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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0eaed08f-cee3-4e69-9eac-6ee6847d478a | energy-minimization-for-active-ris-aided-uav | 2306.10233 | null | https://arxiv.org/abs/2306.10233v1 | https://arxiv.org/pdf/2306.10233v1.pdf | Energy Minimization for Active RIS-Aided UAV-Enabled SWIPT Systems | In this paper, we consider an active reconfigurable intelligent surface (RIS)-aided unmanned aerial vehicle(UAV)-enabled simultaneous wireless information and power transfer(SWIPT) system with multiple ground users. Compared with the conventional passive RIS, the active RIS deploying the internally integrated amplifiers can offset part of the multiplicative fading. In this system, we deal with an optimization problem of minimizing the total energy cost of the UAV. Specifically, we alternately optimize the trajectories, the hovering time, and the reflection vectors at the active RIS by using the successive convex approximation (SCA) method. Simulation results show that the active RIS performs better in energy saving than the conventional passive RIS. | ['Jiangzhou Wang', 'Zhenkun Zhang', 'Cunhua Pan', 'Ruijing Liu', 'Zhangjie Peng'] | 2023-06-17 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [ 7.83066630e-01 7.41781235e-01 -9.33909044e-02 7.04189003e-01
-6.38406053e-02 -9.62323844e-01 2.39495739e-01 -4.01352078e-01
-8.15552324e-02 7.00933933e-01 -3.78307015e-01 -4.35675442e-01
-5.99526584e-01 -8.63057911e-01 -2.88796902e-01 -1.16987455e+00
-5.33645809e-01 -4.27036196e-01 -2.21171722e-01 -4.63166267e-01
-4.62957501e-01 2.87704587e-01 -8.55136096e-01 -9.70207632e-01
1.10308266e+00 1.30182457e+00 9.52457339e-02 5.12305558e-01
9.18122649e-01 4.69285160e-01 -7.92435646e-01 1.38172274e-02
6.93978846e-01 -2.98710287e-01 -9.46600959e-02 -1.80843212e-02
-2.37398610e-01 -5.88305950e-01 -5.71079910e-01 9.72269773e-01
5.82126379e-01 7.69149289e-02 4.14364219e-01 -1.52995574e+00
-7.75535703e-02 4.23418194e-01 -8.19564521e-01 -1.39377192e-01
5.53199500e-02 -3.46921325e-01 4.74055588e-01 -2.55652457e-01
9.77499261e-02 4.94361818e-01 4.37466592e-01 8.22852738e-03
-5.05896151e-01 -7.40389168e-01 5.22530526e-02 -1.07179821e-01
-1.72152889e+00 -5.48043489e-01 5.05153298e-01 2.27233648e-01
4.46928948e-01 6.15116417e-01 1.37529373e+00 1.69136360e-01
3.98338169e-01 4.57077563e-01 5.72632790e-01 -5.19425690e-01
3.12748134e-01 -2.58905649e-01 -3.35147083e-01 1.05875969e+00
1.03416061e+00 3.54167730e-01 -7.60217011e-02 -4.67271477e-01
5.92332780e-01 -1.36158451e-01 -6.38805449e-01 -2.74491638e-01
-1.17226601e+00 3.49023163e-01 5.73706508e-01 4.93074842e-02
-7.24009156e-01 3.33770871e-01 -5.06045341e-01 4.76248056e-01
4.73555535e-01 6.02960944e-01 -2.57806391e-01 3.73270035e-01
-5.85808575e-01 -4.52565670e-01 9.46616471e-01 1.51938701e+00
2.44064793e-01 5.07145107e-01 -1.53536096e-01 6.90888464e-02
6.91951334e-01 1.29922628e+00 -1.96307227e-01 -8.32609594e-01
1.53725863e-01 3.86939943e-01 5.51243484e-01 -7.83604205e-01
-5.52568734e-01 -9.97299612e-01 -8.37773502e-01 -8.84671882e-02
-1.70578957e-01 -1.36742294e+00 -5.81137002e-01 1.32501137e+00
2.74172693e-01 1.10992923e-01 6.44739091e-01 6.74821377e-01
2.63180315e-01 9.50090468e-01 -3.80035818e-01 -1.05402684e+00
9.40103829e-01 -7.91556358e-01 -9.40478325e-01 -4.14201677e-01
1.31668970e-01 -4.42363530e-01 -2.00038984e-01 1.51158273e-01
-9.38681602e-01 3.60465020e-01 -1.74900758e+00 7.05303609e-01
1.97646581e-02 8.03987026e-01 3.88317555e-01 8.48415434e-01
-1.16868961e+00 -1.36877820e-01 -6.35348618e-01 -1.71183288e-01
1.15555681e-01 7.47819185e-01 2.70586938e-01 2.61313260e-01
-9.09254134e-01 5.89930058e-01 -8.12477060e-03 2.85859704e-01
-6.29918158e-01 -5.90799391e-01 -4.76991206e-01 -4.15209606e-02
6.42292917e-01 -9.28809583e-01 1.16528165e+00 -1.02300882e+00
-2.13949180e+00 1.39896227e-02 3.50996584e-01 -5.69273829e-01
5.38953170e-02 -3.60799432e-02 -6.15157075e-02 3.25073034e-01
-4.95306611e-01 1.18863270e-01 8.37664306e-01 -1.07363546e+00
-7.27406204e-01 -5.13213813e-01 3.56535941e-01 5.48503757e-01
-3.13780963e-01 -6.98374987e-01 7.30457157e-02 -3.95583183e-01
1.71888411e-01 -1.45272839e+00 -3.85300696e-01 -1.28215238e-01
-4.47366655e-01 3.45865935e-01 1.16587317e+00 -2.83990860e-01
9.21535611e-01 -2.14763403e+00 3.69809508e-01 2.38629699e-01
-1.42288674e-02 2.27333143e-01 -1.07284062e-01 3.04831505e-01
6.04525745e-01 -4.03825521e-01 1.58553626e-02 1.45844847e-01
-3.92600387e-01 -1.15883484e-01 -4.51760530e-01 7.57544577e-01
-4.22355145e-01 4.19993311e-01 -9.07853782e-01 1.86174259e-01
-1.76680669e-01 3.66967618e-01 -6.89088330e-02 6.75250813e-02
2.50710994e-01 3.59053075e-01 -9.68286335e-01 8.76268446e-01
6.75633669e-01 3.07504773e-01 4.58175540e-01 -4.07794029e-01
-6.08224690e-01 -4.99510258e-01 -5.45030296e-01 1.15219545e+00
-6.18582964e-01 5.51189303e-01 6.79583848e-01 -6.14572704e-01
6.62296951e-01 4.44503605e-01 8.61328781e-01 -4.82729614e-01
4.70574409e-01 1.20744817e-01 -1.67864531e-01 3.36491764e-02
5.53676605e-01 1.36839077e-01 -3.46490294e-02 5.45875616e-02
2.56359488e-01 -8.65721256e-02 -3.84840846e-01 1.02931999e-01
1.25946510e+00 -2.97814220e-01 6.87569380e-01 -4.51485395e-01
1.75010324e-01 4.82122600e-02 4.90751386e-01 3.03837150e-01
-8.46342836e-03 -4.84009504e-01 -2.14181706e-01 3.28112155e-01
-2.57411718e-01 -1.03703701e+00 2.28170231e-01 3.26734453e-01
9.81111586e-01 -1.85278460e-01 -4.86132056e-01 -3.04095507e-01
-1.95843905e-01 6.80603802e-01 1.70026608e-02 -4.39110935e-01
1.89264074e-01 -8.99379373e-01 4.44769830e-01 -2.68005699e-01
1.12627411e+00 5.30282855e-01 -1.29295301e+00 2.13910207e-01
-1.36970386e-01 -1.02222204e+00 -3.71441811e-01 1.07660592e-01
-5.30728400e-01 -8.21314692e-01 -6.91801310e-01 -3.58609766e-01
9.72520769e-01 1.15022969e+00 2.41120413e-01 -2.76726842e-01
-4.73322347e-02 1.27815545e+00 -4.37023431e-01 -1.18888354e+00
4.84139919e-01 -2.59745687e-01 3.68789315e-01 1.38717994e-01
-5.51867425e-01 -2.37655699e-01 -7.93631256e-01 4.78116721e-01
-2.08010510e-01 -3.78555357e-02 5.84704518e-01 5.18483937e-01
5.14593065e-01 5.30983090e-01 5.21185935e-01 -1.54828718e-02
3.35087270e-01 -7.13091850e-01 -1.13120830e+00 4.95085537e-01
-3.36230695e-01 -5.28343856e-01 3.49165946e-01 -2.15070024e-01
-9.66203094e-01 5.24779379e-01 5.08944452e-01 4.68382016e-02
7.42360592e-01 1.79539710e-01 -3.26858521e-01 -1.20409119e+00
4.10139784e-02 7.24198818e-02 4.18708660e-03 5.74437082e-01
2.24753544e-01 8.65734100e-01 1.16770022e-01 2.86679983e-01
1.18461549e+00 4.58674192e-01 6.00939274e-01 -1.56281471e+00
-5.73831022e-01 1.13433599e-02 1.54651895e-01 -8.61866355e-01
4.54894096e-01 -1.41333020e+00 -6.97517693e-01 3.55974644e-01
-8.99853468e-01 -4.20768201e-01 -3.84233862e-01 8.86585057e-01
-4.54811215e-01 3.18726785e-02 3.28574181e-01 -1.50084770e+00
-6.48699641e-01 -6.17031932e-01 8.29879284e-01 5.59513271e-01
2.60054231e-01 -6.11895978e-01 -8.49422067e-02 -9.90822911e-02
5.55467784e-01 7.14582920e-01 2.06074595e-01 2.83988982e-01
-9.88735080e-01 -5.19469619e-01 3.50020260e-01 -1.36722073e-01
2.35925123e-01 -3.64180595e-01 -6.76963687e-01 -7.25035667e-01
9.94739980e-02 4.38424200e-01 3.12528163e-01 6.09502494e-01
4.42206442e-01 -9.21374321e-01 -8.54748011e-01 6.08272016e-01
1.52059662e+00 7.21148908e-01 2.86061913e-01 -3.65928590e-01
8.15898255e-02 1.16624177e-01 1.05076623e+00 7.94001102e-01
2.71580607e-01 5.89205801e-01 9.26833153e-01 -9.89679247e-02
5.81190944e-01 -1.24287881e-01 7.90165722e-01 6.07231259e-01
-4.67010498e-01 -1.08926630e+00 -1.26113519e-01 7.20293447e-02
-1.71309674e+00 -3.37797552e-01 -8.93869027e-02 2.50269389e+00
-1.82582662e-01 -6.68724954e-01 -1.37473136e-01 -3.26521620e-02
4.56927299e-01 3.52504663e-02 -5.73354542e-01 2.65150279e-01
-6.62539825e-02 -3.00313026e-01 1.74580598e+00 2.92050600e-01
-9.52501357e-01 4.00247246e-01 6.29561663e+00 4.15773779e-01
-9.16362464e-01 2.66766131e-01 -9.40403491e-02 -1.81453094e-01
-2.28943661e-01 -2.10211739e-01 -2.18829945e-01 1.12614430e-01
8.79308522e-01 -8.23726535e-01 7.21462667e-01 4.27184045e-01
1.16341442e-01 -6.03307366e-01 -3.73490691e-01 7.93003380e-01
1.02100983e-01 -7.60880172e-01 -3.12879115e-01 4.76181358e-01
6.72438264e-01 -7.69940615e-02 6.86313733e-02 -2.06064597e-01
2.94207603e-01 3.12935375e-02 7.16417015e-01 6.98539734e-01
8.27193618e-01 -9.67571795e-01 5.30707538e-01 4.74168926e-01
-1.47235894e+00 -7.45827019e-01 -2.54911542e-01 -1.77333504e-01
1.77981451e-01 7.28106201e-01 -3.70912671e-01 1.17591739e+00
1.96923807e-01 2.81594962e-01 2.91285515e-02 1.05051911e+00
-2.53279656e-01 4.30801064e-01 -7.43773460e-01 -5.31083763e-01
5.21596037e-02 -6.08738303e-01 1.38664401e+00 1.77475676e-01
1.15247643e+00 1.04148960e+00 -3.88546512e-02 1.96091026e-01
1.45078553e-02 -1.80921420e-01 -1.32598805e+00 -2.22261667e-01
6.87451959e-01 1.29229248e+00 -5.03089845e-01 1.96325630e-01
-3.24946225e-01 1.00441849e+00 -5.89043319e-01 5.89320302e-01
-7.22811997e-01 -8.27129662e-01 6.64046109e-01 -2.12548554e-01
-4.58008312e-02 -8.87075722e-01 4.08348888e-02 -8.71996760e-01
-4.11744922e-01 8.57331753e-02 -5.50386645e-02 -6.49153054e-01
-4.15405750e-01 3.96456242e-01 -1.56649575e-01 -1.69277489e+00
2.83923149e-01 -9.60438699e-02 -5.53085446e-01 2.56788105e-01
-1.23214865e+00 -9.33718681e-01 -4.29241389e-01 4.28629398e-01
2.57242024e-02 -3.63142908e-01 7.49720097e-01 -3.45402956e-03
-5.96121788e-01 2.25634813e-01 4.12509769e-01 -6.45254314e-01
1.08190157e-01 -5.30448914e-01 -4.19420093e-01 1.04041183e+00
-4.28603351e-01 1.91981837e-01 7.04928279e-01 -6.08729184e-01
-2.60516191e+00 -1.27960181e+00 -1.68935165e-01 2.83743352e-01
4.87101644e-01 1.27674729e-01 5.61450958e-01 5.73095679e-01
8.71196985e-01 -4.81534749e-01 9.31053579e-01 -7.44816363e-01
7.21898973e-01 -3.17934155e-01 -1.48961031e+00 7.15567887e-01
1.11906898e+00 3.39061052e-01 2.46291399e-01 3.71879518e-01
7.08621979e-01 -2.99597859e-01 -6.84448838e-01 5.21702826e-01
6.43591523e-01 2.16366887e-01 7.98089027e-01 2.83568978e-01
-6.34583831e-01 -6.23172104e-01 -3.08639079e-01 -1.96715307e+00
-4.83235538e-01 -1.31136954e+00 -3.28124225e-01 9.62306857e-01
2.07847789e-01 -1.15729618e+00 2.29382187e-01 1.14096398e-03
-2.87518531e-01 -4.10288155e-01 -1.26441574e+00 -8.97630095e-01
-7.40537822e-01 4.12529588e-01 3.27698231e-01 4.17688489e-01
5.44427931e-01 6.28532529e-01 -5.34231484e-01 1.02936053e+00
1.08469450e+00 -3.38625163e-01 4.84516919e-01 -1.28804481e+00
4.81969677e-02 3.09820563e-01 -4.16308552e-01 -1.22868478e+00
-9.71072242e-02 -5.08982658e-01 2.31054962e-01 -1.54955292e+00
-6.00991547e-01 -2.23755643e-01 2.61842385e-02 3.80967110e-01
2.64597803e-01 1.79574996e-01 9.94822234e-02 -2.55870610e-01
-3.93057853e-01 9.32433963e-01 1.14723396e+00 -5.28008461e-01
-2.94873089e-01 6.03992522e-01 -7.07452595e-01 7.27760196e-01
8.10163200e-01 -9.49098095e-02 -8.50337505e-01 -4.17248160e-01
2.85255373e-01 5.20448267e-01 1.51606798e-01 -1.45748246e+00
4.87565666e-01 -1.93507254e-01 1.31076157e-01 -3.84371012e-01
4.07240480e-01 -1.57166314e+00 4.43132341e-01 8.73115301e-01
4.66377795e-01 -5.63805521e-01 1.28703102e-01 1.15582120e+00
2.19185218e-01 2.55344689e-01 7.53325224e-01 4.97721046e-01
-7.85907954e-02 4.60889339e-01 -1.19587624e+00 -7.10256517e-01
1.78711498e+00 -1.76179707e-02 -6.82101905e-01 -6.10915899e-01
-1.88647762e-01 7.08648026e-01 8.23807865e-02 1.35655344e-01
4.75315869e-01 -9.46003497e-01 -8.00911486e-02 1.06107406e-01
-1.88627169e-01 -4.02820975e-01 1.65411383e-01 1.22397137e+00
-3.21043096e-02 5.69887459e-01 -1.22383103e-01 -1.71671480e-01
-1.21265149e+00 -1.16341628e-01 4.28641945e-01 4.87864256e-01
-5.12175374e-02 6.08331323e-01 -1.32267028e-01 1.61964178e-01
-9.31631867e-03 6.92490116e-02 -1.87122017e-01 1.47579968e-01
2.45721266e-01 9.36410069e-01 -2.50299782e-01 -2.78691918e-01
-2.82859653e-01 7.68723011e-01 7.86201656e-01 5.20787248e-03
1.04085350e+00 -6.45602882e-01 -6.98909685e-02 -8.72174501e-02
3.86660367e-01 3.15340370e-01 -1.08185780e+00 1.56127185e-01
-7.00834632e-01 -2.93682098e-01 8.10684621e-01 -5.64782500e-01
-1.20665109e+00 -5.94918132e-02 9.42635477e-01 5.70072711e-01
1.53786683e+00 -4.87757921e-01 4.37395811e-01 9.37693596e-01
1.44688261e+00 -8.47659111e-01 -2.86176831e-01 3.50308955e-01
6.07258260e-01 -3.89035821e-01 1.92743495e-01 -8.74219418e-01
-1.87476978e-01 1.17103612e+00 2.11927354e-01 -1.73481211e-01
8.02711606e-01 5.58973670e-01 -2.94960588e-01 -5.70785366e-02
-5.43902993e-01 -3.54832083e-01 -9.12632570e-02 5.79737961e-01
-1.45332098e-01 4.43111211e-01 -5.12330532e-01 7.21836865e-01
4.86857295e-02 -8.55412986e-03 9.17941391e-01 1.25886750e+00
-7.42826939e-01 -5.92799485e-01 -2.82738984e-01 6.10314012e-01
-1.29310980e-01 3.96617085e-01 -7.03755558e-01 3.68084610e-01
1.19976653e-03 1.43034935e+00 1.58686079e-02 -7.10213006e-01
7.61806250e-01 -6.34380639e-01 7.82571197e-01 -1.66493312e-01
1.18342862e-01 -6.25834912e-02 2.68990636e-01 -3.53322953e-01
-5.18662333e-01 -3.95354629e-01 -9.11952734e-01 -1.45659223e-02
-9.98010814e-01 4.93962049e-01 7.64993966e-01 7.93244362e-01
6.16886973e-01 9.33807909e-01 1.45980513e+00 -6.87773705e-01
-1.41441137e-01 -5.30204833e-01 -9.90529060e-01 -1.52334332e+00
5.28311253e-01 -7.57611930e-01 -5.11969090e-01 -4.80150342e-01] | [5.935178756713867, 1.4852954149246216] |
6b80ba10-7f1a-4c26-8177-c96b4edd1a98 | spatio-temporal-attention-mechanism-and | 2108.03543 | null | https://arxiv.org/abs/2108.03543v1 | https://arxiv.org/pdf/2108.03543v1.pdf | Spatio-Temporal Attention Mechanism and Knowledge Distillation for Lip Reading | Despite the advancement in the domain of audio and audio-visual speech recognition, visual speech recognition systems are still quite under-explored due to the visual ambiguity of some phonemes. In this work, we propose a new lip-reading model that combines three contributions. First, the model front-end adopts a spatio-temporal attention mechanism to help extract the informative data from the input visual frames. Second, the model back-end utilizes a sequence-level and frame-level Knowledge Distillation (KD) techniques that allow leveraging audio data during the visual model training. Third, a data preprocessing pipeline is adopted that includes facial landmarks detection-based lip-alignment. On LRW lip-reading dataset benchmark, a noticeable accuracy improvement is demonstrated; the spatio-temporal attention, Knowledge Distillation, and lip-alignment contributions achieved 88.43%, 88.64%, and 88.37% respectively. | ['Nourhan Sakr', 'Omar Abugabal', 'Hadeel Mabrouk', 'Farah Eldeshnawy', 'Hesham M. Eraqi', 'Marian Ramsis', 'Shahd Elashmawy'] | 2021-08-07 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 1.58738151e-01 4.73340265e-02 -4.44223344e-01 -4.01757985e-01
-1.12889314e+00 -1.17856577e-01 6.29577160e-01 -2.17082977e-01
-3.79074663e-01 4.54860955e-01 5.58129489e-01 -1.51344389e-01
3.81779432e-01 -1.91850275e-01 -5.48575759e-01 -5.84028602e-01
2.46036425e-01 -1.13157406e-01 3.72479528e-01 2.90899128e-01
4.01905090e-01 4.25614059e-01 -1.90781856e+00 4.61326391e-01
6.86913192e-01 1.23624766e+00 1.51297331e-01 9.84368920e-01
-3.57204944e-01 5.61157584e-01 -2.16675624e-01 -8.48336667e-02
-5.45459427e-02 -2.23829120e-01 -7.23904908e-01 1.16201676e-01
6.04675055e-01 -4.94814575e-01 -1.93218827e-01 9.31989610e-01
9.80960488e-01 2.04649270e-01 5.56200147e-01 -1.44144952e+00
-4.91179854e-01 6.46171486e-03 -7.68110633e-01 9.37443674e-02
4.85960424e-01 4.43390310e-01 8.01675022e-01 -1.33988464e+00
6.21742070e-01 1.35465062e+00 4.56505984e-01 6.85604930e-01
-8.44945371e-01 -7.07687616e-01 1.25285566e-01 7.37026215e-01
-1.44582283e+00 -1.19749951e+00 1.08600092e+00 -4.37489003e-01
1.11979985e+00 1.02196835e-01 7.00053513e-01 8.37340474e-01
-3.06502074e-01 1.06436014e+00 1.00516462e+00 -6.71683013e-01
3.24801616e-02 8.55882540e-02 -1.34074628e-01 5.63986301e-01
-3.81505579e-01 2.54124045e-01 -1.03151512e+00 3.58456224e-01
4.77685720e-01 -3.24542612e-01 -3.21576804e-01 -3.29063594e-01
-8.12054038e-01 4.24785793e-01 1.71816677e-01 8.70777443e-02
-3.75881642e-01 -5.30273952e-02 6.51518047e-01 -3.09210747e-01
3.55895817e-01 -3.18402708e-01 -2.58968621e-01 -3.99949193e-01
-1.32237220e+00 -3.19395721e-01 3.01253051e-01 9.49093401e-01
5.86768210e-01 2.79923648e-01 -3.74928981e-01 1.12143338e+00
9.01307821e-01 4.86989111e-01 5.66505075e-01 -7.72428930e-01
5.47760487e-01 2.97680914e-01 -2.57757485e-01 -7.73851991e-01
-2.89466888e-01 2.51516346e-02 -6.07287049e-01 4.10730749e-01
6.24307357e-02 1.78166196e-01 -1.13753140e+00 1.68042386e+00
6.04816973e-01 5.50816059e-01 7.87784755e-02 8.84011209e-01
1.13095069e+00 5.35108805e-01 4.15031075e-01 -3.34269643e-01
1.45288825e+00 -1.18443966e+00 -9.77173686e-01 -9.51639414e-02
8.18166733e-02 -9.40637767e-01 1.10747838e+00 2.25202605e-01
-1.32330680e+00 -9.84108925e-01 -9.58187938e-01 -5.37149727e-01
-3.71627182e-01 4.68075275e-01 1.30296558e-01 5.36301374e-01
-1.18941879e+00 -2.15058327e-01 -5.69540679e-01 -3.30031931e-01
6.94494486e-01 2.20936581e-01 -4.62182790e-01 5.35428198e-03
-7.18402445e-01 5.62012970e-01 2.34386966e-01 1.12459503e-01
-9.38969314e-01 -7.14467645e-01 -1.00736320e+00 4.15656492e-02
3.22782934e-01 -4.93602663e-01 1.28742206e+00 -9.56496477e-01
-2.06576324e+00 1.00436556e+00 -7.55386114e-01 -3.13179910e-01
4.89113212e-01 -7.48516023e-02 -4.65548754e-01 5.23776591e-01
-1.96404159e-01 1.12969577e+00 1.19735408e+00 -1.30112612e+00
-7.48920918e-01 -2.17626750e-01 -3.88865918e-01 4.00735438e-01
-1.60450935e-01 1.61877900e-01 -9.19297695e-01 -6.92272067e-01
-5.69596933e-03 -5.54586351e-01 4.86648560e-01 3.02190185e-01
-4.02280480e-01 -3.59130949e-01 1.13200212e+00 -1.07312310e+00
1.14775074e+00 -2.32024980e+00 -7.98425451e-02 -1.46978080e-01
-1.99452303e-02 8.40873361e-01 -2.62987942e-01 8.22910890e-02
-6.41095564e-02 1.18928634e-01 1.25514895e-01 -8.81801844e-01
-6.21541962e-02 -1.31595239e-01 -3.09777588e-01 2.81141579e-01
4.42089349e-01 1.06773448e+00 -6.01691604e-01 -9.16633964e-01
6.68662131e-01 8.21080804e-01 -5.44510126e-01 2.66086876e-01
-1.37645572e-01 1.88445017e-01 8.12654197e-03 9.95502710e-01
8.19195151e-01 2.62871832e-01 -1.42510742e-01 -3.59280646e-01
-2.34059125e-01 3.05764526e-01 -9.77861047e-01 1.78482866e+00
-4.34761405e-01 9.76797402e-01 2.32776657e-01 -6.74968600e-01
7.18031406e-01 5.45260191e-01 3.36368322e-01 -9.85878289e-01
4.91976403e-02 5.22707589e-03 -3.14410836e-01 -8.83734047e-01
3.99941117e-01 1.30617023e-01 5.31558454e-01 1.29233092e-01
1.38759121e-01 8.94639790e-02 -2.81582028e-01 -3.61375511e-02
4.36602712e-01 3.98486316e-01 3.69063079e-01 2.06969559e-01
1.08086479e+00 -4.55101877e-01 5.14620543e-01 3.31269167e-02
-6.23100638e-01 7.11220264e-01 2.63788611e-01 -6.96103200e-02
-8.36210012e-01 -1.01468527e+00 6.22892380e-02 1.15609944e+00
7.15716407e-02 -5.07235527e-01 -8.24107587e-01 -5.92781007e-01
-1.98058292e-01 4.98524636e-01 -4.24456716e-01 -1.66185677e-01
-3.31492066e-01 5.21118492e-02 6.92190289e-01 6.10267162e-01
7.82174587e-01 -1.12822199e+00 -5.68767905e-01 -6.10401891e-02
-1.70285076e-01 -1.27161956e+00 -6.78660214e-01 -4.82314348e-01
-4.67168480e-01 -1.02103877e+00 -8.54691446e-01 -9.88444865e-01
5.38789451e-01 3.64170223e-01 5.80131233e-01 -8.75564888e-02
-4.28291023e-01 4.46719199e-01 -1.45668715e-01 -4.56206024e-01
-3.00685376e-01 -1.61202073e-01 -7.48559088e-02 2.39963531e-01
3.93471181e-01 -4.32129204e-01 -6.57255650e-01 2.04141155e-01
-4.68826205e-01 3.24173599e-01 6.12582147e-01 8.01674247e-01
8.00949335e-01 -5.48311412e-01 7.54087925e-01 1.24248996e-01
4.20920968e-01 -1.13422908e-01 -4.70198274e-01 4.77684975e-01
-4.10758972e-01 -2.05794588e-01 2.84952104e-01 -6.22851789e-01
-1.16383719e+00 2.14182943e-01 -3.90295655e-01 -6.17506504e-01
-5.21537304e-01 -2.99647488e-02 -4.71838713e-01 -6.80410862e-02
5.70062324e-02 4.78596032e-01 2.56335855e-01 -7.12853491e-01
4.34999615e-01 1.10965288e+00 7.65001774e-01 -1.84997603e-01
4.44149584e-01 4.19550866e-01 -2.27694273e-01 -1.28595293e+00
-2.72275686e-01 -5.90328217e-01 -5.46459854e-01 -4.17156726e-01
1.08610129e+00 -9.30115998e-01 -1.06916416e+00 7.23780870e-01
-1.24131882e+00 -2.03080311e-01 -1.56561121e-01 5.05254984e-01
-7.32005358e-01 6.48855746e-01 -1.75895378e-01 -1.13014877e+00
-5.64279675e-01 -1.37198448e+00 1.27818751e+00 4.16759551e-01
5.80584593e-02 -4.45169061e-01 -5.66724874e-02 6.71442509e-01
3.85400802e-01 -2.31080651e-01 7.48031676e-01 -5.28218389e-01
-6.30369365e-01 1.38914347e-01 -5.67057490e-01 4.43304271e-01
7.64596909e-02 2.61460781e-01 -1.44192338e+00 -4.42204140e-02
-6.31915569e-01 -3.98419023e-01 7.43278325e-01 4.84102607e-01
1.10946620e+00 -3.06691289e-01 -2.61212409e-01 6.41072094e-01
8.97986591e-01 3.69191766e-01 6.70278788e-01 -1.05827205e-01
5.98263443e-01 6.65416777e-01 6.65919065e-01 2.70874530e-01
6.17847323e-01 1.09002483e+00 4.16968524e-01 -2.51938701e-01
-1.00598192e+00 -6.38507307e-01 3.79661918e-01 7.51463711e-01
1.02538653e-01 1.66102983e-02 -9.38577175e-01 6.94166958e-01
-1.66842294e+00 -9.94447112e-01 3.85219097e-01 2.05898571e+00
8.96453619e-01 -1.19751029e-01 3.06450039e-01 2.83255368e-01
7.51354992e-01 1.89722911e-01 -5.93980551e-01 -3.15977216e-01
2.59505659e-02 1.41825169e-01 -8.25789664e-03 9.13761795e-01
-1.00071216e+00 1.24583793e+00 5.74943495e+00 9.74988222e-01
-1.44702244e+00 8.69428366e-02 5.71371734e-01 -8.40104073e-02
7.49446675e-02 -3.45630616e-01 -8.92967999e-01 4.20265555e-01
6.67282939e-01 7.58035854e-02 2.77745098e-01 7.32025146e-01
6.39655948e-01 -3.26961011e-01 -7.83024490e-01 1.42031598e+00
3.69049072e-01 -1.17386389e+00 5.70086204e-02 8.28828141e-02
3.09725821e-01 -1.50781736e-01 2.12887034e-01 1.32726327e-01
-3.52064639e-01 -1.13360906e+00 7.41703033e-01 7.09125817e-01
1.23756909e+00 -8.87692928e-01 2.99056590e-01 1.88465551e-01
-1.51191354e+00 8.52218270e-02 3.29580829e-02 3.72381747e-01
2.63752341e-01 -4.35750559e-03 -1.33358586e+00 4.16632444e-01
6.99788451e-01 5.31556666e-01 -4.26298112e-01 1.09131324e+00
-3.65119845e-01 6.11553729e-01 -3.13713431e-01 2.09381014e-01
-1.34441286e-01 2.78054118e-01 4.67238009e-01 1.38855898e+00
5.41331507e-02 -1.84165269e-01 -3.66155766e-02 5.20158827e-01
-1.30609358e-02 1.69504657e-01 -4.75335211e-01 1.49153918e-01
5.47149718e-01 1.23831224e+00 -2.11529523e-01 -1.95468798e-01
-3.70916992e-01 7.82430232e-01 1.55141950e-01 5.89159667e-01
-9.83246326e-01 -4.60938394e-01 1.01775265e+00 4.00994495e-02
3.25597227e-01 -2.08749950e-01 -2.01026961e-01 -6.98525369e-01
1.60888702e-01 -8.44910502e-01 -7.18251839e-02 -1.08462358e+00
-5.86976409e-01 3.46951962e-01 -2.03582510e-01 -9.97644365e-01
-5.00267625e-01 -5.12455761e-01 -4.89699900e-01 9.28102076e-01
-2.10692406e+00 -1.46995747e+00 -4.24213886e-01 7.75039852e-01
1.04180181e+00 -3.45226735e-01 6.40766263e-01 2.73138791e-01
-5.63951552e-01 1.03110266e+00 -2.99357355e-01 3.63230675e-01
8.04809749e-01 -7.14397430e-01 1.41126812e-01 8.59599650e-01
2.13594019e-01 3.51873606e-01 1.62343800e-01 -4.13945824e-01
-1.29147911e+00 -9.37815428e-01 1.11751819e+00 1.20745331e-01
2.05046371e-01 -3.07117969e-01 -8.20265412e-01 3.15543681e-01
4.27905202e-01 2.18974352e-01 6.86758518e-01 -4.05799657e-01
-5.15437424e-01 -3.26735139e-01 -1.21573234e+00 6.42823935e-01
8.84739637e-01 -1.00613046e+00 -5.24454176e-01 -1.98076203e-01
6.11784399e-01 -1.91340983e-01 -5.64028502e-01 3.47706646e-01
8.00978124e-01 -8.34473729e-01 1.23579693e+00 -4.97503400e-01
-2.67763194e-02 -3.92510504e-01 -1.63335755e-01 -8.04460168e-01
1.14092402e-01 -8.32162082e-01 -3.65699351e-01 1.73408532e+00
9.08612832e-02 -3.92598033e-01 4.78308916e-01 2.62843996e-01
-9.43260789e-02 -7.55096018e-01 -1.36559880e+00 -4.19195563e-01
-3.83640349e-01 -6.94496512e-01 4.56806660e-01 5.99499702e-01
1.72771141e-01 3.55648309e-01 -4.36953276e-01 8.94840658e-02
3.78454685e-01 -2.56684065e-01 8.09266806e-01 -9.48409021e-01
-4.89554629e-02 -5.70444345e-01 -4.31976736e-01 -1.27375340e+00
3.04541618e-01 -6.12855554e-01 1.39847323e-01 -1.53280437e+00
-7.27451965e-02 5.61156136e-04 -1.45475268e-01 6.26747489e-01
-3.95588465e-02 2.16737941e-01 4.20680195e-01 -1.72073357e-02
-5.19159377e-01 6.08113289e-01 1.04821312e+00 -1.54766843e-01
-2.88109660e-01 -7.95049071e-02 -2.28026450e-01 7.94521987e-01
8.17015469e-01 -2.39649080e-02 -5.30024827e-01 -3.67133886e-01
-3.75347316e-01 1.89936310e-01 6.37129486e-01 -9.26465034e-01
5.88819146e-01 -2.16034632e-02 3.15451741e-01 -1.02823806e+00
8.39387298e-01 -6.99278474e-01 -2.84651011e-01 1.77373201e-01
-3.75048280e-01 -1.99599087e-01 4.99253720e-01 5.68095148e-01
-4.10308301e-01 1.10604078e-01 1.06131637e+00 4.09771144e-01
-8.96367610e-01 2.72126824e-01 -2.76885957e-01 7.60031939e-02
1.21913505e+00 -5.74743092e-01 -2.56165534e-01 -3.77965868e-01
-5.88610172e-01 3.47198471e-02 1.89828411e-01 6.08075678e-01
1.01423407e+00 -1.33903968e+00 -4.40889835e-01 5.78388870e-01
8.77993926e-02 -5.58303408e-02 5.33284307e-01 9.29845393e-01
-1.20083824e-01 5.03335297e-01 -2.16660574e-01 -6.92493558e-01
-1.80206954e+00 5.24132669e-01 2.38361806e-01 2.24223077e-01
-4.79956955e-01 9.04223382e-01 -5.02425199e-03 1.58673421e-01
8.70321631e-01 -2.94369668e-01 -2.96759188e-01 2.22981483e-01
7.15851188e-01 6.01849914e-01 -9.83954743e-02 -1.01494944e+00
-4.97677863e-01 1.01239324e+00 1.66742280e-01 -4.49704051e-01
9.34862733e-01 -4.21328992e-01 3.88420135e-01 2.33798295e-01
1.27145565e+00 1.72344908e-01 -1.53759241e+00 -2.32692987e-01
-9.60941091e-02 -4.40771967e-01 8.79582167e-02 -1.00097227e+00
-8.95261288e-01 1.62261868e+00 7.52570331e-01 -2.85408437e-01
1.40290952e+00 -1.32110372e-01 6.39306366e-01 -1.40841693e-01
-1.52867749e-01 -1.29663932e+00 2.16211244e-01 4.14353848e-01
1.00262916e+00 -1.24246633e+00 -3.85657817e-01 -3.84119719e-01
-6.68130636e-01 1.04584551e+00 5.84240615e-01 5.28978765e-01
5.87040246e-01 2.27544561e-01 8.50255787e-02 3.66631836e-01
-7.44105577e-01 -5.68409204e-01 6.22902095e-01 1.09924519e+00
2.13712811e-01 -2.89220572e-01 4.72254567e-02 5.14961064e-01
1.52103812e-01 2.83835649e-01 -1.51758105e-01 5.68963528e-01
-5.72049141e-01 -7.87781119e-01 -2.80013770e-01 -2.82890648e-01
-3.04481417e-01 -3.01484227e-01 -3.49053949e-01 6.38237238e-01
1.57015368e-01 1.13522387e+00 -3.97487730e-02 -3.37122291e-01
1.64572448e-01 5.93188524e-01 1.55182183e-01 -3.17766517e-01
-2.49989197e-01 4.51260626e-01 -5.55588230e-02 -6.64792836e-01
-3.47582102e-01 -5.25627196e-01 -1.44753134e+00 -1.03304209e-02
-1.06966712e-01 -2.03238145e-01 9.87948954e-01 6.92992032e-01
8.66028666e-01 4.38861489e-01 4.30005699e-01 -1.00533509e+00
-2.44556308e-01 -9.77215886e-01 -5.28893135e-02 7.13019967e-02
7.04239726e-01 -7.44482934e-01 -2.25205705e-01 3.17872673e-01] | [14.314133644104004, 5.0015411376953125] |
4fa171ff-882c-4a7b-b1f1-b6bd9984f2e9 | learning-deep-video-stabilization-without | 2011.09697 | null | https://arxiv.org/abs/2011.09697v2 | https://arxiv.org/pdf/2011.09697v2.pdf | Deep Motion Blind Video Stabilization | Despite the advances in the field of generative models in computer vision, video stabilization still lacks a pure regressive deep-learning-based formulation. Deep video stabilization is generally formulated with the help of explicit motion estimation modules due to the lack of a dataset containing pairs of videos with similar perspective but different motion. Therefore, the deep learning approaches for this task have difficulties in the pixel-level synthesis of latent stabilized frames, and resort to motion estimation modules for indirect transformations of the unstable frames to stabilized frames, leading to the loss of visual content near the frame boundaries. In this work, we aim to declutter this over-complicated formulation of video stabilization with the help of a novel dataset that contains pairs of training videos with similar perspective but different motion, and verify its effectiveness by successfully learning motion blind full-frame video stabilization through employing strictly conventional generative techniques and further improve the stability through a curriculum-learning inspired adversarial training strategy. Through extensive experimentation, we show the quantitative and qualitative advantages of the proposed approach to the state-of-the-art video stabilization approaches. Moreover, our method achieves $\sim3\times$ speed-up over the currently available fastest video stabilization methods. | ['Tae Hyun Kim', 'Sangjoon Yu', 'Muhammad Kashif Ali'] | 2020-11-19 | null | null | null | null | ['video-stabilization'] | ['computer-vision'] | [ 1.05928116e-01 -7.71943852e-02 -5.33924736e-02 1.56573385e-01
-6.45255148e-01 -3.54483366e-01 7.02533305e-01 -2.99128443e-01
-1.69564486e-01 7.41878748e-01 2.37322897e-01 -1.34694561e-01
2.33309045e-01 -3.86435211e-01 -1.08700180e+00 -1.12851131e+00
1.29188761e-01 -2.41359603e-03 2.02563211e-01 -4.58440214e-01
1.35036856e-01 2.71905243e-01 -1.52538788e+00 2.14001513e-03
6.27814293e-01 8.20573747e-01 4.76129688e-02 5.91691971e-01
2.27854252e-01 9.88537729e-01 -1.33018851e-01 -2.94527411e-01
3.60564232e-01 -7.72057533e-01 -5.39272130e-01 5.20873070e-01
6.91146970e-01 -5.18153250e-01 -5.36964059e-01 1.01111865e+00
5.30106187e-01 3.63015920e-01 4.43722814e-01 -9.79653358e-01
-5.04135966e-01 1.72482178e-01 -5.59224308e-01 1.99742928e-01
4.41544235e-01 2.45271951e-01 6.37690544e-01 -7.63386965e-01
8.63524377e-01 1.21302629e+00 5.89559078e-01 8.59980464e-01
-1.37403882e+00 -3.93270642e-01 1.48083076e-01 2.70679533e-01
-1.04426610e+00 -7.10711479e-01 9.94674146e-01 -5.53052723e-01
6.01771653e-01 -7.59078488e-02 5.24171650e-01 1.26081431e+00
1.11652024e-01 5.81509233e-01 8.55262756e-01 -4.90525901e-01
1.65513292e-01 -2.51446545e-01 -3.29880655e-01 8.51210237e-01
-1.74491983e-02 3.12012374e-01 -4.73143518e-01 3.85808229e-01
9.74512041e-01 1.26911644e-02 -5.52562654e-01 -7.91398108e-01
-1.23526990e+00 7.29535937e-01 3.99471641e-01 3.48925352e-01
-1.47269905e-01 4.32005018e-01 2.93611169e-01 1.46408498e-01
4.20008361e-01 2.13265777e-01 -9.67505798e-02 1.50098622e-01
-1.30321491e+00 3.23067635e-01 3.94628763e-01 8.84138823e-01
8.44444215e-01 8.00797403e-01 -1.37002945e-01 2.04518139e-01
3.15629154e-01 2.76618928e-01 5.10333776e-01 -1.24875391e+00
3.52415442e-01 3.09682608e-01 3.58201027e-01 -1.21045637e+00
-1.24474838e-01 -3.65223318e-01 -1.02921557e+00 5.12403846e-01
5.85814536e-01 -1.88814014e-01 -6.94068015e-01 1.99349511e+00
4.04135764e-01 4.91983682e-01 1.24438882e-01 9.95305419e-01
4.09521341e-01 7.15967536e-01 -1.32369995e-01 -6.95969820e-01
9.62466300e-01 -1.17235410e+00 -7.48636603e-01 -1.09917233e-02
3.46173644e-01 -8.77251267e-01 9.89386559e-01 2.07599357e-01
-1.37664998e+00 -7.90015101e-01 -1.00202739e+00 -1.43670127e-01
1.62257254e-01 -4.49648462e-02 2.58262753e-01 4.96501505e-01
-1.27380657e+00 7.10307658e-01 -1.10944629e+00 -2.31917188e-01
2.55297601e-01 3.25935960e-01 -3.89112532e-01 4.87333238e-02
-9.94701862e-01 7.48886049e-01 9.44954157e-02 1.43674463e-01
-1.27397168e+00 -7.49181867e-01 -1.05482864e+00 8.99556354e-02
3.05235565e-01 -9.29501891e-01 9.94213998e-01 -1.46046770e+00
-1.75172496e+00 6.49592280e-01 -2.83123940e-01 -4.15751606e-01
8.56110156e-01 -4.38498050e-01 4.38504666e-02 2.74143249e-01
-9.78562459e-02 7.13220417e-01 1.40911126e+00 -1.40120399e+00
-3.69671017e-01 5.86423539e-02 1.92958176e-01 1.00387624e-02
-3.51904482e-01 3.54196578e-02 -3.52687359e-01 -9.22525167e-01
-8.25772956e-02 -9.96233761e-01 -2.47026265e-01 2.00826656e-02
-3.25272009e-02 3.47599596e-01 8.65818977e-01 -7.02615917e-01
1.10960269e+00 -2.00618482e+00 8.23236048e-01 -3.53629649e-01
2.33630657e-01 4.70615834e-01 -6.43939078e-02 3.28874081e-01
-3.81407410e-01 -1.38579458e-01 -8.72101560e-02 -6.47161484e-01
-2.67117292e-01 1.00574687e-01 -4.88242507e-01 7.32709348e-01
1.81161672e-01 6.14131808e-01 -9.63773429e-01 -4.67191309e-01
5.02250552e-01 6.97805882e-01 -8.90326619e-01 3.06826949e-01
-2.74713993e-01 9.83862877e-01 -2.07117766e-01 5.21574974e-01
4.54274595e-01 -3.03653115e-03 1.45282313e-01 -3.23948473e-01
-2.41652668e-01 -1.95065305e-01 -1.04852819e+00 2.01774478e+00
-1.77798659e-01 7.16376781e-01 3.05873692e-01 -1.37019563e+00
4.75159585e-01 5.12829185e-01 8.82616758e-01 -3.46999109e-01
4.24611568e-01 1.63362458e-01 -1.64128393e-01 -5.12970686e-01
3.39511454e-01 -1.71721905e-01 3.12379211e-01 1.79818179e-02
1.35706887e-01 -1.08376637e-01 3.42228800e-01 9.33987740e-03
8.65100265e-01 6.83934689e-01 -1.19331919e-01 -2.93739796e-01
8.92678380e-01 -3.20461363e-01 4.62836206e-01 3.03399086e-01
-3.07264507e-01 9.02044117e-01 4.68291968e-01 -4.06370789e-01
-1.38093960e+00 -8.08064759e-01 1.49476022e-01 7.64333129e-01
1.29047006e-01 -4.25172895e-01 -1.10036004e+00 -1.52322993e-01
-5.32077849e-01 2.57510781e-01 -5.22489846e-01 -1.50177404e-01
-9.15478766e-01 -5.65108895e-01 3.18589151e-01 3.03754747e-01
4.08161759e-01 -9.33118939e-01 -6.05917513e-01 2.73566216e-01
-3.36398751e-01 -1.20108318e+00 -4.26570356e-01 -1.65939614e-01
-9.46987092e-01 -1.01650369e+00 -1.08056343e+00 -1.01730454e+00
6.32890165e-01 3.22334737e-01 8.98021936e-01 1.24682710e-01
8.50819126e-02 1.97084486e-01 -1.70529276e-01 3.06985915e-01
-6.46110594e-01 -1.02665454e-01 3.14949334e-01 2.33397156e-01
-5.49198031e-01 -7.44179606e-01 -8.39776218e-01 3.23936135e-01
-1.11507618e+00 3.49534482e-01 1.84160084e-01 1.03870583e+00
4.56950992e-01 -1.92131206e-01 1.97549179e-01 -3.26744616e-01
1.21622674e-01 -1.78418234e-01 -7.09815681e-01 2.91665941e-02
-4.84831423e-01 2.36007348e-01 7.09517062e-01 -4.70462233e-01
-9.05746579e-01 2.27373078e-01 -1.17686465e-01 -9.88474011e-01
1.53917745e-01 2.44111434e-01 -1.93714142e-01 -1.47809833e-01
5.80117047e-01 3.79612923e-01 2.04975039e-01 -1.62096962e-01
4.22531962e-01 -1.45078570e-01 7.92347729e-01 -5.12276590e-01
1.07736611e+00 5.27794063e-01 2.54481018e-01 -7.45983481e-01
-6.67818189e-01 -1.19224645e-01 -5.32683134e-01 -4.56779361e-01
1.06490576e+00 -1.12773085e+00 -5.94779372e-01 7.31414497e-01
-1.09908676e+00 -2.89389759e-01 -1.22781329e-01 3.85616899e-01
-1.04629612e+00 8.00154626e-01 -7.87891448e-01 -4.20225888e-01
-2.08861277e-01 -1.50200558e+00 9.69728291e-01 2.09777027e-01
6.36687502e-02 -9.78854120e-01 2.47525707e-01 3.25098693e-01
3.48914832e-01 4.68224257e-01 6.67710125e-01 1.87334225e-01
-8.19835842e-01 -3.93378623e-02 2.17087880e-01 4.87621784e-01
9.91569236e-02 2.92347789e-01 -7.88049877e-01 -6.80517018e-01
2.12843969e-01 -2.03590512e-01 7.99658239e-01 6.76874697e-01
8.61596346e-01 -3.02009702e-01 -1.26127213e-01 8.32799017e-01
1.50818825e+00 -7.55739808e-02 9.41352248e-01 4.13868934e-01
8.63991320e-01 4.20251489e-01 3.76048446e-01 2.80785650e-01
1.77936628e-02 8.54644477e-01 7.13834107e-01 -2.06341043e-01
-3.14714760e-01 -9.07621309e-02 7.45244801e-01 7.63609052e-01
-5.45083880e-01 -2.61588633e-01 -5.48361063e-01 5.64439893e-01
-2.08319664e+00 -1.29463458e+00 -1.08063119e-02 2.13234043e+00
7.75426924e-01 1.93491429e-02 2.80514117e-02 2.95990765e-01
8.18664789e-01 4.26063597e-01 -2.43161723e-01 -1.41467988e-01
-4.92632054e-02 -4.35861424e-02 1.88853309e-01 6.98590934e-01
-1.25103807e+00 9.86485183e-01 5.90820169e+00 7.33212233e-01
-1.34177530e+00 -9.05423518e-03 5.46896279e-01 -1.73391387e-01
-1.90089121e-01 4.28315401e-02 -5.85985363e-01 5.30013680e-01
8.76778126e-01 1.91946894e-01 4.48246539e-01 6.29537821e-01
6.32078886e-01 8.80974606e-02 -1.06511176e+00 1.00441134e+00
1.72297314e-01 -1.53201282e+00 4.83589508e-02 -1.95827916e-01
1.06482351e+00 -4.22207892e-01 2.16613337e-01 7.86317140e-02
-1.46257028e-01 -7.94163942e-01 1.07003033e+00 6.14362597e-01
5.78359723e-01 -6.03975415e-01 4.86723721e-01 1.61204845e-01
-1.00058353e+00 -4.79629152e-02 -2.06695005e-01 -7.50957951e-02
3.43235701e-01 2.49849379e-01 -1.78361967e-01 6.20950639e-01
6.13465786e-01 9.62605655e-01 -4.50587332e-01 6.43690884e-01
-2.26122335e-01 4.69701678e-01 7.24775344e-02 4.06512976e-01
2.39767566e-01 -4.90932137e-01 6.59788251e-01 7.83820629e-01
4.63524759e-01 -1.10040512e-02 -2.59886105e-02 6.97229803e-01
-2.15334576e-02 -9.14813578e-02 -5.44577897e-01 1.46065757e-01
-2.22062469e-01 1.03550994e+00 -5.98535299e-01 -3.26303720e-01
-4.25370723e-01 9.41979170e-01 6.19287007e-02 4.65560347e-01
-1.16318834e+00 3.22251707e-01 7.13412106e-01 4.19405639e-01
5.15519917e-01 -4.71443534e-01 2.39380710e-02 -1.58281159e+00
-5.03094867e-02 -1.00076902e+00 -1.07271530e-01 -9.65646684e-01
-7.53060997e-01 5.81493855e-01 -1.33104891e-01 -1.62343204e+00
-6.30328059e-01 -2.79835582e-01 -6.58622622e-01 6.85449064e-01
-1.48653650e+00 -1.27568603e+00 -3.95852238e-01 9.35338616e-01
8.01522732e-01 -2.22446159e-01 4.70976382e-01 4.87251133e-01
-6.88558638e-01 5.29855430e-01 4.43671644e-01 -1.04541972e-01
9.27641809e-01 -9.14518237e-01 5.26902974e-02 1.38882816e+00
-1.69448614e-01 4.13764864e-01 1.13685429e+00 -3.73633891e-01
-1.43376064e+00 -9.36437726e-01 6.59385026e-01 -3.15765381e-01
7.25544631e-01 -1.33368999e-01 -7.83650756e-01 5.70680022e-01
5.66495359e-01 5.96136041e-02 1.85563758e-01 -5.34056008e-01
7.73991644e-02 -1.45302877e-01 -7.53817737e-01 8.83349299e-01
8.57697904e-01 -3.42650712e-01 -5.02210200e-01 5.55451363e-02
6.20148718e-01 -4.60729063e-01 -5.75705290e-01 4.08523679e-01
4.10425067e-01 -1.11676395e+00 1.16827941e+00 -4.99901205e-01
1.05671227e+00 -5.96719027e-01 -4.61324528e-02 -1.06307006e+00
-2.66882658e-01 -8.81309032e-01 -2.59592921e-01 1.29116201e+00
-1.08950943e-01 1.49781127e-02 7.97788203e-01 4.12185431e-01
-3.20944846e-01 -4.48469520e-01 -8.28946412e-01 -4.43955660e-01
1.67819843e-01 -5.87383956e-02 -1.51090413e-01 8.46875429e-01
-2.37457424e-01 2.77890563e-01 -8.68195117e-01 1.27507269e-01
5.13622701e-01 1.26259983e-01 8.86856079e-01 -6.77335322e-01
-2.31170788e-01 -3.48035723e-01 -4.70745504e-01 -9.74466205e-01
3.84367287e-01 -3.94123524e-01 1.31397724e-01 -1.04070163e+00
-2.99736429e-02 1.87474728e-01 -4.69645895e-02 8.12942311e-02
-2.40560547e-01 2.97238380e-01 2.90319115e-01 3.84906918e-01
-4.11077231e-01 8.31983626e-01 1.31792438e+00 -2.39613354e-01
-1.14022642e-01 -1.28862172e-01 -2.14441895e-01 7.44046688e-01
4.54792827e-01 -1.66959912e-01 -5.45385480e-01 -4.51600164e-01
7.27875009e-02 3.21429551e-01 4.43801343e-01 -1.28789055e+00
4.47260030e-02 -8.91812816e-02 3.12646955e-01 -2.01105878e-01
2.76599497e-01 -8.04397941e-01 2.96327233e-01 7.56625772e-01
-2.54462749e-01 2.68262684e-01 1.54162332e-01 6.20332420e-01
-3.54083329e-01 1.34121068e-03 1.18246722e+00 -6.65054396e-02
-6.44091427e-01 2.58897781e-01 -6.27306521e-01 -1.52048305e-01
9.65453744e-01 -2.91357368e-01 -3.25583220e-02 -4.40507770e-01
-8.55990171e-01 -1.63970113e-01 8.39014590e-01 3.29927713e-01
3.49055499e-01 -1.19710016e+00 -5.26136100e-01 2.42205769e-01
-4.35540050e-01 -2.34456128e-03 2.35724017e-01 1.02688634e+00
-9.28640306e-01 1.68552265e-01 -6.17358506e-01 -7.16387212e-01
-1.20057952e+00 1.01046741e+00 5.29999018e-01 -1.75195321e-01
-5.50075889e-01 4.69027370e-01 4.69259232e-01 2.34234542e-01
1.66777313e-01 -3.68908316e-01 -1.52055249e-01 -2.39911508e-02
2.76075512e-01 3.37752283e-01 9.88310762e-03 -7.78149724e-01
-9.55407545e-02 7.90048540e-01 3.17896545e-01 -7.81896710e-02
1.20167482e+00 -4.30730134e-01 5.69827221e-02 -4.49736556e-03
9.74913538e-01 3.27804103e-03 -1.83351791e+00 1.14757210e-01
-5.12638450e-01 -4.33564276e-01 -1.02159001e-01 -2.78589204e-02
-1.41509140e+00 9.05785799e-01 8.58530223e-01 -1.27938062e-01
1.16384399e+00 -3.93161684e-01 7.02783167e-01 5.51890358e-02
1.75938070e-01 -7.70175874e-01 3.54610831e-01 3.14604521e-01
7.68085122e-01 -1.27675688e+00 -6.61734715e-02 -2.68659085e-01
-2.90373474e-01 1.08871531e+00 5.16698718e-01 -4.19667125e-01
4.69167173e-01 6.35865470e-03 9.02456567e-02 2.20192000e-01
-5.91322303e-01 -4.72049005e-02 1.93216905e-01 4.53595310e-01
4.49332446e-01 -5.91913283e-01 -5.08051395e-01 3.09630651e-02
1.74453139e-01 -3.14548681e-03 6.33705795e-01 7.17887819e-01
-2.39687294e-01 -1.12823749e+00 -4.10622030e-01 -4.40114051e-01
-5.25057673e-01 -2.75871418e-02 1.06080845e-01 7.90688097e-01
1.02203034e-01 8.06641281e-01 -1.87935695e-01 -1.67373881e-01
4.47104126e-02 -6.67251199e-02 7.17510939e-01 -7.30067044e-02
-4.70909238e-01 4.57108319e-01 -4.06136960e-01 -6.32374108e-01
-1.01432002e+00 -8.08986783e-01 -9.90938187e-01 -6.34853065e-01
-4.52785641e-02 -8.84015709e-02 3.07483315e-01 9.20021236e-01
1.27716720e-01 6.25208080e-01 6.61288977e-01 -1.52460158e+00
-4.14243102e-01 -7.11316526e-01 -1.77693397e-01 7.20573306e-01
6.43347144e-01 -7.68988609e-01 -3.41349691e-01 8.16801965e-01] | [10.665437698364258, -1.3893405199050903] |
a0f9d05c-6b2a-4a87-bd25-6c926f99374a | a-comparison-of-neuroelectrophysiology | 2306.15041 | null | https://arxiv.org/abs/2306.15041v1 | https://arxiv.org/pdf/2306.15041v1.pdf | A Comparison of Neuroelectrophysiology Databases | As data sharing has become more prevalent, three pillars - archives, standards, and analysis tools - have emerged as critical components in facilitating effective data sharing and collaboration. This paper compares four freely available intracranial neuroelectrophysiology data repositories: Data Archive for the BRAIN Initiative (DABI), Distributed Archives for Neurophysiology Data Integration (DANDI), OpenNeuro, and Brain-CODE. These archives provide researchers with tools to store, share, and reanalyze neurophysiology data though the means of accomplishing these objectives differ. The Brain Imaging Data Structure (BIDS) and Neurodata Without Borders (NWB) are utilized by these archives to make data more accessible to researchers by implementing a common standard. While many tools are available to reanalyze data on and off the archives' platforms, this article features Reproducible Analysis and Visualization of Intracranial EEG (RAVE) toolkit, developed specifically for the analysis of intracranial signal data and integrated with the discussed standards and archives. Neuroelectrophysiology data archives improve how researchers can aggregate, analyze, distribute, and parse these data, which can lead to more significant findings in neuroscience research. | ['Dominique Duncan', 'Arthur W. Toga', 'Nader Pouratian', 'Michael Beauchamp', 'John Magnotti', 'Zhengjia Wang', 'Stephen R Arnott', 'Alana Sparks', 'Brendan Behan', 'Scott Makeig', 'Arnaud Delorme', 'Dora Hermes', 'Chris Markiewicz', 'Russell A. Poldrack', 'Benjamin Dichter', 'Yaroslav Halchenko', 'Satrajit Ghosh', 'Samuel Hobel', 'Calvary Fisher', 'Sana Salehi', 'Rachael Garner', 'Samantha L. Cohen', 'Misque Boswell', 'Alex Gray', 'Priyanka Subash'] | 2023-06-26 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-6.84680700e-01 -4.64410365e-01 3.45698118e-01 -5.22844017e-01
-5.12988687e-01 -3.17700058e-01 1.96954176e-01 5.30558407e-01
-6.27748609e-01 8.59014213e-01 5.55839658e-01 -2.61088848e-01
-2.83236682e-01 -4.32749242e-01 -1.09770410e-01 -5.58468580e-01
-4.55776036e-01 7.90758431e-02 8.52258280e-02 -8.19156095e-02
5.70142329e-01 6.82344258e-01 -1.46974099e+00 2.13084430e-01
8.35551620e-01 9.66716111e-01 3.22813570e-01 1.07895099e-01
-7.64166489e-02 1.77300811e-01 -7.28940547e-01 1.03602141e-01
-7.33343214e-02 -1.59591228e-01 -6.61592901e-01 -7.68750310e-01
-5.01526967e-02 -3.18600893e-01 -9.09220576e-02 8.92937899e-01
8.23977828e-01 -6.62461817e-02 2.44750559e-01 -1.22959507e+00
-7.19660044e-01 2.46537909e-01 -2.18318090e-01 1.13457656e+00
3.08622777e-01 1.43201500e-01 2.61667192e-01 -7.14814723e-01
8.04315150e-01 7.12171316e-01 5.12838840e-01 2.43015379e-01
-9.20198083e-01 -1.02862763e+00 -2.42268935e-01 4.83015448e-01
-1.37117076e+00 -5.08807361e-01 2.11835846e-01 -5.75854182e-01
1.06146502e+00 4.49398160e-01 1.05165231e+00 1.25918722e+00
7.51464367e-01 -1.45492643e-01 1.19127059e+00 -2.44310617e-01
3.03768039e-01 1.37833267e-01 9.49365079e-01 -1.47472084e-01
4.91282433e-01 -2.46263534e-01 -9.40442502e-01 -4.22686309e-01
7.20559776e-01 -1.76562384e-01 -5.28173745e-01 2.69529581e-01
-1.06121230e+00 1.89158767e-01 1.29125938e-01 7.48989165e-01
-8.12857926e-01 -3.92917275e-01 8.31547976e-01 4.81306106e-01
6.87955022e-01 1.49195239e-01 -2.39793435e-01 -6.16703510e-01
-8.34375799e-01 3.16724956e-01 5.65988243e-01 7.45334208e-01
1.76183075e-01 -9.59061086e-02 1.52805582e-01 1.00187469e+00
4.36468452e-01 2.80439228e-01 1.08092761e+00 -8.78014982e-01
1.45738691e-01 3.68959397e-01 -1.48705825e-01 -8.25669885e-01
-8.53905678e-01 -1.99717358e-01 -6.44217551e-01 2.45547354e-01
1.03412047e-01 2.46575493e-02 -3.57170403e-01 1.29435229e+00
9.68787074e-02 -3.10557544e-01 -2.73759395e-01 1.02122509e+00
9.52447414e-01 8.33638683e-02 2.22448692e-01 -1.96113974e-01
1.55075109e+00 -2.31115311e-01 -1.18966675e+00 2.38207966e-01
5.50025165e-01 -3.69728476e-01 8.55663478e-01 7.00510323e-01
-1.33183265e+00 -2.58900493e-01 -9.77160454e-01 -2.18742251e-01
-8.60509038e-01 -4.52122629e-01 3.30585003e-01 5.18571734e-01
-1.59670341e+00 3.92090738e-01 -1.16300333e+00 -6.57628357e-01
6.75361037e-01 1.90527141e-01 -8.08188915e-01 4.96920288e-01
-1.01079881e+00 1.43383586e+00 1.15246482e-01 -8.82160887e-02
-5.38619757e-01 -9.13768113e-01 -2.84250319e-01 -1.77925944e-01
-4.86019492e-01 -3.70829552e-01 7.81998277e-01 -3.52428734e-01
-9.20319319e-01 9.41974282e-01 -2.93657575e-02 -4.31768268e-01
4.12436435e-03 -2.87019275e-03 -8.35116565e-01 3.09243828e-01
3.30213547e-01 4.09246981e-01 1.46884769e-01 -5.09815872e-01
-3.24504197e-01 -1.00369895e+00 -7.20723510e-01 -1.15525045e-01
-5.30877650e-01 8.06246638e-01 -2.03062654e-01 -5.55212080e-01
1.36534125e-01 -2.94898421e-01 3.21542412e-01 1.07144937e-01
-8.12756922e-03 -1.14636526e-01 6.39607131e-01 -1.17869782e+00
1.04966271e+00 -2.44652963e+00 -1.13520332e-01 2.96664864e-01
5.52578032e-01 3.17232385e-02 2.28645608e-01 3.02051574e-01
-4.45398659e-01 2.72418439e-01 1.22106276e-01 -3.02074373e-01
-1.36010602e-01 6.89179823e-03 -5.08161373e-02 8.35506439e-01
-3.57182652e-01 3.69331241e-01 -4.62111771e-01 -1.83305442e-01
5.23457341e-02 5.37232161e-01 -2.56200612e-01 -9.89588201e-02
6.72266006e-01 7.93567955e-01 -9.67513397e-02 3.85513991e-01
8.03220928e-01 -1.16188101e-01 3.84494178e-02 -1.13762558e-01
-9.00731504e-01 4.96929795e-01 -9.21087563e-01 1.74523747e+00
-1.70969605e-01 9.09325063e-01 4.54725683e-01 -6.00269556e-01
9.63317454e-01 6.33802235e-01 5.20894885e-01 -9.84699726e-01
3.19812238e-01 3.03133935e-01 2.24213019e-01 -6.77811146e-01
2.61973888e-01 3.17919046e-01 4.90970671e-01 8.11879933e-01
3.52031559e-01 3.18142444e-01 2.04393283e-01 3.17686498e-01
1.07287836e+00 -9.02223140e-02 2.21722946e-01 -7.30048835e-01
1.56475857e-01 -3.19492310e-01 3.97230268e-01 2.36201942e-01
-4.42214221e-01 3.76768500e-01 3.34898233e-01 -2.74875969e-01
-1.01335418e+00 -1.11292434e+00 -9.92263079e-01 8.04466367e-01
-3.05716097e-01 -6.72260523e-01 -7.97484756e-01 4.81662989e-01
-1.28266558e-01 5.49872756e-01 -4.53315496e-01 1.31296426e-01
1.60561070e-01 -7.15879798e-01 5.39510071e-01 4.20968682e-01
4.60123628e-01 -1.18601167e+00 -1.07145226e+00 2.23955452e-01
-2.62478411e-01 -6.50169432e-01 -3.88459116e-02 -5.60847968e-02
-9.16861713e-01 -9.65346336e-01 -6.17487550e-01 -2.48104483e-01
4.72839057e-01 1.10449806e-01 5.41790068e-01 1.76895902e-01
-9.42126215e-01 6.40035391e-01 -2.01241478e-01 -7.55076766e-01
-4.40375358e-02 -2.47674733e-01 3.56256455e-01 -3.08714360e-01
4.04353857e-01 -1.07932973e+00 -6.30058587e-01 -6.91496581e-02
-9.34906363e-01 -7.48784691e-02 2.54234709e-02 5.31130284e-02
5.05881250e-01 -5.76979816e-01 1.12797403e+00 -1.85897261e-01
1.16994059e+00 -9.96020079e-01 -5.08267641e-01 3.89504805e-02
-8.70120645e-01 -5.96747339e-01 1.74557567e-01 3.08650643e-01
-7.52807856e-01 -9.84654129e-01 -1.03262894e-01 -2.41122440e-01
-3.71099114e-01 6.28282189e-01 9.83084738e-02 -9.56541151e-02
5.68701386e-01 5.65407276e-02 3.62386048e-01 -8.31528723e-01
-8.39107931e-02 1.11182272e+00 7.97103345e-01 -1.81763291e-01
-4.29141909e-01 4.30790931e-01 -5.10567009e-01 -7.26688921e-01
1.11723714e-01 -5.76325059e-01 -6.79666758e-01 -3.61134231e-01
1.05919051e+00 -9.63651836e-01 -5.85859835e-01 5.53678334e-01
-1.07581902e+00 3.33065353e-02 2.98533272e-02 8.78606856e-01
-1.04133308e-01 1.50825277e-01 -4.01529700e-01 -5.24046481e-01
-7.89115012e-01 -1.03311622e+00 3.99611682e-01 2.57920563e-01
-6.82738245e-01 -9.45541203e-01 3.10156882e-01 3.36947143e-02
7.89969742e-01 1.92016408e-01 7.63378382e-01 -8.35318685e-01
-9.80062187e-02 -7.31256455e-02 -1.66550130e-01 2.49072462e-01
3.15020569e-02 1.06097795e-01 -9.56868410e-01 -1.41944261e-02
3.27994078e-01 2.12109070e-02 -5.72630204e-02 3.51934344e-01
1.46723390e+00 -3.30246128e-02 -3.02603424e-01 4.93110627e-01
1.00464082e+00 2.88703650e-01 8.40179861e-01 6.00013256e-01
-4.90834750e-02 6.51560843e-01 -1.43135220e-01 7.01552331e-01
6.16721034e-01 4.03272927e-01 2.25790709e-01 3.49717259e-01
-4.19781432e-02 6.86478913e-01 -1.94010027e-02 1.21067739e+00
-4.55133975e-01 4.86729920e-01 -1.20445919e+00 5.68512499e-01
-1.60540628e+00 -7.77880251e-01 -6.73992574e-01 2.29924703e+00
6.97331429e-01 -2.40867481e-01 1.75621018e-01 -1.35575026e-01
5.05351484e-01 -3.43167841e-01 -3.44578832e-01 -4.44903523e-01
-8.86632130e-02 4.63558823e-01 1.83254719e-01 3.72996032e-02
-2.95842677e-01 1.78040087e-01 7.15647602e+00 1.96676031e-01
-1.32165349e+00 8.44764113e-01 1.28626928e-01 -6.47311687e-01
-1.08801104e-01 -2.79339135e-01 -4.63348448e-01 8.61976802e-01
1.74586022e+00 -9.57352579e-01 6.85161531e-01 5.86751997e-01
7.58572280e-01 -4.52170759e-01 -7.11280406e-01 1.26682997e+00
-5.45887351e-02 -1.67497289e+00 -4.10748065e-01 2.96703041e-01
-1.07057355e-01 8.10363233e-01 -2.64054984e-01 -2.09614500e-01
-3.58160704e-01 -7.18848705e-01 7.76850283e-01 1.11668670e+00
7.32220829e-01 -4.17258978e-01 5.48598707e-01 -4.41259108e-02
-5.61912477e-01 1.60807326e-01 -2.91020691e-01 -1.09797463e-01
1.67533174e-01 6.43193901e-01 6.60215914e-02 6.85948372e-01
1.42825282e+00 7.08570063e-01 -6.37121499e-01 1.53598332e+00
3.33683312e-01 3.65221977e-01 -7.42096230e-02 3.10014993e-01
-4.11005080e-01 -3.28888416e-01 5.51621854e-01 8.54289174e-01
3.70237380e-01 2.58109182e-01 -4.71398503e-01 1.16289985e+00
1.29799038e-01 2.30900466e-01 -4.18751627e-01 -8.43415558e-02
6.56726956e-01 1.54849339e+00 -7.53088832e-01 -1.13612592e-01
-5.70235491e-01 4.69774276e-01 2.35369414e-01 7.85329491e-02
-4.15122181e-01 -6.08644843e-01 7.20369816e-01 1.55930802e-01
-5.72359562e-01 -5.38444936e-01 -6.45696044e-01 -8.46254706e-01
1.38303354e-01 -7.91643322e-01 3.81823003e-01 -1.36621511e+00
-1.26305521e+00 7.63616741e-01 3.95773977e-01 -7.85928011e-01
2.03817654e-02 -6.58639908e-01 -7.21183836e-01 1.49038231e+00
-9.38057065e-01 -5.50437689e-01 -5.83319545e-01 8.54420781e-01
-2.83139404e-02 -1.28151104e-01 1.27356207e+00 6.53151929e-01
-7.75261521e-01 -1.30496696e-02 3.03777248e-01 -1.44630805e-01
8.44014525e-01 -7.84145355e-01 4.31200974e-02 2.91687220e-01
-3.44765753e-01 1.06210244e+00 2.90709555e-01 -7.73060918e-01
-1.23540592e+00 -4.67172474e-01 6.93116963e-01 -3.02999169e-01
1.02867389e+00 -4.21651989e-01 -1.29846835e+00 7.37875581e-01
6.30318522e-01 -2.02492654e-01 1.31773663e+00 1.99772403e-01
6.90540075e-02 -2.58618712e-01 -1.12904084e+00 4.59475398e-01
8.41697037e-01 -5.09539783e-01 -1.00643265e+00 5.05753160e-01
1.76333755e-01 -9.46051106e-02 -1.38006592e+00 -3.20095539e-01
6.01467431e-01 -9.67427552e-01 4.44648802e-01 -4.41468239e-01
-7.80657008e-02 9.63590890e-02 2.44244352e-01 -1.47921824e+00
-1.05687886e-01 -6.49208009e-01 4.21384186e-01 1.21432126e+00
1.27692387e-01 -1.36338723e+00 -2.47505188e-01 1.26978135e+00
-6.00337684e-01 -3.92750651e-01 -1.26724970e+00 -6.09198213e-01
9.36272293e-02 -8.36919844e-01 9.07529891e-01 8.08058560e-01
8.45648468e-01 -4.21918631e-01 3.85690153e-01 5.50984079e-03
3.83310109e-01 -5.24915934e-01 3.64692658e-01 -1.34383881e+00
3.91844898e-01 -4.99351680e-01 -6.32909238e-01 2.12918341e-01
-1.74032554e-01 -1.26974571e+00 -5.89695930e-01 -1.74004996e+00
3.84453759e-02 -3.62183005e-01 -4.84643608e-01 7.28334367e-01
4.36398447e-01 1.35079488e-01 4.36284691e-02 7.90359020e-01
-1.97919756e-01 2.35377222e-01 8.50987732e-01 6.54064178e-01
-1.30594447e-01 -7.22289979e-01 -9.24274623e-01 5.81933796e-01
8.41866374e-01 -4.36041713e-01 -1.69322714e-01 -5.23104668e-01
-5.05266301e-02 -2.22534552e-01 7.07097232e-01 -1.33365703e+00
3.79257947e-01 4.10265326e-01 4.45544422e-01 -6.74584925e-01
1.24488614e-01 -5.73125005e-01 5.42251229e-01 9.47941914e-02
-2.57229805e-02 4.82719392e-01 6.97240353e-01 -4.33213562e-02
-1.05420277e-01 -9.21404138e-02 6.19733512e-01 5.78513220e-02
-3.24508876e-01 2.15563446e-01 -7.60786831e-01 -4.05365676e-02
1.02968538e+00 -3.36765617e-01 -9.31403637e-01 -1.41981840e-01
-1.20853484e+00 2.40631223e-01 2.82461911e-01 4.80350077e-01
3.90942484e-01 -1.17754233e+00 -5.08226216e-01 6.19299948e-01
-7.29584470e-02 -5.34910738e-01 5.90398610e-01 1.60739446e+00
-6.36782944e-01 6.44051433e-01 -1.12485266e+00 -2.61653990e-01
-9.96490896e-01 1.72039848e-02 2.04351097e-01 3.90496790e-01
-1.21415544e+00 4.51247633e-01 -3.64338845e-01 -6.01584837e-03
3.42312485e-01 -3.36018473e-01 -2.84440547e-01 3.00686240e-01
1.18039274e+00 6.17623389e-01 6.42561734e-01 -4.27899629e-01
-5.97406507e-01 -4.30953383e-01 -3.35992277e-01 -3.88700008e-01
1.71123040e+00 -3.67208481e-01 -8.28921258e-01 9.02983844e-01
9.36226785e-01 -1.88934788e-01 -5.52036226e-01 5.24146914e-01
-7.05873892e-02 -3.82496655e-01 3.27071130e-01 -9.50783491e-01
-8.78969371e-01 8.25876534e-01 9.62271750e-01 5.01991749e-01
1.06986105e+00 -2.21464500e-01 3.81390780e-01 -2.15779826e-01
6.13168597e-01 -1.25592649e+00 -5.95147073e-01 2.20758334e-01
1.42344809e+00 -3.59000921e-01 1.96216702e-02 1.22525945e-01
-3.95503402e-01 1.32319951e+00 3.69921148e-01 2.10639760e-02
1.13790345e+00 4.91144270e-01 1.17798254e-01 -6.29487276e-01
-6.75635934e-01 4.28854704e-01 -1.32498927e-02 6.84059381e-01
6.51491523e-01 -1.69456005e-01 -1.00765002e+00 1.29225004e+00
-2.90271550e-01 5.32008886e-01 4.56722707e-01 1.00699627e+00
-1.27488226e-01 -9.80733216e-01 -7.10023582e-01 8.58645320e-01
-6.77530885e-01 -2.71873534e-01 -8.20491984e-02 7.61354387e-01
6.42334893e-02 7.23460257e-01 4.92976427e-01 3.66817683e-01
1.11108907e-01 4.38652694e-01 3.05139720e-01 -4.89032656e-01
-7.87481189e-01 -6.52978122e-02 -9.70752351e-03 -7.37614334e-01
-2.06824064e-01 -8.02276075e-01 -1.49918890e+00 -2.83296645e-01
-2.02749316e-02 2.39532202e-01 1.37147081e+00 7.08438277e-01
1.20844662e+00 7.95192242e-01 -8.83629471e-02 -8.02110314e-01
5.62735051e-02 -1.06974101e+00 -1.14889991e+00 -5.08039109e-02
-1.91497609e-01 -7.63046920e-01 -2.78673768e-01 6.75319880e-02] | [13.210068702697754, 3.4437038898468018] |
86758bb9-b8c6-40c9-9222-584a6a05c451 | the-effects-of-input-type-and-pronunciation | 2306.00535 | null | https://arxiv.org/abs/2306.00535v1 | https://arxiv.org/pdf/2306.00535v1.pdf | The Effects of Input Type and Pronunciation Dictionary Usage in Transfer Learning for Low-Resource Text-to-Speech | We compare phone labels and articulatory features as input for cross-lingual transfer learning in text-to-speech (TTS) for low-resource languages (LRLs). Experiments with FastSpeech 2 and the LRL West Frisian show that using articulatory features outperformed using phone labels in both intelligibility and naturalness. For LRLs without pronunciation dictionaries, we propose two novel approaches: a) using a massively multilingual model to convert grapheme-to-phone (G2P) in both training and synthesizing, and b) using a universal phone recognizer to create a makeshift dictionary. Results show that the G2P approach performs largely on par with using a ground-truth dictionary and the phone recognition approach, while performing generally worse, remains a viable option for LRLs less suitable for the G2P approach. Within each approach, using articulatory features as input outperforms using phone labels. | ['Esther Klabbers', 'Jelske Dijkstra', 'Matt Coler', 'Phat Do'] | 2023-06-01 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [ 9.77691934e-02 4.60297354e-02 -9.48462486e-02 -2.88959980e-01
-1.43604100e+00 -9.66592014e-01 8.24617863e-01 -1.90639302e-01
-4.84127611e-01 7.86081970e-01 4.76963282e-01 -9.05229867e-01
1.87578484e-01 -2.52172142e-01 -6.40454650e-01 -5.11512578e-01
4.70601410e-01 7.73685157e-01 -2.69255996e-01 -3.09455752e-01
-1.41195402e-01 4.85736996e-01 -1.35074615e+00 1.45716920e-01
8.25225651e-01 6.62678540e-01 5.51851690e-01 6.93263769e-01
-4.18844700e-01 7.03241110e-01 -4.71110940e-01 -5.02342463e-01
3.56453508e-01 -6.67935550e-01 -7.21316516e-01 -1.20718300e-01
5.71574509e-01 1.25132455e-02 -4.89315689e-02 7.96996176e-01
6.14021122e-01 2.56128907e-01 9.03181374e-01 -5.19279718e-01
-6.77639782e-01 8.23432744e-01 1.67570636e-01 7.21549615e-03
6.39584899e-01 3.02374810e-02 1.11605906e+00 -1.28692544e+00
5.50135851e-01 1.61651421e+00 7.19604015e-01 5.65049827e-01
-1.36792123e+00 -5.69226503e-01 -2.33700171e-01 4.66815978e-02
-1.49738407e+00 -1.23670709e+00 4.13206428e-01 -4.63543326e-01
1.42066801e+00 2.94321060e-01 2.99787194e-01 9.78119671e-01
-2.59209782e-01 5.52395999e-01 1.34957123e+00 -1.17977798e+00
-1.29334154e-02 4.53200489e-01 -4.69618887e-01 6.10984862e-01
-2.54042059e-01 5.18935025e-01 -6.05213344e-01 2.45702550e-01
5.72349608e-01 -8.13396990e-01 -1.91782117e-01 1.73562244e-01
-1.41245711e+00 8.13512623e-01 -2.69064993e-01 6.59933865e-01
-2.37716153e-01 -4.02327590e-02 5.16606092e-01 7.95052111e-01
3.59342009e-01 4.53887790e-01 -7.49532044e-01 -3.62940162e-01
-1.10298169e+00 -2.56376356e-01 8.61912191e-01 9.39106643e-01
6.60830259e-01 7.09357083e-01 9.95465219e-02 1.54225314e+00
2.24414289e-01 1.03134203e+00 8.96232069e-01 -8.43183815e-01
5.15706062e-01 -2.44276688e-01 -9.30359960e-02 -3.75082642e-01
3.59885693e-02 -7.73509070e-02 -1.71900153e-01 6.53149700e-03
6.25251055e-01 -1.44115523e-01 -9.67422307e-01 1.77543437e+00
1.23467065e-01 -1.43013194e-01 5.49911976e-01 3.84612709e-01
6.83868647e-01 1.05370951e+00 6.89985454e-02 -4.92728710e-01
1.13508451e+00 -1.09641850e+00 -8.68579924e-01 -3.24181676e-01
9.93739128e-01 -1.25759327e+00 1.37519753e+00 5.26378095e-01
-1.26834810e+00 -1.03627098e+00 -9.35308218e-01 -2.86451697e-01
-7.01701641e-01 5.31088829e-01 1.61035985e-01 1.15924203e+00
-1.29285228e+00 4.22292233e-01 -5.35541177e-01 -3.87159914e-01
-5.44536293e-01 3.30348760e-01 -5.18345594e-01 -2.19767526e-01
-1.31460857e+00 1.29562986e+00 5.95046520e-01 -3.53757709e-01
-8.78887117e-01 -4.23108339e-01 -1.29967701e+00 -2.01444939e-01
2.33920664e-02 -3.86239365e-02 1.25213265e+00 -9.51340854e-01
-2.12476635e+00 1.04063404e+00 -1.79981351e-01 -3.44392926e-01
2.34708145e-01 -1.64634407e-01 -8.67124379e-01 -1.77222893e-01
-3.13666761e-02 7.21966207e-01 9.44493592e-01 -1.27273703e+00
-7.53711879e-01 2.93233544e-02 -4.99076694e-01 5.72254658e-01
-9.82948765e-02 2.76820660e-01 -2.97698528e-01 -7.23738611e-01
-1.29269525e-01 -1.05904901e+00 4.24260497e-01 -7.80799925e-01
-1.26571432e-01 -5.47378004e-01 5.08603036e-01 -1.38507688e+00
1.17984474e+00 -1.88492644e+00 1.63066044e-01 2.17898488e-01
-6.57163680e-01 7.19811201e-01 -2.92256773e-01 6.35415494e-01
-1.46268710e-01 8.36761966e-02 -1.11954994e-01 -6.28310978e-01
2.39329413e-01 5.95332742e-01 -2.88745135e-01 1.85824707e-01
1.81188032e-01 9.67547238e-01 -8.44647706e-01 -3.38899136e-01
6.84914291e-01 4.72347885e-01 -3.08607727e-01 9.08871219e-02
1.75910011e-01 6.49838090e-01 3.50631863e-01 5.22105515e-01
1.30191490e-01 7.64373779e-01 4.28616852e-01 5.60060889e-02
-4.40845817e-01 1.05066764e+00 -9.56929982e-01 1.69698727e+00
-1.37699568e+00 6.12045825e-01 2.83877105e-01 -9.94812310e-01
1.06448257e+00 6.74264610e-01 -1.04347222e-01 -6.64723575e-01
4.43121418e-03 9.82772410e-01 3.89560610e-01 -2.09863782e-01
3.75983715e-01 -5.52170277e-01 -2.10464969e-01 1.60013333e-01
5.85361421e-01 -7.52135813e-01 -4.23381031e-02 -2.75717407e-01
7.45126009e-01 3.37257475e-01 4.25827742e-01 -5.43503463e-01
7.93188751e-01 -7.56940916e-02 3.38529319e-01 5.20189285e-01
1.60298124e-01 4.56024498e-01 -2.62491435e-01 1.33071885e-01
-9.41575110e-01 -1.32597303e+00 -1.59471631e-01 1.26898968e+00
-5.99152982e-01 -2.87144274e-01 -9.53651011e-01 -4.22365785e-01
-6.35821074e-02 1.32045305e+00 -2.69865114e-02 6.76331818e-02
-8.83380592e-01 1.45811006e-01 8.58360827e-01 2.66815901e-01
-2.06377804e-02 -1.38922584e+00 3.02700311e-01 5.24734080e-01
-3.23918164e-01 -1.32789576e+00 -8.78495038e-01 3.86008143e-01
-4.86480355e-01 -3.75282556e-01 -8.80913198e-01 -1.28257465e+00
1.29502028e-01 -1.38213918e-01 9.86767948e-01 -3.36267680e-01
1.31407648e-01 4.95025307e-01 -4.83593613e-01 -4.10096526e-01
-1.26414776e+00 9.58454609e-02 6.08902216e-01 -1.39633626e-01
3.88395309e-01 -3.42749029e-01 2.47099534e-01 1.49620652e-01
-4.59357262e-01 -3.66882354e-01 5.21275043e-01 7.96814859e-01
6.41193926e-01 -2.44113564e-01 9.75995004e-01 -9.20780003e-01
4.32283282e-01 -1.35857537e-01 -3.93137962e-01 1.66811749e-01
-5.82648396e-01 -1.28409058e-01 1.04555190e+00 -2.83181190e-01
-1.15746427e+00 1.64190575e-01 -8.16561878e-01 -4.01262015e-01
-3.31362695e-01 4.17100102e-01 -6.08416855e-01 2.75080074e-02
5.98005593e-01 2.14352772e-01 -1.61845144e-02 -7.80316532e-01
6.84017003e-01 1.08433580e+00 9.24255550e-01 -7.55892456e-01
9.03979838e-01 -8.39660168e-02 -4.99563307e-01 -1.34957802e+00
-4.51324344e-01 -5.59274435e-01 -7.64439106e-01 -9.89292562e-02
8.25194061e-01 -1.02089131e+00 -1.05553992e-01 4.24800128e-01
-1.12830222e+00 -5.50576270e-01 -6.71126425e-01 9.37187672e-01
-8.87555420e-01 3.20660025e-01 -5.41584671e-01 -7.44399071e-01
-1.51601061e-01 -1.16721094e+00 1.06164658e+00 -3.09042782e-01
-3.46798539e-01 -1.45920277e+00 1.71092525e-01 5.08138478e-01
4.09006685e-01 -2.16609433e-01 1.08113253e+00 -8.74485254e-01
1.13676287e-01 -1.07553571e-01 2.22777829e-01 1.03387356e+00
6.81383312e-01 -3.13492626e-01 -1.15419769e+00 -4.16987062e-01
4.67994511e-02 -4.90115285e-01 3.20707232e-01 2.72238612e-01
3.68038177e-01 -2.82627374e-01 1.43055886e-01 5.70128500e-01
1.14177215e+00 5.23958981e-01 6.17052555e-01 -1.38861567e-01
8.08740318e-01 8.39430153e-01 5.87612092e-01 -2.01932356e-01
4.09238100e-01 7.86212623e-01 -4.32914406e-01 -1.76233351e-01
-9.24775839e-01 -5.00813305e-01 1.07806230e+00 2.10826468e+00
9.17165205e-02 -3.10798675e-01 -1.09059465e+00 8.70638072e-01
-1.18748367e+00 -4.41254765e-01 9.66688320e-02 2.41990662e+00
1.14306784e+00 2.93993205e-03 -6.68622181e-02 1.66856587e-01
7.67382383e-01 -8.06415081e-02 -5.52700050e-02 -1.03741348e+00
-3.09018612e-01 7.86294401e-01 4.68966246e-01 1.06148589e+00
-6.43513381e-01 1.66132748e+00 6.18638659e+00 1.46440649e+00
-1.01150429e+00 4.20695066e-01 1.61587238e-01 3.96085471e-01
-3.92791420e-01 1.59068387e-02 -9.71371651e-01 2.92269051e-01
1.65604889e+00 1.05137944e-01 1.03048635e+00 4.68510032e-01
2.15553880e-01 2.02040672e-01 -1.20855033e+00 1.17371237e+00
2.57793933e-01 -1.00808537e+00 1.80085644e-01 -1.13882065e-01
7.13106155e-01 2.92178571e-01 -3.75477076e-02 6.40961111e-01
4.72250581e-01 -1.05410564e+00 9.64764416e-01 1.84275046e-01
1.50111830e+00 -6.92553341e-01 4.34996873e-01 3.35195214e-01
-1.36048317e+00 4.52965885e-01 -3.01849991e-01 1.88161150e-01
2.52165407e-01 6.77452981e-03 -1.01024044e+00 6.51359916e-01
1.59064069e-01 3.67532581e-01 -1.62187770e-01 4.22253698e-01
-3.44860166e-01 1.12372434e+00 -4.51149195e-01 2.68328875e-01
3.56822401e-01 -4.09270674e-01 6.29150987e-01 1.76728344e+00
6.03100479e-01 -3.68368477e-01 1.55563176e-01 1.13351859e-01
7.73301274e-02 8.16733062e-01 -8.20794821e-01 -3.76813412e-01
6.09816074e-01 8.52120459e-01 -4.98341471e-01 -3.86064738e-01
-4.00652319e-01 1.02819109e+00 2.16365382e-01 1.57463238e-01
-2.60971218e-01 -4.44774747e-01 4.97297764e-01 -9.33775008e-02
3.02767843e-01 -5.01211107e-01 3.41745317e-02 -8.91637325e-01
-2.23161653e-01 -1.25579917e+00 -4.11205227e-03 -7.02267706e-01
-1.20124376e+00 8.33108485e-01 -1.59194037e-01 -1.07727194e+00
-8.67977560e-01 -9.02739286e-01 -2.89776057e-01 1.40486908e+00
-1.72059917e+00 -1.32094550e+00 4.96368885e-01 5.67496359e-01
1.03766596e+00 -5.55263937e-01 1.18112636e+00 3.63366634e-01
-9.50758636e-04 7.98785508e-01 4.07332569e-01 -2.99826767e-02
8.35317552e-01 -1.35130250e+00 6.26423061e-01 7.59382010e-01
7.76309669e-01 5.07091761e-01 4.80368614e-01 -4.89355236e-01
-1.14846754e+00 -1.02861226e+00 1.28825808e+00 -4.42141861e-01
6.77703202e-01 -4.92211014e-01 -8.28915536e-01 6.94962800e-01
5.23946226e-01 -3.87275219e-01 7.99183547e-01 6.71816245e-02
-2.16892675e-01 2.13136394e-02 -9.42753613e-01 5.52062750e-01
1.06778109e+00 -1.18168473e+00 -9.21692550e-01 4.78230625e-01
7.68066168e-01 -1.92790553e-01 -9.71010268e-01 1.18049644e-01
4.70610619e-01 -3.77107114e-01 7.82249749e-01 -3.65906000e-01
-1.92800522e-01 -2.69617200e-01 -6.98074758e-01 -1.94580173e+00
-7.66239241e-02 -1.03682101e+00 4.68019158e-01 1.70478773e+00
6.15912855e-01 -7.68573999e-01 2.70744413e-02 -2.16730237e-01
-6.81157410e-01 5.98126948e-02 -1.18717778e+00 -1.29273486e+00
4.98857170e-01 -6.90626621e-01 3.02639335e-01 1.07822835e+00
-1.61011353e-01 6.84804618e-01 -3.67926598e-01 -3.56439501e-02
1.65628955e-01 -4.14361447e-01 6.02409899e-01 -9.54481781e-01
-5.88831723e-01 -2.85905570e-01 -1.71066001e-01 -9.81795788e-01
6.49393260e-01 -1.41284978e+00 4.02241796e-01 -1.37877333e+00
-8.43549728e-01 -8.39484632e-01 -1.69377491e-01 5.59829056e-01
1.68435648e-01 3.68448585e-01 2.94629246e-01 1.05380952e-01
1.19121708e-01 3.75373244e-01 8.90816391e-01 1.39622733e-01
-3.42231423e-01 -8.74667689e-02 -2.88609147e-01 7.36949444e-01
7.38632023e-01 -4.31461245e-01 -2.96268076e-01 -4.57622379e-01
-2.86051959e-01 2.31239557e-01 -4.12672669e-01 -9.69771445e-01
-2.89096143e-02 1.21568136e-01 -2.42411774e-02 -2.81703025e-01
6.30170882e-01 -5.76520860e-01 1.18818969e-01 2.51554012e-01
-4.62395288e-02 9.67118517e-02 4.95435089e-01 2.95198355e-02
-3.44760567e-01 -5.27878761e-01 8.62941802e-01 -2.20509376e-02
-6.78139210e-01 -2.03889236e-01 -7.97003686e-01 3.28580499e-01
4.78047580e-01 -3.01377714e-01 2.97189727e-02 -5.17382801e-01
-8.45819950e-01 -2.43041560e-01 3.86458308e-01 6.11431062e-01
2.09105924e-01 -1.41452694e+00 -9.48748052e-01 5.69202662e-01
-3.35235242e-03 -7.30654299e-01 -2.51475513e-01 5.21684229e-01
-5.19539177e-01 8.56836021e-01 -4.91554625e-02 -3.25775415e-01
-1.05526435e+00 3.12362194e-01 3.02464694e-01 -1.23856999e-01
-1.73544213e-01 8.88280809e-01 1.39234349e-01 -1.06357360e+00
6.78778663e-02 -4.18694727e-02 -8.01214725e-02 1.10503532e-01
6.17626421e-02 1.83068231e-01 5.09815395e-01 -1.39200401e+00
-3.47079515e-01 5.59110105e-01 7.34552741e-02 -7.21965313e-01
9.17895138e-01 -4.09788132e-01 2.96518117e-01 9.40991223e-01
1.24482405e+00 8.51022422e-01 -7.77543783e-01 -3.23688298e-01
1.31698489e-01 -9.27012563e-02 2.72530347e-01 -9.04833317e-01
-5.49052536e-01 1.08268988e+00 3.89342815e-01 1.86746180e-01
7.91138172e-01 -9.97766629e-02 9.34758961e-01 3.32819462e-01
3.59718949e-01 -1.47598076e+00 -4.79202896e-01 9.07354772e-01
9.57913578e-01 -9.49528039e-01 -5.13227284e-01 -4.26535696e-01
-8.53561461e-01 1.08279514e+00 -1.20404810e-02 1.61350016e-02
7.06743240e-01 2.18230054e-01 5.95269978e-01 3.30038071e-01
-3.89736772e-01 -6.47718608e-01 4.18491930e-01 8.60158026e-01
5.44195175e-01 3.30268741e-01 -2.77564555e-01 1.37886897e-01
-8.71815860e-01 -5.52261412e-01 1.79616287e-01 4.56273466e-01
-2.78199226e-01 -1.62725770e+00 -4.60991859e-01 2.73154110e-01
-4.96789604e-01 -7.51985013e-01 -2.50802249e-01 9.07704175e-01
4.05622631e-01 1.25363743e+00 7.74956793e-02 -3.40277284e-01
2.89869875e-01 6.47117376e-01 5.86806059e-01 -1.14638388e+00
-7.15495884e-01 5.01143038e-01 6.60628021e-01 -1.19959779e-01
-3.23561549e-01 -1.00090706e+00 -1.05483806e+00 9.72532928e-02
-5.30610919e-01 3.59357357e-01 1.02362311e+00 1.14144278e+00
-2.06224680e-01 3.28678936e-01 6.89576209e-01 -8.43292475e-01
-4.98367041e-01 -1.10962725e+00 -7.89673984e-01 8.74143243e-02
2.12295339e-01 -5.18850982e-01 -5.66136539e-01 4.09681797e-01] | [14.318657875061035, 6.972248554229736] |
e50618c7-78ed-40ff-8279-9c4017026825 | instance-search-via-instance-level | 1806.03576 | null | https://arxiv.org/abs/1806.03576v2 | https://arxiv.org/pdf/1806.03576v2.pdf | Instance Search via Instance Level Segmentation and Feature Representation | Instance search is an interesting task as well as a challenging issue due to the lack of effective feature representation. In this paper, an instance level feature representation built upon fully convolutional instance-aware segmentation is proposed. The feature is ROI-pooled from the segmented instance region. So that instances in various sizes and layouts are represented by deep features in uniform length. This representation is further enhanced by the use of deformable ResNeXt blocks. Superior performance is observed in terms of its distinctiveness and scalability on a challenging evaluation dataset built by ourselves. In addition, the proposed enhancement on the network structure also shows superior performance on the instance segmentation task. | ['Wan-Lei Zhao', 'Yu Zhan'] | 2018-06-10 | null | null | null | null | ['instance-search'] | ['computer-vision'] | [ 3.31251770e-01 2.92367786e-01 -2.86967605e-01 -5.97491086e-01
-6.63540721e-01 -5.33999443e-01 6.09657824e-01 3.08427334e-01
-4.76152331e-01 7.60338187e-01 -3.83968494e-04 3.69691253e-01
-5.39919257e-01 -7.93265641e-01 -6.92537248e-01 -4.69921619e-01
-1.94208827e-02 2.50593513e-01 4.87061590e-01 -8.51382315e-02
4.43576306e-01 7.50737667e-01 -1.72520804e+00 2.06140891e-01
9.61759567e-01 1.34746969e+00 3.04005891e-01 2.26669759e-01
-3.85383368e-01 3.03714842e-01 -7.42604136e-01 -1.26786396e-01
5.21609545e-01 -4.88853082e-02 -1.16763532e+00 4.42009628e-01
6.98467016e-01 -1.90409496e-01 -2.22462013e-01 7.59941161e-01
2.52895325e-01 5.00457585e-01 4.13319260e-01 -9.60228741e-01
-1.96533427e-01 5.11207044e-01 -5.49595535e-01 3.25546205e-01
4.97237481e-02 -1.80993564e-02 1.09659266e+00 -6.70782924e-01
9.47469294e-01 9.69502270e-01 3.34503770e-01 3.34340960e-01
-1.06781268e+00 -4.37796950e-01 4.86859053e-01 1.68743491e-01
-1.36038375e+00 -2.47466713e-01 7.87286341e-01 -1.04360975e-01
9.31216657e-01 4.13832247e-01 6.01821303e-01 6.68060124e-01
-2.98854560e-01 1.07308745e+00 9.99237895e-01 -2.59901762e-01
2.39373580e-01 1.67145044e-01 3.90666872e-01 5.72346389e-01
9.09737721e-02 -2.75714725e-01 -2.07292721e-01 1.67731851e-01
8.23535323e-01 2.83702761e-01 -1.36763319e-01 -4.18512613e-01
-1.02732778e+00 6.32527471e-01 1.05395675e+00 6.95114076e-01
-4.36221927e-01 1.19945355e-01 5.33309579e-01 -1.27953529e-01
4.85739172e-01 7.29907513e-01 -6.18688822e-01 -6.47779778e-02
-1.52735388e+00 1.52107507e-01 6.26102149e-01 1.12041569e+00
7.68000841e-01 -6.95817024e-02 -6.03105426e-01 8.30525815e-01
-1.82527706e-01 -1.24368742e-01 4.32434201e-01 -8.48860979e-01
4.05851603e-01 1.26150405e+00 -1.23679571e-01 -8.91433835e-01
-5.53593576e-01 -9.78643715e-01 -6.79360569e-01 -3.36657427e-02
2.67329305e-01 1.86052814e-01 -1.34259033e+00 1.35451972e+00
5.63313961e-01 3.32090348e-01 -2.61245489e-01 7.80708492e-01
1.10026586e+00 3.71244222e-01 2.41075456e-02 1.51797652e-01
1.41507220e+00 -1.31528056e+00 -5.03449142e-01 -1.69350207e-01
3.09810132e-01 -3.97736281e-01 7.53815055e-01 1.17649451e-01
-1.04338992e+00 -6.83628619e-01 -9.09841239e-01 -1.57061502e-01
-7.78506935e-01 2.39337593e-01 7.27544725e-01 5.27212560e-01
-9.59000528e-01 6.52532995e-01 -7.19221950e-01 -3.92408431e-01
1.02576530e+00 6.18441880e-01 -4.33877975e-01 -5.79220913e-02
-8.51535738e-01 4.56763238e-01 7.26610541e-01 1.78952128e-01
-4.04394925e-01 -6.16247952e-01 -9.76671278e-01 2.76837915e-01
5.55811286e-01 -5.43830335e-01 8.20353150e-01 -8.90462816e-01
-1.33371568e+00 7.72284091e-01 -1.03967629e-01 -6.61605835e-01
6.05739057e-01 -4.76309769e-02 -1.29021659e-01 5.01379550e-01
7.75508359e-02 6.81882262e-01 1.09209883e+00 -1.19457662e+00
-8.16963971e-01 -4.96442109e-01 4.81096417e-01 8.54738206e-02
-5.01121044e-01 -3.05900574e-01 -7.97725320e-01 -7.63603032e-01
3.45680088e-01 -8.43571603e-01 -5.72040081e-01 -2.84124911e-01
-5.62223911e-01 -1.84275553e-01 1.10612261e+00 -6.40966296e-01
1.41654932e+00 -2.09450364e+00 8.14046860e-02 4.91459131e-01
2.03543290e-01 4.96792346e-01 -1.28590256e-01 2.44322047e-01
1.28043488e-01 4.36765701e-01 -3.94254863e-01 -4.79802378e-02
1.91286132e-01 9.28066745e-02 7.77868256e-02 2.55034178e-01
5.12174428e-01 1.20327401e+00 -5.64663410e-01 -5.74861288e-01
3.44287246e-01 4.70876873e-01 -4.07111526e-01 1.04848230e-02
-1.27468854e-01 4.73208547e-01 -7.58555889e-01 8.31506193e-01
7.64960706e-01 -4.53471959e-01 -5.98993264e-02 -3.27959567e-01
-6.29082024e-02 1.57003105e-02 -1.26965368e+00 2.04384637e+00
-5.22941291e-01 3.96750420e-01 -8.86082370e-03 -1.12793803e+00
8.68103445e-01 7.50235915e-02 5.80259144e-01 -8.93825829e-01
-5.95317315e-03 1.52129591e-01 -5.11519704e-03 -3.23824555e-01
9.45595443e-01 4.16506797e-01 -1.07393235e-01 3.17930698e-01
5.32053672e-02 -7.18836933e-02 5.82147896e-01 2.66137898e-01
9.68043685e-01 2.14995593e-01 3.14664543e-01 -4.36852545e-01
6.51433170e-01 1.43689066e-01 4.78395760e-01 7.30908513e-01
-1.16771452e-01 6.12007201e-01 2.61365414e-01 -3.94892722e-01
-6.59605384e-01 -5.86261511e-01 -6.86841607e-01 8.60584438e-01
4.59083438e-01 -4.63698208e-01 -9.68380988e-01 -1.07650399e+00
5.79815283e-02 3.06994975e-01 -8.24118316e-01 1.14031896e-01
-6.84148848e-01 -5.36416054e-01 1.82741091e-01 8.70076120e-01
8.67543161e-01 -1.15998089e+00 -8.26653242e-01 2.96744347e-01
1.30795300e-01 -1.20459950e+00 -2.31829390e-01 2.77018875e-01
-1.03111422e+00 -1.07099628e+00 -7.74028003e-01 -9.01193798e-01
8.80416512e-01 1.09915413e-01 1.07453859e+00 3.18856686e-01
-7.33255446e-01 3.30253184e-01 -5.12024522e-01 -1.70174778e-01
1.77213460e-01 5.19520700e-01 -6.38322353e-01 -3.20442766e-02
3.51345539e-02 -2.11733624e-01 -1.01447773e+00 3.89119923e-01
-1.16803336e+00 -2.61104733e-01 6.15975261e-01 8.38737011e-01
9.59993005e-01 1.98693037e-01 5.86489439e-01 -1.04017830e+00
3.39422762e-01 -2.02563271e-01 -3.48328650e-01 2.31678769e-01
-2.51426280e-01 -2.76608374e-02 5.00510991e-01 -9.79819819e-02
-1.06219447e+00 6.13390580e-02 -9.90563110e-02 -1.10903673e-01
-5.07379711e-01 3.73630673e-01 -3.18711028e-02 -2.13173509e-01
2.90635586e-01 1.62887983e-02 -1.96797282e-01 -5.49926877e-01
4.02574986e-01 5.16411245e-01 7.04659000e-02 -4.97262836e-01
7.28399575e-01 5.45971930e-01 -1.74061861e-02 -9.48811948e-01
-6.99335039e-01 -5.75129211e-01 -8.69881928e-01 7.41373599e-02
6.25719845e-01 -6.53628588e-01 -4.35902029e-01 2.50189126e-01
-7.26478279e-01 -2.05292761e-01 -6.87580228e-01 1.28662139e-01
-4.05864507e-01 2.40522981e-01 -4.81512785e-01 -5.12264252e-01
-3.94815594e-01 -1.31867051e+00 1.10513616e+00 5.47602773e-01
1.78159419e-02 -8.13682854e-01 -4.25326318e-01 4.03301626e-01
6.26677454e-01 5.53564429e-01 6.81283534e-01 -9.40668046e-01
-8.76915157e-01 -5.34034550e-01 -5.76254070e-01 3.22325826e-01
1.96533039e-01 1.63149089e-01 -9.15848494e-01 -3.81264985e-01
-3.18872541e-01 -5.88080920e-02 1.12184894e+00 5.24086416e-01
1.62180507e+00 -1.43118864e-02 -5.85329652e-01 5.70817769e-01
1.59686279e+00 -6.28802925e-02 4.61294681e-01 5.38062930e-01
6.33233368e-01 6.11114025e-01 7.44878471e-01 3.94297302e-01
1.80481374e-01 7.07467556e-01 7.10447311e-01 -4.27769870e-01
-1.70815662e-01 1.95249423e-01 -4.44542706e-01 2.07755089e-01
-1.18342645e-01 -1.43571600e-01 -7.96600342e-01 6.58805728e-01
-1.86932719e+00 -7.96001017e-01 -6.95982501e-02 1.93770349e+00
2.80907840e-01 1.31342888e-01 2.63697088e-01 2.35962585e-01
7.60331333e-01 2.86681324e-01 -6.25718713e-01 -4.59541321e-01
1.76892355e-02 6.75821424e-01 4.18332070e-01 2.06954908e-02
-1.42186046e+00 9.18390632e-01 5.91648102e+00 1.12155390e+00
-9.93955374e-01 -5.61836101e-02 8.51129591e-01 -7.48758316e-02
1.28281459e-01 -2.41998285e-01 -8.65196705e-01 4.69592214e-01
4.11318332e-01 4.45188768e-02 5.90556003e-02 8.15023363e-01
-9.45336670e-02 -2.49612108e-01 -8.34883392e-01 6.33410394e-01
-5.70129342e-02 -1.39023161e+00 -4.94055226e-02 5.83041608e-02
8.02041471e-01 -2.62207806e-01 1.80208430e-01 4.13832784e-01
-2.22537816e-01 -1.06334686e+00 5.01362741e-01 3.66856664e-01
6.33840501e-01 -1.07876325e+00 6.85132861e-01 1.89153269e-01
-1.52018273e+00 -2.85825938e-01 -3.01236540e-01 2.19950914e-01
-2.00730525e-02 5.71777225e-01 -7.84727693e-01 8.31444085e-01
7.55753636e-01 7.38153279e-01 -7.80404925e-01 1.30604291e+00
-3.09360139e-02 3.94370735e-01 -5.00606239e-01 2.28041977e-01
5.31888127e-01 -6.88179657e-02 2.61878967e-01 1.52276444e+00
8.34188014e-02 -1.07043549e-01 3.45770597e-01 9.28210616e-01
-2.28648826e-01 3.67407799e-01 -3.26719940e-01 1.04132615e-01
2.26619363e-01 1.69507599e+00 -1.28074229e+00 -4.03343976e-01
-2.89727390e-01 9.43085253e-01 3.79859507e-01 2.80983478e-01
-7.82391548e-01 -5.80146194e-01 4.04346347e-01 7.29789957e-02
7.68226027e-01 3.56061049e-02 -1.89266145e-01 -8.50452602e-01
2.79791415e-01 -5.27785420e-01 3.42313439e-01 8.33408609e-02
-1.02021074e+00 9.15453017e-01 -1.60670221e-01 -1.24160910e+00
-9.33581591e-02 -3.51608574e-01 -5.42074621e-01 6.97043777e-01
-1.70185471e+00 -1.40022838e+00 -4.79265541e-01 5.60732126e-01
7.83296287e-01 1.88153476e-01 5.72052062e-01 2.95326710e-01
-9.09181595e-01 7.82436073e-01 -4.51944545e-02 3.46773528e-02
1.71395123e-01 -1.30945075e+00 4.13914889e-01 6.06453240e-01
1.77372947e-01 7.97454059e-01 2.41497695e-01 -3.46151024e-01
-9.09315109e-01 -1.27495825e+00 3.54558796e-01 -2.74761975e-01
3.75067919e-01 -2.39422530e-01 -1.05127597e+00 3.64596575e-01
2.03264371e-01 5.31432450e-01 5.21239877e-01 -4.70666103e-02
-5.46771884e-02 -5.49357049e-02 -1.51647055e+00 2.59932697e-01
1.11496496e+00 -3.16130221e-01 -3.95262361e-01 3.18838656e-01
5.53448081e-01 -6.01286829e-01 -1.17518270e+00 5.34784973e-01
4.60686684e-01 -9.14651096e-01 1.03493083e+00 -6.01254106e-01
4.37672049e-01 -1.59613252e-01 -1.13829806e-01 -9.64847207e-01
-2.84758806e-01 -2.93426126e-01 -2.01994494e-01 1.48098600e+00
4.82794702e-01 -3.90978307e-01 1.12495482e+00 7.11426795e-01
-4.04374339e-02 -1.21780324e+00 -7.84082830e-01 -8.96435738e-01
-2.05714583e-01 -7.25184828e-02 8.49644423e-01 8.08598459e-01
-2.55029798e-01 -1.15642957e-01 1.94078401e-01 -2.48258887e-03
3.73975903e-01 6.46487892e-01 5.56000650e-01 -1.22983992e+00
5.01361825e-02 -4.24378335e-01 -6.32207811e-01 -9.76626515e-01
1.78389147e-01 -1.01627493e+00 -1.94647744e-01 -1.82263434e+00
1.30794615e-01 -6.22495174e-01 -7.73632884e-01 2.39348099e-01
-1.48902208e-01 5.34995615e-01 3.18412960e-01 7.55037554e-03
-8.55212510e-01 2.69075334e-01 1.37877524e+00 -1.74132884e-01
-3.13418359e-01 3.66258100e-02 -4.38187450e-01 6.22056007e-01
9.80096698e-01 -2.36342922e-01 -2.07944989e-01 -2.80422360e-01
-2.65740156e-01 -2.13238355e-02 2.70520002e-01 -1.06321013e+00
8.22076499e-02 1.73685178e-01 4.73223656e-01 -7.04894125e-01
4.08287734e-01 -1.00247788e+00 -2.73068901e-02 2.69398361e-01
-4.43464249e-01 -9.49392244e-02 4.35190499e-01 6.79696620e-01
-4.62647885e-01 -3.80056381e-01 6.30260587e-01 -3.70966941e-01
-1.14348590e+00 5.06393909e-01 2.43105933e-01 2.22385023e-03
1.32527113e+00 -7.70410061e-01 -1.50691882e-01 2.58210450e-01
-8.01656306e-01 1.58857793e-01 5.36650777e-01 2.99319535e-01
5.10273993e-01 -1.05003762e+00 -3.87785256e-01 1.69957265e-01
2.17543766e-01 2.61676371e-01 4.61042762e-01 7.87592471e-01
-4.01202232e-01 4.87977535e-01 -2.72928745e-01 -6.89327657e-01
-1.12296104e+00 3.86359185e-01 4.67509106e-02 -5.54889321e-01
-7.51514435e-01 9.30479288e-01 2.20925301e-01 -3.35194230e-01
2.42747545e-01 -3.10324192e-01 -3.62251997e-01 3.00666124e-01
3.16919655e-01 4.47923481e-01 2.92681664e-01 -7.36289680e-01
-4.25133586e-01 5.80263138e-01 -4.30507064e-01 4.32440102e-01
1.63126528e+00 -1.00398354e-01 -7.29406849e-02 -2.45191291e-01
1.18773305e+00 -9.46230441e-02 -1.38585186e+00 -2.33778581e-01
3.37481290e-01 -5.24693251e-01 8.89919624e-02 -8.77856255e-01
-1.59004533e+00 6.45193756e-01 5.33118606e-01 3.97738367e-01
1.04819620e+00 -9.04754251e-02 8.27844501e-01 6.24099933e-02
3.56812716e-01 -1.24147844e+00 2.21534856e-02 2.58751512e-01
8.16037476e-01 -1.31758928e+00 1.64137676e-01 -5.16943157e-01
-4.56289709e-01 1.09655762e+00 6.86959445e-01 -3.93808544e-01
5.35190463e-01 1.56089410e-01 -3.35791945e-01 -4.48023707e-01
-2.58029461e-01 -5.11418521e-01 5.00787199e-01 4.20052797e-01
4.10529494e-01 -5.52044027e-02 -5.34508705e-01 5.26681721e-01
1.05308600e-01 -2.05290318e-01 1.89749852e-01 1.08915448e+00
-2.75447041e-01 -9.75170970e-01 -7.70908520e-02 8.60012472e-01
-6.00723684e-01 1.20355442e-01 -3.54107022e-01 1.11145043e+00
1.87520698e-01 7.33000636e-01 1.47365093e-01 -5.71810491e-02
4.39080030e-01 -2.18298361e-01 3.76509815e-01 -7.14096665e-01
-1.23594320e+00 -8.45298991e-02 -1.83798507e-01 -8.63871992e-01
-5.55086017e-01 -4.12232965e-01 -1.42220163e+00 2.03311712e-01
-4.73103195e-01 -2.54513673e-03 8.48147750e-01 8.66567791e-01
5.46325326e-01 8.41316283e-01 5.98352730e-01 -1.05239105e+00
-2.18308151e-01 -6.61938310e-01 -7.15996504e-01 6.19370580e-01
2.48603910e-01 -6.65256560e-01 -1.03278041e-01 -3.62221897e-01] | [9.603459358215332, 0.24133743345737457] |
ef8d4b93-584e-4bc0-894b-475e0ede3e15 | what-truly-matters-in-trajectory-prediction | 2306.15136 | null | https://arxiv.org/abs/2306.15136v1 | https://arxiv.org/pdf/2306.15136v1.pdf | What Truly Matters in Trajectory Prediction for Autonomous Driving? | In the autonomous driving system, trajectory prediction plays a vital role in ensuring safety and facilitating smooth navigation. However, we observe a substantial discrepancy between the accuracy of predictors on fixed datasets and their driving performance when used in downstream tasks. This discrepancy arises from two overlooked factors in the current evaluation protocols of trajectory prediction: 1) the dynamics gap between the dataset and real driving scenario; and 2) the computational efficiency of predictors. In real-world scenarios, prediction algorithms influence the behavior of autonomous vehicles, which, in turn, alter the behaviors of other agents on the road. This interaction results in predictor-specific dynamics that directly impact prediction results. As other agents' responses are predetermined on datasets, a significant dynamics gap arises between evaluations conducted on fixed datasets and actual driving scenarios. Furthermore, focusing solely on accuracy fails to address the demand for computational efficiency, which is critical for the real-time response required by the autonomous driving system. Therefore, in this paper, we demonstrate that an interactive, task-driven evaluation approach for trajectory prediction is crucial to reflect its efficacy for autonomous driving. | ['David Hsu', 'Sifa Zheng', 'Panpan Cai', 'Cunjun Yu', 'Tran Phong', 'Haoran Wu'] | 2023-06-27 | null | null | null | null | ['trajectory-prediction', 'autonomous-vehicles'] | ['computer-vision', 'computer-vision'] | [-1.27003014e-01 -2.03847349e-01 -2.62395352e-01 -4.81775343e-01
-3.16622108e-01 -6.99592590e-01 6.82840824e-01 2.22499445e-01
-4.64640796e-01 6.25751078e-01 2.26479247e-01 -8.08242261e-01
-3.81343096e-01 -8.63942266e-01 -5.97916186e-01 -4.90449548e-01
-1.75451458e-01 4.57469881e-01 5.93124866e-01 -5.30705869e-01
2.39721015e-01 5.93544483e-01 -2.06134629e+00 -2.53544338e-02
7.26109207e-01 7.00915039e-01 3.63010503e-02 7.71761000e-01
1.57828271e-01 8.64273369e-01 -2.64688611e-01 -1.96506847e-02
2.59258062e-01 -1.03331506e-01 -2.38982499e-01 -2.67164230e-01
-9.03808922e-02 -4.27571625e-01 -5.25954425e-01 5.56175709e-01
9.28763300e-02 2.07150444e-01 6.93721414e-01 -1.75274754e+00
2.48397559e-01 3.61272871e-01 1.94157138e-02 2.73608953e-01
1.09739311e-01 7.00650454e-01 7.04369783e-01 -4.42453653e-01
5.51520407e-01 8.71661007e-01 5.71447670e-01 1.93830267e-01
-1.16736317e+00 -6.50575578e-01 2.68609077e-01 4.29368377e-01
-1.16314805e+00 -7.59501576e-01 4.58112836e-01 -8.41811419e-01
6.54080272e-01 2.19052002e-01 5.13197243e-01 8.73361349e-01
5.12072206e-01 3.83379310e-01 6.44929647e-01 1.47223189e-01
4.14215088e-01 4.45458561e-01 2.25909203e-01 1.00468740e-01
9.86337066e-02 7.31289983e-01 -4.43656713e-01 -4.24066884e-03
-2.98118386e-02 -1.26768753e-01 -1.96010336e-01 -3.71857464e-01
-1.16615450e+00 6.06908858e-01 1.03160098e-01 3.68693247e-02
-7.08311439e-01 2.20824167e-01 4.28469896e-01 5.77993691e-01
5.05920686e-02 2.00701460e-01 -3.18441033e-01 -8.91759098e-01
-6.84178889e-01 6.95463240e-01 7.76816428e-01 8.27532649e-01
7.67221928e-01 -5.41862100e-02 -1.50030985e-01 1.42848551e-01
1.18678555e-01 5.55708766e-01 5.12945130e-02 -9.51315165e-01
4.42419112e-01 5.78865528e-01 5.47888458e-01 -1.14034641e+00
-5.40041029e-01 -3.95065427e-01 -1.74575269e-01 5.19944072e-01
7.88943112e-01 -1.65640265e-01 -3.96001875e-01 1.79389977e+00
3.73369485e-01 6.55808523e-02 2.04557523e-01 1.11754525e+00
2.75932819e-01 5.42816222e-01 3.78818274e-01 -7.97026418e-03
9.99766409e-01 -6.09471798e-01 -5.25700867e-01 -1.16028056e-01
1.10795891e+00 -6.83982134e-01 1.01016366e+00 3.67195308e-02
-8.70844424e-01 -5.95912993e-01 -8.80119681e-01 5.64186394e-01
-4.50018644e-01 -5.23947515e-02 7.03992620e-02 4.70791399e-01
-8.05725455e-01 3.47211659e-01 -8.11218679e-01 -5.19375861e-01
-5.64173758e-02 4.64731634e-01 -8.70064870e-02 6.23391792e-02
-1.22253215e+00 1.29136181e+00 -5.10157496e-02 -4.51792292e-02
-8.01998019e-01 -8.79485667e-01 -3.67951274e-01 2.54636873e-02
1.97878480e-01 -2.99345642e-01 1.31412673e+00 -4.95901912e-01
-1.20308852e+00 2.67309994e-01 -2.97501117e-01 -6.03787303e-01
1.12552941e+00 3.15262261e-03 -6.23668134e-01 -5.23779988e-01
2.53560077e-02 4.33833510e-01 3.00933838e-01 -1.29079962e+00
-1.24703145e+00 -3.02185178e-01 2.36578239e-03 3.13143998e-01
1.19848266e-01 -3.87623370e-01 -2.27036193e-01 2.19637677e-01
-3.16900641e-01 -1.24532688e+00 -3.59632075e-01 -3.55368733e-01
8.97607431e-02 -2.79977709e-01 1.16156709e+00 -2.06081510e-01
1.35660553e+00 -2.37644696e+00 -3.32972199e-01 4.50128347e-01
1.25823006e-01 1.53002352e-01 -1.88907236e-01 8.41442168e-01
3.66608024e-01 -1.16863884e-01 2.52726823e-01 -6.39013723e-02
-2.39471346e-02 1.71261951e-01 -6.42766714e-01 4.86859500e-01
5.61139360e-02 7.28368998e-01 -8.27912271e-01 -8.80476460e-02
5.09423792e-01 2.55656302e-01 -3.87357980e-01 1.75276577e-01
-1.56222627e-01 8.39048147e-01 -5.73285162e-01 9.18224752e-02
4.31693763e-01 2.53746659e-01 -5.30565903e-02 1.43071026e-01
-7.74874151e-01 5.55376112e-01 -9.23424006e-01 8.24808776e-01
-5.02798855e-01 9.11383450e-01 -1.00154862e-01 -6.18887305e-01
9.46635783e-01 9.55925360e-02 6.76559448e-01 -1.13939333e+00
9.71404910e-02 2.98263609e-01 4.57523346e-01 -5.09908259e-01
1.00832498e+00 3.78700495e-01 -3.17660309e-02 5.53163052e-01
-9.23736036e-01 1.32883251e-01 3.71248811e-01 2.66760066e-02
1.22312057e+00 2.16875505e-02 -2.53413051e-01 -3.29948574e-01
3.27743709e-01 8.68968248e-01 5.32378435e-01 5.42000830e-01
-5.75750947e-01 9.04896855e-02 7.17846394e-01 -7.12791920e-01
-1.20372176e+00 -7.89397717e-01 -1.65262699e-01 9.43775058e-01
5.81401169e-01 -1.68096960e-01 -5.41605413e-01 -3.47637683e-01
2.83792913e-01 1.20562279e+00 -6.49983704e-01 -3.78999859e-01
-5.97755134e-01 -3.29450190e-01 4.53155130e-01 3.55461329e-01
1.37911327e-02 -8.02008331e-01 -1.24951768e+00 6.61774933e-01
-1.46876007e-01 -1.31659806e+00 -1.81028351e-01 -1.23594955e-01
-4.32047784e-01 -1.13526809e+00 2.42029965e-01 -1.00258306e-01
5.22793233e-01 6.44870639e-01 9.65810716e-01 1.90226153e-01
4.30786252e-01 4.32678610e-01 -1.80361763e-01 -5.35520017e-01
-8.41329098e-01 3.59761953e-01 1.49251930e-02 -1.10941656e-01
5.58547676e-01 -4.77761596e-01 -8.04874837e-01 9.56035912e-01
-5.54974020e-01 1.26565531e-01 4.90547448e-01 2.93684810e-01
3.15271884e-01 2.92081833e-01 7.29474366e-01 -5.54219484e-01
7.50335038e-01 -8.38688374e-01 -8.76556396e-01 4.36976403e-02
-9.99360263e-01 -7.77908936e-02 7.36749709e-01 -3.62429231e-01
-6.67277753e-01 1.37606800e-01 9.96998399e-02 -1.27751246e-01
-3.49291831e-01 4.81387675e-01 2.32800379e-01 2.12531313e-01
6.60524130e-01 3.22386235e-01 4.22746062e-01 1.32300377e-01
1.15142599e-01 7.18938589e-01 3.48103881e-01 -2.37259209e-01
7.14384854e-01 4.23376471e-01 4.54910323e-02 -6.87121809e-01
5.15838563e-02 -3.11623812e-01 -3.57500672e-01 -6.76680505e-01
5.10644853e-01 -9.68206286e-01 -1.08304906e+00 1.29447624e-01
-9.13111925e-01 -8.68483126e-01 -1.34388566e-01 5.91159344e-01
-5.72460115e-01 -2.72130817e-01 -1.23455137e-01 -8.85517895e-01
4.42318991e-02 -1.62333930e+00 6.51174903e-01 9.14213285e-02
-5.57317793e-01 -8.43173981e-01 1.53197601e-01 2.31868654e-01
8.20182264e-01 6.32599294e-02 8.02865624e-01 -4.39965010e-01
-8.64801884e-01 -3.58280718e-01 -2.22711727e-01 -3.58627170e-01
-2.14775562e-01 3.91423106e-01 -6.56885087e-01 -2.46707723e-01
-4.65829730e-01 1.45308837e-01 1.84080407e-01 2.16694847e-01
4.79462802e-01 8.66758674e-02 -6.00849986e-01 -5.48224784e-02
1.19017363e+00 4.62529361e-01 5.83519101e-01 4.69299883e-01
2.24866256e-01 1.17917311e+00 1.13326943e+00 3.13780040e-01
9.82162476e-01 9.36011970e-01 5.89847207e-01 4.53537554e-02
6.53826073e-02 -3.36658448e-01 5.02633214e-01 3.51857215e-01
8.03895444e-02 -2.94422597e-01 -1.31663060e+00 6.87625587e-01
-2.20112276e+00 -9.57816780e-01 -7.15501308e-01 2.50776792e+00
1.23813614e-01 3.61473709e-01 4.83176410e-01 2.85405805e-03
4.35003996e-01 -1.30651027e-01 -7.47401893e-01 -4.51003224e-01
3.36740553e-01 -8.58117819e-01 8.75615716e-01 5.58418989e-01
-5.16447425e-01 7.13592112e-01 6.05296946e+00 4.26003575e-01
-1.49927366e+00 6.84650755e-03 4.93761569e-01 -1.46689042e-01
-4.25388694e-01 4.39017229e-02 -7.66146481e-01 9.31099653e-01
1.53346169e+00 -5.46748579e-01 4.13367718e-01 8.49050045e-01
1.08067846e+00 -2.91100889e-01 -1.19716477e+00 4.55429792e-01
-7.29895651e-01 -1.20080054e+00 -4.41353112e-01 5.27240634e-01
4.39857334e-01 5.61337471e-01 1.68160647e-01 3.68291855e-01
4.51235324e-01 -9.68053460e-01 9.65606689e-01 4.31954026e-01
4.13830787e-01 -8.38226438e-01 7.21419930e-01 6.09574139e-01
-1.11988473e+00 -2.09275603e-01 -5.34682274e-02 -3.28235358e-01
3.70674759e-01 4.23854321e-01 -1.03687549e+00 1.89393316e-03
3.18330497e-01 3.40850145e-01 -2.23789528e-01 7.69156277e-01
1.01047583e-01 7.86673486e-01 -2.84434438e-01 -2.01198444e-01
5.37027717e-01 -1.93558186e-01 6.37104332e-01 9.88948047e-01
2.10521802e-01 1.82805862e-02 1.47130743e-01 5.93206763e-01
4.64212775e-01 -7.65211284e-02 -9.82357442e-01 1.09426312e-01
7.77382910e-01 9.96436179e-01 -4.36694384e-01 -6.48877472e-02
-4.40281063e-01 -9.28965285e-02 8.90239701e-02 5.20639837e-01
-1.11151969e+00 1.52791724e-01 1.48173892e+00 6.40650690e-01
9.44749787e-02 -5.90978861e-01 -5.90074062e-01 -4.11801308e-01
-2.43239719e-02 -6.80215895e-01 -5.77883013e-02 -4.26695496e-01
-5.51524699e-01 5.22899806e-01 -1.56642377e-01 -1.56207573e+00
-5.22836506e-01 -1.20728105e-01 -6.23556495e-01 7.87661612e-01
-1.63736928e+00 -8.70911658e-01 -4.56735939e-01 4.33254957e-01
4.34346765e-01 5.11630846e-04 3.85652810e-01 4.65638340e-01
-4.26568180e-01 3.30900401e-01 1.53716296e-01 -4.31615710e-01
5.04308939e-01 -6.47593021e-01 6.48366570e-01 7.54402876e-01
-3.61005247e-01 1.24520585e-01 1.15052199e+00 -5.58280647e-01
-1.56088412e+00 -1.18316996e+00 8.74403238e-01 -7.56927431e-01
6.49459600e-01 -5.61301969e-02 -9.21640456e-01 3.88283581e-01
-1.32454842e-01 -2.63563544e-01 4.24935371e-01 8.03989917e-02
2.50248671e-01 -2.56671548e-01 -8.99347603e-01 8.74264419e-01
1.02342093e+00 -2.76029706e-01 2.34225005e-01 -1.22425452e-01
4.71574724e-01 -3.81976873e-01 -6.13645315e-01 4.95586038e-01
7.68658519e-01 -1.05474472e+00 4.18875575e-01 -6.41515315e-01
4.37124252e-01 -4.75187093e-01 -1.02942847e-01 -1.22482896e+00
-2.29384929e-01 -2.80179381e-01 1.82430685e-01 8.09247315e-01
8.08498085e-01 -8.71520638e-01 9.65705574e-01 1.32861185e+00
1.92584321e-02 -9.13241625e-01 -7.74934649e-01 -5.63214362e-01
8.06120485e-02 -9.00050819e-01 9.73442316e-01 5.47546089e-01
-7.21905604e-02 -6.98810443e-02 -1.30135909e-01 3.54401827e-01
3.41474324e-01 4.27332371e-02 1.42740548e+00 -9.28262234e-01
3.05652499e-01 -4.96363401e-01 -6.18611157e-01 -9.85250592e-01
6.39919564e-03 -2.74659067e-01 2.16906264e-01 -1.13289821e+00
-2.65699506e-01 -1.09501481e+00 -2.30956431e-02 5.72791249e-02
1.10526986e-01 -2.04929426e-01 1.72807321e-01 5.65748692e-01
-3.52819234e-01 4.29212540e-01 8.07671607e-01 2.00682744e-01
-5.60601115e-01 3.37632626e-01 -2.98878431e-01 3.11961412e-01
9.76522207e-01 -5.21121919e-01 -7.98539460e-01 -4.78012145e-01
2.06602991e-01 2.86734641e-01 2.94505566e-01 -1.17138195e+00
6.44851029e-01 -5.73706031e-01 -4.11311746e-01 -7.20899105e-01
1.42057687e-01 -1.02827406e+00 5.65655768e-01 7.50140607e-01
-2.89309770e-01 3.59693795e-01 2.43903980e-01 6.34036481e-01
-2.19967946e-01 2.77536094e-01 4.52796608e-01 5.36092699e-01
-9.00216818e-01 3.39913815e-01 -9.86067653e-01 -1.76870555e-01
1.38236928e+00 -5.79945445e-01 -3.43132228e-01 -6.82375848e-01
-3.95804286e-01 7.34303176e-01 6.80941284e-01 7.06435442e-01
1.98526785e-01 -1.03140044e+00 -7.54830539e-01 2.46017456e-01
2.76013017e-01 -2.95534164e-01 2.10028902e-01 1.22244179e+00
-3.41559917e-01 5.66217780e-01 -1.46447822e-01 -7.51769483e-01
-1.08799803e+00 1.74565822e-01 3.74149680e-01 -2.00398311e-01
-5.17290950e-01 -4.35465798e-02 1.63085684e-01 -3.29435289e-01
1.22249283e-01 -2.08338335e-01 -1.70656621e-01 -1.10765360e-01
2.56248295e-01 5.50803185e-01 2.09705889e-01 -7.66195655e-01
-2.68027663e-01 1.35111362e-01 6.79917708e-02 -1.83110833e-01
9.43727851e-01 -4.36255187e-01 5.16760647e-01 4.56205994e-01
7.40605533e-01 -7.97716975e-02 -1.53810179e+00 1.06515633e-02
6.63579330e-02 -5.01146257e-01 1.12463526e-01 -5.92360556e-01
-7.46081114e-01 5.32739520e-01 5.50291717e-01 4.00545120e-01
6.49423957e-01 -2.94667989e-01 9.85607564e-01 9.74869877e-02
6.45008087e-01 -1.08038628e+00 -5.48334241e-01 2.43976012e-01
6.28788650e-01 -1.39436162e+00 -3.34478021e-01 -3.70202094e-01
-7.40317762e-01 8.90288889e-01 6.67313337e-01 3.08922082e-01
4.81244355e-01 2.74630606e-01 4.59100515e-01 -7.68058896e-02
-1.24528182e+00 -2.09502533e-01 -2.02489957e-01 5.73313594e-01
1.28362060e-01 3.52998555e-01 -4.17234153e-01 1.30959988e-01
-3.92582476e-01 1.14099510e-01 5.76341689e-01 7.21044660e-01
-4.25480068e-01 -9.94375944e-01 -9.49218795e-02 3.91512752e-01
1.05019711e-01 5.27210772e-01 -8.60054046e-02 8.69816542e-01
5.35898991e-02 1.24692309e+00 3.96668166e-01 -6.97361112e-01
6.30752861e-01 -3.63254368e-01 -2.92201221e-01 -1.08394094e-01
-4.76761699e-01 -7.48412549e-01 4.62241620e-01 -7.18970180e-01
5.07135838e-02 -9.90650833e-01 -1.36973476e+00 -1.01346636e+00
2.76280325e-02 3.66042167e-01 9.97948110e-01 9.81654763e-01
8.88161600e-01 3.63628328e-01 7.17439890e-01 -6.10566795e-01
-6.30079985e-01 -4.73805487e-01 -2.03705788e-01 3.36253494e-01
2.84584105e-01 -7.83089519e-01 -3.30410153e-01 -2.47111425e-01] | [5.748108863830566, 1.0744520425796509] |
d2e6e689-8eb9-4295-a85a-f684ccf3db3d | portfolio-cuts-a-graph-theoretic-framework-to | 1910.05561 | null | https://arxiv.org/abs/1910.05561v3 | https://arxiv.org/pdf/1910.05561v3.pdf | Portfolio Cuts: A Graph-Theoretic Framework to Diversification | Investment returns naturally reside on irregular domains, however, standard multivariate portfolio optimization methods are agnostic to data structure. To this end, we investigate ways for domain knowledge to be conveniently incorporated into the analysis, by means of graphs. Next, to relax the assumption of the completeness of graph topology and to equip the graph model with practically relevant physical intuition, we introduce the portfolio cut paradigm. Such a graph-theoretic portfolio partitioning technique is shown to allow the investor to devise robust and tractable asset allocation schemes, by virtue of a rigorous graph framework for considering smaller, computationally feasible, and economically meaningful clusters of assets, based on graph cuts. In turn, this makes it possible to fully utilize the asset returns covariance matrix for constructing the portfolio, even without the requirement for its inversion. The advantages of the proposed framework over traditional methods are demonstrated through numerical simulations based on real-world price data. | ['Danilo P. Mandic', 'Anthony G. Constantinides', 'Ljubisa Stankovic', 'Bruno Scalzo Dees'] | 2019-10-12 | null | null | null | null | ['physical-intuition'] | ['reasoning'] | [-1.39391780e-01 2.62324572e-01 -1.22815982e-01 -3.23931500e-02
-7.09267259e-02 -1.00005722e+00 3.33934516e-01 1.35713875e-01
2.97412753e-01 9.25501347e-01 -4.11484540e-01 -6.98957324e-01
-1.17716813e+00 -1.22181165e+00 -4.49070543e-01 -7.52366126e-01
-2.21742481e-01 2.67899275e-01 -2.82967597e-01 8.18778295e-03
3.06402266e-01 7.95635760e-01 -9.60932016e-01 -6.19686186e-01
9.58105028e-01 1.15253913e+00 -2.52214849e-01 4.97867726e-02
1.46922439e-01 5.39263785e-01 -1.60458401e-01 -1.00441551e+00
7.40836143e-01 -3.52065533e-01 -2.25125834e-01 5.25351644e-01
-2.30628237e-01 -2.72492737e-01 -3.03863883e-01 1.26127577e+00
2.13038281e-01 2.95836478e-01 8.95301580e-01 -1.13587606e+00
-4.58796114e-01 6.65031314e-01 -7.21252084e-01 1.02287911e-01
-7.68632581e-03 -9.54500064e-02 1.38926613e+00 -5.86151302e-01
3.86328697e-01 7.54079521e-01 4.56606925e-01 -1.25792414e-01
-1.20077193e+00 -2.96449035e-01 1.71995223e-01 -3.05442631e-01
-1.26459944e+00 -1.56194538e-01 1.37800944e+00 -5.07257938e-01
4.87274647e-01 4.96792257e-01 7.50070632e-01 2.87214845e-01
7.47307613e-02 1.46927401e-01 8.94827485e-01 -4.29278642e-01
3.40581715e-01 2.62376249e-01 3.01028907e-01 5.35661757e-01
1.22303724e+00 2.44340062e-01 -8.31846073e-02 -2.85037130e-01
8.01612139e-01 9.53751653e-02 -4.00788218e-01 -1.05218661e+00
-7.35269070e-01 8.49840999e-01 1.64381802e-01 2.21178725e-01
-6.00174010e-01 -3.60329710e-02 -3.14227454e-02 2.82679170e-01
4.38613534e-01 5.25139272e-01 1.11315183e-01 3.24890673e-01
-7.68595576e-01 1.97304592e-01 9.19646382e-01 1.16694617e+00
5.05216599e-01 3.77718389e-01 1.79032937e-01 2.35898688e-01
6.22729242e-01 4.83686715e-01 -2.95130223e-01 -7.72610784e-01
9.11067724e-01 6.13288283e-01 3.54950994e-01 -1.40831137e+00
-2.85654485e-01 -8.44213247e-01 -7.14834034e-01 1.48819059e-01
5.24468303e-01 -4.37805772e-01 -1.79030985e-01 1.39954329e+00
4.46290255e-01 1.18205979e-01 8.96526650e-02 8.08650672e-01
-9.48953703e-02 1.64315134e-01 -3.21957231e-01 -4.25998032e-01
1.15317702e+00 -4.78119582e-01 -4.13134336e-01 2.50362307e-01
3.73340994e-01 -2.05838606e-01 6.49239004e-01 3.70304078e-01
-9.66689527e-01 -6.26506582e-02 -1.20006335e+00 7.18879044e-01
-1.08630590e-01 -1.89601898e-01 1.04373825e+00 1.15417385e+00
-8.63730669e-01 7.02359676e-01 -5.91535926e-01 7.77035905e-03
3.38374496e-01 2.88403064e-01 1.22116402e-01 2.31349856e-01
-1.01722598e+00 5.29289603e-01 4.59474146e-01 5.01178622e-01
-5.10358214e-01 -5.73943317e-01 -5.57061851e-01 4.67650592e-01
6.41889095e-01 -9.83199835e-01 8.01194370e-01 -8.78571928e-01
-1.33871961e+00 2.59652942e-01 3.76399577e-01 -5.97404301e-01
6.97718501e-01 2.67615110e-01 -3.42030764e-01 4.96175617e-01
-3.11547279e-01 -5.69935203e-01 1.00652015e+00 -1.06978106e+00
-2.83401221e-01 -2.79837966e-01 4.53132331e-01 1.30967364e-01
-4.32847172e-01 -3.32763255e-01 7.79620931e-02 -7.61386395e-01
-1.08107580e-02 -5.77339888e-01 -3.97706181e-01 -4.18390810e-01
-5.79096913e-01 2.27391168e-01 1.56703159e-01 -4.11733091e-01
1.17584097e+00 -1.89670384e+00 6.81823865e-02 1.12609601e+00
5.16113162e-01 -1.43367171e-01 3.17806095e-01 6.87087476e-01
-3.02331686e-01 2.14142323e-01 -3.41410547e-01 -1.61213130e-02
3.84316206e-01 -1.80212036e-01 -3.30567688e-01 8.17211092e-01
2.25053772e-01 8.81411612e-01 -6.27336144e-01 -1.16819985e-01
2.41310269e-01 -1.88972965e-01 -5.66358745e-01 -2.01574545e-02
1.52321085e-02 1.75518230e-01 -1.14200246e+00 7.13948905e-01
7.93101251e-01 -3.45154405e-01 7.26651967e-01 1.72186151e-01
3.73204947e-02 3.11008431e-02 -1.39981914e+00 9.05457914e-01
-1.70102328e-01 1.36447027e-01 2.23381639e-01 -1.58801854e+00
9.33278263e-01 1.13636188e-01 6.48611724e-01 -4.02643174e-01
1.55900568e-01 7.70096481e-02 -1.05062304e-02 -3.11006103e-02
2.71823227e-01 -4.49875742e-01 -9.62044224e-02 6.62242949e-01
-2.88027287e-01 1.23011857e-01 4.41540629e-01 5.81548065e-02
8.68106782e-01 -7.80396163e-02 3.33704799e-01 -7.00573623e-01
4.85247850e-01 -1.71073765e-01 3.46120507e-01 5.30337930e-01
2.11259827e-01 1.49026737e-02 9.41588819e-01 -7.97828287e-02
-1.01432800e+00 -1.16537333e+00 -1.66686758e-01 1.60798490e-01
5.07869497e-02 1.51445910e-01 -6.89350605e-01 -5.87298214e-01
5.21000445e-01 6.56661928e-01 -4.42214608e-01 7.05035031e-02
-2.83802241e-01 -1.13509321e+00 -2.89372914e-03 2.45988101e-01
1.91645369e-01 -5.31133056e-01 -5.77094615e-01 2.84992099e-01
3.87254357e-01 -8.42145324e-01 -1.55095547e-01 7.23358020e-02
-1.06625319e+00 -1.23737693e+00 -7.29493320e-01 -2.06201494e-01
8.55589688e-01 3.28196913e-01 1.05820036e+00 6.36490732e-02
-7.48131350e-02 7.93123722e-01 -2.52155006e-01 -2.36026943e-01
9.06640813e-02 -1.27698451e-01 9.34060737e-02 4.60576355e-01
-4.85655479e-02 -9.38316762e-01 -6.57676637e-01 6.12653568e-02
-8.13893974e-01 -2.53146648e-01 6.07955337e-01 5.70614636e-01
5.19459844e-01 7.62581825e-01 1.04157972e+00 -1.05316269e+00
8.13862681e-01 -6.84076846e-01 -1.36431229e+00 6.68241322e-01
-1.02281570e+00 -3.69964838e-02 6.06095910e-01 8.40082299e-03
-1.30774319e+00 -3.03979605e-01 7.04676867e-01 -3.16318840e-01
4.25313920e-01 9.09039795e-01 -4.84154642e-01 -3.99903446e-01
-1.38774076e-02 8.77116099e-02 -2.87739243e-02 -3.15319866e-01
6.64584279e-01 1.22084543e-01 1.88298941e-01 -8.69111061e-01
1.26014245e+00 4.06549811e-01 7.20471740e-01 -5.57236910e-01
-2.60976851e-01 -1.60921842e-01 -7.04687119e-01 -2.78964818e-01
4.57588464e-01 -5.43747187e-01 -8.43798101e-01 1.73425544e-02
-5.45754254e-01 4.07746732e-02 -2.89346337e-01 5.96082270e-01
-5.74960113e-01 6.36713088e-01 -2.49033332e-01 -1.31175005e+00
-2.16316730e-02 -6.04410410e-01 2.93808728e-01 7.53184855e-02
6.63434714e-02 -1.53292024e+00 -9.02792588e-02 2.02191174e-01
7.19462708e-02 7.24700987e-01 1.12728715e+00 -6.96197331e-01
-1.09018207e+00 -2.58106619e-01 -2.90750772e-01 1.73686743e-01
3.64659548e-01 2.97928508e-02 -6.15539193e-01 -2.97095925e-01
4.85610098e-01 3.88479680e-01 4.45076019e-01 3.72447252e-01
9.29277480e-01 -4.68528628e-01 -2.22221091e-01 7.12099910e-01
1.76744068e+00 3.31058502e-01 3.35846722e-01 3.12484801e-01
4.94288415e-01 9.50750828e-01 6.26049936e-01 8.95296872e-01
2.00564742e-01 3.18073511e-01 3.72791380e-01 6.68247491e-02
5.00139713e-01 -1.51190475e-01 -1.50048852e-01 7.46641874e-01
-3.93254310e-01 -5.23386478e-01 -7.32003093e-01 3.04004937e-01
-1.79230154e+00 -9.08930838e-01 -1.11649804e-01 2.40767479e+00
1.95872709e-01 2.73448527e-01 3.36940020e-01 1.32233158e-01
7.64919996e-01 1.14834242e-01 -5.13816774e-01 -2.58305222e-01
-1.39857024e-01 2.22049028e-01 1.02163815e+00 1.98881254e-01
-9.14752305e-01 1.27241462e-01 5.83571815e+00 6.95827246e-01
-5.64552367e-01 -3.73187810e-01 6.08587384e-01 1.25860041e-02
-9.36527908e-01 2.33781800e-01 -4.52012300e-01 4.25587416e-01
6.65802956e-01 -8.52355719e-01 6.39523447e-01 7.68711209e-01
3.47593427e-01 4.85213101e-02 -1.02562964e+00 5.75563908e-01
-2.96456456e-01 -1.17749918e+00 -8.77481550e-02 4.26392287e-01
5.97599268e-01 -8.31862092e-01 2.74678439e-01 7.66821147e-04
2.02948064e-01 -8.20643723e-01 6.57970488e-01 8.73324096e-01
4.27238584e-01 -1.20011055e+00 5.90930402e-01 1.32132962e-01
-1.31315029e+00 -2.03361526e-01 -4.77339923e-01 -1.42572686e-01
2.81967372e-01 8.54386091e-01 -6.73874021e-01 1.39280808e+00
1.71245113e-01 5.92189252e-01 -3.85125071e-01 1.08876467e+00
3.09547726e-02 4.54988360e-01 -3.39774072e-01 -7.91639909e-02
1.43811554e-01 -9.74346459e-01 5.38882494e-01 7.65435219e-01
6.11767769e-01 2.95799285e-01 -2.06747800e-01 1.06564760e+00
-2.34061740e-02 1.75054297e-01 -9.18583572e-01 -5.52148819e-01
4.61999238e-01 1.17099679e+00 -1.04590082e+00 2.63284128e-02
-7.90991902e-01 2.69429743e-01 1.51131809e-01 6.70906663e-01
-8.58717680e-01 -5.23339152e-01 3.22327793e-01 2.51576155e-01
4.32603210e-01 -3.18297923e-01 -5.32181978e-01 -1.02853656e+00
4.32006836e-01 -7.77283549e-01 4.58740324e-01 -1.14150956e-01
-1.50206971e+00 5.77078527e-03 2.17437103e-01 -1.44238353e+00
-3.62770766e-01 -6.89837515e-01 -7.41146207e-01 9.20028150e-01
-1.75262535e+00 -8.14302087e-01 2.00112507e-01 5.15198112e-01
-2.55204767e-01 -2.95873463e-01 2.77874738e-01 1.60546392e-01
-7.89273560e-01 5.17180204e-01 4.85227257e-01 -2.44976394e-02
-8.35119486e-02 -1.28873181e+00 1.27318323e-01 1.00633597e+00
2.05632091e-01 7.91909873e-01 5.30287385e-01 -8.21106315e-01
-1.67742646e+00 -8.69259119e-01 2.87829101e-01 -2.96943784e-01
1.22965252e+00 -1.90640882e-01 -6.94366217e-01 5.22722423e-01
8.06222335e-02 -1.81164563e-01 6.12367213e-01 8.97488222e-02
9.40122306e-02 -4.17045981e-01 -1.19564033e+00 5.31002581e-01
9.35912549e-01 -3.92167002e-01 -6.17427945e-01 2.75724441e-01
5.37816346e-01 8.52634758e-02 -1.23272324e+00 3.13026220e-01
3.96659076e-01 -1.03162646e+00 1.18845689e+00 -5.39901614e-01
-1.33446887e-01 -1.77465081e-01 -1.72662571e-01 -8.68546665e-01
-3.02579671e-01 -9.53009725e-01 -2.01852515e-01 1.46974409e+00
4.91546452e-01 -1.21183670e+00 8.98354948e-01 8.70691895e-01
1.09973028e-01 -5.81474066e-01 -9.23121095e-01 -1.05789173e+00
9.30628628e-02 -2.79420435e-01 9.94067132e-01 1.04330587e+00
2.43659958e-01 -7.36143142e-02 -2.58187890e-01 3.58720452e-01
1.13703609e+00 6.17460668e-01 4.67036396e-01 -1.31831956e+00
-7.66629517e-01 -6.74638927e-01 -2.53510475e-01 -6.45476580e-01
1.63192168e-01 -8.11331987e-01 -5.68746090e-01 -1.07284033e+00
4.63166684e-02 -8.07669997e-01 -5.45042336e-01 -1.24637000e-01
5.89849241e-02 -1.74028739e-01 2.90161848e-01 1.40273988e-01
-2.95406640e-01 5.65815985e-01 1.27541375e+00 9.52204391e-02
-2.70160660e-02 4.12107438e-01 -9.41686571e-01 6.29811645e-01
7.76309311e-01 -2.47085929e-01 -1.00561845e+00 4.12131064e-02
4.16887850e-01 6.54115081e-01 2.81809509e-01 -5.14566600e-01
-1.77639797e-02 -4.75865096e-01 1.36545718e-01 -3.01109046e-01
3.78764072e-03 -1.09443057e+00 6.14434123e-01 1.96325853e-01
6.31962568e-02 4.51518781e-03 -7.26762339e-02 8.86336923e-01
-1.14914134e-01 -5.13383031e-01 2.97835410e-01 1.22858778e-01
-2.87018776e-01 2.90815175e-01 6.09665774e-02 -9.93556529e-02
1.41446757e+00 -2.81981379e-01 -9.51596200e-02 -4.23696399e-01
-6.95510328e-01 2.66384572e-01 4.97936726e-01 -2.24449992e-01
4.12977487e-01 -1.38569021e+00 -3.32525760e-01 2.66493838e-02
-2.94244051e-01 -3.22807640e-01 2.56037831e-01 8.60224366e-01
-4.91863608e-01 8.75626981e-01 3.70282605e-02 -1.08088121e-01
-4.88197118e-01 9.26596284e-01 3.29479307e-01 -4.10184205e-01
-4.95191842e-01 2.91478962e-01 5.06694198e-01 1.12350702e-01
-1.46379009e-01 -3.69536310e-01 -1.50699869e-01 2.14061439e-01
1.03000773e-03 5.13843477e-01 4.06914651e-02 -2.51344651e-01
-2.51557648e-01 7.44246602e-01 4.55942571e-01 3.01489029e-02
1.27729952e+00 -4.58335102e-01 -4.94360067e-02 1.18863158e-01
5.72947562e-01 3.77874225e-01 -1.30165851e+00 -5.56456335e-02
4.65433300e-01 -5.31848848e-01 -2.33396128e-01 -2.28933424e-01
-1.34650159e+00 6.58192337e-01 -7.90808275e-02 9.19923782e-01
1.15767741e+00 -3.81033808e-01 3.49045634e-01 5.13247311e-01
7.24220335e-01 -7.77722836e-01 -2.67240673e-01 -3.23843330e-01
6.88076138e-01 -6.28551185e-01 3.10843766e-01 -7.22417235e-01
-4.85860944e-01 1.22161448e+00 -3.43263336e-02 -3.16356331e-01
9.13677514e-01 -4.31016386e-02 -5.06804109e-01 -3.30992609e-01
-3.23265344e-01 -1.03150167e-01 4.48944092e-01 6.15957201e-01
-2.74379477e-02 4.08548683e-01 -3.81177098e-01 1.02056372e+00
-1.43991515e-01 -4.36646551e-01 6.17150187e-01 7.27965355e-01
-2.82151073e-01 -1.12357605e+00 -2.01090872e-01 4.92887557e-01
-6.06179059e-01 -8.95036384e-03 -3.07225645e-01 1.32132697e+00
-4.22718763e-01 7.60145366e-01 -9.12527889e-02 8.89176652e-02
4.85901922e-01 -2.93260604e-01 4.84768629e-01 -4.99514669e-01
-1.05840191e-01 2.03891203e-01 3.27954203e-01 -1.26900852e-01
-3.36168855e-01 -9.49999630e-01 -8.37363482e-01 -4.28059518e-01
-4.27946448e-01 3.96330118e-01 1.50170252e-01 8.43405366e-01
8.44051838e-02 4.94970739e-01 1.05669820e+00 -5.35219431e-01
-9.81233239e-01 -3.13098401e-01 -1.34941149e+00 -8.38418901e-02
1.73292365e-02 -9.81704533e-01 -5.89761436e-01 -3.35977077e-01] | [5.087348461151123, 4.058304309844971] |
20bac21c-7b08-4c0c-9076-a9ba2361a334 | xsleepnet-multi-view-sequential-model-for | 2007.05492 | null | https://arxiv.org/abs/2007.05492v4 | https://arxiv.org/pdf/2007.05492v4.pdf | XSleepNet: Multi-View Sequential Model for Automatic Sleep Staging | Automating sleep staging is vital to scale up sleep assessment and diagnosis to serve millions experiencing sleep deprivation and disorders and enable longitudinal sleep monitoring in home environments. Learning from raw polysomnography signals and their derived time-frequency image representations has been prevalent. However, learning from multi-view inputs (e.g., both the raw signals and the time-frequency images) for sleep staging is difficult and not well understood. This work proposes a sequence-to-sequence sleep staging model, XSleepNet, that is capable of learning a joint representation from both raw signals and time-frequency images. Since different views may generalize or overfit at different rates, the proposed network is trained such that the learning pace on each view is adapted based on their generalization/overfitting behavior. In simple terms, the learning on a particular view is speeded up when it is generalizing well and slowed down when it is overfitting. View-specific generalization/overfitting measures are computed on-the-fly during the training course and used to derive weights to blend the gradients from different views. As a result, the network is able to retain the representation power of different views in the joint features which represent the underlying distribution better than those learned by each individual view alone. Furthermore, the XSleepNet architecture is principally designed to gain robustness to the amount of training data and to increase the complementarity between the input views. Experimental results on five databases of different sizes show that XSleepNet consistently outperforms the single-view baselines and the multi-view baseline with a simple fusion strategy. Finally, XSleepNet also outperforms prior sleep staging methods and improves previous state-of-the-art results on the experimental databases. | ['Alfred Mertins', 'Minh C. Tran', 'Maarten De Vos', 'Oliver Y. Chén', 'Philipp Koch', 'Huy Phan'] | 2020-07-08 | null | null | null | null | ['sleep-stage-detection', 'sleep-staging'] | ['medical', 'medical'] | [ 4.38166223e-02 -3.07773769e-01 -2.91100919e-01 -7.17635751e-01
-4.60352749e-01 -2.17541918e-01 1.45100832e-01 -3.27145398e-01
-3.74866366e-01 5.62071800e-01 4.32125479e-01 2.70684540e-01
-2.77993698e-02 -3.50957692e-01 -1.20594099e-01 -8.70352685e-01
2.11389195e-02 2.47678161e-01 1.11092098e-01 -3.29222679e-01
-3.03910524e-01 2.20964905e-02 -1.60510886e+00 2.03142807e-01
8.34992409e-01 1.16436386e+00 2.80736029e-01 6.39075041e-01
2.55926400e-01 4.34161007e-01 -5.33117354e-01 -1.13801569e-01
2.39593506e-01 -6.50740802e-01 -3.77055764e-01 1.52383327e-01
5.99767029e-01 -1.22550264e-01 -3.43940198e-01 9.10801411e-01
6.56757295e-01 1.84031710e-01 1.38154358e-01 -1.00733411e+00
-5.27381063e-01 6.71013370e-02 -3.96879524e-01 9.53410566e-01
2.00636283e-01 1.84860140e-01 8.51419687e-01 -8.44250917e-01
1.95312545e-01 7.30804503e-01 7.23936558e-01 1.02725863e+00
-1.34791458e+00 -7.04753935e-01 3.82319421e-01 3.36546063e-01
-1.02769923e+00 -5.96448183e-01 6.62671506e-01 -4.87903468e-02
8.96954536e-01 3.16861987e-01 1.19849777e+00 1.34054363e+00
4.43497717e-01 7.49416888e-01 1.20371687e+00 1.49845317e-01
1.46154523e-01 7.55364224e-02 1.38752371e-01 7.96521962e-01
2.01885104e-01 -1.46501428e-02 -9.74804282e-01 2.24963561e-01
3.53037119e-01 6.57161057e-01 -5.13519347e-01 -3.77313346e-01
-1.11135149e+00 5.42582691e-01 7.00897634e-01 3.97057801e-01
-2.90720254e-01 -1.45507619e-01 5.21399081e-01 3.82646859e-01
6.62772655e-01 3.33922654e-01 -4.93081838e-01 -3.80362391e-01
-1.44237387e+00 -2.39619315e-01 5.65021098e-01 5.64935505e-01
6.74195111e-01 1.74247295e-01 -1.80220529e-01 1.06086111e+00
6.55457079e-02 5.22007227e-01 1.13863587e+00 -8.02327812e-01
4.74450797e-01 8.50091696e-01 -3.10915947e-01 -5.37848771e-01
-9.38081861e-01 -7.19892621e-01 -1.02420962e+00 7.10476050e-03
5.09573072e-02 -4.66013625e-02 -1.10899317e+00 1.87946594e+00
-3.95121872e-02 3.58008623e-01 -4.49685492e-02 7.63644934e-01
8.66186023e-01 3.07318002e-01 -1.65816933e-01 -3.52099180e-01
1.44848955e+00 -1.13163137e+00 -6.67911768e-01 -8.13936889e-01
8.73308629e-02 -1.56766504e-01 1.09490418e+00 4.18116540e-01
-1.16210282e+00 -8.75216186e-01 -1.35583019e+00 -3.63290235e-02
-2.17817098e-01 2.72076756e-01 2.97610968e-01 5.67967474e-01
-1.31607270e+00 8.04378986e-01 -1.50982141e+00 -6.05802357e-01
5.81118703e-01 6.74227893e-01 -3.59336317e-01 -1.25220520e-02
-8.14579785e-01 7.23159850e-01 3.97183634e-02 1.51607975e-01
-7.15853095e-01 -9.04230237e-01 -7.05357909e-01 2.16376573e-01
-3.37512568e-02 -1.15542328e+00 1.00657606e+00 -9.85886633e-01
-1.22059524e+00 8.25361729e-01 -4.00717825e-01 -5.13471723e-01
3.23312730e-01 -3.07568610e-01 -8.17833126e-01 3.94774675e-01
3.16018373e-01 5.13724208e-01 1.23523772e+00 -8.65462065e-01
-6.98072970e-01 -7.81082392e-01 -1.36378556e-01 3.80092740e-01
-6.07057452e-01 -4.27793622e-01 -5.47573447e-01 -5.46085000e-01
2.19056502e-01 -1.03704810e+00 3.29760201e-02 4.33522202e-02
-3.40031795e-02 1.85551822e-01 6.54551685e-01 -5.00958204e-01
1.34122229e+00 -2.48014998e+00 3.80924135e-01 -2.36570776e-01
7.09397435e-01 1.11176901e-01 1.36276320e-01 1.44323632e-01
-1.19141459e-01 -2.65133679e-01 -1.45005748e-01 -7.96280146e-01
-3.49987507e-01 5.63881695e-01 1.18569694e-01 6.21597052e-01
-1.60388410e-01 7.59031117e-01 -1.06886387e+00 -1.79781035e-01
1.12086840e-01 3.94436657e-01 -5.82480490e-01 2.82995164e-01
5.28373599e-01 5.40997922e-01 -1.10112742e-01 5.32800853e-01
2.72607684e-01 -6.61600173e-01 -1.09979464e-03 -3.37405324e-01
2.34122440e-01 3.27581972e-01 -4.98089790e-01 1.95835471e+00
-8.13819528e-01 5.50754607e-01 -1.10245213e-01 -8.54465246e-01
6.86036825e-01 1.97409347e-01 6.09938145e-01 -7.54799068e-01
-1.52221337e-01 9.03070644e-02 -6.15554936e-02 -5.47095776e-01
1.97913423e-01 -6.56569004e-01 6.01548627e-02 4.36291516e-01
5.05635917e-01 2.86649138e-01 2.52353072e-01 -3.14201266e-02
1.09053671e+00 -1.75303563e-01 4.60552722e-01 -6.58499524e-02
5.47893286e-01 -6.01454496e-01 8.45126569e-01 4.91702080e-01
-4.22366589e-01 7.43962288e-01 2.15776384e-01 -6.14337027e-01
-6.62105799e-01 -1.33710778e+00 -8.57725367e-02 1.14789271e+00
1.88406736e-01 -7.15617478e-01 -3.63492489e-01 -1.08006489e+00
-1.77890554e-01 4.81304854e-01 -8.73741567e-01 -7.81554103e-01
-3.79396915e-01 -9.17834997e-01 3.63993585e-01 8.53805661e-01
5.51962376e-01 -8.70793045e-01 -8.89350474e-01 1.33917332e-01
-1.38874263e-01 -1.00946677e+00 -8.04565251e-01 5.33252716e-01
-1.30104852e+00 -9.40441489e-01 -6.92041337e-01 -4.11238223e-01
7.54269123e-01 6.12275124e-01 1.18861616e+00 1.01214163e-01
-2.05778584e-01 5.23449481e-01 -2.57719398e-01 -1.73440680e-01
-3.72562967e-02 1.66765437e-01 5.41294277e-01 3.33973110e-01
5.13272226e-01 -1.25352144e+00 -1.19806325e+00 4.06862557e-01
-7.28237987e-01 -1.45174982e-02 6.19662702e-01 1.04988718e+00
5.22336841e-01 -1.71469048e-01 5.33208489e-01 -7.04469204e-01
4.96601194e-01 -5.05220294e-01 -5.44847809e-02 4.78641167e-02
-1.07070744e+00 5.83961755e-02 8.15939426e-01 -3.69863719e-01
-8.47974539e-01 -1.93426475e-01 -1.08486757e-01 -7.51693308e-01
6.82012970e-03 2.06561372e-01 8.25931281e-02 1.26273572e-01
6.18119061e-01 6.47751331e-01 1.38739601e-01 -4.78286475e-01
1.22019127e-01 3.73254240e-01 5.84874690e-01 -3.60549204e-02
6.00582182e-01 7.10830450e-01 -2.18313083e-01 -7.69490361e-01
-1.30356443e+00 -8.15550447e-01 -6.96944594e-01 -1.63503975e-01
9.97183144e-01 -1.19341457e+00 -4.00631964e-01 3.49775940e-01
-2.42815197e-01 -1.82778746e-01 -5.84402919e-01 4.23133343e-01
-5.00239432e-01 1.48546055e-01 -2.46631145e-01 -3.77338201e-01
-5.29551148e-01 -1.02992964e+00 1.15862083e+00 5.90418518e-01
-3.42012972e-01 -1.21221566e+00 3.45588475e-01 3.08774441e-01
4.20813143e-01 -1.37021229e-01 7.68074870e-01 -6.16824031e-01
-1.89032182e-02 -7.32704327e-02 6.44377172e-02 6.70832813e-01
5.62264323e-01 -6.04559541e-01 -1.20627546e+00 -7.10613132e-01
6.44347370e-01 -2.24508405e-01 1.09910119e+00 3.95893693e-01
9.62938309e-01 -3.77632082e-01 -2.52001584e-01 1.08934236e+00
1.29964042e+00 -5.96806072e-02 4.54356521e-01 3.72214943e-01
5.85424662e-01 1.20910667e-01 -6.87144846e-02 2.93173760e-01
6.18256032e-01 6.30397737e-01 2.16917843e-01 -1.28706589e-01
-3.16127181e-01 -1.01946563e-01 6.92437768e-01 1.18874538e+00
-1.70236707e-01 1.78672850e-01 -4.34645772e-01 3.62322539e-01
-1.68106902e+00 -1.12592435e+00 4.35786605e-01 2.16773057e+00
5.44613957e-01 2.41804466e-01 4.12474692e-01 -3.37156979e-03
1.27025068e-01 5.05782485e-01 -9.85354543e-01 -1.55406520e-01
8.02111477e-02 3.19343060e-01 2.49908134e-01 -9.33587551e-02
-8.53218973e-01 2.99131989e-01 6.12486649e+00 3.40433449e-01
-1.39569283e+00 4.24879342e-01 4.83432263e-01 -8.78370106e-01
1.08061872e-01 -4.42765772e-01 -7.89389133e-01 7.80509472e-01
1.18941736e+00 -4.56697308e-02 8.53564322e-01 7.96079099e-01
2.46289521e-01 -1.64769422e-02 -1.19632125e+00 1.35409236e+00
3.97857040e-01 -1.07459247e+00 -3.34966660e-01 4.16870415e-02
5.68836927e-01 4.66692477e-01 1.58177406e-01 5.92884362e-01
-3.42135251e-01 -7.73673892e-01 5.71183503e-01 7.32701659e-01
8.05833459e-01 -4.01729316e-01 6.90931082e-01 3.06518137e-01
-1.37893462e+00 -4.73257333e-01 -1.70305625e-01 1.50298834e-01
7.89794475e-02 2.55895078e-01 -6.07066214e-01 5.93030214e-01
1.02153826e+00 1.33444953e+00 -1.16173851e+00 9.36141908e-01
-2.21236914e-01 6.75266325e-01 -1.95467681e-01 2.54159063e-01
-2.89447848e-02 -2.69860685e-01 3.41459870e-01 9.39954877e-01
4.03763294e-01 -1.13928564e-01 1.71973675e-01 7.10845172e-01
1.37435738e-03 -4.49982971e-01 -4.24648046e-01 2.55979836e-01
3.20981555e-02 1.43509734e+00 -6.95159853e-01 -2.80609339e-01
-8.20168436e-01 1.17271829e+00 3.60135347e-01 3.59604508e-01
-6.16195917e-01 3.28222886e-02 8.16666424e-01 2.13339642e-01
5.87968707e-01 -2.12797336e-02 -3.17492008e-01 -1.60648739e+00
-1.84968188e-02 -8.29921246e-01 6.11061037e-01 -6.44091070e-01
-1.53191698e+00 9.21730340e-01 -4.08206508e-02 -1.62178731e+00
-3.74795087e-02 -2.58528262e-01 -8.05105925e-01 6.56195104e-01
-1.45388448e+00 -9.47363257e-01 -5.87073743e-01 7.46350050e-01
9.46575403e-01 -2.50863761e-01 6.83743894e-01 3.66400957e-01
-7.80470073e-01 6.59254014e-01 -1.02029122e-01 -2.84278095e-01
7.36794233e-01 -1.53130877e+00 2.39073306e-01 7.25816250e-01
2.88013339e-01 8.29557300e-01 3.84194493e-01 -3.43982577e-01
-1.19375408e+00 -9.89765823e-01 5.51277637e-01 -5.81383944e-01
4.96094108e-01 -3.84025186e-01 -8.85092795e-01 4.44420099e-01
6.44590035e-02 3.70417178e-01 1.06359613e+00 3.04924905e-01
-1.12426981e-01 -8.01096976e-01 -9.30582702e-01 3.57066244e-01
1.03388321e+00 -7.01030910e-01 -6.98186815e-01 -1.06522152e-02
1.27575666e-01 -4.33407396e-01 -8.15414846e-01 2.75264502e-01
6.67751849e-01 -1.42170286e+00 1.02774036e+00 -3.60380471e-01
2.24223942e-01 -1.74945205e-01 -5.73262163e-02 -1.46712518e+00
-4.37479407e-01 -5.37489653e-01 -5.32171786e-01 9.07382071e-01
2.58676231e-01 -6.14990473e-01 8.12200904e-01 3.42058718e-01
-3.03085208e-01 -1.13249624e+00 -1.16162324e+00 -7.10388720e-01
-5.12854636e-01 -1.67327940e-01 4.63257283e-01 3.90156806e-01
-3.89758460e-02 8.39037478e-01 -4.44576383e-01 1.60320356e-01
4.09272641e-01 1.58853874e-01 4.42517936e-01 -1.31869245e+00
-3.09754729e-01 -3.24732542e-01 -5.16403198e-01 -9.15803492e-01
-1.33734867e-01 -1.10578883e+00 -1.07650667e-01 -1.42027497e+00
3.38463545e-01 -9.79101211e-02 -8.67748260e-01 4.98160869e-01
-5.52487493e-01 4.76799637e-01 2.63717681e-01 4.23364908e-01
-6.74918473e-01 7.23432779e-01 1.26313460e+00 -3.99986319e-02
-4.42878515e-01 5.09149671e-01 -8.87525022e-01 8.63196373e-01
6.67125702e-01 -4.24250394e-01 -7.97834516e-01 -2.34272592e-02
1.57561451e-01 3.39099998e-03 2.78283119e-01 -1.39942074e+00
2.17807427e-01 4.40260619e-01 8.18317890e-01 -3.63995522e-01
6.97492182e-01 -7.98984289e-01 2.03502961e-02 4.03847784e-01
-2.39291433e-02 3.38785261e-01 7.25580677e-02 8.36690843e-01
-3.31258923e-02 8.03682506e-02 8.76530826e-01 -3.48442048e-01
-6.67778492e-01 4.53511715e-01 -1.30887210e-01 3.18395257e-01
5.44268608e-01 -5.55741549e-01 -1.84255481e-01 -3.83324534e-01
-1.02605677e+00 3.14616889e-01 4.63386655e-01 5.64752519e-01
7.27893293e-01 -1.32249081e+00 -7.83719569e-02 7.70684063e-01
3.07319075e-01 -2.83418566e-01 4.35026586e-01 1.36155653e+00
8.61255601e-02 8.90417844e-02 -3.81839722e-01 -8.49489748e-01
-1.11507654e+00 4.74393189e-01 5.42597353e-01 -5.02483010e-01
-9.38650668e-01 7.61728346e-01 3.83049130e-01 -1.36234224e-01
1.49207860e-01 -6.43858135e-01 -3.37925404e-01 3.32107008e-01
7.07151890e-01 3.86346459e-01 3.66233647e-01 -5.12693822e-01
-4.09410864e-01 5.97068667e-01 -2.37668693e-01 1.83787555e-01
1.69492340e+00 -4.42048907e-01 1.29664615e-01 8.61356616e-01
1.17563415e+00 -1.38779193e-01 -1.54793394e+00 -7.97499195e-02
-2.18089476e-01 -3.00585449e-01 5.07384315e-02 -7.36224413e-01
-1.41927576e+00 7.49456763e-01 1.06038642e+00 2.29095474e-01
1.55421615e+00 -3.04303896e-02 1.04514074e+00 1.52676180e-01
1.84896618e-01 -8.83983314e-01 3.55495393e-01 1.28494680e-01
6.59211278e-01 -1.18591511e+00 2.31646776e-01 3.09771478e-01
-8.70819032e-01 1.27463222e+00 6.72482371e-01 -1.37449011e-01
6.62341714e-01 -4.99371998e-02 1.77287534e-01 -5.26772082e-01
-7.67298102e-01 -1.50928542e-01 5.53665221e-01 6.44143045e-01
4.55171615e-02 -3.36219631e-02 2.18057752e-01 8.27618837e-01
-3.83540601e-01 8.71142298e-02 2.69040674e-01 6.95757687e-01
-2.53065526e-01 -9.62404728e-01 -3.78362578e-03 9.58781123e-01
-4.39088643e-01 -3.19216326e-02 6.94785491e-02 5.36334276e-01
3.08854431e-01 7.85605490e-01 -5.02786003e-02 -6.48166955e-01
4.91414368e-01 2.40369424e-01 4.46076095e-01 -9.28217649e-01
-6.47992671e-01 2.61399709e-02 -2.02443734e-01 -7.79709518e-01
-7.24203825e-01 -6.85957134e-01 -1.01669943e+00 1.19196504e-01
-5.08185737e-02 -2.94073690e-02 1.57584324e-01 9.56153512e-01
4.65928227e-01 9.21183109e-01 7.46303439e-01 -1.09807634e+00
-5.20217001e-01 -8.40215147e-01 -7.20274270e-01 3.44286114e-01
9.62307811e-01 -7.56183565e-01 -4.03993994e-01 3.37156616e-02] | [13.484896659851074, 3.515169143676758] |
f0960571-44e6-48ac-a997-6bc28ed54c09 | joint-aec-and-beamforming-with-double-talk | 2111.04904 | null | https://arxiv.org/abs/2111.04904v2 | https://arxiv.org/pdf/2111.04904v2.pdf | Joint Neural AEC and Beamforming with Double-Talk Detection | Acoustic echo cancellation (AEC) in full-duplex communication systems eliminates acoustic feedback. However, nonlinear distortions induced by audio devices, background noise, reverberation, and double-talk reduce the efficiency of conventional AEC systems. Several hybrid AEC models were proposed to address this, which use deep learning models to suppress residual echo from standard adaptive filtering. This paper proposes deep learning-based joint AEC and beamforming model (JAECBF) building on our previous self-attentive recurrent neural network (RNN) beamformer. The proposed network consists of two modules: (i) multi-channel neural-AEC, and (ii) joint AEC-RNN beamformer with a double-talk detection (DTD) that computes time-frequency (T-F) beamforming weights. We train the proposed model in an end-to-end approach to eliminate background noise and echoes from far-end audio devices, which include nonlinear distortions. From experimental evaluations, we find the proposed network outperforms other multi-channel AEC and denoising systems in terms of speech recognition rate and overall speech quality. | ['Dong Yu', 'Shi-Xiong Zhang', 'Meng Yu', 'Yong Xu', 'Vinay Kothapally'] | 2021-11-09 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [-1.46098733e-02 -4.69848961e-01 7.51413167e-01 -2.21837431e-01
-1.01486492e+00 -3.53117615e-01 2.19007641e-01 -5.69464684e-01
-4.94276315e-01 3.61903429e-01 1.09787452e+00 -5.01901031e-01
-1.23111248e-01 -1.84633926e-01 -4.26164895e-01 -6.99301481e-01
-2.34572724e-01 -4.57851827e-01 -3.70251499e-02 -5.62966943e-01
-1.58450931e-01 4.24843818e-01 -1.48911548e+00 4.67144430e-01
8.27428818e-01 9.96925712e-01 3.73020649e-01 1.32358360e+00
1.17042914e-01 5.90355754e-01 -1.07785535e+00 -2.02910364e-01
1.40784338e-01 -5.28747261e-01 2.25064114e-01 -5.11066020e-01
3.36783051e-01 -5.79257011e-01 -7.76979327e-01 9.04260099e-01
1.54180014e+00 3.44370365e-01 5.02927601e-01 -7.32194543e-01
-3.26058835e-01 6.98400736e-01 -2.29177624e-01 2.66553283e-01
2.36719936e-01 1.19867958e-01 4.25241262e-01 -1.45144725e+00
-1.81394622e-01 1.45254099e+00 1.29514372e+00 6.56813741e-01
-7.22238541e-01 -1.07533503e+00 -5.94885349e-02 2.37725213e-01
-1.18514264e+00 -1.11817324e+00 7.79772162e-01 6.93758726e-02
1.38647330e+00 4.01080221e-01 4.75267678e-01 1.15741432e+00
1.74217314e-01 6.98023021e-01 9.08625782e-01 -6.60673976e-01
1.83511376e-01 -2.59862512e-01 8.60495344e-02 2.59743333e-01
-3.90281171e-01 7.68346906e-01 -8.78247201e-01 -3.25859636e-02
4.72829670e-01 -3.89796138e-01 -8.26819539e-01 6.70553744e-01
-5.57559967e-01 2.78420001e-01 2.14295313e-01 3.70065600e-01
-4.26791936e-01 4.30576921e-01 3.12041968e-01 7.36223698e-01
3.81325454e-01 1.68522224e-02 -3.65364373e-01 -1.10847712e-01
-9.89080310e-01 -5.50401174e-02 9.94108200e-01 7.82674134e-01
-7.00791851e-02 1.20806825e+00 -2.24582866e-01 1.40562689e+00
6.67593122e-01 1.22412324e+00 6.97502851e-01 -8.31892252e-01
4.73661095e-01 -7.70585895e-01 8.03000033e-02 -9.42566991e-01
-5.96704006e-01 -1.24479330e+00 -1.35883427e+00 1.72692940e-01
-4.62737828e-01 -8.87779057e-01 -9.84729528e-01 1.67697406e+00
2.06184145e-02 4.68865186e-01 2.10952282e-01 1.21383047e+00
9.33141112e-01 1.07276022e+00 -1.13832958e-01 -4.77979660e-01
6.71666503e-01 -1.00553834e+00 -1.61820340e+00 -2.23619133e-01
1.40347138e-01 -1.26060557e+00 3.79608870e-01 6.64722621e-01
-1.26678348e+00 -8.81940842e-01 -1.32721698e+00 8.04950148e-02
-1.92123383e-01 -1.82998832e-02 1.63863767e-02 1.41072464e+00
-1.46589553e+00 2.63049573e-01 -3.86585832e-01 3.84601265e-01
-3.31946611e-01 4.20633137e-01 1.11103788e-01 1.96912289e-01
-1.53067708e+00 7.78799534e-01 -3.14594984e-01 8.17259729e-01
-1.14573431e+00 -8.68397057e-01 -6.81685269e-01 3.36231768e-01
-2.75796056e-02 -4.50671524e-01 1.51075077e+00 -9.87159133e-01
-2.02542949e+00 -2.27285221e-01 -2.98892438e-01 -5.46141744e-01
8.48641172e-02 -8.66144001e-01 -1.45159042e+00 -2.82399058e-01
-5.85335493e-01 5.46849407e-02 1.34691834e+00 -1.33869958e+00
-8.00561666e-01 -1.52509725e-02 -6.96679890e-01 3.69240850e-01
-3.36042941e-01 -2.07436364e-02 -2.80653656e-01 -1.17507517e+00
4.03515212e-02 -4.37255800e-01 -1.30192593e-01 -3.25360537e-01
-1.55004025e-01 3.01474869e-01 8.68175745e-01 -1.21393311e+00
1.79729950e+00 -2.29899931e+00 -1.79430515e-01 3.98173422e-01
-1.04982644e-01 1.01596999e+00 -5.84683776e-01 4.98745412e-01
-4.26442236e-01 -1.99109957e-01 2.07584742e-02 -6.33022368e-01
-1.58138748e-03 -2.82381147e-01 -3.78664345e-01 4.77523774e-01
-1.26611710e-01 2.07166821e-01 -3.21426839e-01 1.62411913e-01
3.83682072e-01 1.01242089e+00 -5.76621652e-01 3.90787721e-01
5.64118505e-01 -5.62890321e-02 2.43985295e-01 4.24395591e-01
1.13858652e+00 8.57522666e-01 -1.82976350e-01 -4.79639798e-01
-3.67602736e-01 2.54047483e-01 -1.52844799e+00 1.52329493e+00
-9.66461718e-01 8.35696101e-01 1.13704240e+00 -7.03135014e-01
8.60453308e-01 8.54902983e-01 1.88969951e-02 -1.16711259e+00
7.52212703e-02 5.98458052e-01 3.30117404e-01 -3.92306924e-01
3.03241044e-01 -3.16432357e-01 3.20774943e-01 1.80497497e-01
4.36274588e-01 1.05951861e-01 -4.65247184e-01 1.75651873e-03
1.13029695e+00 -4.50203121e-01 -2.59781808e-01 -3.57859015e-01
9.79389131e-01 -9.68395889e-01 7.30464458e-01 8.82574677e-01
-1.88752308e-01 6.24986291e-01 -5.23708582e-01 -1.10728259e-03
-6.37730777e-01 -1.04813468e+00 2.91426122e-01 1.31335342e+00
-1.57309175e-01 -2.50483036e-01 -7.00195312e-01 1.90153599e-01
-1.33979753e-01 8.35210919e-01 3.24696675e-02 -3.07810187e-01
-8.44407022e-01 -2.89429277e-01 8.89284492e-01 4.20236707e-01
5.04001856e-01 -6.63134873e-01 1.06367253e-01 7.52717137e-01
-4.54780191e-01 -7.40978062e-01 -1.07446229e+00 4.39691991e-01
-4.79286224e-01 -1.72394067e-01 -1.01670003e+00 -9.10025239e-01
4.90393490e-02 5.15838921e-01 6.70291722e-01 -2.04632938e-01
-8.27833712e-02 6.70515299e-01 -1.90096334e-01 -7.39835203e-01
-3.01305413e-01 -4.35379773e-01 5.61660707e-01 2.38554299e-01
-4.74370494e-02 -7.74868190e-01 -9.79787350e-01 4.18805003e-01
-6.28651321e-01 -4.61291820e-01 7.31799424e-01 7.33900011e-01
5.96275330e-02 2.65964240e-01 8.49938273e-01 -3.20385188e-01
1.34093368e+00 -1.76490694e-01 -3.59430850e-01 9.14477035e-02
-3.91305029e-01 -2.83956647e-01 6.80692554e-01 -5.20615935e-01
-1.83791173e+00 -1.92967162e-01 -1.18015528e+00 -3.49198639e-01
1.70794293e-01 5.30498028e-01 -2.79582262e-01 -8.94797072e-02
6.35315776e-01 4.15624231e-01 -1.90497622e-01 -6.96872830e-01
2.90146351e-01 1.34070134e+00 1.00443554e+00 -5.01035973e-02
8.46649766e-01 -6.34011552e-02 -3.95600855e-01 -1.21053469e+00
-2.74797350e-01 -6.42371058e-01 1.91067681e-02 -5.76761365e-01
4.04915094e-01 -1.34983194e+00 -7.51621306e-01 8.38416755e-01
-1.33974314e+00 -2.38745540e-01 2.89736062e-01 1.10269880e+00
-8.65322053e-02 1.57640621e-01 -7.89761126e-01 -1.40605414e+00
-9.05030668e-01 -7.88303196e-01 6.77972198e-01 2.59189934e-01
3.21489875e-03 -6.08444750e-01 2.23070651e-01 -6.21183850e-02
1.41387284e+00 -5.67704320e-01 5.37596822e-01 -2.01147556e-01
-1.91847429e-01 -2.46251106e-01 4.30210769e-01 8.10028195e-01
-5.06381877e-02 -5.27007103e-01 -1.32388639e+00 -3.80651861e-01
4.97627825e-01 1.25689670e-01 8.50713491e-01 8.85716438e-01
7.23160803e-01 -2.40704000e-01 3.38793807e-02 7.46232927e-01
1.34781337e+00 6.89490080e-01 8.95506620e-01 -4.03738528e-01
1.18288465e-01 1.84310466e-01 -3.67145836e-02 3.63756835e-01
-2.89007962e-01 1.91866562e-01 9.51552689e-02 -4.51547235e-01
-7.59493113e-01 4.06075567e-02 6.81179345e-01 1.67974830e+00
6.11089058e-02 -9.12033439e-01 -4.43717033e-01 3.32664043e-01
-1.33725178e+00 -9.39360023e-01 -3.20387572e-01 2.01356602e+00
6.58838093e-01 7.52410665e-02 -3.45333993e-01 3.15747857e-01
7.06704855e-01 7.67768249e-02 -4.76326078e-01 -7.19760060e-01
-4.24736679e-01 6.23687863e-01 7.21115649e-01 1.07768714e+00
-8.76805305e-01 3.21903557e-01 6.23948145e+00 1.03690815e+00
-1.11892211e+00 4.65039462e-01 3.29669654e-01 -2.51240849e-01
-3.28341514e-01 -7.83829331e-01 -5.83154559e-01 1.20736606e-01
1.42766953e+00 3.14915061e-01 6.71866715e-01 3.36960644e-01
5.59246898e-01 3.55228335e-01 -6.93451941e-01 1.24238050e+00
2.39547521e-01 -8.78320038e-01 -3.42702717e-01 -3.46683025e-01
3.13271612e-01 7.54890591e-02 2.73166806e-01 7.86275506e-01
1.08913623e-01 -9.06129837e-01 7.92965770e-01 7.86976457e-01
8.41951609e-01 -6.93502665e-01 7.54821181e-01 2.52409995e-01
-1.25614786e+00 -4.65368420e-01 -1.04333505e-01 9.04684961e-02
3.54831129e-01 7.81124294e-01 -2.08826095e-01 3.28125536e-01
7.42834866e-01 6.62291329e-03 2.69542366e-01 1.28646517e+00
-5.57427481e-02 1.09136403e+00 -4.58065063e-01 -9.48890373e-02
8.14079195e-02 1.56128272e-01 7.65017211e-01 1.84380257e+00
6.22219443e-01 4.77842659e-01 -5.18552125e-01 2.82764047e-01
-2.58609205e-01 -2.82381147e-01 -6.36458695e-02 3.97219449e-01
6.85127020e-01 9.82043028e-01 7.05854222e-02 1.00133762e-01
-5.73694825e-01 9.20710564e-01 -5.59926510e-01 9.32826340e-01
-7.51056731e-01 -9.39833939e-01 6.52093232e-01 -2.10783541e-01
2.83582628e-01 -2.55955517e-01 -3.66333639e-03 -5.06426513e-01
-3.86917330e-02 -9.99389708e-01 -1.30739078e-01 -8.55039835e-01
-1.11733651e+00 7.25836337e-01 -7.31202960e-01 -1.22478235e+00
-4.56697978e-02 -5.85763097e-01 -8.00353229e-01 1.41634548e+00
-1.52634859e+00 -5.80110073e-01 5.54642715e-02 8.60873342e-01
8.52527916e-01 -3.86607230e-01 8.71531308e-01 1.08372736e+00
-3.35102975e-01 9.86499727e-01 6.17119551e-01 -2.44890247e-02
8.77541959e-01 -1.06509209e+00 2.80467778e-01 1.16351295e+00
-3.67186695e-01 8.25502276e-01 8.80432129e-01 -5.05194247e-01
-1.59849310e+00 -1.09964967e+00 8.17327321e-01 4.28436220e-01
2.65611768e-01 -6.86414182e-01 -7.22327232e-01 3.95669974e-02
8.32675755e-01 -2.36022532e-01 8.71105850e-01 8.26808736e-02
-1.12835489e-01 -8.35216880e-01 -6.80851161e-01 5.61115265e-01
9.90127981e-01 -7.27933645e-01 -3.16097319e-01 7.46088848e-02
7.13233531e-01 -3.15048963e-01 -5.49363732e-01 3.86335999e-01
8.71807814e-01 -8.85782659e-01 1.12178791e+00 4.15979847e-02
-1.06211733e-02 -3.34403425e-01 -4.33301330e-01 -1.83682418e+00
-5.46579123e-01 -1.32869864e+00 -2.51173526e-01 1.39394450e+00
6.70022786e-01 -5.89745462e-01 2.49721184e-01 2.30309725e-01
-9.71944213e-01 -1.68233395e-01 -9.42846835e-01 -6.83207035e-01
-1.91150278e-01 -6.93936467e-01 2.61108100e-01 1.98557228e-01
-2.99747378e-01 5.75222194e-01 -9.84160423e-01 6.67726457e-01
5.43828547e-01 -6.78593755e-01 2.06768736e-01 -7.80285597e-01
-4.78283077e-01 -4.19967830e-01 3.04235399e-01 -1.52140760e+00
-3.38461012e-01 -2.41335571e-01 6.96227610e-01 -1.43410182e+00
-6.89113081e-01 4.80594337e-02 -5.26470900e-01 -2.28290752e-01
-1.54984565e-02 -1.62501842e-01 6.93239942e-02 -2.98081875e-01
-1.61264822e-01 8.52343202e-01 7.98366845e-01 -2.62931287e-01
-2.43889794e-01 3.78694177e-01 -5.00112295e-01 6.39485300e-01
4.86771107e-01 -3.95538121e-01 -4.25808847e-01 -8.26910257e-01
-2.01460823e-01 7.35537052e-01 -1.21677645e-01 -1.27461565e+00
1.02250314e+00 6.57134414e-01 5.16917706e-01 -8.04946244e-01
7.27319360e-01 -1.08552301e+00 1.31695420e-01 6.79265082e-01
-4.84466761e-01 -2.25796357e-01 3.39481622e-01 7.90141821e-01
-4.62767184e-01 4.09764312e-02 6.95778787e-01 1.62488058e-01
-3.07935894e-01 -1.40524492e-01 -1.21447456e+00 -4.97961879e-01
-1.46466764e-02 3.43067460e-02 -2.19419435e-01 -1.16371775e+00
-7.36054778e-01 1.59549370e-01 -1.07103145e+00 4.35777873e-01
1.03108478e+00 -1.31170738e+00 -1.00652361e+00 3.23822290e-01
-6.25263214e-01 -6.64369285e-01 7.45626926e-01 6.10405385e-01
-1.77078143e-01 6.99508607e-01 1.52882800e-01 -2.30102807e-01
-1.28269494e+00 2.12154016e-02 9.42387164e-01 2.18873769e-01
-2.55202621e-01 1.35652184e+00 -1.79929323e-02 -3.87572497e-01
1.11800086e+00 -1.64509729e-01 -1.07755825e-01 -3.70059550e-01
7.46405244e-01 5.50597966e-01 5.24740517e-01 -4.40311700e-01
-2.14198500e-01 4.23068732e-01 -1.08312396e-02 -6.23632729e-01
1.15980959e+00 -4.54742491e-01 2.06117392e-01 6.95388913e-02
1.33435786e+00 2.19976574e-01 -7.98478067e-01 -3.83571208e-01
-5.05111098e-01 -1.83802426e-01 7.14963257e-01 -1.44763958e+00
-1.19025636e+00 1.04096437e+00 1.55536950e+00 -1.24329276e-01
1.88330388e+00 -9.26036179e-01 1.16684091e+00 4.70273197e-01
-7.29257688e-02 -1.42524588e+00 1.17643416e-01 6.47695899e-01
1.24865067e+00 -6.54616475e-01 -4.70540851e-01 1.91944949e-02
-1.76447466e-01 1.20924485e+00 4.69733596e-01 4.20903228e-02
1.32455778e+00 8.85824680e-01 5.07357538e-01 2.97474504e-01
-6.33003175e-01 -2.34144945e-02 1.80585563e-01 7.98165262e-01
5.29394925e-01 -1.43007681e-01 -3.18250358e-01 1.03252423e+00
-1.66439041e-01 -2.89636910e-01 2.37671137e-01 7.10175216e-01
-7.44862199e-01 -8.03447902e-01 -6.98044240e-01 3.23885769e-01
-5.67177892e-01 -4.26977605e-01 -2.64023602e-01 2.58198768e-01
1.11622311e-01 1.72092438e+00 -6.28173202e-02 -7.54214168e-01
8.24738383e-01 -7.59109482e-02 -8.30031857e-02 1.80843934e-01
-1.07050848e+00 9.64427531e-01 2.17684627e-01 -4.25450504e-01
-2.78515399e-01 -3.34790945e-01 -8.59968185e-01 -1.73435673e-01
-6.97698116e-01 1.65237159e-01 9.92841363e-01 6.78404570e-01
3.30474108e-01 1.24356985e+00 8.37683499e-01 -8.23912203e-01
-5.35325229e-01 -1.40635502e+00 -6.61312580e-01 -8.40581208e-02
1.00156689e+00 -1.23293370e-01 -5.05273342e-01 -7.55390450e-02] | [15.002547264099121, 5.967494487762451] |
66b726cd-8ca3-4608-8419-f9da4cf9300f | machine-reading-comprehension-with-enhanced | null | null | https://openreview.net/forum?id=EVV259WQuFG | https://openreview.net/pdf?id=EVV259WQuFG | Machine Reading Comprehension with Enhanced Linguistic Verifiers | We propose two linguistic verifiers for span-extraction style machine reading comprehension to respectively tackle two challenges: how to evaluate the syntactic completeness of predicted answers and how to utilize the rich context of long documents. Our first verifier rewrites a question through replacing its interrogatives by the predicted answer phrases and then builds a cross-attention scorer between the rewritten question and the segment, so that the answer candidates are scored in a \emph{position-sensitive} context. Our second verifier builds a hierarchical attention network to represent segments in a passage where neighbour segments in long passages are \emph{recurrently connected} and can contribute to current segment-question pair's inference for answerablility classification and boundary determination. We then combine these two verifiers together into a pipeline and apply it to SQuAD2.0, NewsQA and TriviaQA benchmark sets. Our pipeline achieves significantly better improvements of both exact matching and F1 scores than state-of-the-art baselines. | ['Xianchao Wu'] | 2021-01-01 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [ 5.01213133e-01 5.02975106e-01 2.72965450e-02 -5.19525826e-01
-1.79102719e+00 -9.76809978e-01 1.48996219e-01 3.49998921e-01
-5.61761618e-01 6.72042310e-01 6.51117802e-01 -7.33247519e-01
1.57583088e-01 -7.69869626e-01 -8.29493344e-01 3.70433624e-03
5.42592525e-01 7.02214837e-01 8.45962584e-01 -3.93706739e-01
3.99145544e-01 -2.01915905e-01 -1.29980946e+00 1.05587304e+00
1.15389037e+00 9.23655629e-01 1.45262837e-01 1.23754358e+00
-5.09870291e-01 1.17389166e+00 -8.07008207e-01 -9.07093883e-01
-2.33465806e-01 -7.46334255e-01 -1.73477066e+00 -6.57477438e-01
7.95115590e-01 -2.68604457e-01 -1.31763756e-01 6.90602601e-01
5.40899813e-01 1.40167192e-01 4.53545511e-01 -3.36777359e-01
-8.74881685e-01 1.09520185e+00 -3.21082920e-01 8.04633200e-01
9.68412519e-01 6.34017363e-02 1.71522403e+00 -8.72502804e-01
6.52464986e-01 1.23175204e+00 5.23091018e-01 5.60464680e-01
-9.15255070e-01 -2.72931933e-01 2.55388588e-01 7.24368453e-01
-7.84483373e-01 -5.17782688e-01 5.58076739e-01 -1.21872358e-01
1.46832240e+00 7.28862226e-01 8.54862630e-02 8.75709057e-01
-9.85419825e-02 1.14145672e+00 6.47619784e-01 -5.22808611e-01
-1.86405852e-01 -3.45732212e-01 8.36147964e-01 7.59783804e-01
-6.65407121e-01 -4.00320023e-01 -4.22612876e-01 5.31105660e-02
-1.76150560e-01 -7.58000791e-01 -5.87754071e-01 7.05489933e-01
-1.04851079e+00 8.36161911e-01 4.82002497e-01 2.81808853e-01
-1.81448132e-01 -1.00264587e-01 3.81935269e-01 5.86527228e-01
6.00064397e-02 7.62510359e-01 -6.86708927e-01 -1.77976087e-01
-8.29501152e-01 5.01786053e-01 1.06976473e+00 8.50982010e-01
3.38240832e-01 -7.27146983e-01 -9.95094359e-01 8.62486660e-01
1.15140723e-02 1.90491378e-01 2.96356291e-01 -1.14329100e+00
1.11780477e+00 7.88699567e-01 8.72135535e-02 -7.87923157e-01
-3.45012367e-01 -4.35705930e-01 -2.40173131e-01 -5.80128074e-01
7.75624156e-01 -1.02915406e-01 -3.77211034e-01 1.62806165e+00
5.42482674e-01 -1.94031209e-01 9.29799750e-02 7.88017094e-01
1.41457069e+00 9.93378043e-01 -6.25996888e-02 -2.81417798e-02
1.74727261e+00 -1.55941057e+00 -5.78886032e-01 -2.49647856e-01
6.59550786e-01 -9.47057307e-01 1.42967987e+00 6.52303407e-03
-1.61445260e+00 -8.38431656e-01 -8.15688789e-01 -8.71202052e-01
1.04114391e-01 5.06144799e-02 -3.46415430e-01 -1.88223012e-02
-7.91485369e-01 4.64969248e-01 -2.79580683e-01 7.02925771e-02
2.22802997e-01 1.60595134e-01 6.77268207e-02 -4.72563803e-02
-1.51546586e+00 1.03999281e+00 9.54219252e-02 1.08060896e-01
-5.13942778e-01 -8.92289877e-01 -7.91541755e-01 3.20425600e-01
2.51492977e-01 -9.61107433e-01 1.83326888e+00 -7.67412841e-01
-1.48350430e+00 1.29830587e+00 -7.43698120e-01 -5.88663578e-01
4.03757781e-01 -4.49362427e-01 -3.28473866e-01 3.85327220e-01
2.44524762e-01 4.68497932e-01 4.77567255e-01 -6.86845720e-01
-8.28117967e-01 -2.69220769e-01 3.75589639e-01 3.21496546e-01
3.92206609e-01 3.95193279e-01 -5.08662760e-01 -1.78376585e-01
2.79624104e-01 -4.53074336e-01 1.82313174e-01 -5.50394714e-01
-7.32195556e-01 -9.37932312e-01 3.85512233e-01 -1.22264671e+00
1.65394938e+00 -1.67483008e+00 3.59429151e-01 -1.42503291e-01
2.32230499e-01 4.25435752e-01 -4.35844272e-01 4.50990856e-01
2.65860528e-01 2.95707006e-02 -2.18437389e-01 -3.62030476e-01
2.84487247e-01 4.19141501e-02 -6.90270007e-01 -2.19794158e-02
4.43278432e-01 1.50626945e+00 -9.24094379e-01 -5.55586696e-01
-4.77160126e-01 -2.00990140e-02 -6.47481978e-01 6.68654740e-01
-7.65260041e-01 5.06097555e-01 -3.85830104e-01 3.21883261e-01
3.38175416e-01 -2.62992740e-01 -9.46699753e-02 -3.33984308e-02
1.05249248e-02 1.45956314e+00 -5.51053762e-01 1.39956152e+00
-3.91682565e-01 4.62804526e-01 7.88426842e-04 -6.08482718e-01
8.23766530e-01 2.04570100e-01 -5.35026312e-01 -9.27403569e-01
7.87957683e-02 3.58719736e-01 1.03253193e-01 -1.04137170e+00
6.64402544e-01 1.76978976e-01 -2.23565444e-01 5.20190597e-01
-5.21004433e-03 -8.92819241e-02 2.85786510e-01 3.00896972e-01
1.42931259e+00 3.42734993e-01 9.80100259e-02 -1.03731416e-01
1.21145260e+00 -1.66325361e-01 4.14824396e-01 9.26537156e-01
-1.85573369e-01 8.90244722e-01 9.82308805e-01 -2.59904176e-01
-6.89225137e-01 -1.02278626e+00 2.38555238e-01 1.69590878e+00
-1.77986890e-01 -2.94880748e-01 -1.12100697e+00 -1.19863093e+00
-1.15850769e-01 1.04934907e+00 -7.93413579e-01 6.18194118e-02
-1.32112837e+00 2.66982108e-01 8.69821727e-01 6.43581450e-01
4.45442945e-01 -1.37023747e+00 -5.52479863e-01 3.84243727e-01
-8.42506886e-01 -9.24296558e-01 -7.54227459e-01 -1.64583754e-02
-3.41665685e-01 -1.23435163e+00 -5.74539363e-01 -1.21137249e+00
2.40334079e-01 -2.22653449e-01 1.71994734e+00 5.19814730e-01
3.88862371e-01 -1.04198247e-01 -4.81991112e-01 -1.83859214e-01
-4.73003834e-01 6.44128144e-01 -7.36780345e-01 -2.19764516e-01
6.79549992e-01 -2.13030308e-01 -7.84993410e-01 2.30168849e-01
-2.96723038e-01 -4.55501638e-02 1.54859304e-01 7.99644709e-01
8.03561985e-01 -7.83875883e-01 1.00983596e+00 -1.19982624e+00
8.31292152e-01 -4.76933569e-01 -5.41455984e-01 7.61873484e-01
9.71657690e-03 1.56582564e-01 7.41240859e-01 -1.26660809e-01
-1.23168421e+00 -4.03214693e-01 -8.94319117e-01 3.75471413e-01
-4.36901674e-02 5.83548307e-01 -2.71672428e-01 6.37851775e-01
7.29088604e-01 -1.00095784e-02 -3.67148340e-01 -5.86787999e-01
6.43878579e-01 6.77681267e-01 1.03187525e+00 -6.18919134e-01
5.11605561e-01 -2.68988937e-01 -5.77080727e-01 -2.44388744e-01
-1.79727650e+00 -6.27237320e-01 -6.56007826e-01 7.63556883e-02
1.13922524e+00 -3.69758010e-01 -9.87912118e-01 2.44363710e-01
-1.74010229e+00 -2.54392743e-01 -3.58341068e-01 -1.08337186e-01
-4.51564640e-01 2.92873234e-01 -9.64995146e-01 -4.47314590e-01
-7.56838858e-01 -9.93015885e-01 9.92582381e-01 4.25290942e-01
-7.12380350e-01 -7.79122353e-01 2.48261899e-01 1.10372353e+00
3.07589591e-01 -1.76040158e-01 1.37448668e+00 -1.06718457e+00
-5.13079286e-01 -6.20765463e-02 -3.05311948e-01 2.89405733e-01
-4.93654668e-01 -2.08582044e-01 -1.00768626e+00 1.21042974e-01
1.76984578e-01 -6.08727038e-01 1.11066175e+00 1.42629752e-02
1.01358461e+00 -7.38027096e-01 8.57534334e-02 2.93156356e-01
8.16542327e-01 -1.78533286e-01 7.24812746e-01 7.04832450e-02
4.16433364e-01 9.81073558e-01 4.21791106e-01 -3.12159866e-01
8.82029951e-01 4.88106340e-01 2.49071628e-01 2.53315985e-01
-4.00615364e-01 -5.81920743e-01 3.00038904e-01 1.38510716e+00
4.35010493e-01 -4.18890923e-01 -8.33479345e-01 7.83992112e-01
-1.79233491e+00 -1.02568007e+00 -6.84994221e-01 1.78510392e+00
1.35063803e+00 2.36514047e-01 -7.03731030e-02 5.17676584e-02
5.86859524e-01 2.10754052e-01 -4.54904884e-01 -7.32829988e-01
-2.35580638e-01 6.98669434e-01 -2.91400373e-01 9.53956962e-01
-9.31014717e-01 1.10300279e+00 5.52400827e+00 6.30176902e-01
-5.17325163e-01 2.17413992e-01 7.05716014e-01 8.98177102e-02
-7.29322970e-01 1.38151482e-01 -1.07406914e+00 3.07770878e-01
1.09630346e+00 1.46900892e-01 6.45683408e-02 3.39937091e-01
-1.57515466e-01 -1.67516977e-01 -1.31662345e+00 2.17711911e-01
2.98587918e-01 -1.55558777e+00 -1.10972583e-01 -8.21768641e-01
6.55690908e-01 4.06431630e-02 1.75767299e-02 6.52730763e-01
1.53024465e-01 -1.20916963e+00 7.43827343e-01 6.27631366e-01
4.81896162e-01 -5.99642158e-01 1.00167513e+00 4.89422560e-01
-1.04853117e+00 1.54165570e-02 -2.11063609e-01 -1.48546681e-01
5.15593290e-01 5.02151325e-02 -6.90870523e-01 3.78021091e-01
5.26314437e-01 1.68700486e-01 -6.25191987e-01 9.25028861e-01
-1.16910005e+00 1.29882514e+00 -1.53027605e-02 -3.86071146e-01
3.28490227e-01 7.75992051e-02 5.81555188e-01 1.21513724e+00
-1.09637015e-01 4.97816414e-01 -1.93506092e-01 1.02847886e+00
-5.18151879e-01 2.96474785e-01 1.75505146e-01 3.66194844e-01
4.97476250e-01 9.93058324e-01 -1.45009384e-01 -4.69525397e-01
-4.77625281e-01 1.10257888e+00 1.01694643e+00 3.80822957e-01
-6.98368013e-01 -7.67972112e-01 1.54986843e-01 6.84146211e-02
5.58358133e-01 3.11647177e-01 -2.93527573e-01 -1.13296020e+00
3.42480004e-01 -1.22755718e+00 8.66982698e-01 -8.77242446e-01
-1.30580854e+00 7.64318645e-01 -6.37448907e-01 -4.72565681e-01
-2.47164667e-01 -4.22681034e-01 -1.11197329e+00 1.33999789e+00
-1.75064600e+00 -1.20915568e+00 -8.83299671e-03 2.64248073e-01
8.07698250e-01 1.58172086e-01 8.61203074e-01 1.18088402e-01
-4.98372316e-01 9.31038380e-01 -3.85134459e-01 4.40297455e-01
4.56384242e-01 -1.39363706e+00 8.38438272e-01 9.78169680e-01
2.48432457e-01 5.92025399e-01 4.83175397e-01 -3.77641201e-01
-8.63936186e-01 -8.65375876e-01 1.90097845e+00 -9.54691172e-01
7.27531374e-01 -1.84132531e-01 -1.51440072e+00 7.12689877e-01
5.49566507e-01 -3.95226866e-01 6.12271786e-01 2.85258442e-01
-6.68162346e-01 1.47048548e-01 -7.23105192e-01 4.58546072e-01
8.21799338e-01 -9.63940322e-01 -1.54547751e+00 2.98554450e-01
1.35280800e+00 -7.00914264e-01 -5.49879789e-01 3.94945383e-01
2.75213867e-01 -9.65727270e-01 7.41728008e-01 -1.14464438e+00
8.26511145e-01 -2.11962819e-01 -4.09409069e-02 -8.40605021e-01
-1.96558863e-01 -8.06759357e-01 -3.59712481e-01 1.52097964e+00
1.05260420e+00 -1.66187510e-01 5.67264915e-01 4.22222376e-01
-4.57238972e-01 -1.13120401e+00 -1.21397650e+00 -1.39549360e-01
5.90995789e-01 -2.96587974e-01 7.47448325e-01 3.39141697e-01
2.79078394e-01 9.44197237e-01 2.73408622e-01 -3.70715698e-03
-2.16265842e-02 5.27113974e-01 4.68724579e-01 -8.09276521e-01
-5.48428237e-01 -5.76806784e-01 2.65724570e-01 -1.77396393e+00
4.81244624e-01 -1.12920082e+00 2.76602119e-01 -1.77807653e+00
1.33520946e-01 -1.76086918e-01 -1.07832253e-01 2.05456078e-01
-7.79775798e-01 -2.06797733e-03 2.51008067e-02 -1.05830468e-02
-1.07917964e+00 3.93667996e-01 1.25015187e+00 -1.63784981e-01
-6.28309846e-02 3.24511439e-01 -7.22062707e-01 7.50970006e-01
5.57771862e-01 -2.88769484e-01 -8.26205090e-02 -9.49386001e-01
5.98139167e-01 4.12609577e-01 -3.21811363e-02 -5.60067773e-01
4.25734699e-01 1.12236828e-01 6.98589459e-02 -9.20845866e-01
-1.27350524e-01 -5.36525659e-02 -6.79737031e-01 2.07471281e-01
-1.11032295e+00 4.15804386e-01 -7.26479515e-02 2.52889723e-01
-2.89837956e-01 -7.37530529e-01 6.07031763e-01 -1.77711114e-01
-1.65514633e-01 -1.31784990e-01 -8.88654739e-02 1.03405511e+00
4.84773308e-01 1.49756312e-01 -8.21310580e-01 -4.69780266e-01
-6.67363644e-01 6.49209738e-01 -4.48457114e-02 3.58645797e-01
6.78105354e-01 -8.54222476e-01 -1.23272645e+00 -3.82640655e-03
4.99473847e-02 1.99832350e-01 3.26322228e-01 8.54579270e-01
-5.65538466e-01 5.27740657e-01 4.05020624e-01 -2.32972875e-01
-1.30153763e+00 3.28892469e-01 4.94090885e-01 -7.31527984e-01
-3.59036833e-01 1.67015588e+00 -1.28143027e-01 -8.13073039e-01
3.64686072e-01 -6.37822330e-01 -5.48251331e-01 1.86257496e-01
8.03662002e-01 3.63218725e-01 1.20595485e-01 -6.88485384e-01
-2.21217528e-01 5.21503091e-01 -2.31641918e-01 -5.28071783e-02
1.01514220e+00 -3.88595551e-01 -4.36283499e-01 3.10083270e-01
1.26296175e+00 3.78011584e-01 -1.02516663e+00 -4.11952257e-01
3.66918445e-01 1.03308134e-01 -5.37434757e-01 -1.14753127e+00
-4.75426167e-01 1.15578592e+00 -2.93870509e-01 1.58328891e-01
7.84003615e-01 5.14122546e-01 1.45875943e+00 4.05084163e-01
-3.38516951e-01 -1.06614399e+00 1.85025305e-01 1.21168566e+00
1.21332550e+00 -9.64264095e-01 -5.16886413e-01 -3.90554607e-01
-5.73744953e-01 1.00619960e+00 8.57205570e-01 1.03720270e-01
1.02612831e-01 -1.65890157e-01 2.39609331e-01 -2.29292363e-01
-9.82824564e-01 -2.61191845e-01 9.48645413e-01 2.36730769e-01
6.31447732e-01 -2.29779154e-01 -5.06607354e-01 1.05925465e+00
-7.77933061e-01 -4.88359034e-01 2.57538617e-01 4.46308881e-01
-7.66649723e-01 -8.47205818e-01 -3.25417109e-02 4.29476470e-01
-7.93565989e-01 -4.13856268e-01 -5.54203689e-01 1.94022939e-01
-3.36569780e-03 1.27661967e+00 1.16686687e-01 -9.71592665e-02
6.64686143e-01 4.25890088e-01 3.72974843e-01 -7.05994725e-01
-1.29752302e+00 -4.10147041e-01 5.97272038e-01 -4.69314158e-01
5.98264821e-02 -6.03605092e-01 -1.60425341e+00 -1.48701388e-02
-4.29754674e-01 3.59634697e-01 9.73639190e-02 1.27171862e+00
3.42644185e-01 6.30322099e-01 3.74755234e-01 1.56147555e-01
-9.90865231e-01 -1.11035323e+00 2.88098097e-01 5.17458141e-01
4.71644491e-01 9.89445671e-02 -2.79091060e-01 -1.11397892e-01] | [11.38524055480957, 8.086621284484863] |
15d54782-6c49-4228-b270-1e2afdd6c8fe | landcover-ai-dataset-for-automatic-mapping-of | 2005.02264 | null | https://arxiv.org/abs/2005.02264v4 | https://arxiv.org/pdf/2005.02264v4.pdf | LandCover.ai: Dataset for Automatic Mapping of Buildings, Woodlands, Water and Roads from Aerial Imagery | Monitoring of land cover and land use is crucial in natural resources management. Automatic visual mapping can carry enormous economic value for agriculture, forestry, or public administration. Satellite or aerial images combined with computer vision and deep learning enable precise assessment and can significantly speed up change detection. Aerial imagery usually provides images with much higher pixel resolution than satellite data allowing more detailed mapping. However, there is still a lack of aerial datasets made for the segmentation, covering rural areas with a resolution of tens centimeters per pixel, manual fine labels, and highly publicly important environmental instances like buildings, woods, water, or roads. Here we introduce LandCover.ai (Land Cover from Aerial Imagery) dataset for semantic segmentation. We collected images of 216.27 sq. km rural areas across Poland, a country in Central Europe, 39.51 sq. km with resolution 50 cm per pixel and 176.76 sq. km with resolution 25 cm per pixel and manually fine annotated four following classes of objects: buildings, woodlands, water, and roads. Additionally, we report simple benchmark results, achieving 85.56% of mean intersection over union on the test set. It proves that the automatic mapping of land cover is possible with a relatively small, cost-efficient, RGB-only dataset. The dataset is publicly available at https://landcover.ai.linuxpolska.com/ | ['Natalia Ziemba-Jankowska', 'Dominik Batorski', 'Adrian Boguszewski', 'Tomasz Dziedzic', 'Anna Zambrzycka'] | 2020-05-05 | null | null | null | null | ['object-detection-in-aerial-images', 'semantic-segmentation-of-orthoimagery'] | ['computer-vision', 'medical'] | [ 3.77584130e-01 -5.93996644e-02 -2.70605564e-01 -2.50911236e-01
-3.68712604e-01 -8.90492916e-01 4.19950187e-01 3.36887777e-01
-6.82290614e-01 1.06451058e+00 -2.00106069e-01 -7.52078176e-01
-1.36116371e-01 -1.62527382e+00 -6.59936249e-01 -8.17623675e-01
-5.80471992e-01 2.92287171e-01 9.61988121e-02 -2.51640052e-01
-2.33292073e-01 7.91165590e-01 -1.64329302e+00 -2.30336711e-01
9.87744927e-01 9.38976586e-01 6.62919164e-01 5.21160901e-01
-5.24084531e-02 -6.44122064e-02 -1.00829326e-01 3.51387151e-02
6.66565478e-01 1.36952758e-01 -9.23016667e-01 3.04695189e-01
4.56351727e-01 -6.75544441e-01 -3.05799525e-02 1.39020514e+00
1.88058451e-01 -2.04531759e-01 5.62915623e-01 -9.30132151e-01
-2.82164246e-01 5.42501032e-01 -1.12256098e+00 8.85946769e-03
-2.05628917e-01 2.88786858e-01 8.55830967e-01 -4.51164126e-01
6.65336370e-01 9.13415611e-01 7.13459432e-01 -3.57932508e-01
-1.18162549e+00 -6.18760407e-01 2.54585326e-01 -4.57844250e-02
-1.83970642e+00 -1.58635587e-01 1.28068537e-01 -4.43676889e-01
7.38002539e-01 6.68572783e-01 9.42796052e-01 1.22010648e-01
1.49863642e-02 4.07385588e-01 1.25425565e+00 -1.73783571e-01
9.51692760e-02 -4.48664278e-02 -4.54904176e-02 6.15824103e-01
5.94167948e-01 7.93610290e-02 4.74924326e-01 1.80542797e-01
9.53876376e-01 3.32226336e-01 -2.02741802e-01 -1.83700725e-01
-1.08691800e+00 6.77079439e-01 1.20054758e+00 3.73269230e-01
-6.64061666e-01 1.20145068e-01 1.74736232e-01 2.34307691e-01
5.96992493e-01 1.89773694e-01 -6.46374166e-01 3.73711318e-01
-1.16765141e+00 2.59683654e-02 2.05985844e-01 8.93916130e-01
1.26749086e+00 -7.70407543e-02 4.93034512e-01 6.39403164e-01
1.14213988e-01 1.12301672e+00 -2.05669422e-02 -9.38435853e-01
5.01344979e-01 8.94256234e-01 3.62663478e-01 -1.22711217e+00
-6.96909428e-01 -2.95165986e-01 -1.42012537e+00 4.94356692e-01
6.80920109e-02 -1.32937536e-01 -1.32710266e+00 1.19933999e+00
2.46406227e-01 -4.22294468e-01 -1.37426360e-02 9.31439638e-01
6.14988148e-01 8.44982326e-01 5.08917391e-01 -3.57989557e-02
1.45604289e+00 -2.57332981e-01 -3.82141709e-01 -3.28003943e-01
5.27923465e-01 -2.17261374e-01 9.28975105e-01 -2.25192364e-02
-4.87094104e-01 -4.80393291e-01 -8.25661838e-01 3.91945392e-01
-9.39357042e-01 5.47086060e-01 1.02299857e+00 4.36755776e-01
-1.08716547e+00 6.07494533e-01 -7.91082740e-01 -6.41690254e-01
9.51707900e-01 2.58726895e-01 -6.86515868e-01 9.67646297e-03
-1.17493606e+00 7.41609812e-01 6.99302197e-01 4.56677258e-01
-5.73855519e-01 -3.86583209e-01 -1.00759459e+00 -4.48455475e-02
1.80503592e-01 -9.55953822e-02 5.73082626e-01 -1.26034939e+00
-7.50973046e-01 1.37100816e+00 3.15639108e-01 -5.80058992e-01
5.54943264e-01 -9.51636957e-06 -4.39676404e-01 1.12008207e-01
6.66167140e-01 9.73733664e-01 3.77446949e-01 -1.16684902e+00
-1.23657882e+00 -6.33527339e-01 1.77774698e-01 1.58308238e-01
-3.99650931e-02 -1.08192742e-01 1.06099181e-01 -4.95702237e-01
3.66329610e-01 -8.04474831e-01 -5.45982778e-01 2.16515284e-04
-2.56295979e-01 2.95024067e-01 4.88857985e-01 -1.01222599e+00
1.12004864e+00 -2.00758529e+00 -2.91070491e-01 2.41149127e-01
-1.49756938e-01 3.27978760e-01 -8.96179155e-02 1.68716595e-01
-1.25145838e-01 5.38983822e-01 -1.13133395e+00 6.47919893e-01
-2.51247883e-01 4.20490921e-01 -2.02528834e-01 5.33580959e-01
2.96503752e-01 8.72105598e-01 -8.77047956e-01 -4.09713805e-01
6.32860661e-01 9.32048187e-02 1.68442860e-01 -2.56600171e-01
3.12127825e-02 1.63181096e-01 -4.42577571e-01 1.12732351e+00
1.31265438e+00 1.64982349e-01 1.28834069e-01 -6.26987889e-02
-7.00701475e-01 -2.58825630e-01 -1.36946642e+00 1.16873574e+00
-3.12389374e-01 8.46461833e-01 2.98207134e-01 -1.01673484e+00
9.01908159e-01 -5.04582860e-02 4.63431895e-01 -5.84822536e-01
-6.08439781e-02 2.01050922e-01 -2.83255845e-01 -4.35040623e-01
6.50071144e-01 -8.73069912e-02 -9.24361274e-02 4.29507755e-02
-3.22872937e-01 -3.42833072e-01 1.42393515e-01 -2.76265174e-01
7.08982468e-01 2.59881973e-01 7.13161469e-01 -4.70887452e-01
1.72935858e-01 7.87900507e-01 3.81328851e-01 4.45701748e-01
-3.44955981e-01 4.54151392e-01 2.66893119e-01 -8.97864997e-01
-1.19332325e+00 -9.47422743e-01 -7.06596911e-01 1.02478957e+00
2.56574541e-01 2.30286509e-01 -5.75257659e-01 -2.33053312e-01
3.65362376e-01 4.61721748e-01 -4.98361707e-01 3.07976454e-01
-3.87997329e-01 -1.27773654e+00 5.06664991e-01 5.43655276e-01
1.36060941e+00 -1.31356728e+00 -6.91197336e-01 2.14161485e-01
-2.11916968e-01 -9.59282696e-01 3.78727019e-01 2.61288673e-01
-1.05678761e+00 -1.05184793e+00 -7.76841044e-01 -6.33349061e-01
8.00287485e-01 5.71324110e-01 1.14243770e+00 -1.16234817e-01
-6.30103290e-01 -8.84811208e-03 -2.26323262e-01 -2.20892444e-01
-1.32716641e-01 3.26237887e-01 -1.84313968e-01 -1.68137506e-01
2.82358319e-01 -6.22087300e-01 -5.98318040e-01 3.66051912e-01
-9.14007425e-01 2.26943642e-01 9.46622074e-01 3.31086040e-01
8.18756521e-01 5.86373866e-01 2.62461334e-01 -6.78169370e-01
-5.05715132e-01 -7.40639806e-01 -9.03531909e-01 2.13851824e-01
-3.54806304e-01 -4.43444341e-01 1.95592180e-01 2.71191180e-01
-8.34558070e-01 6.83794081e-01 2.61809584e-02 4.60873127e-01
-9.94336963e-01 6.16347313e-01 -3.74687493e-01 1.04440711e-02
6.20630324e-01 1.58364370e-01 -4.05416101e-01 -4.53509092e-01
3.55269969e-01 1.10197401e+00 8.12675476e-01 -7.05484226e-02
1.03473198e+00 7.49403179e-01 -8.68099369e-03 -1.24692094e+00
-3.83434117e-01 -7.79517293e-01 -1.24411881e+00 -8.42439532e-02
7.83153653e-01 -1.25408316e+00 -2.12017268e-01 5.19024611e-01
-6.94956720e-01 -5.49223185e-01 -1.44223988e-01 2.40157560e-01
-3.97767991e-01 9.39665362e-02 5.96280172e-02 -6.48376584e-01
-3.73848826e-01 -5.28179705e-01 9.12328660e-01 2.97188908e-01
9.05468613e-02 -7.27886379e-01 -3.21620584e-01 1.03902683e-01
3.43144238e-01 8.29209745e-01 5.57229102e-01 3.40771019e-01
-6.70620322e-01 -1.84438780e-01 -8.96548867e-01 3.99536967e-01
5.58084726e-01 2.42537320e-01 -7.98117340e-01 -1.61466897e-01
-6.48501992e-01 2.04046983e-02 1.32164884e+00 7.60339499e-01
9.81517673e-01 -2.51556009e-01 -5.54620445e-01 7.02053249e-01
1.96813095e+00 2.95877140e-02 8.12687635e-01 6.91775799e-01
5.71157753e-01 6.37746096e-01 9.73923981e-01 4.00894612e-01
1.39192760e-01 1.99505284e-01 1.06320083e+00 -6.49379551e-01
1.71124667e-01 2.60741919e-01 9.02467072e-02 -2.79606968e-01
-6.96652770e-01 2.61899736e-02 -1.43797362e+00 9.90345478e-01
-1.55267560e+00 -1.08531964e+00 -5.45522332e-01 2.12068653e+00
8.17931116e-01 -1.26230955e-01 -5.80395833e-02 2.73572147e-01
8.98946524e-01 2.17777669e-01 -4.58154410e-01 1.56806827e-01
-4.62979734e-01 2.61899412e-01 1.65414655e+00 5.42293847e-01
-1.95504129e+00 1.32981884e+00 5.10339117e+00 5.85805535e-01
-9.59699154e-01 -2.04797909e-01 5.15247762e-01 4.48636711e-01
3.50979884e-04 -1.90362528e-01 -6.93421602e-01 2.95562923e-01
6.17548585e-01 9.85914320e-02 3.00742805e-01 8.71941090e-01
4.42722887e-01 -6.21246696e-01 -3.43035907e-02 6.76052570e-01
-7.28254914e-01 -1.17350543e+00 5.10238595e-02 2.14672431e-01
8.33508551e-01 3.88981253e-01 -2.20492110e-01 -2.40891688e-02
6.15072906e-01 -1.00210190e+00 4.76647168e-01 4.26835388e-01
1.41401660e+00 -7.68051088e-01 9.77117121e-01 2.88285911e-01
-1.48422432e+00 -1.57381341e-01 -6.80360258e-01 -1.94523469e-01
-2.45557308e-01 4.90032971e-01 -4.51417625e-01 6.10221863e-01
1.10971928e+00 9.09140706e-01 -7.62166619e-01 1.01727235e+00
-3.00434321e-01 6.64560199e-01 -5.57709515e-01 4.10603255e-01
7.62442410e-01 -4.78792459e-01 2.71177799e-01 1.16970992e+00
3.58391225e-01 3.98185372e-01 3.08289379e-01 5.06813586e-01
1.33385316e-01 -5.26434109e-02 -8.61926675e-01 1.48100495e-01
2.82498151e-01 1.48693776e+00 -1.11705816e+00 -1.93497121e-01
3.12438551e-02 9.18254852e-01 -1.56141803e-01 3.07925135e-01
-6.73057973e-01 -5.76025486e-01 8.29878330e-01 3.16560000e-01
5.86687550e-02 -3.98627281e-01 -1.56200394e-01 -7.00325012e-01
-5.87537587e-02 -3.29963714e-01 1.51598975e-01 -7.16504693e-01
-6.66638315e-01 2.40573779e-01 1.71855360e-01 -1.28664529e+00
1.39431670e-01 -7.85224259e-01 -2.06229523e-01 9.56371725e-01
-2.00339532e+00 -1.43657088e+00 -9.91456091e-01 1.44019991e-01
3.84768307e-01 -8.32368433e-03 1.07480478e+00 1.01040460e-01
-4.28014189e-01 -2.84880579e-01 4.89827991e-01 3.21414441e-01
1.90346047e-01 -1.46441889e+00 7.51968980e-01 7.26730824e-01
-3.60150933e-01 -1.27334267e-01 2.95784652e-01 -6.12785041e-01
-8.23728144e-01 -1.86688232e+00 6.76909804e-01 7.88032487e-02
5.61364532e-01 1.93228036e-01 -7.81752706e-01 6.71115816e-01
-2.43363425e-01 1.22706413e-01 3.77492607e-01 -2.74662793e-01
2.02669904e-01 -4.51261133e-01 -1.41256928e+00 3.83689374e-01
1.13675916e+00 -1.02395192e-01 -4.89070527e-02 4.86146182e-01
2.92458981e-01 -4.08793241e-02 -7.39367247e-01 5.07217288e-01
6.47383392e-01 -8.28467250e-01 1.04858577e+00 -2.26664290e-01
4.31576759e-01 -4.18588936e-01 -5.03520846e-01 -1.27606952e+00
-7.43459344e-01 4.70450595e-02 9.31484580e-01 1.21881413e+00
2.37363756e-01 -7.00830936e-01 5.71943164e-01 1.40655071e-01
-1.34366110e-01 5.64552564e-03 -7.43736744e-01 -9.82252777e-01
2.42562756e-01 -4.33136970e-01 8.15486968e-01 8.76649380e-01
-8.15248013e-01 -2.51883715e-01 9.88658294e-02 7.40021825e-01
5.58589697e-01 4.24607813e-01 7.23797262e-01 -1.70543313e+00
6.26798987e-01 -5.80766797e-01 -6.85492933e-01 -3.91595006e-01
2.86590811e-02 -6.82794750e-01 -2.40779534e-01 -2.03925872e+00
1.44051328e-01 -8.91202688e-01 3.11728809e-02 1.15535283e+00
-5.45924418e-02 6.61829054e-01 -1.97789282e-01 3.25924397e-01
-2.12236643e-02 1.90976918e-01 8.45023155e-01 -6.64296627e-01
-2.76421875e-01 1.76925763e-01 -4.28922921e-01 9.69554901e-01
1.42558527e+00 -3.82336229e-01 2.61825353e-01 -4.99200761e-01
2.69558400e-01 -3.22820485e-01 8.41441870e-01 -1.07418025e+00
-4.27319616e-01 -5.82216263e-01 5.13788044e-01 -1.05579698e+00
-2.29658931e-01 -1.25994337e+00 4.78784859e-01 7.36644924e-01
2.14474037e-01 -5.05149961e-01 5.13632178e-01 2.58791268e-01
-8.57509524e-02 -9.67252851e-02 9.77611840e-01 -4.20610070e-01
-1.40785801e+00 4.74048853e-01 -4.50376689e-01 -5.31424880e-01
1.24850178e+00 -3.15099925e-01 -3.11985463e-02 2.88891662e-02
-6.70807302e-01 3.59076768e-01 6.71246111e-01 1.71748936e-01
1.77698404e-01 -1.27269912e+00 -1.01461256e+00 3.87243003e-01
3.02836716e-01 1.47308543e-01 6.29979596e-02 4.22484875e-01
-1.24399757e+00 4.90465909e-01 -7.82065332e-01 -6.35225892e-01
-1.18593955e+00 1.61309704e-01 4.80695575e-01 1.32046953e-01
-5.39580166e-01 4.53443110e-01 -4.33288291e-02 -5.92780292e-01
-2.35721216e-01 -5.93180895e-01 -5.21193385e-01 4.73307997e-01
3.62894475e-01 4.30239350e-01 -4.61802334e-02 -9.58331168e-01
-4.16125774e-01 8.50828171e-01 7.33202696e-01 1.93145543e-01
1.55304039e+00 -4.37721848e-01 -4.48494554e-01 3.33968490e-01
7.40680516e-01 -4.83412743e-01 -1.23958719e+00 -1.69484302e-01
9.01699290e-02 -4.99860734e-01 5.02547562e-01 -9.38256621e-01
-1.03533602e+00 8.26860428e-01 1.12380445e+00 4.69928265e-01
1.35465479e+00 -1.15525126e-01 2.63685465e-01 4.75204259e-01
7.23710775e-01 -1.09300768e+00 -1.14084911e+00 2.72396654e-01
1.00654125e+00 -1.80958128e+00 3.61912549e-01 -3.42642367e-01
-3.34895700e-01 1.08226395e+00 2.00156972e-01 5.39883226e-02
5.61389923e-01 1.49384707e-01 2.96839532e-02 2.30593104e-02
1.55621573e-01 -1.13964558e+00 -3.51159200e-02 8.22589040e-01
1.69774786e-01 9.57085192e-01 2.45774295e-02 1.63485140e-01
-2.16123924e-01 -2.01171279e-01 3.18606883e-01 7.52994180e-01
-1.07991087e+00 -4.23721761e-01 -6.77186430e-01 6.87805176e-01
-1.23131603e-01 -3.40716332e-01 -2.65292406e-01 1.13851559e+00
4.52768743e-01 7.23373234e-01 4.24685776e-01 -3.83945405e-02
2.02757314e-01 -3.07319701e-01 1.49236172e-01 -4.87465322e-01
-1.81199312e-01 -2.40248084e-01 -7.04432651e-02 -3.65898043e-01
-6.27296567e-01 -7.62095928e-01 -1.23679519e+00 -6.46867514e-01
-1.03803933e-01 -1.63774669e-01 9.50022221e-01 4.92718130e-01
1.90106835e-02 1.17968291e-01 8.02986026e-01 -1.02697766e+00
-1.49686141e-02 -1.13602054e+00 -1.14216721e+00 -1.99958086e-01
4.14795130e-01 -5.28950810e-01 -2.95201004e-01 2.47563913e-01] | [9.396011352539062, -1.5365546941757202] |
bc9c5cff-9d97-4f05-9477-e6267bef80f7 | noisy-channel-language-model-prompting-for | 2108.04106 | null | https://arxiv.org/abs/2108.04106v3 | https://arxiv.org/pdf/2108.04106v3.pdf | Noisy Channel Language Model Prompting for Few-Shot Text Classification | We introduce a noisy channel approach for language model prompting in few-shot text classification. Instead of computing the likelihood of the label given the input (referred as direct models), channel models compute the conditional probability of the input given the label, and are thereby required to explain every word in the input. We use channel models for recently proposed few-shot learning methods with no or very limited updates to the language model parameters, via either in-context demonstration or prompt tuning. Our experiments show that, for both methods, channel models significantly outperform their direct counterparts, which we attribute to their stability, i.e., lower variance and higher worst-case accuracy. We also present extensive ablations that provide recommendations for when to use channel prompt tuning instead of other competitive methods (e.g., direct head tuning): channel prompt tuning is preferred when the number of training examples is small, labels in the training data are imbalanced, or generalization to unseen labels is required. | ['Luke Zettlemoyer', 'Hannaneh Hajishirzi', 'Mike Lewis', 'Sewon Min'] | 2021-08-09 | null | https://aclanthology.org/2022.acl-long.365 | https://aclanthology.org/2022.acl-long.365.pdf | acl-2022-5 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 4.12224978e-01 1.56603456e-01 -5.56789100e-01 -5.16419768e-01
-1.18977106e+00 -4.93690223e-01 6.65406346e-01 4.30338025e-01
-6.98206544e-01 7.59205639e-01 3.29877108e-01 -3.30026746e-01
1.87892586e-01 -7.10021079e-01 -5.16684830e-01 -5.18926561e-01
2.65517443e-01 4.54513937e-01 1.05498560e-01 -9.87923294e-02
2.03253508e-01 -2.54941553e-01 -1.59244132e+00 9.10035968e-02
4.18710113e-01 1.03244865e+00 9.33310166e-02 9.78238523e-01
-4.64832097e-01 8.46733034e-01 -4.96401519e-01 -3.07107538e-01
-7.99467638e-02 -6.10890865e-01 -8.68931055e-01 2.83775385e-02
-2.31125038e-02 -4.06016588e-01 -3.22443783e-01 8.59246790e-01
7.12056160e-01 3.51746649e-01 6.76233828e-01 -1.17347157e+00
-4.27897245e-01 7.71847546e-01 -2.62075454e-01 3.39391708e-01
1.95420429e-01 8.94652307e-02 1.44738996e+00 -9.63097334e-01
6.38325810e-01 1.06092441e+00 4.91215169e-01 9.76434767e-01
-1.51827097e+00 -6.39508784e-01 5.49633026e-01 1.84099406e-01
-1.16538870e+00 -7.70122826e-01 3.28014702e-01 -3.94983351e-01
9.70527768e-01 -4.38265018e-02 3.43658894e-01 1.39092219e+00
3.61172967e-02 8.23543131e-01 9.49777365e-01 -7.56291509e-01
9.08240616e-01 4.82903838e-01 6.70009613e-01 3.75405699e-01
7.78509453e-02 2.42896490e-02 -9.76924956e-01 -5.10427058e-01
6.91055879e-02 -1.32058552e-02 4.53621373e-02 -1.70842811e-01
-8.44505370e-01 1.18841290e+00 9.10502579e-03 -1.34871304e-02
-3.15776676e-01 4.71566945e-01 4.38483536e-01 3.48217964e-01
5.28490245e-01 3.97470325e-01 -5.31336546e-01 -5.15448272e-01
-8.27197075e-01 2.90292829e-01 8.58917832e-01 1.17548215e+00
8.28241885e-01 -7.21974373e-02 -6.18372738e-01 7.74964273e-01
2.37757161e-01 3.19173604e-01 4.14812773e-01 -7.73496687e-01
4.37238008e-01 2.69975718e-02 2.72048235e-01 -6.80287257e-02
-5.03151298e-01 -3.74117553e-01 -2.51223356e-01 -2.04647183e-01
3.13999623e-01 -5.36731243e-01 -1.13139570e+00 1.85093498e+00
-9.41042602e-02 2.51644909e-01 2.16784030e-01 6.25503957e-01
7.55126655e-01 6.01064384e-01 5.11472106e-01 -4.31378394e-01
1.33711004e+00 -6.76770568e-01 -8.68497312e-01 -4.96024251e-01
1.00369847e+00 -6.58184290e-01 1.28234231e+00 2.31219262e-01
-6.88209593e-01 -1.68332130e-01 -1.00159526e+00 4.31439281e-02
-3.27544361e-01 -3.92887324e-01 7.91868567e-01 8.28001618e-01
-6.10230744e-01 5.21867394e-01 -6.82387590e-01 -4.31438416e-01
4.43484247e-01 1.97877094e-01 3.36768001e-01 -1.23679101e-01
-1.57509041e+00 4.94581491e-01 1.43077135e-01 -5.07540345e-01
-9.35537279e-01 -5.22857726e-01 -6.90510273e-01 2.84139931e-01
4.50864315e-01 -3.57656419e-01 1.83378351e+00 -6.52707100e-01
-1.39970422e+00 4.54404682e-01 -5.19723475e-01 -4.18007791e-01
3.10274988e-01 -5.79770356e-02 -2.46301770e-01 -1.94021255e-01
3.54918912e-02 7.59495258e-01 8.52214694e-01 -9.68827903e-01
-6.93374157e-01 -4.84792627e-02 1.37009397e-01 1.94558486e-01
-6.77376449e-01 -1.87171362e-02 -5.46426296e-01 -2.72313386e-01
3.40379104e-02 -8.12488019e-01 -4.94144768e-01 -2.00562924e-01
-5.29317379e-01 -2.44256929e-01 5.58738351e-01 -1.01491727e-01
1.42574573e+00 -2.14259124e+00 -4.81568992e-01 3.18635076e-01
1.25168860e-01 -7.37173781e-02 -3.63103040e-02 4.23486143e-01
2.67224640e-01 3.23657900e-01 -5.15435860e-02 -4.03919667e-01
1.65536031e-01 3.01240563e-01 -5.07501483e-01 2.36546725e-01
-1.16362929e-01 8.66548479e-01 -9.45322931e-01 -3.09114397e-01
8.47165734e-02 3.01986784e-01 -5.36743283e-01 1.17573433e-01
-3.85316104e-01 1.69481665e-01 -4.44525719e-01 4.61083502e-01
1.89381257e-01 -6.00792825e-01 -4.25264947e-02 2.96215892e-01
1.37389019e-01 5.44296265e-01 -1.26012075e+00 1.44332445e+00
-6.19761527e-01 4.93059158e-01 -1.91265240e-01 -5.90307415e-01
7.01232731e-01 5.83822608e-01 2.22359281e-02 -8.06160331e-01
1.96473733e-01 2.26868577e-02 -1.54248670e-01 -3.15889984e-01
4.54972953e-01 -3.52896452e-01 -1.69950202e-01 9.14150655e-01
1.79885373e-01 7.52584115e-02 1.03112206e-01 6.27952993e-01
1.03076339e+00 -2.98938990e-01 4.31033760e-01 -6.65314496e-02
-2.82112151e-01 -2.20796138e-01 3.80478531e-01 1.47309816e+00
-3.40013295e-01 4.86177981e-01 5.60333490e-01 -5.52672371e-02
-9.52218473e-01 -6.82113171e-01 -1.98853612e-01 1.65965140e+00
2.49666303e-01 -7.82092392e-01 -6.37537301e-01 -5.58049500e-01
-8.73704329e-02 1.05467749e+00 -6.33923590e-01 -2.98089713e-01
1.76808417e-01 -8.17030847e-01 3.69594485e-01 7.60612607e-01
-1.63017005e-01 -7.75862575e-01 -7.54468203e-01 4.02233452e-01
-4.97525260e-02 -1.05173147e+00 -3.31412911e-01 1.03857970e+00
-7.76925504e-01 -6.59269691e-01 -6.80918992e-01 -3.55685830e-01
4.19715196e-01 3.09396803e-01 6.91707432e-01 -4.00944203e-02
-1.89753592e-01 4.91320580e-01 -5.13213336e-01 -6.71136737e-01
-1.34372771e-01 1.05414063e-01 4.81712893e-02 -8.04276951e-03
6.94852889e-01 -1.90749705e-01 -5.17327189e-01 1.18409269e-01
-5.95743179e-01 -2.34223142e-01 3.44400227e-01 1.12418365e+00
4.80164677e-01 -3.61667126e-01 6.59666300e-01 -1.41458631e+00
9.21002865e-01 -5.61730504e-01 -2.19458148e-01 1.87280133e-01
-1.09613109e+00 2.24637330e-01 5.14033318e-01 -4.66059476e-01
-8.60371470e-01 4.91813757e-02 -2.65754610e-01 -2.60730386e-01
-2.89787203e-01 4.92112905e-01 2.27407247e-01 2.28223950e-01
1.07014537e+00 9.19260159e-02 -2.87286371e-01 -3.40695858e-01
5.66051185e-01 9.73131418e-01 1.03239775e-01 -4.76891071e-01
3.43978763e-01 2.93637574e-01 -5.78575671e-01 -5.88892400e-01
-9.99584079e-01 -7.14770913e-01 -3.89222801e-01 -3.10441628e-02
7.88673282e-01 -1.00717521e+00 -4.96225834e-01 1.58677883e-02
-1.06023836e+00 -3.47619593e-01 -3.15896988e-01 5.66305578e-01
-6.07998848e-01 2.30791532e-02 -7.41550446e-01 -1.19305897e+00
-2.78261214e-01 -1.02142859e+00 1.00667846e+00 1.34935006e-01
-6.22194529e-01 -1.04307938e+00 -1.82995528e-01 -3.00327968e-03
5.48669219e-01 -5.32431185e-01 1.14944673e+00 -1.10730958e+00
-2.35282540e-01 -5.17095923e-01 -1.48306742e-01 -2.15945076e-02
-3.90880965e-02 -3.16118896e-01 -1.75019813e+00 -2.34047472e-01
-3.28775823e-01 -6.91281736e-01 1.03185618e+00 4.68770832e-01
1.12377036e+00 -9.74321812e-02 -3.18075180e-01 2.17235655e-01
1.20899034e+00 2.74425179e-01 1.84496939e-01 3.08455992e-02
4.47426021e-01 5.63059151e-01 7.12360799e-01 1.02200568e+00
8.46839026e-02 6.68704033e-01 5.53439111e-02 1.25479475e-01
1.71686098e-01 -4.35417533e-01 2.39570543e-01 4.46163148e-01
4.73133802e-01 -4.65372145e-01 -8.69407177e-01 2.85656154e-01
-1.86521256e+00 -8.09838474e-01 9.42146778e-02 2.36897039e+00
7.19174206e-01 6.41929269e-01 3.38600390e-02 1.22032426e-01
7.60901511e-01 1.82079658e-01 -7.61082649e-01 -5.62388957e-01
1.44660082e-02 2.31226072e-01 6.26005292e-01 7.64747500e-01
-9.86279666e-01 1.08926916e+00 7.30244875e+00 9.38947082e-01
-1.05072176e+00 4.99577820e-01 7.31501758e-01 -3.64135891e-01
-5.72543979e-01 2.29214519e-01 -1.09525228e+00 3.53420496e-01
1.15574932e+00 -2.97349334e-01 4.83734131e-01 9.24428284e-01
1.06364973e-01 -2.29335293e-01 -1.26683950e+00 1.13515019e+00
-1.24084011e-01 -1.34442413e+00 -2.27167517e-01 -3.84808425e-03
6.14674211e-01 5.32484613e-02 -7.81928971e-02 5.80212772e-01
3.61477882e-01 -8.92476737e-01 7.58889437e-01 2.04080015e-01
8.90019536e-01 -6.63353086e-01 5.56766152e-01 5.14190078e-01
-9.22794104e-01 -3.62530321e-01 -4.11846638e-01 -3.29638749e-01
2.35418037e-01 6.05349302e-01 -8.59118581e-01 -6.67788088e-02
4.30879802e-01 4.47149605e-01 -3.13622534e-01 7.69079328e-01
-1.93642214e-01 1.10539496e+00 -2.93492407e-01 -5.15234768e-01
2.64829665e-01 5.54555297e-01 3.21742535e-01 1.31862342e+00
2.07780059e-02 1.36829183e-01 4.11693722e-01 6.56701803e-01
-9.68736410e-02 4.18154240e-01 -6.13157630e-01 -1.44185469e-01
8.62438560e-01 1.08148801e+00 -7.68597424e-01 -6.89125896e-01
-5.79197764e-01 7.23079443e-01 1.74480826e-01 6.45124435e-01
-4.37533051e-01 -4.93684232e-01 3.79582971e-01 -2.24332988e-01
2.57213473e-01 2.85110325e-01 -5.97262502e-01 -9.53648865e-01
-2.78504997e-01 -3.90547931e-01 5.59041083e-01 -5.29205978e-01
-1.28138900e+00 4.00355071e-01 -1.79495439e-01 -1.16557598e+00
-4.62515622e-01 -4.04580355e-01 -7.49783158e-01 8.16734910e-01
-1.50951898e+00 -5.42582989e-01 -1.96741242e-02 4.91960853e-01
7.37729609e-01 -5.04059158e-02 1.19933331e+00 1.61555454e-01
-3.35873008e-01 9.53429043e-01 8.98891538e-02 -1.89918410e-02
9.45873499e-01 -1.11300707e+00 4.24836129e-01 4.76605445e-01
1.73917934e-01 6.06713295e-01 9.25837398e-01 -4.41094130e-01
-1.21267390e+00 -9.69043732e-01 1.11265314e+00 -4.05699998e-01
5.77953160e-01 -8.12778473e-01 -7.57171690e-01 6.45658493e-01
-1.82940990e-01 1.07972629e-01 1.10786474e+00 6.81461513e-01
-5.76912224e-01 1.08253017e-01 -8.44616592e-01 4.97582972e-01
8.30909431e-01 -8.76043260e-01 -1.92478359e-01 3.67657006e-01
9.20772254e-01 -4.47259359e-02 -4.32454020e-01 -1.21653169e-01
5.52506626e-01 -5.86573839e-01 4.07005310e-01 -8.76824260e-01
7.83685446e-02 1.47443891e-01 -5.92464805e-01 -1.28474832e+00
-3.56254011e-01 -6.34237528e-01 -2.55796582e-01 9.71799791e-01
1.01696348e+00 -4.13670450e-01 7.25829482e-01 9.19140697e-01
1.38941780e-01 -6.89378083e-01 -8.65753353e-01 -7.10882723e-01
1.46716610e-02 -1.01554728e+00 2.40095764e-01 8.80406082e-01
4.48642969e-01 7.64391482e-01 -6.93748295e-01 -1.69970185e-01
5.06315947e-01 -8.51821005e-02 4.46225047e-01 -1.27808094e+00
-4.78628963e-01 -8.60634297e-02 -1.94949359e-01 -1.11808610e+00
1.43925130e-01 -8.65869164e-01 3.54844898e-01 -1.35967624e+00
4.95583385e-01 -4.03318167e-01 -5.05599439e-01 6.89202189e-01
-3.03895801e-01 1.17347524e-01 1.55348435e-01 1.85800105e-01
-9.31858718e-01 4.62309033e-01 7.53299415e-01 -1.15649484e-01
-3.76311481e-01 2.65832633e-01 -1.09516335e+00 3.87092680e-01
6.78596199e-01 -6.77515388e-01 -4.42656219e-01 3.53996679e-02
2.72526503e-01 9.38637480e-02 1.28128275e-01 -7.28791833e-01
5.04574418e-01 -1.15269914e-01 2.22719163e-01 -1.13972940e-01
3.44841123e-01 -5.64822078e-01 -5.62794745e-01 3.85148883e-01
-1.17065036e+00 -5.31730533e-01 -7.83132464e-02 8.96536946e-01
1.29069403e-01 -4.59223211e-01 8.48994255e-01 -1.29877210e-01
-9.10927296e-01 3.08770895e-01 -6.97374642e-01 2.60290265e-01
8.17499518e-01 -1.50346398e-01 -2.22796798e-01 -8.74140441e-01
-1.02586186e+00 1.88028976e-01 7.33937472e-02 4.87443388e-01
5.13168991e-01 -1.07463646e+00 -3.36672157e-01 1.73941255e-02
7.85673738e-01 -4.48309809e-01 9.58231091e-02 7.39376962e-01
4.01020676e-01 5.46469092e-01 4.87286925e-01 -4.75244403e-01
-1.02322185e+00 4.91592258e-01 3.25409956e-02 3.72636504e-02
-6.28506899e-01 9.85957325e-01 1.41017184e-01 -2.61336476e-01
8.20196092e-01 -8.17401558e-02 1.38775138e-02 2.85495192e-01
8.30323637e-01 1.86397374e-01 2.22874627e-01 7.99566787e-03
-2.52827615e-01 6.04361445e-02 -3.39531779e-01 -5.21597385e-01
8.35377514e-01 -2.71091372e-01 5.30389488e-01 1.10371995e+00
1.08287358e+00 -2.96368152e-01 -1.18130207e+00 -6.60281837e-01
2.47715833e-03 -2.97744364e-01 3.82578313e-01 -9.31891978e-01
-4.65497285e-01 1.20789707e+00 9.11606729e-01 1.82020009e-01
5.05385041e-01 2.58066148e-01 6.22862101e-01 7.74708211e-01
4.11404878e-01 -1.50196266e+00 -2.18111761e-02 7.78111517e-01
6.00235201e-02 -1.34463143e+00 -2.27986276e-01 -2.35859394e-01
-9.74644244e-01 9.27394092e-01 4.87940878e-01 3.45939070e-01
7.77335286e-01 4.47263867e-01 4.12242174e-01 -1.81439351e-02
-1.36412680e+00 -3.79462928e-01 -2.05979675e-01 4.38123167e-01
5.86908698e-01 -2.93725939e-03 -1.64383963e-01 7.00612724e-01
8.67299214e-02 8.23545177e-03 4.56595033e-01 1.07450116e+00
-9.50343549e-01 -8.86897564e-01 -1.77179828e-01 8.68493140e-01
-3.59023750e-01 -4.70246196e-01 -3.09703767e-01 2.22415745e-01
-2.62959182e-01 1.37377238e+00 6.33515567e-02 -5.31280696e-01
1.97117135e-01 6.37914479e-01 8.97074211e-03 -9.40733254e-01
-3.26503277e-01 3.19031477e-01 2.08445102e-01 -4.05079395e-01
-8.25727545e-03 -3.84498209e-01 -1.50103879e+00 -5.94078377e-02
-6.33055985e-01 2.90136963e-01 7.00848162e-01 1.21766222e+00
3.96319211e-01 3.16254020e-01 6.51776314e-01 -6.47017896e-01
-8.68189991e-01 -9.45685148e-01 -8.20222616e-01 2.02377111e-01
3.23324233e-01 -6.35503232e-01 -5.05387962e-01 -1.24264903e-01] | [10.862394332885742, 7.939383506774902] |
768d51d3-9416-4115-8ba0-fd279db2e024 | alfred-a-benchmark-for-interpreting-grounded | 1912.01734 | null | https://arxiv.org/abs/1912.01734v2 | https://arxiv.org/pdf/1912.01734v2.pdf | ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Tasks | We present ALFRED (Action Learning From Realistic Environments and Directives), a benchmark for learning a mapping from natural language instructions and egocentric vision to sequences of actions for household tasks. ALFRED includes long, compositional tasks with non-reversible state changes to shrink the gap between research benchmarks and real-world applications. ALFRED consists of expert demonstrations in interactive visual environments for 25k natural language directives. These directives contain both high-level goals like "Rinse off a mug and place it in the coffee maker." and low-level language instructions like "Walk to the coffee maker on the right." ALFRED tasks are more complex in terms of sequence length, action space, and language than existing vision-and-language task datasets. We show that a baseline model based on recent embodied vision-and-language tasks performs poorly on ALFRED, suggesting that there is significant room for developing innovative grounded visual language understanding models with this benchmark. | ['Winson Han', 'Roozbeh Mottaghi', 'Luke Zettlemoyer', 'Dieter Fox', 'Yonatan Bisk', 'Mohit Shridhar', 'Jesse Thomason', 'Daniel Gordon'] | 2019-12-03 | alfred-a-benchmark-for-interpreting-grounded-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Shridhar_ALFRED_A_Benchmark_for_Interpreting_Grounded_Instructions_for_Everyday_Tasks_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Shridhar_ALFRED_A_Benchmark_for_Interpreting_Grounded_Instructions_for_Everyday_Tasks_CVPR_2020_paper.pdf | cvpr-2020-6 | ['natural-language-visual-grounding'] | ['reasoning'] | [ 7.37450495e-02 1.05115607e-01 -6.15203707e-03 -4.73660499e-01
-4.09336418e-01 -8.04326236e-01 7.61051118e-01 -1.36359587e-01
-6.08577013e-01 6.58636808e-01 8.48707080e-01 -7.17251182e-01
3.65765333e-01 -3.57689679e-01 -9.46911454e-01 -4.71825123e-01
-1.57555804e-01 2.86828071e-01 -1.00389570e-01 -3.80580097e-01
1.77793801e-01 2.28940576e-01 -1.48559833e+00 7.10455656e-01
5.09990931e-01 3.48238826e-01 6.66031897e-01 1.10259628e+00
-3.35111879e-02 1.68223858e+00 -3.61420155e-01 9.76910517e-02
1.18407793e-01 -5.15420496e-01 -1.17446423e+00 2.96773072e-02
3.62954080e-01 -8.04772496e-01 -5.37097037e-01 7.68066108e-01
4.92084175e-01 5.32247126e-01 5.97878456e-01 -1.54630649e+00
-1.22183025e+00 6.94424212e-01 -3.60803545e-01 -1.75231807e-02
8.81147087e-01 1.00696695e+00 8.92114699e-01 -5.82255185e-01
9.37721968e-01 1.60431564e+00 4.00524855e-01 8.89421821e-01
-1.35884130e+00 -3.79359156e-01 6.20480120e-01 2.15730086e-01
-7.15755582e-01 -6.34129465e-01 2.13428110e-01 -7.47073114e-01
1.62994254e+00 -4.90870839e-03 6.69073939e-01 1.76389110e+00
1.73819244e-01 1.05685413e+00 9.81343329e-01 -3.83214891e-01
4.49110538e-01 -3.32578838e-01 1.02251299e-01 1.01498616e+00
-3.14567573e-02 2.28347704e-01 -5.35658658e-01 1.71623230e-01
8.01444888e-01 -5.78232110e-02 -2.54243910e-01 -8.05362940e-01
-1.77143097e+00 5.23774028e-01 4.10270572e-01 6.49117678e-02
-4.13042665e-01 7.13290691e-01 6.48747146e-01 3.50532949e-01
-5.47696203e-02 7.02834368e-01 -3.63096803e-01 -4.78414953e-01
-1.85824081e-01 4.97793883e-01 8.12454224e-01 1.18270040e+00
7.44027615e-01 7.96123669e-02 -6.33112013e-01 3.10291678e-01
2.01480657e-01 5.50763369e-01 3.68800849e-01 -1.62604856e+00
5.07035732e-01 6.52846515e-01 5.99818945e-01 -3.98359001e-01
-6.30925834e-01 3.54866982e-01 -3.37837726e-01 7.50971735e-01
5.80271900e-01 -5.19737542e-01 -1.11983514e+00 1.86681545e+00
-2.33076233e-02 -3.55769284e-02 4.86913830e-01 1.00031209e+00
1.17823839e+00 7.47624397e-01 4.55167770e-01 2.26159289e-01
1.34636843e+00 -1.47166121e+00 -5.47388375e-01 -5.59292853e-01
1.09324288e+00 -3.40741396e-01 1.82642806e+00 4.63762730e-02
-1.00740123e+00 -7.07340479e-01 -7.46131301e-01 -5.90909004e-01
-3.60795349e-01 2.76750505e-01 7.90999055e-01 5.19648679e-02
-1.37387681e+00 1.92827940e-01 -1.03550029e+00 -7.14305282e-01
1.27251163e-01 -1.32730678e-02 -6.08075082e-01 -2.19571576e-01
-6.95870280e-01 9.75796521e-01 2.72273749e-01 -1.05164021e-01
-1.61479950e+00 -6.64560497e-01 -1.55122733e+00 -7.68205747e-02
3.08402598e-01 -8.53181660e-01 1.85181677e+00 -8.71878684e-01
-1.59214306e+00 1.02136576e+00 4.59015043e-03 -5.65369725e-01
6.85463548e-01 -4.20357108e-01 -8.63119438e-02 -1.33905619e-01
3.31164330e-01 1.18694878e+00 3.04708928e-01 -1.01487446e+00
-5.68495870e-01 -2.41419479e-01 7.48786032e-01 4.70452636e-01
3.45948905e-01 -1.36328325e-01 -1.71211541e-01 -1.59853533e-01
-5.41953743e-01 -1.05979550e+00 -2.57334769e-01 3.01204532e-01
-2.89011389e-01 -2.73195863e-01 6.80159688e-01 -5.50111651e-01
5.69887042e-01 -1.99724984e+00 2.57602453e-01 -5.76741397e-01
1.63913086e-01 1.49800237e-02 -5.99244714e-01 6.30047321e-01
-1.05584055e-01 -2.25648656e-01 1.51062876e-01 -3.24721575e-01
3.30253333e-01 1.95950657e-01 -3.03518623e-01 3.42245102e-01
-1.33476883e-01 1.33593452e+00 -1.43822610e+00 -1.34494022e-01
5.18958092e-01 2.98484206e-01 -6.19332790e-01 5.04725516e-01
-5.75069189e-01 5.35111845e-01 -1.91520080e-01 3.75250429e-01
1.79148301e-01 -2.55597383e-01 1.43434629e-01 -2.20136903e-02
-2.42550418e-01 4.71520126e-01 -7.46879637e-01 2.35105753e+00
-6.66687310e-01 9.50118959e-01 5.87234944e-02 -5.99366248e-01
4.57783580e-01 1.98555991e-01 1.98412523e-01 -1.01113284e+00
-2.02016681e-01 -3.78707021e-01 2.17798650e-02 -1.03453231e+00
4.83752906e-01 2.42575660e-01 -2.12899894e-01 6.59590840e-01
-1.36236265e-01 -3.38460743e-01 4.75655109e-01 3.47674906e-01
1.47992730e+00 9.42312002e-01 3.73896569e-01 -2.10406959e-01
9.19527709e-02 3.06938887e-01 1.40171111e-01 7.65011430e-01
-4.75332290e-01 3.09458017e-01 4.21255678e-01 -7.82488704e-01
-1.09435570e+00 -1.27434325e+00 8.31699312e-01 1.72167337e+00
6.01771325e-02 -4.45895463e-01 -8.43016803e-01 -4.51570958e-01
-1.24790408e-01 1.38996708e+00 -8.35429430e-01 -1.11075141e-01
-4.62812603e-01 3.87933962e-02 5.03562510e-01 9.07931685e-01
6.55970454e-01 -1.75756025e+00 -1.38500571e+00 2.74218637e-02
-2.59925246e-01 -1.14792836e+00 -6.76212728e-01 2.32193127e-01
-3.74683321e-01 -1.14500892e+00 -6.33146763e-01 -8.86744738e-01
6.84939802e-01 3.58722359e-01 1.54078102e+00 -3.66796374e-01
-3.04484278e-01 1.09497607e+00 -3.29779118e-01 -4.36526269e-01
-6.01406753e-01 -2.36239403e-01 1.96410306e-02 -5.63581467e-01
5.26975095e-01 -4.72636223e-01 -7.01177537e-01 5.06885275e-02
-2.69122005e-01 8.00237954e-01 4.78288740e-01 6.80190802e-01
1.22390829e-01 -8.59699726e-01 1.03429765e-01 -4.23418373e-01
7.73457229e-01 -2.21756741e-01 -5.36335111e-01 4.00627494e-01
-9.01829638e-03 3.57388526e-01 4.29049015e-01 -3.95341694e-01
-1.11444736e+00 2.62180507e-01 2.01012149e-01 -2.23355383e-01
-5.07637382e-01 -6.26924634e-02 4.15421836e-03 4.18757111e-01
1.00519001e+00 3.54492426e-01 -6.93521872e-02 -9.41018984e-02
9.74300385e-01 3.01678926e-01 9.37089801e-01 -8.91056061e-01
2.31670544e-01 3.85178596e-01 -4.07801002e-01 -7.20201850e-01
-4.95326817e-01 -2.96086490e-01 -5.56417763e-01 5.74931279e-02
1.26553702e+00 -1.16401660e+00 -1.30343270e+00 3.70736152e-01
-1.29019570e+00 -1.37302494e+00 -5.17888248e-01 4.72358108e-01
-1.25693130e+00 1.27054170e-01 -6.55614197e-01 -7.00513840e-01
-2.71891087e-01 -1.35296535e+00 1.23658967e+00 1.38418406e-01
-7.12629676e-01 -8.77309859e-01 2.45093599e-01 1.39959767e-01
5.62372468e-02 1.92053363e-01 7.84387469e-01 -1.63055792e-01
