.38769519e-01 -7.49752164e-01 -2.09186673e-01
-4.51884001e-01 7.73785710e-02 -7.10349858e-01 -4.08596732e-02
8.90071154e-01 9.76188898e-01 -7.70351708e-01 1.01195538e+00
-5.86520806e-02 6.87938571e-01 1.00814022e-01 -3.77446055e-01
4.02095109e-01 8.76113847e-02 3.18419099e-01 1.38319659e+00
2.16031790e-01 3.24897557e-01 -4.02319372e-01 1.40517426e+00
-1.72277242e-01 -5.03493175e-02 -9.04240847e-01 -2.34478060e-02
-1.56396553e-01 1.49058020e+00 -8.07187736e-01 -5.19950807e-01
-3.06784719e-01 1.10302246e+00 4.98537034e-01 5.30619621e-01
-6.21824205e-01 -6.63559675e-01 2.73704261e-01 -1.08063556e-01
1.72867075e-01 -1.06113665e-01 -2.04777539e-01 -1.29421067e+00
-1.53243050e-01 -6.89874887e-01 -5.04956283e-02 -1.13947177e+00
-1.15108275e+00 1.28522873e+00 -1.17170185e-01 -1.10438061e+00
-3.82325202e-01 -1.16149902e+00 -6.92094624e-01 7.97225535e-01
-1.33879137e+00 -8.32740903e-01 -5.59913039e-01 9.10279632e-01
4.98857141e-01 -4.45860118e-01 9.28989351e-01 -5.23026213e-02
-1.87487155e-01 2.14239374e-01 -1.72670767e-01 2.86571413e-01
3.44532222e-01 -1.15773273e+00 1.15649156e-01 4.24809277e-01
7.75814891e-01 5.62784910e-01 3.38728249e-01 1.06408380e-01
-1.29945278e+00 -8.23365927e-01 5.97340405e-01 -6.34302258e-01
5.55779099e-01 -6.94037616e-01 -1.31740236e+00 7.04968393e-01
8.03448677e-01 1.51101828e-01 4.89586502e-01 -2.26286232e-01
-3.77291232e-01 -1.43207252e-01 -8.72249186e-01 2.59108663e-01
7.36448348e-01 -8.56825829e-01 -1.28328252e+00 4.05151486e-01
6.37618229e-02 2.55669683e-01 -6.38389826e-01 1.04427978e-01
4.89501029e-01 -8.52111340e-01 9.35926855e-01 -8.18256676e-01
2.08343998e-01 1.49309244e-02 -8.25254917e-02 -1.47780657e+00
-5.14369190e-01 -5.73256798e-02 2.42066115e-01 7.14358509e-01
3.49348098e-01 -8.85316968e-01 2.51277149e-01 2.65632004e-01
2.32134253e-01 -4.10497457e-01 -9.60484684e-01 -6.60019517e-01
7.37610087e-02 -3.13212812e-01 9.61009786e-02 7.76970685e-01
3.13713640e-01 4.48031545e-01 -1.07723929e-01 -1.71946615e-01
5.87244511e-01 -9.22954753e-02 1.50029093e-01 -1.45009041e+00
2.24314537e-03 -5.25154054e-01 -1.15500724e+00 -5.59219420e-01
6.49633646e-01 -1.33444297e+00 -8.10117461e-03 -1.25216067e+00
4.92041886e-01 2.00784534e-01 -8.68570387e-01 6.73728406e-01
3.64068270e-01 8.09091747e-01 2.86539495e-01 2.89413899e-01
-4.13330019e-01 3.96513969e-01 8.65540981e-01 -4.08086777e-01
-6.22195601e-02 -1.57738939e-01 -4.60980982e-01 6.41168237e-01
7.26002157e-01 -4.26063627e-01 -1.11526050e-01 -1.77960753e-01
2.64624041e-02 -1.15879089e-01 1.20312738e+00 -1.53580141e+00
4.42887664e-01 4.12283659e-01 1.17392170e+00 -8.08000863e-02
2.63742417e-01 -8.45632195e-01 -9.74089950e-02 6.71581209e-01
-6.37230217e-01 -7.32825473e-02 3.71715099e-01 4.22832340e-01
-4.48983669e-01 -2.11887255e-01 9.09904540e-01 -2.60147899e-01
-8.11438501e-01 -1.52976103e-02 -9.62469399e-01 -2.55300850e-01
1.02519894e+00 -3.16287309e-01 -3.05425644e-01 -1.33203745e-01
-1.10429800e+00 -2.50986993e-01 -6.82433546e-02 4.73066539e-01
9.78115857e-01 -1.24599516e+00 -6.77815318e-01 6.81860566e-01
1.82830840e-01 -8.42326462e-01 -5.32327443e-02 9.52596068e-01
-2.32634038e-01 6.91925287e-01 -9.91331100e-01 -1.03624213e+00
-9.45679963e-01 9.47435200e-01 5.80275953e-01 5.03860474e-01
-7.82421589e-01 6.93686962e-01 3.53291363e-01 1.03484362e-03
3.26699018e-02 -3.95444691e-01 -6.63029194e-01 1.95040926e-01
7.44542658e-01 -3.26664485e-02 2.49105990e-01 -6.52457893e-01
-5.01236022e-01 6.16161048e-01 1.11209303e-01 -1.09599553e-01
1.51343894e+00 2.33695120e-01 -2.53392071e-01 6.44049823e-01
1.53355563e+00 -7.30042219e-01 -9.09927130e-01 6.05112826e-03
2.01658919e-01 -2.74724551e-02 8.03880244e-02 -9.07373965e-01
-1.04469895e+00 1.48256958e+00 1.09439540e+00 4.34457719e-01
1.21832609e+00 2.58707166e-01 8.07725266e-02 7.64622211e-01
4.43625778e-01 -7.76661932e-01 3.68259907e-01 2.79058844e-01
1.29136240e+00 -1.28549445e+00 -4.31212813e-01 3.48199278e-01
-7.37573862e-01 1.54567218e+00 3.88204515e-01 -7.26934254e-01
6.18184865e-01 -1.74291089e-01 9.10205022e-03 -6.10866904e-01
-8.40594947e-01 -2.04020306e-01 6.24273300e-01 5.72914600e-01
3.43051523e-01 3.07099614e-02 2.14104831e-01 3.26825798e-01