-4.93791878e-01 5.55571496e-01 -2.73523808e-01 1.47196546e-01
3.66443992e-01 -1.71258971e-01 -6.86677754e-01 -4.17154163e-01
-4.13993955e-01 -8.98000002e-01 6.66401386e-01 3.06997806e-01
9.02866781e-01 -3.60687017e-01 -4.02187020e-01 6.12421095e-01
9.64190602e-01 5.14728785e-01 7.24306524e-01 4.08933252e-01
7.61388719e-01 3.76421094e-01 4.91315961e-01 2.88873374e-01
8.24781239e-01 4.59087640e-01 5.18258274e-01 -2.16350421e-01
-3.23416561e-01 -4.88434821e-01 8.97756875e-01 1.21078342e-01
-7.73398280e-02 -3.13916594e-01 -1.01911843e+00 7.11547196e-01
-2.03602386e+00 -1.11079264e+00 2.43098214e-01 1.72175765e+00
6.30457342e-01 -1.84413165e-01 8.83271247e-02 -5.48365951e-01
2.31452465e-01 2.93310732e-01 -9.02087390e-01 -7.30802715e-01
2.11122438e-01 -2.94665813e-01 2.48932526e-01 6.31131530e-01
-1.00734115e+00 1.13202751e+00 6.68419456e+00 -9.14938673e-02
-9.38104272e-01 -1.25988334e-01 3.09924603e-01 -7.78230056e-02
-2.65656501e-01 -5.93835376e-02 -4.51956958e-01 3.05163860e-01
7.85959125e-01 -5.64259961e-02 6.95181489e-01 9.83419538e-01
4.62437093e-01 -2.56471157e-01 -1.77712429e+00 1.14082491e+00
-1.05626501e-01 -1.31060207e+00 -4.44644243e-02 -2.35182539e-01
5.77164710e-01 3.97682697e-01 -1.36331901e-01 7.01177061e-01
1.36377418e+00 -1.30310845e+00 8.70833158e-01 3.63385081e-01
8.82132828e-01 -2.15528458e-01 4.01774272e-02 4.86132294e-01
-1.01665401e+00 -2.30017066e-01 -1.50517389e-01 -7.06271589e-01
1.49552405e-01 -4.37499911e-01 -1.07168770e+00 5.43472581e-02
7.79958427e-01 7.65385091e-01 -3.16828191e-01 4.97513771e-01
-4.14339662e-01 1.33976474e-01 8.30304623e-02 -1.53387561e-01
6.96899831e-01 -2.28272989e-01 2.59651095e-01 1.31031322e+00
2.36020014e-01 1.09667502e-01 3.86803508e-01 9.49954629e-01
4.85439971e-02 -1.77209511e-01 -1.20835817e+00 -2.95059800e-01
1.23013481e-01 9.48090553e-01 -4.92056161e-01 -4.16469932e-01
-5.92439055e-01 1.32501161e+00 4.29439217e-01 7.97177017e-01
-8.35611641e-01 -1.34188429e-01 1.19319379e+00 8.61339085e-03
1.11944824e-01 -5.98131001e-01 1.19877450e-01 -1.08631635e+00
-2.13392049e-01 -1.18073046e+00 1.27627760e-01 -1.49538136e+00
-6.52353168e-01 2.84753412e-01 -1.59269385e-02 -9.37020779e-01
-6.61975265e-01 -8.09548080e-01 -5.40209830e-01 6.50637567e-01
-1.07348502e+00 -1.30806005e+00 -6.87389851e-01 6.28057897e-01
1.10770786e+00 -1.27765238e-01 1.07844329e+00 -3.51220191e-01
-1.34047881e-01 1.61371633e-01 -1.65984988e-01 2.12226510e-01
7.44848013e-01 -1.43373775e+00 1.19778347e+00 6.41670406e-01
9.25618261e-02 5.49234569e-01 8.13633978e-01 -3.89373511e-01
-1.54102957e+00 -8.41353357e-01 4.66174006e-01 -9.30590510e-01
4.90201592e-01 -6.84377789e-01 -5.47723830e-01 1.42598605e+00
7.16452718e-01 -1.83410235e-02 2.58176386e-01 -9.06479955e-02
-5.27287424e-01 2.42877752e-01 -8.76140952e-01 1.30793929e+00
1.56612015e+00 -7.63504088e-01 -8.64512742e-01 5.08205473e-01
1.03948534e+00 -5.76738954e-01 -1.78784251e-01 -1.25220616e-03
6.74166799e-01 -1.00637305e+00 1.05436742e+00 -1.07146800e+00
5.82504332e-01 -4.78214234e-01 -2.19580933e-01 -1.61381996e+00
-4.25634384e-01 -8.67536366e-01 -4.44450639e-02 6.79991722e-01
6.31398782e-02 -2.96930701e-01 5.36848068e-01 7.06245124e-01
-1.61734104e-01 -2.14776769e-01 -2.64308393e-01 -5.86018562e-01
-7.44242817e-02 -3.15526009e-01 4.08074141e-01 5.85714579e-01
2.75685906e-01 7.47041881e-01 -2.59371400e-01 -2.11969852e-01
2.77284354e-01 9.44039971e-02 1.14174831e+00 -6.10329747e-01
-2.59158731e-01 -2.89801449e-01 -3.99510600e-02 -1.39667583e+00
5.10782003e-01 -8.10589492e-01 3.70426744e-01 -2.05059814e+00
2.39813149e-01 1.21948145e-01 4.71958667e-02 8.55648935e-01
8.00264478e-02 -3.44088852e-01 3.58098000e-01 -2.09447831e-01
-9.15803432e-01 5.92810333e-01 1.32583177e+00 -4.18289214e-01
-2.97191828e-01 -4.44788963e-01 -7.35472798e-01 8.16949904e-01
4.67244655e-01 1.26698956e-01 -9.89304125e-01 -8.82779658e-01
1.63967490e-01 6.85442612e-03 5.72899699e-01 -9.83870506e-01
2.68477827e-01 -5.56184530e-01 3.11291665e-01 -3.55170369e-01
4.53557521e-01 -5.76215744e-01 -4.45300452e-02 5.62380672e-01
-6.70672238e-01 5.51938593e-01 3.74863833e-01 4.33299750e-01
3.83330584e-01 2.43547603e-01 3.83607835e-01 -6.60360336e-01
-1.55286562e+00 -9.42702740e-02 -5.99582314e-01 1.96247235e-01
1.20498419e+00 -2.20186055e-01 -7.17474341e-01 -8.36523890e-01
-8.19239318e-01 5.77209473e-01 7.88547516e-01 7.84410059e-01
5.58013082e-01 -1.07205462e+00 -6.88058376e-01 -4.38782610e-02
3.06712270e-01 -5.98953431e-03 2.11948290e-01 4.08676893e-01
-7.76297212e-01 5.12396216e-01 -6.08150482e-01 -5.15710652e-01
-1.18573558e+00 9.92711723e-01 4.24758911e-01 -6.01681843e-02
-9.18681383e-01 8.15424681e-01 9.38231766e-01 -7.33717859e-01
3.33008677e-01 -8.11479688e-01 -2.60700863e-02 -3.34427118e-01
7.33524799e-01 3.66683006e-01 -4.95514691e-01 -4.01013225e-01
-4.10175055e-01 3.73989314e-01 7.00315461e-02 -2.10700050e-01
1.20402110e+00 -3.05624247e-01 4.91705835e-02 5.83079636e-01
8.77892613e-01 -3.62991065e-01 -1.94333184e+00 -1.38611300e-03
-3.88334580e-02 2.89145987e-02 -4.12272841e-01 -9.95227993e-01
-4.17967767e-01 1.04863715e+00 6.70347214e-01 -3.67059708e-01
7.19069421e-01 -5.07016107e-02 4.37434673e-01 1.16318297e+00
5.42387784e-01 -1.06541026e+00 6.03860855e-01 8.01394761e-01
1.28190827e+00 -1.33795059e+00 -2.21112534e-01 2.06657529e-01
-1.04617918e+00 7.40757644e-01 9.90875959e-01 8.64365920e-02
4.12368178e-02 4.59409833e-01 3.49550992e-01 -2.25301951e-01
-1.02542055e+00 -2.42160156e-01 -2.02402323e-01 1.20413494e+00
4.79639292e-01 2.30521291e-01 3.22335631e-01 2.14568421e-01
-1.62513077e-01 2.99369156e-01 5.21208167e-01 9.32627678e-01
-2.98588574e-01 -5.13440490e-01 1.27048660e-02 -1.59438565e-01
1.99947536e-01 -1.24344364e-01 -4.06968325e-01 8.12205613e-01
-9.23842043e-02 9.16805267e-01 2.57386893e-01 -9.00226012e-02
4.19656992e-01 1.96782455e-01 7.54972279e-01 -8.14066350e-01
-3.71707082e-01 -6.13109887e-01 2.69059688e-01 -1.12492609e+00
-3.15354943e-01 -4.97628123e-01 -1.61222732e+00 -2.94109821e-01
6.45518005e-01 -3.02515596e-01 4.10342187e-01 8.10531855e-01
5.21975219e-01 7.02206552e-01 -2.15797156e-01 -1.13962209e+00
-5.43663561e-01 -9.26596642e-01 -8.97089839e-02 5.32994449e-01
6.00899041e-01 -4.11124915e-01 4.91987951e-02 3.47277671e-01] | [4.396026611328125, 0.7679980993270874] |
c0c45042-a295-4bdb-9ac4-b508fdf9f3f5 | back-to-optimization-diffusion-based-zero | 2307.03833 | null | https://arxiv.org/abs/2307.03833v1 | https://arxiv.org/pdf/2307.03833v1.pdf | Back to Optimization: Diffusion-based Zero-Shot 3D Human Pose Estimation | Learning-based methods have dominated the 3D human pose estimation (HPE) tasks with significantly better performance in most benchmarks than traditional optimization-based methods. Nonetheless, 3D HPE in the wild is still the biggest challenge of learning-based models, whether with 2D-3D lifting, image-to-3D, or diffusion-based methods, since the trained networks implicitly learn camera intrinsic parameters and domain-based 3D human pose distributions and estimate poses by statistical average. On the other hand, the optimization-based methods estimate results case-by-case, which can predict more diverse and sophisticated human poses in the wild. By combining the advantages of optimization-based and learning-based methods, we propose the Zero-shot Diffusion-based Optimization (ZeDO) pipeline for 3D HPE to solve the problem of cross-domain and in-the-wild 3D HPE. Our multi-hypothesis ZeDO achieves state-of-the-art (SOTA) performance on Human3.6M as minMPJPE $51.4$mm without training with any 2D-3D or image-3D pairs. Moreover, our single-hypothesis ZeDO achieves SOTA performance on 3DPW dataset with PA-MPJPE $42.6$mm on cross-dataset evaluation, which even outperforms learning-based methods trained on 3DPW. | ['Jenq-Neng Hwang', 'Cheng-Yen Yang', 'Wenhao Chai', 'Lei LI', 'Zhuoran Zhou', 'Zhongyu Jiang'] | 2023-07-07 | null | null | null | null | ['pose-estimation', '3d-human-pose-estimation', 'image-to-3d'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-5.20781994e-01 -8.47333223e-02 -1.53047979e-01 -5.82705699e-02
-1.17291093e+00 -7.51033872e-02 4.52999473e-02 -6.03393555e-01
-7.90478468e-01 3.67724001e-01 2.36864015e-01 3.37115705e-01
2.76886493e-01 -3.03159535e-01 -9.43821847e-01 -3.40437979e-01
-1.01170644e-01 1.17836738e+00 2.57953405e-01 -5.73378980e-01
-2.82156944e-01 3.63450766e-01 -1.28066611e+00 7.10195228e-02
2.77602941e-01 1.07795298e+00 9.92351323e-02 8.91077280e-01
4.30412978e-01 3.84597301e-01 -5.51215231e-01 -6.46434486e-01
7.72534132e-01 -9.31629762e-02 -4.29858208e-01 2.06695288e-01
1.12081671e+00 -5.50591052e-01 -6.74922764e-01 6.53418660e-01
1.26210880e+00 1.34020463e-01 4.84023899e-01 -1.33171940e+00
-2.47699738e-01 -2.86686063e-01 -9.44608450e-01 -1.96649581e-01
8.13401997e-01 8.12886119e-01 6.49266243e-01 -1.37224805e+00
9.69193161e-01 1.69059539e+00 1.11308527e+00 6.91842973e-01
-1.14728558e+00 -6.53995156e-01 9.44710821e-02 2.68122822e-01
-1.53698051e+00 -6.47510588e-02 6.75770998e-01 -5.45166612e-01
1.33184004e+00 -1.08427413e-01 1.17451024e+00 1.55994928e+00
4.05586779e-01 1.14813960e+00 9.84775066e-01 -1.33403078e-01
-3.17017734e-01 -3.48323613e-01 -3.60987812e-01 1.16355634e+00
1.27627924e-01 3.13282073e-01 -1.02923882e+00 2.33813643e-01
7.97800541e-01 -1.87303975e-01 -1.09137438e-01 -7.19548106e-01
-1.55336726e+00 5.19605637e-01 5.93533576e-01 -4.26078022e-01
-3.06240350e-01 3.19704890e-01 5.08075237e-01 -4.48142923e-02
5.19574225e-01 5.04917979e-01 -7.81484962e-01 -2.82394528e-01
-9.55348670e-01 9.32820499e-01 6.18966520e-01 1.23828685e+00
4.77384120e-01 9.99734923e-03 -1.38016686e-01 6.84245467e-01
2.12264746e-01 1.11582255e+00 1.72657147e-01 -1.14937794e+00
8.70336533e-01 3.89287144e-01 3.03552866e-01 -1.08261895e+00
-8.66936922e-01 -4.39230949e-01 -7.65013993e-01 2.25845978e-01
7.29959607e-01 -1.28561556e-01 -1.02869034e+00 1.69299209e+00
6.61783993e-01 -2.11350441e-01 -3.22611481e-01 1.36164725e+00
8.49341035e-01 5.52035630e-01 -1.17780551e-01 1.96884364e-01
1.11429477e+00 -1.55949032e+00 -3.79698128e-01 -5.84845245e-01
6.88881338e-01 -6.68188095e-01 1.16554773e+00 4.52022076e-01
-1.19489706e+00 -9.71528053e-01 -8.87537479e-01 -2.97931254e-01
-1.19719356e-01 2.04020977e-01 2.08837435e-01 4.35026050e-01
-7.71926105e-01 3.35028529e-01 -8.81236494e-01 -3.77732903e-01
2.64898121e-01 2.76974916e-01 -8.21851730e-01 -2.63722658e-01
-1.04056859e+00 1.32578051e+00 2.85377175e-01 2.75554448e-01
-1.09087634e+00 -9.27051604e-01 -1.05103827e+00 -6.00252926e-01
7.01473653e-01 -1.05501378e+00 1.04566061e+00 -1.70731515e-01
-1.32962918e+00 1.25901628e+00 8.74914676e-02 -5.22454917e-01
1.15248907e+00 -9.64347005e-01 -6.44779205e-02 5.51640168e-02
1.69498146e-01 1.18767858e+00 7.58441627e-01 -1.24477780e+00
-3.99599850e-01 -5.58395267e-01 -1.82856351e-01 4.20740485e-01
2.05468968e-01 -5.57542086e-01 -1.06702411e+00 -5.31422496e-01
9.61571932e-02 -1.39225185e+00 -2.18434513e-01 4.11980808e-01
-7.17935383e-01 -3.53299268e-02 6.82135582e-01 -9.06683028e-01
8.68777692e-01 -1.75448906e+00 7.08566606e-01 -2.05097571e-01
3.48443419e-01 1.99736640e-01 -1.93997338e-01 1.71210736e-01
2.07165062e-01 -4.68445748e-01 1.16838969e-01 -1.02912235e+00
4.06073868e-01 1.42520800e-01 3.26748967e-01 7.22931325e-01
1.80294394e-01 1.10801244e+00 -8.10936451e-01 -8.28381956e-01
6.04366183e-01 4.58538115e-01 -8.42032433e-01 4.79939848e-01
-2.35756531e-01 6.73142016e-01 -6.27272427e-02 8.60053718e-01
7.05249369e-01 -9.87369195e-02 3.03738490e-02 -8.17274690e-01
1.37660995e-01 -3.50177526e-01 -1.16534662e+00 2.42994809e+00
-4.86644059e-01 4.37324375e-01 -1.70771554e-01 -6.71359599e-01
6.14269555e-01 5.17325476e-04 7.63882101e-01 -5.89297593e-01
2.00722978e-01 3.36245954e-01 -2.12146491e-01 -4.56916153e-01
5.76032579e-01 1.47083759e-01 -4.89139289e-01 -8.75728428e-02
4.97124583e-01 -4.33255732e-01 4.54850160e-02 -3.40506509e-02
9.13677573e-01 7.60799229e-01 8.75422657e-02 2.04356620e-04
1.83884174e-01 2.01996759e-01 7.30029762e-01 5.77648640e-01
-5.84156215e-01 1.00059128e+00 1.38802320e-01 -6.95755124e-01
-1.36658370e+00 -1.31151772e+00 2.56338060e-01 8.57133329e-01
3.61532182e-01 -5.40056348e-01 -8.54526997e-01 -8.95044386e-01
2.43448019e-01 2.54999816e-01 -5.23293734e-01 -6.42870143e-02
-7.49597549e-01 -4.84232128e-01 6.19778275e-01 5.11443913e-01
8.81685078e-01 -5.27527034e-01 -7.41414666e-01 4.84408066e-02
-5.94602227e-01 -1.36674500e+00 -8.80618215e-01 -1.68042734e-01
-4.51644093e-01 -1.15044701e+00 -1.27671325e+00 -4.99969691e-01
3.18150252e-01 -4.79401909e-02 1.42000294e+00 -3.50613326e-01
-5.57598293e-01 6.80988491e-01 -2.26658612e-01 -3.25199455e-01
1.00462988e-01 1.02666371e-01 4.93260831e-01 -3.38217765e-01
3.87898207e-01 -1.91471159e-01 -8.77607822e-01 7.31809139e-01
-4.37456667e-02 5.64956740e-02 6.57094300e-01 8.61301601e-01
9.72319901e-01 -2.80395925e-01 -9.86320749e-02 -3.00836533e-01
3.85454297e-02 -1.67164765e-02 -3.83954793e-01 1.43983915e-01
-2.28982627e-01 4.21651267e-02 1.61240280e-01 -6.51773453e-01
-9.00530636e-01 4.27177697e-01 -4.06815380e-01 -1.00039828e+00
2.57029273e-02 1.23008769e-02 -2.27824464e-01 -1.53812200e-01
9.78792667e-01 -2.30799466e-02 2.67659016e-02 -5.02378881e-01
1.99485704e-01 9.16223153e-02 8.18814218e-01 -5.49396813e-01
9.62465942e-01 5.65224707e-01 1.44668356e-01 -6.36464417e-01
-1.25496793e+00 -6.11372590e-01 -8.86866689e-01 -6.71963990e-01
1.38952339e+00 -1.40436697e+00 -6.93127513e-01 8.54765415e-01
-1.19993377e+00 -6.34168267e-01 -2.39073679e-01 6.73953414e-01
-8.84208500e-01 2.32754484e-01 -5.11000097e-01 -7.31017351e-01
-2.37385437e-01 -1.42034352e+00 1.59442401e+00 -2.26138860e-01
-4.87453520e-01 -7.14060307e-01 1.28708884e-01 7.90330172e-01
-6.66317120e-02 5.63334584e-01 2.16975644e-01 -2.04792261e-01
-4.80891496e-01 -6.43912852e-01 6.50653988e-02 3.25104982e-01
-4.60860461e-01 -5.28626561e-01 -6.43028080e-01 -5.11777043e-01
-2.89554745e-01 -8.80702257e-01 6.27258718e-01 6.20974123e-01
8.50780547e-01 1.79750338e-01 -2.68483877e-01 8.73795211e-01
1.06095541e+00 -6.78827763e-01 5.43909192e-01 2.57943243e-01
1.06855798e+00 4.38308775e-01 1.02984524e+00 6.62739635e-01
7.56636500e-01 1.24173164e+00 4.36103582e-01 -2.56386995e-01
-5.29757738e-01 -7.04774797e-01 5.16224146e-01 6.49528444e-01
-3.98521274e-01 -2.00744178e-02 -8.57997596e-01 2.59885520e-01
-1.79100645e+00 -8.26380849e-01 -4.67432067e-02 2.14867830e+00
6.66148245e-01 4.29767817e-01 5.74307323e-01 -2.32915267e-01
5.10434628e-01 3.71209949e-01 -7.34239101e-01 2.94310391e-01
-1.14306264e-01 1.77839976e-02 7.41191626e-01 3.10833782e-01
-1.18723273e+00 9.16016161e-01 5.54664183e+00 1.01747954e+00
-7.16893256e-01 3.51344854e-01 3.53214145e-01 -6.17715776e-01
4.00485516e-01 -4.40654069e-01 -1.34977853e+00 2.60672987e-01
3.45621139e-01 5.96374810e-01 2.28993863e-01 1.06813776e+00
1.44262925e-01 -3.55924249e-01 -1.28087747e+00 1.72433174e+00
3.33923340e-01 -1.08154237e+00 -1.63832784e-01 1.14114389e-01
7.95945644e-01 1.89087078e-01 4.80573289e-02 4.81438637e-01
3.84690613e-02 -8.20925653e-01 1.13806319e+00 5.36528528e-01
9.10641432e-01 -6.28324926e-01 6.77991748e-01 6.55808628e-01
-1.31454980e+00 1.45874619e-01 -2.81799585e-01 1.63433149e-01
5.99064589e-01 5.08821189e-01 -4.45349962e-01 4.52609748e-01
1.17911005e+00 6.94718361e-01 -4.75952804e-01 8.62106800e-01
-2.01655954e-01 -1.00345686e-01 -4.74188268e-01 1.44507512e-01
1.51728511e-01 1.71641022e-01 7.08555877e-01 1.29661155e+00
3.35520357e-01 -1.65576667e-01 7.02534616e-01 5.22883654e-01
-7.15913028e-02 -3.43622237e-01 -4.11257505e-01 4.38413918e-01
8.26926082e-02 1.07263768e+00 -3.58077407e-01 -1.83425099e-01
-1.18929826e-01 1.33803630e+00 3.58514816e-01 2.56344192e-02
-1.27476251e+00 -2.43361965e-02 7.20267296e-01 4.33375120e-01
3.24756235e-01 -8.26001048e-01 1.82043269e-01 -1.31849420e+00
2.18956202e-01 -1.04398596e+00 3.12532246e-01 -8.31894696e-01
-1.39111757e+00 3.95061731e-01 2.89694369e-01 -1.40773773e+00
-2.23783612e-01 -9.12349641e-01 2.05695719e-01 6.09567106e-01
-1.13581467e+00 -1.38261127e+00 -5.13207793e-01 6.83728814e-01
8.11325848e-01 -6.38107136e-02 5.12181163e-01 4.26448047e-01
-4.37134743e-01 8.20271790e-01 -3.88900012e-01 3.56853634e-01
1.10743427e+00 -1.16055453e+00 6.18510365e-01 6.36921585e-01
1.47434369e-01 7.78565556e-02 7.18570650e-01 -7.62969375e-01
-1.82595265e+00 -9.77262914e-01 7.88164258e-01 -1.11840153e+00
2.27762148e-01 -7.12026894e-01 -3.12310040e-01 4.79191214e-01
-1.16764084e-01 2.65954256e-01 2.30920061e-01 2.08817139e-01
-5.09253263e-01 -4.96507324e-02 -1.04371512e+00 7.20323265e-01
1.71572649e+00 -3.99467379e-01 -4.31686699e-01 7.65236199e-01
7.72315562e-01 -1.20825839e+00 -1.16465783e+00 4.74540293e-01
8.20866764e-01 -1.08484077e+00 1.47404003e+00 -4.31040645e-01
4.67549443e-01 -3.07433605e-01 -4.49152976e-01 -1.26717675e+00
-2.11877808e-01 -5.29247642e-01 -3.32727939e-01 4.06628877e-01
3.58498544e-01 -8.32978487e-02 1.21388304e+00 4.82754767e-01
-1.43811360e-01 -8.86674225e-01 -9.82067287e-01 -1.11171675e+00
-9.03590918e-02 -7.18683004e-01 1.51101112e-01 3.85065526e-01
-4.73160714e-01 1.59747720e-01 -1.08829129e+00 4.69307527e-02
1.03654182e+00 -2.47858584e-01 1.62169206e+00 -6.32273793e-01
-6.13680422e-01 -1.06278799e-01 -5.84772766e-01 -1.60865414e+00
1.74703762e-01 -4.75414753e-01 3.62393379e-01 -1.28752398e+00
8.75006318e-02 -1.77449301e-01 3.64188343e-01 1.56798422e-01
-2.23055229e-01 4.43739265e-01 5.92492759e-01 1.62471294e-01
-9.19558883e-01 6.04214549e-01 1.55003846e+00 -2.52911538e-01
-5.31845465e-02 -1.91020995e-01 1.79985985e-01 8.24939728e-01
2.06153661e-01 -3.69404525e-01 -1.70466572e-01 -5.67325711e-01
8.37166682e-02 6.41361549e-02 9.10466135e-01 -1.46062887e+00
1.83655784e-01 -6.78068819e-03 7.52696991e-01 -1.04705751e+00
1.10313439e+00 -6.36408329e-01 1.82124645e-01 6.67195916e-01
-6.55314922e-02 3.21416318e-01 4.46598232e-02 6.96414888e-01
2.58561294e-03 3.67793918e-01 7.57336259e-01 -5.57205975e-01
-1.11992252e+00 7.38793194e-01 2.49900490e-01 7.87887871e-01
7.49446511e-01 -3.78181458e-01 1.50473714e-01 -4.46645170e-01
-7.11820900e-01 3.45085412e-01 3.77691418e-01 4.66550261e-01
7.67796159e-01 -1.42660499e+00 -8.84381890e-01 -7.04559162e-02
3.26387703e-01 4.09601659e-01 6.40456378e-01 1.08681870e+00
-7.04254925e-01 2.04608306e-01 -2.48570830e-01 -1.03622270e+00
-1.16937339e+00 3.78147811e-01 5.91278374e-01 -5.88008702e-01
-5.98810852e-01 1.26820183e+00 4.21220064e-02 -8.95346880e-01
4.24794704e-01 2.58839615e-02 5.81355333e-01 -2.01473370e-01
1.47789344e-01 6.23374701e-01 -4.84336801e-02 -9.24807131e-01
-4.55140889e-01 9.63088989e-01 2.15776876e-01 -2.31031045e-01
1.08133340e+00 9.65917110e-02 6.04430258e-01 3.57803077e-01
1.49710941e+00 -3.09674740e-01 -1.79643941e+00 -2.85759121e-01
-5.07339060e-01 -6.88000977e-01 -1.74938157e-01 -8.55875015e-01
-9.94750857e-01 9.67158139e-01 7.53458023e-01 -7.53314853e-01
7.65799046e-01 3.86522599e-02 1.20761299e+00 4.89264429e-01
7.64706612e-01 -1.74833488e+00 6.53560102e-01 4.53367859e-01
1.09362257e+00 -1.54538119e+00 2.54697055e-01 -3.07581335e-01
-8.78551006e-01 8.44661772e-01 1.05018938e+00 -1.99268371e-01
5.60989082e-01 5.03674559e-02 2.91525070e-02 -1.57182485e-01
-4.15517628e-01 -2.57518083e-01 6.50025129e-01 7.95872569e-01
1.69043809e-01 6.05172385e-03 2.08169699e-01 2.90302545e-01
-2.82335579e-01 -1.38640329e-01 -2.75468349e-01 9.06281471e-01
-1.59203649e-01 -8.57222319e-01 -6.25683308e-01 1.19730398e-01
8.05689469e-02 1.78667277e-01 -2.02606902e-01 1.26957488e+00
3.76174062e-01 5.76427162e-01 -2.16194630e-01 -9.14213896e-01
9.83348489e-01 -8.29848833e-03 9.79322493e-01 -3.62742335e-01
-5.65561295e-01 1.08005777e-01 2.67080307e-01 -1.16097283e+00
-4.41096187e-01 -6.72590137e-01 -8.94197285e-01 -6.24495447e-01
-2.68202603e-01 -4.88601476e-01 5.71014285e-01 1.01913309e+00
4.29309487e-01 2.70567715e-01 1.61975235e-01 -1.63237643e+00
-6.76410198e-01 -8.28290403e-01 -4.44495499e-01 6.69056654e-01
5.10607064e-02 -1.13651752e+00 -5.20967171e-02 -2.17636049e-01] | [6.944225311279297, -0.9441770315170288] |
4c5f15ac-7110-411c-a116-d0256faf3922 | domain-alignment-and-temporal-aggregation-for | 2211.12036 | null | https://arxiv.org/abs/2211.12036v2 | https://arxiv.org/pdf/2211.12036v2.pdf | Dual Prototype Attention for Unsupervised Video Object Segmentation | Unsupervised video object segmentation (VOS) aims to detect and segment the most salient object in videos. The primary techniques used in unsupervised VOS are 1) the collaboration of appearance and motion information and 2) temporal fusion between different frames. This paper proposes two novel prototype-based attention mechanisms, inter-modality attention (IMA) and inter-frame attention (IFA), to incorporate these techniques via dense propagation across different modalities and frames. IMA densely integrates context information from different modalities based on a mutual refinement. IFA injects global context of a video to the query frame, enabling a full utilization of useful properties from multiple frames. Experimental results on public benchmark datasets demonstrate that our proposed approach outperforms all existing methods by a substantial margin. The proposed two components are also thoroughly validated via ablative study. | ['Dogyoon Lee', 'Sangyoun Lee', 'Seunghoon Lee', 'Minhyeok Lee', 'Suhwan Cho'] | 2022-11-22 | null | null | null | null | ['video-object-segmentation', 'unsupervised-video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.46397272e-01 -1.85429230e-01 -4.85020518e-01 -1.07762896e-01
-7.04600275e-01 -3.37315947e-01 5.62488675e-01 -4.58818153e-02
-5.49435377e-01 5.28664768e-01 3.48385781e-01 2.66155690e-01
1.74713552e-01 -3.12443495e-01 -8.22619021e-01 -7.12215543e-01
1.20856613e-01 -1.06233478e-01 8.33675861e-01 1.15921974e-01
2.45319501e-01 2.96012729e-01 -1.61717153e+00 4.63177323e-01
6.89014137e-01 1.03908896e+00 6.36240244e-01 6.21967614e-01
-2.08059028e-01 1.10210180e+00 -1.63385317e-01 -1.25718981e-01
2.34743252e-01 -4.79947567e-01 -1.09847260e+00 3.78582418e-01
4.94790971e-01 -5.37685692e-01 -2.32647926e-01 1.04439569e+00
1.30677402e-01 5.11251330e-01 3.24353278e-01 -1.31973588e+00
-4.44189399e-01 5.85053861e-01 -9.30733681e-01 7.94705451e-01
5.35111606e-01 2.33141288e-01 8.89186502e-01 -1.05673456e+00
8.94938588e-01 1.24976921e+00 3.95057499e-01 4.49765742e-01
-8.02908182e-01 -3.81937265e-01 5.94545186e-01 5.57882726e-01
-1.32865357e+00 -4.49484676e-01 9.98567700e-01 -4.70724374e-01
8.65144968e-01 2.10320011e-01 7.74684906e-01 8.02438855e-01
1.06221456e-02 1.35230899e+00 8.77290845e-01 -3.38298947e-01
9.27992687e-02 -4.51983549e-02 1.34179652e-01 6.64697468e-01
-9.65983197e-02 -1.45631984e-01 -7.47688890e-01 -4.70003858e-02
8.74739945e-01 2.77502537e-01 -3.87385488e-01 -1.10197671e-01
-1.43611515e+00 5.45278668e-01 3.90646040e-01 4.52774704e-01
-6.73200965e-01 1.51332200e-01 2.58666754e-01 -5.40577322e-02
3.64902675e-01 3.98000069e-02 -2.53596991e-01 1.30662799e-01
-1.13921261e+00 -7.09965751e-02 2.44261071e-01 1.11733592e+00
6.60330594e-01 1.01444498e-01 -5.57305634e-01 5.67317605e-01
7.16259301e-01 3.87981355e-01 5.17812788e-01 -1.25275099e+00
2.64646292e-01 4.90870208e-01 1.87701598e-01 -9.61162686e-01
-1.53343379e-01 5.72666759e-03 -6.32682085e-01 5.38263433e-02
2.97415197e-01 -1.06885076e-01 -1.03075349e+00 1.80726826e+00
7.02753007e-01 7.86035836e-01 -1.10782191e-01 1.22297430e+00
1.20765948e+00 7.16150522e-01 6.50017142e-01 -4.65385526e-01
1.25218487e+00 -1.28897500e+00 -8.62480283e-01 -1.50390998e-01
-1.55974012e-02 -9.33540881e-01 4.45364565e-01 6.92711547e-02
-1.52964878e+00 -1.00977051e+00 -9.13153648e-01 -1.17287636e-01
-2.61732787e-01 -2.55661964e-01 5.08231163e-01 2.11553439e-01
-1.25836325e+00 2.77297974e-01 -7.08798587e-01 -4.78273720e-01
6.21426165e-01 4.01946753e-01 -4.10841823e-01 1.50570199e-02
-9.84123230e-01 3.81802410e-01 5.28117657e-01 8.93499777e-02
-1.04984212e+00 -5.04873395e-01 -9.46280122e-01 -1.60361081e-01
4.83055681e-01 -1.00609398e+00 1.02484691e+00 -1.55737877e+00
-1.26297438e+00 6.56339049e-01 -6.61134183e-01 -5.34180045e-01
3.53857636e-01 -4.75369930e-01 -1.95617676e-01 6.03281438e-01
1.99778020e-01 1.10268581e+00 1.13587916e+00 -1.34922230e+00
-9.72056866e-01 -4.06709462e-02 1.90065324e-01 4.17336166e-01
-6.85050562e-02 3.17400903e-01 -1.10466754e+00 -8.06574941e-01
8.82906094e-02 -6.40997648e-01 -1.43295079e-01 -2.31397018e-01
-3.44477534e-01 -3.49798054e-01 1.24214602e+00 -6.57891393e-01
1.18354607e+00 -1.88939691e+00 4.48954701e-01 -6.15362823e-02
3.37174147e-01 4.00118291e-01 -2.60731220e-01 5.13908528e-02
-6.72134245e-03 -1.13485932e-01 -1.60934329e-01 -4.68510002e-01
-3.94696206e-01 6.49624243e-02 -6.82739615e-02 3.89484018e-01
3.56930137e-01 1.17067313e+00 -1.13991261e+00 -1.09028685e+00
5.79178452e-01 5.12740791e-01 -5.11549234e-01 4.39697772e-01
-2.42468297e-01 8.17936242e-01 -4.50589627e-01 1.03792524e+00
4.64801222e-01 -3.92217308e-01 4.13548946e-03 -5.05430341e-01
-1.06321022e-01 -3.42785776e-01 -1.10246766e+00 1.94729805e+00
8.39027688e-02 6.25160575e-01 -1.39459604e-02 -7.72944689e-01
3.01975429e-01 3.88318509e-01 8.65448356e-01 -5.81361413e-01
3.60348016e-01 -2.14850768e-01 -3.33864838e-01 -8.69919777e-01
5.71907580e-01 3.09783578e-01 2.76982993e-01 1.18678816e-01
4.10852939e-01 5.69493473e-01 4.89059538e-01 4.20753539e-01
6.86357796e-01 4.71285343e-01 3.34661394e-01 -1.96556538e-01
9.35117781e-01 -2.71792382e-01 9.48792160e-01 8.88212264e-01
-6.57365024e-01 8.01104486e-01 -5.30631095e-02 -3.87929201e-01
-8.59981179e-01 -8.21429193e-01 1.58566982e-01 1.24670076e+00
8.53785515e-01 -2.95109242e-01 -7.41961241e-01 -9.33318555e-01
-1.79843768e-01 2.71784127e-01 -9.40266013e-01 2.95285374e-01
-5.39034665e-01 -3.45742643e-01 1.95425539e-03 8.33746910e-01
5.97731411e-01 -1.32178903e+00 -7.87401259e-01 1.13167159e-01
-6.12391949e-01 -1.34596133e+00 -7.62328863e-01 -4.11933810e-01
-8.01063359e-01 -1.16634369e+00 -9.66835856e-01 -8.98487091e-01
5.74834943e-01 8.04377139e-01 1.06089723e+00 2.77190477e-01
-1.04121558e-01 8.51503670e-01 -5.31062245e-01 -5.36106341e-02
-1.38221994e-01 -3.08764398e-01 -5.85607104e-02 5.50457716e-01
2.33213127e-01 -4.36834574e-01 -8.34320068e-01 3.41014206e-01
-8.76359344e-01 4.31679994e-01 7.23291337e-01 6.26569927e-01
9.31947827e-01 -2.98198521e-01 5.04242659e-01 -7.39825606e-01
-8.68116468e-02 -7.64875412e-01 -4.24334735e-01 4.42002743e-01
-1.90617636e-01 -2.77592033e-01 7.26388709e-04 -4.11600351e-01
-1.36509359e+00 2.30559140e-01 1.22136720e-01 -9.08828080e-01
-4.26351041e-01 2.60119736e-01 -1.34142146e-01 -1.42609775e-01
-2.74632201e-02 3.14538389e-01 -2.65565455e-01 -2.98198253e-01
5.32348216e-01 2.90627539e-01 6.81294560e-01 -3.98833662e-01
4.80419099e-01 6.69559121e-01 -4.04151827e-01 -8.33642185e-01
-6.83965445e-01 -9.38586473e-01 -7.72086918e-01 -7.56593168e-01
1.41216362e+00 -1.01977897e+00 -6.29527569e-01 4.65237558e-01
-1.22062194e+00 -1.87137201e-01 -3.62951785e-01 5.66264570e-01
-4.81693208e-01 6.43223524e-01 -6.39323354e-01 -8.40672851e-01
-4.19144422e-01 -1.25344384e+00 1.22195303e+00 8.63002658e-01
-4.22753356e-02 -9.53474641e-01 -1.42919719e-01 6.42842293e-01
2.35881150e-01 3.44921350e-01 7.90435299e-02 -4.05767202e-01
-1.00325191e+00 1.15643755e-01 -4.42086428e-01 1.03250735e-01
1.45910785e-01 1.55201450e-01 -1.07694650e+00 -2.10027352e-01
-1.50464162e-01 -8.15573111e-02 9.94944513e-01 8.20310295e-01
9.65723813e-01 -7.85476565e-02 -3.47478688e-01 5.35794258e-01
1.38706172e+00 4.45668817e-01 6.11540854e-01 2.65086859e-01
1.14832699e+00 3.25112373e-01 8.62700760e-01 3.02050471e-01
3.32024693e-01 5.68611801e-01 5.71989000e-01 -2.15366781e-01
-4.58831042e-01 5.23986109e-02 4.89618242e-01 9.26606894e-01
-5.87972224e-01 -1.99010476e-01 -5.51481545e-01 8.11825812e-01
-2.20404053e+00 -1.34730697e+00 -6.49119318e-02 1.97230136e+00
5.03627241e-01 5.04005291e-02 2.66351581e-01 -2.30245501e-01
9.73495007e-01 2.60852873e-01 -6.46625578e-01 1.11857682e-01
-3.67889255e-01 -2.37481833e-01 3.37906897e-01 4.74769324e-01
-1.48348629e+00 8.59221697e-01 6.40126991e+00 8.32521737e-01
-6.54238522e-01 4.19699967e-01 7.73885906e-01 -2.36845519e-02
-9.37091410e-02 -7.13562313e-03 -6.31299376e-01 6.47868574e-01
4.50905025e-01 6.04463667e-02 1.75757706e-01 6.24234080e-01
2.34725907e-01 -4.52430576e-01 -8.64264727e-01 9.32556868e-01
2.21010998e-01 -1.54934585e+00 1.14616305e-01 -3.26192260e-01
1.20674074e+00 1.90982580e-01 -3.20743807e-02 9.17989314e-02
1.36803999e-03 -5.15732646e-01 1.06422067e+00 8.78431022e-01
3.84940356e-01 -5.70237160e-01 8.87780547e-01 -1.38111398e-01
-1.64921832e+00 -1.11217476e-01 9.66727585e-02 2.69443631e-01
4.41923082e-01 1.42397791e-01 -4.04427499e-01 8.80468965e-01
1.02858710e+00 1.12390316e+00 -6.10677838e-01 1.13802683e+00
-5.01976460e-02 6.63428545e-01 -2.18446776e-01 4.31176305e-01
4.33311969e-01 -1.41558036e-01 8.69702578e-01 1.48691320e+00
-8.85066092e-02 2.03793451e-01 4.03522670e-01 6.52268291e-01
4.24010828e-02 1.38776258e-01 -1.93828121e-01 1.72164261e-01
2.28021994e-01 1.31937921e+00 -1.05449879e+00 -6.98516369e-01
-8.37119102e-01 9.86234963e-01 1.09820105e-02 7.00200617e-01
-1.02502370e+00 4.06659804e-02 4.32406276e-01 -5.37664816e-02
4.88191664e-01 -2.42274776e-02 9.20689330e-02 -1.14319217e+00
-1.94973081e-01 -4.81940389e-01 7.92209625e-01 -1.04350686e+00
-1.07447910e+00 5.06512463e-01 9.32493880e-02 -1.35344493e+00
-6.01288863e-02 3.35499793e-02 -5.55358768e-01 6.82570100e-01
-1.72380590e+00 -1.36329484e+00 -5.15037358e-01 8.89445245e-01
1.17883873e+00 -3.73437218e-02 1.34668812e-01 4.51728076e-01
-6.77534997e-01 2.09941119e-01 -3.48082185e-01 1.95147559e-01
4.54704523e-01 -1.07442284e+00 -2.06757933e-02 1.24019849e+00
3.07910562e-01 4.12977278e-01 4.91384834e-01 -7.91956484e-01
-1.26509333e+00 -1.07984293e+00 4.83102292e-01 -2.81857789e-01
3.87724102e-01 2.71557212e-01 -1.00978470e+00 7.08681405e-01
7.64758945e-01 1.68539494e-01 5.37011027e-01 -3.48734647e-01
-6.17652535e-02 2.26944000e-01 -8.90170872e-01 3.33187044e-01
9.36732054e-01 -3.65397781e-01 -6.57142937e-01 7.50192925e-02
9.01764452e-01 -3.72488528e-01 -8.63670886e-01 5.98554373e-01
5.25801063e-01 -1.02952635e+00 1.10701692e+00 -6.13847494e-01
4.02494371e-01 -8.42891216e-01 -3.00120503e-01 -4.72804099e-01
-4.97511148e-01 -7.72434890e-01 -5.59803724e-01 1.45928264e+00
-1.36765316e-01 -1.07607074e-01 4.30469602e-01 5.41189790e-01
-6.19579293e-02 -6.61552727e-01 -7.27852225e-01 -3.59267503e-01
-5.83696902e-01 -4.69897240e-01 2.67347276e-01 1.08943665e+00
-2.52727032e-01 1.54362008e-01 -5.44769704e-01 3.56460929e-01
6.40789628e-01 3.19943607e-01 5.82779169e-01 -8.82461011e-01
-3.75260651e-01 -5.87947845e-01 -6.25694156e-01 -1.05274820e+00
7.86868203e-03 -6.20774686e-01 -3.54202790e-03 -1.44119704e+00
7.19908357e-01 -7.76251359e-03 -8.03586841e-01 4.20061469e-01
-6.19397044e-01 5.93204975e-01 4.02394176e-01 3.68894428e-01
-1.52035987e+00 4.94780183e-01 1.15083420e+00 -1.50547847e-01
-2.61359483e-01 -3.84030879e-01 -3.04469913e-01 1.08951271e+00
5.81267595e-01 -2.37768456e-01 -3.56678665e-01 -3.90497714e-01
-3.62520039e-01 8.83452445e-02 3.84763986e-01 -1.03372383e+00
4.67240691e-01 -3.50640506e-01 5.80963671e-01 -8.26706588e-01
2.43411198e-01 -8.05181921e-01 1.32066920e-01 1.80709124e-01
-1.97156444e-01 -1.49899721e-01 2.35503376e-01 1.03346193e+00
-3.74040484e-01 6.79782927e-02 7.86695540e-01 9.75302700e-03
-1.49975932e+00 4.28261071e-01 -3.19004208e-01 1.34194186e-02
1.21978223e+00 -2.74276882e-01 7.26542324e-02 -2.89447367e-01
-9.71114755e-01 3.86749208e-01 2.80173659e-01 6.92423761e-01
7.26047993e-01 -1.40942180e+00 -5.98865986e-01 1.20979086e-01
-5.82692437e-02 -6.58726469e-02 7.03998923e-01 1.21285617e+00
-2.46792629e-01 2.89117038e-01 -2.22643137e-01 -1.02538157e+00
-1.45181978e+00 7.49257922e-01 1.01829425e-01 1.05515249e-01
-6.32996380e-01 9.80153143e-01 6.65237546e-01 4.98174906e-01
2.54190922e-01 -1.92788303e-01 -5.58808327e-01 1.21267408e-01
8.71017337e-01 2.16968298e-01 -4.86153901e-01 -1.35788953e+00
-3.43826562e-01 7.86288202e-01 -1.21347822e-01 4.23964560e-02
1.03022206e+00 -6.98666155e-01 4.24477421e-02 4.22880083e-01
1.00684381e+00 -1.81991488e-01 -1.77715099e+00 -4.95190561e-01
-1.81466550e-01 -6.80703819e-01 1.08250059e-01 -5.09181559e-01
-1.51495504e+00 4.23946708e-01 5.73604703e-01 1.36318609e-01
1.52293599e+00 1.40358612e-01 8.70079160e-01 -1.89624175e-01
2.28781253e-01 -1.07450628e+00 1.56752869e-01 4.82121781e-02
6.75110400e-01 -1.49999475e+00 2.15804949e-02 -4.55660194e-01
-7.81105578e-01 8.53521585e-01 6.84386611e-01 4.59724329e-02
5.96778214e-01 -6.21040277e-02 -7.88449496e-02 -2.10292786e-02
-6.54437304e-01 -7.42220759e-01 6.68872714e-01 4.32026267e-01
2.10283697e-01 -4.21865404e-01 -2.24683508e-01 6.42334104e-01
8.88890505e-01 1.13648251e-01 1.24721117e-01 1.04426837e+00
-4.04560745e-01 -7.02744126e-01 -4.97019887e-01 8.97022858e-02
-6.07832551e-01 1.25594720e-01 -9.14954469e-02 6.48195982e-01
4.51475292e-01 1.08254933e+00 3.83294448e-02 -3.59938473e-01
-1.45428240e-01 1.88349430e-02 4.65339720e-01 -1.36748657e-01
-5.08961380e-01 5.92818499e-01 -3.87353837e-01 -9.89473403e-01
-1.34498537e+00 -9.10578668e-01 -1.26044738e+00 7.09801093e-02
-3.42140883e-01 -6.68443963e-02 2.66072422e-01 1.06000197e+00
4.05254364e-01 7.94413030e-01 4.80597466e-01 -1.32588983e+00
5.54661751e-01 -6.46010816e-01 -2.78725445e-01 7.92944729e-01
4.39831257e-01 -8.24059010e-01 -7.63832331e-02 7.07161427e-01] | [9.360130310058594, -0.21430765092372894] |
69c28bd9-ef8d-4bf7-8843-4fb58be5f910 | improving-sparse-representation-based | 1607.01059 | null | http://arxiv.org/abs/1607.01059v6 | http://arxiv.org/pdf/1607.01059v6.pdf | Improving Sparse Representation-Based Classification Using Local Principal Component Analysis | Sparse representation-based classification (SRC), proposed by Wright et al.,
seeks the sparsest decomposition of a test sample over the dictionary of
training samples, with classification to the most-contributing class. Because
it assumes test samples can be written as linear combinations of their
same-class training samples, the success of SRC depends on the size and
representativeness of the training set. Our proposed classification algorithm
enlarges the training set by using local principal component analysis to
approximate the basis vectors of the tangent hyperplane of the class manifold
at each training sample. The dictionary in SRC is replaced by a local
dictionary that adapts to the test sample and includes training samples and
their corresponding tangent basis vectors. We use a synthetic data set and
three face databases to demonstrate that this method can achieve higher
classification accuracy than SRC in cases of sparse sampling, nonlinear class
manifolds, and stringent dimension reduction. | ['Chelsea Weaver', 'Naoki Saito'] | 2016-07-04 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 1.63547292e-01 6.71392679e-02 -5.33031821e-01 -1.24951214e-01
-5.18061399e-01 -3.48029017e-01 4.03356463e-01 -4.63228583e-01
3.52398753e-01 6.32591844e-01 1.59535766e-01 1.60217330e-01
-2.53384292e-01 -6.50305808e-01 -3.05818707e-01 -9.24080610e-01
1.88080668e-02 7.08346367e-01 -1.31182104e-01 1.15635921e-03
1.58929229e-01 9.59689677e-01 -1.67572486e+00 3.89656246e-01
7.24922419e-01 9.42182720e-01 -1.47882178e-01 4.83899117e-01
3.14713977e-02 2.76229888e-01 -6.96212053e-01 -8.35076794e-02
5.68396032e-01 -6.24948919e-01 -5.62382340e-01 8.58692706e-01
8.23579609e-01 1.42840177e-01 -6.03651047e-01 1.14886653e+00
7.57471174e-02 2.11588293e-01 1.12273490e+00 -1.29120243e+00
-7.35567987e-01 4.06327425e-03 -5.17467499e-01 1.01565635e-02
2.23375082e-01 -4.60585713e-01 6.06155217e-01 -1.37779844e+00
7.65366316e-01 1.07817459e+00 6.03762269e-01 5.96349657e-01
-1.54454315e+00 -5.58039010e-01 -2.71377116e-01 2.00021639e-01
-1.79180384e+00 -7.32069552e-01 1.28625011e+00 -5.07536709e-01
5.88502586e-01 4.81154352e-01 7.99229264e-01 6.95859313e-01
9.21115801e-02 4.90842700e-01 8.03085268e-01 -5.55153787e-01
6.52527690e-01 3.01653415e-01 3.24859947e-01 6.31285608e-01
3.02093804e-01 -8.03320482e-02 -3.81031036e-01 -6.91345453e-01
8.21726978e-01 6.81707039e-02 -5.07377267e-01 -9.33510840e-01
-8.22438300e-01 1.04546440e+00 -6.89167902e-02 4.38927829e-01
-4.60892171e-01 -4.64149117e-01 -6.26849607e-02 3.30070972e-01
2.76067972e-01 2.83175200e-01 -1.62054434e-01 2.67230481e-01
-1.32200742e+00 -8.53932649e-02 1.04715228e+00 9.70867872e-01
9.81693447e-01 5.06427348e-01 2.82687455e-01 1.00724709e+00
1.83248043e-01 6.74181044e-01 7.58078158e-01 -1.06384671e+00
1.84761345e-01 6.70010805e-01 -2.86507875e-01 -1.32298172e+00
-7.50197172e-02 -6.54794812e-01 -9.30894315e-01 1.41684785e-01
2.63172150e-01 2.04670608e-01 -6.76228285e-01 1.31969690e+00
6.69195116e-01 3.66997361e-01 2.94763029e-01 7.95338035e-01
4.21154886e-01 5.93846917e-01 -5.26183367e-01 -6.10006332e-01
8.77017260e-01 -5.72181523e-01 -3.58974397e-01 6.01618886e-02
4.82418835e-01 -7.69476652e-01 6.98318303e-01 4.20129120e-01
-6.67318046e-01 -5.94799638e-01 -1.06391442e+00 3.98482352e-01
1.15417846e-01 5.01550436e-01 4.60097194e-01 6.18312657e-01
-8.41957450e-01 5.04512846e-01 -6.54620111e-01 -3.56808573e-01
3.06410909e-01 3.67801160e-01 -7.61304498e-01 -4.31353867e-01
-4.99526381e-01 4.35552150e-01 -3.02724708e-02 -2.89441403e-02
-7.88376570e-01 -6.27481222e-01 -9.17782724e-01 -4.66327332e-02
-2.09990755e-01 -2.91448683e-01 4.26772803e-01 -1.20448065e+00
-1.18599689e+00 7.44185388e-01 -5.53008080e-01 1.61575507e-02
-6.49137720e-02 4.09285814e-01 -6.37578547e-01 4.45171475e-01
2.02307895e-01 1.89754248e-01 1.61647427e+00 -1.23303246e+00
-3.28481704e-01 -5.66517472e-01 -4.87814814e-01 -8.66316110e-02
-6.69735909e-01 -4.49894994e-01 -2.33460248e-01 -6.12165391e-01
7.97218561e-01 -9.85836923e-01 -2.28503570e-02 -1.06534488e-01
-2.87060648e-01 -8.79720896e-02 1.35930908e+00 -4.52340573e-01
1.21214092e+00 -2.65181398e+00 5.96805573e-01 6.11128271e-01
1.50592178e-01 9.45998821e-03 -4.60680515e-01 3.43082547e-01
-5.42394102e-01 -3.61945033e-01 -3.62179190e-01 -1.68533653e-01
-5.00118673e-01 3.08410525e-01 -4.27327335e-01 1.05180275e+00
2.45029643e-01 2.45064154e-01 -6.25844061e-01 -3.82849991e-01
1.65104885e-02 6.19655550e-01 -6.06858611e-01 -2.97621861e-02
2.32834578e-01 2.52620816e-01 -4.89545316e-01 8.87242854e-01
8.20097327e-01 -1.78725377e-01 2.51825541e-01 -5.65257609e-01
3.34349096e-01 -3.36882651e-01 -1.60237932e+00 1.28448069e+00
-1.52276456e-01 6.88914537e-01 1.73615307e-01 -1.41371143e+00
1.32899690e+00 2.57353336e-01 8.46022666e-01 -1.50003940e-01
-2.45586529e-01 3.05931807e-01 3.47533859e-02 -2.92543292e-01
4.73258970e-03 -3.88155133e-02 3.53099555e-01 2.44043648e-01
3.38007987e-01 -2.53939200e-02 1.96197033e-01 5.62948436e-02
7.55249977e-01 -2.10165039e-01 4.55293089e-01 -7.07794011e-01
8.03480864e-01 -1.14175945e-03 6.85697138e-01 1.96648970e-01
-1.00924246e-01 6.94189966e-01 3.78524095e-01 -6.36490464e-01
-1.19606793e+00 -8.80502343e-01 -8.34576368e-01 4.85415101e-01
-1.24113366e-01 -3.56887430e-01 -7.03980446e-01 -9.22137201e-01
2.35243738e-01 4.95482534e-01 -7.43615389e-01 -4.46581542e-01
-5.67920208e-01 -6.51151419e-01 5.61593547e-02 7.48850554e-02
2.58136064e-01 -4.95154023e-01 -1.17129520e-01 7.77157843e-02
1.36067837e-01 -7.34898627e-01 -6.10072851e-01 -1.09802201e-01
-1.19554567e+00 -1.50928605e+00 -6.76461816e-01 -1.14761877e+00
1.38521719e+00 3.71526241e-01 5.92636943e-01 -9.11489278e-02
-2.56529242e-01 7.92172372e-01 -2.79615283e-01 3.07644665e-01
-5.06480932e-01 -2.13468894e-01 3.68188083e-01 6.88978910e-01
4.26513851e-01 -7.06892192e-01 -2.64074504e-01 5.30054033e-01
-6.79066420e-01 -3.25368613e-01 2.14300483e-01 1.17363632e+00
8.06143999e-01 2.62091756e-01 5.89991510e-01 -6.14643633e-01
2.21063271e-01 -6.83500648e-01 -5.07257104e-01 1.58046737e-01
-5.19768059e-01 -6.97070360e-02 7.49226034e-01 -8.06382895e-01
-4.50794190e-01 3.47894013e-01 3.04176807e-01 -9.92817461e-01
-2.60734279e-02 4.64169055e-01 -1.25768721e-01 -5.58021605e-01
9.97239351e-01 7.89187789e-01 3.09028745e-01 -5.09328246e-01
2.69276518e-02 7.08656549e-01 4.19257909e-01 -6.74279571e-01
9.61911976e-01 4.42025930e-01 4.75198835e-01 -1.42102551e+00
-5.09487092e-01 -5.07562220e-01 -9.10278141e-01 -1.94914594e-01
2.50471056e-01 -7.32131004e-01 -2.72320449e-01 -2.20417161e-03
-6.86870277e-01 2.09582284e-01 -8.39676261e-01 6.57180429e-01
-5.01469016e-01 5.16577184e-01 -1.84250966e-01 -6.65440619e-01
-5.88025562e-02 -9.35890913e-01 1.04052961e+00 3.04130535e-03
-2.06271306e-01 -9.77627993e-01 1.57862946e-01 1.38909250e-01
2.28580296e-01 2.66705632e-01 8.96378696e-01 -8.01476538e-01
-2.65014201e-01 -7.78953612e-01 5.20594597e-01 6.05775714e-01
4.39877063e-01 2.54695147e-01 -7.70352006e-01 -8.16380918e-01
6.14472687e-01 -6.96661994e-02 4.93834943e-01 3.76697421e-01
9.25282300e-01 -5.21251380e-01 -4.15372729e-01 8.69595766e-01
1.35137677e+00 2.22705320e-01 4.70303982e-01 2.52870694e-02
7.13529050e-01 5.21365941e-01 3.75124067e-01 3.75487894e-01
-8.66867229e-02 5.96543849e-01 9.45335627e-03 2.68927068e-01
-3.03627372e-01 7.05495756e-03 3.55430156e-01 1.04157639e+00
1.48779914e-01 5.46249330e-01 -7.71005094e-01 4.56660718e-01
-1.57575047e+00 -1.19831419e+00 7.51919076e-02 2.46580434e+00
4.12891656e-01 -3.80802453e-01 4.12950702e-02 6.79791212e-01
7.70190656e-01 -1.37951314e-01 -7.13376224e-01 3.41432095e-02
-3.87829393e-01 2.05152646e-01 -5.26464544e-02 5.20933151e-01
-9.86910045e-01 5.06508529e-01 7.28977966e+00 8.68331552e-01
-1.27788687e+00 -6.64957985e-02 5.59862077e-01 -6.30992278e-02
-1.94137558e-01 -1.39636546e-02 -9.61446166e-01 2.77893662e-01
7.44534552e-01 -3.72884572e-01 6.63244069e-01 1.09399998e+00
-2.04969168e-01 3.12654138e-01 -1.23240173e+00 1.10634220e+00
5.44322193e-01 -1.42190886e+00 4.15337741e-01 1.67231217e-01
1.11844218e+00 -2.49275461e-01 2.06584066e-01 1.64547145e-01
-4.14238483e-01 -7.25260794e-01 4.69713032e-01 4.03441340e-01
8.44783783e-01 -6.45579875e-01 3.00360292e-01 3.56840879e-01
-1.27017283e+00 -2.53928155e-01 -7.89550781e-01 2.49812990e-01
-6.36706173e-01 6.23750150e-01 -7.79093504e-01 1.40826091e-01
2.30246842e-01 1.12709272e+00 -4.93564695e-01 7.87027955e-01
2.69980699e-01 6.54838800e-01 -2.67205477e-01 3.03558797e-01
-8.56471509e-02 -8.71187747e-01 8.57586086e-01 8.21491003e-01
5.73722661e-01 3.89279634e-01 4.22203094e-01 6.52317762e-01
1.56430811e-01 3.04934621e-01 -7.79968977e-01 -1.22152828e-01
7.25071192e-01 1.30511999e+00 -4.78673905e-01 -4.32346135e-01
-4.11495119e-01 7.94743121e-01 2.30693996e-01 7.30679452e-01
-1.18337229e-01 -1.70764551e-01 8.41205239e-01 1.43127680e-01
5.46621859e-01 -3.09091777e-01 -2.76434183e-01 -1.37486565e+00
-2.10979283e-02 -1.16774726e+00 3.58088762e-01 -2.89301217e-01
-1.12623823e+00 6.76538646e-01 -1.73971713e-01 -1.67554784e+00
-3.41250807e-01 -5.76701343e-01 -4.67615217e-01 9.92456675e-01
-8.77816975e-01 -1.08552551e+00 -1.17717562e-02 9.69116569e-01
4.12810832e-01 -9.22983170e-01 1.07173526e+00 1.79289341e-01
-5.20642519e-01 5.92173755e-01 5.57825029e-01 1.00244716e-01
2.14511067e-01 -7.31199622e-01 -3.92471910e-01 7.07673311e-01
2.00704798e-01 7.45606184e-01 4.78201181e-01 -5.71116865e-01
-1.75829160e+00 -1.14861715e+00 7.04352558e-01 -2.50192195e-01
5.33873320e-01 -6.43673465e-02 -1.04056239e+00 6.65002286e-01
-3.83791834e-01 3.55591595e-01 1.01692235e+00 -4.78652716e-02
-4.76987451e-01 -4.35862452e-01 -1.45521128e+00 3.44606698e-01
6.78638279e-01 -7.62496531e-01 -4.20329273e-01 6.47028565e-01
1.27324864e-01 -1.52032152e-01 -1.19476521e+00 2.31153741e-01
4.37787652e-01 -6.33845806e-01 9.57518160e-01 -6.10201120e-01
4.22698222e-02 -5.04470766e-01 -6.61686659e-01 -1.18470502e+00
-6.55921340e-01 -3.11469227e-01 -4.17780966e-01 8.33171666e-01
2.99005732e-02 -6.48885667e-01 1.19971251e+00 3.95430833e-01
-1.21319510e-01 -7.96989024e-01 -1.34897459e+00 -8.74514520e-01
-9.16174427e-02 4.90363345e-05 5.19103408e-01 1.28405428e+00
1.00515790e-01 2.39632830e-01 -2.08497331e-01 3.45633745e-01
1.11962080e+00 5.80298364e-01 6.71055257e-01 -1.42256200e+00
-3.23512435e-01 -2.98093766e-01 -9.19868708e-01 -6.35552227e-01
5.09837806e-01 -1.19859886e+00 -5.07080793e-01 -9.21792448e-01
4.93179820e-03 -4.84918147e-01 1.79951027e-01 3.73886943e-01
2.93295473e-01 3.56819659e-01 -2.74421629e-02 7.77886629e-01
1.47440255e-01 5.97157836e-01 1.41627729e+00 -4.38775510e-01
-3.31909478e-01 9.08896774e-02 -4.53648061e-01 5.46357691e-01
4.78653044e-01 -4.39457297e-01 -4.50199246e-01 -3.01097576e-02
-3.89986306e-01 2.87338704e-01 -4.05025147e-02 -1.16796803e+00
1.88369274e-01 -3.23704153e-01 6.95808291e-01 -3.31026167e-01
5.23301482e-01 -1.03796875e+00 6.53905451e-01 5.36356688e-01
-1.70719564e-01 -2.32974872e-01 -4.25509363e-02 6.46805227e-01
-4.42660362e-01 -3.64391357e-01 1.08526242e+00 2.92323053e-01
-2.76154757e-01 5.71660221e-01 -1.52695596e-01 -3.35027188e-01
1.10772324e+00 -6.51126564e-01 1.00227229e-01 -2.00291634e-01
-9.55974042e-01 -2.74126172e-01 5.90935886e-01 -6.78096637e-02
9.60396945e-01 -1.86522067e+00 -8.27445269e-01 1.13648069e+00
1.51303515e-01 -1.86292842e-01 -1.18163563e-01 6.16656661e-01
-2.78879136e-01 3.40178847e-01 -2.43398845e-01 -8.51746321e-01
-1.12479031e+00 6.45120859e-01 5.44706404e-01 4.70536619e-01
-5.90175867e-01 5.10443330e-01 3.10485363e-02 -3.56001377e-01
-7.18908831e-02 9.94836017e-02 -3.92320901e-01 4.15001996e-02
5.51465631e-01 5.80576658e-01 3.04338224e-02 -1.19696867e+00
-3.19429219e-01 9.91018832e-01 1.62674651e-01 2.16744646e-01
1.25776017e+00 1.79244697e-01 -3.78466308e-01 3.22042346e-01
1.74814653e+00 2.13260308e-01 -9.35761988e-01 -3.62983912e-01
-2.66524076e-01 -7.98665762e-01 9.31777284e-02 -4.22222577e-02
-1.33666337e+00 4.15829867e-01 4.49242800e-01 1.29760876e-01
1.16970110e+00 -5.81350103e-02 1.76278353e-01 4.62577134e-01
4.41561580e-01 -6.86865449e-01 -7.60451565e-03 2.81559557e-01
1.14423716e+00 -9.40707803e-01 5.07366993e-02 -5.14298618e-01
-2.33220950e-01 1.55887759e+00 4.08235520e-01 -6.03416145e-01
1.07963896e+00 -8.56750309e-02 -2.40103230e-01 -1.91877708e-01
-6.43938363e-01 4.38531190e-01 7.98433363e-01 8.26978922e-01
1.71986014e-01 6.39966503e-02 -1.66018829e-01 1.75253183e-01
-1.97326183e-01 -2.83269614e-01 3.52033883e-01 6.49318635e-01
-5.00376761e-01 -8.41653407e-01 -8.73353422e-01 6.62866354e-01
2.29484081e-01 1.95068404e-01 -1.79617479e-01 6.29999459e-01
-9.09555852e-02 7.31213331e-01 2.34656274e-01 -4.67678279e-01
2.41727784e-01 1.19767793e-01 5.61376810e-01 -6.94932699e-01
1.55963242e-01 2.44876474e-01 -3.30991745e-01 -4.24581409e-01
-2.15817645e-01 -1.09569311e+00 -1.01100695e+00 -2.92942405e-01
-2.44648114e-01 6.15704000e-01 5.68400800e-01 7.04088807e-01
4.05889958e-01 -6.52145967e-02 1.22442973e+00 -8.50978851e-01
-8.45985472e-01 -7.50671029e-01 -1.11979866e+00 4.43684757e-01
5.10519266e-01 -5.85229993e-01 -7.58826971e-01 2.74518728e-01] | [12.441650390625, 0.4292293190956116] |
aab174fd-99ad-49b6-98aa-96cae0a6a485 | anomaly-detection-based-unknown-face | 2007.05856 | null | https://arxiv.org/abs/2007.05856v1 | https://arxiv.org/pdf/2007.05856v1.pdf | Anomaly Detection-Based Unknown Face Presentation Attack Detection | Anomaly detection-based spoof attack detection is a recent development in face Presentation Attack Detection (fPAD), where a spoof detector is learned using only non-attacked images of users. These detectors are of practical importance as they are shown to generalize well to new attack types. In this paper, we present a deep-learning solution for anomaly detection-based spoof attack detection where both classifier and feature representations are learned together end-to-end. First, we introduce a pseudo-negative class during training in the absence of attacked images. The pseudo-negative class is modeled using a Gaussian distribution whose mean is calculated by a weighted running mean. Secondly, we use pairwise confusion loss to further regularize the training process. The proposed approach benefits from the representation learning power of the CNNs and learns better features for fPAD task as shown in our ablation study. We perform extensive experiments on four publicly available datasets: Replay-Attack, Rose-Youtu, OULU-NPU and Spoof in Wild to show the effectiveness of the proposed approach over the previous methods. Code is available at: \url{https://github.com/yashasvi97/IJCB2020_anomaly} | ['Pramuditha Perera', 'Vishal M. Patel', 'Poojan Oza', 'Yashasvi Baweja'] | 2020-07-11 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 2.39994869e-01 -1.73914582e-01 2.58257892e-02 -2.87137717e-01
-1.58937737e-01 -6.47319317e-01 7.61595488e-01 2.05652729e-01
-1.75404698e-01 2.16225922e-01 -1.93694204e-01 -4.57883626e-01
2.93522686e-01 -5.82772851e-01 -7.30089068e-01 -5.23063123e-01
-6.24297023e-01 1.83752254e-01 1.84552878e-01 -3.47227842e-01
8.93058181e-02 7.21213400e-01 -1.34330416e+00 5.47955871e-01
5.22512913e-01 1.19816029e+00 -6.60462976e-01 1.01277959e+00
2.78302670e-01 5.34183323e-01 -8.58338356e-01 -9.00350213e-01
5.49841046e-01 -2.72435427e-01 -7.00318873e-01 -5.19932099e-02
7.26761699e-01 -5.99008799e-01 -5.83857059e-01 1.22577691e+00
6.62695110e-01 -1.42319277e-01 5.24321556e-01 -1.72050405e+00
-4.94151682e-01 1.34468228e-01 -8.15508902e-01 6.64599121e-01
6.34857297e-01 3.05341035e-01 7.12310791e-01 -8.67767453e-01
4.88118589e-01 1.60787034e+00 5.89817405e-01 8.75042617e-01
-8.87422383e-01 -1.25779271e+00 -1.18724868e-01 4.04208362e-01
-1.07344806e+00 -7.16826797e-01 1.02662373e+00 -3.09256822e-01
7.97212005e-01 1.76851526e-01 5.46457827e-01 1.90137422e+00
8.82865414e-02 9.63987470e-01 7.71723688e-01 -4.62375373e-01
-1.88018963e-01 7.93997645e-02 1.37236178e-01 8.82569134e-01
4.48553979e-01 4.12368119e-01 -3.58326346e-01 -6.71281397e-01
3.58169883e-01 3.23328972e-01 -2.01944247e-01 -3.35320026e-01
-6.37501001e-01 1.06728888e+00 4.68490541e-01 2.65285462e-01
-1.72690213e-01 1.32508233e-01 9.17665064e-01 7.14808702e-01
3.97856146e-01 2.58756459e-01 -2.17889190e-01 -6.75312756e-03
-8.28595698e-01 3.54087621e-01 8.62701416e-01 3.30675066e-01
3.87390643e-01 2.12069005e-01 -4.26985063e-02 7.37507522e-01
3.68257701e-01 4.56083328e-01 4.84937072e-01 -4.71807867e-01
4.13154215e-01 1.09555267e-01 -6.05557933e-02 -1.41474998e+00
-1.29433170e-01 -2.58386731e-01 -7.17636764e-01 2.71354169e-01
4.52096760e-01 -2.25494236e-01 -1.13673854e+00 1.69622970e+00
3.62289995e-01 7.26755142e-01 -5.55394478e-02 4.69962597e-01
4.12356019e-01 4.51833576e-01 -2.17569452e-02 -8.27755705e-02
1.13559747e+00 -8.80735099e-01 -7.54263222e-01 1.93942990e-02
7.15469539e-01 -8.98251832e-01 8.24339032e-01 5.71954608e-01
-6.64641857e-01 -2.25181699e-01 -1.07326734e+00 4.01223928e-01
-6.73968196e-01 -2.84643024e-01 5.10118544e-01 1.29805923e+00
-9.71702278e-01 7.05578327e-01 -7.47427464e-01 -4.48677450e-01
1.10313046e+00 5.02709627e-01 -6.06338620e-01 -1.49607202e-02
-1.36403155e+00 3.66033375e-01 3.38929862e-01 -1.32007971e-01
-1.42884719e+00 -3.06517333e-01 -8.52616787e-01 -4.85200658e-02
2.41042048e-01 -1.82358429e-01 1.07566774e+00 -1.18705618e+00
-1.30939639e+00 1.04387546e+00 -5.84838651e-02 -6.15994036e-01
5.29450417e-01 -3.02581310e-01 -7.86032677e-01 2.96530902e-01
-2.12499842e-01 2.44436905e-01 1.59398603e+00 -1.17553723e+00
-3.11063200e-01 -5.42855203e-01 7.24700913e-02 -3.35965723e-01
-6.73893571e-01 3.41281831e-01 -5.14719561e-02 -8.46640408e-01
-1.81261167e-01 -8.59237790e-01 1.94010422e-01 6.81036636e-02
-4.74390805e-01 -2.39253879e-01 1.51917315e+00 -8.18017125e-01
1.35073292e+00 -2.30846095e+00 -3.83396864e-01 4.37318474e-01
2.67060429e-01 1.13693380e+00 -3.18529516e-01 4.45491403e-01
-6.26500070e-01 2.71456599e-01 -3.63422722e-01 -5.59871852e-01
-1.17855422e-01 4.33718376e-02 -3.77145737e-01 8.41289818e-01
4.11850661e-01 5.61982810e-01 -9.60642338e-01 -2.03814253e-01
1.13714203e-01 6.00754976e-01 -8.07243526e-01 4.00537610e-01
1.45732939e-01 4.03225154e-01 -7.92156756e-02 1.05852175e+00
1.22204638e+00 2.13356838e-01 -1.96011230e-01 9.74034518e-02
6.31684244e-01 7.14725032e-02 -1.16270149e+00 1.40531123e+00
-2.79790819e-01 6.65171027e-01 2.94803530e-01 -1.17964280e+00
6.59347057e-01 5.47853410e-01 3.26354802e-01 -2.80224591e-01
4.69575107e-01 3.07488710e-01 3.29661429e-01 -5.80922842e-01
-1.03321694e-01 2.73218066e-01 2.76363343e-01 4.80541945e-01
4.78650510e-01 2.74762988e-01 9.50800553e-02 4.37688231e-01
1.22432768e+00 -3.27057719e-01 2.00948700e-01 -7.46088549e-02
8.85093629e-01 -7.69662380e-01 3.31518263e-01 8.90575111e-01
-7.84737945e-01 2.91913152e-01 7.65029371e-01 -6.59125268e-01
-9.23675954e-01 -1.04971015e+00 -3.30661625e-01 1.20277870e+00
-1.26925319e-01 -6.71419084e-01 -8.64400089e-01 -1.49053812e+00
2.24946842e-01 3.73944253e-01 -8.07949960e-01 -2.99747199e-01
-5.88985801e-01 -5.96511960e-01 1.02760136e+00 6.44053891e-02
5.21845222e-01 -1.17649126e+00 -2.28025541e-01 -2.01061279e-01
1.75610390e-02 -1.18293142e+00 -2.25329071e-01 -2.12631464e-01
-3.97386283e-01 -1.28473020e+00 -5.23808062e-01 -6.00280881e-01
6.15264177e-01 8.66806954e-02 6.65299654e-01 5.82522333e-01
-5.63755870e-01 4.79107052e-01 -2.76074767e-01 -7.41496444e-01
-5.27117670e-01 -3.19744319e-01 6.03667319e-01 5.33382118e-01
6.53906465e-01 -6.94267988e-01 -5.86390376e-01 1.20846026e-01
-1.04083920e+00 -7.97551274e-01 2.05399916e-01 9.34314787e-01
-1.04731336e-01 -1.74601078e-01 2.21960902e-01 -9.47317004e-01
7.16770411e-01 -7.56474972e-01 -2.61717349e-01 -7.90935829e-02
-2.46010706e-01 -2.92088509e-01 3.75827402e-01 -6.58975482e-01
-4.74705458e-01 -1.20052330e-01 -3.92775089e-01 -8.79847407e-01
-4.22967136e-01 -2.13255435e-01 -1.87545255e-01 -3.66089165e-01
8.64191294e-01 4.32563387e-02 2.52709508e-01 -2.70461440e-01
5.95138525e-04 8.97920787e-01 2.83828676e-01 -3.91693354e-01
1.22505224e+00 6.72103643e-01 4.18056920e-02 -1.11538506e+00
-3.90062124e-01 -5.45764863e-01 -5.64468026e-01 2.97943074e-02
4.56878781e-01 -6.09883189e-01 -8.85110259e-01 9.73101676e-01
-1.24020195e+00 4.41735983e-02 3.61257084e-02 1.90823272e-01
-1.74007788e-01 1.07473242e+00 -7.15311289e-01 -1.02859163e+00
-5.44471204e-01 -1.14521885e+00 8.45500469e-01 -1.99433863e-02
-6.61584921e-03 -1.06072319e+00 1.37414366e-01 1.21129015e-02
5.04159331e-01 5.74665189e-01 4.36494201e-01 -1.46347153e+00
1.71992615e-01 -6.99111283e-01 -2.13818565e-01 7.75689840e-01
2.13067323e-01 2.04084694e-01 -1.49403417e+00 -8.68738055e-01
9.19208750e-02 -4.70890641e-01 1.12506700e+00 2.44693905e-02
1.55628276e+00 -5.46878994e-01 -3.97080213e-01 8.72700036e-01
9.70801651e-01 5.78541942e-02 5.80684423e-01 7.02498257e-02
8.36879909e-01 5.85797489e-01 1.51328981e-01 5.72568476e-01
-3.13366383e-01 6.87988460e-01 9.48464036e-01 9.65086669e-02
4.71257865e-02 -1.71037793e-01 5.96678078e-01 1.51679805e-02
1.41603664e-01 -5.50557792e-01 -9.88758802e-01 3.00209284e-01
-1.43193734e+00 -1.27224040e+00 1.37785167e-01 2.22203922e+00
2.48513207e-01 1.42882660e-01 4.48370725e-01 5.88546395e-01
8.10172737e-01 4.75275934e-01 -2.11982071e-01 -8.63627017e-01
1.12672579e-02 4.75593567e-01 1.75862134e-01 6.71308935e-01
-1.69462514e+00 1.01543570e+00 5.16861200e+00 7.96877742e-01
-1.23083055e+00 2.39275739e-01 5.94232678e-01 -6.07514866e-02
2.83321381e-01 -2.70274043e-01 -6.93730831e-01 6.18494153e-01
1.07246971e+00 3.48732978e-01 3.74480158e-01 8.02289426e-01
-3.16722363e-01 4.66850877e-01 -9.45369840e-01 1.26267290e+00
4.63797152e-01 -1.01504993e+00 1.90376326e-01 2.33206540e-01
3.73217076e-01 -6.03094026e-02 5.63299596e-01 1.78163424e-01
2.62718081e-01 -1.07470679e+00 2.12461621e-01 -1.67912290e-01
6.73402786e-01 -8.18976939e-01 8.09808731e-01 -4.14055698e-02
-9.83015597e-01 -5.89321196e-01 -3.74248713e-01 1.75631449e-01
-1.03001289e-01 3.50370616e-01 -1.12227094e+00 3.77915949e-01
7.47458816e-01 7.44156599e-01 -6.92836821e-01 8.72414649e-01
-1.69674188e-01 8.55664790e-01 -4.85112727e-01 3.35021108e-01
2.57945299e-01 3.70465696e-01 9.89417613e-01 1.50647426e+00
2.58023143e-01 -3.26286495e-01 7.37361461e-02 3.91207963e-01
-1.41757950e-01 9.62184742e-02 -1.07387888e+00 -2.19438747e-01
2.82730907e-01 1.24637365e+00 -3.43441725e-01 -1.56262070e-01
-2.29147002e-01 1.20691526e+00 2.02749446e-01 2.58809209e-01
-7.51372099e-01 -6.25521243e-01 1.00880241e+00 2.71037221e-01
3.18948805e-01 1.10255256e-02 2.97834933e-01 -1.32798040e+00
-1.87622793e-02 -1.20686603e+00 7.80216396e-01 4.10301834e-02
-1.38626778e+00 7.39398360e-01 -1.35916039e-01 -1.09261072e+00
-5.01663148e-01 -1.03549778e+00 -9.64455664e-01 7.42390513e-01
-1.40409017e+00 -1.22058296e+00 -1.88408583e-01 9.54695821e-01
4.17468458e-01 -7.28082299e-01 9.19018209e-01 4.08192813e-01
-7.89118409e-01 1.20102668e+00 -2.11287141e-01 7.85474539e-01
8.13368618e-01 -9.95150387e-01 7.03631461e-01 1.02953565e+00
2.42604598e-01 6.39876485e-01 5.48266768e-01 -6.32270396e-01
-9.32397723e-01 -7.85475791e-01 4.78084564e-01 -6.46463633e-01
7.43585587e-01 -6.13043427e-01 -1.05466604e+00 5.79391003e-01
2.95060247e-01 8.09754014e-01 8.26681197e-01 -2.34546050e-01
-9.00184214e-01 7.38522112e-02 -1.68434536e+00 2.94873029e-01
8.89363348e-01 -6.02121830e-01 -3.33796293e-02 5.23375511e-01
3.62170517e-01 -2.58987755e-01 -2.69950509e-01 4.74784672e-01
6.49321795e-01 -1.19149077e+00 1.07464170e+00 -1.14717722e+00
9.36217904e-02 1.31778330e-01 -8.81713480e-02 -1.15720439e+00
-1.21594213e-01 -1.03521490e+00 -8.35986018e-01 9.91234422e-01
1.44564882e-01 -8.36185157e-01 7.71651268e-01 -2.09256038e-01
2.92529702e-01 -8.23133111e-01 -1.02240586e+00 -7.82006979e-01
-1.60861444e-02 -4.60686594e-01 5.80634296e-01 1.22565782e+00
-1.25421107e-01 -1.72613680e-01 -5.05854964e-01 4.51827526e-01
8.95472109e-01 -6.84183598e-01 8.57815921e-01 -1.35872889e+00
-3.47065061e-01 -2.61295617e-01 -9.99723852e-01 -6.48917019e-01
3.37406993e-01 -8.70097995e-01 -6.09882295e-01 -4.77111369e-01
-1.12513080e-01 -5.08867614e-02 -5.66954374e-01 4.14408505e-01
-1.40711159e-01 5.53274930e-01 1.80837482e-01 1.10542625e-01
-2.97860652e-01 1.92876279e-01 8.45620036e-01 -1.96106941e-01
1.86397936e-02 3.07209283e-01 -2.62254983e-01 8.32852900e-01
8.81502151e-01 -5.53471029e-01 -2.13867635e-01 -1.42938495e-01
-1.77004963e-01 -3.68335813e-01 4.41244543e-01 -1.14048564e+00
-2.08365664e-01 3.16298753e-01 6.11142218e-01 -1.93928257e-01
5.02875566e-01 -8.57915163e-01 -6.48894668e-01 9.70597804e-01
-1.70068443e-01 3.12152535e-01 1.83760077e-01 6.86084867e-01
9.66807678e-02 -1.96429998e-01 1.15117216e+00 1.27185956e-01
-3.02521259e-01 6.95256352e-01 -3.35263344e-03 -8.54677409e-02
1.19135332e+00 -2.04585403e-01 -4.00524050e-01 -5.64361811e-01
-8.41296732e-01 -1.75414115e-01 2.42248386e-01 5.79006612e-01
8.83784473e-01 -1.27499235e+00 -8.71863186e-01 9.39436376e-01
1.57489896e-01 -6.45232081e-01 6.61643669e-02 7.13457286e-01
-7.13561893e-01 1.86350003e-01 -5.09455800e-01 -6.43215060e-01
-1.58492649e+00 8.66450429e-01 3.71600717e-01 -2.17958808e-01
-3.10012847e-01 1.15508711e+00 3.30584720e-02 -3.69508088e-01
5.17463744e-01 3.50598037e-01 -2.72781253e-01 -4.46000658e-02
8.99378240e-01 3.47700566e-01 -3.05226948e-02 -9.76744652e-01
-4.43462640e-01 7.89735392e-02 -6.42383277e-01 1.01068556e-01
9.70800757e-01 2.48299778e-01 -4.36369777e-02 -1.59918025e-01
1.57652259e+00 1.83685496e-01 -9.06765521e-01 -2.78450876e-01
-6.02888204e-02 -1.02083218e+00 -3.11874390e-01 -3.74738187e-01
-1.23754537e+00 1.26550674e+00 1.26917160e+00 3.92318338e-01
9.58479345e-01 -1.99624985e-01 7.96169341e-01 2.78537780e-01
-1.90492585e-01 -5.45986950e-01 6.16233408e-01 4.83808309e-01
7.81180024e-01 -1.50376070e+00 -2.65507340e-01 -4.43642706e-01
-4.75108147e-01 1.08393276e+00 6.91855550e-01 -3.31958324e-01
1.02871001e+00 1.47422329e-02 -2.27231197e-02 -1.65454328e-01
-2.73533612e-01 2.26618014e-02 7.00203478e-02 1.04965496e+00
2.48826236e-01 -1.75571322e-01 9.73037481e-02 1.14079982e-01
-1.44720092e-01 -4.84600872e-01 3.54152679e-01 1.07621777e+00
-1.37930568e-02 -1.28322327e+00 -4.63636667e-01 3.91699910e-01
-1.07008624e+00 5.07723838e-02 -5.90112746e-01 4.29898024e-01
1.43400744e-01 9.92136955e-01 5.56207746e-02 -6.28067493e-01
1.74898021e-02 3.01863372e-01 3.66672426e-01 -5.35404146e-01
-5.58247447e-01 -3.19996327e-01 -1.76788762e-01 -7.81255484e-01
6.26861677e-03 -4.34833735e-01 -5.14266372e-01 -5.22835791e-01
-3.29477429e-01 1.50634587e-01 6.36915624e-01 7.33847499e-01
4.05369192e-01 1.33722529e-01 1.10200775e+00 -9.41605091e-01
-7.28237510e-01 -1.06707644e+00 -3.58553648e-01 8.49991262e-01
9.51384187e-01 -8.73276711e-01 -7.90951073e-01 -3.84065300e-01] | [13.065220832824707, 1.1554921865463257] |
99878aa2-8319-4b8e-a0a3-5fb3c783b1af | cooperative-trajectory-planning-in-uncertain | 2203.04452 | null | https://arxiv.org/abs/2203.04452v3 | https://arxiv.org/pdf/2203.04452v3.pdf | Cooperative Trajectory Planning in Uncertain Environments with Monte Carlo Tree Search and Risk Metrics | Automated vehicles require the ability to cooperate with humans for smooth integration into today's traffic. While the concept of cooperation is well known, developing a robust and efficient cooperative trajectory planning method is still a challenge. One aspect of this challenge is the uncertainty surrounding the state of the environment due to limited sensor accuracy. This uncertainty can be represented by a Partially Observable Markov Decision Process. Our work addresses this problem by extending an existing cooperative trajectory planning approach based on Monte Carlo Tree Search for continuous action spaces. It does so by explicitly modeling uncertainties in the form of a root belief state, from which start states for trees are sampled. After the trees have been constructed with Monte Carlo Tree Search, their results are aggregated into return distributions using kernel regression. We apply two risk metrics for the final selection, namely a Lower Confidence Bound and a Conditional Value at Risk. It can be demonstrated that the integration of risk metrics in the final selection policy consistently outperforms a baseline in uncertain environments, generating considerably safer trajectories. | ['J. Marius Zöllner', 'Karl Kurzer', 'Philipp Stegmaier'] | 2022-03-09 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [-1.87089347e-04 6.24276698e-01 -2.98923522e-01 -4.67265517e-01
-1.05295169e+00 -5.54458678e-01 9.92028356e-01 2.68771440e-01
-5.55072427e-01 1.15483665e+00 1.08964235e-01 -5.07053256e-01
-1.58996642e-01 -9.46816504e-01 -7.03893661e-01 -7.26813555e-01
-3.81319374e-01 7.83492208e-01 6.87114537e-01 -6.86684996e-02
7.80558540e-03 6.45538688e-01 -1.47925639e+00 -4.00261313e-01
1.11322784e+00 9.15695667e-01 1.12153158e-01 6.65574789e-01
1.62492275e-01 7.92866349e-01 -5.38973093e-01 -2.64613688e-01
1.90051287e-01 -1.28678888e-01 -4.84485805e-01 -1.87561214e-01
-3.64103913e-01 -4.35756564e-01 1.90040544e-02 9.46912527e-01
1.15153171e-01 5.03485680e-01 9.07987833e-01 -1.82703066e+00
4.97907341e-01 4.99037236e-01 6.28481805e-02 -2.20865905e-01
2.28394538e-01 2.92495698e-01 6.77060902e-01 -3.89744759e-01
6.48171723e-01 1.36158752e+00 6.09618068e-01 2.28862718e-01
-1.19838202e+00 -5.33532739e-01 1.80403262e-01 2.58680582e-01
-1.43692660e+00 -4.36289340e-01 3.12909096e-01 -6.34823620e-01
8.56721580e-01 -5.59299393e-03 5.03376067e-01 7.28778899e-01
7.83246577e-01 6.01509631e-01 9.91641402e-01 -7.12102950e-02
9.91007626e-01 1.75532326e-01 -1.36837453e-01 2.71536291e-01
5.34458458e-01 7.79491067e-01 -8.87918919e-02 -4.20431703e-01
1.65458992e-01 -2.01634720e-01 2.59261191e-01 -7.14680076e-01
-9.36556935e-01 9.26954925e-01 3.20964158e-01 -1.39155596e-01
-7.12706387e-01 6.03015780e-01 8.65234211e-02 2.08937470e-02
2.78068125e-01 1.17485188e-01 -8.22612867e-02 -4.25785422e-01
-8.25624764e-01 6.99819088e-01 9.59914088e-01 1.00053358e+00
9.15868342e-01 -9.99863967e-02 -3.20140809e-01 -3.27571891e-02
6.83843911e-01 6.47725880e-01 -5.75224578e-01 -1.38383102e+00
5.05309045e-01 2.96912789e-01 8.36420536e-01 -5.91777623e-01
-3.48372638e-01 -9.31774303e-02 -2.48785570e-01 8.45747292e-01
5.57970583e-01 -7.00228631e-01 -8.84664059e-01 1.62755442e+00
6.72247171e-01 2.65823007e-01 3.63820165e-01 5.23479879e-01
-1.03634410e-01 7.58783281e-01 3.02533209e-01 -1.71006560e-01
6.86945915e-01 -7.27164090e-01 -7.55903304e-01 -1.52029485e-01
6.18816078e-01 -3.51300716e-01 1.35903861e-02 3.53123218e-01
-8.85917723e-01 -1.50426507e-01 -1.10607588e+00 4.72929150e-01
-2.55056053e-01 -3.67953151e-01 4.77812499e-01 9.42811608e-01
-1.08873808e+00 5.90292752e-01 -1.18824136e+00 -3.20060313e-01
3.86864185e-01 2.98405379e-01 -2.22050369e-01 -1.95810497e-01
-1.10625017e+00 1.50880694e+00 3.75122100e-01 -5.81900217e-02
-1.23901534e+00 -4.09953326e-01 -8.08580101e-01 -2.16739580e-01
6.61883116e-01 -3.80368024e-01 1.40853786e+00 -7.50463977e-02
-1.65625226e+00 -1.60847440e-01 -1.19097009e-01 -8.31963420e-01
7.61471510e-01 -4.62870412e-02 -1.85139239e-01 -5.39178066e-02
2.13065699e-01 6.86317801e-01 6.63729727e-01 -1.45254397e+00
-9.78191197e-01 -5.92879616e-02 -2.47511603e-02 2.77715832e-01
5.79163790e-01 -1.18523538e-01 -1.38796106e-01 -8.85790437e-02
-2.00770557e-01 -1.37589824e+00 -9.73367989e-01 -1.52744636e-01
-1.29855469e-01 -3.50473374e-01 7.19471037e-01 -4.11623418e-01
1.06269956e+00 -1.67404842e+00 -1.27212539e-01 5.53166866e-01
-5.64838899e-03 -4.97678891e-02 1.38751775e-01 7.32590318e-01
5.28156161e-01 2.62231939e-02 -1.36387289e-01 -3.02119344e-01
2.27297351e-01 3.14643025e-01 -2.16170609e-01 5.35225153e-01
1.27477467e-01 7.14329720e-01 -1.06593788e+00 -3.53819847e-01
4.06669855e-01 2.64430434e-01 -4.31401700e-01 6.79139867e-02
-4.10895646e-01 4.01221335e-01 -7.96273947e-01 3.09640139e-01
6.21289790e-01 3.94027561e-01 1.07040510e-01 4.11165208e-01
-4.06654507e-01 4.76704985e-02 -1.23447645e+00 1.06737816e+00
-2.07233667e-01 3.70577365e-01 1.68270826e-01 -5.55417359e-01
8.79657686e-01 3.78445715e-01 6.42497241e-01 -9.58383754e-02
3.25680524e-01 1.94413736e-01 -2.04881281e-02 -3.46713141e-02
7.13091612e-01 -2.32051834e-01 -4.12607759e-01 4.22151476e-01
-2.87646979e-01 -6.26885533e-01 -1.24247216e-01 2.99700946e-01
1.37598515e+00 1.91671506e-01 1.82814553e-01 -1.33057088e-01
8.78609344e-02 3.72373283e-01 4.62387711e-01 9.91508186e-01
-6.74270630e-01 7.22860098e-02 5.68348885e-01 -6.06632233e-02
-9.28498328e-01 -1.10690844e+00 2.52572587e-04 4.48764443e-01
3.06992471e-01 -1.58831000e-01 -6.70876682e-01 -7.35745192e-01
2.78154969e-01 1.39139783e+00 -5.29416740e-01 -1.68961033e-01
-2.68364817e-01 -2.95042068e-01 2.80078709e-01 5.26007533e-01
3.15904915e-01 -5.93159497e-01 -9.09776151e-01 5.48062801e-01
5.15354052e-03 -1.02256370e+00 -1.21338226e-01 1.59285158e-01
-5.71604967e-01 -8.28529835e-01 -2.88989186e-01 2.46042728e-01
3.46897542e-01 1.09550089e-01 7.01268733e-01 -2.98055947e-01
1.11888252e-01 6.21292412e-01 -2.74163872e-01 -8.85227621e-01
-7.61513710e-01 -2.39219606e-01 1.37072191e-01 -1.08343251e-01
1.79653287e-01 -2.14916378e-01 -4.27604526e-01 4.99627918e-01
-5.49950659e-01 -2.67636180e-01 4.55369204e-01 4.24836665e-01
3.93852472e-01 5.32852650e-01 6.89856112e-01 -3.52098674e-01
6.19432747e-01 -8.54491889e-01 -9.34610367e-01 1.02908812e-01
-5.58338106e-01 2.57289618e-01 -2.63950136e-02 -1.09317802e-01
-1.38071787e+00 4.70245868e-01 6.43481314e-02 -1.77341655e-01
-2.91631222e-01 4.79239732e-01 -1.36988103e-01 6.22334704e-03
5.10039508e-01 -4.48423177e-01 2.88685083e-01 1.75073877e-01
6.01595879e-01 4.73195463e-01 1.80103615e-01 -6.75514162e-01
8.40866387e-01 5.78670800e-01 3.37062806e-01 -6.68169022e-01
-3.87478411e-01 -2.95050025e-01 -5.23827076e-01 -7.86763668e-01
1.01473069e+00 -8.75749826e-01 -8.09505761e-01 2.41631597e-01
-1.14901304e+00 -4.75955933e-01 -4.82936680e-01 8.56657147e-01
-8.65981698e-01 1.51875794e-01 1.05399273e-01 -1.53838551e+00
4.13395882e-01 -1.28449941e+00 9.56599236e-01 1.37849241e-01
-3.56191158e-01 -8.72627854e-01 1.62267610e-01 1.89414248e-01
4.61398900e-01 4.93806869e-01 5.40136039e-01 -5.31213880e-01
-1.14143634e+00 -6.43438101e-01 1.67742521e-01 -2.98915636e-02
-3.18802774e-01 1.73384752e-02 -6.15207613e-01 -2.09093347e-01
-1.79999277e-01 -1.17779240e-01 7.17109978e-01 6.22065663e-01
1.49583176e-01 -1.10605337e-01 -8.84026289e-01 -2.83171386e-01
1.15583241e+00 5.71245968e-01 5.57193816e-01 1.54817834e-01
5.89144118e-02 9.57114816e-01 1.24840617e+00 4.87266690e-01
7.49766171e-01 7.84261703e-01 6.33414268e-01 5.52780211e-01
3.70735139e-01 -3.67338806e-01 5.08318424e-01 -8.20801556e-02
-1.19939364e-01 -3.49108756e-01 -1.14521599e+00 5.64533114e-01
-2.25248432e+00 -1.14836895e+00 -1.00644581e-01 2.42150998e+00
3.11256886e-01 3.73431593e-01 2.87591577e-01 -2.88947523e-02
5.83646894e-01 -6.09903932e-02 -5.92284620e-01 -2.79797703e-01
4.95539218e-01 -4.32206601e-01 1.12014735e+00 9.35888886e-01
-9.70936954e-01 9.23996925e-01 7.04638386e+00 8.53331149e-01
-4.86892551e-01 1.00257158e-01 2.20055401e-01 8.41764957e-02
-3.01489502e-01 4.61007625e-01 -1.11815071e+00 3.59294862e-01
1.24131656e+00 -3.86567235e-01 1.89903960e-01 8.68146956e-01
5.81746757e-01 -8.49746764e-01 -9.38859403e-01 2.11173356e-01
-5.13589561e-01 -1.14593089e+00 -5.54501951e-01 3.73284072e-01
7.96480477e-01 2.65106171e-01 -1.63064778e-01 4.37752247e-01
1.36543238e+00 -1.09518027e+00 9.69706833e-01 7.90285349e-01
3.29624981e-01 -1.05196285e+00 6.89327538e-01 6.84568286e-01
-1.23477340e+00 -2.24279270e-01 -1.38822906e-02 -1.63058918e-02
7.46937931e-01 5.75185478e-01 -1.21299350e+00 6.38955414e-01
3.87420356e-01 2.14312807e-01 -9.86139476e-02 1.21796966e+00
-2.23637268e-01 6.79463089e-01 -6.39401972e-01 -2.66139805e-01
4.14400339e-01 -2.34597236e-01 9.76194263e-01 7.40397155e-01
4.30148572e-01 7.02665150e-02 4.77096856e-01 8.27311814e-01
7.38855541e-01 -5.08156776e-01 -9.38621342e-01 1.37032941e-01
9.38618660e-01 8.51731718e-01 -6.69860244e-01 -1.85311764e-01
-5.26142791e-02 2.26250082e-01 1.28633261e-01 4.82828945e-01
-9.84529674e-01 -2.36770213e-01 6.72474742e-01 -1.91843972e-01
3.33535939e-01 -5.73666573e-01 -3.45553786e-01 -5.00752807e-01
-2.52277583e-01 -3.05799186e-01 -5.22092637e-03 -5.88136911e-01
-7.61026382e-01 4.78331506e-01 6.76043570e-01 -1.17387807e+00
-8.51355791e-01 -1.14390038e-01 -6.30196214e-01 8.46774936e-01
-1.21025443e+00 -9.94322360e-01 5.13632074e-02 2.22991943e-01
2.17517734e-01 -5.05657196e-02 5.13618946e-01 -2.12962240e-01
-3.51571441e-01 8.58819261e-02 1.50448442e-01 -5.95275044e-01
2.18094945e-01 -1.07404578e+00 4.53954250e-01 8.63218069e-01
-5.26126206e-01 2.11427450e-01 1.22879076e+00 -1.25076818e+00
-1.20301521e+00 -1.25739479e+00 7.73453534e-01 -7.71587789e-01
6.82644188e-01 -1.30614817e-01 -5.65503836e-01 7.03296065e-01
-2.25938000e-02 -1.81992158e-01 1.00433424e-01 -1.54171690e-01
1.22578382e-01 2.10647076e-01 -1.38779008e+00 6.82655632e-01
6.96564198e-01 -3.21586020e-02 -2.18324929e-01 1.07224435e-02
7.65974581e-01 -1.68109044e-01 -8.10833931e-01 5.41957140e-01
4.96193171e-01 -8.16103697e-01 5.93534172e-01 -3.57654542e-01
-2.40913942e-01 -6.34267449e-01 -2.75771469e-01 -1.53855538e+00
-7.05486536e-02 -9.44187760e-01 3.97450104e-02 9.99010384e-01
6.13356829e-01 -7.80787766e-01 8.94295216e-01 1.20378041e+00
-1.90168560e-01 -6.62910938e-01 -1.46307588e+00 -1.09097588e+00
2.95602322e-01 -9.18242931e-01 7.52694190e-01 1.23745091e-01
-5.20082656e-03 -4.36157472e-02 -2.72096425e-01 3.80807906e-01
1.04209316e+00 -3.46906751e-01 9.92598891e-01 -1.29396200e+00
1.49357647e-01 -2.22201258e-01 -4.71839666e-01 -8.23853016e-01
2.81193197e-01 -3.57805818e-01 6.70001090e-01 -1.82533598e+00
-2.68447280e-01 -8.26632440e-01 2.68207461e-01 1.72478646e-01
1.31573573e-01 -6.42291427e-01 2.32150316e-01 -7.91637152e-02
-7.11099088e-01 8.59625399e-01 8.77296090e-01 1.67854384e-01
-2.15112954e-01 5.69715261e-01 -1.19798273e-01 7.66190350e-01
8.64717662e-01 -6.19778335e-01 -5.96588969e-01 2.12534994e-01
1.47198075e-02 6.65858507e-01 3.46098512e-01 -1.18418419e+00
6.39852464e-01 -7.44447947e-01 -2.81272739e-01 -1.12351418e+00
6.40001774e-01 -1.12104344e+00 7.11188257e-01 5.37980855e-01
-2.38182113e-01 -2.22435415e-01 8.75598863e-02 1.21426189e+00
1.03534140e-01 -3.18447024e-01 6.54401660e-01 2.09876314e-01
-6.00007057e-01 1.62579760e-01 -1.07289457e+00 -3.16532940e-01
1.78464329e+00 -4.70408559e-01 -1.11196019e-01 -8.27456772e-01
-7.74103642e-01 7.80556738e-01 4.19723988e-01 3.06335151e-01
4.09626693e-01 -1.27283382e+00 -5.96722007e-01 -2.24561751e-01
-1.89301874e-02 6.00336399e-03 1.27846003e-01 6.93643570e-01
-9.10376757e-02 6.08112633e-01 4.13426310e-02 -4.81306136e-01
-9.25714076e-01 3.18534881e-01 4.28304225e-01 -2.64275283e-01
-2.51389027e-01 4.25298214e-01 -2.76884884e-01 -4.72134054e-01
4.06674266e-01 -3.12175721e-01 -1.78761721e-01 -3.21687125e-02
3.08387130e-01 9.74419653e-01 -2.32885078e-01 -8.07678461e-01
-3.01246971e-01 1.31865919e-01 3.84609252e-01 -8.07907701e-01
9.43666220e-01 -4.56758201e-01 3.86553019e-01 3.51806492e-01
6.16605461e-01 -8.84304270e-02 -1.80102348e+00 7.39919990e-02
2.91339278e-01 -3.45310867e-01 9.22437534e-02 -8.35235059e-01
-4.47611809e-01 3.83140326e-01 4.20206398e-01 1.87234685e-01
2.68180817e-01 1.27279684e-02 5.72222769e-01 4.35183913e-01
1.00467384e+00 -1.16251123e+00 -4.32889640e-01 6.62108600e-01
7.40083158e-01 -1.23962247e+00 5.64656407e-02 -4.79488641e-01
-9.02707398e-01 5.74940801e-01 3.79801363e-01 2.20595989e-02
9.25078988e-01 4.98717040e-01 -1.04257099e-01 2.94162296e-02
-1.03171980e+00 -5.41846275e-01 -3.67046893e-02 9.55604911e-01
-3.40841979e-01 4.11240786e-01 -2.16898337e-01 3.92234713e-01
-1.32761270e-01 1.05800688e-01 6.04109943e-01 1.08019400e+00
-9.15049136e-01 -9.87934530e-01 -5.42010903e-01 4.03955311e-01
3.79684716e-02 6.10518038e-01 -4.70842309e-02 6.49504781e-01
6.12982064e-02 1.64318204e+00 -6.79968372e-02 -3.89712214e-01
3.01379740e-01 -7.99033493e-02 1.63573056e-01 -5.14271855e-01
-1.30071595e-01 -2.90125310e-01 6.92887425e-01 -6.85676336e-01
-2.48285219e-01 -1.18272913e+00 -1.31302071e+00 -2.52898663e-01
-5.22435427e-01 4.39049780e-01 9.16256368e-01 1.05478466e+00
2.97166675e-01 2.18890101e-01 5.58234692e-01 -1.19147611e+00
-9.52479959e-01 -5.27194321e-01 -4.19806033e-01 -3.27506721e-01
2.65533328e-01 -1.12602878e+00 -3.23340088e-01 -4.37898725e-01] | [5.037102699279785, 1.6427311897277832] |
0c424ea9-fdd5-4e52-bf49-f1424766dea6 | dual-domain-filters-based-texture-and | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Yang_Dual_Domain_Filters_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Yang_Dual_Domain_Filters_2015_CVPR_paper.pdf | Dual Domain Filters Based Texture and Structure Preserving Image Non-Blind Deconvolution | Image deconvolution continues to be an active research topic of recovering a sharp image, given a blurry one generated by a convolution. One of the most challenging problems in image deconvolution is how to preserve the fine scale texture structures while removing blur and noise. Various methods have been implemented in both spatial and transform domains, such as gradient based methods, nonlocal self-similarity methods, sparsity based methods. However, each domain has its advantages and shortcomings, which can be complemented by each other. In this work we propose a new approach for efficient image deconvolution based on dual domain filters. In the deblurring process, we offer a hybrid method that a novel rolling guidance filter is used to ensure proper texture/structure separation, and then in the transform domain, we use the short-time Fourier transform to recover the textures while removing noise with energy shrinkage. Our hybrid algorithm that is surprisingly easy to implement, and experimental results clearly show that the proposed algorithm outperforms many state-of-the-art deconvolution algorithms in terms of both quantitative measure and visual perception quality. | ['Yujing Guan', 'Hang Yang', 'Yan Niu', 'Ming Zhu', 'Zhongbo Zhang'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['image-deconvolution'] | ['computer-vision'] | [ 3.52582186e-01 -5.98822951e-01 5.14533341e-01 -5.11717089e-02
-2.94344127e-01 -4.03442591e-01 4.02523547e-01 -5.67448676e-01
-3.24182212e-01 9.74749565e-01 5.32519341e-01 1.78923339e-01
-5.15452266e-01 -5.11329174e-01 -4.41726834e-01 -1.07338357e+00
4.03360367e-01 -7.63404742e-02 4.18399185e-01 -2.65792936e-01
6.56451881e-01 3.44102681e-01 -1.66120148e+00 2.69284975e-02
1.18275130e+00 8.59777749e-01 5.98029256e-01 4.20896202e-01
-1.57039031e-01 7.89772451e-01 -6.55073579e-03 -3.08355521e-02
2.42560148e-01 -6.93773210e-01 -7.27981865e-01 2.48417541e-01
1.49622053e-01 -4.51483667e-01 -3.16061646e-01 1.51350105e+00
5.61983049e-01 3.17095309e-01 6.60209358e-01 -4.77297455e-01
-9.45259511e-01 3.91119644e-02 -7.99464524e-01 3.59917611e-01
3.07191342e-01 -1.53078539e-02 1.70529246e-01 -1.10248458e+00
4.36028332e-01 1.10193121e+00 7.80604243e-01 3.11163992e-01
-1.18622184e+00 -3.43231618e-01 -2.58910567e-01 5.05559206e-01
-1.08520627e+00 -6.73473716e-01 9.82425749e-01 -4.36664999e-01
3.50562125e-01 3.40466291e-01 3.60741019e-01 6.06027424e-01
2.71397799e-01 3.32335770e-01 1.80377543e+00 -5.86684346e-01
9.33238938e-02 5.89588732e-02 2.88134873e-01 5.26287794e-01
2.38470107e-01 1.23637229e-01 -3.38990837e-01 -1.35921299e-01
9.85152066e-01 2.83548057e-01 -9.34666336e-01 -1.38172865e-01
-1.12709868e+00 4.08228457e-01 8.70503113e-02 6.81723714e-01
-5.34506142e-01 -8.69789124e-02 5.58992252e-02 2.03502908e-01
7.64830768e-01 2.31264427e-01 -7.63623565e-02 -3.58318500e-02
-1.23285449e+00 3.28925192e-01 6.77043974e-01 4.49703932e-01
8.30898285e-01 -1.41232476e-01 -2.68382728e-01 1.01450312e+00
3.65999013e-01 5.02037525e-01 7.44551837e-01 -9.05441225e-01
-4.34083976e-02 3.30007374e-02 5.30316472e-01 -9.57587421e-01
2.00819839e-02 -4.23872739e-01 -1.04386163e+00 4.99274790e-01
4.60083812e-01 1.65844690e-02 -8.03148150e-01 1.35421360e+00
2.31765822e-01 5.87722242e-01 7.40450621e-02 1.40107703e+00
7.92014360e-01 6.49895549e-01 -4.47662652e-01 -5.26747823e-01
1.47331095e+00 -9.93775785e-01 -1.09720492e+00 -8.28263443e-03
-2.65370578e-01 -1.48533201e+00 5.09286880e-01 4.71854061e-01
-1.48242819e+00 -5.62838435e-01 -7.92741299e-01 -2.18153328e-01
-2.36708578e-02 2.19434291e-01 2.24269405e-01 4.90432799e-01
-1.18292868e+00 9.30576980e-01 -6.18965089e-01 -4.41086531e-01
3.39754000e-02 2.33809073e-02 -3.45997870e-01 -4.11169797e-01
-8.52288604e-01 1.01968789e+00 9.40933302e-02 3.78429413e-01
-6.73624575e-01 -3.89272511e-01 -5.03698587e-01 1.08629413e-01
2.54108995e-01 -8.04927766e-01 9.08017933e-01 -8.70280981e-01
-1.73990619e+00 6.82068586e-01 -4.65026498e-01 -2.25230023e-01
5.73404074e-01 -3.54277998e-01 -2.90482879e-01 2.99393147e-01
-7.84900971e-04 -1.38209507e-01 1.33641803e+00 -1.35007632e+00
-3.87056381e-01 -4.00597125e-01 -3.39174956e-01 3.60872298e-01
-1.38947174e-01 1.48303747e-01 -2.09624022e-01 -7.68794775e-01
5.21297336e-01 -5.91567576e-01 2.57808100e-02 7.56790042e-02
-1.80302963e-01 2.67659813e-01 9.59374845e-01 -1.21455503e+00
1.03249717e+00 -2.19552231e+00 4.37989354e-01 -2.23466590e-01
2.90673286e-01 4.50415552e-01 2.47388199e-01 3.33328575e-01
-1.02418423e-01 -3.62015247e-01 -5.51343322e-01 -3.73197228e-01
-4.42113519e-01 -2.05487795e-02 -2.84577221e-01 8.50560606e-01
-1.61079511e-01 4.78891164e-01 -7.23563135e-01 -1.64177984e-01
3.76545012e-01 6.94605768e-01 -2.75676399e-01 4.39985245e-01
3.31704706e-01 6.99835241e-01 -3.12609643e-01 3.54914397e-01
1.39536011e+00 -1.46391978e-02 -1.56980500e-01 -5.77882946e-01
-6.79752231e-01 -2.87369549e-01 -1.34174418e+00 1.56089723e+00
-2.00876758e-01 4.71638024e-01 7.34969854e-01 -1.27947116e+00
8.95383060e-01 4.28784341e-01 2.64675438e-01 -3.63693923e-01
5.18289544e-02 6.70559406e-01 -3.52393061e-01 -1.02214694e+00
4.76855218e-01 -5.75390220e-01 8.29673052e-01 2.90063381e-01
-7.60628656e-02 -6.42176196e-02 2.04410590e-02 -4.08354640e-01
7.96540320e-01 3.23913008e-01 1.72321558e-01 -5.34695148e-01
9.80473340e-01 -2.51370519e-01 3.18764061e-01 4.88531113e-01
-6.29820451e-02 1.11721337e+00 -2.99128448e-03 -3.10546011e-01
-1.08341897e+00 -7.58606076e-01 -1.60043001e-01 4.15541381e-01
5.70434093e-01 3.52704257e-01 -9.82850552e-01 -7.83745721e-02
-1.66388452e-01 2.29078472e-01 -3.71540785e-01 6.22621179e-02
-6.03936195e-01 -9.55202997e-01 1.45174623e-01 -1.02323696e-01
1.16222227e+00 -7.56056428e-01 -4.49772865e-01 3.59021246e-01
-4.77308840e-01 -1.01576829e+00 -6.33363605e-01 -1.67457476e-01
-1.08965147e+00 -1.11840081e+00 -1.47231841e+00 -1.07294655e+00
6.77560925e-01 9.80831027e-01 5.59169412e-01 4.87939008e-02
-2.11681440e-01 1.97228894e-01 -3.50173056e-01 2.57662892e-01
-1.63125813e-01 -6.04954898e-01 -1.55463651e-01 6.29766226e-01
-5.07577099e-02 -8.82381380e-01 -9.53819335e-01 4.16031539e-01
-9.79600012e-01 9.66176912e-02 8.58057976e-01 1.12542856e+00
3.43146861e-01 3.92709106e-01 2.15046689e-01 -6.22733414e-01
8.62301230e-01 -3.86099249e-01 -4.53647316e-01 9.86695662e-02
-6.16653562e-01 8.75706822e-02 5.74559033e-01 -1.98713303e-01
-1.71881676e+00 -2.36067489e-01 7.75471553e-02 -4.90211487e-01
-2.17538282e-01 3.96183848e-01 1.80440292e-01 -4.64103192e-01
5.37164032e-01 9.70888793e-01 4.82842565e-01 -1.11346066e+00
1.44061759e-01 7.05483794e-01 8.14397633e-01 -3.54149431e-01
8.03427279e-01 8.61670136e-01 -4.87417653e-02 -9.98035789e-01
-5.02398014e-01 -7.26955652e-01 -3.55299741e-01 -1.76949874e-01
1.02717459e+00 -7.59795904e-01 -6.01940989e-01 1.00607383e+00
-1.33349049e+00 1.50774866e-01 1.38120130e-01 7.30732441e-01
-4.02019143e-01 9.43098545e-01 -7.44356096e-01 -8.82737875e-01
-3.27980906e-01 -1.18076694e+00 6.84923410e-01 5.16785681e-01
4.67777818e-01 -9.61062610e-01 -3.42977904e-02 4.48404700e-01
9.31168020e-01 -2.95480043e-02 6.06856167e-01 1.55935049e-01
-6.12771094e-01 2.90948182e-01 -6.58161938e-01 6.90069795e-01
3.69918108e-01 -5.29348314e-01 -9.50996459e-01 -3.43718946e-01
1.00891387e+00 1.31570786e-01 1.01558971e+00 7.87463665e-01
1.01306844e+00 -3.22435498e-01 -1.74789742e-01 7.90849507e-01
1.80326855e+00 6.73075616e-02 9.76745725e-01 2.99342275e-01
3.70591104e-01 5.58223963e-01 4.34442699e-01 1.39014244e-01
-2.40103547e-02 5.72306573e-01 3.21762383e-01 -3.53319913e-01
-5.53280830e-01 2.88352311e-01 2.59796977e-01 7.72470593e-01
-6.06484115e-01 7.29139224e-02 -4.01656717e-01 6.04341745e-01
-1.98429930e+00 -1.21093059e+00 -5.51451504e-01 2.18716621e+00
8.48093569e-01 -3.26414138e-01 -3.53138179e-01 1.48137093e-01
1.03792107e+00 4.13890332e-02 -1.34156391e-01 8.31020027e-02
-2.86392212e-01 2.03222722e-01 5.71835518e-01 7.07598984e-01
-9.94333506e-01 5.18192172e-01 5.93724775e+00 8.64738047e-01
-1.18191254e+00 4.03505027e-01 4.44981247e-01 3.63270164e-01
-2.00552583e-01 1.28715575e-01 -2.67175525e-01 7.25998402e-01
3.64707440e-01 -1.19754449e-01 8.93635213e-01 2.49145538e-01
7.39687383e-01 -4.63564306e-01 -4.47801828e-01 1.31907630e+00
1.62209347e-01 -1.17475533e+00 -3.89383435e-01 -1.36185154e-01
7.25822926e-01 -3.10249656e-01 -1.88916419e-02 -3.75878185e-01
-2.96479195e-01 -9.17639971e-01 5.53046703e-01 1.00965595e+00
5.28519273e-01 -3.75859022e-01 9.20544147e-01 4.32262063e-01
-8.99701834e-01 1.96622938e-01 -4.03467178e-01 -2.88359672e-01
2.53178120e-01 1.07754815e+00 -7.64485449e-02 8.45371187e-01
8.61963570e-01 8.77525568e-01 -1.23239808e-01 1.46642113e+00
-3.77527811e-02 4.20195609e-01 1.72935590e-01 3.26001465e-01
1.06447048e-01 -9.22351360e-01 9.09253538e-01 1.02595925e+00
7.51780868e-01 4.24472094e-01 -2.18463615e-01 1.15208435e+00
3.74336272e-01 -6.64891750e-02 -4.09800678e-01 3.58271301e-01
8.61899555e-02 1.30292785e+00 -6.08451426e-01 -3.17123055e-01
-4.52272773e-01 1.39054334e+00 -2.66294956e-01 5.89604616e-01
-5.90240538e-01 -4.56405312e-01 4.20230836e-01 1.19840272e-01
4.07481521e-01 -2.92532772e-01 -2.22832665e-01 -1.44113946e+00
1.91071883e-01 -7.10903406e-01 -1.33833870e-01 -1.10736918e+00
-1.25209367e+00 4.84401137e-01 -1.67802036e-01 -1.35196638e+00
3.32579672e-01 -6.39503002e-01 -5.78684807e-01 1.38949907e+00
-2.07302666e+00 -9.43784833e-01 -6.34858608e-01 7.77802110e-01
6.12755060e-01 5.18376045e-02 3.86529446e-01 4.89094138e-01
-1.97169140e-01 -2.64958471e-01 6.57543182e-01 -3.68652344e-01
9.77581084e-01 -9.95795548e-01 -2.66036332e-01 1.09298825e+00
-5.73328614e-01 6.92196131e-01 1.07967925e+00 -5.90151906e-01
-1.40343189e+00 -6.37220562e-01 7.21300662e-01 3.34189355e-01
3.36235225e-01 2.01632559e-01 -1.11707914e+00 2.27469236e-01
4.94742364e-01 1.23235807e-01 2.07470119e-01 -4.58891392e-01
1.31512493e-01 -1.43001094e-01 -1.38857114e+00 2.85786927e-01
7.28623450e-01 -3.44363511e-01 -8.96532595e-01 2.36708745e-01
3.32134038e-01 -3.99268806e-01 -5.56650400e-01 1.98089957e-01
4.91663128e-01 -1.60759199e+00 9.28371787e-01 1.49819374e-01
4.99248803e-01 -8.29437912e-01 6.78849518e-02 -1.45841026e+00
-6.48950279e-01 -7.91460156e-01 8.05560127e-02 1.27688313e+00
-2.07588136e-01 -9.64613378e-01 4.05715942e-01 1.36210725e-01
-2.37140983e-01 -3.73420924e-01 -7.20863342e-01 -6.09473169e-01
-3.92348081e-01 2.23387539e-01 1.81666732e-01 8.78090382e-01
-3.14266711e-01 8.10699388e-02 -8.04696739e-01 1.96430117e-01
1.03621340e+00 2.44894281e-01 2.82688797e-01 -1.07372975e+00
-3.86197567e-01 -4.20820445e-01 -1.71968028e-01 -1.28517616e+00
-3.18523377e-01 -4.38732177e-01 1.75153196e-01 -1.54705000e+00
4.54977542e-01 -1.97655186e-01 -1.09283440e-01 -1.15673281e-01
-1.82645977e-01 2.93187261e-01 -1.23471662e-01 6.50609136e-01
5.69450483e-02 5.77862382e-01 1.53852057e+00 1.15930252e-01
1.39041618e-02 -3.26274149e-02 -6.14595115e-01 7.26676524e-01
3.91492367e-01 -1.94747463e-01 -3.09742123e-01 -6.07334852e-01
-3.29585612e-01 1.31245255e-01 5.94521344e-01 -9.79394734e-01
4.27064061e-01 -9.83095765e-02 2.43688270e-01 -2.79760301e-01
3.86876523e-01 -7.99240768e-01 3.16584677e-01 4.68348652e-01
1.38592601e-01 -3.30787301e-01 -5.34552075e-02 6.87730312e-01
-6.16735160e-01 -4.56668675e-01 1.20486796e+00 -4.78899330e-01
-5.98806620e-01 -7.17410073e-02 -2.83409655e-01 -2.90748417e-01
7.96564639e-01 -3.82517695e-01 -4.46737111e-01 -3.07857990e-01
-6.98436677e-01 -3.32254261e-01 5.03238440e-01 -1.25451130e-03
7.24638045e-01 -1.09682250e+00 -7.95826018e-01 1.05996661e-01
-4.87983912e-01 -3.19453776e-01 6.87133670e-01 1.48935950e+00
-7.95043707e-01 1.75204515e-01 -4.88622487e-01 -4.16430861e-01
-1.31399417e+00 4.39499319e-01 4.39057767e-01 -8.33068602e-03
-6.91864729e-01 8.09226334e-01 3.49771023e-01 1.72508433e-01
-1.04260609e-01 -2.49803573e-01 -4.05391484e-01 -2.16903046e-01
7.49342620e-01 6.85398877e-01 9.20278430e-02 -6.60213232e-01
-1.03227310e-01 1.10869527e+00 1.79092616e-01 -1.69244930e-01
1.56833529e+00 -6.58216357e-01 -8.80600095e-01 -2.79883891e-02
1.08262467e+00 2.29062855e-01 -1.24570405e+00 -3.21242243e-01
-3.31204683e-01 -8.18619788e-01 4.94382024e-01 -7.45885968e-01
-1.00862026e+00 7.47896671e-01 8.70680273e-01 4.07052308e-01
1.55581868e+00 -4.22426462e-01 8.84836137e-01 -3.17725092e-01
1.85472563e-01 -8.65266502e-01 -5.51055335e-02 3.09471995e-01
1.05144119e+00 -1.12684643e+00 4.91615124e-02 -7.75425613e-01
-1.49196699e-01 1.41947448e+00 1.39347047e-01 -3.82245839e-01
6.65432036e-01 -3.12974527e-02 -1.60613075e-01 -1.27351731e-01
-2.25926712e-01 -2.71410286e-01 2.75773376e-01 3.63520831e-01
5.49587309e-01 -3.05132240e-01 -9.61641252e-01 4.16121572e-01
3.81141454e-01 4.09880668e-01 7.01858640e-01 7.11353540e-01
-8.03953230e-01 -1.02470005e+00 -1.06358361e+00 1.09253548e-01
-6.43814087e-01 -1.80807352e-01 1.91522062e-01 1.83680192e-01
2.29101971e-01 1.03889692e+00 -3.18393201e-01 -3.98745909e-02
1.84279129e-01 -3.26550715e-02 6.43654585e-01 -2.51105964e-01
-2.05169678e-01 4.26697850e-01 -1.06328964e-01 -2.83136576e-01
-7.64521182e-01 -6.58489704e-01 -8.32669556e-01 -2.72876740e-01
-4.64577585e-01 2.42430195e-01 6.59767151e-01 9.57738936e-01
1.60269395e-01 3.95912379e-01 5.19043565e-01 -1.18692148e+00
-4.74863976e-01 -1.02119744e+00 -9.26878273e-01 4.92538273e-01
7.87958384e-01 -7.88754284e-01 -7.24330842e-01 2.52522439e-01] | [11.650370597839355, -2.7054152488708496] |
72d3854b-fa7a-4917-9358-6c0688ddd9be | grammatical-error-correction-are-we-there-yet | null | null | https://aclanthology.org/2022.coling-1.246 | https://aclanthology.org/2022.coling-1.246.pdf | Grammatical Error Correction: Are We There Yet? | There has been much recent progress in natural language processing, and grammatical error correction (GEC) is no exception. We found that state-of-the-art GEC systems (T5 and GECToR) outperform humans by a wide margin on the CoNLL-2014 test set, a benchmark GEC test corpus, as measured by the standard F0.5 evaluation metric. However, a careful examination of their outputs reveals that there are still classes of errors that they fail to correct. This suggests that creating new test data that more accurately measure the true performance of GEC systems constitutes important future work. | ['Hwee Tou Ng', 'Muhammad Reza Qorib'] | null | null | null | null | coling-2022-10 | ['grammatical-error-correction'] | ['natural-language-processing'] | [ 1.70140341e-02 1.02710582e-01 2.81975180e-01 -5.78161836e-01
-1.04072702e+00 -7.49503136e-01 6.11568987e-01 7.21052408e-01
-7.63098240e-01 6.60779178e-01 7.56158680e-02 -5.66789031e-01
2.45060399e-01 -4.96285647e-01 -7.20581174e-01 -1.58469811e-01
-1.90438390e-01 5.42847872e-01 2.10754260e-01 -5.00908911e-01
6.83158278e-01 -2.18541056e-01 -1.46025753e+00 3.79979163e-01
1.36403763e+00 2.47494146e-01 -1.24617599e-01 9.99255121e-01
9.84217077e-02 5.88607490e-01 -1.01542294e+00 -8.31248581e-01
-1.98304594e-01 -4.56422001e-01 -1.05187440e+00 -6.82371557e-01
8.03793311e-01 1.90862432e-01 3.57680589e-01 1.24781966e+00
4.92193043e-01 -2.43171304e-03 4.75961477e-01 -7.53581762e-01
-8.99233282e-01 6.69666648e-01 -2.30567213e-02 3.47792715e-01
5.34133792e-01 1.83703184e-01 1.11913061e+00 -9.49548662e-01
7.87242651e-01 1.34816885e+00 8.47577989e-01 7.79555559e-01
-9.36707497e-01 -5.28892994e-01 1.46613464e-01 -1.35122298e-03
-1.32743943e+00 -3.93436372e-01 -3.66641283e-02 -1.05433747e-01
1.79238164e+00 2.70971119e-01 3.06823879e-01 9.16365564e-01
5.10894120e-01 5.69636643e-01 1.01677144e+00 -9.29764032e-01
5.24090677e-02 -1.30267665e-01 5.17730713e-01 8.77237618e-01
6.18221402e-01 2.45294899e-01 -4.54628825e-01 -9.89171676e-03
-1.03302971e-01 -7.75626838e-01 -4.60872591e-01 5.98153353e-01
-9.66722548e-01 7.38727331e-01 1.51366740e-01 4.54936326e-01
4.62817177e-02 4.15030897e-01 6.25336349e-01 6.81871116e-01
5.76965868e-01 8.66485655e-01 -6.80471003e-01 -6.51133180e-01
-7.43818879e-01 5.68964362e-01 1.05282187e+00 1.06646025e+00
2.19491556e-01 1.06656618e-01 -3.33472416e-02 6.52324319e-01
3.51154178e-01 6.53611422e-01 5.65762162e-01 -8.37213635e-01
8.26331317e-01 6.65436387e-01 -7.42390454e-02 -8.56527805e-01
-2.07910076e-01 -5.37994266e-01 -3.57913733e-01 6.03069253e-02
6.50100172e-01 -2.38128453e-02 -9.30428028e-01 1.68424141e+00
-2.91922987e-01 -3.18786800e-01 3.32927710e-04 6.18306398e-01
6.41902268e-01 5.23255587e-01 6.08712792e-01 1.91271052e-01
9.89045560e-01 -5.67183793e-01 -7.42247581e-01 -5.05314052e-01
1.28681421e+00 -9.67053235e-01 1.31485701e+00 4.77396667e-01
-1.09257960e+00 -3.12422037e-01 -1.18760610e+00 -2.26830959e-01
-4.59489077e-01 2.38769408e-02 7.25823760e-01 8.50382805e-01
-1.03146017e+00 8.14633429e-01 -8.61810207e-01 -5.85292757e-01
-1.44806895e-02 4.72486615e-02 -2.99201936e-01 -3.16108465e-01
-1.24112403e+00 1.28487432e+00 4.44743633e-01 -3.34385075e-02
-5.78974187e-01 -6.05148554e-01 -9.46450174e-01 -1.19231112e-01
1.88002527e-01 -2.55791843e-01 1.44260693e+00 -5.59900284e-01
-1.00722384e+00 1.33185565e+00 -3.17329943e-01 -5.38351297e-01
2.94694960e-01 -4.82664317e-01 -7.56893456e-01 -2.44737521e-01
1.98354889e-02 4.56685126e-01 2.75770903e-01 -8.32074821e-01
-9.69265997e-01 -2.97900110e-01 -1.87373266e-01 1.89640224e-01
2.93855369e-01 6.42250717e-01 -2.01680809e-01 -6.47150397e-01
6.83714682e-03 -9.62118745e-01 1.73330843e-01 -5.33246100e-01
-1.75256535e-01 -7.35611796e-01 1.61427394e-01 -8.88597727e-01
1.70308399e+00 -2.09529781e+00 -1.42973810e-01 7.23575801e-02
2.09532350e-01 6.64865077e-01 -2.33620092e-01 3.06660563e-01
-8.30887929e-02 8.44395220e-01 -3.50689948e-01 -6.19847417e-01
-2.40106042e-02 1.50074214e-01 -5.73542491e-02 3.25364023e-01
2.21462339e-01 9.77011859e-01 -1.12867510e+00 -3.01220506e-01
-2.55398273e-01 6.16706982e-02 -6.59935653e-01 -2.54492611e-01
-1.34557188e-01 1.01937838e-02 -3.96945467e-03 7.98517287e-01
3.35672230e-01 -4.27011549e-02 -3.05260867e-02 6.22862697e-01
-2.32329413e-01 8.73997033e-01 -7.37090170e-01 1.54091704e+00
-1.07513309e-01 7.25724757e-01 -2.77310342e-01 -7.32493460e-01
7.71875799e-01 2.81615615e-01 -4.60133970e-01 -6.88732564e-01
1.04278840e-01 7.90899575e-01 6.17475688e-01 -2.31535986e-01
8.00876081e-01 -1.76378433e-02 -3.70239258e-01 4.28836733e-01
3.48348767e-01 -2.24272177e-01 5.07253468e-01 1.73050001e-01
1.33619523e+00 -3.67334932e-01 4.00839984e-01 -6.15036070e-01
3.22234929e-01 3.31097513e-01 8.91742527e-01 1.18284035e+00
-6.25795543e-01 8.56170058e-01 2.45900318e-01 -2.46031821e-01
-9.54644382e-01 -8.31172466e-01 -1.07131474e-01 9.33974683e-01
-3.08168888e-01 -8.80745471e-01 -1.10534012e+00 -9.69094932e-01
2.23321952e-02 1.22708678e+00 -4.56081778e-01 -1.13962613e-01
-7.30782509e-01 -6.89317524e-01 1.06340158e+00 5.39297879e-01
1.97059900e-01 -1.31035876e+00 -3.85601938e-01 3.03149790e-01
-2.46547699e-01 -9.83553290e-01 -4.64616299e-01 2.20740691e-01
-8.17205071e-01 -1.28551924e+00 -1.72247097e-01 -9.59849954e-01
5.77010512e-01 1.01787429e-02 1.62076128e+00 9.16306853e-01
-5.58882719e-03 2.62598634e-01 -7.63022661e-01 -6.06742144e-01
-9.13731575e-01 1.29628599e-01 -1.09895125e-01 -1.01955426e+00
8.61158907e-01 6.46245778e-02 -9.89146158e-02 -6.41737208e-02
-2.86852926e-01 -2.34121799e-01 2.30379269e-01 7.86152005e-01
3.76013607e-01 4.06460017e-02 4.84015584e-01 -1.35458028e+00
8.71850371e-01 -1.87711209e-01 -3.76925081e-01 6.04169905e-01
-1.00594831e+00 -1.25583380e-01 5.43936908e-01 8.69176313e-02
-8.72054577e-01 -6.36649847e-01 -3.98152232e-01 4.31410342e-01
-1.93825379e-01 7.84564853e-01 6.99476451e-02 -3.74692082e-02
8.54735315e-01 -8.70670006e-02 -3.52153450e-01 -4.44456488e-01
5.64172748e-04 5.49521387e-01 7.21639931e-01 -6.89409316e-01
3.94226372e-01 -3.92769665e-01 -4.50015604e-01 -5.19401371e-01
-9.80000854e-01 -1.38489559e-01 -2.19469815e-01 5.38956672e-02
6.15186036e-01 -8.17716837e-01 -4.38945264e-01 5.93721628e-01
-1.64136088e+00 -4.09483373e-01 6.61274567e-02 4.18433309e-01
-1.11450896e-01 1.64159805e-01 -8.29365969e-01 -5.19278884e-01
-6.97353780e-01 -1.03143430e+00 9.31993484e-01 7.68412510e-03
-6.46874607e-01 -1.03301048e+00 8.00361782e-02 2.57137507e-01
4.72057015e-01 7.73403719e-02 1.19388139e+00 -9.17304575e-01
-1.25592276e-01 -5.34501195e-01 9.52355340e-02 5.51900029e-01
-2.44495183e-01 2.43627444e-01 -4.83193427e-01 -3.58994752e-01
-7.28359446e-02 -2.15727612e-01 8.75001490e-01 1.32404089e-01
9.56401885e-01 -2.79425588e-02 -7.02805445e-02 3.59909534e-01
1.42926908e+00 1.95583761e-01 8.15939248e-01 2.86889076e-01
5.09693563e-01 4.68946218e-01 5.75680196e-01 -3.70634437e-01
6.22185647e-01 3.36370766e-01 5.76881543e-02 4.83232081e-01
-3.19462657e-01 -5.40985703e-01 4.17962879e-01 1.22803378e+00
-1.71172544e-01 -8.25506568e-01 -1.41681099e+00 6.92891359e-01
-1.61546445e+00 -7.47584224e-01 -5.86589992e-01 2.21819091e+00
8.58332574e-01 5.36110818e-01 -4.98857975e-01 3.08466524e-01
7.88808942e-01 -2.48697251e-01 -7.34774694e-02 -9.56655443e-01
-3.22798371e-01 6.66279733e-01 3.83842528e-01 6.05234981e-01
-9.16345119e-01 1.39041972e+00 8.00472832e+00 6.61588132e-01
-7.49731541e-01 1.93358902e-02 4.73202735e-01 2.89662570e-01
-3.11798036e-01 9.49001461e-02 -1.13823223e+00 5.12061894e-01
1.31177485e+00 -2.29729131e-01 4.32138175e-01 5.84502816e-01
-9.44535509e-02 -2.20280305e-01 -9.14551258e-01 7.18524039e-01
3.00462872e-01 -8.58007550e-01 -1.72005087e-01 -1.47623479e-01
8.27565253e-01 3.02862793e-01 -1.86486274e-01 9.22069788e-01
5.63613296e-01 -1.09660470e+00 7.12912858e-01 2.65964538e-01
7.23796368e-01 -6.39078796e-01 1.00799346e+00 3.58552665e-01
-6.05040669e-01 2.20194951e-01 -5.59843600e-01 -2.69430131e-01
7.37761408e-02 6.48129642e-01 -4.96673733e-01 3.14257234e-01
7.58667886e-01 5.54293573e-01 -1.19091094e+00 1.04934669e+00
-8.61497819e-01 1.30556536e+00 -1.26822665e-01 -2.93749809e-01
8.74913558e-02 2.26296052e-01 6.20566010e-01 1.40205693e+00
4.02936637e-01 1.23694129e-01 -1.73226193e-01 6.19201541e-01
-2.50717431e-01 1.23087667e-01 -4.28949773e-01 -2.37910822e-01
7.21261740e-01 7.46276796e-01 -4.30639863e-01 -4.61839437e-01
-4.50661510e-01 9.76339042e-01 7.25308716e-01 4.30121832e-02
-5.42951286e-01 -6.40689075e-01 5.37583113e-01 -1.38682991e-01
-9.00006965e-02 -3.35878909e-01 -4.05454159e-01 -1.22979927e+00
9.51935872e-02 -1.34128833e+00 3.48399371e-01 -4.78760004e-01
-1.28871310e+00 5.79247355e-01 -3.47920537e-01 -8.67218614e-01
-3.05684566e-01 -7.60769188e-01 -7.85557508e-01 8.61083388e-01
-1.38487768e+00 -6.37759864e-01 -1.56692997e-01 8.76342729e-02
4.24128681e-01 -2.37972200e-01 1.06025422e+00 3.55011940e-01
-5.10126650e-01 1.13455904e+00 -6.90846369e-02 2.48459354e-01
9.10245717e-01 -1.60785770e+00 1.07806909e+00 1.29595554e+00
1.46704376e-01 9.84266996e-01 7.94188082e-01 -9.02749777e-01
-1.01069069e+00 -9.73198593e-01 1.74098122e+00 -7.74029851e-01
3.85950357e-01 -2.64586121e-01 -1.06969702e+00 8.13316762e-01
9.77852494e-02 -1.04054749e-01 4.58223403e-01 5.15261173e-01
-2.94305146e-01 4.56611454e-01 -1.08866620e+00 3.70320737e-01
1.20368612e+00 -5.12897372e-01 -8.46474707e-01 6.34332076e-02
9.29444849e-01 -6.38951302e-01 -6.98099554e-01 5.98095953e-01
3.65247756e-01 -6.54419780e-01 2.77258068e-01 -8.25619996e-01
5.71713150e-01 -2.62411535e-01 -2.54961044e-01 -1.79379225e+00
-2.38386720e-01 -3.59277964e-01 1.78036943e-01 1.37199593e+00
8.10983658e-01 -5.80691218e-01 4.04347360e-01 7.11661756e-01
-6.15494609e-01 -3.98061097e-01 -5.80925643e-01 -9.95352387e-01
6.36472225e-01 -7.82008410e-01 4.15347546e-01 1.03904343e+00
2.20932767e-01 1.87554777e-01 1.64779201e-01 -1.55344829e-01
4.40393120e-01 -4.94272768e-01 5.01187205e-01 -1.23919284e+00
-1.18214324e-01 -4.40029591e-01 -5.28455079e-01 -6.93177879e-01
2.19334379e-01 -1.12140787e+00 4.16185141e-01 -1.25949705e+00
-2.48774309e-02 -3.84731501e-01 -1.38301373e-01 4.92628634e-01
-7.37605453e-01 1.86758459e-01 1.74305990e-01 -7.77249341e-04
-6.98127210e-01 3.68836284e-01 9.39214706e-01 1.66965965e-02
1.14142224e-01 -3.08085710e-01 -1.02873063e+00 6.37155950e-01
9.13721204e-01 -7.45111346e-01 1.48490012e-01 -1.01430082e+00
4.53517705e-01 -3.35006177e-01 1.19093917e-01 -1.12162375e+00
2.16049254e-01 1.91227973e-01 3.31472792e-02 -2.59807736e-01
-5.46030104e-01 -7.06636235e-02 -2.87423074e-01 4.95789677e-01
-4.73627210e-01 5.90410054e-01 3.44300002e-01 2.15129644e-01
-3.36491942e-01 -6.31542981e-01 5.70093632e-01 -1.83742315e-01
-7.39715397e-01 -1.83202088e-01 -4.61207569e-01 6.49659872e-01
3.89900446e-01 8.85256380e-02 -8.41763496e-01 -6.80989819e-03
-2.39172369e-01 3.49750191e-01 4.73659962e-01 4.73487854e-01
3.04009765e-01 -9.46074784e-01 -1.09827709e+00 -1.57295130e-02
5.14312744e-01 -2.58044064e-01 -2.64200598e-01 4.60575312e-01
-7.27023065e-01 6.35209858e-01 2.62930125e-01 -3.24287713e-01
-1.18639016e+00 -1.06559023e-01 4.54028875e-01 -3.82048994e-01
-5.15830576e-01 1.00960267e+00 -4.46489632e-01 -7.77382314e-01
7.37555996e-02 -3.85868549e-01 -5.89298643e-02 -4.62072939e-01
6.89002812e-01 3.92800659e-01 6.14747703e-01 -5.18391490e-01
-4.39641476e-01 1.64396793e-01 -3.39504331e-01 4.46845479e-02
1.12510121e+00 1.23908902e-02 -1.27667561e-01 3.57279897e-01
9.20128107e-01 4.56311107e-02 -4.27524596e-01 -1.73431516e-01
5.10775030e-01 -4.33925390e-01 -1.31054088e-01 -1.47388494e+00
-6.03907347e-01 6.74489260e-01 4.13498968e-01 -3.80620286e-02
7.61826456e-01 -3.33456188e-01 5.63126326e-01 5.07894456e-01
6.64371848e-01 -1.36734283e+00 -5.01698256e-01 1.32421446e+00
8.70963216e-01 -1.40284336e+00 -1.18936799e-01 -4.28178906e-01
-5.10071695e-01 8.95866334e-01 8.37682784e-01 -2.86869019e-01
5.68468869e-01 4.12454680e-02 5.71844429e-02 -3.84437054e-01
-1.04218018e+00 -9.04239044e-02 1.85240299e-01 4.40904588e-01
1.15187573e+00 3.01295727e-01 -1.20419216e+00 7.16379166e-01
-6.69738054e-01 -3.99663031e-01 5.59432805e-01 8.39324772e-01
-5.31716704e-01 -1.14577615e+00 -4.05292735e-02 5.42144418e-01
-7.28671193e-01 -4.69233960e-01 -3.72294605e-01 1.13027740e+00
6.67974055e-02 1.30709565e+00 -9.93915368e-03 -2.78361320e-01
5.19179881e-01 3.90216053e-01 6.84162796e-01 -1.01000476e+00
-1.00696838e+00 -4.98588443e-01 4.09257323e-01 -5.32937229e-01
5.14189675e-02 -7.05065131e-01 -1.44587064e+00 -6.41657710e-01
-5.70817769e-01 2.65592039e-01 7.17751324e-01 1.05516922e+00
2.73682028e-01 3.05803418e-01 -8.13392252e-02 -1.87211946e-01
-6.86721087e-01 -1.38109827e+00 -4.34625894e-01 5.80205917e-01
-6.33045435e-02 -3.99933100e-01 -6.36877537e-01 -1.17806405e-01] | [11.06196403503418, 10.689937591552734] |
64203665-d945-4db1-aff8-355e87eb522e | end-to-end-3d-dense-captioning-with-vote2cap | 2301.02508 | null | https://arxiv.org/abs/2301.02508v1 | https://arxiv.org/pdf/2301.02508v1.pdf | End-to-End 3D Dense Captioning with Vote2Cap-DETR | 3D dense captioning aims to generate multiple captions localized with their associated object regions. Existing methods follow a sophisticated ``detect-then-describe'' pipeline equipped with numerous hand-crafted components. However, these hand-crafted components would yield suboptimal performance given cluttered object spatial and class distributions among different scenes. In this paper, we propose a simple-yet-effective transformer framework Vote2Cap-DETR based on recent popular \textbf{DE}tection \textbf{TR}ansformer (DETR). Compared with prior arts, our framework has several appealing advantages: 1) Without resorting to numerous hand-crafted components, our method is based on a full transformer encoder-decoder architecture with a learnable vote query driven object decoder, and a caption decoder that produces the dense captions in a set-prediction manner. 2) In contrast to the two-stage scheme, our method can perform detection and captioning in one-stage. 3) Without bells and whistles, extensive experiments on two commonly used datasets, ScanRefer and Nr3D, demonstrate that our Vote2Cap-DETR surpasses current state-of-the-arts by 11.13\% and 7.11\% in CIDEr@0.5IoU, respectively. Codes will be released soon. | ['Gang Yu', 'Tao Chen', 'Yinjie Lei', 'Xin Chen', 'Hongyuan Zhu', 'Sijin Chen'] | 2023-01-06 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_End-to-End_3D_Dense_Captioning_With_Vote2Cap-DETR_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_End-to-End_3D_Dense_Captioning_With_Vote2Cap-DETR_CVPR_2023_paper.pdf | cvpr-2023-1 | ['dense-captioning', '3d-dense-captioning'] | ['computer-vision', 'computer-vision'] | [ 3.47070664e-01 1.73848242e-01 3.81770246e-02 -4.16048765e-01
-1.32363367e+00 -5.62168717e-01 6.90388620e-01 -3.38991791e-01
-2.89363921e-01 6.20707810e-01 2.43971407e-01 -1.65619969e-01
5.10140479e-01 -4.26142126e-01 -1.06406796e+00 -6.05298042e-01
4.28647816e-01 7.31382608e-01 4.97685939e-01 -1.33086920e-01
-1.01516120e-01 1.68343540e-02 -1.38950539e+00 4.26281720e-01
6.28249109e-01 1.18316364e+00 6.60437942e-01 4.80121255e-01
-4.34797667e-02 7.88903654e-01 -2.94987261e-01 -7.87735403e-01
3.00757647e-01 -5.33392906e-01 -3.66633147e-01 2.29053095e-01
7.21536517e-01 -6.04693711e-01 -3.15215200e-01 8.70491385e-01
5.66723645e-01 -1.89587727e-01 7.15505064e-01 -9.63534474e-01
-9.32425439e-01 6.07866883e-01 -9.37371135e-01 7.30752870e-02
1.45368218e-01 5.62312663e-01 1.14993393e+00 -1.44746351e+00
4.95077997e-01 1.29532826e+00 5.40338933e-01 6.81834698e-01
-1.32895482e+00 -8.78605127e-01 2.12290421e-01 -6.27606437e-02
-1.49763560e+00 -5.60319245e-01 6.71619833e-01 -4.62118506e-01
7.06308663e-01 1.14241287e-01 4.35881972e-01 1.18861091e+00
-1.18847087e-01 1.15781760e+00 9.15291131e-01 -7.84172788e-02
1.18418671e-01 1.22401945e-01 -2.79746920e-01 7.09777832e-01
1.98680758e-01 7.03518540e-02 -4.33967471e-01 1.43645883e-01
8.41229618e-01 -1.99659988e-01 -2.08988592e-01 -2.88279712e-01
-1.35320210e+00 7.47862995e-01 7.32275546e-01 -7.92890117e-02
-4.49297786e-01 5.41845500e-01 1.59499466e-01 -2.75026709e-01
5.45133471e-01 1.80093363e-01 -6.15163073e-02 1.26577765e-01
-1.24527943e+00 2.77042150e-01 4.16892886e-01 1.39288330e+00
7.65008926e-01 7.41708502e-02 -5.98801196e-01 7.63203144e-01
5.62955499e-01 9.56070185e-01 -3.14954631e-02 -6.05704248e-01
7.88148642e-01 2.56790131e-01 2.82597721e-01 -8.22214365e-01
-3.53579745e-02 -6.89015508e-01 -8.43376994e-01 -2.84961481e-02
1.39695346e-01 -1.02851607e-01 -1.31244242e+00 1.55981863e+00
2.25531757e-01 3.39972228e-01 -1.09428532e-01 1.30289388e+00
9.49702859e-01 7.37695813e-01 1.46064475e-01 1.45845369e-01
1.51993763e+00 -1.25534260e+00 -5.36228001e-01 -6.14585102e-01
2.01770350e-01 -8.69967341e-01 1.04138935e+00 1.65095642e-01
-1.22713220e+00 -5.96042752e-01 -8.63324523e-01 -3.37711364e-01
9.48625728e-02 1.90055251e-01 5.62334239e-01 3.72227490e-01
-1.19208109e+00 -3.56002524e-02 -5.54426730e-01 -1.70985565e-01
7.95790493e-01 2.42809877e-01 -1.53156117e-01 -2.50345975e-01
-8.18979204e-01 8.19699645e-01 3.29717457e-01 1.78249836e-01
-1.37020302e+00 -5.80446899e-01 -8.90517592e-01 -4.42003505e-03
3.61707062e-01 -9.93342757e-01 1.65726399e+00 -7.17446625e-01
-1.32370377e+00 8.59375536e-01 -2.92338282e-01 -4.57506478e-01
6.06702089e-01 -3.33654791e-01 -1.51566535e-01 2.09016651e-01
2.73717225e-01 1.15339756e+00 6.75753236e-01 -1.51740360e+00
-6.35738790e-01 7.21380040e-02 -7.61274621e-02 3.15346599e-01
-7.50064030e-02 9.37545523e-02 -1.10892808e+00 -7.18125701e-01
5.88951744e-02 -9.49663579e-01 -2.09617481e-01 4.41555381e-01
-5.88266969e-01 8.84930789e-02 6.72309160e-01 -4.85907257e-01
9.34188962e-01 -2.17776084e+00 9.21577960e-03 -2.26316452e-01
2.64022171e-01 3.51706505e-01 -2.87555099e-01 2.60845363e-01
2.64090896e-01 -2.16551218e-02 -4.37939376e-01 -8.07260752e-01
1.69421598e-01 3.47498357e-02 -3.66752177e-01 5.52546740e-01
5.46523631e-01 1.17143869e+00 -1.06188333e+00 -6.49776280e-01
4.44111943e-01 6.25890315e-01 -7.44149625e-01 2.92187124e-01
-6.23367667e-01 3.58053505e-01 -5.27343154e-01 8.15673172e-01
7.45416343e-01 -6.60250664e-01 -1.57878831e-01 -4.24510211e-01
-8.80996659e-02 2.16217130e-01 -9.58721220e-01 1.87285078e+00
-3.41301739e-01 6.41393781e-01 1.39555126e-01 -5.86806536e-01
9.15147901e-01 2.94226050e-01 2.01931655e-01 -7.90786445e-01
3.01623553e-01 4.02465612e-01 -3.34311515e-01 -3.19309205e-01
5.57315648e-01 -2.03133136e-01 -3.54600489e-01 2.46030137e-01
1.91913158e-01 -2.24324703e-01 1.23947620e-01 2.90219516e-01
1.08352625e+00 1.83427751e-01 -8.88471380e-02 -7.82712922e-03
1.79746777e-01 2.20045559e-02 3.55408728e-01 7.80939996e-01
-1.54781565e-01 1.30036736e+00 2.41730034e-01 -2.11975172e-01
-1.22415090e+00 -1.15705001e+00 -9.01893377e-02 8.11571956e-01
3.80413741e-01 -1.77871794e-01 -4.90014613e-01 -6.48982942e-01
-1.39578208e-01 8.48567367e-01 -4.94864255e-01 1.71351269e-01
-4.61932719e-01 -5.20788610e-01 5.05660474e-01 5.83007276e-01
6.50054932e-01 -9.51591969e-01 -3.90705436e-01 2.94343919e-01
-3.26482594e-01 -1.33594072e+00 -8.27854097e-01 1.04902431e-01
-5.47737002e-01 -6.61416590e-01 -1.13800299e+00 -9.07890141e-01
7.09063292e-01 5.74367940e-01 1.30592716e+00 -2.28341490e-01
6.87835272e-03 4.53895740e-02 -4.58582908e-01 -4.87027884e-01
-2.36804739e-01 7.16650188e-02 -3.39017808e-01 5.59006371e-02
2.60054618e-01 -4.96971846e-01 -9.03584838e-01 3.97396266e-01
-8.43678772e-01 6.87859058e-01 1.14786363e+00 7.90080249e-01
8.67765427e-01 -6.09694004e-01 3.00430834e-01 -6.56016588e-01
1.20936483e-01 -6.27701879e-01 -6.68893695e-01 1.27115846e-01
-4.63195264e-01 1.91636682e-01 3.44401509e-01 -2.55567580e-01
-1.00251162e+00 4.66510594e-01 -2.55102813e-01 -7.44734168e-01
-6.03293702e-02 5.26040718e-02 -1.61533102e-01 2.28105664e-01
6.14735544e-01 4.56680626e-01 -2.29989350e-01 -6.64118946e-01
5.95993876e-01 7.21300185e-01 8.48236620e-01 -4.48129565e-01
1.13410628e+00 4.50297832e-01 -3.74685287e-01 -2.13660747e-01
-1.01250851e+00 -3.55732173e-01 -2.13657930e-01 -3.01130056e-01
1.08002341e+00 -1.57708657e+00 -5.45528471e-01 3.28694850e-01
-1.40817094e+00 -1.03709154e-01 -1.58939272e-01 3.29565853e-01
-5.17871499e-01 1.91484634e-02 -5.63583374e-01 -8.02070439e-01
-5.99571288e-01 -1.19329202e+00 1.60361564e+00 2.54512101e-01
3.07594202e-02 -4.31711644e-01 -1.36177599e-01 6.19643629e-01
5.12722075e-01 2.59011954e-01 3.76726747e-01 -4.11262512e-01
-9.04773295e-01 -1.41238034e-01 -7.21483648e-01 2.35960022e-01
-1.54561922e-01 -3.86213064e-01 -1.15021384e+00 -2.52505839e-01
-4.25410926e-01 -4.58441705e-01 8.49602342e-01 1.73610196e-01
1.12548935e+00 -4.09701288e-01 -3.47292960e-01 5.96957684e-01
1.57708740e+00 -1.55582360e-03 5.00617504e-01 6.62964359e-02
8.56261730e-01 1.88541427e-01 5.25700867e-01 4.34209824e-01
5.63873231e-01 8.34956408e-01 8.87339473e-01 -4.24474120e-01
-5.19165456e-01 -6.42271876e-01 4.93939728e-01 6.37411058e-01
1.57729551e-01 -6.97913051e-01 -8.03601086e-01 7.62812793e-01
-1.81024361e+00 -8.72316837e-01 -1.88955382e-01 1.75791883e+00
9.98708189e-01 2.32603028e-01 -2.17129644e-02 -4.26062763e-01
8.28683972e-01 2.99905837e-01 -6.68446541e-01 -1.12145124e-02
-1.82040557e-01 1.55284330e-01 5.90959072e-01 3.80686879e-01
-1.01934552e+00 1.06726503e+00 5.09899569e+00 8.16873133e-01
-1.03264272e+00 2.63156205e-01 6.55688226e-01 -3.42179149e-01
-5.53109646e-01 -1.00479141e-01 -1.07159519e+00 6.53645217e-01
7.97285140e-01 4.69836146e-02 5.77586815e-02 8.73811007e-01
2.02400088e-01 -1.66004021e-02 -1.18790925e+00 1.16864443e+00
2.05175743e-01 -1.43706512e+00 1.39201492e-01 -9.79367793e-02
6.15007162e-01 3.92410010e-01 1.68063328e-01 3.47393394e-01
5.05000055e-01 -8.86886656e-01 1.32700050e+00 2.95460224e-01
1.16702294e+00 -4.19478595e-01 6.76731348e-01 1.81001902e-01
-1.18875802e+00 1.93969727e-01 -2.02998027e-01 1.99310228e-01
6.31208122e-01 7.54086912e-01 -1.13869309e+00 4.40811843e-01
6.90219879e-01 6.24978781e-01 -4.43312824e-01 1.30773556e+00
-3.87214601e-01 6.88141882e-01 -4.37712371e-01 -1.25714922e-02
4.53289002e-01 2.38588452e-01 5.64643919e-01 1.36260462e+00
4.59246963e-01 1.22283399e-01 6.12233132e-02 1.07054508e+00
-2.40628913e-01 -1.98995084e-01 -2.91946799e-01 5.74508682e-02
4.76102918e-01 1.08422661e+00 -6.25507474e-01 -3.23380440e-01
-4.24853384e-01 1.22328818e+00 1.26272932e-01 2.91735202e-01
-1.28433299e+00 -9.14396718e-02 3.74647498e-01 3.80299956e-01
8.23921919e-01 1.00071719e-02 -2.81264812e-01 -1.07477629e+00
1.97570086e-01 -7.63910651e-01 7.74876699e-02 -1.13219821e+00
-1.39729655e+00 8.81769776e-01 -4.84268740e-02 -1.41639209e+00
-8.46506730e-02 -3.64305884e-01 -2.91542321e-01 9.31638181e-01
-1.65465689e+00 -1.51970851e+00 -3.52216393e-01 5.56884646e-01
8.47669661e-01 1.55528530e-01 6.02224886e-01 5.51506519e-01
-6.53177798e-01 7.13039577e-01 -4.69957059e-03 1.55556798e-01
7.43353069e-01 -1.08198452e+00 5.96481204e-01 8.30606282e-01
2.11364225e-01 -1.64712500e-03 8.41390789e-01 -5.10878801e-01
-1.39216375e+00 -1.38556921e+00 9.63134408e-01 -5.15945137e-01
4.04084563e-01 -7.97303259e-01 -5.74205399e-01 6.67154014e-01
3.83466601e-01 1.97991565e-01 2.94343352e-01 -2.34694824e-01
-3.28822166e-01 -2.19353780e-01 -8.91425610e-01 6.38370872e-01
1.05026424e+00 -2.85571367e-01 -4.59460527e-01 4.84840453e-01
8.56557965e-01 -6.41556501e-01 -3.37885290e-01 2.87723064e-01
3.88428390e-01 -8.24910283e-01 9.19239759e-01 -2.39793714e-02
7.70686388e-01 -6.08055651e-01 -2.56622642e-01 -9.04134631e-01
-4.60500717e-01 -7.07637131e-01 -2.59816021e-01 1.25270641e+00
6.62532449e-01 -1.52418479e-01 8.05427194e-01 5.62103152e-01
-5.79304934e-01 -7.46891618e-01 -9.24426973e-01 -6.19918644e-01
-2.02734694e-01 -5.00202298e-01 4.69110131e-01 4.64506149e-01
-5.17083049e-01 7.61492491e-01 -7.17185676e-01 2.96603113e-01
7.51233220e-01 1.69556811e-01 7.81161010e-01 -8.07989240e-01
-4.62941945e-01 -1.90105721e-01 -1.47982746e-01 -1.65942907e+00
-2.72442341e-01 -8.15432429e-01 5.95796227e-01 -1.73626387e+00
4.29584593e-01 -6.47580147e-01 -2.23712742e-01 7.61654317e-01
-2.37285882e-01 6.82134271e-01 3.73911440e-01 4.07854527e-01
-1.02317715e+00 8.50096643e-01 1.44967723e+00 -2.94248402e-01
1.55804247e-01 -2.25150734e-01 -8.87352586e-01 4.79282230e-01
4.37008858e-01 -7.29583323e-01 -4.61619198e-01 -8.96691918e-01
-2.53248755e-02 3.75644416e-02 6.75009191e-01 -1.11479604e+00
2.50500083e-01 4.54379916e-02 4.41865921e-01 -9.71451819e-01
6.70047581e-01 -7.24239767e-01 2.29566723e-01 1.49502486e-01
-7.43730366e-02 7.68424347e-02 3.32653463e-01 7.42147207e-01
-1.43082559e-01 9.25948843e-02 7.57620215e-01 -1.74284607e-01
-7.50263333e-01 5.02660275e-01 -8.68352950e-02 -2.12106686e-02
1.09848297e+00 -2.11624607e-01 -3.09118837e-01 -3.86409014e-01
-2.21878052e-01 4.41693574e-01 4.21501547e-01 5.57626605e-01
6.16343439e-01 -1.30743396e+00 -1.08674443e+00 -9.81605276e-02
2.94978410e-01 5.16266882e-01 2.64857292e-01 6.59723997e-01
-4.85555083e-01 4.82208043e-01 1.49124727e-01 -8.81918252e-01
-8.40935111e-01 5.48000038e-01 2.04171002e-01 -1.66767210e-01
-6.77428186e-01 1.24295533e+00 6.15889072e-01 -2.93469429e-02
2.11237177e-01 -1.60763234e-01 1.90573737e-01 -6.76512197e-02
4.53264207e-01 -2.84357756e-01 -9.41579565e-02 -7.45951712e-01
-4.39676881e-01 3.52557629e-01 -2.76111096e-01 -1.71634123e-01
1.41232312e+00 -1.37753278e-01 3.80215496e-01 1.06366165e-01
1.03844869e+00 -3.46815199e-01 -1.79173160e+00 -3.51852417e-01
-4.76163238e-01 -4.39755291e-01 1.06227398e-01 -9.65909898e-01
-1.25521362e+00 7.92500317e-01 6.18016660e-01 -3.38647157e-01
1.12185884e+00 3.84810686e-01 1.00240254e+00 6.76285848e-02
3.32939982e-01 -7.26893544e-01 2.24230438e-01 3.05796385e-01
1.09989846e+00 -1.29225492e+00 -8.44996497e-02 -4.93774861e-01
-7.81728923e-01 5.44791222e-01 9.07023907e-01 2.14874139e-03
3.58413249e-01 1.91796482e-01 -2.62733195e-02 -2.35746667e-01
-8.91280353e-01 -3.30253929e-01 3.52946043e-01 4.49490994e-01
2.79287010e-01 -1.26113951e-01 -6.50083134e-03 6.34778142e-01
-5.10740019e-02 4.52035703e-02 2.60872066e-01 6.86538696e-01
-4.84176129e-01 -8.74554873e-01 -3.71072292e-01 3.23471218e-01
-2.65436620e-01 -5.11065423e-01 -1.52149022e-01 6.78586423e-01
3.02712679e-01 8.69934082e-01 1.59193054e-01 -4.50679153e-01
4.30754304e-01 -1.95766777e-01 1.97342798e-01 -8.04506063e-01
-5.63326180e-01 1.93849802e-01 1.03531100e-01 -5.20690084e-01
-2.94313282e-01 -5.97102404e-01 -1.03607953e+00 -1.23066776e-01
-5.36389232e-01 -4.19569239e-02 5.32609224e-01 7.86937714e-01
6.72628224e-01 4.22504216e-01 3.92779559e-01 -1.07010877e+00
-4.82792079e-01 -9.60521579e-01 -2.64441937e-01 2.64140785e-01
3.92180592e-01 -5.74115038e-01 -1.58306539e-01 2.55759537e-01] | [10.345929145812988, 0.9399077892303467] |
10328b4f-5deb-45cc-b852-677ab8e13855 | few-shot-text-classification-with-triplet | 2103.07552 | null | https://arxiv.org/abs/2103.07552v1 | https://arxiv.org/pdf/2103.07552v1.pdf | Few-Shot Text Classification with Triplet Networks, Data Augmentation, and Curriculum Learning | Few-shot text classification is a fundamental NLP task in which a model aims to classify text into a large number of categories, given only a few training examples per category. This paper explores data augmentation -- a technique particularly suitable for training with limited data -- for this few-shot, highly-multiclass text classification setting. On four diverse text classification tasks, we find that common data augmentation techniques can improve the performance of triplet networks by up to 3.0% on average. To further boost performance, we present a simple training strategy called curriculum data augmentation, which leverages curriculum learning by first training on only original examples and then introducing augmented data as training progresses. We explore a two-stage and a gradual schedule, and find that, compared with standard single-stage training, curriculum data augmentation trains faster, improves performance, and remains robust to high amounts of noising from augmentation. | ['Shiqi Xu', 'Yu Cheng', 'Soroush Vosoughi', 'Chengyu Huang', 'Jason Wei'] | 2021-03-12 | null | https://aclanthology.org/2021.naacl-main.434 | https://aclanthology.org/2021.naacl-main.434.pdf | naacl-2021-4 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 7.16542602e-01 1.17184319e-01 -4.82340366e-01 -6.04686081e-01
-8.96030426e-01 -4.77445483e-01 9.19952035e-01 5.97141266e-01
-7.55571902e-01 6.82763815e-01 1.51806846e-01 -6.32411659e-01
1.25488207e-01 -5.36390483e-01 -3.46539944e-01 -4.80630070e-01
3.24319214e-01 8.46651673e-01 1.07758000e-01 -3.80475640e-01
1.70286298e-01 2.69182902e-02 -1.56903112e+00 4.60643202e-01
7.01391339e-01 7.66627252e-01 -3.35249126e-01 1.02362633e+00
-5.90800822e-01 8.44509244e-01 -6.93606913e-01 -3.23691279e-01
1.42859459e-01 -4.53789353e-01 -1.00455034e+00 6.88796192e-02
6.23229980e-01 -2.67549843e-01 -3.45719486e-01 6.09055936e-01
4.25972879e-01 5.97572803e-01 6.86034858e-01 -1.35709405e+00
-8.38344753e-01 8.21597636e-01 -5.40704250e-01 4.56354290e-01
1.29651904e-01 4.06579711e-02 1.25238800e+00 -9.61600304e-01
4.83959347e-01 8.45979214e-01 8.52722347e-01 9.55277860e-01
-1.52409208e+00 -7.27161527e-01 2.64242679e-01 -1.08010784e-01
-7.55236208e-01 -4.72017318e-01 6.42148376e-01 -4.04814571e-01
1.16640007e+00 1.98802650e-01 3.85249704e-01 1.32764208e+00
-1.51570931e-01 9.97495770e-01 9.59496915e-01 -7.71961689e-01
1.70980021e-01 2.68875927e-01 1.08614719e+00 4.25835133e-01
1.21788613e-01 -1.19006066e-02 -5.05665839e-01 -2.54384190e-01
-3.36556993e-02 1.85905144e-01 5.53156696e-02 -2.66019404e-01
-1.04919648e+00 1.02029693e+00 3.84750128e-01 4.72673208e-01
1.63515359e-01 1.98768049e-01 7.19518900e-01 7.44466186e-01
8.25707853e-01 7.59416461e-01 -7.77023792e-01 -3.51618141e-01
-9.85183477e-01 1.14689030e-01 8.83958697e-01 1.03754711e+00
6.98690832e-01 1.55889392e-01 -3.78035367e-01 1.01934588e+00
-1.51058644e-01 6.36025965e-02 8.58780980e-01 -3.56101573e-01
6.60354674e-01 6.82507515e-01 -1.75532848e-01 -1.93325996e-01
-4.58712667e-01 -2.11526647e-01 -7.87952244e-01 2.76260376e-02
3.26936334e-01 -5.06204188e-01 -1.51325357e+00 1.54338384e+00
2.42333308e-01 6.28524721e-02 2.37252310e-01 3.99396181e-01
9.92789090e-01 7.00561583e-01 4.56635863e-01 -2.18485355e-01
1.33145595e+00 -1.07399321e+00 -7.01839566e-01 -4.61699039e-01
1.37710476e+00 -4.24623281e-01 1.32170284e+00 8.51512849e-02
-8.59727919e-01 -4.45627868e-01 -1.17027640e+00 -3.89879435e-01
-9.02511239e-01 -2.90622830e-01 7.77884662e-01 9.18152034e-01
-7.09672987e-01 7.21633494e-01 -6.01960719e-01 -5.25194466e-01
6.86294734e-01 3.76555771e-01 -3.41143578e-01 -2.96108633e-01
-1.08592403e+00 9.64222252e-01 5.95491588e-01 -6.17380619e-01
-6.33438349e-01 -1.13918090e+00 -9.52176988e-01 3.19151133e-01
1.22493766e-01 -3.70616585e-01 1.55675769e+00 -7.92439520e-01
-1.20286357e+00 6.49133563e-01 -1.04020365e-01 -5.35529375e-01
2.32964277e-01 -1.13182597e-01 -2.79709905e-01 -1.67089015e-01
-7.53994733e-02 8.66347790e-01 7.00563192e-01 -1.10423851e+00
-7.01275170e-01 -3.10768634e-01 -1.30724758e-01 3.17606211e-01
-1.18869328e+00 1.39305755e-01 5.29444590e-02 -6.85213149e-01
-2.73052990e-01 -7.15072870e-01 -2.72465289e-01 -1.58905625e-01
-1.44202337e-01 -5.40185750e-01 1.24338377e+00 -2.19916999e-01
9.39047873e-01 -1.93499184e+00 -4.11130264e-02 -1.02012634e-01
1.69080749e-01 5.20949483e-01 -4.05128956e-01 1.81672364e-01
-4.41664636e-01 2.85758168e-01 -4.28031296e-01 -7.72321105e-01
4.44798321e-02 3.98084700e-01 -4.09665316e-01 1.26535103e-01
5.43212712e-01 1.17140388e+00 -9.23261285e-01 -3.27407062e-01
7.57585093e-02 1.07190400e-01 -3.71858031e-01 8.14242363e-02
-4.88930166e-01 1.96866453e-01 -1.83331564e-01 4.91067529e-01
2.60435969e-01 -4.35149312e-01 -1.24567814e-01 3.17404568e-01
4.90763746e-02 2.52191305e-01 -8.82331491e-01 1.75102675e+00
-3.20740461e-01 1.07522082e+00 -5.53917527e-01 -1.33020151e+00
8.50296974e-01 5.18747807e-01 4.82587516e-01 -5.29912293e-01
5.05334735e-01 -1.36524305e-01 5.58108911e-02 -2.89138943e-01
5.97543836e-01 -4.44919437e-01 -2.74122536e-01 9.45579886e-01
4.92327601e-01 -3.83214056e-02 4.90432054e-01 3.94467771e-01
1.25462008e+00 -2.23099425e-01 2.15627611e-01 -1.33819610e-01
8.42228830e-02 4.71385211e-01 2.20931157e-01 1.01091838e+00
-2.11441353e-01 5.36985695e-01 4.49037611e-01 -6.63984776e-01
-1.40675735e+00 -3.80180180e-01 -3.73910517e-01 1.88464022e+00
-3.62137556e-01 -2.70014852e-01 -4.72469598e-01 -9.42997217e-01
1.35027424e-01 9.71172750e-01 -1.05012691e+00 -4.15463150e-01
-3.18601042e-01 -1.23483574e+00 4.95296001e-01 8.93776238e-01
2.07951561e-01 -9.04266596e-01 -2.93401301e-01 1.46173060e-01
9.83129144e-02 -1.03993416e+00 -3.37910563e-01 9.27364111e-01
-1.11415565e+00 -8.94086897e-01 -7.52502799e-01 -1.03381228e+00
6.86933577e-01 3.74324262e-01 1.06239724e+00 3.74301463e-01
-4.53254730e-01 2.46181592e-01 -6.59358442e-01 -6.30140066e-01
-5.48261285e-01 4.14539397e-01 -1.15620442e-01 -3.21950525e-01
7.56389737e-01 -3.21080595e-01 -1.58252157e-02 -5.24141081e-02
-7.77923822e-01 1.69134028e-02 3.65229875e-01 1.27161670e+00
-2.52283271e-02 -2.29057193e-01 6.05130970e-01 -1.43496406e+00
7.43559957e-01 -5.50162971e-01 -2.13159725e-01 1.61090329e-01
-8.57268929e-01 -8.75750184e-02 8.61835778e-01 -8.13435078e-01
-9.67069507e-01 -5.62190153e-02 2.65175570e-02 -4.21760350e-01
-3.87586594e-01 4.93688136e-01 4.12465811e-01 -2.49062568e-01
1.22600126e+00 1.65938377e-01 -2.56676469e-02 -3.09373409e-01
5.01813114e-01 8.99502933e-01 2.08699480e-01 -4.58083898e-01
9.26864445e-01 2.77577698e-01 -2.66008526e-01 -8.71416211e-01
-1.29906678e+00 -8.33078980e-01 -8.62973213e-01 2.34220058e-01
5.32451153e-01 -8.17624152e-01 -2.70902276e-01 1.27103880e-01
-9.19431388e-01 -7.86338270e-01 -7.06714869e-01 2.80014783e-01
-2.25880772e-01 1.82090878e-01 -7.24877656e-01 -6.09402835e-01
-5.74288011e-01 -5.28209329e-01 6.72031581e-01 9.64690447e-02
-1.97380528e-01 -1.15311241e+00 2.74043769e-01 2.40812227e-01
3.73544097e-01 -9.67775211e-02 1.09477651e+00 -1.44037366e+00
1.35866448e-01 -5.14014661e-01 -2.09361836e-01 1.52356863e-01
7.97613245e-03 -1.00782260e-01 -1.22780120e+00 -5.73013544e-01
-1.86250299e-01 -9.51699436e-01 9.77957368e-01 4.71149869e-02
1.11815679e+00 -2.49480866e-02 -3.40279818e-01 4.23188239e-01
1.14543712e+00 1.79582760e-01 3.12521040e-01 4.00310695e-01
7.01968133e-01 3.80408525e-01 5.72245896e-01 2.89184958e-01
-4.53014895e-02 2.78821200e-01 1.53402286e-02 -3.10562611e-01
-4.94300053e-02 1.60132393e-01 -1.48156628e-01 7.25611985e-01
3.11582983e-01 -3.57985765e-01 -1.06719804e+00 6.38128698e-01
-1.74398518e+00 -8.34393263e-01 -6.96856454e-02 1.87864435e+00
1.04440165e+00 3.30591321e-01 1.01778440e-01 3.84135038e-01
5.71676075e-01 9.46634933e-02 -6.51677251e-01 -5.48610985e-01
7.98407868e-02 6.11595988e-01 2.90250629e-01 2.97792941e-01
-1.33603525e+00 1.08914280e+00 7.73398972e+00 5.53375483e-01
-7.95800149e-01 2.78019726e-01 6.34189248e-01 -3.79277021e-01
-8.60764682e-02 -3.01678702e-02 -1.04002023e+00 2.81602174e-01
1.31276321e+00 -4.60241735e-01 1.95387639e-02 9.55068409e-01
-3.73734564e-01 2.63129681e-01 -1.35227048e+00 5.09025574e-01
3.11636984e-01 -1.45864213e+00 2.40433484e-01 -1.69619843e-01
1.06371057e+00 2.11377576e-01 -4.52093966e-03 1.03020597e+00
6.84051454e-01 -9.08609807e-01 1.60214789e-02 -1.50375932e-01
8.35970461e-01 -7.56676257e-01 7.05676496e-01 4.43690687e-01
-8.59390020e-01 -4.07623053e-01 -4.59551573e-01 -9.63621289e-02
-2.38167405e-01 1.23282924e-01 -7.61188149e-01 1.78644210e-01
5.15544653e-01 7.22633719e-01 -7.48509347e-01 9.91994977e-01
7.27832783e-03 6.93322718e-01 -1.40141159e-01 -4.17611986e-01
5.08120060e-01 1.94344997e-01 1.21329583e-01 1.42549026e+00
-5.14538810e-02 4.48441744e-01 3.65835339e-01 3.25620502e-01
-4.75027591e-01 6.62562996e-02 -6.26343191e-01 -2.23669022e-01
2.57078111e-01 1.31226897e+00 -6.21952295e-01 -8.47460866e-01
-6.36649728e-01 7.50285447e-01 5.41707039e-01 2.71920353e-01
-5.02170622e-01 -8.20094764e-01 3.37595850e-01 -4.07823086e-01
2.28102878e-01 -9.35707285e-05 -5.73939919e-01 -1.15851891e+00
-3.36823404e-01 -8.37656081e-01 6.88397944e-01 -5.96688092e-01
-1.31516492e+00 4.39687908e-01 -1.75158903e-01 -1.06767833e+00
-2.37888232e-01 -7.69262314e-01 -8.92271161e-01 6.85817599e-01
-1.58146894e+00 -1.11029339e+00 -4.45841253e-01 4.50929850e-01
9.31290627e-01 -3.42736423e-01 1.18517387e+00 1.52170300e-01
-7.48307288e-01 9.58426774e-01 5.67950010e-01 4.72824574e-01
7.71638215e-01 -1.55431306e+00 8.96653295e-01 6.81840420e-01
2.96070725e-01 4.13673431e-01 7.26471722e-01 -4.46099311e-01
-1.05456877e+00 -1.11622894e+00 9.56907153e-01 -6.89112365e-01
1.01500833e+00 -5.36982775e-01 -1.40787435e+00 1.02902460e+00
2.30898336e-01 1.42032832e-01 1.24484491e+00 6.93659902e-01
-6.22953713e-01 1.81435347e-01 -1.08743501e+00 6.02189124e-01
7.01280773e-01 -3.79266262e-01 -1.02650118e+00 6.87238812e-01
9.08343077e-01 -2.97840595e-01 -1.00934815e+00 1.24176197e-01
3.36912572e-01 -2.47854203e-01 6.68416679e-01 -1.48172319e+00
5.71443737e-01 3.92018318e-01 1.21191792e-01 -1.43967366e+00
-3.24753046e-01 -5.46860337e-01 -3.59442681e-01 1.05352807e+00
6.85175121e-01 -4.65071023e-01 1.30915701e+00 7.94671655e-01
-1.96630552e-01 -4.12055492e-01 -7.87233651e-01 -7.18648314e-01
6.10504389e-01 -3.27176988e-01 2.65442729e-01 1.53035438e+00
5.65349698e-01 7.98381329e-01 -2.86258131e-01 -4.29137260e-01
5.96634924e-01 -1.20976949e-02 7.65304804e-01 -1.58530056e+00
6.78258985e-02 -4.50299680e-01 -1.92474663e-01 -9.56430733e-01
2.08613738e-01 -1.10626924e+00 1.84949040e-01 -1.38337982e+00
4.69549567e-01 -4.47428405e-01 -4.90327686e-01 9.83864963e-01
-5.07362902e-01 3.70814711e-01 6.15786873e-02 6.05036691e-02
-4.83403862e-01 6.17352545e-01 1.02298796e+00 -3.01552922e-01
-1.75115570e-01 -3.57878953e-02 -9.22744155e-01 3.16722512e-01
8.62693906e-01 -6.27888799e-01 -4.08595473e-01 -4.67191160e-01
-1.17695644e-01 -3.10384721e-01 -2.32024174e-02 -9.23199058e-01
5.26417136e-01 3.88013236e-02 6.45969093e-01 -5.79091907e-01
3.43873650e-01 -6.66919291e-01 -7.42019236e-01 4.70693260e-01
-9.96844709e-01 -1.88963667e-01 5.66445768e-01 6.53887689e-01
1.38934329e-01 -6.77246094e-01 1.01221788e+00 -6.47592247e-02
-6.56570375e-01 4.00608122e-01 -4.80455399e-01 3.43408197e-01
1.00474811e+00 -2.41976216e-01 -6.53967321e-01 -4.49206829e-02
-8.57029676e-01 5.42911947e-01 1.75102383e-01 5.47207296e-01
2.83096641e-01 -1.30270839e+00 -6.78752184e-01 2.73281932e-01
4.82241482e-01 -1.44326717e-01 -2.03413852e-02 5.60855150e-01
6.57368638e-03 5.63140512e-01 -7.94399008e-02 -5.70860982e-01
-1.47151887e+00 6.59612000e-01 1.80265337e-01 -2.27691278e-01
-8.45969379e-01 9.02491927e-01 -2.93452680e-01 -7.67214358e-01
4.92392570e-01 -5.57237081e-02 -3.97073239e-01 2.96322763e-01
7.71343648e-01 7.45005235e-02 2.65425056e-01 -1.70735896e-01
9.44336951e-02 2.91195810e-01 -8.47797275e-01 -2.21136529e-02
1.23506737e+00 3.94685775e-01 3.96722436e-01 7.60532260e-01
1.37844920e+00 -9.54977989e-01 -1.12948120e+00 -5.95947444e-01
1.64803907e-01 -3.00401330e-01 2.09970102e-01 -8.42751324e-01
-6.15831733e-01 1.10384154e+00 4.21731025e-01 5.64107060e-01
6.87308550e-01 -1.85802370e-01 5.87494969e-01 1.01851451e+00
-2.51818627e-01 -1.29571748e+00 4.71704632e-01 8.08428347e-01
2.60079950e-01 -1.67337191e+00 1.21808693e-01 -8.83202925e-02
-5.78788340e-01 1.15794706e+00 7.43668258e-01 9.80691835e-02
5.75365424e-01 2.26565003e-01 3.88585851e-02 -1.21957555e-01
-1.24243534e+00 -2.52695739e-01 1.95120499e-01 5.86207747e-01
6.20656788e-01 -3.49025846e-01 1.62825644e-01 1.54619858e-01
-6.50374070e-02 -9.99037176e-02 5.80009401e-01 1.33629775e+00
-8.30718815e-01 -9.85493839e-01 -8.45629126e-02 1.01717842e+00
-3.34192783e-01 -3.87397468e-01 -4.23538208e-01 9.95805264e-01
-2.75930017e-01 1.02679288e+00 6.01275861e-01 -1.80729017e-01
3.33816051e-01 7.39710093e-01 1.74560234e-01 -1.15262163e+00
-8.50790143e-01 -2.48101756e-01 1.44985899e-01 2.85959356e-02
-2.01351047e-01 -6.90611422e-01 -1.04245675e+00 -3.75250787e-01
-6.79048061e-01 4.18731183e-01 7.03012526e-01 1.23083985e+00
9.60534662e-02 7.85470307e-01 6.37484848e-01 -7.35997319e-01
-8.84860694e-01 -1.33757114e+00 -3.62436116e-01 4.69716460e-01
4.19560462e-01 -4.30532128e-01 -6.01374328e-01 6.57652840e-02] | [10.764201164245605, 7.925563335418701] |
48bf0ede-3500-4719-b258-73cf49fc8ef8 | the-rumour-mill-making-misinformation-spread | 2002.04494 | null | https://arxiv.org/abs/2002.04494v2 | https://arxiv.org/pdf/2002.04494v2.pdf | The Rumour Mill: Making the Spread of Misinformation Explicit and Tangible | Misinformation spread presents a technological and social threat to society. With the advance of AI-based language models, automatically generated texts have become difficult to identify and easy to create at scale. We present "The Rumour Mill", a playful art piece, designed as a commentary on the spread of rumours and automatically-generated misinformation. The mill is a tabletop interactive machine, which invites a user to experience the process of creating believable text by interacting with different tangible controls on the mill. The user manipulates visible parameters to adjust the genre and type of an automatically generated text rumour. The Rumour Mill is a physical demonstration of the state of current technology and its ability to generate and manipulate natural language text, and of the act of starting and spreading rumours. | ['Leon Derczynski', 'Jeanette Falk Olesen', 'Nanna Inie'] | 2020-02-11 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [ 2.41789967e-01 3.46722126e-01 -9.06852819e-03 2.88027287e-01
-1.33074448e-02 -8.52611363e-01 1.28973067e+00 6.20084479e-02
1.90908954e-01 5.71459651e-01 6.30499065e-01 -5.63556731e-01
4.63747352e-01 -7.64237404e-01 -2.50291467e-01 -7.66129866e-02
-1.01495162e-01 4.71426159e-01 2.36998037e-01 -7.96596169e-01
7.99288034e-01 1.96950033e-01 -1.23521399e+00 3.85814607e-01
4.42796707e-01 2.82399088e-01 2.29539186e-01 9.15451407e-01
-4.10856485e-01 1.26205707e+00 -1.17575192e+00 -2.25083113e-01
-2.41164297e-01 -8.26801300e-01 -6.03478193e-01 -2.14476556e-01
-1.19860671e-01 -6.40128016e-01 -4.69279885e-02 1.13841128e+00
2.22443968e-01 -2.79554367e-01 3.82323831e-01 -1.08636403e+00
-1.20511949e+00 9.95818913e-01 -4.58883762e-01 9.03336108e-02
1.13043046e+00 2.19092280e-01 2.45294243e-01 -5.35339475e-01
1.08804560e+00 1.59949970e+00 6.16703272e-01 5.54327130e-01
-1.15686452e+00 -6.52888954e-01 -4.02855217e-01 -4.39175129e-01
-1.17040884e+00 -2.79095799e-01 8.49317908e-01 -9.59451497e-01
5.16529262e-01 7.92116761e-01 1.04757953e+00 1.28041279e+00
7.34328926e-01 7.10083663e-01 9.38328147e-01 -4.63815629e-01
4.25502300e-01 7.39506006e-01 -3.44349623e-01 8.89278591e-01
5.22675142e-02 8.76919180e-02 -9.01508212e-01 -8.26540112e-01
1.24935293e+00 -2.95853287e-01 -1.68819278e-01 -5.17974198e-02
-1.30772173e+00 7.33923078e-01 3.70102786e-02 4.66795206e-01
-3.95667404e-01 1.55623764e-01 2.56233037e-01 3.41050059e-01
9.17493999e-01 9.82033372e-01 2.07153365e-01 -6.91279888e-01
-5.55812418e-01 4.54666227e-01 1.20753431e+00 6.67977333e-01
-1.34777874e-01 -1.36087924e-01 -4.70991433e-02 6.96226239e-01
3.62439632e-01 6.69276655e-01 5.67533195e-01 -6.94675446e-01
2.73319632e-02 5.79367816e-01 8.26074302e-01 -1.75767672e+00
-2.71444023e-01 5.32406643e-02 -2.21773624e-01 7.50532389e-01
4.25916731e-01 -5.89414239e-01 -2.70385087e-01 1.02268565e+00
3.73714447e-01 -2.08381534e-01 -3.71253788e-01 9.76792634e-01
6.33058906e-01 8.06562126e-01 -9.79470313e-02 -4.06562574e-02
1.11958909e+00 -1.68298155e-01 -1.17982602e+00 -7.53122494e-02
6.20208502e-01 -9.89435196e-01 1.27181005e+00 4.71399367e-01
-1.21463180e+00 1.75872013e-01 -1.36030746e+00 5.24612181e-02
-5.33862770e-01 -5.21850884e-01 6.85951710e-01 9.06852067e-01
-6.71510696e-01 7.92413890e-01 -5.25779665e-01 -3.48153055e-01
4.60197061e-01 -3.13148499e-01 2.09705800e-01 4.22840744e-01
-1.12476313e+00 1.05453551e+00 6.66667446e-02 -2.77781840e-02
-5.28032899e-01 -6.55317187e-01 -4.73495483e-01 -3.24991196e-01
1.29023701e-01 -8.01006794e-01 1.47805381e+00 -1.14424586e+00
-2.22418571e+00 8.27928901e-01 4.15411651e-01 -2.94027627e-01
1.07650340e+00 -7.10234046e-01 -5.47998309e-01 4.39919978e-02
-8.55367556e-02 2.35679463e-01 1.01652479e+00 -1.41869807e+00
-1.07161321e-01 -1.17513612e-01 -1.21903732e-01 -6.94072619e-02
-2.79939715e-02 6.07019126e-01 5.91714561e-01 -7.05169022e-01
-1.06716417e-01 -5.79637825e-01 4.28679660e-02 1.10386334e-01
-6.43589139e-01 2.09460735e-01 1.14139318e+00 -5.79046965e-01
1.15620744e+00 -1.91592538e+00 -3.19434464e-01 -3.63462116e-03
4.69787985e-01 4.33218062e-01 4.59632874e-01 1.05964625e+00
4.47829306e-01 7.04184353e-01 2.84908354e-01 9.78831425e-02
1.70025676e-01 -4.35473979e-01 -3.27780247e-01 4.48672384e-01
-3.23517472e-01 6.90829575e-01 -1.59904838e+00 -2.15984195e-01
9.85646099e-02 5.94619274e-01 -2.84676254e-01 -2.92209350e-02
-5.50302684e-01 3.28139871e-01 -7.82500625e-01 3.25900555e-01
4.87457216e-01 -1.42165869e-01 1.94846541e-01 6.22424603e-01
-4.77632046e-01 6.62878633e-01 -5.44444680e-01 1.28533173e+00
-3.88344705e-01 9.99347091e-01 2.32527807e-01 3.37134808e-01
1.03056264e+00 3.93266916e-01 -1.44428700e-01 -2.83251345e-01
4.55077171e-01 -3.91389579e-02 -5.30067146e-01 -9.25987422e-01
9.38150227e-01 -4.53542143e-01 -3.13358992e-01 1.33518302e+00
-8.06134105e-01 -9.77382660e-01 -4.76589620e-01 5.42627454e-01
1.07676816e+00 7.90159553e-02 4.68888581e-01 -3.16850156e-01
-2.16981918e-01 1.12839624e-01 -5.04483998e-01 9.51042533e-01
1.24761499e-01 3.13367844e-01 4.28198010e-01 -7.52390981e-01
-9.46698070e-01 -1.01259351e+00 2.99068242e-01 1.31982100e+00
2.41427034e-01 -4.45825875e-01 -7.78554440e-01 -3.63486826e-01
-4.08154055e-02 1.37104237e+00 -5.42621553e-01 -2.06255279e-02
-6.45439997e-02 -4.88520533e-01 4.30890262e-01 -2.87836730e-01
2.67468899e-01 -1.30637765e+00 -1.07001781e+00 1.85080752e-01
-9.03356522e-02 -4.14549679e-01 -4.33290392e-01 -8.32580686e-01
-5.94353795e-01 -5.23866832e-01 -3.19921374e-01 -4.74092573e-01
5.80197453e-01 1.62332624e-01 9.08728421e-01 3.45411867e-01
-2.26770759e-01 1.16399094e-01 -3.15664649e-01 -9.11509633e-01
-1.31825948e+00 -4.81712520e-01 -2.54585475e-01 -2.34212168e-02
-2.14963034e-01 -9.04610276e-01 -3.54671061e-01 1.34138271e-01
-9.61603940e-01 8.53329360e-01 -1.11615680e-01 3.55667233e-01
-6.80828452e-01 -2.28138283e-01 5.99073708e-01 -1.05506611e+00
1.39959908e+00 -6.46641076e-01 -1.58506304e-01 -2.89743602e-01
5.07834926e-02 -2.14445427e-01 2.53638208e-01 -7.97922671e-01
-1.06042838e+00 -4.06953573e-01 3.41073364e-01 3.41165245e-01
-1.56113833e-01 3.93278837e-01 3.09213787e-01 2.25864835e-02
1.11708033e+00 -4.43374552e-02 1.68668002e-01 -5.11446834e-01
7.58227468e-01 1.08330941e+00 4.99989897e-01 -1.38762116e-01
5.79829395e-01 6.92374229e-01 -6.22436821e-01 -1.06804395e+00
-2.92735636e-01 9.83118564e-02 -3.97308737e-01 -6.81240499e-01
3.97305876e-01 3.34459357e-02 -9.18899179e-01 6.56446457e-01
-1.46175635e+00 -7.86009192e-01 -1.32762969e-01 3.14208046e-02
-3.95732403e-01 -2.54718401e-02 -8.48525047e-01 -1.26803648e+00
-2.35521391e-01 -3.18885475e-01 7.25476921e-01 2.70312582e-03
-1.09824669e+00 -8.87450814e-01 -2.59411186e-02 1.97359890e-01
7.84180105e-01 6.93335652e-01 8.00173402e-01 -6.05202913e-01
-3.61971706e-01 -5.87139666e-01 -1.62740946e-02 -5.26894987e-01
3.27981174e-01 1.04259990e-01 -6.35383487e-01 1.83117419e-01
1.51092201e-01 -2.41494656e-01 -1.73004866e-01 -4.27313715e-01
3.66498232e-01 -1.15105081e+00 -6.00206435e-01 -2.85496831e-01
8.73254299e-01 3.92454445e-01 7.62928486e-01 4.12335455e-01
4.23225254e-01 5.88328242e-01 1.34040937e-01 6.95527256e-01
-4.31315973e-02 4.91061211e-01 1.99841827e-01 2.62072325e-01
6.32528495e-03 -8.37028384e-01 2.53892303e-01 5.62246621e-01
1.25239920e-02 -4.24876034e-01 -1.03559816e+00 1.28208086e-01
-1.55694032e+00 -1.09967256e+00 -1.37338981e-01 2.34587240e+00
8.63793552e-01 4.02676105e-01 2.80712191e-02 -1.19039990e-01
8.30736518e-01 1.65925190e-01 -2.09674295e-02 -8.86688232e-01
5.83060741e-01 -4.45088714e-01 1.30719870e-01 9.22171772e-01
-5.38211524e-01 7.97040641e-01 7.46313381e+00 3.62629563e-01
-1.43513393e+00 9.57814232e-02 2.96149015e-01 -2.87823886e-01
-6.33266330e-01 -2.92662323e-01 -9.59641933e-02 6.72217846e-01
8.02594185e-01 -5.35908580e-01 6.85876131e-01 2.65453011e-01
7.69019246e-01 -4.18264449e-01 -9.61985946e-01 4.47668850e-01
2.80838430e-01 -1.66688991e+00 2.25917503e-01 -1.16892934e-01
4.47715044e-01 -2.27851197e-01 -6.23407960e-03 -2.32957646e-01
9.00345683e-01 -1.02135181e+00 1.10337257e+00 5.06279111e-01
5.55578709e-01 -4.11079347e-01 9.63029489e-02 7.98764586e-01
5.47585711e-02 6.84972852e-02 1.78943947e-01 -7.78662205e-01
5.78036606e-01 2.53259242e-01 -1.57355368e+00 -4.00703371e-01
2.05516726e-01 -5.28784618e-02 -2.56507486e-01 8.73976529e-01
-2.46391937e-01 5.86472631e-01 -3.88866395e-01 -9.13630784e-01
-1.48054793e-01 8.93094204e-03 1.30785251e+00 1.43871093e+00
1.75610706e-01 1.96871504e-01 -2.06203789e-01 1.28407645e+00
1.99672788e-01 -1.57096665e-02 -1.01264668e+00 -5.10022163e-01
6.20756686e-01 7.28155911e-01 -8.39422226e-01 -1.99842662e-01
3.62990528e-01 1.12479830e+00 -2.84276664e-01 1.25693649e-01
-5.93685746e-01 -4.62116152e-01 2.64651656e-01 6.85037494e-01
-3.06179464e-01 -3.91616821e-01 -6.90912724e-01 -5.87618351e-01
-4.99079794e-01 -9.30207431e-01 -3.44295263e-01 -1.29047012e+00
-8.85841548e-01 2.39313871e-01 -1.13675773e-01 -6.27995372e-01
-4.16346163e-01 2.26721168e-01 -7.47828841e-01 6.08747780e-01
-5.97431883e-02 -9.74447072e-01 -1.01557992e-01 -9.67199504e-02
5.31370938e-01 2.38252759e-01 1.01249838e+00 -3.93919706e-01
1.03768587e-01 1.25122935e-01 4.96446081e-02 -1.37459919e-01
2.34499976e-01 -8.15539002e-01 9.78961945e-01 1.68600053e-01
-6.76683486e-02 7.23842263e-01 1.59248757e+00 -1.13245165e+00
-1.37812066e+00 -3.06417882e-01 9.41495717e-01 -9.77699637e-01
1.23227501e+00 -7.94158697e-01 -9.32067871e-01 6.87158465e-01
4.70198303e-01 -5.74198246e-01 3.54949415e-01 -4.77046296e-02
-3.25447023e-01 6.24522686e-01 -1.57201040e+00 1.18795013e+00
9.55156684e-01 -7.61769593e-01 -8.09716821e-01 5.63337624e-01
6.80847168e-01 -4.68120009e-01 -1.23007134e-01 -5.31958103e-01
9.85220253e-01 -9.91000772e-01 6.62037492e-01 -3.97735447e-01
6.72273576e-01 4.42053266e-02 5.18213630e-01 -1.53856683e+00
-1.57705501e-01 -1.75480962e+00 2.03584835e-01 1.18073106e+00
5.53310692e-01 -8.54243279e-01 5.97051620e-01 7.48840272e-01
3.86664063e-01 -1.96612969e-01 -4.10467535e-01 -3.26835096e-01
-7.59945959e-02 -2.62346655e-01 6.74899161e-01 1.27865541e+00
1.34844697e+00 3.76671255e-01 -4.40373808e-01 1.80804729e-02
2.05262750e-01 -3.76167685e-01 7.39513516e-01 -1.09460020e+00
-3.50910932e-01 -6.56462193e-01 -2.97696143e-01 -9.50020969e-01
-8.52233887e-01 -4.97494847e-01 3.67122352e-01 -1.12910295e+00
4.45085652e-02 -2.57847369e-01 5.21607220e-01 -2.77774427e-02
2.52762377e-01 1.65809453e-01 5.41372597e-01 4.09922302e-01
-3.07203829e-01 8.89599472e-02 1.44616055e+00 9.82706696e-02
-7.75216639e-01 -1.78656384e-01 -6.45714283e-01 1.21054029e+00
9.10906017e-01 -5.63211441e-01 -3.49129260e-01 -3.45045999e-02
9.60195720e-01 3.09247553e-01 5.48978746e-01 -7.10344017e-01
1.71779871e-01 -2.08305717e-01 1.09408736e-01 -1.32501975e-01
2.48513848e-01 -2.95422226e-01 3.01619858e-01 4.35153484e-01
-8.24540317e-01 -2.18580272e-02 2.85878867e-01 5.50029695e-01
7.37213433e-01 -2.24809162e-02 3.05539995e-01 -2.56130278e-01
1.80339009e-01 -7.43048191e-01 -1.41940856e+00 -1.42288089e-01
8.07404518e-01 -9.31841061e-02 -7.83710182e-01 -9.83344853e-01
-4.70646709e-01 -3.29446137e-01 7.49638855e-01 7.65082419e-01
6.12213314e-01 -1.03947258e+00 -5.12968898e-01 1.19725764e-01
-3.63625586e-01 -2.62583196e-01 -2.66695023e-01 2.09858656e-01
-8.77300441e-01 -9.31418687e-03 -2.03080937e-01 -3.53020951e-02
-1.20710123e+00 6.11872435e-01 2.69309223e-01 2.96390086e-01
-1.11584640e+00 8.45338821e-01 -9.35618486e-03 -5.08969128e-02
-2.01784328e-01 -1.59354001e-01 -1.65791541e-01 -3.07136714e-01
1.32414222e+00 6.54134452e-01 -5.18439949e-01 -9.52040777e-02
8.71372968e-02 -2.95106024e-01 -2.74763014e-02 -6.69371426e-01
1.12376475e+00 -4.43730205e-01 -2.13167280e-01 9.82542515e-01
6.58061266e-01 5.44622540e-01 -1.02003002e+00 2.33889788e-01
-1.00769140e-01 -6.86950028e-01 -1.18892983e-01 -1.42093921e+00
-1.99682973e-02 3.80453736e-01 8.24926198e-02 1.05015588e+00
-2.01257179e-03 -1.94398537e-02 1.04269147e+00 1.58171624e-01
6.21429026e-01 -7.38576055e-01 3.94746363e-01 7.12287426e-02
1.54882169e+00 -5.32085478e-01 5.82500324e-02 -5.43333948e-01
-6.33102655e-01 9.17498946e-01 -3.50943878e-02 2.99235377e-02
5.86210132e-01 7.80298114e-01 3.70046616e-01 -6.09157681e-01
-5.07784963e-01 8.65699470e-01 -2.13550001e-01 8.07689667e-01
4.62900698e-01 4.16842073e-01 -4.97064203e-01 2.84952343e-01
-7.28536010e-01 6.78743362e-01 1.01944137e+00 1.08383584e+00
-8.34471881e-01 -5.02050996e-01 -8.20408523e-01 5.16595356e-02
-7.17470050e-01 4.43519503e-02 -1.18532062e+00 4.57220614e-01
-9.29968730e-02 1.34115469e+00 -1.54865041e-01 -5.28914690e-01
-2.08492666e-01 4.65801060e-02 3.46610397e-01 -6.36228919e-01
-6.89764380e-01 -5.48101723e-01 6.24203026e-01 -2.79074371e-01
2.92008936e-01 -5.46402693e-01 -1.15901852e+00 -8.84008765e-01
-1.55599326e-01 1.69818208e-01 8.93536747e-01 8.96134377e-01
3.26270193e-01 -3.09528168e-02 2.54086465e-01 -1.18992877e+00
-2.83960164e-01 -9.98556495e-01 -5.39896965e-01 2.85637856e-01
4.73004997e-01 -1.72798112e-01 -2.31650263e-01 1.83938332e-02] | [9.298699378967285, 6.450888633728027] |
075519a6-be2f-4647-bb56-352b595ab3e9 | generating-multiple-length-summaries-via | 2212.10843 | null | https://arxiv.org/abs/2212.10843v1 | https://arxiv.org/pdf/2212.10843v1.pdf | Generating Multiple-Length Summaries via Reinforcement Learning for Unsupervised Sentence Summarization | Sentence summarization shortens given texts while maintaining core contents of the texts. Unsupervised approaches have been studied to summarize texts without human-written summaries. However, recent unsupervised models are extractive, which remove words from texts and thus they are less flexible than abstractive summarization. In this work, we devise an abstractive model based on reinforcement learning without ground-truth summaries. We formulate the unsupervised summarization based on the Markov decision process with rewards representing the summary quality. To further enhance the summary quality, we develop a multi-summary learning mechanism that generates multiple summaries with varying lengths for a given text, while making the summaries mutually enhance each other. Experimental results show that the proposed model substantially outperforms both abstractive and extractive models, yet frequently generating new words not contained in input texts. | ['Hwanjo Yu', 'Xing Xie', 'Chanyoung Park', 'Xiting Wang', 'Dongmin Hyun'] | 2022-12-21 | null | null | null | null | ['abstractive-text-summarization', 'abstractive-sentence-summarization', 'unsupervised-sentence-summarization'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 4.39149052e-01 4.84894454e-01 -4.75590259e-01 -2.69477665e-01
-1.28075397e+00 -4.99101728e-01 4.76484686e-01 6.18498921e-01
-3.65741223e-01 1.42352021e+00 1.03543866e+00 3.49639505e-02
2.90162545e-02 -7.09567070e-01 -5.47168732e-01 -4.98306423e-01
4.40529674e-01 3.14890683e-01 -2.79430225e-02 2.53271870e-02
9.86803114e-01 -3.68981883e-02 -1.14044213e+00 2.27099165e-01
1.66041005e+00 8.97822827e-02 4.76763129e-01 1.19132888e+00
-2.90678829e-01 8.99613857e-01 -1.18496430e+00 -3.75253469e-01
-3.29437643e-01 -8.78486872e-01 -7.61169732e-01 3.68103087e-01
3.93568456e-01 -6.54413521e-01 -4.49322522e-01 1.06300282e+00
5.49060345e-01 4.82665300e-01 9.02308702e-01 -7.76724398e-01
-9.55877662e-01 1.37319553e+00 -5.58086991e-01 2.68550694e-01
6.80755138e-01 2.81559639e-02 1.39915633e+00 -4.30987239e-01
3.75563681e-01 1.17970049e+00 2.26614252e-01 7.03016877e-01
-9.94427204e-01 -1.81193069e-01 3.99057627e-01 -9.86429304e-02
-6.39720678e-01 -6.60345733e-01 8.43847215e-01 3.70846502e-02
1.21860075e+00 4.76577997e-01 6.66569591e-01 9.38329279e-01
7.09553778e-01 1.30957007e+00 6.29673243e-01 -6.10452771e-01
4.71100509e-01 -1.60067320e-01 4.58747357e-01 4.39561874e-01
6.24364734e-01 -3.51770520e-01 -7.33228862e-01 -1.52201921e-01
1.43555254e-01 -5.63918948e-02 -1.84524938e-01 1.14047460e-01
-1.00045311e+00 1.00635564e+00 -2.38353983e-01 9.00411382e-02
-7.93137252e-01 2.99873799e-01 5.81452250e-01 1.27764344e-01
7.75180757e-01 6.64785922e-01 -6.04588352e-02 -2.84374982e-01
-1.40328419e+00 5.35799563e-01 8.72193694e-01 1.10457754e+00
6.84230983e-01 3.81967455e-01 -7.41836071e-01 5.88783801e-01
1.35996521e-01 6.16560817e-01 7.59385705e-01 -9.91623580e-01
8.49186122e-01 4.52998191e-01 3.03447992e-01 -8.92782271e-01
-1.50079414e-01 -3.94945651e-01 -8.49994898e-01 -4.55800980e-01
-4.02451605e-01 -6.64154649e-01 -6.26400054e-01 1.44642603e+00
-2.18375027e-01 -1.55595437e-01 6.25383198e-01 4.09018099e-01
9.96674895e-01 1.17385840e+00 -1.85499992e-02 -9.41183150e-01
6.71010971e-01 -1.41113627e+00 -1.32907355e+00 -3.70293021e-01
5.33078969e-01 -5.19338071e-01 8.99404585e-01 4.36109930e-01
-1.64365554e+00 -4.19111401e-01 -1.24468601e+00 9.69150290e-02
2.01468989e-01 2.88588166e-01 2.03265131e-01 4.52416360e-01
-1.14803886e+00 8.74000490e-01 -8.42557073e-01 -2.54044026e-01
2.74558514e-01 2.25789517e-01 1.45825401e-01 2.64001757e-01
-1.01536679e+00 8.20261419e-01 7.73932159e-01 -2.51023799e-01
-7.47844517e-01 -2.76527584e-01 -9.48113680e-01 3.91561180e-01
5.47931910e-01 -1.06455779e+00 1.83454204e+00 -6.93820000e-01
-1.95918906e+00 1.36411518e-01 -3.02043974e-01 -6.39999509e-01
3.88579071e-01 -6.76993012e-01 -9.32829827e-02 4.65268910e-01
3.79916847e-01 6.20806694e-01 8.63572896e-01 -1.18793738e+00
-9.23487902e-01 -4.71383007e-03 -1.71047777e-01 5.97380102e-01
-5.17023623e-01 -2.65586376e-01 1.95125174e-02 -7.07133353e-01
-4.07684565e-01 -4.12783176e-01 -6.43163204e-01 -1.05078614e+00
-1.00298631e+00 -4.12467092e-01 6.08769298e-01 -8.14504027e-01
1.77262664e+00 -1.56774187e+00 2.36346036e-01 -2.74291277e-01
2.32730910e-01 3.18366051e-01 -2.67991304e-01 1.01165020e+00
5.21958053e-01 3.06070209e-01 -4.86134410e-01 -6.16242468e-01
-2.13721674e-02 1.57310143e-01 -5.38067639e-01 -2.42679901e-02
2.55045861e-01 9.21939254e-01 -1.24813223e+00 -7.59834230e-01
-6.29587695e-02 -3.57962012e-01 -5.54801524e-01 5.53191841e-01
-5.06415844e-01 7.88504109e-02 -8.09586406e-01 2.82048676e-02
4.26318526e-01 -1.04386136e-01 1.21277019e-01 3.33499730e-01
-2.58211076e-01 4.66193914e-01 -8.62352371e-01 1.81427300e+00
-2.45894626e-01 4.36306596e-01 -3.87984037e-01 -9.47156906e-01
1.06281066e+00 3.56103331e-01 9.83782113e-02 -1.76740900e-01
6.03006594e-02 3.12626585e-02 -2.84663469e-01 -5.91636896e-01
1.47325814e+00 -1.53295577e-01 -2.63959736e-01 8.42845261e-01
8.00678283e-02 -5.61657131e-01 7.76097536e-01 7.22780108e-01
1.06614137e+00 2.86500994e-02 6.98664665e-01 -7.76415318e-02
4.61918056e-01 -1.99070834e-02 5.51148593e-01 1.16969860e+00
-6.82387948e-02 6.29279256e-01 7.36853123e-01 1.90714836e-01
-1.00546038e+00 -1.00189042e+00 6.21538937e-01 7.97691762e-01
1.33349076e-01 -7.17838824e-01 -1.07078159e+00 -8.54324818e-01
-3.03131193e-01 1.32887304e+00 -2.40673274e-01 -4.15205300e-01
-4.49510664e-01 -3.73771340e-01 5.22267401e-01 5.17832577e-01
4.68922526e-01 -1.36518884e+00 -8.63780916e-01 5.01545310e-01
-5.78864157e-01 -6.50866508e-01 -9.11231756e-01 2.21114848e-02
-1.17120433e+00 -4.77874279e-01 -8.37666690e-01 -7.69324660e-01
5.07523239e-01 3.59960347e-01 8.34855556e-01 -7.57243857e-02
4.15387213e-01 5.02505183e-01 -5.10601997e-01 -5.82236588e-01
-8.72956455e-01 7.52098799e-01 5.55666089e-02 -2.34561235e-01
-2.49886476e-02 -4.07140791e-01 -3.63565445e-01 -5.62012076e-01
-1.15587926e+00 2.95883209e-01 8.71724784e-01 8.65463555e-01
3.76345634e-01 -6.53034300e-02 1.49029052e+00 -1.10614336e+00
1.48704100e+00 -3.55093360e-01 2.01237239e-02 3.96505803e-01
-6.64790213e-01 4.04522240e-01 1.01393402e+00 -3.19190234e-01
-1.58701241e+00 -3.79370153e-01 3.45814228e-02 4.93829548e-02
-9.67331603e-02 7.91022718e-01 -4.02103849e-02 9.06416357e-01
4.84489948e-01 7.51904249e-01 -6.68115392e-02 -1.32506311e-01
5.22670329e-01 8.27356637e-01 5.79026878e-01 -4.35825288e-01
4.86095458e-01 -1.71017200e-02 -5.09225309e-01 -9.40754175e-01
-1.08182442e+00 -5.09043634e-01 -5.03644347e-01 -1.28115997e-01
6.86128616e-01 -6.27328694e-01 -1.85067296e-01 1.27270833e-01
-1.39668810e+00 -1.76488891e-01 -5.37444532e-01 5.43412745e-01
-9.89680231e-01 1.05262673e+00 -6.41229868e-01 -1.08974922e+00
-1.04198408e+00 -7.12649882e-01 1.01857352e+00 8.24736059e-01
-7.11950958e-01 -9.05617595e-01 4.30978656e-01 7.60891736e-02
1.08217940e-01 1.87477469e-01 7.74752855e-01 -9.01984751e-01
-2.07830578e-01 -1.67101577e-01 1.67957604e-01 4.64061916e-01
5.22786379e-01 2.56584913e-01 -5.01418710e-01 -2.40589097e-01
-2.53308751e-02 -4.51311767e-01 1.11695123e+00 7.02822745e-01
7.54046679e-01 -9.97571766e-01 -1.58929318e-01 -1.45881817e-01
1.12581277e+00 1.62431702e-01 4.81112450e-01 2.79651731e-01
4.11439687e-01 6.07256949e-01 8.73831689e-01 8.36239994e-01
2.68535972e-01 -3.39886934e-01 8.10365230e-02 2.65819132e-01
1.36843786e-01 -5.31051397e-01 5.73667347e-01 1.40385270e+00
1.76423877e-01 -8.92842531e-01 -3.60322982e-01 7.75878787e-01
-2.23206544e+00 -1.19828689e+00 -6.14024699e-02 1.75909686e+00
1.14321232e+00 4.44885433e-01 5.03864512e-02 -4.44816016e-02
8.43527853e-01 6.11102641e-01 -6.19453549e-01 -8.37006569e-01
-1.14420317e-01 4.19002287e-02 1.31871000e-01 6.26927793e-01
-6.99589431e-01 1.10759282e+00 6.49288559e+00 6.40471756e-01
-6.96416318e-01 -4.79481488e-01 3.84150028e-01 -1.22173145e-01
-7.49139071e-01 5.43604828e-02 -7.41827071e-01 3.60095710e-01
1.04705966e+00 -9.53169167e-01 -8.21177289e-02 6.64021373e-01
7.68107772e-01 -4.28695053e-01 -1.03245890e+00 4.86817688e-01
3.79098088e-01 -1.39760756e+00 7.56740332e-01 -3.16640496e-01
1.27706611e+00 -5.72970748e-01 -2.15141654e-01 3.97327900e-01
3.98381531e-01 -5.77840149e-01 6.92478240e-01 8.95058036e-01
4.76739168e-01 -1.15292263e+00 6.83771849e-01 8.86193335e-01
-6.61676526e-01 4.08346057e-02 -5.75176001e-01 -2.04612851e-01
4.18014288e-01 4.72713411e-01 -7.32214630e-01 9.56536829e-01
-1.46168262e-01 1.00584805e+00 -5.57020307e-01 9.79460061e-01
-5.65812469e-01 8.23760808e-01 3.19415540e-01 -6.06405735e-01
4.80920166e-01 -3.58064592e-01 9.68076825e-01 1.38494897e+00
2.95888901e-01 1.32343546e-01 4.34918702e-01 6.77429557e-01
-2.78968185e-01 3.60535890e-01 -5.27911901e-01 -2.59211719e-01
4.30767477e-01 1.09876454e+00 -6.23660982e-01 -8.20553958e-01
3.82094122e-02 1.03814876e+00 2.81405628e-01 4.27135974e-01
-6.23579681e-01 -8.31332386e-01 -2.03462750e-01 -2.99927652e-01
1.98198974e-01 -9.10748094e-02 -4.56249207e-01 -1.17513072e+00
-1.31621780e-02 -8.43169987e-01 2.26594508e-01 -7.09422529e-01
-1.03850734e+00 4.43207920e-01 1.78383887e-01 -9.91698742e-01
-4.79122192e-01 2.63763309e-01 -1.18168151e+00 5.06053567e-01
-1.54872644e+00 -6.46774471e-01 1.65654644e-01 -1.36851430e-01
1.37123036e+00 -2.90763706e-01 6.27995253e-01 -4.97647673e-01
-7.78710127e-01 3.55593294e-01 4.97151345e-01 -3.64588052e-01
8.28143775e-01 -1.53417957e+00 3.99514347e-01 1.02090251e+00
-1.57943517e-01 6.05794668e-01 1.09104931e+00 -1.19319236e+00
-1.02431953e+00 -1.16097486e+00 1.04558361e+00 3.70465815e-02
3.58725369e-01 9.20720026e-02 -8.55199397e-01 5.50887406e-01
1.05309737e+00 -1.06504166e+00 7.94090390e-01 -2.14537129e-01
3.85374635e-01 1.10065021e-01 -8.11155736e-01 8.60595226e-01
6.56734049e-01 -5.44895381e-02 -1.30370808e+00 2.30135426e-01
1.01342773e+00 -9.86955017e-02 -4.05097663e-01 1.47185428e-02
1.59559906e-01 -7.11674869e-01 1.84149981e-01 -6.38278186e-01
1.04913306e+00 1.96932685e-02 2.96626061e-01 -1.96241868e+00
-3.45477074e-01 -1.06761658e+00 -5.34938455e-01 1.54958117e+00
4.19950485e-01 -3.29631597e-01 9.01143730e-01 2.47741088e-01
-4.84361738e-01 -6.03432715e-01 -4.17811096e-01 -6.88863039e-01
1.91652328e-01 3.07204068e-01 3.92232418e-01 2.66953498e-01
5.08566320e-01 1.01136851e+00 -4.91631448e-01 -1.85375378e-01
6.11484528e-01 1.42960623e-01 7.87977219e-01 -9.16216373e-01
-8.86977315e-02 -5.99679351e-01 4.26253974e-01 -1.48318481e+00
4.85948265e-01 -6.21930122e-01 5.05053520e-01 -2.25512314e+00
6.26759887e-01 3.67440313e-01 1.18887812e-01 1.29991621e-01
-8.42123091e-01 -6.18819952e-01 4.81336042e-02 1.39058635e-01
-1.19244730e+00 1.20311022e+00 1.36459565e+00 -3.98926735e-01
-5.89762270e-01 2.30146945e-01 -1.12890589e+00 5.79133093e-01
1.15043890e+00 -5.83460391e-01 -6.33676291e-01 -2.78560817e-01
-2.09375080e-02 5.55838943e-01 -4.84825075e-01 -7.73198545e-01
4.20997918e-01 -3.65053296e-01 1.35641068e-01 -1.32576752e+00
-2.77546674e-01 7.95318484e-02 -5.35049260e-01 6.14362180e-01
-1.04103529e+00 1.95111796e-01 3.45942751e-02 9.80556726e-01
-1.86635271e-01 -8.91667366e-01 5.46255291e-01 -2.46500790e-01
-9.15435851e-02 -1.87632572e-02 -9.80554998e-01 2.68076509e-01
8.13750088e-01 -2.36302659e-01 -1.36921704e-01 -7.14970589e-01
-3.36239368e-01 5.76287150e-01 9.83126238e-02 1.31560951e-01
8.42209339e-01 -9.84470904e-01 -1.25874019e+00 -4.80238318e-01
-2.43560746e-01 2.14706704e-01 2.77993470e-01 1.79512918e-01
-4.77450550e-01 6.30136549e-01 -1.79475993e-01 -2.35401750e-01
-1.11275542e+00 4.95356411e-01 -3.50613266e-01 -9.00049806e-01
-6.50605500e-01 2.79361308e-01 9.39269066e-02 -1.31749213e-01
2.42787093e-01 -4.68371272e-01 -5.84929109e-01 2.53568292e-01
6.51090324e-01 7.74384260e-01 -1.18123345e-01 1.00896377e-02
2.85234302e-01 1.16315104e-01 -6.57935679e-01 -3.36343080e-01
1.36190796e+00 -4.92180198e-01 1.89054776e-02 6.54206038e-01
8.15215290e-01 2.63507783e-01 -1.20213270e+00 -1.59140944e-01
3.08301926e-01 -1.34407938e-01 -2.63471037e-01 -4.99979824e-01
-4.04089272e-01 6.22375488e-01 -3.89391989e-01 4.13641125e-01
8.73460293e-01 -3.45041156e-01 1.16042030e+00 7.17225730e-01
-1.58092380e-01 -1.49755275e+00 6.48379564e-01 5.30658245e-01
8.88097644e-01 -9.45386827e-01 4.04084295e-01 -2.16461718e-02
-9.27232444e-01 1.24168670e+00 7.63489425e-01 -4.26567972e-01
-2.14569777e-01 -8.65407195e-03 -3.10807794e-01 1.53135136e-01
-9.89740968e-01 7.43201301e-02 -3.09525933e-02 5.89550972e-01
3.54188919e-01 3.24453898e-02 -9.03705835e-01 8.62654626e-01
-4.37276423e-01 -1.75533131e-01 1.35927498e+00 1.06484008e+00
-9.86761928e-01 -1.00271261e+00 -3.03731799e-01 8.01084936e-01
-5.49930096e-01 -1.34830952e-01 -6.60078466e-01 2.20833600e-01
-8.07658911e-01 1.34967160e+00 -1.29563168e-01 2.19521299e-02
4.02242422e-01 1.45950820e-02 3.31082731e-01 -1.16992760e+00
-7.65672743e-01 2.26363048e-01 2.66637474e-01 3.55267107e-01
-2.82072783e-01 -7.18755007e-01 -1.41959333e+00 -1.90182403e-01
-7.28158772e-01 7.76490152e-01 1.85973555e-01 9.24985528e-01
2.10781530e-01 9.14086878e-01 9.35331285e-01 -7.56867766e-01
-1.22489452e+00 -1.25428081e+00 -5.18480778e-01 -3.19598727e-02
5.69835126e-01 -1.83312055e-02 -2.39908159e-01 1.93644717e-01] | [12.528646469116211, 9.436277389526367] |
117dddab-9cf0-4050-8c7c-5d4fbccd9a32 | euclid-towards-efficient-unsupervised | 2210.00498 | null | https://arxiv.org/abs/2210.00498v2 | https://arxiv.org/pdf/2210.00498v2.pdf | EUCLID: Towards Efficient Unsupervised Reinforcement Learning with Multi-choice Dynamics Model | Unsupervised reinforcement learning (URL) poses a promising paradigm to learn useful behaviors in a task-agnostic environment without the guidance of extrinsic rewards to facilitate the fast adaptation of various downstream tasks. Previous works focused on the pre-training in a model-free manner while lacking the study of transition dynamics modeling that leaves a large space for the improvement of sample efficiency in downstream tasks. To this end, we propose an Efficient Unsupervised Reinforcement Learning Framework with Multi-choice Dynamics model (EUCLID), which introduces a novel model-fused paradigm to jointly pre-train the dynamics model and unsupervised exploration policy in the pre-training phase, thus better leveraging the environmental samples and improving the downstream task sampling efficiency. However, constructing a generalizable model which captures the local dynamics under different behaviors remains a challenging problem. We introduce the multi-choice dynamics model that covers different local dynamics under different behaviors concurrently, which uses different heads to learn the state transition under different behaviors during unsupervised pre-training and selects the most appropriate head for prediction in the downstream task. Experimental results in the manipulation and locomotion domains demonstrate that EUCLID achieves state-of-the-art performance with high sample efficiency, basically solving the state-based URLB benchmark and reaching a mean normalized score of 104.0$\pm$1.2$\%$ in downstream tasks with 100k fine-tuning steps, which is equivalent to DDPG's performance at 2M interactive steps with 20x more data. | ['Changjie Fan', 'Yingfeng Chen', 'Jinyi Liu', 'Yujing Hu', 'Yan Zheng', 'Yao Mu', 'Fei Ni', 'Jianye Hao', 'Yifu Yuan'] | 2022-10-02 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [-5.24611883e-02 -2.95422614e-01 -3.59599382e-01 -8.55600759e-02
-5.86500883e-01 -3.16310316e-01 4.28926647e-01 4.38648351e-02
-6.52646661e-01 7.87018478e-01 -2.05207109e-01 -1.75375670e-01
-4.02971864e-01 -6.36156142e-01 -7.59085000e-01 -1.05491495e+00
-4.00718242e-01 5.34072936e-01 4.40410852e-01 -5.03367484e-01
2.83833295e-01 1.20756328e-01 -1.64798045e+00 -2.67122805e-01
1.28020585e+00 9.95316446e-01 7.62061536e-01 4.62813199e-01
2.43012145e-01 3.58771354e-01 -2.93809712e-01 2.66963184e-01
3.53492886e-01 -3.46399307e-01 -4.88981009e-01 -2.99922731e-02
-1.38457507e-01 -3.37773412e-01 -3.08075875e-01 7.37768888e-01
5.74989557e-01 6.49276972e-01 5.51433921e-01 -1.10328615e+00
-1.02805972e-01 5.85936069e-01 -5.02280414e-01 3.65378112e-01
-2.64436960e-01 9.30498660e-01 7.05847621e-01 -3.83737922e-01
3.29694748e-01 9.77246761e-01 2.39529625e-01 6.85626686e-01
-1.20654202e+00 -8.26146662e-01 5.82033396e-01 2.63394922e-01
-9.96468246e-01 -9.97214094e-02 5.49755335e-01 -3.45517337e-01
9.35259938e-01 -1.49240628e-01 8.09620559e-01 1.37723899e+00
2.36369878e-01 7.66139269e-01 1.17003584e+00 1.17698871e-01
5.59584200e-01 -2.53892779e-01 2.93707401e-02 7.67285466e-01
1.21451064e-03 6.15890145e-01 -7.80977488e-01 3.77653316e-02
8.92297924e-01 1.52962320e-02 -2.15674564e-02 -5.57403028e-01
-1.28639627e+00 6.36503994e-01 5.60342550e-01 3.46153267e-02
-3.31572980e-01 3.63789231e-01 4.42393094e-01 3.80409926e-01
1.65845640e-02 5.69964528e-01 -7.27637589e-01 -7.03479111e-01
-4.80773300e-01 5.13346553e-01 4.51367617e-01 1.01017010e+00
8.25810075e-01 3.35209429e-01 -3.51959914e-01 7.12247729e-01
1.14612915e-01 3.66940171e-01 9.53977048e-01 -1.06708217e+00
6.24360025e-01 7.09739029e-01 2.24050790e-01 -6.62716448e-01
-5.14498115e-01 -7.91688085e-01 -7.53246427e-01 3.44092995e-01
4.79424536e-01 -3.25170487e-01 -9.37245846e-01 2.01643157e+00
6.92749798e-01 2.61651367e-01 5.16051464e-02 8.72586548e-01
1.00882143e-01 5.44815958e-01 1.50819018e-01 -2.78533906e-01
1.04884160e+00 -1.36269701e+00 -3.20439845e-01 -3.81085306e-01
7.71338105e-01 -1.13374032e-01 1.51566982e+00 5.64498782e-01
-8.26798320e-01 -8.46098542e-01 -1.01957059e+00 3.21461558e-01
-1.65827662e-01 9.18640047e-02 6.22824967e-01 1.52276903e-01
-6.13308668e-01 9.82430279e-01 -1.20995617e+00 -9.17237699e-02
3.27228516e-01 5.86338222e-01 -3.92968617e-02 2.37203971e-01
-1.07926261e+00 7.30671108e-01 6.72032535e-01 -4.02191766e-02
-1.61743474e+00 -5.84295273e-01 -4.95580375e-01 1.34470940e-01
9.47597206e-01 -4.39102590e-01 1.18770587e+00 -4.83152449e-01
-1.97313023e+00 3.55729721e-02 1.86331928e-01 -6.11363173e-01
5.80061913e-01 -3.91017526e-01 -2.04595998e-02 -1.52925268e-01
8.47422034e-02 6.37735069e-01 1.11406529e+00 -9.67113018e-01
-7.72937238e-01 -3.94710928e-01 7.92699158e-02 5.53445339e-01
-6.53697133e-01 -7.58387029e-01 -5.39476573e-01 -7.27591932e-01
-1.68964997e-01 -1.26255703e+00 -5.69945931e-01 -2.79948443e-01
3.08781513e-03 -3.79376471e-01 9.13475275e-01 -2.55092293e-01
1.27456641e+00 -2.08594418e+00 5.38543940e-01 -9.00824144e-02
1.41436383e-01 2.54222453e-01 -3.08517903e-01 4.49236721e-01
4.01737958e-01 -2.68648893e-01 -4.37948227e-01 -3.10280055e-01
-2.83385515e-02 2.67267436e-01 -1.34450331e-01 1.61107466e-01
1.37434199e-01 8.28399122e-01 -1.14905620e+00 -2.97525287e-01
1.31585479e-01 -3.44917178e-02 -7.04964936e-01 3.63877475e-01
-5.55973351e-01 1.00414610e+00 -9.81066227e-01 5.57370424e-01
3.20428133e-01 -3.28062057e-01 2.40931824e-01 3.23831469e-01
-4.38666567e-02 4.18819413e-02 -1.21258748e+00 2.06385303e+00
-6.63723648e-01 1.69573098e-01 2.36849654e-02 -1.26530218e+00
1.10892868e+00 -1.59269378e-01 6.61779284e-01 -8.10268760e-01
4.77319844e-02 2.75199652e-01 3.51427883e-01 -6.80979192e-01
3.30802649e-01 1.44821033e-01 -3.40111762e-01 4.36067104e-01
1.44871116e-01 -6.82683736e-02 3.60599399e-01 -1.77135751e-01
1.17788458e+00 7.81552613e-01 8.72094408e-02 -3.85058314e-01
4.66491699e-01 1.34946346e-01 7.60268807e-01 8.34502876e-01
-3.88039261e-01 -7.35195279e-02 2.66419828e-01 -1.90608546e-01
-6.59407437e-01 -1.01779318e+00 8.03567767e-02 1.40938509e+00
5.49505532e-01 -3.25132906e-01 -5.41921437e-01 -6.97414339e-01
7.73106515e-02 5.45172334e-01 -6.32831573e-01 -6.09191298e-01
-8.11779857e-01 -7.60514498e-01 2.06122607e-01 5.44990540e-01
6.54162705e-01 -1.48129547e+00 -1.11418712e+00 4.02828842e-01
-8.83145332e-02 -8.59247386e-01 -3.04756671e-01 6.91175222e-01
-1.17434692e+00 -8.86069477e-01 -6.27757668e-01 -7.63477743e-01
5.13416052e-01 1.42423078e-01 6.01503730e-01 -2.58333050e-02
-1.72548100e-01 6.27040341e-02 -4.21175539e-01 2.82548973e-03
-9.28540230e-02 4.04625237e-01 2.96802491e-01 -2.07750514e-01
6.06438294e-02 -8.83142352e-01 -7.06563473e-01 3.92331272e-01
-6.14640296e-01 7.57246092e-02 8.12720776e-01 1.23381400e+00
6.98853374e-01 1.62164513e-02 8.31224680e-01 -6.28224909e-01
5.67682981e-01 -6.25519693e-01 -6.88179135e-01 -2.29865815e-02
-9.02648509e-01 3.95396382e-01 1.10672140e+00 -8.50785732e-01
-1.09802651e+00 -8.97654740e-04 1.75299495e-01 -4.30989116e-01
-1.29502848e-01 4.12211746e-01 4.06627543e-02 1.14941031e-01
6.78318024e-01 7.19041526e-01 2.03552291e-01 -5.81064403e-01
2.43333727e-01 3.28557581e-01 3.11353803e-01 -1.07327664e+00
8.34462225e-01 2.93097377e-01 1.12488471e-01 -5.31008244e-01
-6.06138289e-01 -3.56213629e-01 -4.00865495e-01 -1.78066388e-01
5.20255983e-01 -8.59879136e-01 -1.12359262e+00 5.52264452e-01
-3.02577257e-01 -1.06543696e+00 -3.79977375e-01 5.58275878e-01
-1.02211881e+00 2.01357305e-01 -4.54988658e-01 -7.61915326e-01
-2.71041691e-01 -1.50427556e+00 8.81512523e-01 5.34099162e-01
1.75956473e-01 -6.23181343e-01 6.94535375e-02 2.71916002e-01
5.05044997e-01 1.27325788e-01 9.70511317e-01 -4.07702386e-01
-5.70800483e-01 2.84089267e-01 1.28462270e-01 1.61298439e-01
7.75261223e-02 -4.74459350e-01 -5.26942134e-01 -6.24836981e-01
4.37382236e-02 -7.17812538e-01 8.56869459e-01 2.17294008e-01
1.50169396e+00 -1.48238182e-01 -3.80106151e-01 6.74895108e-01
1.35799575e+00 4.21185851e-01 3.50569278e-01 6.01090372e-01
5.69843650e-01 4.97903258e-01 1.19349670e+00 7.53550589e-01
2.86478907e-01 7.13799477e-01 7.11401701e-01 4.16058064e-01
2.94218004e-01 -4.19206202e-01 5.95349848e-01 6.92642033e-01
-8.01096037e-02 -2.32050400e-02 -7.08568454e-01 5.47921658e-01
-2.25283957e+00 -5.75333416e-01 1.57600358e-01 2.16162968e+00
9.79812741e-01 3.60030413e-01 3.75299662e-01 -1.67998269e-01
3.16071898e-01 6.70413002e-02 -1.34099317e+00 -5.94752766e-02
4.29350436e-01 2.06898138e-01 2.99286902e-01 1.85052425e-01
-9.83857810e-01 1.22469401e+00 4.80754375e+00 1.15558696e+00
-1.21261489e+00 -1.68838099e-01 4.80692923e-01 -1.95442304e-01
2.61092424e-01 1.60552800e-01 -1.02016783e+00 7.53120065e-01
9.01501656e-01 -1.58419475e-01 6.96437955e-01 1.09919488e+00
5.42959869e-01 -2.01185614e-01 -8.56779277e-01 6.60785735e-01
-5.86970091e-01 -9.94126379e-01 -1.31268352e-01 1.30647108e-01
8.00083101e-01 8.87802243e-02 1.71948239e-01 9.06357527e-01
3.85745913e-01 -6.95271552e-01 5.66213906e-01 4.46930736e-01
5.52601218e-01 -7.04900503e-01 2.78707474e-01 9.62036312e-01
-1.14493823e+00 -6.22296989e-01 -4.23306316e-01 -1.89877555e-01
-5.72731160e-02 8.12578872e-02 -7.56988645e-01 5.26522219e-01
8.81094933e-01 9.05962646e-01 -4.58001256e-01 8.37770164e-01
-1.08037755e-01 5.47220230e-01 -2.57046640e-01 -4.79750365e-01
4.90728289e-01 -4.29521441e-01 7.34910309e-01 6.91005707e-01
3.62046063e-01 -1.27336979e-01 5.97672522e-01 5.86933970e-01
2.55401313e-01 1.13194082e-02 -3.63207221e-01 -4.74266708e-02
4.41725105e-01 1.19295526e+00 -6.23996377e-01 -1.86464489e-01
1.57037303e-01 7.18582988e-01 6.51836872e-01 3.48794729e-01
-1.03715491e+00 -2.27403626e-01 6.18100345e-01 -6.57362789e-02
6.49522960e-01 -5.58159113e-01 -2.37469841e-02 -1.09609640e+00
3.01684979e-02 -9.69338000e-01 2.42815465e-01 -2.37639353e-01
-1.05446982e+00 3.52553606e-01 1.28919736e-01 -1.61264777e+00
-3.99331063e-01 -4.49859768e-01 -4.95325178e-01 5.87282300e-01
-1.57418704e+00 -7.73104191e-01 -5.40416241e-01 5.82873881e-01
8.32947552e-01 -2.80708790e-01 6.07752979e-01 -5.65373525e-02
-7.91525602e-01 4.72568482e-01 3.44384253e-01 -4.36326981e-01
6.22344077e-01 -1.29041398e+00 1.33107811e-01 5.89598000e-01
-3.39731097e-01 6.56123400e-01 5.11526406e-01 -5.68801343e-01
-1.60075796e+00 -1.03962815e+00 3.23554091e-02 -1.12596475e-01
6.83512926e-01 -2.29012102e-01 -8.32535267e-01 2.42494643e-01
-1.72967523e-01 -9.60471034e-02 3.71959716e-01 -2.77958135e-03
1.90095797e-01 -3.52736115e-01 -7.94955492e-01 8.97127986e-01
1.41699910e+00 1.15874283e-01 -2.84562886e-01 1.81610778e-01
8.16228211e-01 -4.70729649e-01 -9.40969586e-01 5.13889372e-01
4.37658727e-01 -7.75865376e-01 8.28323960e-01 -7.29282379e-01
5.15865445e-01 -3.04051578e-01 2.23951399e-01 -1.34694123e+00
-2.71916747e-01 -8.27354372e-01 -5.25105238e-01 8.66367340e-01
2.21727595e-01 -6.95995808e-01 9.16878700e-01 1.76728502e-01
-3.89978290e-01 -1.33800137e+00 -8.71064067e-01 -9.66551483e-01
1.12809852e-01 -2.45376378e-01 4.50836033e-01 5.49470961e-01
-4.23466116e-02 3.19431543e-01 -5.07433712e-01 -1.84220552e-01
5.16552925e-01 3.98913532e-01 1.16122472e+00 -9.73781943e-01
-7.58024156e-01 -4.69770491e-01 8.25000033e-02 -1.58460462e+00
1.22668706e-01 -6.81762874e-01 3.71056765e-01 -1.17573583e+00
-6.35194853e-02 -8.06828618e-01 -4.63153392e-01 4.77785200e-01
-4.14064437e-01 -4.80333149e-01 1.72060668e-01 3.15561593e-01
-6.41826689e-01 1.12813270e+00 1.73345697e+00 -5.02627268e-02
-5.85567474e-01 1.91694826e-01 -7.12014854e-01 5.60783565e-01
9.11583185e-01 -5.80428123e-01 -8.88754308e-01 -3.90533298e-01
-2.50993907e-01 2.87635148e-01 1.59323007e-01 -1.28476763e+00
1.20978437e-01 -6.23030186e-01 2.10659459e-01 -3.35392714e-01
4.05389309e-01 -6.40379965e-01 -2.84746885e-01 9.05111849e-01
-4.73651439e-01 -1.17373876e-02 3.50632966e-01 1.23518598e+00
1.63380980e-01 -1.55125991e-01 6.95205867e-01 -1.81185722e-01
-1.11311376e+00 5.73118091e-01 -3.17268580e-01 2.06383392e-01
1.24387157e+00 -1.38579935e-01 -2.59759724e-01 1.83511116e-02
-8.46740544e-01 7.91826010e-01 3.43002111e-01 5.31736255e-01
4.48313266e-01 -9.94045794e-01 -3.30909282e-01 2.27958173e-01
1.52253702e-01 1.83179259e-01 4.93026316e-01 8.57619822e-01
-1.84724942e-01 1.18342519e-01 -6.10114694e-01 -6.84803367e-01
-6.81120098e-01 4.42534596e-01 1.35841504e-01 -9.00816023e-01
-7.17546284e-01 7.58854091e-01 1.73964992e-01 -4.28299844e-01
1.87775254e-01 -3.00620735e-01 -4.05252308e-01 -2.62990385e-01
7.79541060e-02 5.70130587e-01 -1.80512056e-01 -7.83222690e-02
-3.24643701e-01 3.93373668e-01 -7.50230774e-02 -2.26444565e-02
1.37529373e+00 -5.46908863e-02 5.01664996e-01 5.24340153e-01
9.48721647e-01 -4.75738168e-01 -2.01445389e+00 -1.13797180e-01
-5.32608759e-03 -3.12783778e-01 -7.63248876e-02 -6.83893681e-01
-8.83100271e-01 9.09553289e-01 7.08330274e-01 -6.80669993e-02
1.10247743e+00 -4.40102190e-01 1.00939023e+00 7.53518820e-01
7.91130066e-01 -1.45959067e+00 6.38987899e-01 6.44973338e-01
7.62781680e-01 -1.31583869e+00 -1.68945864e-01 8.57493952e-02
-8.27826917e-01 1.01744473e+00 1.35224688e+00 -4.58782315e-01
5.30759573e-01 -3.36494446e-02 -3.77704740e-01 6.03814721e-02
-1.03892219e+00 -2.63651997e-01 3.58221643e-02 5.62754512e-01
-1.38517782e-01 3.61720920e-02 -5.37629724e-01 6.30848944e-01
1.14322929e-02 -5.24929799e-02 1.56331938e-02 1.24334061e+00
-5.70876777e-01 -1.24693322e+00 2.39782199e-01 5.66384375e-01
3.63821574e-02 2.57915378e-01 7.03692883e-02 8.57217491e-01
-5.59530640e-03 6.27991557e-01 -2.48167768e-01 -5.94414532e-01
1.73931733e-01 -1.88882962e-01 4.11514997e-01 -6.23303831e-01
-7.66951680e-01 1.89924687e-01 -1.74538448e-01 -7.46950746e-01
-1.01367310e-01 -5.62770188e-01 -1.43222702e+00 4.24206816e-02
-2.48398051e-01 1.53639317e-01 3.16740036e-01 1.00429797e+00
7.70238698e-01 6.93510950e-01 7.70067036e-01 -9.75555003e-01
-1.21416748e+00 -1.04789972e+00 -4.79022861e-01 3.47912788e-01
2.26519868e-01 -1.21390688e+00 -2.19839573e-01 -9.16459262e-02] | [4.201882362365723, 2.0226352214813232] |
eb224d10-782c-4de5-b873-800dba64c0bb | document-level-time-anchoring-for-timeline | null | null | https://aclanthology.org/P15-2059 | https://aclanthology.org/P15-2059.pdf | Document Level Time-anchoring for TimeLine Extraction | null | ['German Rigau', 'Egoitz Laparra', 'Itziar Aldabe'] | 2015-07-01 | document-level-time-anchoring-for-timeline-1 | https://aclanthology.org/P15-2059 | https://aclanthology.org/P15-2059.pdf | ijcnlp-2015-7 | ['temporal-information-extraction'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.406285762786865, 3.6772236824035645] |
4b363637-9505-4ca5-8a77-afe662a79382 | exploring-current-user-web-search-behaviours | 2104.04501 | null | https://arxiv.org/abs/2104.04501v1 | https://arxiv.org/pdf/2104.04501v1.pdf | Exploring Current User Web Search Behaviours in Analysis Tasks to be Supported in Conversational Search | Conversational search presents opportunities to support users in their search activities to improve the effectiveness and efficiency of search while reducing their cognitive load. Limitations of the potential competency of conversational agents restrict the situations for which conversational search agents can replace human intermediaries. It is thus more interesting, initially at least, to investigate opportunities for conversational interaction to support less complex information retrieval tasks, such as typical web search, which do not require human-level intelligence in the conversational agent. In order to move towards the development of a system to enable conversational search of this type, we need to understand their required capabilities. To progress our understanding of these, we report a study examining the behaviour of users when using a standard web search engine, designed to enable us to identify opportunities to support their search activities using a conversational agent. | ['Gareth J. F. Jones', 'Abhishek Kaushik'] | 2021-04-09 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 1.23725846e-01 4.08447355e-01 -1.20758168e-01 -2.00517908e-01
-2.19079643e-01 -8.39457214e-01 1.14389729e+00 3.99046503e-02
-6.73421264e-01 4.98376310e-01 3.39979738e-01 -9.43058491e-01
-2.81568378e-01 -4.86040652e-01 3.71334612e-01 -2.06954610e-02
2.49099955e-01 7.06681013e-01 4.78459567e-01 -4.79765475e-01
6.97394609e-01 6.47446394e-01 -1.73915708e+00 3.75737578e-01
7.93899417e-01 1.89625666e-01 7.71559656e-01 7.10925996e-01
-4.62390304e-01 7.80180752e-01 -7.80930698e-01 -4.11569998e-02
4.02087010e-02 -4.78246540e-01 -1.19273961e+00 2.66294670e-03
-8.03839788e-02 -5.98308444e-01 -2.07752869e-01 4.95755017e-01
5.21175265e-01 5.62230527e-01 5.38644075e-01 -1.16823506e+00
-8.94603655e-02 1.64839849e-01 4.49171782e-01 6.44488454e-01
1.10576177e+00 1.33462682e-01 7.85298884e-01 -2.59160131e-01
5.74318230e-01 1.15492988e+00 1.82556182e-01 5.28230667e-01
-9.19964135e-01 -4.22956944e-01 -6.85776994e-02 1.71830103e-01
-1.03564918e+00 -8.12261164e-01 4.61637735e-01 -2.57044435e-01
1.44839168e+00 7.44066179e-01 7.16721058e-01 8.06492805e-01
-9.63315219e-02 5.33272922e-01 9.62332666e-01 -1.08231831e+00
5.24597988e-02 7.25238383e-01 4.14026916e-01 6.26282275e-01
2.25512221e-01 -2.01931924e-01 -5.87832630e-01 -6.81679666e-01
4.40355420e-01 -2.24361196e-01 -3.22588086e-01 -1.05673619e-01
-8.40476513e-01 6.80613935e-01 1.01752535e-01 1.06466448e+00
-4.88117397e-01 -3.78554434e-01 9.89680216e-02 5.34632385e-01
2.54237473e-01 1.15650916e+00 -1.66046783e-01 -7.70671248e-01
-2.08711877e-01 3.74113441e-01 1.78117955e+00 7.81721175e-01
4.21072632e-01 -6.01786435e-01 -1.30251069e-02 1.03476369e+00
4.64755952e-01 1.24238074e-01 6.81264162e-01 -9.64133024e-01
1.73154861e-01 1.07803953e+00 4.23939705e-01 -7.91545272e-01
-3.69335860e-01 -5.57540283e-02 2.86990702e-01 2.08844975e-01
3.65326792e-01 -1.29903063e-01 -1.66068450e-01 1.02948177e+00
1.68325469e-01 -6.50407016e-01 4.26861569e-02 6.12425208e-01
5.38733244e-01 3.75512153e-01 4.69586141e-02 -5.23576736e-01
1.60742390e+00 -7.05707490e-01 -6.09678447e-01 -3.13748449e-01
1.16058612e+00 -1.01650298e+00 1.21378863e+00 -4.80033047e-02
-1.15085173e+00 -4.71766144e-02 -7.68715441e-01 9.21963230e-02
-6.07573748e-01 -5.91365576e-01 8.46617758e-01 9.10846591e-01
-1.06382656e+00 1.36803135e-01 -6.78608298e-01 -9.97845113e-01
-6.70857251e-01 3.74581397e-01 1.90233905e-02 1.49114832e-01
-1.28233266e+00 1.47125590e+00 6.24888390e-02 -7.96391964e-02
1.11918047e-01 1.40614480e-01 -5.25304854e-01 3.17057133e-01
3.77839655e-01 -7.86599040e-01 1.85365355e+00 -7.19720125e-01
-1.07013762e+00 7.00136602e-01 -3.09075087e-01 3.84011827e-02
2.71841943e-01 2.53396630e-01 -2.47050509e-01 1.19402349e-01
3.51869106e-01 2.79985219e-01 6.90416768e-02 -1.06979775e+00
-9.96759057e-01 -2.16976330e-01 4.35274869e-01 9.26461101e-01
-4.15698767e-01 4.67904925e-01 -5.27395844e-01 5.06741814e-02
-3.75540018e-01 -1.07557034e+00 8.96081924e-02 -4.04690206e-01
2.27016956e-01 -8.39409530e-01 8.59182537e-01 -5.80806255e-01
1.35675395e+00 -1.66475773e+00 -2.73209244e-01 4.15478259e-01
2.97890953e-03 5.07037222e-01 1.03134133e-01 9.85452235e-01
4.66461837e-01 3.21603805e-01 6.88579261e-01 1.33711118e-02
2.86110163e-01 2.01937571e-01 3.58950227e-01 -1.64569929e-01
-5.91602743e-01 7.07889855e-01 -7.87348747e-01 -7.05925882e-01
-2.95775514e-02 2.63924390e-01 -2.85495102e-01 2.09723189e-01
-2.04312682e-01 -1.44920126e-01 -8.15704465e-01 4.70393538e-01
-2.86858886e-01 -1.86960384e-01 3.23607236e-01 4.60630894e-01
-4.71460134e-01 8.46780598e-01 -6.12893343e-01 9.34866726e-01
-8.59749854e-01 8.38239908e-01 4.15198833e-01 -5.11938870e-01
4.11973387e-01 6.70268297e-01 4.04401571e-01 -1.12168062e+00
-3.96066643e-02 2.52285302e-01 3.65424871e-01 -6.90363705e-01
5.33083498e-01 2.59364188e-01 2.18837157e-01 1.19778049e+00
-6.31117165e-01 -1.14294533e-02 3.46879154e-01 2.28092119e-01
1.41524220e+00 -3.97275358e-01 1.75687909e-01 -1.51558921e-01
4.67817843e-01 3.23263198e-01 -3.01652670e-01 1.10913754e+00
-1.06455468e-01 -4.90370512e-01 -7.56898820e-02 -3.94925952e-01
-6.73115849e-01 -3.31621140e-01 1.84560388e-01 1.38421381e+00
3.85292396e-02 -4.42140341e-01 -6.59745276e-01 -3.67456198e-01
-3.86195481e-01 1.06322789e+00 3.36812228e-01 3.97426262e-03
-3.95146936e-01 -4.00457442e-01 3.04572999e-01 -3.53507735e-02
7.09406912e-01 -1.50584519e+00 -1.05830383e+00 4.08111006e-01
-5.04118681e-01 -7.19193161e-01 -5.15096366e-01 6.12507686e-02
-6.19391263e-01 -1.08287764e+00 -4.58111554e-01 -1.04613197e+00
3.78579587e-01 7.68028796e-01 9.78810847e-01 8.18740845e-01
-2.07954034e-01 1.19268727e+00 -6.57436073e-01 -4.22061503e-01
-5.28889656e-01 5.26522875e-01 -2.54764229e-01 -9.04678047e-01
6.74643278e-01 -4.56071049e-01 -7.58262217e-01 8.16890478e-01
-7.10039556e-01 7.92918503e-02 4.87113595e-01 5.58812380e-01
-6.14580452e-01 4.06905293e-01 4.28422570e-01 -5.40307462e-01
1.79144228e+00 -3.77158761e-01 -3.04126948e-01 4.10912395e-01
-9.57757950e-01 -9.45635736e-02 4.25708154e-03 -3.88464957e-01
-1.30188704e+00 -3.43753099e-01 1.68672889e-01 8.34572136e-01
-1.36553552e-02 7.49085188e-01 1.29481435e-01 -4.47613806e-01
8.78208399e-01 2.41690531e-01 3.50984216e-01 -2.86672652e-01
-1.84816852e-01 1.32658732e+00 -6.70249225e-04 -3.77578974e-01
4.00721461e-01 -1.72382027e-01 -5.27509093e-01 -1.23946559e+00
5.29878512e-02 -1.01465905e+00 6.73167687e-03 -3.51565897e-01
4.75956321e-01 -3.94500166e-01 -1.30519962e+00 1.49483308e-02
-9.95130062e-01 -7.07461059e-01 4.83252198e-01 3.54890615e-01
-5.70988595e-01 3.55716407e-01 -3.57056171e-01 -1.23860502e+00
-2.73336083e-01 -8.69874716e-01 5.55305719e-01 3.54127109e-01
-9.90233183e-01 -1.12375462e+00 1.05294123e-01 1.06849253e+00
7.60550380e-01 -6.12925589e-01 9.11761999e-01 -1.13623631e+00
-6.33908629e-01 -4.99092281e-01 -6.85692355e-02 -2.85803109e-01
4.33453202e-01 -4.93842244e-01 -5.86535096e-01 -1.98096886e-01
1.77402422e-01 -9.55569148e-02 -9.08754393e-02 -3.00245643e-01
2.90643394e-01 -5.51819026e-01 -8.45393181e-01 -3.82957011e-01
6.80251896e-01 1.10266256e+00 3.56588066e-01 8.52767944e-01
-1.90886378e-01 8.53498161e-01 5.51382542e-01 5.13795838e-02
3.18058461e-01 1.19608235e+00 -3.01013768e-01 8.79911408e-02
1.77100644e-01 6.25923043e-03 -7.15671182e-02 2.57037342e-01
-1.73881009e-01 -5.20905554e-01 -9.98732209e-01 2.75691330e-01
-1.82591760e+00 -1.14007592e+00 3.64337981e-01 1.89714837e+00
5.80937743e-01 1.03900798e-01 2.06456035e-01 -1.01201780e-01
5.52728891e-01 -1.19578894e-02 -2.47081757e-01 -6.57086551e-01
8.36733818e-01 -2.04153702e-01 2.49559328e-01 9.88659799e-01
-2.55748600e-01 5.59376776e-01 6.89225292e+00 3.69015746e-02
-5.07416070e-01 -2.55049020e-01 3.58843148e-01 3.72409374e-02
-4.47176367e-01 1.57505870e-01 -6.28530324e-01 4.72082496e-01
8.99851501e-01 -3.85996729e-01 1.05566955e+00 7.82839596e-01
4.71222669e-01 -6.43653572e-01 -1.02432108e+00 5.16829133e-01
4.85138595e-02 -9.98247564e-01 -3.02231073e-01 3.49750251e-01
-1.11253314e-01 -4.00244325e-01 -5.59263051e-01 5.92535257e-01
2.87503451e-01 -6.94800615e-01 -3.05639327e-01 3.54014665e-01
1.36789978e-02 -4.33263481e-01 5.88814497e-01 1.04598463e+00
-8.35903227e-01 -1.89748034e-02 2.97874123e-01 -6.45418942e-01
2.36828774e-01 -4.59347934e-01 -1.58286631e+00 -2.75250107e-01
3.77778590e-01 -5.76026917e-01 -5.04802346e-01 9.92713451e-01
4.92289722e-01 3.54669895e-03 -4.85021353e-01 -9.84938085e-01
1.39344588e-01 -9.40426141e-02 4.86509770e-01 9.78703499e-01
1.16298102e-01 5.86759031e-01 6.04674779e-02 2.07293496e-01
3.82573545e-01 1.26559988e-01 -8.49064946e-01 -3.42817098e-01
9.00911748e-01 9.17184651e-01 -8.72225225e-01 -2.34523177e-01
-5.78821182e-01 9.93209839e-01 -3.05602700e-02 2.99779087e-01
5.83656840e-02 -5.24969399e-01 4.32184428e-01 5.69612384e-01
-4.27482843e-01 -3.26232135e-01 8.10906738e-02 -6.82156384e-01
1.36633202e-01 -1.43629110e+00 1.77802444e-01 -9.82149959e-01
-7.24493206e-01 5.18550634e-01 4.13081974e-01 -2.10627735e-01
-1.38988614e+00 -2.54345745e-01 -7.15500712e-01 1.24626839e+00
-7.48457432e-01 -6.83730245e-01 -3.44841361e-01 4.48155403e-01
7.44924545e-01 -3.65370005e-01 9.93830383e-01 -8.27300549e-02
1.87231451e-01 1.09052844e-01 -2.37231925e-01 -2.88059235e-01
5.48830152e-01 -8.78970623e-01 1.13248184e-01 3.04119796e-01
3.74429300e-02 1.13920939e+00 8.42229903e-01 -8.00286055e-01
-1.51075017e+00 2.59227902e-01 1.53768444e+00 -5.03657818e-01
4.00956213e-01 -2.07547382e-01 -5.83059967e-01 4.76595044e-01
5.11378348e-01 -1.13244259e+00 6.53028011e-01 6.42141581e-01
3.47349077e-01 1.94268376e-01 -1.06310129e+00 1.18246758e+00
1.18399501e+00 -8.96759033e-01 -8.28001082e-01 6.34377122e-01
4.33321297e-01 7.77595714e-02 -3.40332091e-01 -7.38790780e-02
7.89934814e-01 -8.87320459e-01 1.04582679e+00 -3.10730468e-02
-4.31955308e-01 8.51705745e-02 4.39552665e-01 -1.05903172e+00
-2.19598934e-01 -1.02815831e+00 4.85213876e-01 9.92953360e-01
5.30085385e-01 -1.11379600e+00 9.23077106e-01 1.76031196e+00
1.33939879e-02 -2.05970988e-01 -7.65989661e-01 -4.24132198e-01
-6.22827709e-01 -8.76459926e-02 2.72900134e-01 7.26533175e-01
9.95555460e-01 5.36539376e-01 2.63996542e-01 -1.32796437e-01
-4.59385850e-02 -3.20362657e-01 7.80994654e-01 -1.50199032e+00
-2.63490021e-01 -6.82738006e-01 -1.12045929e-01 -1.00839055e+00
5.46816476e-02 -4.46857542e-01 1.47417327e-02 -1.81611216e+00
-7.68279359e-02 -3.27745199e-01 4.55260873e-01 3.94972056e-01
1.16087429e-01 -4.43149060e-01 2.19459906e-01 5.68408549e-01
-3.73564124e-01 -1.45753741e-01 9.56810832e-01 2.05219269e-01
-7.94995785e-01 5.50919592e-01 -1.21208096e+00 5.59550226e-01
7.59125412e-01 -1.26325026e-01 -1.11360705e+00 9.99901816e-02
4.63209301e-01 5.29785991e-01 -2.25834232e-02 -6.41663551e-01
8.88852417e-01 -3.37114573e-01 9.22074690e-02 -2.19688118e-01
3.32289457e-01 -1.04644263e+00 3.05316657e-01 5.38804471e-01
-5.25038600e-01 3.89598012e-01 9.57000405e-02 1.80927604e-01
-2.08239153e-01 -8.00311327e-01 -2.41193190e-01 -5.04004180e-01
-5.35379589e-01 -4.90209371e-01 -1.46465898e+00 -4.50439900e-01
9.92560148e-01 -5.50730765e-01 -4.41902578e-01 -1.04315448e+00
-5.64950407e-01 4.44085002e-01 5.58212340e-01 2.23747805e-01
1.28642753e-01 -8.17966163e-01 1.75665855e-01 -2.33558603e-02
1.84491888e-01 -4.11622941e-01 -4.29557443e-01 3.12764376e-01
-7.56446481e-01 1.26060581e+00 -2.36816734e-01 1.67205632e-01
-1.80404997e+00 3.56163055e-01 3.76633525e-01 -1.93455294e-01
-4.04574662e-01 3.02188963e-01 -1.56759754e-01 -1.46333352e-01
6.33351147e-01 1.73299640e-01 -3.44273180e-01 1.45879269e-01
7.05585837e-01 3.75562221e-01 3.64594497e-02 -7.14905038e-02
-3.11519772e-01 -1.97171539e-01 -2.23207250e-01 -8.52841139e-01
7.71454990e-01 -3.76005799e-01 -1.11079447e-01 7.79854283e-02
7.72880554e-01 -2.11435426e-02 -5.49513102e-01 -2.41680183e-02
3.38276923e-01 -5.73869288e-01 4.66122702e-02 -1.07033598e+00
4.61257324e-02 2.60495823e-02 2.55319238e-01 1.09573328e+00
8.68319690e-01 9.75136384e-02 3.90069395e-01 1.31148434e+00
5.41950464e-01 -1.11771178e+00 5.46243563e-02 2.94068724e-01
8.27397704e-01 -1.07629192e+00 -5.60616627e-02 -3.63172650e-01
-3.91423523e-01 1.13383710e+00 5.22432268e-01 6.06504261e-01
3.79927129e-01 -1.41526192e-01 9.27580222e-02 -5.41459024e-01
-1.05659783e+00 -3.37575048e-01 -5.09434147e-03 6.51687324e-01
7.10911214e-01 -2.76432037e-01 -1.05341566e+00 -2.83390135e-01
-1.22679345e-01 1.50189579e-01 2.88903803e-01 1.25682819e+00
-6.95843160e-01 -1.32593048e+00 -3.60110283e-01 5.75527012e-01
-2.01467082e-01 -1.39644414e-01 -1.12786996e+00 9.24565494e-01
-5.26308417e-01 1.54932141e+00 4.04246785e-02 6.24689534e-02
5.47802269e-01 5.80093503e-01 8.49242061e-02 -4.75624770e-01
-7.69151092e-01 -5.16090356e-03 1.19332659e+00 -3.60331893e-01
-3.81857157e-01 -7.70225644e-01 -5.86452425e-01 -4.45111096e-01
-5.34760416e-01 7.19547391e-01 8.73860717e-01 9.50037599e-01
1.80772319e-01 -1.10150211e-01 1.28951892e-01 -5.15485942e-01
-4.03210551e-01 -1.10967994e+00 -1.83941305e-01 1.65746272e-01
-6.50514215e-02 -4.62577492e-01 -5.16937077e-01 -2.90469795e-01] | [12.251749992370605, 7.768195152282715] |
f730a371-4971-4386-a4f1-cdc086ef2614 | unsupervised-video-summarization-with | null | null | http://openaccess.thecvf.com/content_cvpr_2017/html/Mahasseni_Unsupervised_Video_Summarization_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Mahasseni_Unsupervised_Video_Summarization_CVPR_2017_paper.pdf | Unsupervised Video Summarization With Adversarial LSTM Networks | This paper addresses the problem of unsupervised video summarization, formulated as selecting a sparse subset of video frames that optimally represent the input video. Our key idea is to learn a deep summarizer network to minimize distance between training videos and a distribution of their summarizations, in an unsupervised way. Such a summarizer can then be applied on a new video for estimating its optimal summarization. For learning, we specify a novel generative adversarial framework, consisting of the summarizer and discriminator. The summarizer is the autoencoder long short-term memory network (LSTM) aimed at, first, selecting video frames, and then decoding the obtained summarization for reconstructing the input video. The discriminator is another LSTM aimed at distinguishing between the original video and its reconstruction from the summarizer. The summarizer LSTM is cast as an adversary of the discriminator, i.e., trained so as to maximally confuse the discriminator. This learning is also regularized for sparsity. Evaluation on four benchmark datasets, consisting of videos showing diverse events in first- and third-person views, demonstrates our competitive performance in comparison to fully supervised state-of-the-art approaches.