-1.60634533e-01 2.26139322e-01 5.05676270e-01 4.43836868e-01
-6.21793687e-01 -4.07547116e-01 -2.78857112e-01 6.26366735e-01
-3.87547284e-01 -1.22102998e-01 -2.21922114e-01 6.43883646e-01
-2.44921148e-01 6.79023504e-01 3.96946371e-01 -3.32839519e-01
4.24508601e-01 3.81316811e-01 7.49816120e-01 -4.43140090e-01
-5.81772983e-01 -3.61570925e-01 -6.25386596e-01 -5.54284871e-01
-5.03723681e-01 -3.06097388e-01 -1.06758010e+00 2.40948260e-01
-1.01366460e-01 -2.30541658e-02 8.11433733e-01 9.73716676e-01
9.35480893e-02 6.50255740e-01 6.20165288e-01 -1.27763128e+00
-3.60019177e-01 -1.19126070e+00 -7.02608228e-01 4.22484010e-01
5.97011864e-01 -7.49737978e-01 -5.16963422e-01 2.59863377e-01] | [9.98229694366455, 2.423039674758911] |
7d748ff9-1011-4fc4-b83b-2f8a163d4844 | two-level-attention-with-two-stage-multi-task | 1811.12139 | null | http://arxiv.org/abs/1811.12139v1 | http://arxiv.org/pdf/1811.12139v1.pdf | Two-level Attention with Two-stage Multi-task Learning for Facial Emotion Recognition | Compared with facial emotion recognition on categorical model, the
dimensional emotion recognition can describe numerous emotions of the real
world more accurately. Most prior works of dimensional emotion estimation only
considered laboratory data and used video, speech or other multi-modal
features. The effect of these methods applied on static images in the real
world is unknown. In this paper, a two-level attention with two-stage
multi-task learning (2Att-2Mt) framework is proposed for facial emotion
estimation on only static images. Firstly, the features of corresponding
region(position-level features) are extracted and enhanced automatically by
first-level attention mechanism. In the following, we utilize Bi-directional
Recurrent Neural Network(Bi-RNN) with self-attention(second-level attention) to
make full use of the relationship features of different layers(layer-level
features) adaptively. Owing to the inherent complexity of dimensional emotion
recognition, we propose a two-stage multi-task learning structure to exploited
categorical representations to ameliorate the dimensional representations and
estimate valence and arousal simultaneously in view of the correlation of the
two targets. The quantitative results conducted on AffectNet dataset show
significant advancement on Concordance Correlation Coefficient(CCC) and Root
Mean Square Error(RMSE), illustrating the superiority of the proposed
framework. Besides, extensive comparative experiments have also fully
demonstrated the effectiveness of different components. | ['Fuji Ren', 'Xiaohua Wang', 'Min Hu', 'Muzi Peng', 'Lijuan Pan', 'Chunhua Jin'] | 2018-11-29 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [-3.35301682e-02 -2.51844168e-01 -2.29261001e-03 -6.30547702e-01
-3.72700006e-01 -1.45959839e-01 3.66925001e-01 -3.83140504e-01
-4.41653788e-01 4.63906705e-01 2.27471411e-01 3.93105447e-01
-2.50099003e-02 -2.86783487e-01 -1.92430139e-01 -8.18702877e-01
-1.18859097e-01 -2.89589107e-01 -4.70166624e-01 -3.21676403e-01
2.46708870e-01 5.92383027e-01 -1.82949805e+00 5.58828652e-01
7.04476833e-01 1.61690450e+00 -1.10177331e-01 4.97230351e-01
-3.47206831e-01 6.40241861e-01 -6.10293269e-01 -3.84616643e-01
-5.33523932e-02 -3.55050147e-01 -4.00835216e-01 1.64565489e-01
4.99309273e-03 -4.23022449e-01 -6.25488162e-02 9.08055186e-01
8.16770136e-01 2.01647609e-01 7.33959496e-01 -1.40711045e+00
-6.50830388e-01 -9.95023400e-02 -9.18487787e-01 1.56203523e-01
1.75657913e-01 -1.43688455e-01 6.38187289e-01 -1.19604611e+00
2.04007491e-01 1.35279870e+00 3.30081284e-01 5.37741065e-01
-6.77559078e-01 -9.87957656e-01 4.71979588e-01 4.22319680e-01
-1.34849226e+00 -4.25894320e-01 1.16341758e+00 -2.36973405e-01
9.92280722e-01 -6.88353479e-02 5.38849115e-01 1.24972057e+00
3.84752929e-01 6.89800322e-01 1.34196889e+00 -1.68853417e-01
-8.87230411e-02 5.07827580e-01 6.55633211e-02 6.73286974e-01
-3.47978860e-01 4.55741957e-02 -4.77812827e-01 -1.87593717e-02
6.36584520e-01 1.29859289e-02 -5.38013093e-02 -4.19996604e-02
-6.46644771e-01 7.91258097e-01 2.47702897e-01 6.43145323e-01
-7.08988011e-01 -1.45872697e-01 9.64981973e-01 3.35014075e-01