| ['Behrooz Mahasseni', 'Sinisa Todorovic', 'Michael Lam'] | 2017-07-01 | null | null | null | cvpr-2017-7 | ['unsupervised-video-summarization'] | ['computer-vision'] | [ 4.82162654e-01 2.49007672e-01 -2.28499696e-01 -1.28315061e-01
-1.05538297e+00 -4.85708147e-01 5.09271562e-01 -5.01843765e-02
-1.71457469e-01 5.58831871e-01 7.18224764e-01 2.06007674e-01
4.97147173e-01 -3.70592922e-01 -1.36963165e+00 -9.20218468e-01
2.14553382e-02 4.18803215e-01 -2.07566097e-01 3.36317718e-01
2.97025889e-01 2.06616253e-01 -1.39279473e+00 7.23170340e-01
7.43725657e-01 8.93907368e-01 4.09542352e-01 8.83444905e-01
2.48429462e-01 1.42585194e+00 -9.77907360e-01 -2.14109525e-01
-2.14231387e-02 -9.16985095e-01 -5.67395926e-01 6.04775846e-01
6.82382286e-01 -6.09701455e-01 -1.04809999e+00 1.13777411e+00
3.47753376e-01 4.33097243e-01 8.47169638e-01 -9.76412177e-01
-7.90440738e-01 8.69424403e-01 -1.96701393e-01 6.73132464e-02
6.11502647e-01 1.83179960e-01 9.94881928e-01 -8.72570455e-01
7.36048341e-01 9.65979695e-01 2.72469878e-01 7.16497183e-01
-1.26770079e+00 -3.96448225e-01 1.13300197e-01 4.80254963e-02
-1.17048180e+00 -7.76750028e-01 9.28194284e-01 -4.56180245e-01
7.39070654e-01 3.72950792e-01 6.48408890e-01 1.57154727e+00
3.79983664e-01 1.11674953e+00 4.22638774e-01 4.07814234e-02
5.70867777e-01 -1.57963205e-02 -1.05726898e-01 6.79764748e-01
1.73371434e-02 -2.52660513e-01 -6.65189505e-01 -2.40084872e-01
4.98898715e-01 3.29354882e-01 -4.39413428e-01 -2.50865042e-01
-1.14268863e+00 7.64890552e-01 2.32635260e-01 1.01509124e-01
-7.64664412e-01 3.10704648e-01 5.67705095e-01 3.92734766e-01
6.05217636e-01 4.59755808e-01 5.87268285e-02 2.53720611e-01
-1.45722580e+00 3.33563834e-01 9.03103411e-01 7.65652359e-01
5.25025368e-01 4.90882009e-01 -6.08998895e-01 5.77514410e-01
-1.10378610e-02 4.83454019e-01 7.19071865e-01 -1.26259577e+00
6.48739159e-01 1.29044414e-01 5.12354001e-02 -1.02920437e+00
1.83681250e-01 -3.70174855e-01 -9.91668701e-01 5.02965264e-02
-1.13583244e-01 -2.74300724e-01 -9.41382289e-01 1.65256226e+00
-2.76716232e-01 6.95969343e-01 3.77779841e-01 9.65967834e-01
9.58746552e-01 1.33174789e+00 -8.37561488e-02 -4.59304482e-01
8.49376619e-01 -1.03438687e+00 -3.80953878e-01 -4.36390489e-01
1.82853192e-01 -3.81967098e-01 6.53923512e-01 4.51308370e-01
-1.48751009e+00 -6.17226005e-01 -1.18933594e+00 -3.24214026e-02
3.00302863e-01 3.60937506e-01 -1.79039478e-01 -1.01929329e-01
-1.11475074e+00 7.83584177e-01 -1.03852880e+00 -1.87903523e-01
4.77371931e-01 1.86664134e-01 -5.38056433e-01 1.70650855e-02
-9.23994362e-01 5.32715261e-01 7.09896088e-01 -1.41362220e-01
-1.70733511e+00 -4.43458349e-01 -1.11723959e+00 4.35165644e-01
4.14912611e-01 -1.01536691e+00 9.46391702e-01 -1.63644326e+00
-1.48071849e+00 7.76161015e-01 -2.30647430e-01 -9.69984770e-01
2.77176797e-01 -1.70185462e-01 -1.24483153e-01 5.90995729e-01
2.91380882e-01 5.12900054e-01 1.56897557e+00 -1.50290620e+00
-4.55215812e-01 -1.42936766e-01 -1.09696880e-01 3.46460670e-01
-1.30566373e-01 -3.40660661e-01 -4.12485272e-01 -9.92808402e-01
-1.44192070e-01 -8.56415153e-01 -2.37369105e-01 -6.11608863e-01
-9.04757440e-01 1.86388001e-01 8.99838150e-01 -1.18844318e+00
1.13701260e+00 -2.15704107e+00 9.16956186e-01 -1.01844974e-01
2.63648957e-01 4.16208595e-01 -3.90661329e-01 5.03404975e-01
-3.20537835e-01 -2.10054830e-01 -3.24554920e-01 -9.28604901e-01
-2.85392135e-01 1.19063079e-01 -8.92057419e-01 6.19811237e-01
1.48264855e-01 8.76379430e-01 -9.39476311e-01 -1.96491003e-01
2.57269353e-01 3.86267006e-01 -5.55998862e-01 7.03340113e-01
-5.90947330e-01 6.60821497e-01 -5.48906922e-01 3.57353568e-01
2.73105919e-01 -2.62908995e-01 1.56245336e-01 -2.92640150e-01
1.43582731e-01 1.34302408e-01 -8.27864408e-01 1.97436357e+00
-1.73306867e-01 6.57341897e-01 -2.13893741e-01 -1.52388549e+00
8.26855540e-01 4.64132249e-01 5.02778828e-01 -1.57953575e-01
2.10147053e-01 7.52386674e-02 -5.48586488e-01 -7.84693003e-01
6.47361636e-01 9.88568552e-03 -2.05522358e-01 5.16971648e-01
4.43986148e-01 9.78946909e-02 2.40916982e-01 5.05758941e-01
1.08350682e+00 1.01992991e-02 2.85139412e-01 8.32053050e-02
5.83604217e-01 -3.20504725e-01 3.64679188e-01 8.64247024e-01
1.87458724e-01 9.78717446e-01 7.55703330e-01 -2.25474343e-01
-1.13246381e+00 -1.02412558e+00 7.84604430e-01 9.00458634e-01
-4.31547426e-02 -4.27966088e-01 -9.85200107e-01 -7.93500483e-01
-2.36864746e-01 1.05750835e+00 -5.87465405e-01 -5.10223269e-01
-6.26061976e-01 1.71585679e-01 3.67805064e-01 2.69129783e-01
2.55159378e-01 -1.26574826e+00 -6.95365012e-01 1.85244352e-01
-4.63418216e-01 -1.04351461e+00 -7.43805587e-01 2.70924047e-02
-7.95739889e-01 -7.97044158e-01 -9.47219431e-01 -9.95984197e-01
7.33898044e-01 1.46879688e-01 1.15485442e+00 -1.45268917e-01
2.73288071e-01 4.57184553e-01 -2.44091615e-01 1.38528243e-01
-8.46133530e-01 6.25735745e-02 -1.99285001e-02 4.76709396e-01
-1.15712658e-01 -6.56703830e-01 -4.52988923e-01 -3.52563262e-01
-1.24291253e+00 1.41913071e-01 5.27911842e-01 8.25615227e-01
8.20287943e-01 -8.51430744e-02 5.83575845e-01 -9.27635729e-01
3.30652535e-01 -7.92354703e-01 -2.13272601e-01 2.31144130e-01
2.23341599e-01 1.84007123e-01 1.28628159e+00 -4.60934073e-01
-8.07313323e-01 2.06314340e-01 -9.59890410e-02 -1.08481693e+00
-1.59689933e-01 4.90128040e-01 -3.49458933e-01 4.36215222e-01
4.68856752e-01 9.28819239e-01 2.69324593e-02 -2.38023654e-01
4.40541446e-01 6.37809575e-01 1.07135963e+00 -3.39385450e-01
5.18593013e-01 5.65883994e-01 -3.49167079e-01 -1.02078784e+00
-1.05772531e+00 -1.58569112e-01 -3.09796453e-01 -2.07319334e-01
9.86950874e-01 -1.04864240e+00 -1.53039977e-01 4.47931647e-01
-1.35556436e+00 -1.69700623e-01 -6.63444281e-01 3.42615843e-01
-1.08023715e+00 3.62670183e-01 -4.30097789e-01 -4.25653219e-01
-3.49606961e-01 -1.28626871e+00 1.15896928e+00 1.71657711e-01
-1.94653779e-01 -9.65818048e-01 1.45891085e-01 3.84181023e-01
-5.60692288e-02 6.79496527e-01 7.56327152e-01 -1.09017670e+00
-6.62014365e-01 -2.99937665e-01 3.78299922e-01 8.70522797e-01
6.44680932e-02 1.10313911e-02 -6.34587049e-01 -5.99106133e-01
4.59403962e-01 -5.03605306e-01 1.43865180e+00 7.03805625e-01
1.22996175e+00 -9.75094557e-01 -2.38735780e-01 7.23413467e-01
1.23931050e+00 1.88958794e-01 7.60179639e-01 -8.90205707e-03
8.75971675e-01 2.97555745e-01 4.29655075e-01 4.45427358e-01
1.39709041e-01 1.78188533e-01 5.84498525e-01 2.76121438e-01
-1.75180901e-02 -5.92064798e-01 8.06398511e-01 8.40351939e-01
1.18607596e-01 -6.58719242e-01 -3.42346817e-01 3.38001341e-01
-2.03612494e+00 -1.42259777e+00 5.70605636e-01 2.19066215e+00
4.98582691e-01 -6.11239560e-02 1.74602538e-01 -1.56630293e-01
8.53190780e-01 9.55653787e-01 -7.40711629e-01 -1.92994386e-01
-6.42969608e-02 -5.37291653e-02 1.01886116e-01 3.75334352e-01
-1.19308531e+00 8.65364432e-01 5.25866032e+00 7.43678749e-01
-1.36373007e+00 -3.81482542e-02 8.61257076e-01 -3.53521764e-01
-3.12889189e-01 -9.96017456e-02 -2.67781734e-01 7.89432645e-01
9.76925552e-01 -1.88062936e-01 5.44005752e-01 7.73321629e-01
4.25955623e-01 1.01383179e-01 -1.35920119e+00 8.91724229e-01
7.97121406e-01 -1.67483354e+00 6.52081609e-01 -3.87409240e-01
1.11287344e+00 -1.02530025e-01 2.47143477e-01 2.52323687e-01
-7.95623660e-02 -1.03111768e+00 1.06382608e+00 9.33640659e-01
7.50356674e-01 -9.54542994e-01 4.19308841e-01 5.69562554e-01
-7.40260482e-01 -4.70708385e-02 -4.97177035e-01 2.16802254e-01
2.46872425e-01 4.60729599e-01 -3.23423773e-01 4.25336927e-01
3.32702726e-01 1.17255664e+00 -1.29271999e-01 5.83558023e-01
-2.92440087e-01 5.96691310e-01 2.62826592e-01 3.31393659e-01
3.58345717e-01 -4.17573154e-01 1.31365526e+00 1.22306454e+00
3.46540213e-01 3.40010338e-02 4.16661233e-01 7.83906996e-01
-4.85727519e-01 -2.55004704e-01 -8.44593942e-01 -3.71781081e-01
2.55715847e-01 8.91069710e-01 -5.09089351e-01 -7.87439704e-01
-1.87169537e-01 1.48344123e+00 4.19548661e-01 7.65579045e-01
-9.34705198e-01 -7.18123987e-02 4.62741494e-01 1.58385783e-01
6.55483782e-01 1.51144847e-01 2.17879161e-01 -1.54735923e+00
5.61740883e-02 -1.21330571e+00 3.79935205e-01 -7.81061113e-01
-9.62090731e-01 7.08527863e-01 -1.77083433e-01 -1.32149529e+00
-6.00199461e-01 5.07330447e-02 -9.36480761e-01 5.38002849e-01
-8.89030159e-01 -1.11529183e+00 -8.32618326e-02 6.17893577e-01
1.00127280e+00 -6.62276447e-01 4.40327555e-01 -1.32453859e-01
-5.88902414e-01 2.35817671e-01 4.23410326e-01 6.17991127e-02
5.22620678e-01 -1.06129324e+00 2.32120365e-01 1.21594775e+00
3.50389063e-01 4.43191588e-01 1.10865426e+00 -7.36968994e-01
-1.52955365e+00 -1.63112998e+00 6.36837363e-01 -5.91905825e-02
5.15776336e-01 -1.63566872e-01 -9.33062673e-01 1.11412263e+00
3.56418848e-01 -2.15729058e-01 4.91426915e-01 -7.59124696e-01
-2.08409011e-01 -9.76531673e-03 -9.68568265e-01 5.98627448e-01
5.53268731e-01 -6.00135624e-01 -9.14990604e-01 3.31666738e-01
8.04653466e-01 -4.04296130e-01 -4.60937411e-01 3.13888937e-02
3.26674759e-01 -1.17116964e+00 1.01660621e+00 -8.83318901e-01
1.18311381e+00 -1.40455872e-01 -1.93416551e-01 -1.69992435e+00
-1.13513060e-01 -7.80150652e-01 -7.47620761e-01 1.13861632e+00
-1.70387864e-01 -1.50380462e-01 8.17105114e-01 -3.84472031e-03
-3.87692809e-01 -7.17161238e-01 -6.76585615e-01 -5.16519308e-01
-3.61345470e-01 6.02154396e-02 1.01603620e-01 5.53660035e-01
-3.94992173e-01 5.89606166e-01 -9.31969166e-01 1.65327370e-01
7.82473207e-01 2.68278271e-01 8.04598093e-01 -4.66963500e-01
-6.17261052e-01 -2.22848848e-01 -3.53637934e-01 -1.38288629e+00
8.40277791e-01 -9.92644668e-01 2.04321131e-01 -1.55163038e+00
4.06649262e-01 5.10951459e-01 -1.88786119e-01 4.32579033e-03
-1.90058872e-01 9.62088816e-03 3.38054627e-01 3.51474553e-01
-8.58252585e-01 8.73531103e-01 1.03343546e+00 -5.98965704e-01
-3.24726552e-01 2.80543894e-01 -7.70251632e-01 9.02955592e-01
5.56158900e-01 -7.07728803e-01 -5.12563944e-01 -5.35066485e-01
-2.71671981e-01 8.05773199e-01 5.77389717e-01 -1.07783175e+00
2.98979551e-01 5.97788878e-02 3.68459284e-01 -7.02501118e-01
4.12302703e-01 -5.17970443e-01 2.75351614e-01 6.36199713e-01
-8.30338478e-01 -2.83275813e-01 -2.30714530e-01 7.42229342e-01
-4.27690119e-01 -4.48542058e-01 7.94333220e-01 -2.73167908e-01
-4.77103740e-01 6.03080153e-01 -4.44895536e-01 1.68057814e-01
9.13024426e-01 -1.18938856e-01 -6.40917495e-02 -8.58568668e-01
-9.05085385e-01 7.57208988e-02 4.63913977e-01 1.80542871e-01
9.39909637e-01 -1.36307311e+00 -1.11417115e+00 1.82209760e-01
-1.30597845e-01 1.16966385e-02 3.51347774e-01 4.31945801e-01
-5.94969094e-01 2.36535192e-01 -2.57151157e-01 -4.91386175e-01
-1.09816146e+00 8.45162272e-01 1.83111385e-01 -3.41670007e-01
-7.64660180e-01 6.37136638e-01 4.14852411e-01 3.11217368e-01
3.42669278e-01 -1.88189492e-01 -2.45356858e-01 9.67781767e-02
7.09883630e-01 4.00820017e-01 -3.39519948e-01 -8.34467649e-01
-1.68731920e-02 1.25178456e-01 -8.03567544e-02 -6.47674426e-02
1.46043265e+00 -1.32189125e-01 -2.17682302e-01 6.08842611e-01
1.69363165e+00 3.38419043e-02 -1.53541458e+00 -9.84755158e-02
-3.02957356e-01 -2.29955301e-01 -3.32145780e-01 -1.10623538e-01
-1.16419995e+00 5.88184714e-01 -4.54149134e-02 3.11230361e-01
1.12983823e+00 4.40408774e-02 1.03103280e+00 4.41024542e-01
3.79848741e-02 -6.70819521e-01 4.51715022e-01 3.78437579e-01
1.13819981e+00 -9.92870510e-01 -1.17890567e-01 3.35877806e-01
-1.00870860e+00 1.14641035e+00 1.58751458e-01 -9.09784317e-01
2.68022448e-01 5.94117828e-02 -3.87783766e-01 -1.13552384e-01
-8.14076185e-01 2.66394049e-01 3.96587223e-01 3.66555780e-01
-6.87123761e-02 -3.51767093e-02 2.41512701e-01 7.08541453e-01
-2.44612470e-02 -7.17148632e-02 6.43432260e-01 5.11256814e-01
-6.46555543e-01 -4.30151165e-01 -2.69359529e-01 6.65371239e-01
-5.26865721e-01 8.42743833e-03 -3.71498078e-01 2.65152127e-01
-1.11898251e-01 6.76848292e-01 3.21163267e-01 -5.19105792e-01
1.65711030e-01 -1.48284942e-01 3.22621614e-01 -8.27159226e-01
-4.82846230e-01 9.19714347e-02 -2.68631727e-01 -6.24041200e-01
-4.15624470e-01 -7.12786257e-01 -8.86951625e-01 -2.03919888e-01
2.31733084e-01 3.82245839e-01 1.86742604e-01 1.04223847e+00
1.02379851e-01 6.06078327e-01 9.67500448e-01 -1.30687582e+00
-8.00027788e-01 -7.20596433e-01 -6.12770677e-01 5.43729901e-01
8.14966381e-01 -2.81073228e-02 -5.02043009e-01 6.20844066e-01] | [10.440534591674805, 0.422097384929657] |
2d463c3f-104f-4544-a8f2-789d02d6aa27 | semi-automated-extraction-of-research-topics | 2306.13075 | null | https://arxiv.org/abs/2306.13075v1 | https://arxiv.org/pdf/2306.13075v1.pdf | Semi-automated extraction of research topics and trends from NCI funding in radiological sciences from 2000-2020 | Investigators, funders, and the public desire knowledge on topics and trends in publicly funded research but current efforts in manual categorization are limited in scale and understanding. We developed a semi-automated approach to extract and name research topics, and applied this to \$1.9B of NCI funding over 21 years in the radiological sciences to determine micro- and macro-scale research topics and funding trends. Our method relies on sequential clustering of existing biomedical-based word embeddings, naming using subject matter experts, and visualization to discover trends at a macroscopic scale above individual topics. We present results using 15 and 60 cluster topics, where we found that 2D projection of grant embeddings reveals two dominant axes: physics-biology and therapeutic-diagnostic. For our dataset, we found that funding for therapeutics- and physics-based research have outpaced diagnostics- and biology-based research, respectively. We hope these results may (1) give insight to funders on the appropriateness of their funding allocation, (2) assist investigators in contextualizing their work and explore neighboring research domains, and (3) allow the public to review where their tax dollars are being allocated. | ['John Kang', 'Paul Kinahan', 'Daniel Chen', 'August Anderson', 'Joseph Tsai', 'Peter Beidler', 'Mark Nguyen'] | 2023-06-22 | null | null | null | null | ['word-embeddings'] | ['methodology'] | [-4.08314258e-01 1.78805873e-01 -5.49797177e-01 -1.93871647e-01
-1.05461395e+00 -6.39659226e-01 4.58208382e-01 9.16093171e-01
-5.61039329e-01 7.02993989e-01 1.22563577e+00 -9.99939978e-01
-3.96246284e-01 -5.52910447e-01 -3.80589455e-01 -5.63307941e-01
-2.59654075e-01 3.98008287e-01 -1.63504750e-01 2.57542193e-01
5.03200889e-01 6.31280720e-01 -7.17087865e-01 6.13397956e-02
6.02183104e-01 3.42364937e-01 -5.78931049e-02 2.89013594e-01
-4.37242597e-01 4.01811332e-01 -6.01848900e-01 9.47806612e-02
-2.92560935e-01 -1.89168841e-01 -7.56596684e-01 -3.71617019e-01
2.72883713e-01 1.93466067e-01 -3.52642685e-01 7.15883851e-01
7.06803918e-01 -5.75448200e-02 8.01002085e-01 -7.99528956e-01
-8.57751667e-01 5.30553401e-01 -5.50671637e-01 7.31445312e-01
2.18515664e-01 2.36983225e-01 9.82044041e-01 -7.25824893e-01
1.44140136e+00 9.85666871e-01 6.40933216e-01 5.24643660e-01
-1.20763683e+00 -8.24743807e-01 -5.04622818e-04 1.83774456e-02
-1.21060991e+00 -3.29204232e-01 6.51908755e-01 -1.10516346e+00
8.02436471e-01 2.06658378e-01 8.29841912e-01 1.17645037e+00
6.08732402e-01 -1.20761760e-01 1.23514879e+00 6.44298717e-02
6.45569324e-01 2.88867801e-01 5.61701775e-01 4.15374011e-01
6.03557289e-01 -3.92047226e-01 -3.88212919e-01 -9.49684203e-01
6.19654655e-01 1.87949941e-01 -4.83753264e-01 9.43092108e-02
-1.71004975e+00 8.48555565e-01 3.75829786e-01 6.71823800e-01
-3.87066901e-01 2.74010062e-01 3.71223688e-01 -4.65252548e-02
6.23960793e-01 7.97640026e-01 -2.43748888e-01 -2.85324603e-01
-1.04147327e+00 2.64475405e-01 4.23392355e-01 3.31779838e-01
3.74172837e-01 -4.12895292e-01 -7.83873349e-02 5.03520072e-01
3.16272587e-01 9.25040152e-03 5.26062310e-01 -8.04343820e-01
-1.38099089e-01 6.84825301e-01 7.25203604e-02 -1.05000150e+00
-8.18926871e-01 -3.22906226e-01 -5.02562881e-01 -4.44904668e-03
2.75086582e-01 -2.14858770e-01 -7.26451099e-01 1.42058599e+00
2.89777994e-01 1.24353319e-01 -2.89743453e-01 8.71400237e-01
9.96990383e-01 1.79597452e-01 5.59847474e-01 3.57764885e-02
1.92976880e+00 -1.88769951e-01 -7.52730072e-01 5.60429692e-01
9.52047527e-01 -7.14027464e-01 7.71726429e-01 -3.69496718e-02
-8.17954779e-01 -2.10555866e-01 -7.50484943e-01 -1.06379554e-01
-7.85794199e-01 -2.66665518e-01 5.98014355e-01 6.41974628e-01
-1.39146471e+00 3.89546871e-01 -8.35897148e-01 -7.24630952e-01
9.45536315e-01 1.61843866e-01 -2.21247941e-01 -5.12945689e-02
-1.09400344e+00 7.78521121e-01 -3.15157771e-01 -5.96373320e-01
-8.84762585e-01 -1.52739298e+00 -4.35002565e-01 4.97263996e-03
-3.19483399e-01 -8.35530460e-01 3.73154581e-01 4.52939451e-01
-6.32697999e-01 8.71338367e-01 -3.62000912e-01 -4.92825806e-02
4.01817150e-02 3.42597723e-01 -5.53856909e-01 2.54684597e-01
5.53245187e-01 7.90234268e-01 -1.78221121e-01 -7.74477839e-01
-4.59110171e-01 -5.85238576e-01 -2.68154025e-01 1.46412635e-02
-8.42799544e-01 2.30414331e-01 -9.93021503e-02 -6.33808255e-01
-5.28788473e-03 -7.10632086e-01 -5.74935436e-01 2.79123217e-01
-2.60148019e-01 -5.95464468e-01 7.58122802e-01 -6.64110482e-01
9.94468749e-01 -1.91268885e+00 -2.14790832e-02 7.31557310e-02
9.46119964e-01 -7.10457087e-01 2.38088056e-01 4.65255231e-01
-2.65254647e-01 8.31863880e-01 2.89065212e-01 1.17378183e-01
-2.99546421e-01 -3.08267564e-01 -2.71051228e-01 9.36170042e-01
3.60828787e-02 8.55146110e-01 -1.14817214e+00 -5.80124080e-01
-7.15317056e-02 4.76613462e-01 -5.22754610e-01 -3.06267679e-01
9.20397267e-02 4.79778826e-01 -8.79740417e-01 9.28919256e-01
5.47871768e-01 -5.36856830e-01 2.72886783e-01 -2.24313468e-01
-5.04592001e-01 5.22016823e-01 -5.84989250e-01 1.57189023e+00
-1.22183397e-01 8.16929698e-01 7.56263211e-02 -9.39092100e-01
6.12576425e-01 4.82411474e-01 1.14611542e+00 -2.63799369e-01
-6.59691729e-03 -2.29536314e-02 1.20629936e-01 -3.63616496e-01
2.84015566e-01 -1.84116781e-01 -1.30941227e-01 9.97688890e-01
-3.07843298e-01 1.64391026e-01 -1.65870339e-01 5.17216682e-01
1.63508391e+00 -4.40207779e-01 -4.81701903e-02 -9.39891100e-01
-1.41738728e-01 4.25386459e-01 4.05440360e-01 5.06639600e-01
-6.19185388e-01 3.00187290e-01 7.33351529e-01 -4.48101908e-01
-1.05024529e+00 -1.02678406e+00 -7.70323932e-01 9.18105483e-01
-2.58050621e-01 -3.79768878e-01 -2.13770941e-01 -5.18793702e-01
3.55178297e-01 4.15887654e-01 -1.02619863e+00 1.95903435e-01
-2.89981991e-01 -1.03355777e+00 5.05230188e-01 2.97463834e-01
-2.91535586e-01 -4.74596828e-01 -5.87098777e-01 2.13300884e-01
-2.30561439e-02 -6.97791636e-01 -5.84607363e-01 5.53376675e-02
-9.87105668e-01 -1.20322526e+00 -1.33137906e+00 -5.59151411e-01
7.71206379e-01 2.25720733e-01 6.77768111e-01 -2.74009824e-01
-8.65187407e-01 5.98629594e-01 3.70888114e-02 -5.53315461e-01
-1.18710279e-01 -1.51165664e-01 3.11940879e-01 -7.54965723e-01
5.28830051e-01 -4.68815595e-01 -9.37763393e-01 1.68375090e-01
-6.89859092e-01 -4.85711247e-01 3.48920435e-01 5.59926808e-01
5.65780580e-01 -1.98050082e-01 7.72822440e-01 -9.00500953e-01
9.32293475e-01 -1.01402092e+00 -7.80422464e-02 5.21656051e-02
-9.49025810e-01 -1.46192759e-01 -7.15811923e-02 -2.48990491e-01
-5.05588114e-01 -6.22963727e-01 1.34085536e-01 -2.54095703e-01
-1.27984375e-01 5.90319395e-01 2.41429195e-01 -2.00434163e-01
9.62253511e-01 -1.98887452e-01 -9.86642390e-02 -4.53762591e-01
2.64064997e-01 8.17135334e-01 1.97276231e-02 -7.61855185e-01
2.34706059e-01 9.35087979e-01 -4.34521437e-02 -8.86492610e-01
-2.48803690e-01 -7.89430439e-01 -1.96161401e-02 -1.17874965e-01
1.12526286e+00 -9.34228718e-01 -7.64757693e-01 -2.94896632e-01
-9.98707175e-01 6.86798152e-03 -3.87913138e-01 7.11353421e-01
-5.33973472e-03 3.48208874e-01 -7.37016439e-01 -2.89577812e-01
-2.08253220e-01 -1.28675973e+00 9.68739688e-01 1.77809402e-01
-8.41315269e-01 -1.42867446e+00 4.89558518e-01 3.77669752e-01
6.47745073e-01 4.80551451e-01 1.33677173e+00 -4.92380619e-01
-2.08809897e-01 2.10407764e-01 -4.44304764e-01 -6.88220799e-01
5.16463578e-01 -1.06641054e-01 -8.71198773e-01 -1.41579181e-01
-1.65574700e-01 7.17680380e-02 7.12044060e-01 8.33905458e-01
1.31705701e+00 -4.98683825e-02 -1.02157438e+00 3.72321993e-01
1.20204711e+00 1.99170128e-01 3.54573160e-01 5.28463364e-01
4.55626339e-01 7.04834759e-01 -8.03340524e-02 4.42803472e-01
3.08661908e-01 4.42640752e-01 -1.30633339e-01 -3.14207911e-01
-8.19984302e-02 2.67968893e-01 5.45173250e-02 6.50713921e-01
-3.74248326e-02 1.33825541e-01 -1.35829437e+00 9.81555700e-01
-1.26856041e+00 -6.43647194e-01 -1.42072469e-01 1.73148787e+00
1.07764173e+00 1.77471578e-01 1.37672633e-01 -6.05460227e-01
6.16645515e-01 5.10135368e-02 -6.57854617e-01 -3.42878848e-01
1.29783243e-01 3.99531543e-01 6.78554475e-01 4.04407568e-02
-5.66293001e-01 5.74685574e-01 7.45074749e+00 6.47384703e-01
-1.15933299e+00 5.12341976e-01 1.15332246e+00 -2.34369904e-01
-9.54457402e-01 1.00865262e-02 -6.63990557e-01 4.19158280e-01
1.17195725e+00 -7.01774240e-01 -2.52176881e-01 6.21640921e-01
6.34678781e-01 6.82728440e-02 -1.02707148e+00 5.54581642e-01
-4.57410842e-01 -2.00112486e+00 -1.12778753e-01 7.47657180e-01
7.58053958e-01 2.22553357e-01 1.77206650e-01 -1.29989600e-02
6.18431151e-01 -1.17389822e+00 6.02065176e-02 6.50959909e-01
9.74137902e-01 -3.65552068e-01 5.78762829e-01 -3.80745202e-01
-8.66074681e-01 1.72268242e-01 -4.55454469e-01 9.81349647e-02
5.37977852e-02 1.11525381e+00 -8.95937085e-01 3.38358253e-01
8.42947543e-01 7.45126545e-01 -3.10618371e-01 9.77209389e-01
4.05957133e-01 9.55589354e-01 -1.14676647e-01 -1.70632929e-01
3.34140301e-01 1.69174790e-01 5.39542556e-01 1.49028397e+00
4.27165091e-01 1.02678582e-01 2.59525161e-02 1.06528294e+00
-1.59618959e-01 1.86771810e-01 -4.04967159e-01 -7.80110478e-01
5.09903908e-01 1.27990294e+00 -1.20565927e+00 -2.92426467e-01
-2.68744975e-01 2.97153085e-01 7.88296312e-02 4.63471293e-01
-5.68848133e-01 -5.70421398e-01 1.12832987e+00 5.25881648e-01
-1.57745838e-01 -3.07903528e-01 -7.15505660e-01 -7.35686779e-01
-5.01039505e-01 -3.94454926e-01 5.63341498e-01 -5.21125793e-01
-1.24950898e+00 1.90117121e-01 -1.49852619e-01 -8.12748313e-01
4.10602897e-01 -4.79673326e-01 -7.20987916e-01 1.08729506e+00
-1.03916276e+00 -5.71294665e-01 -1.73309688e-02 7.15018883e-02
5.37151545e-02 -1.31936222e-01 8.77363443e-01 1.69598594e-01
-4.09442842e-01 5.04062414e-01 5.14031231e-01 1.09881712e-02
8.61329734e-01 -1.02977467e+00 1.28879353e-01 -1.70716166e-01
-1.60018384e-01 1.26271093e+00 7.24097252e-01 -9.47460055e-01
-1.46501219e+00 -9.14121926e-01 8.87095451e-01 -7.81128287e-01
1.26793933e+00 -1.49246782e-01 -8.21704447e-01 3.73740524e-01
2.79843241e-01 -1.86284825e-01 1.48256612e+00 7.26261437e-01
-1.71947777e-01 2.56690323e-01 -1.20503366e+00 7.66015887e-01
7.83139825e-01 -4.41473424e-01 -5.90460658e-01 7.75768876e-01
8.34247828e-01 -1.11880086e-01 -1.54138553e+00 -1.42986268e-01
5.17077804e-01 -2.28487939e-01 7.45942891e-01 -9.61376131e-01
4.52357769e-01 4.54986095e-02 4.56944518e-02 -1.15210462e+00
-6.68024242e-01 -2.67401159e-01 3.98725271e-01 8.55634153e-01
3.76784593e-01 -6.71629906e-01 1.05311966e+00 7.52532303e-01
-3.12619030e-01 -1.25237370e+00 -1.20383000e+00 -3.68436188e-01
7.62234151e-01 -3.18384439e-01 3.84464055e-01 1.46256125e+00
6.05373979e-01 -2.66277522e-01 5.17081797e-01 -7.74617940e-02
6.21424556e-01 -7.07816631e-02 1.63828388e-01 -1.17473149e+00
2.78022792e-02 -8.92054141e-01 -7.02942789e-01 -3.25826555e-01
-1.25963822e-01 -9.87503767e-01 -6.55431092e-01 -1.89229929e+00
5.00063717e-01 -7.03655243e-01 -7.52423823e-01 5.35892069e-01
-3.87810990e-02 5.77122122e-02 -3.98762912e-01 5.91832340e-01
-3.18858087e-01 1.75368398e-01 1.01049447e+00 -4.39602911e-01
-1.33664936e-01 -6.71848238e-01 -1.49527335e+00 2.47911602e-01
5.00765324e-01 -6.43689990e-01 1.11661606e-01 -4.71927039e-02
1.29648387e-01 -2.12957337e-01 3.55394065e-01 -1.03532994e+00
3.37968826e-01 -3.34092379e-01 6.14618838e-01 -4.12734151e-01
-8.27430412e-02 -2.52533168e-01 1.87352061e-01 6.13113225e-01
-5.65687060e-01 -1.37804702e-01 5.60998082e-01 6.41383827e-01
1.42951772e-01 1.10091880e-01 3.12465549e-01 -8.42102915e-02
-7.65953138e-02 1.75322920e-01 -7.47575223e-01 9.50819626e-02
9.52343404e-01 -7.81647786e-02 -7.69021511e-01 -2.33860299e-01
-8.48628700e-01 3.37218046e-01 6.59487724e-01 2.03618050e-01
4.41160202e-01 -1.32593083e+00 -7.72693455e-01 -5.55772245e-01
1.24454014e-01 -4.90709066e-01 5.35653710e-01 9.80493784e-01
-3.34757119e-01 7.70297825e-01 -4.91142869e-02 -3.81827623e-01
-9.35681880e-01 4.09411043e-01 -8.77281427e-02 -3.45375508e-01
-6.58348799e-01 9.07037020e-01 4.29594070e-01 -2.74134517e-01
1.83443025e-01 -2.46890500e-01 -2.20758423e-01 3.61218452e-01
4.95878994e-01 5.23259163e-01 -1.33710414e-01 -4.13958222e-01
-7.49084413e-01 4.93173331e-01 -5.05183935e-01 -2.13286459e-01
1.37415564e+00 3.27419817e-01 -1.32917836e-01 5.33151031e-01
1.53996444e+00 2.42589787e-01 -4.79267806e-01 1.22730285e-01
6.61481544e-02 -1.47219701e-02 4.19126898e-01 -5.88173687e-01
-7.71030605e-01 8.47406268e-01 7.35814929e-01 7.28657246e-02
4.62538928e-01 4.84434515e-01 7.34456778e-01 -6.97144195e-02
1.17640376e-01 -1.04054415e+00 -7.80670270e-02 -1.00852847e-01
5.31969786e-01 -6.81086659e-01 4.93952841e-01 -8.02130103e-02
-1.43533945e-01 1.16201985e+00 1.36401892e-01 3.56357813e-01
1.06636810e+00 2.39158735e-01 -7.97565840e-03 -9.72616136e-01
-7.31156647e-01 2.57435888e-01 1.13588341e-01 5.21343231e-01
1.03775167e+00 2.00627804e-01 -6.70651615e-01 7.17696965e-01
-1.31673580e-02 2.70058185e-01 6.20723426e-01 9.95251119e-01
-4.76987600e-01 -9.84888673e-01 -5.43586016e-01 8.92813802e-01
-6.11461103e-01 -2.02496558e-01 -4.70330447e-01 5.19275963e-01
1.97735056e-01 6.54940546e-01 1.78504109e-01 -2.36728266e-01
-1.02622800e-01 4.72077392e-02 -9.53636616e-02 -6.84703171e-01
-3.34175140e-01 1.77215874e-01 -5.64609580e-02 -3.11048120e-01
-2.02469721e-01 -8.66961837e-01 -1.27524722e+00 -2.51846910e-01
1.00748546e-01 5.43784738e-01 9.25052047e-01 5.35676956e-01
1.04580212e+00 9.18063939e-01 1.95342168e-01 -3.87339234e-01
5.48514836e-02 -7.92548478e-01 -5.44638395e-01 6.26825616e-02
-9.33234692e-02 -7.20559597e-01 -4.71634686e-01 -1.10529386e-01] | [9.07591724395752, 8.274869918823242] |
421930a3-6e52-46f6-a9e4-4195781fd84a | unsupervised-temporal-video-grounding-with | 2201.05307 | null | https://arxiv.org/abs/2201.05307v1 | https://arxiv.org/pdf/2201.05307v1.pdf | Unsupervised Temporal Video Grounding with Deep Semantic Clustering | Temporal video grounding (TVG) aims to localize a target segment in a video according to a given sentence query. Though respectable works have made decent achievements in this task, they severely rely on abundant video-query paired data, which is expensive and time-consuming to collect in real-world scenarios. In this paper, we explore whether a video grounding model can be learned without any paired annotations. To the best of our knowledge, this paper is the first work trying to address TVG in an unsupervised setting. Considering there is no paired supervision, we propose a novel Deep Semantic Clustering Network (DSCNet) to leverage all semantic information from the whole query set to compose the possible activity in each video for grounding. Specifically, we first develop a language semantic mining module, which extracts implicit semantic features from the whole query set. Then, these language semantic features serve as the guidance to compose the activity in video via a video-based semantic aggregation module. Finally, we utilize a foreground attention branch to filter out the redundant background activities and refine the grounding results. To validate the effectiveness of our DSCNet, we conduct experiments on both ActivityNet Captions and Charades-STA datasets. The results demonstrate that DSCNet achieves competitive performance, and even outperforms most weakly-supervised approaches. | ['Pan Zhou', 'Zichuan Xu', 'Yu Cheng', 'Kai Zou', 'Xing Di', 'Yinzhen Wang', 'Xiaoye Qu', 'Daizong Liu'] | 2022-01-14 | null | null | null | null | ['video-grounding'] | ['computer-vision'] | [ 1.87083676e-01 -1.28698960e-01 -4.25268859e-01 -4.13508713e-01
-7.12551117e-01 -2.88461864e-01 4.21738803e-01 -1.03989944e-01
-3.34148049e-01 4.36897904e-01 2.38093376e-01 -9.30734426e-02
-9.40660387e-02 -5.70439219e-01 -8.47859204e-01 -5.61957955e-01
1.23749189e-01 9.70245227e-02 5.19792974e-01 -8.77969041e-02
8.79166722e-02 -1.13418244e-01 -1.49520433e+00 3.86454821e-01
9.94279921e-01 1.16715002e+00 5.73333740e-01 1.55701324e-01
-1.93661839e-01 1.15241277e+00 -5.23837924e-01 -2.43873432e-01
1.09090634e-01 -6.81614578e-01 -8.28031957e-01 5.26575804e-01
3.33136767e-01 -3.81865323e-01 -6.31943583e-01 1.20815885e+00
8.26943517e-02 2.55961418e-01 7.32054561e-02 -1.47300065e+00
-3.70628685e-01 7.67383099e-01 -4.42948073e-01 3.75925273e-01
4.80574936e-01 2.02760652e-01 1.17289555e+00 -7.72327781e-01
6.60972714e-01 1.00561666e+00 1.20082654e-01 3.03374648e-01
-7.18871415e-01 -6.86619997e-01 7.40723491e-01 4.92827833e-01
-1.46924388e+00 -3.76077294e-01 1.02762794e+00 -2.47823969e-01
5.76718152e-01 4.09642123e-02 8.83471191e-01 1.18546295e+00
-3.52612793e-01 1.24135542e+00 7.22982168e-01 -4.07688357e-02
2.06950560e-01 -1.48177713e-01 1.47709906e-01 9.55034137e-01
-5.91407083e-02 -4.77600902e-01 -8.60811710e-01 1.85671687e-01
5.51661849e-01 2.37876907e-01 -4.83047187e-01 -2.88890690e-01
-1.37658882e+00 6.31543577e-01 4.49343681e-01 5.33568621e-01
-4.00256872e-01 2.60913879e-01 2.39329651e-01 1.82409793e-01
5.74853778e-01 2.29561254e-01 -4.13373023e-01 -1.55104190e-01
-1.10388231e+00 9.79919955e-02 6.46044374e-01 1.11945212e+00
8.74794662e-01 -2.00037509e-01 -2.79123724e-01 5.50871551e-01
3.71270925e-01 2.44622871e-01 4.56040531e-01 -1.00481522e+00
7.67616510e-01 8.74859810e-01 1.13896905e-02 -1.25104785e+00
-2.04739213e-01 -4.69442129e-01 -4.79281545e-01 -5.08935630e-01
3.33544165e-01 -9.68095437e-02 -7.91829705e-01 1.66320848e+00
2.74585396e-01 7.62283206e-01 -2.42423974e-02 1.25366426e+00
9.39319313e-01 7.43975699e-01 1.59266874e-01 -3.38046640e-01
1.28095853e+00 -1.28712177e+00 -6.66658700e-01 -2.12398604e-01
7.60931015e-01 -4.74710196e-01 1.15223849e+00 3.84360671e-01
-6.34412348e-01 -4.93959606e-01 -1.07152534e+00 -7.32013315e-04
4.49006557e-02 1.44880742e-01 5.64229488e-01 8.45185667e-02
-7.53840029e-01 3.94602001e-01 -1.02820551e+00 -5.62939644e-01
5.13698459e-01 6.36770204e-02 -2.57096112e-01 -3.17097157e-01
-1.20532453e+00 2.96097606e-01 6.35380983e-01 1.49744794e-01
-1.10557711e+00 -4.23261821e-01 -9.10759568e-01 -1.41220689e-02
1.17563403e+00 -5.88816881e-01 1.05764163e+00 -1.33366847e+00
-1.17159784e+00 7.61509657e-01 -3.79630804e-01 -5.98819971e-01
4.49434876e-01 -3.14143717e-01 -4.68857408e-01 6.19393051e-01
3.67307156e-01 6.84550405e-01 6.72420025e-01 -1.08539546e+00
-7.73488879e-01 -3.16458672e-01 4.67629910e-01 4.02469724e-01
-4.48601425e-01 8.07644278e-02 -1.22467339e+00 -7.64368236e-01
3.06167305e-01 -7.86300898e-01 -1.53871283e-01 -1.92590550e-01
-5.57331145e-01 -3.43533933e-01 9.62127447e-01 -7.24761486e-01
1.36622691e+00 -2.25684810e+00 2.49741390e-01 1.05921507e-01
3.59904408e-01 8.49619806e-02 -1.17474414e-01 1.70026228e-01
2.83540040e-01 4.52158302e-02 -1.42449483e-01 -3.18383425e-01
-9.79512036e-02 3.01679254e-01 -2.17880875e-01 3.73691678e-01
1.80645153e-01 1.03349054e+00 -1.16064239e+00 -7.78346717e-01
7.26806298e-02 2.32754976e-01 -6.59775138e-01 3.22972119e-01
-6.87649548e-01 4.73895848e-01 -8.40658844e-01 7.34630883e-01
1.89964578e-01 -6.18272126e-01 2.18887702e-01 -4.03682739e-01
8.56665969e-02 1.40153572e-01 -9.87937152e-01 2.26712084e+00
-2.00807318e-01 4.86913592e-01 -3.70531753e-02 -1.38100886e+00
6.93827987e-01 1.65266976e-01 8.62178445e-01 -6.19738996e-01
1.07587710e-01 8.61404613e-02 -2.39616424e-01 -7.30938733e-01
4.08691376e-01 7.72566870e-02 -1.15424223e-01 2.92321175e-01
3.41699243e-01 4.35223848e-01 3.45593333e-01 5.77439845e-01
1.13477683e+00 2.86181211e-01 4.05975524e-03 -6.88539743e-02
5.88851452e-01 7.27053210e-02 7.74954200e-01 5.04291475e-01
-2.90119320e-01 6.05981946e-01 5.79165280e-01 -3.18151206e-01
-5.31410515e-01 -7.88260877e-01 4.77553219e-01 1.25892282e+00
8.51099610e-01 -8.39979529e-01 -8.89731050e-01 -9.54939842e-01
-4.51034814e-01 3.41396958e-01 -4.12458658e-01 -1.22069500e-01
-4.94396776e-01 -6.28245354e-01 2.98420876e-01 4.73185450e-01
8.62417102e-01 -9.46292102e-01 -3.93388480e-01 1.64515227e-01
-7.48248637e-01 -1.63099718e+00 -6.39481485e-01 -1.81473762e-01
-6.71431899e-01 -1.29736114e+00 -5.06206036e-01 -9.17330086e-01
5.64247668e-01 6.62613630e-01 9.10235763e-01 3.89323682e-01
1.28318503e-01 3.01131725e-01 -7.14817882e-01 -4.43610363e-02
5.45313507e-02 1.79599419e-01 -2.13973865e-01 4.82465208e-01
6.09081447e-01 -2.80601978e-01 -7.59501815e-01 4.79048014e-01
-7.88164020e-01 3.45033318e-01 4.36692476e-01 5.14172256e-01
6.80596530e-01 2.51765907e-01 2.97558069e-01 -6.27003670e-01
2.13541076e-01 -6.34129286e-01 -4.44238961e-01 2.20542505e-01
-2.66576290e-01 -1.63154542e-01 5.02756476e-01 -1.91230044e-01
-9.25488174e-01 1.59982964e-01 3.85981351e-02 -8.57198417e-01
-2.06402138e-01 6.94820046e-01 -5.81399441e-01 3.07444155e-01
2.16823786e-01 4.26882327e-01 -1.86444089e-01 -4.84353095e-01
2.00579792e-01 4.62818772e-01 6.25707269e-01 -5.83133638e-01
7.61824012e-01 6.42193854e-01 -3.55258375e-01 -6.79721296e-01
-1.38308394e+00 -8.21063340e-01 -4.69679177e-01 -3.13546717e-01
1.22690618e+00 -1.16310012e+00 -4.14208770e-01 3.09796423e-01
-9.79770660e-01 -3.18034172e-01 3.41038313e-03 4.66549903e-01
-5.54299712e-01 5.74408352e-01 -4.52738613e-01 -5.30577183e-01
1.34114968e-02 -1.18803668e+00 1.22843099e+00 1.32960722e-01
1.92725714e-02 -6.32073283e-01 -3.55497003e-01 7.61479497e-01
-2.22510606e-01 1.36227116e-01 2.51381159e-01 -7.43781805e-01
-1.09241259e+00 2.48128772e-02 -3.23760509e-01 1.89783990e-01
4.96738441e-02 -1.43880293e-01 -6.98475599e-01 -1.27286807e-01
5.99566475e-03 -2.22471654e-01 1.09797752e+00 1.94438353e-01
1.48441410e+00 -3.99876297e-01 -3.56988162e-01 8.23002338e-01
1.10976124e+00 1.68851107e-01 4.88637805e-01 3.87145191e-01
1.08870578e+00 6.05559170e-01 1.03308594e+00 4.01973099e-01
5.53080916e-01 7.28636444e-01 5.80906391e-01 -8.57267007e-02
3.07038277e-02 -5.77761233e-01 5.41449726e-01 8.14500272e-01
-9.44629163e-02 -3.98255527e-01 -7.14410186e-01 5.15760362e-01
-2.22781706e+00 -1.19815171e+00 1.54426768e-01 1.72817934e+00
4.94756579e-01 1.11552447e-01 2.17520878e-01 1.10118456e-01
8.58152390e-01 4.20854092e-01 -4.42886800e-01 6.06936932e-01
-1.87207341e-01 -2.50228941e-01 2.00056762e-01 1.75010368e-01
-1.38184512e+00 1.18758106e+00 4.51107073e+00 9.29132938e-01
-8.91914248e-01 3.03058952e-01 6.76839828e-01 -2.97130257e-01
-2.49186948e-01 9.98466164e-02 -6.35056257e-01 7.80691266e-01
5.90122700e-01 -1.67009234e-01 3.14715266e-01 8.21417749e-01
5.05362570e-01 -9.36544966e-03 -9.64271426e-01 1.15497220e+00
3.39785576e-01 -1.20307291e+00 1.04021914e-01 -9.89015549e-02
6.47700191e-01 1.03132166e-01 -2.62553602e-01 2.45260805e-01
6.41462877e-02 -6.28272057e-01 1.02027535e+00 3.79903257e-01
4.17083621e-01 -6.27623677e-01 7.50607610e-01 4.33214575e-01
-1.38373697e+00 -1.12186998e-01 -1.71386898e-01 3.64492796e-02
3.61694276e-01 5.37067890e-01 -4.85942781e-01 7.18583643e-01
8.55663300e-01 1.23832035e+00 -5.35475969e-01 1.03038347e+00
-4.28452820e-01 8.13341558e-01 -2.42336139e-01 6.70758188e-02
5.87637722e-01 -2.47373387e-01 4.92372245e-01 1.00205672e+00
3.91880751e-01 2.12270081e-01 7.38647521e-01 7.06678271e-01
-2.37315208e-01 3.17087889e-01 -3.73714715e-01 -1.76907942e-01
3.86477292e-01 1.16174567e+00 -1.08882570e+00 -5.23976803e-01
-6.56906605e-01 1.24395657e+00 1.10169113e-01 5.67681193e-01
-1.12899935e+00 -1.96541250e-01 5.93364239e-01 2.34328061e-01
2.90052265e-01 -2.33869150e-01 1.59448564e-01 -1.74096370e+00
3.06947559e-01 -7.14204192e-01 6.89364493e-01 -8.46375167e-01
-1.03195894e+00 4.29547548e-01 1.69840883e-02 -1.38614678e+00
3.83878537e-02 -1.99875280e-01 -4.41808224e-01 2.95654178e-01
-1.43699002e+00 -1.19205832e+00 -7.53070652e-01 9.32861567e-01
8.27615440e-01 -1.35063648e-01 8.68709236e-02 4.64335054e-01
-8.04376066e-01 3.25994581e-01 -4.11460668e-01 4.41492081e-01
4.82912898e-01 -8.56111586e-01 1.81748252e-02 1.19180489e+00
5.67248225e-01 5.62091351e-01 4.40384388e-01 -6.68763041e-01
-1.38436270e+00 -1.40874541e+00 6.61867678e-01 -2.97439396e-01
8.33141088e-01 -3.10613394e-01 -9.12354112e-01 7.52455056e-01
6.07749261e-03 1.73925906e-01 5.74516296e-01 -1.30755708e-01
-2.34897897e-01 -7.78018981e-02 -5.19683957e-01 4.84208643e-01
1.57448184e+00 -6.40165508e-01 -6.68417990e-01 4.11444098e-01
1.17156720e+00 -3.49471927e-01 -5.92921853e-01 3.46371323e-01
1.61498666e-01 -8.34923327e-01 7.99119532e-01 -4.75841612e-01
4.59787488e-01 -7.51226842e-01 -2.85017222e-01 -9.25964713e-01
7.03027770e-02 -6.44055963e-01 -1.88606307e-01 1.32874513e+00
1.52111799e-01 -1.66744113e-01 1.01671576e+00 2.68983662e-01
-4.88857985e-01 -5.79330921e-01 -6.46300435e-01 -8.22246075e-01
-6.20557487e-01 -8.46175790e-01 5.92581868e-01 1.05684316e+00
-9.08594299e-03 3.11840594e-01 -5.63491046e-01 2.47409448e-01
5.52063465e-01 4.28727537e-01 7.44789779e-01 -9.27155256e-01
-4.37314332e-01 -2.77086705e-01 -5.42478144e-01 -1.41773200e+00
4.05795962e-01 -8.91999722e-01 8.70024338e-02 -1.53448629e+00
3.54208916e-01 -2.43651077e-01 -4.46252882e-01 5.45286238e-01
-4.41940427e-01 3.52905124e-01 3.66899461e-01 4.01495785e-01
-1.46912217e+00 7.30545998e-01 1.15685344e+00 -3.38569254e-01
-2.65976310e-01 -1.72947422e-01 -8.36679935e-01 8.92835200e-01
6.60941362e-01 -2.99509287e-01 -7.80266225e-01 -5.58352828e-01
8.80545676e-02 2.66848933e-02 5.53833723e-01 -1.01637805e+00
2.89821744e-01 -2.69859493e-01 2.28830595e-02 -6.10415936e-01
1.73935205e-01 -7.53030419e-01 2.32804604e-02 1.63997769e-01
-2.21494466e-01 -2.25958794e-01 -2.36179695e-01 8.95847917e-01
-6.69270158e-01 1.19819939e-02 3.59342903e-01 -1.43876716e-01
-1.31490374e+00 7.34992623e-01 -1.94587812e-01 1.56928629e-01
1.23478174e+00 -2.03274891e-01 -1.66763276e-01 -3.72053623e-01
-7.83485591e-01 5.79423487e-01 5.21172583e-01 5.95666826e-01
6.00646734e-01 -1.37073267e+00 -4.32233304e-01 -3.05572785e-02
3.85722578e-01 1.64477661e-01 3.64222884e-01 9.96247947e-01
-4.07797128e-01 3.15721750e-01 2.10116953e-01 -7.90539443e-01
-1.18075788e+00 8.13826442e-01 1.42692357e-01 9.40767024e-03
-7.13668585e-01 7.80138075e-01 5.46812356e-01 2.83507377e-01
2.24953905e-01 -3.52509469e-01 -2.43594751e-01 2.08758995e-01
5.74543238e-01 9.32788700e-02 -1.38776019e-01 -8.12204719e-01
-3.85465980e-01 4.70234185e-01 1.22883253e-01 -2.92995386e-02
1.14720213e+00 -3.59781533e-01 -1.46558480e-02 2.79548407e-01
1.23387218e+00 -2.77019233e-01 -1.38261425e+00 -3.65076661e-01
1.71657950e-01 -5.31216800e-01 -3.64689790e-02 -3.17324519e-01
-1.48007154e+00 8.11072946e-01 1.68639850e-02 2.33652943e-04
1.35154521e+00 3.67089182e-01 9.38493550e-01 3.64103049e-01
4.67334747e-01 -1.00526106e+00 4.86208081e-01 2.78448105e-01
5.74131250e-01 -1.20279479e+00 -2.69380242e-01 -6.85238898e-01
-7.03737497e-01 7.73272336e-01 9.17290866e-01 7.37043321e-02
4.83586490e-01 -1.74534157e-01 -8.82450193e-02 -2.53533512e-01
-7.24713147e-01 -5.49154282e-01 2.94550031e-01 2.98752189e-01
1.19057998e-01 -2.31661066e-01 -2.48051450e-01 8.14290166e-01
2.29971651e-02 1.80809796e-01 2.66681552e-01 8.24436545e-01
-5.91115117e-01 -8.70678306e-01 -1.31847009e-01 3.23522568e-01
-4.83824879e-01 -5.02292579e-03 -3.50427061e-01 5.47013581e-01
3.52385044e-01 1.08342147e+00 -4.17703716e-03 -5.97435534e-01
1.39277980e-01 -6.40973747e-02 2.79323429e-01 -7.04041362e-01
-2.44832546e-01 3.47175956e-01 9.82790738e-02 -9.91543353e-01
-9.23256934e-01 -6.16407812e-01 -1.42483950e+00 -1.12876296e-02
-1.27976015e-01 3.44629973e-01 1.78749844e-01 1.14797163e+00
3.58155102e-01 5.34847796e-01 4.36507195e-01 -6.29242301e-01
5.40444329e-02 -7.06069708e-01 -5.54057956e-01 7.17100322e-01
1.05125211e-01 -7.54649282e-01 -2.70741761e-01 3.23432744e-01] | [9.660659790039062, 0.6798979043960571] |
2b39cc79-7f41-4e7f-8696-03cdbc5482f8 | interpreting-pretrained-source-code-models | 2305.00875 | null | https://arxiv.org/abs/2305.00875v1 | https://arxiv.org/pdf/2305.00875v1.pdf | Interpreting Pretrained Source-code Models using Neuron Redundancy Analyses | Neural code intelligence models continue to be 'black boxes' to the human programmer. This opacity limits their application towards code intelligence tasks, particularly for applications like vulnerability detection where a model's reliance on spurious correlations can be safety-critical. We introduce a neuron-level approach to interpretability of neural code intelligence models which eliminates redundancy due to highly similar or task-irrelevant neurons within these networks. We evaluate the remaining important neurons using probing classifiers which are often used to ascertain whether certain properties have been encoded within the latent representations of neural intelligence models. However, probing accuracies may be artificially inflated due to repetitive and deterministic nature of tokens in code datasets. Therefore, we adapt the selectivity metric originally introduced in NLP to account for probe memorization, to formulate our source-code probing tasks. Through our neuron analysis, we find that more than 95\% of the neurons are redundant wrt. our code intelligence tasks and can be eliminated without significant loss in accuracy. We further trace individual and subsets of important neurons to specific code properties which could be used to influence model predictions. We demonstrate that it is possible to identify 'number' neurons, 'string' neurons, and higher level 'text' neurons which are responsible for specific code properties. This could potentially be used to modify neurons responsible for predictions based on incorrect signals. Additionally, the distribution and concentration of the important neurons within different source code embeddings can be used as measures of task complexity, to compare source-code embeddings and guide training choices for transfer learning over similar tasks. | ['Ali Jannesari', 'Christopher Quinn', 'Zefu Hu', 'Arushi Sharma'] | 2023-05-01 | null | null | null | null | ['vulnerability-detection', 'memorization'] | ['miscellaneous', 'natural-language-processing'] | [ 4.94052261e-01 2.31197730e-01 -6.87277988e-02 -2.70799458e-01
-3.83548826e-01 -9.46714520e-01 3.50178272e-01 5.64996779e-01
-2.61015236e-01 2.70063370e-01 4.06882852e-01 -7.83708930e-01
-1.47914007e-01 -7.76089966e-01 -7.23270178e-01 -4.72392112e-01
-1.22478426e-01 4.85287383e-02 1.55359402e-01 -1.47330165e-01
7.11707711e-01 4.88656700e-01 -1.75255907e+00 8.53091359e-01
7.09395230e-01 5.89236498e-01 7.46338814e-02 5.24484396e-01
-3.13221097e-01 9.98246551e-01 -7.59346426e-01 -2.05095634e-01
-6.86344728e-02 -1.55471489e-01 -7.84748971e-01 -6.45029426e-01
2.06808373e-01 5.20118326e-02 2.14389473e-01 1.44417262e+00
9.04386267e-02 -5.74900694e-02 8.69821012e-01 -1.06630695e+00
-9.67100561e-01 8.38318884e-01 -3.64140838e-01 6.56317592e-01
2.98911601e-01 2.18080357e-01 1.15480673e+00 -6.65297866e-01
4.77521211e-01 1.15960467e+00 9.53176439e-01 5.68442106e-01
-1.65183377e+00 -6.19537234e-01 1.53927162e-01 2.88947709e-02
-1.12359512e+00 -4.23591405e-01 6.71275377e-01 -8.47997308e-01
1.50085807e+00 2.95084983e-01 2.68276125e-01 1.13270903e+00
4.16771978e-01 3.91365319e-01 7.35303581e-01 -4.78845924e-01
3.21089566e-01 2.97067970e-01 5.17393589e-01 8.59964788e-01
6.18069410e-01 1.59932435e-01 -2.89965063e-01 -5.66119075e-01
4.07708317e-01 -2.63192989e-02 -1.63869813e-01 -2.71598279e-01
-1.04374015e+00 1.18338513e+00 4.76668328e-01 4.68971491e-01
-8.40528384e-02 3.52438331e-01 6.39856696e-01 3.13953489e-01
9.24968719e-02 1.26187718e+00 -7.27879167e-01 -3.71204078e-01
-5.03109157e-01 1.62531406e-01 7.51656055e-01 6.47465348e-01
9.13609147e-01 1.05176069e-01 -2.03127339e-01 8.05885971e-01
1.69120774e-01 -1.14926927e-01 7.99376965e-01 -1.00758505e+00
3.94812107e-01 9.56650972e-01 -2.53076255e-01 -1.15612113e+00
-5.17743051e-01 -4.57783282e-01 -2.89260447e-01 3.64179760e-01
2.77595460e-01 1.32498948e-03 -6.33198500e-01 1.89093471e+00
-4.21383888e-01 -3.11925560e-01 1.53690083e-02 4.48836297e-01
2.12037385e-01 3.38091195e-01 2.81888902e-01 3.03109050e-01
1.47204983e+00 -5.38849592e-01 -1.47921965e-01 -5.66686392e-01
1.17345250e+00 -4.31124628e-01 1.51590002e+00 2.27057755e-01
-7.95828760e-01 -3.67704272e-01 -1.06222677e+00 9.79369655e-02
-6.27052426e-01 -1.69522136e-01 8.92609239e-01 8.70927513e-01
-1.06291592e+00 6.84723079e-01 -6.58481896e-01 -2.35082746e-01
6.43300295e-01 5.83874941e-01 -6.85486868e-02 2.75604635e-01
-9.62071657e-01 9.81791198e-01 3.76394182e-01 -4.46047902e-01
-7.21540630e-01 -8.14300656e-01 -9.03902650e-01 5.21915436e-01
-2.03030825e-01 -3.92456710e-01 1.03677523e+00 -1.14345729e+00
-6.96987092e-01 7.18066573e-01 -2.08967999e-01 -2.57584959e-01
-3.13155770e-01 3.32046866e-01 -3.04295212e-01 -1.15770929e-01
2.53903598e-01 7.10473776e-01 5.92077315e-01 -1.16958714e+00
-4.03653562e-01 -2.28930801e-01 2.40045980e-01 -3.73992026e-01
-6.24314129e-01 1.55740529e-01 1.77586138e-01 -7.29840875e-01
-3.70017514e-02 -9.15890336e-01 -5.27095683e-02 -7.75049999e-02
-2.31030926e-01 -9.54194665e-02 3.66128653e-01 -4.60907489e-01
1.20340979e+00 -2.18189764e+00 -9.08655375e-02 3.81157011e-01
3.07951450e-01 1.35579109e-01 -4.43392396e-01 1.27461746e-01
-3.92940640e-01 7.14060187e-01 -4.26732928e-01 1.93833232e-01
-9.59110707e-02 4.52173129e-02 -4.91665423e-01 2.66332299e-01
6.37357414e-01 8.72391462e-01 -8.01479101e-01 -1.12283058e-01
-2.18651161e-01 3.14849317e-01 -1.09069622e+00 9.42753069e-03
-1.26275226e-01 4.20698151e-02 -3.77829373e-01 6.65331602e-01
6.32583797e-02 -1.52640820e-01 -9.41554233e-02 1.65548787e-01
3.49232890e-02 7.53241956e-01 -4.80325133e-01 1.23274934e+00
-5.97627521e-01 9.34937119e-01 -2.99755812e-01 -1.15457070e+00
1.03630340e+00 4.52069682e-04 -2.34916806e-01 -6.78468406e-01
1.46270424e-01 1.46575794e-01 6.21140599e-01 -6.16003811e-01
3.73980254e-01 -3.37695256e-02 -2.65885383e-01 7.00845897e-01
-6.17748275e-02 2.60715872e-01 -9.70261320e-02 -1.88519284e-01
1.65963709e+00 -1.39909476e-01 9.55418199e-02 -7.12751985e-01
2.77406484e-01 1.86903924e-01 4.84659046e-01 9.78849113e-01
-3.12186807e-01 5.74176610e-01 1.02136707e+00 -4.27339941e-01
-1.15367377e+00 -9.00185287e-01 -3.83477598e-01 1.31834042e+00
-2.93210655e-01 -1.36492610e-01 -7.80414283e-01 -5.29713809e-01
2.94109695e-02 1.13733208e+00 -1.09020054e+00 -8.09309423e-01
-4.09796327e-01 -8.43220949e-01 1.03616190e+00 7.98622966e-01
-1.30093947e-01 -1.14382756e+00 -8.84798765e-01 2.35035345e-01
4.12752926e-02 -4.77074027e-01 -9.68119279e-02 1.05212247e+00
-9.78344560e-01 -1.01592684e+00 -1.00595102e-01 -6.62742496e-01
1.01195157e+00 1.99292582e-02 1.29084587e+00 5.94292164e-01
-2.68143803e-01 2.07872123e-01 -3.16499859e-01 -4.12863433e-01
-6.75465524e-01 3.05599779e-01 -2.41035670e-02 -5.14433444e-01
9.30617571e-01 -7.19254434e-01 -3.17320973e-01 9.33627710e-02
-9.31152344e-01 -3.09510589e-01 4.33769584e-01 8.15645576e-01
2.32956167e-02 7.22579286e-02 4.67806548e-01 -1.12332535e+00
1.10341334e+00 -8.45203102e-01 -3.28253299e-01 1.67458087e-01
-6.53065324e-01 7.43995011e-01 7.69276321e-01 -7.67410159e-01
-6.26083910e-01 -1.59918696e-01 -1.17876612e-01 -1.34161487e-01
-2.08147183e-01 6.31469727e-01 -4.36810516e-02 -2.45061561e-01
1.27307701e+00 -2.89099198e-03 -1.17236763e-01 -3.62480819e-01
-4.47205454e-02 4.97188330e-01 2.07104534e-01 -9.98630524e-01
6.08701289e-01 1.28593445e-01 -4.09347445e-01 -1.91603020e-01
-3.43814164e-01 5.68261975e-03 -4.61926222e-01 4.10959631e-01
7.48250604e-01 -5.54719865e-01 -6.98392272e-01 -1.31124437e-01
-1.37349033e+00 -3.06183845e-01 -1.60186514e-01 8.11492279e-02
-3.99588674e-01 1.54724523e-01 -6.63401306e-01 -5.77521801e-01
-1.75444398e-03 -1.57507050e+00 6.59384966e-01 1.31084099e-01
-1.12471366e+00 -1.12220168e+00 3.03348843e-02 -7.68683329e-02
5.08896232e-01 -4.04632427e-02 1.94745266e+00 -1.13501740e+00
-2.93671906e-01 -2.04267442e-01 -3.50162894e-01 1.30292356e-01
-4.47522625e-02 1.14969313e-01 -1.30415285e+00 -1.56099703e-02
4.96738404e-02 -2.01692164e-01 9.85955060e-01 4.68569547e-02
1.37200439e+00 -5.33902645e-01 -5.06425858e-01 5.77370524e-01
1.36085665e+00 4.28850204e-01 5.85478663e-01 5.34065425e-01
6.48679852e-01 8.80171239e-01 -5.32392114e-02 1.33373082e-01
2.01496296e-02 3.68689537e-01 3.88964117e-01 3.23404759e-01
3.13957572e-01 -2.24746048e-01 6.72702312e-01 5.46510100e-01
2.39972755e-01 4.65593450e-02 -1.42735636e+00 8.11629474e-01
-1.46150517e+00 -8.59397113e-01 -4.90392223e-02 2.05057883e+00
9.17596757e-01 4.99365628e-01 -2.48037398e-01 1.63627952e-01
5.78469276e-01 -1.64977148e-01 -5.51231623e-01 -1.27102220e+00
1.13413267e-01 2.90118694e-01 4.78492379e-01 3.40036184e-01
-6.39369607e-01 5.35084128e-01 6.80810070e+00 4.28175896e-01
-8.84181976e-01 5.22141904e-02 7.44207799e-01 4.39867117e-02
-9.29083705e-01 1.45165682e-01 -7.65900254e-01 4.54026967e-01
1.32409108e+00 1.00242443e-01 7.19080031e-01 1.02224004e+00
-1.29598171e-01 -2.09001526e-02 -1.52207088e+00 5.16313314e-01
4.11889702e-02 -1.30770421e+00 1.12750821e-01 1.34517521e-01
5.53036273e-01 1.25811500e-02 2.41256610e-01 5.02872825e-01
5.41660309e-01 -1.24687564e+00 7.41649330e-01 2.99125969e-01
5.97427070e-01 -7.75686324e-01 5.33606172e-01 2.35491082e-01
-8.39824200e-01 -6.10449076e-01 -7.01237023e-01 -4.93934929e-01
-6.55829728e-01 3.84625405e-01 -8.10180187e-01 -6.80315375e-01
7.11561203e-01 2.73091048e-01 -1.11549640e+00 6.56858563e-01
-1.86920121e-01 5.91212153e-01 -1.90210845e-02 -3.89409989e-01
1.46652862e-01 4.08520669e-01 3.06074113e-01 1.31202674e+00
2.93390036e-01 -3.41879070e-01 -4.83955622e-01 1.75348818e+00
2.16711741e-02 -2.59302855e-01 -9.10047352e-01 -2.93873221e-01
6.75348878e-01 1.03419435e+00 -9.62033153e-01 -2.01241836e-01
-4.37007934e-01 7.88632333e-01 5.82177281e-01 4.96127158e-01
-7.32960403e-01 -6.76892579e-01 1.05630410e+00 -2.01748777e-02
2.39240661e-01 1.27271889e-02 -9.08591092e-01 -1.07755339e+00
-1.41024292e-01 -9.96806800e-01 6.26906082e-02 -7.28186488e-01
-1.15961051e+00 6.42658889e-01 -1.47940293e-01 -8.77204418e-01
-4.16873544e-01 -9.53394651e-01 -9.12366509e-01 9.42292273e-01
-1.11043274e+00 -5.72365284e-01 8.71680900e-02 2.08352312e-01
2.64079452e-01 -4.99213427e-01 1.13998699e+00 -2.60080155e-02
-4.82369721e-01 9.73672628e-01 9.43477973e-02 4.36870307e-01
1.81772754e-01 -1.19087923e+00 7.52306283e-01 8.62205863e-01
2.33513772e-01 1.34342062e+00 7.35149145e-01 -5.49280226e-01
-9.97345388e-01 -1.04457963e+00 9.33435082e-01 -9.73876536e-01
7.04382718e-01 -5.70898533e-01 -1.27918422e+00 6.33332968e-01
-2.00164750e-01 -6.55974895e-02 9.51267600e-01 4.30600345e-01
-1.03107381e+00 3.02341789e-01 -1.08527422e+00 7.00737536e-01
9.05748844e-01 -1.04668307e+00 -8.09797943e-01 -1.32680848e-01
8.17030907e-01 2.37874851e-01 -6.92006946e-01 2.33022422e-01
3.91446799e-01 -1.01563811e+00 9.21857536e-01 -9.01454926e-01
6.80357754e-01 -1.35313496e-01 -1.60033986e-01 -1.21558499e+00
-7.32028067e-01 1.03291668e-01 1.40084341e-01 1.29554796e+00
8.78108978e-01 -5.77417552e-01 5.59291244e-01 1.02473044e+00
-1.10376321e-01 -4.81284410e-01 -7.69218504e-01 -4.51553255e-01
3.94441694e-01 -6.67398572e-01 7.33848631e-01 1.08823848e+00
3.91452909e-01 1.13590091e-01 4.17367518e-01 1.27770737e-01
1.77800104e-01 -2.29903296e-01 6.32183701e-02 -1.35344636e+00
-5.52285492e-01 -9.16887999e-01 -5.02847433e-01 -4.21311975e-01
4.95150000e-01 -1.23127508e+00 1.11710489e-01 -9.00515795e-01
3.21565926e-01 -6.08561516e-01 -5.22949278e-01 8.17452013e-01
-1.90436840e-01 2.36387759e-01 2.08453573e-02 3.14954460e-01
-3.12578417e-02 -9.04381201e-02 3.39791209e-01 -3.60621393e-01
-1.38517112e-01 -2.76800931e-01 -1.24767244e+00 8.47846746e-01
9.26370680e-01 -8.06018472e-01 -4.43955868e-01 -5.80491900e-01
8.18041146e-01 -4.12125558e-01 4.98265773e-01 -1.06909001e+00
-5.23894951e-02 -1.03843324e-02 5.58043659e-01 2.30565295e-01
-1.71136990e-01 -8.40774417e-01 -8.00625533e-02 6.86240315e-01
-8.24483454e-01 3.12857121e-01 6.41593695e-01 2.92198300e-01
5.69285490e-02 -8.82130921e-01 6.02876425e-01 -3.80219012e-01
-6.09640181e-01 -4.21949148e-01 -9.22210872e-01 1.76851705e-01
6.03684247e-01 -4.84252274e-01 -4.64926213e-01 5.45009039e-02
-2.24372521e-01 -3.60627681e-01 7.21759617e-01 6.78084254e-01
5.25385678e-01 -1.16193223e+00 -2.05753297e-01 4.03735518e-01
4.95544732e-01 -4.99614745e-01 -2.23737210e-01 4.64733243e-01
-4.46970373e-01 4.34126407e-01 -3.91478717e-01 -3.58795106e-01
-8.81413877e-01 8.60887408e-01 3.93039376e-01 1.24762893e-01
-2.83076704e-01 8.79759550e-01 3.69172752e-01 -5.12653232e-01
9.67408568e-02 -7.58731842e-01 -2.93427408e-01 1.13046743e-01
6.00543737e-01 3.99654284e-02 8.97155330e-02 -2.73930877e-01
-4.56105053e-01 2.58667499e-01 -1.28202319e-01 1.43474162e-01
1.28370106e+00 2.30185971e-01 -2.23582402e-01 6.77160859e-01
1.47971344e+00 -5.02035394e-02 -1.01060951e+00 2.44086921e-01
4.88146156e-01 -2.45484054e-01 -1.10862903e-01 -8.59632432e-01
-7.88366675e-01 1.30258405e+00 6.78787887e-01 1.17514782e-01
8.15586984e-01 -1.00317858e-01 2.30304644e-01 4.67047095e-01
2.67776191e-01 -7.67632246e-01 -8.02960396e-02 5.70811868e-01
5.92985451e-01 -9.72507656e-01 -4.36995268e-01 1.28916517e-01
-2.30585635e-01 1.33074224e+00 7.47661471e-01 -1.17913291e-01
4.16578114e-01 5.98044395e-01 -1.16286196e-01 -1.42429888e-01
-9.14962947e-01 2.23970577e-01 5.33879399e-02 8.91424060e-01
7.22857714e-01 -6.88809007e-02 -3.16687226e-02 8.19211721e-01
-1.69960022e-01 -3.74510258e-01 6.42528474e-01 8.18603992e-01
-6.61096513e-01 -1.04253519e+00 -4.81695682e-01 8.80315065e-01
-5.06009698e-01 -5.98023236e-01 -4.08144027e-01 4.15014416e-01
3.42338115e-01 6.16015792e-01 3.89822423e-01 -7.43487537e-01
1.20000750e-01 4.65306491e-01 1.31143719e-01 -1.03008938e+00
-9.98083174e-01 -5.52792013e-01 -2.59813759e-02 -2.91980207e-01
1.50186196e-01 -6.19822979e-01 -1.41346395e+00 -3.30858409e-01
-2.07647309e-01 2.33377656e-03 3.90459687e-01 8.10019195e-01
5.77713609e-01 5.55254281e-01 8.78375024e-02 -5.76615870e-01
-5.32216311e-01 -7.59693503e-01 -1.48389950e-01 4.91792560e-01
3.01569968e-01 -6.56622469e-01 -4.83260006e-01 1.15841910e-01] | [7.566608428955078, 7.755120277404785] |
c8404ce1-7da5-4ba6-a1bf-1bc8fea0dd18 | follow-the-attention-combining-partial-pose | 1905.04430 | null | https://arxiv.org/abs/1905.04430v2 | https://arxiv.org/pdf/1905.04430v2.pdf | Follow the Attention: Combining Partial Pose and Object Motion for Fine-Grained Action Detection | Retailers have long been searching for ways to effectively understand their customers' behaviour in order to provide a smooth and pleasant shopping experience that attracts more customers everyday and maximises their revenue, consequently. Humans can flawlessly understand others' behaviour by combining different visual cues from activity to gestures and facial expressions. Empowering the computer vision systems to do so, however, is still an open problem due to its intrinsic challenges as well as extrinsic enforced difficulties like lack of publicly available data and unique environment conditions (wild). In this work, We emphasise on detecting the first and by far the most crucial cue in behaviour analysis; that is human activity detection in computer vision. To do so, we introduce a framework for integrating human pose and object motion to both temporally detect and classify the activities in a fine-grained manner (very short and similar activities). We incorporate partial human pose and interaction with the objects in a multi-stream neural network architecture to guide the spatiotemporal attention mechanism for more efficient activity recognition. To this end, in the absence of pose supervision, we propose to use the Generative Adversarial Network (GAN) to generate exact joint locations from noisy probability heat maps. Additionally, based on the intuition that complex actions demand more than one source of information to be identified even by humans, we integrate the second stream of object motion to our network as a prior knowledge that we quantitatively show improves the recognition results. We empirically show the capability of our approach by achieving state-of-the-art results on MERL shopping dataset. We further investigate the effectiveness of this approach on a new shopping dataset that we have collected to address existing shortcomings. | ['Mohammad Mahdi Kazemi Moghaddam', 'Javen Shi', 'Ehsan Abbasnejad'] | 2019-05-11 | null | null | null | null | ['fine-grained-action-detection'] | ['computer-vision'] | [ 4.14952964e-01 -1.35648802e-01 3.13226730e-02 -3.89274448e-01
-6.03076756e-01 -7.16194391e-01 8.43031824e-01 -1.87931836e-01
-5.17538369e-01 3.29469144e-01 1.87013343e-01 1.32946655e-01
9.60964896e-03 -4.40833896e-01 -7.92697906e-01 -8.34752560e-01
2.84926244e-03 3.07137311e-01 1.97500363e-01 -3.45685124e-01
2.08864138e-01 6.48097456e-01 -1.72189128e+00 3.47376704e-01
4.82623070e-01 1.10710967e+00 8.99664909e-02 7.13560343e-01
1.02282830e-01 8.17856669e-01 -4.83270913e-01 -3.57590139e-01
3.29096228e-01 -4.91313279e-01 -4.51174736e-01 4.21207815e-01
2.39004835e-01 -3.05408001e-01 -2.95370910e-02 7.85829484e-01
2.59478748e-01 3.43086779e-01 5.75258732e-01 -1.31390417e+00
-3.55857402e-01 1.52308285e-01 -5.09779572e-01 9.61401314e-02
5.53844035e-01 6.39757454e-01 6.58875465e-01 -6.48670614e-01
5.60959339e-01 1.14303780e+00 5.47581553e-01 4.96880084e-01
-1.36522675e+00 -4.20189232e-01 3.76581311e-01 2.21672848e-01
-1.17076766e+00 -4.94183153e-01 1.07085776e+00 -1.02977969e-01
7.92867601e-01 2.52843648e-01 7.81110346e-01 1.84464407e+00
-3.02604698e-02 1.17098880e+00 9.52907026e-01 -2.93365747e-01
3.40586275e-01 1.41450807e-01 -2.15252981e-01 4.01858091e-01
-6.82852343e-02 1.07343338e-01 -4.98131365e-01 2.70694315e-01
7.69273520e-01 2.11241797e-01 -1.43270776e-01 -5.90519547e-01
-1.20099699e+00 7.47368455e-01 4.58594114e-01 4.16739404e-01
-6.54711545e-01 3.11747164e-01 2.89534956e-01 -6.11572107e-03
7.84609392e-02 4.47916567e-01 -3.23690057e-01 -5.46145499e-01
-7.92851925e-01 2.09960192e-01 9.02909815e-01 6.83051705e-01
2.65438735e-01 -1.20610565e-01 -3.57567072e-02 5.75316072e-01
3.70692939e-01 4.74041015e-01 3.97527993e-01 -9.69372809e-01
2.72607148e-01 5.62528789e-01 3.31435680e-01 -1.17451668e+00
-3.48539472e-01 -3.90278220e-01 -6.30407274e-01 3.04230154e-01
7.97990859e-01 2.67136186e-01 -9.79348004e-01 1.79329205e+00
2.78144330e-01 1.42478615e-01 -7.45248869e-02 1.16162193e+00
1.54383793e-01 5.11521697e-01 2.51411229e-01 7.16527849e-02
1.30345333e+00 -9.76686418e-01 -5.99230111e-01 -4.77775514e-01
4.35738027e-01 -5.74326754e-01 1.19281065e+00 6.39918566e-01
-9.54138577e-01 -6.47511184e-01 -9.81298625e-01 1.77625939e-01
-4.77322757e-01 -5.75566962e-02 7.85380244e-01 6.81611240e-01
-6.18328631e-01 4.53916848e-01 -1.25823748e+00 -4.94720787e-01
4.96212870e-01 3.22122902e-01 -3.85626674e-01 -1.46368265e-01
-9.36529756e-01 8.73377740e-01 1.56574786e-01 3.31431061e-01
-7.71802783e-01 -3.33557487e-01 -1.02885616e+00 -7.34112710e-02
5.84781528e-01 -4.02560711e-01 1.20999134e+00 -1.22906935e+00
-1.48208165e+00 6.01833403e-01 -1.09627778e-02 -5.88663638e-01
7.91067600e-01 -2.76550591e-01 -3.03809792e-01 1.87591121e-01
3.03689241e-02 8.07980716e-01 8.42115343e-01 -1.54077613e+00
-6.73788011e-01 -6.70983791e-01 -7.03276275e-03 1.71716109e-01
-1.48525283e-01 -1.76760614e-01 -7.38094687e-01 -7.18907297e-01
1.22387327e-01 -1.17107511e+00 -1.75121978e-01 -6.92960620e-02
-1.58331737e-01 -1.08858990e-02 9.40714419e-01 -6.37464583e-01
6.92096293e-01 -2.26954198e+00 1.10329114e-01 2.57981420e-01
-7.29930177e-02 2.70368218e-01 -1.31529480e-01 2.42952570e-01
1.42587706e-01 -1.89584270e-01 -8.96704793e-02 -5.47494113e-01
2.14913890e-01 5.17369986e-01 -1.77102387e-01 5.72397709e-01
4.35987681e-01 1.31251919e+00 -8.31258893e-01 -2.22276002e-01
5.82453668e-01 7.67261207e-01 -3.38489175e-01 2.33351916e-01
-1.13469861e-01 8.31218719e-01 -3.56141001e-01 8.12334836e-01
4.50312376e-01 -4.44653779e-02 -7.20944405e-02 -2.74310648e-01
1.58918247e-01 -8.43033269e-02 -1.46236908e+00 1.79816675e+00
-5.55663526e-01 2.86067694e-01 1.61844358e-01 -1.09198225e+00
7.31952071e-01 1.18285917e-01 6.52489483e-01 -1.10930753e+00
2.39382386e-01 9.21880752e-02 -5.03859334e-02 -4.13970768e-01
2.17781067e-01 2.26445533e-02 -4.28074151e-01 4.46641594e-01
4.68601808e-02 1.18526079e-01 1.04423895e-01 -5.22481762e-02
1.12605882e+00 4.80597883e-01 1.29943117e-01 2.47839496e-01
3.44836801e-01 -9.10477415e-02 2.49509633e-01 9.33553457e-01
-5.48955262e-01 5.13932109e-01 1.86677247e-01 -4.27477419e-01
-9.91071880e-01 -1.13242602e+00 1.31914064e-01 1.13599432e+00
3.59281629e-01 1.11572810e-01 -7.07634568e-01 -7.65309751e-01
-1.21505208e-01 7.32134998e-01 -7.59815276e-01 -2.00386122e-01
-7.30502963e-01 -4.98857349e-01 4.92255300e-01 9.10697639e-01
6.29776180e-01 -1.38073707e+00 -1.02830732e+00 3.93765271e-01
-1.05007231e-01 -1.33492780e+00 -3.75814229e-01 3.95154953e-01
-5.09626091e-01 -8.44302297e-01 -7.61732578e-01 -4.39968973e-01
3.94196391e-01 1.18586659e-01 8.78333509e-01 -4.23787236e-01
-4.94926393e-01 5.69022596e-01 -4.33838904e-01 -3.16364557e-01
-3.15681517e-01 -5.26440255e-02 -1.29182741e-01 3.47444415e-01
6.56251669e-01 -6.19858742e-01 -8.15541506e-01 6.68650568e-01
-1.03161216e+00 -1.42896160e-01 9.33992684e-01 6.08582199e-01
3.64549577e-01 3.76066044e-02 3.79001766e-01 -6.07124448e-01
5.42822003e-01 -3.68744403e-01 -3.37229609e-01 2.24711020e-02
-3.84943098e-01 1.28524691e-01 4.77381349e-01 -7.30549216e-01
-1.06240535e+00 6.56882286e-01 -1.55032858e-01 -4.26162481e-01
-7.17977226e-01 1.03193812e-01 -2.70932525e-01 -1.50447125e-02
5.72485566e-01 4.17243451e-01 -9.79235396e-03 -3.83869231e-01
6.96040213e-01 4.82010454e-01 8.76491308e-01 -3.86516273e-01
5.35775661e-01 7.34317958e-01 -7.37625584e-02 -6.30690515e-01
-5.13172805e-01 -7.73735821e-01 -8.49954307e-01 -3.58758509e-01
9.18549716e-01 -6.53996944e-01 -1.13012803e+00 4.60704744e-01
-9.99643862e-01 -1.93403989e-01 -2.38615051e-01 3.18219066e-01
-8.34876716e-01 3.01517189e-01 -5.73725104e-01 -1.15067732e+00
-7.03233527e-03 -1.21079910e+00 1.41342962e+00 6.57738000e-02
-3.56605798e-01 -7.55502284e-01 -1.58826888e-01 5.04038513e-01
4.63516593e-01 5.82852244e-01 2.99240470e-01 -6.77943110e-01
-6.83495343e-01 -3.36140513e-01 -2.47120969e-02 2.84002692e-01
1.56315118e-02 -4.75522876e-01 -1.04173648e+00 -3.43770385e-02
2.20515817e-01 -2.57637262e-01 7.72876799e-01 3.14620554e-01
9.14664328e-01 -1.82546377e-01 -2.62654603e-01 3.78722757e-01
1.18907475e+00 3.95588815e-01 8.49710703e-01 3.85728121e-01
7.84230947e-01 8.54740143e-01 7.72378206e-01 3.82530451e-01
2.59860039e-01 9.79573727e-01 6.24410927e-01 -1.86341703e-01
4.29577492e-02 -2.41541266e-01 4.75453705e-01 1.92009732e-01
-2.05140039e-01 -1.92733109e-01 -6.27997518e-01 5.84130526e-01
-2.01136541e+00 -1.15634620e+00 4.68734205e-02 2.09935880e+00
5.01425564e-01 2.72529036e-01 4.30293888e-01 3.01739305e-01
4.59814966e-01 1.83308218e-02 -6.75298214e-01 -3.72451484e-01
1.52512178e-01 1.54771835e-01 5.93693137e-01 9.96333510e-02
-1.28575110e+00 7.68354714e-01 5.19552898e+00 7.01396883e-01
-1.01234663e+00 9.32229869e-03 5.12610316e-01 -2.30835631e-01
1.04518957e-01 -3.60996217e-01 -6.70455217e-01 5.60367227e-01
9.08828437e-01 5.84282100e-01 7.43493855e-01 9.03804541e-01
3.42519552e-01 -2.79702187e-01 -1.22993386e+00 1.00619066e+00
2.50654280e-01 -7.79186070e-01 -3.88455153e-01 1.98168203e-01
4.51812804e-01 -2.64592737e-01 1.56971216e-01 4.39651817e-01
2.08781570e-01 -1.13968003e+00 8.70877624e-01 5.34816325e-01
1.95762455e-01 -7.97089636e-01 6.38933241e-01 5.77848315e-01
-1.27214706e+00 -2.57670164e-01 3.82822841e-01 6.23486415e-02
5.45594454e-01 1.24037139e-01 -9.45309281e-01 2.68473029e-01
6.89001799e-01 5.38848579e-01 -4.62023228e-01 7.04818606e-01
-2.47581154e-01 3.13203841e-01 -5.69035590e-01 -2.00121328e-01
5.18999040e-01 6.11620359e-02 3.67897213e-01 1.19809151e+00
1.19325303e-01 -8.82048607e-02 1.86762080e-01 9.74569798e-01
3.07062268e-01 -1.81867167e-01 -6.99171424e-01 -1.07652031e-01
4.61779609e-02 1.19063866e+00 -9.77866411e-01 1.54896080e-01
-4.15215284e-01 1.41859722e+00 7.25361109e-02 3.71943980e-01
-9.99216139e-01 -8.14967528e-02 5.62485397e-01 2.40359724e-01
5.26243567e-01 -3.26645404e-01 -1.15525678e-01 -1.00980425e+00
3.70173126e-01 -9.12093461e-01 1.61254734e-01 -7.62816787e-01
-1.25631011e+00 2.57254541e-01 -5.73224165e-02 -1.01292896e+00
-6.83582962e-01 -6.96276367e-01 -1.95076287e-01 6.76168203e-01
-1.36953998e+00 -1.55921674e+00 -4.40803111e-01 6.61406338e-01
8.01308930e-01 3.08902591e-01 7.20157206e-01 3.47290605e-01
-2.73804784e-01 5.02660215e-01 -3.13911706e-01 1.74994871e-01
4.96852160e-01 -1.21043801e+00 4.29652274e-01 8.11765909e-01
3.94857138e-01 6.06528997e-01 7.47602284e-01 -4.97462958e-01
-1.59477425e+00 -8.78472805e-01 4.70227778e-01 -8.27253699e-01
4.97827798e-01 -6.31678939e-01 -6.78204954e-01 7.38696873e-01
-1.30165905e-01 2.34089829e-02 3.66799116e-01 -1.77152798e-01
-1.35753676e-01 2.79958155e-02 -1.20182860e+00 6.40833616e-01
1.09920394e+00 -4.11106229e-01 -5.01340091e-01 -1.19449735e-01
3.35349858e-01 -2.30409563e-01 -7.04861939e-01 4.46103334e-01
7.73753941e-01 -1.12798142e+00 1.11567664e+00 -3.33636314e-01
2.09781066e-01 -2.66189903e-01 -2.76507974e-01 -1.02341580e+00
6.43063262e-02 -5.49223363e-01 -2.30740219e-01 1.09788537e+00
1.96318939e-01 -4.45405811e-01 1.05605853e+00 7.82124281e-01
1.12359010e-01 -8.13998520e-01 -7.65256405e-01 -6.41232848e-01
-4.88608211e-01 -9.05186296e-01 4.01316613e-01 6.80659235e-01
-1.93930730e-01 2.16212183e-01 -5.06419301e-01 1.41640544e-01
4.69336599e-01 -6.92401826e-02 9.76094663e-01 -1.06868565e+00
-5.57214439e-01 -3.17204744e-01 -6.86270237e-01 -1.12914109e+00
-6.52224571e-02 -3.14991713e-01 2.75896281e-01 -1.20651579e+00
4.09145541e-02 -1.64571881e-01 -2.26319104e-01 3.91646922e-01
2.05414772e-01 5.11488140e-01 3.65970135e-01 1.76474839e-01
-6.96668446e-01 3.15356255e-01 1.30479813e+00 -4.06434275e-02
-2.69480675e-01 1.42521575e-01 -6.84037745e-01 5.87674618e-01
3.89167964e-01 -1.76141635e-01 -4.17938977e-01 1.58548839e-02
2.11880170e-02 -3.71977836e-02 7.69737661e-01 -1.02297199e+00
1.89583868e-01 5.56583516e-02 5.91257453e-01 -5.05439520e-01
7.63497114e-01 -1.15264666e+00 1.16788931e-01 4.27921742e-01
-3.05927485e-01 -1.52895704e-01 1.00957744e-01 8.61203372e-01
-1.97897941e-01 -3.77747752e-02 5.06094456e-01 -3.19800079e-01
-1.09136522e+00 -6.78487569e-02 -2.19221607e-01 -3.42202395e-01
1.26756847e+00 -4.54429299e-01 9.58147459e-03 -6.71290636e-01
-1.01336741e+00 1.38894722e-01 4.14366931e-01 7.39451885e-01
3.43653411e-01 -1.21620560e+00 -2.81713396e-01 4.28695321e-01
3.89600158e-01 -2.19518587e-01 2.05880925e-01 9.61319983e-01
-2.92088568e-01 5.19557059e-01 -2.24127665e-01 -8.68049741e-01
-9.10431623e-01 6.61024868e-01 1.83506891e-01 -2.67189354e-01
-4.68863755e-01 6.34169221e-01 6.54114876e-03 -2.87304640e-01
4.26141411e-01 -4.69354928e-01 -3.24864089e-01 -2.62582544e-02
4.58174914e-01 2.16688424e-01 1.49686277e-01 -8.37885439e-01
-3.80411953e-01 4.44217652e-01 -1.22062191e-02 -3.09903592e-01
1.34996796e+00 -2.57956892e-01 7.09773600e-01 3.52766395e-01
1.22586131e+00 -1.93460435e-01 -1.76702511e+00 7.35992864e-02
3.75612788e-02 -5.82513332e-01 -3.38885598e-02 -9.89642322e-01
-9.36489940e-01 8.53761733e-01 9.12622929e-01 4.33296144e-01
1.16519189e+00 1.85875535e-01 9.46548641e-01 1.91833273e-01
4.50369596e-01 -1.11104238e+00 3.64220053e-01 5.79144694e-02
6.19070351e-01 -1.36780179e+00 -4.67005700e-01 -2.67569065e-01
-1.01572096e+00 8.19771528e-01 3.92009854e-01 -1.04836039e-01
1.66651055e-01 2.80580133e-01 2.07892075e-01 -2.07670897e-01
-3.26787114e-01 -4.70079362e-01 1.94990516e-01 7.75654078e-01
1.36177763e-01 -1.64082363e-01 1.61128677e-02 6.66743279e-01
1.09017633e-01 8.04241151e-02 5.32677397e-02 1.02673733e+00
-1.91016644e-01 -9.75701153e-01 -3.55452538e-01 7.12540820e-02
-4.17106062e-01 3.45695168e-01 -3.29787582e-01 9.26322818e-01
1.09804787e-01 9.47546303e-01 6.37324601e-02 -2.59917557e-01
6.35772586e-01 2.48044208e-02 5.42819023e-01 -2.92074502e-01
-4.73296195e-01 2.75722355e-01 -2.53995694e-03 -1.01898205e+00
-5.37966907e-01 -8.35536540e-01 -1.06950498e+00 -5.98108247e-02
-7.56433234e-02 -4.17136371e-01 1.02999854e+00 1.03732610e+00
3.40599537e-01 5.00124753e-01 4.90891188e-01 -1.47654879e+00
-7.22617984e-01 -8.41388106e-01 -4.78189826e-01 9.79410648e-01
7.86384568e-02 -8.55597496e-01 -4.10800636e-01 1.64270967e-01] | [7.879979133605957, 0.14377570152282715] |
a41b6484-e749-49b7-97ce-00c9f95de380 | learning-to-represent-review-with-tensor | null | null | https://aclanthology.org/D16-1083 | https://aclanthology.org/D16-1083.pdf | Learning to Represent Review with Tensor Decomposition for Spam Detection | null | ['Jun Zhao', 'Shizhu He', 'Xuepeng Wang', 'Kang Liu'] | 2016-11-01 | null | null | null | emnlp-2016-11 | ['spam-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.331501483917236, 3.8560943603515625] |
0ae868e9-c245-4521-a7a7-05b60398c4c3 | un-solving-morphological-inflection-lemma-1 | null | null | https://aclanthology.org/2022.acl-short.96 | https://aclanthology.org/2022.acl-short.96.pdf | (Un)solving Morphological Inflection: Lemma Overlap Artificially Inflates Models’ Performance | In the domain of Morphology, Inflection is a fundamental and important task that gained a lot of traction in recent years, mostly via SIGMORPHON’s shared-tasks.With average accuracy above 0.9 over the scores of all languages, the task is considered mostly solved using relatively generic neural seq2seq models, even with little data provided.In this work, we propose to re-evaluate morphological inflection models by employing harder train-test splits that will challenge the generalization capacity of the models. In particular, as opposed to the naïve split-by-form, we propose a split-by-lemma method to challenge the performance on existing benchmarks.Our experiments with the three top-ranked systems on the SIGMORPHON’s 2020 shared-task show that the lemma-split presents an average drop of 30 percentage points in macro-average for the 90 languages included. The effect is most significant for low-resourced languages with a drop as high as 95 points, but even high-resourced languages lose about 10 points on average. Our results clearly show that generalizing inflection to unseen lemmas is far from being solved, presenting a simple yet effective means to promote more sophisticated models. | ['Reut Tsarfaty', 'David Guriel', 'Omer Goldman'] | null | null | null | null | acl-2022-5 | ['morphological-inflection'] | ['natural-language-processing'] | [-1.10380739e-01 5.58798499e-02 4.01557535e-02 -3.55953842e-01
-1.25198543e+00 -1.01978028e+00 5.94470561e-01 4.31396991e-01
-9.91315007e-01 8.79199207e-01 3.49797100e-01 -4.35986817e-01
-1.67328399e-02 -5.28343022e-01 -8.37822914e-01 -4.41044778e-01
-9.97325405e-02 5.98506749e-01 1.61120772e-01 -5.70391536e-01
2.37881429e-02 2.28323311e-01 -1.33631063e+00 5.28191924e-01
9.88508880e-01 5.08542955e-01 1.10517018e-01 3.63834918e-01
-4.79523331e-01 3.41063440e-01 -8.98212373e-01 -8.91494513e-01
2.42789090e-01 -1.70144513e-01 -9.41167474e-01 -6.37272775e-01
8.22364330e-01 2.77306050e-01 2.62891382e-01 9.70564067e-01
6.81756198e-01 1.19115850e-02 4.48004514e-01 -5.41961670e-01
-7.48442590e-01 1.38465106e+00 -4.73403752e-01 4.00254369e-01
2.45586395e-01 2.44123250e-01 1.37993431e+00 -1.01109004e+00
8.65317225e-01 1.34581971e+00 8.16407144e-01 4.27203208e-01
-1.48833776e+00 -6.33857548e-01 4.41732436e-01 -1.42758014e-02
-1.45208716e+00 -3.46954286e-01 3.83668631e-01 -2.30759993e-01
1.43185544e+00 4.34668183e-01 3.54563385e-01 9.32513773e-01
1.01649612e-01 8.60035777e-01 1.20779848e+00 -3.45329463e-01
-5.73742203e-02 3.58572491e-02 2.29852140e-01 3.22108418e-01
4.96758819e-01 -1.94889322e-01 -5.70330262e-01 3.79761830e-02
1.07785024e-01 -7.74945974e-01 -3.28777701e-01 1.37640297e-01
-1.25845063e+00 6.40536666e-01 4.71481204e-01 5.65945387e-01
-2.17092901e-01 -7.32448772e-02 6.94099307e-01 5.07534862e-01
7.85879791e-01 9.58617806e-01 -9.82585609e-01 -2.27680922e-01
-1.07483792e+00 4.67908084e-01 7.73817480e-01 5.94943464e-01
5.15654922e-01 1.74087107e-01 -1.53393567e-01 1.14039028e+00
-2.46835768e-01 4.80993241e-01 6.20837212e-01 -4.81298417e-01
7.59319365e-01 4.57774103e-01 -2.96043813e-01 -4.61901814e-01
-5.38005292e-01 -7.78001010e-01 -6.21038318e-01 8.67232829e-02
9.18622494e-01 -2.43794054e-01 -8.24155450e-01 2.03976607e+00
-3.89446542e-02 -3.75077605e-01 -1.12088300e-01 6.19482994e-01
6.01221204e-01 7.53969908e-01 2.83551872e-01 -1.66054204e-01
1.35544491e+00 -6.24043047e-01 -5.35095215e-01 -4.99218225e-01
8.25084388e-01 -9.04758036e-01 1.64301991e+00 7.92234480e-01
-1.46908760e+00 -5.47265530e-01 -1.18034315e+00 -2.64839292e-01
-7.47389972e-01 -2.89391011e-01 5.47832787e-01 5.81274986e-01
-1.19459081e+00 7.26921976e-01 -6.61041439e-01 -2.70694017e-01
2.10263148e-01 3.48764569e-01 -3.85373414e-01 -1.74533967e-02
-1.15191543e+00 1.22031891e+00 6.69510901e-01 -1.11050814e-01
-3.98320258e-01 -1.25323188e+00 -6.68324351e-01 -6.69298619e-02
2.92258382e-01 -3.99321407e-01 1.13340187e+00 -9.41939712e-01
-1.25214922e+00 1.23852146e+00 6.01072386e-02 -6.27903283e-01
4.85678196e-01 -5.05572259e-01 -3.24254960e-01 -5.36505878e-01
-8.57233331e-02 8.17798436e-01 1.29712820e-01 -1.22170615e+00
-6.51858211e-01 -3.81928802e-01 -1.20297387e-01 1.13924339e-01
-2.14687765e-01 4.24983352e-01 -1.81808248e-01 -7.66410053e-01
2.39849016e-02 -8.40788066e-01 -2.44237091e-02 -6.34826362e-01
-2.76816010e-01 -4.52944785e-01 2.07372978e-01 -7.83394217e-01
1.40393436e+00 -2.01879692e+00 2.41018593e-01 -2.08785787e-01
-1.81444570e-01 3.69437903e-01 -3.27987671e-01 2.52552569e-01
-2.74766415e-01 3.47224265e-01 -6.31173193e-01 -5.27324259e-01
1.59946736e-02 1.56057358e-01 -3.06917846e-01 2.16575637e-01
5.75260103e-01 9.99454319e-01 -8.93178105e-01 -1.13783054e-01
-2.64415115e-01 2.77460873e-01 -5.99313438e-01 -2.62483329e-01
-3.60625833e-01 8.09341203e-03 3.72143090e-01 4.59614575e-01
6.69791579e-01 2.55941510e-01 9.11948010e-02 1.35830760e-01
-5.74703038e-01 8.37152839e-01 -1.06038320e+00 1.93153906e+00
-5.85245967e-01 3.63064557e-01 1.76077321e-01 -4.18732882e-01
8.06075275e-01 4.15177830e-02 -7.70101398e-02 -7.28633404e-01
-1.53821021e-01 6.86957300e-01 6.29185140e-01 1.78126395e-02
7.37044990e-01 -3.36083651e-01 -3.63354921e-01 2.68419504e-01
2.96535566e-02 -4.89397824e-01 4.00891185e-01 1.35167427e-02
1.09371448e+00 7.54261315e-02 3.43817413e-01 -7.71269560e-01
5.02979696e-01 -7.73085877e-02 5.45543969e-01 4.77288038e-01
6.60904869e-02 6.38195872e-01 5.17484307e-01 -3.19147021e-01
-8.50317061e-01 -1.28302276e+00 -3.62926185e-01 1.32971907e+00
-5.21413505e-01 -5.51399529e-01 -8.23428035e-01 -6.77490771e-01
4.03057784e-02 1.06407690e+00 -5.29924333e-01 -1.25812307e-01
-1.06838989e+00 -1.07211339e+00 9.20959353e-01 3.92476827e-01
3.05619091e-01 -1.44067907e+00 -2.67846853e-01 2.28099823e-01
-1.00043707e-01 -8.11409473e-01 -2.98592359e-01 6.56047225e-01
-7.54761219e-01 -5.07786572e-01 -8.07715356e-01 -9.14938629e-01
1.48312360e-01 -3.10365319e-01 1.56573057e+00 -5.51458895e-02
-9.57032889e-02 -1.91600487e-01 -2.53631115e-01 -6.62158549e-01
-4.92363393e-01 6.38066292e-01 -1.09106660e-01 -3.41666311e-01
3.70175213e-01 -3.69893640e-01 -1.49427980e-01 -2.62987405e-01
-8.24369311e-01 -3.87520969e-01 7.43789077e-01 6.55563474e-01
5.47544897e-01 -3.82670730e-01 8.55068624e-01 -9.76016760e-01
7.10149884e-01 -4.23016042e-01 -4.80203658e-01 4.92099151e-02
-5.60448468e-01 1.48837954e-01 8.78479362e-01 -4.00064111e-01
-8.66995513e-01 -2.18854532e-01 -3.43010008e-01 1.12226598e-01
-2.04278514e-01 5.99099100e-01 -4.18487430e-01 4.58192050e-01
7.58807003e-01 -2.20998097e-02 -2.05550611e-01 -9.03145432e-01
4.44177210e-01 3.88972938e-01 6.41579926e-01 -8.08656037e-01
7.51876771e-01 1.37403280e-01 -2.82508731e-01 -8.53493214e-01
-1.01388705e+00 -3.15638870e-01 -5.81019640e-01 4.22955126e-01
8.47249866e-01 -7.75139689e-01 -2.68763006e-01 5.03489554e-01
-1.30392361e+00 -6.50656879e-01 -5.32326460e-01 2.88015306e-01
-1.72351867e-01 1.45698875e-01 -9.57606852e-01 -4.54285383e-01
-4.91508216e-01 -1.07710457e+00 9.42383945e-01 -2.43048489e-01
-4.97825027e-01 -9.54257071e-01 3.73073816e-01 5.70151918e-02
5.66604018e-01 4.15500961e-02 1.48479557e+00 -9.42246795e-01
-2.45810181e-01 2.25448832e-01 -6.15975969e-02 6.01227582e-01
9.34513062e-02 -4.29289676e-02 -9.93588448e-01 -3.00934404e-01
-6.29942715e-02 -3.05195540e-01 1.12844002e+00 1.89782292e-01
9.89383280e-01 -8.65939409e-02 1.24641694e-01 5.88715851e-01
1.32652640e+00 -1.61412343e-01 6.54541135e-01 3.97580117e-01
4.28643316e-01 5.88832557e-01 3.06927472e-01 -1.04907244e-01
3.81713927e-01 6.86259210e-01 2.67219424e-01 -1.39821125e-02
-5.12341082e-01 -1.34606391e-01 6.45847857e-01 1.29027092e+00
1.27215967e-01 -3.14506531e-01 -1.14690566e+00 8.27743411e-01
-1.31428432e+00 -7.44152904e-01 -4.88632232e-01 2.56861997e+00
1.38813937e+00 4.32110339e-01 1.98479936e-01 1.72126636e-01
5.16223013e-01 1.64864227e-01 -1.64051726e-01 -9.27396894e-01
-6.69185221e-01 7.52744079e-01 3.65060776e-01 7.40596831e-01
-8.67178023e-01 1.42399168e+00 6.56832933e+00 1.04504573e+00
-1.18477213e+00 1.04828253e-01 5.19773126e-01 -3.03919286e-01
-4.68042642e-01 -1.36389554e-01 -9.98144031e-01 5.34004986e-01
1.25131559e+00 -2.11847842e-01 3.18454057e-01 3.66111130e-01
-7.88282230e-02 -2.39061862e-02 -1.18432510e+00 6.77410185e-01
1.34326220e-01 -8.33466172e-01 2.31331065e-01 2.28921901e-02
7.47810423e-01 5.10903180e-01 1.74555138e-01 6.74055874e-01
3.61318320e-01 -1.17391169e+00 8.78757417e-01 1.93750471e-01
8.06520641e-01 -9.75396633e-01 6.73714936e-01 2.76554585e-01
-9.13997233e-01 2.48634085e-01 -5.45232713e-01 -2.21634164e-01
1.59779444e-01 7.94708073e-01 -6.43478632e-01 5.09510577e-01
5.43113589e-01 3.18470687e-01 -8.99863839e-01 1.10247636e+00
-4.02066112e-01 1.05259156e+00 -4.92669076e-01 5.86044788e-02
3.36677551e-01 3.91367404e-03 6.11443460e-01 1.81072712e+00
2.06859440e-01 -2.43527427e-01 4.37258556e-02 6.66242063e-01
-2.66387969e-01 7.04298794e-01 -5.36415875e-01 2.92172939e-01
2.50848472e-01 1.24987602e+00 -5.16244888e-01 -2.82084554e-01
-3.34952623e-01 8.63008320e-01 7.50724375e-01 -3.48124802e-02
-7.90956795e-01 -4.54026043e-01 6.84024513e-01 1.24464266e-01
3.29226106e-01 -3.05392593e-01 -5.13759673e-01 -8.10556352e-01
7.92928785e-02 -1.14859819e+00 4.50351954e-01 -4.54481035e-01
-1.43366933e+00 7.62538314e-01 -1.06407013e-02 -6.83333576e-01
-1.04995385e-01 -8.19372058e-01 -3.77686739e-01 9.58904386e-01
-1.50876462e+00 -1.04649758e+00 1.92347482e-01 3.43375981e-01
6.52012706e-01 1.71980634e-02 8.82522106e-01 4.51745093e-01
-3.78028989e-01 9.07844484e-01 -1.69614721e-02 1.41686141e-01
1.07164645e+00 -1.77585125e+00 8.09759855e-01 8.80155623e-01
5.54171503e-01 7.23364294e-01 7.67097592e-01 -5.87575078e-01
-9.76144969e-01 -8.85742903e-01 1.34636605e+00 -7.53479540e-01
8.28345656e-01 -6.13844931e-01 -1.36600888e+00 6.79841518e-01
6.28422976e-01 -1.91872984e-01 6.43122375e-01 6.26746118e-01
-7.39020050e-01 -1.55290306e-01 -8.28631818e-01 7.98465550e-01
1.31032991e+00 -2.90773362e-01 -9.48366940e-01 2.60659456e-01
9.57617283e-01 -3.25100392e-01 -6.27813399e-01 4.01454926e-01
2.11219907e-01 -7.96997070e-01 6.54301465e-01 -7.37129688e-01
4.73649412e-01 -1.11141466e-01 -2.83671170e-01 -1.85757363e+00
-3.52540433e-01 -5.12196124e-01 4.14372981e-01 1.55394554e+00
9.86771882e-01 -6.49929881e-01 6.11472845e-01 1.49536701e-02
-4.68078226e-01 -5.77647984e-01 -1.09707248e+00 -1.16637588e+00
1.07262933e+00 -5.82905769e-01 5.79240859e-01 9.06020164e-01
-5.05721495e-02 6.53056979e-01 2.22265899e-01 -2.06439316e-01
2.49725521e-01 -2.31128052e-01 4.85592961e-01 -1.17846155e+00
-4.44177538e-01 -8.89531672e-01 -2.94297338e-01 -7.83557713e-01
3.99441451e-01 -1.59716308e+00 1.62300423e-01 -1.26520741e+00
1.89782694e-01 -4.40098971e-01 -4.66909051e-01 5.31002164e-01
-3.62207383e-01 3.16885829e-01 3.89502496e-01 -1.72172293e-01
-4.16269898e-01 1.81041062e-01 9.03213978e-01 2.52458593e-03
-1.76017120e-01 -2.18554929e-01 -8.07781935e-01 7.33694613e-01
8.23265076e-01 -4.55389172e-01 -6.76381364e-02 -9.57657576e-01
6.56339824e-01 -6.23546064e-01 3.08724698e-02 -9.73684132e-01
2.85659488e-02 2.08505213e-01 7.29954541e-02 -5.69701910e-01
2.28555232e-01 -3.64196092e-01 -2.31129035e-01 5.73519886e-01
-3.21873397e-01 6.65114462e-01 7.52597749e-01 -2.47479021e-03
-1.64726287e-01 -2.67601430e-01 6.50453269e-01 -2.33572721e-01
-3.95985842e-01 -1.49590105e-01 -2.99478531e-01 7.48824537e-01
6.44605875e-01 5.93598709e-02 -4.21012431e-01 5.41680865e-02
-5.90022445e-01 2.92682368e-02 3.12981725e-01 5.37656069e-01
8.72034431e-02 -9.68070090e-01 -1.29640079e+00 1.46934807e-01
1.27464339e-01 1.68111697e-02 -8.22466537e-02 5.92844844e-01
-3.06050450e-01 4.52118665e-01 4.93518775e-03 -3.99565369e-01
-1.05974185e+00 4.50871795e-01 2.84061849e-01 -6.60697818e-01
-4.55522478e-01 1.11185002e+00 2.78964549e-01 -7.13506997e-01
7.50835314e-02 -6.57156646e-01 7.60482699e-02 4.55424190e-01
4.33419079e-01 3.38183492e-01 6.35125458e-01 -5.20910740e-01
-4.74615157e-01 3.86861652e-01 -3.67536068e-01 -3.16585362e-01
1.42081511e+00 3.33680451e-01 -3.30344856e-01 8.44026864e-01
1.10398364e+00 6.93546474e-01 -6.88839912e-01 -2.80180816e-02
6.20462716e-01 3.74651588e-02 -2.56480187e-01 -1.36836088e+00
-7.83228517e-01 9.02172029e-01 1.36746332e-01 5.33134900e-02
7.94818103e-01 5.42726293e-02 6.90737009e-01 2.71874517e-01
5.02429485e-01 -1.04084051e+00 -3.27664971e-01 1.09318900e+00
1.20445728e+00 -9.51456547e-01 -1.21178798e-01 -2.42998272e-01
-4.35275555e-01 6.37861490e-01 5.75429559e-01 -2.24071741e-01
3.26960027e-01 5.09704292e-01 1.63459778e-01 4.77526197e-03
-9.60682213e-01 -1.96469754e-01 2.14166492e-01 2.93914348e-01
9.32698250e-01 2.40409836e-01 -5.96704066e-01 8.01632345e-01
-8.13890934e-01 -5.93446434e-01 4.42760080e-01 4.56322223e-01
-4.77378756e-01 -1.32476139e+00 -2.12132797e-01 3.53390932e-01
-9.27537739e-01 -6.74305677e-01 -6.37826502e-01 1.22217572e+00
2.45791063e-01 5.91600835e-01 2.69084632e-01 4.08259071e-02
6.04249775e-01 6.06542349e-01 5.76284051e-01 -9.73733962e-01
-1.32190001e+00 -2.08966359e-01 2.92552918e-01 -4.10675108e-01
3.38366270e-01 -9.62577403e-01 -1.25279546e+00 -1.67233467e-01
-1.62438199e-01 6.07872494e-02 6.52954817e-01 7.05508471e-01
1.31142646e-01 4.81362939e-01 -1.40407141e-02 -6.29121006e-01
-8.98280680e-01 -1.20997894e+00 -4.03709471e-01 4.13785934e-01
8.34243968e-02 -2.34356835e-01 -7.15380728e-01 -4.04137105e-01] | [10.629302978515625, 9.796550750732422] |
d0d72799-fbf8-4574-8f01-0e3b76d9ac89 | learning-roi-transformer-for-detecting | 1812.00155 | null | http://arxiv.org/abs/1812.00155v1 | http://arxiv.org/pdf/1812.00155v1.pdf | Learning RoI Transformer for Detecting Oriented Objects in Aerial Images | Object detection in aerial images is an active yet challenging task in
computer vision because of the birdview perspective, the highly complex
backgrounds, and the variant appearances of objects. Especially when detecting
densely packed objects in aerial images, methods relying on horizontal
proposals for common object detection often introduce mismatches between the
Region of Interests (RoIs) and objects. This leads to the common misalignment
between the final object classification confidence and localization accuracy.
Although rotated anchors have been used to tackle this problem, the design of
them always multiplies the number of anchors and dramatically increases the
computational complexity. In this paper, we propose a RoI Transformer to
address these problems. More precisely, to improve the quality of region
proposals, we first designed a Rotated RoI (RRoI) learner to transform a
Horizontal Region of Interest (HRoI) into a Rotated Region of Interest (RRoI).