6.32354975e-01 3.24367695e-02 -4.26859468e-01 -2.29286194e-01
-5.53450286e-01 -4.38641459e-01 2.92112947e-01 5.61086953e-01
5.66871703e-01 4.74086732e-01 -1.55682266e-01 1.10251713e+00
2.80155689e-01 3.83486480e-01 9.05565143e-01 -5.86071670e-01
1.84849706e-02 6.53160512e-01 -2.16094106e-01 -1.48480868e+00
-7.44345188e-01 -3.30185413e-01 -1.20282996e+00 1.42340481e-01
-2.97904223e-01 -4.91552174e-01 -6.25733435e-01 1.95977831e+00
3.84985864e-01 9.20675173e-02 2.35923067e-01 1.10325897e+00
9.40644979e-01 8.14795911e-01 3.60615045e-01 -7.02188909e-01
1.45343339e+00 -7.49175847e-01 -1.19761491e+00 1.39677059e-03
5.55074155e-01 -3.84400606e-01 8.04603875e-01 5.05275965e-01
-6.74641907e-01 -9.00234520e-01 -9.89130795e-01 1.51219696e-01
-5.90596855e-01 4.77070570e-01 9.32652116e-01 7.02342033e-01
-9.15420830e-01 9.60703120e-02 -2.97283977e-01 -3.48914005e-02
3.72729808e-01 4.24461007e-01 -4.49829668e-01 3.49506140e-01
-1.61443460e+00 8.42676222e-01 1.20131232e-01 7.02379584e-01
-7.01498747e-01 -2.82395244e-01 -9.22156215e-01 6.72280937e-02
5.47931604e-02 -2.09770352e-01 6.52361751e-01 -1.79129016e+00
-1.77019060e+00 5.85413933e-01 -1.69035256e-01 2.18670830e-01
-7.79411420e-02 -3.23392332e-01 -7.51754403e-01 3.52571279e-01
-3.94505769e-01 7.55375385e-01 1.02967620e+00 -1.03508484e+00
-2.91491836e-01 -6.71912313e-01 -2.08912805e-01 5.43734491e-01
-8.57246101e-01 2.79444039e-01 -1.71571836e-01 -4.24328625e-01
-2.09516492e-02 -5.43287575e-01 -4.07142797e-03 1.52418297e-02
1.48049474e-01 -3.80030751e-01 9.37788665e-01 -5.92809260e-01
1.32646441e+00 -2.33841372e+00 2.75458157e-01 1.17065527e-01
4.78830710e-02 2.76024222e-01 -4.36205596e-01 -4.01140414e-02
-3.88672799e-01 5.88703761e-03 3.39871734e-01 -1.77792564e-01
-1.25764892e-01 -1.52227670e-01 -1.42719850e-01 4.60762054e-01
4.35184449e-01 7.13326752e-01 -3.87532532e-01 -6.56778157e-01
2.81830072e-01 7.75864542e-01 -2.90591389e-01 2.95670360e-01
2.48865649e-01 2.92014539e-01 -5.57132959e-01 8.37871253e-01
6.68365300e-01 2.77176797e-02 -1.35490103e-02 -7.59714782e-01
-5.26330471e-02 -5.08998752e-01 -1.20548141e+00 1.53729153e+00
-7.53864706e-01 4.23959941e-01 2.72146761e-01 -1.01621282e+00
1.26375329e+00 5.90769470e-01 7.04608262e-01 -9.76239443e-01
5.46247363e-01 -1.47688523e-01 2.50802618e-02 -8.98515821e-01
3.81249219e-01 -2.42338896e-01 -1.67756647e-01 2.78349817e-01
3.74355614e-01 3.34672838e-01 -3.38725090e-01 -2.50877887e-01
4.27855700e-01 1.72369301e-01 3.87267113e-01 4.46782932e-02
6.86768353e-01 -5.69802284e-01 6.51203752e-01 1.02574512e-01
-6.48698568e-01 4.91932370e-02 5.50670505e-01 -3.98089588e-01
-6.42288983e-01 -3.78393054e-01 -1.28187329e-01 1.39062476e+00
1.10233240e-01 -1.33038074e-01 -4.90549892e-01 -4.95679051e-01
-1.91413566e-01 4.59906071e-01 -9.17384267e-01 -5.09937704e-01
-6.39719144e-02 -1.07695627e+00 4.93209213e-01 5.28019130e-01
5.74850619e-01 -1.16118085e+00 -6.41265452e-01 8.37401375e-02
-5.65847754e-03 -1.05056965e+00 -1.08638965e-01 1.63568988e-01
-5.36857963e-01 -5.85374832e-01 -6.55297339e-01 -5.94288290e-01
4.66148794e-01 -3.95662114e-02 5.93791723e-01 -2.85358131e-01
-3.66416425e-01 4.76870060e-01 -2.47743651e-01 -4.32885736e-01
3.52357924e-01 -3.26345205e-01 2.24660754e-01 6.09501064e-01
5.19747615e-01 -5.84821820e-01 -4.40187454e-01 2.65120894e-01
-7.50240862e-01 -1.67422071e-01 8.96490037e-01 9.06200588e-01
4.00639385e-01 -5.49259037e-02 1.06468821e+00 -3.94097656e-01
9.03748214e-01 -5.61579704e-01 -2.26867244e-01 2.17889175e-01
-4.38039154e-01 -2.32609689e-01 6.86470985e-01 -7.99162567e-01
-1.36525500e+00 -1.88045390e-02 -1.02324463e-01 -9.11159933e-01
-2.29265720e-01 4.18778569e-01 -3.84278715e-01 -7.27928281e-02
1.66582406e-01 2.79798895e-01 1.29907966e-01 9.51552540e-02
2.50356674e-01 9.39300656e-01 1.50367260e-01 -4.00765926e-01
8.48222151e-02 1.21829152e-01 -2.36616307e-03 -7.61829078e-01
-7.53524184e-01 -1.75425127e-01 -6.23921692e-01 -6.48335099e-01
8.94283950e-01 -1.18496096e+00 -1.20665288e+00 6.61493361e-01
-1.04774868e+00 1.36764765e-01 3.72432619e-01 6.70321226e-01
-4.58646566e-01 2.02271178e-01 -7.41087139e-01 -1.25701320e+00