Based on the RRoIs, we then proposed a Rotated Position Sensitive RoI Align
(RPS-RoI-Align) module to extract rotation-invariant features from them for
boosting subsequent classification and regression. Our RoI Transformer is with
light weight and can be easily embedded into detectors for oriented object
detection. A simple implementation of the RoI Transformer has achieved
state-of-the-art performances on two common and challenging aerial datasets,
i.e., DOTA and HRSC2016, with a neglectable reduction to detection speed. Our
RoI Transformer exceeds the deformable Position Sensitive RoI pooling when
oriented bounding-box annotations are available. Extensive experiments have
also validated the flexibility and effectiveness of our RoI Transformer. The
results demonstrate that it can be easily integrated with other detector
architectures and significantly improve the performances. | ['Gui-Song Xia', 'Yang Long', 'Qikai Lu', 'Jian Ding', 'Nan Xue'] | 2018-12-01 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 2.16578335e-01 -2.15226650e-01 1.25581726e-01 -2.92328179e-01
-3.58352274e-01 -6.29524171e-01 3.29725415e-01 -1.16102412e-01
-6.44809663e-01 3.16521615e-01 -2.59242237e-01 -3.15358862e-02
1.17010690e-01 -6.82795525e-01 -6.27125740e-01 -7.19386160e-01
-5.16138524e-02 -1.10722028e-01 9.83875513e-01 -1.88428238e-01
5.79579435e-02 7.96487808e-01 -1.73925447e+00 3.85553837e-01
6.51122987e-01 1.20612288e+00 5.31298041e-01 3.95012677e-01
3.38396013e-01 4.15129334e-01 -6.78132117e-01 -1.36394396e-01
6.99256778e-01 -1.10139988e-01 -4.17088330e-01 1.17520019e-02
5.60078144e-01 -5.89401901e-01 -7.76983276e-02 9.21677947e-01
5.52708566e-01 1.54276080e-02 5.96511662e-01 -1.06456268e+00
-3.36549789e-01 4.82397318e-01 -1.12881291e+00 5.04961491e-01
1.82698518e-01 9.57021043e-02 9.39660847e-01 -1.15114725e+00
3.88054073e-01 9.95343208e-01 7.69293547e-01 3.31886470e-01
-1.04045939e+00 -7.05604672e-01 5.86307406e-01 -5.19125164e-02
-1.68474901e+00 -9.26603675e-02 4.64677095e-01 -4.16285604e-01
6.99104846e-01 6.43214107e-01 6.59262419e-01 8.17058325e-01
5.90388402e-02 7.34064817e-01 1.04508424e+00 -3.45285445e-01
6.89592734e-02 3.54377598e-01 -4.58535692e-03 7.63160527e-01
3.91085476e-01 1.04971990e-01 -2.03603152e-02 1.35302261e-01
7.49099851e-01 3.65851432e-01 -4.32216197e-01 -5.40017307e-01
-1.27577150e+00 6.23095751e-01 1.07385898e+00 2.19776183e-01
-1.62313461e-01 -1.16716385e-01 8.75084922e-02 -1.95592254e-01
2.71773189e-01 5.75763047e-01 -3.60061079e-01 7.81099975e-01
-9.58625615e-01 1.62172988e-01 2.41810903e-01 9.82795238e-01
5.09267747e-01 -1.64781094e-01 -4.74163622e-01 6.36633515e-01
3.26071143e-01 5.62843919e-01 3.11008811e-01 -1.74281731e-01
6.05252028e-01 9.68307853e-01 2.87153572e-01 -1.27932668e+00
-8.00093770e-01 -7.19354868e-01 -6.49212003e-01 3.41237456e-01
5.85315049e-01 -3.07410415e-02 -8.12067986e-01 1.36664724e+00
6.01764143e-01 -1.84902564e-01 -2.43376121e-01 1.42989874e+00
8.97253454e-01 6.12403512e-01 -8.11048672e-02 2.32390225e-01
1.63094759e+00 -9.45409358e-01 -1.13319740e-01 -5.56407630e-01
5.64812183e-01 -7.68012047e-01 9.39595580e-01 1.04466349e-01
-6.52182639e-01 -7.64371276e-01 -1.28316772e+00 8.87697786e-02
-4.96820211e-01 9.45956469e-01 4.46547538e-01 5.89784980e-01
-6.50182724e-01 2.92598635e-01 -6.11177146e-01 -4.31344748e-01
4.37113464e-01 5.04022241e-01 -3.76192689e-01 7.50719309e-02
-6.90919459e-01 8.25362027e-01 4.97011632e-01 4.65504736e-01
-7.24277377e-01 -4.06040698e-01 -6.95021093e-01 -8.59926492e-02
5.80772042e-01 -3.25924814e-01 8.07281613e-01 -7.86788404e-01
-1.26513886e+00 7.11751938e-01 3.75043005e-01 -4.44908231e-01
5.89130342e-01 -3.15196872e-01 -1.92831472e-01 1.23267630e-02
4.97193635e-02 9.74710047e-01 1.22732770e+00 -9.32843328e-01
-1.01853704e+00 -5.03251433e-01 3.11223030e-01 2.33460575e-01
-3.62139493e-01 2.96708107e-01 -3.04833502e-01 -7.48372436e-01
3.24321181e-01 -9.72914159e-01 -2.68801242e-01 3.38961631e-01
-2.51529783e-01 -5.62561154e-02 1.01835263e+00 -4.45808470e-01
1.20214152e+00 -2.23536086e+00 1.37374684e-01 -9.27070603e-02
3.07511017e-02 4.92171973e-01 -1.85537323e-01 -5.44633567e-02
-1.11045472e-01 -8.56449082e-02 -1.80531710e-01 9.84308720e-02
-3.19499135e-01 -3.02752674e-01 -5.68809509e-01 7.83132970e-01
6.16815031e-01 7.57614434e-01 -7.27031052e-01 -4.96269464e-01
4.58808243e-01 3.93623531e-01 -5.69231629e-01 2.06704170e-01
7.26884604e-02 1.73719808e-01 -5.41346967e-01 9.42304432e-01
8.57262671e-01 -1.63384676e-02 -1.65141582e-01 -6.19033277e-01
-4.56415623e-01 -6.52062967e-02 -1.45759439e+00 1.15191078e+00
-1.70927018e-01 5.23644805e-01 8.43969174e-03 -6.41628861e-01
1.13556421e+00 -1.11241199e-01 1.58104524e-01 -3.27740163e-01
2.26285085e-01 1.00693867e-01 4.48036715e-02 -1.58417553e-01
5.93865991e-01 1.50131404e-01 -4.29958738e-02 -6.66081905e-02
-1.09953461e-02 -1.75000623e-01 1.09038867e-01 -3.06464702e-01
8.53131592e-01 4.55853373e-01 7.17008114e-01 -2.28274614e-01
5.97036123e-01 -5.84164746e-02 5.84665775e-01 5.80009580e-01
-2.61871487e-01 8.54535103e-01 2.36766264e-01 -6.62387908e-01
-6.45623922e-01 -8.07846963e-01 -4.39987540e-01 1.11098850e+00
4.61624801e-01 -2.57985473e-01 -6.82812631e-01 -1.10657930e+00
-8.11280236e-02 2.86591232e-01 -6.22711301e-01 3.14197652e-02
-5.34024000e-01 -9.83777165e-01 4.24366415e-01 7.21352220e-01
6.92437708e-01 -8.48735750e-01 -1.33931470e+00 -4.55146134e-02
2.19341628e-02 -1.19878340e+00 -4.71240729e-01 4.11468357e-01
-6.36131406e-01 -1.03418410e+00 -8.05001795e-01 -5.38740575e-01
8.79252970e-01 8.70148420e-01 6.55210018e-01 -1.06662542e-01
-8.40254605e-01 5.11308899e-03 -5.58448434e-01 -5.38028538e-01
3.10034037e-01 1.53917233e-02 9.88792777e-02 1.08829945e-01
6.72839209e-02 1.90947577e-02 -8.07552457e-01 1.06679749e+00
-9.32970226e-01 -7.73698166e-02 8.27358842e-01 8.46576631e-01
5.04717290e-01 1.37149787e-03 1.72846690e-01 -3.81243974e-01
-3.31923850e-02 -5.00632599e-02 -1.18245208e+00 2.70043761e-01
-1.85934663e-01 -9.29633528e-02 6.10812545e-01 -4.47271228e-01
-6.79756165e-01 5.20241976e-01 2.92664230e-01 -2.33465746e-01
-1.30323291e-01 9.32010077e-03 -1.90608099e-01 -3.93095732e-01
7.98032463e-01 -1.96971565e-01 -4.02310103e-01 -1.81413174e-01
2.15367258e-01 7.27174938e-01 2.40661278e-01 -3.63179445e-02
9.88067627e-01 5.15640974e-01 -5.73488101e-02 -8.65090072e-01
-9.50347781e-01 -7.05030859e-01 -8.33774686e-01 -3.09090525e-01
8.16377461e-01 -1.11702120e+00 -4.06509250e-01 2.38537863e-01
-1.15067267e+00 -1.43286699e-04 -3.26121122e-01 5.56325972e-01
-3.56260315e-02 2.64402032e-01 -5.97561970e-02 -7.80559957e-01
-2.40603760e-01 -1.34305108e+00 1.35163975e+00 5.32531857e-01
7.94741586e-02 -3.98644954e-01 -4.27818388e-01 2.04174846e-01
2.73510933e-01 3.53139848e-01 3.81395966e-01 -4.35891360e-01
-7.25491703e-01 -5.62270403e-01 -6.60866976e-01 3.28890294e-01
-4.96452972e-02 1.10411562e-01 -1.12088394e+00 -4.29905087e-01
-2.21748307e-01 -1.96011990e-01 9.67586815e-01 3.26068372e-01
1.12625384e+00 -2.13237572e-02 -6.00717902e-01 6.74712956e-01
1.38220215e+00 2.72362605e-02 4.42801982e-01 2.96295077e-01
4.92497683e-01 6.36448026e-01 1.12302887e+00 5.82773626e-01
-8.93627331e-02 1.14122784e+00 7.25411594e-01 -4.52944279e-01
-6.47867396e-02 3.24876569e-02 6.40496135e-01 6.79472275e-03
-4.12971318e-01 -2.83012509e-01 -6.37016177e-01 2.35328794e-01
-1.63413143e+00 -8.63843799e-01 -1.13545932e-01 2.22985315e+00
3.72333288e-01 7.99371973e-02 1.96903899e-01 -1.28547195e-02
8.04424167e-01 2.41029352e-01 -4.04165238e-01 1.69037998e-01
-1.21059604e-01 -2.81370748e-02 1.03665340e+00 -1.18528686e-01
-1.69124353e+00 9.80777442e-01 5.76176977e+00 7.51799643e-01
-1.19621503e+00 -1.41315505e-01 3.23147386e-01 -8.82824883e-02
6.11982167e-01 -2.33981162e-01 -1.43804264e+00 3.14968199e-01
1.92912638e-01 4.01798189e-01 1.81861296e-01 1.26612031e+00
-1.07811913e-01 -1.32657632e-01 -9.07185733e-01 8.55195403e-01
1.20373212e-01 -8.50450635e-01 -1.17513292e-01 -6.95275841e-03
3.96068245e-01 -2.16691364e-02 1.79736063e-01 2.91919261e-01
-1.38110623e-01 -8.33710611e-01 9.89389598e-01 8.21622908e-02
7.18843460e-01 -5.66331983e-01 9.79386210e-01 2.54135221e-01
-1.55166721e+00 -5.78790426e-01 -8.89038980e-01 2.29857527e-02
-2.39544287e-01 3.22950095e-01 -1.04394257e+00 5.62263966e-01
1.01133466e+00 5.76416612e-01 -1.05063605e+00 1.23244619e+00
-1.83997929e-01 1.85215339e-01 -4.79438037e-01 -1.24434255e-01
2.38650993e-01 -1.37840211e-01 5.59572041e-01 1.25405574e+00
4.31692928e-01 -2.63682324e-02 2.96255559e-01 9.21913981e-01
1.12756580e-01 1.31427243e-01 -6.51473343e-01 2.18862146e-01
1.62051216e-01 1.81647241e+00 -1.11105549e+00 3.75554636e-02
-3.16283017e-01 8.83782089e-01 2.56394267e-01 3.97826806e-02
-1.27881861e+00 -4.83044267e-01 3.15652192e-01 3.16119581e-01
8.50805461e-01 -3.98033746e-02 2.22609833e-01 -1.08340788e+00
2.65342981e-01 -6.97654605e-01 3.08620930e-01 -5.13575554e-01
-1.01974177e+00 8.08629155e-01 1.95334196e-01 -1.63975656e+00
3.63968641e-01 -1.01748562e+00 -3.86803031e-01 5.50999582e-01
-1.50375915e+00 -1.31285298e+00 -8.59509587e-01 2.39733636e-01
5.69702148e-01 -1.17831148e-01 6.55621350e-01 2.13863879e-01
-9.17342782e-01 5.55287182e-01 -2.98600614e-01 2.38151371e-01
7.34231174e-01 -1.02464128e+00 1.84380159e-01 1.12713182e+00
2.89808244e-01 3.80473733e-01 4.11789715e-01 -3.12866211e-01
-1.19791186e+00 -1.44829166e+00 9.93230194e-02 -5.28599918e-01
3.32384914e-01 -5.53240478e-01 -6.43393695e-01 4.31314856e-01
-2.69457310e-01 6.20380700e-01 3.77356529e-01 -3.60610217e-01
-3.50568652e-01 -4.19967741e-01 -1.03723216e+00 5.52243710e-01
1.00362742e+00 -3.06165107e-02 -4.29778486e-01 2.90242404e-01
5.77744246e-01 -4.41966236e-01 -5.01493990e-01 7.47790158e-01
5.46724856e-01 -1.22744131e+00 1.21327031e+00 -2.00653672e-01
3.75403203e-02 -9.04517114e-01 -2.63653845e-01 -9.21818376e-01
-4.78200167e-01 -1.95387691e-01 2.12283999e-01 1.11472499e+00
2.73938179e-01 -5.18046796e-01 3.90537024e-01 2.30071381e-01
-8.41024220e-02 -7.80999780e-01 -9.00721908e-01 -8.26440454e-01
-5.72227299e-01 3.05427844e-03 4.84196156e-01 4.78025138e-01
-2.67349660e-01 1.96447894e-01 -1.89637303e-01 6.07489884e-01
3.32665145e-01 2.59386599e-01 7.64500260e-01 -1.19476938e+00
-2.13465065e-01 -2.92820215e-01 -7.96683073e-01 -1.14165854e+00
-3.56855333e-01 -7.72920728e-01 2.47495398e-01 -1.21044815e+00
-2.72982158e-02 -6.19054019e-01 -3.25840235e-01 5.37152052e-01
-2.37922162e-01 6.84400380e-01 2.86642462e-01 6.13874421e-02
-6.36448324e-01 4.13775057e-01 9.46403503e-01 -1.58580124e-01
-3.54628325e-01 2.20624998e-01 -3.31547081e-01 9.16420996e-01
7.12465942e-01 -5.22183955e-01 -9.80759785e-02 -2.30094820e-01
-4.95108180e-02 -4.10263330e-01 7.25355685e-01 -1.47617924e+00
8.78077373e-02 8.67938921e-02 8.78427923e-01 -8.66348267e-01
3.95746261e-01 -1.03596187e+00 -2.28692800e-01 5.82762957e-01
-5.11192530e-02 1.05804697e-01 3.66542548e-01 6.20600045e-01
6.93518296e-03 -3.43602836e-01 1.05069315e+00 -5.05402014e-02
-7.10727453e-01 6.92718253e-02 -1.54297099e-01 -2.79965788e-01
1.60170555e+00 -3.40374947e-01 -2.34116331e-01 4.31849003e-01
-1.65071666e-01 2.73602437e-02 4.98523444e-01 4.57019389e-01
8.03792953e-01 -9.62419808e-01 -5.45094311e-01 4.79007572e-01
3.51047575e-01 2.35098138e-01 -1.49602834e-02 1.05499530e+00
-5.35932362e-01 5.62102914e-01 -3.00637990e-01 -8.07380974e-01
-1.45886195e+00 7.16607153e-01 4.11449701e-01 -1.82352737e-01
-7.08450139e-01 9.38905478e-01 5.37704945e-01 -2.95353204e-01
2.67540902e-01 -9.09364700e-01 -4.26149130e-01 3.02853674e-01
7.22582638e-01 3.04174215e-01 5.85025772e-02 -4.96241093e-01
-5.91562688e-01 8.90467107e-01 -1.76146030e-01 4.08154309e-01
1.16550410e+00 9.12570953e-02 1.01056568e-01 8.37710053e-02
5.71832061e-01 6.02042302e-02 -1.36395597e+00 6.63984381e-03
-1.69381440e-01 -6.72173440e-01 1.21171400e-01 -5.49186170e-01
-1.00292110e+00 7.55215347e-01 9.04339731e-01 3.53448316e-02
1.02404642e+00 -8.55699852e-02 3.39800417e-01 4.00969356e-01
5.43519974e-01 -8.45119596e-01 1.19139880e-01 8.53972882e-03
1.02173495e+00 -1.39210129e+00 4.49031413e-01 -5.77960312e-01
-6.43685102e-01 1.23122942e+00 8.47583055e-01 -1.60823986e-01
2.84960151e-01 2.04877079e-01 -9.07858983e-02 -2.84063183e-02
-2.52214581e-01 -4.89030808e-01 6.76733553e-01 5.67714810e-01
2.38632232e-01 -1.35523109e-02 -5.85327595e-02 5.66714644e-01
1.68397561e-01 -3.36510599e-01 2.72262573e-01 7.38518476e-01
-5.77097118e-01 -6.94037557e-01 -6.81026578e-01 2.07750246e-01
-2.44622335e-01 3.88408229e-02 -4.37634677e-01 1.01030910e+00
4.45927709e-01 8.01941931e-01 -5.78408055e-02 -4.51120406e-01
5.04666924e-01 -3.74647498e-01 4.79730397e-01 -6.51000440e-01
-7.53811717e-01 -4.97022877e-03 -3.45581681e-01 -6.40711248e-01
-3.28255385e-01 -4.57968712e-01 -1.01829970e+00 4.14027661e-01
-1.00338519e+00 -2.15479195e-01 6.76170826e-01 5.42705357e-01
2.84384906e-01 6.37936234e-01 6.20711446e-01 -1.02013862e+00
-7.73360491e-01 -8.52158606e-01 -4.72243845e-01 -8.40322450e-02
1.68480784e-01 -9.42798138e-01 -2.82355815e-01 -1.82151258e-01] | [8.70976448059082, -0.7398063540458679] |
0c4511ac-6ed9-4f91-926b-b40c8ba15a07 | in-the-name-of-fairness-assessing-the-bias-in | 2305.11348 | null | https://arxiv.org/abs/2305.11348v1 | https://arxiv.org/pdf/2305.11348v1.pdf | In the Name of Fairness: Assessing the Bias in Clinical Record De-identification | Data sharing is crucial for open science and reproducible research, but the legal sharing of clinical data requires the removal of protected health information from electronic health records. This process, known as de-identification, is often achieved through the use of machine learning algorithms by many commercial and open-source systems. While these systems have shown compelling results on average, the variation in their performance across different demographic groups has not been thoroughly examined. In this work, we investigate the bias of de-identification systems on names in clinical notes via a large-scale empirical analysis. To achieve this, we create 16 name sets that vary along four demographic dimensions: gender, race, name popularity, and the decade of popularity. We insert these names into 100 manually curated clinical templates and evaluate the performance of nine public and private de-identification methods. Our findings reveal that there are statistically significant performance gaps along a majority of the demographic dimensions in most methods. We further illustrate that de-identification quality is affected by polysemy in names, gender context, and clinical note characteristics. To mitigate the identified gaps, we propose a simple and method-agnostic solution by fine-tuning de-identification methods with clinical context and diverse names. Overall, it is imperative to address the bias in existing methods immediately so that downstream stakeholders can build high-quality systems to serve all demographic parties fairly. | ['Marzyeh Ghassemi', 'Tom Joseph Pollard', 'Shulammite Lim', 'Yuxin Xiao'] | 2023-05-18 | null | null | null | null | ['de-identification'] | ['natural-language-processing'] | [-8.26066956e-02 -1.22431196e-01 -3.58302414e-01 -3.36106867e-01
-1.07402527e+00 -9.89789605e-01 2.78356969e-01 7.37283170e-01
-4.32028800e-01 6.80158615e-01 7.35261500e-01 -6.10166192e-01
-4.85910237e-01 -5.76025009e-01 -2.04818562e-01 -4.37574267e-01
3.53408009e-01 4.86851186e-01 -3.69306117e-01 3.75394166e-01
3.44417542e-01 4.59685773e-01 -8.56818795e-01 3.27816576e-01
1.02398133e+00 4.93401349e-01 -5.44353962e-01 4.04877603e-01
-1.77066624e-01 3.09240192e-01 -7.11014569e-01 -6.61928535e-01
4.31639910e-01 -3.44616830e-01 -8.21894467e-01 -4.98814404e-01
5.37654757e-01 -2.46104911e-01 1.88701227e-02 9.78039443e-01
9.64376092e-01 -6.04951739e-01 5.99245250e-01 -8.03781807e-01
-9.59972501e-01 9.01624441e-01 -4.46994990e-01 2.06789643e-01
4.18801486e-01 3.56562883e-01 1.13518131e+00 -2.84904808e-01
9.73873377e-01 8.41336429e-01 1.26951468e+00 5.13519228e-01
-1.36149979e+00 -9.82550025e-01 -3.62918198e-01 -4.20319617e-01
-1.69567621e+00 -4.05583590e-01 1.31741345e-01 -1.03548968e+00
3.91919583e-01 5.12262583e-01 4.42477167e-01 1.01914012e+00
8.45491812e-02 -4.14655991e-02 1.06380177e+00 -1.56045526e-01
5.97015396e-02 2.84845412e-01 1.85472578e-01 3.67168188e-01
8.92998874e-01 -2.85289079e-01 -2.65995860e-01 -1.07473242e+00
4.80319917e-01 3.84887636e-01 -2.91726619e-01 -6.25890791e-02
-1.50289774e+00 5.99085748e-01 -1.37451783e-01 3.37967545e-01
-2.55347311e-01 -4.30174351e-01 4.25903887e-01 8.26812461e-02
4.59579527e-01 1.09128463e+00 -6.72909975e-01 -3.00715983e-01
-1.13612080e+00 1.63017139e-01 9.72907126e-01 8.14288497e-01
3.55714351e-01 -5.24063110e-01 -4.66440827e-01 9.08615589e-01
-6.14934415e-03 5.10004938e-01 5.89998484e-01 -8.82143140e-01
2.11206809e-01 8.03616226e-01 3.71486723e-01 -9.42816913e-01
-5.54494858e-01 -5.26318371e-01 -5.94738722e-01 -4.66437817e-01
7.99319506e-01 -2.90460765e-01 -6.46978855e-01 1.62854099e+00
2.94148892e-01 -1.32132232e-01 -1.06807597e-01 3.86881649e-01
8.35464895e-01 -7.72252530e-02 5.10759115e-01 -1.54135883e-01
1.68208742e+00 -2.02245295e-01 -8.77858162e-01 4.43238884e-01
9.25724506e-01 -8.77831519e-01 7.07265913e-01 3.92517895e-02
-8.80270779e-01 -1.06223837e-01 -4.41544294e-01 1.04292974e-01
-4.47613090e-01 -1.70952007e-01 2.42310584e-01 1.07522464e+00
-8.01346481e-01 4.93254155e-01 -6.57588065e-01 -6.29577339e-01
6.81231737e-01 2.68379867e-01 -4.56451267e-01 1.16343550e-01
-8.97025049e-01 5.34746766e-01 -1.20331571e-01 -4.79879439e-01
-1.82559848e-01 -1.38413620e+00 -4.52907354e-01 2.15069447e-02
7.54283071e-02 -1.06587601e+00 9.20536518e-01 -4.38568294e-01
-6.21541798e-01 1.27317643e+00 3.54884788e-02 -6.57986179e-02
5.64966679e-01 -4.18859236e-02 -7.57861972e-01 -6.76002949e-02
4.63783294e-01 1.01928204e-01 2.00186357e-01 -9.66059089e-01
-7.13528812e-01 -6.18619502e-01 -3.93385261e-01 -3.57224137e-01
-6.16283476e-01 3.86398137e-01 -2.29222048e-02 -8.00195694e-01
-2.58529261e-02 -9.72326756e-01 -2.41078913e-01 -1.78934664e-01
-5.33997953e-01 -9.00641531e-02 1.05617624e-02 -8.73346746e-01
1.80315685e+00 -2.40860224e+00 -4.66264397e-01 3.15386385e-01
7.44992435e-01 1.17943175e-01 2.26518273e-01 4.49212134e-01
-2.71322075e-02 8.21179688e-01 -6.56610057e-02 -1.15077384e-02
-1.11143351e-01 -2.17264608e-01 -2.23622158e-01 7.39347935e-01
-1.24359101e-01 8.03342342e-01 -9.44811344e-01 -6.68628454e-01
-3.23046952e-01 3.04236561e-01 -9.18887615e-01 1.00020759e-01
2.38974631e-01 4.06842142e-01 -4.00127172e-01 1.00341010e+00
7.05918252e-01 -7.61743546e-01 4.84522104e-01 -5.77830710e-02
-3.75085115e-01 3.97890449e-01 -8.85990977e-01 1.34088540e+00
2.26295546e-01 1.02830425e-01 1.53395563e-01 -1.65435195e-01
7.77205467e-01 4.02773142e-01 9.82548654e-01 9.14179068e-03
3.71829234e-02 5.66086709e-01 2.10771978e-01 -5.70631683e-01
4.77762848e-01 -1.86082006e-01 -4.91412021e-02 6.06331348e-01
-5.09501338e-01 3.28289896e-01 -2.27753863e-01 1.13839038e-01
1.33775568e+00 -5.48801124e-01 4.03586626e-01 -3.74239802e-01
5.43450154e-02 2.30980977e-01 9.88629758e-01 8.84168088e-01
-5.99310398e-01 9.15695727e-01 4.11557138e-01 -2.16399506e-01
-1.08587551e+00 -8.47599566e-01 -7.05314159e-01 7.94268906e-01
-3.46356601e-01 -5.75404763e-01 -6.55555904e-01 -6.60131335e-01
6.76880360e-01 2.75370359e-01 -6.83742285e-01 -5.98217826e-03
-2.90816069e-01 -8.15171719e-01 1.20161819e+00 3.52647603e-01
-9.97635573e-02 -5.45260727e-01 -4.95859921e-01 1.81437463e-01
-2.84427822e-01 -1.01770675e+00 -8.19735229e-01 -3.65702391e-01
-6.78594053e-01 -1.37685871e+00 -9.47943091e-01 -4.10749704e-01
6.72300577e-01 -4.78623621e-02 9.88492966e-01 1.32641286e-01
-5.70566773e-01 5.37211895e-01 -2.40437731e-01 -5.33610225e-01
-5.70371628e-01 5.20924270e-01 8.91948417e-02 3.84729654e-02
7.73103356e-01 -2.95046926e-01 -8.90563369e-01 3.65092605e-01
-8.33008826e-01 -5.17216206e-01 5.49876392e-01 6.32286847e-01
3.54586035e-01 -6.07419193e-01 6.96289778e-01 -1.27454221e+00
6.67199731e-01 -8.09695721e-01 -3.07429463e-01 3.55304480e-01
-1.13572717e+00 -1.16670370e-01 3.48737955e-01 -3.56132269e-01
-6.57787204e-01 -2.60716707e-01 -2.58130226e-02 -1.15395866e-01
-1.85383096e-01 5.49933136e-01 1.14761174e-01 5.41490875e-02
8.12051713e-01 -2.20828623e-01 2.99707919e-01 -7.84198999e-01
-2.24301219e-02 1.19079816e+00 2.88784653e-01 -4.80814427e-01
6.14341259e-01 5.85183322e-01 -4.99828279e-01 -4.13200170e-01
-7.23241329e-01 -8.10140610e-01 -3.95868421e-01 3.86539996e-01
8.57093096e-01 -1.04522216e+00 -7.68262327e-01 2.35584036e-01
-7.48679638e-01 5.18974364e-02 -2.85318375e-01 7.34399080e-01
8.40957314e-02 4.34074491e-01 -6.06380463e-01 -4.72873807e-01
-5.03790498e-01 -1.06631374e+00 8.38632226e-01 6.62862882e-02
-9.34006989e-01 -9.10858870e-01 5.06052911e-01 5.72531641e-01
5.62674105e-01 3.39115858e-01 1.04203641e+00 -1.16547441e+00
-2.30662465e-01 -2.05729693e-01 -3.33179414e-01 -1.49784476e-01
6.82555914e-01 1.99369729e-01 -8.17027390e-01 -3.73638541e-01
-2.22063780e-01 1.77252963e-01 4.28964734e-01 4.65831190e-01
1.15533769e+00 -4.25419062e-01 -7.29160786e-01 1.00258970e+00
1.22383630e+00 9.51286629e-02 3.52217853e-01 4.87844080e-01
7.66342938e-01 5.76390445e-01 1.85624629e-01 8.39737713e-01
5.00850260e-01 3.11776310e-01 -2.79961318e-01 -1.66585252e-01
1.79131597e-01 -1.44414306e-01 -2.56797731e-01 6.67354822e-01
-1.02102328e-02 2.27404252e-01 -1.43321681e+00 8.11608672e-01
-1.35881901e+00 -9.33712065e-01 -3.06025390e-02 2.32476306e+00
1.31820321e+00 -2.85697550e-01 3.12888026e-01 -3.75785619e-01
8.86136234e-01 -1.66889533e-01 -5.41768491e-01 -9.17382091e-02
-2.86360115e-01 2.04976480e-02 9.47740912e-01 2.75103021e-02
-5.98360300e-01 3.63800526e-01 7.63887644e+00 3.16964388e-01
-1.15067196e+00 4.78528179e-02 7.71028280e-01 -2.11548463e-01
-5.10065079e-01 -1.42046809e-01 -9.70065534e-01 6.24347329e-01
9.46262360e-01 -5.51910758e-01 -2.72821765e-02 5.64889252e-01
2.31861383e-01 3.02787542e-01 -1.26062083e+00 1.12553823e+00
-7.86398277e-02 -1.57584691e+00 -1.00299427e-02 4.42996502e-01
7.88330972e-01 -9.07063186e-02 2.06046671e-01 -1.22059025e-01
3.45198393e-01 -1.01264060e+00 2.22822472e-01 5.07520795e-01
1.30882013e+00 -2.72282928e-01 5.92733324e-01 -2.59662896e-01
-6.75827622e-01 -3.54825169e-01 -5.62305935e-02 3.01379442e-01
-5.20308912e-02 7.52924442e-01 -8.37473333e-01 5.42105973e-01
7.28480816e-01 6.67097926e-01 -5.26923954e-01 1.28150022e+00
5.03619909e-01 8.19308162e-01 -2.78952401e-02 4.27324295e-01
-1.39498636e-01 1.16120651e-02 2.83437282e-01 1.21569252e+00
3.29388171e-01 2.26870298e-01 -1.19348668e-01 9.24912870e-01
-2.68918037e-01 3.71169537e-01 -4.60977882e-01 -4.95633632e-01
8.43057275e-01 1.13625693e+00 -5.95433950e-01 -2.30588809e-01
-6.14961207e-01 2.66671687e-01 -1.09413318e-01 2.22395375e-01
-6.49125218e-01 -3.90286773e-01 1.09742641e+00 6.54138029e-01
7.02810958e-02 1.63543761e-01 -6.38752341e-01 -1.06622601e+00
-1.99530706e-01 -1.31542039e+00 1.08316863e+00 -2.09899098e-01
-1.65525985e+00 1.81600735e-01 -2.77786374e-01 -1.15176082e+00
-3.83158773e-02 -2.62148023e-01 5.23442067e-02 1.06228435e+00
-1.29023707e+00 -8.63577366e-01 -1.94764867e-01 4.30168390e-01
-3.70501131e-01 -1.36304632e-01 1.01985824e+00 7.10520685e-01
-6.05724037e-01 1.24259007e+00 5.05990326e-01 5.48436224e-01
1.57452846e+00 -9.32043433e-01 2.45581970e-01 6.46568686e-02
-3.98279309e-01 1.39421785e+00 4.94117618e-01 -1.01158214e+00
-1.13176727e+00 -8.26725245e-01 1.09947670e+00 -1.01440692e+00
6.40994370e-01 -9.12746787e-02 -9.52479661e-01 8.82716179e-01
-8.97029936e-02 -1.42978713e-01 1.64390910e+00 5.84979594e-01
-6.89346194e-01 -5.66046871e-02 -1.44608200e+00 3.77564669e-01
1.05846429e+00 -7.97814071e-01 -5.68022788e-01 1.70849264e-01
5.63994825e-01 -1.19520746e-01 -1.50164700e+00 2.66345382e-01
9.08029139e-01 -7.22560704e-01 9.28502560e-01 -9.18172479e-01
2.92743772e-01 -6.50650039e-02 1.54687822e-01 -9.16932821e-01
-5.69857895e-01 -7.69327641e-01 5.05867779e-01 1.38695097e+00
7.25846529e-01 -1.10788167e+00 7.52244234e-01 1.35487413e+00
1.42754972e-01 -6.70501232e-01 -6.99110210e-01 -5.96864700e-01
3.42771649e-01 2.04572886e-01 1.04814076e+00 1.68863881e+00
4.34540033e-01 -1.57595575e-01 -1.14453703e-01 1.71601996e-01
3.66957933e-01 2.37365499e-01 7.72139728e-01 -1.38094115e+00
-8.89511481e-02 -5.76315403e-01 -3.32255036e-01 -2.77127236e-01
-1.41059071e-01 -8.59921396e-01 -6.05250657e-01 -1.37388480e+00
7.68482089e-01 -9.82763410e-01 -4.85012233e-01 5.17477393e-01
-5.27831078e-01 -1.01212822e-01 -1.68596938e-01 7.28844404e-01
-4.29725379e-01 -2.52294719e-01 9.24309433e-01 1.52236121e-02
-3.80937755e-01 -2.11904779e-01 -1.29897106e+00 3.95710051e-01
6.84305906e-01 -7.59775519e-01 1.68770596e-01 -2.71685421e-01
3.23832512e-01 -1.26804471e-01 1.24663524e-01 -7.26576209e-01
3.21627855e-01 -3.45543236e-01 2.14083925e-01 -4.64822769e-01
-3.34371835e-01 -6.33036077e-01 5.03493845e-01 6.10332727e-01
-5.44375002e-01 3.85419399e-01 1.43306285e-01 2.91519880e-01
5.14139235e-02 1.72592431e-01 4.88062680e-01 -1.49096876e-01
2.03224376e-01 2.66792268e-01 -2.33818412e-01 5.92966378e-01
8.21014345e-01 -3.57542217e-01 -5.39236605e-01 4.54995688e-03
-5.65640032e-01 1.73204511e-01 9.14846659e-01 1.80262536e-01
4.27619480e-02 -1.01475203e+00 -9.25481796e-01 9.97031406e-02
4.54099208e-01 -3.38525951e-01 3.15787613e-01 1.10358524e+00
-4.69182909e-01 4.61429149e-01 -1.73956797e-01 -4.01289970e-01
-1.35499048e+00 6.35276020e-01 3.88935149e-01 -1.95305794e-01
-4.54378247e-01 4.98652637e-01 6.25616163e-02 -5.37631571e-01
1.32003129e-01 -2.28963658e-01 -2.95343921e-02 3.28483105e-01
5.81125259e-01 4.99879837e-01 -6.88237548e-02 -5.09703040e-01
-5.79735398e-01 3.95982534e-01 -1.86137140e-01 4.86458242e-02
1.25886345e+00 -5.77199608e-02 -3.03249478e-01 2.78922409e-01
1.08687901e+00 6.82509840e-01 -5.10913968e-01 -1.15691431e-01
1.16859831e-01 -6.44575894e-01 -5.25802433e-01 -8.14484060e-01
-8.65015924e-01 1.90260947e-01 4.94971782e-01 2.78740823e-01
6.08294010e-01 6.72177598e-03 6.96999252e-01 -1.20110095e-01
5.45931011e-02 -9.45980310e-01 -5.20644069e-01 9.48573127e-02
2.03306213e-01 -1.04985464e+00 2.13118851e-01 -2.82355100e-01
-3.43291730e-01 5.62585354e-01 3.26614946e-01 6.48317695e-01
7.97104359e-01 2.00751662e-01 4.69773084e-01 -3.48862648e-01
-3.99157226e-01 2.52688736e-01 9.77974571e-03 3.58772784e-01
7.03637600e-01 2.67019093e-01 -6.92301035e-01 9.20245588e-01
-2.37938270e-01 1.14029996e-01 5.47560990e-01 8.42708051e-01
7.49907643e-02 -1.25389183e+00 -6.56041563e-01 8.64922643e-01
-1.27069795e+00 -3.53050411e-01 -5.15410364e-01 4.98539537e-01
1.59534067e-01 7.80710578e-01 1.83920674e-02 -2.44673759e-01
2.89933711e-01 3.32697272e-01 -8.25231820e-02 -7.37065017e-01
-1.20376229e+00 -1.74236745e-01 1.61035389e-01 -2.83412874e-01
-1.75589591e-01 -9.66805398e-01 -9.85078335e-01 -5.85583866e-01
-3.08820546e-01 2.29184166e-01 4.70300496e-01 4.84067231e-01
1.12617636e+00 2.37739727e-01 3.69007260e-01 3.17613840e-01
-8.85214210e-01 -6.31642342e-01 -6.24060810e-01 9.96005774e-01
3.17334533e-01 -4.26321208e-01 -5.67270696e-01 -1.62068277e-03] | [6.992888927459717, 6.842841625213623] |
410f297a-07da-4972-a210-dac48a4144c8 | measuring-mathematical-problem-solving-with | 2103.03874 | null | https://arxiv.org/abs/2103.03874v2 | https://arxiv.org/pdf/2103.03874v2.pdf | Measuring Mathematical Problem Solving With the MATH Dataset | Many intellectual endeavors require mathematical problem solving, but this skill remains beyond the capabilities of computers. To measure this ability in machine learning models, we introduce MATH, a new dataset of 12,500 challenging competition mathematics problems. Each problem in MATH has a full step-by-step solution which can be used to teach models to generate answer derivations and explanations. To facilitate future research and increase accuracy on MATH, we also contribute a large auxiliary pretraining dataset which helps teach models the fundamentals of mathematics. Even though we are able to increase accuracy on MATH, our results show that accuracy remains relatively low, even with enormous Transformer models. Moreover, we find that simply increasing budgets and model parameter counts will be impractical for achieving strong mathematical reasoning if scaling trends continue. While scaling Transformers is automatically solving most other text-based tasks, scaling is not currently solving MATH. To have more traction on mathematical problem solving we will likely need new algorithmic advancements from the broader research community. | ['Jacob Steinhardt', 'Dawn Song', 'Eric Tang', 'Steven Basart', 'Akul Arora', 'Saurav Kadavath', 'Collin Burns', 'Dan Hendrycks'] | 2021-03-05 | null | null | null | null | ['math-word-problem-solving', 'mathematical-reasoning', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'natural-language-processing', 'reasoning', 'time-series'] | [ 1.72608525e-01 3.78128678e-01 -1.29112741e-02 -5.26763201e-01
-7.24000692e-01 -9.80351985e-01 3.02220762e-01 3.23496252e-01
-7.61569366e-02 8.69447708e-01 4.29607322e-03 -1.02824438e+00
-4.04837042e-01 -1.20199180e+00 -9.12923515e-01 1.33960560e-01
1.59283608e-01 6.98458970e-01 -4.93063256e-02 -4.26056057e-01
4.16214019e-01 3.66448939e-01 -1.14190662e+00 5.18110871e-01
1.24609113e+00 2.86474168e-01 1.56524584e-01 8.28337371e-01
-3.33318144e-01 1.32961905e+00 -6.22937143e-01 -8.45153332e-01
3.46824557e-01 -5.55451870e-01 -1.21179593e+00 -4.68529791e-01
8.84945571e-01 -3.76286894e-01 -3.93489391e-01 7.96845019e-01
4.26900340e-03 2.36578062e-01 6.84302866e-01 -1.18106246e+00
-9.05514598e-01 1.18316650e+00 -1.64479360e-01 3.80006373e-01
4.14898276e-01 2.56743282e-01 1.28141415e+00 -5.08531213e-01
4.65711594e-01 1.19066072e+00 6.33905411e-01 4.98496264e-01
-1.29012060e+00 -8.78924251e-01 1.47401363e-01 2.96342105e-01
-8.74014497e-01 3.32001895e-02 7.09485352e-01 -4.36404645e-01
7.05965638e-01 3.11107278e-01 1.04653180e+00 5.42161286e-01
-1.59414634e-01 7.14902878e-01 1.17311716e+00 -2.83386499e-01
-6.14920296e-02 8.92049260e-03 4.56720918e-01 9.07631457e-01
2.97676414e-01 -1.69147864e-01 -3.92328382e-01 2.02706188e-01
8.78688395e-01 -1.82633027e-01 -2.28141889e-01 -6.68574683e-03
-1.26956511e+00 7.87849844e-01 4.68392462e-01 3.09395641e-01
1.45327017e-01 4.71865624e-01 -2.07719356e-01 8.17636490e-01
-7.29356632e-02 1.32786489e+00 -7.11928904e-01 -4.12878603e-01
-1.16306782e+00 8.61333430e-01 1.03625846e+00 9.52349544e-01
4.53536451e-01 1.12987615e-01 -2.60752514e-02 2.87546784e-01
-2.58929104e-01 3.01363498e-01 7.42019638e-02 -1.51107943e+00
7.40790606e-01 8.67243707e-01 -7.87664279e-02 -1.02902365e+00
-3.67111921e-01 -8.70558977e-01 -4.77170944e-01 9.22450423e-02
1.12135422e+00 -1.03297599e-01 -6.06478512e-01 1.69849515e+00
-1.61467493e-01 2.37642467e-01 -2.38513246e-01 7.05429494e-01
8.33572447e-01 8.94243598e-01 1.35982215e-01 1.09661341e-01
9.83286619e-01 -8.13901603e-01 -4.21510577e-01 -1.96907401e-01
1.03137481e+00 -7.43307650e-01 1.38861132e+00 9.11192119e-01
-1.73970199e+00 -6.08588099e-01 -1.04649949e+00 -5.49901307e-01
-3.56751829e-01 -1.50948673e-01 1.11975861e+00 6.95722699e-01
-9.30485249e-01 8.73723745e-01 -5.77881932e-01 1.19808890e-01
6.02666974e-01 3.84025604e-01 -2.20500678e-02 -3.45014066e-01
-1.00018358e+00 1.12498999e+00 2.20673129e-01 -3.80562484e-01
-6.59254968e-01 -1.62454367e+00 -6.62778437e-01 5.56096375e-01
2.88939595e-01 -8.16469312e-01 1.26360095e+00 -5.17975688e-01
-1.04064727e+00 7.12593138e-01 2.86020190e-02 -5.31955123e-01
6.03590965e-01 -3.17382440e-02 2.39367500e-01 -3.19479741e-02
-2.40894005e-01 8.69287431e-01 1.42563567e-01 -9.56502259e-01
-5.26928604e-01 -1.68265715e-01 6.00873649e-01 5.89234419e-02
-5.75293660e-01 -1.83905065e-01 -1.84784457e-02 -6.67872667e-01
3.64164591e-01 -6.82324588e-01 -2.31950879e-02 -1.03170574e-01
4.50734757e-02 -4.93048489e-01 1.95632547e-01 -6.50263190e-01
1.13204491e+00 -1.50263894e+00 4.02984232e-01 1.62023574e-01
5.28962255e-01 -5.55705791e-03 -3.09389830e-01 2.68356521e-02
-1.69925436e-01 4.27300513e-01 2.21250921e-01 6.97920239e-03
3.71310830e-01 -1.78357050e-01 -6.09945774e-01 -4.26367521e-02
1.88121691e-01 1.37795603e+00 -9.14718390e-01 -3.78770381e-01
1.85992509e-01 -7.59748146e-02 -1.14316154e+00 -3.66746890e-03
-5.44107199e-01 8.15595090e-02 -2.88319290e-01 5.02313614e-01
5.00756860e-01 -4.89249676e-01 1.69046253e-01 2.22297594e-01
2.78037995e-01 7.07693219e-01 -1.17977428e+00 1.49416196e+00
-5.84289968e-01 8.14107835e-01 -2.38076687e-01 -1.27087796e+00
7.34653711e-01 -9.14387926e-02 9.22284350e-02 -6.10715687e-01
1.10980868e-01 1.99601427e-01 6.13349974e-01 -3.62366676e-01
4.64468122e-01 -2.25515276e-01 2.66488105e-01 6.30907893e-01
-1.07538000e-01 -9.31502879e-01 6.06975555e-01 4.63849187e-01
1.25970101e+00 1.61761671e-01 -4.52394992e-01 -3.81160945e-01
3.11314285e-01 4.71511483e-01 1.58788800e-01 8.16318452e-01
1.50381356e-01 1.62868366e-01 5.89015782e-01 -7.97191620e-01
-1.17804778e+00 -1.11497986e+00 -8.28229859e-02 1.33816135e+00
-2.93506712e-01 -6.66404188e-01 -7.96808660e-01 -3.65723908e-01
1.13665558e-01 1.02001786e+00 -3.05434316e-01 -5.49325272e-02
-8.25579226e-01 -5.24776578e-01 4.74539191e-01 6.43327236e-01
3.45421612e-01 -6.71512961e-01 -3.88189614e-01 9.63318646e-02
-2.32645020e-01 -1.11766624e+00 7.52591938e-02 1.39560193e-01
-1.26045382e+00 -1.05324602e+00 -5.65919936e-01 -8.50427926e-01
7.14465499e-01 1.75253570e-01 1.45981073e+00 6.88328981e-01
-3.04299563e-01 4.71766472e-01 -5.01467921e-02 -7.12032437e-01
-4.20255899e-01 3.99743021e-01 -1.09849311e-01 -1.12657845e+00
4.76371497e-01 -8.14342558e-01 -1.05135955e-01 -7.29439259e-02
-4.88596052e-01 4.12948847e-01 3.35686237e-01 6.08532667e-01
9.74713825e-04 4.11260545e-01 4.49199885e-01 -8.81448925e-01
6.70512199e-01 -2.71626532e-01 -7.54045844e-01 3.11456978e-01
-4.76635814e-01 8.96053314e-02 8.18768084e-01 -4.47842181e-01
-7.36201465e-01 -2.83177465e-01 1.62123770e-01 -1.65448353e-01
4.39452827e-02 6.91103399e-01 4.15839493e-01 -1.44494668e-01
9.22582865e-01 1.49176449e-01 -2.87406206e-01 -2.78565168e-01
3.09722900e-01 -1.05274050e-02 4.59888369e-01 -1.32371390e+00
1.42809248e+00 -2.35091031e-01 1.92669868e-01 -6.71093464e-01
-1.32165670e+00 2.96038210e-01 -3.30309838e-01 5.49372658e-02
4.33171034e-01 -9.08407927e-01 -1.05388296e+00 -9.81087089e-02
-1.10298181e+00 -8.48981082e-01 -1.36917293e-01 4.83236223e-01
-3.54597896e-01 -7.20423386e-02 -8.96490037e-01 -6.36484027e-01
2.15576917e-01 -1.13409519e+00 3.44632983e-01 3.06711376e-01
-6.60502493e-01 -1.18535733e+00 -3.66945148e-01 1.18471503e+00
5.40800512e-01 -6.47899285e-02 1.65118659e+00 -5.54547012e-01
-1.17752242e+00 4.60334979e-02 -2.07685187e-01 1.24286495e-01
-2.90141106e-01 6.65386543e-02 -7.42560685e-01 3.83579694e-02
-7.66208917e-02 -8.48236859e-01 7.29126215e-01 2.63666660e-01
1.85418284e+00 -2.28594229e-01 6.63098916e-02 6.24069154e-01
9.80461717e-01 -9.48371366e-02 5.91051698e-01 3.76264364e-01
5.62144339e-01 6.48675025e-01 3.11624914e-01 2.23330520e-02
7.40290642e-01 2.21734539e-01 -5.41191995e-02 9.35115516e-02
-9.06619206e-02 -5.02947867e-01 6.60709590e-02 7.77218103e-01
-2.64146984e-01 -1.01713911e-02 -1.34324861e+00 4.85849559e-01
-1.27528870e+00 -1.04939079e+00 -3.59676898e-01 1.82044494e+00
1.32559097e+00 5.42772710e-01 9.49965864e-02 4.33280677e-01
7.27980733e-02 -3.21460187e-01 -2.74169087e-01 -4.55517650e-01
2.04519421e-01 9.60093975e-01 -7.35646999e-03 8.36669385e-01
-6.16194427e-01 1.14856148e+00 7.01018572e+00 8.01282465e-01
-8.28615010e-01 -4.85240251e-01 8.19595754e-01 -2.07765684e-01
-7.39782095e-01 -1.25405416e-01 -7.73103774e-01 2.25238547e-01
1.05519331e+00 -4.06502813e-01 9.55371022e-01 7.50224590e-01
-4.93495772e-03 3.18511091e-02 -1.58578086e+00 9.23686981e-01
-1.77209049e-01 -1.77292264e+00 2.57292002e-01 4.05006902e-03
9.07218575e-01 -6.18282795e-01 2.99520105e-01 8.10089588e-01
6.39524400e-01 -1.88760865e+00 3.78944516e-01 3.24468106e-01
4.27562922e-01 -8.11772048e-01 4.39110637e-01 6.52379811e-01
-8.22531283e-01 -2.23002151e-01 -4.96050835e-01 -1.02485144e+00
-3.76828313e-01 3.76214355e-01 -7.77361333e-01 6.84226304e-02
3.01209629e-01 4.89924937e-01 -9.05285954e-01 8.30712199e-01
-3.94724578e-01 8.78894806e-01 -3.64326805e-01 -4.33410257e-01
1.30753621e-01 -1.23214483e-01 -7.77212009e-02 9.10995007e-01
1.58265695e-01 8.58454704e-01 1.11439534e-01 1.44594729e+00
-3.30350429e-01 -8.48225355e-02 -3.64590496e-01 -5.73423207e-01
6.07318759e-01 1.10811460e+00 -6.74491167e-01 -4.73346919e-01
-1.71835020e-01 4.71568167e-01 5.47698677e-01 1.32885873e-01
-9.21443045e-01 -3.39702517e-01 3.03497046e-01 4.03964281e-01
-2.69203037e-01 -6.34304404e-01 -1.04572797e+00 -1.26446021e+00
-7.16780573e-02 -1.23909938e+00 1.78316981e-01 -1.05980384e+00
-9.83623505e-01 -1.82670757e-01 1.13411754e-01 -5.33597410e-01
-1.70438197e-02 -1.04130650e+00 -6.70618474e-01 8.44614565e-01
-1.25202382e+00 -9.38206375e-01 -5.12476265e-01 2.81670690e-01
4.53661174e-01 -1.30521327e-01 7.16132402e-01 7.21797347e-02
-1.48921266e-01 7.67312527e-01 -3.96218002e-01 3.05364400e-01
5.46756625e-01 -1.47335470e+00 3.54817659e-01 5.09564817e-01
4.79491830e-01 1.03197360e+00 8.97825420e-01 -4.66952443e-01
-1.69694889e+00 -6.21818662e-01 1.10874450e+00 -1.05154431e+00
8.60727489e-01 -2.13088289e-01 -8.43222439e-01 1.02204406e+00
3.49427052e-02 -5.21709979e-01 7.70768225e-01 4.59215760e-01
-4.07049567e-01 1.53014474e-02 -8.54670763e-01 7.54526377e-01
1.21383858e+00 -3.64077687e-01 -1.25059092e+00 6.75145686e-01
8.04480135e-01 -8.16307127e-01 -1.03037083e+00 2.50183135e-01
3.94536495e-01 -6.18440211e-01 1.13376820e+00 -1.20863354e+00
1.37191534e+00 3.26206274e-02 -3.54011580e-02 -1.16321719e+00
-3.79016519e-01 -4.96759802e-01 -1.29654139e-01 1.03606939e+00
6.68014228e-01 -1.16632119e-01 1.37232089e+00 1.11914527e+00
1.89988632e-02 -9.15869772e-01 -2.13015646e-01 -7.30105996e-01
1.07670009e+00 -7.79835522e-01 7.03928173e-01 1.32358479e+00
2.78906196e-01 4.40999299e-01 2.77462363e-01 -1.09899670e-01
6.37793779e-01 2.67841458e-01 1.08113337e+00 -1.33745897e+00
-4.89688814e-01 -8.77927959e-01 -2.09610149e-01 -1.14116275e+00
3.18343848e-01 -1.36762536e+00 -6.29124522e-01 -1.52285433e+00
3.14627945e-01 -6.33323252e-01 2.59686038e-02 5.99671721e-01
-2.24077314e-01 1.90984905e-01 4.01798517e-01 -4.06572193e-01
-3.79613638e-01 2.17925161e-01 1.57693517e+00 -2.08339736e-01
4.39978540e-02 -2.14181855e-01 -1.00359154e+00 8.49774897e-01
7.70355642e-01 -2.01502755e-01 -7.14832008e-01 -7.61078417e-01
5.84582865e-01 1.41101226e-01 4.05926824e-01 -1.26775956e+00
3.00309509e-01 -6.26071930e-01 9.45003331e-01 -3.29529285e-01
2.32859731e-01 -5.19303799e-01 -2.54243821e-01 5.76442122e-01
-8.25015187e-01 6.15195259e-02 4.85723168e-01 -1.41864017e-01
9.92335379e-02 -2.99623787e-01 7.85492718e-01 -4.72469866e-01
-7.33583793e-02 8.73301029e-02 -9.79394168e-02 4.89396006e-01
8.32873821e-01 -3.45445983e-02 -5.18325269e-01 -5.28882325e-01
-6.47186935e-01 5.95044672e-01 2.96460807e-01 1.17897756e-01
4.59299326e-01 -1.12959552e+00 -7.95316637e-01 -2.34635442e-01
-4.54479367e-01 1.76629812e-01 7.12089539e-02 3.87092054e-01
-8.82427990e-01 7.15673566e-01 -2.55304009e-01 -2.55158693e-01
-1.13495886e+00 4.74177599e-01 3.08197647e-01 -4.89983469e-01
-3.13170761e-01 1.17499089e+00 -2.54684594e-03 -7.00934291e-01
4.20026213e-01 -7.89432764e-01 4.89296354e-02 -3.17899823e-01
7.15800107e-01 5.04618824e-01 -2.32553840e-01 4.04458851e-01
1.64163172e-01 4.72923487e-01 -8.24440569e-02 3.02595608e-02
1.66645074e+00 5.20001113e-01 -4.77760769e-02 1.60549745e-01
5.79873800e-01 3.70334238e-02 -7.56168962e-01 9.50725973e-02
-3.41499560e-02 -4.04739231e-01 -1.10956825e-01 -1.21153104e+00
-7.57425308e-01 1.33271265e+00 -6.72635436e-02 1.69011846e-01
7.96057820e-01 -3.02714914e-01 6.65190816e-01 1.06298172e+00
3.97509456e-01 -8.60375345e-01 3.60185504e-01 7.20668674e-01
6.84071183e-01 -1.14352834e+00 3.04684281e-01 -5.06431162e-01
-5.01915216e-02 1.36232436e+00 1.01717269e+00 -1.11329772e-01
3.74462962e-01 3.84500116e-01 -4.36615705e-01 -1.68602794e-01
-9.45605338e-01 2.40408748e-01 2.51356900e-01 3.30306858e-01
1.03313816e+00 -2.11885702e-02 -1.18639864e-01 8.91814530e-01
-1.28649700e+00 2.15235159e-01 8.06186020e-01 6.35172188e-01
-7.42801070e-01 -1.20109451e+00 -4.14547890e-01 6.34935677e-01
-4.23062295e-01 -4.08426493e-01 -4.71373767e-01 9.15639937e-01
3.11926566e-03 7.50666678e-01 -2.50557419e-02 -2.76670754e-01
1.33963674e-01 3.28710437e-01 1.29627478e+00 -6.84653580e-01
-5.61610460e-01 -6.69036567e-01 -4.47700918e-03 -2.25963220e-01
1.74867034e-01 -2.90483087e-01 -1.40960097e+00 -1.13247776e+00
1.15600273e-01 2.09269464e-01 3.62731516e-01 1.09795332e+00
-9.61745605e-02 7.39522934e-01 -4.35521826e-02 -4.32497978e-01
-9.09623981e-01 -8.49167407e-01 -1.61057398e-01 2.51036078e-01
-9.08707678e-02 -4.10096318e-01 -4.37559128e-01 2.01365173e-01] | [9.60161304473877, 7.330379486083984] |
b903684f-e7d7-4745-8cb9-d7357d2effe4 | grounding-dino-marrying-dino-with-grounded | 2303.05499 | null | https://arxiv.org/abs/2303.05499v4 | https://arxiv.org/pdf/2303.05499v4.pdf | Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection | In this paper, we present an open-set object detector, called Grounding DINO, by marrying Transformer-based detector DINO with grounded pre-training, which can detect arbitrary objects with human inputs such as category names or referring expressions. The key solution of open-set object detection is introducing language to a closed-set detector for open-set concept generalization. To effectively fuse language and vision modalities, we conceptually divide a closed-set detector into three phases and propose a tight fusion solution, which includes a feature enhancer, a language-guided query selection, and a cross-modality decoder for cross-modality fusion. While previous works mainly evaluate open-set object detection on novel categories, we propose to also perform evaluations on referring expression comprehension for objects specified with attributes. Grounding DINO performs remarkably well on all three settings, including benchmarks on COCO, LVIS, ODinW, and RefCOCO/+/g. Grounding DINO achieves a $52.5$ AP on the COCO detection zero-shot transfer benchmark, i.e., without any training data from COCO. It sets a new record on the ODinW zero-shot benchmark with a mean $26.1$ AP. Code will be available at \url{https://github.com/IDEA-Research/GroundingDINO}. | ['Lei Zhang', 'Jun Zhu', 'Hang Su', 'Jianwei Yang', 'Chunyuan Li', 'Jie Yang', 'Hao Zhang', 'Feng Li', 'Tianhe Ren', 'Zhaoyang Zeng', 'Shilong Liu'] | 2023-03-09 | null | null | null | null | ['referring-expression', 'zero-shot-object-detection'] | ['computer-vision', 'computer-vision'] | [ 5.08483872e-02 1.55254081e-01 -1.80975288e-01 -3.25630814e-01
-1.31448686e+00 -7.30456412e-01 3.70937437e-01 1.06129192e-01
-4.65064496e-01 2.40689889e-01 -1.19105369e-01 -1.51072070e-01
8.48997086e-02 -6.81551158e-01 -8.23336363e-01 -2.21894547e-01
1.83108777e-01 6.74052298e-01 4.04105157e-01 -4.60265905e-01
-1.74523950e-01 -6.40260987e-03 -1.61004508e+00 6.06421828e-01
6.11257613e-01 1.17465734e+00 1.13077872e-01 4.69884574e-01
-1.04319796e-01 6.67217076e-01 -4.56127614e-01 -8.69198918e-01
2.60669738e-01 -1.69503614e-01 -9.10295904e-01 1.61333695e-01
7.05833554e-01 -6.50144368e-02 -3.68741572e-01 1.26322293e+00
7.60657489e-01 1.79925561e-01 5.96600235e-01 -1.62772262e+00
-1.16244662e+00 8.42476606e-01 -4.20069575e-01 3.88013244e-01
5.05300701e-01 4.37426597e-01 1.36246300e+00 -1.63234043e+00
7.02629507e-01 1.57464802e+00 5.64006805e-01 8.27323914e-01
-1.20982599e+00 -9.90118802e-01 1.16111837e-01 2.48892918e-01
-1.84350371e+00 -4.69206452e-01 3.11290801e-01 -4.09305602e-01
9.84061241e-01 2.67878503e-01 2.93941230e-01 1.13451791e+00
-5.13831854e-01 1.22713530e+00 8.99985194e-01 -5.52724183e-01
-1.73504949e-02 2.23601639e-01 4.87401247e-01 8.96696508e-01
2.42496312e-01 3.16701923e-03 -7.02087104e-01 -1.06415048e-01
3.20117950e-01 3.99140269e-02 -3.86796176e-01 -2.48976409e-01
-1.22768629e+00 8.64620209e-01 6.28503680e-01 1.79382235e-01
3.02345175e-02 3.36434484e-01 4.95255888e-01 4.45982069e-01
3.34969789e-01 4.51125979e-01 -4.31290597e-01 2.12095037e-01
-5.96589684e-01 2.55159140e-01 5.87432146e-01 1.64697933e+00
6.62008345e-01 -3.81868631e-01 -7.02684999e-01 9.30771589e-01
4.27023530e-01 7.02113152e-01 3.64247441e-01 -9.53104258e-01
5.63504457e-01 6.06590092e-01 -1.24379478e-01 -2.72086352e-01
-1.97467595e-01 -4.04894352e-01 -4.26124722e-01 -3.01200360e-01
3.96227419e-01 2.38997899e-02 -1.16442204e+00 1.99891531e+00
2.40123168e-01 -1.39827251e-01 1.09935142e-01 7.93031037e-01
1.34586501e+00 4.93824929e-01 2.42999196e-01 2.76284695e-01
1.90822482e+00 -1.13320827e+00 -4.67807770e-01 -1.89299121e-01
1.09788239e+00 -4.79790241e-01 1.45078897e+00 2.38268495e-01
-9.50230300e-01 -5.76257467e-01 -8.19268942e-01 -4.38763887e-01
-7.04669297e-01 -2.81677656e-02 5.71530342e-01 4.94493693e-01
-9.71756876e-01 -8.71688649e-02 -4.27606016e-01 -7.81943321e-01
8.35256636e-01 3.05166870e-01 -2.05243558e-01 -4.06717211e-01
-1.20424211e+00 6.87083423e-01 5.59755385e-01 -3.09725493e-01
-1.09807134e+00 -7.86320210e-01 -9.40419912e-01 1.31462693e-01
8.12311590e-01 -9.24741805e-01 1.49580526e+00 -7.07716644e-01
-7.02333689e-01 1.50716043e+00 -1.60122305e-01 -5.34662843e-01
2.75239885e-01 -2.18335986e-01 -5.36039948e-01 9.19425711e-02
5.34677744e-01 1.30079818e+00 6.62291348e-01 -1.15600097e+00
-8.20270061e-01 -4.76651222e-01 3.69888157e-01 8.95583108e-02
-2.99360186e-01 2.33137295e-01 -8.13277364e-01 -5.98249793e-01
1.58084482e-01 -9.24544990e-01 2.15824112e-01 1.05410196e-01
-6.52860820e-01 -9.57750201e-01 5.71161807e-01 -1.62771717e-01
1.03241634e+00 -2.46273470e+00 -1.99695051e-01 -1.31139904e-01
6.27847672e-01 1.78298295e-01 -2.03440472e-01 1.70223579e-01
-6.53477013e-02 1.16467401e-01 -1.80310369e-01 -4.30659294e-01
3.73531580e-01 1.01135634e-01 -5.67377806e-01 2.07770973e-01
2.19308689e-01 1.29615211e+00 -9.77591753e-01 -5.83197534e-01
-1.62209094e-01 1.31666258e-01 -7.21239328e-01 7.23969862e-02
-4.55766559e-01 2.43812129e-02 -2.01342627e-01 1.12886047e+00
5.54629922e-01 -5.13391137e-01 -2.21231461e-01 -1.45803303e-01
2.03303739e-01 8.25292431e-03 -1.12003744e+00 1.89708602e+00
-2.55773723e-01 5.82737088e-01 -2.00462610e-01 -7.65467525e-01
8.74365628e-01 2.42015526e-01 1.53228134e-01 -8.23797882e-01
1.78384006e-01 4.06129181e-01 -8.01906884e-02 -4.96087134e-01
5.26719630e-01 -1.39598474e-01 -4.56410795e-01 -5.03760614e-02
7.33524323e-01 -5.64777888e-02 3.80724430e-01 6.58044040e-01
1.07956672e+00 -8.69468786e-03 2.54073203e-01 -3.21864724e-01
2.37044305e-01 3.84471305e-02 5.23067474e-01 1.12450719e+00
-2.91305184e-01 6.74070418e-01 3.41495603e-01 -3.78888622e-02
-5.98024905e-01 -1.37768018e+00 -3.88971448e-01 1.78661656e+00
3.07270974e-01 -5.85029125e-01 -5.03059804e-01 -7.19717085e-01
1.35914758e-01 8.40187728e-01 -6.47657573e-01 -3.25999528e-01
-1.63178861e-01 -4.79151070e-01 8.75529349e-01 7.64158726e-01
4.66513813e-01 -8.53040457e-01 -4.43788826e-01 -1.49853140e-01
-3.15459609e-01 -1.11313653e+00 -8.81714582e-01 3.73028070e-01
-4.25739884e-01 -9.91392434e-01 -7.98794389e-01 -1.00963593e+00
3.59873742e-01 3.72337431e-01 1.47702646e+00 -5.30796349e-02
-3.01822841e-01 7.48617709e-01 -4.40575093e-01 -7.56803870e-01
-1.11386329e-01 -4.98167588e-04 -1.67281684e-02 3.55651625e-03
8.20184350e-01 -1.09878257e-01 -3.88269842e-01 5.14281273e-01
-7.61211276e-01 -1.53600216e-01 3.03832620e-01 9.07838941e-01
6.15337193e-01 -6.37093365e-01 6.92716122e-01 -8.64666283e-01
2.72963822e-01 -7.27604747e-01 -7.03367770e-01 4.87500727e-01
-4.90454882e-01 -2.32520737e-02 1.53701425e-01 -4.20459241e-01
-7.81948745e-01 -1.26989782e-01 -2.56338213e-02 -7.72632122e-01
-7.03939423e-02 1.02125242e-01 -5.29528975e-01 2.72059262e-01
9.53403890e-01 1.14404753e-01 -5.10401368e-01 -5.11999309e-01
8.07150245e-01 8.42944682e-01 8.06179464e-01 -7.35774815e-01
6.71420932e-01 5.83721578e-01 -5.61533868e-01 -6.72634661e-01
-1.20453882e+00 -9.95833099e-01 -3.78665626e-01 1.34762540e-01
8.12161803e-01 -1.25463450e+00 -8.34530890e-01 2.42381200e-01
-1.21469998e+00 -2.34293088e-01 -7.13289857e-01 1.96990103e-01
-4.69907224e-01 6.65936694e-02 -4.95545805e-01 -5.43436468e-01
-4.82868344e-01 -1.00696385e+00 1.41357505e+00 2.28549346e-01
-1.03872195e-01 -7.21628129e-01 -2.67410398e-01 5.32986760e-01
4.45447639e-02 -1.43119395e-01 6.70986652e-01 -1.06435108e+00
-6.81807578e-01 -1.35971531e-01 -4.95898634e-01 3.18298072e-01
-3.97238791e-01 -5.28213799e-01 -1.23060715e+00 -4.12100673e-01
-2.45794058e-01 -9.17113364e-01 1.31328905e+00 -1.61689725e-02
1.15447640e+00 7.29067326e-02 -5.51874876e-01 8.30884635e-01
1.36423898e+00 1.21896982e-01 3.51813644e-01 3.30360457e-02
6.65948212e-01 2.52067298e-01 6.97070301e-01 4.27659273e-01
6.16789877e-01 7.30331123e-01 3.38780642e-01 2.92460732e-02
-5.60484231e-01 -3.69723707e-01 3.14072102e-01 3.15507263e-01
2.61794567e-01 -5.86398005e-01 -1.12499416e+00 8.46492171e-01
-1.68298769e+00 -7.25943208e-01 -1.25980690e-01 2.00679660e+00
7.54126310e-01 2.11574420e-01 1.80510566e-01 -2.16779470e-01
8.15388083e-01 -1.99360162e-01 -6.28077507e-01 -5.92031516e-02
-4.56572354e-01 2.12171465e-01 3.69937003e-01 2.93443233e-01
-1.10416567e+00 1.08750033e+00 5.21753597e+00 1.09735847e+00
-6.79640532e-01 7.00293183e-01 3.88807625e-01 -3.56248379e-01
-4.04777646e-01 9.97146368e-02 -1.30791140e+00 2.33948588e-01
5.60495615e-01 -1.78124428e-01 1.85971081e-01 8.89761508e-01
-5.28810740e-01 -2.82471180e-02 -1.47160220e+00 1.29371750e+00
3.81423295e-01 -1.23517907e+00 2.38826941e-03 -8.66920725e-02
6.07826471e-01 4.41574246e-01 2.24146977e-01 1.00896609e+00
4.26743865e-01 -8.63052428e-01 9.87805128e-01 3.52662444e-01
9.39604104e-01 -2.96952575e-01 4.86482501e-01 1.55299634e-01
-1.10497832e+00 -1.69636384e-01 -2.39271745e-01 2.85293370e-01
-1.43028513e-01 2.52714992e-01 -6.84602618e-01 3.76148969e-01
9.09661472e-01 4.89952534e-01 -7.78854191e-01 1.00542545e+00
-2.51647145e-01 7.32424140e-01 -5.38331628e-01 5.95050678e-02
2.17420444e-01 4.49986428e-01 6.80593312e-01 1.35684383e+00
1.17737465e-01 4.86705564e-02 4.16726857e-01 1.16396523e+00
-6.57436430e-01 7.36959651e-03 -5.51164627e-01 2.48119578e-01
5.95726907e-01 9.40464973e-01 -4.78545934e-01 -6.71021700e-01
-7.35549569e-01 9.82048631e-01 4.20707017e-01 3.65973055e-01
-1.11336553e+00 -4.84958470e-01 4.15128291e-01 1.23983420e-01
4.62725222e-01 1.54605553e-01 4.18268181e-02 -1.40000761e+00
9.01698023e-02 -7.60698080e-01 1.07726705e+00 -9.27688539e-01
-1.37300801e+00 3.89026493e-01 3.62906635e-01 -1.12911880e+00
1.42466962e-01 -7.10401893e-01 -2.81695664e-01 7.18922257e-01
-1.24997127e+00 -1.33892214e+00 -2.63897538e-01 7.63434529e-01
7.13396311e-01 -1.31531268e-01 6.32487357e-01 5.47091603e-01
-6.28227472e-01 9.53205287e-01 -1.27231866e-01 4.37377989e-01
7.19277740e-01 -1.29440868e+00 3.79793167e-01 8.13791811e-01
6.17099881e-01 6.65867388e-01 4.77289975e-01 -3.10940623e-01
-1.27960789e+00 -1.26765120e+00 8.20684552e-01 -9.40090001e-01
6.61332726e-01 -7.31226265e-01 -1.00176919e+00 1.13037097e+00
1.84844792e-01 6.60464764e-01 6.28248632e-01 3.12933415e-01
-9.85200405e-01 -2.53207814e-02 -1.02362287e+00 4.56886917e-01
1.45876646e+00 -7.70317554e-01 -9.13717985e-01 6.39326870e-01
1.25179291e+00 -4.58103448e-01 -8.29455376e-01 3.81293654e-01
1.14122711e-01 -5.15108645e-01 9.43534255e-01 -8.73175681e-01
1.47034347e-01 -2.74620593e-01 -6.72574341e-01 -9.26361978e-01
-3.63796622e-01 -6.61550939e-01 -2.35674366e-01 1.50334978e+00
4.99875754e-01 -5.91077805e-01 4.03809935e-01 3.79382044e-01
-3.55200350e-01 -7.09538460e-01 -9.84283149e-01 -1.22107553e+00
5.61923310e-02 -8.36244285e-01 3.68852824e-01 9.35965896e-01
-2.22630869e-03 8.74190390e-01 9.39937308e-02 1.47699580e-01
7.60883689e-01 1.82866827e-01 6.43902898e-01 -8.49020123e-01
-4.07754302e-01 -3.99487466e-01 -4.63202208e-01 -1.32719433e+00
9.41672400e-02 -1.51116681e+00 -4.76435348e-02 -1.27144706e+00
4.91036206e-01 -3.69091958e-01 -4.37597692e-01 8.26599538e-01
-1.67451352e-01 5.40944755e-01 5.57287872e-01 2.39963412e-01
-1.27315605e+00 6.57513618e-01 9.86127853e-01 -4.90499377e-01
-6.42343536e-02 -2.06920296e-01 -9.58213210e-01 7.60218203e-01
5.78630567e-01 -5.99004686e-01 -3.01821947e-01 -6.24402165e-01
2.16388270e-01 -1.58748329e-01 8.13940525e-01 -9.31253791e-01
2.21725330e-01 2.99442559e-01 -1.86140507e-01 -6.44082606e-01
4.13181514e-01 -4.32577044e-01 -5.01328409e-01 3.17567378e-01
-4.33010787e-01 -1.60711914e-01 2.52778888e-01 5.56976855e-01
-1.69916362e-01 -2.76764035e-01 9.17198122e-01 -1.73305854e-01
-1.34707963e+00 2.57662207e-01 2.48273294e-02 7.94232786e-01
9.25709367e-01 -7.21626803e-02 -6.47265136e-01 -6.06662445e-02
-9.65093076e-01 6.25735760e-01 1.42026037e-01 6.44153893e-01
6.80559516e-01 -1.38957024e+00 -8.56771886e-01 -3.20117027e-02
1.03154063e+00 6.30848035e-02 1.43696204e-01 9.07568932e-01
3.29704434e-02 5.52426159e-01 2.41382390e-01 -8.92379999e-01
-1.11348760e+00 9.19072807e-01 4.53851014e-01 1.12140261e-01
-5.19017398e-01 1.51608992e+00 6.72437429e-01 -5.35564899e-01
3.19852144e-01 -4.65439320e-01 1.50020406e-01 7.26124942e-02
4.75951999e-01 2.22392380e-01 8.91580619e-03 -6.05009019e-01
-5.69204628e-01 3.04191023e-01 -2.47315057e-02 -3.50276381e-02
7.65608907e-01 -1.50556490e-01 -4.58710119e-02 6.09736502e-01
1.39785993e+00 -3.01857024e-01 -8.08104873e-01 -7.60890186e-01
1.95396990e-01 -2.13613108e-01 -1.81984648e-01 -8.45795393e-01
-7.71396399e-01 8.00317645e-01 9.47966337e-01 -5.51793836e-02
1.01558506e+00 9.61389303e-01 7.59845376e-01 7.18878567e-01
4.38687861e-01 -8.68886173e-01 2.78668195e-01 7.18112171e-01
1.02426481e+00 -1.50176680e+00 -4.22971338e-01 -5.68089306e-01
-5.81850290e-01 5.59387445e-01 9.28690851e-01 1.58125982e-01
5.97076178e-01 9.06953495e-03 -1.96985126e-01 -5.30640423e-01
-9.55811620e-01 -8.44801605e-01 4.06856775e-01 5.87917507e-01
3.62239361e-01 2.14162782e-01 5.41497543e-02 8.88360322e-01
1.23047875e-02 6.89303130e-02 1.89614177e-01 6.56579256e-01
-5.60222864e-01 -6.59461439e-01 -5.03012180e-01 5.08093297e-01
-2.01132074e-01 -5.20828843e-01 -1.18330106e-01 8.35907042e-01
5.01894712e-01 8.90121162e-01 9.47351679e-02 -3.05772752e-01
7.15860367e-01 2.89317131e-01 4.83943403e-01 -1.06590080e+00
-3.89583468e-01 -2.21221432e-01 1.60529524e-01 -6.05015457e-01
-1.04670666e-01 -7.21292615e-01 -1.31457019e+00 1.35048926e-01
-4.43233281e-01 -1.36113465e-02 1.19443193e-01 8.58296633e-01
4.12259042e-01 4.02555972e-01 7.23535791e-02 -1.86076283e-01
-6.84722841e-01 -1.02286196e+00 -4.72964138e-01 6.87828422e-01
2.08022967e-01 -8.35376918e-01 -1.46213919e-01 -4.09856923e-02] | [9.951682090759277, 1.6696674823760986] |
c8039abe-bd5b-446d-a387-804947e3074e | inductive-attention-for-video-action | 2212.08830 | null | https://arxiv.org/abs/2212.08830v2 | https://arxiv.org/pdf/2212.08830v2.pdf | Inductive Attention for Video Action Anticipation | Anticipating future actions based on spatiotemporal observations is essential in video understanding and predictive computer vision. Moreover, a model capable of anticipating the future has important applications, it can benefit precautionary systems to react before an event occurs. However, unlike in the action recognition task, future information is inaccessible at observation time -- a model cannot directly map the video frames to the target action to solve the anticipation task. Instead, the temporal inference is required to associate the relevant evidence with possible future actions. Consequently, existing solutions based on the action recognition models are only suboptimal. Recently, researchers proposed extending the observation window to capture longer pre-action profiles from past moments and leveraging attention to retrieve the subtle evidence to improve the anticipation predictions. However, existing attention designs typically use frame inputs as the query which is suboptimal, as a video frame only weakly connects to the future action. To this end, we propose an inductive attention model, dubbed IAM, which leverages the current prediction priors as the query to infer future action and can efficiently process the long video content. Furthermore, our method considers the uncertainty of the future via the many-to-many association in the attention design. As a result, IAM consistently outperforms the state-of-the-art anticipation models on multiple large-scale egocentric video datasets while using significantly fewer model parameters. | ['Oswald Lanz', 'Simon See', 'Cheng-Kuang Lee', 'Giuseppe Fiameni', 'Tsung-Ming Tai'] | 2022-12-17 | null | null | null | null | ['action-anticipation', 'video-understanding'] | ['computer-vision', 'computer-vision'] | [ 2.94410855e-01 5.74651966e-03 -6.43231153e-01 -5.18973708e-01
-4.55922306e-01 -1.23627596e-01 6.32957041e-01 -2.76912570e-01
-2.70083129e-01 5.69150507e-01 8.39971364e-01 -2.07406823e-02
-1.53805122e-01 -4.98201013e-01 -8.63692164e-01 -4.80707109e-01
-1.31998256e-01 3.00181378e-03 3.81857514e-01 1.57445833e-01
2.78834671e-01 1.29896298e-01 -1.60000181e+00 5.08565545e-01
5.94764888e-01 1.19862688e+00 6.49637580e-01 6.44463837e-01
1.98186725e-01 1.22548783e+00 -1.90652590e-02 -1.93530336e-01
8.89657289e-02 -3.53533536e-01 -6.46818519e-01 1.95533708e-01
2.80335873e-01 -9.07056093e-01 -8.56279075e-01 7.98066258e-01
7.59952366e-02 3.92346531e-01 3.10343176e-01 -1.30326772e+00
-8.23621392e-01 4.77472216e-01 -3.98238569e-01 5.66336274e-01
7.17990458e-01 4.81528103e-01 9.45146799e-01 -7.42803872e-01
6.60092235e-01 1.24140644e+00 2.06260771e-01 5.00477493e-01
-5.82726002e-01 -5.14094412e-01 9.12606537e-01 1.15352964e+00
-1.11924469e+00 -5.71777165e-01 8.43766868e-01 -3.67304236e-01
1.10779023e+00 1.45367652e-01 9.20413733e-01 1.40583992e+00
1.90561548e-01 1.28785443e+00 4.67927456e-01 -3.35264392e-02
8.54369551e-02 -5.40514588e-01 -1.49247035e-01 4.39635783e-01
-2.95616955e-01 1.62605718e-01 -7.73633063e-01 3.09402972e-01
7.21407235e-01 5.11953592e-01 -3.47531825e-01 6.85541406e-02
-1.55737615e+00 4.91030723e-01 2.19144955e-01 -7.41664618e-02
-1.07271338e+00 4.58014518e-01 3.07921082e-01 -4.92810942e-02
4.91259277e-01 1.08370461e-01 -4.39747959e-01 -6.58456087e-01
-7.75348902e-01 1.18191279e-01 3.04555833e-01 9.87708211e-01
6.47189081e-01 -5.77398837e-02 -5.70700943e-01 9.60433930e-02
4.55848902e-01 3.64823401e-01 3.64287764e-01 -1.38371587e+00
5.58861315e-01 4.24272597e-01 3.93128276e-01 -1.05535960e+00
2.58811451e-02 -1.16251543e-01 -5.28401315e-01 -3.06453496e-01
1.66656554e-01 -3.00925411e-02 -7.44414926e-01 1.76091683e+00
3.16008359e-01 1.11786354e+00 -3.59924398e-02 1.14539087e+00
3.59642565e-01 7.84692705e-01 2.98293054e-01 -5.78469336e-01
1.18164074e+00 -1.02499747e+00 -9.47161198e-01 -5.44920981e-01
4.69292760e-01 -6.40181661e-01 9.18756008e-01 2.71491259e-01
-9.93307650e-01 -7.15440989e-01 -7.44574428e-01 -1.04555503e-01
1.47993162e-01 7.23227635e-02 7.73008168e-01 -1.42459542e-01
-7.89158463e-01 4.33380961e-01 -1.12366104e+00 -3.83296579e-01
4.73515749e-01 7.66036287e-02 -2.53366709e-01 -4.05115604e-01
-1.26693821e+00 7.26134598e-01 5.41548491e-01 2.99590588e-01
-1.16137278e+00 -5.85596740e-01 -8.46722066e-01 3.23381275e-01
7.75236845e-01 -6.97968721e-01 1.23355675e+00 -1.11735535e+00
-1.39661050e+00 2.34018356e-01 -7.06291974e-01 -8.51153016e-01
4.43590462e-01 -7.28206158e-01 -5.40919602e-01 4.06570226e-01
-4.29187082e-02 8.09422731e-01 9.15386379e-01 -6.62885070e-01
-7.40763247e-01 -5.32094426e-02 4.93362486e-01 3.63085926e-01
-2.20457047e-01 -3.38326627e-03 -8.94715071e-01 -6.20365560e-01
1.73880532e-01 -8.23039591e-01 -1.70379132e-01 7.54432976e-02
-6.95221126e-02 -2.86078990e-01 1.06552708e+00 -6.28852844e-01
1.48522139e+00 -2.16891909e+00 1.28812507e-01 -2.86947012e-01
-2.83944961e-02 1.21489123e-01 -1.40760422e-01 3.54173154e-01
-8.55087042e-02 -1.22929439e-01 2.92669058e-01 -2.31000394e-01
-2.48458628e-02 4.60646152e-01 -7.96076119e-01 3.75438958e-01
1.68037772e-01 1.06718898e+00 -1.19881797e+00 -6.21667445e-01
4.03027326e-01 4.54224050e-01 -7.32856154e-01 5.22384822e-01
-5.85159242e-01 5.59462190e-01 -8.61231148e-01 6.30040348e-01
2.04143539e-01 -4.55937892e-01 7.29820598e-03 -3.68798286e-01
-2.58229095e-02 1.35074526e-01 -7.43685246e-01 1.92065358e+00
-2.03999609e-01 7.13838875e-01 -4.08010542e-01 -9.91768658e-01
5.30688167e-01 5.46062291e-01 7.87905633e-01 -6.14646137e-01
-1.98579520e-01 -3.33532870e-01 -1.04996115e-01 -1.11259568e+00
5.16211033e-01 1.93970665e-01 1.77221313e-01 3.90433073e-01
-2.06310734e-01 6.04882598e-01 1.29670903e-01 3.41527224e-01
1.06686437e+00 6.74333215e-01 2.75477648e-01 4.02570069e-01
4.64713752e-01 -2.58858413e-01 1.20034504e+00 7.05100894e-01
-5.52426100e-01 6.29771411e-01 5.35949707e-01 -6.50738657e-01
-6.85470521e-01 -9.11484778e-01 4.20879751e-01 1.13335228e+00
3.91943187e-01 -5.67204118e-01 -4.23640847e-01 -6.55103445e-01
-3.24558645e-01 9.53411281e-01 -6.18744433e-01 -3.71126771e-01
-6.80641353e-01 -2.87304580e-01 -1.25787199e-01 8.57250333e-01
6.13456905e-01 -1.22448552e+00 -8.13623011e-01 2.60675102e-01
-6.98549509e-01 -1.29085767e+00 -7.88560092e-01 -7.44651198e-01
-6.86461985e-01 -1.08880818e+00 -4.72113252e-01 -1.49174914e-01
4.44078714e-01 3.67754638e-01 1.04139423e+00 1.20950930e-01
2.25475401e-01 7.51446187e-01 -5.76245010e-01 -2.99981922e-01
1.99756026e-01 -4.75941896e-01 2.12981626e-01 4.48436826e-01
7.19008327e-01 -6.00645840e-01 -1.09208488e+00 2.65011311e-01
-6.79726422e-01 4.94584084e-01 6.61815822e-01 6.34636641e-01
6.77467287e-01 -1.93362996e-01 5.21964312e-01 -4.15113807e-01
-4.64942269e-02 -8.26408386e-01 -2.18736500e-01 4.04367238e-01
-4.07083333e-01 -1.05427340e-01 4.86065000e-01 -6.49829388e-01
-1.40626752e+00 1.01838976e-01 8.05716813e-02 -9.48249936e-01
-1.41089812e-01 5.73560536e-01 -3.30307215e-01 5.32722950e-01
-9.75755081e-02 3.40710819e-01 -3.37760717e-01 -2.46831104e-01
2.73950040e-01 2.03195333e-01 5.20741403e-01 -4.45385605e-01
1.03910565e-01 6.54478312e-01 -8.78308266e-02 -4.61614221e-01
-1.19311619e+00 -6.03312314e-01 -4.95191604e-01 -6.54963493e-01
9.92273629e-01 -1.01839161e+00 -8.71290267e-01 1.87689051e-01
-1.41001284e+00 -2.10312396e-01 -8.97162482e-02 9.89857316e-01
-8.87029409e-01 5.83148658e-01 -4.30915475e-01 -9.64009881e-01
4.94779274e-02 -1.19666636e+00 9.89002585e-01 2.39329249e-01
-3.18525165e-01 -8.09844851e-01 -3.27243209e-01 3.25651944e-01
1.00749515e-01 4.72387411e-02 4.34020489e-01 -3.25226039e-01
-1.26805818e+00 -2.01559678e-01 -2.27056906e-01 -1.48284331e-01
-3.75813656e-02 7.58599192e-02 -8.46343338e-01 5.10877417e-03
5.18942624e-02 -2.09863316e-02 9.93040442e-01 6.96846902e-01
1.57318187e+00 -5.27329445e-01 -3.87023419e-01 5.77208281e-01
9.75592792e-01 3.39977235e-01 9.49731469e-01 2.04730794e-01
6.76178336e-01 5.31872749e-01 1.23600698e+00 7.30303705e-01
6.78633451e-01 6.23607159e-01 7.08746552e-01 3.90796721e-01
-4.01471145e-02 -5.74124396e-01 6.00119293e-01 6.08357072e-01
-3.45870316e-01 -5.53175688e-01 -6.98539734e-01 6.47987306e-01
-2.47301316e+00 -1.52006269e+00 1.94505647e-01 2.03053403e+00
3.66853476e-01 9.27370489e-02 -2.64540046e-01 -3.55581343e-01
5.55493414e-01 5.49902856e-01 -9.42480028e-01 2.02103466e-01
4.26151082e-02 -5.86800098e-01 1.81136444e-01 4.06016171e-01
-1.09290159e+00 9.68321204e-01 6.04021692e+00 6.33480549e-01
-1.00669062e+00 7.93247223e-02 6.35749400e-01 -3.89736354e-01
-2.38943145e-01 3.32033247e-01 -7.23185718e-01 7.49087989e-01
8.54065597e-01 -2.70916849e-01 3.04016024e-01 8.26254725e-01
7.97464728e-01 -3.53116691e-01 -1.26438987e+00 1.04157031e+00
1.26630977e-01 -1.36990511e+00 3.16674531e-01 -9.27203149e-02
5.44147789e-01 -7.77048618e-02 1.07754968e-01 3.73856336e-01
-1.87901072e-02 -8.02220225e-01 7.10924566e-01 1.36415398e+00
5.10017395e-01 -4.54996496e-01 4.61707324e-01 5.15852273e-01
-1.35293186e+00 -3.75951737e-01 -2.65411228e-01 -4.74826396e-01
7.67553151e-01 2.81372130e-01 -6.27942920e-01 4.10843849e-01
6.39052570e-01 1.42026377e+00 -3.46330464e-01 9.10595715e-01
-3.57662767e-01 7.60789990e-01 -1.13518052e-01 2.47599319e-01
2.90835440e-01 -1.92431942e-01 6.36433780e-01 7.49406934e-01
6.34671509e-01 5.86275220e-01 3.86264712e-01 6.12656593e-01
2.26283848e-01 -3.84840578e-01 -5.35176575e-01 -7.43620023e-02
4.16330278e-01 7.76059270e-01 -4.14787501e-01 -6.27153337e-01
-6.99526489e-01 1.01349020e+00 2.27196977e-01 8.08187485e-01
-1.17127621e+00 2.33492762e-01 9.08003449e-01 -2.35251747e-02
4.81702000e-01 -2.88919896e-01 8.28805119e-02 -1.36985695e+00
1.88556328e-01 -5.68567574e-01 5.70808411e-01 -1.27764976e+00
-9.58662331e-01 2.79905409e-01 6.74898848e-02 -1.55212915e+00
-6.25102997e-01 -2.97922522e-01 -8.33792865e-01 6.37903690e-01
-1.48563182e+00 -1.23853362e+00 -2.93513030e-01 5.18023372e-01
9.93233740e-01 1.71212539e-01 4.44668531e-01 6.99664131e-02
-5.71666479e-01 -6.07521739e-04 -3.23987097e-01 -8.81109908e-02
7.60636866e-01 -7.36614287e-01 2.28616819e-01 1.20261014e+00
1.50240794e-01 5.17108381e-01 8.19295228e-01 -8.65552783e-01
-1.48695672e+00 -1.16409183e+00 9.06204939e-01 -4.54786986e-01
7.16308355e-01 1.57877594e-01 -1.09781086e+00 1.12947667e+00
6.79123551e-02 2.16986388e-01 5.44298768e-01 1.29202064e-02
-1.04636382e-02 -7.98127055e-02 -4.87025827e-01 7.98307538e-01
1.36129820e+00 -6.52928293e-01 -7.00390875e-01 2.68486679e-01
9.22049701e-01 -3.01211983e-01 -7.23094344e-01 3.60664576e-01
7.50645101e-01 -1.01533282e+00 1.12769508e+00 -8.20840359e-01
6.99198246e-01 -3.55829388e-01 -1.67679101e-01 -8.36864114e-01
-4.36462015e-01 -7.03399003e-01 -9.78932202e-01 1.03836727e+00
3.77103724e-02 -2.38997281e-01 8.18762183e-01 1.13795722e+00
-2.93505728e-01 -8.80063117e-01 -8.57703447e-01 -6.08565569e-01
-7.32852697e-01 -8.33176792e-01 7.49078691e-01 6.39581859e-01
2.80589256e-02 -1.82662793e-02 -9.17015433e-01 4.01793838e-01
3.31475317e-01 3.83049041e-01 6.88247144e-01 -6.99348927e-01
-3.93695265e-01 -1.43880412e-01 -4.56112891e-01 -1.78832257e+00
3.83510798e-01 -2.82484174e-01 1.76321223e-01 -1.40896583e+00
3.12842309e-01 2.04060487e-02 -5.80876708e-01 3.12009990e-01
-5.09169459e-01 -2.11571917e-01 3.21219534e-01 3.75169843e-01
-1.17886865e+00 8.36900949e-01 1.41981173e+00 -5.99994324e-02
8.65463540e-02 9.05171223e-03 -3.38593066e-01 9.55968380e-01
5.11988997e-01 -1.84193507e-01 -7.50965476e-01 -5.92402577e-01
2.78831840e-01 4.71861362e-01 5.66775441e-01 -9.13677633e-01
5.38865745e-01 -7.14502275e-01 3.38742942e-01 -9.27872121e-01
7.92963207e-01 -7.85840213e-01 2.21471235e-01 9.10371989e-02
-4.33102638e-01 9.54447985e-02 -2.18509778e-01 1.32584822e+00
-2.71643996e-01 6.41963631e-02 9.66037139e-02 -8.04917738e-02
-1.53769505e+00 8.97850513e-01 -3.13934118e-01 -1.26969993e-01
1.25304818e+00 -3.47510785e-01 -1.79989740e-01 -6.75644636e-01
-7.20597208e-01 5.61600447e-01 4.13611770e-01 5.73514104e-01
8.42830420e-01 -1.30041218e+00 -4.39572215e-01 1.54842911e-02
9.04214606e-02 -3.65907960e-02 7.68361568e-01 1.15546703e+00
4.79095645e-04 4.37765449e-01 -8.29297602e-02 -6.38795435e-01
-1.18174982e+00 9.99181151e-01 -8.63497481e-02 -3.57678980e-02
-7.51012743e-01 6.81639075e-01 6.34908020e-01 5.25419712e-01
2.93639600e-01 -3.43245178e-01 -5.04721999e-01 -9.72868502e-02
8.96224499e-01 2.78149933e-01 -6.62680805e-01 -7.14505076e-01
-2.19577938e-01 3.88223648e-01 -3.95937152e-02 -2.86945160e-02
1.21906674e+00 -6.46650374e-01 2.05942601e-01 5.06824493e-01
9.45239902e-01 -4.80469853e-01 -2.21197224e+00 -3.93385738e-01
-5.86190522e-02 -9.17929471e-01 1.33042768e-01 -5.05099356e-01
-9.74944413e-01 9.33139861e-01 1.16117001e-01 -2.18866542e-01
1.37989569e+00 -4.00767364e-02 1.02790070e+00 3.72075856e-01
3.30232143e-01 -1.11677539e+00 1.90906256e-01 5.74718356e-01
9.15312588e-01 -1.45817900e+00 1.18991837e-01 -3.41650337e-01
-9.19921875e-01 1.07145369e+00 8.82021070e-01 1.76542670e-01
6.15773082e-01 -1.96406052e-01 -2.77025461e-01 -7.12049454e-02
-1.37264049e+00 -2.19355628e-01 4.93693739e-01 3.98383409e-01
1.84745789e-01 -7.66839609e-02 -4.03676964e-02 6.64724231e-01
4.17929143e-01 3.47052842e-01 2.94681072e-01 6.35803521e-01
-2.28946611e-01 -6.32650137e-01 -1.41952351e-01 4.61905032e-01
-3.38933647e-01 -2.24284288e-02 6.32084534e-02 3.67506802e-01
9.44886357e-02 9.31588352e-01 2.98178077e-01 -3.75015974e-01
1.49509152e-02 -4.69959155e-02 1.46871433e-01 -3.19757551e-01
1.62376687e-01 3.84585261e-02 8.07072595e-02 -1.13636827e+00
-8.23221445e-01 -9.08622146e-01 -1.09033477e+00 -1.72045946e-01
-6.99937120e-02 -1.64424151e-01 2.96007190e-02 1.30553818e+00
7.66535342e-01 5.26773036e-01 4.58423615e-01 -9.52107489e-01
-2.06528261e-01 -8.35701168e-01 -1.16168104e-01 5.42259753e-01
3.08030725e-01 -8.94412577e-01 -7.58028477e-02 4.77113098e-01] | [8.205403327941895, 0.41960087418556213] |
5895c579-b442-474d-ab87-9d8f315583dc | mam-masked-acoustic-modeling-for-end-to-end | 2010.11445 | null | https://arxiv.org/abs/2010.11445v2 | https://arxiv.org/pdf/2010.11445v2.pdf | MAM: Masked Acoustic Modeling for End-to-End Speech-to-Text Translation | End-to-end Speech-to-text Translation (E2E-ST), which directly translates source language speech to target language text, is widely useful in practice, but traditional cascaded approaches (ASR+MT) often suffer from error propagation in the pipeline. On the other hand, existing end-to-end solutions heavily depend on the source language transcriptions for pre-training or multi-task training with Automatic Speech Recognition (ASR). We instead propose a simple technique to learn a robust speech encoder in a self-supervised fashion only on the speech side, which can utilize speech data without transcription. This technique termed Masked Acoustic Modeling (MAM), not only provides an alternative solution to improving E2E-ST, but also can perform pre-training on any acoustic signals (including non-speech ones) without annotation. We conduct our experiments over 8 different translation directions. In the setting without using any transcriptions, our technique achieves an average improvement of +1.1 BLEU, and +2.3 BLEU with MAM pre-training. Pre-training of MAM with arbitrary acoustic signals also has an average improvement with +1.6 BLEU for those languages. Compared with ASR multi-task learning solution, which replies on transcription during training, our pre-trained MAM model, which does not use transcription, achieves similar accuracy. | ['Liang Huang', 'Renjie Zheng', 'Mingbo Ma', 'Junkun Chen'] | 2020-10-22 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 3.77410620e-01 1.28478706e-01 1.94573719e-02 -4.27735269e-01
-1.76764512e+00 -6.00957632e-01 5.73016226e-01 -4.07059431e-01
-4.90831971e-01 5.09090602e-01 4.01418000e-01 -7.62712955e-01
7.02730358e-01 -2.16117397e-01 -9.95640814e-01 -5.10891080e-01
4.78523731e-01 5.54371655e-01 8.15576315e-02 -4.00206536e-01
-3.85864705e-01 -1.08515158e-01 -7.51193047e-01 6.62050784e-01
6.76703393e-01 6.80407107e-01 4.01918411e-01 9.47657526e-01
-2.14502558e-01 5.73449969e-01 -7.99037814e-01 -5.61756194e-01
2.04311758e-01 -6.47935033e-01 -6.03990793e-01 1.03095569e-01
1.54848441e-01 -2.97180355e-01 -3.20546597e-01 8.03530037e-01
8.24509144e-01 -1.84901580e-02 2.83353806e-01 -6.58423722e-01
-6.67278707e-01 1.02723467e+00 -5.65123439e-01 -7.46902078e-02
2.86599696e-01 1.49591863e-01 9.36141908e-01 -1.33818579e+00
1.98797956e-01 1.32785761e+00 4.66597974e-01 7.52631247e-01
-1.10112584e+00 -7.25369036e-01 -8.50724131e-02 -7.79293627e-02
-1.43792570e+00 -1.23846877e+00 5.68920970e-01 -1.45828024e-01
1.19212472e+00 3.79127055e-01 -1.97259896e-02 1.50028574e+00
-1.18474148e-01 9.17785645e-01 1.09252012e+00 -5.75568199e-01
6.09168857e-02 2.81650066e-01 -4.10296708e-01 2.91644901e-01
-5.18749893e-01 1.03825806e-02 -7.68094182e-01 4.66758907e-02
4.16786373e-01 -2.74048507e-01 -2.03794554e-01 5.27441740e-01
-1.56413090e+00 4.99608070e-01 -7.71240592e-02 2.96988249e-01
-2.77066082e-01 5.66618554e-02 5.69428027e-01 7.17432022e-01
7.56438732e-01 8.07782542e-03 -7.70693302e-01 -2.93642282e-01
-1.30510139e+00 -3.80565763e-01 5.90092719e-01 1.08156502e+00
6.73085093e-01 5.33803165e-01 -3.53677988e-01 1.28337061e+00
3.69740516e-01 1.01796758e+00 7.87289500e-01 -4.80866790e-01
9.62054074e-01 -1.11270145e-01 -9.16145369e-02 -4.01645191e-02
-8.80404115e-02 -6.82902157e-01 -8.88357639e-01 -2.83457220e-01
1.63367093e-01 -5.86309016e-01 -1.18355632e+00 1.65676320e+00
1.32101461e-01 4.11610715e-02 5.15755177e-01 9.12643909e-01
6.98669076e-01 1.15205181e+00 -3.07992131e-01 -4.81498688e-01
1.07690954e+00 -1.57131326e+00 -9.55020368e-01 -4.96723026e-01
8.09685171e-01 -1.23084629e+00 1.42705381e+00 2.72872210e-01
-1.24917769e+00 -6.64296806e-01 -8.28653336e-01 -1.02216132e-01
6.57828972e-02 5.65200448e-01 -1.49211779e-01 6.80406988e-01
-1.43121326e+00 1.19093165e-01 -8.97320807e-01 -1.67263120e-01
-1.40002802e-01 3.13614011e-01 -3.02872211e-01 2.35382020e-02
-1.17544270e+00 9.35501933e-01 4.91830856e-02 1.31423339e-01
-1.26954460e+00 -4.45970029e-01 -8.19716930e-01 3.65511775e-02
5.33189476e-01 -4.22621697e-01 1.72637868e+00 -1.35153687e+00
-2.37804437e+00 4.36673343e-01 -7.83055842e-01 -4.48554754e-01
4.30898964e-01 -3.72015715e-01 -5.86193740e-01 -1.92090020e-01
-1.59787849e-01 4.63867247e-01 1.10952950e+00 -1.02137589e+00
-3.84500116e-01 -3.24287601e-02 -3.88405532e-01 3.35244447e-01
-3.92313004e-01 5.88625789e-01 -4.86688524e-01 -8.41962695e-01
-5.50214089e-02 -9.29923654e-01 -1.88853413e-01 -5.74885368e-01
-4.43250865e-01 -4.19378914e-02 8.34886789e-01 -1.14767313e+00
1.27165067e+00 -2.14621568e+00 2.02623516e-01 -7.22663105e-02
-2.91253567e-01 4.28603232e-01 -3.96353513e-01 5.15594065e-01
3.85473371e-02 5.44668213e-02 -4.75530386e-01 -1.03884172e+00
-1.49198875e-01 1.35038301e-01 -4.78698909e-01 1.26516670e-01
3.69161189e-01 9.75602567e-01 -6.90972686e-01 -2.23745123e-01
1.69792131e-01 5.58060884e-01 -3.21966350e-01 4.70021993e-01
-1.89825267e-01 7.25788951e-01 4.45704758e-02 4.62015361e-01
3.86895031e-01 -9.98346210e-02 2.66446769e-01 1.05277374e-01
-1.59691453e-01 9.25107896e-01 -7.71732450e-01 1.88904047e+00
-1.13565278e+00 6.69239342e-01 2.93742418e-01 -8.76480162e-01
1.04436314e+00 9.22641575e-01 1.47069991e-01 -6.40194714e-01
-1.51732579e-01 6.72337890e-01 1.19610734e-01 -1.49907127e-01
3.13554198e-01 -2.54869640e-01 -1.86523851e-02 4.61608380e-01
2.63542145e-01 -4.29338925e-02 -2.63798207e-01 6.17118329e-02
1.13607621e+00 1.97448552e-01 7.33762905e-02 1.85837179e-01
5.36648333e-01 -1.81941032e-01 4.46279377e-01 4.03794974e-01
1.73385039e-01 8.61251771e-01 -9.50493217e-02 6.17232434e-02
-1.24464726e+00 -8.84212434e-01 2.97263324e-01 1.35814309e+00
-4.10895616e-01 -4.74806815e-01 -8.23569238e-01 -6.67609751e-01
-5.93050182e-01 8.77672553e-01 1.87268872e-02 1.17739066e-02
-9.51229870e-01 -7.03969777e-01 8.72051895e-01 3.19476426e-01
3.76742303e-01 -8.55476141e-01 2.11684644e-01 4.05039579e-01
-6.43967152e-01 -1.50236607e+00 -9.72406209e-01 2.49288797e-01
-7.01121569e-01 -1.69964097e-02 -1.16026962e+00 -7.54804552e-01
4.13565248e-01 4.08021867e-01 8.82835865e-01 -2.05307856e-01
6.63666487e-01 -6.65584803e-02 -5.23215890e-01 -1.52479440e-01
-1.07939935e+00 3.63452375e-01 1.80769667e-01 3.04440945e-01
1.15338109e-01 -4.73941803e-01 -2.41686538e-01 5.22996008e-01
-5.26860952e-01 2.13481918e-01 9.45766568e-01 8.86851192e-01
5.68973720e-01 -5.40106118e-01 8.87755632e-01 -6.56273067e-01
2.90596843e-01 -3.73020589e-01 -3.25828791e-01 4.41380650e-01
-4.02688861e-01 4.66894545e-02 1.01013398e+00 -4.59782541e-01
-1.14833009e+00 1.98238164e-01 -7.37218261e-01 -6.60534799e-01
-1.08259834e-01 4.50277537e-01 -2.98981547e-01 3.16792518e-01
6.46641672e-01 5.38660109e-01 -2.14542493e-01 -7.14632154e-01
4.32004273e-01 1.38915336e+00 4.52185452e-01 -2.47109905e-01
8.50832224e-01 -1.25633376e-02 -5.78503788e-01 -8.60186040e-01
-7.08764970e-01 -5.62392473e-01 -4.64440137e-01 9.16604921e-02
6.56426847e-01 -1.33508503e+00 -1.14015438e-01 2.90809661e-01
-1.44684362e+00 -6.47238374e-01 1.03656009e-01 7.31299520e-01
-4.74600405e-01 3.73136729e-01 -6.93915606e-01 -1.03466594e+00
-7.08494961e-01 -1.27955329e+00 1.41442311e+00 -4.51832265e-01
-1.51499972e-01 -7.86092877e-01 -8.52726251e-02 6.70645833e-01
7.26626754e-01 -5.65447688e-01 3.68741721e-01 -8.05893302e-01
-3.88854146e-01 1.12032808e-01 1.50171861e-01 7.37799764e-01
3.28937769e-01 -2.71301001e-01 -1.32979631e+00 -5.05053997e-01
9.18149948e-02 -2.38860726e-01 7.36510873e-01 1.77961349e-01
7.92092204e-01 -5.58904707e-01 3.62408124e-02 4.34554547e-01
9.29922462e-01 2.53142208e-01 5.35847843e-01 -1.04324408e-01
7.85506904e-01 5.05463839e-01 5.04526734e-01 -2.44628359e-02
4.18462336e-01 1.03470683e+00 -1.01363614e-01 -3.35726678e-01
-4.88686562e-01 -4.10066098e-01 1.34166574e+00 1.76082337e+00
2.72125691e-01 -5.08033097e-01 -9.30540025e-01 4.86904234e-01
-1.74367678e+00 -5.55933416e-01 -1.09617092e-01 2.27314615e+00
1.23013997e+00 9.65470225e-02 2.33885303e-01 5.52025512e-02
7.88430035e-01 8.66928622e-02 -3.35339367e-01 -5.21926522e-01
-2.34754175e-01 3.68125409e-01 5.98295271e-01 9.10849869e-01
-7.87249506e-01 1.37519109e+00 6.09956646e+00 1.16446543e+00
-1.51337862e+00 7.90361822e-01 5.89997470e-01 -1.61555082e-01
-3.61133903e-01 9.76435691e-02 -8.63404334e-01 5.06060183e-01
1.79025567e+00 2.64392514e-02 6.91209912e-01 6.18711054e-01
5.90010583e-01 3.29222739e-01 -1.02711523e+00 9.38720942e-01
7.17396736e-02 -1.04500568e+00 3.60662527e-02 -2.84429789e-01
6.64319575e-01 3.84029508e-01 9.94931348e-03 5.17095029e-01
2.83874512e-01 -8.97543728e-01 8.73655260e-01 -1.46963105e-01
1.31287289e+00 -4.68321323e-01 7.36805022e-01 6.61891878e-01
-1.12851417e+00 2.81923532e-01 -1.93188250e-01 6.67191222e-02
5.84575355e-01 7.03017175e-01 -1.35101032e+00 7.91072011e-01
3.66657883e-01 4.36092377e-01 -1.09406911e-01 4.38072383e-01
-4.96664166e-01 1.42845809e+00 -3.88928175e-01 3.94020863e-02
2.32361063e-01 -5.53164445e-02 6.04656577e-01 1.67003131e+00
6.22846842e-01 -3.07428390e-01 1.71634942e-01 3.05654317e-01
-3.13654423e-01 3.79339457e-01 -4.13447767e-01 -6.63001090e-02
4.81802762e-01 1.05783939e+00 -3.10486555e-01 -5.77454686e-01
-6.29626453e-01 1.46615195e+00 1.25019670e-01 5.40009975e-01
-7.74296343e-01 -2.42185771e-01 4.64074254e-01 -4.81575215e-03
2.14077845e-01 -5.62099636e-01 -3.11751664e-01 -1.15089500e+00
1.86763033e-01 -1.21767759e+00 -1.18380062e-01 -6.44040585e-01
-9.37986553e-01 9.13798392e-01 -4.82200801e-01 -1.17591906e+00
-5.55264175e-01 -2.50150174e-01 -4.37981427e-01 1.20401931e+00
-1.68202198e+00 -1.42496884e+00 3.01087260e-01 4.64469761e-01
1.23563480e+00 -3.21824610e-01 8.62891912e-01 6.67037010e-01
-6.53758705e-01 1.01015174e+00 2.09766850e-01 1.50484562e-01
1.08498693e+00 -1.02500677e+00 9.73130941e-01 1.12044823e+00
5.13769448e-01 4.19580609e-01 4.98531491e-01 -5.42180955e-01
-1.40186882e+00 -1.35550821e+00 1.21433723e+00 -4.41006690e-01
6.09099388e-01 -7.63970315e-01 -9.75246847e-01 8.84500802e-01
5.76464713e-01 -2.00702757e-01 6.77336335e-01 1.29666328e-01
-2.96964586e-01 -1.42670453e-01 -6.93487227e-01 6.86433613e-01
1.05341780e+00 -8.50832582e-01 -3.95950764e-01 4.61791784e-01
1.41008842e+00 -5.43028414e-01 -6.69393122e-01 2.55600274e-01
2.31746256e-01 -3.70401859e-01 6.38469100e-01 -4.19901490e-01
2.24846274e-01 -4.14823681e-01 -5.00933886e-01 -1.69467616e+00
4.43238840e-02 -1.30635440e+00 -6.98687807e-02 1.32088995e+00
1.07448697e+00 -5.33522725e-01 3.08526665e-01 2.25826036e-02
-7.58522630e-01 -5.50092101e-01 -1.09162724e+00 -9.46345389e-01
7.35711530e-02 -5.67617595e-01 4.30682719e-01 7.24064529e-01
-1.44778222e-01 8.22039068e-01 -8.83589983e-01 3.29716891e-01
2.77695745e-01 -2.61790335e-01 8.76877189e-01 -3.98004860e-01
-7.10596561e-01 -1.24401383e-01 2.32982650e-01 -1.57425249e+00
2.18315318e-01 -1.05589163e+00 4.34631914e-01 -1.36246133e+00
-1.23544976e-01 -2.82396287e-01 -1.45092517e-01 6.60628736e-01
-2.56288171e-01 2.80121446e-01 1.18443206e-01 2.72948563e-01
-4.02059883e-01 7.06232011e-01 1.08702087e+00 6.22868259e-03
-3.31018895e-01 1.11269832e-01 -4.39477623e-01 4.21135753e-01
8.86290848e-01 -6.55736089e-01 -2.35139519e-01 -9.62806404e-01
-9.18225646e-02 2.91627645e-01 -1.37345761e-01 -6.52907133e-01
1.90814942e-01 8.22222754e-02 -4.51306850e-02 -4.96498257e-01
5.53631306e-01 -5.24586320e-01 1.14125766e-01 1.63615271e-01
-4.17302936e-01 1.47316456e-01 1.19229905e-01 1.97040632e-01
-4.32475269e-01 -8.45022202e-02 7.17926741e-01 4.53024544e-02
-1.48556262e-01 9.89728123e-02 -5.97286642e-01 -1.28779352e-01
4.37179983e-01 1.82157621e-01 -7.58475140e-02 -6.88674748e-01
-6.85418546e-01 1.10628530e-01 8.42122436e-02 5.58168352e-01
5.47104657e-01 -1.26258326e+00 -1.18673110e+00 2.20219225e-01
-1.74732283e-01 -1.17250783e-02 -6.76206173e-03 9.54840779e-01
5.85368983e-02 4.05707926e-01 5.32534122e-01 -7.23041296e-01
-1.27386308e+00 3.10797185e-01 2.47701555e-01 -2.40110561e-01
-3.42780173e-01 8.10797334e-01 1.33078486e-01 -6.65777564e-01
1.89753667e-01 -1.46628082e-01 3.95301610e-01 -3.05177093e-01
4.49517459e-01 1.58950061e-01 5.17475486e-01 -7.78541327e-01
-3.37511629e-01 3.54011416e-01 -1.10093825e-01 -8.46893966e-01
1.04789972e+00 -4.44713920e-01 1.63302377e-01 7.50081420e-01
1.17909718e+00 4.43167090e-01 -9.02621031e-01 -5.21658540e-01
1.09765284e-01 -2.06021845e-01 1.73101559e-01 -9.94801641e-01
-8.89393926e-01 1.27158880e+00 2.03810081e-01 -1.55733198e-01
1.08962739e+00 -4.72275242e-02 1.26492810e+00 5.51095903e-01
3.20897758e-01 -1.04095793e+00 1.46259531e-01 7.85596490e-01
1.09490943e+00 -1.35395968e+00 -6.68615758e-01 -3.33148539e-01
-9.21337962e-01 9.91096735e-01 3.27331036e-01 3.24086040e-01
3.96177053e-01 5.05533397e-01 5.88688552e-01 5.62567413e-01
-9.15660679e-01 -2.52877951e-01 2.50084817e-01 2.92834491e-01
7.73788273e-01 1.85074404e-01 -5.30760176e-02 5.55541873e-01
-3.97174299e-01 -2.78656155e-01 3.06184351e-01 5.45366943e-01
-3.96397382e-01 -1.41611314e+00 -4.27743882e-01 1.05381295e-01
-6.45590961e-01 -5.51751494e-01 -4.51859713e-01 3.95528339e-02
-3.73017222e-01 1.46180165e+00 -2.01228276e-01 -6.32776976e-01
3.87046009e-01 3.34829271e-01 1.10880718e-01 -8.80752802e-01
-8.14665437e-01 7.09647477e-01 3.73304784e-01 -3.62733811e-01
-2.34155253e-01 -4.85759348e-01 -1.21386456e+00 -1.74797148e-01
-4.76319790e-01 2.42160216e-01 9.83039677e-01 9.41378474e-01
4.59790260e-01 5.71982205e-01 1.05939865e+00 -7.06028342e-01
-8.12486172e-01 -1.41387844e+00 4.26836424e-02 -6.93672597e-02
7.40132093e-01 -6.60961568e-02 -3.66941273e-01 1.88058242e-01] | [14.524922370910645, 7.135400295257568] |
90d734cd-d40c-43a6-aa8d-63e51e814aeb | variational-interaction-information-1 | 2012.04251 | null | https://arxiv.org/abs/2012.04251v1 | https://arxiv.org/pdf/2012.04251v1.pdf | Variational Interaction Information Maximization for Cross-domain Disentanglement | Cross-domain disentanglement is the problem of learning representations partitioned into domain-invariant and domain-specific representations, which is a key to successful domain transfer or measuring semantic distance between two domains. Grounded in information theory, we cast the simultaneous learning of domain-invariant and domain-specific representations as a joint objective of multiple information constraints, which does not require adversarial training or gradient reversal layers. We derive a tractable bound of the objective and propose a generative model named Interaction Information Auto-Encoder (IIAE). Our approach reveals insights on the desirable representation for cross-domain disentanglement and its connection to Variational Auto-Encoder (VAE). We demonstrate the validity of our model in the image-to-image translation and the cross-domain retrieval tasks. We further show that our model achieves the state-of-the-art performance in the zero-shot sketch based image retrieval task, even without external knowledge. Our implementation is publicly available at: https://github.com/gr8joo/IIAE | ['Kee-Eung Kim', 'Seunghoon Hong', 'Geon-Hyeong Kim', 'HyeongJoo Hwang'] | 2020-12-08 | variational-interaction-information | http://proceedings.neurips.cc/paper/2020/hash/fe663a72b27bdc613873fbbb512f6f67-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/fe663a72b27bdc613873fbbb512f6f67-Paper.pdf | neurips-2020-12 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 2.79269338e-01 4.64673358e-04 -5.26884019e-01 -2.20716596e-01
-1.38305092e+00 -9.93129969e-01 1.10191143e+00 -4.17508572e-01
-8.73584449e-02 6.91966712e-01 3.26452166e-01 -1.73403442e-01
-3.54894131e-01 -5.53894460e-01 -8.72403681e-01 -5.18785417e-01
2.52856493e-01 5.94375253e-01 -2.13684991e-01 -2.71486342e-01
1.30731732e-01 3.91929328e-01 -1.17853642e+00 2.80997962e-01
6.83558643e-01 9.00880039e-01 5.28647266e-02 6.47968590e-01
5.39783984e-02 5.72619915e-01 -2.82897115e-01 -5.12303650e-01
4.61486995e-01 -3.79317671e-01 -9.41791832e-01 1.19975926e-02
3.89192104e-01 -5.52632570e-01 -7.82826662e-01 9.42887068e-01
3.49042475e-01 2.67811716e-02 1.26467979e+00 -1.46884382e+00
-1.41758025e+00 8.81989896e-02 -4.11547840e-01 -5.67508601e-02
2.80754179e-01 -9.79187787e-02 1.34537208e+00 -1.05235445e+00
9.26664710e-01 1.20216703e+00 1.12601452e-01 7.67737269e-01
-1.64696050e+00 -8.26062799e-01 -9.51503441e-02 1.31697327e-01
-1.49291849e+00 -5.96743762e-01 9.64442611e-01 -7.54109204e-01
5.30128002e-01 -7.39386529e-02 2.95534462e-01 1.56453252e+00
-3.53734195e-02 1.00024951e+00 9.24984097e-01 -3.23027313e-01
1.33211702e-01 1.37310401e-01 -2.51151025e-01 6.34252787e-01
2.72584617e-01 2.46807590e-01 -6.34902239e-01 -2.00618565e-01
1.21617103e+00 2.15968993e-02 -2.56706119e-01 -1.11495829e+00
-1.12733412e+00 1.16928470e+00 3.90123636e-01 1.23154186e-01
-1.64785802e-01 3.15659881e-01 4.31530118e-01 6.67891383e-01
4.61005926e-01 5.02510250e-01 -3.35164011e-01 1.57262743e-01
-7.42032647e-01 3.01091135e-01 5.66763878e-01 1.08647001e+00
5.45020700e-01 -8.46911967e-03 -8.35260227e-02 8.28095078e-01
2.34961450e-01 7.17809439e-01 2.68818110e-01 -1.08909416e+00
5.18028378e-01 2.45319694e-01 2.01489478e-01 -7.25344658e-01
4.26006526e-01 -1.02151513e-01 -7.25619078e-01 2.33735159e-01
3.75848800e-01 1.15583837e-01 -8.14580619e-01 2.27230692e+00
-3.36365178e-02 1.00818895e-01 2.43092299e-01 9.38707709e-01
5.74418962e-01 5.60451984e-01 3.09963953e-02 1.79478690e-01
1.33941674e+00 -6.57826781e-01 -5.72005272e-01 -1.91813678e-01
2.68589675e-01 -5.68989992e-01 1.05284584e+00 -3.44088138e-03
-1.39311671e+00 -3.60390186e-01 -1.18916845e+00 -6.09172523e-01
-5.23417234e-01 -5.60289174e-02 5.38774133e-01 1.99840620e-01
-9.78013039e-01 4.10245210e-01 -6.82620347e-01 -2.09644675e-01
7.19825983e-01 2.28476375e-01 -7.43971050e-01 -2.50661790e-01
-1.36561346e+00 9.24300551e-01 7.52418488e-02 -6.02313042e-01
-1.24172521e+00 -9.36222136e-01 -9.17860806e-01 4.17975411e-02
2.46256113e-01 -1.19373500e+00 1.08710933e+00 -1.18745208e+00
-1.44549990e+00 1.27359939e+00 -7.07777813e-02 -2.82326043e-01
5.21454692e-01 -2.93163508e-01 -1.94621027e-01 3.44999075e-01
2.53974497e-01 7.28874683e-01 1.19968188e+00 -1.26727152e+00
3.66272256e-02 -3.65493000e-01 2.34014452e-01 3.10507208e-01
-2.08710104e-01 -2.65364677e-01 -3.68886560e-01 -7.17996061e-01
-2.18158379e-01 -9.08106148e-01 2.96182692e-01 6.16027713e-01
-1.47115633e-01 -8.76414105e-02 5.56341946e-01 -7.18554437e-01
7.44955897e-01 -2.18302321e+00 7.63172746e-01 -7.73422420e-02
3.24402571e-01 2.24763706e-01 -4.93097693e-01 6.32280111e-01
-2.60621488e-01 -1.19000198e-02 -3.95360798e-01 -3.25278372e-01
3.02771062e-01 2.69856662e-01 -7.22463965e-01 5.01594782e-01
5.62222898e-01 1.20892692e+00 -9.76399481e-01 -3.39561522e-01
1.23432145e-01 7.33636498e-01 -6.57650769e-01 4.41635787e-01
-2.45997846e-01 6.11556590e-01 -5.54863691e-01 3.71378005e-01
6.79542899e-01 -4.68961239e-01 2.60155439e-01 -1.31070644e-01
3.96799177e-01 3.54186088e-01 -7.25097954e-01 2.39976335e+00
-5.12223125e-01 8.91186774e-01 1.37635231e-01 -1.20232999e+00
6.66724324e-01 3.41265529e-01 3.39925259e-01 -7.61540473e-01
-4.97625209e-02 2.52805471e-01 -4.21302855e-01 -1.50676280e-01
4.10645932e-01 -3.71818781e-01 -4.07435060e-01 5.67670882e-01
5.09121597e-01 -2.70278394e-01 -1.55253619e-01 6.11486673e-01
8.30395579e-01 4.14306134e-01 4.53189105e-01 -4.44291204e-01
3.13541085e-01 -3.12206268e-01 2.44035572e-01 6.08347714e-01
-7.10954443e-02 7.28111565e-01 5.83539009e-01 -7.84773082e-02
-1.43019760e+00 -1.68411517e+00 -1.13822155e-01 1.05091393e+00
2.77506709e-01 -1.39741153e-01 -5.12923777e-01 -6.34021163e-01
2.35064000e-01 6.16972446e-01 -7.64450490e-01 -3.11670065e-01
-1.47742704e-01 1.26036093e-01 6.97692573e-01 4.13868189e-01
2.63982862e-01 -5.83352268e-01 -1.89761221e-01 -2.67962486e-01
-2.56053567e-01 -9.49847937e-01 -6.63084209e-01 -1.20930433e-01
-6.08225763e-01 -9.86405015e-01 -1.14968336e+00 -6.74937427e-01
5.27178705e-01 3.15736562e-01 1.34986484e+00 -4.26034957e-01
-2.67619342e-01 8.78446519e-01 -6.96283132e-02 -8.66876692e-02
-3.92057687e-01 1.41035467e-02 -5.15519157e-02 -1.54023111e-01
2.25912973e-01 -9.32525396e-01 -8.15439939e-01 1.87119469e-01
-1.11846530e+00 8.68498161e-02 5.06616890e-01 1.10739648e+00
4.89171594e-01 -8.15782487e-01 6.55694127e-01 -7.68562555e-01
7.93796420e-01 -7.06127107e-01 -7.10978270e-01 3.04394096e-01
-3.32724124e-01 5.91717541e-01 3.92850786e-01 -4.68022287e-01
-1.03007317e+00 -1.49066836e-01 2.22687259e-01 -9.19086874e-01
4.37787920e-02 1.47088096e-01 -3.51772428e-01 1.89050660e-01
6.42455339e-01 3.80632490e-01 1.21436968e-01 -3.27252328e-01
1.02825248e+00 4.09692556e-01 5.36612988e-01 -9.95253503e-01
9.76314247e-01 6.53802752e-01 -9.98089612e-02 -5.54799438e-01
-8.31454933e-01 -3.23468924e-01 -6.95155799e-01 1.81020677e-01
9.43437874e-01 -1.35201108e+00 -5.38058937e-01 6.49364432e-03
-1.28591239e+00 -2.60845929e-01 -4.04508501e-01 1.26428351e-01
-1.01085746e+00 2.28403807e-01 -4.60650712e-01 -4.88378137e-01
-2.31083885e-01 -9.54766810e-01 1.35947740e+00 -1.45791292e-01
-1.92273423e-01 -1.17911828e+00 2.71817476e-01 3.28490317e-01
2.96242386e-01 1.51240200e-01 9.59718645e-01 -5.78332007e-01
-8.50336134e-01 5.02944030e-02 -4.61042434e-01 3.65974128e-01
3.25579941e-02 -4.56729919e-01 -9.51930702e-01 -4.67063427e-01
-2.27190211e-01 -7.61319101e-01 1.02457619e+00 1.68977022e-01
1.01928926e+00 -3.84504080e-01 -1.97645068e-01 7.98014402e-01
1.59189272e+00 -1.35854021e-01 7.10983992e-01 -2.92762555e-02
5.34140289e-01 3.70993644e-01 3.10817599e-01 4.36629593e-01
3.74088794e-01 8.00669074e-01 2.18069240e-01 -2.04991270e-02
-2.73355901e-01 -6.86275363e-01 2.75368243e-01 6.44949079e-01
9.51954871e-02 -1.88938811e-01 -5.90645909e-01 8.98101091e-01
-1.75897968e+00 -1.06100595e+00 6.55308366e-01 2.21299529e+00
9.11277533e-01 -3.06726664e-01 -7.54939718e-03 -4.17887747e-01
5.39347231e-01 3.64380181e-01 -8.63114834e-01 -4.25277293e-01
-4.14665677e-02 3.83758068e-01 3.87198418e-01 7.40866780e-01
-1.03878260e+00 9.88334298e-01 6.06294870e+00 8.89304340e-01
-8.50877583e-01 3.29618663e-01 3.07087123e-01 -1.50221780e-01
-7.58383930e-01 -1.80141926e-02 -1.86547264e-01 2.75803208e-01
6.29724324e-01 -5.50710440e-01 7.64362812e-01 8.22380543e-01
-4.38450247e-01 4.55500573e-01 -1.48467016e+00 1.00543487e+00
3.53977457e-02 -1.49827230e+00 3.20647448e-01 2.64339775e-01
8.79057288e-01 -5.02117835e-02 5.64071417e-01 2.50706017e-01
6.68488681e-01 -9.95464265e-01 6.31681204e-01 4.41564888e-01
1.45797384e+00 -5.09669781e-01 -1.10431239e-02 -6.13598078e-02
-9.68981564e-01 8.22030902e-02 -3.71528327e-01 2.29199797e-01
-5.61799556e-02 9.82362628e-02 -5.66753149e-01 6.33556128e-01
1.05547369e-01 8.83052230e-01 1.00432159e-02 2.21807584e-01
-2.92069435e-01 9.02599990e-02 3.49345580e-02 3.89151037e-01
-1.47266239e-02 -2.46999413e-01 7.34829843e-01 1.05577743e+00
2.03182638e-01 6.58695623e-02 -2.10778892e-01 1.40888321e+00
-4.75122184e-01 -3.91816884e-01 -1.07937312e+00 -3.75396132e-01
6.09031737e-01 7.05610335e-01 4.56147231e-02 -3.23548973e-01
-3.98480922e-01 1.51099670e+00 5.71279645e-01 7.35091746e-01
-7.70131290e-01 -4.08238143e-01 1.41790903e+00 -4.36397940e-02
5.07919967e-01 -2.60356456e-01 -1.93261668e-01 -1.56116319e+00
6.34115711e-02 -8.30386519e-01 3.17880750e-01 -8.17394793e-01
-1.64841890e+00 2.29122803e-01 1.31726190e-01 -1.35073686e+00
-4.95368183e-01 -7.29552388e-01 -3.19636613e-01 1.08101666e+00
-1.58085430e+00 -1.42010939e+00 9.73322690e-02 9.45493877e-01
4.57244188e-01 -2.32137620e-01 1.07436252e+00 3.24619174e-01
-2.75210310e-02 8.83351386e-01 5.44427693e-01 2.75052160e-01
8.14255178e-01 -1.03791130e+00 4.98141348e-01 6.21287525e-01
1.82423159e-01 7.36619473e-01 5.46520054e-01 -2.94774950e-01
-1.59183991e+00 -8.44977677e-01 7.52923727e-01 -7.43596733e-01
9.02807593e-01 -7.13282228e-01 -8.25316072e-01 8.90709579e-01
1.85967162e-01 8.28116015e-02 7.41247654e-01 3.99805270e-02
-1.16888011e+00 1.82325110e-01 -1.06421244e+00 6.01963341e-01
1.22774541e+00 -1.29674661e+00 -8.07795823e-01 2.54996508e-01
9.66633976e-01 -1.94738477e-01 -9.30673480e-01 -9.41376388e-02
8.92757952e-01 -7.71572888e-01 1.47505116e+00 -9.84286845e-01
9.98175621e-01 8.49029496e-02 -5.36760032e-01 -1.26416934e+00
-4.01068777e-01 -4.30655092e-01 -1.33845821e-01 1.04874051e+00
3.62984419e-01 -6.74093843e-01 4.50520813e-01 5.96314967e-01
3.52130562e-01 -3.82088363e-01 -9.51059103e-01 -9.62265491e-01
7.86997497e-01 -1.09487824e-01 4.62188691e-01 1.16557562e+00
4.78562154e-02 3.96267056e-01 -5.82933307e-01 6.43925369e-02
7.71058559e-01 3.13804477e-01 7.54174471e-01 -1.10427082e+00
-4.81629401e-01 -4.54030007e-01 -4.29928660e-01 -1.32350659e+00
5.36646962e-01 -1.22867990e+00 -2.71141201e-01 -1.38064551e+00
5.28218269e-01 -1.12166144e-01 -4.35013026e-01 2.08252564e-01
1.26220688e-01 1.36053070e-01 3.84750128e-01 3.82746130e-01
-6.48101449e-01 8.61808002e-01 1.31843364e+00 -2.44029984e-01
2.38799348e-01 -4.86651242e-01 -9.57648218e-01 2.52392411e-01
6.47374153e-01 -4.72236484e-01 -7.01439619e-01 -6.71244264e-01
1.12265170e-01 2.90443927e-01 5.69557130e-01 -3.82103890e-01
-4.41352986e-02 -2.11456910e-01 2.12489411e-01 -3.49563472e-02
7.48420477e-01 -7.01467454e-01 1.82327703e-02 2.68656731e-01
-8.07016194e-01 -2.62099802e-01 1.50730148e-01 7.77817011e-01
-1.76408321e-01 1.72968090e-01 7.16467738e-01 -7.05544874e-02
-6.00769460e-01 4.95982021e-01 -5.11462390e-02 4.60419387e-01
6.79030299e-01 9.70152915e-02 -4.09643620e-01 -6.53173625e-01
-6.66139126e-01 -3.64855044e-02 6.30619526e-01 5.88998377e-01
7.06455708e-01 -1.68484521e+00 -8.22280169e-01 4.45335388e-01
4.12158996e-01 -3.04294288e-01 1.97806671e-01 2.48899892e-01
-1.06976070e-01 6.00693762e-01 -4.65908796e-01 -4.22976047e-01
-9.74635065e-01 5.93893051e-01 2.03745008e-01 -3.55534822e-01
-3.50545734e-01 8.24812829e-01 8.21955323e-01 -3.87719691e-01
-6.93075508e-02 2.61819065e-01 3.58940065e-01 -1.30229667e-01
3.68051678e-01 1.08911119e-01 -4.87559438e-01 -4.68145072e-01
-3.20247173e-01 5.56122780e-01 -2.00696021e-01 -6.28108859e-01
1.12357974e+00 -1.66146070e-01 1.51527971e-01 3.69757622e-01
1.80783391e+00 -3.60949397e-01 -1.47449958e+00 -4.99355495e-01
-3.38133246e-01 -7.11387873e-01 -1.46642923e-01 -7.12054551e-01
-7.82343566e-01 1.25714195e+00 4.54293013e-01 -2.25087851e-01
9.95216429e-01 4.38642055e-01 6.05414689e-01 2.68493623e-01
3.93773735e-01 -7.92298377e-01 4.49456304e-01 3.62194031e-01
1.36452055e+00 -1.28662157e+00 -1.18140593e-01 -1.06442370e-01
-8.25987220e-01 6.02111518e-01 2.26190940e-01 -5.58327436e-01
6.87301219e-01 -7.44645745e-02 -2.91499734e-01 -6.57186434e-02
-9.12126303e-01 -1.51530370e-01 6.59348845e-01 8.95060480e-01
2.94791639e-01 2.38363355e-01 -1.26574829e-01 4.82636034e-01
9.95499790e-02 8.05419832e-02 6.63770586e-02 7.37257361e-01
-8.00010003e-03 -1.30596948e+00 1.30779922e-01 -2.85365507e-02
-4.07690287e-01 -2.53008932e-01 -5.88208020e-01 7.81567574e-01
-4.18282121e-01 5.12323380e-01 1.81345925e-01 -1.37528718e-01
9.87637788e-02 2.64536679e-01 8.51890862e-01 -2.51161546e-01
6.69180676e-02 -2.11968616e-01 -2.10672453e-01 -6.62488520e-01
-2.31858104e-01 -5.14506638e-01 -7.26739466e-01 -3.82744342e-01
2.05672886e-02 1.16004739e-02 6.67072475e-01 6.75980210e-01
8.43822002e-01 3.49754840e-01 5.50759852e-01 -7.72883773e-01
-6.94741607e-01 -5.71171284e-01 -4.87439960e-01 7.91442275e-01
6.18976414e-01 -8.46739292e-01 -2.60079592e-01 1.74807593e-01] | [11.536447525024414, 0.6232125163078308] |
4bc6a274-9a66-4402-9a3c-5b016a93da79 | debatesum-a-large-scale-argument-mining-and | 2011.07251 | null | https://arxiv.org/abs/2011.07251v1 | https://arxiv.org/pdf/2011.07251v1.pdf | DebateSum: A large-scale argument mining and summarization dataset | Prior work in Argument Mining frequently alludes to its potential applications in automatic debating systems. Despite this focus, almost no datasets or models exist which apply natural language processing techniques to problems found within competitive formal debate. To remedy this, we present the DebateSum dataset. DebateSum consists of 187,386 unique pieces of evidence with corresponding argument and extractive summaries. DebateSum was made using data compiled by competitors within the National Speech and Debate Association over a 7-year period. We train several transformer summarization models to benchmark summarization performance on DebateSum. We also introduce a set of fasttext word-vectors trained on DebateSum called debate2vec. Finally, we present a search engine for this dataset which is utilized extensively by members of the National Speech and Debate Association today. The DebateSum search engine is available to the public here: http://www.debate.cards | ['Arvind Balaji', 'Allen Roush'] | 2020-11-14 | null | https://aclanthology.org/2020.argmining-1.1 | https://aclanthology.org/2020.argmining-1.1.pdf | coling-argmining-2020-12 | ['query-based-extractive-summarization', 'extractive-document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.31441289e-01 6.31763816e-01 -8.13601196e-01 -2.75901079e-01
-1.68595040e+00 -7.96759903e-01 1.24868119e+00 7.17397630e-01
-5.99275947e-01 1.21010220e+00 1.43869936e+00 -8.58366728e-01
1.42677091e-02 -5.34850478e-01 -5.47126710e-01 -2.25083217e-01
4.88079756e-01 7.81392992e-01 -5.86267151e-02 -5.66767156e-01
7.82787263e-01 -3.04649174e-01 -1.26348329e+00 9.63547289e-01
1.01349771e+00 7.86922872e-01 -2.63472944e-01 1.08817351e+00
-6.08731151e-01 9.73133862e-01 -1.21092057e+00 -9.21352744e-01
-3.52538414e-02 -5.39057910e-01 -1.30574083e+00 -2.15572998e-01
9.12463486e-01 -5.64949885e-02 -4.70776528e-01 7.56197929e-01
7.52979100e-01 -3.78256440e-02 5.63087881e-01 -8.67475569e-01
-6.18788362e-01 1.49055755e+00 -3.08452129e-01 8.99260104e-01
4.72277910e-01 -3.45733851e-01 1.88743925e+00 -6.83194220e-01
1.16379678e+00 1.25433195e+00 3.77101898e-01 2.79463500e-01
-8.53944898e-01 -5.08057058e-01 1.34432226e-01 5.97268581e-01
-2.44148269e-01 -6.42802358e-01 9.44632292e-01 -1.37315661e-01
1.09839892e+00 5.37618399e-01 8.52115870e-01 1.62490439e+00
2.57142514e-01 1.39325547e+00 1.05075061e+00 -5.10613143e-01
3.69647816e-02 -1.41902030e-01 9.32447910e-01 1.72926798e-01
5.69872141e-01 -4.60985899e-01 -7.67019331e-01 -8.74855578e-01
-6.95463046e-02 -9.64612901e-01 -4.70547408e-01 8.49844441e-02
-1.05179322e+00 1.49783957e+00 -1.78574055e-01 3.33416700e-01
-3.21156830e-01 -3.28160562e-02 1.01093054e+00 6.74016297e-01
9.16777313e-01 6.70153916e-01 -4.73466575e-01 -5.10608077e-01
-1.08479214e+00 8.33498120e-01 1.40900660e+00 2.50471026e-01
3.28558087e-02 -2.32022941e-01 -3.47085923e-01 1.14768350e+00
8.35821554e-02 4.63291168e-01 6.65172815e-01 -1.20611799e+00
1.12702417e+00 5.52682638e-01 -2.16584764e-02 -1.01616883e+00
-2.50430197e-01 -4.91794080e-01 -4.27133948e-01 -3.10555935e-01
4.73053716e-02 -4.36199903e-01 -1.90176770e-01 1.15272975e+00
3.13762158e-01 -4.13681358e-01 4.81864333e-01 4.71629411e-01
1.57136345e+00 7.58824587e-01 -2.55199909e-01 -4.29450661e-01
1.56100249e+00 -1.13005817e+00 -9.86784339e-01 -8.95892009e-02
6.99344277e-01 -1.25404227e+00 7.61970699e-01 3.29072237e-01
-1.38311982e+00 -2.61806436e-02 -1.29710746e+00 -3.52348059e-01
-1.01677693e-01 -4.14751321e-02 6.07260108e-01 3.26469868e-01
-6.70078218e-01 2.38177195e-01 -1.86464265e-01 -2.64317125e-01
6.22103989e-01 -3.46566647e-01 -1.09383408e-02 2.41461352e-01
-1.42474794e+00 1.33493114e+00 6.83405474e-02 -3.73891950e-01
-3.27948242e-01 -4.42332327e-01 -9.82145846e-01 -9.03204829e-02
6.05015159e-01 -7.91069448e-01 1.96006739e+00 -6.60739660e-01
-1.23428833e+00 9.79663432e-01 -1.44573957e-01 -1.03379035e+00
5.77044845e-01 -7.05864310e-01 -4.83213305e-01 1.84771404e-01
4.38681036e-01 1.35663182e-01 4.91610557e-01 -9.36600685e-01
-6.51755929e-01 8.15363750e-02 9.52210128e-02 2.46828631e-01
-5.83477654e-02 3.37401658e-01 -4.34965566e-02 -8.32274199e-01
-2.91183949e-01 -6.98116243e-01 1.70282051e-02 -6.78512335e-01
-7.68202066e-01 -8.35017622e-01 6.88487828e-01 -9.44898307e-01
1.50512910e+00 -1.35922921e+00 1.80382073e-01 -1.88750073e-01
7.36958683e-02 4.10658181e-01 -1.38151377e-01 8.62636805e-01
-4.88118790e-02 5.10934353e-01 -2.39756137e-01 -9.14702937e-02
1.59198269e-01 2.86491573e-01 -8.10813725e-01 4.58884507e-01
-1.08079433e-01 1.08734334e+00 -8.43370080e-01 -6.63116515e-01
-1.86542973e-01 9.82429683e-02 -5.00912726e-01 -2.54229277e-01
-4.21186924e-01 -1.56082988e-01 -7.97712862e-01 3.95662099e-01
3.84004891e-01 -1.12408310e-01 2.56170303e-01 -4.68281209e-02
-4.15585458e-01 1.34184039e+00 -5.96857369e-01 1.67116940e+00
-1.06105052e-01 1.17586637e+00 2.00867906e-01 -1.16425097e+00
5.41531682e-01 4.81754214e-01 2.83728778e-01 -7.43004441e-01
7.18686700e-01 2.20083296e-01 -1.17222667e-01 -3.82474005e-01
1.10320711e+00 1.95194960e-01 -4.78525788e-01 1.02087831e+00
-2.07508177e-01 -6.32791638e-01 7.02512324e-01 7.54016459e-01
1.18360972e+00 -2.94582546e-01 4.20010805e-01 -3.40096831e-01
5.05829275e-01 6.10327065e-01 4.06374395e-01 8.33953083e-01
-3.84914465e-02 3.88158262e-01 6.03490353e-01 -5.31083643e-01
-1.13821721e+00 -6.66454196e-01 -4.06276196e-01 8.35681856e-01
-2.20322743e-01 -8.32480371e-01 -5.05718291e-01 -8.23726237e-01
2.65901804e-01 9.57697451e-01 -7.54725993e-01 4.27086979e-01
-8.57750952e-01 -5.66566586e-01 6.45947516e-01 2.22646266e-01
5.29334068e-01 -1.16139138e+00 -7.87199676e-01 4.89219338e-01
-7.37523735e-01 -8.45313966e-01 -5.16798198e-01 8.53573531e-02
-5.12489140e-01 -1.69724834e+00 -6.26456916e-01 -5.11800289e-01
-1.36951357e-01 1.79414049e-01 1.51659071e+00 9.10131857e-02
-2.10818440e-01 4.40760553e-01 -4.99318689e-01 -7.68887520e-01
-8.10660481e-01 5.44225037e-01 -3.53314936e-01 -6.39291823e-01
4.39540483e-02 -1.71767548e-01 -3.33387464e-01 -3.02252531e-01
-6.82106256e-01 -6.28040230e-04 3.70998979e-01 1.11219382e+00
1.30583197e-01 -6.95247531e-01 7.91505933e-01 -1.00258839e+00
1.69747400e+00 -5.87351918e-01 -1.51894152e-01 3.45698982e-01
-5.89487612e-01 1.09686874e-01 -1.61453858e-01 -1.17593214e-01
-1.07156098e+00 -1.07996440e+00 -4.32184249e-01 5.36450386e-01
2.31427431e-01 9.41369951e-01 4.44255620e-01 6.19685352e-01
7.92793870e-01 -2.86775708e-01 7.54985735e-02 -3.43614668e-01
8.84254277e-01 7.79646993e-01 5.83510995e-01 -7.59176910e-01
5.61098337e-01 1.80597261e-01 -7.95622826e-01 -9.91602063e-01
-1.50946248e+00 -5.55436730e-01 3.72754559e-02 -7.68789724e-02
6.58092082e-01 -7.23188937e-01 -4.13262695e-01 -1.55498445e-01
-1.37029707e+00 -7.15813637e-02 -4.93018270e-01 3.85111839e-01
-3.27725023e-01 6.32347107e-01 -7.55778670e-01 -6.09640002e-01
-1.04740918e+00 -6.26959562e-01 5.63844740e-01 2.09820941e-01
-1.05399334e+00 -9.09625173e-01 4.94473428e-01 1.01566052e+00
3.96475464e-01 2.77339309e-01 8.24466944e-01 -9.76601660e-01
2.79797912e-01 -8.41670483e-02 3.15631814e-02 2.98733771e-01
-2.02726930e-01 3.08870226e-01 -5.58767200e-01 -7.97948539e-02
-8.38892460e-02 -6.47208989e-01 1.40738690e+00 6.28209352e-01
5.27305424e-01 -6.85518384e-01 -1.00171410e-01 -3.09258819e-01
7.09555030e-01 -2.00180948e-01 1.86764300e-01 9.57555830e-01
-2.04128008e-02 6.21560156e-01 4.43456978e-01 4.71290201e-01
7.20179737e-01 2.75376678e-01 3.92026417e-02 2.75346249e-01
-1.22044444e-01 -1.11623086e-01 3.30201268e-01 1.38566494e+00
9.81738940e-02 -6.67796016e-01 -8.19255650e-01 8.33818316e-01
-2.03248453e+00 -1.50422585e+00 -4.99471545e-01 1.41356933e+00
1.06753552e+00 4.87703115e-01 -7.26302490e-02 2.15419680e-01
6.26558781e-01 8.18009079e-01 -2.95920014e-01 -1.14850175e+00
-7.77300715e-01 6.48038745e-01 3.04125220e-01 6.28068924e-01
-1.08467913e+00 6.03996634e-01 6.38812733e+00 7.65153825e-01
-6.39692962e-01 2.17085570e-01 5.91883659e-01 -3.63970518e-01
-8.66246104e-01 9.65021327e-02 -7.00882137e-01 1.88921779e-01
1.01449633e+00 -1.04697478e+00 -3.95600051e-01 6.68021500e-01
1.36856735e-01 -2.11807787e-01 -5.37622094e-01 6.01405442e-01
3.78393292e-01 -2.29010797e+00 1.08271077e-01 6.37069643e-02
9.02605534e-01 3.97888869e-01 -7.86275193e-02 5.83177686e-01
8.28415155e-01 -8.39485824e-01 8.58624279e-01 -1.10594578e-01
1.71136320e-01 -7.25160897e-01 9.23376024e-01 3.35219830e-01
-5.61080992e-01 -5.62303960e-02 -2.63201684e-01 -1.95995882e-01
5.90170264e-01 5.13407886e-01 -5.95709682e-01 7.94074357e-01
4.56091255e-01 6.71560645e-01 -3.14315021e-01 8.38376284e-01
-6.36006951e-01 1.15425265e+00 -1.89871430e-01 -3.91871125e-01
7.46903896e-01 -3.14163091e-03 8.35058153e-01 1.31396365e+00
4.54316288e-02 2.12682530e-01 4.71800715e-02 1.32954150e-01
-6.45919263e-01 5.11119783e-01 -3.01973104e-01 -2.11813524e-01
6.06831133e-01 1.10909677e+00 -1.74694508e-01 -7.87681997e-01
-3.47295076e-01 5.35725713e-01 2.47824818e-01 2.24979669e-02
-7.58616447e-01 -5.15313260e-02 5.43538511e-01 -1.04795732e-01
3.01953495e-01 3.50107998e-02 -1.28339261e-01 -1.29262614e+00
2.06789579e-02 -1.38382924e+00 8.29477072e-01 -4.64636147e-01
-1.32416868e+00 6.06588244e-01 1.69848025e-01 -7.06152916e-01
-3.12768161e-01 -2.61789225e-02 -9.34803426e-01 5.11949539e-01
-1.53589618e+00 -8.29629838e-01 2.20152050e-01 9.32325572e-02
1.23255897e+00 -1.66154876e-01 7.30419219e-01 -5.84727004e-02
-4.44226503e-01 2.86186844e-01 1.25412732e-01 9.73191708e-02
1.02978384e+00 -1.15715015e+00 7.62541711e-01 2.74006844e-01
4.20898408e-01 5.39435267e-01 1.11270249e+00 -6.03099167e-01
-9.56018806e-01 -6.04690909e-01 1.56433749e+00 -7.36059904e-01
1.11096644e+00 1.56049311e-01 -8.64575267e-01 4.63299721e-01
1.37009382e+00 -8.14304769e-01 9.80654478e-01 3.99008811e-01
-4.47810352e-01 4.80549365e-01 -6.34066880e-01 6.20533705e-01
6.23579025e-01 -5.22647560e-01 -1.71030784e+00 3.69943887e-01
4.61486995e-01 -6.00615084e-01 -5.14125228e-01 3.21060508e-01
6.93530679e-01 -4.39445704e-01 7.13292778e-01 -9.01850998e-01
1.04726827e+00 2.72066981e-01 -1.22621909e-01 -1.57579720e+00
1.95770711e-01 -8.02238822e-01 -2.76133358e-01 1.15871918e+00
8.75266194e-01 -6.06959999e-01 5.67116559e-01 1.86923847e-01
-4.19891149e-01 -9.28775251e-01 -1.26659513e+00 -2.24203870e-01
7.01256812e-01 -4.86395270e-01 1.63705260e-01 1.12188923e+00
2.62804151e-01 7.89597929e-01 -1.27040058e-01 -5.88145673e-01
5.90612471e-01 6.06348157e-01 8.72092962e-01 -1.48589170e+00
4.40424271e-02 -8.90380323e-01 2.91471958e-01 -1.20672655e+00
5.14184058e-01 -9.75188434e-01 -2.33187437e-01 -2.17310476e+00
3.65087330e-01 8.97779763e-02 8.46542269e-02 8.27987120e-02
-2.34548166e-01 -1.07793674e-01 -3.84799279e-02 3.04373115e-01
-5.06903052e-01 5.69264412e-01 1.00778830e+00 -7.36900687e-01
1.17738299e-01 -2.66148031e-01 -1.28558600e+00 7.55072713e-01
1.11877716e+00 -3.66991937e-01 -1.26853198e-01 -4.04193074e-01
3.28184724e-01 1.40993431e-01 -1.53386090e-02 -3.61871183e-01
3.24654311e-01 -4.44981530e-02 -9.97155756e-02 -1.17861104e+00
1.80169195e-01 2.46693507e-01 -3.67895663e-01 1.82326451e-01
-8.39576364e-01 4.59689498e-01 2.17086047e-01 5.32123148e-01
-4.72338915e-01 -2.97608197e-01 2.45166570e-01 -6.59264922e-02
-2.35705525e-01 -2.41763473e-01 -7.80600965e-01 8.17598403e-01
5.70627749e-01 1.26399130e-01 -1.11631262e+00 -8.13569486e-01
-4.77102607e-01 5.81425250e-01 1.30861789e-01 5.68674624e-01
3.35942805e-01 -1.08793473e+00 -1.60596013e+00 -6.34843171e-01
-3.29273976e-02 -2.56150275e-01 4.01974060e-02 8.02808464e-01
-2.77029812e-01 5.83172202e-01 2.52737343e-01 -5.32040074e-02
-1.65335155e+00 -1.52730569e-01 -3.10016513e-01 -5.36852241e-01
-8.10250878e-01 8.68254602e-01 -4.62368101e-01 -1.59566537e-01
6.25031739e-02 -5.34260035e-01 -5.10644913e-01 8.39755237e-01
6.85139000e-01 2.73785740e-01 2.70936117e-02 -7.14842498e-01
-6.72556162e-02 1.79819036e-02 -3.17867965e-01 -4.20916200e-01
1.57087946e+00 -2.16136742e-02 -6.68549836e-02 4.86427903e-01
9.73892391e-01 5.91604173e-01 -4.79550242e-01 -2.54877120e-01
4.44707453e-01 1.46977514e-01 1.52618900e-01 -7.30910301e-01
-5.77127516e-01 5.01059473e-01 -3.76416951e-01 5.37238479e-01
1.49174750e-01 3.47930074e-01 1.04397905e+00 6.96342170e-01
-2.63435632e-01 -1.52638054e+00 -1.04354702e-01 9.18980598e-01
1.46208024e+00 -1.08810270e+00 4.51380670e-01 -3.11400257e-02
-7.08718181e-01 1.11092591e+00 -1.80133522e-01 -1.12161785e-01
4.41749245e-01 1.48718283e-01 4.16915625e-01 -7.30540872e-01
-1.27498031e+00 6.49468228e-02 1.84378922e-01 2.73905601e-02
8.51380229e-01 4.63147946e-02 -1.25339150e+00 7.72683263e-01
-7.98302054e-01 -5.79371393e-01 8.76252055e-01 1.05917096e+00
-7.63171196e-01 -1.34156990e+00 -3.21860969e-01 6.16500616e-01
-9.68724370e-01 -2.86073685e-01 -1.08227408e+00 8.94050777e-01
-4.97523665e-01 1.36143696e+00 -7.21231997e-02 1.18193947e-01
4.63272095e-01 1.91805020e-01 2.25081801e-01 -5.90210199e-01
-1.03733647e+00 -1.98157370e-01 1.51775312e+00 -2.76213795e-01
-6.91976249e-01 -9.34050143e-01 -9.40765440e-01 -7.18393683e-01
-4.92213905e-01 9.98913229e-01 6.94180906e-01 8.22122335e-01
1.22127414e-01 5.48406065e-01 1.57997131e-01 -5.01293302e-01
-7.76810706e-01 -1.30221283e+00 -8.71929973e-02 4.23954427e-01
2.42411911e-01 -4.18576658e-01 -5.44839323e-01 -2.07751676e-01] | [12.107755661010742, 9.607613563537598] |
58f42212-505b-479a-9456-937f1d14f033 | a-deep-model-for-partial-multi-label-image | 2207.02410 | null | https://arxiv.org/abs/2207.02410v1 | https://arxiv.org/pdf/2207.02410v1.pdf | A Deep Model for Partial Multi-Label Image Classification with Curriculum Based Disambiguation | In this paper, we study the partial multi-label (PML) image classification problem, where each image is annotated with a candidate label set consists of multiple relevant labels and other noisy labels. Existing PML methods typically design a disambiguation strategy to filter out noisy labels by utilizing prior knowledge with extra assumptions, which unfortunately is unavailable in many real tasks. Furthermore, because the objective function for disambiguation is usually elaborately designed on the whole training set, it can be hardly optimized in a deep model with SGD on mini-batches. In this paper, for the first time we propose a deep model for PML to enhance the representation and discrimination ability. On one hand, we propose a novel curriculum based disambiguation strategy to progressively identify ground-truth labels by incorporating the varied difficulties of different classes. On the other hand, a consistency regularization is introduced for model retraining to balance fitting identified easy labels and exploiting potential relevant labels. Extensive experimental results on the commonly used benchmark datasets show the proposed method significantly outperforms the SOTA methods. | ['Sheng-Jun Huang', 'Ming-Kun Xie', 'Feng Sun'] | 2022-07-06 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 4.20127571e-01 -7.97115043e-02 -7.07529560e-02 -5.26730895e-01
-1.13477087e+00 -3.87096077e-01 8.28156173e-02 2.81313956e-01
-4.95930284e-01 6.73779666e-01 -2.48830631e-01 7.62444660e-02
-2.93413639e-01 -3.32649559e-01 -4.91602957e-01 -9.72949147e-01
7.19239235e-01 6.08296096e-01 1.34653717e-01 1.62830904e-01
1.04059957e-01 8.35511833e-02 -1.54617989e+00 2.43466813e-03
1.18355155e+00 1.19097745e+00 6.52247429e-01 -3.33642252e-02
-6.99340701e-02 6.32225335e-01 -7.02718854e-01 -2.94915020e-01
1.66665718e-01 -2.75887191e-01 -8.32859993e-01 6.19130909e-01
6.78438187e-01 5.38983848e-03 1.50955886e-01 1.51698494e+00
6.03382945e-01 2.94420749e-01 4.63546693e-01 -9.50154066e-01
-4.54390526e-01 4.07225937e-01 -8.14161599e-01 -2.89108735e-02
-1.80376753e-01 -1.19074459e-04 1.10019875e+00 -7.34744012e-01
4.44670439e-01 1.23675454e+00 4.69080657e-01 5.80208480e-01
-1.34780991e+00 -7.09280372e-01 6.90873623e-01 9.39079374e-02
-1.58826160e+00 -4.96892110e-02 1.04421401e+00 -4.41913426e-01
1.16941839e-01 1.72403887e-01 1.09969072e-01 9.29739892e-01
-3.08572024e-01 1.04804647e+00 1.33846784e+00 -3.39894563e-01
-1.87464766e-02 3.30997437e-01 4.48981524e-01 8.11465085e-01
1.41731184e-02 -3.05746019e-01 -2.51054019e-02 -1.04774892e-01
3.93743187e-01 9.87190381e-03 -4.01312262e-01 -3.73696417e-01
-1.01125705e+00 6.11652613e-01 2.76125818e-01 2.75134265e-01
-3.95590439e-02 -2.11006939e-01 3.03688020e-01 1.56096786e-01
6.24697804e-01 3.61597002e-01 -6.75586164e-01 4.66041714e-01
-8.74951661e-01 -2.33316299e-04 3.38440508e-01 9.98996019e-01
1.07824230e+00 -1.51621491e-01 -4.79320258e-01 1.24781573e+00
3.48638117e-01 2.62188882e-01 7.02398658e-01 -7.30055869e-01
4.02934581e-01 6.56053543e-01 5.19070588e-02 -9.52076435e-01
-5.40784121e-01 -9.80329394e-01 -8.19568753e-01 -4.97146025e-02
3.83911878e-01 -1.36052907e-01 -1.01651132e+00 1.82127488e+00
5.21262109e-01 4.29113567e-01 -8.33671838e-02 1.04571545e+00
9.28888261e-01 3.71643573e-01 3.55297118e-01 -2.80535996e-01
1.44498158e+00 -1.17148614e+00 -7.69543707e-01 -4.09425616e-01
7.09995508e-01 -8.97814095e-01 1.13212323e+00 5.65652192e-01
-5.72191536e-01 -7.48617351e-01 -9.03487206e-01 8.96032602e-02
-9.45829898e-02 6.74782455e-01 4.22560930e-01 4.16954398e-01
-7.15138614e-01 4.72785354e-01 -4.23934340e-01 4.26734872e-02
3.20574909e-01 4.54745322e-01 -4.55799848e-02 -2.75859654e-01
-1.19013226e+00 5.72537422e-01 6.26276433e-01 2.02663898e-01
-9.09006119e-01 -2.79470623e-01 -8.11815143e-01 9.31592062e-02
7.74978340e-01 -3.20533156e-01 1.22851527e+00 -1.19200397e+00
-1.43416440e+00 9.30373251e-01 -6.42223805e-02 7.94089809e-02
4.93734390e-01 2.68823840e-02 -3.61938298e-01 -2.22262926e-02
3.19517046e-01 6.97837889e-01 8.55356872e-01 -1.45612967e+00
-7.77523875e-01 -3.45353663e-01 1.57527566e-01 3.73012990e-01
-3.70017111e-01 -1.36730969e-01 -6.13245845e-01 -8.00887465e-01
4.42502648e-01 -8.40610385e-01 -4.04234499e-01 -4.56478834e-01
-4.97283518e-01 -5.14206588e-01 5.96091509e-01 -4.10109758e-01
1.13436759e+00 -2.17794561e+00 1.15870900e-01 1.33211955e-01
1.81669846e-01 3.03843677e-01 -2.42742971e-01 -3.09854209e-01
2.42556501e-02 5.74409254e-02 -1.50760338e-01 -7.12161183e-01
3.10536120e-02 3.33002061e-01 -9.46436599e-02 4.75644171e-01
2.74562091e-01 4.22949702e-01 -1.00434065e+00 -7.59643674e-01
5.10287210e-02 2.54558325e-01 -3.43013346e-01 1.44292966e-01
-4.29736674e-01 8.34956050e-01 -7.44269967e-01 6.97299719e-01
8.83959889e-01 -6.14094317e-01 1.92140654e-01 -1.96443170e-01
1.68711707e-01 3.64827104e-02 -1.53142107e+00 1.81476390e+00
-4.33304161e-01 2.90610809e-02 1.44337103e-01 -1.17441809e+00
9.33768570e-01 4.09340799e-01 3.96097660e-01 -4.58217800e-01
3.51052195e-01 3.33171397e-01 -3.38453710e-01 -4.34405059e-01
3.50383043e-01 -1.54188514e-01 -1.26761019e-01 1.68388829e-01
2.52460510e-01 1.40711576e-01 2.64703453e-01 -1.51954576e-01
6.09905243e-01 6.14645332e-02 1.52205722e-02 -2.72138119e-01
7.67062008e-01 -3.29939127e-02 1.24167800e+00 7.47950673e-01
-4.12992686e-01 6.43844724e-01 1.87736541e-01 -2.40094751e-01
-3.85805398e-01 -5.14445782e-01 -3.67939115e-01 1.18156362e+00
5.42883396e-01 -1.88773721e-01 -6.48769021e-01 -1.10886919e+00
-2.76400864e-01 3.87222499e-01 -1.91108868e-01 2.05634404e-02
-4.53572989e-01 -1.09760642e+00 9.40239802e-02 1.95823684e-01
4.60717142e-01 -8.33066046e-01 -2.71654833e-04 2.54737288e-01
-3.49775076e-01 -1.07740951e+00 -6.73530519e-01 4.11210954e-01
-6.21709287e-01 -1.09345222e+00 -6.31719947e-01 -1.25839031e+00
9.96351540e-01 4.11965221e-01 9.48366106e-01 3.52622718e-01
-6.23374768e-02 -5.63044026e-02 -2.82876223e-01 -1.23366058e-01
-1.76437974e-01 1.25774443e-01 -1.01243779e-01 3.77074003e-01
3.60360235e-01 -3.17154229e-01 -4.28269774e-01 4.96346325e-01
-1.00357628e+00 1.66979700e-01 7.12463081e-01 1.14870071e+00
1.06030738e+00 5.11547625e-01 7.15527713e-01 -1.02710569e+00
3.36245418e-01 -4.25094903e-01 -7.68668115e-01 3.93969566e-01
-6.66887105e-01 1.35709122e-01 6.15422785e-01 -7.64251947e-01
-1.20985520e+00 5.16574621e-01 -6.99954629e-02 -4.56559449e-01
-3.45450997e-01 5.01822472e-01 -5.37292540e-01 -2.42137164e-01
2.41004869e-01 3.93287614e-02 -4.38128978e-01 -8.99745286e-01
2.21583679e-01 6.31597102e-01 3.43571872e-01 -9.08586979e-01
3.53217274e-01 2.00062320e-01 -9.72203538e-02 -2.24130690e-01
-1.63010609e+00 -7.57616222e-01 -6.88956678e-01 -1.54527247e-01
6.67479038e-01 -1.11869502e+00 -5.28611541e-01 4.55906421e-01
-9.96017992e-01 -1.16443299e-01 7.78926387e-02 4.71632957e-01
-2.19136566e-01 5.10636747e-01 -5.27735054e-01 -5.76021016e-01
-7.89935663e-02 -1.63405275e+00 1.26828063e+00 5.06564677e-01
1.37836888e-01 -7.53273845e-01 -1.61690131e-01 6.78831756e-01
-6.08624481e-02 6.09874167e-02 7.62179673e-01 -7.84574330e-01
-5.97139001e-01 -5.52964248e-02 -4.06046659e-01 5.51415920e-01
1.19274601e-01 -4.68667626e-01 -1.04719281e+00 -4.05034840e-01
1.32828265e-01 -6.24091387e-01 1.06944048e+00 2.18758062e-01
1.43723989e+00 -5.29848747e-02 -3.97529542e-01 3.74972075e-01
1.51048529e+00 4.24666218e-02 -6.37113452e-02 2.81824917e-01
8.81334603e-01 4.89719033e-01 8.29991758e-01 2.76051909e-01
4.27196681e-01 6.54008865e-01 3.83989722e-01 -6.57488257e-02
-1.39331162e-01 6.57700077e-02 1.33819133e-02 9.56438661e-01
4.97875780e-01 -2.60403395e-01 -7.80300856e-01 5.83875299e-01
-1.96574295e+00 -3.84801120e-01 -1.38719007e-01 2.04005718e+00
1.19243884e+00 2.11227447e-01 -3.64191443e-01 8.06705430e-02
1.07490230e+00 5.41632846e-02 -6.05079710e-01 2.79456317e-01
-1.80137798e-01 4.97522298e-03 4.05163676e-01 4.59535837e-01
-1.62292182e+00 8.58719707e-01 5.05228519e+00 1.34128559e+00
-1.00560772e+00 3.90916407e-01 8.72959018e-01 1.88797396e-02
-1.03246816e-01 -1.26560610e-02 -1.19712448e+00 5.32763958e-01
3.88685882e-01 3.56341839e-01 6.11378103e-02 8.77929926e-01
-4.46769409e-02 -1.58169091e-01 -8.49090874e-01 1.03389966e+00
9.71947610e-02 -6.30241454e-01 -6.41261190e-02 -3.03052843e-01
1.12079847e+00 -2.49277428e-01 5.37180491e-02 4.38109905e-01
4.28168505e-01 -6.15025699e-01 7.36599207e-01 4.70315427e-01
4.66861904e-01 -7.30575442e-01 7.21361697e-01 5.24118006e-01
-9.16191339e-01 -1.83818504e-01 -4.17452216e-01 1.50037482e-01
-1.20426826e-01 9.53961492e-01 -4.85879838e-01 6.75643384e-01
3.73147637e-01 5.54876745e-01 -6.75854206e-01 1.17750192e+00
-4.06665325e-01 5.96703887e-01 -1.80118173e-01 3.04707795e-01
3.91664386e-01 -2.12075546e-01 3.45279276e-01 9.46331859e-01
2.31801003e-01 -3.74629647e-02 1.00516307e+00 6.10545933e-01
-1.53399378e-01 2.55167186e-01 -4.89537790e-03 4.19363886e-01
4.12727803e-01 1.55988669e+00 -8.85643721e-01 -2.47097492e-01
-4.90427434e-01 8.82627904e-01 5.02795696e-01 3.11534256e-01
-6.34362876e-01 -1.15750469e-01 2.77317405e-01 -2.99003750e-01
1.55922771e-01 1.50759608e-01 -1.92557365e-01 -1.29580259e+00
3.46758291e-02 -7.71145999e-01 5.63012302e-01 -4.72908378e-01
-1.45988905e+00 5.53625286e-01 -2.68495739e-01 -1.31769061e+00
1.73006341e-01 -3.65273327e-01 -1.94462374e-01 8.81948352e-01
-1.85136664e+00 -1.12560439e+00 -2.20500350e-01 3.03355128e-01
5.98136246e-01 5.24778962e-02 6.24741912e-01 7.49113381e-01
-1.10849369e+00 6.25854909e-01 4.07730550e-01 -7.16225291e-03
1.02067816e+00 -1.29903579e+00 -2.92363554e-01 7.76167035e-01
-1.10815903e-02 2.75968850e-01 5.42881787e-01 -5.61860681e-01
-6.77411854e-01 -1.41127145e+00 7.79629469e-01 -6.26685247e-02
4.84372348e-01 -2.41642550e-01 -1.07837665e+00 5.76729655e-01
-4.05300520e-02 2.35662103e-01 6.62487209e-01 1.33659258e-01
-1.92173898e-01 -2.34272912e-01 -1.04668808e+00 2.84520447e-01
7.02113748e-01 -3.07747662e-01 -3.33383441e-01 6.96439743e-01
7.52417684e-01 -6.36973083e-01 -7.41471648e-01 5.41736186e-01
-3.30466628e-02 -4.07226503e-01 8.02579224e-01 -2.49623999e-01
2.72881389e-02 -6.38004184e-01 6.42261095e-03 -1.36010599e+00
-3.69272470e-01 -2.61199027e-01 1.83754370e-01 1.68866658e+00
8.25846717e-02 -2.97040373e-01 6.79071546e-01 5.18461406e-01
-2.68659920e-01 -8.50613475e-01 -9.22372818e-01 -7.03227162e-01
-2.74430722e-01 -1.22126184e-01 5.40640295e-01 1.20896411e+00
-4.54886883e-01 4.66713250e-01 -4.85450000e-01 4.80171084e-01
7.14877963e-01 4.35384542e-01 2.27506399e-01 -1.54066002e+00
-4.19699341e-01 -4.54665720e-01 -1.47733718e-01 -1.23709905e+00
4.84853804e-01 -1.02191925e+00 4.32847142e-01 -1.17360425e+00
4.35619533e-01 -9.37339723e-01 -8.25755358e-01 6.09815896e-01
-6.74446940e-01 6.29396662e-02 2.63952240e-02 2.58449584e-01
-1.13487232e+00 6.05596840e-01 1.46115708e+00 -4.20228004e-01
8.07212014e-03 1.19269513e-01 -8.84911418e-01 8.13731790e-01
7.33378291e-01 -8.02266240e-01 -5.45969546e-01 -6.58672154e-01
1.03072330e-01 -8.54794830e-02 2.76721776e-01 -9.42218959e-01
1.65413067e-01 -2.72123247e-01 1.13032907e-01 -5.33109009e-01
1.12498641e-01 -8.70912433e-01 -1.97752826e-02 2.51069032e-02
-4.87732887e-01 -4.47968811e-01 4.46971767e-02 8.71388614e-01
-3.17973316e-01 -6.58231914e-01 9.77846444e-01 -1.84743375e-01
-8.34172666e-01 3.17231745e-01 -4.89888340e-02 1.70942679e-01
8.72096360e-01 3.56004000e-01 -2.30178937e-01 1.14127241e-01
-8.56587529e-01 7.31657922e-01 2.97978461e-01 2.97899276e-01
2.98713893e-01 -1.51168287e+00 -5.24570584e-01 -1.31654041e-03
8.45537707e-02 4.78809923e-01 5.17094076e-01 7.12578475e-01
1.15010329e-01 1.72140554e-01 1.77373528e-01 -5.91390491e-01
-1.33770335e+00 6.85147822e-01 3.00297379e-01 -6.75108671e-01
-2.72813290e-01 9.99443889e-01 4.75417882e-01 -5.51563084e-01
5.60768008e-01 -1.71827599e-01 -5.35956264e-01 2.10188910e-01
4.31777179e-01 7.08595589e-02 1.63039759e-01 -7.70929754e-01
-1.98436946e-01 6.60115540e-01 -2.50840575e-01 4.12967205e-01
1.03622139e+00 -5.25025249e-01 -1.42661363e-01 4.29287583e-01
1.14708245e+00 -2.54794806e-01 -1.33640981e+00 -7.08453000e-01
2.98208565e-01 -2.30587900e-01 2.85907030e-01 -8.28841567e-01
-1.41872048e+00 6.73621893e-01 7.31328130e-01 -8.26431587e-02
1.32258737e+00 -1.18284024e-01 8.38996887e-01 3.46712589e-01
3.19720209e-01 -1.31949294e+00 1.23661011e-01 4.01246756e-01
3.80853862e-01 -1.54105270e+00 -2.59412229e-01 -5.62271416e-01
-4.81565595e-01 8.82017732e-01 8.66661966e-01 1.72412843e-01
6.10636055e-01 -1.73241124e-01 2.32905746e-01 -1.62445471e-01
-4.70243812e-01 -4.39460158e-01 3.68477523e-01 1.49879023e-01
1.70842990e-01 5.42692021e-02 -4.93867785e-01 1.03466606e+00
4.72820550e-01 -1.94174901e-01 3.12438220e-01 6.79940104e-01
-6.40159667e-01 -1.31180966e+00 -5.11708081e-01 3.72265339e-01
-6.32836103e-01 -6.12767823e-02 5.03431894e-02 4.35206443e-01
6.62300467e-01 1.07767045e+00 -1.96247920e-01 -2.02717155e-01
2.63175339e-01 1.19235583e-01 2.75542557e-01 -7.52853394e-01
-5.33040047e-01 3.99415314e-01 -1.43011853e-01 -1.82122469e-01
-6.18513405e-01 -6.27470613e-01 -1.19454396e+00 3.68947089e-01
-8.26680183e-01 1.93584815e-01 3.45213652e-01 1.11647403e+00
2.88407564e-01 7.54416525e-01 5.70373952e-01 -5.81885457e-01
-7.24745810e-01 -9.65157807e-01 -8.20151746e-01 6.46272361e-01
2.63489842e-01 -1.03928030e+00 -2.19440833e-01 -5.82341626e-02] | [9.537548065185547, 3.9572956562042236] |
ed4c8165-babb-4547-b311-df1667fb0afa | signaltrain-profiling-audio-compressors-with | 1905.11928 | null | https://arxiv.org/abs/1905.11928v2 | https://arxiv.org/pdf/1905.11928v2.pdf | SignalTrain: Profiling Audio Compressors with Deep Neural Networks | In this work we present a data-driven approach for predicting the behavior of (i.e., profiling) a given non-linear audio signal processing effect (henceforth "audio effect"). Our objective is to learn a mapping function that maps the unprocessed audio to the processed by the audio effect to be profiled, using time-domain samples. To that aim, we employ a deep auto-encoder model that is conditioned on both time-domain samples and the control parameters of the target audio effect. As a test-case study, we focus on the offline profiling of two dynamic range compression audio effects, one software-based and the other analog. Compressors were chosen because they are a widely used and important set of effects and because their parameterized nonlinear time-dependent nature makes them a challenging problem for a system aiming to profile "general" audio effects. Results from our experimental procedure show that the primary functional and auditory characteristics of the compressors can be captured, however there is still sufficient audible noise to merit further investigation before such methods are applied to real-world audio processing workflows. | ['Stylianos I. Mimilakis', 'Benjamin Colburn', 'Scott H. Hawley'] | 2019-05-28 | null | null | null | null | ['audio-signal-processing', 'audio-effects-modeling'] | ['audio', 'audio'] | [ 4.16773349e-01 -3.47694337e-01 3.80511463e-01 -2.35719055e-01
-7.57585645e-01 -6.02362692e-01 3.51614535e-01 2.98697144e-01
1.88182238e-02 1.89862818e-01 2.47454748e-01 -3.99073437e-02
-5.07887363e-01 -2.09924772e-01 -6.89607918e-01 -5.57572305e-01
-5.77188015e-01 3.55036080e-01 2.31653601e-01 -2.19766408e-01
2.32599489e-02 6.91753685e-01 -2.19675612e+00 6.49433255e-01
3.27673882e-01 1.15596116e+00 4.08635080e-01 1.19194233e+00
5.16237378e-01 8.05576265e-01 -9.30844009e-01 1.14860974e-01
2.67211437e-01 -4.91182595e-01 -5.26562572e-01 -9.40942243e-02
1.31163880e-01 -1.73692971e-01 -3.66597235e-01 8.67989719e-01
8.18225026e-01 2.66750008e-01 6.40963256e-01 -8.89584124e-01
4.95639980e-01 8.20153713e-01 2.61138409e-01 2.49122590e-01
6.68553889e-01 3.96586895e-01 8.51485372e-01 -3.99807304e-01
5.02147615e-01 1.00573027e+00 6.90279365e-01 1.61369234e-01
-1.40220416e+00 -4.44917142e-01 -4.16853011e-01 2.93083608e-01
-1.20010591e+00 -7.61581421e-01 1.06618285e+00 -8.52010071e-01
7.10947514e-01 5.51777422e-01 6.96639597e-01 1.24777091e+00
1.92721039e-01 5.23302674e-01 1.08872879e+00 -3.50443453e-01
5.47937334e-01 1.65975377e-01 -2.14581937e-01 -6.72845766e-02
-6.15410805e-01 4.99267399e-01 -7.48128772e-01 -4.83201481e-02
2.70628929e-01 -6.09127998e-01 -6.73702359e-01 -1.08428426e-01
-7.17128456e-01 3.16223085e-01 6.13556243e-02 5.12394309e-01
-5.73832035e-01 3.73960704e-01 6.95249677e-01 7.79508471e-01
2.17327520e-01 6.74640119e-01 -5.35536587e-01 -7.59025693e-01
-1.33087337e+00 4.77331072e-01 9.16391253e-01 6.01588130e-01
3.08413446e-01 4.03560013e-01 -1.60254002e-01 7.30606616e-01
1.43556729e-01 2.24526301e-01 4.70718175e-01 -7.62390733e-01
1.92557901e-01 5.83436973e-02 9.75698233e-02 -7.92214274e-01
-5.68749368e-01 -5.60843289e-01 -4.03393477e-01 2.79187649e-01
5.43428838e-01 -1.20498657e-01 -4.08737391e-01 1.53424621e+00
7.37704411e-02 2.97132283e-01 -6.37205169e-02 1.10242999e+00
3.50814193e-01 8.67087424e-01 -1.41007036e-01 -2.57907659e-01
1.22880781e+00 -3.15694124e-01 -7.36967444e-01 -9.13837627e-02
2.76822418e-01 -1.01365471e+00 1.47717798e+00 1.00749660e+00
-1.34860110e+00 -8.51833224e-01 -1.28033864e+00 2.34663293e-01
-8.93084630e-02 1.84437200e-01 8.93666893e-02 6.13838255e-01
-9.06233490e-01 9.50288892e-01 -6.49810255e-01 1.76459476e-01
-1.26441911e-01 3.24455678e-01 2.22435575e-02 3.82661849e-01
-1.35715485e+00 6.28108382e-01 1.18764803e-01 1.19818384e-02
-1.34110999e+00 -1.38687801e+00 -3.41296643e-01 5.37662864e-01
1.96343333e-01 -1.66661620e-01 1.70271337e+00 -9.92835522e-01
-1.93408883e+00 4.80905801e-01 3.51986885e-01 -5.57026744e-01
8.04557621e-01 -2.84681231e-01 -6.81243241e-01 2.50772983e-01
-5.15668929e-01 -2.42604502e-02 1.23545432e+00 -1.22243559e+00
-4.27526236e-01 -3.57069029e-03 -1.58812121e-01 -8.12780038e-02
-3.18633497e-01 3.32512289e-01 -2.51648009e-01 -6.88345373e-01
-3.88701856e-01 -6.92829072e-01 1.19009376e-01 -3.79814565e-01
-3.79831940e-01 4.09756869e-01 6.91723466e-01 -9.26106155e-01
1.53964365e+00 -2.41727328e+00 2.70542413e-01 2.38277093e-01
-1.35856658e-01 2.63327271e-01 1.02914341e-01 8.59423220e-01
-4.49944228e-01 -2.10360453e-01 -1.44892052e-01 -3.32908213e-01
1.92406058e-01 -4.54286039e-01 -6.11750126e-01 4.32194531e-01
1.01419322e-01 1.61871538e-01 -7.72519648e-01 1.02741666e-01
1.73887998e-01 5.78196526e-01 -5.68845689e-01 7.81853795e-01
-4.62581009e-01 7.71957338e-01 -5.20606562e-02 2.84982592e-01
3.35928172e-01 5.54505229e-01 -3.80132124e-02 -6.16589844e-01
-2.17715859e-01 3.33491951e-01 -1.29818285e+00 1.48707664e+00
-9.28882897e-01 8.65964890e-01 4.58110839e-01 -7.26803660e-01
1.15831470e+00 4.82611388e-01 6.83124661e-01 -6.52182281e-01
2.94095188e-01 3.99939060e-01 4.03435796e-01 -9.65989947e-01
4.94378537e-01 -3.38692427e-01 4.94639091e-02 3.43537390e-01
1.87204763e-01 -6.34378731e-01 1.68870613e-01 -2.90838510e-01
1.25367928e+00 7.40036964e-02 1.21696226e-01 -4.29779142e-01
5.92495918e-01 -4.05820906e-01 1.54993758e-01 3.22161466e-01
5.67421280e-02 5.74739814e-01 7.80535400e-01 -3.79649922e-02
-1.15185845e+00 -7.39979327e-01 -5.67915626e-02 1.11300576e+00
-4.70842689e-01 -5.33172905e-01 -1.02687573e+00 1.44533738e-01
-1.58403635e-01 9.20411229e-01 -2.53605336e-01 -5.68332553e-01
-6.16946757e-01 -5.74166775e-02 8.40186536e-01 7.97673464e-02
-2.55083561e-01 -1.09139657e+00 -1.01368582e+00 5.56060195e-01
7.35140964e-02 -9.73114848e-01 -3.30837101e-01 5.99973023e-01
-6.17079318e-01 -8.04100335e-01 -3.36252898e-01 -1.77179486e-01
-2.38970473e-01 -6.58540249e-01 9.98701751e-01 -4.72050339e-01
-4.07113224e-01 6.10907555e-01 -3.66024166e-01 -6.81277275e-01
-1.02078998e+00 -1.39493495e-01 2.26385757e-01 5.85855484e-01
-2.11145133e-01 -8.97006869e-01 -4.37077194e-01 2.58214325e-01
-1.21323907e+00 -3.20707560e-01 4.42425519e-01 3.13354790e-01
3.79971355e-01 5.72847664e-01 3.89613628e-01 -5.62734008e-01
1.17192626e+00 -3.86291891e-01 -7.23725319e-01 -8.53548422e-02
-1.80526093e-01 -6.70958087e-02 1.21014082e+00 -9.39791262e-01
-8.30046237e-01 2.28002459e-01 -2.60024339e-01 -8.77521336e-01
-3.28591347e-01 6.85993969e-01 -5.88182449e-01 2.71865755e-01
8.49827468e-01 4.74871472e-02 -4.72792648e-02 -7.86640286e-01
-8.26654658e-02 9.00145948e-01 7.45235085e-01 -6.16552055e-01
6.00478828e-01 1.01033792e-01 2.15569466e-01 -1.06278539e+00
-2.57005632e-01 -4.41315144e-01 -3.13776731e-01 -6.30688787e-01
3.90937507e-01 -5.58394253e-01 -8.26879561e-01 5.00581801e-01
-1.13965917e+00 -8.17012370e-01 -6.06768727e-01 5.73680222e-01
-1.08577454e+00 -3.04891285e-03 -5.33377945e-01 -1.16817069e+00
-1.99571460e-01 -1.26536667e+00 1.08657432e+00 -2.43107766e-01
-3.81282836e-01 -7.54158080e-01 1.91564828e-01 -1.02154508e-01
5.31656563e-01 2.65138805e-01 9.10128295e-01 -5.79721928e-01
-2.71419168e-01 -4.08902794e-01 5.69074929e-01 6.37165546e-01
-2.65244931e-01 3.33784521e-01 -1.54225624e+00 -1.48501441e-01
6.62539482e-01 -7.53873438e-02 3.01157534e-01 4.72294897e-01
1.17547059e+00 -1.57872275e-01 2.73717672e-01 4.77297574e-01
1.16863990e+00 2.22745076e-01 6.05439305e-01 -1.40462235e-01
1.68955699e-01 8.79273415e-01 7.66716957e-01 6.60003245e-01
-4.17206764e-01 1.12180340e+00 4.41116005e-01 6.95193633e-02
-2.64131278e-01 -3.01406860e-01 6.60380960e-01 9.88943934e-01
1.64652348e-01 -2.46655583e-01 -9.30967987e-01 3.32797855e-01
-1.24902987e+00 -7.85947680e-01 -3.05756599e-01 2.43424106e+00
9.03070450e-01 2.35966206e-01 2.87749976e-01 8.25202882e-01
6.57799661e-01 -1.23912342e-01 -4.22012836e-01 -7.97256470e-01
2.38749176e-01 5.58341086e-01 1.18077934e-01 4.47365433e-01
-8.97583127e-01 3.30895424e-01 5.64611197e+00 8.63210559e-01
-1.47766662e+00 -7.33305737e-02 2.37162799e-01 -2.30398566e-01
-5.17177992e-02 -2.16448590e-01 -3.64285141e-01 7.04299510e-01
1.87691951e+00 -2.17405424e-01 8.36892664e-01 6.26456261e-01
8.73542905e-01 1.77956551e-01 -1.56624031e+00 8.74356806e-01
-3.36542815e-01 -7.73739874e-01 -3.07665646e-01 -1.41197413e-01
1.57058790e-01 -4.32295173e-01 2.80246526e-01 2.71739334e-01
-4.59360451e-01 -1.03021872e+00 1.28286552e+00 7.20027089e-01
1.07370818e+00 -6.13298535e-01 3.83257836e-01 4.30314690e-01
-1.09678328e+00 -3.71866286e-01 -2.84852497e-02 -1.40851930e-01
4.34833288e-01 9.49545681e-01 -9.27900016e-01 3.71986806e-01
4.64746952e-01 3.13121140e-01 -2.34494820e-01 1.29212081e+00
-1.05068482e-01 1.19022918e+00 -2.72705138e-01 6.23001941e-02
-9.39569473e-02 1.85345128e-01 8.80932629e-01 1.45827782e+00
5.53655446e-01 -2.41730079e-01 -2.24930614e-01 8.37324142e-01
8.98596048e-02 1.26731500e-01 -4.02117074e-01 -2.48102769e-01
2.93742597e-01 1.09786522e+00 -2.66194344e-01 -2.41654180e-02
1.22344121e-01 4.35801685e-01 -4.86189216e-01 2.62727231e-01
-8.64829183e-01 -4.29119587e-01 7.38584220e-01 5.58537781e-01
2.00331554e-01 1.71827488e-02 3.08321603e-02 -7.00636744e-01
-1.04599791e-02 -1.22132981e+00 -5.79920672e-02 -9.62732017e-01
-7.73735881e-01 6.13024592e-01 1.34216592e-01 -1.75403845e+00
-7.20527768e-01 -6.33170485e-01 -5.08403718e-01 9.37816441e-01
-1.15044296e+00 -5.65149486e-01 -1.66819260e-01 5.46009004e-01
5.65302014e-01 -7.79194012e-02 8.17315102e-01 5.06361127e-01
-7.05920486e-03 2.44535074e-01 1.34830356e-01 -5.30150115e-01
6.04146659e-01 -1.19180274e+00 -1.86198503e-02 6.82935596e-01
-3.58270369e-02 2.14487448e-01 1.40580916e+00 -3.12133282e-01
-1.60333824e+00 -1.01315713e+00 5.54934204e-01 -7.71260709e-02
1.02077651e+00 -7.46383131e-01 -9.71132398e-01 1.31196216e-01
-7.45026916e-02 -2.08050147e-01 5.20185292e-01 -2.75757998e-01
-5.57511859e-02 -4.77383703e-01 -1.01080358e+00 2.70612419e-01
5.28953075e-01 -7.61626303e-01 -3.39777946e-01 1.33611232e-01
5.59007943e-01 -6.19846702e-01 -1.03463745e+00 2.36920342e-01
4.24965084e-01 -1.13297546e+00 8.31739545e-01 -2.82138973e-01
7.19611466e-01 -2.97325283e-01 -1.35477573e-01 -1.61474121e+00
1.10014372e-01 -1.13816857e+00 -4.16367114e-01 1.33728302e+00
7.69666359e-02 -1.82728902e-01 3.34025443e-01 3.30746263e-01
-2.78896153e-01 -5.23580849e-01 -8.19533885e-01 -8.66563618e-01
-2.57317096e-01 -1.02997386e+00 5.76853633e-01 4.61854249e-01
5.72767928e-02 2.47285843e-01 -3.65259171e-01 3.49649101e-01
1.89122871e-01 -7.81576484e-02 6.82719827e-01 -1.05656719e+00
-8.39013755e-01 -5.87609828e-01 -6.39840186e-01 -7.30748534e-01
-1.36578038e-01 -5.14215827e-01 3.01496297e-01 -5.53220153e-01
-4.72686201e-01 -3.07984203e-01 -2.76466399e-01 -3.54713231e-01
2.52637625e-01 -1.56187609e-01 2.94798583e-01 -1.15493521e-01
1.99874789e-01 3.27193022e-01 8.75545621e-01 -2.25525588e-01
-4.47448075e-01 6.26847863e-01 -9.90539640e-02 4.72060889e-01
7.75854766e-01 -4.83037055e-01 -7.15213776e-01 -9.59256217e-02
2.18387440e-01 6.34283006e-01 2.05584407e-01 -1.50004232e+00
3.16775799e-01 1.26629695e-01 -2.51149446e-01 -2.96874374e-01
5.64898074e-01 -1.04737449e+00 6.30821228e-01 3.92561585e-01
-7.16008067e-01 -2.84516871e-01 4.30420846e-01 2.81958550e-01
-5.46877742e-01 -5.46811044e-01 7.78158963e-01 2.91227341e-01
-4.16442603e-01 4.89137135e-02 -5.58124900e-01 -6.90231547e-02
5.79520762e-01 8.70821327e-02 2.68375307e-01 -7.96628654e-01
-8.69320035e-01 -4.20736372e-01 8.84281769e-02 1.65197983e-01
4.35994416e-01 -9.17090535e-01 -8.29449594e-01 3.49462718e-01
-2.53230073e-02 -2.92192459e-01 4.60568368e-01 7.91911125e-01
-4.00250584e-01 3.06962520e-01 -1.65248066e-01 -6.29330635e-01
-1.32398689e+00 6.88033938e-01 5.69120765e-01 -2.56519318e-01
-4.58916903e-01 6.14320338e-01 -2.74879001e-02 -3.61671560e-02
3.86568397e-01 -5.94870865e-01 -2.37705231e-01 2.02615008e-01
5.18511891e-01 5.04111409e-01 6.28705978e-01 -3.18845123e-01
-1.18709374e-02 2.65496671e-01 4.85823870e-01 -3.68245602e-01
1.37650144e+00 2.75441974e-01 1.21108003e-01 9.93934333e-01
1.45986164e+00 1.47819534e-01 -1.33866560e+00 2.12040111e-01
-2.88493000e-02 -4.45494473e-01 2.77506530e-01 -8.13954890e-01
-9.07805026e-01 1.23326528e+00 7.61436999e-01 7.26053238e-01
1.67186570e+00 -2.32833415e-01 4.42314595e-01 8.43369868e-03
2.34064803e-01 -1.10297513e+00 6.41611367e-02 2.24795341e-01
1.25179565e+00 -9.00995582e-02 -2.07011715e-01 -1.84256330e-01
-3.50371987e-01 1.30294490e+00 1.11592226e-01 -2.15475447e-02
6.82369113e-01 8.97282183e-01 7.31949583e-02 -2.55102552e-02
-8.08919907e-01 7.54539967e-02 3.36881369e-01 6.24324143e-01
2.64580101e-01 6.45818561e-02 4.36104946e-02 9.83960629e-01
-6.64568365e-01 1.45261735e-01 6.02696538e-01 4.97733086e-01
-9.54259038e-02 -9.44414556e-01 -7.25033641e-01 3.52697372e-01
-5.64867437e-01 8.19787160e-02 -2.64578968e-01 5.51307857e-01
-1.84071939e-02 9.41644371e-01 -1.69951051e-01 -7.08839893e-01
9.31226611e-01 7.88583905e-02 4.08612937e-01 -3.22218418e-01
-1.02594006e+00 3.38722855e-01 1.42988294e-01 -6.63663089e-01
9.99952573e-03 -8.39692652e-01 -9.19095993e-01 -4.71907221e-02
-8.63761902e-02 -2.50415839e-02 1.16898370e+00 5.85418403e-01
7.79330507e-02 1.11089253e+00 8.92173827e-01 -1.19744146e+00
-8.83906782e-01 -1.04893839e+00 -8.15224290e-01 5.47172725e-01
4.99604166e-01 -4.50539201e-01 -6.26243889e-01 3.57085466e-01] | [15.516103744506836, 5.817506313323975] |
0820be32-22e0-476a-b14c-7b7e049b1499 | bone-marrow-cytomorphology-cell-detection | 2305.05430 | null | https://arxiv.org/abs/2305.05430v1 | https://arxiv.org/pdf/2305.05430v1.pdf | Bone Marrow Cytomorphology Cell Detection using InceptionResNetV2 | Critical clinical decision points in haematology are influenced by the requirement of bone marrow cytology for a haematological diagnosis. Bone marrow cytology, however, is restricted to reference facilities with expertise, and linked to inter-observer variability which requires a long time to process that could result in a delayed or inaccurate diagnosis, leaving an unmet need for cutting-edge supporting technologies. This paper presents a novel transfer learning model for Bone Marrow Cell Detection to provide a solution to all the difficulties faced for the task along with considerable accuracy. The proposed model achieved 96.19\% accuracy which can be used in the future for analysis of other medical images in this domain. | ['Khandaker Tabin Hasan', 'Raisa Fairooz Meem'] | 2023-05-09 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 1.06415063e-01 4.88631502e-02 3.72087769e-02 -1.39029920e-01
-1.03978217e+00 -2.59583145e-01 2.54559219e-01 7.69141614e-01
-5.93473494e-01 9.14849401e-01 -2.75457382e-01 -6.77216768e-01
-3.44298601e-01 -6.22947395e-01 -4.61824238e-03 -1.14511096e+00
6.90754205e-02 9.79310930e-01 1.78313553e-01 2.90585220e-01
1.88535213e-01 8.05535614e-01 -7.39916265e-01 4.07875478e-01
3.04558247e-01 6.90142810e-01 3.79078627e-01 7.82728493e-01
-3.17804247e-01 9.66296732e-01 -4.68229115e-01 -1.48465917e-01
6.37320131e-02 -3.87891799e-01 -6.60549879e-01 2.14820445e-01
-1.59823418e-01 -3.22441727e-01 6.58600628e-02 4.24109936e-01
9.54015374e-01 -1.33444473e-01 1.02242827e+00 -9.73210216e-01
-6.42331690e-02 6.58086166e-02 -6.45932019e-01 7.25873053e-01
7.94243291e-02 1.53226316e-01 2.51031339e-01 -7.50979483e-01
3.95824522e-01 5.80522835e-01 6.35679901e-01 4.58297908e-01
-9.15314913e-01 -5.74503124e-01 -6.38805568e-01 2.56907523e-01
-1.13967776e+00 -2.90317476e-01 5.19958377e-01 -6.61384165e-01
8.16157341e-01 1.91643815e-02 6.07799172e-01 5.51653266e-01
7.81842589e-01 3.34767640e-01 1.20489717e+00 -6.17361248e-01
1.80890843e-01 3.33365232e-01 4.24875319e-02 5.93938410e-01
3.81205112e-01 7.70368055e-02 -4.01450574e-01 -7.10500926e-02
8.44775081e-01 5.81378043e-02 -1.14183493e-01 -3.13543320e-01
-1.04963446e+00 9.16664243e-01 2.91563004e-01 6.41939342e-01
-4.02317017e-01 -2.16876492e-01 6.93086803e-01 4.30138856e-01
9.53322351e-02 1.27946645e-01 -2.02841610e-01 -5.83377779e-02
-7.80400515e-01 -3.43435854e-01 5.37392557e-01 3.37036580e-01
2.27899536e-01 -2.03488663e-01 2.08623603e-01 6.75043225e-01
2.05320716e-01 2.47561097e-01 5.21009803e-01 -4.85178381e-01
-1.43874243e-01 6.05125666e-01 -1.52380124e-01 -5.66777647e-01
-8.70413184e-01 -4.05415773e-01 -1.08854687e+00 2.83009201e-01
8.24947000e-01 -1.05086723e-02 -9.31767464e-01 9.42483246e-01
4.64980453e-01 -2.43072510e-02 -2.03554749e-01 9.01156723e-01
6.37414694e-01 2.79620111e-01 2.17029840e-01 -2.92364120e-01
1.50997090e+00 -5.07738948e-01 -6.91961825e-01 2.76849121e-01
7.26351738e-01 -8.99501026e-01 6.46648169e-01 5.87652862e-01
-9.42832053e-01 -2.02072412e-01 -8.98657918e-01 1.44185469e-01
-2.82422543e-01 -8.54414701e-02 6.79356694e-01 6.99007332e-01
-1.15883863e+00 3.16483945e-01 -9.24113512e-01 -8.11366260e-01
4.46761608e-01 8.25036585e-01 -4.91647393e-01 -7.28997663e-02
-8.26084852e-01 1.23429739e+00 3.60120982e-01 3.37494850e-01
-4.35918063e-01 -6.61984444e-01 -5.59922755e-01 -3.35423529e-01
1.16473615e-01 -7.87339866e-01 1.09719634e+00 -2.02288523e-01
-1.29606342e+00 1.12432694e+00 7.48179629e-02 -2.74255723e-01
7.02665448e-01 2.17008889e-01 -3.71642143e-01 2.86205202e-01
2.39419062e-02 4.60523129e-01 4.67109382e-01 -8.27772796e-01
-7.80882716e-01 -4.31918025e-01 -4.69064504e-01 8.59889835e-02
2.19141722e-01 5.74788377e-02 -2.42508277e-01 -4.24674571e-01
3.40896510e-02 -8.17427695e-01 -3.52108330e-01 -7.91720152e-02
2.46185943e-01 -2.92209834e-01 6.02033019e-01 -6.50977910e-01
9.06460404e-01 -1.85282540e+00 -3.67466420e-01 1.13095172e-01
2.64937609e-01 3.98355067e-01 4.25930530e-01 3.42483073e-01
-1.32929653e-01 1.02790017e-02 1.48892447e-01 6.71704113e-02
-3.57051015e-01 1.03917740e-01 3.80057722e-01 7.64307916e-01
2.65910983e-01 7.21480489e-01 -9.28415179e-01 -9.73001719e-01
7.63360322e-01 6.29373014e-01 6.94367886e-02 2.96389133e-01
5.09511471e-01 6.06567562e-01 -3.89266431e-01 8.54296148e-01
4.10268545e-01 -4.02387679e-01 1.11993998e-01 5.50974384e-02
2.13291287e-01 -1.12363435e-01 -9.29210067e-01 1.26413572e+00
-4.95086968e-01 5.26439548e-01 1.85463086e-01 -8.95266950e-01
8.40273917e-01 5.99703670e-01 7.22596407e-01 -7.14865208e-01
4.49569881e-01 3.95602018e-01 2.07623765e-01 -7.06220746e-01
-7.27654248e-02 -9.85692739e-01 3.22015375e-01 6.92735493e-01
-1.82106301e-01 -1.27817839e-01 3.75464782e-02 -9.60230827e-03
1.03507984e+00 -2.52932489e-01 4.50530857e-01 -2.53282100e-01
7.32028365e-01 -1.09566562e-01 5.00147581e-01 3.57100278e-01
-6.90463126e-01 5.15071511e-01 1.43594712e-01 -7.85491824e-01
-8.64792287e-01 -1.02044296e+00 -4.51162100e-01 5.76109052e-01
-3.09338808e-01 4.10955757e-01 -2.28998467e-01 -6.83732867e-01
6.61332458e-02 2.87585080e-01 -6.54359758e-01 5.10461405e-02
-4.97985214e-01 -7.77841806e-01 2.99068481e-01 6.11492932e-01
3.86780910e-02 -1.12657583e+00 -9.77420926e-01 5.17529845e-01
-5.21755125e-03 -8.56858552e-01 -3.85455005e-02 5.67228138e-01
-1.00662029e+00 -1.27889633e+00 -1.03471231e+00 -9.41592157e-01
6.44584835e-01 2.01759353e-01 9.55306232e-01 3.56530011e-01
-5.90244651e-01 1.51580125e-01 -2.56915212e-01 -5.62849700e-01
-5.16493022e-01 -5.84287159e-02 -1.50593713e-01 -1.83963314e-01
8.58679831e-01 -1.33846894e-01 -7.54939437e-01 1.93320468e-01
-8.63177001e-01 -1.12094328e-01 9.07062232e-01 1.10107589e+00
6.53629303e-01 9.47028622e-02 9.74465370e-01 -1.10437930e+00
4.36532587e-01 -4.92364943e-01 -2.02715158e-01 3.12851787e-01
-7.28933215e-01 -3.06856424e-01 4.34291393e-01 -1.62405130e-02
-1.05561185e+00 -2.23482311e-01 4.81678098e-02 2.71090623e-02
-3.08452874e-01 4.19389337e-01 3.71926397e-01 -2.17178583e-01
4.24451441e-01 3.02256811e-02 2.07346469e-01 -1.31983802e-01
-5.66625774e-01 7.83081889e-01 3.32544684e-01 -1.48168206e-01
4.03064489e-01 4.77137208e-01 5.45373261e-01 -8.73217106e-01
-2.51832753e-01 -8.26547921e-01 -7.26458967e-01 -2.32608944e-01
6.93578422e-01 -6.35684252e-01 -4.06011581e-01 4.19054896e-01
-7.41109550e-01 -1.94635302e-01 3.70872170e-02 7.23865092e-01
-4.23360825e-01 5.27656138e-01 -1.03124249e+00 -7.76724815e-01
-7.05002904e-01 -1.03593874e+00 7.48352468e-01 1.91333339e-01
-6.07311547e-01 -1.34699965e+00 2.99665749e-01 5.18496633e-01
5.38101792e-01 3.74950796e-01 1.01231110e+00 -6.17460787e-01
-2.39880770e-01 -5.93968868e-01 -3.22047979e-01 -3.58868949e-02
4.52414542e-01 -1.03751076e-02 -8.27167511e-01 -5.07510364e-01
2.82968611e-01 -3.37836742e-01 5.53598642e-01 5.36831975e-01
7.16620624e-01 5.79888582e-01 -3.24846447e-01 2.73724198e-01
1.54599607e+00 7.50647962e-01 3.43351662e-01 5.19227684e-01
2.59233177e-01 6.40053451e-01 7.29123414e-01 4.86639619e-01
1.68762222e-01 1.88559115e-01 7.92289600e-02 -4.06896025e-01
-2.06107665e-02 8.26804042e-02 -4.08631563e-01 7.03601062e-01
7.03338012e-02 -1.75493270e-01 -1.42432189e+00 7.34093845e-01
-1.47056842e+00 -6.25416994e-01 -2.87018359e-01 2.04277253e+00
4.87724066e-01 1.10659242e-01 1.20858051e-01 5.99167049e-01
7.10158288e-01 -5.60382247e-01 -3.92212182e-01 -5.64824939e-01
2.88801074e-01 3.76630813e-01 4.34987307e-01 4.08008039e-01
-6.86405420e-01 2.91247815e-01 7.00141764e+00 5.90796411e-01
-1.32635760e+00 -3.36534753e-02 7.63013780e-01 -7.68356472e-02
1.16565689e-01 -3.13321412e-01 -3.92311096e-01 5.88780046e-01
8.56603086e-01 -2.69079298e-01 -3.65383804e-01 2.59119213e-01
3.84056330e-01 -7.65224218e-01 -1.27533567e+00 9.25629497e-01
-9.15857255e-02 -9.90158200e-01 -1.60289541e-01 1.61371395e-01
2.69748628e-01 -3.99615586e-01 2.86282357e-02 2.17937097e-01
-9.76752788e-02 -1.04760396e+00 -4.23678458e-01 1.36438489e-01
8.94310534e-01 -9.00336266e-01 1.46256459e+00 3.21056634e-01
-7.94617534e-01 4.14282948e-01 -4.06493098e-01 -1.60441950e-01
3.16164978e-02 6.76035583e-01 -1.86201274e+00 3.64374161e-01
4.60001677e-01 1.41899779e-01 -3.17313373e-01 1.17853534e+00
1.54530153e-01 4.61457640e-01 -2.97676206e-01 -2.03880429e-01
1.92277148e-01 -4.56282087e-02 -9.46578234e-02 1.24339008e+00
3.19392830e-01 1.51275516e-01 -1.15885578e-01 -3.86846773e-02
3.37959349e-01 5.16951323e-01 -5.67163587e-01 6.62797242e-02
2.72217155e-01 1.32138157e+00 -1.46667695e+00 -2.26551667e-01
-4.60364163e-01 6.94395244e-01 3.08186144e-01 1.58961527e-02
-6.72310770e-01 -7.78867975e-02 -1.32920966e-01 4.13071185e-01
-6.90056458e-02 3.19341481e-01 -5.12458324e-01 -7.59177804e-01
-3.88844371e-01 -5.77909052e-01 1.00343335e+00 -4.70671028e-01
-1.23799360e+00 3.80710512e-01 -3.83412540e-01 -1.13887465e+00
-4.10778224e-01 -7.61284292e-01 -6.20315313e-01 8.36350739e-01
-1.67618966e+00 -9.50864911e-01 -6.66880086e-02 4.20406878e-01
4.06205803e-01 -2.02401638e-01 1.01093066e+00 3.42591316e-01
-4.65387732e-01 5.92781305e-01 -6.95082024e-02 1.47508144e-01
1.04111660e+00 -1.43428051e+00 -4.34626013e-01 2.85228997e-01
-4.71379399e-01 4.10432547e-01 5.28717399e-01 -3.77789766e-01
-1.19635892e+00 -7.40454376e-01 1.22195828e+00 -4.64310557e-01
6.00228608e-01 3.55889946e-01 -8.67006898e-01 3.63650203e-01
2.45027319e-01 -3.07575446e-02 1.61602509e+00 -4.10319380e-02
2.52286077e-01 -8.48855004e-02 -1.61706996e+00 2.58324504e-01
9.29787457e-02 -4.49044257e-01 -6.38317406e-01 5.40846363e-02
-1.68213472e-01 -3.52345347e-01 -9.50037122e-01 2.68994629e-01
6.29244626e-01 -9.50583518e-01 4.80017155e-01 -3.34898055e-01
2.34128609e-01 -2.03871384e-01 3.20488095e-01 -1.06691504e+00
-3.49824756e-01 -1.18478991e-01 3.09919745e-01 9.32884514e-01
4.03501540e-01 -5.54091036e-01 1.29439843e+00 6.84665859e-01
2.60467231e-01 -9.59310055e-01 -1.08123314e+00 -4.26939368e-01
4.19571698e-01 -2.91278809e-01 1.56295449e-01 9.93806303e-01
2.91039050e-01 2.76546955e-01 1.84912700e-02 -1.45866796e-01
6.36215806e-01 5.86528443e-02 4.12143409e-01 -1.08629000e+00
-1.66222572e-01 -5.04505634e-01 -6.23829246e-01 -4.87554520e-02
-4.09410387e-01 -7.16614127e-01 -2.67222017e-01 -1.63536060e+00
4.41265970e-01 -4.91129994e-01 -7.28523970e-01 3.82000841e-02
-2.11775944e-01 4.67089653e-01 -7.31549710e-02 -5.44156134e-02
-2.76545525e-01 -1.63041726e-01 1.14685738e+00 1.96083449e-03
-4.69930433e-02 1.08043902e-01 -5.14512479e-01 6.07218564e-01
1.10901952e+00 -4.41865474e-01 -3.69009942e-01 -1.48498714e-01
2.63912436e-02 5.34708142e-01 -6.08972088e-02 -9.55159307e-01
4.42853510e-01 -4.80055548e-02 7.49253809e-01 -8.52549076e-01
1.15386859e-01 -1.05217266e+00 3.72968882e-01 1.02452850e+00
6.20687716e-02 1.09281130e-01 1.43711478e-01 3.35390329e-01
-3.16181809e-01 -2.17550591e-01 1.00008678e+00 -3.40623140e-01
-3.83538753e-01 8.24757516e-02 -6.91050887e-01 -3.23588699e-01
1.46379173e+00 -6.87953830e-01 1.16812743e-01 -1.66967452e-01
-9.05942082e-01 4.45623368e-01 4.90352124e-01 -8.75787288e-02
6.18697882e-01 -1.04925036e+00 -9.66899574e-01 8.81235376e-02
7.18090385e-02 2.21273258e-01 5.10250151e-01 1.14823675e+00
-1.11874592e+00 6.73628390e-01 -4.39728945e-01 -7.09139287e-01
-1.34515381e+00 6.95983887e-01 3.30427647e-01 -7.52406240e-01
-4.68489617e-01 7.54528463e-01 -2.10339874e-01 6.39765486e-02
1.97658539e-01 -1.38237821e-02 -1.86989263e-01 3.80506009e-01
5.12564182e-01 5.63852191e-01 5.09229958e-01 -4.40942019e-01
-4.56693798e-01 6.07836366e-01 -4.40553457e-01 7.57699087e-02
1.13842773e+00 -1.49510100e-01 1.74008876e-01 6.19467676e-01
1.04993260e+00 -3.25702429e-01 -7.46594250e-01 -2.20817626e-01
1.34740412e-01 -4.99893069e-01 2.16935799e-01 -8.98520052e-01
-9.12765622e-01 1.14117658e+00 8.09365630e-01 -8.96086544e-02
1.04754627e+00 -1.76287562e-01 5.86947560e-01 2.08225742e-01
5.50031960e-01 -1.13993502e+00 -1.44709721e-01 -1.09107830e-01
4.31812465e-01 -1.65487921e+00 1.45302340e-01 -2.86471248e-01
-5.43612003e-01 1.36822903e+00 3.71038556e-01 8.25259686e-02
7.15758204e-01 5.84778309e-01 7.55530119e-01 -1.11724623e-01
-8.08819056e-01 -4.90827523e-02 -2.80488223e-01 8.96047890e-01
9.92060542e-01 1.39650181e-01 -6.13316536e-01 5.06968647e-02
2.55859524e-01 4.85643208e-01 4.87383515e-01 1.27044606e+00
-4.23953772e-01 -1.09978664e+00 -5.55218697e-01 6.72037899e-01
-9.12054718e-01 4.02550399e-01 -2.06390411e-01 1.01974022e+00
1.62978724e-01 9.86116350e-01 4.14469466e-02 1.11730956e-02
9.03781280e-02 2.62131125e-01 5.13267219e-01 -5.37867665e-01
-5.62025487e-01 4.52306151e-01 -2.39952743e-01 1.10096969e-01
-3.90104473e-01 -7.03031301e-01 -1.33575785e+00 -4.39063847e-01
-3.77191454e-01 2.76946783e-01 3.89959425e-01 7.57736802e-01
-1.57492027e-01 5.44686496e-01 6.10771656e-01 -3.34301054e-01
-3.52945179e-01 -1.05682135e+00 -8.64198148e-01 2.08653659e-01
3.66038144e-01 -5.51308692e-01 -4.32262212e-01 -1.02170203e-02] | [15.044724464416504, -3.011800765991211] |
ced2bb62-70c0-4ac2-a7b8-003d17b22ce4 | multimodal-image-outpainting-with-regularized | 1910.11481 | null | https://arxiv.org/abs/1910.11481v1 | https://arxiv.org/pdf/1910.11481v1.pdf | Multimodal Image Outpainting With Regularized Normalized Diversification | In this paper, we study the problem of generating a set ofrealistic and diverse backgrounds when given only a smallforeground region. We refer to this task as image outpaint-ing. The technical challenge of this task is to synthesize notonly plausible but also diverse image outputs. Traditionalgenerative adversarial networks suffer from mode collapse.While recent approaches propose to maximize orpreserve the pairwise distance between generated sampleswith respect to their latent distance, they do not explicitlyprevent the diverse samples of different conditional inputsfrom collapsing. Therefore, we propose a new regulariza-tion method to encourage diverse sampling in conditionalsynthesis. In addition, we propose a feature pyramid dis-criminator to improve the image quality. Our experimen-tal results show that our model can produce more diverseimages without sacrificing visual quality compared to state-of-the-arts approaches in both the CelebA face dataset and the Cityscape scene dataset. | ['Lingzhi Zhang', 'Jiancong Wang', 'Jianbo Shi'] | 2019-10-25 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 4.16971564e-01 1.54099971e-01 -2.86029056e-02 -2.79167980e-01
-8.93253803e-01 -5.13194919e-01 6.30283952e-01 -6.37310743e-01
-1.87756985e-01 1.17252898e+00 1.21606700e-01 1.44877851e-01
3.72149974e-01 -8.34790587e-01 -1.08060920e+00 -7.87847161e-01
4.79741454e-01 1.62796125e-01 -8.14839303e-02 -1.41905978e-01
-3.02310158e-02 6.95799351e-01 -1.54130089e+00 4.05764222e-01
1.30206144e+00 7.95362532e-01 -7.03956410e-02 6.51732147e-01
2.42317453e-01 8.81799281e-01 -9.38067198e-01 -6.68553650e-01
4.32154059e-01 -1.00378275e+00 -3.66460770e-01 5.45226276e-01
8.32709968e-01 -4.07543570e-01 -3.40912879e-01 1.31855989e+00
5.32026291e-01 -9.06575378e-03 8.64873409e-01 -1.50621963e+00
-8.18471074e-01 4.14917350e-01 -8.55726480e-01 -1.23415859e-02
2.80193537e-01 1.27715394e-01 5.00329733e-01 -9.53329146e-01
7.12374687e-01 1.23636293e+00 4.02732342e-01 7.96391308e-01
-1.67671669e+00 -9.73430872e-01 1.23664893e-01 -1.16326444e-01
-1.36615169e+00 -6.62163794e-01 8.31574023e-01 -3.48255485e-01
2.19325393e-01 3.86773944e-01 5.85705280e-01 1.32529700e+00
1.34375662e-01 8.27518344e-01 1.32674587e+00 -4.17748004e-01
2.41224334e-01 -1.08442381e-02 -6.71793878e-01 5.55572689e-01
3.58486444e-01 2.50059843e-01 -5.81381440e-01 1.75040588e-02
1.00097847e+00 6.89901831e-03 -3.51270527e-01 -2.30552137e-01
-9.61953044e-01 7.62942433e-01 4.61197346e-01 -1.32498788e-02
-4.19109523e-01 2.84855157e-01 -2.65942037e-01 2.73793638e-01
5.02853930e-01 3.81629676e-01 2.05667883e-01 3.86853546e-01
-1.20100713e+00 5.70152104e-01 5.31960130e-01 9.66370583e-01
7.79741406e-01 5.53911567e-01 -3.16666573e-01 7.06473053e-01
4.73241024e-02 8.49340975e-01 9.68422890e-02 -1.20235765e+00
4.95896876e-01 5.92516661e-02 2.78929591e-01 -1.07735217e+00
2.52246946e-01 -1.35179326e-01 -1.09048021e+00 7.16829419e-01
4.93614554e-01 -6.20352685e-01 -1.27621138e+00 1.82431614e+00
1.99540466e-01 4.95281756e-01 2.99195703e-02 8.49817395e-01
6.22896075e-01 7.38336444e-01 -1.25583336e-01 -2.49159679e-01
7.78829634e-01 -9.43124115e-01 -7.41397500e-01 -1.99779093e-01
-2.95158803e-01 -9.14173424e-01 1.00280654e+00 4.36405480e-01
-1.30872035e+00 -5.19823015e-01 -1.06814098e+00 1.05999872e-01
1.72452956e-01 -6.60701096e-03 4.44910526e-01 4.14963394e-01
-1.00545776e+00 6.04027629e-01 -7.14838028e-01 2.35859174e-02
6.93767309e-01 7.34460652e-02 -4.14617032e-01 -1.00448370e-01
-7.66254783e-01 7.57693887e-01 1.92848876e-01 -1.73525125e-01
-1.06346965e+00 -7.02151120e-01 -7.49650240e-01 -1.58196479e-01
2.64025241e-01 -6.83339119e-01 1.01292002e+00 -1.52405274e+00
-1.63473117e+00 7.00044572e-01 2.40029693e-02 -3.23677391e-01
8.77847135e-01 -2.28036627e-01 -2.50221729e-01 8.61603841e-02
1.71461418e-01 1.12844110e+00 1.17236817e+00 -1.66493213e+00
-4.36199158e-01 -1.11756667e-01 5.08013926e-02 3.52478176e-01
-1.31349459e-01 -1.03623308e-01 -4.87927765e-01 -1.15433037e+00
1.04517268e-03 -8.41398001e-01 -3.32653999e-01 3.32834244e-01
-9.03179824e-01 5.01884043e-01 7.80001760e-01 -5.17868221e-01
7.07508445e-01 -2.11022282e+00 4.15376246e-01 -3.96784544e-02
8.89571682e-02 1.15635589e-01 -2.14256853e-01 1.00068413e-01
-1.10518053e-01 1.12711355e-01 -4.67347622e-01 -3.98457736e-01
-2.07311541e-01 3.10197383e-01 -5.98252773e-01 4.35553819e-01
5.92069030e-01 8.03968906e-01 -8.51909578e-01 -5.44626713e-01
2.09524617e-01 5.85353792e-01 -7.41335511e-01 3.81437868e-01
-3.46566707e-01 7.02176034e-01 -3.01036805e-01 6.36394739e-01
9.38153446e-01 2.70238500e-02 -2.19105378e-01 1.71078797e-02
2.03222439e-01 -5.39072752e-01 -1.04552948e+00 1.57802415e+00
-2.99652666e-01 5.35325348e-01 7.10258037e-02 -5.75236976e-01
8.06031942e-01 6.84768930e-02 1.88709438e-01 -3.80024731e-01
1.51711702e-01 1.57499969e-01 -8.92610922e-02 -3.46721977e-01
4.92808223e-01 -4.74345624e-01 1.10567555e-01 3.07271797e-02
-4.66018654e-02 -5.35717010e-01 7.66116232e-02 5.62637709e-02
6.74041688e-01 3.52179617e-01 -1.73774268e-02 -1.82994276e-01
1.77525774e-01 -2.21254155e-01 6.53499603e-01 6.25233173e-01
-1.36972759e-02 1.27986646e+00 6.40931547e-01 -1.02991015e-01
-1.23230648e+00 -1.34719396e+00 7.36172423e-02 4.22349840e-01
2.99487442e-01 4.95864451e-03 -1.04579747e+00 -7.08504021e-01
-7.78973699e-02 7.16634929e-01 -6.61187589e-01 -1.11992791e-01
-4.72861111e-01 -6.39365971e-01 7.32342482e-01 4.09344852e-01
7.58966804e-01 -1.00711119e+00 -4.29400772e-01 -3.22971307e-02
-4.21545237e-01 -1.01601875e+00 -7.39133716e-01 -1.05321854e-01
-5.96034408e-01 -8.22734594e-01 -1.09700000e+00 -7.38648713e-01
9.54096079e-01 -2.03938019e-02 1.17025089e+00 -2.34560907e-01
-2.25323260e-01 -1.89851865e-01 -5.33672646e-02 -3.96926731e-01
-6.39854491e-01 -1.97897285e-01 -1.05919771e-01 2.70168453e-01
2.48105321e-02 -7.30195224e-01 -6.05237186e-01 2.18718007e-01
-1.11115146e+00 3.48097265e-01 5.25823474e-01 9.80225086e-01
8.26816559e-01 -1.70560926e-02 5.47309875e-01 -8.37361455e-01
4.26577836e-01 -3.07099462e-01 -6.42438054e-01 2.81087577e-01
-7.71267563e-02 1.39592499e-01 8.69937241e-01 -7.98840642e-01
-1.16283584e+00 2.06726789e-01 -1.13150589e-01 -7.54404306e-01
-1.34947836e-01 -1.38198465e-01 -5.13685405e-01 -3.47164646e-02
8.60470533e-01 9.03971791e-02 -9.59337037e-03 -3.32205854e-02
5.71562350e-01 2.47350901e-01 7.09064066e-01 -7.45993495e-01
1.06300175e+00 6.25256240e-01 9.72506255e-02 -7.11790621e-01
-7.31277168e-01 3.02286804e-01 -2.22683415e-01 -2.86486447e-01
8.78670156e-01 -9.95929718e-01 -3.45596880e-01 6.76442504e-01
-1.10297513e+00 -3.55615318e-01 -6.06019437e-01 1.78993344e-01
-7.60425627e-01 1.43806651e-01 -5.30339539e-01 -6.26016796e-01
-3.25636901e-02 -1.11100948e+00 1.10438859e+00 4.66645896e-01
1.12304725e-01 -5.84220767e-01 -6.63849339e-02 7.25840032e-02
3.06202173e-01 9.58728194e-01 4.61818874e-01 7.60079101e-02
-8.04323733e-01 2.19727196e-02 -1.40196085e-01 5.48752069e-01
1.86376840e-01 1.07030466e-01 -9.89713669e-01 -1.96219802e-01
-6.02016859e-02 -5.42315483e-01 9.38626170e-01 3.80421072e-01
1.22143209e+00 -4.86585110e-01 -4.21945937e-02 9.34497893e-01
1.50608170e+00 6.50392845e-02 1.02304602e+00 -8.12390745e-02
6.64897561e-01 5.35680532e-01 5.55877268e-01 4.08528537e-01
-1.11921012e-01 5.80473065e-01 4.34452802e-01 -4.02690351e-01
-4.24539447e-01 -5.71232736e-01 4.15233791e-01 2.04384595e-01
-1.20237693e-01 -5.64033866e-01 -5.08498192e-01 4.72541839e-01
-1.72509193e+00 -1.20185232e+00 1.27955064e-01 2.11771488e+00
1.14661717e+00 -1.45044876e-03 -4.41572852e-02 -9.70469043e-02
9.21366930e-01 2.35589504e-01 -6.74352944e-01 -3.54026817e-02
-3.92437845e-01 2.26449832e-01 4.12147492e-01 6.39001429e-01
-9.92319524e-01 1.00430143e+00 6.46489811e+00 8.88691664e-01
-1.19937730e+00 -4.26718220e-02 1.15656149e+00 -2.74182767e-01
-5.23978233e-01 -1.96778089e-01 -6.13185406e-01 6.35021031e-01
5.22666037e-01 -1.29857957e-01 5.06178916e-01 7.10940897e-01
-1.84059795e-02 -2.12051079e-01 -8.50667834e-01 8.75078738e-01
4.61614370e-01 -1.29761028e+00 2.13790268e-01 -1.65220663e-01
1.25536954e+00 -3.88316333e-01 4.06394601e-01 4.20445688e-02
6.55077815e-01 -1.27419674e+00 9.35689211e-01 7.46216536e-01
1.09273243e+00 -1.05696845e+00 2.98633486e-01 2.78458863e-01
-6.13252938e-01 2.22903952e-01 -3.07597399e-01 2.64039636e-01
2.10890502e-01 7.03204691e-01 -4.65092748e-01 4.65575844e-01
5.52039087e-01 5.70780337e-01 -4.50061858e-01 9.05605495e-01
-5.85931778e-01 4.13904458e-01 -1.53825358e-01 4.29856658e-01
-1.34328917e-01 -3.14533144e-01 4.97588634e-01 8.33231390e-01
6.16752684e-01 -1.91187486e-03 2.19839901e-01 1.21012020e+00
-3.65741670e-01 -2.58799881e-01 -8.15101266e-01 -1.46315899e-02
3.99823844e-01 9.10689414e-01 -6.64544463e-01 -2.73356974e-01
-1.24175370e-01 1.40516114e+00 1.52720183e-01 6.41663849e-01
-1.13585889e+00 -5.23347199e-01 7.16055453e-01 1.97635993e-01
3.20052743e-01 2.09513977e-02 -4.12240207e-01 -1.44590783e+00
4.58615273e-02 -9.76443887e-01 -6.82238564e-02 -9.82001483e-01
-1.30765450e+00 8.60728443e-01 6.09102100e-02 -1.18922198e+00
-2.93837070e-01 -1.47639602e-01 -5.49270809e-01 9.80744362e-01
-1.60433388e+00 -1.27729893e+00 -3.33488733e-01 6.76782608e-01
5.92287302e-01 -1.88393235e-01 5.83747745e-01 9.98553708e-02
-4.34817940e-01 5.87976277e-01 1.27191693e-01 3.62651162e-02
8.91537786e-01 -1.12492454e+00 3.51303190e-01 1.29566622e+00
1.97484586e-02 2.62782902e-01 1.01073086e+00 -7.68166661e-01
-8.74726057e-01 -1.33966196e+00 6.78362966e-01 -2.88396657e-01
1.87639743e-01 -4.60871488e-01 -6.33440852e-01 5.47782481e-01
5.22184134e-01 3.77280116e-01 2.50757754e-01 -5.35324275e-01
-3.40055466e-01 -6.42937198e-02 -1.44905853e+00 9.16565597e-01
9.87391710e-01 -2.12212086e-01 -2.89487302e-01 1.18837900e-01
7.19943523e-01 -5.03983140e-01 -4.61674988e-01 3.05227697e-01
2.94509381e-01 -1.22665632e+00 1.07950997e+00 -4.84523714e-01
9.25390244e-01 -4.09460753e-01 -2.07045421e-01 -1.43720484e+00
-5.85688762e-02 -8.04554045e-01 -1.11502245e-01 1.27289116e+00
2.31144518e-01 -2.81417936e-01 1.18376338e+00 4.08732980e-01
3.68366800e-02 -5.51350951e-01 -6.69298708e-01 -8.38605046e-01
3.97529155e-01 1.02876693e-01 6.19632542e-01 7.34466374e-01
-6.16928995e-01 2.09734112e-01 -7.92447269e-01 1.24356955e-01
8.87549281e-01 6.50325492e-02 8.98010790e-01 -7.97311127e-01
-4.95507777e-01 -1.50286943e-01 -3.22740495e-01 -8.56497109e-01
2.21264690e-01 -5.59470952e-01 3.53372037e-01 -1.17958796e+00
1.78584695e-01 -2.44729906e-01 7.47772902e-02 3.06755930e-01
-4.84473735e-01 6.36274636e-01 3.79803091e-01 5.44347018e-02
-2.91391909e-01 7.45056570e-01 1.73899126e+00 -1.82850346e-01
-5.25307022e-02 -1.61693588e-01 -8.30934227e-01 5.51960528e-01
8.04835975e-01 -6.08974338e-01 -7.80717850e-01 -5.21782994e-01
-7.30481818e-02 1.57750160e-01 6.13184631e-01 -1.12695420e+00
-1.31836355e-01 -6.33304358e-01 7.96119988e-01 -5.36885738e-01
6.04114652e-01 -5.82099080e-01 4.06946898e-01 3.15564990e-01
-2.40312412e-01 1.91457011e-02 1.06646523e-01 6.17807567e-01
-3.29615027e-01 7.07621202e-02 1.29744840e+00 -1.20006502e-01
-2.23007202e-01 4.04758424e-01 -2.77559578e-01 4.50963497e-01
1.27829885e+00 -9.46855694e-02 -3.04486990e-01 -6.47314668e-01
-4.37617958e-01 -8.86762440e-02 8.93867612e-01 2.68137008e-01
7.45135605e-01 -1.64260375e+00 -1.02219486e+00 4.37604487e-01
-1.65757477e-01 1.13913164e-01 1.62591457e-01 3.34322274e-01
-7.37493932e-01 -2.33837247e-01 -4.98502970e-01 -5.05955517e-01
-1.18732464e+00 4.41065431e-01 3.56058538e-01 -3.42356525e-02
-4.05375570e-01 1.04570460e+00 4.35510635e-01 -2.71190345e-01
1.62607789e-01 -9.25964788e-02 1.05328359e-01 -2.28762537e-01
5.69434106e-01 4.01994139e-02 -2.94867873e-01 -6.41541243e-01
1.70439649e-02 2.93107063e-01 -2.33432837e-02 -3.82055491e-01
1.18109155e+00 1.24019563e-01 3.82862799e-02 6.45354837e-02
8.92320693e-01 2.09348604e-01 -1.87329352e+00 1.28220797e-01
-5.05365193e-01 -9.84452009e-01 -3.48857909e-01 -7.19481230e-01
-1.33901834e+00 7.60176957e-01 4.80496347e-01 -1.39189780e-01
1.31294918e+00 -2.11299434e-01 5.67585349e-01 5.69060966e-02
3.68651867e-01 -8.50944221e-01 5.50349534e-01 1.84529617e-01
1.14019930e+00 -1.16542614e+00 -7.72926360e-02 -5.55978000e-01
-9.26539421e-01 7.99747109e-01 8.82450700e-01 -5.42989671e-01
4.33432549e-01 5.29424489e-01 1.42238513e-01 3.09835017e-01
-4.62913543e-01 -3.43882106e-03 1.11743078e-01 9.20829713e-01
2.44409278e-01 -3.11265010e-02 -1.00224718e-01 2.50307083e-01
-2.49703288e-01 5.76342382e-02 6.76805794e-01 8.57283652e-01
-2.89247721e-01 -1.19941056e+00 -5.93299329e-01 2.36701801e-01
-5.80959380e-01 -4.43284512e-02 -3.86079907e-01 7.85591245e-01
3.95791471e-01 9.72522020e-01 -1.53124342e-02 -2.89320648e-01
1.59594625e-01 -3.01957130e-01 7.31930733e-01 -5.15362918e-01
-2.54113287e-01 1.95066601e-01 -1.04726061e-01 -4.33351755e-01
-4.51813310e-01 -6.19103432e-01 -9.25384998e-01 -4.15719002e-01
-4.11234170e-01 -9.23614725e-02 3.68684202e-01 6.27866745e-01
2.56155133e-01 5.50150216e-01 6.66468143e-01 -9.40198541e-01
-4.53165561e-01 -6.74340546e-01 -5.49981952e-01 5.26248276e-01
3.90621841e-01 -4.06831503e-01 -1.53664500e-01 3.83824438e-01] | [11.711695671081543, -0.5252341628074646] |
29078ec3-9060-4145-9eef-135ed876a3c5 | musical-features-for-automatic-music | 2004.07171 | null | https://arxiv.org/abs/2004.07171v1 | https://arxiv.org/pdf/2004.07171v1.pdf | Musical Features for Automatic Music Transcription Evaluation | This technical report gives a detailed, formal description of the features introduced in the paper: Adrien Ycart, Lele Liu, Emmanouil Benetos and Marcus T. Pearce. "Investigating the Perceptual Validity of Evaluation Metrics for Automatic Piano Music Transcription", Transactions of the International Society for Music Information Retrieval (TISMIR), Accepted, 2020. | ['Emmanouil Benetos', 'Lele Liu', 'Marcus T. Pearce', 'Adrien Ycart'] | 2020-04-15 | null | null | null | null | ['music-transcription'] | ['music'] | [ 5.83630092e-02 -3.17504555e-01 -2.13469535e-01 1.51964948e-01
-7.92878091e-01 -7.06861377e-01 4.43934411e-01 -1.90774113e-01
6.57640025e-02 5.63186407e-01 3.58345687e-01 1.48576245e-01
-7.03420818e-01 -4.52780388e-02 -1.59647912e-01 -3.55759293e-01
-3.38756233e-01 4.11442667e-02 -3.55992727e-02 -1.57819670e-02
8.05450618e-01 5.75833499e-01 -1.80236733e+00 2.34374702e-01
2.49311298e-01 1.23366141e+00 -7.21545815e-02 1.15910029e+00
5.91204941e-01 4.74882573e-01 -8.37166011e-01 -2.43211523e-01
3.21970880e-01 -6.63908124e-01 -8.32886398e-01 -9.44769308e-02
4.89593595e-01 1.35649651e-01 -2.55222917e-01 1.24134099e+00
7.52184331e-01 2.95615762e-01 8.85871768e-01 -1.21105218e+00
-4.17744756e-01 1.00345385e+00 -6.15435183e-01 5.40161014e-01
6.58045530e-01 -3.90345395e-01 1.20451033e+00 -9.65020001e-01
5.24963200e-01 9.10733521e-01 8.62829447e-01 1.01963274e-01
-9.58711267e-01 -9.76782382e-01 -3.71509343e-01 6.50404394e-01
-1.74671388e+00 -4.98168796e-01 7.98640251e-01 -5.87981701e-01
7.27269173e-01 8.94910395e-01 5.51863015e-01 6.99622691e-01
-1.77576467e-02 8.34612310e-01 1.03711998e+00 -8.96965802e-01
-2.16823921e-01 -1.49143219e-01 2.41840899e-01 9.42982063e-02
2.11462289e-01 3.20879847e-01 -6.00443125e-01 -4.99524623e-01
7.58011162e-01 -8.01433802e-01 -1.50990993e-01 -7.33850300e-02
-1.08534575e+00 3.22366923e-01 -1.01178132e-01 9.18200493e-01
-5.51247656e-01 2.29212046e-01 6.86417699e-01 7.01356113e-01
3.69971283e-02 5.06844938e-01 -2.33979255e-01 -8.00522149e-01
-1.12337279e+00 5.61536372e-01 7.80228078e-01 1.03718269e+00
-1.35747015e-01 3.06187689e-01 -3.90904434e-02 8.14075112e-01
1.99239776e-01 3.92214715e-01 5.31812370e-01 -1.48503137e+00
1.35583669e-01 -8.74209553e-02 4.07084256e-01 -1.12450349e+00
-3.69856395e-02 -1.27302617e-01 -3.46938252e-01 -5.15003689e-02
2.50630500e-03 1.47557929e-01 -2.03839868e-01 1.08214211e+00
-3.23381007e-01 1.19371135e-02 -1.18262477e-01 9.45149422e-01
6.76937282e-01 6.61490440e-01 2.21754797e-02 -5.29704869e-01
1.04625857e+00 -7.71981001e-01 -9.06798720e-01 8.46406281e-01
-1.72049969e-01 -1.52469683e+00 1.06426954e+00 1.27966559e+00
-1.49838388e+00 -1.04250109e+00 -1.44312000e+00 5.43303430e-01
1.51337177e-01 6.75356925e-01 4.70817268e-01 9.14065242e-01
-7.16917217e-01 8.42827260e-01 -1.98745891e-01 -3.29288036e-01
-1.96533635e-01 4.33959275e-01 -2.64645051e-02 3.81488860e-01
-7.86217332e-01 4.32287157e-01 5.04272342e-01 -1.92024559e-01
-6.09279335e-01 -1.73201621e-01 -1.27799198e-01 -1.17752075e-01
-5.17709665e-02 1.90006629e-01 1.62276375e+00 -9.90374029e-01
-1.65837133e+00 8.25919032e-01 1.99501306e-01 -2.85281807e-01
1.57837778e-01 -5.12860239e-01 -1.13654435e+00 4.03422296e-01
5.74021339e-02 3.69708538e-01 6.11172676e-01 -1.05849361e+00
-5.86140156e-01 -1.89833790e-01 -3.36675644e-02 3.66323382e-01
1.41246334e-01 5.89595914e-01 -3.57151330e-01 -1.30554163e+00
1.77535638e-02 -9.82327998e-01 1.20776311e-01 -8.70159626e-01
-3.68264169e-01 -2.47302964e-01 6.81486487e-01 -7.10010767e-01
1.86839211e+00 -2.53704667e+00 -1.49889916e-01 5.70916653e-01
-2.88584977e-01 4.35346127e-01 5.64645454e-02 5.48355103e-01
-3.41192454e-01 1.48616791e-01 7.98487812e-02 4.35504168e-01
2.09598243e-01 -6.07923530e-02 -2.15054736e-01 1.45728752e-01
-6.90068185e-01 4.28564608e-01 -6.02900803e-01 -3.92453343e-01
2.68673033e-01 2.16061428e-01 7.69489491e-03 -1.57111615e-01
4.76373821e-01 2.04991981e-01 -2.71659642e-01 8.30732286e-01
3.34621668e-01 4.62283902e-02 -3.50455455e-02 -3.90790403e-01
-3.90571207e-01 5.24179041e-01 -1.59682643e+00 1.49228489e+00
1.67236641e-01 8.12522590e-01 -6.60185665e-02 -5.12308419e-01
1.00138366e+00 9.47067559e-01 8.33565652e-01 -1.79440334e-01
4.37710047e-01 4.37716126e-01 6.27896786e-02 -3.13185155e-01
6.31360948e-01 -2.28232428e-01 -2.49701574e-01 4.32294071e-01
-1.21019809e-02 -4.15457398e-01 7.38144517e-01 -1.10461287e-01
5.04715621e-01 3.01761955e-01 7.06976831e-01 -3.61792833e-01
8.58468890e-01 -8.77025798e-02 3.35827619e-01 7.08781242e-01
-7.16373146e-01 7.95554817e-01 -4.29770239e-02 -1.37481451e-01
-1.02881145e+00 -1.21544003e+00 -2.47687295e-01 9.86363709e-01
9.25299302e-02 -9.26936209e-01 -7.66195536e-01 -3.81780386e-01
-2.34562412e-01 5.07105529e-01 -2.18185112e-01 4.60197367e-02
-5.04321158e-01 -9.84783328e-05 1.08795357e+00 2.53616244e-01
4.03290093e-01 -1.28919125e+00 -3.89691472e-01 9.64768305e-02
-2.04008907e-01 -8.55681598e-01 -7.92447805e-01 -1.81996077e-01
-8.69736135e-01 -1.02221847e+00 -6.40343189e-01 -7.93057442e-01
-3.70035917e-01 1.04452679e-02 1.05828166e+00 -2.58126140e-01
-4.32530582e-01 6.04692340e-01 -7.05492139e-01 -4.92348701e-01
-4.44309443e-01 -1.05501324e-01 3.89154375e-01 -4.59230870e-01
6.32841587e-01 -7.84216583e-01 -6.45729065e-01 6.12009406e-01
-4.95383143e-01 -4.90355372e-01 6.02337956e-01 3.45752865e-01
7.01046407e-01 -3.86649296e-02 6.19618118e-01 -3.72800440e-01
8.72636080e-01 4.02636200e-01 -2.29732394e-01 -7.77754411e-02
-7.08388627e-01 -5.98905325e-01 2.90971458e-01 -3.98758441e-01
-6.87501669e-01 -2.52045125e-01 -1.20802782e-01 -3.66443306e-01
-6.47622719e-02 3.07963043e-01 3.65417749e-02 -1.29170209e-01
8.49651635e-01 1.12551175e-01 -3.68778765e-01 -4.42950457e-01
1.84076965e-01 1.07711577e+00 9.26902473e-01 -5.92681766e-01
5.94050646e-01 1.29052609e-01 -1.66271821e-01 -9.55963135e-01
-4.56910998e-01 -9.39432979e-01 -4.96411204e-01 -5.23709774e-01
5.74295640e-01 -6.47757053e-01 -6.31620824e-01 2.99072057e-01
-9.03240085e-01 3.62175971e-01 -7.00742185e-01 1.08947456e+00
-9.78070557e-01 4.45512831e-01 -6.98886991e-01 -9.44526851e-01
-5.05893528e-01 -5.65149248e-01 6.13652647e-01 3.78628187e-02
-9.74306762e-01 -5.34921229e-01 6.70525849e-01 3.18686783e-01
-3.60122845e-02 -1.60764694e-01 6.56088293e-01 -4.13637847e-01
-1.09172367e-01 -6.19465470e-01 -8.61032829e-02 6.51609838e-01
1.68865174e-01 3.44135016e-01 -9.76529837e-01 -1.05583444e-01
-6.83598071e-02 -1.81721315e-01 3.29249918e-01 5.11784196e-01
1.03965318e+00 -1.35722101e-01 6.63093999e-02 1.15151845e-01
1.38657510e+00 5.86784482e-01 7.90255845e-01 4.04133856e-01
2.13911355e-01 8.53520259e-02 8.52308691e-01 5.37646651e-01
-3.96755189e-01 8.13843191e-01 -3.19834977e-01 5.89885592e-01
-3.15165341e-01 -3.28844309e-01 3.15187722e-01 1.39515734e+00
-7.67552793e-01 1.86891593e-02 -4.38628465e-01 4.61110234e-01
-1.60579348e+00 -1.14505672e+00 -1.08343445e-01 2.01723766e+00
5.77489793e-01 -8.64358768e-02 4.71481621e-01 1.07006371e+00
6.63452864e-01 1.23179778e-01 1.72893614e-01 -6.25890076e-01
-1.69221297e-01 4.43947524e-01 3.96754056e-01 3.44445854e-01
-1.23318946e+00 6.65311515e-01 7.94214201e+00 9.95377958e-01
-8.96132469e-01 7.48184100e-02 -3.62981223e-02 -6.71396628e-02
1.53114289e-01 -9.33780987e-03 -3.87771875e-01 2.57660449e-01
1.05588043e+00 -5.25166452e-01 5.00466585e-01 5.02684474e-01
3.75135273e-01 -4.40735333e-02 -7.74778247e-01 1.41748691e+00
4.39962149e-01 -9.59375441e-01 9.62711722e-02 -1.08605161e-01
9.32591081e-01 -2.13688478e-01 3.71091217e-01 1.87622100e-01
-3.35226297e-01 -1.09463108e+00 7.13817596e-01 3.86532038e-01
7.47098804e-01 -1.05903041e+00 6.03119552e-01 -8.68185386e-02
-1.17942572e+00 -1.27832413e-01 -2.51677692e-01 -2.28104368e-01
-1.31143183e-01 2.61318381e-03 -3.48337829e-01 6.56561136e-01
6.31635189e-01 5.12059689e-01 -3.19997579e-01 1.34067643e+00
6.73423484e-02 1.04506695e+00 -6.25453740e-02 4.63623106e-02
3.15173268e-02 -1.15036741e-01 9.39303696e-01 1.24596524e+00
5.65714836e-01 2.03628868e-01 -2.59769540e-02 2.71831453e-01
3.27780217e-01 4.74097610e-01 -4.27867353e-01 -3.67080450e-01
3.96576375e-01 9.37145889e-01 -7.33780324e-01 -4.42900151e-01
-7.77837709e-02 6.94954216e-01 -6.04699969e-01 2.40649253e-01
-4.88324523e-01 -6.74457967e-01 2.85818338e-01 2.18152091e-01
2.20353469e-01 -5.99371530e-02 -6.03924394e-01 -5.74890077e-01
-1.06752433e-01 -1.10741663e+00 4.11798716e-01 -8.14600945e-01
-1.26648343e+00 4.60154802e-01 1.51458010e-01 -1.75693488e+00
-2.61926442e-01 -7.03800321e-01 -5.02812684e-01 8.47567618e-01
-3.71275842e-01 -8.05213511e-01 1.37425691e-01 3.22530985e-01
3.67707580e-01 -5.42062044e-01 9.44463015e-01 7.45387614e-01
2.05974042e-01 6.20521963e-01 1.04122162e-01 -4.83360514e-02
7.38548517e-01 -1.07249320e+00 8.03328902e-02 3.34243298e-01
7.99989045e-01 3.54730546e-01 9.64111984e-01 -4.78382111e-01
-1.06712508e+00 -3.05614412e-01 9.95830715e-01 -4.11120832e-01
7.01853216e-01 5.83164990e-01 -4.07251775e-01 6.91528141e-01
4.74191070e-01 -9.18313444e-01 9.41174030e-01 -1.00488342e-01
-2.43108988e-01 -2.05503911e-01 -7.88265049e-01 4.38850075e-01
6.26221001e-01 -5.89804828e-01 -8.62745643e-01 -2.90895114e-03
2.54555911e-01 -7.81423226e-02 -1.11065614e+00 5.57951629e-01
1.08851981e+00 -9.73937571e-01 7.46366322e-01 -2.71557838e-01
7.49039324e-03 -6.93672538e-01 -4.41175222e-01 -5.95844448e-01
-3.39143991e-01 -8.05197597e-01 3.45110029e-01 9.25069392e-01
3.03589970e-01 1.91466913e-01 5.89554608e-01 -2.51477569e-01
-1.88814610e-01 -1.14473954e-01 -1.11364317e+00 -8.39779556e-01
-1.96954325e-01 -9.98126566e-01 2.52301663e-01 8.23561728e-01
6.85773253e-01 5.08573174e-01 -5.49578428e-01 -3.62901002e-01
4.30178702e-01 -6.27842322e-02 8.31900656e-01 -1.18226135e+00
-6.79756403e-01 -8.48581851e-01 -1.07425714e+00 -8.24246049e-01
-2.41080895e-01 -6.65252507e-01 -8.04779008e-02 -1.20063901e+00
2.50905871e-01 -1.12579323e-01 -9.14545894e-01 -1.69722319e-01
2.54836202e-01 8.28667581e-01 5.25521398e-01 4.03755605e-01
-7.37457871e-01 -7.48646408e-02 1.03281033e+00 -1.40617251e-01
-1.62363425e-01 5.23677528e-01 -5.39969742e-01 7.34908402e-01
6.05758488e-01 -1.78076938e-01 -4.44345534e-01 2.50530541e-01
1.94801673e-01 2.68887192e-01 8.76948908e-02 -1.26188970e+00
-1.31023496e-01 -4.73123863e-02 2.07736239e-01 -8.08678150e-01
4.54000264e-01 -7.31336176e-01 6.07097924e-01 2.17223674e-01
-2.99761027e-01 2.95935839e-01 3.57166737e-01 1.65155172e-01
-6.49094343e-01 -4.61526632e-01 6.75710022e-01 -2.42783036e-02
-4.17485118e-01 -1.65406317e-01 -5.55310249e-01 -3.42204183e-01
6.86344266e-01 -5.35348058e-01 1.93699703e-01 -5.57941556e-01
-1.05969667e+00 -4.31267262e-01 1.14217341e-01 5.68003535e-01
5.29402733e-01 -1.70178092e+00 -6.41813695e-01 -5.24775237e-02
6.51745796e-02 -1.11897528e+00 1.48102343e-01 1.10569072e+00
-7.14454770e-01 6.10778868e-01 -5.14113128e-01 -2.80522704e-01
-1.72564971e+00 2.47124061e-01 -1.88788652e-01 -3.18401568e-02
-5.11649013e-01 6.77302837e-01 -4.19951797e-01 7.33324587e-02
8.24423954e-02 -2.15676818e-02 -5.15076935e-01 -6.98600858e-02
6.21285439e-01 6.34137273e-01 -2.57347780e-03 -9.62886691e-01
-2.31419802e-01 7.59933174e-01 2.21674919e-01 -7.21011400e-01
7.27173805e-01 -2.16235191e-01 -1.55509412e-01 1.45955026e+00
7.71439433e-01 3.96945715e-01 -1.70099493e-02 3.67841721e-01
4.10786062e-01 -7.43936241e-01 -3.51015300e-01 -1.04735744e+00
-3.62136662e-01 6.20524764e-01 1.14350867e+00 2.83717453e-01
1.28334987e+00 -2.40143061e-01 7.18047023e-01 3.30391467e-01
4.77896571e-01 -1.42255616e+00 -7.19980672e-02 5.07083476e-01
9.73671138e-01 -3.05351883e-01 1.86082110e-01 -4.20842499e-01
-7.61781633e-01 1.14901578e+00 -1.24501688e-02 -4.05397713e-01
9.80405927e-01 -1.24529382e-04 3.10482085e-01 2.51700729e-01
-6.25708640e-01 -3.59978750e-02 6.98491931e-01 6.18289828e-01
9.06659245e-01 3.38030785e-01 -1.16968715e+00 9.33300793e-01
-8.25853884e-01 2.74065554e-01 3.55148971e-01 7.38801301e-01
-4.56786186e-01 -1.15510857e+00 -6.21910810e-01 2.22021833e-01
-9.50368047e-01 1.41849963e-03 -8.02321732e-01 9.14523780e-01
2.74943709e-01 9.64651048e-01 -3.34392726e-01 -6.87617302e-01
6.92342997e-01 1.30084410e-01 7.83659399e-01 -3.15717608e-01
-9.67202783e-01 7.94601321e-01 3.81366938e-01 -4.75057587e-02
-7.51458824e-01 -8.47782016e-01 -9.14081752e-01 -2.93020576e-01
-2.72309780e-01 1.02448726e+00 6.71392798e-01 5.41848421e-01
4.71695736e-02 4.01356876e-01 4.75316077e-01 -7.52313495e-01
-3.73868018e-01 -1.35039890e+00 -1.18581522e+00 2.31212467e-01
-4.58175801e-02 -6.41972840e-01 -3.76168579e-01 3.05327386e-01] | [15.973363876342773, 5.254083633422852] |
f49d7a81-dd94-4f0b-88d6-9f08338c93ac | dualfl-a-duality-based-federated-learning | 2305.10294 | null | https://arxiv.org/abs/2305.10294v1 | https://arxiv.org/pdf/2305.10294v1.pdf | DualFL: A Duality-based Federated Learning Algorithm with Communication Acceleration in the General Convex Regime | We propose a novel training algorithm called DualFL (Dualized Federated Learning), for solving a distributed optimization problem in federated learning. Our approach is based on a specific dual formulation of the federated learning problem. DualFL achieves communication acceleration under various settings on smoothness and strong convexity of the problem. Moreover, it theoretically guarantees the use of inexact local solvers, preserving its optimal communication complexity even with inexact local solutions. DualFL is the first federated learning algorithm that achieves communication acceleration, even when the cost function is either nonsmooth or non-strongly convex. Numerical results demonstrate that the practical performance of DualFL is comparable to those of state-of-the-art federated learning algorithms, and it is robust with respect to hyperparameter tuning. | ['Jinchao Xu', 'Jongho Park'] | 2023-05-17 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-6.99868619e-01 2.05753390e-02 -4.97883052e-01 -6.83018267e-02
-1.42682791e+00 -6.72308326e-01 2.39119902e-01 8.36610049e-02
-1.03636689e-01 8.53712440e-01 2.50742346e-01 -5.82226157e-01
-5.61920881e-01 -7.60163724e-01 -1.06215990e+00 -7.97720611e-01
-4.92352664e-01 5.67248344e-01 -8.07959616e-01 1.62594989e-02
-2.26442486e-01 2.89950550e-01 -1.28945243e+00 4.01023537e-01
9.23915207e-01 1.10673833e+00 -2.52491444e-01 6.81106508e-01
-2.29351580e-01 9.04057443e-01 -2.58832544e-01 -5.13539553e-01
8.79576087e-01 -1.93570688e-01 -1.10698724e+00 1.05139673e-01
8.05535316e-01 -2.50609428e-01 -4.58940566e-01 1.14244688e+00
3.39120686e-01 8.07040036e-02 -6.46975264e-02 -1.24436462e+00
-4.84618217e-01 7.73084998e-01 -3.89969796e-01 1.20414145e-01
3.12907040e-01 2.16979021e-03 1.08852887e+00 -1.14751410e+00
4.46810901e-01 1.02783048e+00 8.99275064e-01 2.38598406e-01
-9.36313510e-01 -6.01956129e-01 3.19424957e-01 6.75906837e-02
-1.16137350e+00 -4.63755041e-01 3.16246867e-01 -9.63680819e-02
7.70006120e-01 4.12918031e-01 4.32687551e-01 4.93147373e-01
2.63291448e-02 9.25077975e-01 1.19015479e+00 -4.70066220e-01
5.13628364e-01 6.72198460e-02 2.31029361e-01 1.13970268e+00
5.52670896e-01 3.33200395e-01 -9.23041523e-01 -1.05478895e+00
2.52641290e-01 1.99699312e-01 -5.28323472e-01 -4.64247376e-01
-1.05254531e+00 9.37507629e-01 3.11107367e-01 9.57020968e-02
-5.24140775e-01 2.73519993e-01 4.70537126e-01 8.71580482e-01
8.18631887e-01 -5.68831936e-02 -7.92059600e-01 4.56323326e-02
-7.36074150e-01 2.11820766e-01 1.36789441e+00 1.06992316e+00
1.07875514e+00 5.33993319e-02 6.16412759e-02 4.53516245e-01
2.40460169e-02 4.59020674e-01 5.14169335e-01 -1.08971918e+00
8.50454092e-01 5.51526904e-01 1.82853028e-01 -7.27014184e-01
-1.99635208e-01 -5.84010661e-01 -6.19869053e-01 -2.19975505e-02
3.91161740e-01 -7.08642066e-01 -6.27724305e-02 1.51854360e+00
9.78769243e-01 3.87400627e-01 2.32957542e-01 1.19022238e+00
4.79087621e-01 5.95423043e-01 -5.04416287e-01 -4.42305356e-01
9.26179826e-01 -1.52879786e+00 -4.92298871e-01 -3.99026722e-02
1.07970941e+00 -4.10229564e-01 5.77248931e-01 5.78512073e-01
-1.13234377e+00 1.51044458e-01 -8.03473413e-01 3.83248150e-01
-1.69292137e-01 -1.94393724e-01 1.27406383e+00 8.69821608e-01
-1.06395388e+00 6.43258572e-01 -7.83721149e-01 5.58262356e-02
4.38011408e-01 6.45258904e-01 -5.52451074e-01 -5.19839644e-01
-5.69590509e-01 3.02164316e-01 2.05884829e-01 -2.11140320e-01
-1.15621316e+00 -1.15994692e+00 -5.55473924e-01 1.75507113e-01
3.75543118e-01 -1.13290811e+00 1.50622642e+00 -1.10850286e+00
-1.38672543e+00 5.90541542e-01 -6.85513616e-02 -4.97413725e-01
7.31381774e-01 -2.19816402e-01 -1.91055655e-01 1.04619727e-01
-1.79311991e-01 -5.35657644e-01 7.70808220e-01 -9.68279839e-01
-8.29972506e-01 -8.09617341e-01 1.24464206e-01 2.22933993e-01
-1.05175209e+00 -1.36752769e-01 -4.17324662e-01 -4.27514642e-01
-2.76466370e-01 -7.31583536e-01 -5.27233362e-01 1.75658196e-01
7.14852242e-03 -2.15549350e-01 1.08604372e+00 -3.46569359e-01
9.01190162e-01 -2.03119993e+00 1.28536940e-01 3.03862542e-01
4.89200979e-01 4.59888838e-02 -2.78833658e-01 7.16981947e-01
1.13258041e-01 -1.84502795e-01 1.56549606e-02 -5.36114097e-01
1.36746421e-01 1.89965829e-01 -2.55161732e-01 1.32622647e+00
-8.70334327e-01 5.74882627e-01 -1.07453358e+00 -3.33581626e-01
-1.64632604e-01 3.33276689e-01 -8.01892579e-01 3.11199784e-01
-3.06864232e-01 2.35495731e-01 -8.75644982e-01 7.56853163e-01
8.43670547e-01 -5.50238788e-01 5.08907199e-01 1.96097761e-01
-8.51696804e-02 6.10388853e-02 -1.39353371e+00 1.95582116e+00
-9.50242281e-01 5.98345585e-02 1.18457484e+00 -9.58539546e-01
5.85142255e-01 5.46267688e-01 9.15748298e-01 -3.14790040e-01
1.62579224e-01 4.82535988e-01 -8.88531804e-01 -2.99132913e-01
9.32326689e-02 1.44298807e-01 1.98427632e-01 1.15157044e+00
2.02846259e-01 5.22494018e-01 -8.58208090e-02 5.32747209e-01
1.21627283e+00 -3.40545207e-01 3.89371216e-02 -6.06691301e-01
5.57277262e-01 5.06961206e-03 6.35315239e-01 7.68503129e-01
2.39360798e-02 9.66577157e-02 -2.54970491e-02 -8.50904822e-01
-5.39165974e-01 -6.95348322e-01 -7.23623186e-02 1.41234052e+00
1.20989718e-01 -8.64199936e-01 -7.55769312e-01 -1.07218540e+00
6.01547539e-01 1.34248227e-01 -3.90260220e-01 2.45861813e-01
-3.40225041e-01 -7.31567442e-01 1.60482481e-01 1.57496519e-02
3.85745913e-01 -1.13199577e-01 -1.97032288e-01 2.41317213e-01
6.90282881e-02 -7.91056335e-01 -8.15053761e-01 3.20165485e-01
-1.10386312e+00 -1.51009071e+00 -5.33213794e-01 -7.64780402e-01
8.40646207e-01 7.18839526e-01 1.05537391e+00 2.92416573e-01
-2.11904407e-01 8.73001754e-01 -1.45443052e-01 -1.39717767e-02
-1.38824433e-01 -8.17738920e-02 2.63877600e-01 4.45210874e-01
-5.82890622e-02 -3.11433733e-01 -6.87524736e-01 1.56528354e-01
-7.87334502e-01 -3.71072859e-01 1.85232423e-02 1.38030601e+00
6.74095809e-01 7.27265552e-02 3.66695613e-01 -1.20165861e+00
6.65716290e-01 -8.16492736e-01 -7.74548173e-01 4.41428214e-01
-9.12604868e-01 1.33170500e-01 1.24834311e+00 -9.64710861e-02
-9.90042746e-01 3.43339324e-01 4.00182486e-01 -8.84809256e-01
3.82543147e-01 5.50546229e-01 1.64626047e-01 -8.54916871e-01
7.76171267e-01 1.27632290e-01 3.32355022e-01 -6.82107210e-01
5.59253395e-01 6.79637790e-01 4.50583786e-01 -1.19878697e+00
5.59202254e-01 7.24254608e-01 -8.70694146e-02 -3.59443367e-01
-9.44315791e-01 -6.91786528e-01 5.88998795e-02 -1.33598540e-02
-2.12949306e-01 -1.18591905e+00 -9.63661313e-01 5.64978197e-02
-9.21814144e-01 -3.39370668e-01 -3.30451131e-01 5.05255401e-01
-7.23964095e-01 4.55753535e-01 -7.36434460e-01 -6.90065444e-01
-9.83898640e-01 -7.46194839e-01 6.27573848e-01 -1.48182303e-01
4.88166362e-01 -1.47644317e+00 3.50739419e-01 5.17697990e-01
5.71524560e-01 3.02324384e-01 4.22285408e-01 -5.26434541e-01
-5.52743793e-01 -3.20831358e-01 1.31926298e-01 -7.26001188e-02
4.22043689e-02 -3.93343359e-01 -7.30137467e-01 -1.16787708e+00
1.29417643e-01 -7.16235816e-01 6.46503687e-01 4.22810279e-02
1.34263253e+00 -1.37561285e+00 -3.38168889e-01 1.39891887e+00
1.73185372e+00 -9.14184153e-01 -4.32205677e-01 3.71326804e-01
3.08841825e-01 2.70306110e-01 5.40240347e-01 1.27895331e+00
4.04284090e-01 1.79215536e-01 7.99111843e-01 -1.21930830e-01
1.35895222e-01 -2.75455266e-01 3.40398252e-01 7.63156831e-01
3.05728316e-01 1.80605233e-01 -7.53483415e-01 5.86753130e-01
-2.26648545e+00 -6.73123658e-01 -2.05356449e-01 2.28199601e+00
7.21675634e-01 -7.17178047e-01 2.61207223e-01 -1.40338510e-01
6.30237639e-01 -3.96654792e-02 -7.71053195e-01 -5.30178368e-01
-1.87904462e-01 3.10207039e-01 1.01370239e+00 3.57865989e-01
-7.99924672e-01 6.95000291e-01 7.18138027e+00 5.86431980e-01
-1.04372287e+00 6.92968667e-01 5.53457558e-01 -5.28987229e-01
-3.27477247e-01 6.50673348e-04 -5.70245147e-01 2.80004948e-01
9.20410335e-01 -7.51642466e-01 1.07361221e+00 1.37938297e+00
-1.77777305e-01 4.42570180e-01 -1.04832733e+00 1.28353691e+00
-1.24712966e-01 -1.90525973e+00 -3.91786814e-01 7.91421086e-02
1.42859483e+00 5.50433457e-01 9.98617262e-02 1.62427709e-01
9.15573537e-01 -7.88731337e-01 4.95956481e-01 2.49959864e-02
7.01448202e-01 -1.17954528e+00 4.37077641e-01 6.58333361e-01
-1.18855643e+00 -7.12656796e-01 -3.12529743e-01 -1.65457994e-01
-3.21439981e-01 7.02408671e-01 -2.25955844e-01 7.68929601e-01
9.23798501e-01 5.10531187e-01 -1.23028301e-01 1.14633334e+00
2.17748567e-01 6.36263311e-01 -6.09957576e-01 3.19135040e-01
3.51645350e-01 -2.00802654e-01 3.99342746e-01 1.12582624e+00
2.98550546e-01 2.06551820e-01 6.27825618e-01 4.61446166e-01
-4.78312671e-01 5.14998615e-01 -6.12960219e-01 2.24217579e-01
6.16159081e-01 1.43013430e+00 1.51967093e-01 -2.07318276e-01
-6.87214375e-01 8.51840079e-01 9.30626512e-01 3.36964816e-01
-4.19306427e-01 4.47947979e-02 9.45638239e-01 -2.86063224e-01
4.24600154e-01 -1.27198145e-01 -2.91499794e-01 -1.39672124e+00
4.83493246e-02 -1.23414207e+00 1.26898932e+00 2.77520388e-01
-1.68180120e+00 5.00910163e-01 -5.34133971e-01 -8.87583792e-01
-1.52842149e-01 -4.78424668e-01 -6.37054384e-01 5.13720870e-01
-1.65496647e+00 -1.07262063e+00 -2.58426011e-01 1.46207547e+00
-1.12472018e-02 -4.46346611e-01 1.03577006e+00 3.45402062e-01
-6.60571575e-01 1.12855101e+00 7.24755645e-01 -3.84593874e-01
5.45284092e-01 -1.00394833e+00 -2.96093851e-01 7.73792744e-01
1.21918567e-01 4.91856694e-01 5.14636815e-01 -1.38474151e-01
-2.78730965e+00 -1.41835928e+00 6.52807295e-01 -1.68546883e-03
7.65094638e-01 -8.79020393e-02 -3.67757857e-01 8.28254700e-01
1.77982062e-01 9.96033907e-01 9.76633728e-01 3.86712253e-01
-6.64301574e-01 -4.92936850e-01 -1.47599876e+00 2.70100553e-02
8.51928711e-01 -4.99705970e-01 4.90528233e-02 1.26219308e+00
5.77400029e-01 -7.33148456e-01 -1.03183520e+00 -6.06986359e-02
8.65636840e-02 -8.74954522e-01 7.21979260e-01 -7.65147984e-01
-2.27570221e-01 7.83847272e-02 -2.81922370e-01 -1.28223848e+00
-4.39319462e-01 -1.51231182e+00 -1.15181911e+00 6.86109781e-01
5.61976172e-02 -1.03087282e+00 1.28584349e+00 4.70211029e-01
-9.68588442e-02 -9.35787678e-01 -1.17647243e+00 -1.12817192e+00
1.61722317e-01 -2.13054419e-01 8.88096929e-01 1.23067880e+00
5.21980882e-01 -3.69700849e-01 -4.21583325e-01 4.55893725e-01
1.17804241e+00 6.45170927e-01 9.90797698e-01 -8.57913077e-01
-9.68612790e-01 -1.68330699e-01 5.19511029e-02 -1.04299450e+00
5.03131330e-01 -1.37303054e+00 -5.30432522e-01 -1.11446559e+00
1.67967111e-01 -7.56702423e-01 -4.49730515e-01 8.73586714e-01
9.90373269e-02 -9.90581959e-02 2.74268001e-01 3.67411822e-01
-8.76358151e-01 6.40791893e-01 9.94866371e-01 -2.67274648e-01
-5.26070409e-02 5.47819920e-02 -7.72475123e-01 3.39427173e-01
6.16935611e-01 -4.87701952e-01 -4.90244478e-03 -8.65395963e-01
8.53649005e-02 5.11028290e-01 1.01063050e-01 -6.57044828e-01
8.52802753e-01 -3.45702976e-01 -2.57708490e-01 -2.52613455e-01
-9.93010178e-02 -1.11840415e+00 1.74610436e-01 7.05828071e-01
-1.19887635e-01 1.79888189e-01 -8.11804906e-02 7.71193087e-01
-2.52987057e-01 7.16772676e-02 5.72719455e-01 -1.30141094e-01
-2.14510813e-01 9.12762761e-01 -1.70852654e-02 2.54539013e-01
1.28032279e+00 2.53134459e-01 -4.11906898e-01 -4.60427910e-01
-3.43830049e-01 7.60601044e-01 5.21694005e-01 -6.57386258e-02
4.55495983e-01 -1.28559387e+00 -8.39839458e-01 1.98256657e-01
-7.82848746e-02 -5.24415597e-02 5.12524769e-02 9.04691458e-01
-4.59796369e-01 5.41068137e-01 4.25853044e-01 -2.72866309e-01
-1.14864194e+00 8.83869588e-01 5.77520609e-01 -3.57395858e-01
-9.57852960e-01 9.63623047e-01 -3.45498204e-01 -7.02355802e-01
6.74175084e-01 3.99841905e-01 8.29689980e-01 -4.97529149e-01
7.58634508e-01 7.66117990e-01 5.23410976e-01 -1.13642693e-01
-3.08891505e-01 3.44500542e-02 3.05202771e-02 3.94021064e-01
1.55435050e+00 1.09448411e-01 -3.74499708e-01 -2.64072895e-01
1.44589210e+00 2.30449334e-01 -1.29752707e+00 -6.81481302e-01
-3.09702843e-01 -7.95655847e-01 6.05798364e-01 -5.68038821e-01
-1.78722966e+00 5.56417741e-02 2.79960126e-01 -1.24288335e-01
1.01094973e+00 -2.02976406e-01 1.12955582e+00 7.33416319e-01
1.05832362e+00 -1.07312953e+00 -2.08048239e-01 4.74072635e-01
5.56619883e-01 -1.09377337e+00 3.00190836e-01 -2.88283557e-01
-2.72963960e-02 1.36042404e+00 4.98593539e-01 -3.90998781e-01
9.00527537e-01 5.92736542e-01 -3.33101377e-02 -2.59721965e-01
-1.35987389e+00 3.77689868e-01 -1.15648516e-01 2.93940812e-01
1.81360975e-01 5.21247089e-02 -5.18284798e-01 6.22394085e-01
-1.13489322e-01 4.17612866e-02 -2.89665703e-02 1.12185025e+00
-1.99072719e-01 -9.69709337e-01 -7.96638608e-01 2.81679481e-01
-5.88143587e-01 2.13572338e-01 -2.84075856e-01 3.36300135e-01
-1.20529041e-01 1.08530176e+00 -2.46058539e-01 -2.02191010e-01
-2.04162091e-01 -3.05104136e-01 5.42714536e-01 -3.13275546e-01
-1.23436403e+00 -2.52355456e-01 4.31586057e-02 -1.00724423e+00
-2.74026394e-02 -3.44266593e-01 -1.31563520e+00 -9.32814598e-01
-4.88892674e-01 7.65715539e-01 8.50081801e-01 7.33555019e-01
7.50996053e-01 -1.78174138e-01 1.44735861e+00 -5.93039572e-01
-1.51395512e+00 -2.04914525e-01 -7.69518614e-01 3.31476986e-01
5.71621895e-01 -1.10485882e-01 -6.76914215e-01 -4.83071744e-01] | [6.093705177307129, 5.472965717315674] |
fc70414b-cd38-4f9a-b38d-304cc1d7c45e | learning-to-summarize-videos-by-contrasting | 2301.05213 | null | https://arxiv.org/abs/2301.05213v3 | https://arxiv.org/pdf/2301.05213v3.pdf | Learning to Summarize Videos by Contrasting Clips | Video summarization aims at choosing parts of a video that narrate a story as close as possible to the original one. Most of the existing video summarization approaches focus on hand-crafted labels. As the number of videos grows exponentially, there emerges an increasing need for methods that can learn meaningful summarizations without labeled annotations. In this paper, we aim to maximally exploit unsupervised video summarization while concentrating the supervision to a few, personalized labels as an add-on. To do so, we formulate the key requirements for the informative video summarization. Then, we propose contrastive learning as the answer to both questions. To further boost Contrastive video Summarization (CSUM), we propose to contrast top-k features instead of a mean video feature as employed by the existing method, which we implement with a differentiable top-k feature selector. Our experiments on several benchmarks demonstrate, that our approach allows for meaningful and diverse summaries when no labeled data is provided. | ['Arnold Smeulders', 'Cees Kaandorp', 'Artem Moskalev', 'Ivan Sosnovik'] | 2023-01-12 | null | null | null | null | ['unsupervised-video-summarization'] | ['computer-vision'] | [ 3.56165528e-01 1.03128664e-01 -4.59717065e-01 -3.84203523e-01
-9.78299558e-01 -5.30731499e-01 6.52844071e-01 4.85869527e-01
-2.63510764e-01 7.79320002e-01 8.32049429e-01 2.15613708e-01
-5.06155044e-02 -3.16230804e-01 -7.98351586e-01 -5.63096762e-01
1.27546102e-01 8.18407536e-02 1.73280567e-01 3.23942222e-04
5.09503245e-01 1.94634944e-01 -1.61048985e+00 4.58576471e-01
1.04627967e+00 7.69229531e-01 2.58726776e-01 6.88740909e-01
-9.51625109e-02 1.03137541e+00 -6.61683738e-01 -2.54782528e-01
1.47116348e-01 -9.01929975e-01 -8.96996081e-01 7.02489853e-01
7.75709569e-01 -5.00786424e-01 -3.91679823e-01 8.74644279e-01
3.21152747e-01 4.76801604e-01 8.03371549e-01 -1.23272419e+00
-3.23923469e-01 8.67510557e-01 -5.10451019e-01 1.96541101e-01
6.70148194e-01 -4.93859574e-02 1.33600283e+00 -7.15854287e-01
8.71921480e-01 8.96844327e-01 4.71092194e-01 4.54061747e-01
-1.19541347e+00 -4.75964919e-02 4.85777885e-01 3.35379332e-01
-1.09436262e+00 -6.24819875e-01 1.04128504e+00 -5.04925668e-01
5.18495560e-01 5.19626558e-01 7.09295690e-01 9.86614048e-01
-2.07376540e-01 1.15680766e+00 5.88681281e-01 -3.54993880e-01
3.71224612e-01 2.02738177e-02 2.84558624e-01 5.42128384e-01
3.11941445e-01 -8.04618537e-01 -7.16681123e-01 -2.05403611e-01
3.80280286e-01 2.36365333e-01 -5.45242965e-01 -4.57792968e-01
-1.36287451e+00 7.08409786e-01 1.34168431e-01 2.34789938e-01
-5.19256234e-01 1.39942706e-01 7.99036443e-01 2.69980967e-01
6.86141789e-01 7.49001563e-01 -1.68509364e-01 -8.87638032e-02
-1.46114683e+00 3.39338124e-01 8.12252700e-01 1.07648170e+00
9.75135326e-01 -5.50154261e-02 -7.29111016e-01 6.98626280e-01
-1.56526953e-01 7.45453089e-02 3.95229638e-01 -1.29767895e+00
4.05676514e-01 5.91339290e-01 2.06907183e-01 -9.21046019e-01
-8.02590027e-02 -2.12104976e-01 -6.17559433e-01 -3.19207489e-01
-1.25307543e-02 -9.60009396e-02 -6.43781543e-01 1.57118070e+00
1.32090256e-01 2.77745873e-01 -2.73931510e-04 7.67335832e-01
9.23411429e-01 7.52675295e-01 -1.80135265e-01 -6.55194521e-01
1.06430852e+00 -1.35966635e+00 -6.44094586e-01 1.58161148e-01
6.65569425e-01 -4.92620677e-01 1.14509642e+00 3.60332996e-01
-1.10809827e+00 -3.91500890e-01 -1.05320382e+00 -7.45256990e-02
2.43236408e-01 1.93172336e-01 2.76019007e-01 2.83848107e-01
-1.13562894e+00 7.56198287e-01 -7.78011143e-01 -6.01722419e-01
4.20629740e-01 1.48966804e-01 -4.90793526e-01 -1.75245360e-01
-6.88635767e-01 4.13322210e-01 5.94982266e-01 -2.85532236e-01
-8.41386676e-01 -4.17501420e-01 -8.63435447e-01 2.14553460e-01
8.12061250e-01 -9.02383506e-01 1.25389516e+00 -1.36027968e+00
-1.54213786e+00 5.21597207e-01 -3.07468116e-01 -3.15782726e-01
5.23251176e-01 -4.42653477e-01 1.76739246e-01 7.92629838e-01
2.54053146e-01 6.97657526e-01 1.02715731e+00 -1.41513002e+00
-8.01822543e-01 1.90980490e-02 3.53233486e-01 3.92279774e-01
-7.79972553e-01 6.60285819e-04 -4.78939831e-01 -7.77678907e-01
-2.00812757e-01 -7.50964284e-01 -2.88486570e-01 -4.34819102e-01
-6.34322345e-01 -4.00623292e-01 8.00668895e-01 -6.99621201e-01
1.41313612e+00 -2.12520432e+00 6.39943898e-01 -1.01966165e-01
4.42231536e-01 2.30631396e-01 -1.05234936e-01 7.20984459e-01
1.78396285e-01 8.99852812e-02 -2.56953001e-01 -5.96324861e-01
-4.14241925e-02 2.71916017e-03 -1.22563779e-01 4.12716597e-01
1.01557106e-01 6.42258584e-01 -1.14064431e+00 -7.03090966e-01
1.88171063e-02 1.45206064e-01 -8.29479158e-01 3.06924641e-01
-3.78071725e-01 6.50464714e-01 -6.44045115e-01 3.62530559e-01
3.38730872e-01 -3.42213869e-01 8.97593275e-02 -1.96595743e-01
-2.11399883e-01 6.69347346e-02 -9.62405443e-01 1.94411612e+00
-1.66982785e-01 6.01984560e-01 -2.74550527e-01 -1.32041943e+00
7.64030397e-01 3.42441738e-01 7.30477333e-01 -1.27640039e-01
6.05095886e-02 2.30655789e-01 -5.97312212e-01 -6.17273092e-01
7.92233348e-01 9.22669396e-02 -2.84319341e-01 4.60898817e-01
1.70748472e-01 1.62437320e-01 6.20406389e-01 5.71859479e-01
1.25450933e+00 6.94146752e-02 5.32876313e-01 -1.54665396e-01
6.94470704e-01 7.36314282e-02 5.83157241e-01 8.20636272e-01
1.18073173e-01 8.55315864e-01 8.27854812e-01 -2.39985153e-01
-1.01355755e+00 -5.77294350e-01 4.52780515e-01 1.20581210e+00
7.96609744e-02 -9.66621459e-01 -1.08433378e+00 -9.25951898e-01
-2.49769539e-01 5.09826303e-01 -4.05711025e-01 -8.23106095e-02
-7.61301041e-01 -1.79378971e-01 1.71046138e-01 2.92674482e-01
2.67002672e-01 -9.89830256e-01 -8.27142596e-01 7.65620098e-02
-5.01756907e-01 -1.02441573e+00 -7.97648549e-01 -1.58018842e-02
-8.02343965e-01 -8.35351050e-01 -9.31025624e-01 -7.99964309e-01
7.40390062e-01 5.67404389e-01 9.85247314e-01 -4.60558236e-02
1.66440889e-01 7.17708528e-01 -8.97703528e-01 4.57107946e-02
-3.76239121e-01 5.35093188e-01 -4.08864617e-02 3.10100824e-01
-3.62882726e-02 -6.64027989e-01 -5.68474054e-01 -1.44603133e-01
-1.06016874e+00 3.41429263e-01 5.41948020e-01 6.98682666e-01
5.35086513e-01 -1.73509181e-01 8.46606016e-01 -1.03479731e+00
5.12481987e-01 -5.35912156e-01 -8.82949829e-02 2.87118763e-01
-3.43487591e-01 1.80035517e-01 1.07741237e+00 -3.15841973e-01
-8.80758345e-01 1.40471146e-01 1.60249785e-01 -6.45096302e-01
-1.12480029e-01 6.75638378e-01 -2.36115605e-01 3.24825466e-01
3.93709511e-01 3.78568172e-01 -2.27012679e-01 -4.55413193e-01
5.39526343e-01 6.34340525e-01 5.65869391e-01 -4.49813306e-01
5.64722955e-01 3.78460318e-01 -1.97566465e-01 -1.08808041e+00
-9.81775284e-01 -7.17041254e-01 -6.85138404e-01 -3.90848875e-01
7.38682628e-01 -7.58004844e-01 -4.10201937e-01 1.29815675e-02
-9.95513618e-01 -1.03941396e-01 -6.43384576e-01 3.58976275e-01
-1.00000966e+00 7.93634593e-01 -3.27322990e-01 -4.99538660e-01
-3.30308855e-01 -1.05559456e+00 1.09337580e+00 2.36182317e-01
-3.29444408e-01 -6.70079231e-01 1.13230430e-01 3.19680333e-01
1.54132143e-01 4.12735134e-01 6.92434907e-01 -9.84625459e-01
-6.16899490e-01 -1.67081371e-01 6.21781824e-03 4.22374845e-01
3.12244773e-01 1.16949929e-02 -7.43336022e-01 -3.56818527e-01
-8.77352152e-03 -3.65150988e-01 1.28484201e+00 5.13416111e-01
1.12266791e+00 -7.97579527e-01 -9.31153074e-02 5.34079194e-01
1.22561431e+00 -1.48108855e-01 2.99459875e-01 2.11227164e-01
9.12683487e-01 6.88884318e-01 7.38764465e-01 8.01138878e-01
4.01895195e-01 5.57420611e-01 3.10675681e-01 2.92350262e-01
-2.83535700e-02 -4.42927092e-01 5.29041171e-01 1.03765762e+00
-1.82118580e-01 -5.05058408e-01 -5.24890721e-01 5.75294435e-01
-2.29003930e+00 -1.20576704e+00 1.50706008e-01 2.16894007e+00
7.34327316e-01 -9.87943411e-02 4.01934654e-01 1.81673303e-01
8.29934478e-01 5.52128971e-01 -5.00318646e-01 -5.80291413e-02
9.04325843e-02 -3.41508895e-01 1.76643789e-01 2.52305180e-01
-1.25555301e+00 8.10553789e-01 5.70678139e+00 8.89163554e-01
-1.08312404e+00 2.89229825e-02 5.66166639e-01 -3.17326128e-01
-5.85311413e-01 1.74736410e-01 -5.33075988e-01 4.82769907e-01
7.74389267e-01 -4.44904476e-01 2.71269172e-01 8.28843832e-01
5.08335710e-01 -1.11755230e-01 -1.43324018e+00 1.10999703e+00
5.60738981e-01 -1.47195899e+00 2.68950790e-01 -1.73795298e-01
9.85048413e-01 -4.43455756e-01 -1.79923102e-01 2.76700974e-01
-1.65422067e-01 -6.13774002e-01 8.95970047e-01 5.72922885e-01
5.80728471e-01 -7.96477795e-01 6.21374011e-01 4.10094202e-01
-1.05268884e+00 8.31893645e-03 -2.12084264e-01 2.75665373e-02
3.20796609e-01 4.52053159e-01 -6.62067473e-01 7.89095581e-01
2.45513991e-01 1.14472497e+00 -6.26485705e-01 1.18853807e+00
-1.62207767e-01 7.02930629e-01 -4.60113771e-02 -2.64873691e-02
3.42877835e-01 -8.05790722e-02 8.11760008e-01 1.37857044e+00
3.13560098e-01 6.90288693e-02 7.74100721e-01 2.62266845e-01
-4.00025189e-01 4.98072475e-01 -5.09089589e-01 -1.84590176e-01