-6.03113949e-01 -1.01391268e+00 1.11059713e+00 3.65623742e-01
-2.32492387e-02 -9.04478788e-01 -2.14236110e-01 -3.46024823e-03
4.66405064e-01 3.17672163e-01 8.77999365e-01 -4.85047013e-01
1.71425432e-01 -1.55402467e-01 -3.57782900e-01 3.06465358e-01
2.07878590e-01 1.07362255e-01 -1.21579731e+00 9.68550816e-02
2.62644112e-01 -7.98120320e-01 6.14481747e-01 5.30095220e-01
1.49169469e+00 -1.02244087e-01 1.38300238e-02 5.19864440e-01
1.44786036e+00 3.79487902e-01 6.01130307e-01 3.15180682e-02
6.70552671e-01 7.37365842e-01 8.02503884e-01 8.59579444e-01
3.56937170e-01 4.03766602e-01 4.72081482e-01 -2.60333776e-01
3.37835461e-01 2.35592142e-01 4.97516364e-01 9.58170116e-01
-8.19692686e-02 5.86739043e-03 -3.38686764e-01 2.73591042e-01
-1.56875026e+00 -1.06488740e+00 2.50608653e-01 1.61002493e+00
5.71513653e-01 -6.30294234e-02 7.27078095e-02 -7.36264735e-02
8.88863623e-01 2.44137347e-01 -9.26041186e-01 -9.71200347e-01
-2.95660943e-01 -2.13420808e-01 -2.18941271e-01 6.53958097e-02
-1.14964187e+00 6.84621394e-01 5.85669327e+00 8.78098369e-01
-1.47449207e+00 -5.37722781e-02 1.11335874e+00 -2.73933589e-01
-8.27741995e-02 -7.20759511e-01 -6.51316345e-01 4.34668601e-01
1.08164454e+00 -3.13504301e-02 2.32407019e-01 1.06209123e+00
1.84011191e-01 -2.14436874e-02 -6.84681296e-01 1.41120505e+00
4.58543062e-01 -4.69040990e-01 1.32787853e-01 -1.67208478e-01
4.01034147e-01 -5.99168003e-01 4.40501451e-01 7.51565635e-01
-4.92218584e-01 -9.59198475e-01 3.07467252e-01 7.80426443e-01
9.31415558e-01 -1.06618655e+00 8.84355068e-01 9.34455767e-02
-1.23582518e+00 -3.77155960e-01 -4.10992473e-01 -4.59211506e-02
7.80599564e-03 5.02692878e-01 -2.21619070e-01 5.92600763e-01
6.23599589e-01 6.39732957e-01 -3.17860067e-01 3.13920200e-01
1.73517525e-01 1.20245956e-01 -1.68857992e-01 -4.10560340e-01
2.64763296e-01 -2.53525883e-01 1.05155721e-01 1.20612681e+00
3.12837988e-01 5.07914543e-01 -2.38691539e-01 5.15660465e-01
3.19530070e-02 5.55750966e-01 -4.96655673e-01 -1.24781460e-01
1.30634502e-01 1.84591877e+00 -3.96004081e-01 -3.86348248e-01
-5.51628888e-01 1.06789875e+00 4.03696537e-01 2.35634729e-01
-1.10708153e+00 -4.92103517e-01 7.54903138e-01 -6.96678579e-01
2.70089209e-01 6.45904765e-02 -1.10323235e-01 -1.02985656e+00
-8.20991248e-02 -8.48643303e-01 3.80535483e-01 -1.02938724e+00
-1.26935768e+00 1.02189147e+00 -1.92700937e-01 -1.07189286e+00
-2.11022601e-01 -6.91381156e-01 -5.18008530e-01 7.68987715e-01
-1.50414491e+00 -1.11373878e+00 -3.77532691e-01 7.28689551e-01
6.02035701e-01 -9.15571600e-02 1.08134043e+00 3.86118770e-01
-9.37476695e-01 8.16227078e-01 -1.35710090e-01 8.12219009e-02
8.45810473e-01 -7.86650300e-01 -6.34394705e-01 3.66693407e-01
-3.54246646e-01 3.59798282e-01 3.22281092e-01 -3.41823727e-01
-1.37925577e+00 -8.91613781e-01 3.11078012e-01 1.39732927e-01
4.81856167e-01 -3.26077253e-01 -8.67862523e-01 2.46174738e-01
2.76939392e-01 2.23502498e-02 8.34955513e-01 2.72149205e-01
-3.39963347e-01 -3.42785239e-01 -1.08095062e+00 4.92214352e-01
6.20473802e-01 -5.73616445e-01 -2.35675573e-01 -7.04417676e-02
5.34453452e-01 -1.38115048e-01 -1.09753847e+00 7.53018320e-01
9.38822865e-01 -1.06334484e+00 6.02760494e-01 -5.58350623e-01
5.30530572e-01 6.45122230e-02 -2.86582530e-01 -1.26446080e+00
-3.50667208e-01 -4.65286151e-02 -1.23454951e-01 1.33463025e+00
2.09798887e-01 -4.84802365e-01 2.82417923e-01 5.65503240e-01
-8.92981663e-02 -1.25784230e+00 -7.35495687e-01 -1.77883059e-01
-3.18848789e-01 -2.14359045e-01 4.57177043e-01 1.04585731e+00
1.60219043e-01 5.85125208e-01 -6.74694240e-01 9.26462039e-02
1.43634781e-01 2.31538489e-01 3.04841280e-01 -1.08901703e+00
1.02973640e-01 -4.95759636e-01 -5.06560624e-01 -5.69377899e-01
5.78582227e-01 -3.96695673e-01 -1.75962090e-01 -1.00065529e+00
2.60899514e-01 1.09960146e-01 -7.30541229e-01 4.30372328e-01
-2.35317707e-01 1.20679870e-01 -9.20133814e-02 -1.48555130e-01
-7.67419815e-01 1.19106305e+00 1.26422656e+00 6.43753260e-02
-2.03779921e-01 -3.10781091e-01 -7.09379494e-01 6.60201490e-01
6.85712934e-01 -6.98187649e-02 -5.45364976e-01 -2.00230509e-01
6.58238977e-02 2.73990482e-01 1.20617868e-02 -8.85088205e-01
1.01287477e-01 -6.80765659e-02 9.96130586e-01 -5.37399769e-01