2.79472470e-01 1.29569793e+00 -9.99282598e-01 -6.25184119e-01
-4.60789144e-01 1.25370944e+00 2.43464425e-01 3.33137751e-01
-6.81313396e-01 -2.13718489e-01 1.17252236e-02 1.72784626e-01
3.30869108e-01 -7.93000907e-02 7.92677701e-02 -1.47988117e+00
1.04952745e-01 -8.29958737e-01 4.29075301e-01 -4.57009673e-01
-1.06313491e+00 4.46788967e-01 2.98014075e-01 -1.37578821e+00
-3.34339052e-01 7.18157589e-02 -6.59790576e-01 2.24730477e-01
-1.48591292e+00 -1.13128495e+00 -4.08218294e-01 4.25914347e-01
1.06055832e+00 -9.44093522e-03 3.55805546e-01 1.76161677e-01
-5.27517438e-01 3.13293219e-01 1.53931469e-01 -2.43995354e-01
9.56320465e-01 -1.23057079e+00 -4.46994491e-02 8.90300274e-01
8.15709606e-02 3.11376065e-01 1.08549678e+00 -5.69884300e-01
-1.26497459e+00 -1.11959374e+00 8.96200955e-01 -6.11055531e-02
4.66081709e-01 1.01104639e-02 -7.62117863e-01 7.45262086e-01
3.34402412e-01 -2.05454946e-01 7.77286649e-01 -8.17545503e-02
-8.05918202e-02 -1.09454557e-01 -8.10050607e-01 5.61856449e-01
1.03907847e+00 -2.06957072e-01 -6.17346942e-01 4.08703089e-01
9.48501587e-01 1.28036002e-02 -5.85121870e-01 5.59700765e-02
2.61472762e-01 -8.83740366e-01 6.29244268e-01 -6.62626624e-01
7.35618412e-01 -2.51592696e-01 -1.76601231e-01 -1.37967145e+00
-2.72933543e-01 -9.28064585e-01 -3.36458623e-01 1.45549071e+00
5.89288548e-02 -1.38918206e-01 8.71093929e-01 3.13375831e-01
-5.28236747e-01 -5.62604487e-01 -5.55738509e-01 -6.37973845e-01
-5.59132218e-01 1.27740920e-01 2.91517884e-01 8.67515266e-01
3.31889302e-01 4.42492813e-01 -7.20193803e-01 -2.11825445e-01
5.27841926e-01 2.68351853e-01 9.67061996e-01 -1.14908373e+00
-2.35752776e-01 -4.91815209e-01 -4.61777806e-01 -1.22638905e+00
3.32926005e-01 -8.40564609e-01 8.61785412e-02 -1.81223905e+00
7.09713161e-01 -7.20139742e-02 -3.04737031e-01 3.46251369e-01
-3.81390929e-01 7.40461946e-02 3.80328894e-01 3.30613941e-01
-1.35538673e+00 7.38587677e-01 1.08554292e+00 -1.85167983e-01
-4.14323449e-01 3.66542339e-02 -9.05530989e-01 8.45690310e-01
7.09555984e-01 -4.47721809e-01 -6.41589105e-01 -2.77891576e-01
1.17688797e-01 2.22449183e-01 2.02780254e-02 -9.40290689e-01
3.26649368e-01 -3.51154089e-01 -1.18300788e-01 -6.01830363e-01
1.57316685e-01 -5.62974393e-01 -2.83780620e-02 5.25152981e-02
-6.69774055e-01 -1.83133289e-01 -3.76802087e-01 7.14100718e-01
-4.42722589e-01 -5.36877811e-01 5.33794522e-01 -1.62089646e-01
-7.68651187e-01 3.00418019e-01 -4.24919903e-01 1.34342179e-01
1.06914115e+00 -2.44231477e-01 -6.33524805e-02 -6.29019201e-01
-6.22137725e-01 3.17857981e-01 8.01327050e-01 1.73341542e-01
6.03517473e-01 -1.28973269e+00 -9.14238691e-01 -3.46386075e-01
2.46687964e-01 2.37787366e-02 3.76448691e-01 8.73130739e-01
-5.22935212e-01 3.24219674e-01 -1.61936641e-01 -4.86273438e-01
-1.37141836e+00 6.12924039e-01 -3.32369417e-01 -2.22502008e-01
-8.73786688e-01 5.46975732e-01 1.60605669e-01 3.01302761e-01
2.94596881e-01 -1.70232907e-01 -6.02882206e-01 5.47881365e-01
5.86397409e-01 5.87828577e-01 -1.32902056e-01 -7.14557827e-01
-9.08285379e-02 3.34134459e-01 -2.16543138e-01 1.53480554e-02
1.56040967e+00 -4.33639646e-01 3.43284979e-02 6.44669533e-01
1.23096216e+00 2.32748970e-01 -1.45604670e+00 -2.34152168e-01
1.89509809e-01 -4.17931229e-01 -2.26465613e-01 -1.29803807e-01
-8.66455853e-01 4.96791482e-01 -1.99988663e-01 4.21922505e-01
1.29460394e+00 2.32413039e-01 9.30088997e-01 5.73128700e-01
2.20950603e-01 -9.78962362e-01 4.64473069e-01 2.66425580e-01
8.01018298e-01 -1.08130670e+00 2.19626471e-01 -2.66190797e-01
-9.42351997e-01 1.07962155e+00 3.22200209e-01 -3.46813232e-01
8.07704628e-02 -2.82259166e-01 -4.67589110e-01 -5.95207177e-02
-7.71227419e-01 -1.96708068e-01 2.54869550e-01 2.88609385e-01
3.16274524e-01 -5.65316267e-02 -5.73271096e-01 5.51495910e-01
-5.08730039e-02 5.58672138e-02 9.00035679e-01 1.06313360e+00
-9.01015222e-01 -9.73371744e-01 -1.65632069e-01 7.30088532e-01
-4.53141332e-01 7.95136765e-02 -4.25070137e-01 3.86268973e-01
-1.98664233e-01 8.79012465e-01 -1.20379962e-01 -2.60953873e-01
3.01753491e-01 4.90146205e-02 4.32332963e-01 -9.36289012e-01
-4.59411055e-01 2.35797971e-01 1.53851494e-01 -3.83232743e-01
-7.90424943e-01 -8.90702605e-01 -9.96377826e-01 -1.22381829e-01
-1.75448522e-01 3.96569848e-01 1.96207896e-01 9.50932980e-01
2.69137174e-01 5.87241530e-01 8.89831364e-01 -1.27578461e+00
-5.27553320e-01 -7.63455629e-01 -5.39196849e-01 4.61824208e-01
5.98452091e-01 -3.56744260e-01 -3.86769116e-01 5.42380035e-01] | [10.44701862335205, 0.45321765542030334] |
d458f153-f50d-4b47-b3ec-46a63015f4da | associating-objects-with-transformers-for | 2106.02638 | null | https://arxiv.org/abs/2106.02638v3 | https://arxiv.org/pdf/2106.02638v3.pdf | Associating Objects with Transformers for Video Object Segmentation | This paper investigates how to realize better and more efficient embedding learning to tackle the semi-supervised video object segmentation under challenging multi-object scenarios. The state-of-the-art methods learn to decode features with a single positive object and thus have to match and segment each target separately under multi-object scenarios, consuming multiple times computing resources. To solve the problem, we propose an Associating Objects with Transformers (AOT) approach to match and decode multiple objects uniformly. In detail, AOT employs an identification mechanism to associate multiple targets into the same high-dimensional embedding space. Thus, we can simultaneously process multiple objects' matching and segmentation decoding as efficiently as processing a single object. For sufficiently modeling multi-object association, a Long Short-Term Transformer is designed for constructing hierarchical matching and propagation. We conduct extensive experiments on both multi-object and single-object benchmarks to examine AOT variant networks with different complexities. Particularly, our R50-AOT-L outperforms all the state-of-the-art competitors on three popular benchmarks, i.e., YouTube-VOS (84.1% J&F), DAVIS 2017 (84.9%), and DAVIS 2016 (91.1%), while keeping more than $3\times$ faster multi-object run-time. Meanwhile, our AOT-T can maintain real-time multi-object speed on the above benchmarks. Based on AOT, we ranked 1st in the 3rd Large-scale VOS Challenge. | ['Yi Yang', 'Yunchao Wei', 'Zongxin Yang'] | 2021-06-04 | null | http://proceedings.neurips.cc/paper/2021/hash/147702db07145348245dc5a2f2fe5683-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/147702db07145348245dc5a2f2fe5683-Paper.pdf | neurips-2021-12 | ['one-shot-visual-object-segmentation'] | ['computer-vision'] | [ 8.60407054e-02 -2.09144026e-01 -4.53816742e-01 -2.34143749e-01
-1.04400635e+00 -4.33590204e-01 -1.30185768e-01 -1.64996255e-02
-5.37175834e-01 1.05536833e-01 -1.55474976e-01 3.18368152e-02
-1.48185175e-02 -5.97379327e-01 -1.06163085e+00 -5.14012516e-01
-4.49878164e-02 6.35734975e-01 6.88216627e-01 3.22657257e-01
-1.69332951e-01 1.28906429e-01 -1.35974061e+00 3.48008037e-01
5.94025373e-01 1.43885052e+00 3.63457948e-01 6.29855335e-01
-1.18346177e-01 4.35699373e-01 -4.02599752e-01 -7.56310046e-01
4.76654857e-01 -2.31886040e-02 -6.74265444e-01 1.48690954e-01
8.08718681e-01 -4.91507828e-01 -6.45641267e-01 1.07226086e+00
5.37835002e-01 -2.77282357e-01 6.51788235e-01 -1.56485844e+00
-7.94520199e-01 7.73325920e-01 -9.09546614e-01 3.56426507e-01
-7.94548243e-02 4.38898772e-01 1.32235050e+00 -1.09338355e+00
4.04307991e-01 1.21952093e+00 5.95576465e-01 6.88706458e-01
-1.05868137e+00 -9.26239908e-01 4.43296850e-01 3.68777275e-01
-1.51229477e+00 -3.31689507e-01 4.90475595e-01 -3.36553693e-01
7.26695657e-01 2.38363713e-01 8.17153096e-01 7.98742056e-01
-2.29819924e-01 1.51909184e+00 5.47671735e-01 1.44471228e-01
-1.07875884e-01 8.51631910e-02 2.29063019e-01 8.55017722e-01
3.86780381e-01 -3.30679536e-01 -4.45358127e-01 1.80217654e-01
6.08014047e-01 2.63278466e-02 -1.66874468e-01 -2.31671184e-01
-1.44777262e+00 6.70837164e-01 5.18618166e-01 -2.01186184e-02
-1.87959462e-01 5.86269081e-01 6.99744940e-01 9.73335952e-02
2.03893498e-01 -5.42857349e-02 -4.77548897e-01 -3.30743641e-02
-9.78214622e-01 2.16661230e-01 5.25633156e-01 1.43248904e+00
6.30506277e-01 7.89168254e-02 -4.73850489e-01 6.71697617e-01
6.03705049e-01 4.73310053e-01 2.06125200e-01 -8.09332252e-01
9.87244248e-01 5.62737465e-01 -1.84899151e-01 -9.35287893e-01
-1.04631148e-01 -5.74868977e-01 -1.04111552e+00 -2.32898802e-01
3.39236200e-01 -2.01188363e-02 -1.09439003e+00 1.69315374e+00
4.82842147e-01 6.99607730e-01 7.53988996e-02 1.13638282e+00
1.14849019e+00 9.98487234e-01 1.66915283e-01 -5.81229180e-02
1.59658062e+00 -1.42957139e+00 -4.79913652e-01 -4.76209730e-01
5.64171970e-01 -5.89712322e-01 1.00818741e+00 2.00045899e-01
-1.20224226e+00 -8.13515484e-01 -1.04735184e+00 -2.18383044e-01
2.81700888e-03 2.85102308e-01 6.06383860e-01 5.02983868e-01
-7.09558487e-01 3.41238767e-01 -7.46394038e-01 6.78343698e-02
1.00781941e+00 5.60723186e-01 -1.07245833e-01 -1.91564173e-01
-9.57199514e-01 3.70433867e-01 3.32703531e-01 1.07978404e-01
-1.26468658e+00 -9.51026618e-01 -7.50368536e-01 2.43487909e-01
6.62669361e-01 -6.86080575e-01 8.76772404e-01 -6.26190305e-01
-9.71991181e-01 9.31643009e-01 -1.56844527e-01 -4.02096897e-01
5.73327899e-01 -3.53918880e-01 -2.98879981e-01 2.26992726e-01
2.84202158e-01 1.00861347e+00 8.79837751e-01 -1.22251272e+00
-8.52850258e-01 -2.91879117e-01 -8.40013660e-03 1.61052540e-01
-6.25845730e-01 -8.03676713e-03 -1.21617270e+00 -8.11814964e-01
1.97322696e-01 -7.81979561e-01 -5.52298166e-02 5.44239998e-01
-5.28137624e-01 -4.60085779e-01 8.95460427e-01 -5.06449938e-01
1.25292337e+00 -2.38944411e+00 5.24005115e-01 -1.17842712e-01
7.19534099e-01 4.78616178e-01 -2.90763885e-01 -1.67703643e-01
1.39215574e-01 2.11275250e-01 -1.07243039e-01 -6.94674909e-01
2.20497355e-01 1.98809385e-01 -5.73766269e-02 5.49468100e-01
2.21725658e-01 1.31774008e+00 -6.30708635e-01 -9.53349411e-01
-7.77076185e-02 2.12540790e-01 -6.42731428e-01 2.46056050e-01
-1.85923874e-01 -1.27443880e-01 -4.29452926e-01 9.56411600e-01
6.50851607e-01 -6.35629892e-01 -2.34375328e-01 -7.50730515e-01
2.77713895e-01 -9.29289237e-02 -1.36283100e+00 1.63603497e+00
-1.29903793e-01 6.28589988e-01 3.10078282e-02 -1.09853816e+00
6.45610511e-01 1.87945709e-01 6.21440709e-01 -6.24242187e-01
2.40933657e-01 2.74741590e-01 -2.16849566e-01 -5.89328110e-01
2.39927769e-01 1.79064229e-01 -9.80155095e-02 2.20232993e-01
1.88055634e-01 3.04378569e-01 2.77654618e-01 3.57894659e-01
9.61769879e-01 -2.30748594e-01 -2.14582548e-01 -8.46900418e-02
2.39354923e-01 -3.05179894e-01 8.67201746e-01 6.89858615e-01
-4.52812821e-01 5.38737893e-01 4.84183550e-01 -1.98390901e-01
-9.42968071e-01 -1.17033803e+00 7.72605389e-02 1.00284350e+00
7.86757886e-01 -4.72004175e-01 -5.91964602e-01 -9.79667664e-01
2.10265100e-01 2.82430470e-01 -4.75145251e-01 -6.81643561e-02
-7.22668111e-01 -8.03975642e-01 7.17570901e-01 8.36679399e-01
5.87550700e-01 -8.60073447e-01 -5.18214047e-01 2.18773425e-01
-2.35152707e-01 -1.42970812e+00 -9.80646312e-01 -7.30920658e-02
-7.22845256e-01 -9.50464427e-01 -1.01923823e+00 -1.18314397e+00
4.12358195e-01 3.47577870e-01 1.02055776e+00 3.01906690e-02
-3.78322452e-01 2.54915714e-01 -2.61666507e-01 -1.56227291e-01
2.73354948e-02 2.20306516e-01 -4.49420027e-02 2.70525098e-01
1.64928541e-01 -2.94182956e-01 -6.37463272e-01 5.54457068e-01
-8.67338896e-01 1.99598670e-01 7.96755970e-01 7.12480962e-01
8.79677773e-01 -8.36978257e-02 5.96026063e-01 -4.69620675e-01
1.87024977e-02 -6.11207902e-01 -6.31765425e-01 4.80168402e-01
-4.06395435e-01 -1.11665927e-01 4.37905014e-01 -9.21020031e-01
-3.04151863e-01 8.14140812e-02 -1.68213338e-01 -1.02792001e+00
2.01010093e-01 1.85555264e-01 -4.67401892e-01 -1.69797316e-01
1.17462330e-01 4.17648554e-01 -2.23462343e-01 -5.04684865e-01
1.68596596e-01 6.96142733e-01 5.64703584e-01 -5.28489649e-01
1.02177882e+00 3.40782881e-01 -1.26790151e-01 -3.74662519e-01
-9.11290228e-01 -7.21895576e-01 -2.40653425e-01 -2.64248490e-01
1.13375998e+00 -1.20477736e+00 -9.71103907e-01 7.51376033e-01
-1.22836006e+00 -2.97644377e-01 -2.10032791e-01 3.93744618e-01
-4.19864088e-01 4.43689287e-01 -7.27583528e-01 -4.20447975e-01
-5.23165405e-01 -1.45601988e+00 1.22608364e+00 3.12211514e-01
2.63626754e-01 -5.17113328e-01 -3.88591528e-01 5.67999065e-01
2.28496656e-01 -1.78975448e-01 8.10215831e-01 -4.97175545e-01
-1.06224203e+00 -4.87355981e-03 -7.44648874e-01 4.16845292e-01
-1.55292362e-01 -2.02225655e-01 -7.58495808e-01 -4.92362320e-01
-3.32470357e-01 -4.21499014e-01 1.13005781e+00 1.20443113e-01
1.46077824e+00 -2.66142488e-01 -4.76142943e-01 8.66189480e-01
1.40209711e+00 -3.17865573e-02 4.81024325e-01 1.40875593e-01
1.12303674e+00 2.18210116e-01 7.81003237e-01 3.67791981e-01
7.26892233e-01 9.01551306e-01 7.15589583e-01 -3.98935676e-02
-3.24647337e-01 -1.57565445e-01 3.30794901e-01 9.95382845e-01
2.82554358e-01 -5.43440223e-01 -6.35991633e-01 8.85976076e-01
-1.94338942e+00 -6.58265829e-01 -1.90774202e-01 1.91205096e+00
8.45030546e-01 4.51440156e-01 3.09682667e-01 7.03283101e-02
7.43917346e-01 2.88081706e-01 -7.78757751e-01 3.61283541e-01
-1.75885275e-01 -7.23081455e-02 7.39791989e-01 1.20018892e-01
-1.37934399e+00 1.00156021e+00 5.27926731e+00 1.17885792e+00
-9.58200395e-01 3.90151829e-01 6.53428555e-01 -3.06825310e-01
-1.41546831e-01 -2.46388406e-01 -1.30622864e+00 7.68194675e-01
5.31594157e-01 -9.02706534e-02 3.62861514e-01 8.26318741e-01
-3.26923937e-01 2.46537462e-01 -1.09575188e+00 1.36298788e+00
7.73119703e-02 -1.57890546e+00 2.35484049e-01 -8.19735602e-02
6.55427873e-01 1.37592807e-01 1.97524056e-01 4.45680976e-01
-2.56548345e-01 -8.60082448e-01 1.05457985e+00 1.66624457e-01
8.93856704e-01 -4.87548143e-01 5.93004465e-01 1.83939919e-01
-1.78061795e+00 -1.75386295e-01 -3.44035566e-01 5.62478065e-01
4.77839559e-01 4.62109506e-01 -1.33861497e-01 5.46076298e-01
1.03677738e+00 8.89438093e-01 -6.38998687e-01 1.37944567e+00
1.82005957e-01 6.91775322e-01 -6.08894050e-01 6.40390068e-02
4.09325480e-01 1.53368652e-01 5.92731893e-01 1.28919125e+00
2.33097240e-01 6.51925355e-02 3.44329894e-01 7.13532209e-01
-5.37945807e-01 9.76570621e-02 7.88888186e-02 -9.03402343e-02
4.69102770e-01 1.15052831e+00 -7.23164260e-01 -6.93131864e-01
-4.92312670e-01 1.09841239e+00 2.73587853e-01 1.85346261e-01
-1.49025011e+00 -2.27907568e-01 9.10129011e-01 3.41058746e-02
8.89480352e-01 -1.17506839e-01 -2.57180810e-01 -1.28238463e+00
3.55527818e-01 -7.66141355e-01 4.48363334e-01 -5.17726064e-01
-1.27186775e+00 4.88030285e-01 6.52910722e-03 -1.21905565e+00
5.10475814e-01 -5.66870332e-01 -4.08817917e-01 4.48126584e-01
-1.60963571e+00 -1.23136556e+00 -3.24437469e-01 5.07503748e-01
7.07072735e-01 -1.06777303e-01 2.66918391e-01 1.00733173e+00
-8.93996537e-01 1.12526882e+00 -1.33291140e-01 3.74617726e-01
4.80061054e-01 -1.05204678e+00 5.69451630e-01 6.75113618e-01
4.74517077e-01 6.88806176e-02 2.31373012e-01 -5.49972355e-01
-1.69854283e+00 -1.45812106e+00 5.09748340e-01 -1.98496297e-01
5.21033406e-01 -4.14351195e-01 -1.06256175e+00 7.54347324e-01
-1.17121182e-01 5.67057014e-01 4.22796160e-01 -4.15363103e-01
-5.40221870e-01 -4.56826329e-01 -1.02981234e+00 4.92521524e-01
1.39318013e+00 -3.10009718e-01 -2.57867306e-01 3.63932580e-01
1.11304307e+00 -5.98567843e-01 -1.07612014e+00 6.18002176e-01
5.22400320e-01 -5.76039314e-01 1.24020994e+00 -6.23362422e-01
3.39118600e-01 -4.76256877e-01 -3.72823328e-01 -6.54273570e-01
-3.40433329e-01 -5.45406222e-01 -6.67294979e-01 1.35030842e+00
3.37239951e-01 -3.69622439e-01 9.80630338e-01 2.50693709e-01
-2.45926425e-01 -1.26743960e+00 -1.10919392e+00 -9.31946576e-01
-2.35274181e-01 -5.72946370e-01 5.36730826e-01 5.08174181e-01
-6.74703181e-01 2.02383116e-01 -5.74870586e-01 4.56102699e-01
9.08734679e-01 2.17363790e-01 8.06726635e-01 -9.64085281e-01
-5.80158830e-01 -4.69179183e-01 -7.19666004e-01 -1.75806785e+00
5.46819307e-02 -9.28430736e-01 -6.85210824e-02 -1.41367233e+00
4.40522045e-01 -6.78847015e-01 -4.34172064e-01 4.85457331e-01
-4.03144658e-01 6.18703425e-01 6.36070609e-01 2.40029141e-01
-1.17714787e+00 6.87358797e-01 1.31814337e+00 -4.86623228e-01
2.56244149e-02 2.85675060e-02 -5.99600017e-01 4.47893560e-01
5.69768190e-01 -7.33759105e-01 -2.10201696e-01 -8.66413414e-01
1.18642084e-01 -1.36542514e-01 5.36068738e-01 -1.06051421e+00
3.41743022e-01 3.24056260e-02 1.82192326e-01 -7.84101605e-01
5.01927435e-01 -7.92219222e-01 1.29028961e-01 3.92814904e-01
-6.84875846e-02 -2.98601501e-02 3.52078497e-01 7.05312073e-01
-7.49690458e-02 -1.76829025e-01 8.49367917e-01 1.75624013e-01
-9.62217867e-01 9.13879216e-01 1.19856946e-01 4.68296677e-01
1.43271184e+00 -3.21199656e-01 -2.82332033e-01 1.43849580e-02
-6.11645222e-01 6.77691400e-01 1.83403879e-01 6.06223106e-01
8.98168981e-01 -1.36666548e+00 -7.56703794e-01 5.86635694e-02
1.57890365e-01 4.30345565e-01 4.54740644e-01 9.56104279e-01
-3.85464907e-01 1.39411822e-01 5.11871942e-04 -9.29252028e-01
-1.39710271e+00 5.86746335e-01 2.00171441e-01 -1.52583897e-01
-6.90742671e-01 1.20463920e+00 3.41697961e-01 2.19041407e-02
6.08840108e-01 -2.80369580e-01 5.42941317e-03 1.05661117e-01
3.82754534e-01 4.53981370e-01 -3.92853588e-01 -7.15788424e-01
-3.45929563e-01 1.02074850e+00 -2.73230702e-01 2.43055940e-01
1.16696811e+00 -8.54467973e-02 1.54236585e-01 3.17404628e-01
1.50739992e+00 -3.58144701e-01 -1.48315334e+00 -5.00426769e-01
-2.09977135e-01 -5.58319628e-01 -5.41348010e-02 -3.94817621e-01
-1.68865645e+00 8.93564165e-01 7.05430567e-01 -4.31431979e-02
1.05006492e+00 4.24588978e-01 1.32555163e+00 1.71891868e-01
2.39503041e-01 -7.98450768e-01 3.50115716e-01 8.27745050e-02
6.01256251e-01 -1.30134761e+00 -3.20337601e-02 -6.46069050e-01
-7.77518392e-01 8.15201223e-01 9.69343364e-01 3.74221336e-03
6.33510828e-01 2.40239635e-01 -1.68696254e-01 -2.06486583e-01
-6.84213758e-01 -2.06977814e-01 4.86021012e-01 2.50962377e-01
-2.51341462e-01 5.45649976e-02 5.19193336e-02 7.16614425e-01
3.61828744e-01 -1.07841834e-01 8.00727010e-02 5.21815002e-01
-2.60819405e-01 -7.41838694e-01 -4.90633510e-02 7.01893032e-01
-4.77297574e-01 -3.79831642e-02 1.32482231e-01 6.09335780e-01
2.44285390e-01 5.74978471e-01 2.91228406e-02 -5.89426816e-01
3.12160164e-01 -2.90839881e-01 3.34149003e-01 -4.42389429e-01
-5.33130407e-01 5.25627695e-02 -2.12532267e-01 -7.58856118e-01
-2.56990105e-01 -5.86568475e-01 -1.30194628e+00 -1.79111645e-01
-4.85087574e-01 -2.27936700e-01 3.78039181e-01 8.33842933e-01
4.27605152e-01 6.10770643e-01 4.35058892e-01 -8.11206818e-01
-7.12588727e-01 -6.72423303e-01 -3.28274220e-01 3.07188720e-01
3.04973453e-01 -6.60241604e-01 -2.40105614e-01 -9.88353491e-02] | [9.323702812194824, 0.0002035892102867365] |
1a8c6068-944b-4339-b056-f69a536cd95f | giga-ssl-self-supervised-learning-for | 2212.03273 | null | https://arxiv.org/abs/2212.03273v1 | https://arxiv.org/pdf/2212.03273v1.pdf | Giga-SSL: Self-Supervised Learning for Gigapixel Images | Whole slide images (WSI) are microscopy images of stained tissue slides routinely prepared for diagnosis and treatment selection in medical practice. WSI are very large (gigapixel size) and complex (made of up to millions of cells). The current state-of-the-art (SoTA) approach to classify WSI subdivides them into tiles, encodes them by pre-trained networks and applies Multiple Instance Learning (MIL) to train for specific downstream tasks. However, annotated datasets are often small, typically a few hundred to a few thousand WSI, which may cause overfitting and underperforming models. Conversely, the number of unannotated WSI is ever increasing, with datasets of tens of thousands (soon to be millions) of images available. While it has been previously proposed to use these unannotated data to identify suitable tile representations by self-supervised learning (SSL), downstream classification tasks still require full supervision because parts of the MIL architecture is not trained during tile level SSL pre-training. Here, we propose a strategy of slide level SSL to leverage the large number of WSI without annotations to infer powerful slide representations. Applying our method to The Cancer-Genome Atlas, one of the most widely used data resources in cancer research (16 TB image data), we are able to downsize the dataset to 23 MB without any loss in predictive power: we show that a linear classifier trained on top of these embeddings maintains or improves previous SoTA performances on various benchmark WSI classification tasks. Finally, we observe that training a classifier on these representations with tiny datasets (e.g. 50 slides) improved performances over SoTA by an average of +6.3 AUC points over all downstream tasks. | ['Thomas Walter', 'Etienne Decencière', 'Marvin Lerousseau', 'Tristan Lazard'] | 2022-12-06 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 5.76632261e-01 4.36301559e-01 -2.93639779e-01 -1.12296507e-01
-1.43456376e+00 -5.13795555e-01 6.32040381e-01 4.57564414e-01
-5.45201838e-01 9.50025439e-01 3.11621338e-01 -3.26796442e-01
-3.00668534e-02 -8.50373089e-01 -7.98368275e-01 -1.23257041e+00
-1.17111303e-01 6.88404262e-01 4.00506765e-01 -1.15002561e-02
3.43530886e-02 4.43868369e-01 -1.48095667e+00 5.92199266e-01
4.16460156e-01 1.03803742e+00 -7.15443492e-02 1.08524120e+00
-2.09047213e-01 4.11511928e-01 -6.22218668e-01 -1.66575015e-01
5.50387427e-02 -1.33509979e-01 -8.13846409e-01 -1.22872792e-01
4.38706487e-01 9.78422910e-02 -1.87467933e-01 6.46333098e-01
6.92111671e-01 -4.82795894e-01 8.42229784e-01 -9.09657300e-01
-2.82986999e-01 4.22245950e-01 -5.69854140e-01 3.93745452e-01
-2.07420245e-01 1.67822197e-01 1.05761480e+00 -6.85957193e-01
1.00204265e+00 7.12384045e-01 8.49754512e-01 6.43598258e-01
-1.40296113e+00 -5.87724090e-01 -4.78946149e-01 -3.24329168e-01
-1.48952246e+00 -4.32395697e-01 9.46087167e-02 -3.40800494e-01
1.06161785e+00 5.99466801e-01 6.38877690e-01 9.44610417e-01
5.95679939e-01 6.99969709e-01 1.09809732e+00 -4.54378188e-01
2.37756476e-01 9.42995548e-02 1.46410033e-01 6.91898823e-01
5.17134786e-01 -3.41737032e-01 -4.90809053e-01 -3.79213095e-01
6.84001446e-01 8.25468823e-02 -2.11577818e-01 -1.45483180e-03
-1.64073348e+00 7.97389746e-01 4.40891922e-01 6.59738064e-01
2.24320441e-01 2.26718813e-01 7.11714864e-01 2.53818005e-01
8.08227301e-01 7.36804903e-01 -5.37981510e-01 4.08792682e-02
-9.51469660e-01 -1.08321562e-01 4.84804124e-01 4.38745648e-01
7.52825677e-01 -5.44604659e-01 -1.98530987e-01 6.49533927e-01
-2.30263039e-01 2.78244585e-01 8.73970389e-01 -3.01242530e-01
2.29983687e-01 8.15392911e-01 -1.95244998e-01 -6.50174022e-01
-6.55179262e-01 -6.62538409e-01 -1.16761410e+00 -9.91977192e-03
6.46691024e-01 7.72485211e-02 -1.06857979e+00 1.50059247e+00
3.07741702e-01 3.93534571e-01 8.25311244e-03 5.00259757e-01
6.65332556e-01 6.12827718e-01 9.28613171e-02 3.92078497e-02
1.57714999e+00 -5.75353801e-01 -3.94248247e-01 -4.01382297e-02
1.33964288e+00 -5.04597902e-01 1.01497102e+00 2.78210014e-01
-6.37345970e-01 -1.71026587e-01 -1.03906310e+00 -1.97961703e-01
-7.53390968e-01 1.40448719e-01 6.00361526e-01 3.99950922e-01
-1.30308163e+00 3.99322033e-01 -9.31230843e-01 -6.08985901e-01
1.06594181e+00 8.18216860e-01 -8.17284822e-01 -2.18657762e-01
-7.00071633e-01 8.22374463e-01 1.11479469e-01 -2.15380624e-01
-7.09008574e-01 -1.25137913e+00 -6.75445318e-01 1.69728994e-01
4.94223945e-02 -6.63957298e-01 6.18916154e-01 -6.24456108e-01
-1.23103058e+00 1.30817378e+00 -1.16690449e-01 -3.43809873e-01
2.08237723e-01 2.34283313e-01 -1.74593896e-01 9.26984325e-02
9.31047276e-02 7.35772014e-01 3.05010587e-01 -9.01741624e-01
-5.80304205e-01 -4.80820954e-01 -1.80630326e-01 -2.05124825e-01
-7.48948276e-01 -3.40745449e-01 -3.59422266e-01 -4.03646201e-01
-6.90811127e-02 -9.67610180e-01 -3.16384643e-01 2.24231780e-01
-5.62215030e-01 -5.16856089e-02 7.39710689e-01 -3.20668131e-01
1.00198960e+00 -2.12505817e+00 7.77973309e-02 1.55657634e-01
4.86047029e-01 2.86094546e-01 -3.92387807e-01 2.71347821e-01
-1.09180696e-01 3.38007927e-01 -2.41852716e-01 -3.34490836e-01
-2.28034005e-01 2.08572492e-01 9.18454528e-02 7.70699680e-01
3.48582834e-01 1.17974854e+00 -7.52840698e-01 -5.46224654e-01
9.50582996e-02 4.60196495e-01 -4.71411079e-01 2.29789630e-01
-5.71721569e-02 3.96550298e-01 -3.42288464e-01 7.78545082e-01
4.08975959e-01 -7.32463121e-01 2.68515974e-01 -2.90220410e-01
1.33362398e-01 1.35459423e-01 -4.63282764e-01 1.64557457e+00
-5.75545192e-01 7.24063933e-01 -2.95531809e-01 -1.15563345e+00
5.51488996e-01 2.85500675e-01 6.47111416e-01 -3.01997572e-01
2.23232687e-01 3.26351315e-01 5.20263128e-02 -4.25349742e-01
-2.03642935e-01 -2.38196000e-01 -3.72511923e-01 4.11414295e-01
2.01603457e-01 4.09605587e-03 6.87761456e-02 -4.63630147e-02
1.72875750e+00 -5.86939752e-01 5.03489733e-01 -3.89696151e-01
4.55455303e-01 2.75741398e-01 5.23608446e-01 4.56915528e-01
-2.36414541e-02 7.83744872e-01 8.27065468e-01 -7.45038986e-01
-1.19705939e+00 -8.66292179e-01 -4.66939569e-01 1.09833682e+00
-2.70806134e-01 -2.23643601e-01 -6.92151248e-01 -7.52036154e-01
6.94334507e-02 4.45515849e-02 -1.21585488e+00 1.18220774e-02
-4.59850341e-01 -1.45571566e+00 1.09239674e+00 5.22165060e-01
3.74198742e-02 -7.21659124e-01 -4.79455352e-01 1.90550730e-01
1.89619556e-01 -7.89064705e-01 -2.34642133e-01 5.19573212e-01
-6.72990322e-01 -1.17349029e+00 -7.43324459e-01 -6.95472896e-01
1.03150225e+00 1.34023339e-01 1.09388399e+00 4.24639553e-01
-1.07988286e+00 -2.06447437e-01 -2.14643016e-01 -4.41866219e-01
-3.15461457e-01 4.08461034e-01 -1.46933839e-01 -3.93864661e-02
3.63601565e-01 -3.49943459e-01 -7.08725810e-01 -4.96401191e-02
-1.23303890e+00 1.04441941e-01 9.61122990e-01 1.35908735e+00
8.07608783e-01 -2.09562019e-01 6.42024755e-01 -1.41132402e+00
1.83236431e-02 -5.74387729e-01 -3.64635974e-01 3.13156545e-01
-1.91762879e-01 1.88721240e-01 6.61673546e-01 -3.59904557e-01
-6.28187478e-01 3.61397155e-02 -3.18277538e-01 -1.14697397e-01
-1.12807833e-01 5.52626848e-01 3.72182019e-02 -3.33766162e-01
7.24249005e-01 7.75138140e-02 2.59470522e-01 -6.76408932e-02
-5.35383634e-02 6.04900539e-01 3.19474846e-01 -2.48409018e-01
6.02695644e-01 6.71348751e-01 2.22647458e-01 -6.51165664e-01
-9.98422623e-01 -6.56228185e-01 -6.26294613e-01 1.24265969e-01
8.58532548e-01 -7.59112477e-01 -7.53440082e-01 2.14555830e-01
-5.92597663e-01 -7.53955007e-01 -1.99191540e-01 1.39445081e-01
-2.87854135e-01 -5.44778295e-02 -9.31023061e-01 -2.67023236e-01
-4.11780745e-01 -1.01309991e+00 1.64204931e+00 -4.60099168e-02
-3.22437197e-01 -1.14834583e+00 3.11970234e-01 3.02606732e-01
5.26994109e-01 6.79988027e-01 1.36230123e+00 -7.07075477e-01
-3.84842396e-01 -4.49271113e-01 -3.10096234e-01 -2.01591775e-01
4.14750874e-01 -6.16246946e-02 -1.23475409e+00 -4.46611434e-01
-5.67779601e-01 -4.73694026e-01 1.09050107e+00 3.70983064e-01
1.55612195e+00 -2.81717777e-01 -9.50074852e-01 7.44100034e-01
1.86031413e+00 -4.38101053e-01 6.69508874e-01 3.13814610e-01
4.37388539e-01 3.68712693e-01 3.98250401e-01 1.93129241e-01
-2.15833094e-02 4.91845608e-01 3.27673674e-01 -5.45367599e-01
-2.26807803e-01 1.01527758e-01 -1.54285831e-02 6.49443626e-01
-1.33238724e-02 -4.25554901e-01 -1.08690441e+00 7.43533731e-01
-1.46277785e+00 -6.21737957e-01 2.53320131e-02 2.13775420e+00
1.07760990e+00 1.04526959e-01 -1.75177321e-01 3.93331528e-01
3.52528781e-01 -1.31155469e-03 -4.94486719e-01 -1.34134233e-01
-2.41474211e-02 5.68685532e-01 7.96015978e-01 2.13007897e-01
-1.15873873e+00 7.94234216e-01 5.95667934e+00 1.04702747e+00
-1.43486500e+00 3.19926351e-01 1.30492938e+00 -2.75685310e-01
-2.39485279e-01 -4.88357633e-01 -8.11238885e-01 2.73698300e-01
1.26960230e+00 2.29404401e-02 -6.03568703e-02 4.81440485e-01
-8.53945538e-02 -1.22916326e-01 -1.12514043e+00 7.04450011e-01
-3.72622721e-02 -2.02922082e+00 -1.24079166e-02 4.25755799e-01
8.47969592e-01 3.23640764e-01 2.25771412e-01 3.03465307e-01
2.32456431e-01 -1.41938138e+00 -3.42042558e-03 1.80208668e-01
1.41493726e+00 -7.11801171e-01 1.31276977e+00 2.83598453e-01
-8.68597925e-01 1.14394713e-03 -7.38695621e-01 2.77642578e-01
-3.67062926e-01 1.09504271e+00 -1.27889383e+00 3.39479417e-01
5.37558675e-01 5.45569956e-01 -8.26261640e-01 8.40744913e-01
4.12277102e-01 6.99814320e-01 -3.87116760e-01 -4.72816303e-02
2.45510459e-01 4.01106358e-01 -1.17953256e-01 1.54468989e+00
6.32585526e-01 6.00512549e-02 -6.79414123e-02 3.21900845e-01
-3.06830943e-01 1.09865312e-02 -6.51868284e-01 1.58331245e-02
3.26661706e-01 1.62054801e+00 -1.04739070e+00 -4.61002409e-01
-3.34624708e-01 7.89698660e-01 6.28770232e-01 -2.91505810e-02
-6.48791134e-01 -2.32129738e-01 8.56468022e-01 3.46343458e-01
2.10429639e-01 2.14687079e-01 -5.07064104e-01 -8.86959255e-01
-5.15868187e-01 -7.05250502e-01 4.62978482e-01 -2.07096189e-01
-1.53898990e+00 6.02335989e-01 -5.85172176e-01 -1.31766272e+00
4.19199280e-02 -9.81821954e-01 -5.76986313e-01 4.47135985e-01
-1.67585361e+00 -1.37328434e+00 -3.12478155e-01 2.29102507e-01
2.93591082e-01 3.37400846e-02 1.29605782e+00 1.55505717e-01
-5.94395339e-01 7.54578650e-01 4.53577429e-01 4.26906019e-01
7.45881438e-01 -1.24205446e+00 1.23771176e-01 1.03903294e-01
3.39985229e-02 4.25484657e-01 4.17099118e-01 -3.11689615e-01
-1.47190285e+00 -1.48170459e+00 9.96815860e-01 -4.65745360e-01
7.73488045e-01 -7.54610121e-01 -9.40671861e-01 6.16832316e-01
-1.05126575e-01 8.73037934e-01 1.28496587e+00 1.27827097e-02
-3.01536918e-01 -2.81624585e-01 -1.25645256e+00 4.62508261e-01
7.76490569e-01 -3.46951693e-01 -3.66480760e-02 6.38410926e-01
4.15818751e-01 -4.16324764e-01 -1.22376537e+00 3.36754024e-01
5.19403994e-01 -7.73578167e-01 8.64702880e-01 -4.08206463e-01
5.30046344e-01 -2.88938582e-01 7.43535012e-02 -1.22351551e+00
-2.24535614e-01 -3.45579833e-02 4.09736395e-01 9.49167013e-01
7.55124390e-01 -6.98614061e-01 1.12349987e+00 1.95329204e-01
-2.14456782e-01 -1.25816357e+00 -1.30528462e+00 -4.62962091e-01
2.82231063e-01 9.94132236e-02 5.39745688e-01 9.02698696e-01
4.10403609e-01 -3.17679159e-02 3.99505980e-02 1.01877876e-01
4.32179719e-01 5.83486333e-02 7.59792984e-01 -1.21692491e+00
-2.45061398e-01 -3.39247018e-01 -8.02102625e-01 -3.44024837e-01
9.42072570e-02 -1.13594735e+00 2.36676887e-01 -1.37157726e+00
5.45337617e-01 -6.52438581e-01 -5.52521050e-01 7.50874937e-01
-3.39579344e-01 8.78782868e-01 -3.25437963e-01 1.85867906e-01
-5.95166326e-01 -3.14878561e-02 1.14751720e+00 -3.88355315e-01
2.45905727e-01 -4.37049627e-01 -6.60489261e-01 4.14690763e-01
6.82635427e-01 -5.28499186e-01 -1.65672690e-01 -3.83802563e-01
3.81962210e-02 4.16273624e-02 2.11236507e-01 -1.21211088e+00
1.65223613e-01 5.81724681e-02 7.42799938e-01 -1.93722606e-01
2.78978080e-01 -4.78537709e-01 2.24016070e-01 5.33962727e-01
-6.09872162e-01 -3.43885660e-01 3.18979502e-01 5.57320297e-01
-2.28470325e-01 1.65221259e-01 7.85368681e-01 -1.49807140e-01
-1.27809569e-01 5.03270984e-01 -5.11958361e-01 -2.92991579e-01
1.06454086e+00 -1.57872185e-01 -5.43978393e-01 1.70809135e-01
-4.90854174e-01 -1.26103148e-01 4.84539151e-01 -1.19213291e-01
3.24641407e-01 -1.10727930e+00 -7.70666361e-01 2.58558571e-01
5.24460375e-01 3.09240013e-01 1.77398577e-01 1.06830227e+00
-7.23969817e-01 4.98323739e-01 -8.11445713e-02 -7.53593445e-01
-1.25034845e+00 3.49895090e-01 9.15465038e-03 -9.88683939e-01
-6.05577528e-01 1.09158683e+00 4.06780064e-01 -4.44341600e-01
-1.82438269e-01 -5.79545677e-01 6.99965796e-03 -7.05284849e-02
6.26885653e-01 -1.31777182e-01 4.32072431e-01 -3.11398089e-01
-3.71168613e-01 5.34821033e-01 -3.00737441e-01 2.61897683e-01
1.65041637e+00 3.52192760e-01 -2.54406214e-01 4.26577300e-01
1.68324852e+00 -1.78289302e-02 -1.06708896e+00 3.83813940e-02
1.92242995e-01 -3.12662154e-01 -2.26279274e-02 -5.77543676e-01
-1.00747776e+00 9.61809337e-01 4.60525900e-01 -8.99572447e-02
9.96118486e-01 2.16512546e-01 8.16685200e-01 3.11183631e-01
5.25505543e-01 -7.33750045e-01 5.28255962e-02 1.69642866e-01
4.04450029e-01 -1.25505078e+00 1.13611799e-02 -4.60399389e-01
-2.10792094e-01 1.10701478e+00 3.77392918e-01 -2.95027673e-01
5.66317379e-01 8.89252126e-01 1.23065293e-01 -2.99602270e-01
-1.14871323e+00 3.21821608e-02 -6.91722706e-02 4.93388683e-01
6.65294051e-01 1.07303306e-01 2.41547059e-02 5.60659707e-01
-4.04327437e-02 7.95214549e-02 4.59838957e-01 8.19279611e-01
-4.82893825e-01 -1.00740623e+00 -2.26075724e-01 1.10366488e+00
-6.39978290e-01 -3.41809839e-02 -6.76806197e-02 8.56325626e-01
1.53712034e-01 4.64067221e-01 3.24320972e-01 -2.84775138e-01
-2.24665418e-01 2.99151000e-02 2.75997102e-01 -8.36455464e-01
-5.30985057e-01 -3.82944494e-02 -8.37249607e-02 -3.84858251e-01
-3.23311388e-01 -4.41683054e-01 -1.30194545e+00 -2.94924885e-01
-3.55523109e-01 -9.72083211e-02 3.35289955e-01 1.01660335e+00
3.10150653e-01 6.61061525e-01 3.53862792e-01 -8.89006019e-01
-4.78889346e-02 -9.76871908e-01 -6.41003489e-01 3.06190073e-01
4.16916966e-01 -4.33761507e-01 -3.61187190e-01 1.63834602e-01] | [15.11540412902832, -2.963696241378784] |
c650592f-586a-4dac-bd84-d460c3c29a1f | multilingual-part-of-speech-tagging-two | 1401.5695 | null | http://arxiv.org/abs/1401.5695v1 | http://arxiv.org/pdf/1401.5695v1.pdf | Multilingual Part-of-Speech Tagging: Two Unsupervised Approaches | We demonstrate the effectiveness of multilingual learning for unsupervised
part-of-speech tagging. The central assumption of our work is that by combining
cues from multiple languages, the structure of each becomes more apparent. We
consider two ways of applying this intuition to the problem of unsupervised
part-of-speech tagging: a model that directly merges tag structures for a pair
of languages into a single sequence and a second model which instead
incorporates multilingual context using latent variables. Both approaches are
formulated as hierarchical Bayesian models, using Markov Chain Monte Carlo
sampling techniques for inference. Our results demonstrate that by
incorporating multilingual evidence we can achieve impressive performance gains
across a range of scenarios. We also found that performance improves steadily
as the number of available languages increases. | ['Regina Barzilay', 'Benjamin Snyder', 'Tahira Naseem', 'Jacob Eisenstein'] | 2014-01-15 | null | null | null | null | ['unsupervised-part-of-speech-tagging'] | ['natural-language-processing'] | [-3.26615423e-01 5.13939485e-02 -5.55352390e-01 -5.01202464e-01
-1.49064863e+00 -9.34675157e-01 8.58823299e-01 2.57523626e-01
-4.93152320e-01 7.24166453e-01 6.80146575e-01 -6.18288279e-01
1.77286789e-01 -3.73238087e-01 -5.18561721e-01 -4.53821033e-01
-2.38909394e-01 6.84509754e-01 3.55759919e-01 3.69070517e-03
-1.09982029e-01 2.71731168e-01 -9.35973287e-01 1.80652112e-01
5.96351326e-01 2.05672413e-01 2.44633347e-01 5.84240556e-01
-4.47458982e-01 8.33543003e-01 -3.89655679e-01 -4.97762471e-01
2.01704174e-01 -2.60823488e-01 -8.35833609e-01 2.76705056e-01
2.67746747e-01 1.04399614e-01 -1.59870684e-01 1.18925655e+00
1.95988208e-01 3.74602005e-02 6.58674002e-01 -8.18884790e-01
-2.59843200e-01 1.12677121e+00 -5.17828822e-01 1.45155579e-01
4.13831413e-01 -2.66834110e-01 1.38122618e+00 -6.23618901e-01
5.74935913e-01 1.51209271e+00 7.84307063e-01 1.52313774e-02
-1.47120297e+00 -6.17632687e-01 4.79422241e-01 -3.89015883e-01
-1.47971153e+00 -5.76062679e-01 2.86462426e-01 -8.08766365e-01
9.67859030e-01 -1.59116074e-01 3.62687349e-01 7.28716135e-01
1.83161825e-01 7.66447127e-01 1.42264760e+00 -8.67165625e-01
-1.11917751e-02 1.72000214e-01 3.78779143e-01 7.52612054e-01
3.29615831e-01 1.30815908e-01 -5.61470091e-01 -6.61045253e-01
4.38121289e-01 -1.32088155e-01 3.69439036e-01 -3.37451071e-01
-1.02524102e+00 9.64551866e-01 -2.00313672e-01 4.71690208e-01
-1.36802539e-01 1.40373752e-01 2.34797299e-01 -7.97860771e-02
5.98884225e-01 1.96762994e-01 -6.90464079e-01 4.53885198e-02
-1.16715264e+00 -6.69620857e-02 9.30440724e-01 1.00395942e+00
9.14555550e-01 -2.49612957e-01 1.77515626e-01 8.58757257e-01
9.41886067e-01 7.09214568e-01 1.49846196e-01 -1.00857258e+00
3.63494694e-01 -8.75514559e-03 2.77959347e-01 -2.95059919e-01
-8.88122693e-02 -2.41143644e-01 -5.97041193e-03 -2.71772712e-01
6.01164043e-01 -3.56316358e-01 -1.02623749e+00 1.86628163e+00
2.11053461e-01 -9.56440568e-02 1.88794851e-01 3.72244745e-01
3.70988399e-02 6.23004854e-01 5.57949364e-01 -2.83469349e-01
1.67290616e+00 -8.25486243e-01 -8.39364111e-01 -5.01244187e-01
7.12005615e-01 -1.19317126e+00 5.63629031e-01 2.03496590e-01
-8.90354574e-01 -2.45646507e-01 -9.24348593e-01 4.55151051e-02
-2.72314161e-01 -9.52709168e-02 9.71700013e-01 7.34973311e-01
-1.22498739e+00 2.24990413e-01 -1.16974473e+00 -5.31013668e-01
-2.24791273e-01 1.28608182e-01 -2.24408388e-01 -1.43841386e-01
-1.24059415e+00 6.98651314e-01 4.63774383e-01 -2.19717845e-01
-8.04062843e-01 1.39208794e-01 -8.79639685e-01 -4.36982773e-02
1.61313489e-01 -2.96630144e-01 1.69592118e+00 -9.60655630e-01
-1.18011475e+00 8.39554429e-01 -5.15559852e-01 -2.00253800e-01
3.06001678e-02 -4.09853876e-01 -5.69316030e-01 -1.48516344e-02
4.69105065e-01 4.80097383e-01 5.13767719e-01 -1.47532821e+00
-8.77611339e-01 -1.75579160e-01 -5.36787249e-02 1.39463186e-01
-7.19191954e-02 6.52106941e-01 -7.08705723e-01 -5.74481666e-01
2.40606800e-01 -1.10097158e+00 -4.48712528e-01 -5.42210460e-01
-2.33703867e-01 -1.49519876e-01 1.34129107e-01 -8.60041559e-01
1.23638439e+00 -2.01234460e+00 -2.40512431e-01 2.66945928e-01
1.20535828e-01 -1.51583850e-01 -1.23986825e-01 6.84859395e-01
-4.29704934e-02 3.46992135e-01 -3.26121077e-02 -3.52465808e-01
2.94601507e-02 5.05793333e-01 -3.79876375e-01 3.53276730e-01
-5.39655872e-02 6.41974449e-01 -1.19459760e+00 -8.99395943e-01
-2.06088424e-01 3.48373830e-01 -5.44332325e-01 8.12651291e-02
-1.10134125e-01 2.17767894e-01 -3.63654584e-01 4.96261299e-01
4.43614751e-01 -3.04379672e-01 1.18105495e+00 2.34588310e-01
-3.33681047e-01 8.55251551e-01 -1.19169748e+00 1.71521044e+00
-3.62090707e-01 3.72433126e-01 3.34559172e-01 -4.86142606e-01
6.06445074e-01 6.61002696e-01 3.40823144e-01 -2.14545727e-01
-2.64556020e-01 2.73394078e-01 3.19986977e-02 -1.69630542e-01
5.07085621e-01 -5.93974173e-01 -6.74246728e-01 6.71265244e-01
3.63094568e-01 1.14475228e-01 2.96925515e-01 4.59353894e-01
8.41667175e-01 3.32506090e-01 7.10599601e-01 -4.62914139e-01
-1.09980322e-01 1.26430482e-01 7.43447721e-01 1.06017530e+00
-2.93199241e-01 -1.04444146e-01 3.72655183e-01 -1.38849229e-01
-1.14037716e+00 -1.45463192e+00 -2.65545338e-01 1.44201434e+00
-2.71382213e-01 -5.84433973e-01 -5.54398894e-01 -8.81467462e-01
-1.05870351e-01 5.69613218e-01 -1.99886575e-01 6.12426698e-01
-5.58870792e-01 -7.94662416e-01 6.06062889e-01 5.87048829e-01
-1.52688101e-01 -5.99241257e-01 -2.92626359e-02 2.81694770e-01
-3.40010583e-01 -1.29745102e+00 -4.50036228e-01 6.28547668e-01
-8.82525444e-01 -6.49541676e-01 -5.09905040e-01 -9.14196968e-01
4.48238671e-01 2.67310739e-01 1.31659591e+00 -1.74984023e-01
4.21748489e-01 4.82507974e-01 -3.79232705e-01 3.07292938e-02
-8.40790033e-01 1.65401608e-01 2.08073854e-01 -4.21818912e-01
4.75639254e-01 -4.28110898e-01 2.87550464e-02 1.06480725e-01
-7.20729828e-01 -1.69277087e-01 8.70640934e-01 6.63597822e-01
2.79327393e-01 -1.04811929e-01 5.74388765e-02 -1.15693974e+00
3.43286574e-01 -5.81578612e-01 -5.34506321e-01 4.24763381e-01
-4.14380163e-01 4.40638363e-01 1.05252385e-01 -2.68761963e-01
-1.26937330e+00 1.74263522e-01 -1.88409850e-01 1.21006809e-01
-4.16565269e-01 8.52549493e-01 -1.34642154e-01 3.71316075e-01
3.03479761e-01 6.89741224e-02 -3.13547492e-01 -6.31399691e-01
6.33565843e-01 8.16843450e-01 3.83679867e-01 -1.00910699e+00
6.16447508e-01 3.43079060e-01 -3.74293268e-01 -7.92099595e-01
-1.18705773e+00 -9.97657180e-01 -9.20302749e-01 -3.18524055e-02
9.29000854e-01 -1.46126783e+00 -1.45874113e-01 4.29089144e-02
-1.10412931e+00 -3.47806782e-01 1.57014951e-01 7.34436989e-01
-2.00532228e-01 6.71746075e-01 -1.09251511e+00 -1.03006017e+00
3.08507353e-01 -1.01981604e+00 1.16986930e+00 -1.68731928e-01
-3.67998213e-01 -1.44064045e+00 5.68239331e-01 3.45896244e-01
-2.22821385e-01 -3.61293137e-01 8.76010895e-01 -9.35388803e-01
-4.90937322e-01 -1.78057477e-01 1.08143296e-02 -4.58972342e-02
2.24559069e-01 -4.87521999e-02 -9.15820539e-01 -4.71654236e-01
-2.41474077e-01 -2.02523232e-01 7.65771806e-01 2.21313521e-01
1.34689838e-01 -3.37392956e-01 -6.20504558e-01 -1.06463516e-02
1.55816901e+00 1.64343089e-01 2.87418336e-01 1.24441221e-01
6.24858975e-01 6.40150249e-01 5.12834966e-01 2.32411414e-01
6.41025126e-01 7.33966827e-01 -3.23399067e-01 -3.61785553e-02
-4.07743128e-03 -5.92167139e-01 6.44842982e-01 1.60460150e+00
3.04317385e-01 -2.82730877e-01 -1.37307823e+00 8.94394755e-01
-1.80940449e+00 -9.93204057e-01 -8.32196698e-02 2.08112288e+00
1.00012958e+00 2.93651491e-01 2.60650069e-01 -4.60115075e-01
9.14326847e-01 1.68983221e-01 1.93120137e-01 -1.60763845e-01
-2.02364400e-02 1.08141951e-01 6.88902974e-01 1.09012151e+00
-1.16376042e+00 1.21720946e+00 8.09380627e+00 6.60274744e-01
-6.67965412e-01 4.75435823e-01 2.93633580e-01 3.99763167e-01
-4.37634647e-01 7.32223511e-01 -1.23774123e+00 2.86474079e-01
1.25941300e+00 1.82745263e-01 1.74828663e-01 5.32987654e-01
1.08541332e-01 -3.73766720e-01 -7.60353684e-01 2.27270022e-01
-7.13803321e-02 -7.85001278e-01 3.00339591e-02 2.74331570e-01
9.45137024e-01 4.51007187e-01 -2.85329312e-01 1.85094193e-01
1.54049456e+00 -5.69016516e-01 9.14809585e-01 4.10229504e-01
5.56316912e-01 -5.14981270e-01 5.42535961e-01 4.30640638e-01
-1.37662113e+00 1.47727668e-01 -5.84602281e-02 -1.74050495e-01
4.94594574e-01 5.72040439e-01 -1.02600551e+00 5.42942762e-01
4.76055801e-01 3.41927618e-01 -3.52865577e-01 1.00618100e+00
-6.12837791e-01 1.11090088e+00 -3.11621636e-01 2.00842217e-01
3.51023167e-01 2.06991099e-02 4.41881329e-01 1.66139185e+00
1.76887453e-01 -2.41353348e-01 8.75229478e-01 1.98902413e-01
1.87199831e-01 3.56369317e-02 -4.72613156e-01 -2.34691203e-01
8.14848661e-01 1.03570819e+00 -9.59964633e-01 -5.84258854e-01
-8.56085122e-01 6.35505259e-01 4.79304075e-01 4.81624335e-01
-4.16021198e-01 -7.43867233e-02 4.89374578e-01 -1.26781389e-02
6.14362359e-01 -8.05025041e-01 9.90370363e-02 -1.39311600e+00
-4.43053007e-01 -8.10832262e-01 6.67181492e-01 -7.15313911e-01
-1.35536337e+00 3.65212739e-01 3.33379328e-01 -6.98393345e-01
-5.88157713e-01 -5.62029064e-01 -9.62277502e-02 1.05003011e+00
-1.39490199e+00 -1.35441136e+00 5.87830901e-01 1.62630960e-01
3.87862533e-01 4.71406318e-02 9.59575534e-01 2.75312960e-01
-3.26373756e-01 3.17210495e-01 3.83345276e-01 4.94466871e-01
8.83540988e-01 -1.39389575e+00 6.10079348e-01 1.24602246e+00
7.33677506e-01 1.09568560e+00 7.42928743e-01 -1.03565598e+00
-8.75016272e-01 -7.28199184e-01 1.60941434e+00 -6.69416368e-01
1.10952604e+00 -5.92882216e-01 -6.41190171e-01 1.21076679e+00
4.55420971e-01 -2.83285439e-01 1.15123343e+00 9.54012513e-01
-6.90223217e-01 3.62718403e-01 -5.46652317e-01 4.69681352e-01
7.06925690e-01 -1.01475239e+00 -8.26576114e-01 2.05242842e-01
6.94927394e-01 6.41865358e-02 -1.14356661e+00 -7.88083822e-02
6.41768992e-01 -5.88978589e-01 6.23914480e-01 -6.24673843e-01
-1.05013698e-01 -2.98639178e-01 -5.90672970e-01 -1.26492989e+00
-5.33543050e-01 -7.66471982e-01 2.57505208e-01 1.37357950e+00
6.99693084e-01 -5.81145704e-01 3.09409142e-01 2.79681742e-01
-8.37323740e-02 -8.67721662e-02 -6.80370450e-01 -8.26388717e-01
3.73662472e-01 -5.39038181e-01 2.98135459e-01 1.21633768e+00
2.23153159e-01 6.34589434e-01 -2.28272751e-01 5.09817362e-01
6.73595846e-01 4.52017076e-02 5.01999080e-01 -1.31390584e+00
-6.01849258e-01 -4.90615405e-02 -1.40000671e-01 -1.39938545e+00
3.75661433e-01 -9.40636814e-01 4.36582059e-01 -1.41866612e+00
6.87727749e-01 -6.16441667e-01 -3.21427852e-01 7.13559628e-01
-2.53800869e-01 -3.89987715e-02 1.50046423e-01 5.63844740e-01
-8.91313791e-01 -4.81656045e-02 5.84394753e-01 1.31186455e-01
8.72856975e-02 -2.61669576e-01 -7.25977361e-01 8.84907305e-01
4.89001393e-01 -9.33858693e-01 -5.62727116e-02 -4.57056612e-01
3.96042496e-01 3.52131128e-01 3.38205625e-03 -6.03490174e-01
2.07775041e-01 -1.50089145e-01 1.41147226e-01 -5.17158628e-01
1.21879362e-01 -4.39110309e-01 5.52601367e-02 2.74705052e-01
-3.21384728e-01 2.10335493e-01 2.46704251e-01 6.75301373e-01
-2.52874345e-01 -2.32657865e-01 5.43495059e-01 -3.21888030e-01
-7.60084391e-01 -8.34398046e-02 -8.41112494e-01 1.80509850e-01
4.40589815e-01 1.74288005e-01 2.96634734e-02 -3.79188180e-01
-9.66843486e-01 1.04470551e-03 5.76020658e-01 2.45433494e-01
-1.29956111e-01 -1.29821074e+00 -5.25594056e-01 -1.35872096e-01
1.39482737e-01 -7.54734278e-01 -3.00064594e-01 6.85878456e-01
-2.28269756e-01 7.80991316e-01 3.14933151e-01 -6.13714039e-01
-1.11761010e+00 4.67677087e-01 1.56992033e-01 -7.31422961e-01
-1.21385008e-01 5.40048838e-01 2.35470563e-01 -7.94162869e-01
6.15911791e-03 -4.50624414e-02 2.22807098e-02 8.42384025e-02
1.04612991e-01 -5.39353788e-02 -2.53195643e-01 -8.78950715e-01
-3.97418261e-01 2.88602889e-01 -2.61555016e-01 -9.27371442e-01
1.06443763e+00 -4.70414966e-01 -1.92500189e-01 9.09872293e-01
8.23345363e-01 7.63449192e-01 -1.28888261e+00 -5.49090505e-01
5.10012448e-01 -1.98555902e-01 -1.90123711e-02 -8.35713744e-01
-3.87396485e-01 6.82281077e-01 3.20802957e-01 1.45541906e-01
5.07927358e-01 4.49115276e-01 4.29620266e-01 3.85822117e-01
7.01185822e-01 -9.30402815e-01 -2.40531608e-01 7.88628042e-01
2.18726099e-02 -1.05989265e+00 1.54276922e-01 -3.45156193e-01
-5.65809071e-01 6.95253432e-01 -3.78054976e-02 2.35673904e-01
7.90493548e-01 5.55545986e-01 3.75280321e-01 -1.62417516e-01
-7.55945861e-01 -5.46495497e-01 1.06111310e-01 4.10944283e-01
1.01281440e+00 4.49557841e-01 -2.55852968e-01 5.23786843e-01
-1.19358681e-01 -4.83410656e-01 2.60695279e-01 1.20536113e+00
-6.78724229e-01 -1.63639069e+00 -5.08577168e-01 8.44823495e-02
-8.23504150e-01 -5.10544419e-01 -2.36589253e-01 8.67718339e-01
2.69551110e-02 1.06562304e+00 8.16826448e-02 5.16614411e-04
-2.10570008e-01 4.91442770e-01 5.51983953e-01 -9.90339756e-01
-3.21495682e-01 8.23069870e-01 5.54024756e-01 -9.34716314e-02
-6.70542002e-01 -1.28836060e+00 -9.79587436e-01 5.61738275e-02
-4.89127696e-01 7.02858925e-01 6.13372982e-01 1.18877077e+00
9.31923315e-02 1.25376523e-01 3.44722241e-01 -4.42493498e-01
-5.99191010e-01 -9.48977709e-01 -7.35058367e-01 -3.72147374e-02
1.81167707e-01 -6.51294053e-01 -2.35076442e-01 3.34189087e-01] | [10.355708122253418, 9.798215866088867] |
3c63400e-ae9d-4e6c-a684-ccefe0c2d3b9 | textadain-fine-grained-adain-for-robust-text | 2105.03906 | null | https://arxiv.org/abs/2105.03906v3 | https://arxiv.org/pdf/2105.03906v3.pdf | TextAdaIN: Paying Attention to Shortcut Learning in Text Recognizers | Leveraging the characteristics of convolutional layers, neural networks are extremely effective for pattern recognition tasks. However in some cases, their decisions are based on unintended information leading to high performance on standard benchmarks but also to a lack of generalization to challenging testing conditions and unintuitive failures. Recent work has termed this "shortcut learning" and addressed its presence in multiple domains. In text recognition, we reveal another such shortcut, whereby recognizers overly depend on local image statistics. Motivated by this, we suggest an approach to regulate the reliance on local statistics that improves text recognition performance. Our method, termed TextAdaIN, creates local distortions in the feature map which prevent the network from overfitting to local statistics. It does so by viewing each feature map as a sequence of elements and deliberately mismatching fine-grained feature statistics between elements in a mini-batch. Despite TextAdaIN's simplicity, extensive experiments show its effectiveness compared to other, more complicated methods. TextAdaIN achieves state-of-the-art results on standard handwritten text recognition benchmarks. It generalizes to multiple architectures and to the domain of scene text recognition. Furthermore, we demonstrate that integrating TextAdaIN improves robustness towards more challenging testing conditions. The official Pytorch implementation can be found at https://github.com/amazon-research/textadain-robust-recognition. | ['Ron Litman', 'Sharon Fogel', 'Oren Nuriel'] | 2021-05-09 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 2.36877054e-01 -5.53028524e-01 -7.05184415e-02 -6.73256516e-01
-6.56962216e-01 -5.43094516e-01 6.24749064e-01 -1.72660694e-01
-5.35743713e-01 3.54471058e-01 -1.98147725e-02 -3.14827472e-01
-6.69372976e-02 -6.25935256e-01 -8.39563251e-01 -8.93007755e-01
3.43186021e-01 3.52962375e-01 1.82429492e-01 -2.05883775e-02
3.96852523e-01 7.62787104e-01 -1.31355262e+00 5.68091869e-01
5.58633924e-01 1.14318597e+00 -6.33724630e-02 7.40598202e-01
-1.49180189e-01 8.62700224e-01 -8.26015949e-01 -3.94511968e-01
5.54267824e-01 -1.70542419e-01 -5.99112272e-01 2.43367210e-01
8.80068362e-01 -3.88475984e-01 -3.90843749e-01 9.12105441e-01
5.01896024e-01 1.32388070e-01 6.84195876e-01 -1.16139579e+00
-7.88736880e-01 4.34683800e-01 -5.21619916e-01 1.79537863e-01
-1.46902815e-01 3.43139470e-01 9.92825568e-01 -1.19710815e+00
3.11054111e-01 1.10185552e+00 9.45703924e-01 4.43467349e-01
-1.33311462e+00 -5.72133422e-01 2.71781951e-01 1.22084677e-01
-1.29547274e+00 -5.52666485e-01 5.14416873e-01 -4.43008184e-01
1.24342883e+00 3.33927751e-01 1.12436257e-01 1.46003258e+00
4.15068626e-01 1.08666575e+00 9.21395600e-01 -4.74639237e-01
2.08361655e-01 -1.52780443e-01 3.35111529e-01 5.65864623e-01
1.91416025e-01 7.70734400e-02 -5.68794608e-01 3.61249857e-02
5.85862756e-01 1.64125517e-01 -2.27338850e-01 -4.23953265e-01
-1.00252259e+00 5.72834730e-01 3.77953351e-01 2.90447533e-01
-1.84157282e-01 1.35731056e-01 5.01650214e-01 6.27815008e-01
3.25186133e-01 4.58383083e-01 -5.60601473e-01 -1.80801228e-01
-9.47336197e-01 1.14790365e-01 8.84987473e-01 5.79650581e-01
7.46874094e-01 2.73073524e-01 -2.31857240e-01 1.17052722e+00
-7.05095679e-02 5.58319986e-01 6.74681067e-01 -3.64159107e-01
5.89795709e-01 6.38553143e-01 -3.22641283e-01 -1.01880085e+00
-4.09194499e-01 -5.38788080e-01 -1.07375181e+00 4.30465162e-01
7.50193000e-01 -1.81033108e-02 -1.40389740e+00 1.48516822e+00
-1.19963706e-01 -2.20403999e-01 -1.69743240e-01 9.92536306e-01
4.54680175e-01 4.08621848e-01 -1.81033880e-01 4.46638048e-01
8.88702512e-01 -1.01039791e+00 -3.73840988e-01 -5.15719771e-01
6.76410377e-01 -7.67502844e-01 1.38088739e+00 7.02651083e-01
-6.35110140e-01 -3.59620661e-01 -1.20590580e+00 5.49022853e-02
-6.55589819e-01 3.12553942e-01 3.71221840e-01 7.16831267e-01
-8.91056895e-01 7.28328168e-01 -9.86972451e-01 -3.17102611e-01
7.16968060e-01 5.45377076e-01 -4.23408687e-01 -2.55918473e-01
-6.35816038e-01 8.89628649e-01 2.56206334e-01 2.86892027e-01
-7.86488473e-01 -4.47023600e-01 -6.75366759e-01 1.34533182e-01
6.41611218e-01 -1.76922172e-01 1.29923117e+00 -1.40638876e+00
-1.38821077e+00 6.10256135e-01 -3.21821794e-02 -4.79990035e-01
8.11183512e-01 -4.30991232e-01 -3.24136347e-01 -1.38129234e-01
-6.43102974e-02 4.32382494e-01 1.21627474e+00 -9.14400935e-01
-3.54952335e-01 -3.99012297e-01 -4.53347236e-01 -3.01566552e-02
-4.68232453e-01 -6.63513616e-02 -4.72935349e-01 -9.76612031e-01
4.33881767e-02 -7.93884218e-01 -7.13483319e-02 6.83176070e-02
-6.13965571e-01 -1.30440518e-01 1.01026475e+00 -2.72336304e-01
8.80983174e-01 -2.31293488e+00 -1.93925217e-01 3.49991471e-01
1.97548568e-01 4.12336230e-01 -3.78762126e-01 3.41725588e-01
-5.64417653e-02 2.50134826e-01 -2.42304146e-01 -3.17942053e-01
1.96369037e-01 2.80573398e-01 -4.45244998e-01 5.36499381e-01
4.75532711e-01 8.81327808e-01 -4.21310544e-01 4.90616221e-05
3.04139376e-01 3.63078058e-01 -3.95498037e-01 -1.17068090e-01
-1.05091594e-01 -2.72698365e-02 -3.18306059e-01 8.68152797e-01
7.98365593e-01 -4.25594538e-01 -1.98356554e-01 -8.18042234e-02
1.23459816e-01 1.70803815e-01 -1.08316386e+00 1.32265830e+00
-2.75252074e-01 9.05124426e-01 -1.03343628e-01 -1.20765722e+00
1.10586619e+00 -9.07806307e-02 2.22363591e-01 -6.97030604e-01
2.36876443e-01 3.79162014e-01 1.68718114e-01 -2.24472299e-01
3.95482898e-01 1.14915542e-01 7.51647875e-02 5.72126508e-01
2.20837384e-01 -1.51588367e-02 7.95423910e-02 -6.71400800e-02
1.27595556e+00 2.95428000e-02 1.13617696e-01 -2.29845181e-01
3.07136238e-01 -1.28209367e-01 3.79462689e-01 1.30875206e+00
-2.61268556e-01 8.66414368e-01 4.37345594e-01 -8.49659979e-01
-1.16768742e+00 -9.23006237e-01 -3.05333734e-01 1.23132193e+00
-3.09192002e-01 -3.20088893e-01 -6.04964435e-01 -9.38284099e-01
3.52790058e-01 5.41824520e-01 -7.87135243e-01 -2.74024934e-01
-4.44859236e-01 -7.90552378e-01 8.75956595e-01 7.16273963e-01
7.48525083e-01 -1.00987732e+00 -4.12789017e-01 1.31747887e-01
3.44655216e-01 -1.02549624e+00 -5.38560390e-01 6.70485556e-01
-6.44678175e-01 -9.30595636e-01 -7.40202963e-01 -4.97192174e-01
5.90606391e-01 1.84319481e-01 1.04235339e+00 1.34866849e-01
-5.57787538e-01 2.94018894e-01 -4.24047232e-01 -4.38345045e-01
-3.22486132e-01 3.10601920e-01 8.11917037e-02 1.80815682e-01
7.81563401e-01 -3.63946497e-01 -3.80738497e-01 6.51459932e-01
-1.06918073e+00 -2.03811154e-01 7.51545548e-01 1.20500183e+00
3.62699658e-01 -1.43313468e-01 4.03271765e-01 -8.43340993e-01
5.64105093e-01 -2.77173370e-01 -5.27764559e-01 1.99935377e-01
-6.96171343e-01 3.72528210e-02 9.17749345e-01 -6.19759142e-01
-7.44690061e-01 4.26162854e-02 -9.86892730e-02 -5.37370384e-01
-4.33232486e-01 4.44535166e-01 8.21439773e-02 -2.13560298e-01
9.61366594e-01 3.69376004e-01 2.21660361e-01 -5.16925752e-01
-1.00817323e-01 6.77245796e-01 4.11698461e-01 -7.00203776e-01
5.99953771e-01 4.10213768e-01 -2.29290679e-01 -9.59971130e-01
-8.90545189e-01 -2.57434815e-01 -6.24002337e-01 6.46724254e-02
3.61160815e-01 -5.80837131e-01 -5.74792445e-01 1.01332808e+00
-7.69108236e-01 -7.25604057e-01 -2.64494389e-01 2.68636316e-01
-3.84564012e-01 3.25448096e-01 -4.73275453e-01 -5.24650037e-01
-1.55378431e-01 -1.17403495e+00 9.18508589e-01 4.29688357e-02
-1.31202936e-01 -9.22549665e-01 -3.07802290e-01 8.84440690e-02
5.20200193e-01 -2.11747084e-02 5.17667234e-01 -1.20429516e+00
-4.44754481e-01 -4.34051841e-01 -3.08840960e-01 8.65757406e-01
2.62853414e-01 1.56609938e-01 -1.17050815e+00 -3.58186185e-01
-1.60808116e-01 -6.78461254e-01 1.22518873e+00 2.66318828e-01
1.40035856e+00 -3.75976354e-01 -1.82234664e-02 8.64997923e-01
1.22884309e+00 2.50424352e-02 6.25156045e-01 7.32305586e-01
7.21529126e-01 4.32598144e-01 2.50619471e-01 5.61253965e-01
-9.82125998e-02 6.10586345e-01 3.21777552e-01 -2.17934832e-01
-1.06838003e-01 5.43140210e-02 3.18275064e-01 4.64675963e-01
3.29149246e-01 -4.44917619e-01 -1.22841728e+00 3.89930218e-01
-1.83574593e+00 -7.34420180e-01 5.56823723e-02 2.06933808e+00
7.34611630e-01 2.23018229e-01 -4.98122163e-02 1.93662331e-01
6.18184507e-01 1.55181423e-01 -9.10517037e-01 -5.63974082e-01
-5.79345405e-01 2.58178234e-01 7.24142730e-01 2.99585104e-01
-1.13673115e+00 9.72048283e-01 6.03587723e+00 9.78386402e-01
-1.40496564e+00 -1.72673449e-01 8.06266785e-01 -1.00765504e-01
1.27213880e-01 -3.17910135e-01 -7.99388289e-01 3.64114612e-01
5.36820352e-01 1.78151056e-01 4.12892669e-01 8.95708323e-01
-3.13512534e-02 -8.01869407e-02 -1.24392176e+00 9.46593165e-01
1.57665163e-01 -1.39928436e+00 6.95663542e-02 -1.58003829e-02
6.26569986e-01 4.13788825e-01 2.25532055e-01 1.44883916e-01
4.07564193e-01 -1.30924129e+00 7.90985703e-01 3.17673743e-01
7.71040857e-01 -6.13496184e-01 8.11811864e-01 2.29137331e-01
-7.79068291e-01 -2.87508249e-01 -4.60759670e-01 -1.46352239e-02
-5.09191334e-01 6.66486919e-01 -1.00602853e+00 2.00180486e-01
8.32557797e-01 7.40917683e-01 -7.64950454e-01 9.53009367e-01
-1.85023136e-02 8.08087111e-01 -4.80033189e-01 -1.30555987e-01
4.35067326e-01 1.66592136e-01 3.10964882e-01 1.51154602e+00
1.88236445e-01 -3.99823248e-01 1.76066428e-01 8.38357687e-01
-1.00648984e-01 6.86245859e-02 -8.09192657e-01 -6.75735101e-02
3.89586866e-01 9.68307137e-01 -7.76215494e-01 -2.85243988e-01
-3.60122949e-01 1.06755269e+00 3.76296222e-01 4.67521518e-01
-4.99109983e-01 -5.21996558e-01 5.69901407e-01 -2.00236499e-01
5.79542756e-01 -2.08390936e-01 -7.67538309e-01 -1.26987040e+00
3.75485837e-01 -1.21055353e+00 3.59011620e-01 -5.21910727e-01
-1.50664485e+00 5.38156450e-01 -3.43564868e-01 -1.01659930e+00
-8.59981328e-02 -1.18188953e+00 -5.34493327e-01 7.74619222e-01
-1.19877434e+00 -1.02079904e+00 -3.79682004e-01 7.33763158e-01
7.09291101e-01 -3.84147763e-01 6.35342240e-01 1.16964780e-01
-7.95830250e-01 1.11743093e+00 4.17669475e-01 4.86052930e-01
1.00595427e+00 -1.18109691e+00 6.52089417e-01 8.30010653e-01
2.79451698e-01 4.17524040e-01 5.35482466e-01 -4.20856804e-01
-1.36752284e+00 -1.00285900e+00 4.88329798e-01 -4.37058747e-01
7.17905521e-01 -5.83703876e-01 -1.22403228e+00 7.39821017e-01
-6.38215318e-02 2.02398807e-01 4.73620534e-01 2.40846649e-01
-8.46493423e-01 -3.51020515e-01 -9.86827254e-01 6.22818172e-01
8.13616514e-01 -4.76811528e-01 -4.20534492e-01 3.03423196e-01
9.06621143e-02 -4.75353301e-01 -5.02088428e-01 3.58268976e-01
6.97654128e-01 -1.08553469e+00 7.02996075e-01 -5.95981956e-01
5.37228763e-01 -2.71887123e-03 -3.01652610e-01 -1.19956684e+00
-2.13102758e-01 -5.69879889e-01 1.57933712e-01 1.06676745e+00
5.50252974e-01 -9.81833696e-01 9.02119100e-01 4.90075052e-01
-2.31972799e-01 -6.76041007e-01 -9.45905924e-01 -1.13158679e+00
3.35721761e-01 -5.75132847e-01 3.97244155e-01 1.02835226e+00
-2.65842527e-01 1.47097511e-02 -5.08047283e-01 6.78805113e-02
4.63574290e-01 -1.60362646e-01 9.45852280e-01 -9.88836646e-01
-2.63605356e-01 -7.53795445e-01 -6.19151652e-01 -1.11795354e+00
7.17772096e-02 -7.31672525e-01 3.99180561e-01 -8.63454580e-01
5.98749183e-02 -4.89669323e-01 -3.91229719e-01 9.71230268e-01
-7.04059824e-02 3.90383780e-01 3.18699330e-01 3.21848512e-01
-4.43861097e-01 3.40628713e-01 9.92576301e-01 -3.33998382e-01
-8.75684768e-02 -5.84162101e-02 -5.66477299e-01 5.87895274e-01
1.07601655e+00 -4.37605739e-01 -6.26875460e-02 -7.02778399e-01
-3.98624502e-03 -5.48758864e-01 4.81288910e-01 -1.14540339e+00
3.83740604e-01 -1.17835268e-01 6.35806441e-01 -4.41791356e-01
1.93683803e-01 -8.19891334e-01 -3.10933709e-01 3.11586022e-01
-4.11314845e-01 3.46388333e-02 5.25748610e-01 3.01394135e-01
-2.20577225e-01 -2.18352273e-01 8.73792946e-01 2.97620371e-02
-6.81794763e-01 1.55691788e-01 -4.58042353e-01 7.94857517e-02
6.34747028e-01 -5.99055409e-01 -6.06345654e-01 -2.83030778e-01
-4.72954214e-01 8.85265023e-02 4.96043563e-01 4.86619830e-01
6.25457942e-01 -1.07353449e+00 -6.86873257e-01 6.28010631e-01
6.88327178e-02 -3.62214781e-02 1.01272509e-01 8.81038666e-01
-4.96802926e-01 3.36089939e-01 -2.71005094e-01 -9.23436284e-01
-1.21050894e+00 2.24974036e-01 6.37830615e-01 -1.06445327e-01
-7.51918793e-01 8.92456830e-01 2.83860475e-01 -4.51605856e-01
5.38473666e-01 -3.58408779e-01 4.26361680e-01 -2.63150692e-01
5.67793787e-01 1.38413221e-01 5.06009996e-01 -2.50946760e-01
-3.64253193e-01 5.57426989e-01 -6.45855963e-01 1.05510123e-01
1.34438038e+00 1.75887153e-01 1.75975949e-01 4.42949712e-01
1.14197540e+00 -8.12557265e-02 -1.62661576e+00 -3.26676667e-01
1.87383771e-01 -6.13238037e-01 -6.54736310e-02 -1.17254698e+00
-1.09649169e+00 8.85803461e-01 4.97436792e-01 2.31335700e-01
1.22437549e+00 -2.78461128e-01 3.39094311e-01 1.08549392e+00
-5.74021712e-02 -1.16850102e+00 2.65555114e-01 9.89869356e-01
9.14520919e-01 -1.39345849e+00 -1.43667134e-02 1.89402830e-02
-6.76807880e-01 1.47165084e+00 7.44890451e-01 -3.71319085e-01
5.03434300e-01 7.45186567e-01 3.96924436e-01 -2.21581701e-02
-8.54208887e-01 1.13862351e-01 1.48439944e-01 5.52535534e-01
3.55853796e-01 -1.78263575e-01 2.55069435e-01 4.22953486e-01
-6.37650266e-02 -9.48383380e-03 4.54249084e-01 1.06523442e+00
-2.82777548e-01 -1.02826989e+00 -5.25555074e-01 6.77364230e-01
-6.60182536e-01 -2.37944692e-01 -7.88654029e-01 9.41004634e-01
-1.50227278e-01 7.08041131e-01 2.26378798e-01 -5.18868625e-01
3.24799508e-01 1.82144284e-01 2.22086862e-01 -3.87977868e-01
-6.77918136e-01 -3.79162394e-02 -5.31562604e-02 -5.67977548e-01
3.84843275e-02 -6.33092046e-01 -8.99382234e-01 -4.43019032e-01
-3.08441281e-01 -2.40114808e-01 6.82603478e-01 9.42139983e-01
4.07790542e-01 4.72216338e-01 6.67029202e-01 -8.42663586e-01
-1.08908379e+00 -9.51936424e-01 -5.30521274e-01 4.75796312e-01
6.28639460e-01 -4.97025013e-01 -4.64561909e-01 1.95576344e-02] | [11.616665840148926, 2.456207036972046] |
04283cb1-3b71-4b6e-a30a-4d1ecb7a45bb | debiasingword-embeddings-improves-multimodal | 1905.10464 | null | https://arxiv.org/abs/1905.10464v3 | https://arxiv.org/pdf/1905.10464v3.pdf | Debiasing Word Embeddings Improves Multimodal Machine Translation | In recent years, pretrained word embeddings have proved useful for multimodal neural machine translation (NMT) models to address the shortage of available datasets. However, the integration of pretrained word embeddings has not yet been explored extensively. Further, pretrained word embeddings in high dimensional spaces have been reported to suffer from the hubness problem. Although some debiasing techniques have been proposed to address this problem for other natural language processing tasks, they have seldom been studied for multimodal NMT models. In this study, we examine various kinds of word embeddings and introduce two debiasing techniques for three multimodal NMT models and two language pairs -- English-German translation and English-French translation. With our optimal settings, the overall performance of multimodal models was improved by up to +1.93 BLEU and +2.02 METEOR for English-German translation and +1.73 BLEU and +0.95 METEOR for English-French translation. | ['Mamoru Komachi', 'Tosho Hirasawa'] | 2019-05-24 | debiasing-word-embeddings-improves-multimodal | https://aclanthology.org/W19-6604 | https://aclanthology.org/W19-6604.pdf | ws-2019-8 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 3.62000391e-02 -3.09503190e-02 -3.66493374e-01 -2.02604905e-01
-1.08583057e+00 -5.32505929e-01 8.59539807e-01 1.44816697e-01
-8.04322779e-01 8.62177849e-01 2.00792208e-01 -5.11681795e-01
2.06567332e-01 -4.89712387e-01 -5.18595636e-01 -5.68569660e-01
2.96934366e-01 6.48665667e-01 -8.18571597e-02 -3.39306086e-01
1.68221161e-01 6.00568727e-02 -9.57577288e-01 -4.99512739e-02
1.09468472e+00 6.23146176e-01 6.39174208e-02 5.80013871e-01
-4.33916420e-01 -1.02288000e-01 -5.43748915e-01 -8.61569047e-01
2.12248284e-02 -3.64535272e-01 -6.30087554e-01 -2.69308925e-01
4.68060166e-01 -5.45830885e-03 -4.79340315e-01 1.03442490e+00
8.28665078e-01 1.01735562e-01 8.36610615e-01 -1.17147577e+00
-1.31929493e+00 5.18705010e-01 -6.83607340e-01 6.52730018e-02
2.65152991e-01 -2.11248115e-01 1.16759324e+00 -1.30814898e+00
6.26543224e-01 1.29833794e+00 3.91357988e-01 8.32424521e-01
-1.25284982e+00 -4.47092980e-01 -2.34784529e-01 2.15248823e-01
-1.17557609e+00 -2.06369475e-01 4.56646591e-01 -2.45473944e-02
1.22413516e+00 1.36454582e-01 2.92498738e-01 1.51214123e+00
4.96788114e-01 7.94452310e-01 1.09293079e+00 -4.11779344e-01
-1.27162054e-01 4.78610694e-01 7.89237171e-02 5.58396161e-01
2.94024915e-01 -2.44301945e-01 -6.26878858e-01 6.01002062e-03
6.71816766e-01 -2.87815541e-01 -5.41865453e-02 -1.18998893e-01
-1.64904821e+00 9.97549355e-01 3.16824377e-01 4.62575495e-01
-1.73648372e-01 9.65907946e-02 5.17758906e-01 5.26095390e-01
8.01070929e-01 7.17565417e-01 -4.36984479e-01 -3.42627227e-01
-5.75803280e-01 2.49669831e-02 4.69091237e-01 1.01147163e+00
5.84871650e-01 3.97453718e-02 -1.07679479e-01 1.22905827e+00
1.80562288e-01 7.05670655e-01 6.46009624e-01 -3.77098769e-01
9.20527220e-01 5.17425060e-01 2.29648445e-02 -9.73445237e-01
-3.05704594e-01 -2.39499882e-01 -9.59638357e-01 -3.41836810e-01
4.30659026e-01 -3.16772342e-01 -8.78339827e-01 1.84833801e+00
-2.92719295e-03 -1.81978941e-01 3.86479646e-01 1.13341129e+00
7.09006608e-01 1.04520178e+00 3.55609413e-03 -6.86964169e-02
1.44342387e+00 -1.17884505e+00 -9.51841593e-01 -3.42833161e-01
7.15360940e-01 -1.26557207e+00 1.26170850e+00 -4.40970100e-02
-1.21825922e+00 -4.92281139e-01 -8.09209108e-01 -2.61539012e-01
-6.21720076e-01 5.81975318e-02 4.05015916e-01 6.60251498e-01
-9.79069412e-01 1.56604856e-01 -6.85548723e-01 -8.00385654e-01
1.27334908e-01 4.63397235e-01 -7.94088781e-01 -4.19771731e-01
-1.42322373e+00 1.40086770e+00 1.90710537e-02 -2.77120955e-02
-5.39526403e-01 -4.44421172e-01 -7.16207027e-01 -6.82463646e-02
-3.07298247e-02 -6.74386203e-01 9.72075462e-01 -8.00436914e-01
-1.52664506e+00 8.75099838e-01 -4.67693657e-01 -2.07264453e-01
3.18809479e-01 -2.81761914e-01 -6.22535050e-01 -5.43707870e-02
5.35956584e-02 9.91627216e-01 6.94727063e-01 -9.95718360e-01
-3.32857251e-01 -4.28864002e-01 -9.37219858e-02 3.63252610e-01
-8.80018890e-01 1.62709445e-01 -3.93876791e-01 -7.61875153e-01
-1.53030297e-02 -9.27221596e-01 -1.06598057e-01 -1.28900558e-01
-1.62160531e-01 -3.57360274e-01 6.05619192e-01 -6.82753623e-01
1.27128577e+00 -1.92992163e+00 5.77372432e-01 -2.55929857e-01
-5.06374687e-02 3.58815938e-01 -7.31172681e-01 6.40005708e-01
1.31477594e-01 3.99918795e-01 -9.72261876e-02 -5.94375968e-01
2.54590034e-01 3.54970723e-01 -1.55003488e-01 2.58695871e-01
6.42851651e-01 1.29268301e+00 -9.03760195e-01 -3.40473384e-01
4.38138321e-02 5.63051760e-01 -3.08862209e-01 2.93980330e-01
-1.42951636e-02 7.01956823e-02 -1.89367801e-01 7.68558979e-01
7.06950247e-01 1.98945448e-01 1.14431508e-01 -6.24681860e-02
-5.88694401e-02 2.24664837e-01 -5.17151892e-01 1.90159118e+00
-4.81980234e-01 9.63030934e-01 -2.09554881e-01 -7.12336838e-01
1.03710151e+00 4.59137291e-01 1.52424380e-01 -8.32299173e-01
3.48953575e-01 4.79415774e-01 2.20525607e-01 -6.80599988e-01
9.37579751e-01 -5.29154360e-01 -2.24400073e-01 5.88074148e-01
3.56442183e-01 -2.72082407e-02 7.53051192e-02 3.15149291e-03
7.97385097e-01 -1.65373534e-01 -2.37227634e-01 -2.29535729e-01
2.20989943e-01 2.15668961e-01 1.47862494e-01 2.41173252e-01
-2.51627743e-01 7.54821002e-01 3.68097663e-01 -2.37273663e-01
-1.21881557e+00 -1.07723618e+00 -1.07893258e-01 1.26966369e+00
2.91051090e-01 -3.15041959e-01 -7.26696491e-01 -5.25967956e-01
-1.52026728e-01 7.80291438e-01 -5.07278621e-01 -3.84553909e-01
-4.66306686e-01 -1.08972788e+00 8.42537105e-01 5.48180878e-01
3.23171139e-01 -6.28107309e-01 -2.32040137e-02 1.50984123e-01
-3.52359623e-01 -1.07900965e+00 -8.53895187e-01 1.15845986e-01
-1.01355231e+00 -6.36291206e-01 -1.23744869e+00 -1.12120509e+00
6.28138900e-01 3.56303602e-01 9.26580966e-01 -2.48369232e-01
2.00305181e-03 7.63837993e-02 -4.73835886e-01 -2.70284474e-01
-2.69465715e-01 4.17339712e-01 5.52875042e-01 -2.14846376e-02
7.80032814e-01 -1.44577101e-01 -2.80931234e-01 5.14762759e-01
-1.05214059e+00 -1.07326187e-01 5.75583756e-01 1.15172625e+00
1.42985925e-01 -6.96729720e-01 6.73689902e-01 -2.41177946e-01
1.23295069e+00 -4.86629099e-01 -2.07209527e-01 4.08977896e-01
-7.18397677e-01 1.72453433e-01 4.18460429e-01 -9.54777122e-01
-6.92146599e-01 -5.16140819e-01 -9.22416523e-03 -3.25581968e-01
-2.10558455e-02 6.01696014e-01 -3.54202874e-02 1.09620281e-01
4.14043009e-01 2.25152060e-01 -3.30542736e-02 -4.18951213e-01
6.03591621e-01 1.14989543e+00 3.38788897e-01 -5.92491686e-01
7.92254925e-01 -2.14691296e-01 -3.90940279e-01 -7.59747386e-01
-2.71678984e-01 -1.65267989e-01 -6.55871868e-01 1.10876203e-01
1.10333204e+00 -7.98184276e-01 -1.28991470e-01 2.43985638e-01
-1.45733190e+00 5.63042015e-02 1.89102009e-01 7.87987053e-01
-2.27328628e-01 4.35595602e-01 -6.46510482e-01 -6.23947918e-01
-3.73896539e-01 -1.36080241e+00 8.23934615e-01 2.28074163e-01
-4.71099705e-01 -1.17691696e+00 2.68076241e-01 4.74649638e-01
7.30698049e-01 -2.59053081e-01 1.31460321e+00 -7.58905530e-01
-2.40688786e-01 -3.74849021e-01 -4.54130322e-01 1.88199922e-01
2.96075940e-01 -9.84776169e-02 -8.68198395e-01 -3.27816695e-01
-4.22020316e-01 -3.14160258e-01 7.65008509e-01 5.38118482e-02
3.70064467e-01 5.41982651e-02 -1.59556836e-01 3.42622846e-01
1.11399603e+00 1.54916182e-01 6.01375878e-01 3.50140244e-01
4.84112293e-01 4.86502349e-01 5.01720548e-01 -3.53533821e-03
2.53964871e-01 6.23387933e-01 3.82398784e-01 -5.08424044e-02
1.18393395e-02 -2.86183685e-01 6.02336645e-01 1.29062033e+00
5.21661993e-03 -6.73139811e-01 -8.56569946e-01 9.03507888e-01
-1.74846387e+00 -4.51913953e-01 -1.94722086e-01 2.04539037e+00
7.91736484e-01 -1.28337935e-01 -9.11287144e-02 -2.65681893e-01
7.18437135e-01 1.70935661e-01 -3.17598492e-01 -9.69306409e-01
-3.24779958e-01 2.52718657e-01 2.63790697e-01 4.68381464e-01
-7.99504936e-01 1.32788718e+00 5.96373320e+00 6.84356213e-01
-9.66577828e-01 3.46636623e-01 3.36971611e-01 4.87380065e-02
-5.07556140e-01 -1.49635941e-01 -5.59045672e-01 3.00025225e-01
1.08831275e+00 -1.14309810e-01 3.77894461e-01 3.72221589e-01
-3.54791395e-02 4.07982580e-02 -1.11621416e+00 1.15703166e+00
2.92927921e-01 -1.01568842e+00 3.00849169e-01 1.13461517e-01
9.83995616e-01 2.02700421e-02 3.71125817e-01 5.00102043e-01
-1.75196707e-01 -1.28028536e+00 2.16888323e-01 2.94209838e-01
1.03200448e+00 -9.28969443e-01 1.10511184e+00 1.68656841e-01
-6.84243858e-01 2.30454445e-01 -7.64711738e-01 8.01464468e-02
3.23771507e-01 3.10011297e-01 -6.92485571e-01 6.63487911e-01
3.49285752e-01 4.41772372e-01 -4.02290702e-01 7.33828187e-01
-1.19887926e-01 4.30690974e-01 -1.07817322e-01 -4.14422959e-01
6.75031960e-01 -3.87885034e-01 4.17396784e-01 1.22818100e+00
7.35921502e-01 -3.81214768e-01 -3.75619888e-01 6.12461925e-01
-4.41731393e-01 4.15862471e-01 -6.87045455e-01 -6.03033066e-01
2.66964376e-01 1.03714085e+00 -2.03648686e-01 -2.63091803e-01
-4.95382071e-01 1.37560606e+00 3.92526597e-01 4.52193767e-01
-1.01357424e+00 -5.86252928e-01 1.09407210e+00 -1.09707616e-01
-8.71660709e-02 -5.46654999e-01 -2.96911538e-01 -1.40800560e+00
-4.71804366e-02 -7.90791750e-01 -1.05620742e-01 -7.43968546e-01
-1.66522181e+00 9.74483967e-01 -2.55383790e-01 -1.09380519e+00
-4.82350402e-02 -9.52806890e-01 -5.53076208e-01 1.09891510e+00
-1.34960806e+00 -1.08961010e+00 2.50190407e-01 2.72270679e-01
6.98391676e-01 -6.34714723e-01 1.14994895e+00 4.24910128e-01
-7.55510449e-01 9.18739200e-01 4.26738709e-01 9.83825773e-02
1.29232478e+00 -1.05977106e+00 4.66351986e-01 5.65416574e-01
1.97446987e-01 9.46514845e-01 6.22714043e-01 -4.13993239e-01
-1.82723284e+00 -9.01863217e-01 1.62973225e+00 -6.11843526e-01
9.10640836e-01 -4.35366839e-01 -1.04150760e+00 4.27197933e-01
9.39113855e-01 -3.76733661e-01 7.87464082e-01 1.44992650e-01
-5.37319899e-01 3.22631598e-02 -9.05897200e-01 9.78508413e-01
7.80514121e-01 -6.10456347e-01 -7.53791988e-01 2.24003062e-01
7.75635242e-01 -1.04938306e-01 -1.03353286e+00 1.98467538e-01
5.82107961e-01 -3.23488951e-01 8.47621024e-01 -9.25941348e-01
6.07153118e-01 -9.17421132e-02 -4.80454177e-01 -1.65802991e+00
-1.39944792e-01 -4.73923773e-01 1.09606177e-01 1.26966989e+00
8.31739485e-01 -6.76661372e-01 3.08208942e-01 5.11539459e-01
-1.46191895e-01 -8.36089611e-01 -1.28331077e+00 -6.71512127e-01
6.87962055e-01 -3.19464058e-01 3.69196028e-01 1.05872107e+00
3.07366639e-01 7.85066783e-01 -7.37964153e-01 -6.13577068e-02
2.31537640e-01 -1.04839511e-01 7.08386421e-01 -8.53681803e-01
2.15434402e-01 -7.09176958e-01 -2.33548045e-01 -1.25832546e+00
2.03934401e-01 -1.15745509e+00 -1.35044381e-01 -1.79373169e+00
3.44669670e-01 6.75257221e-02 -4.23418045e-01 3.26340884e-01
-3.43986869e-01 4.76499707e-01 9.26029459e-02 2.85084709e-03
-2.30094671e-01 8.58149052e-01 1.17946696e+00 -3.37728620e-01
1.02463439e-01 -4.56201255e-01 -5.97280860e-01 1.38247728e-01
1.14837337e+00 -5.25433302e-01 -2.65984952e-01 -1.20596468e+00
2.33919397e-01 -1.62469655e-01 7.57327527e-02 -5.19147933e-01
3.97314541e-02 -1.26716867e-01 2.59668410e-01 -3.46033692e-01
5.31597435e-01 -7.64610708e-01 -2.94961035e-01 8.39786753e-02
-4.10811901e-01 7.64955223e-01 3.08454275e-01 4.75386739e-01
-3.74767751e-01 -1.06169209e-01 5.72389543e-01 1.93836346e-01
-3.61599177e-01 1.80853605e-01 -6.12814665e-01 -4.77118455e-02
8.45187545e-01 -1.04101092e-01 -4.74297702e-01 -2.51542151e-01
-3.60223264e-01 3.76364022e-01 4.20748591e-01 9.94865596e-01
8.21632087e-01 -1.77949560e+00 -8.06285262e-01 2.33043954e-01
2.67188877e-01 -5.52864909e-01 4.54759561e-02 1.11517513e+00
-3.11323076e-01 5.60618401e-01 -2.83652008e-01 -3.15633446e-01
-1.30931056e+00 4.33738500e-01 3.06691695e-02 -7.90638030e-02
6.10487256e-03 7.48590410e-01 -2.90139973e-01 -7.59450197e-01
2.96787083e-01 -3.80838573e-01 1.67734757e-01 2.51811206e-01
3.05895656e-01 3.15761417e-01 -8.60761479e-02 -7.53144801e-01
-4.40384388e-01 7.65858591e-01 -9.25301760e-02 -3.06442827e-01
1.13530219e+00 -2.28375003e-01 -2.56995291e-01 5.12101054e-01
1.41403377e+00 -2.13813066e-01 -5.34837723e-01 -3.68402302e-01
9.06886458e-02 -4.14695501e-01 -1.53876200e-01 -7.75849342e-01
-6.02518439e-01 1.63607383e+00 6.62723601e-01 -5.02813235e-02
8.40031624e-01 -1.38400868e-01 1.19884205e+00 4.16992217e-01
2.69105494e-01 -1.15259349e+00 1.07797451e-01 9.19732392e-01
6.91170752e-01 -1.47560835e+00 -5.95155656e-01 3.21187675e-01
-7.68787503e-01 1.40345562e+00 5.30588329e-01 2.03852981e-01
5.64237274e-02 -1.83327645e-02 3.60056281e-01 2.52027750e-01
-7.58648932e-01 -1.58568799e-01 4.11905974e-01 5.51387846e-01
8.17163408e-01 7.01370761e-02 -5.64680278e-01 4.86639619e-01
8.77939016e-02 -2.89223164e-01 2.61061996e-01 6.09676600e-01
-4.41109180e-01 -1.57541585e+00 -4.28411514e-01 7.08573014e-02
-3.87640893e-01 -3.18330675e-01 -6.19816124e-01 7.57426798e-01
-1.86723158e-01 1.05439603e+00 4.85129766e-02 -7.26092100e-01
3.13643664e-01 5.12036800e-01 4.90678638e-01 -4.83603060e-01
-4.67873216e-01 2.23304912e-01 5.75111397e-02 -6.35208488e-02
-3.89571667e-01 -2.87082225e-01 -1.01372015e+00 -5.07228076e-01
-3.84256095e-01 2.04694644e-01 6.97343946e-01 6.97680354e-01
4.01501626e-01 2.14859724e-01 3.64639193e-01 -7.64993548e-01
-6.21832788e-01 -1.34655952e+00 -2.98455715e-01 2.48162761e-01
1.05560064e-01 -4.65223342e-01 -2.43520126e-01 -3.38931054e-01] | [11.512602806091309, 10.089245796203613] |
b09dc64b-838b-4ec5-96b3-edcbdfb46c64 | dpc-unsupervised-deep-point-correspondence | 2110.08636 | null | https://arxiv.org/abs/2110.08636v1 | https://arxiv.org/pdf/2110.08636v1.pdf | DPC: Unsupervised Deep Point Correspondence via Cross and Self Construction | We present a new method for real-time non-rigid dense correspondence between point clouds based on structured shape construction. Our method, termed Deep Point Correspondence (DPC), requires a fraction of the training data compared to previous techniques and presents better generalization capabilities. Until now, two main approaches have been suggested for the dense correspondence problem. The first is a spectral-based approach that obtains great results on synthetic datasets but requires mesh connectivity of the shapes and long inference processing time while being unstable in real-world scenarios. The second is a spatial approach that uses an encoder-decoder framework to regress an ordered point cloud for the matching alignment from an irregular input. Unfortunately, the decoder brings considerable disadvantages, as it requires a large amount of training data and struggles to generalize well in cross-dataset evaluations. DPC's novelty lies in its lack of a decoder component. Instead, we use latent similarity and the input coordinates themselves to construct the point cloud and determine correspondence, replacing the coordinate regression done by the decoder. Extensive experiments show that our construction scheme leads to a performance boost in comparison to recent state-of-the-art correspondence methods. Our code is publicly available at https://github.com/dvirginz/DPC. | ['Dan Raviv', 'Shai Avidan', 'Dvir Ginzburg', 'Itai Lang'] | 2021-10-16 | null | null | null | null | ['3d-dense-shape-correspondence'] | ['computer-vision'] | [ 7.51930401e-02 -2.42974255e-02 2.90428489e-01 -2.40736783e-01
-1.10144997e+00 -4.97814596e-01 8.48839998e-01 1.36393920e-01
-1.09576657e-01 4.03526694e-01 -1.66893959e-01 -9.28195491e-02
-1.94204692e-02 -9.36466813e-01 -1.12558734e+00 -4.89212275e-01
1.96792990e-01 1.09025884e+00 5.82995951e-01 -3.60186815e-01
4.47363317e-01 7.93195188e-01 -1.58654857e+00 1.30996734e-01
8.26918364e-01 9.08218384e-01 4.01323348e-01 2.36531988e-01
-3.54070812e-01 3.90419774e-02 -9.49520543e-02 -4.48767751e-01
5.39541900e-01 -2.72251815e-01 -7.49355316e-01 -8.96637440e-02
5.84422112e-01 -4.00161259e-02 -5.13700861e-03 9.19043660e-01
5.57648063e-01 -1.62047371e-01 5.53234994e-01 -1.03381991e+00
-3.68907243e-01 -7.52203837e-02 -6.14690661e-01 -4.06399965e-01
5.34157097e-01 -7.77103752e-02 7.66521394e-01 -1.35623455e+00
7.30406880e-01 1.11362076e+00 1.25818944e+00 3.65914881e-01
-1.53953719e+00 -5.64622343e-01 -3.03185731e-01 -8.31586719e-02
-1.73845208e+00 -4.57447380e-01 8.99296880e-01 -6.41814530e-01
9.59318459e-01 4.10072327e-01 7.04037249e-01 7.86388576e-01
-6.75679594e-02 2.78213799e-01 9.95005012e-01 -4.97767776e-01
2.38872245e-01 -3.92437167e-03 -3.18500042e-01 6.44635499e-01
6.73407968e-03 2.97675937e-01 -2.80149251e-01 -4.25568163e-01
1.02988815e+00 -1.48633197e-01 -1.79746076e-01 -8.03641379e-01
-1.15165234e+00 7.38717735e-01 5.90299129e-01 2.84800291e-01
-2.13927716e-01 1.50933430e-01 1.80758104e-01 2.60675490e-01
7.57563710e-01 3.56206745e-01 -2.93973118e-01 -1.84476182e-01
-1.25488329e+00 4.23219919e-01 6.77826405e-01 1.24296308e+00
1.04436183e+00 -2.91112274e-01 3.84920180e-01 7.31895685e-01
2.89121270e-01 5.49487174e-01 1.68145671e-01 -8.32145929e-01
7.37657547e-01 5.91295660e-01 -7.87220374e-02 -1.38527453e+00
-3.07815671e-01 -3.01395029e-01 -7.69869328e-01 3.76453489e-01
2.98698813e-01 3.64776284e-01 -6.66401625e-01 1.32561004e+00
3.48150581e-01 3.22415113e-01 -1.79028913e-01 7.94404030e-01
6.71804667e-01 5.85029781e-01 -4.51552987e-01 1.76660419e-01
9.08256412e-01 -7.06047952e-01 -2.36936957e-01 4.78819907e-02
4.85659510e-01 -1.12064970e+00 1.01297486e+00 1.35710552e-01
-1.32474589e+00 -5.22757709e-01 -8.64561498e-01 -3.00131500e-01
-3.34475875e-01 1.54951317e-02 2.75232762e-01 1.02073282e-01
-1.32624173e+00 8.71045530e-01 -9.29171801e-01 -3.62567931e-01
3.94541889e-01 5.06188393e-01 -3.45791340e-01 7.77712464e-02
-6.53325677e-01 8.43881547e-01 1.44315258e-01 -1.25159949e-01
-4.78449501e-02 -8.34982872e-01 -8.20737243e-01 -7.17108622e-02
1.24993727e-01 -8.35870326e-01 1.10501242e+00 -7.35954642e-01
-1.54131019e+00 1.04334474e+00 -2.12846309e-01 -1.68338493e-01
8.16457450e-01 -1.01888508e-01 1.15921125e-01 4.66529094e-02
1.80821642e-01 8.68779182e-01 7.17880547e-01 -1.53846169e+00
-3.20634782e-01 -2.90324122e-01 -3.11466932e-01 4.82390001e-02
2.79462516e-01 -9.28453654e-02 -7.56279111e-01 -7.79561639e-01
5.95115840e-01 -1.19925213e+00 -2.07101464e-01 3.51335317e-01
-3.14569175e-01 -1.52111396e-01 8.05480242e-01 -5.20263076e-01
7.32231975e-01 -2.15973330e+00 2.24991262e-01 5.36483586e-01
1.13893755e-01 1.85189962e-01 -2.11327411e-02 6.98236465e-01
-1.42847881e-01 -3.74779291e-02 -5.46387434e-01 -6.84051871e-01
-6.81918189e-02 1.66645542e-01 -4.13766176e-01 6.62616670e-01
2.06033975e-01 9.06367660e-01 -6.42999649e-01 -4.37606037e-01
3.18894923e-01 6.27867103e-01 -7.72653162e-01 9.92043018e-02
-2.65383661e-01 6.85248911e-01 -1.76060632e-01 5.03422737e-01
9.55963612e-01 -3.60492110e-01 -1.66045308e-01 -2.55080462e-01
-4.18032706e-01 2.86517203e-01 -1.16805112e+00 2.10996532e+00
-4.10973936e-01 3.75585884e-01 -8.96856710e-02 -9.64742601e-01
1.25104725e+00 2.55157232e-01 7.18817294e-01 -6.94677234e-01
-3.66567187e-02 6.24132037e-01 -3.18863481e-01 -3.25040035e-02
3.74494433e-01 -1.72018141e-01 9.06505957e-02 2.29678378e-01
-1.72163561e-01 -6.71663105e-01 -2.46814355e-01 -1.55322878e-02
8.10153306e-01 5.61993062e-01 1.77638456e-01 -3.36046785e-01
5.64324021e-01 2.88407266e-01 4.75657135e-01 2.16819257e-01
5.42314053e-01 1.20631242e+00 1.78319857e-01 -4.56736892e-01
-1.37678456e+00 -1.13292217e+00 -3.16146791e-01 2.69896269e-01
3.12650084e-01 -5.83986640e-01 -7.07682669e-01 -1.72166973e-01
1.61310330e-01 5.94996691e-01 -3.00475836e-01 1.68095559e-01
-8.73658597e-01 -1.94185928e-01 2.69232690e-01 5.34895718e-01
3.01370054e-01 -7.17600703e-01 -3.74051929e-01 1.38309091e-01
-1.00640438e-01 -1.14339125e+00 -4.25679117e-01 -7.91532025e-02
-1.19308639e+00 -9.75117028e-01 -6.10850871e-01 -7.24510550e-01
9.93672967e-01 1.96655929e-01 1.27954435e+00 1.30422205e-01
-1.39802814e-01 3.26155365e-01 -1.92802072e-01 -2.31253803e-01
-3.48520130e-01 8.51393640e-02 -3.34130675e-02 -2.20346004e-01
1.27599403e-01 -9.19914782e-01 -4.46147799e-01 5.39670944e-01
-7.10087180e-01 4.38794702e-01 5.28157532e-01 7.10850596e-01
1.07328463e+00 -3.60927194e-01 2.49351293e-01 -8.25087488e-01
2.26470917e-01 -3.36168319e-01 -9.81789291e-01 -3.98427472e-02
-5.56879401e-01 1.14764251e-01 5.15622497e-01 -7.42011368e-02
-6.31304204e-01 5.79731643e-01 -4.50123012e-01 -7.31032073e-01
4.08930853e-02 1.14408433e-01 3.96206789e-02 -3.84629130e-01
6.64277017e-01 1.96526676e-01 4.30072606e-01 -8.63878906e-01
3.14010710e-01 4.00878280e-01 7.22808361e-01 -7.63821661e-01
1.28919673e+00 6.30189538e-01 1.39645413e-01 -7.22946465e-01
-3.78766596e-01 -5.58470130e-01 -1.10200584e+00 -5.80159128e-02
6.53572857e-01 -8.11423600e-01 -4.88332301e-01 2.09507078e-01
-1.59351134e+00 -2.04424456e-01 -4.00509715e-01 3.15613031e-01
-9.05315340e-01 3.75712216e-01 -2.76345611e-01 -2.67555445e-01
-3.55389327e-01 -1.15822959e+00 1.46162522e+00 -2.11450800e-01
-2.55275160e-01 -9.25035238e-01 2.20340431e-01 1.47808626e-01
4.33240414e-01 6.06186688e-01 7.92868078e-01 -2.51455426e-01
-8.90693069e-01 -3.62686872e-01 -2.53070176e-01 9.08480659e-02
1.35357916e-01 -2.63518579e-02 -7.88529754e-01 -3.65517765e-01
-1.06271198e-02 -7.95177072e-02 3.96094888e-01 8.46746713e-02
1.30217683e+00 -2.08904684e-01 -4.80884135e-01 9.11181092e-01
1.68413854e+00 -1.54501021e-01 7.58471429e-01 3.37856770e-01
8.30425203e-01 5.23685098e-01 4.60160911e-01 2.38292798e-01
4.34200883e-01 1.24141014e+00 4.79444325e-01 -4.15594786e-01
-2.16575921e-01 -3.66325974e-01 -2.48624310e-02 1.29047370e+00
-4.32547241e-01 2.21477404e-01 -1.18381917e+00 2.68035382e-01
-1.89665246e+00 -8.27766061e-01 -4.50102597e-01 2.35463142e+00
8.32839191e-01 -1.23907365e-02 2.90032495e-02 1.04454190e-01
5.24748743e-01 -1.95193902e-01 -2.16676787e-01 -3.17608088e-01
1.97505374e-02 5.04766047e-01 4.76101398e-01 6.27020657e-01
-8.65795672e-01 9.25029516e-01 5.89808607e+00 7.48235106e-01
-1.29132092e+00 2.24030629e-01 1.59332737e-01 7.86118358e-02
-3.55285078e-01 1.80032939e-01 -5.97379446e-01 4.71785218e-01
7.04323292e-01 -9.20180827e-02 2.52720118e-01 7.66025066e-01
1.63494021e-01 6.33413866e-02 -1.18715727e+00 1.26740789e+00
-9.91636701e-03 -1.56850684e+00 -1.69302970e-01 1.44923896e-01
5.65234244e-01 3.56659979e-01 -2.19533727e-01 -1.71337485e-01
-3.41365207e-03 -8.93539310e-01 8.98214877e-01 6.57417476e-01
1.08881831e+00 -6.18258476e-01 5.50423980e-01 4.77397412e-01
-1.22165918e+00 5.57273269e-01 -5.62221706e-01 5.62118664e-02
1.60503283e-01 6.47185981e-01 -7.25692213e-01 8.32090974e-01
5.26484370e-01 6.37736320e-01 -6.16790414e-01 1.06177020e+00
1.38848990e-01 2.71022171e-01 -5.82109869e-01 3.90073895e-01
1.27782986e-01 -5.57153881e-01 5.66466153e-01 1.02977657e+00
6.28098071e-01 -3.87813151e-02 2.47734338e-01 1.08842385e+00
1.63400784e-01 1.72646403e-01 -8.77788484e-01 5.02777636e-01
4.31082636e-01 9.65770066e-01 -7.36229181e-01 -1.15678117e-01
-5.08133888e-01 9.92872000e-01 3.75752419e-01 -1.81440730e-02
-8.35034311e-01 -1.58587232e-01 2.92190462e-01 6.63275719e-01
3.51086169e-01 -5.25172830e-01 -6.77960515e-01 -8.47877920e-01
3.27809304e-01 -6.46161020e-01 -9.53387395e-02 -7.17145026e-01
-1.16392016e+00 8.36493313e-01 1.31285205e-01 -1.78764427e+00
-3.43586206e-01 -3.54253858e-01 -4.89439338e-01 1.09898663e+00
-1.43297565e+00 -1.16740251e+00 -5.12492299e-01 5.01864254e-01
3.96934897e-01 -8.44544917e-02 9.30721760e-01 5.03902316e-01
7.13585131e-03 3.85629803e-01 1.74008667e-01 -9.11445022e-02
5.50124764e-01 -1.20999563e+00 8.82972479e-01 4.63487506e-01
1.14498362e-01 5.17937362e-01 5.05440176e-01 -6.85488939e-01
-1.46809995e+00 -1.02096808e+00 1.05946624e+00 -6.12985075e-01
3.61717016e-01 -5.12475789e-01 -1.15913153e+00 4.12515163e-01
-6.97124079e-02 2.62069345e-01 4.14024562e-01 -1.72501296e-01
-1.89563662e-01 -2.19472274e-02 -1.03848040e+00 3.91993731e-01
1.17487776e+00 -4.64312673e-01 -6.30656004e-01 4.00309414e-01
5.49274445e-01 -8.80746007e-01 -8.63874376e-01 5.46654463e-01
3.08053285e-01 -1.00466919e+00 1.08825612e+00 1.75540652e-02
5.44630885e-01 -5.01779258e-01 -1.44074857e-01 -1.19177055e+00
-3.54214013e-01 -6.05432451e-01 1.38994172e-01 1.14751470e+00
3.66408616e-01 -7.12476194e-01 8.15261602e-01 4.23279107e-01
-3.64894122e-01 -1.08602345e+00 -9.37345862e-01 -1.00995195e+00
1.67589322e-01 -3.31005007e-01 6.72670960e-01 1.02451432e+00
-4.72477078e-01 2.11253509e-01 -2.85002198e-02 3.02794039e-01
6.57139361e-01 4.11910534e-01 1.05333674e+00 -1.53125381e+00
-1.07029140e-01 -2.46691331e-01 -5.99902093e-01 -1.07456172e+00
8.29882175e-02 -1.12532890e+00 2.65320204e-02 -1.58334315e+00
-3.07201058e-01 -9.73820746e-01 5.41954637e-01 4.58337754e-01
3.32602233e-01 4.17813241e-01 1.58002988e-01 7.25871205e-01
-3.33598778e-02 6.26180112e-01 1.14044666e+00 9.05412585e-02
-1.92753494e-01 1.38872132e-01 -2.96513736e-01 7.07112551e-01
8.32585037e-01 -5.83329499e-01 -2.53540933e-01 -6.55952156e-01
1.93889484e-01 5.34154065e-02 5.70041955e-01 -1.16121423e+00
4.44118738e-01 6.50830567e-02 1.74406871e-01 -9.71660912e-01
5.91871917e-01 -1.11743128e+00 7.29818285e-01 4.14179057e-01
1.44195423e-01 5.72163761e-01 2.88465500e-01 4.10826832e-01
-3.14056218e-01 -1.87633768e-01 8.72953534e-01 -1.19935818e-01
-2.76544064e-01 3.42122108e-01 2.69396365e-01 -1.38252884e-01
8.92337799e-01 -5.79976499e-01 2.90602390e-02 -1.25551045e-01
-4.90816802e-01 1.61593072e-02 1.00739801e+00 3.53041708e-01
6.40141964e-01 -1.60316825e+00 -8.17390084e-01 3.27419609e-01
3.54165621e-02 6.77928209e-01 -1.65190354e-01 9.10816967e-01
-1.03543758e+00 3.30298543e-01 -2.35927701e-02 -1.08029234e+00
-1.19638240e+00 4.34076250e-01 3.17377478e-01 5.04421182e-02
-1.23948371e+00 5.49961627e-01 4.60393578e-02 -6.68391645e-01
-2.53280159e-02 -2.50335544e-01 2.87671328e-01 -2.15728328e-01
2.74463315e-02 2.27643415e-01 5.27153194e-01 -8.76728773e-01
-3.85835111e-01 1.19461858e+00 1.88277394e-01 -1.34142503e-01
1.53204501e+00 2.32640356e-01 -2.40423948e-01 3.11574548e-01
1.41685200e+00 1.13745943e-01 -1.15297782e+00 -3.96081120e-01
1.71899438e-01 -7.71446586e-01 -7.04371184e-02 -3.31358552e-01
-9.95779574e-01 7.31787860e-01 5.07803321e-01 7.45056570e-02
9.11296010e-01 9.40110087e-02 8.85270894e-01 1.43779263e-01
6.41256392e-01 -9.98225808e-01 -1.83636546e-01 5.39617717e-01
1.35892165e+00 -1.10138834e+00 2.49058723e-01 -9.31503296e-01
-3.27936739e-01 1.17386198e+00 3.39808285e-01 -3.81606847e-01
8.13303292e-01 5.26288033e-01 -3.94341983e-02 -5.09916723e-01
-4.09056246e-01 -7.77416155e-02 4.46259707e-01 4.81188595e-01
3.19828510e-01 -1.00562215e-01 -2.89449722e-01 -7.00213294e-03
-6.05806351e-01 -1.03357986e-01 2.51655690e-02 8.86219501e-01
-1.63127974e-01 -1.50302184e+00 -5.12095392e-01 2.21402407e-01
4.14460339e-03 -1.53604159e-02 -4.64943200e-01 7.40927577e-01
8.71193707e-02 3.52993250e-01 4.14364517e-01 -3.28049839e-01
5.94406426e-01 -1.00265685e-02 5.65670133e-01 -6.72152698e-01
-4.47831750e-01 1.00679718e-01 -1.53564781e-01 -8.57610464e-01
-4.55711305e-01 -8.11045706e-01 -1.31740272e+00 -4.84662831e-01
-2.71625102e-01 1.09323926e-01 8.98100734e-01 6.64404273e-01
7.02758491e-01 1.28478348e-01 8.31545234e-01 -1.36045825e+00
-4.83296186e-01 -5.89676797e-01 -9.73226652e-02 4.78948206e-01
1.54224262e-01 -8.62076461e-01 -1.71181083e-01 7.17510581e-02] | [8.38163948059082, -3.3257265090942383] |
6fe450c1-ffaf-4703-9efa-4e0735d81855 | structured-occlusion-coding-for-robust-face | 1502.00478 | null | http://arxiv.org/abs/1502.00478v2 | http://arxiv.org/pdf/1502.00478v2.pdf | Structured Occlusion Coding for Robust Face Recognition | Occlusion in face recognition is a common yet challenging problem. While
sparse representation based classification (SRC) has been shown promising
performance in laboratory conditions (i.e. noiseless or random pixel
corrupted), it performs much worse in practical scenarios. In this paper, we
consider the practical face recognition problem, where the occlusions are
predictable and available for sampling. We propose the structured occlusion
coding (SOC) to address occlusion problems. The structured coding here lies in
two folds. On one hand, we employ a structured dictionary for recognition. On
the other hand, we propose to use the structured sparsity in this formulation.
Specifically, SOC simultaneously separates the occlusion and classifies the
image. In this way, the problem of recognizing an occluded image is turned into
seeking a structured sparse solution on occlusion-appended dictionary. In order
to construct a well-performing occlusion dictionary, we propose an occlusion
mask estimating technique via locality constrained dictionary (LCD), showing
striking improvement in occlusion sample. On a category-specific occlusion
dictionary, we replace norm sparsity with the structured sparsity which is
shown more robust, further enhancing the robustness of our approach. Moreover,
SOC achieves significant improvement in handling large occlusion in real world.
Extensive experiments are conducted on public data sets to validate the
superiority of the proposed algorithm. | ['Yandong Wen', 'Youjun Xiang', 'Yuli Fu', 'Weiyang Liu', 'Rui Hu', 'Meng Yang'] | 2015-02-02 | null | null | null | null | ['robust-face-recognition', 'sparse-representation-based-classification'] | ['computer-vision', 'computer-vision'] | [ 4.45851892e-01 -1.78664744e-01 -2.89784610e-01 -4.84284967e-01
-6.58343673e-01 -8.36901069e-02 3.31699967e-01 -5.12699902e-01
2.99716681e-01 7.12795377e-01 4.02107835e-01 -2.15355791e-02
-1.76976368e-01 -5.29610276e-01 -5.10142744e-01 -1.08302009e+00
3.67928237e-01 1.47280425e-01 -3.03112596e-01 1.25327915e-01
2.01706171e-01 7.41766751e-01 -1.78477037e+00 1.07550055e-01
9.10084128e-01 1.20143485e+00 1.78627253e-01 -8.99528265e-02
-1.29984394e-01 6.19201779e-01 -4.30923164e-01 -1.00321852e-01
4.70425695e-01 -2.55825371e-01 -3.15268368e-01 7.76493788e-01
6.73289776e-01 -9.14955810e-02 -3.86617184e-01 1.03360474e+00
5.42767644e-01 -1.60081498e-02 5.36062121e-01 -1.03339517e+00
-5.25460958e-01 5.82183264e-02 -9.76395726e-01 1.64840907e-01
4.51046288e-01 -8.22098926e-02 6.62191510e-01 -1.18537307e+00
7.17073143e-01 1.53173280e+00 4.47772503e-01 2.23932475e-01
-1.12999928e+00 -6.95669174e-01 3.02968889e-01 7.85588399e-02
-1.62144291e+00 -9.30856943e-01 1.16907775e+00 -4.63424712e-01
4.06851858e-01 2.69431502e-01 3.12574834e-01 7.85528600e-01
6.99405596e-02 8.11633229e-01 1.31155384e+00 -4.54914510e-01
3.20854634e-01 1.78451449e-01 1.11767977e-01 6.17565989e-01
5.42205334e-01 1.25158966e-01 -5.61742246e-01 -1.99124753e-01
6.53630674e-01 3.59799057e-01 -5.99652588e-01 -5.20777583e-01
-7.64901876e-01 7.59705424e-01 1.37486339e-01 2.52230912e-01
-3.94092172e-01 -2.62411952e-01 2.24213570e-01 2.63673693e-01
4.85376418e-01 -2.05883041e-01 2.59615667e-02 4.73426998e-01
-9.76472318e-01 1.76870637e-03 6.83866382e-01 8.89035225e-01
6.66890562e-01 6.00405991e-01 -1.87735543e-01 1.11668456e+00
5.42770803e-01 5.94987929e-01 4.07778054e-01 -7.36087441e-01
4.87522632e-01 4.61529642e-01 -1.29051253e-01 -1.54096842e+00
-9.23268199e-02 -9.34828162e-01 -1.12449455e+00 2.00938866e-01
1.80593953e-01 2.89204478e-01 -8.49307835e-01 1.61160028e+00
6.11730754e-01 6.11455023e-01 1.34311646e-01 9.41863060e-01
8.14887822e-01 5.53520501e-01 8.96896720e-02 -7.42831767e-01
1.50261223e+00 -7.64695585e-01 -1.07792211e+00 -1.87360987e-01
3.91941369e-01 -9.25559521e-01 4.22003418e-01 5.54456532e-01
-7.05368161e-01 -7.17060328e-01 -1.13766301e+00 9.04671624e-02
-1.58106405e-02 3.25866282e-01 7.45711386e-01 8.03713739e-01
-6.83573365e-01 1.27780050e-01 -4.63095039e-01 -4.03953105e-01
6.59113050e-01 5.16133964e-01 -5.69505870e-01 -5.51951766e-01
-7.37887383e-01 4.88477260e-01 9.87355784e-02 3.44938129e-01
-8.36920619e-01 -4.24785197e-01 -1.04899132e+00 -1.65749174e-02
3.10227096e-01 -2.62252569e-01 4.75717694e-01 -7.96635628e-01
-1.36569440e+00 7.50137031e-01 -7.03605294e-01 -2.65202880e-01
1.95303008e-01 -6.10429514e-03 -5.89987159e-01 1.94963083e-01
5.93795553e-02 4.12864983e-01 1.30128133e+00 -1.47669017e+00
-3.10793132e-01 -5.61214626e-01 -1.72159746e-01 9.07144994e-02
-4.27715510e-01 -1.37924224e-01 -4.48812991e-01 -9.42993164e-01
7.27436185e-01 -8.17095220e-01 -6.28877133e-02 2.14206815e-01
-1.69176400e-01 -1.36941969e-01 1.36304116e+00 -7.06849873e-01
1.11594796e+00 -2.48231530e+00 5.97142018e-02 3.02424133e-01
1.34985819e-01 3.73358846e-01 -1.85094416e-01 2.17949077e-01
-2.48324633e-01 -3.36070091e-01 -3.72800171e-01 -5.56664407e-01
-3.34412456e-01 4.49506670e-01 -5.43327510e-01 9.78911400e-01
1.96331128e-01 5.89237511e-01 -4.24750954e-01 -7.41337240e-01
3.42959851e-01 7.65040219e-01 -6.97075188e-01 1.89435959e-01
2.21694931e-01 7.64050722e-01 -6.16210461e-01 1.20536280e+00
1.09900570e+00 1.44508632e-03 1.20346054e-01 -5.49694777e-01
1.73199587e-02 -2.39460841e-01 -1.76588202e+00 1.73859513e+00
-4.90024388e-01 3.73926163e-01 3.04447532e-01 -1.42583859e+00
1.25498676e+00 4.52085495e-01 5.04944563e-01 -7.05191791e-01
1.42018378e-01 2.70447850e-01 -2.95013338e-01 -5.37597299e-01
-8.59722495e-02 -7.95809478e-02 4.90368456e-01 1.99821234e-01
-8.21103752e-02 2.19493002e-01 1.68186356e-03 -5.56657203e-02
8.28450441e-01 1.43089667e-01 4.16197538e-01 -4.95667100e-01
1.00931358e+00 -3.43773603e-01 9.25420523e-01 3.90526265e-01
-2.61222661e-01 6.40689433e-01 1.93999171e-01 -5.96368611e-01
-5.22808075e-01 -6.19289219e-01 -6.13600314e-01 5.04433751e-01
2.96977699e-01 -5.39699569e-02 -5.24544239e-01 -4.11532283e-01
5.60463518e-02 1.54862687e-01 -4.17152345e-01 1.37854204e-01
-8.65405381e-01 -7.80792534e-01 -1.56913046e-02 2.08474576e-01
6.99868321e-01 -8.11133087e-01 -1.27978057e-01 1.68905601e-01
-6.75392151e-02 -1.11197841e+00 -4.04010147e-01 2.12306418e-02
-8.62997174e-01 -9.65206683e-01 -6.14639819e-01 -1.18998945e+00
9.26466167e-01 7.78742731e-01 7.46874809e-01 1.56237513e-01
-5.02026975e-01 3.44499618e-01 -2.85657257e-01 -1.24168806e-01
3.36338371e-01 -3.34017128e-01 2.17750490e-01 7.59533644e-01
2.66359776e-01 -8.49015772e-01 -7.19356120e-01 4.67429608e-01
-9.35130894e-01 -2.97166198e-01 6.71792150e-01 1.00914192e+00
9.59374130e-01 1.93635210e-01 6.95843101e-01 -9.06263053e-01
2.32260868e-01 -5.39443374e-01 -5.54640710e-01 2.66325712e-01
-4.08479482e-01 -1.35348499e-01 6.22342348e-01 -5.35286248e-01
-1.16950715e+00 4.63381201e-01 -1.11059330e-01 -6.29992783e-01
-1.26395151e-01 2.71786451e-01 -6.02083564e-01 -6.46144331e-01
3.31127048e-01 5.06378591e-01 7.65611678e-02 -7.39622474e-01
1.39930978e-01 7.86068380e-01 3.63080889e-01 -7.24445581e-01
1.14159644e+00 8.51938069e-01 2.73033857e-01 -1.12758064e+00
-8.00020337e-01 -5.26270092e-01 -3.85351092e-01 -1.12376645e-01
4.43229765e-01 -1.14278650e+00 -6.82362437e-01 1.40405193e-01
-9.46185827e-01 4.32069093e-01 -1.40267774e-01 4.41200167e-01
-3.54238808e-01 6.71822429e-01 -1.76888838e-01 -1.08132172e+00
-2.00988561e-01 -1.31977844e+00 1.24821472e+00 1.46922141e-01
1.83836579e-01 -8.06768119e-01 -1.22291885e-01 5.88207603e-01
3.48120660e-01 4.44885075e-01 5.62294424e-01 -3.88857216e-01
-8.06679606e-01 -3.52169245e-01 -3.63381296e-01 2.69621730e-01
2.81340271e-01 -3.92325073e-01 -1.10939527e+00 -5.87049663e-01
7.43126273e-01 -2.56570637e-01 6.77527189e-01 3.80024850e-01
1.33742642e+00 -1.81140125e-01 -3.94471049e-01 7.73727298e-01
1.71248007e+00 3.35984349e-01 7.80054212e-01 -5.68462946e-02
6.90073133e-01 6.76861167e-01 7.43253291e-01 4.95566219e-01
-2.50207204e-02 8.64802122e-01 2.44091496e-01 -1.56276152e-01
-6.06902122e-01 -1.35782912e-01 8.89439806e-02 1.03352809e+00
2.78196603e-01 -1.31538391e-01 -3.73235136e-01 3.16530019e-01
-1.72891724e+00 -8.26520860e-01 -1.73747316e-02 2.15206814e+00
5.90953231e-01 -2.10848033e-01 -3.90197575e-01 4.28557664e-01
7.45328665e-01 4.16716576e-01 -3.91963631e-01 -9.95245799e-02
-4.62472171e-01 2.74677873e-01 1.23872794e-01 5.66267550e-01
-9.76301253e-01 7.18848169e-01 5.62612295e+00 1.28766751e+00
-1.21728039e+00 3.12065929e-01 6.82191432e-01 2.14746058e-01
-1.44596383e-01 1.16930947e-01 -1.18680644e+00 4.99507993e-01
2.06935450e-01 3.10665399e-01 3.22798789e-01 9.19157445e-01
2.78845638e-01 -1.22421980e-01 -9.60475445e-01 1.41838109e+00
4.63522583e-01 -1.06639016e+00 3.42648298e-01 2.20598847e-01
8.32309186e-01 -7.44416118e-01 1.65316358e-01 8.03567320e-02
-2.95021445e-01 -1.03392160e+00 4.97950852e-01 4.25957173e-01
7.58920550e-01 -4.19521779e-01 5.95921755e-01 1.26399070e-01
-1.41779470e+00 1.64622348e-02 -5.97365677e-01 6.31157532e-02
1.02412730e-01 9.91310239e-01 -8.55780020e-02 6.23145044e-01
1.64204746e-01 9.49183166e-01 -3.18081111e-01 1.16047657e+00
-9.94527191e-02 5.13018489e-01 -2.62071520e-01 6.15603209e-01
-8.12838897e-02 -5.10161042e-01 5.66860855e-01 9.75486338e-01
1.67554498e-01 3.96366388e-01 5.41031182e-01 5.87701440e-01
-1.28124626e-02 4.11512971e-01 -8.81965399e-01 4.71367329e-01
4.77147520e-01 1.17779410e+00 -4.82563615e-01 -1.47974581e-01
-5.87204158e-01 6.60896301e-01 2.99074173e-01 5.36299646e-01
-6.94290638e-01 -1.59293368e-01 5.12182295e-01 -6.67092809e-03
4.66784716e-01 -1.71543121e-01 -3.61918449e-01 -1.27617610e+00
3.40225637e-01 -1.07385683e+00 2.57724881e-01 -2.34536394e-01
-1.05343568e+00 5.73269963e-01 -2.56521404e-01 -1.53090644e+00
4.27202016e-01 -4.19106811e-01 -4.36273336e-01 7.64976561e-01
-1.92245257e+00 -1.17303479e+00 -4.43498731e-01 7.29114473e-01
9.64707196e-01 -2.76719719e-01 5.66005647e-01 8.70173931e-01
-9.74525452e-01 6.37214720e-01 1.15674309e-01 -1.87498733e-01
6.60346270e-01 -4.46085304e-01 -4.49223399e-01 9.36263263e-01
2.46025831e-01 7.33603418e-01 5.56455553e-01 -6.63021982e-01
-1.85657394e+00 -1.17192280e+00 7.39609420e-01 8.38400871e-02
2.13902622e-01 -2.74745107e-01 -9.23411131e-01 4.01613176e-01
-1.41961515e-01 3.76509279e-01 7.57652879e-01 -2.03287810e-01
-5.34552038e-01 -5.23257554e-01 -1.35414410e+00 2.78963566e-01
1.18337047e+00 -4.11122650e-01 -4.58507091e-01 6.84371233e-01
3.60602289e-01 -2.86381900e-01 -8.08691800e-01 5.90455055e-01
5.48379183e-01 -7.84335792e-01 1.12782109e+00 -1.09390810e-01
1.63597107e-01 -6.03072584e-01 -6.70748174e-01 -6.78256392e-01
-3.13297361e-01 -5.02483010e-01 -3.42624784e-01 1.52287376e+00
-1.81826204e-01 -9.05785620e-01 9.32115138e-01 2.73206562e-01
-6.88632810e-03 -9.64148521e-01 -1.13772643e+00 -1.07505560e+00
-4.72920805e-01 -6.42712116e-02 4.75013494e-01 9.34131980e-01
-3.78785461e-01 4.46438789e-02 -7.73104131e-01 4.02795106e-01
1.09134650e+00 3.76032114e-01 5.96384108e-01 -1.17721760e+00
-1.84859142e-01 9.81301516e-02 -4.82942730e-01 -1.39315665e+00
3.85686398e-01 -7.38584399e-01 -9.82683152e-02 -1.11218476e+00
1.71765357e-01 -7.85419464e-01 -1.31284177e-01 2.74150848e-01
-2.03759857e-02 4.74386066e-01 -3.29068750e-02 6.02503121e-01
-4.67032135e-01 7.97423661e-01 1.20694005e+00 -3.22887510e-01
-7.68384561e-02 -2.13035822e-01 -7.38712490e-01 5.06917238e-01
5.85006952e-01 -4.97991085e-01 -5.27367175e-01 -5.65686107e-01
-5.20445824e-01 6.46078661e-02 9.48507935e-02 -1.20077837e+00
2.11940110e-01 2.91975383e-02 3.74608338e-01 -5.03797710e-01
5.20205081e-01 -1.12893999e+00 4.34514612e-01 5.41720212e-01
3.11012864e-02 -5.02459228e-01 -2.83377152e-02 8.51092935e-01
-5.45507431e-01 -2.99856424e-01 1.00987613e+00 7.80989602e-02
-5.82229018e-01 4.37742919e-01 6.56446889e-02 -2.12636754e-01
1.08760834e+00 -6.24534488e-01 1.87094510e-03 -1.03986293e-01
-5.69523990e-01 1.45572964e-02 1.40813380e-01 2.07298145e-01
6.80606842e-01 -1.52146387e+00 -6.58419788e-01 5.67428768e-01
-4.40729931e-02 -2.37066343e-01 2.55660951e-01 9.66021299e-01
-3.86647344e-01 6.32416248e-01 -3.37752886e-02 -6.94364905e-01
-1.20623755e+00 6.96501374e-01 -4.05559689e-02 -5.83600923e-02
-6.63997352e-01 6.60689056e-01 4.87134159e-01 -1.40217729e-02
5.99351883e-01 1.21358603e-01 -4.57733214e-01 9.37065035e-02
6.75647974e-01 2.12255165e-01 6.45578057e-02 -9.05274749e-01
-3.76420647e-01 1.10418940e+00 -8.87448490e-02 4.16672707e-01
1.29893327e+00 -2.19944656e-01 -2.28510782e-01 -8.03219974e-02
1.30109990e+00 2.57174313e-01 -1.00775361e+00 -5.11250734e-01
-2.28054121e-01 -9.82510269e-01 1.22856110e-01 -1.32790491e-01
-1.52650034e+00 6.52071416e-01 7.73917794e-01 -1.32123381e-01
1.29447293e+00 -4.17974055e-01 8.31047118e-01 1.48293078e-01
6.01024866e-01 -5.48725963e-01 1.86560601e-02 2.25297585e-02
1.11167908e+00 -1.37999225e+00 3.56555223e-01 -1.11755955e+00
-1.92062706e-01 8.58537972e-01 4.48048949e-01 -1.95955053e-01
8.92655194e-01 2.10545920e-02 -1.31881535e-01 -3.48551840e-01
-3.59882265e-01 -2.13121772e-01 1.35573044e-01 5.58888078e-01
1.57057717e-01 -1.65950090e-01 -5.59869766e-01 2.29509816e-01
2.29146793e-01 -8.25500488e-02 4.46792729e-02 9.36516225e-01
-5.17548323e-01 -1.18435276e+00 -7.72553682e-01 3.40638459e-01
-2.08872557e-01 4.99573573e-02 1.26586303e-01 3.72209281e-01
3.61514866e-01 1.16970789e+00 -3.23828608e-01 -7.33426437e-02
2.13121638e-01 -7.35160336e-02 2.70189881e-01 -6.36981070e-01
-1.57189474e-01 2.74757296e-01 -1.72698304e-01 -5.20371735e-01
-5.01247227e-01 -6.99027658e-01 -6.16880417e-01 1.59160476e-02
-4.73686486e-01 1.74417630e-01 6.06914163e-01 8.35196316e-01
3.84366304e-01 2.63775855e-01 9.60294247e-01 -7.12777078e-01
-3.65879595e-01 -6.06850028e-01 -7.63161421e-01 5.06580234e-01
2.00575292e-01 -1.05625951e+00 -6.03693724e-01 9.81736183e-02] | [12.535969734191895, 0.4053519368171692] |
d948ded0-110d-413d-875e-ba308b794ba2 | head-and-eye-egocentric-gesture-recognition | 2201.11500 | null | https://arxiv.org/abs/2201.11500v2 | https://arxiv.org/pdf/2201.11500v2.pdf | Head and eye egocentric gesture recognition for human-robot interaction using eyewear cameras | Non-verbal communication plays a particularly important role in a wide range of scenarios in Human-Robot Interaction (HRI). Accordingly, this work addresses the problem of human gesture recognition. In particular, we focus on head and eye gestures, and adopt an egocentric (first-person) perspective using eyewear cameras. We argue that this egocentric view may offer a number of conceptual and technical benefits over scene- or robot-centric perspectives. A motion-based recognition approach is proposed, which operates at two temporal granularities. Locally, frame-to-frame homographies are estimated with a convolutional neural network (CNN). The output of this CNN is input to a long short-term memory (LSTM) to capture longer-term temporal visual relationships, which are relevant to characterize gestures. Regarding the configuration of the network architecture, one particularly interesting finding is that using the output of an internal layer of the homography CNN increases the recognition rate with respect to using the homography matrix itself. While this work focuses on action recognition, and no robot or user study has been conducted yet, the system has been designed to meet real-time constraints. The encouraging results suggest that the proposed egocentric perspective is viable, and this proof-of-concept work provides novel and useful contributions to the exciting area of HRI. | ['V. Javier Traver', 'Javier Marina-Miranda'] | 2022-01-27 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 1.76952899e-01 3.76755223e-02 -2.22782254e-01 -2.86521643e-01
-1.32916570e-01 -5.18357493e-02 8.50412667e-01 -5.48704684e-01
-5.88699460e-01 3.02751482e-01 3.39408964e-01 1.00696906e-01
-6.97721094e-02 -4.37821805e-01 -4.22030956e-01 -7.48895347e-01
-6.85744807e-02 2.83374876e-01 -1.32154405e-01 -1.74499825e-01
4.70402032e-01 9.60036457e-01 -1.49346840e+00 6.64549842e-02
-4.03615367e-03 7.95273066e-01 2.07059935e-01 8.44657302e-01
1.91858679e-01 8.54849517e-01 -4.61178929e-01 -7.19720591e-03
1.02276810e-01 -4.61846769e-01 -8.84861529e-01 2.54897505e-01
3.75986636e-01 -6.32543087e-01 -7.60339737e-01 7.41122127e-01
6.73366964e-01 4.89778012e-01 3.96442831e-01 -1.26719677e+00
-2.61935771e-01 8.49277973e-02 -3.94447565e-01 -1.45388329e-02
6.99853480e-01 3.78303736e-01 7.57810116e-01 -8.62458050e-01
8.36193860e-01 1.33855987e+00 2.67950058e-01 5.18654168e-01
-7.51481235e-01 -3.82510245e-01 2.06559584e-01 2.91644216e-01
-1.36988771e+00 -4.83676374e-01 7.82796800e-01 -3.98602307e-01
1.27629900e+00 7.96314627e-02 8.69970858e-01 1.36177063e+00
3.75246644e-01 1.07655954e+00 7.86009848e-01 -5.43412805e-01
1.29800349e-01 -2.03857735e-01 1.21694878e-01 5.94835877e-01
-8.13534632e-02 1.89715654e-01 -7.67848313e-01 2.91559637e-01
1.15278935e+00 2.23415688e-01 -3.41813892e-01 -6.75813138e-01
-1.39451790e+00 5.80534220e-01 5.57892561e-01 7.40320086e-01
-6.52269483e-01 4.35400367e-01 3.79738092e-01 2.43868902e-01
7.52330348e-02 3.47493768e-01 3.01802531e-02 -4.22003984e-01
-6.74573004e-01 -9.62913632e-02 7.20166028e-01 8.92872691e-01
4.16536599e-01 7.78576508e-02 1.30843548e-02 4.39300001e-01
3.85560334e-01 3.97951603e-01 5.67292571e-01 -1.02332973e+00
4.98460829e-01 4.69641715e-01 -2.81493962e-02 -1.32835579e+00
-8.11757326e-01 -8.52483287e-02 -5.87410986e-01 2.63583332e-01
6.02785647e-01 3.72636772e-04 -6.70678258e-01 1.58473146e+00
1.35434389e-01 -1.47228122e-01 1.27998754e-01 1.37905180e+00
6.28571928e-01 2.94544607e-01 1.01090468e-01 -9.23788399e-02
1.35398066e+00 -9.43022907e-01 -8.63100290e-01 -2.05635414e-01
5.60821295e-01 -5.94218373e-01 8.67195487e-01 2.85816580e-01
-1.01369798e+00 -5.77313483e-01 -8.61493230e-01 -1.17520265e-01
-4.74851310e-01 3.58817607e-01 5.96842945e-01 2.11134508e-01
-1.04845715e+00 2.46517271e-01 -1.06344390e+00 -1.25700486e+00
-1.51886269e-01 4.95795578e-01 -6.31722510e-01 1.19733997e-01
-9.03233171e-01 1.17514825e+00 4.10527378e-01 4.66569245e-01
-5.10405481e-01 2.33044237e-01 -1.01867628e+00 8.75708461e-02
3.18844140e-01 -6.88934565e-01 1.09796524e+00 -8.81803632e-01
-2.06302404e+00 6.45263672e-01 -3.47340167e-01 -3.56979251e-01
6.63507164e-01 -3.12140018e-01 -1.82688922e-01 4.05367702e-01
-2.82745957e-01 8.76959920e-01 7.92707562e-01 -1.14099002e+00
-3.52414906e-01 -7.55875170e-01 3.07063818e-01 6.10824704e-01
-3.38155568e-01 1.31257057e-01 -7.47313321e-01 -4.41467285e-01
4.36729789e-01 -1.33063710e+00 4.42563146e-02 -3.23072746e-02
-1.61559895e-01 -1.70585468e-01 8.74016345e-01 -5.84916174e-01
9.85388279e-01 -1.98256111e+00 4.00427997e-01 1.33411646e-01
-8.30408856e-02 3.31758976e-01 -1.20453164e-01 6.22686207e-01
-6.99316338e-02 -2.94687003e-01 1.24601409e-01 -2.61104286e-01
-1.82367191e-01 1.18043073e-01 -6.90039843e-02 7.23064482e-01
-1.79919309e-03 1.08501542e+00 -7.80254006e-01 -2.51666427e-01
7.23685205e-01 6.08947396e-01 -1.42672047e-01 2.34969810e-01
3.76613215e-02 6.57145917e-01 -2.79721469e-01 4.81390715e-01
2.42055789e-01 -3.65092866e-02 3.77207905e-01 -2.57570446e-01
-3.43002170e-01 1.36611328e-01 -9.31379795e-01 1.88668883e+00
-5.69083989e-01 1.03981483e+00 -1.10180303e-01 -9.48297322e-01
7.66047776e-01 5.76760828e-01 5.68412483e-01 -7.56490767e-01
4.93795723e-01 -7.55273178e-02 2.52114842e-03 -9.05133903e-01
5.99679649e-01 1.72147438e-01 1.56107754e-01 5.13237059e-01
3.60867265e-03 2.25362152e-01 3.37863825e-02 -5.47279790e-02
7.96941221e-01 4.68753070e-01 5.93726814e-01 1.29882246e-01
5.32871008e-01 3.50562558e-02 3.24514955e-02 4.75097984e-01
-4.03203279e-01 6.02233827e-01 3.05084199e-01 -5.59451938e-01
-7.63764739e-01 -6.64040685e-01 4.01626229e-01 9.60253894e-01
2.19883278e-01 1.16304599e-01 -5.93449712e-01 -3.67878556e-01
-3.85225773e-01 5.60360491e-01 -4.67424124e-01 -7.70131722e-02
-9.17967916e-01 -1.06911287e-01 4.55259472e-01 7.46350229e-01
8.24417174e-01 -1.32810783e+00 -1.40257585e+00 9.97756720e-02
-2.04053313e-01 -1.26678824e+00 -3.11563164e-01 -1.58741586e-02
-9.57779467e-01 -1.22025955e+00 -1.01487970e+00 -6.56087220e-01
5.83349824e-01 6.49339259e-01 4.58575666e-01 -4.07412171e-01
-4.90270182e-02 9.55322504e-01 -4.78425771e-01 -1.98586471e-02
8.43547955e-02 2.15220600e-02 1.71529517e-01 1.62035257e-01
7.44885385e-01 -4.43855971e-01 -5.73132217e-01 3.90955597e-01
-5.26740491e-01 -6.18810244e-02 5.60491502e-01 7.29282737e-01
5.90985119e-02 -3.76943290e-01 3.62588540e-02 -2.63682514e-01
4.43788797e-01 -6.33508191e-02 -4.35104370e-01 1.08304076e-01
-2.04852924e-01 -1.04244262e-01 3.18772197e-01 -5.17617583e-01
-1.22797692e+00 3.82276535e-01 1.49837866e-01 -6.59946561e-01
-5.14435709e-01 4.58196074e-01 -2.65921026e-01 -3.11224684e-02
2.32256815e-01 1.77154347e-01 1.86223343e-01 -2.95267284e-01
4.84772354e-01 7.81028152e-01 5.29413879e-01 -1.21028192e-01
4.10450786e-01 7.96834171e-01 2.52231006e-02 -1.28514969e+00
-1.35570824e-01 -8.36900353e-01 -1.16038954e+00 -5.43396115e-01
1.11479306e+00 -7.77818263e-01 -1.21518242e+00 5.42915225e-01
-1.49213195e+00 -3.41220766e-01 4.87973019e-02 1.07780421e+00
-9.03789699e-01 3.52747023e-01 -5.11981130e-01 -1.01816821e+00
8.50959402e-03 -1.17725623e+00 1.15575314e+00 2.07721055e-01
-5.42165816e-01 -8.15316677e-01 -1.84919119e-01 3.06767523e-01
1.11132972e-01 2.17339218e-01 4.83109146e-01 -4.83722836e-01
-7.06161857e-01 -4.14633512e-01 -1.87419891e-01 -6.97806329e-02
-3.98061797e-03 -1.65443346e-01 -1.06308293e+00 -2.87968457e-01
1.91871166e-01 -1.69413388e-01 6.59199119e-01 4.18222994e-01
6.01256311e-01 -1.09620005e-01 -3.50354344e-01 3.44439328e-01
1.05283964e+00 6.26245201e-01 8.14688146e-01 5.42590559e-01
8.04978728e-01 9.55876768e-01 6.52191281e-01 4.68601078e-01
5.01119554e-01 1.07331407e+00 3.77494186e-01 -2.62178816e-02
-1.72422737e-01 -9.51278210e-02 5.11607766e-01 5.67988753e-01
-3.11231315e-01 -2.58914322e-01 -9.14470077e-01 4.39452440e-01
-2.02704453e+00 -1.14274931e+00 1.18516013e-02 2.12941122e+00
4.21310179e-02 -2.66767830e-01 5.99880926e-02 -2.92912759e-02
6.73577905e-01 3.21269333e-01 -5.14105320e-01 -5.57581484e-01
7.82189965e-02 -3.06518167e-01 3.08551133e-01 4.21444893e-01
-1.15415347e+00 1.14760125e+00 6.13981009e+00 8.77263024e-02
-1.63864517e+00 -1.53576389e-01 8.34444538e-02 -7.60507062e-02
4.89557952e-01 -1.72216758e-01 -4.57708418e-01 -6.52318448e-02
7.48381913e-01 3.92224580e-01 4.68659133e-01 9.22359586e-01
5.63589513e-01 -4.27644849e-01 -1.35049641e+00 1.19437003e+00
4.31971014e-01 -9.05008256e-01 -1.51416317e-01 4.22244281e-01
1.55112565e-01 -3.83514538e-02 5.79289980e-02 2.72474680e-02
-1.67596474e-01 -8.54755163e-01 5.52775741e-01 7.39755511e-01
6.32049501e-01 -5.26534021e-01 6.96661592e-01 4.00624931e-01
-1.26010478e+00 -1.17052235e-01 -7.15158656e-02 -3.51106405e-01
2.21443906e-01 -9.89457294e-02 -9.46602285e-01 3.07530433e-01
5.92257440e-01 6.30438328e-01 -1.85427129e-01 6.84681892e-01
-1.56023771e-01 9.83661413e-02 -1.80203870e-01 -1.12803623e-01
3.05638582e-01 -1.03589848e-01 6.37947023e-01 1.15739179e+00
1.67209029e-01 3.74038011e-01 -5.42378165e-02 6.42015219e-01
3.43290359e-01 -5.96909411e-02 -1.08536088e+00 -2.04037368e-01
2.02766042e-02 9.30851102e-01 -8.13018322e-01 -2.35265717e-01
-4.92672056e-01 1.30902147e+00 1.44162387e-01 6.16013467e-01
-5.37584901e-01 -3.07828039e-01 5.75677872e-01 -3.01402390e-01
2.14808822e-01 -8.08675468e-01 -3.10773522e-01 -1.19417584e+00
1.63130369e-02 -7.25248456e-01 1.77143291e-01 -8.51145983e-01
-6.03452981e-01 4.59861189e-01 1.51984379e-01 -1.39581800e+00
-7.52809823e-01 -7.87836909e-01 -3.95594656e-01 6.51249707e-01
-1.20692384e+00 -1.23302066e+00 -6.12014890e-01 4.72781658e-01
7.96158791e-01 -1.40196905e-01 8.07811916e-01 -5.94910160e-02
-5.08166909e-01 4.53496039e-01 -1.18816800e-01 2.85838574e-01
5.89608967e-01 -8.01028013e-01 2.79352635e-01 7.60845780e-01
2.15359136e-01 1.00421345e+00 5.71575582e-01 -5.46804130e-01
-1.89385486e+00 -6.50061786e-01 1.09708464e+00 -4.87449914e-01
4.41322356e-01 -1.39702350e-01 -4.32527333e-01 7.20186532e-01
1.99403062e-01 -2.34884217e-01 2.94898689e-01 -7.67989233e-02
-1.11912541e-01 2.36562431e-01 -8.63270521e-01 8.65759492e-01
1.19504166e+00 -8.06736767e-01 -5.48001349e-01 8.88800696e-02
2.81765580e-01 -3.94252211e-01 -7.06930339e-01 2.79954016e-01
1.03538096e+00 -1.02690232e+00 8.18818569e-01 -4.41137344e-01
2.13544667e-01 -1.18177637e-01 -3.11459184e-01 -1.03394544e+00
-6.68802485e-02 -4.40929562e-01 -7.65956491e-02 7.46484458e-01
-1.06308661e-01 -6.33995771e-01 9.02118266e-01 8.17846715e-01
8.79987851e-02 -4.55440193e-01 -8.42482865e-01 -8.29170644e-01
-4.59835649e-01 -5.08056998e-01 2.09762067e-01 7.57771313e-01
3.86106133e-01 3.81318659e-01 -5.07867873e-01 5.91485538e-02
1.05733663e-01 6.13232218e-02 1.23600328e+00 -9.89588976e-01
5.15923537e-02 -4.31617439e-01 -6.04920983e-01 -1.36055863e+00
2.35908449e-01 -4.28675979e-01 1.43295094e-01 -1.49706328e+00
4.16451273e-03 1.95448160e-01 8.90263990e-02 2.20209718e-01
4.15351272e-01 1.88191965e-01 5.50711155e-01 4.33384925e-01
-3.49023730e-01 6.93602860e-01 1.27776468e+00 -1.05547942e-01
-3.86220336e-01 6.58569857e-02 -2.19013579e-02 8.91525686e-01
6.44154727e-01 8.23054984e-02 -3.10125887e-01 -5.30586183e-01
-3.86966795e-01 3.38732988e-01 4.88306612e-01 -9.60352600e-01
7.75861800e-01 -1.62944794e-01 2.04655871e-01 -5.88551521e-01
9.47694540e-01 -1.00493336e+00 -1.10054174e-02 6.00132942e-01
-3.57840389e-01 1.57305107e-01 -1.13834873e-01 5.58244407e-01
-3.97361577e-01 7.85823353e-03 4.82386321e-01 -3.05741400e-01
-1.05422223e+00 1.74620636e-02 -6.24105811e-01 -5.32870173e-01
9.21558082e-01 -7.26414382e-01 -5.31171449e-02 -7.32727468e-01
-7.37235188e-01 -1.18265718e-01 3.99313509e-01 5.64448953e-01
7.79497981e-01 -1.10567200e+00 -1.20594501e-01 2.51124293e-01
2.94268131e-01 -3.28434467e-01 8.36106166e-02 1.13453734e+00
-5.46281040e-01 1.15256500e+00 -3.56311291e-01 -7.75843441e-01
-1.28709960e+00 4.04938847e-01 3.09653938e-01 1.04073137e-01
-9.31975543e-01 4.96154517e-01 2.72132069e-01 -3.26290697e-01
6.59626245e-01 -2.37396017e-01 -5.06415367e-01 1.95325002e-01
3.88783276e-01 5.24096251e-01 -2.26600736e-01 -1.34589231e+00
-4.09769535e-01 7.68337190e-01 2.54274666e-01 -5.70366323e-01
1.08372164e+00 -3.17592233e-01 -1.23847788e-02 5.54739177e-01
1.29795063e+00 -6.05387330e-01 -1.22256768e+00 -2.68409848e-01
2.02121481e-01 -4.97736603e-01 -4.66014445e-02 -4.20651078e-01
-8.92015219e-01 1.16931665e+00 7.68115044e-01 -2.52781838e-01
1.01932847e+00 -1.83515400e-01 6.34192824e-01 8.80824983e-01
5.84566712e-01 -1.16169930e+00 3.92165452e-01 8.92871380e-01
1.04730022e+00 -1.35060382e+00 -1.67095095e-01 -1.07069664e-01
-8.28635991e-01 1.36561406e+00 6.34848654e-01 -1.22216367e-03
3.60439509e-01 -1.27478987e-01 1.53785050e-01 -3.70654762e-01
-3.14126313e-01 -4.61190224e-01 2.23998398e-01 5.79004705e-01
4.76947576e-01 9.04874224e-03 -3.25629294e-01 -5.19881435e-02
-6.30997419e-02 3.30322757e-02 3.66988510e-01 1.04614151e+00
-2.77752787e-01 -7.80179739e-01 -4.47889626e-01 -1.93068281e-01
9.21917111e-02 2.10465237e-01 -4.73072201e-01 1.20812559e+00
-2.07098812e-01 9.89293754e-01 1.09978735e-01 -5.13506770e-01
3.61366034e-01 2.42655545e-01 6.01273239e-01 -3.78637820e-01
-4.32038724e-01 2.00022295e-01 3.09357885e-02 -9.59015429e-01
-8.25273752e-01 -7.19069541e-01 -1.35509968e+00 -1.49126843e-01
-1.35780558e-01 -4.11632329e-01 1.00700879e+00 1.26779580e+00
1.83084846e-01 2.68660516e-01 3.16532761e-01 -1.39810658e+00
-2.11864203e-01 -9.63138163e-01 -3.26341420e-01 2.55065411e-01
2.13845119e-01 -8.09039533e-01 -2.09627603e-03 -4.38570529e-02] | [6.904598236083984, -0.11858619004487991] |
7563776b-203b-4c68-9740-05375511c7eb | learning-multiple-gaits-of-quadruped-robot | 2112.04741 | null | https://arxiv.org/abs/2112.04741v1 | https://arxiv.org/pdf/2112.04741v1.pdf | Learning multiple gaits of quadruped robot using hierarchical reinforcement learning | There is a growing interest in learning a velocity command tracking controller of quadruped robot using reinforcement learning due to its robustness and scalability. However, a single policy, trained end-to-end, usually shows a single gait regardless of the command velocity. This could be a suboptimal solution considering the existence of optimal gait according to the velocity for quadruped animals. In this work, we propose a hierarchical controller for quadruped robot that could generate multiple gaits (i.e. pace, trot, bound) while tracking velocity command. Our controller is composed of two policies, each working as a central pattern generator and local feedback controller, and trained with hierarchical reinforcement learning. Experiment results show 1) the existence of optimal gait for specific velocity range 2) the efficiency of our hierarchical controller compared to a controller composed of a single policy, which usually shows a single gait. Codes are publicly available. | ['Dongjun Lee', 'Bukun Son', 'Yunho Kim'] | 2021-12-09 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [-2.18231142e-01 2.26623639e-01 -7.27348104e-02 2.39900887e-01
4.53954265e-02 -4.27665293e-01 1.30763352e-01 -1.26667693e-01
-3.70057076e-01 1.17493916e+00 -1.84576839e-01 -1.44499354e-02
-7.73785785e-02 -9.30541635e-01 -9.44471121e-01 -9.92001414e-01
-2.70172298e-01 3.91568661e-01 5.32479525e-01 -4.27644253e-01
1.21731803e-01 9.99155715e-02 -1.59472132e+00 -3.86342585e-01
9.74881470e-01 6.01581335e-01 6.33848727e-01 7.51979589e-01
4.35504019e-01 4.96984005e-01 -4.38129812e-01 5.00688851e-01
2.56085932e-01 -7.20165074e-01 -3.84024680e-01 1.91698328e-01
-1.42464027e-01 -2.30497777e-01 -1.68874964e-01 7.30376840e-01
6.40758693e-01 1.87140733e-01 6.12652659e-01 -1.67598712e+00
-1.99414745e-01 7.21459270e-01 -4.37485844e-01 -2.17509925e-01
1.50840670e-01 5.68119168e-01 8.87753069e-01 1.27718166e-01
6.08290017e-01 1.04229939e+00 5.86255670e-01 7.13630855e-01
-1.11776340e+00 -6.42969847e-01 2.74435263e-02 1.98831931e-02
-1.31230295e+00 1.50108695e-01 4.55089152e-01 -3.66567135e-01
7.80415773e-01 -4.02184457e-01 9.63682413e-01 1.07857037e+00
6.71790242e-01 4.57542926e-01 9.26048934e-01 -2.18964174e-01
5.46573997e-01 -3.96164626e-01 -3.94359708e-01 8.86509001e-01
5.24139404e-01 4.09928292e-01 -1.69889510e-01 1.47711098e-01
1.08361363e+00 -4.46579069e-01 -2.44782582e-01 -8.64692450e-01
-1.19409275e+00 6.60123467e-01 6.06854022e-01 2.28242069e-01
-6.68222725e-01 3.90984595e-01 3.12633932e-01 2.56608784e-01
-6.84131384e-01 4.68179107e-01 -7.61104941e-01 -3.57915342e-01
-3.50732774e-01 8.19793522e-01 8.30497921e-01 9.03117537e-01
6.80986702e-01 5.28576672e-01 -6.89944625e-02 5.96680701e-01
1.17353946e-01 7.01002121e-01 5.81776679e-01 -1.01553142e+00
2.12278321e-01 6.75533235e-01 4.18950647e-01 -9.26155269e-01
-9.27159309e-01 -1.82997793e-01 -8.10840666e-01 5.13229549e-01
4.71892536e-01 -6.63750708e-01 -7.32500970e-01 2.00737309e+00
2.93104649e-01 -5.05256653e-02 2.60229439e-01 9.90913033e-01
2.82952972e-02 7.16644943e-01 -8.44625011e-03 -2.60458380e-01
1.07753742e+00 -6.74665511e-01 -6.41582847e-01 -1.02890015e-01
5.47279119e-01 -5.49239144e-02 1.06657839e+00 2.98244476e-01
-8.25504899e-01 -5.15272021e-01 -1.40326691e+00 7.79519379e-01
-2.30967075e-01 2.17494756e-01 -3.56624499e-02 3.93451780e-01
-8.11721981e-01 8.33795846e-01 -1.05311799e+00 -8.13845813e-01
-3.70683759e-01 5.09488046e-01 -9.76114795e-02 4.83094960e-01
-1.17840886e+00 1.01591277e+00 9.70597684e-01 -3.09341967e-01
-8.68725777e-01 -2.70504028e-01 -6.53794885e-01 -6.23712316e-02
4.68409866e-01 -8.96831572e-01 1.34470642e+00 -6.00097179e-01
-1.94388056e+00 1.25834629e-01 3.85116369e-01 -5.81766427e-01
6.06917560e-01 -6.44262359e-02 -5.79798594e-02 1.08677313e-01
2.68331826e-01 6.90568209e-01 9.36062038e-01 -1.35800159e+00
-8.71598661e-01 3.52385566e-02 -1.41831130e-01 3.84820133e-01
-1.83397934e-01 -6.02889359e-01 8.07839707e-02 -3.00971508e-01
-4.04022783e-02 -1.21099222e+00 -1.86141804e-01 -3.44953805e-01
-1.65082276e-01 -2.28053838e-01 1.02407718e+00 -2.33817846e-01
1.24785066e+00 -1.73714280e+00 3.09954375e-01 8.62655565e-02
-3.20499092e-01 1.03983328e-01 1.12007566e-01 6.74863875e-01
1.84168875e-01 2.94030458e-02 -2.64088780e-01 6.06884420e-01
-1.45521229e-02 4.99052882e-01 9.09955651e-02 3.20303440e-01
8.12893435e-02 5.18924177e-01 -1.14450622e+00 -5.21672368e-01
1.71154961e-01 5.77866770e-02 -6.36923373e-01 4.77241844e-01
-4.46348161e-01 5.96055150e-01 -5.63956976e-01 3.57271165e-01
4.19543326e-01 -9.62309316e-02 5.99964976e-01 -3.78750563e-02
-6.43002272e-01 -4.10414875e-01 -1.30047417e+00 1.02596283e+00
-2.98119277e-01 -1.09079741e-01 3.85301262e-01 -1.00694585e+00
1.26668823e+00 3.99055660e-01 8.05925846e-01 -5.59804201e-01
4.15547609e-01 2.62349367e-01 2.97970921e-01 -3.95135671e-01
1.68258339e-01 -1.19398914e-01 -1.58149943e-01 2.72801578e-01
2.56471455e-01 -2.85848409e-01 6.49260759e-01 -2.96569794e-01
1.41364300e+00 6.65896237e-01 4.71347302e-01 -3.15208942e-01
5.28928041e-01 1.14085272e-01 8.00046086e-01 6.67001188e-01
-3.74996394e-01 2.86492884e-01 5.40575802e-01 5.59868803e-03
-1.35657227e+00 -1.07267630e+00 3.28303993e-01 1.01795244e+00
3.32944483e-01 -4.25385028e-01 -5.97128987e-01 -7.72404373e-02
1.23325072e-01 4.44118828e-01 -3.19673508e-01 -3.39243025e-01
-1.00309229e+00 -4.94198173e-01 3.85827214e-01 5.45558512e-01
8.15149784e-01 -1.29995084e+00 -1.52719080e+00 4.75889832e-01
-2.98109233e-01 -1.15005922e+00 -3.12006146e-01 4.14907455e-01
-7.74656415e-01 -1.18729997e+00 -5.98549426e-01 -9.57795560e-01
3.94834489e-01 3.34790815e-03 5.54089606e-01 2.42833853e-01
9.79374349e-02 2.43358657e-01 -4.99521554e-01 -1.74126893e-01
-5.38258374e-01 1.98379666e-01 4.74670082e-01 -2.99646884e-01
-5.92753291e-01 -6.56386554e-01 -5.24511874e-01 6.93235517e-01
-6.70498967e-01 3.10761873e-02 6.87256336e-01 1.04312027e+00
4.24748152e-01 3.58960927e-01 6.05428934e-01 -1.08883224e-01
7.22264528e-01 -2.98882961e-01 -8.03591907e-01 8.05300921e-02
-4.62776184e-01 2.41587326e-01 1.14555752e+00 -4.31183785e-01
-7.83248842e-01 5.16115844e-01 -4.34188917e-02 -1.83328971e-01
-2.44684979e-01 1.83104828e-01 -2.26370916e-02 3.52003761e-02
6.37672424e-01 3.91905755e-01 3.58532041e-01 2.88024005e-02
3.79278213e-01 4.46306616e-01 5.49642265e-01 -9.23817754e-01
8.63591015e-01 -8.10406357e-02 3.33734512e-01 -8.78918469e-01
-5.87181784e-02 -1.93328902e-01 -6.24749005e-01 -5.66050053e-01
8.21422219e-01 -6.01985097e-01 -1.25204980e+00 4.60151941e-01
-8.43817234e-01 -5.75760543e-01 7.30965063e-02 3.15208256e-01
-1.36253524e+00 6.71172217e-02 -5.57815313e-01 -8.15638125e-01
-2.39372671e-01 -9.01754260e-01 6.89753056e-01 6.03815377e-01
-2.77843475e-01 -5.70847154e-01 2.00736761e-01 -4.54051703e-01
3.87297064e-01 6.02519751e-01 9.30477560e-01 -4.41714004e-02
-1.82038218e-01 1.03405930e-01 2.29265809e-01 -1.09831236e-01
9.49279219e-02 6.38852566e-02 -1.93367362e-01 -7.28225172e-01
-4.15787637e-01 -5.48619032e-01 1.71387732e-01 6.51973367e-01
4.76714253e-01 -4.53241855e-01 -4.08236980e-01 1.81241274e-01
1.69575632e+00 6.95110500e-01 3.88423085e-01 7.40560055e-01
3.29168111e-01 3.77494812e-01 8.24840963e-01 6.75791740e-01
2.22771034e-01 7.29014754e-01 6.75224364e-01 3.73855680e-01
8.16270262e-02 -6.19360626e-01 5.03531158e-01 5.14840841e-01
-2.25464374e-01 -7.68407285e-02 -9.48176742e-01 4.88200307e-01
-2.27413416e+00 -9.39678848e-01 1.17918335e-01 2.16075492e+00
5.87732255e-01 2.17688665e-01 8.06697667e-01 3.13584954e-01
9.29660141e-01 -3.15444410e-01 -7.75383651e-01 -4.23210919e-01
6.23749085e-02 -9.75008234e-02 7.33995497e-01 8.65618289e-02
-9.34243083e-01 8.42358708e-01 6.21568871e+00 3.72712970e-01
-1.33784497e+00 -6.12981915e-01 -1.33931652e-01 4.31081474e-01
4.97985870e-01 -1.00924134e-01 -7.47073531e-01 7.07877576e-01
7.98753858e-01 -4.64925140e-01 4.43240345e-01 9.16298926e-01
5.00293016e-01 -2.22279459e-01 -7.01193273e-01 5.91423988e-01
-4.42419678e-01 -8.94990683e-01 -1.54492140e-01 -9.29669067e-02
6.58703804e-01 -4.98561025e-01 -4.25098598e-01 5.31525612e-01
6.41944826e-01 -7.47008443e-01 8.56674373e-01 7.54340068e-02
2.56325185e-01 -9.03766990e-01 5.51081061e-01 8.75009418e-01
-1.51168251e+00 -5.95787168e-01 -2.36513823e-01 -1.92864120e-01
2.33660385e-01 -1.76892892e-01 -9.13748085e-01 6.83898389e-01
6.37495935e-01 5.80481112e-01 -3.26605201e-01 1.36414957e+00
-2.88751453e-01 4.73367304e-01 -4.49942350e-01 -6.56914532e-01
3.05448264e-01 -3.46930444e-01 6.85547113e-01 8.03069890e-01
4.86121595e-01 3.56070958e-02 6.29643738e-01 6.48407161e-01
5.86848438e-01 8.70932564e-02 -7.01251209e-01 1.14690308e-02
3.92710119e-01 1.02702403e+00 -8.29744875e-01 -2.93898918e-02
4.43518125e-02 6.79087877e-01 3.58060747e-01 1.39575422e-01
-1.06639791e+00 -5.79202950e-01 4.31700557e-01 7.51783978e-03
6.66969538e-01 -5.16608119e-01 -8.75230059e-02 -7.35148132e-01
-1.84094876e-01 -7.23280191e-01 3.04638535e-01 -6.38247132e-01
-1.07387614e+00 1.82193205e-01 9.13509056e-02 -1.70296395e+00
-6.10627174e-01 -4.94911462e-01 -5.56598723e-01 2.11046055e-01
-8.01431954e-01 -7.17297971e-01 -2.63163924e-01 4.84892786e-01
2.89769202e-01 -1.38219401e-01 6.26881838e-01 -1.62099674e-01
-4.53714579e-01 3.23177606e-01 -3.63326222e-02 -7.10819438e-02
6.01693630e-01 -1.31997836e+00 -3.43236655e-01 5.63845456e-01
-7.33077228e-01 3.36997956e-01 1.21633017e+00 -8.11418176e-01
-1.45469999e+00 -9.43642735e-01 2.89502263e-01 2.97651291e-01
8.04554224e-01 1.44096017e-01 -6.74510717e-01 3.87206614e-01
2.33795986e-01 -3.86883616e-01 -1.30224273e-01 -4.34318066e-01
4.59563792e-01 -1.76761016e-01 -1.09140337e+00 8.53935778e-01
8.38070989e-01 4.91276205e-01 -4.73059297e-01 -1.22752674e-01
6.31603241e-01 -5.25153697e-01 -8.19878280e-01 6.63133502e-01
8.94408643e-01 -7.40990937e-01 6.60009921e-01 -3.75405639e-01
4.14612710e-01 -7.03266621e-01 7.46502131e-02 -1.74622536e+00
-5.43209851e-01 -5.07253945e-01 -1.37375340e-01 8.60601962e-01
7.64612705e-02 -5.57708383e-01 7.87237048e-01 -1.42654434e-01
-2.72296332e-02 -7.33369768e-01 -8.74099195e-01 -1.37953627e+00
2.24818379e-01 3.26793849e-01 3.96192014e-01 4.69207376e-01
3.40796053e-01 5.74613988e-01 -6.26487017e-01 4.13658440e-01
6.52278960e-01 9.05471593e-02 9.48067367e-01 -8.68361235e-01
-3.11129004e-01 -4.14884955e-01 -4.21994627e-01 -7.80608058e-01
-7.93097615e-02 -5.01787484e-01 6.57205224e-01 -1.64599562e+00
-2.43824854e-01 -2.53628075e-01 -5.02038226e-02 6.86932385e-01
1.37476161e-01 -2.64488935e-01 2.63262779e-01 -9.11145285e-02
-4.25426662e-01 5.29926717e-01 1.52376401e+00 9.94720906e-02
-7.00504065e-01 2.18351901e-01 -2.81683147e-01 5.32600939e-01
1.43807614e+00 -3.25418800e-01 -4.55223680e-01 1.02986127e-01
-2.09320813e-01 6.36082232e-01 7.86481872e-02 -1.70550370e+00
2.21786514e-01 -3.90218765e-01 2.53383309e-01 -4.36437517e-01
2.77885906e-02 -8.22880149e-01 1.22088350e-01 1.23874223e+00
-1.00068964e-01 1.94331482e-01 8.76003951e-02 6.63643301e-01
-3.79683003e-02 -7.55462870e-02 1.07138348e+00 -1.99805096e-01
-8.55672181e-01 1.04261972e-01 -6.99010968e-01 -1.81298763e-01
1.35308886e+00 -2.24536747e-01 -2.01595262e-01 -2.77269036e-01
-5.84122717e-01 6.99661255e-01 3.87852401e-01 5.08621573e-01
3.30215424e-01 -1.31310606e+00 -4.45721686e-01 -3.24840434e-02
1.24130443e-01 -3.45847756e-01 -6.88610226e-02 4.81037647e-01
-6.27198219e-01 4.51746285e-01 -1.07971466e+00 -6.42418861e-01
-1.02351058e+00 4.18105394e-01 3.89814347e-01 -2.37514406e-01
-8.02580953e-01 -4.11398374e-02 -2.04092920e-01 -6.26441121e-01
5.56462593e-02 -3.36449414e-01 -3.29179496e-01 -3.91593903e-01
-4.11998332e-02 4.08547223e-01 -2.80746400e-01 -3.81214172e-01
-2.73789018e-01 8.71166587e-01 5.16671896e-01 -1.09057270e-01
1.10256410e+00 2.34363712e-02 2.51033783e-01 3.78471464e-01
4.71294075e-01 -6.28245234e-01 -1.53435731e+00 4.90094423e-01
-2.63441354e-02 -1.50523141e-01 -6.13025069e-01 -6.72830880e-01
-7.07156479e-01 3.06283832e-01 6.94977462e-01 2.81239390e-01
9.28602874e-01 -5.47078252e-01 6.56738639e-01 4.97883528e-01
8.74421775e-01 -1.44805741e+00 3.92001987e-01 8.22903514e-01
9.23896611e-01 -8.04376841e-01 -1.99734747e-01 -1.18828401e-01
-8.12845349e-01 1.18314326e+00 1.29664922e+00 -6.03562236e-01
4.65790123e-01 4.02911305e-01 -4.54959199e-02 4.02013987e-01
-9.11782324e-01 -4.28845286e-01 -2.91061878e-01 1.02047527e+00
1.05084874e-01 2.21543267e-01 -8.60482514e-01 2.73305118e-01
-4.80038106e-01 1.49006531e-01 6.69454873e-01 1.26316714e+00
-1.03858674e+00 -1.06760895e+00 -3.81366909e-01 5.19746579e-02
1.30548209e-01 6.61138058e-01 7.30776787e-02 1.13639259e+00
2.57048428e-01 9.80252743e-01 -7.57740885e-02 -5.52743614e-01
6.69606268e-01 -6.88498616e-02 6.53789937e-01 -2.94799894e-01
-3.51917386e-01 3.87809006e-04 -8.19952711e-02 -4.58023489e-01
-1.29003786e-02 -4.88155037e-01 -1.75615418e+00 -1.27469808e-01
1.54485870e-02 1.49020180e-01 3.44848812e-01 7.16268182e-01
-1.65619794e-02 6.03685379e-01 6.49299860e-01 -8.65138948e-01
-8.21897626e-01 -7.94892013e-01 -6.96595788e-01 4.04024333e-01
2.75571138e-01 -1.14438522e+00 -1.52362928e-01 2.77459882e-02] | [4.635200500488281, 1.3994934558868408] |
c7ff56b5-8414-4222-bb67-ed06af48a587 | not-end-to-end-explore-multi-stage | 2107.04810 | null | https://arxiv.org/abs/2107.04810v1 | https://arxiv.org/pdf/2107.04810v1.pdf | Not End-to-End: Explore Multi-Stage Architecture for Online Surgical Phase Recognition | Surgical phase recognition is of particular interest to computer assisted surgery systems, in which the goal is to predict what phase is occurring at each frame for a surgery video. Networks with multi-stage architecture have been widely applied in many computer vision tasks with rich patterns, where a predictor stage first outputs initial predictions and an additional refinement stage operates on the initial predictions to perform further refinement. Existing works show that surgical video contents are well ordered and contain rich temporal patterns, making the multi-stage architecture well suited for the surgical phase recognition task. However, we observe that when simply applying the multi-stage architecture to the surgical phase recognition task, the end-to-end training manner will make the refinement ability fall short of its wishes. To address the problem, we propose a new non end-to-end training strategy and explore different designs of multi-stage architecture for surgical phase recognition task. For the non end-to-end training strategy, the refinement stage is trained separately with proposed two types of disturbed sequences. Meanwhile, we evaluate three different choices of refinement models to show that our analysis and solution are robust to the choices of specific multi-stage models. We conduct experiments on two public benchmarks, the M2CAI16 Workflow Challenge, and the Cholec80 dataset. Results show that multi-stage architecture trained with our strategy largely boosts the performance of the current state-of-the-art single-stage model. Code is available at \url{https://github.com/ChinaYi/casual_tcn}. | ['Tingting Jiang', 'Fangqiu Yi'] | 2021-07-10 | null | null | null | null | ['online-surgical-phase-recognition', 'surgical-phase-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.43941414e-01 1.99024349e-01 -6.06644332e-01 -3.12688440e-01
-5.19627273e-01 -2.46281162e-01 4.35173362e-01 -5.91835119e-02
-6.20395482e-01 2.16834903e-01 3.09856713e-01 -5.51145315e-01
-5.27455807e-02 -3.80005091e-01 -5.29264987e-01 -8.13328505e-01
-1.33180976e-01 3.96896422e-01 4.04875875e-01 -1.43533140e-01
2.93192621e-02 3.53719622e-01 -1.13383675e+00 6.75281525e-01
3.22602838e-01 1.14649618e+00 3.04981768e-01 8.74669611e-01
2.51557618e-01 8.46894085e-01 -8.21518376e-02 -3.66535783e-02
5.23139536e-01 -2.97980458e-01 -8.49754512e-01 -7.13543996e-05
1.74558684e-02 -3.05524766e-01 -5.46796262e-01 7.81120777e-01
5.25083363e-01 -1.47238031e-01 3.69511664e-01 -8.93199146e-01
2.14587525e-01 5.30421734e-01 -4.61261451e-01 5.08628666e-01
-6.22827895e-02 3.37328047e-01 7.05426931e-01 -7.66331673e-01
6.27267003e-01 5.70205748e-01 6.67525351e-01 7.27676153e-01
-9.11831856e-01 -7.21764863e-01 2.26132900e-01 2.38573566e-01
-1.22856021e+00 -5.05079269e-01 7.18876541e-01 -4.96421218e-01
4.72809076e-01 2.88969070e-01 6.83821142e-01 1.09569979e+00
6.40028119e-01 9.31673288e-01 1.00653875e+00 -1.33639947e-01
9.88390297e-03 -4.42691714e-01 1.44099399e-01 1.02677214e+00
-2.05998379e-03 3.41413558e-01 -4.10435110e-01 9.45281833e-02
8.18774879e-01 4.77049708e-01 -5.33681214e-01 -5.02756536e-01
-1.54824960e+00 5.83116412e-01 6.20825231e-01 4.08704966e-01
-4.35004115e-01 -9.39879380e-03 3.61614764e-01 2.22319305e-01
2.01765627e-01 4.45687234e-01 -4.01426077e-01 -1.34434968e-01
-1.13640225e+00 -2.72772107e-02 6.94869876e-01 7.98570931e-01
4.24970180e-01 -3.36803257e-01 -5.14625549e-01 6.17438018e-01
1.00956462e-01 -2.76835650e-01 8.87015700e-01 -8.02784681e-01
1.80948362e-01 5.77611089e-01 -2.13161916e-01 -5.08403182e-01
-9.38965142e-01 -7.61301100e-01 -1.36625886e+00 1.48589775e-01
5.50041437e-01 -1.77201733e-01 -1.48360670e+00 1.56441534e+00
1.78952381e-01 4.54799116e-01 -1.25352785e-01 9.78257179e-01
9.50425863e-01 2.96759963e-01 3.64187546e-02 -4.35370803e-01
1.41264486e+00 -1.33087623e+00 -4.58230078e-01 -4.46023315e-01
9.09725845e-01 -5.58249176e-01 6.29392266e-01 4.31416333e-01
-1.03389335e+00 -4.05858904e-01 -9.36470449e-01 1.41792536e-01
4.44300957e-02 2.06399456e-01 5.72449863e-01 1.03124090e-01
-9.87372398e-01 7.24054992e-01 -1.36268413e+00 -1.90983593e-01
5.40211022e-01 6.75687134e-01 -5.44760764e-01 -2.28280440e-01
-7.98358202e-01 7.44645357e-01 3.65544081e-01 5.48715055e-01
-1.22791743e+00 -8.37270081e-01 -7.43891001e-01 1.32756969e-02
5.64716637e-01 -7.86276817e-01 1.52917218e+00 -9.15178061e-01
-1.31278872e+00 9.29130375e-01 -3.24452102e-01 -4.65573519e-01
4.80339110e-01 4.99178888e-03 -1.01168498e-01 2.44638145e-01
-1.71342477e-01 5.00474870e-01 7.57027924e-01 -9.09884512e-01
-8.10465336e-01 -2.85505146e-01 1.71599627e-01 2.37396538e-01
-2.03043014e-01 -1.61900282e-01 -8.35700810e-01 -6.27327144e-01
2.11258158e-01 -1.33749127e+00 -8.27395916e-01 1.94664046e-01
-2.64293075e-01 1.84811920e-01 3.36557180e-01 -4.72772449e-01
1.38618422e+00 -2.20189285e+00 2.87389010e-01 2.54318099e-02
4.50299472e-01 2.43313909e-01 -5.69016561e-02 2.17674095e-02
-3.32659364e-01 1.52993770e-02 -2.49770120e-01 -5.06074727e-01
-5.08660734e-01 2.69565493e-01 2.69081444e-01 6.05051160e-01
1.07674105e-02 6.38738751e-01 -7.86669910e-01 -5.38465500e-01
8.50489363e-02 1.67816803e-01 -8.44628751e-01 4.70924526e-01
1.59999475e-01 6.92431152e-01 -3.88225675e-01 6.07484877e-01
1.82107955e-01 -5.58177829e-01 3.40869904e-01 -4.42913592e-01
4.96189855e-02 1.21188559e-01 -7.45162129e-01 2.14097404e+00
-3.70998770e-01 3.32550704e-01 1.09279610e-01 -1.04806924e+00
3.16592783e-01 5.70875704e-01 9.44402814e-01 -5.69026172e-01
3.18872839e-01 1.50777310e-01 6.17630422e-01 -3.32587659e-01
1.35549545e-01 -1.84082881e-01 -1.13008525e-02 7.56959394e-02
9.89830643e-02 3.24380368e-01 2.86330551e-01 8.47511292e-02
1.52932489e+00 -1.90278292e-02 4.75772947e-01 -1.23368606e-01
4.27440852e-01 1.56117007e-01 8.61412227e-01 6.61817670e-01
-4.11952674e-01 8.54121983e-01 5.74649155e-01 -7.46389449e-01
-8.50649536e-01 -7.34635055e-01 9.33258086e-02 8.87637198e-01
2.81998008e-01 -4.47635949e-01 -3.68051052e-01 -9.40917850e-01
-4.17627811e-01 1.36920244e-01 -9.08957362e-01 -2.87183851e-01
-8.47577810e-01 -7.88981497e-01 1.46754682e-01 5.67735016e-01
9.83577669e-02 -9.83081937e-01 -8.29843760e-01 3.30983132e-01
-3.57898176e-01 -1.26198816e+00 -5.83107293e-01 4.31766808e-01
-1.02775824e+00 -1.22770238e+00 -9.32027698e-01 -9.21080887e-01
8.55642557e-01 3.22372764e-01 9.76583898e-01 3.66487592e-01
-2.28085294e-01 8.57786760e-02 -5.16926050e-01 -3.08869720e-01
-3.60171646e-01 3.59839618e-01 -1.43050835e-01 3.42002441e-03
-4.44199331e-02 -5.03923416e-01 -1.16306198e+00 4.34098423e-01
-8.97349238e-01 7.28775382e-01 9.82063711e-01 1.11433315e+00
6.66825175e-01 -2.76454449e-01 -3.76029834e-02 -1.02986658e+00
-7.20051900e-02 -5.24252295e-01 -3.23154837e-01 2.64417231e-02
-7.02204168e-01 6.33574054e-02 5.98546386e-01 -5.11669695e-01
-5.65186441e-01 4.96398330e-01 -3.64427954e-01 -8.22599351e-01
-7.49955028e-02 7.61363566e-01 3.09178680e-01 3.89444525e-04
4.12808150e-01 2.49009401e-01 3.33903372e-01 -2.93574989e-01
-1.91097498e-01 2.31952623e-01 5.74658394e-01 -1.30226448e-01
7.33143389e-01 4.35953766e-01 1.34124801e-01 -5.07783473e-01
-1.14697051e+00 -8.26497018e-01 -5.20516753e-01 -3.16435635e-01
8.10820639e-01 -9.50878143e-01 -6.43624306e-01 4.68188465e-01
-8.25148880e-01 -6.67136490e-01 -6.49946928e-02 5.59203446e-01
-6.80747390e-01 2.74472296e-01 -8.94202411e-01 -1.56370848e-01
-5.17285526e-01 -1.61191583e+00 8.92447352e-01 2.99733400e-01
-2.41910547e-01 -8.53943288e-01 1.11266941e-01 3.32170963e-01
3.06731492e-01 2.62908369e-01 7.58711755e-01 -8.26121807e-01
-4.99368817e-01 -2.18720943e-01 1.52596012e-01 3.62088084e-02
1.25418961e-01 -2.85559237e-01 -6.54211402e-01 -5.01264155e-01
2.33207759e-03 -6.45158142e-02 1.02423728e+00 5.61290860e-01
1.54116488e+00 -2.22505867e-01 -4.94916379e-01 1.07835579e+00
1.24547625e+00 1.34803608e-01 6.71659708e-01 5.41152298e-01
6.46526456e-01 2.12311670e-01 7.52407968e-01 3.50195557e-01
4.44590241e-01 7.40708113e-01 6.41454637e-01 -3.97648573e-01
-9.69000235e-02 8.99771526e-02 2.72223264e-01 7.66927958e-01
-1.81052923e-01 -9.76228565e-02 -1.15924489e+00 5.93406141e-01
-1.98370326e+00 -9.39696670e-01 2.78720021e-01 2.16319323e+00
1.01855528e+00 3.09541762e-01 9.76139456e-02 1.13085933e-01
2.68927366e-01 3.75498712e-01 -3.92959148e-01 4.51437831e-02
4.93625641e-01 3.14242691e-01 7.12838113e-01 1.86019585e-01
-1.29271758e+00 7.25928843e-01 5.44153547e+00 8.23001981e-01
-1.39293242e+00 5.18913083e-02 7.78092444e-01 -3.62587422e-01
2.86669761e-01 -3.19669656e-02 -7.36330748e-01 3.76486033e-01
1.03114188e+00 2.29716692e-02 6.54146299e-02 6.50750995e-01
3.84886861e-01 -1.03132881e-03 -1.35851133e+00 9.23125923e-01
-7.93431699e-02 -1.64391625e+00 -2.67251641e-01 4.22784314e-02
4.39524233e-01 1.96772844e-01 -1.15154967e-01 4.23532873e-01
1.18578836e-01 -1.04022515e+00 4.49573070e-01 4.45740432e-01
7.20210493e-01 -4.01610732e-01 8.43176246e-01 4.90168452e-01
-1.23927140e+00 -2.25702673e-01 1.38970464e-01 1.63821563e-01
1.73467442e-01 1.71281204e-01 -1.08943605e+00 6.71997726e-01
6.68684423e-01 8.76561463e-01 -6.10831678e-01 1.38940096e+00
-4.45336243e-03 6.25238359e-01 -1.60231769e-01 3.69476646e-01
3.35202187e-01 2.20816568e-01 4.10416812e-01 1.21838152e+00
3.45619380e-01 2.98256665e-01 3.78970355e-01 2.02490345e-01
1.72970928e-02 -2.55358905e-01 -2.92429924e-01 2.84187086e-02
-3.02311908e-02 1.47356200e+00 -8.64048302e-01 -2.56878793e-01
-4.79665995e-01 8.18861127e-01 2.45323136e-01 1.38353214e-01
-7.80608177e-01 3.82095240e-02 4.78415012e-01 3.32000494e-01
2.86863267e-01 -2.70175785e-01 -3.36493820e-01 -1.19429207e+00
-8.81335288e-02 -1.02002203e+00 7.17360437e-01 -5.02758205e-01
-6.74146533e-01 6.98010981e-01 -2.35123798e-01 -1.73454010e+00
-3.81868988e-01 -6.06802642e-01 -6.35431111e-01 4.39350873e-01
-1.57389426e+00 -1.15745139e+00 -5.93145609e-01 6.90438509e-01
9.23668325e-01 7.05753416e-02 8.23886871e-01 3.61503929e-01
-6.37157559e-01 7.27754653e-01 -2.54291207e-01 5.02780318e-01
8.02827954e-01 -9.56727147e-01 3.06420028e-02 1.07911766e+00
-1.25586092e-01 3.75798613e-01 6.97260916e-01 -4.45159227e-01
-1.50512886e+00 -1.07230413e+00 6.01915061e-01 -2.12362394e-01
6.33832335e-01 -6.13907650e-02 -6.56318605e-01 7.78855145e-01
3.05534694e-02 2.20938116e-01 7.88071275e-01 9.88819227e-02
1.27247915e-01 -2.17901692e-01 -6.27571285e-01 7.35540509e-01
9.94358242e-01 -6.12431914e-02 -3.64633709e-01 2.68620461e-01
5.46872854e-01 -1.02610958e+00 -1.16602683e+00 8.95840883e-01
7.25851774e-01 -8.34462404e-01 8.28624845e-01 -5.44091702e-01
8.68303359e-01 -1.81377605e-01 2.19234765e-01 -1.17030549e+00
-5.47232926e-01 -6.17334306e-01 -2.59793758e-01 4.85586226e-01
6.46865726e-01 -2.33093023e-01 1.07645893e+00 6.05637431e-01
-6.39121711e-01 -1.33695936e+00 -8.16668689e-01 -3.92644346e-01
-7.09964260e-02 -1.93298444e-01 5.59430905e-02 6.62959397e-01
-6.86916932e-02 1.94330335e-01 -5.11614084e-01 2.16287807e-01
2.64730006e-01 2.09588856e-01 7.03330219e-01 -9.63383138e-01
-4.54160452e-01 -5.77558994e-01 -4.38057423e-01 -1.15565550e+00
-2.09359214e-01 -8.56872737e-01 2.79838800e-01 -1.53405130e+00
3.38982999e-01 -4.40344989e-01 -7.17237651e-01 8.36038172e-01
-3.80948514e-01 2.46178165e-01 3.57980698e-01 5.51922917e-01
-5.76568782e-01 3.24498385e-01 1.42674553e+00 -2.01158926e-01
-2.17776179e-01 4.70777661e-01 -6.83584094e-01 7.75516033e-01
5.69807708e-01 -6.20802820e-01 -3.11667949e-01 -3.53425235e-01
-1.66299287e-02 4.47934180e-01 2.54699707e-01 -1.08351171e+00
5.71303368e-01 -1.64485589e-01 3.46454680e-01 -3.95060092e-01
3.85534823e-01 -8.68392825e-01 2.09617272e-01 9.38427269e-01
-3.14115584e-01 7.80065358e-02 2.13223085e-01 4.39529121e-01
-3.99214268e-01 -5.95401376e-02 1.01444340e+00 -2.07019731e-01
-7.96601832e-01 1.04830468e+00 -3.93909782e-01 -1.26970515e-01
1.04269135e+00 -2.60398388e-01 -1.40169695e-01 -5.48048019e-02
-1.07396924e+00 3.49198163e-01 4.46552873e-01 4.84124839e-01
6.22705519e-01 -1.09822130e+00 -6.73611462e-01 9.90634784e-02
1.74766764e-01 3.82008195e-01 3.84151012e-01 1.55504763e+00
-6.29218161e-01 3.22671860e-01 -1.83912873e-01 -8.63564491e-01
-1.33513546e+00 6.01017892e-01 4.57557976e-01 -8.95090342e-01
-8.86926770e-01 7.26184726e-01 5.86165786e-01 -9.19263512e-02
3.53394032e-01 -3.54765594e-01 -3.82712215e-01 -2.28892207e-01
3.50150317e-01 -3.51118028e-01 1.37160942e-01 -3.97611290e-01
-4.58334357e-01 4.01540697e-01 -5.40938139e-01 2.24413991e-01
1.41516423e+00 1.67178318e-01 3.46397489e-01 2.50428528e-01
1.03370297e+00 -4.98757690e-01 -1.18691003e+00 -4.23611760e-01
-1.08321138e-01 -1.20978087e-01 1.69603959e-01 -7.18804896e-01
-1.45192909e+00 6.49724007e-01 7.12282836e-01 -2.94127405e-01
1.47841680e+00 -2.63061494e-01 7.97542155e-01 1.27600297e-01
2.54735470e-01 -6.23906791e-01 -4.31228289e-03 5.34334064e-01
6.83895648e-01 -1.35304022e+00 1.74275730e-02 -2.80741900e-01
-7.86425829e-01 1.05085325e+00 7.07364500e-01 -1.56273961e-01
8.50953341e-01 3.64751905e-01 9.28157941e-02 -2.18902424e-01
-1.06785011e+00 -1.30408451e-01 4.30420667e-01 4.17716838e-02
4.41296071e-01 -1.30424544e-01 -3.45911652e-01 6.64636850e-01
-5.92722036e-02 3.42581391e-01 5.68120480e-01 1.12535918e+00
-1.12917863e-01 -1.01744056e+00 1.84970751e-01 6.53614342e-01
-7.26344645e-01 -1.21481918e-01 1.03692912e-01 7.73388326e-01
-2.05013111e-01 6.45284891e-01 -1.30220786e-01 -5.35975754e-01
3.68167728e-01 -1.17754422e-01 2.79839635e-01 -6.95511818e-01
-9.21530485e-01 2.50650883e-01 1.51737332e-01 -1.10772705e+00
-4.94344771e-01 -5.32556415e-01 -1.14398825e+00 -2.54986417e-02
4.68872907e-03 5.93807213e-02 3.77616137e-01 9.53208148e-01
3.25467974e-01 9.27149773e-01 6.69660389e-01 -9.86932874e-01
-3.69536400e-01 -8.77917767e-01 -1.77715048e-01 2.74202496e-01
6.21033490e-01 -6.16830885e-01 -2.23704904e-01 1.48404196e-01] | [14.14635944366455, -3.2945520877838135] |
7bbe9d13-e5cb-46fa-8cb4-ce68519829f0 | fusing-pre-trained-language-models-with | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yu_Fusing_Pre-Trained_Language_Models_With_Multimodal_Prompts_Through_Reinforcement_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yu_Fusing_Pre-Trained_Language_Models_With_Multimodal_Prompts_Through_Reinforcement_Learning_CVPR_2023_paper.pdf | Fusing Pre-Trained Language Models With Multimodal Prompts Through Reinforcement Learning | Language models are capable of commonsense reasoning: while domain-specific models can learn from explicit knowledge (e.g. commonsense graphs [6], ethical norms [25]), and larger models like GPT-3 manifest broad commonsense reasoning capacity. Can their knowledge be extended to multimodal inputs such as images and audio without paired domain data? In this work, we propose ESPER (Extending Sensory PErception with Reinforcement learning) which enables text-only pretrained models to address multimodal tasks such as visual commonsense reasoning. Our key novelty is to use reinforcement learning to align multimodal inputs to language model generations without direct supervision: for example, our reward optimization relies only on cosine similarity derived from CLIP and requires no additional paired (image, text) data. Experiments demonstrate that ESPER outperforms baselines and prior work on a variety of multimodal text generation tasks ranging from captioning to commonsense reasoning; these include a new benchmark we collect and release, the ESP dataset, which tasks models with generating the text of several different domains for each image. Our code and data are publicly released at https://github.com/JiwanChung/esper. | ['Yejin Choi', 'Gunhee Kim', 'Ronan Le Bras', 'Prithviraj Ammanabrolu', 'Rowan Zellers', 'Ximing Lu', 'Jae Sung Park', 'Jack Hessel', 'Heeseung Yun', 'Jiwan Chung', 'Youngjae Yu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 6.06143475e-01 4.71006870e-01 -5.76713160e-02 -3.47549587e-01
-9.83052671e-01 -8.66484582e-01 8.96406233e-01 -2.15607509e-01
-2.37395763e-01 8.47287595e-01 5.66742718e-01 -2.37612218e-01
2.44889498e-01 -6.70412064e-01 -9.64170575e-01 -3.03881437e-01
5.01852989e-01 6.96260989e-01 -2.88764238e-01 -7.46153355e-01
2.67332196e-01 -1.37621284e-01 -1.25427270e+00 7.33771324e-01
9.69322026e-01 8.93805146e-01 -2.07048617e-02 8.62932801e-01
-8.29664990e-03 1.40527892e+00 -5.98221183e-01 -1.01186752e+00
-5.16090617e-02 -7.05248713e-01 -1.11561990e+00 -1.37392059e-01
6.99358463e-01 -6.36435390e-01 -4.83572066e-01 1.04713869e+00
6.87148094e-01 3.36108565e-01 9.81401682e-01 -1.76287150e+00
-1.71836829e+00 1.13341022e+00 -2.58915722e-01 -3.12161535e-01
6.97535098e-01 7.99218714e-01 1.33593261e+00 -6.84241176e-01
8.51301849e-01 1.74501443e+00 4.01092589e-01 1.20122182e+00
-1.51847172e+00 -7.51082003e-01 -1.04241639e-01 5.18110514e-01
-9.65034544e-01 -5.43806553e-01 8.00294876e-01 -3.81915569e-01
9.99252141e-01 1.61495760e-01 2.70777911e-01 2.09208465e+00
-2.38331452e-01 1.13479185e+00 1.28542614e+00 -1.34250447e-01
2.38848299e-01 -5.61107211e-02 -3.44290435e-01 5.81708670e-01
-2.52477467e-01 5.89791220e-04 -6.75080478e-01 -1.88000485e-01
5.35357714e-01 -4.17507082e-01 -2.45311588e-01 -2.12418929e-01
-1.62582564e+00 1.07644606e+00 5.62909663e-01 -1.17813282e-01
-1.73683226e-01 7.70323694e-01 5.25540709e-01 3.46792489e-01
-3.94654311e-02 8.80841196e-01 -9.49147046e-02 -3.21360707e-01
-5.81139028e-01 5.14441133e-01 7.31204391e-01 1.01843596e+00
5.25184929e-01 6.95138648e-02 -5.43237031e-01 1.01075983e+00
1.60635695e-01 1.05875182e+00 4.12631661e-01 -1.59053218e+00
7.02182889e-01 2.36715674e-01 7.28854984e-02 -8.30743670e-01
-4.30661626e-02 3.48948807e-01 -7.59195328e-01 5.64990789e-02
4.74376142e-01 -4.32985783e-01 -8.58891249e-01 2.18593216e+00
-1.81851640e-01 -7.95899853e-02 5.03114462e-01 9.57480669e-01
1.19672966e+00 7.10204542e-01 2.94208229e-01 3.27322632e-01
1.26564252e+00 -8.08055639e-01 -5.31992495e-01 -4.55471963e-01
3.79088283e-01 -7.09285200e-01 1.49766505e+00 4.00558650e-01
-1.25287247e+00 -1.77058771e-01 -5.81209600e-01 -5.37562490e-01
-5.68804801e-01 -6.54738396e-02 5.19863248e-01 1.39384881e-01
-1.18612170e+00 3.81095856e-01 -2.91843265e-01 -6.06507242e-01
5.87635219e-01 -8.63088444e-02 -2.86269695e-01 -1.97880104e-01
-1.65071511e+00 1.05166984e+00 4.45181370e-01 -1.83747754e-01
-1.22647858e+00 -5.29708982e-01 -1.11620378e+00 -1.29294083e-01
3.84034365e-01 -1.18009198e+00 1.57424223e+00 -1.36572886e+00
-1.52250040e+00 1.17777741e+00 6.67799115e-02 -5.89517772e-01
5.68267584e-01 -1.04211748e-01 -1.30572036e-01 3.72133255e-01
1.55819535e-01 1.52057314e+00 8.23018134e-01 -1.43170524e+00
1.83601771e-02 4.28426825e-02 4.15468872e-01 2.90563524e-01
-1.58044383e-01 -8.41106027e-02 1.25018075e-01 -6.07280254e-01
-6.48737967e-01 -9.82360184e-01 2.08116367e-01 1.19839199e-01
-8.22693646e-01 -3.37917387e-01 4.08471525e-01 -8.19737971e-01
5.67735314e-01 -1.95416057e+00 5.49567878e-01 -1.41097054e-01
3.07305735e-02 -3.32775384e-01 -6.52313411e-01 6.02611303e-01
-1.54805146e-02 2.45601162e-01 -3.52186173e-01 -1.62231833e-01
7.23998427e-01 2.34976187e-01 -7.78879404e-01 -1.71137437e-01
5.30783057e-01 1.34294546e+00 -1.09714925e+00 -7.77279317e-01
3.87046076e-02 3.63686085e-01 -6.60638869e-01 1.62844315e-01
-8.89027536e-01 1.45395845e-01 -6.22819290e-02 6.73710287e-01
3.99074852e-01 -4.94037360e-01 -4.67001274e-02 -2.06186414e-01
5.23518920e-01 1.57055691e-01 -4.94179040e-01 2.05119944e+00
-4.61105675e-01 8.38966846e-01 -2.14558035e-01 -6.19748294e-01
6.72007143e-01 3.03461969e-01 -7.15561211e-02 -7.29870737e-01
1.58508241e-01 -1.07882187e-01 -5.25459312e-02 -6.82761192e-01
6.37542725e-01 -4.23489958e-01 -4.94922549e-01 5.96188903e-01
2.25863203e-01 -8.30111325e-01 3.52482498e-01 7.94195831e-01
1.06414521e+00 2.57818699e-01 7.96861053e-02 2.97354072e-01
5.96755967e-02 3.13207865e-01 4.07629460e-02 7.90224016e-01
-3.61934602e-01 4.68990922e-01 8.03543806e-01 1.74215361e-01
-1.08468497e+00 -1.64358306e+00 2.37009332e-01 1.46875787e+00
1.86507821e-01 -8.99770558e-02 -6.75982893e-01 -6.01018608e-01
1.35083228e-01 1.44791222e+00 -8.68154943e-01 -2.80390799e-01
-7.15438426e-02 -3.03759933e-01 1.01697052e+00 5.54875672e-01
4.45125133e-01 -1.54974699e+00 -2.90030390e-01 -1.59786493e-01
-7.73651183e-01 -1.31320846e+00 -4.23311383e-01 -7.72544518e-02
-4.96945411e-01 -8.76332164e-01 -7.51019597e-01 -5.18788576e-01
3.93429786e-01 -2.20981956e-01 1.51134396e+00 -2.22390071e-01
-2.58722633e-01 9.19801116e-01 -3.97596717e-01 -3.19170147e-01
-6.98574364e-01 -3.00331205e-01 -4.63046700e-01 -4.27283436e-01
2.68205136e-01 -4.77933377e-01 -4.91364449e-01 6.60893917e-02
-8.92899096e-01 3.51120800e-01 5.90843856e-01 1.10233378e+00
2.10350618e-01 -6.01235807e-01 7.42193282e-01 -4.10275608e-01
1.08802438e+00 -6.83732569e-01 -3.13752107e-02 3.39209259e-01
-6.06098734e-02 1.53253496e-01 5.59288979e-01 -5.47902822e-01
-1.19290173e+00 -2.93344736e-01 1.87409475e-01 -5.58190703e-01
-4.54205096e-01 1.74151957e-01 -5.57799414e-02 4.64352965e-01
1.04447806e+00 4.21707451e-01 5.81740700e-02 1.51304349e-01
9.65795577e-01 4.70469356e-01 8.59165728e-01 -1.08137763e+00
7.99012363e-01 4.19681698e-01 -2.08844990e-01 -2.21929044e-01
-8.62783551e-01 1.52460709e-01 -7.51169547e-02 -1.30164221e-01
1.25545526e+00 -9.74966824e-01 -1.05698061e+00 1.48423180e-01
-1.55985928e+00 -7.35253453e-01 -2.27791891e-01 2.20035583e-01
-1.14868057e+00 3.63130093e-01 -7.61033714e-01 -6.88824594e-01
-4.24594998e-01 -9.00421381e-01 1.20015872e+00 9.41568762e-02
-6.62224293e-01 -1.08613181e+00 7.76016489e-02 8.93824041e-01
3.85078907e-01 4.88188416e-01 1.16769087e+00 -5.62855244e-01
-4.98167247e-01 4.72496122e-01 -5.24427354e-01 5.23261309e-01
-2.23265126e-01 6.35352507e-02 -1.00073779e+00 1.71661347e-01
-7.15140700e-01 -1.58788741e+00 1.03691959e+00 3.91464867e-02
1.11590290e+00 -4.29637223e-01 1.46569416e-01 3.26308846e-01
1.11347675e+00 -2.76742410e-02 8.12933445e-01 2.93612748e-01
6.17063761e-01 6.91573501e-01 4.22629684e-01 4.67986614e-01
8.66340995e-01 3.49498391e-01 7.62490153e-01 7.90244788e-02
-2.73331732e-01 -5.29658139e-01 7.74960935e-01 1.17419504e-01
-6.00834452e-02 -4.26496983e-01 -1.01061988e+00 7.10225940e-01
-1.91747975e+00 -1.68809152e+00 2.18026221e-01 1.53123021e+00
1.44083941e+00 -3.31503838e-01 1.67092681e-02 -4.27440226e-01
6.89189255e-01 7.73722604e-02 -9.02754843e-01 -8.11048806e-01
-4.05203849e-01 -1.56731624e-02 -3.79158929e-02 4.65398997e-01
-9.08399165e-01 1.22801244e+00 6.06760645e+00 8.36959004e-01
-7.51586378e-01 -4.61093001e-02 5.26072383e-01 -4.37879324e-01
-8.07803869e-01 -1.10633776e-01 -1.90322861e-01 3.27951789e-01
6.64486885e-01 -2.52280921e-01 9.78622735e-01 7.28108525e-01
-1.52442669e-02 -2.10093204e-02 -1.50511980e+00 1.26379013e+00
4.31891471e-01 -1.14282119e+00 4.90732849e-01 -2.50620663e-01
8.68896246e-01 5.55500351e-02 4.77470249e-01 7.04140425e-01
9.80919838e-01 -1.44644928e+00 8.30421388e-01 6.53305531e-01
8.56786072e-01 -6.19921803e-01 3.39296401e-01 1.75065603e-02
-2.93956250e-01 4.27894555e-02 -2.88356930e-01 1.32482842e-01
1.92984119e-01 2.58535713e-01 -1.05074632e+00 1.45457566e-01
5.06284833e-01 6.63349509e-01 -6.08497977e-01 2.58560240e-01
-5.50067604e-01 4.07337368e-01 -1.01464326e-02 -2.70780355e-01
2.77902126e-01 2.27877840e-01 4.81803656e-01 1.27813184e+00
3.22809994e-01 -5.61394244e-02 -1.33419380e-01 1.63675678e+00
-6.20771408e-01 -4.28771898e-02 -7.63604045e-01 -5.69108009e-01
4.02896583e-01 1.11292756e+00 1.42027186e-02 -5.18184543e-01
-1.56125546e-01 1.31047475e+00 3.73439342e-01 5.97982943e-01
-1.15526819e+00 -4.99498934e-01 6.00504816e-01 -2.88690954e-01
1.23884842e-01 1.44396782e-01 -3.36428732e-01 -1.23830044e+00
-3.24235886e-01 -1.21283937e+00 4.11967874e-01 -1.75300825e+00
-1.95434749e+00 2.99070477e-01 1.42902359e-01 -8.32057297e-01
-7.69743860e-01 -8.59903693e-01 -5.69058716e-01 7.27881253e-01
-1.38006771e+00 -1.52740407e+00 -3.50304961e-01 1.13338435e+00
4.08403903e-01 -2.58217305e-01 8.00522029e-01 -4.17699605e-01
-1.13943748e-01 5.74748635e-01 -2.70121813e-01 4.18647349e-01
1.28641868e+00 -1.50312924e+00 2.84516662e-01 2.54022002e-01
1.96274146e-02 5.39978206e-01 8.96103501e-01 -6.31228566e-01
-1.33089137e+00 -7.58610845e-01 6.44799829e-01 -8.54449153e-01
1.15076983e+00 -3.38794410e-01 -5.32064676e-01 8.74676645e-01
9.97085333e-01 -3.98676842e-01 9.43777621e-01 9.63544622e-02
-1.08958650e+00 3.03057253e-01 -1.25139606e+00 1.08707857e+00
1.15371883e+00 -7.89575219e-01 -9.06560838e-01 4.92576122e-01
9.20567870e-01 -3.72372925e-01 -8.29380155e-01 -2.44563054e-02
4.48159009e-01 -8.27344358e-01 9.12819624e-01 -8.98986340e-01
1.49952924e+00 -1.02992050e-01 -4.41504449e-01 -1.62996674e+00
-6.46032318e-02 -7.14464247e-01 -4.20417562e-02 1.22299504e+00
6.29547238e-01 -3.76290947e-01 3.75618905e-01 6.73991382e-01
1.73597306e-01 -2.52222091e-01 -7.84682751e-01 -7.00914443e-01
4.89561200e-01 -4.54908282e-01 5.60906529e-01 1.27644467e+00
4.07908559e-01 6.47042394e-01 -4.57378268e-01 -2.26996958e-01
7.46048212e-01 1.01168752e-01 8.09893250e-01 -7.03812778e-01
-6.44229174e-01 -6.42376065e-01 -1.53601736e-01 -7.40589321e-01
6.65142715e-01 -1.23712933e+00 1.70140058e-01 -1.74437702e+00
5.91486096e-01 1.40791550e-01 -3.76968384e-02 9.38386083e-01
-1.52111262e-01 2.71186531e-01 5.08564174e-01 -1.57748789e-01
-8.93199921e-01 8.00090730e-01 1.53523529e+00 -5.66946805e-01
2.48556718e-01 -8.12829256e-01 -1.08657813e+00 6.19538248e-01
9.30560291e-01 -9.23229232e-02 -6.25980675e-01 -7.07715750e-01
7.36086428e-01 7.84658343e-02 1.20092356e+00 -5.65453827e-01
-3.88686582e-02 -5.49557328e-01 4.87298816e-01 -1.89118460e-01
5.52739143e-01 -4.73926127e-01 -2.99489975e-01 7.02795833e-02
-1.01278293e+00 -1.09332599e-01 3.77156049e-01 4.68801826e-01
-1.55084684e-01 -2.02666655e-01 6.51037157e-01 -2.15071633e-01
-8.05317342e-01 -2.09242970e-01 -4.01022770e-02 7.02866971e-01
8.55094492e-01 1.42102629e-01 -1.08907902e+00 -1.04054368e+00
-8.11790884e-01 4.18336183e-01 5.37800312e-01 5.18778324e-01
8.69439602e-01 -1.59773982e+00 -1.00304163e+00 -8.12149107e-01
5.38896084e-01 -3.91624480e-01 4.38902080e-01 6.37120426e-01
-9.06869322e-02 1.91021100e-01 -3.79997551e-01 -4.66781437e-01
-1.04483092e+00 6.74400866e-01 2.65891701e-01 2.94960812e-02
-1.14958808e-01 9.43531752e-01 2.35160992e-01 -6.64395869e-01
-8.06781575e-02 -1.53933480e-01 2.14248180e-01 -1.78548679e-01
4.39538985e-01 1.20958954e-01 -6.98282778e-01 -4.86058116e-01
-2.76003808e-01 2.75757521e-01 -6.78472966e-02 -5.52173316e-01
9.46901858e-01 4.49909689e-03 -1.64021388e-01 4.40425903e-01
1.03252840e+00 -2.82179266e-01 -1.21046138e+00 -6.72861859e-02
-4.11927879e-01 -1.81276649e-01 -3.38778168e-01 -1.59765494e+00
-5.73124349e-01 1.13387299e+00 2.22074613e-01 -1.35527298e-01
1.09079611e+00 4.02434528e-01 8.27363014e-01 8.53292465e-01
4.17887680e-02 -1.31232381e+00 7.69102037e-01 7.08728492e-01
1.48825884e+00 -1.59185970e+00 -4.07315731e-01 1.79778665e-01
-1.48709977e+00 9.99085963e-01 8.55382740e-01 7.03976378e-02
-1.46144554e-01 9.09362510e-02 -8.36989656e-03 -2.11049449e-02
-1.27121079e+00 -3.77506524e-01 2.92052686e-01 1.10984969e+00
5.44155598e-01 1.79815471e-01 2.26889044e-01 6.30339086e-01
-5.61866581e-01 1.19524688e-01 6.33937836e-01 5.91214418e-01
-2.46681020e-01 -7.50078678e-01 -5.11425912e-01 1.32430226e-01
-1.10650524e-01 -5.97047627e-01 -1.04278648e+00 6.42596781e-01
7.74458423e-02 1.04346406e+00 -2.74905600e-02 -3.33497882e-01
1.76588088e-01 2.74592727e-01 8.26274097e-01 -4.33970898e-01
-4.31853563e-01 -4.83847678e-01 3.98292392e-01 -5.94385564e-01
-3.66756082e-01 -6.32965922e-01 -1.49637413e+00 -5.70636094e-01
4.11357403e-01 -2.69274563e-01 5.09814203e-01 6.28983796e-01
2.88810462e-01 3.37988347e-01 1.20173395e-01 -6.59642875e-01
-7.56485105e-01 -8.46399486e-01 -2.13087231e-01 9.67646420e-01
2.30408385e-01 -3.47443134e-01 -4.48643088e-01 5.85478246e-01] | [10.86749267578125, 1.580081582069397] |
26a85e77-e597-4b05-8bf6-5f005054a0e4 | acrobat-optimizing-auto-batching-of-dynamic | 2305.10611 | null | https://arxiv.org/abs/2305.10611v1 | https://arxiv.org/pdf/2305.10611v1.pdf | ACRoBat: Optimizing Auto-batching of Dynamic Deep Learning at Compile Time | Dynamic control flow is an important technique often used to design expressive and efficient deep learning computations for applications such as text parsing, machine translation, exiting early out of deep models and so on. However, the resulting control flow divergence makes batching, an important performance optimization, difficult to perform manually. In this paper, we present ACRoBat, a framework that enables efficient automatic batching for dynamic deep learning computations by performing hybrid static+dynamic compiler optimizations and end-to-end tensor code generation. ACRoBat performs up to 8.5X better than DyNet, a state-of-the-art framework for automatic batching, on an Nvidia GeForce RTX 3070 GPU. | ['Todd C. Mowry', 'Phillip B. Gibbons', 'Tianqi Chen', 'Pratik Fegade'] | 2023-05-17 | null | null | null | null | ['code-generation'] | ['computer-code'] | [-3.35058063e-01 -4.08149093e-01 -4.18356270e-01 -5.07006228e-01
-4.61821288e-01 -7.22284555e-01 4.43171024e-01 2.61041015e-01
-4.98171747e-01 4.63805258e-01 2.10048020e-01 -1.16513574e+00
6.55651212e-01 -7.15584397e-01 -6.63772345e-01 -4.21329230e-01
-3.01497191e-01 3.78738552e-01 -1.31674707e-01 -1.58709049e-01
2.94024110e-01 6.83712244e-01 -1.38496184e+00 7.44719803e-01
4.60039198e-01 9.83184695e-01 -1.40281543e-01 1.35514557e+00
-6.18990064e-01 1.22256029e+00 -2.93604076e-01 -6.16087377e-01
2.88456231e-01 1.86454765e-02 -9.61367309e-01 -3.49387020e-01
6.26585484e-01 -5.81453741e-01 3.30109186e-02 7.06891775e-01
6.08771443e-01 -1.06572524e-01 3.31985876e-02 -8.92604172e-01
7.14252442e-02 1.16186106e+00 -4.83342111e-01 4.25476462e-01
-1.18331634e-01 5.05453050e-01 1.14330149e+00 -6.71131670e-01
3.73580694e-01 1.24377966e+00 5.32332122e-01 5.60571551e-01
-1.21268141e+00 -5.39026022e-01 -1.00298658e-01 -2.10734919e-01
-1.07305753e+00 -5.78461885e-01 5.21368980e-01 -5.71407974e-01
1.49601221e+00 3.32128137e-01 5.85328281e-01 1.02722013e+00
5.53788483e-01 1.03539991e+00 8.66544425e-01 -4.01114851e-01
5.78960598e-01 -4.51496750e-01 1.77413374e-01 1.06890702e+00
-2.33183727e-01 3.77767086e-01 -2.44508296e-01 -4.87995416e-01
5.21965861e-01 -4.97826517e-01 3.26672047e-01 1.73137993e-01
-1.28935468e+00 9.39810455e-01 2.89795041e-01 7.07824379e-02
-2.39542425e-01 8.35597157e-01 1.48482358e+00 4.15422082e-01
3.60030562e-01 5.00135124e-01 -1.07500446e+00 -7.49129713e-01
-8.01806688e-01 5.28740406e-01 1.09492612e+00 7.39936888e-01
7.10129797e-01 5.20480037e-01 -4.49970245e-01 5.31595528e-01
-4.67120036e-02 1.33474395e-01 5.55715978e-01 -1.04301989e+00
5.98268509e-01 5.33130765e-01 -3.62606019e-01 -5.95516920e-01
-6.01181746e-01 -3.55331808e-01 -1.02674413e+00 3.21827531e-01
2.71578401e-01 -6.90191984e-01 -7.06764817e-01 1.21397543e+00
6.20997488e-01 -7.55347610e-02 3.36772855e-03 9.41407740e-01
5.06426394e-01 8.54161024e-01 2.44331673e-01 2.02057660e-01
1.31470895e+00 -1.41286802e+00 -2.19586313e-01 -2.89502412e-01
1.50857365e+00 -1.07590127e+00 1.47668016e+00 6.43294632e-01
-7.92246521e-01 -5.04841328e-01 -8.19699943e-01 -5.14419854e-01
-2.63067726e-02 3.75071198e-01 1.71462762e+00 6.78840578e-01
-1.20353174e+00 6.90638959e-01 -1.37241411e+00 4.56490725e-01
3.62201095e-01 5.90835273e-01 5.36565706e-02 2.90054917e-01
-6.24808550e-01 2.44344130e-01 7.05088854e-01 2.57185139e-02
-8.51448357e-01 -1.17152190e+00 -6.45369232e-01 1.61738217e-01
3.56799662e-01 -7.79965937e-01 1.68267334e+00 -1.23203266e+00
-2.12924361e+00 9.72364724e-01 3.26649006e-03 -7.62237787e-01
5.40744007e-01 -6.12627506e-01 4.15675342e-02 -4.49206352e-01
-2.64582157e-01 5.15212059e-01 9.27890778e-01 -3.46569896e-01
-5.32564819e-01 -2.08155215e-01 2.17835784e-01 -1.44859016e-01
1.23766400e-02 4.53952372e-01 -3.72722268e-01 -6.02235973e-01
-6.90368354e-01 -1.06914055e+00 -6.28490686e-01 -2.13402405e-01
-5.70986629e-01 -2.09902361e-01 6.35147572e-01 -6.10215485e-01
1.01395559e+00 -1.98950982e+00 2.23921612e-01 -1.37379438e-01
4.18093562e-01 4.34460521e-01 2.95687914e-02 -6.31074384e-02
-4.81476001e-02 8.21918622e-02 4.86775348e-03 -2.55393654e-01
1.83505699e-01 1.97582707e-01 -5.43094218e-01 2.67691135e-01
3.80420834e-02 8.00403595e-01 -8.50824893e-01 -3.96900654e-01
1.08556189e-01 2.53928810e-01 -1.05790198e+00 4.23137605e-01
-6.47190869e-01 4.37161982e-01 -4.20328617e-01 4.81287956e-01
4.86595690e-01 -2.29020622e-02 1.18358083e-01 -1.61301326e-02
-5.82611322e-01 5.63468695e-01 -7.52938807e-01 1.83294678e+00
-1.05568242e+00 8.97088587e-01 1.70585573e-01 -7.24112451e-01
6.38190269e-01 -1.19341426e-01 2.81673908e-01 -6.20186687e-01
5.04423380e-01 2.43224144e-01 1.04660407e-01 -1.81224078e-01
1.07896185e+00 3.53928149e-01 -3.39813143e-01 5.97647130e-01
-1.72117382e-01 -1.88483909e-01 4.57457662e-01 1.51963934e-01
1.09487200e+00 2.07086772e-01 1.94952622e-01 -5.26058733e-01
5.80340922e-01 2.58402973e-01 3.92281294e-01 5.92120290e-01
2.49401256e-01 1.13836132e-01 1.11197710e+00 -1.13122690e+00
-1.65269136e+00 -4.97833192e-01 6.54759631e-02 2.00700855e+00
-8.82742107e-01 -8.36523354e-01 -9.79839742e-01 -5.22280097e-01
-2.47984067e-01 7.78070450e-01 -3.35002273e-01 7.71723315e-02
-1.21598327e+00 -1.00519001e+00 8.39854479e-01 6.52798057e-01
5.43668985e-01 -7.81291008e-01 -9.06113327e-01 5.37457287e-01
3.90591949e-01 -1.19205832e+00 -7.18108773e-01 3.20226550e-01
-1.28654790e+00 -4.45966870e-01 -1.66010991e-01 -5.35002947e-01
3.31279874e-01 -3.82365108e-01 1.74567449e+00 3.26736122e-01
-3.90768230e-01 -3.21527570e-01 -1.98652059e-01 4.86029516e-04
-8.43613982e-01 6.67590976e-01 -2.58938432e-01 -3.23255479e-01
4.93446738e-02 -4.05228734e-01 -4.92742956e-01 -2.28352174e-01
-8.05076301e-01 5.99245369e-01 4.62710917e-01 1.11955237e+00
6.38707817e-01 -3.04296166e-01 -4.40942615e-01 -1.10630894e+00
6.17288709e-01 -1.37808457e-01 -1.38744843e+00 -1.55584648e-01
-4.53013301e-01 7.28467524e-01 1.32901859e+00 -5.57326317e-01
-8.86914313e-01 1.93903461e-01 -7.41276383e-01 -2.72819400e-01
7.36569345e-04 3.44757229e-01 1.19175948e-01 -2.62217462e-01
8.92623901e-01 -1.72277614e-01 -2.06655204e-01 -3.77306074e-01
5.47262490e-01 4.59223479e-01 4.48694766e-01 -1.16001952e+00
1.81222498e-01 2.50610620e-01 1.31152943e-01 -7.66240001e-01
-6.96710527e-01 -1.65013075e-01 -5.29032588e-01 3.28364789e-01
6.96931422e-01 -8.07547390e-01 -1.10442615e+00 7.81441212e-01
-1.38636053e+00 -1.23138630e+00 -3.93541008e-02 8.58344883e-02
-4.65765536e-01 4.75260317e-02 -1.04375613e+00 -2.72992194e-01
-1.07981193e+00 -1.63884223e+00 1.37972891e+00 -4.73063141e-02
-1.55017033e-01 -1.14912558e+00 9.49169770e-02 2.23341465e-01
6.01492882e-01 2.11975411e-01 9.90347147e-01 -3.46830875e-01
-5.99342585e-01 2.84768548e-02 -2.84502447e-01 4.46625501e-01
-6.61623538e-01 5.30706167e-01 -7.26258695e-01 -2.54020542e-01
-2.80767173e-01 -3.26904565e-01 6.80666089e-01 1.26034424e-01
1.72543299e+00 -6.54629946e-01 -2.78051081e-03 1.39535046e+00
1.46729052e+00 -1.84143946e-01 6.42293036e-01 1.91441223e-01
1.26803029e+00 1.14950247e-01 2.06206381e-01 8.30146670e-01
2.76982725e-01 6.29601777e-01 2.93319941e-01 -5.39203286e-02
1.68522462e-01 1.39766693e-01 5.41531265e-01 1.10221720e+00
7.81178772e-02 1.79067001e-01 -1.22354746e+00 1.27148867e-01
-1.39362192e+00 -2.29102403e-01 -4.08359587e-01 1.92951524e+00
1.14792252e+00 1.79702103e-01 -2.70413514e-03 8.50124517e-04
1.72377467e-01 3.51627320e-02 -4.04723376e-01 -1.44432211e+00
3.73589307e-01 8.65427613e-01 9.83398736e-01 4.79270101e-01
-1.15989351e+00 1.65373230e+00 5.41680527e+00 9.62458074e-01
-1.79590905e+00 2.91201442e-01 1.01131690e+00 -2.47953162e-02
-1.06345549e-01 1.34880587e-01 -1.11471701e+00 2.43894264e-01
1.35403681e+00 1.63818181e-01 6.93309903e-01 1.47352564e+00
-2.80191675e-02 -1.57932006e-02 -1.29290366e+00 8.28729391e-01
-4.69818324e-01 -1.74661422e+00 -3.28329831e-01 -2.64316965e-02
7.53411174e-01 3.49129498e-01 -5.84253259e-02 6.58308268e-01
7.27314770e-01 -9.42435682e-01 7.64761269e-01 -1.61880359e-01
8.03497374e-01 -9.83750224e-01 6.39858127e-01 -1.22910906e-02
-1.01183844e+00 3.41574401e-02 -2.71416932e-01 -2.58497983e-01
-7.96191394e-03 8.85782063e-01 -9.19727981e-01 -2.43809059e-01
6.62943721e-01 7.57298619e-02 -5.81201375e-01 6.07874632e-01
-5.61030321e-02 8.65113020e-01 -1.90202296e-01 -3.79050106e-01
6.52074754e-01 3.86455283e-02 2.55196840e-01 1.69882989e+00
-2.27452274e-02 -1.56796992e-01 2.37069294e-01 6.96319818e-01
-2.45014250e-01 3.40952933e-01 -1.98094517e-01 -3.02836329e-01
-6.77587688e-02 1.35440385e+00 -8.80590558e-01 -7.51138985e-01
-2.33301058e-01 1.08526826e+00 4.75526333e-01 -3.16621624e-02
-1.05199027e+00 -4.81187075e-01 9.76685286e-01 -1.86057836e-01
3.50388348e-01 -8.31560254e-01 -9.16468203e-01 -1.45219064e+00
-1.10506900e-01 -1.15104818e+00 1.05785452e-01 -1.84905037e-01
-7.30853319e-01 6.12880707e-01 -2.33257115e-01 -5.29864609e-01
-5.03010988e-01 -9.75747764e-01 -6.66826785e-01 6.26888931e-01
-1.02886856e+00 -9.30179179e-01 -1.20463267e-01 2.57515877e-01
4.63357717e-01 -2.37564877e-01 7.37552106e-01 4.33463097e-01
-6.18109167e-01 9.67521667e-01 1.79126665e-01 3.04129392e-01
1.33230519e-02 -1.37344956e+00 1.26918435e+00 8.30377698e-01
-2.38447875e-01 5.66684186e-01 6.64564073e-01 -5.06057382e-01
-2.14245653e+00 -1.23805892e+00 5.71463704e-01 1.00092843e-01
1.09730232e+00 -6.47359610e-01 -7.21103549e-01 5.93327105e-01
1.43699229e-01 2.28173018e-01 5.23797274e-01 1.89437121e-01
-4.36003625e-01 -3.21146071e-01 -6.49698019e-01 8.60978544e-01
6.93036318e-01 -3.57491612e-01 7.11704493e-02 6.67454481e-01
1.13397050e+00 -1.24461150e+00 -9.37502563e-01 -4.21312898e-02
5.25711298e-01 -9.33165133e-01 8.37157607e-01 -7.06744552e-01
6.81496084e-01 1.85879052e-01 -1.00474201e-01 -1.02069008e+00
-5.11485934e-02 -1.16880357e+00 -3.86805803e-01 1.04247320e+00
2.46417582e-01 -3.21745187e-01 9.60983515e-01 4.08480227e-01
-5.37910223e-01 -6.85474038e-01 -6.42091691e-01 -3.81210536e-01
3.69395375e-01 -8.80272150e-01 8.55421960e-01 8.48917544e-01
-5.26831627e-01 4.92271215e-01 -3.33505332e-01 -2.69962966e-01
4.09842938e-01 7.76384175e-02 1.18781543e+00 -6.77013993e-01
-6.56349421e-01 -7.98267186e-01 -1.94902778e-01 -1.04991817e+00
5.18700302e-01 -8.59695315e-01 -1.25409663e-01 -5.21363437e-01
-2.95995563e-01 -6.99270010e-01 5.06694257e-01 6.20336652e-01
1.11149345e-02 2.07371116e-02 4.82257344e-02 -1.27705678e-01
-2.47891366e-01 3.34020078e-01 1.15545487e+00 -9.66156423e-02
-4.04383779e-01 -1.29490018e-01 -5.11821806e-01 6.62170291e-01
9.34746623e-01 -3.76244605e-01 1.11558095e-01 -8.64441574e-01
7.23920584e-01 2.64058322e-01 1.57778844e-01 -8.02307725e-01
-1.17276005e-01 -3.35738391e-01 1.82757769e-02 -3.66010010e-01
-1.53340161e-01 -2.42575005e-01 -1.02501467e-01 6.70483351e-01
-4.75108474e-01 6.52280211e-01 6.06784582e-01 -2.60731578e-01
8.86511896e-03 -2.80116260e-01 7.93255687e-01 -3.75498950e-01
-9.62029696e-01 4.84523237e-01 -5.02380133e-01 3.52347910e-01
6.36685610e-01 5.01576304e-01 -2.18992263e-01 2.87643999e-01
-2.90790975e-01 1.91724545e-03 3.39510500e-01 1.84994861e-01
3.64267707e-01 -1.04192638e+00 -8.23347509e-01 3.21004540e-01
-3.05223167e-01 2.48637304e-01 2.37350807e-01 7.49531984e-01
-1.59577799e+00 4.40124840e-01 -2.31331751e-01 -6.53164029e-01
-1.12627459e+00 6.52951002e-01 2.27038205e-01 -7.43664265e-01
-8.54292870e-01 9.69043076e-01 -5.12228161e-02 -5.37372589e-01
4.85625677e-02 -8.61266375e-01 3.10692638e-01 -3.27180535e-01
6.37087941e-01 3.86436522e-01 5.67053258e-01 -2.78969347e-01
-1.28652811e-01 1.41819805e-01 -1.37252837e-01 4.78655957e-02
9.57667351e-01 4.66858327e-01 -8.41499567e-01 1.10143408e-01
1.56766963e+00 -5.73861264e-02 -9.86306906e-01 1.27844259e-01
5.50592989e-02 -1.83536887e-01 6.25406921e-01 -3.62725914e-01
-1.35269761e+00 1.04066849e+00 5.91028869e-01 -2.97529966e-01
1.01610041e+00 -5.10893404e-01 1.32057631e+00 7.60405421e-01
2.79220939e-01 -1.10456252e+00 -7.53291696e-02 1.17213643e+00
5.18033445e-01 -1.17052126e+00 -1.06348116e-02 -9.59444940e-02
-4.31729525e-01 1.47835910e+00 5.75533569e-01 -2.29591608e-01
5.10719001e-01 1.02097237e+00 -2.17209071e-01 1.46754026e-01
-1.17433822e+00 2.00793594e-01 1.42217517e-01 1.61532015e-01
8.87035251e-01 3.67972940e-01 -3.44729424e-01 2.71690972e-02
-7.02438474e-01 9.10164937e-02 5.90185583e-01 9.22949493e-01
8.75108093e-02 -1.15992427e+00 -3.55670094e-01 4.60284978e-01
-9.41227257e-01 -5.55595815e-01 1.15377307e-01 5.36093056e-01
-1.72838524e-01 2.48697937e-01 4.91054535e-01 -6.04822099e-01
-3.20779741e-01 -1.99960023e-01 4.03373599e-01 -5.04725695e-01
-1.32930672e+00 -4.77240756e-02 3.90764713e-01 -7.70428836e-01
3.15514445e-01 -3.21221709e-01 -1.27171326e+00 -9.04245794e-01
8.13686028e-02 -9.44397375e-02 1.10126936e+00 7.92801857e-01
5.76224089e-01 6.79139853e-01 5.47504485e-01 -1.12489450e+00
-5.56708097e-01 -6.33261144e-01 -1.21892162e-03 -3.19447257e-02
2.92121261e-01 -7.57305622e-02 -1.44485787e-01 1.27477422e-01] | [8.46313762664795, 3.25398325920105] |
7b73fd0c-f6eb-4163-a277-fda635e93a42 | making-the-most-of-tweet-inherent-features | 1503.07405 | null | http://arxiv.org/abs/1503.07405v1 | http://arxiv.org/pdf/1503.07405v1.pdf | Making the Most of Tweet-Inherent Features for Social Spam Detection on Twitter | Social spam produces a great amount of noise on social media services such as
Twitter, which reduces the signal-to-noise ratio that both end users and data
mining applications observe. Existing techniques on social spam detection have
focused primarily on the identification of spam accounts by using extensive
historical and network-based data. In this paper we focus on the detection of
spam tweets, which optimises the amount of data that needs to be gathered by
relying only on tweet-inherent features. This enables the application of the
spam detection system to a large set of tweets in a timely fashion, potentially
applicable in a real-time or near real-time setting. Using two large
hand-labelled datasets of tweets containing spam, we study the suitability of
five classification algorithms and four different feature sets to the social
spam detection task. Our results show that, by using the limited set of
features readily available in a tweet, we can achieve encouraging results which
are competitive when compared against existing spammer detection systems that
make use of additional, costly user features. Our study is the first that
attempts at generalising conclusions on the optimal classifiers and sets of
features for social spam detection over different datasets. | ['Procter Rob', 'Liakata Maria', 'Zubiaga Arkaitz', 'Wang Bo'] | 2015-03-25 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 1.48052976e-01 -3.19577992e-01 -1.27417305e-02 -3.80817562e-01
-4.56903636e-01 -4.74851370e-01 1.19711065e+00 6.14011586e-01
-7.65605748e-01 4.78498399e-01 2.42184736e-02 -6.28146708e-01
-1.55854583e-01 -1.09047961e+00 1.07821807e-01 -4.50574160e-01
-2.12204278e-01 4.76511151e-01 8.14646006e-01 -6.11024797e-01
4.32048589e-01 4.96912628e-01 -1.80196154e+00 2.68362403e-01
6.11920476e-01 8.91918063e-01 -1.01365753e-01 5.59583485e-01
-3.58792365e-01 6.85362339e-01 -8.01167965e-01 -3.15908611e-01
2.69190073e-01 -3.60543281e-01 -7.44335771e-01 1.08303182e-01
1.92427948e-01 -1.70820937e-01 -2.09308982e-01 9.31866229e-01
3.38207453e-01 1.64425075e-01 4.33632791e-01 -1.28239548e+00
7.58786723e-02 3.48495513e-01 -3.02300960e-01 5.93741417e-01
5.36543727e-01 5.06837554e-02 7.69771218e-01 -4.08607960e-01
4.80066478e-01 1.33026230e+00 7.56263614e-01 2.43046075e-01
-1.30567873e+00 -7.20047951e-01 2.84187533e-02 -9.77367349e-03
-9.89178658e-01 -4.71107423e-01 4.49786514e-01 -3.38327467e-01
6.74215019e-01 6.28598571e-01 4.34233814e-01 8.46871138e-01
-7.12472498e-02 5.07295847e-01 1.36313593e+00 -4.26546156e-01
5.12653887e-02 4.74842131e-01 3.72697592e-01 2.73366123e-01
4.08197522e-01 7.89093897e-02 -3.86140347e-01 -8.93911064e-01
3.09346728e-02 1.74302295e-01 5.33284321e-02 -1.45591851e-02
-7.32340813e-01 1.24268126e+00 3.71170193e-01 5.44769466e-01
-1.55963868e-01 -2.06649721e-01 7.63540983e-01 9.17831838e-01
1.00494432e+00 5.33457041e-01 -2.80083120e-01 -5.64200170e-02
-1.14841592e+00 2.20558956e-01 1.20585608e+00 5.43314636e-01
6.73040032e-01 -2.67597884e-01 1.77352384e-01 8.20833147e-01
2.75065720e-01 5.10035813e-01 8.84463191e-01 -4.15893227e-01
2.99492300e-01 9.27162051e-01 2.39612520e-01 -1.29755199e+00
-5.44201195e-01 -2.01930001e-01 -2.80609667e-01 -3.55368815e-02
6.84954762e-01 -1.51404619e-01 -6.35983825e-01 1.25593114e+00
4.53587919e-01 1.45955989e-02 -3.33909541e-01 6.96006536e-01
7.37275600e-01 3.22555661e-01 1.22551762e-01 -3.63221884e-01
1.34304142e+00 -3.53230268e-01 -3.91258299e-01 -3.71775210e-01
7.74595976e-01 -9.82656240e-01 9.09445226e-01 2.62568831e-01
-5.89010894e-01 -2.74524111e-02 -6.67018354e-01 3.79414558e-01
-9.44305718e-01 -4.79645878e-01 6.06767297e-01 1.05851483e+00
-6.22549832e-01 8.94646108e-01 -5.55664718e-01 -8.14971030e-01
2.72308677e-01 6.54256046e-01 -2.73875624e-01 2.95552045e-01
-1.33577180e+00 1.02982903e+00 1.27221242e-01 -1.72870219e-01
-1.55451655e-01 -4.79120612e-02 -3.47012579e-01 -1.46297678e-01
3.48735571e-01 -1.83072865e-01 1.34263778e+00 -1.38529670e+00
-9.97342229e-01 1.09568048e+00 -1.22990362e-01 -6.73776209e-01
9.00959373e-01 1.36653453e-01 -6.40869319e-01 1.89949319e-01
3.22159201e-01 -1.25126541e-01 1.17546487e+00 -1.11077285e+00
-9.51747775e-01 -3.53553236e-01 -3.37336421e-01 -3.30589950e-01
-5.96228600e-01 6.40388012e-01 1.13296080e-02 -5.99247456e-01
1.87425062e-01 -1.04598713e+00 -2.69399732e-01 -5.48075378e-01
-1.25522062e-01 -2.48708159e-01 1.24379647e+00 -2.98640877e-01
1.43718958e+00 -1.89023376e+00 -5.42472184e-01 6.53813362e-01
3.21163297e-01 7.11881876e-01 1.45832896e-01 8.67042840e-01
6.82854354e-02 2.01734647e-01 -1.64892599e-02 -2.98585504e-01
-1.80479527e-01 1.29874825e-01 -4.31840241e-01 8.34108710e-01
1.66086815e-02 6.37952566e-01 -1.19240820e+00 -5.12227178e-01
2.70931661e-01 -5.43694645e-02 -2.96436585e-02 1.40944600e-01
2.68521100e-01 6.30218536e-02 -6.49905205e-01 3.08721095e-01
4.87573922e-01 -7.83644468e-02 8.33730921e-02 5.61420023e-01
-7.15697557e-02 6.08352363e-01 -1.01080239e+00 5.11415184e-01
-4.95664030e-01 8.71376634e-01 2.14528114e-01 -1.05433011e+00
9.71349478e-01 1.68607831e-01 4.31989968e-01 -4.07416612e-01
6.40370429e-01 4.57853526e-01 -1.01821208e-02 -2.62022436e-01
4.13284868e-01 -3.13888580e-01 -1.91226795e-01 9.87146258e-01
-2.44638160e-01 1.18584614e-02 5.40861726e-01 4.34718847e-01
1.37745440e+00 -7.57443666e-01 2.88462400e-01 -5.15872419e-01
7.94501543e-01 2.24958491e-02 -5.00760935e-02 1.18548667e+00
-2.93449670e-01 1.89720184e-01 6.08067214e-01 -3.19160402e-01
-9.38600302e-01 -7.45311141e-01 -2.74721473e-01 1.44213426e+00
7.26284459e-02 -3.95898342e-01 -5.93369007e-01 -8.67606878e-01
4.53890473e-01 5.02750933e-01 -4.91942644e-01 -5.55871911e-02
-6.47499025e-01 -1.36214066e+00 3.59659731e-01 -1.66379079e-01
1.32480621e-01 -1.08715546e+00 -4.11440134e-01 4.69555706e-01
1.30325586e-01 -8.54847848e-01 -2.01943874e-01 2.19831973e-01
-1.03072059e+00 -1.36810195e+00 -1.37700856e-01 -5.68908870e-01
6.98786438e-01 8.03619087e-01 9.99122739e-01 6.48170292e-01
-1.33665457e-01 2.12841272e-01 -6.23744845e-01 -6.46604240e-01
-8.07969213e-01 3.65963399e-01 5.04912771e-02 2.05061242e-01
9.81894195e-01 -5.89477539e-01 -2.28388697e-01 6.77758753e-01
-9.48849142e-01 -4.72641587e-01 3.06357175e-01 8.58958304e-01
-5.67078948e-01 7.79989809e-02 1.25462234e+00 -1.44147062e+00
1.21325350e+00 -8.50223958e-01 -5.00733554e-01 -3.44449818e-01
-7.58663356e-01 -2.65400350e-01 7.42551506e-01 -4.67459738e-01
-7.94809520e-01 -4.16616350e-02 -1.16058150e-02 3.27562451e-01
-1.51625555e-02 2.50259519e-01 4.36902940e-01 -2.67668068e-01
1.05934680e+00 9.62871462e-02 4.74955767e-01 -4.66288805e-01
1.25945374e-01 1.47991848e+00 2.57963445e-02 3.36922752e-03
8.94940853e-01 6.73317850e-01 -1.84904575e-01 -1.21429658e+00
-7.23968744e-01 -1.39219856e+00 -6.04421020e-01 -7.97511935e-02
-1.21595412e-01 -4.35168326e-01 -6.56630993e-01 6.15092456e-01
-5.92837989e-01 2.19175458e-01 -1.89164698e-01 9.42729414e-02
-3.60233307e-01 7.13469744e-01 -6.04537427e-01 -1.02665603e+00
-4.25443083e-01 -6.62729263e-01 8.33640456e-01 -2.38692194e-01
-6.68943465e-01 -1.01143646e+00 -2.46532723e-01 2.59113729e-01
5.93910515e-01 9.93851721e-02 2.84652382e-01 -1.30878913e+00
-2.51533628e-01 -1.18384814e+00 -4.14364815e-01 3.53039145e-01
3.86570990e-01 -2.99525052e-01 -9.81190205e-01 -5.41917026e-01
1.86028510e-01 -1.53554425e-01 1.10932982e+00 -8.82531852e-02
6.85053110e-01 -3.34045082e-01 -5.40431559e-01 -1.91650346e-01
1.00460804e+00 -2.44985193e-01 2.99951285e-01 7.41644979e-01
1.69506133e-01 8.83808911e-01 5.56427538e-01 3.55024487e-01
-9.88629162e-02 2.70567268e-01 3.55554283e-01 6.38001561e-02
3.91885817e-01 -2.07204670e-02 1.45930976e-01 5.21649301e-01
3.06116164e-01 -3.12872738e-01 -7.09839702e-01 7.94181108e-01
-2.12715578e+00 -1.08555138e+00 -5.84070861e-01 2.37243390e+00
6.71905339e-01 4.26145583e-01 7.54497707e-01 5.16185641e-01
9.92498457e-01 1.27300099e-01 1.08105890e-01 -4.27898765e-01
1.83862522e-01 4.90321741e-02 9.42449450e-01 4.58863556e-01
-1.42296791e+00 7.44324148e-01 6.80698299e+00 7.78819919e-01
-1.00679410e+00 1.04543261e-01 1.44118994e-01 -2.25591883e-01
1.86787665e-01 -3.43964249e-03 -7.92298675e-01 9.37546015e-01
1.45370018e+00 -4.24464434e-01 2.02079177e-01 7.37881124e-01
7.65855908e-01 -5.51157653e-01 -7.21640468e-01 5.42589068e-01
-1.90988723e-02 -9.62334931e-01 -2.23828271e-01 1.31477907e-01
5.25949180e-01 2.63655216e-01 -2.86911577e-01 9.25640836e-02
2.90729642e-01 -7.69247174e-01 3.99944276e-01 3.07181925e-01
3.19426842e-02 -6.29617035e-01 1.05279374e+00 6.65853500e-01
-6.07917309e-01 -4.95683879e-01 6.86690360e-02 -3.65701526e-01
3.47025007e-01 7.90827572e-01 -1.21802926e+00 3.14051360e-02
5.77542841e-01 5.26988208e-01 -8.72095585e-01 1.19037259e+00
6.73732013e-02 8.58499885e-01 -7.35733926e-01 -6.02797270e-01
4.69128340e-01 2.13089865e-03 6.62553191e-01 1.59799564e+00
6.85429154e-03 -2.64782757e-01 2.15983763e-01 2.86235929e-01
9.43152905e-02 4.28716958e-01 -6.71264648e-01 6.66389391e-02
3.78012240e-01 1.25252593e+00 -1.05685282e+00 -4.54934299e-01
-4.83277470e-01 4.76581722e-01 1.09836005e-01 -1.92803174e-01
-1.47177473e-01 -5.80913067e-01 4.14416075e-01 8.51228654e-01
-1.20448083e-01 -1.41282558e-01 -1.49228841e-01 -9.18878913e-01
-1.49484277e-01 -8.66017401e-01 4.35332328e-01 -2.84374468e-02
-1.78286183e+00 3.82071316e-01 4.62285057e-03 -1.23252654e+00
-3.99662495e-01 -6.69077039e-01 -8.57833326e-01 8.62195253e-01
-1.17826295e+00 -1.03538120e+00 -1.54768035e-01 1.99704200e-01
1.77736521e-01 -3.22901905e-01 5.77727854e-01 2.99647093e-01
-2.41880491e-01 2.05293179e-01 5.13845265e-01 1.41247272e-01
8.32260013e-01 -1.12409472e+00 6.83329999e-01 6.11889839e-01
-2.21730515e-01 7.18948245e-01 1.04387224e+00 -8.16683769e-01
-8.52165103e-01 -9.60298419e-01 1.20412385e+00 -8.67363751e-01
1.30474281e+00 -3.51197213e-01 -9.87542272e-01 3.76042366e-01
-3.30080152e-01 -1.62744179e-01 6.03555322e-01 2.70891041e-01
5.59553169e-02 1.68822035e-01 -1.25613904e+00 2.76423067e-01
7.28385448e-01 -5.39099336e-01 -5.75896561e-01 5.87532103e-01
2.20675897e-02 1.23527676e-01 -4.67582583e-01 -1.87583547e-02
5.59606969e-01 -9.84359741e-01 7.82206595e-01 -8.85872364e-01
-8.64318237e-02 -1.23168476e-01 3.04333091e-01 -1.16646171e+00
-1.92488000e-01 -7.09921539e-01 2.14404866e-01 1.13641882e+00
3.59771848e-01 -1.06328714e+00 9.14880872e-01 3.69956851e-01
4.21936601e-01 -3.89761567e-01 -8.50083292e-01 -8.66756618e-01
-1.00421615e-01 -4.86534506e-01 3.68815988e-01 8.30190539e-01
2.63592899e-01 3.01782906e-01 -2.96670169e-01 -3.22040856e-01
4.31792468e-01 1.67489685e-02 9.91402805e-01 -1.74699974e+00
2.13097930e-01 -4.38492358e-01 -7.08842635e-01 -6.34648502e-01
1.02277353e-01 -7.73490191e-01 -1.29318237e-01 -9.73716140e-01
2.86549423e-02 -5.79752982e-01 2.01240964e-02 6.15378991e-02
-1.90641612e-01 5.48931181e-01 1.27572447e-01 4.45307583e-01
-5.71663082e-01 5.99710569e-02 7.10452020e-01 -6.72321245e-02
-4.20812875e-01 6.17723525e-01 -6.55002952e-01 9.47982132e-01
8.50353599e-01 -7.51846731e-01 -1.04729496e-01 3.15088004e-01
1.76088870e-01 -5.90691030e-01 3.57385576e-01 -5.20761132e-01
2.64780313e-01 -1.11109301e-01 3.95773724e-02 -3.07669938e-01
1.05862461e-01 -7.13730991e-01 -3.43756050e-01 5.14372349e-01
-1.17198333e-01 -1.25830725e-01 -4.27144766e-02 8.39929283e-01
-1.47706017e-01 -8.11606407e-01 9.93782043e-01 -4.48289037e-01
-4.78827000e-01 -1.03715688e-01 -9.28634465e-01 7.13657066e-02
1.08148110e+00 -2.77735829e-01 -3.86022389e-01 -6.70114577e-01
-7.54373193e-01 2.86022484e-01 5.73672831e-01 6.40271962e-01
-5.82947172e-02 -9.44976807e-01 -6.93059862e-01 3.49794954e-01
2.22223490e-01 -6.76888943e-01 -3.43868852e-01 8.62471461e-01
-5.65159142e-01 3.63422334e-01 1.04547814e-01 -4.78997052e-01
-1.61207592e+00 4.48955894e-01 2.43258625e-01 -1.57102928e-01
-4.95132923e-01 4.58464533e-01 -5.87048650e-01 -4.27363724e-01
-3.59338685e-03 3.21989834e-01 -3.21970612e-01 6.78797126e-01
5.41555762e-01 6.86580837e-01 4.18525994e-01 -1.01680744e+00
-2.40527138e-01 -2.00186774e-01 -3.01702052e-01 2.96110827e-02
1.15278876e+00 -2.62363374e-01 -4.08871591e-01 3.86887550e-01
1.18615866e+00 1.22479759e-01 -4.31984156e-01 -4.04758036e-01
6.31400883e-01 -9.82549906e-01 1.38273105e-01 -1.71906039e-01
-5.44213653e-01 4.43541437e-01 4.12686408e-01 1.32659912e+00
8.03848982e-01 -5.92542216e-02 8.52687478e-01 3.75590652e-01
2.24345669e-01 -1.42269266e+00 -1.67283416e-01 4.72932786e-01
4.46798474e-01 -1.58656597e+00 3.24588209e-01 -5.12885273e-01
-2.55209833e-01 9.75654125e-01 1.10361896e-01 -3.33617479e-01
9.43524659e-01 -1.37130946e-01 1.27378225e-01 -2.40646824e-01
-5.21269619e-01 -5.23659825e-01 3.17526422e-02 3.75556350e-01
3.95643324e-01 1.55105917e-02 -9.08900619e-01 1.96650222e-01
-3.62218291e-01 -3.16878140e-01 7.66839981e-01 1.23960710e+00
-1.08656204e+00 -1.41176820e+00 -5.28825104e-01 1.37361598e+00
-8.44098806e-01 1.55312896e-01 -9.19765651e-01 7.98024356e-01
-1.14433743e-01 1.47617567e+00 -6.04552180e-02 -7.60343820e-02
4.44471091e-01 1.07937455e-01 -9.78013724e-02 -8.68565083e-01
-9.60835814e-01 -2.57307291e-01 6.36488318e-01 -1.70701519e-01
-4.34032977e-01 -9.47542906e-01 -7.34596968e-01 -8.28981578e-01
-1.02202094e+00 4.66098756e-01 5.80424786e-01 9.89973664e-01
1.49397105e-01 2.40404792e-02 9.59369004e-01 -8.49040806e-01
-8.74013305e-01 -1.19161391e+00 -8.32755208e-01 8.67339432e-01
4.64136869e-01 -5.37805498e-01 -8.47264946e-01 -2.31488720e-01] | [7.910271644592285, 10.039311408996582] |
494c9d47-2551-4e9a-a25e-43c54af24c68 | evaluation-of-speaker-anonymization-on | 2305.01759 | null | https://arxiv.org/abs/2305.01759v1 | https://arxiv.org/pdf/2305.01759v1.pdf | Evaluation of Speaker Anonymization on Emotional Speech | Speech data carries a range of personal information, such as the speaker's identity and emotional state. These attributes can be used for malicious purposes. With the development of virtual assistants, a new generation of privacy threats has emerged. Current studies have addressed the topic of preserving speech privacy. One of them, the VoicePrivacy initiative aims to promote the development of privacy preservation tools for speech technology. The task selected for the VoicePrivacy 2020 Challenge (VPC) is about speaker anonymization. The goal is to hide the source speaker's identity while preserving the linguistic information. The baseline of the VPC makes use of a voice conversion. This paper studies the impact of the speaker anonymization baseline system of the VPC on emotional information present in speech utterances. Evaluation is performed following the VPC rules regarding the attackers' knowledge about the anonymization system. Our results show that the VPC baseline system does not suppress speakers' emotions against informed attackers. When comparing anonymized speech to original speech, the emotion recognition performance is degraded by 15\% relative to IEMOCAP data, similar to the degradation observed for automatic speech recognition used to evaluate the preservation of the linguistic information. | ['Marie Tahon', 'Anthony Larcher', 'Denis Jouvet', 'Pierre Champion', 'Hubert Nourtel'] | 2023-04-15 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [-6.76417202e-02 6.94156408e-01 2.26322562e-01 -6.23793662e-01
-7.18069196e-01 -8.80908132e-01 8.11150610e-01 1.04062051e-01
-4.49965358e-01 5.29267490e-01 7.55746901e-01 -3.86911213e-01
3.23186487e-01 -1.09300837e-01 -3.72183293e-01 -5.76998472e-01
1.86076611e-01 4.88232709e-02 -1.59551591e-01 -2.09955350e-01
-1.34161338e-01 8.52027833e-01 -1.62254357e+00 4.27270621e-01
4.41404462e-01 6.90179706e-01 -4.55884159e-01 6.52824521e-01
-1.83006763e-01 4.94924724e-01 -1.18353415e+00 -7.80674696e-01
4.45739537e-01 -2.15540126e-01 -9.45563495e-01 -2.88510263e-01
2.83744812e-01 -2.99348384e-01 -4.85764712e-01 1.40248084e+00
8.42745364e-01 8.14127177e-02 2.25203037e-02 -1.74047625e+00
-4.11630310e-02 6.64230347e-01 2.41371438e-01 7.53009021e-02
4.47574496e-01 9.20038968e-02 6.15686655e-01 -2.47162387e-01
7.75957048e-01 1.44805467e+00 4.95898515e-01 9.75968063e-01
-1.39139295e+00 -7.11438000e-01 -6.98109642e-02 -1.15715571e-01
-1.55005407e+00 -1.32195675e+00 6.37910485e-01 -2.23561898e-01
9.10391271e-01 1.05145419e+00 1.57231972e-01 1.58550286e+00
-2.85482556e-01 4.78721887e-01 8.61010134e-01 -3.96810919e-01
4.01348174e-01 1.10166490e+00 4.29780841e-01 8.02694932e-02
7.11007267e-02 1.86058775e-01 -6.80186152e-01 -7.82858014e-01
-3.83973867e-01 -7.74932504e-01 -5.95329940e-01 -5.65130971e-02
-6.46069705e-01 6.06638908e-01 -3.57467145e-01 4.65755373e-01
-1.45246014e-01 -2.97739476e-01 8.74299228e-01 5.73166132e-01
5.37967563e-01 5.56374311e-01 -1.85348690e-01 -4.25334483e-01
-6.31224751e-01 2.57951319e-01 1.40573931e+00 1.03057051e+00
3.84460479e-01 1.11828789e-01 -1.30150110e-01 8.56430173e-01
4.06408191e-01 6.27660334e-01 7.17896461e-01 -9.85840380e-01
6.40812874e-01 4.71415631e-02 1.96851894e-01 -1.06687534e+00
-1.16788847e-02 -9.37911347e-02 -2.57336706e-01 3.06336969e-01
4.07793611e-01 -3.91744941e-01 -4.74727243e-01 1.91148913e+00
2.15942636e-01 -4.01632190e-01 6.01954818e-01 6.20501101e-01
6.73314214e-01 6.61859334e-01 2.89248735e-01 -2.56052017e-01
1.39543724e+00 -2.32934669e-01 -1.32550597e+00 -1.75583646e-01
5.85245788e-01 -8.11484754e-01 9.94508982e-01 5.98481409e-02
-7.60877490e-01 1.12333456e-02 -1.07502890e+00 9.92117729e-03
-6.92181230e-01 -3.72870386e-01 2.66876906e-01 1.90296829e+00
-1.18506801e+00 3.72128218e-01 -6.10471725e-01 -5.70535660e-01
1.67457327e-01 5.80949783e-01 -7.71710813e-01 4.68801349e-01
-1.63357747e+00 8.17060232e-01 1.25442311e-01 -2.37412646e-01
-5.02851725e-01 -6.03303373e-01 -7.94530988e-01 2.11658299e-01
-1.36911839e-01 -3.93046848e-02 1.34163558e+00 -1.07542562e+00
-1.60047483e+00 1.16875076e+00 -2.45490164e-01 -8.16591024e-01
8.31763983e-01 -7.41623435e-03 -1.20765078e+00 -9.96357724e-02
1.70028612e-01 4.12260234e-01 9.55568135e-01 -1.15429759e+00
-5.17382920e-01 -6.02573156e-01 -3.93646121e-01 1.41006604e-01
-4.95010436e-01 4.76664245e-01 -2.88581312e-01 -5.74093521e-01
-2.40123987e-01 -1.10749805e+00 3.35537255e-01 -2.75128424e-01
-7.39109457e-01 3.45427424e-01 1.28319502e+00 -1.10432076e+00
1.18481874e+00 -2.77335572e+00 -3.88738304e-01 3.46653253e-01
-1.19940504e-01 5.93425870e-01 2.45218441e-01 3.95038754e-01
-4.07663643e-01 6.51085079e-01 -4.26330697e-03 -8.68134916e-01
3.26431334e-01 2.11897984e-01 -7.77012646e-01 6.85441077e-01
-4.76197422e-01 3.22093517e-01 -3.63668054e-01 -1.33766875e-01
1.80267006e-01 5.97134650e-01 -3.20786864e-01 8.98350552e-02
1.32652432e-01 1.29698426e-01 -8.58382359e-02 2.49312207e-01
8.17705154e-01 8.87071729e-01 2.35784993e-01 1.80706471e-01
-1.35152951e-01 6.32577240e-01 -8.42663944e-01 1.16647124e+00
-1.60940975e-01 8.44800234e-01 8.70230973e-01 -1.16546646e-01
8.60260963e-01 8.85098815e-01 1.19026817e-01 -3.97678763e-01
1.66455075e-01 1.48512274e-01 -2.30627647e-03 -4.85263199e-01
5.67192793e-01 -1.89261571e-01 -2.17604861e-01 3.68498802e-01
-1.65858224e-01 -2.44943704e-02 -5.68453491e-01 2.44874924e-01
8.64047766e-01 -6.84553266e-01 1.45345211e-01 -4.66053814e-01
7.54894674e-01 -5.09204805e-01 6.81751072e-01 4.81810480e-01
-9.17749941e-01 2.56243914e-01 5.00377595e-01 2.00019136e-01
-7.51756132e-01 -9.42053437e-01 -5.06147705e-02 8.43134582e-01
-2.48066828e-01 -4.83417451e-01 -1.45078003e+00 -6.91362023e-01
1.23047143e-01 1.53748393e+00 -3.97647798e-01 -4.85512197e-01
-3.02150279e-01 -2.09319770e-01 1.45703149e+00 -1.30119979e-01
4.55617964e-01 -7.98970520e-01 -2.11399719e-01 -1.08210452e-01
-5.42869627e-01 -1.39159143e+00 -7.89086163e-01 -1.18404508e-01
-3.23573560e-01 -5.13874590e-01 -2.06173807e-01 -3.14212352e-01
1.77836955e-01 -5.04383072e-02 3.12635928e-01 -4.48582798e-01
1.23385809e-01 7.79204607e-01 -2.20622122e-01 -8.03175092e-01
-1.28523564e+00 2.96566576e-01 4.34625775e-01 4.90715027e-01
6.64770126e-01 -6.02151871e-01 4.90232073e-02 2.65381396e-01
-6.43539429e-01 -6.39681935e-01 -2.19960019e-01 2.75568932e-01
-8.96894410e-02 1.99650750e-01 5.50566554e-01 -9.13585424e-01
1.11371207e+00 -2.51484007e-01 -2.34518275e-01 1.58001617e-01
-5.20271897e-01 1.40028551e-01 5.65279722e-01 -2.71230787e-01
-1.29982531e+00 4.49011773e-02 -4.93333489e-01 -3.16071719e-01
-6.01202369e-01 -3.22631449e-01 -1.26959276e+00 -2.06676096e-01
4.83508378e-01 2.05821663e-01 3.78292263e-01 -5.92058003e-01
3.40561986e-01 1.43467963e+00 7.49865949e-01 -2.64487088e-01
5.87851822e-01 4.16307628e-01 -7.07988739e-01 -1.53691375e+00
-3.87206636e-02 -4.09537494e-01 -1.55584186e-01 -1.16126336e-01
7.39361703e-01 -6.93354011e-01 -7.92301416e-01 6.98263168e-01
-1.30445755e+00 2.23417521e-01 -3.75265956e-01 3.22362632e-01
-4.05883968e-01 6.73792660e-01 -3.27366561e-01 -1.11790490e+00
-4.99800801e-01 -1.17689967e+00 7.39123106e-01 -2.21348092e-01
-8.39896739e-01 -6.64268494e-01 -5.66377640e-02 7.59661973e-01
4.15179372e-01 -1.04477601e-02 7.96081662e-01 -1.30478489e+00
1.52849525e-01 -5.68616867e-01 4.51044053e-01 7.33871520e-01
3.49785328e-01 -1.83187395e-01 -1.72430849e+00 -3.26430410e-01
5.88489652e-01 2.82384008e-01 2.49547228e-01 -1.77381128e-01
6.14550412e-01 -9.10212815e-01 -1.37282506e-01 6.45173907e-01
6.84835672e-01 5.48608899e-01 9.95955527e-01 1.95745125e-01
3.91062915e-01 1.13877082e+00 1.09841242e-01 3.65365416e-01
-1.13563538e-01 7.72029757e-01 2.92990804e-01 3.88041347e-01
4.19152305e-02 -4.81259555e-01 6.52204752e-01 4.82658505e-01
5.32279313e-01 -3.63448828e-01 -6.13471806e-01 4.04826283e-01
-1.33377302e+00 -1.15499353e+00 1.09680541e-01 2.44813395e+00
4.66564626e-01 -7.95819983e-02 2.15697557e-01 3.44591767e-01
9.96368170e-01 3.84158432e-01 -2.50523418e-01 -1.08407915e+00
-1.59568980e-01 -2.47442141e-01 8.07601988e-01 1.00845337e+00
-1.06404901e+00 1.03083205e+00 5.87622356e+00 4.12823290e-01
-1.14747548e+00 9.88871828e-02 4.37185973e-01 -2.48740524e-01
-3.60041678e-01 -2.00139284e-01 -6.77477062e-01 6.78920090e-01
1.52000189e+00 -6.96674883e-01 6.24784589e-01 7.21428573e-01
5.60004413e-01 3.83933872e-01 -9.97228324e-01 1.00969505e+00
2.20740855e-01 -8.23560953e-01 -4.38264124e-02 4.99762982e-01
-1.48862749e-01 -1.84005514e-01 2.32417867e-01 1.14959851e-01
1.67887554e-01 -7.81807125e-01 9.13917840e-01 2.59539664e-01
7.51826644e-01 -1.14674509e+00 5.88435233e-01 1.79203644e-01
-6.08186722e-01 2.61723865e-02 4.81381975e-02 2.80746192e-01
2.10689142e-01 1.06689088e-01 -1.18281031e+00 1.35233179e-01
7.11714327e-01 -7.25500099e-03 -3.18029970e-01 2.44041353e-01
-1.49240971e-01 7.39049256e-01 -4.28619117e-01 6.44628927e-02
-8.43151510e-02 -3.45799811e-02 1.24200261e+00 1.11900890e+00
6.41173273e-02 -1.30881453e-02 -5.17354548e-01 7.12867379e-01
-2.90755332e-01 2.67667949e-01 -9.07294214e-01 -4.56922621e-01
8.26057494e-01 6.06318414e-01 3.38327475e-02 -2.70960629e-02
-3.74361910e-02 1.38233030e+00 -1.57214403e-01 3.72620225e-01
-4.89172399e-01 -5.94959915e-01 1.45624089e+00 1.94940120e-01
-1.92239985e-01 2.38510042e-01 -2.67797023e-01 -7.88718045e-01
3.04343164e-01 -1.39033103e+00 2.12938070e-01 -3.81241143e-01
-8.27397764e-01 7.16353714e-01 -2.64599621e-01 -7.17615247e-01
-4.38243479e-01 -1.66307107e-01 -3.41500908e-01 1.18330920e+00
-6.94048285e-01 -8.21671546e-01 2.13844612e-01 6.24349892e-01
8.45761150e-02 -6.12884879e-01 1.31964624e+00 2.58120835e-01
-5.45115113e-01 1.12296033e+00 1.80637777e-01 4.04098332e-01
6.86327338e-01 -9.42317009e-01 7.86235988e-01 1.01093161e+00
-9.19321775e-02 8.63984287e-01 1.17463827e+00 -6.80148184e-01
-1.20255172e+00 -1.03091943e+00 1.41931558e+00 -5.05925417e-01
4.25128490e-01 -8.28706682e-01 -1.05499542e+00 8.86474609e-01
2.63378739e-01 -3.97466183e-01 9.57426012e-01 -9.73880477e-03
-6.62616491e-01 -1.93292603e-01 -1.84366167e+00 5.76006114e-01
5.83635211e-01 -1.28559875e+00 -6.10175848e-01 -1.83982719e-02
1.03350389e+00 4.38876487e-02 -6.35355890e-01 -3.11539173e-01
4.74542618e-01 -9.99460578e-01 7.79742122e-01 -7.35853732e-01
-6.48088336e-01 -1.80877820e-01 -5.28756440e-01 -1.09444606e+00
2.68785924e-01 -1.30580151e+00 1.92359358e-01 2.01842284e+00
3.83402497e-01 -1.20795035e+00 9.20969725e-01 1.49782908e+00
1.66601256e-01 4.64563102e-01 -1.37549281e+00 -9.70747232e-01
7.22131319e-03 -4.86854553e-01 8.67995679e-01 1.15246308e+00
2.16362923e-01 1.83009818e-01 -4.28048640e-01 6.91118002e-01
6.02899611e-01 -7.58482516e-01 8.83391142e-01 -9.21852350e-01
1.58970077e-02 -1.93029642e-01 -7.21301138e-01 -3.19095999e-01
6.32528186e-01 -8.97125661e-01 -1.77965164e-01 -6.01696372e-01
-5.04114628e-01 1.49626747e-01 1.65364459e-01 2.85590500e-01
3.28178048e-01 -6.21126413e-01 4.23302621e-01 -2.07244642e-02
2.05569610e-01 6.05397880e-01 9.49881524e-02 -2.02244729e-01
-4.66489851e-01 4.47125942e-01 -8.38412464e-01 5.32402813e-01
1.11420703e+00 -7.17683136e-01 -5.98966897e-01 3.17227468e-02
-5.39113998e-01 -7.06596151e-02 2.38193989e-01 -1.10501444e+00
2.61325717e-01 2.44538367e-01 -4.13572878e-01 -1.04158878e-01
6.25081956e-01 -1.18781722e+00 4.10616398e-01 4.40673769e-01
-5.37650824e-01 -2.65416145e-01 5.84479034e-01 5.32651961e-01
-2.38063022e-01 -2.26825371e-01 8.32878470e-01 1.11666329e-01
-4.08818364e-01 -1.32928997e-01 -9.26904321e-01 8.62920433e-02
9.35602129e-01 -2.13179350e-01 -3.66107881e-01 -7.49644101e-01
-9.11302388e-01 -2.36622300e-02 6.96431518e-01 6.94841623e-01
3.22261930e-01 -8.59762967e-01 -4.84402716e-01 4.23974633e-01
1.72150701e-01 -9.26412582e-01 1.47351906e-01 2.15662792e-01
-3.05259079e-01 4.62208927e-01 -5.84465712e-02 -3.63808163e-02
-2.14019680e+00 6.08897269e-01 7.80206800e-01 3.55290115e-01
-5.92502654e-01 5.94286740e-01 -1.01252005e-01 -7.00309396e-01
6.04357958e-01 1.22669376e-02 1.56698674e-01 1.23757318e-01
8.61181617e-01 6.02518976e-01 2.59904444e-01 -1.09469962e+00
-6.96413636e-01 -4.35254723e-01 -2.97781676e-01 -7.42181599e-01
9.49544311e-01 -4.38403070e-01 -1.31123409e-01 1.90941975e-01
1.71721280e+00 5.08065403e-01 -5.88692307e-01 1.13525696e-01
2.14551136e-01 -5.45206606e-01 5.95218353e-02 -8.29282463e-01
-7.40154982e-01 7.17623115e-01 9.35298920e-01 3.64248544e-01
6.91068828e-01 -3.84804547e-01 5.55454493e-01 3.45723242e-01
3.78576666e-01 -1.29434633e+00 -1.00715435e+00 2.53037572e-01
1.01463640e+00 -9.18227255e-01 -3.86550874e-01 -6.65478349e-01
-9.54290390e-01 6.28578842e-01 1.70556098e-01 7.79583514e-01
6.91320717e-01 3.50553125e-01 6.78111374e-01 1.19964056e-01
-3.42752844e-01 4.01962548e-01 -1.79198205e-01 1.06194746e+00
2.20320709e-02 4.20343518e-01 -7.79274181e-02 8.36029291e-01
-8.06626022e-01 -6.03757739e-01 6.52667284e-01 7.70074487e-01
-1.75202459e-01 -1.09962356e+00 -5.27295291e-01 2.83140037e-02
-7.16818511e-01 -5.24325930e-02 -1.30463636e+00 6.15146697e-01
-2.21191570e-01 1.37584555e+00 -9.06030163e-02 -5.51092029e-01
6.83973312e-01 6.77212775e-01 -4.01649773e-01 -3.34131718e-01
-1.03554487e+00 -3.74273717e-01 8.20408225e-01 -7.13142931e-01
-3.04207038e-02 -1.18447566e+00 -1.07390547e+00 -7.95969009e-01
1.34306654e-01 5.41957080e-01 1.06678259e+00 5.80150664e-01
6.50334179e-01 1.15954228e-01 6.45711243e-01 -1.02243967e-01
-7.62346327e-01 -6.11449063e-01 -8.97685647e-01 6.27947390e-01
4.68166918e-01 3.52748558e-02 -9.26122963e-01 -1.33082971e-01] | [13.966504096984863, 5.857754230499268] |
bd8730fe-c562-4ae5-9cf6-63d62bb8ea3f | recurrent-neural-networks-with-specialized | 1706.09569 | null | http://arxiv.org/abs/1706.09569v2 | http://arxiv.org/pdf/1706.09569v2.pdf | Recurrent neural networks with specialized word embeddings for health-domain named-entity recognition | Background. Previous state-of-the-art systems on Drug Name Recognition (DNR)
and Clinical Concept Extraction (CCE) have focused on a combination of text
"feature engineering" and conventional machine learning algorithms such as
conditional random fields and support vector machines. However, developing good
features is inherently heavily time-consuming. Conversely, more modern machine
learning approaches such as recurrent neural networks (RNNs) have proved
capable of automatically learning effective features from either random
assignments or automated word "embeddings". Objectives. (i) To create a highly
accurate DNR and CCE system that avoids conventional, time-consuming feature
engineering. (ii) To create richer, more specialized word embeddings by using
health domain datasets such as MIMIC-III. (iii) To evaluate our systems over
three contemporary datasets. Methods. Two deep learning methods, namely the
Bidirectional LSTM and the Bidirectional LSTM-CRF, are evaluated. A CRF model
is set as the baseline to compare the deep learning systems to a traditional
machine learning approach. The same features are used for all the models.