8.35851550e-01 -8.89294326e-01 -1.60538867e-01 1.32474348e-01
-3.97810996e-01 1.86042577e-01 5.18804729e-01 4.02707428e-01
-4.85909522e-01 1.06516287e-01 7.26532519e-01 2.09499113e-02
-1.03324223e+00 4.69191700e-01 -3.41006964e-01 -4.60597008e-01
1.19174302e+00 -2.29420617e-01 -4.15738188e-02 -5.02488673e-01
-8.21740329e-01 1.71263233e-01 -2.26878703e-01 5.23566425e-01
8.65634739e-01 -1.62573969e+00 -5.17698765e-01 3.96465421e-01
1.33858100e-01 -6.62658751e-01 8.99611175e-01 8.87093902e-01
2.93380439e-01 3.06933165e-01 -7.09968507e-01 -3.88807118e-01
-1.18124437e+00 7.42712975e-01 5.41840494e-01 -1.79406125e-02
3.65503877e-02 7.62759387e-01 4.17718738e-01 -2.09343076e-01
1.35153592e-01 1.03911094e-01 -7.51500249e-01 6.71560466e-01
6.45998597e-01 7.07560107e-02 -8.20804760e-02 -8.76419544e-01
-4.70524251e-01 7.45051563e-01 -2.10048258e-01 -3.74068879e-02
1.35558963e+00 -3.56675208e-01 -1.62180066e-01 6.92715943e-01
1.52110946e+00 -4.28835183e-01 -1.17229235e+00 -3.46936174e-02
-3.73947978e-01 -1.74015194e-01 2.12182492e-01 -7.59552121e-01
-1.21059930e+00 1.00583971e+00 9.91868198e-01 -1.67144492e-01
1.72264719e+00 -3.93961251e-01 4.25823987e-01 2.72504091e-01
-1.43049927e-02 -1.35676455e+00 4.98267949e-01 2.53575683e-01
8.85980546e-01 -1.21759915e+00 -2.20493883e-01 -9.82764289e-02
-1.15388048e+00 1.19028854e+00 1.16887963e+00 2.79812217e-02
8.23745191e-01 2.03553066e-01 2.55945116e-01 -3.00076038e-01
-1.10068154e+00 -1.18417978e-01 3.18002403e-01 3.86320293e-01
4.79452968e-01 -3.55848789e-01 -1.47580832e-01 1.05213749e+00
2.49949276e-01 -6.20804392e-02 1.35382578e-01 5.25338173e-01
-4.96579647e-01 -4.10839230e-01 -3.00697714e-01 4.57177162e-01
-4.75592494e-01 6.66088313e-02 -2.22657815e-01 5.32498538e-01
9.06870961e-02 9.45770144e-01 2.17593282e-01 -6.58686638e-01
2.03410983e-01 2.83652425e-01 4.31704640e-01 -4.25777473e-02
-4.82420981e-01 3.38102281e-01 -3.74643505e-02 -7.45888114e-01
-7.86763251e-01 -4.91644621e-01 -1.03902149e+00 1.48343474e-01
-4.19582635e-01 1.23718798e-01 9.53734994e-01 8.65846932e-01
6.56554937e-01 6.98250234e-01 1.21253502e+00 -9.48178232e-01
-2.73682624e-01 -1.27431190e+00 -7.36076534e-01 3.15278828e-01
2.04149961e-01 -8.89704704e-01 -4.42185283e-01 -1.84923753e-01] | [13.6426362991333, 1.8166791200637817] |
78d781d1-853d-4aa1-8497-503d1b4622f3 | learning-transferable-visual-models-from | 2103.00020 | null | https://arxiv.org/abs/2103.00020v1 | https://arxiv.org/pdf/2103.00020v1.pdf | Learning Transferable Visual Models From Natural Language Supervision | State-of-the-art computer vision systems are trained to predict a fixed set of predetermined object categories. This restricted form of supervision limits their generality and usability since additional labeled data is needed to specify any other visual concept. Learning directly from raw text about images is a promising alternative which leverages a much broader source of supervision. We demonstrate that the simple pre-training task of predicting which caption goes with which image is an efficient and scalable way to learn SOTA image representations from scratch on a dataset of 400 million (image, text) pairs collected from the internet. After pre-training, natural language is used to reference learned visual concepts (or describe new ones) enabling zero-shot transfer of the model to downstream tasks. We study the performance of this approach by benchmarking on over 30 different existing computer vision datasets, spanning tasks such as OCR, action recognition in videos, geo-localization, and many types of fine-grained object classification. The model transfers non-trivially to most tasks and is often competitive with a fully supervised baseline without the need for any dataset specific training. For instance, we match the accuracy of the original ResNet-50 on ImageNet zero-shot without needing to use any of the 1.28 million training examples it was trained on. We release our code and pre-trained model weights at https://github.com/OpenAI/CLIP. | ['Ilya Sutskever', 'Gretchen Krueger', 'Jack Clark', 'Pamela Mishkin', 'Amanda Askell', 'Girish Sastry', 'Sandhini Agarwal', 'Gabriel Goh', 'Aditya Ramesh', 'Chris Hallacy', 'Jong Wook Kim', 'Alec Radford'] | 2021-02-26 | null | null | null | null | ['zero-shot-transfer-image-classification', 'open-vocabulary-attribute-detection', 'object-categorization', 'zero-shot-cross-modal-retrieval', 'meme-classification'] | ['computer-vision', 'computer-vision', 'computer-vision', 'miscellaneous', 'natural-language-processing'] | [ 4.34002548e-01 -8.49220678e-02 -5.13251126e-01 -5.98350346e-01