Results. We have obtained the best results with the Bidirectional LSTM-CRF
model, which has outperformed all previously proposed systems. The specialized
embeddings have helped to cover unusual words in DDI-DrugBank and DDI-MedLine,
but not in the 2010 i2b2/VA IRB Revision dataset. Conclusion. We present a
state-of-the-art system for DNR and CCE. Automated word embeddings has allowed
us to avoid costly feature engineering and achieve higher accuracy.
Nevertheless, the embeddings need to be retrained over datasets that are
adequate for the domain, in order to adequately cover the domain-specific
vocabulary. | ['Massimo Piccardi', 'Inigo Jauregi Unanue', 'Ehsan Zare Borzeshi'] | 2017-06-29 | null | null | null | null | ['clinical-concept-extraction'] | ['medical'] | [ 1.93469360e-01 6.20231628e-02 -5.63296378e-01 -3.61575782e-01
-8.46841693e-01 -1.73149839e-01 6.71626329e-01 5.97929239e-01
-9.75513220e-01 1.04221511e+00 3.51323783e-01 -6.93161309e-01
-2.21552715e-01 -6.91143453e-01 -3.40653658e-01 -5.51001906e-01
-3.54252383e-02 7.61678815e-01 -2.92401254e-01 -9.07125399e-02
1.09463096e-01 6.07265294e-01 -1.18023980e+00 2.11350679e-01
7.64821172e-01 7.60132909e-01 3.83239269e-01 4.41348314e-01
-4.48046029e-01 7.15990901e-01 -5.57007730e-01 -1.52823210e-01
-2.75520951e-01 -1.05779447e-01 -1.04276860e+00 -5.71046352e-01
-2.07548607e-02 -4.82289121e-02 -3.12560052e-01 7.89846897e-01
8.35880458e-01 -1.26752164e-02 9.01487947e-01 -5.95682681e-01
-1.02937567e+00 6.60270274e-01 -8.96395892e-02 6.15214519e-02
4.73004997e-01 -1.80073321e-01 8.03659678e-01 -9.21804726e-01
1.01538670e+00 9.15083766e-01 8.14825177e-01 9.11305070e-01
-1.04057837e+00 -6.77517354e-01 -1.98778778e-01 2.68265396e-01
-1.54814875e+00 -1.32436797e-01 2.85622627e-01 -5.81428528e-01
1.76220620e+00 -2.46365592e-02 5.56712687e-01 1.52763546e+00
6.86459839e-01 3.31188768e-01 8.63128603e-01 -6.25035584e-01
1.66931137e-01 5.09053946e-01 3.31171423e-01 4.76953298e-01
3.36987197e-01 3.58415097e-01 -2.99973071e-01 -4.61949408e-01
5.51585436e-01 3.26596469e-01 -3.09149981e-01 9.77773443e-02
-1.07872629e+00 1.31175148e+00 1.00075632e-01 9.53690827e-01
-6.38474107e-01 -4.94783185e-02 6.07516885e-01 7.88378641e-02
5.55745482e-01 6.17520332e-01 -9.34203029e-01 -2.01120377e-01
-9.03358817e-01 -6.05044700e-02 8.89208972e-01 9.36148524e-01
3.17117661e-01 -1.89237893e-02 -2.19473690e-01 1.06874216e+00
4.36406016e-01 3.59884173e-01 1.04634488e+00 -6.81596249e-02
2.31033683e-01 4.50769722e-01 -1.89733639e-01 -9.18692827e-01
-6.52694464e-01 -4.53067988e-01 -9.58544910e-01 -2.87018657e-01
-1.75763387e-02 -2.38220349e-01 -1.54235923e+00 1.54105771e+00
-7.34065101e-02 -9.84977037e-02 4.05446708e-01 3.94066215e-01
1.23491800e+00 5.68777025e-01 5.12535274e-01 -3.04892398e-02
1.73986280e+00 -5.21687269e-01 -1.22087967e+00 1.27016738e-01
1.25783658e+00 -9.21114564e-01 5.15734673e-01 3.00855070e-01
-5.66231906e-01 -3.42292696e-01 -9.22023714e-01 -2.04871997e-01
-1.07120919e+00 1.59860775e-01 6.79147661e-01 7.61433363e-01
-9.53402519e-01 7.40838408e-01 -7.79305279e-01 -5.46699047e-01
6.09485030e-01 5.74629784e-01 -7.53706098e-01 -1.78981826e-01
-1.79297090e+00 1.52144074e+00 4.38686371e-01 -3.46090570e-02
-7.87022591e-01 -7.82391906e-01 -1.02393842e+00 -8.93671811e-02
-2.07465112e-01 -8.18393290e-01 9.06765521e-01 -5.54363966e-01
-1.46414530e+00 7.82816589e-01 -1.69748127e-01 -6.02656186e-01
3.23654264e-02 -2.82859921e-01 -7.26984918e-01 3.14638540e-02
-1.68017205e-02 7.09230185e-01 1.22165479e-01 -6.78253412e-01
-3.67884934e-01 -2.29633093e-01 -5.50767481e-01 -2.03320965e-01
-5.21791697e-01 2.01831728e-01 9.13335681e-02 -7.50911951e-01
-4.16441411e-01 -8.30757678e-01 -4.71072882e-01 -4.87666100e-01
-5.16019225e-01 -6.40520513e-01 3.62746119e-01 -8.12497675e-01
1.52153933e+00 -1.86061561e+00 -2.00768471e-01 1.17227957e-01
2.39924148e-01 6.49607182e-01 -7.15147927e-02 5.27863860e-01
-5.28365493e-01 4.77392167e-01 -1.01577461e-01 -3.40578824e-01
-3.36144120e-01 4.52462137e-01 -1.57173201e-01 4.30196941e-01
3.11699837e-01 8.39969695e-01 -8.15019131e-01 -5.78555405e-01
2.07130030e-01 8.81717622e-01 -4.59238797e-01 1.58119664e-01
-1.26854405e-01 1.12695746e-01 -4.42254752e-01 5.32919884e-01
4.50208127e-01 -2.96400160e-01 3.14922363e-01 -1.37778565e-01
-1.41379997e-01 4.89471257e-01 -8.18948507e-01 1.80777407e+00
-5.63438177e-01 4.96029794e-01 -7.66604245e-01 -9.85015512e-01
1.08327663e+00 8.75309229e-01 5.76455474e-01 -6.58589184e-01
2.03149855e-01 5.32247126e-01 2.50271745e-02 -6.15347922e-01
4.92984116e-01 -3.26474518e-01 6.56137392e-02 -1.10268369e-01
4.18839127e-01 2.49756441e-01 -2.37350658e-01 2.25862041e-02
1.09946585e+00 1.25205651e-01 6.62223995e-01 -3.04470301e-01
4.38321143e-01 2.10763082e-01 7.15279877e-01 6.36991084e-01
5.13431318e-02 4.52391297e-01 1.11137658e-01 -6.84021592e-01
-8.69413674e-01 -7.42922246e-01 -8.95249844e-01 5.74134588e-01
-5.66481829e-01 -4.12783027e-01 -4.89653111e-01 -6.16437197e-01
-6.70223236e-02 8.26433301e-01 -8.12277794e-01 -4.32707742e-02
-4.31580037e-01 -9.15659428e-01 9.06038582e-01 4.03743654e-01
-4.52306531e-02 -1.17052865e+00 -2.25560918e-01 7.16655731e-01
8.53384510e-02 -1.01976442e+00 -3.15732569e-01 6.87561929e-01
-8.69452477e-01 -8.98076594e-01 -9.56996679e-01 -9.75999534e-01
4.24307495e-01 -5.24247527e-01 1.14818680e+00 -1.91071227e-01
-6.41482890e-01 -3.55449654e-02 -5.34678578e-01 -5.54648042e-01
-3.71955335e-01 2.94173360e-01 1.81198537e-01 -5.72011352e-01
1.15709352e+00 -3.14782202e-01 -4.22795713e-01 7.49291806e-03
-1.06860375e+00 -4.11579132e-01 8.15179348e-01 1.20231354e+00
6.24199986e-01 -4.64754939e-01 9.73720789e-01 -1.27709591e+00
7.55672514e-01 -6.29611492e-01 -1.97941110e-01 2.83094049e-01
-9.63470697e-01 3.47573668e-01 6.58399463e-01 -5.87873876e-01
-7.21097529e-01 -3.73680480e-02 -7.85909891e-01 -1.12302534e-01
-3.18603486e-01 7.75956273e-01 1.83316380e-01 2.61928111e-01
8.04293990e-01 1.28033087e-01 -1.74936846e-01 -7.94982731e-01
3.91453892e-01 1.21796906e+00 5.77301718e-02 -2.47511685e-01
1.74696818e-01 1.60592981e-02 -3.60524833e-01 -9.07915890e-01
-5.46981871e-01 -6.51147902e-01 -4.02714401e-01 5.97086966e-01
1.21291578e+00 -8.53247464e-01 -5.11282980e-01 2.64158398e-01
-1.46692336e+00 1.15350887e-01 -3.31023633e-01 1.06678295e+00
-2.94956237e-01 1.98797181e-01 -7.55471349e-01 -4.89634752e-01
-7.08981693e-01 -1.23753250e+00 8.35838854e-01 1.82956114e-01
-5.86190581e-01 -1.37230027e+00 6.41567111e-01 -1.39605030e-01
6.08146250e-01 2.84950256e-01 1.16138983e+00 -1.38326240e+00
2.01947540e-01 -2.26110503e-01 -1.37612671e-01 3.85859638e-01
4.65671718e-01 -4.77451347e-02 -1.03300953e+00 -8.25448930e-02
-1.38720751e-01 -1.30678371e-01 9.16915894e-01 5.18908024e-01
9.90823686e-01 -2.08291739e-01 -7.78644204e-01 5.27521372e-01
1.57066333e+00 7.05909967e-01 9.20218349e-01 3.99462759e-01
5.48948228e-01 2.88059801e-01 2.98577994e-01 2.90108383e-01
2.52897441e-01 6.12322271e-01 -1.63754150e-01 -4.54235107e-01
-5.09428196e-02 -1.99200526e-01 1.29044428e-01 9.67595398e-01
1.06133893e-01 -1.72216222e-01 -1.01714504e+00 8.52878332e-01
-1.43384945e+00 -6.17724299e-01 5.03434911e-02 1.94887841e+00
1.26830828e+00 -4.57627187e-03 -2.25836992e-01 -1.93092570e-01
5.70745707e-01 -1.50953963e-01 -1.53790027e-01 -9.79193389e-01
3.47248875e-02 9.62425590e-01 6.45733118e-01 2.30733454e-01
-9.76413190e-01 9.15019035e-01 5.88592482e+00 9.63329494e-01
-1.22263277e+00 2.57162213e-01 4.74689931e-01 1.90537691e-01
-4.34804201e-01 -2.62419581e-01 -1.20835996e+00 4.72753882e-01
1.49772787e+00 1.20479465e-01 -1.60125971e-01 8.15164983e-01
1.52454851e-02 2.35245124e-01 -1.24580300e+00 9.46755707e-01
7.66957030e-02 -1.70220912e+00 1.83930680e-01 2.93967068e-01
3.98071826e-01 4.10155267e-01 1.83974784e-02 4.66879874e-01
2.99883515e-01 -1.61353803e+00 -1.85893387e-01 6.83082938e-01
1.27509308e+00 -6.02889121e-01 1.36603570e+00 -1.20404311e-01
-8.50465715e-01 1.68903470e-01 -4.70273942e-01 4.97798204e-01
1.39490798e-01 9.55609441e-01 -1.17369378e+00 9.13592577e-01
4.61113721e-01 8.96753788e-01 -2.36116916e-01 8.65625501e-01
1.10848546e-02 3.23934466e-01 -2.33239114e-01 -4.38863754e-01
4.09643143e-01 1.77056298e-01 2.09447920e-01 1.63064826e+00
2.79900670e-01 -1.00560211e-01 -1.80241615e-02 6.72892928e-01
-2.09354937e-01 5.05855083e-01 -8.23110461e-01 -4.47477221e-01
3.76045853e-01 1.05339754e+00 -2.97203213e-01 -3.96432251e-01
-4.51093465e-01 6.05346203e-01 9.72562954e-02 1.74059764e-01
-6.82072759e-01 -8.44383597e-01 7.40156949e-01 3.14185955e-02
3.03159684e-01 2.39805542e-02 -1.01059385e-01 -9.81640637e-01
-3.51899564e-01 -8.69643390e-01 5.50391197e-01 -1.64648816e-01
-1.54454923e+00 1.22553730e+00 -1.57463223e-01 -1.25303578e+00
-4.58402842e-01 -8.58923793e-01 -1.03901610e-01 1.14556694e+00
-1.74447620e+00 -9.28813696e-01 3.82297903e-01 4.41986620e-01
2.89143711e-01 -4.48766053e-01 1.66758764e+00 8.44324172e-01
-4.95121807e-01 6.48867726e-01 3.78495067e-01 4.04102147e-01
8.94784331e-01 -1.07938397e+00 2.12177008e-01 1.49076268e-01
6.08794205e-02 1.10217774e+00 3.34767729e-01 -7.13649988e-01
-9.36535835e-01 -1.06254697e+00 1.53126776e+00 -3.45689118e-01
5.79396427e-01 -2.85537004e-01 -8.78915846e-01 5.56139290e-01
2.15626568e-01 1.09162241e-01 1.28437304e+00 2.66606212e-01
-2.64177501e-01 2.03233168e-01 -1.29608548e+00 1.60748348e-01
6.12045646e-01 -6.82124853e-01 -9.88025665e-01 5.36899149e-01
8.56718004e-01 -1.09657966e-01 -1.44670188e+00 5.03135264e-01
5.43834627e-01 -8.69052783e-02 9.46378827e-01 -1.18574858e+00
5.02570689e-01 -9.35035720e-02 -1.72597885e-01 -1.31272781e+00
-2.31814340e-01 -3.47621828e-01 2.43318960e-01 1.07002807e+00
9.04656172e-01 -1.07303143e+00 3.59927356e-01 4.37092006e-01
-8.31357986e-02 -1.21627629e+00 -1.21277511e+00 -7.40709662e-01
4.55687493e-01 -3.23098928e-01 4.04542625e-01 1.26487994e+00
1.81436658e-01 4.18291748e-01 -2.47233152e-01 -1.94338039e-01
5.49333058e-02 -2.65580237e-01 1.37436792e-01 -1.17217493e+00
-1.36431590e-01 -2.55351484e-01 -6.59155488e-01 -7.66050637e-01
1.52491376e-01 -1.09038591e+00 -1.77357718e-01 -1.76365411e+00
3.10900301e-01 -5.87293148e-01 -5.48012197e-01 7.07456112e-01
9.10529718e-02 -1.24690279e-01 -4.05273408e-01 -4.68045957e-02
-7.81782568e-02 4.02417034e-01 8.60238910e-01 -1.74323276e-01
-3.35865766e-01 -2.55914629e-01 -7.10641503e-01 4.77584571e-01
6.75209701e-01 -1.06156552e+00 1.30559467e-02 -1.70060292e-01
3.69083397e-02 3.69353034e-02 -1.70618519e-01 -4.22308803e-01
2.77565598e-01 1.52550846e-01 5.23139775e-01 -5.09892106e-01
3.11921500e-02 -6.95965588e-01 8.03952590e-02 7.19736874e-01
-3.02643210e-01 1.84652075e-01 3.42694134e-01 4.74915028e-01
-3.70178550e-01 -4.23678547e-01 7.57172763e-01 -1.11605123e-01
-3.36894691e-01 2.57084191e-01 -6.05353773e-01 5.05206734e-02
7.68156707e-01 -2.83497013e-02 -3.43854606e-01 1.15007930e-01
-9.58912075e-01 -2.77181994e-02 -8.17504674e-02 5.21943569e-01
7.40791917e-01 -1.21506703e+00 -7.32196450e-01 8.84691626e-02
2.52063453e-01 -1.76533237e-01 2.00495765e-01 6.76109850e-01
-5.79815686e-01 1.03897595e+00 -8.35940465e-02 -4.54045236e-01
-1.10929620e+00 6.91061139e-01 4.02154356e-01 -8.87868285e-01
-6.63099706e-01 6.99781179e-01 -1.47305414e-01 -5.92444122e-01
2.80562520e-01 -5.40096760e-01 -8.43601644e-01 2.97101974e-01
5.28970659e-01 -7.98247531e-02 4.04757649e-01 -3.72506261e-01
-7.60649323e-01 4.54404652e-01 -5.77851653e-01 2.99773037e-01
1.75245368e+00 4.35800105e-01 -8.51577520e-03 1.22776322e-01
1.66619253e+00 -2.45160922e-01 -2.57863775e-02 -2.18646139e-01
4.38911706e-01 -3.79320346e-02 2.54893869e-01 -1.00139463e+00
-8.29420686e-01 8.90523851e-01 1.04210174e+00 -9.29816514e-02
8.19419682e-01 -2.10435867e-01 1.09370625e+00 2.99911410e-01
2.56126702e-01 -1.12262630e+00 -4.01428223e-01 5.05245507e-01
5.14729083e-01 -1.00752413e+00 4.69576046e-02 1.54210040e-02
-4.97785270e-01 1.43081081e+00 -5.34013547e-02 1.58843800e-01
9.62574363e-01 4.15749252e-01 1.03079684e-01 -4.58746940e-01
-5.38641572e-01 -7.62361959e-02 3.34530205e-01 4.69868541e-01
8.47152472e-01 1.03729650e-01 -5.80131769e-01 7.65010893e-01
3.26415263e-02 4.55605268e-01 4.43157554e-01 7.42023647e-01
-1.86219290e-02 -1.64324415e+00 -1.88150108e-02 6.87133491e-01
-1.00571144e+00 -5.22903860e-01 -8.71448964e-02 9.41059470e-01
1.43237397e-01 8.09265554e-01 -1.79754242e-01 -3.37940842e-01
3.85167509e-01 3.32181364e-01 2.07994282e-01 -9.52132940e-01
-8.38114917e-01 -2.00204447e-01 3.54817450e-01 -3.12512428e-01
-5.05736172e-01 -2.79277384e-01 -1.28754926e+00 5.22661163e-03
-6.62866652e-01 3.25602770e-01 8.93650591e-01 1.02729928e+00
5.87342381e-01 7.44679987e-01 3.02284420e-01 -2.10750625e-01
-6.00044191e-01 -1.21640956e+00 -4.73193496e-01 2.49537572e-01
2.47761220e-01 -7.35485017e-01 -2.60067970e-01 -1.69438809e-01] | [8.444485664367676, 8.701774597167969] |
2e08833b-9929-47f8-a1fe-8ab4faa5de20 | idiap-tiet-lt-edi-acl2022-hope-speech | null | null | https://aclanthology.org/2022.ltedi-1.49 | https://aclanthology.org/2022.ltedi-1.49.pdf | IDIAP_TIET@LT-EDI-ACL2022 : Hope Speech Detection in Social Media using Contextualized BERT with Attention Mechanism | With the increase of users on social media platforms, manipulating or provoking masses of people has become a piece of cake. This spread of hatred among people, which has become a loophole for freedom of speech, must be minimized. Hence, it is essential to have a system that automatically classifies the hatred content, especially on social media, to take it down. This paper presents a simple modular pipeline classifier with BERT embeddings and attention mechanism to classify hope speech content in the Hope Speech Detection shared task for Equality, Diversity, and Inclusion-ACL 2022. Our system submission ranks fourth with an F1-score of 0.84. We release our code-base here https://github.com/Deepanshu-beep/hope-speech-attention . | ['Petr Motlicek', 'Muskaan Singh', 'Deepanshu Khanna'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-6.01519287e-01 2.91790932e-01 -9.41785052e-02 -5.56004196e-02
-6.87200427e-01 -5.72162211e-01 7.34144270e-01 2.59021848e-01
-4.98181909e-01 5.46938360e-01 8.45073104e-01 -3.53928864e-01
4.44725931e-01 -3.85901123e-01 -3.04212626e-02 -3.72958601e-01
2.86601990e-01 4.70660487e-03 5.28314672e-02 -5.39675593e-01
2.58931369e-01 1.32104024e-01 -9.78053153e-01 2.46139660e-01
7.36535132e-01 3.90311331e-01 -2.53666252e-01 9.15262580e-01
1.11004241e-01 1.12318528e+00 -9.06350255e-01 -1.06805325e+00
-6.56344071e-02 -4.65057105e-01 -8.95102382e-01 -4.18735981e-01
3.88235062e-01 -5.62913954e-01 -7.89700091e-01 1.37459528e+00
8.27184677e-01 3.83748449e-02 4.03632849e-01 -1.35916078e+00
-1.27097309e+00 8.78716886e-01 -4.25729007e-01 5.94511926e-01
2.59547234e-01 2.79740006e-01 9.35941160e-01 -8.27933311e-01
5.48111558e-01 1.18161726e+00 4.55250561e-01 8.87130380e-01
-9.28803623e-01 -8.42725039e-01 -5.37281632e-01 2.05446258e-01
-1.39342892e+00 -7.84643054e-01 7.11537302e-01 -7.29678035e-01
9.34391439e-01 6.36378646e-01 8.16797256e-01 1.61608171e+00
1.84215829e-01 7.88277864e-01 6.60301149e-01 -1.18851177e-01
-1.37890756e-01 5.20144343e-01 1.94158658e-01 6.62631273e-01
5.76239862e-02 -4.16066498e-01 -5.76060474e-01 -2.42434025e-01
4.90177190e-03 -6.63325936e-02 -2.93863356e-01 4.83145505e-01
-8.19733322e-01 1.14896119e+00 3.35457981e-01 4.87503648e-01
8.57775211e-02 -1.22688167e-01 6.79191947e-01 1.42760694e-01
8.67574871e-01 5.81159234e-01 -2.18096245e-02 -7.13102221e-01
-8.83525252e-01 1.97303444e-01 7.29458153e-01 4.10523444e-01
1.72514483e-01 -2.54872829e-01 -2.21835390e-01 1.08450019e+00
2.07514331e-01 5.57478428e-01 5.80898702e-01 -6.69779480e-01
2.43629992e-01 5.45864105e-01 -1.60665452e-01 -1.36401713e+00
-4.03117687e-01 -4.38637167e-01 -9.27980781e-01 -1.33984849e-01
3.50594640e-01 -4.75929588e-01 -3.93452376e-01 1.66590071e+00
3.12774450e-01 -2.66630173e-01 -2.74200499e-01 9.06495810e-01
9.34538305e-01 8.10426891e-01 1.13849618e-01 -8.47187862e-02
1.50479543e+00 -9.69409227e-01 -1.03721774e+00 -4.21149395e-02
8.19915414e-01 -8.93781483e-01 1.03620100e+00 2.45664120e-01
-7.33034492e-01 -9.34491456e-02 -9.81714725e-01 -7.17516363e-01
-6.82165861e-01 -2.42035702e-01 3.37184519e-01 1.02089381e+00
-7.23019004e-01 5.80043674e-01 -4.03542727e-01 -5.22757590e-01
7.22880661e-01 -1.01986587e-01 -4.11519229e-01 4.05347526e-01
-1.49726939e+00 9.42971468e-01 -5.93038555e-03 -1.77926347e-02
-6.02941155e-01 -4.70407307e-01 -5.82123339e-01 -1.55013353e-01
-3.61679830e-02 -7.42280707e-02 1.19244504e+00 -9.23369884e-01
-1.12588322e+00 1.06707048e+00 2.79893696e-01 -4.00884241e-01
5.73246658e-01 -7.09969759e-01 -5.77720702e-01 -1.65711567e-01
1.02696978e-01 5.27341247e-01 8.40468526e-01 -5.85545421e-01
-1.44896388e-01 -2.68725097e-01 -5.93044125e-02 -1.79505914e-01
-1.01699257e+00 6.80142403e-01 1.99050531e-01 -5.02519786e-01
-6.57945991e-01 -1.05385268e+00 3.94111544e-01 2.66130809e-02
-8.62255514e-01 -4.60551172e-01 1.14308083e+00 -1.24585068e+00
1.74020183e+00 -2.57293844e+00 1.22357301e-01 -4.08981264e-01
7.45411456e-01 5.37713706e-01 1.97994441e-01 6.45491123e-01
7.42797032e-02 8.60414207e-01 4.70023789e-02 -5.91892719e-01
8.44348744e-02 -3.40959698e-01 -2.04758674e-01 7.77181447e-01
1.88619554e-01 5.59737504e-01 -1.02550554e+00 -3.01591218e-01
-7.74153247e-02 8.47938597e-01 -6.47432029e-01 2.73130417e-01
1.32351622e-01 2.49340221e-01 -2.87574172e-01 3.75994503e-01
6.00400686e-01 -1.85539603e-01 -1.71882406e-01 3.08599174e-01
-4.28604275e-01 3.64833146e-01 -3.64990592e-01 1.37777948e+00
-1.56856608e-02 1.22998905e+00 1.83162928e-01 -1.07658796e-01
8.22945058e-01 3.07901829e-01 1.89851090e-01 -3.75348061e-01
6.47930861e-01 5.15603423e-02 2.04473421e-01 -7.08638549e-01
6.93068504e-01 5.71684241e-02 -3.62877220e-01 2.44714931e-01
-5.24848737e-02 1.22259483e-01 -1.99827909e-01 6.66725636e-01
1.22607398e+00 -4.65232521e-01 2.60149390e-01 -1.95292234e-01
4.46674198e-01 -1.71960428e-01 4.48274612e-01 2.47133449e-01
-9.65031266e-01 6.11590028e-01 8.59066844e-01 -3.46229732e-01
-1.21932817e+00 -6.86979711e-01 -8.13264549e-02 1.30606234e+00
-3.19943756e-01 -8.02855015e-01 -8.35831881e-01 -6.27991498e-01
-4.93859239e-02 7.88878918e-01 -6.00055456e-01 -3.27837527e-01
-2.19860569e-01 -5.25458515e-01 9.14781809e-01 -2.00875640e-01
4.96301383e-01 -9.39077258e-01 -4.01590407e-01 -2.03377947e-01
-3.57652962e-01 -8.70413661e-01 -6.88320577e-01 -2.68695533e-01
1.70516148e-01 -9.04015660e-01 -8.28470826e-01 -4.59697515e-01
9.04244557e-02 2.45512843e-01 4.97326165e-01 4.08528239e-01
-2.39660487e-01 -1.28997475e-01 -5.61342835e-01 -4.12830144e-01
-5.67964733e-01 3.63726258e-01 1.85967144e-02 1.16347179e-01
4.87463653e-01 -4.14671451e-01 -5.90274572e-01 -1.51946887e-01
-5.06427646e-01 2.90346682e-01 -6.15565628e-02 5.54468513e-01
-3.48533392e-01 -1.98736489e-01 4.99714285e-01 -6.02129817e-01
9.04474258e-01 -9.47679400e-01 2.43178103e-02 -3.41444433e-01
8.20097327e-02 -4.82339978e-01 5.39535701e-01 -2.67294049e-01
-5.77111840e-01 -2.19343543e-01 -4.62899238e-01 -2.85197228e-01
-1.14976846e-01 -4.63997386e-03 -1.16424588e-02 3.06367218e-01
6.92382514e-01 -4.61343005e-02 9.47951674e-02 -6.41241431e-01
5.00283480e-01 1.53326714e+00 1.14643522e-01 1.32239833e-01
6.99121296e-01 1.64456144e-01 -7.28744507e-01 -1.55667615e+00
-8.96430433e-01 -4.84389424e-01 -2.02338517e-01 -4.18521434e-01
1.34187353e+00 -9.30939674e-01 -7.26746976e-01 6.79857969e-01
-1.56196356e+00 -1.10896021e-01 1.86776832e-01 1.84080854e-01
1.20288402e-01 4.46935415e-01 -8.79842818e-01 -1.07864869e+00
-6.07240558e-01 -8.18458676e-01 6.49186790e-01 2.61914968e-01
-7.21785307e-01 -5.53082645e-01 4.89731818e-01 9.05469477e-01
6.16029859e-01 2.59438366e-01 3.54809970e-01 -1.25634813e+00
1.23466421e-02 -1.73117206e-01 -1.92928031e-01 5.33490777e-01
8.01704824e-02 3.33595634e-01 -1.31472456e+00 -1.72868371e-01
-1.89114898e-01 -5.34376740e-01 7.11919010e-01 -2.10648850e-01
1.06463683e+00 -8.96187067e-01 -1.11944517e-02 3.07423830e-01
7.54804015e-01 -2.35648409e-01 7.07128823e-01 2.55810887e-01
7.74897873e-01 4.58863795e-01 -1.31285474e-01 7.33106852e-01
3.53714287e-01 4.93954569e-01 3.15565467e-01 3.55998367e-01
-1.85554907e-01 -4.38208789e-01 6.84138298e-01 1.27156472e+00
9.71878916e-02 -3.54611546e-01 -1.07304478e+00 5.72636604e-01
-1.59577882e+00 -1.39429533e+00 -3.10085922e-01 1.77308178e+00
8.63072693e-01 -1.00895250e-02 5.86048901e-01 2.29073446e-02
9.57908213e-01 5.61584115e-01 -1.43772317e-02 -8.65631282e-01
2.42548026e-02 -2.68788189e-01 6.64011687e-02 7.00759709e-01
-1.32548630e+00 8.72986138e-01 4.84259653e+00 7.75462031e-01
-1.21629024e+00 7.63636291e-01 7.77649820e-01 -3.14453691e-01
-8.90618488e-02 -3.54010403e-01 -7.08586454e-01 9.98959780e-01
1.34869742e+00 -2.27315262e-01 4.25516903e-01 7.65435219e-01
3.93119663e-01 3.13556373e-01 -5.61236680e-01 9.14569795e-01
2.26045966e-01 -1.08810115e+00 -4.48422939e-01 1.89230770e-01
3.91193837e-01 3.89070094e-01 2.49199286e-01 3.76695633e-01
1.51093856e-01 -9.60364878e-01 7.27114081e-01 2.08680794e-01
5.17845690e-01 -6.37944698e-01 6.40663862e-01 4.86023933e-01
-4.19290930e-01 -6.95513412e-02 -1.49812385e-01 -1.51721224e-01
1.91702574e-01 6.90004587e-01 -6.59618795e-01 -1.38134241e-01
6.60726249e-01 6.23785496e-01 -7.23126769e-01 8.31261277e-01
-3.84370208e-01 6.34733498e-01 -1.49831280e-01 -4.91737694e-01
1.20010197e-01 3.00140619e-01 8.31503391e-01 1.57874167e+00
7.83277974e-02 -1.78408787e-01 -2.42335588e-01 7.52856493e-01
-4.00777698e-01 2.38997295e-01 -8.53496969e-01 -7.84039557e-01
5.28048635e-01 1.50372612e+00 -2.52670348e-01 -9.27467048e-02
-3.76543254e-01 1.16726732e+00 5.23368597e-01 -2.87799448e-01
-1.17096806e+00 -4.84815776e-01 9.02591348e-01 3.42312247e-01
-2.57049292e-01 -3.61727327e-01 4.98235039e-02 -1.18119097e+00
-2.57428467e-01 -9.26058710e-01 1.77470580e-01 -6.04114115e-01
-1.46018445e+00 4.99579400e-01 -5.29036701e-01 -7.58434713e-01
1.16055585e-01 -2.32221782e-01 -5.25733113e-01 5.33670962e-01
-8.33909035e-01 -1.06929553e+00 -2.78580096e-02 2.66065389e-01
6.37996554e-01 -1.83310911e-01 7.02889800e-01 5.18501043e-01
-1.05109978e+00 8.11151981e-01 -3.27082686e-02 6.56014025e-01
7.29021788e-01 -9.40930068e-01 3.97625029e-01 6.98562086e-01
-5.80295511e-02 5.34113646e-01 8.33059490e-01 -6.43399179e-01
-1.17891335e+00 -8.59835029e-01 1.36031711e+00 -7.50101566e-01
1.31736660e+00 -7.78641939e-01 -8.95904183e-01 5.06015360e-01
9.01920974e-01 -3.82833719e-01 1.03856575e+00 2.66379058e-01
-6.21067703e-01 1.98196784e-01 -1.14417851e+00 6.88068867e-01
8.72664213e-01 -7.98435688e-01 -5.80936253e-01 8.47808480e-01
9.57826674e-01 -6.39959797e-02 -6.80653274e-01 -5.20671189e-01
5.54844618e-01 -1.00008464e+00 3.49786729e-01 -7.92466342e-01
7.35736549e-01 7.60036632e-02 -3.59954014e-02 -1.23991466e+00
-5.08191228e-01 -1.19342041e+00 -2.48066429e-02 1.55006468e+00
4.13220406e-01 -5.43841422e-01 4.65305269e-01 6.93541944e-01
-1.09370135e-01 -2.94363976e-01 -1.25436425e+00 -3.39916438e-01
3.48070055e-01 -2.08748892e-01 6.84888288e-02 1.51478255e+00
5.23503125e-01 7.50684679e-01 -8.75991523e-01 -1.65688258e-03
4.12965685e-01 -7.73834765e-01 7.35493124e-01 -8.86630356e-01
-1.00090027e-01 -6.64759696e-01 -5.30476272e-01 -3.84559393e-01
1.02971248e-01 -1.09874272e+00 -3.68830502e-01 -1.20937717e+00
5.61489701e-01 7.55947903e-02 -2.50172820e-02 6.39645934e-01
1.45028904e-01 2.96114624e-01 4.28924561e-01 1.50386676e-01
-7.62007475e-01 6.27448976e-01 9.64830637e-01 -2.03078151e-01
-6.52989373e-02 -4.57480878e-01 -9.24884558e-01 5.19055843e-01
1.23461902e+00 -6.88864231e-01 5.27497679e-02 -2.60386080e-01
6.68706715e-01 -3.56161982e-01 4.64829862e-01 -9.75085318e-01
1.22112110e-01 7.87789375e-02 1.19743794e-01 -2.39943147e-01
5.40848732e-01 -4.34252828e-01 -1.41504943e-01 6.92874908e-01
-5.51179588e-01 2.79883631e-02 9.62917581e-02 1.61077872e-01
2.23653316e-01 -1.43184796e-01 8.68775427e-01 1.85287520e-01
-1.64890215e-01 8.32289308e-02 -6.72696769e-01 4.44097787e-01
7.62221098e-01 3.87521744e-01 -1.16727388e+00 -5.82597256e-01
-7.60041177e-01 7.93375373e-02 3.64486217e-01 8.49985301e-01
4.34529215e-01 -1.20610952e+00 -1.13792825e+00 9.17947479e-03
4.26973291e-02 -7.13483632e-01 4.25268382e-01 7.79941797e-01
-5.18199265e-01 1.24125034e-01 -1.84658751e-01 6.97645918e-02
-1.40184927e+00 4.93393332e-01 2.68375129e-01 2.23676249e-01
-6.02057517e-01 1.03456414e+00 -2.47672468e-01 -3.20312768e-01
-1.91688705e-02 2.65433460e-01 -6.99411556e-02 2.87253439e-01
1.07555342e+00 5.53817749e-01 -2.86043882e-01 -1.06858778e+00
-6.17311656e-01 -1.97565317e-01 -2.64753461e-01 5.19603714e-02
1.39729500e+00 9.99762416e-02 -3.39010984e-01 4.67444658e-01
1.54262233e+00 5.17903149e-01 -5.70279062e-01 4.01425928e-01
-2.64818519e-01 -6.34872139e-01 2.07048103e-01 -7.34479010e-01
-8.80087078e-01 9.52309251e-01 5.31357467e-01 8.81249905e-01
3.81892234e-01 2.24210680e-01 9.72042501e-01 1.32906968e-02
-2.66297847e-01 -1.09632432e+00 1.58431873e-01 6.25538945e-01
1.25697148e+00 -1.33316267e+00 -5.49936779e-02 -1.91639781e-01
-7.53159523e-01 7.57583737e-01 5.24264395e-01 1.08775524e-02
8.54479253e-01 1.06932901e-01 -9.12645236e-02 -1.76358506e-01
-7.37941504e-01 7.79524222e-02 1.17887687e-02 5.07434487e-01
7.97766566e-01 3.72605503e-01 -4.38344568e-01 8.24565768e-01
-4.65210557e-01 -3.52096379e-01 7.60138214e-01 5.14884412e-01
-8.32285285e-01 -5.61375082e-01 -2.15236083e-01 1.41512096e-01
-9.50541496e-01 -1.68767914e-01 -1.00934327e+00 5.69698811e-01
2.32123956e-01 1.04210579e+00 1.01913223e-02 -8.44921768e-01
5.04809991e-02 2.71107286e-01 -1.22829884e-01 -4.77897316e-01
-8.38083625e-01 -1.35166511e-01 6.18628621e-01 -2.68348753e-01
3.01246233e-02 -5.27008355e-01 -8.59024882e-01 -1.22186625e+00
-1.14365943e-01 2.13974230e-02 7.08121061e-01 4.74093199e-01
6.58741832e-01 1.24827102e-01 5.01084208e-01 -4.59440917e-01
-4.28504616e-01 -1.20844448e+00 -4.23538029e-01 4.41053331e-01
5.12382329e-01 -3.85290086e-01 -8.17192793e-01 -4.19523567e-01] | [8.764692306518555, 10.590055465698242] |
933bf8a7-6889-4e36-aa4e-1cd4d636721a | idiap-submission-lt-edi-acl2022-hope-speech | null | null | https://aclanthology.org/2022.ltedi-1.54 | https://aclanthology.org/2022.ltedi-1.54.pdf | IDIAP Submission@LT-EDI-ACL2022 : Hope Speech Detection for Equality, Diversity and Inclusion | Social media platforms have been provoking masses of people. The individual comments affect a prevalent way of thinking by moving away from preoccupation with discrimination, loneliness, or influence in building confidence, support, and good qualities. This paper aims to identify hope in these social media posts. Hope significantly impacts the well-being of people, as suggested by health professionals. It reflects the belief to achieve an objective, discovers a new path, or become motivated to formulate pathways.In this paper we classify given a social media post, hope speech or not hope speech, using ensembled voting of BERT, ERNIE 2.0 and RoBERTa for English language with 0.54 macro F1-score (2^{st} rank). For non-English languages Malayalam, Spanish and Tamil we utilized XLM RoBERTA with 0.50, 0.81, 0.3 macro F1 score (1^{st}, 1^{st},3^{rd} rank) respectively. For Kannada, we use Multilingual BERT with 0.32 F1 score(5^{th})position. We release our code-base here: https://github.com/Muskaan-Singh/Hate-Speech-detection.git. | ['Petr Motlicek', 'Muskaan Singh'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-4.87105578e-01 4.64760184e-01 -4.94783998e-01 5.66942915e-02
-5.40828824e-01 -2.55180687e-01 9.44314063e-01 6.69000864e-01
-3.12271714e-01 1.00376272e+00 9.97666836e-01 -3.97916764e-01
-1.42437488e-01 -8.24793577e-01 -1.90934435e-01 -4.80744839e-01
2.30892152e-02 2.73158215e-02 -3.16878021e-01 -5.99430978e-01
6.65587902e-01 -7.20316991e-02 -9.43290472e-01 1.64116338e-01
7.52787709e-01 2.96891898e-01 -1.87643200e-01 8.15266252e-01
-8.04289132e-02 1.20127583e+00 -5.01236916e-01 -6.62680924e-01
-3.97308975e-01 -6.47651136e-01 -7.79416382e-01 -5.15173316e-01
-1.21703498e-01 -2.20738519e-02 -1.08586825e-01 1.02072453e+00
6.31496668e-01 4.27943580e-02 7.56126225e-01 -1.01194894e+00
-8.60756218e-01 9.21028972e-01 -5.57305038e-01 2.47023731e-01
7.14756906e-01 -7.83028603e-02 6.57747030e-01 -9.32837069e-01
6.04590595e-01 1.29480243e+00 7.31043339e-01 4.22598571e-01
-8.69850874e-01 -7.67241657e-01 -4.43613648e-01 -1.45972371e-01
-1.22606146e+00 -6.84632838e-01 6.84073150e-01 -6.89063609e-01
7.25141108e-01 5.56715727e-01 9.03813481e-01 1.25375712e+00
2.96122789e-01 3.86351436e-01 1.49949789e+00 -3.05501521e-01
9.91046801e-02 4.97815251e-01 2.37877429e-01 7.47015059e-01
-1.42805010e-01 -3.69082272e-01 -6.55322909e-01 -5.00592232e-01
4.09470707e-01 1.43956989e-01 -2.47352105e-02 1.02256751e+00
-1.14824033e+00 1.22555280e+00 3.76800776e-01 5.05681515e-01
-5.85912943e-01 -2.10676670e-01 8.56456086e-02 2.90320516e-01
7.77484119e-01 8.20515826e-02 -1.54390959e-02 -6.52678311e-01
-8.52810264e-01 -1.21339366e-01 7.13270664e-01 1.29299536e-01
4.18153346e-01 -1.49373576e-01 -1.23186320e-01 1.23612177e+00
4.09666061e-01 6.87284768e-01 4.80198026e-01 -8.04359376e-01
-6.77149519e-02 5.47830284e-01 -1.39883667e-01 -1.37962973e+00
-5.40433705e-01 -5.32431066e-01 -1.02073526e+00 -8.83047357e-02
2.18891233e-01 -4.32286173e-01 -3.03837568e-01 1.70089018e+00
2.33438596e-01 -2.59054184e-01 2.95060605e-01 4.26158160e-01
1.13700330e+00 8.36175442e-01 1.88921001e-02 -6.19315326e-01
1.42493844e+00 -8.04653406e-01 -6.92512751e-01 -1.18683442e-01
4.29469407e-01 -1.28207362e+00 9.88390982e-01 4.69687641e-01
-1.08171690e+00 -2.00526521e-01 -7.01108694e-01 1.04519404e-01
-3.09509277e-01 -2.13900223e-01 3.93887758e-01 8.82532239e-01
-9.96976256e-01 5.65182984e-01 -3.11881870e-01 -8.45328808e-01
4.25969511e-01 -2.04418190e-02 -3.60239774e-01 1.84696555e-01
-1.12746358e+00 8.17486167e-01 -6.06236346e-02 -4.42774564e-01
-3.94093543e-01 -2.95613497e-01 -3.79311085e-01 -5.04316449e-01
7.40910647e-03 -4.58679736e-01 6.71810389e-01 -7.65021682e-01
-1.26404762e+00 9.19337809e-01 -1.34827957e-01 -2.00873017e-01
3.36796224e-01 -2.64010429e-01 -9.48152721e-01 3.55824977e-02
2.76357204e-01 6.09674692e-01 4.66385037e-01 -7.98933029e-01
-2.26742998e-01 -4.20235366e-01 5.71141466e-02 -1.12752967e-01
-5.95049739e-01 6.26648009e-01 4.87718105e-01 -6.64179981e-01
-6.31847233e-02 -7.18752086e-01 -3.10773626e-02 -3.75195235e-01
-7.28139579e-01 -2.92072386e-01 4.62105125e-01 -1.01344323e+00
1.52448833e+00 -2.03026438e+00 -3.65396380e-01 8.08715522e-02
4.70070362e-01 5.72284944e-02 3.56506288e-01 9.14962411e-01
2.67765611e-01 6.76207006e-01 1.55769914e-01 5.61187603e-02
-1.82739064e-01 6.74076900e-02 1.93670720e-01 5.09895444e-01
7.09720254e-02 3.92371356e-01 -1.00112224e+00 -4.02022362e-01
-1.97110981e-01 8.28793645e-01 -5.71523428e-01 -1.53740913e-01
4.64918613e-01 3.68988603e-01 -4.76327807e-01 5.99113882e-01
4.92566913e-01 -4.00323749e-01 -5.66904806e-02 3.30621541e-01
-7.29604721e-01 4.00342345e-01 -7.69445658e-01 8.63699377e-01
-1.93098009e-01 5.91829240e-01 -6.43455535e-02 -6.34288013e-01
1.35025346e+00 2.87429124e-01 1.68396086e-01 -4.02911842e-01
3.03760707e-01 1.60424709e-01 1.93794385e-01 -5.64707518e-01
5.28324425e-01 -3.49098802e-01 1.08392769e-03 4.46210712e-01
-4.33568150e-01 1.69981882e-01 -1.95816770e-01 5.34060776e-01
9.76597846e-01 -6.76107705e-01 6.65562987e-01 -4.12939698e-01
4.24889117e-01 -2.66538173e-01 2.87948370e-01 5.64581454e-01
-4.29297537e-01 4.14546341e-01 8.70784581e-01 2.16942765e-02
-8.36855650e-01 -8.32038522e-01 -3.27012390e-01 1.10102761e+00
-3.68244916e-01 -4.62435216e-01 -4.55377072e-01 -1.16996564e-01
-3.55001390e-01 1.22799671e+00 -5.18395603e-01 -1.01417705e-01
3.23501416e-02 -6.79830194e-01 6.45813942e-01 -1.89973742e-01
7.70058095e-01 -9.22351718e-01 -3.24622571e-01 3.22280675e-02
-4.04954821e-01 -4.30462152e-01 -9.85004827e-02 -2.27954149e-01
-5.83244383e-01 -7.04796016e-01 -8.98241341e-01 -4.92068052e-01
4.21783149e-01 -1.84492406e-03 8.77093554e-01 2.34289765e-01
1.59312621e-01 1.45293146e-01 -4.68977123e-01 -4.76449966e-01
-6.64066255e-01 -2.36315608e-01 2.38944635e-01 -2.91836053e-01
1.94264784e-01 -5.89152157e-01 -5.97426653e-01 4.02595066e-02
-4.15284187e-01 3.38259935e-01 1.46807864e-01 7.02957690e-01
-8.81270245e-02 -1.50095418e-01 9.26422715e-01 -7.05607772e-01
7.54965246e-01 -1.07432806e+00 5.12461483e-01 -1.76639214e-01
-6.41698301e-01 -5.13685048e-01 1.17907733e-01 -2.92550266e-01
-7.12221205e-01 -6.24305069e-01 -5.87286353e-01 3.17050427e-01
-1.17956638e-01 7.84033477e-01 4.57044482e-01 4.81740385e-01
8.42847824e-01 -8.36415216e-03 1.47054702e-01 -4.11225975e-01
1.69722915e-01 1.16060019e+00 1.34684324e-01 -1.35718271e-01
4.53226566e-01 3.47912431e-01 -4.76844132e-01 -1.33966780e+00
-6.80837691e-01 -4.93782997e-01 -1.75389469e-01 -5.58744967e-01
1.02718985e+00 -8.37309778e-01 -7.00050533e-01 4.44964379e-01
-1.05318749e+00 -7.68385753e-02 2.77593315e-01 6.31176651e-01
-1.19596682e-02 1.42680749e-01 -6.54104173e-01 -1.36622667e+00
-7.49405563e-01 -5.92597485e-01 2.41156876e-01 4.88039285e-01
-9.01161373e-01 -1.04331374e+00 2.88401097e-01 8.86526525e-01
5.08840442e-01 6.36398375e-01 6.04911745e-01 -8.35307956e-01
4.31465268e-01 -4.35537621e-02 -9.41073075e-02 2.69388467e-01
1.28326356e-01 1.90681994e-01 -8.93810689e-01 6.53464720e-02
7.52084702e-02 -5.31127572e-01 5.52884519e-01 1.86394259e-01
4.36210304e-01 -9.34871614e-01 7.67777860e-02 -1.81860343e-01
9.56820726e-01 7.05104768e-02 7.81681001e-01 3.79179984e-01
2.45945275e-01 5.30965507e-01 2.26055816e-01 9.71296430e-01
4.49845612e-01 7.30777010e-02 1.82751805e-01 -2.21315844e-04
1.15460955e-01 -4.07675534e-01 9.12943304e-01 1.09858048e+00
-4.10567790e-01 -5.76089695e-02 -1.24515283e+00 4.45894748e-01
-1.36219454e+00 -1.23164928e+00 -7.43924916e-01 2.07033706e+00
9.64548171e-01 3.66946012e-01 3.92417997e-01 2.18292803e-01
6.67719603e-01 2.26151600e-01 1.58305224e-02 -7.46100008e-01
-2.55102634e-01 9.86306593e-02 6.74561486e-02 7.57639527e-01
-5.89214623e-01 6.65169120e-01 4.85930252e+00 8.04326653e-01
-1.19919610e+00 3.55222940e-01 1.05127501e+00 -1.30888686e-01
-4.94348526e-01 1.04026370e-01 -5.01561880e-01 4.33289468e-01
1.06771636e+00 -4.14786845e-01 3.19190294e-01 4.52582330e-01
7.59113729e-01 -4.02137965e-01 -3.51622552e-01 8.54745150e-01
1.18579179e-01 -1.20269620e+00 -5.77223897e-01 1.83451355e-01
7.65792966e-01 2.78748602e-01 2.95991302e-01 2.99804389e-01
1.90501153e-01 -1.12454224e+00 6.27719581e-01 6.26539052e-01
5.81036150e-01 -6.24037981e-01 6.12497091e-01 6.37179971e-01
-4.87283885e-01 1.82289869e-01 -2.02291906e-01 -6.70517385e-01
8.88126120e-02 1.04239798e+00 -9.08344030e-01 1.25085473e-01
4.35615182e-01 7.10254252e-01 -4.62654769e-01 5.67034423e-01
-3.01136851e-01 1.07792521e+00 -3.36735487e-01 -6.02795124e-01
2.79524624e-01 -2.51559336e-02 7.86780238e-01 1.23528850e+00
5.27739763e-01 2.20499396e-01 -5.95857054e-02 5.79905450e-01
1.23423927e-01 5.80661297e-01 -4.97647226e-01 -5.33777893e-01
3.00083816e-01 1.29059458e+00 -8.33065510e-01 -2.95647979e-01
-7.29748681e-02 5.69217801e-01 -1.95745118e-02 2.19698921e-02
-7.06433415e-01 -1.54290020e-01 2.25092396e-01 4.79130775e-01
-4.33917284e-01 -1.42013198e-02 -4.12668854e-01 -8.77328932e-01
-5.40218413e-01 -8.64884138e-01 2.71808594e-01 -7.32963562e-01
-1.16058230e+00 2.77614236e-01 -3.08338344e-01 -6.53916180e-01
4.72295731e-02 -2.43204921e-01 -9.20017838e-01 8.08393598e-01
-7.89615810e-01 -9.15702641e-01 3.25204693e-02 4.27513778e-01
3.65079433e-01 -1.17675222e-01 8.94915938e-01 2.50135273e-01
-5.64232528e-01 3.55997950e-01 -2.31092386e-02 1.91189442e-02
7.73159385e-01 -7.81256199e-01 -1.81697577e-01 2.65676677e-01
-3.33421081e-01 6.88339472e-01 8.17056775e-01 -8.46941710e-01
-9.83833134e-01 -6.30401671e-01 1.79230654e+00 -3.08517396e-01
1.00113881e+00 4.59534749e-02 -5.39460123e-01 1.33255899e-01
5.73771417e-01 -7.46508658e-01 1.19605362e+00 5.93683302e-01
-1.39440566e-01 3.23381186e-01 -1.49154806e+00 8.57068241e-01
7.55850136e-01 -5.24149001e-01 -3.04665148e-01 7.35158503e-01
4.93603528e-01 2.82242030e-01 -1.15068173e+00 -3.20045829e-01
6.93394005e-01 -1.37824953e+00 6.95327580e-01 -4.38585937e-01
1.06224108e+00 3.54030907e-01 -2.85765976e-01 -1.09591246e+00
-4.42025930e-01 -9.66489196e-01 3.30093682e-01 1.48627675e+00
6.64127529e-01 -8.08686197e-01 4.83486205e-01 4.87291038e-01
3.77655253e-02 -8.02925944e-01 -9.44244385e-01 -1.17552802e-01
4.22980309e-01 -4.89530861e-01 1.11615099e-02 1.55918145e+00
6.48829341e-01 6.00149453e-01 -5.89254618e-01 -1.89093739e-01
2.34385908e-01 -6.40723288e-01 6.28568411e-01 -1.08745885e+00
-2.96239644e-01 -7.13814795e-01 -1.55206800e-01 -3.07639539e-01
-4.74030972e-01 -9.08674061e-01 -6.76239014e-01 -1.72620642e+00
6.40222013e-01 -3.07283044e-01 -2.77137041e-01 6.63972676e-01
4.29921672e-02 3.01749855e-01 1.93115190e-01 1.72932684e-01
-2.85794765e-01 2.39638716e-01 1.06539869e+00 -7.90209398e-02
-3.71658772e-01 -5.35710938e-02 -1.18330801e+00 9.45026934e-01
1.53225386e+00 -5.22408247e-01 -1.98089853e-01 2.22320780e-01
9.41192567e-01 3.10144246e-01 4.59718972e-01 -8.40832114e-01
-5.55339716e-02 -4.21941787e-01 2.59126931e-01 -5.08271217e-01
4.43021566e-01 -1.85929909e-02 3.63159865e-01 1.00756061e+00
-3.61268789e-01 2.35175770e-02 -1.96174100e-01 3.79588753e-02
2.05145612e-01 -3.89910489e-01 6.96267664e-01 -1.22148477e-01
8.14508796e-02 -2.66017377e-01 -5.63205898e-01 1.29737794e-01
6.62344694e-01 -1.06180742e-01 -7.12048113e-01 -9.22232449e-01
-9.27575469e-01 -7.57363886e-02 3.11595768e-01 3.36851716e-01
6.08880758e-01 -1.15226984e+00 -1.20463157e+00 -1.37231871e-01
-7.28081539e-02 -8.65563214e-01 1.89139456e-01 1.45608592e+00
-2.48441949e-01 -1.55971572e-01 -1.77094430e-01 -3.22768651e-02
-1.28532100e+00 -1.77166332e-02 -3.16759586e-01 7.53674060e-02
-2.08397791e-01 9.37404573e-01 -2.58580238e-01 -1.82839155e-01
-2.36089900e-01 2.54855961e-01 -4.74587202e-01 4.88586813e-01
5.92791378e-01 1.02149558e+00 -4.35501575e-01 -9.56518173e-01
-4.75729197e-01 6.09730296e-02 1.58485491e-02 -3.74087691e-01
1.38302994e+00 2.41071009e-03 -5.20355225e-01 8.78963947e-01
1.17603683e+00 5.99175036e-01 -1.46607921e-01 4.03383225e-01
-1.87548593e-01 -4.41895276e-01 7.13859051e-02 -9.62287724e-01
-3.99856925e-01 4.73073304e-01 5.49563766e-01 5.98689973e-01
5.85230589e-01 2.64730960e-01 6.05381668e-01 6.63313866e-02
-1.69799894e-01 -1.05101717e+00 3.23911220e-01 6.63201809e-01
1.14306271e+00 -1.22133601e+00 -1.02028605e-02 -1.40058875e-01
-7.46781528e-01 9.96334255e-01 4.80327457e-02 2.01645836e-01
9.12392735e-01 -1.83256552e-01 5.91607541e-02 -2.37877011e-01
-7.47092605e-01 1.35493249e-01 -6.04958124e-02 3.72155964e-01
1.16659260e+00 5.71697891e-01 -1.03317440e+00 7.22227275e-01
-7.14327574e-01 -4.90503721e-02 7.39777923e-01 4.97861356e-01
-8.40468347e-01 -9.00752783e-01 -5.97285151e-01 5.70697844e-01
-8.66327584e-01 -2.29661718e-01 -7.39168704e-01 5.20517945e-01
3.21006894e-01 1.58507144e+00 -7.74375647e-02 -6.78187191e-01
-2.50711769e-01 1.74112692e-01 -1.49825841e-01 -1.98716640e-01
-7.79398799e-01 4.59593952e-01 8.81161034e-01 2.16758296e-01
-5.19974411e-01 -8.13793719e-01 -1.19194007e+00 -1.00647891e+00
-1.86947331e-01 3.39991927e-01 6.19191825e-01 6.06941938e-01
2.01206371e-01 6.52514771e-02 6.02294445e-01 -8.99797752e-02
-7.57791102e-02 -1.20495868e+00 -4.44848627e-01 -7.99535885e-02
1.15767084e-01 -3.21090102e-01 -4.09604937e-01 -2.50783265e-01] | [8.975335121154785, 10.714468002319336] |
9f952e21-c9d7-470f-abcc-0972a9372be9 | ab-ba-analysis-a-framework-for-estimating | 2204.08474 | null | https://arxiv.org/abs/2204.08474v1 | https://arxiv.org/pdf/2204.08474v1.pdf | AB/BA analysis: A framework for estimating keyword spotting recall improvement while maintaining audio privacy | Evaluation of keyword spotting (KWS) systems that detect keywords in speech is a challenging task under realistic privacy constraints. The KWS is designed to only collect data when the keyword is present, limiting the availability of hard samples that may contain false negatives, and preventing direct estimation of model recall from production data. Alternatively, complementary data collected from other sources may not be fully representative of the real application. In this work, we propose an evaluation technique which we call AB/BA analysis. Our framework evaluates a candidate KWS model B against a baseline model A, using cross-dataset offline decoding for relative recall estimation, without requiring negative examples. Moreover, we propose a formulation with assumptions that allow estimation of relative false positive rate between models with low variance even when the number of false positives is small. Finally, we propose to leverage machine-generated soft labels, in a technique we call Semi-Supervised AB/BA analysis, that improves the analysis time, privacy, and cost. Experiments with both simulation and real data show that AB/BA analysis is successful at measuring recall improvement in conjunction with the trade-off in relative false positive rate. | ['Benjamin L. Bullough', 'Thibaud Senechal', 'Vasistakrishna Baderdinni', 'Raphael Petegrosso'] | 2022-04-18 | null | https://aclanthology.org/2022.naacl-industry.4 | https://aclanthology.org/2022.naacl-industry.4.pdf | naacl-acl-2022-7 | ['keyword-spotting'] | ['speech'] | [ 3.16233605e-01 1.80837408e-01 -1.93768442e-01 -5.10587513e-01
-1.43304574e+00 -7.11273372e-01 4.59145784e-01 3.82576913e-01
-4.75318074e-01 6.56545401e-01 -1.18610255e-01 -4.64925766e-01
2.95304712e-02 -2.77226508e-01 -8.95275295e-01 -5.33007085e-01
1.41680494e-01 2.79692978e-01 3.36835414e-01 3.41098696e-01
-8.43078047e-02 3.03236037e-01 -1.04532945e+00 1.67831451e-01
7.18942583e-01 1.20873642e+00 1.40569195e-01 7.16878295e-01
9.46486220e-02 5.40356159e-01 -9.87268329e-01 -6.00771725e-01
4.99255508e-01 -8.92222300e-02 -4.03572649e-01 -1.16437271e-01
2.91814804e-01 -5.63409388e-01 -7.70651773e-02 1.04474926e+00
5.09494483e-01 -9.02406946e-02 3.46235007e-01 -1.50306487e+00
-1.14229642e-01 6.10353649e-01 -4.62099046e-01 1.05834818e-02
3.23183537e-01 -1.05893917e-01 1.08808196e+00 -8.43591034e-01
2.27741897e-01 9.74873900e-01 4.97470379e-01 5.01291335e-01
-1.22222757e+00 -1.10091400e+00 1.14403129e-01 -8.00606012e-02
-1.69389653e+00 -7.85234988e-01 4.46375549e-01 -1.43258005e-01
7.43806124e-01 7.00339913e-01 1.94474995e-01 1.10258353e+00
-3.00553709e-01 1.17613733e+00 1.04542279e+00 -5.51654816e-01
5.65007448e-01 7.23329782e-01 3.46504033e-01 3.48164469e-01
3.38615716e-01 -5.01257144e-02 -1.05594909e+00 -7.48217702e-01
-3.26404482e-01 -3.38413209e-01 -6.75138116e-01 -3.25701594e-01
-7.98580647e-01 5.19250214e-01 -4.52297896e-01 -8.89847949e-02
-6.73238188e-02 -3.21665257e-01 3.25533658e-01 3.17520648e-01
6.30807936e-01 2.37449139e-01 -6.86635137e-01 -1.65005356e-01
-1.40267539e+00 2.70673722e-01 9.98929620e-01 1.39303935e+00
5.51196814e-01 -2.80982345e-01 -2.31632635e-01 8.04436445e-01
3.52822542e-01 7.82682002e-01 6.08126998e-01 -4.02110606e-01
7.35434055e-01 2.41337836e-01 2.76806533e-01 -7.67928183e-01
-7.44613307e-03 -5.24611592e-01 -3.20443898e-01 -4.16826129e-01
3.86534631e-01 -8.88145566e-02 -6.67273402e-01 1.82575846e+00
2.93104619e-01 2.14713246e-01 1.77865908e-01 6.49426997e-01
3.09657007e-01 6.63797200e-01 -1.64419577e-01 -6.76915050e-01
1.32256544e+00 -6.61244750e-01 -9.72676933e-01 -2.06945643e-01
7.49377251e-01 -6.95076168e-01 1.15136147e+00 4.24731731e-01
-8.46450329e-01 6.13336526e-02 -1.12300825e+00 3.27992141e-01
-3.08282971e-01 2.75273740e-01 2.57705241e-01 1.02490735e+00
-7.91733801e-01 4.19192426e-02 -6.27293289e-01 -9.02798101e-02
9.09209996e-02 3.57875317e-01 -3.38714987e-01 1.35470599e-01
-1.25377274e+00 5.83593249e-01 1.79439634e-01 -8.15514382e-03
-8.39295745e-01 -6.54124141e-01 -6.02073908e-01 1.77936003e-01
7.81922162e-01 5.32344133e-02 1.63068104e+00 -6.40774608e-01
-1.27530181e+00 4.22514617e-01 -5.11085272e-01 -8.83418560e-01
5.99275053e-01 -2.04217836e-01 -6.40035331e-01 -1.44925350e-02
-1.16126321e-01 1.84708104e-01 9.99488652e-01 -1.31390059e+00
-7.32285202e-01 -2.77148575e-01 -2.18884945e-01 2.23422181e-02
-5.98231375e-01 3.99784977e-03 -7.09481776e-01 -5.58761716e-01
-4.71965335e-02 -8.81310582e-01 1.29910931e-01 -2.68369913e-01
-7.48415351e-01 -2.97563486e-02 9.55220461e-01 -8.36330056e-01
1.45314503e+00 -2.25378871e+00 -5.89481115e-01 7.79349923e-01
1.94922179e-01 6.96934998e-01 1.65493950e-01 4.61588562e-01
3.02652478e-01 2.82914668e-01 -1.75488502e-01 -6.03115559e-01
-2.34584836e-03 8.27501118e-02 -6.07104182e-01 3.76395613e-01
1.18034586e-01 4.58288997e-01 -6.64114356e-01 -3.66302967e-01
-9.93929878e-02 -9.22181364e-03 -2.02504918e-01 5.21495104e-01
-2.00967580e-01 -7.84991235e-02 -2.96603709e-01 6.28922880e-01
7.34362066e-01 -1.15018357e-02 3.42567474e-01 -1.18060164e-01
2.44073778e-01 4.99615878e-01 -1.27600253e+00 8.39438975e-01
-5.84931552e-01 6.00724638e-01 2.38558874e-01 -7.78368533e-01
7.67109156e-01 4.02480870e-01 -1.78220913e-01 -5.90420783e-01
5.95558388e-03 1.54102176e-01 -1.88261926e-01 -3.00228268e-01
6.09790027e-01 4.48481500e-01 9.04531777e-02 5.02397358e-01
-8.86501744e-02 1.20455198e-01 -4.95603122e-02 3.26768279e-01
9.49459851e-01 -5.53375661e-01 2.83402383e-01 -4.04826514e-02
3.92757237e-01 -4.78556335e-01 6.76280081e-01 1.04129291e+00
-2.69962847e-01 4.74276543e-01 6.57243013e-01 3.21237475e-01
-7.05114424e-01 -7.91391015e-01 -1.18416384e-01 7.73196876e-01
4.01933342e-02 -6.78917527e-01 -1.04576528e+00 -8.26471210e-01
1.35796815e-01 7.76820838e-01 -2.70487696e-01 -2.60811508e-01
-5.35145476e-02 -5.48023939e-01 8.99435759e-01 1.31486371e-01
2.28801265e-01 -3.59295577e-01 -4.70193386e-01 -8.19371492e-02
-2.09052205e-01 -1.48031425e+00 -7.49275029e-01 4.33673114e-01
-1.69532269e-01 -1.02830243e+00 -5.18926024e-01 -3.22170347e-01
6.50209010e-01 3.07064593e-01 6.23581946e-01 -2.01903298e-01
6.64207563e-02 3.25171113e-01 -4.80499804e-01 -6.09922707e-01
-7.13963509e-01 7.40118474e-02 3.20663631e-01 3.14454526e-01
5.34749031e-01 -1.16201513e-01 -3.13675791e-01 4.04747456e-01
-7.96711564e-01 -1.86426625e-01 4.75084066e-01 8.69364083e-01
5.12331128e-01 -5.95465489e-02 6.98528647e-01 -1.01581740e+00
8.43127668e-01 -4.31821674e-01 -6.63802385e-01 7.25433588e-01
-1.00600672e+00 -2.69467719e-02 5.13997316e-01 -6.79647148e-01
-7.30554998e-01 -1.00926571e-01 2.20297985e-02 -3.91705185e-01
1.43535674e-01 2.56337315e-01 -4.74773347e-01 -1.56633541e-01
3.57331514e-01 3.98389757e-01 -4.62383665e-02 -5.09977698e-01
4.82644290e-02 1.42319477e+00 2.18798980e-01 -2.35111788e-01
6.16950929e-01 3.83883901e-02 -5.20882368e-01 -9.54320312e-01
-6.79260015e-01 -7.60270953e-01 9.72621795e-03 -5.75196445e-02
7.77032152e-02 -1.05073559e+00 -6.55813694e-01 3.44358027e-01
-8.68393958e-01 1.42160980e-02 1.56864375e-02 6.07444167e-01
-1.64614514e-01 5.24415314e-01 -2.67621875e-01 -1.46740675e+00
-5.17376781e-01 -1.22562063e+00 1.12342250e+00 -1.73180491e-01
-4.27363962e-01 -4.75094378e-01 -4.78993982e-01 4.36418772e-01
2.07088813e-01 -3.29861313e-01 5.54853737e-01 -1.54912555e+00
-2.97597975e-01 -7.01766908e-01 -5.96681535e-02 6.58814847e-01
5.34667522e-02 -8.53116810e-02 -1.28124917e+00 -4.75587964e-01
-1.40243992e-02 -2.70136029e-01 4.18894410e-01 -6.13724031e-02
1.40879273e+00 -7.55469978e-01 -2.98895061e-01 1.98776469e-01
8.73535335e-01 3.51226717e-01 -2.63718143e-02 -2.26980790e-01
2.75873393e-01 5.96609056e-01 1.03638089e+00 6.74059749e-01
1.82853237e-01 7.63126850e-01 -7.70739242e-02 9.85291153e-02
3.35748404e-01 -5.88484585e-01 3.74757379e-01 9.16280210e-01
8.74213338e-01 -6.97892904e-01 -7.74856925e-01 5.50116599e-01
-1.57088578e+00 -4.49262261e-01 2.11042672e-01 2.70195270e+00
9.77319181e-01 4.36781347e-01 1.44085109e-01 5.12603521e-01
8.05385649e-01 -1.40429050e-01 -4.47247356e-01 -3.10246050e-01
-7.65647888e-02 4.90216166e-02 1.15919995e+00 6.23924494e-01
-7.71054983e-01 6.79498494e-01 5.84532070e+00 1.13856578e+00
-1.11747694e+00 1.46948516e-01 9.02127624e-01 -2.47524410e-01
-5.22674799e-01 1.59712676e-02 -1.13733494e+00 7.83160150e-01
1.32205665e+00 -3.72978777e-01 2.78962493e-01 7.83095658e-01
2.76035607e-01 -1.81554869e-01 -1.13169944e+00 9.60298240e-01
1.85471743e-01 -8.84952307e-01 -1.41524509e-01 2.03783456e-02
1.12261213e-01 -2.08834484e-01 9.62259397e-02 1.40125617e-01
1.51763394e-01 -4.97642308e-01 1.08474851e+00 1.96095869e-01
7.40574718e-01 -8.22472513e-01 6.96191370e-01 7.90912867e-01
-7.41818309e-01 2.07431032e-03 -5.92902489e-03 3.57765406e-01
-1.68431047e-02 8.97056758e-01 -1.54114425e+00 2.84720451e-01
5.99277556e-01 -7.27835670e-02 -4.27499861e-01 7.12873459e-01
-1.13976328e-02 1.12125468e+00 -6.59308553e-01 -4.55179125e-01
-2.85251229e-03 2.37305254e-01 5.72434545e-01 1.49078465e+00
2.15812370e-01 -1.42452851e-01 1.59805402e-01 4.04743314e-01
-2.23534197e-01 3.99388403e-01 -5.62475204e-01 -1.14864372e-01
1.17781091e+00 1.06592393e+00 -4.00131702e-01 -1.79813489e-01
-3.03832799e-01 7.42344201e-01 2.46292993e-01 4.59377944e-01
-6.85640335e-01 -2.26926029e-01 4.72630799e-01 1.35994002e-01
8.39938372e-02 -4.36749123e-03 8.35208397e-04 -8.96884382e-01
4.86020863e-01 -1.11690474e+00 2.00365290e-01 -4.87774283e-01
-8.49011302e-01 4.24621254e-01 1.58557713e-01 -1.14128375e+00
-3.48954678e-01 -2.50756443e-01 -9.49869677e-02 8.21640134e-01
-1.33632410e+00 -9.01897967e-01 2.59631664e-01 3.64396393e-01
3.64245266e-01 -1.90292209e-01 6.75538123e-01 3.95211756e-01
-6.38770223e-01 1.60176480e+00 2.96364903e-01 -2.18335502e-02
7.91114390e-01 -8.72272313e-01 3.52134883e-01 1.13912821e+00
2.71731287e-01 7.65078068e-01 7.27144599e-01 -5.97739756e-01
-1.31340659e+00 -1.07885337e+00 1.02769721e+00 1.90078914e-02
3.71626705e-01 -9.87657189e-01 -9.29869413e-01 4.28115636e-01
-4.32106048e-01 4.28781807e-02 8.72396648e-01 7.05102086e-02
-4.94014949e-01 -3.84791821e-01 -1.55281913e+00 4.23285186e-01
4.95203674e-01 -9.55113351e-01 -5.87354228e-02 3.74906719e-01
1.00119746e+00 -3.57566059e-01 -4.72816676e-01 3.31963092e-01
6.17625952e-01 -4.75518376e-01 5.65971673e-01 -4.44815278e-01
-4.47710037e-01 -2.64058590e-01 -4.29118901e-01 -9.43602204e-01
4.44822878e-01 -8.54386270e-01 -4.18310374e-01 1.78195632e+00
8.26512158e-01 -5.85351110e-01 6.56976819e-01 9.66898620e-01
3.49855959e-01 -5.82873464e-01 -1.24426687e+00 -9.47386324e-01
-5.49461603e-01 -8.17828774e-01 6.75769210e-01 7.91302383e-01
2.62824651e-02 -7.92715251e-02 -5.90692222e-01 6.71467543e-01
4.42259520e-01 -3.41672748e-01 8.51238012e-01 -5.70463359e-01
-4.21632737e-01 1.63980931e-01 -3.28221142e-01 -1.01761639e+00
3.03247243e-01 -5.73854148e-01 2.92877078e-01 -5.16720653e-01
4.14109975e-03 -5.19169509e-01 -2.17100412e-01 4.50374752e-01
-7.29385242e-02 -7.24187493e-02 2.25534737e-02 6.43268004e-02
-5.24884641e-01 3.49496365e-01 4.35522825e-01 -6.79321289e-02
-1.06508948e-01 4.60023582e-01 -5.35001397e-01 6.01984024e-01
7.57243156e-01 -6.44766152e-01 -5.89322448e-01 -7.52234235e-02
-1.01661518e-01 1.40337214e-01 7.38243014e-02 -7.74206221e-01
3.76861900e-01 -7.05482140e-02 -2.07801387e-01 -4.99952972e-01
3.26546162e-01 -9.77743745e-01 -1.97848137e-02 1.68987453e-01
-8.35380077e-01 -2.12344378e-01 -4.24493439e-02 9.08463836e-01
-1.46708757e-01 -4.16448206e-01 5.34161091e-01 2.98791677e-01
-3.17981690e-01 1.09438144e-01 -2.73899645e-01 -5.57419881e-02
1.08491290e+00 -5.39719127e-02 -1.01371385e-01 -6.82917416e-01
-2.60417044e-01 4.33575064e-01 1.13183931e-01 3.65085512e-01
4.46397543e-01 -8.16320956e-01 -4.55317020e-01 3.24883193e-01
5.22993207e-01 -3.08075011e-01 -1.44597918e-01 5.46737432e-01
-4.06537317e-02 3.35607409e-01 8.72007132e-01 -5.13463855e-01
-1.81085253e+00 4.61046755e-01 7.14016557e-02 -1.50887623e-01
-2.26304799e-01 9.90888000e-01 8.61107260e-02 -3.01264793e-01
8.21178734e-01 -3.86715621e-01 2.86515415e-01 -6.22871779e-02
6.91012681e-01 2.56440103e-01 5.84112465e-01 -5.06097913e-01
-4.53895479e-01 -2.18433037e-01 -3.88764471e-01 -3.00405443e-01
6.24599516e-01 -3.94994348e-01 3.10486197e-01 5.24832904e-01
1.25608182e+00 4.54659998e-01 -9.17739213e-01 -5.81089497e-01
2.97256470e-01 -6.24222517e-01 7.69220516e-02 -1.00004542e+00
-8.29992950e-01 7.25814760e-01 7.15879261e-01 2.77392030e-01
1.10447490e+00 -2.20696196e-01 9.52991545e-01 5.15435874e-01
3.60716254e-01 -1.30828559e+00 -4.87838209e-01 5.00970781e-02
5.60833454e-01 -1.26857495e+00 -1.91918716e-01 -5.02530456e-01
-7.74789095e-01 5.76576889e-01 3.98167163e-01 6.94045901e-01
6.82438612e-01 6.77021861e-01 1.27288342e-01 2.98943520e-01
-9.61328745e-01 1.71984673e-01 2.53516257e-01 3.73107940e-01
1.14764348e-01 2.85180748e-01 -3.45064908e-01 9.86592650e-01
-2.38123238e-01 -2.33385980e-01 3.99834931e-01 9.32987511e-01
-2.49544203e-01 -1.05647469e+00 -2.62070596e-01 7.37161160e-01
-8.49132538e-01 -2.40337536e-01 -5.07059813e-01 4.41681981e-01
-1.86059415e-01 1.25286460e+00 -2.96410657e-02 -6.43914759e-01
4.09896851e-01 1.86632141e-01 -7.32007772e-02 -3.65844220e-01
-4.81712967e-01 9.97113530e-03 4.25998539e-01 -3.28757793e-01
-2.38975696e-02 -5.69280088e-01 -9.02486980e-01 -2.31361419e-01
-9.74010348e-01 6.49146318e-01 8.66340756e-01 1.03402567e+00
3.71817917e-01 3.91372517e-02 1.04814267e+00 7.52422363e-02
-9.69474316e-01 -9.69763219e-01 -6.79291308e-01 1.13138877e-01
5.80988169e-01 -4.04351771e-01 -5.44598281e-01 -1.85532585e-01] | [14.241880416870117, 6.316101551055908] |
f0f453c9-f91b-4662-a8b3-2aa327d93285 | self-supervised-ppg-representation-learning | 2212.04902 | null | https://arxiv.org/abs/2212.04902v2 | https://arxiv.org/pdf/2212.04902v2.pdf | Self-Supervised PPG Representation Learning Shows High Inter-Subject Variability | With the progress of sensor technology in wearables, the collection and analysis of PPG signals are gaining more interest. Using Machine Learning, the cardiac rhythm corresponding to PPG signals can be used to predict different tasks such as activity recognition, sleep stage detection, or more general health status. However, supervised learning is often limited by the amount of available labeled data, which is typically expensive to obtain. To address this problem, we propose a Self-Supervised Learning (SSL) method with a pretext task of signal reconstruction to learn an informative generalized PPG representation. The performance of the proposed SSL framework is compared with two fully supervised baselines. The results show that in a very limited label data setting (10 samples per class or less), using SSL is beneficial, and a simple classifier trained on SSL-learned representations outperforms fully supervised deep neural networks. However, the results reveal that the SSL-learned representations are too focused on encoding the subjects. Unfortunately, there is high inter-subject variability in the SSL-learned representations, which makes working with this data more challenging when labeled data is scarce. The high inter-subject variability suggests that there is still room for improvements in learning representations. In general, the results suggest that SSL may pave the way for the broader use of machine learning models on PPG data in label-scarce regimes. | ['Marcel J. T. Reinders', 'David M. J. Tax', 'Ramin Ghorbani'] | 2022-12-07 | null | null | null | null | ['sleep-stage-detection'] | ['medical'] | [ 3.61517400e-01 2.59002149e-01 -5.97875535e-01 -6.11246049e-01
-8.12718987e-01 -3.97927731e-01 8.42888430e-02 7.45996684e-02
3.95839773e-02 7.50839591e-01 4.85257149e-01 1.19030029e-02
-2.86195632e-02 -4.05922681e-01 -4.79048789e-01 -9.21813250e-01
-1.62967414e-01 5.35011617e-03 -3.97656441e-01 3.48752737e-01
-3.22546571e-01 2.66017586e-01 -1.43275511e+00 1.91210315e-01
4.81188059e-01 1.42417765e+00 -4.62737400e-03 3.54444474e-01
1.01072676e-01 5.24959087e-01 -6.75342739e-01 2.12033316e-01
1.83510095e-01 -8.10257792e-01 -5.40243566e-01 1.62948325e-01
3.55515897e-01 1.67751499e-02 -1.41737223e-01 7.43920386e-01
9.59530056e-01 -1.15433015e-01 3.78373623e-01 -1.09072077e+00
-1.09116703e-01 4.76796716e-01 -1.46655306e-01 5.08414388e-01
3.18131238e-01 1.82534933e-01 1.08241975e+00 -5.18442988e-01
1.87398553e-01 8.36691380e-01 9.65059459e-01 6.26198471e-01
-1.43264663e+00 -7.22427726e-01 -2.09291428e-02 -2.25960254e-03
-1.21657372e+00 -4.91698056e-01 9.45722580e-01 -3.90549570e-01
6.94628656e-01 3.13502997e-01 9.18036819e-01 1.51772583e+00
1.15687139e-01 5.09380221e-01 1.57048321e+00 -3.71196210e-01
3.54170829e-01 1.79065153e-01 2.54408300e-01 4.51645017e-01
3.73784095e-01 1.76575825e-01 -8.16438675e-01 -1.41848654e-01
5.56626201e-01 1.61146343e-01 -3.68953109e-01 -7.61425868e-02
-1.04945016e+00 6.52553737e-01 3.43275219e-01 5.78455865e-01
-4.99031931e-01 9.18268040e-02 3.48890364e-01 2.85258859e-01
9.21789050e-01 8.01972926e-01 -5.23200572e-01 -3.74329895e-01
-1.38714933e+00 -6.10447768e-03 9.50919807e-01 3.87447774e-01
6.04843736e-01 8.45424458e-02 -5.10175824e-01 8.78594160e-01
3.36632609e-01 4.99446929e-01 5.76553524e-01 -9.99800384e-01
1.79110169e-01 5.50528824e-01 -8.56869966e-02 -8.79252791e-01
-9.48381245e-01 -1.01157904e+00 -7.20127285e-01 -2.62371659e-01
4.11400676e-01 -3.44618827e-01 -7.09502518e-01 1.78804886e+00
-1.05836585e-01 3.99205387e-01 -3.01725268e-01 8.79790246e-01
1.01577687e+00 5.16282856e-01 2.48563692e-01 -6.06178701e-01
1.28638971e+00 -5.10064542e-01 -8.21703851e-01 -6.74201727e-01
4.31881428e-01 -1.74400777e-01 1.08374083e+00 4.86267418e-01
-6.95460379e-01 -4.72950965e-01 -1.03497803e+00 3.58572513e-01
-5.63320750e-03 3.34267616e-01 9.16372240e-01 1.04696691e+00
-8.24470580e-01 7.78318107e-01 -8.29293013e-01 -5.86866140e-01
8.28748226e-01 4.75264341e-01 -1.15594789e-01 -1.83590382e-01
-1.17136824e+00 5.41383326e-01 2.26601809e-01 2.47859254e-01
-6.87238455e-01 -7.24855661e-01 -5.46899617e-01 1.59439996e-01
-3.64448614e-02 -4.80517626e-01 8.21432531e-01 -8.13562572e-01
-1.47426164e+00 1.09201264e+00 -4.71562892e-02 -5.53018093e-01
1.27582625e-01 -3.18157263e-02 -7.21739173e-01 2.42359608e-01
-9.99326855e-02 4.86613482e-01 1.14214683e+00 -8.20581138e-01
-1.96599856e-01 -5.39043009e-01 -4.29513246e-01 9.63045936e-03
-2.69190609e-01 -1.02486037e-01 -3.88121419e-02 -5.38971484e-01
3.45392942e-01 -1.03611708e+00 1.02723420e-01 -1.07461892e-01
-1.87991098e-01 -4.35030788e-01 3.26917857e-01 -6.03590071e-01
1.13005877e+00 -2.40620685e+00 -3.67206961e-01 2.05557048e-01
5.64946115e-01 3.53630483e-01 1.01713926e-01 1.32967487e-01
-2.20679596e-01 -3.35330069e-02 -1.66634788e-06 -3.13934535e-01
-1.56016946e-01 3.21104676e-01 -1.45140171e-01 5.25085807e-01
1.13984205e-01 9.59770322e-01 -8.92183423e-01 -9.53389779e-02
1.31617680e-01 3.40108424e-01 -3.04456979e-01 3.55808914e-01
-3.71857584e-02 9.13186252e-01 -2.29938269e-01 6.78014040e-01
3.76239806e-01 -5.28567612e-01 1.59652963e-01 -3.84086400e-01
2.54447669e-01 7.16301799e-01 -7.68247902e-01 1.83416927e+00
-4.29945827e-01 8.39503348e-01 -3.57966781e-01 -1.53562343e+00
1.17001307e+00 4.05266345e-01 7.76090205e-01 -5.81169665e-01
4.05518487e-02 2.33220011e-01 2.91445315e-01 -7.53216147e-01
-4.37151045e-01 -3.84572774e-01 3.78678292e-02 5.36014795e-01
3.24033201e-01 1.41840115e-01 -1.22905031e-01 -2.08721265e-01
1.29418349e+00 -3.00423875e-02 2.14055493e-01 -1.69974104e-01
1.23301961e-01 -3.09312314e-01 8.56709838e-01 9.13154781e-01
-4.52840239e-01 6.43232107e-01 4.03245926e-01 -4.46525931e-01
-3.83730352e-01 -1.00121641e+00 -4.23653543e-01 9.24520433e-01
-3.09031129e-01 -5.78264594e-01 -4.54651743e-01 -6.16432488e-01
-1.65180601e-02 3.44739825e-01 -5.42639971e-01 -5.00456452e-01
-3.26009870e-01 -1.18705738e+00 5.68684161e-01 6.15548611e-01
5.05490184e-01 -1.25739539e+00 -5.05796909e-01 3.88345093e-01
-2.02715054e-01 -1.19364631e+00 -1.38070732e-02 4.36152697e-01
-1.33142257e+00 -8.83181691e-01 -6.80508077e-01 -4.32396114e-01
4.20388162e-01 -8.90784413e-02 1.20625579e+00 -2.76246756e-01
-3.33209664e-01 6.62444293e-01 -3.75183344e-01 -5.27828753e-01
-1.18091747e-01 2.40972638e-01 1.47069439e-01 3.30815673e-01
7.78147757e-01 -9.36306536e-01 -9.18875337e-01 4.31684554e-02
-1.86534137e-01 -3.75517875e-01 6.47827029e-01 6.82264447e-01
5.38523018e-01 -3.21101904e-01 1.21034634e+00 -8.71674836e-01
7.17698514e-01 -5.63404083e-01 1.89816859e-02 -7.34569207e-02
-8.94032300e-01 -2.68533565e-02 2.72910178e-01 -5.45348585e-01
-6.50303483e-01 -1.15251236e-01 -4.47806194e-02 -3.28120887e-01
-2.87006408e-01 5.26704788e-01 -1.12466320e-01 9.17643979e-02
8.93402636e-01 2.21306041e-01 1.59720451e-01 -5.57062984e-01
-8.70128721e-02 9.60463524e-01 4.11976337e-01 -4.20276642e-01
2.75693595e-01 1.26652926e-01 1.19000569e-01 -9.40896571e-01
-1.27716708e+00 -4.78002816e-01 -2.01175272e-01 -3.29348117e-01
7.05764115e-01 -1.15252662e+00 -4.23450381e-01 2.06469044e-01
-5.14323771e-01 -3.68145704e-01 -7.43437290e-01 8.41433108e-01
-5.47000647e-01 9.66990814e-02 -3.57678622e-01 -8.68677437e-01
-5.01805842e-01 -7.44671345e-01 1.02063251e+00 1.68854892e-01
-6.61308050e-01 -9.64185357e-01 1.40966237e-01 6.47856116e-01
5.37755966e-01 2.50833184e-01 6.77337050e-01 -7.84882605e-01
-1.74239844e-01 -3.95248920e-01 4.10355404e-02 6.23223126e-01
5.72888196e-01 -8.05079520e-01 -1.44942641e+00 -3.30273747e-01
3.19566309e-01 -5.51068604e-01 7.81050324e-01 6.51024520e-01
1.45953465e+00 -2.16392070e-01 -2.83020914e-01 6.73537850e-01
8.78914595e-01 -6.70779571e-02 5.99203527e-01 -1.28612369e-01
5.63439369e-01 4.82162535e-01 2.17361256e-01 4.43983912e-01
2.16248870e-01 5.36606431e-01 9.39371958e-02 -1.95696592e-01
-4.64395672e-01 -1.23561680e-01 3.21498960e-01 7.27298915e-01
-7.02333152e-02 1.68778032e-01 -6.88460708e-01 2.80561030e-01
-1.56315935e+00 -7.67485261e-01 1.67991117e-01 2.31639910e+00
8.42732906e-01 -7.04769418e-02 3.89525265e-01 5.51148295e-01
4.27684426e-01 3.08169216e-01 -7.05738425e-01 -1.67671382e-01
6.72895089e-02 4.21821415e-01 2.43503615e-01 -1.01294301e-01
-1.13518023e+00 2.85908401e-01 6.79208851e+00 3.89213413e-01
-1.58675647e+00 2.93700755e-01 6.33274972e-01 -2.01754928e-01
1.90504700e-01 -3.31229061e-01 -6.45778179e-01 8.74465764e-01
1.44831598e+00 1.71157315e-01 5.34467101e-01 5.73491514e-01
6.29598141e-01 -1.52587378e-02 -1.31679833e+00 1.69335234e+00
2.16464013e-01 -1.08876181e+00 -6.42373323e-01 1.37859499e-02
4.59058315e-01 1.85628489e-01 -1.31548285e-01 3.92707914e-01
-4.81792897e-01 -1.03160346e+00 -1.60355363e-02 5.14745593e-01
8.06078076e-01 -2.19054341e-01 6.45267963e-01 2.98565209e-01
-7.69925356e-01 -1.79531693e-01 -1.49879873e-01 -2.54981279e-01
-2.22711116e-01 1.10559976e+00 -8.36441636e-01 1.76808566e-01
8.18729460e-01 1.14511287e+00 -7.61638641e-01 1.18224025e+00
-3.76844317e-01 1.45381403e+00 -4.07260060e-01 5.29696830e-02
-2.89467543e-01 -4.95898798e-02 3.69363993e-01 1.16535282e+00
3.03024828e-01 -1.56038748e-02 2.17537344e-01 9.82711613e-01
-2.39918977e-01 -4.47926521e-02 -6.78388715e-01 -4.15785581e-01
2.60143816e-01 1.20043588e+00 -7.03227103e-01 -1.87332451e-01
-4.55746204e-01 6.67180240e-01 5.41351959e-02 3.80616933e-01
-4.03996885e-01 1.12380177e-01 3.86841089e-01 4.13766682e-01
-1.09026425e-01 -3.81641020e-03 -6.98033750e-01 -1.21828938e+00
2.02424694e-02 -8.06046128e-01 4.32598263e-01 -5.69126487e-01
-1.50185812e+00 3.46157312e-01 -8.23310614e-02 -1.37726462e+00
-3.78621459e-01 -3.59844446e-01 -5.35629988e-01 6.26281321e-01
-1.28743684e+00 -7.04810381e-01 -4.52277362e-01 5.34088552e-01
3.54753137e-01 -1.35203809e-01 1.05783153e+00 4.70399052e-01
-6.45805001e-01 5.50194204e-01 -2.02780902e-01 1.27553418e-02
8.62482607e-01 -1.10579586e+00 -2.33808503e-01 5.41446865e-01
4.66166794e-01 5.70948899e-01 4.64422017e-01 -3.84191066e-01
-1.19520593e+00 -1.08251858e+00 8.06527674e-01 -3.32949519e-01
1.51103243e-01 -4.91559535e-01 -7.37933159e-01 5.46801925e-01
-1.15472041e-01 7.03882515e-01 1.21374428e+00 3.98733318e-01
-6.50371984e-02 -7.17540622e-01 -1.16829813e+00 -1.41245067e-01
9.70830917e-01 -6.58794641e-01 -6.40674174e-01 4.72846448e-01
1.93676770e-01 -1.77016035e-01 -9.62260008e-01 3.38634849e-01
5.98629057e-01 -6.00889921e-01 8.63169372e-01 -4.34558153e-01
-5.06619215e-02 -2.40182281e-02 2.35649608e-02 -1.43806350e+00
-2.44886100e-01 -5.22612751e-01 -4.41002816e-01 1.26175368e+00
3.31031829e-01 -8.33566964e-01 1.00981426e+00 7.95880675e-01
-4.10272889e-02 -7.82278419e-01 -8.14618468e-01 -8.26418757e-01
-3.09969395e-01 -6.17894709e-01 2.14300364e-01 9.63076174e-01
1.02708064e-01 5.74298382e-01 -6.08993232e-01 -5.34667335e-02
6.27140522e-01 1.76178232e-01 2.72162497e-01 -1.79149270e+00
-4.93705451e-01 1.13951236e-01 -5.96172333e-01 -6.62108064e-01
2.11020097e-01 -1.15030229e+00 2.55198218e-02 -1.46622658e+00
2.40362380e-02 -6.03881240e-01 -7.92110324e-01 6.47346616e-01
-1.05038196e-01 5.71686804e-01 6.67097867e-02 3.14797968e-01
-4.10515726e-01 3.23072463e-01 9.07715023e-01 -2.24993572e-01
-5.51259279e-01 5.73419213e-01 -8.48509431e-01 7.16628492e-01
1.01840842e+00 -6.78221405e-01 -5.45862436e-01 1.33306801e-01
1.26282662e-01 -3.09623796e-02 2.37829566e-01 -1.25472414e+00
-6.77916110e-02 3.20290715e-01 7.98770368e-01 -1.55311584e-01
3.88212264e-01 -7.00797617e-01 -1.15554310e-01 3.89091879e-01
-5.40359855e-01 -6.79892778e-01 -4.45379131e-02 5.47134995e-01
-5.37632704e-02 8.46689474e-03 5.75751543e-01 -2.25633472e-01
-3.61593097e-01 2.94940948e-01 -4.12115335e-01 3.33938062e-01
4.24133629e-01 -2.79179305e-01 5.57727553e-02 -4.81781453e-01
-1.14358270e+00 -2.29874015e-01 -2.71513790e-01 4.30714548e-01
3.60821873e-01 -1.29218745e+00 -4.33830887e-01 4.91073072e-01
2.46554703e-01 -2.75320321e-01 3.38438340e-02 9.97127831e-01
1.03255861e-01 3.79901499e-01 -1.63192138e-01 -9.49091852e-01
-7.97157288e-01 2.28968829e-01 4.44642723e-01 -1.95958614e-01
-8.85630667e-01 5.82775116e-01 -5.00102118e-02 -9.23230201e-02
3.14642817e-01 -5.26739895e-01 -3.71578276e-01 2.77057230e-01
4.37012494e-01 4.11347270e-01 4.59227473e-01 -1.88152358e-01
-4.46257383e-01 3.87866080e-01 2.50769943e-01 3.04288030e-01
1.48324323e+00 -7.48301521e-02 7.88753629e-02 8.35918009e-01
1.02673936e+00 -2.28900775e-01 -1.13411343e+00 -2.10936069e-01
1.40768915e-01 -3.02217841e-01 2.63668120e-01 -9.16346490e-01
-1.25251329e+00 1.03135872e+00 1.28221977e+00 2.82735713e-02
1.31136715e+00 2.18708768e-01 4.95055825e-01 3.26663017e-01
4.29147601e-01 -9.76226926e-01 1.14651591e-01 -6.38275445e-02
5.90114176e-01 -1.29591060e+00 2.66746487e-02 -2.88907230e-01
-4.86330599e-01 9.41419899e-01 2.66767949e-01 -2.50366092e-01
7.48303652e-01 -5.41274510e-02 -1.12702791e-02 -2.92664617e-01
-4.42370206e-01 -2.46666372e-01 4.47344542e-01 8.93993497e-01
6.28996372e-01 2.05812052e-01 -4.55669373e-01 1.01203048e+00
-1.64050937e-01 3.91043931e-01 3.51723880e-01 5.67737281e-01
-2.05056474e-01 -1.19831645e+00 -1.61890388e-01 1.10623050e+00
-8.01474869e-01 3.00780963e-02 -1.80889040e-01 3.12377423e-01
3.14708412e-01 1.28648961e+00 -1.97205544e-02 -2.19607875e-01
1.69308439e-01 5.95930994e-01 5.98913491e-01 -1.05755985e+00
-5.94099045e-01 1.06946141e-01 1.71679065e-01 -7.15196967e-01
-7.39888549e-01 -6.87956452e-01 -8.59120369e-01 5.10335386e-01
-8.98461044e-02 -4.06242795e-02 4.64440912e-01 1.20939445e+00
5.28164268e-01 5.90933859e-01 6.84538007e-01 -7.72227168e-01
-4.08915788e-01 -1.15961051e+00 -8.97905946e-01 4.44537401e-01
4.06930745e-01 -7.29581356e-01 -4.96334463e-01 -8.77307877e-02] | [13.465763092041016, 3.435053586959839] |
2ceed2e4-db28-4e57-b47b-9eedb269f593 | topic-sensitive-neural-headline-generation | 1608.05777 | null | http://arxiv.org/abs/1608.05777v1 | http://arxiv.org/pdf/1608.05777v1.pdf | Topic Sensitive Neural Headline Generation | Neural models have recently been used in text summarization including
headline generation. The model can be trained using a set of document-headline
pairs. However, the model does not explicitly consider topical similarities and
differences of documents. We suggest to categorizing documents into various
topics so that documents within the same topic are similar in content and share
similar summarization patterns. Taking advantage of topic information of
documents, we propose topic sensitive neural headline generation model. Our
model can generate more accurate summaries guided by document topics. We test
our model on LCSTS dataset, and experiments show that our method outperforms
other baselines on each topic and achieves the state-of-art performance. | ['ZiYun Wang', 'Ayana', 'Lei Xu', 'Maosong Sun', 'Zhiyuan Liu'] | 2016-08-20 | null | null | null | null | ['headline-generation'] | ['natural-language-processing'] | [ 2.31835648e-01 2.85410166e-01 -5.46339035e-01 -4.22623336e-01
-1.23366094e+00 -4.74346936e-01 9.07226145e-01 5.91093004e-01
-4.82702814e-02 9.97254074e-01 1.37801540e+00 2.22863480e-01
6.61531463e-02 -7.94225812e-01 -5.06652176e-01 -4.42684799e-01
5.02696000e-02 5.40629625e-01 3.32218617e-01 -3.17880183e-01
8.22474241e-01 -2.31846906e-02 -1.09255862e+00 6.77130759e-01
1.32491970e+00 3.83475184e-01 3.17048699e-01 9.59494531e-01
-3.93088520e-01 4.77365762e-01 -1.26329136e+00 8.37099925e-02
-1.87331408e-01 -8.02775919e-01 -9.60891724e-01 4.30178344e-02
9.43656266e-01 -4.88149971e-01 -3.46410751e-01 8.62305164e-01
6.97548926e-01 4.97548610e-01 1.04975605e+00 -8.32039297e-01
-7.38165617e-01 1.13059330e+00 -7.85547376e-01 3.51231903e-01
3.51574719e-01 -5.24995148e-01 1.20634162e+00 -4.81019974e-01
7.52464533e-01 1.39807701e+00 3.87628824e-01 5.87640107e-01
-1.07081091e+00 -3.23743105e-01 6.09037280e-01 -5.95985427e-02
-7.73237705e-01 -3.92785102e-01 7.34554708e-01 -2.66435780e-02
1.07949209e+00 2.24580124e-01 4.93021250e-01 1.04511225e+00
5.36987305e-01 1.26347649e+00 2.54317373e-01 -3.07450533e-01
2.03795344e-01 -2.15530396e-02 6.88413620e-01 2.10874841e-01
5.81339478e-01 -7.89261758e-01 -6.57620251e-01 -3.58151674e-01
1.46682918e-01 -2.21671745e-01 -6.07571244e-01 8.21525455e-02
-1.28405786e+00 1.21901631e+00 3.29606116e-01 3.73072863e-01
-5.08275449e-01 1.25380188e-01 8.13547134e-01 -3.24885361e-02
9.12788868e-01 9.41516995e-01 -9.56298336e-02 1.00219205e-01
-1.51485395e+00 7.71191120e-01 9.39907849e-01 1.12106609e+00
4.15624559e-01 1.13611169e-01 -9.52306628e-01 1.03979516e+00
-4.61096466e-02 3.84332627e-01 9.89153802e-01 -8.79997373e-01
6.88971817e-01 3.87754142e-01 6.92999642e-03 -1.12166619e+00
-2.92686909e-01 -4.38408077e-01 -1.09331357e+00 -7.55857706e-01
-3.44067067e-01 -4.01704878e-01 -9.44145441e-01 1.42262971e+00
-7.32242465e-02 -1.98980197e-01 2.92968243e-01 3.73398840e-01
1.31600571e+00 1.41625059e+00 -1.44534424e-01 -5.87601185e-01
1.00435090e+00 -1.26648867e+00 -9.80457067e-01 -5.64572588e-02
6.37184143e-01 -5.85934281e-01 7.09668994e-01 1.28759921e-01
-1.43442714e+00 -3.44918489e-01 -1.03592467e+00 -1.93202272e-01
-1.73654392e-01 1.23494208e-01 3.47333342e-01 1.20405003e-01
-1.47026575e+00 6.59690320e-01 -6.49347186e-01 -7.37367392e-01
3.07972521e-01 1.04011543e-01 8.98363963e-02 2.91908026e-01
-1.13027084e+00 4.45157051e-01 7.16595232e-01 -4.29726273e-01
-7.14312255e-01 -6.30888343e-01 -8.69475186e-01 3.81812364e-01
7.14745075e-02 -1.18792963e+00 1.79562426e+00 -3.02802354e-01
-1.39490235e+00 2.82380372e-01 -6.73292041e-01 -7.53774345e-01
6.10497706e-02 -4.19509619e-01 -1.72625676e-01 4.42294061e-01
4.45974678e-01 9.71136212e-01 6.14054859e-01 -1.33376074e+00
-9.40635502e-01 -8.84430110e-02 -2.72637874e-01 5.40691376e-01
-6.04646206e-01 -8.40548985e-03 -3.61839682e-01 -8.76325488e-01
-6.09568581e-02 -6.18513465e-01 -2.37649307e-01 -8.30761135e-01
-1.23887551e+00 -6.02739811e-01 1.08476889e+00 -6.95705950e-01
1.52041698e+00 -1.39296031e+00 4.77272272e-02 -2.09676072e-01
3.21787447e-02 1.15217894e-01 -2.37239718e-01 8.73317420e-01
3.27769279e-01 3.27443749e-01 -2.23802224e-01 -3.93060952e-01
1.94825493e-02 -3.72313887e-01 -1.00520968e+00 -4.89798486e-02
-1.07070766e-01 8.81902039e-01 -8.93795013e-01 -6.81847095e-01
-3.67482692e-01 1.48740932e-02 -5.82843900e-01 2.45058447e-01
-5.98910630e-01 5.77799268e-02 -7.81792581e-01 2.20514208e-01
4.15434897e-01 -2.51799226e-01 -2.66342908e-01 -4.01028022e-02
-2.14131782e-03 8.26084018e-01 -4.47869778e-01 1.71660054e+00
-1.64333805e-01 1.04570317e+00 -6.15307629e-01 -7.81555414e-01
9.23493087e-01 3.65872651e-01 2.45812133e-01 -3.21044683e-01
-9.35593918e-02 -9.96807218e-02 -2.42835328e-01 -1.69862539e-01
1.56753957e+00 3.31374556e-01 -1.99579090e-01 9.90124226e-01
-6.26449436e-02 -4.25797880e-01 7.23448277e-01 8.59555721e-01
7.98304141e-01 -3.75679761e-01 4.70613152e-01 -3.74836624e-01
1.11407392e-01 1.70757264e-01 1.74423575e-01 1.21316540e+00
2.61640102e-01 9.23924267e-01 5.98678291e-01 2.65150107e-02
-1.06425178e+00 -8.44563544e-01 1.10630989e-01 1.23508167e+00
1.24694712e-01 -7.63137043e-01 -9.05349851e-01 -7.47944474e-01
-1.85686588e-01 1.12735760e+00 -6.52583063e-01 -1.89746991e-01
-7.74052143e-01 -7.91517794e-01 4.33154374e-01 5.51671982e-01
5.89223981e-01 -1.10383511e+00 -3.80168676e-01 2.87860632e-01
-4.41916585e-01 -7.19623625e-01 -8.93047690e-01 -2.66503066e-01
-1.33012021e+00 -4.95788813e-01 -1.36602354e+00 -9.05664980e-01
5.64488351e-01 7.48354971e-01 1.12616670e+00 -2.54001170e-01
2.78231680e-01 3.37530732e-01 -3.72172236e-01 -5.83207309e-01
-6.00695670e-01 9.57054794e-01 -1.36466339e-01 -5.75673997e-01
5.37610054e-03 -3.11053276e-01 -6.68124914e-01 -1.88371345e-01
-9.30319786e-01 1.44765198e-01 3.86485100e-01 7.22864389e-01
2.77274907e-01 -1.65132463e-01 1.09592807e+00 -1.08826721e+00
1.53816354e+00 -5.77722192e-01 1.04272924e-01 4.23063040e-01
-5.33319473e-01 1.12127922e-01 5.66347361e-01 -2.96826810e-01
-1.44905400e+00 -7.07623541e-01 2.26681069e-01 1.93761498e-01
-1.71270847e-01 7.99442291e-01 1.01783864e-01 8.20057690e-01
5.96722603e-01 4.58556294e-01 -4.30195153e-01 -3.76931220e-01
5.14752448e-01 8.17165673e-01 6.27250910e-01 -3.70801866e-01
3.26084971e-01 3.96248490e-01 -4.31268990e-01 -1.06505179e+00
-1.03786743e+00 -5.98109782e-01 -3.81604195e-01 1.83934998e-02
6.25658810e-01 -7.81267047e-01 1.03282314e-02 3.74198139e-01
-1.62759590e+00 -1.89864598e-02 -3.32802266e-01 3.30530316e-01
-4.50290293e-01 6.14305913e-01 -7.11675644e-01 -4.48758841e-01
-1.16449964e+00 -6.46886349e-01 1.22264683e+00 6.61741793e-01
-7.12289333e-01 -1.10544586e+00 4.85084206e-01 -1.49554536e-01
5.09039581e-01 1.06530175e-01 9.50840533e-01 -1.23383176e+00
-2.44551122e-01 -2.36523375e-01 -9.72008035e-02 1.20536750e-02
4.69922066e-01 1.81279942e-01 -6.75754189e-01 -4.79133219e-01
-2.22761989e-01 -2.11631611e-01 1.56351578e+00 1.06537330e+00
1.08702922e+00 -8.50332916e-01 -7.76348591e-01 2.16976911e-01
9.72677350e-01 2.09145844e-01 5.17410755e-01 3.80322784e-01
6.47352993e-01 7.22595751e-01 4.92543399e-01 4.43048626e-01
5.05722761e-01 3.51936579e-01 -1.25951231e-01 -2.16947645e-02
-6.22388907e-02 -4.01063472e-01 3.64295453e-01 1.10847402e+00
4.15845156e-01 -1.12387466e+00 -5.92438340e-01 1.02815855e+00
-2.03247237e+00 -1.17279899e+00 -1.59753099e-01 1.73382616e+00
9.42784011e-01 9.20121521e-02 2.16011718e-01 -4.08806473e-01
1.03659451e+00 6.81882560e-01 -5.27345598e-01 -7.80178189e-01
-1.92920566e-01 -2.61832803e-01 2.01609537e-01 4.33308691e-01
-1.06135380e+00 1.11035788e+00 7.21489716e+00 7.17120349e-01
-8.54859412e-01 -4.90744412e-01 6.65808856e-01 -3.33945870e-01
-5.73957145e-01 -2.72544414e-01 -1.10795736e+00 4.50300932e-01
8.48227501e-01 -9.95060742e-01 -4.11934406e-01 7.37150252e-01
4.37587738e-01 -2.75946289e-01 -1.06838512e+00 3.14885557e-01
5.97713113e-01 -1.68853891e+00 7.87280381e-01 -1.50168642e-01
1.41499710e+00 -9.42853466e-02 -5.92389107e-02 2.47315183e-01
6.16263568e-01 -6.73867702e-01 2.88105786e-01 4.64804262e-01
2.56886959e-01 -8.54231536e-01 7.34604597e-01 1.95827022e-01
-7.17822492e-01 3.38332891e-01 -5.33487976e-01 3.34529370e-01
5.37322581e-01 6.00952208e-01 -1.14812243e+00 6.46566689e-01
4.40524250e-01 1.00861311e+00 -4.61017191e-01 1.34841430e+00
-1.09057620e-01 8.32380295e-01 8.49924684e-02 -3.80024582e-01
4.55028802e-01 -1.32369911e-02 9.31176245e-01 1.59301817e+00
4.51417595e-01 -2.00822987e-02 9.41079110e-02 6.43435955e-01
-2.49504700e-01 2.78215080e-01 -6.63937628e-01 -2.10100971e-02
6.09118998e-01 9.29491997e-01 -7.28193820e-01 -9.10450697e-01
1.41236693e-01 9.92241681e-01 1.18207648e-01 5.31018734e-01
-3.55945796e-01 -8.52626622e-01 2.24644825e-01 -1.72943339e-01
2.72631526e-01 2.21522339e-02 -2.97279984e-01 -9.55561161e-01
-1.52074829e-01 -6.07750416e-01 4.80862677e-01 -7.96280563e-01
-1.10388660e+00 6.87670112e-01 5.42752802e-01 -9.01601017e-01
-7.49200225e-01 2.70088226e-01 -1.36345136e+00 5.94730735e-01
-1.31708157e+00 -8.65137517e-01 -1.70704722e-01 3.17604728e-02
1.23697877e+00 -3.35614532e-01 5.91874123e-01 -5.67215502e-01
-4.72099930e-01 5.05723953e-01 5.78285515e-01 -2.23856866e-01
9.78100479e-01 -1.41775596e+00 8.73423338e-01 7.40161419e-01
-1.79558665e-01 7.63695896e-01 9.93189752e-01 -1.00677717e+00
-5.84890485e-01 -1.33011913e+00 1.18029273e+00 1.48525322e-02
3.19746524e-01 3.68468426e-02 -1.10596859e+00 6.58138931e-01
1.09100342e+00 -1.28840923e+00 7.04461694e-01 1.44383520e-01
5.03904969e-02 2.08302408e-01 -7.10280597e-01 8.23883712e-01
6.70471191e-01 -7.78539330e-02 -1.01975727e+00 6.91917062e-01
1.30588198e+00 -2.80985236e-01 -4.30425882e-01 1.73126608e-01
2.37401068e-01 -6.56275451e-01 5.63273847e-01 -7.50519156e-01
8.86810720e-01 6.37736917e-02 3.23332429e-01 -1.99613881e+00
-3.12350661e-01 -7.15538681e-01 -2.79739261e-01 1.57241535e+00
4.68813658e-01 -4.86828536e-01 7.12370515e-01 3.12994391e-01
-5.62227488e-01 -5.91177225e-01 -3.86266440e-01 -6.25234842e-01
2.85186589e-01 3.38715464e-01 6.20668173e-01 7.28716433e-01
4.88793641e-01 7.66812265e-01 -2.97622710e-01 -2.41124466e-01
2.91967154e-01 4.04040098e-01 9.18828547e-01 -1.16086173e+00
1.22444503e-01 -8.07418823e-01 2.46463239e-01 -1.42754436e+00
3.46949279e-01 -7.87113190e-01 2.99457490e-01 -2.40555668e+00
6.81854784e-01 1.92201078e-01 1.33272767e-01 1.78576931e-01
-5.03883481e-01 -2.31731191e-01 3.39442305e-03 4.41050768e-01
-9.56675112e-01 1.00305068e+00 1.22752380e+00 -4.64888364e-01
-5.96225500e-01 -2.88352021e-03 -1.18974233e+00 5.54976165e-01
1.09880853e+00 -4.15498793e-01 -4.68810022e-01 -4.49009120e-01
-3.95150669e-02 -9.62604508e-02 -3.40788245e-01 -8.38569224e-01
5.56059599e-01 -8.39696601e-02 2.33668223e-01 -1.39671230e+00
8.13528057e-03 1.45065531e-01 -5.12505114e-01 3.57555121e-01
-1.17798924e+00 8.22270513e-02 2.03920677e-01 6.83710754e-01
-3.69175524e-01 -3.91653955e-01 4.39390957e-01 -1.06250569e-01
-2.60885924e-01 1.69483304e-01 -7.04268932e-01 1.59633771e-01
4.80938286e-01 -1.99065968e-01 -7.13073075e-01 -1.00902772e+00
-2.15691864e-03 5.66222906e-01 3.91641587e-01 4.77815181e-01
6.66585386e-01 -1.20285738e+00 -1.14942634e+00 -4.20973003e-01
1.37601480e-01 3.27679455e-01 2.61862129e-01 2.08521619e-01
-2.72940040e-01 1.02631688e+00 7.77350143e-02 -5.28646111e-01
-1.21106398e+00 2.63810337e-01 1.80054754e-02 -5.04994214e-01
-6.98132455e-01 6.19189501e-01 5.61130524e-01 -3.85177135e-02
3.15742671e-01 -4.89648134e-01 -7.02201843e-01 3.74468923e-01
9.25961912e-01 6.57678962e-01 -7.61905313e-02 -3.35160822e-01
1.87889159e-01 4.31973666e-01 -9.74259198e-01 -2.66297519e-01
1.27059031e+00 -2.20115975e-01 -1.67504728e-01 4.31735426e-01
1.23373508e+00 -9.08277258e-02 -8.43614578e-01 -2.44656339e-01
6.68903887e-02 -9.59379226e-02 6.82399124e-02 -5.47006845e-01
-6.87201023e-01 5.62315762e-01 -2.34037057e-01 4.79294717e-01
9.73539710e-01 1.47608086e-01 1.25596821e+00 6.52347565e-01
-2.52310753e-01 -1.27647877e+00 3.18665266e-01 7.67629325e-01
1.25708413e+00 -1.03096807e+00 3.95564318e-01 -1.64549157e-01
-7.99381495e-01 1.02173710e+00 5.68032622e-01 -3.72838527e-01
6.96289614e-02 -2.48372301e-01 -2.09633902e-01 -2.76992172e-01
-1.06319737e+00 2.76419193e-01 5.83101630e-01 5.51187873e-01
6.17987096e-01 -1.25382811e-01 -6.63810849e-01 3.87863010e-01
-5.99323630e-01 -3.69369835e-01 1.00956368e+00 7.99715579e-01
-1.02962017e+00 -7.29089379e-01 -2.81832695e-01 9.11883652e-01
-4.78349566e-01 -1.79529533e-01 -7.91418552e-01 5.84778309e-01
-9.42223966e-01 1.10798478e+00 2.55514383e-01 -5.49956635e-02
1.59436792e-01 -9.51937959e-02 5.03653064e-02 -9.26540911e-01
-5.81037521e-01 2.85349071e-01 2.76436687e-01 -6.88529015e-02
-2.37421155e-01 -8.08352649e-01 -1.15490782e+00 -3.32120568e-01
-4.60026801e-01 6.27965212e-01 5.84099889e-01 6.67004645e-01
5.93939960e-01 7.33490527e-01 7.57907629e-01 -1.04487276e+00
-4.66771811e-01 -1.46686554e+00 -4.92495924e-01 1.71779960e-01
5.07850170e-01 -2.30381861e-02 -1.78772449e-01 1.37848601e-01] | [12.481456756591797, 9.491965293884277] |
6372c39c-ca8b-4dc4-8f3b-b90eec206e49 | local-feature-descriptor-learning-with | 1706.05358 | null | http://arxiv.org/abs/1706.05358v1 | http://arxiv.org/pdf/1706.05358v1.pdf | Local Feature Descriptor Learning with Adaptive Siamese Network | Although the recent progress in the deep neural network has led to the
development of learnable local feature descriptors, there is no explicit answer
for estimation of the necessary size of a neural network. Specifically, the
local feature is represented in a low dimensional space, so the neural network
should have more compact structure. The small networks required for local
feature descriptor learning may be sensitive to initial conditions and learning
parameters and more likely to become trapped in local minima. In order to
address the above problem, we introduce an adaptive pruning Siamese
Architecture based on neuron activation to learn local feature descriptors,
making the network more computationally efficient with an improved recognition
rate over more complex networks. Our experiments demonstrate that our learned
local feature descriptors outperform the state-of-art methods in patch
matching. | ['Kwang-Ting', 'Yan-Ying Chen', 'Cheng', 'Chong Huang', 'Qiong Liu'] | 2017-06-16 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 1.58601806e-01 -4.37145770e-01 -3.73849154e-01 -4.75877553e-01
-7.18433678e-01 -1.22163162e-01 2.11992458e-01 4.81235087e-02
-4.90451634e-01 6.15514278e-01 -1.48221359e-01 2.39500701e-01
-7.13540494e-01 -9.86002028e-01 -7.12080300e-01 -8.83131027e-01
-2.51626402e-01 2.57032394e-01 4.72529888e-01 4.05566767e-02
4.33705151e-01 1.11464572e+00 -1.73615456e+00 2.56155342e-01
6.79883659e-01 1.40683985e+00 1.57782361e-01 2.94307441e-01
-2.55768374e-02 5.46908140e-01 -6.14871144e-01 3.82098034e-02
5.96913517e-01 -3.58775735e-01 -7.50938356e-01 -3.18531960e-01
8.80591691e-01 -9.31811854e-02 -2.67476797e-01 1.20804918e+00
4.73354578e-01 3.73727351e-01 6.62670135e-01 -9.27623451e-01
-6.69038355e-01 2.96850801e-01 -2.36773416e-01 3.77573252e-01
1.01098113e-01 -1.32418677e-01 1.09645975e+00 -9.18847680e-01
5.58691740e-01 1.04763448e+00 9.83034849e-01 2.28029191e-01
-1.02337742e+00 -6.43204391e-01 -1.34606734e-02 3.52282941e-01
-1.94051850e+00 -4.25646424e-01 7.41530299e-01 7.53279403e-02
1.22702777e+00 -3.05710118e-02 7.76763618e-01 7.07244098e-01
4.64843452e-01 5.55246592e-01 5.32042205e-01 -5.45303285e-01
2.32829094e-01 -3.26751620e-01 5.43668680e-03 9.88650262e-01
3.43024671e-01 3.68651837e-01 -6.49769604e-01 -3.03963542e-01
1.22133803e+00 4.14387256e-01 8.21364596e-02 -6.04872346e-01
-9.36261535e-01 1.11203396e+00 8.42533886e-01 6.34934604e-01
-3.92138451e-01 2.79720068e-01 3.57976556e-01 5.85612178e-01
2.71080762e-01 7.06115246e-01 -4.30608869e-01 -2.01566190e-01
-1.02357817e+00 2.70865083e-01 8.29404771e-01 6.11829162e-01
1.11839569e+00 6.24094754e-02 8.52575377e-02 1.21852696e+00
1.04739845e-01 3.69246095e-01 5.14682174e-01 -1.06853461e+00
3.63578498e-01 8.23182046e-01 -4.47740138e-01 -1.52399480e+00
-3.75056475e-01 -4.17584181e-01 -9.36620891e-01 2.88256288e-01
3.79416883e-01 1.64929539e-01 -8.94413590e-01 1.39281774e+00
6.43617138e-02 1.24457315e-01 -3.04602593e-01 9.20587361e-01
4.41029370e-01 5.83636820e-01 -1.96346208e-01 2.78382510e-01
7.29675472e-01 -1.01597571e+00 -3.25246125e-01 -2.35752583e-01
6.42775774e-01 -6.15383267e-01 6.39197350e-01 1.25308320e-01
-9.80894804e-01 -7.30605125e-01 -1.32562780e+00 -1.87186673e-02
-4.88233536e-01 -2.84426123e-01 7.56009161e-01 3.70844960e-01
-9.79150534e-01 9.84644294e-01 -8.72851908e-01 -3.21901768e-01
4.76121306e-01 7.59806693e-01 -3.87109339e-01 -7.42568448e-02
-1.09283710e+00 8.74251246e-01 4.62013006e-01 2.74060875e-01
-5.00976920e-01 -5.25147557e-01 -9.26454425e-01 3.20957959e-01
2.44908612e-02 -5.46138346e-01 7.51644254e-01 -1.13088572e+00
-1.47276759e+00 6.01708293e-01 -6.66003600e-02 -4.99339074e-01
5.18885022e-03 -3.09166568e-03 -1.89881399e-01 2.39891320e-01
-2.84621388e-01 8.80863547e-01 9.65990245e-01 -6.14115238e-01
-6.09225214e-01 -2.11013794e-01 -1.85407892e-01 1.00772336e-01
-4.97473210e-01 2.60249078e-01 -1.35674223e-01 -7.30888546e-01
3.73292476e-01 -8.95976663e-01 -2.04652250e-01 4.72964108e-01
2.69641846e-01 -5.30257404e-01 7.76113033e-01 -8.22390988e-02
1.11987484e+00 -2.15289617e+00 -1.42892078e-01 7.34572351e-01
8.04032981e-02 3.33339006e-01 -5.08720577e-01 2.47399777e-01
-1.26876026e-01 3.17183286e-02 -1.47314638e-01 9.38458815e-02
-1.85125470e-01 3.35482419e-01 -9.52268168e-02 6.56468987e-01
1.33879215e-01 7.89307117e-01 -6.13989353e-01 -7.10999608e-01
3.47477309e-02 5.01353562e-01 -6.00431323e-01 -1.30507380e-01
4.04432356e-01 -4.20378089e-01 -7.71732926e-01 8.42133462e-01
7.07522452e-01 -1.14658564e-01 -2.25814044e-01 -4.66357842e-02
-6.42809793e-02 1.54533267e-01 -1.42731595e+00 1.69345748e+00
-2.98537731e-01 7.30119407e-01 1.49429422e-02 -1.40307534e+00
1.35883164e+00 9.32000130e-02 6.08512521e-01 -8.28664660e-01
2.31817558e-01 4.21072304e-01 -5.62369525e-02 -8.06169808e-02
2.10455775e-01 -9.02734101e-02 1.49586126e-01 3.06556940e-01
1.92864493e-01 1.59958959e-01 1.01923600e-01 -3.85392785e-01
1.08415735e+00 -2.50672340e-01 2.17645377e-01 -3.81862104e-01
5.35392880e-01 -1.63855180e-01 7.80867398e-01 9.73371148e-01
-2.39428952e-01 6.94519639e-01 1.19360633e-01 -9.96997118e-01
-8.66430342e-01 -8.37768435e-01 -5.00267982e-01 9.98609245e-01
1.70302823e-01 -3.25005531e-01 -6.11222684e-01 -5.31490326e-01
1.95554689e-01 -2.77946323e-01 -4.82401341e-01 -5.15106142e-01
-9.76398349e-01 -3.03156197e-01 6.06794059e-01 6.63504422e-01
5.93795836e-01 -1.33580649e+00 -6.45470262e-01 2.87964046e-01
3.05824667e-01 -7.46597826e-01 -5.62225640e-01 5.39137185e-01
-1.36112607e+00 -9.82557714e-01 -7.93914974e-01 -1.25014913e+00
9.97761667e-01 1.11674167e-01 9.03889954e-01 4.21067983e-01
-6.15184724e-01 5.00239246e-02 -2.24718004e-01 -4.52999733e-02
-5.93088195e-02 4.65275973e-01 5.14404774e-02 -2.42381677e-01
6.36300862e-01 -5.64841986e-01 -7.31705010e-01 4.46146935e-01
-8.12716544e-01 -5.60507774e-01 9.09764528e-01 1.19721913e+00
8.08376431e-01 6.36725798e-02 3.57327670e-01 -4.71572876e-01
7.79812753e-01 -1.43679991e-01 -6.25908256e-01 2.32046694e-01
-6.34924948e-01 5.36926866e-01 8.87106121e-01 -6.66081667e-01
-5.59833288e-01 3.15916330e-01 -5.12920320e-02 -3.82076800e-01
-1.62287176e-01 6.04042232e-01 2.87158847e-01 -8.83926690e-01
8.96830022e-01 1.71074033e-01 -2.51425570e-03 -4.97752458e-01
-6.68200999e-02 4.08367634e-01 8.75814632e-02 -7.39026010e-01
8.39493871e-01 2.00062573e-01 2.86536187e-01 -8.03270400e-01
-4.96474117e-01 -5.10829568e-01 -9.32359099e-01 2.44658872e-01
4.57003832e-01 -5.13604045e-01 -5.79874337e-01 4.02289540e-01
-9.53946352e-01 -4.28904556e-02 -4.46781278e-01 3.14882815e-01
-4.88805622e-01 3.51061910e-01 -4.41598624e-01 -2.56305188e-01
-3.90546858e-01 -1.04264569e+00 7.02407122e-01 3.92385393e-01
-2.47220412e-01 -1.13283169e+00 2.85982519e-01 -2.04126105e-01
8.46016884e-01 7.00111836e-02 7.63965189e-01 -5.44680655e-01
-6.54518306e-01 -7.20158398e-01 -3.45983922e-01 5.16679466e-01
7.60078430e-02 9.43522975e-02 -5.58084846e-01 -4.22867924e-01
-1.81970820e-01 -4.46104497e-01 8.37938309e-01 4.68484521e-01
1.27848661e+00 -2.94513285e-01 -3.84639263e-01 1.04210770e+00
1.45543981e+00 8.96959603e-02 5.08985817e-01 4.98036355e-01
2.78849125e-01 3.07870299e-01 2.82678336e-01 3.67247194e-01
-1.12616001e-02 6.30561769e-01 6.27061278e-02 -2.08642080e-01
-1.61430582e-01 -1.64863482e-01 5.68417944e-02 8.81380558e-01
-1.86630175e-01 3.15983981e-01 -8.39549065e-01 5.81645846e-01
-1.85416472e+00 -1.01098502e+00 5.78990281e-01 2.24311423e+00
7.39168882e-01 3.41045484e-02 -1.20064430e-01 -5.30330688e-02
5.74209869e-01 3.76470566e-01 -5.57246864e-01 -6.34718180e-01
-2.65134871e-01 5.36390126e-01 4.95359868e-01 2.98295528e-01
-1.07407558e+00 9.09759521e-01 7.73651981e+00 9.15927887e-01
-1.24325240e+00 -2.84210980e-01 2.92781353e-01 5.72487935e-02
1.48968041e-01 -7.88019970e-03 -9.02495027e-01 1.70643523e-01
8.33466947e-01 2.29522344e-02 4.43828970e-01 9.77834940e-01
-2.95821577e-01 -2.52606056e-04 -1.05621982e+00 1.30907142e+00
6.55056760e-02 -1.44765961e+00 2.83480167e-01 -1.29912572e-03
8.76968205e-01 2.34699130e-01 -7.63078108e-02 2.52985775e-01
-1.33865878e-01 -1.15344381e+00 3.07952523e-01 7.46399641e-01
3.93426239e-01 -1.09699106e+00 7.72399783e-01 2.46825576e-01
-1.41824841e+00 -2.99029827e-01 -1.15388548e+00 1.31773027e-02
-3.61612469e-01 4.03434247e-01 -3.71160626e-01 -3.84102352e-02
7.62679875e-01 6.19112074e-01 -6.91092193e-01 1.58019853e+00
1.79925635e-01 1.89591587e-01 -8.21271956e-01 -4.00298417e-01
4.78920043e-01 -1.81666717e-01 2.11386785e-01 1.12198591e+00
3.34367692e-01 -1.06282331e-01 3.23310405e-01 8.94455254e-01
-1.74633875e-01 1.83642238e-01 -7.30341911e-01 8.38223938e-03
5.43677688e-01 1.10091841e+00 -6.13589644e-01 8.87287632e-02
-4.66424167e-01 8.96924198e-01 6.70987606e-01 2.95400620e-01
-3.73064637e-01 -1.12721944e+00 6.84848726e-01 2.99913995e-02
4.91145462e-01 -4.23977077e-01 -2.01749161e-01 -8.71285260e-01
1.64476842e-01 -8.13206911e-01 4.40557599e-01 -1.77146435e-01
-1.30994654e+00 6.16688430e-01 -2.28231892e-01 -1.09658396e+00
-3.48901898e-01 -8.09847176e-01 -7.17078209e-01 7.82935739e-01
-1.43592584e+00 -8.00703228e-01 -2.38600239e-01 8.19767833e-01
1.91595331e-01 -3.10458153e-01 9.15756106e-01 3.71328175e-01
-2.36024395e-01 1.02492046e+00 3.55254531e-01 2.90921092e-01
7.54068017e-01 -8.29819798e-01 1.06807508e-01 5.66383719e-01
3.39050919e-01 9.09382939e-01 1.84576914e-01 -3.03448945e-01
-1.26277125e+00 -5.62747061e-01 1.00499177e+00 -5.37803993e-02
5.91066241e-01 -2.59707093e-01 -9.91717160e-01 2.16647908e-01
-2.22350881e-01 5.94450951e-01 4.89442021e-01 3.34602654e-01
-3.56819242e-01 -6.76975846e-01 -1.19773114e+00 2.22099304e-01
1.11690009e+00 -8.53144646e-01 -5.97629726e-01 2.33717278e-01
1.68291152e-01 -2.06711337e-01 -1.05522919e+00 3.78895372e-01
7.70436585e-01 -7.25959539e-01 1.22522211e+00 -4.75871176e-01
-3.27845626e-02 -1.15259610e-01 -1.74549952e-01 -9.29938197e-01
-6.95295930e-01 -3.58688414e-01 1.31548122e-02 7.72176027e-01
5.51705182e-01 -5.67577004e-01 1.10816944e+00 7.12922573e-01
3.23609449e-02 -1.06950557e+00 -1.34160244e+00 -1.07488692e+00
1.51399165e-01 -1.28033618e-02 4.78628129e-01 7.99685895e-01
-7.25405142e-02 -1.28773883e-01 -1.74446404e-01 -2.13879254e-02
4.76712137e-01 4.14295465e-01 3.05211872e-01 -1.47094381e+00
1.05090536e-01 -7.70623624e-01 -9.01167929e-01 -1.29703951e+00
2.87860841e-01 -9.05304968e-01 7.88270682e-02 -1.29465032e+00
2.31213942e-01 -5.99737585e-01 -7.35248923e-01 7.88849652e-01
1.53946757e-01 3.24778140e-01 8.99550319e-03 2.35643268e-01
-7.81218469e-01 6.37221813e-01 1.09950805e+00 -2.05614448e-01
-2.08164290e-01 4.10044892e-03 -3.45713317e-01 6.02916658e-01
6.18590891e-01 -8.80060852e-01 -5.58630675e-02 -4.59629148e-01
2.48036459e-01 -3.67661268e-01 1.57478377e-01 -1.25507390e+00
8.54207933e-01 -1.62918583e-01 6.63567662e-01 -3.76729518e-01
2.92718291e-01 -9.77593541e-01 -3.71521652e-01 6.20294631e-01
-3.12604636e-01 2.44273871e-01 1.20425932e-02 2.72606194e-01
-7.95912564e-01 -6.74196422e-01 9.66395378e-01 -6.29623085e-02
-6.42941236e-01 5.76928735e-01 -3.45327079e-01 -1.38588578e-01
7.89130688e-01 -8.67618740e-01 -7.68863633e-02 -5.04106805e-02
-4.10372555e-01 -2.90399063e-02 3.90417516e-01 4.73679483e-01
9.25895631e-01 -1.61605740e+00 -4.00591940e-01 5.88996768e-01
-1.21056531e-02 -8.71595368e-02 -6.16118871e-02 3.77119303e-01
-8.79844010e-01 5.75337946e-01 -5.77815711e-01 -4.82568920e-01
-1.03827739e+00 3.53619337e-01 6.48516715e-01 -5.15234470e-02
-7.00528204e-01 9.16413069e-01 -1.41072065e-01 -5.31478465e-01
5.44529855e-01 -4.41551864e-01 5.57254031e-02 -1.55225769e-01
4.54303563e-01 5.65609872e-01 6.67440444e-02 -4.00349289e-01
-5.33815145e-01 1.17515886e+00 -1.98971093e-01 5.11193812e-01
1.56860101e+00 2.54527032e-01 -2.34159783e-01 -1.21925727e-01
1.73019791e+00 -4.27321285e-01 -1.17489684e+00 -4.17234480e-01
2.35051319e-01 -5.00335097e-01 3.60404968e-01 -3.12089413e-01
-1.28738618e+00 6.87249005e-01 9.75901961e-01 1.69489114e-03
9.77611840e-01 -1.95871860e-01 9.24441993e-01 1.31684017e+00
3.69836986e-01 -1.50543272e+00 5.58051020e-02 9.89943981e-01
8.98158967e-01 -1.17165065e+00 2.41611674e-01 7.38044530e-02
-1.02699004e-01 1.63748288e+00 4.82485265e-01 -7.67625868e-01
9.62422907e-01 2.54505992e-01 9.86103266e-02 -3.08300883e-01
-3.73270780e-01 1.75090760e-01 6.52217388e-01 5.00293791e-01
2.91153252e-01 -3.69511753e-01 -9.45845097e-02 1.46218002e-01
1.06793426e-01 -1.81490853e-01 -2.48958215e-01 9.44366634e-01
-7.68755317e-01 -1.21713114e+00 -2.67577678e-01 4.26065534e-01
-3.60958070e-01 -4.47345600e-02 -3.17357510e-01 6.39302373e-01
1.34480923e-01 4.48978812e-01 1.51808515e-01 -3.96631211e-02
3.43425483e-01 4.12473502e-03 6.56922877e-01 -3.42360258e-01
-6.84622467e-01 -1.64308429e-01 -2.83760875e-01 -8.80202949e-01
-4.65648323e-01 -4.48944330e-01 -1.21668613e+00 -3.34473789e-01
-6.15356743e-01 8.17653388e-02 5.11690855e-01 9.45116222e-01
4.28447336e-01 1.68125108e-01 6.48425341e-01 -9.79506969e-01
-7.01632321e-01 -6.84335828e-01 -7.09304512e-01 2.90201366e-01
3.48035932e-01 -6.26105547e-01 -3.90069097e-01 -3.90331388e-01] | [10.18256664276123, 0.013409408740699291] |
27c0c95f-90fe-4f94-9343-f4a4152f962f | hue-modification-localization-by-pair | 1903.01735 | null | http://arxiv.org/abs/1903.01735v1 | http://arxiv.org/pdf/1903.01735v1.pdf | Hue Modification Localization By Pair Matching | Hue modification is the adjustment of hue property on color images.
Conducting hue modification on an image is trivial, and it can be abused to
falsify opinions of viewers. Since shapes, edges or textural information
remains unchanged after hue modification, this type of manipulation is
relatively hard to be detected and localized. Since small patches inherit the
same Color Filter Array (CFA) configuration and demosaicing, any distortion
made by local hue modification can be detected by patch matching within the
same image. In this paper, we propose to localize hue modification by means of
a Siamese neural network specifically designed for matching two inputs. By
crafting the network outputs, we are able to form a heatmap which potentially
highlights malicious regions. Our proposed method deals well not only with
uncompressed images but also with the presence of JPEG compression, an
operation usually hindering the exploitation of CFA and demosaicing artifacts.
Experimental evidences corroborate the effectiveness of the proposed method. | ['Giulia Boato', 'Quoc-Tin Phan', 'Michele Vascotto'] | 2019-03-05 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 8.23996186e-01 1.32844709e-02 2.30939299e-01 7.83679634e-02
-2.72057876e-02 -8.02430868e-01 3.69439602e-01 3.22362095e-01
-3.53902757e-01 5.43663383e-01 -3.76449198e-01 -3.51769567e-01
9.27729160e-02 -1.14576602e+00 -8.36418152e-01 -9.72811282e-01
-2.00938240e-01 -3.17421794e-01 2.98726499e-01 -1.46437734e-01
6.76244855e-01 7.48877764e-01 -1.31346607e+00 3.28851253e-01
7.14851141e-01 1.10777640e+00 -8.90162289e-02 6.46380424e-01
2.63804048e-01 5.66142976e-01 -8.16910028e-01 -5.54768741e-01
6.72304273e-01 -4.80336696e-01 -1.26937196e-01 2.63552368e-01
7.26893306e-01 -4.89695251e-01 -2.00651243e-01 1.76142871e+00
3.11822176e-01 -6.57957047e-02 5.92483103e-01 -1.21975791e+00
-8.50648224e-01 3.49554777e-01 -7.80144513e-01 9.01622400e-02
2.98059523e-01 1.73038527e-01 5.31873703e-01 -7.09344149e-01
7.44283378e-01 7.28177071e-01 4.54984128e-01 -1.08810710e-02
-1.11684203e+00 -3.98472071e-01 -3.18899870e-01 3.13676119e-01
-1.52077317e+00 -4.56695519e-02 1.41069603e+00 -5.81397377e-02
2.20295057e-01 6.38703465e-01 9.07501400e-01 8.54855657e-01
4.72642183e-01 5.26399434e-01 1.50663519e+00 -5.08528173e-01
3.19194019e-01 3.82128298e-01 -5.40751100e-01 7.31910050e-01
5.14079154e-01 1.08793415e-01 -3.29094917e-01 2.65837647e-02
7.53395259e-01 -1.61306206e-02 -6.99310124e-01 -7.47853696e-01
-9.98698592e-01 5.24930120e-01 7.86137998e-01 2.37137169e-01
-3.98873359e-01 -8.54629055e-02 2.47299939e-01 3.19993705e-01
1.75684560e-02 5.29449880e-01 1.34376585e-01 2.51772761e-01
-1.15497577e+00 -2.44403675e-01 5.32866657e-01 5.88661671e-01
7.98141718e-01 1.44298747e-01 9.83296782e-02 3.70872110e-01
2.02537272e-02 7.56850243e-01 2.59479642e-01 -5.89516163e-01
2.09677666e-01 5.29380381e-01 -3.49514559e-02 -1.81389987e+00
-1.77575380e-01 -1.91941246e-01 -1.07502174e+00 9.23376262e-01
5.95198631e-01 1.06377207e-01 -8.59601200e-01 1.33261395e+00
3.09922338e-01 -5.39581291e-02 -5.85727841e-02 1.06131577e+00
2.75078684e-01 6.38163626e-01 -1.83277562e-01 -1.96354225e-01
1.19546318e+00 -8.76226842e-01 -7.37944484e-01 4.37240638e-02
-5.18652909e-02 -9.04568076e-01 9.51641917e-01 8.90772164e-01
-1.04042149e+00 -4.90570605e-01 -1.54608548e+00 -6.17419779e-02
-8.01455796e-01 -1.34517655e-01 1.98536515e-01 9.80308235e-01
-7.75913537e-01 6.87776089e-01 -4.00000393e-01 -1.53436720e-01
1.86534062e-01 1.65565565e-01 -4.18536931e-01 -3.20604779e-02
-1.03553760e+00 7.30107486e-01 4.65295374e-01 3.94064844e-01
-4.33620900e-01 -4.17163193e-01 -5.48478067e-01 2.53566951e-01
2.88225144e-01 -3.18447575e-02 3.13261807e-01 -1.33284485e+00
-1.21436095e+00 7.75817394e-01 3.89206290e-01 -3.94596994e-01
9.45062995e-01 7.36147761e-02 -7.87425220e-01 4.61534560e-01
-2.86162049e-01 3.33293796e-01 1.70617461e+00 -1.42648673e+00
-5.57206452e-01 -2.35039577e-01 -1.54108733e-01 -3.00233774e-02
-5.49919009e-01 -4.31304395e-01 -3.64327371e-01 -8.55920076e-01
1.83578998e-01 -5.80388904e-01 1.58743843e-01 4.51877356e-01
-6.34514570e-01 5.67075551e-01 1.03690851e+00 -9.60100770e-01
1.30964625e+00 -2.30621243e+00 -2.78221667e-01 1.09596682e+00
1.05346151e-01 3.68017226e-01 -5.43736666e-02 2.86949635e-01
-2.58534163e-01 7.04934746e-02 -5.56152761e-01 3.95824522e-01
-9.45696086e-02 -1.33651838e-01 -2.31554285e-01 9.11535025e-01
2.08785996e-01 5.69665909e-01 -4.87789005e-01 -3.56561959e-01
2.85445303e-01 5.69291949e-01 -2.73455977e-01 -2.68881619e-02
4.06728126e-02 7.66740292e-02 -8.22541267e-02 8.53493154e-01
1.26407290e+00 2.11152270e-01 3.90293270e-01 -6.86659694e-01
-8.89188796e-02 -4.34841037e-01 -1.33633232e+00 1.09958363e+00
-1.11521296e-01 7.83577681e-01 -1.80323087e-02 -9.08405900e-01
1.12806714e+00 -1.33864358e-01 1.72181666e-01 -1.09797955e+00
2.94988185e-01 1.91451460e-01 -3.14156830e-01 -5.76870143e-01
6.90075099e-01 1.25976443e-01 -2.68835071e-02 4.33673501e-01
-4.71803457e-01 8.40327330e-03 4.76694405e-02 -1.05846487e-01
7.34509826e-01 -1.07297994e-01 4.70491081e-01 -1.89751700e-01
5.69597244e-01 -2.29123622e-01 4.31822650e-02 6.91231430e-01
-2.54255354e-01 5.04209220e-01 5.22873163e-01 -4.32470977e-01
-1.23817778e+00 -1.10003746e+00 -1.11007251e-01 8.20688963e-01
6.72625840e-01 1.25400469e-01 -7.77903736e-01 -7.19467938e-01
7.49714375e-02 4.93262738e-01 -8.18922520e-01 -3.43502074e-01
-5.61676979e-01 -4.39854115e-01 5.84887028e-01 -4.37912624e-03
7.89427280e-01 -1.01284266e+00 -1.08933759e+00 -2.64465739e-03
1.31270751e-01 -7.32558131e-01 -2.95239031e-01 2.92836010e-01
-5.53655505e-01 -1.25945663e+00 -8.86444867e-01 -7.59342730e-01
9.22475934e-01 3.83336425e-01 7.69494772e-01 2.75516361e-01
-3.61447126e-01 1.65184617e-01 -4.02711600e-01 -4.21830546e-03
-6.02697313e-01 -2.98466682e-01 -2.85261482e-01 4.41248775e-01
2.39610657e-01 -5.76329768e-01 -7.87254989e-01 1.79569840e-01
-1.39130878e+00 -2.43913263e-01 6.12026989e-01 6.93971694e-01
3.02450478e-01 6.84109151e-01 -8.69318098e-02 -8.38083804e-01
6.70085728e-01 8.59051384e-03 -6.92166030e-01 6.20765984e-01
-8.40815306e-01 -1.15469493e-01 7.16351449e-01 -3.48507166e-01
-1.06432080e+00 -9.17033572e-03 2.32758433e-01 -3.17940414e-01
-3.73157769e-01 2.00649932e-01 -2.21406490e-01 -7.73055255e-01
7.08458006e-01 2.61518598e-01 9.43452269e-02 -2.84075826e-01
4.43253815e-01 4.61307019e-01 8.47582161e-01 5.80013357e-02
1.35811424e+00 6.43296361e-01 2.31596857e-01 -7.94506848e-01
8.02535638e-02 5.26092425e-02 -5.16376555e-01 -5.32914817e-01
7.22426176e-01 -2.87021667e-01 -5.74921250e-01 4.67567235e-01
-7.75921702e-01 5.71354181e-02 -1.00279525e-01 -7.08614662e-02
-6.18573576e-02 9.36221302e-01 -4.44703788e-01 -6.22153103e-01
-2.07698941e-01 -8.78678024e-01 4.32613879e-01 3.47701073e-01
1.68731868e-01 -8.31506550e-01 -2.67491132e-01 1.33748837e-02
4.71362859e-01 5.21188557e-01 1.19726801e+00 -2.80562043e-01
-6.49714947e-01 -3.39853972e-01 -3.05292726e-01 3.33508134e-01
1.48413390e-01 3.69229376e-01 -8.99537086e-01 -3.76355678e-01
4.17168103e-02 -4.37077181e-03 7.96039879e-01 1.26819566e-01
9.99351680e-01 -4.11471456e-01 9.81119499e-02 7.85732627e-01
1.69713652e+00 4.47479576e-01 1.09261787e+00 8.70334506e-01
4.77715969e-01 4.86879081e-01 5.47099710e-01 4.38280582e-01
-5.13603330e-01 4.40650910e-01 8.25364709e-01 -5.64663231e-01
-1.09252572e-01 -2.08102435e-01 2.47378111e-01 2.66067445e-01
-1.22137815e-01 -2.29048043e-01 -4.65866566e-01 2.04565153e-01
-1.15090001e+00 -8.70117724e-01 -8.76151249e-02 2.20367908e+00
5.60503006e-01 2.20170587e-01 -1.51645929e-01 4.32752401e-01
9.27271128e-01 2.15232521e-01 -5.20177126e-01 -5.81699848e-01
-4.21558648e-01 2.93283165e-01 7.81805933e-01 2.77053982e-01
-1.22412348e+00 5.44610500e-01 5.41098309e+00 9.94714320e-01
-1.59447861e+00 -3.98593545e-01 5.13275206e-01 2.35152960e-01
-3.88711363e-01 -2.25971058e-01 6.13971688e-02 5.13097525e-01
3.11492980e-02 1.34455219e-01 7.18518615e-01 7.44148254e-01
-1.62187312e-02 -4.99913335e-01 -5.40374935e-01 8.95717502e-01
4.27218139e-01 -9.82543468e-01 1.29234314e-01 -1.01892650e-01
6.11003280e-01 -7.57263243e-01 7.43634224e-01 -3.85546744e-01
-3.22473794e-01 -8.14278126e-01 7.56199777e-01 3.18304837e-01
7.77141869e-01 -9.25838351e-01 6.59423590e-01 -2.71655560e-01
-9.22112226e-01 -1.41338274e-01 -4.59148407e-01 3.27455461e-01
1.00615375e-01 4.79475081e-01 -8.49409938e-01 3.66218567e-01
6.76389158e-01 2.97136962e-01 -9.59522188e-01 1.29972553e+00
-4.16476548e-01 2.53858715e-01 -3.03958118e-01 1.10833198e-01
1.28628418e-01 -5.20344317e-01 6.33467436e-01 1.21305490e+00
6.10859454e-01 -4.20996368e-01 -3.62552434e-01 9.33281183e-01
-4.23893481e-02 2.72246212e-01 -6.72264993e-01 -8.76185894e-02
2.62354076e-01 1.31495905e+00 -1.16955698e+00 -2.41999596e-01
-2.26486683e-01 1.46953714e+00 -5.57241999e-02 6.02892697e-01
-8.80344689e-01 -6.27205133e-01 1.06333889e-01 -1.06041312e-01
6.81851327e-01 -4.00666036e-02 -3.69520247e-01 -8.70962560e-01
3.27245072e-02 -1.03161871e+00 2.53034711e-01 -8.64255130e-01
-1.03077888e+00 4.19185162e-01 -3.82979274e-01 -1.54387641e+00
1.88732952e-01 -6.45671427e-01 -5.77003300e-01 3.60467017e-01
-1.44967902e+00 -9.80382800e-01 -4.65862393e-01 8.70671213e-01
2.42270213e-02 -1.54332332e-02 5.40027797e-01 2.78747052e-01
-3.98000658e-01 7.62102664e-01 2.58573920e-01 1.08330540e-01
9.03084338e-01 -1.04822063e+00 -1.56413674e-01 1.27897489e+00
1.25932217e-01 4.80506122e-01 1.00870931e+00 -7.06380069e-01
-1.32840276e+00 -8.38107705e-01 3.70794952e-01 3.65972221e-01
5.09258389e-01 -2.15522230e-01 -9.81844306e-01 2.14452475e-01
5.68733513e-01 -3.78811270e-01 4.16421086e-01 -5.93144894e-01
-6.46609426e-01 -3.47372711e-01 -1.43345678e+00 6.14350915e-01
3.95957708e-01 -6.72207713e-01 -2.29225591e-01 8.88208821e-02
7.41688088e-02 -5.90655468e-02 -8.15529644e-01 1.97093025e-01
7.46787548e-01 -1.38548517e+00 9.45912182e-01 -1.34748086e-01
5.38317263e-01 -5.17122090e-01 -1.26782641e-01 -1.11666179e+00
-2.60432720e-01 -3.96464527e-01 1.82321221e-01 1.06430316e+00
3.56626540e-01 -4.02139008e-01 6.80052757e-01 3.46115887e-01
3.29909176e-01 -1.68018341e-02 -8.18692446e-01 -5.59381485e-01
-2.82278627e-01 1.28371147e-02 5.05180776e-01 1.16657364e+00
-1.12756640e-01 -4.39596355e-01 -7.41556704e-01 5.56267023e-01
8.70180011e-01 2.46647537e-01 6.34929359e-01 -6.11106575e-01
-4.63419229e-01 -5.04074931e-01 -5.47722518e-01 -4.96219665e-01
-4.73786831e-01 -4.60156113e-01 -5.18136844e-02 -7.49840796e-01
-1.15681738e-01 -1.78382993e-01 -3.99498492e-01 4.59192008e-01
4.86010313e-02 1.01690340e+00 2.54080385e-01 2.79280078e-02
-8.66983831e-03 2.55075574e-01 1.22569704e+00 -4.72197950e-01
-2.65514720e-02 -3.55545282e-01 -3.17226529e-01 6.49862647e-01
9.12475765e-01 -3.49767655e-01 -2.61345029e-01 -2.08757147e-02
4.22048897e-01 -1.98863819e-01 4.89278346e-01 -1.19703710e+00
2.44249091e-01 -6.67536398e-03 7.67441511e-01 -3.71328413e-01
9.88694802e-02 -1.44103205e+00 3.12503546e-01 6.90409660e-01
-2.16255248e-01 4.76348260e-03 -9.91801266e-03 6.31437242e-01
-2.82765090e-01 -4.73561168e-01 1.01603556e+00 -6.50713444e-02
-1.05040562e+00 -2.72180229e-01 -5.79555094e-01 -5.62030494e-01
1.28037763e+00 -7.00192094e-01 -5.48160315e-01 -5.01449406e-01
-5.34405172e-01 -5.89178622e-01 7.90275753e-01 -2.62272805e-02
7.57294357e-01 -1.16219687e+00 -1.13877349e-01 8.66217792e-01
7.04072863e-02 -6.17968857e-01 4.51361179e-01 7.91743815e-01
-1.08070159e+00 -2.26672649e-01 -8.60700130e-01 -1.69076264e-01
-1.47699547e+00 1.14999557e+00 1.43844157e-01 2.00315654e-01
-6.92460835e-01 5.73744535e-01 -3.55013907e-02 2.30189428e-01
2.03599945e-01 -2.66512573e-01 -2.62977511e-01 1.61519274e-01
4.73952860e-01 3.73818696e-01 1.74049456e-02 -3.92313570e-01
-7.36057013e-02 6.72534645e-01 2.70123333e-02 1.54198483e-01
1.02926028e+00 -3.42375726e-01 -2.63771385e-01 -5.42842746e-02
1.28290105e+00 5.76152503e-01 -1.11616910e+00 1.53710291e-01
-2.91304201e-01 -9.00857031e-01 -2.93747522e-02 -8.90052021e-01
-1.40870225e+00 7.85739243e-01 9.57059622e-01 8.13687503e-01
1.52333939e+00 -6.98169649e-01 5.98836303e-01 1.68037176e-01
1.91201285e-01 -1.25796568e+00 7.48240501e-02 -2.96375006e-01
8.06083262e-01 -9.51727033e-01 6.70526028e-02 -3.85628760e-01
-4.81127322e-01 1.45458198e+00 3.06186855e-01 -3.05178344e-01
3.97561699e-01 2.29435220e-01 3.01015884e-01 -3.98968011e-02
1.02312796e-01 1.68022975e-01 2.93306351e-01 7.39809155e-01
-2.04171836e-01 2.31039152e-03 -3.46585780e-01 -2.05332801e-01
-6.31053373e-02 -2.60296732e-01 8.29781592e-01 9.48062122e-01
-5.63876390e-01 -7.44766057e-01 -9.24163401e-01 7.28978738e-02
-1.86059579e-01 -1.17545396e-01 -5.02701104e-01 9.84245062e-01
3.64596337e-01 6.87826574e-01 5.96216395e-02 -4.16685551e-01
1.48475766e-01 -6.85103685e-02 4.05052811e-01 4.12716925e-01
-4.30838913e-01 3.32698226e-01 -1.86399326e-01 -6.67955220e-01
-3.23628306e-01 -2.12312087e-01 -6.94635451e-01 -4.39973384e-01
-6.90672081e-03 -1.65155277e-01 6.78284347e-01 4.68042433e-01
-1.12874575e-01 4.25352693e-01 7.41026163e-01 -6.30602717e-01
-3.22956115e-01 -4.75838423e-01 -1.18169272e+00 8.82392645e-01
4.51591641e-01 -4.89064485e-01 -6.07993186e-01 2.84479231e-01] | [11.17641544342041, -1.417871117591858] |
fb6f5416-10d3-4598-ab49-153eefd7926c | interpretable-propaganda-detection-in-news | 2108.12802 | null | https://arxiv.org/abs/2108.12802v1 | https://arxiv.org/pdf/2108.12802v1.pdf | Interpretable Propaganda Detection in News Articles | Online users today are exposed to misleading and propagandistic news articles and media posts on a daily basis. To counter thus, a number of approaches have been designed aiming to achieve a healthier and safer online news and media consumption. Automatic systems are able to support humans in detecting such content; yet, a major impediment to their broad adoption is that besides being accurate, the decisions of such systems need also to be interpretable in order to be trusted and widely adopted by users. Since misleading and propagandistic content influences readers through the use of a number of deception techniques, we propose to detect and to show the use of such techniques as a way to offer interpretability. In particular, we define qualitatively descriptive features and we analyze their suitability for detecting deception techniques. We further show that our interpretable features can be easily combined with pre-trained language models, yielding state-of-the-art results. | ['Preslav Nakov', 'James Glass', 'Mitra Mohtarami', 'Giovanni Da San Martino', 'Seunghak Yu'] | 2021-08-29 | null | https://aclanthology.org/2021.ranlp-1.179 | https://aclanthology.org/2021.ranlp-1.179.pdf | ranlp-2021-9 | ['propaganda-detection'] | ['natural-language-processing'] | [ 1.46539941e-01 3.17946881e-01 -1.79901302e-01 -3.87535930e-01
-5.08725584e-01 -7.11807847e-01 1.18825471e+00 7.23247588e-01
-3.34830225e-01 5.72387636e-01 4.28533584e-01 -4.88129079e-01
2.31031805e-01 -6.34020030e-01 -4.54065055e-01 -1.25797808e-01
1.45767167e-01 2.15382770e-01 1.16686195e-01 -4.86120939e-01
6.63572550e-01 4.91947860e-01 -1.27956271e+00 4.63703930e-01
9.28421378e-01 8.27974081e-01 -5.15839100e-01 6.96760774e-01
-8.08616057e-02 1.09020901e+00 -8.33114982e-01 -9.25873637e-01
-2.25292549e-01 -5.47630787e-01 -9.00307119e-01 1.22350529e-01
1.52606800e-01 -6.36368513e-01 5.87764643e-02 1.28088164e+00
4.12907600e-02 -1.07997820e-01 8.52407157e-01 -8.66002619e-01
-8.22884262e-01 5.82935929e-01 -1.75168350e-01 1.22861311e-01
9.16775584e-01 -1.03191815e-01 7.70738900e-01 -4.04235750e-01
5.15424967e-01 1.32501304e+00 3.35396945e-01 4.90552932e-01
-9.58108664e-01 -3.57127458e-01 -3.19023430e-02 1.49249554e-01
-8.58973861e-01 -5.43508530e-01 8.34565580e-01 -5.45736730e-01
3.37501854e-01 6.78434610e-01 6.42377853e-01 1.68060398e+00
3.57931167e-01 7.08422422e-01 1.18313742e+00 -6.28939748e-01
3.20566148e-01 8.24081779e-01 2.64647007e-01 7.86355436e-01
5.55164397e-01 -2.63064355e-01 -4.82012689e-01 -5.70741296e-01
1.38276055e-01 -8.07636902e-02 -3.30744922e-01 7.50198439e-02
-7.34195411e-01 1.30934572e+00 3.25706482e-01 7.44154394e-01
-2.37868682e-01 -3.19602378e-02 5.87131381e-01 2.72401780e-01
7.75830030e-01 7.82873809e-01 2.32847065e-01 -2.69152969e-01
-8.12869191e-01 2.08227098e-01 1.17744100e+00 4.21312064e-01
3.74911353e-02 -2.92148888e-01 8.63040462e-02 4.64500844e-01
4.56058145e-01 4.97109741e-01 5.13651192e-01 -4.16077822e-01
1.61706477e-01 7.39523172e-01 4.07353878e-01 -1.76077366e+00
-2.53790706e-01 -5.13099611e-01 -6.46752775e-01 6.48644120e-02
5.05075455e-01 2.64994293e-01 -3.59455556e-01 9.79317963e-01
1.14560567e-01 -6.39619887e-01 -3.81489158e-01 8.69809449e-01
1.43651456e-01 5.93686402e-01 1.29097298e-01 -1.37978345e-01
1.60804927e+00 -4.47469771e-01 -9.08322394e-01 -1.02004632e-01
9.75881338e-01 -9.29553807e-01 1.07742906e+00 8.25141191e-01
-8.55905950e-01 1.01000592e-01 -1.18954301e+00 -9.12825689e-02
-7.13123441e-01 -1.08770266e-01 2.76732326e-01 1.07087016e+00
-4.70229447e-01 7.80528843e-01 -6.12781048e-01 -4.16398406e-01
4.80588347e-01 -2.14920476e-01 -5.21601811e-02 1.98268071e-01
-1.29373157e+00 1.48561263e+00 9.07389298e-02 2.22747549e-01
-4.78369385e-01 -3.81132215e-02 -5.98268032e-01 -2.10089326e-01
4.00682032e-01 -3.65555227e-01 1.30351448e+00 -1.31665111e+00
-1.32053685e+00 1.13308012e+00 8.32427517e-02 -6.96971357e-01
1.10207963e+00 -5.01447082e-01 -8.14762294e-01 2.79001862e-01
-8.42122957e-02 -2.99156517e-01 1.22326303e+00 -1.10118151e+00
-3.10566247e-01 -4.47075069e-01 1.15813531e-01 -2.04052016e-01
-6.65228844e-01 4.82260227e-01 3.82824123e-01 -5.44983745e-01
-2.28369340e-01 -7.45631397e-01 2.09682137e-01 2.07310215e-01
-7.71492004e-01 -6.96291029e-02 8.56526017e-01 -1.02992344e+00
1.39426744e+00 -1.92217159e+00 -2.75068700e-01 3.75177622e-01
7.34886587e-01 6.12903833e-01 4.88466322e-01 5.74887812e-01
4.25515294e-01 7.87660658e-01 4.66728732e-02 -2.01798916e-01
2.46297285e-01 -2.52230559e-02 -3.82482886e-01 9.30313468e-01
7.59722739e-02 6.50587320e-01 -1.03761041e+00 -2.01059327e-01
2.16685712e-01 5.45306861e-01 -1.69306800e-01 -2.80210096e-02
-1.76915735e-01 2.03876376e-01 -8.95678461e-01 4.40205634e-01
2.61894226e-01 -3.21489662e-01 1.39949918e-01 1.46384269e-01
-1.91612288e-01 7.60510087e-01 -4.99917954e-01 5.71563721e-01
-4.45577800e-01 8.22471201e-01 3.71923059e-04 -7.37930238e-01
6.11285806e-01 1.71394587e-01 -2.80016363e-01 -6.76295459e-01
6.88029945e-01 4.64811265e-01 -1.00474678e-01 -6.24124110e-01
6.25595510e-01 -7.03306571e-02 -1.52848482e-01 5.77400267e-01
-6.12152040e-01 1.16266847e-01 -1.69153199e-01 4.39889878e-01
7.92493403e-01 -2.26682812e-01 7.69244790e-01 -3.58871788e-01
7.69199848e-01 -2.66160909e-02 -2.67669857e-01 6.91100717e-01
-3.24598044e-01 3.99885029e-02 6.31297886e-01 -6.94371700e-01
-9.40786541e-01 -5.82175016e-01 -5.91892563e-03 1.03471601e+00
3.94530930e-02 -3.44183803e-01 -8.98290873e-01 -1.05599666e+00
-1.73350543e-01 1.45846069e+00 -7.53315151e-01 -3.48966420e-01
-2.04187065e-01 -4.11201656e-01 6.10579729e-01 -8.44835788e-02
4.08106118e-01 -6.46788359e-01 -1.10246253e+00 2.97256470e-01
-1.87504530e-01 -9.60844457e-01 -2.37450957e-01 -2.14288890e-01
-4.68765259e-01 -1.11796391e+00 -3.62323523e-01 -5.35984896e-02
5.62553763e-01 1.70911014e-01 8.42775881e-01 5.23656189e-01
-4.83232662e-02 2.24672914e-01 -7.51901627e-01 -5.92721045e-01
-1.40731382e+00 -2.41565071e-02 7.64856162e-03 1.42084330e-01
2.60136783e-01 -2.16025889e-01 -3.73508811e-01 1.42992780e-01
-1.17872787e+00 5.50036319e-03 3.46940398e-01 5.28109133e-01
-3.52617174e-01 -2.54935324e-01 3.55540931e-01 -1.33185196e+00
9.56565619e-01 -3.79294127e-01 -4.44933832e-01 2.01602012e-01
-7.59173632e-01 8.64285529e-02 8.78261268e-01 -3.11853826e-01
-7.22807288e-01 -4.71478909e-01 -3.79771501e-01 3.43389869e-01
-2.33622313e-01 4.45013791e-01 1.13954984e-01 -3.31331074e-01
1.03732026e+00 2.32031643e-01 1.49158135e-01 -3.67450505e-01
5.13017654e-01 9.47606742e-01 5.26065826e-02 -8.93148184e-02
9.66136813e-01 5.16579211e-01 -3.03001523e-01 -1.07599890e+00
-1.19918871e+00 -4.29229915e-01 -3.28658819e-01 -4.05456364e-01
4.86615509e-01 -4.20211047e-01 -7.33016133e-01 3.43494475e-01
-1.42480826e+00 2.72433043e-01 3.91582042e-01 4.70359847e-02
-1.63774624e-01 6.71094239e-01 -6.11230135e-01 -1.29786980e+00
-3.13090533e-01 -7.09268391e-01 7.43956685e-01 -1.97540268e-01
-7.71596611e-01 -1.21983373e+00 -3.15579981e-01 6.06860995e-01
6.61125481e-01 5.30114233e-01 8.15365374e-01 -1.28917277e+00
-8.71673375e-02 -6.59196794e-01 -7.30516762e-02 3.95936608e-01
1.12966858e-01 1.23939574e-01 -9.32317197e-01 -4.76121679e-02
4.04749513e-01 -3.35104406e-01 2.81287670e-01 -1.90233558e-01
8.22799027e-01 -1.17600942e+00 -2.52172291e-01 -2.47840554e-01
1.12246490e+00 -1.66274354e-01 7.17591822e-01 5.01146317e-01
3.11416656e-01 7.13157713e-01 2.73029000e-01 4.79219496e-01
1.91529155e-01 5.93054354e-01 3.90081167e-01 1.65453076e-01
3.81609589e-01 -3.76584083e-01 4.04182017e-01 4.85946536e-01
-2.50562858e-02 -4.30870235e-01 -6.96236074e-01 3.20759982e-01
-1.62295544e+00 -1.12915027e+00 -5.07258236e-01 2.16124773e+00
5.57475746e-01 5.64480960e-01 2.91187257e-01 3.79355192e-01
4.60905433e-01 1.29293963e-01 -4.77655865e-02 -8.70790720e-01
2.23965600e-01 -3.25219750e-01 4.88415658e-01 4.08384919e-01
-9.59685445e-01 5.45431554e-01 6.19879723e+00 5.98529696e-01
-1.10636973e+00 1.79623798e-01 8.32978308e-01 2.84582019e-01
-3.21227729e-01 -3.65983218e-01 -4.08481061e-01 6.07455730e-01
1.10886180e+00 -1.17521286e-01 2.07788631e-01 8.88964593e-01
5.73867619e-01 -2.34341770e-01 -8.45970094e-01 7.03664958e-01
4.50279564e-01 -1.25846219e+00 7.48647284e-03 1.42702058e-01
2.87935227e-01 -4.75257874e-01 -9.27328132e-03 1.16043061e-03
5.77215329e-02 -9.67900276e-01 1.13116312e+00 3.03116351e-01
3.85887735e-02 -6.66592836e-01 8.16265881e-01 6.83012843e-01
-2.33219266e-01 2.00929791e-02 -2.88519021e-02 -4.02344406e-01
1.62633017e-01 8.61453652e-01 -8.65445971e-01 3.42823327e-01
1.30988911e-01 3.10438335e-01 -6.80326879e-01 7.17066705e-01
-5.08875430e-01 7.08773315e-01 -2.34082207e-01 -8.73366833e-01
3.98043245e-01 2.47922204e-02 6.50172591e-01 1.27807522e+00
1.00272566e-01 -2.21822023e-01 -1.10302053e-01 9.96324599e-01
-2.89756572e-03 2.14216515e-01 -6.71643198e-01 -6.57099187e-01
1.28162771e-01 1.21578479e+00 -7.55549788e-01 -3.62406582e-01
-2.30736539e-01 1.27499533e+00 3.39034200e-01 -1.27279758e-01
-8.42831850e-01 -1.11957490e-01 2.42463470e-01 7.10633337e-01
-2.50087231e-01 -1.29493907e-01 -2.42509738e-01 -1.19036627e+00
-1.46328034e-02 -1.14924896e+00 1.87987462e-01 -3.67381394e-01
-1.38507831e+00 6.04220510e-01 -2.74111181e-01 -8.58309090e-01
-3.52168620e-01 -7.45037615e-01 -3.48716617e-01 4.80994672e-01
-1.30959988e+00 -1.14612508e+00 1.70923416e-02 1.38749033e-01
3.01113814e-01 3.18887472e-01 7.92416573e-01 2.94642188e-02
-2.84132481e-01 3.20167840e-01 -1.42085040e-03 1.92881107e-01
4.06526923e-01 -1.04323161e+00 3.56836945e-01 8.20813894e-01
2.28140712e-01 7.22672820e-01 1.29958665e+00 -5.59033692e-01
-1.25335121e+00 -8.01906526e-01 1.31048214e+00 -7.28065491e-01
1.07386172e+00 -6.29135072e-01 -8.67034018e-01 4.82837647e-01
-6.81756018e-03 -5.52712977e-01 6.04530334e-01 1.90336183e-01
-7.05558360e-01 4.08117235e-01 -1.21677935e+00 7.21185505e-01
6.36191666e-01 -6.13873243e-01 -9.40089881e-01 6.23533309e-01
7.17166215e-02 -1.24428988e-01 -3.54805082e-01 -2.16544449e-01
6.30239367e-01 -1.20380175e+00 4.71941352e-01 -6.54337645e-01
6.70651495e-01 1.75543800e-02 3.24839056e-01 -1.23678637e+00
-1.67456016e-01 -7.58177996e-01 -9.68262479e-02 8.99517119e-01
4.48732138e-01 -9.01344895e-01 2.91787297e-01 8.44522893e-01
1.94976479e-01 -4.29945260e-01 -7.01804757e-01 -6.65293396e-01
-1.36276633e-01 -3.74770910e-01 2.55790837e-02 1.03588104e+00
5.12190521e-01 3.82237256e-01 -6.86767757e-01 7.58717060e-02
4.03806210e-01 -1.54066786e-01 5.00874400e-01 -1.24152696e+00
-1.41419873e-01 -6.69722915e-01 -5.59032261e-01 -8.41986775e-01
-2.28080794e-01 -4.88061994e-01 -8.62757862e-02 -1.19319606e+00
2.02061206e-01 9.77790877e-02 2.81512320e-01 1.13496453e-01
5.60645424e-02 3.48261446e-01 1.11562043e-01 1.92477420e-01
-6.40269756e-01 2.07808048e-01 9.81412888e-01 -2.57814825e-02
1.05509646e-01 1.16524257e-01 -1.05668736e+00 1.02947199e+00
7.19900131e-01 -6.72031999e-01 -1.31444102e-02 9.92400572e-02
7.38327205e-01 -3.14179331e-01 7.67901182e-01 -6.51444852e-01
-6.63059801e-02 5.77744320e-02 2.85109371e-01 -1.44021302e-01
1.36241436e-01 -6.97332740e-01 -1.03552580e-01 7.77017474e-01
-4.69887555e-01 -1.23610221e-01 -2.53318876e-01 7.59498835e-01
-3.78950797e-02 -4.73756701e-01 8.24942410e-01 -2.58126527e-01
-1.08849257e-01 -2.95226604e-01 -1.03615940e+00 -9.92408842e-02
9.33018863e-01 -5.71740828e-02 -4.88413543e-01 -8.89148772e-01
-3.74847323e-01 -1.05190091e-01 5.38373947e-01 3.69311661e-01
3.50400180e-01 -8.04496348e-01 -6.14980876e-01 -1.76377118e-01
2.17508957e-01 -9.17452097e-01 -1.19399503e-01 7.87785411e-01
-8.66980433e-01 4.97081459e-01 5.61185665e-02 -6.88755289e-02
-1.18179440e+00 6.36273265e-01 1.94135845e-01 -1.90808117e-01
-5.38111210e-01 3.24342161e-01 -3.26221377e-01 1.43173501e-01
1.11507192e-01 -2.83452988e-01 -2.71383166e-01 4.96759638e-02
1.14922178e+00 4.54779744e-01 1.51673362e-01 -7.04517066e-01
-2.74618924e-01 -2.14372575e-01 -3.38875651e-01 1.50535524e-01
1.08023667e+00 -3.21484953e-01 -5.53873107e-02 4.82502609e-01
1.11085761e+00 4.42730308e-01 -4.67323601e-01 2.38111883e-01
4.27083492e-01 -6.36148214e-01 1.05168531e-02 -1.09695268e+00
-3.60374868e-01 6.87612295e-01 -1.32775828e-02 1.19578910e+00
6.18410826e-01 -6.95825294e-02 7.89915502e-01 2.79802144e-01
3.60301018e-01 -1.02262676e+00 8.05358961e-03 -1.47452634e-02
1.02622139e+00 -1.21042705e+00 1.79673016e-01 -7.74538755e-01
-5.91854811e-01 1.34723806e+00 -1.96112230e-01 4.34573879e-03
3.78214955e-01 -1.56825870e-01 7.54372776e-02 -4.62720782e-01
-2.21071959e-01 2.58691639e-01 5.36564827e-01 3.41398478e-01
5.95706105e-01 1.88051805e-01 -8.53188157e-01 5.36438107e-01
-1.96136326e-01 -1.96437836e-01 8.59314382e-01 7.13568985e-01
-5.92310071e-01 -9.86072004e-01 -4.45928216e-01 4.37574416e-01
-9.86493945e-01 1.08954638e-01 -9.86778617e-01 8.23559940e-01
-2.34779358e-01 1.25218403e+00 -5.82298338e-01 -1.95374355e-01
1.30012184e-01 1.64939046e-01 1.98285669e-01 -3.23423773e-01
-8.18724453e-01 -1.94095075e-01 7.19193161e-01 -5.75390399e-01
-3.01184386e-01 -4.00804162e-01 -6.24390781e-01 -6.07886374e-01
-4.37898248e-01 1.85393959e-01 7.70674527e-01 1.36175764e+00
2.13964418e-01 1.19231820e-01 4.47140604e-01 -3.83486211e-01
-1.09635329e+00 -8.24330211e-01 -5.08055329e-01 8.55886340e-01
3.10272276e-01 -4.91029501e-01 -6.69940174e-01 -7.06943795e-02] | [8.219870567321777, 10.127910614013672] |
340d4bcb-d1f2-4a26-853e-0861263e7070 | improving-unsupervised-image-clustering-with | 2012.11150 | null | https://arxiv.org/abs/2012.11150v2 | https://arxiv.org/pdf/2012.11150v2.pdf | Improving Unsupervised Image Clustering With Robust Learning | Unsupervised image clustering methods often introduce alternative objectives to indirectly train the model and are subject to faulty predictions and overconfident results. To overcome these challenges, the current research proposes an innovative model RUC that is inspired by robust learning. RUC's novelty is at utilizing pseudo-labels of existing image clustering models as a noisy dataset that may include misclassified samples. Its retraining process can revise misaligned knowledge and alleviate the overconfidence problem in predictions. The model's flexible structure makes it possible to be used as an add-on module to other clustering methods and helps them achieve better performance on multiple datasets. Extensive experiments show that the proposed model can adjust the model confidence with better calibration and gain additional robustness against adversarial noise. | ['Meeyoung Cha', 'Seunghoon Hong', 'Sungkyu Park', 'Danu Kim', 'Sundong Kim', 'Sungwon Han', 'Sungwon Park'] | 2020-12-21 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Park_Improving_Unsupervised_Image_Clustering_With_Robust_Learning_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Park_Improving_Unsupervised_Image_Clustering_With_Robust_Learning_CVPR_2021_paper.pdf | cvpr-2021-1 | ['image-clustering', 'unsupervised-image-classification'] | ['computer-vision', 'computer-vision'] | [ 1.08776495e-01 2.84390569e-01 -3.55295181e-01 -7.19017982e-01
-6.15585625e-01 -2.83327609e-01 4.33636695e-01 -1.83643296e-01
-2.89198905e-01 7.48196602e-01 -1.85311794e-01 -1.10957669e-02
-2.59780943e-01 -5.10411561e-01 -6.13427997e-01 -9.68486845e-01
1.67913467e-01 2.91278422e-01 3.92495215e-01 3.03432852e-01
4.99600500e-01 1.00711338e-01 -1.56027663e+00 4.88739669e-01
1.21232510e+00 7.54447043e-01 1.14325292e-01 2.63294965e-01
1.63212612e-01 1.00331461e+00 -7.97204971e-01 -5.67648590e-01
2.96230733e-01 -9.77131724e-02 -7.75138497e-01 3.87875676e-01
4.74274009e-02 3.06283850e-02 2.22630337e-01 1.34035623e+00
3.03189754e-01 -5.15799746e-02 5.33661067e-01 -1.66049159e+00
-8.82630587e-01 1.18859637e+00 -6.07761323e-01 -1.25321403e-01
9.23448056e-02 2.13910028e-01 3.23546350e-01 -6.57804906e-01
3.44880372e-01 1.23926330e+00 1.06697452e+00 7.61193395e-01
-1.26379216e+00 -1.02920961e+00 4.03332174e-01 4.48933780e-01
-1.64814985e+00 -1.71139210e-01 8.88830960e-01 -2.81139314e-01
2.78085768e-01 3.99340928e-01 5.88986725e-02 1.47840250e+00
-8.33677202e-02 7.45762885e-01 1.34241796e+00 -2.92870820e-01
3.25757712e-01 6.40390158e-01 1.85474947e-01 3.82940948e-01
2.57445037e-01 1.93566129e-01 -1.60589777e-02 7.81935267e-03
2.33227476e-01 1.24476224e-01 -4.94683385e-01 -2.10333198e-01
-1.04785383e+00 7.30986655e-01 4.80303138e-01 1.34397388e-01
-5.65691665e-03 -1.00254260e-01 5.31739257e-02 1.53080732e-01
2.86821038e-01 4.99850839e-01 -3.70471686e-01 1.71505406e-01
-8.03893745e-01 -6.51957765e-02 4.24509317e-01 1.04660928e+00
6.65256858e-01 1.18957505e-01 -2.62320787e-02 1.01779735e+00
4.48721021e-01 3.80671442e-01 1.02867520e+00 -1.14309633e+00
7.77066499e-02 7.33180344e-01 -6.95817918e-02 -1.19708753e+00
-2.40946233e-01 -7.39020944e-01 -9.57943320e-01 2.14238197e-01
-1.30728289e-01 -5.48804961e-02 -1.10681903e+00 1.64318180e+00
2.74089128e-01 7.51683950e-01 4.60945815e-02 7.74196267e-01
6.95643365e-01 4.25077230e-01 1.48808718e-01 -2.86134243e-01
6.28962636e-01 -1.08178389e+00 -6.82927072e-01 6.45984113e-02
2.84232050e-01 -6.77521765e-01 7.99794197e-01 7.19135880e-01
-5.41890025e-01 -1.03024006e+00 -1.27895725e+00 6.66363239e-01
-4.55173731e-01 -3.77552770e-02 4.44116145e-01 1.02660131e+00
-9.79221702e-01 5.89013398e-01 -5.94537735e-01 -9.17149335e-02
5.36012113e-01 4.41632807e-01 -1.81133524e-01 6.88767955e-02
-9.31925178e-01 7.39587009e-01 9.71076429e-01 1.11729324e-01
-6.77713990e-01 -4.81424809e-01 -5.12271106e-01 -1.75373748e-01
5.78862071e-01 -1.75559521e-01 9.68970001e-01 -1.35689044e+00
-1.58517265e+00 4.96025234e-01 3.62252355e-01 -5.30078113e-01
7.51729369e-01 -3.00382171e-02 -7.69191384e-01 8.73487741e-02
7.95021206e-02 8.59021723e-01 1.05259347e+00 -2.01964402e+00
-6.89492404e-01 -2.40996957e-01 -5.49571633e-01 -1.69452634e-02
-2.48501807e-01 -2.96891212e-01 -5.42321920e-01 -9.34378088e-01
2.41232261e-01 -9.43184972e-01 -4.58565414e-01 -4.16512311e-01
-6.06522799e-01 -2.30146460e-02 1.22172534e+00 4.05145548e-02
1.17172945e+00 -1.92850804e+00 -4.47037995e-01 6.48584008e-01
2.66710413e-03 6.10025167e-01 -1.33422375e-01 -4.25035469e-02
-2.57657617e-01 5.28030872e-01 -4.04933155e-01 -3.34984392e-01
-3.32816988e-01 5.09550095e-01 -1.67734623e-01 1.80812821e-01
2.80838966e-01 5.69432259e-01 -8.52076530e-01 -7.07071960e-01
2.49770835e-01 2.49677271e-01 -3.29702377e-01 2.07454756e-01
8.51770118e-02 5.51736414e-01 -2.62625009e-01 6.40687644e-01
1.13823009e+00 -3.27567518e-01 2.32648641e-01 -1.40961364e-01
2.73885012e-01 -5.04712760e-01 -1.51512671e+00 1.04785514e+00
1.70547202e-01 2.05096647e-01 -1.12224810e-01 -1.42703187e+00
1.09345853e+00 2.03105271e-01 3.85622859e-01 -1.71747878e-01
2.97954649e-01 -9.64474306e-02 -1.18085682e-01 -6.87809169e-01
3.32043469e-01 5.00869714e-02 9.54731032e-02 2.36754969e-01
1.09815616e-02 6.28856793e-02 -3.93421739e-01 1.72972605e-01
5.70096076e-01 4.89899516e-02 4.88897227e-02 -2.51944274e-01
8.22633207e-01 6.22620285e-02 1.19343865e+00 1.07500374e+00
-4.13656503e-01 6.63982153e-01 -2.07632743e-02 -1.96045846e-01
-6.59821987e-01 -1.11825800e+00 -3.36764365e-01 5.25453806e-01
3.70014936e-01 -2.31504276e-01 -7.96644628e-01 -1.13068807e+00
-1.18206203e-01 8.24947715e-01 -6.20955288e-01 -4.89530057e-01
-1.41553938e-01 -1.01833856e+00 6.46096647e-01 6.13763034e-01
8.46246958e-01 -7.23015845e-01 -1.12909339e-02 -5.71223237e-02
-3.72810125e-01 -1.05932570e+00 -1.64253280e-01 1.02092996e-01
-8.39041650e-01 -1.39860070e+00 -1.00557350e-01 -8.34814131e-01
1.09050214e+00 3.51466894e-01 9.00834620e-01 3.49926740e-01
-1.28990104e-02 4.83319402e-01 -7.30158508e-01 -7.13921428e-01
-8.08123946e-01 -1.74353868e-01 3.63661438e-01 2.76345700e-01
4.74556267e-01 -3.56086224e-01 -4.54606712e-01 8.79406035e-01
-1.10879612e+00 -1.96390703e-01 3.60340953e-01 1.00304127e+00
6.12231433e-01 5.60056150e-01 9.27170992e-01 -1.48376203e+00
5.68129778e-01 -6.07910812e-01 -6.41372263e-01 4.53932911e-01
-1.49518657e+00 -7.18387663e-02 8.31624329e-01 -8.51905286e-01
-1.44811893e+00 1.76530555e-01 2.35782996e-01 -6.34967029e-01
-2.94619828e-01 2.06065297e-01 -1.84990227e-01 -2.48980105e-01
5.04569829e-01 7.90556241e-03 2.09823072e-01 -4.17782992e-01
3.06943744e-01 9.37454700e-01 8.95677686e-01 -5.79449177e-01
9.58789527e-01 3.81887496e-01 -3.40816289e-01 -3.44950736e-01
-7.52805591e-01 -3.68677676e-01 -7.04234660e-01 -4.72257763e-01
6.06307447e-01 -1.00453091e+00 -7.74887383e-01 5.35389960e-01
-6.19675279e-01 -1.21536762e-01 1.39452159e-01 5.09662986e-01
-2.50625759e-01 8.09647620e-01 -3.00350457e-01 -7.77585387e-01
-1.46848857e-01 -1.23311877e+00 3.66879910e-01 4.36668098e-01
-4.65128385e-02 -8.19635272e-01 -3.12978089e-01 7.11132288e-01
2.61917621e-01 2.42461801e-01 6.18731380e-01 -9.40996885e-01
-5.21007895e-01 -2.89918602e-01 1.19183891e-01 9.00029480e-01
1.17224835e-01 4.99520868e-01 -1.07665420e+00 -2.87416786e-01
2.45818332e-01 -4.28607851e-01 8.71784568e-01 1.85958017e-02
1.72460818e+00 -4.81965065e-01 -4.25946772e-01 6.64746165e-01
1.42972565e+00 3.18203866e-01 6.74605429e-01 4.96397525e-01
4.52857405e-01 5.38380742e-01 9.23764288e-01 3.75399113e-01
3.52853745e-01 3.67475599e-01 5.93259096e-01 9.27735642e-02
1.98882654e-01 -2.60715246e-01 1.78126201e-01 7.50462890e-01
1.25776473e-02 -5.68737760e-02 -6.98332429e-01 2.71366805e-01
-2.13375592e+00 -1.11011088e+00 -1.64971009e-01 2.17508101e+00
9.02039051e-01 2.82081902e-01 -2.78840512e-01 2.80406207e-01
1.08495486e+00 -2.23518014e-01 -4.78494108e-01 -8.69992077e-02
-2.00215876e-01 -1.05207607e-01 4.49053586e-01 4.16154355e-01
-1.22611058e+00 9.31613147e-01 7.27194357e+00 8.75829339e-01
-9.18845773e-01 -4.76367921e-02 1.00021064e+00 2.24842146e-01
-1.24502324e-01 1.20410688e-01 -5.54972291e-01 8.42737615e-01
6.67246521e-01 3.30368340e-01 1.39052585e-01 1.07347882e+00
7.52494782e-02 -1.87952712e-01 -9.16959763e-01 9.99195814e-01
3.07797253e-01 -1.15104866e+00 1.16605394e-01 -3.13618332e-01
1.01282752e+00 -3.07501286e-01 3.37844700e-01 3.87913167e-01
7.26498306e-01 -9.54766572e-01 5.28700769e-01 6.93829656e-01
3.13192815e-01 -9.17986512e-01 1.15052962e+00 4.93737131e-01
-5.79137206e-01 -3.48459870e-01 -5.08918941e-01 1.76853597e-01
-2.66750604e-01 5.07384777e-01 -1.17159653e+00 4.72853869e-01
9.68747854e-01 5.89629114e-01 -1.09502852e+00 9.41357434e-01
-3.24846178e-01 8.90138149e-01 -1.34566709e-01 1.09211549e-01
5.79951070e-02 -2.07159787e-01 3.57347608e-01 1.06189966e+00
-2.03342084e-02 5.00122048e-02 4.58110064e-01 7.36606717e-01
-2.32065581e-02 -1.60061657e-01 -3.80088717e-01 5.03678501e-01
9.45964873e-01 1.16995358e+00 -8.98671329e-01 -4.39079374e-01
-6.06808886e-02 9.08414900e-01 2.51386076e-01 3.38742763e-01
-1.14706421e+00 -1.27638966e-01 1.83078811e-01 -3.10708702e-01
1.23636626e-01 1.97580472e-01 -4.42550451e-01 -9.61389422e-01
-1.87636495e-01 -1.01641285e+00 5.76486647e-01 -8.37325394e-01
-1.57156980e+00 5.41964889e-01 -4.05813120e-02 -1.48854959e+00
-9.59750032e-04 -4.54030514e-01 -6.22478127e-01 1.28003120e-01
-1.27885187e+00 -1.17020476e+00 -4.68237370e-01 9.24013436e-01
4.02825713e-01 -5.14581084e-01 6.79544151e-01 5.77421375e-02
-8.22306752e-01 1.04628015e+00 2.03216940e-01 2.69153416e-01
1.20811439e+00 -1.21240234e+00 -5.30790031e-01 1.05671597e+00
1.84515100e-02 6.14938974e-01 6.01038635e-01 -7.19949305e-01
-5.60413659e-01 -1.48566043e+00 2.05433175e-01 -7.45733023e-01
3.07533503e-01 6.55276421e-03 -9.89911437e-01 6.00977361e-01
4.01299559e-02 -1.50687650e-01 1.14205706e+00 -6.89656436e-02
-4.77604538e-01 -5.10635555e-01 -1.59168184e+00 4.39724535e-01
6.91588938e-01 -1.37044951e-01 -6.79837048e-01 1.16972931e-01
6.97403848e-01 -9.11652073e-02 -9.87514794e-01 6.94620609e-01
1.00049518e-01 -1.12358689e+00 8.74980807e-01 -4.85149950e-01
1.69917747e-01 -4.79914844e-01 -2.76434477e-02 -1.17969906e+00
-5.76262534e-01 -4.86471623e-01 1.38526589e-01 1.71962893e+00
5.06870806e-01 -5.84774911e-01 7.63118088e-01 7.35764444e-01
-7.57411420e-02 -4.47883040e-01 -7.56551981e-01 -7.90938497e-01
-2.80153632e-01 -6.13729715e-01 7.67355025e-01 1.54563761e+00
-6.16449863e-03 8.65927041e-02 -6.01642132e-01 6.14378870e-01
9.23694730e-01 -1.57534227e-01 8.01510334e-01 -1.35334456e+00
-6.74112737e-02 -2.54608661e-01 -5.20984709e-01 -4.65341032e-01
2.08116412e-01 -7.00421393e-01 3.40346038e-01 -7.21981406e-01
2.29783162e-01 -6.11258984e-01 -5.79980791e-01 6.19865954e-01
-6.00376129e-01 4.27228510e-01 1.13692150e-01 4.11208928e-01
-8.34014416e-01 4.12495226e-01 8.16235483e-01 -3.54439855e-01
-1.56493768e-01 3.85258347e-01 -9.06227171e-01 8.67793798e-01
1.00932789e+00 -6.86664999e-01 -7.09302068e-01 -2.47884214e-01
-3.70536000e-01 -5.43019891e-01 2.00641751e-01 -1.21600378e+00
4.86988604e-01 -2.42874771e-01 6.55627012e-01 -5.61876833e-01
-1.61610737e-01 -1.26980507e+00 2.97120005e-01 4.16613132e-01
-4.27987397e-01 -5.45965657e-02 -1.86174735e-01 9.49841201e-01
-2.96491861e-01 -4.89425153e-01 1.12815058e+00 -2.23622650e-01
-7.33003020e-01 7.51561001e-02 -3.18752944e-01 -3.07791024e-01
1.40644670e+00 -5.96805573e-01 -2.52512157e-01 -2.74646193e-01
-6.36185288e-01 6.98932290e-01 5.00533462e-01 7.00977206e-01
7.92062938e-01 -1.26559484e+00 -4.90568608e-01 1.89408630e-01
1.09289080e-01 1.09825283e-01 1.11981742e-01 6.10980034e-01
-2.27234185e-01 -3.02150007e-02 -3.86779867e-02 -9.05419946e-01
-1.34690475e+00 8.92661393e-01 1.55217424e-01 -1.72745325e-02
-1.67838261e-01 8.02343845e-01 -4.36172277e-01 -6.91915870e-01
5.99411786e-01 4.63634245e-02 -2.73743719e-01 -2.68818974e-01
4.19720918e-01 4.46057647e-01 -2.75636911e-01 -5.86806655e-01
-2.63595164e-01 4.92582411e-01 -3.12310159e-01 2.62043029e-01
1.24201202e+00 -4.48365778e-01 -1.53845206e-01 3.73182297e-01
7.12575138e-01 -1.07540861e-01 -1.31591833e+00 -2.74607599e-01
2.42174566e-01 -5.26212394e-01 -1.32594049e-01 -1.18665826e+00
-1.18018079e+00 3.50340545e-01 1.03735912e+00 1.63032513e-04
1.22525811e+00 -2.15937957e-01 2.48877525e-01 5.74282587e-01
2.92416185e-01 -1.72039831e+00 3.33337277e-01 5.71581684e-02
3.87988210e-01 -1.79570282e+00 1.13948517e-01 -5.24374723e-01
-1.03251219e+00 9.94133174e-01 1.13823962e+00 1.36365533e-01
8.40577543e-01 2.40419015e-01 4.35988128e-01 1.98509187e-01
-6.35510147e-01 1.65125728e-01 2.73282882e-02 1.27626646e+00
-6.64168522e-02 7.87217095e-02 -2.84347296e-01 9.72369134e-01
6.42551780e-02 -2.79710032e-02 5.44208407e-01 6.11479998e-01
-4.73216772e-01 -1.02430367e+00 -6.69685185e-01 2.04367220e-01
-5.22154033e-01 2.81628311e-01 -4.26708907e-01 7.56831884e-01
6.58701777e-01 1.49831772e+00 -2.29035303e-01 -9.35580015e-01
-1.26391416e-02 4.23292816e-02 -5.82254864e-02 -3.20434183e-01
-5.84916055e-01 -5.81541136e-02 -3.32960010e-01 -5.83852410e-01
-9.57872212e-01 -2.56582141e-01 -1.19762695e+00 -9.71870590e-03
-6.18635952e-01 4.73903239e-01 2.85376042e-01 7.99126625e-01
3.49845380e-01 3.49244237e-01 1.11821318e+00 -3.93047303e-01
-8.20386887e-01 -9.66124654e-01 -3.03020120e-01 8.93678904e-01
-4.31001261e-02 -5.94925284e-01 -3.81763726e-01 2.98939884e-01] | [14.90660572052002, 1.2018327713012695] |
b83921e3-2010-4147-86a2-2a9b0db9b22d | stylealign-analysis-and-applications-of-1 | 2110.11323 | null | https://arxiv.org/abs/2110.11323v2 | https://arxiv.org/pdf/2110.11323v2.pdf | StyleAlign: Analysis and Applications of Aligned StyleGAN Models | In this paper, we perform an in-depth study of the properties and applications of aligned generative models. We refer to two models as aligned if they share the same architecture, and one of them (the child) is obtained from the other (the parent) via fine-tuning to another domain, a common practice in transfer learning. Several works already utilize some basic properties of aligned StyleGAN models to perform image-to-image translation. Here, we perform the first detailed exploration of model alignment, also focusing on StyleGAN. First, we empirically analyze aligned models and provide answers to important questions regarding their nature. In particular, we find that the child model's latent spaces are semantically aligned with those of the parent, inheriting incredibly rich semantics, even for distant data domains such as human faces and churches. Second, equipped with this better understanding, we leverage aligned models to solve a diverse set of tasks. In addition to image translation, we demonstrate fully automatic cross-domain image morphing. We further show that zero-shot vision tasks may be performed in the child domain, while relying exclusively on supervision in the parent domain. We demonstrate qualitatively and quantitatively that our approach yields state-of-the-art results, while requiring only simple fine-tuning and inversion. | ['Dani Lischinski', 'Eli Shechtman', 'Yotam Nitzan', 'Zongze Wu'] | 2021-10-21 | stylealign-analysis-and-applications-of | https://openreview.net/forum?id=Qg2vi4ZbHM9 | https://openreview.net/pdf?id=Qg2vi4ZbHM9 | iclr-2022-4 | ['image-morphing'] | ['computer-vision'] | [ 4.79190171e-01 3.86797965e-01 -1.16286665e-01 -3.45751256e-01
-6.69904113e-01 -7.47295320e-01 8.24065745e-01 -4.53992903e-01
-1.57242671e-01 6.82120085e-01 1.70107737e-01 -6.16705827e-02
1.80897772e-01 -7.35844970e-01 -9.87449765e-01 -7.50429690e-01
3.46410066e-01 6.36046410e-01 -6.84068948e-02 -3.66378307e-01
-4.28133123e-02 1.49367586e-01 -1.28819370e+00 2.17481509e-01
6.60455465e-01 6.46306813e-01 1.37890220e-01 5.35747111e-01
-2.21927203e-02 5.88581026e-01 -2.24909365e-01 -8.45649600e-01
3.72586340e-01 -9.01805162e-01 -9.45508540e-01 3.13172013e-01
6.45007193e-01 -2.99658626e-01 -2.42905065e-01 1.11049032e+00
1.83560789e-01 -1.58925578e-01 7.80363798e-01 -1.42391431e+00
-9.57912385e-01 4.60719705e-01 -4.84062880e-01 -2.45190576e-01
1.58555955e-01 1.12382695e-01 1.07744801e+00 -7.29425669e-01
8.74220848e-01 1.12268698e+00 4.72533196e-01 7.91028917e-01
-1.64855945e+00 -6.03489220e-01 7.54572079e-02 -9.02294293e-02
-1.13401306e+00 -5.97867966e-01 8.12755167e-01 -6.37166321e-01
4.57957596e-01 -1.39970127e-02 6.28032506e-01 1.28030276e+00
-5.83741218e-02 6.74745679e-01 1.31210828e+00 -6.55365229e-01
8.16345885e-02 2.49770597e-01 -1.96178988e-01 8.49818647e-01
1.15584977e-01 2.12084845e-01 -6.66694880e-01 1.60732761e-01
7.71445453e-01 -3.11218619e-01 -2.44041696e-01 -8.77212167e-01
-1.37437987e+00 9.73021686e-01 4.08768356e-01 2.95390993e-01
-9.89714414e-02 1.40238598e-01 1.25263050e-01 4.20805573e-01
4.24121976e-01 4.96797353e-01 -2.98397362e-01 1.07735589e-01
-8.80686283e-01 2.48566493e-01 7.18829572e-01 1.23047721e+00
1.17255723e+00 5.15046641e-02 6.13214783e-02 5.51125884e-01
1.08521685e-01 4.76853877e-01 5.12176156e-01 -1.10729730e+00
1.92436159e-01 1.82633936e-01 -7.20221996e-02 -7.67708659e-01
1.02889359e-01 -3.38749647e-01 -8.23610902e-01 1.99892983e-01
3.66986275e-01 -3.68500464e-02 -1.01275909e+00 2.20458269e+00
1.38929278e-01 3.33631605e-01 2.03655675e-01 5.43332517e-01
5.51648200e-01 3.60617548e-01 -1.07062116e-01 5.44068180e-02
1.37740207e+00 -1.18600285e+00 -3.12259912e-01 -3.94150943e-01
3.67410094e-01 -7.04960406e-01 1.16095245e+00 1.00715421e-01
-1.09901452e+00 -5.75234532e-01 -1.02011037e+00 -1.43245310e-01
-1.93074927e-01 -8.85915011e-02 5.34975946e-01 5.02868295e-01
-1.19685459e+00 6.78789556e-01 -8.08987200e-01 -7.74866283e-01
3.99892181e-01 8.14952180e-02 -5.38707078e-01 -9.32719335e-02
-9.38308358e-01 9.24457312e-01 1.49557114e-01 -3.99299711e-01
-1.06338477e+00 -6.89042211e-01 -9.16428924e-01 -1.33568332e-01
1.95322648e-01 -1.19687152e+00 1.35854518e+00 -1.30185914e+00
-1.50893354e+00 1.42194200e+00 -3.54830176e-01 -3.29888374e-01
5.60454965e-01 -7.35334828e-02 -1.47253364e-01 1.03600904e-01
1.86793789e-01 1.05557489e+00 1.14081419e+00 -1.43923962e+00
-4.17674839e-01 -4.37034488e-01 2.31072441e-01 9.63614509e-02
-3.51327866e-01 -6.28782138e-02 -4.59748417e-01 -7.45222449e-01
-1.21205650e-01 -1.12313795e+00 8.50030184e-02 1.46213949e-01
-3.53070945e-01 2.29521021e-01 6.27906859e-01 -4.20101821e-01
7.15559900e-01 -2.15164399e+00 6.85448885e-01 -5.99122457e-02
2.98853546e-01 3.06316968e-02 -3.38186234e-01 5.02890944e-01
-1.65986627e-01 7.39837959e-02 -5.78631520e-01 -5.97199440e-01
1.19956046e-01 3.90421480e-01 -6.00079954e-01 4.14414674e-01
3.17160875e-01 1.25643492e+00 -7.78454304e-01 -4.33460474e-01
1.15369726e-02 3.47321540e-01 -6.75535262e-01 3.57008487e-01
-2.03300029e-01 9.32990730e-01 -1.97786391e-01 4.01196808e-01
4.60308045e-01 -3.81106615e-01 1.61638930e-01 -3.54397446e-01
1.26556516e-01 1.89056680e-01 -7.91367829e-01 2.00753689e+00
-4.23719555e-01 7.53853559e-01 1.23801775e-01 -1.15315533e+00
7.48352468e-01 1.46561608e-01 3.09695482e-01 -6.39931679e-01
-3.81680094e-02 2.51654416e-01 -8.82417262e-02 -2.71539986e-01
3.36753666e-01 -5.74509084e-01 -5.67413382e-02 6.69535756e-01
4.12232727e-01 -2.62620986e-01 1.29560396e-01 1.49941772e-01
7.55591094e-01 4.65822786e-01 4.01419491e-01 -2.52724379e-01
2.34971702e-01 7.94668775e-03 3.36880207e-01 6.97985947e-01
8.14852677e-03 7.41990209e-01 4.42944258e-01 -1.63053274e-01
-1.35602665e+00 -1.32787800e+00 6.75156116e-02 1.08656633e+00
7.35190585e-02 -2.98560411e-01 -1.06237209e+00 -6.92097306e-01
-1.50268704e-01 5.52224934e-01 -8.55320930e-01 -2.15307489e-01
-6.03349984e-01 -4.16053563e-01 6.44411743e-01 4.33109283e-01
5.22086740e-01 -9.22302008e-01 -6.30262852e-01 -2.48721555e-01
-1.36646822e-01 -1.23027110e+00 -4.69976872e-01 5.52414618e-02
-6.87295675e-01 -9.25353765e-01 -9.64013338e-01 -8.95846188e-01
6.42012894e-01 3.18051368e-01 1.47021341e+00 3.07496786e-02
7.25011379e-02 6.91703737e-01 -3.17413211e-01 -2.21177593e-01
-7.59562433e-01 2.64383942e-01 -1.07600220e-01 1.78069085e-01
1.43460914e-01 -8.79145265e-01 -3.82099837e-01 2.93931901e-01
-1.04050553e+00 3.95843804e-01 6.22187555e-01 8.98094893e-01
4.92038280e-01 -5.57630002e-01 2.81430334e-01 -1.12193167e+00
3.34100872e-01 -3.41215134e-01 -5.64554393e-01 2.80419916e-01
-5.40978253e-01 2.63871998e-01 5.06933510e-01 -2.86970198e-01
-8.15969169e-01 1.30973518e-01 2.47660894e-02 -5.04699588e-01
-3.38347048e-01 2.99895376e-01 -3.17592621e-01 -9.28961784e-02
5.18425107e-01 3.06145877e-01 1.95030689e-01 -4.51905042e-01
7.36229777e-01 3.79920632e-01 7.79201627e-01 -7.57292330e-01
1.13313556e+00 7.81751096e-01 -3.02499291e-02 -7.78140724e-01
-9.24014151e-01 3.44205722e-02 -8.87487471e-01 9.51898769e-02
1.04966438e+00 -8.95781338e-01 -1.68432102e-01 5.35492420e-01
-1.18159807e+00 -5.54489255e-01 -5.32571495e-01 1.94059834e-01
-9.63104069e-01 2.75184512e-01 -4.55716759e-01 -2.34045088e-01
-4.51312661e-02 -1.15017235e+00 1.29643714e+00 6.47876561e-02
-3.56938303e-01 -1.18561232e+00 2.76843578e-01 2.62245566e-01
2.57366389e-01 1.75515309e-01 1.10329926e+00 -6.42328680e-01
-5.43293417e-01 2.88181692e-01 -1.54897496e-01 3.51200104e-01
3.01152200e-01 -2.38419876e-01 -1.07373583e+00 -3.98770094e-01
9.27092731e-02 -3.05414051e-01 9.14771318e-01 1.60816714e-01
7.78613508e-01 -7.69596025e-02 -2.54698366e-01 1.00466430e+00
1.34210277e+00 -1.14633851e-01 6.43336535e-01 3.02767307e-01
8.17911148e-01 8.37704480e-01 3.41648608e-01 1.38498591e-02
5.13060451e-01 8.75444710e-01 3.84861350e-01 -2.69008428e-01
-4.32218879e-01 -5.86202741e-01 3.59266073e-01 6.77748978e-01
2.92967949e-02 -6.66528195e-02 -6.52085781e-01 6.41604960e-01
-1.64898849e+00 -8.55979860e-01 2.16513172e-01 2.12680054e+00
9.46206510e-01 -1.19312741e-01 2.34327525e-01 -2.80330539e-01
7.17954457e-01 3.49380434e-01 -5.50087810e-01 -1.61457792e-01
-2.23457679e-01 4.93667126e-01 2.50276119e-01 5.32620728e-01
-1.03856456e+00 1.17423999e+00 6.95938826e+00 5.85615754e-01
-1.18570578e+00 1.80861130e-01 5.26523650e-01 5.02350517e-02
-6.33089364e-01 3.40183854e-01 -5.06002188e-01 3.00830930e-01
6.92219317e-01 -1.40989646e-01 6.52367353e-01 7.21064627e-01
-2.38637447e-01 1.67527571e-01 -1.54228866e+00 8.35895836e-01
2.39754260e-01 -1.38277948e+00 1.18137889e-01 2.16058254e-01
9.62440610e-01 -8.38104486e-02 1.92746267e-01 1.70479536e-01
4.47913438e-01 -1.11910844e+00 8.41044426e-01 3.56566757e-01
1.05178034e+00 -4.25130010e-01 2.24023581e-01 1.80233300e-01
-9.26453114e-01 3.35959673e-01 -2.30052918e-01 3.02188173e-02
1.80878147e-01 3.73771340e-01 -3.81363869e-01 5.88605464e-01
4.96947408e-01 8.86808634e-01 -4.90576178e-01 4.86392707e-01
-4.09309417e-01 4.32088614e-01 -4.56260182e-02 5.28511405e-01
5.99410711e-03 -4.55442905e-01 4.52301711e-01 9.96298313e-01
5.58587432e-01 -2.20342100e-01 -6.53104438e-03 1.29094028e+00
-2.19164386e-01 -2.34459147e-01 -8.76489520e-01 -1.55533582e-01
2.64068961e-01 1.11579001e+00 -4.69566584e-01 -4.46389079e-01
-5.51761925e-01 1.36968601e+00 2.97474384e-01 3.80931675e-01
-1.02029812e+00 -8.24429765e-02 8.88719797e-01 2.68412810e-02
5.99004984e-01 -3.11522186e-01 -2.97651380e-01 -1.39871633e+00
-2.39661876e-02 -1.01998854e+00 1.21261962e-01 -8.49407673e-01
-1.29180110e+00 6.86959565e-01 1.35506928e-01 -1.18638277e+00
-5.40674329e-01 -5.80460072e-01 -5.19483566e-01 8.68039548e-01
-1.61018348e+00 -1.57283103e+00 -3.66829276e-01 7.45237827e-01
5.03217280e-01 -1.81202605e-01 9.87754822e-01 1.55216500e-01
-4.64624465e-01 5.56247890e-01 1.75139438e-02 2.66039908e-01
8.82876575e-01 -1.17166340e+00 6.43451691e-01 8.47296715e-01
6.16180778e-01 7.05866396e-01 7.70352483e-01 -4.34655070e-01
-1.30300570e+00 -1.05113208e+00 7.30222702e-01 -4.95555520e-01
6.70003355e-01 -4.40403104e-01 -8.01679611e-01 1.06813788e+00
5.68680704e-01 -1.46915644e-01 6.62317276e-01 -3.88925336e-02
-7.17467010e-01 3.99310775e-02 -9.08545375e-01 6.53658926e-01
1.21612573e+00 -7.17354178e-01 -6.73853278e-01 9.09703150e-02
7.32510149e-01 -3.84407043e-01 -7.37234712e-01 1.93640694e-01
7.13144600e-01 -1.22955966e+00 9.21697319e-01 -8.81255090e-01
8.41796875e-01 -1.42168388e-01 -4.20267910e-01 -1.45968211e+00
-1.95605516e-01 -5.65738678e-01 6.36408897e-03 1.40484762e+00
2.44671598e-01 -6.25223756e-01 7.82081008e-01 3.80681366e-01
-7.13965893e-02 -5.35313129e-01 -5.90784907e-01 -9.31358576e-01
4.18800414e-01 -2.16154233e-01 7.06569493e-01 1.02019632e+00
-4.45863992e-01 5.18663704e-01 -5.17021775e-01 -3.63455266e-02
6.09707355e-01 5.18623054e-01 1.03522110e+00 -1.07947361e+00
-6.45400524e-01 -4.29060489e-01 -3.00393939e-01 -1.21962810e+00
4.75636691e-01 -1.01300907e+00 -2.33100913e-02 -1.16900194e+00
4.12379563e-01 -2.01832294e-01 -1.23410448e-01 4.58582431e-01
6.00356497e-02 6.78115606e-01 3.88642311e-01 4.11525875e-01
-2.53130615e-01 5.49426496e-01 1.15115666e+00 -2.06943020e-01
2.37723589e-01 -1.56675145e-01 -9.92986798e-01 8.10029328e-01
7.09908128e-01 -4.62073922e-01 -4.83037323e-01 -7.40395069e-01
1.52298808e-01 -2.18545407e-01 5.70549726e-01 -7.27957606e-01
-4.60537970e-02 -3.12898368e-01 3.17289419e-02 3.19705680e-02
3.67115438e-01 -6.44358039e-01 2.20850751e-01 2.67131269e-01
-3.87639016e-01 7.57441595e-02 -3.40789417e-03 4.56369489e-01
-3.09130251e-01 -2.93986648e-01 9.55123544e-01 -2.75374562e-01
-7.70032823e-01 2.75501788e-01 -3.59212495e-02 1.99990675e-01
1.02166343e+00 -2.12161422e-01 -2.68623739e-01 -5.93359351e-01
-7.04696774e-01 -2.44821131e-01 1.06432581e+00 3.99957597e-01
3.81841511e-01 -1.42452395e+00 -7.18641102e-01 3.67348313e-01
2.90482521e-01 1.57648493e-02 -5.22599816e-02 8.53952587e-01
-2.42196143e-01 3.05446059e-01 -4.78564978e-01 -7.03361511e-01
-1.10600913e+00 6.20623529e-01 3.27067971e-01 -2.60059416e-01
-4.98112023e-01 6.93092167e-01 7.84250438e-01 -3.99739236e-01
-1.44214496e-01 -4.54552658e-02 2.65019536e-01 -3.04305665e-02
1.74357563e-01 3.24023771e-03 -1.31320298e-01 -8.13221216e-01
-2.02540874e-01 9.17416632e-01 1.29317820e-01 -4.06701118e-01
1.35568953e+00 -2.22691238e-01 -2.99974889e-01 4.25524354e-01
1.23468864e+00 1.20002910e-01 -1.43933499e+00 -2.77057081e-01
-1.45297676e-01 -4.63676453e-01 -3.71280193e-01 -5.19454420e-01
-1.16169357e+00 1.11987078e+00 3.05835009e-01 9.25708786e-02
1.19960320e+00 3.75149518e-01 7.42369175e-01 1.11944109e-01
4.05575007e-01 -6.57608032e-01 2.00926796e-01 4.02186126e-01
8.02639902e-01 -1.18949485e+00 -1.96486875e-01 -3.66971105e-01
-7.66240537e-01 7.79509723e-01 4.06960100e-01 -2.37764701e-01
3.81255329e-01 1.12703897e-01 -6.62743747e-02 -4.62339930e-02
-5.83899140e-01 -3.10320407e-01 2.83913493e-01 9.35799003e-01
3.34129602e-01 -5.78541160e-02 7.98723996e-02 3.03046554e-01
-4.44201648e-01 -9.99140218e-02 5.04724741e-01 7.41989613e-01
-1.33562744e-01 -1.52669764e+00 -1.68362156e-01 8.06928985e-03
-2.21118584e-01 -1.84219226e-01 -6.32271886e-01 9.03776586e-01
4.40434963e-02 6.24973357e-01 1.88004822e-01 -3.87693077e-01
1.05901986e-01 2.07954884e-01 9.37093198e-01 -7.01681495e-01
-1.12759456e-01 -6.49005324e-02 -1.31676242e-01 -4.23359632e-01
-6.88985527e-01 -5.80538988e-01 -8.69155288e-01 -3.34022969e-01
5.99293560e-02 -1.29939362e-01 6.23381555e-01 1.05298591e+00
5.09021938e-01 3.77172321e-01 4.62128550e-01 -9.34814036e-01
-5.00848293e-01 -7.76647687e-01 -4.95408446e-01 6.72699988e-01
4.54678714e-01 -6.20129287e-01 -1.58663273e-01 4.85868514e-01] | [11.579248428344727, -0.19636306166648865] |
1298ed86-7164-450a-a021-c04bfa759d5d | adcraft-an-advanced-reinforcement-learning | 2306.11971 | null | https://arxiv.org/abs/2306.11971v2 | https://arxiv.org/pdf/2306.11971v2.pdf | AdCraft: An Advanced Reinforcement Learning Benchmark Environment for Search Engine Marketing Optimization | We introduce AdCraft, a novel benchmark environment for the Reinforcement Learning (RL) community distinguished by its stochastic and non-stationary properties. The environment simulates bidding and budgeting dynamics within Search Engine Marketing (SEM), a digital marketing technique utilizing paid advertising to enhance the visibility of websites on search engine results pages (SERPs). The performance of SEM advertisement campaigns depends on several factors, including keyword selection, ad design, bid management, budget adjustments, and performance monitoring. Deep RL recently emerged as a potential strategy to optimize campaign profitability within the complex and dynamic landscape of SEM but it requires substantial data, which may be costly or infeasible to acquire in practice. Our customizable environment enables practitioners to assess and enhance the robustness of RL algorithms pertinent to SEM bid and budget management without such costs. Through a series of experiments within the environment, we demonstrate the challenges imposed by sparsity and non-stationarity on agent convergence and performance. We hope these challenges further encourage discourse and development around effective strategies for managing real-world uncertainties. | ['Jonah White', 'Jeffrey Roach', 'Owen Levin', 'Maziar Gomrokchi'] | 2023-06-21 | null | null | null | null | ['marketing', 'management'] | ['miscellaneous', 'miscellaneous'] | [-1.75404489e-01 -1.66284770e-01 -5.96169889e-01 -1.33558899e-01
-9.31874514e-01 -8.07197452e-01 6.62289202e-01 1.71567231e-01
-5.75136006e-01 7.13918686e-01 3.61725599e-01 -5.93220055e-01
-5.60128570e-01 -6.24072134e-01 -6.01786852e-01 -4.06566232e-01
-5.94709158e-01 6.44051254e-01 -2.06835002e-01 -4.70849395e-01
7.55239129e-02 2.86781043e-01 -1.41853118e+00 -1.73005417e-01
5.07938504e-01 1.14859629e+00 1.57521889e-01 4.02099937e-01
1.62971482e-01 7.28386760e-01 -6.70024872e-01 -2.59988695e-01
7.77389646e-01 -1.04576558e-01 -1.36048183e-01 2.02419892e-01
-8.04919824e-02 -3.85420710e-01 -3.45987111e-01 9.39785838e-01
5.90535104e-01 3.49048302e-02 3.08705628e-01 -1.13785887e+00
-3.50844562e-01 9.06696975e-01 -6.42552912e-01 5.28386772e-01
-7.21995533e-02 4.69017148e-01 1.29394233e+00 -2.15951800e-01
4.61966455e-01 1.13697481e+00 4.34023589e-01 -1.26994506e-01
-1.39704025e+00 -6.95929170e-01 5.55339634e-01 2.37663969e-01
-9.35446620e-01 -4.49153811e-01 5.82900167e-01 -2.50000566e-01
5.62279880e-01 2.49752879e-01 9.90227222e-01 1.17276073e+00
5.08983552e-01 7.73590386e-01 1.37805057e+00 -1.69269428e-01
6.00995719e-01 2.24223360e-01 -1.81442156e-01 2.67087489e-01
2.31633931e-01 8.29938054e-01 -6.41542733e-01 -4.71361369e-01
5.57395101e-01 -4.50131744e-01 1.78470686e-01 -5.43875933e-01
-8.58102560e-01 1.37810326e+00 2.77406543e-01 -3.89105558e-01
-8.34936678e-01 4.32138145e-01 1.88900009e-01 7.08237410e-01
3.43569428e-01 8.37283373e-01 -5.74368000e-01 -4.28755492e-01
-7.31427908e-01 6.46638453e-01 1.05480659e+00 7.58603036e-01
3.49541336e-01 3.02953035e-01 -5.36983572e-02 5.91215014e-01
3.83911103e-01 5.74405074e-01 3.84928405e-01 -1.15810502e+00
3.41517776e-01 2.16486543e-01 5.69794714e-01 -9.97116029e-01
-4.54046100e-01 -9.83242333e-01 -6.74816072e-02 -2.82939151e-02
3.54075521e-01 -4.86579418e-01 -5.94698966e-01 1.81210160e+00
3.59884262e-01 -1.89106166e-01 -1.28208935e-01 9.29067969e-01
1.00600660e-01 3.56358677e-01 1.83392018e-01 -5.29605329e-01
1.23934817e+00 -5.83488882e-01 -6.14458799e-01 -6.52365327e-01
2.49201089e-01 -7.98923373e-01 1.10802853e+00 3.58629137e-01
-1.32695925e+00 1.64400950e-01 -8.96141171e-01 8.56440842e-01
-2.71475583e-01 -3.61497074e-01 1.16955411e+00 6.06393874e-01
-6.46404386e-01 2.67209828e-01 -8.22920322e-01 -5.10499179e-02
4.03534561e-01 4.63023335e-01 5.86389303e-01 2.36192062e-01
-1.33293176e+00 1.03400898e+00 -3.63623053e-01 5.24158031e-02
-1.24794686e+00 -8.25772047e-01 -4.67688709e-01 2.46988550e-01
1.04811525e+00 -3.78895789e-01 1.75084186e+00 -7.96422660e-01
-1.57751346e+00 1.54901713e-01 4.91348982e-01 -8.04631770e-01
7.14507937e-01 1.35155380e-01 -4.68423396e-01 -2.27579400e-01
1.00093246e-01 2.41808593e-01 9.24848974e-01 -9.56233799e-01
-7.01832116e-01 -4.92873490e-01 3.02159768e-02 5.01957953e-01
-7.63887018e-02 -7.07562864e-02 -6.57804012e-02 -6.11398816e-01
-2.76082307e-01 -1.04887259e+00 -6.47743881e-01 -8.69460523e-01
-5.45465201e-03 -9.86748859e-02 3.52666676e-01 -4.72966254e-01
1.10999346e+00 -1.74756002e+00 -6.80555403e-02 5.91052890e-01
3.59971188e-02 -3.28545779e-01 -1.77841991e-01 5.95293462e-01
4.42790627e-01 3.23779210e-02 4.92537558e-01 3.17410439e-01
4.36869740e-01 4.35629487e-02 -3.19106311e-01 3.67298007e-01
-1.37517571e-01 1.00072634e+00 -7.14887619e-01 -1.11452632e-01
-1.49229825e-01 -1.24459334e-01 -6.50578022e-01 -3.47917937e-02
-7.86447823e-01 6.46807700e-02 -7.30910420e-01 9.68467176e-01
1.83898032e-01 -4.53408241e-01 6.40656769e-01 2.06067756e-01
-1.68276981e-01 3.79388660e-01 -1.39616787e+00 1.01027644e+00
-6.22352362e-01 3.38279068e-01 5.47498047e-01 -8.52220058e-01
4.99706298e-01 -2.72234857e-01 6.00658953e-01 -1.40344095e+00
2.33733729e-01 1.48038402e-01 2.60623664e-01 -2.03651890e-01
6.52145326e-01 2.50976741e-01 -2.57974535e-01 9.28109050e-01
-5.98502040e-01 -1.95882935e-02 4.71190751e-01 1.22531965e-01
1.27064359e+00 -1.71546683e-01 -1.07274065e-02 -5.72844148e-01
-1.24983720e-01 4.55226839e-01 4.82031107e-01 1.05085194e+00
-2.15926155e-01 -4.18483168e-01 6.02328181e-01 -3.75789165e-01
-9.34454322e-01 -9.69304442e-01 -1.45157486e-01 1.38891518e+00
7.96373338e-02 -1.21298447e-01 -2.95131266e-01 -3.52445215e-01
6.48082733e-01 7.26750076e-01 -5.22394657e-01 -1.26458853e-01
-4.19212013e-01 -1.16922283e+00 -1.44232795e-01 1.59728751e-01
3.63686383e-01 -9.04756963e-01 -8.82336438e-01 6.52049601e-01
1.35478094e-01 -9.65039372e-01 -7.54299045e-01 2.99961090e-01
-4.79136556e-01 -1.05373836e+00 -4.02839184e-01 -4.00741160e-01
2.36208752e-01 2.30166703e-01 1.34528422e+00 -3.77920091e-01
-2.34520137e-01 9.55039263e-01 -1.12314685e-03 -6.59317672e-01
-3.98315132e-01 5.39002456e-02 1.11107655e-01 -1.92039818e-01
2.35378399e-01 -3.94215137e-01 -1.04807854e+00 6.94694042e-01
-7.11938560e-01 -4.80964065e-01 7.24893868e-01 7.82090485e-01
6.09982431e-01 3.18906844e-01 1.05822325e+00 -6.83625162e-01
1.42578495e+00 -6.73946023e-01 -1.48284888e+00 2.54452407e-01
-1.24089456e+00 1.10070914e-01 1.19655907e-01 -7.10731864e-01
-8.31888497e-01 -2.79672027e-01 7.56581724e-01 1.18859345e-02
6.79052114e-01 9.35690284e-01 3.26494157e-01 -9.89490524e-02
6.12413704e-01 -5.22588976e-02 5.02176940e-01 -2.97154099e-01
3.55506748e-01 5.36808133e-01 6.30863905e-02 -5.15700042e-01
5.88016987e-01 2.69334644e-01 -8.73354971e-02 -5.64171672e-01
-6.70107007e-01 -7.29106367e-02 5.17616987e-01 -2.24377960e-01
2.15973645e-01 -1.14089024e+00 -1.20669866e+00 -7.36384764e-02
-3.69443387e-01 -6.53288901e-01 -4.11364555e-01 4.91969407e-01
-6.08339489e-01 -1.02875814e-01 -4.31500465e-01 -8.26564729e-01
-1.22834668e-01 -1.36501777e+00 5.88267684e-01 2.52226472e-01
-1.48624167e-01 -8.42789650e-01 1.21676870e-01 5.60767829e-01
8.79597664e-01 -4.63844910e-02 9.27780092e-01 -5.61259329e-01
-8.44331086e-01 -1.57732815e-01 1.03596516e-01 -1.57621577e-02
-1.61598817e-01 -3.07344615e-01 -2.65358448e-01 -4.79585707e-01
-9.29314196e-02 -3.52098376e-01 3.67592275e-01 7.23013878e-01
8.86971831e-01 -5.81684768e-01 -6.50087148e-02 1.97242230e-01
1.20993590e+00 3.09970886e-01 1.45526160e-03 1.00371230e+00
-2.18256101e-01 5.80945432e-01 9.18326080e-01 9.02458668e-01
7.08767116e-01 9.09535766e-01 6.37167275e-01 1.16883263e-01
3.21118474e-01 -2.96628982e-01 3.67649645e-01 3.72891992e-01
3.83526772e-01 -1.04809999e-01 -5.99900186e-01 3.56390148e-01
-1.76187313e+00 -8.06112766e-01 5.54876328e-01 2.26245236e+00
1.06637025e+00 2.71355033e-01 5.32339573e-01 -4.57317263e-01
4.29480016e-01 1.58111900e-01 -1.06383657e+00 -4.88455713e-01
-1.79136172e-01 7.20586907e-03 1.17043352e+00 3.58984113e-01
-6.78297639e-01 8.44446540e-01 7.05182028e+00 5.76144695e-01
-7.52056360e-01 -3.65701802e-02 6.12463892e-01 -5.60605109e-01
-5.42283654e-01 -9.40944105e-02 -7.65185416e-01 6.24492645e-01
1.23821163e+00 -6.16741061e-01 1.19591188e+00 8.31198215e-01
7.26016581e-01 -2.64893204e-01 -8.41872156e-01 6.44992590e-01
-3.30981344e-01 -1.68073761e+00 -4.99427944e-01 6.16359711e-01
7.51725197e-01 3.06302130e-01 5.13115048e-01 5.48309743e-01
1.23504639e+00 -8.72348249e-01 7.08520055e-01 1.41474977e-01
2.53377318e-01 -9.45085168e-01 5.16326249e-01 8.50104839e-02
-7.15731263e-01 -7.50854790e-01 -1.16435520e-01 6.56733885e-02
1.79534808e-01 3.27754468e-01 -8.89822185e-01 -1.61078721e-01
6.34632766e-01 1.50152624e-01 -4.24270034e-01 8.12012136e-01
1.54394627e-01 6.61291480e-01 -4.70147222e-01 -4.91446048e-01
6.09651685e-01 -4.29985464e-01 5.81344068e-01 5.03978431e-01
1.15014389e-01 -1.69463187e-01 4.52196181e-01 7.29428947e-01
-5.34743816e-02 5.87586202e-02 -2.70934582e-01 -6.11913860e-01
7.33692944e-01 1.13051331e+00 -4.22334909e-01 1.73718408e-01
-2.72783875e-01 -7.59979934e-02 -1.00717835e-01 5.54299414e-01
-8.83423090e-01 2.62881070e-01 8.82504523e-01 2.90315926e-01
4.27045196e-01 -3.85967046e-01 -1.16974413e-01 -7.18986094e-01
-1.24223597e-01 -1.66368735e+00 6.39019787e-01 -3.01667571e-01
-1.11010516e+00 2.82160133e-01 1.20285891e-01 -8.40667427e-01
-5.07492185e-01 -2.55534768e-01 -8.55123922e-02 3.20425391e-01
-1.72974646e+00 -6.75761104e-01 1.29778665e-02 4.18642253e-01
5.58796108e-01 -5.60514152e-01 2.93416083e-01 7.73972720e-02
-5.34779668e-01 4.00100559e-01 3.78095150e-01 -7.13587463e-01
3.61286968e-01 -9.31565225e-01 -1.50280008e-02 3.00385743e-01
1.03571862e-01 1.96956426e-01 1.00101054e+00 -7.61075675e-01
-2.01355314e+00 -6.93334043e-01 1.94390699e-01 -2.28974417e-01
1.36566210e+00 -2.83901185e-01 -9.50699747e-02 5.83772004e-01
2.05150291e-01 -6.85707033e-01 5.35509944e-01 3.63013804e-01
3.31065655e-02 -4.89953548e-01 -9.00081456e-01 9.79292512e-01
6.75303161e-01 3.02298423e-02 -3.64755094e-02 7.53065825e-01
6.68208003e-01 -2.99765885e-01 -8.51873457e-01 2.32686266e-01
5.34489453e-01 -5.33324778e-01 8.90236318e-01 -7.06596315e-01
-1.93308201e-02 2.27429792e-01 -2.67948031e-01 -1.47061098e+00
-8.58730152e-02 -1.15836394e+00 -4.81872819e-02 7.36832500e-01
6.46985590e-01 -8.65314305e-01 1.06562841e+00 6.65290475e-01
3.74103069e-01 -6.76892102e-01 -8.43806624e-01 -8.01076829e-01
-6.70947880e-02 -4.80784088e-01 7.90238917e-01 3.84828866e-01
-2.56735355e-01 1.40293539e-01 -3.30473155e-01 2.60581747e-02
9.54677165e-01 4.01598006e-01 7.50807762e-01 -9.48530376e-01
-6.30514681e-01 -5.98525763e-01 7.80598894e-02 -7.72759736e-01
-1.47801563e-02 -5.47677636e-01 -1.83562607e-01 -9.37772572e-01
-1.25115782e-01 -7.59507358e-01 -2.85805762e-01 1.87000539e-02
3.85323346e-01 -1.43579692e-01 2.80193508e-01 3.68218631e-01
-7.13926017e-01 6.41062200e-01 1.13563752e+00 -2.75806069e-01
-5.98109066e-01 5.09321928e-01 -1.10785091e+00 4.84164566e-01
8.24117362e-01 -2.89098531e-01 -5.58019817e-01 -2.30777279e-01
5.94682276e-01 3.93096179e-01 1.75678417e-01 -3.49918932e-01
1.86324850e-01 -6.67631209e-01 5.15420735e-03 -5.31232655e-01
2.13800490e-01 -8.75938952e-01 2.82318771e-01 5.10809600e-01
-6.97566390e-01 6.48579538e-01 1.22075245e-01 9.58616555e-01
4.97356690e-02 1.54178008e-01 5.19602180e-01 -3.00858796e-01
-5.27523518e-01 3.08243722e-01 -6.12056971e-01 3.15965593e-01
1.16090643e+00 2.02704191e-01 -5.11357665e-01 -1.03707480e+00
-5.74363589e-01 9.83170092e-01 6.48676232e-02 3.91004205e-01
1.84170529e-01 -1.13860226e+00 -5.90964019e-01 -5.63144721e-02
-2.62310952e-01 -5.59322238e-01 -4.31231260e-02 7.72087634e-01
-2.07760587e-01 4.56761718e-01 9.26804021e-02 -3.32761228e-01
-8.36028337e-01 3.87164801e-01 3.63999158e-01 -6.14771903e-01
-3.14285755e-01 4.46735024e-01 -2.55641311e-01 -1.31502137e-01
6.18363261e-01 6.40608296e-02 -3.65963019e-02 3.05149496e-01
1.92643061e-01 4.37521130e-01 3.81964669e-02 9.25018862e-02
-8.86955336e-02 -9.00488570e-02 -2.85640121e-01 -4.26246494e-01
1.56513309e+00 -2.81690508e-01 4.07240719e-01 -6.77161962e-02
4.90031213e-01 -4.29786034e-02 -1.54719448e+00 -3.28839362e-01
4.37735587e-01 -4.38198924e-01 5.79397202e-01 -1.14496946e+00
-1.12824845e+00 -9.59740579e-02 5.62311172e-01 6.75970972e-01
8.23982775e-01 -1.46548703e-01 6.92751765e-01 3.72453928e-01
3.38684857e-01 -1.79254949e+00 2.51464367e-01 5.73400185e-02
8.61942053e-01 -1.17375529e+00 3.00237268e-01 1.52995884e-01
-8.96007657e-01 5.60510874e-01 3.82608742e-01 3.69354077e-02
8.42980444e-01 4.20477957e-01 1.71968371e-01 -3.82819474e-01
-1.26197386e+00 -1.97439089e-01 -1.50103018e-01 2.96373755e-01
-2.05386758e-01 2.97777563e-01 -5.47645986e-01 4.83560055e-01
-2.82780617e-01 -3.59251291e-01 6.19527519e-01 9.77000892e-01
-3.08372617e-01 -1.15934002e+00 -1.77731365e-01 7.22218871e-01
-4.11352694e-01 -1.00328527e-01 -1.64875388e-01 1.09638727e+00
-6.61578357e-01 1.07213295e+00 5.33369668e-02 -1.05117083e-01
4.46008682e-01 -3.67268831e-01 1.66382566e-01 -2.48115063e-01
-6.54719591e-01 6.14780605e-01 3.94220591e-01 -8.03647518e-01
-9.74618495e-02 -1.01698732e+00 -6.77914798e-01 -2.99154520e-01
-3.17095041e-01 4.40713435e-01 1.16876125e+00 5.83591521e-01
6.62504673e-01 4.04359549e-01 1.07927501e+00 -3.63671362e-01
-1.88697624e+00 -4.96759713e-01 -8.28521013e-01 -9.76480618e-02
5.29428050e-02 -9.76217449e-01 -7.01782227e-01 -6.63613200e-01] | [4.224656581878662, 2.6918954849243164] |
c5ba862e-0eaa-4802-8037-db889ad27bd9 | interpretable-computer-vision-models-through | 2307.02500 | null | https://arxiv.org/abs/2307.02500v1 | https://arxiv.org/pdf/2307.02500v1.pdf | Interpretable Computer Vision Models through Adversarial Training: Unveiling the Robustness-Interpretability Connection | With the perpetual increase of complexity of the state-of-the-art deep neural networks, it becomes a more and more challenging task to maintain their interpretability. Our work aims to evaluate the effects of adversarial training utilized to produce robust models - less vulnerable to adversarial attacks. It has been shown to make computer vision models more interpretable. Interpretability is as essential as robustness when we deploy the models to the real world. To prove the correlation between these two problems, we extensively examine the models using local feature-importance methods (SHAP, Integrated Gradients) and feature visualization techniques (Representation Inversion, Class Specific Image Generation). Standard models, compared to robust are more susceptible to adversarial attacks, and their learned representations are less meaningful to humans. Conversely, these models focus on distinctive regions of the images which support their predictions. Moreover, the features learned by the robust model are closer to the real ones. | ['Delyan Boychev'] | 2023-07-04 | null | null | null | null | ['image-generation', 'feature-importance'] | ['computer-vision', 'methodology'] | [ 2.02681005e-01 6.00889981e-01 3.92010272e-01 -2.51437426e-01
1.90012738e-01 -7.78876185e-01 9.86598015e-01 -1.92956969e-01
-3.57047945e-01 8.31413686e-01 -3.90121453e-02 -2.62797862e-01
-1.74940869e-01 -8.29775393e-01 -8.08558166e-01 -7.25683749e-01
-3.20362598e-01 8.53056684e-02 8.06219727e-02 -5.53077102e-01
3.79171640e-01 1.06680942e+00 -1.49864686e+00 2.71231115e-01
8.21711063e-01 8.09004426e-01 -3.05290639e-01 5.70813298e-01
1.52246773e-01 7.88373411e-01 -9.12984848e-01 -6.56699777e-01
4.91713315e-01 -2.95071334e-01 -6.95254028e-01 -2.95896173e-01
6.69375181e-01 -3.73033106e-01 -3.48870724e-01 1.53126383e+00
2.32249480e-02 -8.13146234e-02 8.85173857e-01 -1.63972068e+00
-1.41587687e+00 3.16834211e-01 -1.39209703e-01 6.67587891e-02
1.90730542e-01 3.72276217e-01 5.11032820e-01 -5.90476394e-01
7.77386010e-01 1.59239507e+00 7.48047888e-01 6.77118182e-01
-1.30400527e+00 -6.80035353e-01 1.80345118e-01 2.38117471e-01
-1.05515349e+00 -2.33985960e-01 9.98134315e-01 -2.47861102e-01
4.26158220e-01 7.25368142e-01 6.74027085e-01 1.48287177e+00
4.88893986e-01 1.47230700e-01 1.34453356e+00 -4.65338886e-01
2.37738296e-01 5.95472872e-01 6.91174343e-02 6.13231301e-01
6.09837294e-01 6.02883518e-01 -3.26163232e-01 8.24378580e-02
6.78010285e-01 9.40991044e-02 -4.40441489e-01 -5.60167611e-01
-7.63105571e-01 9.07979012e-01 8.35049391e-01 4.44887459e-01
-4.23552305e-01 2.86927372e-01 3.17380995e-01 6.45990670e-01
4.08449471e-01 1.00138199e+00 -1.56003460e-01 1.69438288e-01
-5.76375961e-01 2.78561205e-01 5.65879405e-01 5.97647369e-01
6.44817710e-01 4.37124997e-01 1.09005824e-01 3.31923999e-02
4.36534911e-01 4.58999962e-01 5.13923645e-01 -9.15385485e-01
9.24193487e-02 7.87748277e-01 -3.50508727e-02 -1.58065271e+00
-1.11384608e-01 -7.00995564e-01 -1.00256884e+00 1.10480237e+00
4.42762733e-01 8.14340338e-02 -8.50144923e-01 1.71528792e+00
3.16580869e-02 2.67268457e-02 2.91892320e-01 8.16783190e-01
5.10167301e-01 4.70023036e-01 9.49765518e-02 2.61336058e-01
9.55102324e-01 -6.25791669e-01 -8.76169503e-01 -1.19672403e-01
3.41863930e-01 -4.75507826e-01 1.04271138e+00 3.34095299e-01
-7.70717442e-01 -7.42923260e-01 -1.44001818e+00 1.41474813e-01
-9.77732837e-01 -3.30774218e-01 4.65827644e-01 9.98939872e-01
-1.09456670e+00 1.05791330e+00 -5.65136969e-01 -1.73324406e-01
7.40048468e-01 1.88625067e-01 -6.81005895e-01 3.35343570e-01
-1.34581244e+00 1.29124320e+00 5.82874715e-01 2.43325800e-01
-1.00305212e+00 -6.84649825e-01 -8.47342074e-01 1.04861580e-01
-2.08505198e-01 -4.79368389e-01 6.15361750e-01 -1.62507510e+00
-1.19749701e+00 8.29877377e-01 3.09051871e-01 -7.31637955e-01
1.04876840e+00 -4.03380573e-01 -2.98683733e-01 3.17738146e-01
-3.78307641e-01 6.21026039e-01 1.17276001e+00 -1.84306169e+00
-3.85798216e-02 -3.68572026e-01 2.98255622e-01 -3.00050110e-01
-5.59160352e-01 -1.59728572e-01 4.93415296e-01 -8.12283397e-01
4.10629772e-02 -8.42691898e-01 -2.33963192e-01 6.67067409e-01
-6.20442808e-01 3.06811273e-01 1.34767139e+00 -6.27469301e-01
7.74054408e-01 -2.32649803e+00 -5.82116144e-03 4.43824649e-01
4.09456432e-01 7.31304049e-01 -7.43559971e-02 2.69080192e-01
-7.59744048e-01 6.79748058e-01 -1.66387022e-01 -1.67604193e-01
1.25986394e-02 4.35950696e-01 -6.35941029e-01 5.80004275e-01
5.67203164e-01 7.54642010e-01 -7.29570806e-01 -3.09952885e-01
3.09352756e-01 6.60764694e-01 -2.84559280e-01 2.06335619e-01
1.82776541e-01 5.86834729e-01 -5.18547535e-01 2.19248846e-01
9.55013335e-01 3.57603490e-01 -1.11876793e-01 -3.11106235e-01
1.04490720e-01 -3.94014448e-01 -8.56341541e-01 1.00142050e+00
-2.79329032e-01 1.14235926e+00 -2.81945169e-01 -1.00620127e+00
1.08218503e+00 1.45878762e-01 -2.24689871e-01 -6.93118215e-01
2.11921424e-01 -1.32437265e-02 1.72039479e-01 -5.64399540e-01
3.38803887e-01 -3.53034921e-02 3.43011111e-01 3.09129626e-01
-1.74365565e-01 -1.24141209e-01 -1.78072646e-01 2.47206882e-01
7.60920882e-01 1.33734405e-01 2.39947826e-01 -4.74402070e-01
5.99259555e-01 -1.18350141e-01 2.90671319e-01 7.59709179e-01
-3.00714761e-01 6.75715327e-01 7.79073238e-01 -8.54814768e-01
-1.07084227e+00 -1.28779876e+00 -3.81552242e-02 4.40589160e-01
1.07629918e-01 4.30547670e-02 -1.19964457e+00 -7.89334178e-01
1.61431637e-02 1.12273002e+00 -1.17634273e+00 -8.20053458e-01
-4.08977151e-01 -3.08533400e-01 9.27083552e-01 3.17843080e-01
6.35325789e-01 -1.23349297e+00 -1.03879976e+00 -2.08705291e-01
3.49618495e-01 -7.74063587e-01 2.55968988e-01 -4.07490246e-02
-8.69456291e-01 -1.16536307e+00 -4.06441718e-01 -2.64318973e-01
1.09104097e+00 -2.08029509e-01 1.28924179e+00 5.19577086e-01
-2.69655854e-01 2.66617030e-01 -5.30516505e-01 -8.29898059e-01
-1.04260945e+00 -3.28201979e-01 1.16116039e-01 2.09347188e-01
-1.23492233e-01 -7.38742948e-01 -4.32095200e-01 1.80166885e-01
-1.41414440e+00 -1.81863189e-01 4.48321939e-01 7.29810953e-01
1.63640231e-01 -2.04015747e-01 4.71071422e-01 -9.98454273e-01
9.15214956e-01 -1.73924536e-01 -3.20249408e-01 4.03130740e-01
-7.16161311e-01 1.83357120e-01 1.15912879e+00 -6.06360793e-01
-1.05203724e+00 -1.55193657e-01 1.38232991e-01 -6.30068243e-01
-2.88938671e-01 2.03811765e-01 -2.35081986e-01 -4.76330161e-01
1.13238955e+00 7.89019540e-02 3.38602245e-01 -3.32600832e-01
4.92461801e-01 2.56415427e-01 4.46927309e-01 -4.04352158e-01
1.45701420e+00 5.62006533e-01 2.40550667e-01 -7.98976421e-01
-4.16963339e-01 5.71954131e-01 -8.24323416e-01 -3.66787165e-01
7.18231201e-01 -2.94936478e-01 -5.21018803e-01 5.33631444e-01
-1.34398973e+00 -2.95949467e-02 -6.85666800e-01 3.10196280e-02
-3.88508797e-01 4.46444839e-01 -7.79901817e-02 -8.57664883e-01
-3.01303983e-01 -1.22181273e+00 3.97313923e-01 3.23875606e-01
-3.55701953e-01 -1.24896812e+00 -1.32328734e-01 -5.25404811e-02
6.72673404e-01 9.01493907e-01 1.09129965e+00 -9.00534332e-01
-4.36485916e-01 -3.82509857e-01 -2.52299905e-01 7.96447694e-01
1.41085997e-01 5.50231516e-01 -1.42735565e+00 -2.00524643e-01
2.37947777e-01 -1.69033855e-01 6.22700870e-01 -4.99622375e-02
1.36775172e+00 -7.84381926e-01 -6.16290681e-02 6.66814268e-01
1.15581346e+00 2.48793572e-01 9.88518476e-01 5.10661125e-01
5.46861112e-01 8.22936654e-01 5.40518701e-01 -5.91904633e-02
-4.10500467e-01 4.03892219e-01 1.08183825e+00 -4.53751832e-01
-7.42228404e-02 -2.81627566e-01 3.68533134e-01 1.84477925e-01
-3.36229831e-01 -1.14691891e-01 -7.03635275e-01 1.79838017e-01
-1.72469389e+00 -9.64886427e-01 -1.48009241e-01 1.99751210e+00
4.81600702e-01 4.04489905e-01 -3.93310040e-01 2.77861029e-01
6.21914089e-01 1.39869496e-01 -5.31213939e-01 -8.41891885e-01
-1.66417286e-01 1.72363073e-01 2.67623156e-01 2.90661097e-01
-9.43835855e-01 6.42205179e-01 6.67822790e+00 6.92457616e-01
-1.26361489e+00 -1.14275992e-01 1.08476484e+00 1.17483243e-01
-5.36331594e-01 -6.14797547e-02 1.10212103e-01 4.11742002e-01
9.41145182e-01 -1.42657205e-01 1.97594881e-01 1.10820472e+00
2.19521806e-01 1.80670619e-01 -1.31065726e+00 6.77733660e-01
8.45920071e-02 -1.49589574e+00 4.58301723e-01 1.12038918e-01
5.04816830e-01 -5.32746375e-01 5.40047526e-01 7.89881796e-02
1.50572866e-01 -1.48547328e+00 1.11509466e+00 8.99764657e-01
6.41761303e-01 -8.27042282e-01 9.66788530e-01 9.80906337e-02
-6.13570988e-01 -7.38206506e-02 -4.55123276e-01 -6.74783438e-02
-3.03404331e-01 2.20488921e-01 -6.98435545e-01 4.51922923e-01
7.28983223e-01 3.50952715e-01 -1.17948782e+00 6.05655134e-01
-3.92794609e-01 3.31682831e-01 5.71252815e-02 2.31461581e-02
2.36328736e-01 -2.16518149e-01 6.98044956e-01 8.17255437e-01
1.12985469e-01 -4.54118788e-01 -3.44561994e-01 1.35933816e+00
1.40796393e-01 -2.66243935e-01 -1.23311949e+00 -3.02871205e-02
3.02272826e-01 1.17474067e+00 -6.95826530e-01 -1.98306620e-01
1.41187459e-01 9.29260612e-01 2.88967252e-01 4.41399962e-01
-8.90260100e-01 -3.90262574e-01 7.86534369e-01 -9.90089308e-03
-2.25747690e-01 -2.49024313e-02 -5.70450842e-01 -7.53778219e-01
-5.14382459e-02 -1.16534758e+00 -1.60628662e-01 -9.11005259e-01
-1.32755816e+00 8.88132155e-01 -2.21983623e-02 -1.19880295e+00
-3.33467633e-01 -9.21230674e-01 -1.02125275e+00 9.38381970e-01
-1.30924368e+00 -1.17991865e+00 -2.12029591e-01 6.98800385e-01
1.41379610e-01 -3.25163037e-01 1.06893146e+00 -1.34812966e-01
-4.67464626e-01 8.84895563e-01 1.04090311e-01 2.18708321e-01
2.91241854e-01 -1.13864470e+00 5.46696842e-01 1.08217871e+00
1.32842153e-01 8.97405505e-01 1.08386385e+00 -2.67380983e-01
-9.97607172e-01 -9.99949932e-01 3.73209476e-01 -5.86001635e-01
5.73186338e-01 -2.74181008e-01 -1.12926459e+00 5.95544636e-01
3.06658626e-01 1.13616966e-01 3.73496771e-01 -2.15230182e-01
-5.46236277e-01 -2.15839103e-01 -1.65893817e+00 9.02382314e-01
9.26659882e-01 -5.23357570e-01 -7.69825459e-01 7.41958842e-02
7.67118037e-01 6.04732335e-02 -6.29052877e-01 5.99075332e-02
5.52547634e-01 -1.41075015e+00 1.12252319e+00 -8.99694979e-01
4.49588567e-01 -1.59531713e-01 1.62316889e-01 -1.56785405e+00
-1.83653221e-01 -5.69753647e-01 1.14410564e-01 1.24590147e+00
5.01234293e-01 -1.10136056e+00 6.16040051e-01 8.30882072e-01
-4.88089491e-03 -5.57114899e-01 -1.10049880e+00 -9.06772733e-01
1.63182542e-01 -1.26450369e-02 5.78881323e-01 1.14814270e+00
-4.10095006e-01 -2.61534214e-01 -1.59057200e-01 2.71966070e-01
5.86479545e-01 -2.60113508e-01 7.32645988e-01 -1.36877286e+00
9.59072039e-02 -3.81060004e-01 -9.47794378e-01 2.53976882e-02
2.27906600e-01 -4.55888242e-01 -3.37884724e-01 -1.06635427e+00
-2.86080301e-01 -2.23527685e-01 -3.50340307e-01 6.25964105e-01
5.94542921e-02 4.16519165e-01 6.39152288e-01 5.20104319e-02
-9.54158977e-02 6.07842565e-01 1.39881325e+00 -2.02650085e-01
2.48792544e-01 -1.34777248e-01 -8.34727764e-01 1.03501952e+00
9.23224151e-01 -6.45307720e-01 -5.69061995e-01 -2.76180953e-01
2.18356863e-01 -5.18318951e-01 8.08538437e-01 -1.19334853e+00
-2.74581492e-01 -7.61437640e-02 7.16885984e-01 5.19872792e-02
2.86597461e-01 -1.19677901e+00 3.18036884e-01 9.70328689e-01
-3.67418110e-01 3.06813836e-01 3.67540330e-01 3.62985045e-01
-2.26083204e-01 -7.19138086e-01 1.00250506e+00 -1.04673490e-01
-5.12237787e-01 -5.13792895e-02 -3.91826987e-01 -1.62191734e-01
1.24516940e+00 -6.10808849e-01 -3.58514041e-01 -5.31986177e-01
-7.72344291e-01 -4.25134391e-01 5.40476263e-01 5.20527005e-01
8.10379684e-01 -1.15224123e+00 -5.96408665e-01 4.04014051e-01
-1.29719093e-01 -9.89443436e-02 2.01423213e-01 2.76476711e-01
-1.08006644e+00 1.03433736e-01 -8.80514801e-01 -2.64664203e-01
-1.30551875e+00 6.80834711e-01 5.90498090e-01 -6.68903813e-02
-4.84514296e-01 6.88714445e-01 3.03823441e-01 -3.28395307e-01
8.71235430e-02 -2.15847075e-01 -3.00032496e-01 -1.18494965e-01
4.84765410e-01 3.19676429e-01 -1.66240796e-01 -4.59032416e-01
-1.84148937e-01 2.04296559e-01 -2.62098700e-01 2.21689448e-01
1.36090970e+00 3.21302623e-01 -2.05992386e-01 8.34054202e-02
9.89176095e-01 -1.09264486e-01 -1.41204369e+00 2.70344615e-01
-1.57378718e-01 -7.40004897e-01 -1.56459332e-01 -7.64623165e-01
-1.27268541e+00 1.05852783e+00 1.01582766e+00 8.02969635e-01
1.06965458e+00 -5.10465860e-01 -2.09123567e-02 4.77599084e-01
1.59209028e-01 -7.45277047e-01 1.97310477e-01 2.46054128e-01
1.44024277e+00 -1.01479614e+00 2.01629810e-02 -4.52737629e-01
-5.44741929e-01 1.35610449e+00 6.33973897e-01 -4.26017582e-01
5.76884925e-01 -1.94394458e-02 2.88451731e-01 -1.30013570e-01
-2.52054036e-01 4.70817804e-01 2.53894389e-01 1.15434372e+00
-3.40304784e-02 -1.73454449e-01 1.22304000e-02 3.36645067e-01
-5.07827759e-01 -4.16838527e-01 7.08719730e-01 6.32831752e-01
-4.15452980e-02 -8.90199125e-01 -7.15057433e-01 6.98764175e-02
-3.57218027e-01 -4.07408550e-02 -7.71075189e-01 1.15861070e+00
1.99838653e-01 8.98062289e-01 -2.97352858e-02 -4.09407735e-01
3.27964425e-01 2.03959774e-02 1.37580499e-01 -1.63483217e-01
-8.64803493e-01 -7.81171024e-01 -4.75266203e-02 -6.38560534e-01
-3.48334283e-01 -3.35319519e-01 -9.98606026e-01 -5.12488067e-01
-1.09789386e-01 -3.29352058e-02 7.97898471e-01 8.63248944e-01
5.04240632e-01 7.86224544e-01 6.81182921e-01 -9.42433774e-01
-8.60639453e-01 -8.59296918e-01 -4.31305766e-01 9.63558078e-01
3.28477502e-01 -6.93045974e-01 -6.56317353e-01 1.33981079e-01] | [5.685698509216309, 7.843221664428711] |
b4b83611-08dd-4851-93a8-e7ba30e3a765 | human-pose-estimation-in-monocular | 2304.08186 | null | https://arxiv.org/abs/2304.08186v1 | https://arxiv.org/pdf/2304.08186v1.pdf | Human Pose Estimation in Monocular Omnidirectional Top-View Images | Human pose estimation (HPE) with convolutional neural networks (CNNs) for indoor monitoring is one of the major challenges in computer vision. In contrast to HPE in perspective views, an indoor monitoring system can consist of an omnidirectional camera with a field of view of 180{\deg} to detect the pose of a person with only one sensor per room. To recognize human pose, the detection of keypoints is an essential upstream step. In our work we propose a new dataset for training and evaluation of CNNs for the task of keypoint detection in omnidirectional images. The training dataset, THEODORE+, consists of 50,000 images and is created by a 3D rendering engine, where humans are randomly walking through an indoor environment. In a dynamically created 3D scene, persons move randomly with simultaneously moving omnidirectional camera to generate synthetic RGB images and 2D and 3D ground truth. For evaluation purposes, the real-world PoseFES dataset with two scenarios and 701 frames with up to eight persons per scene was captured and annotated. We propose four training paradigms to finetune or re-train two top-down models in MMPose and two bottom-up models in CenterNet on THEODORE+. Beside a qualitative evaluation we report quantitative results. Compared to a COCO pretrained baseline, we achieve significant improvements especially for top-view scenes on the PoseFES dataset. Our datasets can be found at https://www.tu-chemnitz.de/etit/dst/forschung/comp_vision/datasets/index.php.en. | ['Gangolf Hirtz', 'Dipankar Nandi', 'Yukti Adya', 'Roman Seidel', 'Tobias Scheck', 'Jingrui Yu'] | 2023-04-17 | null | null | null | null | ['keypoint-detection', '3d-human-pose-estimation', '2d-human-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.65924777e-02 -1.07586280e-01 5.86151242e-01 -2.93996006e-01
-3.10558259e-01 -3.55910540e-01 4.12446290e-01 -2.84573197e-01
-8.57024670e-01 4.64943677e-01 5.24110459e-02 -4.31850180e-02
1.74499273e-01 -8.49658787e-01 -1.03451431e+00 -4.81847554e-01
-5.94934635e-03 6.46104336e-01 -5.99840358e-02 -1.05091602e-01
-4.34155703e-01 6.44119501e-01 -1.58336949e+00 3.29066850e-02
8.03603157e-02 9.91201580e-01 3.06517184e-01 1.19172442e+00
8.35606694e-01 2.66271472e-01 -7.46410131e-01 -3.71414721e-01
7.98453867e-01 1.61006898e-01 -2.36400038e-01 2.52465099e-01
9.84491825e-01 -4.66743320e-01 -6.72022462e-01 9.17080283e-01
9.94647264e-01 2.15997517e-01 7.83252567e-02 -1.28331316e+00
-1.32213254e-02 -2.86203511e-02 -3.52119893e-01 1.33956596e-01
8.80792797e-01 2.48866215e-01 5.15187860e-01 -9.46517766e-01
5.94275713e-01 1.26723444e+00 8.78781736e-01 6.10085070e-01
-7.08748043e-01 -4.22286272e-01 1.56737357e-01 1.93894789e-01
-1.42961442e+00 -2.85283208e-01 6.97067916e-01 -2.76935816e-01
7.12131441e-01 5.11301696e-01 1.14426887e+00 1.51374447e+00
4.39889356e-02 6.39571786e-01 7.98938215e-01 -1.24390058e-01
1.53381899e-01 1.47532836e-01 -1.72792792e-01 8.48307967e-01
4.90303367e-01 1.84218273e-01 -4.28930700e-01 2.33512416e-01
1.06573069e+00 5.03609896e-01 -5.43614566e-01 -7.07976520e-01
-1.51467240e+00 3.62481415e-01 9.22437131e-01 -6.83910549e-02
-3.44085455e-01 1.33804813e-01 2.38434494e-01 1.91802859e-01
8.81129056e-02 3.37941736e-01 -5.45160651e-01 4.78040799e-03
-5.31633973e-01 5.68057299e-01 5.55772305e-01 1.07075310e+00
3.47836614e-01 -1.56576559e-01 -7.83548653e-02 6.39238298e-01
2.53802240e-01 9.35498357e-01 5.29377460e-02 -9.67151284e-01
6.02805138e-01 6.32930160e-01 5.39134324e-01 -1.17645621e+00
-8.00611913e-01 -7.51706839e-01 -9.12968338e-01 8.70554373e-02
4.66890067e-01 -4.06065315e-01 -8.03891778e-01 1.40032518e+00
5.64225495e-01 -8.43051746e-02 -1.79819986e-01 1.31099403e+00
9.82815802e-01 4.85461324e-01 -3.72689754e-01 2.59389430e-01
1.55489087e+00 -1.09524071e+00 -2.32211605e-01 -5.25451541e-01
1.86006933e-01 -3.10470492e-01 8.71301949e-01 6.65471733e-01
-9.17214811e-01 -9.02876675e-01 -9.75117683e-01 3.02950293e-02
-4.68446881e-01 5.31666577e-01 4.39715147e-01 5.99729478e-01
-9.71753895e-01 1.07302018e-01 -8.30400050e-01 -7.46927559e-01
1.96330011e-01 3.79417270e-01 -7.05791056e-01 -3.07372689e-01
-1.03478289e+00 5.45445383e-01 1.31030247e-01 4.42486346e-01
-1.17881227e+00 -4.62701380e-01 -9.98651445e-01 -2.15942636e-01
4.02566910e-01 -1.13938904e+00 1.15284753e+00 -3.77148151e-01
-1.03248036e+00 8.79238486e-01 3.33619900e-02 -5.42698979e-01
1.16079569e+00 -7.11788714e-01 -3.06229770e-01 6.97727874e-02
3.17235082e-01 8.52677405e-01 5.87864816e-01 -1.26973879e+00
-7.64948666e-01 -7.76698411e-01 3.91835451e-01 4.27377999e-01
-7.28506744e-02 -2.13129848e-01 -8.09895337e-01 -3.79929304e-01
8.90470073e-02 -1.12537658e+00 -3.88766855e-01 1.32967800e-01
-6.58330560e-01 2.71684080e-01 6.96846247e-01 -6.02774799e-01
5.10722697e-01 -1.83011055e+00 8.19030255e-02 -4.83300500e-02
1.73290655e-01 6.83777481e-02 1.93148553e-01 1.28234878e-01
1.54203810e-02 -4.87705767e-01 1.51285201e-01 -7.68344700e-01
-1.02731567e-02 1.56642906e-02 2.76229054e-01 7.68142402e-01
-4.63840127e-01 6.98986351e-01 -8.89452040e-01 -1.78570762e-01
7.54880846e-01 8.04924130e-01 -6.04194343e-01 3.54864985e-01
1.70963764e-01 6.83814824e-01 -6.75978214e-02 8.22012901e-01
7.86325455e-01 -1.75196037e-01 1.17118247e-01 -2.42723659e-01
-1.45515546e-01 -6.08717427e-02 -1.61326551e+00 1.67919326e+00
-5.38198292e-01 5.36682844e-01 2.57936746e-01 -6.90410852e-01
6.68884277e-01 2.86635756e-01 4.74602550e-01 -6.62075698e-01
3.05292517e-01 -1.21815942e-01 -4.37675416e-01 -4.19114321e-01
5.58834791e-01 4.00172234e-01 -2.34866574e-01 -1.50480956e-01
3.15207131e-02 2.23591253e-01 2.47296482e-01 -1.39761284e-01
1.22304940e+00 5.17888926e-02 1.42313927e-01 1.05147943e-01
3.75668615e-01 -1.46237686e-01 5.54379046e-01 8.93044889e-01
-3.48067850e-01 8.54485333e-01 -1.71052396e-01 -1.11430025e+00
-1.00709414e+00 -1.29329956e+00 -2.01096069e-02 9.88675058e-01
1.98362529e-01 -4.44013268e-01 -7.86429167e-01 -4.76327807e-01
-6.87772259e-02 2.49458566e-01 -6.39786303e-01 2.10574821e-01
-7.18234301e-01 -6.85503066e-01 4.26431865e-01 6.24284625e-01
1.12556291e+00 -9.38421130e-01 -1.32912183e+00 -1.16032369e-01
-4.82954383e-01 -1.40222299e+00 -2.37136483e-01 9.81430486e-02
-3.09806526e-01 -1.35403621e+00 -9.31583226e-01 -4.65124726e-01
8.09643149e-01 6.29544795e-01 1.17191362e+00 -3.32435340e-01
-5.99840760e-01 7.49674976e-01 -1.85324289e-02 -6.50545716e-01
3.89927387e-01 -4.66904454e-02 3.99007887e-01 -1.00663543e-01
2.54704237e-01 -4.56968248e-01 -1.16902530e+00 4.70219433e-01
-3.10095817e-01 1.83183983e-01 3.65472108e-01 3.94194305e-01
4.97085571e-01 -5.67287132e-02 -3.30918103e-01 -4.31807488e-01
9.55057293e-02 -9.56960693e-02 -7.53262758e-01 -8.01687092e-02
2.62907613e-02 -6.92483604e-01 5.16376317e-01 -9.05637592e-02
-7.65036821e-01 4.44679022e-01 -1.75546512e-01 -5.71296394e-01
-6.81366503e-01 -1.30512312e-01 -3.76375109e-01 1.24615580e-01
7.49455392e-01 1.14733174e-01 -4.86879587e-01 -3.39204282e-01
9.61848199e-02 4.57523495e-01 8.49558651e-01 -3.23602974e-01
9.35870588e-01 8.03716183e-01 -1.55157484e-02 -1.20014429e+00
-8.41664433e-01 -6.32192791e-01 -7.28153229e-01 -6.45343125e-01
1.04559696e+00 -1.40289068e+00 -9.73015785e-01 6.07486844e-01
-1.08871627e+00 -4.37925100e-01 -2.45459393e-01 4.85316843e-01
-4.34176147e-01 -8.95429999e-02 -4.06276822e-01 -7.72146225e-01
-2.83559918e-01 -1.05277300e+00 1.43136859e+00 3.73087525e-01
-4.43342254e-02 -7.93220580e-01 2.56637037e-02 7.85400927e-01
1.75844789e-01 5.81730545e-01 -1.64711416e-01 -1.87470376e-01
-5.02881110e-01 -4.03204411e-01 6.86940774e-02 1.13260962e-01
-6.20665401e-02 -4.67310607e-01 -1.04234040e+00 -6.84659421e-01
-1.15086898e-01 -1.75252438e-01 5.70752859e-01 5.00618100e-01
1.00738490e+00 -2.44555816e-01 -4.27109182e-01 9.62559223e-01
1.21721077e+00 4.81792092e-02 4.38612312e-01 6.78585112e-01
1.03132284e+00 3.60637039e-01 5.51388860e-01 6.91386104e-01
6.89524174e-01 9.82595086e-01 7.36229062e-01 -2.15370476e-01
-4.86230059e-03 -2.34674275e-01 3.17376018e-01 8.13264176e-02
-5.60241818e-01 -2.30288833e-01 -1.09194243e+00 4.48903978e-01
-1.65592706e+00 -9.35580313e-01 7.05993688e-03 2.33227658e+00
6.53261840e-02 7.62965232e-02 2.81251878e-01 2.55901605e-01
6.85307264e-01 1.81622356e-01 -3.66219729e-01 4.48265404e-01
2.32775077e-01 -2.33074680e-01 6.70119286e-01 4.33640271e-01
-1.48541045e+00 5.83936095e-01 4.53208208e+00 -6.94083199e-02
-1.12114990e+00 -5.58485240e-02 3.37474734e-01 -5.25340319e-01
5.11525095e-01 -7.24715173e-01 -1.05055177e+00 3.03941607e-01
7.16806471e-01 6.38464212e-01 3.00705135e-01 1.21565485e+00
3.49234611e-01 -2.84100045e-02 -1.04340649e+00 1.43324649e+00
1.09001815e-01 -9.39686418e-01 -3.15654278e-01 1.81556553e-01
5.98617733e-01 3.35449934e-01 7.69032612e-02 2.87607402e-01
2.12814286e-01 -9.06360388e-01 8.66840839e-01 4.01769996e-01
7.32702136e-01 -7.78924286e-01 8.33528101e-01 4.45191562e-01
-1.41454089e+00 -3.24046493e-01 -4.84100819e-01 -1.97466418e-01
1.19466387e-01 4.57006156e-01 -9.33833420e-01 5.35425127e-01
1.49751437e+00 6.22699320e-01 -8.25884879e-01 1.05094039e+00
-3.10297161e-01 1.37815829e-02 -5.02528906e-01 1.12531699e-01
9.97041073e-03 -8.47963467e-02 7.06169665e-01 1.22507405e+00
4.16391373e-01 -2.02705353e-01 3.33982259e-01 4.50153321e-01
-1.05354488e-01 -4.58311588e-01 -8.32538843e-01 7.39679277e-01
2.83280492e-01 1.37769842e+00 -6.24689758e-01 -2.11853564e-01
-2.39192009e-01 1.14013612e+00 5.06237112e-02 3.56620729e-01
-1.01727092e+00 -1.96516559e-01 6.46844506e-01 5.51794291e-01
2.96210438e-01 -3.58667582e-01 2.70946950e-01 -1.37214541e+00
3.19731861e-01 -9.01008666e-01 2.34887764e-01 -8.84643734e-01
-8.02358031e-01 5.65559685e-01 1.11239098e-01 -1.12620223e+00
-1.21288724e-01 -9.42010403e-01 -3.53502423e-01 4.04096067e-01
-1.01285088e+00 -1.31393194e+00 -1.26785004e+00 9.25718367e-01
7.63343692e-01 -2.81792469e-02 6.34029329e-01 5.11941433e-01
-6.19291127e-01 4.93383259e-01 -2.00034961e-01 6.13030910e-01
5.26790202e-01 -1.46205711e+00 6.77237272e-01 9.98920441e-01
8.52069706e-02 5.75689614e-01 8.29288602e-01 -4.41941530e-01
-1.51133776e+00 -1.45551276e+00 5.95767140e-01 -8.18986416e-01
-7.48882517e-02 -8.59982073e-01 -3.42950076e-02 1.00895262e+00
-1.57496497e-01 4.31336194e-01 3.36837828e-01 -9.00874957e-02
-1.39277205e-02 -2.84481645e-01 -1.17296469e+00 6.90576673e-01
1.54973423e+00 -6.32638037e-02 -1.86253786e-01 5.64506233e-01
4.30649072e-01 -9.49432492e-01 -6.07325137e-01 4.74011064e-01
7.00878382e-01 -1.37743318e+00 1.41399431e+00 -3.64758186e-02
1.20588407e-01 -4.94879842e-01 -4.73862529e-01 -1.10747504e+00
-2.55297571e-01 -2.60197163e-01 -1.54397264e-01 5.65941215e-01
-8.14761594e-02 -5.21320760e-01 1.10203981e+00 1.77334636e-01
-1.37794003e-01 -5.77998161e-01 -1.00664461e+00 -6.53466523e-01
-6.60568893e-01 -6.38918161e-01 3.98150921e-01 3.45867425e-01
-6.38221800e-01 3.44636232e-01 -4.30784762e-01 6.18592083e-01
8.95223498e-01 -3.06486994e-01 1.44526076e+00 -1.22477782e+00
-3.29275966e-01 1.23458371e-01 -6.01147056e-01 -1.20094585e+00
-3.82791519e-01 -3.15767735e-01 -2.71628387e-02 -1.76759386e+00
-1.90827809e-02 -1.26874298e-01 6.45115897e-02 1.69734418e-01
1.26754209e-01 4.35047597e-01 2.82363951e-01 -3.90199870e-02
-8.04888189e-01 3.40103626e-01 9.76095915e-01 -2.14807644e-01
-1.54492706e-02 4.70970631e-01 -3.08453858e-01 8.99265945e-01
7.08536446e-01 -1.13151483e-01 -3.26082438e-01 -5.23337066e-01
2.59777427e-01 1.20831251e-01 1.22307408e+00 -1.96758354e+00
2.70133138e-01 2.66665906e-01 1.07234085e+00 -7.67221749e-01
8.36973250e-01 -1.07064319e+00 4.10817862e-01 7.63424337e-01
2.62232441e-02 4.88414466e-01 -9.11229476e-03 4.75881666e-01
1.58108115e-01 3.44301432e-01 7.82987237e-01 -7.59117901e-01
-7.35742569e-01 5.17525673e-01 4.99853492e-03 -1.80127680e-01
9.59436119e-01 -3.07005197e-01 -2.32582167e-01 -5.04126549e-01
-7.19389379e-01 1.58434108e-01 5.72270632e-01 5.36601067e-01
8.57110500e-01 -1.28524292e+00 -5.41569233e-01 3.69834155e-01
1.12990841e-01 5.22787392e-01 3.23013991e-01 5.96639931e-01
-8.24901819e-01 7.70537555e-01 -3.35381776e-01 -8.02123845e-01
-1.41652405e+00 4.56570208e-01 7.83885121e-01 -1.94459990e-01
-6.96998775e-01 9.33888555e-01 2.53701389e-01 -8.93339396e-01
5.66328645e-01 -2.22856343e-01 -1.90324709e-01 -3.12244087e-01
6.02415800e-01 6.02535605e-01 2.24141076e-01 -8.11661601e-01
-5.89022040e-01 6.26719177e-01 2.40557894e-01 -1.52647868e-01
1.39357936e+00 -5.43722436e-02 2.81819701e-01 1.38267756e-01
9.22720134e-01 6.65758029e-02 -1.57592547e+00 3.21678743e-02
-5.87458253e-01 -5.35036027e-01 -2.93676406e-01 -6.32555187e-01
-1.04443789e+00 8.00701976e-01 1.10907960e+00 -1.43064275e-01
8.92178237e-01 -6.30282164e-02 4.83573109e-01 7.94737279e-01
7.94065297e-01 -1.00723791e+00 3.18590492e-01 3.70855868e-01
9.59234238e-01 -1.41925311e+00 -8.24861377e-02 -2.06975088e-01
-3.81897062e-01 8.84583652e-01 1.00558627e+00 -7.89621100e-02
2.92358011e-01 1.28141105e-01 2.09449962e-01 -4.16240484e-01
-2.64678359e-01 -1.28096536e-01 1.20877856e-02 9.87682104e-01
1.39590442e-01 2.39005968e-01 5.86990178e-01 2.35359684e-01
-7.57667661e-01 -2.12524325e-01 3.40589702e-01 9.47938979e-01
-2.17906907e-01 -4.56370264e-01 -7.65960455e-01 2.60323957e-02
-2.86339611e-01 3.72215867e-01 -2.09382042e-01 9.03585136e-01
5.08401275e-01 9.42718983e-01 7.92746618e-02 -5.74326098e-01
8.47374737e-01 -3.86113256e-01 4.78108138e-01 -4.58134651e-01
-4.64481950e-01 -2.50474602e-01 7.71695301e-02 -7.97117472e-01
-2.74373323e-01 -7.33761787e-01 -7.76778042e-01 -2.68046319e-01
2.69624472e-01 -9.28050578e-02 7.27164507e-01 5.45615613e-01
1.82868198e-01 7.98411191e-01 3.64475220e-01 -1.52437639e+00
-1.12489901e-01 -8.51059079e-01 -2.85069555e-01 2.63508916e-01
3.74124885e-01 -6.32624686e-01 -2.17650339e-01 1.56368804e-03] | [7.2758402824401855, -0.9189152121543884] |
99ecc765-debb-44ee-ad9e-ec70fc96d12e | cgam-click-guided-attention-module-for | 2307.01015 | null | https://arxiv.org/abs/2307.01015v1 | https://arxiv.org/pdf/2307.01015v1.pdf | CGAM: Click-Guided Attention Module for Interactive Pathology Image Segmentation via Backpropagating Refinement | Tumor region segmentation is an essential task for the quantitative analysis of digital pathology. Recently presented deep neural networks have shown state-of-the-art performance in various image-segmentation tasks. However, because of the unclear boundary between the cancerous and normal regions in pathology images, despite using modern methods, it is difficult to produce satisfactory segmentation results in terms of the reliability and accuracy required for medical data. In this study, we propose an interactive segmentation method that allows users to refine the output of deep neural networks through click-type user interactions. The primary method is to formulate interactive segmentation as an optimization problem that leverages both user-provided click constraints and semantic information in a feature map using a click-guided attention module (CGAM). Unlike other existing methods, CGAM avoids excessive changes in segmentation results, which can lead to the overfitting of user clicks. Another advantage of CGAM is that the model size is independent of input image size. Experimental results on pathology image datasets indicated that our method performs better than existing state-of-the-art methods. | ['Won-Ki Jeong', 'Seonghui Min'] | 2023-07-03 | null | null | null | null | ['interactive-segmentation'] | ['computer-vision'] | [ 0.25994125 0.15417607 0.02190206 -0.4836698 -0.76641285 -0.18948425
0.14157687 0.32354987 -0.8060035 0.48931256 -0.14835371 -0.51327425
0.02662818 -0.5895083 -0.48552567 -0.6866268 0.35371348 0.2520121
0.61562425 0.08266272 0.3057379 0.2692862 -1.0754656 0.27922475
1.1848422 1.0944993 0.51412016 0.4305041 -0.46374574 0.1904211
-0.43407914 -0.34445688 0.11125512 -0.20967238 -0.8401348 0.13078867
0.20099677 -0.35395092 -0.10238072 1.369944 0.38758844 -0.09094232
0.43598983 -1.0417252 -0.63738054 0.38353807 -0.6734211 0.35540995
-0.20178376 0.14856066 0.85167885 -0.5923038 0.54372627 0.8554943
0.5882566 0.46422103 -1.3030423 -0.7032012 0.2962954 0.00860437
-1.3416305 0.04150026 0.56429744 -0.5875336 0.5522961 0.36133254
0.78304094 0.5719734 0.4899371 0.920776 0.8990883 -0.21932295
0.2168477 0.19221115 0.3035753 0.8076631 0.17600109 -0.33048072
-0.1750164 0.0730943 1.1326804 0.2732369 -0.18337815 -0.37323532
-1.1247195 1.0573356 0.95143706 0.50530815 -0.2685639 -0.06438066
0.35541993 -0.2075085 0.6180691 0.43588617 -0.1750324 0.03761528
-1.1327614 -0.06798379 0.36941084 0.5585501 0.4254856 -0.40954843
-0.31205586 0.6427026 0.31446484 -0.05491156 0.64718366 -0.52542406
0.2828481 1.0029819 0.02216727 -0.8727841 -0.5926561 -0.80072254
-0.80830014 0.27737677 0.67083216 -0.02284228 -1.3547827 1.4869543
0.36540368 -0.25149095 -0.5192446 1.0776973 0.89999574 0.03530114
0.33042207 0.01442108 1.3741046 -1.1633129 -0.93535006 -0.2606671
0.76694953 -0.43967035 1.3886911 0.19829819 -0.94125855 -0.38612705
-0.8429997 -0.10052411 -0.44500202 0.2210866 0.881116 0.5733142
-1.0208116 0.47528902 -1.1473119 -0.40103558 0.82502353 0.51762515
-0.17465284 0.38779676 -0.94775826 0.62863934 0.358238 0.17720258
-0.5089913 -0.9466977 -0.5387581 0.09369756 0.41338497 -0.5191668
1.4735571 -1.0231414 -1.264214 0.86802346 -0.1110107 -0.39661285
0.75365704 -0.19102009 -0.04696485 0.29216766 0.1433823 0.9828665
0.41131023 -1.051271 -0.74324536 -0.4354151 -0.02508731 0.2556532
-0.40861797 -0.290295 -1.0711275 -0.46703413 0.35856366 -0.72185606
-0.75589305 0.47120538 -0.67155224 -0.13163431 0.8146155 -0.7297232
1.3432242 -2.2015781 -0.2564657 0.32264915 0.525745 0.18414348
0.19232737 -0.12535289 0.01626839 0.40189227 -0.22921905 -0.28135243
-0.18424651 -0.08342934 0.22549036 0.47777498 -0.03730143 1.0129356
-0.6750108 -0.94323343 0.18088661 0.36290416 -0.5869237 0.14045909
-0.21463108 0.71736103 -0.5438304 0.74168575 0.658135 -0.7259728
-0.00938262 -0.3659168 0.02548063 0.01337083 -0.8155071 1.8412211
-0.20454513 0.5748208 0.1802065 -0.9547448 0.5401161 0.14883178
0.74356943 -0.66660935 0.35490188 0.04562562 0.14364581 -0.60181963
0.3026985 0.04903286 0.08713607 0.38498002 -0.208929 0.01107203
0.33743674 0.2559108 0.9420542 -0.25816563 0.18324587 -0.2086752
0.35908997 0.01820762 0.57014316 0.93331456 -0.43028715 0.62510705
0.6783464 -0.2951935 -0.8647662 -0.88363546 -0.21346587 0.97172767
0.2861948 -0.04942974 -1.148232 -1.0039321 -0.13058437 0.46600947
-0.936925 0.19671415 -0.34127834 -0.72772646 0.30657896 0.62837565
0.6603013 -1.0525398 -0.7519264 0.16325897 -0.18551742 -1.0657176
-0.6910741 0.03850776 -1.1283114 -1.0778825 -0.91722214 -0.79440933
1.3579619 0.21025807 0.77556866 0.40700328 -0.779501 -0.07744223
-0.20443708 -0.40232834 -0.19040337 0.29745546 -0.633638 -0.03599186
0.5231495 -0.18092406 -1.0283631 0.341486 -1.1269078 0.5335297
0.93553096 0.9092532 0.7642208 0.02998534 0.3409414 -1.1776001
0.6152299 -0.08511721 -0.6351122 0.08837686 -0.68758434 -0.1636158
0.31194904 -0.35407752 -0.90721864 0.21795008 -0.17462158 -0.13502029
-0.22366616 0.6586389 0.21493205 -0.23343235 0.45701298 0.17843935
0.3888704 -0.2919347 0.07567663 0.6351407 0.36358696 0.10566812
0.4347712 0.5655511 -0.23736103 -0.50666183 -0.94969475 -0.65674394
-0.65695876 -0.26271972 1.1653283 -0.43231124 -0.8750567 0.45917183
-0.8147008 -0.31384155 0.02342107 0.42948243 -0.28707013 0.44345495
-0.786588 -0.46627462 -0.5762387 -1.6918696 0.87777215 0.5844219
-0.09869462 -1.0375663 -0.45254523 0.5875676 0.5708744 0.19386148
1.04285 -0.78043395 -0.66270727 -0.548194 -0.59676784 0.04321703
0.27926564 -0.03133137 -0.6449445 -0.2563185 -0.20215508 -0.14391287
0.7935827 0.7578396 1.9015348 -0.06906924 -0.82692295 0.6073802
1.3193419 0.21924475 0.5330731 0.3974296 0.5363128 0.34819865
0.6756131 0.18777357 0.12523825 0.4171242 0.59082735 -0.8447087
-0.00947947 -0.02659347 -0.4752333 0.2807842 0.17984927 0.0369514
-1.0001295 0.5314753 -1.8057245 -0.53929687 -0.14095181 2.08991
1.0926265 0.4832708 -0.08523726 -0.06201665 0.70923996 -0.21654719
-0.8136228 0.02409588 0.50415194 0.05011632 0.61233354 0.29085508
-1.2656416 0.804517 6.5024486 0.96958995 -1.2326375 0.15182121
0.9353624 -0.06697285 -0.03157393 -0.32416895 -0.8236625 0.5148411
0.25968322 0.16483758 -0.07315607 0.8825542 0.53158927 -0.48631185
-1.0715383 0.84607095 -0.18584555 -1.4979293 -0.10667706 0.18322687
0.5163615 -0.12995522 0.17058471 0.1082576 0.14532535 -1.1166939
0.34249055 0.43797126 0.8232146 -0.4969597 0.9835397 0.26323628
-0.76625276 0.02670692 -0.06502768 0.45272842 0.12881267 0.65576595
-1.182698 0.1978564 0.7142591 0.3064399 -0.7656314 1.4774355
-0.03330225 0.57540655 -0.23038183 -0.22703144 0.39158764 -0.06798939
0.26165998 1.2360022 -0.03344742 -0.06443139 0.25398844 1.0233482
0.07501183 0.29703078 -0.05310443 -0.14835596 0.21562013 1.5260142
-1.3177416 -0.22534762 -0.42267367 0.74096274 0.3226189 0.51191336
-0.93442506 -0.51042247 0.20544098 0.33043754 0.2513871 -0.14754663
-0.84362465 -0.7330546 -0.15393175 -0.65937465 0.40181354 -0.48526883
-1.0162851 0.5323539 -0.33089644 -1.0452768 0.24651687 -0.5890349
-0.67735505 0.82132024 -1.2279396 -1.0605845 -0.69877255 0.40201932
0.78528255 0.20678772 0.655615 0.30236012 -0.7000435 0.828833
-0.02545824 0.34238103 0.54333425 -1.270604 0.01826732 0.6656172
-0.2557803 0.55688286 0.65138274 -0.77831197 -0.850508 -0.953743
0.39275685 0.01611067 0.6288632 -0.2503865 -0.84860045 0.45791823
0.24700186 0.10085573 1.0634586 0.01372907 0.2680383 0.05537573
-1.1422743 0.80572695 0.67980224 -0.08539715 -0.16026968 0.3609331
0.6410632 -0.7957734 -0.742014 0.39968982 0.37291887 -0.86698294
0.6935561 -0.47512844 0.4286347 -0.2031193 0.30572754 -1.1389655
-0.40680936 -0.1773887 0.256094 0.70787597 0.80852073 -0.51009226
1.0777847 1.0411435 -0.2509914 -1.1855124 -0.83619183 -0.32710958
-0.02071757 -0.05847668 0.3195493 0.6300233 -0.01204122 -0.00673135
0.21270148 0.02326329 0.5302509 0.02527426 0.5601767 -1.0556725
-0.15852965 -0.7017072 -0.28271872 -0.94879836 -0.24835427 -0.84858125
0.07314589 -1.8588868 0.36121887 -0.47832522 -0.40913054 0.5753697
-0.44169018 0.38433874 0.03932214 0.24703719 -0.6461208 0.2801559
1.6106908 -0.15576635 -0.34392282 0.13411435 -0.83851063 0.7794472
0.7841576 -0.35766232 -0.28447184 -0.3721759 -0.14884652 -0.03736831
0.35642365 -0.7743323 0.51576245 -0.14414479 0.6246452 -0.9859374
0.16508548 -0.7178615 -0.07250964 0.7101878 -0.59835196 -0.3171271
0.21484645 0.55202985 -0.23159102 -0.3486729 0.88999444 -0.28367877
-0.36445782 0.33604148 -0.46018466 -0.16512352 1.1200362 -0.5485182
-0.11620022 -0.27381453 -0.94464326 0.35455394 0.18983562 0.18025252
0.38761386 -1.0495397 -0.37734228 0.11687574 0.00750141 0.25237852
0.3915199 1.3069738 -0.7189396 0.57147384 0.11043295 -0.8337987
-1.3753998 0.35335714 0.4728474 -0.62641495 -0.74247545 0.9218483
0.4704765 -0.18448494 0.55988544 -0.4172264 -0.28498012 -0.2668549
0.40875342 -0.02230762 0.13490517 -0.008482 -0.22315128 0.15457878
-0.6796259 0.01723585 1.1002648 -0.12533952 0.1319236 0.08457003
0.916317 -0.44893646 -1.3766841 -0.12933852 -0.18941104 -0.5527478
0.42210034 -1.0957663 -1.3953885 0.8042301 0.88638395 0.3066211
1.0157175 -0.0828834 0.8787285 0.06536155 0.04750921 -1.3537726
0.246364 0.12154422 0.60464257 -1.5083811 -0.16141123 -0.62933487
-0.59638596 1.0038675 1.0288987 0.15113297 0.6446574 0.27953812
0.30614847 -0.23755112 -0.293173 -0.04567347 0.28874147 0.29246846
0.65704155 0.00831775 -0.71191674 0.7765667 -0.07836743 0.22263855
0.22689392 0.9554772 -0.5753291 -0.8340252 -0.21276845 0.8773596
-0.7777782 -0.10015512 -0.0846156 0.8628714 -0.15189259 0.74054646
0.07785253 -0.0106274 0.02109898 -0.17045908 0.07122775 -0.5830341
-0.6148685 0.33902746 -0.22551915 -0.5504796 0.0509157 -0.3176225
-1.4404715 0.1960469 -0.7397983 0.18779768 0.85401106 0.9237602
0.29593575 0.8130592 0.28123832 -0.5788273 -0.49086624 -1.00889
-0.43499607 0.5193546 0.04503269 -0.6297772 -0.11953341 0.08638937] | [14.611607551574707, -2.3582446575164795] |
7755366d-9d18-48a7-bb3f-49007872ad08 | time-space-tradeoff-in-deep-learning-models | 1901.10503 | null | http://arxiv.org/abs/1901.10503v1 | http://arxiv.org/pdf/1901.10503v1.pdf | Time-Space tradeoff in deep learning models for crop classification on satellite multi-spectral image time series | In this article, we investigate several structured deep learning models for
crop type classification on multi-spectral time series. In particular, our aim
is to assess the respective importance of spatial and temporal structures in
such data. With this objective, we consider several designs of convolutional,
recurrent, and hybrid neural networks, and assess their performance on a large
dataset of freely available Sentinel-2 imagery. We find that the
best-performing approaches are hybrid configurations for which most of the
parameters (up to 90%) are allocated to modeling the temporal structure of the
data. Our results thus constitute a set of guidelines for the design of bespoke
deep learning models for crop type classification. | ['Vivien Sainte Fare Garnot', 'Sebastien Giordano', 'Nesrine Chehata', 'Loic Landrieu'] | 2019-01-29 | null | null | null | null | ['crop-classification'] | ['miscellaneous'] | [ 1.47494022e-02 -3.81941140e-01 -1.37460589e-01 -2.82896668e-01
-2.76866198e-01 -8.52751255e-01 5.37341952e-01 3.26263130e-01
-3.58386576e-01 4.08816099e-01 8.90249573e-03 -6.53629839e-01
-5.19988537e-01 -9.83959734e-01 -5.25902510e-01 -7.73560703e-01
-6.06643438e-01 -2.08607465e-01 -2.05756381e-01 -7.16815412e-01
-2.51801908e-01 7.69251823e-01 -1.77577007e+00 3.01819295e-01
6.50074422e-01 1.25854623e+00 2.35507011e-01 6.53270483e-01
2.90467720e-02 7.35927343e-01 -3.87424260e-01 4.21563946e-02
2.53705442e-01 -7.46562928e-02 -7.21563637e-01 -1.15974665e-01
1.83776468e-01 -1.08087793e-01 -4.39724959e-02 6.89856827e-01
5.01713514e-01 -1.20345421e-01 6.27312481e-01 -8.36097538e-01
-2.44069532e-01 7.15852082e-01 -3.10474992e-01 5.07295787e-01
-1.86493546e-01 3.36798802e-02 8.00656438e-01 -5.84890306e-01
3.29661041e-01 7.28722215e-01 1.26853275e+00 -1.12899736e-01
-9.85687792e-01 -3.65542442e-01 4.11297411e-01 1.09922916e-01
-1.33301151e+00 -3.88287634e-01 6.16065800e-01 -7.11523771e-01
1.08654928e+00 1.07590593e-01 9.30307209e-01 9.80972171e-01
4.73620266e-01 6.66313767e-01 1.12449014e+00 -3.88811678e-01
2.15611234e-01 -1.75184995e-01 5.06268084e-01 3.05827528e-01
1.57297000e-01 4.29470420e-01 -2.84318328e-01 -3.12715709e-01
5.30237198e-01 -8.14494491e-02 1.46691885e-03 -2.53128380e-01
-8.46787691e-01 8.38611066e-01 5.76643288e-01 5.35769105e-01
-9.50215220e-01 1.09683676e-03 5.92197418e-01 3.92759651e-01
9.54226792e-01 3.80354017e-01 -9.99680102e-01 1.97744206e-01
-1.01631808e+00 3.27110618e-01 3.95632505e-01 4.56500977e-01
7.17950702e-01 3.35164815e-01 -3.47434059e-02 8.44440758e-01
5.84680103e-02 6.03957236e-01 1.34805605e-01 -5.68653107e-01
1.33534178e-01 3.94492775e-01 1.34008721e-01 -9.60377336e-01
-9.10378098e-01 -7.98396885e-01 -1.12946761e+00 -1.79743692e-01
1.19172402e-01 -6.32472992e-01 -1.00706089e+00 1.48361838e+00
-1.36340708e-01 -4.32554958e-03 2.63279349e-01 5.35309315e-01
7.20933616e-01 7.43121088e-01 3.19184333e-01 -1.75244272e-01
1.22851384e+00 -2.96649784e-01 -6.77243650e-01 -2.21654505e-01
7.75071263e-01 -5.65073848e-01 4.51606810e-01 2.88761803e-03
-5.89286923e-01 -5.54146707e-01 -8.36047709e-01 5.21875441e-01
-8.66905510e-01 5.25748909e-01 1.02592528e+00 4.61232007e-01
-1.02341807e+00 7.19577610e-01 -8.01215410e-01 -6.18555605e-01
1.46254763e-01 7.59557858e-02 4.80517633e-02 2.87705302e-01
-1.35956538e+00 8.06303322e-01 6.39406800e-01 8.49728286e-01
-6.77023113e-01 -8.45992982e-01 -6.43872559e-01 9.04753432e-02
-1.90470517e-01 -2.55071104e-01 1.06873262e+00 -1.09041297e+00
-1.05082405e+00 9.17412341e-01 5.34361834e-03 -7.73773015e-01
-1.05490591e-02 4.40004803e-02 -4.91738379e-01 -1.81845695e-01
-8.21748897e-02 5.14352441e-01 5.47784507e-01 -1.10059595e+00
-6.89175487e-01 -4.69514072e-01 1.12355053e-01 -5.58767803e-02
-3.97321433e-01 3.31178576e-01 4.64741290e-01 -7.68936694e-01
4.44803294e-03 -1.09316587e+00 -4.84567821e-01 -3.62494111e-01
2.10262060e-01 -7.69572780e-02 7.53353238e-01 -8.21968853e-01
1.19230664e+00 -2.26849055e+00 -4.36816225e-03 5.44317067e-02
-1.99259102e-01 5.02549231e-01 -4.12257731e-01 7.09032536e-01
-2.74753541e-01 1.29338622e-01 -2.98229128e-01 2.35078007e-01
-1.90818369e-01 2.98826069e-01 -5.58952868e-01 3.10991406e-01
4.78773326e-01 1.00987613e+00 -6.50690615e-01 1.65700465e-01
2.09272534e-01 5.16067505e-01 -4.08129254e-03 8.60107392e-02
-3.47702891e-01 9.63451639e-02 -4.63779867e-01 6.05602086e-01
8.37958515e-01 -1.66648939e-01 3.71338367e-01 -2.97948241e-01
-6.32583618e-01 2.68662393e-01 -8.32120001e-01 1.05489850e+00
-3.71494502e-01 7.01780498e-01 4.15897220e-02 -1.09722126e+00
8.89714658e-01 3.71574640e-01 8.23935091e-01 -6.69065893e-01
-7.12615401e-02 3.15703720e-01 9.45902616e-02 -3.24917704e-01
5.50250351e-01 2.54942447e-01 -4.96142684e-03 1.29301280e-01
8.67574364e-02 1.10048093e-01 1.48599103e-01 -4.84504431e-01
7.51205146e-01 1.54458225e-01 2.25155294e-01 -6.89836621e-01
1.03527129e-01 1.69130057e-01 4.70722198e-01 6.62355840e-01
-1.84208930e-01 4.44542110e-01 7.60181546e-01 -1.03925371e+00
-1.10505188e+00 -4.66861755e-01 -3.28386158e-01 1.17132246e+00
-4.90906775e-01 -2.33342554e-02 -3.40872139e-01 -3.51890475e-01
-2.82888068e-04 3.26038182e-01 -8.53036046e-01 1.42721757e-01
-3.94601882e-01 -1.60653055e+00 8.46556127e-01 5.67111969e-01
5.39191782e-01 -1.22365499e+00 -9.68624294e-01 3.42779100e-01
-1.50431782e-01 -9.30015624e-01 4.22914654e-01 8.59542072e-01
-9.38325763e-01 -9.63998079e-01 -7.07434893e-01 -4.92101818e-01
-1.09810583e-01 4.17248189e-01 1.07808769e+00 -4.23990428e-01
3.92939113e-02 5.76997511e-02 -7.11193025e-01 -5.10101199e-01
-6.51281551e-02 6.66056812e-01 -3.91126513e-01 -1.14012919e-01
3.48137319e-01 -4.44083929e-01 -3.95502388e-01 5.03592715e-02
-9.81629908e-01 -1.87111244e-01 5.32764018e-01 7.60286927e-01
2.90050864e-01 1.51819631e-01 4.97319430e-01 -3.96341860e-01
4.29482877e-01 -7.96989501e-01 -7.58183122e-01 4.08273935e-01
-5.00695050e-01 -1.64172202e-01 3.60049397e-01 -2.53927976e-01
-8.24058115e-01 2.24965110e-01 -2.46677399e-01 -1.91373825e-01
-3.74689966e-01 1.33224249e+00 4.45859134e-01 -1.69406399e-01
6.85449541e-01 2.13856101e-01 -1.47163868e-01 -5.04038274e-01
-1.84116997e-02 5.45488119e-01 4.63241488e-02 -2.77679771e-01
3.10955048e-01 3.39977115e-01 -1.21748179e-01 -1.26669669e+00
-9.82942700e-01 -3.65444571e-01 -7.88812578e-01 -3.39919925e-01
6.50764883e-01 -1.15161252e+00 -1.83691740e-01 1.08647740e+00
-1.02853644e+00 -7.05713689e-01 -9.48681235e-02 3.74709636e-01
-3.44406217e-01 -1.93800833e-02 -5.45485318e-01 -9.16715384e-01
-3.45951319e-01 -8.42458189e-01 1.07328010e+00 -3.51789407e-02
1.65037543e-01 -1.19231796e+00 1.56783193e-01 -3.13696742e-01
7.73594797e-01 6.19863808e-01 9.28485870e-01 -2.15256721e-01
-1.67654589e-01 1.64445322e-02 -3.28080028e-01 1.74887791e-01
1.77661523e-01 4.00147945e-01 -1.19375658e+00 -2.54576296e-01
-2.33000636e-01 -2.83876330e-01 1.12731814e+00 1.02986038e+00
1.11560285e+00 7.80151635e-02 -4.06637453e-02 8.47514749e-01
1.57457137e+00 2.13375106e-01 5.75955212e-01 3.37157995e-01
4.12803560e-01 7.78397083e-01 6.09831631e-01 5.35623550e-01
3.51768881e-01 4.73056942e-01 5.71520746e-01 -4.78676021e-01
3.29078943e-01 1.19127683e-01 1.50606841e-01 5.97745121e-01
-2.75910079e-01 -2.94666171e-01 -1.34971714e+00 7.82921672e-01
-1.87207437e+00 -1.12922382e+00 -3.21932912e-01 1.93232560e+00
4.34019834e-01 -2.62509555e-01 2.42435828e-01 -1.05183497e-02
3.78233403e-01 7.48297513e-01 -4.10300881e-01 -1.69169605e-01
-7.03716099e-01 2.91242957e-01 9.69047725e-01 2.21451297e-02
-1.65247738e+00 1.12808073e+00 7.71764708e+00 5.10773063e-01
-1.43735528e+00 -1.23453505e-01 5.74518919e-01 1.84881404e-01
-3.54989134e-02 -1.44677982e-01 -7.53002226e-01 1.21170640e-01
1.38083613e+00 1.55922472e-01 2.72792011e-01 3.09613496e-01
4.72586513e-01 -8.38861316e-02 -4.08581823e-01 2.43227243e-01
-4.21734124e-01 -1.43791592e+00 -2.74417162e-01 9.75466445e-02
6.39122009e-01 6.35517061e-01 9.02530253e-02 2.10655659e-01
3.46416831e-01 -8.54212999e-01 5.85791767e-01 5.00743926e-01
3.96564066e-01 -6.90976262e-01 1.04524291e+00 4.81376313e-02
-1.45200408e+00 -4.00306672e-01 -3.24703872e-01 -1.95095137e-01
-2.24668786e-01 6.00716352e-01 -1.71968073e-01 1.03142989e+00
1.01679993e+00 1.10616350e+00 -6.11975074e-01 9.94653642e-01
3.53738755e-01 7.37497270e-01 -4.56351340e-01 7.10694268e-02
7.50835180e-01 2.75316685e-02 5.58464564e-02 1.26376545e+00
4.05505598e-01 5.21281585e-02 2.32190385e-01 4.91100341e-01
5.40840685e-01 -1.37192709e-02 -8.36047232e-01 -4.07568991e-01
3.21381122e-01 1.09695947e+00 -5.75038075e-01 1.88655227e-01
-2.91151762e-01 2.46390358e-01 1.63998723e-01 5.90224862e-01
-5.88280320e-01 -3.47498842e-02 9.03223574e-01 -1.94726974e-01
4.87273842e-01 -4.93374974e-01 -1.44214302e-01 -1.01855445e+00
-9.05022956e-03 -9.16240990e-01 5.64965606e-01 -8.64162326e-01
-9.83679295e-01 6.96180224e-01 9.80860740e-02 -1.13591838e+00
-3.61580282e-01 -7.79958546e-01 -4.12623137e-01 1.02211797e+00
-1.97799098e+00 -1.50341797e+00 -4.24746752e-01 1.93515554e-01
2.13811204e-01 -2.47720852e-01 1.19948876e+00 1.56029567e-01
-5.85148335e-01 -7.83046186e-02 3.88316244e-01 1.28891006e-01
1.67493522e-01 -8.02214205e-01 6.30050600e-01 7.56803513e-01
-1.85432270e-01 1.87751099e-01 5.67382216e-01 -4.58513886e-01
-1.18237531e+00 -1.31118941e+00 8.64963651e-01 9.96204093e-02
8.24686348e-01 -6.99281618e-02 -9.46419299e-01 6.68104470e-01
3.90092313e-01 -5.24700806e-02 6.17336512e-01 5.35241485e-01
-2.59287566e-01 -3.38572234e-01 -6.66215122e-01 6.80877268e-02
6.57900453e-01 -5.38735390e-01 -2.96775214e-02 3.43895704e-01
5.60397983e-01 -1.62893280e-01 -1.19080329e+00 9.67441022e-01
6.37829363e-01 -9.28194940e-01 6.69220090e-01 -5.91563106e-01
3.20533514e-01 -1.85177624e-02 -3.33983034e-01 -1.58116853e+00
-5.92739105e-01 -2.59376973e-01 2.90166229e-01 8.40828657e-01
4.60526049e-01 -6.09206498e-01 5.41399598e-01 -2.41783082e-01
-4.18777540e-02 -5.87778389e-01 -5.50881028e-01 -5.90207160e-01
4.32189137e-01 -3.94903719e-01 6.96374655e-01 1.07994819e+00
-5.42701781e-01 -1.43232703e-01 -2.51944870e-01 5.28984785e-01
2.57086396e-01 2.93792218e-01 2.18141422e-01 -1.39475155e+00
4.09846276e-01 -5.15791178e-01 -1.71708569e-01 -3.26788336e-01
3.15210074e-01 -3.64254415e-01 -3.82129252e-01 -1.09974563e+00
-2.32510626e-01 -6.70696318e-01 -6.19529307e-01 7.92922497e-01
1.37812302e-01 -7.81664550e-02 -5.03612570e-02 1.41927108e-01
2.42056735e-02 5.53246439e-01 4.18602735e-01 -1.98148206e-01
-2.45727941e-01 3.47692490e-01 -2.00695291e-01 4.15063769e-01
1.24873352e+00 -1.22994088e-01 -1.77879781e-01 -7.15381742e-01
4.81373876e-01 1.14256412e-01 6.91675544e-01 -9.47865963e-01
-2.44897917e-01 -4.12769675e-01 2.29396746e-01 -1.03903186e+00
-1.54215842e-03 -6.30376041e-01 4.54920918e-01 3.81144941e-01
-4.33989704e-01 2.74832070e-01 6.86438680e-01 6.84312657e-02
-3.52298290e-01 -9.45671648e-02 6.71656489e-01 -1.44356802e-01
-9.54010785e-01 4.06627506e-01 -7.44266987e-01 -2.50682682e-01
6.66714489e-01 1.33592963e-01 -2.01711133e-01 -1.43045723e-01
-7.62087464e-01 2.83242345e-01 9.16650966e-02 7.45257139e-01
1.46416664e-01 -1.11179376e+00 -8.89403939e-01 3.28815073e-01
3.68399560e-01 -3.81958455e-01 4.20229316e-01 6.08132839e-01
-5.56540430e-01 8.10525298e-01 -4.99337643e-01 -7.05504477e-01
-1.01200175e+00 2.69189000e-01 9.63338912e-01 -3.23991030e-01
-2.61235714e-01 5.26056409e-01 -3.13516915e-01 -7.31751025e-01
6.60009980e-02 -4.01605606e-01 -7.13067770e-01 7.07082212e-01
6.45292327e-02 -1.56008229e-02 4.04226393e-01 -4.63111907e-01
-2.65025258e-01 4.11455721e-01 2.52739847e-01 1.66304335e-01
1.82429862e+00 2.15544388e-01 -2.37340569e-01 7.17849255e-01
9.35846627e-01 -6.60383642e-01 -1.19844139e+00 -2.94038117e-01
1.55359253e-01 8.50657448e-02 4.94737357e-01 -7.60722756e-01
-1.18573105e+00 6.76909447e-01 1.10791314e+00 8.36767077e-01
1.41153514e+00 -4.55950826e-01 1.87678829e-01 4.88603592e-01
3.74674290e-01 -1.02994239e+00 -7.38413215e-01 9.90120471e-01
7.39870012e-01 -1.11141098e+00 -1.91971660e-01 -1.27710581e-01
-3.59130621e-01 1.15168929e+00 2.58559048e-01 -3.61985937e-02
1.01031065e+00 3.89630675e-01 2.39984557e-01 -3.55109423e-01
-9.45782304e-01 -7.31970847e-01 9.90662873e-02 6.86751306e-01
7.35627711e-01 3.88402641e-01 -7.95877352e-02 2.26674620e-02
-1.78590789e-02 6.62248433e-02 2.82860070e-01 9.20836151e-01
-3.22383046e-01 -7.90619671e-01 -1.78922206e-01 4.46819723e-01
-2.94703215e-01 -1.87797800e-01 -2.63896614e-01 7.20698059e-01
-1.10567361e-02 8.49390030e-01 3.27960253e-02 -2.67125845e-01
2.99517453e-01 3.56360115e-02 1.33623138e-01 -2.31988713e-01
-1.01378465e+00 9.38840210e-03 2.29088187e-01 -3.24716657e-01
-8.16393197e-01 -9.37870979e-01 -3.15439820e-01 -2.84143567e-01
-1.85415030e-01 -1.01454109e-01 7.17103183e-01 8.97597373e-01
4.70816940e-01 6.68417335e-01 8.04891646e-01 -1.01386285e+00
-4.89795476e-01 -1.10888922e+00 -6.56416118e-01 -1.96282223e-01
5.68104863e-01 -4.10438478e-01 -1.13494009e-01 -2.42290407e-01] | [9.451417922973633, -1.574607491493225] |
d5d85cd0-f86a-4db0-9df0-6473d4dedbb2 | automated-problem-setting-selection-in-multi | 2104.09967 | null | https://arxiv.org/abs/2104.09967v2 | https://arxiv.org/pdf/2104.09967v2.pdf | Multi-target prediction for dummies using two-branch neural networks | Multi-target prediction (MTP) serves as an umbrella term for machine learning tasks that concern the simultaneous prediction of multiple target variables. Classical instantiations are multi-label classification, multivariate regression, multi-task learning, dyadic prediction, zero-shot learning, network inference, and matrix completion. Despite the significant similarities, all these domains have evolved separately into distinct research areas over the last two decades. This led to the development of a plethora of highly-engineered methods, and created a substantially-high entrance barrier for machine learning practitioners that are not experts in the field. In this work we present a generic deep learning methodology that can be used for a wide range of multi-target prediction problems. We introduce a flexible multi-branch neural network architecture, partially configured via a questionnaire that helps end-users to select a suitable MTP problem setting for their needs. Experimental results for a wide range of domains illustrate that the proposed methodology manifests a competitive performance compared to methods from specific MTP domains. | ['Willem Waegeman', 'Bernard De Baets', 'Dimitrios Iliadis'] | 2021-04-19 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 4.78003860e-01 3.54080163e-02 -5.44300199e-01 -5.67833960e-01
-9.87700701e-01 -2.57925212e-01 4.79018688e-01 3.06022227e-01
-1.13389678e-01 8.73288512e-01 -1.87929705e-01 -3.16935517e-02
-5.68787932e-01 -6.10964060e-01 -3.15968156e-01 -7.91941643e-01
1.37053905e-02 8.44930530e-01 5.82936294e-02 -2.41832942e-01
1.24761082e-01 3.66518348e-01 -1.79053330e+00 4.22669649e-01
6.49594009e-01 1.23966730e+00 5.76220360e-03 2.88882494e-01
-4.21002842e-02 8.23600113e-01 -4.43500727e-01 -7.60641277e-01
1.34814546e-01 1.30719736e-01 -8.29216361e-01 -1.25110865e-01
3.19700181e-01 2.58401126e-01 1.82510629e-01 6.26378357e-01
6.50227666e-01 2.58540958e-01 9.38898921e-01 -1.68080628e+00
-3.98910224e-01 4.12453890e-01 -5.99553168e-01 -1.20948866e-01
3.47976424e-02 -2.15731263e-01 1.41837895e+00 -8.71077657e-01
3.09811980e-01 1.24460900e+00 9.52758074e-01 3.77026886e-01
-1.54805362e+00 -6.66811109e-01 -1.60837620e-01 4.97856587e-01
-1.07318091e+00 -2.30318829e-01 9.34650600e-01 -7.62885153e-01
9.02296662e-01 3.07139188e-01 7.99664110e-03 1.54211926e+00
2.94909447e-01 8.19556773e-01 1.14543653e+00 -5.19745827e-01
2.25427613e-01 4.20430005e-01 4.95103002e-01 3.97931486e-01
-2.62780100e-01 8.63624513e-02 -5.56337535e-01 -4.81635153e-01
2.84112275e-01 3.13287765e-01 2.25127339e-01 -6.56074822e-01
-1.17319202e+00 1.28332520e+00 -8.19550678e-02 3.74983102e-01
-4.00015384e-01 -2.42013872e-01 6.39773905e-01 5.55920005e-01
6.88490212e-01 4.97068316e-01 -8.49741280e-01 -1.25853509e-01
-7.57490337e-01 3.45894575e-01 1.07599008e+00 7.04799175e-01
8.30222130e-01 -7.92314038e-02 -2.88632929e-01 1.32506609e+00
2.25842655e-01 6.07919283e-02 5.51474810e-01 -7.95558572e-01
5.03579140e-01 6.47502720e-01 2.90215790e-01 -1.02475274e+00
-7.67409086e-01 -5.31147361e-01 -1.09015739e+00 3.07289541e-01
5.69617093e-01 -2.79694021e-01 -4.14423525e-01 1.66013587e+00
5.53333521e-01 9.85079110e-02 -9.54484567e-02 5.61867237e-01
5.69701195e-01 6.32137716e-01 2.28330672e-01 -2.93600768e-01
1.14875972e+00 -1.31408989e+00 -4.12759602e-01 -3.82149965e-01
5.78692257e-01 -6.42472327e-01 8.23963344e-01 7.06092775e-01
-4.60408777e-01 -5.52688301e-01 -7.98118889e-01 1.41880736e-01
-6.23715222e-01 -1.62688881e-01 8.17702889e-01 6.81138873e-01
-7.10080802e-01 9.72840369e-01 -4.70065922e-01 -3.83120000e-01
2.24984139e-01 4.79415148e-01 -3.10204834e-01 -9.59783942e-02
-1.39715517e+00 1.20592296e+00 5.75263321e-01 -2.41148904e-01
-9.50013399e-01 -7.07981527e-01 -5.95816970e-01 1.67314976e-01
3.60150665e-01 -4.81337309e-01 1.20874226e+00 -1.09950364e+00
-1.48653328e+00 9.70520973e-01 1.94919258e-01 -3.14368427e-01
3.93777370e-01 -3.07101995e-01 -4.64206845e-01 -4.60013151e-01
3.57227139e-02 3.59442860e-01 1.02589035e+00 -7.96311140e-01
-9.18915629e-01 -5.43792725e-01 -1.32252872e-01 2.29874402e-01
-5.40887058e-01 3.99845719e-01 4.29751985e-02 -5.38004279e-01
-5.49966753e-01 -7.27410614e-01 -5.17735183e-01 -3.51750642e-01
-4.02445167e-01 -5.78140259e-01 6.81836188e-01 -3.34203243e-01
1.11959302e+00 -2.06593776e+00 5.57969272e-01 -1.84716165e-01
2.17362508e-01 2.31057107e-01 -3.75992924e-01 7.42083967e-01
-5.54034829e-01 -2.40475148e-01 -7.55113140e-02 -5.60787916e-01
1.99321195e-01 9.86201465e-02 -1.42383054e-01 4.07910615e-01
7.62704015e-02 7.58875132e-01 -8.07475984e-01 -4.28337067e-01
2.11373523e-01 2.79806107e-01 -1.06231332e-01 2.43238047e-01
-4.25661415e-01 3.12725484e-01 -3.10135931e-01 8.29724789e-01
3.97297949e-01 -5.78572810e-01 1.88923627e-01 1.01933971e-01
2.09701620e-02 -2.58028597e-01 -1.18548775e+00 1.65906656e+00
-3.47792506e-01 4.22309220e-01 -9.83768925e-02 -1.60353434e+00
1.13542640e+00 4.75431263e-01 8.87533545e-01 -3.52764696e-01
1.48816779e-01 1.01052478e-01 -1.78914383e-01 -5.68879604e-01
3.04346204e-01 -3.77716333e-01 -2.56293476e-01 4.03190404e-01
3.89572293e-01 4.86917436e-01 1.72907099e-01 -3.85711312e-01
1.02942896e+00 -1.16763180e-02 7.24389553e-01 -7.52997473e-02
4.15726215e-01 -1.51671069e-02 7.76834607e-01 7.76039243e-01
-3.96117717e-01 2.76062042e-01 5.73665202e-01 -8.81805420e-01
-9.02437866e-01 -7.32062638e-01 -1.89139470e-01 1.99283397e+00
-1.61798313e-01 -2.26820245e-01 -1.40406847e-01 -5.66260874e-01
1.75155208e-01 5.83296895e-01 -7.42303967e-01 -7.28027662e-03
-2.75108308e-01 -1.06972456e+00 2.41768330e-01 2.19054714e-01
1.70838870e-02 -1.19498169e+00 -5.47723711e-01 5.14538765e-01
-1.95726305e-01 -8.31486106e-01 2.08737120e-01 6.73987031e-01
-8.35381091e-01 -1.04496443e+00 -8.51692140e-01 -9.18987691e-01
-2.03165591e-01 4.75156270e-02 1.35270739e+00 -4.89843160e-01
-2.99267024e-01 2.23548740e-01 -4.43660587e-01 -5.74889779e-01
-4.30506140e-01 4.31838691e-01 1.72304511e-01 3.68170649e-01
6.83870614e-01 -8.88293207e-01 -1.78949967e-01 5.16514361e-01
-6.70282662e-01 8.50082114e-02 6.62853658e-01 1.08894432e+00
4.71115172e-01 -1.09605357e-01 9.10726905e-01 -1.06401229e+00
7.99771309e-01 -9.34370995e-01 -6.09746933e-01 7.05730498e-01
-8.24359477e-01 -9.74700414e-03 6.13551199e-01 -7.89752245e-01
-8.71772110e-01 1.27896503e-01 2.36514080e-02 -4.21676069e-01
-6.29709423e-01 7.23928273e-01 -9.93616953e-02 -1.28590450e-01
9.67544258e-01 1.29365519e-01 5.05979583e-02 -6.63920403e-01
2.98756123e-01 6.86364353e-01 -1.83418572e-01 -4.17137086e-01
4.18426067e-01 -9.05513689e-02 2.35935375e-01 -4.50763285e-01
-1.12882030e+00 -6.41537547e-01 -8.33512247e-01 -1.77000150e-01
7.78760433e-01 -7.64192760e-01 -8.86162460e-01 6.63935959e-01
-1.03369439e+00 -3.79249066e-01 9.81941223e-02 2.66521603e-01
-5.15216887e-01 2.04268396e-01 -5.11626959e-01 -6.78225338e-01
-4.81816918e-01 -1.13213587e+00 9.53315496e-01 1.26319885e-01
-3.05216372e-01 -1.15233719e+00 3.97970676e-01 5.65795243e-01
5.24118721e-01 4.64878827e-01 1.22502136e+00 -1.21566093e+00
-1.82779923e-01 -2.64615595e-01 -1.51096851e-01 3.03529322e-01
-9.29560214e-02 -3.01893413e-01 -1.09957957e+00 -1.84400201e-01
-1.32667542e-01 -8.25902283e-01 6.65284812e-01 3.38007629e-01
1.02831399e+00 -4.38531861e-02 -5.10150731e-01 4.48831975e-01
1.44509161e+00 2.13581488e-01 1.81940258e-01 5.07456601e-01
5.98580658e-01 7.96502411e-01 7.61820674e-01 6.81790411e-01
2.89939702e-01 8.67758751e-01 6.11754060e-01 2.35053543e-02
4.74856555e-01 2.90728807e-01 1.15474239e-01 5.84490299e-01
-2.52098553e-02 -1.03404999e-01 -1.22438347e+00 3.47581834e-01
-2.18103361e+00 -9.57504034e-01 -2.02948198e-01 2.03076124e+00
6.95907176e-01 -9.68910605e-02 4.08110142e-01 3.11472900e-02
7.40612268e-01 2.83515424e-01 -6.97156966e-01 -3.13697755e-01
1.13189511e-01 9.81685966e-02 2.37438753e-01 1.80491582e-01
-1.53689754e+00 7.53784716e-01 6.68373108e+00 1.31690609e+00
-1.03906322e+00 4.05911267e-01 7.78151035e-01 4.73596044e-02
3.03319305e-01 -2.71038502e-01 -1.05607474e+00 3.92953634e-01
1.16130686e+00 -1.49948359e-01 2.59012997e-01 1.22309494e+00
-1.70452595e-01 1.41876996e-01 -1.38538063e+00 1.16299164e+00
-8.83250907e-02 -1.10123432e+00 -3.66714269e-01 6.49949629e-03
5.75448930e-01 4.08716381e-01 2.97367513e-01 8.48387063e-01
3.30852807e-01 -1.10002661e+00 1.83157980e-01 3.79161507e-01
5.97442091e-01 -6.59623682e-01 5.81203580e-01 6.44345641e-01
-8.41156900e-01 -5.44425905e-01 -4.72828776e-01 -3.68407041e-01
2.98017673e-02 7.01339304e-01 -8.09821725e-01 6.68448389e-01
4.56454873e-01 8.08913946e-01 -3.60408962e-01 1.14693153e+00
1.36649862e-01 6.45833194e-01 -6.80498630e-02 -1.69409424e-01
1.63329691e-01 -3.53806645e-01 2.48415381e-01 9.73237693e-01
8.99555087e-02 -3.28802675e-01 2.98290640e-01 4.12932098e-01
8.60423893e-02 5.36253572e-01 -6.20703757e-01 4.11359593e-02
2.32625440e-01 1.50316203e+00 -3.75139058e-01 -2.15694532e-01
-8.12116742e-01 7.50263333e-01 7.75125623e-01 3.32711160e-01
-9.61961031e-01 -2.44807661e-01 6.25783741e-01 -3.57834756e-01
1.85663477e-01 1.85568377e-01 -4.45076764e-01 -1.07836378e+00
-3.01014423e-01 -1.22463477e+00 7.27060139e-01 -3.15979213e-01
-1.77319396e+00 5.47278404e-01 -1.40740901e-01 -1.26145029e+00
-3.92499685e-01 -7.08247066e-01 -3.87326479e-01 8.04870903e-01
-1.42993402e+00 -1.18554962e+00 -1.72338322e-01 6.32290423e-01
5.51897347e-01 -5.60982883e-01 1.42103004e+00 5.71004093e-01
-7.60394871e-01 4.23835307e-01 5.83516121e-01 -1.20719336e-01
8.64366710e-01 -1.20808756e+00 4.06380109e-02 2.70380259e-01
1.06625505e-01 1.99190136e-02 6.33273065e-01 -2.92187870e-01
-1.00793874e+00 -1.08896792e+00 8.93365502e-01 -4.97792214e-01
8.95156026e-01 -4.26269472e-01 -7.83091545e-01 7.35677302e-01
1.65235698e-01 -1.85782641e-01 1.25833821e+00 8.14136446e-01
-5.32117665e-01 -4.10047263e-01 -9.35628593e-01 2.45056525e-01
4.86656040e-01 -2.89245129e-01 -4.33108509e-01 7.61869550e-01
5.87440252e-01 -6.78676888e-02 -1.18753159e+00 4.58131969e-01
5.96399784e-01 -8.91977131e-01 1.05330825e+00 -9.78331208e-01
6.48544610e-01 3.97800595e-01 -2.92139798e-01 -1.32455564e+00
-7.07880437e-01 -5.34449816e-01 -3.40870678e-01 1.04837167e+00
6.17493212e-01 -5.28816223e-01 8.65447104e-01 8.72268319e-01
8.37459713e-02 -1.08153331e+00 -9.86510098e-01 -5.16973317e-01
1.63979471e-01 -3.95209014e-01 2.76904017e-01 1.29288816e+00
9.27022994e-02 8.74009311e-01 -9.95833337e-01 -1.57963574e-01
8.07646811e-01 4.17195082e-01 7.36251175e-01 -1.95995474e+00
-6.04524255e-01 -6.10573828e-01 -2.70424992e-01 -7.30019450e-01
2.27528319e-01 -9.42537487e-01 -3.36105347e-01 -1.16698384e+00
2.94125617e-01 -3.95185292e-01 -6.53774083e-01 7.47156262e-01
-9.42494646e-02 -8.06366354e-02 2.68094651e-02 2.23798349e-01
-7.66977787e-01 5.32402813e-01 7.77460515e-01 -1.17424138e-01
-1.28622636e-01 5.04390776e-01 -6.53528750e-01 6.33699656e-01
8.24578464e-01 -7.07161963e-01 -2.21534044e-01 -1.32582545e-01
3.58983040e-01 3.70433033e-01 1.02294438e-01 -9.82555509e-01
8.98532048e-02 -4.96956468e-01 1.72711492e-01 -4.30655688e-01
6.21032894e-01 -9.64432716e-01 2.31447086e-01 2.73540527e-01
-4.37767446e-01 -8.33990350e-02 -2.51332670e-02 5.65197110e-01
-3.12683165e-01 -3.61827731e-01 8.64206314e-01 -5.38782179e-02
-8.45216036e-01 4.11814839e-01 -2.08134815e-01 -8.03692117e-02
1.48387718e+00 7.97303915e-02 -1.78319067e-01 -1.48962528e-01
-9.87111390e-01 1.40756354e-01 -1.44795820e-01 5.43693125e-01
3.17907542e-01 -1.20712912e+00 -6.78907216e-01 -8.87009948e-02
2.57824659e-01 -4.16509092e-01 2.58333087e-01 9.86074626e-01
-1.78058334e-02 4.52004641e-01 -4.08599555e-01 -4.68920499e-01
-1.34615839e+00 8.51442754e-01 3.00509691e-01 -9.69695032e-01
-3.35448921e-01 8.57735813e-01 9.74943303e-03 -6.76511884e-01
4.87694532e-01 2.82610118e-01 -4.54565704e-01 5.68311870e-01
4.99380380e-01 5.20832896e-01 -2.29462031e-02 -3.53507251e-01
-1.53604642e-01 1.42491892e-01 -1.37340605e-01 3.15568388e-01
1.82099366e+00 1.19651906e-01 -1.55535026e-03 9.87078786e-01
1.14685035e+00 -6.66656196e-01 -9.96436238e-01 -5.61642647e-01
3.78722399e-01 -1.64344192e-01 -2.16978863e-01 -1.06898272e+00
-8.02151263e-01 9.88071024e-01 6.47407949e-01 4.56302315e-01
9.99817908e-01 -1.00378200e-01 4.09457296e-01 5.69679320e-01
4.83329326e-01 -1.14086592e+00 2.20822230e-01 7.79572487e-01
7.28513420e-01 -1.68005872e+00 -3.91571641e-01 -1.17382124e-01
-7.03346670e-01 1.17201412e+00 6.38750732e-01 1.39118254e-01
7.80079365e-01 -2.52181459e-02 -8.55233371e-02 -1.57630324e-01
-1.01097190e+00 1.45781022e-02 3.13712597e-01 6.49321377e-01
5.34721732e-01 6.77733719e-02 -1.71198711e-01 7.31649876e-01
2.51790315e-01 1.92571338e-02 1.54583037e-01 6.23056948e-01
-5.41558802e-01 -1.31202126e+00 -4.24448103e-01 6.78675950e-01
-5.03491819e-01 -9.75624546e-02 -1.07437819e-01 7.23854423e-01
1.22963138e-01 9.22071040e-01 -3.57993364e-01 -4.36389297e-01
9.56920013e-02 5.63882709e-01 1.57684579e-01 -5.78387201e-01
-7.80018628e-01 -1.77510157e-01 1.74297303e-01 -4.47110385e-01
-4.88383770e-01 -5.41783035e-01 -3.35503280e-01 -2.55478561e-01
-2.19204426e-01 1.13273494e-01 6.04036927e-01 9.79448855e-01
4.41377074e-01 2.78189182e-01 6.48533821e-01 -9.17906463e-01
-8.30481350e-01 -1.41939509e+00 -8.57902765e-01 4.48021680e-01
2.25194581e-02 -1.01956439e+00 -8.51926059e-02 -1.12894326e-01] | [9.153657913208008, 4.204662322998047] |
7e10b990-a63b-4af7-94b8-ea6215406530 | exploring-social-influence-for-recommendation | 1109.0758 | null | https://arxiv.org/abs/1109.0758v1 | https://arxiv.org/pdf/1109.0758v1.pdf | Exploring Social Influence for Recommendation - A Probabilistic Generative Model Approach | In this paper, we propose a probabilistic generative model, called unified model, which naturally unifies the ideas of social influence, collaborative filtering and content-based methods for item recommendation. To address the issue of hidden social influence, we devise new algorithms to learn the model parameters of our proposal based on expectation maximization (EM). In addition to a single-machine version of our EM algorithm, we further devise a parallelized implementation on the Map-Reduce framework to process two large-scale datasets we collect. Moreover, we show that the social influence obtained from our generative models can be used for group recommendation. Finally, we conduct comprehensive experiments using the datasets crawled from last.fm and whrrl.com to validate our ideas. Experimental results show that the generative models with social influence significantly outperform those without incorporating social influence. The unified generative model proposed in this paper obtains the best performance. Moreover, our study on social influence finds that users in whrrl.com are more likely to get influenced by friends than those in last.fm. The experimental results also confirm that our social influence based group recommendation algorithm outperforms the state-of-the-art algorithms for group recommendation. | ['Wang-Chien Lee', 'Xingjie Liu', 'Mao Ye'] | 2011-09-04 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-3.32013845e-01 1.10183179e-01 -2.03046381e-01 -4.02268857e-01
-3.21659148e-01 -1.97743341e-01 8.47738922e-01 -2.97729492e-01
-3.01881015e-01 8.02436650e-01 6.11277103e-01 -2.49414593e-01
-5.47629356e-01 -1.39066601e+00 -7.97592461e-01 -6.93208933e-01
-1.14784501e-02 6.37366354e-01 2.91344404e-01 -2.37428471e-01
4.72848684e-01 -2.08806768e-01 -1.54684126e+00 3.27374667e-01
8.40210617e-01 1.47862926e-01 7.00232029e-01 4.51415569e-01
-2.36796543e-01 6.32427812e-01 -4.07962769e-01 -6.01987839e-01
4.16055098e-02 -4.55540448e-01 -4.16415870e-01 1.52718112e-01
-1.37950122e-01 -2.62397140e-01 -3.52671951e-01 7.42854059e-01
6.87045395e-01 3.76239091e-01 5.78599751e-01 -1.08067942e+00
-7.63858020e-01 1.34766817e+00 -4.87758368e-01 6.72527701e-02
5.00408828e-01 -6.21527851e-01 9.84324455e-01 -8.92061591e-01
6.05203927e-01 1.54685903e+00 4.71688300e-01 1.18428543e-01
-8.77892673e-01 -4.90061343e-01 6.68622255e-01 8.78084451e-02
-1.26205981e+00 -3.18155773e-02 5.21279454e-01 -1.21654175e-01
6.75381839e-01 1.70389041e-01 7.51826644e-01 8.96866560e-01
1.42055497e-01 1.25497782e+00 1.05338180e+00 -3.30918759e-01
9.68472958e-02 1.90411627e-01 1.93311378e-01 3.79830956e-01
4.89492655e-01 -1.74712047e-01 -7.51750588e-01 -7.05542266e-01
6.55128658e-01 5.30047417e-01 -9.51669440e-02 3.03948671e-02
-9.44027603e-01 1.12666750e+00 2.67308876e-02 5.62398314e-01
-4.84213918e-01 2.33742326e-01 -4.33682382e-01 4.78638083e-01
9.71328557e-01 6.16826564e-02 -2.89562047e-01 -9.91934910e-02
-9.02649403e-01 2.18436066e-02 1.29495978e+00 8.95276845e-01
7.51383245e-01 -5.21137714e-01 1.34844959e-01 8.97022665e-01
1.26729310e+00 6.13446891e-01 1.15940988e-01 -5.43642163e-01
8.67304429e-02 3.33994508e-01 1.06740169e-01 -1.36748838e+00
-1.44349992e-01 -7.60480344e-01 -6.36086345e-01 -6.64177418e-01
1.12710007e-01 -5.00263393e-01 -2.12334588e-01 1.28543460e+00
4.00058717e-01 5.66865027e-01 -1.69760481e-01 6.15277052e-01
7.29848623e-01 8.26089740e-01 -3.96408662e-02 -4.35455859e-01
8.18697870e-01 -1.09370637e+00 -8.93368423e-01 2.37948284e-01
6.98033214e-01 -1.02819574e+00 4.17299867e-01 4.15150613e-01
-9.67660129e-01 -5.69033921e-01 -9.39709663e-01 6.37051642e-01
-6.70743436e-02 -6.36164099e-02 1.12326109e+00 8.29402983e-01
-1.23760688e+00 7.87022948e-01 -7.94432640e-01 -4.22169089e-01
5.98751307e-02 4.25196886e-01 2.38305852e-01 -2.59661764e-01
-1.30955863e+00 3.71419430e-01 -1.07647195e-01 -1.62984729e-01
-6.80703938e-01 -4.20692444e-01 -7.40140229e-02 -1.43915907e-01
5.35269380e-01 -1.10536182e+00 8.87079179e-01 -6.80274606e-01
-1.53229594e+00 2.07772568e-01 -4.53639805e-01 -2.32641399e-01
2.55690187e-01 -4.98821467e-01 -5.09157121e-01 -2.54391462e-01
-1.00136116e-01 -4.26017493e-02 5.98869979e-01 -1.41979241e+00
-8.70953500e-01 -2.52863199e-01 1.61062792e-01 1.22356668e-01
-5.24201810e-01 2.30462000e-01 -7.22818315e-01 -4.49316204e-01
-2.53748205e-02 -1.18980169e+00 -5.93954444e-01 -1.05204272e+00
-3.23631197e-01 -5.92308879e-01 4.49662954e-01 -1.57705948e-01
1.77284133e+00 -1.75117302e+00 4.63179462e-02 6.98221624e-01
3.88722032e-01 3.61570083e-02 2.91638095e-02 8.28557611e-01
5.53369224e-01 4.09601361e-01 4.21607435e-01 -4.41658109e-01
1.62416145e-01 2.67942697e-01 -1.84321776e-01 3.78199130e-01
-7.51281440e-01 9.33665395e-01 -9.49277878e-01 -2.47698754e-01
3.28357294e-02 6.44232631e-01 -7.24715590e-01 1.34192437e-01
-1.46577433e-01 3.41607243e-01 -1.04146469e+00 4.77956235e-02
7.74112523e-01 -7.13517666e-01 6.59612656e-01 2.14391023e-01
-4.43067439e-02 4.88012910e-01 -1.40175378e+00 1.38711739e+00
-3.67537707e-01 -4.91035245e-02 2.43419707e-02 -3.57041627e-01
1.02081156e+00 1.66723743e-01 6.51445568e-01 -4.07627746e-02
1.56038254e-01 -8.38853642e-02 -4.53711953e-03 -1.96222931e-01
6.57601237e-01 2.20944226e-01 1.17938854e-01 1.54501283e+00
-1.75903425e-01 4.39551264e-01 3.64785999e-01 8.69589388e-01
9.93439972e-01 -1.17140427e-01 7.78706372e-02 -3.79419595e-01
3.10595274e-01 -5.21257639e-01 3.41481984e-01 1.35459924e+00
4.01775777e-01 9.51146036e-02 5.09933047e-02 -5.25515154e-02
-6.05214298e-01 -7.13212371e-01 3.72383222e-02 1.51223993e+00
1.77756667e-01 -1.13904810e+00 -6.15616679e-01 -6.42809987e-01
1.34702280e-01 7.22760737e-01 -6.10066772e-01 2.27933183e-01
-2.31221721e-01 -1.51635802e+00 -9.42023024e-02 1.48653701e-01
3.52076828e-01 -6.31139934e-01 5.58238566e-01 4.74292815e-01
-1.08157694e-01 -6.40621662e-01 -3.40293318e-01 -4.56956655e-01
-9.37415004e-01 -9.28612292e-01 -4.59808141e-01 -3.21098417e-01
7.92449117e-01 7.74661601e-01 1.31877518e+00 5.33899844e-01
5.48519671e-01 8.92359495e-01 -9.93916690e-01 -4.81249928e-01
-2.99753338e-01 1.16658352e-01 1.08737357e-01 3.33212256e-01
8.96549225e-01 -8.74290764e-01 -6.95799828e-01 1.05848312e+00
-8.46854508e-01 3.14128287e-02 5.06102443e-01 1.55083939e-01
5.43747008e-01 2.57752001e-01 7.33926237e-01 -1.21687603e+00
9.78083432e-01 -1.07219040e+00 -1.44922197e-01 9.65762585e-02
-1.12306130e+00 -8.53983611e-02 8.11175257e-03 -3.66103828e-01
-1.12284482e+00 -4.59480226e-01 -1.33589968e-01 3.53832126e-01
1.41394604e-02 7.73603261e-01 -1.14430122e-01 2.82542497e-01
2.89558500e-01 -3.01947296e-02 -1.99543357e-01 -8.77899051e-01
5.56755960e-01 9.22501981e-01 -4.47488606e-01 -4.16216969e-01
6.29537761e-01 7.59479046e-01 -4.31612164e-01 -6.56027377e-01
-1.04898024e+00 -6.57513499e-01 -5.31558096e-01 -4.11091477e-01
5.61814070e-01 -9.17741776e-01 -4.19886053e-01 3.49513531e-01
-8.14094186e-01 -6.60324004e-03 3.94970685e-01 9.09007728e-01
-3.19383740e-01 4.01649177e-01 -8.11841786e-01 -9.25900519e-01
-3.30868989e-01 -6.61883175e-01 6.21203542e-01 9.37717333e-02
-1.63077772e-01 -1.36165679e+00 3.36622238e-01 4.80745286e-01
5.86392641e-01 -2.68311501e-01 1.37775347e-01 -8.73687685e-01
-8.78383279e-01 -7.94517174e-02 1.65335149e-01 -7.41241947e-02
5.49250804e-02 1.53149337e-01 -4.41072375e-01 -3.92486185e-01
-1.18363842e-01 4.90834713e-01 8.07859838e-01 6.04166985e-01
6.02669120e-01 2.71206931e-03 -8.20739388e-01 1.39386132e-01
1.09827793e+00 1.04343541e-01 8.01002324e-01 4.03827220e-01
6.55951917e-01 3.76459062e-01 9.17665601e-01 7.99625039e-01
6.68477535e-01 5.39970636e-01 1.31545305e-01 -9.39082727e-03
1.33933485e-01 -5.69764137e-01 5.59336603e-01 1.67630780e+00
-6.08977616e-01 -6.23970449e-01 -3.51659477e-01 2.72685021e-01
-2.37095141e+00 -1.08622527e+00 -7.46723831e-01 2.13975906e+00
4.40954745e-01 -2.73240119e-01 2.53200591e-01 -2.40980819e-01
9.87289190e-01 -2.99987681e-02 3.96727249e-02 -3.25140841e-02
-2.21757919e-01 7.00190058e-03 4.38102931e-01 6.63897693e-01
-8.56471419e-01 9.03447688e-01 6.64374018e+00 8.74192715e-01
-2.18210205e-01 4.99258846e-01 3.68346870e-01 -4.86031950e-01
-6.64791524e-01 7.55077675e-02 -1.37124443e+00 6.63558841e-01
1.18014443e+00 -4.84696060e-01 4.19760197e-01 6.70748889e-01
5.82442045e-01 -2.00965345e-01 -7.16206551e-01 5.53955495e-01
3.20283502e-01 -1.12319446e+00 -1.04425043e-01 5.95109224e-01
1.47580504e+00 4.06027943e-01 -2.15403467e-01 6.63283169e-02
9.97897267e-01 -6.52474523e-01 1.09748743e-01 1.06747317e+00
-4.27689850e-01 -9.41898048e-01 8.85037124e-01 6.10242546e-01
-8.59943688e-01 1.50626883e-01 -8.56280744e-01 -2.33863384e-01
7.79996872e-01 1.46218610e+00 -7.10754693e-01 7.70019770e-01
6.43489361e-01 1.17344737e+00 -2.43272126e-01 1.03054142e+00
-7.37058163e-01 1.16858280e+00 -3.10382754e-01 -7.52414763e-01
2.77194032e-03 -5.73805273e-01 5.86097300e-01 1.14433277e+00
5.67189217e-01 5.34621440e-02 2.02231511e-01 5.04235804e-01
-1.23599648e-01 4.88663465e-01 -3.16544026e-01 -5.56431673e-02
3.18225175e-01 1.16722739e+00 -5.98598361e-01 -5.23733199e-01
-4.88285184e-01 5.88076770e-01 -4.56306245e-03 2.47528151e-01
-7.05241323e-01 2.28564754e-01 3.22833657e-01 5.55450380e-01
6.19317293e-01 -3.49449277e-01 1.36972010e-01 -1.16766071e+00
-4.23663437e-01 -6.16880953e-01 1.96873680e-01 -5.58623016e-01
-1.47498870e+00 5.12552142e-01 6.56377450e-02 -9.34832036e-01
-3.79296720e-01 -2.03688279e-01 -6.09509945e-01 7.15568602e-01
-1.10815656e+00 -7.96275795e-01 4.45973501e-02 6.60441995e-01
1.41571164e-01 -2.51217037e-01 5.70176363e-01 7.40473986e-01
-3.07668000e-01 2.32171893e-01 4.44275081e-01 -3.37701917e-01
7.94430554e-01 -9.97698903e-01 2.14492425e-01 5.80290973e-01
3.79227430e-01 1.16376960e+00 5.61527312e-01 -9.93801415e-01
-1.48525906e+00 -9.41162229e-01 1.18084025e+00 -7.66367197e-01
8.92052591e-01 -1.59538433e-01 -4.26090032e-01 7.31880665e-01
2.68211216e-01 -7.76151538e-01 1.44076788e+00 8.86260748e-01
1.91016987e-01 3.07041407e-01 -7.52801180e-01 5.20093620e-01
1.44082510e+00 -1.24764547e-01 -6.08003855e-01 5.13246357e-01
5.82417905e-01 4.89671439e-01 -1.16288054e+00 1.63875759e-01
5.86320579e-01 -1.01862228e+00 7.30365694e-01 -5.38895249e-01
4.17789876e-01 -5.91491044e-01 -2.84379303e-01 -1.59937751e+00
-9.10526991e-01 -7.35694110e-01 -2.28108898e-01 1.31905556e+00
7.21550047e-01 -7.50515282e-01 6.80218458e-01 5.32548845e-01
2.88462669e-01 -4.90296274e-01 -1.00277610e-01 -3.79187793e-01
-1.87105998e-01 -6.69335485e-01 7.01199710e-01 1.02926290e+00
-2.53149439e-02 4.98117626e-01 -7.50564337e-01 -4.13393788e-02
7.22038567e-01 4.26527977e-01 1.21519446e+00 -1.37677908e+00
-1.03929508e+00 1.23516716e-01 -4.69968207e-02 -1.58778799e+00
-1.32231712e-01 -9.34719086e-01 -4.72422421e-01 -1.96451020e+00
7.31503427e-01 -5.52071869e-01 -5.61776161e-01 1.37554437e-01
-2.71472722e-01 1.73241511e-01 1.49752408e-01 4.86886382e-01
-1.18378723e+00 4.47641790e-01 1.59857786e+00 2.38559380e-01
-3.74182850e-01 4.25089598e-01 -1.24495471e+00 7.32259393e-01
6.82941139e-01 -8.80317271e-01 -3.85804951e-01 -2.51090378e-01
8.97805870e-01 -4.23917323e-01 -4.53817874e-01 -2.68678278e-01
4.69422221e-01 4.56114672e-02 1.66763570e-02 -8.63992333e-01
-1.32673636e-01 -4.43453312e-01 5.84870458e-01 5.00939712e-02
-1.55982554e-01 -2.83352107e-01 -5.00120819e-01 8.79559398e-01
4.66958154e-03 -2.82586008e-01 7.61483461e-02 -1.10257469e-01
-7.69373178e-02 3.23060781e-01 -7.41306782e-01 -6.49138927e-01
5.79515636e-01 2.32244477e-01 -2.20847890e-01 -9.82077062e-01
-1.21547365e+00 2.53864616e-01 2.17090979e-01 5.70793688e-01
3.21772486e-01 -1.32469165e+00 -1.23321986e+00 -3.23833793e-01
-9.55895409e-02 -6.81077898e-01 1.71474248e-01 1.20196736e+00
3.05729479e-01 3.49151075e-01 4.39388573e-01 -3.25150579e-01
-1.40191674e+00 4.29082602e-01 -2.69682437e-01 -5.69435060e-01
-7.19674155e-02 7.18036175e-01 6.20890483e-02 -6.71235263e-01
-1.37702644e-01 3.12962651e-01 -3.36319208e-01 -3.03866602e-02
9.00154233e-01 7.05601037e-01 -1.99768424e-01 -4.53479946e-01
6.77898005e-02 2.50726670e-01 -1.75732344e-01 -5.01192994e-02
1.74602711e+00 -6.87551916e-01 -5.40018320e-01 3.93353969e-01
8.67978334e-01 6.15102172e-01 -6.07443690e-01 -5.17057717e-01
-2.78060764e-01 -5.43799281e-01 5.09243369e-01 -7.49782324e-01
-1.01905298e+00 1.19804479e-01 1.13767885e-01 5.30021369e-01
5.76366365e-01 3.41921359e-01 5.71967781e-01 5.47912836e-01
8.63543272e-01 -1.17030704e+00 -2.86167443e-01 6.10198259e-01
6.04293227e-01 -8.83636713e-01 2.77824372e-01 -7.04123914e-01
-2.36485973e-01 6.07375681e-01 5.74601591e-02 -3.53024989e-01
1.39923143e+00 2.60270834e-01 -2.40981802e-01 -2.54257351e-01
-1.17996979e+00 -3.85342747e-01 3.49338472e-01 1.42670006e-01
7.66553640e-01 3.39549452e-01 -1.08505690e+00 9.73107576e-01
-1.38070852e-01 6.18505739e-02 4.31956559e-01 8.88632774e-01
-9.15135145e-01 -1.73212540e+00 -5.49077153e-01 7.88915396e-01
-5.75796187e-01 -1.75584748e-01 -7.86944985e-01 4.25067276e-01
-4.41436917e-02 1.70881259e+00 -4.81245108e-02 -8.44072163e-01
-1.12096354e-01 -1.74818203e-01 4.14950043e-01 -5.63449442e-01
-6.19126856e-01 6.72129035e-01 3.91504765e-01 -3.08058292e-01
-9.31594908e-01 -7.86797464e-01 -8.50656450e-01 -8.17515373e-01
-6.92165911e-01 7.07145631e-01 9.41937089e-01 9.85664189e-01
7.07686543e-01 4.30729628e-01 9.41528797e-01 -6.56855822e-01
-4.78442498e-02 -1.42572594e+00 -8.33433747e-01 2.69973814e-01
-7.89483428e-01 -7.35201240e-01 -6.93686724e-01 -2.29871809e-01] | [9.94609546661377, 5.656081676483154] |
be372652-8029-4045-b38f-5dd31ec8e87d | who-you-play-affects-how-you-play-predicting | 2303.16741 | null | https://arxiv.org/abs/2303.16741v1 | https://arxiv.org/pdf/2303.16741v1.pdf | Who You Play Affects How You Play: Predicting Sports Performance Using Graph Attention Networks With Temporal Convolution | This study presents a novel deep learning method, called GATv2-GCN, for predicting player performance in sports. To construct a dynamic player interaction graph, we leverage player statistics and their interactions during gameplay. We use a graph attention network to capture the attention that each player pays to each other, allowing for more accurate modeling of the dynamic player interactions. To handle the multivariate player statistics time series, we incorporate a temporal convolution layer, which provides the model with temporal predictive power. We evaluate the performance of our model using real-world sports data, demonstrating its effectiveness in predicting player performance. Furthermore, we explore the potential use of our model in a sports betting context, providing insights into profitable strategies that leverage our predictive power. The proposed method has the potential to advance the state-of-the-art in player performance prediction and to provide valuable insights for sports analytics and betting industries. | ['Vikram Krishnamurthy', 'Rui Luo'] | 2023-03-29 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [-4.47125793e-01 -2.42819384e-01 -4.78458941e-01 -8.41358379e-02
-1.99968457e-01 -3.59975487e-01 1.46924421e-01 2.10799128e-01
-2.99941123e-01 3.27017605e-01 4.63863254e-01 -2.47960538e-01
-3.76838386e-01 -1.28413999e+00 -5.40084660e-01 -2.49757722e-01
-5.11231899e-01 5.16791701e-01 3.55396807e-01 -6.36457980e-01
1.94003761e-01 2.19149619e-01 -1.50841296e+00 3.92270565e-01
6.03446722e-01 9.44509625e-01 -1.28227398e-01 8.51792574e-01
1.17623001e-01 1.46787119e+00 -7.85888255e-01 -8.95567775e-01
3.30292076e-01 -4.12102342e-01 -6.08622849e-01 -1.17077060e-01
-7.82581717e-02 -2.00493902e-01 -1.04610634e+00 4.38853681e-01
3.44299704e-01 6.48037374e-01 2.78201312e-01 -1.22490883e+00
-4.46711540e-01 9.21874940e-01 -6.34704649e-01 8.18701804e-01
3.65927309e-01 6.13833010e-01 1.35216880e+00 1.42395377e-01
4.35350090e-01 8.00679803e-01 8.11300099e-01 2.87457734e-01
-9.96094227e-01 -8.00664365e-01 2.96780169e-01 6.80423498e-01
-8.50329578e-01 1.70816675e-01 8.02409172e-01 -5.83836317e-01
8.47873747e-01 2.37623259e-01 1.45036733e+00 1.00423336e+00
1.85641572e-01 1.18949640e+00 5.46963453e-01 5.05917557e-02
-2.10833356e-01 -6.63137734e-01 4.95563418e-01 5.51222146e-01
-2.53774881e-01 2.56520003e-01 -1.12658691e+00 2.23527595e-01
8.33695471e-01 2.57064164e-01 2.32664466e-01 2.50693977e-01
-6.15984559e-01 8.55910003e-01 5.68292499e-01 2.67787538e-02
-6.04632080e-01 6.89515591e-01 6.64232433e-01 2.05186859e-01
8.23985338e-01 4.75193143e-01 1.56591125e-02 -1.02592683e+00
-8.92947197e-01 6.29025578e-01 8.56536388e-01 4.79257941e-01
1.06001928e-01 1.49605110e-01 -8.23090255e-01 7.50548840e-01
-2.11695388e-01 -2.06410646e-01 2.62491882e-01 -9.73222494e-01
4.95082468e-01 8.69345009e-01 -3.41223568e-01 -1.16799343e+00
-5.50445080e-01 -7.81272948e-01 -3.33300382e-01 -3.20095956e-01
4.29464787e-01 1.54343635e-01 -5.51259458e-01 1.64752913e+00
-5.17943129e-02 7.79118657e-01 -4.39629227e-01 9.39440727e-01
9.69049454e-01 4.56575423e-01 1.38311505e-01 3.78288895e-01
1.03242087e+00 -1.11909485e+00 -4.92658347e-01 -1.92933813e-01
4.98497248e-01 -1.16743341e-01 1.14372301e+00 3.34242553e-01
-1.22872484e+00 -5.80618382e-01 -5.59923530e-01 3.51537503e-02
9.43415090e-02 -1.53348029e-01 1.24792981e+00 3.30516636e-01
-6.65789902e-01 9.31900203e-01 -1.26916230e+00 -4.81256433e-02
6.22945786e-01 3.51065367e-01 1.70659259e-01 2.37999067e-01
-1.35102379e+00 6.37953341e-01 3.39685678e-01 -1.16183152e-02
-8.79011393e-01 -9.52429056e-01 -6.28216386e-01 4.51752633e-01
7.34418452e-01 -5.23243248e-01 1.19158173e+00 -5.09475291e-01
-1.51136708e+00 7.52494872e-01 1.16491549e-01 -8.01665604e-01
6.54591322e-01 -4.19887036e-01 -2.25336224e-01 -2.91153669e-01
1.61030382e-01 -5.63726649e-02 7.04924986e-02 -3.76587719e-01
-6.11885726e-01 -3.14754874e-01 2.70972908e-01 3.41135472e-01
-5.58761358e-01 3.41457059e-03 -1.17132020e+00 -6.31178916e-01
-2.97600836e-01 -8.52363646e-01 -3.49035293e-01 -7.88692534e-01
-4.12743747e-01 -5.78474879e-01 2.30748609e-01 -8.52905929e-01
1.98164582e+00 -1.99386954e+00 3.27111840e-01 4.05841440e-01
5.68437040e-01 1.92759961e-01 -7.74703398e-02 7.40165651e-01
2.76798278e-01 -9.22556445e-02 5.91337919e-01 -2.91413903e-01
1.31307226e-02 2.26341650e-01 -6.60303012e-02 2.26783186e-01
-7.56769776e-02 1.58762741e+00 -7.96872318e-01 -7.87172243e-02
9.38154161e-02 7.63062090e-02 -9.05627072e-01 2.60113150e-01
-1.92590028e-01 4.68072742e-01 -6.51611090e-01 3.77825558e-01
3.88037972e-02 -2.74724573e-01 2.59226590e-01 1.94225997e-01
-8.73078033e-02 6.79563165e-01 -7.12826252e-01 1.44947696e+00
-1.37472406e-01 8.71181786e-01 -2.25931749e-01 -1.19502223e+00
8.57459128e-01 -2.53844678e-01 8.95593762e-01 -8.59565377e-01
3.66752625e-01 -3.93657237e-01 3.40974480e-01 -5.97136080e-01
8.18883359e-01 2.45094180e-01 -5.10815620e-01 6.77661479e-01
-3.99234779e-02 4.72892940e-01 6.27487659e-01 4.73531485e-01
1.51383352e+00 7.12269619e-02 -1.11480042e-01 1.05919689e-01
-4.28820476e-02 8.57927352e-02 6.87380135e-01 9.68089521e-01
-1.71842709e-01 2.58670691e-02 9.99747336e-01 -7.32624769e-01
-8.76397431e-01 -7.48649836e-01 6.17996216e-01 1.72420228e+00
1.03664041e-01 -7.79906392e-01 -4.87504900e-01 -1.86954647e-01
3.86562556e-01 2.95362711e-01 -8.93691838e-01 -5.36769092e-01
-8.32410038e-01 -4.75602537e-01 7.70057499e-01 9.64335918e-01
5.45846820e-01 -1.19069910e+00 -4.55678344e-01 5.19135177e-01
-3.92811358e-01 -9.29551363e-01 -6.19731247e-01 -1.15781829e-01
-6.33358955e-01 -1.26127183e+00 -2.90901065e-01 -3.92949522e-01
-4.00537133e-01 3.53389122e-02 1.59832108e+00 3.46505255e-01
-1.87926635e-01 4.62986529e-01 -4.09161329e-01 -4.85900164e-01
-1.23978868e-01 5.45685530e-01 -6.58079535e-02 -1.30122960e-01
7.19001293e-01 -7.61142850e-01 -6.69645846e-01 2.57984519e-01
-5.13143539e-01 1.76293343e-01 4.58051302e-02 7.42780387e-01
3.91756088e-01 1.32243663e-01 3.42725098e-01 -1.09301019e+00
1.15703094e+00 -7.47499585e-01 -3.64471197e-01 -5.80414459e-02
-5.18834889e-01 -3.21397543e-01 2.94652134e-01 -6.98968709e-01
-5.62366724e-01 -4.67254877e-01 -4.78867851e-02 -5.91641903e-01
3.99856180e-01 9.67109442e-01 3.53862315e-01 1.82584718e-01
5.04545331e-01 2.05719516e-01 1.47052929e-01 -5.61722815e-01
2.78000951e-01 1.70026332e-01 8.96880627e-01 -7.02487588e-01
5.27081728e-01 2.74224758e-01 8.44797790e-02 -3.43315244e-01
-1.07496369e+00 -8.21083307e-01 -3.70370716e-01 -6.91814661e-01
8.80530655e-01 -9.84994233e-01 -1.78114164e+00 6.25586689e-01
-6.44176185e-01 -6.04942620e-01 -4.62505043e-01 3.46990168e-01
-5.92254281e-01 3.32169910e-03 -1.34613073e+00 -9.24252868e-01
-1.99719369e-01 -6.82904840e-01 6.15095973e-01 5.22448897e-01
-3.06921899e-01 -1.10690832e+00 1.88215613e-01 9.91464198e-01
1.15487717e-01 3.16284001e-01 5.54377615e-01 -8.73157024e-01
-6.51740849e-01 -3.87960374e-01 7.09771216e-02 -1.01180054e-01
-3.30644935e-01 -1.38747871e-01 -4.80559468e-01 -1.52100427e-02
-7.32866883e-01 -2.75624126e-01 1.18867934e+00 6.80696011e-01
1.55758965e+00 8.28782991e-02 -2.92446941e-01 8.36795151e-01
7.89076746e-01 -4.87991273e-02 6.83663547e-01 5.38071394e-01
1.00833297e+00 4.81488734e-01 6.71022117e-01 7.64744341e-01
8.34116399e-01 7.86398411e-01 4.46550250e-01 -1.41605079e-01
1.97288804e-02 -8.37084115e-01 1.61110505e-01 8.16706240e-01
-9.61037040e-01 -4.97606725e-01 -9.05463159e-01 4.62829500e-01
-2.27177906e+00 -1.39520013e+00 -3.00847471e-01 1.73736894e+00
3.48570853e-01 4.31176662e-01 1.00284457e+00 1.82976887e-01
5.98360181e-01 2.93183953e-01 -8.47994447e-01 -4.67856526e-01
9.83845666e-02 5.46173871e-01 5.85194290e-01 -3.67314517e-02
-9.99322474e-01 1.47321427e+00 6.66024828e+00 1.10699987e+00
-7.81959772e-01 3.88228036e-02 5.16955853e-01 -7.47868717e-01
2.56708451e-02 -1.97386399e-01 -5.20723641e-01 5.48900723e-01
8.81376326e-01 -3.41113836e-01 5.75553060e-01 7.01821864e-01
5.36616027e-01 9.25248936e-02 -8.58355343e-01 8.52300107e-01
-1.14402279e-01 -1.62227869e+00 -2.49889851e-01 5.70168018e-01
4.92540240e-01 -7.79248849e-02 2.17862770e-01 8.22007477e-01
7.97665954e-01 -1.20819366e+00 6.56405032e-01 8.41340959e-01
4.42847371e-01 -1.06952286e+00 6.33479357e-01 4.56463903e-01
-1.44914424e+00 -4.37738687e-01 -4.39462401e-02 -9.59986329e-01
2.92772889e-01 6.81129321e-02 -6.00805402e-01 3.68326873e-01
8.55733335e-01 1.33006299e+00 -3.97808194e-01 1.01694047e+00
-4.88329232e-02 1.18776608e+00 -6.32046610e-02 -2.57668912e-01
2.82367259e-01 -4.64094400e-01 5.16242087e-01 8.31444025e-01
8.33790973e-02 4.07309175e-01 5.14758170e-01 9.23312664e-01
-2.24058196e-01 1.05685949e-01 -2.98292130e-01 -6.52736008e-01
4.39572692e-01 9.18364465e-01 -6.49306476e-01 -1.24068469e-01
-3.67550939e-01 7.38346815e-01 7.26536691e-01 1.78185970e-01
-1.08481610e+00 7.98058212e-02 1.05945468e+00 5.72504461e-01
2.02308148e-01 -2.86079407e-01 -5.01988769e-01 -1.11578953e+00
-1.60614878e-01 -7.34997034e-01 8.79182994e-01 -4.60281044e-01
-1.39274251e+00 1.16311587e-01 -2.34983742e-01 -1.07425010e+00
-2.27374420e-01 -3.15364033e-01 -1.22962129e+00 7.48332918e-01
-8.75041902e-01 -1.16474259e+00 -2.90894449e-01 5.78809798e-01
5.03489196e-01 -5.25967717e-01 1.46043792e-01 2.44091555e-01
-8.32347751e-01 7.68609166e-01 -2.82942086e-01 5.24540961e-01
2.33565062e-01 -9.80937600e-01 5.87041259e-01 6.31958246e-01
3.35473150e-01 2.64906347e-01 4.88065064e-01 -8.28477144e-01
-1.23675668e+00 -8.80717039e-01 3.38941693e-01 -5.08257210e-01
1.06763864e+00 -2.05902576e-01 -7.93618917e-01 7.94329226e-01
-1.65245786e-01 -3.67996544e-01 1.24263000e+00 9.92396951e-01
-4.79477011e-02 -1.26971239e-02 -2.96081722e-01 5.63546121e-01
1.43548918e+00 -5.44724941e-01 -3.34522247e-01 1.22464925e-01
5.64956605e-01 -7.36836374e-01 -1.08937883e+00 1.43574879e-01
7.27695525e-01 -9.17541504e-01 1.02064157e+00 -1.30386817e+00
9.02586877e-01 3.99341404e-01 3.78675014e-01 -1.41846859e+00
-8.02225828e-01 -5.08913875e-01 -4.32988852e-01 8.84709537e-01
3.03128242e-01 -2.36214980e-01 1.51361465e+00 9.61483359e-01
-2.35224083e-01 -1.14560318e+00 -7.23693669e-01 -6.61885202e-01
4.33101021e-02 -1.05264699e+00 7.83964694e-01 8.26785326e-01
3.00075561e-01 1.17544904e-01 -8.45293581e-01 -2.91872621e-01
3.09365481e-01 3.55697542e-01 1.05565751e+00 -1.33435071e+00
-1.04842186e+00 -9.32934880e-01 -9.28002477e-01 -1.31468976e+00
1.86458811e-01 -9.50108767e-01 -5.24983525e-01 -1.32332826e+00
4.78218406e-01 -2.48793676e-01 -4.57913816e-01 4.30058539e-01
-4.09296155e-01 6.01216257e-01 4.86910820e-01 1.40151426e-01
-9.08428848e-01 4.91602719e-01 1.46357477e+00 -1.46306738e-01
-5.79986453e-01 5.50126076e-01 -7.95019031e-01 5.14665902e-01
6.70529306e-01 -2.52249837e-01 -2.09547698e-01 -3.59632641e-01
4.86602455e-01 2.47542888e-01 3.50550026e-01 -8.27863514e-01
3.61444741e-01 -3.73209059e-01 1.09308869e-01 -3.49832237e-01
6.02357030e-01 -6.91710338e-02 5.07350266e-03 2.76666194e-01
-5.48051536e-01 4.72770482e-02 1.75254136e-01 8.68346810e-01
-3.07673484e-01 3.01299542e-01 8.09303820e-02 8.58643837e-03
-8.13949108e-01 9.20062363e-01 -3.78168970e-01 3.36899012e-01
1.05386388e+00 -4.00441349e-01 -1.81071460e-01 -9.42456424e-01
-1.06083918e+00 8.42498958e-01 -2.67026313e-02 4.54326928e-01
3.48286569e-01 -1.40753007e+00 -8.06418121e-01 -1.31719075e-02
7.86736161e-02 -4.63041782e-01 5.97425163e-01 7.70796120e-01
-5.80713928e-01 -4.20605354e-02 -2.39448667e-01 -4.21963632e-01
-1.34597111e+00 2.29644224e-01 3.11113536e-01 -7.13974655e-01
-7.67261744e-01 1.05220199e+00 -1.37687370e-01 -7.93530792e-02
1.51432171e-01 -3.34915216e-03 -6.01111650e-01 1.94734335e-03
2.69443035e-01 7.45743632e-01 -3.66276741e-01 -5.25755584e-01
-7.18229860e-02 5.58866002e-02 8.53172168e-02 9.58136469e-02
1.75754249e+00 2.21618444e-01 2.84416318e-01 6.61239147e-01
5.02107263e-01 -7.63135850e-02 -1.51124585e+00 -2.58372366e-01
-5.24907671e-02 -6.93877459e-01 9.44870934e-02 -5.38683057e-01
-1.57536769e+00 7.96390891e-01 -3.38073261e-02 5.24231315e-01
9.56979036e-01 9.98001322e-02 1.16438448e+00 3.25693898e-02
2.95245439e-01 -1.24603987e+00 2.90209740e-01 6.92094088e-01
4.09101307e-01 -1.07402849e+00 -1.20457105e-01 -2.65257329e-01
-1.06689525e+00 9.17505026e-01 7.75659561e-01 -5.26281238e-01
6.98143542e-01 1.83723360e-01 -1.48978204e-01 -6.40566528e-01
-9.47793841e-01 -5.44091761e-01 6.27239585e-01 3.25247347e-01
4.92709279e-01 5.17712176e-01 -3.57690662e-01 1.26540327e+00
-7.58235455e-01 1.49578571e-01 3.88140082e-01 4.93239254e-01
-9.49737653e-02 -1.06149507e+00 2.63321310e-01 1.02346718e+00
-6.73112452e-01 -5.89585677e-02 -4.64760929e-01 7.57749200e-01
6.71954593e-03 8.92937720e-01 3.28896046e-01 -1.11643314e+00
6.53917730e-01 -7.52570257e-02 3.76728147e-01 -7.02280462e-01
-1.12923801e+00 -1.04272634e-01 2.61566877e-01 -9.64939892e-01
-2.34039072e-02 -5.97799361e-01 -9.57230449e-01 -1.22113633e+00
-1.43120244e-01 1.38891891e-01 1.27710953e-01 9.63083804e-01
3.42319936e-01 1.14992666e+00 3.86620492e-01 -7.53452837e-01
-1.33315071e-01 -9.72930253e-01 -1.00346708e+00 6.26084566e-01
-4.05016899e-01 -8.99296582e-01 1.93761215e-01 -3.71812165e-01] | [6.698396682739258, 0.3401778042316437] |
4771fce6-7e65-417e-a707-3f33b710f968 | time-conditioned-generative-modeling-of | 2301.08951 | null | https://arxiv.org/abs/2301.08951v3 | https://arxiv.org/pdf/2301.08951v3.pdf | Time-Conditioned Generative Modeling of Object-Centric Representations for Video Decomposition and Prediction | When perceiving the world from multiple viewpoints, humans have the ability to reason about the complete objects in a compositional manner even when an object is completely occluded from certain viewpoints. Meanwhile, humans are able to imagine novel views after observing multiple viewpoints. Recent remarkable advances in multi-view object-centric learning still leaves some unresolved problems: 1) The shapes of partially or completely occluded objects can not be well reconstructed. 2) The novel viewpoint prediction depends on expensive viewpoint annotations rather than implicit rules in view representations. In this paper, we introduce a time-conditioned generative model for videos. To reconstruct the complete shape of an object accurately, we enhance the disentanglement between the latent representations of objects and views, where the latent representations of time-conditioned views are jointly inferred with a Transformer and then are input to a sequential extension of Slot Attention to learn object-centric representations. In addition, Gaussian processes are employed as priors of view latent variables for video generation and novel-view prediction without viewpoint annotations. Experiments on multiple datasets demonstrate that the proposed model can make object-centric video decomposition, reconstruct the complete shapes of occluded objects, and make novel-view predictions. | ['Bin Li', 'Chengmin Gao'] | 2023-01-21 | null | null | null | null | ['video-generation'] | ['computer-vision'] | [ 1.93070754e-01 2.14874417e-01 -3.68552841e-02 -5.55670023e-01
-4.61961061e-01 -6.40912354e-01 7.02491105e-01 -6.15047872e-01
3.46765667e-01 4.01210904e-01 4.71353173e-01 3.14213663e-01
7.82555193e-02 -4.98793721e-01 -9.05491292e-01 -7.82400191e-01
5.03364325e-01 8.40812027e-01 -2.36463454e-02 1.45467103e-01
9.84728858e-02 4.34775591e-01 -1.77538943e+00 8.37870955e-01
4.24188465e-01 6.59956872e-01 6.22463107e-01 7.47894049e-01
7.00696558e-02 1.00507367e+00 -2.97596812e-01 -3.76769364e-01
2.11750954e-01 -1.44740254e-01 -4.36080098e-01 9.24680412e-01
4.91499752e-01 -7.13697553e-01 -5.47205806e-01 9.66569304e-01
-6.94856420e-02 2.45662659e-01 7.14828372e-01 -1.26365900e+00
-9.66779053e-01 3.60353947e-01 -5.61980009e-01 -9.90724564e-02
6.19366765e-01 1.05182119e-01 9.65107799e-01 -1.19560122e+00
8.42236459e-01 1.61754727e+00 -1.78254172e-01 6.70264423e-01
-1.28729832e+00 -2.29921505e-01 1.05289376e+00 5.68489373e-01
-1.21889973e+00 -6.71966672e-01 1.02518559e+00 -7.16229916e-01
8.52890015e-01 2.67898589e-01 7.52275407e-01 1.39059162e+00
2.72900820e-01 9.66208756e-01 7.92250037e-01 1.32207602e-01
5.86411543e-02 2.19815105e-01 -4.12060559e-01 5.38038671e-01
1.82185903e-01 -1.40694035e-02 -5.06758451e-01 -5.95183671e-02
9.23930585e-01 7.68117964e-01 -4.19125617e-01 -8.50185454e-01
-1.52453303e+00 6.54169619e-01 -2.09412947e-02 -5.70143126e-02
-4.99956101e-01 -7.86746740e-02 -8.37891176e-02 3.68999615e-02
6.51921809e-01 1.63340848e-02 -6.26959264e-01 2.47700512e-01
-5.88851511e-01 2.19575122e-01 5.04691005e-01 1.44226086e+00
4.88915473e-01 3.13874871e-01 -8.86234045e-02 2.37885416e-01
5.62727690e-01 7.10759699e-01 1.69807643e-01 -1.24139071e+00
5.87976336e-01 6.94659472e-01 3.65660518e-01 -9.23050821e-01
6.07178658e-02 -2.81308264e-01 -9.45214927e-01 1.73503488e-01
2.13504750e-02 2.16997311e-01 -8.77880573e-01 1.65242589e+00
3.82591307e-01 1.33130625e-01 3.42599601e-02 9.40697908e-01
6.82955801e-01 7.61331081e-01 -3.11618418e-01 -5.77657521e-01
1.29495943e+00 -1.09679890e+00 -8.07917893e-01 -2.50444680e-01
-1.62316397e-01 -7.60990441e-01 6.31708324e-01 7.11636305e-01
-1.36741090e+00 -1.00225019e+00 -6.90414310e-01 -4.39184576e-01
1.84774429e-01 2.50981838e-01 3.99935007e-01 7.27368146e-02
-8.93877089e-01 2.23907083e-01 -1.16131675e+00 -2.34058928e-02
3.65643770e-01 6.93018660e-02 -5.74275553e-01 -5.63656032e-01
-3.67849112e-01 6.70961678e-01 3.31909567e-01 1.82483077e-01
-1.50689137e+00 -4.95304704e-01 -9.25183713e-01 2.19242662e-01
6.78206444e-01 -1.30564487e+00 9.56929982e-01 -9.13564920e-01
-1.36456096e+00 7.23470509e-01 -7.25371420e-01 9.52382684e-02
5.70101857e-01 -3.56408060e-01 -3.30766141e-01 3.09351712e-01
1.53604478e-01 5.80154240e-01 1.27330148e+00 -1.74020338e+00
-6.26656353e-01 -6.80244207e-01 3.96084726e-01 5.73295474e-01
2.53512740e-01 -4.69965994e-01 -3.84074777e-01 -6.01849794e-01
7.82023668e-01 -8.57153773e-01 -1.16804503e-01 2.20222086e-01
-2.92381555e-01 -8.95238519e-02 9.37052071e-01 -8.49582195e-01
4.42083716e-01 -2.08868313e+00 9.11158681e-01 -5.77229083e-01
3.90587181e-01 -3.43581825e-01 8.02516416e-02 4.04257059e-01
-1.53259233e-01 -2.51832694e-01 1.87080652e-01 -6.84545815e-01
-1.80142328e-01 5.19524872e-01 -7.40527451e-01 4.14930850e-01
3.33179869e-02 6.46709621e-01 -9.23125982e-01 -7.15233982e-02
5.06830335e-01 6.70570195e-01 -8.14140975e-01 5.25662422e-01
-5.27639151e-01 9.87383425e-01 -4.15316433e-01 5.51248610e-01
7.13105857e-01 -5.13067842e-01 3.04054201e-01 -5.31071782e-01
1.09070331e-01 1.03744216e-01 -1.16501963e+00 2.03978324e+00
-3.88966799e-01 3.14820319e-01 -6.61781058e-02 -9.49767411e-01
5.29309094e-01 6.24995112e-01 3.31077784e-01 -4.10443582e-02
-1.13555655e-01 -3.35132390e-01 -1.77129224e-01 -8.63305926e-01
3.01970363e-01 -5.17958522e-01 3.31390768e-01 4.76381749e-01
2.48340935e-01 -1.46782096e-03 -8.80458578e-02 2.78382957e-01
3.73681992e-01 7.08960176e-01 3.85410368e-01 2.17884228e-01
4.19833243e-01 -4.64235902e-01 7.02366889e-01 3.43008965e-01
2.53051400e-01 9.61893916e-01 1.87551618e-01 -7.29132950e-01
-1.06294572e+00 -1.52233732e+00 3.48207057e-01 9.85660374e-01
3.25438023e-01 -1.63984731e-01 -1.78438380e-01 -5.44409752e-01
-2.88715482e-01 1.05978906e+00 -5.76879025e-01 -6.33592978e-02
-5.43133736e-01 -3.27989817e-01 -4.38547790e-01 6.84679568e-01
-5.49833737e-02 -8.86757910e-01 -5.10427535e-01 6.36561736e-02
-7.12926686e-01 -1.34400189e+00 -4.79618579e-01 -2.67543733e-01
-1.23187494e+00 -1.13904572e+00 -4.83666450e-01 -5.66149592e-01
1.23509324e+00 9.71805394e-01 9.96668339e-01 -2.86753565e-01
-5.25558591e-02 6.69793725e-01 -1.72459021e-01 -1.32382125e-01
-3.42144072e-01 -6.46615386e-01 2.72260189e-01 5.05774558e-01
2.27113478e-02 -8.75128210e-01 -5.08955598e-01 2.86975831e-01
-8.88128221e-01 7.35357642e-01 4.85667169e-01 7.68620193e-01
7.47034907e-01 -8.40114150e-03 6.04298748e-02 -8.72679710e-01
-3.93609762e-01 -4.97741401e-01 -4.89315629e-01 4.94144648e-01
-5.26337959e-02 6.23130463e-02 6.40669584e-01 -6.62883639e-01
-1.66381502e+00 1.10116743e-01 2.79833853e-01 -1.13424289e+00
-3.60655457e-01 1.14490524e-01 -7.24524021e-01 7.41842926e-01
9.50306877e-02 5.04980266e-01 -1.27360508e-01 -6.65009975e-01
6.88627958e-01 -1.63193554e-01 8.97517726e-02 -5.40109992e-01
7.79440403e-01 1.03885925e+00 -2.20936581e-01 -6.02529049e-01
-1.07805264e+00 -3.12696368e-01 -1.07964039e+00 -6.43149093e-02
1.06924355e+00 -1.41684830e+00 -6.82935417e-01 1.59970820e-01
-1.37937760e+00 3.28773081e-01 -3.47171336e-01 5.85440874e-01
-8.62766743e-01 5.56667387e-01 -4.20462728e-01 -8.14387619e-01
2.39558026e-01 -1.19042480e+00 1.33068645e+00 4.73054126e-02
7.56755173e-02 -1.04380453e+00 -3.26386541e-01 9.04788017e-01
-1.91906735e-01 1.51772901e-01 1.02162278e+00 -3.26665461e-01
-1.24368334e+00 -4.63031456e-02 7.06808418e-02 2.50539809e-01
3.86109382e-01 -1.65503044e-02 -1.10840201e+00 -5.40068388e-01
7.80164421e-01 -2.63888054e-02 7.52514720e-01 6.89689755e-01
1.11594880e+00 -5.12525678e-01 -3.95788431e-01 6.66816413e-01
1.18214929e+00 3.53943586e-01 2.60277957e-01 -5.10161877e-01
9.89384174e-01 7.58276343e-01 4.67486054e-01 5.94410837e-01
6.62797987e-01 4.55454409e-01 7.42118835e-01 4.27982539e-01
2.27139518e-02 -5.81374466e-01 4.44067091e-01 1.15339804e+00
-4.35393006e-01 -5.66135287e-01 -3.13027442e-01 4.49079156e-01
-1.82207739e+00 -1.34375751e+00 3.28286961e-02 2.03918815e+00
9.74735916e-02 -4.75802347e-02 -3.26234311e-01 -2.73984492e-01
7.57984698e-01 3.38770449e-01 -8.74386430e-01 3.06465536e-01
-7.22927565e-04 -5.93469918e-01 -3.13831627e-01 5.10339260e-01
-8.19860697e-01 6.85742497e-01 5.47035551e+00 2.07493529e-01
-6.47831917e-01 2.73843288e-01 4.53682691e-01 -4.93765295e-01
-6.90090656e-01 3.91762555e-01 -8.34789097e-01 3.47272977e-02
1.75389633e-01 -8.89085382e-02 5.80653906e-01 8.92983258e-01
1.34470165e-01 1.52549550e-01 -1.56472957e+00 1.17027760e+00
7.05211461e-01 -1.11213219e+00 5.76238930e-01 1.72942713e-01
9.38945293e-01 -3.54644567e-01 2.19363004e-01 2.43653089e-01
1.65935099e-01 -6.40892684e-01 1.14035594e+00 9.36802745e-01
6.32947743e-01 -3.34693044e-01 2.06492186e-01 9.28356111e-01
-1.13940489e+00 -2.86597162e-01 -7.68969834e-01 -3.22723418e-01
7.14077711e-01 2.99816906e-01 -6.44781828e-01 8.51720273e-01
4.72830236e-01 1.12421989e+00 -2.45478705e-01 5.41205347e-01
-3.79387558e-01 9.97780729e-03 1.34465367e-01 6.91211522e-01
-2.50549644e-01 -2.94885576e-01 7.83264577e-01 2.19455674e-01
5.54148078e-01 5.08158565e-01 2.50968188e-01 1.17438138e+00
2.47568905e-01 -4.31477249e-01 -6.46986425e-01 6.44841939e-02
-9.60445628e-02 1.15987456e+00 -5.24151266e-01 -4.95965332e-01
-6.06651068e-01 1.11125958e+00 1.87530413e-01 7.43099511e-01
-9.08991158e-01 7.89776266e-01 5.50119162e-01 1.58008412e-01
6.16141260e-01 -3.43483299e-01 2.95126606e-02 -2.03884816e+00
1.99432582e-01 -7.19951153e-01 2.20059410e-01 -1.39282095e+00
-1.39669085e+00 5.89103878e-01 3.37573320e-01 -1.62286973e+00
-2.41776288e-01 -5.67504883e-01 -2.73027301e-01 6.81998551e-01
-1.13121855e+00 -1.53312266e+00 -2.64720082e-01 6.69265807e-01
1.42621207e+00 -1.98639721e-01 7.00234354e-01 6.16814010e-02
-1.27885684e-01 -2.52984554e-01 -1.89672470e-01 -3.18011910e-01
5.17426252e-01 -1.03943431e+00 2.60941982e-01 8.23660553e-01
5.99066436e-01 6.48856878e-01 6.61290288e-01 -7.70194113e-01
-1.61364245e+00 -1.00392473e+00 6.81719720e-01 -9.25642371e-01
2.33211502e-01 -5.97482383e-01 -8.43088031e-01 1.32725000e+00
1.39297292e-01 1.43736705e-01 6.89955711e-01 -4.14763615e-02
-4.18873101e-01 9.60745886e-02 -6.62683010e-01 5.88105917e-01
1.18858075e+00 -5.72877645e-01 -7.92516232e-01 4.24347430e-01
9.30913746e-01 -4.52891886e-01 -4.91957843e-01 2.90068299e-01
7.70592749e-01 -1.15360820e+00 1.11484885e+00 -9.70730066e-01
5.19552946e-01 -4.01792288e-01 -5.81519604e-01 -1.11177993e+00
-6.11192763e-01 -1.71655431e-01 -5.60237408e-01 9.84518766e-01
-6.54438213e-02 -2.92075664e-01 5.18644273e-01 5.66778541e-01
-2.34072089e-01 -5.54239750e-01 -7.31966853e-01 -4.08231288e-01
-3.72764289e-01 -2.00702682e-01 4.50694621e-01 8.55813086e-01
-4.79517341e-01 4.34243888e-01 -8.54072332e-01 6.99537396e-01
8.23167622e-01 7.64213026e-01 7.90225029e-01 -1.25562024e+00
-6.97509706e-01 1.15327470e-01 -2.40297973e-01 -1.48514128e+00
1.45438641e-01 -5.95400333e-01 -2.05970462e-02 -1.63277984e+00
7.02762067e-01 2.94561267e-01 1.91687476e-02 4.81494796e-03
-2.22974822e-01 -1.69953972e-01 4.50667351e-01 4.60671693e-01
-6.46984696e-01 9.03336763e-01 1.78443730e+00 -1.55488133e-01
1.06504388e-01 3.73231977e-01 -6.11780286e-01 1.21204388e+00
1.99842408e-01 -4.19523209e-01 -1.00074506e+00 -9.45024192e-01
4.62544650e-01 8.02212834e-01 5.77179372e-01 -5.87482750e-01
4.31163758e-02 -5.61682343e-01 6.90607488e-01 -1.11079276e+00
8.62234950e-01 -1.06000507e+00 6.70196235e-01 1.59752443e-01
-2.34454468e-01 1.03880860e-01 -3.14787179e-01 1.26443088e+00
-2.03812972e-01 1.39952496e-01 4.23855901e-01 -6.28374994e-01
-4.84823346e-01 6.56155348e-01 -2.61595935e-01 -1.75116464e-01
9.89058912e-01 -3.43381196e-01 -1.43798798e-01 -5.48739552e-01
-1.46189857e+00 -1.92885604e-02 6.03044569e-01 8.34753335e-01
1.00531065e+00 -1.42580116e+00 -6.37680233e-01 5.95751107e-01
1.46175385e-01 4.22667176e-01 1.03760660e+00 6.39765322e-01
-1.81647226e-01 2.94667065e-01 -2.04112068e-01 -9.64577734e-01
-1.11666679e+00 1.16728675e+00 -1.45615134e-02 6.66060969e-02
-8.34201574e-01 8.52576792e-01 1.33451331e+00 -3.26059312e-01
-4.97165471e-02 -4.88251328e-01 -2.44020596e-01 -4.09083925e-02
5.40801287e-01 2.64128864e-01 -5.61275959e-01 -9.16144013e-01
7.29562342e-02 5.60003638e-01 -1.28484681e-01 -1.24218769e-01
1.45326555e+00 -7.06422210e-01 -6.38201460e-02 1.03960645e+00
9.15402472e-01 -1.41851097e-01 -1.82125342e+00 -2.79470652e-01
-7.24723935e-01 -8.49248230e-01 -4.17969853e-01 -4.88328159e-01
-1.06836998e+00 1.39516795e+00 1.11494370e-01 -1.88667238e-01
9.61462498e-01 1.16715930e-01 4.81088221e-01 2.76907206e-01
5.58103085e-01 -4.67797160e-01 4.66637194e-01 1.74546406e-01
1.28539753e+00 -1.17636287e+00 1.68360174e-01 -6.09484196e-01
-9.21734452e-01 1.14270592e+00 8.20435226e-01 -2.51709651e-02
6.69149160e-01 -2.16818124e-01 -3.34871024e-01 -2.29107738e-01
-1.43441486e+00 3.07422310e-01 5.20331740e-01 5.77386260e-01
3.36884409e-02 1.25126436e-01 5.88053584e-01 6.92506552e-01
1.14996560e-01 -4.17526215e-01 6.51379585e-01 5.85684061e-01
-3.22079152e-01 -7.39604235e-01 -3.30312759e-01 8.36720541e-02
-1.15531258e-01 1.26531377e-01 5.07569779e-03 3.92999679e-01
2.91886091e-01 7.02107489e-01 1.76086295e-02 -4.03488763e-02
9.34506506e-02 1.83286250e-01 8.96142840e-01 -1.09881794e+00
2.94451892e-01 6.56722128e-01 -4.00530607e-01 -4.94739801e-01
-6.69536889e-01 -8.71789515e-01 -7.75040090e-01 1.77397326e-01
-2.08907858e-01 -2.66563267e-01 2.08464831e-01 1.19779515e+00
3.45521599e-01 6.32419407e-01 6.14617705e-01 -1.18728638e+00
-5.50243318e-01 -7.04956830e-01 -6.76087856e-01 5.93890905e-01
3.97500157e-01 -8.44450712e-01 -4.66267228e-01 8.36071193e-01] | [8.825374603271484, -2.9253485202789307] |
207093f2-e452-45df-918b-bc6b1f93aa0c | detecting-object-states-vs-detecting-objects | 2112.08281 | null | https://arxiv.org/abs/2112.08281v2 | https://arxiv.org/pdf/2112.08281v2.pdf | Detecting Object States vs Detecting Objects: A New Dataset and a Quantitative Experimental Study | The detection of object states in images (State Detection - SD) is a problem of both theoretical and practical importance and it is tightly interwoven with other important computer vision problems, such as action recognition and affordance detection. It is also highly relevant to any entity that needs to reason and act in dynamic domains, such as robotic systems and intelligent agents. Despite its importance, up to now, the research on this problem has been limited. In this paper, we attempt a systematic study of the SD problem. First, we introduce the Object State Detection Dataset (OSDD), a new publicly available dataset consisting of more than 19,000 annotations for 18 object categories and 9 state classes. Second, using a standard deep learning framework used for Object Detection (OD), we conduct a number of appropriately designed experiments, towards an in-depth study of the behavior of the SD problem. This study enables the setup of a baseline on the performance of SD, as well as its relative performance in comparison to OD, in a variety of scenarios. Overall, the experimental outcomes confirm that SD is harder than OD and that tailored SD methods need to be developed for addressing effectively this significant problem. | ['Theodore Patkos', 'Dimitris Plexousakis', 'Antonis Argyros', 'Filippos Gouidis'] | 2021-12-15 | null | null | null | null | ['affordance-detection'] | ['computer-vision'] | [ 3.44066858e-01 1.39778838e-01 -1.88443437e-01 -2.38122255e-01
-1.91927224e-01 -5.53428531e-01 9.53631282e-01 7.37624019e-02
-6.00215375e-01 4.48933691e-01 3.72940190e-02 -1.30119264e-01
-1.07172348e-01 -2.13390812e-01 -5.37615359e-01 -6.17809892e-01
-3.55887979e-01 5.06911993e-01 9.53636944e-01 -7.35626891e-02
3.43298912e-01 7.87189245e-01 -1.86700165e+00 -9.77123827e-02
3.49825561e-01 1.06228292e+00 1.74323395e-01 7.42742062e-01
3.87735277e-01 9.58642662e-01 -4.92815971e-01 -1.05639055e-01
2.32166499e-01 -1.90713733e-01 -1.11961830e+00 3.04067582e-01
4.83059406e-01 -4.49456185e-01 -3.09349567e-01 1.06174517e+00
4.08287525e-01 3.24369907e-01 4.80386883e-01 -1.55019915e+00
-8.96742865e-02 1.29729077e-01 -1.81123301e-01 5.50089896e-01
4.71551657e-01 5.61487675e-01 9.14897621e-01 -3.81127298e-01
7.94951916e-01 1.47821760e+00 1.71689555e-01 6.05386317e-01
-1.16953468e+00 -2.60213733e-01 3.50355059e-01 3.69513899e-01
-7.80078650e-01 -6.25589907e-01 4.87094551e-01 -5.46146214e-01
1.07843149e+00 1.17414504e-01 5.59515715e-01 1.33101881e+00
-2.27272466e-01 1.19361973e+00 1.25422263e+00 -5.19608736e-01
5.60298502e-01 2.00193003e-01 3.53904665e-01 5.21097600e-01
4.26409841e-01 3.24886978e-01 -3.05576086e-01 6.19548336e-02
6.31753325e-01 -2.49275342e-01 -3.93969454e-02 -1.00072432e+00
-1.46234107e+00 6.14882767e-01 2.67796010e-01 2.60242492e-01
-1.70823008e-01 2.79179037e-01 4.87620413e-01 2.67620176e-01
-1.25153251e-02 5.19673526e-01 -4.75147933e-01 -4.42298710e-01
-4.53024358e-01 5.51238477e-01 9.53288555e-01 6.95216119e-01
3.46413672e-01 -2.46152177e-01 -5.97402975e-02 4.30064857e-01
4.37314510e-01 2.41162464e-01 3.05686384e-01 -1.07849240e+00
2.56581336e-01 8.31233501e-01 4.32180375e-01 -7.45492876e-01
-6.39366090e-01 -5.32658510e-02 -4.09927636e-01 6.53320432e-01
6.62674248e-01 5.00890985e-03 -8.46955717e-01 1.81884110e+00
6.70940042e-01 1.35498390e-01 1.37329742e-01 1.13493657e+00
5.96928835e-01 3.37598681e-01 8.98227394e-02 -4.68052551e-02
1.37877524e+00 -8.31986070e-01 -5.70592701e-01 -5.90126634e-01
5.63332200e-01 -4.58952099e-01 9.49642897e-01 3.88854563e-01
-8.32887530e-01 -4.25014138e-01 -1.06911516e+00 2.30723303e-02
-5.34741461e-01 8.57915655e-02 8.83995414e-01 4.58810270e-01
-8.44453514e-01 4.79998499e-01 -1.41891623e+00 -1.06203520e+00
4.06683952e-01 4.40704018e-01 -3.97876710e-01 9.65418145e-02
-9.16186690e-01 1.48490739e+00 6.56583786e-01 -5.76588959e-02
-1.10863876e+00 3.10807489e-02 -9.49350178e-01 -2.92216569e-01
8.99994135e-01 -2.80824572e-01 1.50583243e+00 -6.75200105e-01
-1.38541055e+00 9.14513409e-01 1.13355741e-01 -6.12767875e-01
6.28173470e-01 -2.87888825e-01 -4.06207740e-01 1.47236794e-01
3.56952250e-02 6.31152153e-01 7.09046185e-01 -1.04123998e+00
-8.52101147e-01 -6.25957310e-01 5.10243237e-01 1.59464762e-01
-2.52213906e-02 4.08569843e-01 -2.98827618e-01 -1.94238082e-01
9.00884643e-02 -1.33536506e+00 -3.57849568e-01 2.04711676e-01
-2.71147907e-01 -4.55016315e-01 9.88780022e-01 -1.58172280e-01
9.19161797e-01 -2.24880505e+00 2.64912337e-01 -2.21380383e-01
4.91099544e-02 5.47389030e-01 -2.69330367e-02 4.03225660e-01
-3.94831300e-02 -4.15185422e-01 -1.74366891e-01 -1.94975734e-01
3.07235479e-01 3.65769774e-01 -2.69414872e-01 7.36645103e-01
5.80022097e-01 8.05828989e-01 -9.62896168e-01 -3.91809225e-01
4.20952737e-01 1.36918062e-02 -3.55547041e-01 8.43076333e-02
-3.80937815e-01 3.68817568e-01 -6.19499505e-01 6.36810958e-01
1.91438213e-01 -2.60000855e-01 1.16891474e-01 4.00784425e-03
-3.22463274e-01 4.30185586e-01 -1.59623814e+00 1.44749212e+00
6.46837726e-02 8.04704547e-01 -5.79841286e-02 -1.18437743e+00
5.68610072e-01 8.34046677e-02 5.29148042e-01 -4.74362165e-01
2.21128896e-01 1.28288552e-01 3.61388773e-01 -7.07806766e-01
5.55137932e-01 1.83369339e-01 -2.40114287e-01 5.04364848e-01
8.72693732e-02 1.43813284e-03 5.56877375e-01 8.76912326e-02
1.34763217e+00 2.89844602e-01 5.28966725e-01 -1.96522653e-01
4.19678181e-01 2.76253313e-01 2.23716691e-01 6.75490737e-01
-9.10262704e-01 2.07157388e-01 6.75317168e-01 -5.77407300e-01
-7.37592280e-01 -8.22226107e-01 -1.76836088e-01 9.65255857e-01
5.37836015e-01 -7.55693465e-02 -6.30718172e-01 -7.21723080e-01
8.58263150e-02 5.12974858e-01 -6.47018075e-01 -3.77529919e-01
-4.88182843e-01 -7.14666665e-01 3.38025302e-01 5.82446337e-01
5.49275756e-01 -1.45095301e+00 -1.40433693e+00 1.42840356e-01
2.37114727e-01 -1.55660546e+00 1.67861715e-01 4.47417647e-01
-7.73861647e-01 -1.28972363e+00 -3.14054400e-01 -5.63087404e-01
2.43949637e-01 1.84523985e-01 1.03913224e+00 -1.92243278e-01
-3.58772457e-01 6.25736892e-01 -4.96487349e-01 -7.51543581e-01
-5.88469148e-01 1.74106210e-01 2.23700240e-01 -2.85678599e-02
5.34202933e-01 -1.94025338e-01 -4.87356991e-01 5.12142539e-01
-9.45395112e-01 -1.87443122e-01 7.74446130e-01 4.30223912e-01
1.91411953e-02 -1.30106509e-01 2.88025856e-01 -4.08474237e-01
4.54993069e-01 -1.37233853e-01 -9.56735730e-01 8.13502222e-02
-5.15303910e-01 1.70420572e-01 -8.51521790e-02 -5.49065232e-01
-7.49150395e-01 4.43439841e-01 7.26256296e-02 -1.57226562e-01
-6.48188293e-01 2.67186224e-01 -1.62075162e-01 3.41864750e-02
7.70802140e-01 1.25123456e-01 2.52067626e-01 -4.50393021e-01
1.61080092e-01 6.64477408e-01 4.91604894e-01 -2.83230573e-01
4.88264740e-01 7.21895576e-01 1.56098112e-01 -9.21705365e-01
-9.08339918e-01 -6.95882201e-01 -7.66111016e-01 -2.28900656e-01
9.34695721e-01 -6.30477369e-01 -7.58814752e-01 6.69200718e-01
-9.55386758e-01 -6.60233974e-01 -2.99396873e-01 5.01451671e-01
-7.34102190e-01 4.46539640e-01 -2.78833479e-01 -9.09729183e-01
2.57400662e-01 -1.30470514e+00 1.29431486e+00 1.39672533e-01
-3.02308381e-01 -7.67445564e-01 1.50194332e-01 2.84752071e-01
1.08086631e-01 3.25778604e-01 5.09943008e-01 -8.32150400e-01
-7.03208327e-01 -2.82812715e-01 -1.76921144e-01 3.76987755e-01
1.14256693e-02 -5.62520884e-02 -9.88884926e-01 -3.46923321e-01
-1.27888888e-01 -5.28571248e-01 8.07919562e-01 1.40762269e-01
5.91627896e-01 1.31610245e-01 -4.31500018e-01 -7.37767518e-02
1.07758749e+00 1.90623760e-01 6.22951865e-01 6.41335428e-01
3.91294956e-01 6.77755058e-01 9.04328406e-01 4.75309104e-01
3.60859782e-01 1.05346584e+00 8.90533209e-01 1.21390767e-01
-1.76054344e-01 1.77488044e-01 5.09152889e-01 -5.04011870e-04
-1.27316386e-01 -1.07815847e-01 -9.54203010e-01 4.90378827e-01
-2.04985976e+00 -1.04193437e+00 -1.27375588e-01 2.17210031e+00
5.01380563e-01 3.59235317e-01 4.11282003e-01 3.53415012e-01
5.14326692e-01 2.20171332e-01 -8.13952148e-01 -2.37871543e-01
2.03935757e-01 -1.92662328e-01 3.73289704e-01 7.82811120e-02
-1.56495786e+00 9.46063757e-01 6.73298597e+00 2.34279796e-01
-1.11317813e+00 -5.37481084e-02 6.86891675e-02 6.98398799e-02
7.30361104e-01 3.55227254e-02 -9.86871541e-01 3.73373717e-01
8.86853755e-01 2.10755363e-01 4.20314103e-01 1.01201522e+00
1.40102714e-01 -6.50716841e-01 -1.36763144e+00 7.60182679e-01
6.01554662e-02 -7.38797724e-01 -4.92098302e-01 6.48607165e-02
5.77161491e-01 2.77771652e-01 -1.14320919e-01 4.75521803e-01
5.17607450e-01 -4.90004718e-01 1.01298738e+00 1.81433350e-01
2.28629634e-01 -2.56712049e-01 6.10844195e-01 6.36488199e-01
-8.19670498e-01 -3.77019763e-01 -2.07911029e-01 -4.07451510e-01
1.08258858e-01 2.33796000e-01 -8.36529911e-01 1.01490751e-01
7.61886239e-01 7.57420361e-01 -7.24717557e-01 1.01796436e+00
-2.63293356e-01 4.00235832e-01 -3.28561544e-01 -1.45346090e-01
3.08358699e-01 2.17383966e-01 7.36432731e-01 9.70198274e-01
-1.51566669e-01 -6.60996810e-02 3.42656374e-01 6.81776583e-01
3.66048485e-01 -4.87644792e-01 -6.10991716e-01 -2.66788930e-01
9.52063948e-02 1.18699372e+00 -9.70807672e-01 -4.89040405e-01
-3.93395364e-01 7.81009734e-01 3.29609662e-01 7.35931844e-02
-7.80916691e-01 1.82985198e-02 9.84430671e-01 5.20173013e-02
4.68154192e-01 -5.40767848e-01 1.46814674e-01 -1.15140665e+00
8.05945992e-02 -8.69143486e-01 4.05955136e-01 -6.88303173e-01
-8.79961789e-01 2.79362202e-01 3.77708942e-01 -1.26193500e+00
-4.18525130e-01 -9.67097819e-01 -3.79987061e-01 3.53333622e-01
-1.70193791e+00 -9.15869474e-01 -4.36391622e-01 4.95908648e-01
5.18685877e-01 9.95061174e-02 6.32113695e-01 1.31334335e-01
-8.14242125e-01 2.07802989e-02 -2.07566872e-01 1.45127237e-01
5.42032421e-01 -1.24280977e+00 4.41335291e-01 1.02063859e+00
1.68910474e-01 4.07770187e-01 9.26168680e-01 -5.10368526e-01
-1.63874269e+00 -9.59999979e-01 5.50254762e-01 -9.07616019e-01
8.66208017e-01 -3.88827235e-01 -6.40263915e-01 7.77349710e-01
-1.18928544e-01 3.97724867e-01 -2.46302839e-02 -4.89806756e-02
-3.18373218e-02 2.16026440e-01 -9.10883844e-01 6.34163737e-01
1.31113088e+00 -3.16071033e-01 -7.79451132e-01 3.74761790e-01
5.38188577e-01 -5.17271280e-01 -5.35148501e-01 5.23910761e-01
5.22881329e-01 -1.18596947e+00 8.09764087e-01 -7.89525807e-01
1.91708475e-01 -4.84117448e-01 -9.81748775e-02 -9.58472013e-01
-2.92874515e-01 -4.61759090e-01 -4.98482853e-01 9.99652565e-01
8.30070674e-02 -6.65877640e-01 8.33884299e-01 6.71429992e-01
-1.79597765e-01 -5.48921168e-01 -8.16108644e-01 -1.09719694e+00
-4.73398507e-01 -5.19432247e-01 2.38782018e-01 7.04645574e-01
-1.80134252e-01 4.10599858e-01 5.14928736e-02 3.20790738e-01
3.26265931e-01 2.08730236e-01 1.08752143e+00 -1.42563891e+00
-1.39888152e-01 -5.04592180e-01 -1.01132607e+00 -9.38693285e-01
4.97106314e-02 -5.74395597e-01 2.50656605e-01 -1.59352136e+00
4.51000370e-02 -2.65043080e-01 -3.59020531e-01 5.58047891e-01
-2.11145170e-02 1.22448942e-02 2.29354009e-01 3.07628036e-01
-1.11238599e+00 3.74280065e-01 1.23961294e+00 -5.00745997e-02
-1.21476293e-01 1.22897446e-01 -3.91733766e-01 8.27699125e-01
5.67996025e-01 -3.76259178e-01 -2.64043957e-01 -1.41796485e-01
6.64947480e-02 -2.48845533e-01 7.96122551e-01 -1.15028083e+00
1.13418743e-01 -2.10330248e-01 -9.82992500e-02 -4.41553891e-01
4.66484100e-01 -9.21689093e-01 -2.90168524e-01 8.00163567e-01
-2.85050601e-01 -1.58064678e-01 1.63902283e-01 6.74289584e-01
-1.00546189e-01 -2.38065243e-01 8.51452351e-01 -2.16045678e-01
-1.49149942e+00 2.21639164e-02 -3.96753788e-01 2.37222891e-02
1.49009824e+00 -3.83092523e-01 -2.04072431e-01 -7.25011900e-02
-7.46760428e-01 1.55882955e-01 4.77532655e-01 6.83408678e-01
1.10693440e-01 -1.01389241e+00 -3.72555763e-01 1.09950043e-01
4.77184325e-01 -7.63380975e-02 -1.53207436e-01 1.08932900e+00
-2.68890738e-01 5.48421800e-01 -2.67693073e-01 -8.94023478e-01
-1.11087191e+00 7.39818215e-01 1.63977548e-01 -2.27990165e-01
-6.12407625e-01 5.42206466e-01 3.52751510e-03 -2.43553892e-01
4.62704271e-01 -6.85004115e-01 -4.49328274e-01 8.39819387e-02
4.72722471e-01 5.06049991e-01 8.68060365e-02 -6.84100688e-01
-4.69698101e-01 2.07731813e-01 1.16163976e-01 -1.07871201e-02
1.34934473e+00 -9.59102809e-02 9.38689243e-03 5.69013536e-01
8.23605716e-01 -6.77921474e-01 -1.57596636e+00 -1.19150266e-01
5.33236027e-01 -4.00604606e-01 -3.13711353e-02 -7.37397909e-01
-6.63866460e-01 6.76930249e-01 8.68643761e-01 5.96149623e-01
8.20836365e-01 3.70127320e-01 3.15026671e-01 6.12491071e-01
4.90569562e-01 -9.97254848e-01 3.61684978e-01 6.40948653e-01
7.69759774e-01 -1.70832992e+00 7.00227991e-02 -3.10658514e-01
-5.75471878e-01 7.40314007e-01 6.22029781e-01 -2.14655071e-01
2.86580503e-01 2.68959224e-01 -1.65560454e-01 -4.17533785e-01
-8.80089521e-01 -6.07962489e-01 1.31682605e-01 5.73175609e-01
-2.72609480e-02 -1.52158394e-01 9.84825846e-03 1.78727284e-01
1.23858005e-01 2.43580833e-01 5.13232410e-01 1.39910913e+00
-5.40453672e-01 -9.70881760e-01 -3.37576866e-01 3.58756602e-01
-2.48033762e-01 5.08540809e-01 -5.46407759e-01 9.57425535e-01
2.58407760e-02 8.62709343e-01 -1.54038928e-02 -8.10495764e-02
6.02688074e-01 -9.02803093e-02 5.84659815e-01 -7.44535089e-01
-1.89559907e-01 -3.79632831e-01 1.97310582e-01 -9.06222701e-01
-8.90741944e-01 -1.17804956e+00 -1.09668303e+00 2.76623726e-01
-4.83805060e-01 -4.81350034e-01 9.16069329e-01 1.16723955e+00
2.80107379e-01 3.95331562e-01 2.17241079e-01 -1.06641340e+00
-8.60399246e-01 -9.33089852e-01 -5.16093254e-01 5.28434813e-01
4.26623076e-01 -1.22311914e+00 -3.48224550e-01 3.57983150e-02] | [8.24169921875, 0.09785018861293793] |
ec761a36-d054-4c68-9acc-34c4ee29bd6f | form-follows-function-a-different-approach-to | 2306.03337 | null | https://arxiv.org/abs/2306.03337v1 | https://arxiv.org/pdf/2306.03337v1.pdf | Form Follows Function: A Different Approach to Neuron Connectivity | It may be possible to discover much of the organization of synaptic connections in nervous systems by designing simple logic circuits that can perform a single, biologically advantageous function. This method has led to neuronal networks that can generate neural correlates of phenomena central to color vision, olfaction, short-term memory, and brain waves. One of the network designs is a family of general information processors that exhibit major features of cerebral cortex physiology and anatomy. A similar logic circuit approach applied to two primitive ganglia that have been studied extensively led to discoveries of how the ganglia can produce lobster peristaltic action and lamprey locomotion. For each network design, all neurons, connections, and types of connections are shown explicitly. The neurons' operation depends only on explicitly stated, minimal properties of excitement and inhibition. This operation is dynamic in the sense that the level of neuron activity is the only cellular change, making the networks' operation consistent with the speed of most brain functions. Conclusions that the networks can generate neural correlates of known phenomena are not claims; they are theorems that follow from the models' explicit architectures and minimal neuron capabilities. The logic circuit designs can be implemented with electronic components. A few of the designs are apparently new to engineering, filling gaps and providing improvements in well-known families of logic circuits. A novel transformation can convert certain electronic logic circuit designs to neuronal network designs, and vice versa. | ['Lane Yoder'] | 2023-06-06 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 1.15483120e-01 1.98817015e-01 8.39115456e-02 6.81134313e-02
9.67594683e-01 -7.65988231e-01 8.07863533e-01 -2.32800528e-01
-3.70452732e-01 9.98420715e-01 -1.04045413e-01 -3.91793281e-01
-2.38047391e-01 -9.53205526e-01 -6.69870555e-01 -8.78669977e-01
-4.43751812e-01 -4.98779900e-02 5.37258148e-01 -7.13653803e-01
5.10565817e-01 7.98774004e-01 -1.73824096e+00 3.45304042e-01
2.91338742e-01 9.63761330e-01 2.61941612e-01 7.93856382e-01
-2.01075315e-01 9.64508176e-01 -3.16123873e-01 -2.92615645e-04
1.35518223e-01 -9.08388734e-01 -5.35081089e-01 -3.71850282e-01
-2.81544358e-01 1.50633350e-01 -2.58779347e-01 6.93121552e-01
1.16624631e-01 -4.11549509e-01 6.78916872e-01 -1.06704092e+00
-7.97499120e-01 5.10390639e-01 1.58965424e-01 1.13495000e-01
5.58692217e-02 2.77906299e-01 7.41440356e-01 -2.78971791e-01
6.09871686e-01 1.14654660e+00 7.00539231e-01 1.01132667e+00
-1.48597932e+00 -4.76411760e-01 -1.21713929e-01 -2.11857762e-02
-1.03009129e+00 -5.56565344e-01 1.66388616e-01 -3.39124680e-01
1.15104210e+00 1.60574138e-01 1.50042593e+00 4.54322457e-01
9.70901847e-01 1.55440599e-01 1.16254973e+00 -2.96671689e-01
4.36262339e-01 1.96702685e-02 -1.26010418e-01 9.10215199e-01
8.05607736e-01 6.77332729e-02 -6.29504561e-01 1.28125429e-01
1.22080195e+00 -1.59905076e-01 -5.17635107e-01 5.97111285e-02
-1.18413854e+00 2.58402109e-01 3.20313096e-01 6.99916005e-01
2.45809127e-02 9.32884276e-01 1.35530278e-01 4.47183400e-01
-5.19219100e-01 9.42876756e-01 -4.30049598e-01 8.90359879e-02
-3.06844383e-01 1.58355936e-01 1.16740811e+00 5.51502824e-01
6.25653327e-01 2.26286501e-01 4.18246120e-01 3.19423884e-01
2.46183440e-01 5.07619798e-01 5.58775961e-01 -1.24235880e+00
-5.13896942e-01 7.40330637e-01 -4.18369144e-01 -7.13411629e-01
-6.39126897e-01 -2.59616882e-01 -6.90765381e-01 7.34198570e-01
5.61613262e-01 -1.78059936e-01 -5.99035025e-01 1.73058951e+00
-4.21254247e-01 -2.01564878e-01 2.67681152e-01 3.53860676e-01
5.58450103e-01 5.95610678e-01 -2.97943562e-01 -2.21580938e-01
1.22994995e+00 -1.09180473e-01 -6.06720924e-01 5.76140508e-02
4.87084717e-01 -2.91781694e-01 6.15342855e-01 4.49131072e-01
-1.11696827e+00 -7.21541196e-02 -1.27600169e+00 3.74831557e-01
-6.83096170e-01 -4.63408560e-01 1.20406258e+00 7.09854662e-01
-1.43604100e+00 7.09415674e-01 -6.50803983e-01 -5.16994059e-01
4.10607725e-01 6.53541148e-01 -3.01706940e-01 5.10104954e-01
-8.92767966e-01 1.16302776e+00 3.36781502e-01 5.76516800e-03
-7.05515683e-01 -3.31722349e-01 -1.40592575e-01 1.84382096e-01
-2.98265517e-01 -1.01202488e+00 1.01712370e+00 -1.12851334e+00
-1.59691179e+00 7.99713552e-01 -1.00858152e-01 -5.43659151e-01
-2.42655218e-01 5.76932967e-01 -7.61817172e-02 3.83150339e-01
-3.06725353e-01 1.04817152e+00 1.12455100e-01 -1.16402674e+00
-3.40753943e-01 -1.00520395e-01 -9.22998413e-02 -1.31892502e-01
-2.37637803e-01 -3.12460124e-01 1.39173701e-01 -1.88666776e-01
3.37255746e-01 -7.14604199e-01 -2.37389371e-01 5.14481962e-01
-1.12664420e-02 5.91172390e-02 4.92183357e-01 2.63123989e-01
6.30135894e-01 -1.93169534e+00 -2.20728945e-02 3.43759179e-01
2.29520082e-01 -1.83201253e-01 -9.17224139e-02 5.66474617e-01
6.54444993e-02 3.43068331e-01 -2.28611618e-01 5.79433262e-01
-2.99272269e-01 4.13657010e-01 -3.05008441e-01 3.43339503e-01
1.73235357e-01 1.02121818e+00 -7.02238560e-01 -1.85967907e-01
-1.71922177e-01 5.10734200e-01 -5.94566762e-01 -3.67543429e-01
-2.39467666e-01 1.70368612e-01 -2.69360274e-01 5.34735978e-01
2.51942202e-02 -6.18462712e-02 2.25047827e-01 9.03077200e-02
-5.17967820e-01 2.68066406e-01 -6.27424777e-01 1.20440829e+00
-7.16132927e-04 1.12393904e+00 9.35871601e-02 -9.84600782e-01
8.41075957e-01 5.44237554e-01 2.93131441e-01 -7.61698544e-01
3.42171520e-01 4.77600634e-01 8.57205927e-01 -4.83571470e-01
-2.56247520e-01 -5.53106487e-01 3.26455295e-01 5.44611394e-01
6.34579435e-02 -6.95087194e-01 3.74341071e-01 1.63835630e-01
1.34827423e+00 -8.24288055e-02 9.53426808e-02 -6.53682232e-01
2.13965550e-02 2.90245339e-02 4.70690161e-01 6.23510838e-01
1.25931740e-01 1.95826381e-01 6.73413694e-01 -4.11762565e-01
-1.03127468e+00 -1.10955000e+00 -2.60996550e-01 5.24828851e-01
2.69874424e-01 -6.44584000e-02 -6.82850480e-01 6.67600572e-01
-1.97002664e-01 4.16377366e-01 -6.63071096e-01 -2.67799348e-01
-2.27887154e-01 -8.51193428e-01 1.03526843e+00 1.83144927e-01
5.71590185e-01 -1.34590459e+00 -1.31256282e+00 3.99329454e-01
5.16358614e-01 -8.05355191e-01 4.20367330e-01 8.44959736e-01
-1.15644121e+00 -1.08930337e+00 -2.00577408e-01 -8.98195148e-01
8.12666833e-01 1.09366283e-01 6.66750669e-01 5.14958858e-01
-6.14225209e-01 3.01772743e-01 8.04903731e-02 -5.14034450e-01
-2.39395857e-01 -4.21217740e-01 3.09109390e-01 -5.04294217e-01
1.01749972e-02 -1.20338559e+00 -4.19219315e-01 -8.16932414e-03
-8.81081223e-01 2.36839354e-01 6.23193920e-01 6.94989383e-01
4.44825530e-01 8.80559608e-02 8.02651346e-01 -5.27976394e-01
6.37301564e-01 -2.35405490e-01 -2.46011719e-01 1.03580534e-01
-4.22168344e-01 3.64414275e-01 7.08212078e-01 -2.21590564e-01
-6.84983969e-01 2.33661793e-02 6.39425218e-03 5.23256719e-01
-1.14748046e-01 2.43097663e-01 -2.19565239e-02 -4.35292184e-01
6.94064260e-01 5.50911546e-01 3.14868838e-01 1.45601943e-01
2.29781747e-01 8.32451954e-02 5.84682167e-01 -4.74105507e-01
4.09761995e-01 8.23074758e-01 5.48573673e-01 -1.09761214e+00
9.88629833e-02 3.06663036e-01 -3.52950484e-01 -2.90698081e-01
7.14783490e-01 -4.42474812e-01 -1.20668030e+00 5.49652100e-01
-1.34166169e+00 -5.10099947e-01 -4.21104521e-01 5.34475267e-01
-8.59217227e-01 -2.64213890e-01 -7.55971849e-01 -8.57479215e-01
5.14113195e-02 -7.47010708e-01 7.32285455e-02 4.70645905e-01
-2.43174121e-01 -1.07657242e+00 2.10146263e-01 -6.47459030e-01
6.79922044e-01 2.19314873e-01 1.35871494e+00 -2.70123512e-01
-8.97292495e-01 8.34254455e-03 1.18774856e-02 1.25783950e-01
1.90644652e-01 4.31921750e-01 -8.80397201e-01 2.52405852e-01
7.58975744e-02 -2.57456124e-01 1.07352531e+00 4.58103657e-01
7.47922063e-01 -3.32956553e-01 -6.10495269e-01 6.46448553e-01
1.55023718e+00 9.44008887e-01 1.03099382e+00 2.77488112e-01
-1.08314436e-02 7.62758911e-01 -7.25277603e-01 -6.63350970e-02
-1.30619988e-01 1.64376516e-02 4.66551304e-01 -6.56198850e-03
-1.21366277e-01 1.25754237e-01 4.58444238e-01 8.75346959e-01
-6.09504163e-01 -4.88987230e-02 -6.20784104e-01 3.99239987e-01
-1.45397377e+00 -1.42332256e+00 -2.72212029e-01 1.84346843e+00
9.73135769e-01 5.35690427e-01 -1.86031088e-01 2.49359787e-01
5.90846241e-01 -5.45309186e-01 -6.69929206e-01 -7.58230686e-01
-7.04825997e-01 4.51415479e-01 7.18880355e-01 3.61814529e-01
-1.82687178e-01 6.36666238e-01 8.51855564e+00 3.07995826e-01
-1.08329606e+00 -3.17889005e-01 3.59381616e-01 -2.70128131e-01
-7.24487126e-01 1.24720365e-01 -6.15724385e-01 3.06562960e-01
1.02627373e+00 -1.38190240e-01 8.66514623e-01 3.04216385e-01
2.29620799e-01 -4.42362934e-01 -1.30846047e+00 8.05213749e-01
-2.14750528e-01 -1.91109347e+00 1.92654371e-01 1.07425638e-01
5.98110020e-01 -1.98473647e-01 -5.22460938e-02 -1.76139310e-01
2.39972234e-01 -1.26925099e+00 6.82833195e-01 8.10772121e-01
6.02683067e-01 -4.80388135e-01 2.24885866e-01 1.83340564e-01
-9.89650726e-01 -3.76484215e-01 -3.74067694e-01 -7.17511594e-01
2.08094008e-02 5.03016174e-01 -3.55459034e-01 -5.33285737e-01
5.03705382e-01 4.58648711e-01 -3.79045635e-01 1.30957127e+00
-1.04297757e-01 4.32350188e-01 -4.94870216e-01 -7.52352834e-01
1.07432537e-01 -6.44937158e-02 4.76290077e-01 1.09258068e+00
2.08292663e-01 4.64266211e-01 -9.92192090e-01 1.30225253e+00
1.38582379e-01 -3.15558702e-01 -1.21332157e+00 -3.83457571e-01
6.55340374e-01 1.13182962e+00 -1.41588080e+00 -2.64903724e-01
-2.04339981e-01 3.93890589e-01 -3.27894062e-01 3.32795203e-01
-4.29992497e-01 -7.88100839e-01 8.59114587e-01 2.92082578e-01
1.47844151e-01 -3.74293238e-01 -7.25919187e-01 -7.92150378e-01
-4.81324345e-01 -4.42675561e-01 -5.97724259e-01 -6.93948388e-01
-9.99091923e-01 3.36523831e-01 -3.88124824e-01 -6.78300619e-01
1.14247855e-02 -1.03896451e+00 -9.60305393e-01 4.12234873e-01
-9.02949393e-01 -7.02692270e-01 1.02772661e-01 5.50518215e-01
3.12040597e-02 -3.18523079e-01 8.87059510e-01 -2.04695448e-01
-1.80415243e-01 1.20430648e-01 2.09285393e-01 1.92703325e-02
2.83473507e-02 -1.00178695e+00 2.01045185e-01 4.52872306e-01
9.42878872e-02 7.93917537e-01 6.32305562e-01 -2.33620405e-01
-1.48685730e+00 -5.13794303e-01 7.30199993e-01 -1.29090771e-01
5.67057550e-01 -4.53078568e-01 -5.70504725e-01 4.11400914e-01
3.30632329e-01 -4.70597565e-01 5.99232972e-01 -3.87944669e-01
-2.29221076e-01 -2.59218127e-01 -9.56020236e-01 9.21187401e-01
1.17814565e+00 -3.50317448e-01 -6.80866599e-01 2.01986711e-02
5.31604588e-01 4.42424446e-01 -2.53075331e-01 4.31142598e-02
1.27844286e+00 -1.01145577e+00 6.74111664e-01 -4.83671248e-01
4.12499875e-01 -5.61937749e-01 1.17892800e-02 -1.19481897e+00
-3.04361910e-01 -5.72888255e-01 3.09816748e-01 7.62713253e-01
7.36999989e-01 -1.29618919e+00 3.19844514e-01 5.99926174e-01
-3.30041826e-01 -7.45331883e-01 -8.15403640e-01 -9.23027456e-01
2.03871071e-01 3.32048759e-02 2.12827608e-01 6.32556438e-01
7.69486070e-01 3.52214247e-01 4.86738414e-01 -6.25051141e-01
4.80374575e-01 -1.85697556e-01 1.87119663e-01 -1.34600949e+00
-3.60022098e-01 -1.06968331e+00 -8.94561172e-01 -7.30078876e-01
-2.35056773e-01 -9.72327650e-01 8.04502591e-02 -1.67845035e+00
9.37546790e-02 -2.68707484e-01 -2.53575325e-01 6.36361718e-01
7.01169789e-01 3.13047349e-01 6.04078099e-02 1.52603179e-01
-2.26992425e-02 1.73934735e-02 1.22791922e+00 1.78139791e-01
-2.00592294e-01 -4.04290855e-01 -9.69857037e-01 9.39571917e-01
1.11446130e+00 -4.15138930e-01 -5.94964981e-01 -3.48176867e-01
8.14599037e-01 -2.71753013e-01 4.33928072e-01 -1.41141546e+00
5.78840137e-01 -1.41302049e-01 6.97105348e-01 4.10729162e-02
2.19451651e-01 -6.32975757e-01 3.95277947e-01 1.21089125e+00
-4.47748572e-01 1.10544920e-01 2.45334223e-01 3.48955512e-01
1.27048373e-01 -9.33509395e-02 1.03111744e+00 -6.47313774e-01
-6.67533636e-01 -3.20574045e-01 -1.46681988e+00 -1.68290019e-01
1.11033297e+00 -8.94470870e-01 -8.69587243e-01 -2.07983270e-01
-5.66735566e-01 -1.45466089e-01 6.18729293e-01 -2.60676831e-01
7.21870005e-01 -1.04249072e+00 -2.21496850e-01 1.36761874e-01
-4.32860911e-01 -4.85974610e-01 -3.94369543e-01 7.04327226e-01
-1.07928777e+00 6.91483140e-01 -1.15410650e+00 -3.49886507e-01
-5.60997486e-01 2.24656731e-01 7.84734726e-01 5.07365048e-01
-3.12398612e-01 9.91241336e-01 3.85614753e-01 1.70222566e-01
-1.09306954e-01 -5.00928462e-01 -1.90392449e-01 -2.50663280e-01
5.57988226e-01 1.29349083e-01 -5.54866612e-01 5.27253821e-02
-4.61650997e-01 6.48397923e-01 4.59363490e-01 -3.35889965e-01
1.36543024e+00 7.33329728e-02 -8.00485373e-01 7.82333314e-01
5.80440640e-01 -2.44328082e-01 -9.06821072e-01 7.13979602e-01
-3.04433793e-01 3.52026552e-01 -3.97901416e-01 -8.22741866e-01
-9.55241978e-01 1.03031743e+00 1.59758657e-01 3.31670523e-01
1.11795616e+00 -9.67126936e-02 4.24300373e-01 9.62662339e-01
6.50638759e-01 -8.88737857e-01 1.18717782e-01 5.73766172e-01
6.48803473e-01 -1.83027610e-01 -1.13462701e-01 -1.19541157e-02
4.35221847e-03 1.52749383e+00 6.30005300e-01 -5.08477390e-01
8.10606360e-01 8.98342609e-01 -3.63485664e-01 -3.34387720e-01
-1.51824009e+00 -2.10704878e-01 -4.33783442e-01 6.97954893e-01
6.25346065e-01 -4.94335145e-02 -6.41273141e-01 3.60821903e-01
-1.78492174e-01 2.65097648e-01 7.51673460e-01 9.15177643e-01
-1.33851373e+00 -8.35849643e-01 -1.73469767e-01 4.23647165e-01
-2.41863757e-01 -1.98443919e-01 -8.20639133e-01 7.33242452e-01
5.34723818e-01 6.31603837e-01 2.88292140e-01 -4.76493359e-01
-2.90534962e-02 1.44162804e-01 9.82343554e-01 -4.97872293e-01
-5.75980127e-01 -1.89910114e-01 -1.33013785e-01 -3.05253893e-01
-4.36629593e-01 -2.34509051e-01 -1.97158742e+00 -6.16144836e-01
-1.83158904e-01 2.59846509e-01 7.61497736e-01 8.22623849e-01
1.60744697e-01 5.41606426e-01 4.82590050e-02 -8.16814005e-01
2.28086740e-01 -4.36235666e-01 -9.48420942e-01 -1.66941628e-01
1.38640106e-01 -3.18079352e-01 -4.02172893e-01 5.94943225e-01] | [8.033976554870605, 2.971726894378662] |
11c51d5f-c7dd-4938-a883-f4caca944aec | sample-based-distributional-policy-gradient | 2001.02652 | null | https://arxiv.org/abs/2001.02652v1 | https://arxiv.org/pdf/2001.02652v1.pdf | Sample-based Distributional Policy Gradient | Distributional reinforcement learning (DRL) is a recent reinforcement learning framework whose success has been supported by various empirical studies. It relies on the key idea of replacing the expected return with the return distribution, which captures the intrinsic randomness of the long term rewards. Most of the existing literature on DRL focuses on problems with discrete action space and value based methods. In this work, motivated by applications in robotics with continuous action space control settings, we propose sample-based distributional policy gradient (SDPG) algorithm. It models the return distribution using samples via a reparameterization technique widely used in generative modeling and inference. We compare SDPG with the state-of-art policy gradient method in DRL, distributed distributional deterministic policy gradients (D4PG), which has demonstrated state-of-art performance. We apply SDPG and D4PG to multiple OpenAI Gym environments and observe that our algorithm shows better sample efficiency as well as higher reward for most tasks. | ['Keuntaek Lee', 'Rahul Singh', 'Yongxin Chen'] | 2020-01-08 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-2.69686669e-01 4.02939767e-02 -5.13636053e-01 -4.94756289e-02
-8.05455983e-01 -4.41188216e-01 8.31283629e-01 3.43547873e-02
-8.78297210e-01 1.33797204e+00 4.17018294e-01 -2.80540973e-01
-5.53964913e-01 -8.48141253e-01 -9.22708988e-01 -9.20684636e-01
-4.46262449e-01 5.66395104e-01 5.38704805e-02 -2.62007147e-01
4.67680931e-01 2.48354882e-01 -1.47843301e+00 -2.04550862e-01
9.67850983e-01 5.22172928e-01 4.04541880e-01 7.59441137e-01
5.51247820e-02 9.80522990e-01 -5.27626514e-01 4.59460765e-02
2.80634284e-01 -5.86205065e-01 -6.12841904e-01 -3.37176055e-01
-1.36089951e-01 -6.11052215e-01 -4.58669662e-01 9.76004720e-01
8.73641133e-01 6.29169106e-01 9.20615613e-01 -1.46754384e+00
-7.30245113e-01 9.55532849e-01 -5.03921926e-01 6.31397516e-02
2.34150141e-01 3.46843272e-01 8.77475441e-01 -5.09201705e-01
4.88949746e-01 1.62962568e+00 3.12582672e-01 7.15222418e-01
-1.47280324e+00 -3.03225636e-01 1.87465087e-01 2.34086767e-01
-7.36501038e-01 2.89235264e-01 6.39690638e-01 -2.75101244e-01
7.04416454e-01 -2.74879605e-01 7.93571532e-01 1.35895622e+00
5.40849745e-01 1.46715605e+00 1.55747604e+00 -3.72638285e-01
1.00682497e+00 -3.13221574e-01 -3.50191742e-01 3.62632990e-01
1.01492763e-01 8.83755386e-01 -4.71272111e-01 -2.82948256e-01
9.10648942e-01 1.27515107e-01 4.27316017e-02 -8.52015436e-01
-9.92016494e-01 1.16012931e+00 2.77641624e-01 -9.87204313e-02
-7.49333799e-01 6.89326823e-01 3.63187969e-01 4.05754060e-01
3.35957348e-01 1.83966160e-01 -2.99147457e-01 -7.80079901e-01
-6.52682245e-01 1.04649019e+00 9.14390981e-01 7.75680542e-01
5.52296638e-01 2.93789059e-01 -8.73005629e-01 8.22304606e-01
3.83239955e-01 7.71477640e-01 7.53445983e-01 -1.25277174e+00
2.42675141e-01 -4.86679897e-02 5.04092813e-01 -3.09144676e-01
-1.85700908e-01 -2.38932416e-01 -3.41976196e-01 7.38250077e-01
4.82548118e-01 -4.90773559e-01 -6.57074153e-01 1.95825982e+00
3.13841045e-01 2.15748072e-01 1.15349039e-01 5.59460044e-01
3.30898887e-03 4.15182531e-01 2.30741650e-01 -4.85099107e-02
4.25026357e-01 -7.70878077e-01 -4.14151162e-01 -4.44345139e-02
5.21223783e-01 -2.01273680e-01 1.49643278e+00 4.69680429e-01
-9.19149697e-01 -3.71970654e-01 -8.29400063e-01 3.57130229e-01
-1.19199166e-02 -5.07817455e-02 4.86154169e-01 5.83088517e-01
-1.16682458e+00 1.21362996e+00 -9.51097548e-01 -2.48371854e-01
5.53490877e-01 1.65461749e-01 2.31762230e-01 -1.85356304e-01
-8.82963359e-01 7.94310570e-01 3.39598745e-01 -4.70012248e-01
-1.68960965e+00 -6.27144456e-01 -7.39118993e-01 -1.40559271e-01
5.64987600e-01 -4.01887506e-01 1.77760923e+00 -3.21488649e-01
-2.34283948e+00 2.00504884e-01 3.74049485e-01 -8.65801394e-01
7.51957655e-01 -5.33875346e-01 2.97605276e-01 -5.54600842e-02
1.05663076e-01 7.26471961e-01 9.15143073e-01 -9.88053918e-01
-7.07805216e-01 -4.03507650e-01 4.87559214e-02 4.59792346e-01
-3.46123264e-03 -6.75970435e-01 6.46476924e-01 -5.80958366e-01
-5.80676377e-01 -1.00802231e+00 -3.98040801e-01 -3.67979735e-01
-7.72167230e-03 -7.23803103e-01 5.29232144e-01 -2.78292030e-01
1.02736211e+00 -1.96954548e+00 3.39575827e-01 -3.23783383e-02
-1.53639123e-01 8.64261836e-02 -2.40239292e-01 8.20657969e-01
3.33853424e-01 -2.28895888e-01 -2.14507043e-01 -1.15746014e-01
6.43054724e-01 6.79523230e-01 -6.87785745e-01 4.62051988e-01
-1.39101878e-01 1.01866865e+00 -1.50831676e+00 -1.52055562e-01
2.31796309e-01 5.83084710e-02 -7.59973764e-01 3.14907372e-01
-6.00529134e-01 3.53290975e-01 -5.87034822e-01 3.54349315e-01
3.48235846e-01 3.44028145e-01 8.78231898e-02 7.27210104e-01
-1.28517255e-01 4.56419885e-02 -1.05186749e+00 1.76356184e+00
-2.74378240e-01 6.53698444e-02 -3.23756009e-01 -1.13665271e+00
9.66654599e-01 3.68059985e-02 7.77868986e-01 -5.79844177e-01
2.03160197e-02 3.20674747e-01 -3.98844145e-02 -3.73306781e-01
6.91238105e-01 -3.16299081e-01 -1.04861371e-01 6.40064359e-01
2.73689866e-01 -4.89987880e-01 3.60559464e-01 1.34108171e-01
1.07960474e+00 1.06002021e+00 1.96117073e-01 -5.28548181e-01
-4.29871157e-02 -2.07410514e-01 3.95356894e-01 1.26790571e+00
-4.91873175e-01 3.82227376e-02 7.32326090e-01 -5.90341948e-02
-9.59828377e-01 -1.42805207e+00 1.97934777e-01 1.27103961e+00
-1.07424043e-01 -1.05156310e-01 -6.95125163e-01 -7.05010533e-01
5.51900983e-01 1.08719409e+00 -7.75999248e-01 -3.96999419e-01
-3.81203651e-01 -5.26502967e-01 3.75926942e-01 5.91168821e-01
4.65686262e-01 -1.59543705e+00 -8.44534516e-01 5.24913609e-01
1.86205253e-01 -4.22462195e-01 -2.90736705e-01 2.39228249e-01
-1.10013759e+00 -8.79608691e-01 -9.89181101e-01 -2.88982034e-01
1.00846142e-01 -1.72686920e-01 1.13659227e+00 -7.35867202e-01
-1.13405041e-01 1.01256013e+00 -3.64047170e-01 -7.70979345e-01
-3.07158023e-01 -1.24394722e-01 3.36921901e-01 -4.01959389e-01
3.13429326e-01 -6.81568146e-01 -6.27128780e-01 -7.06110373e-02
-9.17728186e-01 -5.39363384e-01 6.12990916e-01 1.10130787e+00
7.17956483e-01 -1.98218063e-01 9.20145810e-01 -6.22070432e-01
1.41527820e+00 -5.64367056e-01 -6.93389177e-01 7.77712837e-03
-8.51282358e-01 5.58567524e-01 4.53524768e-01 -5.67414761e-01
-8.75944972e-01 -3.82467955e-01 9.31179617e-03 -3.13571125e-01
-5.60543360e-03 2.76537001e-01 3.13348204e-01 3.33136171e-01
7.06564426e-01 4.92095321e-01 2.74952471e-01 -2.58426875e-01
7.02913880e-01 4.42646682e-01 3.44853401e-01 -1.24839020e+00
2.68521309e-01 2.18742207e-01 1.34967476e-01 -5.78825593e-01
-6.59788489e-01 -1.01266578e-01 -1.57459691e-01 -3.00899148e-01
6.83650553e-01 -5.59768260e-01 -1.12894595e+00 5.39884150e-01
-3.77605587e-01 -1.20996892e+00 -1.09417641e+00 8.68260145e-01
-1.65515578e+00 3.93134467e-02 -5.57734132e-01 -1.32275784e+00
-6.12849928e-02 -1.02757597e+00 1.01013756e+00 3.48671615e-01
2.58563668e-01 -1.04122293e+00 7.11207151e-01 -5.43950438e-01
4.84717727e-01 3.64667386e-01 6.96766317e-01 -3.62814814e-01
-2.27852196e-01 2.39734545e-01 4.24541622e-01 5.63639879e-01
5.71214221e-03 -4.96418476e-01 -5.63591063e-01 -6.45126581e-01
-4.67045791e-02 -8.77631247e-01 7.73150027e-01 9.33644116e-01
1.13511801e+00 -2.62618288e-02 9.26744118e-02 2.21129373e-01
1.46850240e+00 1.88876301e-01 6.26668930e-01 4.76511002e-01
2.23687634e-01 2.76236087e-01 1.06811154e+00 1.00670147e+00
3.51498961e-01 2.17860028e-01 5.90305030e-01 4.61342871e-01
2.00008884e-01 -9.69693065e-01 8.08719575e-01 1.36551991e-01
-3.67227495e-02 5.91996349e-02 -4.39612746e-01 5.83643615e-01
-2.33277822e+00 -1.30978382e+00 5.53007126e-01 2.45476890e+00
1.07836747e+00 -2.04997323e-02 8.09382617e-01 -1.29103154e-01
4.87205923e-01 6.95060045e-02 -1.12211943e+00 -6.22830808e-01
2.82683909e-01 5.25546551e-01 6.40777171e-01 4.55246210e-01
-6.87430024e-01 9.39520121e-01 6.89862537e+00 1.10813570e+00
-6.99232996e-01 -1.01323165e-01 6.02432609e-01 -2.07417890e-01
-2.75253654e-01 -1.98262766e-01 -8.20969045e-01 5.09171426e-01
8.82987142e-01 -5.61683953e-01 9.07035530e-01 1.22813082e+00
3.69370490e-01 -4.10578966e-01 -1.03925598e+00 8.74225855e-01
-3.58457834e-01 -9.39083993e-01 -3.70450974e-01 3.24167937e-01
9.42019343e-01 2.61669248e-01 3.91808569e-01 1.06827593e+00
1.30768549e+00 -9.46867168e-01 8.05416942e-01 6.49258792e-01
3.58770758e-01 -1.15405381e+00 4.27579403e-01 7.23089218e-01
-7.09005654e-01 -5.67306936e-01 -8.02186608e-01 -2.40765288e-01
-4.23985906e-02 3.61956835e-01 -8.09673548e-01 3.42312455e-01
6.45359814e-01 7.59092689e-01 -2.65936311e-02 1.02224636e+00
-6.06301248e-01 9.17069674e-01 -2.24837884e-01 -6.17961168e-01
6.75375104e-01 -5.67873061e-01 5.74384630e-01 7.51715243e-01
3.61521006e-01 -4.23681647e-01 5.68389237e-01 9.34248090e-01
1.19264998e-01 5.89152947e-02 -8.90894890e-01 -1.99425220e-01
3.88402760e-01 8.95591319e-01 -2.93881923e-01 -1.41440853e-01
1.26845583e-01 6.71424329e-01 6.10983491e-01 3.29051882e-01
-7.97088385e-01 -2.64990389e-01 7.23208606e-01 -1.71771735e-01
4.82936382e-01 -4.74972576e-01 9.78325903e-02 -7.67952323e-01
-1.09898649e-01 -7.89876461e-01 2.80193239e-01 -3.78371328e-01
-1.43523633e+00 7.38546923e-02 3.98407489e-01 -1.23944831e+00
-9.17160749e-01 -5.61226189e-01 -3.97199064e-01 7.11689353e-01
-1.63778770e+00 -6.24276280e-01 3.10561627e-01 7.17477977e-01
5.51936090e-01 -2.22159341e-01 7.80829191e-01 -1.78841874e-01
-1.44658551e-01 3.45326275e-01 8.33357811e-01 -3.62967581e-01
6.62791252e-01 -1.88244283e+00 3.11471503e-02 4.10495937e-01
-9.79415029e-02 2.16325611e-01 7.48163164e-01 -7.18511820e-01
-1.46566629e+00 -8.70565772e-01 -6.43201126e-03 -1.95723236e-01
7.91025817e-01 -1.03991702e-01 -4.19428170e-01 5.38718641e-01
2.63750017e-01 -1.17912553e-01 4.20177430e-01 -9.75991637e-02
2.09281087e-01 1.61576029e-02 -1.35598338e+00 7.25109577e-01
9.63371336e-01 -1.61460981e-01 -5.43779612e-01 1.90211445e-01
5.95453799e-01 -3.77244830e-01 -7.65057921e-01 3.01545225e-02
4.75033253e-01 -9.53908682e-01 7.44616568e-01 -7.77268887e-01
5.18411934e-01 2.98174582e-02 -2.99503416e-01 -2.08516073e+00
-3.34427416e-01 -8.91891718e-01 -3.86682332e-01 8.64770651e-01
-1.45202398e-01 -6.49589479e-01 6.40144408e-01 3.61667842e-01
-1.14789471e-01 -1.12335503e+00 -9.00125206e-01 -1.30727339e+00
7.07945585e-01 -3.05294871e-01 5.84861636e-01 3.12359393e-01
1.30188584e-01 -8.17117542e-02 -5.60875118e-01 -5.09482503e-01
9.59647954e-01 1.27473354e-01 8.69058371e-01 -8.21378350e-01
-7.36112475e-01 -4.28249508e-01 -6.19577914e-02 -1.19410098e+00
2.42448196e-01 -8.06991518e-01 3.86686146e-01 -1.62075162e+00
-5.92562854e-02 -2.85049051e-01 -4.40501243e-01 1.96796030e-01
-2.02905551e-01 -3.65192384e-01 2.24329591e-01 -1.96352527e-01
-6.33708537e-01 1.35550737e+00 1.44993830e+00 9.43661034e-02
-4.83015984e-01 2.18797028e-01 -5.34095824e-01 5.26924193e-01
1.08113599e+00 -4.91768390e-01 -9.24594223e-01 -1.12883449e-01
-8.99552256e-02 8.93039629e-02 2.10546970e-01 -9.27537680e-01
-1.99293599e-01 -5.57049155e-01 1.85308054e-01 -4.38343316e-01
2.14225911e-02 -2.36055970e-01 -5.17318428e-01 8.61494482e-01
-6.74041212e-01 2.87806392e-01 -9.52084288e-02 1.00287509e+00
7.15081170e-02 -3.01675647e-01 7.78793752e-01 -2.75310159e-01
-6.52279258e-01 5.75360656e-01 -5.15317500e-01 4.69078392e-01
1.02165961e+00 8.68996084e-02 -1.24705926e-01 -5.06931961e-01
-6.88625932e-01 3.37790787e-01 1.52733132e-01 2.76227832e-01
8.29491556e-01 -1.62246120e+00 -8.43236983e-01 -1.49156019e-01
-1.14162602e-01 -2.42837936e-01 -3.53608876e-02 5.91608703e-01
-1.47514999e-01 2.74465829e-02 -3.33690166e-01 -4.17482436e-01
-5.41098475e-01 5.85577846e-01 1.59908235e-01 -6.58546329e-01
-7.33246505e-01 5.15763640e-01 -2.73678422e-01 -6.65857315e-01
3.32970500e-01 -4.63006347e-01 -4.18979153e-02 -3.42994720e-01
5.98079801e-01 7.39759922e-01 -5.00215709e-01 1.85316890e-01
1.77914739e-01 1.94410644e-02 1.67223111e-01 -7.44189382e-01
1.69624805e+00 1.79013655e-01 3.24749470e-01 7.37860322e-01
5.94542146e-01 -3.86994928e-01 -2.00687790e+00 -1.50521502e-01
4.68962975e-02 -6.10041380e-01 -4.31352854e-02 -7.19646454e-01
-5.59485674e-01 7.87403762e-01 8.65580082e-01 1.14748009e-01
6.65510297e-01 -2.98333377e-01 5.69144845e-01 4.71230417e-01
1.00892651e+00 -1.58159697e+00 4.70291615e-01 7.40165353e-01
8.68910968e-01 -1.00641322e+00 -1.40727758e-01 6.23989940e-01
-9.41420078e-01 7.75238514e-01 4.99322385e-01 -8.92077088e-01
7.27078795e-01 1.98117062e-01 -3.65439653e-01 1.93686768e-01
-7.03495920e-01 -3.62568080e-01 -2.80370176e-01 1.02173162e+00
1.79171637e-01 4.68470305e-01 -6.12499297e-01 2.53925651e-01
-3.23945612e-01 3.00393492e-01 5.04092991e-01 1.26483738e+00
-6.48692846e-01 -1.56673682e+00 -6.25594556e-02 4.56033081e-01
-3.46926183e-01 1.08382419e-01 5.04615866e-02 6.90490425e-01
-4.21560913e-01 8.37067664e-01 -1.99331865e-01 -1.01991192e-01
1.85327545e-01 -7.59845972e-02 1.07764328e+00 -4.24309433e-01
-3.03673327e-01 -2.85529494e-02 -4.37801510e-01 -7.42472649e-01
-3.51476461e-01 -8.55233669e-01 -1.42540419e+00 -1.86596125e-01
1.77992880e-01 2.15492755e-01 6.41875923e-01 8.60905170e-01
3.80605400e-01 5.16728401e-01 7.72231519e-01 -9.40143287e-01
-1.70666456e+00 -1.15064108e+00 -1.21837044e+00 3.32109779e-01
7.80999437e-02 -1.07241714e+00 -3.55930291e-02 -5.03007472e-01] | [4.035818576812744, 2.3749656677246094] |
59359406-dd05-46e1-bf08-53a650f2ec9f | sp2-a-second-order-stochastic-polyak-method | 2207.08171 | null | https://arxiv.org/abs/2207.08171v1 | https://arxiv.org/pdf/2207.08171v1.pdf | SP2: A Second Order Stochastic Polyak Method | Recently the "SP" (Stochastic Polyak step size) method has emerged as a competitive adaptive method for setting the step sizes of SGD. SP can be interpreted as a method specialized to interpolated models, since it solves the interpolation equations. SP solves these equation by using local linearizations of the model. We take a step further and develop a method for solving the interpolation equations that uses the local second-order approximation of the model. Our resulting method SP2 uses Hessian-vector products to speed-up the convergence of SP. Furthermore, and rather uniquely among second-order methods, the design of SP2 in no way relies on positive definite Hessian matrices or convexity of the objective function. We show SP2 is very competitive on matrix completion, non-convex test problems and logistic regression. We also provide a convergence theory on sums-of-quadratics. | ['Robert M. Gower', 'Deanna Needell', 'Martin Takáč', 'William J. Swartworth', 'Shuang Li'] | 2022-07-17 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [-1.35245174e-01 1.90198049e-01 -5.14852218e-02 -1.52322948e-01
-1.13595796e+00 -4.69770312e-01 2.13090792e-01 -8.93272832e-02
-3.11215132e-01 8.30782652e-01 2.50369869e-02 -2.62402594e-01
-2.80534625e-01 -3.13583404e-01 -8.85147810e-01 -9.10797060e-01
-4.17630188e-02 5.99858582e-01 -8.88640657e-02 -4.64577258e-01
2.88336009e-01 4.37351316e-01 -8.22042286e-01 1.60294194e-02
1.03062940e+00 7.73584485e-01 3.86409014e-02 6.63188219e-01
2.03445330e-01 3.19674909e-01 1.08998142e-01 -4.53092128e-01
7.51237214e-01 -4.73477632e-01 -7.21311808e-01 9.52814594e-02
5.33129871e-01 -3.17896791e-02 2.38288511e-02 1.05021489e+00
4.80930597e-01 1.66162059e-01 7.04793751e-01 -1.33081722e+00
-5.28437138e-01 4.53149557e-01 -7.11115181e-01 -1.68346345e-01
3.64612699e-01 -2.01933384e-01 1.13208210e+00 -1.58239567e+00
5.30146837e-01 1.27151179e+00 1.31653726e+00 1.69492349e-01
-1.89181697e+00 -1.54279605e-01 -2.05813214e-01 -1.05314605e-01
-1.66983378e+00 -2.96710789e-01 4.16509181e-01 -5.79388440e-01
5.61203122e-01 6.28477573e-01 4.72190142e-01 7.28449464e-01
1.84792206e-01 8.32630575e-01 1.24615562e+00 -5.90972126e-01
9.31332037e-02 3.37755173e-01 2.07442969e-01 7.67762482e-01
1.39869481e-01 -1.45168483e-01 -3.19945693e-01 -5.01139343e-01
8.76268864e-01 -4.08294141e-01 -3.69236916e-01 -5.85386574e-01
-1.14093029e+00 1.16012633e+00 8.93301219e-02 4.13044132e-02
-3.41418296e-01 8.17736611e-02 3.20932925e-01 3.13414514e-01
6.62675560e-01 6.17390633e-01 -5.57545781e-01 -3.83176319e-02
-1.06541812e+00 4.59655732e-01 1.09990025e+00 1.00243258e+00
8.22477043e-01 7.53987208e-02 -1.88757449e-01 8.86874676e-01
3.22296172e-01 5.63319266e-01 2.19729748e-02 -1.31760287e+00
5.51125169e-01 1.39401704e-01 3.28371823e-01 -1.02407849e+00
-5.79670012e-01 -5.23514569e-01 -8.13149035e-01 1.52710229e-01
6.12295330e-01 -2.09907711e-01 -3.03879738e-01 1.62812686e+00
4.04805601e-01 7.77248591e-02 -1.93823591e-01 7.50253797e-01
-3.54325026e-02 8.64528775e-01 -4.20537055e-01 -5.58097661e-01
7.31861115e-01 -1.04579723e+00 -6.68440521e-01 3.38340193e-01
8.34493458e-01 -6.15320861e-01 9.75394070e-01 7.64379561e-01
-1.52958667e+00 -3.28066319e-01 -7.91294515e-01 -2.77711630e-01
-1.77426279e-01 3.73960882e-01 5.09116709e-01 5.71180940e-01
-1.56128240e+00 9.08465147e-01 -7.40270853e-01 9.33025405e-02
-8.46066326e-02 6.64817333e-01 -3.05105925e-01 3.19017619e-01
-8.81455898e-01 9.36852932e-01 8.08593258e-03 3.96974593e-01
-4.98489022e-01 -1.09792745e+00 -8.61737311e-01 2.00017765e-02
2.61900157e-01 -7.06286371e-01 9.46730316e-01 -9.54913020e-01
-1.72997868e+00 7.40620732e-01 -3.96905124e-01 -3.72138441e-01
9.52436924e-01 -1.95928812e-01 1.11331172e-01 -1.31054029e-01
1.14305004e-01 2.65342087e-01 9.85752940e-01 -1.02819836e+00
-2.48933658e-01 -3.08522999e-01 -1.33804187e-01 1.88735485e-01
-6.63990453e-02 8.38338509e-02 -3.00318033e-01 -4.56301033e-01
1.19748116e-01 -1.29022396e+00 -7.56932080e-01 1.25203058e-01
-3.85418057e-01 -9.67008546e-02 1.60986513e-01 -9.76166725e-01
1.34081960e+00 -2.16480064e+00 7.05725133e-01 5.48273206e-01
1.61956385e-01 -7.17337579e-02 -1.49965346e-01 6.43738866e-01
-3.27222139e-01 1.52520239e-01 -6.66862488e-01 -6.45416677e-01
1.76160663e-01 1.89972937e-01 -2.53151625e-01 6.82447493e-01
2.01255143e-01 5.85149765e-01 -7.88586140e-01 -4.25452739e-01
-2.21037015e-01 4.98651952e-01 -8.81705880e-01 -3.09772909e-01
1.19629353e-01 3.71170580e-01 -2.55935252e-01 8.69120434e-02
1.14768100e+00 -1.27392516e-01 -7.94399381e-02 -1.15784645e-01
-4.94870603e-01 -8.11686832e-03 -1.73961484e+00 1.74712121e+00
-5.93800485e-01 4.80195582e-01 4.46224868e-01 -1.16592586e+00
6.16591513e-01 1.72073215e-01 6.89680219e-01 2.93882079e-02
-3.05929273e-01 6.15736604e-01 -5.17787278e-01 -2.98195809e-01
3.81630659e-01 -1.96304873e-01 2.40726650e-01 -2.04948068e-01
-1.62735745e-01 -1.03314497e-01 4.40553039e-01 1.28219172e-01
7.49833226e-01 2.47347295e-01 2.42072687e-01 -8.07500303e-01
8.85145545e-01 1.00982592e-01 5.22803187e-01 6.60943985e-01
2.96700776e-01 8.61592114e-01 8.82353008e-01 -2.22754672e-01
-1.14442658e+00 -1.09762859e+00 -5.73972702e-01 8.63485813e-01
-3.82608443e-01 -5.65025866e-01 -6.70970082e-01 -5.04401743e-01
2.23944649e-01 5.70346415e-01 -7.11909413e-01 2.43013233e-01
-6.34251416e-01 -8.05269063e-01 2.56253332e-01 3.98613751e-01
1.52511597e-01 -1.42298818e-01 2.84562320e-01 1.80522919e-01
-3.17185782e-02 -8.41764808e-01 -8.62600744e-01 2.46536985e-01
-9.53146577e-01 -8.85712266e-01 -9.81362641e-01 -7.92488813e-01
8.77796650e-01 1.63315311e-02 8.58064413e-01 -2.02688634e-01
2.52505541e-01 3.49930555e-01 2.00670063e-02 -4.67764363e-02
-3.82378250e-01 -1.86169326e-01 2.75628358e-01 2.62326896e-01
-2.73719758e-01 -4.63661373e-01 -3.49676192e-01 4.27840471e-01
-7.67157376e-01 1.08396992e-01 3.98004591e-01 1.12536705e+00
8.69282544e-01 -2.42234424e-01 1.98370919e-01 -1.03984118e+00
6.82984471e-01 -5.60137331e-01 -1.01689744e+00 2.25716054e-01
-6.84762180e-01 4.45128471e-01 8.63785386e-01 -4.29087311e-01
-7.12048233e-01 3.90886158e-01 -1.77474588e-01 -4.41279590e-01
7.72763133e-01 7.92226613e-01 1.48137197e-01 -5.44535100e-01
6.90509737e-01 1.02392033e-01 1.34103730e-01 -6.75114155e-01
3.52228552e-01 3.42126191e-01 3.16883057e-01 -6.86946630e-01
9.03888404e-01 4.83541071e-01 5.71695328e-01 -9.78024364e-01
-7.21628189e-01 -6.34624779e-01 -6.69983149e-01 1.48230821e-01
6.56007111e-01 -8.76892447e-01 -7.57654548e-01 2.00245455e-01
-1.05428004e+00 -4.97279853e-01 -4.99299318e-01 4.40797031e-01
-9.24936593e-01 4.29395735e-01 -6.33994102e-01 -8.19711804e-01
1.87637823e-04 -1.08772361e+00 1.08272231e+00 -2.91889668e-01
1.43099055e-01 -1.25181198e+00 4.14623439e-01 -9.39894561e-03
3.28967065e-01 2.72687376e-01 3.90798360e-01 -2.70049930e-01
-1.68252423e-01 -1.95465550e-01 -1.00488447e-01 7.57384181e-01
-4.10271317e-01 1.86701715e-01 -4.85962749e-01 -3.10227066e-01
3.25035512e-01 1.47661582e-01 5.27370989e-01 6.88058853e-01
9.66583788e-01 -7.31040776e-01 -1.45102650e-01 9.79261041e-01
1.67120886e+00 -5.78003526e-01 4.10590768e-01 7.87349567e-02
8.59401584e-01 3.87057513e-01 4.53814805e-01 5.64862013e-01
3.85619670e-01 1.06963468e+00 -1.42928183e-01 -1.70454413e-01
1.90114886e-01 -1.73520252e-01 6.56073391e-01 1.01854169e+00
-1.94118112e-01 5.12857974e-01 -8.27436507e-01 3.53177339e-01
-1.98118639e+00 -8.69962156e-01 -1.02539575e+00 2.56027007e+00
1.03218293e+00 -2.38272399e-01 3.94971162e-01 5.30268289e-02
4.61300284e-01 -4.35442507e-01 -2.10100830e-01 -6.99569046e-01
-2.50861734e-01 2.51015902e-01 1.04685235e+00 1.10133076e+00
-8.92009079e-01 5.93650639e-01 7.52362108e+00 9.16049659e-01
-7.42113829e-01 3.06913823e-01 3.04667532e-01 2.18296558e-01
-3.65247101e-01 2.65106469e-01 -1.00487208e+00 5.38114548e-01
8.89243305e-01 -1.51908264e-01 6.77260458e-01 8.49914551e-01
5.96856236e-01 -6.24062456e-02 -1.20257747e+00 8.53527129e-01
-7.67266750e-02 -1.08127165e+00 -4.03064162e-01 3.15626919e-01
1.22854924e+00 -6.22563958e-02 2.18120560e-01 1.75618276e-01
2.81425178e-01 -8.91948998e-01 6.87581420e-01 5.07827282e-01
6.77715957e-01 -6.17350459e-01 5.00166118e-01 3.02891493e-01
-1.05717576e+00 3.11176479e-02 -6.53879106e-01 -3.49171683e-02
1.80699453e-01 6.76073551e-01 -5.24986863e-01 4.78884995e-01
3.13150078e-01 5.76904237e-01 -3.92064929e-01 1.22500193e+00
-2.72950362e-02 5.51871836e-01 -7.37449586e-01 2.20101625e-01
3.08679342e-01 -9.58421230e-01 6.99304104e-01 1.35819554e+00
3.49229753e-01 5.86975962e-02 1.13333635e-01 8.63818645e-01
4.08824533e-01 5.38454890e-01 -4.41207498e-01 3.57274801e-01
-1.42526403e-01 1.12571466e+00 -1.99830398e-01 -5.13300896e-02
-5.49343348e-01 1.00352502e+00 2.41431355e-01 5.20216882e-01
-8.01922262e-01 -8.91080648e-02 5.20021617e-01 7.65444189e-02
2.99140841e-01 -7.19632030e-01 -5.31918824e-01 -1.34211040e+00
1.95289940e-01 -8.79492581e-01 2.05887109e-01 -5.94931781e-01
-1.22931719e+00 2.45975211e-01 1.50879234e-01 -1.12666130e+00
-2.80119151e-01 -8.60773742e-01 -3.04298937e-01 1.05155945e+00
-1.16607881e+00 -9.43053663e-01 1.91961929e-01 5.90559006e-01
3.25732470e-01 4.61196452e-01 7.44522989e-01 3.30035001e-01
-6.06855035e-01 6.46548808e-01 7.71773398e-01 -3.48727852e-01
6.39540672e-01 -1.55209363e+00 1.84052363e-01 8.18217993e-01
-1.54575154e-01 6.29979610e-01 1.07244945e+00 -3.78028601e-01
-1.63494909e+00 -6.05649948e-01 1.06661069e+00 -5.56306005e-01
1.03934908e+00 -4.36871499e-01 -8.29907298e-01 6.95218146e-01
-9.51195210e-02 -2.56261956e-02 4.20296282e-01 1.82411581e-01
-4.82096411e-02 -2.66217947e-01 -9.77610469e-01 5.81923187e-01
7.93487370e-01 -4.44245994e-01 2.92834528e-02 8.94744635e-01
4.69065785e-01 -6.32002056e-01 -1.05882490e+00 8.48932341e-02
3.90551597e-01 -6.32986069e-01 1.08645117e+00 -6.72646999e-01
2.63772666e-01 -3.19668531e-01 -2.41963968e-01 -1.27722847e+00
-4.33201224e-01 -1.26174843e+00 8.37718323e-02 9.68143940e-01
8.08946252e-01 -8.67126226e-01 5.46736717e-01 8.73428881e-01
-2.83189714e-01 -8.04490685e-01 -8.55543792e-01 -1.12299132e+00
3.73111725e-01 -3.26312840e-01 -8.61582235e-02 9.67811346e-01
3.30471277e-01 1.09218046e-01 -5.59680283e-01 1.51127338e-01
6.12362981e-01 -3.52763087e-01 8.08115304e-01 -1.03236496e+00
-7.86518753e-01 -3.40544879e-01 -1.39442399e-01 -1.25663841e+00
1.30587682e-01 -9.97121990e-01 3.83260734e-02 -1.16860282e+00
-4.33843508e-02 -6.46340370e-01 -2.72341538e-02 3.20352018e-01
-1.35995641e-01 1.41558215e-01 1.71442643e-01 2.24347681e-01
-1.91230133e-01 3.72582704e-01 1.24427128e+00 1.02437258e-01
-5.55066586e-01 3.86080235e-01 -4.86785799e-01 6.69527173e-01
5.55985272e-01 -3.79319906e-01 -2.14809701e-01 -2.65526503e-01
9.11075115e-01 3.11029404e-01 1.71709895e-01 -7.96909451e-01
2.41950989e-01 -1.44561082e-01 -8.09108019e-02 -3.17049414e-01
3.66464496e-01 -7.28146970e-01 3.65518779e-01 3.82192075e-01
-4.58505750e-01 3.39176714e-01 4.50677127e-02 3.36795300e-01
-3.64496298e-02 -6.50125682e-01 7.45983958e-01 1.35216594e-01
3.04238219e-02 1.89113766e-01 -2.07193568e-01 2.14665055e-01
6.31112874e-01 -1.17937177e-01 2.21784264e-01 -4.12370741e-01
-9.92987514e-01 3.38016123e-01 5.85229635e-01 -2.56098658e-01
2.63446182e-01 -1.47460866e+00 -1.08396184e+00 1.27512366e-01
-4.38139468e-01 -4.60595191e-02 -1.82115077e-03 1.74318087e+00
-7.70070553e-01 3.39720756e-01 2.43485570e-01 -5.01834095e-01
-9.81928766e-01 6.13134027e-01 3.03733289e-01 -4.66359049e-01
-2.10230067e-01 9.11098599e-01 2.69846290e-01 -2.99420595e-01
8.34032670e-02 -3.54451954e-01 2.41245374e-01 -1.09367967e-02
3.24911773e-01 7.29847848e-01 -1.94240790e-02 -5.41800499e-01
-3.11875761e-01 6.83625221e-01 2.87338883e-01 -4.82488930e-01
1.40017283e+00 -2.45413586e-01 -4.11148787e-01 6.07316375e-01
1.59861112e+00 4.50298697e-01 -1.34481108e+00 -2.03505880e-03
-4.09256220e-02 -3.04567337e-01 -7.94860348e-02 -3.21795106e-01
-8.52124810e-01 6.14483654e-01 7.22199678e-02 2.28993073e-01
1.04823864e+00 -5.01190126e-01 4.15364593e-01 3.11967582e-01
2.13409290e-01 -1.08720517e+00 -4.40029025e-01 5.73327541e-01
1.25703168e+00 -1.04599643e+00 2.94047028e-01 -8.29290450e-01
-5.81447780e-01 1.19060433e+00 -1.27289649e-02 -5.15817285e-01
7.44796157e-01 8.69242027e-02 -2.89006323e-01 3.18646222e-01
-4.50242668e-01 6.72225654e-02 6.09037578e-01 4.48612839e-01
3.29567999e-01 6.42716810e-02 -1.11520839e+00 2.38045618e-01
-2.02941090e-01 -1.95561603e-01 5.96988022e-01 3.65535498e-01
-4.77879345e-02 -1.32499886e+00 -4.30544585e-01 1.18051313e-01
-3.86773467e-01 -4.40070480e-01 -1.73683912e-01 7.58580685e-01
-1.92134112e-01 7.62078762e-01 -4.14873242e-01 -1.92119092e-01
1.84312016e-01 9.06132683e-02 5.71091115e-01 -3.22228134e-01
-5.87879121e-01 1.19567215e-01 1.64439440e-01 -7.31088459e-01
-9.32123810e-02 -1.00884008e+00 -1.01189208e+00 -4.39926356e-01
-6.02338195e-01 3.79526079e-01 9.71190870e-01 7.59172916e-01
1.85676053e-01 -1.98499069e-01 1.00530112e+00 -9.84243989e-01
-1.16614771e+00 -6.37920439e-01 -7.25258827e-01 9.23728719e-02
3.39686632e-01 -3.69152635e-01 -6.32618546e-01 3.64578590e-02] | [6.929137706756592, 4.377682209014893] |
49a66483-4856-4d61-b5c6-e8fd496c19e7 | opinion-tree-parsing-for-aspect-based | 2306.08925 | null | https://arxiv.org/abs/2306.08925v1 | https://arxiv.org/pdf/2306.08925v1.pdf | Opinion Tree Parsing for Aspect-based Sentiment Analysis | Extracting sentiment elements using pre-trained generative models has recently led to large improvements in aspect-based sentiment analysis benchmarks. However, these models always need large-scale computing resources, and they also ignore explicit modeling of structure between sentiment elements. To address these challenges, we propose an opinion tree parsing model, aiming to parse all the sentiment elements from an opinion tree, which is much faster, and can explicitly reveal a more comprehensive and complete aspect-level sentiment structure. In particular, we first introduce a novel context-free opinion grammar to normalize the opinion tree structure. We then employ a neural chart-based opinion tree parser to fully explore the correlations among sentiment elements and parse them into an opinion tree structure. Extensive experiments show the superiority of our proposed model and the capacity of the opinion tree parser with the proposed context-free opinion grammar. More importantly, the results also prove that our model is much faster than previous models. | ['Guodong Zhou', 'Yue Zhang', 'Zhongqing Wang', 'Xiaotong Jiang', 'Xiaoyi Bao'] | 2023-06-15 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [ 7.67809376e-02 1.99059680e-01 -1.96126118e-01 -6.79442227e-01
-6.61226451e-01 -7.80802727e-01 3.71449977e-01 1.49293765e-01
1.71849560e-02 2.61711150e-01 4.64933336e-01 -7.19427407e-01
4.07252282e-01 -1.22023284e+00 -2.91138858e-01 -5.51840365e-01
2.67485678e-01 3.86762828e-01 -1.02047808e-01 -5.12227476e-01
3.83144110e-01 -1.19599707e-01 -1.46242714e+00 2.63562202e-01
9.73907292e-01 1.11782539e+00 -2.34930083e-01 6.13546312e-01
-7.05698967e-01 7.41806388e-01 -6.46639824e-01 -1.02697277e+00
-2.12059468e-02 -4.77851987e-01 -6.85677469e-01 1.30831748e-01
2.17907932e-02 -1.03527270e-02 5.14435530e-01 1.28995049e+00