-8.45022261e-01 -8.15109372e-01 1.04752266e+00 4.52413596e-02
-6.97924435e-01 4.93921667e-01 2.81964242e-01 -2.31860667e-01
4.40064639e-01 -4.57461536e-01 -1.23114848e+00 -4.18037325e-01
2.68743008e-01 5.27544677e-01 3.40705603e-01 -2.00195070e-02
3.37471515e-01 1.94218799e-01 -1.71713853e+00 6.18744552e-01
1.73385963e-01 1.04375386e+00 4.88819301e-01 7.83825517e-01
-1.05409004e-01 9.84445751e-01 -3.80608201e-01 -4.44423854e-01
3.16797972e-01 -2.02080578e-01 -9.49885428e-01 4.67453003e-01
8.60817313e-01 -5.09441674e-01 -5.04792213e-01 9.29873168e-01
1.55966118e-01 5.32063954e-02 7.64030695e-01 -1.22419643e+00
-9.58550096e-01 3.46119642e-01 -5.98123312e-01 3.48713249e-01
1.97594762e-01 4.59791452e-01 9.90231693e-01 -1.12984490e+00
8.54198158e-01 1.01516128e+00 5.99970996e-01 6.77250385e-01
-1.17289984e+00 -5.43261528e-01 9.52307805e-02 3.07101846e-01
-1.34600282e+00 -6.33769989e-01 3.74458641e-01 -6.44701779e-01
9.42197978e-01 -4.34930176e-02 3.89056206e-01 1.45329380e+00
-2.80438125e-01 7.73561478e-01 9.52384710e-01 -4.65061814e-01
4.93329987e-02 3.22442859e-01 2.53996909e-01 7.15948522e-01
1.24489948e-01 -6.67361915e-02 -5.03624201e-01 6.83256686e-02
5.51891863e-01 1.95410743e-01 -2.28224725e-01 -5.40911376e-01
-1.44255435e+00 7.40456879e-01 5.83086073e-01 1.54236749e-01
5.75720845e-03 4.67765331e-01 5.57644725e-01 1.87569275e-01
3.94726038e-01 4.16620225e-01 -6.22273147e-01 -9.76238698e-02
-9.59433794e-01 -7.98277855e-02 6.63283229e-01 1.20792460e+00
1.09035611e+00 -1.73332468e-01 -2.46524140e-01 7.23841071e-01
4.82537597e-02 5.56715429e-01 6.07258856e-01 -1.05897212e+00
5.37476659e-01 2.92987436e-01 2.72667930e-02 -5.02400339e-01
2.21887566e-02 -2.42538005e-01 -5.88234067e-01 -1.77245885e-02
5.59967458e-01 1.78447403e-02 -1.44548500e+00 1.52604318e+00
5.62274903e-02 3.56705040e-01 -2.81531028e-02 8.23541522e-01
7.06177294e-01 8.10310543e-01 1.49763346e-01 3.78178656e-01
1.37455106e+00 -1.19753671e+00 -6.68935552e-02 -6.39749169e-01
5.86034060e-01 -7.28572488e-01 1.26785493e+00 5.37414402e-02
-6.47533774e-01 -5.81293106e-01 -7.57357419e-01 -4.08960253e-01
-7.61730075e-01 2.26788044e-01 5.37069857e-01 2.48903126e-01
-1.17182076e+00 4.68588769e-01 -7.20543206e-01 -7.19473243e-01
7.06662178e-01 1.28300861e-01 -6.11802876e-01 -5.43791413e-01
-8.27838302e-01 8.82671833e-01 4.02606726e-01 -3.74565750e-01
-1.22412062e+00 -7.65987813e-01 -1.01649380e+00 -4.68900980e-05
4.14296091e-01 -5.78736842e-01 1.37244868e+00 -1.52790499e+00
-1.09851837e+00 1.36428881e+00 -9.97995585e-02 -6.23846531e-01
5.08303046e-01 -2.01859266e-01 -3.01783606e-02 3.31408322e-01
4.52091783e-01 1.20946789e+00 1.17151856e+00 -1.10443139e+00
-6.02920413e-01 -1.77870229e-01 1.96540937e-01 3.85903120e-02
-3.83855551e-01 1.31811425e-01 -8.31234992e-01 -5.10727882e-01
-4.99540687e-01 -1.05085957e+00 1.36805801e-02 2.76520431e-01
-2.99323857e-01 -2.95094937e-01 7.30768919e-01 -5.05243897e-01
5.05340517e-01 -2.30836153e+00 -1.07307687e-01 -1.79222107e-01
2.36036088e-02 2.22172126e-01 -5.40412664e-01 4.63241458e-01
-1.53353766e-01 5.88206723e-02 -2.01074824e-01 -3.92265201e-01
-1.12642005e-01 1.55808270e-01 -5.13583899e-01 5.93761861e-01
4.04062569e-01 1.04892123e+00 -9.08815503e-01 -3.89803022e-01
4.23671216e-01 4.37836111e-01 -5.21571040e-01 3.02469760e-01
-4.43964154e-01 2.81521142e-01 -1.85997173e-01 5.24684966e-01
2.10774496e-01 -7.58147240e-01 -1.13682725e-01 -2.34723583e-01
3.45019326e-02 1.24839395e-01 -6.84806049e-01 1.86771464e+00
-4.81074512e-01 1.06388950e+00 -3.20570439e-01 -1.34847939e+00
6.12235844e-01 9.84410793e-02 3.79010946e-01 -4.63639379e-01
4.98516038e-02 -4.34752777e-02 -2.81175405e-01 -5.47418773e-01