2.25693166e-01 -2.33623192e-01 3.85927737e-01 -1.06478679e+00
-7.72077382e-01 8.23364854e-01 -7.22271502e-01 -2.13060632e-01
3.79406273e-01 -3.83563191e-01 1.63134909e+00 -6.46136820e-01
6.21119499e-01 1.14469779e+00 5.97085834e-01 3.79045963e-01
-7.24849224e-01 -5.23240209e-01 8.71778488e-01 -9.21362266e-02
-6.46477282e-01 -9.16946381e-02 1.04349351e+00 -8.51613507e-02
1.20113456e+00 1.42596945e-01 9.89749730e-01 8.85900378e-01
6.42328143e-01 8.78223836e-01 1.23977637e+00 -3.46450627e-01
4.16906327e-01 1.41376808e-01 7.00055361e-01 7.21444130e-01
5.46663582e-01 -5.71614146e-01 -5.08365512e-01 -8.35888013e-02
1.48328215e-01 7.28492662e-02 2.11655751e-01 -4.14360583e-01
-5.96454501e-01 9.82052505e-01 -7.82736838e-02 2.80945808e-01
-3.43637168e-01 -1.28985539e-01 4.41406488e-01 1.89570427e-01
8.37888300e-01 2.57889748e-01 -9.74912345e-01 -1.64926648e-01
-6.57289505e-01 -5.42529635e-02 1.20470059e+00 1.02127624e+00
9.75323677e-01 1.33180479e-02 1.67337786e-02 5.80913186e-01
5.36380172e-01 8.10529530e-01 5.23019731e-01 -5.18604815e-01
4.41698939e-01 1.15127146e+00 -4.96883571e-01 -1.15634751e+00
-4.41942960e-01 -5.74234605e-01 -7.79135942e-01 -5.74174412e-02
-1.81763515e-01 -3.84738743e-01 -9.54200208e-01 1.49886823e+00
4.47524250e-01 -3.64962161e-01 3.45842093e-01 3.02853853e-01
8.60788405e-01 3.92101556e-01 1.07494228e-01 -2.72664279e-01
1.81500471e+00 -1.09816837e+00 -7.29670703e-01 -5.99568248e-01
9.05507743e-01 -7.71535277e-01 1.00738347e+00 2.77420551e-01
-8.43347907e-01 -2.81921834e-01 -1.05440462e+00 -7.43610710e-02
-5.42795122e-01 -8.03452134e-02 1.35050106e+00 1.16823494e+00
-1.03819680e+00 1.08252615e-01 -8.21070671e-01 -1.37540564e-01
2.94417053e-01 3.64914030e-01 -2.36984298e-01 1.14677116e-01
-1.02570617e+00 3.25642943e-01 9.21357945e-02 2.61880457e-01
-4.34101932e-02 -4.00714189e-01 -1.28988135e+00 1.41341984e-01
5.05650520e-01 -9.14032638e-01 1.36964190e+00 -1.13077509e+00
-1.68361604e+00 6.30227923e-01 -7.92572498e-01 -5.16287982e-02
-2.84263104e-01 -3.02706599e-01 -4.28847194e-01 -1.39762893e-01
2.45702222e-01 2.47630313e-01 6.11944318e-01 -1.19971251e+00
-7.61007547e-01 -6.07951522e-01 5.00793815e-01 2.63426423e-01
-5.91462970e-01 8.00927207e-02 -7.26706207e-01 -5.98194242e-01
2.49748945e-01 -8.60553503e-01 -5.25189936e-01 -8.42366755e-01
-4.21320885e-01 -2.77637959e-01 5.51067293e-01 -2.69458264e-01
1.44744575e+00 -1.78659701e+00 -1.95519015e-01 2.63039440e-01
1.87998384e-01 -9.53435227e-02 -2.65922338e-01 2.23553181e-01
-7.15880319e-02 3.43774736e-01 -3.06780577e-01 -5.74953914e-01
2.03445271e-01 3.60483736e-01 -4.70992625e-01 -3.18221331e-01
3.25423330e-01 1.30940795e+00 -7.42478430e-01 -4.18452203e-01
-1.34656683e-01 3.98549825e-01 -9.01890934e-01 6.33589998e-02
-4.21721369e-01 1.62758101e-02 -8.69448125e-01 9.22557950e-01
7.84423351e-01 -3.00811023e-01 6.14832342e-01 -9.57148150e-02
1.09191783e-01 5.81553996e-01 -8.67528081e-01 1.49071586e+00
-6.60580218e-01 1.42543688e-01 -1.13516495e-01 -8.59459400e-01
1.05515242e+00 1.18569080e-02 3.31037521e-01 -7.29894102e-01
4.02405828e-01 -3.77433971e-02 -1.35584131e-01 -2.28322163e-01
8.73183191e-01 -3.67175221e-01 -4.73502815e-01 7.82533705e-01
5.77319898e-02 -4.04414892e-01 6.21334434e-01 2.66850322e-01
8.48156214e-01 3.32596377e-02 4.09162462e-01 -1.79753974e-01
8.26686561e-01 -1.10184714e-01 7.45900393e-01 3.99768233e-01
5.15875109e-02 4.07406420e-01 1.16745329e+00 -4.68047023e-01
-5.20040154e-01 -6.18479967e-01 1.54547036e-01 1.20216012e+00
-8.31873622e-03 -1.08677387e+00 -8.02487671e-01 -1.25922322e+00
-4.70602304e-01 6.91838205e-01 -7.10517824e-01 9.80602950e-02
-5.12492001e-01 -1.25187981e+00 -5.99180795e-02 7.28120327e-01
3.12046766e-01 -9.78456557e-01 -7.88001269e-02 1.36033088e-01
-1.00307569e-01 -1.25260258e+00 -2.02058405e-01 4.46217030e-01
-1.02923465e+00 -1.00882983e+00 -1.56923771e-01 -8.49847257e-01
8.58702123e-01 1.08866870e-01 1.51304960e+00 3.35903431e-04
4.06215221e-01 1.80035725e-01 -7.65324950e-01 -6.24006391e-01
-2.16919363e-01 5.19209504e-01 -2.44761974e-01 -6.50773048e-02
9.40963745e-01 -6.65476799e-01 -5.20364523e-01 -1.44198313e-02
-1.01973915e+00 -1.07208520e-01 7.27983892e-01 5.90453088e-01
8.41752827e-01 2.31781170e-01 4.59008366e-01 -1.56840611e+00
9.63809967e-01 -3.21485728e-01 -5.08699596e-01 2.05695689e-01
-1.13906980e+00 2.44127020e-01 8.14551115e-01 7.12633505e-02
-1.18223345e+00 -1.82433560e-01 -5.90493083e-01 4.46458727e-01
3.63922515e-03 9.06113386e-01 -4.36069131e-01 3.59075099e-01
-8.13981444e-02 1.35504767e-01 -5.04840434e-01 -2.01690555e-01
5.19091904e-01 5.77399611e-01 1.58935770e-01 -5.81876695e-01
6.21601582e-01 6.18564248e-01 -1.47589833e-01 -3.53481442e-01
-1.36814737e+00 -3.07191461e-01 -6.61838174e-01 2.16606304e-01
8.43257785e-01 -8.13622057e-01 -4.04651254e-01 4.43614513e-01
-1.12608647e+00 1.17254481e-01 -2.24385396e-01 1.50318190e-01
-3.02778035e-01 5.17401040e-01 -6.27361834e-01 -8.08218718e-01
-8.81257594e-01 -1.04678118e+00 1.31020010e+00 4.56845582e-01
-3.46237838e-01 -1.29731691e+00 4.91850764e-01 4.64793354e-01
3.30514252e-01 -6.06723391e-02 1.15127981e+00 -6.58065438e-01
-3.77371550e-01 -4.85809326e-01 -6.67337030e-02 3.80297780e-01
1.37514994e-01 1.39022619e-01 -8.00150931e-01 1.03688434e-01
2.68158197e-01 -6.29422292e-02 9.16506231e-01 1.87027141e-01
7.60024011e-01 -6.75697103e-02 -6.04733787e-02 6.39669240e-01
1.23071802e+00 1.66924208e-01 5.49750745e-01 5.04566967e-01
6.50183022e-01 6.67756021e-01 5.31057596e-01 2.28583798e-01
9.97076273e-01 -1.13012930e-02 1.64951071e-01 -7.21127167e-02
2.31000826e-01 -3.90388459e-01 6.44690335e-01 1.73464119e+00
-4.62626442e-02 -3.89283299e-01 -5.52061200e-01 5.15429437e-01
-1.65066671e+00 -4.62476671e-01 -2.57608891e-01 1.52277553e+00
6.35839045e-01 4.67285156e-01 -3.14588368e-01 4.84728850e-02
1.52656391e-01 6.16068661e-01 -4.69055265e-01 -9.82710779e-01
-2.81467795e-01 7.87735820e-01 1.66387036e-01 4.32255030e-01
-1.14221013e+00 1.39370716e+00 6.98373938e+00 5.60025334e-01
-1.00699699e+00 -7.32611418e-02 6.51163459e-01 1.74232155e-01
-9.70321476e-01 3.54942352e-01 -1.02517796e+00 5.23433164e-02
8.07896733e-01 -2.68893957e-01 -1.56009421e-01 1.16235375e+00
-9.22128037e-02 -6.59854859e-02 -7.77833462e-01 5.09475350e-01
3.37710261e-01 -9.68434989e-01 3.92121941e-01 5.00785224e-02
9.21646178e-01 -2.49123707e-01 5.17644510e-02 4.31325227e-01
5.19532919e-01 -8.02343428e-01 4.42882031e-01 2.55043745e-01
3.63904506e-01 -9.58160877e-01 1.26510668e+00 -1.31284535e-01
-1.66015506e+00 1.66160494e-01 -3.80398244e-01 -4.02385414e-01
3.76859426e-01 1.03016758e+00 -1.60147324e-01 8.57026279e-01
6.78313136e-01 8.67977977e-01 -7.34628022e-01 1.87110573e-01
-9.15987909e-01 6.44321680e-01 1.73103623e-02 -3.55482489e-01
1.56837448e-01 -4.20913696e-01 2.02879712e-01 1.09371102e+00
2.72692084e-01 9.93267354e-03 -1.29156634e-01 4.37463641e-01
-6.08310290e-02 4.97692794e-01 -3.82541507e-01 -2.63214260e-01
-9.64671150e-02 1.76567185e+00 -9.90181923e-01 -5.01907527e-01
-8.16018403e-01 8.53739619e-01 3.00351888e-01 2.75731981e-01
-5.08179128e-01 -3.57217818e-01 9.52795386e-01 -4.35402781e-01
7.20999241e-01 -2.18950883e-01 -6.09467864e-01 -1.55912113e+00
4.25334156e-01 -1.16166282e+00 2.53588408e-01 -7.33726382e-01
-1.22106588e+00 9.86883402e-01 -4.44344729e-01 -8.71812046e-01
-5.80545604e-01 -8.66696000e-01 -7.52565861e-01 6.60248697e-01
-1.65807807e+00 -1.16397440e+00 3.89197543e-02 2.67973483e-01
5.62619209e-01 4.74395342e-02 1.08917296e+00 -1.11774683e-01
-6.99197888e-01 6.57909155e-01 -3.03640038e-01 2.49988988e-01
4.62273717e-01 -1.41087747e+00 8.72678876e-01 8.75751793e-01
3.41552019e-01 1.08500528e+00 4.58051354e-01 -7.62042463e-01
-1.54340339e+00 -8.53253186e-01 1.20784688e+00 -8.02640319e-01
1.04699266e+00 -5.45636892e-01 -6.66129887e-01 6.82882011e-01
4.02549386e-01 -5.99842250e-01 1.24811471e+00 6.47164047e-01
-6.89464569e-01 -9.45160463e-02 -4.75134194e-01 6.59252763e-01
9.40124989e-01 -5.13621032e-01 -8.64677250e-01 -7.86962286e-02
9.53395188e-01 -1.03752777e-01 -7.42988706e-01 5.57267487e-01
8.62077236e-01 -1.19973457e+00 5.01710832e-01 -5.88885725e-01
8.22499633e-01 -2.47445643e-01 -8.06721076e-02 -1.20428026e+00
-9.88983661e-02 -4.32787329e-01 -1.99519187e-01 1.40847266e+00
9.80127275e-01 -8.14972281e-01 1.16923189e+00 6.52655542e-01
2.98826043e-02 -1.08202839e+00 -3.94885987e-01 -3.16821218e-01
2.49974310e-01 -9.65279102e-01 8.80674362e-01 7.32976139e-01
2.03738570e-01 9.06038165e-01 -1.24685235e-01 4.54203784e-02
4.02580380e-01 7.50598788e-01 7.02644467e-01 -1.19257736e+00
-2.37194180e-01 -5.81457913e-01 -1.93645641e-01 -1.26835918e+00
3.10253888e-01 -6.19980335e-01 -1.17426425e-01 -1.92220652e+00
4.03896034e-01 -2.62361735e-01 -2.89887041e-01 4.75815028e-01
-6.52176857e-01 1.61456168e-01 -9.24110636e-02 -1.51313022e-01
-8.50228369e-01 5.87875426e-01 1.25016427e+00 -1.00086167e-01
-4.24003564e-02 1.03561692e-01 -1.63950825e+00 1.11276603e+00
8.04420590e-01 -5.49582899e-01 -4.20826524e-01 -4.26360846e-01
1.16933858e+00 -4.43907112e-01 -4.62096542e-01 -4.25119936e-01
3.06577742e-01 9.90165025e-03 4.74819280e-02 -9.36479986e-01
-7.50350878e-02 -6.52061403e-01 -4.20720130e-01 8.04022625e-02
1.17450498e-01 3.99430037e-01 6.55374452e-02 4.81846660e-01
-5.78036189e-01 -2.39524350e-01 5.39549962e-02 -3.56440470e-02
-6.11909389e-01 2.62512773e-01 -4.06301618e-01 4.17275913e-02
6.27688110e-01 -2.26156954e-02 -2.59356260e-01 -4.97364014e-01
-3.98614347e-01 2.66786903e-01 4.45128322e-01 5.87086439e-01
3.28633696e-01 -1.01756537e+00 -2.78306752e-01 5.59835553e-01
1.76570982e-01 1.31495208e-01 1.30137220e-01 5.34229279e-01
-2.66028166e-01 5.14580548e-01 3.03641409e-01 -3.21013838e-01
-1.19989347e+00 5.30812621e-01 -4.01999950e-02 -9.45417583e-01
-1.51370257e-01 7.52471566e-01 4.17959541e-01 -7.85884202e-01
-2.70912170e-01 -7.09610462e-01 -6.12897038e-01 2.55928010e-01
4.55355376e-01 -2.55731076e-01 2.20067084e-01 -5.72002411e-01
-3.51772815e-01 1.22317147e+00 -1.45803496e-01 -4.65465821e-02
1.37819350e+00 -2.40564659e-01 -6.72895849e-01 2.42968217e-01
9.44793344e-01 5.92325807e-01 -6.03948712e-01 1.16376877e-02
9.97386798e-02 5.06802835e-02 -7.71931186e-02 -5.49608827e-01
-1.30658102e+00 7.77913094e-01 -1.32448241e-01 5.00951171e-01
1.50191188e+00 9.88159180e-02 1.11759067e+00 4.48471397e-01
2.55404651e-01 -1.04177094e+00 -1.53403968e-01 9.79964972e-01
2.73677915e-01 -1.10877645e+00 7.09805787e-02 -7.22443283e-01
-7.03370392e-01 1.05569685e+00 4.60196823e-01 -1.86846461e-02
7.47584820e-01 5.89817584e-01 6.25199616e-01 -3.42383474e-01
-1.05441463e+00 -5.29578805e-01 2.31869757e-01 5.07250965e-01
6.90203786e-01 5.34191504e-02 -5.53853393e-01 1.22821963e+00
-8.06264937e-01 -2.18071476e-01 2.78129578e-01 8.95636618e-01
-3.23530436e-01 -1.71524084e+00 -4.09530476e-02 3.95619988e-01
-6.92173958e-01 -6.90922737e-01 -6.83729649e-01 4.69121128e-01
-5.31933096e-04 1.28497660e+00 -4.97418903e-02 -5.77678561e-01
5.32850444e-01 2.70234793e-01 9.66951698e-02 -7.58753598e-01
-7.59282947e-01 1.94113284e-01 1.26809463e-01 -6.09233320e-01
-6.65644646e-01 -4.98910725e-01 -1.13856339e+00 -1.80685267e-01
-3.56022269e-01 4.34383333e-01 8.32156539e-01 1.07907128e+00
5.66629052e-01 6.32837236e-01 6.25648677e-01 -1.53159574e-01
5.15336217e-03 -7.67135978e-01 -4.45004523e-01 1.37725621e-01
-1.00609958e-01 -1.10806748e-01 -5.12845278e-01 -1.65686198e-02] | [11.438981056213379, 6.7012410163879395] |
e66423d9-5932-489b-8c72-a3d821b7573b | table-based-fact-verification-with-self-2 | null | null | https://aclanthology.org/2022.coling-1.120 | https://aclanthology.org/2022.coling-1.120.pdf | Table-based Fact Verification with Self-labeled Keypoint Alignment | Table-based fact verification aims to verify whether a statement sentence is trusted or fake. Most existing methods rely on graph feature or data augmentation but fail to investigate evidence correlation between the statement and table effectively. In this paper, we propose a self-Labeled Keypoint Alignment model, named LKA, to explore the correlation between the two. Specifically, a dual-view alignment module based on the statement and table views is designed to discriminate the salient words through multiple interactions, where one regular and one adversarial alignment network cooperatively character the alignment discrepancy. Considering the interaction characteristic inherent in the alignment module, we introduce a novel mixture-of experts block to elaborately integrate the interacted information for supporting the alignment and final classification. Furthermore, a contrastive learning loss is utilized to learn the precise representation of the structure-involved words, encouraging the words closer to words with the same table attribute and farther from the words with the unrelated attribute. Experimental results on three widely-studied datasets show that our model can outperform the state-of-the-art baselines and capture interpretable evidence words. | ['Peng Yang', 'Guangzhen Zhao'] | null | null | null | null | coling-2022-10 | ['table-based-fact-verification'] | ['natural-language-processing'] | [ 2.84198951e-02 1.88507155e-01 -6.71104729e-01 -5.09083152e-01
-1.02025723e+00 -7.09518135e-01 7.56090581e-01 4.72451419e-01
1.20998964e-01 5.25791824e-01 5.58124959e-01 -4.04039562e-01
2.30542064e-01 -6.58890426e-01 -8.42947781e-01 -5.57079852e-01
4.07870471e-01 4.31150258e-01 -2.05671713e-02 -4.06063855e-01
5.31608760e-01 -3.20053846e-02 -8.22373331e-01 5.96401155e-01
1.14635313e+00 9.26269591e-01 -3.68910164e-01 1.23813776e-02
-4.41971481e-01 1.04307878e+00 -7.51705945e-01 -1.11766076e+00
2.87679344e-01 -5.37016690e-01 -5.88302553e-01 2.25890458e-01
7.61560500e-01 -2.45252311e-01 -3.31997722e-01 1.38871729e+00
3.81662488e-01 -3.28091681e-01 5.97898424e-01 -1.42892718e+00
-1.16105306e+00 1.12492859e+00 -8.83237362e-01 2.22786084e-01
5.84113717e-01 5.67407012e-02 1.65134668e+00 -1.06737053e+00
4.21010524e-01 1.29760599e+00 4.88665581e-01 1.94874093e-01
-8.44795704e-01 -1.00013244e+00 7.40122974e-01 6.35839522e-01
-1.37483478e+00 -2.67347544e-01 1.36951125e+00 -2.08996400e-01
3.17340732e-01 1.92799985e-01 5.12612462e-01 1.49980927e+00
3.17747951e-01 9.38529670e-01 1.35798240e+00 -2.55772293e-01
-5.02912663e-02 4.62813795e-01 3.86105329e-01 9.60705042e-01
6.03128374e-01 -3.27240467e-01 -7.42645383e-01 -3.51395249e-01
9.02667493e-02 -1.21001936e-01 -3.93980414e-01 -5.54487169e-01
-1.21247113e+00 9.20906544e-01 5.98140001e-01 1.68150499e-01
-2.14816049e-01 -1.63162157e-01 4.48462665e-01 1.63794756e-01
5.82520664e-01 4.16613191e-01 -3.91483694e-01 3.86132509e-01
-7.34390140e-01 1.37011573e-01 6.37233436e-01 1.09574783e+00
5.85199773e-01 -1.54791534e-01 -5.32498300e-01 2.48405010e-01
5.73119640e-01 3.57191801e-01 4.06489819e-01 -2.08104596e-01
1.15643990e+00 1.27869558e+00 -3.28158773e-02 -1.67499661e+00
-5.73732108e-02 -8.43327045e-01 -9.35357094e-01 -2.48408243e-01
2.12107897e-01 3.18308830e-01 -6.45475566e-01 1.44008923e+00
5.11999428e-01 1.93622500e-01 9.74080861e-02 1.00049436e+00
9.95960653e-01 3.68629336e-01 -2.08719060e-01 5.30843697e-02
1.86027682e+00 -1.16122758e+00 -9.94860351e-01 -5.51666975e-01
7.05335557e-01 -7.28472412e-01 1.28977287e+00 2.87920445e-01
-7.57516265e-01 -4.39817309e-01 -1.54899383e+00 -3.19026500e-01
-5.77064157e-01 1.20848544e-01 5.94958425e-01 6.21407509e-01
-4.25327808e-01 9.31418464e-02 -2.25522235e-01 3.04154664e-01
5.92678666e-01 -1.31576300e-01 -3.96355063e-01 -1.61227629e-01
-1.77223182e+00 1.00366521e+00 3.97453457e-01 2.28290930e-01
-6.03948474e-01 -4.97138560e-01 -1.05947900e+00 5.31211719e-02
7.74459183e-01 -7.76272655e-01 6.29542053e-01 -9.08868253e-01
-1.05566823e+00 1.06295776e+00 -7.16736764e-02 -3.41851026e-01
5.84009171e-01 -2.66158879e-01 -7.15720952e-01 -3.51266488e-02
3.75900686e-01 9.71611869e-03 9.99365985e-01 -1.63116860e+00
-2.06360266e-01 -6.44218206e-01 2.00675115e-01 2.45237827e-01
-3.12783241e-01 -1.37712240e-01 -3.03765148e-01 -9.81821835e-01
3.02654386e-01 -4.94064599e-01 1.51478440e-01 -1.10615425e-01
-1.07328498e+00 -1.47817805e-01 7.16329217e-01 -9.55850303e-01
1.47486758e+00 -1.85577619e+00 2.28681028e-01 3.26339990e-01
6.75566554e-01 1.22226611e-01 1.26598049e-02 4.33185130e-01
-2.61898994e-01 2.13598207e-01 -1.03124812e-01 -3.62789333e-01
2.62186807e-02 5.34350425e-02 -7.56507516e-01 6.47055566e-01
1.74638391e-01 1.08023739e+00 -1.05019379e+00 -6.25780702e-01
-2.71177292e-01 1.33315593e-01 -1.49216339e-01 1.00878082e-01
-1.30689248e-01 3.58960658e-01 -6.12122715e-01 8.28004837e-01
8.79842520e-01 -6.70465291e-01 3.09288621e-01 -6.33951545e-01
5.21911979e-01 7.06839204e-01 -1.04562759e+00 1.55754864e+00
-6.98103830e-02 2.97097743e-01 -8.17516968e-02 -7.83901811e-01
1.13916957e+00 9.60984603e-02 -1.25339940e-01 -5.03965139e-01
1.21146895e-01 1.34095684e-01 -5.36042899e-02 -4.65178996e-01
5.22245705e-01 -5.99315241e-02 -1.71410084e-01 4.48658854e-01
-1.85720205e-01 -1.14992224e-01 -2.84198374e-01 6.87089086e-01
7.31621027e-01 -7.81327114e-02 5.78551710e-01 -2.00860739e-01
9.19300258e-01 -2.19547749e-01 5.72211683e-01 4.66011912e-01
-3.30588073e-01 2.41707966e-01 8.33378673e-01 -4.16957557e-01
-7.14056134e-01 -8.14475536e-01 3.16546440e-01 6.92303658e-01
7.59660423e-01 -6.78481817e-01 -5.12480438e-01 -1.43259168e+00
1.60577506e-01 1.03592825e+00 -9.33200061e-01 -5.19956946e-01
-3.86925757e-01 -3.66749734e-01 5.46428204e-01 5.29934108e-01
7.71484256e-01 -4.65868115e-01 3.07746530e-01 -1.75844938e-01
-6.19060457e-01 -1.19776893e+00 -8.82570684e-01 -1.22125924e-01
-3.88226926e-01 -1.18467689e+00 7.98199102e-02 -5.61028719e-01
9.61511254e-01 4.54124480e-01 1.22672403e+00 3.66821080e-01
3.85063618e-01 -1.63405687e-02 -3.76312405e-01 -2.94901311e-01
-4.21129316e-01 -8.85270163e-02 -1.28416419e-01 4.57970530e-01
5.19213676e-01 -3.75401527e-01 -5.57230592e-01 3.22111219e-01
-8.36223722e-01 1.69120505e-01 6.98054373e-01 9.04181778e-01
3.75875473e-01 -1.51235759e-01 3.57981563e-01 -1.19630945e+00
6.92203045e-01 -5.58288276e-01 -3.16035748e-01 7.42363334e-01
-9.41131711e-01 3.04849356e-01 6.95566833e-01 -4.72806722e-01
-6.11188233e-01 -4.18837935e-01 2.40588024e-01 -5.27390778e-01
3.91228378e-01 6.72211051e-01 -8.71652305e-01 1.57182559e-01
3.20508510e-01 4.64044094e-01 -2.47303829e-01 3.12513188e-02
6.65435791e-01 4.89009112e-01 3.69571418e-01 -6.02715254e-01
1.22522831e+00 5.95772445e-01 -1.58099502e-01 1.14799030e-01
-1.38150513e+00 -3.65640968e-01 -5.29622912e-01 -1.63796738e-01
6.35729373e-01 -9.51776624e-01 -5.99568069e-01 2.16298625e-01
-1.38577569e+00 3.82619083e-01 3.15559888e-03 1.93908978e-02
-5.46077155e-02 6.74137473e-01 -4.55923796e-01 -7.49373138e-01
-3.07067007e-01 -1.12043715e+00 1.23814881e+00 7.17888325e-02
-1.59975603e-01 -9.77083623e-01 -1.37126833e-01 9.95212913e-01
-1.74404141e-02 1.75574496e-01 1.09897768e+00 -1.19233155e+00
-7.31379092e-01 -5.30261993e-01 -3.41172069e-01 2.53564000e-01
4.00723279e-01 -6.33454695e-02 -7.59101927e-01 -1.23374104e-01
2.62678444e-01 -4.07013595e-01 8.61079693e-01 -3.01285654e-01
1.09311938e+00 -7.98233092e-01 -2.59194732e-01 2.71604002e-01
9.67480302e-01 -2.62733072e-01 6.20018363e-01 3.29569161e-01
1.11512089e+00 6.38379872e-01 7.02856183e-01 1.32490426e-01
9.01097953e-01 5.61202943e-01 5.97684801e-01 -8.90234392e-03
1.14985175e-01 -9.25507545e-01 3.54190350e-01 1.01060855e+00
3.64724040e-01 -5.97907662e-01 -6.61736488e-01 3.10954303e-01
-1.79485977e+00 -9.43277419e-01 -2.82251984e-01 1.87035203e+00
8.78900766e-01 5.23836195e-01 -2.28457898e-01 2.24102825e-01
8.86866629e-01 6.65374875e-01 -4.14036334e-01 -3.94824483e-02
-5.21230876e-01 -4.06234056e-01 3.02323282e-01 5.62935829e-01
-1.06437373e+00 9.34656441e-01 5.42292690e+00 1.11242068e+00
-8.56078386e-01 5.15099689e-02 9.05676425e-01 4.53942746e-01
-9.96502101e-01 2.61131346e-01 -6.97969854e-01 5.36742866e-01
2.86707669e-01 -1.41406775e-01 -2.44935956e-02 7.94466317e-01
-1.68452099e-01 3.59052151e-01 -9.20104325e-01 7.05460429e-01
6.38046741e-01 -1.27225280e+00 6.33754790e-01 6.18346594e-02
6.02217853e-01 -7.31143355e-01 2.83587188e-01 2.89266407e-01
2.39427432e-01 -9.49077308e-01 9.02817190e-01 3.71322244e-01
5.36513984e-01 -6.01072431e-01 9.64384854e-01 2.32357129e-01
-1.14135003e+00 2.37978786e-01 -2.87873894e-02 6.30518720e-02
-9.40835178e-02 6.33786261e-01 -7.22019196e-01 1.04297292e+00
1.17447205e-01 7.19541371e-01 -7.30266929e-01 1.55089706e-01
-9.15583193e-01 4.96188998e-01 3.57374728e-01 -2.02986732e-01
2.20940471e-01 -3.03578198e-01 6.01515114e-01 8.95297110e-01
-1.62106261e-01 3.19997333e-02 3.89063418e-01 1.02759790e+00
-4.98070151e-01 1.18851006e-01 -5.90168417e-01 -1.34136930e-01
5.53749979e-01 1.29622364e+00 -4.94229764e-01 -5.29480457e-01
-5.23795903e-01 1.12497258e+00 6.04163468e-01 1.38171688e-01
-1.01576924e+00 -1.12903737e-01 4.36730802e-01 -1.30008990e-02
3.01897734e-01 -7.04037771e-02 -5.92330515e-01 -1.49675822e+00
4.11995053e-01 -1.23863161e+00 5.06752074e-01 -8.04929137e-01
-1.66089749e+00 5.87169230e-01 -2.81989217e-01 -1.51496279e+00
1.80825457e-01 -5.41414499e-01 -7.39139557e-01 8.93230438e-01
-1.75262105e+00 -1.57700264e+00 -2.44956553e-01 5.67837536e-01
2.90517837e-01 -1.87311783e-01 6.03226364e-01 -7.26768048e-03
-6.56679392e-01 1.02975845e+00 -4.08023536e-01 5.19733191e-01
8.72409701e-01 -1.29812157e+00 4.67878640e-01 1.18565750e+00
5.08500278e-01 1.01543248e+00 7.78124988e-01 -9.75709140e-01
-1.40862656e+00 -8.41140687e-01 1.11096275e+00 -8.13449502e-01
1.02304578e+00 -6.99893892e-01 -1.21475768e+00 6.52497828e-01
3.28677326e-01 1.77150950e-01 9.48260486e-01 1.60922736e-01
-1.33916926e+00 -2.04516396e-01 -1.01497209e+00 6.75413072e-01
8.62983465e-01 -9.03907239e-01 -8.57777297e-01 4.50814992e-01
8.84953618e-01 -5.23574471e-01 -3.48978937e-01 4.18062329e-01
2.20614389e-01 -8.44332218e-01 7.81643689e-01 -1.03545117e+00
9.09563720e-01 -4.56181556e-01 -1.60937369e-01 -1.27635515e+00
-3.40485811e-01 -4.68576521e-01 -5.66388547e-01 1.59101009e+00
5.94704688e-01 -6.58667445e-01 6.21179700e-01 3.21213782e-01
5.60475606e-03 -8.96467745e-01 -1.05076885e+00 -5.57793558e-01
-1.32096499e-01 -3.97523865e-02 6.87182128e-01 1.49374616e+00
2.28191704e-01 7.91834891e-01 -5.34520924e-01 5.40144920e-01
4.70932782e-01 5.44206977e-01 7.26372719e-01 -8.20462704e-01
-2.79377013e-01 -4.90923375e-01 -3.98002863e-01 -1.16636181e+00
4.52466786e-01 -9.90235507e-01 -1.73595309e-01 -1.34958935e+00
5.14063895e-01 -1.57463662e-02 -3.89860392e-01 2.83633292e-01
-1.05646014e+00 -1.74432993e-01 -1.36112899e-01 1.83217824e-01
-5.90488911e-01 9.19553936e-01 1.41451871e+00 -6.27931893e-01
4.45712149e-01 -1.65309429e-01 -1.37547767e+00 6.33645415e-01
5.35809219e-01 -3.73985887e-01 -4.63388324e-01 -3.18797946e-01
5.07762730e-01 -6.66368753e-02 4.55966264e-01 -3.71105522e-01
4.72459793e-01 -1.30280331e-01 2.36339763e-01 -6.60803378e-01
1.98530853e-01 -8.88794422e-01 -4.67756093e-01 2.99849033e-01
-5.76372445e-01 3.44822466e-01 -9.08538625e-02 1.05133474e+00
-3.11041921e-01 2.45946031e-02 3.49251240e-01 -1.27858773e-04
-2.31071085e-01 3.47302973e-01 1.49155214e-01 2.91512877e-01
8.34182203e-01 -2.78117862e-02 -8.60297918e-01 -5.53678513e-01
-4.26624939e-02 3.38000476e-01 2.39023462e-01 4.81968015e-01
7.93400288e-01 -1.55750430e+00 -1.03826344e+00 7.45182708e-02
5.70431709e-01 -1.27636284e-01 2.36765340e-01 6.75166607e-01
-1.33206397e-01 9.98854414e-02 1.42139047e-01 -1.84580594e-01
-1.21645570e+00 9.88641679e-01 1.90206587e-01 -8.08594882e-01
-3.36888641e-01 9.53208864e-01 1.62926093e-01 -3.30517352e-01
1.08994238e-01 -1.91076025e-01 -2.17450947e-01 1.36204273e-01
5.58966517e-01 1.74136143e-02 2.38261074e-02 -6.86232388e-01
-5.35042107e-01 3.16876978e-01 -6.15174174e-01 1.12914450e-01
6.97582364e-01 -1.58999786e-01 -3.68826866e-01 5.12575865e-01
9.51918602e-01 7.04093635e-01 -8.13101768e-01 -5.75692415e-01
3.99895757e-02 -7.07741261e-01 -1.07589893e-01 -8.74142051e-01
-9.71700728e-01 8.04414392e-01 -1.89328603e-02 2.11834669e-01
8.59162748e-01 2.05110526e-03 7.02322900e-01 8.26038793e-02
2.30320066e-01 -6.26562536e-01 4.16801333e-01 2.81463563e-01
1.14578974e+00 -1.49663377e+00 2.61966348e-01 -9.06215310e-01
-9.24860299e-01 9.41278696e-01 9.20956314e-01 -3.11273132e-02
2.56946653e-01 3.03188097e-02 2.00458840e-01 -4.93485451e-01
-5.85787296e-01 2.64150411e-01 6.98367834e-01 3.03296655e-01
3.17515135e-01 -1.26507245e-02 -2.71832168e-01 9.47885036e-01
-1.96360931e-01 -8.04966211e-01 3.47518235e-01 5.66287637e-01
-8.77671465e-02 -9.69858229e-01 -1.84409037e-01 2.63391256e-01
-3.91232818e-01 -5.55730164e-01 -1.09084654e+00 6.54810250e-01
1.10766394e-02 1.00547028e+00 -2.96973914e-01 -8.18175316e-01
2.00687781e-01 4.60207872e-02 1.67854205e-01 -3.26539278e-01
-7.82115757e-01 -3.76230389e-01 1.07749991e-01 -3.68297100e-01
-2.99871564e-01 -3.56288970e-01 -9.98052597e-01 -4.37409669e-01
-6.45535588e-01 2.27968022e-01 3.04154307e-01 1.17003953e+00
3.98195386e-01 4.85823959e-01 8.41013849e-01 5.21483608e-02
-8.19342494e-01 -9.03789997e-01 -4.66740936e-01 5.13312995e-01
4.08095658e-01 -5.66489518e-01 -5.29854357e-01 -2.93804646e-01] | [9.105487823486328, 7.915650844573975] |
e22d1fa9-1e47-4b7f-a48a-d6689d293783 | a-new-dataset-and-boundary-attention-semantic | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/6832 | https://ojs.aaai.org/index.php/AAAI/article/view/6832 | A New Dataset and Boundary-Attention Semantic Segmentation for Face Parsing | Face parsing has recently attracted increasing interest due to its numerous application potentials, such as facial make up and facial image generation. In this paper, we make contributions on face parsing task from two aspects. First, we develop a high-efficiency framework for pixel-level face parsing annotating and construct a new large-scale Landmark guided face Parsing dataset (LaPa). It consists of more than 22,000 facial images with abundant variations in expression, pose and occlusion, and each image of LaPa is provided with an 11-category pixel-level label map and 106-point landmarks. The dataset is publicly accessible to the community for boosting the advance of face parsing.1 Second, a simple yet effective Boundary-Attention Semantic Segmentation (BASS) method is proposed for face parsing, which contains a three-branch network with elaborately developed loss functions to fully exploit the boundary information. Extensive experiments on our LaPa benchmark and the public Helen dataset show the superiority of our proposed method. | ['Tao Mei', 'Xiaobo Wang', 'Yue Si', 'Hao Shen', 'Hailin Shi', 'Yinglu Liu'] | 2020-04-03 | null | null | null | proceedings-of-the-aaai-conference-on-1 | ['face-parsing'] | ['computer-vision'] | [ 4.09910947e-01 3.60095143e-01 -2.60871828e-01 -1.04104114e+00
-1.12042427e+00 -4.22560275e-01 2.67777920e-01 -6.64167583e-01
-1.47710860e-01 3.93062413e-01 -1.01777904e-01 4.76384833e-02
4.12703693e-01 -5.77914059e-01 -7.75429428e-01 -5.82907677e-01
1.18507698e-01 4.31096792e-01 -5.54741211e-02 1.04879007e-01
-7.09019974e-02 7.24117279e-01 -1.47494423e+00 2.03395784e-01
6.06515944e-01 1.55504072e+00 -1.13542743e-01 3.60895038e-01
-2.55521148e-01 3.24815571e-01 -1.68731928e-01 -8.16246867e-01
2.55280554e-01 -4.63746071e-01 -9.09510791e-01 3.65446597e-01
8.63293469e-01 -4.76182550e-01 1.26116440e-01 1.07998264e+00
5.46862066e-01 -1.87623680e-01 3.00264418e-01 -1.41369033e+00
-2.29029685e-01 3.01964402e-01 -9.73173380e-01 -2.46584475e-01
1.43374890e-01 -6.63472041e-02 8.84016216e-01 -1.09629130e+00
6.77328408e-01 1.58846414e+00 7.07757354e-01 1.04281616e+00
-9.74700749e-01 -1.10066831e+00 2.16911674e-01 -1.49149373e-01
-1.17559505e+00 -8.50860119e-01 1.10498536e+00 -2.06314340e-01
3.06882650e-01 -2.73738559e-02 3.35795403e-01 9.63095069e-01
-4.28914517e-01 6.63773119e-01 1.08035946e+00 -3.00683856e-01
1.95280574e-02 -2.61393458e-01 -1.18823454e-01 1.40918124e+00
-7.34886527e-02 -2.27909729e-01 -5.26268065e-01 -7.66199529e-02
8.96801889e-01 -2.59520084e-01 7.17939213e-02 -3.24424118e-01
-5.03185630e-01 8.28725159e-01 5.67978442e-01 -2.36979455e-01
-4.23851609e-02 2.72351891e-01 2.70963043e-01 -2.88639545e-01
7.82234907e-01 -2.81204760e-01 -5.22449136e-01 1.83467552e-01
-7.14012802e-01 1.62067175e-01 4.69459027e-01 9.48637366e-01
8.88623357e-01 -1.84517637e-01 -6.75660819e-02 1.21602190e+00
5.23519754e-01 5.60214698e-01 -1.28138632e-01 -1.51471162e+00
5.22421598e-01 5.44930220e-01 -3.33692372e-01 -9.31758940e-01
-5.04966676e-01 1.99223056e-01 -6.95385575e-01 4.90707234e-02
3.84378642e-01 -2.39150867e-01 -1.02390313e+00 2.10614109e+00
7.55966842e-01 3.89836431e-01 -2.43723318e-01 7.58080125e-01
1.36521721e+00 4.43661660e-01 4.91807073e-01 -1.09456584e-01
1.72102022e+00 -1.15720689e+00 -6.04184687e-01 -3.01419824e-01
2.42845118e-01 -4.28524464e-01 9.27001297e-01 1.26990899e-01
-1.19936585e+00 -4.99528855e-01 -5.87734997e-01 -4.26129967e-01
-1.29168123e-01 3.55977833e-01 1.02799642e+00 8.97216797e-01
-1.27297175e+00 2.68595248e-01 -7.00173259e-01 -2.68120795e-01
1.39456224e+00 5.31921506e-01 -6.58662915e-01 -2.22329199e-01
-7.14923799e-01 6.35592490e-02 -3.20177816e-04 4.49453771e-01
-6.93170011e-01 -6.27414227e-01 -1.30398774e+00 6.07557185e-02
5.23558617e-01 -4.98638004e-01 1.18102968e+00 -1.16502416e+00
-1.76016700e+00 1.50465560e+00 -4.66258913e-01 2.80293912e-01
4.03783232e-01 -1.57559682e-02 -1.77256111e-02 4.89302963e-01
3.46166551e-01 1.42811060e+00 8.78307879e-01 -1.21866632e+00
-4.69108999e-01 -8.78890038e-01 -2.27174401e-01 -9.53467861e-02
-2.80644923e-01 3.98377836e-01 -1.25449347e+00 -5.75668395e-01
1.49529696e-01 -8.14939082e-01 -1.49289593e-01 4.05963689e-01
-5.10593176e-01 -4.10165548e-01 6.97931767e-01 -8.58400404e-01
6.11024141e-01 -2.13376665e+00 -4.92873862e-02 1.38514951e-01
2.71522515e-02 1.08403809e-01 -4.48201269e-01 -2.72119224e-01
-5.65785430e-02 2.14752033e-01 -6.05914474e-01 -8.78382921e-01
-1.28486812e-01 6.19970188e-02 -1.93722725e-01 3.48449588e-01
7.12507963e-01 1.13256919e+00 -5.19506931e-01 -8.40456843e-01
-1.47214413e-01 7.10727870e-01 -8.66153061e-01 1.79955423e-01
-3.18218768e-01 6.85387790e-01 -7.05489814e-01 1.33308530e+00
1.06734061e+00 -1.44986808e-01 5.85889965e-02 -1.76629171e-01
4.06901896e-01 -2.11453170e-01 -6.00605249e-01 1.95737898e+00
-3.36997628e-01 2.30212644e-01 5.68592548e-01 -6.50603950e-01
8.71626019e-01 1.16487026e-01 6.11423135e-01 -5.75006247e-01
3.10757995e-01 -4.58081206e-03 -6.41586244e-01 -2.90783733e-01
-1.51681274e-01 -7.18972459e-02 -1.28287718e-01 1.75050378e-01
1.95599779e-01 -1.73119709e-01 4.26159091e-02 2.84935925e-02
7.18308508e-01 4.49555546e-01 -1.53795198e-01 -1.22997500e-01
7.48992980e-01 -3.80060375e-01 9.46259081e-01 4.96963151e-02
-4.17783141e-01 7.22499430e-01 9.47134793e-01 -3.65630984e-01
-5.41811585e-01 -8.37036967e-01 -3.94362777e-01 1.36098909e+00
9.62174609e-02 -3.21023405e-01 -1.39844084e+00 -1.07537973e+00
-1.26581416e-01 -7.41186645e-03 -7.60182083e-01 3.63184780e-01
-6.81290388e-01 -8.05347443e-01 5.87492824e-01 7.47997761e-01
8.47399414e-01 -1.42822981e+00 -1.52250612e-02 -2.69501716e-01
-2.74634093e-01 -1.61163735e+00 -6.31556094e-01 -2.56058782e-01
-8.32568109e-01 -1.17350626e+00 -5.71116269e-01 -1.13180923e+00
1.03359544e+00 -8.15991387e-02 1.13476396e+00 2.68267274e-01
-6.51545227e-01 3.41088980e-01 -8.51070955e-02 -2.14520693e-01
1.14363998e-01 -1.15929671e-01 -4.15646017e-01 1.79704875e-01
2.17969611e-01 -2.30957910e-01 -4.71650124e-01 3.02505732e-01
-5.88262916e-01 -3.06033175e-02 4.40225065e-01 7.70354509e-01
9.05636072e-01 -2.83229291e-01 6.99591517e-01 -1.31301498e+00
3.73742431e-02 -2.57259399e-01 -8.22927833e-01 2.25914791e-01
-1.38422281e-01 -3.42119396e-01 2.15727970e-01 1.51492864e-01
-1.38258684e+00 5.38741887e-01 -6.33329272e-01 -1.38136715e-01
-2.28540823e-01 -7.49288127e-02 -9.30890381e-01 -5.07758021e-01
-1.60418540e-01 -2.29594499e-01 8.80379081e-02 -5.60856104e-01
5.61318398e-01 3.76861542e-01 8.27413917e-01 -1.03381252e+00
6.63424611e-01 6.69560194e-01 2.35437244e-01 -6.94346905e-01
-1.05576861e+00 -7.65734613e-02 -8.19152772e-01 -7.25657940e-02
1.11056232e+00 -1.02229643e+00 -7.84385383e-01 7.18149722e-01
-1.26563585e+00 -4.66399312e-01 2.29883179e-01 -2.96446919e-01
-6.40511990e-01 3.51052612e-01 -7.74633348e-01 -6.62006140e-01
-4.39984143e-01 -1.39373147e+00 1.68348026e+00 4.97627527e-01
3.25385004e-01 -7.18793273e-01 -3.70256156e-01 8.93590212e-01
-5.46625964e-02 4.58124489e-01 8.04490447e-01 -2.10705489e-01
-5.31917870e-01 1.65941730e-01 -6.32319450e-01 3.45017612e-01
-2.89435405e-02 7.44811296e-02 -1.34355354e+00 -8.36746767e-02
-2.56114364e-01 -7.14585125e-01 9.46219742e-01 3.92983764e-01
2.00662541e+00 -2.99368296e-02 -3.67582977e-01 1.28716898e+00
1.15404475e+00 -6.77717675e-04 5.31491876e-01 -1.32758573e-01
9.38898981e-01 1.10169744e+00 5.13003886e-01 2.57458091e-01
5.17590106e-01 7.12523758e-01 5.03580928e-01 -4.33148861e-01
-3.21494341e-01 -3.86270344e-01 2.11579859e-01 2.44248688e-01
-2.46066810e-03 6.01043291e-02 -6.37069225e-01 2.09209621e-01
-1.27334380e+00 -5.85644960e-01 6.96159229e-02 1.68035984e+00
7.90150940e-01 -3.54935288e-01 1.81157514e-01 -4.22084630e-01
8.13501298e-01 1.67782277e-01 -7.57354796e-01 -2.25399509e-01
-7.40139037e-02 6.29877865e-01 1.16388641e-01 4.31074053e-01
-1.45231140e+00 1.46002698e+00 6.03268480e+00 8.13010335e-01
-1.01471293e+00 1.77740738e-01 1.62125945e+00 2.06957698e-01
4.88805510e-02 -4.14330274e-01 -1.14706063e+00 5.30935407e-01
5.30796707e-01 5.21936297e-01 2.84523904e-01 1.10179102e+00
-3.80842499e-02 1.11814193e-01 -8.33200932e-01 1.16255713e+00
3.06710690e-01 -1.19034016e+00 -4.44814637e-02 2.75336429e-02
6.79814279e-01 -2.85984069e-01 2.36655176e-01 1.20777048e-01
7.30680898e-02 -1.30131316e+00 3.49494457e-01 1.38690710e-01
1.36056662e+00 -1.05716097e+00 4.92065817e-01 -1.68288752e-01
-1.33859742e+00 1.52670052e-02 -3.18436563e-01 3.93066198e-01
1.05533578e-01 2.57123590e-01 -2.86301285e-01 1.22106835e-01
9.12667513e-01 5.94060242e-01 -6.16193593e-01 4.39547807e-01
-4.70518857e-01 5.83485126e-01 -3.17647696e-01 5.69123149e-01
1.38867542e-01 -3.85945171e-01 -1.34764940e-01 9.45212007e-01
2.38817677e-01 4.55294758e-01 1.58439562e-01 8.41707647e-01
-7.45793760e-01 2.52311230e-01 -2.95710534e-01 7.26320446e-02
4.15054083e-01 1.82173836e+00 -1.01177597e+00 -9.30571929e-02
-4.82222557e-01 1.04860258e+00 5.11694193e-01 2.54771292e-01
-8.69120300e-01 8.00511241e-03 8.43705177e-01 3.80319767e-02
4.50914681e-01 2.23021910e-01 -3.38684916e-01 -7.99327493e-01
1.29260063e-01 -5.81712604e-01 4.04375762e-01 -5.94946206e-01
-1.22588575e+00 7.85426676e-01 -1.94759130e-01 -2.92903215e-01
4.56383228e-02 -8.51177096e-01 -6.72050595e-01 7.40876198e-01
-1.91154706e+00 -1.68314755e+00 -5.85204542e-01 5.77488542e-01
5.31911731e-01 -1.62210047e-01 7.03255653e-01 4.00874555e-01
-1.10143530e+00 9.06691790e-01 -5.25019228e-01 6.37778699e-01
6.38854444e-01 -8.81087065e-01 4.70386535e-01 6.12581611e-01
5.02495170e-02 3.40739012e-01 -1.02648340e-01 -4.81536478e-01
-1.26602817e+00 -1.46864080e+00 7.53921926e-01 -4.65500742e-01
2.42838547e-01 -7.76078582e-01 -6.95151329e-01 7.83447087e-01
-2.22428352e-01 6.50227249e-01 5.33431292e-01 -8.59215930e-02
-4.14566994e-01 -3.28517854e-01 -1.54637420e+00 3.83972794e-01
1.35436904e+00 -2.51093626e-01 1.23897120e-01 4.74819779e-01
5.65073431e-01 -4.95791495e-01 -6.32949412e-01 7.17938304e-01
4.35309410e-01 -1.04665673e+00 1.11179984e+00 -3.72295558e-01
5.91770887e-01 1.23746768e-01 9.03694890e-03 -6.56499922e-01
5.11088967e-03 -7.13554502e-01 3.14913273e-01 1.74753976e+00
1.10075727e-01 -5.41140258e-01 1.23215461e+00 7.31683671e-01
1.17414549e-03 -1.14039969e+00 -1.02612293e+00 -1.99079335e-01
1.25145748e-01 -3.57471317e-01 7.50030339e-01 5.37356198e-01
-6.02334023e-01 1.90980062e-01 -1.47176966e-01 1.28072292e-01
8.33513856e-01 1.29702881e-01 8.08923662e-01 -1.30746531e+00
2.19070122e-01 -4.97407764e-01 -3.06779534e-01 -1.21607077e+00
9.03894901e-01 -9.49462414e-01 1.93304405e-01 -1.08088255e+00
4.66333777e-01 -6.60863876e-01 3.69403027e-02 9.00261223e-01
-3.15380126e-01 9.84979630e-01 -6.39950261e-02 -1.63835675e-01
-6.17290258e-01 5.90146542e-01 1.33360898e+00 2.90419795e-02
2.92558044e-01 -3.40062752e-02 -8.44008207e-01 1.01622319e+00
4.33955967e-01 -3.20083499e-01 -3.16430420e-01 -4.56393898e-01
-1.71884701e-01 6.07617125e-02 4.46207494e-01 -5.55243850e-01
-1.19287401e-01 -3.52735966e-02 6.09480381e-01 -5.16964614e-01
5.85581243e-01 -6.24213338e-01 -3.22031945e-01 8.84981826e-02
-1.88666165e-01 -1.32253438e-01 2.16955721e-01 4.79089081e-01
-2.11471453e-01 -2.78128348e-02 1.18572330e+00 -6.65676296e-02
-9.79231954e-01 9.52949226e-01 6.48617446e-01 3.05505991e-01
1.16383696e+00 -2.06454415e-02 -1.05676934e-01 -9.62402448e-02
-6.19115174e-01 3.26297402e-01 4.46240544e-01 3.92156601e-01
5.04700363e-01 -1.33751047e+00 -7.41601229e-01 5.89859188e-01
3.22820134e-02 4.27513003e-01 3.09395581e-01 5.47317326e-01
-5.49936056e-01 1.67836338e-01 -1.22292817e-01 -6.22154117e-01
-1.47295880e+00 5.52924126e-02 2.47893304e-01 8.06310102e-02
-5.80518961e-01 1.49399710e+00 6.43860698e-01 -7.20759749e-01
2.95368969e-01 -6.42324612e-03 -1.74564153e-01 -7.35316053e-02
6.68032050e-01 5.91481142e-02 -8.23658705e-02 -1.07620597e+00
-5.73249817e-01 1.22298789e+00 1.74521133e-01 1.10767931e-01
1.37990141e+00 -3.75088362e-04 -4.93675232e-01 -2.58575320e-01
1.36680365e+00 -7.72637203e-02 -1.68226230e+00 1.32984261e-03
-1.31824300e-01 -5.30951381e-01 -4.52939495e-02 -8.12748075e-01
-1.76350224e+00 9.80545580e-01 4.24289346e-01 -5.53495288e-01
1.40496588e+00 3.26576889e-01 9.62802112e-01 1.82850566e-02
4.27122921e-01 -8.55455935e-01 1.76074654e-01 3.06948036e-01
7.69677579e-01 -1.57198358e+00 -2.26759493e-01 -1.18967891e+00
-5.54548800e-01 9.93572831e-01 7.91188538e-01 9.75987390e-02
6.46137416e-01 2.78612405e-01 2.51277506e-01 -2.08374456e-01
-3.25139284e-01 -8.39487910e-02 2.22300991e-01 7.31955528e-01
3.99500728e-01 -1.05699748e-01 -2.99236514e-02 9.03375864e-01
-1.48295060e-01 -1.42609641e-01 -2.21658602e-01 4.69697922e-01
-2.07223192e-01 -1.28194880e+00 -3.23548466e-01 8.34224373e-02
-5.85624874e-01 7.29517639e-02 -4.84225452e-01 7.45141625e-01
2.37701297e-01 8.41464400e-01 2.03589067e-01 1.54636264e-01
-8.40484351e-02 1.29454881e-01 6.03799820e-01 -7.30705202e-01
-1.33127868e-01 1.10451080e-01 -8.28422159e-02 -1.07042038e+00
-3.16190481e-01 -6.70920491e-01 -1.53561735e+00 1.73554057e-04
5.04528433e-02 -7.91302547e-02 1.01892054e+00 7.76005983e-01
3.89704227e-01 2.50306636e-01 8.21498632e-01 -1.09561574e+00
-1.39443710e-01 -6.25643969e-01 -5.35556793e-01 4.78447199e-01
8.22501443e-03 -8.43674779e-01 -1.21696323e-01 2.25446910e-01] | [13.486132621765137, 0.6509491801261902] |
bef7c26e-7d80-4069-b4d1-e50a72673ce1 | synthaspoof-developing-face-presentation | 2303.02660 | null | https://arxiv.org/abs/2303.02660v2 | https://arxiv.org/pdf/2303.02660v2.pdf | SynthASpoof: Developing Face Presentation Attack Detection Based on Privacy-friendly Synthetic Data | Recently, significant progress has been made in face presentation attack detection (PAD), which aims to secure face recognition systems against presentation attacks, owing to the availability of several face PAD datasets. However, all available datasets are based on privacy and legally-sensitive authentic biometric data with a limited number of subjects. To target these legal and technical challenges, this work presents the first synthetic-based face PAD dataset, named SynthASpoof, as a large-scale PAD development dataset. The bona fide samples in SynthASpoof are synthetically generated and the attack samples are collected by presenting such synthetic data to capture systems in a real attack scenario. The experimental results demonstrate the feasibility of using SynthASpoof for the development of face PAD. Moreover, we boost the performance of such a solution by incorporating the domain generalization tool MixStyle into the PAD solutions. Additionally, we showed the viability of using synthetic data as a supplement to enrich the diversity of limited authentic training data and consistently enhance PAD performances. The SynthASpoof dataset, containing 25,000 bona fide and 78,800 attack samples, the implementation, and the pre-trained weights are made publicly available. | ['Naser Damer', 'Marco Huber', 'Meiling Fang'] | 2023-03-05 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 1.53977633e-01 -7.24695250e-02 5.36997020e-02 -2.65161753e-01
-7.32604325e-01 -7.54064143e-01 6.30784273e-01 -3.50133598e-01
-4.00290675e-02 6.62570834e-01 -1.59951463e-01 -9.43655595e-02
-1.88918963e-01 -7.62418628e-01 -4.77480859e-01 -6.31410837e-01
-3.86295408e-01 1.82001919e-01 -3.34557593e-02 -3.37262481e-01
1.09029837e-01 9.59559500e-01 -1.82214606e+00 6.08848155e-01
5.29187560e-01 1.31098616e+00 -5.89631677e-01 2.99378455e-01
1.76868930e-01 -3.75238173e-02 -9.61804688e-01 -1.09603226e+00
6.42907500e-01 -1.97105601e-01 -1.99196845e-01 -3.01518321e-01
8.51777673e-01 -6.52333915e-01 -3.58503282e-01 7.42064834e-01
9.72216725e-01 -5.28361976e-01 4.16153342e-01 -1.91248715e+00
-4.65029389e-01 3.82683456e-01 -4.17677879e-01 -1.82870403e-01
6.20189369e-01 2.91573822e-01 3.67730111e-01 -1.12779343e+00
7.77549863e-01 1.37407804e+00 7.51934648e-01 1.00983489e+00
-8.52717876e-01 -1.31481647e+00 -5.02088487e-01 1.35045260e-01
-1.53492677e+00 -6.99355781e-01 9.60465431e-01 -3.12652916e-01
3.78142029e-01 2.25640744e-01 3.24944288e-01 1.83350325e+00
-8.84194300e-02 3.93010557e-01 1.10536540e+00 -5.07700086e-01
2.03876883e-01 6.13473237e-01 1.26804903e-01 3.56085330e-01
5.44191599e-01 3.02783996e-01 -6.97735429e-01 -6.70077562e-01
2.99830109e-01 -2.77826577e-01 -3.18562806e-01 -7.75737017e-02
-3.22837025e-01 6.60557210e-01 -2.78838098e-01 1.70233116e-01
-2.21096754e-01 -4.46605414e-01 4.32220429e-01 3.59062523e-01
7.58536234e-02 3.38311970e-01 -1.27451539e-01 -2.10806006e-03
-8.06029618e-01 3.65530640e-01 9.96856391e-01 8.80819023e-01
4.29045707e-01 2.87261069e-01 -1.64140448e-01 7.42258191e-01
1.57777488e-01 8.28464985e-01 2.71759212e-01 -5.20232320e-01
4.70072687e-01 4.71620768e-01 6.77702203e-03 -1.22747886e+00
6.88375533e-02 1.51169911e-01 -4.65120077e-01 3.45146596e-01
5.69710433e-01 -3.32143933e-01 -7.98112035e-01 1.81993937e+00
1.81651801e-01 1.46214515e-01 1.57614112e-01 4.40864444e-01
6.69879198e-01 4.07489717e-01 7.95736462e-02 -5.95274456e-02
1.50641167e+00 -1.00903869e-01 -6.86020672e-01 2.45509818e-01
2.46759251e-01 -1.00514531e+00 1.20475709e+00 7.61685371e-01
-8.40651572e-01 -4.95502114e-01 -1.25037801e+00 6.86874747e-01
-6.07477427e-01 2.32488394e-01 4.62708592e-01 1.90882623e+00
-1.06388092e+00 1.64180130e-01 -1.24418795e-01 -3.09983820e-01
9.57956314e-01 5.72458804e-01 -8.70305896e-01 -3.22230816e-01
-1.49938858e+00 5.20371139e-01 1.39770269e-01 -2.42495444e-02
-9.10996675e-01 -8.93286645e-01 -7.17566192e-01 3.49085964e-02
2.38421351e-01 2.59351522e-01 6.07343495e-01 -6.13858104e-01
-1.44522834e+00 8.47613752e-01 4.38626498e-01 -4.72585082e-01
4.55638111e-01 -3.70900854e-02 -1.11307812e+00 3.23540211e-01
-3.82660180e-01 4.54280257e-01 1.23131776e+00 -1.24522257e+00
-1.49224743e-01 -3.91125143e-01 -1.15836330e-01 -7.60639846e-01
-1.12251699e+00 4.79301810e-01 -1.09705448e-01 -5.43190181e-01
-6.80496395e-01 -8.06477606e-01 4.17712420e-01 1.84141859e-01
-1.77063957e-01 1.42578855e-01 1.48377359e+00 -7.04518557e-01
1.20951223e+00 -2.47519875e+00 -7.01108098e-01 7.23893166e-01
-1.53086081e-01 1.08991659e+00 -4.28669631e-01 7.93238580e-01
-3.47790956e-01 2.99857229e-01 3.59404758e-02 -1.84501871e-01
1.03747830e-01 -2.23110393e-02 -7.47662902e-01 2.69688874e-01
6.60016596e-01 3.27568382e-01 -3.85503024e-01 -3.53210926e-01
-7.93422665e-03 6.24115884e-01 -6.51718140e-01 2.19572246e-01
2.30105780e-02 1.57077163e-01 -4.45152640e-01 9.85317230e-01
1.55539024e+00 4.78158146e-01 1.69119000e-01 -9.69066843e-02
2.28942961e-01 -2.62051255e-01 -1.47500956e+00 1.10929787e+00
-1.48226097e-01 3.23236108e-01 2.73385972e-01 -4.66832817e-01
1.32544255e+00 4.70587492e-01 6.51161611e-01 -5.45060277e-01
3.27405870e-01 2.12856159e-01 -1.02235332e-01 -4.21423584e-01
2.92593867e-01 3.31842899e-01 -7.31852874e-02 5.96840084e-01
2.65688241e-01 2.42795348e-01 2.58705288e-01 1.45984545e-01
9.65519309e-01 -1.04245111e-01 -9.61186066e-02 -5.09191267e-02
9.48113084e-01 -4.29374427e-01 5.10106921e-01 4.50086951e-01
-3.97846311e-01 5.13961017e-01 6.64472282e-01 -2.91293472e-01
-6.81725204e-01 -1.32281816e+00 -3.87979865e-01 4.34681416e-01
-3.51960540e-01 -7.18407273e-01 -1.16260326e+00 -8.68871868e-01
1.73143357e-01 3.84073853e-01 -4.78986859e-01 -2.59746253e-01
-6.54549181e-01 -6.90451801e-01 1.51422787e+00 1.51519045e-01
7.62853622e-01 -9.88119960e-01 -2.26528153e-01 -1.98905542e-01
3.25350851e-01 -1.22558677e+00 -2.95064837e-01 -5.29447198e-01
-1.71602319e-03 -1.44592953e+00 -4.33456063e-01 -6.00004077e-01
5.30347586e-01 -1.19044088e-01 6.53915226e-01 5.80727421e-02
-8.15399170e-01 5.12491167e-01 -1.95206970e-01 -6.97417140e-01
-6.19703889e-01 -1.96592420e-01 4.01862442e-01 5.38889229e-01
5.62787294e-01 -4.31515753e-01 -4.06278163e-01 6.06929660e-01
-9.68273163e-01 -8.24051440e-01 3.19988400e-01 8.08988035e-01
-1.22070201e-01 -6.52451664e-02 1.00381649e+00 -9.14032698e-01
8.61532569e-01 -4.05907989e-01 -6.17570221e-01 3.18281710e-01
-4.33762223e-01 -4.58777905e-01 6.36253655e-01 -8.64642262e-01
-1.11762488e+00 -5.39975427e-02 -4.30350006e-01 -4.77241546e-01
-1.83729991e-01 2.75182966e-02 -7.57419229e-01 -5.04140258e-01
1.01357388e+00 9.36352238e-02 6.06964827e-01 -3.13280702e-01
1.97241649e-01 1.02802360e+00 4.87234414e-01 -9.69893575e-01
1.06433833e+00 2.42602497e-01 -5.39355120e-03 -8.86398017e-01
1.08124033e-01 3.42857689e-02 -1.62279412e-01 -3.37765396e-01
1.91942051e-01 -6.68746769e-01 -1.04761124e+00 1.02126563e+00
-8.65352511e-01 1.77184597e-01 -5.04477322e-02 2.17124492e-01
-1.71437278e-01 6.00227535e-01 -4.65049952e-01 -1.02497113e+00
-5.54357290e-01 -1.16084552e+00 8.62123370e-01 2.63196319e-01
-6.41385987e-02 -4.88877356e-01 -1.80058405e-01 2.26231709e-01
7.10785449e-01 6.90177202e-01 6.96447074e-01 -8.53642941e-01
-2.54595965e-01 -7.25493133e-01 -5.33688851e-02 6.65548146e-01
2.67844796e-01 4.23387378e-01 -1.44266474e+00 -5.39229155e-01
5.50086014e-02 -5.68330407e-01 1.43798247e-01 -2.70146400e-01
1.18559182e+00 -3.96868378e-01 -1.35193348e-01 4.62622851e-01
1.11490345e+00 3.43351483e-01 1.06160045e+00 5.91104031e-02
5.61858155e-02 8.90209436e-01 7.03456223e-01 7.11010039e-01
-1.32967696e-01 8.25314403e-01 2.85615087e-01 1.08030163e-01
-1.81247160e-01 -2.66682416e-01 5.25152564e-01 -2.69428995e-02
2.28134185e-01 -1.63942501e-01 -9.17379618e-01 1.16937913e-01
-9.43547726e-01 -1.19192088e+00 1.97198391e-01 2.24471521e+00
8.55607808e-01 2.21568663e-02 3.20500374e-01 4.70396638e-01
7.41024792e-01 -8.43000188e-02 -1.65890813e-01 -4.20680612e-01
-3.23939979e-01 8.95426035e-01 2.39892215e-01 4.57053035e-02
-1.07248282e+00 8.36462677e-01 5.83690691e+00 1.06997573e+00
-1.50536585e+00 -1.08932592e-01 4.55414355e-01 7.38468990e-02
8.31664205e-02 -4.19257849e-01 -1.17116022e+00 7.07274139e-01
1.19132340e+00 -2.79701650e-01 1.77032918e-01 1.03346121e+00
-8.20181370e-02 4.24774617e-01 -9.45959508e-01 1.07523465e+00
3.89997959e-01 -1.31516194e+00 2.27645427e-01 3.09734970e-01
2.99772888e-01 -8.01507711e-01 7.78802693e-01 1.43866509e-01
-6.31235018e-02 -9.36250925e-01 3.14967513e-01 1.04685888e-01
1.20229089e+00 -9.14355934e-01 6.24702871e-01 -9.95109789e-03
-8.71103525e-01 -3.43602747e-01 -3.01050991e-01 4.48874950e-01
-2.88442582e-01 2.12911278e-01 -1.12401986e+00 6.76847219e-01
6.03936553e-01 1.98420137e-01 -8.03840041e-01 7.71456540e-01
9.07796621e-02 6.75400496e-01 -5.83709538e-01 1.64373964e-01
-2.24054322e-01 4.17091958e-02 3.00066084e-01 1.10653675e+00
5.89085162e-01 -2.08185688e-01 -3.25946063e-01 6.61040246e-01
-2.16698721e-01 2.54944861e-01 -8.74560237e-01 -2.24198461e-01
8.64039540e-01 1.48386955e+00 -4.72242385e-02 1.80142194e-01
-1.94538891e-01 3.41854066e-01 -2.24006578e-01 1.49717033e-01
-8.78513694e-01 -4.08168644e-01 9.88651037e-01 3.61038387e-01
1.00880072e-01 1.62081808e-01 -9.51192603e-02 -9.26714242e-01
2.03346983e-01 -1.56206810e+00 6.28871262e-01 -3.71545643e-01
-1.41590059e+00 9.43071067e-01 6.06886707e-02 -1.45954621e+00
-3.74828756e-01 -1.01796234e+00 -7.16145813e-01 1.02055609e+00
-1.18343294e+00 -1.52068865e+00 -4.29920912e-01 9.88281429e-01
-4.24820073e-02 -9.66200769e-01 1.15997970e+00 7.68818855e-01
-6.85895979e-01 1.53242099e+00 -1.46489069e-01 2.90366173e-01
1.03256118e+00 -5.30518055e-01 4.36676860e-01 7.94979393e-01
-1.95168197e-01 1.00999212e+00 2.47767791e-01 -5.33373773e-01
-1.68806982e+00 -8.77745688e-01 3.72225463e-01 -3.64610553e-01
4.36088294e-01 -9.64375615e-01 -9.34153557e-01 1.84551612e-01
7.60502517e-02 1.01498701e-01 1.03534353e+00 -3.86734009e-01
-7.88026035e-01 -4.68162119e-01 -2.14299250e+00 5.99717677e-01
7.65439987e-01 -5.36458969e-01 -1.37783051e-01 1.44308746e-01
1.27385572e-01 -1.62406728e-01 -1.03100812e+00 4.47955847e-01
7.29020417e-01 -7.73680091e-01 1.14494491e+00 -5.40549040e-01
7.90107921e-02 -1.16419934e-01 -1.60687178e-01 -7.65415728e-01
4.88569468e-01 -9.05770004e-01 -2.08327129e-01 1.90311575e+00
2.48280510e-01 -9.46268737e-01 1.14625251e+00 8.08207512e-01
4.17983174e-01 -4.30431753e-01 -9.73181009e-01 -9.09551978e-01
1.73831563e-02 -2.44306043e-01 1.10118878e+00 1.01568031e+00
-1.44440740e-01 -5.22200048e-01 -6.02737129e-01 2.69679606e-01
7.00625181e-01 -5.58588207e-01 1.21463561e+00 -1.23381901e+00
-2.27068681e-02 8.44437480e-02 -6.94252491e-01 -2.54767425e-02
1.17154919e-01 -5.93900681e-01 -5.40716946e-01 -3.71740550e-01
-4.45340365e-01 -5.04870296e-01 -9.90737453e-02 5.58841825e-01
1.71524510e-01 4.53161061e-01 2.94103980e-01 -1.31996095e-01
3.77179265e-01 1.61626905e-01 7.65887260e-01 -6.31069466e-02
6.43650219e-02 3.82754840e-02 -7.07515776e-01 3.37726474e-01
7.19677925e-01 -3.47858191e-01 -6.88052893e-01 3.08187395e-01
-4.87337381e-01 -1.81225047e-01 1.89986125e-01 -1.14270079e+00
-6.34894893e-03 -2.44435109e-02 5.63984692e-01 -2.97785282e-01
6.28047943e-01 -1.03776765e+00 3.38994175e-01 4.81848836e-01
2.15585455e-02 -8.56562182e-02 5.12863934e-01 3.20662670e-02
-1.59465134e-01 -8.67575109e-02 8.28732073e-01 3.85715753e-01
-6.02738142e-01 5.70741475e-01 -5.10067865e-02 -3.29835601e-02
1.48527670e+00 -6.04988158e-01 -7.39415646e-01 -1.49148434e-01
-3.61383706e-01 -1.84250638e-01 4.04997468e-01 5.79562843e-01
8.35861504e-01 -1.31119084e+00 -7.30111241e-01 1.20453167e+00
2.54421264e-01 -8.21664870e-01 3.22812438e-01 9.15227756e-02
-4.70627666e-01 6.81259781e-02 -9.60834324e-01 -2.50006914e-01
-1.70949221e+00 6.77744865e-01 1.59152806e-01 1.67252466e-01
-3.90594989e-01 4.63704139e-01 -2.29203224e-01 -2.74389714e-01
3.08073074e-01 4.91852432e-01 -1.71311915e-01 1.62642986e-01
1.00952923e+00 2.27940962e-01 2.36220285e-01 -4.62620318e-01
-4.86375868e-01 2.45755911e-01 -3.44264776e-01 -1.96801126e-01
1.12437594e+00 4.68013942e-01 -2.22251583e-02 -6.07973754e-01
1.13424158e+00 4.94265795e-01 -1.03399980e+00 7.17279390e-02
-1.55844749e-03 -1.07219315e+00 -6.60267353e-01 -8.36292982e-01
-1.13621366e+00 8.05894554e-01 1.00015664e+00 9.97808203e-02
1.21049976e+00 -6.02359831e-01 6.43851757e-01 3.18089992e-01
6.86681747e-01 -8.34445179e-01 2.35399321e-01 7.03527778e-02
8.91960025e-01 -8.01269412e-01 -3.01086307e-01 -8.55721474e-01
-5.70539653e-01 1.19780219e+00 8.15230489e-01 7.97590613e-02
7.52616823e-01 6.94488466e-01 2.07788140e-01 8.22178200e-02
-4.83601362e-01 5.08242905e-01 -7.86437020e-02 1.38391650e+00
1.86564233e-02 -4.50274736e-01 -3.85288775e-01 9.48936582e-01
-3.37200135e-01 1.54315621e-01 5.00837266e-01 1.01136875e+00
1.39418855e-01 -1.79199255e+00 -6.72552407e-01 2.54338771e-01
-6.83825314e-01 1.66613370e-01 -6.19485855e-01 9.74190891e-01
1.58787087e-01 8.87141645e-01 -3.72892886e-01 -5.50517023e-01
4.65772718e-01 2.56407201e-01 3.55531543e-01 -5.52283004e-02
-8.78994346e-01 -5.82810521e-01 2.13460207e-01 -3.00578684e-01
2.22373441e-01 -5.24479508e-01 -6.93425775e-01 -8.25376689e-01
7.13211894e-02 2.64477611e-01 7.88887084e-01 3.33472103e-01
5.72607696e-01 -8.68418068e-02 1.11289668e+00 -4.74209458e-01
-9.36655343e-01 -6.91721022e-01 -5.37272453e-01 5.98617792e-01
-1.20805301e-01 -7.26384401e-01 -2.71179259e-01 -1.57422200e-02] | [13.039863586425781, 1.0860024690628052] |
017eb0be-5fc8-44c7-8000-da3849816cc5 | deep-word-embeddings-for-visual-speech | 1710.11201 | null | http://arxiv.org/abs/1710.11201v1 | http://arxiv.org/pdf/1710.11201v1.pdf | Deep word embeddings for visual speech recognition | In this paper we present a deep learning architecture for extracting word
embeddings for visual speech recognition. The embeddings summarize the
information of the mouth region that is relevant to the problem of word
recognition, while suppressing other types of variability such as speaker, pose
and illumination. The system is comprised of a spatiotemporal convolutional
layer, a Residual Network and bidirectional LSTMs and is trained on the
Lipreading in-the-wild database. We first show that the proposed architecture
goes beyond state-of-the-art on closed-set word identification, by attaining
11.92% error rate on a vocabulary of 500 words. We then examine the capacity of
the embeddings in modelling words unseen during training. We deploy
Probabilistic Linear Discriminant Analysis (PLDA) to model the embeddings and
perform low-shot learning experiments on words unseen during training. The
experiments demonstrate that word-level visual speech recognition is feasible
even in cases where the target words are not included in the training set. | ['Georgios Tzimiropoulos', 'Themos Stafylakis'] | 2017-10-30 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 2.43284963e-02 1.47000933e-02 -5.38542807e-01 -1.15342595e-01
-7.85231531e-01 -1.82232276e-01 5.06224394e-01 -4.25126463e-01
-4.97857183e-01 2.42838323e-01 2.81586230e-01 -3.79533738e-01
4.30614293e-01 -1.08419038e-01 -5.32648861e-01 -9.60291207e-01
1.06051952e-01 -1.13910072e-01 1.18555818e-02 2.12459385e-01
1.43169612e-01 6.98624194e-01 -2.04327726e+00 1.79932222e-01
2.60108650e-01 1.24608076e+00 3.36430579e-01 7.54277885e-01
-3.38096857e-01 4.82986957e-01 -7.72237003e-01 -2.40092399e-04
-8.50814134e-02 -1.23251222e-01 -2.82417834e-01 2.00973719e-01
5.20728886e-01 -5.34892082e-01 -6.22595429e-01 8.02532673e-01
9.63635623e-01 3.27798605e-01 7.71016598e-01 -1.08306491e+00
-1.24020076e+00 -1.49735928e-01 -2.17508897e-01 3.14268887e-01
3.61918993e-02 2.55158842e-01 8.41747820e-01 -1.65805113e+00
3.86610270e-01 1.25033998e+00 3.08308542e-01 1.13107502e+00
-9.30340827e-01 -5.79951286e-01 1.31826878e-01 4.00702804e-01
-1.47317362e+00 -1.37406349e+00 8.39432597e-01 -3.14894497e-01
1.43019485e+00 7.44266659e-02 3.02612424e-01 1.78266668e+00
5.34074754e-03 9.48941112e-01 6.75803602e-01 -7.50302970e-01
3.25617671e-01 3.47573668e-01 3.08633775e-01 6.10900581e-01
4.90088835e-02 3.74486983e-01 -1.00728726e+00 1.14644347e-02
3.79336268e-01 -1.54174864e-01 -3.17502528e-01 -5.82012236e-01
-8.08873177e-01 8.43716323e-01 2.16211468e-01 3.22480649e-01
-2.60411620e-01 1.09375335e-01 3.34972680e-01 -6.69843629e-02
6.32543027e-01 -1.96484804e-01 -1.56475425e-01 6.53039068e-02
-9.08055961e-01 -5.67208529e-01 6.03529453e-01 8.61975551e-01
4.68308032e-01 5.01421511e-01 -2.32976362e-01 1.32275856e+00
7.51705647e-01 7.45278358e-01 1.07630801e+00 -5.63869476e-01
2.34705023e-02 1.01624383e-02 -9.23935696e-02 -5.69310963e-01
1.22985439e-02 -6.43272251e-02 -3.59473348e-01 3.57292771e-01
7.02543482e-02 3.55952866e-02 -1.61389971e+00 1.69380605e+00
6.21407144e-02 2.65142471e-01 2.63158530e-01 8.34654748e-01
1.29296434e+00 6.78009868e-01 2.13104948e-01 -3.43875736e-01
1.44756567e+00 -1.06401730e+00 -1.03126991e+00 -5.72732687e-01
2.48889029e-01 -7.11626828e-01 9.72949386e-01 2.33734008e-02
-7.61696398e-01 -6.76839769e-01 -1.04496527e+00 -3.87735307e-01
-7.47484863e-01 4.15136576e-01 9.86592993e-02 8.33553851e-01
-1.30270088e+00 -8.81452337e-02 -7.61041164e-01 -5.89883208e-01
4.31393236e-01 3.51571500e-01 -3.89943570e-01 -2.81236297e-03
-1.10467815e+00 9.96861219e-01 -1.44505367e-01 1.04541801e-01
-1.31769764e+00 -4.67429787e-01 -1.30134416e+00 1.15497433e-01
-3.86763327e-02 -1.96680591e-01 1.06965780e+00 -8.29389989e-01
-1.71457243e+00 1.12257779e+00 -9.47254777e-01 -1.74900979e-01
-1.90702919e-02 1.04755126e-01 -6.53244257e-01 2.23903269e-01
-1.85492262e-01 9.28348005e-01 1.52743268e+00 -1.15399277e+00
-9.21910405e-02 -3.67706537e-01 -4.88311917e-01 -9.36301798e-02
-6.40400589e-01 2.15793863e-01 -6.33533597e-01 -5.82044721e-01
-1.97967574e-01 -6.69138670e-01 4.17112291e-01 4.08494294e-01
-1.87648669e-01 -5.24137080e-01 1.10664439e+00 -6.90242648e-01
8.54795814e-01 -2.63617110e+00 -1.96301758e-01 -2.74754912e-01
-6.58008680e-02 7.41417825e-01 -5.40461719e-01 1.68621570e-01
-2.08888680e-01 1.22911520e-01 1.64584950e-01 -1.10422945e+00
8.92964303e-02 3.87810796e-01 -4.60206866e-01 7.20382154e-01
3.60425830e-01 9.64259207e-01 -3.95934522e-01 -4.95889902e-01
5.47201216e-01 1.01986182e+00 -1.96814001e-01 4.19353127e-01
4.47312370e-02 -2.15688884e-01 -3.91244553e-02 9.62587655e-01
6.07860386e-01 1.37507811e-01 -1.79161429e-01 -2.80620039e-01
3.53904180e-02 3.17060053e-01 -7.16021359e-01 1.62900198e+00
-6.39606893e-01 1.19835865e+00 3.89545470e-01 -9.00015950e-01
1.01197207e+00 5.39598405e-01 2.35904176e-02 -7.00976014e-01
2.39290655e-01 7.22355768e-02 -1.65851921e-01 -8.29492033e-01
2.62999147e-01 -1.69374362e-01 4.44848001e-01 3.32907587e-01
5.77106893e-01 3.10738981e-01 -3.98972720e-01 -2.24502683e-01
5.89185655e-01 -3.03362548e-01 1.57412499e-01 -9.82910246e-02
4.84516323e-01 -5.83685458e-01 1.89758793e-01 3.49575847e-01
-7.25324869e-01 5.97589910e-01 1.48312226e-01 -1.85392022e-01
-8.89106572e-01 -1.15557778e+00 -4.23520595e-01 1.25922763e+00
1.10422753e-01 -8.31198692e-02 -7.21520305e-01 -4.02668148e-01
-8.96531716e-02 8.84777844e-01 -8.10276270e-01 -3.01906437e-01
4.97998856e-02 -3.45543236e-01 5.57246149e-01 5.35873234e-01
3.21421772e-02 -1.31538010e+00 -4.84789133e-01 -2.33731940e-01
7.97122419e-02 -1.36513376e+00 -7.34422266e-01 2.05093741e-01
-3.20054352e-01 -8.52122784e-01 -9.41420734e-01 -1.30098546e+00
6.03352547e-01 4.64372456e-01 5.35102129e-01 -1.59057111e-01
-6.16837502e-01 7.18577206e-01 -1.61049098e-01 -5.55705786e-01
-3.47753197e-01 -1.85605675e-01 3.58046860e-01 3.03755552e-01
9.51722264e-01 -2.51425356e-01 -4.71154094e-01 3.14878285e-01
-5.71604133e-01 -6.45918012e-01 3.51409018e-01 1.11689126e+00
4.03533816e-01 -5.61565578e-01 5.69144547e-01 2.27376580e-01
6.62697375e-01 -2.84601092e-01 -3.16095173e-01 1.99869558e-01
-3.20388287e-01 4.40121479e-02 6.93714172e-02 -9.47277129e-01
-7.74083376e-01 7.51218721e-02 -2.66939253e-01 -9.56424356e-01
-4.47634041e-01 -1.70364857e-01 -3.05895925e-01 -1.81602836e-01
3.91420603e-01 5.37884057e-01 5.75272620e-01 -6.34922504e-01
3.82370323e-01 1.38867056e+00 3.09940785e-01 5.48051633e-02
3.21176469e-01 5.02985716e-01 -5.37869036e-01 -1.65046728e+00
-4.58148032e-01 -6.74886823e-01 -5.86230874e-01 -1.49943292e-01
1.05009580e+00 -9.85411644e-01 -7.20113754e-01 5.85865915e-01
-1.40182602e+00 -1.99119523e-01 -2.54189253e-01 5.63765943e-01
-4.37543124e-01 3.02343071e-01 -3.49111229e-01 -1.32034254e+00
-3.45538765e-01 -1.39253902e+00 1.26551557e+00 8.71216431e-02
1.02778971e-02 -9.97213960e-01 -1.28450757e-02 2.92983353e-01
4.80632335e-01 -4.86619771e-01 6.39728427e-01 -1.01752055e+00
-1.67065307e-01 -1.33306995e-01 -1.10729784e-01 8.17459643e-01
3.57390076e-01 -2.37937775e-02 -1.82510519e+00 -2.88592339e-01
1.83566630e-01 -4.31761354e-01 1.15516949e+00 5.60614109e-01
9.67796028e-01 -1.71123296e-01 -3.76993656e-01 6.18108332e-01
1.04264593e+00 3.13837409e-01 6.52179718e-01 -1.25880256e-01
5.00609636e-01 7.03439653e-01 1.04617052e-01 3.01673673e-02
1.07101969e-01 6.97324336e-01 5.37556827e-01 -2.88462818e-01
-6.46615028e-01 -3.02935481e-01 6.93521202e-01 6.96975529e-01
5.77136099e-01 -3.35782170e-01 -6.35312319e-01 1.06825948e+00
-1.26225281e+00 -8.33129644e-01 5.60869515e-01 2.04441071e+00
4.94260728e-01 -2.35082537e-01 5.81677966e-02 4.13614213e-02
9.35587227e-01 5.78986466e-01 -6.62529707e-01 -8.36351991e-01
-8.80834907e-02 3.59242052e-01 2.35608220e-01 8.96476567e-01
-1.01968074e+00 1.17474365e+00 6.79379511e+00 7.63437271e-01
-1.45409393e+00 3.46359104e-01 4.37758863e-01 -3.02166313e-01
-9.14306045e-02 -6.19739711e-01 -9.09723520e-01 3.53956342e-01
1.09619582e+00 2.70767719e-01 3.26060712e-01 1.00442755e+00
2.45612979e-01 2.71292508e-01 -9.95643437e-01 1.24923205e+00
9.53639090e-01 -1.02111316e+00 1.46430042e-02 2.04249099e-01
1.64324820e-01 3.14618945e-01 4.34329063e-01 2.57845670e-01
-3.23895067e-01 -1.28367937e+00 4.83026445e-01 2.78016418e-01
1.18936694e+00 -4.04082000e-01 3.59087110e-01 3.20327133e-01
-8.66320968e-01 -8.24308842e-02 -5.24789989e-01 3.24490875e-01
9.63998493e-03 5.54849245e-02 -9.83530343e-01 -3.57432753e-01
4.98954028e-01 4.43069428e-01 -2.09918052e-01 6.70238793e-01
-2.54212558e-01 5.14760077e-01 -2.38448426e-01 -3.08078974e-01
1.45279512e-01 4.36027646e-01 5.31908572e-01 1.22013938e+00
2.98872381e-01 -9.82685909e-02 -2.54101783e-01 8.21905673e-01
-2.15984523e-01 6.07230254e-02 -9.59959388e-01 -2.55284578e-01
4.62466776e-01 9.30904269e-01 -2.70779561e-02 -1.02829143e-01
-5.55564880e-01 1.02309382e+00 1.67949677e-01 8.19628417e-01
-5.78105152e-01 -4.07367557e-01 1.29062533e+00 -1.30048856e-01
8.97836208e-01 -3.51307213e-01 1.09657221e-01 -1.03818762e+00
9.73253325e-02 -5.85120499e-01 -1.23228416e-01 -8.66432726e-01
-1.26219535e+00 6.61718369e-01 -3.88073236e-01 -8.77406299e-01
-2.95455456e-01 -1.23558187e+00 -6.18967056e-01 1.11780846e+00
-1.96478665e+00 -1.14304328e+00 -1.03588186e-01 6.25199676e-01
9.98831332e-01 -5.14117479e-01 1.25392163e+00 1.43759146e-01
-6.79769039e-01 9.47739661e-01 1.27671182e-01 2.82449484e-01
7.36862719e-01 -6.62050247e-01 4.25238669e-01 7.57961094e-01
4.27316129e-01 6.49451435e-01 3.59359771e-01 -1.67977437e-01
-1.52445197e+00 -7.63952971e-01 1.13586974e+00 -2.42089614e-01
6.90014482e-01 -8.88918996e-01 -8.71340394e-01 5.66041231e-01
3.34605068e-01 5.80499530e-01 9.62950945e-01 -1.52914658e-01
-6.23058558e-01 -1.25015363e-01 -1.07534480e+00 3.93899739e-01
7.68541694e-01 -1.20331323e+00 -8.58408689e-01 1.86102197e-01
8.39844406e-01 -1.14803419e-01 -3.49794835e-01 3.41173969e-02
7.14876473e-01 -5.37486613e-01 1.21280098e+00 -8.75922441e-01
-1.31193832e-01 1.28075453e-02 -4.62743640e-01 -1.15508270e+00
8.62511694e-02 -6.10239208e-01 -3.26405525e-01 1.23446763e+00
5.09446025e-01 -6.68289244e-01 6.45770252e-01 4.36107844e-01
-1.32466368e-02 -5.89912295e-01 -1.52889717e+00 -9.52134907e-01
7.81759396e-02 -6.19946420e-01 1.35354802e-01 5.46724975e-01
9.45830345e-02 3.59362572e-01 -4.26630646e-01 2.95694560e-01
6.27606511e-01 -2.57911414e-01 3.59188557e-01 -8.37730110e-01
1.62543297e-01 -1.88469186e-01 -5.22678554e-01 -1.25017297e+00
8.03772986e-01 -5.94759285e-01 4.53406781e-01 -1.42643726e+00
-1.36934265e-01 1.18193381e-01 -4.36050177e-01 5.22679865e-01
1.12566747e-01 1.93517193e-01 4.17296775e-02 5.51486276e-02
-1.93372712e-01 8.35728884e-01 9.68648970e-01 -5.40640831e-01
-1.29306894e-02 -2.50616759e-01 -3.21183711e-01 5.09590864e-01
5.37702739e-01 -2.10691378e-01 -4.31149691e-01 -5.05816460e-01
-8.97269607e-01 -3.13186675e-01 4.14552301e-01 -5.32007456e-01
4.90542591e-01 9.58513990e-02 2.26741180e-01 -4.08438385e-01
1.00739145e+00 -7.99717069e-01 -6.21352732e-01 2.61906207e-01
-3.89748126e-01 -4.84767199e-01 5.64320862e-01 5.00345111e-01
-2.20624343e-01 -2.58487105e-01 9.30315137e-01 3.43573064e-01
-8.93625140e-01 2.19502732e-01 -4.99762505e-01 -7.54779726e-02
1.13195431e+00 -3.21986169e-01 -4.44079518e-01 -3.18702847e-01
-1.00776458e+00 -2.38358647e-01 3.04551065e-01 7.40433574e-01
9.82103527e-01 -1.24637294e+00 -4.25542831e-01 7.39602566e-01
3.59802634e-01 -6.70641541e-01 1.55809850e-01 5.03073454e-01
1.79812647e-02 6.76862061e-01 1.22892283e-01 -6.60590827e-01
-1.41079009e+00 9.43906784e-01 5.83936095e-01 6.38121486e-01
-4.56809640e-01 1.07885575e+00 2.67562151e-01 -5.79149288e-04
8.64177346e-01 -3.03231835e-01 -1.85522869e-01 3.58999550e-01
8.05766821e-01 3.02654896e-02 8.87017101e-02 -9.55064535e-01
-7.97680259e-01 6.52049422e-01 5.95683604e-02 -1.37670472e-01
1.15555191e+00 -2.22383112e-01 3.26802462e-01 6.49158597e-01
1.71727252e+00 -6.33758083e-02 -1.09743357e+00 -2.74400592e-01
-3.98136050e-01 -3.21585625e-01 4.74178582e-01 -5.72851121e-01
-9.41193879e-01 1.57236981e+00 1.15147495e+00 4.62499075e-02
8.65946233e-01 2.38380119e-01 6.96321070e-01 2.62023687e-01
-1.52713895e-01 -1.06845856e+00 2.46749118e-01 3.46065342e-01
9.80355263e-01 -1.50626874e+00 -5.68456233e-01 -1.16320215e-01
-6.55015826e-01 1.08280146e+00 2.18046769e-01 4.13073562e-02
9.33506906e-01 3.15339416e-01 4.05865312e-01 1.85672212e-02
-8.80046487e-01 -5.05079925e-01 4.79128867e-01 1.05625308e+00
1.89541355e-01 -1.32827595e-01 3.15105498e-01 3.43669742e-01
3.05813521e-01 -2.81247318e-01 4.56037633e-02 5.00427008e-01
-5.62172830e-01 -7.63269067e-01 -2.55943030e-01 -7.32805431e-02
-3.03546727e-01 -4.84581023e-01 -5.30559421e-01 6.74248636e-01
-1.21142946e-01 1.16420424e+00 4.15711582e-01 -3.08945924e-01
2.31176481e-01 6.20168567e-01 1.98588327e-01 -6.97786152e-01
6.88886568e-02 1.64574355e-01 5.30861877e-02 -5.47992647e-01
-2.41973370e-01 -5.61681449e-01 -9.10052240e-01 1.38631538e-01
-4.00698811e-01 -1.59763411e-01 1.13520586e+00 8.76931369e-01
5.44628024e-01 4.18126941e-01 5.38683355e-01 -9.78085697e-01
-5.91719091e-01 -1.12361431e+00 -6.01071775e-01 9.18390416e-03
1.08651018e+00 -9.33303297e-01 -9.77060318e-01 6.92638084e-02] | [14.276989936828613, 4.967987537384033] |
a42368c7-8201-4aea-964a-47c7e333bb9f | ebms-vs-cl-exploring-self-supervised-visual | 2206.14355 | null | https://arxiv.org/abs/2206.14355v1 | https://arxiv.org/pdf/2206.14355v1.pdf | EBMs vs. CL: Exploring Self-Supervised Visual Pretraining for Visual Question Answering | The availability of clean and diverse labeled data is a major roadblock for training models on complex tasks such as visual question answering (VQA). The extensive work on large vision-and-language models has shown that self-supervised learning is effective for pretraining multimodal interactions. In this technical report, we focus on visual representations. We review and evaluate self-supervised methods to leverage unlabeled images and pretrain a model, which we then fine-tune on a custom VQA task that allows controlled evaluation and diagnosis. We compare energy-based models (EBMs) with contrastive learning (CL). While EBMs are growing in popularity, they lack an evaluation on downstream tasks. We find that both EBMs and CL can learn representations from unlabeled images that enable training a VQA model on very little annotated data. In a simple setting similar to CLEVR, we find that CL representations also improve systematic generalization, and even match the performance of representations from a larger, supervised, ImageNet-pretrained model. However, we find EBMs to be difficult to train because of instabilities and high variability in their results. Although EBMs prove useful for OOD detection, other results on supervised energy-based training and uncertainty calibration are largely negative. Overall, CL currently seems a preferable option over EBMs. | ['Damien Teney', 'Anton Van Den Hengel', 'Anthony Dick', 'Ehsan Abbasnejad', 'Violetta Shevchenko'] | 2022-06-29 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [-1.87921468e-02 2.41521806e-01 -2.60972053e-01 -5.94242573e-01
-1.08013666e+00 -7.00683475e-01 7.19160557e-01 1.73071817e-01
-5.69736838e-01 6.41893983e-01 2.17273772e-01 -2.47343928e-01
2.71162480e-01 -4.17963654e-01 -9.92368758e-01 -5.88805556e-01
1.77209124e-01 6.31912529e-01 1.74660549e-01 -3.85988206e-02
-2.24254876e-01 3.32635671e-01 -1.72246861e+00 2.60923564e-01
7.67570138e-01 9.70563829e-01 2.06283301e-01 8.13361406e-01
-2.81428397e-01 1.07823312e+00 -4.91795301e-01 -3.29192370e-01
-1.00050338e-01 -5.03391027e-01 -9.39478755e-01 -2.61089369e-03
8.57343435e-01 -3.94341111e-01 -2.55887359e-01 6.23060942e-01
6.44989252e-01 3.75287652e-01 1.01769829e+00 -1.42928350e+00
-8.56420338e-01 1.89243451e-01 -4.43804502e-01 4.37366396e-01
1.22748978e-01 5.40659070e-01 1.03323078e+00 -9.63906050e-01
8.64338040e-01 1.31022382e+00 5.80628991e-01 1.03785479e+00
-1.38013875e+00 -3.40314329e-01 1.82385296e-01 2.19649926e-01
-1.01234281e+00 -7.88363397e-01 5.19389272e-01 -4.12563145e-01
1.32832646e+00 8.87910873e-02 3.38260263e-01 1.42208087e+00
-2.22064517e-02 1.05530548e+00 1.01186216e+00 -6.62167192e-01
3.39065492e-01 5.87563634e-01 1.11571588e-01 9.86449063e-01
3.13622504e-02 3.19818348e-01 -6.51213884e-01 -2.55258791e-02
3.60614419e-01 -4.76609856e-01 -2.19993398e-01 -7.42720485e-01
-9.49985445e-01 9.24172044e-01 7.60932922e-01 2.01358229e-01
-1.01561226e-01 6.29019141e-01 3.40539724e-01 1.81518525e-01
4.14685071e-01 4.26183194e-01 -2.89673388e-01 9.95788816e-03
-1.02661192e+00 -4.42665294e-02 5.78728735e-01 7.42167294e-01
7.66546547e-01 2.91116744e-01 -4.58189815e-01 9.15327430e-01
5.30588031e-01 5.62056899e-01 3.52046937e-01 -1.18246591e+00
1.02892831e-01 3.12996864e-01 -1.25355542e-01 -5.59035182e-01
-4.33446676e-01 -1.64651826e-01 -4.80442137e-01 6.93875253e-01
4.86783445e-01 -1.78519040e-01 -1.53511429e+00 1.83665371e+00
6.85755908e-02 -2.30411798e-01 1.48681641e-01 1.00038421e+00
1.40263879e+00 6.75011814e-01 6.10846937e-01 6.71847397e-03
9.54141319e-01 -1.16006219e+00 -6.72476411e-01 -5.66375911e-01
4.79300141e-01 -3.78371656e-01 1.20550263e+00 3.36296350e-01
-1.22170901e+00 -5.20078182e-01 -1.09697390e+00 -4.08663601e-01
-3.78206104e-01 1.51915833e-01 8.04225922e-01 7.88958907e-01
-1.34408700e+00 3.84948909e-01 -8.74761343e-01 -5.59544027e-01
6.74695373e-01 2.56908685e-01 -3.05091083e-01 -8.08600187e-02
-8.90351057e-01 1.26323926e+00 2.00994000e-01 -1.44647852e-01
-1.44512594e+00 -5.19735694e-01 -1.12349784e+00 -6.97314665e-02
3.30703735e-01 -7.56758690e-01 1.40150714e+00 -1.24021423e+00
-1.22507966e+00 1.23232555e+00 -1.77375704e-01 -4.94691789e-01
2.70682752e-01 1.17869243e-01 -1.67658925e-01 3.96512061e-01
-5.19674607e-02 1.52026439e+00 1.08324039e+00 -1.72285652e+00
-2.11960211e-01 -1.39794186e-01 2.26484686e-01 4.43374455e-01
-1.77328870e-01 -2.46094301e-01 -5.97596288e-01 -2.82207429e-01
-3.84751946e-01 -6.51260912e-01 -1.00492448e-01 2.91606963e-01
-2.23780647e-01 -2.65291005e-01 6.38282478e-01 -5.77297211e-01
7.09432900e-01 -1.92685866e+00 1.40142113e-01 1.07996166e-01
1.74707651e-01 1.27069131e-01 -2.55399227e-01 1.00682722e-02
-1.94682740e-02 9.92083848e-02 -3.36027771e-01 -5.53065419e-01
1.41820614e-03 4.47051525e-01 -9.88764688e-02 2.65046358e-01
6.07959032e-01 1.51242864e+00 -8.05640876e-01 -8.18776786e-01
4.05530691e-01 5.61207652e-01 -5.56410670e-01 3.43042403e-01
-5.78365445e-01 3.94627094e-01 4.53126989e-02 9.96039450e-01
2.61777759e-01 -6.62124455e-01 -5.16708083e-02 -3.61752778e-01
2.93082327e-01 -1.56921502e-02 -6.92536652e-01 1.85452056e+00
-5.61750948e-01 1.04025781e+00 2.82561630e-01 -1.05652249e+00
5.12430906e-01 1.12909213e-01 1.69924840e-01 -9.96197939e-01
2.46688232e-01 -5.20179532e-02 -1.43267006e-01 -6.40233874e-01
2.95169622e-01 -1.96161360e-01 2.74240941e-01 4.03577864e-01
8.17357659e-01 -3.25550735e-01 4.86047342e-02 5.04424751e-01
8.87588501e-01 4.10828263e-01 -2.86945850e-02 1.97430383e-02
7.54431039e-02 3.41425717e-01 -6.48482330e-03 7.75169373e-01
-4.02732790e-01 7.70186961e-01 1.09462664e-01 7.09719062e-02
-9.84091878e-01 -1.20585644e+00 -2.03974113e-01 1.24654341e+00
2.81374365e-01 -2.59458721e-01 -6.63450778e-01 -9.59439456e-01
-1.83726493e-02 8.89736652e-01 -5.90732872e-01 -2.28675127e-01
-7.99621493e-02 -8.28845680e-01 4.55115557e-01 8.35604608e-01
3.61181408e-01 -1.21495354e+00 -6.82654142e-01 -1.34985104e-01
-2.35052183e-02 -9.76964474e-01 -2.69761205e-01 5.91555238e-01
-7.39678681e-01 -9.37243462e-01 -8.36494803e-01 -7.89188981e-01
6.54522359e-01 7.33543783e-02 1.56800854e+00 5.67255989e-02
-5.77184021e-01 1.31129813e+00 -8.82201567e-02 -5.80749333e-01
-4.58767146e-01 -1.53223068e-01 -1.68987453e-01 -3.38501781e-01
4.00333703e-01 -9.69351307e-02 -6.68503106e-01 1.36479661e-01
-8.31159234e-01 -3.07980388e-01 5.50906003e-01 1.03490877e+00
7.90201783e-01 -6.07875466e-01 4.36544091e-01 -6.92571461e-01
5.68044424e-01 -4.57327664e-01 -4.63841736e-01 4.28867817e-01
-8.48192692e-01 4.64698523e-01 -3.64661217e-02 -4.25554365e-01
-1.29741096e+00 2.57951945e-01 -2.22608924e-01 -7.49001324e-01
-1.97225392e-01 3.13149720e-01 5.27901687e-02 -3.08921784e-01
1.19932115e+00 -2.78634548e-01 1.05067022e-01 -1.81774408e-01
7.33892858e-01 4.03227657e-01 6.36141956e-01 -4.09397751e-01
6.95755601e-01 4.36599016e-01 -2.99207270e-01 -6.03871405e-01
-7.69928575e-01 -3.32228512e-01 -1.25034064e-01 -3.47167611e-01
1.26885211e+00 -9.98217225e-01 -6.53418362e-01 -6.14529327e-02
-8.82278502e-01 -7.86001325e-01 -6.00906074e-01 3.67693514e-01
-5.53576589e-01 3.45879853e-01 -4.09646958e-01 -9.92932916e-01
-3.17221314e-01 -1.05215251e+00 1.14794207e+00 4.56426054e-01
-8.94717425e-02 -1.32350159e+00 2.46356547e-01 5.71164072e-01
4.23875093e-01 8.41214880e-02 8.41205060e-01 -5.35450399e-01
-5.77175260e-01 2.54609913e-01 -2.88438797e-01 5.92691660e-01
-1.49796844e-01 5.38646579e-02 -1.53537345e+00 -1.83790445e-01
-3.64469409e-01 -1.22386134e+00 1.32549751e+00 5.91997683e-01
1.09312642e+00 3.06201100e-01 -3.81265432e-01 5.03590763e-01
1.32427537e+00 -8.95789266e-02 4.30996180e-01 1.62209168e-01
7.11176395e-01 6.11910701e-01 4.87512618e-01 -2.96143353e-01
4.85644102e-01 5.49718857e-01 6.60562694e-01 -4.45196748e-01
-5.56567729e-01 -2.37802312e-01 5.21620631e-01 2.83120662e-01
-1.77519321e-02 -4.11157668e-01 -1.09859586e+00 5.63627660e-01
-1.63373625e+00 -8.29217136e-01 6.48205653e-02 1.99386311e+00
7.04664826e-01 6.49545491e-02 4.14362215e-02 -2.91406542e-01
4.25482988e-01 1.02442063e-01 -8.80842507e-01 -5.27752399e-01
-1.80083126e-01 3.96750033e-01 5.29921889e-01 6.74988270e-01
-1.20579004e+00 8.29595685e-01 7.53587437e+00 6.43877685e-01
-8.84037256e-01 3.48604500e-01 8.64085436e-01 -1.68894425e-01
-5.77895582e-01 -2.25625500e-01 -6.59088552e-01 1.98476855e-02
1.12218475e+00 2.81083256e-01 4.75240976e-01 8.89741838e-01
-2.18169436e-01 -4.47466791e-01 -1.43000710e+00 1.38419187e+00
4.08003747e-01 -1.42029226e+00 -7.79747218e-02 -2.20466092e-01
7.83329964e-01 3.43321353e-01 1.21805340e-01 6.42306745e-01
5.28118789e-01 -1.40223658e+00 4.93205488e-01 4.88708019e-01
1.05659425e+00 -4.20063853e-01 5.88101029e-01 1.32952988e-01
-7.29809225e-01 1.86104290e-02 -3.85365635e-01 4.97875035e-01
1.06013283e-01 1.98732033e-01 -6.83818877e-01 2.89412916e-01
8.71251583e-01 4.10926431e-01 -8.85385871e-01 8.71695876e-01
-1.83289647e-01 6.22690976e-01 -2.37343416e-01 7.13784993e-02
2.07301989e-01 1.53321669e-01 2.13613823e-01 1.13928485e+00
-6.94616660e-02 -2.40806509e-02 -2.88929008e-02 1.08811831e+00
-2.49585107e-01 -5.07962592e-02 -7.42003381e-01 -2.73851365e-01
7.14577958e-02 1.37551701e+00 -4.37302381e-01 -3.56226772e-01
-5.51584363e-01 1.05715072e+00 5.43363750e-01 5.25197387e-01
-8.34933698e-01 3.20602246e-02 2.35680580e-01 -2.40567893e-01
1.52606145e-01 -1.81526914e-02 -9.68288258e-02 -1.33943808e+00
-2.38388389e-01 -7.87268460e-01 5.33036709e-01 -1.25419343e+00
-1.50553262e+00 5.99322915e-01 1.26107931e-01 -8.79889071e-01
-7.03371882e-01 -8.89130712e-01 -4.82236862e-01 6.51004553e-01
-1.71492267e+00 -1.04301929e+00 -6.06300950e-01 6.86808109e-01
6.11857116e-01 -1.45909354e-01 9.72740948e-01 9.15953740e-02
-3.71284038e-01 6.23001218e-01 -1.44108273e-02 -5.41207828e-02
8.76591325e-01 -1.52906859e+00 7.04609230e-02 5.59186578e-01
6.83972478e-01 1.38626143e-01 5.80924273e-01 -4.44600672e-01
-1.38957810e+00 -7.81008899e-01 2.78065622e-01 -9.51694012e-01
2.71997601e-01 -2.93908119e-01 -9.19614851e-01 7.38691747e-01
6.54826343e-01 2.17907175e-01 5.71674407e-01 -3.90809588e-02
-3.85392100e-01 7.71529004e-02 -1.20442200e+00 2.46959001e-01
1.05617666e+00 -7.88395822e-01 -7.00563252e-01 4.84139413e-01
3.78875494e-01 -4.05233383e-01 -5.89072108e-01 4.23623264e-01
3.22642982e-01 -9.45608735e-01 1.04740620e+00 -6.37933433e-01
2.50503749e-01 -2.38326769e-02 -8.43443051e-02 -1.45850062e+00
1.03237629e-01 -1.08722076e-01 -3.54845345e-01 1.16504371e+00
5.39532065e-01 -1.03539772e-01 8.66478622e-01 8.57354045e-01
1.02086000e-01 -4.59644526e-01 -7.17902958e-01 -6.75623298e-01
3.13561261e-02 -4.69766557e-01 -3.40977460e-01 7.45569527e-01
-2.36359313e-01 7.04803586e-01 -1.61037773e-01 -6.67536333e-02
6.82538748e-01 -1.98059082e-01 5.42370260e-01 -1.11900306e+00
-4.67601359e-01 -2.77968436e-01 -2.50181347e-01 -7.23461449e-01
2.92494118e-01 -1.01620376e+00 3.05092692e-01 -1.85764647e+00
3.71854097e-01 -2.84880310e-01 -2.17602924e-01 6.20880604e-01
-1.66181877e-01 4.92267609e-01 4.55817729e-02 -1.68564450e-02
-6.77380502e-01 6.31841183e-01 9.30474460e-01 -6.58348382e-01
-2.71589786e-01 -4.73130614e-01 -4.67420340e-01 5.51626623e-01
5.94692707e-01 -1.89773217e-01 -7.70119488e-01 -5.84757507e-01
8.59636515e-02 -2.39373207e-01 6.84168518e-01 -7.53677785e-01
3.93586280e-03 9.15842578e-02 7.98037946e-01 -4.14262205e-01
5.57699680e-01 -6.80024445e-01 -2.73025185e-01 1.66430980e-01
-4.89364564e-01 -2.91863084e-01 5.97493231e-01 6.22466743e-01
-2.74618417e-01 -5.21596611e-01 7.07452595e-01 -2.48737752e-01
-1.18561125e+00 1.37911767e-01 -1.82288975e-01 2.82980531e-01
8.84536743e-01 -2.62325499e-02 -4.18476164e-01 -8.25671315e-01
-1.05921710e+00 5.52453279e-01 3.98079604e-01 4.30923969e-01
7.82755911e-01 -1.24263334e+00 -4.46261257e-01 -1.27343878e-01
4.38171744e-01 -9.67611149e-02 1.51113600e-01 6.48498058e-01
-2.72428304e-01 2.60921329e-01 -1.86488226e-01 -9.40812230e-01
-1.20010412e+00 6.91169620e-01 6.59837365e-01 -3.65149193e-02
-3.62741858e-01 1.04391360e+00 3.23093385e-01 -3.38672191e-01
6.38758123e-01 -1.16023973e-01 -2.45409593e-01 2.10098267e-01
3.14621747e-01 2.69672237e-02 1.09333418e-01 -3.68624032e-01
-5.22359550e-01 4.55392569e-01 9.00174007e-02 -4.51101393e-01
1.12391269e+00 -1.95120201e-01 4.47285593e-01 5.75023651e-01
1.08483124e+00 -1.83069810e-01 -1.47823441e+00 1.23717906e-02
-1.73080608e-01 -7.03646019e-02 4.55427021e-01 -1.04894590e+00
-1.11864889e+00 1.21955526e+00 1.15508533e+00 -9.66561735e-02
1.03410470e+00 4.98296469e-01 1.38695955e-01 5.66388726e-01
6.47073165e-02 -1.21799421e+00 5.36379874e-01 9.58431438e-02
8.16591680e-01 -1.91951478e+00 -5.45149408e-02 -6.49234802e-02
-1.20040119e+00 8.66188526e-01 9.30566847e-01 9.96753424e-02
3.79846781e-01 7.95364231e-02 2.89832056e-01 -4.25813735e-01
-8.34826469e-01 -7.17832804e-01 5.82847118e-01 9.29372907e-01
3.84256542e-01 -4.24870014e-01 2.03296617e-01 2.36347988e-01
3.16381067e-01 -5.55225126e-02 2.76153654e-01 7.83911586e-01
-3.03372353e-01 -1.00848782e+00 -2.27045015e-01 4.28955793e-01
-1.47392079e-01 -1.36612266e-01 -5.09432614e-01 9.11038935e-01
1.84981339e-02 1.04069448e+00 2.08059683e-01 -1.92561537e-01
1.42608419e-01 4.68462229e-01 8.87981355e-01 -6.84075058e-01
-5.39361119e-01 -9.23942998e-02 4.11857873e-01 -7.06511617e-01
-7.73465574e-01 -3.87450367e-01 -1.40762138e+00 2.08741650e-01
-3.65917504e-01 -9.62315425e-02 8.33584368e-01 8.36011410e-01
2.28542432e-01 5.50250709e-01 2.60555089e-01 -9.09110308e-01
-4.08734858e-01 -7.40925729e-01 -1.96869537e-01 4.74523902e-01
4.93721038e-01 -7.67780602e-01 -3.62417310e-01 1.78659365e-01] | [10.840490341186523, 1.6965606212615967] |
bad46f28-ba64-41fd-b7f1-07dd6174654e | low-field-magnetic-resonance-image | 2304.13385 | null | https://arxiv.org/abs/2304.13385v1 | https://arxiv.org/pdf/2304.13385v1.pdf | Low-field magnetic resonance image enhancement via stochastic image quality transfer | Low-field (<1T) magnetic resonance imaging (MRI) scanners remain in widespread use in low- and middle-income countries (LMICs) and are commonly used for some applications in higher income countries e.g. for small child patients with obesity, claustrophobia, implants, or tattoos. However, low-field MR images commonly have lower resolution and poorer contrast than images from high field (1.5T, 3T, and above). Here, we present Image Quality Transfer (IQT) to enhance low-field structural MRI by estimating from a low-field image the image we would have obtained from the same subject at high field. Our approach uses (i) a stochastic low-field image simulator as the forward model to capture uncertainty and variation in the contrast of low-field images corresponding to a particular high-field image, and (ii) an anisotropic U-Net variant specifically designed for the IQT inverse problem. We evaluate the proposed algorithm both in simulation and using multi-contrast (T1-weighted, T2-weighted, and fluid attenuated inversion recovery (FLAIR)) clinical low-field MRI data from an LMIC hospital. We show the efficacy of IQT in improving contrast and resolution of low-field MR images. We demonstrate that IQT-enhanced images have potential for enhancing visualisation of anatomical structures and pathological lesions of clinical relevance from the perspective of radiologists. IQT is proved to have capability of boosting the diagnostic value of low-field MRI, especially in low-resource settings. | ['Daniel C. Alexander', 'Delmiro Fernandez-Reyes', 'Judith Helen Cross', 'Ikeoluwa Lagunju', 'David W. Carmichael', 'Biobele J. Brown', 'Lisa Ronan', 'Stefano B. Blumberg', 'Ryutaro Tanno', 'Godwin Ogbole', "Felice D'Arco", 'Matteo Figini', 'Hongxiang Lin'] | 2023-04-26 | null | null | null | null | ['image-enhancement'] | ['computer-vision'] | [ 5.60815334e-01 9.93957892e-02 2.42097899e-01 -1.13267876e-01
-5.98435879e-01 -2.33670831e-01 4.15922612e-01 3.84631567e-02
-6.79874301e-01 7.89418995e-01 1.61829919e-01 -3.67942542e-01
-7.16798723e-01 -5.35655856e-01 -6.60580218e-01 -7.72021234e-01
-6.41951323e-01 7.74256051e-01 4.72198009e-01 1.06734894e-01
1.16359949e-01 4.54193503e-01 -9.65159357e-01 3.23447615e-01
1.04006398e+00 6.05920255e-01 1.03322244e+00 3.68498802e-01
4.19332922e-01 9.98877406e-01 7.77027085e-02 1.62538067e-02
6.95493519e-02 -3.50373894e-01 -1.22727132e+00 7.43187172e-03
1.08391546e-01 -4.64735627e-01 4.13865596e-02 9.70160007e-01
7.46317863e-01 1.73214361e-01 4.97127235e-01 -4.18294698e-01
-2.36534595e-01 5.54580390e-01 -8.74549508e-01 9.27037477e-01
1.29819289e-01 3.05291682e-01 -3.16795051e-01 -7.48199284e-01
9.96161520e-01 7.82550156e-01 5.60417652e-01 3.18762898e-01
-1.19333470e+00 -4.02309179e-01 -4.27205056e-01 2.81010300e-01
-1.08961952e+00 -1.67713836e-01 1.89188778e-01 -5.77756882e-01
8.64221573e-01 1.58586949e-01 4.80037242e-01 4.45555717e-01
8.89969051e-01 1.21036798e-01 1.86800778e+00 -4.08126116e-01
2.61604059e-02 -2.59298503e-01 -4.67876524e-01 4.41360414e-01
2.76693732e-01 1.80403635e-01 -9.00551751e-02 -1.95715025e-01
8.55199695e-01 -2.29244709e-01 -5.63663363e-01 -4.12089676e-01
-1.40307713e+00 4.97431248e-01 3.79985541e-01 9.96447206e-01
-8.88373673e-01 -9.99613926e-02 2.01050162e-01 -1.04572155e-01
5.70996881e-01 1.63741902e-01 -7.38283843e-02 1.04946703e-01
-8.99881601e-01 -1.02880768e-01 8.64926074e-03 2.57681578e-01
1.45083860e-01 1.51440520e-02 -1.60843372e-01 7.28485823e-01
1.05687320e-01 6.74515724e-01 7.85370231e-01 -9.47819173e-01
5.83996214e-02 -3.20374608e-01 -8.70607495e-02 -6.79960132e-01
-5.63917935e-01 -6.11941576e-01 -8.67163241e-01 6.46730512e-02
1.54345006e-01 -7.09727481e-02 -9.83045638e-01 1.42759132e+00
5.14089227e-01 2.73777157e-01 -2.35432252e-01 1.33769107e+00
3.84084672e-01 4.24434513e-01 2.91535139e-01 -5.88897347e-01
1.16231787e+00 -5.11400700e-01 -8.37765634e-01 -2.78551430e-01
7.06984401e-01 -7.42947996e-01 4.74327952e-01 2.24624008e-01
-1.67522109e+00 -7.65079781e-02 -7.72897840e-01 7.00065613e-01
3.03455681e-01 -5.60438812e-01 3.57267410e-01 6.52638912e-01
-1.13577294e+00 8.63025546e-01 -1.07778192e+00 -3.65645774e-02
3.18276256e-01 4.87431645e-01 -6.05342627e-01 -5.87917686e-01
-1.38444543e+00 1.47511625e+00 1.90220401e-01 1.00260369e-01
-6.82673216e-01 -1.27573144e+00 -6.97965801e-01 -2.84640342e-01
3.37572545e-01 -3.71584445e-01 9.18806612e-01 -7.38834441e-01
-1.10663426e+00 8.64896834e-01 1.55872703e-01 -4.69700485e-01
5.34995735e-01 2.68634975e-01 -5.40255904e-01 7.69184828e-01
4.83068436e-01 4.21942353e-01 7.45477438e-01 -1.15462065e+00
-8.82353336e-02 -6.68734968e-01 -4.49701637e-01 2.20661640e-01
6.24409735e-01 5.35171151e-01 3.16598862e-01 -4.01587784e-01
3.50956112e-01 -9.27572727e-01 -3.53863597e-01 -5.18512070e-01
1.04186796e-01 8.11986029e-01 6.05747044e-01 -1.28355181e+00
7.66877770e-01 -1.58329260e+00 -1.97415292e-01 4.90577459e-01
3.10574144e-01 2.70793080e-01 2.07299963e-01 -2.64310181e-01
-5.03449917e-01 -1.31424338e-01 -6.09138370e-01 8.24971855e-01
-7.08397508e-01 -5.93480952e-02 5.10932326e-01 9.24902439e-01
-1.33583605e-01 9.03360784e-01 -1.17258537e+00 -6.48200631e-01
4.89903986e-01 7.68873572e-01 -3.69524896e-01 -6.43800274e-02
5.97018540e-01 1.26727104e+00 -4.34183687e-01 6.12482615e-02
1.11997259e+00 -1.27346233e-01 6.74710035e-01 -3.76698136e-01
-4.55933869e-01 -2.43986830e-01 -9.25387502e-01 1.49174237e+00
-6.79573417e-01 9.66609418e-02 6.02079272e-01 -9.22818899e-01
3.11727971e-01 5.27975857e-01 8.58435929e-01 -1.45367920e+00
2.18896672e-01 4.55777526e-01 6.75140440e-01 -7.54843354e-01
-8.70217904e-02 -9.87507701e-01 6.48447931e-01 5.86064994e-01
-6.95987567e-02 -4.80408579e-01 5.58086112e-02 1.30253315e-01
8.38362396e-01 -5.66573516e-02 -1.06119014e-01 -7.60891318e-01
6.96183145e-01 -4.41540778e-01 1.23994753e-01 7.58061826e-01
-3.73725832e-01 5.63857377e-01 3.68443727e-02 -2.34942853e-01
-1.16911244e+00 -1.00417340e+00 -8.67589414e-01 3.93718898e-01
1.35296416e-02 4.74542439e-01 -7.10006535e-01 -2.77808368e-01
-3.68992448e-01 5.16796768e-01 -4.47912425e-01 5.19911498e-02
-9.61502075e-01 -1.20296872e+00 1.82940960e-01 1.98985472e-01
4.36748445e-01 -1.03497446e+00 -1.00851750e+00 3.25412959e-01
-5.47437370e-01 -1.13853526e+00 -2.74100810e-01 2.69453168e-01
-1.14483488e+00 -6.49114192e-01 -1.49615204e+00 -4.82115835e-01
5.34676731e-01 -4.55408804e-02 1.23951316e+00 -1.16507392e-02
-2.73637474e-01 4.47636575e-01 -1.67708263e-01 -9.76594761e-02
-5.10654986e-01 -3.20864230e-01 1.67869523e-01 -1.74911782e-01
-4.82126266e-01 -6.43251538e-01 -9.39320624e-01 3.95307362e-01
-1.17337573e+00 -4.77563255e-02 6.63696885e-01 8.92620027e-01
6.18757486e-01 1.04711339e-01 3.41534972e-01 -7.59511471e-01
5.44176340e-01 -5.47771394e-01 -2.98857301e-01 2.78053135e-01
-6.04297519e-01 1.01966523e-01 3.39156300e-01 -3.87510389e-01
-1.44234324e+00 -5.38314939e-01 -1.57884911e-01 -5.62468357e-02
-1.33686155e-01 3.72396678e-01 5.10974944e-01 -7.01436520e-01
6.14912271e-01 6.79812729e-02 2.61110336e-01 -2.17899173e-01
-1.77168936e-01 4.67714995e-01 5.46744406e-01 -8.47739875e-01
3.21160048e-01 7.12985277e-01 5.12992680e-01 -7.02668667e-01
-3.19010556e-01 -1.09809548e-01 -6.80586100e-01 -5.25002122e-01
9.18147206e-01 -4.38788921e-01 -6.17430389e-01 3.64225447e-01
-5.67532659e-01 -3.59404892e-01 -1.82231963e-01 1.13181937e+00
-5.34229398e-01 4.98238534e-01 -6.94448113e-01 -5.14728487e-01
-3.53834569e-01 -1.78134310e+00 6.77469492e-01 1.60618797e-01
1.76776037e-01 -1.25671506e+00 -1.03816167e-01 4.36605513e-01
1.00061405e+00 4.84548062e-01 1.08935177e+00 5.21014482e-02
-3.47254068e-01 3.53091896e-01 -4.31815125e-02 2.45887712e-01
-1.87094033e-01 -9.87607539e-01 -7.01392472e-01 -2.85195559e-01
6.07394278e-01 -3.54730248e-01 2.68083930e-01 1.28907418e+00
1.14227068e+00 4.15498853e-01 8.43123998e-03 6.36260569e-01
1.44176793e+00 4.78217214e-01 7.37146258e-01 4.77291256e-01
2.90191323e-01 8.01114798e-01 5.20298123e-01 1.56488970e-01
2.63500884e-02 6.22373462e-01 2.37564132e-01 -4.10289615e-01
-2.00247362e-01 1.81961834e-01 -3.51695061e-01 6.39297366e-01
-5.57168782e-01 4.90792483e-01 -1.29428172e+00 7.41333008e-01
-1.14024448e+00 -8.60950708e-01 -2.79044241e-01 2.31521153e+00
6.68030918e-01 -1.90733492e-01 -1.37881309e-01 -4.02305983e-02
1.13831341e+00 -2.10448444e-01 -3.52638483e-01 -3.29970539e-01
-9.73300114e-02 5.81807971e-01 7.75632858e-01 6.74552739e-01
-5.84441602e-01 9.80812460e-02 6.19070101e+00 8.14240932e-01
-1.35033762e+00 8.22840929e-01 1.04654527e+00 6.98881522e-02
-2.65248358e-01 -6.59750551e-02 2.89429247e-01 4.59630787e-01
1.10477149e+00 -1.01563081e-01 4.05485183e-01 1.15541533e-01
3.41813296e-01 -7.93829381e-01 -3.61758739e-01 5.89809656e-01
-2.06090569e-01 -1.00453186e+00 -4.82969612e-01 9.51364115e-02
7.26228654e-01 2.09968641e-01 3.73862982e-01 -1.16275862e-01
-3.93247396e-01 -1.29777586e+00 4.48251933e-01 3.50895226e-01
1.44894683e+00 -8.07714701e-01 8.52525830e-01 3.42337310e-01
-6.37804270e-01 3.70282501e-01 1.12034567e-01 2.58964777e-01
5.63723683e-01 6.10598803e-01 -6.84158981e-01 7.60579050e-01
9.26792800e-01 -1.21040635e-01 -2.00758204e-01 8.34053516e-01
4.08789128e-01 3.09539258e-01 -4.14460391e-01 6.57311916e-01
5.40507257e-01 -3.88446957e-01 3.84064704e-01 9.37029481e-01
2.80768275e-01 4.81125653e-01 -1.54137790e-01 5.46285391e-01
3.88899505e-01 8.76354724e-02 -4.87773746e-01 5.33791840e-01
-3.20931971e-01 1.08574855e+00 -1.08657038e+00 -2.60618895e-01
-2.83899337e-01 6.21690571e-01 -1.89093247e-01 2.62147844e-01
-5.93220651e-01 8.33875835e-02 -3.89411211e-01 8.77686620e-01
-8.80183373e-03 1.08081311e-01 -7.87737370e-02 -9.40879345e-01
-1.15670420e-01 -6.98031485e-01 2.36563146e-01 -1.05221057e+00
-8.10765326e-01 7.62413144e-01 7.72272229e-01 -7.54634798e-01
-4.18952465e-01 -2.12713361e-01 -1.93955868e-01 1.34310520e+00
-1.55736494e+00 -7.02856004e-01 1.19443975e-01 4.23014581e-01
-1.03413522e-01 4.70048249e-01 3.21752816e-01 5.43071270e-01
2.56668329e-01 -9.03450549e-02 2.56606579e-01 -3.78492683e-01
4.10076141e-01 -1.04415631e+00 -2.98626035e-01 6.63113654e-01
-8.73619258e-01 6.39728665e-01 5.57959378e-01 -1.01951575e+00
-1.05227447e+00 -6.08494222e-01 6.23955369e-01 5.11534289e-02
5.30912101e-01 2.51218736e-01 -9.13405120e-01 5.33699930e-01
2.49035120e-01 6.30685568e-01 2.93201476e-01 -6.66110218e-01
6.68688059e-01 2.52607644e-01 -2.09055042e+00 5.49700111e-02
6.06891870e-01 -2.36193061e-01 -4.57068324e-01 2.05685318e-01
4.79184724e-02 -8.00652206e-01 -1.36635518e+00 7.66075850e-01
5.51726460e-01 -1.07547069e+00 1.13776553e+00 -1.15317620e-01
4.66929555e-01 9.93875787e-02 2.97735095e-01 -1.44363034e+00
-4.10271317e-01 -3.97917807e-01 3.56146455e-01 4.60175723e-01
-6.25156909e-02 -7.40023851e-01 3.96723390e-01 5.99931180e-01
3.42835039e-02 -6.41746700e-01 -1.12881136e+00 -5.62876463e-01
4.63011771e-01 -3.32940072e-01 2.12078452e-01 9.72792685e-01
-2.66930103e-01 -1.29841939e-01 -2.73129612e-01 6.25588149e-02
1.00809503e+00 -2.86901772e-01 -3.93179625e-01 -9.10170913e-01
-2.57577598e-01 6.72903284e-02 -2.68389314e-01 -2.28660420e-01
1.90819517e-01 -1.15098488e+00 -1.47352889e-01 -1.40107632e+00
5.13553083e-01 -6.80002987e-01 -1.96844995e-01 -1.29174337e-01
-7.59629067e-03 3.50601345e-01 1.81275070e-01 1.04007959e-01
4.10731174e-02 1.18715048e-01 2.05859542e+00 8.05273652e-02
2.90857345e-01 -3.22521538e-01 -2.06793770e-01 6.92114055e-01
3.91490161e-01 -6.61611795e-01 -3.18076581e-01 -3.63075495e-01
-1.06285743e-01 7.88658440e-01 3.51052016e-01 -9.88837421e-01
-1.06943421e-01 -1.08550735e-01 7.48979688e-01 -2.19576299e-01
-1.47998825e-01 -9.49956179e-01 4.53014612e-01 9.04004812e-01
-2.92780008e-02 1.52568534e-01 1.72508463e-01 -3.29245180e-01
9.12120845e-03 -5.75402796e-01 1.45230377e+00 -7.75298774e-01
-3.28439116e-01 1.98012993e-01 -5.19012153e-01 2.80488700e-01
6.95312560e-01 7.07426891e-02 -3.68271917e-02 -3.99692476e-01
-1.05222607e+00 -1.58634543e-01 3.11103761e-01 -1.31647184e-01
6.66781068e-01 -9.42232847e-01 -8.02810550e-01 2.08166584e-01
-4.02659595e-01 -1.79004401e-01 1.03360415e+00 1.73609555e+00
-9.37738180e-01 5.08119524e-01 -6.85770631e-01 -7.74481058e-01
-6.82974935e-01 4.85167921e-01 9.37880516e-01 -7.29235232e-01
-8.25926840e-01 4.63353217e-01 4.98174459e-01 -4.98370260e-01
-5.55123806e-01 1.37507552e-02 -3.85224782e-02 -6.03327692e-01
6.09695554e-01 4.33365554e-01 6.31129980e-01 -1.34430778e+00
-5.08302331e-01 7.79798388e-01 -1.21166714e-01 -3.94585758e-01
1.35349822e+00 -3.84525299e-01 -1.19653083e-01 -8.18814803e-03
8.87723148e-01 -3.38443190e-01 -1.02067184e+00 7.36573115e-02
-1.41683608e-01 -6.16592824e-01 8.23883235e-01 -1.01321387e+00
-9.87480819e-01 7.72706985e-01 1.09096014e+00 -2.28545830e-01
1.20532298e+00 -9.78803337e-02 6.08876586e-01 -4.58642006e-01
7.19828904e-01 -7.65483141e-01 -4.16146278e-01 1.79051310e-01
7.10446894e-01 -8.78620386e-01 1.03336647e-01 -2.53564537e-01
-7.64572263e-01 7.95174539e-01 2.98153877e-01 1.46617800e-01
5.09808898e-01 7.89003670e-01 4.14027609e-02 -3.13046008e-01
-3.47908169e-01 -4.82561402e-02 1.93331078e-01 9.43904817e-01
7.34166801e-01 -2.10296176e-02 -6.67951822e-01 -9.94282588e-02
3.45123112e-01 2.37248942e-01 7.22717941e-01 1.08350599e+00
-8.10155421e-02 -8.27413142e-01 -7.66940236e-01 5.59540331e-01
-1.22728407e+00 -1.66721418e-01 6.17413521e-01 5.69085538e-01
2.91565269e-01 6.02673590e-01 -8.75971168e-02 4.53747243e-01
2.69006431e-01 -1.63481995e-01 1.13346589e+00 -3.10849011e-01
-8.20582330e-01 3.50972235e-01 -2.23052204e-01 -4.84789550e-01
-7.80090868e-01 -6.34446919e-01 -1.27862322e+00 -1.58146694e-01
-3.91202467e-03 2.21330523e-01 1.11226702e+00 9.83293355e-01
-1.14166159e-02 5.82504570e-01 4.65572178e-01 -9.74910855e-01
-1.27624944e-01 -1.05240905e+00 -9.86359298e-01 4.66051549e-01
7.42712617e-02 -7.83204854e-01 -1.68980882e-01 -3.87124360e-01] | [13.58226490020752, -2.407409429550171] |
987ce365-e959-4270-bfd8-ae91ac75a12b | a-copy-augmented-generative-model-for-open | null | null | https://openreview.net/forum?id=9RHCjj-vhq3 | https://openreview.net/pdf?id=9RHCjj-vhq3 | A Copy-Augmented Generative Model for Open-Domain Question Answering | Open-domain question answering is a challenging task with a wide variety of practical applications. Existing modern approaches mostly follow a standard two-stage paradigm: retriever then reader. In this article, we focus on improving the effectiveness of the reader module and propose a novel copy-augmented generative approach that integrates the merits of both extractive and generative readers. In particular, our model is built upon the powerful generative model FiD \citep{FiD}. We enhance the original generative reader by incorporating a pointer network to encourage the model to directly copy words from the retrieved passages. We conduct experiments on the two benchmark datasets, Natural Questions and TriviaQA, and the empirical results demonstrate the performance gains of our proposed approach. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['triviaqa'] | ['miscellaneous'] | [ 5.32164350e-02 1.37070954e-01 1.79480135e-01 -1.85310394e-01
-1.20838606e+00 -6.17203951e-01 7.96926379e-01 -3.95156741e-01
-3.45869631e-01 8.71652901e-01 4.91271377e-01 -3.37803781e-01
-3.82683218e-01 -8.78051519e-01 -6.13520324e-01 -5.02472818e-01
7.21850097e-01 6.77042663e-01 5.41653991e-01 -6.14088595e-01
4.05482173e-01 -2.31336176e-01 -1.07118368e+00 1.12982355e-01
1.43445933e+00 6.88571393e-01 3.88838023e-01 6.05184615e-01
-6.60405695e-01 7.38767922e-01 -8.39359820e-01 -8.39276314e-01
-4.13756967e-02 -7.16231942e-01 -8.41283679e-01 -1.70736894e-01
-1.34342358e-01 -5.61128914e-01 -5.45339882e-01 8.65456045e-01
6.99955046e-01 4.18207914e-01 5.62439799e-01 -9.75618482e-01
-1.41075385e+00 8.24228942e-01 -5.67730188e-01 5.14014006e-01
5.05545855e-01 8.42400417e-02 1.11128318e+00 -8.29443157e-01
3.04333568e-01 1.27837229e+00 2.08324134e-01 5.03505468e-01
-7.40815401e-01 -6.23424649e-01 2.84071594e-01 4.07519192e-01
-1.35868895e+00 -2.22013682e-01 9.55614090e-01 1.39781237e-02
8.33439171e-01 1.75717890e-01 1.49977073e-01 9.28512990e-01
-1.29769266e-01 8.76169384e-01 9.81646419e-01 -3.91639858e-01
1.39276326e-01 -1.25664221e-02 3.12789321e-01 4.14586663e-01
2.22936153e-01 -1.69477850e-01 -1.56194329e-01 -3.52521151e-01
6.57775760e-01 1.27082050e-01 -5.88134110e-01 6.93017915e-02
-8.89737844e-01 1.01511621e+00 5.02587199e-01 1.26538917e-01
-4.66414511e-01 1.03594586e-02 8.50034952e-02 2.38123342e-01
4.15333420e-01 3.53455514e-01 -1.25925317e-01 -1.16289429e-01
-8.23891222e-01 4.90964204e-01 9.95260537e-01 1.28712511e+00
6.06540382e-01 -3.84046495e-01 -6.54264569e-01 9.45879400e-01
6.90481305e-01 5.42266130e-01 6.48877382e-01 -7.40472734e-01
6.03208244e-01 6.80068433e-01 2.57059503e-02 -8.75492334e-01
2.19036415e-01 -7.52720714e-01 -5.73012292e-01 -6.71620607e-01
5.05668856e-02 -2.80391812e-01 -7.93500185e-01 1.62964106e+00
4.24356699e-01 8.72439798e-03 2.58699626e-01 8.04162443e-01
1.24714816e+00 1.15597939e+00 1.08237930e-01 -2.81241201e-02
1.39604652e+00 -1.37396252e+00 -9.44171071e-01 -4.07650679e-01
4.02530789e-01 -9.69244242e-01 1.20318198e+00 -3.27968523e-02
-1.14933252e+00 -5.38683653e-01 -7.90209532e-01 -4.99222696e-01
-1.98400766e-01 6.31128699e-02 1.85516417e-01 3.87022436e-01
-7.51094282e-01 2.23435223e-01 -3.89725804e-01 -2.45652884e-01
3.60732436e-01 -3.78075130e-02 1.37910783e-01 -4.37037528e-01
-1.38556588e+00 5.51181614e-01 6.39317989e-01 1.40245810e-01
-6.16584778e-01 -5.61199903e-01 -4.29670870e-01 2.79466331e-01
7.02597380e-01 -1.32190561e+00 1.56684029e+00 -4.04361993e-01
-1.51530683e+00 4.28124309e-01 -2.58093983e-01 -1.62542537e-01
3.35896730e-01 -6.04252219e-01 -3.21010470e-01 2.81172693e-01
-4.08540517e-02 4.15703058e-01 6.51060820e-01 -1.20257580e+00
-4.71481442e-01 -2.42915839e-01 4.18208659e-01 3.08830947e-01
-3.67363095e-01 1.78308524e-02 -6.53108239e-01 -7.96623111e-01
-4.09063548e-02 -7.96886563e-01 1.72844362e-02 -2.97866881e-01
-4.65592593e-01 -6.43124700e-01 6.82811201e-01 -9.66468275e-01
1.63348007e+00 -2.09112644e+00 1.04203239e-01 8.56996607e-03
3.99707258e-01 4.21594501e-01 -1.84959143e-01 7.60937274e-01
2.68325120e-01 9.17063057e-02 -2.56134301e-01 -8.20159763e-02
3.17535758e-01 2.96451181e-01 -6.85123324e-01 -2.65492857e-01
8.86829719e-02 1.31338012e+00 -8.69188786e-01 -6.73682928e-01
-3.96750510e-01 2.88770407e-01 -6.43296659e-01 6.56647801e-01
-7.17240930e-01 3.45685519e-02 -9.45177734e-01 3.61568302e-01
7.48664141e-01 -5.77441514e-01 -1.82113260e-01 7.75856003e-02
4.05493468e-01 5.59682846e-01 -8.38093400e-01 1.77548099e+00
-3.06207836e-01 1.14517607e-01 -2.27627546e-01 -3.91907632e-01
1.05096793e+00 1.70735300e-01 -3.03872019e-01 -7.77779400e-01
1.71403512e-01 2.01364234e-01 7.98894539e-02 -6.78406179e-01
8.40944946e-01 -4.79575284e-02 1.17040753e-01 8.25052023e-01
-1.44028813e-02 8.05089548e-02 2.76817113e-01 8.18378150e-01
1.07380903e+00 3.78055304e-01 9.86605361e-02 -1.81933731e-01
5.36378324e-01 -4.03278321e-02 1.70056552e-01 7.31388748e-01
8.01121816e-02 5.19022167e-01 5.13752818e-01 3.60769182e-01
-7.59243011e-01 -1.05490637e+00 4.68404651e-01 1.34982133e+00
2.64027864e-01 -4.48536068e-01 -8.23338926e-01 -8.42653513e-01
-3.58576417e-01 1.04312372e+00 -5.72914064e-01 -3.95816594e-01
-6.69833064e-01 -6.36298597e-01 5.20535350e-01 7.24402726e-01
8.07146311e-01 -1.07556701e+00 -1.31938234e-01 8.94237086e-02
-7.71426678e-01 -8.98119390e-01 -7.87489772e-01 -3.60823095e-01
-8.35059583e-01 -8.36439908e-01 -8.93617928e-01 -1.02275300e+00
5.90984464e-01 5.36152124e-01 1.29777563e+00 3.51061046e-01
3.03468764e-01 4.03307557e-01 -7.54214048e-01 -2.98446298e-01
-3.32014680e-01 5.57931185e-01 -8.22402656e-01 -1.13443084e-01
4.10470486e-01 -6.22692585e-01 -8.76046300e-01 2.54656702e-01
-1.26955938e+00 4.60803546e-02 9.58131790e-01 8.09884310e-01
4.99315888e-01 -3.86517167e-01 1.01642847e+00 -1.00816453e+00
1.25607896e+00 -9.17481124e-01 -2.25847855e-01 5.68723619e-01
-5.74973762e-01 3.16676289e-01 5.57721853e-01 -4.42966640e-01
-1.53111863e+00 -7.23082542e-01 -5.37207365e-01 -2.80382633e-01
1.13143764e-01 7.82935858e-01 -2.95626849e-01 3.55677515e-01
4.39883292e-01 5.52278757e-01 -1.02377206e-01 -7.22947359e-01
6.42741442e-01 7.83577263e-01 6.48888826e-01 -7.08894491e-01
1.06080031e+00 5.77769950e-02 -6.46052659e-01 -3.18243772e-01
-8.96617353e-01 -4.79871333e-01 2.23760623e-02 -5.50060458e-02
5.50389111e-01 -7.31527984e-01 -4.19523984e-01 1.48991078e-01
-1.32318580e+00 1.65722400e-01 -3.83243352e-01 2.00230062e-01
-2.49638155e-01 5.67801714e-01 -5.64084113e-01 -6.52896762e-01
-9.14913476e-01 -7.34024227e-01 8.83558929e-01 7.32127905e-01
1.00009479e-01 -1.09199095e+00 4.08864617e-01 5.71549714e-01
7.17791021e-01 -1.64211214e-01 1.14614928e+00 -1.18677282e+00
-8.62306237e-01 -1.26530394e-01 -3.81450593e-01 2.00454712e-01
-6.43678084e-02 -3.12208384e-01 -7.73975849e-01 -3.89214605e-02
4.57665101e-02 -3.78880352e-01 9.56593513e-01 -2.17374936e-01
9.39519227e-01 -4.43977147e-01 -2.52480149e-01 2.98388124e-01
1.41685712e+00 1.93821594e-01 9.96958613e-01 2.22405851e-01
5.05498588e-01 4.44111437e-01 5.22188604e-01 2.13965207e-01
7.16889203e-01 1.62481487e-01 2.38113195e-01 2.45315656e-01
-1.43525273e-01 -7.07804561e-01 1.74955964e-01 1.33468807e+00
8.17099139e-02 -6.73021078e-01 -5.80844939e-01 6.61019325e-01
-1.81655169e+00 -8.82598400e-01 1.59428548e-02 1.72234273e+00
9.67511296e-01 5.11373729e-02 -1.79074332e-01 -1.88301072e-01
7.08219826e-01 1.96504042e-01 -4.53849047e-01 -1.99532658e-01
-4.29849513e-03 4.74044263e-01 -2.64179647e-01 2.71905303e-01
-4.82732028e-01 7.48459935e-01 6.22464561e+00 1.05996859e+00
-3.36872578e-01 1.68235958e-01 2.41856515e-01 1.11613855e-01
-7.42863894e-01 2.18467027e-01 -9.36137319e-01 5.63717604e-01
7.82127619e-01 -5.85426748e-01 2.92022198e-01 6.52390718e-01
-2.11489022e-01 1.36823719e-02 -8.19924116e-01 6.83273196e-01
4.00547713e-01 -1.07543373e+00 4.73438591e-01 -1.36617362e-01
5.71146846e-01 -3.42861205e-01 -8.71515721e-02 8.45923305e-01
3.40687126e-01 -7.23427534e-01 3.37663591e-01 8.78594100e-01
2.39181578e-01 -7.83827543e-01 7.50042796e-01 6.66293681e-01
-8.56279671e-01 6.52106777e-02 -4.40514952e-01 1.57503024e-01
3.48119318e-01 3.15849423e-01 -4.99760151e-01 8.71046007e-01
5.00045896e-01 2.85237044e-01 -7.44939864e-01 1.15097523e+00
-4.36002582e-01 8.26165736e-01 -2.40698710e-01 -4.56588626e-01
2.71474838e-01 -3.17081839e-01 5.28960943e-01 8.93100083e-01
3.21186692e-01 5.95108271e-01 -7.28583010e-03 1.16691732e+00
-3.62274319e-01 3.31548542e-01 -2.02915087e-01 -1.80395812e-01
5.95579624e-01 1.23867285e+00 -2.58374363e-01 -4.41030800e-01
-5.05832493e-01 8.48423660e-01 3.88512313e-01 2.60826260e-01
-1.07583237e+00 -7.69877672e-01 -7.96042159e-02 -1.60487164e-02
6.38506591e-01 2.56981589e-02 1.97913036e-01 -1.09075010e+00
2.77481079e-01 -9.93384659e-01 6.66045904e-01 -1.17500329e+00
-1.47111595e+00 6.70933962e-01 8.30003172e-02 -8.34334612e-01
-2.97381520e-01 -3.53859067e-02 -8.47814560e-01 1.09531903e+00
-1.80633020e+00 -1.06951129e+00 -4.90277082e-01 7.13783979e-01
5.64505696e-01 1.62537783e-01 5.90233922e-01 3.40396047e-01
-6.05226815e-01 6.77129686e-01 4.18765187e-01 1.81682408e-01
6.27572238e-01 -1.10165894e+00 4.17156935e-01 9.18297470e-01
-3.80523950e-02 1.09749126e+00 5.26380718e-01 -6.57717884e-01
-1.31601739e+00 -1.11591506e+00 8.76438677e-01 -3.92444998e-01
6.01923823e-01 -6.67388365e-02 -1.22004604e+00 7.84223318e-01
8.51911426e-01 -5.50524950e-01 7.40388989e-01 -1.34412199e-01
-3.05355430e-01 5.62341064e-02 -8.20024729e-01 6.18664205e-01
9.24324870e-01 -2.50905991e-01 -1.40212274e+00 3.64073873e-01
1.13716483e+00 -4.21699166e-01 -6.82085693e-01 2.12943852e-01
2.98870951e-01 -6.43177688e-01 8.21171045e-01 -5.43050706e-01
7.03124762e-01 -2.61047900e-01 1.72760598e-02 -1.24563158e+00
-3.27263176e-01 -8.13966274e-01 -4.74327624e-01 1.64636576e+00
2.83830374e-01 -8.49359274e-01 4.70899254e-01 4.33511883e-01
-2.48670787e-01 -8.51030827e-01 -6.11367941e-01 -6.49378896e-01
2.46249095e-01 7.83687383e-02 8.48038912e-01 4.24435049e-01
-2.50233650e-01 9.30763543e-01 -6.56228364e-02 -1.17219873e-01
2.36503288e-01 3.53363782e-01 8.20654571e-01 -9.18781281e-01
-7.01310217e-01 -2.24017039e-01 3.38703811e-01 -1.83652413e+00
-2.77664512e-01 -8.34789693e-01 1.98347211e-01 -1.90037966e+00
5.60256004e-01 -2.41417944e-01 -1.23975769e-01 5.54170832e-02
-8.28158796e-01 -1.15361795e-01 5.99337555e-02 3.36776137e-01
-9.72575307e-01 9.66670752e-01 1.52588797e+00 1.04987465e-01
-1.36404634e-02 -1.32066116e-01 -1.37935638e+00 4.82571602e-01
6.83998346e-01 -4.28523362e-01 -8.36522758e-01 -8.65489483e-01
3.06108862e-01 -5.32089993e-02 4.09757614e-01 -6.29096746e-01
4.15760994e-01 6.44975752e-02 4.17129546e-02 -8.85517120e-01
8.28749835e-02 -5.21659315e-01 -8.48440006e-02 1.33187845e-01
-4.57340658e-01 3.74000430e-01 9.30042788e-02 8.55454743e-01
-2.31288359e-01 -4.59151596e-01 4.28533226e-01 -1.40574604e-01
-4.71820831e-01 2.97951102e-01 9.50732529e-02 7.31948972e-01
8.23506474e-01 2.32697219e-01 -8.64768863e-01 -5.55437624e-01
-2.73086369e-01 4.79649395e-01 -3.36224586e-02 5.27861357e-01
8.19250464e-01 -1.28905427e+00 -8.55901182e-01 -1.81972176e-01
2.09933668e-01 8.21324289e-02 4.47762161e-01 6.26490355e-01
-5.07392287e-01 5.40336847e-01 3.11624318e-01 -2.41336405e-01
-7.57935822e-01 7.15994298e-01 1.64011508e-01 -6.19783819e-01
-5.10292888e-01 8.55223417e-01 3.32345814e-01 -3.72160643e-01
3.25659961e-02 -5.68168014e-02 -3.74247968e-01 -2.49755412e-01
6.14709973e-01 4.13481355e-01 -1.44637242e-01 -2.56036401e-01
3.09913531e-02 2.56816536e-01 -6.16269827e-01 -6.22711778e-02
1.00318718e+00 -3.62040699e-01 -2.14863643e-01 1.12849504e-01
1.13126457e+00 3.20692994e-02 -7.42720842e-01 -7.13130236e-01
-3.01864535e-01 -3.43653649e-01 -6.31282628e-02 -9.10527468e-01
-9.09560084e-01 8.32542181e-01 2.41951570e-01 2.59733915e-01
1.24782503e+00 2.96031237e-01 1.46312034e+00 5.54824710e-01
7.18320683e-02 -8.07842016e-01 2.76811272e-01 5.31817734e-01
9.51258957e-01 -8.25521767e-01 -1.34882927e-01 -4.56253260e-01
-4.04539734e-01 7.50308514e-01 7.68729746e-01 -2.39354953e-01
4.17295665e-01 -1.84731632e-01 6.30413182e-03 -2.83855706e-01
-8.47424507e-01 -3.24993968e-01 2.76165158e-01 3.57039392e-01
2.75611103e-01 -3.88374686e-01 -7.58801103e-01 1.13662744e+00
-2.53070951e-01 1.14434280e-01 3.09596270e-01 1.22273684e+00
-6.93577170e-01 -1.07190514e+00 -3.02239299e-01 5.73369741e-01
-4.52035636e-01 -4.26872373e-01 -3.35864425e-01 7.65062094e-01
-2.38779604e-01 1.06345499e+00 -1.17736578e-01 -1.59814194e-01
3.89310688e-01 2.23168507e-01 3.29351306e-01 -4.71025109e-01
-6.24766767e-01 8.09519142e-02 -1.04824454e-01 -6.53175414e-02
-3.37803572e-01 -4.27457124e-01 -1.08153975e+00 -4.17115569e-01
-8.16465139e-01 5.44948936e-01 2.75272638e-01 1.14664209e+00
6.46500468e-01 6.99472666e-01 4.96578634e-01 1.26675338e-01
-9.48992431e-01 -1.24587333e+00 -3.30203205e-01 2.15199843e-01
-4.63532209e-02 -5.46251416e-01 -2.75639176e-01 -9.53254290e-03] | [11.275111198425293, 8.035415649414062] |
d44897ea-4f65-4f5c-849e-9354576b8ca8 | fcsr-gan-joint-face-completion-and-super | 1911.01045 | null | https://arxiv.org/abs/1911.01045v1 | https://arxiv.org/pdf/1911.01045v1.pdf | FCSR-GAN: Joint Face Completion and Super-resolution via Multi-task Learning | Combined variations containing low-resolution and occlusion often present in face images in the wild, e.g., under the scenario of video surveillance. While most of the existing face image recovery approaches can handle only one type of variation per model, in this work, we propose a deep generative adversarial network (FCSR-GAN) for performing joint face completion and face super-resolution via multi-task learning. The generator of FCSR-GAN aims to recover a high-resolution face image without occlusion given an input low-resolution face image with occlusion. The discriminator of FCSR-GAN uses a set of carefully designed losses (an adversarial loss, a perceptual loss, a pixel loss, a smooth loss, a style loss, and a face prior loss) to assure the high quality of the recovered high-resolution face images without occlusion. The whole network of FCSR-GAN can be trained end-to-end using our two-stage training strategy. Experimental results on the public-domain CelebA and Helen databases show that the proposed approach outperforms the state-of-the-art methods in jointly performing face super-resolution (up to 8 $\times$) and face completion, and shows good generalization ability in cross-database testing. Our FCSR-GAN is also useful for improving face identification performance when there are low-resolution and occlusion in face images. | ['Shiguang Shan', 'Xilin Chen', 'Jiancheng Cai', 'Hu Han'] | 2019-11-04 | null | null | null | null | ['facial-inpainting'] | ['computer-vision'] | [ 3.00097585e-01 -5.22923507e-02 2.24671587e-01 -5.29577732e-01
-1.08398652e+00 -1.54092178e-01 3.69284302e-01 -9.54173267e-01
-9.90165994e-02 7.60766685e-01 -1.46769598e-01 2.54449725e-01
9.62204859e-02 -9.16191339e-01 -1.00013697e+00 -9.03794825e-01
2.56432354e-01 4.11150575e-01 -2.99968004e-01 -2.85069376e-01
-3.60881269e-01 6.78992152e-01 -1.76472127e+00 3.34882945e-01
7.99536645e-01 1.12383711e+00 6.10070070e-03 4.95193541e-01
1.29155293e-01 6.04101002e-01 -7.22239912e-01 -8.69186640e-01
6.35265589e-01 -5.22263348e-01 -2.60383010e-01 1.35142133e-01
8.12585652e-01 -7.62551129e-01 -4.08422798e-01 9.85837638e-01
8.61529171e-01 -3.10440473e-02 5.95330834e-01 -1.12036431e+00
-9.81115937e-01 1.00610577e-01 -1.06587899e+00 -1.53030297e-02
4.91223842e-01 2.13137865e-01 2.75016189e-01 -1.05754423e+00
4.79480296e-01 1.73908281e+00 7.18026042e-01 1.04551327e+00
-1.32305598e+00 -1.11394393e+00 -1.79634601e-01 -1.54124007e-01
-1.69765019e+00 -7.80901253e-01 8.16318214e-01 -2.53270119e-01
3.50701869e-01 9.62078124e-02 7.67545030e-02 1.34175861e+00
8.27332900e-04 1.52976632e-01 1.21432996e+00 -4.44354005e-02
-1.17337421e-01 7.33375624e-02 -7.33809948e-01 6.87046945e-01
1.05301984e-01 2.33643502e-01 -3.99145067e-01 -1.96071863e-01
1.28596807e+00 1.62544355e-01 -3.93908203e-01 1.28729433e-01
-4.73055691e-01 7.65242159e-01 3.45542014e-01 -1.89620927e-02
-4.47042346e-01 -1.82567567e-01 5.88395409e-02 4.49958175e-01
7.71032631e-01 -1.64684847e-01 -1.20416671e-01 6.24852657e-01
-1.06085396e+00 3.03070694e-01 3.46680015e-01 9.15110588e-01
7.22125173e-01 4.72557187e-01 -3.48634362e-01 1.29744589e+00
2.20157996e-01 7.83318341e-01 2.63829917e-01 -8.19319069e-01
4.49231595e-01 1.42020404e-01 1.98802650e-01 -7.91375816e-01
1.25790700e-01 -3.08097780e-01 -1.17453730e+00 5.01525819e-01
2.00308204e-01 -2.25132272e-01 -9.57281709e-01 2.01316619e+00
4.87077773e-01 3.94038826e-01 2.19566211e-01 1.00096738e+00
1.04305279e+00 5.24192333e-01 5.72985224e-03 -4.65525776e-01
1.36835563e+00 -6.36773348e-01 -9.37056422e-01 -1.94903985e-01
-4.62303877e-01 -9.14008081e-01 7.95273364e-01 2.09919631e-01
-1.38985264e+00 -9.46691215e-01 -8.34281802e-01 -9.84432548e-02
8.86097327e-02 2.40054652e-01 2.75108188e-01 7.59592533e-01
-1.16783452e+00 3.26083481e-01 -3.95467550e-01 1.19146876e-01
9.22913551e-01 4.29916590e-01 -7.72290766e-01 -5.06892085e-01
-1.07016790e+00 6.24682486e-01 -1.07372992e-01 3.29077095e-01
-1.35452747e+00 -9.52380180e-01 -8.63830924e-01 -1.18690766e-01
2.99656391e-01 -7.07935095e-01 6.58677340e-01 -1.36438310e+00
-1.67504787e+00 1.08924532e+00 -4.10090871e-02 -1.82765305e-01
8.51077676e-01 -1.89408451e-01 -6.57186985e-01 2.62082607e-01
5.14433719e-02 5.86022913e-01 1.53039777e+00 -1.41982186e+00
-2.35425189e-01 -6.27460897e-01 -2.05860376e-01 6.72793388e-02
-1.07587583e-01 2.83929825e-01 -5.38260460e-01 -7.59243608e-01
-3.43273491e-01 -5.46953797e-01 1.99618772e-01 1.66550666e-01
-2.72636972e-02 4.71268259e-02 1.07376206e+00 -1.19944394e+00
6.34073675e-01 -2.19648314e+00 2.53888309e-01 -1.85459301e-01
-8.74241367e-02 5.22097588e-01 -5.33504725e-01 -1.81579724e-01
-3.26595217e-01 1.86494403e-02 -4.22978610e-01 -7.23950684e-01
-4.03058022e-01 7.04173893e-02 -3.13452423e-01 6.54649913e-01
5.09948432e-01 6.83417559e-01 -4.55029339e-01 -3.49542052e-01
1.93820149e-01 1.40684223e+00 -5.08680046e-01 7.78199136e-01
-2.04697251e-01 8.64042997e-01 -1.90801308e-01 7.89412618e-01
1.31687987e+00 1.88014925e-01 -7.06282333e-02 -3.00056607e-01
3.37110043e-01 -4.73871708e-01 -1.16973174e+00 1.57210958e+00
-5.85664332e-01 1.15588151e-01 5.92382133e-01 -5.43508351e-01
1.14434242e+00 4.62928653e-01 2.47863665e-01 -7.13716090e-01
7.08541572e-02 1.06525637e-01 -3.75466853e-01 -3.88822109e-01
1.16121471e-01 -4.18312997e-01 2.47947469e-01 3.77250724e-02
2.53052205e-01 1.25370517e-01 -1.86079785e-01 -2.13848516e-01
7.22992539e-01 2.65121937e-01 -6.23919778e-02 6.11585900e-02
7.74927735e-01 -9.33258176e-01 7.67255545e-01 1.43227383e-01
7.51105174e-02 1.01922572e+00 2.39146620e-01 -2.39163101e-01
-1.21013653e+00 -1.07866740e+00 -2.53210336e-01 1.04484677e+00
-2.83113211e-01 2.12069780e-01 -9.62080300e-01 -4.75680113e-01
-2.48461068e-02 2.36414447e-01 -5.90441167e-01 -6.18354091e-03
-7.05474615e-01 -7.82905459e-01 6.40047014e-01 3.77326757e-01
9.55116034e-01 -1.18672144e+00 4.99702580e-02 -1.35239854e-01
-2.27970466e-01 -1.35234523e+00 -6.32736981e-01 -5.14500558e-01
-4.85679835e-01 -9.47930276e-01 -1.02066934e+00 -7.92516291e-01
7.86171973e-01 8.96719694e-02 1.01679599e+00 4.78829965e-02
-4.95761216e-01 2.12405816e-01 -3.43913287e-02 4.95677255e-02
-5.04567683e-01 -4.79095012e-01 -6.74930066e-02 6.11932278e-01
-1.22426916e-03 -6.73534930e-01 -7.49169290e-01 3.84603232e-01
-1.00965142e+00 -3.17434400e-01 6.00001693e-01 1.07118738e+00
7.78486609e-01 3.12921345e-01 7.36294746e-01 -8.89519393e-01
2.94217169e-01 -4.56873745e-01 -7.60914624e-01 2.47810856e-01
-2.87803769e-01 -2.31651202e-01 7.68086433e-01 -5.21108031e-01
-1.55303383e+00 1.09144486e-01 -4.57185805e-01 -1.02089572e+00
-2.35769764e-01 -2.51740009e-01 -9.05441940e-01 -3.17136556e-01
5.42298138e-01 2.34000757e-01 1.94392785e-01 -5.08935273e-01
3.45284313e-01 5.77521324e-01 7.58539498e-01 -4.63038802e-01
1.02957046e+00 5.26013732e-01 1.56538069e-01 -8.37068856e-01
-6.12751424e-01 6.70809075e-02 -3.02238017e-01 -3.09183877e-02
8.57625663e-01 -1.43752468e+00 -7.91463733e-01 7.24004805e-01
-9.36290860e-01 -2.51916170e-01 -1.63402975e-01 1.80309676e-02
-5.17754138e-01 1.64792627e-01 -5.45181751e-01 -9.54585731e-01
-6.37859344e-01 -1.17261922e+00 1.32306492e+00 4.56548840e-01
5.35481572e-01 -7.00459123e-01 -2.91778028e-01 7.61962652e-01
6.62908852e-01 6.26686096e-01 4.49706703e-01 -2.93798484e-02
-5.22647798e-01 6.45957747e-03 -2.20284700e-01 6.76530421e-01
1.37763903e-01 -1.45829037e-01 -1.31898260e+00 -7.74805129e-01
1.70073032e-01 -5.27435303e-01 8.81164312e-01 3.30371827e-01
1.41016972e+00 -5.50863802e-01 9.42211822e-02 1.11393797e+00
1.62058556e+00 8.40037875e-03 1.03378844e+00 -4.74388003e-01
7.78798342e-01 5.89885533e-01 4.12542820e-01 3.98844540e-01
1.06642112e-01 8.36123884e-01 5.24124384e-01 -3.25705141e-01
-4.65482563e-01 -2.79486418e-01 4.25836712e-01 7.45540857e-02
-2.48664752e-01 -1.64986596e-01 -3.19465518e-01 2.66567469e-01
-1.37488949e+00 -1.12393332e+00 3.73207092e-01 2.35367084e+00
9.03339863e-01 -4.43349063e-01 1.72883064e-01 -8.09586421e-02
1.13403010e+00 1.82481319e-01 -7.41092384e-01 2.76348703e-02
-4.32813823e-01 6.06308699e-01 2.97497928e-01 5.40436566e-01
-9.97710526e-01 9.19329941e-01 5.33971214e+00 1.04084826e+00
-1.09546804e+00 4.97219563e-01 9.83864725e-01 -2.25497350e-01
-1.45520046e-01 -5.92754185e-01 -9.69766319e-01 4.99456674e-01
7.99061179e-01 3.72402012e-01 8.80501628e-01 7.57590711e-01
-5.89066632e-02 3.56890678e-01 -7.52070427e-01 1.29331100e+00
5.43216348e-01 -8.43976140e-01 2.16403037e-01 -1.05554964e-02
7.72240996e-01 -4.94418621e-01 5.81340969e-01 2.18855217e-01
9.30895060e-02 -1.53211951e+00 4.18304920e-01 5.47327220e-01
1.65466940e+00 -1.10454726e+00 5.68769276e-01 -1.11161262e-01
-1.20972753e+00 -1.25843287e-01 -5.37384808e-01 5.50553322e-01
1.00906529e-02 5.03504097e-01 -2.73968756e-01 7.03388512e-01
6.68824792e-01 3.51665169e-01 -3.81213337e-01 4.06950504e-01
-2.61497349e-01 1.24078639e-01 -2.65907403e-02 9.22874451e-01
-4.14100707e-01 -3.09642613e-01 5.06402016e-01 8.98859739e-01
2.88932830e-01 2.83854306e-01 -6.03802949e-02 1.04402483e+00
-6.08321607e-01 -1.17646202e-01 -5.34722090e-01 3.38733226e-01
4.74961907e-01 1.48098576e+00 -2.07343668e-01 5.90361729e-02
-3.65785986e-01 1.25369263e+00 3.29064935e-01 4.22738373e-01
-1.01722145e+00 5.38027799e-03 8.10556114e-01 3.55268955e-01
5.10084867e-01 3.50881785e-01 1.28187373e-01 -9.45512474e-01
2.49370277e-01 -1.13432002e+00 2.63100475e-01 -6.75710380e-01
-1.46396065e+00 9.96825278e-01 -2.39386097e-01 -9.40546989e-01
-2.86601573e-01 -3.22518826e-01 -4.54722971e-01 1.29072678e+00
-1.71387005e+00 -1.64455378e+00 -6.28162861e-01 1.04906523e+00
5.11736393e-01 -6.81183994e-01 8.03035200e-01 7.52781808e-01
-7.06420600e-01 1.17299962e+00 -1.50382966e-01 2.26697206e-01
9.29270506e-01 -7.62047470e-01 1.70618549e-01 8.35234821e-01
-3.41060966e-01 4.54824269e-01 6.18313789e-01 -6.18145287e-01
-1.55113363e+00 -1.55659664e+00 1.09278679e-01 -1.57421008e-01
-1.62450463e-01 -4.55682963e-01 -1.06028318e+00 6.05067134e-01
1.33268848e-01 5.45564294e-01 5.71896553e-01 -3.56206357e-01
-6.73514903e-01 -5.19373417e-01 -2.01459384e+00 2.57404089e-01
1.01744533e+00 -5.85827768e-01 -1.13553405e-01 2.18550786e-01
6.05248213e-01 -4.99525666e-01 -1.10934746e+00 5.92075825e-01
6.55585051e-01 -1.03140569e+00 1.30577958e+00 -2.07724229e-01
5.11835515e-01 -3.13610852e-01 -2.92122066e-01 -1.08660102e+00
-2.07517251e-01 -7.41434276e-01 -7.12117366e-03 1.66005003e+00
-2.34869309e-02 -5.11436999e-01 4.99028385e-01 4.01750326e-01
1.81594044e-01 -4.84157890e-01 -9.94788885e-01 -6.18904769e-01
4.01666276e-02 3.51154476e-01 8.80415440e-01 7.03610539e-01
-1.08082211e+00 2.04285860e-01 -8.46892715e-01 3.70948941e-01
1.11154211e+00 -1.18371285e-01 8.49155664e-01 -1.06030595e+00
-3.57112229e-01 -9.16957036e-02 -1.28914401e-01 -5.31952262e-01
5.46966076e-01 -4.55473602e-01 -1.56971991e-01 -9.69047546e-01
3.12038809e-01 -2.05925748e-01 5.16674258e-02 3.24566931e-01
-2.02381656e-01 7.00614035e-01 -6.43453002e-03 3.54323387e-02
-2.17805594e-01 7.57795572e-01 1.40853548e+00 4.50610444e-02
-1.83284767e-02 -8.54534376e-03 -7.56463051e-01 4.89330441e-01
4.79683459e-01 -2.10146710e-01 -3.42122495e-01 -3.92824590e-01
-3.19658220e-01 4.35758203e-01 5.03688335e-01 -8.46084177e-01
-1.24193996e-01 -1.58889797e-02 9.45681572e-01 -3.49214852e-01
7.84208596e-01 -8.82848680e-01 7.01107919e-01 7.02781975e-02
-6.05787970e-02 -1.88899785e-01 2.70565718e-01 5.01338005e-01
-1.76492214e-01 2.72368014e-01 1.55981135e+00 -1.65392369e-01
-1.89430952e-01 7.63934135e-01 5.07661760e-01 8.82128403e-02
9.35801148e-01 4.89420481e-02 -3.95145148e-01 -3.15193713e-01
-4.99468118e-01 -1.63812533e-01 4.91598517e-01 5.37988842e-01
9.20769691e-01 -1.44989336e+00 -1.32107532e+00 7.97080994e-01
-2.07721472e-01 1.49431497e-01 7.21116841e-01 3.24034989e-01
-2.64006466e-01 -1.35727212e-01 -5.93508363e-01 -3.42282981e-01
-1.38756955e+00 6.42477095e-01 5.60685158e-01 -1.04620844e-01
-5.53573012e-01 7.86999524e-01 4.13512319e-01 -2.57149398e-01
1.94915399e-01 5.67060173e-01 -1.51334673e-01 -7.43351355e-02
9.55620885e-01 2.69973993e-01 -9.76079423e-03 -1.05998993e+00
-1.94215313e-01 7.73480058e-01 -1.14287540e-01 -8.68040312e-04
1.48331296e+00 -3.71472985e-02 -2.00492427e-01 -2.05512732e-01
1.25170100e+00 -4.99719344e-02 -1.62783921e+00 -1.44729599e-01
-9.41926122e-01 -8.62846971e-01 -2.86489893e-02 -7.66339302e-01
-1.71790123e+00 7.01763391e-01 1.05742884e+00 -3.20862293e-01
1.63806546e+00 -2.72972822e-01 6.25673354e-01 -3.69150311e-01
4.63732809e-01 -5.92185497e-01 1.55898944e-01 -8.96611251e-03
1.37478614e+00 -1.34474707e+00 -1.17737263e-01 -4.72437948e-01
-6.14099324e-01 6.89868391e-01 7.65060723e-01 -2.82933325e-01
5.29695690e-01 4.05951321e-01 -1.09564617e-01 1.58508420e-01
-4.43813473e-01 -5.50556295e-02 1.87789589e-01 9.17053640e-01
2.58003056e-01 -9.73986462e-02 4.28311676e-02 6.10438704e-01
-5.76009341e-02 -4.60465364e-02 1.38078764e-01 1.99585810e-01
4.46554385e-02 -9.56338048e-01 -6.85401857e-01 1.62323758e-01
-8.97638381e-01 7.84435272e-02 3.15557085e-02 5.66499949e-01
4.28065121e-01 9.84915912e-01 1.30202219e-01 -2.26271391e-01
3.06155086e-01 -6.73273429e-02 7.20049620e-01 -3.24804157e-01
-6.28494203e-01 2.85063893e-01 -3.52270275e-01 -6.22896791e-01
-3.27115953e-01 -4.41039234e-01 -7.33747125e-01 -4.89083320e-01
-1.17577650e-01 -1.92254394e-01 5.72719753e-01 4.88691986e-01
3.59694004e-01 5.95355034e-01 9.92764831e-01 -8.97040367e-01
-5.67796350e-01 -1.14537108e+00 -6.56841278e-01 5.89803398e-01
5.44303238e-01 -6.69931710e-01 -4.31258887e-01 1.49611279e-01] | [12.824518203735352, 0.011312554590404034] |
c3450b62-d439-42d7-a392-ba8348f7111a | a-partition-filter-network-for-joint-entity | 2108.12202 | null | https://arxiv.org/abs/2108.12202v8 | https://arxiv.org/pdf/2108.12202v8.pdf | A Partition Filter Network for Joint Entity and Relation Extraction | In joint entity and relation extraction, existing work either sequentially encode task-specific features, leading to an imbalance in inter-task feature interaction where features extracted later have no direct contact with those that come first. Or they encode entity features and relation features in a parallel manner, meaning that feature representation learning for each task is largely independent of each other except for input sharing. We propose a partition filter network to model two-way interaction between tasks properly, where feature encoding is decomposed into two steps: partition and filter. In our encoder, we leverage two gates: entity and relation gate, to segment neurons into two task partitions and one shared partition. The shared partition represents inter-task information valuable to both tasks and is evenly shared across two tasks to ensure proper two-way interaction. The task partitions represent intra-task information and are formed through concerted efforts of both gates, making sure that encoding of task-specific features is dependent upon each other. Experiment results on six public datasets show that our model performs significantly better than previous approaches. In addition, contrary to what previous work has claimed, our auxiliary experiments suggest that relation prediction is contributory to named entity prediction in a non-negligible way. The source code can be found at https://github.com/Coopercoppers/PFN. | ['Zhongyu Wei', 'Qi Zhang', 'Jinlan Fu', 'Chong Zhang', 'Zhiheng Yan'] | 2021-08-27 | null | https://aclanthology.org/2021.emnlp-main.17 | https://aclanthology.org/2021.emnlp-main.17.pdf | emnlp-2021-11 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 2.43431494e-01 2.17656076e-01 -3.28740031e-01 -5.80247700e-01
-6.41380906e-01 -5.95533192e-01 5.16894817e-01 2.47917458e-01
-5.20402908e-01 8.98551583e-01 3.07328850e-01 -3.29718441e-02
-1.86831176e-01 -7.62556672e-01 -7.94401467e-01 -4.59764779e-01
-2.11299658e-01 3.79772455e-01 3.10861826e-01 -1.02156125e-01
-4.49808687e-02 -1.76477637e-02 -1.42245185e+00 6.62105799e-01
6.69039547e-01 9.83546376e-01 1.77198872e-01 5.29519796e-01
-1.09398641e-01 8.60930443e-01 -5.42251945e-01 -5.13339639e-01
3.04763377e-01 6.63729152e-03 -1.27160048e+00 -1.90149799e-01
-1.90167442e-01 9.82245281e-02 -4.26520348e-01 8.05160522e-01
5.06316721e-01 1.79277599e-01 6.34491324e-01 -1.40443361e+00
-7.71440506e-01 9.59001124e-01 -4.25188690e-01 6.48063570e-02
-2.58966386e-02 2.40950696e-02 1.46835792e+00 -7.45453775e-01
6.18416965e-01 9.58411336e-01 6.67096376e-01 3.07913721e-01
-1.37187266e+00 -8.88889194e-01 4.41283882e-01 -2.09670290e-02
-1.53371012e+00 -5.32029033e-01 4.40546215e-01 -7.06140697e-01
1.48109007e+00 2.69222081e-01 4.45710361e-01 1.11059666e+00
5.22965074e-01 8.36165130e-01 6.63870037e-01 3.48216891e-02
-1.50895610e-01 1.25476450e-01 5.83167434e-01 4.17587548e-01
4.14668530e-01 -1.28742501e-01 -7.59213448e-01 -4.28501755e-01
5.30441344e-01 1.25107065e-01 -5.38818061e-01 -2.51400858e-01
-1.32603669e+00 6.59274459e-01 3.22989047e-01 5.23581326e-01
-5.19386053e-01 1.85000584e-01 4.97349262e-01 5.11826873e-01
5.48222601e-01 5.21459043e-01 -1.12578785e+00 -1.73153356e-01
-6.53142750e-01 1.24531761e-01 9.43975747e-01 1.10073328e+00
9.99281466e-01 -5.67126632e-01 -5.55687904e-01 7.06039310e-01
1.87523097e-01 -3.54348570e-02 6.12849832e-01 -4.94600713e-01
6.70746446e-01 9.05262113e-01 -8.34301859e-02 -9.06659126e-01
-5.97492158e-01 -5.53613603e-01 -9.50272679e-01 -3.34021360e-01
4.13226813e-01 -3.40120107e-01 -7.68523157e-01 2.02249336e+00
1.63299531e-01 -8.06841180e-02 -4.60642911e-02 5.24466574e-01
8.73728871e-01 3.21831167e-01 2.21254036e-01 6.08329894e-03
1.78666353e+00 -9.00487542e-01 -6.68183982e-01 -4.23701495e-01
9.30632651e-01 -8.13952982e-01 7.33650982e-01 2.29290560e-01
-8.72854650e-01 -5.55243850e-01 -9.09698009e-01 -5.08708596e-01
-4.61956739e-01 3.83769691e-01 7.87847519e-01 3.36987972e-01
-8.31355929e-01 5.20406604e-01 -7.57763028e-01 -1.36231974e-01
6.44643247e-01 7.56593823e-01 -6.25501156e-01 1.57389387e-01
-1.58070076e+00 7.76124537e-01 5.05734503e-01 1.04111500e-01
-4.61729854e-01 -6.48915708e-01 -8.16120923e-01 3.72194171e-01
4.78920698e-01 -9.11467373e-01 1.22271252e+00 -5.27338922e-01
-9.90436554e-01 6.80013478e-01 -5.27009130e-01 -3.30342948e-01
1.00305155e-01 -3.37378770e-01 -1.67483658e-01 -4.49016154e-01
1.40677646e-01 5.45642555e-01 5.81152439e-01 -9.68347788e-01
-8.56132865e-01 -3.52057904e-01 1.05656929e-01 2.86002487e-01
-4.29252058e-01 -2.08765995e-02 -7.78596759e-01 -4.48103458e-01
-7.24501582e-03 -6.85990870e-01 -1.79803684e-01 -3.86431158e-01
-6.99347317e-01 -6.25987113e-01 5.22151947e-01 -3.83088499e-01
1.66451657e+00 -2.07213426e+00 1.06763080e-01 9.76212248e-02
7.51381218e-01 1.15589477e-01 -1.28719777e-01 4.61985588e-01
-2.51366585e-01 3.97500008e-01 -2.54208863e-01 -5.46873152e-01
2.46968493e-02 6.95235804e-02 -1.62249014e-01 3.58335137e-01
5.86080074e-01 1.11309087e+00 -7.90503204e-01 -2.55141228e-01
-4.71653312e-01 6.28772914e-01 -3.39957148e-01 2.04125606e-02
1.02386832e-01 2.76506662e-01 -7.90073037e-01 4.34935391e-01
5.38617849e-01 -6.92631066e-01 3.07298601e-01 -2.64329702e-01
-4.30011749e-03 9.09367383e-01 -1.09011400e+00 1.59756744e+00
-2.58469850e-01 4.87350166e-01 -9.65925958e-03 -1.02996016e+00
7.64719307e-01 5.64694464e-01 4.99621004e-01 -4.78177547e-01
1.82646379e-01 1.30858263e-02 1.89527974e-01 -2.28408694e-01
6.23831570e-01 1.56238928e-01 -2.25565851e-01 7.11387634e-01
3.37101400e-01 3.85825276e-01 3.04504842e-01 2.78286874e-01
1.43080163e+00 2.81277020e-02 5.66547871e-01 -2.36025259e-01
3.30259144e-01 -1.48614779e-01 1.04365492e+00 7.75859594e-01
-2.10406080e-01 5.37364721e-01 8.75463665e-01 -4.45406199e-01
-7.20970571e-01 -5.97743034e-01 -1.73481733e-01 1.36612260e+00
-6.44082576e-02 -7.88459659e-01 -4.40824121e-01 -9.68299925e-01
1.93304747e-01 3.54370594e-01 -9.86924350e-01 -3.03693414e-01
-3.86904120e-01 -7.89186060e-01 7.02705204e-01 6.67315185e-01
2.15009928e-01 -1.05377984e+00 -4.29044843e-01 3.28431666e-01
-3.79183054e-01 -9.19673383e-01 -5.79021811e-01 9.78881359e-01
-5.56299269e-01 -1.28526807e+00 -4.65216994e-01 -5.68270385e-01
5.56065619e-01 3.13927084e-01 1.30459785e+00 1.19852006e-01
-7.67513514e-02 1.05341682e-02 -3.06338727e-01 -3.94368529e-01
2.04745874e-01 7.14592457e-01 -1.06016740e-01 -9.47502032e-02
6.91183984e-01 -5.38674116e-01 -4.93156195e-01 5.17475724e-01
-7.06122279e-01 1.19506381e-01 6.56348348e-01 1.10551906e+00
5.25217950e-01 6.91147596e-02 4.52324927e-01 -1.27963638e+00
6.09036207e-01 -8.60205233e-01 -1.03188470e-01 3.52299988e-01
-5.23557484e-01 1.11547373e-01 4.37341392e-01 -2.73617893e-01
-9.49252367e-01 3.25240254e-01 1.51301727e-01 -1.43374026e-01
-1.39319360e-01 5.55453062e-01 -4.40356046e-01 4.00563449e-01
5.57823062e-01 1.99742541e-01 -2.35042065e-01 -4.22006309e-01
1.58566892e-01 6.10031962e-01 1.08823910e-01 -6.78368390e-01
6.32319868e-01 1.39312133e-01 -4.80550528e-01 -3.70778590e-01
-7.87605703e-01 -4.62882191e-01 -7.30738461e-01 3.64153028e-01
9.26656067e-01 -1.12373519e+00 -8.27817738e-01 4.66637552e-01
-1.34363055e+00 -3.95628065e-01 -4.93132263e-01 5.02183378e-01
-1.70940712e-01 -3.90421562e-02 -6.72631681e-01 -4.65297282e-01
-3.29973161e-01 -1.27280271e+00 1.04771352e+00 7.69501776e-02
-4.23675954e-01 -9.10616815e-01 -1.02439791e-01 2.26935908e-01
3.16941440e-01 -3.57982248e-01 8.14647675e-01 -1.22753477e+00
-4.15343493e-01 -2.45588203e-03 -3.85459155e-01 1.71223506e-01
3.72571677e-01 -2.10531667e-01 -1.13115740e+00 -2.76512891e-01
-9.24888849e-02 -3.38921040e-01 1.20311224e+00 2.56195426e-01
9.76542056e-01 -1.59150317e-01 -7.34188914e-01 5.32154679e-01
8.77992451e-01 1.30738933e-02 3.47584248e-01 1.53731689e-01
8.37256789e-01 7.33639419e-01 4.16503936e-01 4.69834656e-01
8.10659945e-01 5.19029915e-01 5.83946034e-02 -1.79595396e-01
-8.16351026e-02 -1.84624754e-02 3.16784263e-01 7.24488676e-01
2.20450130e-03 -4.67420697e-01 -1.08332515e+00 6.13959312e-01
-2.24678636e+00 -6.98054135e-01 -3.48761410e-01 2.13599801e+00
1.14555585e+00 1.01534858e-01 -1.38033897e-01 -6.83202001e-04
5.40381849e-01 2.41063759e-02 -4.73099381e-01 -1.95162147e-01
-1.52897149e-01 6.60692602e-02 4.85409200e-01 2.47428611e-01
-1.27919710e+00 9.00654674e-01 5.77407217e+00 8.78113270e-01
-9.48224902e-01 2.92949796e-01 8.74976933e-01 -1.67291969e-01
-4.36237127e-01 9.33988094e-02 -1.04596221e+00 5.14350414e-01
7.30069935e-01 -3.78174752e-01 1.18002214e-01 4.95222390e-01
-3.14554840e-01 -8.54564160e-02 -1.31797051e+00 4.99621660e-01
-3.89705181e-01 -1.05939651e+00 -1.35349020e-01 2.51677930e-01
5.04651964e-01 3.37518096e-01 -1.47530809e-01 5.04265964e-01
6.93412423e-01 -1.02284908e+00 6.21039152e-01 5.29488623e-01
5.02160728e-01 -7.34616160e-01 8.34269822e-01 3.31611514e-01
-1.68806779e+00 5.35079278e-03 -4.33399320e-01 -2.05166966e-01
5.58319874e-02 9.93069828e-01 -6.59174919e-01 8.01092982e-01
6.91699088e-01 6.28229618e-01 -5.52673340e-01 7.43238807e-01
-3.09023231e-01 5.96962810e-01 -1.80560350e-01 3.60155433e-01
-2.44833156e-02 2.03268513e-01 4.13471729e-01 1.27491093e+00
7.94555023e-02 -2.45383307e-02 3.74194890e-01 7.68983901e-01
-2.59782165e-01 1.21058978e-01 -6.84140265e-01 -1.72987014e-01
7.02628136e-01 1.29792595e+00 -5.88619471e-01 -3.93263757e-01
-7.05024838e-01 8.86078835e-01 7.58517802e-01 4.15278196e-01
-7.08857358e-01 -6.70289099e-01 9.96000588e-01 -1.01080038e-01
6.27632558e-01 -7.05975294e-02 -5.26979387e-01 -1.18150675e+00
1.66136652e-01 -5.36396980e-01 4.36131716e-01 -1.44586354e-01
-1.58600461e+00 8.04830015e-01 -3.09626609e-01 -1.00025904e+00
-1.09669209e-01 -3.34803194e-01 -4.29613620e-01 1.26309526e+00
-1.32524383e+00 -1.35967934e+00 1.62947163e-01 5.50185800e-01
2.12917700e-01 -5.57502992e-02 8.80897641e-01 4.02901322e-01
-8.18199098e-01 8.75477493e-01 -1.89616844e-01 3.17278117e-01
9.83981371e-01 -1.09906602e+00 7.48666465e-01 6.45411491e-01
1.04231812e-01 1.00965774e+00 6.74849823e-02 -8.67913425e-01
-1.19662070e+00 -1.21137917e+00 1.63658881e+00 -5.78152835e-01
5.70278525e-01 -7.34347224e-01 -1.04794288e+00 8.87456417e-01
1.66632414e-01 2.58241206e-01 1.14123070e+00 7.52616882e-01
-6.34173334e-01 1.16020150e-01 -7.78005540e-01 1.51716784e-01
1.19401288e+00 -5.70859253e-01 -4.73357856e-01 1.02119386e-01
1.01701701e+00 -2.68674523e-01 -1.03905094e+00 3.69343281e-01
5.96420765e-01 -6.72867596e-01 6.82644129e-01 -8.51422608e-01
4.87579584e-01 -3.19524527e-01 -1.71198130e-01 -1.29922605e+00
-7.43449271e-01 -4.86144662e-01 -2.34422445e-01 1.42359757e+00
9.33247983e-01 -8.91243100e-01 4.86394405e-01 7.41203666e-01
-1.60979673e-01 -1.11185312e+00 -6.80553615e-01 -6.92311108e-01
6.08266778e-02 -3.85923117e-01 8.08324933e-01 1.33987594e+00
5.78364395e-02 8.70352328e-01 -3.28281879e-01 4.15235162e-02
1.97352588e-01 1.39454633e-01 5.60078919e-01 -1.45292437e+00
-3.66581559e-01 -4.43169475e-01 -7.61865005e-02 -1.00294769e+00
2.71921098e-01 -1.23393965e+00 1.01140775e-01 -1.44402254e+00
6.87633455e-01 -7.35036373e-01 -7.46007502e-01 1.17468417e+00
-6.60177052e-01 1.67789355e-01 9.36358199e-02 3.77776980e-01
-7.26028085e-01 5.01098514e-01 1.19625652e+00 2.21514199e-02
-2.88270652e-01 9.70096700e-03 -1.30860949e+00 4.74156111e-01
8.36142063e-01 -8.24430645e-01 -4.50592518e-01 -6.32889926e-01
3.66572231e-01 -3.09377611e-01 5.88388070e-02 -8.19482565e-01
4.99231458e-01 7.52034560e-02 5.60825408e-01 -3.47512722e-01
1.41249940e-01 -6.70440793e-01 1.25474647e-01 9.11599398e-02
-4.62175757e-01 -1.93759516e-01 7.95299187e-02 4.96488094e-01
-2.96793401e-01 6.44432828e-02 1.78283259e-01 8.61574635e-02
-3.14969331e-01 5.36140323e-01 -2.57149667e-01 8.31774771e-02
8.09906483e-01 -7.03062639e-02 -4.64488477e-01 -7.38370046e-02
-7.92140365e-01 4.65742737e-01 2.16671199e-01 4.98299122e-01
1.02816090e-01 -1.30529320e+00 -7.70430923e-01 2.26387918e-01
1.52090222e-01 2.58727282e-01 3.20619345e-01 1.03815162e+00
4.65026110e-01 5.30057907e-01 3.89470048e-02 -3.20909142e-01
-1.06324744e+00 3.11222583e-01 -1.24690406e-01 -9.30821061e-01
-3.22962284e-01 1.29125762e+00 7.86476016e-01 -6.75760388e-01
1.93842113e-01 -3.02850008e-01 -1.74048051e-01 4.38220203e-01
4.95154351e-01 4.14672904e-02 2.86678493e-01 -6.62116468e-01
-5.20803154e-01 1.44188061e-01 -5.87500274e-01 2.46657163e-01
1.49111271e+00 -7.34480545e-02 -2.18655378e-01 4.94915128e-01
1.23645651e+00 -8.76317322e-02 -1.18206596e+00 -5.50007522e-01
5.95408841e-04 -1.53291211e-01 -6.01853766e-02 -7.95020163e-01
-1.24095416e+00 7.18154728e-01 8.74541327e-02 2.36944079e-01
1.15029109e+00 7.46317133e-02 7.83666253e-01 3.41289282e-01
2.86421955e-01 -8.83955240e-01 -3.25442821e-01 9.06832039e-01
6.91518903e-01 -1.06967866e+00 2.53821071e-02 -7.37581551e-01
-7.94436693e-01 7.15263307e-01 8.38949859e-01 9.43399966e-02
8.21341753e-01 6.55157030e-01 -3.70787978e-01 -2.85270512e-01
-1.38692808e+00 -3.27348828e-01 5.09363174e-01 4.97645408e-01
1.09628808e+00 2.00567737e-01 -5.08790016e-01 1.55666053e+00
-6.05821200e-02 6.65150881e-02 -5.67281619e-02 1.08080363e+00
-1.94167420e-01 -1.54636645e+00 3.41333561e-02 9.45146859e-01
-5.39362371e-01 -4.79222745e-01 -4.20010835e-01 5.95663965e-01
3.87049198e-01 7.67414808e-01 2.73658007e-01 -7.35167503e-01
2.93550879e-01 3.91359687e-01 1.25051513e-01 -8.94817650e-01
-1.21905959e+00 -7.31281489e-02 3.35761666e-01 -7.93345392e-01
5.42325675e-02 -5.32808363e-01 -1.31973279e+00 -4.27114889e-02
-3.93502742e-01 1.94534153e-01 1.82950765e-01 8.05359423e-01
9.27702367e-01 9.36226606e-01 4.30301398e-01 -6.38161421e-01
-3.78207296e-01 -1.00393462e+00 -5.87663949e-01 1.91539064e-01
1.16200529e-01 -8.45277250e-01 3.69182229e-02 3.03280242e-02] | [9.264056205749512, 8.726468086242676] |
71fd6398-d3bd-4497-a1f0-fc0045238b22 | heavy-tailed-features-and-empirical-analysis | 1210.7215 | null | http://arxiv.org/abs/1210.7215v2 | http://arxiv.org/pdf/1210.7215v2.pdf | Heavy-Tailed Features and Empirical Analysis of the Limit Order Book Volume Profiles in Futures Markets | This paper poses a few fundamental questions regarding the attributes of the
volume profile of a Limit Order Books stochastic structure by taking into
consideration aspects of intraday and interday statistical features, the impact
of different exchange features and the impact of market participants in
different asset sectors. This paper aims to address the following questions:
1. Is there statistical evidence that heavy-tailed sub-exponential volume
profiles occur at different levels of the Limit Order Book on the bid and ask
and if so does this happen on intra or interday time scales ?
2.In futures exchanges, are heavy tail features exchange (CBOT, CME, EUREX,
SGX and COMEX) or asset class (government bonds, equities and precious metals)
dependent and do they happen on ultra-high (<1sec) or mid-range (1sec -10min)
high frequency data?
3.Does the presence of stochastic heavy-tailed volume profile features evolve
in a manner that would inform or be indicative of market participant behaviors,
such as high frequency algorithmic trading, quote stuffing and price discovery
intra-daily?
4. Is there statistical evidence for a need to consider dynamic behavior of
the parameters of models for Limit Order Book volume profiles on an intra-daily
time scale ?
Progress on aspects of each question is obtained via statistically rigorous
results to verify the empirical findings for an unprecedentedly large set of
futures market LOB data. The data comprises several exchanges, several futures
asset classes and all trading days of 2010, using market depth (Type II) order
book data to 5 levels on the bid and ask. | [] | 2015-04-22 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-6.19043469e-01 -4.07195091e-01 4.68944833e-02 -3.57734740e-01
-5.42389274e-01 -1.21143794e+00 1.01458466e+00 5.06584942e-01
-2.70886242e-01 6.67167068e-01 3.33619535e-01 -7.35820949e-01
-7.82022953e-01 -9.85545278e-01 -4.20007378e-01 -5.26089549e-01
-5.28263986e-01 8.79612148e-01 5.03710270e-01 -4.48282063e-01
5.92833579e-01 7.38731265e-01 -8.35007310e-01 1.61457993e-02
1.49074957e-01 1.28638792e+00 -4.65658009e-01 4.14886624e-01
-1.67075723e-01 6.13839030e-01 -4.89131629e-01 -2.86053628e-01
7.73332477e-01 -2.04274416e-01 -1.90147519e-01 1.44809708e-01
-2.09705740e-01 -3.79313856e-01 -3.18631530e-02 7.25028574e-01
1.30566210e-01 -1.00724638e-01 6.15079045e-01 -8.62635314e-01
-2.62570441e-01 7.83690810e-01 -6.14342868e-01 1.07707024e+00
-8.53498355e-02 3.29930604e-01 1.22643948e+00 -3.48163277e-01
2.43094608e-01 7.18643963e-01 5.12441278e-01 -5.07718086e-01
-1.26441133e+00 -4.98014122e-01 -4.79248948e-02 -3.22773159e-01
-8.48753750e-01 2.27046371e-01 4.21572983e-01 -6.79133296e-01
6.47724152e-01 4.09318507e-01 5.15734315e-01 6.73385262e-01
6.72107995e-01 -7.34503865e-02 1.27256298e+00 1.61331296e-02
2.44870618e-01 2.84219682e-01 3.42267364e-01 -5.69331467e-01
4.68174368e-01 5.05820692e-01 -3.15809071e-01 -7.29092240e-01
7.39111662e-01 -3.42617258e-02 -1.02557398e-01 2.83968925e-01
-8.54303002e-01 1.18495595e+00 -2.02920824e-01 5.84930778e-01
-6.55335605e-01 -3.20480585e-01 3.65966737e-01 9.23385024e-01
6.51798248e-01 2.51589268e-01 -1.25279951e+00 -4.97720420e-01
-8.11999679e-01 7.59938419e-01 1.38073087e+00 8.32446873e-01
4.02155221e-01 -5.97289018e-02 -1.08860508e-01 1.93882927e-01
2.12626338e-01 5.20222425e-01 3.72330904e-01 -7.79154301e-01
7.94253051e-01 3.57686803e-02 3.86689007e-01 -2.94236839e-01
-5.67141891e-01 -4.02060360e-01 -1.46926418e-01 3.26012343e-01
9.00708854e-01 -3.85919988e-01 -5.27212396e-04 1.31093383e+00
1.33810401e-01 -5.25396466e-01 -2.01133251e-01 3.74703556e-01
-2.10641362e-02 5.39863288e-01 -1.52928352e-01 -5.68713307e-01
1.64978099e+00 1.06679559e-01 -4.07685459e-01 3.57334673e-01
3.18336666e-01 -8.96186531e-01 6.19337916e-01 3.22223425e-01
-1.26899266e+00 -2.72013992e-01 -4.84503925e-01 6.27458751e-01
-2.00493619e-01 -8.68906379e-01 3.61147612e-01 5.70149183e-01
-3.54975313e-01 9.05008495e-01 -6.41054273e-01 1.64306760e-01
-4.03926373e-01 1.69088870e-01 2.82934785e-01 8.10830712e-01
-1.25298333e+00 6.50645077e-01 1.49403617e-01 -8.43424350e-02
-2.56883085e-01 -1.09148824e+00 -2.11680025e-01 2.73606718e-01
2.46666998e-01 2.49687228e-02 1.31937861e+00 -5.52474082e-01
-9.07365978e-01 2.53783852e-01 7.27533638e-01 -8.07818055e-01
8.39494109e-01 1.65933505e-01 -7.23428190e-01 -1.79809481e-02
-5.92446923e-02 -5.52474141e-01 3.55888873e-01 -6.95400476e-01
-9.02448654e-01 -7.16712058e-01 -1.99161410e-01 -1.58195436e-01
3.33263695e-01 3.21325362e-01 5.91131568e-01 -1.03533721e+00
1.27072886e-01 -8.09508920e-01 3.52999151e-01 -8.40172768e-01
-2.68972870e-02 -3.04489285e-01 5.85194945e-01 -5.69377184e-01
1.05469775e+00 -1.94859815e+00 -9.91401851e-01 5.34042716e-01
-2.39519402e-01 -5.08129895e-01 4.95759189e-01 1.19536221e+00
-2.78135240e-01 1.67821914e-01 -3.38811730e-03 5.90071857e-01
5.37159085e-01 5.75414300e-02 -7.64089704e-01 6.53612733e-01
-7.46543854e-02 6.22454524e-01 -1.45595163e-01 3.16793293e-01
-5.09803295e-02 -1.08297512e-01 -1.74605086e-01 2.46514287e-02
-2.67835230e-01 2.57994145e-01 -2.89890438e-01 3.61928403e-01
7.75673211e-01 -1.24010205e-01 -1.23029746e-01 1.27325550e-01
-8.95749211e-01 5.93967795e-01 -1.11012340e+00 2.48253301e-01
2.39941210e-01 4.50678200e-01 9.97460335e-02 -5.82926154e-01
9.41552281e-01 2.99042791e-01 3.70358407e-01 -1.05000603e+00
2.34308951e-02 6.29122734e-01 4.09505695e-01 -7.72617012e-02
4.15915579e-01 -6.88378096e-01 -2.65048333e-02 9.53053117e-01
-3.06060731e-01 -1.64755434e-01 3.03785503e-01 -1.33722663e-01
9.00184631e-01 -4.06655550e-01 -2.76562959e-01 -6.68089688e-01
-3.08543500e-02 -2.90701449e-01 4.45291936e-01 5.24206042e-01
1.21318556e-01 4.65349197e-01 1.20227885e+00 -3.28702569e-01
-1.11051750e+00 -1.18632340e+00 -8.25787187e-01 7.65187144e-01
-2.16601282e-01 6.69427142e-02 -1.38910607e-01 -1.28931999e-01
6.98608994e-01 9.70554531e-01 -6.64254546e-01 2.08072782e-01
-3.75820458e-01 -1.25845051e+00 -1.24318145e-01 2.62924016e-01
4.17298228e-01 -1.03143620e+00 -7.91961670e-01 4.19637382e-01
5.67036927e-01 -9.44459558e-01 -5.80229580e-01 4.79735762e-01
-1.11327815e+00 -9.42456067e-01 -6.12791240e-01 -1.52332084e-02
-1.76430233e-02 -4.48074967e-01 1.24919379e+00 -4.14457262e-01
1.35658607e-01 5.35379946e-01 -2.00033858e-01 -8.43424976e-01
-2.59955525e-01 -1.90925747e-01 -2.41832927e-01 3.20401669e-01
5.78194141e-01 -6.30204082e-01 -8.29304814e-01 4.44688886e-01
-1.00507367e+00 -1.12342334e+00 2.44520470e-01 2.01501489e-01
2.59141594e-01 4.56035525e-01 7.29944050e-01 -8.26735198e-01
8.36783588e-01 -7.00332999e-01 -1.28465128e+00 5.16316369e-02
-1.22744238e+00 -5.94543703e-02 2.25807622e-01 -2.73412347e-01
-1.00783896e+00 -7.76108384e-01 3.00788999e-01 9.11327004e-02
7.42551833e-02 6.10314369e-01 3.74789625e-01 2.93098032e-01
4.42027524e-02 2.55278200e-01 2.00389266e-01 -1.00229800e+00
-3.19371939e-01 4.63564485e-01 3.40134442e-01 -5.97692430e-01
1.01551020e+00 5.92382610e-01 -6.59013540e-02 -7.17437446e-01
-3.69124562e-01 -3.35493118e-01 -3.60714376e-01 3.21640462e-01
6.00995183e-01 -6.90416157e-01 -9.62261200e-01 5.50955951e-01
-5.02668500e-01 -5.39879978e-01 -6.18550479e-01 7.90697932e-01
-4.23905909e-01 1.98628977e-02 -1.18197322e+00 -1.12113047e+00
-2.00347394e-01 -5.81969380e-01 3.83139342e-01 2.24320322e-01
-5.67309618e-01 -1.29373229e+00 4.42393631e-01 1.13709942e-01
5.22529006e-01 5.12925506e-01 1.22411501e+00 -1.53786778e+00
-6.25544429e-01 -4.05082554e-01 -4.90029529e-02 1.98228627e-01
2.03979567e-01 2.04739764e-01 -3.70937884e-01 -3.56130540e-01
6.51971400e-01 1.90717950e-01 3.20983231e-01 5.59879422e-01
1.47296160e-01 -3.38582158e-01 3.13958734e-01 2.17591003e-01
1.26232493e+00 6.89047515e-01 1.88096732e-01 8.27290356e-01
-3.19876671e-01 7.57409096e-01 5.35110235e-01 9.26893830e-01
3.37678522e-01 3.07139486e-01 2.54813254e-01 3.75757158e-01
7.71770358e-01 1.03346467e-01 3.22629362e-01 4.05757010e-01
2.47704741e-02 2.51969904e-01 -7.98720181e-01 6.76859021e-01
-1.12335014e+00 -1.09216464e+00 -5.61285853e-01 2.45125961e+00
7.05081582e-01 5.97835541e-01 8.49237740e-01 -1.55547544e-01
4.03559208e-01 2.96480536e-01 -4.77814257e-01 -5.30105233e-01
-2.46787593e-01 1.35914549e-01 9.05392826e-01 3.31499189e-01
-4.29748595e-01 -1.13661729e-01 6.05718136e+00 3.31424892e-01
-9.13398445e-01 -2.52202898e-01 9.21336532e-01 -2.24263147e-01
-6.60728335e-01 1.93731517e-01 -1.06972539e+00 1.07777560e+00
1.31052840e+00 -6.48430228e-01 1.79209024e-01 1.83978766e-01
4.32058811e-01 -3.67273867e-01 -1.04350114e+00 4.20041412e-01
-7.20146954e-01 -1.16039479e+00 -4.92017746e-01 9.42552865e-01
6.37199759e-01 1.76849127e-01 2.88743973e-01 2.65540063e-01
1.34380221e-01 -4.89074409e-01 1.01841545e+00 5.83925903e-01
5.24028428e-02 -1.02487946e+00 8.11518550e-01 2.78721184e-01
-9.55267787e-01 -3.76016125e-02 -1.49080276e-01 3.42488592e-03
3.82906497e-01 8.15368652e-01 -3.33779424e-01 4.91215885e-01
8.02833557e-01 -8.44262466e-02 -2.42202282e-01 8.06167305e-01
5.39657712e-01 8.24568868e-01 -7.17733443e-01 5.26983254e-02
5.60961485e-01 -8.20199192e-01 4.65561450e-01 6.26049042e-01
4.11818653e-01 3.61479461e-01 -5.45549572e-01 1.07092154e+00
2.35220477e-01 -2.58383244e-01 -3.05162102e-01 -3.01077992e-01
3.88367593e-01 9.80260253e-01 -8.44254851e-01 -1.46892364e-03
-8.10115993e-01 2.16188192e-01 -6.57095551e-01 3.20398808e-01
-6.20086432e-01 -3.31969202e-01 4.91310269e-01 9.28773403e-01
7.12181926e-01 -2.83295691e-01 -5.00231922e-01 -8.02349925e-01
3.22869360e-01 -7.43080318e-01 7.95855403e-01 -1.10095546e-01
-1.58872628e+00 1.90098137e-01 1.18528560e-01 -1.12664616e+00
-5.73201776e-01 -4.25258785e-01 -1.05316317e+00 1.06338608e+00
-1.42086732e+00 -4.96426746e-02 3.99744570e-01 7.25892901e-01
8.71027261e-02 -3.74981254e-01 1.02445602e-01 -1.28172278e-01
2.39004537e-01 1.04798302e-01 7.52943158e-01 2.82708466e-01
2.00851232e-01 -1.56643772e+00 6.68786407e-01 3.58325318e-02
1.92196935e-01 5.32197297e-01 8.33471417e-01 -1.02298331e+00
-8.55809689e-01 -5.19173443e-01 8.03101540e-01 -7.07311571e-01
1.43439984e+00 -1.51376158e-01 -1.05627275e+00 7.76124239e-01
4.14017916e-01 -2.20790148e-01 6.91598535e-01 -6.56593293e-02
4.44134995e-02 -1.87806547e-01 -1.12593365e+00 2.78281756e-02
1.34038001e-01 -5.51055729e-01 -8.70121956e-01 4.22156096e-01
4.43332106e-01 6.99436814e-02 -1.46712852e+00 4.28285033e-01
6.19071186e-01 -1.33623469e+00 4.70215410e-01 -4.25099283e-01
-1.76682517e-01 9.56290215e-02 -2.69225597e-01 -1.03710973e+00
-1.57090962e-01 -1.05460978e+00 1.33802682e-01 1.38805449e+00
6.21818662e-01 -1.36778450e+00 4.47318256e-01 1.13504457e+00
5.12802422e-01 -6.00494683e-01 -1.27230930e+00 -1.16339302e+00
6.75906241e-01 -8.28721970e-02 8.81645977e-01 6.79151833e-01
-2.37784594e-01 3.44483294e-02 2.12742358e-01 -6.55971020e-02
7.60787964e-01 7.84788907e-01 3.14534783e-01 -1.37763846e+00
-4.75477278e-01 -4.86666203e-01 -2.04974152e-02 -4.97112364e-01
-4.17598814e-01 -4.55969900e-01 -7.31438816e-01 -6.07968330e-01
-9.27248970e-02 -2.46618226e-01 -2.87443995e-01 -4.86618340e-01
4.39590633e-01 -4.80106443e-01 3.39622974e-01 4.44300264e-01
4.94859330e-02 4.11189854e-01 7.54043579e-01 2.94442475e-01
-8.09524283e-02 6.42983556e-01 -3.38137150e-01 3.12845737e-01
7.20751464e-01 -1.96205303e-01 -1.46658808e-01 3.49086702e-01
3.96728426e-01 6.92669272e-01 3.06974024e-01 -5.99385083e-01
-6.56821057e-02 -2.13476419e-01 3.04452002e-01 -1.03381693e+00
-1.36991292e-01 -8.15987945e-01 4.30729955e-01 5.27822137e-01
-2.33333021e-01 7.84460008e-01 3.57243046e-02 6.19673610e-01
-2.84381151e-01 -3.51229817e-01 5.04552722e-01 -3.72477829e-01
2.07319260e-01 3.62581223e-01 -5.40705860e-01 4.91504848e-01
1.18634582e+00 -1.15511082e-01 -1.39639050e-01 -6.93165481e-01
-6.98301554e-01 4.00555223e-01 4.14973795e-01 4.01471198e-01
-3.32718372e-01 -1.12793946e+00 -9.03096795e-01 1.70424253e-01
-3.76833349e-01 -4.69840050e-01 -4.45396453e-03 1.07364297e+00
-3.67843300e-01 5.87693274e-01 1.65991887e-01 -2.53545523e-01
-5.86915374e-01 2.64764726e-01 5.36870480e-01 -3.16335678e-01
-6.23809695e-01 5.61858714e-01 2.22441509e-01 -6.04800209e-02
-1.70970976e-01 -5.06597936e-01 1.43788546e-01 9.37657356e-01
5.61777711e-01 6.48216784e-01 1.89157531e-01 -3.97543669e-01
-4.55672070e-02 4.60520089e-01 -2.21200436e-01 -4.53368247e-01
1.70337224e+00 -1.84637457e-01 -2.23888844e-01 1.15658927e+00
1.13576007e+00 1.10947445e-01 -1.49674821e+00 -1.15311388e-02
9.06908810e-01 -2.13794589e-01 -4.66445357e-01 -6.11002684e-01
-1.05292892e+00 4.84518945e-01 3.28387797e-01 1.31874084e+00
4.04534519e-01 3.10223013e-01 9.07101810e-01 -1.92070663e-01
4.26942110e-01 -1.47917092e+00 -4.62137848e-01 4.49911982e-01
1.16265965e+00 -7.75917649e-01 -9.33960825e-02 3.25725555e-01
-5.54819465e-01 1.35275900e+00 -1.50444508e-01 -4.40569401e-01
1.31132698e+00 3.84710163e-01 6.30625412e-02 -5.22294700e-01
-1.01989996e+00 3.15514475e-01 5.73319802e-03 -2.49455959e-01
2.59254038e-01 1.57527417e-01 -7.38735199e-01 6.34674907e-01
-6.17975235e-01 -4.18819606e-01 7.95514107e-01 9.33773696e-01
-3.44159424e-01 -9.66510952e-01 -4.11773771e-01 7.61292279e-01
-1.11093378e+00 -4.56723478e-03 -3.50400537e-01 1.25734615e+00
-3.58528972e-01 5.36827028e-01 8.35427642e-01 1.63608879e-01
4.94352162e-01 2.20701247e-01 1.10935234e-02 -3.59008610e-01
-1.10152364e+00 5.09327233e-01 8.28728527e-02 -2.28679270e-01
1.74010992e-01 -1.42827976e+00 -1.17719710e+00 -5.85874438e-01
-3.10093105e-01 5.11472702e-01 3.32837015e-01 9.41244721e-01
-3.05688381e-02 -1.63575113e-01 1.01193774e+00 -4.22251523e-01
-1.51381624e+00 -1.05980170e+00 -1.60103643e+00 3.83330971e-01
5.59747994e-01 -2.60718971e-01 -1.15960383e+00 -4.43814576e-01] | [4.737300395965576, 4.0869460105896] |
9c76f119-a604-4b5a-adbd-6cb95112d4c8 | social-media-analysis-for-organizations-us | 1803.09133 | null | http://arxiv.org/abs/1803.09133v1 | http://arxiv.org/pdf/1803.09133v1.pdf | Social Media Analysis For Organizations: Us Northeastern Public And State Libraries Case Study | Social networking sites such as Twitter have provided a great opportunity for
organizations such as public libraries to disseminate information for public
relations purposes. However, there is a need to analyze vast amounts of social
media data. This study presents a computational approach to explore the content
of tweets posted by nine public libraries in the northeastern United States of
America. In December 2017, this study extracted more than 19,000 tweets from
the Twitter accounts of seven state libraries and two urban public libraries.
Computational methods were applied to collect the tweets and discover
meaningful themes. This paper shows how the libraries have used Twitter to
represent their services and provides a starting point for different
organizations to evaluate the themes of their public tweets. | ['Matthew Collins', 'Amir Karami'] | 2018-03-24 | null | null | null | null | ['public-relations'] | ['miscellaneous'] | [-4.87726331e-01 2.17327908e-01 -4.06435311e-01 -2.18786672e-02
-9.00171220e-01 -6.38248026e-01 9.16317165e-01 1.02668345e+00
-4.77088541e-01 8.72151554e-01 8.96891415e-01 -4.11816597e-01
2.89843440e-01 -1.10755241e+00 -3.14868763e-02 -3.09422910e-01
-6.37151077e-02 3.80827896e-02 5.96827976e-02 -5.10353684e-01
8.69185984e-01 4.98386562e-01 -1.24920678e+00 3.20745111e-01
5.65016031e-01 6.40480399e-01 1.71531975e-01 2.74390787e-01
-9.93392587e-01 7.69231558e-01 -6.96709156e-01 -1.62251994e-01
-1.28990263e-01 -8.70523006e-02 -7.42238700e-01 8.35024342e-02
-2.06336305e-01 5.65550774e-02 -2.14341283e-01 9.13250387e-01
4.53301638e-01 -2.59515364e-02 2.50149909e-02 -1.10376465e+00
-7.88411573e-02 1.04507577e+00 -3.46754730e-01 4.31557417e-01
7.90876210e-01 4.99866754e-02 9.25812840e-01 -6.90684557e-01
1.21588981e+00 1.09019578e+00 4.14575428e-01 -6.09306633e-01
-6.86182797e-01 -9.38038647e-01 -2.00195178e-01 -2.33905137e-01
-1.55471182e+00 -2.04841971e-01 5.98599970e-01 -6.94885254e-01
6.15952551e-01 3.93445611e-01 1.02233136e+00 7.85398543e-01
3.28207552e-01 4.59998012e-01 1.60864758e+00 -3.25984389e-01
-2.81415712e-02 5.84060311e-01 5.42963326e-01 2.14259237e-01
1.65141195e-01 -9.80911136e-01 -6.29890144e-01 -7.66268611e-01
-1.32152019e-02 1.97662964e-01 1.36769131e-01 1.02252567e+00
-1.42686939e+00 1.21945679e+00 4.20322657e-01 8.83904099e-01
-6.60098612e-01 -2.78940886e-01 5.24467945e-01 -1.00761242e-01
9.93002236e-01 4.49807823e-01 2.77685702e-01 -5.48936129e-01
-5.74581802e-01 2.13733315e-01 1.11465716e+00 7.34723866e-01
1.08668184e+00 -3.65356505e-01 1.53793871e-01 4.64782268e-01
6.38178766e-01 4.36921328e-01 3.61454189e-01 -3.80059987e-01
4.08850193e-01 1.13358331e+00 1.83119953e-01 -1.74291492e+00
-4.78135586e-01 -2.92825043e-01 -1.24433830e-01 -7.47439027e-01
2.26902142e-01 -2.87517428e-01 -2.71479219e-01 6.52634442e-01
2.96356708e-01 -2.66099811e-01 -2.43173063e-01 5.60310841e-01
1.17935622e+00 8.25674951e-01 1.73656374e-01 -2.47677535e-01
1.34394801e+00 -2.18374655e-01 -1.19222271e+00 2.06141204e-01
4.14869130e-01 -1.26168334e+00 9.17742431e-01 -1.52075097e-01
-9.45001841e-01 6.32586852e-02 -5.47789276e-01 1.47585422e-01
-9.93939757e-01 -5.78073502e-01 5.31853020e-01 6.16526425e-01
-6.80201173e-01 2.07240626e-01 -5.54605305e-01 -9.68076408e-01
4.45154250e-01 2.30801199e-02 -4.90567051e-02 1.61697254e-01
-1.07704449e+00 6.05374098e-01 3.50142181e-01 -3.15067768e-01
-1.17182314e-01 -2.45872498e-01 -5.03117681e-01 -3.35600406e-01
1.82142891e-02 2.03381762e-01 8.01393509e-01 -3.68517220e-01
-7.85693705e-01 1.05178189e+00 -3.06513488e-01 -7.38319606e-02
4.39848095e-01 2.11012110e-01 -5.48927605e-01 1.22154452e-01
5.77938080e-01 -2.07282022e-01 -3.87128770e-01 -1.07048869e+00
-7.18653381e-01 -3.03361535e-01 -1.98181480e-01 -5.03567040e-01
-9.38681602e-01 8.90979350e-01 -1.18782096e-01 -3.80665064e-01
4.39904183e-01 -9.41266835e-01 -1.74266979e-01 -9.66592669e-01
-7.61530221e-01 -2.86225021e-01 9.22076762e-01 -6.46826982e-01
1.51126015e+00 -2.22154903e+00 -8.18478823e-01 5.42612195e-01
3.09676945e-01 -6.17182672e-01 7.77920425e-01 1.37704778e+00
5.62396109e-01 7.06084669e-01 2.04730377e-01 -1.94023296e-01
-2.53300276e-02 1.41157433e-01 -3.95175844e-01 8.48323286e-01
-3.53550404e-01 4.29524660e-01 -1.16077495e+00 -6.64625168e-01
-3.06019336e-01 3.11779886e-01 2.31799837e-02 -2.47084111e-01
1.69606283e-01 4.52715605e-01 -8.66661668e-01 6.58661902e-01
5.33076346e-01 -5.09610295e-01 1.79409921e-01 2.74125725e-01
-1.25577092e+00 6.11400425e-01 -7.42764831e-01 9.03293431e-01
-8.09089914e-02 1.50674593e+00 -1.07169617e-02 -1.33834258e-01
1.15817702e+00 1.09255224e-01 8.35227251e-01 -4.88709033e-01
6.07987761e-01 2.97673851e-01 -2.26225927e-01 -4.33664829e-01
1.16481888e+00 4.13001254e-02 -5.81033468e-01 1.08681631e+00
-7.72020221e-01 9.63495895e-02 1.91013217e-01 4.29311246e-01
7.29309499e-01 -5.43208838e-01 1.68020234e-01 -5.62244773e-01
1.80888772e-01 4.99051660e-01 1.96499288e-01 3.39802682e-01
-1.66163549e-01 1.70396373e-01 3.34320486e-01 -5.05174696e-01
-1.00283527e+00 -2.60557026e-01 -4.92006540e-01 9.55949426e-01
2.08449773e-02 -8.61151934e-01 -2.77559072e-01 2.42089748e-01
-5.90882599e-02 6.00233495e-01 -2.43965611e-01 7.07839191e-01
-3.50034177e-01 -8.77362430e-01 2.39267915e-01 -4.02964026e-01
7.98472345e-01 -8.35222542e-01 -4.62430328e-01 3.85891825e-01
-8.54691565e-01 -1.43115020e+00 1.01294570e-01 -2.29966864e-01
-3.66031229e-01 -8.93071294e-01 -2.54654139e-01 -5.13173401e-01
7.63770103e-01 5.62457144e-01 6.35018468e-01 -1.28518611e-01
-1.79106250e-01 2.33129978e-01 -4.39115852e-01 -9.84550655e-01
-4.07523245e-01 5.92737019e-01 -1.90618500e-01 1.66500911e-01
6.34212375e-01 -2.62644172e-01 -2.57670581e-01 3.54889005e-01
-8.15855503e-01 -8.68630558e-02 -3.06671143e-01 -1.43326968e-01
9.47246850e-02 8.04401562e-02 6.28758192e-01 -9.96713758e-01
9.60998297e-01 -1.36091053e+00 -2.55635381e-01 -4.18585837e-01
-3.82402062e-01 -6.56774700e-01 1.54618025e-01 4.58572321e-02
-6.42260671e-01 -3.99341762e-01 1.90697402e-01 6.06612802e-01
2.43384331e-01 1.19334924e+00 4.41514611e-01 -6.56617284e-02
4.89306450e-01 -6.75414279e-02 3.21154967e-02 -2.39490092e-01
-3.86172146e-01 1.24143183e+00 -1.86630279e-01 -3.15025508e-01
8.37033689e-01 8.16033244e-01 -3.08491051e-01 -1.24200785e+00
-6.32894278e-01 -8.79737675e-01 -3.48289341e-01 -9.86287951e-01
7.76807606e-01 -9.19152200e-01 -1.18642426e+00 1.34291038e-01
-8.16685438e-01 1.19878180e-01 -5.02497169e-05 5.70913374e-01
1.31368920e-01 -2.45019551e-02 -5.14190137e-01 -7.61524796e-01
-5.21761226e-03 -7.33347416e-01 3.81252199e-01 2.90220410e-01
-6.35586202e-01 -1.06450438e+00 -1.69055000e-01 4.32326257e-01
8.75152588e-01 8.99615169e-01 4.07016844e-01 -9.77107286e-01
-6.29579544e-01 -4.43366438e-01 -4.06299859e-01 -5.10873139e-01
3.82055610e-01 4.45471793e-01 -9.60983694e-01 -1.83363318e-01
-1.56360984e-01 5.88529557e-02 9.02674124e-02 1.03536276e-02
9.77723837e-01 -8.87564659e-01 -4.24469799e-01 -1.15766838e-01
1.42468262e+00 1.02996238e-01 5.92174053e-01 1.51486862e+00
2.74348408e-01 6.49509311e-01 5.17795503e-01 1.35998464e+00
6.94443762e-01 -8.03692788e-02 2.57310838e-01 2.69391924e-01
5.88945270e-01 -3.55306901e-02 2.54260182e-01 1.39351404e+00
-4.23360914e-01 1.09601751e-01 -1.66127908e+00 5.79268813e-01
-1.53545702e+00 -1.02230978e+00 -7.84338892e-01 1.84865057e+00
8.15091372e-01 2.51257926e-01 2.30028927e-01 4.93110158e-03
7.93421805e-01 3.34228396e-01 1.88582465e-01 -1.81354940e-01
-2.60657489e-01 -2.69996196e-01 8.39441955e-01 3.38917732e-01
-6.94062352e-01 4.15058643e-01 6.66862297e+00 2.28454888e-01
-1.27270758e+00 2.30984226e-01 7.62987971e-01 9.75362360e-02
-6.15229845e-01 2.88109541e-01 -1.15141213e+00 6.10645235e-01
1.24582887e+00 -1.02853823e+00 -1.53985217e-01 6.91303968e-01
8.08724284e-01 -5.34946382e-01 -1.21949397e-01 3.93534273e-01
-1.25510227e-02 -1.78037703e+00 -3.18158060e-01 6.79202557e-01
1.05397224e+00 3.24274868e-01 -1.70698181e-01 -1.58660617e-02
2.09758893e-01 -6.06096089e-01 5.17047346e-01 7.03704655e-01
2.35744387e-01 -6.53133869e-01 7.24701941e-01 3.86135608e-01
-8.95449877e-01 -2.19997559e-02 -3.12714577e-02 -5.43708026e-01
2.99104750e-01 6.73912108e-01 -1.45194149e+00 2.47146770e-01
6.59434915e-01 5.65934420e-01 -4.63972032e-01 1.08705902e+00
2.31642216e-01 7.23580658e-01 -1.72065228e-01 -6.58969462e-01
8.11802149e-01 -1.69973463e-01 6.64626956e-01 1.55484331e+00
3.59099567e-01 -2.17029694e-02 5.45841932e-01 2.97220737e-01
-2.77824908e-01 5.78316331e-01 -6.25831425e-01 -5.81826866e-01
7.77370751e-01 1.24870718e+00 -1.38831353e+00 -2.97777534e-01
-4.61180210e-01 -1.43363789e-01 -2.24907354e-01 2.30449587e-01
-6.04967117e-01 -5.30000627e-01 9.55074728e-02 1.04914796e+00
-5.61809599e-01 -8.25389922e-01 -2.96908200e-01 -6.33414209e-01
-5.21872878e-01 -4.99353915e-01 9.76586491e-02 -4.76051778e-01
-1.11730468e+00 4.37380940e-01 7.52062201e-02 -1.04164052e+00
1.51208311e-01 2.79497392e-02 -5.41981995e-01 8.20660412e-01
-1.49809599e+00 -1.06524765e+00 -8.80617201e-01 4.16973889e-01
1.57159328e-01 -1.57365650e-01 5.31151116e-01 5.26485384e-01
-4.80044127e-01 -1.73470497e-01 3.77105117e-01 3.26553166e-01
6.51446760e-01 -7.29767859e-01 3.65873729e-03 3.08151633e-01
-8.50376129e-01 9.68022406e-01 7.88080990e-01 -7.37330854e-01
-1.70830083e+00 -9.44307625e-01 1.71476948e+00 1.71806826e-03
1.44067693e+00 -1.48209885e-01 -4.98448104e-01 8.62779438e-01
8.70709658e-01 -4.74672288e-01 1.40483999e+00 1.67604864e-01
3.75282019e-01 2.83378392e-01 -1.06787944e+00 6.96732342e-01
3.13056409e-01 -7.45440066e-01 -3.49070907e-01 9.58928585e-01
3.95650506e-01 -5.66626973e-02 -1.14503467e+00 -6.06842518e-01
4.23572451e-01 -5.86631060e-01 5.55873871e-01 -2.70281971e-01
5.50577283e-01 8.22593048e-02 -1.76358163e-01 -9.12127078e-01
-1.12097614e-01 -1.13175440e+00 6.60254061e-01 1.49361634e+00
3.23176861e-01 -1.19727910e+00 4.76821899e-01 1.14685047e+00
-1.41027361e-01 -1.87984481e-01 -5.39716423e-01 -2.57756948e-01
-2.57937819e-01 -4.80991334e-01 5.45161188e-01 1.52780414e+00
3.04760724e-01 -3.79517004e-02 3.45966637e-01 -3.26599717e-01
5.31676590e-01 2.08695799e-01 1.15613258e+00 -1.38831067e+00
6.89238489e-01 -4.95410293e-01 -4.23764616e-01 -3.14221770e-01
7.56733492e-02 -9.28123415e-01 -5.42560935e-01 -1.78685665e+00
4.84412491e-01 -6.26366913e-01 1.64875150e-01 2.25140065e-01
3.41637790e-01 3.29174429e-01 2.15044349e-01 6.94760323e-01
-3.14137936e-01 -4.18852121e-02 1.27221048e+00 3.21703553e-02
-3.26305658e-01 -1.54690668e-01 -1.02725601e+00 6.89978242e-01
8.30370486e-01 -4.69532907e-01 5.57361186e-01 1.60954863e-01
1.21959472e+00 -3.82744104e-01 -1.09129632e-02 -5.93264222e-01
6.12526655e-01 -3.31321895e-01 1.00401394e-01 -9.37066913e-01
-1.17218181e-01 -6.75711632e-01 3.02486509e-01 3.98655385e-01
-2.83190072e-01 2.97086656e-01 2.85564631e-01 -5.60029931e-02
-5.89554548e-01 -1.55648002e-02 3.58782500e-01 -3.15613449e-01
-4.03288722e-01 -1.10359319e-01 -1.13980663e+00 -1.06831737e-01
1.06176579e+00 -3.17037731e-01 -6.99020922e-01 -6.31403506e-01
-7.67138839e-01 1.43407375e-01 4.04925078e-01 -4.11817506e-02
2.97291040e-01 -1.31594145e+00 -8.92056167e-01 -3.13715816e-01
1.98090270e-01 -2.99370706e-01 -2.60658234e-01 9.60611045e-01
-6.41407907e-01 4.06824231e-01 -3.79589647e-01 -3.13505352e-01
-1.06096303e+00 -3.14697385e-01 -4.57265586e-01 1.10276297e-01
-6.39635086e-01 2.49346837e-01 -6.30069911e-01 -1.18783623e-01
-7.95897767e-02 -8.32764804e-02 -4.93178993e-01 1.18650627e+00
5.73180258e-01 6.49274349e-01 -1.97996274e-01 -1.01366174e+00
-1.18099019e-01 2.29714829e-02 2.39721641e-01 -2.25363016e-01
1.73320484e+00 -5.49784124e-01 -6.33981764e-01 1.18413341e+00
1.48183501e+00 4.42782551e-01 -1.28864899e-01 -1.85109094e-01
4.52260345e-01 -1.15804613e+00 1.86283901e-01 -7.09543824e-02
-5.18927395e-01 2.06224188e-01 -1.20206214e-01 1.06439304e+00
7.00197518e-01 1.53316125e-01 8.48234117e-01 1.36645332e-01
3.35532874e-01 -1.41988099e+00 -8.96511078e-02 9.39092636e-01
7.79784679e-01 -1.15995383e+00 2.84947187e-01 -4.17840898e-01
-3.84416312e-01 1.48767507e+00 -1.96889967e-01 2.30503187e-01
9.85803068e-01 2.40452304e-01 -1.23945773e-01 -5.10935962e-01
-2.24636793e-01 -1.86464787e-01 -3.58272135e-01 -2.76550412e-01
9.42111671e-01 1.01684749e-01 -9.51878607e-01 3.83549333e-01
-5.66142201e-01 -2.24019587e-01 1.27896166e+00 1.51030707e+00
-8.35736632e-01 -8.28507662e-01 -8.74114573e-01 5.14168262e-01
-1.13306034e+00 1.27124533e-01 -5.65822899e-01 9.92715061e-01
-3.92401785e-01 9.42103088e-01 2.70021200e-01 -1.93838656e-01
2.24518101e-03 4.27162042e-04 -4.92210567e-01 -5.80873907e-01
-1.13769186e+00 2.19099939e-01 5.61519504e-01 -6.63685724e-02
-6.50115490e-01 -1.10718501e+00 -1.71130729e+00 -1.05268407e+00
-1.02830157e-01 5.82152247e-01 1.44938314e+00 5.73290944e-01
3.71254355e-01 3.38286608e-01 1.15070212e+00 -6.53216064e-01
4.05503392e-01 -1.01637566e+00 -5.85076272e-01 3.59683126e-01
1.47837818e-01 -2.78798789e-01 -5.04138410e-01 -7.26967379e-02] | [10.531303405761719, 7.093046188354492] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.