2.59982497e-01 2.70527273e-01 -1.54024318e-01 3.93546909e-01
5.95834672e-01 7.42648020e-02 3.86666358e-01 3.75010043e-01
1.14535367e+00 3.11915576e-01 2.91539222e-01 -2.01313838e-01
2.82883048e-01 4.12640989e-01 1.20100237e-01 9.27699804e-01
-6.80357516e-02 8.59690666e-01 9.80850011e-02 -4.64803308e-01
-1.35584474e+00 -9.57291424e-01 -1.00608692e-01 1.56027675e+00
5.00333915e-03 -3.31609994e-01 -7.18953788e-01 -7.11573303e-01
2.02157740e-02 6.04502380e-01 -8.14767540e-01 9.55498666e-02
-3.20520163e-01 -1.19955145e-01 4.64850783e-01 6.51840925e-01
5.74223638e-01 -1.09949899e+00 -4.28836435e-01 -1.45933867e-01
-1.37080550e-01 -1.25999665e+00 -6.74483299e-01 1.91590443e-01
-5.32221079e-01 -1.19145191e+00 -8.67140710e-01 -8.38301241e-01
8.21334779e-01 5.80989361e-01 1.31242895e+00 1.53426751e-01
-4.96725351e-01 8.68848860e-01 -4.07753617e-01 -3.82700056e-01
-2.54708052e-01 1.42914265e-01 -1.62799612e-01 1.22603193e-01
7.40119338e-01 -1.65312141e-01 -6.19389653e-01 1.04472645e-01
-9.60421264e-01 1.37325734e-01 6.57018244e-01 8.25856686e-01
5.01611114e-01 -5.28654277e-01 3.19252610e-01 -8.56445074e-01
5.99061809e-02 -6.65481150e-01 -6.49237692e-01 3.31387967e-01
-1.66782066e-01 2.41507411e-01 6.43316805e-01 -5.99249959e-01
-7.50485897e-01 6.07810497e-01 1.68987691e-01 -6.87349200e-01
-6.05696440e-01 2.51688123e-01 2.45813861e-01 5.65012135e-02
7.35333443e-01 4.56644714e-01 -1.23835422e-01 -4.93964672e-01
7.73798466e-01 8.06193411e-01 7.76071548e-01 -3.47116262e-01
8.99027765e-01 5.92741132e-01 -4.11818832e-01 -8.35938513e-01
-1.24045670e+00 -8.32976878e-01 -7.64781952e-01 -4.86570038e-02
1.04510462e+00 -1.16173017e+00 -3.25073510e-01 2.60878116e-01
-1.18104124e+00 -6.23598874e-01 -3.35024059e-01 3.88211012e-01
-7.82281041e-01 2.00101897e-01 -4.66981679e-01 -2.49304980e-01
-2.28056177e-01 -9.24072504e-01 1.23508263e+00 6.64476529e-02
3.30791324e-02 -8.32782745e-01 -1.30361974e-01 3.86807978e-01
3.49127114e-01 -1.12235650e-01 5.44909418e-01 -7.40710914e-01
-7.00149000e-01 -1.97762355e-01 -4.86229748e-01 5.01085579e-01
5.29037006e-02 -1.46462888e-01 -1.08438277e+00 -3.59046608e-01
-2.93621838e-01 -8.74727845e-01 1.21246672e+00 2.70184070e-01
1.46936142e+00 -2.80041248e-01 -3.95371497e-01 6.74365819e-01
1.61588752e+00 -3.48413497e-01 4.34740663e-01 4.56608057e-01
7.89089620e-01 5.82670093e-01 5.75855553e-01 2.50308007e-01
3.70185763e-01 5.58525622e-01 3.81736487e-01 -6.18126541e-02
-3.43663156e-01 -3.83570254e-01 4.38218176e-01 2.38779217e-01
-2.10008249e-02 -1.44938484e-01 -1.00032485e+00 8.23750496e-01
-1.71858823e+00 -1.08398688e+00 2.72191495e-01 2.14356947e+00
1.00257969e+00 -3.59508507e-02 1.94620062e-02 -3.52964163e-01
8.92083347e-01 1.72461957e-01 -7.04983354e-01 4.59579304e-02
1.73400089e-01 1.60373583e-01 1.03497672e+00 2.02836096e-01
-1.35871089e+00 1.20108092e+00 6.11863613e+00 6.42930388e-01
-1.27432084e+00 2.40376264e-01 7.67417729e-01 -1.48297593e-01
1.62249506e-01 -4.91951667e-02 -9.21178639e-01 4.48654443e-01
1.08988202e+00 -9.68476683e-02 4.99155045e-01 1.08112144e+00
-9.16307345e-02 -9.15389955e-02 -1.51995158e+00 1.17565215e+00
3.81768644e-01 -1.64841163e+00 2.07541898e-01 -7.26421624e-02
7.55806684e-01 7.86245406e-01 4.07805704e-02 2.81431019e-01
5.83595574e-01 -1.05555809e+00 8.49607706e-01 4.86092418e-01
1.24715769e+00 -1.41644672e-01 4.24784392e-01 2.71063894e-01
-9.40414429e-01 -1.12600997e-01 -5.54076731e-01 2.36279657e-03
-3.80995907e-02 1.36520281e-01 -8.25398266e-01 4.46252003e-02
8.64792824e-01 1.06546664e+00 -1.00383782e+00 9.89297092e-01
-2.02817455e-01 7.23755479e-01 -3.30138266e-01 1.04019143e-01
4.71753269e-01 2.26759419e-01 1.60844281e-01 1.46638846e+00
2.43542224e-01 -3.48192453e-02 2.51297742e-01 6.29687011e-01
-4.79241103e-01 -9.36949551e-02 -8.93311560e-01 -1.36334479e-01
3.51127058e-01 1.18745995e+00 -6.12267554e-01 -7.22428143e-01
-8.87001395e-01 1.11177886e+00 5.01632214e-01 5.10897458e-01
-7.67463207e-01 -3.91588777e-01 5.51084936e-01 3.52613449e-01
6.47202313e-01 -1.67388543e-01 1.56902343e-01 -1.48980069e+00
-9.89664495e-02 -7.56282806e-01 3.97987336e-01 -1.07431328e+00
-1.49382198e+00 5.16080618e-01 3.58105227e-02 -1.18749058e+00
-4.92409557e-01 -9.25569654e-01 -3.69839847e-01 5.52383423e-01
-1.62077379e+00 -1.29108071e+00 -4.40549940e-01 9.98224556e-01
1.00739491e+00 -2.54979581e-01 7.37674415e-01 1.71544403e-01
-1.56142145e-01 4.07571763e-01 1.94042280e-01 6.11286700e-01
1.07723641e+00 -1.21199107e+00 4.63063687e-01 7.86488056e-01
5.85995257e-01 2.66499817e-01 6.80513203e-01 -3.69817287e-01
-1.32262361e+00 -1.41543984e+00 5.71322501e-01 -7.96707392e-01
9.94261205e-01 -6.23273969e-01 -7.77280450e-01 1.06544209e+00
4.19068038e-01 5.11640608e-01 4.33866173e-01 -9.70359221e-02
-7.75076807e-01 -9.63619724e-02 -7.20479786e-01 3.66182327e-01
1.05506563e+00 -7.34595835e-01 -7.09787607e-01 6.69074118e-01
7.22947419e-01 -1.17824994e-01 -5.22614956e-01 2.81191673e-02
3.31458241e-01 -6.71855867e-01 1.02378726e+00 -8.55386615e-01
7.74230778e-01 -2.32897937e-01 -2.56338060e-01 -1.06105030e+00
-2.55237103e-01 -2.75368482e-01 1.26393974e-01 1.13136137e+00
4.32603091e-01 -4.71272171e-01 7.12309182e-01 5.13576686e-01
-3.22571285e-02 -3.95231754e-01 -6.13267422e-01 -8.10007393e-01
-6.28362820e-02 -3.88837576e-01 1.47155821e-01 9.76051807e-01
-3.64396960e-01 5.55435061e-01 -4.69121695e-01 -5.97581733e-03
7.34295011e-01 1.06850624e-01 9.61850166e-01 -1.02263761e+00
-4.03168797e-01 -1.19482189e-01 -5.99375069e-01 -1.14869654e+00
3.13322514e-01 -1.04850328e+00 1.97346538e-01 -1.54492342e+00
4.12629187e-01 -3.01121503e-01 -3.31608862e-01 7.90250778e-01
7.39016756e-02 7.18179762e-01 3.31638157e-01 4.98730630e-01
-9.78433251e-01 2.75640488e-01 7.88417578e-01 -4.16020006e-01
2.05660775e-01 -2.10347295e-01 -5.92621922e-01 7.11346447e-01
8.30773175e-01 -5.71432054e-01 -4.47193891e-01 -7.90287912e-01
-1.91193521e-01 -2.58118212e-01 6.71871662e-01 -9.87501681e-01
2.38597095e-01 -1.51274040e-01 4.98354197e-01 -3.18500161e-01
3.89457494e-01 -7.41298437e-01 -2.74070859e-01 1.91822961e-01
-6.27875328e-01 -2.02926755e-01 7.11783320e-02 7.34452486e-01
-2.09857672e-01 -4.38740909e-01 8.07395279e-01 -3.27630579e-01
-1.27989817e+00 4.62615967e-01 -2.93644130e-01 2.14003846e-01
1.14903855e+00 -8.44707489e-02 -5.36546290e-01 -4.18786377e-01
-6.68719471e-01 5.44542484e-02 7.26789415e-01 7.22506404e-01
3.65061641e-01 -1.05556858e+00 -5.95365167e-01 -5.31198410e-03
6.09835684e-01 -3.03888291e-01 -1.00892991e-01 4.37002003e-01
-4.33176577e-01 7.01874852e-01 -3.71014208e-01 -7.82632589e-01
-1.14059401e+00 9.07455742e-01 1.85892791e-01 4.51726615e-02
-7.44304836e-01 7.24582672e-01 4.58932519e-01 -2.13022962e-01
2.22867906e-01 -1.86967909e-01 -2.25405227e-02 -2.43985027e-01
7.72702813e-01 -1.43301547e-01 -1.64466456e-01 -7.44485795e-01
-3.24816614e-01 5.52973032e-01 -2.14451402e-01 -1.39091507e-01
1.39182734e+00 -1.55296832e-01 1.20809034e-01 4.40576255e-01
1.39317667e+00 -4.44871426e-01 -1.62486303e+00 -4.99523550e-01
1.53587401e-01 -5.18964946e-01 -4.46703769e-02 -7.01678097e-01
-9.92124140e-01 1.06876945e+00 5.73626935e-01 -4.66430038e-02
9.68602240e-01 5.14358222e-01 3.63040149e-01 8.52955878e-01
4.25749421e-01 -1.07943881e+00 3.23504657e-01 5.59771001e-01
7.50837505e-01 -1.66642845e+00 -2.05355525e-01 -9.76958275e-02
-7.85092056e-01 1.04817045e+00 5.41537464e-01 -3.89824063e-01
5.88044584e-01 -1.63055714e-02 4.52336334e-02 -8.51819664e-02
-8.50845337e-01 -4.88524228e-01 2.61675626e-01 7.76116133e-01
3.69058311e-01 -1.82176009e-01 1.75331920e-01 1.35609761e-01
6.61208779e-02 1.95724487e-01 5.85171819e-01 8.96189451e-01
-5.21381855e-01 -8.77142191e-01 -8.87808129e-02 6.51512444e-01
-5.08095980e-01 -4.13338661e-01 -3.05306107e-01 6.56812668e-01
-5.67232743e-02 6.95834994e-01 3.37078154e-01 -6.97254464e-02
-9.23168734e-02 1.41296864e-01 4.22364205e-01 -1.10797393e+00
-9.48583558e-02 -2.72990793e-01 -1.96870342e-02 -7.76968122e-01
-7.00145602e-01 -6.16797090e-01 -9.29315507e-01 -1.60997622e-02
-1.29306670e-02 -2.09393725e-01 8.89221430e-01 9.38963056e-01
4.75630760e-01 1.03377759e-01 3.56378198e-01 -1.11893451e+00
-4.80338186e-01 -9.78743970e-01 -2.52874911e-01 7.15861142e-01
3.72525930e-01 -5.96210063e-01 -2.93136716e-01 7.11137176e-01] | [10.084856033325195, 1.826259970664978] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.